0 views

Uploaded by Prateek_14

- Arfken Weber Math Selected Problem Solution
- PEO CO
- PDF
- Legendre Equation
- Locally Anisotropic Kinetic Processes and Thermodynamics in Curved spaces
- Power Series With Solution
- A Harmonic Sequence
- Matlab-Signals and Systems
- Iq 3514961502
- 02 Matlab Tutorial
- Infinite Series
- Metric
- 7 Fourier Series
- Dorrie 16
- syb
- Maximal generalization of Lanczos' derivative using one-dimensional integrals
- Schw en Maple 13
- qf08
- excalc
- taylor

You are on page 1of 575

ii

Contents

Introduction xi

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

1.5 Homework Assignment 2 . . . . . . . . . . . . . . . . . . . . . . . 19

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

2.5 Homework Assignment 3 . . . . . . . . . . . . . . . . . . . . . . . 40

2.6 Homework Assignment 4 . . . . . . . . . . . . . . . . . . . . . . . 42

2.7 Homework Assignment 5 . . . . . . . . . . . . . . . . . . . . . . . 43

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

3.5 Homework Assignment 6 . . . . . . . . . . . . . . . . . . . . . . . 53

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

4.7 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

4.11 Homework Assignment 7 . . . . . . . . . . . . . . . . . . . . . . . 80

4.12 Homework Assignment 8 . . . . . . . . . . . . . . . . . . . . . . . 81

4.13 Homework Assignment 9 . . . . . . . . . . . . . . . . . . . . . . . 83

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

5.9 Homework Assignment 11 . . . . . . . . . . . . . . . . . . . . . . 108

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

6.8 Homework Assignment 12 . . . . . . . . . . . . . . . . . . . . . . 159

6.9 Homework Assignment 13 . . . . . . . . . . . . . . . . . . . . . . 159

6.10 Homework Assignment 14 . . . . . . . . . . . . . . . . . . . . . . 160

6.11 Homework Assignment 15 . . . . . . . . . . . . . . . . . . . . . . 161

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

7.3 Homework Assignment 16 . . . . . . . . . . . . . . . . . . . . . . 191

7.4 Homework Assignment 17 . . . . . . . . . . . . . . . . . . . . . . 192

CONTENTS v

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

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

10.1 Matrix Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261

10.2 Orthogonal matrices and the rotational operator . . . . . . . . . 264

10.3 Eigenvalues and Eigenvectors . . . . . . . . . . . . . . . . . . . . 265

10.4 Physical applications . . . . . . . . . . . . . . . . . . . . . . . . . 270

11.1 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

12.1 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

13.1 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

14.1 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

15.1 Inverse Laplace Transform . . . . . . . . . . . . . . . . . . . . . . 438

15.2 Applications of Laplace transforms . . . . . . . . . . . . . . . . . 440

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

16.8 Fourier Transforms and Convolution . . . . . . . . . . . . . . . . 473

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

18.1 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

19.1 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

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.

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

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

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

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)

1

X 1

rk = (1.11)

1 r

k=0

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

6 CHAPTER 1. SERIES AND CONVERGENCE

Figure 1.3: The motion of the bird and the two trains.

Following a similar procedure and referring to Fig. 1.3, the distance the

bird traveled when it ‡ies back to the …rst train, x2 ; can be expressed as

x2 = D1 v0 t 2 (1.17)

so that using

x2 x2

vo = ) t2 = (1.18)

t2 vo

we …nd

x2

x2 = D1 v0 t2 = D1 ) x2 = D1 : (1.19)

1+

Substituting Eq. (1.16) into Eq. (1.19), one …nds

1

x2 = D0 : (1.20)

1+ +1

x3 = D2 : (1.21)

1+

Using

2

1 1

D2 = D1 = D0 (1.22)

+1 +1

1.2. TESTING SERIES FOR CONVERGENCE 7

we …nd

2

1

x3 = D0 : (1.23)

1+ +1

Then the total distance the bird ‡ies before the two trains crash can be

expressed as an in…nite series,

1

X = x1 + x2 + x3 + ::: = D0 + D0

1+ 1+ +1

2

1

+ D0 :::

1+ +1

" #

0 1 2

1 1 1

)X= D0 + + + ::: ; (1.24)

1+ +1 +1 +1

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.

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

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

8 CHAPTER 1. SERIES AND CONVERGENCE

(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

(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)

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.

1 2 6 24

1+ + + + ; :::; (1.35)

6 120 5040 362; 880

1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 9

(b) Use the ratio test to check the convergence of this series.

Solution:

(a) We …rst rewrite the series using some general expression in terms of n

that generates all the terms in the series. In most cases we look for some

kind of relation that can be expressed using factorial and/or exponential

function. To this end, we note that

1 2 6 1 1 2 1 3 2 4 3 2

1+ + + + ::: = + + + + :::

6 120 5040 0! 3! 5! 7! 9!

0! 1! 2! 3! 4! n!

= + + + + :: :::; (1.36)

1! 3! 5! 7! 9! (2n + 1)!

X1

1 2 6 24

1+ + + + ; ::: = an ; (1.37)

6 120 5040 362; 880 n=0

where

n!

an = : (1.38)

(2n + 1)!

(b) We use the ratio test to determine if the series is convergent or divergent.

Using the result determined in (a), Eqs. (?? and (??, the ratio, ;

(n+1)! (n+1)(n!)

an+1 (2(n+1)+1)! (2n+3)(2n+2)(2n+1)!

= lim = lim n!

= lim n!

n!1 an n!1

(2n+1)!

n!1

(2n+1)!

n+1 n 1

= lim ) ' lim = lim =0

n!1 (2n + 3) (2n + 2) n!1 4n2 n!1 4n

which shows, < 1; and the series is a convergent series according to Eq.

(1.34).

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).

10 CHAPTER 1. SERIES AND 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

1

X 1 dn f (x)

f (x) = bn (x a)n , where, bn = (1.42)

n=0

n! dxn x=a

1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 11

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

is expressible as

X1

( 1)n+1 n

ln(1 + x) = x : (1.46)

n=1

n

12 CHAPTER 1. SERIES AND CONVERGENCE

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)

(

( 1)n+1

bn = n ; for n = 1; 2; 3::: (1.53)

0 n=0

1

X X1

( 1)n+1 n

ln(1 + x) = bn xn = x : (1.54)

n=0 n=1

n

1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 13

(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)

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

X1 X1

( 1)n+1 n 1

x = (1.61)

n=1

n n=1

n

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

ing.

14 CHAPTER 1. SERIES AND CONVERGENCE

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)

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

1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 15

x. Show that, for small charges Q and large lengths L, the equilibrium

position is given by

1=3

Q2 L

x= : (1.71)

2 0 mg

Solution: If the balls are at equilibrium the net force acting on each ball must

be zero. There are three forces acting on each of the balls. These are

the repulsive electrostatic force ( Fe ), the tension force ( T ), and earth’s

gravitational force ( W ) [Fig. 1.4].The magnitude of the electrical ( Fe )

1 Q2

Fe = and W = mg;

4 0 x2

respectively. Since the balls are at equilibrium, the net force on each balls

must add up to zero. Considering the ball on the right side in Fig. 1.4,

for the net for in the y-direction, we have

1 Q2 1 Q2

F1x = Fe Ty = T cos ) T cos = (1.73)

4 0 x2 4 0 x2

2

sin 4 0 mgx

tan = = 2

: (1.74)

cos Q

16 CHAPTER 1. SERIES AND CONVERGENCE

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

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

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

1.3. SERIES REPRESENTATIONS OF REAL FUNCTIONS 17

so that

1=3

Q2 x Q2 Q2 L

sin ' 2

) ' 2

)x' : (1.86)

4 0 mgx 2L 4 0 mgx 2 0 mg

18 CHAPTER 1. SERIES AND CONVERGENCE

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.

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!

1

X n

(x 1)

:

n=0

2n

5. Prove that

n

X n (n + 1)

i=

i=1

2

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:

(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.

(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

2 4 6

+ ::: = 1:

3! 5! 7!

20 CHAPTER 1. SERIES AND CONVERGENCE

Chapter 2

Complex Numbers,

functions, and series

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)

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

The Cartesian and Polar coordinates, as shown in Fig. 2.1, are related by the

equations

x = r cos ; y = r sin ; (2.7)

x + iy:

relates an exponential function and trigonometric functions, is given by

complex function. Using Euler’s equation one can also represent the complex

number in Eq.(2.6); in an exponential form as

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

plex number, z = x + iy; is denoted by z and de…ned as

z =x iy (2.11)

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,

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:

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

24 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

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 )

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.

2.1. COMPLEX NUMBERS 25

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)

(5 + i) x 5yi = 4 + 3i (2.32)

Solution: To determine the values of the real quantities x and y; we …rst need

to write the left hand side in the form z = a + ib

Two complex number z1 and z2 are equal if and only if their corresponding

Re and Im parts are equal. Thus

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

26 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

complex for some media such as metals and should be expressed, generally,

as

n = nR + inI ; (2.36)

where both nR and nI are real. Note that Eq. (2.36) shows the wave ~k in

Eq. (2.35) is also possibly complex and should be written as

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)

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 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)],

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)

!

k = (nR inI ) : (2.43)

c

2.1. COMPLEX NUMBERS 27

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)

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

28 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

This result shows that the intensity decays exponentially as the wave pen-

etrates inside the medium.

Convergence

Complex In…nite Series: An in…nite series

1

X

zn = z0 + z1 + z2 + z3 :::zn + ::: (2.52)

n=0

zn = an + ibn ; (2.53)

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

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.

power series

1

X

an xn = a0 + a1 + a2 + a3 :::an + ::: (2.56)

n=0

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

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!

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.

30 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

Example 2.7 Find and sketch the circle of convergence in the complex plane

for the following complex in…nite series:

1

X 1

X n

n (z 2 + i)

cn z = (2.67)

n=0 n=0

2n

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

(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 )

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)

2

[exp (i )] = exp (2i ) = cos (2 ) i sin (2 ) (2.85)

2 2 2

[exp (i )] = [cos ( ) i sin ( )] ) [exp (i )] = cos2 ( ) sin2 ( )

2i sin ( ) cos ( ) (2.86)

32 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

which leads to

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)

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

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

34 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

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

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)

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)

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 :

Z

I= sin ( x) cos ( x) dx (2.115)

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

36 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

which leads to

1 cos [( + ) x] cos [( ) x]

I= + (2.121)

2 +

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

2.4. ALGEBRAIC VS TRANSCENDENTAL FUNCTIONS 37

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

we have

(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 =

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

38 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

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)

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

2.4. ALGEBRAIC VS TRANSCENDENTAL FUNCTIONS 39

Example 2.16 Find the value for the following complex expression

2 !12 3

p

3+i

z = sin 1 4 p 5 (2.153)

3 i

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

40 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

1. For the complex number p i

z= 2e 4

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

2.5. HOMEWORK ASSIGNMENT 3 41

(a) Find the velocity (v) and acceleration (a) of the particle

dz d2 z

v= ;a = 2

dt dt

(b) Find the magnitude (absolute value) of the velocity and acceleration

(c) Describe the motion of the particle.

42 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

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

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.

2.7. HOMEWORK ASSIGNMENT 5 43

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 ),

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

44 CHAPTER 2. COMPLEX NUMBERS, FUNCTIONS, AND SERIES

1 1 1

= + :

Z12 Z1 Z2

(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

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

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 ~

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)

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.

hand rule is valid. For example in Cartesian coordinate system we have

^{ ^ |^

|^ = k; k^ = ^{; k^ ^{ = |^: (3.8)

~ ^

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

placed at a position ~r with respect to an axis (through the origin) is given

by ~ = ~r F~ .

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

Solution:

48 CHAPTER 3. VECTORS, LINES, AND PLANES

~ = ~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)

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)

= (3.16)

2

the magnitude of the torque would be

To …nd the direction use the right hand rule and you will …nd the direction

to be along the negative-z direction. Then the torque may be expressed

as

~ = 8N m^ z (3.18)

~ = ~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)

y (3.21)

3.3. EQUATIONS OF A STRAIGHT LINE 49

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

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

~ = 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)

50 CHAPTER 3. VECTORS, LINES, AND PLANES

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)

~r ~r0 = [x (t) 1] x

^ + [y (t) 2] y^ + [z (t) 3] z^ (3.34)

~v = 0^

x + (2m=s) y^ + (1m=s) z^: (3.35)

This means

^ + [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

3.4. EQUATION OF A PLANE 51

Then

(~r0 + ~v t0 ) ~v t0 = 0

0 02

(~r0 ~v ) t + (~v ~v ) t = 0 (3.44)

Noting that

2 2

~v ~v = (2m=s) + (1m=s) = 5m=s (3.45)

~r0 ~v = r0x vx + r0y vy + r0z vz = 2 2+3 1=7

Therefore

(~r0 ~v ) t0 + (~v ~v ) t02 = 0 (3.46)

becomes

7

7t0 + 5t02 = 0 ) t0 = 0; t0 = (3.47)

5

The acceptable value is t0 = 0. Thus the shortest distance from the origin

is given by the magnitude of the vector

p p

~r0 = ~r0 + ~v t0 = ~r0 ) r0 = 12 + 22 + 32 = 14 (3.48)

Example 4 Find the shortest distance from the origin to the plane

3x 2y 6z = 7 (3.49)

Solution: Comparing

3x 2y 6z = 7 (3.50)

to the equation of a plane

ax + by + cz = d (3.51)

we have

~ = 3^

N x 2^

y 6^

z (3.52)

which is a vector normal to the plane. The unit vector along normal vector

is given by

~

^ = N = q 3^

N

x 2^ y 6^z ^ = 3x

)N ^

2

y^

6

z^ (3.53)

~

N 2 2 7 7 7

32 + ( 2) + ( 6)

We now pick any point on the plane. This point must satisfy the equation

3x 2y 6z = 7: (3.54)

52 CHAPTER 3. VECTORS, LINES, AND PLANES

~r = 3^

x + 1^

y (3.55)

Then the distance from the origin to the plane is the component of this

~ ; because the distance

vector along the vector perpendicular to the plane, N

is measured by the length of the normal to the plane. We can write this

distance as

d = ~r N^ = 3 3 2 1 ) d = 1units (3.56)

7 7

Example 5 Find the symmetric and parametric equations of the line through

(0; 2; 4) and (3; 2; 1).

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).

3.5. HOMEWORK ASSIGNMENT 6 53

Solutions: We …rst construct two vectors A ~ and B~ that lie on the plane formed

by these three points. Consider the vector pointing from (0; 1; 1) to (2; 1; 3)

and the vector from (0; 1; 1) to (4; 2; 1) and let denote these vectors by A ~

~

and B, respectively. The we can write

~ = (2

A 0) x

^ + (1 1) y^ + (3 1) z^ = 2^

x + 2^

z (3.64)

~ = (4

B 0) x

^ + (4 1) y^ + (1 1) z^ = 4^

x + 3^

y

The vector, N ~ , that is normal to the plane formed by these two vectors

(i.e. the three points) can be determined using the cross product of the

two vectors

x

^ y^ z^

~ =A

~ ~ = ~ =x 2 2 0 2 0 2

N B 0 2 2 )N ^ y^ + z^

3 0 4 0 4 3

4 3 0

(3.65)

~ =

N 6^

x + 8^

y 8^

z (3.66)

We recall that equation of a plane is given by

a (x x0 ) + b (y y0 ) + c (z z0 ) = 0 (3.67)

where

~ = a^

N x + b^

y + c^

z (3.68)

and

~r0 = x0 x

^ + y0 y^ z0 z^ (3.69)

is a point on the plane as measured from the origin. Then we can write

the equation of the plane formed by the three points as

6 (x x0 ) + 8 (y y0 ) 8 (z z0 ) = 0 (3.70)

where

~r0 = x0 x

^ + y0 y^ z0 z^ (3.71)

is given by ~r0 = (0; 1; 1), (2; 1; 3), or (4; 2; 1):

~=

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.

