As per new UGC syllabusCopyright © 2003 New Age International (P) Ltd., Publishers NEW AGE INTERNATIONAL (P) LIMITED, PUBLISHERS 4835/24, Ansari Road, Daryaganj, New Delhi-11002 Offices at: Bangalores Chennai * Guwahati + Kolkata Hyderabad * Lucknow » Mumbai This book or any part thereof may not be reproduced in any form without the written permission of the publisher. This book cannot be sold outside the country to which it is consigned by the publisher without the prior permission of the publisher. ISBN : 81-224-1500-8 Published by K.K. Gupta for Nel Age International (P) Ltd.. 4835/24, Ansari Road, Daryaganj, New Delhi-110002 and printed in india at Hindustan Offset, Delhi-110 051Contents 2.14 Lattice Specific Heats 2.23 2.15 Einstein Theory 2.24 2.16_Debye Theory of Atomic Heat 2.28 ” 2.17 Summary 2.36 2.18 Solved Examples 237 2.19 Problems __2.42 ‘ = 3 2.21 Objective Questions 2.44 3._Motion of Electrons 3.13.49 3.41__Introduction __34 3.2 Conduction Electrons 32 ‘1, Crystalline State and Bonding —1.1-1.70 33 Free Electrons 33 11 Overview ql 34 Characteristics of Conduction 1.2 _ Crystal Structure 47 Electrons 35 13 X-ray Diffraction 1.35 35 Blectron Collisions 38 14 Bonding 1.40 36 Boltzmann Equation 341 15 Solved Examples 1.60 3.7 Relaxation Time 3.13 16_Problems 161 38 Mean Free Path 314 1.7 Review Ouesti 1 39 Conductivity and Ohm's Law 3.15 18 Objective Questions 1.68 3.10 Effect of Impurities on p 3.19 2. Lattice Vibrations and Specific 311 _The Allowed Energies 321 Heats 212.46 3.12 Fermi Characteristics 323 21__Introduction ___21 313 Density of States * 3.26 22 Second Order Expansion 3.14 Fermi-Dirac Statistics 330 of Potential Energy 22 3.15 Summary 337 2.3__Lemnard-Jones Potential. _2.3 3.16 Solved Examples 3.38 2.4__Harmonic Motion _23 3.17 Problems 3.45 25 Simple Harmonic Oscillator 2.4 3.18_Review Questions é 347 2.6__Normal Modes of a Lattice 2.6 3.19 Objective Questions 347 2.1_Vibrations of _One-Dimen- 4._Energy Bands 41-459 sional Monatomic Chain __2.8 41 Introduction 44 2.8 Vibrations of Qne-dimen- 42 Periodic Potential 46 sional Diatomic Chain ___2.11 43 Bloch Theorem 48 29 Density of States 217 44 Kronig Penney Model 414 2.10 Interactions Beyond Nearest 45 Effective Mass of Electron 4.19 Neighbours 2.18, 46 NFE Model 424 Pe ee 218 7 allan 2.12 Three Dimensional Lattice _2.18 48 Density of States 429 2.13 _Phonons _____2.19 49° Fermi Surface 431 410 Electrical Conductivity and FS 4,37 4.11 Metals, Semiconductors and Insulators 439 412 Summary 445 413 Solved Examples 447 414 Problems 455 415 Review Questions 456 4.16 Objective Questions 457 ._ Superconductivity 5.1-5.61 5.1__Introduetion __s1 52. Discovery 52(ii) 5.3 Occurrence 53 6.12 Objective Questions 635 5.4 Transition Temperature 55 7. Semiconductors 7AW738 5.5 Theoretical Survey 56 71 Introduction 1a 5.6 Empirical Criteria 56 72 Intrinsic Semiconductors 12 5.7__Energy Gap 58 73 Electrons and Holes 73 58 Energy Gap Dependent 74 Temperature Dependence of Properties 5.10 Carrier Concentration a 5.9__Meissner Effect 5.18 75_Feumi Lew! ___710 5.10. Trapped Magnetic Flux 5.20 76 Conductivity and Mobility 713 SAL Critical Field 523 17 Doped Semiconductors 77 5.12 Types of Superconductors 5.26 78 __Law of Mass Action _7.21. 5.13_Critical Cusrents 533 79 Impurity States 12 5.14 Critical Superconducting 710 Hall Effect 125 Susie ba 711 Summary 1.29 S.15_Two-fluid Model __5.34 7,12. Solved Examples 129 5.16 Penetration of Magnetic Fields 5.35 713 Problems 135 5.17 Pippard Coherence 537 114_Review Oucstions 136 5.18 Isotope Effect 5.38 7A5_ Objective Questions 136 5.19 Origin of Energy Gap 5.40 8 Semiconductor Devices 8.18.49 5.20 Cooper Pairs 5.40 81 p-n Junctions 81 5.21_BCS Theory 5.46 82 Currents across Junction in 5.22 Summary 5.50 Equilibrium 8.1 5.23 Solved Examples 551 83 Biased Junctions 81 1.24 Problems 5.57 84 Diode 8.14 Review Questions 5.58 85 Zener Diode 819 5.26 Objective Questions 5.59 86 Tunnel Dinde Magnetic rties 6:16.36 8.7_Transistor 8.25 1 In i 61 88 Mctal-Semiconductor Junction 8.28 6.2 Atomic Magnetic Moment 63 82 Photosensitive Devices __ 8.31 63 Diamagnetism 66 810 Solar Cells 831 6.4 Paramagnetism 68 811 Light-Emitting Diode 834 6.5 Ferromagnetism 6.14 8.12 Summary 839 66 Antiferromagnetism 6.24 813. Solved Examples 841 6.7 Ferrimagnetism 620 B14 Problems 8 4G. 68 Summary 6.28 815 Review Oucstions 847 69 Solved Examples 6.29 8.16 Objective Questions 8.47 6.10 Problems 633 Ol Appendix ALAix 6.11 Review Questions __634 OV Index 410. 10.16 Hybrid Equivalent Model 10.26 10.17. Partial Derivative Definition of h-parameters 1031 10.18, Transistor Biasing 10.33 10.19. Q—Point 10.34 95 Full-Wave Rectifier 012 1020. Load Line 10.35 9.6 — Introduction to Filters 9.16 10.21, Variation in h-parameters 10.6 9.7 Inductor Filter 9.16 1022. Bias Stabilisation 1038 98 Capacitor Filter 9.18 1023. Thermal Runaway 10.41 9.9 L-Section Filter (LC Filter) 9.20 10.24, Summary 10.46 9.10 Multiple L-Section Filter 9.22 10.25. Solved Examples 10.47 9.11 x-Section Filter 9.23 10.26. Problems 10.53 912 x-Section Filter with R 10.27. Review Questions 10.54 replacingL 9.24 10.28. Objective Questions 10.55 9.13 Summary of Filters 9.24 11. Field Effect Transistors __‘11.1-11.35, 9.14 Zener Diode 9.25 1.1_Introduction 9.15 Power Supply Characteristics 9.26 112__Iunction Ficld-Effect Transistor 11.1 9.16 Shunt Voltage Regulators 9.28 11.3. Comparison of FET and BJT 113 9.17 Series Voltage Regulators 931 114. FET Operation 1.4 9.18. IC Volt: 935 1.5_JEET V-I Characteristic Curves 119 9.19 Rey Power Suj 935 116 Pinch-off Voltage 11.10 9.20 Summary 9.38 11.7_ Theory of Current 9.21 Solved Examples 9.40 Characteristics of KET __1.11 9.22, Problems 9.47 118 Transfer Characteristics __11.13 9.23 Review Questions 9.48 119 FET Parameters 114 9.24 Objective Questions 9.49 11,10 Relation between g, and Vp 11.15 Bipolar Junction Transistors 101-1058 11.11 MOSFET 15 10.1. Introduction 10.1 1112 _D-MOSFET. 1116 10.2, ‘Transistor Construction 102 1113 E-MOSFET 1118 103, Bipolar Transistor 10.4 11.14 Graphic Symbols for MOSFETs 11.22 10.4 Open-Circuited Transistor 106 11.15 Summary of Transisitor 105 Transistor Biasing 106 Properties uz 10.6, ‘Transistor Operation 10.7 11.16 Summary 123 10.7 Current Components 108 11.17 Solved Examples 11.26 108 Current Gain ____109 11.18 Problems 11.30 10.9. ‘Transistor Configurations and 11.19 Review Questions 11.31 ‘sti 11.20 Objective Questions 11.32 10.10 Subscripted Symbols 10.10 12, FET Biasing 121-12.22 10.11_C-B Configuration 10.11 121_Introduction __121 10.12. CE ration 10.13 122. Fixed-Bias Configuration 121 10.13. C=C ation 10.17 123 Self-Bias Configuration 123 10.14 BJT Modelling 10.18 124 Voltage Divider Biasing 125 10.15 r,-Model 10.22 125 Variable Voltage Resistor 12.5(iv) 126 ‘Transconductance Factor 126 13.10, Emitter Follower 13.15 12.7 _Biasing for Zero Current Drift 128 13.11. Classification of Amplifiers 13.16 128 _Biasing Against Device Variation 12.9 13.12 Distortion in Amplifiers. 13.16 12.9 Energy level Diagram in 13.13 Decibel 13.17 ‘n-Channel FET 12,10 13.14 Miller Effect Capacitance 13.18 12.10 Source Follower Configuration 12.10 13.15 Miller Effect in Bipolar 12.11_D-MOSFET Biasing IB Transistor 13.19 12.12_E-MOSFET Biasing 1213 13.16 Frequency Response of an 12.13 Feedback Biasing 12.13 Amplificr 13.20 12.14 Summary x 13.17 R—C Coupled Transistor 12.15 Solved Examples 12.15 Amplifier 13.22 12.16 Problems 1219 __13.18. Transformer coupled Amplifier 13.25 12.17 Review Questions 1220 _13.19 Bode Plot __ ‘134.26, 12.18 Objective Questions 1220 13.20. Equivalent Circuit at Low, 13, Small SI 13.1-13.50 , Medium and High Frequency 13.27 Ba 13.21. Feedback 13.31 13.2. General Principle of Operation 13.2 13.22 Applications of Negative 13.3. Practical Circuit of Transistor Feedback 13.32 ‘Amplifier 3 13.23. n-Stage Cascaded Amplifier 13.34 13.4. Various Currents of Transistor 13.24. Multistage Frequency Effects 13.35 Amplifier 33 13.25. Common-Source Amplifier 13.36 13.5. Various Voltages of Transistor 13.26. Noise of Electronic Circuits 13.37 Amplifier 34 13.27, Summary 13,38 13.6. Effects of Ry and R, on 13.28 Solved Examples 13.42 Amplifier BS 13.29. Problems 13.46, 13.7 Transistor as an Amplifier B8 13.30 Review Questions 13.47 138 Common Emitter Amplifier 13.11 1331, Objective Questions 13.48 Bo Common-Base Amplifier 1314) Index 14Preface The book originates from the model curriculum recently proposed by the Course Development Committee of UGC for undergraduate students for the paper entitled “Solid State Physics, Solid State Devices and Electronics”. Looking into the width of the coverage of the subject, the book has been devided into two parts : Part 1 covers Solid State Physics and Devices, while Part 2 comparises Electronics. ‘The purpose of the book is to treat the fundamental principles in sufficient detail to bring out a coherent physical framework for all the subject matter. The approach is fairly pragmatic throughout, aiming to provide a physical as well as mathematical understanding of the wide range of phenomena, We will assume that the student has already made himself well-versed with ordinary evel quantum mechanics. ‘The book endeavours to present a systematic and comprehensive coverage with indepth treatment of the subject matter. Although the book aims to target B.Sc. students, yet the readership could easily be broadened to include chemists, material scientists and electrical engineers. The salicnt features of the book are : 1. A unified presentation of the essentials of Solid State Physics and Electronics through casy and lucid language. . 2, Basic ideas are emphasised to help students clearly comprehend the theoretrical meaning of the essentials and recognize their experimental foundations. 3. About 20 solved examples in each chapter illustrate the underlying theoretical concepts described there in. 4, A large ‘number of problems and review questions in each chapter would test the degree of understanding a student has acquired. 5. Inclusion of sufficient number of objective questions would help the students a great deal in facing competitive examinations like NET, SLET, ICS and State Civil Services. A short summary of the plan of the contents of the book will be in order. Chapter 1 deals with the crystal state and bonding and thereby includes crystal structure and X- ray diffraction. The lattice vibrations and specific heats form the subject matter of Chapter 2, where a basic discussion on second order expansion of potential energy is given and specialized for Lennard-Jones potential. Description’ of harmonic motion and normal modes of a lattice forms the bass of vibrations of monotomic and diatomic one-dimensional lattices. Einstcin and Debye theories of specific heats are presented as applications of lattice vibration. Chapter 3 develops the free electron theory of metals and is applied to derive the Fermi Characteristics and explain the Fermi-Dirac Statistics. Chapter 4 introduces the subject of electron dynamics in terms of periodic potential and Bloch wave. It also draws a basic distintion between a free electron and a Bloch electron. NFE model illustrates that the existence of energy gap is a direct consequence of the characteristics of the Bloch electrons,aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.14 Solid State Physics and Devices measured. Most useful metal objects, however, are not made from just one single crystal, but are_made up of ‘many thousands of tiny individual crystals ("grains") joined together by atomic forces at grain "boundaries", further- more, these very small crystals are usually randomly oriented, and thus are mixed up so that the directional variations are averaged out. A substance so formed is called polycrystalline. This results in a body which can be regarded as, homogeneous (every small part of the body from Wherever it is taken is identical with every other part), continuous (containing no voids) and isotropic (having the same _ physical properties in all directions). This is known as an botropic continuum. 1.1.3 Liquid crystals : In 1888, Reinitzer observed that, as he heated cholsteryl benzoate, the solid melted at 149°C into a liquid of a white, turbid, murky appearance. When this liquid is heated further, it underwent another transition at 179°C, this time into a clear, transparent liquid. The substance was further investigated by Lehman, who found that the liquid actually exhibits optical anisotropy—i.e., birefringence —when in the turbid region, much as a crystal does. The unusual fact that the substance has the mechanical properties of a liquid—eg, ability to flow and low viscosity—and the anisotropic optical properties of a crystal prompted Lehman to coin the descriptive named liquid crystals, which has been retained ever since. The liquid crystalline phase is often called the mesophase, and a substance having such character a mesogen. A liquid crystal is a state of matter that mixes the properties of both the liquid and solid states. Liquid crystals may be described as condensed fluid states with spontaneous isotropy. In liquid crystals, there is only two- or one-dimensional regularity. Such substances actually flow and will rise in a capillary tube. Liquid crystals are not rare substances. A few bundred examples of liquid crystals are known, most of them being organic compounds (many of which contain aromatic molecules in their structure), such as ammonium oleate Cy7H33COONH,. Liquid crystals are characterised by (i) their ability to form monocrystals with the application of a normal magnetic or electric field; Gi) an optical activity of a magnitude without parallel in cither solids or liquids; and (iii) a temperature — sensitivity which results in a colour change in a certain liquid crystals. For molecular crystals that tend toward irregular cigar shape we may observe a phase transition between 0 K and the MP at which the translation symmetry is lost but the orientational symmetry is retained. This phase is known as liquid crystal phase, which represents the “intermediate” states of matter between a crystal and a liquid. The molecules in liquid crystals are long and rodlike, a typical length range being 15-40 A. The large anistropy of the molecule is essential for the appearance of the mesophase, as we shall sec. The appearance of a liquid crystal, resembling that of a colloidal solution, led to early suggestions that such a substance is also a colloidal solution. We know that this is incorrect because a liquid crystal has well fixed lower and upper temperatures, an indication that we are dealing here with true phase transition, and hence a distinct phase of matter. X-ray studies also clearly cstablish that the substance has orientational order in the mesophase. The mesogens discussed here are thermotropic, ie., they are obtained by heating solids, There are also lyotropic liquid crystals, formed by dissolving certain crystals in suitable solutions. This latter group is relevant to the structure of biological membranes, which also appear to have a liquid crystalline structure. Classification : We have said that the molecules in a liquid crystal are long. In the mesophase these long molecules tend to align parallel to cach other along a certain preferred direction. There are also additional structures present, on the basis of which Friedel divided liquid “crystals imo three different phases: nematic, cholesteric, and smectic.Cystalline State and Bonding 1S (i) The nematic phase has the simplest structure. The molecules are parallel to cach other, but otherwise their spatial distribution is random, as in a liquid (Fig. 1.3a). There is thus an orientational order, but the molecules are able to move around from one region to another as in a liquid—a fact responsible for the low viscosity. Each molecule is, of course, free to rotate around its axis, because of its rodlike shape. A liquid in the nematic phase also has a turbid appearance. An example of a nematic crystal is p-azoxyanisole, whose temperature Tange of existence is 116-136°C. Na HHL aM i sll Hin ihe VAL Wy) Mh NE TMI (@) (b) © Fig 13 The (a) nematic, (b) choleseric, and (c) smectic phases of liquid ena. (i) In the cholesteric phase, the molecules are also aligned ‘parallel to each other, but the direction of alignment twists progressively, resulting in a helical structure (Fig. 1.3b). Thus the substance consists of parallel sheets, or laycrs. In cach shect the molecules are aligned parallel to each other. The pitch of the helix is typically around 2000 A, but this can be lengthened by the application of suitable external fields. Because of the helical structure, a cholesteric substance exhibits optical activity, ie, the plane of polarization of a light beam is rotated as it travels in the substance in a direction parallel to the axis of the helix. The amount of the optical activity is enormous in some cases, cg,, in activity of 6 x 10%/mm has been observed. That is, the plane of polarization is rotated through an angle of 6 x 10° in a plate 1 mm thick, which can be compared with an activity of only 300%mm in an ordinary organic compound, Chemically, cholestrogens are usually ester cholesterol, a fact responsible for the name "cholesteric phase.” An example is cholesteryl cinnamate, whose range of existence is 156—197°C. Mechanically, a cholesteric liquid has a somewhat higher viscosity than a nematic one. Gili) The structure of the smectic phase is illustrated by Fig. 1.3(c). It consists of a series of layers, in which the molecules are all parallel to each other and normal to the layer plane. The layers interact only weakly, and can readily slip past each other, or be made to rotate relative to each other. It is these motions which are responsible for the liquid-like mechanical properties. In addition to the orientational order, the molecules also exhibit a regularity in their distribution pattern within their own layer, ie, a certain amount of spatial order exists in each layer. The type of spatial pattern as well as the amount of tilt of the direction of alignment relative to the layer plane (the direction: of alignment is not always normal to the plane) have led to further finer subclassifications of the smectic phase. In the smectic A phase, the axis is normal to the layers, while in the C mesophase the axis is tilted. In the smectic B phase the molecules are thought to have an hep structure within the layers, and in the D phase the structure is known to be cubic. An example of a smectogen is ethyl Prazoxybenzoate, whose existence range is 114-120°C, Of all mesophases, the smectic is closest to a solid structure. The only difference between a smectogen and a solid is the lamellar structure of the smectogen, which permits the slip and rotational motions. The smectogen is essentially a two- dimensional solid. Some substances exhibit more than one type of mesophase, depending on the temperature, For instance, 44'-dien-heptyloxya- noxybenzene is smectic in the range 74—95°C and nematic in the range 95~—124°C. Above 124°C, the compound turns into a regular isotropic liquid. The fact that the smectic phase ‘occurs at a lower temperature than the nematic is expected, in as much as the figure has a higher order.Liquid crystals have many applications. Within the liquid crystal phase, some materials (cholesterol extcrs) show a change in colour with change in temperature, and this property is used in liquid crystal thermometers, Some liquid crystals (alkyl cyanobiphenyl) can change the plane of polarisation of light, and are used in liquid crystal displays (LCD). These displays are used in electronic wrist watches, clocks, microcalculators, panel meters, thermometers and industrial products. They can be used to record, store and display images which can be projected onto a large sereen. They also have potential use as television displass. 1.1.4 Glasses, polymers and composites : Some oxide ceramics are crystalline or polycrystalline, but others, notably ordinary glass, arc not. In glass, the atoms have atomic “short-range order’ and a restricted number of nearest neighbours (about nine) but do not achieve the long-range order of crystals. Long-chain polymers have groups of atoms that are joined together continuously to form a long row, but this can curve and fold back on itself rather like a long, tangled necklace of beads, some regions within a polymer material can achieve a crystalline regularity, and some polymers can be prepared in a completely crystalline form. Many useful polymers (c.g. "polythene” and “nylon") have a “glass transition temperature’, Tp, below which they behave in a hard brittle way like glass, and above which they rapidly become plastic or even rubbery. Over the glass transition temperature range, the elastic modules of such polymers drops by a factor of 10°, Bricks and concrete, and the composite materials wood, paper and bone, are really mixtures of crystalline substances bonded intimately with one or more amorphous or noncrystalline components. Glasses, polymers and composites are also often treated as isotropic continua, but this may not always be a valid assumption. Solid State Physics and Devices A glass has the random structure of the liquid from which it is derived by cooling below the freezing point, without crystallization. Also, a glass has the clastic properties of an isotropic solid. By general agreement, we say that a liquid on being cooled becomes a glass when the viscosity equals 10’ poise, where a poise is the CGS unit of viscosity. This defines the glass transition temperature Ty. At temperatures above T,, we have a liquid, below Ty we have a glass. The transition is not a thermodynamic phase transition, only a transition for "practical purposes.” Relatively few liquids can be cooled fast enough in the bulk to form a glass before crystallization intervenes. Molecules of most substances have high enough mobility in the liquid so that on cooling a liquid—solid melting transition occurs a long time before the viscosity increas to 10" poise or 10! ep. By contrast, liquid water has a viscosity 18 cp at the freezing point; the viscosity increases enormously on freezing. We can often make a glass by depositing a jet of atoms of a substrate cooled to a low temperature, a process which will sometimes produce " an amorphous layer with like propertics. Amorphous ribbons of some metal alloys may be produced in this way in industrial quantities. The value 10” poise used to define 7, is arbitrary, but not unreasonable. If we bond a slab of glass 1 cm thick to two plane parallel vertical surfaces, the glass will flow perceptibly in one year under its own weight when the viscosity drops below 10" poise. (For comparison, the viscosity of the mantle of earth is of the order of 107 poise.) Figure 1.4 shows the difference between (a) a ceystal and (b) a glass. In certain cases, the solid state may correspond to a supercooled liquid in which the molecular arrangement of the liquid state is frozen in; because of rapid cooling and a high viscosity of the liquid, crystals may not have hadCrystalline State and Bonding a Fig. L4 Schematic illustration of the difference between. (a) @ crystal and (b) a glass. time to grow and a glassy material results (see Fig. 14b). Upon anncaling, such glassy substances may crystallize (devitrify), as is well known to any experimentalist who has worked with quartz. ‘Table 1.1 illustrates the characteristics of glasses and compares these to those of metals. ‘Table 11 Glass Metal Structure amorphous | crystalline Bonding covalent metallic Yield-stress | almost ideal | non-ideal | Workability poor, brittle | good, ductile Hardness very high —_| low to high Optical transparent | opaque transmission Thermal poor ‘very good conductivity [Resistance high very low very good Poor to good 17 115 Metallic Glasses : A metallic glass has conductivity similar in magnitude to that of solid and liquid metals and optical behaviour typical of metals. The precise information thus known for the metallic glass is : ( It is homogencous, non-crystalline, but disordered structure. (i) It is densely packed as in a compact crystalline structure. (iii) Interatomic distances are not those found in foc and hep structures. (iv) Non-metallic atoms are spread out in metallic matrix, obeying a unique structural role—ice, not to be in hard contact with another nonmetallic atom, Metallic glasses can be used as low temperature thermometers having resistivity of about 60 4V/K, Since the resistance of metallic glasses is high and does not change appreciably with temperature, they can be used as resistance standards. Chalcogenide glasses are used in threshold (cut-off) switching, optical memory devices and now-a-days in lasers as source. Important applications of metallic glasses are magnets for fusion reactors and magnetically leviated trains. 1.2 Crystal Structure 1.2.1 Crystal periodicity : The geometrical form of a crystal is best described in terms of its crystallographic axes; these are generally three lines meeting at a point, which are chosen so as to bear a definite relationship to the characteristic features of the crystal. For example, the crystallographic axis may coincide with, or be parallel to, the axes of symmetry (the lines about which a crystal if notated remains invariant), or the edges between principal faces, or be normal to the planes of symmetry (a plane which acts as a mirror with respect to which two atoms in the crystal bear image-object relationship). The choice of the axes is to some extent arbitrary, although it isaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Crystalline State and Bonding It may be noted that: i) Every basis is identical in composi- tion, arrangement and orientation. (ii) No basis contains fewer atoms than a primitive basis contains. iii) The basis consists of the atoms, their spacing, and bond angles, which recur in an identical fashion about each lattice point such that every atom in the crystal is accounted for. (iv) The complete structure of a crystal is specified by giving both the unit cell and its contents, specified by the basis vectors desctibing the atomic positions. If the basis vectors from the origin. of the unit cell to an atom is ya, + Myf, + sy, it is usual to shorten the notation and say that the atom is (i445 Hz, Hs); the complete set of basis vectors is called the basis of the structure. (v) Fewer atoms are’ needed to describe the basis set for a primitive lattice than when non-primitive vectors arc used, and for some simple elemental solids (such as argon or sodium) the basis consists of a single atom. Other elements crystallise in structures with several atoms per primitive basis (c.g, a basis of two atoms for silicon, four atoms for gallium) and, of course, for a compound the basis must comprise at least one molecule. In some complex organic compounds many thousands of atoms are required for a single basis set. A space lattice or Bravais lattice is not an arrangement of atoms. It is purely a geometric arrangement of points in space. In order to describe a crystal stricture, we must cite both the lattice and the symmetry of the basis of atoms associated with each lattice point. ‘The crystal structure is thus given by two specifications: 1. the lattice, 2, the assembly of atoms that repeats itself. Hence, the logical relation is lattice + basis = crystal soucture This is not a thcoretical equation, but just a concise way to write the above definition. I 1 Figure 18 illustrates the distinction between a lattice and a crystal structure for a two-dimensional case. The basis of two atoms is depicted by a small and a larger open circle. Symbolically, ()=@) +) - 8 ee - 8 ee (a) Space lattice ° ° (b) Basis, containing two different ions of of of of of? ot? of of? °° of? of of? ioe? o°° o*° o°° (©) Crystal structure Fig, 1.3 The cnstal structure (c) may be considered to be formed ty the addiion of the basis (b) 10 every latice point of the space lattice (a). It should be remembered that the lattice points of a space lattice do not generally represent a single atom, but rather a group of atoms or a molecule. In addition to the symmetry properties of the lattice points themselves, then, one must consider the symmetry properties of the molecules or groups of atoms at each lattice point about the lattice points themselves inorder to completely ‘enumerate all possible crystal structures. When this is done, it is found that there are 230 basically different repetisive pattems in which such elements can be arranged to form possible crystal structures in three dimensions. 1.23 Fundamental translation vectors : The basis vectors a, 6, ¢ for a lattice are a set of linearly independent vectors and are not unique, but they must satisfy the requirementLR Solid State Physics and Devices that all the equivalent points in the lattice are given by R,, of eq. (1.5) with a suitable choice of nj (i=1,2,3). In addition, all integral choices of n; must represent a lattice point, A suitable set of basis vectors may be obtained as follows : a is taken as the shortest possible periodic Jength in the lattice; bis the next shortest length not parallel to a; and © is the next shortest length not coplanar with b and c. It is to be noted that (i) In general, a, 6, ¢ are not orthogonal. (ii) The set (@, 6, €) is not unique. Any lincar combination with integer coefficients gives another valid set. In Fig. 1.7 (a), Ry = 2a + 3b + Se (as) where 1, = 2, ny =3 and ng =5. Here Ry is a lattice vector as it takes us from one lattice point to the other lattice point. Thus, any lattice vector R, can be written as a linear combination of a, 6 and c (using integers, m), 2 and ny). Conversely, it is also true that any lincar combination of a, 6 and c (using integers, ny, my and m3) is a lattice vector Ry. Such vectors a, 6 and c are called basis vectors or fundamental vectors ot unit translation vectors of the lattice, Consider the lattice shown in Fig. 1.9(a). Let us choose the origin of coordinates at a certain lattice point, say A. Now the position vector of any lattice point can be written as R, = nye +ngh, (14a) where a, b are the two vector shown, and (ny m) is @ pair of integers whose values depend on the lattice point. Thus for the point D, (13, m2) =(0, 2); for B, (ty, mg) = (1, 0), and for F, (ny, m2) = (0, ~ 1). The two vectors a and 6 (which must be noncolinear) form a set of basis vectors for the lattice, in terms of which the positions of all lattice points can be conveniently expressed by the use of (La). The set of all vectors expressed by the equation is called the lattice vectors. We may also say that the lattice is invariant under the group of all the translations expressed by (1.4a). This is often rephrased by saying that the lattice has a translational symmetry under all displacements specified by the lattice vectors R,, ‘The choice of basis vectors is not unique. Thus one could equally well take the vectors @ and 6' (=a +6) as a basis (Fig. 1.9a). Other possibilities are also evident. The choice is usually dictated by convenience, but for all the lattices we shall meet in this text, such a choice has already been made, and is now a matter of convention. 12.4 The unit cell : The area of the parallelogram whose sides are the basis vectors and 6 is called a unit celf of the lattice (Fig. 1.9a), in that, if such a cell is translated by all the lattice vectors of (14a), the area of the whole lattice is covered once and only once. The unit cell is usually the smailest area which produces this coverage. Therefore the lattice may be viewed as composed of a large number of equivalent unit cells placed side by side, like a mosaic pattern. The choice of a unit cell for one and the same lattice is not unique, for the same reason that the choice of basis vectors is not unique. Thus the parallelogram formed by @ and 6’ in , 1.9a is also an acceptable unit cell; once ‘again the choice is dictated by convenience. ‘The following remarks may be helpful. (i) All unit cells have the same area, Thus the cell formed by a, b has the area S= ax], while that formed by a, 6° has the area S’ = Ja x b'| = |ax(a+6)| Ja x b| = S, where we used the result axa =0, Therefore the area of the unit cell is unique, even though the particular shape is not. (ii) If you are interested in how many lattice points belong to a unit cell, refer to Fig. 1.9(a). The unit cell formed by a x b hasCrystalline State and Bonding four points at its corners, but each of these points is shared by four adjacent cells, Hence cach unit cell has only one lattice point. Conventionally, a unit cell may be defined as an clementary unit of volume, which builds up the whole crystal by repetitive operations. Thus, the elementary volume must have all the propertics of the crystal. In other words, a crystal can be built up by the juxtaposition’ of such elementary units. Fig, 1.9(0) Vectors @ and b are Bost vectors of the lice Ycwr a and ' form another set of basis vector: Shaded and haiched areas ate unit cells corresponding to first and second set of basis vectors, respectively. Thus, a crystal can always be divided into building blocks called the unit cells. Each unit cell has the same shape, the same volume, and the same contents. By completing the parallelopiped formed by a, 6 and ¢, a volume a : (6 x c) is enclosed which, if continuously uanslated parallel to itself by R,, (eq. 1.2) will fill all space. This volurie, which contains one lattice point, is called a primitive unit cell. The scheme of calculating the effective number of lattice points in a unre dimensional lattice is as follows : (@ If a lattice point lies at the comer of a unit cell, it is considered to be 1/8th in each cell. (i) A lattice point lying on the edge of a unit cell contributes 1/4th to the cell. (ii) A lattice point on the face of a unit cell is only half in the cell. 413 (iv) A lattice point situated in the bulk of the unit cell is entirely its own property. We define the term unit cell as a cell also formed by three vectors, in a manner similar to that as above, which will fill all space when translated by multiples of these vectors. The difference is that while a primitive unit cell contains one lattice point, a unit cell may contain more. Cells such as these, containing more than one lattice point, or also called nonprimitive or multiply primitive unit cells. Figure 1.9(b) shows examples of primitive and nonprimitive unit cells. The primitive cell in the lower left is slightly displaced and thus more clearly shows that it only contains one lattice point, Note that there arc an_ infinite number of possible primitive unit cells, but they all must have the same volume jin three dimension (or arca in two dimension as in the figure). However, there are conventions as to OS: P = primitive unit cell NP = nonprimitive unit cell Fig. 1.9(b) Examples of primitive and nonprimitive want cells The unit cell is not of course, unique. One obvious way to define it is as the parallelopiped bounded by the basic vectors @, } and cas done in the above case, Even then the midpoint of the cell can be arbitrarily chosen (as in Fig. 1.10a) although it is usual to centre it on of the atomic positions. Primitive versus nonprimitive cells : Mis sometimes more convenient, however, to deal with a unit cell which is larger, and which exhibits the symmetry of the lattice more cleary than the primitive celt. The idea is illustrated by the Bravais lattice in Fig. 1.10(b). Clearly, the vectors a}, a> can be chosen as a basis set, inFig. 