Theo Phys I Math Methods PDF

Theoretical Physics I-Mathematical Methods
Daniel Erenso and Victor Montemayor
February 25, 2019

ii
Contents
Introduction xi
I Mathematical Methods in Physics I 1

1 Series and Convergence 3
1.1 Sequence and series . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Testing series for convergence . . . . . . . . . . . . . . . . . . . . 7
1.3 Series representations of real functions . . . . . . . . . . . . . . . 9
1.4 Homework Assignment 1 . . . . . . . . . . . . . . . . . . . . . . . 18
2 Complex Numbers, functions, and series 21

2.1 Complex numbers . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2 Complex In…nite Series and the Circle of Convergence . . . . . . 28
2.3 Powers and Roots of complex numbers . . . . . . . . . . . . . . . 32
2.4 Algebraic vs transcendental functions . . . . . . . . . . . . . . . . 34
3 Vectors, Lines, and Planes 45

3.1 An Overview of Vector Fundamentals . . . . . . . . . . . . . . . 45
3.2 Vector Multiplication . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.3 Equations of a straight line . . . . . . . . . . . . . . . . . . . . . 49
3.4 Equation of a plane . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4 Matrix and determinants 55

4.1 Important Terminologies . . . . . . . . . . . . . . . . . . . . . . . 55
4.2 Matrix Arithmetic and Manipulation . . . . . . . . . . . . . . . 56
4.3 Matrix representation of a set of linear equations . . . . . . . . . 57
4.4 Matrix application: solving a set of linear equations . . . . . . . 58
4.5 Determinant of a square matrix . . . . . . . . . . . . . . . . . . . 63
iii
iv CONTENTS
4.6 Cramer’s Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

4.7 The adjoint and inverse of a matrix . . . . . . . . . . . . . . . . . 67
4.8 Orthogonal matrices and the rotation matrix . . . . . . . . . . . 72
4.9 Linear independence . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.10 Gram-Schmidt orthogonalization . . . . . . . . . . . . . . . . . . 78
5 Introduction to di¤erential calculus-I 85

5.1 Partial di¤erentiation . . . . . . . . . . . . . . . . . . . . . . . . 85
5.2 Total di¤erential . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
5.3 The multivariable form of the Chain rule . . . . . . . . . . . . . . 90
5.4 Extremum (Max/Min) problems . . . . . . . . . . . . . . . . . . 94
5.5 The Method of Lagrangian Multipliers . . . . . . . . . . . . . . . 97
5.6 Change of Variables . . . . . . . . . . . . . . . . . . . . . . . . . 101
5.7 Legendre Transformations . . . . . . . . . . . . . . . . . . . . . . 106
5.8 Homework Assignment 10 . . . . . . . . . . . . . . . . . . . . . . 107
6 Introduction to Di¤erential Calculus II 109

6.1 Linear ordinary di¤erential equations . . . . . . . . . . . . . . . . 109
6.1.1 First-order LODE with constant coe¢ cients . . . . . . . . 110
6.2 The …rst-order DE and exact di¤erential . . . . . . . . . . . . . . 117
6.3 First-order DE and none exact total di¤erential . . . . . . . . . . 122
6.4 Higher-Order ODE’s . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.4.1 Homogeneous ODE’s with constant coe¢ cient . . . . . . . 127
6.4.2 None homogeneous ODE’s with Constant Coe¢ cients . . 137
6.5 The method of superposition and the particular solution. . . . . 141
6.6 The method of successive integration . . . . . . . . . . . . . . . . 145
6.7 Partial Di¤erential Equations . . . . . . . . . . . . . . . . . . . . 154
7 Integral calculus-scalar functions 163

7.1 Integration in Cartesian coordinates . . . . . . . . . . . . . . . . 163
7.1.1 Single and multiple integrals . . . . . . . . . . . . . . . . 163
7.1.2 Physical Applications . . . . . . . . . . . . . . . . . . . . 171
7.2 Integration in curvilinear coordinates . . . . . . . . . . . . . . . . 183
7.2.1 One- and two–dimensional curvilinear Coordinates . . . . 184
7.2.2 3-D Curvilinear coordinates-cylindrical . . . . . . . . . . 186
7.2.3 3-D Curvilinear Coordinate Systems: spherical Coordinates 188
CONTENTS v
8 Vector Calculus 195

8.1 Vector Products . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
8.1.1 Physical applications . . . . . . . . . . . . . . . . . . . . . 196
8.1.2 Vectors derivatives . . . . . . . . . . . . . . . . . . . . . . 198
8.2 Di¤erential Operators . . . . . . . . . . . . . . . . . . . . . . . . 202
8.2.1 The gradient operator and directional derivative . . . . . 202
8.2.2 The Divergence, the Curl, and the Laplacian . . . . . . . 206
8.3 Single vector function integrals . . . . . . . . . . . . . . . . . . . 217
8.3.1 Line vector integrals . . . . . . . . . . . . . . . . . . . . . 217
8.3.2 Conservative vector …elds . . . . . . . . . . . . . . . . . . 219
8.3.3 Conservative Fields and Exact Di¤erentials . . . . . . . . 221
8.4 Multiple vector integrals . . . . . . . . . . . . . . . . . . . . . . . 222
8.4.1 Green’s Theorem . . . . . . . . . . . . . . . . . . . . . . . 222
8.4.2 The Stokes’Theorem . . . . . . . . . . . . . . . . . . . . 226
8.4.3 The Divergence Theorem . . . . . . . . . . . . . . . . . . 231
8.4.4 More Examples on Divergence and Stokes theorems . . . 233
8.4.5 Applications of the Divergence and Stoke’s theorems . . . 238
II Mathematical Methods in Physics II 241

9 Introduction to the Calculus of Variations 243
9.1 Geodesic and stationary points . . . . . . . . . . . . . . . . . . . 243
9.2 The general problem . . . . . . . . . . . . . . . . . . . . . . . . 247
9.3 Appplications: The Brachystochrone Problem . . . . . . . . . . . 250
9.4 Applications: classical mechanics . . . . . . . . . . . . . . . . . . 253
9.5 Physical application of the Euler-Lagrange equation . . . . . . . 254
10 Introduction to the Eigenvalue Problem 261

10.1 Matrix Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
10.2 Orthogonal matrices and the rotational operator . . . . . . . . . 264
10.3 Eigenvalues and Eigenvectors . . . . . . . . . . . . . . . . . . . . 265
10.4 Physical applications . . . . . . . . . . . . . . . . . . . . . . . . . 270
11 Special functions 279

11.1 The factorial and gamma function . . . . . . . . . . . . . . . . . 279
11.2 The Beta Function . . . . . . . . . . . . . . . . . . . . . . . . . . 283
11.3 Stirling’s Formula . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
11.4 The Error Function . . . . . . . . . . . . . . . . . . . . . . . . . . 288
11.5 Elliptic Integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
11.6 The Dirac delta function . . . . . . . . . . . . . . . . . . . . . . . 294
vi CONTENTS
12 Power Series Solutions to Di¤erential Equations 303

12.1 Power series substitution . . . . . . . . . . . . . . . . . . . . . . . 303
12.2 Orthogonal vectors and Dirac Notation . . . . . . . . . . . . . . . 307
12.3 Complete Sets of Functions . . . . . . . . . . . . . . . . . . . . . 310
12.4 The Legendre Di¤erential Equation . . . . . . . . . . . . . . . . . 313
12.5 The Legendre Polynomials . . . . . . . . . . . . . . . . . . . . . . 316
12.6 The Generating Function for the Legendre Polynomials . . . . . 319
12.7 Legendre Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
12.8 The Associated Legendre Di¤erential Equation . . . . . . . . . . 329
12.9 Spherical Harmonics and the addition theorem . . . . . . . . . . 333
12.10The Method of Frobenius and the Bessel equation . . . . . . . . 341
12.11The orthogonality of the Bessel Functions . . . . . . . . . . . . . 348
12.12Fuch’s theorem for 2-nd DE . . . . . . . . . . . . . . . . . . . . . 354
13 A partial di¤erential equation 359

13.1 PDE in physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
13.2 Laplace’s equation in Cartesian coordinates . . . . . . . . . . . . 361
13.3 Laplace’s equation in spherical coordinates . . . . . . . . . . . . . 367
13.4 Laplace’s equation in cylindrical coordinates . . . . . . . . . . . . 373
13.5 Poisson’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . . 380
14 Functions of Complex Variables 387

14.1 Complex variables and functions . . . . . . . . . . . . . . . . . . 387
14.2 Analytic Functions . . . . . . . . . . . . . . . . . . . . . . . . . . 389
14.3 Important Terminologies . . . . . . . . . . . . . . . . . . . . . . . 391
14.4 Contour Integration and Cauchy’s Theorem . . . . . . . . . . . . 396
14.5 Cauchy’s Integral formula . . . . . . . . . . . . . . . . . . . . . . 398
14.6 Laurent’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 404
14.7 Residues and the Residue Theorem . . . . . . . . . . . . . . . . . 405
14.8 Methods of …nding residues . . . . . . . . . . . . . . . . . . . . . 407
14.9 Applications of the Residue Theorem . . . . . . . . . . . . . . . . 412
14.10Standard Methods of Integration using Contour Integrals . . . . 418
14.11The Kramers-Kronig Relations . . . . . . . . . . . . . . . . . . . 419
14.12The revised Residue Theorem . . . . . . . . . . . . . . . . . . . . 420
15 Laplace transform 433

15.1 Inverse Laplace Transform . . . . . . . . . . . . . . . . . . . . . . 438
15.2 Applications of Laplace transforms . . . . . . . . . . . . . . . . . 440
16 Fourier series and transform 449

16.1 Introduction to Fourier Series . . . . . . . . . . . . . . . . . . . . 449
16.2 The Fourier Series Expansion of a Periodic Function . . . . . . . 452
16.3 Dirichlet Conditions . . . . . . . . . . . . . . . . . . . . . . . . . 455
16.4 Fourier series with spatial and temporal arguments . . . . . . . . 457
16.5 The Fourier Transform . . . . . . . . . . . . . . . . . . . . . . . . 460
16.6 The Dirac Delta Function and the Fourier transform . . . . . . . 465
CONTENTS vii
16.7 Applications of the Fourier Transform . . . . . . . . . . . . . . . 467

16.8 Fourier Transforms and Convolution . . . . . . . . . . . . . . . . 473
III Mathematical Methods in physics III 477

17 Manifolds 479
17.1 What is a Manifold? . . . . . . . . . . . . . . . . . . . . . . . . . 479
17.2 Curves and surfaces in a Manifold . . . . . . . . . . . . . . . . . 481
17.3 Coordinate transformations and summation convention . . . . . . 484
17.4 The Riemannian geometry . . . . . . . . . . . . . . . . . . . . . . 488
17.5 Intrinsic and extrinsic geometry and the metric . . . . . . . . . . 491
17.6 Length, area, and volume . . . . . . . . . . . . . . . . . . . . . . 498
17.7 Local Cartesian coordinates and tangent space . . . . . . . . . . 506
17.8 The signature of a manifold . . . . . . . . . . . . . . . . . . . . . 511
18 Vector Calculus on manifolds 515

18.1 The tangent vector . . . . . . . . . . . . . . . . . . . . . . . . . . 515
18.2 The basis vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
18.3 The metric function and coordinate transformations . . . . . . . 517
18.3.1 Raising and lowering vector indices . . . . . . . . . . . . . 523
18.4 The inner product and null vectors . . . . . . . . . . . . . . 524
18.5 The a¢ ne connections . . . . . . . . . . . . . . . . . . . . . . . . 525
18.6 Local geodesic and Cartesian coordinates . . . . . . . . . . . . . 533
18.7 The gradient, the divergence, the curl on a manifold . . . . . . . 536
18.8 Intrinsic derivative of a vector along a curve . . . . . . . . . . . . 538
18.9 Parallel transport . . . . . . . . . . . . . . . . . . . . . . . . . . . 539
18.10Null curves, non-null curves, and a¢ ne parameter . . . . . . . . . 540
18.11Refreshment from theoretical physics: the calculus of variation . 542
18.12The geodesic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544
18.13Stationary property of the non-null geodesic . . . . . . . . . . . . 546
19 Tensor Calculus on manifolds 549

19.1 Tensors …elds and rank of a tensor . . . . . . . . . . . . . . . . . 549
19.2 Mapping tensors into tensors . . . . . . . . . . . . . . . . . . . . 554
19.3 Elementary tensor operations . . . . . . . . . . . . . . . . . . . . 555
19.4 Tensors and coordinate transformations . . . . . . . . . . . . . . 557
19.5 Tensor equations and the quotient theorem . . . . . . . . . . . . 558
19.6 Covariant derivatives of a tensor . . . . . . . . . . . . . . . . . . 559
19.7 Intrinsic derivative . . . . . . . . . . . . . . . . . . . . . . . . . . 561
viii CONTENTS
Preface
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
Series and Convergence
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.
1.1 Sequence and series

A sequence: a sequence is an ordered list of objects (or events). Like a set, it
contains members (also called elements, or terms), and the number of ordered
element (possibly in…nite) is called the length of the sequence. Unlike a set,
order matters, and exactly the same elements can appear multiple times at
di¤erent positions in the sequence. A sequence is a discrete function.
A series: The sum of terms of a sequence is a series. More precisely, if
(x1 ; x2 ; x3 ; :::) is a sequence, one may consider the sequence of partial sums
(S1 ; S2 ; S3 ; ::Sn :), with
n
X
Sn = x1 + x2 + x3 :: + xn = xk : (1.1)
k=0
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
2; 4; 8; 16; 32; :::; (1.2)

ii. A bouncing ball that raises each time 2/3 of the height of the previous
bounce:
2 4 8 16
1; ; ; ; ; :::; (1.3)
3 9 27 81
Suppose we want to determine the total number of bacteria or the total height
the ball raises in a given period of time we add the terms in the above expres-
sions. If we denote the sum of those terms by Sn , we may write the above
3
4 CHAPTER 1. SERIES AND CONVERGENCE
Figure 1.1: Bacteria growth in diluted human blood sample.
Figure 1.2: A bouncing ball.
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
is known as geometric series. Using

1
X
rk = 1 + r + r2 + r3 + r4 + ::: (1.8)
k=0
1.1. SEQUENCE AND SERIES 5
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)
Then for jrj < 1; we have

1
X 1
rk = (1.11)
1 r
k=0
so that the geometric series becomes

1
X a
S = lim Sn = a rk = : (1.12)
n!1 1 r
k=0
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
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
In a similar way for x3 ; we can easily show that
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
which can be put in the form

1
X
X=a rn ; (1.25)
n=0
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.
1.2 Testing series for convergence

If an in…nite series, like the geometric series we studied in the previous section,
has a …nite sum it is said to be convergent otherwise it is divergent. There are
di¤erent ways of testing whether a given in…nite series,
1
X
a0 + a1 + a2 + :::: = an ; (1.28)
n=0
is a convergent or divergent series. Here we discuss some of the commonly used

to test the convergence of an in…nite series
(1) Preliminary test: If an does not tend to zero as n ! 1;
lim an 6= 0; (1.29)
n!1
the in…nite series is a divergent series.

(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
is a convergent series, the in…nite series in Eq. (1.28) is a convergent series.

(3) The comparison test: Suppose there exists a convergent in…nite series with
all the terms positive,
1
X
bn = b0 + b1 + b2 + ::::; (1.31)
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)
The series in Eq. (1.28) is convergent if this integral is …nite otherwise it is

a divergent series. In Eq. (1.32), only the upper limit of the integration is
included. This means the integral need to be evaluated only at the upper
limit to test the convergence.
(5) The ratio test: In the ratio test, for the in…nite series in Eq. (1.28), we
…rst evaluate the ration de…ned by
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.
Example 1.2 Consider the in…nite series given by

1 2 6 24
1+ + + + ; :::; (1.35)
6 120 5040 362; 880
1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 9
(a) Rewrite this series using standard summation notation.
(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)!
so that the series can be expressed as
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).
1.3 Series representations of real functions

This section introduces to how we express a given well de…ned real function in a
series form. We focus on three form of series representation of a function P (x):
Power series, Taylor, and Maclaurin’s series. The series representation of this
function could be convergent or divergent. The convergence or divergence is
determined by the domain of the variable, x, that the function depends on.
We will see how to determine the range of the domain for which the series is
convergent (i.e. the interval of convergence).
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
exists for all n 0 and x 2 R domain, we can write
1
X 1 dn f (x)
f (x) = bn (x a)n , where, bn = (1.42)
n=0
n! dxn x=a
Taylor Series for some common functions:

X1
( 1)n 2n+1
sin x = x
n=0
(2n + 1)!
x3 x5
=x + + ::: convergent for all x 2 R (1.43a)
3! 5!
X1
( 1)n 2n
cos x = x
n=0
(2n)!
x2 x4
=1 + :::convergent for all x 2 R (1.43b)
2! 4!
X1
1 n
ex = x
n=0
n!
x2
=1+x+ + :::convergent for all x 2 R (1.43c)
2!
X1
( 1)n+1 n
ln(1 + x) = x
n=1
n
x2 x3
=x + :::convergent for 1<x 1; (1.43d)
2 3
1
X p
(1 + x)p = xn
n
n=1
p(p 1) 2
= 1 + px + x + :::convergent for all jxj < 1 (1.43e)
2!
where
p p (p 1) (p 2) (p 3) ::: (p n + 1)
= (1.44)
n n!
and it 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 1.3
(a) Show that the Maclaurin’s Series expansion for the function
f (x) = ln(1 + x) (1.45)
is expressible as
X1
( 1)n+1 n
ln(1 + x) = x : (1.46)
n=1
n
(b) Find the interval of convergence for this series.

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)
so that one can easily see that

(
( 1)n+1
bn = n ; for n = 1; 2; 3::: (1.53)
0 n=0
There follows that

1
X X1
( 1)n+1 n
ln(1 + x) = bn xn = x : (1.54)
n=0 n=1
n
This is the Maclaurin’s Series.

(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)
We note that at x = 1, the series becomes

1
( P
X ( 1)n+1 n 1 ( 1)n+1
x = n=1P1 n 1 ; for x = 1 (1.59)
n=1
n n=1 n for x = 1
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
shows that for x = 1, the series

X1 X1
( 1)n+1 n 1
x = (1.61)
n=1
n n=1
n
is a divergent series. However, for x = 1 the series becomes an alternating

series
X1 1
( 1)n+1 n X ( 1)n+1
x = (1.62)
n=1
n n=1
n
which is a convergent series.
Exercise: Prove that an alternating series of the form
1
X
( 1)n+1 an (1.63)
n=1
with an > 0 is a convergent series if an is monotonically decreas-

ing.
Therefore, the interval of convergence for the Maclaurin’s Series is 1<x 1:

Example 1.4 In Special Relativity the total energy of an object of (rest) mass
2 1=2
m0 moving with a speed v is given by E = m0 c2 , where = 1=(1 ) ,
8
= v=c, and c = 3:0 10 m=s is the speed of light in a vacuum. This
total energy is the sum of the kinetic energy T and the rest-mass energy
Eo = m0 c2 (this is the object’s energy when the object is not moving, so
v = 0 and = 1):
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)
Applying the series expansion

1
X p p(p 1) 2
(1 + x)p = xn = 1 + px + x :::; (1.67)
n 2!
n=1
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
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 )
Figure 1.4: Two charged balls suspended from a massless string.
and the gravitational ( W ) forces are given by
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
F1y = Ty W = T sin mg = 0 ) T sin = mg (1.72)
and in the x direction

1 Q2 1 Q2
F1x = Fe Ty = T cos ) T cos = (1.73)
4 0 x2 4 0 x2
Dividing Eq. (1.72) by (1.73), we …nd

2
sin 4 0 mgx
tan = = 2
: (1.74)
cos Q
In terms of the length of the threads, L; and the separation distance, x,

we may write
q
2
L2 x4 2L x2
1=2
tan = = 1 : (1.75)
x=2 x 4L2
Applying the series expansion
1
X
p p p(p 1) 2
(1 + x) = xn = 1 + px + x ::: (1.76)
n 2!
n=1
we can express Eq. (1.75) as

1=2 2
#
1 1
2L x2 2L 1 x2 2(2 1) x2
1 = 1 + + ::: :
x 4L2 x 2 4L2 2! 4L2
(1.77)
For large L >> x we can drop all the terms in the series that involve L
and approximate Eq.(1.77) as
1=2
2L x2 2L
tan = 1 ' (1.78)
x 4L2 x
Substituting Eq. (1.78) into Eq. (1.74) and solving for x, we …nd
2 2 1=3
4 0 mgx 2L 4 0 mgx Q2 L
tan = ) ' )x' (1.79)
Q2 x Q2 2 0 mg
Another simple way: For the …gure shown belowwe may write
F1y = Ty W = 0 ) T cos = mg (1.80)
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
sin ' and cos ' 1: (1.85)

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
1.4 Homework Assignment 1

1. Land for sell: Imagine you are given a speci…c procedure to sell a big
chunk of land to investors. This land is triangularly shaped plateau. The
triangle can be approximated as an equilateral triangle. The size of this
land, approximately, about square kilometers. The procedure for selling
the land reads as follows:
(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.
2. Find the limit of the sequence an given by
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!
4. Find the interval of convergence for the series

1
X n
(x 1)
:
n=0
2n
5. Prove that
n
X n (n + 1)
i=
i=1
2
is true for any integer n > 0 (even or odd).

1.5. HOMEWORK ASSIGNMENT 2 19

For the functions listed in problem 1 and 2: (a) Find the …rst four none zero
terms of the Maclaurin’s series; (b) Find the general term and write the series in
summation form; (c) Check your result in (a) by computer; (d) Use a computer
to plot the functions and several approximating partial sum of the series.
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) Find F=W as a series of power of .

(b) In the play ground it is not easy to estimate or measure the angle but
we can estimate x and l. Find F=W as a series of power of x=l.
4. Evaluate the following indeterminate functions using ’L’Hospital’s rule
(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
5. Use the series you know to show that

2 4 6
+ ::: = 1:
3! 5! 7!
Chapter 2
Complex Numbers,
functions, and series
2.1 Complex numbers

If you punch the real number, x = 1; into your calculator and try to …nd
its square roots, your calculator displays an error. That is because in reality
the square root for a negative number does not exist. For Mathematicians and
Physicists the square roots of a negative number do exist. It is just not in a set
of real numbers but rather in more universal set of numbers known as complex
numbers. A complex number, z, is de…ned by
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)
Rectangular and polar representation: For any complex number, z = a +

ib, we can use a rectangular representation using the Cartesian coordinates
(x; y) or polar representation using the polar coordinates (r; ). In rectangular
representation the real part a is the x coordinate and the imaginary part b is
the y coordinate,
a = x; b = y: (2.4)
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
z = x + iy = r cos + ir sin : (2.6)
The Cartesian and Polar coordinates, as shown in Fig. 2.1, are related by the
equations
x = r cos ; y = r sin ; (2.7)
Figure 2.1: Rectangular and Polar representation of a complex number z =

x + iy:
Euler’s equation and exponential representation: The Euler’s Equation, that

relates an exponential function and trigonometric functions, is given by
exp (i ) = cos ( ) + i sin ( ) : (2.8)
We will derive Euler’s equation in section 2.2 when we introduce to series of

complex function. Using Euler’s equation one can also represent the complex
number in Eq.(2.6); in an exponential form as
z = x + iy = r (cos ( ) + i sin ( )) = r exp (i ) : (2.9)
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
Complex conjugate of a complex Number : The complex conjugate of a com-

plex number, z = x + iy; is denoted by z and de…ned as
z =x iy (2.11)
The magnitude of a complex number, z; is related to its complex conjugate by

p p p
jzj = zz = (x + iy) (x iy) = x2 + y 2 : (2.12)
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,
3 + 4i (3 + 4i) (3 + 4i) 9 + 12i + 12i 16 7 + 24i

z= = = =
3 4i (3 4i) (3 + 4i) 32 + 42 25
7 24
= + i ) Re z = 7=25; Im z = 24=25 (2.14)
25 25
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:
The complex conjugate, z ; is given by replacing i by i

p
3 2i sin
z = (2.16)
4 + 2e 4i
Applying Euler’s formula, we can write
4i
2e = 2 cos (4 ) 2i sin (4 ) (2.17)
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
complex conjugate. The complex conjugate of 2 cos (4 ) 2i sin (4 ) is

2 cos (4 ) + 2i sin (4 ) . Thus
p
3 2i sin (2 cos (4 ) + 2i sin (4 ))
z =
(2 cos (4 ) 2i sin (4 )) (2 cos (4 ) + 2i sin (4 ))
p
3 2i sin (4 + 2 cos 4 + 2i sin 4 )
= 2 (2.19)
(4 + 2 cos 4 ) + 4 sin2 4
p p
3 (4 + 2 cos (4 )) + 4 sin (4 ) sin + 2i 3 sin (4 ) 2i sin (4 + 2 cos (4 ))
= 2
(4 + 2 cos (4 )) + 4 sin2 (4 )
(2.20)
p
3 (4 + 2 cos (4 )) + 4 sin (4 ) sin
) z = 2
(4 + 2 cos (4 )) + 4 sin2 (4 )
p
2 3 sin (4 ) 2 sin (4 + 2 cos (4 ))
+i 2 : (2.21)
(4 + 2 cos (4 )) + 4 sin2 (4 )
Therefore, the real and imaginary parts of z are
p
3 (4 + 2 cos 4 ) + 4 sin 4 sin
Re (z ) = 2 ; (2.22)
(4 + 2 cos (4 )) + 4 sin2 (4 )
p
2 3 sin 4 2 sin (4 + 2 cos 4 )
Im (z ) = 2 (2.23)
(4 + 2 cos (4 )) + 4 sin2 (4 )
Example 2.3 Consider the complex function

3
Z(x) = 2f (x) ig(x) (2.24)
2
where f (x) and g(x) are possibly complex functions of the real variable x.
(a) Find an expression for the complex conjugate of the function Z(x).
(b) Evaluate Z (x) for x = 1 if
p
f (x) = 3ix2 and g(x) = 2x + 4 2i (2.25)
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.
(b) We are given p

f (x) = 3ix2 , g(x) = 2x + 4 2i (2.27)
from which we …nd
p
f (x) = 3ix2 ; g (x) = 2x 4 2i (2.28)
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)
Example 2.4 Find the values of the real quantities x and y if
(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
(5 + i) x 5yi = 5x + i (x 5y) = 4 + 3i (2.33)
Two complex number z1 and z2 are equal if and only if their corresponding
Re and Im parts are equal. Thus
5x = 4; x 5y = 3 ) x = 4=5; y = 11=25: (2.34)
Example 2.5 The Intensity of a Traveling Electromagnetic Waves: An elec-

tromagnetic wave is described by two mutually perpendicular oscillating
~ and magnetic …eld (B)
electric (E) ~ vectors. An electromagnetic wave
traveling in the positive z-direction the electric …eld can be expressed as
~ t) = E0 exp [i (kz
E(z; !t)] x
^;
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
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
~k = n ! z^ = ! (nR + inI ) z^: (2.37)

c c
~ usually determined from the electric …eld
The magnetic …eld vector (B)
~ using the relation
vector (E)
~ t) = ~k
B(z; ~ (z; t):
E (2.38)
The “Poynting vector” is an important physical quantity in describing

electromagnetic waves. It is proportional to the cross product of the elec-
tric and magnetic …eld vectors,
~ t) = 1 E(z;
S(z; ~ t) ~ t):
B(z; (2.39)
0
The direction of Poynting vector shows the direction of propagation of

the traveling wave and the magnitude gives the intensity of the wave as a
function of position and time in the medium that the wave is propagat-
ing in. For a an electromagnetic wave traveling in a medium where the
refractive index is complex [Eq. (2.36)],
(a) …nd the complex conjugate of the electric …eld vector
(b) Find the magnetic …eld vector
(c) Shown that the magnitude of the Poynting vector is given by

q p
~ ~ ~ n2R + n2I 2 2!
S(z; t) = S(z; t) S (z; t) = E0 exp nI z : (2.40)
0 c c
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)
Taking the complex conjugate of Eq. (2.37), we have

!
k = (nR inI ) : (2.43)
c
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)
(c) Using the electric …eld

h ! i
~ t)
E(z; = E0 exp [i (kz !t)] x
^ = E0 exp i (nR + inI ) z !t x^
h c i
~ t) = E0 exp ! !
) E(z; nI z exp i nR z !t y^; (2.46)
c c
the result in part (b) for the magnetic …eld, the pointing vector
~= 1E
S ~ ~
B (2.47)
0
can be expressed as
~ = 1 n ! h ! i o
S E0 exp nI z exp i nR z !t x ^
0 c c
n! ! h ! i o
(nR + inI ) E0 exp nI z exp i nR z !t y^
c c c
) S ~ = 1 ! (nR + inI ) E02 exp !
2 nI z (^x y^) (2.48)
0 c c
) S ~ = 1 ! (nR + inI ) E02 exp !
2 nI z z^ (2.49)
0 c c
Then the magnitude of the pointing vector (the intensity)
q
~ t) = S(z;
S(z; ~ t) S ~ (z; t); (2.50)
becomes
~ t) = 1 ! ! 1 !
S(z; (nR + inI ) E02 exp 2 nI z z^ (nR inI ) E02
0 c c 0 c
! i1=2
exp 2 nI z z^
p c
! n2R + n2I 2 2!
) jS(z; t)j = E0 exp nI z : (2.51)
0c c
This result shows that the intensity decays exponentially as the wave pen-
etrates inside the medium.
2.2 Complex In…nite Series and the Circle of

Convergence
Complex In…nite Series: An in…nite series
1
X
zn = z0 + z1 + z2 + z3 :::zn + ::: (2.52)
n=0
is said to be a complex in…nite series if
zn = an + ibn ; (2.53)
where an and bn are real number. An in…nite complex series is convergent if

and only if
X1
an = a0 + a1 + a2 + a3 :::an + ::: = A (2.54)
n=0
and
1
X
bn = b0 + b1 + b2 + b3 :::bn + ::: = B; (2.55)
n=0
where A and B are real constants.

Test of convergence: For a complex in…nite series the convergence can be
tested using absolute convergence test. This means
(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.
Series of Complex function: In the previous chapter we have seen a real

power series
1
X
an xn = a0 + a1 + a2 + a3 :::an + ::: (2.56)
n=0
where an and x are real. A power series

1
X n 2 3 n
cn (z z0 ) = c0 +c1n (z z0 )+c2 (z z0 ) +c3 (z z0 ) :::cn (z z0 ) +:::
n=0
(2.57)
2.2. COMPLEX INFINITE SERIES AND THE CIRCLE OF CONVERGENCE29
is a series of complex function when z is a complex variable and z0 or cn are

complex.
Radius of convergence: Consider the complex power series in Eq. (2.57).
This series converges, applying the ratio test, for
n+1
cn+1 (z z0 ) cn+1
= lim n < 1 ) lim jz z0 j < 1 (2.58)
n!1 cn (z z0 ) n!1 cn
For example, for
cn+1 n
= (2.59)
cn n+1
we have
cn+1
lim =1 (2.60)
n!1 cn
which leads to
= jz z0 j < 1: (2.61)
For z = x + iy and z0 = a + ib; we note that
z z0 = x + iy (a + ib) = (x a) + i(y b)
p 2
) jz z0 j = (x a)2 + (y b)2 < 1 ) (x a) + (y b)2 < 12 : (2.62)
We recall the equation
(x a)2 + (y b)2 = R2 ; (2.63)
de…nes a circle with radius, R, centered at (a; b) : Thus the series converges for
all complex numbers, z; inside the circle (disk) of radius R = 1 on the complex
plane: This circle (disk) is known as circle (disk) of convergence for the in…nite
complex series in Eq. (2.57) and the radius of the circle (disk) is called the
radius of convergence.
Example 2.6 Determine the convergence of the following complex in…nite se-
ries
1
X i 1 i 1
zn = 1 + (2.64)
n=0
2 4! 6! 8!
Solution: This series can be expressed as

1
X X1 n
( i)
zn = (2.65)
n=0 n=0
(2n)!
Applying the ratio test we have
,
n+1 n n+1
zn+1 ( i) ( i) ( i) (2n)!
lim = lim = lim n
n!1 zn n!1 (2 (n + 1))! (2n)! n!1 (2 (n + 1))! ( i)
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.
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
Solution: A complex series

1
X n
an (z z0 ) (2.68)
n=0
is convergent when
n+1
an+1 (z z0 )
= lim n < 1: (2.69)
n!1 an (z z0 )
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
Example 2.9 Derive

(b) Euler’s Formula
ei = cos ( ) + i sin ( ) (2.76)
using series expansion of complex variables.
(b) the double-angle formulae for sin(2 ) and cos(2 )
sin(2 ) = 2 sin ( ) cos ( ) ;

cos(2 ) = cos2 ( ) sin2 ( ) : (2.77)
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!

2 4 3 3
exp (i ) = 1 + + ::: + i i +i ::: (2.80)
2! 4! 3! 5!
2 4 3 3
exp (i ) = 1 + + ::: + i + ::: (2.81)
2! 4! 3! 5!
Applying the the series expansions for sine and cosine functions
X1
( 1)n 2n+1 x3 x5 x7
sin x = x =x + ::: convergent for all x
n=0
(2n 1)! 3! 5! 7!
(2.82)
X1
( 1)n 2n x2 x4 x6
cos x = x =1 + :::convergent for all x (2.83)
n=0
(2n)! 2! 4! 6!
we may write
exp (i ) = cos ( ) + i sin ( ) (2.84)
(b) using the result above we may write

2
[exp (i )] = exp (2i ) = cos (2 ) i sin (2 ) (2.85)
But we also know that

2 2 2
[exp (i )] = [cos ( ) i sin ( )] ) [exp (i )] = cos2 ( ) sin2 ( )
2i sin ( ) cos ( ) (2.86)
Therefore, we can see that
cos (2 ) i sin (2 ) = cos2 ( ) sin2 ( ) 2i sin ( ) cos ( ) (2.87)
which leads to
cos (2 ) = cos2 ( ) sin2 ( ) ; sin (2 ) = 2 sin ( ) cos ( ) : (2.88)
2.3 Powers and Roots of complex numbers

The powers and roots of a complex number is easily determined by applying the
Euler’s formula. Let’s consider a complex number
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)
and show the solutions on a complex plane.
Solution: Noting that

z 5 = 32 = r exp [ ] (2.95)
where
r = 25 ; = 0 (2.96)
the solutions to
z 5 = 32 (2.97)
given by
zk = r1=5 exp (i k ) = r1=5 (cos ( k ) + i sin ( k )) ; (2.98)
2.3. POWERS AND ROOTS OF COMPLEX NUMBERS 33
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
z0 = 2 [cos (0) + i sin (0)] = 2 + i0

2 2
z1 = 2 cos + i sin = 0:62 + i1:90
5 5
4 4
z2 = 2 cos + i sin = 1:62 + i1:18
5 5
6 6
z3 = 2 cos + i sin = 1:62 i1:18
5 5
8 8
z4 = 2 cos + i sin = 1:62 i1:90 (2.100)
5 5
These roots are shown on the complex plane in Fig. 2.2

2.4 Algebraic vs transcendental functions

The logarithm and the exponential function are examples of transcendental func-
tions. Transcendental function is a term often used to describe the trigonometric
functions, i.e., sine, cosine, tangent, cotangent, secant, cosecant, and so forth. A
function that is not transcendental is said to be algebraic. Examples of algebraic
functions are rational functions and the square root function. A composition of
transcendental functions can give an algebraic function. For example
p
f (x) = cos (arcsin (x)) = 1 x2 (2.101)
Euler’s Equation, trigonometric, and hyperbolic Functions: We recall the
Euler’s formula in terms of a real variable x is given by
exp (ix) = cos (x) + i sin (x) (2.102)
and taking the complex conjugate we have
exp ( ix) = cos (x) i sin (x) (2.103)
so that adding the two equations we …nd
eix + e ix
cos (x) = (2.104)
2
and subtracting
eix e ix
sin (x) = : (2.105)
2i
If x is a complex variable z, using Euler’s Equation, we can express the sine and
cosine functions as
eiz + e iz
cos (z) = (2.106)
2
and
eiz e iz
sin (z) = (2.107)
2i
If z is a pure imaginary complex number z = iy, the sine and cosine functions
become
e y + ey
cos (iy) = ) cos (iy) = cosh (y) (2.108)
2
and
e y ey ey e y
sin (iy) = =i ) sin (iy) = i sinh (y) (2.109)
2i 2
where
ey + e y ey e y
cosh (y) = ; sinh (y) = (2.110)
2 2
are called the cosine hyperbolic and sine hyperbolic functions. It can be gener-
alized for a complex number z as
z
e + ez ez + e z
sinh (z) = ; cosh (z) = : (2.111)
2 2
2.4. ALGEBRAIC VS TRANSCENDENTAL FUNCTIONS 35
Other hyperbolic functions: named and de…ned in a way similar to the

trigonometric functions:
sinh (z) 1
tanh (z) = ; coth (z) = (2.112)
cosh (z) tanh (z)
1 1
csc h (z) = ; sec h (z) =
sinh (z) cosh (z)
Logarithms: Consider a complex number z = rei , then the logarithm of z

to the base e which is denoted by ln z can be expressed as
ln z = ln rei = ln r + ln ei = ln r + i ln e ) ln z = ln r + i (2.113)
The principal value: the principal value of a complex number z expressed

in polar coordinates as
z = rei (2.114)
is the value of the complex number when 0 2 : The principal value of
ln z is the value of ln z = ln rei for 0 2 :
Example 2.11 Evaluate the following integral applying Euler’s formula,

Z
I= sin ( x) cos ( x) dx (2.115)
where and are positive, nonzero integers.

Solution: Applying Euler’s formula we can write
ei x
e i x
ei x
+e i x
sin ( x) = ; cos ( x) = (2.116)
2i 2
so that
ei x
e i x
ei x
+e i x
sin ( x) cos ( x) =
2i 2
ei( + )x
e i( + )x
+ ei( )x
e i( )x
= (2.117)
:
4i
Then substituting the above expression for the integrand, we have
Z h i
1
I= ei( + )x e i( + )x + ei( )x
e i( )x
dx (2.118)
4i
so that applying Euler’s formula, once again, one can write
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
which leads to
1 cos [( + ) x] cos [( ) x]
I= + (2.121)
2 +
Noting that cosine is an even function,
cos (x) = cos ( x)
we can easily see that Eq. (2.121) reduces to

cos [( + ) ] cos [( + ) ] cos [( ) ] cos [( ) ]
I= + =0
+
(2.122)
Therefore Z
I= sin ( x) cos ( x) dx = 0: (2.123)
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
Then substituting the above expression for the integrand, we have

Z
1
I= fcos [( ) x] cos [( + ) x]g dx (2.130)
2
1 sin [( ) x] sin [( + ) x]
)I= : (2.131)
2 +
We recall sin is an odd function, which means
sin ( x) = sin (x) (2.132)
we have
sin [ ( ) ]= sin [( ) ] ; sin [ ( + ) ] = sin [( + ) ]

(2.133)
so that
sin [( ) ] sin [( + ) ]
I= (2.134)
+
Since and are positive, none zero integers, we always 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 =
which means the sine function
fj (x) = sin (jx) (2.138)
form an orthogonal set of functions in the domain ( ; ):

Example 2.13 Evaluate the complex expression
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
and using the relation

ab = eb ln a ; (2.142)
for a = 2; b = 1 (or 1), we have
ln 2 1
e =2 = 0:5; eln 2 = 21 = 2 (2.143)
so that
3=2 3
cos [i ln 2] = = = 0:75: (2.144)
2 4
Example 2.14 Find the principal value of the complex logarithm
z = ln (1 i) (2.145)
Solution: To …nd the principal value …rst we need to express z = 1 i using

polar coordinates. To this end, we note that
q p
2 7
r = j1 ij = 12 + ( 1) = 2; = +2 n (2.146)
4
where n = 0; 1; 2:::But to …nd the principal value for ln z we use = 47 :
Thus p 7
z = 1 i = 2ei 4 (2.147)
which leads to
7 i 1 7 i
ln z = ln 21=2 + ) ln z = ln 2 + (2.148)
4 2 4
Example 2.15 Find the principal value of the complex quantity
i
z = (1 i) (2.149)
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
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
Solution: Let the value of this expression be so that

p !12
3+i
sin ( ) = p : (2.154)
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

1. For the complex number p i
z= 2e 4
(a) Find x; y; r; and write it in the forms
z = (x; y); z = x + iy; z = (r; ); and z = r(cos + i sin ):
(b) Find z and also write it in the forms
z = (x; y); z = x + iy; z = (r; ); and z = r(cos + i sin ):
(c) Plot both z and z and label it.

2. Write the complex expression
1
z= 2
(2 3i)
in the form z = a + bi:
3. Prove that
z1 z1
=
z2 z2
(i.e. the conjugate of the quotient of two complex numbers is the quotient
of the conjugates.) Also prove that
(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
z = cos (3t) + i sin(3t)

(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.

1. Test each of the following series for convergence.
(a)
1
X 1 1 1 1
n =1+ + 2 + :::: n + :::
n=0
(1 + i) (1 + i) (1 + i) (1 + i)
(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
4. Find the roots for

z = i1=5 :
and show your results on the complex plane.
5. Find the formula for sin (3 ) and cos (3 ) :
Hint: refer to the example for double angle formula that we did in class.

Find the values in rectangular form x+iy for the complex expressions in problem
1-4
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 ),
capacitor (ZC ), and inductor (ZL ), are given by
ZR = R; ZC = 1= (i!c) ; and ZL = i!L;
where R is the resistance of the resistor, C is the capacitance of the ca-

pacitor, L is the inductance of the inductor, and is the angular frequency
of the ac-source voltage, V (t) = Vmax sin (!t) in the circuit. In ac cir-
cuit for series connection of two impedances, Z1 and Z2 the equivalence
impedance Z12 is given by
Z12 = Z1 + Z2
where as for parallel connection

1 1 1
= + :
Z12 Z1 Z2
(a) Find the equivalent impedance of the circuit

(b) Find the magnitude of the equivalent impedance of the circuit.
(c) A circuit is said to be at resonance when the total impedance of the circuit
is real. Find ! in terms of R, L, and C at resonance.
Chapter 3
Vectors, Lines, and Planes
3.1 An Overview of Vector Fundamentals

A vector : A vector is a quantity that can be described by a magnitude and
direction. We shall represent a vector by a boldface letter (for example A) or
~ Here we will use
by a letter with an arrow on top of the letter (for example A).
the second way of representation. The component of a vector is represented by
~
a subscript (for example Ax represent the x component of the vector A).
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
3.2 Vector Multiplication

Two vectors can be multiplied in two di¤erent ways. One way of multiply-
ing two vectors leads to a scalar. Such multiplication of vectors is referred as
Scalar product (dot product). The other way of multiplying two vectors leads
to another vector and is known as cross product (vector product)
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 ~
Application: Suppose you pull an object on a frictionless surface by ap-

plying a force and moved the object horizontally for some distance. If the
force applied is F~ and the displacement of the object is S
~ , the work done
W is given by
W = F~ S ~ = F~ S~ cos ( ) (3.6)
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)
and the direction can be determined using the right-hand rule.

3.2. VECTOR MULTIPLICATION 47
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.
Right-Handed Coordinate Systems: in right-hand coordinate system the right

hand rule is valid. For example in Cartesian coordinate system we have
^{ ^ |^
|^ = k; k^ = ^{; k^ ^{ = |^: (3.8)
If we know the components of the two vectors, A ~ = Ax^{ + Ay |^ + Az k^ and

~ ^
B = Bx^{ + By |^ + Bz k; then you can …nd the components of the vector
product C~ =A~ B ^ in Cartesian coordinates, using
~ = Cx^{ + Cy |^ + Cz k;
Cx = Ay B z Az B y (3.9)
Cy = Az Bx Ax Bz
Cz = Ax B y Ay B z
Application of cross product: torque due to a force F~ on a ridged object

placed at a position ~r with respect to an axis (through the origin) is given
by ~ = ~r F~ .
Example 1 A force of magnitude 4:0N pointing in the positive-x direction is

applied to a large solid object at the position (0; 2; 0)m. Find the resulting
torque exerted on the object about the origin using the
(a) the magnitude and direction form of the cross product.
(b) component form of the cross product.
Solution:
(a) The torque is given by

~ = ~r F~ (3.10)
where the force, pointing along the positive-x direction is given by
F~ = (4; 0; 0) N = 4N x
^ (3.11)
and the position vector
~r = (0; 2; 0)m = 2m^

y (3.12)
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)
Noting that the angle between ~r and F~ is
= (3.16)
2
the magnitude of the torque would be
j~ j = F~ j~rj sin ( ) = 8N m: (3.17)
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)
(b) The torque is given by

~ = ~r F~ (3.19)
where the force, pointing along the positive-x direction is given by
F~ = (4; 0; 0) N = 4N x
^ (3.20)
and the position vector
~r = (0; 2; 0)m = 2m^

y (3.21)
3.3. EQUATIONS OF A STRAIGHT LINE 49
We apply the relation

x^ y^ z^
~
A ~ =
B Ax Ay Az (3.22)
Bx By Bz
which gives
x
^ y^ z^
2 0 0 0 0 2
~ = ~r F~ = 0 2 0 =x
^ y^ + z^ (3.23)
0 0 4 0 4 0
4 0 0
~ = ~r F~ = 8N m^
z
3.3 Equations of a straight line

Symmetric equations (in a two dimensional space):
x x0 y y0 z z0
= = (3.24)
a b c
where a; b;and c 6= 0: When c = 0 we should write the symmetric equation of
the straight line as
x x0 y y0
= ; z = z0 (3.25)
a b
Parametric equations:
8
< x = x0 + at
~ or
~r = ~r0 + At y = y0 + bt (3.26)
:
z = z0 + ct
The symmetric equations can be obtained from the parametric equations:
x x0 y y0 z z0
t= = = (3.27)
a b c
3.4 Equation of a plane

~ = ai + bj + ck is normal (perpendicular) to a plane, then the dot product
If N
~ and the vector ~r ~r0
of the vector N
~r ~r0 = (x x0 ) x
^ + (y y0 ) y^ + (z z0 ) z^ (3.28)
is zero,
a (x x0 ) + b (y y0 ) + c (z z0 ) = 0 (3.29)
This de…nes the equation of the plane. It can be rewritten as
ax + by + cz = d (3.30)
where
d = ax0 + by0 + cz0 (3.31)
Example 2 The position of a particle at time t = 0 is given by (1; 2; 3)m. The

particle moves with a constant velocity (0; 2; 1)m=s. What is the equation
of the particle’s trajectory?
Solution: The initial position of the particle can be expressed, using Cartesian
unit vectors, as
~r0 = x
^ + 2^
y + 3^z (3.32)
at a later time t, let the position of the particle is described by the vector
~r = x (t) x
^ + y (t) y^ + z (t) z^ (3.33)
If the particle is traveling by a constant velocity, the vector
~r ~r0 = [x (t) 1] x
^ + [y (t) 2] y^ + [z (t) 3] z^ (3.34)
must be proportional to the velocity vector
~v = 0^
x + (2m=s) y^ + (1m=s) z^: (3.35)
This means
~r ~r0 = t~v ) [x (t) 1] x

^ + [y (t) 2] y^ + [z (t) 3] z^ (3.36)
= (2m=s) t^
y + (1m=s) t^
z
) 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
~r0 (~r0 ~r0 ) = 0: (3.41)
Since ~r0 describes a point on the trajectory of the particle given by
~r (t) = ~r0 + ~v t (3.42)

3.4. EQUATION OF A PLANE 51
at which the particle becomes closest to the origin, we may write
~r0 = ~r0 + ~v t0 ) ~r0 ~r0 = ~v t0 : (3.43)
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)
We may pick (3; 1; 0) which we write as
~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).
Solution: The symmetric equations of a line are given by

x x0 y y0 z z0
= = (3.57)
a b c
where
~ = a^
A x + b^
y + c^
z (3.58)
is a vector parallel to the line passing through (0; 2; 4) and (3; 2; 1) .
Noting that a vector parallel to the line passing through can be expressed
as
~ = (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).
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):

~=
1. Find the angles between A 2i + j ~ = 2i
2k and B 2j
~ = 2i j + 4k and B
2. Show that the vectors A ~ = 5i + 2j 2k are orthogonal
(perpendicular) and …nd a third vector perpendicular to both.
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.
4. A particle is traveling along the line (x 3)=2 = (y + 1)=( 2) = z 1:

~ Find the distance
Write the equation of this path in the form ~r = ~r0 + At.
of closest approach of the particle to the origin (that is, the distance from
the origin to the line). If t represents time, show that the time of the
closest approach is
A~ ~r0
t= :
A2
Use this value to check your answer for the distance of closest approach.
If P is the point of closest approach, what is A ~ ~r?
5. The vectors A~ = ai + bj and B

~ = ci + dj form two sides of a parallelogram.
Show that the area of the parallelogram is given by the absolute value of
the following determinant
a b
c d
Chapter 4
Matrix and determinants
4.1 Important Terminologies

A matrix : A rectangular array of quantities, usually inclosed in large parenthe-
ses with n rows and m columns.
2 3
a11 a12 a1m 1 a1m
6 a21 a22 a2m 1 a2m 7
6 7
6 7
6 7
A=6 6 7 (4.1)
7
6 7
6 7
4 an 11 an 12 an 1m 1 an 1m 5
an1 an2 anm 1 anm
when n = m the matrix is called a square matrix.

Transpose of a Matrix : The transpose of a Matrix A in Eq. (4.1) which we
denote by AT is obtained by simply write the rows of matrix A as columns.
2 3
a11 a21 an 11 an1
6 a12 a22 an 12 an2 7
6 7
6 7
6 7
T
A =6 6 7 (4.2)
7
6 7
6 7
4 a1m 1 a2m 1 an 1m 1 anm 1 5
a1m a2m an 1m anm
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
The Identity Matrix; I (The Unit Matrix, U )
IA = AI = A (4.4)
4.2 Matrix Arithmetic and Manipulation

Consider the three matrices:
0 1 0 1
2 4 2 1 3
2 3 1
A = ;B = @ 1 1 A;C = @ 4 1 2 A ; (4.5)
2 1 0
3 1 1 0 1
0 1
2 0 1
D = @ 1 1 2 A
3 1 0
Multiplication by a Scalar : Any matrix can be multiplied by a scalar:
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
Addition and subtraction: Two matrices can be added or subtracted if and

only if they have the same dimensions. From matrices A; B; C, and D we can
add/ subtract only matrices C and D
0 1 0 1
2 1 3 2 0 1
C +D = @ 4 1 2 A+@ 1 1 2 A (4.7)
1 0 1 3 1 0
0 1 0 1
2 2 1+0 3+1 0 1 4
= @ 4+1 1 1 2+2 A=@ 5 2 0 A
1+3 0+1 1+0 2 1 1
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
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 ; (4.10)
k=1
where (ab)ij is the element of the product matrix AB:

Commutativities: For any two multiplyable matrices C and D;
CD 6= DC (4.11)
Commutator : For square matrices C and D the Commutator [C; D] is de-

…ned as
[C; D] = CD DC (4.12)
For any three matrices, F; G;and H that can be multiplied we can write
The Associative Law :
F (GH) = (F G)H (4.13)
The Distributive Law :
F (G + H) = F G + F H (4.14)
4.3 Matrix representation of a set of linear equa-

tions
A set of linear equations
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 :
can be expressed using matrices as

2 32 3 23
a1 b1 c1 d1 x c1
6 a2 0 c2 d2 7 6 7 6 c2 7
6 76 y 7=6 7
4 0 b3 c3 0 54 z 5 4 c3 5 : (4.16)
a4 0 0 0 w c4
or
M r = k; (4.17)
where 2 3
a1 b1 c1 d1
6 a2 0 c2 d2 7
M =6
4 0
7 (4.18)
b3 c3 0 5
a4 0 0 0
is called the coe¢ cient matrix,

2 3
x
6 y 7
r=6 7
4 z 5; (4.19)
w
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
which is referred as the Augmented Matrix.
4.4 Matrix application: solving a set of linear

equations
With application of matrices we can solve a set of linear equations. There are
two di¤erent methods. The …rst involves the Gaussian Elimination method and
Row Echelon Form where as the second method involves Cramer’s Rule. We
…rst discuss the …rst method. The second method will be discussed after we get
introduced to square matrices and determinants.
Gaussian Elimination method (named after German mathematician and sci-
entist Carl Friedrich Gauss): in this method we follow the following basic steps
to solve a set of linear equations:
1st Write the set of linear equations using matrices.
2nd Obtain the augmented matrix.
3rd Apply any of the following elementary row operations
(a) Multiply any row of the matrix by a non-zero constant,
(b) Interchange any two rows,
(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.
Example 1 The matrix 2 3

0 1 4 0
6 0 0 1 0 7
B=6
4 0
7 (4.22)
0 0 1 5
0 0 0 0
in Row Echelon Form and in a reduced row form where as the matrix
2 3
1 1 1 1
C=4 0 9 0 2 5 (4.23)
0 0 0 3
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.
Rank of a Matrix : the number of nonzero rows remaining when a matrix

has been row reduced is called the rank of the matrix.
(a) If (rank M )<(rank A), the equations are inconsistent and there is no solu-
tion.
(b) If (rank M )=(rank A)= n (number of unknowns), there is one solution.