54 CHAPTER 3. VECTORS, LINES, AND PLANES

~ 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?

~ = 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

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

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

IA = AI = A (4.4)

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

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

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

row i of A times column j of B. In index notation

n

X

(ab)ij = aik bkj ; (4.10)

k=1

Commutativities: For any two multiplyable matrices C and D;

CD 6= DC (4.11)

…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)

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 :

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

58 CHAPTER 4. MATRIX AND DETERMINANTS

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

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:

(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.

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.

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.

60 CHAPTER 4. MATRIX AND DETERMINANTS

(c) If (rank M )=(rank A)= R < n, then R unknowns can be found in terms of

the remaining n R unknowns.

Example 2 Consider the circuit shown. The circuit element values are R =

1:0W and V = 2:0V .

(a) Applying Kirchho¤’s voltage and current rules …nd set of linear equations.

(b) Find the coe¢ cient and augmented matrices for the set of linear equations.

(c) Find the values of the three branch currents I1 ; I2 ; and I3 using Gaussian

elimination method.

(d) Find the rank of the coe¢ cient and augmented matrices. Is the ranks

consistent with what we discussed earlier?

Kirchho¤ ’s Voltage Law : Start at one point in a circuit and go around any

closed loop in the circuit. The voltage at the starting point of the loop must be

the same as the voltage at the ending point, since they are the same point in

the circuit. Another way of stating this is simply that the net change in voltage

as you go around any closed loop in the circuit must be zero (so the starting

potential must be the same as the ending potential). This is just a restatement

of the conservation of energy.

Kirchho¤ ’s Current Law :The net current ‡owing into any node of a circuit

must equal the net current ‡owing out of that same node. This law says that

there is no place from which charge magically appears in the circuit, and no

place that the charge disappears. There is a net conservation of charge in an

electric circuit.

Solution:

(a) Applying Kirchho¤’s voltage law for the closed loop abef a we can write

V

V +I1 R+2I2 R 2 V +I1 R = 0 ) 2I1 R+2I2 R = V ) I1 +I2 =

2R

(4.25)

and for closed loop bcdeb

I3 R 2 V + I3 R + 2 V 2I2 R = 0 ) 2I3 R 2I2 R = 0 I2 + I3 = 0:

(4.26)

4.4. MATRIX APPLICATION: SOLVING A SET OF LINEAR EQUATIONS61

I1 = I2 + I3 ) I1 I2 I3 = 0 (4.27)

I1 + I2 = 1 (4.28)

I2 + I3 = 0

I1 I2 I3 = 0

or

I1 + I2 + 0I3 = 1 (4.29)

0I1 I2 + I3 = 0

I1 I2 I3 = 0

M I = K; (4.30)

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

2 3

1 1 0 1

A=4 0 1 1 0 5 (4.35)

0 2 1 1

62 CHAPTER 4. MATRIX AND DETERMINANTS

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

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

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

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

64 CHAPTER 4. MATRIX AND DETERMINANTS

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

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)

0 1

2 2 0

A=@ 0 2 2 A (4.67)

1 1 1

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

66 CHAPTER 4. MATRIX AND DETERMINANTS

2 2 0 2

jAj = 2 2 (4.70)

1 1 1 1

) jAj = 12 (4.71)

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)

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.)

2 3

1 1 0

M =4 0 1 1 5 (4.74)

1 1 1

2 3

1 1 0 1

A=4 0 1 1 0 5: (4.75)

1 1 1 0

2 3

1 1 0

M1 = 4 0 1 1 5; (4.76)

0 1 1

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

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

Adjoint of a Matrix : The adjoint of a square matrix, A, is given by

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.

2 3

a11 a12 a13

A = 4 a21 a22 a23 5 (4.87)

a31 a32 a33

68 CHAPTER 4. MATRIX AND DETERMINANTS

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

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

b. The adjoint matrix approach

2 3

1 2 3 1 0 0

4 2 0 4 0 1 0 5 (4.92)

1 1 1 0 0 1

2 3

1 0 0 a11 a12 a13

4 0 1 0 a21 a22 a23 5 (4.93)

0 0 1 a31 a32 a33

0 1

a11 a12 a13

A 1 = @ a21 a22 a23 A : (4.94)

a31 a32 a33

4.7. THE ADJOINT AND INVERSE OF A MATRIX 69

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

2 3

1 2 3 1 0 0

4 2 1 0 1 1 1 5 (4.96)

0 1 4 1 0 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

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

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

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

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

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

70 CHAPTER 4. MATRIX AND DETERMINANTS

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

subtract row 3 from row 1

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

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

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

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

4.7. THE ADJOINT AND INVERSE OF A MATRIX 71

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

T

1 [cof (A)]

A = (4.122)

det jAj

72 CHAPTER 4. MATRIX AND DETERMINANTS

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)

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)

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.

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

0

y = x sin ( ) + y cos ( )

z0 = z

4.8. ORTHOGONAL MATRICES AND THE ROTATION MATRIX 73

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

0 1

cos ( ) sin ( ) 0

R = @ sin ( ) cos ( ) 0 A (4.127)

0 0 1

~ (I used Mathematica)

The inverse of the matrix R

0 1

cos ( ) sin ( ) 0

RT = @ sin ( ) cos ( ) 0 A : (4.128)

0 0 1

Therefore, since

1

R = RT (4.129)

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)

74 CHAPTER 4. MATRIX AND DETERMINANTS

Solution: We need to check the two conditions stated above. First condition:

2

O(aA) = (aA) + 3; (4.133)

2 2

aO(A) = a A + 3 = aA + 3a ) O(aA) 6= aO(A)

You can concluded the function is not linear because if one of the condi-

tions is violated then the function is not linear. But let’s check also the

second condition,

2

O(A+B) = (A + B) +3 ) O(A+B) = A2 +AB +BA+B 2 +3 (4.134)

which shows

O(A + B) 6= O(A) + O(B) (4.136)

A 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.

the remaining n m unknowns.

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

and

I1 + I2 = 1; I2 + I3 = 0; I1 + I3 = 1; (4.140)

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

0 1

1 1 0 1

A=@ 0 1 1 0 A (4.142)

1 0 1 1

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

76 CHAPTER 4. MATRIX AND DETERMINANTS

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.

the three equations we obtained in the example above

~ = (1; 1; 0) ; B

A ~ = (0; 1; 0) ; C

~ = (1; 0; 1) (4.146)

~+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^:

~ 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^

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.

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.

~

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

~ 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.

78 CHAPTER 4. MATRIX AND DETERMINANTS

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)

~r1 = (9; 0; 7) ; ~r2 = (0; 9; 13) (4.160)

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)

(b) Subtract this component from A

~

B ~ A^ A^

B

^=

B : (4.163)

~

B ~ A^ A^

B

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)

(b) Subtract these components from C

~

C ~ A^ A^

C ~ B

C ^ B

^

C^ = : (4.165)

~

C ~ A^ A^

C ~ B

C ^ B

^

A^ and B^ it also has a unit magnitude.

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

(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) :

80 CHAPTER 4. MATRIX AND DETERMINANTS

The third orthonormal basis vector, following step 3(a) and (b), becomes

~

C ~ A^ A^

C ~ B

C ^ B

^

(0; 0; 3; 4) 3 4

C^ = = = 0; 0; ;

~

C ~ A^ A^

C ~ B

C ^ B

^ j(0; 0; 3; 4)j 5 5

(4.171)

Function Spaces and Linear Independence of Sets of Functions: If f1 (x) ; f2 (x) ; f3 (x) ;

:::fn (x) have derivatives of order n 1, and if the determinant

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)

= det

! cos (!t) 2! sin (!t) cos (!t)

= 2! sin2 (!t) cos (!t) ! sin2 (!t) cos (!t)

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:

class using row reduction of matrices

(a) The circuit diagram for this problem is shown in the …gure below

4.12. HOMEWORK ASSIGNMENT 8 81

(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.

82 CHAPTER 4. MATRIX AND DETERMINANTS

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

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

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.

4.13. HOMEWORK ASSIGNMENT 9 83

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

0 1

1 1 4 3

B 3 1 10 7 C

M =B

@ 4

C:

2 14 10 A

2 0 6 4

~ = (1; 4; 5); B

A ~ = (5; 2; 1); C

~ = (2; 1; 3); D

~ = (3; 6; 11)

~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

p p !

1+i 3 p3 (1 +

4 2 2p

i)

M= p

p3 (1 + i) 3+i

2 2 4

1

M = MT :

84 CHAPTER 4. MATRIX AND DETERMINANTS

Chapter 5

Introduction to di¤erential

calculus-I

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.

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

= 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

= lim = lim ...y is kept constant

@x x!0 x x!0 x

(5.2)

and

= 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 ; :::xn ) = lim

@xi xi !0 xi

...all x 6= xi are kept constant (5.4)

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

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

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.

following partial derivative

@z

: (5.8)

@r x

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

88 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

Example 2b An ideal gas is de…ned to be a gas that obeys the ideal gas equa-

tion of state

P V = nRT (5.11)

and

3

U= nRT (5.12)

2

where p is the pressure, V is the volume, n mass of the ideal gas mea-

sured in moles, T is the temperature, U is the internal energy, and R

the universal gas constant. Note that P; V; T; and U are the thermody-

namic variables. Obviously, as we can see from these two equations are

not independent set of variables. Find expressions for the following partial

di¤erentiations:

@P @P @P @2P @2P

; ; ; ; and : (5.13)

@V T @T V @V U @V 2 T @T 2 V

3

P V = nRT; U = nRT (5.14)

2

we may write

nRT 3 2U nR 2U 2U

P = ; U = nRT; T = )P = = (5.15)

V 2 3nR V 3nR 3V

so that

@P @P (V; T ) @ nRT nRT

= = = ; (5.16a)

@V T @V T @V V V2

@P @P (V; T ) @ nRT nR

= = = ; (5.16b)

@T V @V V @T V V

@P @P (V; U ) @ 2U 2U

= = = ; (5.16c)

@V U @V U @V 3V 3V 2

@P @P (U; V ) @ 2U 2

= = = ; (5.16d)

@U V @U V @U 3V 3V

= =

@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

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

dV:

Solution: We recall that volume of a cylinder is given by

V (R; H) = R2 H (5.24)

@V @V

dV = dR + dH (5.25)

@R @H

90 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

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 = 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)

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.

z = exp x2 y2 (5.37)

where

x(t) = et and y(t) = e t : (5.38)

Find

dz

: (5.39)

dt

Solution: Since x and y are a function of the same variable, t; we can write

dz @z dx @z dy

= + (5.40)

dt @x dt @y dt

Noting that

@z

= ( 2x) exp x2 y2

@x

@z

= ( 2y) exp x2 y2

@y

dx dy

= et = x; = et = y (5.41)

dt dt

we …nd

dz

= ( 2x) exp x2 y 2 x + ( 2y) exp x2 y 2 ( y)

dt

dz

) = 2 y 2 x2 exp x2 y 2 (5.42)

dt

Example 5 Show that the trajectory of the projectile motion of a ball with

mass m thrown from the ground (x0 = 0; y0 = 0) at an angle (measured

from the horizontal (+x-axis)) with a speed v0 is de…ned by equation of

hyperbola given by

y (x) = bx ax2 ; (5.43)

92 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

where

g

a= ; b = v0 tan ( ) : (5.44)

2v0 cos2 ( )

Note that x is the distance of the ball measured on the ground and y(x)

is the height of the ball from the ground.

Solution: From Newtons second law

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)

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

this can be put in the form

where

g

a= ; b = v0 tan ( ) : (5.55)

2v0 cos2 ( )

Example 6 Implicit Di¤ erentiation and application of the total di¤ erential :

Consider the equation

y 3 x2 y = 8 (5.56)

where y is a function of x. Evaluate dy=dx and d2 y=dx2 at the point

(x; y) = (3; 1), where y = y (x; y)

dy d2 y

and (5.57)

dx dx2

is to apply implicit di¤erentiation. To this end, let

z (x; y) = y 3 x2 y = 8 (5.58)

so that

dz @z @z dy dz dy

= + =0) = 2xy + 3y 2 x2 =0

dx @x @y dx dx dx

dy 2xy

) = 2 (5.59)

dx 3y x2

dy

= 1: (5.60)

dx

To …nd the second derivative of the function we use the result for the …rst

derivative. Noting that

dy 2xy

= 2 = f (x; y) (5.61)

dx 3y x2

we can write

d2 y @ @ dy

2

= f (x; y) + f (x; y) : (5.62)

dx @x @y dx

Upon carrying out the partial di¤erentiations, we have

@ 2y 4x2 y

[f (x; y)] = 2 + 2 (5.63)

@x 3y x2 (3y 2 x2 )

94 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

and

@ 2x 12xy 2

[f (x; y)] = 2 2; (5.64)

@y 3y x2 (3y 2 x2 )

so that Eq. (5.62) becomes

!

d2 y 2y 4x2 y 2x 12xy 2 dy

= + 2 + 2 : (5.65)

dx2 3y 2 x2 (3y 2 x2 ) 3y 2 x2 (3y 2 x2 ) dx

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

" #

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

dy

=1 (5.68)

dx

we …nd from Eq. (5.122)

d2 y 1 8

= 1 1 1= : (5.69)

dx2 3 3

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.

Solution: As shown in Fig. 5.1, the length, width, and height of the aquarium

are l; w; and h, respectively. The the volume, which is …xed, is related to

l; w; and h by

V (l; w; h) = lwh = V0 : (5.73)

The aquarium has rectangular sides made of glass. The surface area,

A (l; w; h) of the …ve sides (excluding the top side which is open) is given

by

A (l; w; h) = lw + 2 (lh + wh) : (5.74)

We are interested in constructing an aquarium with least amount of glass.

It means the area of the glass sides that the aquarium is constructed from

must be minimized while the volume remains constant. As we can see

96 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

from Eq. (5.1) l; w; and h are not independent variables. We can reduce

the area from a function of three variable to a three variable by expressing

h in terms of l and w using Eq. (5.1) as

V0

h= : (5.75)

lw

Substituting Eq. (5.75) into Eq. (5.74) we can then express the area in

terms of two independent variables

1 1

A (l; w) = lw + 2V0 + : (5.76)

l w

For the minimum area we must have

@A (l; w)

=0 (5.77)

@l

and

@A (l; w)

=0 (5.78)

@w

so that using Eq. (??), one …nds

2V0

w =0 (5.79)

l2

and

2V0

l =0 (5.80)

w2

respectively. Now combining these two equations, we …nd

w4 w3 1=3

w 2V0 =0)w 1 = 0 ) w = 0; w = (2V0 ) (5.81)

4V02 2V0

But the acceptable solution is

1=3

w = (2V0 ) : (5.82)

Substituting combining this result and Eqs. (??) and (5.1) the correspond-

ing length and height of the aquarium are

2V0 2V0 1=3

l = = = (2V0 ) (5.83)

w2 (2V0 )

2=3

1=3

V0 V0

h = 1=3 1=3

= (5.84)

(2V0 ) (2V0 ) 4

Thus the proportions can be expressed as

l:w:h 1

1=3

=1:1:

(2V0 ) 8

These are the dimension of the Aquarium for least amount of glass for

construction and a volume V0 .