1.10(a) General two-dimensional space lattice showing basis vectors, and two forms to the unit cell as @ parallelogram with different centres. which case the unit cell is the parallelogram 5. However, the lattice may also be regarded as a set of adjacent rectangles, where we takes the vectors a and 6 as basis vectors. The unit cell is then the area Sp formed by these vectors. It has one lattice point at its centre, in addition to the Points at the corner. This cell is a nonprimitive ‘unit cell. The reason for the choice of the nonprimitive cell Sp is that it shows the rectangular symmetry most clearly. Although this symmetry is also present in the primitive cell Sy (as if must be, since both refer to the same lattice), the choice of the cells somehow obscures this fact. Note the following points. (@ The area of the nonprimitive cell is an integral multiplic of the primitive cell. In Fig. 1.10(b), the multiplication factor is two. No connection should be drawn between nonprimitive cells and non- Bravais lattices. The former refers to the particular (and somewhat arbitrary) choice of basis vectors in a Bravais lattice, while the latter refers to the physical fact of noncquivalent sites. 125 WignerSeiz cell Another convenient way of choosing the unit cell is to take the volume bounded by the planes which perpendicularly bisect the nearest R,. Examples of these (wo methods of choosing the unit cell are shown in Figures 110 and 1.11 for Gi) Solid State Physics and Devices Fig. 1.1006). Area 51 is a primitive unit cell: area Sr is a nonprimitve unit cell two-dimensional lattices. The unit cell in the latter case is referred to as the Wigner-Seitz cell. Fig. 111 Triangular plane lattice. \a| = |b| and the angle between them is x13. Two forms of the unit cell are shown—one as a parallelogram and the other at a ‘hexagon obtained by draning perpendicular bisects to ‘ich lac rector from a cruel te The location of a point in the unit cell is specified by the three numbers u, v and w, for which each coordinate describes the distance from an origin in units of the cell dimension(s). ‘Thus, if one corner of the unit cell is denoted as the position 000, the opposite corner is 111, and fractional numbers are required for any imtermediate point within or on the surface of the cell. In this notation, + 30 is one of the face centre positions and 5 4 is the centre of the cell 126 Macroscopic symmetry : The symmetry of external shapes of perfectly developed crystals is apparent at first sight, This symmetry, however, may also be found in irregularity bound crystals by measuring theirCrystalline State and Bonding geometric and physical properties. From the macroscopic point of view, single crystals differ from amorphous substances ey virtue of anisotropy, ie. by the dependence of their physical properties on direction. If the physical properties are measured for their dependence on direction, we find that some directions are physically equivale a nt, indisingathable on physically equivalent directions ma FON Into selPcoinedence by certain movements in the crystal (transformations). These transformations are not arbitrary and are called the operations of the macroscopic symmetry. The operation of macroscopic symmetry leads to a change in crystal coordinates yz) — @'y'z') which is not accompanied by a compression, extension or a shift of any point in the crystal or of the crystal as a whole. In other words, macroscopic symmetry operations do not alter the spacings between the individual points of the crystal and at least one point remains at rest. The fundamental symmetry operations are the mirror reflection in a plane, rotation about an axis and the inversion through a point. The corresponding symmetry clements are plane of symmetry, axis of symmetry and inversion centre. The plane of symmetry : If the crystal may be divided by a mirror plane into two pars so that one part is the mbror picture of other in the mirror plane (M 2 the the element is called the plane of symmety. When this plane of symmetry is chosen as the coordinate plane YZ ig the crystal, then we can say that the planc of symmetry is such that a transformation transforms any point (yz) in the crystal to the point (ry 2), ic. @yz)— @y2) The corresponding operation is denoted by My. The axis of symmetry : A crystal has n-fold axis of symmetry if its physically cquivalent directions are brought into self- coincidence and its physical state is not altered by a rotatioa uf 2n/m around an axis passing through the crystal. Since the crystal possesses a three-dimensional periodicity, it has only two-, three-, four-, and sixfold axes of symmetry, ie. n = 2, 3, 4 and 6. Lis The axes of symmetry corresponding to these values of n are respectively called diad, wiad, tetrad and hexad axes. The inversion centre : A crystal structure possesses a centre of inversion if any point in the ‘crystal is associated with another point situated symmetrically about a fixed point which is called the centre of inversion (centre of symmetry). In this transformation (denoted by 1) an arbitrary point of the crystal (yz) is transformed into an equivalent point (x'y’z’) according to the relation (yz) — @92) In other words, a lattice point is said to be the centre of symmetry, if a line from any other lattice point to it and extending an equal distance on the other side encounters an exactly identical point. This means that the replacement of every interatomic vector r by —r should leave the crystal structure invariant. These three elementary operations of symmetry, or elements of symmetry, may be formed from a single element of symmetry— the plane of symmetry. The rotation by the angle @ is equivalent to the mirror reflection on two Planes with the angle a/2 bewween them. The centre of inversion may be created by a mirror reflection on three planes perpendicular one to another. Inspite of this fact, the three elementary ‘operations of symmetry which are simple and illustrative arc used as independent simple operations from which the other more complicated elements of symmetry are composed. ‘Any of the three elements of symmetry may be arbitrarily repeated so that the various combinations of the elementary operations of symmetry yield altogether 32 various operations of symmetry. Any of the 32 operations of symmetry contains as its elements certain movements by which the physically equivalent directions may be brought into self-coincidence. ‘These elements form a closed set with certain properties which in mathematical terminology is called a group. For this reason, the 32 independent operations of symmetry are also called point groups. For example, the 4-{old axis116 of symmetry is a point group with four elements of symmetry, namely, the rotations by 90°, 180°, 270° and 360°. The number of elements in a point group increases with the growing symmetry of the crystal. Crystals having any particular point group symmetry are said to belong toa particular crea ‘The symmetry of a crystal point group is transferred to the physical properties of the crystal according to Neumann's principle, which states that “any physical property of a crystal has at least the symmetry of its point group.” A crystal may have one or more planes, and one or more axes of symmetry, but never more than one centre of symmetry, in fact, many crystals are not centro-symmetrical, since they develop differently at opposite ends (d) (e) (0 Fig. 1.12 Elements of symmeny of a cube. The diagrams in Figure 1.12 serve to illustrate the significance of the various elements of symmetry of the-cube: (a) shows a rectangular plane of symmetry of which there are three, the others being in directions at right angles to the one shown; (b) is a diagonal plane of symmetry, there being a total of six such planes passing diagonally through the cube; (c) indicates one of the three axes of four-fold symmetry, at right angles to each other; (d) represents an axis of three-fold symmetry, passing through opposite corners, of which the cube possesses four; (e) depicts one of the six axes of two-fold symmetry, emerging from opposite edges; and finally (f) Solid State Physics and Devices shows one centre of symmetry at the body centre of the cube. A perfect cube has, therefore, a total of twenty-three clements of symmetry; this aumber is exceptionally large, because of the high degree of symmetry possessed by a cubical body. 1.2.7 Allowed rotations : An axis of rotational symmetry is an axis such that the lattice takes up an identical position on rotation through a suitable angle about it. The angle of rotation must be of the form 2z/n, whee n is an integer and such an axis is called an n-fold axis. Any real crystal _ exhibits both translational and rotational symmetry. The mere fact that a crystal must have translational symmetry places restrictions on the types of rotational symmetry that one can have. A crystal can have only one-, two-, three-, four, and six-fold axes of symmetry. The proof of this therorem is facilitated by the geometrical construction shown in Fig. 1.13. Fig. La3 Let A in Fig. 1.13 be one of the points of a crystal lattice through which an axis of symmetry passes (perpendicular to the plane of the diagram). If B is another point of the lattice separated by one lattice vector from A, then another identical axis of symmetry must pass through B, Now, rotate the crystal through an angle @ = 2nin (n being an integer) about the axis through A. Thea, the point B, as well as the axis through B, takes the position 2’. Similarly, a rotation about B takes A to A’. But since, by assumption A and B are lattice points of the same crystal and hence can replace cach other by a parallel displacement, it is apparentCrystalline State and Bonding LIT that A’B’ must also be a translational period of the lattice. Thus, if @ is the shortest period in the given direction, then the distance A’B' = pa, where p isan integer. From the figure, = 3) =a-2acos O=pa ie. cos@ = (1-p)/2 A'B! = A'C + CD + DB! Since |cos 9} < 1, p can only have the values p = 0, + 1, 2, 3. ‘These values then lead to @ = 2t/n with n = 1,2,3,4 or 6, which establishes the assertion that only these values but not = 5,7,8, etc. are allowed. 128 2-Dimensional Bravais Lattices : The general two-dimensional lattice shown in Fig. 1.14 is an infinite array of points, The points obey the lattice condition that every point should have the same environment in the same orientation. It is apparent that @, b indicated in the figure arc a possible choice for the primitive translation vectors of the lattice. The general parallelogram defined by primitive a, forms a primitive cell, which is a type of unit cell. The crystal translation operations applied to a unit cell cover all points of the plane. A unit cell which contains lattice points at corners only is called a primitive cell. = asin (0-3) ta tasin[ b eee af} ws oe eer eer? 7 8 Fig, 1.14 General vblique latice in ovo dimensions, showing 4 choice of the fundamental translation vactors a, band a unit cel. It is useful to introduce special relations connecting the primitive vectors a, b. The clations will define four special lattices with symmetry properties not present in the general oblique lattice, Fig. 1.14. To enumerate the special lattices, we exploit the symmetry operations associated with two-dimensional point groups. The relevant symmetry operations are five axes (n = 1,23,46) and the mriror line (m). The requirement that a lattice should be invariant under a rotation operation 2n/n, where n = 1,2,3,4 or 6, or under the mirror operation, places restrictions on the primitive translation vectors a, 6, except for the point operations 1 and 2. The general oblique lattice, Fig. 1.14, is invariant under the point operations 1 and 2 without further restriction, All operations are always considered to be applied through a lattice point, ‘The point operation 4 obviously requires a square lattice, Fig, 1.15a. The point operations 3 and 6 require a hexagonal lattice, Fig. 1.15b. It is easily seen that this is invariant under a rotation 2/6 about an axis through a lattice point and normal to the plane. (b) Hexagonal lattice Ublzy = 120° (@) Square lattice (al = [Bly = 90° lal = (c) Rectar lattice | (d) Centered rectc lar emir |Ogertonm, for both the primitive Bi ‘and for the rectangular amit call for whjoh Fig LAS Special svo-dimeniional lattices. These foun together with the oblique tatice (Fig. 1.10) form the fire ovo-dimensional Braveis latices The resricions on the primitive wanslation vectors are given below each latice. The mirror reflection m is a little more interesting. We write the primitive translationLIB Solid State Physics and Devices vectors a, b in terms of the unit vectors § j along the’x, y axcs: a=a, taf by P+ by st If the primitive vectors are mirrored in the x-atis, we have after the reflection operation (us) ata on an bi =bdy +b, fF ‘We wish a’, b’ to be lattice vectors. One possible choice of a, b to accomplish this is a=ai; } oy; is rectangular, Fig, 1.15. There is a second possibility, as b' will be lattice vector if (1.7) the resulting lati a'=a-5, ot, by using eqs. (1.5) and (1.6), bp =ay—b b= a) —by (18) We have a solution if a= O;a, = 2b,; thus another possible choice of primitive translation vectors is a=a ja Raby h (19) This choice gives a centred rectangular lattice, Fig. 1.15d. A centred lattice is denoted by C, and a primitive lattice by P. We have now exhausted the two-dimensional lattices which are required by the point-group operations applied to lattice points. Such lattices are called Bravais or space lattices. 1.2.9 Microscopic symmetry : So far, in the interpretation of symmetry, we have regarded the crystal as a finite macroscopic body, Since by translating the crystal we change its position in such a body, a translation (macroscopic) as a symmetry operation is totally excluded. In order to able to consider even the translation as an operation of symmetry, we must regard the crystal as a three- dimensional discontinuous medium such as, from the microscopic point of view, it actually is, To get the coincidence of a crystal when translating it by integral multiples of its lattice vectors @, b, ¢, it is necessary to consider an infinite” crystal of which the lattice points form the entire lattice space. Such a crystal is called ideal. In an ideal crystal, the translations by multiples of lattice vectors a, b, ¢ are simultaneously operations of symmetry. Now let us see how an actual crystal approximates (o an ideal one. In a crystal of volume 1 cm’, the interplaner spacings are 5A, say, so that there are 2x10" lattice parameters per centimetre. If the crystal is translated by some tens or hundreds of lattice parameters, the relative change of its position is 10~°—10~5 cm and, consequently its mew position cannot be distinguished from the initial position even under the best microscopes, Thus, from the microscopic point of view, every real crystal conforms to the definition of an ideal crystal. 12.10 Lattice ypes : In order to find the positions of atoms and lattice points when identifying the structure of a crystal, it is necessary to choose a suitable coordinate system, The simplest choice of the rectangular system is not in general advantageous, as it does not yield a true picture of the symmetry of a crystal and in some cases has no relation to its structure, Morcover, geometrical relations in crystallography would be too complex and not very iilustrative. Therefore, the coordinate system in a crystal are chosen so as to agree with its symmetry. If a crystal has a certain element of symmetry, we choose a coordinate system so that this clement forms the origin, i, an axis or a plane of the crystallographic coordinate system. If, for example, a crystal has an axis of symmetry, we choose it as one of the coordinates. Similarly, if it bas a plane ofCrystalline State and Bonding 119 symmetry, we choose this pline as a coordinate plane. The coordinate system of asymmetrical crystals with elements of symmetry 1 and I is oblique. The unit cell formed by three unit vectors of system is a general parallelepiped and the system is called triclinic. A growing number of elements of symmetry enlarges simuliancously the symmetry of crystal systems and their corresponding unit such a cells. For example; if a crystal has a four-fold axis of symmetry and. if we associate with it with the vector ¢ of the unit cell, then a, b must be perpendicular to each other and also to the vector ¢. If a and b have to ‘coincide when rotated by 90°, it must also"hold true that |a | = |b |. The unit cell will thus have the form of a rectangular quadrangular prism and the coordinate system is called the tetragonal, Cubic Hetagonal Tetragonal Trigonal / Orthorhombic Monoclinic Triclinic Fig. LAG The seven crystal sysicms ix tree dimensions, obtained by successive distortions of a cube.1.20 If we confine ourselves only to macroscopic elements of symmetry, ie if we consider the crystal as an anisotropic continuum medium, the hexagonal and rhombohedral systems cannot be distinguished from one another and, therefore, sometimes introduced as two equivalent systems leading to six possible crystal systems. However, we shall treat them separately and altogether seven crystal coordinate systems arc introduced. The conventional unit cells for these seven systems can be envisaged as arising from Solid State Physics and Devices progressive distortions of cubic symmetry, as sketched in Fig. 1.16. In two dimensions, ten permissible point groups are associated with five different Bravais lattices, In three dimensions it turns out by parallel arguments that the thirty-two permissible point groups require fourteen different Bravais or space lattices. The three-dimensional Bravais lattices are listed in Table 1.2. The fourteen Bravais lattices are grouped into seven crystal systems: triclinic, monoclinic, orthorhombic, tetragonal, cubic, trigonal, and Table 1.2 The fourteen bravais lattices and conventional unit cells Number of lattices in system. System Nature of unit- cell axes and Total no. of parameters Lengths and angles to be angles specified ‘Triclinic 1 azbec axpay # 90° a,b.e By? Monodiinic avb#c a,b,c B Orthorhombic a,b,c alysis lasaylay *Lattices symbols : P- c- opposite to cach other) Primitive (Lattice points are at the corners of the unit cell only) Side centred or Base centred (Lattice points are at the comers and at two face centres 1 Body centred (Lattice points are at the corners and at the body centre) F- Face centred (Lattice points are at the corners and at the six face centres)Crystalline State and Bonding 121 hexagonal. The division into system is summarised conveniently ia terms of the axial relations for the conventional unit cells. The axes a, b, ¢ and angles a, 6, y are shown in Fig. 1.5, The fourteen space lattices are illustrated in Fig. 1.17 by a conventional unit cell of cach. The conventional unit cells are not always primitive, as sometimes a multiply primitive cell has a closer connection with the symmetry elements than a primitive cell. PF ee e Trielinic-P Monoclinic-P Monoclinic-B Orthorhombic-P uo Tg” ae Tairagonal-P Orthorhombic-I Orthorhombie-F “i yr" 1 Tetragonal-1. Hexagonal-P Trigonal-R Cubic-P Cubie-I Cubic Fig. L17 The conventional unit cell for each of the 14 Bravais lattices. (For the hexagonal lattice the unit cell is outlined in solid lines, but a hesogonal priem is also showm, doued, 10 help indicete the angles of the unit cel)1.22 Solid State Physics and Devices In the iricline system, the single space lattice has a primitive (P) unit cell, with three axes cf unequal lengths and unequal angles. The symmetry of the lattice is 1, all space lattices being invariant under inversion. The examples are CuSQ,.5H,0, — NaAlSigOg, Ca$,036H,0 and RbjFe(CN),.2H,0. In the monoclinic system, there are two space lattices, one with a primitive (P) unit cell and the other (C) with a conventional cell which’ is, base-centred (non-primitive), with lattice points at the face centres of the cell faces normal to the c-axis. The primitive cell of the monoclinic C space lattice is an oblique thombic prism, LisSO,.H,O is an example of C unit cell, while tie examples of P unit cell arc Gypsum (CaSO,.2H0), Borax (NayB,07.10H,0) and C,oHg (Naphthalene). In the orthorhombic system, there are four ‘space lattices; one lattice (P) has a primitive cell; one lattice is base-centred (C); one is body-centred (I); and one is face-centred (F). The examples are as follows : P: CaH, BaHy, StH, MnP, As3Fe, As,Ni C : PuBrs, ZrSin, HfSi> 1: NbBa, UAly, a-AgeTe F : Ale, GeS», SizTi, MgyCu In the tetragonal system, the simplest unit is a right square prism; this is a primitive unit cell and is associated with a tetragonal P space lattice. A second tetragonal lattice I is body centred. The examples are as follows : P : UpSini Cug5b, B-Sn, CaP», TiO2, AlbZrs 1: In, AgFeSp, CdGaySes, AginTes, CdGeAsy In the cubic system, there are three space lattices: the simple cubic P lattice which is primitive; the body-centred 1 lattice; and the face-centred F lattice, The characteristics of the three cubic space lattices are summarised in Table 13. ‘Table 1.3 Characteristics of cubic lattices. Body. centred Simple Face- centred Unit celi volume Lattice points per cell Volume of primitive cell Lattice points per unit volume Nearest acighbour distance Number of nearest neighbours Second neighbour distance |Number of second. neighbours The examples of cubic lattices are as follows: Mn, B-W, AlAus, AuV, CryRu, As,Cdy Na, K, Rb, Cs, Ba, Cy Nb, Mo, Ta F : Al, Pb, Ag Au, Cu, Ar Ki Ne, Ca, Ni, Pd, Pt In the trigonal system a rhombohedron is usually chosen as the unit cell. The lattice is primitive, but it is usually denoted by R, rather than P, so that we speak of trigonal R space lattice. The examples of this system are calcite (CaCO), B-quartz (SiO) and tourmaline, In the hexagonal system the unit cell coventionally chosen is a right prism based on a rhombus with an angle of 60°. The lattice is primitive. The relation of the above cell with a hexagonal prism is shown in Fig. 1.18. This cell is used sometimes also for trigonal lattices. The examples of the hexagonal system are Be, Mg Zn, Cd and graphite. Although the symmetry of the fourteen Bravais lattices varies considerably they all have one clement of symmetry in common in that they all possess a centre of symmetry. This isCastalline State and Bonding Fig LIN Relation of the pnumuive cell m the hexagonal system (hewy tines) 10 a prion of hexagonal smmmenn: defined as a point in the interior of the lattice through which every lattice point can be taken to an equidistant laitice point on the other side of the lattice. Tum now from these purely geometric networks of lattice points to the actual crystal structure exhibited by materials. In general, each lattice point is associated with a group of atoms or molecules rather than just one atom, and the many different geometric arrangements possible mean that there are far more than just fourteen different types of crystal structure, In many cases the resulting crystal structure docs not show the full symmetry associated with the basic crystal structure. For example, in cases where complex molecules with symmetric shapes are associated with the lattice points the resulting crystal structure may no longer have a centre of symmetry. This has important consequences because the absence of a centre of symmetry is a necessary requirement for such phenomena as pyroelectricity, piezoelectricity and optical activity. 1211 Lattice planes : Although the description of atom positions in a unit cell is a complete description of the crystal structure, it is useful to have a means of describing a plane, such as the cube faces in the face-centred cubic (fe.c.) structure. For this purpese, a system of caystallographic indices known as Miller indices has been developed. Since primary interest in lattice planes is confined to the major planes in 1.23 the unit cell of a crystal they are defined in terms of their intercepts. on the three translation vectors. The scheme of determination of the Miller indices is as follows: (1) Take as the origin any atom in the crystal and erect coordinate axcs from this atom in the direction of the basis vectors. (2) Find the intercepts of a plane belonging to the system, expressing them as integral multiples of the basis vectors along the crystal axes. (3) Take the reciprocals of these numbers and reduce to the smallest triad of integers h, k, 1 having the same ratio. ‘The quantity (i k J) is then the Miller indices of that system of planes. Note that the Miller indices are written by simply enclosing the three numbers in parentheses. There may be a number of systems of planes whose Miller indices differ by permutation of numbers or of minus signs, yet which are all crystagraphically equivalent so far as density of atoms and interplanar spacings are concerned. For example, in an orthogonal lattice the planes (hkl), (ikl), (Aki), (KD, (hkl), Gh ed, stc., obtained by assigning various combinations of minus signs to the Miller indices, are all equivalent in this sense. (It is conventional to write the minus signs above rather than before the Miller indices.) Likewise, in a cubic lattice, all the planes represented by permutations of the three Miller indices among themselves, such as (hk), (khi), (hk), etc, as well as those obtained by (aking various combinations of minus signs, are all crystallographically equi- valent. When referring to the complete set of erystallographically equivalent planes of which (hk) is a member, it is customary to enclose the Miller indices in curly brackets, thus: {kt}. If a plane is parallel to one or two of the vectors @, b, ¢, one or two of the intercepts will be at infinity; the corresponding Miller indices are then zero.1.24 x Fig 1.19 Designation of @ erstal plane by Miller indices: Any plane A’B'C’ (see Fig. 1.19) can be defined by its intercepts OA’, OB’ and OC’ on the three principal axes of the unit cell OA, OB and OC. The usual notation for Miller indices is given by taking the reciprocals of the ratios of these intercepts to the unit cell parameters. Thus, the plane A’B'C’ is denoted by (OA/OA' OB/OB’ OC/OC’) which, in general, is written as (hkl). If OA' = 4/2, OB" = b/4 and OC! = c/2 then the plane A'B'C’ is the (1/4 1/V%4 1/¥2) or the (242) plane. Similarly a plane (ABC) with intercepts OA, OB and OC is the (OA/OA OB/OB OC/OC) or (111) plane, and the plane DFBA is the (OA/OA OB/OB OC/) or (110) plane, ie. a plane parallel to an axis is taken to intercept that axis at infinity. If the intercept is negative then the corresponding Miller index is also negative and is denoted by placing a bar over the numerical value, eg. 2. The indices of the side of a unit cell of a cubic lattice are (100), (100), (0,1,0), (010), (001) and (001) and these are equivalent planes of the same form and are collectively represented by {100}. Further examples of planes in the cubic system are illustrated in Fig. 1.20, In the hexagonal system, an alternative indexing system, which has four numbers in each set of indices, is often used, because it more readily shows the hexagonal symmetry. The four axes used to define the indices are the vertical axis of the hexagonal prism and the three axes in the base, the latter being 120° Solid State Physics and Devices z (122) plane y * @) (110) planes » (120) planes) x Fig. 1.20 Lattice planes in the cubic system. apart as illustrated in Fig. 1.21. The indices of a plane are denoted by A, k, i and / where these indices are the reciprocals of the intercepts on the a b, d and ¢ axes respectively. Since only Fig 21 Four axes in she hexagonal system.Cystalline State and Bonding Fig. 1.22. Direction angles of the normal to the (hid) plane. three noncoplanar axes are necessary to specify a plane in space, these four indices cannot be independent and they must satisfy the additional condition that & +k = —i. Thus, the base plane in a hexagonal unit cell is the (0001) plane and the vertical prism planes being to the {1.0 T 0} family. Wt should be noted that all these indices represent a set of equidistant parallel planes, since our reference unit cell could be chosen anywhere in the crystal. In X-ray crystallo- graphy, interest lies in the interplanar spacing of these planes, which is defined as the perpendicular distance from the origin to the plane, ie. a similar imaginary plane is assumed to pass through the origin. The indices (ni, nk and nf) where n is an integer, therefore represent a family of planes which lie parallel to the family (fk!) but with an interplanar spacing 1/n of that of the latter. Note that as far as the external morphology of the crystal is concerned, the (111) and (222) planes, for example, are identical and so it is common in this case to reduce the indices to their lowest form, Spacing between planes of the same Miller indices : In connection with x-ray diffraction from a crystal, one needs to know the inter- planar distance between planes labelled by the same Miller indices, say (hk1). Let us call this distance dy. The actual formula depends on the crystal structure, and we confine ourselves to the case in which the axes are orthogonal. We can calculate this by referring to Fig. 1.22, visualizing another plane parallel to the one shown and passing through the origin. The distance between these planes, dy, is simply the length of the normal line drawn from the 125 origin to the plane shown, Suppose that the angles which the normal line makes with the axes arc a, B and y, and that the intercepts of the plane (Ak!) with the axes are x, y and z. Then it is evident from the figure that dha = £605 = ycosB = 2.05 y. From the above equation and the identity cosa + cos’B + cos’y we get (19a) Now x, y and z are related to the Miller indices h, k and |. If one reviews the process of defining these indices, onc readily obtains the relations. a e 7 han’, kant, lan$, (19) y where 1 is the common factor used to reduce the indices to the smallest integers possible. Solving for x, y and z from (1.9) and substituting into (1.9a), one obtains n pit which is the required formula, Thus the interplanar distance of the (111) planes in a simple cubic crystal is d= na/V3, where a is the cubic edge. For the nearest planes, n = 1 and then (1.10) For a cubic crystal, c= 8 =a, and hence a dna = GI a +k +P) Figure 1.23 shows interplanar spacings for some standard planes in a cubic crystal.1.26 @ (b) &) Fig. 1.23. Spacing berween erstal planes in cubic tance Using eq, (1.10), we-obtain a dee 0 (24040) A100 400 Henee dyyo = and dha = Observe that the spacing of planes with larger Miller indices is closer than the spacing Solid State Physics and Devices of the principal symmetry planes, planes. Table 1.4 summarizes formulae for the interplanar different crystal systems, The Miller indices have a clear geometrical meaning. From a comparison of equations of a plane in the intercept and the normal forms, it is evident that the Miller indices are proportional to the direction cosines of the normal to the corresponding plane. Since the Miller indices are relatively prime numbers, in the latice space they determine that plane which lies nearest to origin from the complete system of equivalent lattice planes, To find all other lattice planes parallel to this plane, we displace it by an integral multiple of | the interplanar spacing d in the direction of its normal. ie. the {100} the _ various spacings of Table 1.4 Interplanar spacings of various crystal systems. System htt Cubic a +i + PV? ‘Tetragonal Orthorhombic | “12 2 2 Hexagonal 40300 + hk +H) Pe a e 22 Rhombohedral wg ay yt cor = Boose) [0:7 + +P) sin?e + 2 (ik + kl + Ih) (cos*a—cosa)y 12 sin? 2 Monoclinic sin B [é 4 Ksin'B LT a 2 2 2 ac cP ve Triclinic dows = Vidi H + dank? + Spal” + rahe + Wagkl + Worhl\ 7’? where 1? = a?b%e7(1 — costar — cos” B — cos*y + 2cosex cos cos y) and d= bc2sin2a@ 333 = ab sin? 423 = abe (cos B cos y ~ cos «) 822 = a7 2 sin? B - 812 = abc? (cosarcos B— cosy) 331 = abe (cos y cos a — cosh)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Qysialline State and Bonding 129 ik ss 2nd layer © y sites Fig. 1.25 Close packing of identical spheres. are nearest nieghbours to cach other. If we imagine the atoms to be "hard" spheres, then we see that the nearest neighbours “touch” cach other, and the distance between the centres of nearest neighbour atoms is the diameter of the atoms, Thus, in this hypothetical crystal, the diameter of the atoms is 5.00.4. In certain calculations (e.g. in resolved shear stresses), it is necessary to determine the angle between two different crystal directions. The formula for a general case like triclinic is very complicated. Therefore in Table 1.5, we enlist the formulae for some crystal systems which may be frequently used in calculations. The formula for cubic system is the simplest of all which can be determined by dot product. 