100%(2)100% au considerat acest document util (2 voturi)

2K vizualizări606 paginiAug 29, 2011

© Attribution Non-Commercial (BY-NC)

PDF, TXT sau citiți online pe Scribd

Attribution Non-Commercial (BY-NC)

100%(2)100% au considerat acest document util (2 voturi)

2K vizualizări606 paginiAttribution Non-Commercial (BY-NC)

Sunteți pe pagina 1din 606

Learning without thought is labour lost; thought without learning is perilous. Confucius

Preface

These notes, chapters, lecture notes or what you will have become concrete with my interaction with my students. Though many good books exist on the subject (Jordan & Balmain 1968 W. H. Hayt & Buck 2001), (to name only two) I would like to present this book to my undergraduate students as my view of the subject. In this book my motive is to somehow excite the interest of the reader. The subject is presented as a complete whole, starting from the mathematical principles involved to electromagnetic theory in general. I have included a detailed treatment of the mathematics required, since I feel that the application of the mathematics to engineering situations requires more explanation and solved problems. I include in this book some examples (but not too many) and some exercises which I think may be important. Also I have adopted a conversational style: somewhat like a professor delivering a lecture. Many times I have asked questions, like the questions a student would ask, and then proceeded to answer them. My other advice to the reader is to read and refer to as many texts as possible. The reader is also encouraged to solve as many problems as possible from dierent books and from as many sources as possible. I have also attempted to provide careful and complete explanations of the material, while at the same time maintaining a writing style which is succinct and to the point. Questions for self assessment are given at the end of each chapter. These are (i) Review questions : Theoretical in nature, these questions are without answers. (ii) Numerical problems: The practice questions are given in this section with hints and answers. (iii) Short answer type questions: Questions with answers which will help students in fetching 2-4 marks in examinations are given in this section. (iv) Objective type questions: Multiple choice questions (One question with four options to choose) are given in this section. (v) Open book exam question: Theoretical/parctice question (With hints for reference to particular page/section for answers) are given in this. Sunil Bhooshan, JUIT, Waknaghat

ii

Aclnowledgements

Agastya Bhooshan Gagan Gupta Vinay Kumar Rohit Sharma Lyx Xg Inkscape Latex

iii

Copyright Sunil Bhooshan 2008. Paper or electronic copies for non-commercial use may be made freely without explicit permission of the author. All other rights are reserved.

iv

Contents

Frequently Used Reference Material 0.1. Table of Fundamental Constants . . 0.2. Units . . . . . . . . . . . . . . . . . . 0.3. The Greek Alphabet . . . . . . . . . . 0.4. SI Prexes . . . . . . . . . . . . . . . 0.5. Dielectric Constants of Materials . . 0.6. Relative Permeabilities of Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 1 2 2 2 3

I.

Introductory Material

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4

5 5 5 7 13 16 17 21 24 31 31 33 37 43 43 43 44 48 49 50 51 54 57 63 66 71 74

1. Scalars and Vectors 1.1. Chapter Goals . . . . . . . . . . . . . . . . 1.2. Introduction . . . . . . . . . . . . . . . . . 1.3. Scalars . . . . . . . . . . . . . . . . . . . . 1.4. Vectors . . . . . . . . . . . . . . . . . . . . 1.4.1. Vector Addition . . . . . . . . . . 1.4.2. A Handy Technique . . . . . . . . 1.4.3. Dot Product or Scalar Product . . 1.4.4. Cross Product or Vector Product . 1.5. Units and Dimensions . . . . . . . . . . . 1.5.1. The Basis of Units and Dimensions 1.6. List of Formulae . . . . . . . . . . . . . . . 1.7. Practice Problems and Self Assessment . .

2. Coordinate Systems and Fields 2.1. Chapter Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Scalar and Vector Fields . . . . . . . . . . . . . . . . . . . . . . . 2.4. The Rectangular Coordinate System . . . . . . . . . . . . . . . . 2.4.1. Distance Between Two Points . . . . . . . . . . . . . . . . 2.4.2. Direction Cosines . . . . . . . . . . . . . . . . . . . . . . . 2.4.3. Vector Equation of a Straight Line . . . . . . . . . . . . . 2.4.4. Equation of a Plane . . . . . . . . . . . . . . . . . . . . . . 2.5. Cylindrical Coordinate System . . . . . . . . . . . . . . . . . . . 2.5.1. Equations of Surfaces and Lines in Cylindrical Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6. The Spherical Coordinate System . . . . . . . . . . . . . . . . . . 2.7. List of Formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8. Practice Problems and Self Assessment . . . . . . . . . . . . . . .

Contents

3. Vector Calculus 3.1. Chapter Goals . . . . . . . . . . . . . . . 3.2. Basic 3-Dimensional Calculus . . . . . . 3.2.1. Dierential Element of a Line . . 3.2.2. Line Integral . . . . . . . . . . . 3.2.3. Dierential Element of a Surface 3.2.4. Surface Integral . . . . . . . . . 3.2.5. The Volume Integral . . . . . . . 3.3. Dierential Calculus Concepts . . . . . 3.3.1. The Del or Nabla Operator . . . 3.3.2. Gradient . . . . . . . . . . . . . . 3.3.3. The Curl . . . . . . . . . . . . . . 3.3.4. Divergence . . . . . . . . . . . . 3.4. Maxwells Equations . . . . . . . . . . . 3.5. Units and Dimensions of EM Fields . . 3.6. List of Formulae . . . . . . . . . . . . . . 3.7. Practice Problems and Self Assessment . 79 79 79 79 83 88 91 94 96 99 101 104 110 112 113 114 117

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

II. Electrostatics

4. The Electric Field and Gausss Law 4.1. Chapter Goals . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Electrostatics: An Introduction . . . . . . . . . . . . . . . . 4.3. Charge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Coulombs Law and the Electric Field . . . . . . . . . . . . 4.5. The Electric Field due to a System of Point Charges . . . . 4.5.1. Electric Dipole . . . . . . . . . . . . . . . . . . . . . 4.5.2. Electric Field Due to Any Number of Point Charges 4.6. Electric Field due to Continuous Charge Distributions . . 4.6.1. Innite Line Charge . . . . . . . . . . . . . . . . . . 4.6.2. Innite Sheet Charge . . . . . . . . . . . . . . . . . . 4.7. Electric Displacement and Flux Density D. . . . . . . . . 4.8. Gausss Law . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.9. Gausss Law Applied to Cases of Spherical Symmetry . . . 4.9.1. Gausss Law Applied to a Point Charge . . . . . . . 4.9.2. Gausss Law Applied to a Charged Sphere . . . . . 4.10. Gausss Law Applied to Cases of Cylindrical Symmetry . . 4.11. Gausss Law Applied to Cases of Rectangular Symmetry . 4.12. List of Formulae . . . . . . . . . . . . . . . . . . . . . . . . . 4.13. Practice problems and Self Assessment . . . . . . . . . . . . 5. Energy and Potential 5.1. Chapter Goals . . . . . . . . . . . . . . . . . . . . . . 5.2. Potential Due to a Point Charge . . . . . . . . . . . . 5.3. Equipotential Surfaces . . . . . . . . . . . . . . . . . 5.4. Potential Energy . . . . . . . . . . . . . . . . . . . . . 5.5. Potential Due to a System of Point Charges . . . . . 5.6. Potential Due Any Continuous Charge Distribution 5.7. List of Formulae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123

124 124 124 125 129 139 139 143 145 147 152 156 157 161 161 163 166 169 170 172 178 178 178 185 187 193 195 203

vi

Contents

5.8. Practice Problems and Self Assessment . . . . . . . . . . . . . . . 206 6. The Electric Field and Material Media 6.1. Chapter Goals . . . . . . . . . . . . . . . . . . 6.2. Current and Current Density . . . . . . . . . 6.3. Continuity Equation . . . . . . . . . . . . . . 6.4. Conductors, Semiconductors and Dielectrics 6.4.1. Conductors and Resistance . . . . . . 6.4.2. Relaxation Time for Conductors . . . 6.4.3. The Method of Images . . . . . . . . . 6.4.4. Semiconductors . . . . . . . . . . . . . 6.4.5. Dielectrics . . . . . . . . . . . . . . . . 6.5. Capacitance . . . . . . . . . . . . . . . . . . . 6.5.1. Parallel Plate Capacitor . . . . . . . . 6.5.2. Coaxial Line . . . . . . . . . . . . . . . 6.5.3. Two Conductor Line . . . . . . . . . . 6.6. Relation Between Capacitance and Resistance 6.7. Boundary Conditions for Electrostatic Fields 6.8. Energy Stored in the Electric Field . . . . . . 6.9. List of Formulae . . . . . . . . . . . . . . . . . 6.10. Practice Problems and Self Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 212 212 217 219 221 223 228 230 232 233 234 236 238 241 243 248 252 256 264 264 265 266 266 271 271 274 280 281 283

7. Laplaces and Poissons Equations 7.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2. Uniqueness Theorem . . . . . . . . . . . . . . . . . . . . . 7.3. Laplaces Equation . . . . . . . . . . . . . . . . . . . . . . 7.3.1. Some One Dimensional Solutions . . . . . . . . . 7.3.2. Two Dimensional Solutions to Laplaces Equation 7.3.2.1. Analytic Functions . . . . . . . . . . . . . 7.3.3. Separation of Variables . . . . . . . . . . . . . . . . 7.3.4. Numerical Techniques . . . . . . . . . . . . . . . . 7.4. Poissons Equation . . . . . . . . . . . . . . . . . . . . . . 7.4.1. One Dimensional Solutions . . . . . . . . . . . . .

III. Magnetostatics

8. The Magnetic FieldAmperes Law 8.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 8.2. The Biot-Savart Law . . . . . . . . . . . . . . . . . . . 8.2.1. Types of Current . . . . . . . . . . . . . . . . . 8.3. Amperes Law . . . . . . . . . . . . . . . . . . . . . . . 8.3.1. Amperes Law Applied to a Thin Straight Wire 8.4. Amperes Law Applied to a Wire of Radius a . . . . . 8.5. Amperes Law Applied to an Innite Solenoid . . . . 8.6. The Magnetic FieldRigorous Calculations . . . . . 8.6.1. Magnetic Field of a Straight Wire . . . . . . . . 8.6.2. Loop of Wire Carrying a Current . . . . . . . . 8.6.3. Magnetic Field Due to a Current Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

284

285 285 285 289 291 293 296 299 303 303 307 308

vii

Contents

9. The Vector Potential 9.1. The Magnetic Flux Density . . . . . . . . . 9.2. The Biot-Savart Law . . . . . . . . . . . . 9.3. Various Results . . . . . . . . . . . . . . . 9.3.1. Current Carrying Conductor . . . 9.3.2. Two Current Carrying Conductors 9.4. Far Field Approximation . . . . . . . . . . 9.4.1. Square Current Loop . . . . . . . . 312 312 316 317 317 317 318 320

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

10. Magnetic Forces 323 10.1. The Lorentz Force . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 10.2. Electron Moving in a Steady Magnetic Field . . . . . . . . . . . . 324 10.3. A Straight Wire Carrying a Current in a Magnetic Field . . . . . 326 10.4. Other Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . 328 10.5. Loop Carrying a Current in a Constant Magnetic Field . . . . . 328 10.6. Torque on Loop Carrying a Current in a Constant Magnetic Field 329 10.7. Force between Two Current Elements . . . . . . . . . . . . . . . 331 11. Inductance, Magnetisation and Magnetic Circuits 11.1. Inductance . . . . . . . . . . . . . . . . . . . . . 11.1.1. Inductance of a Coil . . . . . . . . . . . 11.1.2. Inductance of a Coaxial Line . . . . . . 11.1.3. Magnetic Energy . . . . . . . . . . . . . 11.1.4. Inductance of a Circular Loop . . . . . 11.1.5. Mutual inductance . . . . . . . . . . . . 11.2. Magnetic Materials and Magnetic Circuits . . . 11.2.1. Magnetisation . . . . . . . . . . . . . . . 11.2.2. Magnetic Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333 333 334 335 336 338 342 344 344 346

12. Time Dependant Fields 12.1. Chapter Goals . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2. List of Formulae . . . . . . . . . . . . . . . . . . . . . . . . . 12.3. Faradays Law . . . . . . . . . . . . . . . . . . . . . . . . . . 12.4. A Maxwell Equation from Faradays Law . . . . . . . . . . 12.5. The Displacement Current Density . . . . . . . . . . . . . . 12.6. Time-Dependent Maxwells Equations . . . . . . . . . . . . 12.6.1. Point form of the Equations . . . . . . . . . . . . . . 12.7. Integral Form of Maxwells Equations . . . . . . . . . . . . 12.8. The Fundamental Equations of Radiation and Propagation 12.9. Time Domain Wave Equation . . . . . . . . . . . . . . . . . 12.10. requency Domain Wave Equation . . . . . . . . . . . . . . F 12.10.1.Phasors . . . . . . . . . . . . . . . . . . . . . . . . . . 12.11. he Wave Equation . . . . . . . . . . . . . . . . . . . . . . . T 12.12. hapter Summary . . . . . . . . . . . . . . . . . . . . . . . . C 12.13. hort Answer Questions . . . . . . . . . . . . . . . . . . . . S 12.14. roblems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

350

351 351 351 353 357 358 361 361 362 362 363 369 369 372 375 375 376

viii

Contents

13. Electromagnetic Waves 13.1. Uniform Plane Wave . . . . . . . . . . . . . 13.2. Wave Polarisation . . . . . . . . . . . . . . . 13.2.1. Circular Polarisation . . . . . . . . . 13.2.2. Elliptical Polarisation . . . . . . . . . 13.3. Wave Propagation in Conducting Media . 13.3.1. Low Conductivity Materials . . . . . 13.3.2. High Conductivity Materials . . . . 13.4. Boundary Conditions . . . . . . . . . . . . . 13.5. Reection and Refraction of Waves . . . . . 13.5.1. Reection from a Metal Surface . . . 13.5.2. Refraction from a Dielectric Surface 13.6. Poynting Vector and the Flow of Power . . 13.6.1. Poyntings Theorem . . . . . . . . . 13.6.2. Poynting Vector . . . . . . . . . . . . 379 379 386 386 389 391 396 398 400 404 404 407 409 410 413 417 418 421 425 427 427 432 435 437 446 451 452 452 458 459 467 472 472 476 485 492 495 499 502 502 502 504 506 506 508 508

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14. Transmission Lines 14.1. Time Domain Equation . . . . . . . . . . . . . . . . . 14.2. Frequency Domain Equation . . . . . . . . . . . . . 14.3. Solutions to the Transmission Line Equation . . . . 14.3.1. Power Considerations . . . . . . . . . . . . . 14.3.2. Reections from Discontinuities . . . . . . . 14.3.3. Standing Wave Ratio . . . . . . . . . . . . . . 14.3.4. Input Impedance Anywhere Along the Line 14.4. Transmission Line Charts . . . . . . . . . . . . . . . 14.5. Transformer Matching . . . . . . . . . . . . . . . . . 14.6. References . . . . . . . . . . . . . . . . . . . . . . . . 15. Waveguides 15.1. The Parallel Plate Waveguide 15.2. TEM mode Waveguides . . . 15.3. The Rectangular Waveguide . 15.4. The Circular Waveguide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

16. Radiation from Currents 16.1. Wave Equation due to Charges and Currents . . . . 16.2. Radiation from a Current Element . . . . . . . . . . 16.3. The Half-Wave Dipole Antenna . . . . . . . . . . . . 16.4. Basic Antenna Concepts . . . . . . . . . . . . . . . . 16.5. Directivity . . . . . . . . . . . . . . . . . . . . . . . . 16.6. Eective Aperture and Friis Transmission Formula 17. Introduction to Antennas 17.1. Chapter Goals . . . . . . . . . . . . 17.2. Introduction . . . . . . . . . . . . . 17.3. Linear Antenna Arrays . . . . . . . 17.4. Linear Array with Equal Currents 17.4.1. The Array Factor . . . . . . 17.4.2. Nulls and Sidelobes . . . . 17.4.3. Beam Pointing Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Contents

17.5. Fareld Pattern . . . . . . . . . . . . . . 17.6. Aperture Antennas . . . . . . . . . . . . 17.7. Horn Antennas . . . . . . . . . . . . . . 17.7.1. Introduction . . . . . . . . . . . . 17.8. Parabolic Reector . . . . . . . . . . . . 17.9. List of Formulae . . . . . . . . . . . . . . 17.10. ractice Problems and Self Assessment . P . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513 515 518 518 520 521 521 522 522 524 525 533 533 535 539 540 542 542 542 547 548 550 551

18. Radio Wave Propagation 18.1. Introduction . . . . . . . . . . . . . . . . . . . . . 18.2. Ground Wave Propagation . . . . . . . . . . . . . 18.3. Earth Reection . . . . . . . . . . . . . . . . . . . 18.4. The Surface Wave . . . . . . . . . . . . . . . . . . 18.4.1. The Surface Wave for the Vertical Dipole 18.4.2. Wave Tilt of the Surface Wave . . . . . . . 18.5. Surface Wave for a Horizontal Dipole . . . . . . 18.6. Approximations for Ground Wave Propagation . 18.7. Tropospheric Propagation . . . . . . . . . . . . . 18.7.1. Spherical Earth Considerations . . . . . . 18.7.2. Tropospheric Waves . . . . . . . . . . . . 18.8. Ionospheric Propagation . . . . . . . . . . . . . . 18.8.1. The Ionosphere . . . . . . . . . . . . . . . 18.8.1.1. Plasma Oscillations . . . . . . . 18.8.1.2. Wave Propagation in a Plasma .

A. List of Symbols 557 A.1. Commonly Use Symbols and Nomenclature . . . . . . . . . . . 557 B. Coordinate Systems B.1. Rectangular to Cylindrical, Cylindrical to Rectangular . . . . . B.2. Rectangular to Spherical, Spherical to Rectangular . . . . . . . . B.3. Spherical and Cylindrical Coordinates . . . . . . . . . . . . . . . B.4. Grad, Div, Curl and Laplacian in Dierent Coordinate Systems B.4.1. Cartesian Coordinate . . . . . . . . . . . . . . . . . . . . B.4.2. Cylindrical Coordinates . . . . . . . . . . . . . . . . . . . B.4.3. Spherical Coordinates . . . . . . . . . . . . . . . . . . . . C. Mathematical Reference C.1. General . . . . . . . . . . . . . . . C.1.1. Important Constants . . . C.1.2. Taylors Series Expansion C.2. Vector Identities . . . . . . . . . . C.2.1. General . . . . . . . . . . . C.2.2. Gradient . . . . . . . . . . C.2.3. Curl . . . . . . . . . . . . . C.2.4. Divergence . . . . . . . . C.2.5. Double . . . . . . . . . . . C.3. Complex Variables . . . . . . . . C.3.1. General . . . . . . . . . . . C.3.2. Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564 564 565 565 566 566 566 567 568 568 568 568 568 568 568 568 569 569 569 569 569

Contents

C.3.3. Complex conjugates . . . . . . . . . . C.3.4. Eulers Identity . . . . . . . . . . . . . C.4. Trigonometry . . . . . . . . . . . . . . . . . . C.4.1. Basic formulae . . . . . . . . . . . . . C.4.2. Sum and dierence formulae . . . . . C.4.3. Double angle formulae . . . . . . . . . C.4.4. Half angle formulae . . . . . . . . . . C.4.5. Product to sum formulae . . . . . . . C.4.6. Sum and dierence to product . . . . C.4.7. Triangle Formulae . . . . . . . . . . . C.4.8. Powers of the trigonometric functions C.5. Dierentiation . . . . . . . . . . . . . . . . . . C.5.1. Rules . . . . . . . . . . . . . . . . . . . C.5.2. Dierentiation of Functions . . . . . . C.6. Integration . . . . . . . . . . . . . . . . . . . . C.6.1. Common Substitutions . . . . . . . . . C.6.2. Indenite Integrals . . . . . . . . . . . Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570 570 570 570 570 570 571 571 571 571 571 571 572 572 573 573 573 576

xi

List of Figures

1.1. Figure for example 1.4 . . . . . . . . . . . . . . . . . . . . . . 1.2. Figure of vector showing its head and tail . . . . . . . . . . . 1.3. The rectangular coordinate system . . . . . . . . . . . . . . . 1.4. Unit vectors in the rectangular coordinate system . . . . . . 1.5. The force shown in rectangular coordinates . . . . . . . . . . 1.6. Figure showing vector addition and subtraction . . . . . . . 1.7. Addition of two vectors . . . . . . . . . . . . . . . . . . . . . 1.8. Figure for Example . . . . . . . . . . . . . . . . . . . . . . . . 1.9. Dot Product between two vectors . . . . . . . . . . . . . . . . 1.10. Work done and dot product . . . . . . . . . . . . . . . . . . . 1.11. Figure showing the vectors resulting from the cross product 1.12. Right hand thumb rule and the cross product . . . . . . . . . 1.13. Figure to calculate the cross products between unit vectors . 1.14. A method to memorise the calculation of the cross product . 1.15. Figure showing a spinning object . . . . . . . . . . . . . . . . 2.1. 2.2. 2.3. 2.4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 13 14 14 15 16 18 19 21 22 24 25 26 28 29 45 45 48 49 51 53 55 56 57 60 60 64 65 66 68 71 80 81 82 83 86 88

Figure showing the coordinates for the pressure scalar eld . . . The scalar eld g(r) for Earth . . . . . . . . . . . . . . . . . . . . Depiction of right- and left-handed coordinate systems . . . . . To calculate the dierence vector between two points in rectangular coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. A straight line in rectangular coordinates . . . . . . . . . . . . . 2.6. Figure used to calculate the vector equation of a straight line . . 2.7. Figure to calculate the equation of a plane . . . . . . . . . . . . . 2.8. 3D plot of the plane of the example . . . . . . . . . . . . . . . . . 2.9. The cylindrical coordinate system . . . . . . . . . . . . . . . . . . 2.10. a , a , and az . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.11. (ax , a y ) and (a , a ) of rectangular and cylindrical coordinates respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.12. Surfaces in cylindrical coordinates. . . . . . . . . . . . . . . . . . 2.13. The helix of the example . . . . . . . . . . . . . . . . . . . . . . . 2.14. The spherical coordinate system. . . . . . . . . . . . . . . . . . . 2.15. Relationship between (a , az ) and (ar , a ) of the cylindrical and spherical coordinate systems respectively . . . . . . . . . . . . . 2.16. Coordinate surfaces in spherical coordinates . . . . . . . . . . . 3.1. 3.2. 3.3. 3.4. 3.5. 3.6. Dierential element of a line . . . . . . . . . . . . . . . Dierential linear elements in cylindrical coordinates Dierential linear elements in spherical coordinates . Figure illustrating the scalar line integral . . . . . . . Figure illustrating the line integral . . . . . . . . . . . The dierential surface element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xii

List of Figures

3.7. Dierential surface elements in rectangular Coordinates . . . . . 3.8. Dierential surface elements in cylindrical coordinates . . . . . 3.9. Dierential surface element in spherical coordinates of the surface of a sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.10. Surface element lying on the surface of a cone . . . . . . . . . . . 3.11. Figure showing the calculation of the scalar surface integral . . 3.12. Figure showing the calculation of the vector surface integral . . 3.13. Dierential Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.14. The volume element in cylindrical coordinates . . . . . . . . . . 3.15. The volume element in spherical coordinates . . . . . . . . . . . 3.16. Details of the calculation of the volume integral . . . . . . . . . 3.17. Figure showing two neighbouring points . . . . . . . . . . . . . 3.18. shown on the surface (x, y, z) = (x0 , y0 , z0 ) . . . . . . . . . 3.19. Properties of the curl. (a) When the surface enclosed is at. (b) When the surface enclosed is bulging . . . . . . . . . . . . . . . 3.20. Properties of the curl. When the surface enclosed is closed . . . 3.21. Divergence Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 89 89 90 90 91 92 92 94 95 95 97 102 105 109 111

4.1. Charge distributions (a) Point Charge (b) Line Charge (c) Surface Charge (d) Volume Charge . . . . . . . . . . . . . . . . . . . . . . 125 4.2. Coulombs Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 4.3. Figure illustrating Coulombs Law . . . . . . . . . . . . . . . . . 130 4.4. Figure for Example 4.6. . . . . . . . . . . . . . . . . . . . . . . . . 132 4.5. Exners goldleaf electroscope . . . . . . . . . . . . . . . . . . . . 133 4.6. Coulombs law appicable to three charges . . . . . . . . . . . . . 134 4.7. Electric eld at an arbitrary eld point due to a point charge . . 136 4.8. Force felt on a charge q due to an external eld E(x, y, z). F = qE. 137 4.9. The E eld due to a point charge. . . . . . . . . . . . . . . . . . . 137 4.10. Figure to analyse a dipole . . . . . . . . . . . . . . . . . . . . . . 140 4.11. Field plot for the dipole. d = 1 m . . . . . . . . . . . . . . . . . . 142 4.12. eld plot for two equal but similar charges of magnitude Q and d= 1 m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 4.13. Electric eld due to a system of charges. . . . . . . . . . . . . . . 143 4.14. Charges placed at the corner of an equilateral triangle . . . . . . 144 4.15. The minuscule electric eld, dE produced by a minuscule charge dQ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 4.16. Figure showing a line of charge of densityl C/m from to along the z axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 4.17. Field plots for the innite line charge: Streamline plot of the innite line charge in the xy plane . . . . . . . . . . . . . . . . . 149 4.18. Field plots for the innite line charge: Streamline plot for the innite line charge in a plane passing through the z axis . . . . . 149 4.19. Computation of the electric eld on the axis of a short line charge 150 4.20. Computation of the electric eld on the axis of a ring of charge . 151 4.21. Plot of the E eld along the axis of a ring of charge . . . . . . . . 152 4.22. Calculation of the Electric eld due to an innite sheet of charge 153 4.23. Two sheet charges placed at z = d/2 . . . . . . . . . . . . . . . 154 4.24. Charged disk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 4.25. A at ring of charge . . . . . . . . . . . . . . . . . . . . . . . . . . 156 4.26. Faradays concentric spheres . . . . . . . . . . . . . . . . . . . . 156

xiii

List of Figures

4.27. Gausss Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.28. The relationship of ux to charge . . . . . . . . . . . . . . . . . . 4.29. Gausss Law applied to a point charge . . . . . . . . . . . . . . . 4.30. Applying Gausss Law to a uniformly charged hollow sphere with radius R0 . The Gaussian surface is a sphere with radius r0 . 4.31. Uniformly charged sphere . . . . . . . . . . . . . . . . . . . . . . 4.32. Plot of Dr versus r for a uniformly charged sphere . . . . . . . . 4.33. Gausss Law applied to an innite line charge . . . . . . . . . . . 4.34. An innite cylindrical hollow tube of radius 0 with surface charge s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.35. Gausss law applied to an innite sheet of charge . . . . . . . . 4.36. Diagram for Problem 5 . . . . . . . . . . . . . . . . . . . . . . . . 5.1. 5.2. 5.3. 5.4. 5.5. 5.6. 158 160 161 163 165 166 167 168 169 177

Potential dierence for a point charge . . . . . . . . . . . . . . . 180 (x) for Example 5.5 . . . . . . . . . . . . . . . . . . . . . . . . . . 183 Potential due to a point charge placed at r . . . . . . . . . . . . 184 The surface V = x2 y2 . . . . . . . . . . . . . . . . . . . . . . . . 187 Work done when moving Qt towards Q . . . . . . . . . . . . . . 188 [1, 0, 0] m moving with a velocity of 6107ax m/sec in the presence of q = 3 106C at [0, 0, 0] . . . . . . . . . . . . . . . . . . . . . . 189 5.7. Electron microscope picture of a spider (photo taken taken from a website of University of Minnesota, http://umn.edu, 2010) . . 191 5.8. Figure to calculate the potential for a system of point charges . . 193 5.9. A dipole aligned along the z axis . . . . . . . . . . . . . . . . . . 194 5.10. dV(r, r ) due to a minuscule charge dQ(r ) . . . . . . . . . . . . . 196 5.11. Potential calculation for a volume charge distribution . . . . . . 197 5.12. Calculation of the potential for a pair of parallel charged lines with charge l and l . . . . . . . . . . . . . . . . . . . . . . . . . 198 5.13. Lines of constant electric potential for a pair of innite line charges200 5.14. Electric eld plot of the electric eld for two parallel innite line charges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200 5.15. Charged ring and electron. . . . . . . . . . . . . . . . . . . . . . . 201 5.16. Calculation of potential for Problem 4 . . . . . . . . . . . . . . . 209 5.17. Charged disk for OBEQ Problem 4 . . . . . . . . . . . . . . . . . 211 6.1. 6.2. 6.3. 6.4. 6.5. Denition of current . . . . . . . . . . . . . . . . . . . . . . . . . The current and J. . . . . . . . . . . . . . . . . . . . . . . . . . . . Concepts concerning charge transport . . . . . . . . . . . . . . . Figure illustrating the continuity equation . . . . . . . . . . . . . The energy levels of the outermost shell of materials. The gure shows the valence and conduction band in (a) Metals, (b) Semiconductors, and (c) Dielectrics . . . . . . . . . . . . . . . . 6.6. Ohms law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7. A non-uniform resistor . . . . . . . . . . . . . . . . . . . . . . . . 6.8. The motion of electrons under the inuence of an external eld 6.9. Conductor in the presence of an external eld . . . . . . . . . . . 6.10. A charge Q enclosed by a spherical shell . . . . . . . . . . . . . . 6.11. A point charge near an innite ground plane . . . . . . . . . . . 6.12. Method of images applied to a single charge and a sphere. . . . 213 214 215 217 220 220 222 224 226 227 228 229

xiv

List of Figures

6.13. Polarisation of a single molecule under the inuence of an external eld. (a) molecule when the eld is absent (b) molecule when the eld is present . . . . . . . . . . . . . . . . . . . . . . . 232 6.14. Two metal bodies representing a capacitor . . . . . . . . . . . . . 234 6.15. Geometry and elds of the parallel plate capacitor (a)3-D view (b) Cross- section (c) Rough sketch of the electric eld . . . . . . 234 6.16. Capacitance of a coaxial line . . . . . . . . . . . . . . . . . . . . . 236 6.17. Geometry of the two conductor line . . . . . . . . . . . . . . . . 238 6.18. A charge Q enclosed by a spherical shell . . . . . . . . . . . . . . 240 6.19. Construction of the Leyden jar . . . . . . . . . . . . . . . . . . . 241 6.20. A coaxial resistor . . . . . . . . . . . . . . . . . . . . . . . . . . . 242 6.21. The behaviour of the electric eld near a dielectric boundary . . 243 6.22. Behaviour of the electric eld near dielectric-metal boundary . . 246 6.23. Electric eld in the presence of a dielectric occupying a half-space247 6.24. Moving charges into a region of space . . . . . . . . . . . . . . . 249 6.25. Figure for Problem 17 . . . . . . . . . . . . . . . . . . . . . . . . . 258 6.26. Measurement of for a liquid conductor. . . . . . . . . . . . . . 258 6.27. A capacitor with two dielectrics . . . . . . . . . . . . . . . . . . . 259 6.28. Circuit for Short Answer Question 1 . . . . . . . . . . . . . . . . 259 6.29. Three layer dielectric construction . . . . . . . . . . . . . . . . . 261 6.30. Concentric spheres with dielectric 0 r (r) . . . . . . . . . . . . . 261 7.1. Laplaces equation applied to two innite plates . . . . . . . . . 267 7.2. Figure for Exercise 7.3 . . . . . . . . . . . . . . . . . . . . . . . . 270 7.3. Contour plots of the real (sin(x) cosh(y)) and imaginary (cos(x) sinh(y)) parts of the function sin(z): Real part of sin z . . . . . . . . . . . 272 7.4. Contour plots of the real (sin(x) cosh(y)) and imaginary (cos(x) sinh(y)) parts of the function sin(z): Imaginary part of sin z . . . . . . . . 273 7.5. The real and imaginary parts of the function f (z) = z2 . . . . . . 273 7.6. Contour plot of u = x2 y2 along with the electric eld superimposed on the potential eld. . . . . . . . . . . . . . . . . . . . . . 274 7.7. Laplaces equation applied to a rectangular region . . . . . . . . 277 7.8. Grid of a region of space where Laplaces has to be solved. . . . 280 7.9. The grid method applied to a rectangular domain of width of 6 cm and height 7 cm. . . . . . . . . . . . . . . . . . . . . . . . . . 282 8.1. 8.2. 8.3. 8.4. 8.5. 8.6. 8.7. Figure illustrating the Biot-Savart law . . . . . . . . . . . . . . . z-directed lamentary current at the origin . . . . . . . . . . . . Types of Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The direction of the magnetic eld vis-a-vis a straight conductor carrying current . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.8. Straight wire of radius a carrying a current I . . . . . . . . . . . 8.9. The innite helix . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.11. A torus with a winding . . . . . . . . . . . . . . . . . . . . . . . . 8.12. Applying the Biot-Savart law to a current carrying straight conductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 287 290 291 292 293 295 297 299 300 302 303

xv

List of Figures

8.13. Field lines of the magnetic eld for the straight wire carrying current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 8.14. Magnetic eld on the central axis due to a current loop . . . . . 307 8.15. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 9.1. Magnetic Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2. Figure showing the geometry of how the vector potential calculated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3. Figure to derive the far eld vector potential . . . . . . . . . 9.4. Magnetic eld due to a square loop of current. . . . . . . . . . . is . . . . . . 313 315 319 320 323 324 327 329 330 331 334 335 336 339 340 343 344 346 347 348 348 353 354 355 357 359 363 368 370 375 376 377 379 384 385 387

10.1. The v B force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2. Single electron moving perpendicularly to a steady magnetic eld. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3. A long straight wire in a steady magnetic eld . . . . . . . . . . 10.4. A loop carrying current in a constant magnetic eld . . . . . . . 10.5. Torque on a loop carrying a current . . . . . . . . . . . . . . . . . 10.6. Calculation of the force between two current elements . . . . . . 11.1. Area of integration for calculating the ux linked by a single turn, m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2. Inductance of a coaxial line . . . . . . . . . . . . . . . . . . . . . 11.3. Calculation of inductance from the eld point of view . . . . . . 11.4. A circular loop of wire carrying a current I . . . . . . . . . . . . 11.5. Ax and A y integrands for R = 2, r = 1, = /2 . . . . . . . . . . . 11.6. Mutual inductance of two coils . . . . . . . . . . . . . . . . . . . 11.7. Magnetic moment denition . . . . . . . . . . . . . . . . . . . . . 11.8. A coil wound round a core . . . . . . . . . . . . . . . . . . . . . . 11.9. A magnetic circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.10. agnetic circuit with an air-gap . . . . . . . . . . . . . . . . . . . M 11.11. quivalent circuit of a magnetic core with an air-gap . . . . . . . E 12.1. Stokess theorem applied to E = B/t . . . . . . . . . . . . 12.2. Induced current in a time-varying magnetic eld. . . . . . . . . 12.3. Faradays law applied to generators . . . . . . . . . . . . . . . . 12.4. Ignition system of a car . . . . . . . . . . . . . . . . . . . . . . . . 12.5. Setup to show how the displacement current comes into play . 12.6. Typical communication setup with transmitting and receiving antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.7. A travelling wave . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.8. Showing the relation between time-domain and frequency domain entities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.9. Rotating hoop immersed in a constant B eld. . . . . . . . . . . 12.10. oving magnet and stationary coil . . . . . . . . . . . . . . . . . M 12.11. oving slider on rails . . . . . . . . . . . . . . . . . . . . . . . . M 13.1. The geometry of the uniform plane wave . . . . . . . 13.2. Electric and Magnetic elds of a uniform plane wave 13.3. The electromagnetic frequency spectrum . . . . . . . 13.4. Figure illustrating left circular polarisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xvi

List of Figures

13.5. Figure showing the advance of the wave in a left circular polarisation plane wave . . . . . . . . . . . . . . . . . . . . . . . . . . 13.6. Linear and elliptical polarisations . . . . . . . . . . . . . . . . . . 13.7. The case for = /6, E y0 = 2 and Ex0 = 1.5 . . . . . . . . . . . . . 13.8. Loss tangent for a dielectric . . . . . . . . . . . . . . . . . . . . . 13.9. Prole of a wave propagating from air to a conducting medium 13.10. kin depth for copper as a function of frequency . . . . . . . . . S 13.11. he Behaviour of electromagnetic elds near a boundary conT sisting of a change of medium . . . . . . . . . . . . . . . . . . . . 13.12. wave obliquely incident from air on a metal ( ) . . . . . A 13.13. wave obliquely incident from air (1 ) on a dielectric (2 ) . . . A 13.14. igure Illustrating Poyntings Theorem . . . . . . . . . . . . . . F 13.15. oynting theorem applied to the case of a wire carrying a steady P current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

388 389 392 393 394 400 401 405 407 412 413

14.1. The equivalent circuit of a transmission line . . . . . . . . . . . . 418 14.2. Examples of Transmission lines . . . . . . . . . . . . . . . . . . . 418 14.3. Figure to analyse a transmission line . . . . . . . . . . . . . . . . 418 14.4. Cross-section of the micro strip line . . . . . . . . . . . . . . . . . 425 14.5. Transmission line showing the forward and reverse voltage waves.425 14.6. Transmission line with a load impedance . . . . . . . . . . . . . 428 14.7. A plot of the magnitude of the voltage along a line for VL = 1 V and ZL /Z0 = 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432 14.8. A plot of the magnitude of normalised current along the line for IL = 0.5 A and ZL /Z0 = 2 . . . . . . . . . . . . . . . . . . . . . . . 435 14.9. Input impedance of a transmission line. . . . . . . . . . . . . . . 435 14.10. shown in the complex plane . . . . . . . . . . . . . . . . . . . . 437 14.11. plane with r = 1, r = 0, x = 1 and x = 0 circles . . . . . . . . . . 439 14.12. he Smith chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440 T 14.13. mith chart depiction for this example . . . . . . . . . . . . . . . 441 S 14.14. mith chart showing how to match a load . . . . . . . . . . . . . 443 S 14.15. ircuit for the example being considered . . . . . . . . . . . . . 445 C 14.16. atching the line with a parallel stub . . . . . . . . . . . . . . . 446 M 14.17. atching with a transformer . . . . . . . . . . . . . . . . . . . . 447 M 15.1. The parallel plate waveguide . . . . . . . . . . . . . 15.2. Fields for the parallel plate waveguide. (a) E y (b) Hz 15.3. The parallel plate transmission line . . . . . . . . . . 15.4. 3 dimensional view of a rectangular waveguide . . 15.5. The circular waveguide . . . . . . . . . . . . . . . . . 15.6. The rst few Bessel functions of the rst kind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452 456 459 460 467 469 474 475 476 478 481 483 484

16.1. Radiation from a current source . . . . . . . . . . . . . . . . . . . 16.2. A current carrying conductor . . . . . . . . . . . . . . . . . . . . 16.3. An elemental wire carrying a current . . . . . . . . . . . . . . . . 16.4. The relationship between Az , A and Ar . . . . . . . . . . . . . . 16.5. Direction of E and H shown for a spherical wave . . . . . . . 16.6. Polar plot of the normalised power radiated by an innitesimal current element. . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.7. Equivalent circuit of the half wave dipole . . . . . . . . . . . . .

xvii

List of Figures

16.8. The dipole antenna. (a) close-up details (b) Details with respect to the far eld . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.9. Showing the approximation |R R | r z cos . . . . . . . . . 16.10. agnitude of the electric eld radiated by a half-wave dipole, M shown as a polar plot. Only half the plot is shown. . . . . . . . . 16.11. ormalised power pattern of a half-wave dipole as a function N of . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.12. xamples of antennas . . . . . . . . . . . . . . . . . . . . . . . . . E 16.13. ar elds of an antenna . . . . . . . . . . . . . . . . . . . . . . . . F 16.14. he power pattern of an antenna . . . . . . . . . . . . . . . . . . T 16.15. ower pattern of the same antenna shown in 3-D rectangular P coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.16. enition of a solid angle . . . . . . . . . . . . . . . . . . . . . . D 16.17. eciprocity property of antennas . . . . . . . . . . . . . . . . . . R 16.18. erivation of the Friis transmission formula . . . . . . . . . . . D 17.1. Design criteria of antennas. . . . . . . . . . . . . . . . . . . . . . 17.2. An equi-spaced linear array . . . . . . . . . . . . . . . . . . . . . 17.3. (Top gure) Plot of sin(N/2)/ sin(/2) for N = 10 (Bottom gure) Numerator and Denominator of Equation 17.12 . . . . . . . 17.4. 3-D view of a broadside pattern, where the antenna is oriented vertically. Note the dierence in the scale of the z-axis which has been broadened to show greater particulars of the main beam and sidelobe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.5. Pointing the main beam in dierent directions. (a) 0 = /2, = 0 (b) 0 = /4, = 0.7071kd (c) 0 = 0, = kd. For all these plots, kd = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.6. Figure to calculate the fareld pattern of current sources . . . . 17.7. Plot of normalised E versus for various values of kl . . . . . . 17.8. Wave propagation based on Huygens principle . . . . . . . . . 17.9. An aperture with elds shown . . . . . . . . . . . . . . . . . . . 17.10. etal sheet with rectangular aperture . . . . . . . . . . . . . . . M 17.11. xamples of horn antennas . . . . . . . . . . . . . . . . . . . . . E 17.12. arabolic reector . . . . . . . . . . . . . . . . . . . . . . . . . . . P 17.13. araboloid of revolution . . . . . . . . . . . . . . . . . . . . . . . P

485 486 489 492 493 494 494 495 496 499 500 503 503 507

511 512 513 515 515 516 517 519 520 521

18.1. Radio wave propagation paths over the earth. . . . . . . . . . . 522 18.2. Two antennas of heights h1 and h2 in communication . . . . . . 524 18.3. Consideration of a wave obliquely incident on a dielectric interface.526 18.4. Magnitude and phase of the reection coecient, R , of a plane wave whose E eld is perpendicular to the plane of incidence. r = 15; = 12 103; x = 18 103/ fMHz . is the approach of the wave above the horizon. . . . . . . . . . . . . . . . . . . . . . 528 18.5. Magnitude and phase of the reection coecient, R , of a plane wave whose E eld is parallel to the plane of incidence. r = 15; = 12 103. is the approach of the wave above the horizon. . 529 18.6. Magnitude and phase of the reection coecient, R|| , of a plane wave whose E eld is parallel to the plane of incidence. r = 15; = 12 103; x = 18 103/ fMHz . is the angle of grazing incidence of the wave above the horizon. . . . . . . . . . . . . . 531

xviii

List of Figures

18.7. Magnitude and phase of the reection coecient, R|| , of a plane wave whose E eld is perpendicular to the plane of incidence. r = 15; = 12 103. is the approach of the wave above the horizon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.8. Coordinate system for the surface wave. . . . . . . . . . . . . . . 18.9. (Upper graph) Magnitude of the factor a. p = (R/)a; (Lower graph) The phase constant b. For these graphs, r = 15, = .012 18.10. raph of the approximate value of the ground wave attenuation G factor A versus the numerical distance p for various values of b. 18.11. ave tilt for a surface wave . . . . . . . . . . . . . . . . . . . . . W 18.12. ave tilt for a surface wave-calculation for three frequencies: W 0.5, 10 and 100 MHz. r = 15, = 0.012. . . . . . . . . . . . . . . 18.13. pherical and plane earth . . . . . . . . . . . . . . . . . . . . . . S 18.14. ending of rays in the troposphere . . . . . . . . . . . . . . . . . B 18.15. igure to calculate the curvature of rays in the troposphere . . . F 18.16. onversion of curved paths to straight paths using an eective C radius for the earth . . . . . . . . . . . . . . . . . . . . . . . . . . 18.17. ertical composition of the ionosphere . . . . . . . . . . . . . . . V 18.18. lasma Oscillations of an innite slab . . . . . . . . . . . . . . . P 18.19. high frequency wave travelling in the ionosphere . . . . . . . A 18.20. iagram for fMUF max . . . . . . . . . . . . . . . . . . . . . . . . D

532 534 536 537 537 539 542 543 544 545 549 550 554 556

xix

List of Tables

1.1. The basic SI units . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Cross products of the cylindrical coordinate system . . . . . . . 2.2. Table of dot products between ax , a y , az and a , a , az . . . . . . 2.3. Table of dot products in Spherical Coordinates . . . . . . . . . . 7 59 62 69

3.1. Table showing the units of the various electromagnetic quantities 114 3.2. Summary of Properties of Grad, Div and Curl . . . . . . . . . . 116 13.1. IEEE microwave band designations . . . . . . . . . . . . . . . . 386

14.1. Calculation of L and C for two wire (conductor radius=a, spacing between centres=b) and the coaxial (radius of inner conductor=a, inner radius of outer conductor=b) lines. . . . . . . . . . . . . . 424 15.1. Characteristic impedance of a Microstrip line . . . . . . . . . . . 459 18.1. Typical relative dielectric constants of various geological materials comprising the surface of the earth Davis & Annan (1980) . 527 18.2. Table showing the daytime properties of the ionospheric layers. In all the layers there is a great day-night variation of the electron density. Note that on the ground, the pressure is about 1000 mB. (The pressure is measure of the number of molecules/cc) . . . . 549

xx

Students need to refer to some material quite often in the course of reading the text. I have found that students are more likely to refer to reference material in the opening chapter of a book rather than an Appendix. Therefore some material is included here.

Symbol c e g G 0 Z0 h k me me mn mp 0 NA rB Name Velocity of light Charge of an electron Acceleration due to gravity Newtons gravitational constant Permittivity of vacuum Characteristic impedance of free space Plancks constant Boltzmanns constant Electron mass Electron mass Neutron mass Proton mass Permeability of free space Avogadros constant Bohr radius Value 2.9979 109 (ms1 ) 1.6022 1019 (C) 9.8067 (ms2 ) 6.6726 1011 (N m2 kg2 ) 8.8542 1012 (Fm1 ) 376.99 () 6.6261 1034 (J s) 1.3807 1023 (J K1 ) 9.1094 1031 (kg) 9.1094 1031 (kg) 1.6749 1027 (kg) 1.6726 1027 (kg) 1.2566 105 (Hm1 ) 6.022 1023 (mole1 ) 0.53 1010 (m)

0.2. Units

Base quantity length mass time electric current thermodynamic temperature amount of substance luminous intensity Name meter kilogramme second ampere kelvin mole candela Symbol m kg s A K mole cd

, A, alpha , Z, zeta , , lambda , R, rho , X, chi , varrho , B, beta , H, eta , M, mu , , sigma , , psi (si) , , gamma , , , theta , N, nu , T, tau , , omega , , delta , I, iota , , xi (zi) , , upsilon , , E epsilon , K, kappa , , , pi , , phi

0.4. SI Prexes

Symbol a f p n m d Name atto femto pico nano micro milli deca 10n 1018 1015 1012 109 106 103 101 Symbol h k M G T P E Name hecto kilo mega giga tera peta exa 10n 102 103 106 109 1012 1015 1018

Material Air Alumina Bakelite Fused quartz Fused silica (glass) Gallium Arsenide (GaAs) Germanium Glass Ice (pure distilled water) Paper Paran Polystyrene Sea Water Silicon Soil Teon Vacuum (free space) Vinyl Water Wood r 1.00054 9.6-10 3.7 3.8 3.8 13.1 16 4 - 10 3.2- 4.15 3.0 2-3 2.55 80 11.7 - 12.9 2.55-2.59 2.1 1.00000 2.8 - 4.5 80-88 1.2 - 2.1 Loss tangent 0.0002 0.00006-0.0002 0.0016 0.0009-0.12 0.0001-.0003 4-5 0.0017-0.0062 0.00028 0.04-0.15 0.03-0.04

Material(2001 n.d.) Electrical Steel Mu-metal Permalloy Ferrite (NiZn) Ferrite (MnZn) Steel r 4000 20,000 8000 6-650 >650 700 Material Nickel Platinum Alluminum Vacuum Copper Water r 100 1.0 1.0 1.0 1.0 1.0

Part I.

Introductory Material

Advice is seldom welcome; and those who want it the most always want it the least. Earl of Chesterfield

This chapter introduces the student to the very, very basic math involved in taking a course in electromagnetics. In particular the topics covered are 1. Manipulating scalars with special emphasis on units. 2. Doing calculations and the special care required in doing them. 3. Estimating the result of a calculation with emphasis on the order of magnitude of the result of those calculations. 4. Understanding the basic concept of a vector and the unit vector. 5. Addition and subtraction of two vectors. 6. Scalar and vector products of two vectors. 7. Units and dimensions of engineering quantities.

1.2. Introduction

Electromagnetics is a subject which is today being rediscovered and applied in an interdisciplinary sense to many areas of engineering which include wireless and wire-line transmission and communication, circuits, computer interconnects, optical ber links and components, antennas, plasmas, wave propagation in the ionosphere, lasers and many others. Concerning the historical aspects of electromagnetics, the bending of rays of light in dielectrics, the law of reection and the existence of metal mirrors was known in antiquity. The use of loadstone in compasses was well-known to the Chinese and convex lenses have been discovered in Carthage (present day Tunesia). Though investigation of electromagnetic phenomena started many centuries ago1 the bulk of the laws of electromagnetic theory were discovered in the eighteenth and nineteenth century. Some of the names of investigators of that time include the names of Count Alessandro Giuseppe Volta (1745-1827), Andre Marie Ampere (1775-1836), Michael Faraday (1791-1867), James Clerk Maxwell (1831-1879), Heinrich Hertz (1857-1894), and many others.

1 For

example, William Gilbert (1544-1603), conducted experiments for 18 years with magnetic materials, wrote his book De Magnete.

Did you know? Faraday, who was the son of a blacksmith, was educated only upto the school level. (He only attended day school till the age of thirteen). Even though he had this limited academic background, he was considered one of the greatest experimentalist of his time due to his path-breaking work: his far reaching contributions to electromagnetism and electrochemistry. His work is of such a great deal of historical signicance that a unit of charge and a unit of capacitance has been named after him. This period was followed by intense research into the engineering applications of the theory in the twentieth century. This was especially true during the second world war, when electromagnetic theory received a sudden burst of attention with applications in radar and communications. After WW II most of the focus of the top universities and corporations was in this area and a huge volume of research was published and made available. Fundamental to the study of electromagnetism is the study and use of vector and scalar elds. Before a study of the subject is undertaken, the student must acquire a good knowledge of vector analysis. Part I gives a basic introduction to the mathematics required and also a general introduction. The student is also advised to consult a standard text as supplementary reading {See, for example Kreyszig (2003) Spiegel (1974)} Before the study of this subject is commenced, the student is encouraged to do undertake a simple survey or project where he nds the answers to the following questions 1. What will I study in this subject? 2. Where will I apply it? 3. Of what use is it? Electromagnetics is applied in military defense applications, high-speed electronics, ultrahigh-speed photonic integrated circuits, microcavity laser design, light switching in femtoseconds and imaging of the human body. Furthemore, microwave circuits typically process bandpass signals at frequencies above 3 GHz. Common circuit features include microstrip transmission lines, directional couplers, circulators, lters, matching networks, and individual transistors. Circuit operation is fundamentally based upon electromagnetic wave phenomena... Digital circuits typically process low-pass pulses having clock rates below 2 GHz. Typical circuits include densely packed, multiple planes of metal traces providing ow paths for the signals, dc power feeds, and ground returns. Via pins provide electrical connections between the planes. Circuit operation is nominally not based upon electromagnetic wave eects...2 Once these questions are answered the student will know the goals which are to be attained. In discussions with students I found that many found this subject far too dicult to grasp due to the underlying mathematical complexity. My advice to the reader who shares the same view is 1. That due attention must be paid to the mathematical basics which have been explained in this Part, after which the student will nd that electromagnetic theory may prove to be less intractable.

2 Allen

Taove, Why Study Electromagnetics: The First Unit in an Undergraduate Electromagnetics Course

Table 1.1.: The basic SI units

Base quantity length mass time electric current thermodynamic temperature amount of substance luminous intensity

Symbol m kg s A K mole cd

2. Whenever any new equation or concept is studied, it should be immediately applied to some thought situation. Any simple problem may be conjured up and the equation or concept should then be applied to it. 3. Problems at the end of each chapter are given for the reader to solve. After solving these problems he or she will then truly grasp the essential concepts. Apart from the introductory chapters on mathematics, the subject is introduced in a phased manner: Electrostaticsthe study of the electric eld produced by static charges; Magnetostaticsthe study of the magnetic eld produced by steady currents; Radiation and Propagationthe study the radiation and propagation of electromagnetic waves. In the last Part an introduction to antennas is also included.

1.3. Scalars

As has often3 been quotedmathematics is the language of science. As engineers, wherever we nd that we need to apply the results of any subject, we rst need to describe it in general terms through equations and then proceed those equations to concrete situations. These statements apply to all engineering sciences, and in this particular case, to electromagnetics. In our applications we need to use real numbers which are called scalars. As the reader would be well aware, a scalar is a real number (and sometimes a complex one) which describes a physical quantity. 10.2 kg describes the mass of something. That something may be sand or milk or gas. 32.4 million metric tonnes may describe the mass of rice produced by some country. 22.4 litres of gas at STP describes the volume occupied by one mole of a gas at STP. As the student probably already knows, 10.2 kg, 32.4 million metric tonnes and 22.4 litres, are scalars. The units used in this book are the SI units and are shown in Table 1.1. Scalars are manipulated according to well known rules: Rule 1. Scalars a, b, . . . may be added together to yield a third scalar. For example c = a+b

3

(1.1)

It is essential to note some points If a is 20 grams of sand and is added to b equal to 30 grams of sand then the addition gives us c equal to 50 grams of sand. The word grams implies that the scalars possess the same units. So the rule is: before adding two numbers make sure that they both have the same units. In this case the units are grams. So, 20 grams of sand cannot be added to 22.4 litres of gas, even though both are scalars. Similarly 20 grams of some material cannot be added to 0.5 pounds of the same material (even though both are masses) because both these numbers belong to dierent systems of units . The rule is that when adding scalars a, b, c, . . . attention must be paid to the fact that all those scalars must be of the same type4 and, furthermore, all the scalars must be quantied in the same system of units, for example (g) or (kg) or (l), etc. For a quick glance at the SI system of units Barrow (1966) the reader can look at the reference material included in the start of the book.5 It is because of this that units and dimensions play a very important role in engineering. Therefore before adding two numbers the student must ensure that the two numbers must not only be of the same type but they also must have the same units. Simply stated, grams must be added to grams, pounds to pounds, etc. The most important point here is to be vigilantly aware of what one is doing with numbers. Rule 2. Scalars can multiplied by other scalars or real numbers. But here, even though the two scalars do not necessarily have to be of the same type (i.e. having the same units and dimensions) they must belong to the same system of units. i.e., the British system, or the cgs system or the SI system, and so on. c = ab Here a can be equal to 22.4 (dimensionless) and b = 1 litre. So c = 22.4 litres a can be 10 meters/sec, b, 10 sec making c equal to 100 meters. An important point to be noted is that not only the two scalars get multiplied, but also the two units do. In the rst example c has the dimensions of litres since: (no units) (litres) gives litres. In the second case (meters/sec) (sec) is meters. Rule 3. Subtraction can be tricky. The rst point is that when subtracting numbers the guidelines given in rule 1 must be followed. The second point is that in mathematics negative numbers may be allowed, but in engineering one must be clear about the physical interpretation of a negative number. Thus: c = ab (1.2)

(1.3)

is an equation which makes sense in mathematics, but in engineering: if a = 10 litres and b = 20 litres then what is the meaning of c = 10 litres?

4 i.e., 5 On

all masses or all volumes etc. the Internet the reader may refer to the URL http://physics.nist.gov/cuu/Units/units.html.

It is obvious that when any answer with a negative sign occurs in the equations the answer is admissible if only the negative answer has physical meaning. Rule 4. Division of scalars can be done irrespective of the units of the two quantities involved in the division but the units of the result is equal to the division of the units of the two numbers. If c = a/b (1.4)

and a is 10 meters and b is 2 seconds then c is 5 meters/second. Students make mistakes of various kinds when doing scalar manipulation and calculations. When doing a calculation it is important to keep in mind the order of magnitude6 of various terms. It is better to do a calculation with only two or three decimal places of accuracy and get the correct result rather than doing a calculation with eight or nine decimal places and obtaining an answer which is completely o the mark. . Sometimes students punch in wrong numbers into their calculators and as a result get completely wrong results. It is important to keep a strict watch on the numbers as they are crunched, as the following example shows. EXAMPLE 1.1 10 lbs of a substance A of volume 1.233 l is added to 9.221 kg of a substance B of volume 2555 cc. If on adding the two substances there is a contraction of volume by 12%, what is the density of the material at the end? Solution: To start with do a preliminary computation with only one decimal point (or only whole numbers) to get an idea of what the answer is. Step 1. To solve this problem, we need to quantify everything to a set of common units. Hence 10 lbs = 10 lbs = 4.546 kg ? 2.2 lbs/kg

Is this answer reasonable? Yes because 10 rounded o is 10, and 2.2 rounded o is 2, hence 10/2 5 which is close to 4.546. Step 2. Next we add the two masses Total mass = 4.546 + 9.221 = 14.767 kg ? The answer is not reasonable, because 5+9 is 14, but 14.767 rounded o is 15. Something is wrong. Doing the calculation again Total mass = 4.546 + 9.221 = 13.767 kg Step 3. To calculate the total volume in litres. All values must be in litres 1.233 + .2555 = 1.4885 l Is this okay? Look at things carefully. 2555 cc is 2.555 litres, not 0.2555 l. So re-calculating 1.233 + 2.555 = 3.788 l

6

The answer seems reasonable. 1+3=4. Going to tenths, 0.2+0.5 is 0.7. Yes it seems okay. Step 4. Calculating the 12% contraction, 10% of 3.788 l is .3788 l. The answer should be around .3788 l 12 % of 3.788 = 0.4546 0.1 of 3.7 is 0.37; 0.02 is around 0.07. So .37+.07=0.44, which is close to the answer given above. So the answer is reasonable. We now calculate the contraction in the volume. 3.788 0.4546 = 3.3335 l 3.8-0.5 is 3.3. The answer seems okay. Step 5. Calculating the density Density = 13.767 kg = 4.13 kg/l 3.3335 l

14/3=4.66. Possibility of an error. Recalculate. We get the same result. This is the answer. EXERCISE 1.1 Follow the procedure given in the previous example to do the following question. A car races around a circular wedge which consists of a circular arc connected by two radii to the centre of the circle. The angle subtended by the arc is 40 . The length of the arc is 10 miles, and the car takes one hour to complete one circuit. What is the speed of the car in km/hr? Ans. 61.95 km/hr EXAMPLE 1.2 If the circumference of a circle is 1600 m. Estimate the diameter. What is the maximum and minimum value of the diameter? Compute these values then proceed to do an accurate calculation. It is important to get some idea of the answer before doing an accurate calculation. Solution: Step 1. If d is the diameter of a circle and c the circumference then d = c Step 2. We must get some idea of the value of the diameter. To estimate the diameter, since 3 4 the diameter lies between 1600 1600 (= 400 m) d (= 533.3 m) 4 3 We can say that the value of the diameter will be close to 533 m since 3. Step 3. Doing an accurate calculation d= 1600 = 509.3 m

10

Did you know? Archimedes (287-212 BCE ?) A student of Euclid, was one of the greatest Greek mathematicians of his time. He did a great deal of work in the elds of mathematics and engineering. One of his coontributions is the approximation to the calculation of . He concluded that 3 10/71 < < 3 1/7. EXERCISE 1.2 If the major and minor axes, of an ellipse, a = 20 m and b = 10 m estimate the area. Ans. 150 < A < 200; A = 157.07 m2 EXAMPLE 1.3 Multiply two complex numbers. a = 3 + j2 b = 15 + j3 before the multiplication, the two numbers have to be converted into polar form, and then multiplied. How will you estimate the polar forms? Solution: Step 1. Estimating the magnitudes of a and b and then calculating the exact values: |a| = 9 + 4 = 13 4 |a| = 9 + 4 = 13 = 3.606 (exact value) |b| = And the phases: a = tan1 a = tan1 2 45 3 2 = 33.7 3 3 0 15 3 = 11.31 15 ab 6045 225 + 9 225 = 15 |b| = 225 + 9 = 15.297 (exact value)

(exact value)

(exact value)

Step 2. Now that we know the approximate value, we proceed to calculate the exact value of the multiplication. The exact value of ab = 55.15445.

11

EXAMPLE 1.4 Solution: How will you estimate the result of the sum 1 ? n2 n=1 To do this problem, rst spend some time thinking about ways and means to proceed along a quick way to do an estimation of this innite sum. Think of what is similar to summation. Even draw a graph of the various terms.

1.2

0.8

0.6

0.4

0.2

Step 1. First write out the numbers. 1, 0.25, 0.11, etc. Seeing the analogy between 1/n2 and 1/x2 is it possible to estimate the series from the integral of 1/x2 ? Step 2. To solve this problem we study Figure 1.1. The gure shows a graph of the function 1/x2 . Three rectangles have been drawn. The three rectangles hve areas 1, 1/22 = 0.25, and 1/32 = 0.111. Notice that the sum of the area of the rectangles is greater than the area under the curve. After the third term the area of the rectangles and that of the area under the curve become comparable. Step 3. Therefore the sum of the series can therefore be approximated by S 1 + .25 + .111 + Shaded area the shaded area is given by

4

dx = 0.25 x2

hence the sum can be approximated as 1 1.611 n2 The actual answer is 2 /6 = 1.645. The approximation is only 2% lower than the correct answer.

12

Head

Tail

Figure 1.2.: Figure of vector showing its head and tail

1.4. Vectors

One of the fundamental mathematical entities used in the study of electromagnetic theory, are vectors (See Spiegel (1974)). Vectors are dened to be entities which possess both magnitude (like scalars) and direction. The depiction of these entities is typically in the form of directed line segments, as shown in Figure 1.2. Did you know? The parallelogram law for the addition of vectors is probably part of the work of Aristotle (384322 B.C.?), a Greek philosopher. This law is formulated in the book Mechanics of Heron of Alexandria around rst century A.D. and is mentioned as a corollary in the famous Principia Mathematica by Isaac Newton. Though Newton dealt almost wholly with forces and velocities (which are vectors), in the Principia, he never formally proposed the concept of a vector. The systematic study of vectors was carried in the 19th and early 20th centuries. The gure shows a line segment shaped like an arrow, labelled A with a tail and a head. The tail-head combination describes a direction while the length of the arrow gives the magnitude of the vector. This is the popular conception of vectors, that they are directed line segments. This statement is true and yet untrue. Vectors are abstract entities whose representation takes the form of directed line segments. Take the concept of force. As we all know, that a force cannot be seen nor heard nor smelt, but only its eects can be felt. Thus fundamentally vectors are abstract entities. Just like you cannot see the number 1, similarly vectors cannot be seen. But their mathematical manipulation leads to real results which can be corroborated by experiment. Next we consider the rudiments of the rectangular coordinate system in 3-space with which, it is hoped, that the reader is familiar. The rectangular coordinate system is shown in Figure 1.3. A point in 3D-space consists of three numbers (x, y, z) which correspond to the three distances cut o by perpendiculars from the point in question to the three axes: x, y and z. This is the position vector of the point and it is the vector joining the origin to the point under consideration. The unit vector in rectangular coordinates in the x-direction is ax . Similarly we can specify unit vectors a y and az in the yand z-directions respectively. Referring to Figure 1.3, the position vector of the point (1, 2, 2) is notationally the vector

13

z

y 1 2

z

Arbitrary Point

r(1, 2, 2) (ax + 2a y + 2az) The magnitude of r r is given by the well-known distance formula: r= 12 + 22 + 22 = 3

(1.5)

(1.6)

and it is the distance of the point (1, 2, 2) from the origin. Before we proceed any further it is important to point out the nature of the unit vectors in the rectangular coordinate system: (i) The unit vectors are constants. i.e. their directions are constant, no matter which point in 3-space is chosen. (ii) The unit vectors are orthonormal7 . This is shown in Figure 1.4 . That is, the vectors are perpendicular to each other and their magnitude is unity. Any vector, specied in rectangular coordinates, has the representation: A = Ax a x + A y a y + Az a z

7A

(1.7)

set of orthonormal vectors have the property that they are perpendicular to each other, and each of the vectors of the orthonormal set has unit magnitude.

14

2 1 2

Where Ax , A y and Az are real or complex numbers. With this notation, the force F = ax + 2a y + 2az (Nt), is shown in Figure 1.5. Thus we can see that Ax Fx = 1 N, A y F y = 2 N and Az Fz = 2 N. That is, the unit vector ax is multiplied by 1, a y is multiplied by 2 and az is multiplied by 2. Or in other words the component of the force in the ax direction is 1, that in the a y direction is 2 and so on. It is clear from the above discussion that scalar multiplication of vectors is the natural outcome of the denition of unit vectors. This is so because unit vectors have to be multiplied by real numbers to give us vectors which are directed in any general direction. It is also clear that unit vectors are dimensionless, but the scalars have dimensions. Another way to corroborate this statement is by observing that if a vector A is given, then the unit vector in the direction of A is A = A/|A| (1.8)

Where8 the hat notation is used to denote a general unit vector,9 and |A| A is the magnitude or length of A. |A| = A2 + A2 + A2 x y z (1.9)

In this text, A may often be written instead of |A| when there is no ambiguity. For example for the force which we just considered, the magnitude of the force is F = 12 + 22 + 22 = 3 while the unit vector is F = F/F = 1ax + 2a y + 2az /3

8 The

9 Not

reader should prove that this is indeed a unit vector included are the unit vectors of coordinate systems

15

d

Figure 1.6.: Figure showing vector addition and subtraction

EXERCISE 1.3 A vector A = 5ax + 5a y + 5az is given. Find A and A. Ans. A = 5 3 and A = ax + a y + az / 3.

Vectors can be manipulated according to well-dened rules. Examples of operations between scalars are +, , and . The simplest operation is the scalar multiplication of a vector. Thus if A is a vector, then 5A is another vector, whose length (or more accurately, magnitude) is multiplied by 5, but its direction is unchanged10 . In rectangular coordinates if A Ax , A y , Az then 5A 5Ax , 5A y , 5Az (1.10)

A particularly important binary operation between two or more vectors is vector addition. Vectors can be added. Here the concept of + (vector addition) is dierent than in the case of scalars. To form A + B, The vector B is translated parallel to itself and the tail of B is attached to the head of A as shown in the Figure 1.6. The vector equation reads: C = A+B (1.11)

We can perform the operation the other way, attaching the tail of A to the head of B and get the same result.(This is not shown in the gure, but the student can easily prove it for himself). This last statement says that vector addition is commutative: C = A+B = B+A (1.12)

10 The

16

that is whether A is added to B or B is added to A, it gives us the same result. Thus if A = 2ax + 3a y + 4az and B = 5ax + 6a y + 7az then A + B = 2ax + 3a y + 4az + 5ax + 6a y + 7az = 7ax + 9a y + 11az and B + A = 5ax + 6a y + 7az + 2ax + 3a y + 4az = 7ax + 9a y + 11az Vectors are associative as well.For three vectors A, B and C : (A + B) + C = B + (A + C) (1.13)

The equation says that the order of addition is unimportant because it leads to the same result. That is if we add A to B rst and add the result to C, we get the same result when we add A to C and then add the result to B. This property is analogous to associativity in scalars.To get an idea of the correct nature of this result, choose any three arbitrary vectors and apply the previous equation. Similarly one can subtract vectors. B is B reversed in direction (and this is just another vector) and so B is added to A: C = A + (B) = A B (1.14)

This is also shown in Figure 1.6. It must be remembered however that vectors which are added or subtracted must of the same type, just as with scalars.

Whenever we look at a polygon of vectors as in addition or subtraction we can easily compute the nal result by using a certain technique which is explained below. On examining Figure 1.6 on the preceding page we can start from any point, say a, and go to any other neighbouring point connected by a vector, say point d. Going from a to d the vector involved is B and since we are going in the direction of the arrow we write this vector as it is: (B) From point d we go to point o going against the direction of the arrow, so we change the sign of the vector and add it to the previous vector: (B) (A B) Note the change in sign of the second term. From o let us go back to a, hence we add A to the previous sum. Since we have come back to the starting point we make the sum of the terms equal to zero: (B) (A B) + A = 0 This is a valid vector equation. We can apply this technique to any other polygon of vectors, as long as we reach the starting point. Applying the

17

(a)

(b)

technique to the circuit: o-d-a-c-o 11 the vector equation becomes: (A B) (B) + (B) (A + B) = 0 The reader may be wondering how this technique helps formulation of vector equations? Well, let us apply this technique to the circuit o-a-c and let call the vector o-c C. Then we nd that C = A+B EXAMPLE 1.5 Let us add two vectors A of 10 N to B of 15 N at an angle of 30 to A as shown in the Figure 1.7 and let the result be denoted as C. Find the magnitude and direction of C Solution: Method (a) Step 1. First draw a sketch of the two vectors in the notebook. [See Figure 1.7(a)] Step 2. From Figure 1.7 we can see that vectors A,B and C form a triangle. We calculate the length of the third side based on the law of cosines. The third side gives us the magnitude of C: |C| = we nd that |C| = 24.18 N, Step 3. If we apply the law of sines to the triangle again the angle which C makes with the horizontal is 18.010 sin sin 150 = |B| |C| C(= sin ) = 18.01 Method (b) Step 1. Draw a sketch as shown in Figure 1.7(b) |A|2 + |B|2 2 |A| |B| cos(180 30)

11 o-d (A B);

18

Step 2. Since the second method is to decompose the vector B into two perpendicular vectors, the vector parallel to the horizontal is given by Bx = 15 cos(30 )ax N. = 12.93ax Nt The vector in the vertical direction is given by B y = 15 sin(30 )a y N. = 7.5a y Nt Step 3. We add Bx to A giving Cx and C y = B y Cx = 22.93ax N C y = B y = 7.5a y N C = |C| = 22.932 + 7.52 = 24.18 N, which gives us the same answer as before. The angle that C makes with the horizontal is given by arctan(C y /Cx ) = 18.01 EXAMPLE 1.6 Add two vectors A and B such that the angle from A to B is 150 , A = 10 newtons and B = 20 newtons. (See Figure 1.8) Solution: This example shows us how to split a vector into two perpendicular components, one along the direction of a second vector and the other component perpendicular to it. Step 1. Referring to a sketch (Figure 1.8), B is split into two parts one along A: B

to A

19

and the other perpendicular to A B to A = B sin A = 10 A Nt = BB

to A

Where A is the unit vector in the direction of A and A is the unit vector perpendicular to A but lying in the plane enclosed by B and A. Step 2. Since A+B = C adding B|| to A we get C = (10 17.32)A = 7.32A C = 10A Hence C = 7.32A + 10AN

EXAMPLE 1.7 Find the dierence of the two vectors: A B given that A = 1ax + 3a y + 5az N and B = 5a y N in rectangular coordinates. Solution: Step 1. We have to calculate C = A B. With A = 1ax + 3a y + 5az N, and B = 0ax + 5a y + 0az N. Step 2. Doing the actual calculation, C = = = AB (1 0)ax + (3 5)a y + (5 0)az ax 2a y + 5ax

EXERCISE 1.4 Subtract B from A. The angle from A to B is 150 , A = 10 Nt and B = 20 N. Ans. Magnitude 29.09; 22.32 with A. EXERCISE 1.5 Find that vector, B, which when added to A = 30ax gives us C = 15ax + 15a y N Ans. 15ax + 15a y N. EXERCISE 1.6 If C = A + B, and A has a magnitude of 10, C also has a magnitude of 10 and is at an angle of 90 to A, nd B. Ans. B = 10 2; B = 135

20

(a)

(b)

There are two other important operations between vectors12 . The dot product (or scalar product between two vectors) which results in a scalar; and the vector product of two vectors resulting in another vector. Both will be considered in turn. The dot or scalar product is given by denition to be: A B = |A||B| cos (1.15)

Where is the operator representative of the scalar product and is the angle between A and B in accordance with Figure 1.9. Notice that though two vectors are involved in the product, the result is a scalar. From the denition it is clear that the dot product is commutative: A B = |A| |B| cos() = B A (1.16)

When we examine Figure 1.9 we can see that the product is positive for values of 00 || 900 and it is negative for 900 || 1800. When is zero, that is, when the vectors are in the same direction the magnitudes of the two vectors get multiplied. On the other hand, when the two vectors are opposite in direction then though the magnitudes of the two vectors get multiplied but the sign is negative. When the two vectors are perpendicular to each other ( = 900 ) then the scalar product is zero. From the denition of the scalar product (substituting A for B :) |A| = AA (1.17)

EXAMPLE 1.8 (a) Find the work done when a constant force of F of magnitude 100 N is applied on a mass as shown, moving it over a distance of 100 m. The force is applied at an angle of 300 with the horizontal. (b) Calculate the work done to move the mass by 100 m again, but here the magnitude of the force is varied with displacement according to the law: |F(x)| = 10x, N the direction remaining the same as before. (see Figure 1.10) Solution: Part (a)

12 SeeThomas &

Finney (1996)

21

100 meters

Step 1. We know that the work done by a force is a dot product: Fd where F is the force and d is the displacement vector. Step 2. So to work out this part we take the dot product of F with the displacement d = 100ax m. Note that F=100 N. W = work done = F d = F (100ax)

Part (b) Step 1. A innitesimal amount of work dW is equal to F dl where F is the force and dl is the innitesimal displacement. So W= Step 2. To be specic

x=100

F dl

W = work done =

x=0 x=100

F(x) dx

=

x=0

(10x) cos(300)dx

x=100

EXERCISE 1.7 (a) A force of 10y2 ax + a y N is applied to a mass to move it by a distance of 15 m in the a y direction starting from (0,0). Find the work done. (b) If the resistance from the surroundings is such that the mass does not move at all, what is the work done? Ans. (a) 11.25 kJ (b) 0.

22

EXAMPLE 1.9 Find the scalar products between all pairs of unit vectors belonging to the rectangular coordinate system, ax , a y , and az Solution: Step 1. We rst write all the possible dot products. These dot products are ax ax , ax a y , ax az , a y ax , a y a y , a y az , az ax , az a y and az az . Step 2. We know that (i) All the unit vectors are perpendicular to each other, and (ii) Each unit vector is of magnitude equal to 1. Step 3. From the denition of the scalar product (Equation 1.15) it is clear that for two vectors if cos = 0 (or = 90 ) the dot product is zero. Hence we can conclude that ax a y = ax az = 0;

a y ax = a y az = 0 az ax = az a y = 0

We equate all these products to zero on our list. Step 4. Since the magnitude of each of the the unit vectors (ax , ...) is 1 and taking the dot product of each unit vector with itself, we get ax ax = a y a y = az az = 1 In all these cases cos = 1, (because = 0.) Using the results of the last example we can nd the dot product between two vectors when they are specied in rectangular coordinates. If A = (Ax , A y , Az ) and B = (Bx , B y , Bz ) then A B = (Ax ax + A y a y + Az az ) (Bxax + B y a y + Bz az ) (1.18)

Since all dot products of the unit vectors with other unit vectors are zero and dot products of the unit vectors with themselves are 1 then

(Ax ax +A y a y +Az az )(Bx ax +B y a y +Bz az ) = Ax ax (Bx ax +B y a y +Bz az )+A y a y (Bx ax +B y a y +Bz az ) + Az az (Bx ax + B ya y + Bz az ) considering the right hand side term by term Ax ax (Bx ax + B y a y + Bz az ) = Ax ax Bx ax + Axax B y a y + Ax ax Bz az

= Ax B x A y a y (Bx ax + B y a y + Bz az ) = A y a y Bx ax + A y a y B y a y + A y a y Bz az = Ay By Az az (Bx ax + B y a y + Bz az ) = Az az Bx ax + Az az B y a y + Az az Bz az = Az B z

23

z

Figure 1.11.: Figure showing the vectors resulting from the cross product

Therefore A B = Ax B x + A y B y + Az B z In Equation 1.18 if we substitute ax for B then Ax = A a x Similarly Ay = A ay Az = A a z (1.21) (1.22) (1.20) (1.19)

Another binary operation used by mathematicians and which is useful in the study of natural phenomena is the cross product or vector product. The vector product involves two vectors and the result is a third vector. Thus (see Figure 1.11) in symbolic notation: C = AB where the magnitude of C is given by: |A B| = |A||B|sin (1.24) (1.23)

And the direction of C is given by the well-known right hand thumb rule. The right hand thumb rule states that the direction of the vector product is perpendicular to both A and B and is given by the direction of the thumb when the right hand is held in a position as it were holding an imaginary stick with the thumb along the direction of the stick. The hand is held in such a way that the ngers are curled from A to B. This can be seen from Figure 1.12. It is important to note that the cross product is anti-commutative. C = A B = B A (1.25)

24

Figure 1.12.: Right hand thumb rule and the cross product

EXAMPLE 1.10 Find the cross products between all unit vectors in rectangular coordinates: ax , a y , and az Solution: Step 1. As in the previous example, we need to nd the various products: ax ax , ax a y , ax az , a y ax, a y a y , a y az , az ax, az a y and az az . Step 2. We know that the cross product is proportional to sin , the angle between the two vectors. Based on this we can safely say that: ax ax = a y a y = az az = 0 In all these cases sin = 0 ( = 0.) Step 4. Also since the magnitudes of both ax and a y are both 1, therefore: ax a y = |ax | a y sin = 1 Step 5. Since ax a y has a magnitude 1 and is in the direction of az , we realize that we are talking about the unit vector az . ax a y = az Step 6. Since the cross product is anti-commutative a y ax = az (1.28) (1.27) Step 3. Further with reference to Figure 1.11, ax a y is in the direction of az . (1.26)

Similarly by this type of reasoning we can easily obtain the other products. Step 7. We can also memorise these relations by using Figure 1.13. If we go along the direction of the arrow, ax a y = az or az ax = a y . But if we go against the direction of the arrow, a negative sign is required: a y ax = az or ax az = a y . We now consider a very simple example based on the above solved example.