5.5. THE METHOD OF LAGRANGIAN MULTIPLIERS 97

In the Example 7 the volume of the rectangular aquarium, which is a function

of three variables, is speci…ed,

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)

df = 0 ) dx + dy + dz = 0 (5.87)

@x @y @z

and the necessary and su¢ cient condition is

= = = 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 @ @f @ @f @

) + dx + + dy + + dz = 0

@x @x @y @y @z @z

98 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

@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

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

along with the constraint equation

(x; y) = constant

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)

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

(5.101)

100 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

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)

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

1=3

V0 1=3

4h3 = V0 ) h = ) w = l = 2h = (2V0 ) (5.109)

4

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

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.

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.

102 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

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)

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

5.6. CHANGE OF VARIABLES 103

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

In the next example we will see how we make change of variables for the two

dimensional Laplace equation in Cartesian to Cylindrical coordinates.

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

Solution: There are two di¤erent way of transforming this equation. I will use

the short way. We note that

@V (r; ) @V (x; y) @x @V (x; y) @y

= + (5.134)

@r @x @r @y @r

so that using expressions for x and y in Eq. (5.133), one …nds

@V (r; ) @V (x; y) @V (x; y)

= cos ( ) + sin ( ) : (5.135)

@r @x @y

104 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

= +

@ @x @ @y @

@V (x; y) @V (x; y)

= r sin ( ) + r cos ( ) (5.136)

@x @y

= sin ( ) + cos ( ) : (5.137)

r @ @x @y

…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

= 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

= sin ( ) + : (5.143)

@y @r r @

@ @ cos ( ) @

= sin ( ) + ; (5.144)

@y @r r @

@ @ sin ( ) @

= cos ( ) (5.145)

@x @r r @

5.6. CHANGE OF VARIABLES 105

@2 @ cos ( ) @ @ cos ( ) @

= sin ( ) + sin ( ) + (5.146)

@y 2 @r r @ @r r @

and

@2 @ sin ( ) @ @ sin ( ) @

= cos ( ) cos ( ) : (5.147)

@x2 @r r @ @r r @

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 @ @

…nds

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)

= sin2 ( ) +

@y 2 @r 2 r @r@ r2 @

2 2 2

cos ( ) @ cos ( ) @

+ + : (5.150)

r @r r2 @ 2

@2 @2 @ 1 @

= cos2 ( ) sin ( ) cos ( )

@x2 @r2 @r r @

sin ( ) @ @ sin ( ) @ @

cos ( ) + sin ( ) (5.151)

r @ @r r2 @ @

that leads to

= cos2 ( ) +

@x2 @r 2 r2 @ r @r@

2 2

sin ( ) @ sin ( ) @ 2

+ + : (5.152)

r @r r2 @ 2

106 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

@2 @2 @2 1 @ 1 @2

+ = + + (5.153)

@x2 @y 2 @r2 r @r r2 @ 2

which can be rewritten as

@2 @2 1 @ @ 1 @2

+ 2 = r + (5.154)

@x2 @y r @r @r r2 @ 2

Therefore the Laplace equation in polar coordinates can be expressed as

1 @ @V (r; ) 1 @ 2 V (r; )

r + =0 (5.155)

r @r @r r2 @ 2

Legendre Transformation of a function f (x) denoted by g (p) is given by

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:

W = work done on the system; Q = heat energy into the system;

dU = dQ + dW = T dS P dV (5.159)

F = Helmholtz F ree Energy; F = F (T; V )

5.8. HOMEWORK ASSIGNMENT 10 107

Thermodynamics by deriving the form of the thermodynamic potential

Helmholtz Free Energy such that F = F (T; V ).

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

dU = T dS P dV (5.161)

we …nd

@U

= T; (5.162)

@S V

@U

= P: (5.163)

@V S

written as

@U @U

F = F (T; V ) = F ;V =F ; (5.164)

@S V @S V

@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

108 CHAPTER 5. INTRODUCTION TO DIFFERENTIAL CALCULUS-I

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

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

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.

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

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.

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

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)

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.

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.

112 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

(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

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

6.1. LINEAR ORDINARY DIFFERENTIAL EQUATIONS 113

where

g g

b0 = 0; b1 = 1; f (t) = C1 + exp ( kt)) (6.22)

k k

Applying Eq. (6.17) we can write the solution to Eq. (6.17) as

b0

t Z t

b0

t e b1 b0 0

y (t) = C2 e b1

+ f (t0 ) e b1 t dt0 ; (6.23)

b1 t0

Z th i

g g

y (t) = C2 + C1 + exp ( kt0 )) dt0 ; (6.24)

t0 k k

g g

y (t) = C2 + C1 + t + 2 [exp ( kt) 1] (6.25)

k k

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

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

114 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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

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.

(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:

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

6.1. LINEAR ORDINARY DIFFERENTIAL EQUATIONS 115

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

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

116 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

where

l2 B 2

a0 = ; a1 = 1; f (t) = 0

mR

We recall the solution to Eq. (6.45) is given by

a0

t Z t

a0

t e a1 a0 0

v (t) = C1 e a1

+ f (t0 ) e a1 t dt0 ; (6.46)

a1 t0

which gives

l2 B 2 t

a0

t

v (t) = C1 e a1

= C1 e mR (6.47)

Noting that initial velocity,v (t = 0) = v0 ; the velocity of the rod becomes

l2 B 2 t

v(t) = v0 e mR : (6.48)

(d) The energy delivered to the resistor per unit time (power) is given by

dW 2

P = = Iin (t) R (6.49)

dt

Using the results we found for the induced current in Eq. (6.42) along

with the velocity in Eq. (6.48), Eq. (6.42) can be put in the form

2l2 B 2 t

dW l2 v 2 (t) B 2 B 2 l2 v02

= R = e mR (6.50)

dt R2 R

or

dW

= f (t) : (6.51)

dt

Then the total energy delivered to the resistor is

Z1 2l2 B 2 t

B 2 l2 v02 B 2 l2 v02 mR

W = e mR dt ) W =

R R 2l2 B 2

0

1

) W = mvo2 (6.52)

2

Example 3 Discuss the variation in the intensity of radiation as it passes

through a stellar atmosphere assuming a plane-parallel model for the at-

mosphere.

Starting Point: Consider a di¤erential slab of stellar atmosphere of thickness

dz at the position z above the reference position z = 0. Let (z) be the

density of the atmosphere at that position, and let I(z) be the intensity

of radiation incident on the slab. Let Io be the intensity of radiation at

the reference position z = 0 within the atmosphere. Find an expression

for I(z) for z > 0.

6.2. THE FIRST-ORDER DE AND EXACT DIFFERENTIAL 117

of the atmosphere by

dI (z) = K (z)dz (6.53)

Z I(z) Z z Z z

dI (z) = K (z)dz ) I (z) = I0 K (z)dz (6.54)

I0 0 0

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

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.

118 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Solution: The total di¤erential, dF; for the function of two variables, F (x; y) ;

is given by

dF = P (x; y) dx + Q (x; y) dy: (6.60)

For the given function in Eq. (6.59), we have

@F @

P (x; y) = = 3x2 + 2x sin (2y) = 6x + 2 sin (2y) (6.61)

@x @x

and

@F @

Q (x; y) = = 3x2 + 2x sin (2y) = 4x cos (2y) (6.62)

@y @y

so that total di¤erential becomes

where

@F @F

P (x; y) = = 6x + 2 sin (2y) and Q (x; y) = = 4x cos (2y) : (6.65)

@x @y

Taking the partial derivative of the function P (x; y) with respect to y and the

function Q (x; y) with respect to x, we …nd

@P @F @

= = [6x + 2 sin (2y)] = 4 cos (2y) (6.66)

@y @y@x @y

@Q @F @

= = [4x cos (2y)] = 4 cos (2y) : (6.67)

@x @x@y @x

This shows that for the total di¤erential

we found

@P (x; y) @Q (x; y)

= : (6.69)

@y @x

Such kind of di¤erential is called an exact di¤ erential. If the di¤erential is an

exact di¤ erential, then there is a function F (x; y) such that

@F @F

P (x; y) = ; Q (x; y) =

@x @y

Example 5 For the …rst-order DE

dy

x3 + 2y = 3x (2 xy) ; (6.70)

dx

6.2. THE FIRST-ORDER DE AND EXACT DIFFERENTIAL 119

(a) Find the total di¤erential for a function of two variable F (x; y) =constant

in form

dF (x; y) = P (x; y) dx + Q (x; y) dy = 0 (6.71)

(b) Show that the total di¤erential is an exact di¤erential and …nd the function

F (x; y) =constant.

(c) Using the function F (x; y) =constant, …nd the solution to the given di¤er-

ential equation.

Solution:

(a) Multiplying the given …rst-order di¤erential equation by dx, we …nd

where

P (x; y) = 3x (2 xy) ; Q (x; y) = x3 + 2y : (6.74)

(b) Now taking the partial derivative of P with respect to y and Q with respect

to x, we …nd

@P @

= [3x (2 xy)] = 3x2 (6.75)

@y @y

and

@Q @

= x3 + 2y = 3x2 : (6.76)

@x @x

There follows that

@P @Q

= (6.77)

@y @x

Thus, the total di¤erential is an exact di¤ erential and there exists a func-

tion F (x; y) =constant, such that

@F @F

P = and Q = : (6.78)

@x @y

Now using

@F

P = = 3x (2 xy) (6.79)

@x

and integrating with respect to x

Z x Z x

@F (x0 ; y) 0

dx = 3x0 (2 x0 y) dx0 (6.80)

x0 @x0 x0

we …nd

F (x; y) F (x0 ; y) = 3x2 x3 y 3x20 + x30 y: (6.81)

120 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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)

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

6.2. THE FIRST-ORDER DE AND EXACT DIFFERENTIAL 121

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

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)

@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

122 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

@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.

ential

We recall that the di¤erential equation

@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

dy (x) a0 (x) f (x)

= y (x) (6.112)

dx a1 (x) a1 (x)

that can be put in the form

dy (x)

Q (x; y) = P (x; y) : (6.113)

dx

There follows the total deferential

where

a0 (x) f (x)

P (x; y) = y (x) ; Q (x; y) = 1: (6.115)

a1 (x) a1 (x)

Here we note that

@P (x; y) @ a0 (x) f (x) @Q (x; y)

= y (x) = y; = 0; (6.116)

@y @y a1 (x) a1 (x) @x

and

@P (x; y) @Q (x; y)

6= (6.117)

@y @x

that mean the resulting DE from the …rst-order ODE is not an exact di¤erential

and the method we developed in the previous section does not work. However,

this method can be used if we can …nd a function that can make the ODE in Eq.

(6.111) to lead to an exact di¤erential. Next we will determine this function

which is referred as "The integration factor".

The Integration Factor: The …rst-order ODE with none constant coe¢ cient

in Eq. (6.111) can put in the form

dy

+ P (x)y = Q(x); (6.118)

dx

where we introduced the functions de…ned by

a0 (x) f (x)

P (x) = ; Q(x) = : (6.119)

a1 (x) a1 (x)

This is the generic …rst-order ODE. Let’s say this ODE does not lead to an exact

di¤erential. Let the function V (x) be the function that when it multiplied to

the ODE in Eq. (6.118) leads to a total di¤erential that is an exact di¤erential.

Thus multiplying Eq. (6.118) by V (x) ; we have

dy

V (x) + V (x) P (x)y = V (x) Q(x)

dx

dy

) V (x) + V (x) P (x)y V (x) Q(x) = 0 (6.120)

dx

124 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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)

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

a0

x Z

a0

x e a1 a0

y (x) = Ce a1

+ f (x) e a1 x dx: (6.130)

a1

dy

+ P (x)y = Q(x); (6.131)

dx

where

a0 f (x)

P (x) = :Q (x) = : (6.132)

a1 a1

Substituting this functions into Eq. (6.128), we …nd

R a0 Z R a0 R

a1 dx

f (x) a0

y(x) = e e a1 dx dx + Ce a1 dx

: (6.133)

a1

that leads to a0

x Z

a0

x e a1 a0

y(x) = Ce a1

+ e a1 x f (x) dx: (6.134)

a1

which the same as Eq. (6.130).

dy

2x + 4x2 = y (6.135)

dx

dy

+ P (x)y = Q(x); (6.136)

dx

(b) Using the general solution for such kind of …rst-order DE (Eq. (6.118)),

…nd the solution for y (x) subject to the boundary condition,

26

y (x0 = 1) = : (6.137)

3

Solution:

dy 1 dy

+ 2x = y) + P (x) y = Q (x) ; (6.138)

dx 2x dx

we have

1

P (x) = ; Q (x) = 2x: (6.139)

2x

126 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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.

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

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

a0 + a1 k + a2 k 2 + a3 k 3 ::::an 1k

n 1

+ an k n Aekx = 0: (6.149)

a0 + a1 k + a2 k 2 + a3 k 3 ::::an 1k

n 1

+ an k n = 0: (6.150)

This equation is called the indicial equation. The solution to the indicial equa-

tion in Eq. (6.150) gives n roots that could be real or complex. If these roots

are none degenerate (i.e. no repeated identical roots) and are given by

k = k1 ; k2 :::kn ; (6.151)

the general solution to the HODE with constant coe¢ cients in Eq. (6.147) is

given by

y (x) = C1 ek1 x + C2 ek2 x + :::Cn ekn x (6.152)

As we have seen in the in the solutions to …rst-order ODE, the constants

C1 ; C2 ; :::Cn ; are determined from the boundary conditions pertinent to the

particular problem.

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).

Q (t) dI (t)

+L =0 (6.153)

C dt

128 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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!

6.4. HIGHER-ORDER ODE’S 129

Combining Eqs. (6.160) and (6.161), one can easily show that

1 I0 1 I0

A= Q0 + ;B = Q0 : (6.162)

2 i! 2 i!

Therefore, the charge at a given time t can be expressed as

1 I0 1 I0

Q(t) = Q0 + exp (i!t) + Q0 exp ( i!t) ; (6.163)

2 i! 2 i!

or

exp (i!t) + exp ( i!t) I0 exp (i!t) exp ( i!t)

Q(t) = Q0 + :

2 ! 2i

(6.164)

Applying Euler’s formula Eq. (6.164) can be put in the form

Q(t) = A cos (!t) + B sin (!t) ; (6.165)

where

I0

C = Q0 ; D = : (6.166)

!

Introducing the constants de…ned by

I0

Q0 = E sin '; = E cos '; (6.167)

!

Eq. (6.165) can also be expressed as

Q (t) = E [sin ' cos (!t) + cos ' sin (!t)] = E sin (!t + ') ; (6.168)

with s

2

Q0 I0

tan (') = ;E = Q20 + : (6.169)

I0 =! !