(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
Using Kirchho¤ ’s Current Law at node b; we may write
I1 = I2 + I3 ) I1 I2 I3 = 0 (4.27)
(b) Substituting R = 1and V = 2V in the results we obtained above, we have
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
so that we can put these equations using matrices as
M I = K; (4.30)
where coe¢ cient matrix M is given by

2 3
1 1 0
M =4 0 1 1 5 (4.31)
1 1 1
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
Subtract row 1 from row 3 :

2 3
1 1 0 1
A=4 0 1 1 0 5 (4.35)
0 2 1 1
Multiply row 2 by 2 and subtract from row three:

2 3
1 1 0 1
A=4 0 1 1 0 5 (4.36)
0 0 3 1
Divide row 3 by 3:
2 3
1 1 0 1
A=4 0 1 1 0 5 (4.37)
1
0 0 1 3
Subtract row 3 from row 2:
2 3
1 1 0 1
A=4 0 1 0 1
3
5 (4.38)
1
0 0 1 3
Divide row 2 by 1:
2 3
1 1 0 1
A=4 0 1 0 1
3
5 (4.39)
1
0 0 1 3
Subtract row 2 from row 1:
2 2
3
1 0 0 3
A=4 0 1 0 1
3
5 (4.40)
1
0 0 1 3
Therefore the values of the three currents would be
2 3 2 2 3
I1 3
4 I2 5 = 4 1 5 (4.41)
3
1
I3 3
which means
2 1 1
I1 = A; I2 = A; I3 = A (4.42)
3 3 3
(d) The number of nonzero rows for the row reduced augmented matrix
2 3
1 0 0 23
A=4 0 1 0 13 5 (4.43)
0 0 1 13
is 3 and the rank is 3. For the reduced coe¢ cient matrix (the …rst three
columns in the reduced augmented matrix)
2 3
1 0 0
M =4 0 1 0 5 (4.44)
0 0 1
we have also 3 nonzero rows and the rank is 3. There are 3 unknowns,
this means that (rank M )=(rank A)= 3 (number of unknowns), therefore
we …nd one solution for each unknowns.
4.5. DETERMINANT OF A SQUARE MATRIX 63
4.5 Determinant of a square matrix

The determinant of a 2 2 matrix : Suppose we are given a 2 2 matrix
a11 a12
A= ; (4.45)
a21 a22
the determinant is given by
2
X
jAj = ( 1)i+j aij Mij ; (4.46)
j=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 : 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
so that noting that

M11 = a22 ; M12 = a21 (4.48)
one …nds
jAj = a11 a22 a12 a21 : (4.49)
The determinant of a 3 3 matrix: Consider a 3-dimensional matrix
2 3
^{ |^ k^
C = 4 Ax Ay Az 5 (4.50)
Bx By Bz
constructed from two 3-dimensional vectors in Cartesian coordinates and the
^ Using Eq. (4.59) we may write the determinant of this
unit vectors, ^{, |^, and k.
matrix as
3
X
jAj = ( 1)1+j aij Mij (4.51)
j=1
= ( 1)1+1 a11 M11 + ( 1)1+2 a12 M12 + ( 1)1+3 a13 M13 ;

where we chose i = 1. Note that for a 3-dimensiona square matrix i = 1; 2; or
3: Noting that
a11 = ^{; a12 = |^; and a13 = k^ (4.52)
the corresponding Minors M11 ; M12 , and M13 are given by
Ay Az
M11 = = Ay Bz Az B y ; (4.53)
By Bz
Ax Az
M12 = = Ax B z Az B x ; (4.54)
Bx Bz
Ax Ay
M13 = = Ax B y Ay B x ; (4.55)
Bx By
we …nd for Eq. (4.51)

jCj = ( 1)1+1^{ (Ay Bz Az By ) + ( 1)1+2 |^(Ax Bz Az Bx )
+( 1)1+3 k^ (Ax By Ay Bx )
) jCj = ^{ (Ay Bz Az By ) + |^(Az Bx Ax Bz ) + k^ (Ax By Ay B(4.56)
x)
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
(b) two rows (or columns) are identical

Example 5
a11 a12 = a11 a11 a21
A= or A = ) det A = 0
a21 a22 = a21 a21 = a11 a22 = a21
(4.64)
(c) two rows (or columns) are proportional
Example 6
a11 a12 = ka11 a11 a12
B= or B = ) det A = 0;
a21 a22 = ka21 a21 = ka11 a22 = ka12
(4.65)
where k is the proportionality constant.
3. Interchanging a row or a column: If two rows (or two columns) of a square
matrix are interchanged, the determinant of the matrix remain the same.
Example 7
a11 a12 a12 a11
A = ,B =
a21 a22 a22 a21
) det A = det B (4.66)
Example 8 Evaluate the determinant of the matrix A:

0 1
2 2 0
A=@ 0 2 2 A (4.67)
1 1 1
Solution: Applying the relation above we can express the determinant as

1+1 2 2 1+2 0 2
jAj = 2 ( 1) + 2 ( 1) (4.68)
1 1 1 1
0 1
2 2 0
1+3 @ 0
+0 ( 1) 2 2 A (4.69)
1 1 1
2 2 0 2
jAj = 2 2 (4.70)
1 1 1 1
jAj = 2 (( 2 1) (2 1)) 2 ((0 1) (2 1)) = 8 + 4

) jAj = 12 (4.71)
4.6 Cramer’s Rule

Let a system of simultaneous equations be represented by the augmented matrix
A, and let M be the corresponding coe¢ cient matrix. Also, let Mj be the
coe¢ cient matrix with its j th column replaced by the constant column from
Aaug . We de…ne the following determinants:
D = jM j ; Dj = jMj j (4.72)
The value of the j th unknown, xj ; is then equal to
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.)
Solution: We recall that the coe¢ cient matrix is given by

2 3
1 1 0
M =4 0 1 1 5 (4.74)
1 1 1
and the corresponding Augmented matrix by

2 3
1 1 0 1
A=4 0 1 1 0 5: (4.75)
1 1 1 0
Replacing the …rst column of M by the last column of A; we …nd

2 3
1 1 0
M1 = 4 0 1 1 5; (4.76)
0 1 1
the second column of M by the last column of A

2 3
1 1 0
M2 = 4 0 0 1 5 (4.77)
1 0 1
4.7. THE ADJOINT AND INVERSE OF A MATRIX 67
and the third column of M by the last column of A

2 3
1 1 1
M3 = 4 0 1 0 5: (4.78)
1 1 0
For the determinants of the matrices M; M1 ; M2 ; and M3 , one can easily
…nd
1 1 0 1
jM j = ) jM j = 2 + 1 = 3; (4.79)
1 1 1 1
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
4.7 The adjoint and inverse of a matrix

Adjoint of a Matrix : The adjoint of a square matrix, A, is given by
adj(A) = [cof (A)]T (4.84)
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.
a. Row reduction in this approach for the matrix, for example,

2 3
a11 a12 a13
A = 4 a21 a22 a23 5 (4.87)
a31 a32 a33
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
so that the inverse of the Matrix A is given by

0 1
b11 b12 b13
A 1 = @ b21 b22 b23 A : (4.90)
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
a. Row reduction approach

b. The adjoint matrix approach
Solution: a. In the row reduction approach we start from

2 3
1 2 3 1 0 0
4 2 0 4 0 1 0 5 (4.92)
1 1 1 0 0 1
and try to get

2 3
1 0 0 a11 a12 a13
4 0 1 0 a21 a22 a23 5 (4.93)
0 0 1 a31 a32 a33
so that we can get the inverse matrix

0 1
a11 a12 a13
A 1 = @ a21 a22 a23 A : (4.94)
a31 a32 a33
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
Add row 3 to row 2:

2 3
1 2 3 1 0 0
4 2 1 0 1 1 1 5 (4.96)
0 1 4 1 0 1
Multiply row 2 by 2 and subtract the result from row 1

2 3
5 0 3 1 2 2
4 2 1 0 1 1 1 5 (4.97)
0 1 4 1 0 1
Divide row 1 by -5 and row 2 by 2:

2 3
1 0 3=5 1=5 2=5 2=5
4 1 1=2 0 1=2 1=2 1=2 5 (4.98)
0 1 4 1 0 1
Subtract row 1 from row 2

2 3
1 0 3=5 1=5 2=5 2=5
4 0 1=2 3=5 3=10 1=10 1=10 5 (4.99)
0 1 4 1 0 1
Divide row 2 by 1/2

2 3
1 0 3=5 1=5 2=5 2=5
4 0 1 6=5 3=5 1=5 1=5 5 (4.100)
0 1 4 1 0 1
subtract row 2 from row 3

2 3
1 0 3=5 1=5 2=5 2=5
4 0 1 6=5 3=5 1=5 1=5 5 (4.101)
0 0 14=5 2=5 1=5 4=5
Divide row 1 by 3=5, row 2 by 6=5, and row 3 by 14=5

2 3
5=3 0 1 1=3 2=3 2=3
4 0 5=6 1 1=2 1=6 1=6 5 (4.102)
0 0 1 1=7 1=14 2=7

2 3
5=3 0 1 1=3 2=3 2=3
4 0 5=6 0 5=14 5=21 5=42 5 (4.103)
0 0 1 1=7 1=14 2=7
2 3
5=3 0 0 10=21 25=42 20=21
4 0 5=6 0 5=14 5=21 5=42 5 (4.104)
0 0 1 1=7 1=14 2=7
Divide row 1 by 5=3 and row 2 by 5=6

2 3
1 0 0 1=7 5=14 4=7
4 0 1 0 3=7 2=7 1=7 5 (4.105)
0 0 1 1=7 1=14 2=7
Then the inverse matrix is given by

0 1
2=7 5=14 4=7
A 1 = @ 3=7 2=7 1=7 A : (4.106)
1=7 1=14 2=7
Using Matematica
N.B. Use Mathematica only to check your result!!!

b. Here we …rst need to …nd the adjoint matrix of A. In order to …nd the
adjoint, …rst we need to …nd the cofactor matrix. Let this cofactor
matrix be 0 1
c11 c12 c13
cof (A) = @ c21 c22 c23 A (4.107)
c31 c32 c33
where the elements
i+j
cij = ( 1) Mij (4.108)
in which Mij is the minor for the matrix A. Now using
0 1
1 2 3
A=@ 2 0 4 A (4.109)
1 1 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
Then the inverse matrix

T
1 [cof (A)]
A = (4.122)
det jAj
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)
4.8 Orthogonal matrices and the rotation ma-

trix
Orthogonal Matrices: matrices that make an orthogonal transformation of vec-
tors. In an orthogonal transformation of vectors the magnitude of the vectors
remains the same. For an orthogonal matrix
1
M = MT (4.124)
The Rotation Operator : Consider a rectangular coordinate system (XY Z)

rotated by an angle about the z axis in the counterclockwise direction to give
another coordinate system (X0Y 0Z0) as shown in the …gure below.
Consider a vector ~r which has components (x; y; z) in the XY Z coordinate

system. We want to …nd the components of this vector in the X 0 Y 0 Z 0 coordinate
system. From the …gure above we can write that the new components (x0 ; y 0 ; z 0 )
can be expressed as
x0 = x cos ( ) + y cos (90 ) = x cos ( ) + y sin ( ) (4.125)

0
y = x sin ( ) + y cos ( )
z0 = z
4.8. ORTHOGONAL MATRICES AND THE ROTATION MATRIX 73
Using matrix representation we can write that

0 0 1 0 10 1
x cos ( ) sin ( ) 0 x
@ y 0 A = @ sin ( ) cos ( ) 0 A @ y A ) r0 = Rr (4.126)
z0 0 0 1 z
where the matrix R( ) = R is the rotation matrix given by

0 1
cos ( ) sin ( ) 0
R = @ sin ( ) cos ( ) 0 A (4.127)
0 0 1
~ (I used Mathematica)
The inverse of the matrix R
which is the same as the transpose of the matrix, RT given by

0 1
cos ( ) sin ( ) 0
RT = @ sin ( ) cos ( ) 0 A : (4.128)
0 0 1
Therefore, since
1
R = RT (4.129)
the rotation matrix is orthogonal.

Linear Operators: The operator/function F (x) is a linear operator/function
if
F (ax) = aF (x) (4.130)
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)
Is this a linear operator?

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)
O(A) + O(B) = A2 + 3 + B 2 + 3 ) O(A) + O(B) = A2 + B 2 + 6 (4.135)
which shows
O(A + B) 6= O(A) + O(B) (4.136)
4.9 Linear independence

A set of linear equations or vectors could be linearly dependent or independent.
In order to obtain a complete solution to a set of linear equations using ma-
trices the equations must be linearly independent. Sometimes in some physical
problems it may be necessary to express a given vector as a set of linearly in-
dependent vectors. In this section we will see how linear independency in set of
linear equations, vectors, and functions can be veri…ed and how we can obtain
a set of linearly independent orthogonal vectors.
Linear independency-set of linear equations: The existence of the solutions
to a set of linear equations with n unknowns depends on the linear independency
of the equations which is determined by the Rank of the Augmented (A) and
Coe¢ cient matrices (M ). The Rank of a Matrix is the number of nonzero rows
remaining when a matrix has been row reduced.
a. If (rank M )<(rank A), the equations are inconsistent and there is no solution.
b. If (rank M )=(rank A)= n (number of unknowns), there is one solution.
c. If (rank M )=(rank A)= m < n, then m unknowns can be found in terms of

the remaining n m unknowns.
Example 12 [Linear independence-set of linear equations: in this example the

current rule is excluded in order to illustrate linear independency in a set
of linear equations.] Consider the circuit shown below. The circuit element
values are R = 1:0 and V = 2:0V . Find the values of the three branch
currents I1 ; I2 , and I3 by solving the system of three simultaneous equa-
tions obtained by applying only Kirchho¤’s voltage law to three di¤erent
loops.
4.9. LINEAR INDEPENDENCE 75
Solution: Applying Kirchho¤’s voltage law to three di¤erent loops we can write
V + I1 R + 2I2 R 2 V + I1 R = 0 ) 2I1 R + 2I2 R = V; (4.137)
2 V 2I2 R + I3 R 2 V + I3 R = 0 ) 2I2 R + 2I3 R = 0; (4.138)

and
V + I1 R + I3 R 2 V + I3 R + I1 R = 0 ) 2I1 R + 2I3 R = V: (4.139)
Substituting the values for R and V , we …nd
I1 + I2 = 1; I2 + I3 = 0; I1 + I3 = 1; (4.140)
which may be put in a matrix form

0 10 1 0 1
1 1 0 I1 1
@ 0 1 1 A @ I2 A = @ 0 A : (4.141)
1 0 1 I3 1
There follows that for the Augmented matrix, we have

0 1
1 1 0 1
A=@ 0 1 1 0 A (4.142)
1 0 1 1
and for the coe¢ cient matrix

0 1
1 1 0
M =@ 0 1 1 A: (4.143)
1 0 1
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
and subtract row three from row one

0 1
0 0 0 0
@ 0 1 1 1 A: (4.145)
1 0 1 1
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.
Linearly dependent and independent vectors: If we de…ne three vectors from

the three equations we obtained in the example above
~ = (1; 1; 0) ; B
A ~ = (0; 1; 0) ; C
~ = (1; 0; 1) (4.146)
we can see that

~+B
A ~ ~ =0
C (4.147)
These three vectors are linearly dependent vectors. In general for a set of vectors
~1; A
A ~2; A
~ 3 ; :::A
~ n ; if
~ 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^:
Using matrix representation, we may write

~ 1 + k2 A
k1 A ~ 2 + k3 A
~ 3 + ::: + kn A~n = 0
0 10 1
k1 a1 k1 a2 k1 a2 x^
) @ k2 b1 k2 b2 k2 b2 A @ y^ A = 0: (4.150)
k3 c1 k3 c2 k3 c2 z^
This equality holds if and only if
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.
In general, any set of linearly independent vectors either is already a

mutually orthogonal set of vectors or else the vectors in the set can be combined
to form a set of mutually orthogonal vectors (in a process called Gram-Schmidt
orthogonalization).
A Basis Set of Vectors: these are set of linearly independent vectors which
span a vector space.
A Spanning Set of Vectors: a set of vectors spans a space if all the vectors
in the space can be written as a linear combination of the spanning basis set of
vectors.
Example 14 Consider the four set of vectors given by

~
A = ~ = (5; 2; 1) ;
(1; 4; 5) ; B (4.155)
~
C = ~ = (3; 6; 11)
(2; 1; 3) ; D
and the two dimensional vector space (a plane) formed by the two vectors
~r1 = (9; 0; 7) ; ~r2 = (0; 9; 13) (4.156)
and the origin. Show that

~ B;
(a) The four vectors A; ~ C;and
~ ~ are a spanning set of vectors in the two
D
dimensional space formed by the two vectors ~r1 and ~r2 .
(b) Show that the vectors ~r1 and ~r2 are a basis set of vectors for the two
dimensional vector space.
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)
in terms of the basis set of vectors

~r1 = (9; 0; 7) ; ~r2 = (0; 9; 13) (4.160)
4.10 Gram-Schmidt orthogonalization

We call a set of vectors orthonormal if they are mutually orthogonal (perpen-
dicular), and each vector is normalized (that is its norm is one). For example
the vectors x^; y^; z^ form an orthonormal set. Gram-Schmidt orthogonalization
is a systematic process of obtaining orthonormal basis vectors from a set of
basis vectors. Suppose we have basis vectors A, ~ B,
~ and C;~ we can obtain an
orthonormal set of basis vectors applying the following procedure:
~ you get the …rst orthonormal basis vector
1. Normalize A-
~
A
A^ = : (4.161)
~
A
2.
~ along the direction of A
(a) Find the component of the second vector, B; ~
~A = B
B ~ A^ A:
^ (4.162)
~ and normalize the resulting vector,

(b) Subtract this component from A
~
B ~ A^ A^
B
^=
B : (4.163)
~
B ~ A^ A^
B
This is the second orthonormal vector in the set. It is orthogonal to A^

and it has a unit magnitude.
4.10. GRAM-SCHMIDT ORTHOGONALIZATION 79
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)
~ and normalize the resulting vector,

(b) Subtract these components from C
~
C ~ A^ A^
C ~ B
C ^ B
^
C^ = : (4.165)
~
C ~ A^ A^
C ~ B
C ^ B
^
This is the third orthonormal vector in the set. It is orthogonal to both

A^ and B^ it also has a unit magnitude.
4. For a space of higher dimensions, we continue to implement this procedure.

Example 15 For the given set of basis vectors, use Gram-Schmidt method to
…nd an orthonormal set
~ = (0; 2; 0; 0) ; B
A ~ = (3; 4; 0; 0) ; C
~ = (1; 2; 3; 4) : (4.166)
~
Solution: Following the …rst step we normalize vector A
~
A
A^ = = (0; 1; 0; 0) : (4.167)
~
A
and using step 2(a), we have
B~A = B ~ A^ A^ = [(3; 4; 0; 0) (0; 1; 0; 0)] (0; 1; 0; 0) = (0; 4; 0; 0)

(4.168)
so that following step 2(b) the second orthonormal vector in the set can
be expressed
~
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) :
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
f1 (x) f2 (x) ::: fn (x)

f10 (x) f20 (x) ::: fn0 (x)
: : :
W = det 6= 0 (4.172)
: : :
: : :
f1n 1 (x) f2n 1 (x) :: fnn 1 (x)
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)
sin (!t) sin2 (!t)

= det
! cos (!t) 2! sin (!t) cos (!t)
= 2! sin2 (!t) cos (!t) ! sin2 (!t) cos (!t)
W = ! sin2 (!t) cos (!t) 6= 0

and the functions are independent.

1. Solve the following set of equations for the four unknowns by reducing the
matrix using elementary row operations..
x 2y 4 = 0; 2z + w = 5; w + y = 2; 3x 1 = 5z:
2. Solve the following dc circuit problems from the algebra-based physics

class using row reduction of matrices
(a) The circuit diagram for this problem is shown in the …gure below
(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!

1.
(a) Given the following sets of equations

1 p 1 p
x0 = x + y 3 ; y0 = x 3 + y ; and z 0 = z
2 2
and
00 1 p 00 1 0p 00
x = x0 + y 0 3 ; y = x 3 + y 0 ; and z = z 0
2 2
write each set as a matrix equation and then solve for x00 and y 00 in terms
of x and y by multiplying matrices. Theses equations represent rotations
of axes about the z-axis. By comparing the equations and matrices with
the rotation matrix derived in class, …nd the rotation angles and check
your result.
(b) Prove that the matrices corresponding to the two sets of equations given
in (a) above can correspond to rotation matrices by demonstrating that
they are orthogonal matrices.
2. Show that the function
~ ~r + 3
f (~r) = A
is linear or none linear function.
3. Show that the vector function
~
f (~r) = A ~r
is a linear or none linear vector function.

4. De…ne the integral with respect to x from 0 to 1; the object being operated
on are function of x. Show that this integral operator is linear or none
linear operator. (This means we want to show that the operator
Z1
() dx
0
is a linear or none linear operator.)

5.
(a) For the matrix 2 3
1 0 5i
A=4 2i 2 0 5
1 1+i 0
…nd the transpose, the inverse , the complex conjugate, and the transpose
conjugate.
(b) Show that the product AAT is a symmetric matrix.

1. For each of the following problems write and tow reduce the augmented
matrix to …nd out whether the given set of equations have exactly one
solution, no solution, or an in…nite set of solutions.
(a)
2x + y z = 2
4x + y 2z = 3
(b)
2x + 5y + z = 2
x + y + 2z = 1
x + 5z = 3
2. Find the rank of the matrix

0 1
1 1 4 3
B 3 1 10 7 C
M =B
@ 4
C:
2 14 10 A
2 0 6 4
3. Write the vectors

~ = (1; 4; 5); B
A ~ = (5; 2; 1); C
~ = (2; 1; 3); D
~ = (3; 6; 11)
as a linear combination of the vectors

~a = (9; 0; 7) ; ~b = (0; 9; 13) :
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
5. Show that the matrix

p p !
1+i 3 p3 (1 +
4 2 2p
i)
M= p
p3 (1 + i) 3+i
2 2 4
is a unitary matrix. That means we need to show that

1
M = MT :
Chapter 5
Introduction to di¤erential
calculus-I
In advanced undergraduate or graduate physics courses we often deal with dif-

ferential calculus. In this chapter we will get introduced to the basics in dif-
ferential calculus. These includes partial di¤erentiation and total di¤erential of
multivariable functions and its applications to real physical problems.
5.1 Partial di¤erentiation

Consider a well-de…ned function z(x; y) that de…nes the z coordinate of a point
in space. This function depends on two coordinates (x and y). This function is
di¤erentiable with respect to both x and y for all x and y. In the case where
z depends only one coordinates, for example z (x) ; the di¤erentiation of this
function with respect to x is given by
dz (x) z (x) z (x + x) z (x)

= lim = lim : (5.1)
dx x!0 x x!0 x
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
@z (x; y) f (x; y) f (x + x; y) f (x; y)

= lim = lim ...y is kept constant
@x x!0 x x!0 x
(5.2)
and
@z (x; y) f (x; y) f (x; y + y) f (x; y)

= lim = lim ...x is kept constant.
@y y!0 y y!0 y
(5.3)
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
@ f (x1 ; x2 ; :::xi + xi ::xn ) f (x1 ; x2 ; :::xi :::xn )

f (x1 ; x2 ; :::xn ) = lim
@xi xi !0 xi
...all x 6= xi are kept constant (5.4)
Example 1 Consider the function z(x; y) = x2 + y 2 , where x and y are inde-

pendent.
(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
Using the function z(x; y) = x2 + y 2 ; we can evaluate the partial di¤erenti-

ations at point (x; y) = (1; 0) as:
@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
Important notation: In some branches of physics, such as in thermodynamics

and statistical physics, a function may depend on set of variables which are not
necessarily independent as it may be related by one or more variables. However,
it may physically be important to …nd the partial di¤erentiation with respect
to one variable by keeping one or more of the other variables constant. In such
cases, the partial di¤erentiation is expressed using subscripts for the variables
that need to be kept constant. For example
@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.
Example 2a For z (x; y) = x2 + 2y 2 ; x = r cos ( ) ; and y = r sin ( ) ;…nd the

following partial derivative
@z
: (5.8)
@r x
Solution: First the function z (x; y) needs to be expressed as z (x; r) : To this

end, we note that
p
x2 + y 2 = r2 ) y = r2 x2 (5.9)
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
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
Solution: Using the ideal gas equations
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
@2P @ 2 P (V; T ) @ @P (V; T )

= =
@V 2 T @V 2 T @V @V
@ nRT 2nRT
= 2
= ; (5.17)
@V V V3
@2P @ 2 P (V; T ) @ @P @ nR
= = = = 0:(5.18)
@T 2 V @T 2
V @T @T @T V
5.2. TOTAL DIFFERENTIAL 89
5.2 Total di¤erential

Consider a straight wire with length L that lies along the positive x-axis with one
end at the origin. Imagine you glued a charge that is not uniformly distributed
along the wire. That means the charge Q on the wire depends on the coordinate
x, (Q (x)). Thus the in…nitesimal charge Q (x) on the wire over the length in
between points x and x + x depends on the length ( x) and how the charge
changes over x ( Q(x) x ),
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
H H = 7:6 0:2cm; R R = 3:7 0:2cm (5.23)
Find an approximate value for the uncertainty in the cylinder’s volume,

dV:
Solution: We recall that volume of a cylinder is given by
V (R; H) = R2 H (5.24)
so that the total di¤erential can be expressed as

@V @V
dV = dR + dH (5.25)
@R @H
replacing dV; dR; and dH by V; R; and H; respectively, we may ex-

press the uncertainty in volume as
@V @V
V = R+ H: (5.26)
@R @H
Evaluating
@V @
= R2 H = 2 RH (5.27a)
@R @R
@V @
= R2 H = R2 (5.27b)
@H @H
at H = 7:6 and R = 3:7; we have
@V
= 176:6 (5.28a)
@R H=7:6;R=3:7
@V
= 42:9 (5.28b)
@H H=7:6;R=3:7
so that using these results along with R = 0:2 and H = 0:2; the
uncertainty in the volume becomes
V = 176:6 0:2 + 42:9 0:2 = 44:76 (5.29)
Since uncertainty is expressed using one signi…cant …gure, we must write
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)
5.3 The multivariable form of the Chain rule

Function of one variable: In a function of a single variable, y (x), the variable
x could be a function of another variable such as time, t; (x (t)). In such cases
to …nd the time dependence of the function y (x) ; we need to use the chain rule
for a function of one variable. We recall from the previous section, the total
di¤erential for y (x) can be written as
dy
dy = dx (5.33)
dx
so that diving this by dt, we …nd the Chain rule for a function of single variable
dy dy dx
= : (5.34)
dt dx dt
5.3. THE MULTIVARIABLE FORM OF THE CHAIN RULE 91
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.
Example 4 Consider the function
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)
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
Fx = max and Fy = may (5.45)
Since for a projectile motion there is no force on the x direction, Fx = 0;

we …nd
dvx dx
ax = = 0 ) vx (t) = v0x ) = v0x ) x (t) = x0 + v0x t
dt dt
) x (t) = x0 + v0 cos ( ) t: (5.46)
This shows that the x coordinate depends on time t. In the y direction

there is gravitational force, Fy = mg that leads to
dvy dy
ay = = g ) vy (t) = v0y gt ) = v0 sin ( ) gt: (5.47)
dt dt
Using the chain rule for a function of one variable, one can write
dy dy dx
= : (5.48)
dt dx dt
From the result in Eq. (5.46), we have
dx x (t) x0
= v0 cos ( ) and t = (5.49)
dt v0 cos ( )
so that substituting these equations along with the result in Eq. (5.47),
one may write Eq. (5.48) as
x (t) x0 dy
v0 sin ( ) g = v0 cos ( ) (5.50)
v0 cos ( ) dx
There follows that
dy g
= v0 tan ( ) (x x0 ) : (5.51)
dx v0 cos2 ( )
Upon integrating this equation with respect to x,
Z Z
g
dy = v0 tan ( ) (x x0 ) dx (5.52)
v0 cos2 ( )
we …nd
g x2
y (x) = y0 + v0 tan ( ) x x0 x : (5.53)
v0 cos2 ( ) 2
5.3. THE MULTIVARIABLE FORM OF THE CHAIN RULE 93
Assuming the initial position of the projectile is at the origin (x0 = 0; y0 = 0) ;

this can be put in the form
y (x) = bx ax2 ; (5.54)
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)
Solution: The easy way to …nd both
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
Then at (x; y) = (3; 1); we …nd
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 )
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

2
d2 y 1 2yx 1 2xy
= +
dx2 x 3y 2 x2 y 3y 2 x2
" #
2
1 2xy 3 2xy dy
+ (5.66)
y 3y 2 x2 x 3y 2 x2 dx
and using the result in Eq. (5.59), we …nd

" #
2 2
d2 y 1 dy 1 dy 1 dy 3 dy 2xy
= + + (5.67)
dx2 x dx y dx y dx x dx 3y 2 x2
Recalling the result we obtain for Eq. (5.59) at (x; y) = (3; 1)

dy
=1 (5.68)
dx
we …nd from Eq. (5.122)
d2 y 1 8
= 1 1 1= : (5.69)
dx2 3 3
5.4 Extremum (Max/Min) problems

For a function of single variable y (x), the …rst derivative
dy
f (x) = (5.70)
dx
tells us the slope of the tangent line at the point on the curve with the coordinate
(x; y (x)). The tangent line to an extremum point on a curve is a horizontal
line, the slope of which is zero. Therefore, if the x coordinates of the extremum
point on the curve de…ned by y (x) is xi ; since the slope of the tangent line at
(xi ; y (xi )) is zero,
dy
f (xi ) = = 0: (5.71)
dx x=xi
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.
Figure 5.1: The aquarium has length l, width w and height h:
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
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
5.5 The Method of Lagrangian Multipliers

In the Example 7 the volume of the rectangular aquarium, which is a function
of three variables, is speci…ed,
V (l; w; h) = lwh = V0 : (5.85)
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)
@f (x; y; z) @f (x; y; z) @f (x; y; z)

df = 0 ) dx + dy + dz = 0 (5.87)
@x @y @z
and the necessary and su¢ cient condition is
@f (x; y; z) @f (x; y; z) @f (x; y; z)

= = = 0: (5.88)
@x @y @z
Noting that for the constraint
@ @ @
d (x; y; z) = dk = 0 ) dx + dy + dz = 0 (5.89)
@x @y @z
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
F (x; y; z) = f (x; y; z) + (x; y; z) = f (x; y; z) + k ) dF = df = 0

@f @ @f @ @f @
) + dx + + dy + + dz = 0
@x @x @y @y @z @z
There follows that

@f @ @f @ @f @
+ = 0; + = 0; and + = 0: (5.90)
@x @x @y @y @z @z
Here we want x; y; and z so need not be determined. This method would fail
if the coe¢ cient of vanish at the extremum,
@ @ @
= = = 0:
@x @y @z
The following steps can be used to …nd extremum points for a function f (x; y; z; :::)
with a constraint (x; y; z; :::) = k
(1) Construct the function F (x; y; z; :::) = f (x; y; z; :::) + (x; y; z; :::), where
is the (unknown) Lagrangian multiplier.
(2) Compute the partial derivatives of F with respect to each of the variables
x; y; z; ::: and set them equal to zero.
(3) Solve the resulting system of simultaneous equations generated in Step 2
along with the constraint equation
(x; y; z; :::) = constant
Example 8 Consider a wire bent to …t a curve de…ned by y = 1 x2 as shown

in the …gure below.
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
(3) Solve the resulting system of simultaneous equations generated in Step 2

along with the constraint equation
(x; y) = constant
to determine the coordinates (x; y) at which the function f (x; y) takes a

maxima or minima values:
1 y
p = 2 ;p = ; y + x2 = 1 (5.95)
x2 + y2 x2 + y2
These three equations can be combined to …nd the values for x and y
1 2y
p = p ;y = 1 x2 ) 1 = 2 1 x2 (5.96)
x2 + y2 x + y2
2
1 1
) x = p ;y = (5.97)
2 2
Then the minimum distance is found to be
p
dmin = 3=2: (5.98)
Example 8 (a) The aquarium problem : 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: In order to determine the minimum area of the glasses used using
the method of Lagrangian multiplies we need to write the constraint. The
constraint is the volume
V0 = lwh ) (l; w; h) = lwh = V0 = constant. (5.99)
The surface area of the aquarium is
A (l; w; h) = lw + 2 (wh + lh) ) f (l; w; h) = lw + 2 (wh + lh) : (5.100)
Constructing the function F (l; w; h);
F (l; w; h) = f (l; w; h)+ (l; w; h) ) F (l; w; h) = lw +2 (wh + lh)+ lwh:

(5.101)
We then minimize this function by partially di¤erentiation it with respect

to l; w; and h
@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)
to determine l; w; and h: Since we do not need the solution for ; from

Eq. (5.102) we can express as
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
Now substituting these results in Eq. (5.105), we …nd

1=3
V0 1=3
4h3 = V0 ) h = ) w = l = 2h = (2V0 ) (5.109)
4
Example 8 (b) Particle in a box-Quantum mechanics:Consider the quantum

mechanical problem of a particle with mass m in a box. The box is a
rectangular box with length l; width w; and height h: The ground state
energy energy is given by
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
Solution: Constructing the function F (l; w; h);

1 ~ 1 1
F (l; w; h) = f (l; w; h)+ (l; w; h) ) F (l; w; h) = 2
+ 2 + 2 + lwh:
l 4m w h
(5.110)
We then minimize this function by partially di¤erentiation it with respect
to l; w; and h
@F (l; w; h) ~
=0) + wh = 0; (5.111)
@l 2ml3
@F (l; w; h) ~
=0) + lh = 0; (5.112)
@w 2mw3
and
@F (l; w; h) ~
=0) + lw = 0: (5.113)
@h 2mh3
We solve the above three equations along with the constraint
lwh = V0 : (5.114)
to determine l; w; and h: Since we do not need the solution for ; using
the …rst equation, we have
~
= (5.115)
2ml3 wh
so that using this expression for ; one can write
~ ~ ~
3
+ lh = 0 ) 3
= )w=l (5.116)
2mw 2mw 2ml2 w
and
~ ~ ~
3
+ lw = 0 ) 3
= )h=l (5.117)
2mh 2mh 2ml2 h
Now substituting these results into the constraint for the volume, we …nd
1=3 1=3
l3 = V0 ) l = (V0 ) ) w = l = h = (V0 ) (5.118)
The box must be a cube, w : l : h = 1 in order to minimize the energy of
the particle.
5.6 Change of Variables

Physical problems in most branches of physics requires solving di¤erential equa-
tions. These di¤erential equations could be a function of single or multivariable.
In order to …nd the solutions to these di¤erential equations change of variables
sometimes is necessary for di¤erent reasons. In order to change the di¤erential
equation from one variable to another when the di¤erential equation is a func-
tion of multivariable, we need to apply partial di¤erentiation of the function. In
the next examples we illustrates application of partial di¤erentiation in change
of variables in a second order di¤erential equation for single and multivariable
function.
Example 9 The Bessel di¤ erential equation is given by:

d2 y dy
x2 2
+x (1 x) y = 0 (5.119)
dx dx
In this equation y = y(x). Find the corresponding equation with the
following change of variable:
p
u = 2 x; y = y (u) (5.120)
Solution: Using partial di¤erentiation, we can express

dy @y du
= (5.121)
dx @u dx
since y = y (u) that depends only in one variable u, we may rewrite
dy dy du dy 1
= = p : (5.122)
dx du dx du x
Using the result above the second derivative in the bessel equation can be
expressed as
d2 y d dy 1 1 d dy dy d 1
= p =p + p : (5.123)
dx2 dx du x x dx du du dx x
Noting that
d dy d dy du d dy d2 y 1
= ) = 2p (5.124)
dx du du du dx dx du du x
and
d 1 1
p = p (5.125)
dx x 2x x
we …nd
d2 y 1 d dy dy d 1 1 d2 y 1 dy
2
=p + p = p : (5.126)
dx x dx du du dx x x du2 2x x du
Then substituting Eqs. (5.122) and (5.126) into Eq. (5.119) the Bessel
equation takes the form
1 d2 y 1 dy dy 1
x2 p +x p (1 x) y = 0 (5.127)
x du2 2x x du du x
which can simpli…ed into
p
d2 y x dy
x 2
+ (1 x) y = 0 (5.128)
du 2 du
From Eq. (5.120) we have x = u2 =4 and the Bessel equation becomes
u2 d2 y u dy u2
+ 1 y=0
4 du2 4 du 4
d2 y dy
) u2 2 + u + u2 4 y = 0: (5.129)
du du
In classical or quantum mechanical studies of the electrical or magnetic prop-

erties of particles or systems in physics often involve solving a second order
partial di¤erential equation for some form of multivariable a scalar or vector
function. One of these that commonly used is the Laplace equation. The
Laplace equation in Cartesian coordinates for some scalar function V (x; y; z)
is given by
@ 2 V (x; y; z) @ 2 V (x; y; z) @ 2 V (x; y; z)
+ + = 0:
@x2 @y 2 @z 2
In order to determine the solution to such kind of second order partial di¤er-
ential equations it may be necessary to change the variables from Cartesian
to cylindrical (r; '; z) or to spherical (r; ; ') coordinates if the system has a
cylindrical or spherical symmetry, respectively. Such change of variables for the
given partial di¤erential equation are determined by the equations that relates
a point in space de…ned by the Cartesian coordinates to Cylindrical coordinates
x = r cos ; y = r sin ; z = z (5.130)
and to spherical coordinates
x = r cos ' sin ; y = r sin ' sin ; z = r cos . (5.131)
In the next example we will see how we make change of variables for the two
dimensional Laplace equation in Cartesian to Cylindrical coordinates.
Example 10 In classical electrodynamics in a two-dimensional space where

there is no free charges the electrical potential, V (x; y) ; satis…es the two-
dimensional Laplace Equation given by
@ 2 V (x; y) @ 2 V (x; y)
+ =0 (5.132)
@x2 @y 2
Suppose the two dimensional space is a dielectric disk of radius R and you
are given a boundary condition for the potential at the edge of the disk is
zero V (x; y) = 0; for x2 + y 2 = R2 . For such kind of problem it is much
easier if we solve the Laplace equation in polar coordinates. Transform
the Laplace equation into polar coordinates using
x = r cos ( ) :y = r sin ( ) ; V = V (r; ) : (5.133)
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
Similarly, for partial di¤erentiation with respect to ; we have
@V (r; ) @V (x; y) @x @V (x; y) @y

= +
@ @x @ @y @
@V (x; y) @V (x; y)
= r sin ( ) + r cos ( ) (5.136)
@x @y
dividing this equation by r; one …nds
1 @V (r; ) @V (x; y) @V (x; y)

= sin ( ) + cos ( ) : (5.137)
r @ @x @y
Now multiplying Eq. (5.135) by cos ( ) and Eq. (5.137) by sin ( ) ; we

…nd
@V (r; ) @V (x; y) @V (x; y)
cos ( ) = cos2 ( ) + sin ( ) cos ( ) (5.138)
@r @x @y
and
sin ( ) @V (r; ) @V (x; y) @V (x; y)
= sin2 ( ) + sin ( ) cos ( ) : (5.139)
r @ @x @y
Subtracting Eq. (5.139) from Eq. (5.138) one …nds
@V (x; y) @V (r; ) sin ( ) @V (r; )

= cos ( ) : (5.140)
@x @r r @
Similarly,multiplying Eq. (5.135) by sin ( ) and Eq. (5.137) by cos ( ) ;
we have
@V (r; ) @V (x; y) @V (x; y)
sin ( ) = sin ( ) cos ( ) + sin2 ( ) (5.141)
@r @x @y
and
cos ( ) @V (r; ) @V (x; y) @V (x; y)
= sin ( ) cos ( ) + cos2 ( ) : (5.142)
r @ @x @y
and adding these two equations, we …nd
@V (x; y) @V (r; ) cos ( ) @V (r; )

= sin ( ) + : (5.143)
@y @r r @
From Eqs. (5.140) and Eq. (5.143) follows that
@ @ cos ( ) @
= sin ( ) + ; (5.144)
@y @r r @
@ @ sin ( ) @
= cos ( ) (5.145)
@x @r r @
so that one can write
@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 @
On carrying out the partial di¤erentiation in the …rst square bracket on

the second square bracket in Eq. (5.146) we have
@2 @2 @ 1 @ cos ( ) @ @
= sin2 ( ) + sin ( ) cos ( ) + sin ( )
@y 2 @r2 @r r @ r @ @r
cos ( ) @ @
+ cos ( ) (5.148)
r2 @ @
so that upon performing the di¤erentiations in the square brackets, one

…nds
@2 @2 sin ( ) cos ( ) @ 2 sin ( ) cos ( ) @

2
= sin2 ( ) 2 +
@y @r r @r@ r2 @
2 2
cos ( ) @ sin ( ) cos ( ) @ sin ( ) cos ( ) @ cos ( ) @ 2
2
+ + + :
r @r r @ @r r 2 @ r2 @ 2
(5.149)
that simpli…es into
@2 @2 2 sin ( ) cos ( ) @ 2 2 sin ( ) cos ( ) @

= sin2 ( ) +
@y 2 @r 2 r @r@ r2 @
2 2 2
cos ( ) @ cos ( ) @
+ + : (5.150)
r @r r2 @ 2
Following a similar procedure in Eq. (5.147), we have
@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 sin ( ) cos ( ) @ 2 sin ( ) cos ( ) @ 2

= cos2 ( ) +
@x2 @r 2 r2 @ r @r@
2 2
sin ( ) @ sin ( ) @ 2
+ + : (5.152)
r @r r2 @ 2
Now adding Eqs. (5.150) and (5.152), we …nd
@2 @2 @2 1 @ 1 @2
+ = + + (5.153)
@x2 @y 2 @r2 r @r r2 @ 2
@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
5.7 Legendre Transformations

Legendre Transformation of a function f (x) denoted by g (p) is given by
g (p) = xp f (x) (5.156)
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:
P = pressure; V = volume; T = temperature(K); S = entropy

W = work done on the system; Q = heat energy into the system;
The First Law of Thermodynamics:
dU = dQ + dW = T dS P dV (5.159)
Internal Energy and the Thermodynamic Potentials:
U = Internal Energy; U = U (S; V ); H = Enthalpy, H = H(S; P );

F = Helmholtz F ree Energy; F = F (T; V )
Example 11 Demonstrate a Legendre Transformation on the First Law of

Thermodynamics by deriving the form of the thermodynamic potential
Helmholtz Free Energy such that F = F (T; V ).
Solution: Let the internal energy, U , is a function of entropy, S; and volume,

V: Then the total di¤erential for the internal energy can be written as
@U @U
dU (S; V ) = dS + dV: (5.160)
@S V @V S
Comparing this equation with the …rst low of thermodynamics
dU = T dS P dV (5.161)
we …nd
@U
= T; (5.162)
@S V
@U
= P: (5.163)
@V S
The Helmholtz free energy which is a function of T and V , can then be

written as
@U @U
F = F (T; V ) = F ;V =F ; (5.164)
@S V @S V
which is the Legendre Transformation of the internal energy U
@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)

1. For the function p
z (u; v; w) = ln u2 + v 2 + w2
…nd the following partial derivatives:
@z @z @z
; ; and :
@u @v @w
2. For the function

u (x; y) = ex cos (y)
verify that
(a)
@2u @2u
=
@x@y @y@x
(b)
@2u @2u
+ 2 = 0:
@x2 @y
If
z (x; y) = x2 + 2y 2 ; x = r cos ( ) ; y = r sin ( )
…nd the following partial derivatives
3.
@z
@ x
4.
@z
@y@
5. Use di¤erential to show that for large n and small a
p p a
n+a n' p :
2 n

1. The acceleration of gravity can be found from the length l and period T
of a pendulum; the formula is g = 4 2 l=T 2 : Find the relative error in g in
the worst case if the relative error in l is 5%, and the relative error in T
is 2%:
2. For an idea gas of N molecules, the number of molecules with speeds v
is given by the formula
Z
4a3 N v 2 a2 x2
n (v) = p x e dx
0
where a is a constant and N is the total number of molecules. If N = 1026 ,

estimate the number of molecules with speeds between v = 1=a and 1:01=a
y dz
3. Given z = xe ; x = cosh (t) ; y = cos (t) ; …nd dt :
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
In the previous chapter we got introduced to the basics in di¤erential calculus

and its applications. Generally, we introduced to how to determine the di¤er-
entiation of functions of single and multivariable variables and its application
maximization or minimization in physical problems. In this chapter, we will
introduce to a sort of the reverse process in di¤erential calculus and its applica-
tion in real physical problems. We will learn the di¤erent methods in di¤erential
calculus used to determine a single or multivariable variable functions satisfying
a di¤ erential equation. A di¤erential equation (DE) is an equation involving
derivatives of functions with respect to one or more independent variables. In
real physical problems we could …nd linear or none linear di¤erential equations
for scalar (e.g. an electrical potential) or vector (e.g. an electric …eld) func-
tions. In real physical problems we could …nd linear or none linear di¤erential
equations for scalar (e.g. an electrical potential) or vector (e.g. an electric
…eld) functions. In this chapter we will consider primarily focus only in linear
di¤erential equations.
6.1 Linear ordinary di¤erential equations
As we just stated a di¤erential equation is an equation involving derivatives of

functions with respect to one or more independent variables. If there is only
one independent variable, then the equation is said to be an ordinary di¤ erential
equation (ODE). The order of the ODE or any DE is the order of the highest-
order derivative in the DE. An nth order ODE for the function y (x) (with x as
109
110 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II
independent and y as dependent variable) has the generic form
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.
6.1.1 First-order LODE with constant coe¢ cients

First-order LODE with constant coe¢ cients, generally, has the form
dy (x)
a0 y (x) + a1 = f (x) (6.3)
dx
The solutions to …rst order LODE can generally be determined by a method of
integration. For a1 6= 0; one can put Eq. (6.3) in the form
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

d h a0 i 1 a0
y (x) e a1 x = f (x) e a1 x : (6.8)
dx a1
Integrating Eq. (6.8), we may write,
Z x Z x
d h a0 0 i
a1 x
1 a0 0
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
where we have replaced the result of the integral when it is evaluated at x0 by

a constant C (x0 ) : Thus Eq. (6.10) can be expressed as
Z
a0
a1 x
a0
a1 x0
1 a0
y (x) e = y (x0 ) e C (x0 ) + f (x) e a1 x dx: (6.11)
a1
Thus by introducing another constant C for
a0
C = y (x0 ) e a1 x0 C (x0 ) ; (6.12)
the general solution to the …rst-order LODE in Eq. (6.3) is given by

a0
x Z
a0
x e a1 a0
y (x) = Ce a1
+ f (x) e a1 x dx: (6.13)
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.
Example 1 An object is dropped from rest and has a vertically downward

acceleration whose magnitude is given by a = g exp ( kt), where k is a
(real) positive constant and g is the acceleration due to gravity at the
earth’s surface.
(a) Find the general solution for the distance fallen as a function of time.
(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
From Eqs. (6.15) and (??), we note that
a0 = 0; a1 = 1; f (t) = g exp ( kt)
and the solution becomes

Z t
v (t) = C1 + g exp ( kt0 ) dt0 ; (6.18)
0
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
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
so that substituting the values in Eq. (6.22), we have

Z th i
g g
y (t) = C2 + C1 + exp ( kt0 )) dt0 ; (6.24)
t0 k k
and upon carrying out the integration, one …nds

g g
y (t) = C2 + C1 + t + 2 [exp ( kt) 1] (6.25)
k k
(b) We are given the initial conditions (boundary conditions) (v (t = 0) = 0)

and assume that the object is dropped from an initial position (y (t = 0) = y0 ) :
Then evaluating Eqs. (6.19) and (6.25) at t = 0, we …nd
C1 = 0; C2 = y0
and the distance fallen as function of time becomes

g g
d (t) = y0 y (t) = t + 2 [1 exp ( kt)] (6.26)
k k
and the speed
g
v (t) = (1 exp ( kt)) (6.27)
k
(c) For small t (using Taylor series expansion), we have
1 1 2
exp ( kt) ' 1 kt + kt2 ) 1 exp ( kt) ' kt kt (6.28)
2 2
so that
g g 1 2 1
d (t) ' t + 2 kt kt ) d (t) ' gt2 : (6.29)
k k 2 2
and the speed
g 1 2
v (t) '
(kt kt ) ) v (t) ' gt (6.30)
k 2
where we have dropped the higher order in t. For large time; the velocity
g
v (t) = (1 exp ( kt)) (6.31)
k
becomes
g
v (t) ' = v0 : (6.32)
k
For the distance

g g
d (t) = y0 y (t) = t + 2 [1 exp ( kt)] ; (6.33)
k k
if we let the large time be t = t0 ; we …nd
g
d (t) = d (t) = y0 y (t) = v0 t0 + : (6.34)
k2
Example 2 A conducting bar illustrated in the …gure below moves on two

frictionless parallel semi-in…nite rails connected by a conducting wire of
length l in the presence of uniform magnetic …eld directed into the page.
The bar has mass m and its length is also l. The bar is given an initial
velocity, v0 ; at t = 0 directed to the right.
(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.
(b) Find the magnetic force on the conducting bar.
(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:
(a) According to Faraday’s law a changing magnetic ‡ux, ; induces voltage,

Vin given by
d
Vin = : (6.35)
dt
Over a time interval t = t and t = t+dt, the position of the conducting rod
changes from x0 to x0 +dx, as the as it slides along the positive x-direction
with a speed v (t). This distance can expressed as
dx = vv (t) dt: (6.36)

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
dA = ldx = lvdt: (6.37)
The magnetic ‡ux, ; for a uniform magnetic …led B directed perpendic-

ular to the area, A; bounded a closed conducting wire can be expressed
as
= BA: (6.38)
Since the change in the magnetic ‡ux with time is due to the change in
area, we can write
d = BdA: (6.39)
so that using Eq. (6.37) one …nds for the change ‡ux over the time interval
dt
d = Blvdt: (6.40)
The induced voltage due to the change in magnetic ‡ux with time becomes
d
Vin = = Blv: (6.41)
dt
Applying Ohm’s law the resulting induced current, Iin can be expressed
as
Vin Blv
Iin = = : (6.42)
R R
According to Lenz’s law the induced current must oppose the cause for
the induction of the current. The cause is increase in magnetic ‡ux. To
oppose this increase, the induced current must ‡ow in a counterclockwise
direction.
(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
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
Solution: The change in intensity due to absorption is related to the density

of the atmosphere by
dI (z) = K (z)dz (6.53)
where K is a constant of proportionality.