25

Figure 1.13.: Figure to calculate the cross products between unit vectors

EXAMPLE 1.11 Let A = 2ax + 4a y and B = ax + 7a y which lie on the x-y plane. Find A B. Corroborate that |A B| = AB sin and perpenicular to both A and B. Step 1. We write A B in terms of the unit vectors ax , ... etc. A B = (2ax + 4a y) (ax + 7a y) Step 2. We next multiply term by term = 2ax ax + 2ax 7a y + 4a y ax + 4a y 7a y Step 3. We use the results just obtained: ax ax = 0; ax a y = az , etc. = 0 + 14az + (4az) + 0 = 10az It is clear from this example that A B is perpendicular to both A and B and has a magnitude of 10. Step 4. To calculate the value of , the angle between A and B we proceed as follows, using the dot product. A B = (2ax + 4a y) (ax + 7a y) = 30 = AB cos with A= 22 + 42 = 20 B = 12 + 72 = 50

Step 5. We calculate next using the dot product formula = AB = arccos[(A B)/AB] = arccos(30/ 1000) = 18.41 Step 6. We now use the cross product formula. sin(18.41) = 0.3159 with A and B calculated earlier |A B| = AB sin = 50 20 0.3159 = 31.62 0.3159 = 10

26

Based on the previous example we are in a position to calculate the cross or vector product in rectangular coordinates. Let A = Ax ax + A y a y + Az az and B = Bx ax + B ya y + Bz az then A B = (Ax ax + A y a y + Az az ) (Bxax + B y a y + Bz az ) = Ax ax (Bxax + B y a y + Bz az ) +A y a y (Bxax + B y a y + Bz az ) +Az az (Bxax + B y a y + Bz az ) The rst term of Equation 1.29 reads: Ax ax (Bx ax + B ya y + Bz az ) The rst product, Ax Bx ax ax is zero. The second product: Ax ax B y a y is equal to Ax B y az and the third product: Ax ax Bz az equals Ax Bz a y . Hence the rst term is Ax ax (Bx ax + B ya y + Bz az ) = Ax B y az AxBz a y Similarly, the second and third terms respectively are: A y a y (Bxax + B y a y + Bz az ) = A y Bx az + A y Bz ax Az az (Bxax + B y a y + Bz az ) = Az Bx a y Az B y ax Finally collecting terms A B = (A y Bz B y Az )ax +(Az Bx Bz Ax )a y +(Ax B y BxA y )az The cross product may be written as A B = ax simplifying this becomes ax A B = Ax Bx ay Ay By az Az Bz Ay By Az Az + ay Bz Bz Ax Ax + az Bx Bx Ay By

(1.29)

(1.30)

(1.31)

To memorise these equations we examine Figure 1.14 which shows a triangle with its vertexes marked x, y and z in an anti-clockwise order. This gure is the same one used previously. If C = A B and we want to calculate Cx then we go to the point marked x and write Cx =

27

z

then we go towards y along the arrow and write A y . We have up to now: Cx = A y We continue on to the point marked z and multiply A y by Bz . We have so far: Cx = A y Bz Since we are at the point marked z we write Az . Cx = A y Bz Az From here we come back to point marked y (moving against the direction of the arrow) we Az multiply by B y . Cx = A y Bz Az B y Since we have moved against the direction of the arrow, a negative sign must be placed in front of this product. Cx = A y Bz Az B y We can continue to write the other components in the same manner. EXAMPLE 1.12 Find the vector product (A B) of two vectors in rectangular coordinates given by A = (4, 3, 0) and B = (3, 4, 0). Conrm the formula |A B| is given by ABsin Solution: Step 1. First we can see that both vectors lie in the x-y plane. This is so because both vectors have no z-component. The magnitudes of A and B are 42 + 32 = 5. B = 32 + 42 = 5

A=

Step 2. The dot product between A and B is used to calculate the angle between the two vectors: A B = 4 3 + 3 4 = 24 ABcos

28

x

Figure 1.15.: Figure showing a spinning object

where is the angle between the two vectors. Hence cos = 24/(AB) = 24/25 = arccos(24/25) = 16.26 Step 3. and we now use |A B| = AB sin : AB sin() = 5 5 0.28 = 7 The cross product has this magnitude and its direction is given by the right hand thumb rule and therefore perpendicular to both. (Figure 1.12) Using Equation(1.30) we get

Cz = (A B)z = Ax B y A y Bx = 4 4 3 3 = 7

Cx = (A B)x = A y Bz Az B y = 3 0 0 4 = 0 C y = (A B) y = Az Bx Ax Bz = 0 3 4 0 = 0

From these calculations it is clear that C is perpendicular to both A and B and has a magnitude of 7. Which conrms our result . EXERCISE 1.8 The vertices of a regular tetrahedron are O=(0,0,0), A=(0,2,0), B=( 3, 1,0) and C=(1/ 3,1,2 2/3). Find the unit vector perpendicular to face ABC. Ans. 0.41ax 0.71a y + 0.29az EXAMPLE 1.13 A spinning object, O, spinning at the rate of = 60 rpm is spinning about the vertical axis as shown. Find the velocity of any particle at any point r(x, y, z). Solution:

29

Step 1. First we must get to the basics of the problem. Characterising the rotation as a vector, the angular velocity may be written as: = 60az (rev/minute) = 2az (rad/sec) Step 2. The position vector of any point is given by r = xax + ya y + zaz . (m) Step 3. In scalar notation v = , but we are dealing with vectors so the correct formula is v = r (Check: v is perpendicular to both as well as r) v = r = (2az ) (xax + ya y + zaz ) = 2x (az ax) + 2y az a y + 2 (az az ) = 2(xa y yax) m/s Step 4. Verfy whether your answer is correct! Notice that the velocity of the particle is independent of the z-coordinate. The equation implies that all particles on a straight line parallel to the z-axis possess the same velocity, which from our intuition we can justify. Step 5. Satisfy yourself that the answer is correct (take the magnitude of both sides of the equation): |v| = 2 x2 + y2 = 2(= ) m/s

being the perpendicular distance of the point from the z-axis. This is also intuitively correct. Step 6. Consider special cases. Consider a point on the x-axis and nd out the direction of the velocity. If r = xax then v = 2xa y Note the direction which is correct. This is what we expect Another aspect of vectors which we must be sure to remember is that the cross product is not associative. for example (ax ax ) az = 0 az = 0 and ax (ax az ) = ax (a y) = az So clearly

30

ax (ax az )

A (B C)

(1.32)

We now take a look at the triple scalar product ABC which can be interpreted in only one way: the cross product has to be taken rst, and then only the dot product. Since ax Bx BC = Cx or B C = ax (B y Cz Bz C y ) + but A B C = Ax (B y Cz Bz C y ) + comparing the two previous equations, we can see that ax , a y and az has been replaced by Ax , A y and Az . Therefore Ax Bx Cx Ay By Cy Az Bz Cz ay By Cy az Bz Cz

ABC =

(1.33)

1.5.1. The Basis of Units and Dimensions

Units and dimensions play an important role in electromagnetic theory. Throughout this book, the fundamental units used will that of the SI units13 . These fundamental units are: meters (m), for the dimension of length [L], kilogramme (kg) for the dimension of mass [M], seconds (s), for the dimension of time [T] and coulombs (C), for the dimension of charge [Q]. With these units (and dimensions) we can derive the units and dimensions of any physical quantity14 . The basic units of the SI scheme are given in Table 1.1 on page 7.

13

14 Therefore

From the french, Le Systme International dUnits all other physical units are called derived units.

31

Did you know? There were many distinct units of measurement, such as meters, feet and yards. However, historically, all measurement required a standard, rough and ready, unit to begin which people could use in day to day living. For example, the length of a foot lead to the measure of a foot. A yard was the distance from the tip of the nose to the tip of the middle nger of the outstretched arm. Similarly, for the Romans, one thousand paces was a millia (Latin for a thousand) from where we get our mile. Let us start in a simple way and nd the unit and dimension of velocity. v = velocity = dr/dt (1.34)

Where r is the position vector of a particle in meters and t is time in seconds. Since we are dividing dr which is in meters, by dt which is in seconds, hence the units of velocity is meters (meters)(second)1 second and the dimension of velocity is [L][T]1. Observe that the unit and dimension of velocity is obtained through a formula. Similarly the unit of acceleration is derived from the formula: a = acceleration = d dr dv = dt dt dt (1.35)

Which is d/dt (units of s1 ) dr/dt (having the units of ms1 ), therefore the unit of acceleration is therefore s1 ms1 ms2 , and the dimension is [L][T]2 . Similarly from the formula F = ma (1.36) the unit of force is (kg)(m)(s)2 , (the student can work this out for himself). Due to the complexity of the unit of force, it is given a new name: newton (N). Incidentally the dimension of force is [M][L][T]2. In this way we can dene the units of any physical quantity. The really important point which I have noted in students studying electromagnetic theory is their confusion about units of various physical quantities and this confusion adds to their confusion about the subject in general. Generally, in electromagnetics, the units used for the various physical quantities are in terms of Coulomb (C), amperes (A C/s), volts (V), ohms (), farads (F) and henrys (H).15 In this book, when the physical quantity is discussed, then we shall also discuss its units. In the meanwhile, we will consider ampere and volt to be the more fundamental of these ve quantities, then from Ohms law V = IR The unit of ohms is volt/ampere.

15 It

(1.37)

is important for the student to remember that the short form of a unit is written in capital letters when the unit is named after a person. For example (C) or (N), but the unit itself is written in small letters: newton.

32

Let us look at the impedance equation Z = jL (1.38)

Where Z is the impedance in , is the radian frequency in rad/sec (which is essentially s1 , since radians is dimensionless) and L is the inductance in henry. Then unit of henry, (H) is /sec1 = sec, since radians have no dimensions. Finally we can derive the unit of farad from the admittance equation: Y = jC (1.39)

where Y is the admittance in and C is the capacitance in farad (F). The the unit of farad is then /sec1 = sec, and so on.

The position vector r of a point x, y, z in cartesian coordinates is r(x, y, z) xax + ya y + zaz Any vector A is given by A = Ax a x + A y a y + Az a z Its magnitude is given by |A| = A2 + A2 + A2 x y z

and its direction by the unit vector A = A/|A| When a vector is multiplied by a scalar (say by 5) then 5A 5Ax ax + 5A y a y + 5Az az When two vectors are added A + B = (Ax + Bx ) ax + A y + B y a y + (Az + Bz ) az Vector addition is commutative A+B = B+A Vector addition is associative (A + B) + C = A + (B + C)

33

Vector subtraction A B = (Ax Bx ) ax + A y B y a y + (Az Bz ) az The dot or scalar product is given by denition to be: A B = |A||B| cos Where is the operator representative of the scalar product and is the angle between A and B. Also A B = Ax B x + A y B y + Az B z The dot product is commutative AB = BA The vector product is a vector perpendicular to both vectors in the direction given by the right hand thumb rule |A B| = |A||B|sin Or ax A B = Ax Bx ay Ay By az Az Bz

A (B C) Ax A y Az A B C = Bx B y Bz Cx C y Cz

(A B) C

Chapter Summary

In this book the SI units will be used. (m), (kg), (A), (C) and ( K). (See Table 1.1) When adding or subtracting two numbers we must take care to make both numbers have the same units. That is, a kg b kg is ne; but not a kg b lbs. Also we must pay attention to look at the order of magnitude of both terms, and, most importantly, the order of magnitude of the result. That is 1 104 2 104 is 1 104.

When multiplying or dividing two numbers we must pay attention to the units of the two numbers: they must belong to the same system of

34

units. For example a (m) b (s) c (hours) is incorrect We must pay great attention to the order of magnitude of the numbers and that of the result. For example 1 104 2 104 is the order of 100 .

In any equation learn to check the units and dimensions of all terms. Learn to make sure that all the terms belong to the same system of

In particular:

units; Learn to look at the units and dimensions of the nal answer ;

Learn to approximate the nal answer and then only do the nal

calculations. Vectors are entities which possess both magnitude and direction. The position vector of the point (x, y, z) is notationally the vector r (xax + ya y + zaz) The magnitude of r r is given by the well-known distance formula: r= x 2 + y 2 + z2

and it is the distance of the point (x, y, z) from the origin. Any vector, specied in rectangular coordinates, has the representation: A = Ax a x + A y a y + Az a z Where Ax , A y and Az are real or complex numbers. The unit vector in the direction of A is A = A/|A| Where the hat notation is used to denote a general unit vector, and |A| A is the magnitude or length of A. |A| = A2 + A2 + A2 x y z

In this text, A may often be written instead of |A| Thus if A is a vector, then 5A is another vector A Ax , A y , Az then 5A 5Ax , 5A y , 5Az To form A + B, The vector B is translated parallel to itself and the tail of B is attached to the head of A. Then the vector from the tail of A to the head of B is C C = A+B

35

Addition is commutative and associative A+B = B+A (A + B) + C = A + (B + C) The vector -B is the vector B reversed in direction. In rectangular coordinates A + B = [Ax + Bx , A y + B y , Az + Bz ] The dot or scalar product is given by denition to be: A B = |A||B| cos = B A Where is the operator representative of the scalar product and is the angle between A and B. In rectangular coordinates A B = Ax B x + A y B y + Az B z From the denition of the scalar product (substituting A for B :) |A| = A A Work done is W is the work done, F is the force, d is the displacement. Or W expressed as an integral is W= F dl W = work done = F d

For i, j = x, y, z then ai a j = 1 if i = j. Otherwise it is zero. The vector or cross product is |C| = |A B| = |A||B|sin where is the angle from A to B and the direction of C is given by the right hand thumb rule (Refer Fig. 1.12). The cross product is ant-commutative: A B = B A The cross product can be obtained from: ax A B = Ax Bx ay Ay By az Az Bz

36

The cross product is not associative: (A B) C The scalar triple product is: ABC = Ax Bx Cx Ay By Cy Az Bz Cz A (B C)

Review Questions

1. Name two common applications of electromagnetic theory and write a short note (about 10 lines) on each application. 2. Why is the study of scalars operations imperative in engineering electromagnetics? In particular, why do we to pay attention to small details (like units and order of magnitude) in problem solving? 3. Dene a vector and a unit vector. 4. Why do we need to know about vector operations? Write down the rule for a) Scalar multiplication of a vector. Does scalar multiplication of a vector change the direction of the vector? b) Addition of two vectors. c) Sutraction of two vectors. 5. Why is the dot product important in vector operations? (Discuss with respect to work done) Why is it called scalar product? Discuss the sign of the dot product with respect to angle between the two vectors. Why is the dot product commutative based on the denition? 6. How do you arrive at the cross product between two vectors? Explain the right hand thumb rule. 7. When we take the cross product given in terms ax , a y and az explain how we obtain the cross product using the determinant notation. Using the same notation explain why the cross product is anti-commutative. 8. Explain how there are two types of units: fundamental and derived.

Problems

1. An electron moves with a velocity of 2 107m/s nd its kinetic energy. (a) First write the formula (b) then write the units of the K.E. (c) estimate the order of magnitude of the result and the answer (d) then use the calculator to calculate the answer. Follow this procedure every time you

37

do a calculation. This procedure seems very involved, but in the end it will save you a lot of trouble. Ans. (a) 1/2 me v2 (b) Joules (c) 18 1017 (J) (d) 1.8218 1016 (J)

2. How many meters are there in a mile? Follow some procedure to estimate and check your answer. Ans. 1609.3 m to a mile

(3/2)

3. Do two integrations (a) 0 sin(x)dx and (b) 0 cos(x)dx. How will you check whether your results are correct? Ans. (a) 2 (b) -1 4. If a ball is dropped from a certain height from rest and takes 1 s to hit the oor and then return to its initial position, estimate the velocity with which it hits the oor. Also estimate the height from which it is dropped. Think of various ways to do this estimation. Finally do an accurate calculation. Ans. 9.8 m/s; 4.9 m 5. Is the sum 1 n

n=1

convergent or divergent? Hint: See Ex. 1.4 Ans. Divergent. 6. In any calculation think before you leap do not blindly proceed to solve the given problem. Take a moment o and concentrate on the solution. What is the result of the integral

xex dx ?

7.

8.

9.

10.

This is an example of how some time spent initially on thinking of the solution will save you a lot of trouble. In simple integrations of this type a graph of the integrand is very helpful. This method of approach in the long run will be very benecial. Hint: What is the type of function which is being integrated? Ans. 0 Vector A has a magnitude of 1 m , B has a magnitude of 2 m at an angle of 30 to A. Find the magnitude and angle of the resultants A + B and A B. Ans. |A + B| = 2.909; (A + B) = 20.1 with A; |A B| = 1.239; (A B) = 233.8 with A. Vector A has a magnitude of 1 m, B has a magnitude of 2 m. The angle between A and B is not specied. Find the minimum and maximum value of |A + B| and |A B| for all possible angles. Ans. |A + B|max = 3.0; (A + B)max = 0 ; |A + B|min = 1; (A + B)min = 180 ; |A B|max = 3.0; (A B)max = 180 ; |A B|min = 1; (A B)min = 0 . A is vector of length 1 m. B is to A. (a) What should be the length of B to make |A + B|= 2 m? (b) |A B|= 2 m? Ans. (a) 3 (b) 3 Vector A has a magnitude of 1 m, B has a magnitude of 2 m. The angle

38

between A and B is not specied. Find the angle between A and B to make |A + B|= 1.5 m Ans. 46.56 with A. 11. Show that B is B reversed in direction, and then show that B + (B) = 0. 12. Find the position vectors of the two points P1 = (1, 1, 2), P2 = (2, 3, 2) and the vectors P1 P2 and P2 P1 . Ans. P1 = ax a y + 2az ; P2 = 2ax + 3a y + 2az ; vector P1 P2 = P2 P1 ; vector P2 P1 = P1 P2 . 13. Show that for two vectors A and B using the diagrammatic representation of vectors that A+B = B+A 14. Find the position vector of the mid point of P1 P2 where P1 and P2 are two points given in Problem 12. Let this point be Pm . Ans. Pm = (1/2)(P1 + P2)=1.5ax + a y + 2az . 15. Find the angles between the three vectors OP1 , OP2 and OPm . Where O is the origin and where P1 and P2 are two points given in Problem 12. Ans. Angle between OP1 and OP2 = 72.71. Angle between OP1 and OPm = 46.9. Angle between OP2 and OPm = 25.7. 16. For three vectors A, B and C using the diagrammatic representation of vectors show that (A + B) + C = B + (A + C) 17. Find the unit vector which is perpendicular to the plane OP1 P2 where P1 and P2 are two points given in Problem 12. Ans. (P1 P2) / |P1 P2|= (1/ 93)[8, 2, 5]. 18. Show that the work done by a force F operating over the displacement d must be F d. 19. Show that in in the case of a planet having an circular orbit around the sun, no work is done at any time. 20. Show that in in the case of a planet having an elliptical orbit around the sun, no work is done over one cycle of the orbit. 21. Show that |A B| is equal to the parallelogram whose two sides are A and B. Hint: calculate the area of a parallelogram. 22. Given A = 5ax + 3a y, B = 5ax 3a y and C = 6az . Compute A B C and A (B C). Show that A (B C) = (A C) B (A B) C using the values given. Ans. A B C = 180; A (B C) = 96az 23. Show that A B C is equal to the volume of the parallelepiped whose sides are A, B and C. 24. Why is the radian dimensionless? Hint: How do you dene radians? 25. Show that the unit of force (N) is (kg)(m)(s)2 Hint: What are the units of mass? Acceleration?

39

26. Find the units of the angular velocity, torque, energy and power in SI units. Ans. Angular velocity = rad/sec; Torque = N-m; Energy = N-m; Power = N-m/sec. 27. Derive the units of refractive index of a dielectric. Start with the denition of refractive index. Ans. Dimensionless 28. Investigate the units of heat and temperature. Ans. Heat = Energy, Temperature = dimensionless, ... K 29. What are the units of pressure in SI units? Ans. N/m2 .

1. A scalar of magnitude 5, is added to 2 g of iron. The answer written is 7. Is this correct? Justify your answer. Ans. 5+2 g is not equal to 7 since the dimensions of the two added quantities are not the same. 5 is dimensionless and 2 is in grams. 2. A scalar of magnitude 5, is multiplied by 2 g of iron. The answer written is 10 g. Is this correct? Justify your answer. Ans. 5 2 g = 10 g is correct. The interpretation is Five 2 g pieces of iron... etc. 3. A 5 g of iron, is subtracted from 2 g of iron. The written is -3 g. Is this correct? Justify your answer. Ans. What is the meaning of -3 g of iron? Physically, there is no such thing. Therefore -3 g is incorrect. 4. The dot product of a vector A of unit m with a vector B of unit s gives a result C (m-s). Is this correct? Justify your answer. Ans. No the answer is incorrect. A B is always a scalar, and C is a vector. 5. The cross product of a vector A of unit m with a vector B of unit s gives a result C (m-s). Is this correct? Justify your answer. Ans. Yes, C (m-s) may be the correct answer if C = A B.

In the following questions one or more choices may be correct. Sometimes choices may be separated with an or and somtimes with an and. 1. A unit vector A is equal to (a) A/A (b) A/A (c) A-A (d) None of these Ans. (a) 2. A boat is travelling across a N-S river, perpendicular to the bank, with a velocity 10 m/s E. The current of water is moving at 10 m/s N. If the width of the river is 100 m, where does the boat reach on the other bank? (a) Directly opposite, E (b) 50 m N of the opposite point. (c) 100 m S of the opposite point (d) 100 m N from the opposite point. Ans. 100 m N from the opposite point.

40

3. Which of the following vectors are perpendicular to each other: (a) ax , a y (b) (ax + a y ), (ax a y ) (c) (A + B), (A B). (a) a and b. (b) a and c (c) c only (d) all. Ans. (a) 4. A top is spinning at 2 rev/s, what is the velocity of a point on the axis of rotation? (a) 0 m/s (b) 1 m/s (c) 2 m/s (d) 3 m/s. Ans. a. 5. If two vectors A and B are such that A B = 0. Niether A nor B is 0. what is the angle between the vectors? (a) 0 (b) 45 (c) 90 (d) 270 Ans. c or d. 6. If |E1 | = 1 and |E2 | = 4 and E1 E2 = 6 then: (a) The angle is 0 (b) Impossible answer (c) The angle is either 45 or 45 . (d) The maximum value of E1 E2 = 3. Ans. (b). 7. If |E1 | = 1 and |E2 | = 4 and E1 E2 = 4 then: (a) The angle is 0 (b) Impossible answer (c) The angle is either 45 or 45 . (d) The maximum value of E1 E2 = 3. Ans. (a). 8. If |E1 | = 1 and |E2 | = 4 and E1 E2 = 2.828 then: (a) The angle is 0 (b) Impossible answer (c) The angle is either 45 or 45 . (d) The maximum value of E1 E2 = 3. Ans. (c). 9. If E B1 = E B2 then (a) E=0 (b) E (B1 B2 ) = 0 (c) B1 = B2 (d) All of the above Ans. (d) 10. A force F = 3ax + 4a y + 5az (N) moves a particle from P=(1,1,1) to Q=(4,6,2). What is the work done? (a) 35 J (b) 34 J (c) 33 J (d) None of the above Ans. (b) 11. If A and B are two non co-linear vectors with an angle from A to B and if C is dened by AB C= AB sin (a) C is perpendicular to A and B (b) C is a unit vector (c) A B is not a vector (d) all of the above Ans. (a) and (b). 12. A = 2ax + a y + 2az and r = xax + ya y + zaz then r A = 5 represents (a) a curve (b) a region of 3-space (c) a plane (d) a sphere Ans. (c).

1. State the law of triangle of velocities. Hint: See Sect. 1.4.1

41

2. In the case where a planet is moving around the Sun show how and why the planet speeds up and slows down. Hint: See Ex. 1.8. The force is directed toward the Sun. When F l is positive the planet gains potential energy. etc. 3. When considering the position vector of two points P1 and P2 . What part of the vector OP1 is in the direction of OP2 ? Hint: Consider the dot product and the unit vector of OP2 ! 4. Using the right hand thumb rule show why the directions of AB and B A are opposite to each other. Why is A A = 0 even though A 0. 5. Explain why if three vectors A, B and C lie on a plane then the determinant Ax Bx Cx Ay By Cy Az Bz Cz =0

42

I hear, and I forget. I see, and I remember. I do, and I understand. Chinese Proverb

In this chapter the student claries concepts related to electromagnetic elds and their description using various commonly used coordinate systems. In particular the topics covered are 1. A basic idea of scalar and vector elds. 2. Right and left handed coordinate systems. 3. A discussion of the rectangular coordinate system with emphasis on: a) The unit vectors of a rectangular coordinate system, ax , a y and az . b) Distance between two points. c) Equation of a straight line. d) Equation of a plane. 4. The cylindrical coordinate system with emphasis on: a) The unit vectors of a cylindrical coordinate system, a , a and az . b) Conversion between rectangular coordinates and cylindrical coordinates. c) Equations of lines and surfaces in cylindrical coordinates. 5. Discussion of the spherical coordinate system, with emphasis on a) The unit vectors of a spherical coordinate system, ar , a and a . b) Conversion between rectangular coordinates and spherical coordinates. c) Conversion between cylindrical coordinates and spherical coordinates.

2.2. Introduction

The need for the use of more than one coordinate system arises from the fact that electromagnetic phenomena are easier calculated or understood in a system that is appropriate to that application. Frequently, it is necessary to transform from one coordinate system to another.

43

For the denition of the rectangular coordinate system in three dimensional space, one denes three mutually perpendicular lines or axes in three dimensional space. The labelling of these axes is done in two ways: a right-handed (or left-handed) orthogonal set. Why are coordinate systems important? They are important because our vectors (electric or magnetic) are functions of coordinates and so a study in this direction is imperative. The most important coordinate systems (from the viewpoint of electromagnetic theory) are the rectangular, the cylindrical and spherical coordinate systems. Generally the coordinate system is chosen keeping in mind the problem being discussed. So for more complicated problems, more complicated coordinate systems are required (See, for example I. P. Slalskaya & Uyand (1979).)

In electromagnetic theory various vectors and scalars, the electric eld or magnetic eld or the electric potential, are dened over some region of space. In this example, the rst two are vectors, while the third is a scalar. A vector or scalar eld is a vector or a scalar which is a function of coordinates and perhaps time. To talk about a vector or scalar eld therefore one has to rst dene an origin, a coordinate system and then the function or functions which dene the vector for a vector eld, and scalar for the case of a scalar eld. As a result of this denition the electric eld and magnetic eld are vector elds and the electric potential is a scalar eld. Consider the example of atmospheric pressure. We know that 1. Pressure above the surface of the earth is a function of the distance from the surface. The pressure higher up is lower than the pressure closer to the ground. 2. Also as we move transversely at any level, the pressure does not change. To write an equation describing the pressure, we must dene a coordinate system whose origin is at some point on the ground. Obviously, the pressure must be a function of the height, h. As mentioned earlier, in (2) the pressure does not change transversely. This is shown in Figure 2.1. Therefore when we write the equation for the pressure, both (1) and (2) must be satised. Since only the height is involved, only one coordinate is needed and we can safely say that the pressure is only a function of the height, P P(h) where h is the height from the origin placed on some point on the surface of the earth (at mean sea level). The equation for the pressure is: P(h) = P0 eh where h is the height above the surface of the earth (within about a km); is some constant which has to be experimentally determined, (or theoretically calcu-

44

P(h)

Figure 2.1.: Figure showing the coordinates for the pressure scalar eld

Earth

lated)1 and P0 is the pressure on the surface of the earth. (that is, P = P0 when the height, h = 0). First let us see whether both conditions (1) and (2) given above are both satised. (1) is satised since the pressure does indeed reduce when we go up, but the exact form of the equation is to be corroborated experimentally. (2) is satised since when we move transversely, the pressure remains the same. P(h) is a scalar eld because it is a scalar which is a function of coordinates; in this case a single coordinate, h. EXAMPLE 2.1 At what height is the pressure 9/10 of the pressure at the surface of the earth? Solution Step 1. Write out the formula: P(h0 ) = P0 e1.19310

4 h 0

= 0.9P0

1 1.19278 104

45

Step 2. Calculate h0 : h0 = ln(0.9)/(1.193 104) = 883 m In another example, (see Figure 2.2) the origin of the coordinate system is located at the centre of the earth. r is the distance of the observation point from this origin, where the scalar eld is being evaluated. A little thought tells us that the coordinate system being used is essentially the spherical coordinate system. Using this coordinate system, the acceleration due to gravity g for the earth may be dened as: g(R) = GME /r2 R0 r Where g(r) is the acceleration due to gravity2 of the earth at a distant point from the earths centre, in meters/sec2 ; r is the distance of a distant point from the earths centre in meters; R0 is the radius of the earth, in meters; G is the gravitational constant, in SI units; and ME is the mass of the earth, in kilo-grams. From these two examples we can see that a scalar eld is essentially a scalar which is some physical quantity which changes from point to point in space. Or in the language of mathematics, it is a physical quantity which is a function of coordinates. In general we can characterise a scalar eld by the equation (r) (2.1)

Where is (obviously) a scalar and r is the position vector of the point where the scalar exists, and which we are interested in. Thus in our two examples, in the rst case we interested in the pressure of the atmosphere somewhere in and above the clouds; and in the second case we are interested in the acceleration due to gravity much above the surface of the earth. For rectangular coordinates a scalar eld may be written as: (x, y, z) or in the case of cylindrical and spherical coordinates we have, respectively, the two equations: (, , z) (r, , ) Where the sets (, , z) and (r, , ) are the cylindrical and spherical coordinates. The previous examples have given us a fair idea about what a scalar eld is. Similarly a vector which is a function of coordinates is a vector eld. Generally the coordinate system will be three dimensional but sometimes it could be one

2

Note that though the acceleration due to gravity is a vector, here it is modelled as a scalar

46

or two dimensional in nature. The dierence between just a vector and a vector eld is that a vector eld maybe dierentiated or integrated with respect to the coordinates, but a simple vector may not be dierentiated or integrated (with respect to the coordinates) because it is a constant, and does not change from point to point. A vector or scalar eld, however, changes from point to point and therefore calculus may be applied to these entities. Thus g at a distance of r = 7 106 meters from the centre of the earth will be given by the formula given above. However, when g is evaluated at another point, r = (7 106 + 1) meters its value will be slightly dierent from the earlier result. Let us take a specic example of a vector eld. In rectangular coordinates, for example, the vector eld E in a particular case may be specied by the equation: E= K(xax + ya y + zaz )

3/2

x 2 + y 2 + z2

Where K is a constant, and x, y and z are the coordinates of the point in question. This equation is actually three equations: Ex = Kx

3/2

(x2 + y2 + z2 ) Ky

Ey =

3/2

Ez =

3/2

Let us proceed to examine this vector eld. At the coordinate point (1,1,1). E is given by: K(xax + ya y + zaz )

3/2

E=

x 2 + y 2 + z2

(x,y,z)=(1,1,1)

K = (ax + a y + az ) 27 = 0.1925(ax + a y + az )K And at the point (2,2,2), the E eld is E = 0.04811(ax + a y + az )K So it is clear that at dierent points the eld has dierent values and as we move away from the origin, in this case, the E eld reduces drastically.

47

Generalising these results we can write the E vector eld in general terms as: E E(r) Where r is the position vector of any general point. As we did for scalars, we can write the representations of E(r) in dierent coordinate systems as: E E(x, y, z) for rectangular coordinates; E E(, , z) for cylindrical coordinates; and E E(r, , ) for spherical coordinates. Did you know? The mathematical concept of vector elds arose originally 19th century physics. They were proposed by Michael Faraday, in his conceptualisation of magnetism in terms of lines of force. Faraday also underscored the fact that the eld itself should be an object of study, due to the value of such a concept in explaining physical phenomena. Today the concept of vector elds are used in practically all branches of egineering and physics.

Coordinate systems play a central role in the application of electromagnetic theory to various engineering situations. A particular coordinate system is chosen because in that coordinate system the equations reduce to a particularly simple ones. Though the rectangular coordinate system is most often used, the cylindrical and spherical coordinate systems have proved to be useful in various applications. To start with we take a closer look at the rectangular coordinate system and then we proceed to other coordinate systems.

z

y x x z

Before we proceed we must rst decide whether we will be working with a left handed or right handed coordinate system. (See Figure 2.3.) We can write our equations in either coordinate system since both are equally valid. How do we characterise these systems? In the right-handed rectangular coordinate system

48

z

Figure 2.4.: To calculate the dierence vector between two points in rectangular coordinates

if we go from ax to a y then we get az , following the right hand thumb rule, while in the left-handed rectangular system when we go from ax to a y then we get az , but following the left hand thumb rule. In this book, however, we shall only be using right handed coordinate systems only.

We will be required to do various mathematical manipulations in the rectangular (or rectangular) system. For example if two points: (x0 , y0 , z0 ) = (1, 2, 2) and (x1 , y1 , z1 ) = (2, 3, 2) are given, then how do we nd the distance between these two? On the other hand if two lines are given how do we nd the angles between the two? To answer these questions we can use vector algebra with great advantage. Referring to Figure 2.4 with (x0 , y0 , z0 ) = (1, 2, 2) and (x1 , y1 , z1 ) = (2, 3, 2) the dierence vector is given by: R{point 0 to point1} = r{point 1} r{point 0} Or R01 = r(x1 , y1 , z1 ) r(x0, y0 , z0 ) = r1 r0 For two points (x0 , y0 , z0 ) = (1, 2, 2) and (x1 , y1 , z1 ) = (2, 3, 2) R01 = r1 r0 = (2, 3, 2) (1, 2, 2) = (1, 1, 0)

(2.2)

49

and r1 = 12 + 22 + 22 = 17

If we take the dot product of R01 with r1 R01 r1 = (1, 1, 0) (2, 3, 2) = 5 then we can nd the angle between the two vectors: Or cos = 5/ 34 = 0.8549 Or = 30.90 Similarly we can nd the angle with the other position vector. 2 17 cos = 5

If a vector V in rectangular coordinates is given and we wish to nd the angle the vector makes with the coordinate axes then we can do so in the following way. We rst nd the unit vector in the direction of V (V = V/V) and then take its dot product with the coordinate unit vectors ax , a y and az . Then the three direction cosines are these three dot products. Therefore if the unit vector V makes angles of , and with the x-, y-, and z-axes respectively then the direction cosines (denoted by , , ) of V are given by = V ax = V ay = V ax = cos = cos = cos

(2.3)

EXAMPLE 2.2 Find the direction cosines of and angles that the position vector r = (2, 3, 4) makes with the coordinate axes Solution: Step 1. The unit vector r is r= = (2ax + 3a y + 4az) 22 + 32 + 42 (2ax + 3a y + 4az)

50

z

Step 2. The direction cosines are the cosines of the angles which the unit vecor makes with the three axes. They are therefore = 0.37139; = 0.55709; and = 0.74278 Step 3. The corresponding angles are cos1 , cos1 and cos1 : 68.2; 56.2 and 42 EXERCISE 2.1 Find the direction cosines and angles of the point (-5,0,5). Ans. Direction cosines: 1/ 2, 0, 1/ 2; Angles: 3/4, /2, /4.

Very often one requires the vector equation of a straight line passing through a vector, or through a unit vector which is attached to some point in 3-space. As we know, any straight line can have only one degree of freedom.3 Generally the easiest equation of a straight line is the parametric equation. The simplest example of a parametric equation is that of one of the axes. Take the x-axis. The parametric equation of the x-axis is r = ax t where t is the parameter, and t . Examining the previous equation in some more detail we nd that r = ax for t = 1; r = 0 for t = 0; and r = ax for t = 1. This denes the x-axis. To obtain the parametric equation in a more systematic manner, we examine Figure 2.5. The gure shows that that if a vector r is given, attached to a point

3 The

number of degrees of freedom corresponds to the number of independent variables in any equation or mathematical description. For example in two dimensions the general equation of a straight line: y = mx + c, where m and c are constants has only one degree of freedom. If x is specied, y is uniquely determined; and, on the other hand if y is specied then x is uniquely determined. A region in the 2-dimensional plane has however two degrees of freedom. There, x and y can be independently specied

51

is another point which lies on the straight line shown. t is a parameter whose positive values lead to points toward the head of the vector r while negative values of t lead to points on the straight line toward the tail of the vector. The position vector of a point on the straight line is: r(x, y, z) = tR + r0 ( < t < ) (2.4)

Note that this vector equation is consists of three equations: the rst having x on the left hand side, and two more having y and z on the left hand side x y z = tRx + x0 = tR y + y0 = tRz + z0 ( < t < )

EXAMPLE 2.3 Find the parametric equation of the line which passes through (0,0,0) and (1,1,1) Solution: Step 1. In this example (x0 , y0 , z0 ) = (0, 0, 0) and (x1 , y1 , z1 ) = (1, 1, 1) Step 2. Then R = r1 r0 = (1, 1, 1) Step 3. Using the theory which has been developed above r = r0 + tR x = x0 + t(x1 x0) = t y = y0 + t(y1 y0 ) = t z = z0 + t(z1 z0 ) = t EXAMPLE 2.4 The two points (1,1,1)(cm) and (1,2,3)(cm) are connected by a straight line and extended on both sides. (a) Find a point on the straight line which is a distance of 5 cm from (1,1,1) toward (1,2,3) and also away from (1,2,3) (in the other direction.) (b) Find the coordinates of the points of intersection of this line with the x-, y- and z-planes. Solution: Part (a) Step 1. First of all we calculate R (see Figure 2.6). R is given by: R = r1 r0 = (1, 2, 3) (1, 1, 1) = (0, 1, 2) cm

Step 2. Next we write the vector equation for the straight line joining the two points:

52

Origin

Figure 2.6.: Figure used to calculate the vector equation of a straight line

(2.5)

Step 3. Notice that for = 0 we at the point (1,1,1) and for = 1 we land up at (1,2,3). To nd the distance between points this equation is not helpful. Step 4. To get the correct equation we need to calculate the unit vector R in the direction of R. The unit vector is calculated by: R = R/R = (0, 1, 2)/ 5 = (0, 1/ 5, 2/ 5) Step 5. We can now write a new equation involving the parameter t and ar : r(x, y, z) = t (0, 1/ 5, 2 5) + (1, 1, 1) ( < t < )

We can clearly see that the parameter t is the distance parameter. When t is 0, we are at (1,1,1), but when t=1 then we are at one unit vector away (along the line) from (1,1,1) or 1 cm away. Step 6. Now we can get the correct answer for this part: Put t=5 cm in the previous equation. r(xa , ya , za ) = 5 0, 1/ 5, 2 5 + (1, 1, 1) = 0, 5/ 5 , 2 5/ 5 + (1, 1, 1) = (0, 5, 2 5) + (1, 1, 1) = 1, (1 + 5), (1 + 2 5) cm

When we put t=-5 cm we go in the other direction and get the coordinates of the second point: r(xb , yb , zb ) = 5 (0, 1/ 5, 2 5) + (1, 1, 1) = 0, 5/ 5, 2 5/ 5 + (1, 1, 1) = (0, 5, 2 5) + (1, 1, 1) = 1, (1 5), (1 2 5) cm

Part (b) Step 1. In this part we have to nd where the line meets the three planes. Let us consider the x-y plane. The x-y plane is described by the equation z=0 (Note

53

that when z=0, x and y can be anything). Here we have to put z=0 on the left hand side of the equation: r(x, y, z) = t (0, 1/ 5, 2 5) + (1, 1, 1) ( < t < )

0 = t 2/ 5 + 1 t = 5/2 Step 2. Using this value of t and utilising the other two equations: xxy plane yxy plane =1 = 1/2 cm

Step 3. Let us proceed and calculate where this line meets the x-z plane described by y=0 and substituted into the second equation: 0 = t 1/ 5 + 1 t = 5 With which we calculate: xxz plane zxz plane =1 = 1 cm

Step 4. In the same manner we proceed to calculate the meeting point of the line with the third plane the y-z plane described by x=0. The appropriate equation is: 0 0

???

=1?

= t0+1

Here after careful thought we conclude that the line does not meet the y-z plane!

EXERCISE 2.2 Find the parametric equation of the line which passes through the points (-1,-1,-1) and (1,1,1). Ans. r = (1, 1, 1) + t(2, 2,2)

Very often in the application of the electromagnetic equations, the equation of a plane is desired. Again we can obtain the equation of one of the coordinate planes. Take the y-z plane. All over the y-z plane the value of x is zero. Hence the equation of the y-z plane is x = 0.

54

Referring to Figure 2.7, suppose R1 and R2 are two vectors given in 3-space, then we would like to calculate the equation of a plane which encloses both R1 and R2 and also passes through the point r0 (x0 , y0 , z0 ). Let the coordinates of any point lying on the plane be r(x, y, z). To proceed further we rst calculate two vectors: (shown in the Figure) Rp = R1 R2 and We can clearly see that rp is perpendicular to the plane, and R must lie on the plane. To make R lie on the plane, we put the restriction that it be perpendicular to Rp : R Rp =0 (r r0) Rp =0 (2.6) (r r0 ) (R1 R2 ) = 0 This equation is a single equation in three variables: x,y and z and therefore has two degrees of freedom. The equation apples particularly in two cases: (i)When two vectors R1 , R2 and r0 are given (a plane parallel to two given vectors and passing through a point) and (ii)When Rp and r0 are given (a plane passing through a point and perpendicular to a given vector.) R = r r0

EXAMPLE 2.5 Find the equation of a plane passing through (1,0,0), (0,1,0) and (0,0,1) in the rectangular coordinate system. Also nd the unit vector perpendicular to this plane. Solution: Step 1. Two vectors which lie on this plane are: R1 = (0, 1, 0) (1, 0, 0) = (1, 1, 0) and R2 = (0, 0, 1) (1, 0, 0) = (1, 0, 1)

55

Step 2. If we take the cross product R1 R2 we get a vector perpendicular to the plane: Rp = R1 R2 = (1, 1, 0) (1, 0,1) = ax (1, 0, 1) + a y (1, 0, 1) On further simplication: Rp = a y + az + ax (1, 1, 1) Step 3. By inspection the unit vector is: Rp = (1/ 3)(ax + a y + az ) Step 4. Using the equation derived in the this section: (x, y, z) (1, 0, 0) (1, 1, 1) = (x 1) + y + z = 0 Hence the desired equation is: x+ y+z= 1

EXERCISE 2.3 Find the plane passing through (0,0,0), (1,0,0) and (0,1,0) Ans. z = 0

56

Application. A practical application of the rectangular coordinate system as applied to the Earth GPSa is the UTMb locator system. With UTM , the earth is divided by parallel lines which are E-W and N-S on the Mercator projection map of the world. Similar to latitude longitude lines, these parallels create a matrix of 60 columns by 20 rows. The columns are labeled by number 1 through 60 (starting from the tip of Siberia near Alaska going east) and the rows are labeled C through X, starting from the south pole. O and I, are ommited so they are not confused with the numbers zero and one. Each UTM zone is therefore expressed by a number and letter, for example New Delhi, India would be in R43.

a Global b Universal

z a b d c e

g O f y

The cylindrical coordinate system is shown in Figure 2.9. Any point in cylindrical coordinates is described by three scalars: , , and z. From the point in question (i.e., whose coordinates are , , and z) a perpendicular is dropped onto the x-y plane. As in the rectangular case, the length of this perpendicular is the z coordinate. The foot of the perpendicular (point f) is joined to the origin (O). This length of this line is the coordinate. The angle between this line and the x-axis is the coordinate. The z coordinate is the same as in the case of the rectangular coordinate system. By observing the gure which shows the cylindrical description superimposed on the rectangular coordinate system we get the following set of equations: x= y= z= cos sin z

(2.7)

57

obtained by squaring and adding the rst two equations, while is obtained by dividing the second equation by the rst. The inverse transformation equations are: = x2 + y2 (2.8) = arctan(y/x) z= z EXAMPLE 2.6 Find the cylindrical coordinates of the point (x = 1, y = 1, z = 1) given in rectangular coordinates. Solution: Step 1. The rectangular coordinates are x = 1, y = 1 z = 1 Step 2. the coordinate is given by = Also = arctan and z=1 The coordinates in cylindrical coordinates are , , z = (1.4142, 0.78540, 1) EXAMPLE 2.7 Find the rectangular coordinates of the point (=1, =1,z =1) given in cylindrical coordinates. Solution: Step 1. The cylindrical coordinates of the point in question are , , z = (1, 1, 1) the three rectangular coordinates are x = cos = 1 0.54030 y = sin = 1 0.84147 z=1 x2 + y2 = 2 = 1.4142

y = = 0.78540 x 4

EXERCISE 2.4 Find the cylindrical coordinates of the rectangular coordinate points (x = 1, y = 1, z = 1), (-1,1,1), (1,-1,1) and (-1,-1,1) Ans. In all cases, z = 1 and = = 2, = /4. 2, = /4; = 2, = 3/4, = 2, = 5/4,

58

Table 2.1.: Cross products of the cylindrical coordinate system

a a az

a 0 az a

a az 0 a

az a a 0

An important point to remember is that the unit vectors in the cylindrical coordinate system are dependent on the position of the point in space. Examining Figure 2.9 we can see that the along straight line d e only changes. On this line the z and are constants. The unit vector a is parallel to this line. The circle a b c is a circular line on which only the coordinate changes; on this circle z and are constants, and the unit vector a is tangent to it. Similarly on the straight line f g only the z coordinate changes and the unit vector az lies parallel to it. It is important to note that the three unit vectors, a , a and az are orthogonal to each other. Referring again to Figure Figure 2.9 on page 57, we nd that a is parallel to the x-y plane and along the line joining the origin to the perpendicular from , , z to the x-y plane. It makes an angle with the x-axis and an angle (/2 ) with the y-axis.

Therefore a and a are perpendicular to az . Thus we see that the a , a , az form a right-handed, orthonormal set of vectors. a a a az az a = az = a = a

Since a is parallel to the x-y plane, it is perpendicular to az . Further, unit vector a is perpendicular to a in the increasing direction of phi and again parallel to the x-y plane. By observing the gure, we can see that a = az a .

(2.9)

The relations given above are put down in the form of a table and an aid to their memorisation is through the use of the accompanying diagram, Figure 2.10. Scrutinising the gure we can see that a a moving along the arrow gives us az and so on. (See 1.4.4) a , a , ax , and ay lie on the same plane, a plane parallel to the xy plane. The three unit vectors for the cylindrical coordinates are related to the unit vectors for rectangular coordinates. Let us start with a . (Refer to Figure 2.11)

59

a makes an angle of with ax and (/2 ) with a y , therefore a = cos ax + sin a y Also a = az a, therefore a = cos a y sin ax These relations are given below

a = cos ax + sin a y a = cos a y sin ax az = az And they may be written in the form of a matrix, [T]rc , as: a a az cos sin 0 ax sin cos 0 a y = 0 0 1 az

[T]rc

(2.13)

60

matrix T1 between rectangular-cylindrical coordinates) are rc ax cos sin 0 a a sin cos 0 a y = 0 0 1 az az

[T]1 rc

(2.14)

With the inverse related to [T]rc by [T]1 = [T]t rc rc The superscript t implies transpose Figure 2.11 shows these relations in threedimensional space. The position vector r is given by:

r = xax + ya y + zaz = cos ax + sin a y + zaz Or r = a + zaz If we take the dot product of Equation 2.10 with ax and a y we get: a ax a a y Similarly Equation 2.11 gives us: a ax a a y = sin = cos = cos = sin

(2.15)

(2.16)

(2.17)

(2.18)

These two equations are extremely useful because if some vector A is given both in rectangular as well as cylindrical coordinates then: A A a + A a + Az a z = Ax a x + A y a y + Az a z Then A = a (A a + A a + Az az ) = a (Ax ax + A y a y + Az az ) = Ax (a ax ) + A y(a a y ) + Az (a az ) = Ax cos + A y sin (2.19)

Using this technique, we can now write the general matrix equation for conversion: A A Az cos sin 0 Ax sin cos 0 A y = Az 0 0 1 61

(2.20)

Table 2.2.: Table of dot products between ax , a y , az and a , a , az

a a az And

ax cos() -sin() 0

ay sin() cos() 0

az 0 0 1

or

Ax a x a A a a y = y Az az a

ax a a y a az a

ax az a y az az az

The individual terms of the matrix have been evaluated from Equations 2.17 and 2.18 where the dot products are given. These dot products can be put down in the form of a table, Table 2.2. EXAMPLE 2.8 Compute ax , a y and az in terms of a , a and az . Solution: Step 1. ax = (1, 0, 0) in rectangular coordinates. In Equation 2.21 we insert (1, 0, 0) instead of Ax , A y , Az and get A A Az cos sin 0 1 cos sin cos 0 0 sin = = 0 0 0 0 1 ax = cos a sin a +

A A

A A Az

(2.21)

(2.22)

Hence

0 az

Az

Step 2. We notice that the unit vector ax is a function of coordinates. Thus it depends on the coordinates of the point where the unit vector is evaluated. For example, if we take the point (1, 0, z) in cylindrical coordinates then ax = a . On the other hand at the point (1, /2, z) ax = a . Step 3. In a similar manner, a y = sin a + cosa and az is the same in both coordinate systems.

62

EXAMPLE 2.9 The vector eld A(x, y, z) at the point (1,1,1) in rectangular coordinates is (1,1,1). Find the value of the vector eld at the same point in cylindrical coordinates. Solution: Step 1. The point (1,1,1) (rectangular coordinates) translates to , , z = (1.4142, 0.78540, 1) (cylindrical coordinates. See Example 2.6) Step 2. The point (1,1,1) in rectangular coordinates, using the matrix to convert to cylindrical coordinates is A A Az cos sin 0 1 sin cos 0 1 = 0 0 1 1

Step 3. Since = 0.7854r, cos(.7854) = sin(.7854) = 0.7071. So A = 1.4142, A = 0, Az = 1. EXAMPLE 2.10 The vector eld A(, , z) (cylindrical coordinates) in the region 0 1 is given by A = a Find the value of the vector eld in the same region in rectangular coordinates. Solution: Step 1. The components of the vector eld A are A = 0, A = , Az = 0. We need to nd Ax , A y , Az . So using the conversion matrix Ax cos A sin y = 0 Az sin cos 0 0 0 1 A A Az cos sin 0 0 sin sin cos 0 cos = = 0 0 0 0 1 y = x 0

Step 2. We use =

Or

x2 + y2 y x2 +y2 Ax A = Ay = x2 + y2 x 2 x2 +y Az 0 A = yax + xa y.

The simplest surface in cylindrical coordinates corresponds to = constant. A little reection will tell us that this equation is that of a cylinder. An example of such a surface is shown in Figure 2.12. The gure shows three surfaces dened by the equation = constant, z = constant and = constant. In this

63

gure z = constant planes are planes parallel to the x-y plane. When we put = 0 = constant then it describes a half-plane passing through the z-axis and at an angle of 0 .

Surface

Surface Surface

EXAMPLE 2.11 Find the parametric equation of the cylinder = 2 in rectangular coordinates. Solution: Step 1. The equation set from , and z to x, y and z are given by Equation set 2.7 :

x = cos y = sin z=z These are in fact the parametric equations of x, y and z for any value of (or in terms of) , and z. Step 2. In these equations we substitute = 2 and we get the parametric equations of the cylinder in rectangular coordinates as a function of and z as parameters. x = 2 cos y = 2 sin z =z Step 3. Since these equations have two degrees of freedom in 3-space, they constitute a surface4 . Let us now take a look at the mathematical description of lines in cylindrical coordinates. Generally the lines which can easily be described are those lines which lie on the surface of a cylinder (though, of course other curves are possible). As in the case of rectangular coordinates, lines in cylindrical coordinates

4 In 3-space

or 2-space, a curve is always by an equation with one independent variable. A surface is always described by two independent variables, and so on. These independent variables are called parameters. Thus equations in one parameter represents a line (in 2-space or 3-space); in 3-space equations in two parameters represents a surface.

64

possess only one degree of freedom. That is we equations of the type (t); (t); and z z(t) (2.23)

are curves. Here t is the parameter in question. A simple example is that of a helix (spiral, screw). A helix is a curve in space which lies on a cylinder and advances spirally [see chapter 11]Thomas & Finney (1996). A little reection tells us while the x- and y- values lie on a circle, the z-coordinate advances in such a way that the whole curve lies on a cylinder. When the curve advances in the direction by 2 then the z-coordinate advances in the z-direction by L. The parametric equation for a right-handed helix, where the distance between turns is 2 and the radius of the helix is 0.5, is given by the equations (in cylindrical coordinates): z = 0.5 =t =t (2.24)

A plot of the helix described by these parametric equations for two turns is presented in Figure 2.13

65

z q a b c p g

f s

Did you know?/Application. In the mathematical description of the movement of planets or comets about the sun, polar coordinates are used. a The orbit of a planet or comet is a conic section, which is given by = l ; z=0 1 + e cos

where e is the eccentricity and l is the semi-latus rectum (the chord through a focus parallel to the conic section directrix is called the latus rectum,). If e > 1, this equation denes a hyperbola; if e = 1, it is a parabola; and if e < 1, the curve is an ellipse. When e = 0 we obtain a circle of radius l.

a Note

The Spherical coordinate is represented by three parameters: (r, , ) which are shown in the accompanying gure. To describe any point, the r parameter is the distance from the point in question to the origin. This the length O p . Therefore: r |r| = x 2 + y 2 + z2 (2.25)

where x, y, and z are the rectangular coordinates of the point. The coordinate is the angle between the two lines: the z-axis and the line O p. (i.e., the line joining the origin to the point). (See Figure 2.14.) From observing the Figure, it is clear that z = rcos (2.26)

66

or = arccos(z/r) = arccos z x2 + y2 + z 2 (2.27)

similarly it is clear that the coordinate is angle between the x-axis and the line joining the origin to the foot of the perpendicular5 from the point p to the x-y plane . Therefore: = arccos(x/) = arccos x x2 + y 2 (2.28)

Where = x2 + y2 . These relations have give the values the r, and coordinates in terms of the of x, y and z coordinates. The inverse transformation is given by: x = r sin cos y = r sin sin z= r cos

(2.29)

The unit vectors in the spherical coordinate system are depicted in Figure 2.14. ar is outward along the line O p. On the line O p and are both constant and only r varies. On the circle a b c r and are both constants, while is variable. Tangent to this circle is the unit vector a . The third unit vector a is tangent to the circle q p s on which r and are both constants and is variable. The circle a b c is like the line of latitude on the surface of the earth, while the circle q p s is like a line of longitude. These three unit vectors are perpendicular to each other with ar a = a (2.30)

a a = ar

a ar = a

To compute the unit vectors of the spherical coordinates in terms of the unit vectors of the rectangular and cylindrical coordinate systems, we take a look at Figure 2.15. The unit vectors a , ar , a and az lie on the same plane as shown in the gure. The unit vector ar is connected to az and a by ar = az cos + a sin Since a = ax cos + a y sin

5

(2.31) (2.32)

The perpendicular is p f or g f

67

x

Figure 2.15.: Relationship between (a , az ) and (ar , a ) of the cylindrical and spherical coordinate systems respectively

a

(rectangular to spherical)

(2.33)

The important point to remember is that the spherical and cylindrical coordinates share the same , and a , coordinate and unit vector respectively. Thus observing the plane = constant the unit vectors az , a (which are perpendicular to each other) and ar , a (which are also perpendicular to each other) lie on this plane as diagrammed. Since a = a ar and using the results of the cross products of cylindrical coordinates: a = a ar = a (az cos + a sin ) = a az cos + a a sin )

a az

= (ax cos + a y sin ) cos az sin The third unit vector a is as for cylindrical coordinates a = a y cos ax sin

(2.34)

(2.35)

Therefor the unit vectors of the spherical coordinate system in terms of the rectangular coordinates are given by: ar = sin cos ax + sin sin a y + cosaz . a = cos cos ax + cos sin a y sin az

a = sin ax + cos a y

(2.36)

68

Table 2.3.: Table of dot products in Spherical Coordinates

ax ay az

a sin cos 0

These results can be put in the form of a matrix equation (with the transformation matrix Trs ) : ar sin cos sin sin cos a cos cos cos sin sin = a sin cos 0

Trs

ax a y az

(2.37)

We can also nd the inverse transformation (with T1 ) rs ax sin cos cos cos a sin sin cos sin y = cos sin az

T1 rs

sin cos 0

ar a a

(2.38)

Note that

T1 = Tt rs rs

(2.39)

The three unit vectors of the spherical coordinate system are orthonormal, and a table of their dot products with ax etc., are shown in Table 2.3 We can compute the components of a vector in either of the two coordinate systems Ar A A ar ax ar a y ar az a a a a a a = x y z a ax a a y a az sin cos sin sin = cos cos cos sin sin cos ax a a y a az a ax a a y a az a ax a a y a az a Ax A y Az cos Ax A sin y 0 Az

(2.40)

Ax A y = Az =

sin cos cos cos sin sin cos sin cos sin

sin cos 0

Ar A A

(2.41)

Ar A A

69

EXAMPLE 2.12 Compute ax , a y and az in terms of a , a and a . Solution: Step 1. ax is (1,1,10 in rectangular coordinates. Therefore in Equation 2.40 we insert (1, 0, 0) instead of Ax , A y , Az and get Ar A A ar ax a a = x a ax ar a y a a y a a y ar az a az a az 1 0 0

which translates to Ar A A

Step 2. Hence

sin cos sin sin cos = cos cos cos sin sin sin cos 0

Ar A A

Step 3. In the same manner we get a y = sin sin ar + cos sin a + cos a az = cos ar sin a The spherical coordinate system is especially suited mathematical descriptions involving spheres, cones or half planes passing through the origin. Thus r = 3 is a sphere whose radius is three; = /4 is a cone whose cone angle is /2; and = /6 is a half plane passing through the origin at an angle of 300 to x z plane. These surfaces are shown in Figure 2.16. As we have noted earlier that any equation with two degrees of freedom is a surface, we can rewrite the equation of a sphere as: r= 3 = anything = anything (2.42)

Which has only two degrees of freedom (only two variables and may be specied independently. r is xed, equal to 3). Just as in the case of the cylindrical coordinate system we can write parametric equations of a space curve which are functions of a single variable in the spherical coordinate system: r r (t) ; (t); and (t) Where t is the parameter and r (t) , (t) and (t) are functions of t. (2.43)

70

z z z

y x (a) x (b)

y x (c)

R01 , the position vector from point 0 at r0 to point 1 at r1 is given by R01 = r1 r0 The direction cosines of a vector V are given by = V ax ay =V = V ax = cos = cos = cos

The parametric equation of a straight line where R is vector parallel to the line, r0 is the position vector of a specic point on the line and r is the position vector of any point on the line is r(x, y, z) = tR + r0 where t is a parameter. The equation of a plane in rectangular coordinates is given by (r r0 ) (R1 R2 ) = 0 where r is the position vector of a point on the plane; r0 is the position vector of a specic point on the plane and R1 and R2 are two vectors lying on the plane. If , and z are the cylindrical coordinates of a point then the rectangular coordinates are given by x = cos y = sin z= z If x, y and z are the rectangular coordinates of a point then the cylindrical ( < t < )

71

coordinates are given by x2 + y2 = = arctan(y/x) z= z The cross products of the unit vectors in cylindrical coordinates are given by a a = az a az = a az a = a The position vector of a point in cylindrical coordinates is r = a + zaz When the same vector is expressed in cylindrical and rectangular coordinates then A cos sin 0 Ax A sin cos 0 A = y Az 0 0 1 Az Ax cos sin 0 A A y = sin cos 0 A 0 0 1 Az Az

and

Spherical coordinates use three coordinates: r, and . In terms of rectangular coordinates these are: x 2 + y 2 + z2 z = arccos 2 + y 2 + z2 x x = arccos x2 + y2 r= r sin cos r sin sin r cos

If the coordinates are expressed in spherical coordinates then rectangular coordinates the equations are x= y= z=

a a = ar

a ar = a

72

If a vector is given in rectangular coordinates then in spherical coordinates its components are Ar A A sin cos cos cos = sin sin sin cos cos sin sin cos 0 Ax A y Az

If a vector is given in spherical coordinates then in rectangular coordinates its components are Ax sin cos A sin sin y = Az cos cos cos sin cos sin cos sin 0 Ar A A

Chapter Summary

Since electromagnetism is expressed in terms of vectors and scalars which change from point to point in space we have to express our equations in terms of scalar and vector elds. Scalar and vector elds are scalars and vectors which are functions of coordinates and time. Scalar and vector elds may be dierentiated and integrated with respect to coordinates and time. The type of coordinate system we choose to express our equations in aids in making those equations simple in that particular coordinate system. In the following equations r = x, y, z The distance between two points with position vectors r1 and r2 is |r1 r2 |. The vector equation of a straight line in rectangular coordinates is r = tR + r0 where t is a parameter, R is a vector along the line and r0 is the position vector of a particular point on the line and r is the position vector of any point on the line. The equation of a plane in rectangular coordinates is (r r0 ) (R1 R2 ) = 0 where r is the position vector of any point on the plane; r0 is the position vector of a point on the plane and R1 , R2 are two non-colinear vectors on the plane. Cylindrical coordinates are used where the geometry is either a cylinder, = 0 , or a semi-innite half-plane, = 0 or a plane z = z0 or a combination of these three. Spherical coordinates are used where the geometry is either a sphere, r = r0 , or a cone, = 0 , or a semi-innite half-plane, = 0 or a combination of these three.

73

Review Questions

1. Dene (a) A scalar eld and (b) A vector eld. Why do we need scalar and vector elds to study electromagnetic theory? 2. Why is the pressure above the Earths surface a scalar eld? Is the dielectric constant of air above the Earths surface a scalar eld? Research this question. 3. Are scalar and vector elds dierentiable with respect to coordinates? With respect to time? 4. What is the dierence between a right-handed and left-handed coordinate system? Are equations written in the two coordinate systems exactly same? Explain your answer with an example. 5. What is the diernce between the equation of a curve and that of a surface? 6. Explain with examples where you would use Cartesian, cylindrical and spherical coordinate systems? 7. In the spherical coordinate system are the unit vectors constant vectors? Explain with an example.

Problems

1. The scalar eld V(z) = 100/z for z > 1000 km. Find V at 1000 and 1010 km. Ans. 104 and 0.9901 104 2. Find dV/dz of Prob. 1 at 1000 km. Ans. 1010 3. A vector eld is given by A = 3xyax + 3yza y + 3zxaz. Compute Ax A y Az + + x y z what is the result? Ans. 3(x + y + z); A scalar eld. 4. The vector eld A of Prob. 3. Compute ax ay az /x /y /z Ax Ay Az Ans. 3yax 3za y 3xaz; A vector eld. 5. A scalar eld is given by = 3xyz compute A= ax + ay + az x y z

74

at the point (1,1,1) Ans. A = 3(ax + a y + az ) 6. In rectangular coordinates nd the equation of a straight line which passes through the point (1,1,1) and is parallel to the x-axis. Ans. (x, y, z) = (1 + t, 1, 1). Other solutions are also possible. 7. In rectangular coordinates nd the equation of a plane which passes through the point (1,1,1) and is parallel to the xy-plane. Ans. z 1 = 0 8. Find the equation of the straight line joining the origin to (1,1,1). Ans. (x, y, z) = (1 + t, 1 + t, 1 + t). Other solutions are also possible. 9. Find the equation of a plane passing through (1,2,3) and perpedicular to the vector 2ax + 3a y + 4az . Ans. 2x + 3y + 4z 20 = 0 10. Find the equation of a line passing through the points (1,2,3) and (2,3,4). Ans. (x, y, z) = (1 + t, 2 + t, 3 + t). 11. Find the equation of a circle in the x-y plane with radius 5. Ans. (x, y, z) = (5 cos t, 5 sin t, 0) with 0 t 2. 12. What does (x, y, z) = (5 cos t, 5 sin t, 4) with 0 t 2 represent? Justify. Ans. The equation of a circle with radius 5 on the plane z = 4. 13. Find the equation of an innite cylinder of radius 2 with its axis coincident with the z-axis in cylindrical coordinates. Ans. = 2. 14. Find the equation of a circle of radius 2 with its center coincident with the origin and lying on the xy-plane, in cylindrical coordinates Ans. = 2, z = 0, = t 0 t 2. 15. Find the equation of the half-plane x = 0 y 0 in cylindrical coordinates. Ans. = /2. 16. For the rectangular coordinate system nd the point (x = 1, y = 2, z = 3) in cylindrical coordinates. Ans. = 5, = 63.4, z = 3. 17. The points (0,0,0), (1,0,0) and (0,1,0) form a triangle. Find the unit vector normal to the surface of the triangle. Ans. az . 18. Find the unit vector to the surface x2 + y2 + z2 = 25 at the point (0,0,5) Ans. az 19. Find the unit vector to the surface x2 + y2 + 2z = 5 at the point (0,0,5/2) and (1,1,3/2) Ans. az and ax + a y + az / 3 20. Consider the vector eld A = sin a + cos za + az Find the normal and tangential component at the point ( = 2, = 1r , z = 1) on the cylinder = 2. Ans. Normal component 0.8415a; Tangential component 0.5403a+ 2az .

75

1. Find the unit vector of the point (x = 0, y = 0, z = k) to the point (, , 0) in the cylindrical coordinate system. Ans. u = cos ax + sin a y kaz / 2 + k2 is the unit vector in the 2. Transform 3ax 3a y 2az into spherical coordinates at the point (x = 3, y = 1, z = 3). Ans. We observe the point (-3,1,3). At this point = 0.8117r and = 2.8198r. We now use Equation 2.40 to convert from rectangular to spherical coordinates. After conversion, the vector is 4.129ar + 1.451a + 1.897a. 3. The equation for A written in spherical coordinates is A= ejkr ar r rectangular coordinates. On further conversion u = a kaz / 2 + k2

rewrite in rectangular coordinates. Ans. r = x2 + y2 + z2 . ar = cos az + sin cos ax + sin sin a y . cos = z/r; sin = x2 + y2/r; cos = x/ x2 + y2 ; sin = y/ x2 + y2 . Using these results 2 2 2 ejk x +y +z A= 2 xax + ya y + zaz x + y 2 + z2 or ar = r/r. 4. Express A = k/r2 ar in cylindrical coordinates. Ans. r2 = 2 + z2 ; ar = a + zaz / 2 + z2 giving k a + zaz 2 + z2

3/2

A=

5. The vector eld A = ka / given in cylindrical coordinates is to be expressed in spherical coordinates. Ans. = r2 z2; the unit vector a remains the same. So ka A= r2 z2

In the following questions one or more choices may be correct. Sometimes choices may be separated with an or and somtimes with an and. 1. The expression for a y in spherical coordinates at the point P(r = 4, = 0.2, = 0.8) is (a) 0.4ar + 0.5a 0.81a (b) 0.35ar + 0.48a 0.81a (c) 0.48ar + 0.48a 0.81a (d) 0.48ar + 0.35a + 0.81a Ans. (b)

76

2. The surface = 2, = 4, = 45 , = 135 , z = 3 and z = 4 form a closed surface. The enclosing area is (a) 34.29 (b) 32.27 (c) 20.7 (d) 16.4 Ans. (b) 3. The vector B = (10/r) ar + r cos a + a in rectangular coordinates at the point (-3,4,0) is (a) ax 2a y (b) 2ax + a y (c) 1.36ax 2.72a y (d) 2.72ax + 1.36a y Ans. (b) 4. A vector eld in Cartesian form is given as D = xax + ya y nd the cylindrical form. (a) a / (b) a (c) a (d) a Ans. (b) 5. A vector eld in Cartesian form is given as D = yax xa y nd the cylindrical form. (a) a / (b) a (c) a (d) a Ans. (d) 6. A vector eld in Cartesian form is given as D = xax ya y nd the cylindrical form. (a) a / + a (b) cos(2)a sin(2)a (c) cos a + sin a (d) None of these Ans. (b) 7. A vector eld in Cartesian form is given as D = yax + xa y nd the cylindrical form. (a) cos(2)a + sin(2)a (b) cos(2)a sin(2)a (c) cos a + sin a (d) None of these Ans. (a) 8. A vector eld is given by D = k/r2 ar in spherical coordinates. Find the Dz component of the vector eld in rectangular coordinates. (a) kz/ x2 + y2 + z2 (b) kxy/ x2 + y2 + z2 (c) kz/ x2 + y2 + z2 of these Ans. (c)

3/2

(d) None

9. A vector eld is given by D = k/r2 ar in spherical coordinates. Find the Dz component of the vector eld in cylindrical coordinates. (a) kz/ 2 + z2 (b) k2 sin cos / 2 + z2 (c) kz/ 2 + z2 these Ans. (c)

3/2

(d) None of

77

10. The unit of ar in spherical coordinates is (a) m (b) no dimension (c) ft (d) None of these Ans. (b)

1. Give one example each of a (a) A scalar eld and (b) A vector eld. What is the advantage of using elds in electromagnetic theory? Hint: See Sec. 2.3. 2. When do we use a particular coordinate system? For example to describe a plane which coordinate system will we use? Hint: See Sec. 2.4. 3. If we want to nd the equation of a straight line in 3-space what are the possible ways of dening the straight line? Hint: See Sec. 2.4.3. 4. For dening a helix, why do we choose the cylindrical coordinate system? Hint: See Sec. 2.5.1. 5. Which are those surfaces which are special in terms of coordinate = constant in spherical coordinates? Hint: See Sec. 2.6. ***Chapter Complete***

78

3. Vector Calculus

Do not worry about your diculties in Mathematics. I can assure you mine are still greater. Albert Einstein

1. The student is the introduced to the concepts of line integral, surface integral and volume integral in a very basic manner and starting from the fundamentals. 2. Based on the above denition the concept of grad, curl and div are introduced. The physical interpretation of these operations is made clear. 3. Divergence theorem and Stokes theorem are also discussed. 4. At the end of the chapter Maxwells equations are introduced and the units of various electromagnetic quantities are discussed.

The analysis of vector and scalar elds is facilitated through the use of vector calculus. Scalar and vector elds are important since they are the words of the mathematical language which describes the majority of electromagnetic phenomena. Luckily, the cornerstone on which the whole of vector analysis rests are a few important but simple concepts, which once understood leads to a mastery of the subject. These shall be presented here, hopefully in a clear and concise manner.

To understand vector calculus a clear concept of dierential elements must be acquired. These elements generally occur in the various coordinate systems which we have considered in a previous chapter. Because of their importance we need to know the dierent types of dierential elements which occur in the description of EM elds and how to manipulate them. To start with we concentrate on linear elements. Figure (3.1) shows a curve in space on which two closely placed points are shown. The coordinates of the point A are (x, y, z) in rectangular coordinates. Similarly the point B which is very close to the point A has the coordinates of (x + x, y + y, z + z). The position vector r of the point A is xax + ya y + zaz . In particular, r(A) r(x, y, x) in rectangular coordinates, r(, , z) in cylindrical coordinates, and so on. Similarly the position vector

79

3. Vector Calculus

B A y

x

Figure 3.1.: Dierential element of a line

of the point B is r(B) r(x + x, y + y, z + z) in rectangular coordinates, and r( + , + , z + z) in cylindrical coordinates, etc. The point B is so close to the point A that the linear element AB can be considered to be a straight line even though both points may lie on a curve. From this it is clear that the dierential elements of the coordinates in the rectangular coordinates is the set (x, y, z) Where the x can be read as a very small increment of x.1 And the dierential element of a line (the vector AB) is the vector: l r = xax + ya y + zaz (3.1)

The equation given above says that the vector l is the sum of three vectors: a small vector x in the ax direction, y in the a y direction and z in the az direction. Where x = xax , y = ya y and z = zaz . The length of this vector is l |r| = x2 + y2 + z2 (3.2)

It is obvious that the unit vector along the line is t= l xax + ya y + zaz = l x2 + y2 + z2 (3.3)

On a little reection it should be clear that this unit vector is along the direction of the tangent to the curve on which the dierential element lies.