If we de…ne another constant de…ned in a slightly di¤erent way

I0

Q0 = G cos '; = G sin '; (6.170)

!

one can rewrite Eq. (6.165) as

Q(t) = G cos ' cos (!t) + G sin ' sin (!t) = G cos (!t ') ; (6.171)

where s

2

I0 =! I0

tan (') = ;G = Q20 + : (6.172)

Q0 !

The current

dQ

I (t) = (6.173)

dt

can be determined using any one of the equations for the charge in Eqs.

(6.164), (6.165), (6.168), or (6.171). If we chose Eq. (6.168), we …nd

I (t) = I0 cos (!t) Q0 ! sin (!t) (6.174)

for the current in the circuit.

130 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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)

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 + ') :

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:

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

6.4. HIGHER-ORDER ODE’S 131

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

= 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)

132 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

2

+ k2 = 0 ) = ik (6.187)

: (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)

we …nd the indicial equation

2

q2 = 0 ) = q (6.193)

(x) = A1 eqx + A2 e qx

: (6.194)

qx

(x) = A2 e : (6.195)

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

6.4. HIGHER-ORDER ODE’S 133

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

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

m1 1

; (6.203)

Pm 2 1 (x) = C00 + C10 x + C20 x2 ::: + 0

Cm2 1 xm2 1 ;

134 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Example 11 In each of the following parts, …nd the indicial equation, list the

roots of the indicial equation {ki }, and write out the general solution to

the DEs:

(a)

d3 y dy

3

+ = 0: (6.204)

dx dx

(b)

d4 s d3 s

2 3 = 0: (6.205)

dt4 dt3

Solution:

dy d3 y

y (x) = Aekx ) = ky (x) ) 3 = k 3 y (x) (6.206)

dx dx

into the given DE, the indicial equation becomes

k 3 + k = 0: (6.207)

k 3 + k = 0 ) k k 2 + 1 = 0 ) k1 = 0; k2 = i; k3 = i: (6.208)

The roots do not show any degeneracy in this case and the solution to the

DE can be written as

y (x) = A + Beix + Ce ix

: (6.209)

or

y (x) = A + D cos (x) + E sin (x) (6.210)

(b) Substituting

d3 s d4 s

s (t) = Aekt ; ) = k 3

s (t) ) = k 4 s (t) (6.211)

dt3 dt4

in the given DE, the roots to the indicial equation are found to be

3

2k 4 3k 3 = k 3 (2k 3) = 0 ) k1 = ; k2 = k3 = k4 = 0: (6.212)

2

Since the the indicial equation has three identical roots, the solution can

be written as

3

2k 4 3k 3 = 0 ) k 3 (2k 3) = 0 ) k1 = ; k2 = k3 = k4 = 0; (6.213)

2

6.4. HIGHER-ORDER ODE’S 135

This shows that the three roots (k2 ; k3 ; k4 ) are degenerate with degeneracy,

m = 3; and one none degenerate root (k1 ). Thus applying the relation in

Eq. (6.202), the general solution can written as

where a is the value for the degenerate roots and k1 is the value for the

none degenerate root. Using these values the general solution to the DE

becomes

3

y (x) = A + Bx + Cx2 + Ae 2 x : (6.215)

N.B. You may use Mathematica to check you results. Here is the result for

the previous two examples

Here C[1], C[2], C[3], and C[4] are constants of integration. note that -1 in

8

the …rst solution can be absorbed in C[2] and 27 in the second solution can also

be absorbed in C[1].

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.

136 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Solution: Since the mass initially was at equilibrium and also consider the dis-

placement from this equilibrium position, we do not consider the buoyant

and gravitational force on the mass. Therefore, the equation of motion for

the mass is given by Newton’s second law

d2 y

Fnet = may = m : (6.216)

dt2

The net force acting on the mass is given by

dy

Fnet = ky bvy = ky b (6.217)

dt

which leads to

d2 y dy d2 y dy

m = ky b ) 2 +2 + !2 y = 0 (6.218)

dt2 dt dt dt

where

k b

; !2 = = : (6.219)

m 2m

The indicial equation would then be

2

+2 + ! 2 = 0; (6.220)

1 = + D; 2 = D; (6.221)

where p

D= 2 !2 (6.222)

Case 1 ( = !) we have

p

D= 2 !2 = 0 (6.223)

1 = 2 = (6.224)

t

y(t) = (At + B) e (6.225)

Case 2 ( > !) we note that D is real

p

D= 2 !2 < (6.226)

( D)t ( +D)t

y(t) = Ae + Be : (6.227)

6.4. HIGHER-ORDER ODE’S 137

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)

( 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)

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

4-th Any boundary conditions given to the problem must then be applied to

the general solution y(x).

138 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

d2 s ds

3 + 2s = 16: (6.236)

dt2 dt

Solution:

Step 1 : Find the homogeneous solution. That means determine the solution

for

d2 s ds

2

3 + 2s = 0: (6.237)

dt dt

Noting that the indicial equation is

k2 3k + 2 = 0 ) k1 = 2; k2 = 1 (6.238)

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 4 : Apply the boundary conditions if there are any. Here we are not

given any.

Example 14 Find the general solution to the following di¤erential equation:

d2 R

+ 4R = sin (3x) (6.244)

dx2

Solution:

Step 1 : Find the homogeneous solution for

d2 R

+ 4R = 0: (6.245)

dx2

6.4. HIGHER-ORDER ODE’S 139

k 2 + 4 = 0 ) k1 = 2i; k2 = 2i (6.246)

(an educated guess) to the particular solution that has the form

Note that this guess will be valid because 3i is di¤ erent from the roots of

the indicial equation. See the next example.

Since in the particular solution we are not allowed to have arbitrary constant

of integration, we must determine C: Upon substituting Eq. (6.248) into

Eq. (6.244), this constant can be determined to be

9C sin (3x) + 4C sin (3x) = sin (3x) ) 5C sin (3x) = sin (3x)

1

) C= : (6.249)

5

Thus the particular solution becomes

1

Hp = sin (3x) : (6.250)

5

Step 3 : Write the general solution

1

R (x) = Rh (x) + Rp (x) ) R (x) = A cos (2x) + B sin (2x) sin (3x) :

5

(6.251)

Step 4 : Apply the boundary conditions if there are any. Here there are

no any.

Example 16 Find the general solution to the following di¤erential equation:

d2 y

+ 4y = sin (2x) (6.252)

dx2

Solution:

Step 1 : Find the solution for the homogeneous LDE

d2 y

+ 4y = 0: (6.253)

dx2

The roots to the indicial equation

k2 + 4 = 0 (6.254)

140 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

are

k1 = 2i; k2 = 2i: (6.255)

Then the homogeneous part of the solution can be expressed as

or preferably

yh (x) = A0 e2ix + B 0 e 2ix

: (6.257)

Step 2 : When the non-homogenous part f (x) is a trigonometric or hyper-

bolic function (i.e. sin (ax) ; cos (ax) ; sinh (ax) ; or cosh (ax)), before we

try to …nd the particular solution it is recommended to express the func-

tions as exponential function applying Euler’s formula. Thus for function

in Eq. (6.252), one should write

1 2ix 2ix

f (x) = sin (2x) = e e : (6.258)

2i

Now comparing Eqs. (6.257) and (6.258), we note that values to the ex-

ponents to the Homogeneous solution (the roots of the indicial equation)

are the same as the exponents to the function, f (x) : Under these circum-

stances, one can make an educated guess to the particular solution based

on the property of an exponential function and the order of the di¤eren-

tial equation. For example, for the DE in Eq. (6.252), if one guesses a

particular solution

yp (x) = C1 e2ix + C2 e 2ix

which has the same form as the homogeneous solution in Eq. (6.257), we

…nd

d2 yp

+ 4yp = 4yp + 4yp = 0;

dx2

as one would have guessed. Therefore, we must guess a function that must

have a di¤erent form the homogenous solution. In this case, our educated

guess be

yp (x) = x C1 e2ix + C2 e 2ix ; (6.259)

that leads to

dyp

= C1 e2ix + C2 e 2ix + 2ix C1 e2ix C2 e 2ix

dx

d 2 yp

) = 2i C1 e2ix C2 e 2ix + 2i C1 e2ix C2 e 2ix

dx2

4x C1 e2ix + C2 e 2ix (6.260)

2

d yp

) + 4yp = 4i C1 e2ix C2 e 2ix

: (6.261)

dx2

Substituting this result into Eq. (??), we …nd

1 2ix

4i C1 e2ix C2 e 2ix

= e e 2ix

:

2i

6.5. THE METHOD OF SUPERPOSITION AND THE PARTICULAR SOLUTION.141

1

C1 = C2 =

: (6.262)

8

Therefore, the particular solution becomes

1

yp (x) = xe2ix + xe 2ix

; (6.263)

8

that can be put in the form

x e2ix + e 2ix

x

yp (x) = = cos (2x) : (6.264)

4 2 4

x

y (x) = A cos (2x) + B sin (2x) cos (2x) (6.265)

4

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

where

yp1 (x) = Pn (x) e2ix ; yp2 (x) = Qn (x) e 2ix

: (6.267)

with Pn (x) = C1 x and Qn (x) = C2 x are polynomial functions with degree,

n = 1: To determine the particular solution, we also wrote the function, f (x) =

sin (2x) ; as a sum of two functions such that

d2 yp (x)

+ 4yp (x) = f (x) = f1 (x) + f2 (x) ; (6.268)

dx2

where

f1 (x) = Pm (x) e2ix ; f2 (x) = Qm (x) e 2ix

; (6.269)

1 1

with Pm (x) = 2iand Qm (x) = are polynomial of degree, m = 0: In a

2i

similar way, the particular solutions to

d2 yp0 1 (x) 0

d2 yp0 2 (x)

+ 4yp1 (x) = f 1 (x) and + 4yp0 2 (x) = f2 (x) ; (6.270)

dx2 dx2

can be guess to be

yp0 1 (x) = Pn0 (x) e2ix ; yp0 2 (x) = Q0n (x) e 2ix

: (6.271)

142 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

with Pn0 (x) = C10 x and Q0n (x) = C20 x are polynomial functions with degree, n =

1; respectively. Upon adding these two equations, we …nd a none homogeneous

ODE in Eq. (6.272) has the same form as Eq. (6.268)

d2 yp0

+ 4yp0 = f1 (x) + f2 (x) = f (x) ; (6.272)

dx2

where

yp0 (x) = 2yp0 1 (x) + 2yp0 2 (x) = Pn00 (x) e2ix + Q00n (x) e 2ix

; (6.273)

with Pn00 (x) = 2C10 x = C100 x and Q00n (x) = 2C20 x = C200 x are still the same

polynomials with the same degree, n = 1 obtained the previous example by

directly considering the function f (x). This can be generalized for an n-th

order none homogeneous ODE

X dn y

an = f (x) (6.274)

n

dxn

X dn y

an = f1 (x) + f2 (x) : (6.275)

n

dxn

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.

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)

d2 y

y = 2 sin (x) + 4x cos (x) (6.287)

dx2

as

d2 y eix e ix

eix + e ix

y=2 + 4x (6.288)

dx2 2i 2

d d

+1 1 = (2x i) eix + (2x + i) e ix

(6.289)

dx dx

or

d d

+1 1 = f1 (x) + f2 (x) ; (6.290)

dx dx

where

f1 (x) = (2x i) eix ; f2 (x) = (2x + i) e ix

; (6.291)

and

a = 1; b = 1 (6.292)

Then for the non-homogenous DE with the function f1 (x)

d d

+1 1 = f1 (x) = (2x i) eix (6.293)

dx dx

144 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

noting that

c = i 6= a; b (6.294)

we may write the particular solution as

that gives

dyp1

= Beix + i (A + Bx) eix = Beix + iyp1 (x) ; (6.296)

dx

d 2 y p1 dBe ix

dyp1

= +i = iBeix + i Beix + iyp1 (x)

dx2 dx dx

= 2iBeix yp1 (x) (6.297)

d2 yp1 (x)

yp1 (x) = f1 (x) = (2x i) eix : (6.298)

dx2

we …nd

= (2x i) eix (6.299)

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!

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

146 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

in the form

d2 d d d

a2 + a1 + a0 y = a b y = F (x) ; (6.312)

dx2 dx dx dx

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

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)

(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

6.6. THE METHOD OF SUCCESSIVE INTEGRATION 147

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)

(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

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).

homogeneous ODE (F (x) = 0)

d2 y dy

a2 2

+ a1 + a0 y = 0; (6.328)

dx dx

148 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

a2 k 2 + a1 k + a0 = 0 (6.329)

the roots are given by Eq. (6.313). When these roots are none degenerate

(a 6= b), we saw that the general solution to DE is given by

Z x

00

y (x) = C1 e(a b)x dx00 + C2 ebx ; (6.331)

y (x) = + C2 ebx = + C2 ebx

a b a b

= C1 eax + C2 ebx : (6.332)

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)

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

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

6.6. THE METHOD OF SUCCESSIVE INTEGRATION 149

one can write the function, F (x) ; that makes the ODE none homogeneous,

as

X n

F (x) = ecx Pn (x) = ecx Ai x i : (6.337)

n=0

Z " Z #

x n

X x00

00

yp (x) = ebx e(a b)x

Ai e(c a)x0 0i

x dx0 dx00 (6.338)

n=0

Z " Z #

n

X x x00

bx x00 x0 0i 0

= Ai e e e x dx dx00

n=0

where

=c a and =a b: (6.339)

Next we will evaluate Eq. (6.338) to derive the particular solutions given

in Eq. (6.335).

condition one can use the relation

di x0 x 0i

[e ]=e x ; (6.340)

d i

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

n

X

bx di e( + )x

yp (x) = e Ai : (6.343)

i=0

d i +

150 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Noting that

d0 e( + )x

e( + )x

= = C 0 e( + )x

;

d 0 + +

" #

d e( + )x

d 1 x

= 2 + e( + )x

d + d ( + ) +

= (C0 + C1 x) e( + )x ;

" #

d2 e( + )x

1 x

= 2 + e( + )x

d 2 + ( + ) +

" #

2 x x x2

= 3 2 2 + e( + )x

( + ) ( + ) ( + ) +

= C0 + C1 x + C2 x2 e( + )x

; (6.344)

one can easily show that

Xn Z

di x ( + )x00 00

Ai i e dx = C0 + C1 x + C2 x2 + :::Cn xn e( (6.345)

+ )x

i=0

d

= Qn (x) e( + )x

;

where Qn (x) is a polynomial of degree n. Note that the constants Ai are

absorbed in the constants Ci :Thus, substituting, = c a and = a b

in Eq. (6.345), the particular solution in Eq. (6.343) becomes

yp (x) = ebx Qn (x) e(c a+a b)x

= Qn (x) ecx : (6.346)

(ii.) When c = a & c 6= b: for this case we have = c a = 0 and Eq.

(6.338) becomes

n

X Z x" Z x00 #

bx x00

yp (x) = Ai e e x dx dx00 :

0i 0

(6.347)

n=0

Substituting this into Eq. (6.327), the particular solution can be expressed

as

Z

Ai bx x h x00 00 i+1 i 00

Xn

yp (x) = i+1

e e (x ) dx (6.348)

i=0

n+1

X Z x

Aj 1 x00 j

= ebx j

e (x00 ) dx00 ;

j=1

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)

6.6. THE METHOD OF SUCCESSIVE INTEGRATION 151

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

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.)