Z I(z) Z z Z z
dI (z) = K (z)dz ) I (z) = I0 K (z)dz (6.54)
I0 0 0
6.2 The …rst-order DE and exact di¤erential

In the previous section, we have seen how the solution to a …rst-order LODE
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
dF (x; y) = P (x; y) dx + Q (x; y) dy = 0; (6.57)
which is the total di¤erential of some constant function, F (x; y) =constant,

that depends on both x and y if
@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.
Example 4 Find dF for the function
F (x; y) = 3x2 + 2x sin (2y) : (6.59)

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
dF = [6x + 2 sin (2y)] dx + [4x cos (2y)] dy (6.63)
In Example 4 we found the total di¤erential
dF = P (x; y) dx + Q (x; y) dy; (6.64)
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
dF = P (x; y) dx + Q (x; y) dy; (6.68)
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
(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
x3 + 2y dy = 3x (2 xy) dx: (6.72)
This can be put in the form
dF (x; y) = P (x; y) dx + Q (x; y) dy = 0; (6.73)
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)
This equation can be rewritten as

F (x; y) = 3x2 x3 y + C (y) ; (6.82)
where we introduced the function
C (y) = F (x0 ; y) 3x20 + x30 y; (6.83)
is a constant with respect to x but not with respect to y. Now substituting
Eq. (6.82) into
@F
Q (x; y) = = x3 + 2y ; (6.84)
@y
we …nd
@
3x2 x3 y + C (y) = x3 + 2y
@y
dC (y) dC (y)
) x3 + = x3 2y ) = 2y (6.85)
dy dy
Now integrating with respect to y,
Z y Z y
dC (y 0 ) 0
dy = 2 y 0 dy 0 ; (6.86)
y0 dy 0 y0
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)
(c) The equation

3x2 x3 y y2 = C (6.89)
can be rewritten as
y 2 + x3 y 3x2 + C = 0; (6.90)
or
ay 2 + by + c = 0; (6.91)
which is a quadratic equation with
a = 1; b = x3 ; c = 3x2 + C: (6.92)
Using the well known solutions for a quadratic equation,
p
b b2 4ac
y= (6.93)
2a
we …nd
1h 3 p 6 i
y (x) = x x + 4 (3x2 C) (6.94)
2
Example 6 Solve the DE

dy
y2 3x2 e3y = 2xe3y + ex : (6.95)
dx
in the same ways as Example 6.
Solution:
(a) Multiplying the given …rst-order di¤erential equation by dx, we …nd
y2 3x2 e3y dy = 2xe3y + ex dx: (6.96)
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)
(b) For an exact di¤erential, we must be able to show that

@P (x; y) @Q (x; y)
= : (6.99)
@y @x
Using Eq. (6.98), we …nd
@P (x; y) @
= 2xe3y + ex = 6xe3y ;
@y @y
@Q (x; y) @
= y 2 3x2 e3y = 6xe3y ; (6.100)
@x @x
that shows
@P (x; y) @Q (x; y)
= : (6.101)
@y @x
Therefore, the di¤erential is an exact di¤ erential and there exists a func-
tion F (x; y) (=Constant, in this case) such that
@F @F
P (x; y) = ; Q (x; y) = : (6.102)
@x @y
This leads to
Z x Z x
@F 0 0
dx = 2x0 e3y + x0 ex dx0
x0 @x0 x0
) F (x; y) F (x0 ; y) = x2 e3y + xex x20 e3y + x0 ex0 + C1 (y) (6.103)
or
F (x; y) = x2 e3y + xex + C1 (y) :
where we introduced the function de…ned by
F (x; y) = x2 e3y + xex + C1 (y) : (6.104)

Substituting this result into

@F
= Q (x; y) = y2 3x2 e3y = 3x2 e3y y2 ; (6.105)
@y
we …nd
@ dC1 (y)
x2 e3y + xex + C1 (y) = 3x2 e3y y2 ) = y2 : (6.106)
@y dy
Integrating this equation with respect to y yields
Z
y3
C1 (y) = y 2 dy + K ) C1 (y) = + C2 ;
3
where C2 is a constant of integration with respect to y and it is independent
of both x and y. Now substituting this result into Eq. (6.104), the solution
to the exact di¤erential in Eq. (6.104) is found to be
1 3
F (x; y) = x2 e3y + xex y = C: (6.107)
3
Note that we have de…ned C representing all the constants independent
of the both variables x and y.
6.3 First-order DE and none exact total di¤er-

ential
We recall that the di¤erential equation
dF = P (x; y) dx + Q (x; y) dy (6.108)
is said to be an exact di¤ erential equation, when

@P (x; y) @Q (x; y)
= : (6.109)
@y @x
For an exact di¤ erential there exists a function, F (x; y); such that
@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
This ODE can be rewritten as

dy (x) a0 (x) f (x)
= y (x) (6.112)
dx a1 (x) a1 (x)
dy (x)
Q (x; y) = P (x; y) : (6.113)
dx
There follows the total deferential
dF (x; y) = P (x; y) dx + Q (x; y) dy = 0; (6.114)
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
which can put in the form
dF (x; y) = [V (x) P (x)y V (x) Q(x)] dx + V (x) dy = 0: (6.121)
We recall that for di¤erential to be an exact di¤erential the partial derivative of

the coe¢ cient of dx with respect to y and the partial derivative of the coe¢ cient
of dy with respect to x must be equal. This means for the di¤erential in Eq.
(6.121), we must have
@ @
[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)
where C is the constant of integration. Dividing this equation by the integration

factor; we …nd
R
Z h R i R
y(x) = e P (x)dx e P (x)dx Q(x) dx + Ce P (x)dx : (6.128)
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
the solution were shown to be

a0
x Z
a0
x e a1 a0
y (x) = Ce a1
+ f (x) e a1 x dx: (6.130)
a1
Noting that Eq. (6.129) can rewritten as Eq. (6.118)
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).
Example 7 For the …rst-order DE

dy
2x + 4x2 = y (6.135)
dx
(a) Express the DE in the form
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:
(a) Noting that the DE can be rewritten as
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
(b) Using the function P (x) ; we …nd

Z Z
1 1
P (x)dx = dx = ln x: (6.140)
2x 2
Substituting this result and Q (x) = 2x into Eq. (6.118), we have
Z h i
1 1 1
ln x
y(x) = e 2 e 2 ln x ( 2x) dx + Ce 2 ln x
p
Z p p
= 2eln x eln(1= x) xdx + Celn x (6.141)
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.
6.4 Higher-Order ODE’s

In the previous section we have introduced to various methods for solving a …rst-
order ODE. Next we will get introduced to how to solve higher-order (n > 2)
ODEs. We recall that an nth order ODE is given by
dy d2 y d3 y dn 1 y
a0 (x) y + a1 (x) + a2 (x) 2 + a3 (x) 3 ::::an 1 (x) n 1 (6.146)
dx dx dx dx
dn y
+an (x) n = f (x) :
dx
We will develop the methods for solving this equation in two parts. In the
…rst part we will consider various techniques for Homogenous ODE. Generally,
a Homogeneous DE is a DE in which the function, f (x) = 0. In the second
part, we will develop the techniques for none-Homogeneous ordinary di¤erential
equations (HODE) that will be built upon the techniques for HODE.
6.4. HIGHER-ORDER ODE’S 127
6.4.1 Homogeneous ODE’s with constant coe¢ cient

From Eq. (6.146) an nth order HODE when all the coe¢ cients, a0 (x) =
a0 ; a1 (x) = a1 ; a2 (x) = a2 ::::an (x) = an ; can be written as
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
y(x) = Aekx : (6.148)
Substituting this into Eq. (6.147), one can easily …nd
a0 + a1 k + a2 k 2 + a3 k 3 ::::an 1k
n 1
+ an k n Aekx = 0: (6.149)
There follows that
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.
Example 8 An inductor and capacitor are connected in a series circuit, as

shown. At t = 0, the charge on the capacitor plates is Qo and the current
‡owing in the circuit is Io . Find an expression for the charge on the
capacitor, Q(t).
Solution: Using Kircho¤’s rule we may write
Q (t) dI (t)
+L =0 (6.153)
C dt
and recalling that I(t) = dQ=dt, we may write

d2 Q 1
+ Q = 0: (6.154)
dt2 LC
Eq. (6.154) is a HODE with constant coe¢ cients. Thus, we can assume a
solution of the form
Q (t) = Aekt : (6.155)
Substituting this into Eq. (6.154), we …nd
d2
Aekt + ! 2 Aekt = 0 ) k 2 + ! 2 = 0 (6.156)
dt2
the solution of which is given by
k1 = i!; k2 = i! (6.157)
where we introduced the constant de…ned by
1
!=p : (6.158)
LC
The general solution can then be expressed as
Q(t) = A exp (i!t) + B exp ( i!t) : (6.159)
At t = 0 the capacitor is fully charged and the charge is Q0 ; this means
Q(0) = A exp (i!t) + B exp ( i!t)jt=0 = Q0 ) A + B = Q0 : (6.160)
We are also given that the current at t = 0 is I0 : Thus using the de…nition
of an electrical current, we may write
dQ
I(t = 0) = = I0
dt t=0
d I0
) [A exp (i!t) + B exp ( i!t)] = I0 ) A B= (6.161)
dt t=0 i!
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.
In Example 8, we saw that the solution to the second-order DE in Eq. (6.154)

can be expressed in four di¤erent forms given by Eqs. (6.164), (6.165), (6.168),
and (6.171). In view of this results, we can make two generalizations to the
solution of a second-order HODE that can be put in form
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)
are complex, k1 = i! and k2 = i! and the general solution can be expressed

in four di¤erent forms given by:
8
>
> A exp (i!x) + B exp ( i!x) ;
<
C cos (!x) + D sin (!x) ;
y (x) = (6.178)
>
> E sin (!x + ') ;
:
G cos (!x + ') :
For the minus case

d2 y
! 2 y = 0; (6.179)
dx2
the roots to the corresponding indicial equation
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:
Example 10 The Schröedinger Equation: The one-dimensional time-independent

Schröedinger equation for a particle traveling along the x-direction with a
potential energy function U (x) is given by
~2 d2 (x)
+ U (x) (x) = E (x); (6.182)
2m dx2
where ~ = h=2 (h is a constant called Planck’s constant), m is the mass of

the particle, and E is the total energy of the particle. Consider a particle
of mass m and energy E traveling in the potential shown in the …gure.
The x-axis is divided into two regions by the potential energy function.
Region I has U (x) = 0, which means that there are no forces acting on
the particle. In this region the particle is a free particle. Region II has
U (x) = Uo , where Uo is a constant such that Uo > E. This is a forbidden
region for the particle according to classical physics. The signi…cance of
the wave function (x) that is the solution to the Schröedinger equation
2
above for a given potential energy function is that j (x)j dx gives the
probability of …nding the particle within an interval dx at each value of x.
Solve the Schröedinger equation for the wave function of the particle in
each of the two regions I and II shown in the …gure. Comment on your
answers.
Solution: In region I, the potential energy is zero, U (x) = 0; and the Schröedinger
Equation becomes
~2 d2 (x) d2 (x) 2mE

= E (x) ) = (x) (6.183)
2m dx2 dx2 ~2
which we may write as
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)
we …nd the indicial equation

2
+ k2 = 0 ) = ik (6.187)
so that the solution can be written as
(x) = A1 eikx + A2 e ikx

: (6.188)
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)
2
q2 = 0 ) = q (6.193)
so that the solution can be written as
(x) = A1 eqx + A2 e qx
: (6.194)
But physical condition requires A1 = 0, which means

qx
(x) = A2 e : (6.195)
Indicial equation with degenerate roots: As we stated earlier the gen-

eral solution in Eq. (??) is valid when there are no repeated roots (none-
degenerate roots). However, it is possible that a HODE with constant coe¢ -
cients could lead to an indicial equation with degenerate roots. For example,
for the second-order HODE
d2 y dy
2a + a2 y = 0; (6.196)
dx2 dx
the roots of the quadratic indicial equation
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
y(x) = A + Bx + Cx2 ::: + Dxn 1

eax (6.201)
(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
Pm 1 1 (x) = C0 + C1 x + C2 x2 ::: + Cm1 1x

m1 1
; (6.203)
Pm 2 1 (x) = C00 + C10 x + C20 x2 ::: + 0
Cm2 1 xm2 1 ;
are polynomials of degree m1 1 and m2 1; respectively.

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:
(a) Upon substituting
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)
The is a cubic equation with three distinct roots
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
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
y (x) = A + Bx + Cx2 eax + Aek1 x ; (6.214)
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].
Example 12 Damped Simple Harmonic Motion: An ideal spring of spring con-

stant k hangs vertically from a support rod and has a mass m attached to
its lower end. The mass hangs suspended in a viscous ‡uid. The mass is
pulled down some distance from its equilibrium position at y = 0 and re-
leased. Discuss the resulting motion of the mass if the viscous ‡uid exerts
a damping force on it given by fy = bvy , where b is a damping constant.
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)
and the solutions

1 = + D; 2 = D; (6.221)
where p
D= 2 !2 (6.222)
Case 1 ( = !) we have
p
D= 2 !2 = 0 (6.223)
and the roots to the indicial equation becomes
1 = 2 = (6.224)
and the general solution is given by

t
y(t) = (At + B) e (6.225)
This describes a Critically damped motion

Case 2 ( > !) we note that D is real
p
D= 2 !2 < (6.226)
which means the solution is given by

( D)t ( +D)t
y(t) = Ae + Be : (6.227)
This shows over-damped motion

Case 3 ( < !) we may write

p p
D= 2 !2 = (! 2 2) =i (6.228)
where p
= !2 2 (6.229)
which means the roots for the indicial equation becomes
1 = +i ; 2 = i (6.230)
and the solution is given by

( i )t ( +i )t t
y(t) = Ae + Be ) y(t) = e Aei t
+ Be i t
(6.231)
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)
It shows under-damped oscillatory motion.
6.4.2 None homogeneous ODE’s with Constant Coe¢ cients

th
An n -order none homogeneous ODE with constant coe¢ cient has the form
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
y(x) = yh (x) + yp (x): (6.235)
4-th Any boundary conditions given to the problem must then be applied to
the general solution y(x).
Example 13 Find the general solution to the following di¤erential equation:
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)
the homogeneous solution can be written as
sh = Ae2t + Bet : (6.239)
Step 2 : By inspecting the none homogeneous ODE, I guess a particular

solution of the form
dsp d2 sp
sp = C ) = = 0; (6.240)
dt dt2
so that when we substitute it back into the none homogeneous ODE in
Eq. (6.236), we …nd
2C = 16 ) C = 8: (6.241)
Thus the particular solution is found to be
sp = 8: (6.242)
Step 3 : Write the general solution
s (t) = sh (t) + sp (t) ) s (t) = Ae2t + Bet + 8: (6.243)
Step 4 : Apply the boundary conditions if there are any. Here we are not
given any.
d2 R
+ 4R = sin (3x) (6.244)
dx2
Solution:
Step 1 : Find the homogeneous solution for
d2 R
+ 4R = 0: (6.245)
dx2
Noting that the roots to the indicial equation are
k 2 + 4 = 0 ) k1 = 2i; k2 = 2i (6.246)
the homogeneous solution can be written as
Rh (x) = A cos (2x) + B sin (2x) : (6.247)
Step 2 : By inspecting the given none homogeneous ODE, I made a guess

(an educated guess) to the particular solution that has the form
Rp = C sin (3x) : (6.248)
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.
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)
are
k1 = 2i; k2 = 2i: (6.255)
Then the homogeneous part of the solution can be expressed as
yh (x) = A cos (2x) + B sin (2x) ; (6.256)
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
There follows that

1
C1 = C2 =
: (6.262)
8
Therefore, the particular solution becomes
1
yp (x) = xe2ix + xe 2ix
; (6.263)
8
x e2ix + e 2ix
x
yp (x) = = cos (2x) : (6.264)
4 2 4
Step 3 : Write the general solution, y (x) = yh (x) + yp (x)

x
y (x) = A cos (2x) + B sin (2x) cos (2x) (6.265)
4
Step 3 : In this example also we are not given boundary conditions.
6.5 The method of superposition and the par-

ticular solution.
In the previous example the particular solution in Eq. (6.259) that we came up
with an educated guess that can be put in the form
yp (x) = yp1 (x) + yp2 (x) ; (6.266)
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)
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
that can be written as

X dn y
an = f1 (x) + f2 (x) : (6.275)
n
dxn
The particular solution can be determined the principle of superposition that

states
y (x) = yh (x) + yp1 (x) + yp2 (x) (6.276)
where yp1 (x) and yp2 (x) satisfy the ODE with the non-homogeneous terms
f1 (x) and f2 (x),
X dn y
an n = f1 (x) (6.277)
n
dx
and
X dn y
an = f2 (x) ; (6.278)
n
dxn
respectively.
From what we discussed in relation to the previous example, we can also
state that for a second-order Non-homogenous second-order ODE that can be
expressed as
d d
a b y (x) = f1 (x) = ecx Pn (x) ; (6.279)
dx dx
where Pn (x) is a polynomial of degree n; the particular solution of such kind of
non-homogenous LDE is given by
8 cx
< e Qn (x) ; ( if c 6= a & c 6= b)
yp (x) = xecx Qn (x) ; (if c = a or b & a 6= b) (6.280)
: 2 cx
x e Qn (x) ; (if c = a = b)
6.5. THE METHOD OF SUPERPOSITION AND THE PARTICULAR SOLUTION.143
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.
Example 18 Using the principle of superposition solve the following di¤eren-

tial equation:
d2 y
y = 2 sin (x) + 4x cos (x) (6.282)
dx2
Solution: For the 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)
For the particular solutions …rst we rewrite the DE
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
noting that
c = i 6= a; b (6.294)
we may write the particular solution as
yp1 (x) = (A + Bx) eix ; (6.295)
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)
Substituting these equations into
d2 yp1 (x)
yp1 (x) = f1 (x) = (2x i) eix : (6.298)
dx2
we …nd
2iBeix 2yp1 (x) = (2x i) eix ) 2iBeix 2 (A + Bx) eix

= (2x i) eix (6.299)
) [ 2Bx + 2Bi 2A] eix = (2x i) eix : (6.300)

There follows that
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!
IS THAT NOT MUCH EASIER?
6.6 The method of successive integration

In the previous section the methods introduced to determine the general solu-
tions whether for homogeneous or none homogeneous ODE with constant coe¢ -
cients is based on an educated guess. In this section, we will derive a general an
integral expression for the general solution to a second-order none homogeneous
ODE by successively integrating the ODE. The general solutions determined by
an educated guess in the previous sections will be shown that it can be derived
from this general integral expression. To this end, we put the second-order ODE
with constant coe¢ cients
d2 y dy
a2 2
+ a1 + a0 y = F (x) (6.311)
dx dx
in the form
d2 d d d
a2 + a1 + a0 y = a b y = F (x) ; (6.312)
dx2 dx dx dx
where a and b are the roots for a quadratic equation given by

p p
a1 a21 4a2 a0 a1 + a21 4a2 a0
a= ;b = : (6.313)
a2 a2
Introducing a new variable u (x) de…ned by
d
u (x) = b y (x) ; (6.314)
dx
Eq. (6.312) can be rewrite as
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)
where C1 is a constant of integration. Now substituting Eq. (6.319) into Eq.

(6.314), one can write
Z x
dy ax 0
by = e e ax F (x0 ) dx0 + C1 eax ; (6.320)
dx
bx
and multiplying by e , we …nd
Z x
bx dy bx (a b)x ax ax0
e be y=e e e F (x0 ) dx0 + C1 e(a b)x
: (6.321)
dx
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 again C2 is a constant of integration. Thus the solution to the DE in Eq.

(6.312) can be expressed as
Z " x Z 0 #
x
b)x00 ax0
y (x) = ebx e(a e F (x0 ) dx0 dx00 (6.324)
Z x
b)x00
+C1 ebx e(a dx00 + C2 ebx ;
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
y (x) = yH (x) + yp (x) ; (6.325)
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).
A. HODE with none degenerate roots: We recall that for a second-order

homogeneous ODE (F (x) = 0)
d2 y dy
a2 2
+ a1 + a0 y = 0; (6.328)
dx dx
that leads to an indicial equation
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
y (x) = C1 eax + C2 ebx : (6.330)
The result in Eq. (6.324), for F (x0 ) = 0; becomes

Z x
00
y (x) = C1 e(a b)x dx00 + C2 ebx ; (6.331)
and upon carrying out the integration, one …nds
C1 ebx e(a b)x C1 eax

y (x) = + C2 ebx = + C2 ebx
a b a b
= C1 eax + C2 ebx : (6.332)
where we have absorbed the constant 1/ (a b) in C1 : This is the same

as Eq. (6.330).
B. HODE with degenerate roots: For a homogeneous ODE that leads to an
indicial equation with degenerate roots (i.e. a = b), Eq. (6.331) becomes
Z x
y (x) = C1 ebx dx00 + C2 ebx = (C2 + C1 x) ebx ; (6.333)
which is in agreement with Eq. (6.199).

C. None HODE: We have seen that the particular solutions, yp (x) ; for the
second-order none HODE
d2 y dy d d
a2 2
+ a1 + a0 y = a b = ecx Pn (x) ; (6.334)
dx dx dx dx
with a polynomial Pn (x) of degree n; were determined based on an edu-

cated guess. For such none HODE, depending on the degeneracy of the
values for a; b; and c; we saw that the particular solutions are given by
8 cx
< e Qn (x) ; ( when c 6= a & c 6= b) ;
yp (x) = xecx Qn (x) ; (when c = a or b & a 6= b) ; (6.335)
: 2 cx
x e Qn (x) ; (when c = a = b) :
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
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
Then the particular solution in Eq. (6.327) can be expressed as,
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).
(i.) When c 6= a & c 6= b: Under this condition, =c a 6= 0: Under this

condition one can use the relation
di x0 x 0i
[e ]=e x ; (6.340)
d i
and show that

Z x00 Z x00
x0 0i 0 di x0 di x00
e x dx = e dx0 = e : (6.341)
d i 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
Upon carrying out the integration, we …nd
n
X
bx di e( + )x
yp (x) = e Ai : (6.343)
i=0
d i +
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
where we have introduced the summation index de…ned by j = i + 1.

Applying the relation in Eq. (6.340), one can easily show that
X Aj 1 dj Z x 00
n+1 n+1
X Aj 1 dj e x
bx x 00 bx
yp (x) = e e dx = e :
j=1
j
d j j=1
j
d j
(6.349)
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
Upon integrating this, we …nd

n+1
X Aj 1 j+1
yp (x) = ebx (x00 ) ; (6.353)
j=1
j (j + 1)

yp (x) = C2 x2 + C2 x3 + :::Cn xn+2 ebx = x2 ecx Qn (x) ; (6.354)
where Qn (x) is a polynomial of degree n. This result is also the same as
the solution in Eq. (6.335) when c = a =b.
Example 17 A forced Harmonic oscillator : Consider an ideal spring with
spring constant k = 8N=m with a mass m = 2kg attached to its one end
with the other end …xed. The spring and mass are sitting on a frictionless
surface. The mass is constantly acted by a periodic external force given by,
F~e = 2 sin (3t) x
^:What would be the amplitude of the oscillation. Initially
the mass is at rest with the spring at its equilibrium position. (i.e. the
spring is neither stretched nor compressed.)
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;
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)
From Eq. (6.356), we have

e3it e 3it
a = 2i; b = 2i; and F (t) = sin (3t) = ; (6.359)
2i
so that Eqs. (6.357) becomes
Z t
2it 00 C1 2it
xH (t) = C1 e e4it dt00 + C2 e 2it
= e + C2 e 2it
: (6.360)
4i
Using Euler’s formula the homogeneous solutions can be expressed as
C1
xH (t) = [cos (2t) + 2i sin (2t)] + C2 [cos (2t) 2i sin (2t)]
4i
C1 C1
= + 2iC2 cos (2t) + 2iC2 sin (2t) ; (6.361)
4i 4i
we can rewrite as this equation as
xH (t) = A cos (2t) + B sin (2t) : (6.362)
Substituting the values for a, b, and F (t) from Eq. (6.359) into Eq.
(6.358), the particular solution, becomes
Z t" Z t00 0
! #
2it 4it00 2it0 e3it e 3it0
xp (t) = e e e dt0 dt00
2i
Z " Z 00 #
e 2it t 4it00 t 0
= e eit e 5it0 dt0 dt00 : (6.363)
2i
Upon carrying out the inner integral, we have

" !#
2it Z t it00 5it00
e e 00 e
xp (t) = e4it + dt00
2i i 5i
Z " 5it00 00
!#
e 2it t e e it
= + dt00 ; (6.364)
2i i 5i
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
Therefore, with the given initial conditions, the position is given by
3 1
x (t) = sin (2t) sin (3t) : (6.370)
10 5
and the velocity

1
v (t) = [2 cos (2t) 3 cos (3t)] : (6.371)
5
Using Mathematica:
6.7 Partial Di¤erential Equations

In the previous sections we saw that physical properties of some real systems
can be described by homogeneous or none homogeneous ordinary di¤erential
equations (ODE) of unknown functions of one variable. The techniques for
solving such kind of di¤erential equation were also discussed in these sections.
Generally, most real physical systems or process are described by unknown
functions of two or more variables that must be determined from a homogeneous
or none homogeneous Partial Di¤ erential Equation (PDE ).
A Partial Di¤ erential Equation (PDE ). PDE is a di¤erential equation in-
volving partial derivatives of an unknown function with respect to two or more
independent variables. There are various physical processes that requires use of
partial di¤erential equations. Some are listed below:
1. The gravitation potential of a given mass distribution described by given

mass density or electrical potential of a given charge distribution with a
given charge density satisfy the partial di¤erential equation (for example
in Cartesian coordinates) given by
@ 2 V (x; y; z) @ 2 V (x; y; z) @ 2 V (x; y; z)

+ + = (x; y; z); (6.372)
@x2 @y 2 @z 2
where V (x; y; z) is the gravitational potential (electrical potential) of a

given mass (charge) distribution and (x; y; z) is the mass (charge) density.
If you know the density, then by solving this partial di¤erential equation
you can determine the potential at some point in space. This PDE is
known as the Poisson’s equation. If (x; y; z) = 0; it is called Laplace’s
equation
6.7. PARTIAL DIFFERENTIAL EQUATIONS 155
2. The Schrödinger equation:
~2 @ 2 (x; y; z; t) @ 2 (x; y; z; t) @ 2 (x; y; z; t)

+ +
2m @x2 @y 2 @z 2
@
+U (x; y; z) (x; y; z; t) = i~ (x; y; z; t); (6.373)
@t
where (x; y; z; t) is the wave function U (x; y; z) is the potential energy.
3. The wave equation:
@ 2 u(x; y; z; t) @ 2 u(x; y; z; t) @ 2 u(x; y; z; t) 1 @2

+ + = u(x; y; z; t);
@x2 @y 2 @z 2 v 2 @t2
(6.374)
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.
1st . Assume a product form for the solution.
2nd . Substitute the product function into the original PDE and simplify.
3rd . Divide both sides of the resulting DE by the product function.
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!
8th . Apply any BC’s to the general solution.
Example 18 Traveling Waves on a Stretched String:
(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
where s
T
v= ; (6.376)
T is the tension on the string, and is the linear mass density.
(b) Solve the 1-D traveling wave equation for a string of length L having both
ends …xed. Which means, apply the BC’s
u(y; t = 0) = u(y = 0; t) = u(y = L; t) = 0: (6.377)
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
transverse force on the segment between y and y + y can be expressed

as
F = T sin 0 sin ( ) (6.378)
If we assume the displacement is small, we can replace sin by tan,
0
F = T tan tan ( ) : (6.379)
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
We recall that the de…nition of the …rst derivative of a function u(y) at y
@u u(y + y) u(y)
= lim (6.382)
@y y=y
y!0 y
and the second derivative

2 3
@u @u
@2u 6 @y
y+ y
@y
y7
= lim 4 5 (6.383)
@y 2 y=y
y!0 y
" #
@2u @u @u
) y = lim (6.384)
@y 2 y=y
y!0 @y y+ y @y y
so that for small y; we can write
@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
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)
Equation (6.389) is the one-dimensional wave equation. It describes a

wave traveling in the y-direction with a speed, v; given by Eq. (6.390)
that depends on the magnitude of the tension on the string, T; and the
linear mass density of the string, :
(b) Let’s assume the solution of the PDE can be expressed as
u(y; t) = Y (y) T (t) (6.391)
so that we can write
d2 Y (y) Y (y) d2 T (t)
T (t) = (6.392)
dy 2 v2 dt2
dividing both sides by Y (y) T (t)
1 d2 Y (y) 1 d2 T (t) 2 d2 Y (y)
= = k ) + k 2 Y (y) = 0;
Y (y) dy 2 v 2 T (t) dt2 dy 2
d2 T (t)
+ k 2 v 2 T (t) = 0 (6.393)
dt2
The solutions to these HDEs is given by
Y (y) = A cos (ky) + B sin (ky) ; T (t) = C cos (kvt) + D sin (kvt) (6.394)
Therefore the general solution can be expressed as
X
u(y; t) = [Ak cos (ky) + Bk sin (ky)] [Ck cos (kvt) + Dk sin (kvt)]
k
(6.395)
Now we apply the boundary conditions. For u(y = 0; t) = 0; we have
X
u(0; t) = Ak (Ck cos (kvt) + Dk sin (kvt)) = 0 (6.396)
k
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
Then applying the second boundary conditions for u(y; t = 0) = 0; we …nd

X
u(y; t = 0) = (Bk sin (ky)) Ck = 0 ) Ck = 0: (6.399)
k
Thus the further simpli…ed solution can be rewritten as

X X
u(y; t) = Bk Dk sin (ky) sin (kvt) = Bk0 sin (ky) sin (kvt) (6.400)
k k
Then, applying the second boundary condition u(y = L; t) = 0; we have

X
u(L; t) = Bk0 sin (kL) sin (kvt) = 0 ) sin kL = 0
k
n
) kL = n ) k = ; where n = 0; 1; 2:::: (6.401)
L
Therefore, the general solution becomes
X
u(y; t) = Bk0 sin (kn y) sin (kn vt) : (6.402)
n=0
where
n
kn = ; n = 0; 1; 2:3::: (6.403)
L
This concludes the chapter to the introduction to di¤erential equations.

1. Boas 8.1 #6
2. Boas 8.1 #7
3. Boas 8.2 #6
5. Boas 8.2 #32

1. Consider the …rst-order di¤erential
dy
2x + y = 2x5=2
dx
(a) Show that the di¤erential equation is an exact di¤erential equation or not.
(b) Write the di¤erential equation in the form
dy
+ P (x) y = Q (x)
dx
and determine the solution using the relation derived for such di¤erential
equation.
(c) Verify your solution using computer programing (Mathematica is recom-

mended).
2. Do part (a)-(c) in problem 1 for the di¤erential equation
dy
+ y tanh x = 2ex
dx
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:

1. Find the general solution for the None-homogeneous DE
d2 y dy
2
+ 2y = e2x :
dx dx
d2 y dy
5 +6 + 2y = x2 + 6x:
dx2 dx
d2
+ 1 y = 8x sin x:
dx2
4. Using the method of superposition …nd the general solution to the DE

d2 y dy
5 + 6y = 2ex + 6x 5:
dx2 dx

1. In classical physics, the motion of a particle can be described for all time
once the “initial conditions” (position, velocity) of the particle’s motion
are known. The so-called “equation of motion” in this case is Newton’s
second law. In order to apply Newton’s second law, one must be able to
specify all of the forces acting on the particle in question. In quantum me-
chanics, the equation of motion is called the Schröudinger equation. This
equation is a partial di¤erential equation involving a function called the
wave function for the particle, denoted (x; t) (for 1-D problems), which
contains all observable information about the particle’s motion (position,
velocity, momentum, energy,...). Instead of precisely specifying all future
2
motion of the particle, in quantum mechanics the quantity j (x; t) is
used to predict the probability of a particular future motion. The 1-D
Schröudinger equation is given by
~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 :
Chapter 7
Integral calculus-scalar
functions
In this chapter we will be introduced to single and multiple integrals involving

scalar functions. We focus on the methods for carrying out such integrals. In
almost all branches of physics, integrations of scalar functions are often used
to determine various quantities describing the static or dynamic properties of
an object. The object could be in a microscopic particle like an electron in an
atom or massive objects like a planet in the solar system. The simplicity of
carrying out integrations of scalar functions depends on the geometry of the
object and system that object is existing in. The methods of integrations in
Cartesian coordinates for rectangular geometry and the methods of integrations
in curvilinear coordinates for curved geometry can greatly simplify the integra-
tions. We will be introduced the methods of integrations of single and multiple
integrals in Cartesian and curvilinear coordinates and its applications in real
physical problems.
7.1 Integration in Cartesian coordinates

We are interested in the methods of integrations of functions of single, two, and
three variables in Cartesian coordinates.
7.1.1 Single and multiple integrals

A. Single integral and arc length
Consider the two points P1 and P2 in the x-y plane shown in Fig. 7.1.1
described by the Cartesian coordinates (x1 ; y1 (x1 )) and (x2 ; y2 (x2 ))), respec-
tively. You took two paths to get from the …rst to the second point. The …rst
is a straight line path shown in red and the second is a curved path shown by
the black curve. The slope
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)
That can also be put in the form

s s
2 2
y x
l= x +1= y + 1: (7.6)
x y
7.1. INTEGRATION IN CARTESIAN COORDINATES 165
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)
The points along the length of the rectangle, l, are de…ned by the equation
x (y) = y1 = constant or x (y) = y2 = constant (7.12)
and those along the width, w, by
y (x) = x1 = constant or y (x) = x2 = constant. (7.13)
There follows that

dy dx
= 0; =0 (7.14)
dx dy
Thus the length and the width of the rectangle can be expressed using the
arc-length relation that can be expressed as
v" #
Zx2u
u dy 2 Zx2
l= t + 1 dx = dx (7.15)
dx
x1 x1
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
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:
(a) The height is the change in y coordinate ( y),

Z x2
y= m (x) dx; (7.22)
x1
where m (x) is the slope of the wire given by

dy (x)
m (x) = :
dx
The function de…ning the curve shown in the …gure is
p
y (x) = R2 x2
and slope becomes
dy (x) x
m (x) = = p : (7.23)
dx R2 x2
The distance traveled along the x-direction can then be written as
Z 0 p !
xdx 0 3
2 2 1=2
y= p = R x = 1 R: (7.24)
R=2 R 2 x 2 R=2 4
(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
Using the result in Eq. (7.23), we may write

v" #
ZRu
u 2 ZR
t x Rdx
l= p + 1 dx = p dx: (7.26)
R 2 x2 R 2 x2
0 0
Introducing the transformation de…ned by
x = R sin ( ) ) dx = R cos ( ) (7.27)
the length of the wire is found to be
Z=2 Z=2
R2 cos ( ) R
l= q d =R d = : (7.28)
2
R2 1 sin ( ) 2
0 0
Note that we have expressed the limits of integration in terms of the

variable : That means for x = 0
x = R sin ( ) ) =0
and x = R
1
x = R sin ( ) ) = sin (1) = =2
Homework problem: Using the expression

s
Zy2 2
dx
l= 1+ dy; (7.29)
dy
y1
show that you will …nd the same result.

(c) The area is given by the surface integral
ZZ
A= dxdy: (7.30)
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
Introducing the transformation of variable de…ned by

y = R sin ) dy = R cos d (7.32)
and noting that
y=0) = 0; y = R ) = =2 (7.33)
we may rewrite the area as
Z=2p Z=2
2 2 R2
A= R2 R2 sin R cos d = R cos2 d = : (7.34)
4
0 0
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
Integrating with respect to z,

8 9
Z < y=Z 2x+2
x=1
=
V = (2 + x + y) dy dx; (7.38)
: ;
x=0 y=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
and with respect to x; results in

Z
x=1
1
V = [6 6x] dx = 6x 3x2 0
= 3cm2 : (7.40)
x=0
7.1.2 Physical Applications

There are several physical quantities that require carrying out single and mul-
tiple integrals. Here are some examples in electrostatics and mechanics.
I. Electrostatics
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)
In view of Eq. (7.41), one can then express dV (r) as
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
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
This is a surface integral that requires multiple integration. Note that

the in…nitessimal area, da0 ; on the surface of the sphere with radius r0 , in
spherical coordinates can be expressed as
0
da0 = r02 sin d 0 d'0 (7.48)
Volume charge density: Similarly, for a charge described by a volume

charge density,
dq 0
(~r0 ) = 0 ) dq 0 = (~r0 )dv 0 ; (7.49)
dv
the electrostatic potential at a point ~r for the entire volume of charge can
be expressed as
ZZ Z
1 (~r0 )
V (r) = dv 0 : (7.50)
4 0 j~r ~r0 j
V olum
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
and
ZZ Z
Q= (~r0 )dv 0 ; (7.53)
V olum
respectively.
II. Classical Mechanics
A. Moment of inertial, I: In classical mechanics, moment of inertia is also

called mass moment of inertia, rotational inertia, polar moment of iner-
tia of mass, or the angular mass, (SI units kg m2 ). It is a property of a
distribution of mass in space that measures its resistance to rotational ac-
celeration about an axis. Newton’s …rst law, which describes the inertia of
a body in linear motion, can be extended to the inertia of a body rotating
about an axis using the moment of inertia. That is, an object that is ro-
tating at constant angular velocity will remain rotating unless acted upon
by an external torque. In this way, the moment of inertia plays the same
role in rotational dynamics as mass does in linear dynamics, describing
the relationship between angular momentum and angular velocity, torque
and angular acceleration. The symbols I and sometimes J are usually
used to refer to the moment of inertia or polar moment of inertia. For a
discrete mass that consist of N entities each with mass mi , the moment
of inertial is given by
XN
I= mi ri2 ; (7.54)
i=1
for continuum mass distribution in volume, V , described by the mass

density, (r) ; the moment of inertia can be determined using the volume
integral
Z ZZZ ZZZ
I= dmr2 = (r) r2 dv = (x; y; z) r2 dxdydz (7.55)
M V V
where ri (r) is the perpendicular distance from the axis of rotation and
dm dm
(r) = = (7.56)
dv dxdydz
is the volume mass density in Cartesian coordinates.
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
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)
For a mass distributed on a volume with volume mass density

dm dm
(x; y; z) = = ) dm = (x; y; z) dxdydx (7.59)
dv dxdydz
the center of mass can expressed as
ZZZ ZZZ
x (x; y; z) dxdydx y (x; y; z) dxdydx
x = ZZZ ; y = ZZZ ;(7.60)
(x; y; z) dxdydx (x; y; z) dxdydx
ZZZ
z (x; y; z) dxdydx
z = ZZZ : (7.61)
(x; y; z) dxdydx
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.
(a) What are the MKS units of the constant A?

(b) Find an expression for A if the total charge on the rod is Q.
(c) Point P is located a distance b from the origin along the negative x-axis.
Find an expression for the electrostatic potential at the point P , V (r).
Solution:
(a) the unit must be the unit of charge=length2 = C=m2
(b) The total charge for a line charge distribution is given by
Z Z L
0 0 AL2 2Q
Q= (~r )dr = Axdx = )A= 2 (7.66)
line 0 2 L
(c) For a line charge distribution, the potential is given by

Z L
1 (~r0 )
V (r) = dr0 (7.67)
4 0 0 j~r ~r0 j
Using the charge density (~r0 ) = Ax0 and

0
~r ~r = jb ( x
^) x0 (^
x)j = b + x0
one can write Z L

A x0
V (r) = dx0 : (7.68)
4 0 0 b + x0
Noting that
x0 b
=1 ; (7.69)
b + x0 b + x0
one can put the integral in the form
Z "Z Z #
L L L
A b A dx0
V (r) = 1 0
dx0 = 0
dx b
4 0 0 b+x 4 0 0 0 b + x0
(7.70)
which leads to
x=L
A A b+L
V (r) = [L b ln (b + x)] = L b ln : (7.71)
4 0 x=0 4 0 b
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
Example 4 A semi-circular plate of radius R is arranged in the upper-half x-y

plane with its center-of-curvature at the origin. The plate has a surface-
charge density given by (x; y) = Ax2 y, where A is a constant.
(a) What are the MKS units of the constant A?
(b) What is the total charge on the plate in terms of A?
(c) Re-do part (b), switching the order of integration.
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 :
(b) The total charge is given by surface integral

2 3
y 0 (x0 )
Z Zx2 Z
6 7
Q= (~r0 )da0 = 4 (x0 ; y 0 )dy 0 5 dx0 (7.73)
area
x1 y1 (x0 )
Using equation of a circle
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)
For the area shown in the …gure the limits of integration

p would then be
x1 = R to x2 = R and y10 (x0 ) = 0 to y20 (x0 ) = R2 x02 and the
integral becomes
2 0p 13
Z R RZ2 x02
6 02 B C7
Q=A 4x @ y 0 dy 0 A5 dx0 ; (7.76)
R
0
where we used the surface charge density

(x0 ; y 0 ) = Ax02 y 0 : (7.77)
Upon carrying out the integration with respect to y …rst, we have
Z R " p
R 2 x2
# Z
2 y 2
A R
Q=A x dx = R2 x2 x4 dx; (7.78)
R 2 0 2 R
so that integration with respect to x0 leads to

" R
#
A R 2 x3 x5 R5 R5 2AR5
Q= =A = (7.79)
2 3 5 R 3 5 15
(c) If we switch the order of integration, we have

0
2 (y )
Z Zy2 xZ
Q= (~r0 )da0 = [ (x0 ; y 0 )dy 0 ] dx0 (7.80)
area
y1 x1 (y 0 )
Referring to the …gure,p the limits of integration p would then be y1 = 0 to

y2 = R and x1 (y 0 ) = R2 y 02 to x2 (y 0 ) = R2 y 02 and the integral
becomes
2 p 3
R2 y 02 2 p 2 02 3
Z R Z Z R R y
6 7 x 03
Q = A y0 64 x02
dx07
5 dy 0
= A y 04 5 0
p 2 02 dy
0 p 2 02 0 3 R y
R y
2 3
Z R
3
2 R2 y 02 2
= A y 4
0 5 dy 0 (7.81)
0 3
The above integration can be carried out using transformation of variables.

If we introduce
du
u = R2 y 2 ) = ydy; (7.82)
2
one can write
2 3
Z 3
Z 3
2 R2 y 2 2 u2 2 5
y 4 5 dy = du = u2 (7.83)
3 3 15
Therefore, the charge becomes

2 3
Z R 3
R
2 R2 y 2 2 2A 5 2AR5
Q=A y4 5 dy = R2 y2 2
= (7.84)
0 3 15 0 15
Example 5 A right-circular cylinder has a height H, a radius R, and a mass M

distributed uniformly throughout its volume. Find the moment of inertia
of the cylinder about its axis of symmetry.
Solution: Since the mass is distributed uniformly in the cylinder, the volume
density, (r) ; is a constant given by
dm M M
(r) = = = : (7.85)
dv V R2 H
The moment of inertia about the axis of symmetry is an axis that passes
through the center of the cylinder. For a cylinder center about the z-axis,
we can take the z-axis as the center of symmetry. Thus using Eq. (7.55),
the moment of inertia can be expressed using a triple integral in Cartesian
coordinates as
Z ZZZ ZZZ
2 2 M
I = dmr = (r) r dxdydz r2 dxdydz: (7.86)
R2 H
M V V
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
y = R sin ( ) ) dy = R cos ( ) d (7.92)
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
Note that we have re-expressed the limits of integration in terms of

(i.e. R x R) =2 =2) : Using the result obtained by Math-
ematica (which you can also easily show),
the integral becomes

Z =2
3
cos2 ( ) + 2 sin2 ( ) cos2 ( ) = : (7.94)
=2 4
The moment of inertia is then found to be

M R2
I= : (7.95)
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.
Solution:
(a) The total mass of the wire is given by

Z
M = dm (7.96)
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)
Noting that from the result in (a)

Z
R
M = dm = (7.100)
2
The x and y coordinates of the center of mass can be expressed as

R Z Z
xdm 2 2
x= = xdm = xdl
M R R
and R Z Z
ydm 2 2
y= R = ydm = ydl; (7.101)
dm R R
where we replaced dm = dl and take into account that the density is
constant. From Eq. (7.10), we note that
v" # s
u
u dy 2 dx
2
dl = t + 1 dx = 1 + dy (7.102)
dx dy
we may write v"

Z u #
2 R u dy 2
x= xt + 1 dx; (7.103)
R 0 dx
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
Upon carrying out the integrations,
2 p R
2 p R
x= R2 x2 ;y = R2 x2 ; (7.109)
0 0
and one …nds

2R
x=y= : (7.110)
The center of mass is ( 2R ; 2R ; 0).

(c) The area can be determined using Eqs. (7.19)
2 3
Zx2 yZ2 (x)
6 7
A= 4 dy 5 dx: (7.111)
x1 y1 (x)
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

x = R sin ) dx = R cos d ; (7.116)
we may write
Z Rp Z R
A= R2 R2 sin2 R cos d = R2 cos2 d : (7.117)
0 0
Employing the relation

cos (2 ) + 1
cos (2 ) = cos2 ( ) sin2 ( ) ) cos2 ( ) = (7.118)
2
the area is found to be
Z =2 =2
2 cos 2 + 1 sin 2 R2
A=R d = + = : (7.119)
0 2 4 2 0 4
1
Note that the limits of integration in terms of = sin (x=R) ;
x=0) = 0; and x = R ) = (7.120)

2
are used in the integration.
7.2. INTEGRATION IN CURVILINEAR COORDINATES 183
(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)
and you can determine the volume by evaluating the integral

Z Z p 2 2 Z p 2 2 2
R R x R (y +z )
V = p dzdydx: (7.122)
R R 2 x2 0
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)
where y can be related to the radius of the sphere by

p
y = R 2 x2 : (7.124)
The volume will then be

Z R Z R
R3 2 R3
V = y 2 dx = R2 x2 dx = R3 = : (7.125)
0 0 3 3
7.2 Integration in curvilinear coordinates

We now consider the methods of integrations in curvilinear coordinates. We …rst
consider one and two dimensional curvilinear coordinates. For three-dimensional
coordinates, integrations in cylindrical and spherical coordinates will be consid-

ered. For problems with curved geometries, the simplicity of carrying out the
integrations in curvilinear instead of Cartesian coordinates will also be demon-
strated in this section. The demonstrations will re-consider some of the problems
we saw as examples in the previous section.
7.2.1 One- and two–dimensional curvilinear Coordinates

In polar coordinates a point in the x-y plane with coordinates (x; y) can be
described by (r; ). These two system of coordinates are related by
x = r cos ; y = r sin : (7.126)
The in…nitessimal arc-length in Cartesian coordinates that we saw in the previ-

ous section p
dl = dx2 + dy 2
in polar coordinates can then be written as
q q
2 2 2 2
dl = [d (r cos )] + [d (r sin )] = [cos dr r sin d ] + [sin dr + r cos d ] :
This can be simpli…ed into

q p
2
dl = cos2 + sin2 dr2 + r2 sin2 + cos2 d = dr2 + r2 d 2 :
By factoring dr or rd ; one can establish the relations

s
2
d
ds = 1 + r dr (7.127)
dr
or s
2
1 dr
ds = + 1rd : (7.128)
rd
respectively. Similarly for the in…nitessimal area in Cartesian coordinates,
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 = dxdy = (cos ( ) dr r sin ( ) d ) (sin ( ) dr + r cos ( ) d ) : (7.132)

There follows that

2 2
da = cos2 ( ) rdrd sin2 ( ) rdrd + sin ( ) cos ( ) (dr) sin ( ) cos ( ) (rd ) :
h i (7.133)
2 2
da = rdrd 2 sin2 ( ) rdrd + sin ( ) cos ( ) (dr) (rd ) : (7.134)
which is given by
da = rdrd : (7.135)
Example 7 Using curvilinear coordinates, repeat Example 6a-d

(a) The total mass of the wire is given by
Z Z
M = dm = L ds (7.136)
In polar coordinates this can be expressed as

s
Z =2 2
1 dr
M= L + 1rd (7.137)
0 rd
Referring to the …gure, we have
dr
r=R) =0 (7.138)
d
so that Z =2
LR
M= L Rd = (7.139)
0 2
(b) The area is given by Z Z Z
A= da = dxdy (7.140)
Noting that the in…nitesimal area da can be written as
da = rdrd (7.141)
we may express the area as

Z =2 Z R
R2
A= rdrd = (7.142)
0 0 4
Example 8 (This is Example 4b) A semi-circular plate of radius R is arranged

in the upper-half x-y plane with its center-of-curvature at the origin. The
plate has a surface-charge density given by (x; y) = Ax2 y, where A is a
constant.
Determine the total charge on the plate using polar coordinates.
Solution:
7.2.2 3-D Curvilinear coordinates-cylindrical

A point p in space described by the Cartesian coordinates (x; y; z) can be de-
scribed in cylindrical coordinates by (r; '; z). These coordinates are related
by
x = r cos (') ; y = r sin (') ; z = z (7.143)
In cylindrical coordinates an in…nitesimal volume dv can be expressed as
dv = dr (rd') dz = rdrd'dz (7.144)
in…nitesimal area da on the curved surface
da = dz (rd') = rdzd' (7.145)
on the planar surface

da = (dr) (rd') = rdrd' (7.146)
Moment of Inertia:The moment of inertia of an in…nitessimal mass dm in

the in…nitessimal volume dv about a given axis of rotation is given by
Z Z
I = r? dm = r? dv (7.147)
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)
Example 9 (This is the same problem considered in Example 17.4 ) A right-

circular cylinder has a height H, a radius R, and a mass M distributed
uniformly throughout its volume. Find the moment of inertia of the cylin-
der about its axis of symmetry using cylindrical coordinates.
Solution: Moment of inertia of a solid object is given by

Z Z
I= r2 dm = r2 (x; y; z) dxdydz (7.149)
vol vol
The mass is distributed uniformly and therefore the density is constant

Z Z Z
M M
I= r2 dxdydz = r2 dxdydz = r2 dxdydz (7.150)
V R2 H
vol vol vol
The axis of symmetry is the z-axis if we assume the cylinder is vertical.