12.43 Atomic packing : The role of interatomic bonding forces in the structure of materials will now be considered. The atoms in a solid state may be regarded for most purposes as rigid spheres with all the atoms of ‘one element having identical size. An obvious starting point is, therefore, the consideration of the kinds of structure that can be built up by packing identical spheres together as closely as possible, This is especially valid for metallic and van der Waals bonded solids of identical atoms where the bonding in nondirectional. A. single layer of spheres, pushed together, with themscives in a hexagonal closc packed network with six nearest neighbours as illustrated in Fig. 1.25. This layer offers recesses into which additional spheres can rest to form a Note: The ace issimple hexagenal LAT with aes 00) and (23 1/312) associated with ‘each lattice point. Unit cell (Spheres represented by lattice points) Fig. 1.26 The hexagonal close packed structure. second layer. It can be seen, however, that this second layer will occupy all the X sites or all the Y sites (occupation of the X sites is illustrated in Fig, 125). A third layer is again presented with altematives, either the spheres can be placed directly over those in the first layer or over the previously occupied Y sites. If the spheres in the third layer are positioned vertically above those in the first layer, a hexagonal lattice is formed which is known as the hexagonal close packed (h.cp.) structure with a stacking arrangement ABABAB --- - This is illustrated in Fig, 1.26, Placing the third layer of spheres in the Y sites ‘necessitates the fourth layer dircctly over the first layer. Thus, the stacking arrangement is now ABCABCABC-. etc. and a ‘face-centered cubic structure is built up. The axes of the cube, however, are neither in, nor1.30 Solid State Physics and Devices Close ‘packed planes ‘ Fig. 1.27 Face centered cuinc lumice showing the close packed planes perpendicular to, the plane of the close packed layers and actually the layers are (111) planes. Note the face-centered cubic lattice illustrated Fig, 1.27 is not the simplest building unit but Taken as the unit cell because it shows the maximum symmetry. Each atom has twelve nearest neighbours in both of these close packed structures; six in the close packed plane and three in each of the adjacent close packed planes above and below. ‘The number of nearest neighbour is called the coordination number. These two close packed structures should be favoured by solids with non-directional forces, ic. metallic or van der waals bonding, and indeed many metals do crystallise in the fee. structure, as do all the rare gases except helium which adopts the hcp. structure. Among the metals the hep. strucure is exemplified by cobalt and zine. Although these are the only two ways in which identical spheres can be closely packed together, a very important structure, where the packing ‘is slightly looser, is the body centered cubic (b.c.e.) structure. In this case cach atom has only cight nearest neighbours and this is indicative of the relatively loose packing. In fact, the packing fraction, or the fraction of the unit cell occupied by the sphereical atoms is 0.74 for both fec. and A.cp. but only 0.68 for b.cc. Inspite Of this looser packing many metals crystallise in the bc.c, form including all the alkali metals, and iron below 910K. Two possible reasons for the formation of this structure rather than the close packed structures are (a) the thermal energy of the atoms (c.g. alkali metais tend to change to the ce, phase at low temperatures and (b) partial covalent bonding A quantitative measure of the closeness of packing in a crystal is provided by the packing fraction, f, defined as _ volume occupied by the atoms (hard sphere) “volume of the unit celi of die structure The theoretical calculations of f required the knowledge of number of atoms, N, per unit cell and atomic radius, R, in terms of ‘a’, the Iength of a side of a cubic lattice. We have displayed in Table 1.6 the values of N, Ra and f alongwith the number (,) of — nearest neighbours and that (Nya) for next nearest neighbours for simple cubic (sc), body-centered cubic (bec), face-centered cubic (fec) and hexagonal close-packed (hep) structures Table 1.6, Certain Data for Common Structures. se | bee fee hep N 2 4 2 Nn 8 12 12 6 6 ‘The “closeness” of packing of the atoms in any given structure can be inferred from the number of nearest neighbours that surround each atom, This number, which is often referred to as the coordination number of the crystal, can range from a maximum of 12 in dlose-packed structures down to 4 in diamond and vincblend, whose structures arc relatively open, 1.2.14 Common crystal structures : Simple cubic structure. The unit cell for the simple (sc) lattice is shown in Fig. 1.25. OK = a, OB = a, and OC = a; can be chosen as the basis vectors and are given byCrystalline State and Bonding 131 aqaai a,-ap (1.12) as=ak and the lattice vector is given by Ry = nya, + paz + 343 Salm itm ftns& (1.13) For the sc lattice, we may choose the unit cell to be a cube of side a. The choice of position of the lattice point within the unit cell is arbitcary, We may arrange the cubes so that cack cube contains one lattice point at the centre (sce Fig. 1.28). Alternately, we may arrange the cubes so that the lattice points are at the corners of the cube (see Fig, 1.28b). The second choice is the conventional unit cell. Each of these unit cells contains one lattice point, This is obviously true for the unit cell in Fig. 1.28a which contains one lattice point in the centre. But the conventional unit cell in Fig. 1.28 appears to contain eight lattice ——* @ (b) Fig. 1.28 (a) The unit cell for the sc lattice with a lauice Point at the cenre and (b) with a tatice point at each corner (conventional unit cel) points, one at cach corner. Actually, cach of these lattice points is shared by eight ncighbouring unit cells that adjoin at the corner : 5 a] sameeren so that cach unit cell contains “¢ lattice point” at the corner. Eight such lattice points give us a total of one lattice point in the unit cell Just as the choice of basis vectors for a Jattice is not unique, the choice of unit cells is also not unique. For example, we could just as well choose the unit cell shown in Fig. 129. We only require that the unit cells be identical to each other and fiil all space. There are infinite number of ways to choose the unit cell. However, the conventional unit cell for the sc lattice is the cube shown in Fig, 1.29. The distance ‘a’ between adjacent atoms in the x. 3: or z direction is called the lattice parameter or constant. a Fig, 1.29 An alternate choice of unit cell for she sc lattice. Body-centred cubic structure. We can form a new lattice which is different from the sc lattice by placing an additional lattice point at the centre of the unit cell of Fig. 1.28. The resulting lattice is called body-centred cubic (bcc). Note that every lattice point is equivalent to every other lattice point. Each of the ori sc lattice points is also in the body-centred Position of eight of the new lattice points. The conventional unit cell for the bec lattice is a cube of side ‘a’ as shown in Fig. 1.30. We see that there are two lattice points in fie ane 1 this unit cella lattice point in the centre and 4132 Solid State Physics and Devices Fig. 1.30 The conventional unit cell of the bec lattice. lattice point at each of the eight corners. This is not the smallest possible unit cell which can be constructed for the bcc lattice. The smallest possible unit cell, called the primitive unit cet, contains only one lattice point. Its volume is half the volume of the conventional unit cell shown in Fig. 1.30. The choice of the primitive unit cell is not unique. There are an infinite number of possible ways to choose it. However, we usually use the primitive unit cell called the Wigner-Seitz cell. For the bec lattice, the Wigner-Seitz cell is a ‘truncated octahedian’ with a volume of a°/2. These cells nest together and fill all space. This cell contains one lattice point at its centre and contains the region of space that is closer to that point than to any other lattice point. The typical basis vector of a bee lattice can be obtained by referring to Fig. 1.27, where 2-0 - ORs MT-ORs RAH PS Fete SS vis or a= 5+ vis Rie Thus, a very common way to choose the basis vectors of a bee lattice is (1.14) In the bec lattice, points along the [111] direction are nearest neighbours. The distance between them is the length of the vector Ry aya (i+ f+ ® (1.15) which is V3a/2 = 0.866a. Face-centred cubic structure : Yet another lattice can be formed from the sc lattice by placing a lattice point at the centre of each face of the unit cell shown in Fig. 1.28b. The resulting lattice is called face-centred cubic (fcc). As with the bee lattice, every lattice point in the fcc lattice is equivalent. For conventional unit cell for the fec lattice is a cube of side ‘a’ shown in Fig. 131. Let us count the number of lattice points in this unit cell. There are eight corners (cach containing i lattice point) and six faces (each containing } lattice point) which gives us a total of four lattice points. Fig. L3L The conventional unit cell of the fee latice This, of course, is not a primitive, unit cell for the fcc lattice. The Wigner-Scitz primitive unit cell for the fcc lattice is a rhombic dodecahedron. It contains one lattice point, and thus its volume is a°/4.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Crystalline State and Bonding l.glg.l 004;030;300 Each atom has as nearest neighbours six atoms of the opposite kind, so that the coordination number is 6. Representative crystals having the NaCI arrangemem include those in the immediately following table. Crystal aA) Crystal a4) Lik 408 NHg = 724 NeCl 563 Agr 5.77 KBr 659 MgO 4.20 RBI 133 MnO 443 PbS 5.92 uo 492 CsCI structure : ‘The cesium chloride structure is shown in Fig. 134. There is one molecule per unit cell, with atoms in the body-cenired positions : The space lattice is simple cubic. Each atom is at the centre of a cube of atoms of the ‘opposite kind, so that the coordination number is 8. Representative crystals having the CsCl arrangement include those in the immediately following table. Crystal «= a(A) | Cystal a4) cscr 4.11 | CuZn Bbrass) 294 Br 397 | AgMg 3.28 ™ 420 | Lig 329 NHACt 387 | AINE 288 RBCI(I90°C) 3.74 | BeCu 270 Fig L346 CsCl crystal stucture 135 13 X-ray Diffraction 13.1 Introduction Tf we shine electromagnetic radiation into a crystal, cach atom in the crystal individually scatters some of the radiation. Thus, a crystal with N atoms becomes N different sources of radaition. The waves radiated from these N sources interfere with each other, and we find at some angles that these waves constructively interfere, giving rise to intense spots on a screen. In general, interference effects of waves can usually be observed in any object. which contains a number of repeating units (atoms in a crystal), For the effect to be most easily measurable, we must usc waves with a wavelength comparable to the distance between the repeating units, We need to use the wavelength 2 =d, where d is the interplanar distance. If we use A > d, there is no solution to the Bragg’s law except for n = 0 so that we only observe the central spot at @=0. We observe no diffraction spots. If we use A < < d, then we see from Bragg’s equation that the diffraction spots are very close together. Overlap of spots may cause them to be unresolvable. ‘When the atoms of a crystal are exposed to electromagnetic radiation, such as X-rays, they experience electrical forces due to the interaction of the charged particles of the atoms with the electric field vector of the clectro- magnetic wave. The atomic electrons are therefore vibrated harmonically at the frequency of the incident radiation, thus undergoing acce- eration. These accelerated charges, according to electromagnetic theory, reradiate clectro- magnetic energy at the frequency of vibration, that is, at the incident wave frequency. At visible light frequencies, where the incident wavelength is much larger than the interatomic distances, the superposition of the waves thus reradiated or scattered by the individual atoms ‘of the crystal simply gives rise to the well- Known effects of optical refraction and reflec- tion, At x-ray frequencies, however, the incident wavelength is comparable to the interatomic spacing, and the diffraction of the radiation by the atoms of the crystal can be observed. 1. Diffraction by a crystal was first observed for x-rays,1.36 Solid State Physics and Devices 2. x-ray diffraction is the method most commonly used to determine the structure of a crystal. 3. For x-rays, A~ 14 =d. The diffraction condition is satisfied and hence x-rays are diffracted by crystals. 4. For A=1A, Ex 10 eY x-rays can, therefore, penetrate deep down in: the bulk of ‘the crystal and hence extract information from there. 5, x-rays scatter mainly from electron charge density of the atom, 6. x-rays interact much less strongly with atomic nuclei than electrons. Diffraction of x-rays from H-atom-aiom is hardly observed because of its very low charge dersity. 7. Magnetic fields have no effect on x-rays at all. Therefore magnetism associated with atomic nuclei gives no response to x- rays. 8. Each part of the atom scatters x-rays difforently. 9. Since every primitive unit cell in the crystal is identical in contents, cach scatters x-rays in an identical manner. 1.3.2 Bragg’s Law : A crystal consists of a regular array of atoms, each of which can scatter em. waves. The mechanism of scattering is straightforward. An atom in a constant electric field becomes polarized since its negatively charged electrons and _ positively charged nucleus experience forces in opposite directions. Thesc forces arc small compared with the forces holding the atoms together, and so the result is a distorted charge distribution equivalent to an clectric dipole. In the presence of the alternating electric field of an em. wave of frequency v, the polarization changes back and forth with the same frequency v. An oscillating clectric dipole is thus created at the expense of some of the energy of the incoming wave. The oscillating dipole in turn radiates e.m. waves of frequency v, and these secondary waves go out in all directions except along the dipole axis. (In an assembly of atoms exposed to unpolarized radiation, the latter restriction does not apply since the contributions of the individual atoms are random.) Fig, 138 The scatcring of lecro-magnetic radiation by @ group of atoms. Incident plane waves are reemitted ‘spherical waves In wave terminology, the secondary waves have spherical wave fronts in place of the plane wave fronts of the incoming waves (Fig. 1.35). The scattering process, then, involves atoms that absorb incident plane waves and rcemit spherical waves of the same frequency. ‘A monochromatic beam of x-rays that falls upon a crystal will be scattered in all directions inside it. However, owing to the regular arrangement of the atoms, in certain directions the scattered waves will constructively interfere with one another while in others they will destructively interfere. The atoms in a crystal may be thought of as defining familics of parallel planes, with each family having a characteristic separation between its component planes. This analysis was suggested in 1913 by WL. Bragg, in honour of whom the above planes are called Bragg planes. The conditions that must be fulfilled for radiation scattered by crystal atoms to undergo constructive interference may be obtained from a diagram like that in Fig. 1.36. A beam containing x-rays of wavelength A is incident upon a crystal at an angle 8 with a family of Bragg planes whose spacing is d. The beam goes past atom A in the first plane and atom B in the next, and cach of them scatters part of the beam in random directions. Constructive interference tekes place only between those scattered rays that are parallel and whose paths differ by exactly 4, 2%, 3, and so on. That is, the path difference must be nA, where n is an integer. The only rays scattered by A and B for which this is true are those labelled I and II in Fig. 1.36.Crystalline State and Bonding ig. 1.36 Xray scatering from a cubic crystal. The first condition on I and I is that their common scattering angle be equal to the angle of incidence 6 of the original beam. (This condition, which is independent of wavelength, is the same as that for ordinary specular reflection in optics: angle of incidence = angle of reflection.) The second condition is that Qdsin@=nd n= 1,235" (1.17) since ray I must travel the distance 2d sin 0 farther than ray I. The integer n is the order of the scattered beam, and also called the order of the corresponding’ reflection. Since sin@ < 1, then 4s 24, gives a maximum value for the wavelength of the x- rays that can be used. From this we see why visible light cannot be used. Figure 1.37%(a) shows how the Bragg angle is determined experimentally as just half the angle between the incident and reflected beam. From eq. (1.17) for 1=2 we would measure a larger value of @ than for the values obtained from n= 1. It is not too difficult to show that the angle corresponding to this n = 2 reflection corresponds to the angles from other prominent crystallographic planes. Such a possibility ean be seen in Fig. 137(b) for the identical structure as in Fig. 1.29. Then Brage’s law can be reinterpreted to be 2d! sin =A (1.18) where d’ = din is the spacing between planes of the crystal. It is clear from Fig. 137(b) that there are an infinite number of such planes, although the ones for very large n have a low density of atoms and hence low scattering amplitude. 137 Fig. 1.37 (a) The Bragg angie 8 which is half the angle Dreween the incident and scanered beam. (b) The same dattice as in Fig 1.36, but emphasizing a different set of planes, Clearly there are an infinite ‘01 of planes. In order that Bragg’s law can be satisfied and X-ray diffraction can occur the various values of @ and 1 must be related. It is necessary in practice to vary either the angle of inclination of the specimen to the beam or the wavelength of the radiation. The schematic design of an x-ray spectrometer based upon Bragg’s analysis is shown in Fig. 1.38. A narrow beam of x-rays falls l Coltimator Fig, 1.38 Xray specnometer.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Crystalline State and Bonding |, 2, 3 in eq, (1.20b), we get ay Ak = 2, ay Ak = 20, ay-Dk = 20; These are the alternative forms of Laue equations. These are valuable for their simple geometrical interpretation: The first of these equations a,-Ak = 2xv, tells us that Ak lies on a certain cone about the direction of a;. The second equation tells us that Ak lies on a cone about az as well, and the third equation requires that Ak lies on a cone about a3. Thus, at a reflection Ak must satisfy all three equations; it must lie at the common lin of: intersection of three cones, which is a stringent condition that can be satisfied only by systematic sweeping or searching in wavelength or crystal orientation —or else by sheer accident. Finally, it may be stated that the Laue equations provide one useful avenue for discussions of crystal symmetry and structure. If the atomic scattering centres are on the (hkl) planes, vy, vz and v3 can be respectively identified with A, k and / to make it possible to write the Laue equations as ay-Ak = 2h ay: Ak = Ink ay: Ak = 2a! and (1.20) It must be noted that for a given value of 4 and an arbitrary direction of incidence My, it is in general not possible to find a direction which satisfies (1.20a). In other words, for a monochromatic x-ray beam falling on a crystal with an arbitrary direction of incidence, in general no diffraction is observed, This may readily be understood by remembering that for a two-dimensional lattice there exist only ic directions for a diffracted beam and ese directions are in general not part of the direction cones determined by ‘the third condition required for the. three-dimensional case, Thus only for particular angles of incidence will diffraction be observed: it is exactly this limitation that makes’ x-ray 139 diffraction a useful tool for investigating crystal structures. This poiat will become more clear when we discuss the Bragg formula below. Before doing this, it may be useful to rewrite (1.20a) for the case of a simple cubic lattice. Assuming a = b = c, we obtain from (1.20a) by squaring and adding cos Bcos By ~ cos y cos yo) a+ e+e? Now, if ¢ represents the angle between the incident and scattered beams, we may write 2(1 — cos ¢) = 4sin? (/2) 2a*(1 — cos acos ay — 22) ace ya A = 15 (yr +20 + v3) In this form von Laue equation is closely related to the Bragg formula, Experimental Method of Laue Spats : (1) The Laue method is used for routine orientation of single crystal samples. (2) In this method a stationary single crystal is irradiated with a continuous wavelength x-ray source. 3) The various sets of planes in the crystal select, from the white radiation, the appropriate i for the diffraction condition to hold. The particular @ then determines the exposure on the film, (4) The resulting pattern (Laue spots) shows the symmetry properties of the crystal. (5)_In this manner crystals ‘canbe oriented. For example, if the crystal is mounted so that it is @ degrecs from the 6-fold axis, the resulting pattern shows this very clearly. Then the goniometer can be rotated by— a degrees to obtain the proper crystal direct (6) Typically such oriented single crystals are then cut by a saw and later used in experiments, or the crystals could be used while mounted on the goniometer. (7) Laue method can be used for a rapid determination of the symmetry and orientation of a single crystal. ‘The Laue method is very simple in concept and operation. A narrow pencil beam of x-rays is allowed to fall on a single crystal, as illustrated in Figure 1.40, This beam contains140 Solid State Physics and Devices xrays over a wide wavelength range, and for any wavelength satisfying the Bragg condition a diffracted beam will emerge. The figure shows two positions in which a photographic plate can be placed to record a set of diffraction spots’ which is characteristic of the structure and orientation of the crystal. The set of diffraction spots can appear very confusing when the structure of the crystal is ‘not known, because several wavelengths, corresponding ‘with different orders for the same set of planes, may be reflected to the same location on the plate. For this reason, the Laue method is rarely used for investigation of new structures. Its main use lies in the routine orientation of known crystals. Thus Figure 1.40(a) shows the sample crystal attached to a tiple-axis goniometer, the familiar array of spots (Figure 1.40(b)) produced after a crude attempt at orientation can be measured to determine what fine adjustment of angle is necessary for perfect alignment of a major crystal plane with the axis of the Laue camera. Flat films are placed in front of and behind the specimen. Since A covers a continuous range, the crystal selects that particular wavelength which satisfies Bragg’s law al the present orientation, and a diffracted beam emerges at the corresponding angle. The diffracted beam is then recorded as a spot on the film. But since the wavelength correspon- ding to a spot is not measured, one cannot determine the actual values of the interplanar spacings—only their ratios. Therefore one can determine the shape but not the absolute size of the unit cell. A typical Laue photograph is shown in Fig. 1.40(b): Note that if the direction of the beam is an axis of symmetry of the crystal, then the diffraction pattern should exhibit this symmetry. Figure 1.40(b) shows the G-fold symmetry of the symmetry axis in Mg, which has the hexagonal structure. 1.4 Bonding 14.1 Introduction : The problem of solid cohesion has proved itself to be one of the most difficult to answer satisfactorily. Qualita- Collimator Back reflection plate position o) Fig. 1.40 The Lave method: (a) Experimental arrangement. (8) Laue pattem for an Mg crystal, with the x-7ay ‘beam parallel to the 6-fold symmeny axis. tively it is certainly possible to understand why solids occur. There are short-range attractive forces between atoms and if the temperature is, sufficiently low so that these forces are morc important than the effects of thermal agitation the atoms will condense to form a solid. One of the fundamental questions a complete solid state theory must answer is ‘Why docs an assembly of atoms generally form a periodic structure in the lowest energy state?’ To date no complete answer has been given and it is necessary to accept the periodicity as a fact and then attempt to calculate the energy of the lattice as a function of the lattice parameter. ‘The cohesive energy of a solid is defined as the energy of the crystal in the lowest energy state at absolute zero minus the sum of the ground state energies of the isolated atoms and is a negativequantity. The magnitude of the cohesive energy is called the binding energy (or heat of sublimation) and is equal to the work required toCrystalline State and Bonding separate the crystal into its constituent atoms. For a macroscopic crystal for which the surface effects may be ignored, the cohesive energy is proportional to the crystal volume and may sensibly be expressed in electron volts per atom or per molecule. (Sometimes expressed in kcal per mole and 23.052 kcal per mole is equivalent to 1 cV per molecule). Various properties of solids can be explained by differcnt types of bonds that exist between the atoms constituting them. The bonds in solids are produced by forces between atoms, ions and molecules. The binding forces can be divided into two classes: (i) interatomic, the examples of which are ionic, covalent, hydrogen bond and repulsive forces, (ji) inter molecular, that is, the van dor Waals forces. An additional type of force, called the exchange force and which is attractive in nature, results when elections are not localized near one atomic centre but their position alternatively varies between more atomic centres, ic. the electrons are shared by a number of centres, In principle the crystal encrgy may be obtained by finding the expectation value of the many-particle Hamiltonian (including both electrons and ions) regarding the ions as fixed. In this way tal energy is found as a function of the ionic separation. Of course, to do this accurately it is necessary to express the many particle wave function as a sum of determinantal wave functions. This procedure has proved most difficult im practice and resort has to be made to more approximate methods Solids may be roughly classified by the strength and nature of their cohesive forces into five major types. They are: (a) molecular crystals such as solid nitrogen with binding energies of up to 0.1 eV per molecile; hydrogen-bonded crystals such as ice with binding energies of about 0.5 eV per ‘molecule; ionic crystals such as sodium chloride with binding energies of about 5—10 eY per molecule; valence crystals of which diamond is a good example with binding energies of the order of 10 eV per molecule, and ) © ) 141 (©) metals with binding energics of 1-5 cV per molecule. This classification is somewhat artificial and many solids fall into intermediate categories. An example is zinc sulphide which is both ionic and valence-like in character. The rest of this chapter is devoted to a semi-quantitative description of these five main types of solid. 1.4.2 Repulsive Forces : When compresing solids we find that an increasing pressure causes a decrease in their volume and initiates a mutual repulsion of particles and thus an increasing resistance to the compression. When the pressure (acting within clastic limits) is lowered, the substance returns to the initial state. From this we may derive that in a solid repulsive forces begin to act with the attraction of particles under pressure. The equilibrium state of a solid is determined by an equilibrium of attractive and repulsive forces. The action of the repulsive forces rapidly increases with decreasing distance between atoms and is usually expressed by the relation between the potential energy Uj, and the atomic spacing r : ’ pep = > hep = From the theoretical point of view, the exponential equation for Upgp is more correct: r thy =ben (-2) where b, n, p are parameters describing the repulsive forces. After combining the formulae for the energies Ugur = “% and Upp of the attractive and repulsive forces, respectively, we get a relation for the resulting energy : a,b r tne 53s] Ot [ la where @ is the attractive force constant, a, b > 0, To satisfy the empirical condition stating that attractive forces act at greater distances and repulsive forces only at short distancesLa Solid State Physics and Devices Fig Lat, around the equilibrium state, we must have n>m. The values of the exponent m do not fall below 1, and the values of n are between 2 and 10. The variation of the attractive and repulsive forces and their resulting energy between two particles as a function of their separation r is shown in Fig. 1.41. For a certain distance r, determining the equilibrium position of particles of a solid, the resulting energy has a minimum value. 1.43 Interatomic Potential : The potential between a pair of atoms is referred to as the interatomic potential. ‘The attraction between two atoms brings them closer together until the individual electron clouds begin to overlap and a strong repulsive force arises to comply with Pauli's exclusion principle. When the attractive force and the repulsive force between any two atoms are equal, the two atoms should be in a stable situation with a minimum potential energy, corresponding to the equilibrium interatomic distance, r. Consider two atoms in their ground states with an infinite separation, and consequently an interaction potential energy equal to zero. As the atoms are brought together, the potential cnergy of cither atom will be given by pn [Aree in potetl enya cxresion = + increase in potential energy due to repulsion Any valid interatomic potential, must essentially satisfy the following conditions () Asr+0,U+@ Gi) In the limit r> 0, U>0 (ii) U must show a minimom for the solid to be stable, The minimum must correspond to attraction. From the very existence of solids one may draw two general conclusions: (1) there must act attractive forces between the atoms or molecules in a solid which keep them together, (2) there must be repulsive forces acting between the atoms as well, since large external pressures are required to compress a solid to any appreciable extent. (Both conclusions also apply to liquids.) In order to illustrate the importance of both types of forces, let us consider the simplest system in this respect, viz., a single pair of atoms A and B which form a stable chemical compound. Without paying attention to the physical origin of the forces between the two atoms, let us assume that the potential energy of atom B duc to the presence of atom A is given by an expression of the type a b +7 where a, b, n, m are constants characteristics for the AB molecule of the metal and r is the interatomic distance. This is a case of a simple type of interatomic potential, which has been employed for other solid types as well. ‘The zero of energy is chosen such that for infinite separation p=0. The first term, which is nogative, corresponds to the cncrgy associated with the forces of attraction, the second (positive) term corresponds to the forces of repulsion. In order to give a minimum potential energy condition at the equilibrium separation to, ‘m in equation (1.21) must be greater than n so that BU 5, sit <> is positive op Pes u (121)Cystalline State and Bonding The resultant force of interaction, F is given by -F (12) so that na mob Peal nal (1.23) At the equilibrium separation, the net force is zero, ie. na mb ee eT (1.28) The energy and the force between two atoms A and B which form a chemical compound are represented in Fig. 142. The stable configuration for the system corresponds to the minimum in the U(”) curve, which occurs for a particular separation" r=7. The corresponding energy U((o) is negative; thus the positive quantity D = — U(r) is the dissociation energy of the molecule, ic., the energy required to separate the two atoms. Dissociation may occur, for example, at high temperatures or as a result of other processes in which the molecule can absord suficient energy. The dissociation energies are of the order of onc or a few electron volts. ‘Assuming that the energy curve exhibits a minimum, one may express the equilibrit distance "7, and the corresponding binding F t @ ® Fig. 1.42 Schematic representation of (a) the energy and (b) force between 180 aioms as function of their seperation r. The dashed curves are the sums of the ‘atnactive and repulsive curves. 143. energy U(?,) in terms of the constants a, b, m, and n by making use of the condition a) &) =0, (2). e-fa)) os According to (1.24) this condition is equivalent to the requirement that the attractive and repulsive forces balance, ie, F(ra) = 0. Substituting from (1.23) into (1.21) one obtains for the energy in the equilibrium state ie, Ut % Note that although the attractive and repulsive forces are equal in equilibrium, the attractive and repulsive energies arc not equal since nem, In fact, if m>n, the total binding energy is essentially determined by the energy of attraction —a/7. ‘As one may oxpect already by looking at Fig. 1.42, a minimum in the energy curve is possible only if m >; thus the formation of a chemical bond requires that the repulsive forces be of shorter range than the altractive ones. This may be shown readily by employing the condition that (@°U/d?),.), > 0 if U(?) must have a minimum at 7, In fact, this condition leads to =n(uta , m(m + ib me gee 7 which upon substitution of ry from (1.25) immediately gives (27) m>n (1.28) Although the energy can in general not be represented accurately by a power function of the type (1.21), the above treatment provides some useful qualitative conclusions which may be extended to solids. The forces acting between the atoms in solids are electrostatic in nature; they are determined essentially by the way in which the outer electrons of the composing atoms are1d Solid State Physics and Devices distributed in space. The physical properties of solids are determined to a large extent by the electron distribution, and it is thus possible on an empirical basis to divide solids into different groups corresponding to different types of electron distributions. One may distinguish between the following extreme types: 1. Tonic erystals (NaCl, KF) 2. Valence erystals (diamond, SiC) 3. Metals (Cu, 4g Mg) 4, van der Waals crystals (argon, many organic crystals)It should ‘be said from the outset that many intermediate cases occur and in general one must be somewhat careful in employing very specific labels. 