152 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Solution: In order to …nd the amplitude of oscillation one needs to solve the

equation of motion of the mass. Applying Newtons’s second low one can

write the equation of motion for the mass, m, as

d2 x

m = Fe kx = 2 sin (3t) kx: (6.355)

dt2

Using the values for the spring constant, k = 8N=m; and mass, m = 2kg;

that can be put in the form

d2 x d2 d d

+ 4x = +4 x= + 2i 2i x = sin (3t) : (6.356)

dt2 dt2 dt dt

In view of the general solutions in Eqs. (6.326) and (6.327) that we de-

rived by successively integrating the second-order none homogeneous ODE

given in Eq. (6.312), the homogeneous and the particular solutions to Eq.

(6.356) can be expresses as

Z t

00

bt

xH (t) = C1 e e(a b)t dt00 + C2 ebt (6.357)

and Z t" Z #

t00

bt (a b)t00 at0 0 0

xp (t) = e e e F (t ) dt dt00 : (6.358)

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

6.6. THE METHOD OF SUCCESSIVE INTEGRATION 153

" !#

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

3 1

x (t) = sin (2t) sin (3t) : (6.370)

10 5

1

v (t) = [2 cos (2t) 3 cos (3t)] : (6.371)

5

Using Mathematica:

154 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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:

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

+ + = (x; y; z); (6.372)

@x2 @y 2 @z 2

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

+ +

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.

+ + = 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.

2nd . Substitute the product function into the original PDE and simplify.

4th . Try to separate the variables, getting only one variable on one side of the

equation.

5th . The two sides of the equation must then be equal to the same constant,

called the separation constant. (Think about what to name this constant—

you can save yourself a lot of needless work by choosing a good name!)

6th . Solve the resulting ordinary DE’s using techniques that you were intro-

duced to in the previous sections.

7th . The general solution is then a linear combination of all possible product-

form solutions to the PDE, including all possible values of the separation

constant!

(a) Show that the string supports a wave motion by deriving the wave equation.

@ 2 u (z; t) 1 @ 2 u (z; t)

= ; (6.375)

@z 2 v2 @t2

156 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

where s

T

v= ; (6.376)

(b) Solve the 1-D traveling wave equation for a string of length L having both

ends …xed. Which means, apply the BC’s

Solution:

(a) Because of the weight of the mass hanging the string experiences a tension

force T . When is displaced from the equilibrium by the vibrator as shown

in the …gure below:If the string is displaced from the equilibrium, the net

6.7. PARTIAL DIFFERENTIAL EQUATIONS 157

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

@u u(y + y) u(y)

= lim (6.382)

@y y=y

y!0 y

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

@2u

F 'T y (6.385)

@y 2

Noting that for the transverse motion, Newton’s second law can be written

as

@2u

F = ma = m 2 (6.386)

@t

If the string has a linear mass density , we can write m = y so that

@2u

F = y (6.387)

@t2

Combining the two expressions for the force F; we …nd

@2u @2u

T y= y (6.388)

@y 2 @t2

158 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Which leads to

@2u @2u @2u 1 @2u

= or = (6.389)

@y 2 T @t2 @y 2 v 2 @t2

where s

T

v= : (6.390)

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

6.8. HOMEWORK ASSIGNMENT 12 159

X

u(y; t = 0) = (Bk sin (ky)) Ck = 0 ) Ck = 0: (6.399)

k

X X

u(y; t) = Bk Dk sin (ky) sin (kvt) = Bk0 sin (ky) sin (kvt) (6.400)

k k

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.

160 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

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

2. Find the general solution for the None-homogeneous DE

d2 y dy

5 +6 + 2y = x2 + 6x:

dx2 dx

3. Find the general solution for the None-homogeneous DE

d2

+ 1 y = 8x sin x:

dx2

d2 y dy

5 + 6y = 2ex + 6x 5:

dx2 dx

6.11. HOMEWORK ASSIGNMENT 15 161

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 :

162 CHAPTER 6. INTRODUCTION TO DIFFERENTIAL CALCULUS II

Chapter 7

Integral calculus-scalar

functions

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.

We are interested in the methods of integrations of functions of single, two, and

three variables in Cartesian coordinates.

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)

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)

166 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

The points along the length of the rectangle, l, are de…ned by the equation

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

7.1. INTEGRATION IN CARTESIAN COORDINATES 167

as 2 3

Zx2 xZ2 (y)

6 7

A= 4 dy 5 dx; (7.20)

x1 x1 (y)

Note that even though the order of integrations are switched, the result should

be the same.

C. Triple integral-Volume

ZZZ

V = dxdydx: (7.21)

Example 1 Line and surface integral : Consider a wire bent into the arc of a

circle of radius R, as shown in the …gure below.

(a) You climbed along this wire. Determine the height of the elevation that

you climbed if you have moved a distance of half of the radius along the

x-direction.

(b) Find the length of the wire

(c) Find the area bounded by the curve, the x and y axes

Solution:

168 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

Z x2

y= m (x) dx; (7.22)

x1

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

7.1. INTEGRATION IN CARTESIAN COORDINATES 169

(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

v" #

ZRu

u 2 ZR

t x Rdx

l= p + 1 dx = p dx: (7.26)

R 2 x2 R 2 x2

0 0

Z=2 Z=2

R2 cos ( ) R

l= q d =R d = : (7.28)

2

R2 1 sin ( ) 2

0 0

variable : That means for x = 0

x = R sin ( ) ) =0

and x = R

1

x = R sin ( ) ) = sin (1) = =2

s

Zy2 2

dx

l= 1+ dy; (7.29)

dy

y1

(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

170 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

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

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

Z

x=1

1

V = [6 6x] dx = 6x 3x2 0

= 3cm2 : (7.40)

x=0

7.1. INTEGRATION IN CARTESIAN COORDINATES 171

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)

dq 0

dV (r) = : (7.43)

4 r ~r0 j

0 j~

Line charge density : Suppose you are given a charge that is distributed on

a …nite length of wire with length, L. Also the charge is not uniformly

distributed on the wire and the charge distribution is described by a line

a line charge density,

dq 0

(~r0 ) = ) dq 0 = (~r0 )dr0 : (7.44)

dr0

Using this equation, the total electrostatic potential due to the entire

charge over the entire length can be determined from Eq. (7.43) using the

172 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

line integral

Z Z

dq 0 1 (~r0 )

V (r) = = dr0 (7.45)

4 r ~r0 j

0 j~ 4 0 L j~r ~r0 j

Surface charge density: Any external charge that one may put on a metal-

lic ball is distributed only on the surface of the sphere no matter the

magnitude of the charge. This charge distribute itself uniformly or none

uniformly depending on the absence or presence of charges in surround-

ing regions of the sphere. let’s consider the general case none uniform

distribution that can be described by a surface charge density,

dq 0

(~r0 ) = ) dq 0 = (~r0 )da0 ; (7.46)

da0

where da0 is the in…nitessimal area over which the in…nitessimal charge

dq 0 is distributed on: In this case the electrostatic potential at a point ~r

due to the entire charge on the surface of the sphere is given

Z Z

1 (~r0 )

V (r) = da0 : (7.47)

4 0 j~r ~r0 j

Area

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)

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

7.1. INTEGRATION IN CARTESIAN COORDINATES 173

and

ZZ Z

Q= (~r0 )dv 0 ; (7.53)

V olum

respectively.

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

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

B. The total mass: The total mass, M , in a volume with mass density, (r)

is given by

ZZ Z

M= (r)dxdydz: (7.57)

V ol m

174 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

C. Center of Mass: The center of mass (gravity) for a given mass with uni-

formly or nonuniform distribution is a point at which the mass can be

at rotational equilibrium about an axis passing through this point. The

center mass, in Cartesian coordinates, (x; y; z) can be determined using

Z Z Z Z Z Z

x dm = xdm; y dm = ydm; and z dm = zdm: (7.58)

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.

7.1. INTEGRATION IN CARTESIAN COORDINATES 175

(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

Z L

1 (~r0 )

V (r) = dr0 (7.67)

4 0 0 j~r ~r0 j

0

~r ~r = jb ( x

^) x0 (^

x)j = b + x0

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

176 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

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

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.

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 :

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 )

x02 + y 02 = R2 (7.74)

the equation of curved part of the semi-circle in the …gure can be written

as p

y 0 (x0 ) = R2 x02 : (7.75)

7.1. INTEGRATION IN CARTESIAN COORDINATES 177

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

(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

" R

#

A R 2 x3 x5 R5 R5 2AR5

Q= =A = (7.79)

2 3 5 R 3 5 15

0

2 (y )

Z Zy2 xZ

Q= (~r0 )da0 = [ (x0 ; y 0 )dy 0 ] dx0 (7.80)

area

y1 x1 (y 0 )

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

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

178 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

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

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

7.1. INTEGRATION IN CARTESIAN COORDINATES 179

we may write

Z =2

M R2 + 2R2 sin2 ( )

I = 2 cos2 ( ) d

=2 3

Z =2

2M R2

= cos2 ( ) + 2 sin2 ( ) cos2 ( ) d : (7.93)

3 =2

(i.e. R x R) =2 =2) : Using the result obtained by Math-

ematica (which you can also easily show),

Z =2

3

cos2 ( ) + 2 sin2 ( ) cos2 ( ) = : (7.94)

=2 4

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.

180 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

Solution:

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)

Z

R

M = dm = (7.100)

2

7.1. INTEGRATION IN CARTESIAN COORDINATES 181

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

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

2 p R

2 p R

x= R2 x2 ;y = R2 x2 ; (7.109)

0 0

182 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

2R

x=y= : (7.110)

(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

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) ;

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)

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)

p

y = R 2 x2 : (7.124)

Z R Z R

R3 2 R3

V = y 2 dx = R2 x2 dx = R3 = : (7.125)

0 0 3 3

We now consider the methods of integrations in curvilinear coordinates. We …rst

consider one and two dimensional curvilinear coordinates. For three-dimensional

184 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

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.

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

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 ] :

q p

2

dl = cos2 + sin2 dr2 + r2 sin2 + cos2 d = dr2 + r2 d 2 :

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

7.2. INTEGRATION IN CURVILINEAR COORDINATES 185

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)

(a) The total mass of the wire is given by

Z Z

M = dm = L ds (7.136)

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)

da = rdrd (7.141)

Z =2 Z R

R2

A= rdrd = (7.142)

0 0 4

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:

186 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

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)

da = (dr) (rd') = rdrd' (7.146)

7.2. INTEGRATION IN CURVILINEAR COORDINATES 187

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)

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.

Z Z

I= r2 dm = r2 (x; y; z) dxdydz (7.149)

vol vol

Z Z Z

M M

I= r2 dxdydz = r2 dxdydz = r2 dxdydz (7.150)

V R2 H

vol vol vol

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

circular cone of base radius R and height H to be

R2 H

V = : (7.153)

3

Z Z Z

dv = rdrd'dz ) V = rdrd'dz (7.154)

188 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

(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

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

" H

#

R2 2 z3

2 R2 H

V = H z Hz + = : (7.158)

H2 3 0 3

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

7.2. INTEGRATION IN CURVILINEAR COORDINATES 189

density given by

'

( ; ') = 0 sin ( ) cos (7.161)

4

where 0 is a constant. Find the total charge.

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

190 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

Recalling that

Z

sin 2

sin2 ( ) d = = (7.166)

0 2 4 0 2

and Z 2 2

' '

cos d' = 4 sin =4 (7.167)

0 4 4 0

2

Q=2 0R : (7.168)

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!)

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

2

A= da =

1

0 1

7.3. HOMEWORK ASSIGNMENT 16 191

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

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:

192 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

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

(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:

7.5. HOMEWORK ASSIGNMENT 18 193

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

(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

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,

(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:

194 CHAPTER 7. INTEGRAL CALCULUS-SCALAR FUNCTIONS

Chapter 8

Vector Calculus

introduced to calculus of vector functions (di¤erentiation and integration). But

…rst we begin by reviewing vector products and their applications in physics

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)

~

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 ".

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

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)

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

198 CHAPTER 8. VECTOR CALCULUS

~ ~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

, where (i; j; k) = (1; 2; 3) (8.23)

~a1 (~a2 ~a3 )

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.

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

8.1. VECTOR PRODUCTS 199

d ~

cA ~ dc + dA c

~ =A (8.26)

dt dt dt

Derivative of scalar vector product:

d ~ ~ ~ ~

~ dB + dA B

A B =A ~ (8.27)

dt dt dt

Derivative of vector product:

d ~ ~

dB ~

dA

A ~ =A

B ~ + ~

B (8.28)

dt dt dt

Example 8.1 The position of an object at time, t; is described by the two

dimensional vector, ~r (t) : This vector, using Polar coordinates, (r; ) ; can

be expressed as

~r (t) = r cos ( ) x

^ + r sin ( ) y^: (8.29)

Find the general expressions for the velocity, ~v (t) ; and acceleration, ~a (t) ;

and express your answer in Polar Coordinates.

Solution:

The velocity: ~v (t) ; is given by

d~r d d

~v (t) = = (r cos ( )) x

^+ (r sin ( )) y^ (8.30)

dt dt dt

Assuming the general case where both r and are time dependent, upon

di¤ erentiating, we …nd

dr d cos ( ) dr d sin ( )

~v (t) = cos ( ) x

^+r x

^+ sin ( ) y^ + r y^

dt dt dt dt

dr d dr d

) ~v (t) = cos ( ) x

^ r sin ( ) x ^+ sin ( ) y^ + r cos ( (8.31)

) y^

dt dt dt dt

that can put in the form

dr d

~v (t) = (cos ( ) x

^ + sin ( ) y^) + r ( sin ( ) x

^ + cos ( ) y^) : (8.32)

dt dt

Noting that in two-dimensional Polar coordinates, the unit vectors along

the radial direction (i.e. along ~r) is

d~r d~r

r^ = = cos ( ) x

^ + sin ( ) y^; (8.33)

dr dr

and along the angular direction is

^ = d~r d~r

= sin ( ) x

^ + cos ( ) y^; (8.34)

d d

200 CHAPTER 8. VECTOR CALCULUS

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

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

~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

8.1. VECTOR PRODUCTS 201

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

d2 r dr ^;

~a (t) = r! 2 r^ + 2 !+r (8.50)

dt2 dt

where we used

dr d! d2 r

at = ! + r; ar = 2 ; vt = !r (8.51)

dt dt dt

and the angular acceleration de…ned by

d2

= (8.52)

dt2

For a circular orbit where r is constant of motion (i.e. ~r = R^

r), we have

dr d2 r

= 0; 2 = 0 (8.53)

dt dt

and

~v (t) = vt ^ = !R^ (8.54)

so that

vt2

~a (t) = R! 2 r^ + R ^ = r^ + R ^ = ~ac + ~ajj (8.55)

R

where

vt2

~ac = r^ (8.56)

R

and

d!

~ajj = R ^ = R ^ (8.57)

dt

are the centripetal and tangential accelerations, respectively.

202 CHAPTER 8. VECTOR CALCULUS

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)

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

df @f @f @f

= nx + ny + nz

ds @x @y @z

@f @f @f

= x

^+ y^ + z^ (nx x

^ + ny y^ + nz z^) (8.71)

@x @y @z

df

= rf n

^;

ds

where n^ is the unit vector along the direction of the vector S ~ that is tangent

to the surface de…ned by the function, f (x; y; z) at a point on the surface with

coordinates (x; y; z).