1 Note

80

3. Vector Calculus

z a b c

z

d e

g O f

x

Figure 3.2.: Dierential linear elements in cylindrical coordinates

In the same way the in cylindrical coordinates, shown in Figure 3.2 the dierential elements of the coordinates are: (, , z) Referring to the gure the dierential element z is the same as that in rectangular coordinates; the element is a small linear element in the direction of the unit vector a and which is a short extension of the line d-e in the gure; while is a small linear element tangent to the circle a-b-c and which is in the direction of a . The dierential element, the vector AB of Figure 3.1 on the previous page, is the vector: l r = a + a + zaz (3.4)

From the equation given above, we can see that the vector l is the sum of three small vectors: a , a and zaz . By varying the values of , and z we can reach any neighbouring point. The length of this vector is |r| l = 2 + 2 2 + z2 (3.5)

In Spherical coordinates, shown in Figure 3.3 on the following page, the dier-

81

3. Vector Calculus

z q a b c z d p g O f s y

x

Figure 3.3.: Dierential linear elements in spherical coordinates

ential coordinates are (r, , ) The spherical coordinate system is exactly the same system which is used to describe the geography of the earth. q-p-s in this gure is a line of longitude passing through the point with position vector r. On this circle r and are constant. a-b-c is a line of latitude passing through our point and on it r and are constant. The element r is a linear element which is in the direction of the unit vector ar and an extension of the r coordinate keeping and constant. r is a small element of the line of longitude q-p-s. And, nally, the dierential linear element r sin is an element of the line of latitude a-b-c. The general linear element, the vector AB of the gure on page 80, l r = rar + ra + r sin()a its length is l |r| = r2 + r2 2 + r2 sin2 ()2 (3.8) (3.7)

and the unit tangent vector to the curve at this point is t= rar + ra + r sin()a r2 + r2 2 + r2 sin2 ()2 (3.9)

82

3. Vector Calculus

z

y A

Given a scalar or vector eld, (r) or F(r) we can dene two types of line integrals:

B

(r)dl

A

and

B A

F(r) dl

(3.10)

To understand the scalar line integral let us look at Figure 3.4. In reference to this gure, we wish to compute the integral

B

(r)dl

A

over the curve A B. To do this, we divide the curve into a number of points (N + 1, to be exact) starting with r0 = (x0 , y0 , z0 ), to rN = (xN , yN , zN ). Two typical points on this curve are also shown, ri (xi , yi , zi ) and ri+1 (xi+1 , yi+1 , zi+1 ). The integral is computed according to i=N1 lim (xi , yi , zi )li N

i=0

(3.11)

where

(3.12) (3.13)

83

3. Vector Calculus

In the limit as N the sum (Equation 3.11) becomes an integral: B B i=N1 lim (xi , yi , zi )li = (x, y, z) dx2 + dy2 + dz2 (3.14) (r)dl = N

i=0 A A

B B

= = =

(3.15)

(r)dR =

A A

(3.16)

B B

(r)dl =

A A

[r(t)]

dr dt dt

(3.17)

This equation applies to any coordinate system. EXAMPLE 3.1 Find the length of the straight line from A(0, 0, 0) to B(1, 1, 1) Solution: Step 1. First we need to nd the parametric equation of the straight line. By the methods we have discussed in the last chapter, this is x=t y=t z=t As a check, we notice that at t = 0 the coordinates are (0,0,0) and at t = 1 the coordinates become (1,1,1). Step 2. The length of the line is given by A dR, where A(0, 0, 0) is the point when t = 0, and the point B(1, 1, 1) is reached when t = 1. Using Equation 3.17 we can identify by inspection that (r) = 1

B

84

3. Vector Calculus

dr = 3 dt Step 3. We substitute these values in the line integral. Therefore

1

3dt =

EXAMPLE 3.2 Find the length of one turn of the helix =1 = 2t z = 2t Solution: Step 1. Just as in the previous example, one turn implies that t goes from t = 0 to t = 1. (That is goes from = 0 to = 2). Then using Equation 3.17 we nd that (r) = 1 d dr(t) d dz = a + a + az = 0a + (1)(2)a + 2az dt dt dt dt Hence dr(t) = 0a + (1)(2)a + 2az = 42 + 4 dt Step 2. Therefore

B B

dR =

A A

dr dt = dt

t=1 t=0

42 + 4 dt = 42 + 4

We proceed to the vector variety of the line integral. By scrutinising the Figure 3.5 the vector line integral i=N1 F(xi , yi , zi ) li lim N

i=0 B F(r) dr A

where: Or

i=0 B B

(3.19)

F(r)dl =

A A

85

3. Vector Calculus

z

y A

B A B

F(r) dr =

Fx

A

dy dx dz + F y + Fz dt dt dt dt

(3.21)

B A B

F(r) dl =

dr dt dt

(3.22)

An example of the line integral is where it is used to nd the work done by a particle in a eld. Thus for instance

B

W=

A

F(r) dl

(3.23)

is the work done by a variable force F exerted on a particle. Note that the expression is the integral generalisation of the formula W = Force distance . EXAMPLE 3.3 Find the work done to move a particle in the force eld F(r) = 103 ar nt r2

along the straight line from r = (10, 0, 0) m in spherical coordinates, to r = (100, 0, 0) m Solution:

86

3. Vector Calculus

Step 1. The straight line can be characterised by the parametric set of equations r = 10t =0 =0 where t goes from t = 1 to t = 10. Step 2. By examining Equation 3.22 we can identify (in the spherical coordinate system) that 103 103 ar for 1 t 10 (3.25) F(r) = 2 ar = r 100t2 d d dr dr = ar + r a + r sin a = 10ar dt dt dt dt Step 3. Hence

B A B

(3.24)

F(r) dr =

F(r)

dr dt = dt

t=10 t=1

1030 dt 100t2

t=10

t=1

t=10

=

t=1

EXAMPLE 3.4 Given the vector eld G = yax 2.5xa y + 3zaz nd the line integral of G along the straight line from (1,1,1) to (0,0,0) Solution: Step 1. The parametric equations of the straight line are x = 1 t; y = 1 t; z = 1 t Step 2. The line integral is L= =

(1,1,1) 1

0t1

G dl

(0,0,0) (0,0,0)

(1,1,1)

=

t=0 1

1 t=0

87

3. Vector Calculus

z

d b

C c y

B x

Did you know?/Application. Integration could have possibly be used to calculate the volume of frustum of square pyramid as exhibited in the Moscow Mathematical Papyrus of ancient Egypt perhaps written around 1850 BC. Archimedes used a method similar to integration to nd the area of a circle by inscibing the circle with a polygon of a greater and greater number of sides. Eudoxus (about 370 BC), systemetised such a method applying it to volumes. Similar methods were developed in ancient times in China and India. The Indian mathematician Aryabhata used a similar method to nd the volume of a cube.

Before we take a look at surface integrals we must understand the concept of a small element of area. Looking at the accompanying gure (Figure 3.6) ABCD is some curved surface on which lies a small element of area abcd. The element of area, of magnitude S, is so small that though it lies on the curved surface, it can be considered locally plane. This is like the case of a plane on the surface of the earth: though the earth is spherical the plane appears at. Since the surface is plane-like it has a unit vector associated with it: S which is perpendicular to the elemental surface area . The surface element thus has vectorial properties. The magnitude of S is |S| = S (3.26)

while the unit vector associated with it is perpendicular to the surface at the point r. S = S(r) (3.27) Examples of such surfaces are given below in Figures 3.7, and 3.8. In rectangular coordinates one such basic surface element would be S = xy and the unit vector associated with this element would be S = S = az . Such an element would lie on a plane parallel to the xy plane. Another example of a surface element lying on a plane parallel to the yz plane would be S = yz and and its direction would be S = ax . These elements are shown in Figure 3.7.

88

3. Vector Calculus

z

z

Similarly, in cylindrical coordinates the basic surface elements (S) are multiplications of z, and . Figure 3.8 shows these surface elements. Finally, in spherical coordinates the elemental lengths are r, r and r sin . S consists of multiplication of these elements. For example in Figure 3.9 a surface element has been shown lying on the surface of a sphere. The linear elements are r sin and r which are perpendicular to each other, and hence S = (r0 sin )(r0 ) and S = ar . In the other surface element, diagrammed in Figure 3.10 the two linear elements, perpendicular to each other, are r sin 0 , and r, S = (r sin 0 )(r) and S = a .

89

3. Vector Calculus

z

Surface of a sphere

Figure 3.9.: Dierential surface element in spherical coordinates of the surface of a sphere

Surface of a cone

90

3. Vector Calculus

C A D

z axis y axis B

x axis

Figure 3.11.: Figure showing the calculation of the scalar surface integral

The concept of a surface integral may now be introduced. Referring to Figure 3.11 we can calculate the integral (r)dS

ABCD

where ABCD is the surface S over which the integral is evaluated and (r) is a scalar eld. The surface is rst divided into a number of surface elements (N + 1) starting with S0 , to SN . A typical surface element Si is also shown. The integral is computed according to the sum

N

lim

(3.28)

where (ri ) is the scalar eld evaluated at the centre of the surface element, Si . (Si is a typical surface element, as discussed in the previous section.) In the limit as N the sum becomes an integral: (r)dS

ABCD

(3.29)

This expression applies to any coordinate system. EXAMPLE 3.5 To begin with let us take a simple example. For (r) = 1/r (spherical coordinates) nd the surface integral of over the surface of a sphere with radius equal to 2. Solution: Step 1. The surface integral is I= where = 1/r.

2

dS

91

3. Vector Calculus

C A D

z axis y axis B

x axis

Figure 3.12.: Figure showing the calculation of the vector surface integral

Step 2. An element of surface on a sphere is dS = r2 sin dd. So Step 3. Since the surface has r = 2 I = 0.5 sin dd

= (2) (1 1) = 2

= ( =0 ) ( cos |= ) =0

=2

The vector surface integral may be calculated using the same principles. F dS (3.30)

ABCD

F is the vector eld, and its dot product is taken with the dierential surface vector. This expression may be approximated by (see Figure 3.12): i=N1 lim F(ri ) Si N

i=0

(3.31)

where

Si = Si Si

(3.32)

92

3. Vector Calculus

The physical meaning of the vectorial surface integral can be understood from observation of Figure 3.13 showing the dierential ux3 . The gure shows that the eld F is at an angle to the normal S. The eld can therefore be split into two parts: one along S (perpendicular to the surface) which is F F cos and the other perpendicular to S which is equal to F F sin (parallel to the surface). The parallel part does not pass through the surface at all, while perpendicular part passes completely through the surface. Therefore the dierential ux, is = F cos()S = F S (3.33)

And the total ux through some surface S is given by an integration of the above dierential =

S

F dS

(3.34)

EXAMPLE 3.6 Find the ux, , through the surface of a sphere with unit radius for (i)F = (1/r) a (ii) F = (1/r) a and (iii) F = (1/r) ar in spherical coordinates. Solution: Step 1. The ux is given by =

S

F dS

dS in all the three cases is dS = ar r2 sin dd r=1 = ar sin dd Step 2. in the case of (i) F dS = (1/r)a (ar r2 sin dd) = 0 So the ux is zero. In the case of (ii) also F dS = [(1/r)a] (ar r2 sin dd) = 0 but in the third case F dS = [(1/r)ar] (ar r2 sin dd) = r sin dd r=1 = sin dd The ux is therefore =

,

3

sin dd = 4

93

3. Vector Calculus

z

Before we discuss the volume integral we must understand the meaning of an element of volume. In rectangular coordinates a volume element is V = xyz (3.35)

and consists of the three linear elements multiplied together. Using the same concept the element of volume in the cylindrical coordinate system is V = ()()(z) (3.36)

Such a volume element is shown in Figure 3.14. Similarly, in the spherical coordinate system the volume element is V = (r)(r)(r sin ) This is shown in Figure 3.15 The volume integral usually integrates a scalar eld (r). Using the earlier method of converting a volume, V, into smaller regions (N + 1 regions of elemental volume Vi , i = 0, . . .N ) and then summing over i, the volume integral is: i=N1 (r)dV = lim (ri )Vi (3.38) N

V i=0

(3.37)

This is shown in Figure 3.16. Though the elemental volumes are shown to be small cubes, they can be of arbitrary shapes, as long as they t into one another and span the whole volume.

EXAMPLE 3.7 Find the volume integral, I, of a scalar eld V = 1/r over a sphere with unit radius in spherical coordinates.

94

3. Vector Calculus

........

95

3. Vector Calculus

Solution: Step 1. The volume integral I which is of interest is given by I=

V

(1/r)dV

Where V is the region occupied by a sphere of unit radius. Step 2. A volume element in spherical coordinates is dV = r2 sin dddr. So I=

V

r=1 =2

=

V

= 4

Once the notion of a eld has been understood, we can proceed to apply calculus concepts to elds.4 First let us approach a problem in one-dimension in a very simple way. Let f (x) be a function, which we evaluate at two points f (x0 ) and f (x1 ) with x1 > x0 but x1 almost equal to x0 . Because x1 x0 therefore f (x0 ) f (x1 ). For example let the function f (x) be x3 with x0 = 3 and x1 = 3.01 then f (x0 ) = f (3) = 33 = 27 and f (x1 ) = f (3.01) = 3.013 = 27.271. If we set x = x1 x0 , then x = x1 x0 = 3.01 3 = 0.01 Then by the same token we can call f (x) = f (x1 ) f (x0). So: f (x) = f (x1 ) f (x0) = 27.271 27.0 = 0.271 (3.39)

By convention, f (x) /x is an approximation to the derivative of f (x) at x = 3. Here by numerical computation f (x) /x = 0.271/0.01 = 27.1

4 Here

96

3. Vector Calculus

z

The exact answer, as we know is 27. We can do the same computation making x1 to be still closer to x0 and we will get a result which is still closer to the exact result. For instance x0 = 3; x1 = 3.0001; x = 0.0001 f (x0 ) = 27; f (x1 ) = 27.0027001; f (x) = 0.0027001 f (x) /x = 0.0027001/0.0001 = 27.001 27.0 Let us proceed to understand the meaning of the partial derivative. Look at the scalar eld f (x, y, z) = xey sin(z) which is a function of three variables: x, y and z. We can evaluate the function f (x, y, z) at (x0 , y0 , z0 ) and (x1 , y0 , z0 ). The two points are assumed to be very close 5 to each other. If we call x f (x, y, z) = f (x1 , y0 , z0 ) f (x0, y0 , z0 ) x = x1 x0 f (x1 , y0 , z0 ) f (x0, y0 , z0 ) x f (x, y, z) = x x1 x0 In the limit as x1 x0 ,6 we can see that the denition leads to the partial derivative: f (x, y, z x

6 5 That

=

x0 ,y0 ,z0

f (x1 , y0 , z0 ) f (x0, y0 , z0 ) x1 x0

x1 x0

is x1 > x0 but x1 x0 In the limit as x1 x0 means that x1 becomes closer and closer to x0 but x1 is never equal to x0 .

97

3. Vector Calculus

In the same manner we can dene the other partial derivatives: f (x, y, z y and f (x, y, z z =

x0 ,y0 ,z0

=

x0 ,y0 ,z0

f (x0 , y1 , z0 ) f (x0, y0 , z0 ) y1 y0

y1 y0

f (x0 , y0 , z1 ) f (x0, y0 , z0 ) z1 z0

z1 z0

Where are the partial derivatives useful? Let us take some examples. The rst use that we can put partial derivatives to use is in the denition of nabla: . Nabla is a vector operator whose denition is: For the function f (x, y, z)

f f f x + y y + z z x x0 ,y0 ,z0

ax + a y + az x y z

(3.40)

f (x, y, z) x

0 ,y0 ,z0

x0 ,y0 ,z0

(3.41) The vector r of course is purely arbitrary. For example one particular value of r may be: r = 0.005ax + 0.00007a y + 0az Or another value may be: r = 0.0005ax + 0a y + 0.00009az and so on. Thus for dierent values of r we get dierent values of f . Let us take a particularly simple (but very interesting) case. Let us look at the surface: f (x, y, z) = f (x0 , y0 , z0 ) (constant) where x0 , y0 , z0 is a point on the surface. 7 Suppose that r is allowed to take values so that it lies only on this surface. Then f , must be zero, since f is equal to a constant on the surface. More clearly: f (x, y, z) x

7 This 0 ,y0 ,z0

= f (x, y, z) r x

0 ,y0 ,z0

= f (x, y, z) r x

0 ,y0 ,z0

=0

denes a surface because only two of the three variables, x, y and z may be independently specied.

98

3. Vector Calculus

when r lies on the surface f (x, y, z) = f (x0 , y0 , z0 ) After a little reection it is also clear that r lying on f (x, y, z) = f (x0 , y0 , z0 ) really means that it lies on the tangent plane to the given surface, at (x0 , y0 , z0 ). And since f (x, y, z) r x ,y ,z = 0

0 0 0

therefore these two vectors are perpendicular to each other. Hence f (x, y, z) is perpendicular to the tangent plane to the surface at r0 = (x0 , y0 , z0 ). The mathematical entity f is so important, that it is called the gradient of the scalar eld f (x, y, z)8

What is an operator? An operator is a mathematical entity which operates on a function or another entity. A very simple example of an operator is the dierentiation operator d/dt (3.42) Notice that d/dt in itself has no meaning. When it attacks or operates on a function then only meaning develops. Thus when it operates on x(t) then the operation has meaning d [x(t)] (3.43) dt which obviously is the derivative of x(t). Another important point to be remembered is that an operator may operate on another operator to give us a completely new operator. For example d2 d d = 2 dt dt dt (3.44)

Which may now operate on a function to give us a double derivative. Another example is that of a matrix 1 2 1 1 (3.45)

Or it may operate on another matrix (an operator) to give us a new matrix (operator) 1 1 3 2 5 6 = (3.47) 2 1 2 4 8 8 Let us go on to the case of the Del or nabla operator which plays an extremely fundamental role in electromagnetic theory. The Del operator in rectangular

8

99

3. Vector Calculus

coordinates takes the form = ax + a y + az x y z (3.48)

Notice that the vector notation may be omitted without loss of understanding. may operate on mathematical entities in various ways: 1. (a vector) may operate on the scalar eld to give a vector eld [] (3.49)

which is called the gradient (or grad) of . In rectangular coordinates the gradient takes the form = ax + a y + az (x, y, z) = ax + ay + az x y z x y z (3.50)

2. The Del operator may operate on a vector eld through the dot () product. In this case the result would be a scalar eld. In rectangular coordinates the product becomes F = = ax + a y + az Fx ax + F y a y + Fz az x y z Fx F y Fz + + x y z

(3.51)

and is called the divergence (or div) of the vector F. 3. may operate on a vector through the vector or cross () product to give us another vector eld. The product in rectangular coordinates is: ax F =

x

ay

y

az

z

Fx

Fy

Fz (3.52)

= ax

F y Fx Fx Fz Fz F y + ay + az y z z x x y

and is called the circulation (or curl) of the vector eld F. 4. The operator may operate on itself also in two ways, the rst of which is = 2 = = ax + a y + az ax + a y + az x y z x y z 2 2 2 + 2+ 2 2 x y z (3.53)

100

3. Vector Calculus

may operate on a scalar eld to give us another scalar eld 2 = 2 2 2 + + x2 y2 z2 (3.54)

The Laplacian may also operate on a vector 2 A = 2 Ax ax + 2 A y a y + 2 Az az 5. The operator may operate on itself through the cross product = 0 is identically zero! The student may verify this for himself. (3.56) (3.55)

3.3.2. Gradient

The gradient (or grad) is dened by the operation of the Del operator on a scalar eld (r) (3.57) The gradient in the rectangular coordinate system when applied to a scalar eld which is a function of rectangular coordinates is = ax + ay + az x y z (3.58)

EXAMPLE 3.8 Find the gradient of the scalar eld = ze cos . Solution: Step 1. Obviously the scalar eld is given in cylindrical coordinates. Therefore using the formula for the gradient in cylindrical coordinates, = therefore z = ze cos a e sin a + e cos az 1 a + a + az z

101

3. Vector Calculus

z

Note that grad() has been introduced before in Section 3.3 on page 96 where it was shown that the gradient of a scalar eld at a particular point (x0 , y0 , z0 ) is perpendicular to the surface (x, y, z) = (x0 , y0 , z0 ) ( constant). This result holds good for all coordinate systems. The second result which is very important is to do with the line integral of the gradient. Let = E (3.61) be a vector eld which we want to integrate between two points A (xi , yi , zi ) and B (x f , y f , z f ) over any curve L. It does not matter what the shape of the curve is. Then the line integral

B A L B

[ dl] = =

A L B A L B

=

A L

= (x f , y f , z f ) (xi, yi , zi )

(3.62)

In other words, the line integral of the gradient of a scalar is independent of the path

102

3. Vector Calculus

of integration9 . This is the most important result concerning the gradient.

B A L

The other results all follow from this result! 1. From Equation 3.61 we can say that if any eld is the gradient of a scalar then its line integral from A (xi , yi , zi ) to B (x f , y f , z f ) is independent of the path of integration. i.e. if

B

E( ) then

A L

(3.63)

2. When the gradient of a scalar is integrated over a closed curve, i.e., when A (xi , yi , zi ) is equal to B (x f , y f , z f ) (which means that (xi , yi , zi ) = (x f , y f , z f )) [ dl] = 0 (3.64)

any C

This is true because from Equation 3.62 (x f , y f , z f ) = (xi , yi , zi ) and hence the result follows. 3. We can proceed to infer another important result viz., that if the closed line integral of a vector eld E is always identically zero then the vector is the gradient of some scalar eld E dl = 0 then E = (3.65)

EXAMPLE 3.9 For scalar eld = xyz investigate the line integral of along the straight line joining the points (0,0,0) and (1,1,1). Solution: Step 1. Using the formula for the gradient in rectangular coordinates, = ax + ay + az x y z

= yzax + xza y + xyaz Step 2. Since the line integral is to be evaluated on the straight line joining (0,0,0) and (1,1,1) I=

L

9 Important

dl =

result

103

3. Vector Calculus

Step 3. To perform this line integration, we need to use the parametric equations for the straight line x(t) = t; y(t) = t; z(t) = t giving dx = dt; dy = dt; dz = dt or I=

L

= = =

L t=1

t=0 31 t 0=

= (1, 1, 1) (0, 0, 0) Other results involving the gradient are (where 1 and 2 are scalar elds and a and b are constants.) a1 + b2 = a1 + b2 1 2 = 2 1 + 1 2 1 /2 =

n

2 1 1 2 2

2 (n1)

= n

(3.69) (3.70)

= 0

The curl of a vector A ( A) plays a very important role in electromagnetic theory. In rectangular coordinates the curl of A is given by ax ay az A = /x /y /z = ax y Az z A y +a y (z Ax xAz )+az x A y y Ax Ax Ay Az (3.71)

104

3. Vector Calculus

(a)

(b)

Figure 3.19.: Properties of the curl. (a) When the surface enclosed is at. (b) When the surface enclosed is bulging

where x /x etc. In other coordinate systems the curl is given below. (Please refer to the Reference Material on page 1 for all important formulae) A Az A 1 Az A 1 A A = a + a + y az z z

(3.72)

(3.73) for spherical coordinates. An important relation for the curl is (the relation is given without proof, and the reader can see for example Spiegel (1974) for more detail). A dl =

L

a + 1 (rA ) Ar a r r

( A) dS

(3.74)

where (see Figure 3.19(a)) L is a closed curve over which the line integral is calculated, and S is the surface enclosed by the line. The result which we have given is due to Stokes, and therefore is called Stokes theorem. The line integral in Equation 3.74 is evaluated in the counter-clockwise sense, while the vector associated with the surface is given by the right hand thumb rule10 . The ngers curl in the direction of the line integral, while the vector associated with the enclosed surface is in the direction of the thumb. Referring to the gure we can apply the above equation. In the (a) part of the gure. The gure shows a closed curve L enclosing a at surface S. The unit vector associated with the surface is given by the right hand thumb rule. Then A dl =

L

10

( A) dS

See the Figure 1.12 on page 25 which shows how the hand is held.

105

3. Vector Calculus

We may also apply the above equation without hesitation to the case where the surface concerned is bulging rather than at as is the case of Figure 3.19 (b). If A is the gradient of a scalar eld , i.e., A = then

L

A dl = 0

(3.75)

The proof of this is seen through Stokes theorem: consider a surface S bounded by a curve L as shown in Figure 3.19. On this surface A dS = = =0

L L

A dl dl

=0

Or The inverse is also true. For every case if in some region R, A dl = 0 then A is the gradient of a scalar eld

L

A = = 0

(3.77)

(3.78)

in R. Also if

A = 0

everywhere, then A is the gradient of a scalar eld. A = EXAMPLE 3.10 Verify in spherical coordinates that for any scalar eld (r, , ), = 0. Solution: Step 1. Using the formula for in spherical coordinates = 1 1 ar + a + a r r r sin

106

3. Vector Calculus

Step 2. A is

Step 3. substituting

1 1 Ar rA A 1 sin A A = ar + r sin r sin r Ar = in the previous equation sin 1 1 r sin 1 r = r sin Step 4. or simplifying, 1 1 1 = r sin r r which is zero. 1 1 ; A = ; A = r r r sin

1 (rA ) Ar a a + r r (3.79)

r 1 1 r ar + r sin 1 1 r r r + r r

1 r sin

r a a

1 1 1 2 ar + sin r sin r r 1 + r r

r

EXAMPLE 3.11 Apply Stokes theorem to the vector eld A = x2 yzax + y2 zxa y + z2 xyaz to the surface bounded by the four points (0,0,1), (1,0,1), (1,1,1) and (0,1,1) and verify the theorem. Solution: Step 1. The formula for the curl in rectangular coordinates is A = where A y Ax Az A y Ax Az ax + ay + az y z z x x y Ax = x2 yz; A y = y2 zx; Az = z2 xy

Step 2. Substituting these into the previous formula: A = (x2 yz) (z2 xy) (y2 zx) (x2 yz) (z2 xy) (y2 zx) ax + ay + az y z z x x y

107

3. Vector Calculus

which gives A = z2 x y 2 x a x + x 2 y z2 y a y + y 2 z x 2 z a z Step 3. Stokes theorem says A dS = A dl

If we observe the region of integration we nd that it is a plane area on the x-y plane, and when traversed in the counterclockwise sense, an element of are is given by: dS = dxdyaz Step 4. so A dS = y2 z x2z dxdy but the plane of integration is the z = 1 plane, so: A dS|z=1 = y2 z x2z hence

x=1,y=1 z=1

dxdy = y2 x2 dxdy

A dS = =

x=1 x=0

dy

=

y=0

(1,0,1) (1,1,1) (0,1,1) (0,1,1)

I=

L

A dl =

(0,0,1)

Ax dx +

(1,0,1)

A y dy +

(1,1,1)

Ax dx +

(0,1,1)

A y dy

which becomes:

(1,0,1)

I=

(0,0,1)

x2 y y=0 dx +

(1,1,1) (1,0,1)

y2 x x=1 dy +

(0,1,1) (1,1,1)

x2 y y=1 dx +

(0,1,1) (0,1,1)

y2 x x=0 dy

108

3. Vector Calculus

Figure 3.20.: Properties of the curl. When the surface enclosed is closed

which simplies to

y=1

I=

y=0

y2 xdy + x /3

3 0 1

x=0 x=1

x2 dx

= y /3 =0

1 + 0

We go another route and apply Stokess theorem to the case when the line integral is vanishingly small (see Figure 3.20), and the surface is bulging as in the gure. Then obviously, A dl = 0

L0

(3.80)

This is true for any A since the length of the line L is zero. However by Stokess theorem A dl =

L0

S f or L0

( A) dS =

( A) dS = 0

(3.81)

But this holds for any A, therefore the surface integral of the curl of A over any closed surface is identically zero! ( A) dS = 0 (3.82)

This result will be discussed in the next section. EXAMPLE 3.12 For the vector eld A = r2 sin ar + r2 sin cos a , nd the surface integral of A over a sphere of radius r0 .

109

3. Vector Calculus

Solution: The curl of A in the spherical coordinates is 1 1 Ar rA A 1 sin A A = ar + r sin r sin r + which becomes in our case: E = A = (r2 sin cos ) 1 1 (r2 sin ) 1 ar + a r sin r sin 3 (r2 sin ) 1 r sin cos + a r r Er = r sin Taking A dS = =

=0,=0

1 (rA ) Ar a r r

(3.83)

Since we will be integrating over a sphere, we need to consider only the ar component. Which simplies to

r sin r2 sin dd

Sphere =,=2

r3 sin sin dd 0

=0

3.3.4. Divergence

The divergence of vector eld, A, = A, in rectangular coordinates is given by Ax A y Az A = + + (3.84) x y z In other coordinate systems the divergence of A is A = A = 1 A 1 A Az + + z (3.85)

(3.86)

110

3. Vector Calculus

and is well known as the divergence theorem. Here V is any volume and S is the surface enclosing the volume. The theorem is illustrated in Figure 3.21. Referring to Equation 3.87 and the gure, the integration here is over any volume V enclosed by the closed surface S. The unit vector S is the outward normal to the closed surface.

Applying this theorem to the case when the vector eld is the curl of some vector eld we nd ( A)dV = ( A) dS = 0

(3.88)

The rst equality is true due to the divergence theorem (Equation 3.87) and the second equality is true due to Equation 3.82 on page 109. Since this equation is valid over any volume, this must imply that. ( A) = 0 Conversely if the divergence of a vector eld is always zero, A = 0 then A must be the curl of another vector A = B since then ( B) = 0 There is an important note to be added here, (without proof) that a vector eld is fully specied only when both its divergence as well as its curl are specied. (3.91) (3.90) (3.89)

111

3. Vector Calculus

EXAMPLE 3.13 In cylindrical coordinates show that A = 0. Solution: Step 1. In cylindrical coordinates A Az A 1 Az A 1 A A = a + a + az z z B = on comparing equations A Az A 1 A 1 Az A ; B = ; Bz = B = z z 1

1 Az

1 B Bz 1 B + + z

A z

B =

Az

1 =

A z

1 1 z z + + z z z z

1 2 3 1 2 3

2 A

1 A

1 + + z 2 1 A 2 A A 2 A 1 2 A 1 z + + + z z z z

cancels cancels

A z

Az

2 A

2 A

1 A

2 A

All the terms cancel (1-1, 2-2 and 3-3 all cancel) and so we have proved our result.

Maxwells Equations, which are the fundamental equations on which the whole of electromagnetic theory is built, in general apply to four types of vector elds (r is the position vector of any point in space, and t is time): 1. The Electric eld E(r, t) 2. The Magnetic eld H(r, t) 3. The Electric ux density D(r, t) 4. And, the Magnetic ux density B(r, t)

112

3. Vector Calculus

Why the rst two are called elds and the other two are called ux densities, will be explained later in the book. In the study of electrostatics only the elds 1 and 3 will be considered (E(r, t) and D(r, t)), and that too when these elds are not functions of time: E(r) and D(r). Similarly, in magneto-statics the elds 2 and 4 will be considered (H(r, t) and B(r, t)), when these elds are time-independent functions, H(r) and B(r). The general Maxwells equations, which are both functions of position as well as time are given below, for future reference: D = v E = B/t B = 0 and H = D/t + J (3.95) (3.92) (3.93) (3.94)

These equations are in SI units. Here v is the volume charge density in C/m3 and J is the current density in A/m2 . In addition there are the two equations: D = E (3.96)

Which is a relation between D and E in a dielectric. Here is the permittivity of the dielectric medium, and its unit is F/m. Another equation is applicable to magnetic materials is B = H (3.97) Where is the permeability of the magnetic material in H/m. The last equation (and which is very important) is the continuity equation: J = v /t (3.98)

Keeping the above discussion in mind, let us take a look at the units of the various terms of Maxwells equations. The rst of Maxwells equations is: D = Dx D y Dz + + = v x y z (3.99)

The D eld is called the Electric ux density. v is the volume charge density and its unit is C/m3 . Hence we can say that D has the units of C/m2 . Next, let us look at the equation: H = D/t + J (3.100)

11

This section may be a little advanced. The student may read it as and when required

113

3. Vector Calculus

Table 3.1.: Table showing the units of the various electromagnetic quantities

Symbol D E H B v J

Name Electric Flux Density Electric Field Magnetic Field Magnetic Flux Density Volume Charge Density Current Density Permittivity Permeability

Unit C/m2 = (F/m)(V/m) V/m A/m (Vs)/m2 = (H/m)(A/m) = tesla C/m3 A/m2 F/m H/m

The unit of D/t is Cm2 s1 = (Cs1 )m2 = A/m2 . Hence the unit of J is A/m2 and H is A/m. Let us focus next on the relation between D and E: D = E (3.101)

Where the unit of is F/m. Hence the unit of E is (C/m2 )/(F/m) = (C/F)m1 = V/m, since C/F is volt. Hence the unit of the E eld is V/m. Re-looking at D, it has the alternate units of (F/m)(V/m). Finally examination of the equation: E = B/t (3.102)

straightaway gives us the unit of B: (V s)/m2 = (H/m)(A/m), where we have converted V to H by using V = Ldi/dt. The unit of B is also called Tesla (T).12 The results of this discussion can be put down in the form of a table. (Table 3.1)

The dierential element of a line in rectangular coordinates is l r = xax + ya y + zaz The dierential element of a line in cylindrical coordinates is l r = a + a + zaz The dierential element of a line in spherical coordinates is l r = rar + ra + r sin()a The line integral of a scalar eld is

B B

(r)dl =

A A

[r(t)]

dr dt dt

12 1

Tesla= 1 Weber/m2

114

3. Vector Calculus

The line integral of a vector eld is

B A B

F(r) dl =

dr dt dt

S

F dS

The gradient of a scalar eld (r) at an arbitrary point r0 =

ax + a y + az x y z

(x0 , y0 , z0 ) is perpendicular to (r) = (r0 ). The line integral of the gradient of a scalar eld ,

B A L

[ dl]

When the gradient is integrated along any closed curve C, gives the

from point A to B is independent of the path of integration and will give the same result when integrated along any line L.

any C

The rst result the surface integral of the curl of a vector connected

L

( A) dS

where the line is a closed curve L which encloses the surface S. This is Stokess Theorem.

The surface integral of the curl of any vector eld over a closed surface = 0. Here is any scalar eld.

S is always zero.

S

( A) dS = 0

115

3. Vector Calculus

The divergence A

eld A is always equal to the surface integral of that eld over the enclosing surface S

V

( A) dV =

A dS

This is the divergence theorem. The divergence of the curl of a vector is always zero. ( A) = 0 To specify a vector eld completely through equations, both its divergence and curl must be specied.

Chapter Summary

Table 3.2.: Summary of Properties of Grad, Div and Curl

Operator

Integral Property

Dierential Property

Gradient

b dl = La L

dl = 0

(a) (b)

= 0

Curl

A dl = S ( A) dS S ( A) dS = 0

V AdV

( A) = 0

Divergence

S A dS =

The gradient of a scalar eld (r) at an arbitrary point r0 = (x0 , y0 , z0 ) is perpendicular to (r) = (r0 ). The line integral of the gradient of a scalar eld ,

B [ dl], from point A L

When the gradient is integrated along any closed curve C, gives the same [ dl] = 0. result always, which is zero.

any C

A to B is independent of the path of integration and will give the same result when integrated along any line L.

116

3. Vector Calculus

The rst result the surface integral of the curl of a vector connected to its line integral A dl =

L S

( A) dS

where the line is a closed curve L which encloses the surface S. This is Stokess Theorem. = 0. Here is any scalar eld. ( A) dS = 0. The surface integral of the curl of any vector eld over a closed surface S is always zero.

S

Review Questions

1. Dene the linear elements in the rectangular, cylindrical and spherical coordinate systems. 2. Dene the following in rectangular coordinates: a) The nabla operator b) Gradient of a scalar eld c) The curl of a vector eld d) The divergence of a vector eld 3. Give the physical signicance of the gradient (of a scalar eld), curl and divergence (of a vector eld). 4. In rectangular coordinates, state the divergence and Stokes theorems. 5. What is the connection between the ux of a vector eld and divergence?

Problems

1. Find the normal vector to the surfaces r = r0 , = 0 and = 0 where r0 , 0 and 0 are constants. Ans. ar , a and a . 2. The vector eld A = yax xa y + zaz is given in rectangular coordinates. Find the surface where |A| is a constant. Ans. x2 + y2 + z2 = a constant. 3. In a particular application, the electric eld in the spherical coordinate system is given by ar E=K 2 r where K is a constant. Find E for this eld. Ans. Everywhere zero except at r = 0.

117

3. Vector Calculus

4. In a particular application, the electric eld in the spherical coordinate system is given by ar E=K 2 r where K is a constant. Find E for this eld. Ans. Everywhere zero except at r = 0. 5. A scalar eld V in the spherical coordinate system is given by V=K 1 r

where K is a constant. Find V/x, V/y and V/z. Ans. Kx/r3 ; Ky/r3 ; Kz/r3 . 6. The magnetic eld H in cylindrical coordinates is given by H= K a

where K is a constant. Compute H. Ans. Everywhere zero except at = 0. 7. The magnetic eld H in cylindrical coordinates is given by H= K a

8.

9.

10.

11. 12.

where K is a constant. Compute H. Ans. Everywhere zero except at = 0. What is the method of obtaining the equation of any plane in spherical coordinates? How would you go about doing it? Ans. Write the equation in rectangular coordinates, then substitute the values of x, y, z for sperical coordinates. The equation of a cone in spherical coordinates is = 0 . Find the equation in rectangular coordinates. Ans. z = tan 0 x2 + y2. Find the volume of a sphere using the spherical coordinate system through integration. Ans. (4/3)r3 Find the surface area of a cylinder of radius a and length l through integration. A scalar eld 2 2 2 V(r) = e(x +y +z )

nd the volume integral of this eld over all space. 13. The vector eld (cylindrical coordinates and K is a constant) A= K a

is to be integrated in the clockwise direction over the boundary of a unit circle lying on the x-y plane. Find the line integral of this vector eld. By

118

3. Vector Calculus

looking at the structure of A can you guess the answer? 14. Find the gradient of K V= r where V is given in spherical coordinates, and K is a constant. Find the line integral of this vector eld over the boundary of a unit circle lying on the x-y plane. 15. Find of the function V of Problem 14. 16. Using any vector eld A(x, y, z) show that ( A) = 0 a) The line integral of this vector eld from (1, 1, 1) to (1, 2, 3) b) Find the surface integral of this eld over the square region which is described by R = 1 x 1 and 1 y 1 18. For the vector eld D=K xax + ya y + zaz x 2 + y 2 + z2

3/2

nd the divergence of D, K being a constant. Going over to spherical coordinates nd the surface integral of D over a sphere of any radius. How do you explain the results using D dV = D dS ?

19. If the unit of the charge density v is C/m3 then from the equation D = v show that the unit of D is C/m2 . 20. Using the unit of E, which is V/m show that has the unit of F/m.

1. Find the normal to the surface z x2/4 = 8 at the point z = 9, x = 2. Ans. The gradient gives the normal at any point. The grad. of the surface in general is x n = az ax 2 and the unit normal at the point in consideration is n = (az ax) / 2. 2. For the vector eld A = z sin a + z cos a + sin az nd the line integral of the vector eld along the unit circle in the z=1 plane in the anti clockwise direction.

119

3. Vector Calculus

Ans. The line integral is

2 c

A dl =

z cos

c

=1, z=1

d =

=0

cos d = 0

3. Find the ux of A = z sin a + z cos a + sin az out of the closed cylinder bounded by the three surfaces z = 1 and = 2. Ans. Outward ux from the z = 1 surface is

=2 =2

=0 =0 0

=2 =2

=0 =0

0

z=1 =2

(z sin )ddz

z=1 =0

=2

=0

so the ux out of the closed surface is zero. 4. In spherical coordinates a vector eld is given by A= Kar r2

nd the ux out of a closed sphere with radius R. Ans. Outward ux from the r = R surface is

= =2

=0 =0

K R2 sin dd = 2K [ cos ] = 4K 0 R2

5. In rectangular coordinates show that A = 0 where A is any vector. Ans. A = ax y Az z A y + a y (z Ax x Az ) + az x A y yAx and A = x y Az z A y + y (z Ax xAz ) + z x A y y Ax = xy Az xz A y + yz Ax yx Az + zx A y zy Ax =0 the overbar, underbar and underbracket terms cancel each other.

120

3. Vector Calculus

1. The length of the vector dl along the curve x = 2y = 4z2 at the point (1,1/2,1/2) is (a) dx (b) 1.1dx (c) 1.263dx (d) 1.266dx Ans. (d) 2. The length of the the curve x = y = z from (0,0,0) to (1,1,1) is (a) 1 (b) 2 (c) 3 (d) 2 Ans. (c) 3. The length of the the curve x = 2y = 4z2 from (0,0,0) to (1,1/2,1/2) is (a) 2.661 (b) 3.661 (c) 1.2688 (d) 2.2688 Ans. (c) 4. Find the line integral of the vector A = 3ax + 4a y + 5az along the curve x = y = z from (0,0,0) to (1,1,1). The value of the integral is (a) 12 (b) 11 (c) 10 (d) 9 Ans. (a) 5. Find the line integral of the vector A = 3ax + 4a y + 5az along the curve x = 2y = 4z2 from (0,0,0) to (1,1/2,1/2). The value of the integral is (a) 5.5 (b) 7.5 (c) 6.5 (d) 2 Ans. (b) 6. The magnitude of gradient of the scalar eld = ze cos at the point z = 1, = 1, = 1 is (a) 2.2214 (b) 0.4181 (c) 0.5 (d) Ans. (b) 7. The surface integral of the vector A = xyax on the xy plane for 2 x 2 and 2 y 2 is (a) 0 (b) 4 (c) -4 (d) 16 Ans. (a) 8. The surface integral of the vector A = sin az on the xy plane for 0 2 and 0 2 is (a) 2 (b) 0 (c) -4 (d) 16 Ans. (b) 9. The curl of the vector A = sin az is (a) 2 (b) cos az sin a (c) cos a sin a (d) cos a + sin a Ans. (c) 10. The divergence of the vector A = sin az is (a) 0 (b) cos az sin a (c) (d) sin Ans. (a)

1. Find the linear element dl between the points a) x = 1, y = 1, z = 1 and x = 1, y = 1.01, z = 1.001 b) = 1, = 1, z = 1 and = 1, = 1.01, z = 1.001 c) r = 1, = 1, = 1 and r = 1, = 1.01, = 1.001 Hint: See Section 3.2.1.

121

3. Vector Calculus

2. Find the length of the curve x = y2 = z2 from (1,1,1) to (4, 2, 2) Hint: Consider the integral

(4,2,2)

(1,1,1)

G = yax 2.5xa y + 3zaz nd the line integral of G along the straight line from (1,1,1) to (2,3,4) Hint: Consider Example 3.4. 4. If A = 2xyax + 3xya y + 4xyzaz prove Stokes theorem in the at region enclosed by the straight lines (0,0,0) to (1,0,0) to (1,1,0) to (0,1,0) and back to (0,0,0). 5. If A = 2xyax + 3xya y + 4xyzaz prove the divergence theorem in the volume enclosed by the planes z = 0 z = 1; x = 0 x = 1 and y = 0 y = 1. ***Chapter complete*** END OF PART I

122

Part II.

Electrostatics

123

If you wish to reach the highest, begin at the lowest Syrus: Maxims

This chapter introduces the student to the fundamentals of electromagnetism. We start with electrostatics which is the compuation of the electric eld due to stationary charges. The topics covered are: 1. The concept of charge: the idealised point charge and other distributed charges such as line charge, surface charge and volume charge. 2. Next, Coulombs law is introduced which involves the calculation of forces between point charges. 3. Inroduction to the eld concept, that of the electric eld. 4. Calculation of the electric eld due to many charges, and its application to a dipole. 5. Calculation of the electric eld for continuous charge distributions. 6. We discuss the concept of electric ux and the electric ux density vector.

Electrostatics is the study and analysis of the eects of stationary (and almost stationary) charges. These charges may be treated as point charges or distributed charges residing on the surfaces of conductors or dielectrics, or idealised as placed in some region of space. In most cases when we apply the concepts developed in the study of electrostatics we assume that the eects due to the motion of the charges are neglected which in some cases may not be strictly true, but which give us results which are corroborated by experiments. The knowledge of static electricity dates back to the earliest civilisations, but a systematic study of this phenomenon started much later after the development of the scientic method. For example, in 425 BC Democritus proposed that all matter is made of small indivisible particles called atoms. This theory was based more on intuitive reasoning rather than scientic investigation. In the eighteenth century electrostatic machines were developed. In 1747 Benjamin Franklin conducted experiments that demonstrated that one type of electrication could be neutralised by the other type. This indicated to him

124

+ + + + + (a) (b)

+ + ++ + + + +

+ + + ++ + +

(c)

(d)

Figure 4.1.: Charge distributions (a) Point Charge (b) Line Charge (c) Surface Charge (d) Volume Charge

that the two types of electricity were not just dierent; they were opposites and he called one type positive and the other negative. This was followed by experiments conducted by Joseph Priestley in 1766 who proposed that the force between electric charges follows an inverse square law. In 1777 Charles de Coulomb invented a torsion balance to measure the force proposed by Priestly between electrically charged objects (Coulombs law). When a material becomes charged by rubbing or by other means, it has either a surplus or a decit of electrons. A body with a surplus of electrons is said to be negatively charged; a body with a deciency, positively charged. The amount or quantity of charge on a body is expressed in coulombs (positive or negative). A coulomb is an enormous amount of charge, as will be clear through the examples in the book, and, in most electrostatic situations governed by static electricity charge levels of an extremely small fraction of a coulomb give rise to signicant eects. Electrostatic forces exist between charged bodies. Bodies with like charge have repulsive forces between them, while oppositely-charged bodies experience attraction.

4.3. Charge

From the study atomic physics and chemistry, we know that the atom consists of electrons and protons which are negatively and positively charged respectively. Thus the smallest unit of charge, which is the charge on a proton, also called the elementary charge, is e = 1.602 1019 C. We know that macroscopic objects consist of atoms which are neutral and normally, the number of protons and electrons are equal in number in an atom. If, for some reason, (such as rubbing an object with silk) there is surplus of charge on such an object, it exhibits electrical properties. The phenomena in such cases are referred to as static electricity 1 . The SI unit of electric charge is the coulomb, which represents approximately 6.24 1018 elementary charges2 . The coulomb is dened as the quantity of charge that has passes through the cross-section of a conductor carrying one ampere of current within one second. The symbols Q and q are used to denote a quantity of electric charge.

1 2 The

See http://en.wikipedia.org reasoning is as follows: n e = 1 C where n is the number of elementary charges. Hence the result follows.

125

The electrical charge is to the electric eld as mass is to the gravitational eld. Electrostatic elds are produced because of the existence of charge just like gravitational elds are produced due to the existence of mass. In electrostatic formulations charges have four types of idealisations, as shown in Figure 4.1 1. Point charges. The idealisation here is that the full charge Q is concentrated at a point. The unit of such a charge is Coulombs. In this book such charges will be designated by Q or q. 2. Line charges. The idealisation here is that the charge is distributed over a line or a curve. The unit of the charge distributed over a line is C/m and referred to as linear charge density. To calculate the line charge density, l , at a point we take a very small linear element l on the charged curve and nd the charge, q coulombs on it on it. It is intuitively obvious that the charge residing on the linear element l will be proportional to l itself. Thus if l is comparatively large then the charge residing on it, q, will also be large. And if l is small then q will be small. In essence, since the charge q will be proportional to l q l l Or more precisely l = lim

l0

q l

(4.1)

at that point. This can be done for all points on the curve to give a l which is a function of coordinates. 3. Surfaces charges. Here the idealisation is that the charge is distributed over some surface. The unit of such a charge distribution is C/m2 . The mathematical denition of s at a point on the surface is obtained in a similar manner as earlier s = lim q S (4.2)

S0

where q is the charge on a very small surface S. This is done at all points on the surface to give a complete picture of the charge all over the surface. 4. Volume charges. We take a small element of volume V and nd the total charge q contained in it. Then the volume charge density v in C/m3 at that point is q v = lim (4.3) V0 V EXERCISE 4.1 How would you express a point charge Q, mathematically, in rectangular coordinates? Hint: Use the Dirac delta function. Ans. Q(x)(y)(z) EXAMPLE 4.1 One million electrons are equally distributed over a linear region of 1 cm. Find the charge density.

126

Solution: Step 1. We are required to nd the charge density, which consists of 1 million (106 ) electrons (-ve charge) uniformly distributed over 1 cm (.01 m). Step 2. The charge of an electron is 1.602 1019 C. So 106 elctrons have a charge of 1.602 1019 106 = 1.602 1013 (C) Step 3. Since the charge is distributed over a region of 1 cm (=0.01 m) l = 1.602 1013 0.01 = 1.602 1011 (C/m) EXERCISE 4.2 One million electrons are equally distributed over a region of 1 cm3 . Find the charge density. Ans. v = 1.602 107 C/m3 EXAMPLE 4.2 The surface distribution of charges on a plane z = 0 is given by s = 1012 e|x| e| y| Find the total charge in the region R bounded by 5 x 5 and 3 y 3 Solution: Step 1. The total charge, Q, is given by Q=

R

s (x, y)dxdy

R

R

= 1012

5 5

e|x| dx =

0 5

ex dx +

5 0

ex dx = 2

5 0

0 5

ex dx +

5 0

ex dx

0 3

e y dy +

3 0

ey dy

= 1012 2 1 e5

2 1 e3

(C)

127

EXERCISE 4.3 The volume distribution of charges in a region R given by 3 x 3, 3 y 3 and 3 z 3 is given by v = 1014 cos (x/3) e|x| e| y| e|z| Find the total charge in the region R. Ans. 1014 2 1 e3

2

1.0014 C/m3

nd the total charge enclosed in a volume of the sphere of radius r Solution: Step 1. The total charge Q is given by Q=

R

EXAMPLE 4.3 In a region in space the charge density is given by er0 er (C/m3 ) when 0 r r 0 v = 0 when r > r0

v dV

where R is a sphere of radius r. Step 2. When r r0 there is a charge density given by the statement of the problem and when r > r0 there is no charge denity. Step 3. Draw a rough sketch of the charge density. (We will not do this but the student should) Step 4. Since we are considering spherical coordinates, we must write out the volume element in spherical coordinates, which is dV = r2 sin dddr as given in Section 3.2.5. Step 5. The amount of charge contained in such an element is v r2 sin dddr. Step 6. Case 1. r < r0 Therefore the total charge contained in a sphere of radius r < r0 is v r2 sin dddr =

rr0 =0 2 =0 r r=0

2 r

=

=0

sin d

=0 3 r0

d

r=0

(er0 er )r2 dr

r r=0

r e r2 2 r + 2 er = (2)(2) 3 r3 er0 r2 2 r + 2 er + 2 = 4 3

128

Step 6. Case 2. When r r0 the integral remains the same except that the limits change v r2 sin dddr =

r=r0 =0 2 =0 r0 r=0

r0 r=0

It was known to the ancient Greeks as long ago as 600 B.C. that amber, (the Greek word for amber is elektron) rubbed with wool, acquired the property of attracting light objects. In 1747 Benjamin Franklin (1706-1790) supposed that electric re (read charge) is a common element existing in all bodies. If a body had more than its normal share of this re it was positively charged, if less it was negatively charged. Then twenty years later in 1767 it was proposed by Priestley, that the law of electrical attraction was the same as that of gravitational attraction, namely, that the strength of electrical attraction between oppositely charged bodies varies as the inverse square of the distance. It was much later, during the 1780s, that a French engineer, Charles Coulomb (1736-1806) investigated the quantitative relation of forces between charged objects. Using a torsion balance, created by Coulomb himself, he determine how an electric force varies as a function of the magnitude of the charges and the distance between them. Referring to Figure 4.2, what Coulomb found was that 1. The force exerted on a charge q by the charge Q was proportional to the product of the two charges and inversely proportional to the square of the distance between the two F Qq R2 (4.4)

He also found that 2. The force was a repulsive when the two charges possessed the same polarity and attractive when the two charges were opposite in polarity. The direction of the force was along the line connecting the two bodies. We can implement these ideas, using vector notation, into an equation. Figure 4.3 depicts two charges,

129

x

Figure 4.3.: Figure illustrating Coulombs Law

Q1 , and q are the two charges with r1 (Can be read as the position vector from point 1 to point occupied by q) which is the position vector of q with respect to Q1 Then the Coulombs law says, that the force, F, felt by q charge due to the presence of Q1 is given by Q1 q (r r1 ) 1 1 Q1 q R1 = F= 4 40 R2 0 |r r1 |3 1 (4.5)

where : 1/40 = 8.9898 109 is the constant of proportionality and 0 is a constant characterising the medium (which in this case is vacuum3 ). Its unit is F/m. Did you know?/Application: In electrostatic spray painting or powder coating, a charged spray of tiny particles comprising the paint are electrically charged which causes them to repel each other while exiting the nozzle and spread evenly. The spray is charged of one polarity while the surface which is being painted is charged with the opposite polarity. Due to this, the paint is attracted to the object and gives an even coat. The method also ensures that the paint adheres to the surface, reaches hard-to-reach areas and in general gives a far superior coating when compared to ordinary spray painting. EXAMPLE 4.4 Find the force felt by a 1 nC test charge q due to a 1 nC charge Q at distances of 1 cm and 10 cm. Solution: Step 1. The two charges repel each other with a force in accordance with Coulombs law. (they repel each other because they are of the same polarity) Step 2. Since no coordinate system is given we place one of the charges at the

3 Air

130

origin and the other at a distance of 1 cm from it F= q Q 40 R2 1 109 1 109 = 1.1127 1010 (0.01)2 = 8.987 105 (N)

at a distance of 1 cm the force felt by either of the charges is 89.87 N. Step 3. At a distance of 10 cm the force will be 100 times weaker due to the inverse square law. One does not have to do the calculation a second time. EXERCISE 4.4 Find the force felt by a 1 pC test charge q due to a 1 C charge Q at distances of 1 cm and 10 cm. Ans. 89.87 N at 1 cm. and 898.7 nN at 10 cm. EXAMPLE 4.5 In a silver sphere weighing 1 gm, 1 % of all the outermost single electrons are removed. Find the charge density inside the sphere. If the silver sphere was to act like a point charge, nd the force on another charge of 1 nC placed 1 m from it. Solution: Step 1. First we have to calculate the number of atoms in 1 gm of silver. So we proceed as follows: silver has an atomic weight of 107.9; so 107.9 gm of silver has 6.022 1023 (= NA ) atoms. 1 gm of silver has N= 6.022 1023 = 5.581 1021 atoms 107.9

Step 3. These atoms have an excess charge contributed by one proton (e = 1.602 1019C). Therefore the excess charge is Q = 5.581 1019 1.602 1019 = 8.542 C Step 4. Next we need to calculate the volume occupied by a gram of silver. The density of silver is 10.5103 kg/m3 . From this we compute that 1 gm(= 0.001 kg) occupies a volume of V= 0.001 = 95.23 109 m3 10.5 103

Step 5. The volume charge density, v , is therefore v = Q 8.852 = 89.691 MC/m3 = V 95.23 109

131

1 m.

2 cm.

Figure 4.4.: Figure for Example 4.6.

Step 6. The force exerted by Q =8.542 C on q =1 nC charge at a distance of 1 m is given by Coulombs law to be: F= q Q 40 R2 1 109 8.542 = 2 1 1.1127 1010 = 75.79 N

EXAMPLE 4.6 Two metallic balls of 1 gm. each are suspended by 1 m. long threads from a hook as shown in Figure 4.4. Equal charges are placed on each of the two balls, and they are found to be separated from each other by 2 cm. Find the charge on each ball. Solution: Step 1. As a rst step, we need to nd the coulombic force on each ball. Using the diagram on the right, if T is the tension in the string, m (= .001 kg) the mass of each ball, F the coulomb force and g (= 9.8 m/s2 ) the acceleration due to gravity, then the two equations which dene the forces of the problem are T = mg cos(/2) + F sin(/2) F cos(/2) = mg sin(/2) Step 2. There are two equations: which equation should we use? The rst can be rejected since it contains a second variable, T. So from the second equation F = mg tan(/2)

132

Metal rod Insulator Gold leaf Charge Glass housing

+ + + +

+ + + +

Scale

Stand

Step 3. We still need to calculate . From geometrical considerations and using sin (/2) tan (/2) when is small, we get F = mg tan(/2) 0.001 9.8 Step 4. If Q is the charge on each ball then F = 9.8 105 = or Q2 40 (0.02)2 0.01 = 9.8 105 N 1

Q = 2.088 nC By using this method or some allied technique one can accurately measure charges. Application: The Gold Leaf Electroscope. An application of measurment of charge based on Example 4.6 is the gold leaf electroscope by Exner (19101920) (See Figure 4.5). In essence the basis on which the instrument works is the charging of two thin, extremely delicate gold leaves by an external charge introduced through a metal rod. Since the charge causes the two leaves to repel each other a scale is placed at the back of the assembly which reads o the charge residing on the leaves. Generally such an instrument would be caliberated through an experimental procedure rather than through analysis. EXAMPLE 4.7 In the hydrogen atom compare the force of gravity (between the electron and proton) with the Coulombic force of attraction. Step 1. Since the two particles are of opposite polarity both kinds of forces (due to gravity and due to electrostatic attraction) are attractive. If the distance

133

z

x

Figure 4.6.: Coulombs law appicable to three charges

between the particles is one Bohr radius, rB = 0.53 1010 (m) then the force due to gravity is me mp FG = G 2 rB where G is the gravitational constant and me (= 9.10941031 kg) and mp (1.6749 1027 kg) are the masses of the electron and proton respectively. Using G = 6.6726 1011 (mks units) FG = 3.61 1047 (N) Step 2. On the other hand, the force due to Coulombic attraction is given by FC = and 1 e2 = 8.212 108 (N) 40 r2

B

The Coulombic force of attraction is 1039 times the gravitational force. When we examine Coulombs law, we know that there is force on one charge due to another charge. What happens when there are three charges: q (the test charge on which the force is felt), Q1 and Q2 as shown in Figure 4.6. Now the force on q due to Q1 is Q1 q (r r1 ) 1 Q1 q 1 (4.6) F1 = 3 R1 = 4 40 R 0 |r r1 |3 1

where r1 is the vector directed from Q1 to q, r is position vector of q and r1 is the position vector of Q1 . Similarly (and using similar notation) Q2 q (r r2 ) 1 1 Q2 q (4.7) F2 = 3 R2 = 4 40 R 0 |r r2 |3 1

134

is the force on q due to Q2 . So the total force on the charge q is: FT = F1 + F2 = q Q1 (r r1 ) Q2 (r r2 ) + 40 |r r1 |3 |r r2 |3 (4.8)

EXAMPLE 4.8 Find the force on a charge q = 1 nC located at the mid-point of two equal charges of 1 C located at (1,1,1) and (5,3,2) in rectangular coordinates. Step 1. First we nd all the position vectors. r1 (Location of Q1 ) = ax + a y + az r2 (Location of Q2 ) = 5ax + 3a y + 2az and r(Location of the mid point) = And so R1 = r r1 = 2ax + a y + 0.5az (r1 + r2 ) = 3ax + 2a y + 1.5az 2

Step 2. Draw a sketch and show all the vectors on the sketch.

R2 = r r2 = 2ax a y 0.5az where R1 , R2 are the position vectors from Q1 , Q2 to q. Step 3. We now use some reasoning using Coulombs law: since R1 = R2 and Q1 = Q2 therefore F1 = F2 therefore FT = 0. If we extend this concept to N charges, Q1 , . . . QN located at r1 , . . . rN exerting a coulomb force on q at r then FT = Q1 q (r r1 ) Q1 q (r r2 ) Q1 q (r rN ) q + + 3 3 40 |r r1 | |r r2 | |r rN |3 (4.9)

If we examine Equation 4.5 once more, the fact that a force is felt by Q2 must imply some sort of physical connection between the two. Eect (which is the force felt by Q2 ) must have a cause. Taking the case of gravity, we nd that the Earth exerts a pull on the Moon, and the pull is mysterious, since, there is no visible physical connection. Because of these arguments we proceed to take a fresh look at Coulombs law, in which we concentrate on those terms which do not involve q. Rewriting Coulombs Law: 1 Q 1 R1 F{Force on q} = q 4 2 0 R1 = qE1 The term in the square brackets, which we have called E1 , is F 1 Q1 R1 E1 = = q 40 R2 1 135

(4.10)

(4.11)

is independent of q. In the bracketed portion, there is no part played by q. However R1 being a unit vector, gives the full term, the character of a vector. We also know that as we move q around, it will always tend to feel a force created by Q1 , no matter where it is moved. These arguments therefore make us tend to believe that the charge Q1 produces an invisible force eld around itself, and any charges introduced into this eld are subjected to Coulombs force. This force eld we call the electric eld.

z Field point

Figure 4.7.: Electric eld at an arbitrary eld point due to a point charge

Referring to Figure 4.7, where no charge is placed at r, the electric eld is: E= Q1 Q 1 R1 1 1 R1 = 40 R2 40 R3 1 1

where R1 = r r1 . Regarding the units of E, Force U = N/C Charge The electric eld in electromagnetic units is also Volt/meter. Equation 4.11, also says that if the charge is positive then the electric eld points radially away from the principal charge, in the direction R1 while if the charge is negative then the direction of the electric eld points radially towards the principal charge, that is, in the direction R1 . The unit of the electric eld must obviously be the units of4.5 and 4.11 leads to a relation between the force F which q experiences, and the external electric eld E F = qE (4.12)

This equation is valid when the electric eld is produced not only by a single point charge, but by any external system of charges4 . Figure 4.8 shows such a situation. There is an electric eld, E(x, y, z), produced by some system of charges not shown in the gure. If charge q is immersed in this eld, E, it feels a force F = qE(r).

4

Note that a charge does not feel a force due to the electric eld produced by itself!

136

Figure 4.8.: Force felt on a charge q due to an external eld E(x, y, z). F = qE.

EXAMPLE 4.9 (a) Find the distance where the E eld is 1.0 V/m due to a 1 C point charge. (b) If another 1.0 C point charge is placed at this point, nd the force on it. See Figure 4.9. Step 1. (a) If we place the 1 C principal charge Q at the origin, the E eld due to it is given by Q 1 E(r) = ar 40 r2 where r is the position vector of the eld point or observation point. Step 2. We have to nd that distance R where |E| is 1.0 V/m, due to this charge. Hence |E| = 1 Q 40 R2 Q 1 = 2 40 R

137

R2 = 1/40 = 8.988 109 (m). Which implies that R = 94.8 km. From this result it is clear that 1 C is a very large charge indeed. A quick calculation shows that the E eld at a distance of 1 m from this charge is 8.988 109 V/m, which is a very intense eld! Step 3. (b) The force exerted on the 1 C test charge Qt at a distance of 94.8 km is given by F = Qt E Since E is 1.0 V/m and Q is 1.0 C, hence F is 1 N. Continuing with our discussion (refer to Figure 4.3 on page 130) Equation 4.5 which applies to the gure gives the force F felt on the charge q. But the charge q generates an electric eld Eq , of its own, and and Q1 feels a force due to q charge eld. By the force equation: Fon Q1 = Q1 Eq Where (R1 )q 1 40 R2

1

(4.13)

Eq =

(4.14)

Notice the negative sign on R1 . By examining the two equations, Equation 4.11 and Equation 4.14 we can see that though the E elds are dierent from each other the forces on the two charges are equal but opposite each other. The force that the principal charge, Q1 , exerts on the test charge, q, is equal and opposite to the force that test charge q exerts on the principle charge! This situation is the same as what we nd in the case of gravity.5 EXAMPLE 4.10 Find the electric eld at (1,2,3) (rectangular coordinates) due to a 1 nC point charge (Q) located at (1,1,1) Step 1. In this example we rst calculate all the position vectors. Draw a rough sketch. r = ax + 2a y + 3az r1 = ax + a y + az (r is the observation point) (r1 is the position of the charge)

5 The

(V/m)

gravitational eld is such that the force that the Sun exerts on the Earth is the same as that the Earth exerts on the Sun. Though the forces are same, it is the Earth which moves around the Sun, and not the other way around, because of the dierence of the two masses.

138

Did you know? When we want to measure large elctric elds from a moving vehicle, such as an aircraft, the eld mill is one of the best means for such a measurement. The eld mill makes use of a rotating shutter or vane which exposes and then shields an electrostatic eld-sampling probe. This method produces a pulsating voltage on the probe that is proportional to the eld strength. The signal is then averaged and amplied to give the relative eld strength.

The electric eld obeys the principle of superposition. This is so since the Coulombic force is additive. Thus if more than one charge is present, then each charge produces its own electric eld everywhere. The total eld at any point in space is the vector sum of the electric eld of all the charges. To be specic, if single charge Q1 produces an electric eld E1 at the point r and the charge Q2 produces an electric eld E2 at exactly the same point, then it is found that the total electric eld at that point is the vector sum of the two elds ETotal = E1 + E2 (4.15)

this is clear by examining Equation (4.8): This principle can be extended to the case of systems of charges. In other words, if one system of charges, 1 , produces an electric eld E1 (no other elds being present) at the point r and another system of charges, 2 , produces an electric eld E2 at exactly the same point (no other elds being present), then the the two systems of charges, 1 and 2 , acting together, produce a eld ETotal = E1 + E2 (4.16)

at the same point r. With the application of this principle we are able to compute the electric eld, with our present knowledge, due to any system of point charges distributed in space. We are of course considering only the electrostatic case, but this principle is applicable to the general case as well. We now apply superposition to the case of the computation of the electric eld due to the presence of only two point charges based on our knowledge of the electric eld of a single point charge.

In this section we examine the well known problem of the electric dipole. The electric dipole consists of two charges, illustrated in Figure 4.10 of equal magnitude, one a positive one, Q, placed at r1 = (0, d/2, 0) and the other a negative charge, Q, placed at r2 = (0, d/2, 0). We have to nd the electric eld everywhere due to them. The study of the dipole has important consequences in the study of dielectrics which we will encounter later on.

139

y

Referring to the gure, we observe that there is symmetry about the y-axis. If we pass an arbitrary plane through the y axis, the charge conguration will be identical in that plane as it is in the x-y plane. It is because of this that the elds also will be identical in that plane as they will be in the x-y plane. We conclude from this observation that the electric eld has a cylindrical symmetry about the y-axis. The reader is guided to note that in every problem we encounter we look for specic symmetries, which will be rewarded by a great simplication of the problem! Due to the above argument, the electric eld is calculated only in the x-y plane, giving us the condition that the observation point has the position vector r = x, y, 0 (z = 0.) We proceed by calculating the eld due to each charge and then nd the vector sum of the two at the eld point. The electric due to the positive charge, E+ in the x-y plane is given by E+ = where R+ = r r1 = [x, y, 0] [0, d/2, 0] = x, y d/2 , 0 Q R+ 40 R2 + (4.17)

(4.18)

R+ = and

x, y d/2 , 0 x2 + y d/2

2

(4.19)

R2 = x2 + y d/2 + Q R 40 R2

(4.20)

140

Where the terms are R = r r2

2 2

R =

(4.25)

EXAMPLE 4.11 For a dipole nd the resultant electric eld on the x axis and y-axis. Step 1. (a) On the x axis both y and z must be set to zero. (y = 0, z = 0), in the original equation for Er so xax + y d/2 a y xax + y + d/2 a y Q Er = 3/2 3/2 40 2 2 2 x + y d/2 x2 + y + d/2 = Q 40 xax d/2a y xax + d/2a y 3/2 3/2 2 2 2 x + (d/2) x2 + (d/2) da y x2 + (d/2)2

3/2 y=0

Q 40

(4.26)

Step 2. (b) on the y axis (x = 0, z = 0). xax + y d/2 a y xax + y + d/2 a y Q Er = 40 2 2 3/2 2 3/2 x + y d/2 x2 + y + d/2 x=0 y d/2 a y y + d/2 a y Q = 3 40 y d/2 3 y + d/2 = Q 40 da y y d/2

3

It is dicult to visualise the electric eld in any particular case. If the charge distribution is simple we can visualise the electric eld. For example for the

141

case of a lone positive charge at the origin, we know that the electric eld streams away in the ar direction, that is, radially away. Since engineers need to see, or rather get an idea of the eld, numerical procedures exist which plot the direction of the eld in any desirable plane. These are called streamline or eld plots which dened by the equations dx dy dz = = Ex E y Ez (4.27)

If, for example, we are interested in a eld plot in the x-y plane, then we put z = 0 and plot only dy E y (4.28) = dx Ex

0. 8

Q

0. 4

-0. 4

-Q

-0. 8

-0. 8

-0. 4

0. 4

0. 8

Figure 4.11 shows the eld plot of the electric eld for a dipole with inter-charge distance d = 1m. The ax and a y components of the electric eld, namely, Ex and E y are computed on grid points which cover the whole region. The unit vector in the direction of the electric eld is attached to the grid point as shown in the gure. In our case, the region of interest is a square of 2 m 2 m in extent and is covered by a grid of 14 14. One streamline is drawn as an example. The streamline is calculated as follows. A small distance is moved in the direction of y/x calculated from Equation 4.28 by a small amount, Then at the new point the process is repeated till the streamline ends. Let us see how the eld lines look when the two charges have equal values: each of the two charges is a positive charge Q. The eld plot is illustrated in Figure 4.12. The following is the set of observations which one can make. 1. The electric eld streams out from each charge, but the lines do not cross each other. 2. Each set of eld lines are identical because the two charges are identical. 3. From the gure, and other arguments, it is clear that there will be a point,

142

midway between the two charges where the electric eld will be zero.6

2

Q

0

2 0 2

Figure 4.12.: eld plot for two equal but similar charges of magnitude Q and d= 1 m

Field Point

....

z o y x

Figure 4.13.: Electric eld due to a system of charges.

We consider next, the total eld due to a system of point charges at an arbitrary point r(= x, y, z ). Consider a set of point charges Q1 , Q2 , . . . QN placed at positions r = x , y , z , i=1, ... N,7 as depicted in Figure 4.13. The electric eld i i i i due to Q1 placed at r and observed at the eld point r is given by 1 E1 (r) = Q 1 R1 40 R3 1 R1 = r r 1

between the two charges the electric eld due to the upper charge will be equal and opposite to the electric eld due to the lower charge. 7 All throughout the book the prime notation, r or r , will be used for the position vectors of the i charges which produce the eld.

6 Midway

143

y

similarly, the electric eld due to the second charge, Q2 located at r and 2 observed at the same eld point r is given by E2 (r) = Q 2 R2 40 R3 2 R2 = r r 2

in this way, the electric eld, Ei , at r, due to the ith charge Qi whose location is r is i Q i Ri Ri = r r (4.29) Ei (r) = i 40 R3 i but the total electric eld at the observation point r is a superposition of all these elds E (r) = E1 + E2 + + EN i=N = i=1 Ei or

i=N

E(r) =

i=1

Q i Ri 40 R3 i

Ri = r r i

(4.30)

i=N

E(r) =

i=1

Qi 40

x x a x + y y a y + z z a z i i i x x i

2

y y i

z z i

(4.31)

EXAMPLE 4.12 Find the electric eld at the centre of an equilateral triangle whose corners have equal charges Q. Step 1. The charges are placed at the corner of an equilateral triangle. Set up the coordinate system with the origin at the centre of triangle. Draw a sketch

144

of the arrangement. Step 2. Let the three charges be of value Q each with position vectors r , r and r . 3 1 2 If r makes an angle with the x axis and its magnitude is A, then 1 r = A cos ax + A sin a y 1 r = A cos( + 2/3)ax + A sin( + 2/3)a y 2

Step 3. Since we need to nd the electric eld at the origin, r = 0. The Ri i = 1, . . .3 of Equation 4.30 are given by R2 = A cos( + 2/3)ax A sin( + 2/3)a y R3 = A cos( + 4/3)ax A sin( + 4/3)a y and the electric eld is E= = Q R3 Q R2 Q R1 + + 40 R3 40 R3 40 R3 2 3 1 Q (R1 + R2 + R3 ) = 0 40 A3

=0

R1 = A cos ax A sin a y

Step 4. Why? Write out the expression for R1 + R2 + R3 and simplify: since cos( + 2/3) = cos cos(2/3) sin sin(2/3) 3 1 sin = cos 2 2 similarly 3 1 cos( + 4/3) = cos + sin 2 2 cos + cos( + 2/3) + cos( + 4/3) = 0 in the same way we can show that sin + sin( + 2/3) + sin( + 4/3) = 0

so

From Coulombs law we obtained the electric eld due to any arbitrary charge distribution of point charges, Q1 . . .QN by the method of superposition. Although this result looks insignicant, it is highly signicant. We can obtain, by

145

using this result, the electric eld due to any continuous charge distribution, such as any system of volume, surface and linear charge distributions.

Miniscule E Field

dEat r =

1 40

r r r

dQ Miniscule Charge

r

y x

Figure 4.15.: The minuscule electric eld, dE produced by a minuscule charge dQ.

To proceed we have to rst concentrate on the the electric eld produced at a point r by a minuscule charge dQ placed at the position vector r . This is shown in Figure 4.15. Since the minuscule charge produces a eld as in the case of a point charge, its eld is also innitesimal dEat r = dQat r |r r | 1 40 |r r |3

If we look at this equation and contrast it with the equation of a point charge, it is the same except that the charge and the eld are both minuscule quantities. The eld, as usual, is a radial eld with the eld lines emanating from the charge. The charge dQ will be dierent for dierent kinds of charge distributions. If we want to compute the electric eld due a linear charge distribution l then dQ = l dL ; if we are considering charges on a surface with a surface charge distribution s then, dQ = s dS ; or if we have a volume charge distribution, v , then dQ = v dV . Putting down these facts in a single equation l (r )dL dQ = s (r )dS v (r )dV for a linear charge distribution for a surface charge distribution for a volume charge distribution

(4.32)

For the case of a line charge, with l being the linear charge density, E(r) = l l (r ). 146 l dL (r r ) 40 |r r |3 (4.33)

In the same manner we can dene the electric eld at a distant point if the charge is a surface charge or a volume charge.

In this section we apply the concepts and equations just discussed to a specic case, which is a innite line charge with a linear charge density l (C/m) = constant. The line charge is placed along the z axis of a coordinate system, shown in the accompanying gure, which is Figure 4.16. The innite line charge is the the most basic of all cylindrical structures. The other cylindrical structure like the charged innite straight wire of nite thickness has a eld which is similar to one produced by this structure. From this conguration we may also derive the electric eld of other congurations for example, the two conductor lines.

Line Charge z

dL = dz r (0, 0, z )

r r r(x, 0, 0)

x

Figure 4.16.: Figure showing a line of charge of densityl C/m from to along the z axis

To compute the electric at an arbitrary point we have to apply Equation 4.33 to the conguration shown in the gure. In the most general case r = x, y, z is the eld point and r = x , y , z is point where the charge producing the eld lies. So r r = x x, y y, z z . The three components of the electric eld from the general equationEquation

147

4.33 may be spelt out as Ex = Ey = Ez = l dz 40 l dz 40 l dz 40 x x

3/2

3/2

3/2

But we know that r is constrained to lie on the z axis since all the charge is along the z-axis. This translates to the equation: r = x = 0, y = 0, z . The integration is performed over the z coordinate. (l dL l dz ). Furthermore, from the symmetry of the problem it is clear that the electric eld will be the same along any plane parallel to the x-y plane. Hence we can safely evaluate the eld specically, on the x-axis: r = [x, 0, z = 0] and obtain values of the eld everywhere. Using these values: r r = [x, 0, z]. The integration is performed in the interval z [, ]. E= l (r r ) dz

Looking at these terms one at a time, we examine the ax part of the eld. (Keep x constant!), the indenite integral (See Appendix) dz z = (x2 + z2 )3/2 x2 z2 + x2 therefore l xdz z 2 + x 2

3 2

40 |r r |3 l dz x = 40 2 z + x2

z

3 2

, 0,

z z 2 + x 2

3 2

(4.34)

(4.35)

= l x

dz z 2 + x 2

3 2

l x z x 2 z 2 + x 2

(4.36)

which the reader may verify by dierentiation. Putting the integration limits at z = we get l /x. Ex is: Ex = The third term l 2 0 x l z dz z 2 + x 2 2

3

(4.37)

(4.38)

is zero since the integrand is an odd function of z . Thus the total electric eld

148

-4

-8

-8

-4

Figure 4.17.: Field plots for the innite line charge: Streamline plot of the innite line charge in the xy plane

-4

-8

-8

-4

Figure 4.18.: Field plots for the innite line charge: Streamline plot for the innite line charge in a plane passing through the z axis

is E=

l ax 2 0 x

(4.39)

in the Cartesian coordinate system. Due to the rotational symmetry of the problem, cylindrical coordinates will yield a simpler result. E= l a 2 0 (4.40)

Why? Due to symmetry, ax will become a ; also x is the distance from the line charge to the observation point, and so x may be replaced by . Note that the result is the same as the one we obtained by application of Gausss law. To help the reader visualise the elds, the eld plots for the innite line charge in the x-y plane and also the plot for any plane passing through the z axis are shown. The eld is visualised as streaming away from the charged line cylindrically since there is one component, namely, the directed component. Since 1 E As we approach the z axis ( = 0) the eld becomes more and more intense and tends to innity as 0. As we increase the value of , the eld falls o toward zero.

149

Figure 4.19.: Computation of the electric eld on the axis of a short line charge

EXAMPLE 4.13 Compute the electric eld at point P, on the axis of a short line charge of length 2a as shown in Figure 4.19. Step 1. The electric eld is given by E(r) = l dl (r r ) = 40 |r r |3

z=a z=a

l dz (xax z az ) 40 x2 + z2 3/2

z=a

xdz

3/2 z=a x2 + z2

l z = 40 x z2 + x2

z =a

=

z =a

l a 20 x a2 + x2

using the tables of integrals in the appendix. Step 3. Also from arguments from earlier in the chapter Ez = 0 Step 4. In cylindrical coordinates, by inspection, x becomes and ax becomes a l a E = 20 a2 + 2 Step 5. If the line becomes innitely long then a then E = l 20

EXERCISE 4.5 Compute the electric eld at point P, on the axis of a short line charge (as shown in Figure 4.19) but of length a + b when the charge exists from a z b . Calculate the elds for the special case of a = 0 and b =

150

Figure 4.20.: Computation of the electric eld on the axis of a ring of charge

z=b z=a

l dz (xax z az ) 40 x2 + z2 3/2

EXAMPLE 4.14 Compute the electric eld at point on the axis of a ring of charge of radius a as shown in Figure 4.20. What happens to the eld as a z? 151

0.4

0.35

0.3

0.25

0.2

0.15

0.1

0.05

0 0 1 2 3 4 5

Figure 4.21.: Plot of the E eld along the axis of a ring of charge

Step 1. To calculate the electric eld at r = (0, 0, z) due to a ring placed on the xy plane we use E(r) = l dl (r r ) l = 3 40 |r r | 40 ad zaz aa a2 + z2

3/2

Step 2. Symmetry: due to symmetry of the problem, the a component will be zero and therefore E(z) = Step 3. As a z (but a Q = 2al : l l az 2azaz = 3/2 40 a2 + z2 20 a2 + z2 a 3/2 z

E(z) = lim

a0

which is the electric eld at (0, 0, z) of a charge located at the origin. A typical plot of the eld is shown in Figure 4.21 EXERCISE 4.6 Do Ex. 4.14 using rectangular coordinates.