Then we may express the inertia about this axis as
Z Z
M
I= r2 dm = r2 rdrd'dz (7.151)
R2 H
vol vol
ZH ZR Z2
M 1
I= r3 drd'dz = M R2 (7.152)
R2 H 2
z=0x=0 0
Example 10 Using Cylindrical coordinates prove that the volume of a right-

circular cone of base radius R and height H to be
R2 H
V = : (7.153)
3
Solution: In cylindrical coordinates an in…nitessimal volume dv is given by

Z Z Z
dv = rdrd'dz ) V = rdrd'dz (7.154)
Noting that for a right circular cone, we have
(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
Upon integrating with respect to ' and then r, one …nds

Z H
(H z)R Z H
r2 H
R2
V =2 dz = H2 2Hz + z 2 dz: (7.157)
0 2 0 H2 0
There follows that

" H
#
R2 2 z3
2 R2 H
V = H z Hz + = : (7.158)
H2 3 0 3
7.2.3 3-D Curvilinear Coordinate Systems: spherical Co-

ordinates
A point p in space described by the Cartesian coordinates (x; y; z) can be de-
scribed in spherical coordinates by (r; ; '). These coordinates are related by
x = r cos (') sin ( ) ; y = r sin (') sin ( ) ; z = r cos ( )
In spherical coordinates an in…nitessimal volume dv can be expressed as
dv = dr (rd ) (r sin ( ) d') = r2 dr sin ( ) d d' (7.159)
in…nitessimal area da on the surface
da = (rd ) (r sin ( ) d') = r2 sin ( ) d d' (7.160)

Example 11 A thin spherical dielectric shell of radius R has a surface-charge

density given by
'
( ; ') = 0 sin ( ) cos (7.161)
4
where 0 is a constant. Find the total charge.
Solution: An in…nitesimal area on the surface of a sphere of radius R is given

by
da = R2 sin ( ) d d' (7.162)
Thus the total surface charge given by
Z Z
Q= ( ; ') da (7.163)
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
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
The total charge is found to be

2
Q=2 0R : (7.168)
Example 12 A zone of a sphere is de…ned to be that section of the sphere

that lies between two parallel planes intersecting the sphere. Consider
two planes separated by a distance h that intersect a sphere of radius
R. Prove that the surface area of a zone cut out by these two planes is
A = 2 Rh. Prove this somewhat surprising result (surprising in that it
does not depend on where the slice of the sphere is taken— toward the
equator of the sphere or closer to the top or bottom!)
Solution: An in…nitesimal area da on the surface of the sphere with radius R

is given by
da = R2 sin d d': (7.169)
Then the surface area of the zone is given by
Z Z 2 Z 2
R2 sin d d' = 2 R2 cos ( )

2
A= da =
1
0 1
= 2 R2 [cos ( 1 ) cos ( 2 )] : (7.170)

where the limits of integration, 1 and 2 represent the angular positions

of the two planes forming the zone. Suppose in Cartesian coordinates, the
upper plane is de…ned by z = z1 and the bottom by z = z1 ; we can easily
see that
z1 z2
cos ( 1 ) = ; cos ( 2 ) = : (7.171)
R R
There follow that
R [cos ( 1 ) cos ( 2 )] = z1 z2 = h; (7.172)
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.

1. Consider an area on the x-y plane bounded by the y-axis, the line de…ned
p
by the equation, x = 2; and a curve de…ned by the equation, y = x;.
Using integration in Cartesian coordinates, …nd
(a) the length of the curved side of the area,
(b) the area.
2. The area in the previous problem is rotated about the y-axis to form a
volume with curved surface. Using integration in Cartesian coordinates,
…nd the surface area for
(a) the curved surface,
(b) the plane surface.
3. For an area (A) bounded by the x-axis (y = 0), y = sin (x) ; and x = =2
evaluate the integral
Z Z
I= 6y 2 cos xdxdy:
A
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:

1. A rope of length, L = 10cm, is stretched along the positive y-axis with one
end at the origin. The linear mass density, (y) ; over this rope increases
uniformly from 4 gm=cm at its one end at the origin to 24 gm=cm at the
other end: Find
(a) the linear mass density,
(b) the total mass of the rope,
(c) the center of mass,
(d) the moment of inertia about an axis passing through the center of mass,
(e) the moment of inertia about an axis passing through the heavy end.
2. A solid disk with radius, r = a; and negligible thickness is sitting on the
x-y plane centered at the origin. It has a uniform surface mass density,
0 : Using integration in Cartesian coordinates, …nd
(a) the circumference of the circle,

(b) the area of the disk,
(c) the centroid (or center of mass) for the portion of the disk in the …rst
quadrant,
(d) the moment of inertia of the disk about an axis along the diameter of the
disk,
(e) the centroid for the portion of the side of the disk in the …rst quadrant,
3. The surface charge density ( ) on the area bounded by the x-axis (y = 0)
and curve (y = 16 x2 ) is proportional to y: That means
(y) = ky:
Find the total charge (Q) on the area.
4. A light with an intensity, I0 = 42J=m2 ; is incident on a square silvered
mirror sitting on the x-y plane. The square has side length 2 m and
is centered about the origin. The fraction of the incident light that is
2
re‡ected at a point on the mirror (x; y) is (x y) =4. In other words, the
intensity of the re‡ected light at the point is
2
IR (x; y) = I0 (x y) =4
Find the fraction of the re‡ected light from the entire mirror.
5. Using double integrals determine the volume formed below the surface
de…ned by z = py(x + 2) and above the surface bounded by x + y = 0,
y = 1; and y = x:

1. Repeat Problem 2 in Homework Assignment 16 using Polar coordinates.
1. Consider the volume of the space bounded by the curved surface of the
cylinder de…ned x2 + y 2 = 4, z = 2x2 + y 2 ; and the x-y plane. Find the
volume of the cylinder using integrations in cylindrical coordinates.
3. Consider a solid sphere of radius, r = a and mass, M . The mass is
distributed uniformly over the volume of the sphere. Using integrations
in spherical coordinates, …nd
(a) The centroid of a solid half ball,

(b) the moment of inertia of the of the sphere about an axis passing through
the diameter (along the z-axis)
(c) the volume of the sphere
4. In problem 3 we considered a solid sphere of radius, r = a; and mass,

M that is distributed uniformly over the volume of the sphere. Now let’s
imagine the sphere is hollow and the same mass is distributed over the
surface of the sphere. Using integrations in spherical coordinates, …nd
(a) the surface area of the sphere,
(b) the centroid of the curved area of the upper hemisphere,

(c) the moment of inertia of the of the sphere about an axis passing through
the diameter (along the z-axis)
5. Consider a sphere with constant mass density, 0 and radius r = 2a:
Determine the gravitational attraction force on a unit mass (m = 1) at
the center of the sphere due to the portion of the mass of the sphere
occupying the volume above the plane de…ned by, z = a:
Chapter 8
Vector Calculus
We were introduced to the basic to vector algebra in chapter 3. Here we will

introduced to calculus of vector functions (di¤erentiation and integration). But
…rst we begin by reviewing vector products and their applications in physics
8.1 Vector Products

Two vectors A~ = Ax x ~ = Bx x
^+Ay y^ +Az z^ and B ^+By y^+Bz z^ can be multiplied in
two di¤erent ways one resulting a scalar and the other resulting a third di¤erent
vector.
Dot product (Scalar product): The dot product of A ~ and B ~ is given by
~ 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)
Taking into account the relation

~
A ~ =
C ~
C ~
A;
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 ".
8.1.1 Physical applications

Work: Suppose a constant force, F~ ; is applied on an object to move it for from
a position ~r1 to a position ~r2 for a displacement, d~ = ~r = ~r2 ~r1 : The angle
between the force and the displacement vectors is : We have seen that the work
done by this force, W; is given by the scalar product of the vectors F~ and d;~
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
Angular and linear velocities: Consider an object under a rotational motion.

Suppose the object is rotating with a constant angular velocity, ! ~ ; about some
axis of rotation. Over a time interval, t = t2 t1 ; the position of the object
relative to the axis of rotation has changed from ~r1 to ~r2 and the linear displace-
ment becomes, d~ = ~r = ~r2 ~r1 : The linear velocity, ~v ; of the object, which
is a constant for a constant angular velocity, !~ ; can be determined using the
vector product of the two vectors expressed as
~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)
Torque: The mechanical torque, ~ ; is vector physical quantity that we often

use to describe the rotational motion of an object acted by a force, F~ . Suppose
the force acted on the objected has rotated the object about an axis such that the
position relative to the axis has changed from ~r1 to ~r2 for a total displacement,
d~ = ~r = ~r2 ~r1 : For a constant force the torque about the axis of rotation is
given by the vector product,
~ = d~ F~ : (8.17)
However, when the force is not constant and depends on the position of the
object, the torque given by the integral,
Z ~
r2
~= d~r F~ : (8.18)
~
r1
Replace the linear velocity, ~v = !

~ ~r; in the expression for the linear
momentum p~ = m~v ; the angular momentum can be expressed as a triple vector
product. Z
J~ = m d~r (~
! ~r) (8.19)
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
Solid state Physics: Reciprocal lattice vectors {~bi }
~bi = ~aj ~ak

, where (i; j; k) = (1; 2; 3) (8.23)
~a1 (~a2 ~a3 )
{~ai } is called a lattice vector. It is used to describe crystal structure in solids.

The reciprocal lattice vectors are useful to study waves in solids (lattice waves)
known as phonons. These waves are described by the wave vector, ~k; expressed
in terms of the reciprocal vectors.
8.1.2 Vectors derivatives

Next we shall see how we carry out the derivatives of a vector functions. To
this end, let’s consider time dependent scalar function, c (t) ; and two vector
functions that can be expressed in Cartesian coordinates as,
~ (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
Derivative of a vector multiplied by a scalar:
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
the instantaneous velocity can be put in takes the form

dr d
~v (t) = r^ + r^: (8.35)
dt dt
In terms of the angular speed, !;
d
!= (8.36)
dt
and the radial speed
dr
vr = ; (8.37)
dt
the instantaneous velocity can be expressed as
~v (t) = vr r^ + !r^: (8.38)
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
~v (t) = vt ^ = !r^: (8.42)
This is in agreement with what we learned in into physics I. Acceleration:

~a(t) is given by the time derivative of the instantaneous velocity,
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
so that the acceleration can be put in the form

dvr dvt ^ dvr
~a (t) = r^ + vr ! ^ + vt !^
r= vt ! r^
dt dt dt
dvt ^:
+ vr ! + (8.46)
dt
In terms of the magnitude of the acceleration in the radial and angular
directions
dvr d2 r
ar = = 2 (8.47)
dt dt
and
dvt d (r!) dr d!
at = = =! + r; (8.48)
dt dt dt dt
the instantaneous acceleration becomes
~a (t) = (ar vt !) r^ + (vr ! + at ) ^: (8.49)
This can also be put in the form
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.
8.2 Di¤erential Operators

8.2.1 The gradient operator and directional derivative
In the previous section we considered the derivative of scalar and vector func-
tions of one variable, time t. In most physical problems we often encounter
products of scalars and vectors that are functions of two or more variables.
Suppose we are given the scalar function in Cartesian coordinates, f (x; y; z).
We recall that the in…nitesimal change in this function, df (the total di¤eren-
tial), is given by
@f @f @f
df = dx + dy + dz: (8.58)
@x @y @z
~ and B
In view of the expression for scalar vector product of two vectors A ~ in
Cartesian coordinates
~ 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)
there follows that

dx dy dz
dsnx = dx; dsny = dy; dsnz = dz ) nx = ; ny = ; nz = : (8.69)
ds ds ds
For a function f (x; y; z); the directional derivative is the derivative along the
direction n
^ . It is given by
df @f dx @f dy @f dz
= + + : (8.70)
ds @x ds @y ds @z ds
Using the results above, we may write
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
and the directional derivative can then be expressed as
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).
The Gradient Operator in curvilinear Coordinates: In the previous chapter

we have introduced to two curvilinear coordinate systems, namely Cylindrical
and spherical coordinates system. Using the relation between the Cartesian
coordinates and these curvilinear coordinates, it can be derived that the gradient
in Cylindrical Coordinates is given by
@f '
^ @f @f
rf = r^ + + z^ (8.72)
@r r @' @z
and in Spherical Coordinates by
@f ^ @f '
^ @f
rf = r^ + + ; (8.73)
@r r@ r sin @'
where (^ ^ ; z^) and r^; ^; '

r; ' ^ are the unit vectors in Cylindrical and spherical
coordinates, respectively. Note that with the exception of z^; these unit vectors
are not constant unlike the Cartesian unit vectors.
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
where 0 is the electrical permittivity of a free space. In classical electro-

dynamics, the electric …eld is related to the electrostatic potential, (r) ;
by the equation
E~ (~r) = r (r) : (8.76)
Find an expression for the electric …eld at point P using
(a)
@ 1 @ '
^ @
r = r^ +^ + (8.77)
@r r@ r sin @'
(b)
@ @ @
O= x
^+ y^ + z^ (8.78)
@x @y @z
(c) Find the directional derivative of the electrostatic potential at ( 1; 2; 1) in

~ = 2^
the direction S x 2^ y + z^
Solution:
(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
(b) In Cartesian coordinates, the electric potential is given by
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
~ (~r) = @ (r) @ (r) @ (r)

E r (r) = x
^+ y^ + z^ : (8.82)
@x @y @z
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
Similarly,the partial derivatives with respect to y and z, we have
@ (~r) kQ y kQ @ (~r) kQ z kQ
= = y; = = z: (8.84)
@y r2 r r3 @z r2 r r3
so that the gradient of (~r) becomes
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
(c) We are given

~ = 2^
S x 2^
y + z^ (8.88)
and we want to …nd
d (r)
= r (r) n
^ : (8.89)
ds (1;2; 1)
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)
Using the expression for r in Cartesian coordinates, we have
d (~r) kQ 2 2 1
= 3=2
x y+ z (8.94)
ds (x2 + y2 + z2) 3 3 3
(1;2; 1)
so that the directional derivative for the electrostatic potential becomes

d (~r) kQ 2 4 1 kQ
= p = p : (8.95)
ds 6 3 3 3 6
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?
~
8.2.2 The Divergence, the Curl, and the Laplacian

We have seen the gradient operator on a scalar function. It can provide infor-
mation regarding how the scalar function changes along a given direction. Now
we will consider how the gradient operator acts on a vector functions, such as
the electric …eld due to a point charge we determined in the previous example,
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
~ = @Vx + @Vy + @Vz .

r V (8.97)
@x @y @z
Note that the divergence of a vector function results in a scalar function. For the
~ (x; y; z) ;, is denoted by r V
curl of the vector, V ~ . In Cartesian Coordinates,
~
r V is given by the determinant expression
x
^ y^ z^
r ~ = det
V @ @ @
(8.98)
@x @y @z
Vx Vy Vz
that leads to
~ = @Vz @Vy @Vx @Vz @Vy @Vx

r V x
^+ y^ + z^: (8.99)
@y @z @z @x @x @y
~ (x; y; z) is expressible as a gradient of some scalar
Suppose the vector …eld V
function, (x; y; z) ;
~ (x; y; z) = r (x; y; z) ;
V (8.100)
the divergence of this vector becomes
~ = r [r (x; y; z)] = (r r) (x; y; z) = r2 (x; y; z) ;
r V (8.101)
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
that can be expressed using the Laplacian as
~ (~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)
~ ; one can establish the

for the curl of the curl of the vector function, V
relation
r r V ~ =r r V ~ ~:
r2 V (8.111)
that involves the divergence and the Laplacian of the a vector function.
Thus for the Laplacian of a vector function one can write
~ =r r V
r2 V ~ r r ~
V
~ ; 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 :
(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
For f (x; y; z) = xyz;
@ (xyz) @ (xyz) @ (xyz)

rf = x
^+ y^ + z^ = yz x
^ + xz y^ + xy^
z: (8.115)
@x @y @z
Upon evaluating this at (x; y; z) = (1; 1; 1); we …nd
rf = x
^ + y^ z^: (8.116)
~ = (Vx ; Vy ; Vz ), is given by
(b) The divergence of the vector function, V
~ = @Vx + @Vy + @Vz :

r V (8.117)
@x @y @z
Noting that
Vx = xy; Vy = yz; Vz = xz
one …nds
~ = @ (xy) + @ (yz) + @ (xz) = y + z + x:

r V (8.118)
@x @y @z
At the point (1; 1; 1); the divergence is found to be
~ = 1:
r V (8.119)
(c) For r ~ ; we have

V
x
^ y^ z^
r ~ = det
V @ @ @
(8.120)
@x @y @z
xy yz xz
that leads to
~ = @ (xz) @ (yz) @ (xy) @ (xz)

r V x
^+ y^
@y @z @z @x
@ (yz) @ (xy)
+ z^
@x @y
)r ~ = (0
V y) x
^ + (0 z) y^ + (0 x) z^ = y^
x z y^ x^
z : (8.121)
Evaluating this at the point (1; 1; 1);
r ~ =x
V ^ y^ z^: (8.122)
(d) For the Laplacian of the scalar function, f ,

@2f @2f @2f
r2 f = + + (8.123)
@x2 @y 2 @z 2
we …nd
@ 2 (xyz) @ 2 (xyz) @ 2 (xyz)
r2 f = + + = 0: (8.124)
@x2 @y 2 @z 2
Noting that
r2 f = r rf; (8.125)
one can also …nd the same result
@ (yz) @ (xz) @ (xy)
r2 f = r (yz x
^ + xz y^ + xy^
z) = + + = 0; (8.126)
@x @y @z
using the expression for the gradient of the function determined in (a).
(e) Here we want to determine the Laplacian of a vector function instead of a
scalar function. In such cases, one should evaluate
~
@2V ~
@2V ~
@2V
~ =
r2 V + + (8.127)
@x2 @y 2 @z 2
where
~ = Vx x
V ^ + Vy y^ + Vz z^:
Thus for
~ = xy^
V x + yz y^ + zx^
z
one …nds
~
@2V @ 2 (xy^
x + yz y^ + zx^
z)
= =0 (8.128)
@x2 @x 2
~
@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
The Laplacian becomes

~ = 0:
r2 V (8.131)
We can also apply the relation
~ =r r V
r2 V ~ r r ~
V
along with the results we obtained earlier

~ =y+z+x
r V (8.132)
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
Using these results, one also …nds
~ =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)
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
and
1 @ @f 1 @2f @2f
r2 f = r + 2 2
+ 2; (8.144)
r @r @r r @' @z
respectively. For a vector function,
~ (r; '; z) = Vr (r; '; z) r^ + V' (r; '; z) '

V ^ + Vz (r; '; z) z^;
the divergence and the curl can determined using
~ = 1 @ (rVr ) + 1 @ (V' ) + @Vz

r V (8.145)
r @r r @' @z
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 @'
and the Laplacian as
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,
r + V (r; '; )^ + V' (r; '; )^

V (r; '; ) = Vr (r; '; )^ ';
the divergence and the curl are given by
~ = 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.
Example 8.4 Consider an in…nitely long cylindrical wire of radius a carrying

a current I along the positive z-direction and this current is uniformly
distributed.
The magnetic …eld is given by
~ =
B 0 Ir=2 a2 '
^ for r < a (8.151)
at a distance r from the wire in cylindrical coordinates ( r; '; z); 0 is

magnetic permeability of a free space.
(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)
(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
Noting the current density, which is de…ned as

Current
J= ; (8.155)
Area
is
I
J= (8.156)
a2
we may write
)r ~ =
B 0J z
^ = ~
0J (8.157)
which is known as Ampere’s law.
(c) The divergence in cylindrical coordinates is given by
~ = 1 @ (rVr ) + 1 @ (V' ) + @Vz

r B
r @r r @' @z
and using
0 Ir
Br = 0; B' = ; Bz = 0:
2 a2
we …nd
~ =0
r B
[Electricity and Magnetism I...PHYS 4310]
Example 8.5 In example 8.2 we have shown that the electric …eld of a point
charge, Q, position at the origin, is given
~ (~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
8.3 Single vector function integrals

8.3.1 Line vector integrals
Example 8.6 Consider the vector …eld
~ r) = x2 y^
A(~ x + xy 2 y^ + a3 e y
x cos ( ) z^; (8.162)
where a; ; and are known constants. Evaluate the line integral

I
I= ~ r) d~r
A(~ (8.163)
C
around the curve C shown in the diagram below.
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
Let’s represent a point on any of these line segments by a vector
~r = x^
x + y y^ (8.165)
so that
d~r = dx^
x + dy y^ (8.166)
For a point on side s1
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)
Therefore, using the three line integrals on s1

Z
~ r) d~r = 0
A(~ (8.174)
s1
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
8.3.2 Conservative vector …elds

Conservative Vector Fields: a vector …eld
~ (x; y; z) = Vx (x; y; z) x
V ^ + Vy (x; y; z) y^ + Vz (x; y; z) z^ (8.178)
is said to be conservative if any one of the following conditions is satis…ed:
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
If the curl of the vector …eld is zero,
r ~ (x; y; z) = 0:
V (8.180)
If the vector …eld can be expressed as a gradient of some scalar function

U (x; y; z);
V~ (x; y; z) = rU (x; y; z): (8.181)
Gravitational and electric …elds are conservative …elds.
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)
Figure 8.1: Moon’s orbit around the earth.

(a) Find the gravitational force
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)
There follows that
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)
Then the work done

I
GM m
W = r^ d~r = 0 (8.193)
C a2
8.3.3 Conservative Fields and Exact Di¤erentials

Exact vs Inexact Di¤ erentials: Consider the scalar function U (x; y; z). we recall
that the di¤erential for this function is given by
@U @U @U
dU =dx + dy + dz
@x @y @z
@U @U @U
dU = x
^+ y^ + z^ (dx^x + dy y^ + dz z^) = rU d~r: (8.194)
@x @y @z
Let’s assume that the scalar potential U (x; y; z) be a potential function for some
conservative vector …eld, V ~ (x; y; z) = Vx (x; y; z) x
^ + Vy (x; y; z) y^ + Vz (x; y; z) z^;
~
(i.e. V (x; y; z) = rU (x; y; z)) then we can write
~ 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
we may write
@Vz @Vy @2U @ 2 U @Vx @Vz @2U @2U

= or = ; = or = ;
@y @z @y@z @z@y @z @x @z@x @x@z
@Vy @Vx @2U @2U
= or = : (8.199)
@x @y @x@y @y@x
These relations are known as the Reciprocity Relations. The di¤erential
@U @U @U
dU = dx + dy + dz; (8.200)
@x @y @z
is said to be an exact di¤erential if the Reciprocity Relations are fully satis…ed.

If not, then the di¤erential is inexact. We derived the Reciprocity Relations
assuming the scalar function U is a function that its gradient gives a conservative
~ (x; y; z) = rU (x; y; z)). Therefore, we can conclude that
vector …eld V (i.e. V
a di¤erential
@U @U @U
dU = dx+ dy + dz = f (x; y; z)dx+g(x; y; z)dy +h(x; y; z)dz (8.201)
@x @y @z
is an exact di¤erential if the vector …eld constructed as
~ (x; y; z) = f (x; y; z)^

V x + g(x; y; z)^
y + h(x; y; z)^
z (8.202)
is a conservative vector …eld. Which means we can also say if

I I I I
~ (x; y; z) d~r =
V f (x; y; z)dx + g(x; y; z)dy + h(x; y; z)dz = 0;
C C C C
(8.203)
the di¤erential is exact otherwise it is inexact.
8.4 Multiple vector integrals

8.4.1 Green’s Theorem
Let’s consider two functions P (x; y) and Q(x; y). If these two functions and its
…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] ; (8.204)
A @x @y C
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
so that using the relation

Z b
df (t)
dt = f (b) f (a) (8.207)
a dt
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
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
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
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
8.4.2 The Stokes’Theorem

Stokes’ Theorem:
I Z
~ d~r =
D r ~ n
D ^ dA: (8.222)
Curve Surf ace A
bounding A
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)
~ that depends on the x and y coordinates

For a vector D
~ (x; y) = Dx (x; y) x
D ^ + Dy (x; y) y^ + Dz (x; y) z^ (8.225)
the curl in Cartesian coordinates is given by

x
^ y^ z^
~ = det @ @ @ @Dz @Dy @Dx @Dz
r D @x @y @z = x
^+ y^
@y @z @z @x
Dx Dy Dz
@Dy @Dx
+ z^ (8.226)
@x @y
One can then write
~ d~r = Dx (x; y) dx + Dy (x; y) dy;
D
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
Some Notes about Stokes’ Theorem:

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
~ through the open area A.

is just the ‡ux of the curl of the vector D
3. The integral I
~ d~r
D (8.230)
C
~ around the closed curve C. If D

is the line integral of the vector D ~ is a
force, then this integral is just the work done by that force around the
curve C. If the force is a conservative force, then the integral is zero.
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.

~
D(x; y; z) = 3y; xz; yz 2 (8.234)
and the paraboloid

2z = x2 + y 2 (8.235)
bounded by the plane z = 2.
(a) Evaluate the ‡ux integral directly

Z
r ~ n
D ^ dA (8.236)
A
(b) Evaluate the ‡ux integral indirectly by exploiting Stokes’ theorem.

Solution:
~
(a) The curl of the vector D(x; y; z) = 3y; xz; yz 2 is given by
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)
Using the integral
we …nd Z
r ~ n
D ^ dA = 20 : (8.242)
A
(b) Applying Stokes’ theorem

Z I
r ~ n
D ^ dA = ~ d~r:
D (8.243)
Surf ace A Curve
bounding A
we have
~
D(x; y; z) d~r = 3y; xz; yz 2 (dx; dy) = 3ydx xzdy (8.244)
Using the circular curve of radius R = 2 on the plane z = 2 bounding the

circular surface we considered in part a, we may write
I Z 2 h p p i
~ d~r
D = 3 4 x2 dx 2xd 4 x2
C 2
Z 2 h p p i
+ 3 4 x2 dx 2xd 4 x2 (8.245)
2
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
8.4.3 The Divergence Theorem

~ (x; y; z)
For a vector …eld V
ZZZ ZZ
~
r V (x; y; z) dxdydz = ~ (x; y; z) n
V ^ da (8.250)
volum e Surface
enclosing

~ (x; y; z) = (xy; yz; zx)
V (8.251)
(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)
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
Noting that Vx (0; y; z) = 0 and Vx (a; y; z) = ay; we …nd

Z c Z b
ab2 c
vx = aydydz = : (8.255)
0 0 2
Similarly for the surfaces normal to the y axis, we may write

ZZ ZZ
vy = ~
V (x; b; z) y
^ dxdz + ~ (x; 0; z) ( y^) dxdz
V (8.256)
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
Noting that Vy (x; 0; z) = 0 and Vy (x; b; z) = bz; we …nd

Z c Z a
abc2
vy = bzdxdz = (8.258)
0 0 2
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
(b) According to Divergence theorem

ZZZ ZZ
r V ~ (x; y; z) dxdydz = ~ (x; y; z) n
V ^ da
volum e Surface
enclosing
(8.261)
Thus the ‡ux can be expressed as
ZZZ
v = r V ~ (x; y; z) dxdydz
volum e
Z c Z bZ a
@Vx @Vy @Vz
= + + dxdydz
0 0 0 @x @y @z
Z c Z b Z a
@ (xy) @ (yz) @ (zx)
= + + dxdydz
0 0 0 @x @y @z
Z c Z b Z a
= (y + z + x) dxdydz (8.262)
0 0 0
which leads to
a2 bc ab2 c abc2
v = + + : (8.263)
2 2 2
8.4.4 More Examples on Divergence and Stokes theorems

The divergence in cylindrical coordinates:
~ = 1 @ (rV ) + 1 @ (V' ) + @ (Vz )

r V (8.264)
r @r r @' @z
Example 8.10 Verify the divergence theorem for the vector …eld
~ = ar^
V r + b^
' + cz z^ (8.265)
(where a; b; and c are constants) and the cylinder shown below.
Solution:
Noting that
Vr = ar; V' = b; V z = cz (8.266)
the divergence
~ = 1 @ (rV ) + 1 @ (V' ) + @ (Vz )

r V (8.267)
r @r r @' @z
becomes
~ = 1 @ r2 a + 1 @ (b) + @ (cz) = 2a + c
r V (8.268)
r @r r @' @z
Now applying the Divergence theorem

ZZZ ZZ
r V~ (x; y; z) dxdydz = ~ (x; y; z) n
V ^ da
volum e Surface
enclosing
(8.269)
For the volume integral, we …nd
ZZZ
(2a + c) dxdydz = (2a + c) R2 H: (8.270)
volum e
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)
~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
The Divergence in spherical Polar Coordinates:
~ = 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
~ = 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)
and the curve C shown in the diagram below.
Solution:
We are given the magnetic …eld in spherical coordinates with components
Br = 4; B = 3; B' = 2 (8.287)
so that the curl of the magnetic …eld
~ = 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
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
We recall the position vector in spherical coordinates is given by
~r = r^
r ) d~r = dr^
r + rd^
r (8.296)
where
r^ = sin ( ) cos (') x

^ + sin ( ) sin (') y^ + cos ( ) z^: (8.297)
Since the closed curve is on the x-y plane ( = =2), we have
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
The two results for the magnetic …eld vector
r + 3^
~ = 4^
B 2^
' (8.303)
verify that Stokes’ theorem

I Z
B~ d~r = r ~ n
B ^ dA: (8.304)
C A
is valid.
8.4.5 Applications of the Divergence and Stoke’s theo-

rems
Gauss’ Law
~ =
r E (8.305)
0
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
This is the di¤erential form of Ampere’s law.

Example 8.13 Who is wrong (Amperes or Stokes)?[From PHYS 4330]: Con-
sider a parallel plate capacitor connected to a battery as shown in the …gure
below
If we consider the ‡at circular area for the Amperian loop shown in the …gure
the current enclosed is just Ienc
Z
Ienc = J~ d~a = I: (8.311)
f lat surf ace
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
Figure 8.2: A capacitor connected to a dc current source.
which leads to a di¤ erent result

I Z
~
B d~r = r B ~ n^ dA (8.314)
C A
Z
J~ n
^ dA = 0 Ien = 0
A
Who is wrong here? Neither Ampere nor Stoke! However, in the

case considered above the current is not steady because of the build
up of charge on the capacitor. This suggested that Ampere’s Law
need to be …xed so that it can also be used when the current is not
steady and it was …xed by Maxwell. You want to learn more about
this? TAKE THEORETICAL II (PHYS 3160) in spring 2013 and
then E & M I & II (PHYS 4310 & 4330)!!!
Part II
Physics II
241
Chapter 9
Introduction to the
Calculus of Variations
9.1 Geodesic and stationary points

Geodesic: The curve along a surface which marks the shortest distance between
two neighboring points. Finding geodesics is one of the problems which can be
solved using the calculus of variation.
Stationary point: A point with coordinates, (x0 ; f (x0 ));on a curve de…ned
by the function f (x) is said to be a stationary point when
df (x)
= 0: (9.1)
dx x=x0
Figure 9.1: Stationary points.
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
dy (t) v0y v0 sin ( o )

vy (t) = =0)t= = (9.4)
dt g g
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.)
Figure 9.2: Geodesic and none geodesic paths.
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
We may rewrite this distance as

s
Z (2) 2
dy
L= 1+ dx: (9.7)
(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 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
Y (x; ) = y (x) + (x) ; (9.8)
where (x) is an arbitrary function which must satisfy the condition
(x1 ) = (x2 ) = 0 (9.9)
so that at the two end points ( x = x1 = x2 ), we …nd
Y (x; ) = y (x) : (9.10)
Here is the constant of variation. It is this constant that determines by

how much Y (x) di¤ ers from y (x). Now in terms of Y (x), we may write
Z (2) p
L( ) = 1 + Y 0 2 dx; (9.11)
(1)
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)
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
In view of these results, one …nds for the stationary point

Z (2)
dL ( ) 1 1 dY 0 ( )
= p (2Y 0 ) dx
d =0 (1) 2 1 + Y 02 d
=0
Z (2)
y 0 (x) 0 (x)
= p dx = 0: (9.20)
(1) 1 + y 02 (x)
Using integration by parts

Z Z
udv = uv vdu (9.21)
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
Since (x) is an arbitrary function, for the integral to be zero, we must

have !
d y0 y0
p = 0 ) p = c; (9.27)
dx 1 + y0 2 1 + y0 2
where c is a constant. Upon solving for y 0
1=2
c2
y0 = = m: (9.28)
1 c2
Note that we have introduced another constant in terms of the constant c.

There follows that
dy
= m ) y (x) = mx + b; (9.29)
dx
which is equation of a straight line.
9.2 The general problem

In the previous section we saw the application of the calculus of variation to
show that the shortest path connecting two points on a plane (the geodesic) is
a straight line
dy
= m ) y (x) = mx + b: (9.30)
dx
Next we shall consider the application of the calculus of variation to the general
problem. In an Euclidean space a surface is de…ned by the function, F (x; y; z) ;
where it depends on the Cartesian coordinates x; y; and z: Instead of the Euclid-
dy
ean space let’s consider a surface de…ned by the function, F x; y (x) ; y 0 (x) = dx :
This surface could, for example, be a surface in phase space (in Classical me-
chanics) if we replace
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
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
Y (x; ) = y (x) + (x) ; (9.33)
(x1 ) = (x2 ) = 0 (9.34)
so that at the two points (x = x1 = x2 ), we …nd
Y (x; ) = y (x) : (9.35)
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
For the Geodesic the integral

Z x2
I( )= F (x; Y (x; ) ; Y 0 (x; )) dx; (9.39)
x1
must be stationary, that means

Z x2
dI ( ) d
= [F (x; Y (x; ) ; Y 0 (x; ))] dx = 0: (9.40)
d =0 x1 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
Then the integral for the Geodesic line becomes

Z x2
dI ( ) @
= F (x; y (x) ; y 0 (x)) (x) dx + (9.44)
d =0 x @y
Z1 x2
@
+ F (x; y (x) ; y 0 (x)) 0 (x) dx: (9.45)
x1 @y 0
Using integration by parts the second integral can be rewritten as

Z x2 x2
@ 0 0 @ @F
F (x; y (x) ; y (x)) (x) dx = (x)
x1 @y 0 @x @y 0 x1
Z x2
@ @F
(x) (x) dx; (9.46)
x1 @x @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
Thus the stationary integral can be put in the form

Z x2
dI ( ) @
= F (x; y (x) ; y 0 (x)) (x) dx+
d =0 x1 @y
Z x2
@ @F
(x) (x) dx = 0: (9.49)
x1 @x @y 0
or Z x2
dI ( ) @ @F @F
= (x) dx = 0: (9.50)
d =0 x1 @x @y 0 @y
There follow that
@ @F @F
=0 (9.51)
@x @y 0 @y
where F = F (x; y (x) ; y 0 (x)):
Example 9.3 Do Example 9.2 using Euler-Lagrange Equation.
9.3 Appplications: The Brachystochrone Prob-

lem
The Brachystochrone Problem: If two points A and B are given, at di¤erent
heights but not lying one above the other (as shown in the …gure below), it is
required to …nd among all possible curves connecting them, that one along which
a material point slides from A to B under the in‡uence of gravity (neglecting
friction) in the shortest possible time. This curve is called a Brachistochrone
curve (Gr. o&, brachistos - the shortest, o o&, chronos - time), or
curve of fastest descent.[From Wikipedia, the free encyclopedia]
This problem occupied at the time of the leading mathematicians in the
whole of Europe: Newton, Leibniz, Bernoulli, L’Hospital, and others. From then
on, the calculus of variations developed as a special mathematical discipline.
9.3. APPPLICATIONS: THE BRACHYSTOCHRONE PROBLEM 251
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
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
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)
Figure 9.3: Cycloid motion.
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)
On the other hand if the circles center is
x = r ;y = r: (9.67)
then the trajectory looks like the …gure shown below
9.4 Applications: classical mechanics

Multivariable Functions: Suppose we are given a function, F , that depends on
y; z; dy=dx; dz=dx, and x, and we want to …nd y = y(x) and z = z(x) which
make Z x 2
I= F (x; y; z; y 0 ; z 0 ) dx (9.68)
x1
stationary, then we must solve the Euler-Lagrange equations

@ @F @F
= 0; (9.69)
@x @y 0 @y
@ @F @F
= 0
@x @z 0 @z
9.5 Physical application of the Euler-Lagrange

equation
~ The Hamiltonian ( H): The sum of the kinetic energy (T ) and potential
energy (V )
H = T + V: (9.70)
~ The Lagrangian ( L): The kinetic energy minus the potential energy
L=T V: (9.71)
~ 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
~ Hamilton’s Principle: The motion of a given system from time t1 to time t2

is such that the classical action
Z t2
S= Ldt (9.73)
t1
9.5. PHYSICAL APPLICATION OF THE EULER-LAGRANGE EQUATION255
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
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
Therefore, using the Lagrangian we …nd
@ @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.
Figure 9.4: Atwood’s Machine.

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
V = m1 gy1 + m2 gy2 (9.81)
Then the Lagrangian

L=T V (9.82)
of the system can be written as
1 1
L= m1 y_ 12 + m2 y_ 22 m1 gy1 m2 gy2 : (9.83)
2 2
This equation appears to be a function of two variables. However, because
of the constraint
y1 + y2 + l = C; (9.84)
where C is a constant, we end up with a Lagrangian that depends only on
one variable. If we replace
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
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)
and the kinetic energy becomes

1 2
T = m r_ 2 + r2 _ : (9.96)
2
The potential energy is related to the central force by
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)
Then the Lagrangian can be expressed as

Z
1 2
_ ;_ =T
L t; r; r; U= m r_ 2 + r2 _ + f (r) dr (9.99)
2
Then using the Euler-Lagrange’s equation
@ @L @L
=0 (9.100)
@t @ q_i @qi
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
Chapter 10
Introduction to the
Eigenvalue Problem
10.1 Matrix Review

Matrix Arithmetic and Manipulation: Consider the following matrices:
0 1
2 4
2 3 1
A = ;B = @ 1 1 A
2 1 0
3 1
0 1 0 1
2 1 3 2 0 1
C = @ 4 1 2 A;D = @ 1 1 2 A (10.1)
1 0 1 3 1 0
Multiplication by a Scalar : Any matrix can be multiplied by a scalar:
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
~ Addition and subtraction: Two matrices can be added or subtracted if and

only if they have the same dimensions. From matrices A; B; C, and D we
can add/subtract only matrices C and D
0 1 0 1
2 1 3 2 0 1
C +D =@ 4 1 2 A+@ 1 1 2 A
1 0 1 3 1 0
0 1 0 1
2 2 1+0 3+1 0 1 4
=@ 4+1 1 1 2+2 A=@ 5 2 0 A
1+3 0+1 1+0 2 1 1
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
where (ab)ij is the element of the product matrix AB:

~ Commutativities: For any two multiplyable matrices C and D;
CD 6= DC (10.6)
~ Commutator : For square matrices C and D the Commutator [C; D] is de…ned

as
[C; D] = CD DC (10.7)
~ For any three matrices, F; G;and H that can be multiplied, we can write
The Associative Law :
F (GH) = (F G)H (10.8)
The Distributive Law :
F (G + H) = F G + F H (10.9)
~ The Identity Matrix; I (Boas: The Unit Matrix, U )

IA = AI = A (10.10)
~ Transpose of a Matrix : The transpose of the matrix A is denoted by AT :

0 1
2 2
2 3 1
A = ; AT = @ 3 1 A
2 1 0
1 0
0 1
2 4
2 1 3
B = @ 1 1 A ; BT = (10.11)
4 1 1
3 1
10.1. MATRIX REVIEW 263
~ Adjoint of a Matrix : The adjoint of a square matrix, A, is given by

adj(A) = [cof (A)]T (10.12)
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 : (10.13)
~ Inverse of a (Square) Matrix : A 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
a. Row reduction approach

b. The adjoint matrix approach
Solution: a. In the row reduction approach we start from

2 3
1 2 3 1 0 0
4 2 0 4 0 1 0 5 (10.21)
1 1 1 0 0 1
and try to get 2 3
1 0 0 a11 a12 a13
4 0 1 0 a21 a22 a23 5 (10.22)
0 0 1 a31 a32 a33
so that we can get the inverse matrix
0 1
a11 a12 a13
A 1 = @ a21 a22 a23 A : (10.23)
a31 a32 a33
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
10.2 Orthogonal matrices and the rotational op-

erator
Matrices that make an orthogonal transformation of vectors. In an orthogonal
transformation of vectors the magnitude of the vectors remains the same. For
an orthogonal matrix
M 1 = MT (10.26)
The Rotation Operator : For a counter-clockwise rotation about the z-axis
by an angle , we denote the rotation matrix by R: Rz ( ) = R. Then,
r0 = Rr (10.27)
0 0
1 0 10 1
x cos sin 0 x
) @ y 0 A = @ sin cos 0 A@ y A
z0 0 0 1 z
10.3. EIGENVALUES AND EIGENVECTORS 265
10.3 Eigenvalues and Eigenvectors

In general, a linear transformation of a vector, ~r; to a vector, ~r0 ; using matrix
can be expressed as
~r0 = M ~r: (10.28)
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
Under this transformation if the vector ~r0 is expressible as
~r0 = M ~r = ~r; (10.30)
where is a constant, then the vector ~r is called the eigenvector (characteristic

vector) and is the eigenvalue (characteristic value) of the Matrix M;
0 10 1 0 1
M11 M12 M13 x x
@ M21 M22 M23 A @ y A = @ y A: (10.31)
M31 M32 M33 z z
Using the identity matrix, I;

0 1
1 0 0
I=@ 0 1 0 A; (10.32)
0 0 1
one can put Eq. (10.31)

0 10 1
M11 M12 M13 x
@ M21 M22 M23 A @ y A
M31 M32 M33 z
0 10 1 0 10 1
1 0 0 x 0 0 x
= @ 0 1 0 A@ y A = @ 0 0 A@ y A (10.33)
0 0 1 z 0 0 z

20 1 0 13 0 1
M11 M12 M13 0 0 x
4@ M21 M22 M23 A @ 0 0 A5 @ y A
M31 M32 M33 0 0 z
=0 (10.34)
0 10 1
M11 M12 M13 x
)@ M21 M22 M23 A @ y A = 0:
M31 M32 M33 z
The eigenvalues are obtained from the condition
M11 M12 M13

M21 M22 M23 = 0; (10.35)
M31 M32 M33
which is known as the eigenvalue equation (characteristic equation). To …nd the

eigenvectors we substitute the eigenvalues and solve the resulting equations.
From The VNR Concise Encyclopedia of Mathematics (Van Nostrand Rein-
hold Co., publishers, 1977):
Eigenvalues: Eigenvalue problems are important in many branches of physics.
They make it possible to …nd coordinate systems in which the transformations
in question take on their simplest forms. In mechanics for instance, the prin-
cipal moments of a rigid body are found with the help of the eigenvalues of the
symmetric matrix representing the inertia tensor.... Eigenvalues are of central
importance in quantum mechanics, in which the measured values of physical
“observables” appear as the eigenvalues of certain operators. The term “trans-
formation” is used predominantly in pure mathematical (geometrical) context,
whereas “operator” is more customary in applications (physics, technology).
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
Solution: The eigenvalue equation

0 1 0
0 1 0
1 0 0 =0) =0
0 0
0 0 0
3
) + =0) 1 = 0; 2 = 1; 3 = 1: (10.37)
The corresponding eigenvectors are determined from
0 10 1
M11 i M12 M13 xi
@ M21 M22 i M23 A @ yi A = 0: (10.38)
M31 M32 M33 i zi
For 1 = 0, we …nd
0 10 1
0 1 0 x1
@ 1 0 0 A @ y1 A = 0 ) x1 = 0; y1 = 0: (10.39)
0 0 0 z1
We use the bra-ket notation to represent eigenvectors. The eigenvector for
an eigenvalue, ; is represented using a ket-vector, j i ; which is expressed
as a column matrix 0 1
x
j i = @ y A: (10.40)
z
The eigenvector for 1 can then be written as a column matrix
0 1
0
j 1i = @ 0 A : (10.41)
z1
The corresponding bra-vector, h j ; generally, is the transpose conjugate
of the ket-vector. But here since we will consider only real values, the
bra-vector is written as
h 1j = x y z : (10.42)
Thus we can write
h 1j = 0 0 z1 : (10.43)
Both bra and ket vectors must be normalized (be unit vectors). Thus for
any eigenvectors, we must have
h j i = 1:
Normalizing this vector
0 1 0 1
0 0
h 1 j 1i =1) 0 0 z1 @ 0 A = 1 ) z1 = 1 ) j 1i =
@ 0 A:
z1 1
(10.44)
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)
and the eigenvector becomes
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:
Eigenvalue equation for Hermitian Matrices: A Hermitian Matrix is a matrix

that is equal to its transposed and conjugated matrix
T
(M ) = M: (10.51)
The eigenvalues for a Hermitian matrix are all real and the corresponding eigen-
vectors are orthogonal. This means for a Hermitian matrix with the eigenvalue
equation
M j ii = i j ii ; (10.52)
we always …nd
i = i (10.53)
and
0 i 6= j
h j j ii = = ij (10.54)
1 i=j
The Similarity Transformation: The similarity transformation of the matrix
M is given by
D C 1 M C: (10.55)
where C is a matrix whose columns are the eigenvectors of the Eigenvalue equa-
tion for matrix M . The matrix D is a diagonal matrix where the diagonal
elements are the eigenvalues.
10.4 Physical applications

Normal modes of Vibration for couple harmonic oscillators
Example 10.4 Consider a system consisting of two equal masses m connected

by three identical springs of spring constant k.
Figure 10.1: Coupled harmonic oscillators.
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.
Step 1: The Equations of Motion:

Step 2: Similarity Transformation (…nd the eigenvalues and eigenvectors)
Step 3: Solving the De-coupled Transformed Equations of Motion
Step 4: The Propagator Matrix, U
Step 5: The Normal Modes of Vibration
Solution:
Step 1 The Equations of Motion: To …nd the equations of motion I will use
Euler-Lagrange equations. You can use Newton’s second law if you want
to. To use the Euler-Lagrange equation we need to …nd the kinetic energy
and the potential energy. Suppose at a given instant of time the posi-
tion of the …rst mass is x1 and the second mass is x2 as measured from
their corresponding equilibrium positions. Then the corresponding kinetic
energy can be expressed as
1 1
K1 = mx_ 21 ; K2 = mx_ 22 (10.56)
2 2
10.4. PHYSICAL APPLICATIONS 271
so that the total kinetic energy be

1
K= m x_ 21 + x_ 22 : (10.57)
2
The potential energy (elastic) is due to the displacement of the springs
from their equilibrium position. If the …rst mass is displaced by x1 from
the equilibrium and the second mass by x2 in the positive x direction,
then the …rst spring will be stretched by x1 ; the second spring will be
compressed by x1 and at the same time it will be stretched by x2 ; as
a result the net displacement will be x2 x1 : The third spring will be
compressed by x2 : Therefore, the total potential energy will be
1 2 1 2 1 2
K= kx + kx + k (x2 x1 ) : (10.58)
2 1 2 2 2
Then the Lagrangian
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
Using the Euler-Lagrange’s equation
@ @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)
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.
Step 2: Similarity transformation: In a matrix form the two equations

can be put in the form
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)
we know that the matrix, C; constructed from the normalized eigen

vectors j i ; leads to a similarity transformation for the matrix M.
That means
1
C MC = D
1
a11 a12 M11 M12 a11 a12
)
a21 a22 M21 M22 a21 a22
1 0
= ; (10.73)
0 2
where
a11 a11
j 1i = ;j 2i = (10.74)
a21 a22
are the normalized eigen vectors. Noting that
1 1
C C = CC =I
one can write the matrix for the equation of motion as

:: ::
1
~r = M~r ) ~r = M CC ~r (10.75)
1
and multiplying from the left by the matrix C ; one can put the
equation of motion in the form
::
1 1 1
C ~r = C MC C ~r (10.76)
and introducing a new vector de…ned by
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
It is important to note that C is a matrix whose columns are the

eigenvectors of the Eigenvalue equation for matrix M . The matrix
D is a diagonal matrix in which the diagonal elements are the eigen-
~ such that
values. Suppose if we can …nd the eigenvectors, R;
~ = R;
MR ~ (10.78)
or using the bra-ket notation (also known as Dirac notation)
Mj i= j i; (10.79)
then for the eigenvalue, one can written
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)
The corresponding eigenvectors are obtained from
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
Solving these equations for, 1 = ! 2 ; we …nd

!2 !2 X1
=0) X1 + X2 = 0 ) X2 = X1
!2 !2 X2
(10.83)
for, 2 = 3! 2 ;
!2 !2 X1
= 0 ) X1 + X2 = 0 ) X2 = X1 : (10.84)
!2 !2 X2
Then the two eigenvectors becomes
~ 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:
Step 3: Solving the De-coupled Transformed Equations of Motion: The matrix

C and its inverse C 1 can be expressed as
! !
p1 p1 p1 p1
2 2 1 2 2
C= p1 p1
;C = p1 p1
: (10.88)
2 2 2 2
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)
so that the solutions can be expressed as

p p
y1 = A cos (!t) + B sin (!t) ; y2 = C cos 3!t + D sin 3!t : (10.93)
Now substituting these back into
1 x1 y1 1 x1 y1
C = ) CC =C
x2 y2 x2 y2
x1 y1
) =C (10.94)
x2 y2
we …nd
!
p1 p1 A cos
x1 (t)
= 2 2 p (!t) + B sin (!t)
p
x2 (t) p1 p1 C cos 3!t + D sin 3!t
2 2
(10.95)
Initially (t = 0) the …rst mass is displaced, x1 (0) = x10 and the second
mass displaced x2 (0) = x20 ; and both masses released from rest, x_ 1 (0) =
x_ 2 (0) = 0: Using these initial conditions we …nd
!
p1 p1 x10 A
2 2 =
p1 p1 x 20 C
2 2
1 1
) A = p (x10 + x20 ) ; C = p (x10 x20 ) (10.96)
2 2
and
!
p1 p1 !A sin
2 2 x_ 1
= p p (!t) + p
!B cos (!t) p
p1 p1 x_ 2 3!C sin 3!t + 3!D cos 3!t
2 2
)
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)
which can be put using matrices as

p p
x1 (t) 1 cos (!t) + cos p3!t cos (!t) cos p3!t
=
x2 (t) 2 cos (!t) cos 3!t cos (!t) + cos 3!t
x10
; (10.101)
x20
or
x1 (t) x1 (0)
= U (t) ) ~r = U (t)~r (0) ; (10.102)
x2 (t) x2 (0)
where
x1 (0) x10
~r (0) = = ; (10.103)
x2 (0) x20
and
p p
1 cos (!t) + cos p3!t cos (!t) cos p3!t
U (t) =
2 cos (!t) cos 3!t cos (!t) + cos 3!t
is called the propagator matrix.
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
= p
x2 (t) 2 2 cos (!t) cos 3!t cos (!t) + cos 3!t
1
(10.106)
1
which leads to
x1 (t) 1 cos (!t)

=p ) x1 (t) = x2 (t) : (10.107)
x2 (t) 2 cos (!t)
The two masses oscillate with a frequency, !; in the same direction. On

the other hand, if the initially state of the two masses is given by the
second eigenvector
x10 1 1
~r (0) = =p : (10.108)
x20 2 1
then we have
p p
= 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
11.1 The factorial and gamma function

The Factorial, n!: We recall the factorial function, n!; for a positive integer or
zero is de…ned as
n! = n (n 1) (n 2) (n 3)(n 4):::3 2 1 0!; (11.1)
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

Z 1
1!
x1 e x
dx = 2
: (11.6)
0
279
280 CHAPTER 11. SPECIAL FUNCTIONS
For the second derivative (n = 2)

Z 1 Z 1
@2 x @2 1 @ x @ 1!
e dx = ) xe dx =
@ 2 0 @ 2 @ 0 @ 2
Z 1
2 1! 2!
) x2 e x
dx = 3
= 3
: (11.7)
0
For the third derivative (n = 3)