204 CHAPTER 8. VECTOR CALCULUS

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

@f ^ @f '

^ @f

rf = r^ + + ; (8.73)

@r r@ r sin @'

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

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

~ = 2^

the direction S x 2^ y + z^

Solution:

8.2. DIFFERENTIAL OPERATORS 205

(a) Since the electric potential depends only on r, using spherical coordinates

is easier than Cartesian. Using the gradient in spherical coordinates, one

can then write the electric …eld as

@ kQ 1 @ kQ 1 @ kQ

~ (~r) =

E r (r) = r^ +^ +'^

@r r r@ r r sin @' r

~ (~r) = @ kQ kQ

)E r^ = 2 r^ (8.79)

@r r r

kQ

(r) = ; (8.80)

r (x; y; z)

where p

~r = x^

x + y y^ + z z^ ) r = x2 + y 2 + z 3 : (8.81)

Then we may write

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

@ (~r) kQ y kQ @ (~r) kQ z kQ

= = y; = = z: (8.84)

@y r2 r r3 @z r2 r r3

kQ kQ kQ kQ kQ

r (~r) = x^

x y y^ z z^ = (x^

x + y y^ + z z^) = ~r:

r3 r3 r3 r3 r3

(8.85)

Noting that the unit vector along ~r is

~r

r^ = (8.86)

r

the electric …eld is found to be

~ (~r) = kQ

E r (~r) = r^: (8.87)

r2

206 CHAPTER 8. VECTOR CALCULUS

~ = 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)

d (~r) kQ 2 2 1

= 3=2

x y+ z (8.94)

ds (x2 + y2 + z2) 3 3 3

(1;2; 1)

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?

~

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,

8.2. DIFFERENTIAL OPERATORS 207

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

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

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

208 CHAPTER 8. VECTOR CALCULUS

~ (~r) = @2 @2 @2 (x; y; z)

r E r r (x; y; z) = 2

+ 2 + 2 =

@x @y @z 0

2

)r (x; y; z) = F (x; y; z) : (8.106)

where

(x; y; z)

F (x; y; z) = : (8.107)

0

Equation (8.106) is known as Poison’s equation. In a free space where there is

no charge, (x; y; z) = 0; we …nd

r2 (x; y; z) = 0; (8.108)

which is known as Laplace’s equation. Often you encounter Poison’s and Laplace’s

equations in classical electromagnetism and mechanics theories.

You also encounter, more frequently, in most branches of theoretical physics

that could involve the gradient, the divergence, and the curl of a scalar and/or

a vector functions. Thus it worth to remember the following relations:

~ ; multiplied

(a) The divergence of a vector and a scalar : For a vector function, V

by a scalar functions f; the divergence can be expressed as

~

r: f V ~ rf + f r:V

=V ~ (8.109)

(b) The curl of the curl : In view of the relation for triple vector product

~

A ~

B ~ =B

C ~ A

~ C

~ ~ A

C ~ B

~ (8.110)

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 :

8.2. DIFFERENTIAL OPERATORS 209

(a) rf

~

(b) r V

(c) r ~

V

(d) r2 f

~

(e) r2 V

~

(f ) r2 f V

Solution:

(a) The gradient of the scalare function, f , given by

@f @f @f

rf = x

^+ y^ + z^: (8.114)

@x @y @z

rf = x

^+ y^ + z^ = yz x

^ + xz y^ + xy^

z: (8.115)

@x @y @z

rf = x

^ + y^ z^: (8.116)

~ = (Vx ; Vy ; Vz ), is given by

(b) The divergence of the vector function, V

r V (8.117)

@x @y @z

Noting that

Vx = xy; Vy = yz; Vz = xz

one …nds

r V (8.118)

@x @y @z

~ = 1:

r V (8.119)

V

x

^ y^ z^

r ~ = det

V @ @ @

(8.120)

@x @y @z

xy yz xz

210 CHAPTER 8. VECTOR CALCULUS

that leads to

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)

r ~ =x

V ^ y^ z^: (8.122)

@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

8.2. DIFFERENTIAL OPERATORS 211

~ = 0:

r2 V (8.131)

We can also apply the relation

~ =r r V

r2 V ~ r r ~

V

~ =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

~ =r r V

r2 V ~ r r ~

V =0 (8.137)

(f ) Noting that

~

r2 f V ^r2 (f Vx ) + y^r2 (f Vz ) + z^r2 (f Vz )

=x (8.138)

and

2 2 2

@ 2 (xy) z @ 2 (xy) z @ 2 (xy) z

2

r (f Vx ) = + + = 2y 2 z + 2x2 z

@x2 @y 2 @z 2

(8.139)

212 CHAPTER 8. VECTOR CALCULUS

2 2 2

@ 2 (yz) x @ 2 (yz) x @ 2 (yz) x

2

r (f Vy ) = + + = 2z 2 x + 2y 2 x

@x2 @y 2 @z 2

(8.140)

2 2 2 2 2 2

@ (zx) y @ (zx) y @ (zx) y

r2 (f Vz ) = + + = 2z 2 y + 2x2 y

@x2 @y 2 @z 2

(8.141)

we …nd

~

r2 f V ^ 2y 2 z + 2x2 z + y^ 2z 2 x + 2y 2 x + z^ 2z 2 y + 2x2 y (8.142)

=x

Using Mathematica

The gradient, the divergence, the Laplacian, and the curl that we have seen

so far is for scalar and vector functions expressed in terms of Cartesian coordi-

nates. Depending on the problem we want to solve, sometimes it is convenient

to evaluate these quantities in curvilinear coordinates by expressing the func-

tions using the corresponding transformations for the Cartesian to the conve-

nient curvilinear coordinates. Thus it is important to know the expression for

these quantities in the two curvilinear coordinates we studied (cylindrical and

spherical coordinates). These expressions can be derived from its corresponding

expressions in Cartesian. For now we list these expressions in cylindrical and

spherical coordinates.

(a) Cylindrical coordinates (r; '; z): For a scalar function, f (r; '; z); the gradi-

ent and Laplacian are given by

@f 1 @f @f

rf = r^ +'

^ + z^ (8.143)

@r r @' @z

8.2. DIFFERENTIAL OPERATORS 213

and

1 @ @f 1 @2f @2f

r2 f = r + 2 2

+ 2; (8.144)

r @r @r r @' @z

respectively. For a vector function,

V ^ + Vz (r; '; z) z^;

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 @'

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,

V (r; '; ) = Vr (r; '; )^ ';

~ = 1 @ r2 Vr +

r V

1 @

(sin ( ) V ) +

1 @V'

(8.149)

r2 @r r sin ( ) @ r sin ( ) @'

and

1 @ @V 1 @Vr 1 @

r ~ = r^

V (sin ( ) V' ) +^ (rV' )

r sin ( ) @ @' r sin ( ) @' r @r

1 @ @Vr

+^' (rV ) ; (8.150)

r @r @

respectively.

214 CHAPTER 8. VECTOR CALCULUS

a current I along the positive z-direction and this current is uniformly

distributed.

~ =

B 0 Ir=2 a2 '

^ for r < a (8.151)

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)

8.2. DIFFERENTIAL OPERATORS 215

(b) We note that the components of the magnetic …eld in cylindrical coordinates

are

Ir

Br = 0; B' = 0 2 ; Bz = 0:

2 a

Then using the expression for the curl of a vector in cylindrical coordinates

~ = r^ 1 @Bz

r B

@B'

+'

^

@Br @Bz

+ z^

1 @

(rB' )

@Br

r @' @z @z @r r @r @'

(8.152)

we …nd

~ @B' 1 @

r B = r^ + z^ (rB' )

@z r @r

@ 0 Ir 1 @ Ir

= r^ 2

+ z^ r 0 2 (8.153)

@z 2 a r @r 2 a

2

)r ~ = z^ 1 @

B 0 Ir

= 0I

z^ (8.154)

r @r 2 a 2 a2

216 CHAPTER 8. VECTOR CALCULUS

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

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

8.3.1 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)

I

I= ~ r) d~r

A(~ (8.163)

C

Solution:

The closed line integral can easily be carried out if we divide the integral into

three line integrals corresponding to the three sides ( s1; s2;and s3),

I Z Z Z

~ r) d~r =

A(~ ~ r) d~r +

A(~ ~ r) d~r +

A(~ ~ r) d~r

A(~ (8.164)

C s1 s2 s3

~r = x^

x + y y^ (8.165)

so that

d~r = dx^

x + dy y^ (8.166)

218 CHAPTER 8. VECTOR CALCULUS

x = 0 ) ~r = y y^ ) d~r = dy y^ (8.167)

for s2 p p

y= 2 ) ~r = x^

x+ 2^

y ) d~r = dx^

x (8.168)

and for s3

p p dx

y= x ) ~r = x^

x+ x^ x + p y^:

y ) d~r = dx^ (8.169)

2 x

The function

~ r) = x2 y^

A(~ x + xy 2 y^ + a3 e y

x cos ( ) z^; (8.170)

also takes di¤ erent expressions on these three sides of the curve. On s1

~ r) = 0

x = 0 ) A(~ (8.171)

on s2

p p p

y= ~ r) = x2 2^

2; z = 0 ) A(~ y + a3 e

x + 2x^ 2

x cos ( ) z^ (8.172)

and on s3

p p

~ r) = x5=2 x

y = x; z = 0 ) A(~ ^ + x2 y^ + a3 e x

x cos ( ) z^ (8.173)

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. SINGLE VECTOR FUNCTION INTEGRALS 219

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)

If the line integral over a closed curve is zero independent of the path,

I

~ (x; y; z) d~r = 0:

V (8.179)

C

r ~ (x; y; z) = 0:

V (8.180)

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)

220 CHAPTER 8. VECTOR CALCULUS

F~ = r U (8.183)

on the moon using the gravitational potential energy.

(b) Show that gravitational force is a conservative force using the work done

by the gravitational force when the moon makes one complete revolution

around the earth. I

W = F~ d~r

C

Solution:

(a) The gravitational force is given by

F~ = r U (8.184)

using the gradient in spherical coordinates

@U ^ 1 @U 1 @U

rU = r^ + +'

^ (8.185)

@r r @ r sin @'

we …nd

@U @ GM m GM m

F~ = r^ = r^ = r^ (8.186)

@r @r r r2

(b) The work done is given by

I I

GM m

W = F~ (x; y; z) d~r = r^ d~r (8.187)

C C a2

for a circular orbit r is a constant. That means in spherical coordinates

the position vector

~r = a^

r (8.188)

becomes

d~r = ad^

r: (8.189)

Recalling that in spherical coordinates

~r = r sin ( ) cos (') x

^ + r sin ( ) sin (') y^ + r cos ( ) z^

the unit vector along the radial direction

@~r @~r

r^ =

@r @r

= sin ( ) cos (') x

^ + sin ( ) sin (') y^ + cos ( ) z^ (8.190)

we may write

d^

r = [cos ( ) cos (') d sin ( ) sin (') d'] x

^

+ [cos ( ) sin (') d + sin ( ) cos (') d'] y^ sin ( ) d z^: (8.191)

8.3. SINGLE VECTOR FUNCTION INTEGRALS 221

r^ d~r = sin ( ) cos ( ) cos2 (') d sin2 ( ) sin (') cos (') d'

+ sin ( ) cos ( ) sin2 (') d + sin2 ( ) sin (') cos (') d'

sin ( ) cos ( ) d :

) r^ d~r = 0 (8.192)

I

GM m

W = r^ d~r = 0 (8.193)

C a2

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

222 CHAPTER 8. VECTOR CALCULUS

we may write

= 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

@U @U @U

dU = dx + dy + dz; (8.200)

@x @y @z

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

V x + g(x; y; z)^

y + h(x; y; z)^

z (8.202)

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.

8.4.1 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

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

224 CHAPTER 8. VECTOR CALCULUS

order (i.e. …rst with respect to x and then with respect to y) for the

integral

ZZ Z d Z xr

@Q(x; y) @Q(x; y)

dxdy = dx dy (8.209)

A @x c xl @x

as shown in the …gure below,for the same area then we can write

ZZ Z d Z d Z d

@Q(x; y)

dxdy = [Q(xr ; y) Q(xl ; y)] dy = Q(xr ; y)dy Q(xl ; y)dy

A @x c c c

Z d Z c ZZ I

@Q(x; y)

= Q(xr ; y)dy + Q(xl ; y)dy ) dxdy = Q(x; y)dy

c d A @x C

(8.210)

Here also the line integral over the same curve C is performed in the coun-

terclockwise direction like the previous integration. Therefore, combining

the two relations

ZZ I ZZ I

@P (x; y) @Q(x; y)

dydx = P (x; y)dx; dxdy = Q(x; y)dy;

A @y C A @x C

(8.211)

we …nd

ZZ I

@Q(x; y) @P (x; y)

dxdy = [P (x; y)dx + Q(x; y)dy] : (8.212)

A @x @y C

Example 8.8 Let the area A be the region bounded by the curves y = x2 and

x = y 2 . Verify Green’s theorem for the integral

I

I= 2xy x2 dx + x + y 2 dy (8.213)

C

8.4. MULTIPLE VECTOR INTEGRALS 225

Solution:

Comparing the given integral with

ZZ I

@Q(x; y) @P (x; y)

dxdy = [P (x; y)dx + Q(x; y)dy] : (8.214)

A @x @y C

we want to verify that the result of

I I

I1 = [P (x; y)dx + Q(x; y)dy] = 2xy x2 dx + x + y 2 dy

C C

(8.215)

is equal to

ZZ

@Q(x; y) @P (x; y)

I2 = dxdy

A @x @y

ZZ " #

@ x + y2 @ 2xy x2

= dxdy (8.216)

A @x @y

for the area shown in the …gure belowAnother short cut for the line integral

is the following:

Z

I1 = 2xy x2 dx + x + y 2 dy

Cl

Z

+ 2xy x2 dx + x + y 2 dy

Cu

226 CHAPTER 8. VECTOR CALCULUS

For the lower curve y = x2 ) dy = 2xdx and for the upper curve x =

y 2 ) dx = 2ydy and noting that the two curves intersect at two points

(0; 0) and (1; 1), we can write

Z 1 Z 0

I1 = 2x3 x2 dx + x + x4 2xdx + 2y 3 y 4 2ydy + 2y 2 dy

0 1

(8.217)

Z 1 Z 0

I1 = 2x3 x2 + 2x2 + 2x5 dx + 4y 4 2y 5 + 2y 2 dy (8.218)

0 1

Z 1 Z 0

5 3 2

I1 = 2x + 2x + x dx + 4y 4 2y 5 + 2y 2 dy (8.219)

0 1

1 0

x6 x4 x3 4y 5 y6 2y 3 1

I1 = + + + + = (8.220)

3 2 3 0 5 3 3 1 30

For the area integral

ZZ " # Z 1 Z py

@ x + y2 @ 2xy x2

I2 = dxdy = (1 2x) dxdy

A @x @y 0 y2

Z 1 p Z 1

2 y p

= x x y2 dy = y y y 2 + y 4 dy

0 0

2 3 5 1

2 3=2 y y y 2 1 1 1 1

= y + = + = (8.221)

3 2 3 5 0 3 2 3 5 30

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)

For a vector D

~ (x; y) = Dx (x; y) x

D ^ + Dy (x; y) y^ + Dz (x; y) z^ (8.225)

8.4. MULTIPLE VECTOR INTEGRALS 227

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

228 CHAPTER 8. VECTOR CALCULUS

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

is just the ‡ux of the curl of the vector D

3. The integral I

~ d~r

D (8.230)

C

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)

2z = x2 + y 2 (8.235)

bounded by the plane z = 2.