Let us consider another case: that of an innite sheet of charge. We would like to compute the electric eld produced by such a sheet. As shown in the accompanying gure, Figure 4.22, the sheet coincides with the z=0 plane and the charge on it is s C/m2 , which for the case under consideration, has a constant value. The sheet covers the entire x-y plane. We expect the eld to be the same in the x and y directions. As we move in either the x or y directions, but maintaining the same distance from it, the eld should be unchanged. Therefore the eld will not be a function of either the xor y- coordinates, but it will be a function of only the z coordinate. To calculate

152

z

dS = dx dy

r r

r(0, 0, z)

...

O

...

Sheet of charge

...

...

r (x , y , 0)

Figure 4.22.: Calculation of the Electric eld due to an innite sheet of charge

is to be applied in the proper manner. In this equation, dS = dx dy where dS is a small element of area on the sheet of charge. The general form of the vectors r (the eld point) and r (position vector of the charges) are r = x, y, z and r = x , y , z . In our specic case r and r reduce to r = [0, 0, z] and r = x , y , 0 , since the sheet occupies the z = 0 plane which is innite in extent, we expect Ex and E y to be zero. Using the two vectors r and r for Ez Ez =

xy plane

s dx dy 40

z x2 + y2 + z2

3/2

This vector has to be integrated over the whole x-y plane and where the variable z must be treated as a constant. Or

, x =, y =

zdx dy z2 + y2 + x2

3 2

Integrating with respect to x (please refer to the integral given in Equation 4.35): s dy 40

x=

zdx z2 + y2 + x2

3 2

y2 ).

is z/(z2 + Substituting these limits, the integral becomes We now integrate this expression with respect to y between the limits y = y0 to y = +y0 and then let yo . The integral The value of this integral at its upper limit, and lower limits x 2z/(z2 + y2 ).

s dy 40 z2 + y2

x z z2 + y2 + x

x=

(4.41)

153

z z=d/2 + y x _ z=-d/2

s 40

+y0 y0

s y0 2 zdy = 4 arctan 2 + y2 40 z z

(4.42)

This result can be veried by referring to the Appendix. This expression becomes y0 lim 4 arctan y0 z 4 /2 for z > 0 = 4 /2 for z < 0 when z > 0 when z < 0 (4.43)

the other components of the electric eld being zero. This is the same result which we obtained by application of Gausss law. It is important to note that that for all structures, the electric eld points away from the positive charges. We shall return to these results later in the book. See Richard P. Feynman & Sands (2001) Jordan & Balmain (1968). EXERCISE 4.7 Two charged sheets, innite in extent are placed at z = (d/2) with surface charge densities of s (at d/2) and s C/m2 (at d/2) respectively. Show, using superposition that there is an electric eld between the two sheets and none outside. Obtain the value of the electric eld between the sheets and its direction. Remember this result for the case of a parallel plate capacitor, which is considered later in the book Hint: The electric eld everywhere is the superposition of the electric eld due to the two sheet charges. The upper sheet produces a E eld given by /(2 )a s 0 z Eu = /(2 )a s 0 z z > d/2 z < d/2

0

(4.44)

Similarly, the electric eld due to the lower sheet produces a E eld given by /(2 )a z > d/2 s 0 z El = /(2 )a s z < d/2 0 z 154

With reference to Figure 4.23 we see that there are three regions. 1. z > d/2. 2. d/2 > z > d/2. and 3. z < d/2. In Regions 1 and 3 the eld is zero. In Region 2 the eld is E = s /0 az d/2 > z > d/2

EXAMPLE 4.15 Obtain the electric eld along the axis of a charged disk with s (C/m2 ). Step 1. From the gure r = a (with 0 a 0 2) and r = zaz Step 2. So r r = zaz a and |r r |3 = 2 + z2 Step 3. The electric eld is given by (Set E = 0) Ez =

disk 3/2

s d d 40

z

3/2 2 + z2

=

=0

s d 20

z 2 + z2

3/2

s 20

z 1 2 + a2 z

EXAMPLE 4.16 Find the E eld on the axis of a at ring of charge with charge density s (C/m2 ) and with inner outer radii equal to a and b respectively as shown in the gure. Step 1. From the gure r = a (with a b 0 2) and r = zaz r r = 2 + z2 r r = zaz a

3 3/2

155

Small amount of flux Dielectric

-Q

Step 2. The electric eld along the axis of the ring is given by Ez =

ring b

s d d 40 s d 20

z

3/2 2 + z2

=

=a

z

3/2 2 + z2

s z 20

1 1 2 + a2 2 + b2 z z

Faraday conducted experiments with concentric spheres and found some astonishing results. A sphere of with charge Q was placed within a larger sphere containing dielectric. The outer sphere was then earthed for a short while and

156

the inner sphere was removed. It was found that the outer sphere contained the same amount of charge as the inner sphere (but of opposite sign). These results held no matter whatsoever the size of the spheres or whatever the dielectric. Faraday hypothised that there was an electric ux, which moved from the the inner sphere to the outer one equal in magnitude to the charge independant of the size of the spheres and the dielectric : =Q (C) (4.45)

It was further hypothised that there was an electric displacement from the inner to the outer sphere and on any sphere between the two (shown dotted in the gure) there was a displacement density D whose magnitude would be D= Q = 4r2 4r2 (C/m2 ) (A)

Using these ideas a further development would be that the ux and displacement density would be related by = D S and = D dS (4.47) (4.46)

After a little reection we could visulaise that Equation A above can be written in vector form as Q ar D= (4.48) 4r2

Carl Friedrich Gauss formulated, in 1835, a law applying to electrostatic elds which connected arbitrary charge distributions to the electric elds produced. Gausss law is a law relating the distribution of charge to the resulting electric eld. Gausss law states that: The electric ux through any closed surface is equal to the enclosed electric charge. In mathematical terms, the electric ux = =Q d =

S

D dS

(4.49) (4.50)

(Enclosed charge)

157

and using the results of Section 4.3 Q=

V

v dV

we have as a result

V

.D dV =

v dV

V

(4.52)

The equation, applies equally to both situations where the eld is time varying and the charges non-stationary and the other case of electrostatics, where the charges are stationary and the elds are time-independent. The equation will be applied here to the latter case. D is called by various names: the electric displacement density and the electric ux density, with units of C/m2 . v is the volume charge density with units of C/m3 .

Surface

Charges Volume

Equation 4.51 says that the volume integral of the divergence D over a volume V is equal to the surface integral of D over the closed surface S enclosing the volume V. . . . dS represents the integral over a closed surface enclosing the volume. Such a volume-surface combination is shown in Figure 4.27. If we examine Figure 4.27 and apply Gausss law to it, then the volume integral of the divergence of D is equal to the total charge (Expression 4.54) enclosed. v dV

V

(4.54)

We can perform these integrations provided we know both D and v . Provided that the eld and charge distribution are the correct ones, there should be equality between the two integrations. Though only the volume charge, v , has been specically mentioned it includes the cases of surface, linear and point charges.8

8

The case of point charges, surface charges and linear charges can be converted into volume

158

What is the relation between D and E? These two elds are rather closely related. The relation between the two is given by D = E (4.55)

at every point in space. is the permittivity of the medium being considered. If the medium is not vacuum, the for a medium with relative dielectric constant r = 0 r (4.56) To be specic D = 0 E for free space. EXAMPLE 4.17 The electric eld in a region of space is D = (k/r)ar 0 r 1 (spherical coordinates). Find the ux leaving the region R bounded by 0 /2, /4 /2 and 0.25 r 0.5. Also nd the total charge inside this region. Step 1. Let us rst draw and understand the bounding surfaces. R is bounded by six surfaces: 1. and 2. 0 /2, /4 /2 with r = 0.25 or r = 0.5 with dS ar 3. and 4. /4 /2 and 0.25 r 0.5 with = 0 or = /2 with dS a (4.57)

5. and 6. 0 /2, and 0.25 r 0.5 with = /4 or = /2 with dS a Step 2. Out of these surfaces only surfaces 1 and 2 have ux coming in or going out, since the surface vector dS ar . Surfaces 3-6 have dS a or a . Step 3. So for surface 1 (ux entering is negative) 1 = DdS = k ar ar r2 sin r dd = 0.25k 1 = 0.556k 2 2

S1

S1

r=0.25

S1

D dS =

S2

k ar ar r2 sin r

r=0.5

dd = 0.5k

1 = 1.112k 2 2

Step 4. Algebraically adding the two uxes, the total ux leaving the region is 1.11k 0.556k = 0.556k. Step 5. To nd the charge enclosed we need to rst nd D: D =

2 1 r Dr r2 r k = 2 = v r

charges by means of the Dirac delta function. For example a uniform surface of s C/m2 covering the x-y plane can be written as s (z) C/m3 .

159

and DdV = 1 k 2 r sin dddr = k 0.25 = 0.556k 2 r2 2

And we have shown that the total ux leaving is equal to the enclosed charge.

B Sphere 2

c Sphere 1 A a b

Going back to the ux equation, Equation 4.49, we know from examining it, that the total ux leaving a closed surface is equal to the total charge inside it. Figure 4.28 illustrates this point. The volume charge, v , produces a D eld all around it, a small portion of which is shown in the gure. Two Gaussian spheres9 are drawn, one which does not enclose the charge, while the other encloses it. The dierential ux (= D S), through the dierential surface A is equal to that through surface B. This implies that the small ux entering b on the surface of Sphere 2 leaves through the small surface c. In this way all the ux entering the surface of Sphere 2 leaves through some other part of the surface of the sphere. If we give a positive sign to the ux leaving the surface as positive then the ux entering it is negative, and the net ux leaving is zero. This is not the case with Sphere 1. Here, the ux only leaves the surface and none enters it and the net ux leaving the is equal to the total charge inside the sphere. The most important fact to remember from Gausss law is that the ux density diverges from positive charges!

160

z Gaussian sphere

y Q

x

Figure 4.29.: Gausss Law applied to a point charge

Gausss law is best understood by applying it to various situations. With a minimum of eort one obtains spectacular results and the mathematics will be fairly straightforward, but the reasoning required to obtain the results will be acute. Also, the results obtained in this chapter will corroborate the results of the next chapter. This aspect will give great satisfaction and will further increase our motivation to study the subject. As a start let us apply Gausss to Figure 4.29. A little discussion will help. Since we are talking about a closed surface, and a point charge, we draw that closed (or Gaussian) surface which is equidistant from the point charge viz., a sphere with radius r0 . We do this to from the point of view of the following reasoning: since all points are equidistant from the charge, 1. We expect that the D eld will be equal at all points on the surface. 2. Furthermore we also expect that D will have only a radial eld component. Therefore, in the spherical coordinate system, (r, , ) D = Dr ar (4.58)

To calculate D dS, an element of area on the sphere is (see Figure 3.9 on page 90) dS = ar (r0 sin d)(r0 d) (4.59)

161

so that D dS = (Dr ar ) ar r2 sin dd 0 = Dr r2 sin dd 0

sin x 0

Integrating this over the sphere. (Note that constants in this case)

2 sphere =0 2

D dS =

=0

Dr r2 sin dd 0

= Dr r2 0

d

=0 =0

sin d

sphere

(4.61)

(4.62)

(4.63)

But r0 can assume any value, so letting r0 be replaced by the more general value r Q (4.64) Dr = 4r2 If we compare the electric eld at the same point in the same situation. (Note: r = 0 and r = r in Equation 4.11 on page 135.) 0 E= Which means that D = 0 E (4.66) Q ar 40 r2 (4.65)

EXAMPLE 4.18 Compute D for all values of r for the case of a point charge. Do not include the point r = 0. Explain your result. Solution: For a point charge D is (in spherical coordinates) D= Q ar 4r2 (A)

162

and A in spherical coordinates for any general vector is A =

2 1 (sin A ) 1 A 1 r Ar + + 2 r r sin r sin r

in this formula we substitute only Dr as per Equation (A) given above (D , D = 0) Q Dr = 4r2 and 2 1 r Dr 1 (Q/4) D = 2 = 2 =0 r r r r Why? This is so because there are no charges anywhere, except at the origin, where there is a point charge.

z Charged Sphere with Radius R0

Unit Vector ar

Figure 4.30.: Applying Gausss Law to a uniformly charged hollow sphere with radius R0 . The Gaussian surface is a sphere with radius r0 .

Let us apply Gausss law to yet another case, that of a hollow sphere whose surface is uniformly charged, shown in Figure 4.30. The charged sphere has a radius R0 while the Gaussian sphere has a radius r0 . The Gaussian sphere may be expanded or contracted at will, to consider dierent situations, by changing the value of r0 . The total charge on the surface of the sphere is considered to be Q. Two cases can be clearly distinguished Case 1. In this case, the Gaussian sphere has a radius less than the radius of the charged sphere r0 < R0 . D dS = 4Dr r2 0 (4.67)

SGaussian Sphere

Since the Gaussian sphere has a radius less than the charged sphere, no charge

163

is enclosed. So v dV = 0

Gauss sphere

(4.68)

by equating the two results, Dr = 0, D = 0. There is no eld inside the sphere. Case 2. The Gaussian sphere has a radius which is larger than the charged sphere. In this case D dS = 4Dr r2 0 when r0 > R0 (4.69)

V

(4.70)

Dr =

Q Q 4r2 4r2 0

(4.71)

Where we have replaced r0 by the more general value of r. Using r, D = (Q/4r2 )ar . To sum up 0 inside the charged sphere r < R0 D= Q (4.72) 2 ar outside the charged sphere r > R0 4r The important thing to remember is that Gausss law can be applied successfully whenever there is a symmetry about the problem. EXAMPLE 4.19 If an electric eld exists at the surface of the earth which is about 110-150 V/m, radially downward. What is (a) The total charge dispersed on the surface of the Earth? And (b) the charge density at the surface? Using the spherical coordinate system, (r, , ), the electric eld has only one component, namely, the radial component. Er = 110 V From this section the ux density at the surface of the earth is Dr = 0 Er = 8.854 1012 (110) = 9.7394 1010 C/m2

where we have taken the lower value of the eld. The radius of the Earth is 6350 km, so 9.7394 1010 = Q 4r2 E

164

rE is the Earths radius. Solving for Q Q = 439502 C And the surface charge density is s = Q = 9.7394 1010 C/m2 4r2 E

Notice that the surface charge density, s , is equal to the Dr ! EXERCISE 4.8 Find D everywhere for the hollow sphere with radius R0 and surface charge s . Explain your result. EXAMPLE 4.20 A sphere of radius R0 is uniformly charged, throughout its entire volume with a a charge density v . Find the electric ux density eld for all values of r in spherical coordinates, by using a Gaussian sphere. Solution: Step 1. The charged sphere and the gaussian sphere are shown in Figure 4.31. Since the problem has spherical symmetry, we expect only one component of D, namely D = Dr (r)ar therefore applying

Gaussian Sphere

D dS =

v dV

V

or Dr (4r2 ) =

165

Gaussian Sphere

D dS = Dr (4r2 )

and v so

4 dV = v R3 3 0 Dr = v R3 0 3r2

The plot of Dr versus r is shown in Figure 4.32. EXAMPLE 4.21 How will you check whether the answer in the previous example is correct? Solution: We know that D should be equal to v inside the charged sphere and zero outside. Therefore for r < R0

2 1 r Dr D = 2 r r r r2 v 3 1 = 2 r r = v

Let us consider another couple of problems which exhibit symmetry. The rst of these is the innite line charge coincident with the z-axis and with a linear charge density of l (C/m) as shown in Figure 4.33. The obvious thing to do

166

z Infinite line charge

Gaussian surface

here is to choose a cylindrical Gaussian surface whose radius is 0 and whose end surfaces lie between z = z0 and z = z1 . From the symmetry of the gure it is clear that the D eld will be streaming radially away. In the cylindrical coordinate system, , , z , D will be a D : that is, there will be no a or az component of D. Applying Gausss law to the surface (see Figure 3.8 on page 89) we nd that the surface can be split up into three parts: the top, the bottom and the side. The intergration over the top of the cylinder is zero since D dS = (a D ) (az dd) = 0 (a az = 0) (4.73)

top

bottom

Therefore the contribution from the side of the cylinder is the total contribution D dS = =

z=z0 =0 z1 2

side

(a D ) (a ddz)

z1 2

D 0 ddz ddz

z=z0 =0

= D 0

(4.75)

167

Figure 4.34.: An innite cylindrical hollow tube of radius 0 with surface charge s

cylinder

D dS = 2D 0 (z1 z0)

(4.76)

z1

v dV =

cylinder z=z0

l dz dz (4.77)

z1

= l

z=z0

= Total charge enclosed = l (z1 z0 ) Equating these terms D 0 (z1 z0 ) (2) = l (z1 z0 ) D = l 20

(4.78)

(4.79) (4.80)

But, as earlier, 0 can be anything, so replacing 0 by the more general value of l D = (4.81) 2 And the E eld is a E where E = D 0 = l 20 (4.82)

168

EXAMPLE 4.22 Find the electric eld everywhere for the case of an innite hollow cylindrical tube of radius 0 , (as shown in Figure 4.34) with a surface charge of s C/m2 . Use a cylindrical Gaussian surface. We know from the symmetry of the gure that the D eld will have only a component D = D ()a and using the Gaussian cylindrical surface shown dotted in the gure for the case > 0 we apply Gausss law D ddz(= 2 l D) = Charge enclosed = s l 20 0

,z

then D = s

> 0

On the other hand when < 0 the charge enclosed is zero and D = 0 < 0

z Gaussian surface infinite sheet charge at z=0

...

...

above the sheet

... x

...

below the sheet

Let us apply Gausss law to a case of an innite sheet charge. Figure 4.35 shows an innite sheet of charge, coincident with the x-y plane extending in all directions to innity. The Gaussian surface is a rectangular box extending along the x axis from x = x0 to x = x1 ; along the y axis from y = y0 to y = y1 ; and centred along the z axis from z = z0 to z = +z0 . The charge on the sheet is a constant, = s C/m2 . From symmetry, on the walls of the Gaussian surface, the

169

D eld is assumed to have the form [0, 0, Dz ] for z > 0 D = [0, 0, Dz] for z < 0 Dz = constant on the walls of the top and bottom of the Gaussian surface (4.83) With this structure of D the ux out of the sides of the box will be zero, while the ux from the top and bottom will be top = (x1 x0 ) y1 y0 Dz bottom = (x1 x0) y1 y0 Dz so the total ux will be total = top + bottom = 2 (x1 x0 ) y1 y0 Dz The total charge enclosed by the Gaussian surface is Qtotal = s (x1 x0) y1 y0 equating the right hand sides of the previous two equations 2 (x1 x0) y1 y0 Dz = s (x1 x0 ) y1 y0 or Dz (x, y, z0) = and Dz (x, y, z0) = Dz (x, y, z0) = The electric eld will be s Ez = 20s 2 for z > 0 for z < 0 (4.91) s 2 s 2 (4.88) (4.87) (4.86) (4.84) (4.85)

(4.89)

(4.90)

the other eld components are zero. Some good references for this chapter are Jordan & Balmain (1968), W. H. Hayt & Buck (2001), and Richard P. Feynman & Sands (2001).

Distributed charges are calculated as per

The line charge: l = lim

l0 q l q S q V

S0

V0

170

Coulombs law says, that the force, F, felt by a charge q placed at r due to the presence of Q1 at r1 is given by F= Q1 q (r r1 ) 1 40 |r r1 |3

The total electric eld at the observation point r due to Q1 at r , Q2 at 1 r . . . and QN at r is a superposition of all the elds produced by the N 2 charges alone E (r) = E1 + E2 + + EN i=N = i=1 Ei or E(r) =

i=1 i=N

Q i Ri 40 R3 i

Ri = r r i

The electric eld due a continuous charge distribution is given by Eat r = where l (r )dL dQ = s (r )dS v (r )dV for a linear charge distribution for a surface charge distribution for a volume charge distribution dQat r |r r | 1 40 |r r |3

Gauss law states that the ux leving a region of space is equal to the charge enclosed: =Q = = (C) v dV D dS

171

Chapter Summary

In this chapter the following points are of considerable interest: There are various types of charge such as the idealised point charge, Q, the line charge, l , surface charge, s , and volume charge, v . The SI unit of charge is the Coulomb which is also 1 A/s. The net charge can be calculated from the charge line distribution as Q= or the surface charge distribution as Q= Coulombs law is written as F= q1 q2 aR 4o R2 s dl l dl

The student learns to obtain the electric eld for continuous charge distributions. Gauss law is next considered D = v D dS = (in point form) v dV (in integral form)

Review Questions

1. Write a short note on Gauss law and the signicance of electric ux. 2. Write a short note on Coulombs Law. 3. What is the relationship between the force on a point charge and the electric eld? 4. Why is Gauss law only applied to cases of symmetry? 5. For non-symmetrical cases which equations will we use to calculate the electric eld?

Problems

1. A 2 C point charge is located at A(4, 3, 5) in free space. The electric eld at P (8,12,2) (in cyclindrical coordinate) is ? Ans. E = 159.7a + 27.4a 49.4az 172

2. The charge density on the z axis varies as 108 / z2 + 1 C/m for < z < . How much charge lies on the z axis from 10 z 10? Ans. 2 108 tan1 (10) C. 3. The charge density on the z axis varies as 108 z2 for 1 < z < 1 C/m l (z) = 0 otherwise 4. The charge density on the z axis varies as 108 / z2 + 5 C/m for < z < . How much charge lies on the z axis from z ? Ans. / 5 C. 5. The charge density on the z axis varies as 108 / 4z2 + 25 C/m for < z < . How much charge lies on the z axis from z ? Ans. /10 C. 6. The surface charge density in cylindrical coordinates varies as s ( = 1, , z) = 1/ z2 + 1 (C/m2 ) for < z < How much charge lies on the z axis from 10 z 10? Ans. (2/3) 108 C.

How much charge lies on the innite cylinder? Ans. 22 C. 7. The surface charge density in cylindrical coordinates varies as s (, = 0, z) = e|z| / 2 + 1 (C/m2 )

11. Two charges, Q1 = 1 nC and Q2 = 1 C are placed at point 1= (0,0,0) and point 2= (3,4,0) respectively. Find the force of the charge placed at 1 on 2 and vice versa. Ans. 0.3590 106 3ax + 4a y /5 N and 0.3590 106 3ax + 4a y /5 N. 12. Find the electric eld in both cases for Problem 8. Ans. 8.9875 ax V/m and 8.9875ax V/m.

9. Two charges, Q1 = 1 nC and Q2 = 1 C are placed at point 1= (0,0,0) and point 2= (1,0,0) respectively. Find the force of the charge placed at 1 on 2 and vice versa. Ans. 8.9875 106ax N and 8.9875 106ax N. 10. Two charges, Q1 = 1 nC and Q2 = 1 C are placed at point 1= (0,0,0) and point 2= (1/ 2,1/ 2,0) respectively. Find the force of the charge placed at 1 on 2 and vice versa. Ans. 8.9875 106 ax + a y / 2 N and 8.9875 106 ax + a y / 2 N.

How much charge lies on the half plane? Ans. /2 C. 8. Equal charges of 1 nC are placed at point 1= (0,0,0) and point 2= (1,0,0). Find the force of the charge placed at 1 on 2 and vice versa. Ans. 8.9875 109 ax N and 8.9875 109ax N.

173

13. Find the electric eld in both cases for Problem 11. Ans. 0.3590 103 3ax + 4a y /5 V/m and 0.3590 3ax + 4a y /5 V/m.

14. Equal charges of magnitude q = 1 nC are placed on the corners of a square of side 1 m. coincident with the xy plane. Find the force on a 1 nC charge placed on the (a) centroid of the square (b) 0.5 m vertically above the centroid. Ans. (a) 0. (b) -5.084108az N. 15. 10 nC charges are placed at (1,0,0) and (-1,0,0). Two opposite charges (-10 nC) are placed at (0,1,0) and (0,-1,0). Find the force on the charges at (1,0,0) and (0,1,0). Ans. 4.1082 1007ax N and 4.1082 1007a y N. 16. A charge Q0 is placed at the origin. At a distance of 1 m from it the electric eld is 100 V/m. Find Q0 . Ans. 1.113 108 C

17. Two charges, q, are placed at (a, 0, 0) and (a, 0, 0). (a) Where in all 3 space is E = 0? (b) Ex = 0? Ans. (a) (0,0,0) (b) yz plane 18. Four charges q are placed at the corners of a square in the xy plane, two opposite corners of which are (a, 0, 0) and (a, 0, 0). Find the electric eld E at (2a, 0, 0). Ans. 1.4202q / 40 a2 ax V/m. 19. Two equal charges of 10 nC are placed at (1,2,3) and (-3,-2,-1). Find the electric eld at the origin. Ans. 3.431(ax az ) V/m.

23. On the surface of a sphere of radius a, the electric eld is given by E = aar + a2 sin a + 5a nd the charge enclosed. Ans. 4a3 C

22. Calculate the divergence of E = rar + r2 sin a + 5a at (r = 1, = /2, = 0). If r = 1 at that point then what is the charge density, v , there? Ans. E = 3 + 2r cos; v = 30

21. In cylindrical coordinates two equal charges (q) are placed at ( = 1, = /2, z = 3) and ( = 1, = /2, z = 3). Find the electric eld at the origin. Ans. 5.6839 108qa y + 1.7052 109qaz V/m

20. Two equal charges of 10 nC are placed at (r = 1, = /2, = 3) and (r = 3, = /2, = 1). Find the electric eld at the origin. Ans. 83.58ax 4.28a y V/m.

1. If two charges, one of 1 mC of mass 1 g and the other of 1 C of mass 1 kg are placed in a uniform eld of E =1 V/m what will be forces and accelerations felt by them? Solution. The 1 mC charge will feel a force of 1 mN and the 1 C charge of 1 N. (F = qE). The accelerations of both will be the same. 1 103/103 = 1 m/s2 . (a = F/m)

174

2. If two charges q1 (>0) and q2 (<0) of opposite polarity are placed a distance d apart, will there be a point on the straight line joining them where the electric eld is zero? Solution. Suppose there is such a point. Place a small charge q > 0 at that point. Now q1 will repel it toward q2 and q2 will attract it toward itself. Therefore q will always feel a force toward q2 . Hence there will be no such point. 3. A charged circular disk of radius a is placed on the x-y plane, with its axis coinciding with the z axis and having a charge density s . On intuitive grounds what should be the electric eld at 0 < z a (very small) and z a (very large)? Solution. The electric eld on the axis of the disk is given by Ez = s 20 1 z z2 + a2

For 0 < z a z2 can be neglected compared to a2 so Ez = s /20 1 When z is very large we expect Ez = a2 s Q a2 = = s 40 z2 40 z2 40 z2 z a

4. A charge Q is placed at the center of a thin (thickness=) bakelite sphere of outside radius a. What would be the electric eld at a radius of 2a? A charge Q is placed on the surface of the bakelite sphere. What would be the electric eld now? Solution. The D eld at 2a would be given by D= so E= Q ar 4(2a)2

Q ar 40 (2a)2

when Q is placed on the bakelite sphere, the total charge enclosed in the sphere of radius 2a is zero. So D = E = 0 5. Two hollow cylindrical charged innite cylinders with radii 0 < a < b with surface charge densities sa and sb C/m2 are placed concentrically. Find the electric eld at = (a + b)/2. For a hollow cylinder of radius 0 and surface charge s D = s 0 > 0 < 0

D = 0

175

using superposition D = sa

2a a a+b

1. The electric eld E1 at point r due to Q1 at point r1 is given by (a) E1 = (1/40)Q1 r1 /|r1 |3 (b) E1 = (1/40)Q1 (rr1 )/|rr1 |3 (c)E1 = (1/40)Q1 (r r1 )/|r1 |3 (d) E1 = 1/40 Q1 r/(|r r1|3 ) Ans. (b) 2. A force of attraction exists between (a) Two positively charged bodies (b) Two negatively charged bodies (c) Two oppositely charged bodies (d) When one body has a charge which is twice than that of the rst one Ans. (c) 3. What is not the unit of 1/40 ? (a) [F/m] (b) m[C/V]1 (c) [mF]1 (d) None of These Ans. (c) 4. A uniform volume charge with density of 0.2C/m2is present throughout the spherical shell extending from r = 3 to r = 5 cm and = 0 everywhere else. The total charge present will be (a) 41.05 pC (b) 257.92 pC (c) 82.1 pC (d) 129.6 pC Ans. (c) 5. A dipole having Qd/40 = 100 SI units is located at origin in free space and aligned so that its moment is in the az direction. The electric eld at point (r = 1, = 45 , = 0 ) is (see Section 5.5) (a) 158.11 V/m (b) 194.21 V/m (c) 146.21 V/m (d) 167.37 V/m Ans. (a) 6. A cylindrical surface = 8 cm contains the surface charge density s = 5e20|z| (nC/m2 ). The ux that leaves the surface = 8 cm ;1cm < z<5cm and 30 < < 90 is (a) 270.07 pC (b) 9.45 pC (c) 270.7 nC (d) 9.45nC Ans. (b)

1. What is importance of the analysis of dipoles in electromagnetic theory? Hint: Research the modelling of dielectrics. 2. Can any three charges be placed on a straight line be in equilibrium? That is each charge feel no net force due to the action of the other two charges. Hint: Use Coulombs law. 3. Find the electric eld everywhere due to a spherical shell of inner radius a and outer radius b of uniform charge density v (C/m3 ) Hint: Use Gauss law. 4. Find the electric eld everywhere due to a cylindrical shell of uniform charge density v (C/m3 ) and which has an inner radius a and outer radius b. The cylindrical shell is innite in its length.

176

z r

a R

Sphere

Hint: Use Gauss law. 5. By using the results of disk of charge with surface charge density s = v z, nd the electric eld of a uniformly charged sphere. Hint: See Figure 4.36. ***Chapter complete*** Part2BElectrostatics

177

Let early education be a sort of amusement; you will then be better able to nd out the natural bent Plato

In this chapter the student is introduced to the concept of potential. The potential is very useful concept related to potential and kinetic energy and also the electric eld. The student learns the following 1. To calculate the the work done on a charge required to move it in a region immersed in an electric eld from a point A to B. 2. To calculate the potential dierence between points A and B. 3. The concept of the potential at innity for a point charge. 4. How the scalar potential and the electric eld are connected. 5. The importance of equipotential surfaces. 6. The concept of potential energy. 7. The calculation of the potential when there are a number of point charges. 8. The calculation of the potential due to continuous charge distributions.

Looking work and energy, we know that if a charge Q is immersed in an electric eld, it experiences a force, F F = QE (5.1) If we expect to move this charge by application of an external force, (F) we need to work against this force, and the incremental work done will be dW = F dl = QE dl the negative sign tells us that the force applied is working against eld, if dW is to be positive. Or the work done by the external force to move the charge from rA to rB will be

rB rB

W=

rA

QE dl = Q

rA

E dl

(5.2)

EXAMPLE 5.1 In a region of space the electric eld E = xy2 ax xz2 a y + xyzaz nd the work done to move a charge of 1 nC from (1,1,1) to (0, 1,1)

178

Step 1. First draw a sketch of the path of integration. The work done is W = Q where P is the integration path. E dl

Step 2. We realise from the sketch that path of integration is along a line parallel to the x axis; y and z are xed with y = z = 1.

(0,1,1)

Step 3. Putting y = z = 1 we carry out the the integration. The work done is W = Q E dl = Q = Q (xy2 ax xz2a y + xyzaz) ax dx xdx (putting y = z = 1) x2 2

0 1

= 1 109

EXAMPLE 5.2 For the electric eld of Example 5.1 nd the work done to move a charge of 1 nC from (3,5,6) to (0, 0,0) along a straight line. Step 1. We rst draw a sketch of the straight line, on the coordinate system etc. Step 2. We next nd the parametric equation of a straight line joining the two points. In accordance with 2.4.3 the vector joining r0 = (3, 5, 6) to r1 = (0, 0, 0) is R01 = (3, 5, 6). The parametric equation of the straight line joining these two points is r = r0 + tR01 0 t 1 x = 3(1 t) y = 5(1 t) z = 6(1 t) Step 3. The work done is

r1

W = Q = Q = Q = Q

r0 r1 r0 r1

E dl (xy2 ax xz2a y + xyzaz) (dxax + dya y + dzaz ) xy2 dx + xz2dy + xyzdz 15(1 t)3(3dt) + 18(1 t)3(5dt) + 90(1 t)3(6dt) (1 t)3dt = 168.75 (nJ)

r0 t=1 t=0

t=1

= 675Q

t=0

179

initial point

Line of integration

final point Q

The concept of the electric potential in the electrostatic eld springs directly from the concept of work. Using Equation 5.2 we have VAB = W = Q

rB rA

E dl

remember that Q is the charged which is moved in the electric eld. EXAMPLE 5.3 For a point charge Q located at the origin nd the potential between two points rA = (rA , /2, 0) (initial point) and rB = (rB , /2, 0) (nal point) in spherical coordinates. Step 1. Draw a sketch pertaining to the problem. We notice from the looking at the problem that we are integrating only in the r-direction. Step 2. We next look at the The electric eld for a charge located at the origin is given by Q ar E= 40 r2 dl = drar since and are constants along the path of integration (d = d = 0). So

rB rB

VAB =

rA

E dl =

rA

Q Q dr = 2 40 r 40 r

rB

=

rA

Q 1 1 40 rB rA

Generalising these results we look at the potential dierence between two points rA (initial point) and rB (nal point) as shown in Figure 5.1. Let the curve be parametrised as (r(t), (t), (t)) with rA = (rA , A , A ) and rB = (rB , B , B ). Then

B rB

VAB =

E dl =

rA

Q ar (drar + da + da ) 40 r2

rB

= or VAB =

rA

Q 1 1 Qdr = 2 40 rB rA 40 r

(5.3)

Q 1 1 40 rB rA

180

which is a statement of the fact that in an electrostatic eld of a point charge, the potential dierence is independent of the path of integration. Why? Let us look at Equation 5.3. We see that in the integration, the path is independent of the parameter t and depends only on the r coordinate: the path may be anything. Furthermore if rA is a point at innity, rA and rB r then we can dene an absolute potential Q Vr = (5.4) 40 r with the point at innity as having zero potential: V = 0 And VAB = VB VA = Vn Vinit (5.6) Another important consequence of this property is that if we go from A to B along one path and back from B to A along another path, then VAA = VAB + VBA = (VB VA ) + (VA VB) = 0 in terms of the line integral of the electric eld this is E dl = 0 (L is any closed path) (5.8) (5.7) (5.5)

Enclosed surface

or E = 0 (5.10)

Referring back to the section concerning the curl, Section 3.3.3, one knows that when the curl of a vector eld is zero like the previous equation then it must be the gradient of a scalar eld. Having said this we look at and compare the right hand sides of the two equations: dV = E dl = Ex dx E ydy Ezdz and dV = V V V dx + dy + dz x y z (5.11)

(5.12)

181

from which we get V x V Ey = y V Ez = x Ex = or E = V (5.14)

(5.13)

The electric eld, E, is the gradient of the potential. In this book, the variable V is used both for the volume as well as the electric potential, but this should not cause confusion. What are the units of V ? Units of V(r) = Units of {E l} = V/m m = V which is volts. EXAMPLE 5.4 The electric potential in a region of space is V = kx2 yz. Find the electric eld and charge in that region. The region is air. Step 1. The electric eld E is given by E = V = V V V ax ay az = 2kxyzax kx2za y kx2 yaz x y z (V/m)

U U

Step2. From the E eld we can obtain the D eld: D = 0 E = k0 2xyzax + x2 za y + x2 yaz Step 3. and v = D = Dx D y Dz + + = 2k0 yz x y z (C) (C/m2 )

EXAMPLE 5.5 The charge density (x) L x L is shown in Figure 5.2. Find the electric eld and potential as a function of x. The medium has a dielectric constant r . Step 1. First: D = and so Dx D y Dz + + = x y z dDx = (x) dx

182

x

Dx (x) =

L

(x)dx = 0

x x1 x1 L x

(0)dx +

x1

x2 m x2 2 x x2 m 2 xdx = x 1 2 x2 1 1 m m 2 x dx = x2 + x x2 2 2x2

Dx = 0

From the ux density we can get the electric eld 0 m x 2 2 2 x2 x2 ax D 1 1 E= = x1 0 r 0 r m x2 x2 a 2 x 2x2 0 Step 3. Since there is only one component Ex and dV = Ex dx we can obtain V(x) from Ex by

x

L x x1 0 x x2 x1 x 0

x2 x L

V(x) =

Ex dx

L

183

The reader must note that the absolute potential is unknown to the extent of an additive constant, V0 , but the functional form of the potential is correct. For the point charge, V0 is taken to be zero for r equal to innity. However it should be noted that for dierent congurations, V0 will be dierent. Generally, it is taken to be zero. Proceeding now along a dierent tack, we would like to nd the potential for a point charge placed at an arbitrary point r . Referring to Figure 5.3, if the potential at the point (X, Y, Z) due to a point charge placed at the origin of the (X, Y, Z) coordinate system is V(R) = V(X, Y, Z) Q = 40 R Q = 40 X2 + Y2 + Z2

(5.15)

then what is the potential V(r) due a point charge placed at r = (x , y , z )? Referring to the gure once again, we know that in the (x, y, z) coordinate system, the same potential, V(R) is V(r) = Q Q = = 40 R 40 |r r | Q 40 (x x )2 + y y 2 + (z z )2

1/2

since R = r r . if we compute V(= E) of the above expression V = (V/x) ax + V/y a y + (V/z) az = 40 (x x)2 + y y 2 + (z z )2 40 |r r |3 184 Q (r r ) Q (x x) ax + x y a y + (x z) az

3/2

(5.16) (5.17)

which is the electric eld at an observation point due to a charge placed at (x , y , z ). So we can conclude that the potential V(r) due a point charge placed at r = (x , y , z ) is indeed V(r) = Q 40 r r 0 (5.18)

Equipotential surfaces are those surfaces on which the potential is constant. What is so important about an equipotential surface is that the eld is normal to the surface at all points on it. This was explained in Section 3.3. If we examine the eld for the three cases considered in the previous chapter the point charge, the innite line charge and the innite plane charge, we can evaluate the equipotential surfaces for these three cases to give us an insight on equipotential surfaces in general. For the point charge at (x , y , z ) the potential at (x, y, z) is 0 0 0 V(r) = 40 if we equate V to a constant Q 40 then x x 0

2 2 2 1/2

Q x x 0

2

y y 0

z z 0

2 1/2

(5.19)

y y 0

2

= k1

(5.20)

z z 0

2 1/2

x x + y y + z z 0 0 0

Q =k 40 k1

(5.21)

2 2 2

= k2

(5.22)

is the equation of a sphere with radius k and centre (x , y , z ). Since the electric 0 0 0 eld is normal to this surface at all points it is radially streaming out of r . 0 Proceeding to examine the eld of a line charge, Figure 4.33 on page 167, the electric eld has only one component E , , z = l 20 (5.23)

185

Equating this component to the a component of V in cylindrical coordinates l V = 20 dV = d or V() = d l 20 l 1 = ln 20

(5.24)

(5.25)

This is the potential function. Because only dierences in potential are important, the function is zero at = 1. This is point with which compare all other values V are compared to. The equipotential surfaces are computed by l 1 ln = k1 20 20 1 ln = k1 l = k2 = ek2 =k (5.27) where k1 , k2 and k are constants. The previous equations are a family of cylinders cylinders whose axes coincide with the z-axis. Their cross-sections are circles with centres (0, 0, z). On any one such surface, the electric eld is normal to the surface and equal at all points on it. The electric eld streams out radially away from the z axis. To compute the potential eld for the case of the innite plane charge, Figure 4.35, the electric eld is given by s az when z > 0 (5.28) E = 20s a when z < 0 2 z

0

(5.26)

for z > 0

or

s dV = dz 20

(5.29)

V=

s z 20

(5.30)

the V = constant surfaces are planes z = k surfaces, which are parallel to the x-y plane. (z = 0 is the x-y plane).

186

-10 -5 0 5 5 -5 10 -10 0 10

EXAMPLE 5.6 For the potential function V = x2 y2 nd the equipotential surfaces Step 1. The surface V = x2 y2 is shown in Figure 5.4.

Step 2. Equipotential surfaces are those surfaces when V = constant On the base of the gure are shown curves x2 y2 = k where k is a constant. Each of these contours may be imagined to be a surface which extends from z = to z =

Next we investigate the question: What is the physical meaning of the line integral of the electric eld? We know that the line integral of the force F dl is the work done. So let us nd the work done on a test charge Qt when moving it in an electric eld due to the point charge Q placed at the origin. First let the test charge be far away at r = . Let us further suppose that both Q and Qt are both positive charges, and let us move Qt towards Q along a straight line. See Figure 5.5. Since both Q and Qt are both of the same polarity we expect a resistance to the motion when we move Qt since the charges repel each other. The force on the

187

z

4 3 2 1

1 2 3 4

x

Figure 5.5.: Work done when moving Qt towards Q

test charge is F = Qt Eprinciple charge Q = Qt r 40 r2 The work done to move the test charge from r = to r = r0 is given by

r0 r= r0 r0

F dl =

r=

Qt E dl = Qt

r=

E dl = Qt V(r0 )

(5.31)

V(r0 ) is the potential due to the charge Q at the point r = r0 . In fact it is Qt (V(r0 ) V()) but since V() = 0 the work done is equal to Qt V(r0 ). We have done work on the charge, but it seems to have made no dierence to it! Where has the work done gone? Notice that Qt and Q are of the same polarity so obviously to maintain its position we are exerting a force. If we release Qt it will move towards r = with increasing velocity. That is it: it will gain in kinetic energy and so it must have had gained potential energy. We can write a general formula Gain in Potential Energy=Qt [V(Final Position) V(Initial Position)] Also we may write (from what has been said earlier) that

B

(5.32)

E dl

188

EXAMPLE 5.7 An electron has an initial velocity of v = 2 107ax m/sec when it is at the coordinate point [1, 0, 0] m. The electron is moving in the electric eld of a charge q = 3 106 C placed at the origin of a coordinate system. Describe the motion of the electron. Let us rst draw a diagram of the conguration

z

4 3 2 1

q e

1 2 3 4

Figure 5.6.: [1, 0, 0] m moving with a velocity of 6 107 ax m/sec in the presence of q = 3 106 C at [0, 0, 0]

Step 1. The potential at [1,0,0] is1 Vin = q/(40 rin ) = (3 106)/(40 ) = 2.696 103V volts. In the previous equation rin is the initial position of the charge and Vin is its initial potential. Step 2. The initial potential energy Pin of the electron is Pin = eVin = 4.31411 1015J (note the negative potential energy) where e = 1.6 1019C the charge of an electron. As the electron moves it slows down due to attraction to the charge at the origin. Step 3. Its initial kinetic energy is 1 Kin = me v2 in 2 = 1.6398 1015J

1

The subscript in will be used for initial quantities while f in will be used for nal quantities.

189

Step 4. The initial total energy is ET = Pin + Kin = 2.6743 1015 J As the electron moves it comes to rest. Step 5. Hence ETin = ET f in = ET P f in = ET = 2.6743 1015J But P f in = eq 40 r f in

which gives the nal position of the electron r f in = 1.6131 m After this the electron reverses motion and moves toward the charge at the origin.

EXAMPLE 5.8 In the previous example beyond which velocity does the electron escape from the clutches of the 3 106C charge? Step 1. The initial potential energy of the electron is Pin = eVin = 4.31411 1015J as earlier. Step 2. When the electron reaches r = its nal potential energy becomes zero and its kinetic energy also becomes zero. Which means that its total energy must be zero initially itself. Step 3. We have to solve the equation

ET = Kin + Pin = K f in + P f in = 0 or or setting ET = 0 we have 1 me v2 + (4.31411 1015) = 0 in 2 solving for vin . vin = 9.713231 107m/sec

190

Figure 5.7.: Electron microscope picture of a spider (photo taken taken from a website of University of Minnesota, http://umn.edu, 2010)

Step 4. Interpretation: this result can be stated as follows: if the electron has an initial velocity greater than 9.713231 107m/sec then the electron does not return toward the charge at the origin. Did you know?/Application The electron microscope (as the name implies) uses a beam of electrons to illuminate the specimen and create a magnied image of it. From quantum theory, electrons are dual in nature being waves or particles; the wavelength of an electron may be calculated according to, = h/(mev) where h is the Planks constant and v is the velocity of the electrona . Obviously the greater the velocity of an electron, the smaller its wavelength, so very small wavelengths through very high velocities are acheived. By these means wavelengths about 100,000 times shorter than light, are possible which give magnications of up to about 2,000,000x. (Microscopes using light are limited to about 2000x magnication.) Electron microscopes use electrostatic lenses to focus the electron beam. Typically, electrons are emitted by an electron gun, (something like an electron gun in a TV tube) which are accelerated through a potential dierence of 40 to 400 keV and transmitted through the specimen. The electrons which emerge are used to form an image. Figure 5.7 shows one such image.

a De

Broglie equation

From the above example we realise that potential energy of a charge may be negative! Hence we should rather be concerned about the dierences in potential and potential energy rather than their absolute values. The potentials and potential dierence between two points in space, A

191

A x0 , y0 , z0 to B B x1 , y1 , z1 is equal to

B

V =

A B

(E) dl V dl (5.33)

=

A

Thus for a charge of 2 pC placed at the origin, the potential dierence between the points A [1, 1, 1] and B [2, 2, 2] V(A) = 0.0179755V V(B) = 0.0051891V V(A) > V(B) The electric eld goes from A to B. If a positive charge Qt is placed at A and no restrictions are placed on it then it will feel a force from A to B, and tend to go from A to B, losing potential energy = Qt (0.012786)J. On the other hand, a negative charge placed at B will feel a force towards A. If the charge at the origin is a -2 pC charge then V(A) = 0.0179755V V(B) = 0.0051891V V(B) > V(A) The electric eld goes from B to A. In this case if a positive charge is placed at B and no restrictions are placed on it then it will feel a force from B to A. On the other hand, a negative charge placed at A will feel a force towards B. We come to two important conclusions 1. The electric eld is directed from a higher potential to a lower potential. And 2. Positive charges feel forces in the direction of the electric eld, and when they move they go from a higher potential towards a lower potential, losing potential energy. Negative charges feel forces in a direction against the electric eld. They move from a lower potential to a higher potential again losing potential energy.

192

The potential at r = (x, y, z) due to a single charge Q1 placed at the point r = 1 (x , y , z ) is given by 1 1 1 V(r) = = 40 Q1 40 r r 1

2

Q1

2 2

(5.34)

x x + y y + z z 1 1 1

....

z o y x

Figure 5.8.: Figure to calculate the potential for a system of point charges

which we already know from the early part of this chapter, but how do we calculate the potential due to many point charges? As usual we use superposition. For example if there are two point charges Q1 and Q2 placed at r = (x , y , z ) 1 1 1 1 and r = (x , y , z ) then the potential at r = (x, y, z) is 2 2 2 2 V(r) = V1 (r) + V2(r)

2

=

i=1

Qi 40 x x + y y + z z i i i

2 2 2

Similarly if charges Q1 , Q2 , . . . , QN are placed at r , r , . . . , r respectively, then N 1 2 each of these charges produces a potential Vi (r) at the eld point r. The total potential, therefore, due to all these charges placed at all these positions is

i=N

V(r) =

i=1

Vi (r)

or

i=N

V(r) =

i=1 4

Qi 2 2 2 + y y + z z x xi i i 193

(5.35)

And from which expression we can evaluate the electric eld by E = V (5.36)

In the previous equation, the nabla operator operates only on the r variable. That is, in rectangular coordinates E = V = =

ax + a y + az V x , y , z , x, y, z x y z V V V ax + ay + az x y z (5.37)

In the previous equation dierentiation must be carried out only on the x, y, z trio, and not on the x , y , z set of variables. The real value of the potential expression is here we have to perform dierentiations rather than integrations to compute the electric eld. Let us revisit the dipole this time aligned along the z-axis as shown in the gure.

z Distant point

r1

Q d Q

r r2

y

d cos

x

Figure 5.9.: A dipole aligned along the z axis

EXAMPLE 5.9 Find the far eld for the case of a dipole as shown in Figure 5.9. Step 1. Using the formula of Section 5.5, the potential due to the two identical charges is Q V(r) = 4 0 1 z

d 2 2 2 2 + y +x

z+

d 2

1 2 2 + x2 +y

(5.38)

194

Step 2. Going over to spherical coordinates x = r sin cos ; y = r sin sin ; z = r cos and simplifying Q 1 1 20 2 4 d r cos + d2 2 + 4 d r cos + d2 4r 4r 1 1 Q 2 4 d r cos + d2 2 + 4 d r cos + d2 20 4r 4r Q 1 1 20 4 r2 4 d r cos 4 r2 + 4 d r cos Q 1 1 = 40 r 1 ( d/ r ) cos 1 + ( d/ r ) cos we get V= 1 1x = 1 x + ... 2 p = Qdaz (5.39)

V(r) =

(5.40)

Step 4. Using

p ar Qd cos = 40 r2 40 r2

Because of the absence of there is a symmetry of the E eld about the z axis. E 1 V Qd sin = r 40 r3 E = 0 (5.42) (5.43)

EXERCISE 5.1 Show all the steps related to the Equation 5.40, going from rectangular to spherical coordinates. This will hone your math skills.

In the Section 4.6 on page 145 we obtained the electric eld due any continuous charge distribution, we now obtain the scalar electric potential due any contin-

195

uous charge distribution. To do so we nd the innitesimal potential dV at a point in space, r due an elemental charge placed at r as shown in the gure is

Miniscule Potential dV(r, r )

r r r

z Miniscule Charge, dQ

r

y x

dV(r, r ) =

dQ 40 (xx )2 +

( yy )2 +(zz )2

dV(r, r ) = 40

dQ(r ) (x x)2 + y y 2 + (z z )2

(5.44)

where r is the position vector of the eld point and r is the position vector of the charge dQ. Notice that this is the same formula as that for a point charge, except that the point charge Q has been replaced by dQ. Now to nd the potential due to any charge distribution we integrate the above equation over the region where the charge is V(r) =

R

dV(r) dQ(r ) 40 (x x )2 + y y 2 + (z z )2

(5.45) (5.46)

=

R

Where R is the region occupied by the charge. Now dQ will be dierent for dierent charge distributions v (r )dV for a volume charge distribution (5.47) dQ = s (r )dS for a surface charge distribution l (r )dL for a linear charge distribution Therefore V(r) =

V S L v (r )dV 40 (xx )2 +( yy ) +(zz )2 s (r )dS (xx )2 +( yy ) +(zz )2 l (r )dL (xx )2 +( yy ) +(zz )2

2 2 2

40 40

196

V(r) =

2

r r dV

r

y x

For a volume charge distribution the accompanying gure, (Figure 5.11) shows the various regions; and the integrations are performed on V . We now apply our newly derived formulae to obtain the potential and the electric eld to the case of a pair of line charges shown in Figure 5.12. The gure shows a couple of charged lines piercing the x-y plane at [0,-d/2,0] and [0,-d/2,0]. The lines are charged with a with a constant line charge of l (z ) and l (z ) C/m. From symmetry considerations, we compute the potential only on the x-y plane, as the elds on any other plane would be the same. With these remarks in mind r = [x, y, 0] (5.49) and r = [0, d/2, z] r = [0, d/2, z] l (r )dL

L

(5.50) (5.51)

V(r) =

The left integral is over the positively charged line while the right integral is over the negatively charged line. The integral may be written in a slightly

40 (x x)2 + y y 2 + (z z )2 l dz + = 40 x2 + y d/2 2 + z2 40

l 2 x2 + y + d/2 + z2 dz

(5.52)

197

z To Infinity

dz

r

[0,d/2,0] [0,d/2,0] y

r r l C/m

Chrged Lines

l C/m

d

Figure 5.12.: Calculation of the potential for a pair of parallel charged lines with charge l and l .

z0

Where

V(r) = lim + z0 2 2 2 + y d/2 + z2 2 + y + d/2 + z2 40 x 0 x z =z0 l z z sinh1 sinh1 lim = 2 2 4 e0 z0 d d 2 2 y 2 +x y 2 +x z =z0 l 2z0 2z0 sinh1 sinh1 = lim z0 2 e0 4 y2 4 d y + 4 x2 + d2 4 y2 + 4 d y + 4 x2 + d2 l dz l dz sinh1 (p) = log p2 + 1 + p (5.53)

The reader is referred to the Appendix where a table of such integral has been provided. After taking the limit z0 the potential in the x-y plane becomes l log V(x, y) = 2 0 y + d/2 2 + x2 2 2 y d/2 + x (5.54)

198

If we wish to compute the equipotential surfaces for this example l log V(x, y) = 2 0 y + d/2 2 + x2 y d/2 2 + x2 y + d/2 =V 1 2 y d/2 + x

2 2

+ x2

(5.55)

a constant. Or

= exp

V1 2 0 =k l

(5.56)

2 2 1 + k2 d 2 k2 1 1 + k2 yd + x2 + y2 + 2 1 k2 d = 4 1 k2 + 2 1 k2 d 1 k2 2 2 2 d 1 + k2 = d k x2 + y + 2 2 1 k2 1 k2 dk = rk 1 k2

(5.57)

and are valid for 0 < k < 1. If we substitute 1/k for k < 1 then we arrive a family of complementary circles with the same radii r= but centres dk = r1/k 1 k2 (5.58)

d 1 + k2 0, yk = 2 1 k2

These plots are shown here. We can compute the electric eld for this case by E = V

d 1 + k2 0, y1/k = + 2 1 k2

199

k=0.6

k=0.5 k=0.4

2 2 1 0 1

1 0 -1 -2 -1 0 1

Figure 5.13.: Lines of constant electric potential for a pair of innite line charges

Though the expressions are quite complex, but the eld plot is particularly simple, as shown in the gure. The electric eld streamlines are circles which cut the equipotential lines at right angles. Recall that the equipotential lines are also circles. In fact in any electrostatic case, the electric eld lines and the equipotential lines always cut each other at right angles.

Figure 5.14.: Electric eld plot of the electric eld for two parallel innite line charges

EXAMPLE 5.10 Find the potential due to an innite line charge with charge density l along the z-axis (z = ) Step 1. The electric eld due to such a conguration is given by Equation 4.40. E= l a 2 0

200

Step 2. Since

2 2

V1 to 2 =

E dl =

l 1 l d = ln 2 0 2 0 2

EXAMPLE 5.11 An electron at rest is released from the position of 1 m along the axis of a ring of diameter 1 m which is charged with a linear charge density of l = 1 pC/m as shown in the gure. Describe the motion of the electron.

z Electron at [1,0,0]

r r V(0, 0, z) =

2 0

l 0 z2 +2 0

l = 1 pC/m

x

Figure 5.15.: Charged ring and electron.

Step 1. The motion of the electron is best described by computing the potential function of the charged ring. The potential V(r) is given by V (r) = =

Ring

l dL Ring 40 |r r | 4 0

l dL y y 2 + (x x)2

(z z )2 +

(5.59)

Where l is the charge density of 1 pC/m. Both r and r are shown clearly in the gure. Step 2. We want to nd the potential on the z-axis only. So r = [z, 0, 0] and for the ring (5.60)

201

r = [x , y , 0]

(5.61)

where x = 0 cos and y = 0 sin where , the parameter, is the angle and 0 = 0.5 m. Step 3. Now since the geometry of the problem is cylindrical, namely, the charge lies on a circle with radius 0 (= 0.5 m) and the electron is on the z axis, which is the axis of the ring, we go over to the cylindrical coordinate system. Using x = 0 cos and y = 0 sin

2

V (r) =

l 0 d 4 0 z2 + y 0 sin

2

+ x 0

cos

(5.62)

Step 4. Since the potential function needs to be calculated only on the z axis, x and y are both equated to zero

2

V (z) =

2 2

0 2

=

0

l 0 d 4 0

= 2 0

l 0 z2 + 2 0

Step 5. Now we come to the electron. The initial kinetic energy of the electron is zero, and its initial potential energy and total energy is l 0 e 2 0 z2 + 2 0

z=1 m

Pin =

(5.64)

(5.65)

Interpretation. The electric eld has only one component, namely, the z component, as we have shown in Example 4.14. And therefore the electron feels a strong attractive force toward the centre of the ring. As it moves its kinetic energy increases while its potential energy decreases keeping its total energy

202

constant, which is equal to the initial value of total energy ET = Kinetic Energy + Potential Energy = K+P = Kin + Pin = 4.0458 1021 J (5.66) (5.67)

Step 5. When the electron is at a distance z from the origin its potential energy is l 0 e (5.68) P(z) = 2 0 z2 + 2 0 and its kinetic energy is 1 K = me v2 2 l 0 e 1 + me v2 = 4.0458 1021 (= ET ) K+P = 2 2 0 z2 + 2 0

(5.69) (5.70)

1 4

az

(5.71)

m2 2 z2 + 2 e 0 0

Important note: To check this result, we substitute the values of ET , 0 , e, me , 0 , and also z = 1, which gives a value of zero. If we want to nd the velocity of the electron when it reaches the centre of the ring then we substitute z = 0 in the above equation. v(z = 0) = 1.047 105az m/s (5.72)

The work done by the external force to move the charge from rA to rB will be r

B

W = Q

rA

E dl

The concept of the electric potential in the electrostatic eld springs directly from the concept of work. The potential dierence between points A and B is rB W VAB = = E dl Q rA The potential dierence between two points, rA and rB , in the eld created by a point charge is 1 Q 1 VAB = 40 rB rA

203

The absolute potential for a point r in the eld created by a point charge is Q V(r) = 40 r and where V() = 0 The potential dierence between two points A and B is VAB = VB VA = Vn Vinit The line integral of the electric eld on a closed path is zero. E dl = 0 (L is any closed path)

Maxwells equation for electrostatic elds is E = 0 The relation between the electric eld E and the potential V is E = V The potential at r for a point charge Q located at the point r is V(r) = Q = 40 |r r | Q 40 (x x)2 + y y 2 + (z z )2

1/2

An equipotential surface is one where the potential is constant. The electric eld is always perpendicular to an equipotential surface. The work done to move a charge from r = to r = r0 is given by

r0 r=

F dl = QV(r0 )

(5.73)

Gain in potential energy for a charge Q is Gain in Potential Energy=Q [V(Final Position) V(Initial Position)] The potential at a point r due to N charges Qi at points r , i = 1, ..., N in i rectangular coordinates is i=N Qi V(r) = 2 2 2 + y y + z z i=1 40 x xi i i

Consider a dipole consisting of charges (Q, Q) with a separation d, oriented along the z-axis, and located at the origin. Then the potential V(r)

204

is V= Qd cos 40 r2

The electric eld for such a dipole is Qd cos 20 r2 Qd sin E = 40 r2 E = 0 Er = The potential at the point r due to continuous distribution of charges is V(r) =

R

dQ(r ) 40 |r r |

where v (r )dV dQ = s (r )dS l (r )dL for a volume charge distribution for a surface charge distribution for a linear charge distribution

The potential at any point (x, y, z) for two innitely long lines with charge distribution l and l is l log V(x, y) = 2 0 y + d/2 2 + x2 2 2 y d/2 + x

Chapter Summary

In this chapter, The student learns to calculate the work done to move a charge from point A to B. The concept of potential is introduced: what is the potential due to a point charge at a point r. We explain how the electric eld and potential are related: E = V The student learns Maxwells equation: E = 0 for electrostatic elds. The concept of equipotential surfaces is introduced. We explain how to calculate the gain or loss of potential energy of a charge: Q V f in Vin .

Using the principle of superposition, we learnt to calculate the potential due to many point charges. We applied the principle of superposition to the dipole and calculated the potential function and electric eld at distant points.

205

We learnt how to calculate the potential for continuous charge distributions.

Review Questions

1. Explain how work has to be done to move a charge from point A to B. If the work is positive what does it mean? If it is negative what is the implication? 2. Explain why electric eld lines are perpendicular to equipotential surfaces? 3. Left alone in a static electric eld with no external force applied, an electron gains momentum. What happens to is potential energy? 4. If we calculated the line integral of a static electric eld along a closed contour what should be the answer? Justify. 5. If we calculate the potential at a point in the presence of a number of point sources what is the law we use and why? 6. Derive an expression for the potential at any point due to a point charge.

Problems

1. Show that the electric eld for a line charge and sheet charge satises the equation E = 0 2. The electric eld near the the origin of a Cartesian coordinate system is E = (yz + 2xy2z2 )ax (xz + 2yx2z2 )a y (yx + 2zy2x2 )az . Find the work done on a 10 nC charge to move it by 6 mm from (1,1,1) in the x, y and z directions respectively. Ans. 108 0.018 (W) in all three cases. 3. For the E eld of Problem 2 nd the work done to move a 10 nC charge to move it by 6 mm from (1,1,1) in the ax + a y + a y direction. Ans. 108 0.031 (W) 4. For the E eld of Problem 2 nd the work done to move a 10 nC charge to move it from (0,0,0) to (1,1,1) along a straight line. Ans. 2 108 (W) 5. For the E eld of Problem 2 nd the work done to move a 10 nC charge to move it from (0,0,0) to (1,1,1) along the following zig-zag path (0,0,0) to (1,0,0) to (1,1,0) to (1,1,1). From the answer what do you infer? Ans. 2 108. Probably the E eld is conservative. E = 0. 6. In free space V = x2 y2 (z + 3). Find E at (3, 4, -6). Ans. 288ax + 864a y 144az

7. Let V = xy2 z , calculate the energy expended in transferring a 2C point charge from (1, 1, 2) to (2, 1, 3). Ans. 16 J

206

8. Determine the electric eld due to the following potentials (a) V = x + y2 + 2z2 (b) V = (x2 + y2 + z2 )1/2 y Ans. (a) [1, 2 y, 4 z] (b) [ 2 x2 2 , 2 2 2 , 2 z 2 2 ]

z +y +x z +y +x z +y +x

11. Find the work done in carrying a 5C charge from P(1, 2, 2)to R(0, 0, 0) in an electric eld E = 2xzax x2 az V/m

10. Determine the electric eld due to the following potential V = er sin cos cos() er cos() sin() er , ] Ans. [cos er sin () , r r

9. Determine the electric eld due to the following potential V = 2 zsin Ans. [2 sin z, cos z, sin 2 ]

Ans. 5 W

12. The electric eld in cylindrical coordinates is E = sin za cos za sin az determine the work done in moving a 4 nC charge from (a) A(1, 0, 0) to B(4, 0, 0) (b) B(4, 0, 0) to C(4, 30, 0) (c) C(4, 30, 0) to C(4, 30 , 2) (d) A to D Ans. (a) 0. (b) 0. (c) 16 nJ (d) 16 nJ 13. In free space, V = x2 yz V. Find (a) E at (3, 4, 6) (b) the charge within the cube 0 < x, y, z < 1 Ans. (a) 72ax+54a y-36az (b) 0 /2 C

17. For a disk of charge with constant surface charge s , with inner radius b and outer radius a nd the potential along the axis of the disk. z2 + a2 z2 + b2 Ans. s /20 18. For a disk of charge with surface charge s and radius a nd the potential along the axis of the disk. Ans. s /20 a tan1 a z z

14. Three point charges Q1 = 1 mC, Q2 = 2 mC , Q3 = 3 mC are respectively located at (0, 1, 4) (2, 5, 2) and (3, 6, 6). (a) Find the potential Vp at P(1, 1, 1) (b) Calculate potential dierence VPQ if Q is (1, 2, 3) Ans. (a) 1.45106 V (b) 1.18106 V 15. To verify that E = [2 x y (z + 3), x2 (z + 3), x2 y] is truly an electrostatic eld, show that (a) E =0 (b) L E dl = 0 where L is the closed loop of (0,0,2) (3,0,2) (3,3,2) (0,3,2). 16. For a disk of charge with constant surface charge s and radius a nd the potential along the axis of the disk. z2 + a2 z Ans. s /20

207

1. For the case of two concentric spheres, the inner sphere has a radius rin with charge Q on it, while the outer sphere has radius rout with charge Q. An electron leaves the outer sphere and rushes toward the inner sphere. Find KE of the electron when it reaches the inner sphere. Ans. The potential of the outer sphere is Vout = Q 40 rout

similarly the potential of the inner sphere is Vin = the dierence in PE is Q 40 rin

2. A potential eld exists in a region where = f (x). Find the value of 2 V if v = 0 throughout the region. Ans. D = E and D = v = 0 D = ( f (x)V) = f V f V

So,

2 V = 1/ f (x) (d f /dx)(V/x)

3. If V = (10/r2) sin V in free space then what is v at point P (r = 2, = 30 = 0 ) ? Ans: Use 2 V = v /0 ; E = V 20 sin () 10 cos () , , 0] =[ r3 r3 (0 E) = v 10 = 0 4 r sin () and at the point under consideration one can evaluate the expression.

208

O

Figure 5.16.: Calculation of potential for Problem 4

4. Derive an expression for the potential at any point due to a point charge. Solution. Let the curve be parametrised as (r(t), (t), (t)) with rA = (rA , A , A ) and rB = (rB , B , B ). Then dl = drar + da + da

B rB

VAB =

E dl =

rA

Q ar (drar + da + da) 40 r2

rB

= or VAB =

rA

Q Qdr 1 1 = 40 r2 40 rB rA

Q 1 1 40 rB rA

1. Which of the following sentences is incorrect? (a) The line of force or ux lines are always normal to equipotential surfaces (b) The conductivity of metals generally decrease with temperature (c) Work done in moving a charge along a closed path in electrostatic elds is zero (d) None of the above Ans. (d) 2. In an electrostatic eld, which of the following statements are correct. Work done in moving a charge Q from point A to B is (a) dependant on the path chosen (b) dependant on the initial and nal potentials (c) equal to Q A E dl (d) equal to Q A V dl Ans. (b) and (d) 3. When moving a small charge q = 1 pC in the eld of a charge Q = 1 C located at the origin (in air) along the path (x = 1, y = 1, z = 1) to (1,2,1) to (1,2,2) and back to (1,1,1) (all in meters) the work done is: (a) 10 pJ (b) 10 nJ (c) 10 J (d) 0 J (e) None of the above Ans. (d) 4. For a negative point charge Q (Q > 0) the point which is taken as V = 0 is (a) Not dened (b) at innity (c) at 1 m from the charge (d) none of the

B B

209

above Ans. (b) 5. For a negative point charge Q (Q > 0) at a distance of 1 m from the charge (a) V is negative (b) V = Q/40 (c) V is positive (d) V is zero Ans. (a) and (b) 6. For two equal point charges Q, the potential is V everywhere, and (a) The potential at the mid-point between the two charges may be taken as zero (b) V = 0 (c) the potential is the sum of the potentials of each of the charges (d) none of the above Ans. (a), (b) and (c) 7. Work done in moving a charge Q in a straight line in an electrostatic eld which is varying with coordinates is (a) Force distance (b) F dl (c) F dl (d) Q E dl Ans. (c) and (d) 8. The electric lines are (a) directed from a higher potential to a lower potential (b) directed from a lower potential to a higher potential (c) perpendicular to equipotential surfaces (d) None of the above Ans. (a) and (c) 9. When a free electron is placed in an electrostatic eld it moves (a) from a higher potential to a lower potential (b) from a lower potential to a higher potential (c) perpendicular to equipotential surfaces (d) None of the above Ans. (b) and (c) 10. Equipotential surfaces (a) may intersect at right angles (b) never intersect (c) may intersect at any angle (d) V is perpendicular to them Ans. (b) and (d)

1. If there is a potential eld V = ax + by + cz where a, b, c are constants, then show that the electric is constant, and nd the direction in which the electric eld points. Hint. Use E = V. 2. Can electric eld lines of electrostatic elds be closed curves? Research this topic. Hint. Use E = V and the concept of equipotential surfaces. 3. For a single charge E = 0. From this concept and the principle of superposition show that E applies to elds generated by all kinds of charges. Hint. Consider Eat r = Eat r = 1 40 1 40 |r r |3 dQat r |r r | |r r |3 V

V

dQat r |r r |

4. Find the potential along the axis of a circular disk of charge of surface charge s and radius a. V (r) = =

Disk

s dS Disk 40 |r r | 4 0

s dS

s d d 4 0 4 0 s d d s 2 2 z +a z 20

(z z )2 + y y 2 + (x x)2 z2 + y 2 + x 2 z2 + 2

= =

211

No matter how good teaching may be, each student must take the responsibility for his own education. John Carolus S.J.

In this chapter we will discuss how the electric eld interacts with materials. After reading the chapter the student will understand various concepts and also learns how elds interact with materials. The student learns the following 1. The concept of current and current density. 2. How charge and current are related. 3. The concept of velocity of mobile charge carriers and its connection with the current density. 4. The concept of mobile charge density. 5. The concept of conservation of charge and the continuity equation. 6. The criteria under materials become conductors, semiconductors and dielectrics. 7. The relationship between the electric eld and the current density. 8. How to calculate the capacitance. 9. How the electric elds behave at the boundary of two media: he learns how elds behave. 10. What is the energy stored in the electrostatic eld.

Broadly there are three types of media which interact with electric elds giving rise to dierent types of results. These are conductors, semiconductors and dielectrics. In conductors and semiconductors electric elds give rise to the ow of charge or electric currents. In conductors the charges are predominantly electrons, while in semiconductors charge ow is due to both electrons and holes. In dielectrics, the charges are not free to move and the electric eld polarises the molecules comprising the nucleus and valence electrons. First we will consider conductors which support currents, and, as a prelude to charge ow we need to understand come key concepts.

212

Charge ow of any type, whether of the positive or negative variety gives rise to currents. Current is dened as the total charge ow per second through the cross-section of a conductor. To be specic, consider a straight conductor of cross-sectional area A through which a charge dQ passes in time dt. Then the current I is given by I = dQ/dt (6.1)

Area A

I(t)

Straight Wire

The equivalent straight wire carrying a current is also shown. EXAMPLE 6.1 A wire carries a constant current of 1 A. Find the total charge passing a plane perpendicular to the cross-section of the wire in 1 s. Step 1. The charge crossing any cross-section of the wire in time dt is given by the current dQ = Idt. Step 2. So if Q is the total charge crossing the cross-section in 1 s is

t0 +1 t0 +1

Q=

t0

Idt = I

t0

dt = 1 (C)

EXAMPLE 6.2 A wire carries a current of I = 1 exp{3t} (A). Find the total charge passing a plane perpendicular to the cross-section of the wire from t = 0 to t = 3 s. Step 1. The charge crossing any cross-section of the wire in time dt is given by the current dQ = Idt Step 2. So if Q is the total charge crossing the cross-section, it is calculated by

3 3

Q=

Idt =

1 exp{3t}dt =

1 e9 = 0.33 (C) 3

Since current is calculated from the movement of charge, and the cross-section of the conductor, we dene a vector eld, J, which is a current density in amp/m2 , and which should give us the current. That is the small current which passes through the small surface element S is I = J S = J nS = J =J

to n nS to n + J to n

213

J n n J J

to n

J to n

dS (a)

dS (b)

Figure 6.2.: The current and J.

The relation between I, J and S are shown in the accompanying gure. It is important to note here that there may be current in a non-conducting medium which may be due to motion of charge between two points, say in air1 , which may also lead to a current, which is termed as a convection current. Furthermore, when the average velocity of a small volume of charge is v, then the current density should be proportional to the instantaneous velocity of that volume Jv (6.6)

Looking at the units. The units of J are (C/s).(1/m2) while the units of v are m/sec so U C/(sec.m2 ) = [x](m/sec) where [x] are the units of some quantity. Then [x] = C/m3 But C/m3 are the units of the charge density. So [x] = m . J = m v (6.7)

U

Where m is the mobile charge density. The meaning of term mobile will become clear shortly. The equation says that at any point, the current density J is equal to the product of the charge density and the velocity with which the charge is moving at that point. This equation applies to both convection as well as conduction currents. EXAMPLE 6.3 In a straight wire of diameter 1 mm the current is 1 A nd the magnitude of the current density if the charges are uniformly distributed across the cross-section of the wire. Step 1. If I is the current, A the area of cross-section and J the current density, then I=

Area o f CS

1

J dS = JA

214

Step 2. So J = I/A = 1/( (0.001/2)2) = 1.2732 106 (A/m2) x x

z x y

z x

(a) 3D View

If we now refer the next gure we can understand some important concepts concerning m , the mobile charge density, n, the density of charge carriers and J the current density. We2 visualise a situation when a conducting medium containing mobile charge carriers is under the inuence of an external electric eld in the ax direction. We concentrate on a small rectangular parallelepiped, buried in the medium, with sides x, y andz. Referring to the gure, which shows such a parallelepiped, we can see that a small number of charge carriers, N, are enclosed by it. For a metal, the mobile charge carriers would be electrons. Now if each carrier caries a charge q then the total charge in the parallelepiped would be equal to Nq. Or the charge enclosed by this volume, V = xyz, would be Q = Nq (6.8) And the density of charge carriers and charge would be n= and m = N V (6.9)

Q = nq V

(6.10)

respectively. Please note that m must not be confused by the excess charge density v . For example, in the case of a metal, the medium is electrically neutral. That is, the charge contribution due to the number of mobile free electrons is equal to the charge contribution due to the metal nuclei, making the medium neutral. In the case of a metal m is the density of the charge contributed by the conduction band electrons. Note that generally no eld

2 The

explanation given below is an extremely simplied one to help the reader visualise the concepts involved

215

may be produced by m 3 . In most cases D m (6.11)

The parallelepiped should be placed in such a way that the side x should be oriented along the direction of motion of the carriers as shown. The average velocity4 is v ax x/t (6.12) If in small time t charges get displaced by a distance x then the parallelepiped enclosing the charges moves parallel to itself as shown. Since the charge carriers have moved, the total (elemental) current through the face with sides y, z is I = J S I= where J = m x ax = m v t (6.15) (6.16) J S (6.13) (6.14)

S = yzax

EXAMPLE 6.4 Estimate the mobile charge density, m , of silver. Step 1. The density of silver is 10.5 103 kg/m3 ; silver has an atomic weight of 107.9. So 107.9 gm have NA = 6.022 1023 atoms. Step 2. Or the number of atoms per kg is 6.022 1023/107.9 103 = 5.58 1024 atoms/Kg (6.17)

Step 3. Multiplying the number of atoms/kg by the density of silver (kg/m3 ), we get the number of atoms/m3 . 5.58 1024 10.5 103 = 5.86 1028 atoms/m3 (6.18)

Step 4. The valency of silver is 1. So if each atom contributes a maximum of 1 electron to the conduction band. m = 5.86 1028 1.602 1019 = 9.39 10 C/m

9 3

(6.19) (6.20)

It may so happen that these mobile charge carriers may accumulate somewhere and cause excess charges. In that case an electric eld will be produced. 4 We are assuming that all the charge carriers are moving in the same direction

3

216

EXAMPLE 6.5 Using m for silver, and J from Example 6.3 estimate the magnitude of the velocity v of the mobile charge carriers. Step 1. We use the Formula 6.7. J = m v = 9.39 109v Step 2. From Example 6.3, J = 1.2732 106 so v = 1.36 104 (m/s)

n J V

J dS = I=

v /t dV

J dS = dQ/dt

Let us apply the divergence theorem to J. Referring to Figure 6.4, in a region where there are currents, JdV = J dS (6.21)

But

S

dQ dt

(6.22)

The negative sign indicates that the charge inside the volume is decreasing. But Q = v dV. So v dV (6.23) JdV = t Or since space coordinates and time are independent of each other we can change the order of dierentiation and integration: t v dV = v dV t (6.24)

217

Using the last result: JdV = Which becomes J = v t (6.26) v dV t (6.25)

This is the continuity equation. It connects the current density with the rate of change of charge in any region of space. And the integral form is J dS = v dV t (6.27)

The importance of the contiuity equation is that it tells us that charge is conserved. Charge is neither created nor destroyed. EXAMPLE 6.6

5

Obtain the continuity equation from Maxwells equations We can get this result from Maxwells equations as well, which has a current term present in it, Step 1. Namely, H = D/t + J (6.28) The equation says that the curl of H is equal to the current density plus the rate of change of the ux density. Step 2. If we take the divergence of this equation ( H) = (D/t + J) (6.29)

Step 3. Using the identity ( A) = 0, and (D/t) = /t( D) (since space and time are independent of each other) the previous equation becomes J = D t v = t

(6.30)

where we have used Gausss Law, namely, D = v . EXAMPLE 6.7 Apply the continuity equation to a straight wire carrying a current I and circular cross-sectional area A. Step 1. Let the wire be oriented in the z direction. Step 2. Considering a Gaussian surface: lower and upper surfaces: z = z0 , z1 ; side surface = a (= A/)

5 This

is a little advanced

218

Step 3. (as shown in Figure 4.33) we nd that J dS = Step 4. With J dS = 0 (No current leaves through the side) J dS =

Bottom

Lower surface

J dS +

Upper surface

J dS +

Side

J dS

Side

Top

J dS = IA

Step 5. Also since there is no accumulation of charge with time dQ =0 dt Q = constant = 0 Step 6. So J dS = v dQ dV = dt dt

The band theory of materials postulates that the outermost shell of atoms have two types of energy bands. Referring to Figure 6.5, 1. A valence band, where the electrons are tightly bound to the nucleus. These electrons do not take part in charge transport. Under the inuence of an external electric eld the nuclei and the valence band electrons take part in formation of minuscule dipoles. 2. A conduction band whose lower edge is generally higher than the upper edge of the valence band. Here if electrons are present, they take part in current formation under the inuence of external elds. In conductors, such as metals, the conduction band overlaps the valence band and at room temperature the charge carriers are electrons. This is shown in the (a) part of the gure. In semiconductors (part (b) of the gure) the two bands do not overlap: there is about a 1 eV 6 gap between the upper edge of the valence band and lower edge of the conduction band. Normally the conduction band is empty but at room temperature some electrons acquire enough energy migrate from the valence band to the conduction band giving rise to two types of charge carriers: conduction band electrons and valence band holes. In the case of dielectrics (part (c) of the gure) the band gap is of the order of 6-7 eV. In this case electrons never acquire enough energy to move from the valence band to the conduction band.