Z 1 Z 1
@3 x @3 1 @ @ 2!
e dx = ) x2 e x
dx =
@ 3 0 @ 3 @ 0 @ 3
Z 1 Z 1
3 x 3 2 1! 3!
) x e dx = 3
) x3 e x
dx = 3
: (11.8)
0 0
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
Now we set = 1, we …nd

Z 1
n! = xn e x
dx; (11.10)
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
The Recursion Relation: If we replace p by p + 1 in the expression for the

Gamma function Z 1
(p) = xp 1 e x dx: (11.14)
0
we …nd Z 1
(p + 1) = xp e x
dx: (11.15)
0
11.1. THE FACTORIAL AND GAMMA FUNCTION 281
If we denote
u = xp ; dv = e x
dx ) du = pxp 1
;v = e x
; (11.16)
then using integration by parts

Z Z
udv = uv vdu (11.17)
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)
Example 11.1 Show that

1 p
= (11.22)
2
and Z 1 1
u2 2
e du = : (11.23)
0 2
Solution: Using the de…nition of the Gamma function, we can write
Z 1
1 1
= p e x dx. (11.24)
2 0 x
Introducing a new variable de…ned by
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
Squaring both sides, we have

Z 1 Z 1 Z 1 Z 1
1 2
v2 2 2
2
=4 e u du e dv =4 e (u +v ) dudv
2 0 0 0 0
(11.27)
so that introducing the polar coordinates de…ned by
u = r cos ; v = r sin ) dudv = rdrd ; u2 + v 2 = r2 (11.28)
as shown in the …gure below

the above integral can be put in the form
Z 1 Z =2
2 1 r2
=4 e rdr d : (11.29)
2 0 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
Now substituting this result into

Z 1 Z 1
1 1 x u2
= p e dx = 2 e du; (11.31)
2 0 x 0
we can see that Z r

1
u2
e du = : (11.32)
0 2
11.2. THE BETA FUNCTION 283
11.2 The Beta Function

For p > 0 and q > 0, the beta function B(p; q) is de…ned by a de…nite integral
Z 1
q 1
B(p; q) = xp 1 (1 x) dx: (11.33)
0
There are di¤erent forms of representations of the beta function. These

includes the following
~ Replace x = y=a ) dx = dy=a

Z Z
1 a y p 1 y q 1 1 a
q 1
B(p; q) = 1 dy = yp 1
(a y) dy:
a 0 a a ap+q 1 0
(11.34)
~ Replace
x = sin2 ( ) ) dx = 2 sin ( ) cos ( ) d
which gives
Z =2
q 1
B(p; q) = sin2p 2
( ) 1 sin2 ( ) 2 sin ( ) cos ( ) d
0
Z =2
) B(p; q) = 2 sin2p 1
( ) cos2q 1
( )d (11.35)
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)
Solution: For the Gamma functions

Z 1 Z 1
(q) = xq 1 e x dx, (p) = yp 1
e y
dy (11.39)
0 0
introducing the transformation of variables de…ned by
u2 = x ) 2udu = dx; v 2 = y ) 2vdv = dy (11.40)
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
Multiplying the two functions, we have

Z 1Z 1
2 2
(p) (q) = 4 u2q 1 v 2p 1
e (u +v ) du dv, (11.42)
0 0
so that using the polar coordinates
u = r cos ; v = r sin ) dudv = rdrd ; u2 + v 2 = r2 (11.43)
we …nd
Z 1 Z =2
2q 1 2p 1 r2
(p) (q) = 4 (r cos ) (r sin ) e rdrd . (11.44)
0 0

Z 1 Z =2
r2 2p 1 2q 1
(p) (q) = 4 r2(p+q 1)
e rdr (sin ) (cos ) d .
0 0
(11.45)
The …rst integral
Z 1 Z 1
r2
4 r2(p+q 1) e rdr = 2 R(p+q 1)
e R
dR = 2 (p + q) (11.46)
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
11.3 Stirling’s Formula

We recall the Gamma Function
Z 1
(p) = xp 1
e x
dx; (11.50)
0
when p is zero or positive integer, gives the factorial function

Z 1
p! = (p + 1) = xp e x dx: (11.51)
0
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)
or if we take the natural logarithm of both sides

p 1
ln(p!) = ln pp e p
2 p = ln (pp ) + ln e p
+ ln (2 p) 2
1 1
= p ln (p) p ln (e) + ln (2 p) ) ln(p!) = p ln p p+ ln (2 p)
2 2
If p is very large, the last term is very small as compared to the …rst two terms
and the Stirling’s formula is given by
ln(p!) = p ln p p: (11.53)
Proof: Introducing a new variable de…ned by

p p p
x = p + y p ) dx = pdy; x = 0 ) y = p; x ! 1 ) y ! 1;
(11.54)
the function Z 1
p! = (p + 1) = xp e x
dx (11.55)
0
can be put in the form
Z 1
p p p p
p! = p
(p + y p) e (p+y p) pdy: (11.56)
p
Noting that
p p p p p
(p + y p) = eln[(p+y p) ] = ep ln(p+y p) ; (11.57)
one can write

Z 1 p p
Z 1
p p p p
p! = p ep ln(p+y p) e (p+y p) pdy = p p
ep ln(p+y p) p y p
dy:
p p
(11.58)
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
we …nd an approximate expression

p
p py y2
ln (p + y p) = ln (p) + : (11.61)
p 2p
Then the approximate expression for the factorial becomes

Z 1
p p p
p! = p p ep ln(p+y p) p y p dy
p
Z 1 n h p
y2
i p o
Z 1 y2
p p ln(p)+ p
py
p y p p p ln(p) p
= p p
e 2p
dy = p p
e 2
dy
p p
Z 1 Z 1
p y2 p y2
= pep ln(p) p
p
e 2 dy ) p! = pep ln(p) p
p
e 2 dy: (11.62)
p 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
The second integral

Z p
p
y2
lim e 2 dy ! 0: (11.66)
p!1 1
Hence, the Stirling’s approximation for the factorial function will be

p p p p
p! ' 2p ep ln(p) p = 2p eln(p ) e p ' 2p pp e p : (11.67)
or with further approximation

p 1
ln(p!) ' ln 2p pp e p
= (ln + ln p + p ln p p)
2
) ln(p!) ' p ln p p: (11.68)
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
n = 0:5N (N 0:1%) ) n = 0:499N (11.72)

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
ln [P (n)] = N ln N N (n ln n n) ((N n) ln (N n) (N n))

+N ln p = N ln N N n ln n + n N ln (N n) + n ln (N n) + N n
+N ln p = N [ln N ln (N n) + ln p] + n [ln (N n) ln n]
N N n
= N ln p + n ln
N n n
N n N n
) ln [P (n)] ' n ln N ln : (11.76)
n Np
Substituting the values
N =3 1027 ; n = 0:499N ) N n = 0:501N; p = 0:5; (11.77)
we …nd
and
P (n) = exp[ 6 1021 ]: (11.78)
11.4 The Error Function

The error function is de…ned as the area under
Z x
2 2
erf(x) = p e t dt (11.79)
0
There are also other related integrals which sometimes referred as error func-
tions. These includes
11.4. THE ERROR FUNCTION 289
(a) The standard normal or Gaussian cumulative distribution function, (x);

Z x p
1 2 1 1
(x) = p e t =2 dt = + erf x= 2 : (11.80)
2 1 2 2
Noting that
Z x Z 0 Z x
1 2 1 2 1 2
p e t =2 dt = p e t =2 dt + p e t =2 dt
2 1 2 1 2 0
Z 1 Z x r Z x
1 2 1 2 1 1 2
=p e t =2 dt + p e t =2 dt = p +p e t =2 dt
2 0 2 0 2 2 2 0
Z x Z x
1 2 1 1 2
)p e t =2 dt = + p e t =2 dt (11.81)
2 1 2 2 0
The error function can also be expressed as
p Z x
1 1 1 t2 =2
erf x= 2 = (x) =p e dt: (11.82)
2 2 2 0
(b) The complementary error function:

Z 1 p
2 t2 =2
erf c(x) = p e dt = 1 erf x= 2 ,
x
r Z 1
x 2 t2 =2
) erf c( p ) = e dt (11.83)
2 x
Example 11.3 Consider a criterion that either is or is not satis…ed. We look

at a system that has many elements, each of which satis…es or does not
satisfy the criterion.
For example: Consider a test with many multiple-choice questions. Criterion:
the answer to a test question is correct. Each answer on the test is either
correct (satis…es criterion) or is incorrect (does not satisfy the criterion).
We then look at a large number of these systems (for example, a large number
of tests consisting of multiple-choice questions). We let x represent the
number of elements within a given system satisfying the criterion. We
then de…ne the following:
x The average number of elements satisfying the criterion
The standard deviation about the mean of the number of elements satis-
fying the criterion.
The probability that any one system will have x to x + dx elements satisfying
the criterion is then given by the Gaussian distribution:
1 (x x)2
(x)dx = p e 2 2 dx
2
Figure 11.1: Gaussian distribution.
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).
Solution: The probability that one system will have x to x + dx elements

satisfying the criterion is
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
Introducing a new variable de…ned by

x x p
p = y ) dx = 2 dy (11.86)
2
and noting that for x1 = x n and x2 = x + n
x1 x x n x n
y1 = p = p = p
2 2 2
x2 x x+n x n
y2 = p = p =p (11.87)
2 2 2
11.4. THE ERROR FUNCTION 291
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
If we split the integral in to two regions pn2 ; 0 and 0; pn2 ; we have

"Z Z pn #
0
1 2 2 2
Pn (x) = p e y dy + e y dy (11.89)
pn 0
2
and noting that

Z 0 Z 0 Z n
p
2
y2 y2 y2
e dy = e d ( y) = e dy (11.90)
pn n
p 0
2 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%
11.5 Elliptic Integrals

The Complete Elliptic Integral of the First Kind
Z =2
d'
K( ) p : (11.94)
2
0 1 sin2 '
Example 11.5 Consider a simple pendulum with a mass m suspended from

the end of a light rigid rod of length l. We pull the pendulum to the side
by an angle and release it from rest. Find an expression for the period
of the pendulum, T , assuming that = 0 and d =dt > 0 at t = 0, where
is the angle of the pendulum from the vertical. Then …nd an approximate
expression for the period of the pendulum for not-so-small amplitudes of
motion.
Figure 11.2: A simple pendulum. At the initial time, t = 0; the mass m was
displaced by an angle, ; from the vertical.
Solution: Using conservation of Mechanical energy,
M EI = M E (11.95)
where M EI is the initial mechanical energy (when the pendulum is pulled

to the side by an angle ) which is just only the gravitational potential
energy given by
M EI = mghmax = mgl(1 cos ) (11.96)

11.5. ELLIPTIC INTEGRALS 293
and M E is the mechanical energy at some instant of time (i.e. at an

angle ) which is the sum of kinetic and potential energy given by
2
1 1 d
M E = mgh + mv 2 = mgl(1 cos ) + ml2 : (11.97)
2 2 dt
Then
2
1 d
M EI = M E ) mgl(1 cos ) = mgl(1 cos ) + ml2
2 dt
2
r
1 d d 2g
) g cos = g cos ) + l ) = (cos cos ):
2 dt dt l
(11.98)
Noting that the period is the time for one complete oscillation which we
can express as Z t1=2
T =2 dt (11.99)
0
where t = 0 is the time at which the pendulum at the maximum displace-
ment from the vertical ( = ) and t = t1=2 is the time at which the
pendulum reached to the position ( = ). Therefore, using
r
d 2g d
= (cos cos ) ) dt = q (11.100)
dt l 2g
(cos cos ) l
we can write
Z t1=2 Z
d
T =2 dt = 2 q (11.101)
0 2g
l (cos cos )
Using the
cos = 1 2 sin2 ; cos =1 2 sin2 (11.102)

2 2
we have
r s
2g 2g
(cos cos ) = 1 2 sin2 1 + 2 sin2
l l 2 2
s
2g
= 2 sin2 2 sin2 (11.103)
l 2 2
so that
Z
d
T = 2 q
2g 2
l 2 sin 2 2 sin2 2
s Z s Z d
l d l sin( 2 )
= q = r (11.104)
:
g sin2 sin2 g sin2 ( 2 )
2 2 1 sin2 ( 2 )
Introducing the transformation

s
sin 2 sin2 2
sin ' = ) 1 = cos '
sin 2 sin2 2
1 cos 2 d 2 cos ' 2 cos '
) cos 'd' = d ) = d' = q
2 sin 2 sin 2 cos 2 1 sin2 2
d 2 cos '
) =q ; (11.105)
sin 2 1 sin2 sin2 '
2
= ) sin ' = 1)'= (11.106)

2
The expression for the period can be put in the form
s Z s Z
=2 =2
l d' l d'
T =2 q =4 p ;
g =2 1 sin2 2 sin2 ' g 0 1 2
sin2 '
where we introduced the constant

= sin (11.107)
2
and take into consideration the fact that sin2 ' is an even function such
that
Z =2 Z =2
d' d'
q =2 q :
2 2
=2 1 sin 2 sin ' 0 1 sin2 2 sin2 '
Therefore, the period is given by

s
l
T =4 K( ) (11.108)
g
where Z =2
d'
K( )= p (11.109)
2
0 1 sin2 '
is the elliptic integral of the …rst kind.
11.6 The Dirac delta function

Let’s consider a point in space that can be described by the Cartesian coordi-
nates (x; y; z) or the spherical coordinates (r; ; ') as shown in Fig. The position
vector
~r = x^
x + y y^ + z z^
= r sin ( ) cos (') x
^ + r sin ( ) sin (') y^ + r cos ( ) z^ = r^
r (11.110)
11.6. THE DIRAC DELTA FUNCTION 295
Figure 11.3: A point in space in spherical and Cartesian coordinates.
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
spherical coordinates can be expressed as

1
f (r; ; ') = ; for r > 0; (11.116)
r
and one can construct a vector …eld
~ =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
On the other hand if one determines using the divergence theorem

ZZZ ZZ
r V ~ d = ~ (x; y; z) d~a
V (11.122)
volum e Surface
enclosing
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
The function, (~r) ; de…ned as

r^ 1 1 0 for r 6= 0
4 (~r) = r = r r = r2 = (11.125)
r2 r r 1 for r = 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
For one dimensional case,

0 for x 6= 0
(x) = (11.127)
1 for x = 0
and
Z1 Z1
(x) dx = 1; f (x) (x) dx = f (0) (11.128)
1 1
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:
1. The Gaussian function:

n
n (x) = p exp( n2 x2 ); (11.129)
See Fig. 1.
2. Lorentz function
n 1
n (x) = ;
1 + n2 x2
See Fig. 2.
3. The sinc function:

Z n
sin (nx) 1
n (x) = = eixt dt
x 2 n
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
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
This potential for a point charge becomes

Z1Z Z2
1 (~r0 ) r02 sin ( ) dr0 d 0 d'0 1 q
=
4 0 j~r ~r0 j 4 0r
0 0 0
Z1Z Z2
1 (~r0 ) 02 1
) r sin ( ) dr0 d 0 d'0 = : (11.135)
j~r ~r0 j q r
0 0 0
From the property of the Dirac Delta function
Z1Z Z2
0
f (~r0 ) (~r0 ~r0 ) r02 sin dr0 d 0 d'0 = f (~r0 ) (11.136)
0 0 0
one can easily …nd
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
(~r) = q (~r ~r0 ) ; (11.139)
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)
The electric potential for a volume charge distribution is given by

ZZZ
1 (~r0 ) d 0
V (~r) = ; (11.141)
4 0 j~r ~r0 j
V
where ~r0 is the position of the in…nitesimal charge dq 0 = (~r0 ) d 0 ; in an

in…nitesimal volume d 0 , (~r0 ) is the charge density in the volume V .
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
In Cartesian coordinates, we have
~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)
and
(~r0 ~r0 ) = (x0 x0 ) (y 0 y0 ) (z 0 z0 )

0 0 0
= (x ) (y ) (z a) (11.144)
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)
Now applying the property of the Dirac Delta function

Z1
f (x) (x a) dx = f (a) (11.147)
1
one can easily see that

Z1 Z1
0 0 0 0 0
f (x ; y ; z ) (x ) dx = f (x0 ; y 0 ; z 0 ) (x0 0) dx0
1 1
0 0 1
= f (0; y ; z ) = q : (11.148)
2 2
x2 + (y y0 ) + (z z0)
The electric potential becomes

Z1 Z1
q 0 0
V (~r) = (z a) dz f (0; y 0 ; z 0 ) (y 0 ) dy 0 : (11.149)
4 0
1 1
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)
and the expression for potential reduces to

Z1
q
V (~r) = f (0; 0; z 0 ) (z 0 a) dz 0 : (11.151)
4 0
1
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
Power Series Solutions to

Di¤erential Equations
12.1 Power series substitution

The di¤erential equations we seek to …nd the solution to are linear di¤erential
equation like those we studied in PHYS 3150. However, unlike those equa-
tions here the coe¢ cients in the di¤erential equations we shall consider are not
constants and depend on the variable, x; like the following di¤erential equation
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.
Example 12.1 A mass m is attached to a horizontal spring of spring constant,

k; whose other end is attached to a rigid vertical wall. The mass slides
on a horizontal, frictionless surface. At time, t = 0; the mass is stretched
from its equilibrium position by a distance x = xmax and released from
rest. Find the equation for the position of the mass as a function of time,
x(t).
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
The di¤erential equation can then be rewritten as
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
From the above equation we …nd the following recursion relation
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)
For the …rst three odd terms, we have

1 1
a1 ( 1) 2 a1 ( 1) 2
n = 1 ) a3 = ! = !
3 2 3!
1 2 1
a1 ( 1) !
) n = 1 ) a(2 1+1) =
[(2 1) + 1]!
2 2
a3 2 a1 ( 1) ! 4 a1 ( 1) ! 4
n = 3 ) a5 = ! = =
5 4 5 4 3 2 1 5!
2 2 2
a1 ( 1) !
) n = 3 ) a(2 3 1) =
(2 2 + 1)!
3 3
a5 ! 2 a1 ( 1) ! 6 a1 ( 1) ! 2 3
n = 5 ) a7 = = =
7 6 7 6 5 4 3 2 1 7!
3
a1 ( 1) ! 2 3
) n = 5 ) a(2 4 1) = (12.13)
(2 3 + 1)!
Then for the odd terms we can write

m
a1 ( 1) ! 2m
a(2m+1) = ; (12.14)
(2m + 1)!
where m = 0; 1; 2; 3:::Thus the general solution to the di¤erential equation

1
X 1
X 1
X
x (t) = an tn ) x (t) = a2m t2m + a2m+1 t2m+1 ; (12.15)
n=0 m=0 m=0
using the results obtained above, can be expressed as

1
X X1 1
a0 ( 1) ! 2m 2m X a1 ( 1) ! 2m 2m+1
m m
x (t) = an tn = t + t
n=0 m=0
(2m)! m=0
(2m + 1)!
X1 1
a1 X ( 1)
m m
( 1) 2m 2m+1
) x (t) = a0 (!t) + (!t) (12.16)
m=0
(2m)! ! m=0
(2m + 1)!
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
where A = a0 and B = a1 =! are constants determined by the initial

conditions. Since initially (t = 0) the spring is stretched by xmax and the
mass is released from rest, we have
x (t = 0) = a0 cos (0) a1 ! sin (0) = xmax ) a0 = xmax (12.20)
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.
12.2 Orthogonal vectors and Dirac Notation

~ and B;
We recall that if the scalar product of two real vectors, A ~ is zero, given
by
X3
~ B
A ~ = Ai Bi = 0; (12.24)
i=1
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
~ is the complex conjugate of A:

where A ~ Using Dirac notation any vector A ~ is
~
denoted by a ket vector jAi and its complex conjugate A by a bra vector hAj :
Then the dot product of two vectors, using Dirac notation, can be expressed as
3
X
hA jBi = Ai B i (12.26)
i=1
and if the vectors are orthogonal,

3
X
hA jBi = Ai Bi = 0: (12.27)
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
fA1 (x) ; A2 (x) ; A3 (x) :::An (x) :::g (12.29)
that meet the requirement

Z b
0 m 6= n;
An (x)Am (x)dx = (12.30)
a C m = n;
where C is a constant, are said to form an orthogonal set of functions. Using

the Kronicker delta
0 n=6 m
nm = ; (12.31)
1 n=m
and Dirac notation, we may write
hAn (x) jAm (x)i = Cn nm (12.32)
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
These set of functions are said to form an orthonormal set of functions.
Example 12.2 Show that the set of functions de…ned by

1
Fn (x) = p sin (nx)
for n = 1; 2; 3::::form an orthonormal set of functions for all x ( ; ).

Solution: Using Euler’s formula, we can express
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)
whether m = n or m 6= n the second term is always zero and we can write
sin [(m n) ]
hFn (x) jFm (x)i = = nm ;
(m n)
where we have applied applied L’Hospital’s rule. But it would be shorter

if one applies the double angle relations
cos ( + ) = cos ( ) cos ( ) sin ( ) sin ( ) ;

cos ( ) = cos ( ) cos ( ) + sin ( ) sin ( ) ;
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
whether m is equal to or di¤erent from n; sin ((n + m) ) = 0; thus
1 sin ((n m) )
hFn (x) jFm (x)i = = nm
n m
12.3 Complete Sets of Functions

We recall a set of functions, Fn (x) ; form an orthonormal set for all x (a; b) ;
when
Z b
a
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)
or using Dirac notation

hxn j xm i = nm : (12.44)
Any vector ~r R3 can be expressed in terms of these three unit vectors
3
X
~r = ri x
î (12.45)
i=1
or using Dirac notation

3
X
jri = ri jxi i : (12.46)
i=1
Multiplying both sides from the left by hxj j ;
3
X
hxj jri = ri hxj jxi i : (12.47)
i=1
Since the vectors jxi i form an orthonormal set of vectors
3
X
hxj jri = ri ji ) rj = hxj jri : (12.48)
i=1
If one of the vectors, xî ; 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
for all x (a; b) can be expressed as a linear combination of these orthonormal

functions. That means
1
X
jg (x)i = cn jFn (x)i ; (12.49)
n=0
where the expansion coe¢ cient, cn ; is determined using the the property for
orthonormal set of functions
Z b
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
so that using Eq. (12.50), we …nd

1
X Z b
hFm (x) jg (x)i = cn nm = cm ) cm = Fm (x) g (x) dx: (12.52)
n=0 a
The completeness relation: A set of vectors or functions that form a complete

set, using the Dirac equation, are de…ned by the completeness relation as
1
X
jxn i hxn j = 1; for a complete set of vectors
n=0
Z b
dx jFn (x)i hFn (x)j = 1; for a complete set of functions
a
You will often use completeness relations in quantum mechanics.
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)
Thus the set of functionsfFn (x)gform a complete orthonormal set. We

shall see this is nothing but Fourier series expansion. As an example let’s
consider the step-function de…ned by
1 for 0 < x <

f (x) = (12.56)
1 for <x<0
(Example 16.4). The series expansion of this function can be show to be
1
4 X sin [(2n + 1) x]
f (x) = : (12.57)
n=0
2n + 1
Figure 12.1: Fourier series expansion of the step-function.
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
12.4 The Legendre Di¤erential Equation

Series substitution method is generally used to solve a linear di¤erential equation
of the form
dn y dn 1 y dn 2 y
an (x) + an 1 (x) + an 2 (x)
dxn dxn dxn 2
dy
+:::a1 (x) + a0 (x) = 0: (12.58)
dx
For the case where the coe¢ cients an (x) are independent of x, we have seen the
application of this method by solving the harmonic oscillator problem. Next
we shall see how this method is used to determine the solution to the Legendre
di¤erential equation where some of the coe¢ cients, an (x) ; depend on x. The
Legendre di¤erential equation, for x [ 1; 1]; is given by
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
The …rst derivative of this function

1
X X1 1
X
d
y 0 (x) = an (xn ) = an nxn 1
) 2xy 0 (x) = 2nxann : (12.61)
n=0
dx n=0 n=0
The second derivative

1
X 1
X
00 d
y (x) = an n xn 1
= an n (n 1) xn 2
n=0
dx n=0
1
X
00
) x2 y (x) = an n (n 1) xn : (12.62)
n=0
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
where we used a di¤erent dummy summation index, m. Now introducing an-

other dummy summation index de…ned by
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
Now substituting the expressions for y; xy 0 ; x2 y 00 ; and y 00 into the Legendre

di¤erential equation
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
Upon simplifying this expression, we …nd

1
X
fan+2 (n + 2) (n + 1) + an [l (l + 1) n (n + 1)]g xn = 0: (12.66)
n=0
There follows that
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)
one can write

(l n) (l + n + 1)
an+2 = an : (12.68)
(n + 2) (n + 1)
12.4. THE LEGENDRE DIFFERENTIAL EQUATION 315
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
y (x) = ye (x) + yo (x)
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!
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].
12.5 The Legendre Polynomials

The Legendre polynomials are the solution to the Legendre di¤erential equation.
These polynomials are determined from, y(x); that we determined earlier when
we impose the condition for convergence to the series. Any function that is a
solution to any di¤erential equation must be …nite. This means the series form
of the function must be convergent in the given domain, x [ 1; 1]: Using the
ratio test
an+1
lim x < 1: (12.77)
n!1 an
and the recursion relation we derived earlier, we have
(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!
and the odd series

(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 + ::: : (12.81)
7!
12.5. THE LEGENDRE POLYNOMIALS 317
Therefore, the interval of convergence for the function
y (x) = ye (x) + yo (x) (12.82)
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
we can conclude that the series
y (x) = ye (x) + yo (x) (12.84)
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
Is this series a convergent series? Using the integral test, we …nd

Z 1 1
dn 1
= ln (2n + 1) = 1; (12.89)
0 2n + 1 2 0
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
and the solution to the Legendre di¤erential equation, for l = 0 is
y (x) = ye (x) = a0 (12.91)
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
which is a divergent series and we must have, a0 = 0; so that
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!
For example, l = 2 and l = 3,
5 3
l = 2 ) y (x) = a0 1 3x2 ; l = 3 ) y (x) = a1 x x (12.99)
3
If we chose a0 and a1 such that y (x) = 1 at x = 1; we …nd
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
These polynomials, Pn (x) ; are called called the Legendre Polynomials.
12.6 The Generating Function for the Legendre

Polynomials
We recall that the legendre polynomials for l = 0; 1; 2; 3 are
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)
We want to examine the Taylor series expansion for the function

1=2
(x; h) = 1 2xh + h2 (12.103)
about, h = 0 which is given by

1
X 1 dl d h
(x; h) = (x; h) hl = (x; 0) + (x; 0)
l! dhl h=0 dh 1!
l=0
2 2
d h d3 h2
+ (x; 0) + (x; 0) + ::: (12.104)
dh2 2! dh3 3!
For l = 0
(x; 0) = 1 = P0 (x) ; (12.105)
l=1
d x h d
(x; h) = 3=2
) (x; h) = x = P1 (x) ; (12.106)
dh (1 2xh + h2 ) dh h=0
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
we can then conclude for the lth term in the series

dl
(x; h) = l!Pl (x) (12.109)
dhl h=0
Therefore, the Taylor series expansion for the function

1=2
(x; h) = 1 2xh + h2 (12.110)
about h = 0 can be written as
1
X 1
X
1 dl
(x; h) = (x; h) hl = Pl (x) hl : (12.111)
l! dhl h=0
l=0 l=0
The function (x; h) is the generating function for the Legendre polynomials.
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
If we introduce the variables

r0
h= ; x = cos (12.115)
r
we have
1 1 1=2
= 1 2xh + h2 (12.116)
j~r ~r0 j r
so that using
1
X
1=2
1 2xh + h2 = Pl (x) hl (12.117)
l=0
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
when r0 < r: On the other hand when r0 > r; we must write
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
The Rodrigues’ Formula: it gives Legendre Polynomials:
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
Leibniz’Rule: it is useful when we need to …nd a higher order derivatives of

a product (for example when we apply Rodrigues’ Formula). It is given by
XN
dN N dn u dN n v
(uv) = (12.127)
dxN n dxn dxN n
n=0
Example 12.6 Use Leibniz’ Rule to …nd

d8
x2 e2x : (12.128)
dx8
Solution: For N = 8 Leibniz’Rule can be expressed as
X8
d8 8 dn u d8 n v
(uv) = : (12.129)
dx8 n dxn dx8 n
n=0
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
If we use
X 1
1
p = Pl (x) hl (12.136)
1 2xh + h2 l=0
we can write the above expression as

P1 1
X
(x h) l=0 Pl (x) hl
= lPl (x) hl 1
1 2xh + h2
l=0
1
X 1
X
) (x h) Pl (x) hl = 1 2xh + h2 lPl (x) hl 1
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
There follows that
(l + 1) Pl+1 (x) x (1 + 2l) Pl (x) + lPl 1 (x) = 0

) (l + 1) Pl+1 (x) = x (1 + 2l) Pl (x) lPl 1 (x) : (12.141)
This is the recursion relation for the Legendre polynomials.
12.7 Legendre Series

Orthogonality of the Legendre polynomials: Legendre polynomials form an or-
thogonal set of functions. The orthogonality conditions is given by
2
hPl (x)j Pm (x)i = lm (12.142)
2l + 1
for x [ 1; 1] :
Completeness of the Legendre Polynomials:The Legendre polynomials also
form complete set on the interval [ 1; 1]. That means any function f (x) can
be expressed as a linear combination of the Legendre polynomials
1
X
jf (x)i = am jPm (x)i : (12.143)
m=0
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
Example 12.7 Expand the function f (x) in a Legendre series, for
1 1 x 0
f (x) = (12.147)
+1 0 x 1
f(x)
0
-1.0 -0.5 0.0 0.5 1.0
x
-1
Figure 12.2: The step-function.
Solution: The Legendre series is given by
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
the expansion coe¢ cients, al can be expressed as

2 0 3
Z Z1
2l + 1 4
al = Pl (x) f (x) dx + Pl (x) f (x) dx5
2
1 0
2 1 3
Z Z0
2l + 1 4
= Pl (x) dx Pl (x) dx5
2
0 1
2 1 3
Z Z 0
2l + 1 4
= Pl (x) dx Pl ( x) d ( x)5
2
0 1
2 0 3
Z Z1
2l + 1 4
= Pl ( x) dx + Pl (x) dx5
2
1 0
2 1 3
Z Z1
2l + 1 4
= Pl ( x) dx + Pl (x) dx5
2
0 0
2 1 3
Z
2l + 1 4
) al = [Pl (x) Pl ( x)] dx5 : (12.150)
2
0
Using the Rodrigues formula for the Legendre polynomials
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)
we …nd
Z1 Z1
3 7 7
a1 = 3 xdx = ; a3 = 5x3 3x dx = (12.156)
2 2 3
0 0
Therefore, the Legendre series is
3 7
f (x) = P1 (x) P2 (x) + ::: (12.157)
2 3
Least-Squares Fit and Legendre Series: Suppose we are given a curve or a

graph that is described by a function f (x) in the interval ( 1; 1) and we want
to …nd a polynomial
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
12.8 The Associated Legendre Di¤erential Equa-

tion
The Associated Legendre Di¤erential Equation is given by
m2
1 x2 y 00 2xy 0 + l (l + 1) y = 0:
1 x2
where m2 l2 . For m = 0, we …nd the Legendre Di¤erential equation
1 x2 y 00 2xy 0 + l (l + 1) y = 0
the solution of which we found to be the Legendre polynomials
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)
There follows that

00 (m+2)=2 m=2
1 x2 y (x) = 1 x2 u00 (x) 2m 1 x2 xu0 (x)
m=2 (m 2)=2
m 1 x2 u (x) + m (m 2) 1 x2 x2 u (x) : (12.163)
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)
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)
Using the results above we can write the DE
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
) 1 x2 u00 (x) 2 (m + 1) xu0 (x)

m 2 x2 2mx2 2mx2 m2
+ l (l + 1) m+ 2
+ u (x) = 0
1 x 1 x2 1 x2
) 1 x2 u00 (x) 2 (m + 1) xu0 (x)
" #
m2 x2 1
+ l (l + 1) m+ u (x) = 0 (12.171)
1 x2
and the associated di¤erential equation can be put in the form
1 x2 u00 (x) 2 (m + 1) xu0 (x) + [l (l + 1) m (m + 1)] u (x) = 0: (12.172)
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
(a) m=0 : we may rewrite the di¤erential equation

1 x2 u00 (x) 2xu0 (x) + l (l + 1) u (x) = 0: (12.173)
in the form
d2 f0 (x) df0 (x)
1 x22
2x + l (l + 1) f0 (x) = 0: (12.174)
dx dx
which is the Legendre di¤erential equation and the solution is given by
the Legendre polynomials
d0 Pl (x)
u (x) = f0 (x) = Pl (x) = : (12.175)
dx0
(b) m=1 : the di¤erential equation
1 x2 u00 (x) 2 (m + 1) xu0 (x)
+ [l (l + 1) m (m + 1)] u (x) = 0: (12.176)
becomes
1 x2 u00 (x) 4xu0 (x) + [l (l + 1) 2] u (x) = 0: (12.177)
If we di¤erentiate the di¤erential equation we found for m = 0 with respect
to x, we …nd
d d2 f0 (x) df0 (x)
1 x2 2x + l (l + 1) f0 (x) = 0
dx dx2 dx
d3 f0 (x) d2 f0 (x) d2 f0 (x) df0 (x)
) 1 x2 3
2x 2
2x 2
dx dx dx2 dx
df0 (x)
+l (l + 1) =0
dx
d3 f0 (x) d2 f0 (x) df0 (x)
) 1 x2 3
4x 2
+ (l (l + 1) 2) = 0 (12.178)
dx dx dx
df0 (x) d1 Pl (x)
f1 (x) = =
dx dx1
we …nd
d2 f1 (x) df1 (x)
1 x2 4x + (l (l + 1) 2) f1 (x) = 0
dx2 dx
and comparing this with the di¤erential equation for m=1
d2 u (x) du (x)
1 x2 4x + [l (l + 1) 2] u (x) = 0: (12.179)
dx2 dx
one can easily see that
d1 Pl (x)
u (x) = f1 (x) = :
dx1
(c) m=2 : the di¤erential equation

1 x2 u00 (x) 2 (m + 1) xu0 (x) + [l (l + 1) m (m + 1)] u (x) = 0:
(12.180)
becomes
1 x2 u00 (x) 6xu0 (x) + [l (l + 1) 6] u (x) = 0: (12.181)
Now let’s di¤erentiate the di¤erential equation we found for m = 1 with
respect to x
d d2 f1 (x) df1 (x)
1 x2 4x + (l (l + 1) 2) f1 (x) = 0
dx dx2 dx
d3 f1 (x) d2 f1 (x) df1 (x)
) 1 x2 3
6x + (l (l + 1) u0 (x) 6) =0
dx dx2 dx
df1 (x) d2 Pl (x)
f2 (x) = =
dx dx2
we …nd
d2 f2 (x) df2 (x)
1 x2 6x + [l (l + 1) 6] f2 (x) = 0: (12.182)
dx2 dx
which is the same as the di¤erential equation we found for m = 2
d2 u (x) du (x)
1 x2 6x + [l (l + 1) 6] u (x) = 0: (12.183)
dx2 dx
and the solution is given by
df1 (x) d2 Pl (x)
u (x) = f2 (x) ==
dx dx2
Following the same procedure one can easily establish that the solution to
the di¤erential equation
1 x2 u00 (x) 2 (m + 1) xu0 (x) + [l (l + 1) m (m + 1)] u (x) = 0:
can be expressed as
dm Pl (x)
u (x) = : (12.184)
dxm
Substituting this into
m=2
y (x) = 1 x2 u (x) (12.185)
the solution to the Associated Legendre Di¤erential equation
m2
1 x2 y 00 2xy 0 + l (l + 1) y = 0: (12.186)
1 x2
can then be expressed as
m=2 dm
y (x) = 1 x2 Pl (x) (12.187)
dxm
12.9. SPHERICAL HARMONICS AND THE ADDITION THEOREM 333
The Associated Legendre Functions: The solutions to the Associated Legen-

dre Di¤erential equation are the Associated Legendre functions and are given
by
m
m=2 d
Plm (x) = 1 x2 Pl (x) : (12.188)
dxm
Note that when m = 0; we …nd the Legendre Polynomials
Pl0 (x) = Pl (x) : (12.189)
Orthogonality of the Associated Legendre Functions:
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.
12.9 Spherical Harmonics and the addition the-

orem
Consider the second order di¤erential equation
d2 (')
+ m2 (') = 0; (12.191)
d'2
the solution which is given by
(') = A (m) eim' : (12.192)
Let’s also consider the associated Legendre di¤erential equation
d2 y (x) dy (x) m2
1 x2 2x + l (l + 1) y (x) = 0 (12.193)
dx2 dx 1 x2
which we determined the solution to be

m=2 dm
y (x) = Plm (x) = 1 x2 Pl (x) : (12.194)
dxm
Now introducing the transformation de…ned by
y (x) = ( ) ; x = cos ( ) ) sin2 ( ) = 1 x2 ; dx = sin ( ) d (12.195)

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 ( ))
The set of functions de…ned by
Ylm ( ; ') = C (l; m) Plm (cos ( )) eim' ;
are known as Spherical harmonics, where

s
m 2l + 1 (l m)!
C (l; m) = ( 1) ;
4 (l + m)!
are determined from the normalization condition

Z Z 2
Ylm ( ; ') Ylm ( ; ') sin ( ) d d' = 1:
=0 '=0
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
or form a complete set of functions,
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
Thus any …nite function, f ( ; ')
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
Using the completeness relation for the continuos angular space

Z Z2
0 0 0
; '0 ; '0 sin d 0 d'0 = 1; (12.205)
0 0
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
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
Using the Dirac’s notation

0 0
Ylm ( ; ') = h ; ' jYlm i ; Ylm ; '0 = hYlm j ; '0 ;
0 0
f ; '0 = ; '0 f i ; (12.209)
one can write
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
are the expansion coe¢ cients.

The Inverse-Distance Between Two Points: Consider two points described
by the vectors, ~r and ~r0 ; as shown in Fig. 12.3. In spherical coordinates these
two points are described by (r; ; ') and r0 ; 0 ; '0 ;respectively. The angle
between these two vectors, ; (which is related to the angles describing each of
the two vectors) can be used to determine the inverse-distance between these two
points. In view of the equation we determined that relates the inverse distance
and the Legendre polynomials, one can write
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
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 :
The Addition Theorem for Spherical Harmonics: A mathematical result of

considerable interest and the use is called the addition theorem for spherical
Harmonics. The addition theorem states that
l
X
4 0
Pl (cos ( )) = Ylm ( ; ') Ylm ; '0 ; (12.214)
2l + 1
m= l
where Pl are the Legendre polynomials and cos ( )

0 0
cos ( ) = cos ( ) cos + sin ( ) sin cos (' '0 ) : (12.215)
We can prove this theorem using
~r = r sin ( ) cos (') x
^ + r sin ( ) sin (') y^ + r cos ( ) z^ (12.216)
and
0 0 0
~r0 = r0 sin cos ('0 ) x
^ + r0 sin sin ('0 ) y^ + r0 cos z^: (12.217)
We note that
p
j~r~r0 j = (x x0 )2 + (y y 0 )2 + (z z 0 )2
n
2
= r sin ( ) cos (') r0 sin 0 cos ('0 )
o
0 2 0 2
+ r sin ( ) sin (') r0 sin sin ('0 ) + r cos ( ) r0 cos (12.218)

0
j~r ~r0 j = r2 sin2 ( ) cos2 (') + r02 sin2 cos2 ('0 )
0 0
2rr0 sin ( ) sin cos (') cos ('0 ) + r2 sin2 ( ) sin2 (') + r02 sin2 sin2 ('0 )
0 0
2rr0 sin ( ) sin sin (') sin ('0 ) + r2 cos2 ( ) + r02 cos2
0
2rr0 cos ( ) cos (12.219)
j~r ~r0 j = r2 sin2 ( ) cos2 (') + sin2 (') + cos2 ( )

0 0
+r02 sin2 cos2 ('0 ) + sin2 ('0 ) + cos2
0 0
2rr0 sin ( ) sin [cos (') cos ('0 ) + sin (') sin ('0 )] + cos ( ) cos
(12.220)
Applying the relations
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)
j~r ~r0 j = r2 + r02 2rr0 cos ( ) ; (12.223)
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:
Example 12.8 A solid sphere of radius R has a constant volume-charge density

. The sphere is centered at the origin of coordinates. The point P is a
distance d > R from the center of the sphere. Find an expression for the
electrostatic potential at the point P , P , due to the charged sphere.
Solution: The electrostatic potential at point P is given by the integral ex-

pression
Z
1 dq 0
V (~r) = : (12.225)
4 0 j~r ~r0 j
vol
For a uniform charge density, ; we have
dq 0 0
= ) dq 0 = d 0
= r02 dr0 sin d 0 d'0 (12.226)
d 0
Figure 12.4: A charged solid sphere with uniform charge density, :
where we used spherical coordinates for the in…nitessimal value d 0 : Using

this relation the potential can then be expressed as
Z R Z Z 2
1 0
V (~r) = r02 dr0 sin d 0 d'0 : (12.227)
4 0 0 =0 '=0 j~r ~r0 j
The relation
X1 l
1 r<
= P (cos ) ;
l+1 l
(12.228)
j~r ~r0 j r
l=0 >
for a point outside the sphere (r = d > r0 ) give
1
X
1 r0l
= Pl (cos ) : (12.229)
j~r ~r0 j dl+1
l=0
and the potential becomes
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
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
We note that for, l = m = 0, the relations,

s
0 m 2l + 1 (l m)! m 0 0
Ylm ; '0 = ( 1) P cos eim' (12.234)
4 (l + m)! l
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)
Now using the orthonormality relation for the spherical harmonics,

Z Z 2
0 0 0
hYlm j Yl0 m0 i = Ylm ; '0 Yl0 m0 ; '0 sin d 0 d'0
=0 '=0
= ll0 mm0 (12.237)
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
so that the potential becomes

p 1 l Z
4 X X 4 R
r0l+2 0
V (~r) = 0l 0m Ylm ( ; ') dr :
4 0 2l + 1 0 dl+1
l=0 m= l
(12.239)
All the terms in the summation becomes zero except when l = 0 and
m = 0 because of the the Kronecker delta function. Thus we …nd for the
potential
p Z p
4 Y00 ( ; ') R 02 0 4 Y00 ( ; ') R3
V (~r) = r dr =
0 d 0 0 3d
R3
) V (~r) = : (12.240)
3 0d
In terms of the total charge, Q; for a uniform distribution the charge
density is given by
Q
4 3
= ; (12.241)
3 R
and the potential becomes
Q
V (~r) = ; (12.242)
4 0d
which is the potential of a point charge.
12.10 The Method of Frobenius and the Bessel

equation
When the Standard Power Series Solutions fail : The power series solution
1
X
y (x) = an xn = a0 + a1 x + a2 x2 + a3 x3 :::; (12.243)
n=0
where n 0. This is applicable when the di¤erential equation involves no

singular point for 8 x " R: There are di¤erential equations which are not de…ned
for 8 x " R. The solutions of such di¤erential equations involves a factor with
negative or fractions as exponent for x. In such cases the standard power series
solution fails and we need to use a slightly modi…ed method known The Method
of Frobenius which we will see next. But before we do so we …rst consider some
simple di¤erential equations where the solutions do not satisfy the standard
power series solution.
Example 12.9 Solve the di¤erential equation

y
y0 + =0 (12.244)
x
Solution: First we note that this di¤erential equation is not de…ned at x = 0.

But it can be solved easily as follows:
y dy 1 dy dx
y0 + =0) = y) =
x dx x y x
Z y Z x 1
dy dx y x x
) = ) ln = ln = ln
y0 y x0 x y 0 x0 x0
y x0 y x0
) ln = ln ) =
y0 x y0 x
1
) y = (x0 y0 ) x (12.245)
Noting that y0 x0 = a1 (constant), the solution can be written as
1
y = a1 x ; (12.246)
which clearly show that x has a negative exponent and the standard power
series expansion does not work!
3y
y0 =0 (12.247)
2x
Solution: Here also the di¤erential equation is singular at x = 0. The solution
is given by
Z y Z
3y dy 31 dy 3 dx dy 3 x dx
y0 =0) = y) = ) =
2x dx 2x y 2 x y0 y 2 x0 x
3=2 3=2
y 3 x x y x
) ln = ln = ln ) =
y0 2 x0 x0 y0 x0
3=2
) y = y0 x0 x3=2 (12.248)
3=2
Noting that y0 x0 = a1 (constant), the solution can be written as
y = a1 x3=2 = x1=2 (a1 x) : (12.249)

which also involve x with a fraction exponent.
The Method of Frobenius: When we face a di¤erential equation which are
not de…ned for all x " R like the examples we saw above, we use the method of
Frobenius. We used a generalized power series of the form
1
X 1
X
y (x) = xs an xn = an xn+s ; (12.250)
n=0 n=0
where s is a number to be determined along with the expansion coe¢ cients, an .

It may be positive, negative, fraction, or even complex number although we do
not consider the complex number case here.
Example 12.11 The Bessel Di¤ erential equation: As an application of the

Method of Frobenius we will solve the Bessel di¤erential equation
x2 y 00 + xy 0 + x2 p2 y = 0; (12.251)
where p is a constant parameter characterizing the di¤erential equation.