Z

r ~ n

D ^ dA (8.236)

A

8.4. MULTIPLE VECTOR INTEGRALS 229

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)

230 CHAPTER 8. VECTOR CALCULUS

we …nd Z

r ~ n

D ^ dA = 20 : (8.242)

A

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)

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

8.4. MULTIPLE VECTOR INTEGRALS 231

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

~ (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)

232 CHAPTER 8. VECTOR CALCULUS

Z c Z b Z c Z b

) vx = ~ (a; y; z) x

V ^dydz ~ (0; y; z) x

V ^dydz (8.253)

0 0 0 0

Z c Z b Z c Z b

) vx = Vx (a; y; z)dydz Vx (0; y; z)dydz (8.254)

0 0 0 0

Z c Z b

ab2 c

vx = aydydz = : (8.255)

0 0 2

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

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

8.4. MULTIPLE VECTOR INTEGRALS 233

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

The divergence in cylindrical coordinates:

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)

Solution:

Noting that

Vr = ar; V' = b; V z = cz (8.266)

the divergence

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

234 CHAPTER 8. VECTOR CALCULUS

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)

8.4. MULTIPLE VECTOR INTEGRALS 235

~s (r; '; z) = V

V ~ (r = R; '; z) = aR^

r + b^

' + cz z^ (8.275)

so that

ZZ ZZ

~b (r; '; z) n

V ^ b da = (ar^

r + b^

') ( z^) da = 0 (8.276)

bottom bottom

ZZ ZZ

~t (r; '; z) n

V ^ t da = (ar^

r + b^

' + cH z^) z^da

top top

ZZ

= cHda = c R2 H (8.277)

top

and

ZZ ZZ

~s (r; '; z) n

V ^ s da: = (aR^

r + b^

' + cz z^) r^da

side side

ZZ ZZ

= aRda = aR (2 RH) ) ~s (r; '; z) n

V ^ s da = 2a R2 H:

side side

(8.278)

Therefore the integral over the closed cylindrical surface

ZZ ZZ

~

V (r; '; z) n

^ da = ~b (r; '; z) n

V ^ b da

Surface bottom

enclosing

ZZ ZZ

+ ~t (r; '; z) n

V ^ t da + ~s (r; '; z) n

V ^ s da: (8.279)

top side

becomes

ZZ

~ (r; '; z) n

V ^ da = c R2 H + 2a R2 H: = (2a + c) R2 H

Surface

enclosing

ZZZ

= ~ (x; y; z) dxdydz

r V (8.280)

volum e

~ = 1 @ r 2 Vr +

r V

1 @

(sin ( ) V ) +

1 @

(V' ) (8.281)

r2 @r r sin ( ) @ r sin ( ) @'

Example 8.11 Verify the divergence theorem for the vector …eld

~ = ~r

V (8.282)

and the sphere shown below.

Solution: For the ‡ux, we …nd

ZZ Z2 Z

~ (r; ; ') n

V ^ da = r r^R2 sin ( ) d d'

R^

surf ace

0 0

ZZ

~ (r; ; ') n

V ^ da = 4 R3

surf ace

236 CHAPTER 8. VECTOR CALCULUS

~ = r^

The divergence of the vector V r

~ = 1 @ r 2 Vr +

r V

1 @

(sin ( ) V ) +

1 @

(V' ) (8.283)

2

r @r r sin ( ) @ r sin ( ) @'

becomes

~ = 1 @ r2 Vr = 1 @ r3 = 3

r V (8.284)

r2 @r r2 @r

Then the integral becomes

ZZZ Z2 Z

~ (r; ; ') d =

r V 3r2 dr sin ( ) d d' = 4 R3 :

volum e

0 0

(8.285)

Example 8.12 Verify Stokes’ theorem for the magnetic …eld vector

r + 3^

~ = 4^

B 2^

' (8.286)

Solution:

Br = 4; B = 3; B' = 2 (8.287)

8.4. MULTIPLE VECTOR INTEGRALS 237

~ = 1 @ @

r B (sin ( ) B' ) (B ) r^

r sin ( ) @ @'

1 @ 1 @ 1 @ @

+ (Br ) (rB' ) ^ + (rB ) (Br ) ' ^

r sin ( ) @' r @r r @r @

(8.288)

becomes

1 @ 1 @ 1 @

r ~ =

B ( 2 sin ( )) r^ ( 2r) ^ + (3r) '

^

r sin ( ) @ r @r r @r

2 cos ( ) 2 3 ~ = 2 cot ( ) r^ + 2 ^ + 3 '

= r^ + ^ + ' ^)r B ^ (8.289)

r sin ( ) r r r r r

Noting that the in…nitesimal area on the x-y plane in spherical coordinates

can be expressed as

h i

^ da = z^r sin ( ) drd' = cos ( ) r^ sin ( ) ^ r sin ( ) drd'

n (8.290)

then

2 2 3 h i

~ n

r B ^ da = cot ( ) r^ + ^ + '^ cos ( ) r^ sin ( ) ^ r sin ( ) drd'

r r r

(8.291)

~ n 2 2

)r B ^ da = cot ( ) cos ( ) sin ( ) r sin ( ) drd': (8.292)

r r

Noting that on the x-y plane = =2; we have

r ~ n

B ^ da = 2drd': (8.293)

so that the integral Z

r ~ n

B ^ dA (8.294)

A

238 CHAPTER 8. VECTOR CALCULUS

for the surface bounded by the curve shown in the …gure can be written as

Z Z R Z =2

r ~ n

B ^ dA = 2drd' = R : (8.295)

A 0 0

~r = r^

r ) d~r = dr^

r + rd^

r (8.296)

where

^ + sin ( ) sin (') y^ + cos ( ) z^: (8.297)

r^ = cos (') x

^ + sin (') y^ ) d^

r = [ sin (') x

^ + cos (') y^] d' = '

^ d'

(8.298)

so that

~r = r^r ) d~r = dr^

r + r'

^ d': (8.299)

Then

~ d~r = Br dr + B' rd' = 4dr

B 2rd' (8.300)

which leads to

I Z Z Z

~

B d~r = (4dr 2rd')+ (4dr 2rd')+ (4dr 2rd') (8.301)

C l1 l2 l3

where l1 and l3 are the straight lines (on this lines d' = 0) and l2 is the

curved line (on this line r = R and dr = 0). Therefore

I Z R Z =2 Z 0

~ d~r =

B 4dr 2Rd' + 4dr = R : (8.302)

C 0 0 R

r + 3^

~ = 4^

B 2^

' (8.303)

I Z

B~ d~r = r ~ n

B ^ dA: (8.304)

C A

is valid.

rems

Gauss’ Law

~ =

r E (8.305)

0

8.4. MULTIPLE VECTOR INTEGRALS 239

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

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

240 CHAPTER 8. VECTOR CALCULUS

I Z

~

B d~r = r B ~ n^ dA (8.314)

C A

Z

J~ n

^ dA = 0 Ien = 0

A

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

Mathematical Methods in

Physics II

241

Chapter 9

Introduction to the

Calculus of Variations

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

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

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.)

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

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

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)

246CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

we may write

dY dy d

= + ) Y 0 (x; ) = y 0 (x) + 0

(x) (9.16)

dx dx dx

so that

dY 0 ( ) d 0

= [y (x) + 0 (x)] = 0

(x) : (9.17)

d d

There follows that

Y 0 (x; )j =0 = y 0 (x) (9.18)

and

dY 0 ( ) 0

= (x) : (9.19)

d =0

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)

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

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

There follows that

dy

= m ) y (x) = mx + b; (9.29)

dx

which is equation of a straight line.

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

248CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

that we want to solve applying the calculus of variation, we know the coordi-

nates of two di¤ erent points (x1 ; y (x1 ) ; y 0 (x1 )) and (x2 ; y (x2 ) ; y 0 (x2 )) on the

surface de…ned by F = F (x; y (x) ; y 0 (x)): From the in…nitely many trajectories

that can connect these two points, there is only one trajectory on this surface

that is the shortest (the Geodesic). Finding the Geodesic is the general problem

that can be solved applying the calculus of variation.

The surface is de…ned by the function F (x; y (x) ; y 0 (x)): The distance be-

tween these two points determined by evaluating the integral

Z x2

I= F (x; y (x) ; y 0 (x)) dx: (9.32)

x1

To determine the equation that the function F is governed by so that we …nd the

shortest length joining the two points, let the function for any path connecting

the two points be Y (x) : From these in…nite number of functions there is only

one function that gives the minimum distance between the two points. If this

function is y (x) ; then we may write Y (x) in terms of y (x) as

We also have

dY (x; ) dY ( )

= (x) ) = (x) ; (9.36)

d d =0

9.2. THE GENERAL PROBLEM 249

and

dY dy d

= + or Y 0 (x; ) = y 0 (x) + 0

(x) ) Y 0 (x; )j =0 = y 0 (x) ; (9.37)

dx dx dx

which gives

dY 0 ( ) d 0 0 0 dY 0 ( ) 0

= [y (x) + (x)] = (x) ) = (x) : (9.38)

d d d =0

Z x2

I( )= F (x; Y (x; ) ; Y 0 (x; )) dx; (9.39)

x1

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

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

250CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

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

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)):

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

252CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

we have

r

2

Z 2

p Z 2 1 + dx Z y2

p

dx2 + dy 2 dy 1 + x02 1

I= p = p p dy = p

dy;

1 2gy 1 2gy

y1 y 2g

(9.54)

so that the stationary value of this integral is determined from the Euler-

Lagrange equation

@ @F @F

0

= 0; (9.55)

@y @x @x

where p

1 1 + x02

F (y; x (y) ; x0 (y)) = p p ; (9.56)

2g y

and

dx (y)

x0 (y) = : (9.57)

dy

Noting that

@F

=0 (9.58)

@x

and

@F 1 x0

= p p p (9.59)

@x0 2g 1 + x02 y

so that

!

@ 1 x0 x0 p

p p p =0) p p = c; (9.60)

@y 2g 1 + x02 y 1+x 02 y

where c is a constant. Solving for x0 ; we …nd

x0 p

p p = c ) x02 = c 1 + x02 y

1 + x02 y

r Z yr

dx cy cy

) x02 (1 cy) = cy ) = )x= dy: (9.61)

dy 1 cy 0 1 cy

Introducing the transformation de…ned by

1 1

cy = sin2 = (1 cos ( )) ) dy = sin cos d (9.62)

2 2 c 2 2

we have

Z r Z s

y

cy 1 sin2 2

x = dy = sin cos d

0 1 cy c 0 1 sin2 2

2 2

Z Z

1 sin 1

= 2

sin cos d = sin2 d

c 0 cos 2

2 2 c 0 2

Z

1 1 1

) x= (1 cos ( )) d ) x = ( sin ( )) : (9.63)

c 0 2 2c

9.4. APPLICATIONS: CLASSICAL MECHANICS 253

Therefore, the trajectory of the bead that takes the smallest possible time

is given by

1 1

x= ( sin ( )) ; y = (1 cos ( )) : (9.64)

2c 2c

Cycloid: Consider a circle of radius r rolling along the positive x-axis with a

constant angular velocity staring from the origin. If you mark the point on the

circle coinciding with the origin at the initial time (as shown in the Fig. 9.3)

and follow the trajectory of this point its x and y coordinates of this point are

given by

x = r( sin ( )) ; y = r (1 cos ( )) ; (9.65)

where is the angle that the circle (the point) rotated. For example the …gure

below shows this trajectory for a point on a circle of unit radius (r = 1).

For a given , the circle’s centre lies at

x = r ; y = r: (9.66)

x = r ;y = r: (9.67)

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

254CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

@ @F @F

= 0; (9.69)

@x @y 0 @y

@ @F @F

= 0

@x @z 0 @z

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

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

256CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

Since L = L (x; y; x;

_ y;

_ t) which is a function of two variables and we must

have two Euler-Lagrange equations

@ @L @L @ @L @L

= 0; = 0: (9.78)

@t @ x_ @x @t @ y_ @y

@ @L @L @ @U (x)

=0) (mx)

_ + =0

@t @ x_ @x @t @x

@U (x)

) m•

x=

@x

@ @L @L

= 0 ) m•

y = 0 (Zero acceleration) (9.79)

@t @ y_ @y

Example 9.6 The Atwood’s Machine: A string of length, l , passes over a fric-

tionless pulley connecting two masses, m1 and m2 . Find an expression

for the acceleration of the masses in the system.

9.5. PHYSICAL APPLICATION OF THE EULER-LAGRANGE EQUATION257

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

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

258CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

Recalling that

y_ 2 = y_ 1 (9.91)

the acceleration of the second mass becomes

m1 m2

a1 = g: (9.92)

m1 + m2

The minus sign indicates the …rst mass is accelerating in the negative y-

direction.

Example 9.7 Central Forces: Describe the properties of the motion of a mass

m moving under the in‡uence of a central force (that is, a force acting

only along the radial direction) given by

F~ = f (r)^

r (9.93)

for some function f (r). Assume that the motion is con…ned to a plane.

Solution: The kinetic energy

1

T =mv 2 : (9.94)

2

Using polar coordinates the magnitude of the velocity can be expressed as

2

v = r_ 2 + r2 _ (9.95)

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)

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

9.5. PHYSICAL APPLICATION OF THE EULER-LAGRANGE EQUATION259

we have

@ @L @L @ @L @L

= 0; =0 (9.101)

@t @_ @ @t @ r_ @r

so that using

@L @L @L 2 @L

= 0; = mr2 _ ; = mr _ + f (r); = mr_ (9.102)

@ @_ @r @ r_

we …nd

@ @L @L

=0) = const ) mr2 _ = cont ) I! = cons:

@t @_ @_

(Conservation of Ang. Mom.)