219

C Energy C

1 eV

C C: Conduction Band BG

> 6 7 eV

BG

V (a)

V (b)

V (c)

Figure 6.5.: The energy levels of the outermost shell of materials. The gure shows the valence and conduction band in (a) Metals, (b) Semiconductors, and (c) Dielectrics

Area A

Conductivity =

Current I Length l + Voltage V

Figure 6.6.: Ohms law

220

We all know Ohms Law which applies to a conductor V = IR (6.31)

V and I are the voltage across and through the conductor while R is the resistance of the conductor. If l is the length of the conductor, A the area of cross-section and the conductivity of the material R= l A (6.32)

Going over to eld quantities. The electric eld E = V/l The conduction current density is J = I/A then El = JA E= J J = E l A (6.34) (6.33)

(6.35)

The two quantities E and J are actually vectors while is a scalar. The vector equation is J = E (6.36) This is Ohms law in vector form. In a good conductor can take very large values. We can look upon a metal in a microscopic form. Due to the presence of an electric eld, electrons are accelerated. However on acceleration they are slowed down by collinsions with the atoms of the lattice. So v = a (6.37)

where v is the average velocity of electrons, a is the acceleration and is the mean free time between collisions. As we know, a= so v =

6

eE me eE me

(6.38)

(6.39)

which is 40kT. The average energy of electrons in the material is at most only few kT

221

or J = m v = so =

m eE me

(6.40) (6.41)

m e me

EXAMPLE 6.8 Find the current density in copper for an electric eld of 1 V/m. What is the current due to this eld in a wire of d =1 mm diameter? Step 1. Copper has a conductivity of 58.14 106 /m. So an electric eld of E = 1 V/m gives J = E = 58.4 106 A/m2 Step 2. This is a very high current density. A 1 mm diameter wire has an area of 2 d = 1.96 107 m2 A= 2 Step 3. The current in the wire is therefore I = JA = 11.46 A A perfect conductor is one whose conductivity, , tends to innity. Copper practically satises the denition. EXAMPLE 6.9 Find the resistance of the non-uniform resistor whose geometry is shown in Figure 6.7. The resistor consists of deposited material of conductivity and thickness t in a coaxial mode. Step 1. The current enters the resistor at = a and leaves at = b. For the small section between and + d the resistance dR is dR = d t

222

Step 2. Since all these miniscule resistors are in series along the path of the current, they have to be added together, R

innite number

d t

or R=

1 t

b a

d 1 b = ln t a

(6.42)

Suppose in a conductor charge accumulates somewhere in the interior of a conductor, for some reason. What happens? First of all from Gausss law a D eld develops. Also since we are working in a conductor, the charge dissipates. From Equation 6.28, H = D/t + J Since there is no magnetic eld, then at any point D/t + J = 0 Also at that point, D = 0 E = (0 /) J J + J=0 t 0 J = J0 e

t

0

(6.43)

(6.44)

Equation 6.44 says that after the accumulation of charge at any point, a current density develops which dissipates the charge till no further remains. The rate at which the charge dissipates is proportional to exp{/0t}. EXAMPLE 6.10 A charge of 1 C accumulates in copper. Find the time it takes to dissipate to 1018 C. Step 1. Since Step 2. Therefore Step 3. For copper J = J0 e v (t) = v0 e

t

0

t

0

= 1e

t

0

223

we can see that even if a 1 C charge accumulates in copper, it dissipates in the twinkling of an eye! Is there an electric eld in a conductor due to accumulation of charge? Observing Equation 6.44 and using J = E, the electric eld dissipates at the same rate E = E0 e(/0 )t (6.45)

Electron

E vd v = E

Figure 6.8.: The motion of electrons under the inuence of an external eld

The mechanism by which actual conduction takes place is as follows. Electrons are under constant thermal motion. As they move about in the conductor lattice they impinge upon the stationary metal atoms, rebound, then hit another atom, and so on, somewhat as shown in the gure. Under the inuence of an external eld, they continue to move in the manner discussed, but along with that they constantly drift slowly against the eld. The total motion of such a large number of these mobile electrons together constitutes a current. The relationship between the drift velocity, vd , and the electric eld is a linear one for any particular material vd = e E (6.46) where e is a constant called the mobility. The negative sign is there since the motion of electrons are in the opposite direction of the eld.

EXAMPLE 6.11 Estimate the drift velocity of electrons for silver in a conductor of diameter 1 mm carrying a current of 1 Amp and compare the result with the velocities attained due to thermal motion. Also estimate the mobility for silver. Step 1. The conductor carries a current of 1 A. So the current density, J is J= I r2 = 1.27 106amp/m2 J = m vd

(6.47) (6.48)

224

and the value for m = 9.39 109 C/m3 for silver from Example 6.4. So vd 1.36 104 m/s The drift velocity. Step 3. The electric eld is related to J by J = E (6.50) (6.49)

Step 4. The experimental value of for silver is, = 6.25 107 /m, which gives the value of the electric eld to be E 0.02032 V/m Step 5. Since vd = e E so the mobility for silver is e 6.65 103 m2 /(V.s) (6.52) (6.51)

(If we compare the drift velocity of electrons with the thermal velocity of electrons we nd that there is many orders of magnitude dierence between the two as shown now.) Step 5. The kinetic energy of electron under normal thermal motion is about a few kT. Step 6. Let us estimate the velocity of an electron whose kinetic energy is exactly 1 kT, just to get an idea of the velocities involved. Step 7. For T = 3000 K 1 me v2 = kT th 2 vth = 2kT me

9.5 104m/s Compare this value with vd 1.36 104 m/s. Suppose a conductor was suddenly immersed in an electric eld. What would happen? Initially there would be a eld inside the conductor, which would set up large currents. In the case of Figure 6.9 (a), the electrons would move very quickly against the direction of the electric eld and on the left side of the conductor. The right side would be positively charged due to the stationary nuclei bereft of electrons. The migration of the electrons to the left as shown in Figure 6.9 (b) would result in a zero resultant eld inside the conductor. The argument is that if there was a eld inside there would be a perpetual migration of charges producing heat which is not seen in practice. On the surface of the conductor, for the same reason, there would be no tangential electric eld since the charges would migrate on the surface perpetually. There would be a charge on the surface but distributed in such a way that only a normal resultant electric

225

E=0 ++

++

+ + + + +

eld would be present. The surface of the conductor would be a equipotential surface. These results are summarised 1. No electrostatic eld is present inside a conductor 2. On the surface of the conductor electric elds are only normal to the surface 3. Charges are present on the surface but distribute themselves so as to make the surface an equipotential one

EXAMPLE 6.12 A spherical metal shell with inner radius r0 and outer radius r1 has a very small hole in it through which a positive charge Q is introduced as shown in the gure. 6.10 (a) Find the elds everywhere and discuss. (b) If the shell is grounded, then what happens? Example 6.12 part (a) Step 1. We neglect the small hole as it will not change the elds greatly. We proceed to apply all we have learnt to the case of a charge Q enclosed by a spherical shell. Using a spherical Gaussian surface with the charge Q as centre and radius r < r0 , if we recall, the electric eld is given by E= Q r r < r0 V/m 40 r2 (6.53)

Step 2. We now increase the radius of the Gaussian sphere. When r1 > r > r0 the electric eld inside the metal is zero. So, by Gausss law, the total charge enclosed must be zero. Which means that the inside surface of metal shell must have a uniform surface charge whose total value is Q. The surface charge 0 (r = r0 ) = Total Charge Q C/m2 = Total Surface Area 4r2

0

(6.54)

226

Hole to introduce charge + + 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 r1 11111111111111111111 00000000000000000000 r0 11111111111111111111 00000000000000000000+ 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + Q 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + 11111111111111111111 00000000000000000000 11111111111111111111 00000000000000000000 + + 11111111111111111111 00000000000000000000 +

Figure 6.10.: A charge Q enclosed by a spherical shell

Step 3. The eld satises all the conditions discussed: the eld inside the metal is zero; it is normal to the metal surface r = r0 , and since there are no tangential components of the E eld, the surface r = r0 is an equipotential surface. Step 4. Mobile electrons have moved to the inner surface of the shell. So the outer surface develops a surface charge which is positive. Since the inner surface has a total charge equal to Q, the outer surface must have a total charge of Q consisting of the immobile nuclei whose outer electrons have migrated to the inner surface of the shell. Now on increasing the radius of the Gaussian surface still further, r > r1 , the total charge enclosed is once again Q: Q at the centre, Q on the inner surface and Q on the outer surface. Step 5. The electric eld then becomes E= Q r r > r0 V/m 40 r2 (6.55)

and the surface charge on the outer surface is 1 (r = r1 ) = Example 6.12 part (b) Step 1. When we ground the outer surface which is positively charged, negative charge rushes in from ground and neutralises all the positive charge on the outer surface. The inner surface is still negatively charged because of the presence of the charge at the centre, and the electric eld inside the shell is Q C/m2 4r2 1 (6.56)

227

unchanged. Outside the inner surface there is no electric eld anywhere Q r < r0 2r (6.57) E = 40 r 0 r>r

0

Normally charges are present in situations where other material media are also present. What happens when we introduce a large ground plane near a positively charged point charge? Obviously 1. The surface of ground plane will become charged with negative charges; 2. The ground plane will become an equipotential surface. So 3. The electric eld will be be perpendicular to the ground plane at all points on the surface of the ground plane.

Q E field streamlines +

Q

Figure 6.11.: A point charge near an innite ground plane

In fact whenever a conductor is introduced near any charge distribution, the what has been outlined in the enumerated points 1 to 3 will always take place. The situation will be somewhat as shown in Figure 6.11. The gure rings a bell. If we recall the electric eld of a dipole the elds look similar not only is the electric eld similar, but the eld above the ground plane is indeed the same. The solution of such types of problems is done by a well known method, namely, the method of images. In this method, 1. We start with some set of charges which give us the electric eld and the potential everywhere. 2. If some potential surface is replaced by a thin metal plane then the electric eld and the potential eld are not disturbed because the metal acts like an equipotential surface. 3. We have now a new problem which consists of the same set of charges and the metal plane which we have just put in place

228

Sphere

Figure 6.12.: Method of images applied to a single charge and a sphere.

4. The problem outlined in the enumerated 3 above has the solution as in 1. EXAMPLE 6.13 Apply the method of images to the dipole problem Step 1. Using the formula of Section 5.5, the potential due to the two identical charges is Q V(r) = 4 0 1 z d 2

2

+ y2 + x2

d z+ 2

Step 2. In this equation let z = 0. Immediately we can see that V(r)|z=0 = 0 which is an equipotential surface. Step 3. Between the positive and negative charges, at half the distance between them, the potential surface V = 0 is an innite plane. Step 4. An innite metal plane can now be introduced right there. The elds everywhere are the same as before. The lower charge can be removed, and the elds above the plane are still the same, while the elds below the plane vanish. EXAMPLE 6.14 Apply the method of images to the problem of a single charge placed inside of a sphere as shown in Figure 6.12 The method of images may be applied to a sphere as analysed by Tikhonov (1963) who considers the case of a point charge in a hollow metallic sphere. Step 1. Referring to the gure, we would like to nd the potential, V(r), inside a sphere of radius R, centred at O, due to a point charge q inside the sphere at position a. Step 2. The stipulation is that the image of this charge with respect to the rast charge for the sphere is placed at (R/a)2a. It has a charge of qR/a. Step 3. Hence for the conguration of the two charges shown,

1 2 + y2 + x2

(6.58)

229

V(r) =

q (qR/a) 1 + (anywhere) 40 |r1 | |r2 | q (qR/a) 1 = 40 r2 + a2 2ar cos 4 2 2 + R 2R ar cos r 2 2 a a q q 1 = 2 2 40 r + a 2ar cos 2 a2 r + R2 2ar cos

R2

(6.59)

Step 5. Therefore the potential inside the sphere for a single charge q is given by Equation 6.59.

6.4.4. Semiconductors

A semiconductor behaves just like a conductor does, except that there are two kinds of charge carriers. This is so because valence electrons acquire enough energy to enter the conduction band, and they interact with the electric eld and give rise to currents. On the other hand atoms bereft of electrons in the valence band become positively charged, they also become charge carriers as well. These are called holes. Since there are two types of charge carriers Je = me ve = me e E

(6.60)

Jh = mh vh = mh h E where Je is the electron current density Jh is the hole current density me is the mobile electron charge density mh is the mobile hole charge density ve is the electron drift velocity vh is the hole drift velocity e is the electron mobility of the material

(6.61)

230

h is the hole mobility of the material and E is the electric eld within the semiconductor In any conducting material m = nq (6.62) m , being the mobile charge carrier density; n, the carrier concentration (number/m3 ); and q, the value of the charge: q = e for electrons and q = e for holes. Using these expressions J = Jh + Je = nh evh + ne eve

(6.63)

Where ne is the electron concentration and nh is the hole concentration. The velocity of the holes is in the direction of the electric eld, while the velocity of the electrons is in the direction opposite to the eld. vh = h E ve = e E (6.64) (6.65)

(6.66) (6.67)

(6.68)

EXAMPLE 6.15 A sample of silicon is doped with a type 3 element (having 3 outermost electrons) with an eective concentration of 3 1023 atoms per m3 . Find the conductivity of the material. e = 0.14 and p = 0.05 m2 /(Vs). e = 1.6 1019C. Step 1. The sample is silicon, a type 4 element and a semiconductor. Step 2. The impurity is type 3 so the material is a p-type semiconductor. nh = 3 1023. Step 3. Since nh ne (due to the presence of an impurity) nh h |e| = 2400. /m

231

Did you know? John Bardeen (1908 1991), an American physicist and engineer, is the only person to have won the Nobel Prize in Physics twice. First with W. Shockley and W. Brattain in 1956 for the invention of the transistor, and then 1972 when he proposed a fundamental theory of superconductivity (in collaboration with others.) The transistor as we all know has fundamentally changed society with the development of micro-miniaturised electronic components and has made available many modern electronic devices: the telephone, television and computers among many others. Bardeens proposed theory is also used in MRIs a .

a Magnetic

Resonance Imaging

6.4.5. Dielectrics

Referring to Figure 6.5 on page 220, dielectric materials are those materials which have a minimum band gap of 6-7 eV between the valence band and conduction band. For conduction to take place, electrons must be present in the conduction band. If one remembers that 1 eV corresponds to an energy dierence of about 40 kT at room temperature, a quick calculation based on the Maxwell-Boltzmann model shows that, very very few electrons have enough energy to enter the conduction band. Any book on semiconductor physics will help the reader understand this point (Tyagi 2004 Sze 1969). As a result, practically all the electrons are present in the valence band and are tightly bound to the nucleus. A detailed treatment of dielectrics will not be given here, but the basic idea is that a dielectric polarises under the inuence of an external eld into dipole moments at the atomic/molecular levels. A gure of how this takes place is included.

Nucleas

- - +- - -

(b)

Figure 6.13.: Polarisation of a single molecule under the inuence of an external eld. (a) molecule when the eld is absent (b) molecule when the eld is present

The gure shows the two cases, the rst case (a) when no eld is present. An electron cloud exists and spherical symmetry is maintained. In the second case (b), the electron cloud shifts to the left as the eld is directed toward the right. The dipole that is created has a dipole moment p = qd C m (6.69)

where d is the position vector directed from the negative charge to the positive charge.

232

Since we are talking at an atomic level we can talk of a polarisation density, P, which is the number of dipole moments/m3 , but which have been added vectorially. If in a very small volume V there are N dipoles then 1 P = lim V0 V

N i=1

Then, it turns out that under the inuence of an external eld E, the relationship between E and P is a linear one P = 0 e E (6.71)

pi

(6.70)

where e is called the electric susceptibility. The ux density in the dielectric material is related to E and P by the relation D = 0 E + P = 0 E (1 + e) = 0 r E r = 1 + e is the relative dielectric constant. EXAMPLE 6.16 Find the polarisation density in a dielectric material of r = 1.5 immersed in an electric eld of 1 V/m. Step 1. First Step 2. As a second step e = r 1 = 0.5 P = 0 e E = 0 0.5 C/m2 (6.72) (6.73) (6.74)

6.5. Capacitance

If we connect two metal bodies by an emf source like a battery we nd that current ows from one body to the other, through the source. As the current ows one of the masses becomes negatively charged while the other becomes the positively charged. Since each of the two masses is a conductor, each becomes an equipotential surface, with potentials which we designate as V+ and V . It is obvious that V = V+ V , where V is the emf of the battery. When this potential dierence is reached, the current stops owing and charge is established on each body, as shown in the gure, according to the capacity of the system. The capacitance of this conguration is dened as C= Q V (6.75)

233

V = V+ V V+

+Q

V

Q

The capacitance of various congurations are very important from the viewpoint of electrical engineering. In low frequency (from d.c all the way up to hundreds of megahertz) circuits, capacitors are required and their design is of paramount importance. In this section we investigate the approximate capacitance of the parallel plate capacitor, with plate area A, separation d and lled with a dielectric with dielectric constant r . The accompanying gure, Figure 6.15, shows such a parallel plate capacitor.

1111 0000 1111 0000 1111 0000 1111 0000 1111111 0000000 1111 0000 1111111 0000000 1111 0000

z b a 2h y x (a)

Area A

+Q d Q (b)

(c)

Figure 6.15.: Geometry and elds of the parallel plate capacitor (a)3-D view (b) Crosssection (c) Rough sketch of the electric eld

To investigate this conguration, let the top and bottom plates of the capacitance have a charge Q and Q, respectively. The surface charge density, s and s , on the inner surfaces of the top and bottom plates may be approximated

234

by Q/A and Q/A respectively. We draw a Gaussian surface which would be a rectangular parallelepiped with sides a= depth, b = width and height = 2h. Let only half the height, h, be toward the bottom plate and the other half above the top plate. The electric eld lines would run from the inner surface of the top plate to the inner surface of the bottom plate (see 154). Applying Gausss theorem to the the parallelepiped

D dS =

sides

D dS +

top surface

D dS +

bottom surface

D dS

(6.76)

Recalling the case of a sheet of charge, we expect that, to a rst approximation, on the inner surface of the upper plate, D will have the form D = [0, 0, Dz ]. On the upper surface of the top plate, however, D will be approximately zero.

sides

D dS =

left side

D y a y dydz + +

behind

right side

Dx (ax ) dydz +

front

Now from the formulation Dx and D y on the sides will be approximately zero. Hence D dS 0

sides

The eld on the upper surface of the top plate would also be zero. So D dS 0

top surface

x=x0 +a bottom surface y=y0 +b y=y0

D dS =

x=x0

= Dz ab The charge enclosed is s ab where s is the surface charge density on the inner Q surface. Recall that s = A Q ab A Dz = 0 r Ez = Q/A Q Ez = (0 r A)

Dz ab = s ab =

(6.78)

235

The potential dierence between the plates would be V = Ez d = and C= Qd (0 r A) (6.79)

Q 0 r A V d

(6.80)

which is the approximate value of the capacitance of a parallel plate capacitor. In this formulation, we have neglected the fringing elds at the extreme ends of the capacitor plates, and the elds on the outer surfaces of the both the top and bottom plates.

We look at another example, that of a coaxial line. The importance of the capacitance (per meter) of a coaxial line is important in transmission line theory. Much of the television cables laid by cable companies consist of coaxial lines. Computer cables also consist of two- and four-wire lines. In this section we investigate the capacitance of the coaxial line shown in Figure 6.16. The inner and outer conductors have a radii of a and b respectively. Between the the two conductors there is a dielectric of dielectric constant r . The surface charge density on the inner surface is s C/m2 . Radius b

C=

1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000

Radius a

2r 0 loge b a

F/m

We assume that the elds are radial, that is, they diverge from the positive charge on the inner conductor to the negative charge on the outer conductor. Therefore the ux density eld, D is assumed to have a structure, D = D , 0, 0 in the cylindrical coordinate system. The eld is also assumed to be uniform throughout the length of the of the coaxial line. A Gaussian surface is drawn as shown in Figure 6.16. The surface is cylindrical with a radius, , a < < b

236

and the length of the cylinder is h. Integrating on the Gaussian surface D dS = =

cylinderical surface

sides

D dS +

cylinderical surface

D dS

D dS

=

cylindrical surface z0 +h 2 0

D ddz

=

z0

D ddz (6.81)

z0 0

dSdz

z0 +h 2

s a ddz (6.82)

= h sa2 equating these two equations 2hD = hs a2 = Q s a s a E = r 0 D = The potential dierence between the inner and outer conductor is

b

Vab =

E dl =

s a r 0

b s a s a b loge = loge r 0 r 0 a a

(6.86) (6.87)

Q = Vab

hs a2 loge b a

2r 0 h loge

b a

This is the capacitance of a length h of the structure. For h = 1 m, the capacitance/meter is C= 2r 0 loge

b a

F/m

(6.88)

237

Let us look at the example of a two conductor line illustrated in Figure 6.17. We nd the capacitance of such a structure using a slightly convoluted approach to a structure which we have tackled earlier. Each conductor has radius a and their centres are separated by a distance D. The surface charge density on the left and right conductor are such that the the total charge per meter on each is l and l Coulombs/m respectively.

L L

S 1111 0000

1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000

1111 0000 S 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 1111 0000 2a

C=

loge

D2 4 a2 +D 2a

(F/m)

To solve this problem we consider another problem, that of the dual line charge considered on page 198 . We know that the equipotential surfaces are cylinders. Using the method of images (Section 6.4.3 on page 228) we can replace each of the equipotential cylinders by a metal surface. Referring to Figure 6.17, lines L and L are the two line charges, with lines charges l and l coulombs/m; the equipotential surfaces are S and S , which are replaced by metal cylinders. Using the analysis outlined in Section 5.6 on page 198, if V1 is potential on the left equipotential cylinder, the centre line of the cylinder is located at d 1 + k2 x k = , y = 0, z = z 21 2 k k = exp

where

V1 2 0 VL 2 0 = exp l l l VL = loge k 2 0

VL is potential on the left conductor. Similarly the potential on the right conductor is VR 2 0 V2 2 0 = exp 1/k = exp l l

7

238

l l loge (1/k) = loge k 2 0 2 0 l VL VR = loge k 0

VR =

To calculate k we proceed as follows. The x coordinate of the centre of the left conductor translates to 2 D d 1+k (6.89) = 2 2 k2 1 Solving for d k2 1 D k2 + 1 (6.90)

d=

solving for a and substituting the value of d a= Solving for k D 2 4 a2 + D k= 2a using this value of k. C= 0 = loge k

kD k2 + 1

(6.92)

0

D2 4 a2 +D 2a

(F/m)

(6.93)

loge

EXAMPLE 6.17 8 Find the capacitance of two concentric spheres and from there nd also the capacitance of a single sphere (Figure 6.18) The electric eld between the two spheres is E=

8

Q ar 4r2

r0 < r < r1

Important example

239

+ + 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 0000000000000000000 1111111111111111111 1111111111111111111 0000000000000000000 r1 1111111111111111111 0000000000000000000 r0 1111111111111111111 0000000000000000000+ 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 Q 1111111111111111111 + 0000000000000000000 +Q 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + 1111111111111111111 0000000000000000000 1111111111111111111 0000000000000000000 + + 1111111111111111111 0000000000000000000 +

r0

V=

r1

Q dr 4r2

Q r0 4r r1 Q 1 1 = 4 r0 r1

1 r0

r1 1

(F)

if the outer sphere is removed with r1 then the capacitance of a sphere of radius r0 (in air) is C = 40 r0 Did you know? Records indicate a German scientist named E. G. Von Kleist invented the capacitor in 1745. However, about the same time P. van Musschenbroek, from the University of Leyden, Holland, invented a capacitor in the form of a glass jar which became famous as the Leyden jar. (See Figure 6.19) The Leyden jar consisted of a glass jar lined on the inside and outside with metal foil, and partially lled with water. The water really had no part to play, though at that time it was thought to play an important role. The glass in fact was the dielectric, and the charge was stored on the metal foils. To convey the charge to the inner foil, there was a metal chain connecting the innerfoil to the top of the jar.

240

Metal rod Stopper

Glass jar

Metal chain

There is very strong relation between capacitance and resistance. Let us study Figure 6.14. The capacitance is given by C= Q = V

A D dS L

E dl

A E dS L

E dl

(6.94)

where L is the integration along any line connecting one conductor to the other; the line may be curved or straight. A is the area on the surface of any one of the conductors; the integral D dS

A

is equal to the total charge residing on the conductor, and is the permitivity of the medium. Let us consider the same geometry but with the case where the medium has a conductivity . In this case the current I is given by I=

A

J dS =

E dS

which ows alond the electric eld lines. The denition of voltage dierence is exactly the same as before. So V = I

L

R=

A J dS

E dl

A E dS

E dl

(6.95)

241

The integrations being performed are indentical. Therefore, RC = or R= Let us take the following example EXAMPLE 6.18 Find the resistance of length l of a coaxial line (geometry as shown in Figure 6.20) where the current moves from the inner conductor to the outer conductor of the line. The inner and outer conductors have radii a and b respectively. The medium between the inner and outer conductors is a material of conductivity . Step 1. The capacitance per meter of a coaxial line is (Equation 6.88) C= 2r 0 loge

b a L

A E dS

E dl

A E dS L

E dS

(6.96)

(6.97)

Step 2. The capacitance for length l is c = Cl Step 3. The resistance for length l is given by (Equation 6.97) R= r 0 = c l 2r 0 / log e = loge

b a

b a

2l

242

What happens when there is an electric eld in a region where more than one media are present? Suppose there is an electric eld present in a region comprising air, dielectrics and metal bodies. What is the behaviour of the eld in that region?

y z x

1 2

R

1

a

2

d c b

y x Dielectric Boundary

n

E field (a) (b)

t1

n 1

Dielectric Boundary

(c)

Figure 6.21.: The behaviour of the electric eld near a dielectric boundary

In this section we will examine the behaviour of the eld on the very boundary of two dissimilar media. The cases of interest are the dielectric-dielectric boundary and the dielectric-metal boundary. A dielectric (region 2) is immersed in an external electric eld as illustrated in Figure 6.21. The region outside the dielectric (region 1) consists of a medium with permittivity 1 (for example if the medium is air then 1 0 ). The immersed dielectric has a permittivity of 2 . We concentrate on a very small region R on the boundary of the two dielectrics shown in Figure 6.21(a) and apply the integral form of the electrostatic Maxwells equation to the interface. E dl = 0 (6.98)

This is shown in Figure 6.21(b). The gure depicts a close-up of the interface where a loop a-b-c-d has been shown, on which the line integral of the electric eld is to be computed. The arrows superimposed on the loop show that integration is being carried out in the counter-clockwise sense as per the requirement. From the small coordinate system shown on top of the gure, n az is the normal to the boundary. The normal is directed toward region 1. t1 is the tangent to the boundary and in this case it is equal to ax . n, t1 , t2 in that order, form a right-handed coordinate system. We assume that d-a and b-c are vanishingly small, while a-b and c-d are small enough so that that the eld is approximately constant over these lengths. length (b c) = length (d a) 0 length (a b) = length (c d) |x| 243

Applying Equation 6.98 to the loop a-b-c-d, E1 ax |x| + E2 (ax ) |x| = 0 where E1 and E2 are the electric elds in regions 1 and 2 respectively. The integrations over the line lengths b-c and d-a are assumed to be negligible and so have been dropped from the equation. The previous equation now becomes Ex1 Ex2 = 0 (6.99)

In the same manner by considering a similar loop going into the page the longer sides of which lie in medias 1 and 2, while the shorter side cuts across the two media. We perform a similar line integral over it to get E y1 E y2 = 0 which can be summarised as Et1 Et2 = 0 (6.101) (6.100)

The subscript t is used to signify the tangential component. Et1,2 are the tangential components of the electric eld next to the boundary but in media 1 and 2 respectively. Et1,2 = E2 + E2 x1,2 y1,2

The tangential component of the electric eld is continuous across a dielectric boundary. Considering the electric eld as consisting of a normal and tangential components where n az , t1 ax t2 a y the eld is E1,2 = nEn1,2 + t1Et1 1,2 + t2Et2 1,2 (6.102)

The component En1,2 is the normal component of the eld in regions 1 and 2 respectively; Et1,2 1,2 are tangential components of the eld in directions t1,2 and in regions 1 and 2 respectively. If we take the cross product of the previous equation by n n E1,2 = n nEn1,2 + n t1 Et1 1,2 + n t2 Et2 1,2 = 0 + n t1 Et1 1,2 + t2 Et2 1,2 = n tEt1,2 (6.103)

1 2

t2 are perpendicular to each other). We know that n t is again a tangential component. So (6.105) n (E1 E2 ) = 0 This equation is written in this form because it is independent of coordinate

244

notation and because of the importance of its comparison with E = 0 We will come back to this kind of notation when we consider other boundary conditions. Let us now consider the normal component of the eld. If we apply Gausss law D dS = v to the conguration shown in Figure 6.21(c). The gure shows a close up of region R in Figure 6.21(a). We draw a small pill-box as shown in (c) whose height is negligible and the top and bottom areas are so small that the eld is assumed not to change much in those regions height of pill box 0 Area of the top and bottom A Applying Gausss law to this pill-box with the knowledge that that there is no accumulated charge on the interface (s = 0) D1 az A + D2 (az ) A = 0 Dz1 Dz2 = 0 (6.106) (6.107)

or in other words, the normal component of the ux density is continuous across a dielectric-dielectric boundary. Dn1 Dn2 = 0 n (D1 D2 ) = 0 Compare this last equation with D = 0 Let us investigate what the behaviour of the eld is near a dielectric-metal boundary. We know from our earlier discussion Section 6.4.1 on page 221 that there can be no electric eld inside a conductor. In the presence of a eld, a surface charge develops and makes the surface an equipotential one. The Figure 6.22(a) shows a metal body of conductivity immersed in a dielectric of permittivity 1 . We once again apply Maxwells Equation 6.98 on page 243 to the dielectric-metal body E dl = 0 Applying this equation to the loop a-b-c-d in Figure 6.22(b) and, using the same arguments as earlier in this section, (disregarding the fact that we are dealing with a metal) we have (6.108)

245

y z

1 1

x z d c b

n t1 n

y x DielectricMetal Boundary

1

Metal

n (E1 E2) = 0

Et1 = Et2

(6.109)

but there can be no eld inside the metal; that is Et2 = 0. Therefore, just outside the metal, the tangential electric eld must be zero. Et = 0 Just outside a metal surface E=0 inside a metal (6.110)

Similarly applying Gausss law to the pillbox shown on the boundary interface of Figure 6.22(c). D dS = v (6.111) height of pill box 0

Area of the top and bottom A then, applying Gausss law to this pill-box, whose volume is V D1 az A + D2 (az ) A = v V (6.112)

where v is the volume charge density accumulated near the surface. However we know from earlier arguments that the charge exists only on the surface v V = s A

246

where s is the surface charge density D1 az A + D2 (az ) A = s A Dz1 Dz2 = s and so in general we have, independent of coordinate notation n (D1 D2 ) = s (6.115) (6.113) (6.114)

Where n is the normal to the surface but directed toward the outside from the metal, that is medium 1; s is the charge on the surface of the metal; and D1,2 is the ux density outside and inside the metal, respectively. That this equation is indeed true is borne out by observing the units. Note the comparison with D = v Coming back to the earlier equation, we know that inside a metal D2 = 0 so Dn = s Just outside a metal surface D=0 inside a metal (6.117) (6.116)

EXAMPLE 6.19 A uniform electric eld, E1 meets a dielectric with dielectric constant r lling the half space as shown in Figure 6.23. Find the eld inside the dielectric. Step 1. The eld is split into two components, tangential and normal to the dielectric boundary, E1 = E1t + E1n Step 2. Quantifying these components E1t = E1 cos E1n = E1 sin

247

also D1t = 0 E1 cos D1n = 0 E1 sin Step 3. The tangential component at the boundary is continuous so E2t = E1t = E1 cos Step 4. The normal component of the D is continuous, hence D2n = D1n = 0 E1 sin Step 5. The electric eld is therefore (in the dielectric) is E2n = D2n E1 sin = 0 r r

Step 6. Find what the eld looks like in the dielectric: the magnitude rst |E2 | = = E2 + E2 2t 2n E2 cos2 + E2 1 1 sin2 2 r

= E1 cos2 + (sin/r )2 the magnitude of the electric eld in the dielectric is less than that in air. Step 7. Visulaising the electric eld in the dielectric to be E2 = ax E2t + a y E2n then the angle from the x axis is = tan1 which is less than . E(sin /r ) = tan1 (tan /r ) E cos

Let us calculate what happens as we move point charges into a region of space. Figure 6.24 shows how charges from innity are moved into a region of space. The gure shows charges Qi , i = 1, 2, . . . , N being moved to positions given by ri , i = 1, 2, . . . , N from innity. For example, the rst charge of magnitude Q1 is moved to the position R1 (0) (the origin) from R = . The amount of work done is zero, since no other charges are present to produce an electric eld. The work done to move charge number 2 from innity to r2 is equal to the potential at r2 multiplied by the magnitude of the charge: W2 = Q1 Q2 40 |r2 r1 | 248 (J)

z (2) (1)

Q2

...

...

Q1 Q3

(3)

And the total work (WT,3 for three charges) done is WT,3 = W2 + W3 Q1 Q3 Q2 Q3 Q1 Q2 + + = 40 |r2 r1 | 40 |r3 r1 | 40 |r3 r2 | How do we represent this mathematically? Let us try

i=3 j=3

WT,3 =

i=1 j=1

Qi Q j 40 ri r j

i,j=3

Qi Q j 40 ri r j

i, j = 1 i j

249

This would be WT,3 =

that is every term would occur twice. Therefore the correct formula would be 1 WT,3 = 2

i,j=3

Q2 Q1 Q1 Q2 + + 40 |r2 r1 | 40 |r1 r2 | Qi Q j 40 ri r j

i, j = 1 i j

In the same way, for N charges, the work done would be WT,N = 1 2

i,j=N

Qi Q j 40 ri r j

(6.118)

i, j = 1 i j

Let us play around with this equation to get a signicant result. The equation rewritten is j=N i=N Qj 1 WT,N = Qi 2 40 ri r j

i=1 j=1 i j

the term in the square brackets is the potential at ri due to all the other charges. If we write the square bracket summation as

j=N

Vi =

j=1

Qj 40 ri r j 1 = 2

i=N

with j

(6.119)

then WT,N

Q i Vi

i=1

(6.120)

Now instead of large charges, let the charges accumulated be small, that is, each Qi is replaced by Qi and N can made very large WT,N = 1 2

i=N

(Qi ) Vi

i=1

this accumulation can done by replacing Qi = vi vi where vi vi is a very small volume of charge which is moved into the region where the accumulation

250

is taking place9 . Then WT,N = 1 2

i=N

vi vi Vi

i=1

V

(6.121)

The volume V is over the v . That is where v is present, there is a contribution to the volume integral, and where there is no charge density, there is no contribution. We can therefore safely integrate over all space. WT = Now we know that D = v and (DV) = V D + D V therefore WT = = = 1 2 1 2 1 2 ( (DV) D V)dv (DV) dv + 1 2 (D V) dv D (V) dv (Divergence Th.) 1 2 v dv V

all space

(6.122)

all space

all space

all space

surface at

(DV) dS +

1 2

all space

= 0+ = 1 2

1 2

all space

D (V) dv

all space

D Edv

(E = V)

A surprising result. The work done on the charges has been stored in the electrostatic eld! We = 1 2 D EdV = 1 2 |E|2 dV (J) (6.124)

all space

all space

The term we = |E|2 is the energy density in J/m3 at all points permeated by the electric eld.

9 Note

that here we vi instead of Vi for the dierential volume so as not to confuse it with the potential Vi .

251

EXAMPLE 6.20 Find the approximate energy energy stored in a parallel plate capacitor of capacitance C = (A/d) where A is the area of the plates and d is the separation between the plates. The electric eld between the capacitor plates is given by |E| V d

where V is the potential dierence across the capacitor plates. The total stored energy is therefore V 2 1 .|E|2 (volume between the capacitor plates) 2 (Ad) 2 2d A V2 . = 2 d 1 2 = V C 2

The current is related to the charge in a wire by I = dQ/dt where dQ/dt is evaluated at any cross-section of the wire. The current density J and the current I are connected by the relation I=

cs

J dS

where cs is the crossection of the wire. At any point in a conducting medium, if the charge carriers have a velocity v and the charge density of these mobile carriers is m then J = m v If the number of carries per unit volume is n (No./m3 ) and each charge has a charge of q (C), then the charge density is = nq The continuity equation states that J = t

The continuity equation in integral form says that the amount of current leaving a closed surface is equal to the total time rate of depletion of charge in the volume enclosed J dS = dV t

252

Ohms law V = IR where V is the voltage or potential dierence across the resistor of value R, and is I the current through it. Ohms in point form is J = E where J is the current density at a point and E is the electric eld at that point and is the conductivity. The resistance of a piece of material is given by R= l A

where l is the length of the material (the direction in which the current will travel), A is the area of cross-section and is the conductivity. The relaxation time, , for a material is the amount of time charge takes to dissipate in a material from its initial value (Q0 ) to Q0 e1 . for conductors is (/)1 . The mobility () in a conductor or semiconductor is dened by the relation v = E where v is the drift velocity of charge carriers in the material and E is the electric eld. The dipole moment, p of two charges q and q is given by p = qd where the vector d is a position vector directed from the negative charge to the positive charge. The polarisation density, P, is the number of dipole moments/m3 in a dielectric is dened by 1 P = lim V0 V

N i=1

where pi are the dipole moments in the dielectric and V is a small volume while N is the number of dipole moments in V. Under the inuence of an external eld E, the relationship between E and P is a linear one P = 0 e E where e is called the electric susceptibility. For a dielectric e = 1 r In a dielectric D = 0 E + P

pi

253

The capacitance of a capacitor is for two metallic surfaces is C= Q V

where Q is the charge residing on one of the surfaces and V is the potential dierence between them. The approximate value of a parallel plate capacitor is C 0 r A d

where A is the area of the plates and d is the separation between them. The capacitance per metre of a coaxial line is C= 2r 0 loge

b a

(6.125)

where a, b are the inner and outer radii of the line respectively and r is the dielectric constant of the material lling the region between the conductors. The capacitance per metre of a two conductors which are in parallel is C= loge

0

D2 4 a2 +D 2a

(6.126)

where D is the distance between the centers of the two conductors and a is the radius of each conductor. The capacitance of two concentric spheres is C= 4

1 r0

r1 1

where r0 and r1 are the radii of the inner and outer sphere respectively and is the permittivity of the material lling the region between the spheres. The capacitance of a single sphere in air is C = 40 r0 where r0 is the radius of the sphere. At a dielctric-dielectric boundary, the tangential E elds and the normal D elds are continuous E1t = E2t D1n = D2n At a dielectric-metal boundary, the tangential E eld is zero and the

254

normal D eld is equal to the surface charge density Et = 0 Dn = s no elds can exist inside a metal. The energy, WT,N stored in an electrostatic eld for N charges, Q1 . . . QN at position vectors r1 . . . rN is WT,N 1 = 2

i,j=N

Qi Q j 40 ri r j

i, j = 1 i j

all space

all space

Chapter Summary

In this chapter, the student learns the connection between the current, current density and charge. (Section 6.2) the continuity equation is discussed. (Section 6.3) students lears the connection between Maxwells equations and the continuity equation. (Example 6.6) the student is introduced to the dierences and similarities between conductors, semiconductors and dielectrics. (Section 6.4) a detailed treatment of of these three types of materials is given. the student learns about the relaxation time for charge in a metal. the student learns about how the eld behaves in the presence of a metal. the method of images is dicussed. properties of semiconductors are discussed. the properties of dielectrics in the presence of an electrostatic eld is discussed. The concepts of polarisation P and electric susceptibilty, e , are explained. the electrostatic concept of capacitance is explained. Three important capacitances are discussed in detail: the parallel plate capacitor, the capacitances per unit length of coaxial and two wire lines. the boundary conditions for electrostatic elds are discussed for both a dielectric and metal. the student learns the concept of energy stored in the electrostatic eld.

255

Review Questions

1. Explain the concept of ux of a vector eld out of a surface and link it with current density J and the current I. 2. Explain why the current density J is directly proportional to the velocity of the charge carriers (v) and and also the mobile charge density, m . 3. Explain the concept of charge neutrality in a metal. 4. Write a short note on the continuity equation and why does it predict the conservation of charge charge? 5. Write a short note on conductors, dielectrics and semiconductors. 6. Why is it that in a conductor, electrons in the presence of an electric eld, are only described as moving with constant drift velocity? (Ordinarily electric elds accelerate electrons.) How is this related to heat dissipated in a conductor? 7. Why is it that metals immersed in an electrostatic electric eld have no elds in their interior? 8. If a current is present above an innite ground plane, explain how this combination may be replaced by the original current and an image current. 9. Explain the concept of polarisation density, P, and explain how it is related to the electric eld? 10. How is energy stored in a capacitor? 11. Write a short note on the boundary conditions of electrostatic elds for dielectrics and metals.

Problems

1. Given the current density J = e2y ax + e2y a y A/m2 (a) nd the total current crossing the plane y = 0.5 in the a y direction in the region 1 < x < 1, 2 < z < 2. (b) nd the total current crossing the plane x = 0.5 in the ax direction in the region 1 < y < 1, 2 < z < 2. (c) nd the total current crossing the plane z = 0.5 in the az direction in the region 1 < y < 1, 2 < x < 2. Ans: (a) 2.943 (b) 14.5074 (c) 0 2. Given the current density J = e2y ax + e2y a y A/m2 . Find the total current leaving the region 0 < x, y < 0.5, 2.5 < z < 3 by: (a) integrating J dS over the surface of the cube; (b) employing the divergence theorem. Ans. -0.15803 A for both. 3. Let the current density be J = 2a sin a A/m2 within a region of space. Find the total current I crossing the surface (a) = 2, 0 < < 2, 0 < z < 5, in the a direction. (b) Evaluate J at P( = 2.4, = 0.08, z = 6.05). Ans. 4. In spherical coordinates 1 J = 20 sin cos ar + a A/m2 r

256

(a) Find the total current owing through the surface r = 1. (b) Find the total current owing through the surface r = 1, 0 < < /2, 0 < < /2. Ans. 5. If a beam of electrons occupying a region space with density 51020 (No/cc) are moving with uniform velocity v = 2106ax (m/s) (a) what is the charge density, ? (b)the value of J? If the beam is passing through a circular pipe of 1 cm radius, what is the current? Ans. 6. A circular metal wire of radius a is placed with axis coinciding with the z axis. The current density is maximum on circumference with value J0 az and decays exponentially along the radius at a rate proportional to . Find the expression for the current density. Ans. 7. If J = 20e1000(a) az for 0 < < a is the this current density in a circular wire with radius a oriented along the z-axis, what is the current in the wire? For which value of is 98% of the current contained? Ans. 8. Find the resistance per meter of ten strands of copper twisted together with the diameter of each strand equal to .05 mm. Ans. 9. Find the resistance of 1 m of wire with the current density given in Problem 7. 10. A current I = I0 et t > 0, charges one plate of a capacitor of value C. Find the charge on the capcitor plate, and the voltage across the plates as a function of time. Ans. 11. A plane glass is coated with a resistive material of = 105 S/m. The thickness of the coating is 10 m. and the width is 1 mm. Find the resistance of 1 mm of this conductor. Ans. 12. The potential in a region of space is V = ln( x2 + y2 ). The point (1,0) is on a conducting surface. Find the E eld and s at that point. Ans. 13. The potential in a region of space (r = 1) is V = x5 + xy4 10x3 y2 . The point (0,0) is on a conducting surface, z = 0. Find the E eld and s at that point and the equation of the surface in the neighbourhood of (0,0) Ans. 14. Two point charges each with charge 1 pC are embedded in a dielectric of dielectric constant r and placed 1 m apart. Find the force of repulsion when r is equal to (a) 1, (b) 2 and (c) 10. Ans. 15. Calculate the average shift in the electron cloud of a single molecule when oxygen gas O2 (at STP) is placed in an electric eld of 1 V/m. r for oxygen is 1.0005 at 0 C. Hint. 16 g of Oxygen at STP occupies 22.4 l. 16 g of Oxygen also contains 1 Avagadro number of molecules. Calculate the number of molecules per cubic meter. Ans.

257

dielectric

Liquid

Figure 6.26.: Measurement of for a liquid conductor.

16. A sphere of radius 10 cm is charged with 1 C of charge. Another sphere with 5 cm radius without any charge on it is connected by a wire to the rst sphere. Find the charges on both spheres. Ans. 17. A slab of dielectric of large dimensions (width and height very much greater than the thickness, t = 1 cm) is placed at an angle of 45 with respect an electric eld of 1 V/m, as shown in Figure 6.25. Find the electric eld in the dielectric (r = 10) and on the other side. Ans. 18. A liquid of unknown conductivity is placed in a large trough of glass as shown in Figure 6.26. Two rods are immersed in the liquid. The dimensions of the rods are: the diameter of each rod, 1 cm; the distance between the rods, 15 cm, and the depth of immersion is 15 cm. If the voltage applied is 10 V and the current is found to be 100 mA, nd the

258

Metal plates

I C

conductivity of the liquid. 19. For Problem 13, nd the energy stored in the electric eld in the region 0 < x < 1, 0 < y < 1 and 0 < z < 1. The region is air. 20. For a capacitor with two dielectrics calculate the capacitance with respect to the parameters given in Figure 6.27.

1. A coaxial line of length L with inner radius a and outer radius b, (Figure 6.16) is lled with a material with dielectric constant r and conductivity . If a battery of V V with source resistance Rs is connected between the inner and outer conductors nd the current as a function of time. Ans. The capacitance of the coaxial line is given by (refer Equation 6.88) C= 2r 0 L loge

b a

on the other hand the resistance if the line is (Equation 6.97) 2L R= = C log b e a 1

This resistance and capacitance are in parallel. The circiut is as shown in Figure 6.28. When the battery is connected the capacitor acts like a short circuit and the initial current is I(0) = V/Rs . The nal current is zero, and the time constant is R|| C where R|| = R||Rs = Rs R Rs + R

259

so I(t) = IC (t) =

V t/CR|| e Rs

now the current through the resistance is zero at t = 0 and V/(Rs + R) at t = so the current as a function of time is IR = the total current is therefore IT (t) = V t/CR|| V e + (1 et/R|| ) Rs Rs + R V (1 et/R|| ) Rs + R

2. In a circular wire of radius a if the current density across the wire is J = Jz0 1 e/a az 1e

calculate the resistance of length l. Ans. In a tubular section of wire between and + d the current that ows is 1 e/a dI = 2dJz0 1e so the total current is I= = 2Jz0 1e

a 0 x2

(1 e/a)d

a 0

2 2Jz0 l a 0

2Jz0 a2 = e1

x 2Jz0 a x a2 e a 1e 2

a 2x 2 2Jz0 l 2 a x a2 e a x x2 2 2 ax a ea + = 4 2 (1 e)2 =

2 2Jz0 l

(1 e)2

(1 e/a)2 d

(1 e)2

(0.5969a2)

2 (2Jz0 l)/[(1 e)2] (0.5969a2)

2Jz0

2 a2 /(e 1)

0.5969l 2a2

260

Metal plates

3. Find the capacitance of a capacitor with three layers of dielectrics. Ans. Each capacitor has a capacitance Ci = and the capacitance, C, is 1 = C i li A 1 Ci

4. Find the capacitance of two concentric speres (Figure 6.30) with the dielectric between the spheres being a function of r. The radii of the inner and outer spheres is r0 and r1 respectively. The value of the dielectric is 0 r1 at the surface of the outer sphere and is 0 at the surface of the inner sphere and varies linearly with r. Ans. The dielectric can be modelled as r (r) = ar + b 0 0 r1 a= r0 r1 0 r1 r0 0 r1 b= r0 r1

261

Note that in these expressions, if r1 is set equal to 1 then a = 0 and b = 0 . Now if the total charge on the outer sphere is Q and inner sphere is Q, the surface, so the surface charge densities on the outer and inner spheres is Q Q and 4r2 4r2 0 1 The D eld between the two spheres is D= and the electric eld is E= The energy density is Q (ar ) 4r2 Q 4(r)r2

1 E DdV 2 where the integration is to be performed over the volume between the spheres. E= = = E= 1 2 4Q2 2(4)2 Q 4(r)r2

r1 r0

r1

r2 dr (r)r4

Q2 a ln (a r + b) a ln (r) 1 8 br b2 b2

(check by dierentiation)

r0

r0 (ar1 + b) 1 1 1 Q2 a ln 8 b2 r1 (ar0 + b) b r1 r0

setting the energy stored on a capacitor to Q2 /2C Q2 Q2 = a ln r0 (ar1 + b) 1 1 1 2C 8 b2 r1 (ar0 + b) b r1 r0 4 C= r0 (ar1 +b) a ln r (ar +b) 1 r1 r1 b b2

1 0 1 0

in these expressions if a = 0 and b = 0 , then the capacitance reduces to that two concentric spheres.

1. If the current in a wire is I(t) A, then the charge passing through the cross-section of the wire in one second is 1 (a) I C (b) 1/I C (c) 0 I(t) dt (d) none of the above. Ans. (c) 2. If J = eax by cz ax how much current crosses the x-y plane? (a) 0 (b) a (c) 1/a (d) none of the above Ans. (a)

2 2 2

262

3. If J = eax by cz ax , is charge accumulating at (0,0,0)? (a) yes (b) no (c) increasing at a constant rate (d) decreasing at a constant rate Ans. (b) 4. If J = eax by cz ax in a region of space and the mobile charge density is 1 mC/m3 at (0,0,0) what is the drift velocity of the charge? (a) 1 106 m/s (b) 1 104 m/s (c) 1 103 m/s (d) none of the above Ans. (c)

2 2 2 2 2 2

5. If no charge has to accumulate at a point in space, then (a) v /t = t (b) v /t = 0 (c) none of the above Ans. (b) 6. The mobile charge density in a metal is (a) directly proportional to the number of valence electrons (b) inversely proportional to the number of valence electrons (c) there is no connection with the number of valence electrons (d) none of the above Ans. (a) 7. The continuity equation says that (a) matter is neither created nor destroyed (b) the current density is created through the presence of an electric eld (c) charge is niether created nor destroyed (d) the total current leavind a closed surface is equal to the rate of depletion of charge in that surface. Ans. (c) and (d) 8. Which statements are true? (a) The continuity equation can be obtained from Maxwells equations. (b) One of Maxwells equations can be otained from the continuity equation (c) The continuity equation has nothing to do with Maxwells equations. (d) None of the above. Ans. (a) 9. In a semiconductor (a) holes are present in the conduction band which contribute to the current (b) electrons are present in the conduction band which contribute to the current (c) electrons and holes are present in the conduction band which contribute to the current (d) holes are present in the valence band which contribute to the current Ans. (b) and (d) 10. / has the units of (a) charge (b) length (c) mass (d) time. Ans. (d) 11. /2E E (a) has the units of Jm3 (b) is the energy density of the electric eld (c) is a nonsensical quantity (d) has meaning but is nothing to do with energy density Ans. (a) and (b)

***Chapter Complete***

263

The origin of thinking is some perplexity, confusion or doubt. John Dewey

7.1. Introduction

In many situations in electrostatic problems we may know only the boundary conditions of some engineering problem. This may take various forms. For example we may know the voltage on some boundary (as in the form of a voltage on a set of plates) and we may be expected to nd the potential eld in the region enclosed. Or we may be given the charge distribution in a region of space and be expected to obtain the electric eld in that region. In such situations we may have to set up a partial dierential equation and extract the solution. The partial dierential equations and their solution is well known in mathematics as boundary value problems and their solutions. To set up such a partial dierential equation, we start with Gausss law D = v where D is the electric ux density and v is the charge density. Using D = E, being the permittivity of the (homogeneous) medium, we can write the equation in terms of the electric eld, E = v (7.1)

To get a equation involving the potential V we use the gradient relationship between the electric eld and the potential, E = V V = 2 V = v (7.2)

which is Poissons equation. In a charge free region of space this becomes Laplaces equation 2 V = 0 The operator 2 is called the Laplacian. (7.3)

264

A very important point in solving partial dierential equations is that once we have found one solution of the equation for given boundary conditions then do we continue to look for other solutions? Let us take Laplaces equation. Suppose two solutions exist, namely, V1 (r) and V2 (r) both solutions to Laplaces equation in the region V. At the boundary, S, V1b = V2b = V(r)|r=rb (7.4) Where rb is the position vector of the boundary, the surface S. Then let a function be dened as the dierence of these two solutions: (r) = V1 V2 (7.5) (7.6)

=0 =0 2

(7.7)

(7.8)

as both terms on the right are solutions to Laplaces equation. Hence Equation 7.7 becomes (7.9) () = 2 Integrating this equation over the volume V, 1

V

()dv =

2 dv

(7.10)

And also integrating the left of Equation 7.9 over the surface S enclosing V

V

()dv =

dS 2 dv

(7.11)

So

S

dS =

(7.12)

S

dS = 0

1 Note

265

and so

V

2 dv = 0

but

but at the boundary So the constant k is zero. Therefore V1 (r) = V2 (r) everywhere (7.17)

From this it is clear that once a solution of Laplaces equation is obtained, it is the only solution. The same reasoning applies to Poissons Equation. The only place where there is a dierence is a departure in Equation 7.8, where each of the two terms are equal to v / and not zero.

7.3.1. Some One Dimensional Solutions

Out of the two Equations 7.2 and 7.3 the simpler of the two is Laplaces equation, obviously because the is absent. In Cartesian coordinates Laplaces equation becomes 2 V 2 V 2 V 2 V = + + =0 (7.18) x2 y2 z2 Let us consider the simplest of simplest of all cases: let V be a function of only one dimension, namely, x. That is V V(x). This implies that there is no variation in either the y or z directions. Laplaces equation then becomes 2 V =0 x2 performing two integrations d2 V dV = k1 dx = dx dx2 dV dx = V(x) = k1 x + k2 dx (7.19) (7.20)

and

Where k1 and k2 are constants which will be determined by the boundary conditions. Looking at the solution carefully we realise that this solution would apply to two innite plates with a potential dierence of V0 as in the

266

case of the parallel plate capacitor as shown in Figure 7.1. Why? let us compute the E eld E = V = k1 a constant

y

V=0 V = V0 V=

V0 d x

E field

z x=0 x=d x

Infinite Plates

Figure 7.1.: Laplaces equation applied to two innite plates

The electric eld would be a constant only in the case where there would be two parallel plates! To proceed, we apply the solution which we have just obtained to the gure. We set V = 0 at x = 0 and V = V0 at x = d. From the rst condition k1 x + k2|x=0 = 0 k2 = 0; k1 x|x=d = V0 V0 . k1 = d so the total solution is V= V0 x d V0 d (7.21) (7.22)

E = V =

we can see that we have obtained the solution to Laplaces equation for one dimension in the Cartesian coordinate system. To check the answer, we put x = 0 and x = d in the rst of the above equations and we can see that we have obtained the correct solution to Laplaces equation for one dimension in the Cartesian coordinate system. EXAMPLE 7.1 Use the solution to Laplaces equation to nd the capcitance of a

267

parallel plate capacitor. Step 1. The parallel plate capacitor consists of two plates, each of area A placed in parallel, with a separation of d and lled with a dielectric . Step 2. Suppose the upper and lower plates have charges of Q and Q respectively, with voltages of V0 of the upper plate and 0 of the lower plate. Step 3. Since we have two plates, placing our x coordinate on the lower plate, the equation of the voltage is V0 x V= d which satises the boundary condition, V = 0 at x = 0 and V = V0 at x = d. The equation was obtained from the analysis conducted earlier. Step 4. The upper and lower plates have surface charges, s = Q/A. Step 5. Recall that at the metal boundary, the normal component of the electric ux density, D, is given by Dn = s and Dn = En so s = Q = En A Q = AEn

and s is the charge on any plate per m2 . Step 7. The normal electric eld is given by En = Step 6. Therefore the capacitance is C= (AV0 /d) A Q = = V0 V0 d (7.23) dV V0 = dx d

Let us examine Laplaces equation in other coordinate systems and apply it to some specic congurations. In the cylindrical and spherical coordinate systems, Laplaces equation is 2 V = 2 V = 1 V 1 2 V 2 V + + 2 2 z2 (Cylindrical) (7.24)

(Spherical) (7.25)

EXAMPLE 7.2 Find the general solution to Laplaces equation for concentric cylinders.

268

Step 1. Using Laplaces equation in cylindrical coordinates, 2 V = 1 V 1 2 V 2 V + =0 + 2 2 z2

Step 2. We want the solution V such that V = 0 on the inner conductor and V = V0 of the outer conductor of two coaxial cylinders with a and b as the inner and outer radii Step 3. let V V() and V/ = V/z = 0. We want a solution where there is no dependance on the and z coordinates. So, 2 V = Step 4. Integrating this twice d dV dV = k1 d = d d d dV k1 = d dV d = V = k1 ln + k2 d Step 5. Using the boundary condition, V = 0 on the inner conductor ( = a) and V = V0 of the outer conductor ( = b), Step 5a. Then V|=b = k1 ln + k2 =b = V0 k1 ln b + k2 = V0 k1 ln a + k2 = 0 (1) (2) V|=a = k1 ln + k2 =a = 0 dV 1 d =0 d d (7.26)

In the above equations we subtract (2) from (1) to give k1 ln b = V0 a V0 k1 = ln(b/a) V0 ln b ln(b/a)

269

Step 7. By the uniqueness theorem, this solution is the only one and correct one. There are no other solutions. Step 8. The electric eld is E = V0 dV = d ln(b/a)

The reader is encouraged to apply his common sense (ask, how do we check this solution? Where will this be applied) and check that the solution is correct. Also the reader can obtain an expression of the capacitance per meter of the coaxial line. EXERCISE 7.1 Show that the capacitance per meter of the coaxial line with inner radius equal to a and outer radius equal to b is given by C= 2 ln (b/a) (F/m)

What about two concentric spheres maintained at potentials V = 0 and V = V0 ? EXERCISE 7.2 Obtain the one dimensional solution to the Laplaces equation in r for spherical coordinates, apply it two concentric metal spheres of radii a and b. Let the inner sphere be at a potential V0 and the outer one at a potential V = 0. Also devise a method to calculate the capacitance of this combination.

y

Innite cone at = 0 with V = V0 Innite cone at = 1 with V = 0

Figure 7.2.: Figure for Exercise 7.3

EXERCISE 7.3 Find the solution to Laplaces equation for the conguration shown in Figure 7.2.

270

7.3.2.1. Analytic Functions One of simplest ways of solving Laplaces equation, but a method which is indirect in nature, is to use an analytic function of a complex variable. Analytic functions of the complex variable z f (z) = u(x, y) + j

Real part

v(x, y)

Imaginary part

(7.27)

where z = x + jy (7.28) u(x, y) and v(x, y) which are real and imaginary parts of f (z), have the property that they have to satisfy the Cauchy-Riemann partial dierential equations, (Narayan 2001) u v = x y v u = y x (7.29) (7.30)

dierentiating the rst of these equations with respect to x and the second one with respect to y we get 2 u 2 v = 2 xy x 2u 2 v = yx y2 by adding the two equations given above, we get 2 u 2 u 2 v 2 v + 2= xy yx x2 y but at a point which is analytic 2 v 2 v = xy yx we get Laplaces equation in two dimensions 2 u 2 u + =0 x2 y2 similarly we can show that 2 v 2 v + =0 x2 y2 (7.35) (7.34) (7.33) (7.31) (7.32)

(7.36)

271

1

1 .25 .5 0.75 0.75 .5 1 .25

0.5 0 0.5 1

Figure 7.3.: Contour plots of the real (sin(x) cosh(y)) and imaginary (cos(x) sinh(y)) parts of the function sin(z): Real part of sin z

therefore, the real and imaginary parts of every analytic function, f (z), can be a potential source of a solution to an engineering problem. Let us take some concrete examples. Take the case of the function f (z) = sin(z) = sin(x + jy) = sin(x) cosh(y) + j cos(x) sinh(y) where we have used the two relations sin( jy) = j sinh(y) cos( jy) = cosh(y) If we plot the real and imaginary parts of this function u = sin(x) cosh(y) (The real part) v = cos(x) sinh(y) (The imaginary part) as contour plots u(x, y) = 1.0, 0.75, 0.5, 0.25 and 0 v(x, y) = 1.0, 0.75, 0.5, 0.25 and 0 for the region 0 x and 1 y 1, then some very interesting observations follow. Both sub-gures represent a possible potential eld conguration. The usefulness of these plots is that they give us a feel for what the potential looks like in complex 2-dimensional cases. For example if we were to place a metal curved plane specially shaped and innite in extent along the z-direction along the wedge-shaped u = 1 curve (the centre of which is at x = /2, y = 0) and an (7.37) (7.38)

272

1 .75

1 0.5 0

.25 .5 .75 1

.25 .5 .75 1

0.5 1

Figure 7.4.: Contour plots of the real (sin(x) cosh(y)) and imaginary (cos(x) sinh(y)) parts of the function sin(z): Imaginary part of sin z

innite plane along the u = 0 at x = an engineer gets an idea of how the potential changes near a wedge placed at potentials of V = 1 and V = 0 respectively. From the right hand graph, we can see how the potential changes near a right-angled corner. Another interesting function is f (z) = z2 . Both the real and imaginary parts of the function are shown plotted on the same graph. The labels with V apply to the real part of the function while the ordinary labels apply to the imaginary part of the function.

100 50

50V

100 50

25V 25V

25V

50 100 25

50V

50 100

6 4 2 0 2 4 6

Figure 7.5.: The real and imaginary parts of the function f (z) = z2

273

50V

0V

6 4 2 0 2 4 6

25V

50V

Analysing the function in some more detail, z2 = x + jy

2

u = x2 y2 (The real part) v = 2xy (The imaginary part) both parts of the function are hyperbolas (Richard P. Feynman & Sands 2001). The real part of the function z2 with the electric eld plot is shown in gure

= x2 y2 + 2 jxy

50V 25V

0V 25V

0

-2

-4

25V

-6

50V

-6 -4 -2 0 2 4 6

6 4 2 0 2 4 6

6 4 2 0 2 4 6

Figure 7.6.: Contour plot of u = x2 y2 along with the electric eld superimposed on the potential eld.

Another method and which is a direct, analytical method, is the well-known separation of variables technique. Let us consider Laplaces equation in two dimensions (though the method can be applied to 3-dimensional problems as well) 2 V 2 V + =0 x2 y2 Let V = X(x)Y(y) a multiplication of two functions, X which is a function of x

274

50V

25V

50V

while Y is a function of y. Then d2 X 2 V =Y 2 x2 dx d2 Y 2 V =X 2 y2 dy 2 V 2 V d2 X d2 Y + 2 = Y 2 +X 2 x2 y dx dy which gives Y 1 d2 X X dx2

A function only of x

d2 X d2 Y +X 2 = 0 2 dx dy 1 d2 Y Y dy2 =0

(7.39) (7.40)

A function only of y

the second equation is obtained by dividing the rst equation by XY. Now comes a clever argument: a function of x is added to a function of y to give zero! f (x) + g(y) = 0 What does this mean? This only means that the following must be true f (x) = k (a constant: could be zero, negative or positive) g(y) = k (the same constant with a negative sign!) therefore f (x) = 1 d2 X =k X dx2 (7.41) we know that (for a and b for k>0 for k<0 for k=0

d2 X = kX dx2

the reader may dierentiate the above expressions to satisfy himself that they

from the theory of ordinary dierential equations constants) a sinh( k x) + b cosh( k x) X = a sin( k x) + b cos( k x) ax + b

(7.42)

275

indeed satisfy the previous dierential equation. He may use the relations: d sinh( k x) = k cosh( k x) (k>0) dx d cosh( k x) = k sinh( k x) (k>0) dx d sin( k x) = k cos( k x) (k<0) dx d cos( k x) = k sin( k x) (k<0) dx Similarly we can proceed with the y relation g(y) = 1 d2 Y = k Y dy2 (7.43)

d2 Y = kY dy2 or (with $c$ and $d$ constants) c sin( k y) + d cos( k y) for k>0 Y = c sinh( k y) + d cosh( k y) for k<0 cy + d for k=0 so

(7.44)

Quite complex. Instead of hyperbolic functions, we may use instead use exponential functions. An alternative formulation is a exp( k x) + b exp( k x) c sin( k y) + d cos( k y) for k>0 V = XY = a sin( k x) + b cos( k x) c exp( k y) + d exp( k y) for k<0 (ax + b) cy + d for k=0 (7.46) Let us apply these relations to an actual problem. We apply Laplaces equation to the 2-dimensional layout shown in Figure 7.7. The gure shows a rectangular region of width 2a and height b, bounded by metal plates which are maintained at constant potential. The lower and upper plates at y = 0 and y = b are maintained at a potential of V = 0, while the two side plates at x = a and x = a are maintained at V = 20 V and V = 20 V respectively. To analyse this conguration the potential function can be as that given in Equation set 7.45. Since the bottom plate and top plate are both at V = 0 the only function which ts the y direction is sin(y) where is to be determined so as to make the top

276

y

V=0

y=b

Figure 7.7.: Laplaces equation applied to a rectangular region

x=a

plate have a potential of zero. Note that at y = 0 the function is zero. That is sin(y) y=b = sin(b) = 0 or b = m m = b m = 1, 2, 3, . . . m = 1, 2, 3, . . . (7.48) (7.47)

Observing the other boundary condition we must use sinh b as the function since it is an odd function and the boundary condition is 20 V at the left plate and 20 V at the right plate. So the complete solution is

m=

V(x, y) =

m=1

m = dm

Let us see whether the function above ts the bill 1. It satises Laplaces equation (most important) 2. It satises the lower and upper boundaries 3. dm are to be determined to satisfy the left boundary. If it satises the left boundary, it will satisfy the right boundary since sinh( ) is an odd function. 277

So we must proceed to satisfy the left or right boundary.

m=

V(x, y) =

m=1 m=

dm sin m y sinh (m x)

x=a

=

m=1

= V0

m= m=1

dm sin m y sinh (m a) = V0

(7.52)

m= m=1

A little thought tells us that if we multiply both sides of the equation by sin n y and integrate both sides in the interval [0, b] then the orthogonality property of the sine function will give us dm . Lets try this: multiply both sides by sin(n y) with n = 1, 2, . . .m, . . . m= sin n y dm sin m y sinh (m a) = sin n y (V0 )

m=1

m= m=1

dm sinh (m a)

b 0

sin n y sin m y dy = V0

=0 for m n

b 0

sin n y dy

we nd that when n

b 0

sin n y sin m y dy = 0

so every term of the innite sum becomes zero on the left hand side except for the m = n term b b 2 sin n y dy sin n y dy = V0 dn sinh (n a)

0 0

278

the integral

b 0

sin2 n y dy =

b 0

sin n y dy =

dn

n=1,2,3...

we can see that for n = 2, 4, 6, . . . cos ( n) = 1 and for n = 1, 3, 5 . . . cos ( n) = 1 so 4V0 n sinh (n a) dn = 0 using these values of dn

n=

dn = and

n=1,3,5...

n=2,4,6,...

V(x, y) =

n=1,3,

n = dn =

anticipating the next section, let a = 3, b = 7 cm and V0 = 20 V let us calculate dn for n = 1, 3, ... n dn 1 -14.213 3 -0.2991 5 -0.0121 7 -.0006

Let us calculate the potential at (x = 1 cm, y = 5 cm). The top row shows the number of terms in the summation, while the bottom row shows how fast the sum converges.

279

n, number of terms Vn (V) 1 term -5.1562 2 terms -5.3887 3 terms -5.3337 4 terms -5.3337

With the advent of the digital computer most applications involving partial dierential equations and boundary value problems are solved numerically, and Laplaces equation is no exception.

y+h

...

...

...

xh

x+h

The basis of the numerical solution of Laplaces equation, is the following. At a point (x, y) and its neighbourhood (x + h, y), (x h, y), (x, y + h) and (x, y h) the potential function V(x, y) takes on the values V(x + h, y), V(x h, y), V(x, y + h) and V(x, y h). The potential function at a neighbourhood point, say (x + h, y) is given by the Taylor series expansion V(x + h, y) = V(x, y) + h V h2 2 V h3 3 V + + O(h4) + x 2! x2 3! x3 (7.53)

This is the standard expansion where y is treated like a constant. Similarly at the point (x h, y) V(x h, y) = V(x, y) h V h2 2 V h3 3 V + O(h4) + x 2! x2 3! x3 (7.54)

...

280

adding these two equations V(x + h, y) + V(x h, y) = 2V(x, y) + 2 h2 2 V + O(h4) 2! x2 (7.55)

we can similarly obtain an equation for the y-neighbourhood V(x, y + h) + V(x, y h) = 2V(x, y) + 2 adding the above and previous equations V(x + h, y) + V(x h, y) + V(x, y + h) + V(x, y h) 4V(x, y) + h2 but 2 V 2 V + =0 x2 x2 2 V 2 V + x2 x2 (7.57) (7.58) h2 2 V + O(h4) 2! y2 (7.56)

therefore V(x + h, y) + V(x h, y) + V(x, y + h) + V(x, y h) 4V(x, y) or V(x, y) 1 V(x + h, y) + V(x h, y) + V(x, y + h) + V(x, y h) 4 (7.59)

(7.60)

To apply this equation the region where Laplaces is to be solved is divided into a large number of grid points, taking care that the grid points coincide with the boundaries where the potential is specied. The method was applied to a rectangular domain of width 6 cm and height 7 cm as shown in Figure 7.9. The left and right plates were maintained voltages of 20 and -20 V respectively, while the top and bottom plates were maintained at 0 V. After a number of iterations the values stabilised to the ones shown in the gure. The results have been shown after rounding o to the nearest tenth of a volt. At the grid point x=1 cm, y=5 cm, shown circled (=-5.3 V which is -5.26 rounded o), the comparison with the accurate value of -5.3337 V is encouraging.

As already discussed Equation 7.2 in Section 7.1 is the Poissons equation. v (7.61) The equation states that the potential function V(r) satises Equation 7.61, given a set of charges v (r) then the electric eld E maybe obtained everywhere in that region by E = V (7.62) V = 2 V =

281

0 20 0 0 0 0 0

8.8

3.5

3.5

8.8

20

20

11.6

5.3

5.3

11.6

20

20

12.5

5.9

5.9

12.5

20

V=20 V=20

20 12.5 5.9 0 5.9 12.5 20

20

11.6

5.3

5.3

11.6

20

20

8.8

3.5

3.5

8.8

20

V=0

Figure 7.9.: The grid method applied to a rectangular domain of width of 6 cm and height 7 cm.

282

Let us consider Poissons equation where there are two plates of area A with a separation d; the material between the plates is a volume charge of constant magnitude v = c. In Cartesian coordinates Poissons equation becomes 2 V = 2 V 2 V 2 V v c + + = = x2 y2 z2 (7.63)

The boundary conditions are that the lower plate is at a poential V = 0 and the upper plate is at V = V0 . From the symmetry of the problem, let V be a function of only one dimension, namely, x. That is V V(x). This implies that there is no variation in either the y or z directions. This solution is an approximate solution to the problem; Poissons equation then becomes 2 V c = x2 performing two integrations dV c d2 V dx = = x + k1 2 dx dx dV x2 dx = V(x) = c + k1 x + k2 dx 2 (7.64) (7.65)

and

Where k1 and k2 are constants which will be determined by the boundary conditions. At x = 0 when we apply V = 0, we get k2 = 0. Looking further, at x=d d2 (7.66) V0 = c + k 1 d 2 or V0 cd k1 = (7.67) d 2 and hence the potential is V(x) c V0 cd x2 + x 2 d 2 (7.68)

283

Part III.

Magnetostatics

284

8.1. Introduction

The earliest knowledge of magnetism can be traced back to the Chinese, around the year 1000, who discovered naturally occurring magnets called load stone, made up of iron-rich ore. This state of aairs existed till about 1820. It was believed around that time that the Earth was a giant load stone, magnetised in the same way, and the greatest puzzle at was the slow variation of the Earths magnetic eld. That the Earths magnetic eld was changing was conrmed by the fact that at any one place the direction of the compass needle slowly shifted over time. Around 1820, Hans Christian Oersted noted the connection between changing currents and the movement of a compass needle. But he was not able to explain the phenomenon. Experimental investigations by Andre-Marie Ampere in France on two parallel wires carrying current showed that they interacted magnetically. Parallel wires carrying currents in the same direction attracted and anti-parallel currents repelled each other. Ampere further showed that the force between two long straight parallel currents was (a) inversely proportional to the distance between them and (b) proportional to the intensity of the current owing in each. It was up to Maxwell to connect the two type of forces electric and magnetic who neatly tied up both, mathematically, into the now famous Maxwells equations.