1.4.4 Lennard-Jones Potential : The inert gases form the simplest crystals. The clectron distribution is very close to that of the free atoms. The crystals are transparent insulators, weakly bound, with low melting temperatures. The atoms have very high ionization energies. The outermost electron shells of the atoms are completely filled, and the distribution of electron charge in the free atom is spherically symmetric. In the crystal the incrt gas atoms pack together as closely as possible: the crystal structures are all cubic close-packed (fcc), except He? and He’, ‘What holds an inert gas crystal together? The electron distribution in the crystal cannot be significantly distorted from the electron distribution around the free atoms, because the cohesive energy of an atom in the crystal is only 1 percent or less of the ionization energy of an atomic electron. ‘Thus not much energy is available to distort the free atom charge distributions. Part of this distortion gives the van der Waals interaction. van der Waals—London Interaction Eyen in an atom or a molecule which has on the time-average a zero electric dipole moment there will be a fluctuating dipole moment associated with the instantaneous position of the electrons in the atom, The instantaneous electric field associated with the moment will induce a dipole moment in neighbouring atoms. The average interaction of the original moment and the induced moment gives rise to an attractive force between the atoms. This idea of small but finite attraction between atoms gains support from the existence of solid and liquid phases of the rare gases. These forces arc known as dispersion or van der waals forces, and have been treated on a quantum mechanical basis. Consider two identical inert gas atoms at a scparation R large in comparison with the radii of the atoms. What interactions exist between the two neutral atoms? If the charge distributions on the atoms were rigid, the interaction between atoms would be zero, because the electrostatic potential of a spherical distribution of electronic charge is cancelled outside a neutral atom by the electrostatic potential of the charge on the nucleus, Then the inert. gas aloms could show no cohesion and could not condense. But the atoms induce dipole moments in each other, and the induced moments cause an attractive interaction between the atoms. ‘As a model we consider two identical linear harmonic oscillators 1 and 2 separated by R. Each oscillator bears charges te with separations x, and x, as in Fig. 143. The particles oscillate along the x axis. Let p, and Pz denote the momenta. The force constant is k. Then the hamiltonian of the unperturbed system is a1 =a 1 1 1 Hy ft sits Pet zee (129) hr OWLS @WWNLO b—— 2 ——1——— — 4h Fig 1.43 Coordinates of the two oscillators.Crystalline State and Bonding Each uncoupled oscillator is assumed to have the frequency w, of the strongest optical absorption line of the atom, Thus k = mw. Let Ay be energy of the two oscillators, The geometry is shown in the figure. The internuclear coordinate is R. Then =f pe R™ Rtx, =x, R+x, Rox’ (1.30) in the approximation |x;|, |x| <.R we expand (1.30) to obtain in lowest order : the coulomb interaction 2x, H,=- 131) 1 7B ¢ The total Hamiltonian with the approximate form (131) for H, can be diagonalized by the normal mode transformation xm # ete); x e @-=), (132) or, on solving for x, and x, + 1 Hees a= bem) — 033) The subscripts s and a denote symmetric and antisymmetric modes of motion, Further, we have the momenta p,, pz associated with the two modes : =p) (134) The total Hamiltonian Hy + Hy after the transformations (1.33) and (1.34) is 1 1 PLS 7 OstPa)i Pa= Te + [Bet ala] (1.35) The two frequencies of the coupled oscillators are found by inspection of (1.35) to be 145 (1.36) with w, given by (kim). In eq, (1.36) we have expanded the square root. The zero point energy of the system is 5 (w,+ v9); because of the interaction the sum is lowered from the uncoupled value 2: 3a, by au= in (Aa + Aavg) » 5,2)? ee fos = 37) 8 lar? This attractive interaction varies as the minus sith power of the scparation of the two oscillators. This is called tht van der Waals interaction, known also as the London interaction or “the induced dipole-dipole interaction. It is the principal attractive interaction in crystals of inert gases and also in crystals of many organic molecules. The interaction is a quantum effect, in the sense that AU>0 as 7-0. Thus the zero point energy of the system is lowered by the dipole-dipole coupling of Eq. (1.31). The van der Waals interaction does not depend for its existence on any overlap of the change densities of the two atoms. An approximate value of A for identical atoms is given by Nwg2*, where Tuy is the energy of the strongest optical absorption line and a is the electronic polarizability. Repulsive interaction : As the two atoms are brought together their charge distributions gradually overlap (Fig. 1.44), thereby changing1.46 Solid State Physics and Devices Fig. 44 Eleewonic charge distributions overlap as atoms approach. The solid circles denote the nuclei the electrostatic energy of the system. At sufficiently close separations the overlap energy is repulsive, in large part because of the Pauli exclusion principle. The clementary statement of the principle is that two electrons cannot have all their quantum numbers equal. When the charge distributions of two atoms overlap, there is a tendency for electrons from atom B to occupy in part states of atom A already occupied by electrons of atom A, and vice versa.The neighbouring ions, in equilibrium, are so close that their clectron clouds overlap with each other. It may be recalled that the Pauli’s exclusion principle forbids the simultaneous occurrence of two electrons inthe same state of motion and at the same location in space. According to the Pauli’s principle the only way in which two equivalent electrons may occupy the same space is that one of them is excited to a higher energy state. Now the energy generally required for such an excitation of one of the ions is not available to the other one. Therefore, there is no likelihood that the two electrons may occupy the same space. Consequently, the outermost electronic clouds tend to push each other towards the respective nucleus, This clucidates the existence of repulsive forces which are necessary for the stability of the lattice. The repulsive forces are sometimes referred to as the overlap forces, because their magnitude depends on the amount of overlap of the wave-functions of the ®@. @- + + tS) N @- v ¥ Ast 2s & IsPisy 6 Total spin one (a) () Total electron Total electron energy : ~78.98 eV energy: -3938 eV Fig, L45 The effect of Pauli principle on the repulsive ‘nergy: in an extreme example, ovo hydrogen atoms ‘are pushed together until the protons are almost in contact. The enengy of the electron sysiem alone ‘can be taken fram observations on atomic He, which has two ‘electrons. tn (a) the electrons have ‘antiparalie! spins and the Pauli principle has no effec: the clecirons are bound by —78.96. e¥- in (b) the spins are parallel: the Pauli principle forces the promotion of an electron from a 1s t orbital of H 1 a 2s} erbitel of He. The electrons now are bound by —59.38 ev; less than (a) by 19.60 ‘eK This & dhe amount by which the Pauti principle hhas increased the repulsion. outer electrons of the interacting neighbouring ions. The repulsive forces are, therefore, short range and non-clectrostatic in nature. Thus the clectron overlap increases the total cnergy of the system and gives a repulsi contribution to the interaction. An extreme example in which the overlap is complete. is shown in Fig. 1.45, If atoms were subjected only to attractive forces all atoms would coalesce. Thus it is clear that for very short distances of scparation some type of repulsive force must operate. In ionic systems the attractive force arises from the electrostatic charge and the repulsive force from the electron shells, which act as a sort of tough clastic sphere resisting further compression. The repulsion may be described as arising from two effects. First, theCrystalline State and Bonding penetration of one electron shell by the other, which means that the nuclear charges are no longer completely screened and therefore tend to repel one another. The other effect arises from the Pauli exclusion principle which states that two electrons of the same energy cannot occupy the same region of space. For them to be in the same space (overlapping) the energy of one must be increased — this is equivalent to a force of repulsion. With covalent and van der Waals forces the repulsion and attraction are all part of a single mechanism and it is not correct to consider them as arising from separate mechanisms; however, for the purpose of simple calculation it is very’ convenient to do so. We therefore describe the long range attraction by a term of the form A F=r. (1.38) 2 : and the repulsion by a relation of the form B Fo-, 1.39) ae f where B is a constant and n has a value of the order 8 to 10. In fact the power law relation is not very good and there are theoretical reasons for preferring a relation for the repulsive force of the form Be“? where B and p are suitable constants. In the following calculation we shall for simplicity retain the power law for both attraction and repulsion, The resultant force is then Fa4-8 (1.40) and this is plotted in Figure 1.46, The equilibrium separation x, occurs when F=0, This gives B = Ax,"~", Hence eq, (1.40) becomes: . oem FoA [: (1.41) We may at once note that for small displacements from the equilibrium position the restoring force is : 147 5 Ny tL S| Nour 3] Neuraction 3 ° F potential energy U ¢ attraction © Fig, 146 (a) Force ancl (b) potential eney curves for wo ‘aioms (or molecules) as @ function of separation. There is a longrange atwactive force and a shortrange repulsive force which operates only when the molecules come close sogether The equilibrium separation x, occurs when the net force is zero o when the potential energy is a minimum. aF=A eg ae Since the force is proportional to the displacement the motion of the displaced particle will be simple harmonic motion. This conclusion is, of coursc, not restricted to power law relations. Any relation of the form F = f(x) will lead lo a similar conclusion. Born and Mayer suggested the following formula for the repulsive term of the energy : (143) (1.42) U, = Acxp (rp), where A and p are usually known as the potential parameters. The potential expressed by eq. (1.43) is also called the exponential potential.148 We make no attempt here to evaluate the repulsive interaction from first principles. Experimental data on the inert gases can be fitted well by an empirical repulsive potential of the form B/R', where B is a positive constant, when used together with a long-range attractive potential of the form of (1.37). The constants A and B are empirical parameters determined from independent measurements made in the gas phase; the data used include the virial Coefficients and the viscosity. It is usual to write the total potential energy of two atoms at separation R as 2 6 2 o w= [a -( 3) where © and o arc the new parameters, with 4e0° =A and 4eo!? = B . The potential (1.44) is known as the Lennard-Jones potential, Fig. 1.47. The force between the two atoms is given by —dU/dR. Values of ¢ and o can be obtained from gas-phase data, so that calculations on properties of the solid do not involve disposable parameters. (1.44) 6>—— 5 p 3 UR te 2 02 04 0.6 08 10-12 o> Fig 1.47 Form of the Lennard-Jones potential which describes the interaction of wo iver gas atoms. The minimum occurs at R/o = 2 = 1.12 Notice how sicep the curve is inside the minimum, and how flat itis ouside the minuuan The valuc of U atthe minimum is ~ e; and U = 0 at R =a. Other empirical forms for the repulsive interaction are widely used, in particular the exponential form 4 exp(—R/p), where p is a measure of the range of the interaction. ‘This is Solid State Physics and Devices generally as easy to handle anal inverse power law form. Equilibrium lattice constants + If we neglect the kinctic energy of the inert gas aloms, the cohesive energy of an inert gas crystal is given by summing the Lennard-Jones potential (1.44) over all pairs of atoms in the crystal. If there are N atoms in the crystal, the total potential energy is lly as the 1 Ue = ZNO) * ie (1.48) where pyR is the distance between reference atom i and any other atom j, expressed in terms of the nearest neighbour distance R. The factor 3 occurs with the V to compensate for counting twice each pair of atoms. The summations in eq. (1.45) have been evaluated, and for the fee structure = 12.13188; 3" pj i There are 12 nearest-neighbour sites in the fec structure; we see that the series are rapidly converging and have values not far from 12. The nearest neighbours contribute most of the interaction energy of inert gas crystals, The corresponding sums for the hep structure are 12.1329 and 14,45489, If we take Ug in (1.45) as the total ‘energy of the crystal, the equilibrium value Rg is given by requiring that Uj, be @ minimum with respect to variations in the nearest neighbour distance R. 14.45392 (1.46) aw, “R7? oe of = = 2Ne }(12)(12.13) 75 ~ 6(14.45) | (47) whence R/o = 109, (1.48)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.154, overlap between the ions is small and so there is a large gap between the filled and empty electron bands. This explains why sodium chloride is an insulator. The ions are arranged in the lattice so that the Coulomb attractive force between unlike ions exceeds the repulsion between like ions, It is well-known that 90 stationary configuration of charges can be in stable equilibrium under their own electrostatic forces (Earnshaw’s theorem). The additional force (as in the case of molecular crystals) is the quantum mechanical short range repulsion between the ions which opposes the overlap of the electron distributions. Of coruse there are other forces as well, such as the van der Waals forces and also the van der Waals type forces produced by the polarization of the ions by the charges on the other ions. These are not two-body interactions and are more difficult to deal with, but fortunately their effects are small. In the latter case the symmetry of the lattice makes the fick! due to the neighbours cancel to a large extent and in particular the field vanishes at the centre of the ions.Within the two-body interaction approxi- mation the total crystal potential energy is Us = ZU; ayy (1.57) where Uj is the interaction energy between the i and j ions and the sum includes all pairs. The short range repubion term is duc to the overlap of the ion structures and may sensibly be represented by Je“? where 4, p are constants and rj is the distance between the ions. (Originally the interaction energy was taken as Any" with n ~10.) 4 (ope sp * Gant if U; (1.58) ¢ is the magnitude of the charge on each ion and the positive and negative signs are taken for unlike and like charges respectively. If the crystal contains N molecules each with two ions Solid State Physics and Devices then ignoring surface effects the electrostatic contribution to the total energy is obtained by summing the first term in and is (1.59) ris some lattice parameter and may be taken to be the nearest neighbour distance, @ is a positive aumber called the Madelung constant. (160) @ does not depend upon the lattice distance but only upon the crystal structure. A typical value is a= 1.74% as for the NaCt structure. The summation (1.60) is only conditionally convergent and the arrangement of terms is important. This is a consequence of the long range characteristic of the Coulomb force law. ‘The exponential term in (1.58) falls off rapidly and is only important for nearest neighbours. This non-Coulombic contribution to the total energy is =NMle~"? 1.61) where M is the coordination number (M = 6 for NaC). With this approximation the total lattice potential energy 2 ner ven|na tue] (1.62) ‘The quantity in brackets is the cohesive energy per molecule. This expression for U contains three parameters r, A and p. At the equilibrium separation © = 0, This condition may be used to climinate A and writing r = ry leads to Neva [: _ A fo For a given crystal a and ry are known and so it remains only to find p. This is done indirectly by extrapolating the observed compressibility to absolute The compressibility f is defined by Us (1.63) are9% vero.Crystattine State and Bonding =“ (164) where P and V are the pressure and crystal volume respectively. From the first law of thermodynamics dU = 60 - Pav and at very low temperatures 69 ~0 and dU = —PdVv (1.65) The compressibility at absolute zero is then 1_ av ae 1.66) Bar (166) The crystal volume Vis equal to the number of ions multiplied by the volume associated with each ion, For NaCi V= (2Ny° and then from (166) and (1.62), after some algebraic manipulation, it is scen that | 2+ 4) (167) lo 18 (Are ro Bi If B, is found from experiment then the ratio p/r, can be calculated and used in (1.63) to calculate the lattice potential energy. This value may thea be compared with the observed cohesive energy and provides a check on the theory. For the alkali halides p/ry ~ 1/10 and in this case Nea 9 U=~ azegr, 10 (1.68) The lattice energy is equal to about 0.9 of the Madelung electrostatic energy. For NaCl this is about 8.5 eV per molecule and is in good agreement with experiment. The Burn-Madelung model described above gives the lattice energy as composed of two terms, F and Lowdin (1962) have pointed out that this is not correct and that the total energy involves several terms of the same order as the Madelung energy. Briefly their argument is as follows. 15S The virial theorem states that in a conservative system in equilibrium, if the potential is a homogencous function of the co-ordinates of degree n then the average total potential energy V is 2 cm> or n= 77K 103 em = 1.7710 mm Example 7, Zinc has an hep structure where the height of the unit cell is 4.944, and the nearest neighbour distance is 2.7 4. If the atomic weight of zinc is 65.4, calculate the volume V of the unit cell and deasity p of zinc. Solution, Here @ =2.74,c = 4.94.4 ‘Therefore the volume of the unit cell is p= BvBa%e _ 3V3(2.7%107)?(4.94x107 9) 2 2 or V=94x 10°? mn? Now, the unit cell has 6 atoms. Hence 6x654x107> (94x10 ”)x(6x107) (9 = 6957 kg/m? Example 8. The lattice constant of Al is 4.05 A. (a) How many unit cells are there in an At-foil 0.1 mm thick and side 20 cm square? (b) If the foil weighs 10 g, how many atoms are present? Solution. (a) Volume of foil = 20x 20 x 0.01 =4c.. Volume of unit cel “. No. of unit cells 4 (4.05)>x10-% 66.43 _ 400 - 3 = Boag * 0? = 6.02 x 10" (b) No. of atoms N 6.02x103 = 9p 10 = SS x 10 atoms _ 02 x 10 = 224x103 atoms Example 9. Calculate the angle ¢ between [111] and [TT] directions in a cubic crystal. Solution. For a cubic crystal Ayhotkykothls (Otte) (b+ +B)? Aysky=h=1, hg=kp=-1h=1 es A(—1) + (-1) + 11) ~ corps DF +10) Oo Capo ai : “ poy? n= = 100.5° Example 10. The distance between adjacent atomic planes in CaCO3 is 3 A. Find the smallest angle Bragg scattering for 0.3 A serays, Solution. Bragg’s law is 2d sin 6 = nk @ = Gp is smallest when 2 sin Oy =A cos @ Example 11. Electrons are accelerated by 844 V and are reflected from a crystal. The reflection maximum occurs when the glancing angle is 58°. Find the interplanar spacing in the crystal. Solution, x-ray wavelength is given by a= UBa- vial pe - oH Am yn 042 = 4201071 m 2d sin = nA A 4 FsinB Now (a=1) _ 42x107! * Oxsin SF = 0.248 x 107 m = 0248.4Crystalline State and Bonding 1.63 Example 12, Consider a set of crystal 1 planes which are separated by 1.95.4. If we use tand=> x-rays of wavelength = 1.542.A, find all possible Bragg angles for reflection from these planes. Solution, 2d sin@ = nd g f n=1>0=6; and z A 1542 7 sin 4 = 9g = 2x1 = as: 1 n au Again, sin @) = 2 (a) = 0.7906 1-4 v5, 2dsin 8 = nd sin 63 can’t be greater than 1. yeaa Hence, possible Bragg angles are 23° and 52°, WS Example 13. Consider planes shown for sing => the two-dimensional square lattice in Fig. If a = 4.00 4 and we use x-rays of wavelength, _ 4 9 <6, and sin 0, = 4V5 = LSM2X2%4 1.542 A, at what Bragg angles would we » Z 1 la 2x4 observe reflections from these planes? sin 0 = 04318 6, = 26° n=2: sin 0 = 0.8636 = 60 n=3: sin 3 = 1.2954 As sin 83 can’t be greater than 1, third order reflection is not possible. Example 14. On an xray powder photograph of a cubic substance with copper a Kz radiation (A= 1.542 A), lines are observed woe at the following Bragg angles (@ in degrees): 123, 14.1, 20.2, 24.0, 25.1, 29.3, 32.2, 33.1. Determine the nature of the space lattice, and Rekgtion:s calculate the unit cell dimension. Solution, From Bragg’s law . B 7), the force is then attractive and tries to bring the pair back to the equilibrium position. If we push this pair together (r < ro), the force is repulsive and again tries to bring the pair back to the equilibrium position. The force: between the Na*- CT pair is thus a restoring force. A displacement of the ions from their equilibrium position causes the force to try to restore them back to equilibrium.24 Solid State Physics and Devices Fig 24 The net force beoveen a Na*—Cl_ pair New r= 15, this force is approximated by 6 straight line If we pull the ions apart and let them go, they will oscillate about the equilibrium position. Consider what would happen if the restoring force F were simply proportional to the displacement, r— 1, from the equilibrium position ry, i, = -a(r-n) (24) We recognize this equation as Hooke’: Law The two ions behave as though they were connected by a spring of length ro and “spring constant” @. The resulting motion is harmonic oscillation, r=rtAsinot 2.2) The amplitude A depends on how much energy the Na*~Cr pair has, and the frequency @ depends on the magnitude of the spring constant a, as well as the masses of the two ions. A restoring force such as that in eq. (2.1) would be plotted as a straight line on a graph such as Fig. 2.1. As we can sec, the actual restoring force for the Na*— CI pair is not a straight line and, therefore, does not obey Hooke's Law. However, near the equilibrium position (r = 1), the force can be approximated by a straight line, as shown by the dashed line in Fig. 21. (That line passes through F = 0 at r = 7% and has the same slope as F at r = ro) For small displacements from equilibrium, the force very nearly obeys Hooke’s Law. Thus, small oscillations should be harmonic, as given in eq. (2.2). Since we will consider only small oscillations, we can replace real forces between atoms by ‘springs," each with the appropriate spring constant &. As we saw above, a pair of atoms connected by a spring simply oscillates at some well defined frequency. In real crystals, there are forces in all three directions, and the motion of each atom is three-dimensional motion. However, great insight can be gained by studying a one-dimensional situation, and the analysis is much less complicated. 2.5 Simple Harmonic Oscillator Atoms on a mathematically perfect lattice should occupy exactly the sites to which they are attached. Of course, no crystal is perfectly rigid, since it can be deformed by finite forces; hence, it must be possible to displace ajoms from their ideal sites with a finite expenditure of energy. The forces holding atoms in position in the lattice are, in fact, weak enough so that the thermal energy of the atoms is sufficiently large to cause them to be displaced from equilibrium by sizable distances. The magnitude of this thermal vibration, some 5 to 10 per cent of the interatomic spacing at ordinary temperatures, is readily calculable (with suitable simplifying assumptions), Furthermore, several external manifestations of these vibrations, eg. heat capacity, diffusion, and electrical resistivity, can be treated with sufficient accuracy to make elementary discussions meaningful. A preliminary idea of the vibrations of atoms in a solid can be obtained by observing the nature of the interatomic forces. The equilibrium positions of the atoms are determined by a balance between the attractive and repulsive forces. At equilibrium the potential energy of the solid must be a minimum, The potential cnergy of an atom is sketched as a function of the lattice spacing ofLattice Vibrations and Specific Heats 25 lattice spacing —> binding energy per atom Fig 22 The potential energy of an aiom, Eg, in a solid containing N atoms, as & ferction of the tatice spacing, the solid in Fig, 2.2. For large lattice spacing (Ge, large total volume) the potential energy is atbitrarily set at zero, since the atoms then interact negligibly with cach other. As the lattice spacing decreases, the potential cnergy decreases. That this must occur is clear, since solids exist even in the absence of external restraints. Finally, at some critical spacing the potential energy passes through a minimum and increases rapidly with further decrease’in lattice spacing. That this increase must occur is also clear: solids have finite density. The important region is that near the minimum, where no net force is cxerted on an atom by its neighbours. If all the atoms of the solid except one are fixed in position and the atom in question moves in the x direction against its neighbours, as shown in Fig. 2.3, then unbalanced forces (represented by the extension and compression of springs) develop between the neighbouring atoms along the x axis. Since the distance to neighbours in the y and z directions is assumed to be unchanged, the unbalanced force in these directions is zero. Figure 2.2 represents the energy change when the distances to all the neighbours of an atom vary; if only one varies, the energy Eg(x) must be divided by z, where z is the number of nearest neighbours. (We assume that the potential energy for atoms farther away than @ & Fig. 23-(a) The auom in us equllibum postion. (0) The aiom displaced a distance u in the x direction The Springs in they and z drection do net change thei deg toa fis approximation. the ncarest neighbour can be neglected.) The change in encrgy, AE, which accompanies a change in atom position from x9 to x is AE = 2 {Eglty —Uu) + Egy + 0) ~ 2E(%9)} (2.3) where u=x—x9. The first term is the energy of the bond with the right neighbour and the second is the energy of the bond with the left. The factor 2 appears because the movement of an alom against its neighbour causes both the energy of the displaced atom and that of its neighbour (o change by an equal amount. Since the function £,(x) can be expanded in a Taylor series about ils minimum, the total energy change of the atom and its neighbours for small displacements is OE, | @&~myP= Fw (2a) % An atom bound to a definite site by the potential-energy law of eq. (24) is a simple harmonic oscillator. Hence, the vibrating atoms execute harmonic oscillations about their equilibrium positions. The force F on cach atom as a function of its displacement is given by26 pa 4). au @5) The differential equation describing the motion is then (2.6) where m is the mass of the atom. A solution is ‘a\ 2 4 = Aycos (5) t 7) ‘The angular frequency @ of the atom is thus given by o7()" and the frequency (called the Einstein frequency) by 1 (a\@ -3(n) Since ‘the masses of atoms are known from experimental observations, the frequency of vibration can be calculated if the magnitude of the force constant, a, is known. Fortunately « may be estimated simply. Suppose that Hooke’s Law can be applied to the unit cube of Fig. 24. The extension Aa which results when a force F is applied to the unit cube is given by the expression (28) Y: (29) a Ala In this expression ag? is the area of the face of the cube of edge ao. a is defined as that force fo stretch the cube unit distance. Since, Young’s modulus Y is of the order 10! N/m, ais approximately 25 Nim. The vibrational frequencies. can now be calculated. For atoms in the middle of the periodic table, Cx, Fe Ge, etc, the atomic masses are about 10> kg. The frequency of vibration, v, of these atoms is therefore about Solid State Physics and Devices Fig, 24 The unit cube eround the tom, 3 x 10” eps, according to this picture. More refined estimates of a place it’a litle higher than 25 N/m so that v is nearer to 10" cps. It can be shown that the amplitude of oscillation (at room temperature) is a few hundredths of am angstrom unit, ie., a few per cent of the interatomic spacing. The model of atomic vibrations just described is adequate for many purposes. The assumption that the neighbours to a given atom are fixed is, of course, not true for real crystals: neighbouring atoms move as well, When all possible motions of atoms and groups of atoms ~ arc considered, many other frequencies are found to’exist, frequencies ranging from nearly the value of v just calculated down to. that of the fundamental acoustical mode of oscillation of the picce of solid. The vibrations at the lower frequencies are, in fact, just what we know as sound waves. Each separate mechanical oscillation at a given frequency is called a phonon. The displacement of any atom in the crystal is found by summing over all the phonons, cach phonon having an cnergy (ie, amplitude of oscillation) as well as a frequency. 2.6 Normal Modes of a Lattice A normal mode is a correlated motion of the atoms that has a characteristic wave-vector and frequency. If this motion is started, it will continue for ever, ignoring friction, i.e. anharmonic terms. The atoms of a solid take part in thermal vibrations around their equilibrium positions.Lattice Vibrations and Specific Heats Because of a strong interaction between them, the nature of those vibrations turns out to be extremely complex and an accurate description of it presents enormous difficulties. Therefore approximate methods and various simplifica- tions are used to solve this problem. Instead of describing the individual vibrations of the particles the practice is to consider their collective motion in the crystal, which is a spatially ordered structure. This simplification is based on the fact that powerful bonds immediately transmit the vibrations of one particle to other particles and a collective motion in the form of an elastic wave involving all the particles of the crystal is excited in it. Such collective motion is called the aormat ‘mode of a lattice. The number of normal modes coincides with the sumber of degrees of freedom, which is 3N, if N is the number of particles constituting the crystal. Figure 25(a) represents, a one-dimensional model of a solid—a linear chain of atoms separated by a distance a and able to vibrate in the direction perpendicular to the chain. Such a chain may be regarded as a string. If the ends of the chain are fixed, the fundamental mode corresponding to the lowest frequency pin is represented by the standing wave with a node at each end (Fig. 2.5(b), curve 1). The sccond mode is represented by the standing wave with an additional node in the centre of the chain (curve 2). The third mode, or third harmonic, is represented by the standing wave with two additional nodes that divide the chain in three equal parts (curve 3), etc. The length of the shortest wave in such a chain is evidently equal to twice the distance between the atoms of the,chain (Fig. 25(c)) : Amin = 2a (2.10) The corresponding maximum frequency dgyax is Omar = Amin =AV/a— (2AM) where v is the velocity of wave propagation (of sound) along the chain, This maximum frequency is a parameter of the chain’s material and is determined by the 27 Fig. 25 Normal modes of a linear chain made up of identical atoms : (a}—tinear chain; (b)—normal modes of the chain; (¢)—nonnal modes of the ‘chain corresponding to shortest wavelength, interatomic distance and the velocity of wave propagation. Should we set a = 3.6 x 107? m (the lattice parameter of copper) and v = 3550 is (the velocity of sound in copper) we would obtain @max * 3 x 10% 5-1, which corresponds to the frequency of atomic vibrations in a solid. To describe wave processes one usually uses the wave vector k whose direction coincides with that of wave propagation and whose absolute value is k=2/d (2.12) It follows from eq, (2.11) that HA = wiv. Therefore = k=0/v,orw=kv (2.13) Mf there are N primitive cells in the crystal and z atoms per primitive cell, then there are N values of k and % branches, that is, for cach K-value there are % different frequencies for the 3 dispersion curves. For any one atom, a general (ic, arbitrary) displacement involves all 3zN different modes, Thus, a description of the motion in terms of individual atomic displacements is complicated. However, the problem is drastically simplified by using normal modes: these are the appropriate linear combinations of the atomic28 displacements that completely diagonalize the terms in ther Hamiltonian up to second order in the displacements. Once these are found, the total harmonic Hamiltonian of the system is the sum of individual (ie, uncoupled) simple harmonic Hamiltonians, one for each of the 3:N normal modes (all of which commute with fone another). The normal mode concept is very important. Classically, if only one mode is set in motion it continues to oscillate indefinitely while the others remain unexcited. Quantum mechanically the eigenstates of this 2N-body system can be written as a product of one-dimensional harmonic oscillator _ wave functions, one for cach normal mode amplitude. This normal mode decomposition is possible only because the 3zNV equations of motion for the atomic displacements are linear in all of the displacements. This is only an approximation, arising from the second-order term in a Taylor series expansion of the potential energy about the equilibrium internuclear distance. For most purposes it is an excellent starting approxi- mation, However, higher order terms are required for the understanding of many important processes. The apparent breakdown of the harmonic approximation leads to measurable effects, such as phonon-phonon interactions. = Since each normal mode yields a single simple harmonic Hamiltonian, the energy for onc normal mode as well as the sum of all of them can be written immediately. The energy of the Ath normal mode ¢, is quantized and given by Ke = pre (Q.14a) where mm = 0,1, while the total energy of the system is just the sum of the individual energies, Eyota) = 2 &- x One peculiarity of the harmonic oscillator sohition is the occurrence of the zero-point energy. 1 exo = Hox (2.14b) However, this constant amourit of energy for cach normal mode docs not affect most physical results. For example, since @o is Solid State Physics and Devices independent of temperature, the calculation of , is independent of the zero-point energy. This is typical of non-relativistic mechanics, where only energy differences are important 2.7 Vibrations of One-Dimensional Monatomic Chain (a) Modet : The basic assumptions are that 1. Each atom is located at an equilibrium position in the crystal structure, and 2. it can oscillate about this equilibrium position with an amplitude that is small compared to the internuclear distance. From these assumptions it follows that the potential energy of the system depends quadratically on the displacements of the atoms from their equilibrium positions for small displacements. Hence, the force in Newton's equation depends linearly on the displacements. Thus, the system behaves as if all the particles are joined by harmonic springs, and the motion is reduced to a classical problem. The solution, for small oscillations, may be described in terms of nonnal modes. Although vibrations can be treated classically, the problem is difficult’ even in simple crystals with high symmetry. 