The Bessel di¤erential equation, usually, derived from the Laplace equa-
tion in cylindrical coordinates which we will see in the next chapter.
Solution: Using a generalized power series
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
so that substituting these expressions into the Bessel di¤erential equation,

we …nd
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)
Expanding the …rst series in the above expression

1 h
X i h i h i
2 2 2
(n + s) p2 an xn+s = (0 + s) p2 a0 xs + (1 + s) p2 a1 xs+1
n=0
1 h
X i
2
+ (n + s) p2 an xs+n
n=2
h i 1 h
X i
2 2
= s2 p2 a0 xs + (1 + s) p2 a1 xs+1 + (n + s) p2 an xn+s
n=2
(12.257)
and replacing n = m + 2 and noting that for n = 2 ) m = 0; we can write

1 h
X i h i
2 2
(n + s) p2 an xn+s = s2 p2 a0 xs + (1 + s) p2 a1 xs+1
n=0
1 h
X i
2
+ (m + 2 + s) p2 am+2 xm+s+2 : (12.258)
m=0
Since m is a dummy variable we can replace it by n and rewrite the above

expression as
1 h
X i h i
2 2
(n + s) p2 an xn+s = s2 p2 a0 xs + (1 + s) p2 a1 xs+1
n=0
1 h
X i
2
+ (n + 2 + s) p2 an+2 xn+s+2 : (12.259)
n=0
Now substituting this expression into

1 h
X i 1
X
2
(n + s) p2 an xn+s + an xn+s+2 = 0 (12.260)
n=0 n=0
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
There follows that

h i
2
s2 p2 a0 xs = 0; (1 + s) p2 a1 xs+1 = 0;
1 nh
X i o
2
(n + 2 + s) p2 an+2 + an xn+s+2 = 0; (12.263)
n=0
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)
Upon solving the …rst equation, we …nd
s= p (12.265)
and substituting this value into the second equation

h i
2
(1 + s) p2 a1 = 0 (12.266)
we …nd h i
2
(1 p) p2 a1 = 0 ) [1 2p] a1 = 0: (12.267)
For this to be true independent of p, we must have
a1 = 0: (12.268)
The third equation

h i
2
(n + 2 + s) p2 an+2 + an = 0 (12.269)
leads to the recursion relation

1
an+2 = 2 an : (12.270)
(n + 2 + s) p2
This recursion relation gives to two di¤erent functions for the two di¤erent
values of s (= p).
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)
where m = 1; 2; 3::Next we shall consider the …rst few terms:

a0
m = 1 ) a2 =
22 1! (1 + p)
a2 a0
m = 2 ) a4 = = 4
22 2 (2 + p) 2 2! (1 + p) (2 + p)
a4 a0
m = 3 ) a6 = 2
= 6
(12.273)
2 3 (3 + p) 2 3! (1 + p) (2 + p) (3 + p)
Recalling that the Gamma function
(p + 1) = p (p) ; (p + 2) = (p + 1) (p + 1) = p (p + 1) (p) ; (12.274)
(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)
which can be simpli…ed into

1 1 x 2 1 x 4
y (x) = a0 xp (p + 1) +
(p + 1) 1! (p + 2) 2 2! (p + 3) 2
1 x 6
+ ::: : (12.280)
3! (p + 4) 2
Noting that
(1) = (2) = 1; n! = (n + 1) (12.281)
the above expression can be rewritten as

x p 1
y (x) = a0 2p (p + 1)
2 (1) (p + 1)
1 x 2 1 x 4 1 x 6
+ + ::: :
(2) (p + 2) 2 (3) (p + 3) 2 (4) (p + 4) 2
Introducing the function, Jp (x) ; de…ned by
y (x)
Jp (x) = (12.282)
2p a0 (p + 1)
we …nd
x p 1 1 x 2
Jp (x) =
2 (1) (p + 1) (2) (p + 2) 2
1 x 4 1 x 6
+ + ::: : (12.283)
(3) (p + 3) 2 (4) (p + 4) 2
1
X n 2n+p
( 1) x
Jp (x) = : (12.284)
n=0
(n + 1) (n + 1 + p) 2
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
If p is not an integer, Jp (x) is a series starting with xp and J p (x) is a series

starting with x p . Then Jp (x) and J p (x) are two independent solutions and
the linear combination of these two functions is a general solution. The com-
bination is is called either the Neumann or Weber function denoted by either
Np (x) or Yp (x) and is given by
cos ( p) Jp (x) J p (x)
Np (x) = Yp (x) = : (12.286)
sin ( p)
However, if p is an integer, then the …rst few terms in J p (x) are zero because
(n + 1 p) is the gama function of a negative integer, which is in…nite.
J p (x) = ( 1)p Jp (x) (12.287)
and Jp (x) and J p (x) are not independent solutions.
The zeroes of the Bessel function: The zeroes of the bessel function f 1; 2; 3;
:::g are the values of x at which
Jp (x) = 0: (12.288)
in the …gure below the points on the x axis where the Jp (x) intersects (i.e.
Jp (x) = 0) are the zeroes of the Bessel function.
Figure 12.5: Bessel function, Jp (x) ; for p = 0; 3; 6
12.11 The orthogonality of the Bessel Functions

The Orthogonality and Normalization: In the previous section we did solve the
Bessel Di¤erential Equation
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
Figure 12.6: The zeros of the bessel function.
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)
Integrating this equation with respect to x from 0 to 1

Z 1 Z 1
d d d d
Jp ( j x) x Jp ( i x) dx Jp ( i x) x Jp ( j x) dx
0 dx dx 0 dx dx
Z 1
+ 2i 2
j Jp ( i x) Jp ( j x) xdx = 0: (12.298)
0
Applying integration by parts,

Z 1 Z 1
dv 1 du
u dx = uvj0 v dx; (12.299)
0 dx 0 dx
one can write
Z 1 1
d d d
Jp ( j x) x Jp ( i x) dx = Jp ( j x) x (Jp ( i x)) (12.300)
0 dx dx dx 0
Z 1
d d
x (Jp ( i x)) (Jp ( j x)) dx
0 dx dx
and
Z 1 1
d d d
Jp ( i x) x Jp ( j x) = Jp ( i x) x (Jp ( j x)) (12.301)
0 dx dx dx 0
Z 1
d d
x (Jp ( j x)) (Jp ( i x)) dx:
0 dx dx
Noting that i are the zeros of the Bessel function and
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
There follows that, clearly, for i 6= j; since

2 2
i j 6= 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
Then for i = j + ; Eqs. (12.300) & (12.301) become

Z 1
d d
Jp ( j x) x Jp (( j + ) x) dx (12.311)
0 dx dx
Z 1
d d
= x (Jp (( j + ) x)) (Jp ( j x)) dx
0 dx dx
and
Z 1 1
d d d
Jp (( j + ) x) x Jp ( j x) = Jp (( j + ) x) x (Jp ( j x))
0 dx dx dx 0
Z 1
d d
x (Jp (
j x)) (Jp (( j + ) x)) dx
0 dx dx
Z 1
d d d
= Jp (( j + )) (Jp ( j x)) x (Jp ( j x)) (Jp (( j + ) x)) dx:
dx x=1 0 dx dx
(12.312)
Now substituting Eqs. (12.310), (12.311), and (12.312) into Eq. (12.298), we
…nd
Z 1
d
Jp (( j + )) (Jp ( j x)) + 2 j + 2 Jp (( j + ) x) Jp ( j x) xdx = 0:
dx x=1 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 + )
The Taylor series expansion for f ( ) about = 0, we have

2
df ( ) d2 f ( ) 3
d3 f ( )
f ( ) = f ( )j =0 + + + + ::: (12.315)
d =0 2! d 2 =0 3! d 3 =0
then for f ( ) = Jp ( j + ) ; we have
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:
For x << 1 : We recall the Bessel Function

x p 1 1 x 2
Jp (x) =
2 (1) (p + 1) (2) (p + 2) 2
1 x 4 1 x 6
+ + ::: : (12.322)
(3) (p + 3) 2 (4) (p + 4) 2
x 2
so that dropping all higher order terms beginning from 2 for x << 1; we
…nd
x p 1
Jp (x) ' : (12.323)
2 (1) (p + 1)
Using
(1) = 1; (p + 1) = p! (12.324)
we …nd
1 x p
Jp (x) ' : (12.325)
p! 2
Also using
cos ( p) Jp (x) J p (x)
Np (x) = Yp (x) = : (12.326)
sin ( p)
and
x p 1 1 x 2
J p (x) =
2 (1) (1 p) (2) (2 p) 2
1 x 4 1 x 6
+ + ::: : (12.327)
(3) (3 p) 2 (4) (4 p) 2
we can show that p
(p 1)! 2
Np (x) ' (12.328)
x
For x >> 1
r r
2 2p + 1 2 2p + 1
Jp (x) ' cos x ; Np (x) ' sin x
x 4 x 4
Note:Many di¤erential equations occur in practice that are not of the stan-
dard form of the Bessel di¤erential equation that we saw here but whose solution
can be written in terms of the Bessel functions. For example the di¤erential
equation
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.
12.12 Fuch’s theorem for 2-nd DE

So far we have seen a second order di¤erential equation of the form
y 00 + f (x) y 0 + g (x) y = 0 (12.331)
can be solved using the standard power series expansion method (e.g. the Legen-
dre equation) or the more general Frobenius method (e.g. the Bessel equation).
If you think that the Frobenius method can be used to solve any second order
di¤erential equation, then you are wrong. There is a necessary condition that
must be satis…ed if the di¤erential equation is solvable using Frobenius method.
Fuch’s Theorem determines the necessary and su¢ cient condition to apply the
Frobenius Method. It states that for the di¤erential equation
y 00 + f (x) y 0 + g (x) y = 0; (12.332)
the solution can be obtained using the Frobenius method if and only if the
functions de…ned by F (x) = f (x) y 0 and G (x) = g (x) y can be expressed as
power series (i.e. a convergent power series)
1
X 1
X
F (x) = f (x) y 0 = an xn ; G (x) = g (x) y = bn x n : (12.333)
n=0 n=0
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
Solution: We note that the di¤erential equation has the form
y 00 + f (x) y 0 + g (x) y = 0; (12.339)
where
1 2
f (x) = ; g (x) = (12.340)
x2 x3
Using the the given solution
y(x) = e1=x (12.341)
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)
Then the gravitational potential energy can be expressed as
U= mg [l cos ( (t)) l0 ] (12.345)
and the kinetic energy

1 2
T = m l2 _ + l_2 2
: (12.346)
2
Then the Lagrangian becomes
1 2
L=T U= m l2 _ + l_2 + mg [l cos ( ) l0 ] : (12.347)
2
Figure 12.7: The lengthening pendulum.
Using Euler-Lagrange equations

d @L @L
= 0; (12.348)
dt @_ @
we …nd
d d 2_
ml2 _ + mgl sin ( ) = 0 ) l + gl sin ( ) = 0: (12.349)
dt dt
The length is increasing at a constant rate (v), which means
dl
l_ = v ) dt = (12.350)
v
then transforming the DE from t to l; we …nd
d d d2 d
v l2 v + gl sin ( ) = 0 ) l2 v 2 2
+ 2lv 2 + gl sin ( ) = 0
dl dl dl dl
d2 2d g
) 2
+ + 2 sin ( ) = 0: (12.351)
dl l dl lv
For small angle , we have sin ( ) ' and the DE becomes
2
d 2d g
2
+ + 2 = 0: (12.352)
dl l dl lv
We now consider the DE
d2 y 1 2a dy 1 2 a2 p2 c2
+ + bcxc + y=0 (12.353)
dx2 x dx x2
12.12. FUCH’S THEOREM FOR 2-ND DE 357
the solution of which is the Bessel function given by
y (x) = xa Zp (bxc ) : (12.354)
replacing x by l and y by ; we have
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
and one can write p

1 gl
(l) = p Z1 2 ; (12.360)
l v
which is the solution to the DE for the lengthening pendulum.
Chapter 13
A partial di¤erential
equation
A partial di¤erential equation (PDE) is an equation in some unknown function

of more than one variable, say f = V (x; y; z; t), involving possibly di¤erent
orders of partial derivatives of that function. The following are some examples
of PDEs that would …nd in di¤erent branches of physics.
13.1 PDE in physics

Gauss’Law for the Electric Field : The di¤erential form of Gauss’law states that
~ (x; y; z) of a volume charge distribution (x; y; z) satis…es
the electric …eld, E
the PDE
r E ~ (x; y; z) = (x; y; z) : (13.1)
0
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
we may write Gauss’Law as
@Ex (x; y; z) @Ey (x; y; z) @Ez (x; y; z) (x; y; z)

+ + = : (13.3)
@x @y @z 0
~ (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
Gauss’s Law can be expressed as

(x; y; z) (x; y; z)
r [ rV (x; y; z)] = ) r2 V (x; y; z) = (13.5)
0 0
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
This PDE is known as Poisson’s equation.

Laplace’s Equation: If the charge density (x; y; z) = 0, the Poisson’s equa-
tion becomes Laplace equation. It is given by
@ 2 V (x; y; z) @ 2 V (x; y; z) @ 2 V (x; y; z)

+ + = 0: (13.7)
@x2 @y 2 @z 2
The di¤ usion or heat ‡ow equation: Suppose one has a function u which
describes the temperature at a given location (x; y; z). This function will change
over time as heat spreads throughout space. The heat equation is used to
determine the change in the function u over time. In a special case of heat ‡ow
in an isotropic and homogeneous medium in a 3-dimensional space, the heat
equation is given by
@u @ 2 u (x; y; z; t) @ 2 u (x; y; z; t) @ 2 V (x; y; z; t)

+ + = 0: (13.8)
@t @x2 @y 2 @z 2
where is the thermal di¤usivity, a material-speci…c quantity depending on the
thermal conductivity, the mass density, and the speci…c heat capacity
Wave equation:Consider a string with linear mass density under tension
force T displaced from the equilibrium as shown in the …gure below:
If wave propagating on this string is described by the wave equation
@2f 1 @2f
= (13.9)
@z 2 v 2 @t2
where r
T
v= : (13.10)
is the speed of the wave.

13.2. LAPLACE’S EQUATION IN CARTESIAN COORDINATES 361
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.
13.2 Laplace’s equation in Cartesian coordinates

Before we begin to solve Laplace’s equation in Cartesian coordinates, it is im-
portant to revise how we determine the solution for second order HLDE with
constant coe¢ cients.
d2 f (x)
= k 2 f (x) : (13.14)
dx2
Solution: Assuming a solution of the form

x
f (x) = Ae (13.15)

2
k 2 = 0: (13.16)
For the plus case the solutions are
1 = k; 2 = k (13.17)
and the general solution is given by

kx
f (x) = A1 e + A2 ekx : (13.18)
For the minus case the solution to the indicial equation are
1 = ik; 2 = ik (13.19)
and the general solution becomes
f (x) = B1 cos (kx) + B2 sin (kx) : (13.20)
Laplace’s equation in Cartesian coordinates is given by
@ 2 V (x; y; z) @ 2 V (x; y; z) @ 2 V (x; y; z)

+ + = 0:
@x2 @y 2 @z 2
One can express V (x; y; z) as a product of three independent functions A (x) ;

B (y) ; and C (z)
V (x; y; z) = X (x) Y (y) Z (z) : (13.21)
Substituting this expression into Laplace’s equation, we have
d2 X (x) d2 Y (y) d2 Z (z)

Y (y) Z (z) 2
+ X (x) Z (z) 2
+ X (x) Y (y) = 0: (13.22)
dx dy dz 2
and dividing by X (x) Y (y) Z (z) ; we …nd
1 d2 X (x) 1 d2 Y (y) 1 d2 Z (z)

+ + = 0: (13.23)
X (x) dx2 Y (y) dy 2 Z (y) dz 2
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
1 d2 X (x) 1 d2 Y (y) 1 d2 Z (z)

= kx2 ; = ky2 ; = kz2 ; (13.24)
X (x) dx2 Y (y) dy 2 Z (y) dz 2
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
Similarly, for the y dependent part, we …nd
d2 Y (y)
+ ky2 Y (y) = 0: (13.27)
dy 2
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
For the z dependent part
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 ;
Then the general solution to Laplace’s equation in Cartesian coordinates is given

by
XXX
V (x; y; z) = Xkx (x) Yky (y) Zkz (z) (13.31)
kx ky 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
kx2 + ky2 kz2 = 0: (13.32)
We shall see how these constants are determined in the next example.
Example 13.2 Consider an in…nitely long, rectangular waveguide of width, w;

and height h with the top surface held at the constant potential, V (y =
h) = V o, and with all other sides grounded. A vacuum pump has been
connected to the waveguide to remove most of the air within it. Find the
electrostatic potential, V (x; y; z); everywhere within the waveguide. Inside
the the wave guide the electric potential satis…ed Laplace’s equation
@ 2 V (x; y; z) @ 2 V (x; y; z) @ 2 V (x; y; z)

+ + = 0:
@x2 @y 2 @z 2
Figure 13.1: A rectangular wave guide with width w and height h. It is in…nitely
long along the z direction.
Solution: We already determined solution to Laplace’s equation in Cartesian

coordinates. To …nd the potential inside the waveguide, we need to state
the given boundary conditions and impose to the general solution to de-
termine the constants. To this end, we note that solution is given by
X
V (x; y; z) = Xkx (x) Yky (y) Zkz (x) (13.33)
kx ;ky ;kz
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
V (0; y; z) = V (w; y; z) = V (x; 0; z) = 0; V (x; h; z) = V0 : (13.35)
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
Using the solutions in Eqs. (13.28), we …nd
Xkx (0) = 0 ) [Ckx cos (kx x) + Dkx sin (kx x)]x=0 = 0

) Ckx = 0
Xky (w) = 0 ) [Ckx cos (kx x) + Dkx sin (kx x)]x=W = 0
) Ckx + Dkx sin (kx w) = 0 ) Dkx sin (kx w) = 0
) sin (kx w) = 0
n
) kx = ; where n = 0; 1; 2; 3:::
w
Yky (0) = 0 ) Cky cos (ky y) + Dky sin (ky y) y=0
=0
) Cky = 0 (13.39)
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
) lim Zkz (z) = Constant (13.40)

z! 1
and using the result in Eq. (13.30), one …nds
lim Akz ekz z + Bkz e kz z

= lim Akz ekz z = Constant ) kz = 0:
z! 1 z! 1
(13.41)
n
Substituting kx = w ; and kz = 0 into
kx2 + ky2 kz2 = 0: (13.42)
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)
Note that we have used the relation
ei(ix)
e i(ix) ex e x
ex e x
sin (ix) = = =i
2i 2i 2
) sin (ix) = i sinh (x)
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
where n = 1; 2:::.Multiplying both sides by sin nw x and integrating with

respect to x over the width of the waveguide, we have
1
X Z W Z W
h n m m
Hn sinh n sin x sin x dx = V0 sin x dx
n=1
w 0 w w 0 w
(13.47)
so that noting that
Z W Z
n m w W n m x
sin x sin x dx = sin x sin x d
0 w w 0 w w w
Z
w w w
= sin (nu) sin (mu) d (u) = mn = mn (13.48)
0 2 2
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 )
Then the potential becomes

1
4V0 X sin x y
w (2n + 1) sinh w (2n + 1)
V (x; y; z) = h
:
n=1
(2n + 1) sinh w (2n + 1)
13.3. LAPLACE’S EQUATION IN SPHERICAL COORDINATES 367
13.3 Laplace’s equation in spherical coordinates

We have seen an example in the previous lecture illustrating how we solve the
Laplace’s equation in Cartesian coordinates
@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
In spherical coordinates, the Laplacian is given by
Figure 13.2: A point in space described by spherical coordinates, r; ; and ':
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)
Substituting the product function
V (r; ; ') = R (r) ( ) (')

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
( ) (') d dR (r) R (r) (') d d ( )

r2 + sin
r2 dr dr r2 sin d d
R (r) ( ) d2 (')
+ 2 2 = 0: (13.54)
r sin d'2
Dividing the entire equation by R (r) ( ) ('), we have
1 d dR (r) 1d d ( )
r2 + sin
r2 R (r) dr dr r2 sin ( )d d
1 d2 (')
+ 2 = 0; (13.55)
r2 sin (') d'2
and multiplying by r2
1 d dR (r) 1 d d ( )
r2 + sin
R (r) dr dr sin ( )d d
2
1 d (')
+ 2 = 0: (13.56)
sin (') d'2
F1 (r) + F2 ( ; ') = 0 (13.57)
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,
F1 (r) = l (l + 1) ; F2 ( ; ') = l (l + 1) : (13.59)
There follows that

1 d dR (r) d dR (r)
r2 = l (l + 1) ) r2 l (l + 1) R (r) = 0
R (r) dr dr dr dr
d2 R (r) dR (r)
) r2 + 2r l (l + 1) R (r) = 0 (13.60)
dr2 dr
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
where m is a constant. Simplifying these equations, we …nd
d2 (')
+ m2 (') = 0; (13.63)
d'2
the solution of which are given by
Am cos (m') + Bm sin (m')
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
This is the associated Legendre di¤erential equation, which we determined the

solution in the previous chapter, to be the associated Legendry function Plm (x) :
Therefore, for the dependence, we …nd
lm ( ) = Plm (cos ) : (13.68)
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
Then the di¤erential equation
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
There follows that
(n + s) (n + s + 1) l (l + 1) = 0 ) s2 +(2n+1)s+n(n+1) l(l+1) = 0 (13.73)

and solving for s, we …nd

s1 = (n + l + 1) ; s2 = l n (13.74)
Therefore, the solution to the di¤erential equation becomes for s1 = [n + (l + 1)]
X1 P1
( n=0 an ) Al
R1 (r) = an rn+s1 = l+1
= l+1 (13.75)
n=0
r r
and for s2 = l n
1 1
!
X X
n+s2
R2 (r) = an r = an r l = Al r l : (13.76)
n=0 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.
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)
for an electric …eld pointing along the positive z-direction, we have

~
E = Eo z^ = rV (x ! 1; y ! 1; z ! 1)
) V (x ! 1; y ! 1; z ! 1) = Eo z: (13.83)
We need to express this potential in spherical coordinates
V (r ! 1; ; ') = Eo r cos : (13.84)
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)
Using this result we may want to rewrite the potential as

1 X
X l
a2l+1
V (r; ; ') = Bl r l Plm (cos )
rl+1
l=0 m= l
(Am cos (m') + Bm sin (m')) : (13.88)

13.4. LAPLACE’S EQUATION IN CYLINDRICAL COORDINATES 373
Now using the second boundary condition
V (r ! 1; ; ') = E0 r cos (13.89)
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
Expanding the series and including the constant A0 into Bl ; we have

1
X
B0 + B1 r cos + lim Bl rl Pl0 (cos ) = E0 r cos : (13.93)
r!1
l=2
comparing the right and left hand side of this equation we …nd
B0 = 0; B1 = E0 ; Bl = 0 for l > 1: (13.94)
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
13.4 Laplace’s equation in cylindrical coordinates

Consider the Laplace equation in cylindrical coordinates,
1 @ @V 1 @2V @2V
r2 V = s + 2 2
+ =0 (13.97)
s @s @s s @' @z 2
Using separation of variables
V (s; '; z) = R (s) (') Z (z)

Figure 13.3: A point in space described by spherical coordinates, s; '; and z:
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
1 d dR (s) 1 d2 Z (z) 1 d2 (')

s2 s + + = 0; (13.99)
sR (s) ds ds Z (z) dz 2 (') d'2
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
We can write the di¤erential equation
1 d dR (s) 1 d2 Z (z)
s2 s + = p2 (13.101)
sR (s) ds ds Z (z) dz 2
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
which may put in the form

" #
2
2 d2 R (s) dR (s) qs 2
(qs) 2 + (qs) d (qs) + p R (s) = 0 (13.106)
d (qs) r0
Introducing the transformation of variable de…ned by r = qs; we have

d2 R (r) dR (r) r2
r2 +r + p2 R (r) = 0 (13.107)
dr2 dr ro2
If the radius of the cylinder is, ro ; then we can introduce a dimensionless variable,
r
x= ; dr = ro dx; R (r) = y (x) (13.108)
ro
so that
d2 y (x) dy (x)
x2 +x + x2 p2 y (x) = 0: (13.109)
dx2 dx
Since
r
0 r ro ; and x = (13.110)
ro
the solution to the di¤erential equation
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
the solution is given by
(') = Ap cos (p') + Bp sin (p') (13.114)
and for the z-dependence

2
1 d2 Z (z) q
= (13.115)
Z (z) dz 2 r0
the solution is given by

q q
Z (z) = Cq e r0 z + Dq e r0 z
: (13.116)
The trigonometric functions must be single valued, this requires,
cos (p') = cos [p (' + 2 )] ; sin (p') = sin [p (' + 2 )] (13.117)
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
J n (x) = ( 1)n Jn (x) : (13.118)
Note that the Laplace’s equation
1 d dR (s) 1 d2 Z (z) 1 d2 (')

s2 s + + = 0; (13.119)
sR (s) ds ds Z (z) dz 2 (') d'2
using
2
1 d2 Z (z) q 1 d2 (')
= ; = p2
Z (z) dz 2 r0 (') d'2
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; '; z) = A0 ln (s)

XX q
z q
z qs
+ Cq e r0
+ Dq e r0
(An cos (n') + Bn sin (n')) Jp :
q>0 n>0
ro
(13.122)
Example 12.4 A right, circular conducting cylindrical shell of radius r0 and
length L has its axis coincident with the z-axis and its ends at z = 0 and
z = L. All sides of the cylinder are grounded except for the face at z = 0,
which is maintained at a position-dependent potential speci…ed by
V (s; '; z = 0) = s cos ' (13.123)
Find the electrostatic potential everywhere inside the cylinder.

Solution: The general solution for the Laplace equation is given by
V (s; '; z) = A ln (s)

XX q
z q
z qs
+ Cq e r0
+ Dq e r0
(An cos (n') + Bn sin (n')) Jn :
q>0 n>0
ro
(13.124)
Now we write the boundary conditions
V (s; '; 0) = s cos (') ; V (s; '; L) = 0; and V (r0 ; '; z) = 0; (13.125)
Applying the …rst boundary condition, V (s; '; z = 0); we have
V (s; '; 0) = A0 ln (s)

XX qs
+ (Cq + Dq ) (An cos (n') + Bn sin (n')) Jn = s cos (') :
q>0 n>0
ro
(13.126)
which leads to
qs qs
A0 = 0; n = 1; ) Jn = J1 (13.127)
ro ro
Therefore, we can rewrite the potential as

X q q qs
V (s; '; z) = Cq e r0 z + Dq e r0 z
cos (') J1 : (13.128)
q
ro
where we absorbed A1 into the constants. Using the second boundary

condition, V (s; '; z = L) = 0;
X q q qs
V (s; '; L) = Cq e r0 L + Dq e r0 L cos (') J1 = 0: (13.129)
q
ro
we …nd
q q q
Cq e r0 L + Dq e r0 L
= 0 ) Dq = Cq e2 r0 L : (13.130)
Then the simpli…ed expression for the potential can be rewritten as

X q q q qs
V (s; '; z) = Cq e r0 z e 2 r0 L e r0 z
cos (') J1 : (13.131)
q
ro
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
Now we apply the last boundary condition, V (s = r0 ; '; z) = 0

X q
V (r0 ; '; z) = F q sinh (L z) cos (') J1 (q) = 0: (13.133)
q
r0
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
To …nd Fi , we recall the orthonormality of the Bessel function

Z 1
Jp0 (qi ) Jp0 (qj )
hJp (qi x)j xJp (qj x)i = Jp (qi x) Jp (qj x) xdx = ij ij :
0 2
(13.137)
and the boundary condition
V (s; '; z = 0) = s cos ' (13.138)
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
Using the result
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)
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 qj are the zeroes of the Bessel function.
13.5 Poisson’s Equation

In the previous three lectures we have seen several examples from electrostatics
to determine the electric potential of a given charge distribution in a region
where there is no charge with speci…c boundary conditions. In such cases the
electric potential satisfy the Laplace’s equation. We have seen how to solve
Laplace’s equation in Cartesian coordinates where it is given by
@2 @2 @2
V (x; y; z) + V (x; y; z) + V (x; y; z) = 0; (13.145)
@x2 @y 2 @z 2
in cylindrical coordinates
1 @ @ 1 @2 @2
r V (r; '; z) + 2 V (r; '; z) + V (r; '; z) = 0; (13.146)
r @r @r r @'2 @z 2
and in spherical coordinates
1 @ @
r2 V (r; ; ')
r2 @r @r
1 @ @ @2
+ sin sin V (r; ; ') + V (r; ; ') = 0: (13.147)
r2 sin @ @ @'2
We know see how we determine the electric potential in a region where there
is some volume charge distribution, (~r). In such cases the electric potential,
V (x; y; z) ; satis…es the Poisson’s Equation given by
(~r)
r2 V (x; y; z) = ; (13.148)
0
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)
which we already know how to determine the solution. Suppose we determine

the particular solution, Vp , then the solutions is given by
V (x; y; z) = Vh (x; y; z) + Vp (x; y; z) : (13.154)
We can easily show that this solution satis…es the Poisson’s equation
r2 V (x; y; z) = r2 [Vh (x; y; z) + Vp (x; y; z)] = r2 Vh (x; y; z) + r2 Vp (x; y; z)

(13.155)
Since Vh is the solution to Laplace’s equation, we have
r2 V h = 0 (13.156)
and Vp is the particular solution
(~r)
r2 Vp = (13.157)
0
we can see that

(~r)
r2 V (x; y; z) = : (13.158)
0
Once we determine the homogenous and particular solutions we then write the
general solution and apply the boundary conditions.
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.
Figure 13.4: A point charge, Q, and a grounded conducting sphere of radius, R.
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
and we recall that the general solution is given by

1 X
X l
Cl
Vh (r; ; ') = Bl r l + Plm (cos ) (Am cos (m') + Bm sin (m')) :
rl+1
l=0 m= l
(13.161)
We recall that for a point charge, Q; located at a position, ~r0 ; the electric
potential at a distance, ~r; is given by
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
is that of the electrical potential of this point charge
1 Q
Vp (r; ; ') = VQ (r; ; ') = 1=2
: (13.166)
4 0 (r2 + a2 2ra cos )
The general solution to the Poisson’s equation can then be written as
V (r; ; ') = Vh (r; ; ') + Vp (r; ; ')

1 X
X l
Cl
V (r; ; ') = Bl r l + Plm (cos )
rl+1
l=0 m= l
1 Q
(Am cos (m') + Bm sin (m')) + 1=2
:
4 0 (r2 + a2 2ra cos )
(13.167)
Now we will apply the boundary conditions.

1. The electric potential must be …nite as, r ! 1:

2. At any point on the surface of the sphere the electric potential must be
zero since the sphere is grounded. That means
V (R; ; ') = 0: (13.168)
Applying the …rst boundary condition V (r ! 1; ; ') must be …nite (ba-

sically it must be zero since there are no other charges) gives
1 X
X l
Cl
lim V (r; ; ') = lim Bl r l + Plm (cos )
r!1 n!1 rl+1
l=0 m= l
1 Q
(Am cos (m') + Bm sin (m')) + 1=2
:
4 0 (r2 + a2 2ra cos )
1 X
X l
= lim Bl rl Plm (cos ) (Am cos (m') + Bm sin (m')) (13.169)
n!1
l=0 m= l
since this expression diverges, we must have
Bl = 0: (13.170)
Therefore, we may rewrite the electric potential as

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
+ 1=2
: (13.171)
4 0 (r2 + a2 2ra cos )
Now using the second boundary condition V (r = R; ; ') = 0, we have

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
+ 1=2
= 0: (13.172)
4 0 (R2 + a2 2Ra cos )
Recall that Generating Function for the Legendry Polynomials

X1 l
1 1 r<
=p = Pl (cos ) (13.173)
j~r ~r0 j r2 + r02 2rr0 cos rl+1
l=0 >
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
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
The above equation to be zero independent of '; we at least have the

coe¢ cients for Plm (cos ) must be independent of cos (m') and sin (m') :
This requires
Am = 0; Bm = 0 for m 6= 0: (13.176)
and the above expression becomes
1
X 1
Cl Q X Rl
V (R; ; ') = P l (cos ) + Pl (cos ) = 0: (13.177)
Rl+1 4 0 al+1
l=0 l=0
1
X Cl Q Rl
V (R; ; ') = 0 ) l+1
+ Pl (cos ) = 0: (13.178)
R 4 0 al+1
l=0
There follows that
Cl Q Rl Q R2l+1
l+1
+ = 0 ) Cl = : (13.179)
R 4 0 al+1 4 0 al+1
Therefore, the electric potential is given by
1
Q X 1 R2l+1 1 Q
V (r; ; ') = l+1 l+1
Pl (cos )+ 1=2
:
4 0 r a 4 0 (r2 + a2 2ra cos )
l=0
(13.180)
If we rewrite the series term as
1 1
Q X 1 R2l+1 QRa
X 1 R2l
Pl (cos ) = Pl (cos )
4 0 rl+1 al+1 4 0 r l+1 al
l=0 l=0
0 1
X
q r0l
= Pl (cos ) (13.181)
4 0 rl+1
l=0
2
with r0 = Ra and q 0 = QR
a ; we can apply the generating function for
Legendry polynomials
1
X
1 1 r<
=p = Pl (cos ) (13.182)
j~r ~r0 j r2 + r02 2rr0 cos r>
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)
Then the electric potential can be expressed as
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
14.1 Complex variables and functions

Real, Imaginary, and Complex Numbers: numbers that have the form
z = a + ib; (14.1)
with a and b real numbers and

p
i= 1 ) i2 = 1 (14.2)
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)
Rectangular and Polar Representation of Complex Numbers: In the rectan-

gular complex plane the real part a is the x coordinate and the imaginary part
b is the y coordinate
a = x; b = y (14.4)
In polar coordinate the complex number z = a + ib is represented by
a = r cos ; b = r sin (14.5)
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
z = x + iy = r exp (i ) = r cos + ir sin (14.6)
387
388 CHAPTER 14. FUNCTIONS OF COMPLEX VARIABLES
Figure 14.1: Rectangular and Polar representation of a complex number z =

x + iy:
where the magnitude and the phase are given by

p y
1
jzj = r = x2 + y 2 ; = tan (14.7)
x
Complex Conjugate and Magnitude of a Complex Number : The complex

conjugate of a complex number z = x + iy is denoted by z and is given by
z =x iy (14.8)
The magnitude of a complex number z is related to its complex conjugate by

p p
jzj = zz = x2 + y 2 (14.9)
Example 13.1 Consider the complex function
2z + 1
f (z) = (14.10)
z i
Find the real and imaginary parts of this function.
Solution: To …nd the real and imaginary part of a complex function for the
complex variable
z = x + iy (14.11)
we must be able to express the function f (z) as
f (x + iy) = u (x; y) + iv (x; y) : (14.12)

14.2. ANALYTIC FUNCTIONS 389
In order to do that we substitute, z = x + iy; into the given function

2 (x + iy) + 1 2x + 1 + i2y [2x + 1 + i2y] [x i (y 1)]
f (x + iy) = = =
(x + iy) i x + i (y 1) [x + i (y 1)] [x i (y 1)]
[2x + 1 + i2y] [x i (y 1)]
= 2
x2 + (y 1)
(2x + 1) x + 2y (y 1) + i [2xy (2x + 1) (y 1)]
= 2
x2 + (y 1)
(2x + 1) x + 2y (y 1) + i [ y + 2x + 1]
) f (x + iy) = 2 (14.13)
x2 + (y 1)
There follows that
(2x + 1) x + 2y (y 1) 2x y+1
u (x; y) = 2 ; v (x; y) = 2 (14.14)
x2 + (y 1) x2 + (y 1)
Example 13.2 Find the real and imaginary parts of the complex function
f (z) = z 1=4 : (14.15)
Solution: Whenever we are given exponential complex functions it is better to

use polar coordinates. Let
z = r exp (i ) ; (14.16)
where r and are related to Cartesian coordinates x and y by

p y
r = x2 + y 2 ; = tan 1 : (14.17)
x
The function can then be expressed as
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
u (r; ) = r1=4 cos ; v (r; ) = r1=4 sin (14.19)

4 4
14.2 Analytic Functions

A function f (x) is analytic (or regular or holomorphic or monogenic) in a region
of the complex plane if it has a (unique) derivative at every point of the region.
The statement "f (z) is analytic at a point z = z0 " means that f (z) has a
derivative at every point inside some small circle about z = z0 :
df f (z0 + z) f (z0 )
= lim (14.20)
dz z=z0
z!0 z
While the functions

d z2 d (cos z) d p z
= 2z; = sin z; 1 + z2 = p (chain rule)
dz dz dz 1 + z2
(14.21)
are analytic, the function
2
f (z) = jzj (14.22)
is not. So how one can determine whether a given function of complex variable
is analytic of not. We use The Cauchy-Riemann Conditions. It states that if
the complex function
f (z) = u (x; y) + iv (x; y) (14.23)
is analytic in a region, R; then in that region the following two conditions must
be satis…ed
@u @v @v @u
= ; = : (14.24)
@x @y @x @y
Next we shall derive these conditions. To this end, we note that for z = x + iy,
we can write
@f @f @z @f @ @f @f
= = (x + iy) ) = ;
@x @z @x @z @x @z @x
@f @f @z @f @f @f @f
= ) =i ) = i : (14.25)
@y @z @y @y @z @z @y
which leads to
@f @f
= i (14.26)
@x @y
And using f (z) = u (x; y) + iv (x; y) ; we may also write
@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
14.3 Important Terminologies

The following terminologies are important in the description of functions of
complex variable.
(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
@2v @2u @2u @2v

= ) = : (14.34)
@x@y @y 2 @y 2 @x@y
so that adding the two equations, one …nds
@2u @2u @2v @2v

+ = = 0 ) r2 u = 0 (14.35)
@x2 @y 2 @x@y @x@y
which is Laplace’s equation in two dimension. Functions satisfying Laplace’s
equation are called harmonic functions. Thus solutions of Laplace’s equa-
tion which are real and imaginary parts of a function f (z) are called
conjugate harmonic functions.
An Important Theorem: Let f (z) be analytic within a region R, with one or

more singularities on the borders (and possibly outside) of R. Then f (z) has
derivatives of all orders that exist at any point zo inside R, so that f (z) can
be expanded in a Taylor series about the point zo . This series converges for all
points within a circle centered at zo that extends to the closest singularity of
f (z).
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)
Solution: We recall that the Tailor series of a function f (z) is given by
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
From the above expression we note that

1
lim !1
z0 ! 2i z0 + 2i
which shows that the function f (z) is singular at z0 = 2i: Therefore,

the radius of convergence is
jz0 j = 2:
Which means the series is convergent for all z inside the circle of radius
R = 2 centered about the origin.
Example 13.4 Polarization of EM waves (Extra Example): Consider an EM

wave propagating along the positive z-direction. We can express the elec-
tric …eld of this wave using complex functions as
~ = (E0x x
E ^ + E0y y^) ei(kz !t)
; (14.40)
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.
(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
E0x = jE0x j ei'x ; E0y = jE0y j ei'y (14.41)
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)
If the phase ' = 0, the electric …eld becomes

~ = jE0x j ei(kz
E !t)
^ + jE0y j ei(kz
x !t)
y^ (14.45)
which leads to
Ey = jE0x j ei(kz !t)

= jE0x j cos (kz !t) + i jE0x j sin (kz !t) ; (14.46)
Ex = jE0y j ei(kz !t)

= jE0y j cos (kz !t) + i jE0y j sin (kz !t) : (14.47)
For ' = =2, we still have the same form for the x-component
Ex = jE0y j ei(kz !t)

= jE0y j cos (kz !t) + i jE0y j sin (kz !t) : (14.48)
but for the y-component we …nd
Ey = jE0x j ei(kz !t)

= jE0x j cos (kz !t) + i jE0x j sin (kz !t) ;
Ey = jE0y j ei(kz !t)

= jE0y j cos (kz !t =2)+i jE0y j sin (kz !t =2) :
(14.49)
Using
cos ( =2) = cos ( ) cos ( =2) sin ( ) sin ( =2) = sin ( ) (14.50)
and
sin ( =2) = sin ( ) cos ( =2) cos ( ) sin ( =2) = cos ( ) (14.51)
we may write the y-component of the electric …eld as
Ey = jE0y j sin (kz !t) i jE0y j cos (kz !t) : (14.52)
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.
(c) For ' = 0 (z = 0 ) using the results above, we may write
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) :
and the real electric …eld vector would be
Re E ~ = Re (Ex ) x^ + Re (Ey ) y^ = jE0x j cos (!t) x

^ + jE0y j cos (!t) y^
(14.54)
Plotting the y vs the x components for di¤erent we …nd
This kind of EM wave (light) is called Linearly (plane) polarized EM

(light).
For ' = =2 (z = 0 ) using the results in part b,
Ex = jE0x j cos (!t) i jE0x j sin (!t) ; (14.55)
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)
and the real electric …eld vector would be

~ = Re (Ex ) x
Re E ^ + Re (Ey ) y^ = jE0x j cos (!t) x
^jE0y j sin (!t) y^
(14.57)
Plotting the y vs the x components for di¤erent we …nd
This kind of EM wave (light) is called Elliptically (left or right ) polarized

EM (light). If jE0x j = jE0y j ; is is called Circularly (left or right ) polarized :
14.4 Contour Integration and Cauchy’s Theo-

rem
Cauchy’s theorem: Consider a function, f (z); that is analytic inside a region,
R; on the complex plane. Within the region, R; if there is a closed curve C
that does not cross itself, and that has a …nite number of sharp corners, then
14.4. CONTOUR INTEGRATION AND CAUCHY’S THEOREM 397
Cauchy’s theorem states that

I
f (z) dz = 0: (14.58)
C
The line integral on this curve is always zero.
Proof: Using
z = x + iy (14.59)
for a complex function
f (z) = u (x; y) + iv (x; y) (14.60)
the closed integral can be expressed as

I I
f (z) dz = [(u(x; y) + iv (x; y)) (dx + idy)]
C C
I I
= [(u(x; y)dx v (x; y) dy)] + i [(v (x; y) dx + u(x; y)dy)]
(14.61)
:
C C
Applying Green’s theorem [that you were introduced to last semester]

I ZZ
@Q @P
[(P (x; y) dx + Q (x; y) dy)] = dxdy: (14.62)
@x @y
C A
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
For analytic function, f (z) = u (x; y) + iv (x; y), the Cauchy-Riemann

condition states that
@u @v @v @u
= ; = : (14.65)
@x @y @x @y
Applying these conditions, one can easily …nds
I ZZ ZZ
@u @u @v @v
f (z) dz = + dxdy + i dxdy = 0:
@y @y @y @y
C A A
(14.66)
The Equation of a Circle on the Complex Plane: Consider a circle on a

complex plane centered about a point, z0 = x0 + iy0 ; shown in Fig. 14.2.
Any point on this circle, z = x + iy; can be described by the equation
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 14.2: A circle centered about z0 on a complex plane.
14.5 Cauchy’s Integral formula

Suppose the complex function, f (z) ; is analytic inside a region enclosed by a
curve C and also on the curve itself (e.g. Fig. 14.3). The value for this function
at a point, z = z0 ; inside the curve C; f (z0 ) can be determined using the line
integral on on the closed curve, C, given by
I
1 f (z)
f (z0 ) = dz: (14.68)
2 i z z0
C
14.5. CAUCHY’S INTEGRAL FORMULA 399
Figure 14.3: A circle centered about z0 : The function, f (z) ; is analytic every-
where inside and on the closed curve, C:
Proof : Consider the function

f (z)
g (z) = (14.69)
z z0
which is analytic everywhere inside the the closed curve C except z = z0 : We
created a region enclosed by a new closed curve, C 0 , that consist of four curves
(C1 ; C2 ; l1 ; l2 ) as shown in the Fig. 14.4 so that the singular point z = z0 can
be excluded.In this region, the function g (z) is analytic. Then according to
Cauchy’s Theorem, we must have
I
g (z) dz = 0; (14.70)
C0
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.
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
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
Note: A line integral in the complex plane is called a contour integral.
Example 13.5 Evaluate the contour integral

Z 2+4i
I= z 2 dz (14.80)
1+i
along the paths indicated.

(a) A straight line joining the points z1 = 1 + i and z2 = 2 + 4i.
(b) Two straight lines: the …rst from the point z1 = 1 + i to zo = 2 + i, and
the second from the point zo to the point z2 = 2 + 4i.
(c) Find the integral
I
I= z 2 dz (14.81)
C
for the closed triangular curve

Solution:
(a) The equation of the line joining the two points can be expressed as
y y1 y2 y1 y 1 4 1
= ) = = 3 ) y = 3x 2: (14.82)
x x1 x2 x1 x 1 2 1
Using
z = x + iy (14.83)
we may write
Z 2+4i Z 2+4i
2
z dz = x2 y 2 + 2ixy (dx + idy)
1+i 1+i
Z 2+4i
= x2 y 2 + 2ixy dx + 2xy + i x2 y2 dy (14.84)
1+i
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
which is the same result as in part a.

(c) Using Cauchy’s theorem

I
f (z) dz = 0 (14.88)
C
for the function
f (z) = z 2 (14.89)
which is analytic everywhere inside and on the triangle, we can easily …nd
I
z 2 dz = 0: (14.90)
C
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
Example 13.6 Evaluate the contour integral

I
ez
I= dz (14.93)
z (z + 1)
for the closed path C given by jz 1j = 3.
1 1 1
= (14.94)
z (z + 1) z z+1
we may write
I I I
ez ez ez
I= dz = dz + dz (14.95)
C z (z + 1) C z Cz+1
For the contour de…ned by the curvejz 1j = 3 is shown in the …gure

below. Noting that the function
f (z) = ez (14.96)
is analytic inside the curve bounded by C and z = 0 and z = 1 are inside
this curve, applying Cauchy’s integral formula we can write
I z I
e ez
dz = 2 if (0) = 2 i; dz = 2 if ( 1) = 2 ie 1 (14.97)
C z Cz+1
so that I
ez 1
I= dz = 2 i 1 + e : (14.98)
C z (z + 1)
Figure 14.5: A circle with radius r = 3 centered at z = 1:
14.6 Laurent’s Theorem

Consider a region R between two circles C1 and C2 centered at the same point,
z = zo, and let f (z) be analytic in R.
Then f (z) can be expanded in a series of the form
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 )
that converges for any value of z within the region R.

Some important terminologies
(a) A regular point: If bn = 0 for all values of n, then f (z) is analytic at z = zo ,
and zo is said to be a regular point of f (z).
(b) The residue: the coe¢ cient b1 in Laurent’s series is called the residue of
f (z).
(c) A pole: If the principal part of the series has terms only up to bn (i.e.
bn 6= 0, but all the bn+1 = 0); then f (z) is said to have a pole of order n
at the point z = zo
(d) A simple pole: If the principal part of the series has only the single term
b1 (i.e. bn+1 > 0 for all n > 1), then f (z) is said to have a simple pole at
z = zo.
14.7. RESIDUES AND THE RESIDUE THEOREM 405
(e) Essential singularity: If there are an in…nite number of b0 s di¤erent from

zero, f (z) has an essential singularity at z = z0
14.7 Residues and the Residue Theorem

The Residue Theorem: Let zk be an isolated singular point of f (z) inside a
closed curve de…ned by C; the residue theorem states that
I
f (z) dz = 2 iR (zk ) ;
C
where the integral is in a counterclockwise direction and R (zk ) is the residue of

the function f (z) at z = zk :
Proof: If we de…ne a small circle of radius r = , as shown in the …gure below
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 )
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
which may want to put in the form
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
for all n > 2; we …nd

I
f (z) dz = 2 ib1 = 2 iR (zk ) ; (14.109)
C
where
R (zk ) = b1 (14.110)
is the residue of the function f (z) at z = zk

Note: If there are more than one residue, the residue theorem states that
I n
X
f (z) dz = 2 i R (zk ) : (14.111)
k=1
14.8 Methods of …nding residues

Method 1 : Expand f (z) about the point zo and simply read the value of the
residue o¤ of the series. Recall Taylor Series expansions for some basic function
(PHYS 3150): Any function, f (x) that is di¤ erentiable for all values of x in the
speci…ed domain, can be expressed in Taylor or Maclaurin series. That means
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
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
Solution: We recall that the series expansion of the function

f (z) = cos (z) (14.116)
is
z2 z4 z6
cos (z) = 1 + ::: (14.117)
2! 4! 6!
so that
1 1 1 1 1 1 1
cos =1 + ::: (14.118)
z z 2 2! z 4 4! z 6 6!
and
1 11 1 1 1 1
f (z) = z cos =z + : (14.119)
z z 2! z 3 4! z 5 6!
There follows that for zk = 0
1
R (zk ) = b1 = : (14.120)
2
Example 13.8 Find R(1) for the complex function

ez
f (z) = 2: (14.121)
(z 1)
Solution: The function has isolated singular point at zk = 1 and we need to

expand the function
g (z) = ez (14.122)
about this point. The series expansion about this point is is given by
1 2 1 3 1 4
g (z) = ez = e 1 + (z 1) + (z 1) + (z 1) :::: (14.123)
2! 3! 4!
then
ez
f (z) = 2: (14.124)
(z 1)
becomes
e 1 1 2 1 3 1 4
f (z) = 2 1+(z 1) + (z 1) + (z 1) + (z 1) ::::
(z 1) 1! 2! 3! 4!
e e 1 1 1 2
= 2 + (z + e + (z 1) + (z 1) ::::
(z 1) 1) 2! 3! 4!
There follows that the residue for zk = 1;
R (zk ) = b1 = e: (14.125)
Example 13.9 Find R( ) for the complex function

sin z
f (z) = (14.126)
z
Solution: The function has an isolated singular point at

zk = (14.127)
and we need to expand the function sin (z) in series about this point
3 5 7
(z ) (z ) (z )
sin (z) = (z )+ + ::: (14.128)
3! 5! 7!
so that we can write
sin z
f (z) = (14.129)
z
as
2 4 6
(z ) (z ) (z )
f (z) = 1+ + ::: (14.130)
3! 5! 7!
There follows that for zk =
R (zk ) = b1 = 0: (14.131)
Method 2: If f (z) has a simple pole (a pole of order 1) at zo , then multiply

f (z) by (z zo ) and evaluate the result at z = zo (or take the limit as z
approaches zo ):
R (zo ) = lim [(z z0 ) f (z)] (14.132)
z!zo
Example 13.10 Find R( 1) for the complex function

z
f (z) = : (14.133)
(z + 1) (z + 2)
Solution: We note that f (z) has a simple pole (a order of 1) at zo = 1. Then

we can use the second method discussed above to …nd the residue zo = 1
z
R ( 1) = lim [(z + 1) f (z)] = lim (z + 1)
z! 1 z! 1 (z + 1) (z + 2)
z 1
) R ( 1) = lim = = 1: (14.134)
z! 1 (z + 2) ( 1 + 2)
Example 13.11 Find R( ) for the complex function
f (z) = cot (z) (14.135)
Solution: The cotangent function is given by

cos (z)
f (z) = cot (z) = : (14.136)
sin (z)
Let’s assume that this function has a simple pole at z0 = which is true
as we will see shortly. Then using the second method the residue at z0 =
can be expressed as
cos (z)
R ( ) = lim [(z ) f (z)] ) R ( ) = lim (z ) (14.137)
z! z! sin (z)
As z ! , the limit becomes 0=0 and we can apply L’Hospital rule
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)
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)
Example 13.13 Find R( 2) for the complex function

e2z
f (z) = 3 (14.144)
z (z + 2)
Solution: The function has a pole at zo = 2 of order n = 3. Using the third
method for m = n = 3, we have
( )
1 d3 1 3 e2z 1 d2 e2z
R ( 2) = (z + 2) 3 =
(3 1)! dz 3 1 z (z + 2) 2! dz 2 z z= 2
z= 2
1 d 2e2z e2z 1 4e2z 2e2z
2e2z 2e2z
= = + 3
2! dz z z2 z= 2 2! z z2 z 2 z z= 2
2e2z 2e2z e2z 1 1 4 13 4
= + = 1 e ) R ( 2) = e :
z z2 z3 z= 2 2 8 8
(14.145)
14.9 Applications of the Residue Theorem

The Residue Theorem: we recall that if fz1 ; z2 ; ::zk :::zn g are singular points of
f (z) inside a closed curve de…ned by C; the residue theorem states that
I n
X
f (z) dz = 2 i R (zk ) ; (14.146)
C k=1
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
Solution: Consider the a semicircle of radius R on the upper half complex

plane and the complex function
1
f (z) = : (14.148)
z4 +1
Now let’s integrate the complex function over the semicircular contour in
the upper half plane in the counterclockwise direction as shown in the
…gure belowwhich we write as
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
For the curved part, we have
z = R exp[i ] ) dz = iR exp[i ]d ; (14.150)
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
since x4 + 1 is an even function we can write

Z 1 Z 1 Z 1 I
dx dx dx 1 dz
4+1
=2 4+1
) 4+1
= 4+1
: (14.153)
1 x 0 x 0 x 2 z
The integral of the complex function
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 n = 0; 1; 2; 3 [Phys 3150 Roots of a Complex number ], the poles for

the function f (z) are found to be
p
i4 2
z1 = e = cos + i sin ) z1 = (1 + i) ;
p 4 4 2
3 2
z2 = ei 4 = ( 1 + i) ;
2p
5 2
z3 = ei 4 = (1 + i) ;
p 2
7 2
z4 = ei 4 = (1 i) (14.157)
2
Remember, we are integrating on the upper half of the complex plane.

Hence, we are interested in the poles with a phase within 0 :
These poles are
p p
i4 2 i 34 2
z1 = e = (1 + i) ; z2 = e = ( 1 + i) : (14.158)
2 2
We now determine the residue of these poles. We recall that the residue
of a pole z = zk of order n is given by
1 dm 1
m
R (zk ) = f(z zk ) f (z)g (14.159)
(m 1)! dz m 1
z=zk
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
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)
Now using the residue theorem

I n
X
f (z) dz = 2 i R (zk ) (14.167)
k=1
we can see that

I
dz 1 1 i 1+1 i
4
=2 i p + p = i p
z +1 2 2 (1 i) 2 2 (i + 1) 2 2
I
dz 2i
) = i p =p : (14.168)
z4 + 1 2 2 2
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:
Example 13.15 Evaluate the real integral
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
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
where z = exp (i ) ; we can write
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
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
The …rst pole z0 = z1 = 0 is not a simple pole. It has order n = 3. Thus

we use the third method to …nd the residue for this pole
1 dm 1
m
R (zo ) = f(z z0 ) f (z)g ; (14.181)
(m 1)! dz m 1
z=zo
where m > n: For m = 3 we may write
( )
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)
"
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
Therefore the integral gives

I n
X 65 i21
f (z) dz = 2 i R (zk ) = 2 i i + =
48 16 12
k=1
Z2 I
cos (3 )
) d = f (z) dz = (14.184)
5 4 cos ( ) 12
0
Mathematica result:
14.10 Standard Methods of Integration using Con-

tour Integrals
Method I: A real integral over the entire x-axis
Method 2 : A real integral involving sines and cosines from 0 to 2

14.11. THE KRAMERS-KRONIG RELATIONS 419
14.11 The Kramers-Kronig Relations

The Residue Theorem: we recall that
I
f (z) dz = 2 iR (zk ) (14.185)
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
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.
14.12 The revised Residue Theorem

It states that
I " n n
#
X 1X 0
f (z) dz = 2 i R (zk ) + R (zk ) : (14.190)
2
k=1 k=1
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
For the modi…ed contour, we can write

I Z z10 Z z10 + Z z20 Z z20 +
f (z) dz = f (z) dz + f (z) dz + f (z) dz + f (z) dz
R z10 z10 + z20
Z R
+ f (z) dz (14.192)
z20 +
14.12. THE REVISED RESIDUE THEOREM 421
Example 13.15 Evaluate the integrals

Z 1 Z 1
sin (x) cos (x)
I1 = dx; I2 = dx (14.193)
0 x 1 x
Solution: Consider the function

eiz
f (z) = (14.194)
z
For a semi-circular closed curve of radius, R; with diameter on the x-axis,
this function has pole at zk0 = 0 which is exactly on the curve. Therefore,
using the applying the revised residue theorem, we may write

I iz " n n
#
e X 1X 0
dz = 2 i R (zk ) + R (zk ) (14.195)
z 2
k=1 k=1
I
eiz
dz = iR (zk0 = 0) (14.196)
z
This pole is …rst order pole and we can …nd the residue using
R (zk0 ) = lim [(z zk0 ) f (z)] (14.197)
z!0
which gives
eiz
R (0) = lim z =1 (14.198)
z!0 z
Therefore the integral becomes
I
eiz
dz = i (14.199)
z
Now we note that in the limit R ! 1; we have

I Z1 Z Z1 Z1
eiz eix eiz cos (x) sin (x)
dz = dx + = dx + i dx
z x C z x x
1 1 1
Z
eiz
+ (14.200)
C 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
Dispersion (The Frequency Dependence of permittivity) The electrons in a

nonconducting medium are bound to speci…c molecules. These electrons os-
cillates about the equilibrium position with a small amplitude. The electrons
experiencing this kind of motion can be modeled as a harmonic oscillator. Then
the electron experiences a spring force Fs given by
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
^
it will experience a driving force
Fd = qE = qE0 cos (!t)
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)) :
Substituting a particular solution of the form
x
~p (t) = x
~0 exp ( i!t)
into the di¤erential equation, we …nd
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 !
and the particular solution becomes
qE0 =m
x
~p (t) = exp ( i!t)
! 20 !2 i !
Therefore the complex displacement of the electron is given by
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 !
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 !
P~ = 0 eE
~
where
N q2 X fj
e = 2 :
0 m j=1 ! j !2 i !
Recalling that the complex permittivity
~= 0 (1 + e)
and the complex dielectric constant can be expressed as
~ 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 :
where the wave number k~ is complex and is given by

! 1 p
=p ) k~ = ~ 0!
~
k ~ 0
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 :
The wave velocity is given by

!
v=
kRe
and the refractive index of the medium

c ckRe
n= =
v !
The complex wave number
r
p ~ p !
k~ = ~ 0! = 0 0! = r
0 c
2 31=2
! N q2 X fj
) k~ = 41 + 2
5
c 0 m j=1 ! j !2 i j !
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!
For just one particular molecule (j th molecule) we may write that
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!
Introducing the transformation de…ned by the dimensionless variables

! j
= ; =
!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).
Refractive index decreases for ! 1 < ! < ! 2 , 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 less than one for ! > ! 2 which means the wave speed wave
exceeds c. This is not an alarm since energy does not travel at the wave velocity
but rather at a group velocity. Moreover we considered only one molecule.
Far away from resonance ( ! >> ! j ): For the case where we are far away
2
from resonance we can ignore the damping term and we can write ! 2j ! 2 +
2 2
j! ' ! 2j !2 : Then index of refraction can be expressed as
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
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
If we express the frequency in terms of the wave length in a vacuum
! = 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
f (!) = u (!) + iv (!)