The source of magnetic elds are currents. The word source has been put in inverted commas, because currents are only moving charges, and simply put when charges move, they produce magnetic elds. How are currents quantitatively related to magnetic elds? We already know from Coulombs law that electric elds are produced by charges involving an inverse square law. That is 1 E 2 R where R is the distance from the point charge. Similarly magnetic elds are also inversely proportional to the distance from a point current. And what is this point current? Let us write the Biot-Savart law and then understand this. Referring to the Figure 8.1, we see that in a conductor carrying a current there is

285

dH =

I dl rr 4|rr |2

P (r)

dl r

r r

Current, I y x

Figure 8.1.: Figure illustrating the Biot-Savart law

a minuscule current element I dl , located at a position vector vector r where I is the current and dl is a minuscule part of the current carrying conductor, This current element produces a minuscule magnetic eld dH at the point P (r) (at a position vector r). dH = I dl r r 4 |r r |2 (8.1)

Here the source notation has been used: I dl , which is the source term, has been set in the prime notation. r r is the unit vector in the direction of the vector r r . If we substitute rR R = r r Then the equation may be written in a simpler form dH = I dl arR R 4r2 R R (8.3)

(8.2)

The vector rR R may be read as the position vector from R to Rthe source point to the eld point. The equation says that the magnetic eld at the eld point r is proportional to the cross product of the of the lamentary current, I dl , with the unit vector in the direction of the position of the current, arR R ,

286

r r =0

y

I dz a

z

x

Figure 8.2.: z-directed lamentary current at the origin

and is inversely proportional to the square of the distance from the eld point r to the source point r : r2 R . R Let us proceed to evaluate this expression in Cartesian coordinates. In general with no restriction whatsoever, dl = ax dx + a y dy + az dz r = [x, y, z] r = [x , y , z ] r r = [x x, y y , z z ] [x x, y y, z z ] r r = (x x)2 + y y 2 + (z z )2

(8.4)

Using these results, the three dierential components of the magnetic eld are given below in the rectangular coordinate system: dHx = 4 dH y = 4 dHz = 4 I dy (z z ) dz(y y) (x x)2 + y y 2 + (z z )2 I [dz (x x) dx(z z )] (x x)2 + y y 2 + (z z )2 I dx (y y) dy(x x) (x x)2 + y y 2 + (z z )2

3

(8.5)

These expressions are fairly complicated, but to give us a better understanding we apply these expressions to specic cases in Chapter 8.6, for specic current distributions. To get the magnetic eld, we will integrate these expressions. But for the present we apply these equations to a z-directed lamentary current

287

placed at the origin of a coordinate system as shown in Figure 8.2. The gure depicts a lamentary current I dz az placed at the origin of a coordinate system. Here, because the lamentary current is at the origin, r = 0. Plugging in the values with the following reasoning x = 0 since r = 0 therefore y = 0 z = 0 and dl = az dz dx = 0 so dy = 0 I dz y 4 4 I (dz x) x 2 + y 2 + z2

3

dHx = dH y =

x 2 + y 2 + z2

dHz = 0 If we go over to the spherical coordinate system then x = r sin cos y = r sin sin r= and ax = ar sin cos + a cos cos a sin x 2 + y 2 + z2

(8.6)

x 2 + y 2 + z2

x 2 + y 2 + z2

a 3 y

288

Now yax = r sin sin ar sin cos + a cos cos a sin xa y = r sin cos ar sin sin + a cos sin + a cos = (1/2)arr sin2 sin 2 + (1/4)ar sin 2 sin 2 + ar sin cos2 adding these two terms I I yax + xa y = a r sin 4r3 4r3 going back to the magnetic eld there is only one component, namely, dH = I dz sin (A/m) 4r2 (8.7)

the other components being zero. This goes to support the important result that z-directed currents produce -directed magnetic elds. The minuscule eld is proportional to the current I and sin , and inversely proportional to the square of the distance from the source dl. The sin term is present due to the fact that there is a cross product term in the Biot-Savart law. The units of the magnetic eld can be established from the Biot-Savart law. Rewriting the law 4 |r r |2 U Am (r r has no units) Units of dH = m2 = A/m

U

dH =

I dl r r

In electromagnetic theory one will meet with several types of idealisations of the current. Referring to the Figure 8.3, minuscule amounts of these currents are I , J and J . We can write alternate forms of the Biot-Savart law using these s currents. If we recall we have already written out this law for the term I . In some cases the current exists on the surface of a conductor. The idealisation then is that the current is innite thin, but covering the surface. J is the notation s we will use for the surface current idealisation; it has the units of amp/m and is the surface current density. The Biot-Savart law using J is s dH = = 4 |r r |2 J r r s dS dS (8.8)

J (r r ) s 4 |r r |3

289

I dl

J dS s

J dV

S

J (r r ) s 4 |r r |3

dS (A/m)

(8.9)

r is the position vector of the surface current which produces the eld.

r is the position vector where the eld is evaluated dS is a dierential element of surface which contains J s H is evaluated at the eld point r. In the same manner, when the current is a current density, J (amp/m2 ), then J r r

dH =

dV 4 |r r |2 J (r r ) dV = 4 |r r |3 J (r r ) 4 |r r |3 dV

H=

V

dV is a dierential volume element containing the current density J . r is the position vector where the eld is evaluated and r is the position vector of the current density. r and r are shown in Figure 8.4.

290

Element of current

H (r)

J dS or J dV s r

r r r

dH =

or

dH =

J rr s dS 4|rr |2 J rr dV 4|rr |2

z y x

Figure 8.4.: .

Before we read this section, we need to brush up our concepts on the curl (Section 3.3.3) and the line integral (Section 3.2.2). If we go back to Maxwells Equations, we nd that one of the equations is H = D +J t (8.13)

(This equation is given on page 113) Since we are considering the steady elds, the D/t may be dropped and we have H = J (8.14)

This is Amperes Law in dierential form. The law states that the curl of the magnetic eld is equal to the current density at a point in space. If at that point, no current exists, then the curl of the magnetic eld is zero. Integrating both sides of this equation over a surface S as shown in Figure 8.5

S

( H) dS =

J dS

(8.15)

Notice that the gure has two parts (a) and (b). In (a) part of the gure, the surface S bulges out and in part (b) of the gure, S is at. In both cases, L is counter-clockwise and encloses the S. In both cases, the equation given above applies. We recall from vector analysis Equation 3.74, Stokess theorem is, H dl = ( H) dS (8.16)

Where S is the surface of integration, and L is the line which encloses the surface. The line integration has to be carried out in the anti-clockwise sense,

291

J(r) J(r)

J(r) S

S S J(r) S

(a)

(b)

Figure 8.5.

which is indicated by the symbol . So continuing with the previous equation, but incorporating the results of the above equation ( H) dS = H dl = J dS

But J dS = Current through the surface = Ienclosed where Ienclosed is the current enclosed by the line and therefore passing through the enclosed surface. The full expression then becomes H dl = J dS = Ienclosed (8.17)

This is Amperes law in integral form. The line integral of the magnetic eld H over a closed contour in the anti-clockwise sense1 is equal to the current crossing the surface which enclosed by the contour.

1 The anti-clockwise

sense is in the sense of the right hand thumb rule. The integration is carried out in the direction of the ngers while the unit vector of the surface is in the direction of the thumb.

292

Enclosed Surfaces a

c x

Figure 8.6.

Amperes law is like Gausss law: when applied properly to certain symmetrical congurations. It can give us quick results with the minimum of eort. Also Amperes law gives a very sound feel for how the magnetic eld is produced by a current carrying conductors. One of the most popular applications of Amperes law is its application to the straight innite wire of Figure 8.6, which we consider now. The gure shows an innite wire going from z = to z = + carrying a current I . To correctly apply Amperes law, a closed contour must be chosen to the conguration in question. The line integral of the magnetic eld must be equated to the current passing through the enclosed surface. To obtain the current, we may have to do a surface integration. Various loops (or contours) are shown in the gure (for the purpose of carrying out the line integral) which can be considered for application of the law. These will now be considered in turn. Loop a is not good enough because it encloses no current and application of the law will yield no result. The Loop b is not symmetrical though it does enclose the current. Loop c serves our purpose. It has the required symmetry. Loop c is a circle lying on a plane parallel to the x-y plane with radius . On this loop we expect the magnetic eld to have the same value everywhere because of the symmetry of the loop. Using the integral denition of Amperes law H dl = Ienclosed = J dS

293

situation, the left side of the equation leads to H dl = = H d + Hz dz + Hd H d

2

(dz and d are both zero) (H and are both constant on the loop) (8.18) (8.19)

= H

(8.20)

(8.21)

Therefore we can see that the magnetic eld has only one component, namely, the component H = 0, H , 0 = a H Re-examining the magnetic eld, since H 1/, H and H 0 as The magnetic eld curls around the current as shown in Figure 8.7 as per the right hand thumb rule. If the wire were to be held in the right hand, then the magnetic eld curls around the wire in the direction of the ngers, while the thumb points in the direction of the current. The dierential statement of Amperes law states that the curl of H at a point is equal to volume current density, J, at that point. For example for the straight wire, if we were to nd the curl of H at any point except the line = 0, then as 0 (8.22)

294

Straight conductor with direction of current

Figure 8.7.: The direction of the magnetic eld vis-a-vis a straight conductor carrying current

=0 H =0 Hz =0 H ()

(8.23)

The rst equation above is zero because Hz is zero and H is not a function of z. The second equation is zero because H and Hz are both zero. The third

295

equation is zero because H is a constant and so H =

I 2

=0

The second part of the third equation is zero because H = 0. Why is the curl of the magnetic eld zero everywhere? It is zero because everywhere, (except along the z-axis, where the current exists) the current density is zero! There is a magnetic eld, but its curl is zero. On the other hand if we were to consider the integral form of Amperes law and apply it to an innite wire carrying a constant current I H dl = J dS = Ienclosed

H dl = = =

0

2

I d 2 I d 2 (which is = Ienclosed )

=

0

=I

(8.24)

We now apply Amperes law to straight conductor of circular cross-section of radius a and of innite length as shown in Figure 8.8. The wire caries a current I . Given these conditions, the volume current density is J = I az a2 (8.25)

To apply Amperes law in accordance with Equation 8.17, we choose a circular contour of radius b < a along which to carry out the line integration as shown in the gure. Then

2 L

H dl =

H bd

2

= H b

d (8.26)

= 2bH

296

Radius b Radius a x z

Current I

Figure 8.8.: Straight wire of radius a carrying a current I

H is assumed to be constant on the contour, from symmetry considerations. The surface integration is now carried out

=b,=2

J dS =

=0,=0

I az ddaz a2 dd

=b,=2

=0,=0

(8.27)

2bH =

I b2 a2 I b H = 2a2

(8.28)

297

When b > a then

2 L

H dl =

H bd

2

= H b

d (8.30)

=a,=2

J dS = =

=0,=0

I az ddaz a2 dd

I a2

=a,=2

=0,=0

I a2 = 2 a = I

(8.31)

In the above equations the upper limit in the integral sign is = a since the current is only present in the conductor, which is dened by the region 0 2 and 0 a. In addition, it is not surprising that the result is equal to I since the total current crossing the surface is I . Equating the two equations as earlier 2bH = I or H = I 2b

but b may be treated as a variable, which is , in the above equation. Which then gives H = I 2 (8.32)

Outside the conductor, this result is the same as in the case of an innitely thin wire. I 2 a inside the conductor 2a I H = H a = 2 a outside the conductor I a at the boundary

2a

298

0 , 0 , z0 1 , 0 , z0

0 , 0 , z0

0 , 0 , z0

Figure 8.9.: The innite helix

We next apply Amperes law to the innite helix or solenoid shown in Figure 8.9. The radius of the cross-section of the Helix is r and a current I ows through it. To apply Amperes law, we draw a closed rectangular loop, namely, a-b-c-d which will be the path of integration, which is in the counterclockwise sense. The points a,b,c,d in cylindrical coordinates are a in cylindrical coordinates is 0 , 0 , z0 , the point b, 0 , 0 , z0 ,there is the point c is 1 , 0 , z0 and Applying the right hand thumb rule to the path of integration we come to the conclusion that the current enclosed by the loop is the total current going into the plane of the paper. If the helix current is I and the number of turns in the loop is n then nI is the total current enclosed. Then if the number of turns/m of the helix be N, then the total number of turns enclosed is 2Nz0 turns, (because 2z0 is the length of the loop) and therefore the total current enclosed is 2Nz0 I (A). The line integral H dl =

abcd ab

the point d is 1 , 0 , z0 .

Hz dz +

H d +

bc cd

Hz dz +

H d

da

(8.33) (8.34)

= 2Nz0 I (a constant)

299

Figure 8.10.

We can now play with this line integral. But rst we have to have an idea as to the elds existing inside and outside the solenoid. We can get an idea of the elds by imagining that one holds one turn of the solenoid with the right hand with the thumb pointing in the direction of the current. The magnetic eld curls around the turn. Now as we go all round the turn, we can draw a rough sketch of the elds. Such a sketch is shown in the gure for one turn. From the sketch it is clear that the elds in the plane of the loop will be the elds which will present for the innite solenoid. This is so because for the innite number of turns, the eld conguration in the plane of the loop will be repeated again and again. H should be zero because of symmetry considerations. H will also be zero inside and outside the solenoid because there is no H in the plane of the loop for a single turn. Only Hz should be present everywhere inside, and perhaps outside. It is also clear that Hz must be a function of only. Integration 1. Referring to the Figure 8.9, we make line c-d to be inside the helix itself, that is 0 < 1 < r, (where r is the radius of the solenoid). Using this path, no current is enclosed

H dl =

abcd ab

Hz dz + Hz dz +

H d +

bc cd

Hz dz +

H d

da

=

ab

cd

Hz dz since H = 0

300

which implies that Hz 0 = Hz 1 = c a constant (8.35)

This is magnetic eld inside the solenoid. Integration 2. Now we allow the lines a-b and c-d both outside the solenoid, such that r < 0 < 1 , and again no current is enclosed

H dl =

abcd ab

Hz dz +

cd

Hz dz (8.36)

which implies that Hz outside the solenoid is also a constant. But we expect the magnetic eld at to be zero, and the elds are expected to decay slowly. But from the previous equation Hz (0 ) = Hz (1 ) = constant (8.37)

The only explanation which explains this fact is that the magnetic eld outside the solenoid is zero. Hz (0 ) = Hz (1 ) = constant = 0 (8.38)

Integration 3. Now we allow a-b to be inside and c-d to be outside the solenoid, that is 0 < r < 1 , so that

H dl =

abcd ab

Hz dz + Hz dz

cd

Hz dz

=

ab

= 2z0 Hz 0 = 2Nz0 I But the right hand side must be equal to the total current enclosed Current enclosed = 2Nz0 I From the previous equations 2z0 Hz 0 = 2Nz0 I 2Nz0 I Hz (0 ) = 2z0 = NI This is only component of the magnetic eld, the other components being zero. Hence

301

cross-section a

EXAMPLE 8.1 Apply Amperes law to a torus. Step 1. Amperes law applied to a torus is H dl = NI (8.40)

[0, 0, NI] H= 0

(8.39)

where the integration is to be performed along the path shown. N is the total number of turns and I is the current. Note that the current enters into surface enclosed by the path. Step 2.

Step 3. If the circular closed path is chosen outside ( > min + 2a) or inside the torus ( < min ), no current crosses the surface and the eld in these regions is zero.

302

y x

r = [x, y, 0] observation point

Figure 8.12.: Applying the Biot-Savart law to a current carrying straight conductor

8.6.1. Magnetic Field of a Straight Wire

We now calculate the magnetic eld more rigorously to show the student how to go about it. We will rst apply the Biot-Savart Law to a current carrying straight, innite, conductor, shown in the Figure 8.12. dl = az dz r = [x, y, 0] r = [0, 0, z ] r r = [x, y, z ] [x, y, z] r r = x2 + y2 + z2 Using these expressions (8.42)

303

dHx = dH y =

I dz y 4 4 I [dz x]

x2 + y2 + z2 x2 + y2 + z2

dHz = 0

(8.43)

Somewhat simpler. Now we have to keep x and y constant, while integrating over z from to . Let us rst integrate dz z 2 + y 2 + x 2 Using the value of (x, y, z ) I y (x, y, z ) (x, y, z ) 4 I x (x, y, z ) (x, y, z ) Hy = 4 Hz = 0 Hx =

3 2

x, y, z =

= y2 + x2

z z 2 + y 2 + x 2

(8.44)

(8.45)

(x, y, z ) =

1 y2 + x2 1 (x, y, z ) = 2 y + x2 2 y2 + x2

(8.46)

(8.47)

(8.48)

These are the elds everywhere. If we move to any plane parallel to the z= 0 plane, the conguration of the conductor will look the the same and the elds will be identical to the ones we have obtained. We now go over to the

304

cylindrical coordinate system H= I y I x ax + a y + 0az 2 y2 + x2 2 y2 + x2 I y I x cos a sina + = 2 y2 + x2 2 y2 + x2

ax

sin a + cos a

ay

2 y2 + x2 I = a 22 I = a 2

I sin I cos sin a + cos a cos a sin a + 2 2 y2 + x2 I sin cos a + sin2 a + sin cos a + cos2 a = 2 + x2 2 y = I sin2 a +

cancels I cos2 cancels

2 y2 + x2

We can use another way to corroborate this equation. At [x,0,0] a = ax and a = a y ; = x. The elds are therefore Hx = 0 H = 0 I I Hy = H = 2x 2 Hz = 0 (8.49)

also evaluating the elds at [0,y,0]. = y and a = ax which give us the same

Hz = 0

(8.50)

What does the magnetic eld for a straight conductor look like? Scientists and engineers have thought of a simple way to visualise the eld pattern. If one holds the conductor in the right hand with the thumb in the direction of the current, then the magnetic eld lines have a direction as that of the ngers, as shown in Figure 8.7. This is called the right hand thumb rule. We can also plot the eld lines of the magnetic eld. The eld line equations

305

-4

-8

-8

-4

Figure 8.13.: Field lines of the magnetic eld for the straight wire carrying current

are dy dx = Hx H y dx

I y 2 y2 +x2 ) (

dy

I x 2( y2 +x2 )

x2 = y2 + c2

where c is a constant. The eld lines consist of a family of circles with centre [0,0,z], which are shown in Figure 8.13. Some important remarks are required here on the comparison of eld lines of the electric and magnetic elds 1. Electric eld lines start at some charge or start at innity, or some metal surface. 2. Electric eld lines end at a charge, at innity or a metal surface. 3. Magnetic eld lines, on the other hand, form closed loops. Or 4. If a magnetic eld line starts at innity then it also ends at innity.

306

radius

0

r = [x , y , 0] source point

y I Current loop x

Figure 8.14.: Magnetic eld on the central axis due to a current loop

In the next example we apply our Biot-Savart law formulae to nd the magnetic eld along the axis of a current loop shown in the Figure 8.14. The current I ows in the wire and the loop is placed at on the x-y plane. The radius of the loop is 0 . Scrutinising the gure it is clear that r = [x , y , 0] x = 0 cos() y = 0 sin() dx = 0 sin()d

3 3

(8.51)

(8.52)

dHx =

I z0 cos()d 4 2 + z2 0 2 + z2 0 I 2 d 0 4 2 + z2 0

3 3

dH y =

I z0 sin()d 4 =

3

dHz =

I dx (y ) dy(x ) 4 x2 + y2 + z2

3

(8.53)

307

r r dS = dx dy

...

O

J = [Jsx , 0, 0] s

...

Figure 8.15.

Sheet of current y

r = [x , y , 0] source point

2

Hx =

I z0 cos()d 4 2 + z2 0

3

0 2

=0

Hy =

I z0 sin()d 4 2 + z2 0 I 2 d 0 4 2 + z2 0

3 3

0 2

=0 I 2 0 2 2 + z2 0

3

Hz =

(8.54)

The magnetic eld along the axis of the loop has only a single component, namely, the z component, and the eld line starts at and proceeds to . Note that at z = 0 the loop produces a eld Hz = I 20

Referring to Figure 8.15 we can use some of these concepts applied to the case of the magnetic eld produced by a current sheet. The gure shows a

308

current owing in the ax direction with J = Jsx , 0, 0 . If we use Equation 8.9 s the the various terms of this equation (in the most general case, in rectangular coordinates, are) J = ax Jsx + a y Jsy + az Jsz s

r = [x, y, z] r = [x , y , z ] r r = [x x, y y , z z ] [x x, y y, z z ] r r = (x x)2 + y y 2 + (z z )2

Jsy (z z) Jsz (y y) dS

(8.55)

(x x)2 + y y 2 + (z z )2

Jsz (x x) Jsx(z z) dS

(x x)2 + y y 2 + (z z )2

Jsx (y y) Jsy(x x) dS

(x x)2 +

y y 2 + (z z )2

(8.56)

J = ax Jsx s r = [0, 0, z]

= [x , y , z] [x , y , z] r r = x2 + y2 + z2 dS = dx dy

r = [x , y , 0] r r = [x x, y y, z z ]

(8.57)

Here r = [0, 0, z] since the eld on the z axis will be the same as the eld anywhere. Using these equations dHx = 0 dH y = dHz =

Jsx zdx dy

4 4

x2 + y2 + z2 x2 + y2 + z2

Jsx (y )dx dy 3

(8.58)

309

Integrating these equations

Hx = constant(= c)(?)

x,y=

Hy =

x,y= x,y=

Jsx zdx dy

x2 + y2 + z2

Hz =

x,y=

Jsx (y )dx dy

x2 + y2 + z2

(8.59)

In the very rst equation a question mark appears. That is so because the integration is a denite integral and the denite integration of zero is zero. The third integral is zero because the integrand is an odd function of y . So we are left with the second integration Hy =

Jsx zdx dy 3

x2 + y2 + z2

(8.60)

Jsx zdx 3

= 4

4 so

x2 + y2 + z2

Jsx zx

z2 + y 2

z2 + y 2 + x 2

(8.61)

Jsx zx dy Hy = y = 4 z2 + y 2 z2 + y 2 + x 2

y =

x =

Jsx zdy y = 2 z2 + y 2

y =

x =

(8.62)

310

2 Jsx

y z

y =

y Jsx = arctan 2 z

y = y = y =

(8.63)

Hz = 0

If we consider the region above the sheet, the unit vector of the surface toward the region above is az . That is, this is the unit vector of the surface toward the upper region. Then az (Habove Hbelow ) = az = Jsx ax = Js (8.65) Jsx Jsx ay ay 2 2 = az (Jsx ) a y

311

9.1. The Magnetic Flux Density

Let us consider another of Maxwells equations, namely, B = 0 where B is the magnetic ux density. B = H (9.2) (9.1)

What does the previous equation really say or mean? As it stands, in the dierential form, it does not convey much. But if integrate this equation over some volume V enclosed by a surface S and apply the divergence theorem B dV = B dS = 0 (9.3)

then we can understand something more from the above equation. The second equality says that the surface integral of B over the closed surface S is zero. The surface integral of a vector basically gives you the total ux of that out of that surface of that vector. (See Figure 9.1) Here the term B dS

is the total magnetic ux out of the surface S. Therefore, the total magnetic ux out of any closed surface is always zero. Let us compare this with an analogous equation involving the electric eld, which is Gausss law: D dS = Total charge enclosed

If we compare these two equations, then we can come to the surprising conclusion that since the total magnetic ux out of any closed surface is always zero, therefore there are no magnetic charges! Going back to Equation 9.1 we know from our knowledge of vector analysis that when the divergence of a vector is zero, that vector eld must be the curl of some other vector. In other words if B = 0

312

Flux =

S B dS S B dS =

field

Surface

S B

field

Closed surface

S

(a)

Figure 9.1.: Magnetic Flux

(b)

then where A, in theory can be any vector eld. Thus given any vector eld A it will, by the previous equation give us a possible B eld. A is called the vector potential. Can two vector potentials give us the same B eld? Let us investigate this. Let two elds A1 and A2 give us the same B eld B = A1 = A2 therefore recall that when the curl of a vector eld is zero then that vector must be the gradient of a scalar A1 A2 = A1 = A2 + (9.7) (9.8) (A1 A2 ) = 0 (9.5) (9.6) B = A (9.4)

So, vector potential is not unique: two dierent A elds can give us the same B eld. The dierence between the two possible vector potentials is the gradient of a scalar. How do we get a practical or real vector potential? That is given a set of currents. how do we proceed to obtain the vector potential? To do this let us look at another of Maxwells equations, namely Amperes law in magnetostatics H = J

313

or substituting Equation 9.4 in the above equation and using ( A) = ( A) ( ) A we get ( A) = 0 J ( A) ( ) A = 0 J (9.10) B = 0 J (9.9)

(9.11)

Any vector is not fully determined until both is curl and divergence are both specied we choose (for a reason) A = 0 so that the previous equation becomes ( ) A = 0 J (9.12)

2 A = 0 J

(9.13)

On examination of the above equation we nd that we are dealing with three equations 2 Ax = 0 Jx 2 Az = 0 Jz

2 A y = 0 J y

(9.14)

Each of these equations are mathematically similar to the potential equation 2 V = v /0 The solution to the above equation is V (r) = 1 40 v (r ) dv |r r | (9.16) (9.15)

where V(r) is the potential at the eld point r due to charges v (r ) in the volume V whose position vector is specied by the r1 . Using the above result we can assert that2 0 Ji (r ) dV Ai (r) = (9.17) i=x,y,z 4 |r r |

V

where Ai (r) is the ith component (i=x, y or z) of the vector potential at the eld point r due to currents Ji (r ) in the volume V whose position vector is specied

2 Note 1 Note

that dv is used instead of dV to avoid confusion with V, the potential that instead 0 in the denominator there is 0 in the numerator

314

A (r) =

0 4 J(r )dv V |rr |

r r

dV

z

r V

y x

Figure 9.2.: Figure showing the geometry of how the vector potential is calculated

Figure 9.2 depicts the various variables. We can formulate the above equations a bit dierently. Instead of the term Jv dV , we can use other formulations A (r) = 0 4 0 4 Js (r ) dS |r r | I (r ) dl |r r | (9.19)

A (r) =

(9.20)

315

Let us take the curl of the vector potential at a eld point. 0 J (r ) dv (R) A (r) = (R) 4 |r r | = 0 4 0 4

V V V

(R) (R)

1 J (r ) dv |r r |

J (r ) dv |r r |

(9.21)

the nomenclature (R) is the nabla operator operating on the r coordinates only, (which is where the eld point is) and not the r coordinates (which are the coordinates of where the currents are). In the last equation we have used (aA) = a A + a A Now r r (R) r r

1

= (x x)2 + y y 2 + (z z )2 = =

(1/2)

|r r |3 0 4 0 4 (r r )

(r r )

(x x)2 + y y 2 + (z z )2

(3/2)

V V

J (r ) (r r ) |r r |3

Let us apply this expression to a wire of cross-section A carrying a current I. dv = Adl Then then B(r) = 0 4 I(r )dl (r r )

V

JAdl = Idl

(9.22)

|r r |3

(9.23)

316

The important aspect about the vector potential is that we can take the results of the scalar potential and apply then directly. Let us take some examples.

Let us look at a straight wire along the z-axis of a coordinate system, carrying a steady current I. The potential of such a wire with a charge density l is given by l ln y2 + x2 V(x, y) = 2 0 since the vector potential is proportional to the current, and the directions to be properly chosen, we can straight away write (l Iaz and 1/0 0 ) Az = I0 ln y2 + x2 2 I0 ln = 2

Note that in this case only one component of A, namely Az is present since the current is z-directed and the vector potential is in the direction of the current (Examine equation (9.20)). the other components being zero. Therefore using cylindrical coordinates A A A 1 Az A Az a + 1 A = a + z az z

0 0 0

Az a = I0 B= 2

(9.24)

We can proceed along the same lines and nd the eld for two innite wires parallel to the z axis placed on the x-y plane at (x=0,y=d/2) and (x=0,y=-d/2). The potential from electrostatics is l V(x, y) = log 2 0 y + d/2 2 + x2 2 2 y d/2 + x

(9.25)

317

where the wires are charged l and l . Using l 0 the analogous vector potential is I0 log Az = 2 y + d/2

2 2

Iaz , l + x2

Ax = 0 Ay = 0 then I0 B x = y Az = 2

2 y d/2 + x

(9.26)

y + d/2 y + d/2 2 + x2 x

I0 B y = x Az = 2 Bz = 0

y + d/2 2 + x2

y d/2 2 y d/2 + x2

x 2 2 y d/2 + x

(9.27)

we can see that we have obtained quite complicated expressions in a simple manner.

Very often we would like to nd the magnetic eld far from where the currents are placed, the far eld. In such cases it is much easier to nd the eld using the vector potential. Figure 9.3 shows the geometry of the where the currents along with the region of the eld point. The concerned accurate equation for calculation of the vector potential is A (r) = the term r r 0 4 J (r ) dV |r r | 1 r2 2 r cos r + r 2

by the cosine law. is the angle between r and r . The far eld is dened by R R . 1 r2 2 r cos r + r 2 = r 12

r r

cos +

r 2 r

since R /r is much smaller than 1. Therefore we can use a Taylor series expan-

318

A(r)

0 4R2 r V J (r ) ( r ) R dV

z

r >> r r r r

y

x

Figure 9.3.: Figure to derive the far eld vector potential

sion around R /R = 0 1 r2 2 R cos r + R 2 using this expansion we write 0 A (r) = 4 1 r cos 3 cos2 1 r 2 dV + J (r ) + r r2 2 r3 J (r ) dV = 0

V

1 r

First term

3 cos2 1 r 2 r cos + r2 2 r3

Second term Third term

generally we use only the rst two terms. However the term

since the steady current always ows in closed loops3 . The second term always contributes 0 A(r) J (r ) r cos dV 4r2

V

3 This

319

r, observation point

(a/2,b/2)

(a/2,b/2)

(a/2,b/2)

y I x

(a/2,b/2)

r , source point

V

J (r ) ( r ) r dV r

Consider a square loop of steady current as shown in Figure 9.4. Mathematically, the current can be described by the set of equations on the xy plane. Idxax Idya y JdV = Idl = Idxa x Idya y JdV = =I for a/2 > x > a/2, y = b/2 for b/2 > y > b/2, x = a/2 for a/2 > x > a/2, y = b/2 for b/2 > y > b/2, x = a/2

Loop

Idl ax dx + a ydy dx + a y dy

Loop

= I ax =0

Loop

Loop

320

Let us now calculate the second integral J (r ) ( r ) r dV = ax r r Jx (r ) ( r ) R dV + a y r J y (r ) ( r ) R dV

a/2 a/2

r x2 + (b/2)2

a/2 a/2

xx y(b/2)

ax

V

r Jx (r ) ( r ) r dV =

ax R

(I) xx + y(b/2) dx +

(I) xx y(b/2) dx

ax = Iyab R

Similarly for the second integral on the right r = xax + ya y + zaz , r = (a/2)ax + y a y r r = ay r J y (r ) ( r ) r dV = R

b/2 b/2

r (a/2)2 + y2

b/2 b/2

x(a/2) yy

ay

V

(I) x(a/2) + yy dy +

ay R

(Ixab)

(I) x(a/2) + yy dy

Or A= = = = = 0 Iab ax y + a yx 4r3 0 Iab (az r) 4r3 0 Iab R (az r) 4r3 0 Iab az az cos + a sin 4r2 0 Iab a sin 4r2

(9.28)

(We can of course obtain the same result by the long method:

321

a y = ar sin sin + cos sin a + a cos x = r sin cos y = r sin sin using these results

ax y = ar cos sin + cos cos a a sin r sin sin a y x = ar sin sin + cos sin a + a cos r sin cos

= ar r cos sin sin2 + ar sin cos sin cos ar sin sin2 = ar r cos sin sin2 + ar sin cos sin cos + ar sin cos2

the other components being zero.) Based on this vector potential we can obtain the B eld 1 sin A B= r sin 1 sin A = r sin = A 1 1 Ar rA ar + r sin r 1 rA ar + a r r 1 (rA ) Ar a a + r r (9.30)

These expressions are identical with the expressions obtained for the electric eld for an electric dipole given in Section ?? on page ??, except that Qd, the dipole moment is replaced by the expression Iab, which is the current multiplied by the area of the loop. Because of this similarity, Iab is called the magnetic dipole moment. EXERCISE 9.1 Work out the far eld magnetic eld for a current loop of radius a. Show that the expressions for the eld are the same ones which we obtained above except that ab is replaced by a2 .

322

10.1. The Lorentz Force

Forces appear in electrodynamics in two varieties. The rst one consists of the force on a charged particle due to the presence of an electric eld. The law here is F = qE (10.1) where q is the charge, E is the electric eld where the charge is located and F is the force felt by the charge. The rst point to be noticed is that the force on the charge is in the direction of the electric eld. The second point is that the force is proportional to the magnitude of the charge and the magnitude of the electric eld. The second kind of force, experimentally veried, is the force on a moving charge. Referring to Figure 10.1, if a charge q is moving with a velocity v in a magnetic eld with ux density B then the magnitude of the force experienced by the charge is given by F = qvB sin (10.2) where is the angle between v and B. The direction of F is perpendicular to the two vectors: in the direction of the thumb given by the right hand rule, when the ngers are curled from v to B. All this is conveniently written in vector notation as F = qv B (10.3) The total force on a charge in the presence of both the electric as well as the magnetic elds is given by F = q(E + v B) this force is called the Lorentz force. F = qv B (10.4)

11111 00000 11111 00000 11111 00000 v 11111 00000 00000 90 11111 11111 00000 containing Plane 11111 00000 11111 00000 and v q

both B

Figure 10.1.: The v B force

323

B e F v

trajectory y

B = B0 a z

The next question which arises is: where do we apply this new law? (The v B law?) First we require a magnetic eld. Then we require some moving charges so that we can see the eect of the force. Let us start with a single electron moving perpendicular to a magnetic eld. We want it to move perpendicularly, so that the eect of v B is simplied. The situation is depicted in Figure 10.2 which shows B = B0 a z v = v0 a y x = some value which we specify later y=0 z=0 F = ev B The equations of motion are dvx = ev y Bz dt dv y = evx Bz me dt dvz me =0 dt me The third equation integrates to vz = 0 (10.12) (10.9) (10.10) (10.11) (10.7) at t=0 (10.8) (10.5) (10.6)

= ev0 B0 ax

324

Using Laplace transforms with the notation a(t) A(s) the rst two equations can be written as me [sVx (s)] = eV y (s)B0 me sV y (s) v0 = eVx (s)B0 (10.13) (10.14)

L

substituting the rst equation into the second one, the rst equation becomes Vx (s) = or me sV y (s) v0 = eB0 sV y (s) v0 = eB0 sV y (s) + me V y (s) s2 + or V y (s) = taking the inverse Laplace transform v y (t) = v0 cos(c t) (Check: at t = 0 the value of v y = v0 ) and where c = eB0 me (10.16) (10.15)

2

eV y (s)B0 me s eV y (s)B0

2

eB0 me

me s V y (s) s

V y (s) s

2

= v0 = sv0 sv0

eB0 me

s2 +

eB0 2 me

Note that the right hand side is positive since e is negative. c is the cyclotron frequency. vx can now be calculated, since me dv y dt = evx B0 therefore (10.17)

vx =

1 dv y = v0 sin (c t) c dt

Notice that the magnitude of the velocity is constant v2 + v2 = v0 x y (the initial velocity) (10.18)

325

which means that the kinetic energy of the electron neither increases nor decreases and no work is done by this force. Let us nd the trajectory of the electron vx = v0 sin (c t) or dx = v0 sin (c t) dt v0 v0 cos (c t) + x0 where + x0 is x at t=0 x= c c v y = v0 cos(c t) or v0 y= sin (c t) y = 0 at t = 0 c To obtain the geometry of the trajectory of the electron x x0 = v0 cos (c t) c v0 y= sin (c t) c v0 c

2

(10.19)

Hence (x x0)2 + y2 =

(10.20)

the electron is moving in a circle with radius v0 /c , and centre (x0 , 0) with a frequency of c . No work is done by this force! Let us take a look at the general case. Let a charge q move with a velocity v in a magnetic eld with ux density B. Then the dierential work done in time dt is dW = q (v B) dr dt dt (10.21)

where r is the position vector of the charge. But dr =v dt therefore dW = q (v B) vdt = 0! Charges moving in magnetic elds neither gain kinetic energy nor lose it. (10.23) (10.22)

We next consider the case of a long straight metallic wire immersed in a uniform magnetic eld of magnetic ux density B = B0 az shown in Figure 10.3. To get a better feel for the problem, we will analyse the problem from the micro level, and then move on to the macro level. In a metal conductor the mobile charge carriers are electrons. If the density of

326

B = B0 a z F

y

vm = ne

Infinite wire

I x

vd

B0 a z

electrons are n electrons/m3 then the mobile volume charge density is vm = ne (C/m3 ) (10.24)

(10.25)

Let vd = vd a y be the drift velocity, then the force on the wire per meter F is F = lm vd B = lm vd B0 ax (Nt/m) (10.26)

the negative sign occurs since the conventional current direction and the electron current direction are of opposite sign. Therefore F = IB0 ax (Nt/m)

But the current is related to the line mobile charge and the drift velocity through the equation I = l vd (10.27)

and the force on a length of straight line L (in vector notation, I is in the direction L) is F = ILB0 ax = IL B

327

A simpler manner in which we can approach this problem is to start from the equation F = qv B Since the wire is straight, qv may be replaced by IL. So F = IL B (10.28)

We can use the basic equation F = qv B and make it suit our needs: dF = dQ v B = J BdV (10.29) (dQ = v dV) (J = v v) (10.30) (10.31)

= (v dV) v B

In the above equations dF is the elemental force which the charge feels; dQ is an elemental charge; v is the charge density; dV is a volume element; J is the current density and B is the external magnetic ux density. Or alternatively another formulation is the force felt by a dierential current element dF = Idl B (dQv = Idl) (10.32)

Let us consider yet another case: that of a loop in a constant magnetic ux density B = B0 az , shown in Figure 10.4. In this case we use the formulation dF = Idl B so F=I

L

dl B dl B

=I =0

L

328

dl I I

B = B0 a z

since dl = 0

L

(10.34)

L L

since we are integrating on a loop, let the parametric equations for x, y, z be x(t), y(t), z(t). Then

t=t f

dl =

L t=ti

t=t f t=ti

(x = dx/dt, etc.)

= ax x(t f ) x(ti) + a y y(t f ) y(ti) + az z(t f ) z(ti) but since we are considering a loop, x(t f ) = x(ti ); y(t f ) = y(ti ) and z(t f ) = z(ti ). Therefore dl = 0)

L

(10.35)

Therefore no force is felt by a loop carrying a steady current I and immersed in a steady magnetic eld.

The total force on a loop carrying a current is zero, but that does not mean that the loop does not feel a torque. Take the example shown in Figure 10.5. The

329

N

F I a b B o

z + y x origin moves to

I d

S

Figure 10.5.: Torque on a loop carrying a current

gure shows a square loop abcd immersed in a constant magnetic eld B and it is allowed to rotate about a central axis as shown. The magnetic eld (read ux density) is directed from the north pole of the magnet to the south pole, is modelled as B = az B0 . The length of the side of the loop (ab) is L and as a vector it is L = a y L. The other side is of length d. The current owing through the loop is I. The coordinate system chosen is placed at the centre of the loop at the point o with ab along the y axis. Applying now the magnetic force equation along the side of the loop ab: Fab = IL B = I(a y L) (azB0 ) = ax ILB0 similarly the force along the side cd is Fcd = IL B = I(a yL) (azB0 ) = ax ILB0

(10.36)

(10.37)

both these forces tend to rotate the loop in the counter-clockwise sense when

330

viewed from the side ad. The torque about the axis of rotation is due Fab is Tab = r Fab d = (az cos ax sin ) (axILB0 ) 2 d = a y cos ILB0 2 and Tcd = r Fcd d = (az cos + ax sin ) (axILB0 ) 2 d = a y cos ILB0 2 hence the total torque is T = a y ILdB0 cos = a y IAB0 cos where A = Ld is the area of the loop. Notice that the forces on sides bc and da are equal and opposite and pass through the centre of the loop.

(10.38)

(10.39)

I2 dl2 r2

z

r1 I1 dl1

r2 r1

y x

Figure 10.6.: Calculation of the force between two current elements

We now proceed to compute the force which a current element exerts on another one. Ampere, was the rst to explain this eect: he showed that two parallel wires carrying current, attracted each other if the currents owed in the same

331

direction and repelled each other if the currents owed in opposite directions. Consider Figure 10.6 where two current elements exist. The rst element I1 dl1 produces a magnetic eld dH2 at the eld point r2 which is given by the BiotSavart Law: I1 dl1 r2 r1 (10.40) dH2 (r2 ) = 4 |r2 r1 |2 Therefore the eld at point 2 due to the circuit 1 would be an integration of the above equation: I1 dl1 r2 r1 H(2) = 2 L1 4 |r2 r1 |

Normally if a eld at point 2 there is a magnetic eld H(2) then the current element I2 dl2 would feel a minuscule force dF2 = 0 I2 dl2 H(2) Therefore the minuscule element on circuit 2 feels a force dF2 = 0 I2 dl2 H(2) = 0 I2 dl2

L1

(10.41)

I1 dl1 r2 r1 4 |r2 r1 |2

And the total force that circuit 2 feels due to the current ow in circuit 1 is F2 = 0 I2 = 0 dl2 I1 dl1 r2 r1

L1

L2

I2 I1 4

dl2 dl1 r2 r1

L2 L1

4 |r2 r1 |2 |r2 r1 |2

332

11.1. Inductance

As in the case of non-dissipative electric circuit element which store energy, namely, capacitances, the analogous magnetic elements which are non-dissipative and store magnetic energy are inductances. Thus in the case of inductances as we try to increase the current in an inductor, a back emf1 is formed opposing the change. The back emf, vL is given by vL = L dI dt (11.1)

where L is the inductance of the inductor, and I is the current through it. We know that the inductor does not dissipate energy so the power vL I that is being supplied to the inductor is being stored as magnetic energy, in the magnetic eld. The magnetic energy, Wm , stored at any time is

t

Wm =

t=0

dI Idt dt

vL t

=L

t=0 i

dI dt dt

=L =L

i=0 I2

IdI 2 (11.2)

Based on this denition we can get the value of an inductor from L= 2Wm I2 (11.3)

While the current is increasing, at the eld level we nd that the magnetic eld is also on the increase, and energy calculated above is being stored in the eld. Generally an inductor consists of N turns, and each turn links a certain amount

1 emf:

333

Field

Area of integration

Figure 11.1.: Area of integration for calculating the ux linked by a single turn, m

crosssection

B dS

(11.4)

therefore the total ux linked by N turns () is, = Nm The inductance of the inductor is L= Nm = I I (11.6) (11.5)

As a start let us do a quick calculation for an inductance in the form of a coil. Referring to Section 8.5 we know that a solenoid carrying a current I and having n turns per meter produces a uniform magnetic eld H in its core, given by H = nI A/m (11.7)

Now consider an inductor whose length is l with N turns wound on a cylinder with area A. Then n(turns/m) N(turns)/l(m). The magnetic eld in the core of the inductance is therefore (approximately) H NI l (11.8)

The magnetic ux density, B is related to the magnetic eld by B = H 0 The ux linked per turn is m 0 The total ux linked is = Nm 0 NIA l N2 IA l (11.10) NI l (11.9)

(11.11)

334

Radius a

Figure 11.2.: Inductance of a coaxial line

N2 A 0 (11.12) I l of course this value of inductance is only approximate, since the magnetic eld was borrowed from that of an innitely long solenoid where the eld in the core is uniform, but the approximation is a fairly good one and can be relied upon as the starting point of an inductance calculation. L=

Let us do some more inductance calculations, this time the inductance per meter of a coaxial line. The conguration is shown in Figure 11.2, which also shows the area of integration to calculate the ux m . For a current I owing into the centre conductor, the magnetic eld between the inner and outer conductor is given by I H= a (11.13) 2 in a coordinate system with the z-axis pointed down the line and whose origin is placed at the centre of the inner conductor. B is therefore B = 0 I a 2

=b, z=1

m =

=a, z=0

0 0 I b ln 2 a

335

Radius a

Figure 11.3.: Calculation of inductance from the eld point of view

In Section 6.8 we studied about the energy stored in the electric eld in a region of space V is given by 1 D E dV (J) (11.16) We = 2 where

V

1 we = D E (J/m3 ) (11.17) 2 is the electrical energy density at any point in space. The electrical energy stored in a capacitor was 1 (11.18) We = CV 2 2

where V was the potential dierence across the capacitor plates. In a similar manner we now would like to investigate the energy stored in an inductor, from the eld point of view. To do this, we take the the coaxial line as a case in point. Taking a small slice of the coaxial line of length dz along the z-axis, (see Figure

336

11.3 on page 336) the magnetic energy stored in this small section is dWm = 1 dL I2 2 1 dm 2 = I 2 I =b 1 = ( B ddz) I 2 =a

dm =2

1 2

B ddz (

=0

H d)

Ampere s Law

0 ( 2

(11.19)

where dL is the innitesimal inductance of that part; in the last step, the two integrations are taken together. Therefore Wm = Generalising this result, Wm = 0 2 H H dV (J/m) (11.21) 0 2

=b, =2, z=1 2 H dddz (J/m) =a, =0, z=0

(11.20)

1 wm = B H (J/m3 ) 2

(11.22)

Let us therefore calculate the inductance from the eld point of view for a coaxial line. The inductance is L= 2Wm = 0 I2 |H|2 dV (H) I2 (11.23)

337

Applying this formula to a coaxial line, H = I/2 a . So |H|2 dV = I2 = 2 4 = = the inductance is therefore L = 0 0 b |H|2 dV = ln (H/m) 2 2 a I (11.25) I2 42 I2 dddz 42 2

=b, =2, z=1

1 dddz

2 ln

b a (11.24)

b I2 ln 2 a

which is the same result as earlier. (see page 336) From Equation 11.3, we can get another formula for the inductance. Consider 1 2 and from so B HdV = 1 2 ( A) HdV (11.26)

(A H) = H ( A) A ( H) H ( A) = (A H) + A ( H) [ (A H) + A ( H)] dV

S

(11.27) (11.28)

B HdV = =

A H dS +

A JdV

(11.29)

the integral on the right is zero since we may take the surface S at innity where the elds are zero, so Wm = 1 2 B HdV = 1 2 A JdV (11.30)

VC

Consider a loop of radius R as shown in Figure 11.4 of circular cross-section, where a is the radius of the wire. The vector potential A near the wire is given

2 An

advanced topic

338

z

I

cross-section a

R

x

source point

Figure 11.4.: A circular loop of wire carrying a current I

observing a point in the wire we can see that there is only component of J, namely J and J = J ax sin + a y cos (11.31) where J is the magnitude of the current density and is uniform throughout the cross-section. Writing Considering a point on the x-z plane, here = 0 r r = = and 3 Ax = Ay = Az = 0 0 IR 4 0 IR 4

2 0

=0

(11.33)

the rst equation is zero since the integrand has odd symmetry about , the second equation is not zero and is A and the third equation is zero since Jz = 0. The Ax and A y integrands are shown in Figure 11.5. The second equation can be expressed in terms of the complete elliptic integrals K and E, A = where k2 =

3 see

0 4

(2 k2)K(k) 2E(k) k2

(11.34)

(11.35)

Jackson J.D., Classical Electrodynamics, 1999, John Wiley and Sons, Section 5.5

339

0.5 sin(x)/sqrt(4+1-2*2*1*cos(x)) 0.4 0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 0 1 cos(x)/sqrt(4+1-2*2*1*cos(x)) 0.8 1 2 3 4 5 6

0.6

0.4

0.2

-0.2

-0.4 0 1 2 3 4 5 6

340

for r = R, = /2 we nd that k 1. On closer examination we discover that K(k) becomes innity but E(k) remains nite and 1. The functions near this point 4 can be approximated by 1 z1 1 + O((z 1)2 ) + ln 4 + (1 ln4) (z 1) + O((z 1)2 ) K(z)|z1 ln(1 z) 1 2 4 4 (11.36) z1 z1 1 + 2 ln 4 + 1 + O((z 1)2 ) (11.37) E(z)|z1 1 + ln(1 z) 2 4 4 we can see from these equations that as z 1 the largest terms in these expansions are 1 K(z)|z1 ln(1 z) + ln4 2 E(z)|z1 1 which we introduce into Equation 11.34. A 0 1 16 2I ln 2 4 2 (1 k) (11.40) (11.38) (11.39)

where we have substituted r R in the loop. We now proceed to evaluate 1 k on the inside of the wire. We allow r = R + where < a (the radius of the loop). Here we nd the Taylor series expansion of k is k 1 3 2 + 3 + ... 4R2 4R (11.41)

substituting in Equation 11.40, and taking only the 2 term 1 k (/2R)2 A 0 I 2 ln 8R 2 (11.42)

I (2) a2

x2 ln(8R/x) 3x2 2 4

4 http:

341

therefore 1 2 A J d d = = integrating along the wire Wm = 2R = 0 I 2 4 ln(8R/a) 3 2 (11.44) 0 I I a2 3 (2) ln(8R/a) 2 4 a 2 2 0 I 2 4 ln(8R/a) 3 2 (11.43)

0 I 2 3 R ln(8R/a) 2 2 I2 2

=L

(11.45)

There are various other results given in the literature where instead of the constant 3/2 they have either 2 or 7/8. This dierence is chiey due to the level of approximation of the K and E functions.

If a loop or coil produces a magnetic eld, B1 and a second loop or coil links the eld of the rst one then the ux linked by the second loop or coil may be called m21 and read as the ux linked in coil(loop) 2 due to eld 1 is m21 =

S2

B1 dS2

(11.46)

where S2 is the region where the integration is performed. Then the mutual inductance of the second loop or coil is given by M21 = N2 m21 I1 (11.47)

where N2 are the number of turns in the second coil and I1 is the current in the rst coil or loop. Similarly if the second loop has a current I2 in it which produces a eld B2 then the mutual inductance of coil 1 due to current 2 is given by N1 m12 M12 = (11.48) I2

342

I coil 1 coil 2

Torus

EXAMPLE 11.1 Find the mutual inductance of the two coils with turns N1 and N2 wound on a torus as shown in Figure 11.6. Find also the self inductance of coil 1. From Equation 8.41, if the current I is owing in coil 1 then H1 N1 I 2 min < < min + 2a

The result of Equation 8.41 applies even though the torus is partially wound since for 0 the magnetic eld is conned entirely to the inside of the torus. Or N1 I B1 = H1 = 0 = min + a 20 then the ux linked by a single turn of coil 2 is m1 = B1 Area of cross section = and the total ux linked is N2 m1 = so the mutual inductance is M21 = N2 m1 N1 N2 a2 = I 20 N1 N2 Ia2 20 N1 I a2 20

343

By reciprocity M12 = M21 To calculate the (self) inductance of coil 1, the ux linked per turn is again m1 = so the total ux linked in coil 1 is N1 m1 = hence the inductance of coil 1 is: L1 =

2 2 N1 m1 N1 a = I 20 2 N1 Ia2

N1 Ia2 20

20

11.2.1. Magnetisation

The magnetic dipole of a loop is dened as m = IS where I is the loop current, S is the area of the loop and S = SS (11.51) (11.50)

is the vector normal to the area S. This is shown in Figure 11.7. If we take a look at the expression of the torque on a current loop as considered in Section 10.6 we nd that T = a y IAB0 cos = a y mB0 cos which gives us the important formula T = mB 344 (11.52)

where m is the magnetic dipole. If we consider the far elds of a magnetic dipole using the far eld approximation the magnetic ux density B is given by (see Equation 9.30. B 0 m (2 cos ar + sin a ) 4r3 (11.53)

where m = Iab = IS is the magnitude of the magnetic dipole of the square loop. S = ab is the surface area of the loop. Similar results are obtained for a circular loop. If we take a look at the far elds of an electric dipole the electric eld is E p (2 cos ar + sin a ) 40 r3 (11.54)

p = qd being the magnitude of the electric dipole. A comparison of Equations 11.53 and 11.54 shows why the magnetic dipole is so named. We now consider the nature of magnetic materials. One of the ways in which a material exhibits magnetic properties is that in those materials with unpaired electrons orbit around the nuclei in circular5 orbits with magnetic moment,mo . In the presence of a magnetic eld these orbiting electrons feel a torque as outlined in Equation 11.52, which tends to align the magnetic eld (which is in the direction of mo ) produced by the orbiting electron with the external magnetic eld. Another way by which electrons exhibit magnetic properties is through spin whose magnetic moments can be designated as ms . The result of these two vectors can be denoted as mr with mr = mo + ms (11.55)

Since this a introductory level text book, complex explanations of the theory magnetic materials will be avoided, and only a simple introduction will be given. With this in mind, most materials can be considered as one of three types: diamagnetic, paramagnetic or ferromagnetic. For diamagnetic materials, mr = 0. These are weakly repelled by magnetic elds; Since mr is zero, external magnetic elds have little eect on them. Most elements in the periodic table, including copper, silver, and gold, are diamagnetic. In the second case mr 0. These substances are weakly attracted to magnetic elds and external elds are enhanced by their presence. These are paramagnetic substances. An example of a paramagnetic element is potassium. After the eld is removed, the material also does not exhibit any magnetic properties. Finally we consider ferromagnetic materials. In these materials |ms | |mo | (11.56) (11.57)

mr ms 0

In ferromagnetic materials adjacent atoms crystallise in similar arrangements in regions which are known as magnetic domains. These materials are strongly attracted by external magnetic elds. And after the eld is removed some of these materials retain their magnetism.

5

345

I

Magnetic Material N turns

H

Figure 11.8.: A coil wound round a core

In a magnetic material, immersed in an external eld, in a small volume V let there be N small magnetic dipole moments. Then the magnetisation M is given by

N

M = lim

V0

i=1

mi V

(11.58)

and has the same units as the magnetic eld H. Without going into greater detail we can show that B = 0 (H + M) = 0 r H (11.59) (11.60)

where H is the magnetic eld in the absence of the material. Assuming a linear model we can say that M = m H (11.61) where m is the magnetic susceptibility. m > 0 m 0 m < 0 for paramagnetic materials for ferromagnetic materials and for diamagnetic materials

A magnetic circuit is the analysis of magnetic setup using electrical circuit concepts. Figure 11.8 shows a winding on one leg of a rectangular core made

346

of magnetic material (generally 0 ). Applying Amperes law (assuming a uniform eld within the core) NI = Hl (11.64)

where I is the current in the winding; N is the number of turns of the winding; H is the magnetic eld in the core and l is the length of the contour. H is therefore NI (11.65) H= l the ux m is given by NI m = BA = A (11.66) l In this formulation l m (11.67) NI = A if F = NI then F = Rm where F is called the magneto-motive force or mmf (analogous to emf); R is the reluctance (analogous to resistance)6 (11.69) and R= l A (11.68)

m = F /R

+

F = NI

R=

l A

Let us analyse another magnetic circuit, one with an air-gap in it, as shown in Figure 11.10. If we employ Amperes law and apply it on the circuit around the core and air-gap we have NI = Ha la + Hc lc where Ha is the magnetic eld in the gap (which we assume to be constant)

6

(11.70)

347

Magnetic Material

N turns

H

Figure 11.10.: Magnetic circuit with an air-gap

la is the length of the air-gap Hc is the magnetic eld in the core (which also we assume to be constant) and lc is the length of the contour in the core The B eld in the two regions is Ba = 0 H a Bc = c H c (11.71)

where Ba and Bc are the ux densities in the air and core respectively and 0 and c are the permeabilties of air and core respectively From the conditions on the boundary between the core and air: the normal component of B is continuous Ba = Bc ( B) (11.72)

m Rc F Ra

348

so re-writing (11.70) NI =

F

B B la + lc 0 c la +BA 0 A

Ra

= BA

m

lc c A

Rc

F = m (Ra + Rc )

(11.73)

the same result may be obtained by applying the Magnetic Ohms law to the circuit shown in Figure 11.11 on page 348.

349

Part IV.

350

12.1. Chapter Goals

This chapter introduces the student to Maxwells equations in their nal form. In particular the topics covered are 1. Faradays law as it was historically obtained. 2. One of Maxwells equation as obtained from Faradays law. 3. The concept of the displacement current density as a fundamental requirement in Amperes law, and the Maxwells equation derived from the law. 4. A discussion of Maxwells equations in point and integral forms. 5. Time and frequency domain wave equations. 6. A discussion of phasors. 7. Time and frequency domain wave equationsdue to sinusoidally varying charges and currents.

1. Faradays law: emf generated = E = 2. Faradays law for an N turn coil: E = N 3. Maxwells equations in point form: D = v m t m t

E = B/t B = 0

H = J + D/t

351

4. Maxwells equations in integral form: D dS = E dl = v dV B dS t

B dS = 0

L

H dl =

J dS +

D dS t

= 1/ v2 2 E/t2 2 Ex 2 Ex 2 Ex + + 1/ v2 x2 y2 z2 2 E y x2 + 2 E y y2 + 2 E y z2 1/ v2 2 Ex =0 t2 2 E y t2 =0

2 Ez 2 Ez 2 Ez + + 1/ v2 x2 y2 z2

2 Ez =0 t2

E = jH

352

as sources: 2 A 2 A 2 A + + 2 k2 A = J x2 y2 z 2 Ax 2 Ax 2 Ax + + k 2 Ax x2 y2 z2 2 A y x2 + 2 A y y2 + 2 A y z2 k2 A y Jx J y Jz

= = =

2 Az 2 Az 2 Az + + k 2 Az x2 y2 z2

So far we have worked with static electromagnetic elds. We now turn our attention to time varying phenomena. In 1831 Michael Faraday and Joseph Henry independently came to the conclusion that when a changing magnetic eld was linked by a coil of wire connected to a galvanometer, it was found that the needle moved thus showing that an emf was generated in the coil. The conclusion was that a changing magnetic eld linking a circular wire creates an emf. Faradays law states The induced electromotive force, E, in any closed circuit is equal to the time rate of change of the magnetic ux through the circuit. Faradays1 law is a general statement of the fact that an emf is induced in a closed circuit where a time varying magnetic eld is present. B(R, t) B(R, t)

S L L

wire (b)

(a)

With this in mind we concentrate on the set-up shown in Figure 12.1(b) which consists of a wire in the shape of the closed curve L and having a voltmeter in the circuit. The voltmeter shows a voltage which is given by Faradays and

1 Michael

353

Lenzs laws. emf generated = E = m t (12.1)

Lenzs law2 gives us the direction of the current in the closed loop and the sign of the emf produced. That is Lenzs law The induced current ows in a direction such that the current opposes the change that induced it. The law of electromagnetic induction is used in the design of generators, where the motion of the conductors causes the ux enclosed by a closed loop to change and thereby cause an induced emf.

z y

x

B = az B0 f (t)

1 2

Figure 12.2.: Induced current in a time-varying magnetic eld.

EXAMPLE 12.1 For the conguration shown in Figure 12.2, the magnetic ux density is given by the equation a B tu(t) for t = 0 to 1 s z 0 B = az B0 f (t) = a B z 0 for t 1 s Find3 the induced voltage and direction of current in the loop of wire. The ux in the loop is equal to m (t) =

area of loop

where

2 Heinrich 3

354

The induced emf, E, is given by

The line integral of the induced E eld is in the anti-clockwise sense so as to make the unit vector of the surface in the az direction; since the line integral is negative, (= abB0 ) the potential at terminal 1 is higher than the potential at terminal 2, and therefore the current ows from 1 to 2. Notice that the induced current produces a B eld in the az direction, which opposes the change as per Lenzs law. It must be noted that when there is a coil of N turns instead of just a single turn then m E = N (12.2) t

m abB0 = E = 0 t

for t = 0 to 1 s for t 1 s

N

I a b az F B o

y x +

F d ax

origin moves to

S

Figure 12.3.: Faradays law applied to generators

Next we apply Faradays law to the case where the conductor itself is moving. EXAMPLE 12.2 Find the emf generated for the case of Figure 12.3 across the terminals marked + and .

355

Referring to Figure 12.3 we nd that we have a complicated arrangement where a single loop of wire is being turned in the anti-clockwise sense (as seen from the side ad) immersed in a uniform and constant magnetic eld, B = az B0 . The loop is being rotated at an angular speed = d/dt. To apply Faradays law, we need to compute the vector normal to the plane abcd. Two unit vectors, which are perpendicular to each other, and which lie on the plane are u1 = a y and u2 = az cos ax sin So the vector normal to the plane is n = a y u2 = ax cos + az sin The ux through this plane is m =

abcd

(12.3)

(12.4)

B dS =

abcd

(12.5)

where A is the area of the loop4 . The induced emf is therefore E= d d (B0 A sin ) = B0 A cos = B0 A cos (V) dt dt (12.6)

A current is induced in the loop which develops a torque (due to the two forces F in the wires) to turn the loop in the opposite direction of the angular motion. (Lenzs law) In the case which we have just discussed, the magnetic eld is constant and the conductors are moving, causing the ux linked by the loop to change, which causes the emf to be induced. Another way of looking at the same problem is to view it from the point of view of the Lorentz force. The conductors, which move in the constant magnetic eld, contain charges which feel a force given by F = qv B = qEm (12.7)

The vector Em can be looked upon as a motional electric eld. The charges feeling the force then set up currents. EXAMPLE 12.3 A circular loop of 10 turns of conducting wire of radius r = 5 cm and resistance R = 10 is placed in slowly varying a uniform magnetic eld. The magnetic eld makes an angle of 45 with respect to the direction of the surface of the loop. The magnetic eld magnitude is given by B = cos(2t) (T) What is the emf? What is the current?

4 As

356

Spark A T

B 12 V

The emf generated is given by E = N where N = 10, and m = BA cos 45 = cos(2t) (.05)2 cos 45 = 5.9722 103 cos(2t) or the emf generated is |E| = 10 2 5.9722 103 cos(2t) = 0.37524 cos(2t) (V) and the current is 37.524 cos(2t) mA. Did you know? The ignition system of a car uses the concepts developed above to create a spark in the spark plug. Referring to Figure 12.4, we see that when the contact A is opened the magnetic eld of the primary winding slumps to zero. Due to rate of change of ux, the secondary very high voltages (about 40,000 V) which in turn causes a spark in the spark plug. m t

Equation 12.1 can be used to derive an equation in vector calculus. The emf generated in a closed loop is (see Figure 12.1) emf generated = E =

L

E dl

(12.8)

357

The ux m through that loop is m =

S

B dS

(12.10)

We have studied Amperes law which applies to steady magnetic elds which are produced by steady currents. What about cases where the current is time varying? Take the case of a resistor-capacitor combination -series connectedto a battery as shown in Figure 12.5. In the (a) part of the gure we apply Amperes law: H = J (12.12) and we nd that the line integral shown indeed gives us the enclosed current, but when we apply Amperes to the (b) part of the gure (the bulging surface), we nd that the line integral gives us a zero result (there being no current between the plates of the capacitor). Though the line integral is the same as earlier, the two results are dierent. So obviously something is missing. Let us now approach Amperes from a dierent manner: taking the divergence of Equation 12.12, ( H) = J (12.13)

but div(curl(H))=0 for any vector and J = v /t, the continuity equation, therefore v = 0! t which is an absurd result! Let us therefore modify Equation 12.12 to include another current density, J . H = J + J (12.14)

358

H dl = I

dielectric

capacitor

line integral

V0

+ _

resistor

battery

(a) (D/t) dS

line integral

H dl =

resistor

V0

+

battery

(b)

Figure 12.5.: Setup to show how the displacement current comes into play

(12.16)

359

Rewriting Amperes law H = J + D/t (12.17)

the boxed part being the time dependent part. This is the last of Maxwells equations. EXAMPLE 12.4 Apply Maxwells Equation 12.17 to Figure 12.5. From circuit theory we know that the current through the circuit is I= V0 t/RC e R (12.18)

So the line integral in part (a) of the gure, where the enclosed surface is at, is H dl = I (12.19)

but in part (b) of the gure, the surface is bulging and no current is enclosed, but the line integral is the same as before. That is H dl is the same in both gures. Let us pause a moment and calculate the charge accumulated on the capacitor plates Q(t) = V0 t/RC t e dt = CV0 et/RC 0 = CV0 1 et/RC t=0 R

t

(12.20)

The charge produces a voltage across the plates V(t) = Q(t)/C = V0 1 et/RC which produces an electric eld E = V/d = V0 1 et/RC d (12.22) (12.21)

where d is the distance between the plates. The D eld is therefore D = E = and V0 1 et/RC d (12.23)

V0 1 t/RC e A (12.24) d RC where A is the area of the capacitor plates. Recalling that C = A/d the expression given above becomes (D/t) dS = (D/t) dS = V0 A t/RC V0 t/RC e = e = I(t)! Rd(A/d) R (12.25)

360

The current is the same as given in Equation 12.18. The analysis given above is rather crude, but the idea is correct. The fringing elds are taken into account by the bulging surface! So in summary in part (a) of the gure H dl = I and in part (b) H dl = (D/t) dS = I (12.27) (12.26)

12.6.1. Point form of the Equations

So far we have studied the time-independent electromagnetic elds. The relevant equations (that is Maxwells equations) including the time term are given below D = v E = B/t B = 0 H = J + D/t (12.28) (12.29) (12.30) (12.31)

In the second and fourth equations given above, there are time varying terms which are shown boxed. The rst equation is Gausss law, the last is Amperes law, the second equation is Faradays law and the third equation generated the vector potential and tells us that there are no magnetic charges. The two boxed terms are the time dependent terms which add to the static equations and complete all of Maxwells equations. EXAMPLE 12.5 If A is the vector potential and V is the scalar potential then another form of Faradays law is A V E= t Since or hence E= E = B A = t t A )=0 t

(E A V t

(since curl(grad(...))=0

361

We can write Maxwells equations in Integral form from the point form of the equations. 1. Integrating Equation 12.28 over a volume V and using the divergence theorem we get DdV = D dS = v dV (12.32)

where S is the surface enclosing the volume V. 2. Next we integrate Equation 12.29 over a surface S enclosed by the closed curve L and using Stokess theorem we get E dS =

L

E dl =

B dS t

(12.33)

the counter-clockwise direction on the line integral shows us the direction of the line integral. 3. Integrating the third equation: B = B dS = 0 (12.34)

In any situation where we have to communicate between a source and a receiver we usually use high frequency signals and either have wires or cables connecting the two, or we radiate energy from the one to the other. Figure 12.6 illustrates such a communication setup where antennas are used to transfer the signal. The gure displays a transmitter and a receiver. The transmitter is connected to a transmitting antenna through a transmission line, which sends information to a distant receiver in the form of radiated signals. A typical communication link consists of an information source, consisting of electrical signals, suitably modulatedfor example using AM, or FMsent via the antenna. The receiver has the same conguration as that of the transmitter. To properly design the transmission lines or the antennas, which are the parts of the equipment concerning the electromagnetic part, we need to understand the concepts of electromagnetic waves, whether these travel through transmission lines or through space itself. Surprisingly, both these phenomena are governed

362

Radiating Antenna

....

Transmission Line Receiver

Figure 12.6.: Typical communication setup with transmitting and receiving antennas

by the same kind of equations. In this chapter we will examine these equations in greater detail.

To begin with, since we are dealing with electromagnetic phenomena, we must take Maxwells equations into account. Remember that all electromagnetic phenomena are governed by Maxwells equations whether these phenomena be time-independent or time-dependent. Keeping this in mind let us then begin with the basic electromagnetic equations of Maxwell introduced and used frequently (and with which the reader must be familiar by now) and work with them to get wave equations. The equations, this time, are the complete equations which include the time dependent part. Maxwells equations are reproduced here for the convenience of the reader.

D = v

H = D/t + J

E = B/t B = 0

To recapitulate: D is the electric ux density, v is the volume charge density, and these two are connected by the rst equation, which is the divergence of D is equal to the volume charge density. E is the electric eld and B is the magnetic ux density. The curl of E is the negative of the rate of change of B. Next we have the equation, the divergence of B is always zero, and nally the curl of H, the magnetic eld, is equal to the displacement current density D/t plus J is the current density. The relations between D, E on one hand and B, H on the other are D = E (12.40)

363

B = H

(12.41)

where the constants of proportionality and are the permittivity and permeability respectively, of the medium in which the waves travel. Some of these vector elds may or may not be needed depending on the physical situation. To obtain the wave equation of the E eld, we take the curl of Equation 12.37: ( E) = (B/t) = ( B)/t (12.42)

In the second equality /t and were transposed since the two operators are independent of each other and their transposition gives the same result. Using the vector identity ( A) = ( A) ( ) A and applying it to the left hand side of the previous equation, we get ( E ) ( ) E = ( B ) t = (H ) (B = H ) t ( is a constant ) = ( H ) t = (D/t + J ) (Equation 12.39 ) t

(12.43)

In the second equation, of the above equations, we have used Equation 12.41. The last equation we have substituted Equation 12.39, in the right side of the equation. In a region where the wave is propagating, we assume that there are no charges or currents, v = 0, and J = 0. So that D = v = 0; which in turn implies that E = 0, (remember that D = E). This simplies the equation to the time domain wave equation (D/t + J) t ( E = 0 no free charges) ( ) E = (D/t + J) t (J = 0 no currents present) = (D/t) t 2 = 2 D t 2 = 2 (E) (D = E) t 2 E (12.44) = 2 t

( E) ( ) E =

364

Investigating the units of 1/ 1 Units of = 1 (H/m) (F/m) m

= HF m = ( s) ( s) m = s

which we denote by v, velocity. A simple calculation of 1/ 0 0 5 gives 2.998 108 m/s, which we recognise to be c, the velocity of light in free space. Using this notation the wave equation just derived becomes 2 E 1/ v2 2 E/t2 = 0 In Cartesian coordinates the operator 2 = = (12.45)

ax + a y + az ax + a y + az x y z x y z (12.46)

2 2 2 + 2+ 2 2 x y z

= 1/ v2 2 E/t2

This equation is a vector equation so it is actually three equations in Ex , E y and Ez . In Cartesian coordinates these equations are: 2 Ex 2 Ex 2 Ex + + 1/ v2 x2 y2 z2 2 E y x2 2 E y y2 2 E y z2 1/ v2 2 Ex =0 t2 2 E y t2 =0 (12.48)

(12.49)

2 Ez 2 Ez 2 Ez + + 1/ v2 x2 y2 z2

2 Ez =0 t2

(12.50)

5 Remember that

0 and 0 are the , the permeability and , the permittivity of free space,

365

EXAMPLE 12.6 Show that Ez (x, t) = Ex0 cos(t x) with v = / satises the wave equation, just derived. 2 Ez = Ex0 2 cos(t x) 1/ v2 2 Ez = (/v)2 Ex0 cos(t x) t2 = 2 Ex0 cos(t x)

and

substituting both these terms into the wave equation, we nd that it is satised. These partial dierential equations look quite complex. In order to gain a picture of the solution to these equations, we do two things. (a) We look at an analogous scalar equation, and (b) We reduce the number of independent variables. To get a scalar equation we rst concentrate on just one of these three equations: the very rst one in Ex and replace Ex by . To reduce the number of independent variables we set 2 /y2 and 2 /z2 , equal to zero, (which is the equivalent of saying that is not a function of either y or z) 2 x2 1/ v2 2 t2 =0 (12.51)

This is the scalar wave equation in one coordinate and time. We now go over to the body of work done in the past which deals with such an equation. Mathematicians have studied this equation and have found that any function whose argument is of the type x vt, i.e., (x vt), satises the equation. Let us do a quick computation. Let the variable a a (x, t) = x vt. Then a/t = v and a/x = 1 We know that Using the notation that = d/da. (x vt) = (a)

d a (a) = = 1 = x da x

366

2 = 2 x x x = x d a = da x = 1 = Similarly, d a (a) = = (v) = v t da t 2 = t t t2 (v ) (from the previous equation) = t d a = v da t = (v) (v) = v2 Substituting these relations into the scalar wave equation, 2 2 [1/v2] 2 = [1/v2] [v2 ] = 0 x2 t we nd that the scalar wave equation is satised. EXAMPLE 12.7 Let (x vt) = sin [ (2/) (x vt) ], where is the wavelength and v the velocity. Show that this function satises the wave equation. Taking the rst derivative = and, = 2 2 cos (x vt) 2 2 sin (x vt) (12.52)

(12.53)

Now, 2 /x2 and 2 /t2 are obtained in a similar manner 2 2 = cos (x vt) = x 2 2 = x2

2

sin

2 (x vt) =

367

2 2 = v cos (x vt) = v t 2 2 = v2 2 t

2

sin

2 (x vt) = v2

Substituting these terms into the scalar wave equation we nd that it is identically satised. The next question that we ask is that why is this a wave equation? The solution to the wave equation should be a wave. And what is a wave? A wave is a signal carrying energy which travels. Observe the signal (x vt), which is a solution of the wave equation. At t = 0 the signal has a spatial prole given by (x). Now observe the signal at time t. The spatial prole of the signal is now (x vt). Recall that if f (x) (some function) is compared to f (x x0), then the second function is shifted to the right by x0 . Applying this argument, (x vt) is (x) shifted to the right by the vt. But t, time, is increasing all the time, so (x) is continuously shifting to the right. In eect (x) is travelling to the right, a distance of vt in time t. And the velocity with which the wave travels is v. This situation is pictured in Figure 12.7, where a triangular pulse is travelling to the right with a velocity v. A wave travelling toward the right (the positive x direction) is a function of the type (x vt), as discussed, while a wave travelling in the -x direction is any function of the type (x + vt). This can be logically understood by applying the reasoning given earlier. The reader can substitute this function into the wave equation and verify that this function is also a solution to the wave equation.

vt (x) (x vt)

Initial pulse

Figure 12.7.: A travelling wave

EXAMPLE 12.8 Verify that the function (x + vt) is indeed a solution of the wave equation. Also verify that this is a wave travelling to the left. Substituting the function in 2 = v2 t2

368

Similarly

2 = x2 2 x2 [1/v2] 2 t2 =0

Since the argument of is x + vt where there is a + sign, the wave has to be travelling in the x direction. What about the wave equation for the magnetic eld? We nd that we can obtain it in a similar fashion to give 2 H 1/ v2 2 H/t2 = 0 (12.54)

EXERCISE 12.1 Obtain the wave equation for the magnetic eld from rst principles.

Generally, engineering applications require sources which produce sinusoidally varying voltages and currents, which in turn produce electromagnetic elds. In this section we take a look at Maxwells equations when the elds vary sinusoidally.

12.10.1. Phasors

Let us see how the wave equation appears when using sinusoidal quantities, and to do that let us rst take a look at a sinusoidally varying quantity such as (t) = 0 cos(t + ) (12.55)

(see Figure 12.8) Where is the frequency in rad/s, is a constant phase in radians and 0 is the amplitude of the signal. The tilde- signies that the entity is an sinusoidal signal. Notice that tilde looks like a sine wave. This equation can be written as (t) = 0 e j(t+) = 0 e j e jt = e jt (12.56)

Here the symbol means the real part of and is a complex number given a special namephasorcorresponding to (t) = 0 e j (12.57)

The relation between the sinusoidally varying quantity, in time domain, and the phasor is the Equation 12.56. The two can be symbolically shown connected

369

Complex plane

e jt = 0 e j(t+) t t

t=0

= 0 e j

b

0 cos( + t)

0 cos()

Figure 12.8.: Showing the relation between time-domain and frequency domain entities

as

Dierentiating Equation 12.55 with respect to t, time (t) = 0 sin(t + ) Is this result the same as dierentiating Equation 12.56? Let us dierentiate this equation d d (t) = 0 e j (t+) dt dt d ? 0 e j (t+) = dt = j0 e j (t+) = j0 cos (t + ) + j sin (t + ) = j0 cos (t + ) 0 sin (t + ) = 0 sin (t + ) (12.59)

Ph (t)

(12.58)

There is a question mark against the second equality. Can we interchange and d/dt? The answer is yes, because d/dt as calculated above and that calculated by the previous equation are indeed equal. Put down more clearly,

370

when (t) is a sinusoidal function, and is the corresponding phasor, d d = e jt dt dt d e jt = dt = je jt In the same manner we can show that if (12.60)

(t) dt = = = =

EXAMPLE 12.9 Find the phasors for (a) 10 cos(200t) (b) 20 sin(2106t) (c) 13 cos(105t+ 1.5) (d) 25 sin(100t + 45) (a) The term 10 cos(200t) = 10e(200t+0)

That is = 0. So the phasor is 10e j0 = 100 (b) The term 20 sin(2 106t) = 20 cos(/2 2 106t) = 20e j (210

6 t/2)

so the phasor is 20e j(/2) = 20(/2) (c) 13 cos(105 t + 1.5) is by inspection = 13e j(10 13e j1.5 . (d) The phasor is by inspection 25e j (45/2) .