1, The linear chain is an infinite one- dimensional array of identical atoms. 2. Ima one-dimensional chain of similar atoms, there is one atom per primitive unit cell, 3. We assume further that the forces act only between nearest neighbours, However, if this assumption is not made the results are very similar. 4, The nearest neighbours are connected by harmonic forces through the springs. 5. In our treatment, it would be assumed that the motion is only along the direction of the chain (longitudinal waves). If motion is allowed perpendicular to the chain (trans- verse waves) the same results are obtained except with different force constants.Lattice Vibrations and Specific Heats 29 a Myet t ins? Fig 26 A lincar chain of ike atoms n (a) equitibraen; (b) displaced from equilibruan by amour up, () Dispersion : Figure 2.6(a) shows a linear chain of identical atoms, each with mass m, an equilibrium distance a apart, and connected to each nearest neighbour by a spring with spring constant @. When these atoms are displaced longitudinally along the chain from their equilibrium positions by an amount u,, as in Fig 2.6(b), the force on the n-th atom due to (1 +1)th atom is Funes = @ Wns — Ma) and duc to (n — 1)th atom is Fant = 4 tn ~ naa) Therefore, the act force on the n-th atom is Fn = Fast + Fap-1 orm =a ltnes ~ May ~ He Medd (2.15) In eq. (2.15), the two terms come from the two springs that are attached to the n-th atom, We expect wave-like behaviour and try travelling wave solutions of the form Un = uexpli(nka — wt) (2.16) This has wave-like behaviour along the chain, expli(nka)}, with amplitude u, and a time dependence exp(iot). Substituting this form into eq. (2.15) gives 0 wa 2nla wave vector k > Fig. 2.7 ¥s.k for longiuadinal waves of a linear chain ala mou = alexp (ika) + exp(-ika) — Zu 8 (217) Clearly this holds for any value of u provided that w= Ja — cos ka) m = (S2}s#?aaa (2.18a) a\ 2 or o=2(f) | sin (ka/2)}—(2.18b) This is called the dispersion relation because it relates w to k and Fig, 2.7 shows the result. The positive root of eq. (2.18a) is taken because physically we want real frequencies. There are many important points that should be noticed about the w vs k result in eq. (2.186) or Fig. 2.7 : 1. There is symmetry between k and —K. This is called the reflection symmetry This corresponds to the fact that waves travelling to the right or left are identical, ie. @ (k) =@(—k) 2. The results repeat with a period of 2x/a This is called the periodic symmetry : eo(k + 2nfa) = ok) 3. There is a maximum or cutoff frequency, @pay above which the chain of atoms cannot sustain a travelling wave, Wnge OCCUTS atk = (ald) + (p-2n/a), where p is an integer and its value is Wn mae = 2 (5) (2.186)2.10 4, The index n does not appear in formula for @, s0 a single relation gives o vs k. 5. The linear relation between @ and k does not hold. This is equilvalent to saying that the crystalline medium is dispersive, and hence the w ~ k graph is called dispersion relation. Nonlinearity of the graph shows that the Debye model does not hold here. 6, However, expanding eq. (2.18b) for small ka, we have wv on (3) ak m a » i (2 2 = velocity = (a) a (2.19) Thus, in the long wavelength limit, o is linear with k. This can be scen in Fig. 2,7, and the velocity is given by the spring constant, the mass, and the lattice constant. This makes physical sense since for small & we cxpect "not (0 notice’ the individual atoms, which corresponds to the Debye approximation (je, treating the system as a continuous medium). However, as k becomes larger (A shorter), the discreteness of the chain affects the dispersion relation, () Group velocity : The group velocity (ie, the velocity of a wave packet) is vy dw/dk. For this monatomic linear chain, between k = 0 and 2/a, it is given by a\2 (ri) acos (ka/2) (2.20) This result is plotted in Fig. 28 For small k, vg ith the value given by eq. (2.19) and for k = n/a it is zero. More generally vy is 2er0 when- ever k= 1la+p. 2n/a where p is any integer. This k-periodic dispersion curve is the result of the translational symmetry of the crystal, The vibrational modes discussed here are called acoustic modes and the w vs. k is called the acoustic branch because at small k there is @ constant nonzsro group velocity as in a continuous medium. Solid State Physics and Devices t Wa ala ‘e i Fig 28 The group velocity for a monstomie linear chain (@) Brillouin zones : If we put k= kp + 2ap/a (where p = 1, 2, ...) in eq. (216), then it follows from eq. (2.18b) that the frequencies corresponding to the modes k and ky are identical. In other words, the state of the vibration of the array of mass points corresponding to a wave-vector k, is the same as that for any other k + 2np/a. In order to obtain a unique relationship between the state of vibration of lattice, and the wave-vector k, the lattice must be confined to a range of values 2x/a. One selects the range such that Re,ek -Fsest (221) The positive * values correspond to wave propagating in one direction, while the negative valtes represent waves going in opposite direction. The region of k values defined by eq, (221) is called the first Brillouin zone (BZ). The second BZ, consists of wo regions. -Hers-% @ a Pd @ ‘one on each side of the first zone, The two regions Bens-2 @ @ ens a @ and a Zeke a and form the third BZ. Then the definition of higher order zones follows immediately. However, because of the periodic nature of w — k relation, it is ‘sufficient to confine to the first BZ to completely describe the wave propagation. The first BZ contains the entire spectrum.Lattice Vibrations and Specific Heats It may be noted that the dispersion curve flattens gradually as the BZ-cdge is approached. We. may think of the flattening as being due to a component of a reflected (reversed) wave, which gradually grows until at k=+2/«, the amplitude is equal to the incident (forward) component. A wave corresponding to the BZ-boundary is a standing wave rather than a travelling wave. It has a wavelength appropriate for Bragg reflection through an angle of 180°. There are N atoms and one degree of freedom for the atomic motion (longitudinal motion along the chain), so there are N k-values in this zone. If we allow transverse, as well as longitudinal vibrations then there are 3N degrees of freedom. Since we assume that the chain is rotational symmetric the two transversely polarized vibrations are degenerate; their @ vs k curves are identical to each other and they have a shape that is similar to the longitudinal case. However, the transverse force constants are generally smaller so their w values are less than corresponding longitudinal values. This same very important result for the Brillouin zones applies to lattice vibrations in three dimensions. That is, in three dimensions, a crystal with one atom at each of the N lattice points has 3N degrees of freedom, and the 3N Aevalues all can be confined to the first Brillouin zone with onc longitudinal and two transverse branches. 2.8 Vibrations of One-dimensional Diatomic Chain (a) Model : The physical situation is described by (i) The lattice consists of two types of atom having masses M and m. This corresponds to a lattice with a basis of ‘two atoms in the primitive cell. This is a one-dimensional analogue of three dimensional NaCl-lattice, Let M > m. (i) The atoms of mass M and m arc alternately placed. 2u (iii) The particles are numbered in such a way that the even numbers correspond to mass m and odd ones to M. Only the ncarest neighbour restoring forces are important. ‘The displacements of atoms are confined to the clastic range of the Hooke’s law. Here we are considering a chain of 2N atoms with N unit cells but with 2 atoms in cach cell so there-are 2N atoms. Each unit cell is of length d = 2a, where @ is the distance between the atoms with dissimilar masses. For two atoms per unit cell, we will find two dispersion curves. One is just the acoustic branch obtained for monatomic chain; the other, called the optic branch, has a finite frequency at k = 0, as we will see. (b) Dispersion : Imagine a one-dimen- sional crystal as depicted in Fig. 29. This differs from Fig. 2.6 only in that atoms of masses M and m are alternately placed. We shall suppose that m < M, Now if a longitudinal disturbance is encouraged to travel along the crystal, the displacements of the two kinds of atom will usually have different amplitudes Uy, = A exp {il2kna ~ wil} Money = B exp {i[(2a +1)ka — wi}} i we suppose that only the nearcst-ncighbour restoring forces are. important and that displacements are in the elastic range of Hooke’s Law, then the force equations for the displacements 13, and ty, 41 can be written the same fashion as with a monatomic chain @) W) (i) (i) Co) (2.22) = maruy = m(Puy,/ dt) = Blanes + Man — Zon] = Mors) = M (Pum 4)/d?) = altang2 + Han Wonea] (223) Substitution of the two portions of eq. (2:22) into eq. (2.23) produces the two si equations wultancous242 Qn2) rt) 2m Gath) td Vig. 29 A. sequence of neighbouring atoms in a linear diatomic chain. The smaller atoms are assumed ‘0 have a mass m which is less than the mass M of the larger atcms. Note thet the repeat distance is now 2a. This will affect the size of the Brillouin zone. —m oA = able + #4] - 204 —Mo = ale! + #4) — 208 which we can rearrange as A(2a ~ mw?) = 2aB cos (ka) (225) B(2a — Ma*) = 2ad cos (ka)| ‘These can be solved together to eliminate A and B, expressing a dispersion relationship between k and @ : (2a — mw\(2a — Me) = 407cos?(ka) (2.26) so that | (2.24) aaa 2 va Bae 4 sin’ (ka) so[(t+a) -*004 Whereas we previously found that for a monatomic linear chain there was one solution for k > 0 (a wave travelling to the right) and one for k < 0 (a wave travelling to the left), we now have two values of o/for a given k even in the positive quadrant, The spectrum: of the result for w as a double-valued function of k is shown in Fig. 2.10. (c) Acoustic branch : The lower branch ‘of the spectrum seen in Fig. 2.10 is described by the negative sign choice in eq. (2.27). This branch is usually called the acoustic branch and corresponds fairly well with what we have already seen in monatomic lattices, with two exceptions : (221 Solid State Physics and Devices (-n2a) 0 wave vector k ——» (w/2a) Fig 210 Dispeion relationship for the propagation of hela wore tne man ces The Brillouin zone comprises the range of k-space fa wich \k|-s (12a). The lower branch (which be usually called the optical branch of the vibrational specnun. 1, Any set of displacements can now be described in terms of a wave-vector with an absolute value no larger than (7/22), compared with a Brillouin zone boundary at +t (sc/a) for a linear monatomic chain. This reminds us that the range covered by a Brillouin zone is determined by the periodic repeat distance 2a, not by the nearest-neighbour distance. 2. The maximum possible angular frequency for acoustic mode vibrations is from eq. (2.27), and this appears to be independent of the lighter atoms in the chain. The reason for the second of these observations can be understood if we look into the amplitudes of the two types of atoms as a function of frequency. Equations. (2.25) can be arranged as (228)Lattice Vibrations and Specific Heats mplitude of heavy mass M] _ [B Amplitude of light massm:} ~ [A _ |2a-me* | Qacos (ka) (2.29) [2atcos (kay 2a - Mor Thus the amplitude ratio is approximately unity (all atoms moving in the same way) for the long wavelength, low frequency, acoustic waves which satisfy the following set of conditions : wave vector [k] s (é) wn aa? speed vy = E = | (2.30) w 2a angular frequency w = kag << [3% Looking at the term on the far right of eq. (229), it can be seen that as k and « both increase in the acoustic branch, so docs the ratio (B/A). In the limiting case, for the mode with x katy wave vector gular frequency » = w v2 (2.31) phase velocity 2 = (a kM ay (42 group velocity (@) =0 then (B/A) > + © as w->a,. This means that A must be zero regardless of the amplitude B. Since the lighter atoms are not in motion for the limiting mode at frequency w,, it is no longer surprising that the magnitude of m should not enter into equation (2.28) for w). Note that above the angular frequency thete is a frequency range which has no 213 solutions for real k, just as we found for a monatomic chain for aif frequencies above @pax- ThE range of frequencies for which & must be complex forms a forbidden band, since any wave initiated at a frequency in this interval will be heavily damped. The interesting additional feature we find from eq. (2.27) and (0 is that with a diatomic chain a second ied range of frequencies exists, the range of optical mode vibrations. (d) Optical branch : The optical branch of laitice vibrations (the set of optical modes) is so named because modes of this type can be excited with light of an appropriate frequency in solids which are at least partly ionic. This can happen because the amplitude ratio (B/A) is negative throughout the range of frequencies encompassed by the optical branch; nearest neighbours move out of phase with cach other. Such motion can be excited by the transverse electric vector of a suitable em. wave when nearest neighbours carry electric charges of opposite sign, For a linear diatomic chain the optical branch is described by the positive sign choice in eq. (2.27). Long wavelength optical modes (which occupy the region of k-space close to the origin) appear to satisfy the curious set of conditions : wave vector: k +0 angular frequency : v-o=[o(eeal) phase velocity : (@/k) > group velocity : (dax/dk) -» 0 (232) By manipulation of eq. (2.29), we can find that the amplitude ratio of the two kinds of atoms is k=0 a> = (233) ~|214 Solid State Physics and Devices (e) Relative Amplitudes : Figure 2.11 shows the variation of B/A as a function of k. It could be observed that B/A Fig. 211 A/B as @ function of k for acoustic and optic branches. @® In the acoustic branch, in general, the displacement of two types of particles are in the same direction, as B/A is always positive. (i) At k =0, the amplitudes of the light and heavy particles in the acoustic branch are equal, as at this value B/A = 1. The nomenclature of the acoustic branch originates from the fact that this branch can be stimulated by some kind of force that makes all the atoms go in the same direction, such as a compressed wave or sound wave. At the zone boundary (k st/2a), in the acoustic branch, B/A = ©, ie. A=0.In other words, the light particles of mass ‘m (with amplitude A) are at rest. (¥) The optic branch is so named, because modes of this type can be excited with light of an appropriate frequency in solids which arc at least partly ionic. This happens because 4/B is negative through- out the range of frequencies lying in the Gil) (iv) optic branch; the nearest neighbours are in antiphase with each other. If the particles are ions with opposite charges, then the type of motion represented by the optical branch may be excited by an electromagnetic wave. This maximum frequency represents the fundamental infrared absorption frequency, (node tei —-ted- foie bb bb (Of @1e 40 Te 1e-Te- le 1 tot 1 1 1 1 1 toa 1 1 CHOTA fie t e+ ' rt rout tf tra (OO 16 @ 4-4 Te GF - ' t . er Fig. 212 4 schematic diagram showing the atomic displacements for (a) k = 0 optic branch; (b) ‘acoustic branch for small but certainly not zero kj (c) and (4) show the zone toundary modes for the diatomic chain at k = x/a, that is, the zone ‘boundary for diatomic chain. Atk =+ 7/2a, for the optic branch B/A = 0, or B = 0, which shows that the heavy particles of mass M are at rest. (wi) (ii) Figure 2.10 shows that the optic branch can have a k = 0 mode, but the acoustic ‘branch cannot. (f) Longitudinal Modes : Refer to Fig, 2.12, () For the acoustic branch A = B, so the light and heavy atoms in each unit cell are moving in phase as in a sound wave. Of course, if all unit cells moved simultaneously to the right then the crystal would move to the right; this happens at k identically zero. (i) However, in the limit of small (but not zero) k, there are large regions where the light and heavy atoms are moving to the rightLattice Vibrations and Specific Heats 215 and then other regions where they are both moving to the left. Figure 2.12(b) is a poor attempt to show this. For the optic branch at k~0 the phase of the light and heavy atoms are opposite, m4 = —MB. Any two neigh- bouring atoms will oscillate around their equilibrium positions, toward and away from each other (Fig. 2.12a). Thus, if the m and M atoms have opposite charges, as in an ionic crystal, this motion will set up a long wavelength (small k) oscillating dipole moment, which can interact strongly with light of the same frequency. (For atoms in crystals, this is an infrared frequency.) Hence, the name optic branch and the motion is called the optic mode. ‘Thus for these long wavelength vibrations, neighbouring atoms are moving in opposite directions in such a way that the common centre of mass for a neighbour pair docs not move. (g) Transverse modes : Up to this point we have considered only the Jongitudinal vibrations of a diatomic chain, but the essential difference between the requirements of acoustic and optical modes for displacements of neighbouring atoms can be seen more clearly by plotting transverse displacements. This has been done in Fig. 2.13 for 2 wave of intermediate wave-vector. Whether a vibrational mode is longitu- dinal or transverse, a given atomic amplitude of motion requires much more energy for a long wave optical mode than for an acoustic mode of the same wavelength. For it will be appreciated from Fig. 2.13 or its longitudinal equivalent that optical modes minimise changes in second-nearcst-neighbour distances by maximising changes in the ncarcst-ncighbour separations. Since the _ nearest-neighbour stiffness constant is usually considerably larger than any others, cach harmonic oscillator must store a considerable amount of energy. The amplitude ratio (B/A) remains negative throughout the optical branch. From eq. (2.29) we can determine that as k increases Hig. 213 The sequence of omic displacements involved in the wansmission of a transverse wave along a diatomic chain. The difference between the fcoutic and optical modes can be sen more clearly when transverse displacements are skeiched rather than a lon wave and as @ in consequence decreases to its short wave limit of @, then (8/A) approaches zero from the negative sido, Thus for the shortest possible wavelength {= 44a) in the optical branch, the heavy atoms do not move at all and the wave involves the light atoms alone. For the frequency limit. -- the group velocity (dax/dk) again becomes zero. Note that waves of frequency #, or w are standing waves of zero group velocity reflected through 180° by Bragg diffraction. The mass ratio (m/M) determines both the width of the forbidden frequency range and the width of the optical branch. When m and M are not too dissimilar, the forbidden range is a narrow one, and the optical branch covers, a frequency range of up to 1.4 :1. However, if M is considerably larger than m, the gap is a wide ‘one, and all of the optical branch vibrations are cramped into a narrow frequency range at Gy = (2a/mn)”?,2.16 ‘Solid State Physics and Devices (h) Frequency gap : There is a frequency gap in the dispersion diagram (Fig. 2.10) in which k is complex and the waves arc damped, It is interesting to know the frequency difference of the two branches for a given value of k, Atk = 0, this difference is maximum and = @, which reduces as k is increased, until at k=+.2/2a, i becomes minimum, equal 10 2}. This difference in frequencies at the zone boundary is called the frequency gap. The gap is obviously a function of the mass ratio M/m. A frequency lying in the gap @, — w; is not a normal mode of the crystal. It is true that an atom, in principle, can be shaken to any frequency, but the rate of shaking will propagate down the linear chain only if it is a normal mode. Otherwise, the oscillation of successive atoms will cancel out due to phase changes. (i) Effect of mass ratio : Keeping m constant, the effect of change in Mim on frequency gap can be studied. By increasing m/M the width, «3 — w» of the optical band can be made narrower, which results in the decrease of the frequency gap. Thus, as we increase m/M (towards unity), we first change the numerical value of the curvature of the optical branch, and then the sign of curvature which further changes with the change in m/M, and finally when m/M = 1, C and C' respectively coincide with D and D', while A comes over at 0 (Fig. 2.10). Thus, in this limiting case, the optical branch completely covers the acoustic branch. In other words, switching over from a diatomic care to a monatomic one, the two branches degenerate into a single one and the distinction between them disappears. If the value of the heavy mass is increased to the limit M >, the acoustic branch disappears and the optical branch flattens in sych a way that there is a single frequency corresponding to- all K-values (Fig. 2.144). Physically this means that each atom oscillates at its own natural frequency uncoupled to its neighbours. § Branch k wa ao = Fig. 2.14(a) M,-> & Acoustic branch of a diatomic lattice disappears leaving @ set of independent oscillators of constant frequency. The optical branch disappears upward in the limit m > 0 (Fig. 2.14b). In this situation, the unchanged acoustic branch refers to the monatomic lattice. The correspondence can be made exact by letting 22a, as the lattice constant is now doubled. acoustic branch wa 0 mat Fig, 2.14(b) m+ 0. Optical ranch disappears upward, leaving (he dispersion relaion of @ monatomic lace @ Brillouin zones : Equation (227) reveals that @ is a periodic function of k, Thus, the complete description of the dispersion curve can be achieved between the range of k values %eyeh 2a cs or The region corresponding to this interval is known as the first BZ for a one-dimensional diatomic lattice. The second BZ corresponds to two intervals : k ele a a oe 2a and sk Bla n ele Higher order zones could be accordingly defined.Lattice Vibrations and Specific Heats 2.9 Density of States Although any vibrational pattern is possible for an clastic continuum, there is limit to the number of distinguishable modes. of vibration for a lattice of finite size consisting of discrete atoms. For a linear chain of N + 1 atoms and length, L = Na, if the end atoms are clamped, we have x, =xy,1 = 0. Then either longitudinal or transverse vibrations can exist for which there arc 1, 2,3, .. or N half-wavelengths in the distance ‘Na Corresponding to these allowed vibrational modes, there are N’ linearly spaced values of wave-vector : poh, 2, x Na’ Na’ Na’ @ We refer g(k) as the density of states with respect to k and define it as the number of states per unit length per unit interval of k : maa: A 8) = Waj@r7ay ~ = Then the number of states in the range koe + dk is given by 1 8().dk = Fd + 234) ‘The number states per unit length of one dimensional crystal is sdk=1ak, ksx/e =0 » kena In using the density of states for the calculation of physical characteristics like energy or specific heat by summing over all normal mode oscillators, we require the density of states per unit frequency range g(w); if dn is the number of states covering a range of dk centred at k and a range of da of w, we obviously have dn = g(k) dk = g(w) do ‘i ak 2 gw) = 9) 1 gtk) 8) = Ga7an = xdarda (235) 217 which, on using eq. (2.20), gives 1 1 (m\?_ 4 Be) =F le cas(ka/2) (2.36) On substituting the value of cos (ka/2), obtained from ¢q. (2.18b), in eq, (236), we obtain “12 doe au) = 2 (# - «| 237) Equations (2.37) and (2.18c) may be combined to give (may > Ym) 2 Oa ha Thus, g(w) does have an explicit dependence on «, and, in fact, becomes infinite at the upper limit. Iw +0, 1 Ww vo As the zone boundary is approached, w -* w, and (2) > © ‘The density of states plotted in Fig. 2.15 incorporates this behaviour. If we ignore the dispersion of sound waves, we can write @ = vk or dw/dk =v =a Va7m, and from eq, (235), we get eq. (239). This corresponds to the constant density of states shown by the broken fine in the figure, for which the total number of states does not reach the value of N until the frequency @ =x Va7m. (2.38) (2.39) g(a) Wolm nfl Fig 215 Density of states for a chain of atoms .2 2.10 Interactions Beyond Nearest Neighbours ‘The exclusion of interactions beyond the nearest neighbours is only a convenient approximation and a more accurate description is obtained if interactions over larger distances are considered. Once more, @ is periodic in k with period 2a/a (Fig. 2.16) and inside the Brillouin zone, o” is an even function and there is a two-fold degeneracy (except at the origin). The maximum frequency docs not necessarily ‘occur at the points + zt/a although the dispersion curve must have a stationary value at these points. ala 0 ae Fig. 216 Dispersion curve with many neighbour interactions. 2.11 Transverse Waves If the atoms can vibrate in a plane perpendicular to their line, then transverse modes are possible. For small displacements, it can readily be shown that the equations of motion are very similar to those for longitudinal motion [eq. (2.15)]}. There is of course a new force constant. All the analysis of the longitudinal motion will apply and for a linear lattice with cylindrical symmetry, there will be two dispersion curves (Fig. 2.17). 2.12 Three Dimensional Lattice The concept of the division of the vibrational modes into acoustic and optical branches encountered in a diatomic linear Solid State Physics and Devices @ we Fig. 217 Tranmere and longiudinal waves in a ‘one-dimensional monatomic lance with ey and 0 the corresponding cutoff frequencies. lattice can be generalised to be applicable to three-dimensional structures. However, then the number of transverse modes is double that of the longitudinal modes, Thus, for sN atoms in a solid ‘with a polyatomic basis (where s = number of atoms in the primitive basis and N= number of basis groups in the entire crystals), the number of modes in various classes will be as follows: longitudinal acoustic modes = N transverse acoustic modes = 2N longitudinal optical modes = (5 — 1) transverse optical modes = 2(5 - 1)N total modes = N+2N+(s — 1)N + 26 ~ )N = 3sN This shows that the numbers of longitudinal and transverse acoustic modes are always 1 and 2 per atom, irrespective of the basis structure. Also, the optical modes are more aumerous than acoustic modes. For a crystal with s = 2, there will be one longitudinal optical mode and two transverse modes, as shown in Fig. 2.18. However, for a general wave vector, there is no degeneracy and there is no definite distinction between longitudinal and transverse modes. For each branch of the spectrum ok) = 0(— &) If there is a basis of s atoms per ccll, there will be 3s modes for each k out of which there will be 3 acoustic and 3s—3 optical modes.Lattice Vibrations and Specific Heats Tig, 218 Diperion ration foro dee dimensional ena with two atoms per unit cel In_ three-dimensional waves, there are three different polarisations possible for a given direction of propagation k. For example, in Cu, all the lattice waves are acoustic (the lattice is monatomic), and thus in Fig, 2.19, we label the curves LA for longitudinal acoustic and TA for transverse acoustic. In the [100] and [111] directions, both TA branches in Cu have the same dispersion curve and are thus drawn on top of cach other as onc curve. However in {110} direction, we have one LA and two TA. VEE Fig, 219 Dispersion curves along three symmetry directions incu Consider next, a crystal of KBr. The direct lattice (Bravais) is fec, and the reciprocal lattice is bec, as in Cu. The dispersion curves are shown in Fig. 2.20 for three symmetry directions in the crystal. The crystal is diatomic, and thus we observe optical branches, labelled LO and 70 for the longitudinal and transverse polarisations, respectively. Acoustic branches go to zero at k = 0. Optical branches do not. As in the one-dimensional diatomic lattice, we observe a frequency gap between the acoustic and optical branches. 219 Fig. 220 Dispersion cures along three symmeny directions in KBe 2.13 Phonons Our analysis of the atomic vibrations can be looked at in a slightly different way. We can consider the crystal lattice as one large system which is able to pulsate at various frequencies like a lump of jelly. Previously we have suggested that these vibrations were generated by hypothetical oscillators, but of course it is, the system as a whole that is generating its own oscillations —there is really no need to invent an external driving mechanism, The superposition of all the calculated displacements at the various points within the system must be the same as the atomic displacements at those points, So while carlier in this chapter we discussed the vibration of isolated atoms, we now see that because they are coupled together, a more logical point of view is one in which we are thinking in terms of the various modes of vibration of a single vibrating system—the whole assembly. But even this composite system must still obey the rules of quantum mechanics. Its energy can only increase in discrete steps of ‘hy, only now these quanta of vibrational energy involve displacements of all the atoms rather than just one of them. By analogy with photons (the quanta of electromagnetic radiation) these quanta of lattice vibrational energy iv are called phonons. From cq. (2.43), we sce that the cncrgics of the normal modes are quantized and equally spaced. Thus, the nyth excited state of the kth normal mode can be viewed as a state of ny2.20 Solid State Physics and Devices identical excitations in this mode, each having energy /v,. These excitations or quanta we call phonons in strict analogy with the qnanta in an electromagnetic field which we call photons, A quantum state of a crystal with the set of quantum numbers {nj} is said to have "ny phonons in the kth mode," each phonon contributing an cnergy a4. Thus, a phonon is a quantum of excitation in a normal mode. When exciting the Ath normal mode from ny to my+1 phonons, a phonon is "created" (or emitted) and the energy of the vibrational system increases by Hay. When de-exciting a normal mode from nz to my — 1, a phonon is “annihilated” (or absorbed) and the energy of the vibrational system decreases by Tay, The notion of phonons as particle-like is streng- thened because they possess another particle- like property, crystal momentum 71k, in analogy with the ficld momentum #k of a photon. Thus, phonons are often called quasipantctes. Shortly, we will discuss the specific heat of solid materials in terms of quantum harmonic oscillators, and calculate the internal energy by ascertaining the number of oscillators, on the average, in each allowed energy state. If the vibrations of the lattice can be treated in this way, then there must exist quanta of vibrational energy which are emitted or absorbed when transitions between one quantum state and another occur. It is a result of the quantum mechanics of the harmonic oscillator that to a first order of approximation, transitions occur only between adjacent states ‘of the system; in other words, in any such transition the quantum number of the oscillator may change only by + 1. Accordingly the energy change AE of the oscillator, corres- ponding to the energy of the quantum’ which is absorbed or emitted must be AE = Anhy = + hy (2.40) where v is the frequency of the oscillator. The plus sign in (240) refers to absorption, the minus sign to emission. Such a quanum of acoustical energy is commonly referred to as a phonon. If one type of oscillator is cxcited into its mth state, we can say that n phonons of this type are present. If the oscillator passes from nth to its (n + 1)th state in a transition, we say that a phonon has been created in the process; in passing from the nth to (n - 1)th state, a phonon is said to be destroyed. 1. Thermal energy given to a solid is absorbed by springs connecting atoms as elastic energy or elastic waves. Lattice vibrations or clastic waves quantise energy. A phonon is just the quantum of energy in an clastic wave. In fact, a phonon represents the particle concept associated with the quantisation of lattice waves. The situation is quite analogous to the electromagnetic waves where the quanta arc called photons, 2. The Debye result for the specific heat can indeed be obtained by regarding the lattice vibrations of the crystal as a gas of phonons. 3. Since the phonons are indistinguishable particles, the Bose-Einstein distribution function must be uscd to describe the distribution of particles among the energy states of the system. in other words, phonons are bosons, and hence any number of these could populate a quantum state of the system. It is interesting to note that phonons, which are associated with the catire lattice, are obviously ‘particles’ which are indistinguishable. They should, therefore, be controlled by the Bose-Einstein statistics, since they have zero spin. The probability that a state of energy E is occupied by a phonon is then given by 4% E fraté)= {em (2) -a} ‘The mean encrgy of the state associated with a frequency v and an cnergy hv is, therefore, the energy X the probability of occupation, which is hy exp (hv/kT) — This is exactly the same as the temperaturc- dependent part of the expressionLattice Vibrations and Specific Heats hy jap) = i which will be derived [eq. (2.43)] as the mean energy of a localised atomic oscillator. There is therefore no inconsistency in treating all the vibrations of the individual atoms as if they are a superposition of vibrational quanta (phonons) which are associated with the entire assembly of atoms. 1 git 4. A phonon with a definite wave-vector k is an excitation of a normal mode of the entire crystal involving the atoms in the entire crystal. 5. Two types of phonons arc distinguished from each other by the following definitions on the basis of dispersion curves ie o Hii minimum for acoustic =0 lonons, im a maximum for optical k>0 phonons 6. Both optical and acoustic phonons have continuous cnergy spectrum, the former has higher energies than the latter with a forbidden energy gap between the two. 7, The number of phonons in a crystal is temperature dependent : (a) At finite temperatures, the crystals are full of phonons (b) At low temperatures, acoustic modes are more populated than the optical modes. (©) If the temperature of the crystal is raised, the amplitude of the atomic vibrations increases, and in quantum terms this can be thought of as being due to an increase in the number of phonons in the system. (d) On the other hand, the drop in temperature is responsible for the destruction of phonons of the system. 