@ @L @L 2

r = mr _ + f (r)

= 0 ) m• (9.103)

@t @ r_ @r

260CHAPTER 9. INTRODUCTION TO THE CALCULUS OF VARIATIONS

Chapter 10

Introduction to the

Eigenvalue Problem

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

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

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

~ Commutativities: For any two multiplyable matrices C and D;

CD 6= DC (10.6)

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)

IA = AI = A (10.10)

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

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)

1 1

A A = AA =I (10.14)

We can determine the inverse of an invertible matrix (det jAj =

6 0) using

row reduction or the adjoint matrix.

a. Row reduction in this approach for the matrix, for example,

2 3

a11 a12 a13

A = 4 a21 a22 a23 5 (10.15)

a31 a32 a33

we start from 2 3

a11 a12 a13 1 0 0

4 a21 a22 a23 0 1 0 5 (10.16)

a31 a32 a33 0 0 1

and do elementary row operation until we end up with

2 3

1 0 0 b11 b12 b13

4 0 1 0 b21 b22 b23 5 (10.17)

0 0 1 b31 b32 b33

so that the inverse of the Matrix A is given by

0 1

b11 b12 b13

A 1 = @ b21 b22 b23 A : (10.18)

b31 b32 b33

b. Using the adjoint matrix: Using the adjoint matrix the inverse can be ex-

pressed as

T

[cof (A)]

A 1= (10.19)

det jAj

Example 10.1 Find the inverse of the matrix

0 1

1 2 3

A=@ 2 0 4 A (10.20)

1 1 1

using

264 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

b. The adjoint matrix approach

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

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

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

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

0 1

1 0 0

I=@ 0 1 0 A; (10.32)

0 0 1

266 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

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

M21 M22 M23 = 0; (10.35)

M31 M32 M33

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

10.3. EIGENVALUES AND EIGENVECTORS 267

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)

268 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

Similarly for 2 =1

0 10 1

1 1 0 x2

@ 1 1 0 A @ y2 A = 0 ) x2 + y2 = 0; x2 y2 = 0; z2 = 0

0 0 1 z2

) x2 = y2 ; z2 = 0 (10.45)

0 1

1

j 2i = x2 @ 1 A ; (10.46)

0

0 1

1

h 2 j 2i = 1 ) x22 1 1 0 @ 1 A = 1 ) x2 = p1 (10.47)

0 2

1 0

1

1

) j 2i = p @ 1 A (10.48)

2 0

and for 3 = 1

0 10 1

1 1 0 x3

@ 1 1 0 A @ y3 A = 0 ) x3 + y3 = 0; x3 + y3 = 0; z3 = 0

0 0 1 z3

) x3 = y3 ; z3 = 0 (10.49)

1 0 0 1

1 1

1 @

) j 3 i = y3 @ 1 A ) j 3i = p 1 A: (10.50)

0 2 0

Using Mathematica:

10.3. EIGENVALUES AND EIGENVECTORS 269

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.

270 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

Normal modes of Vibration for couple harmonic oscillators

by three identical springs of spring constant k.

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 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

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

@ @L @L

= 0 for i = 1; 2 (10.61)

@t @ x_ i @xi

we …nd

m•

x1 [ kx1 + k (x2 x1 )] = 0 ) m•

x1 = 2kx1 + kx2 (10.62)

and

m•

x2 [ kx2 k (x2 x1 )] = 0 ) m•

x2 = 2kx2 + kx1 : (10.63)

If we introduce a constant r

k

; != (10.64)

m

then the above two equations can be put in the form

x

•1 = 2! 2 x1 + ! 2 x2 (10.65)

272 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

and

• 2 = ! 2 x1

x 2! 2 x2 : (10.66)

These two equations describe the equations of motion for the two masses.

As you can see the two equations are coupled second order di¤erential

equations.

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)

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

10.4. PHYSICAL APPLICATIONS 273

:: ::

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)

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

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)

Mj i= j i; (10.79)

2! 2 !2 2 2

det = 0 ) 2! 2 + !2 =0

!2 2! 2

) 2! 2 + !2 2! 2 + + !2 = 0 ) 1 = !2 ; 2 = 3! 2 :

(10.80)

2! 2 1 !2 X1

=0 (10.81)

!2 2! 2 1 X2

and

2! 2 2 !2 X1

= 0: (10.82)

!2 2! 2 2 X2

274 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

!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:

10.4. PHYSICAL APPLICATIONS 275

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)

276 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

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)

10.4. PHYSICAL APPLICATIONS 277

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

Step 5: The Normal Modes of Vibration: Suppose the initial state of the two

masses is described by the …rst eigenvector. That means

x10 1 1

~r (0) = =p (10.104)

x20 2 1

then

~r = U (t)~r (0) (10.105)

gives

p p

x1 (t) 1 cos (!t) + cos p3!t cos (!t) cos p3!t

= p

x2 (t) 2 2 cos (!t) cos 3!t cos (!t) + cos 3!t

1

(10.106)

1

which leads to

=p ) x1 (t) = x2 (t) : (10.107)

x2 (t) 2 cos (!t)

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

278 CHAPTER 10. INTRODUCTION TO THE EIGENVALUE PROBLEM

then we have

p p

x1 (t) 1 cos (!t) + cos p3!t cos (!t) cos p3!t

= p

x2 (t) 2 2 cos (!t) cos 3!t cos (!t) + cos 3!t

1

(10.109)

1

which gives

p

x1 (t) 1 cos p3!t

=p ) x1 (t) = x2 (t) : (10.110)

x2 (t) 2 cos 3!t

p

The two masses oscillate with a frequency 3! out of phase by .

The two modes of vibrations we saw above are called Normal Modes of vibration.

Chapter 11

Special functions

The Factorial, n!: We recall the factorial function, n!; for a positive integer or

zero is de…ned as

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

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

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

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

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)

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)

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

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

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)

282 CHAPTER 11. SPECIAL FUNCTIONS

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

Z 1 Z 1

1 1 x u2

= p e dx = 2 e du; (11.31)

2 0 x 0

1

u2

e du = : (11.32)

0 2

11.2. THE BETA FUNCTION 283

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

includes the following

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)

284 CHAPTER 11. SPECIAL FUNCTIONS

Z 1 Z 1

(q) = xq 1 e x dx, (p) = yp 1

e y

dy (11.39)

0 0

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

Z 1Z 1

2 2

(p) (q) = 4 u2q 1 v 2p 1

e (u +v ) du dv, (11.42)

0 0

we …nd

Z 1 Z =2

2q 1 2p 1 r2

(p) (q) = 4 (r cos ) (r sin ) e rdrd . (11.44)

0 0

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

We recall the Gamma Function

Z 1

(p) = xp 1

e x

dx; (11.50)

0

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)

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)

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)

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)

286 CHAPTER 11. SPECIAL FUNCTIONS

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

p

p py y2

ln (p + y p) = ln (p) + : (11.61)

p 2p

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

Z p

p

y2

lim e 2 dy ! 0: (11.66)

p!1 1

p p p p

p! ' 2p ep ln(p) p = 2p eln(p ) e p ' 2p pp e p : (11.67)

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

288 CHAPTER 11. SPECIAL FUNCTIONS

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

+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

we …nd

and

P (n) = exp[ 6 1021 ]: (11.78)

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

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

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

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

290 CHAPTER 11. SPECIAL FUNCTIONS

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).

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

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

"Z Z pn #

0

1 2 2 2

Pn (x) = p e y dy + e y dy (11.89)

pn 0

2

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%

292 CHAPTER 11. SPECIAL FUNCTIONS

The Complete Elliptic Integral of the First Kind

Z =2

d'

K( ) p : (11.94)

2

0 1 sin2 '

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.

M EI = M E (11.95)

to the side by an angle ) which is just only the gravitational potential

energy given by

11.5. ELLIPTIC INTEGRALS 293

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

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 )

294 CHAPTER 11. SPECIAL FUNCTIONS

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

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 '

= 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 '

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.

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

where

@~

r

@r

r^ = sin ( ) cos (') x

^ + sin ( ) sin (') y^ + cos ( ) z^ = @~

r

(11.111)

@r

is the unit vector along the radial direction. In spherical coordinates we recall

the gradient and the Laplacian for a scalar function, f (r; ; '), are given by

@f ^ 1 @f 1 @f

rf = r^ + +'

^ ; (11.112)

@r r@ r sin @'

and

1 @ @f 1 @ @f 1 @2f

r2 f = r2 + sin ( ) + 2 ; (11.113)

r2 @r @r r2 sin ( ) @ @ r2 sin ( ) @'2

respectively. On the other hand for a vector …eld,

V r + V (r; '; )^ + V' (r; '; )^

~ (r; '; ) = Vr (r; '; )^ '; (11.114)

instead of the gradient, most often, we are interested in the divergence of the

vector …eld. In spherical coordinates the divergence of a vector …eld is given by

~ = 1 @ r2 Vr +

r V

1 @

(sin ( ) V ) +

1 @V'

: (11.115)

r2 @r r sin ( ) @ r sin ( ) @'

Suppose we have some physical quantity described by some scalar function that

depend on the radial distance between two points. A good example of such phys-

ical quantity is electrical and gravitational potentials. Let’s say this function in

296 CHAPTER 11. SPECIAL FUNCTIONS

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

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

r^ 1 1 0 for r 6= 0

4 (~r) = r = r r = r2 = (11.125)

r2 r r 1 for r = 0

11.6. THE DIRAC DELTA FUNCTION 297

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

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:

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.

298 CHAPTER 11. SPECIAL FUNCTIONS

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

11.6. THE DIRAC DELTA FUNCTION 299

For 3D case

Z1Z Z2

(~r) r2 sin ( ) drd d' = 1

0 0 0

Z1Z Z2

f (~r) (~r) r2 sin ( ) drd d' = f (0)

0 0 0

Z1Z Z2

f (~r ~r0 ) (~r ~r0 ) r2 sin ( ) drd d' = f (~r0 ) (11.131)

0 0 0

Example 11.6 Form introductory physics, the electric potential, V (~r) ; due to

a point charge located at the origin (0; 0; 0) (i.e. r = 0) at a point in space

described by the position vector, ~r; is given by

1 q

V (~r) = :

4 0 r

Show that the volume charge density, (~r) ; for this point charge can be

expressed in terms of the Dirac delta function

dq

(~r) = = q (~r) = q (x) (y) (z) ; (11.132)

d

where dq an in…nitessimal charge in an in…nitesimal volume d :

Solution: The electric potential, dV (~r) of an in…nitessimal charge dq 0 in a

volume d 0 as shown in Fig. can be expressed as

ZZZ

1 dq 0 1 (~r0 ) d 0 1 (~r0 ) d 0

dV (~r) = 0j

= 0j

) V (~r) = ;

4 0 j~

r ~

r 4 0 j~

r ~

r 4 0 j~r ~r0 j

V

(11.133)

Using spherical coordinates, we can write

Z1Z Z2

1 (~r0 ) r02 sin ( ) dr0 d 0 d'0

V (~r) = (11.134)

4 0 j~r ~r0 j

0 0 0

Z1Z Z2

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

300 CHAPTER 11. SPECIAL FUNCTIONS

Z1Z Z2

0

f (~r0 ) (~r0 ~r0 ) r02 sin dr0 d 0 d'0 = f (~r0 ) (11.136)

0 0 0

1 (~r0 )

f (~r0 ) = ; (~r0 ~r0 ) =

j~r ~r0 j q

1 1

) f (~r0 ) = = ) ~r0 = 0 (11.137)

j~r ~r0 j r

which leads to

(~r0 )

= (~r0 ) ) (~r0 ) = q (~r0 ) : (11.138)

q

Example 11.7 The volume charge density, (~r) ; of a point charge, q, placed

at a point on the z-axis, ~r0 = a^

z ; can be expressed as

where (~r) is the Dirac Delta function. Show that the electric potential,

V (~r) ; due to this point charge is given by

1 q q 1

V (~r) = = q : (11.140)

4 0 j~r ~r0 j 4 0 2

x2 + y 2 + (z a)

ZZZ

1 (~r0 ) d 0

V (~r) = ; (11.141)

4 0 j~r ~r0 j

V

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

~r0 = x0 x

^ + y 0 y^ + z 0 z^; ~r = x^x + y y^ + z z^; ~r0 = a^

z

q

0 2 2 2

) j~r ~r j = (x x0 ) + (y y 0 ) + (z z 0 ) (11.143)

11.6. THE DIRAC DELTA FUNCTION 301

and

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)

Z1

f (x) (x a) dx = f (a) (11.147)

1

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)

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)

302 CHAPTER 11. SPECIAL FUNCTIONS

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

Di¤erential Equations

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.

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

1

X 1

X

(n + 2) (n + 1) an+2 tn + ! 2 an tn = 0

n=0 n=0

1

X

) (n + 2) (n + 1) an+2 + ! 2 an tn = 0: (12.9)

n=0

an

(n + 2) (n + 1) an+2 + ! 2 an = 0 ) an+2 = !2 :

(n + 2) (n + 1)

Now let’s examine the …rst few terms for this recursion relation. First we

consider when n is even

a0 2

n = 0 ) a2 = !

2 1

1

a0 ( 1) 2 1

) n = 0 ) a2 1 = !

2!

2 2

a2 2 a0 ( 1) ! 2 2

a0 ( 1) ! 2 2

n = 2 ) a4 = ! = =

4 3 4 3 2 1 4!

a0 2

) n = 2 ) a2 2 = ( 1) ! 2 2

(2 2)!

3 3

a4 ! 2 a0 ( 1) ! 2 3 a0 ( 1) ! 2 3

n = 4 ) a6 = = =

6 5 6 5 4 3 2 1 6!

3 2 3

a0 ( 1) !

) n = 4 ) a2 3 = : (12.10)

(2 3)!

From the results we see above it is not hard to come up with the relation

m

a0 ( 1) ! 2m

a2m = (12.11)

(2m)!

that generates the values for the even term coe¢ cients in the series for

m = 0; 1; 2; ::.

Next we shall consider the odd terms for the recursion relation

an

an+2 = !2 : (12.12)

(n + 2) (n + 1)

306CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS

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)!

m

a1 ( 1) ! 2m

a(2m+1) = ; (12.14)

(2m + 1)!

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

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

conditions. Since initially (t = 0) the spring is stretched by xmax and the

mass is released from rest, we have

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.

~ 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

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

3

X

hA jBi = Ai Bi = 0: (12.27)

i=1

308CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS

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

Z b

0 m 6= n;

An (x)Am (x)dx = (12.30)

a C m = n;

the Kronicker delta

0 n=6 m

nm = ; (12.31)

1 n=m

and Dirac notation, we may write

For m = n,

hAn (x) jAn (x)i = Cn (12.33)

so that the for the set of functions de…ned by

An (x)

Fn (x) = p (12.34)

Cn

one can write

Z b

hFn (x) jFm (x)i = Fn (x) Fm (x) dx = nm : (12.35)

a

1

Fn (x) = p sin (nx)

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)

sin [(m n) ]

hFn (x) jFm (x)i = = nm ;

(m n)

if one applies the double angle relations

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

1 sin ((n m) )

hFn (x) jFm (x)i = = nm

n m

310CHAPTER 12. POWER SERIES SOLUTIONS TO DIFFERENTIAL EQUATIONS

We recall a set of functions, Fn (x) ; form an orthonormal set for all x (a; b) ;

when

Z b

hFn (x) jFm (x)i = Fn (x) Fm (x) dx = nm : (12.42)

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)

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

^i (12.45)

i=1

3

X

jri = ri jxi i : (12.46)

i=1

3

X

hxj jri = ri hxj jxi i : (12.47)

i=1

3

X

hxj jri = ri ji ) rj = hxj jri : (12.48)

i=1

If one of the vectors, x^i ; is missing, then we can not express any vector ~r R3

in terms of the remaining two vectors. But if we have all these three vectors

any vector ~r R3 can be expressed in terms of these vectors. Then we say that

the set of vectors f^

xi g = f^ x1 ; x ^3 g is a complete orthonormal set in R3 :

^2 ; x

We now consider the function space where we have the in…nite set of or-

thonormal functions fFn (x)g = fF0 (x) ; F1 (x) ; F2 (x) ; F3 (x) ; :::g for all x

(a; b) : These set of functions form a complete set if a function, g (x) ; de…ned

12.3. COMPLETE SETS OF FUNCTIONS 311

functions. That means

1

X

jg (x)i = cn jFn (x)i ; (12.49)

n=0

where the expansion coe¢ cie