5 t+1.5)

EXAMPLE 12.10 Convert the phasors (a) 1010 (b) 1e j15 to their real-time counterparts. (a) 1010 = 10e j10 . The conversion to time domain is 10e j(10+t) . The real-time function is therefore [10e j(10+t) ] = 10 cos(10 + t) (b) 1e j15 converts to 1 cos(15 + t).

371

Though we have considered the real part of the complex phasor, we could have worked with the imaginary part ((. . .)) as well. Applying this knowledge to sinusoidally varying elds, let E (R, t) be the elds in time but sinusoidally varying, and E (R) be the corresponding (complex) phasor, then from denition E (R, t) = Ee jt The other eld vectors are similarly written. A single Maxwells equation may written in the phasor notation as E(R, t) = E (R) e jt E (R) e jt B (R, t) t = B (R) e jt t B (R) e jt = t

(12.61)

E (R) e jt = jB (R) e jt

E = jB

(12.62)

In the last equation we have dropped e jt and this factor is understood to be there. Note that all of Maxwells equations remain the same except that /t is replaced by j.

EXERCISE 12.2 Show that Maxwells equations in phasor notation maybe written as in Equations 12.63 to 12.66. Every term in these equations is a phasor. Equations 12.36 to 12.39 become D = v E = jH B = 0 H = jE + J (12.63) (12.64) (12.65) (12.66)

372

We will make no distinction between the notation (that is for example E) used for sinusoidal and time-domain mathematical quantities since the discussion itself will clearly indicate whether the reference is toward the rst set or the second. To obtain the wave equation we take the curl of Equation 12.64: ( E) = j H = j( jE + J) Using the identity we have: ( A) = ( A) ( ) A (12.67)

( E) ( ) E = j( jE + J)

(12.68)

In a region where the wave is propagating, we assume that there are no charges or currents. Therefore there v = 0 (Which implies that E 0) and J = 0. ( ) E = j( j)E Furthermore, ( j)( j) = 2 . Therefore ( ) E = 2 E The phase velocity vp of the wave is related to and by v2 = 1/() p Hence, 2 = 2 1/ 1/ = 2 /v2 = (2 f )2 /( f )2 = (2/)2 = k2 p (12.70) (12.69)

Where is the wavelength of the wave in meters and k is the free-space propagation constant in rad/m. Using these relations, Equation 12.68 becomes ( ) E = k2 E (12.71)

Which is the wave equation. The wave equation for sinusoidal functions is called the Helmholtzs Equation. This equation is in fact three dierent equations (as earlier): 2 Ex 2 Ex 2 Ex + + + k2 E x = 0 x2 y2 z2 2 E y x2 + 2 E y y2 + 2 E y z2 + k2 E y = 0 (12.72)

(12.73)

2 Ez 2 Ez 2 Ez + + + k2 E z = 0 x2 y2 z2

(12.74)

373

Similarly, the frequency domain scalar wave equation in one variable is d2 + k2 = 0 dz2 (12.75)

where k = . In the above equation we have replaced Ex by in the rst of the three previous equations and /y, /x are equated to zero, that is is not a function of these coordinates. If we observe Equation 12.75 we note that its solution is of the sinusoidal type of function [i.e., sin(. . .) or cos(. . .) etc.] if = 0 sin kz then 2 = k2 z2

is satised. Let us look at a specic function of the type that is, let = Aejkz where A is a constant. This function satises the wave equation. If we pause a moment and further examine the nature of the full function A exp j(t kz) we realise that it must be a travelling wave. Since we are looking at a phasor, in real time it must be (z, t) = Ae j(tkz) = A cos (wt kz) Denoting the phase of this travelling wave by = t kz, the at some time t = t0 and position z = z0 , |t=t0 , x=x0 = 0 = t0 kz0

As in the case of the time domain scalar wave which we examined earlier, we allow a dierent set of values of z and t but changed in such a way that the phase remains the same, that is we travel with the phase 0 = (t0 + t) k(z0 + z) = t0 + t kz0 kz

= (t0 kz0 ) + t kz = 0 + t kz 0 = t kz

(12.76)

Notice that vp is the phase velocity since it is the velocity with which the phase travels. We can derive the magnetic eld Helmholtzs equation in a similar

374

(a)

(b)

way

( ) H = k2 H

(12.79)

This chapter introduced the student to Maxwells equations in their nal form. In particular the topics covered were 1. Faradays law as it was historically obtained and the of Maxwells equation as obtained from Faradays law by Maxwell. 2. The concept of the displacement current density as a fundamental requirement in Amperes law, and the Maxwells equation derived from the law. 3. A discussion of Maxwells equations in point and integral forms. 4. Introduction to the scalar and vector wave equations in time and frequency domains. 5. A discussion of phasors and their connection with the frequency domain wave equations. 6. The frequency domain wave equation, with sources which are sinusoidally varying charges and currents.

1. State Faradays law. 2. State Lenzs law. 3. Looking at Figure 12.9 which consists of a loop of metallic wire immersed in a constant B eld. Does a current ow in the metal hoop as shown in (a) part of the gure? Ans. No 4. Looking at Figure 12.9 does a current ow in the metal hoop as shown in (b) part of the gure? Ans. No 5. The magnet shown in Figure 12.10 is moved as shown in (a) part of the gure. In which direction does the current go in the upper part of the coil? Does the current go into the plane of the paper or come out of it? Ans. Into the plane of the paper.

375

coil of wire

S N N S

(a)

(b)

6. The magnet shown in Figure 12.10 is moved as shown in (b) part of the gure. In which direction does the current go in the upper part of the coil? Does the current go into the plane of the paper or come out of it? Ans. Into the plane of the paper. 7. Does (x/v t) satisfy the wave equation? v is the velocity of the wave. Ans. Yes. 8. E/t is the displacement current. Ans. No. 9. What is the angle of the phasors (a) j (b) 1 (c) 1 + j1? Ans. (a) /2 (b) 0 (c) /4 10. In which quadrant is 1 j1 lie? What is its angle? Ans. 4th quadrant; /4.

12.14. Problems

1. In Figure 12.2 the B eld is given by B = t2 u(t)az (Wb/m2 ) = 4 (Wb/m )

2

t 4 (s)

0 t 2 (s)

Find the induced emf V12 (t) generated. 2. In Figure 12.2 the B eld is given by B = cos(100t)az (Wb/m2 ) Find the induced emf V12 (t) generated. 3. In Figure 12.2 the B eld is given by B = cos(100t ky)a y (Wb/m2 ) where k is a constant. Find the induced emf V12 (t) generated. 4. In Figure 12.2 the B eld is given by B = cos(100t kz)ax (Wb/m2 ) where k is a constant. Find the induced emf V12 (t) generated. 5. In Figure 12.2 the B eld is given by B = cos(100t kz)xy(ax + a y + az ) (Wb/m2 )

376

z y

a v, velocity b L

x

Figure 12.11.: Moving slider on rails

where k is a constant and a = 1 and b = 2. Find the induced emf V12 (t) generated. 6. The magnetic eld H(x, y, z, t) of a plane wave is H0 sin(t kx)az and is incident on the antenna as shown in Figure 12.2. Find the magnitude of the emf |V(t)| between terminal 1 and 2. Show that it is: H 0 0 2 sin

ak 2

cos (t)

7. A slider moves on rails with a velocity v as shown in Figure 12.11. The whole arrangement is immersed in a constant magnetic eld B with a direction going into the plane of the paper as shown. If R2 = 0 show that the induced emf Vab is Vab = BvL 8. In Figure 12.11, If B = 0.1 (T), the position of the slider in the y direction is given by y(t) = 0.1t (m), and L = 0.1 (m), nd the emf generated in the loop. 9. In Problem 8 if R1 10 and R2 = 20 nd the voltage drops across the two resistors, and the current. 10. In Figure 12.11, If B = 0.1 (T), the position of the slider in the y direction is given by y(t) = 0.1t + .05t2 (m), and L = 0.1 (m), if R2 = 0, and R1 = 1 , nd the emf generated and the current in the 1 resistor. 11. In Figure 12.11, If B(x, y) = 0.1xy (T), the position of the slider in the y direction is given by y(t) = 0.1t (m), and L = 0.1 (m), if R2 = 0, and R1 = 1 , nd the emf generated and the current in the 1 resistor and when the y-position of the slider is 1 m. 12. From Maxwells equation in integral form Equation 12.33 show how both Faradays law as well as Lenzs laws can be obtained. 13. In a region of free space = 0 and = 0 , the electric eld is a constant, E = E0 az . What is the displacement current density?

377

14. In a region of free space = 0 and = 0 , the electric eld is, E = E0 cos(110t kx)az . What is the displacement current density? What is the conduction current density? 15. In a region = 106 , = 105 , = 105 the electric eld is, E = E0 cos(110t kx)az . What is the displacement current density? What is the conduction current density? 16. By using the arguments given in Section 12.5 show that for a parallel plate capacitor with a dielectric lling of , the input current to the capacitor is equal to the displacement current plus the conduction current within the capacitor.

378

When we set up a communication link between a transmitter and receiver without the help of cables we usually set up two antennas, a transmitting antenna and a receiving antenna. The transmitting antenna transmits electromagnetic waves which are essentially waves whose nature is that of light waves but at much, much lower frequency. These waves are generally sinusoidal, described by frequency and the velocity of propagation, which in turn depends on the medium in which the waves propagate. Thus if the waves propagate in a dielectric, then the medium is described by its permittivity, = r 0 . If the medium is a conducting, then the conductivity, , comes into play, and so on. In this chapter we concentrate on the fundamental properties of electromagnetic waves as they propagate in various types of media, and their interaction with matter.

In the very beginning we consider the simplest of waves, namely, the uniform plane wave. Let a sinusoidally time-varying wave travel in the +z direction but with no variation in either x or y directions. Figure 13.1 depicts how the wave travels. The shaded region shows a single phase front, perpendicular to the direction of propagation.

x

E

...

...

Direction of propagation z

H

...

...

379

The procedure which we now outline is standard for most wave phenomena. Since we are considering a wave, the electric and magnetic elds must satisfy the wave equation. And since their nature is oscillatory, having a exp( jt) time dependence, the elds must satisfy Helmholtzs equation1 . Helmholtzs equation for the electric eld is Equation 12.71. 2 E + k2 E = 0 where k =

If we recall the results of the last chapter, a travelling wave must have a functional dependence, E(R, t) = E0 e j(tz) (13.1)

where is the propagation constant of the wave and is the frequency of oscillation. Here may or may not be equal to k. In the case of a uniform plane wave, E0 is a constant but real phasor, E0 = ax Ex0 + a yE y0 + az Ez0 (Ex0 , E y0 , Ez0 are real and constant) (13.2)

?

?

E0 e jt 2 ejz = k2 E0 e j(tz) 2 E = k2 E Which is satised if = k. The term exp{ j(t kz)} ensures that the wave is a travelling wave, travelling in the +z, direction. The phase of the wave at time t0 is given by t0 kz = constant = K which is the equation of a plane in three dimensions which is parallel to the z = 0 plane. z = (t0 K)/k = Another constant = K1 Now working only with the phasor E = E0 ejkz

(13.3)

1 See

380

The sinusoidal elds in time are Ee jt = E0 ejkz e jt = E0 e j(tkz) = E0 cos (t kz) (13.4)

The term E0 can be pulled out of the bracket since it is real and constant. If this electric eld is to represent an actual wave, it must satisfy Maxwells equations. The rst Maxwell equation which has to be satised is E = v

But we know that the wave is travelling in a region free of charges. So v must be zero. E = 0

(13.5)

In other words there is no electric eld in the direction of propagation; there are only transverse components. E = ax Ex0 ejkz + a y E y0 ejkz (13.6)

The second Maxwell equation which the electric and magnetic elds have to satisfy is E = jH (13.7) Which is Equation 12.64 on page 372. Writing out the various components starting with the x component ( E)x = jHx

Ez E y = jHx y z 0 + jkE y0ejkz = jHx Similarly we can get the other components ( E) y = Ex Ez = jkEx0 ejz = jH y z x E y Ex = 0 = jHz ( E)z = x y

Straightaway, from the last equation, we get Hz = 0 or the magnetic eld is also

381

transverse to the direction of propagation. Figure 13.1 clearly shows how both the electric and magnetic elds lie on z = constant planes. From the above equations, the magnetic eld components are jHx = jE y0 ejkz j E y0 Hx = j = E y0 (13.8)

2

Units of

= = = = =

Since

Z is called the characteristic impedance of the medium. For air or vacuum, a quick calculation of the characteristic impedance, Z Z0 gives 2 Z0 =

2

0 377 0

0 /0 =

(13.14)

4 362 100 = 120

382

The magnitude of the magnetic eld is H= = 1 Z E = Z

2 Hx + H 2 y

E2 + E2 y x (13.15)

Taking a dierent tack, we calculate E H for the uniform plane wave. E H = ax E y Hz Ez H y + a y (Ez Hx Hz Ex ) + az Ex H y E y Hx = a z Ex H y E y H x = a z Ex = az (Ez and Hz are both zero) (From the previous set of equations) E y Ex Ey Z Z

1 2 E + E2 y Z x

|E H| =

It is important to note that the direction of E H is the direction of propagation. To compute the angle, , between E and H we must compute the dot product between the two. Since E H = EH cos Using Equations 13.11 and 13.12 E H = ax Ex0 ejkz + a yE y0 ejkz ax Hx0 ejkz + a yH y0 ejkz ej2kz ax Ex0 + a yE y0 ax E y0 + a y Ex0 Z ej2kz = Ex0E y0 + E y0 Ex0 Z =0 =

(13.17)

Now E 0 and H 0, therefore cos must be zero. In other words, the angle from E to H is /2. We reach an important conclusion: Not only are the electric and magnetic eld perpendicular to the direction of propagation, but they are perpendicular to each other; the vector E H is directed toward the propagation direction. Figure 13.2 depicts the advance of the plane wave where the electric eld is drawn as the solid line while the magnetic eld is the dotted line. Notice that the magnetic eld is always in phase with the electric eld; that is when the electric eld increases, the magnetic eld increases, and when the electric eld

383

decreases the magnetic eld does the same, keeping the ratio of the electric to the magnetic eld constant (=Z, the characteristic impedance).

E field H field

Direction of propagation

1 0

Emax 1 1

1 Hmax

EXAMPLE 13.1 Find the H eld of of an plane wave whose E vector is given by E = (10az + 20a y) cos(2 107t kx) in air. Find the value of k, and the wavelength . From inspection of the E eld it is clear that = 2 107 t(= 2 f ) where f is the frequency of the wave. So f = 107 Hz. Since the velocity c = 3 108 m/s the wavelength is = c/ f = 3 108/107 = 30 m from we can calculate k by k = 2/ = 0.20944 rad/m Again from inspection it is clear that the wave is travelling in the +x direction. So 1 1 ax E = (10a y + 20az) cos(2 107t kx) A/m H= Z0 377 EXAMPLE 13.2 A plane wave caries a power density of 1 MW/m2 . Find the magnitude of the electric and magnetic eld vectors. The power density S is given by S = E2 /Z0 = 106 W/m2 so E= 106 377 = 1.9416 104 V/m

384

and H = E/Z0 = 51.501 A/m Electromagnetic waves in actual use for dierent purposes are characterised by their frequency f and wavelength, . These bands of frequencies have been given names by engineers, and there is an agreement all over the world about their use. Figure 13.3 shows the frequency bands, their names and their uses.

Lumped Shortwave Radio Voice frequency Distrbuted Optical

AM Radio

300 KHz

300 GHz

300 THz

30 MHz

30 KHz

30 THz

300 Hz

3 MHz

3 KHz

3 THz

Micro Waves

I.R.

30300 KHz LF

0.33 MHz MF

330 MHz HF

30300 Hz SLF

330 Hz ELF

In the gure the frequency spectrum is depicted on a logarithmic scale from 3 Hz to 3 PHz. (Read as petahertz). At the top the approximate analytical technique is outlined: for example the lumped element approach is used from DC to a few hundred megahertz; the distributed element approach is used from a few hundred megahertz to about a few hundred gigahertz, and so on. Below the logarithmic spectrum is the rough demarcation of the frequency bands: radio waves, microwaves, infrared and visible and beyond. At the bottom of the gure, the designated frequency bands are named: VLF, very low frequency; LF, low frequency; MF, medium frequency; HF, high frequency; VHF, very high frequency; UHF, ultra high frequency and EHF extremely high frequency and so on. Apart from the radio frequency bands there is yet another set of designations which is primarily due to the US military, but now adopted worldwide. These bands are shown in Table 13.1

385

3 PHz

30 Hz

3 Hz

Band L S C X Ku K Ka V W mm Frequency 1-2 GHz 2-4 GHz 4-8 GHz 8-12 GHz 12-18 GHz 18-27 GHz 27-40 GHz 40-75 GHz 75- 110 GHz 110-300 GHz

The electric eld in the case just discussed always moves on a plane as the eld advances in the direction of propagation. Examination of Figure 13.2 illustrates this point. The gure shows the electric eld in the x direction while the magnetic eld moves in phase in the y direction, and a x Ex a y H y = a z Ex H y (13.18)

Let us examine the solution to the wave equation, Equation 13.6 under dierent cases and conditions. So far we have stipulated that the amplitudes are real and we found that the electric eld moved along a plane as the wave advanced (Again study Figure 13.2). Let us now consider the case when one amplitude of the set [Ex , E y ] is complex, and the other one real. Taking a specic case Ex0 = A0 e j0 E y0 = A0 e E0 = ax Ex0 + a y E y0 Where A0 is of course real. The sinusoidal elds in time are E(R, t) = E0 e j(tkz)

j(/2)

= A0 ax e j(tkz) + a ye j(tkz+/2)

= ax Ex0 + a y E y0 e j(tkz)

386

The magnitude of the eld in time is E (R, t) = A2 cos2 (t kz) + sin2 (t kz) 0

= A0

(13.23)

Which is constant. How do we interpret this result? Let us analyse this situation in greater detail. Ey

t=3T/4

t = 3/2

E(t = 3T/4)

t=T t=T/2

t =

E(t = T/2)

t=0

t = 2 Ex t = 0

t=T/4

t = /2

Figure 13.4.: Figure illustrating left circular polarisation

The original electric eld in time is E = a x Ex + a y E y = A0 ax cos(t kz) a y sin(t kz) (13.24)

We get rid of the z coordinate by setting it to zero. By setting z to zero we are analysing the behaviour of the electric eld on the plane z = 0. E(x, y, z = 0, t) = A0 ax cos(t) a y sin(t) (13.25)

This seems simpler. Let us now proceed to calculate the electric eld E at dierent times: at t = 0, t = T/4, t = T/2, and t = 3T/4, where T is the time period of one cycle. Since = 2/T, t = 0, /2, , and 3/4 at these times. E E E E = A0 a x = A0 ax = A0 a y = A0 a y (At t = 0) (At t = T/4) (At t = T/2) (At t = 3T/4)

We notice that the vector rotates in the clockwise direction, with the tip of the vector moving on a circle with radius A0 . Figure 13.4 illustrates this point.

387

E field helix E field on the xy plane Direction of Propagation

a z

a y 1 x d d o

c b c

A0 A0

Figure 13.5.: Figure showing the advance of the wave in a left circular polarisation plane wave

The next gure Figure 13.5 shows how the wave advances helically in a polarised wave. The gure shows the wave at time t = T. The tip of the electric eld vector at t = 0 is at point a; at t = T/4 it is at b. In this way it traverses a full circle in time T, going through points c and d. In time T dierent parts of the wave advances by varying amounts. The electric eld at point a advances to a which is one wavelength away. The electric eld at point b advances to b three quarters of a wavelength away, and so on. The electric eld vectors all lie parallel to the xy plane but all their tips lie on a helix. On the xy plane the electric eld traces a clockwise circle. This type of polarisation is called left circular polarisation because as the left hand is held in the form of a st, with the thumb extended, the ngers curl in the direction of motion of the electric eld while the thumb points in the direction of the propagation of the wave. Consider the case of a right hand circularly polarised wave. For such a wave, advancing in the +z direction, the electric eld rotates in the anti-clockwise direction on any plane parallel to the xy plane and the wave travels in the direction of the thumb in accordance with the right hand thumb rule. The electric eld then can be written as Ex0 = A0 e j0 E y0 = A0 ej(/2) E0 = ax Ex0 + a y E y0 E = E0 ejz (13.26)

388

The sinusoidal time dependent eld is E = Ee jt

= A0 ax e j0 + a yej(/2) e j(tkz)

) 2 (13.27)

choosing z = 0 as the plane on which the electric eld is evaluated E(z = 0, t) = A0 ax cos(t) + a y sin(t) (13.28)

This is vector which is rotating in the counter-clockwise direction and obeys the right hand thumb rule.

2 1 0 1 2 2 1 0 1 2 (a) x=cos(t), y=2*cos(t) 2 1 0 1 2 2 1 0 1 2 (b) x=cos(t), y=2*cos(t+ )

2 1 0 1 2 2 1 0 1 2

2 1 0 1 2 2 1 0 1 2

We now have studied two types of polarisations, linear and circular polarisation. Let us take a look at the most general type of polarisation, namely

389

elliptical polarisation. Let E = a x Ex + a y E y = ax Ex0 cos (t kz) + a y E y0 cos (t kz + )

(13.29)

where Ex0 and E y0 are the wave amplitudes in the x and y directions and is the phase angle with which E y leads Ex . Choosing the plane z = 0 to evaluate the polarisation E(z = 0) = ax Ex0 cos (t) + a y E y0 cos (t + ) = ax Ex0 cos t + a yE y0 (cos t cos sin t sin ) Ex Ex0 1 Ex Ex0

2

(13.30)

cos t = sin t = Ey E y0

(13.31) (13.32)

2

sin

(13.33)

2

2 sin

(13.34) (13.35)

Let us discuss these equations. Case 1 . = 0 or ; Ex0 and E y0 can be anything. This case gives us linear polarisation. To illustrate this statement we use Equation 13.34 since Equation 13.35 is un-

390

suitable. Letting sin = 0 we have Ey E 2 2 E + Ex cos 2 y Ex cos = 0 E E Ex0 y0 y0 x0 2 Ey Ex E E cos = 0 y0 x0 Ey Ex = cos E y0 Ex0 E y0 E for = 0 Ex0 x y = E Ey0 E for = x E

x0

(13.36)

which is a straight line between the variables E y and Ex . This shown in Figures 13.6 (a) and (b). Case 2. = /2, Ex0 and E y0 can be anything. This gives elliptical polarisation with with the major and minor axes of the ellipse along the coordinate axes. We use Equation 13.35 with cos = 0 and sin = 1 Ey E 2 + Ex Ex0 y0

2

=1

(13.37)

which is the equation of an ellipse with major and minor axis along the x and y directions respectively. This is shown in Figures 13.6 (c) and (d). If E y0 = Ex0 then we obviously get circular polarisation. Case 3. , E y0 , Ex0 all of which can be anything. In this general case, we have elliptical polarisation, with the major axis of the ellipse tilted in any general direction. The case for = /6, E y0 = 2 and Ex0 = 1.5 is shown in Figure 13.7.

In conducting media, as the wave propagates currents are set up. These currents circulate in the medium and produce heat. Thus energy carried by the wave must necessarily be diminished. We already know that where an electric eld is present in a medium with conductivity , there the current density generated at a point in the medium is related to the electric eld through the equation J = E

391

1.5*cos(t), 2*cos(t+pi/6) 2 1.5 1 0.5

Ey

= jE + E j E j E

= j 1 = jC E

Where C is the complex dielectric constant which is so called because of the above set of equations, where C is introduced to simplify things. The conducting medium can be modelled as a dielectric but one with a complex permittivity. C = j = j

(13.42) (13.43)

The ratio /() is dimensionless (the same as r ). This ratio is called the dissipation factor, D 1 for a good dielectric (13.44) D= 1 for a good conductor 392

Generally, the dissipation factor for lossy dielectrics is also called its loss tangent. (13.45) = D= Why it is so called may be seen from examining Figure 13.8.

D = tan

/ or or

Figure 13.8.: Loss tangent for a dielectric

2

2 C = 2 2 + j2 = 2 2 + j2 2 j Equating the real and imaginary parts, and solving the resulting equations we get = R =

2 + 1

= 2 2 + j2

1+ 2

(13.46)

= I =

1+ 2

2 1

(13.47)

What is the meaning of a complex propagation constant C ? Let us examine a wave travelling in the z direction, with the electric eld in the x direction. The

393

wave in complex notation is then Ex = Ex0 ejC z = E ej(j)z

x0

= Ex0 ez ejz

(13.48)

Air

Skin Depth

Figure 13.9.: Prole of a wave propagating from air to a conducting medium

Using sines and cosines, the wave in the real world is Ex = Ex0 ez ejz e jt = Ex0 ez e j(tz) = Ex0 ez cos(t z) (13.49)

Observing the previous equation we can see that as the wave progresses in the z direction, the electric eld is still oscillating (the cosine factor), but its amplitude decays (the exponential factor). This is shown in Figure 13.9. The gure shows a snapshot at some time instant of a wave incident from air into a conductive medium. As the wave progresses into the medium, the following happens 1. The wavelength of the wave decreases. This is so because 1+ 2

2 + 1

2 = Cond Med = C

2 > (= )

where C is the wavelength in the conducting medium. 2. The wave amplitude decays as exp(z). The term exp(z) forms an envelope of the cosine term. Concentrating on the decaying term, the amplitude of the wave falls by

394

Ex0 exp(1) in a distance given by zskin depth = = 1/

Ex0 e = Ex0 e1 = 0.3679Ex0 is called the skin depth and depending on the value of , the skin depth varies. To obtain the value of the magnetic eld. Using the notation x /x etc., H= 1 E j 1 az x E y y Ex + ax y Ez z E y + a y (z Ex xEz ) = j 1 = az x E y yEx +ax y Ez z E y +a y (z Ex x Ez ) j z =j , Ez =0

x =0, y =0 y =0, E y =0

C

In the rst factor the E eld is not a function of either x or y. The same argument applies to the next factor taking into account the additional fact that E y = 0. In the last bracket, dierentiation with respect to z is the same as multiplication by jC . Therefore H= 1 a y jC Ex j C Ex = ay C Ex (13.50)

= ay

The meaning of a complex characteristic impedance is that the electric and magnetic elds are not in phase with each other. Thus, in a propagating wave, propagating in the n direction if the electric eld is E, the magnetic eld is (1/Z)n E. Let us apply what we have learnt to dierent materials.

395

Low conductivity materials are generally lossy dielectrics with a small dissipation factor. Consider such a dielectric. In the following equations we have used the Taylor series expansions, where x 1 x 1+x 1+ 2 1 1x 1+x (13.51) (13.52)

= =

1+ 2 1+ 1 2 2 1+

2 + 1

1 2 4 1 2 1 + 8

(13.53)

= = 2

1+ 2 1+ 1 2 2

2 1

(13.54)

396

The characteristic impedance Z= = = C 1 1 j(/) (applying the taylor series expansion)

1+ j 2

Applying these results to a case of a wave travelling in a dielectric with a loss tangent of 0.05 and a dielectric constant of 2. Let the frequency of the wave be 10 GHz. The loss tangent is 0.05 1. The permeability 0 ; the permittivity = r 0 = 20 . = 2 f = 2 106 rad/s. = 0.05 1 0 = 0 = 298.3412 = 0 1 + 1 8

2

= 298.4345

397

For the case of high conductivity materials like metals, 1 and = 2 1+ 2

2 + 1

+1

(since

1)

2 2

(since

1) (13.55)

1+ 2 2

2 1

2 = 2

(13.56)

398

The characteristic impedance is given by Z= = = 0 C 0 0 j (/) 0 0 1 1 j 0 0

0

Z0 = Z0

j 1+ j 2

Where Z0 is the characteristic impedance of free space( 377 ). Taking the example of copper with a conductivity = 5.814 107 /m; 0 at a frequency of f = 1 MHz, ( = 6.28 106 rad/s) /(0 ) = 1.045 1012 1 and so = The skin depth = 1/ = 6.6 105 m The wave decays to e1 = 0.369 in a distance of .066 mm . The wavelength in air is c 3 108 air = = = 300 m f 1 106 2 2 = 15150 = 4.15 104 m! = 15150 2

A typical example of the skin depth as a function of frequency is shown in Figure 13.10. As we can see from the gure, the skin depth rapidly decreases

399

Skin depth for copper as a function of frequency 0.1

0.001

0.0001

1e05

10

100

100000

1e+06

1e+07

with increasing frequency, and at dc, (zero frequency) the skin depth becomes innity. On the other hand as the frequency is increased into the GHz range, the skin depth goes into the m range. It is for this reason that in many microwave components, the conducting surface is coated with a very thin layer of gold (in which the wave penetrates) to minimise the energy dissipated. In high conductivity materials the wavelength and skin depth are related by material = 2material The characteristic impedance for copper is given by Z Z0 1+ j 0 2 = 377 145 9.57 1013 = 3.61 101045 (13.57)

When we consider electromagnetic problems we need to look at the conditions of the electromagnetic elds at the boundary between two media. Principally what we are interested in is: what are the relations between the electromagnetic eld components in the two media which are (a) tangential to the surface separating the two media? And (b) normal to this surface?

400

z

1 , 1 , 1

y x

2 , 2 , 2

1 , 1 , 1 2 , 2 , 2

z a d b

n

y x Medium Boundary

2 , 2 , 2

c

t2 t1

n 1 , 1 , 1

Figure 13.11.: The Behaviour of electromagnetic elds near a boundary consisting of a change of medium

There is one special case which we always have to keep in mind: the case of a perfect electric conductor, the conductivity, . In this case, no time varying elds can exist inside a conductor: E and H will both be zero. In this case there will be both a surface charge, s , as well as a surface current, Js , at the boundary of the two media. To investigate these problems we take a look at the case where electromagnetic elds are present inside and outside a body with permittivity and permeability (2 , 2 , 2 ). This we call region 2. The body is immersed in a region with permittivity and permeability (1 , 1 , 1 ), called region 1. This conguration is shown in Figure 13.11. In the (a) part of the gure the macro level diagram is shown where a small region labelled R is shown. This region is shown blown up in parts (b) and (c) of the gure. The elds in regions 1 and 2 must satisfy Maxwells equations. D = v E = B = 0 H =

B t (13.58)

D +J t

These are the dierential form of the equations. The integral forms are

401

DdV = E dS = BdV = H dS =

D dS = E dl = B dS = 0 H dl =

v dV B dS t D dS + t

J dS

(13.59)

We rst take a look at the rst of the above equations (Maxwells equations in integral form). The equation is applied to the pillbox shown in part (c) of the gure. The mathematical description of the pillbox is height of pillbox 0 Area of the top and bottom A Applying Gausss law to this pill-box with the knowledge that there may or may not be accumulated charge on the interface. There is no surface charge (s = 0) if 2 is nite. If we have a perfect conductor then 2 then there will be a surface charge, s . Writing Gausss law in the most general case (that is if a surface charge exists) D1 az A + D2 (az ) A = s A Dz1 Dz2 = s (13.60) (13.61)

in other words, the normal component of the ux density is dis-continuous across a dielectric-dielectric boundary by the amount of the surface charge density. Writing the above equation in a more general form Dn1 Dn2 = s n (D1 D2 ) = s (13.62)

If there is no surface charge, (s = 0), then the right hand side is replaced by zero and in that case, the normal component of the D eld is continuous. Let us take the case where the second medium is a pec (perfect electric conductor). Then there will be no elds there and Dn2 = 0 but there will be a surface charge, s . So Dn1 = s (13.63)

Now considering the the second of the Maxwells equations in integral form (the second of the Equation set 13.59) to part (b) of Figure 13.11. In this case length (b c) = length (d a) 0

Applying the line integral to the loop a-b-c-d in the anti-clockwise direction, E1 ax |x| + E2 (ax ) |x| = 0 where E1 and E2 are the electric elds in regions 1 and 2 respectively. The right hand side of the integral B dS t B |dS| = 0 t

since one side enclosing the area is vanishingly small, the area is vanishingly small, that is, |dS| x length (b c) = 0 (13.64) is zero. The condition on the electric eld may then be summarised as Et1 Et2 = 0

(13.65)

The subscript t is used to signify the tangential component. Et1,2 are the tangential components of the electric eld next to the boundary but in media 1 and 2 respectively. Writing this in a more compact form n (E1 E2 ) = 0 Notice that the normal goes from medium 2 to medium 1. Let us take the case where the second medium is a pec (perfect electric conductor). Then there will be no elds there and Et2 = 0. So Et1 = 0 (13.67) (13.66)

we apply the same arguments as for the D vector and get Bn1 Bn2 = 0 Or n (B1 B2 ) = 0 (13.68)

(13.69)

Let us take the case where the second medium is a pec (perfect electric conductor). Then there will be no elds there and Bn2 = 0. So Bn1 = 0 (13.70)

403

Finally we treat the last Maxwell equation D dS + t

H dS =

H dl =

J dS

Integrating over the small loop abcd in the anticlockwise sense Hx1 |x| Hx2 |x| = D y t A + J yA + Jsy |x|

Here |x| is the longer side of the loop and A is the area of the loop. J y is the y-directed volume current while Jsy is the surface current just at the boundary. The line integrals of H over the shorter sides of the loop are equated to zero since there lengths are virtually zero. Using the same argument, A is also considered to be zero. Therefore Hx1 |x| Hx2 |x| = Jsy |x| or Ht1 1 Ht1 2 = Jst2 (13.71) t1 , t2 and n form a right-handed orthogonal coordinate system. t1 , t2 and n are akin to ax , a y and az . Scrutinising the gure, Jst2 is the y- (or t2 -) directed surface current. Writing equation in vector notation az (H1 H2 ) = Js or n (H1 H2 ) = Js (13.73) (13.72)

Let us take the case where the second medium is a pec (perfect electric conductor). Then there will be no elds there and H2 = 0. But in this case there will be the surface current Js . So n H1 = Js (13.74) note that n will be from the pec into medium 1.

13.5.1. Reection from a Metal Surface

Electromagnetic waves having the same nature as light and therefore reect from metallic objects and suer refraction in the presence of dielectrics. The simplest case is one where a plane wave falls on a metal or dielectric plane surface. Let us consider rst the case of a uniform plane wave obliquely incident on a air metal boundary. For the sake of convenience the metal may be considered to be a perfect one with . The wave may approach the 404

boundary in either of two congurations: where the electric eld is perpendicular to the plane of incidence or when it is parallel to the plane of incidence. Both congurations are shown in Figure 13.12. We need to look at a plane wave travelling in a general direction whose unit vector is n. The equation of a plane is f (x, y, z) = ax + by + cz = constant = p then the normal3 to this plane is f = aax + ba y + caz and the unit normal to the plane is aax + ba y + caz n= a2 + b2 + c2 (13.77) (13.76) (13.75)

note that a/ a2 + b2 + c2 is the direction cosine of the vector n in the x-direction, etc. The equation of the plane, Equation 13.75, may now be written as p nr = a2 + b2 + c2 (13.78)

p/ a2 + b2 + c2 is the shortest distance of the plane from the origin. With this introduction, the equation of a uniform plane wave travelling in the n direction is E = E0 ejknr (13.79) where E0 is perpendicular to n. n r describes dierent planes all parallel to to

3

405

each other. We can see for instance if the wave travels in the a y direction then a y r = y, then E = E0 ejky and so on. Referring now to Figure 13.12(a), ni = az cos i + a y sin i nr = az cos i + a y sin i E = a E ejk( y sin i z cos i )

i x x0

where R, the reection coecient is to be determined. At the boundary of the metal, the tangential electric eld must vanish as per the previous section. (Ei + Er )z=0 = 0 ax Ex0 ejk( y sin i z cos i ) + ax REx0 ejk( y sin i +z cos i ) =0

z=0

or or

R = 1 the reected electric eld is therefore for perpendicular incidence, Er = ax Ex0 ejk( y sin i +z cos i )

(13.87)

We now analyse the case of parallel incidence where the magnetic eld is perpendicular to the plane of incidence as shown in Figure 13.12(b). The electric eld is given by E0i = Z1 H0i ni

where Z1 = /1 , (377 ), the intrinsic impedance of free space. We now proceed to apply the boundary condition at the metal boundary: the tangential electric eld is zero at the boundary. E y0r = Z0 Hx0i cos i so E0r = Z0 Hx0i a y cos i + az sin i (13.90) (13.89)

406

z Medium 1 y Medium 2

and the reected magnetic eld is nr E0r /Z0 H0r = az cos i + a y sin i Hx0i a y cos i + az sin i = Hx0i cos2 i + sin2 i ax = Hx0i ax (13.91)

We notice that the reected magnetic eld remains as it is at z = 0! The incident and reected magnetic elds are Hi = ax Hx0 ejk( y sin i z cosi ) H = a H ejk( y sin i +z cosi )

r x x0

(13.92) (13.93)

We now consider a wave obliquely incident on a dielectric surface as shown in Figure 13.13. First let us consider the case where the electric eld is perpendicular to the plane of incidence as shown in the (a) part of the gure. The incident, reected and transmitted electric elds are given by Ei = ax Ex0i ejk( y sin i z cosi ) E = a E ejk( y sin i +z cos i )

r x x0r

Et = ax Ex0t e

407

at the dielectric interface, (z = 0) the tangential elds must be continuous, so we equate the sum of the elds in Region 1 to the elds in Region 2. Ex0i + Ex0r = Ex0t (13.97)

similarly, the incident, reected and transmitted magnetic elds are given by Hm = 1 nm Em Zj (13.98)

where Z j = 0 / j characterises the medium (1 or 2) and m characterises the wave: incident, transmitted or reected waves. So Ex0i a y cos i + az sin i ejk( y sin i z cos i ) Z1 Ex0r a y cos i + az sin i ejk( y sin i +z cos i ) Hr = Z1 Ex0t Ht = a y cos t + az sin t ejk( y sin t z cos t ) Z2 Hi = (13.99) (13.100) (13.101)

at the dielectric interface, the tangential magnetic elds must also be continuous (Ex0i Ex0r) cos i Ex0t cos t = Z1 Z2 using Equation 13.97, (Ex0i Ex0r ) cos i (Ex0i + Ex0r ) cos t = Z1 Z2 or Ex0r (cos i /Z1 + cost /Z2 ) = Ex0i (cos i /Z1 cost /Z2 ) Ex0r cos i /Z1 cost /Z2 = Ex0i cos i /Z1 + cost /Z2 cos i Z1 cos t /Z2 R = cos i + Z1 cos t /Z2 = cos i (2 /1 ) cos t (2 /1 ) cos t cos i + (13.104) (13.105) (13.106) (13.107) (13.103) (13.102)

In these equations, R = is the reection coecient. Also from Snells law sin i = sin t which makes cos t = 1 sin2 t = 1 (1/2 ) sin2 i (13.109) 2 1 (13.108)

408

which gives Er cos i R = = Ei cos i + (2 /1 ) sin2 i (2 /1 ) sin2 i

Proceeding to analyse the case of parallel polarisation, the magnetic elds fro the incident, reected and transmitted waves are: Hi = ax Hx0i ejk( y sin i z cos i ) H = a H ejk( y sin i +z cos i )

r x x0r

Ht = ax Hx0t ejk( y sin t z cos t ) equating the tangential magnetic elds at the boundary Hx0i + Hx0r = Hx0t we now nd the corresponding electric elds Ei = Z1 Hx0i a y cos i + az sin i ejk( y sin i z cos i )

(13.113)

Er = Z1 Hx0r a y cos i + az sin i ejk( y sin i +z cos i ) Et = Z2 Hx0t a y cos t + az sin t ejk( y sin t z cos t ) equating the tangential electric elds at the boundary Z1 (Hx0i cos i Hx0r cos i ) = Z2 Hx0t cos t substituting Equation 13.113 in this equation Z1 cos i (Hx0i Hx0r) = Z2 cos t (Hx0i + Hx0r ) Hx0r (Z1 cos i + Z2 cos t ) = Hx0i (Z1 cos i Z2 cos t ) Hx0r (Z1 cos i Z2 cos t ) = Hx0i (Z1 cos i + Z2 cos t ) from where we get using Snells law Er (2 /1 ) cos i + = R|| = Ei (2 /1 ) cos i + (2 /1 ) sin2 i (2 /1 ) sin2 i

(13.117)

If we stand out in the sun we know that our skin is warmed by the rays from the sun. How did the energy from the sun reach us? It is obvious that the light which falls on our skin has warmed our skin. And light is a form of

409

electromagnetic radiation the same waves which we have studied in Chapter 13. But our studies have not given us a clue as to how waves carry energy and power. All we know as a hint is that the product |E| |H| has the units of power density: (V/m)(A/m) = W/m2 From the knowledge which we obtained from that chapter, we can speculate that the dot product will not do, since the electric and magnetic elds are perpendicular to each other EH = 0 (13.121) Though the dot product is zero, the wave still caries power. We can consider the cross product, which denoted by P is P = EH (13.122)

The units are right. That is, though P 4 has the units of power density, but power density which is a vector? To understand about the concept of Power which is a vector we need to study the basic equations in more detail. The basic work on P, called the Poynting vector, was carried out by Henry Poynting (1852 1914) in 1884. He was an English physicist, and a professor of physics at Mason Science College (now the University of Birmingham) from 1880 until his death. The Poynting vector describes the direction and magnitude of electromagnetic energy ow and is used in the Poynting theorem, a statement about energy conservation for electric and magnetic elds.

Let us concentrate on the vector identity (E H) = H E E H The units of this equation is W/m3 throughout. Now H t E +J H = t E = (13.124) (13.125) (13.123)

4 In

410

the previous identity (E H) = H Identifying each term H H t H E +J E t t (13.126)

Is the power density, (W/m3 ), of the magnetic eld at any point in space. If we integrate over any given volume we will get the total power of the magnetic eld in that region. Obviously, the term H H Is proportional to the energy density stored in the magnetic eld (J/m3 ) at any instant of time, and at a given point in space. Similarly E E t

is the power density of the electric eld, and E E is proportional to the energy density stored in the electric eld. The last term of Equation 13.126, EJ

is the ohmic power density. Integrating Equation 13.126 over a region in space in accordance with Figure 13.14 H E + J dV E t t |H|2 |E|2 + + E J dV 2 t 2 t Heat H

Mag. Fld. Elec. Fld.

(E H) dV =

(E H) dS =

Poynt. Vect.

(13.127)

411

S E H dS

2 |E| V 2 t dV

V

|H|2 V 2 t dV

V E JdV

S (E H) dS =

|H|2 V 2 t

+ |E| + E J dV 2 t

To correctly interpret Equation 13.127 (refer to Figure 13.14) the term on the left is the total power entering the surface S which encloses the volume V6 . The terms on the right are (in that order): the total power gained (a) by the magnetic eld, and (b) by the electric eld. The last term is the power dissipated in ohmic losses. Hence we can write this equation as Power entering S = Power gained by {E in V + H in V} + Heat dissipated in V The two terms

5 To

H H, E E 2 2

1 Ai = t 2 1 = 2 2 Ai t 2Ai

i=x,y,z

i=x,y,z

i=x,y,z

Ai t

= LHS

6 The

surface integral denotes the total ux leaving the surface. The negative sign implies that the power is entering.

412

are the energy densities, J/m3 , stored in the electric and magnetic elds respectively.

The Poynting vector, P, is associated with the ow of power. Does the vector have actual physical signicance? Famous scientists have expressed doubt about its reality. Sir James Jeans states in his classic text, "The Mathematical Theory of Electricity and Magnetism", published by Cambridge University Press in the fth edition of 1933. Page 519. "The integral of the Poynting Flux over a closed surface gives the total ow of energy into or out of a surface, but it has not been proved, and we are not entitled to assume, that there is an actual ow of energy at every point equal to the Poynting Flux. (Italics mine). For instance, if an electried sphere is placed near to a bar magnet, this latter assumption would require a perpetual ow of energy at every point in the eld except the special points at which the electric and magnetic lines of force are tangential to one another. It is dicult to believe that this predicted circulation of energy can have any physical reality... The italicised part states that Poyntings theorem has meaning, but not the Poynting vector. That is the Poynting vector has no real physical signicance whatsoever. However, in the engineering world, and in most areas of electromagnetics especially in antennas the Poynting vector has been used with great success. We look at the case of the energy ow of the case of wire carrying a steady current.

I x

E

z

H S

y

Figure 13.15.: Poynting theorem applied to the case of a wire carrying a steady current

Consider the case of a current carrying conductor of conductivity and which carries a current I shown in Figure 13.15. The conductor has a steady electric eld Ez . The magnetic eld in the cylindrical coordinate system, at a point (, , z) outside (and just inside the conductor) is given by H = Now I = Jz a2 = a2 Ez 413 (Jz = Ez ) I 2

where the radius of the conductor is a, and Jz is the (constant) current density inside the conductor. Therefore Ez = I a2 I Jz = 2 a

We apply Poyntings theorem, Equation 13.127, to the Gaussian surface shown as a cylinder S in the gure. The Poynting vector is P = EH = a 0 0 a 0 H az Ez 0

= a Ez H It is interesting to note that the power ow is inward from the surface of the cylinder. This due to the negative sign with the unit vector a . The only region where the electric eld and magnetic eld are present together is inside the conductor. The Gaussian surface (shown as S in the gure) is made to coincide with the surface of the conductor, but innitesimally smaller and its length is d. Integrations on the two at surfaces contribute nothing, because the direction of the Poynting vector is parallel to those surfaces. The Pointing vector contribution is from the curved surface, an element of area of which is dS = a addz On the surface of the conductor the a part of the Poynting vector is P = Ez H = I I I2 = 2 3 2a a2 2 a

S =2,z=d 2 I addz =0,z=0

22 a3

414

If we calculate the total E J term

V

We notice that the two results are equal. Let us look at another example, that of a uniform plane wave travelling in space in the az . The electric and magnetic elds in phasor notation are given by Ex = E0 ejkz E0 jkz Hy = e Z0

(13.128)

Where k is the notation used for the propagation constant, , for free space and Z0 is the characteristic impedance of free space. We cannot work with phasor quantities since the Poynting theorem has been dened for real time variables. These elds in the real time are Ex = E0 cos (t kz) E0 cos(t kz) Hy = Z0 Then P = EH

(13.129)

ax a y = Ex 0 0 Hy = a z Ex H y = az E2 0

az 0 0

(13.130)

We notice that (a) The Poynting vector travels in the direction of wave propagation, which is intuitively satisfying; (b) the Poynting vector is always positive but pulsating, increasing from 0 to E2 /Z0 and then back to 0, and so on, sinu0 soidally; and (c) the average value of the P is E2 /(2Z0 ), since the average value 0

415

is given by, (T = 2) 1 Pav = T = =

T 0

2 1 E0 T T Z0 2 E2 0

2Z0

These ideas expressed above are reminiscent of circuits where the average power supplied, for example, to a load is given by P = VI cos where V and I are the rms phasor voltage across and current through the load and is the angle between them. Similarly working with electromagnetic phasors (but not rms values) 1 P = (E H) (13.131) 2 Because these are not rms values, the factor 1/2 has to be included. P is the called the Poynting vector, and E and H are the phasor electric and magnetic elds. On the other hand the time averaged Poynting vector is given by 1 Pav = {(E H)} 2 (13.132)

Let us nd out whether we are right. Using the Equation set 13.128 and applying the denition given above 1 Pav = {(E H )} 2 a ay 1 x = Ex 0 2 0 H y

az 0 0

416

In various communication equipment when the frequency is high (> 100 MHz) and the size of our lines is comparable to the wavelength of propagation, concepts of transmission lines have to be employed. Generally, for low frequencies, the length of the wires connecting the components can be ignored, and the voltage and current along the connecting wires can be assumed to be same at all points along the line. This is the lumped model of electrical circuits and is normally employed in regular circuit theory. For direct current this is strictly true, but for alternating current, if the length is comparable to the wavelength, the length becomes important and connecting wires must be treated as transmission lines. Examining Figure 12.6 on page 363 we can see that at suciently high frequencies, the line connecting the antenna to the receiver must be modelled as a transmission line. To be specic, take the example of a television connected to an antenna on the roof of a house. Typically the frequency of the received signal is of the order of a hundred megahertz. So (using = c/ f ), 3 m. The connecting line between the antenna and TV will be about 20 m which is about 7 wavelengths. Thus we see that to model the connecting line as a short lumped segment would lead to a faulty design. Let us consider another case, that where the frequency is only 30 kHz. Here the wavelength is about 10000 m, so 20 m (the length 1/500 ) may be modelled as a short segment of wire, and the lumped model may be applied without loss of accuracy. In general, the length of the wire is important when the signal includes frequency components with corresponding wavelengths which are comparable to the length of the wire. Transmission line theory must therefore be used in the design of circuits when the conditions outlined above apply. A common rule of thumb in connection with the length of the line is that the cable or wire should be treated as a transmission line if the length of the line is greater than 1/10 of the wavelength. At these lengths the phase delay and the interference of any reections on the line become important and can lead to unpredictable behaviour in systems which have not been carefully designed using transmission line theory. How do we proceed to analyse transmission lines? The rst point is that in general transmission lines have to be treated as if the line is composed of inductances, capacitances and resistors which are distributed rather than lumped. A lumped component has a value like L = 1 mH, for an inductor or C = 10 F for a capacitor, but distributed elements must be described in terms L = 1 mH/m, or L = 10 F/m, and their values are per meter rather than xed. The schematic of a transmission line is shown in the accompanying gure, Figure 14.1, while actual transmission lines are shown in Figure 14.2.

417

L R

.....

The gure shows two conductors, an upper conductor and a lower conductor with a distributed inductance of L (H/m) along the line, a distributed capacitance of C (F/m) between the upper and lower conductor, a resistance of R (/m) along the line and a conductance of G ( /m) in parallel with the distributed capacitance.

Conductor

Inner conductor

a

Lz Rz

b

I(z + z)

+

V(z)

I(z)

+

Cz Gz V(z + z)

z a

b

Figure 14.3.: Figure to analyse a transmission line

To analyse a transmission line we take a short section of such a line of length z, between two planes a a and b b on the line in accordance with Figure 14.3. The plane a a is a at coordinate point z while the plane b b is at a

418

coordinate point z + z where z is a very small distance as compared to the wavelength. To obtain the voltage equation, the voltage at plane b b , we proceed as follows: to the voltage, V(z, t), at plane a a we subtract the two voltage drops (=LzI/t) of the series inductor and series resistor (=RzI) between the two planes and thereby obtain the voltage at the plane b b , V(z + , t). The equation is V(z + z, t) = V(z, t) (Lz)

1

I(z, t) (Rz)I(z, t) t

2 3

(14.1)

Here term 1 is the voltage at plane a a in the gure. Term 2 is the voltage drop across the inductance, Lz, which is the total series inductance between the two planes. The third term, term 3, is a voltage drop across a resistance, Rz, in the short section of length z. By taking V(z, t) to the left side of the equation I(z, t) (Rz)I(z, t) t I(z, t) V(z, t) z L z RI(z, t)z z t I(z, t) V(z, t) L RI(z, t) z t

and cancelled the z term on both sides. The concern which we have with the previous equation is that there is a voltage term in the left hand side of the equation while a current term is present in the right hand side. Since the equation under discussion is a dierential equation, we would like voltage terms to be present on both sides of the equation. With this in mind, we obtain an analogous current dierential equation using the current at the two planes a a and b b . To obtain this equation, the current at plane b b , I(z + z, t), is equal to the current at plane a a , I(z, t), minus the two leakage currents through the shunt capacitance (=CzV/t) and resistance (=GzV). I(z + z, t) = I(z, t) Cz V(z, t) GzV(z, t) t (14.3)

Here (Cz) (V(z, t)/t) is the shunt current from the upper line to the lower line due to the capacitance Cz, and (Gz)V(z, t) is the total shunt current loss due to a lossy dielectric between the conductors. As earlier I(z + z, t) I(z, t) I(z, t) z z (14.4)

419

gives us the current equation V(z, t) I(z, t) = C GV(z, t) z t (14.5)

To reduce the number of terms and to consider only the most important terms we neglect (a) R, the series resistance, contributing to the copper loss and (b) G the conductance which contributes to the losses in the dielectric separating the two (upper and lower) lines. Then the equations become V(z, t) I(z, t) = L z t and I(z, t) V(z, t) = C z t (14.7) (14.6)

We have here two partial dierential equations in two independent variables (z, t) and two dependent variables (V, I). Seems complicated? Let us dierentiate the rst equation with respect to z 2 V(z, t) I(z, t) = L z t z2 = L = L = LC I(z, t) t z V(z, t) C t t 2 V(z, t) t2 (changing the order of dierentiation) (substitution from the other equation) (14.8)

Where we changed the order of the dierentiation and also used the second of the two equations. Let us take a look at the units of LC. L has units of ( sec)/m; C has the units of ( sec)/m; so the unit of LC is (sec/m)2 . A little reection tells us that 1/ LC has the units of velocity, v. i.e., v = 1/ LC. Equation 14.8 now becomes 2 V(z, t) 2 V(z, t) = LC = 1/v2 z2 t2 2 V(z, t) t2 (14.9)

This is a wave equation in one dimension, which we can corroborate by comparing this equation with Equation 12.51 on page 366. We observe that both equations are identical, and therefore we understand that V(z, t) denes a voltage wave travelling with a velocity v. This is the most important nding. When a voltage is applied to one end of a transmission line it travels down the line with a velocity v. In a similar manner we can dene the governing equation for the current I(z, t) 2 I(z, t) 2 I(z, t) 2 I(z, t) = LC = 1/v2 (14.10) z2 t2 t2

420

which is also a wave equations. In a transmission line, both the voltage and current are travelling waves.

To obtain the frequency domain equation for transmission lines is one where the voltages and currents are sinusoidal oscillations. As earlier, the time dependence is of the type exp jt and we have to dene distributed impedances and admittances based on inductances, capacitances. For example the distributed series impedance is Z = jL + R (14.11) where L and R have been dened earlier in the previous section. R is the resistance along the length of the line, contributing to copper losses. Similarly the shunt admittance is Y = jC + G (14.12) where G is the conductance describing the dielectric loss. Then V(z + z) = V(z, t) j (Lz) I(z) (Rz) I(z) V(z + z) V(z, t) = j (Lz) I(z) (Rz) I(z) V z = j (Lz) I(z) (Rz) I(z) z (14.13)

In these equations exp( jt) is implicit, and has not been included in the equation. That is V(z, t) = V(z)exp( jt) and I(z, t) = I(z)exp( jt) where V(z, t) and I(z, t) are the total voltage and current. Now proceeding as earlier dV(z) = ( jL + R)I(z) dz Similarly we can obtain dI(z) = ( jC + G)V(z) dz (14.15) (14.14)

Dierentiating the previous equation with respect to z followed by substituting the above equation we have d dV(z) d = ( jL + R) (I(z)) = ( jL + R)( jC + G)V(z) dz dz dz Which gives d2 V(z) = ( jL + R)( jC + G)V(z) dz2 = 2 V(z) (14.17) (14.18) (14.16)

421

In these equations ( jL + R)( jC + G) is the complex propagation constant = + j = ( jL + R)( jC + G)

This kind of realisation comes from comparing with the scalar Helmholtz Equation (12.71). is called the attenuation constant in neppers/m and is the propagation constant. Coming back to the second equation, we get the equation for the current d2 I(z) = ( jL + R)( jC + G)I(z) dz2 (14.19)

In these equations if we set R G 0 for low loss lines, the two equations become d2 V(z) = ( jL)( jC)V(z) = 2 LCV(z) dz2 d2 I(z) = ( jL)( jC)I(z) = 2 LCI(z) dz2 if we set then these equations become

d2 V(z) dz2 d2 I(z) dz2

(14.20)

2 = 2 LC

(14.21)

= 2 V(z) = 2 I(z)

(14.22)

which are two one-dimensional Helmholtzs wave equations for V(z) and I(z) respectively. The solution to these equations which are travelling waves are V(z) = V0 e j(tz) I(z) = I0 e j(tz) (14.23)

And if we drop the exp( jt) (this term is conventionally understood to be present) then V(z) = V0 ejz I(z) = I0 ejz (14.24)

Some notes are in order: The solution to the wave equation where both R and G are not neglected is: V(z) = V01 ez + V02ez

422

Where = + j = ( jL + R)( jC + G) The solution may be veried by substitution. For a loss less line where R = G = 0, = j LC = j and the functional dependence is exp( jz) or exp( jz). In general = + j a complex number, which implies that ez = ez ejz so the voltage wave decays as it progresses. (The arguments are same as given on page 394) One particular case is of great interest, that of the distortion less line where R L = = Z2 0 G C In this case = = ( jL + R)( jC + G) ( jZ2 C + Z2 G)( jC + G) 0 0 (14.26) (14.25)

= Z0 ( jC + G) = j LC + RG = RG = LC

(14.27)

Note that the real part is frequency independent and the imaginary part is proportional to the frequency which means that the velocity of propagation, v = / = 1/ LC (14.28)

is constant for all frequencies. Therefore a pulse sent down the line arrives at the other end undistorted. Next we ask the all important question: how do we get the values of L, C, R and G for physical lines? We give here a table (Table 14.1) of the L and C values for the case of the two wire line and the coaxial line, to calculate these two important parameters. (Since R and G are only present in lossy lines, these parameters are not presented here. The reader is referred to Jordan & Balmain, page 238, Table 7.1. EXAMPLE 14.1 Calculate the inner radius of the outer conductor for a 75 coax line whose inner conductor is of radius is 1 mm. The dielectric between the inner and outer conductor is polystyrene. Calculate the L and C for this line. Polystyrene has a dielectric constant of 2.56. So using the formula from Ta-

423

Table 14.1.: Calculation of L and C for two wire (conductor radius=a, spacing between centres=b) and the coaxial (radius of inner conductor=a, inner radius of outer conductor=b) lines.

Two wire line L C vp ble 14.1 we have Z = (1/2) / ln(b/a) (/) cosh1 (b/2a) ()/ cosh1 (b/2a) (1/) / cosh1 (b/2a) 1/

Z=

L/C

75 = (1/2) 0 /0 r ln(b/a) = (60/ 2.56) ln(b/a) = 37.5 ln(b/a) b/a = 7.3891 b = 7.3891 mm The capacitance/m for this line is given by C = (2)/ ln(b/a) = 27.81 pF/m while the inductance is L = 0.156 H/m EXAMPLE 14.2 Calculate the distance between the conductors for a 300 two wire line whose conductor radius is 1 mm. From Table 14.1 Z = (1/) / cosh1 (b/2a) 300 = (1/) 0 /0 cosh1 (b/2a) = 120 cosh1 (b/2a) b/2a = 6.1323 b = 12.265 mm Another transmission line which is used in a large number of applications is the micro strip line whose cross-section is depicted in Figure 14.4. The characteristic impedance of this line is given in Table 15.1. The investigation of this line is due to Wheeler 1977.

424

w Metallisation t

r

Figure 14.4.: Cross-section of the micro strip line

Since most of the time we will be working with sinusoidal quantities, we will consider the frequency domain equation rather than the time domain equation. The wave equation in the frequency domain, if we recall correctly, is the onedimensional Helmholtz equation with 2 = LC = g (14.29)

Where V+ and V are amplitudes of the forward and backward waves respectively and z is the distance along the line with respect to some origin. An illustration of these waves are shown in Figure 14.5.

We can conrm that plays the role of the propagation constant in a transmission line by taking a look at the units of . g is the wavelength of the voltage or current wave in the transmission line. Working with the rst equation of the previous equation set, the solution is V ejz for a wave travelling in the + z direction, or + V(z) = one of V e+jz for a wave travelling in the z direction

V e jz

V+ ejz

z

Figure 14.5.: Transmission line showing the forward and reverse voltage waves.

In any section of the line, for a coordinate z along the line, the voltage can be

425

written as a sum of these two solutions. That is, in particular, V(z) = V+ ejz +

Incident wave

V e+jz

Reected wave

(14.30)

For each of these solutions, the corresponding current waves are I ejz + I(z) = one of +jz I e for a wave travelling in the + z direction, or for a wave travelling in the z direction

To get the relation between I+ and V+ we use Equation 14.14 V(z) = jLI(z) z V+ ejz = jLI(z) z jV+ ejz = jLI(z)

or I+ = V+ Z0 (14.31)

Note that for a loss less line, Z0 is always real. By the same technique, we can show that I = V Z0 (14.33)

These two relations are very important, in that the forward and reverse voltage waves are linked to the corresponding current waves. That is if V(z) = V+ ejz + V e+jz then I(z) = 1 V+ ejz V e+jz Z0 (14.34)

426

Furthermore we introduce a new (but important) variable, (z), the reection coecient along the line for any value of z (z) = where V+ (z) = V+ ejz V (z) = V e+jz (14.36) V (z) V 2 jz e = V+ (z) V+ (14.35)

Since we have placed no restriction on V+ and V , these two coecients can, in general, be complex.

Let us take a look at the power in the forward/backward wave. The forward/backward wave carries a total average power of (since we are not considering rms values of the voltage and current, and Z0 is assumed to be real) 1 Pav = {VI } 2 1 = V ejz I ejz 2 1 = V ejz I ejz 2 V 1 = V 2 Z0 = |V |2 2Z0

(14.37)

A little explanation is needed. The subscript is the indicator of the forward/backward wave. The power, P+ , in the forward wave is positive, while the power in the reected wave, P , is negative. The ratio of the reected to the incident power is V P |V |2 /2Z0 = = P+ |V+ |2 /2Z0 V+

2

= ||2

(14.38)

We have obtained the solution to the transmission line equation, but we are not nearer to getting the values of V+ and V . How do we get these, to apply our solution to real problems? Let us consider the case where the incident electromagnetic wave, characterised by incident voltage and current waves meets a load impedance ZL , shown in

427

V+ ejz

V e jz

Generator

ZL

Load

z = L

z = l

z=0

Figure 14.6. The gure shows that the load is placed at z = 0, while the generator is at z = L. The incident parameters are V+ (z) = V+ ejz I+ (z) = I+ ejz (14.39)

After meeting the load a reected electromagnetic wave is set up characterised by V (z) = V ejz I (z) = I ejz The total voltage and currents are V(z) = V+ (z) + V(z) for z < 0 V+ (z) V(z) I(z) = I+ (z) + I(z) = Z0

(14.40)

for z < 0

(14.41)

428

We know that the total voltage and current across the load must satisfy ZL = = V(0) I(0) V+ ejz + V e+jz I+ ejz + I e+jz

z=0

V+ ejz + V e+jz = (V+ /Z0 ) ejz + (V /Z0 ) e+jz = Z0 V+ ejz + V V+ ejz V e+jz

z=0

e+jz

z=0

(14.42)

z=0

V+ = (0) V

(14.43)

ZL = 1 + L Z0 ZL L = 1 + L Z0 ZL ZL 1 = L +1 Z0 Z0 L =

ZL Z0 ZL Z0

1 +1

(14.44)

L =

V ZL Z0 = V+ ZL + Z0

(14.45)

429

V = V+ which means that the total incident power is reected1 . When the transmission line is terminated in a short circuit: ZL = 0 V ZL Z0 = V+ ZL + Z0 0 Z0 = 0 + Z0 = 1 V = V+ L = again we can see that all the incident power is reected. When the transmission line is terminated in the characteristic impedance: ZL = Z0 V ZL Z0 = V+ ZL + Z0 Z0 Z0 = Z0 + Z0 =0 V = 0 L = since there is no reected wave all the incident power is consumed. This is the case of maximum power transfer: the condition of a matched load. We also know that if the load is purely imaginary (ZL = jX, X being real) no power is consumed by the load. In this case too all of the incident power must be reected L = V ZL Z0 = V+ ZL + Z0 jX Z0 = jX + Z0 = 1ej2 tan

1 (X/Z 0) 0) 1 (X/Z

V = 1ej2 tan

V+

(14.46)

430

which conrms our surmise. EXAMPLE 14.3 Find the percentage of reected power when a load of value ZL = 40 + j40 terminates a 75 (= Z0 ) transmission line.

L =

2

Now |L | = and V V+

= .19056

So 19% of the power is reected. The reection coecient along the line is from an earlier equation is (z) = V 2z e V+ = L e2z

which tells us that the magnitude of (z) remains constant, = |L |, while the phase changes by 2l. Using = 2/ g (l) = 4l where l = l/ is the distance in wavelengths towards the generator. If we study this equation carefully we nd that the complex reection coecient has a constant magnitude, |L |, and it returns to its original value every half a wavelength: that is for l = 0.5. (14.49)

431

Referring to Figure 14.6, the voltage along the line at a point z = l is given by V(l) = V+ ej(l) + V e j(l) = V+ (cos l + j sin l) + V(cos l j sinl) = (V+ + V ) cos l + j sin l (V+ V ) = VL cos l + jZ0 IL sin l Z0 sin l = VL cos l + j ZL

(14.50) (14.51)

Which is a mathematical statement of the fact that the total voltage along the line is a sum of the forward and backward voltage waves. Here VL and IL are the voltage and current across and through the load. VL = V(0) = V+ ej(l) + V e j(l) l=0 = V+ + V IL = I(0) = I+ ej(l) + I e j(l) l=0 = I+ + I = VL = IL ZL

V+ V Z0 (14.52)

1.4 1.2 Normalised voltage 1 0.8 0.6 0.4 0.2 0 3 2.5 2 1.5 1 Distance in wavelengths 0.5 0

ZL /Z0 = 2

Figure 14.7.: A plot of the magnitude of the voltage along a line for VL = 1 V and ZL /Z0 = 2

A plot of the magnitude of the normalised voltage, |V(l)/VL|, along the line for VL = 1 V and ZL /Z0 = 2 is shown in Figure 14.7. The gure shows the resultant sum of the forward and backward voltages. A study of the gure shows that

432

1. There is a voltage maxima at the load. 2. The periodicity of the standing wave pattern is /2 instead of the usual . 3. Vmax /Vmin = 2. Let us investigate this analytically. (ZL is real) V(l) = VL cos l + j |V(l)| = VL cos2 l + Z0 sin l ZL Z0 ZL

2

sin2 l

(14.53)

2 cos( l) sin( l) z

sin2 ( l) z

To nd the maximas and minimas we set the numerator to be zero, which gives sin(2l) = 0 2l = n n l= 2 2 n = 4

(14.54)

Hence maximas and minimas occur at l = 0, /4, /2, 3/4, ...At these lengths l = Making a table l |V(l)| 0 1 /2 Z0 /ZL 1 3/2 Z0 /ZL ... ... 2 n n = 4 2

Comparing values in the second row, Z > Z then there is a maximum at the load (1 > Z /Z ) L 0 0 L if Z < Z then there is a minimum at the load (1 < Z /Z ) L 0 0 L

and maxima and minima alternate thereafter at intervals of /4. The ratio Vmax /Vmin is called the voltage standing wave ratio (VSWR) s (VSWR) = Vmax /Vmin (14.55)

We now proceed to obtain an analytic expression for s. The voltage along the

433

line is given by V(z) = V+ ejz + V e jz (14.56) Due to some reection, a standing wave is formed. Let the reection coecient be L . Then V = L V+ (14.57) or the previous equation may be written as V(z) = V+ ejz + L e jz |V(z)| = |V+ | ejz + L e jz (14.58)

Now the maximum value of the above equation is when the maximum of each individual term adds and the minimum value is when the two terms in the bracket subtract (remember that |L | 1). That is |V(z)|max = |V+ | |(1 + |L |)| |V(z)|min = |V+ | |(1 |L |)| so s (VSWR) = |V(z)|max 1 + |L | = |V(z)|min 1 |L | (14.60) (14.59)

Applying this formula to the above case, that is, the case of the gure, ZL /Z0 = 2 ZL /Z0 1 L = ZL /Z0 + 1 = 1/3 VSWR, s =

1 13 2 3

=2

For the current, we may proceed along similar lines. The current along the line is given by I(l) = I+ ej(l) + I e j(l) = I+ (cos l + j sin l) + I(cos l j sinl) = (I+ + I ) cos l + j sin l (I+ I ) VL = IL cos l + j sin l Z0 ZL sin l = IL cos l + j Z0

(14.61) (14.62)

A plot of the magnitude of the normalised current, |I(l)/IL|, along the line for IL = 0.5 A and ZL /Z0 = 2 is shown in the accompanying gure. These are the voltage and current along the line in terms of the load parameters and load impedance.

434

3 2.5

Normalised current

1.5

0.5

Distance in wavelenths

Z0 ZL /Z0 = 2

Figure 14.8.: A plot of the magnitude of normalised current along the line for IL = 0.5 A and ZL /Z0 = 2

z= l

ZL

Zin

Figure 14.9.: Input impedance of a transmission line.

We now take a look at the input impedance at a point z = l shown in Figure 14.9. This impedance must be the ratio of the voltage to the current along the line. Zin = Z(l) V(l) = I(l)

435

We now substitute Equations 14.51 and 14.62 into the above equation Zin = VL cos l + j

Z0 ZL sin l Z IL cos l + j ZL sin l 0 Z0 ZL cos l + j ZL sin l Z cos l + j ZL sin l 0

(since ZL = VL /IL )

1 Z0

Z0 cos l + jZL sin l ZL + jZ0 tan(l) Z0 + jZL tan(l) ( Dividing the Num. and Den. by cos l) ZL + jZ0 tan(l) Z0 + jZL tan(l)

= Z0

Zin (l) = Z0

(14.63)

Let us apply this formula to various cases. Open circuit: ZL = Zin = Z0 = Z0 ZL + jZ0 tan(l) Z0 + jZL tan(l) ZL jZL tan(l) (ZL Z0 )

= jZ0 cot l Short circuit: ZL = 0 Zin = Z0 = Z0 ZL + jZ0 tan(l) Z0 + jZL tan(l) jZ0 tan(l) Z0 (ZL = 0)

= jZ0 tan l Matched load ZL = Z0 Zin = Z0 = Z0 = Z0 ZL + jZ0 tan(l) Z0 + jZL tan(l) Z0 + jZ0 tan(l) Z0 + jZ0 tan(l) (ZL = Z0 )

436

{} r = 0, x = 1, = j

Complex Plane

Unit circle

jv (z) r = 0, x = 0, = j r = 1, x = 0, = 0 4l u

{} r = , x = 0, = 1 + j0 l Toward generator

(z l)

r = 0, x = 1, = j1

Figure 14.10.: shown in the complex plane

It is clear that the number of equations involved in transmission line problems increases the complexity of the computation, hence a graphical procedure has been evolved using transmission line charts. We know that the terminating impedance ZL is complex, as is with the case with Zin , but for any sense to be made of this parameter, it must be normalised with respect to the characteristic impedance, Z0 , of the line. So we dene a new parameter Zin (l) = Zin (l) Z0 = r + jx (14.64) (14.65)

Using the new parameter Zin Z0 Zin + Z0 (Zin /Z0 1) u + jv = (Zin /Z0 + 1) Zin 1 = Zin + 1 =

( = u + jv) (14.66)

437

This is a mapping between the complex variables and Zin . And the relation between u and v as functions of r and x are u= x2 + (r 1) (r + 1) x2 + (r + 1)2 2x x2 + (r + 1)2 (14.67) (14.68)

v=

Thus for every r and x we get a u and a v. (remember = u + jv). With reference to Figure 14.10, we know that since generally || = V (z) 1 V+ (z) (14.69)

then the region contained within the unit circle (which is || = 1) contains all values of r and x. For example, using the above equations, 1. r = 0, x = 0 maps to u = 1, v = 0. 2. r = 1, x = 0 maps to u = 0, v = 0. 3. r = , x = 0 maps to u = 1, v = 0. 4. r = 0, x = 1 maps to u = 0, v = 1 5. r = 0, x = 1 maps to u = 0, v = 1 The transformation in the other direction is more useful Zin = 1+ 1 (14.70)

r=

(1 u) (u + 1) v2 v2 + (1 u)2 2v v2 + (1 u)2

(14.71) (14.72)

x=

(1 + r) v2 + r + (1 + r)u2 2ru = 1 2r 1 r + u2 u= v2 + 1+r 1+r 1+r r 2 1 r r 2 v2 + u = 1+r 1+r 1+r 1+r r 2 1 r r v2 + u = + 1+r 1+r 1+r r+1 1 2 r 2 v2 + u = 1+r 1+r

(14.73)

which is the equation of a circle with centre [r/(1 + r), 0] and radius 1/ (1 + r).

438

Similarly the other equation gives (u 1)2 + v 1 x

2

1 x

(14.74)

which is the equation of a circle with centre (1, 1/x) and radius 1/ |x|. On the gamma plane let us draw the ve circles with r = 1, r = 0, x = 1 and x = 0 which are shown in the gure. for Z = 1 + j1 = 0 + j1

unit circle x=1

Z = 1 + j1

r = 0 circle

r=1

= 1 + j0

x=0

=0

= 1 + j0

x=1

Complex plane

= 0 j1

Figure 14.11.: plane with r = 1, r = 0, x = 1 and x = 0 circles

Let us see how these plots were obtained. 1. The r = 1 circle by drawing a circle with centre (1/2, 0) and radius 1/2. 2. The r = 0 circle has its centre at (0, 0) and has a radius of 1. 3. The x = 1 circles have their centre at (1, 1) and radii of 1. 4. The x = 0 circle has its centre at (1, ) and radius of , which is a straight line. In this way we can get every point on the gamma plane in terms of (r, x), which is the Smith chart Smith (1939). Let us apply the Smith chart to some simple cases.

EXAMPLE 14.4 The a 20 m 50 line operated at 350 MHz is terminated with a 75 load. Find the input impedance and the complex reection coecient at the input end. At the outset we nd the wavelength and the normalised impedances. The

439

440

GSMC printout

source file: Date: Tue Jul 1 11:38:15 2008

+j1.0

+j0.5

+j2.0

+j0.2

in

0.2 0.5 1.0 1.000+j0.000

2.0

Mark value of ZL

z0= 1.00

f0= 350.00M

the normalised load impedance plotted on the Smith chart gives the position marked start in Figure 14.13. The origin of the complex gamma plane connected to this point is L (as is shown). L = 0.2 we have to move towards the generator by the amount l = 20 m

441

to give the new value of the reection coecient (l) = L ej2l = L ej (4/) l = L ej4l the length of the line in wavelengths is l = 20/0.85714 = 23.33 in terms of phase this is = 4l = 2 46.66 rad which is 46 complete revolutions and followed by 238 = 180 + 58 in the clockwise sense. Having decided the angle, one moves on the Smith chart in the clockwise sense by on the circle shown. One moves on a circle since the magnitude of the reection coecient is a constant, as one moves toward the generator. The nal value of (l) is in = L ej = 0.11 + j0.17 this value of the reection coecient falls on the normalised input impedance Zin which when read out from the Smith chart gives us Zin = 0.8 + j0.27 this is the value of Zin at the end of the 20 m line. The actual, un-normalised impedance is Zin = 50 Zin

= 38 + j13.5

= 50 0.8 + j0.27

EXAMPLE 14.5 With the load given above, and the transmission line, we would like to somehow match the load to the line, so that the power delivered to the load be maximum. How do we go about doing this? To match the load, we will have adopt a strategy. Referring to Figure 14.14 we move along the line towards the generator, by a distance l along the physical line and from a to b on the Smith chart, such that the point b lies on the r = 1 circle. Doing this we nd that the normalised impedance at b is Zin = 1 j0.43 442

GSMC printout

source file: Date: Tue Jul 1 11:38:15 2008

+j1.0

+j0.5

+j2.0

+j0.2

a

0.2 0.5 1.0 1.000+j0.000 2.0

l

z0= 1.00 f0= 350.00M

Zs = j0.43 ZL = 1.5

Zin = 1

Zin = 1 j0.43

at this point we cancel out the reactive part of Zin by adding a series reactance Zs = j0.43 so that beyond that point, Zin = 1. turns out to be (from the Smith chart) = 39 Corresponding to the distance l turns out to be l = 0.11

443

if we want more line, l can be increased by half a wavelength. This is because one revolution of the Smith chart corresponds to exactly half a wavelength

l = 0.11 + 0.5 = 0.61 Three questions arise from the consideration of the above examples. 1. Is it always possible to match any load? 2. How are we going to conveniently obtain a series reactance? 3. Is it possible to use a shunt reactance? Let us answer these questions one by one. 1. The answer to the rst question is yes, by the method outlined above, it is possible to match any load. 2. The next question is answered by considering a transmission line terminated in an open or short circuit. A transmission line terminated in a short circuit has an input impedance given by Zin = jZ0 tan(l) or Zin = j tan(l) for example to realise Zs = j0.43 we equate tan(l) = 0.43 l = 0.4061 2l = 0.4061 l = 0.064633 l = 5.54 cm The circuit for the example is shown in Figure 14.15 for the example being considered. In the same manner if we use an open circuited line, the required length would be obtained from cot(l) = .43 tan(l) = 2.3256

3. To consider admittances (which for physical reasons is more convenient) we need to consider the Smith chart with admittances as the parameters. To this

444

9.4 cm

5.54 cm

Z0 = 50

Figure 14.15.: Circuit for the example being considered

75 Ohm

If however without going from to we want to read the admittance directly, then we need a Smith chart which has both the impedance and admittance on the same chart. We will use this kind of chart to match the load of the example which we have been considering. The start point is the load of 75 which gives ZL = 1.5 and YL = 0.667. We now go down the line by a distance of l = 0.142 which brings us to Zin1 = 0.855 j0.348 and the corresponding Yin1 = 1.0 + j0.409. We now add a short circuited stub of value Ysp = j0.409 in parallel with the line. The length of the stub is 0.188 . The nal end point is Yin2 = 1.0 + j0.