8. When a lattice vibration is involved in an interaction, not only is its energy always an integral multiple of Av, but also at a time, only a single phonon is absorbed or 221 excited. This is confirmed from the fact that in the neutron diffraction, neutron energy changes by No. In neutron scattering also, only one phonon is absorbed or emitted. (a) Phonon-neutron interaction : A neutron experiences the existence of a crystal lattice with the nuclei of the atoms of the lattice. It is well-known that the elastic scattering (in the Bragg’s diffraction) of any particle (for example, neutron) by crystals requires the conservation of energy alongwith the condition that the neutron of initial wave-vector kj is scattering to an equivalent final wave-vector ky. If the scattering of neutron by phonon is elastic in nature, then a phonon of wave-vector q and frequency @, is created or annihilated such that where hw» is the energy of the phonon involved in the process and M, is the mass of the neutron. (b) Phonon-phonon interaction : In the simplest model of a solid, neighbouring atoms may be interacting with each other by a spring connecting them, When the solid is heated, the heat energy appears as elastic energy of the spring. If the displacement x of the spring is small, its potential energy U is expressed by the harmonic model : Usar and we say that the two oscillating atoms interact weakly. Thus, at very low temperatures the interaction of phonons with one another is negligible. At higher temperatures, when the atomic displacements are larger, the anharmonic terms become appreciable and the two waves (representing two oscillating atoms) can interact and some portions of their energy2.22 can be converted into a wave with different wave-vector and frequency. The smallest quantum which can be ‘removed’ from each wave is to form the third is a phonon, and this process is described by saying that two phonons (one from each of the two individual waves) have combined to form the resultant third phonon. The major portions of the two waves, however, actually passed on without interaction and these may be ignored in any discussion on scattering processes. The conservation combinations of phonons are: (1) energy is conserved, ie. laws for the No, + Nw =v; (2) wave vector is conserved, i.e. ky thg= hy At higher temperatures, strong phonon-phonon interaction leads to the anharmonic term in the potential energy : Usaxr—pr This is called the anharmonic model, which only explains the physical origin of thermal expansion. Thus, the existence of phonon-phonon interaction is responsible for the thermal oxpansion of solids (specially, nonmetals). The interaction between two phonons may be treated as a scattering ‘collision between two particles. (c) Electron-phonon interaction : A free electron in a perfect crystal structure propagates freely without attenuation. However, the perfect periodicity is destroyed by thermal vibrations of the atom. This partial deviation from periodicity causes the electron wave to have a certain probability of being reflected (or scattered). Since the electron is interacting with the vibrating atoms this is called the electron-phonon (or electron-latice) interaction. In very pure metals, it is the interaction that determines the resistivity. Solid State Physics and Devices In fact, it is quite possible to treat the interaction between a phonon and an electron as a scattering collision" between two particles. (d) Crystal momentum : Fundamental to our discussion of phonons has been the fact that the potential energy of the system depends quadratically on the atomic displacements. This leads to the harmonic Hamiltonian, and phonons. However, it must be remembered that when the potential energy of the crystal is expanded in a Taylor series of displacement u, about the atomic equilibrium positions, higher order cubic, quadratic, terms (ie, anharmonic terms) exist, and give rise to important processes in real crystals such as thermal expansion, finite lifetime of phonons (and hence finite mcan free paths), sound absorption, and others. We shall not consider these anharmonic processes in any detail, but will consider what a general anharmonic interaction term must look like and in this way be led to the very important concept of crystal momentwn and conservation of crystal__momentum. This conservation law, along with the conservation of energy, plays an important role — in phonon-phonon interactions, electron-phonon interactions, and even the interaction of phonons with neutrons and photons (both x-rays, and light). (¢) Thermal conductivity : If we sct up a temperature gradient, then clearly the amplitude of the atomic vibrations at the hot end of the specimen will be greater than at the cold end. Phonons will be more dense at the hotter end of the crystal than at the colder end and we can think of the process of heat conduction as being due to the diffusion of phonons down a concentration gradient to the cold end of the specimen. In order to develop a theory of thermal conductivity we need to know the rate at which phonons can transfer heat along a crystal. ‘These phonons are continually scattered by various barriers or interactions like: @) | phonon-phonon interaction; (ii) scattering by impurities;Lattice Vibrations and Specific Heats scattering by dislocations, scattering by the boundaries of the specimen or crystallites. We may assume that the phonon travels some average distance /, the mean free path, between these collisions, It turns out that we can adapt the classical kinetic theory of the heat conductivity of gases to the problem of heal transport by the phonon quasi-particles and we may use the kinetic formula for the thermal conductivity where c, is the specific heat per unit volume of the phonons and v is their velocity. Since phonons are elastic waves within the crystal, v is effectively the velocity of sound. Here ¢, is a lattice specific heat which is a measure of the phonon density in the crystal 2.14 Lattice Specific Heats So far we have been concerned with the particle-like aspects of lattice vibrations, because we were primarily concemed with dispersion laws for which it was usually convenient to adopt a wavelike nomenclature. Armed with this information about dispersion laws and densities of modes with respect to k or w, We now find it appropriate to invoke the quantized particle idea of the phonon once again, in order to sce how the vibrational energy of a crystal depends on temperature. If U denotes the entire vibrational energy of the crystal (per Kilogram, per cubic mete, or per mole), then we arc. interested in C,=(2U/9T),, the specific heat at constant volume. For purposes of actual measurement, the specific heat at constant pressure, C,, is much more convenient, but (C,—C,) is fortunately very small for a solid because the work performed in thermal expansion is so small. Specific heat is really a measure of the number of degrees of freedom. Since degrees of freedom imply freedom to absorb potential or 2.23 kinetic energy, the problem is to know how many modes are there for the system to absorb the energy. The system of our concern is the oscillating lattice, but its analysis is quite difficult, particularly in three dimensions. Instead, we would carry out the development of the theory through models, cach of which bears some resemblance to reality. In most solids, the energy given to lattice vibrations is the dominant contribution to the specific heat. Thus, a knowledge of the frequency spectrum of lattice vibrations of the solid is an essential ingredient of any theory built for the explanation of the lattice heat. For the purpose of illustration, one considers a one-dimensional lattice for which the coupling together of atomic vibrations leads to a band of normal mode frequencies ranging from zero up to some maximum value. Calculation of the lattice energy and specific heat of a solid therefore falls into two parts; the evaluation of the contribution of a single oscillator; and the summation over the frequency distribution of oscillators. The two basic experimental facts about the heat capacity of solids which any theory must explain are : @ Near temperature, the heat capacity of most solids, is approximately =3R =249J mol! K* =5.96 cal deg” mol", At low temperatures, the heat capacities decrease and vanish at T = 0. The experimental results show that the heat capacities are of the form CaP +yT where @ and y are constants. For an insulator, y = 0, we have a pure T°—behaviour for the heat capacity. For metals, y #0 and hence C, also contains a term linear in 7. This linear term is a contribution to the heat copes. from the conduction electrons. The T-term in both cases is due to the lattice vibrations. The classical theory of specific heat has as its foundation the law of Dulong and Petit room @)224 (1819): the specific heat is the same for all ele- mentary solid substances, about 6 calimole K. Fifty years later, Boltzmann pointed out that this result could be rationalized in terms of the energy of vibrating atoms in the solid. From the kinetic theory of gases, the average kinetic cnergy of a moving particle resolved along one coordinate direction is 1 KE, = 3547 Thus the average kinetic energy per particle in three dimensions is and the energy of a mole (N) of moving particle is KE mole = NET = - the internal energy of a perfect gas. For bound aioms, vibrating in a solid, there are three additional components of potential energy, cach equal, on the average, to 3 AT. Therefore, the internal energy per mole of solid is E = 3RT calimole and CG, dE) _3p= (@), 3R = 5.96 cal Imol K = 2494 J/mol/K the value obtained by Dulong and Petit. Thus, the classical theory predicts that the heat capacity per mole is the same (= 5:96 cal. mol K") for all substances at ail temperatures. This is the Dulong-Petit’s Jaw. The Dulong and Petit law is, generally, correct at room temperature and above for elements of atomic weight greater than 40, (i) At very low temperatures, the specific heats of all elements approach zero. A number of light, high melting point clements (B, Be, C and Si, for example) show very much lower Solid State Physics and Devices room temperature values than the Dulong and Petit law predicts. (i) Finally, for a number of very electropositive metals (Na, Cs, Ca and Mg), C, rises with increasing temperature considerably above 3R. A few values for molar heat capacities at constant pressure (1 atmosffaere) and 298°K are shown in Table 2.1. Comparing the valucs with 249 J mot’ K", we find that at high temperatures, the Dulong Petit’s law only holds approximately and can be quite wrong. Table 2.1. Molar heat capacity C, (J mol! K#) at 298 K Au 254 ‘At low temperatures, the agreement between experiments and Dulong-Petit’s law fails altogether. Experimentally, it is found that the heat capacity approaches zero as T ~ 0. In fact, the heat capacity for non-metals approaches zero as T, which is characteristic of the lattice vibrations. In summary, it may be remarked that the disagreement referred to above, is absolutely comprehensible as it is well known that at temperatures around absolute zero, the quantum effects become more pronounced and any effort to explain any property in that range of temperatures on the basis of a classical theory is bound to unsucceed. 2.15 Einstein Theory Basic postulates of the theory are : (i) A crystalline solid consists of N atoms, cach of which vibrates independently; there is no interactions even between nearest neighbours. Each atom vibrates in the three independent directions with a constant frequency vg, frequencies of the atoms being the same because of their assumed identical surroundings. theLattice Vibrations and Specific Heats (iii) To assign quantum character to the vibrations, it was postulated that cach atom is equivalent to three mutually perpendicular Planck oscillators. Thus, in all there are three 3N harmonic oscillators. (iv) Following —Planck’s. quantum hypothesis, cach of the 3N harmonic oscillators can have the quantised energy (n = 0, 1, 2,-++) (+3) wv so that the average energy E in thermal equilibrium at temperature T is given by E=Zp,E, n Here p, is the probability of the oscillator being in the level £, : this is given by a Boltzmann factor so that E (o+3) mee [- (+d) mr] E, = (241) nz0 E = = ep [- (= + 5) von n=0 = nhy exp(— nhv/kT) = iw +e aay exp(— nhv/kT) n=0 ater [. Lap(- al 2 aT |" veo But SeP WHT oye WNT gg WAT. no so that Fal hy E=t™+imaapoy 2) This mcan cnergy is readily interpreted in terms of phonons. For bosons of energy hv, which 2.25 are not conserved, the average number present in thermal equilibrium at temperature T is 1 0) = SoGw7eT) =a (2.44) Multiplication of eq. (2.44) by Av gives the second term in cq. (243) as the contributions of the phonons to the energy. The fist term is eq, (243), Fav, is the zero-point energy, which cannot be frozen out because of the uncertainty principle. Figure 221(a) shows that the mean energy tends to this value in the low temperature limit kT << hy. At high temperatures kT > > hv, we can expand the exponential to obtain hy 2 1 (av [3(8) * 1 1m) = pert 9 Be] -ofo-(H | so that the classical equipartition value is obtained in this li note that a one-dimensional harmonic oscillator gives kT, not 447, because the potential energy, as well as the kinetic energy, is a contribution to the total energy. 1 E=Sh+ quadratic The specific hea C is found by differentiating eq. (2.43) with respect to temperature : (2.45) x (% ae x. ( (ePe/T _ 17 where 6, = Avik, This is shown in Fig, 2.21(b); the specific heat vanishes exponentially at low temperatures and tends to classical value k at ~ high temperatures. The gencral features of Fig. 2.21 are common to all quantum systems: the energy2.26 ) Fig, 2.21 Mean energy and specific heat of @ harmonic oscillator. tends to the zero-point energy at low temperatures and to the classical equipartition value at high temperatures; the specific heat tends to zero at absolute zero and rises to the classical value in such a way that the ‘missing area’ under the classical specific heat curve (shaded in Fig. 2.21b) is equal to the zero-point energy. Considering 3N independent oscillators (N atoms, three dimensions), we can generalise eq, (2.45) to obtain the molar specific heat as Og)? [__ eT C= INC = 3R. @) wry 09 o Gent where the Einstein function is defined by Solid State Physics and Devices re (%) = (%) ee Cd el G T(K) Fig: 222 Tpical curve showing dependence of specific heat of solids on temperature. If we now plot CG, versus T using this equation, we obtain a curve of the same general shape as Fig. 2.22, which indicates that the theory is now in agreement with experiment, at least qualitatively, over the entire temperature range. Note in particular that C, > 0 as T-> O'K, a new and important feature of eq. (2.46) which was lacking in the classical theory. The temperature Og is an adjustable parameter choscn to produce the best fit to the measured values over the whole temperature range. Figure 2.23 illustrates the procedure for copper, where Of is found to be 240°K. The fact that such a good agreement is obtained over such a wide temperature range by adjusting only one parameter is indeed impressive. CuO = 240K Cy ca/gmol. K T(K) 100200300" Ma 2 pele et 0 Cle yams eres ‘The represent experimental values, and the curve is ey he Einaen pasionLattice Vibrations and Specific Heats 227° We can calculate the Einstein frequency Yg once we have determined the temperature 6g, Thus, for 6, = 240"K, the frequency v, = 16,/h is about 25 x 10°}, which is in the infrared region. Let us now examine the behaviour of C,, as given by eq. (2.46) in extreme temperature limits : In the high-temperature limit, where T >, one may expand the exponential fet in a power series of 65/7. Carrying this out and retaining only the largest terms in the series, one finds that C, ~ 3R, which is the classical result. This is to be expected, of course, because in the high-temperature region the quantum aspects of the problem become irrelevant, as we have previously indicated. In the low-temperature range, in which T<6,, the exponent 7 in eq. (2.46) is much larger than unity. The expression for C, then reduces to y= se (22) ee v T = Be eT 47 where B(T) is a function relatively insensitive to temperature. Because of the exponential ¢~e/7, the specific heat approaches zero very rapidly exponentially, in vanishes at T = 0 K. Although the fact that C,-» 0 as T OK agrees with experiment, the manner in which this is approached is not. Equation (2.47) indicates that C, approaches zero exponentially, while experiments show that C, approaches zero as, T°. The decrease predicted by eq. (2.47) is much faster than warranted by experiment, and this, as we shall see, is the basic weakness of the Einstein model. The Einstein model may be summarized as follows: (@ At high temperature, the oscillator is fully excited, acquiring an average energy equal to KT, which leads to a molar specific heat C, = 3R. Gi) On the other hand, at low temperature, the oscillator is essentially unexcited, and hence C, = 0; in other words, the oscillator is "frozen" in its ground state. This “freezing” is also the reason why the vibrational modes in diatomic molecules, such as H;, do not contribute to specific heat, except at high temperatures. In most respects, the Einstein model has been a remarkable success; its results are in good agreement with experiment over most of the temperature range. Nevertheless, the model is incorrect at very low temperatures, at which it predicts a Asoc heat that is much smaller than the observed value. This disagreement could be explained as follows: The atoms in the Einstein model were assumed to oscillate independently of each other. Actually, the idea of independence here is not a viable one because, since the atoms do interact with cach other, the motion of one atom affects its neighbours. The motion of these in turn affects their neighbours, and so forth, so that the motion of one atom anywhere in the solid, in fact, affects all other atoms present. Thus we need to consider the motion of the lattice as a whole, and not a single independent atom. That is, we must consider the collective lattice modes. The most familiar example of such collective modes is the sound waves in solids. When-a sound wave in a solid, the atoms do not oscillate independently, their motions are orchestrated in such a manner that they all move with the same amplitude and with a fixed phase relationship. ‘The unrealistic assumption about the existence of a single frequency of vibrations of a_ solid was first realised by Nernst and who showed that the ‘rimental data of specific heat could be. better fitted by a modified Einstein formula. Instead of using a single frequency v, they assumed that half the oscillators vibrate with the frequency v and the other helf with frequency v/2. Furthermore, Madelung suggested that one should not associate just a single wavelength with atomic vibrations, but should include a whole spectrum of acoustic waves. Thus, by 1911, it had been suggested that Einstcin’s formula could be improved by employing a large number of frequencies, and that the frequencies were to be associated with elastic (sound) waves.2.28 Solid State Physics and Devices These considerations led Debye to remove the disagreement between theory and experiment in the form of Debye model, to which we now turn, 2.16 Debye Theory of Atomic Heat (a) Postulates : The Debye model is based on the following assumptions: (@ Solid is an isotropic clastic continuum, (i) The motion of an atom influences its neighbours’ motion, ie., the atoms are coupled and vibrate collectively, Debye took into account the interaction between vibrating particles by hypothesizing a spectrum of elastic standing waves traversing the crystal, somewhat like a bow! of vibrating jelly, that approximates the thermal vibrations of atoms, (iii) Atomic vibrations “have complete spectrum of frequencies which vary in a continuous fashion. (iv) The medium is dispersionless, ie., the velocity of sound is independent of the frequency. (v) The number of standing sound waves which can be fitted in the solid with frequencies in the range between v and v + dy determines the spectrum, which is characterised by a cut-off frequency vy (known as the Debye frequency). (vi) Polarisation : The waves produce atomic displacements which can be resolved so as to be parallel to the wave vector (longitudinal waves) and in two directions at right angles to it (transverse waves). Thus, each density of state really corresponds to three independent waves—two transverse and one longitudinal. Debye frequency vp is taken to be the same for both the longitudinal and transverse branches. Debye realised that there are not only high-energy modes, as in the Einstein distribution model, but also very low-energy modes as well. For example, sound waves have very low energy and very long wavelength. Thus, there will be contribution to C, at temperatures, corresponding to that of the energies of the modes. Thus, the low-energy modes cannot be ignored. Debye suggested that a more accurate way to characterise the vibrational modes of a crystal would be to treat the crystal as a continuous medium that is isotropic. For elastic waves in a three-dimensional isotropic solid, one longitudinal and two transverse vibrational modes can propagate. (The displacements are parallel to the propagation direction for longitudinal vibrations and perpendicular to transverse vibration.) We usc the subscripts / and ¢ for longitudinal and transverse waves, respectively. The linear dependence of @ on k, that is, @ =vk in the Debye model for both transverse and longitudinal waves should be emphasized. In the Einstein distribution v(=v,) is a constant high frequency value that represents the independent oscillation of all the atoms; vz has no k dependence. In the Debye model, the atoms play no direct role; the solid is assumed to behave as a continuous medium. (The bchaviour of a rcal solid is in between these extremes) Since the Debye approximation considers very long wavelength, that is, small k-modes, the longitudinal (¢,) and transverse (c,) velocities can be related to the elastic constants of the solid, from which it could be deduced that oe for real crystals. It should be noted that in the Debye model c, and c, are independent of the wavelength, as in the continuum. (b) Density of states : We pointed out that the discrepancy between the Einstein thoory and experimental results in the low temperature region was a consequence of the oversimplified model employed by Einstein. In 1912, Debye tackled the problem from a different point of view and, as we shall see,Lattice Vibrations and Specific Heats with great success. Debye realized that it is possible to propagate waves through solids covering a wavelength region extending from low frequencies (sound waves) up to short waves (infrared absorption). The essential difference between the Debye model and the Einstcin model is that Debyc considers the vibrational_modes of a crystal as a whole, whereas Einstein's starting point was to consider the vibration of a single atom, assuming the atomic vibrations to be indepentont of each other. Now, we shall deal with the vibrational modes of a continuous medium, because the results are basic to the "continuum theories" of the specific heat. Let us first consider for simplicity the vibrational modes of a one-dimensional continuous string of length L. Suppose u(x,t) represents the deflection of the string at the point x at the instant £. The waves may then be described by the one-dimensional ‘wave equation. eu 23 (2.48) where ¢, is the velocity of propagation of the waves. If it is assumed that the end points of the string are fixed, the solutions of cq. (2.48) are those corresponding to standing waves : UG, f) =A sin (nxx/L) cos mvp (2.49) where n is a positive integer>1. The wavelengths and frequencies of the possible vibrations represented by eq. (2.49) are given by Ay = 2L/n and Yq = C/A = gn /2L The. frequency spectrum is discrete, one frequency corresponding to cach integer value n..Note that for the one-dimensional string the frequency spectrum corresponds to an infinite number of equidistant lines, as illustrated in Fig, 2.24a. The number of possible modes of vibration in a frequency interval dv is, on the average, equal to (2.50) dn= ¢) a (251) In the three-dimensional case, the wave equation reads 229 au, ou, Fu oy wiete ae Assuming 2 continuous medium inthe shape of ae be a the poaal pee i Solutions ur, a eualegy rites GN U(%, 42, £) =A sin (n,70x/L) sin (ny ty/L) X sin(nwz/L) cos Inve (253) where now fy, my, and ng are positive integers = 1, Substituting this solution into the differential juation , the i eq 2 (2.52), one om felosiss leraea at) (a) tv): (b) Fly 2.24 (0) Frequency spectrum for a foe cons wis according 10 (251). (b) Frequency spectre tree drosonal conan, receding (37),2.30 Thus the possible wavelengths and frequencies are determined by three integers in this case. Let g(v) dv be the number of possible modes of vibrations in the frequency interval between v and y+dy. To calculate this, consider a network of points, cach point being determined by three Cartesian positive integer coordinates ry my and nz, Writing Reantntnt _ 4 3 (255) it is evident that the number of points in a shell between R and R + aR is janR aR (2.56) Here cach point occupies on the average a unit volume in the integer space. Now, each point corresponds to a set of three integers ng my my and cach sct of imtegers determines, according to eq. (2.53), a possible mode of vibration; hence eq. (2.56) immediately gives the umber of possible modes of vibration in a given range. Expressing R in terms of v in eq, (256) one thus finds 0) d 3 amr’) =F Rar a? and Solid State Physics and Devices ‘4 = ( 3] vdv (2.57) where V= L? is the volume of the solid. For a perfect continuum, the possible frequencies vary between 0 and ®, the number ‘of such possible vibrations increasing with the square of the frequency [see Fig. 2.24(b)]. Now remember that each frequency v is associated with one longitudinal (/) and two transverse (!) modes of vibration so that the total density of states is fv) = 2 SD =LO +O sol = gv) + 23(v) eae a 3) (2.58) a @ (c) Debye approximation : Consider an clastic wave propagating in a crystal of volume V. As long as the wavelength of the wave is large compared with the interatomic distances, the crystal "looks like" a continuum from the point of view of the wave. The essential assumption of Debye is now that this continuum model may be employed for all possible vibrational modes of the crystal, Furthermore, the fact that the crystal actually consists of atoms is taken into account by limiting the total number of vibrational modes to 3N, N being the total number of atoms. In other words, the frequency spectrum corresponding to a perfect continuum is cut off so as to comply with a total of 3N modes (see Fig. 2.25a). The Debye cut-off procedure leads to a maximum frequency Vp (the Debye frequency) common to the transverse and longitudinal modes; it is defined byLattice Vibrations and Specific Heats *D 2021 “> S fw) dv raw 4] [Pay 0 1G} 0 = (2.59) (2.60) Putting eq. (2.60) back into eq. (2.57) yields M=ae 61) Figure 225(a) shows the _individual contributions of longitudinal and transverse modes of vibrations, whereas Fig. 225(b) illustrates flv) = f,(v) + f,(v) as a function of v. Here f,(v) = 2 (¥). Mig, 225 (a) Consibuion of | and « modes 0 f(y. (b) flv) from Debye and Einstein models. (@) Intemat eneny U and C, : The total internal energy on the Debye model is ue} fv) hv dv 4 ap (hv/kT) — 1 231 ee = 3000 g 2000 & > 1000 0'T00 200 300 400 500 600 |) > Fig, 2.26 ‘Temperature dependence of internal energy. ‘> ya f Ye ag) vB SP GWwAT)—1 Figure 226 shows the graph to represent the temperature dependence of internal energy U. It is convenient to express vp in terms of the Debye characteristic temperature, defined by kOy = hp (2.63) Also, we can simplify the notation by putting x= Av/kT c= hdv/kT hy, and w= te In terms of the new notation, eq. (2.62) for the energy assumes the form (264) > 4 an{a fice =a [3 f see} (265)232 Solid State Physics and Devices ext a ‘ ¢-1F or C= 3R3 & c. 4, « Gar, (4) i %/T R 2 (2.6) where the Debye function Fp is given by 4 Ee a) T OUT)" \h) 9 =e Thus, in the Debye model, the heat capacity depends on only a single parameter, the Debye temperature @5. It is not possible to evaluate the integral in eq. (2.65) analytically but the function in curly parcnthcses is available in literature in the tabulated form. Figure 2.27 shows a typical Debye C, alongwith the Einstein C, and the measured data. plot To compare eq. (2.65) with experimental sr 7/18 > Fig. 2.27 Calculation of she latice specific heat by Einstein land Debye modds, with she data for Al and Cu results, one must know the Debye temperature. @y. This is determined by choosing the value which, when substituted into eq. (2.65), yields the best fit over the whole temperature range. We see from eq. (2.65) that if C, is plotted versus a reduced temperature 7/6,, then the t 5 x“ =4 © Cu, 343K :. + Ag, 226K 2 = Pb, 103K 32 8 + ©, 1860 K Jt 0 05 1 Ty —> Fig. 2.28 Specific heats versus reduced temperature for four same curve should be obtained for all substances. That is, there is a universal curve for specific heat. This observation is tested experimentally in Fig 228 for four widely different substances, where we see that it holds true to a remarkable degree. The values for 65 have been tabulated in the literature, and a short list for important substances is given in Table 2.2. The value of @, depends on the given substance, but, as Table 2.2 shows a typical value is about 300 K. The corresponding Debye frequency vy =kO,/h is about 10 Hz, which lies in the infrared region of the spectrum, Table 22 Debye temperatures Li 335 NaCl 280 Na 230 K 470 cu 680 AR 255 Au Al Ga Pb Ge Si CcLattice Vibrations and Specific Heats 2.33 High-temperature limit. The is defined by T > >On 0% < <1 high tem- In this case, we rewrite the integrand in q, (2.65) in the form te a -1" ae ne -1) af Then in this limit, eq. (2.65) becomes 3° oon i va or C,=3R (T>>6,) (2.67) which is Dulong and Petit law. Thus, at high temperatures, Debye theory conforms to the Dulong-Petit law. This can also be seen from the following intuitive argument: When T>>@y, every mode of oscillation is completely excited, and has an energy = KT. Therefore, "> u= J eN {ea ° Note that the saturation in specil heat is already evident when T=6,. Since 85 is typically close to room temperature, we see that , begins to approach its classical value 3R for T equal to room temperature and above. Ce Goulesthalf mole deg’) Fg. 229 Companion ofthe T° law and observed values for ket Low-temperature C,. The low-temperature limit is defined by T < <@,i€,.25 > > 1. Since the integrand in eq. (2,64) goes to zero rapidly for large values of x (it goes like x‘ e~*), we can replace the upper limit by infinity in this case. Equation (2.65) then becomes for T< < Op. T)*|, 7 atéde =3R 3 & (é) [ 4 sea (T<< 6) (2.68) Now, the standard integral xiédr__ ant o¢-y 5 3 Ga (@) (2.69) which is known as Debye T’-iaw for atomic heat, and states that the Debye specific heat decreases as at low temperatures, in agreement with experiments. ‘The 77-law is illustrated in Fig. 2.29 for KC1, representing results obtained by Keesom and Pearlman. In the low temperature region, from eq, (2.69), 12x, (T)* u=Jcat=f an (g) aT2.34 4prt or 50 This shows a T* dependence for the internal energy in the low temperature region. These results are reasonably accurate for temperatures below 65/10. If z is the number of atoms in one molecule, then For z = 1, this leads to r)! = 4644 (F) cal/mol K ad = 1944 (2) J/mol K The cubic dependence of C, may be appreciated from the fact that at low temperatures, only a few of the modes are excited, There are the modes whose quantum energy Av is less than kT. The reason for the error in the Einstein model at low temperature is now evident. This model ignores the presence of the very low- frequency, long-wavelength modes which can absorb heat even at very low temperature, because their energies of quantization are very small. The exponcatial freczing of the modes does not actually occur, and the specific heat has a finite value, small though it may be. (e) Elaborations of the Debye model : The Debye theory agrees well with the experiment, because at low temperatures as well as at high temperatures it is exact. It is, in fact, an interpolation formula between two correct limits. However, at intermediate temperatures, the fit is not so good. In this range, the rise of C, with T can often be described with tolerable accuracy by either an Einstein or a Debye model. Solid State Physics and Devices 0s 10182035 ™ — Fig 230 Temperature dependence of the Debye characteristic temperature Oy (as measured from specific heat) for metalic indium. The success of the Debye model at very low temperatures can be understood from the fact that only low frequencies arc excited which is consistent with the treatment of elastic continuum. The failure of the model in intermediate temperature range originates from its inability to incorporate the actual crystalline nature of the solid, and the assumption of common frequency cut-off for both the modes as also the neglect of the temperature dependence of Yp, Which is shown in Fig. 2.30 for indium. Since the Debye model involves a rather drastic. simplification of the actual f(v), we may expect that the value of @, which fits C, at a certain temperature will not necessarily be the optimum value for all temperatures. However, for any pair of measurements (C, and 7), it is possible to calculate the equivalent parameter 6,, A series of such measurements then permits us to plot a curve of 8 against T. There is always some variation of 6, with temperature, though for most solids the variation is smaller in magnitude than that shown by the curve of Fig. 2.30. The principal characteristics of the curve in Fig. 230 (that 6 should first decrease and then increase again on warming) are accountedLattice Vibrations and Specific Heats for by a number of theoretical models based on the Einstein and Debye approaches. As we have noted, Blackman made some of the early attempts at calculating /(v) for simple lattices and deducing the corresponding curves of GC, versus T and @p versus temperature. Blackman also showed that @, has a minimum at an intermediate temperature if one represents all the acoustic modes in a diatomic lattice by a Debye distribution and all the optical modes by an Einstein function, Much