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
where P indicates the Cauchy principal value which is given by

Z 1
u (!)
P d!
1 ! !0
Z !0 Z 1
u (!) u (!)
= lim d! + lim d!
!0 1 ! !0 !0 ! + !
0
!0
Dispersion: The Kramers-Kronig Relations has a wide application in a dis-

persive medium (frequency dependent) for Electromagnetic waves. In such
medium the propagation vector k is a complex quantity which leads to a complex
refractive index
4 4
n2 (!) = 1 + i ) n2 (!) 1=i
! !
According to the Kramers-Kronig Relations, we can write
Z 1
2 2 ! Im n2 (!) 1
Re n (!) 1 = P d!
0 ! 2 ! 20
Z 1
2 2 ! 0 Re n2 (!) 1
Im n (!) 1 = P d!
0 ! 2 ! 20
where ! 0 is a general real angular frequency.
Chapter 15
Laplace transform
Integral Transforms: Frequently in physics we encounter pairs of functions re-

lated by an integral of the form
Z b
F (p) = K (p; t) f (t) dt (15.1)
a
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)
for su¢ ciently large t, t > t0 :

2
f (t) = et : (15.5)
433
434 CHAPTER 15. LAPLACE TRANSFORM
(iii) The Laplace transform fails to exist when the function has a strong
singularity as t ! 0
f (t) = tn for n 1: (15.6)
(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
L [af (t) + dg (t)] = aL [f (t)] + dL [g (t)] ; (15.7)
where a and d are constants.
Example 14.1 Show that the Laplace Transform is linear
Solution: Consider the function
h (t) = af (t) + dg (t) (15.8)
so that the Laplace transform of this function can be expressed as

Z b
pt
H (p) = L [h (t)] = lim e [af (t) + dg (t)] dt
b!1 0
Z b Z a Z a
pt pt pt
= lim e af (t) dt + lim e dg (t) dt = a lim e f (t) dt
b!1 0 a!1 0 a!1 0
Z a
+d lim e pt g (t) dt = aL [f (t)] + dL [g (t)]
a!1 0
) H (p) = L [af (t) + bg (t)] = aL [f (t)] + dL [g (t)] (15.9)
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
(a) Recalling that Z 1

pt
F (p) = L [f (t)] = e f (t) dt: (15.14)
0
we …nd Z 1
pt 1
L [1] = e dt = : (15.15)
0 p
for p > 0:
(b) Z 1
!t (p !)t 1
F (p) = L e = e dt: = (15.16)
0 p !
for p > !
(c) Noting that
e!t + e !t
e!t e !t
cosh (!t) = ; sinh (!t) = (15.17)
2 2
we may write
e!t + e !t
F (p) = L [cosh (!t)] = L (15.18)
2
Recalling that Laplace transform is linear, one can write
1
L [cosh (!t)] =L e!t + L e !t : (15.19)
2
and applying the result in the previous example, we …nd
1 1 1 p
L [cosh (!t)] = + = : (15.20)
2 p ! p+! p2 !2
Similarly
e!t e !t
1
F (p) = L [sinh (!t)] = L =L e!t L e !t
2 2
1 1 1 !
) L [sinh (!t)] = = 2 (15.21)
2 p ! p+! p !2
(d) Recalling that

cos (!t) = cosh (i!t) ; sin (!t) = i sinh (i!t) (15.22)
and applying the result above we can easily write that
p p
L [cos (!t)] = L [cosh (i!t)] = 2 = p2 + ! 2
p2 (i!)
i (i!)
L [sin (!t)] = L [ i sinh (i!t)] = iL [sinh (i!t)] = 2
p2 (i!)
!
L [sin (!t)] = 2 (15.23)
p + !2
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)
Solution: The Laplace transform

Z 1
pt
F (p) = L [f (t)] = e f (t) dt: (15.26)
0
for the unit step function can be rewritten as

Z a Z 1 Z 1 pt 1
pt pt pt e
L [f (t)] = e f (t) dt + e f (t) dt = e dt =
0 a a p a
pa
e
) L [f (t)] = : (15.27)
p
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
and show that the general relation can be written as

h i dn f (t)
L f (n) (t) = L = pn L [f (t)] pn 1
f (0) pn 2 0
f (0)
dtn
00
pn 3
f (0) ::: f (n 1)
(0) : (15.31)
Solution:
(a) Recalling that

Z 1
pt
L [f (t)] = e f (t) dt: (15.32)
0
437
for f 0 (t) ; one can write

Z 1
pt df (t)
L [f 0 (t)] = e dt (15.33)
0 dt
so that applying integration by parts we may write

Z 1 Z 1
1
L [f 0 (t)] = e pt f (t) 0 + p e pt f (t) dt = f (0) + p e pt
f (t) dt
0 0
) L [f 0 (t)] = pL [f (t)] f (0)
(b) Similarly for the second derivative

Z 1
d2 f (t)
L [f 00 (t)] = e pt
dt (15.34)
0 dt2
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)
(c) For the third derivative

Z 1 3 2 1 Z 1 2
000 pt d f (t) pt d f (t) pt d f (t)
L [f (t)] = e dt = e +p e dt
0 dt3 dt2 0 0 dt2
) L [f 000 (t)] = pL [f 00 (t)] f 00 (0) (15.38)
and using the result in part (b), one …nds
L [f 000 (t)] = p3 L [f (t)] p2 f (0) pf 0 (0) f 00 (0)
From the above result we may write the general relation

h i dn f (t)
L f (n) (t) = L = pn L [f (t)] pn 1
f (0) pn 2 0
f (0)
dtn
00
pn 3
f (0) ::: f (n 1)
(0) : (15.39)
15.1 Inverse Laplace Transform

The inverse of a Laplace transform is represented by
1
L [F (p)] = f (t) : (15.40)
In order to perform inverse Laplace transform it is important to know the
Laplace transform of some basic functions that we saw in the examples earlier.
f (t) 1 e!t sin (!t) cos (!t) sinh (!t) cosh (!t) U (t a)
1 1 ! p ! p e ap
L [f (t)] p p ! p2 +! 2 p2 +! 2 p2 ! 2 p2 ! 2 p
Example 14.4 Evaluate the following inverse Laplace transforms

(a)
1 5
L (15.41)
p+2
(b)
1 4p 3
L (15.42)
p2 + 4
Solution:
(a) Noting that
1 5 1 1
L = 5L (15.43)
p+2 p ( 2)
and recalling
1
L e!t = (15.44)
p !
we can write
1 5 2t
L = 5e : (15.45)
p+2
(b) Here also we want to write
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
Example 14.5 Evaluate the integral

Z 1
sin (tx)
f (t) = dx (15.49)
0 x
applying Laplace and inverse Laplace transform
Solution: We …rst apply Laplace transform
Z 1 Z 1
sin (tx)
L [f (t)] = e pt dx dt (15.50)
0 0 x
which can also be expressed, by interchanging the order of the integration,
as
Z 1 Z 1
1
L [f (t)] = sin (tx) e pt dt dx
0 x 0
Z 1 Z 1 itx Z 1 Z 1
1 e e itx pt 1 e (p ix)t e (p+itx)
= e dt dx = dt dx
0 x 0 2i 0 x 0 2i
Z 1 1
1 e (p ix)t e (p+itx)
= + dx
0 2ix (p ix) p + ix 0
Z 1 Z 1
1 1 1 1 p + ix p + ix
= dx = dx
0 2ix p ix p + ix 0 2ix p 2 + x2
Z 1
dx
) L [f (t)] = = (15.51)
0 p + x2
2 2p
where we used the result we determined in using the Residue theorem.
Noting that
Z 1 Z 1
sin (tx) 1 1 sin (tx)
f (t) = dx = L [L [f (t)]] = L L dx
0 x 0 x
(15.52)
for the inverse Laplace transform, we have
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)
one can write

Z 1
sin (tx)
f (t) = dx = for t < 0: (15.56)
0 x 2
For t = 0, obviously, we …nd
Z 1
sin (tx)
f (t) = dx = 0: (15.57)
0 x
Therefore we can write
8
Z 1 < t>0
sin (tx) 2
dx = [2U (t) 1] = 0 t=0 (15.58)
0 x 2 :
2 t<0
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
15.2 Applications of Laplace transforms

In this section we will see applications of Laplace Transforms by solving several
di¤erential equations. To this end we recall the following "Table of Laplace
Transforms" we determined in the previous class.
1 1 ! p ! p e ap
L [f (t)] p p ! p2 +! 2 p2 +! 2 p2 ! 2 p2 ! 2 p
and
h i dn f (t) 00
L f (n) (t) = L = pn L [f (t)] pn 1 f (0) pn 2 f (0)
dtn
pn 3 000
f (0) ::: f (n 1)
(0) : (15.59)
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
Solution: We note that using Laplace transform we have

d2 y
L + L [y (x)] = L [1] (15.61)
dx2
so that with the help of the results given earlier, one can write
1 d2 y
L [1] = ;L = p2 L [y (x)] py (0) y 0 (0) (15.62)
p dx2
15.2. APPLICATIONS OF LAPLACE TRANSFORMS 441
Substituting the given initial conditions
y (0) = 1; y 0 (0) = 0 (15.63)
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
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
Simplifying this expression we …nd
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)
There follows that
A + B = 2; 2A + 3B + C = 7; A + 2B + 2C = 7 (15.73)
Solving these equations using Mathematica
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:
There follows that

" #
1 1 1 1 1 2
y (t) = L +L +L 2 (15.75)
p+2 p+1 (p + 1)
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)
one …nds for the solution to the di¤erential equation

" #
1 1 1 1 1 2
y (t) = L +L +L 2 (15.78)
p+2 p+1 (p + 1)
becomes
2t t t 2t
y (t) = e +e + 2te =e 1 + et (1 + 2t) (15.79)
Using Mathematica:
By directly solving the di¤ erential equation 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)
Solution: Applying Kircho¤’s voltage rule, we can write
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)
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:
Using these results, we can write

5 1 5 p 5 5
L [I (t)] = + I (0) + (15.86)
2 p+5 2 p2 + 5 2 2 p2 + 5 2
so that the inverse Laplace transform gives
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
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:
Note that we can write

5 5t 5 5t 5t
+ I (0) e = + I (0) e = c[1]e (15.90)
2 2
Example 14.8 Damped Harmonic oscillator : Consider a mass m oscillating

under the in‡uence of a spring, spring constant k and damped by a friction
force which is proportional to the velocity (Ff = bv). Assuming that
the particle starts from rest at x (0) = x0 ; x0 (0) = 0. Find the equation of
motion for the mass m and determine the position of the mass as function
of time using Laplace transformation.
Solution:
b
x (t) = x0 e( b=2m)t
cos (! 1 t) + sin (! 1 t) (15.91)
2m! 1
where
k b2
! 21 = (15.92)
m 4m2
Chapter 16
Fourier series and transform
In this chapter we will cover Fourier Series and Transform....
16.1 Introduction to Fourier Series

Periodic Functions : A function f (x) is said to be a periodic function when the
function repeats itself after a period T; which means
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
Periodic Motion: Periodic motion is when the motion of an object contin-

ually repeats itself. This can be repeatedly moving back and forth or it could
be moving in a circular orbit or rotation. Since the Law of Inertia states that
an object moves in a straight line unless acted upon by a force, periodic motion
requires some sort of force to create this special type of motion. Characteristics
of periodic motion are the velocity of the object, the period of motion and the
amplitude of the motion. Example of Periodic motion include circular motion (
e.g. when you swing an object around you that is held on a rope or string, mo-
tion of the planets around the sun) and back-and-forth motion (e.g. a bouncing
ball, pendulum and spring).
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
Introducing the variable

dx
x = !t ') = dt (16.11)
!
and
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 '
using the relation

sin (2x) = 2 sin (x) cos (x) (16.15)
we get
2
Im 2 ' sin (2 ') cos (2 ')
I 2 (t) =
2 2 2
2
Im ' sin ( ') cos ( ')
: (16.16)
2 2 2
Noting that
sin ( ') = sin (') ; cos ( ') = cos (')

sin (2 ') = sin (' 2 ) ; cos (2 ') = cos (' 2 ) (16.17)
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
16.2 The Fourier Series Expansion of a Periodic

Function
Any function of periodicity 2 can be expanded in a Fourier series of the form
X1
1
f (x) = a0 + [an cos (nx) + bn sin (nx)] (16.22)
2 n=1
where the coe¢ cients are given by

Z
1
an = f (x) cos (nx) dx; n = 0; 1; 2; 3:::; (16.23)
Z
1
bn = f (x) sin (nx) dx; n = 1; 2; 3::: (16.24)
Using Euler’s formula

einx + e inx
einx e inx
cos (nx) = ; sin (nx) = (16.25)
2 2i
we may
X1
1 einx + e inx
einx e inx
f (x) = a0 + an + bn
2 n=1
2 2i
X1
1 an ibn an + ibn
= a0 + einx + e inx
2 n=1
2 2
16.2. THE FOURIER SERIES EXPANSION OF A PERIODIC FUNCTION453
Replacing n by n in the second series and noting that
a0 = a0 ei0 x
(16.26)
one can write
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
More concisely the expression for cn can be determined from
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
1 for 0 < x <

f (x) = (16.30)
1 for <x<0
which is repeated inde…nitely with a period of 2 .
Solution: Using
Z
1
an = f (x) cos (nx) dx; n = 0; 1; 2; 3::: (16.31)
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)
Therefore, the Fourier series expansion of
1 for 0 < x <

f (x) = (16.34)
1 for <x<0
is given by
1
4 X sin [(2n + 1) x]
f (x) = : (16.35)
n=0
2n + 1
16.3. DIRICHLET CONDITIONS 455
16.3 Dirichlet Conditions

The Dirichlet conditions are su¢ cient conditions for a real-valued, periodic func-
tion f (x) to be equal to the sum of its Fourier series at each point where f is
continuous. Moreover, the behavior of the Fourier series at points of discon-
tinuity is determined as well. These conditions are named after Johann Peter
Gustav Lejeune Dirichlet. The conditions are:
(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)
If a function has a periodicity of 2 , has a …nite number of maxima, minima,

and discontinuities within a single period, and if it is bounded within that
period (that is, it doesn’t become in…nite), then 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.
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
f (x+ ) = lim f (y) ; f (x ) = lim f (y) (16.37)

y!x+ y!x
denotes the right/left limits of f .
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
Solution: We note that
(a) f (x) have a …nite number of extrema (maxima, minima) in any given in-
terval, <x< :
(b) f (x) has one discontinuity at x = 0 in the interval <x< :
(c) f (x) is absolutely integrable over <x< .
(d) f (x) is bounded within <x< :
We must be able to show that

1
X 1
cn einx = (f (x+ ) + f (x )) : (16.39)
n= 1
2
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
16.4 Fourier series with spatial and temporal ar-

guments
For x unitless (or in radians) and periodic on 2 , f (x + 2 ) = f (x),
X1
1
f (x) = a0 + [an cos (nx) + bn sin (nx)] (16.45)
2 n=1
where the coe¢ cients are determined using

Z
1
an = f (x) cos (nx) dx; n = 0; 1; 2; 3:::
Z
1
bn = f (x) sin (nx) dx; n = 1; 2; 3::: (16.46)
Spatial Arguments: If x represents a position in space, and f (x + ) = f (x),

then, for k = 2 = ,
X1
1
f (x) = a0 + [an cos (nkx) + bn sin (nkx)] (16.47)
2 n=1
where the coe¢ cients

Z=2
2
an = f (x) cos (nkx) dx; n = 0; 1; 2; 3:::
=2
Z=2
2
bn = f (x) sin (nkx) dx; n = 1; 2; 3::: (16.48)
=2
Temporal Arguments: If t represents time, and f (t + T ) = f (t), then, for

! = 2 =T ,
X1
1
f (t) = a0 + [an cos (n!t) + bn sin (n!t)] (16.49)
2 n=1
where the coe¢ cients

T =2
Z
2
an = f (t) cos (n!t) dt; n = 0; 1; 2; 3:::
T
T =2
T =2
Z
2
bn = f (t) sin (n!t) dt; n = 1; 2; 3::: (16.50)
T
T =2
Example 15.6 A voltage signal in the form of a half-recti…ed sine-wave is fed

into a linear circuit whose response to a simple sinusoidal signal is known.
In order to determine the response of the circuit to the input signal, we
wish to …nd the Fourier series expansion of that signal. Assume that the
input signal is given by
T
0 2<t<0
V (t) = (16.51)
Vm sin (!t) 0 < t < T2
Hint: You may …nd the following integrals useful.
T =2
Z
cos [(1 m) !t] cos [(1 + m) !t]
cos (m!t) sin (!t) dt =
2 (1 m) ! 2 (1 + m) !
T =2
T =2
Z
sin [(1 m) !t] sin [(1 + m) !t]
sin (m!t) sin (!t) dt = (16.52)
2 (1 m) ! 2 (1 + m) !
T =2
Solution: Using Fourier expansion coe¢ cients for temporal arguments

T =2
Z
2
an = f (t) cos (n!t) dt; n = 0; 1; 2; 3::: (16.53)
T
T =2
T =2
Z
2
bn = f (t) sin (n!t) dt; n = 1; 2; 3::: (16.54)
T
T =2
for the given function one can write

T =2
Z T =2
Z
2 Vm 2 Vm
an = cos (n!t) sin (!t) dt; bn = sin (n!t) sin (!t) dt:
T T
0 0
(16.55)
Applying the trigonometric relation
sin ( ) = sin cos cos sin
cos ( ) = cos cos sin sin (16.56)
16.4. FOURIER SERIES WITH SPATIAL AND TEMPORAL ARGUMENTS459
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
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
16.5 The Fourier Transform

The Fourier transform of a periodic or non periodic function, f (t) ; that satis…es
all the Dirchlet conditions does exist and is given by
Z 1
1
F [f (t)] = p f (t) e i!t dt = F (!) : (16.67)
2 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
The resulting Fourier series can be written as

ZT 1 ZT
1 0 1 X
0 n n 0
f (t) = f (t ) dt + cos t f (t0 ) cos t dt0
T T n=1 T T
T T
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
We now let the parameter T approach in…nity, transforming the interval [ T; T ]

into the in…nite interval ( 1; 1) : In this limit we have
! ! d!; ! n ! ! (16.74)
and if the function f satis…es the Dirchlet conditions, we can replace the sum-
mation by integration
Z 1 Z1
1
f (t) = d! f (t0 ) cos [! (t t0 )] dt0 : (16.75)
0
1
Noting that cos [! (t t0 )] is an even function and sin (! (t t0 )) is an odd

function of !; we have
Z 1 Z1 Z 1 Z1
0 0 1
0
d! f (t ) cos [! (t t )] dt = d! f (t0 ) cos [! (t t0 )] dt0
0 2 1
1 1
(16.76)
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
which is the Inverse Fourier Transform.

Note: If f (t) has a discontinuity, then the inverse Fourier transform will
return a value for f (t) that is at the midpoint of the discontinuity at the x-value
at which the discontinuity occurs.
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
Note: The function

sin (x)
f (t) = (16.84)
x
is known as the sinc function commonly expressed as sinc (x) :
Example 15.8 Find the Fourier Transform of a Gaussian function

a2 x2
f (x) = e (16.85)
Solution: For the given function we can write

Z 1
1 2 2
F [f (x)] = p e (a x +i!x) dx (16.86)
2 1
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
To evaluate the integral

Z 1 2
a2 (x+ 2a
i!
2)
I= e dx (16.89)
1
consider the integral

I
a2 z 2
e dz (16.90)
i! i!
for the rectangular closed loop with vertices T; T; T + 2a 2; T + 2a 2 . For
this integral applying the residue theorem we can write

!
I ZT Z2a2
a2 z 2 a2 x2 a2 (T +iy)2
e dz = e dx + i e dy
T 0
ZT 2
Z0
a2 (x+ 2a
i!
2) a2 ( T +iy)2
+ e dx + i e dy
!
T
2a2
!
ZT Z2a2 ZT 2
Z0
2
a x 2
a2 (T +iy)2 a2 (x+ 2a
i!
2) a2 (T iy)2
= e dx + i e dy + e dx + i e dy
0 !
T T
2a2
(16.91)
Noting that in the limit T ! 1;

!
Z2a2 Z0
a2 (T +iy)2 a2 (T iy)2
e dy = e dy = 0 (16.92)
0 !
2a2
and there is no poles inside the curve, we may write
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
so that applying the result

0 12
Z1 Z1 Z1 Z1 Z1 Z 2
2 2 2
u2 v2 r2
@ e y
dy A = e du e dv = e (u +v ) dudv = e rdrd
0
1 1 1 1 0
Z1 Z1
r2 r2
1
y2
p
= 2 e rdr = e = ) e dy = (16.95)
0
0 1
we …nd
Z1 2
p
a2 (x+ 2a
i!
2)
e dx = : (16.96)
a
1
Therefore the Fourier integral transform

!2 Z 1
e 4a2 a2 x2
F [f (x)] = p e dx (16.97)
2 1
becomes
!2
e 4a2
F [f (x)] = p = F (!) : (16.98)
a 2
16.6 The Dirac Delta Function and the Fourier

transform
We recall the inverse Fourier transform
Z1
1
f (t) = p F (!) ei!t d! (16.99)
2
1
so that substituting
Z 1
1 i!t0
F (!) = F [f (t)] = p f (t0 ) e dt0 (16.100)
2 1
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
16.7 Applications of the Fourier Transform

The Time-Dependent Schrödinger Equation
~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
Then the total energy for a free particle
p2
E= (16.115)
2m
can be expressed as
k 2 ~2
E= : (16.116)
2m
Using this result one …nds
d2 (x) 2mE d2 (x)

= (x) = k 2 (x) ) + k 2 (x) = 0 (16.117)
dx2 ~2 dx2
the solution of which is given by
(x) = A (k) eikx : (16.118)
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
Substituting the integral expression for (x) ; we …nd

2 32 1 3
Z1 Z1 Z
0
4 A (k 0 ) e ik x dk 0 5 4 A (k) eikx dk 5 dx = 1
1 1 1
0 1
Z1 Z1 Z1
1 ix(k k 0
) dxA dk 0 dk = 1 :
) A (k) A (k 0 ) @ e (16.121)
2 2
1 1 1
Applying the de…nition of the Dirac Delta function

Z1
0 1 i! (t t0 )
(t t)= e d!; (16.122)
2
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
and using the relation
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
Example 15.9 For a particle described by a Gaussian Wavepacket

1=4
x2 2 x2
(x) = Ae = e (16.131)
Some Useful Integrals

Z1 r Z1 p
ay 2 y2
e dy = ; y2 e dy = (16.132)
a 2
1 1
(a) Show that the momentum eigen function is given by
A k2 =4
(k) = p e (16.133)
2
(b) Show that the expectation values
hki = 0; k 2 = : (16.134)
(c) Find the Expectation values hxi and x2 :
(d) Show that the Heisenberg Uncertainty Principle
~
x px = (16.135)
2
is satis…ed.
Solution:
(a) We recall
Z1
1
(x) = p (k) eikx dk (16.136)
2
1
which is the inverse Fourier transform of the momentum eigen function

(k) ; we can write
Z1
1 ikx
(k) = p (x) e dx: (16.137)
2
1
Then substituting
1=4
x2 2 x2
(x) = Ae = e (16.138)
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
Using the result in the previous example or Mathematica
we can easily show that

A k2
(k) = p e 4 : (16.142)
2
(b) In the momentum space we can express the expectation values
Z1 Z1
2
hki = (k) k (k) dk; k = (k) k 2 (k) dk (16.143)
1 1
and using the momentum eigen function, we may write

2 Z1 2 Z1
A k2
2 A k2
hki = p ke 2 dk = 0; k = p k2 e 2 dk:
2 2
1 1
(16.144)
Using the transformation of variable
k p
x= p ) dk = 2 dx (16.145)
2
we have
2 Z1 p 2 Z1
2 A 2 x2 A 3=2 x2
k = p 2 x e 2 dx = p (2 ) x2 e dx:
2 2
1 1
(16.146)
so that using the relation

Z1 p
y2
y2 e dy = (16.147)
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
Using the wave function

1=4
x2 2 x2
(x) = Ae = e (16.152)
the expectation values become

Z1 Z1
2 2 x2 2 2 2 x2
hxi = A xe dx = 0; x =A x2 e dx: (16.153)
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
(d) The Uncertainty in position and momentum

q q
2 2
x = x = hx2 i hxi ; px = ~ k = ~ k = ~ hk 2 i hki (16.158)
would then be
1 p
x= x = p ; px = ~ k = ~ k =~ : (16.159)
2
Then the uncertainty principle
1 p ~
x px = p ~ = (16.160)
2 2
is satis…ed. This means the Gaussian wave packet represent the minimum
uncertainty state.
16.8 Fourier Transforms and Convolution

Convolution: For two functions f (x) and g (x) with corresponding Fourier
Transforms, F (!) and G (!) ; respectively, the convolution of these two func-
tions over the interval ( 1; 1), denoted by f g; is de…ned as
Z 1 Z 1
1 1
f g= p g (y) f (x y) dy = p f (y) g (x y) dy = g f:
2 1 2 1
(16.161)
This form of integrals appears in probability theory in the determination of the
probability density of two random, independent variables. One good example
is the solution of the Poisson’s equation
Z
1 1
V (~r) = (~r0 ) d3 r 0 : (16.162)
4 0 j~r ~r0 j
This may be interpreted as convolution of a charge distribution (~r0 ) (as g (y))and
1
a weighting function (4 0 j~r ~r0 j) (as f (x y)). The operation of convolu-
tion sometimes referred to as the Faltung, the German term for "folding"
The Convolution theorem: the convolution theorem states that the convolu-
tion of two or more functions is the same as the Fourier inverse transforms of a
product of Fourier Transforms. That means
1
f1 f2 f3 :::fn = F [F1 (!) F3 (!) F3 (!) ::Fn (!)] (16.163)
Proof: We shall proof this theorem by considering the case of two functions
Z 1
1
f1 f2 = p f2 (y) f1 (x y) dy (16.164)
2 1
Let Fourier Transform of the function f1 (x y) is F1 (!) ; then may write

Z 1
1
f1 (x y) = F 1 [F1 (!)] = p F1 (!) ei!(x y) d! (16.165)
2 1
so that the convolution integral can be put in the form

Z 1 Z 1
1 1
f1 f2 = p f2 (y) p F1 (!) ei!(x y) d! dy (16.166)
2 1 2 1
Z 1 Z 1
1 1
f1 f2 = p F1 (!) p f2 (y) e i!y dyei!x d! (16.167)
2 1 2 1
Noting that Z 1
i!y
F2 (!) = f2 (y) e dy (16.168)
1
we can write the above expression as

Z 1
1
f1 f2 = p F1 (!) F2 (!) ei!x d! = F 1
[F1 (!) F2 (!)]
2 1
(16.169)
Example 15.10 Discuss, compute, and plot the convolution of the functions
f (x) and g(x) given below.
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
we have to be careful in evaluating this integral because of the function

f (y) :
16.8. FOURIER TRANSFORMS AND CONVOLUTION 475
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)
Case 2 : 10 < x < 0

For this case the range of y would be ( 20; 0) :In this interval we have f (y) =
0 for y < 10 and f (y) = 1 for 10 < y < 0; and we …nd
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
Case 3 : 0 < x < 10

For this case the range of y would be ( 10; 10) :In this interval we have
f (y) = 1; and we …nd
Z x x
1 1 y2
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
Case 4 : 10 < x < 20

For this case the range of y would be (0; 20) :In this interval we have f (y) = 0
for 10 < y < 20 and f (y) = 1 for 0 < y < 10; and we …nd
Z 10 10
1 1 y2
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
The graph of the convoluted functions is shown below

2.0
1.5
f*g
1.0
0.5
0.0
-20 -10 0 10 20 30
Important Properties of Convolution
(a) f (x) g(x) = g(x) f (x)

(b) convolution tends to smooth out f (x)
(c) convolution tends to spread out f (x)
Part III
physics III
477
Chapter 17
Manifolds
17.1 What is a Manifold?

Consider a ridged meterstick pinned at the north and south poles inside a hollow
sphere as shown in Fig.17.1. Initially, t = 0; the sphere is at rest and suddenly
Figure 17.1: Rotation of the coordinate axes.
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 17.2: The degenerate point on a plane in polar coordinates.
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.
17.2 Curves and surfaces in a Manifold

Both curves and surfaces on a Manifold are de…ned parametrically. That means
we use some common parameters. For example, a curve in the phase space
that we saw earlier, can be de…ned in terms of the time parameter, t: Another
example, a curve in the 4D Minkowski spacetime manifold is de…ned by the
interval
ds2 ( ) = c2 d 2 = c2 dt2 dx2 dy 2 dz 2 : (17.1)
Generally, we use a parameter, u; to de…ne a curve.

A curve: a curve in a Manifold of dimension N is de…ne by a parametric
equation
xa = xa (u) ; where a = 1; 2; 3:::N: (17.2)
For example the curve shown in pink in Fig. 17.4 is de…ned by
Figure 17.4: A curve and a surface in a 3D manifold.
x1 (u) = cos5 (u); x2 (u) = 0:4u2 ; x3 (u) = 0:4u3 ; (17.3)
and it needs only one parameter, u:

A surface: a surface in a Manifold of dimension N (which also referred as a
submanifold) has M degrees of freedom which is always less than the dimension
of the Manifold (M < N ) and therefore it depends on M parameters that we
represent by (u1 ; u2 ; u3 ; :::uM ) and is de…ned by the parametric equation
xa = xa u1 ; u2 ; u3 ; :::uM ; where a = 1; 2; 3:::N: (17.4)
Hypersurface: a surface of dimension M in a Manifold of dimension N with

M = N 1: In this case the N 1 parameters can be eliminated from the N
equations and you can …nd one equation
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)
where we used (x1 ; x2 ; x3 ) for (x; y; z).Introducing the parameters u1 ; u2 ;
Figure 17.5: A spherical hypersurface embedded in a 3D manifold. The radius

of the sphere is a.
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)
which is the property of a hypersurface in a manifold.
Therefore a point is restricted to lie in a hypersurface (M = N 1 di-

mensional submanifold embedded in N -dimensional Manifold), then the points
coordinate must satisfy Eq. (17.5). We come up with a similar generalization
to this for a point that belong to any surface with dimension M in a Manifold
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
17.3 Coordinate transformations and summation

convention
Let’s consider the 3-D Euclidean Manifold. A point in this Manifold can be
represented using Cartesian coordinates (x; y; z) which we shall represent by
(x1 ; x2 ; x3 ). This same point can also be represented using spherical coordinates
(r; ; ') that shall represent by (x01 ; x02 ; x03 ). Now the question is how we relate
the Cartesian coordinates with the spherical coordinates or vice versa. In terms
of these notations, one can write
r ! r (x; y; z) ; or x01 ! x01 x1 ; x2 ; x3 ; (17.11)
02 02 1 2 3
! (x; y; z) ; or x ! x x ;x ;x ;
03 03 1 2 3
' ! ' (x; y; z) ; or x ! x x ;x ;x ;
or
x ! x (r; ; ') ; or x1 ! x1 x01 ; x02 ; x03 ; (17.12)
2 2 01 02 03
y ! y (r; ; ') ; or x ! x x ;x ;x ;
3 3 01 02 03
z ! z (r; ; ') ; or x ! x x ;x ;x :
Suppose we have a function, g (x; y; z) or g (r; ; '), which can be expressed by
g x1 ; x2 ; x3 and g x01 ; x02 ; x03 ; respectively, then one can write [Theoretical
Physics I ],
@g @g @x01 @g @x02 @g @x03
= + + ; (17.13)
@x1 @x01 @x1 @x02 @x1 @x03 @x1
@g @g @x01 @g @x02 @g @x03
= + + ; (17.14)
@x2 @x01 @x2 @x02 @x2 @x03 @x2
@g @g @x01 @g @x02 @g @x03
3
= 01 3
+ 02 3
+ : (17.15)
@x @x @x @x @x @x03 @x3
Using matrices [Theoretical Physics I ] this can be put in the form
2 @g 3 2 @x01 @x02 @x03 3 2 @g 3
@x1 @x1 @x1 @x1 @x01
4 @g2 5 = 6 4 @x01 @x02 @x03 7 4 @g 5 :
5 (17.16)
@x @x2 @x2 @x2 @x02
@g @x01 @x02 @x03 @g
@x3 @x3 @x3 @x3 @x03
17.3. COORDINATE TRANSFORMATIONS AND SUMMATION CONVENTION485
For the inverse case, following a similar procedure, we can write

2 @g 3 2 @x1 @x2 @x3
32
@g 3
@x01 @x01 @x01 @x01
@x1
6 1 2 3
7
4 @g02 5 = 4 @x02 @x02 @x02 5 4 @g2 5
@x @x @x @x @x
(17.17)
@g 1 2 3 @g
@x @x @x
@x03 @x03 @x03 @x03 @x3
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
There follows that

2 32 @x1 @x2 @x3
3
@x01 @x02 @x03
@x1 @x1 @x1 @x01 @x01 @x01
6 @x01 @x02 @x03 76 @x
1
@x
2
@x
3
7
4 @x2 @x2 @x2 54 @x02 @x02 @x02 5 = 1: (17.20)
@x01 @x02 @x03 @x
1
@x
2
@x
3
@x3 @x3 @x3 @x03 @x03 @x03
This means the matrix

2 3
@x01 @x02 @x03
@x1 @x1 @x1
1 6 @x01 @x02 @x03 7
A =4 @x2 @x2 @x2 5 (17.21)
@x01 @x02 @x03
@x3 @x3 @x3
must be the inverse matrix for

2 @x1 @x2 @x3
3
@x01 @x01 @x01
6 @x
1
@x
2
@x
3
7
A=4 @x02 @x02 @x02 5 (17.22)
1 2 3
@x @x @x
@x03 @x03 @x03
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
is the transformation matrix that transforms the coordinates (x1 ; x2 ; x3 ) to

(x01 ; x02 ; x03 ).
For a Manifold of dimension N , the transformation matrix is given by
2 3
@x01 @x01 @x01
1
@x02 2
@x02 ::: @xN
6 @x @x : @x02 7
6 @x1 @x 2 @x3 7
@x
0a 6 : 7
6 : : : 7
= 6 : 7: (17.26)
@xb 6 : : : 7
6 : 7
4 : : : 5
0N 0N 0N
@x @x @x
@x1 @x2 ::: @xN
As we recall from [Theoretical physics I], a Matrix is invertible provided the

determinant is di¤erent from zero. Therefore, the inverse transformation is
possible provide the determinant of the transformation matrix which is known
as the Jacobian, J; is di¤erent from zero
@x01 @x01 @x01
@x1 @x2 ::: @xN
@x02 @x02 : @x02
" 0a
# @x1 @x2 @x3
@x : : : :
J = det = : 6= 0 (17.27)
@xb : : :
: : : :
0N 0N 0N
@x @x @x
@x1 @x2 ::: @xN
The inverse transformation Matrix can be written as

2 1 1
3
@x1 @x @x
:::
6 @x 01 @x 02
2
@x 0N
2 7
6 @x012 @x : @x 7
a 6 @x @x02 @x03 7
@x 6 : : : : 7
6
0b = 6
7 (17.28)
@x : : : : 7
6 7
6 : : : : 7
4 N N
5
@xN @x @x
@x01 @x02 ::: @x0N
and the Jacobian J 0

1 1
@x1 @x @x
@x01 @x02 ::: @x0N
2 2
@x2 @x : @x
a @x01 @x02 @x03
@x : : : :
J 0 = det = (17.29)
@x0 b : : : :
: : : :
N N
@xN @x @x
@x01 @x02 ::: @x0N
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
Figure 17.6: Point P and Q on a surface
Noting that for independent coordinates

0a
@x 0; a 6= c
= (17.31)
@x0c 1 a=c
we can generally write
N
X @x0a @xb a
= c: (17.32)
@xb @x0c
b=1
Consider two points P and Q on a Manifold with dimension N . Suppose these
points are separated by in…nitesimal interval so that if the coordinates of P is
xa and that of Q is xa + dxa ; then one can write
N
@x0a 1 @x0a 2 @x0a N X @x0a b
dx0a = dx + dx ::: N dx = dx ; (17.33)
@x1 @x2 @x @xb
b=1
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
here also the summation is evaluated at P .

Einstein’s summation convention: whenever an index occurs twice in an
expression, once as subscript and once as a superscript, imply a summation
over the index. An index should not occur more than twice. For example,
according to Einstein’s summation convention, the summations in Eq. (17.33)
and (17.34) can be expressed
N
X
0a @x0a @x0a b
dx = dxb = gb (x) dxb = dx (17.35)
@xb @xb
b=1
and
N
X @xa 0b @xa 0b
dxa = 0b
dx = gb (x0 ) dx0b = dx : (17.36)
@x @x0b
b=1
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 .
17.4 The Riemannian geometry

The local geometry of a Manifold : The local geometry of a manifold is de-
termined by de…ning the invariant ’distance’ or (as we saw in the Minkowski
spacetime Manifold) the interval ds between points P with coordinate xa and
Q with coordinates xa + dxa : This distance can be assigned in general to be a
well-behaved function g (xa ; dxa ) of the coordinates xa and dxa
ds2 = g (xa ; dxa ) : (17.37)
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
Figure 17.7: A curved surface in a 3D manifold.
these two points, ds, can be expressed as
@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
This cane be rewritten as
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
Again using Einstein’s summation convention and the notation

@g @g
gab (x) = (17.43)
@xa @xb
we may write
ds2 = gab (x) dxa dxb : (17.44)
which is the metric equation that we de…ned earlier. Note that the metric tensor
in this case is a 3 3 matrix given by
2 @g @g @g @g @g @g 3
@x1 @x1 @x1 @x2 @x1 @x3
G=4 @g
@x2
@g
@x1
@g
@x2
@g
@x2
@g
@x2
@g
@x3
5 (17.45)
@g @g @g @g @g @g
@x3 @x1 @x3 @x2 @x3 @x3
and we can easily see that this matrix is symmetric as
@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
In general theory of relativity, we are interested in a Manifold where the interval

ds can be described by the equation of the form
N X
X N
ds2 = gab (x) dxa dxb ; (17.47)
a=1 b=1
or simply using the Einstein’s summation convention
ds2 = gab (x) dxa dxb : (17.48)
A geometry of a Manifold de…ned by Eq. (17.48) is known as the Riemannian

geometry if ds2 > 0. As we have seen in the case of Minkowski spacetime
manifold the interval ds2 can also be negative (spacelike) or zero (lightlike).
In such cases the geometry is referred as pseudo Riemannian geometry and the
manifold can be referred as pseudo Riemannian. The function g (x) is known
as the metric function, where gab (x) represent the element of a metric tensor.
Transformation of the interval : applying the relation in Eq. (17.36), we can
express
@xa 0c @xb 0d
dxa = dx ; dxb
= dx (17.49)
@x0c @x0d
so that the interval can be transformed as
@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
17.5 Intrinsic and extrinsic geometry and the

metric
A given geometry of dimension M de…ned by the metric equation
ds2 = gab (x) dxa dxb : (17.53)
and embedded in a higher dimension manifold of dimension N , (N > M ) is said

to be:
(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.
Example 2.2 Extrinsic geometry: One simple example of extrinsic geometry is

2-D cylindrical geometry. In order to see that let’s consider a plane geom-
etry in a 3-D Euclidean Manifold. Let’s this plane depends on (x1 ; x2 ).
We recall from theoretical physics I generally a plane:
~ = ai + bj + ck is normal (perpendicular) to a plane,
Equation of a plane: If N
then the scalar product of the vector N~ and the vector !r ~r0
~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)
This de…nes the equation of the plane. It can be rewritten as
Figure 17.8: A plane geometry
ax + by + cz = d; (17.56)
where
d = ax0 + by0 + cz0 : (17.57)
Using the notation x1 ; x2 ; x3 for (x; y; z) , we may write
ax1 + bx2 + cx3 = d; (17.58)
where
d = ax10 + bx20 + cx30 : (17.59)
1 2
For a plane that depends on only (x ; x )we may write the interval between
(x1 ; x2 ) and (x1 + dx1 ; x2 + dx2 ) as

2 2
ds2 = dx1 + dx2 : (17.60)
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
g (a; '; z) = g a; x02 ; x03 : (17.61)
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)
For the cylinder with radius a shown in Fig. 17.9
x1 = a cos x02 ; x2 = a sin x02 ; x3 = x03
we …nd
2 2 2 2 2
ds2 = dx1 + dx2 + dx3 = a2 dx02 + dx03 : (17.63)
Figure 17.9: A surface with cylindrical geometry.
This is a 2-D surface embedded in a 3-D manifold. It has cylindrically

curved geometry when it is viewed in this 3-D Euclidean manifold. But
you can actually obtain this geometry from the plane geometry from Eq.
(17.40) by simply substituting
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..
Example 2.3 Intrinsic geometry: One simple example of intrinsic geometry is

2-D spherical geometry embedded in a 3-D Euclidean Manifold. In Fig.
17.5 we see a 3-D in…nitesimal volume. Assume a sphere with radius a:
Then the surface de…ned by a pair of points on this sphere separated by

a distance ds can be expressed as
2 2 2 2
ds2 = (ad ) + (a sin ( ) d') = a2 (d ) + a2 sin2 ( ) (d') : (17.64)
or using the notation x02 ; x03 for ( ; ') ; we can write

2 2
ds2 = a2 dx02 + a2 sin2 x02 dx03 : (17.65)
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)
For a two dimensional spherical geometry with radius a; we have

2 2 2
x1 + x2 + x3 = a2 ; (17.68)
Using (17.68), we can write

r
2 2
x3 = a2 (x1 ) + (x2 ) ; (17.69)
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 )
If we consider a point in the neighborhood of the pole, we may set
x1 = x2 ' 0
and the metric in Eq. (17.70) reduces to the Euclidean form

2 2
ds2 = dx1 + dx2 : (17.71)
Let’s use the coordinates (x01 ; x02 ; x03 ). de…ned by the transformation
x1 = x01 cos x02 ; x2 = x01 sin x02 ; x3 = x03
For any point on the surface of the sphere, we have

2 2 2
x1 + x2 + x3 = a2
r q
2 2 2
) x3 = a2 (x01 ) + (x02 ) = a2 (x01 ) ; (17.72)
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 )
so that after a little algebra, we …nd

0 12
01 01
2 2 2 x dx
ds2 = dx01 + x01 dx02 + @q A (17.74)
2
a2 (x01 )
which simpli…es into

2
a2 dx01 2 2
ds2 = 2 + x01 dx02 (17.75)
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.
Example 2.5 Determine the metric for a three-dimensional sphere of radius a

embedded in a 4-D Euclidean space: We can write the interval
2 2 2 2
ds2 = dx1 + dx2 + dx3 + dx4 ; (17.78)
For a three dimensional spherical geometry with radius a we have

2 2 2 2
x1 + x2 + x3 + x4 = a2
r
2 2 2
) x4 = a2 (x1 ) + (x2 ) + (x3 ) (17.79)
so that
x1 dx1 + x2 dx2 + x3 dx3
dx4 = r :
2 2 2
a2 (x1 ) + (x2 ) + (x3 )
Then the metric can be expressed as

2
2 2 2 x1 dx1 + x2 dx2 + x3 dx3
ds2 = dx1 + dx2 + dx3 + ; (17.80)
2 2 2
a2 (x1 ) + (x2 ) + (x3 )
We introduce the coordinates (r; ; ') which we represent (x01 ; x02 ; x03 ).
and de…ned by the transformation
x1 = r sin ( ) cos (') ; x2 = r sin ( ) sin (') ; x3 = cos ( ) ;
or
x1 = x01 sin x02 cos x03 ; x2 = sin x02 sin x03 ; x3 = cos x02 :
Now referring to Fig.17.5the "distance" squared between point P and Q

can easily be determined using the Pythagorean theorem. First …nd the
length of the hypotenuse of the green triangle. Suppose if we call this
length dss ; we note that
2 2
ds2s = (r sin ( ) d') + (rd ) : (17.81)
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)
Note that I referred this distance as ds0 because it represents distance
ds02 = dx2 + dy 2 + dz 2
We are considering a 3-D sphere in a 4-D Euclidean manifold, where the

interval between point P and Q is given by
2 2 2 2 2
ds2 = ds02 + dx4 = dx1 + dx2 + dx3 + dx4 ; (17.83)
with the constraint

r
4 2 2 2
x = a2 (x1 ) + (x2 ) + (x3 ) :
which we may write, in terms of the polar coordinates, as

p
x4 = a2 r2 : (17.84)
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

a2 2
ds2 = dr2 + r2 d + r2 sin2 ( ) d'2 (17.87)
a2 r2
or
2 2 2
ds2 = g11 dx01 + g22 dx02 + g33 dx03 ; (17.88)
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.
17.6 Length, area, and volume

Length: Suppose two points P and Q on a manifold of dimension N are con-
nected by some curve. The length of the curve connecting these two points is
given by
Z Qp Z Qq
LP Q = 2
jds j = jgab (x) dxa dxb j (17.89)
P P
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
Area and volume:

Generally the area can determined using
ZZ ZZ q q
A = ds1 ds2 = jgab (x) dxa dxb j jgcd (x) dxc dxd j; for c; d 6= a; b
ZZ q ZZ s
dxa dxc 2
= jgab (x) gcd (x) dxa dxb dxc dxd j = gab (x) gcd (x) b d (dxb dxd )
dx dx
ZZ s
dxa dxc
) A= gab (x) gcd (x) b d dxb dxd : (17.92)
dx dx
17.6. LENGTH, AREA, AND VOLUME 499
For an orthogonal set of coordinates, xp
dxp pq 1; p=q
= = (17.93)
dxq 0; p 6= q
and the area becomes

ZZ r ZZ p
ab cd
A= gab (x) gcd (x) dxb dxd = jgbb (x) gdd (x)jdxb dxd ;
note that, d 6= b: As an example let’s consider a 2D hypersurface embedded in

a 3D manifold. Then the surface area is given by
2
X ZZ p ZZ p
1 d
A = jg11 (x) gdd (x)jdx dx + jg22 (x) gdd (x)jdx2 dxd
1
ZZ p
+ jg33 (x) gdd (x)jdx3 dxd :
ZZ p ZZ p
= jg22 (x) g11 (x)jdx2 dx1 + jg33 (x) g11 (x)jdx3 dx1
ZZ p ZZ p
1 2
+ jg11 (x) g22 (x)jdx dx + jg33 (x) g22 (x)jdx3 dxd
ZZ p ZZ p
+ jg11 (x) g33 (x)jdx1 dx3 + jg22 (x) g33 (x)jdx2 dx3
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
which can simply be written as

ZZ p
A= jg22 (x) g11 (x)jdx2 dx1
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)
and the in…nitesimal area dA of the surface de…ned by
xa = constant; (17.95)
for, a = 3; 4:::N; is given by

p
dA = jg11 g22 jdx1 dx2 (17.96)
and for 3-D volume in the x1 ; x2 ; x3 de…ned by
xa = constant (17.97)
for a = 4; 5:::N; the in…nitessimal volume is given by

p
dV = jg11 g22 g33 jdx1 dx2 dx3 (17.98)
Example 2.5 For the two-dimensional sphere of radius a embedded in a 3-D

Euclidean space consider a surface de…ned by a radius, = R: For this
surface …nd
(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
We recall the length of a curve between two points, P and Q, on a manifold

in terms of the metric tensor is given by
Z Qq
LP Q = jgab (x) dxa dxb j: (17.101)
P
For curve on the surface of the sphere shown in red in Fig. 17.6, the
coordinates for point P is x01 = = 0; x02 = = constant and for point
Q x01 = = R; x02 = = constant which leads to
dx02 = d = 0:
Then the length becomes
Z Qp Z R
a 1 R
D= jg11 (x0 ) dx01 dx01 j = p d = a sin :
P 0 a2 2 a
(17.102)
(b) Along the circumference (the curve shown in pink Fig. 17.6), we know that
x01 = = R =constant which leads to
dx01 = d = 0 (17.103)
Thus the expression
Z Q q
LP Q = jgab (x0 ) dx0a dx0b j (17.104)
P
for the circumference becomes
Z 2 r Z 2 p
2
C= g22 (x0 ) (dx02 ) = jg22 (x01 = = R)j dx02
0 0
Z 2
)C= Rdx02 = 2 R (17.105)
0
(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)
Then the surface area becomes

ZZ s ZZ
a2 a
A= 2 d d = p d d : (17.107)
a2 2
a 2 2
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
Homework : In Example 2.5, express the circumference (part b) and the

area (part c) in terms of the distance D (part a) and determine D for the the
maximum circumference and area. Find the maximum circumference and area
of this sphere. Is your answer is consistent with what you know about the
maximum circumference and surface area of a sphere with radius a.
Example 2.6 For the three-dimensional sphere embedded in a 4-D Euclidean

space by considering a 2-D sphere of coordinate radius r = R; …nd
(a) the distance from its center to the surface of this sphere along constant
and constant .
(b) the circumference across the equator

(c) the area of the spherical surface.