1 (14.75) 1+ in this transformation, if we replace by we obtain the original transformation (the previous equation) and the new chart can be read just like the old one with r g and x b. For example if we are at the normalised impedance point Z = 1 + j2 then the complex refection coecient for this point is the straight line joining the origin (which is Z = 1 + j0) to Z = 1 + j2. Let = || . To go over to the admittance Smith chart, we plot (= || ( + )) on the chart and we can read out the normalised admittance value of Y = 0.2 j0.4 from the chart, treating r as g and x as b. To conrm, we can also do the mathematics 1/(1 + j2) = 0.2 j0.4. Yin = g + jb =

445

GSMC printout

source file: Date: Tue Jul 1 15:24:44 2008

+j1.0

+j0.5

+j2.0

+j0.2

1.0 2.0 50.000+j0.000

-j0.200

-j2.000

-j0.500

-j1.000

z0= 1.00

f0= 1.00M

0.142 50 75 0.188

Suppose we have two lines of characteristic impedances Z01 and Z02 and we would like to match these two lines. How would we go about doing it? Starting from Zin (l) = Z0 ZL + jZ0 tan(l) Z0 + jZL tan(l)

446

GSMC printout

source file: Date: Tue Jul 1 11:38:15 2008

+j1.0

+j0.5

+j2.0

+j0.2

ZL = 0.8 + j0.8

Zin1 = 2.44 + j0

0.2 0.5 1.0 1.000+j0.000 2.0

z0= 1.00

f0= 350.00M

50

78

50 0.113

ZL = 0.8 + j0.8

0.25

Figure 14.17.: Matching with a transformer

we want Zin (l) = Z01 and ZL = Z02 then Z01 = Z0 Z02 + jZ0 tan(l) Z0 + jZ02 tan(l)

consider a line of length /4 which implies that l = (2/)(/4) = /2 or tan(/2) then Z01 Z02 = Z2 0

447

or Z0 = Z01 Z02 (14.76)

This formula can be used to match two lines as we are considering or even any given load. Thus to match a load, for example of ZL = 40 + j40 (which is ZL = 0.8 + j0.8) we move toward the generator a distance of l = 0.113 which brings to the point Zin1 = 122 + j0 (Zin1 = 2.44 + j0) which lies on the x = 0 line on the Smith chart. Now we take a section of quarter wavelength line of characteristic impedance Z0 = 50 122 = 78 and the load is matched to the 50 line. The results are shown on a Smith chart. The method of transformer matching has been extended to wide band matching by many researchers. EXAMPLE 14.6 We would like to look at a practical but simple example of the application of the theory which we have developed here. A folded dipole antenna receives EM energy and must be connected to a television receiver, so that we may be able to see our TV programs. How do we go about it? The rst point which we must decide is which is the generator and which is the load. After a little reection, we come to the conclusion that the antenna is the generator, since the EM energy must go form the antenna to the TV receiver. Next we must model the TV receiver as a load, as well the antenna as a generator. If we search for the specications of a folded dipole we nd that it has an input impedance of 300 . Most modern televisions have an input jack marked 50 , so the conguration is now as shown in the gure below.

Since we always start with the load and the transmission line, what should be the characteristic impedance of the line? As a rule of thumb, the transmission line must be matched to the generator. Hence the line should be a two wire ribbon line with Z0 = 300 . However the line must be connected to a 50 load, so there must be a match. We now use a transformer. We can use a quarter wavelength transformer with a characteristic impedance of 123 , but the modern solution is to use a 6 : 1 turns ratio transformer which operates into the MHz region. So the nal set-up is shown below.

448

2.5 m

1m Z1=j50

1.5 m

f=30 MHz

From the gure = 3 108/3 107 = 10 m

= 2/ = 0.628 rad/m = 36 /m At the load L = 0.321 + j.377 Moving toward the generator by 1.5 m we arrive at Z(l = 1.5) = 229 + j114 (l = 1.5) = 0.458 + j0.118 Here we add j50 to the impedance in series. The new impedance and reection coecient are Z(l = 1.5) = 229 + j164 (l = 1.5) = 0.514 + j0.244 449

The star notation l = 1.5 is used to indicate the parameter values on the side of the discontinuity on the generator side. Next we move toward the generator by 1 m. The new parameters are Z(l = 2.5) = 125 j152 Y(l = 2.5) = 0.00322 + j0.00393 At this point we add an admittance of j0.2 are (l = 2.5) = +0.390 j0.414 in shunt. So the new parameters

Y(l = 2.5) = 0.00322 j0.0161 Finally we move toward the generator by 2.5 m and Z(l = 5) = 32.21 j161 (l = 5) = 0.390 j0.742 We know from (l = 5) that |(l = 5)|2 100 = 70.27% is the percentage of power which is reected and

1 |(l = 5)|2 100 = 29.73% is absorbed by the system. But, except for the load, all the other elements are reactive, therefore this is also the percentage of the power absorbed by the load. But to compute the incident power, we need to compute V+ (z) = V+ exp jz and V (z) = V exp jz along the line. The important point to note is to move between transmission line concepts and lumped concepts as and when required. Let us now start at the generator. The voltage across Z(l = 5) and the current through this impedance is V(l = 5) = I(l = 5) = (100)(32.21 j161) = 8.7422 j2.4633 = V+ (5) + V(5) 50 + (32.21 j161)

V+ (5) V(5) (100) = 0.025156 + j0.049266 = 50 + (32.21 j161) Z0 V+ (5) V(5) = 2.5156 + j4.9266

450

V+ (5) = 5.6289 + j1.2317 V (5) = 3.133 j3.695 I+ (5) = V+ (5)/Z0 I (5) = V (5)/Z0 Hence the incident power is |V+ (5)|2 /Z0 = 0.33 W We know that about 30% is absorbed. Therefore only 0.1 W is delivered to the load. To continue our calculations, we move toward the load by 2.5 m which corresponds to a length of /4 V+ (2.5) = V+ (5)ej(/4) = 1.2317 j5.6289

V (2.5) = V (5)e j(/4) = 3.695 + j3.133 The total voltage is V(2.5) = V+ (2.5) + V(2.5) The total current is I(2.5) = {V+ (2.5) V(2.5)}/Z0 = 0.024633 j0.087422

This current gets divided between to admittances of values j0.02 (in shunt) and 0.00322 + j0.00393 which is the input impedance at this point. Using the current divider theorem I(2.5) = (0.024633 j0.087422) (0.00322 + j0.00393) j0.2 + 0.00322 + j0.00393

While the voltage remains the same beyond the shunt element V(2.5) = V+ (2.5) + V(2.5) = V+ (2.5) + V(2.5) From these equations we compute V+ (2.5) and V (2.5), and proceed toward the load. In this manner we continue on.

14.6. References

451

15. Waveguides

Energy transfer in the gigahertz range must necessarily be done through propagation of waves travelling in hollow pipes called waveguides. The method of energy transfer by using an open structure consisting of wires is inecient the radiation losses are prohibitive. These hollow pipes, waveguides, were rst proposed by Sir J. J. Thompson; they were experimentally veried by O. J. Lodge, and mathematically analysed by Lord Raleigh in the 1890s. As a start to our understanding of wave guides, we take a look at the parallel plate waveguide.

x a z o y

Referring to Figure 15.1, the parallel plate waveguide consists of two metal plates whose inner surfaces are placed parallel to each other at a distance of a m apart. Each of the plates is innite in extent in the y and z directions. The inner surface of the lower plate is coincident with the yz plane and is described by x = 0 while the upper surface is described by the the equation x = a. It is clear that due to the innite extent of the surfaces, the parallel plate waveguide cannot be really constructed and is, in reality, an exercise in the understanding of (a) waveguides in general, and (b) the method applied to tackle them mathematically. In the parallel plate waveguide, the two metal plates are assumed to have innite conductivity. This assumption is not very

452

15. Waveguides

far o from the truth, since if we construct the waveguide from copper plates, 107 , which is very close to innity. In the waveguide, an electromagnetic wave is assumed to be travelling and the direction of the ow of power, in this case, is assumed to be in the +z direction. We suppose that the wave is based on sinusoidal oscillations with a frequency and since the wave is oscillatory there is an implicit exp jt dependence in the electromagnetic elds. Since we are considering a travelling wave we must apply the Helmholtz equation, which applies to waves with sinusoidal oscillations. Equation 12.71, applied to the electric, E and magnetic, H elds is 2 E = k2 E

2 H = k2 H

(15.1)

where k (= ) is the free-space propagation constant and E and H are phasors. The z and time dependence, t, of the E and H elds is e j(tz) , which represents a travelling wave moving in the +z direction with a propagation constant = 2/ g . g being the wavelength in the waveguide. In the y direction the waveguide is innite in extent, hence the elds do not depend on the variable y. Therefore E = E0 (x)e j(tz) H = H0 (x)e j(tz) Working with the x component of the electric eld Helmholtz equation 2 Ex = k2 Ex Since /y gives zero, (no dependence on y) and Ex = E0x (x)e j(tz) = jE0x (x)e j(tz) = jEx z z which reduces to multiplication of Ex by j. The previous equation becomes 2 Ex = 2 Ex 2 Ex + + x2 y2

=0

(15.2)

2 Ex z2

=(j)(j)Ex

= k2 Ex

2 Ex + (k2 2)Ex = 0 x2

2 Ex 2 Ex = k2 Ex x2 (15.3)

453

15. Waveguides

The solution to this equation is (if k > ) Ex (x, z) = Ex01 sin( k2 2x) + Ex02 cos( k2 2 x) ejz (15.4)

Note that exp jt is implicit (understood to be present). To simplify the expression we can set kx = k2 2 (15.5) The x dependence of the other elds is similar. Next we apply the boundary conditions of the waveguide. The tangential electric eld and the normal magnetic eld on a perfect conductor are both zero. Working with E y which is tangential to the two planes, E y (x, z) = E y01 sin (kx x) + E y02 cos (kx x) ejz (15.6)

must be zero for x = 0 and x = a for all values of z. Applying the rst condition rst, E y01 sin (kx x) + E y02 cos (kx x) ejz =0

x=0

which is E y01 sin (kx 0) + E y02 cos (kx 0) ejz = 0 E y02 = 0 since cos( k2 2x) = 1 for x = 0. Therefore E y = E y01 sin (kx x) ejz Applying the second boundary condition at x = a gives sin(kx a) = 0 kx a = m kx = therefore E y = E y01 sin In the same manner Hx = Hx01 sin(kx x)ejz Ez = Ez01 sin(kx x)ejz (15.11) m jz x e a m=1,2. . . (15.10) m=1,2,. . . m k2 2 = a m=1,2,. . . (15.9) (15.8)

(15.7)

since Hx is normal and Ez is tangential to the two planes and therefore both must be zero on the two planes. The concept of modes now comes into play. For example when m = 1 1 those

1

(when m = 0 for a sin(.) function means that that particular eld vanishes and in the case of cos(.)

454

15. Waveguides

elds which have a sin(mx/a) dependence have a single maximum in the x direction, and they vanish at the end points. While in a similar case when m = 2, then the eld will have one maximum and one minimum; and so on. To proceed further, Maxwells equations have to be taken into account. Working rst with the curl of the electric eld E = jH which is Ez E y = jHx y z Ex Ez = jH y z x E y Ex = jHz x y

(15.12)

Since /y gives zero, and /z reduces to multiplication by j these equations become jE y = jHx jEx

Ez = jH y x E y = jHz x

(15.13)

Working next with the curl of H H = jE which become jH y = jEx jHx Hz = jE y x H y = jEz x

(15.14)

Looking at the internal structure of the elds given above, E y , Hz and Hx form one set of elds while Ez , H y and Ex form another set. If one set is present the other set need not be present. We dene the rst set as Transverse Electric (Ez = 0) elds (for obvious reasons) while the second set are called the Transverse Magnetic (Hz = 0) elds. This general manner of division is useful in analysis of other waveguide structures as well. 1. TE modes. Let us start with E y = A sin

the eld reduces to a constant)

m x ejz a

(15.15)

455

15. Waveguides

then Hx = A m jz sin x e a m jz m cos x e Hz = A ja a

x

(15.16)

Ey

Hz

(a)

1111111111111 0000000000000

(b)

Figure 15.2.: Fields for the parallel plate waveguide. (a) E y (b) Hz

m x ejz a

(15.17)

(15.18)

Observing the elds of the rst TE mode (m = 1) we nd that E y is proportional to sin(x/a) and Hz is proportional to cos(x/a). These two elds are plotted in Figure 15.2. E y is zero at x = 0 and x = a, as it should be, since the tangential electric eld must be zero at the walls of the metallic conductor. On the other hand Hz is zero at x = a/2 and has maximum and minimum values at the two walls. Due to the presence of Hz there is a -y directed surface currents on the two walls at x = 0 and x = a. These currents change direction every half a wavelength.

456

15. Waveguides

Going over to the examination of the modes is given by the formula = k2 m a

2

(15.19)

What happens when (k = 2/ = /c) becomes less than m/a? We nd that then becomes imaginary, = j and ejz = ej(j)z = ez (15.20)

and the wave decays. Such a mode is called an evanescent mode and is said to be cut o. What is the cut o condition? Starting with zero frequency, the value of k = /c is zero, and the wave is cut o. As the frequency increases, there comes one frequency, c when the wave just begins to propagate. That is when passes from being imaginary to being real. At that frequency = 0 = kc = k2 c m a

2

=0 (15.21)

m a 2 m = c a 2a c = m

(15.22)

where c is the free space wavelength corresponding to the cut o frequency c . When the frequency is lower than the cut o frequency, that particular mode is evanescent. 3. TEM mode. (both Ez = 0 and Hz = 0) Is there a possibility of m = 0? If we examine the TE modes we nd that there is no travelling wave which satises Maxwells equations. However when we put m = 0 in the TM modes we nd that H y = Aejz (15.23) and Ex = A jz e (15.24)

The other elds being zero. Maxwells equations are indeed satised. Furthermore m 2 = k2 =k (15.25) a

m=0

the propagation constant will coincide with the free space propagation con-

457

15. Waveguides

stant. So H y = Aejkz k Ex = A ejkz jkz e =A jkz e =A = AZejkz Z = Z0 = 120 = 377 for free space. (15.26)

As we saw in the previous section that apart from supporting higher order modes, the parallel plate wave guide also supports the TEM mode. TEM modes exist on two conductor lines and may be analysed as transmission lines for frequencies starting from dc to the start of the rst higher order mode. These waveguides have equivalent transmission line equivalent circuits. The most important of these are 1. Two conductor lines (see Figure 14.2 on page 418) Here there are two parallel conductors with conductor radius=a and the spacing between centres=b. Then the transmission line parameters for the two wire line are L = (/) cosh1 (b/2a) C = ()/ cosh1 (b/2a) (15.27)

2. Coaxial lines (see Figure 14.2 on page 418) In this waveguide there are two concentric conductors, with the inner conductor having a radius=a and the inner surface of the outer conductor having a radius=b. L = (/2) ln(b/a) C = (2)/ ln(b/a)

(15.28)

3. Parallel plate transmission line (See Figure 15.3) Here there are two plates placed a m apart and each of width w. Then the parameters L and C are (approximately) L a/w C w/a 4. Microstrip Line (See Figure 14.4 on page 425) The Microstrip line has a characteristic impedance given by

(15.29)

458

15. Waveguides

w/h < 1 Zm = 60 e f f =

er +1 2

ef f

w/h 1

1/2 2

h ln 8 w + 0.25 w h

Zm = e f f =

1 h + er 2 1 + 12 w

+ .04 1 w h

Zm is the characteristic impedance of the Microstrip line r is the dielectric constant of substrate, w is the width of the strip, h is the thickness (height) of substrate and

The rectangular waveguide is a hollow pipe of rectangular cross-section, capable of conveying travelling waves. A three dimensional view of the waveguide is shown in Figure 15.4. The inner surface of the waveguide is made of some high conductivity material like copper or gold and is rectangular in crosssection. The dimensions of the rectangle are a in the x-direction and b in the y-direction as shown with a > b. The wave is assumed to be travelling in the z direction with a time dependence of the form exp( jt) and a z dependence of the type exp( jz) where (= 2/ g ) is the propagation constant. To analyse the waveguide mathematically we need to make a slight detour on the analysis of the Helmholtzs equation. We will take a look at the scalar Helmholtz equation 2 + k2 = 0 (15.30) 459

15. Waveguides

y x a b z

Figure 15.4.: 3 dimensional view of a rectangular waveguide

k = . may be replaced by any of the eld components, Ex , E y . . . or Hz . Since the coordinates of the guide are regular we assume that is of the form = X(x)Y(y)ejz (15.31)

that is X is a function of x alone, Y is a function of y and both are multiplied together to give the elds in the cross-section of the waveguide2 . On substituting this functional dependence of into the Helmholtz equation 2 X(x)Y(y)ejz + k2 X(x)Y(y)ejz = 0 X(x)Y(y)ejz + k2 X(x)Y(y)ejz = 0 (15.32)

2 2 2 + 2+ 2 2 x y z Or

2 2 2 X(x)Y(y)ejz + 2 X(x)Y(y)ejz + 2 X(x)Y(y)ejz +k2 X(x)Y(y)ejz = 0 2 x y z (15.33) Or Y(y)ejz 2 Y(y) 2 X(x) +X(x)ejz +( j)( j)X(x)Y(y)ejz +k2 X(x)Y(y)ejz = 0 x2 y2 (15.34)

dividing throughout by X(x)Y(y) exp( jz) and simplifying 1 2 Y(y) 1 2 X(x) + + (k2 2 ) = 0 X(x) x2 Y(y) y2 (15.35)

2 This

460

15. Waveguides

Now substituting 1 2 X(x) X(x) x2 1 2 Y(y) f2 (y) = Y(y) y2 f1 (x) = we get f1 (x) + f2(y) + (k2 2) = 0

(15.36)

(15.37)

Suppose for the moment that for some value of the pair (x, y) = (x0 , y0 ) the equation is satised. f1 (x0 ) + f2 (y0 ) + (k2 2) = 0 (15.38) Now let x increase to x0 + x then y must change to keep the equation satised, unless each of the two functions is equal to a constant: f1 (x) = k2 (a constant) x

(15.39)

(15.40)

and going back to the earlier equations 1 2 X(x) = k2 x X(x) x2 1 2 Y(y) = k2 y Y(y) y2

(15.41)

The solutions to these equations is (as in the case of the parallel plate waveguide) are X(x) = X01 cos(kx x) + X02 sin(kx x) Y(y) = Y01 cos(k y y) + Y02 sin(k y y) (15.42)

kx and k y will be assume denite values on applying the conditions at the boundaries of the waveguide, and this will be clear a little later in the discussion. Let us now consider Maxwells equations and the electromagnetic elds whose time dependence is exp jt and the z dependence is exp jz (which is the condition of a wave travelling in the z direction). is the frequency of oscillation of the elds in radians/sec, and (= 2/ g ) is the propagation constant. The curl of E is E = jH

461

15. Waveguides

which becomes on replacing /z by j Ez + jE y = jHx y Ez = jH y jEx x E y Ex = jHz x y Similarly, the curl of H is H = jE Hz + jH y = jEx y Hz jHx = jE y x H y Hx = jEz x y

(15.43)

(15.44)

We manipulate these equations in such a manner that Ex , E y , Hx and H y are written in terms of the derivatives of Ez and Hz (this is left as an exercise for the reader to do) j Hz j Ez (k2 2) x (k2 2 ) y j Ez j Hz + 2 Hx = 2 2 ) x (k (k 2 ) y j Hz j Ez + 2 Ey = 2 2 ) y (k (k 2 ) x j j Ez Hz 2 Hy = 2 2 ) y (k (k 2 ) x Ex =

(15.45)

where k = (= 2/0 ). From these equations it is clear that both Ez and Hz cannot both be zero. We work with the Transverse Magnetic (TM) modes rst (Hz = 0). 1. Transverse Magnetic (TM) Modes (Hz = 0) Again we use the separation of variables technique3 . Starting from Ez we let Ez = XEz (x)YEz (y)ejz where XEz (x) = A cos(kx x) + B sin(kx x) YEz (y) = C cos(k y y) + D sin(k y y)

3 See

(15.46)

page 459

462

15. Waveguides

where A, B, C, and D are constants. Similarly all the other eld components have a similar functional dependence. Applying the boundary condition (tangential E, Ez = 0) on guide wall at y = 0 for all values of x and z is (as in the case of the parallel plate waveguide) C cos(k y y) + D sin(k y y) y=0 = 0 C cos(0) = 0 therefore C = 0 Similarly at y = b we set Ez = 0 D sin(k y y) y=b = 0 D sin(k y b) = 0 or k y = n/b therefore Ez = {A cos(kx x) + B sin(kx x)} D sin(k y y)ejz where k y = n/b (15.48) n=1,2,. . . (15.47)

We now proceed to apply the boundary condition at x = 0 and x = a in a similar manner. Setting Ez = 0 at x = 0 for all values of y and z {A cos(kx x) + B sin(kx x)}|x=0 = 0 and for x = a we must have B sin(kx x)|x=a = 0 or kx = m/a Hence Ez is Ez = BD sin(kx x) sin(k y y)ejz m n = Ez0 sin x sin y ejz a b which implies that A = 0 (15.49)

m=0,1,2. . .

(15.50)

(15.51)

463

15. Waveguides

plugging in this value of Ez into Equation set 15.45 j m n m cos x sin y ejz a (k2 2) a b j m n n sin x cos y ejz Hx = Ez0 2 2 ) b (k a b j n n m x cos y ejz E y = Ez0 sin 2 2 ) b (k a b j m m n H y = Ez0 cos x sin y ejz 2 2 ) a (k a b m 2 n 2 m = 1, 2, . . . 2 2 + where k = + a b n = 1, 2, . . . Ex = Ez0

(15.52) (15.53)

In these formulae neither m nor n is zero since then Ez is zero and all the elds become zero. For the rectangular waveguide there is no TM01 or TM10 mode 2. Transverse Electric (TE) Modes (Ez = 0) Here we start with Hz Hz = XHz (x)YHz (y)ejz where XHz (x) = A cos(kx x) + B sin(kx x) YHz (y) = C cos(k y y) + D sin(k y y) (15.55)

(15.54)

where A , B , C , and D are constants. Since the boundary condition of Hz yields no particular results, we obtain the other eld components from Equation set 15.45. j j Hz XHz (x)YHz (y)ejz = 2 (k2 2 ) y (k 2 ) j Hz j X (x)YHz (y)ejz Hx = 2 = 2 2 ) x (k (k 2 ) Hz j j Hz X (x)YHz (y)ejz = 2 Ey = 2 2 ) x (k (k 2 ) Hz j Hz j Hy = 2 = 2 XHz (x)YHz (y)ejz 2 ) y (k (k 2 ) Ex = Where

XHz =

(15.56)

(15.57)

464

15. Waveguides

Applying the boundary condition on Ex at y = 0 and y = b we nd that D = 0 and k y = n/b n = 0, 1, 2 . . .. Similarly applying the boundary condition on E y at x = 0 and x = b we nd that B = 0 and kx = m/b m = 0, 1, 2 . . .. When we apply the boundary conditions on the other magnetic eld components (the normal components of the magnetic eld are zero on the walls of the waveguide) we do not go any further since H y Ex and Hx E y . Therefore Hz = A C cos(kx x) cos(k y y) = Hz0 cos(kx x) cos(k y y) m m = 0, 1, 2, . . . kx = a n ky = n = 0, 1, 2, . . . b The complete elds are j m n n cos x sin y ejz 2 2 ) b (k a b j m n m sin x cos y ejz Hx = Hz0 a (k2 2 ) a b j n m m x cos y ejz E y = Hz0 sin a (k2 2) a b j n n m x sin y ejz H y = Hz0 cos 2 2 ) b (k a b m = 0, 1, 2, . . . 2 2 m n 2 2 where k = + + n = 0, 1, 2, . . . a b but both m, n 0 Ex = Hz0 = k2 m a

2

(15.61)

(15.62)

These elds and equations do appear complex! The propagation constant in the waveguide is n b

2

(15.63)

Each mode propagates with a certain cut-o frequency. Since the mode starts propagating when = 0, therefore kcmn = cmn = cmn = c cmn = c m a m a m a m a

2

+

2

n b n b n b n b

(15.64)

+

2

+

2

(15.65)

where cmn is the cut-o radian frequency for the m, nth TE or TM mode and c

465

15. Waveguides

is the velocity of light in the medium lling the waveguide. The dominant mode (that is the rst mode with the lowest cut-o frequency) is the TE10 mode. For example for a waveguide with a = 2b = 1 cm, m 2 n 2 + c10 = c a b

m=1, n=0

8

For the next higher order mode, for this waveguide, is the TE01 mode. Its cuto frequency is m 2 n 2 + c01 = c a b

m=0, n=1

the cuto wavelength is 2a. The elds of the TE10 mode are Hz = Hz0 cos Ex = 0 Hx = Hz0 j sin x ejz 2 2 ) a (k a j sin x ejz E y = Hz0 2 2 ) a (k a x ejz a (15.66)

Hy = 0

(15.67)

Degenerate modes are those TE and TM modes which have the same cuto frequency but have dierent values of m and n. For example the TE20 (m=2, n=0) mode and the TE01 (m=0, n=1) for the a = 2b waveguides are degenerate.

466

15. Waveguides

a x z

y

Figure 15.5.: The circular waveguide

The circular waveguide is a hollow pipe of circular cross-section made of some high conductivity material like copper. To analyse such a waveguide we would have to write the Helmholtz equation in cylindrical coordinates. The wave equation in cylindrical coordinates applied to a scalar is 2 1 2 2 1 + + + = k2 2 2 2 z2 (15.68)

where k = . We use the same technique as earlier, namely the method of separation of variables. is written as the multiplication of three functions (, , z) = R()F()ejz Substituting this function in the previous equation 2 1 2 2 1 + + + = k2 2 2 2 z2 1 2 2 2 R()F()ejz + 2 2 R()F()ejz + 2 R()F()ejz 2 z 1 + R()F()ejz = k2 R()F()ejz which becomes F()ejz 2 R() R()ejz 2 F() + + ( j)( j)R()F()ejz 2 2 2 + F()ejz R() = k2 R()F()ejz (15.70) (15.69)

(15.71)

467

15. Waveguides

dividing throughout by R()F() exp( jz) 1 R() 1 2 F() 1 2 R() + + k2 2 = 0 + 2 R() 2 R() F() 2 multiplying this equation throughout by 2 R() 2 2 R() + + 2 k2 2 2 R() R()

A function of

1 2 F() =0 F() 2

A function of

hence since the two parts are functions of dierent variables therefore 1 2 F() = n2 F() 2 R() 2 2 R() + + 2 k2 2 = n2 R() 2 R() the rst of these equations integrates to F() = A cos(n) + B sin(n) (15.73) (15.72)

where n = 0, 1 . . .. n must be an integer since for any xed value of , as we increase n from 0 to 2 the elds must repeat themselves: they must be identical from 2 to 4, 4 to 6, ... as they were from 0 to 2. The second equation becomes 2 R() 1 R() n + + R() k2 2 = R() 2 if we let = k2 2 then 2 R() 1 R() n + + R()2 = R() 2 or or or 2 R() 1 R() n + + R() = R() 2 2

2 2

(15.74)

2

=0

(15.75)

468

15. Waveguides

1 0.8 0.6 0.4

Jn (x) J1 (x) J2 (x) J3 (x) J0 (x)

0.2 0

x

Figure 15.6.: The rst few Bessel functions of the rst kind

10

12

14

dened by 2n Zn x dZn Zn+1 Zn1 = 2 dx Zn+1 + Zn1 = and the dierential equation which denes Bessel functions is x2 dy d2 y + x + (x2 n2 )y = 0 2 dx dx (15.77)

(15.76)

This equation, being a second order equation has two solutions, namely, Bessel functions of the rst and second kind. Bessel functions of the second kind are innite at the origin, so will be of no use to us (None of the eld components are innite anywhere within the guide). Bessel functions of the rst kind are depicted in Figure 15.6 for the rst few orders. The notation for these functions is Jn (x) (the Bessel function of the rst kind of order n). The very rst Bessel function is 1 at the origin while it oscillates and decays slowly when the value of x is increased. All the other higher order Bessel functions of the rst kind are zero at the origin while showing similar behaviour as we move away from the origin. Keeping in mind the discussion given above, each eld component has the functional form E , E . . . Hz = A cos(n + )Jn()ejz (15.78) where A and are constants; = k2 2. By orienting the x and y axes correctly, may be set to be zero. The other eld components in terms of Ez

469

15. Waveguides

and Hz are j Ez j Hz 2 2 j Ez j Hz + 2 E = 2 j Hz j Ez 2 H = 2 j Ez j Hz 2 E = 2 H = 1. Transverse Magnetic modes (TM) (Hz = 0) In the transverse magnetic modes we start with Ez . Ez = Ez0 cos(n)Jn ()ejz (15.80)

(15.79)

from which we derive the other eld components. The boundary condition on Ez is Ez |=a = Ez0 cos(n)Jn (a)ejz = 0 which means that Jn (a) = 0 (See Figure 15.6 paying attention to the zeros of J0 (x) and J1 (x) ) First zero 01 a = 2.405 11 a = 3.85 Second zero 02 a = 5.52 12 a = 7.02 (15.81)

TM01 (01 a = 2.41), TM11 (11 a = 3.85), TM02 (01 a = 5.52) and TM12 (01 a = 7.02) (15.82) modes, starting from the lowest to the highest modes. Taking a particular example that of the TM01 mode, since 01 a = 2.405 01 = k2 2 = 2.405/a 01 01 = k2 (2.405/a)2 (15.83)

When k > 2.405/a we have a propagating mode, otherwise we have an evanescent (or cuto) mode.

470

15. Waveguides

The other eld components are jn sin(n)Jn ()ejz 2 jn E = Ez0 2 sin(n)Jn ()ejz j cos(n)Jn ()ejz H = Ez0 j E = Ez0 cos(n)Jn ()ejz H = Ez0

(15.84)

Here Jn () is the Bessel function dierentiated with respect to its argument. 2. Transverse Electric Modes (TE) (Ez = 0)

Here

(15.85)

and the other eld components are j cos(n)Jn ()ejz j E = Hz0 cos(n)Jn ()ejz jn H = Hz0 2 sin(n)Jn ()ejz jn E = Hz0 2 sin(n)Jn ()ejz H = Hz0 Applying the boundary condition on E E |=a = Hz0 Then j cos(n)Jn (a)ejz = 0

Jn (a) = 0

(15.86)

(15.87)

(See Figure 15.6 paying attention to the maximas and minimas of J0 (x) and J1 (x) ) First max or min a = 3.83 01 a = 1.84 11 Second max or min a = 7.02 02 a = 5.33 12

n=0 n=1

These waves are the TE11 , TE01 , TE12 and TE02 modes, starting from the lowest to the highest modes. Comparing the TE and TM modes, we can see that the TE11 ( a = 1.84) is the 11 dominant mode of the circular waveguide.

471

16.1. Wave Equation due to Charges and Currents1

The above discussion applies well to the case where the wave is travelling in a region free of currents and charges, but what about the case when the the wave is produced by currents and charges? In this case we have to proceed along a slightly dierent path because the problem becomes more complex. Let us start from Equation 12.65 on page 372 B = 0 Since the divergence of B is always zero, B must be equal to the curl of some vector2 B = A (16.1) A has a special name it is called the vector potential. Since B = H H= 1 A (16.2)

1 A (16.3)

= j A Therefore E + j A = 0 E + jA = 0

E + jA = E = jA

(16.4)

where the term is placed on the right for the general case, since the curl of the gradient is always identically zero3.

section can be read at a later point in time, when antennas are considered is always zero. 3 i.e., E = jA = j A

1 This

2 ( A)

472

Substituting these equations in Equation 12.66 on page 372 H = jE + J On simplifying 1 A = j jA + J (16.5)

( A) = 2 A j + J ( A) ( ) A = 2 A j + J

So far the curl of A has been dened through Equation 16.1 but not its divergence and a vector eld is not fully dened if both are not dened. Therefore we use a particularly appropriate form of the divergence of A to simplify the above equation 4 : A = j (16.7) By cancelling ( A) with j we get ( ) A = 2 A + J (16.8)

Which is the frequency domain wave equation with a source term J. Once A is obtained, we can get H from Equation 16.2. From H, applying Maxwells Equation 12.66 we can get E. More will be said about this when we actually apply the previous equation. Concentrating now on the previous equation Equation 16.8 becomes (k2 = 2 ) 2 A 2 A 2 A + + 2 k2 A = J x2 y2 z Which is three equations 2 Ax 2 Ax 2 Ax + + k 2 Ax x2 y2 z2 2 A y x2 + 2 A y y2 + 2 A y z2 k2 A y = = = Jx J y Jz (16.10) (16.9)

2 Az 2 Az 2 Az + + k 2 Az x2 y2 z2

We will use the equations derived in this chapter in the forthcoming chapters on propagation and radiation since propagating waves as well as radiating waves are governed by these equations. Next we consider radiation from sinusoidally varying currents, for which we need to apply the equation derived in Section 16.1. In that section, the essential

4

473

equations involved in the computation of elds when the source consists of a sinusoidally varying current density owing in a conductor are derived. The essential equation, given in terms of the vector potential, is Equation 16.8, and which is reproduced here ( ) A = 2 A + J = k2 A + J

(16.11)

z J(r )

Field Point

rr

dV

x

Figure 16.1.: Radiation from a current source

Figure 16.1 shows the region of application of the equation A is the vector potential at the eld point r. r is usually far away from the source. If the source is of the order of a few wavelengths, the eld point is usually much more than tens of wavelengths away. (The importance of the vector potential A is that B = A) J is the current density, which is the source term, shown shaded in the gure. Usually J ows in the metal of the conductor comprising the antenna. The solution to this equation, the vector potential, A(r), is A= 4 where J(r ) is the source current density phasor from which the wave emanates is a function of primed coordinates: the position vector r J(r )ejk|rr | dV |r r |

(16.12)

474

The integration is to be performed over the volume containing the sources which is indicated by dV In this equation V is the region where the currents lie. Let us examine this equation in a little more detail. If we convert the integral to a summation (which is the reverse process) then A 4 J(r )ejk|rr | V |r r |

Which states that the vector potential A consists of a summation of outgoing spherical waves of the type J(r )ejk|rr | |r r | This factor is a spherical wave with centre r (due to the term) ejk|rr | the wave is streaming away from the source J(r ) located at r . The wave amplitude diminishes far away from the source and is proportional to 1 |r r | The total wave comprising the vector potential is a vector sum of all these tiny wavelets. Area = A

J

Figure 16.2.: A current carrying conductor

Let us apply this equation to the most simple of all cases, namely, a small current lament. But before we do this, we need to convert the equation for the vector potential A to another equation which does not contain the current density but rather one which contains a current: in other words, how do we use the current I instead of the current density J in Equation 16.12? To do this, consider Figure 16.2 where a small shaded section of a current carrying conductor is shown. The current density in the conductor is J assumed to be uniform over the cross-section. Consider the term JdV Since dV = dSdl, where dS is an element of the area of cross-section and dl is a small length along the wire, JdV = JdSdl

475

integrating now over the area of cross-section, A, of the wire JdSdl = JAdl = JA Jdl = Idl (I = JA; dl = Jdl)

where I is the current in the wire and dl is a small section of the wire in the direction of the current ow. (The integral over the cross-section is JdS = JA since J is constant over the cross-section). The element dl may be treated as a very small vector. Equation 16.12 becomes A= 4 I(r )ejk|rr | dl |r r |

(16.13)

EXERCISE 16.1 Find the current density J in a wire of diameter 1 mm carrying a current of 1 A.

z Field Point

r Il y

x

Figure 16.3.: An elemental wire carrying a current

Now we are in a position to apply Equation 16.13 to a small wire carrying a current. In particular, we apply the above equation to an elemental wire of length l carrying a current I as shown in Figure 16.3. The gure shows the orientation

476

of the wire in the z direction

dl laz

(16.14)

Also in the equation for the vector potential A the integration sign may be removed since the region of integration is small: 4 4 I(r )ejk|rr | dl |r r | az (l)

A=

L Iejk|rr |

|r r |

Because the wire is very small lengthwise, the value of r r |r| since r r (r 0).Therefore r r |r| = r (16.15)

The r being considered in the above equation is the r coordinate of spherical coordinates. Therefore A = Iejk|rr | az (l) 4 |r r | I (l) ejk|r| az 4 |r| I (l) ejkr az 4 r

So far we have been considering a mixed coordinate system. In the above equation I (l) ejkr A = a z Az az 4 r Expressing Az in spherical coordinates we directly obtain (See gure) Ar = Az cos I (l) ejkr cos 4 r I (l) ejkr sin 4 r (16.16)

A = Az sin A = 0

477

Az Ar = Az cos az a A = Az cos(/2 + ) = Az sin

Figure 16.4.: The relationship between Az , A and Ar

ar /2

B = A

1 (rA ) Ar a a + r r (16.17)

The above equation is the equation of the curl in spherical coordinates. We can proceed to simplify it because we know that A = 0. Therefore A A A 1 1 Ar r 1 sin B= ar + sin r r sin r = 1 (rA ) Ar a a + r r

Next we know that neither of A nor Ar are functions of 1 1 Ar 1 1 (rA ) Ar A ar + a a + r sin r sin r r 1 (rA ) Ar = a (16.18) r r

B=

478

Therefore the rst term is sin (rAz ) 1 (rAz sin ) = r r r r jkr sin I (l) e = 4r r = jk sin I (l) ejkr 4r and the second term is Az (r) sin 1 (Az (r) cos ) = r r = sin I (l) ejkr 4 r2

(16.19)

therefore the magnetic ux density B at a eld point (r, , ) due to a current Ilaz located at the origin at r = 0 is B= sin [I (l)] jk 1 jkr e a + 4 r r2 (16.20)

We know the relation between B and H: B = H so H is H= Using Maxwells equation H = jE 1 1 Hr rH H 1 sin H jE = ar + r sin r sin r + 1 (rH ) Hr a r r a sin [I (l)] jk 1 jkr + e a 4 r r2 (16.21)

Where the right hand side is the curl of H in spherical coordinates. Since

479

H = Hr = 0 & H & H & 1 sin H 1 1 &r rH jE = ar + sin r r sin r & &r H H & 1 r& a r r 1 rH 1 sin H = ar + r sin r r +

Term 1 Term 2

(16.22) (16.23)

Considering each term: the rst term (Term 1) is 1 sin H r sin sin2 [I (l)] jk 1 jkr 1 + e ar ar = r sin 4 r r2 =

and the second term (Term 2) is 1 rH r r sin [I (l)] 1 1 jk + ejkr a = r r 4 r = sin [I (l)] 4r r jk + 1 jkr e r = a a

sin [I (l)] 2 jk 1 jkr k 2 e a 4r r r 2 jk 1 sin [I (l)] k = 2 3 ejkr a 4 r r r Hence E= H [I (l)] = j j cos jk 1 jkr sin k2 jk 1 jkr + 3 e e ar a 2 r2 r 4 r r2 r3 (16.24)

Generally we will be interested in the far elds. Neglecting the components which are proportional to 1/r2 and 1/r3 (which become small as r becomes

480

z

ar H E

y

E

x

H ar

large) we nd that cos sin 2 [I (l)] 2 3 jkr 2 3 jkr jk/r + 1/r e ar (k /r jk/r 1/r )e a E= j 2 4 neglected

neglected

(16.25)

neglected

(16.26)

so the electromagnetic elds are E H = k2 sin [I (l)] / j4r ejkr a = jk sin [I (l)] /4r ejkr a (16.27)

The elds are shown in Figure 16.5. Some important points are to be noticed in the above equations

481

1. Let us compute the ratio jk sin [I (l)] jkr E k2 sin [I (l)] jkr = e e / H j4r 4r ( jk)( jk)/ j k2 /( jk) = j jk k = = 0 = (For free space) = 0 = = Z0 (Characteristic impedence for free space, 377 ) which is the same result that we got in the case of a plane wave travelling in free space. 2. Let us compute the Poynting vector which denotes the direction of power ow. Since we are considering phasors the formula for the Poynting vector is 5 P= substituting the values of E and H jk sin [I (l)] jkr 1 k2 sin [I (l)] jkr e a e a 2 j4r 4r

1 (E H) 2

P=

where the star indicates complex conjugation. Converting all the j s to j s in the second bracket P= jk sin [I (l)] jkr 1 k2 sin [I (l)] jkr e a e a 2 j4r 4r

In the next step of the computation, we do further simplication 1 k sin P = Z0 I (l) ar 2 4r l sin 1 I = Z0 2 2r

2 2

a. The power streams out in the ar direction. It is proportional to the squares of Il and k; and inversely proportional to the square of r.

5

Note that since we are not considering rms values, the factor of 1/2 is present

482

z

1 0.5 0 0.5 1 1 0.5 0 0.5 1 0 0.5

x

1

1.5

Figure 16.6.: Polar plot of the normalised power radiated by an innitesimal current element.

b. The power streams out directionally: at = 0 and no power is radiated, while at = /2 maximum power is radiated. This 3-dimensional half-plot of the normalised power density (W/m2 ) as a function is shown in Figure 16.6. The power radiated per steradian (solid angle) is6 1 l sin U(, ) = P(r, , )r2 = Z0 I 2 2

2

3. The total power radiated is the ux of the Poynting vector through a sphere of radius r PT = = Since P = U/r2 therefore PT = = U(, ) sin dd

=,=2 =0,=0

P dS

1 l sin I Z0 2 2

sin dd

6A

solid angle subtended by some area at the centre of a sphere is the area on the surface of the sphere divided by the square of the radius. Thus an elemental solid angle d = dS/r2 = sin dd

483

Carrying out the integration over which is 2 and collecting terms l 1 PT = Z0 I 2 2 = Z0 I l 2

2 =,=2 =0,=0 2 = =0

sin3 dd

sin3 d

2 2 = =0 2

sin3 d

l Z0 2

I 2

4 3

Here I/ 2 is the rms value of the current, and the rest, Rr , is a resistance called the radiation resistance, which is= (total power radiated)/(rms current squared):

2 PT = Irms Rr

(16.28)

2

Rr =

l 2 Z0 3

2

= 80

(16.29)

The radiation resistance of an antenna represents the total power radiated away. If we examine the gure we can see that the source supplies power to the antenna and sees a loss. Therefore the source and antenna can be modelled as a Thevenin equivalent circuit, which is shown in Figure 16.7. Referring to the gure, Rs is the internal resistance of the source and the antenna is modelled as a lossy resistance, Rr , the radiation resistance, which is represents the total power radiated away.7

Equivalent circuit of antenna Wave

Irms

Antenna

Rs Irms

Source

Rs Rr

Source

7 In

actual practise the antenna is modelled as an impedance Z = Rr + jX. See for example Kraus (1988) Jordan & Balmain (1968) for details.

484

Dipole z z Current Profile

z=L

r = z az

Insulator Rs

r r= r ra r

Distant Point

I(z) z = z

y

r = z az

th

Source

(a) x (b)

Figure 16.8.: The dipole antenna. (a) close-up details (b) Details with respect to the far eld

The dipole antenna consists of two thin but sti wires of equal length (= L), and which are bent into a shape as shown in Figure 16.8. The ends of the two wires are embedded in an insulator which insulates them from each other, and are fed by a source of microwave frequency. Generally the feeder is coaxial line, the centre conductor of which is connected to one of the wires of the dipole, while the outer conductor is connected to the other. Since the wires are limited in extent, the currents at the ends must be zero. Also the currents are sinusoidally distributed on each of the wires (analogous to the case of transmission lines). The most probable current distribution on the upper wire is therefore I = I0 sin(kz + ) where k is the free-space propagation constant (= 2/). Now at z = L, I = 0 so sin(kL + ) = 0, or kL + = 0, or = kL, or kL So I = I0 sin(kz kL) or I = I0 sin(kz + kL)

Function 1 Function 2

Now observing the two functions we know that in the rst function (Function 1) as z is decreased from its value at the end, that is z < L, then the argument of the sine function k(z L) becomes negative and the value of the current, I, is negative. However, this is not true for the second function (Function 2).

485

Therefore, on the upper conductor I = I0 sin(kz + kL) = I0 sin [k(L z)]

for 0 z L

(16.30)

By the same reasoning for the lower conductor the current, I, is I = I0 sin [k(L + z)] for L z 0 (16.31)

Once we have obtained the current distribution on the antenna, we can proceed to apply Equation 16.13 on page 476 A= 4 I(r )ejk|rr | dl |r r | B

z

z=L r = z az z= z A C

os zc r | r |r

Distant Point

D

r=

ra r

th

z cos

x Dipole

z = L

Two approximations are made, which are both quite reasonable: (please refer to Figure 16.9) (1) The term |r r | is approximated by r z cos . How? We notice from the 486

In this equation we use the value for the current we just derived earlier, and r and r are shown in Figure 16.8(b). r = ar r and r = az z . The equation then becomes: z =L I(z )ejk|ar raz z | A= az dz 4 z =L |ar r az z |

gure that r = ar r (spherical coordinates), and r = az z . But the magnitude |r r | is the distance AB or (approximately) the distance CD since AB CD,8 therefore r r CD = OD OC = r OC = r z cos az A 4 I(z )ejk(rz cos) dz r z cos z =L

z =L

(2) In the second part of the approximation the denominator r z cos r since the integral with or without this approximation gives practically the same result. With the second approximation A az 4 I(z )ejk(rz r z =L

z =L z =L z =L

cos)

dz (16.32)

az ejkr = 4r

I(z )e jkz

cos

dz

z =0 z =L

Current for Lz 0

cos

dz

az ejkr + 4r

z =L z =0

Current for 0z L

cos

dz

(16.33)

z =0 z =L

az ejkr 4r

cos

z =L z =0

az ejkr 4r

cos

dz = (16.35)

8 To

corroborate this, let L = 1m, r = 1000m and = /2 then AB = 1000.0005m and CD = 1000m. For the same case for = 0 AB = 999m and CD = 999m.

487

Adding these two integrals we get A= = az ejkr I0 4rk sin2 az ejkr I0 4rk sin2 ej k cos L + e j k cos L 2 cos(kL) {2 cos(kL cos ) 2 cos(kL)} (16.36)

Since we are interested in a half-wave dipole, we set L = /4 or kL = (2/)(/4) = /2 az ejkr I0 cos( cos ) 2 A= (16.37) 2rk sin2 We now use the Equation set 16.16 on page 477, which gives the value of Ar and A in terms of Az , to give Ar = ejkr I0 cos( cos ) cos 2 2rk sin2 ; A = ejkr I0 cos( cos ) 2 2rk sin (16.38)

and then use the denition of the curl, in the spherical coordinate system, (Equation 16.17 on page 478) H = A =

1 (rA ) Ar a r r

(16.39)

where we have dropped all the terms which are zero9 . Next we examine each of the two terms carefully. The term (1/r)(Ar/) leads to a 1/r2 term, (by inspection) which may be dropped (since the eld is expected to decay as 1/r as in the case of the innitesimal current element). Therefore jejkr I0 cos( cos ) 1 (rA ) 2 a = a H = r r 2r sin jejkr I0 cos( cos ) 2 H = 2r sin 1 sin H jE = r sin

Term 1

(16.40)

Next we obtain E from the magnetic eld using Equation 16.23 on page 480 1 rH ar + r r

Term 2

From inspection, Term 1 decays as 1/r2 , and can be dropped, as before. On the

9 these

488

other hand 1 rH jE = r r E = kejkr I0 cos( cos ) 2 a = a 2r sin

(16.41)

z

1 0.5

y

0 0.5 1 1 0.5

Dipole

0 0.5 1 0 0.5

x

1

1.5

Figure 16.10.: Magnitude of the electric eld radiated by a half-wave dipole, shown as a polar plot. Only half the plot is shown.

From the computation of these elds it is clear that (1) The electric and magnetic elds are proportional to each other; (2) Far away from the source (which is the dipole) the elds constitute a local plane wave: that is the electromagnetic elds are perpendicular to each other and also perpendicular to the direction of travel. To be specic E is perpendicular to H and both are perpendicular to ar , the direction of travel; and (3) In dierent directions the amplitudes of the elds are dierent. That is, in the directions of = 0, there is no radiation at all10 , while in the direction = /2 there is maximum radiation. These remarks are illustrated in Figure 16.10.

10 consider

489

Did you know? To measure a sinusoidally varying electric eld, probes are used. Probes are designed to measure the electromagnetic eld with minimal of the eld perturbation. To accomplish this, they are physically and electrically small as possible so that coupling with the eld of interest is minimised. Eeld probes usually use an electrically small dipole coupled with a diode as a sensora

a Making Better Electric Field Measurements, Steve King, White Paper available on http//www.ets-

lindgren.com

The Poynting vector can now be calculated. The Poynting vector at a distant point is 1 P = E H 2 1 = (E a ) H a 2 1 1 = (E a ) ( E a ) 2 Z0

H

1 (E E )ar 2Z0

(since ar = a a )

(16.43)

1 Z0 I0 cos( 2 cos ) 2Z0 2r sin 2

ar =

2 Z0 I0

82 r2

ar

W/m2

(16.44)

Thus we nd that the power streams radially away from the origin, and it is also directional in nature. We are also now in a position to compute the radiation resistance of the antenna, by rst computing the the total power radiated by the dipole. The total power radiated, PT , is the total outward ux of the Poynting vector integrated over a sphere of radius r where r 11 PT =

Sphere

P dS

(16.45)

11 r

490

Choosing such a sphere, PT =

2 Z0 I0 =, =2 =0, =0

82

2 Z0 I0

= =

82

2 Z0 I0

2

=

=

ar ar r2 sin dd

dS 2

sin d

=0

J

=0

=

J=

=0

d = 1.218827

(16.47)

So PT =

2 Z0 I0

1.218827

2

I0 60 2

1.218827

2 = 73.12962Irms (W)

2 Rr = PT /Irms 73

(16.49)

If P(r, , ) is the magnitude of Poynting vector, then the normalised power pattern of the antenna, Pn (, ) = P(r, , )/Pmax is

82 r2 82 r2

2 Z0 I0 2 Z0 I0

Pn (, ) =

max

(16.50)

To nd the maximum of the denominator, we plot the function for the range 0 . The function has zeros at = 0, and has a single maximum at

491

= /2, which is equal to one. (See Figure 16.10 on page 489) That is cos( cos ) 2 sin So: Pn (, ) =

2

=

max

2

=1

=/2

(16.51)

Normalised Power Radiated by a HalfWave Dipole 1 0.9 Normalised Power Radiated 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0

/6 /3 /2 2/3 5/6

Antennas, or aerials, are structures designed to radiate (or receive) electromagnetic radiation in a given direction or directions. Usually antennas operate in air or vacuum (as in the case of antennas used in outer space) but may be also operated in submarine mode, from under water. As in the case of a half-wave dipole, when an antenna is used in the radiating mode, the antenna is excited by a source of electromagnetic waves, and the antenna radiates the radiation as eciently as possible. In this chapter we will look at the fundamentals of antennas and those concepts which will be useful to a communication engineer. Typical antennas are shown in Figure 16.12. When an antenna radiates, it does so directionally. That is it radiates more in one direction than another one. The far eld electromagnetic elds, in the

492

(a) Dipole

(c) Yagi

spherical coordinate system and in phasor notation, are invariably of the type K fn (, ) jkr e r K fn (, ) jkr e H(r, , ) = Z0 r E(r, , ) = where K is a complex constant; fn (, ) is a real function whose maximum value for all values of and is unity. For example, fn (, ) for an innitesimal current element is sin and in the case of the half-wave dipole it is cos[(/2) cos]/ sin Z0 = 377 = 120, the intrinsic impedance of free space and k is the free-space propagation constant. As one can see, the wave decays far away from the source as 1/r. The directions of E and H are such that the electric eld is perpendicular to the magnetic eld, and both are perpendicular to ar . And furthermore, the Poynting vector, P = E H is in the direction of ar . That is P = P(r, , )ar (W/m2 ) (16.53)

(16.52)

And if is the angle (which is /2) from the electric eld vector to the magnetic eld vector, then 1 1 K fn (, ) jkr P = EH sin = e 2 2 r (/2) K fn (, ) jkr e Z0 r

493

z Direction of propagation

ar r E=

x

H E

y

Antenna

K fn (,) jkr e , r

H=

K fn (,) jkr Z0 r e

Main lobe

Back lobes

y Side lobes

Figure 16.14.: The power pattern of an antenna

A typical power pattern of an antenna, in a 3-D polar form, is given in Figure 16.14. The gure shows the main beam of an antenna pointing in the direction of the x-axis, and having sidelobes or minor lobes pointing in other directions. Some radiation is also in a direction opposite to the main lobes, and these are called back lobes. Figure 16.15 shows the same pattern in 3-D rectangular coordinates. Very often the patterns are shown in the principal planes. The principal planes are two perpendicular planes which pass through the axis of the main beam. Thus the principal planes of the pattern given above are the xy (z = 0) or xz ( = 0) planes, and the electric eld or power pattern is often given in these planes, and the eld or power functions are often given in terms of normalised

494

6 5 4 3 2 1 0 1 2 0 1

Side lobe

0.5

1.5

2.5

Figure 16.15.: Power pattern of the same antenna shown in 3-D rectangular coordinates

functions En (r, , ) = Pn (r, , ) = E(r, , ) E(r, , ) maximum value P(r, , ) P(r, , ) maximum value =

K fn (,) jkr e r K jkr re 2

= fn (, )

(16.55) (16.56)

= fn (, )

both these functions are of course, unitless. We may recover the electric eld and the power pattern function from E(r, , ) = Emax En (r, , ) P(r, , ) = Pmax Pn (r, , ) The radiation intensity is dened as U(, ) = r2 P(, ) (Units= W/str) (16.59) (V/m) (W/m )

2

(16.57) (16.58)

16.5. Directivity

The directivity of an antenna is dened by the formula D= Maximum power density radiated (W/m2 ) Average power density radiated (W/m2 ) (16.60)

The maximum power density radiated, will be in some specic direction given by (0 , 0 ). The average power radiated by an antenna, Pave , is Pave = 4r2 (=area Total power radiated (W) of the surface of a sphere) (W/m2 ) (16.61)

495

and the total power radiated is an integral of the Poynting vector over the same sphere PT =

sphere of radius r

P(r, , ) r2 sin dd

an element of area

(16.62)

"Area" AngleAngle

rd dA = r2 sin dd

Solid angle d = sin dd

r sin d y

x

Figure 16.16.: Denition of a solid angle

the area r2 sin dd subtends a solid angle d = sin dd at the centre of the sphere. The analogy is the circle. An arc rd subtends an angle d at the centre of a circle, in the same way, the area r2 sin dd subtends a solid angle sin dd at the centre of a sphere. These ideas can be better understood with reference to Figure 16.16.) Therefore if is the solid angle subtended at the centre of the sphere, then PT =

sphere of radius r

= and

1 4

Pmax Pn (, )d

(16.64)

Pmax Pn (, )d

4 Pn (, )d

(16.65)

496

the term on the right, in the denominator, is called the beam area, Beam area = = so D= note that the directivity, D 1. 4 Pn (, )d

(16.66)

Of what use is the directivity? If we examine the denition of the directivity, Equation 16.60, we nd that the denominator gives us the power density of an isotropic radiator, that is one where the power spreads equally in all directions. Pave = Pisotropic = Total output power (W) 4r2 (W/m2 )

so if Pmax is the power density at a distance r in the direction of the main beam, (0 , 0 ), then Pmax = DPisotropiuc (16.67) EXAMPLE 16.1 An antenna, which radiates 1 MW of power, has a directivity of 23 dB. Find the power density radiated in the direction of the main beam 10 km from the source. If the antenna was an isotropic radiator, it would radiate Pisotropic = The directivity in dB is DdB = 10 log10 (D) so D = 10(DdB /10) = 102.3 = 199.53 106 PT = 7.96 104 (W/m2 ) = 2 4r 4 (104)2

Using this value of directivity PMain beam = DPisotropic = 7.96 104 199.53 = 0.159 (W/m2 ) EXAMPLE 16.2 Find the directivity of an innitesimal dipole of length 0.1.

497

The directivity of the innitesimal dipole is D= Pmax Pav

1 sin 2 Z0 2 1 4 1 sin 2 Z0 2 2 sin max 3

I I

l l

2 max 2

sin dd

sin dd 4 3 = = 2(4/3) 2 The directivity is 1.5 regardless of the length. EXAMPLE 16.3 Find the beam area and directivity of an isotropic radiator. An isotropic radiator radiates equally in all directions. So Pmax = PT 4r2

1 4

where PT is the total power radiated. Also the average power radiated is Pave = So therefore D = Pmax /Pave = 1 The beam area may be calculated by considering Pn (, ) = 1 (Equal power density in all directions) therefore

=, =2

PT 4r2

(16.68)

=

=0, =0

Pn (, ) sin dd

=, =2

=

=0, =0

sin dd

=

= 2

=0

sin d

= 4 EXERCISE 16.2 Find the directivity of a half wave dipole. (Ans: 1.64)

498

Direction of maximum radiation Direction of maximum reception

0

Antenna y Antenna

0

y

0

x x

Antennas are reciprocal elements: the direction in which they radiate best, is the same direction in which they receive best. Thus if an antenna has its main beam in the direction (0 , 0 ), then when used as a receiving antenna it registers maximum radiation from the same direction, namely, the direction (0 , 0 ). This is well illustrated in Figure 16.17. When an antenna is used in the receiving mode, it is immersed in electromagnetic radiation coming toward it in the form of a plane wave12 from the direction (0 , 0 ). The Poynting vector magnitude of the plane wave is P watts/m2 . Hence to trap power (watts), the antenna must have what is called an eective aperture, Ae (m2 ). Therefore the received power, Pr , is P r = Ae P (W) (16.69)

It turns out that the eective aperture and directivity are linked by Ae = 2 D 4 (m2 ) (16.70)

for every antenna. With this formula we can derive the power received by an antenna when it is transmitted by another. Figure 16.18 illustrates the conguration of two antennas placed a distance r apart directed in such a way that the transmitting antenna points its main beam

12 Though

the wave is a spherical wave, coming from a great distance, it appears to have a plane phase front

499

Pr = Pt Dt Transmitting antenna Pt r

Figure 16.18.: Derivation of the Friis transmission formula

2 4r Dt Dr

Dr Receiving antenna Pr

in the direction of the receiving antenna, and the receiving antenna is oriented to receive maximum radiation. If the transmitting antenna of directivity Dt broadcasts Pt watts of power, then at the receiving end, the Poynting vector magnitude is Pt P(r) = Dt (W/m2 ) (16.71) 4r2 Let the receiving antenna have an eective area Aer then the power received, Pr , is Pt Pr = (W) (16.72) Dt Aer 4r2 or substituting the value of Aer Pr == Pt Dr 2 Dt 4 4r2

Aer

(16.73)

2

Dt Dr

(W)

(16.74)

This is the Friis transmission formula. EXAMPLE 16.4 Two half-wave dipoles are placed 1 km apart, and the transmitting antenna transmits 1 W of power at 100 MHz. Find the power received by the receiving antenna Both dipoles have directivities of 1.64. Dt = Dr = 1.64 The wavelength of the wave in air is 3 108 c = 3 (m) = f 100 106

500

the eective area of the receiving antenna is Aer = 1.64 Dr = 3 = 0.39152 (m2 ) 4 4

and the Poynting vector at the receiving antenna is P= Pt 1 D = 1.64 = 1.3051 107 (W/m2 ) 2 t 4r 4 (103)2 Pr = PAer = 1.3051 107 0.39152 = 51.11 nW

501

17.1. Chapter Goals

1. xxxx

17.2. Introduction

Every communication engineer has, at some point, to deal with using an antenna for the purpose of communication. To broadcast messages, an antenna has to convert a current which is the end point in the communication process to electromagnetic waves, which have to beamed toward a receiver. On the other hand to receive messages the antenna has to be properly oriented to receive maximum radiation. The various design criteria which the communication engineer has to take into account are illustrated by Figure 17.1. For the transmitting antenna

Is the shape of the beam properly designed? (Point A on the gure) Does the antenna have the required directivity? The distance to the

receiving antenna has to be taken into account so that the received power is sucient. (Point A on the gure) Is the polarisation correct? For example in some applications a circular polarisation may be required in case there are many receiving antennas with dierent polarisations. (Point A on the gure)

Is the sidelobe level as it should be? High sidelobes implies that the

power is radiated away in other directions (Point A) Is the antenna correctly matched to the circuit? An incorrect match implies that the transmitting circuit does not deliver adequate power to the antenna. (Point B on the gure)

Does the antenna have the required bandwidth? If the antenna has to

be used for high bandwidth applications then this is an important point. (Point A and B on the gure) For the receiving antenna

Does the antenna have the required directivity?

This point has already been made with respect to the transmitting antenna. (Point C on the gure) Is the polarisation correct? The receiving antenna has to match the polarisation of the transmitting antenna. (Point C on the gure)

502

Circuit

Circuit

Figure 17.2.: An equi-spaced linear array Is the sidelobe level as it should be? With high sidelobe levels the

receiving antenna will receive radiation from other sources which will behave like noise. (Point C) Is the antenna correctly matched to the circuit? The received power is precious. With improper match the power received from the antenna will not be delivered to the circuit (Point D)

Does the antenna have the required bandwidth? If the bandwidth

of the received beam is greater than the bandwidth of the receiving antenna then information will be lost. (Point C and D)

503

A linear antenna array is modeled as a series of isotropic radiators linearly placed (on the x axis for simplicity) with a inter-element1 spacing of d as shown in Figure 17.2. Since each radiator is identical, the electric eld of the far eld produced by each radiator, and when it is placed at the origin, is E(r) = E0t ejkr r (17.1)

where theexp( jt) term has been omitted. Equation 17.1 says that the far eld for each radiator is (a) independent of and coordinates of spherical coordinates, and hence by implication the radiator is isotropic; (b) the wave is propagating in the ar direction due to the exp( jkr) term, and (c) the radiation obeys the inverse square law due to the 1/r term. Furthermore, the vector E0t is a constant vector oriented in some arbitrary direction but transverse to ar . From studying equations from the last chapter, it should be clear the both the electric as well as magnetic elds are proportional to the current feeding the element, we can re-write Equation 17.1 as E(r) = I0 C t ejkr r (17.2)

where is the unit vector in the direction of E0t , I0 is the current feeding t the antenna and C is some constant. Since each radiator is identical we may normalise the far eld without loss of generality and write the far elds for each radiator as ejkr (17.3) E(r) = I0 t r Now the problem which we have to solve is that when the N radiators are placed in a straight line, as shown, what is the far eld pattern? Let us rst investigate the problem with only two such radiators. Let the rst radiator be placed at the origin of the x axis as shown in Figure 17.2, and let E1 be its far eld electric eld. Similarly let E2 be the far eld electric eld of the second radiator when radiating when at the position vector dax with respect to the rst. Then ejk|r| |r| jk|rax d| e E2 (r) = I2 t |r dax| E1 (r) = I1 t ET (r) = E1 (r) + E2(r) = I1 t

(17.4)

1 An

element is a radiator

504

where ET (r) is the total electric eld at a distant point. Now |r dax| = =r r2 2rd cos + d2 12 d d cos + r r

2

(17.5)

where the angle is the angle from the x-axis, and is not to be confused with the of spherical coordinates. Treating d/r as a single variable, for d r d |r dax| r(1 cos ) and r 1 1 d 1 (1 + cos ) |r dax| r(1 (d/r) cos) r r 1 r Using these approximations ET (r) = I1 t = I1 t = t ejk|rax d| ejk|r| + I2 |r| |r dax|

(17.6)

A B

We notice that there are two terms. Term A is common to both elements and it is the eld pattern of an isotropic radiator, while term B is what shapes the pattern. For now we will drop using the A term, knowing it is there and its importance will be discussed later. By convention, the B term is called E for the electric eld and it is the factor contributed by the geometry of the array (notice d and cos ). EXAMPLE 17.1 Calculate the error of |r r | being replaced by r r cos where is the angle between r and r . If the angle between r and r is , then and |r | = |r|/100 then r r = r2 + r2 2rr cos ( r r cos ) = r 1 + x2 2x cos x = r /r = 0.01 = r 1.0001 .02 cos comparing the values for cos = 1, 0, 1 ,

505

cos -1 +1 0 a = r r cos r(1.01) r(0.99) r b= r2 + r2 2rr cos r(1.01) r(0.99) r(1.0001) % of error 100 |(b a)|/a 0 0 0.01

Equation 17.7 was obtained using only two elements. If we consider N elements, then by simple logic E = I1 + I2 e jkd cos + I3 e j2kd cos + + IN e j(N1)kd cos (17.8)

Hence we can say with a great deal of condence that |r r | r 2rr cos .

17.4.1. The Array Factor

Let us consider Equation 17.8 and substitute equal current amplitudes, I0 , but phases which increase in the form of an arithmetical progression: 0, , 2, . . . (N 1). Then In = I0 (n 1) = I0 e

j(n1)

n = 1, 2 . . . N (17.9)

Then E = I0 e j0 + e j e jkd cos + e j2 e j2kd cos + + e j(N1) e j(N1)kd cos = I0 1 + e j(kd cos+) + e j2(kd cos +) + + e j(N1)(kd cos +) = I0 1 + e j + e j2 + + e j(N1) (17.10)

where = kd cos + . We realise that the above equation is a geometric progression. For a geometric progression, of a, ar, ar2 . . . arN1 the sum to N terms is 1 rN (17.11) a 1r In Equation 17.10, a = 1 and r = exp( j). So E= = = 1 e jN 1 e (Taking e jN/2 out of the bracket) e j/2

506

1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -4

-3

-2

-1

Figure 17.3.: (Top gure) Plot of sin(N/2)/ sin(/2) for N = 10 (Bottom gure) Numerator and Denominator of Equation 17.12

507

The function (1/N)| sin(N/2)/ sin(/2)| is plotted in Figure 17.3. The term 1/N is added to the function to normalise it, and the function is periodic with a period of 2. Let us further plot the numerator and denominator separately. (The lower plot) We can see from the gure that the numerator | sin(N/2)| for N = 10 has 10 maximas, where the function takes the value 1, and 11 minimas, where the function becomes zero in the range . The maximas are at N 3 = , , . . . , (2m + 1) , . . . for m = 0, 1, . . . 2 2 2 2 = , 3 , . . . (2m + 1) , . . . N N N

(17.13)

which means that the maximas are at odd multiples of /N. The minimas (zeros) of the numerator occur when N = 0, , 2, . . .m . . . for m = 0, 1, . . . 2 = 0, 2 , 4 , . . . 2m . . . N N N or it is clear the the zeros are at even multiples of /N.

(17.14)

The rst maximum of the numerator occurs at = /N. But this maximum does not gure in the eld pattern, it is absorbed in the main beam. The rst sidelobe maxima of the eld pattern occur at 3/N, as shown by the dashed line in the gure. Similarly, the second sidelobe maxima occur at 5/N and so on. Our interest in the sidelobes is mainly in the rst sidelobe adjacent to the principle lobe.

If we consider the zeros of the eld pattern or nulls, then the zeros of the numerator are necessarily zeros of the eld pattern, except where the denominator also has a zero. Near = 0 the numerator is sin(N/2) N/2 and the denominator is sin(/2) /2. At = 0 the eld pattern has a magnitude of N, which is a global maximum and constitutes the main beam or principal lobe.. The other zeros of the numerator are nulls in the eld pattern, which occur at = 2m.

We are also interested in answering the following points what is the direction of the main beam or principal lobe? the placement of the rst null and and eectively the beamwidth. the magnitude of the rst sidelobe with respect to the main beam. It should also be clear that the magnitude of the rst sidelobe can be obtained from this function (Eqn 17.12) but all placements have to be determined from = kd cos + (17.15)

508

The normalsied magnitude of the rst sidelobe is at Esl sin(N/2) N sin(/2)

=3/N

For large N, the rst sidelobe in a uniform is about 13.46 db down from the principal lobe EXAMPLE 17.2 Find the rst sidelobe level for N = 4. Step 1. The array factor is E= sin(2) 4 sin(/2)

Step 2. For the rst sidelobe, the sidelobe maximum is at 2 = 3/2. So Eslmax which is -9.031 dB down. We now look at several aspects of the uniform array. First and foremost what should be the values of the phase progression to point the beam at a particular direction 0 ? To do that we know from the eld pattern E() that the main beam points to = 0. So from Equation 17.15, kd cos 0 + = 0 or Hence, the array factor, Equation 17.12 becomes En () = = sin[N(kd cos + )/2] N sin[(kd cos + )/2] sin[Nkd(cos cos0 )/2] N sin[kd(cos cos0 )/2] (17.17) = kd cos 0 (17.16) 1 = 0.3536 4 sin(3/4)

2 20log(2/3) =

13.46

509

Let us nd the position of the rst null. From equation 17.14, the rst null is at = 2. Or kd(cos f n cos 0 ) = 2 (17.18)

Since f n is close to 0 , f n = 0 + f n . From Taylors series expansion of f (x) at the point x = a (x a)2 + ... 2! cos = cos(0 ) sin(0 )( 0) cos(0 )( 0 )2 /2 + . . . f (x) = f (a) + f (a)(x a) + f (a)

2

or

(17.19)

cos f n = cos(0 ) sin(0 )( f n 0 ) cos(0 )( f n 0 ) /2 + . . . or cos f n = cos(0 ) sin(0 ) f n cos(0 )( f n )2 /2 + . . . substituting in Equation 17.18, 2 kd(cos f n cos0 ) = N cos(0 )( f n )2 2 kd sin(0 ) f n 2 N

(17.20)

The negative sign is taken on the right since when increases, decreases. Now for a broadside array (where 0 = /2), sin 0 = 1, cos 0 = 0 f n = 2 = kdN dN (17.21)

Note that the width of the beam in inversely proportional to the length of the antenna, (= dN), also called the antenna apterture. For an endre array (where 0 = 0), sin 0 = 0, cos 0 = 1, f n = and the BWFN is 2 f n = 2 2 dN (17.24) 4 = kdN 2 dN (17.23)

Figure 17.4 shows a three dimensional view of the electric eld pattern. Notice that the pattern has circular symmetry since the radiation at an angle, , from a linear array is conical. In this pattern, kd = , = 0 and N = 10. Next we consider how the main beam points in dierent directions. If we change the value of in accordance with Equation 17.16, we can change the

510

Other sidelobes First sidelobe Main beam

0.1

0.05

-0.05

Antenna orientation

0.5

...

Figure 17.4.: 3-D view of a broadside pattern, where the antenna is oriented vertically. Note the dierence in the scale of the z-axis which has been broadened to show greater particulars of the main beam and sidelobe

beam pointing angle. This is shown in Figure 17.5. In the (a) part of the gure is set equal to zero (= kd cos 0 = kd cos(/2)) the beam points at the broadside direction. = 0 means that all the currents are in phase. Similarly, in (b) and (c) the beam is shown pointing at = 45 and = 0 . Notice how the beam broadens when pointing in the end-re direction (0 = 0 ). EXAMPLE 17.3 For 10 istropic radiators, with an inter-element spacing of 0.5, what is the progressive phase-shift, , to point the beam at = 45 angle? What is the BWFN of this antenna? Step 1. We apply Equation 17.16, = kd cos 0 Step 2. k = 2/, d = /2 so kd = . cos 0 = 0.7071. So = 0.7071 = 2.22r = 127 Step 3. To obtain the BWFN for this antenna, we solve Equation 17.20, sin(0 ) f n + cos(0 )( f n )2 2 2 Nkd

Step 4. The solution to this equation is (for 0 = /4, N = 10, and kd = ) f n = 0.2512r = 14.4 hence the BWFN is 28.8.

511

10 8 6 4 2 0 2 4 6 8 10 10 5 0 5 10

2 1 3

... N

Antenna

(a)

8 3

6 2 4 1 2

2 1 4 2 6

8 10 5 0 5 10

3 10 5 0 5 10

(b)

(c)

Figure 17.5.: Pointing the main beam in dierent directions. (a) 0 = /2, = 0 (b) 0 = /4, = 0.7071kd (c) 0 = 0, = kd. For all these plots, kd = .

512

z Field Point

x

Figure 17.6.: Figure to calculate the fareld pattern of current sources

EXAMPLE 17.4 Design a uniform antenna array with an inter-element spacing of 0.5, and the BWFN is to be 10 , and the beam should point at = 45 direction. How many elements should we use? If the BWFN is to be 10 then f n = 5 . The required equation is Equation 17.20. (0.7071)(0.0873) + (.07071)(0.0873)2/2 = 2 N 2 N= 0.0620 N = 32.25

We know that as currents oscillate, they radiate electromagnetic waves. The vector potential A is given by Equation 16.13, and which applies to Figure 17.6. The equation is reproduced here for convenience: A= 4 J(r )ejk|rr | dV |r r |

(17.25)

513

Using the results of the last section, using Equation set 17.6 r r r r cos

(17.26)

J(r )ejk(rr r

cos )

dV

(17.27)

and since we are integrating over the primed coordinates A ejkr 4r J(r )e jkr

V

cos

dV

which gives a very important result in spherical coordinates: A= ejkr 4r J(x , y , z )e jk(x

V

cos cos+y cos sin +z cos)

dx dy dz

(17.28)

since r cos = r r, and r = ax cos cos + a y cos sin + az cos . EXAMPLE 17.5 Consider the very simple linear current density ca for l < z < l; c = constant z Jl (z ) = 0 elsewhere ejkr 4r ejkr 4r

A= =

J(x , y , z )e jk(x

V l l

dx dy dz

caz e jkz

cos

dz

and using H = A rst, neglecting all terms proportional to 1/r2 and 1/r3 then computing E = H ar H = E = j k ej k r sin() sin (kl cos ()) r cos() j k ej k r sin() sin (kl cos ()) Z0 r cos()

514

1

0.8

0.6

0.4

0.2

-0.2

A' A B' B

C' C

A normalised plot of E vs , 0 < < is shown in Figure 17.7. If we scrutinise the gure what strikes straight away is that as the length of current element is increased, the beam becomes narrower, and the directivity of the antenna improves. Notice that the pattern resembles the sin x/x function in the fareld.

In his book Trait de la Lumire published in 1690, Huygen discussed a new principle whereby if there is a wavefront of light at t = t0 then each point on the wave front acts as a source of a spherical wave which propagates with speed of light. Since each point on the wavefront acts like a source, the new wavefront is an envelope of all these little spheres as shown in Figure 17.8. We can see from the gure that at t = t0 the wavefront is ABC. At every point on the wavefront as shown at A, B and C small spheres are drawn whose radii are c(t1 t0 ) and the wavefront advances to A , B and C . The new wavefront is A B C . The surface A B C is slightly dierent from ABC. Aperture antennas are based on the fact that if a travelling electromagnetic wavefront is present on some surface then if those elds act like a source of

515

Radiated fields

Aperture

waves then what the fareld would like. A good example is that of a waveguide antenna. A waveguide antenna consists of a waveguide which is open and in ush with a metal plane. The open end of the wavegude radiates