more complicated curves for (v) are now available, thanks to computer-aided theoretical calculations of vibrational modes, and thanks to neutron scattering and other new experimental techniques. Such work is unlikely to add in large measure to what we know about lattice specific heats; however, for C, and Cy are indifferent to these small details, When we use temperature as a variable in scanning lattice energy, this is spectroscopy with very coarse resolution. The more recent and complicated curves for f(v) can, however, tell us significant things about important groups of phonons: their energy, wave-vector, directionality, polarization, and ability to interact with other phonons and other excitations. (f) The Debye Temperature : The concept of the “Debye temperature 65 is useful in connection with many topics in solid state physics other than specific heats, a fact which may come as a surprise. It may have seemed in our discussion up to this point that the Debye approach to specific heat ignored the periodicity of a crystal lattice and the Brillouin zone limitation on the physically realizable ranges uf wave vector and frequency. This is not entirely the case, since the w, of Debye’s model is automatically comparable with the angular frequencies of phonons whose wave-vectors are close to the zone boundaries. These are the majority of phonons which are excited at temperatures higher than 6). Conversely, at temperatures well below the Debye temperature, the only phonons excited 235 are those for k quite close to the centre of the zone and well away from zone boundaries, ‘Thus, phenomena such as thermal conduction (controlled by the anharmonic coupling of phonons to other phonons) and electrical conduction (controlled by the scattering of electrons by phonons) take different forms above and below the Debye temperature. Above the Debye temperature, most phonons have wavelengths of only a few atomic spacings. However, at temperatures well below the Debye temperature the most probable phonon wavelength is of the order of (a9,/T). ‘This wavelength may be hundreds, or indeed thousands, of atomic spacings at sufficiently low temperatures. (g) Shortcomings : Although the Debye theory of specific heats is extremely useful in fitting a broad sweep of crystalline solids, it does have defects. In fact, despite its impressive success, the Debye model also. remains only an approximation, The nature © of the approximation, as pointed out previously, lies in assuming the continuum dispersion relation to hold true for all possible modes of excitation. Experimentally, the approximate nature of the Debyc model is shown by plotting @ versus T over a wide temperature range, where 8, is found at each temperature by matching the experimental value for C, with eq. (2.66) at that temperature. If the Debye model were strictly valid, the value of 6, so obtained should be independent of T. Instead one finds that 6, varies with T as shown in Fig. 2.30, the variation reaching as much as 10% or even more in some cases.. Even in the 1920's it was appreciated that 6, obtained from fitting C, was not always in agreement with values derived from the elastic constants, as Table 23 shows. As more accurate values of C, were determined experimentally, it was observed that the calculated C, did not fit the experimental results exactly. This is usually demonstrated by calculating, at each temperature, the effect2.36 values of 6, that are necessary to fit the experimental data. For most materials, we find that as the tcmperature is lowered, 6, is constant to about 6,/2; below this temperature it starts to decreasc, usually having a minimum 704/10, then increasing at lower temperatures and again becoming constant at very low temperatures. The valucs obtained at high and low temperatures are usually not the same, Table 23 Debye temperatures (in K) for some materials In the table thermal temperature specific heat. In order to improve on the Debye model, one needs to remove the long-wavelength approximation and use, instead, the correct persion relation and the corresponding density of states. refers to low 2.17 Summary © A. particle is said to execute harmonic motion if its potential energy is proportional to the displacement. © If the particle's potential energy contains cubic and higher order terms in addition to the quadratic term, its motion is anharmonic. © A normal mode is a correlated motion of the atoms that has a chracteristic wave-vector and frequency. If this motion is started, it will continue for ever, ignoring friction, i.e. anharomic terms. © In the long-wavelength limit, a crystal may be treated as a continuous medium, because the wavelength is much greater than the lattice constant. The dispersion relation is linear : wack where c, is the velocity of sound, Solid State Physics and Devices ® Consideration of boundary conditions shows that the modes of oscillations—the sound wave —have discrete values of k and w, The density of modes is wo) = (2) ma\e @ Elastic energy of sound waves in solids is quantised, and the quantum unit is called the Phonon. The phonon carries an cnergy E=ho and a momentum p=nk © The dispersion relation for a onc dimensional monatomic lattice, with nearest-neighbour interaction, is © = Gm Sin (ka/2) where the cut-off irequency Wp = (4a/m)!2 (i) The dispersion curve is linear near k =0, the long-wavelength regime, and saturates at large value of k. (ii) The lattice acts as a low-pass filter : only waves whose frequencies are lower than @,, are transmitted; modes with frequencies exceeding mare heavily attenuated. @ The dispersion curves fora one-dimensional diatomic lattice consist of two branches : the lower one the acoustic, and the upper the optical branch. The character of the acoustic branch is similar to that of the monatomic case, while the optical is essentially flat throughout the k-space. There is a frequency gap between the two branches, and thus the lattice acts as a band-pass filter. @ The dispersion relation in a three dimensional lattice is an extension of the ‘one-dimensional case. The wave vector k is now a three—dimensional vector, and the frequency is a function of both theLattice Vibrations and Specific Heats 2.37 magnitude and direction of k. Thus the dispersion has the form o=axk) (i) A Bravais lattice has three acoustic branches ‘one longitudinal and two transverse. (i) A non-Bravais lattice, with s atoms per unit cell, has 3s branches, three of which are acoustic and the remainder optical. Gi) The dispersion curve exhibits. symmetry properties in k-space : (a) The translational symmetry : @; k +G) = 0,(k) (where G is the reciprocal lattice vector) enables us to restrict consideration to the first Brillouin zone only. (b) The inversion symmetry : @i(-k) = (ke) and the rotation symmetry establishes the relation between various regions of the Brillouin zone. © Two basic experimental facts about the heat capacity of solids which any theory must explain arc (i) Near room temperature, the heat capacity of most solids is approximately = 3R = 249F mol! K7! = 5:96 cal mot”! K7! (ii) At low temperatures, the heat capacities decrease and vanish at T = 0. The experimental results show that the heat capacities are of the form C Solution, C,=A ral 8 Dividing the above equations, we haveLattice Vibrations and Specific Heats =a §) os 38x10? \S) 125 eB. "2 x= 355 x 38X10) = 304 ig 3 = 5X10 = 2.43x 1073 J/mot .K Example 16. At low temperatures, the Debye @ for NaCi and KCl, which have the same crystal structure, are 330 K and 220 K respectively. The lattice specific heat of KCI at 5K is 3.8 x 10” JimoLK. Estimate the lattice specific heat of NaCl at 10K. Solution. = 9.0x 10-7 Jimol .K Example 17. 6p for a substance is 320 K. Calculate its internal energy U for 10 motes at 8K. Solution. For T < 9/10, we have “G ry? 1944 (2) J/mol .K r r 2 Usn f Cat = 5 pyr 0 % 0 8 219s [7A] _ 1944s a i 4x 320)° é 3 8 U = 19440x2x (&} =1944x2x [= 241 60.15x10777 64x 10° Example 18, Potassium has the bec. structure with lattice constant 5.3 4. Calculate the Debye temperature on the assumption that the transverse and longitudinal elastic waves in potassium have the same velocity ¢ = 1.5 kenis. Solution, Doo 3! nv? v) = 4nV [+S | = Ae (3 ay 8 = 6, We have 4nV ( = ay Bowne _ 30 0° EV ~ SVN) Now V/N =a°/2 for bec. Aa. (:) 3 > ar/a'/2 \a) “2 1500 or Gy = 102{o.02x10™) ° (3x107%) (138x107) a 30 = 102x557 or 6) = 143K Example 19, Prove that a_ solid comprising N atoms in a volume V will have a Debye temperature given by hes yy O= (= (o? ) where cy = ¢;. Solution, We have242 Solid State Physics and Devices aka) 92°? ede to ae vnn Be ie oa Here ynin = 2a @ . tim it 7 9, = (ann 2 >= ° Be usR_ ft [f-14¢] Example 20. Show that the total Do & sibrational energy of a linear monatomic chain re, is proportional to T? at low temperature and U=RE- J xe de hence that Cy « 7. Solution. According to Debye model, the Pretare |[ee ta internal energy of a monatomic lattice is a % = feeE = hy =[-e“b =1 U=f soya. | ° aT p U=Rk> where from eq, (2.51) : 8 Hence, at low temperatures, veP where c, is the velocity of the elastic wave, and ad Cw We Ea T L = Na = length of the chain. "aT \% = Na = length of the chain, wo Ger ya 2Ne phd 2.19 Problems C4 mT - 1. The force between Na* and CI~ pair Letting o x leads to dv = Ta oar oquiibreas igi by . wd HB - +Bexp(-1/R) ap fee dneg? 0 €-H where B = 545x10~ N and R = 0.321A. %» Find the spring constant of the pair. Ho fz (101 Nim}Lattice Vibrations and Specific Heats 2 10. In an assembly of 10% simple harmonic oscillators, each of frequency 10° Hz, calculate (ignoring the zero-point energy), the mean thermal energy of the system at 20 K. [63x107 7 The Young’s modulus and density of diamond are respectively 8.310" dynelcm? and 352 gicm?. Estimate the Einstein temperature from these data. (1308 kK] In a copper cube 1.0 cm on a side, what is the number of transverse elastic vibrational waves having wavelengths between 1.0x10~* and 1.1x107> om. The speed of transverse elastic waves in copper is 2.3X10? mis. [21x10] A system of oscillators at a temperature 300 K has a certain number of them carrying an energy, E= 1077! J, and a certain number having twice this energy. Caleulate the ratio of the number of oscillators in the two energy states, and also the ratio of the energies carried by the two groups. [1.27; 0.635] Show that, for a one-dimensional diatomic lattice, for the acoustic branch, for small ka, the group velocity coincides with the phase velocity. Prove that for a diatomic linear chain, the gradient of the optical branch of the dispersion curve at the centre of the Brillouin zone is zero. Calculate the Einstcin temperature of cesium using the ion plasma frequency = 1.724x10” Hr, [82.4 K] 6, for Na metal is 150 K. Find the molar specific heat at 10K, [0.574 Jimol.K] ‘The speed of the transverse elastic waves through copper is 2.32x10? m/s and of the longitudinal waves is 4.76%10? m/s. The number density of copper atoms is 8.49% 108 atomsim?, (a) Find the Debye frequency, vp. (b) Calculate the Debye temperature, 8p. [(a) 7.11x10" He; (b) 341 K] 11. 12. 13, 14, 15. 16. 1. 243 The number density of sodium atoms is 252x108 atomsim> and 6p = 150 K. Find the velocity of propagation of the clastic waves in Na, if these have the same value for longitudinal as well as transverse waves. [1.71x103 m/s] The average velocity of elastic waves in Cu is 26x10? mis and a =234. Caleulate 6, for Cu. B19] 9p for diamond is 2000 K, its density is 3500 kgim?, and its atomic weight is 12, whereas its spacing is 15 4. Calculate the velocity of sound in diamond. Make an estimate of the dominant phonon wavelength at 300 K and calculate the frequency of lattice vibration to which this corresponds, [1.2x104 m/s; 104; 1.2x10" Hz] Diamond (atomic weight= 12) has Young’s modulus of 10" Nm~? and a density of 3.5 gcm~?. Ignoring crystalline anisotropy and the difference between longitudinal and transverse modes, calculate the Debye temperature, [2700 x] Estimate 6, for gold, if its atomic weight is 197, the density is 1.9x10* kg m™> and the velocity of sound is 2100 ms™*, [169 x] Prove that the Debye radius is given by p= (ny? where 1 is the concentration of atoms in the solid. In contrast to common frequency in the Debye model, Bom proposed (0 cut-off the spectrum in such a manner that the longitudinal and transverse modes have a common minimum wavelength, Aynin- Following the Debye procedure for the Born model, prove that the specific heat of a solid is given bySolid State Physics and Devices o=a[n() 2m ()] where the Debye function 6p) _ (7)> 7 tahae r (#)-3 (8) fey and with s =/,¢ he, Kdmin . Qualitatively the melting point of a crystalline solid can be defined as the temperature 7,, at which the amplitude of simple harmonic vibrations of the individual atoms about their lattice sites is about ten per cent of the interatomic spacing, a. Show that on this picture, the frequency v of the vibrations of a monatomic crystal (atomic weight M) satisfies the proportionality relation ue a= 1 (Tm 6elG ‘The Einstein temperatures of A/ and of Pb are 240 K and 67 K respectively. Guess orders of magnitude or look up data to decide whether — these temperatures and the above formula are in reasonable agreement. {From the Binstein temperatures for Ai This is what one should expect since @ Pb has a low melting point compared with Al; (i) Pb has a large atomic weight compared with 7; (iii) the” Pb atoms being large, Pb will have a larger lattice spacing than Ai. The data for Pb and AI substituted in the above formula give wad v(PB) +3] 2.20 Review Questions 1. 4, 10. . Explain Distinguish between a harmonic and an anharmonic motion by giving suitable mathematical expressions. Define a simple harmonic oscillator. By expanding the potential energy of an atomic oscillator, derive a formula for the Einstein frequency. the meaning of the normal modes of a lattice. Discuss the salient features of the dispersion relation for the vibrations of a one-dimensional monatomic chain, How do they differ from that of a diatomic chain? . Discuss the formation of Brillouin zones of vibrations one- dimensional lattice, and define them. . Define the deasity of states and draw the corresponding graph. Discuss the salient features of vibrations of a three- dimensional lattice. Define a phonon and give its characteristics. ~ . Draw a typical experimental C,-T graph for a solid. Explain its shape on the basis of Einstein and Debye theory. Discuss the Debye theory of atomic heat and give its shortcomings. 2.21 Objective Questions 1. The knowledge of lattice vibrations is useful for providing information about (A) specific heat (2) thermal expansion (3) thermal conductivity (4) all the above Einstein's theory of lattice vibration is based on the quantum theory of (J) Schrodinger (2) Dirac (3) Planck. (4) HeisenbergLattice Vibrations and Specific Heats 3. The density of states curve are singular whenever (1) the phase velocity vanishes (2) the group velocity vanishes (3) the phase velocity and group velocity are different and non- zero (4) the phase velocity and group velocity are equal and non-zero . If according to Born cut-off procedure, the Debye frequency @p is related to wave velocity ¢ by ¢ os @%=£ — @&=z ¢ a GB) => 4) =— . The spring constant a and Young modulus Y in solids are related by ¥ Qa=Ya Qa=7 aay @a=Ya* . Phonon is the quantum of (1) electromagnetic wave (2) elastic wave @) polarisation wave (4) magnetisation wave 7. A phonon interacts with (1) another phonon (2) a neutron }) an electron 4) all the above . For a linear diatomic chain, the long wavelength satisfies the condition: ()k>0 (2) phase velocity -> 2 (3) group velocity > 0 (4) all the above ). If there are two atoms in the primitive basis, then in the phonon dispersion will have (1) one acoustic and one optical branch (2) one acoustic and two optical branches (3) two acoustic and two optical branches (4) three acoustic and three optical branches 11. 12, 15 245 . In the harmonic model of lattice vibrations (1) there is no thermal expansion (2) there is a thermal expansion (3) the heat capacity is not constant at higher temperature (@) isothermal and elastic constants are different For a one-dimension in the Debye model (CC, «T Qg«P OGqer Mca In the Debye model for a three-dimensional crystal, the internal energy U at low temperatures is given by (1) UeT QU«T? QU«T? @U«T* . The experimental curve for 6, vs. T is Gg Tk @ Kw 7a) ® 7) Tk o 0 |, Debye temperature could be derived from (1) specific heat (2) elastic data (3) both the above (8) none of the above At temperatures well below the Debye temperature, most phonons have wavelength of the order of og oa 06, 9, o- OF246 16. 17. The specific heat at low temperature in the Einstcin’s model of lattice vibrations (1) remains constant with temperature (2) decreases exponentially with decrease of temperature (3) increases exponentially with decrease of temperature (4) drops linearly with decrease of temperature u. u. o “ The quantum mechanical expression for 21. In the Debye theory, a solid is regarded the mean energy of the atomic oscillator is as hw. (1) an isotropic discrete Q)av @ oes (2) an anisotropic discrete (3) an isotropic continuum: B) ear O para < rae y (4) an anisotropic continuum 22, Debye theory of specific heat is valid at i the Debs model, ev is expressed (in foe ‘enperaure my (2) low temperatures (1) @p = 0.710" "8, (3) intermediate temperatures (2) wp = 1.38%10? 6, (4) all temperatures @G) wy = 138x104 8, 23. The Debye model works for (4), =0.7x 107%, wo sryaali sles sas Baw non-crystalline sol + aa ee @) ft be abe Q)T=6, 0 tone @) T<0,/50 24, Which is incorrect as a Debyc-formula? (@) T=8/10 (1) G, = 4644 (7/0p) cal/mol.K |. Internal energy U vs T graph for Debye (2) G, = 1944 (7/0p)? Jmol K mee * (3) C, = (12145) R(T/8p)> uv (4) C, = (504712) R(T/6p)? 25, For a non-dispersive medium ()o=vk Qa=ve 7 r B)o=ve (4) @ = vk (n #1) w @ Answer-Sheet ooo3 Motion of Electrons 3.1 Introduction ‘The free electron theory of metals refers to the case in which the atomic valence electrons are treated as if they are free rather than being bound to the lattice points. Our assumption amounts, in fact, to supposing that the electrons move in a uniform potential rather than the true periodic potential provided by the positive ions. The basic assumption of the theory is that a metal is equivalent to a gas of free electrons in an otherwise empty box. The fact that electrons reside inside the lattice is altogether disregarded in this model. The idea that the large electrical and thermal conductivity of metallic substances might be explained by the presence of large concentrations of mobile free electrons in these materials was first proposed by Drude in 1900. The implications of this hypothesis were exhaustively investigated subsequently by Lorentz. Drude and Lorentz assumed that the free electrons in a metal could be treated as an ideal gas of free particles which, when in thermal equilibrium, would obey Maxwell- Boltzmann statistics. In order to examine what might happen when electric or thermal currents were allowed to flow, thus establishing a nonequilibrium state, it was necessary to investigate how the equilibrium distribution would be modified by a (small) electrical or thermal current. It was also necessary to consider the kinetic behaviour of the electrons as being that of free particles subject to instantaneous collisions which serve to return the distribution to the equilibrium condition, and to express the final result for electrical and thermal conductivity in terms of a mean free path or mean free time between these randomizing collisions. The Drude—Lorentz theory accounted satisfactorily for the well— known experimental law of Wiedemann and Franz, according to which the ratio of electrical to thermal conductivity for most metals is nearly the same, In addition, the magnitudes of the electrical and thermal conductivity could be obtained using values for the mean free path which were quite reasonable. The free-electron theory in its simplest form, however, led to a prediction of the electronic component of the specific heat which (assuming that the Debye theory for the lattice component is right) was in serious disagreement with experimental results. This difficulty (along with certain others) was resolved by Sommerfeld by the use of Fermi-Dirac statistics rather than the classical Boltzmann statistics. The Fermi-Dirac free electron picture scrves as avery simple and conceptually quite direct way of discussing and visualizing transport effects in metals. Since it starts with a preconceived notion of what a metal is like, it is of no value in explaining just why some substances are metals and have free electrons in abundance and others are insulators with few, if any, free electrons. For an explanation of — these underlying questions, one must begin by examining the quantum-mechanical behaviour of electrons in periodic potential fields, thus discussing essentially the same subject on a much ‘more fundamental level, Then the properties of metals and insulators can be understood in terms of quantum theory, and that the foundations upon which the free-electron theory rests, are, for the’ most part compatible with’ the requirements of quantum mechanics.3.2 It is possible to apply-the Schrodinger’s wave equation to the gas of free electrons to obtain the various electronic states of the crystals. At absolute zero, the first electron will be in the iowest energy state and subsequent electrons would occupy the next higher states, so that if there are N electrons altogether, the N lowest energy states will be filled up to a maximum cnergy Ep, called the Fermi energy. The corresponding maximum velocity vp, of the electrons is known as Fermi velocity and is given by (12)mv} = E,. The shape of Fermi surface inamomentum space (k-space) for the free electron model is perfectly spherical, whose radius~the Fermi radius ke, is given by Wke/tm = Ey. The ' electron energy level defined by E=E, is called the Fermi level, which is an important characteristic of any solid. At temperature in excess of absolute zero, some of the electrons will be excited to states of energy higher than E,, lying outside the Fermi surface and there will be some unoccupied states within the Fermi surface, The transition between full and empty states. will then no longer be a sharp onc. Nevertheless, if the temperature is small compared to the Fermi temperature (T, = E,/k), the region of transition will be quite thin. It is pertinent to remark that all the properties of electrons of solids are explained on the basis of the participation of the electrons residing in this transition rogion. 3.2 Conduction Electrons A metallic solid contains a large number of electrons which are free to move throughout the solid. These electrons give rise to the most distinguishing property of a metal—the electrical conductivity, These electrons are thus called conduction electrons. Let us provide the description of conduction electrons by an examples, using the simplest metal, Na, as illustration. Consider first Solid State Physics and Devices an Na gas, which is a collection of free atoms, cach atom having 11 electrons orbiting around the nucleus. In chemistry these electrons are grouped into two classes. The 10 core electrons which comprise the stable structure of the filled first and second shells (Bohr orbits), and a valence electron looscly bound to the rest of the system. This valence electron, which occupies the.third atomic shell, is the electron which is responsible for most of the ordinary chemical properties of Na. In chemical reactions the Na atom usually loses this valence electron—it being loosely bound—and an Na‘ ion formed. This is what happens, for example, in NaCl, in which the electron is transferred from the Na to the Ci atom. The radius of the third shell in Na is 19.4. Let us now bring the Na atoms together to form a metal. In the metallic state, Na has a bee structure, and the distance between nearest neighbours is 3.7 A. We see from Fig. 3.1 that in the solid state two atoms overlap slightly. From this observation it follows that a valence electron is no longer attached to a particular ion, but belongs to both neighbouring ions at the same time, This idea can be carried a step further: A valence electron really belongs to the whole crystal, since it can move readily from one ion to its neighbour, and then the neighbour's acighbour, and so on. This mobile clectron, which is called a valence electron in a free atom, becomes a conduction electron in a solid. Of course, cach atom contributes its own conduction electron, and cach of these electrons belongs to the whole crystal. These are called conduction electrons because they can carry, an electric current under the action of an electric ficld. The conduction is possible +) (-¥ <>< 3s electron Fig. 3.1 Overlap of the 3s orbitals in solid sodium.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.36 temperature conductivity of from 106 to 108 ohm! m7), ‘The conductivity for a sampling of metals is given in Table 32. This table progresses from monovalent metals through multivalent elements to transition elements. Table 3.2 Electrical conductivity and Lorentz number for some metallic elements T=100K | T=237K Metal Electrical | Electrical Conductivity | Conductivity a o (ohm! m7) | (om m4) Copper 2.9x108 65x10" Gold 1.6x108 50x10" Zinc 62x10" 18x10" Cadetiem 43x10" 15x10" Aluminum 21%108 40x10" Lead 15x10" 5.2x10° Tungsten 98x10" 21x10" Iron 0x107 11x10? 3. For pure specimens of metals, the electrical conductivity (0) and thermal conductivity (K) vary with temperature as follows : (i) For T > 8 (Debye temperature) oT and K = constant so that K/oT is independent of temperature (Wiedemann-Franz law); Gi) For T <4 o«T > andKe T? The relation p« T° is known as the Bloch—Gruneisen T° iaw. In general, at low temperature pet" where n ~ 3 to 5 and therefore p70, asT+0 Solid State Physics and Devices 4. For metals that exhibit the pheno- menon of superconductivity, their resistivity dis- appears at temperatures above 0 K and below the critical temperature 7, for the superconducting phase transition (7, = 4.15 K for Hg). 5. A metal also has a large electronic thermal conductivity K,. Wiedemann and Franz. (1853) observed that a good thermal conductor is also a good electrical conductor, and the consistency of (K,/c) among metals at a given temperature is still referred to as the Wiede- mann-Franz law. At ordinary temperatures, K, is usually relatively temperature-independent, whereas ¢ is likely to display a T~! behavior. (This is the case for the copper data of Fig. 3.4). In 1881 Lorenz noted that (K,/o 7) has a temperature-independent value shared in common by many metals, and the value of L = (K,/oT) is known as the Lorenz number.” 6. An increase of K, and a faster increase of 9, are observed when a metal is cooled well below 2 characteristic temperature (which is related to the characteristic temperature @, for thermal capacity). For copper (Fig. 3.4) and some other monovalent metals, o varies as T~$ in the steepest part of the range. 7. The pressure dependence. Resistivity decreases as pressure increases and for alkali metals eventually passes through a minimum, 8. For impure metals or dilute alloys, Matthiessen’s rule, that the resistivity change due to a small amount of impurity in solid solution is independent of temperature. This rule is ‘more generally interpreted by the equation P= Ppure + Pimp G3) where Pye is temperature dependent and Pip is not. At a. sufficiently low temperature, o reaches a platcau value controlled by impuritics.Motion of Electrons 10" 1 3 Ke6T) voltky? (ohne! ney and HY Ke Whi-K) > id to* § 0 won |'" 4 i 108! 10° § i 6? 0 1 10 100 Ta06” TK) aaa Fig. 3.4 Temperaure dependence ofthe electrical conductivity and the electronic thermal conductirity Ke for highly pure copper and. lattice imperfections. The electrical resistivity (the inverse of conductivity, measured in ohm-meter) is often found to approximate the behaviour of Matthiessen’s Rule, that the contribution of impurities and imperfections to the resistivity is the same at all temperatures: 1 1 1 sine 64 3) Ping * Gpure of course, the contribution of cj, is seen most readily at very low temperatures when Gpure is heading for infinity. 9. Magnetic effects in ferromagnetic metals and alloys also contribute to the electrical resistivity. 10. The free electron gas has a very small electronic specific heat, one which is proportional to the absolute temperature. The electron gas also has a very small paramagnetic susceptibility, which is temperature independent. 37 11. The resistivity of alloys that exhibit order-disorder transition shows pronounced minima corresponding to ordered phases. 12 In the presence of a combination of electrical, magnetic, and thermal gradients, a variety of galvano-thermo-magnetic effects arise, where, for example, a temperature gradient may produce electrical currents of _ potential differences. Under such circumstances the relation between a given field and the resultant phenomena must. be described by a magneto-thermo- conductivity tensor Most of these higher order effects are quite small in a metal and are not casy to measure with preci 13. For very pure samples of single crystals, a variety of orientation dependent phenomena can be observed in the presence of a very large magnetic field. Several of these phenomena are oscillatory functions of magnetic ficld strength. According to the theory of metals, a metallic crystal may be envisaged as the superposition of : 1. A periodic array of positive ion cores. Each core consists of an atomic nucleus and the electrons which make up closed shells. From the Pauli exclusion principle, we know that an electron in a closed-shell orbit around an atom is to be associated exclusively with that atom. 2. A quasi-uniform negative charge density due to all outer electrons. These electrons form an “electron gas,” of which the particles are moving at thermal speeds, By their collective action the clectrons in the gas bind the solid together. The free electron density may range from one electron per atom (210% m~3) upwards, depending on the valence of the atoms,38 ‘This combination is shown in part (a) of Fig. 3.5, and the other parts of the figure shows the conclusions we may draw from this picture. The average charge density of the combination in Fig, 35(a) must be zero for electrical neutrality of the solid. Yet if we look at the local situation, we can surmise that there must be @ periodic sequence of positive humps in the charge density p, superimposed upon an. essentially uniform negative sea of charge from the electron gas. This oscillation of p is sketched in Fig, 3.5(b). Since the closed-shell core of an atom is usually considerably smaller than the interatomic distance, the cores will customarily occupy only a small fraction of the total volume, and the positive excursions of p will be of large magnitude over a small volume. 3.5 Electron Collisions Conduction electrons move throughout the metal, They continually interact with the atoms as well as with cach other. We treat the conduction clectrons as point masses which are moving about in 2 “box” filled with “obstacles” of some kind, The electrons frequently “collide” with these obstacles. Thus, the electrons experience collisions (by an unspecified interaction), or that the electrons are scattered. () The collisions are treated as instantancous scattering events which means the time for the scattering to take place is much shorter than any other times in the problem. It is through these collisions that electrons achieve thermal equilibrium corresponding to the metal temperature T. It is assumed that the electrons emerge from collisions with no memory of their velocities before and are then tandomly directed with velocities appropriate to Z (iii) Between collisions the electrons travel in straight lines obeying Newton’s laws. For example, if an electric field € is applied in the a-direction, then we have ‘Solid State Physics and Devices o. "No © electron gas ° ® ° ® ° o @ Qharge Density ©) electrostatic potential ¥—> distance —» © Fig 35 (a) two-dimensional slice through a metalic crystal, indicating «periodic array of relatively compact Positive ion cores. The “gas” of free electrons is ‘envisaged as being uniformly distributed over all the spece. (b) The variation of charge density with (cris for's tnt amar pos Beene segsones of positive ion cores. Along such a tine, p alternates {from positive to negauive and beck again. If we choose 10 examine p along a poral line which does not intersect any cores, we shell find that p is wiformly negative along the enitre line (c) The periodic ‘fluctuation of electrostatic potential required by the charge density curve of part (b).aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Motion of Electrons wine (62) (} : x10-s 329) Relaxation times calculated from eq. (3.29) and the resistivities of Table 3.3 are displayed in Table 3.4, Note that at room temperatures t is typically 107 to 107s. In considering whether this is a reasonable number, it is more instructive to contemplate the mean free path, 1= vor, where vo is the average electronic speed. The length / measures the average distance an electron travels between collisions. In Drude’s time it was natural to estimate vg from classical equipartion of energy Zmvi=2kT. Using the known electronic mass, we find a vg of order 107 emis at room temperature, and hence a mean free path of 1 to 10 A. Since this distance is comparable to the interatomic spacing, the result is suite consistent with “Drude’s original view ‘that collisions are due to the electron bumping into the large heavy ions. However, the above conclusion is. superseded by the following comments : (i) Sphisticated theories show that this classical estimate of vo is an order of magnitude too small at room temperatures, Gi) Furthermore, as_ the lowest temperatures in Table 3.4. 1 is an order of magnitude larger than at room temperature, while vp is actually temperature-independent. This can raise the low-temperature mean free path to 10° or more angstroms, about a thousand times the spacing between ions Today, by working at sufficiently low temperaturs with carefully prepared samples, mean free paths of the order of centimeters 10° interatomic spacing) can be achieved. This is strong evidence that the electrons do not simply bump off the ions, as Drude supposed. 