(d) The volume bounded by the spherical surface.
Solution:
(a) We recall the metric for a 3-D sphere embedded in a 4-D Euclidean space
is given by
2 2 2
ds2 = g11 dx01 + g22 dx02 + g33 dx03 ; (17.109)
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 ( ) (17.110)
a2 r2
or in terms of x01 ; x02 ; x03
a2 2 2
g11 = 2 ; g22 = x01 ; g33 = x01 sin2 x02 : (17.111)
a2 (x01 )
For x02 = =constant and x03 = =constant, we have
dx02 = dx03 = 0: (17.112)
so that the distance D
Z Q q
LP Q = jgab (x0 ) dx0a dx0b j (17.113)
P
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
for the circumference, becomes

Z r Z 2 p
2
C = g33 (x0 ) (dx03 ) = jg33 (x01 = r = R; x02 = =2)jdx03
0
Z 2 q Z 2
) C= R2 sin2 ( =2) dx003 = Rdx003 = 2 R (17.117)
0 0
(c) The spherical surface is de…ned by x01 = R =constant: Therefore, in view of

to Eq. (17.96), the in…nitessimal area on this surface should be expressed
as p
dA = jg22 g33 jdx02 dx03 : (17.118)
Noting that for x01 = r = R
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
r = a sin ( ) ) dr = a cos ( ) d
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
Then the volume becomes

8 s 9
< R R R
2=
1
V = 2 a3 sin 1 (17.129)
: 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
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)
Homework: Determine the metric for a three-dimensional sphere with imag-

inary radius a = ib embedded in a 4-D Euclidean space: and by considering a
sphere de…ned by r = R …nd
(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
17.7 Local Cartesian coordinates and tangent

space
Generally ds2 ; can be positive, negative, or zero as we saw in pseudo-Riemannian
spaces, like the Minkowsky spacetime. For now we shall consider what normally
refer as Riemannian where the metric
ds2 = gcd (x) dxc dxd ; (17.135)
is positive. It is not possible, in general, to …nd a coordinate transformation

that transforms the metric into Euclidean form
2 2 2
ds2 = dx01 + dx02 + ::: + dx0N = 0a 0b
ab dx dx ; (17.136)
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
Thus in the neighborhood of point P , we have

h i
0 2
gab (x0 ) = ab + O x0 x0p
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
Considering only up to the second order terms,

dg(x0 ) 1 dg(x0 )
g(x) = g(x) + (x xp ) + (x xp )2 + ::: (17.140)
dx0 x0 =xp 2! dx0 x0 =x p
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 )
which can can be put in the form

" 2
# " 2
#
2 x1 2 x2 2
ds = 1 + 2 2 dx1 + 1+ 2 2 dx2
a2 (x1 ) (x2 ) a2 (x1 ) (x2 )
2x1 x2 1
+ 2 2 dx dx2 : (17.143)
a2 (x1 ) (x2 )
where x1 ; x2 ; x3 corresponds to the usual Cartesian coordinates (x; y; z):

If one pic a point P with coordinates x1p ; x2p so that one may introduce
the transformation de…ned by
x1 = x01 x1p ; x2 = x02 x2p ) dx1 = dx01 ; dx2 = dx02
and express the metric as
2 2
ds2 = g11 dx01 + g22 dx2 + g12 dx01 dx02 :
where
2
x01 x1p
g11 x01 ; x02 = 1+ 2 2;
a2 x01 x1p x02 x2p
2
x02 x2p
g22 x01 ; x02 = 1+ 2 2
a2 x01 x1p x02 x2p
2
2 x01 x1p x02 x2p
g12 x01 ; x02 = 2 2
a2 x01 x1p x02 x2p
Show that the metric is Locally Cartesian at point P .
Solution: For a locally Cartesian, one must be able to show that at point P
0
0 @gab (x0 )
gab (xaP ) = ab ; = 0: (17.144)
@x0c xa
P
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
Tangent Space to Manifolds: for an arbitrary point P in an N dimensional

Riemannian manifold we can …nd a space that consist of coordinates (xa ) such
that in the Neighborhood of P the line element is Euclidean. That means the
line element is
2 2 2
ds2 = dx1 + dx2 + ::: dxN
this space is called the tangent space, T p . You can see in Fig.17.10, three
di¤erent two-dimensional Tangent spaces (shown in red with green rectangular
grids) at three di¤erent points on a 2D surface embedded in a 3D manifold.
Fig.17.10 In each of these tangent spaces in Fig. 17.10, note that
2 2
ds2 = dx01 + dx02 : (17.148)
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
x3 x1 ; x2 = sin2 (x1 ) + cos(x2 ) (17.149)
which one may write as
g x1 ; x2 ; x3 = sin2 (x1 ) + cos(x2 ) x3 = 0 (17.150)

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
x3 x1 ; x2 = 2 sin(x1p ) cos(x1p ) x1 x1p sin(x2p ) x2 x2p + x3p : (17.157)
Introducing the coordinate transformation de…ned by
x01 = x1 x1p ; x02 = x2 x2p ; x03 = x3p

) dx01 = dx1 ; dx02 = dx2 ; dx03 = 0 (17.158)
then the metric in the tangent space becomes

2 2 2 2 2 2
(ds) = dx01 + dx02 + dx03 = dx01 + dx02 (17.159)
17.8 The signature of a manifold

Let’s consider an arbitrary point P in the pseudo-Riemannian manifolds de-
scribed by the coordinates, xa ; and make a transformation to local Cartesian
coordinates, x0a ; at this point. Suppose this transformation, which satis…es the
conditions in Eq. (17.137) and (17.138), is given by
a b
0 @x @x
gcd (x0 ) = gab (x0 ) : (17.160)
@x0c @x0d
Then at point, P; if the coordinates are x0p = x1p ; x2p ; x3p ; :::xN
p ;
a b
0 @x @x
gcd (xp ) = gab (xp ) ; (17.161)
@x0c xp @x0d
xp
and if the transformation leads to Local Cartesian, we must have

0
0 @gab (x0 )
gcd (xp ) = cd c ; = 0; (17.162)
@x0c x0 =xp
for pseudo-Riemannian manifolds in general. In fact c = 1 for Local Cartesian

in a Riemannian manifolds. Therefore, the transformation metric
a b
0 @x @x
gcd (x0 ) = gab (x0 ) ; (17.163)
@x0c @x0d
for local Cartesian at point P; one can write

a b
0 @x @x
gcd (xp ) = cd c = gab (x0 ) (17.164)
@x0c xp @x0d
xp
a b
0 @x @x
) gcd (xp ) = gab (xp ) : (17.165)
@x0c xp @x0d
xp
This can be put using matrices as
G0 = X T GX: (17.166)
One can easily show this by considering, N = 2, and the matrices
1 0 g11 x1p ; x2p g12 x1p ; x2p

G0 = ;G =
0 2 g21 x1p ; x2p g22 x1p ; x2p
2 1 01 1 3
@x (x ;x02 ) @x (x01 ;x02 )
6 @x01 @x02 7
6 x1p ;x2p x1p ;x2p 7
X = 6 2 2 7;
4 @x (x01 ;x02 ) @x (x01 ;x02 ) 5
@x01 @x01
x1p ;x2p x1p ;x2p
2 1 2 3
@x (x01 ;x02 ) @x (x01 ;x02 )
6 @x01 @x01 7
6 x1p ;x2p x1p ;x2p 7
XT = 6 1 2 7 (17.167)
4 @x (x01 ;x02 ) @x (x01 ;x02 ) 5
@x02 @x01
x1p ;x2p x1p ;x2p
Since the matrix G0 is a diagonal matrix, the transformation is a similarity

transformation (Mathematical Methods II). This means X forms the eigen vec-
tor matrix and G0 forms the eigenvalue matrix and
XT = X 1
) G0 = X 1
GX: (17.168)
Using the inverse transformation matrix, X 0 ; and the corresponding inverse

matrix, X 0 1 ; one must recover the matrix G,
G = X0 1
G0 X 0 = X 0 1
X 1
GXX 0 (17.169)
Therefore we must have
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
Since the metric tensor is symmetric it can be diagonalized by a similarity trans-

formation provided we chose the columns of X to be the normalized eigenvectors
of the matrix G. This means
2 3
1 0 ::: 0
6 0 2 ::: 0 7
6 7
6 : : 7
0
G =6 6 7; (17.172)
6 : : : : 77
4 : : 5
0 0 ::: N
where 1; 2 ::: N are the eigenvalues of the matrix G: The metric
ds2 = gcd (x0 ) dx0c dx0d ; (17.173)
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)
where jgj is the determinant of the matrix

2 3
g11 0 ::: 0
6 0 g22 ::: 0 7
6 7
6 : : 7
G = [gab ] = 6
6 :
7
7 (17.179)
6 :: ::: : 7
4 : : 5
0 0 ::: gN N
Chapter 18
Vector Calculus on
manifolds
18.1 The tangent vector

The tangent vector, ~t; at point p on a Manifold is the vector that lies in the
tangent space, Tp ; at that point, p1 ; and is given by
~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-
Figure 18.1: A tangent vector at point P in the tangent space.
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
18.2 The basis vectors

At each point P on a manifold we can de…ne a set of linearly independent basis
vectors, eâ (x) ; for the tangent space, Tp : The number of the basis vectors is
equal to the dimension of Tp : Any vector …eld at point P; ~v (x) ; can then be
expressed as linear combination of these basis vectors
~v (x) = v a (x) eâ (x) ; (18.2)
where v a (x) are known as the controvariant components of the vector …eld,
~v (x) ; in the basis eâ : For any set of basis vectors we can de…ne another set of
basis vectors known as the dual basis vectors, eâ (x) ; de…ned by
a
eâ (x) e^b (x) = b: (18.3)
The dual basis vectors, eâ ; and the basis vectors, e^b ; form a reciprocal system
of vectors. The local vector …eld ~v (x) can also be expressed in terms of the dual
basis vectors as
~v (x) = va (x) eâ (x) ; (18.4)
where va (x) are known as the covariant components of the vector …eld, ~v (x) ;
in the dual basis vectors, eâ (x) : The controvariant and covariant components
of the vector …eld can be determined using Eq. (18.3)
a
~v (x) e^b (x) = v a (x) eâ (x) e^b (x) = v a (x) b = v b (x) : (18.5)
Similarly
va (x) = ~v (x) eâ (x) : (18.6)
The coordinate basis vectors: in any particular coordinate system xa ;we
can de…ne at every point P of the manifold a set of N coordinate basis vectors
~s
eâ = lim ; (18.7)
a x !0 xa
where ~s is the in…nitessimal separation vector between point P and some nearby
point Q with coordinate separation xa from P . For exampleThus eâ is the
tangent vector to the xa coordinate curve at the point P: As an example we
reconsider the 2D (blue) surface embedded in a 3D Manifold shown in Fig.
18.2. We recall that this surface is de…ned by
x3 x1 ; x2 = sin2 (x1 ) + cos(x2 ) (18.8)
where x1 ; x2 ; and x3 are the usual Cartesian coordinates (x; y). The tangent
plane at point P is given by
x3 x1 ; x2 = 2 sin(x1p ) cos(x1p ) x1 x1p sin(x2p ) x2 x2p + x3p : (18.9)
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
Figure 18.2: The basis vectors in the tangent space.
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
and noting that

x1 x1
2
= 2 =0
x x
we …nd
^ + 2 sin(x1 ) cos(x1 )^
e^1 = x z ; e^2 = y^ sin(x2 )^
z (18.14)
18.3 The metric function and coordinate trans-

formations
In…nitesimal vector separation: Consider two points P and Q on a manifold
with coordinates xa and xa + dxa ; where dxa is none zero for all a, then the
in…nitessimal vector separation between these two points is given by
d~s = eâ (x) dxa (18.15)

The metric function-the covariant components: The equation that deter-

mines the elements of the metric tensor in the metric
ds2 = gab (x) dxa dxb (18.16)
can be obtained from the inner product of the in…nitessimal vector separation.
We note that
ds2 = d~s d~s = eâ (x) dxa e^b (x) dxb = eâ (x) e^b (x) dxa dxb
) ds2 = gab (x) dxa dxb ; (18.17)
where
gab (x) = eâ (x) e^b (x)
is the metric function
Example 3.1 Let’s reconsider the 2D sphere of radius, a, in a 3D Euclidean

space. For an origin set at the north pole of the sphere, a point P is
described by the vector
p
^ + sin (') y^ + a2
~s = cos (') x 2z
^
Figure 18.3: 2D sphere of radius a embedded in 3D Euclidean space. The origin

of the Cartesian coordinates (x1 ; x2 ; x3 ) [or (x; y; z)] is set at the north pole.
(x01 = ; x02 = '):
(a) Find the basis vectors e^01 and e^02 in the tangent space at point P.
(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)
There follows that

~s
e^ = lim
!0
" #
'
= cos (') x
^ + sin (') y^ p z^ + [ sin (') x
^ + cos (') y^] :
a2 2
(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)
Figure 18.4: The basis vectors (^

e ; e^' ) at a point P in the tangent space of a
2D sphere embedded in a 3D Euclidean space.
(b) The metric elements can be determined using the basis vectors
gab (x) = eâ (x) e^b (x) ; (18.25)
and one …nds

2
a2
g11 = g = e^ e^ = 1 = ;
a2 2 a2 2
2
g22 = g'' = e^' e^' = ; g12 = g21 = 0 (18.26)
Homework: Consider the 3D sphere of radius, a, in a 4D Euclidean space.

A point P on this 3D sphere is described by the vector
p
^ + r sin ( ) sin (') y^ + r cos ( ) z^ + a2 r2 w
~s = r sin ( ) cos (') x ^ (18.27)
(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.
The metric function-the controvariant components: noting that the in…n-

itessimal distance between point P and Q can be expressed using the in…nites-
simal vector separation
d~s = eâ (x) dxa (18.28)
we have
b
e^b (x) d~s = e^b (x) eâ (x) dxa = a dx
a
= dxb : (18.29)
Expressing the in…nitessimal vector separation using the dual coordinates basis
vectors as
d~s = eâ (x) dxa (18.30)
we also …nd
b
e^b (x) d~s = e^b (x) eâ (x) dxa = a dx
a
= dxb : (18.31)
Thus the metric
ds2 = d~s d~s = eâ (x) dxa e^b (x) dxb = g ab (x) dxa dxb ; (18.32)
where the controvariant components of the metric tensor can be de…ned as
g ab (x) = eâ (x) e^b (x) (18.33)
Orthonormal basis vector: at a point on a manifold an orthonormal basis

vectors are de…ned by
eâ (x) e^b (x) = ab ; (18.34)
where 2 3
1 0 ::: 0
6 0 1 ::: 0 7
6 7
6 : : 7
[ ab ] = 6
6
7:
7 (18.35)
6 : : : : 7
4 : : 5
0 0 ::: 1
or in short [ ab ] = diag ( 1; 1; :::; 1) is the Cartesian line element of the
tangent space, Tp ; and depends on the signature of the pseudo-Riemannian
manifold.
Basis vectors and coordinate transformations: Suppose we make a
coordinate transformation from xa where the basis vectors are eâ to a new an-
other coordinates x0a we want to …nd how the new basis vectors are transformed
into, e^0a . If we consider point P and another point Q at an in…nitesimal distance
away from point P . The in…nitesimal displacement, d~s; between these points is
independent of the coordinate transformation and we must have
d~s = eâ dxa = e^0c dx0c (18.36)
Using
@xa 0b
dxa = dx ; (18.37)
@x0b
we have
@xa
d~s = eâ dxa = eâ dx0b = e^0c dx0c : (18.38)
@x0b
Noting that
e^b e^b = 1
we have
@xa
eâ dx0b = e^0b e^0b e^0c dx0c = e^0c e^0b e^0b dx0c = e^0b bc dx0c (18.39)
@x0b
@xa
) eâ dx0b = e^0b dx0b : (18.40)
@x0b
so that
@xa
e^0b =
eâ : (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
~v = v a eâ ; ~v 0 = v 0a e^0a (18.43)
or in terms of the dual basis vectors
~v = va eâ ; ~v 0 = va0 e^0a (18.44)
But the vector is the same since it describes a geometrical entity that is inde-
pendent of the coordinate system. Therefore, we must have
~v = v a eâ ; = v 0a e^0a ; (18.45)
so that taking the inner product of ~v and e^0b , we can write

b
v a e^0b eâ = v 0a e^0b e^0a ) v 0a a = v a e^0b eâ ) v 0b = v a e^0b eâ : (18.46)
Applying the relation in Eq. (18.42), one can write
@x0b c
e^0b = e^ ; (18.47)
@xc
so that
@x0b c @x0b @x0b a
c
e^ eâ = 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
18.3.1 Raising and lowering vector indices

The scalar product: The scalar product of two vectors at a point, P; on a
manifold
~v = v a eâ : (18.50)
and
w
~ = w^
eb : (18.51)
is given by
~ = v a eâ wb e^b = gab v a wb :
~v w (18.52)
where
gab = eâ e^b : (18.53)
is the covariant components of the metric tensor. We can also use the dual basis
vectors to express the vectors ~v (x) and w
~ (x)
~v = va eâ : (18.54)
and
~ = wb e^b :
w (18.55)
so that the inner product becomes
~ = va eâ wb e^b = g ab va wb :
~v w (18.56)
where
g ab = eâ 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 eâ wb e^b = eâ e^b v a wb =
~v w b v wb = v a wa (18.58)
or
a
~ = va eâ wb e^b = eâ 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)
Similarly, from Eqs. (18.56) and (18.59), we …nd
~ = 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
eâ = gad e^d (18.62)

and
e^c = g cb e^b : (18.63)
Then noting that the inner product
a a
eâ 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)
This means the metric tensor [g ab ] with the controvariant components, g ab ; is

the inverse matrix of the metric tensor [gab ] with the covariant elements, gab ; :
Thus
GG~ = GG~ = I;
~ = [g ab ] is its inverse.
where G = [gab ] is the metric tensor and G
18.4 The inner product and null vectors

The scalar product of two vectors at a point on a manifold which can be ex-
pressed in four di¤erent ways
gab v a wb = va wa = g ab va wb = v a wa (18.65)
Suppose we take the scalar product of vector, ~v ; with itself, we have
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
ds2 = gab (x) dxa dxb
which could be zero or negative. To accommodate such kind of vectors we de…ne

the length of a vector, ~v , as
q q p q
v = jgab v a v b j = jg ab va vb j = jv a vb j = jva v b j: (18.67)
As is the case in pseudo-Riemannian manifold, the length of a vector can be

zero without the vector being a zero vector (i.e. va 6= 0). Vectors with length
(magnitude) zero with none zero components is known as Null vectors.
The cosine angle between vectors: The angle between two non-null
vectors at a point on a manifold is de…ned by:
v a wa
cos ( ) = p p (18.68)
jvb v b j jwc wc j
In the pseudo-Riemannian manifold, Eq. (18.68) can lead to jcos ( )j > 1:

18.5. THE AFFINE CONNECTIONS 525
Orthogonal vectors: two vectors
~v = v a eâ : (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)
18.5 The a¢ ne connections

It is important to know how vectors changes as the coordinate or the parameter
that de…nes the coordinate changes . For example, in Mankawski space time we
may be interested in the 4D momentum, P~ how it changes with time (the proper
time, ) so that one can explain the condition for conservation of momentum
in general relativity. In such cases for the 4D momentum expressed in terms of
its controvariant components as
P~ = pa eâ (18.72)
one must be able to determine
dP~ d a dpa d^
ea
= (p eâ ) = eâ + 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
= eâ + 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)
@' @
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
Combining this relation with the basis vector, e^ ;
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
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 eâ
= 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
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
@ê ' @^
e '
= e^ + e^' ; = 'e
^ + 'e
^' ;
@ @'
@^
e' ' @ê' '
= ' e^ + ' e^' ; = '' e
^ + '' e
^' ; (18.89)
@ @'
you are going to determine
' ' ' '
; ; '; '; ' ; ' ; '' ; and '' : (18.90)
Also in this case the origin is at the center of the sphere.

(2) Find all the elements for the a¢ ne connection, bac ; for a point on a 3D
sphere embedded in a 4D Euclidean space. Note that in the expressions
Taking the scalar product of Eq. (18.88) and the dual basis vectors,
@^
ea
e^d = b d
ac e
^ e^b ; (18.91)
@xc
applying the properties of the dual basis vectors
d
e^d e^b = b (18.92)
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 eâ (18.94)
@xc
From now on we will use the notation @c eâ ;
@^
ea
@c eâ = : (18.95)
@xc
Di¤erentiating Eq. (18.92) with respect to the coordinate, xc ; and applying the
notation, we …nd
b
eâ @c e^b + e^b @c eâ = @c a = 0 ) e^b @c eâ = eâ @c e^b (18.96)
Using the de…nition of the derivative of the basis vectors and our notation
@^
ea b d
@c eâ = = bac e^b ) e^d @c eâ = b
ac e
^b e^d = ac b = d
ac (18.97)
@xc
d
) ^d @c eâ
ac = e (18.98)
applying this relation,
e^b @c eâ = eâ @c e^b ) e^b @c eâ = a

bc ; (18.99)
and noting that

e^b e^b = 1 (18.100)
we …nd
e^b @c eâ = a
bc e
^b e^b = e^b a b
bc e
^
a a b
) @c e^ = bc e
^ (18.101)
The a¢ ne connection under coordinate transformation: Suppose we

make the coordinate transformation xa ! x0a ; for the a¢ ne connection
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
0b @x0b @xf @xg d @^ ef @ 2 xf @x0b d

ac = d 0a 0c
e^ g
+ e^ e^f (18.106)
@x @x @x @x @x0c @x0a @xd
Now applying
@^
ef f
e^d = d
fg; e^d e^f = d
@xg
Eq. (18.106) becomes
0b @x0b @xf @xg d @x0b @ 2 xd

ac = fg + (18.107)
@xd @x0a @x0c @xd @x0c @x0a
Homework: If we swap the derivatives with respect to x and x0 in the expression
@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
0b @x0b @xf @xg d @xd @xf @ 2 x0b

ac = fg (18.108)
@xd @x0a @x0c @x0a @x0c @xd @xf
The a¢ ne connection and the metric: we recall the a¢ ne connection
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)
is known as the torsion tensor. We will consider a torsionless manifolds, for

which
b
Tac = 0 ) bac = bca : (18.112)
We will determine the relationship between the a¢ ne connection and the metric
for a torsionless manifold. We recall the metric
gab = eâ e^b
so that
@gab
= @c gab = @c (^
ea e^b ) = eâ @c e^b + e^b @c eâ : (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 + eâ ^d = dac e^b
bc e e^d + d
bc e
â e^d
d d
) @c gab = ac gbd + bc gad : (18.115)
Similarly
@b gca = @b (^
ec eâ ) = e^c @b eâ + eâ @b e^c
d d
) @b gca = e^c ab e
^d + eâ 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 ;
g pa (@c gab + @b gca @a gbc ) = 2 d pa

bc g gad (18.120)
From Eq. (18.64), we have

a
g ab gbc = c

d p
g pa (@c gab + @b gca @a gbc ) = 2 bc d (18.121)
Since the summation is over d, replacing d by p, we …nd

p
g pa (@c gab + @b gca @a gbc ) = 2 bc (18.122)
Now relabeling the index a by d

p
g pd (@c gdb + @b gcd @d gbc ) = 2 bc (18.123)
and then the index p by a; we can write
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)
Applying the relation in Eq (18.124), we can express Eq. (18.125) as
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
g = gab cof [G]ab ) @c g = cof [G]ab @c gab : (18.135)
Noting that the cofactor matrix can be expressed as
g ab g = g ab gab cof [G]ab ) cof [G]ab = g ab g (18.136)
we may write
@c g = cof [G]ab @c gab = gg ab @c gab : (18.137)
18.6. LOCAL GEODESIC AND CARTESIAN COORDINATES 533
Now using the relation (18.132), we can write

@c g = gg ab @c gab = gg ab ( abc + bac ) : (18.138)
We recall
d
abc = gad bc (18.139)
so that
@c g = gg ab ( abc + bac ) = gg ab gad d
bc + gbd d
ac
b d a d
= g g ab gad d
bc + g ab gbd d
ac =g d bc + d ac
b a
= g bc + ac : (18.140)
Taking into account that a and b are dummy indices, we can replace b by a so
that
@c g = 2g aac (18.141)
This can be rearranged as
a 1 1 p
ac = @c g = @c ln jgj = @c ln jgj: (18.142)
2g 2
Now for the sake of convenience if we replace c by b, we may write the above
equation as p
a
ab = @b ln jgj (18.143)
The modulus is for the case where the manifold is seudo-Riemannian where the
metric elements can be negative.
18.6 Local geodesic and Cartesian coordinates

Let’s consider a manifold with coordinates system, xa ; and a point P on this
manifold with coordinates, xaP : Let’s now de…ne a new system of coordinates,
x0a ; in terms of xaP ; and the coordinates xa as
1
x0a = xa xaP + a
bc (P ) xb xbP (xc xcP ) : (18.144)
2
We must know that under this transformation how the a¢ ne connection trans-
formed. We recall that under coordinate transformation xa ! x0a ; the a¢ ne
connection is transformed according to

ac = fg : (18.145)
In order to determine this at point P, we need to di¤erentiate Eq. (18.144)
@x0a @xa 1 a @
= + bc (P ) xb xbP (xc xcP )
@xd @xd 2 @xd
@xa 1 a @xb @xc
= + bc (P ) (xc xcP ) + xb xbP ;(18.146)
@xd 2 @xd @xd
where we used the fact that abc (P ) ; xaP ; xbP ; and xcP are constant at point P.
The coordinates xa are independent coordinates,
@xa a @xb b @xc c

= d; = d; = d (18.147)
@xd @xd @xd
so that
@x0a 1 n o
a a b c
= d + bc (P ) (xc xcP ) d + xb xbP d (18.148)
@xd 2
which results in
@x0a a 1 a
= d + dc (P ) (xc xcP ) + a
bd (P ) xb xbP (18.149)
@xd 2
The summation indices are dummy indices and therefore we can replace b by c
in the second term so that
@x0a a 1 a
= d + [ dc (P ) (xc xcP ) + a
cd (P ) (xc xcP )] (18.150)
@xd 2
Since for a torsionless manifold
a a
dc (P ) = cd (P ) ;
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
Similarly, the inverse is also given by
@xa a
= d (18.153)
@x0d P
Di¤erentiating Eq. (18.151) with respect to xe , we have
@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

ac = fg (18.155)
18.6. LOCAL GEODESIC AND CARTESIAN COORDINATES 535
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)
which simpli…ed into
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
where [ ab ] = diag ( 1; 1; :::; 1) : We have also learned that the number of

positive entries (N+ ) minus the number of negative entries (N ) in [ ab ] is called
the signature of the manifold. For geodesic coordinates Eq. ( 18.159) can easily
be shown applying the equation that relates the connection with the metric in
Eq. (18.115). For the x0a coordinates, Eq. (18.115) can be written as
0 0d 0 0d 0
@c gab = ac gbd + bc gad (18.160)
When this equation is evaluated at point P

0
0 @gab (x0 ) 0d 0 0d 0
@c gab = = ac (P ) gbd (P ) + bc (P ) gad (P ) (18.161)
@x0c P
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
x00a = Xx0a (18.162)
where Xba are constants.

18.7 The gradient, the divergence, the curl on a

manifold
Before we see how the a gradient of a scalar, the divergence or the curl of a vector
is determined on a manifold, …rst we need to know how a vector is di¤erentiated.
Consider a vector, ~v , in terms of its controvariant components
~v = v a eâ ; (18.163)
where eâ 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 eâ ) = eâ @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 ) eâ + 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
â (18.166)
so that
@b~v = (@b v a ) eâ + v c a
cb e
â = (@b v a + a c
cb v ) e
â (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
@b~v = (rb v a ) eâ (18.169)
For geodesic coordinates where the a¢ ne connection vanishes,

a
cb =0
the covariant derivative reduces to
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 eâ (18.171)
show that
c
rb va = @b va ab vc (18.172)
18.7. THE GRADIENT, THE DIVERGENCE, THE CURL ON A MANIFOLD537
The covariant derivative of a scalar function: for a scalar function

the covariant derivative is
rb = @b (18.173)
The gradient: The gradient of a scalar function is given by
r = (@a ) eâ (18.174)
The divergence: The divergence of a vector …eld expressed in terms of its

covariant components
~v = v a eâ (18.175)
is given by
r ~v = ra v a (18.176)
Using the relation we obtained
cb v (18.177)
for b = a and replacing c by b, we …nd
r ~v = ra v a = @a v a + a b
ba v : (18.178)
Using the relation

a
p 1 p
ab = @b ln jgj = p @a jgj (18.179)
jgj
we can write
1 p 1 h p i
r ~v = @a v a + v a p @a jgj = p @a v a jgj (18.180)
jgj jgj
The Laplacian: We recall that in the Euclidean space the Laplacian of the
scalar function, ; is given by
r2 = r r : (18.181)
Applying Eq. (18.174) we can write
r2 = r [(@a ) eâ ] : (18.182)
But in order to apply the relation we derive for the divergence in Eq. (18.180),
we need the vector
~v = (@a ) eâ = va eâ : (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
~v = v a eâ = g ab (@b ) eâ : (18.184)

Then the Laplacian becomes
r2 = r g ab (@b ) eâ : (18.185)
Now applying the relation

1 h p i
r ~v = p @a v a jgj (18.186)
jgj
we may write the Laplacian

1 hp i
r2 = ra ra = p @a jgjg ab @b : (18.187)
jgj
The Laplacian symbol r2 is used in the usual 3-D Euclidean space or in an

N-D manifold. In 4-D spacetime manifold, as you will see (or have seen) in the
relativistic electrodynamics in Theoretical Physics IV (Electricity & Magnetism
II), r2 is replaced by 2 known as the d’Alembertian operator.
Curl: The curl is de…ned as a rank-2 antisymmetric tensor with components
(curl ~v )ab = ra vb r b va : (18.188)
Using the relation

c c
rb va = @b va ab v (18.189)
we can express the curl as
c c c c
(curl ~v )ab = @a vb ab v @ b va + ab v = @a vb @b va : (18.190)
18.8 Intrinsic derivative of a vector along a curve

We will encounter vector …elds that does depend on a curve instead of the entire
or some region of the manifold. In such cases the curve may be de…ned by the
coordinates xa that depends on some parameter, u: Let’s consider a vector, v;
expressed in terms of its controvariant components. Since the coordinates on
this curve depends on the parameter, u, this vector can be expressed in terms
of this parameter as
~v (u) = v a (u) eâ (u) : (18.191)
The derivative of this vector along this curve is given by
d d a d^
ea (u) dv a (u)
~v (u) = [v (u) eâ (u)] = v a (u) + eâ (u) (18.192)
du du du du
d ea (u) dxb
d^ dv a (u)
) ~v (u) = v a (u) + e
â (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 + eâ (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
â + eâ = + cb v eâ ; (18.197)
du du du du du
which we put in the form
d Dv a
~v (u) = eâ ; (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) eâ (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
18.9 Parallel transport

In order to understand the idea of parallel transport of a vector on a manifold
let’s consider motion of a particle in space. Suppose the position of the particle
depends on time, t, then the displacement is parametrized by time t, D ~ (t) : The
velocity of the particle is given by
~
dD
~v = (18.204)
dt
and the acceleration by

d~v
~a = : (18.205)
dt
Suppose you plot the displacement of the particle at di¤erent times, then you
would get generally a curve. The particle would have a constant velocity through
this curve provided its acceleration is zero.
d~v
~a = = 0: (18.206)
dt
This means the particle travels along this curve with a constant velocity. The
velocity would have the same magnitude and direction. The velocity vector
remains parallel at each point on the curve describing the displacement of the
particle as a function of time.
On a curve C on a manifold, a parallel transport of a vector
~v = va (u) eâ (u) (18.207)
is when the intrinsic derivative of this vector is zero

c
Dva dva b dx
= ac vb =0 (18.208)
Du du du
18.10 Null curves, non-null curves, and a¢ ne

parameter
We recall that the tangent vector t at point p on a Manifold is the vector that
lies in the tangent space Tp at that point and is given by
~t = lim @~s = d~s ; (18.209)

u!0 u du
where @~s is the in…nitessimal separation vector between the point P and some
nearby point Q on the curve on the manifold corresponding to the parameter
value u + u
In a given coordinate system xa with basis vectors eâ , we can express the
in…nitessimal separation vector d~s as
dxa
d~s = eâ (18.210)
du
so that the tangent vector becomes
a
~t = dx eâ (18.211)
du
Recalling that in pseudo-Riemmanian manifold the length of a vector ~v is given
by q q q
p
jvj = jgab v a v b j = jg ab va vb j = jv a vb j = jva v b j (18.212)
18.10. NULL CURVES, NON-NULL CURVES, AND AFFINE PARAMETER541
Figure 18.5: A tangent vector at point P in the tangent space.
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
ds2 = g ab dxa dxb
Eq. (18.213) becomes

ds
jtj = (18.214)
du
The non-null vectors and the a¢ ne parameter: for the non-null vec-
tors jtj 6= 0: This means according to Eq. (18.214) the distance ds at all points
on the curve must be di¤erent from zero and therefore it depends on the para-
meter u at all points on the curve, s = s (u). If parameter u and the distance s
are related by
u = as + b (18.215)
for a; b 6= 0; the parameter u is called the a¢ ne parameter on the curve.
The null vectors: if the tangent vector is a null vector, then
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.
18.11 Refreshment from theoretical physics: the

calculus of variation
Geodesic: The curve along a surface which marks the shortest distance between
two neighboring points. Finding geodesics is one of the problems which we can
solve using the calculus of variation.
Stationary point: A point on a given function f (x) is said to be stationary
point when
df (x)
= 0:
dx
Geodesic on an Euclidean 2-D space: 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.
Let’s consider two points on the x-y plane. Let …rst point be (x1 ; y2 ) and
the second point be (x1 ; y2 ) : Then the distance between these points is given
by the integral
Z (2)
I= ds (18.217)
(1)
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)
(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 may write Eq. (18.229) as

Z x2 Z !
(2) x1
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
(18.232)
Using the condition
(x1 ) = (x2 ) = 0
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
18.12 The geodesic

We recall that a local geodesic at point P is where the a¢ ne connection is zero
0b b b
ac (P ) = ac (P ) ac (P ) = 0: (18.236)
Suppose we have a set of points in Euclidean space de…ning a geodesic curve,

then at all of these points the a¢ ne connection must vanish.
f
@c e^b = bc e
^f =0 (18.237)
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 eâ (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) eâ ; (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) = eâ = + acb v c eâ ; (18.240)
du Du du du
for the tangent vector intrinsic derivative, we have
d~t Dta dta a c dx

b
= eâ = + cb t eâ ; (18.241)
du Du du du
so that substituting this into Eq. (18.239), we …nd
dta a c dx
b
dxa
+ cb t = (u) (18.242)
du du du
Noting that
a
~t = dx eâ = ta eâ (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
Therefore in general for a privilege parameter called the a¢ ne parameter the

equation for the geodesic in Eq. (18.244) can be written as
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
18.13 Stationary property of the non-null geo-

desic
Let’s consider the curve C in our manifold de…ned by the coordinates xa (u) :
Suppose we have two points 1 and 2 on this curve and we are interested in the
length along this curve joining these two points, this length can be determined
from
Z 2 Z 2q Z 2s
a b
dxa dxb
L= ds = jgab dx dx j = gab du (18.253)
1 1 1 du du
or using the notation

dxa
= x_ a (18.254)
du
this length can be expressed as
Z 2 q Z 2
L= jgab x_ a x_ b jdu: = F du; (18.255)
1 1
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
Using Eq. (18.256), we have

q
@F @ x_ a x_ b @c gab x_ a x_ b @c gab
= jgab x_ a x_ b j = p =
@xc @xc 2 jgab x_ a x_ b j 2s_
q
@F @ gab @ x_ a @ x_ b
= jgab x_ a x_ b j = p x_ b c + x_ a c
@ x_ c @ x_ c 2 jgab x_ a x_ b j @ x_ @ x_
g x_ b gab x_ a gcb x_ b gac x_ a
= p ab a
c + p b
c = p + p
2 jgab x_ a x_ b j 2 jgab x_ a x_ b j 2 jgcb x_ c x_ b j 2 jgac x_ a x_ c j
@F gcb x_ b gac x_ a
) c
= p + p (18.258)
@ x_ 2 jgcb x_ c x_ b j 2 jgac x_ a x_ c j
If we replace b by a in the …rst term, we can write

@F gca x_ a gac x_ a gac x_ a gac x_ a
= p + p = p =
@ x_ c 2 jgca x_ c x_ a j 2 jgac x_ a x_ c j jgac x_ a x_ c j s_
so that the Euler-Lagrange Equation becomes
d gac x_ a x_ a x_ b @a gab
= 0: (18.259)
du s_ 2s_
For the …rst terms we may write
d gac x_ a x_ a dgac gac dx_ a d 1 x_ a dgac gac dx_ a s•

= + +gac x_ a = + gac x_ a 2
du s_ s_ du s_ du du s_ s_ du s_ du s_
(18.260)
and noting that
dgac @gac dxb
= = (@b gac ) x_ b (18.261)
du @xb du
we …nd
d gac x_ a 1 s•
= (@b gac ) x_ a x_ b + gac x
•a gac x_ a (18.262)
du s_ s_ s_
Now substituting Eq. (18.262) into Eq. (18.259), we …nd
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_
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
(@b gac ) x_ a x_ b = (@a gbc ) x_ b x_ a (18.265)
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_
which simpli…es into
1 s•
•d + g dc [@b gac + @a gbc
x @a gab ] x_ a x_ b = x_ d : (18.271)
2 s_
Using the expression for the a¢ ne connection in terms of the metric
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
we see that these equations are equivalent to one another.

Chapter 19
Tensor Calculus on
manifolds
19.1 Tensors …elds and rank of a tensor

In order to understand what a tensor is and what is its rank is, it is important
to have a better understanding of a vector …eld, ~v ; on a manifold. How do we
de…ne a vector …eld on a manifold. We have learned that a vector …eld at a
given point, P , on a manifold is de…ned by the tangent plane, TP ; at that point
on the manifold. This tangent plane is de…ned by the tangent vector, ~t. The
tangent plane is de…ned by the basis vectors, eâ . We can denote the number
produced by the action of the vector, ~t on the vector, ~v ; (the component of the
vector …eld, ~v ; on the tangent space) by the scalar product
~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)
Homework: Show that 2nd rank tensor is linear.

Components of a tensor : We recall that the tangent plane is de…ned in terms
of the basis (~ea ) or dual basis vectors (~eb ). When vectors are expressed in terms
of basis or dual basis vectors we can determine the components of a tensor in
di¤erent forms. But …rst let’s consider if the vectors, ~v and ~u; are just the basis
or the dual basis vectors. In this case we have
(a) 1st rank tensor
t (~ea ) = ~t ~ea = ta ; (19.3)

b
t ~e = ~t ~eb = tb : (19.4)
(b) 2nd rank tensor
t ~ea ; ~eb = tab ; t (~ea ; ~eb ) = tab (19.5)

b
t ~ea ; ~e = tba ; a
t (~e ; ~eb ) = tab (19.6)
Example 4.1 Let’s reconsider the 2D sphere in the 3D manifoldFor the vector
on the tangent plane shown in Fig. (…g4.000)
t (~ea ) = ~e1 + ~e2 or t (~ea ) = ~e1 + ~e2

19.1. TENSORS FIELDS AND RANK OF A TENSOR 551
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
Now we can apply the linearity of tensors to determine the components of

a tensor for the general case where we have two vectors, ~v and ~u; expressed in
terms of the covariant or controvariant components in terms of the basis and
dual basis vectors as
~u = ua~ea ; ~v = vb~eb ; (19.7)
a b
~u = u ~ea ; ~v = v ~eb : (19.8)
(a) 1st rank tensor
t (~u) = t (ua~ea ) = t (~ea ) ua = ta ua ; (19.9)
a a a
t (~u) = t (u ~ea ) = u t (~ea ) = ta u (19.10)
(b) 2nd rank tensor

t (~u; ~v ) = t ua~ea ; vb~eb = t ~ea ; ~eb ua vb = tba ua vb ; (19.11)
a b a b
t (~u; ~v ) = t ua~e ; v ~eb = t (~e ; ~eb ) ua v = tab ua v b ; (19.12)
t (~u; ~v ) = t ua~ea ; v b~eb = t (~ea ; ~eb ) ua v b = tab ua v b : (19.13)
Symmetries of a tensor: A tensor can be symmetric or antisymmetric.

For a second ranked tensor t (~u; ~v )
t (~v ; ~u) Antisymmetric

t (~u; ~v ) =
t (~v ; ~u) symmetric
Any tensor can be expressed as a sum of symmetric and antisymmetric tensor.

Again if we consider a second rank tensor with elements tab , we can express this
elements as
1 1
tab = (tab + tba ) + (tab tba ) : (19.14)
2 2
Introducing the notations for the symmetric part
1
t(ab) = (tab + tba ) (19.15)
2
and the antisymmetric part
1
t[ab] = (tab tba ) (19.16)
2
we can write
tab = t(ab) + t[ab] : (19.17)
N th ranked tensor : for an N th ranked tensor the symmetric and antisymmetric
covariant components of the tensor, ta1 a2 :::aN are given by
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)
Homework: Find the explicit form of t[ab](cd)

The metric tensor revisited: A good example of a 2nd rank tensor is the
metric tensor, g; that we already have seen. Here more generally we de…ne the
metric tensor as a linear map of two vectors ~u and ~v into the number that is
the inner product
g (~u; ~v ) = ~u ~v (19.26)
We can easily see that the metric tensor is symmetric tensor as
g (~u; ~v ) = g (~v ; ~u) = ~v ~u (19.27)
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
tabc = ~ea ~eb ~ec (19.30)
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 :
19.2 Mapping tensors into tensors

We have learned that tensor maps vectors into real numbers. We have seen that
1st rank tensor t (~u) maps the vector ~u into real numbers in the tangent space.
t (~u) = t (ua~ea ) = ua t (~ea ) = ta ua (19.36)
and the second rank tensor t (~u; ~v ) maps the vectors ~u and ~v into real numbers
in the tangent space
t (~u; ~v ) = t ua~ea ; v b~eb = t (~ea ; ~eb ) ua v b = tab ua v b :

19.3. ELEMENTARY TENSOR OPERATIONS 555
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)
so that if we express the vector, w;

~ in terms of its controvariant components
~ = wc~ec
w (19.38)
we …nd
~ = ~ea ~eb wc~ec = (~ea ~eb ~ec ) wc = tabc wc = sab :

t (~ea ; ~eb ; w) (19.39)
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
~ = ua~ea ~eb wc~ec = (~ea ~eb ~ec ) ua wc = tabc ua wc = vb ;

t (~u; ~eb ; w) (19.42)
where we …nd a 1st rank tensor, sb : This means the 3rd rank tensor mapped the
two vectors into 1st rank tensor (a vector).
19.3 Elementary tensor operations

Adding, subtracting, and multiplying by a scalar :
sab = s (~ea ; ~eb ) = t (~ea ; ~eb ) r (~ea ; ~eb ) = tab rab : (19.43)
Outer product: Consider two …rst rank tensors (two vectors) u (~

p) and v (~q) :
The outer product or (tensor product) of these two tensors, which is denoted
by u v is given by
u v (~p; ~q) = u (~
p) v (~q) ; (19.44)
which is a second rank tensor. Suppose (~
p; ~q) ! (~ea ; ~eb ) ; then
u v (~ea ; ~eb ) = u (~ea ) v (~eb ) = ua vb : (19.45)
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)
The tensor h is a 3-rd rank tensor. Using the basis and dual basis vectors we
may express the components as
t s (ea ; eb ; ec ) = t (ea ; eb ) s (ec ) = habc (19.47)

t s (ea ; eb ; ec ) = t (ea ; eb ) s (ec ) = habc (19.48)
a b
t s e ; e ; ec = t e ; e s (ec ) = hab
a b
c (19.49)
c
t s (ea ; eb ; e ) = t (ea ; eb ) s (ec ) = hcab (19.50)
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)
Sol: The components of this 2-nd rank tensor are given by
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
and using a matrix this is expressed as

2 1 1 3
p q p1 q 2 p1 q 3
u v (~ p; ~q) = 4 p2 q 1 p2 q 2 p3 q 3 5
p3 q 1 p3 q 2 p3 q 3
NB: from what we learned in all mathematical or physics courses up to this

point, what we know is that inner product of two 1-st rank tensor (two vectors)
is commutative. However, that generalization does not apply to higher ranked
tensor. So from now on we must keep in mind that inner product of tensors in
general is not commutative including the 1-st rank tensor for the reason described
in terms of tensor contraction. For example, if t is a 2-nd rank tensor and s is
a …rst rank tensor, the inner product
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)
The controvariant components of t can then be expressed as

c d
t = ea eb ec ; ed = a b: (19.56)
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 = tab (ea eb ) = tba (ea eb ) = tab ea eb (19.58)
is true for any ranked tensor.
19.4 Tensors and coordinate transformations

We recall the coordinates basis and the dual basis vectors, under the coordinate
transformation xa ! x0a ; are transformed as
@xa
e^0c = eâ : (19.59)
@x0c
and
@x0a c
e^0a =
e^ : (19.60)
@xc
1-st rank tensor: Suppose we have a 1-st rank tensor, t
t (ea ) = ta ea = ta ea : (19.61)
in the x0a coordinate system is given by

@x0a @x0a c
t0a = t0 (e0a ) = t (e c
) = t (19.62)
@xc @xc
c c
@x @x
t0a = t0 (e0a ) = 0a
t (ec ) = tc (19.63)
@x @x0a
2-nd rank tensor: Suppose we have a 2-nd rank tensor, t ea ; eb ; t ea ; eb ;

and t (ea ; eb ) in the x0a coordinate system is given by
@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 @x0c f

t (e0a ; e0b ; e0c ) = t (19.67)
@x0a @x0b @xf cd
19.5 Tensor equations and the quotient theorem

A tensor equation which holds in one coordinate system must hold in another
coordinate system. Suppose we have an equation that states two second rank
tensors, t and s are equal in the xa coordinate system. That means
tcd = scd (19.68)
Multiplying both sides of this equation 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
t0ab = s0ab : (19.70)
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
tensors (i.e.sab = tabc v c ). Under coordinate transformation xa ! x0a , these set of

elements, using the transformation relations,
@x0a c
t0a = t (19.71)
@xc
@xc
t0a = tc (19.72)
@x0a
in the new coordinate system are given by
@x0a d @xe @x0a @xe d @x0a @xe d f
s0a
b = s se = se = t0a 0c
bc v = t v : (19.73)
@xd @x0b @xd @x0b @xd @x0b ef
In the relation
@x0f c
v 0f =
v (19.74)
@xc
by switching the the primes with the none primes, we can write
@xf 0c
vf = v (19.75)
@x0c
Substituting this expression into Eq. (19.73), we …nd
@x0a @xe d @xf 0c @x0a @xe @xf d

t0a 0c
bc v = t v ) t0a t v 0c = 0 (19.76)
@xd @x0b ef @x0c bc
@xd @x0b @x0c ef
There follows that for an arbitrary vector components, v 0c ;
@x0a @xe @xf d
t0a
bc = t : (19.77)
@xd @x0b @x0c ef
We made a contraction to the set of elements tabc and still we found a tensor.
Therefore the set of element tabc must be the component of the 3-rd rank tensor.
That is the quotient theorem.
19.6 Covariant derivatives of a tensor

Suppose we consider the 1-st rank tensor v 0a in the x0a coordinates. The deriv-
ative of this tensor can be expressed as
@v 0a @v 0a @xc @xc @v 0a
0b
= c 0b
= (19.78)
@x @x @x @x0b @xc
Using the transformation for 1-st rank tensor, we have
@x0a d
v 0a = v (19.79)
@xd
so that
@v 0a @xc @ @x0a d @xc @x0a @v c @xc @ 2 x0a c
0b
= 0b c d
v = 0b d c
+ v (19.80)
@x @x @x @x @x @x @x @x0b @xc @xd
We recall the covariant derivative from chapter 3
cb v (19.81)
so that the gradient of the …rst rank tensor can be expressed as
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)
Using the product rule, we have
@c t = @c tab ea eb + tab (@c ea ) eb + tab ea (@c eb ) (19.85)
and recalling that

f
@c e^b = bc e
^f (19.86)
one can write
f
@c t = @c tab ea eb + tab f
ca e
^f eb + tab ea cb e
^f (19.87)
f
@c t = @c tab (ea eb ) + tab f
ca (^
ef eb ) + tab cb (ea e^f ) (19.88)
If we interchange the indices f and a in the second term and f by b in the third
terms, we have
@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 eâ 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 eâ = 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)
19.7 Intrinsic derivative

Like vectors (1-rank tensor), tensor of rank 2 or more can depend on a subman-
ifold instead of the entire manifold. For example, a given tensor t can depend
on a curve C on the manifold that de…ned by some parameter u. This curve
may be de…ne by the equation xa (u) : The tensor, t; expressed in terms of its
controvariant components, then in terms of this parameter that the curve, C;
depends on can be expressed as
t (u) = tab ea (u) eb (u) :
The intrinsic derivative will then be
dt (u) d ab
= t ea (u) eb (u)
du du
dtab dea (u) deb (u)
= ea (u) eb (u) + tab eb (u) + tab ea (u)
du du du
In terms of the coordinates xa (u) ; we can write
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
dt (u) Dtab Dtab a Dtab a

= ea (u) eb (u) = e (u) eb (u) = e (u) eb (u) :
du Du Du Du
(19.109)
Like vectors parallel-transported, we can say tensors are parallel-transported
when
Dtab
= 0: (19.110)
Du
Suppose we pretend that the tensor depends on the entire manifold instead of
submanifold de…ned by some curve C, we can write
dtab @tab dxc dxc

= c
= @c tab
du @x du du
so that the intrinsic derivative can be expressed as
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

Theo Phys I Math Methods PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Theo Phys I Math Methods PDF

Uploaded by

Copyright:

Available Formats

Theoretical Physics I-Mathematical Methods

Daniel Erenso and Victor Montemayor

February 25, 2019

I Mathematical Methods in Physics I 1

2 Complex Numbers, functions, and series 21

3 Vectors, Lines, and Planes 45

4 Matrix and determinants 55

4.6 Cramer’s Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5 Introduction to di¤erential calculus-I 85

6 Introduction to Di¤erential Calculus II 109

7 Integral calculus-scalar functions 163

7.5 Homework Assignment 18 . . . . . . . . . . . . . . . . . . . . . . 193

8 Vector Calculus 195

II Mathematical Methods in Physics II 241

10 Introduction to the Eigenvalue Problem 261

11 Special functions 279

12 Power Series Solutions to Di¤erential Equations 303

13 A partial di¤erential equation 359

14 Functions of Complex Variables 387

15 Laplace transform 433

16 Fourier series and transform 449

16.7 Applications of the Fourier Transform . . . . . . . . . . . . . . . 467

III Mathematical Methods in physics III 477

18 Vector Calculus on manifolds 515

19 Tensor Calculus on manifolds 549

Series and Convergence

1.1 Sequence and series

2; 4; 8; 16; 32; :::; (1.2)

Figure 1.1: Bacteria growth in diluted human blood sample.

Figure 1.2: A bouncing ball.

is known as geometric series. Using

Then for jrj < 1; we have

so that the geometric series becomes

In a similar way for x3 ; we can easily show that

which can be put in the form

1.2 Testing series for convergence

is a convergent or divergent series. Here we discuss some of the commonly used

the in…nite series is a divergent series.

is a convergent series, the in…nite series in Eq. (1.28) is a convergent series.

The series in Eq. (1.28) is convergent if this integral is …nite otherwise it is

Example 1.2 Consider the in…nite series given by

(a) Rewrite this series using standard summation notation.

so that the series can be expressed as

1.3 Series representations of real functions

exists for all n 0 and x 2 R domain, we can write

Taylor Series for some common functions:

f (x) = ln(1 + x) (1.45)

(b) Find the interval of convergence for this series.

so that one can easily see that

There follows that

This is the Maclaurin’s Series.

We note that at x = 1, the series becomes

shows that for x = 1, the series

is a divergent series. However, for x = 1 the series becomes an alternating

with an > 0 is a convergent series if an is monotonically decreas-

Therefore, the interval of convergence for the Maclaurin’s Series is 1<x 1:

Applying the series expansion

Figure 1.4: Two charged balls suspended from a massless string.

and the gravitational ( W ) forces are given by

F1y = Ty W = T sin mg = 0 ) T sin = mg (1.72)

and in the x direction

Dividing Eq. (1.72) by (1.73), we …nd

In terms of the length of the threads, L; and the separation distance, x,