319 Table 34 Drude relaxation times in units of 1974 s Element | 77K [| 273K | 31K Li 73 0.88 0.61 Na 7 3.2 K 18 4 Rb 4 28 Cs 86 24 Cu a 27 19 Ag 20 40 28 Au 2 30 24 Be 051 0.27 Mg 67 i 0.74 Ca 22 1s Sr 14 0.44 Ba 0.66 0.19 Nb 21 0.42 0:33 Fe 32 0.24 0.14 Zn 24 0.49 034 Ca 24 0.56 Hg O71 Al 65 0.80 0.55 Ga 084 0.17 In 17 038 0.25 Tt 091 0.22 0.15 Sn 11 023 0.15 Pb 057 0.14 0.099 Bi 007 =| 0023 | oo Sb 027 0.055 | 0.036 3.10 Effect of Impurities on p In a pure metal, we expect that p> 0 as T+ 0. However, experimental observations show that p for metals decreases to a small but finite value and remains practicaly constant and indepeadent of temperature in the range 10 K to OK (Fig. 3.7). The resistivity at O K is called the residual resistivity of the metal. This resistivity may be attributed to (i) presence of impurity atoms in small concentrations, (ii) geometrical imperfections like point defects, grain boundaries and stacking faults in the metal.3.20 Solid State Physics and Devices x10 P(TH/p (290 K) —> vow ew © 6 10 14 18 2 TK) (@) pvp (290k) —> 0 20 40 60 80. 100 TK (b) Fig, 37 The normalized resstviy p(T)/p(290 K) versus T {for Na is (a) the low-temperature regian, and (b) at higher temperatures. p(290) = 2.10x10-*Q-m. ‘We may recall that the electron undergoes a collision only because the lattice is not perfectly regular. It is convenicnt to group the deviations from a perfect lattice into two classes : (a) Lattice vibrations (phonons ) of the ions around their equilibrium position due to thermal excitations of the ions. (b) Alll static imperfections, such as foreign impurities or crystal defects. Of this latter group ‘we shall take foreign impurities as an example. To arrive at a result for p : (i) We assign relaxation time 1}, t2 ~ to each type of impurity or imperfection in the metal. Gi) The probsbilities of electrons being scattered by phonons and by impurities are additive, since .these two mechanisms are assumed to act independently. (ii) Since the scattering probability is inversely proportional to the relaxation time, the reciprocal of the effective relaxation time is the sum of the reciprocals of the individual relaxation times. ‘We may, therefore, write alti 7 aa tr where the first term on the right is due to phonons and the second is due to impurities = (3.30) ( tpn is expected to depend on T. (i) 7, depends on impurities, but not on 7. Using the ~— general = formula p=l/o = m/ne*r, we can write mi,m =p, te,()=—52+5 4 31) Pet PO) aT Toh (a) Matthiessen rule : The above equation tells us that the resistivity in the presence of several distinct scattering mechanisms is simply the sum of the resistivities one would have if each alone were present. This is known as Matthiessen’ rule. In other words, this rule states that p can be split into two parts, one of which is independent of 7. From eq. (3.31), we note that p has split into two terms : (@ a terms p, due to scattering by impurities (which is independent of 7), called the residual resistivity, (i) another terms —pp,(T) due to scattering by phonons; hence it is temperature dependent, and is called the ideal resistivity, in that it is the resistivity of a pure specimen.Motion of Electrons (b) Temperature dependence (@ At very low T, scattering by phonons is negligible, because the amplitudes of oscillations are very small, in that region tpn? 2 Pph 20 and hence P =p, = constant This is in agreement with Fig. 3.7. (ii) As T increases, scattering by phonons becomes more effective, and pp,(7) increases; this is why p increases. (iii) Whea 7 becomes sufficiently large, scattering by phonons dominates and P = Ppx(T). (iv) In the high temperature region, Ppn(7) increases linearly with T. This is again in agreement with experiment, as shown in Fig. 37. Finally, we expect that p, should increase with impurity concentrations, and indeed it could be shown that for small concentrations Pe m where 1, is the impurity concentration. ‘We also remark that, for small n; Pon > > P, (except at very low T) 3.11 The Allowed Energies Let us consider an electron moving in a cubical box with side L. The electron will satisfy the schrodinger’s ates vp ast where Vis the potential energy of the electron. Since the electron is free, V = 0 and hence the Schrodinger’s equation assumes the form : 7 E-VYyp=0 (332) vy + ont yeo (333) Writing the expanded operator V* we have form of Laplacian 3.21 Pe, Py, Py , me wet a at We assume the solution in which the variables can be separated : ot Ey=0 334) v2) =X@) YY) Z@) (3.35) Let the electron move along x-axis. Then ¥@) = X@) (3.36) and the corresponding Schrodingcr equation will be 2x. amE, Bx Ul an 2x ImE, ox of fsx 3.31 7 7 337) Let the solution of this equation be X(x) =Csin (ck) (3.38) with the condition that X=0 at x=0 and x=L Csin (eg) = 0 or yh = nga 3.39) ne ne int o ass ann 12° Normalisation of X is expressed by L SP) aca 0 L CS sin? (ek) deo 0 L Sha = cos 2ck,) de = 1 0322 Solid State Physics and Devices or cas Therefore, the x-component of solution is v2 oe ve X@)= () sin Fn = 1,2,3, Likewise y- and 2- components of the solution will be v2 2)" 27% w= (2) sin 8 ny =3,2,35 and wm 2)? x20, 20 (2) sin Fn = 123° with Aly? Pia mL? and natn? Therefore Mi yen (2) Man a ae 3.40) ‘We can take the total energy E as the sum of its components along x, y and z direction : E=E, +E, +E por tanta or E= Ck t+ny+m) BAI) ml? From eq, (339) and similar equations for y- and z- directions, BE Dest G42) Shy Pt ky PEK, R By My» Ty Eitpityt Also from eq. (3.42) Seltgen= B+R+2= |b? 64) Putting eq. (3.43) in eq. (3.41), we have oni nm 2 ear el -Ez wt 12 +np+ndy) 3.44) where my, 1, and n, all have to be positive non-zero integers. The least values of te my. mz = (1, 1, 1) vill refer to the ground state quantum level : 3 Ean= aM * ae Degenerate siates. Each of aumbers My My n, is called a quantum number and it can take only positive integral values. quantum state, while several states having the same energy are said to be degenerate. For cxample, if n quantum states have the samc energy then itis referred to as n-fold degenerate states. The first excited state is defined by Ime This has a degeneracy of 3 as 214, 121, 112 are all equivalent, Again ng + ny +nz = 27 vill be a 4-fold degenerate state corresponding to 333), (511), (151), (115) My Nyy Mz =aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Solid State Physics and Devi Next higher level is ny 2 kp : : z # for which the energy is Elis tay ye 2 F-Pt P+ ea, ot Sma 8 ma 3h 10 3h? __3x(6.626x1074)? De" Bima 8x9.11x10 "x 0.01)? 1.8x1073 = 18x10°8 J = el 16x10" Example 11. Show that the wavelength = 113x107 “ev associated with an electron having an energy Example 9. Express electrical cond- ““#! to the Fermi energy is given by uctivity in terms of Ep. nex Solution : 9 =—— m fi = Earn Example 12. Prove that the compressibility of the free electron gas is given by 3 p= Example 10. Show that the average DE, kinetic energy of the electrons in a metal is ap’ R= 4 k Solution: f= [v3 - For free electron gas Solution : By definition kinetic energy of all electrons Be in Fermi sphere ** “volume of the Fermi sphere Ke ay SS (anah bm ¢ ) ke | f anitak aaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Motion of Electrons 347 32, 33. Calculate the energy of a sixfold _ degenerate state for free electrons in a solid cube of side Lom. [528x107 ev] If g(E) were a decreasing function of E near the top of a band, prove that Ep would increase slightly with 7 according to the relation - © eT) (ag EMD = EO) Fey (3), “t, 3.18. Review Questions 1. 10. 11. }. State the Matthicsscs . Derive State the assumptions of free electron model of a metal. Discuss the achievements and failures of the model. the scattering electron that contribute to the resistivity of a metal . State the salient features of the Drude model and derive the formula for electrical conductivity on its basis. rule for total resistivity and differentiate between the residual and ideal resistivity. . Derive a formula for energy spectrum of electrons moving in a box. Explain the meaning of a quantum state. . Define Fermi energy and explain how is it related with chemical potential. . Define Fermi wave vector and derive a formula for it in terms of concentration of conduction electrons. the formula experessing the variation of Fermi energy with temperative. . Derive the formulae. J=of e and on tt Define the density of states and derive the formula Explain the meaning of Fermi-Dirac statistics and discuss its low temperature expansion. 3.19. At Using the density of occupied electron states derive the formula for electron heat capacity per mole in a metal : CeayT peNe QEp Here N is the Avogadro’s number Objective Questions where . In. free electron theory of metals, (1) kinetic energy of electrons is zero (2) potential energy of electrons is zero (3) both kinetic and potential energy are 7er0 (4) total energy is independent of temperature . The classical expression for the electrical conduct () mner (3) ne? z/m y of a metal is (2) (net/n) (4) ne? /m . The relaxation time for electrons in metals is of the order of (a) 1078s (2) 107 5 (3) 10725 (4) 1074s low temperatures, the electrical resistivity of metals varies as ar ar @r @r . The density of allowed states between E and E + dE is proportional to (1) a7? @ev (3) £32 (4) E . At temperature above 0 K, for E<> 8 31. The electronic specific heat at constant ip volume is proportional to Q) Pon e5 (yr (2) T? @ or? rs OQ) Pp 32, If my ny mz= (123), the number of 1? degenerate states for the electron in a (3) Ppa & (3) cubical box is, (i (2)3 6)? 4 (6 4) Pon (7) 33. If mis the dimensions of a crystal, then . We define the electron density parameter aE) ES 7, a8 the radius of sphere occupied by an electron. Then whe aie: ()t=Q@-1) @x=@-2 2) kp R @- 2 2225 Answer-Sheet ooo4 Energy Bands 4.1 Introduction ‘To begin with Iet us remind oursclves of the restrictions which govern the energy levels in an isolated atom. These can be summarized by the following statements : 1. The properties of the electrons in an atom arc determined by four parameters, or quantum numbers, which are denoted m, 1, m and m,.n is the principal quantum number, | takes integer values from 0 to (n - 1), my is allowed integer values from -1 to +1 and m, canbe ~4or +3. 2. The electrons are only allowed to occupy certain discrete energy levels, which we label using the quantum numbers n and J (since the energy of an electron does not usually depend on my or ms. The letters s. p. d and f are used to denote the states /= 0, 1, 2 and 3 respectively. 3. The occupancy of these levels is determined by the exclusion principle which states that cach electron must possess a unique set of quantum numbers. If we apply these rules to a sodium atom, which has 11 electrons, we find that two electrons occupy cach of the 1s and 25 levels, six occupy the 2p level, leaving one electron in the 3s level, as shown in Fig. 4.1(a). However, in a solid the atoms are not isolated. In fact they are in close proximity to a large number of other atoms, What happens in this case ? Let us first of all consider the situatioi with just two sodium atoms. When thc atoms are far part they behave like isolated atoms, but ‘TD isn" detent my 1 + TITEL, 3s fant vac 1s hand eapacty 2N, 3 ZILIA. fi tgn eopaects (a) (b) (a) The occupation of the electron energy levels in tan isolated sodium atom. (b) The occupation of the energy bands ino sodium crystal. as we move them closer together, the outer (ie. 3s) electrons begin to interact, which affects the energies of these electrons. If we plot the positions of the energy levels as a function of the separation between the atoms, we obtain a picture as in Fig. 4.2. At the equilibrium scparation, ag, we, therefore, have two states, ‘one of which is at a lower energy than the 3s state in an isolated atom, and one which is at a higher energy. We have to be very careful how we interpret this result. The cnergy levels can no longer be considered to belong to a specific atom, rather they are a property of the pair of atoms considered as a whole. Consequently, this picture does not imply that the 3¢ electron attached to one atom has a higher energy than that on the other‘atom. In fact, since there are two allowed values of m, for each state, both of the electrons can occupy the lowest energy level (Fig. 4.2b). : If we cxtend this same argument to a system of N’ sodium atoms, we should get N discrete 3s energy levels, as shown in Fig. 43(a). However, in a macroscopic crystal the number Fig 4aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Energy Bands p——— Ie oe Separation Fig. 4.4 The formation of 1s, 25, 2p and 3s energy bands in sodium. The bands become progressively narrower Jor elecrans which are more tightly Bound 10 the Parent nucleus which shows the splitting for solid hydrogen, and Figure 4.6, for solid sodium. Only the envelopes of the energy levels are shown because the levels themselves are too closely spaced. The several bands in metals like sodium may overlap, but in many materials they do not. The spaces between the bands are called forbidden energy gaps, since electrons in the solid may not have these energies. Some features of the band of levels are quickly apparent (i) The binding energy of the solid as a whole must be caused by shifts in the levels of the electrons, just as for chemical bonding. Thus, as the solid is formed, the energy levels, must shift downward on the average. (ii) The electrons which should be affected most are the outermost, or valence, electrons, since they are closest to the neighbouring ions. (iii) The equilibrium lattice spacing must be at an cncrgy minimum, with the energy levels rising at a distance of closer separation. These three requirements are taken into account in the sketch in Fig. 4.5 for a solid 43 whose lattice spacing varies from large to small values. Energy ——> Interatomic seperation ——> Fig. 45 The energy levels for solid hydrogen as a function of interatomic distance. The atomic levels can be considered to have broadened into bands of allowed energy. Note: the higher levels spit fist. Let g, be the degeneracy of the n-th energy level of a single atom. Then the entire system of N atoms has degencracy Ngy at this level. (iv) As the atoms are brought together, as in Fig. 4.6, the states of the original system must deform continuously; the number of states in the solid must be the same as the number in the original separated atoms, Thus, since the “crystal” with large spacing has Ng, states, the band at the distance of separation of the real crystal must also contain Ng, states. As s band therefore holds 2 electrons per atom, a p band 6, ad band 10, ete. (v) Figure 4.6 gives a picture of energy bands in solid Na, in which the charge clouds of electrons in the deeper atomic levels hardly44 Solid State Physics and Devices Observed value of a, y= 3.674 0 3 10 15 Interatomic spacing (A) —> Fig. 4.6 Energy bands in sodium meta. overlap at all and so lead to very narrow bands with 4 levels packed into very narrow energy range. The noteworthy feature of the bands structure of Na is that the 3p band docs overlap the top of the 3r band. General features of the bands levels are : 1. The states having the largest principal quantum number n split first; then the state with the second largest , and so on. The above behaviour of splitting could be observed from Fig. 45. If a, is the maximum distance between two neighbouring atoms at which the splitting occurs, then ay> 02 >a, 2, For the same value of 1, the states having largest angular quantum numbers / split first, then the state with the second largest /, and so on. If gy is the maximum distance between ‘two neighbouring atoms at which the splitting ‘occurs, then 2 > 2n > no where 1=0 is s-state, /=1 the p state and 1= 2 the d-state. 3. The degeneracy factor associated with a given energy level in the energy band would be the same as that of the corresponding level in the free atom, Accordingly, there would be a 1s band corresponding to the Is level, a 2s band corresponding to the 2s level, and so on. ‘The number of levels in cach band would be N. 4. The filling of the bands follows a simple rule : state of lowest energy are filled first, the next lowest, and so on. Finally, all the electrons have been accommodated. The energy of the highest filled state is called the Fermi level, or the Fermi energy, E,. The magnitude of Ey depends on the number of electrons per unit, volume in the solid, as the latter determines how many’ electrons must go into the bands. At OK, all states up to E, are full, and all states above E, are empty. At higher temperatures, the random thermal energy will empty a few states below Ey by clevating a few electrons to yet higher energy states. No transitions to states below E, will occur, since they are full. Thus, an electron cannot change its state unless enough energy is provided to take it above E;. 5. Valence electrons at or above the Fermi level, of course, arc in one of a number of energy states for the crystal, However, the' Energy Bands Dist i le 9. Distance of separation of muclei Valence ! level i Filled level Distance = a —__ | Filledlevel Fig. 4.7 Energy of elecirons in several levels as a function of the distance between atoms in a solid The ‘minimum in energy occurs at seperation a, the ‘atomic diameter allowable energy states in any band are so great in number and so closely spaced in magnitude that they form, essentially, a continuous range of allowed energy, called the energy band, A Fermi level electron, moving through the crystal, may therefore be regarded as a free election. 6. If the electron moves with a given velocity it will possess an associated kinetic energy and wave vector, |k |. Electrons in a solid will, however, be subjected to certain restrictions because of the existence of the forbidden energy gaps between the bands, as would be seen later. The continuity of the parabolic relationship between kinetic energy (KE.) and &, shown in Fig. 48, must be disrupted at the energy gap because, at the gap, a large discontinuous energy is required for the electron to pass from one state to the next. The preceding argument is completely empirical except for the ideas from quantum KE Fig. 4.8 Parabolic relation besween KE. and k of the free electron, mechanics. To understand the crystal band in any real physical sense, we must understand in detail at least these three things : () the basis for the forces which attract the atoms together. i) the forces of repulsion when they are too close, and the degree of splitting caused by the interactions. (iii) The answer to the first point is complex because it is different for different crystals, The second point is morc easily understood because it is the same for all solids. Atoms repel one another when they are too close, basically because the electrons of a given state must have a fairly well-defined volume to themselves. The Pauli exclusion principle states that the same wave functions from different atoms cannot occupy cxactly the same space, for then they would be exactly the same state. When the atoms approach so closely as to compress the wave functions into smaller and smaller regions so that extensive overlap docs occur and the exclusion principle cannot be satisfied, then the energy rises because of the uncertainty principle, As the volume of the wave function decreases, Ax decreases. Then Ap and hence p must increase, thereby increasing the kinetic energy. The total energy thus rises as the atoms are pushed too close together. This is equivalent to a repulsive force. The third point forms the subject of the most fundamental theorem concerning electrons in solids, namely, that the electrons in @ band of energy levels are mobile and are not localized onaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Solid State Physics and Devices 416 This relation is called the Kronig Penney equation, As an cxample, we have represented in Fig. 4.18, the left hand side of this equation as function of ea for the value of P= 31/2 Observation of Fig, 4.18 leads tothe following interesting conclusions : (i) Since & is proportional to E [eq. (4.19), the abscissa is a measure for energy. (ii) lt is important to realise that the right-hand side of eq. (4.28) can accept only values between —1 and +1, as indicated by the horizontal lines in Fig. 4.18, Therefore, the condition (4.28) can be satisfied only for values of ca for which the lefl-hand side lies between +1. Hence the limits of the allowed values of aa are given by coska = +1 or ka=na— n=4,23-++ (4.29) Obviously eq. (4.29) defines the boundaries of the Brillouin zones of a one-dimensional lattice. (iii) Energy bands. Equation (4.29) gives the relation between energy and a : Re? E Fig 418 The left hand side of eg (428) for P= 3n/2, plotted as function of wa. The allowed regions are heavily drawn, This equation shows that the allowed values of E exactly follow the limiting values of aa. These values are shown by heavily drawn discrete lines in Fig. 4.18. Hence, these constitute a spectrum of possible energy values of an electron propagating in a lattice, so that the continuous portions are separated by finite intervals. Energy continuous regions separaicd by finite intervals are called energy bands. The finite intervals are called energy gaps. (iv) The dotied fines along abscissa of Fig. 418 represent those solutions which have no physical meaning. (v) The width of the allowed energy bands increases with increasing values of aa, ic. with increasing energy; this is a consequence of the fact that the first term of eq. (4.28) decreases on the average with increasing aa. (vi) The width of a particular allowed band decreases with increasing P, ic. with increasing “binding energy” of the electrons. (vii) In the extreme case for which P<, the allowed regions become infinitely narrow and the energy spectrum becomes a line spectrum. In that case, eq. (4.28) has only solutions if sinaa = 0 ic. eatne (1= 1,23.) According to this result and eq. (4.19), the energy spectrum is given by i mi lim (E,) = |=75] 1? 4, oe (23) 430) which one recognizes as the energy levels of a particle in a constant potential box of atomic dimensions. Physically, this could be expected because for large P, tunnelling through the barriers becomes improbable. (vill) Energy gaps. E and k are connected to @ respectively by eq. (4.28) and (4.19). Thus, the results of Fig, 4.18 could be used to get £~k dispersion curve, shown in Fig. 4.19 as follows : The discontinuities in the E versus k curve occur for k (1 =1,2,309 (431)| Energy Bands 417 These k-values define the boundaries of . x 2 the first, second, etc, Brillouin zones. Thus, another part extending between - 7 to ~ 2 (a) the first zone extends from Figure 419 shows the _ periodic x x -= 0 +2 a a (b) Similarly, the second zone consists of two parts ; a a one part extending from 2 ee \ \ \ First Brillouin zone ‘Second Brillouin zone- dependence of E on k and demonstrates the existence of the forbidden energy gaps. Electron velocity. Recalling that the group velocity is defined by vy=do/dk and that E =Tiw, we define the velocity of the electron, when it is represented by a wave packet Energy bands Fig. 4.19 E-k relation for P = 31/2 of the Kronig-Penney model y . (b) Fig. 4.20. Velocity and acceleration as a function of kin the first Brillouin zone of a linear lattice418 Solid State Physics and Devices centered about the energy and wave number kas 1 vag (432) Comparing this equation with the slope of the curve in Fig. 4.19, we conclude that the velocity of the electron varies within the first Brillouin zone, as shown in Fig. 4.20(a), with similar results for the other zones, The velocity is zero both at the bottom and at the top of the band. Al intermediate regions in the band, it is very close to the free-electron velocity, 7 k/nt. Given that F is the external force on the electron, the work done by this force on the electron’ during the time a is Fy dt. This results in a change dé in the energy of the electron. Thus we may write Fvdt = dE = (dE/dk)dk. Introducing the expression for v given by Eq. (432a), we obtain (4.326) This can be taken as the equation of motion of the clectron in the lattice. Comparing. this equation with the classical equation of motion, F = dp/dl, we conclude that 71k plays the same role for clectron motion through a lattice under an external force as the momentum p plays for electron motion in free space under the same force. For that reason 71k is sometimes called the lattice momentum of the electron. To describe the motion of an electron (when it is represented by a wave packet) as it moves through the lattice under an external force, we must use both Eqs. (432a) and (432b). Equation (4.32b) gives the effect of the external force on the electron’s wave number k and Eq. (432a) gives the velocity of the electron resulting from that force plus the clectron’s interaction with the lattice. Let us first consider an clectron in the first” Brillouin zone and suppose, for example, that the electron initially has k = 0; therefore its velocity is also zero. When the force F is applied, k increases according to Eq. (4.32b), and from 4.20 we sec that ¥ also increases; that is, the electron accelerates, But when k reaches a certain value close to 2¢/a, the velocity begins to decrease (ic, the electron decelerates even if the force remains the same). This effect, of course, is due to the interaction of the electron with the lattice. When k=2/a, the velocity becomes zero but the wave packet representing the electron suffers a Bragg reflection in the lattice, k becomes —2/a, and the propagation is, in the opposite direction. Obviously the velocity now becomes negative. If the applied force remains the same, eq. (4.32b) tells us that & continues changing in the same direction, and therefore, since it is negative, decreases in absolute value. Figure 4.20(a) indicates that the velocity first increases in a direction opposite to the external force, but very soon its magnitude mE = oF Wig. 421 Electronic motion in first and second. Brillouin zones in a linear latice. Under the action of an ‘extemal faree, the wave number and the energy of the electron vary within a zone, as shown by the arom. begins to decrease, since the force is opposed to the velocity; ie., the electron decelerates, until eventually the velocity again becomes zero when &=0. From then on the cycle repeats itself. The acceleration of the electron during the whole cycle is shown in Fig. 4.20(b). Representing the state of the electron by a dot in an E versus k diagram (Fig. 421), we can illustrate the process by moving the dot, as indicated by the arrows in the figure. If the electron starts al O, its representative point moves from © to A, then jumps to B and back to 0. Let us now consider an electron in the second Brillouin zone, which covers the ranges ~2n/asks-n/a and 2/ask< n/a. If the electron initially has a k-valve slightly above x/a, the applied force increases the value of kEnengy Bands until k= 21/a, A Bragg reflection suddenly reverses the momentum and k jumps to the value —27/a. If the force continues to act on the electron, & continues to increase ( actually decreasing in absolute value) until it reaches the value ~2c/a, when another Bragg reflection takes. place. and the momentum is again reversed, k changing to the value 2/a. From then on the process continues in a cyclic manner. Hence the electron describes the cycle C+D+E~>F-+C (Fig. 4.21) The variation of the velocity and the acceleration of the electron in the second Brillouin zone is shown in Fig. 4.22. Similar logic applies to the other Brillouin zones. ‘An important conclusion to be derived (b) Fig, 422. Velocity and acceleration as w function k in the second Brillouin zone of a linear lavice. from the above kinematical description is that an external force cannot remove an electron from a Brillouin zone; therefore the electron 419 remains in the same energy band unless cnough energy is gained in a single process (as by absorbing a photon) to Moss the energy gap and pass to the next zone, (There is, however, a certain probability that an clectron under an applied force will bridge the gap at k = +nz/a; this is called the Zener effect.). 4.5 Effe e Mass of Electron (a) Formula, Let us first derive the equation of motion of an clectron in an cnergy band. Suppose that the wave packet is made up of wave functions near a particular wave vector k, The frequency associated with a wave function of energy E is : we a The group velocity ¥, (433) The effects of the crystal on the electron motion are contained in the dispersion relation E(&). The work 0 done on the electron by the electric field & in the interval dr is OE = — cb vgct (434) Using eq. (4.33) we observe that sE= (2| dk =H dk (435) On comparing eq. (434) with (4.35), we have _ (ee sen (2) dk o = Bene (436) Equation (4.36) can be written in terms of the external force F as dk a F (437) This is an important relation : (i) In a ‘crystal %dk/dt is equal to the external force on the electron,aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.4.22 Solid State Physics and Devices not near the edge of a band where the E versus & curve is parabolic. In most conductors, m= as the band is only partially filled. In semiconductors, insulators, and certain conductors (e.g, bismuth) where full or almost full valence bands are involved, the effective mass differs from m. ery Molecirane Region 4 | (7 Region 3 7 ALL, ‘Classical \ benaviony Fig 423 Effective mass. The use of eg. (4.40) in the E verms relation As shown in Fig. 423 : (a) region 1, the parabolic relation is good, and behaviour is therefore classical. (b) In region 2, the curvature is negative and therefore the effective mass is negative. (©) In region 4, the curvature is positive but larger than that of region 1; therefore, the effective mass is smaller, according to eq. (4.40). To appropriately describe the effective mass, the relationship between E and k near the top of, say, the first band needs to be explained in more detail. In quantum mechanics the velocity of a particle is given by the group velocity dev/dk = (1/N) dE/dk of the wave. This means that the electron velocity is proportional to the slope of the Ek curve. We therefore sce that starting at 0 the velocity first increases with E. This is duc to the ordinary increase in kinetic energy. Near the top of the band, however, where the slope of the curve decreases and becomes horizontal, the velocity decreases inspite of the fact that the energy is still increasing. The explanation for _ this behaviour is that for an electron in a periodic lattice E has both kinetic and potential energy components, whercas in the free electron model =m Fig. 424 The mass and relaied quantities of an electron in acnsatEnergy Bands components, whereas in the free electron model it is entirely kinetic energy. So although near the top of the band the velocity (and kinetic energy) of the electron decreases, its potential energy still increases, until at the very top of the band, £ is entirely potential energy. Figure 4.24(a) represents the E-k curve of an clectron under a periodic potential in the region 0 $k <2/a. dE/dk, PE/ak? and m” are respectively plotted in Fig. 424 (b), (c) and (d). For small values of k, E—k curve resembles the free electron curve for which m =m. (i) At A, gradient of the curve and hence group velocity v, is maximum. Here @E/dk* = 0 and hence m” becomes infinite. Gi) In the interval between A and B, @E/dke is negative and hence m* is also negative, i) When the value of k approaches sa/a, v,>0 and m*=—m. This physically means that when k is exactly equal to x/a, then as a result of an accelerating field Bragg reflection occurs. Thus, under the influence of an ficld applied along a given direction an electron, as a consequence of its interactions with the crystal, attains a momentum in the opposite direction and behaves as if it has a negative mass. This startling behaviour of the electron. is really due to the diffraction effects which dominate the behaviour of the electron as the forbidden energy is approached. ~(iv) If the energy in a band depends only slightly on k, then the effective mass will be very large. That is, m'/m>>1 because @E/di is very small. (d) Demonstration of negative mass : In Fig. 4.25 (a), an electron beam is incident on the crystal. Its potential (or energy) and associated momentum is either too large or too small compared to the one required for diffraction. Hence the electron beam, not Bragg reflected, is transmitted through the crystal. A grid placed infront of the sample crystal 4.23 Beam Brogg-reflected 1 1 1 1 1 1 i ' ' ' ' v Fig 425 Explanation of negarive effective masses which ‘occur near a Brillouin zane boundary produces appropriate positive or potential for the diffraction, In Fig. 425(b), a small accelerating or decelerating potential applied on the grid leads to the required condition for Bragg reflection Then the electron beam will be reflected by appropriate set of crystal planes, Thus, a small change in energy of electron beam would induce a large change in the momentum associated with the small effective mass, If the grid potential accelerates an clectron beam of negative small energy, the effective mass is negative (Because the acceleration occurs in a direction opposite to change in momentum). If the grid potential decelerates the electron beam of high energy, then the effective mass is positive.4.248 Solid State Physics and Devices 4.6 NFE Model In Nearly Free Electron (NEE.) Model, the crystal potential / (7) is assumed to be very weak compared with the clectron kinetic energy. This does not correspond to the motion of valence variations of potential energy are generally comparable to the kinetic energy. Even so, the method is worth describing as it is one of the limiting cases (the other is tight-binding); and is the basis of more sophisticated approaches. The approximation will, apply to the motion of a beam of fast electrons ejected into a crystal. In the NFE model, unperturbed electron wave functions are assumed to be given by plane waves. The periodic crystal potential is regarded as a small perturbation. In this model, a nearly parabolic E vs. k cruve is obtained with energy gaps at the boundaries of the Brillouin zones where Bragg reflections of the electron wave functions from the crystal planes strongly affect the electronic states. The resulting E vs k curve can be “folded” back into the first Brillouin zone (i.e. the reduced zone scheme). This simple picture provides a good qualitative description of the band structure, In quantitative analysis, however, there are many problems related to very large values of the crystal Coulomb potential near the centres of the ion cores and its rapid variation in this region. This requires the representation of wave functions as sums of plane waves. A very large number of plane waves must be used in order to approach the correct solution. To some extent, the method can be improved if care is taken in the selection of plane waves of proper symmetry properties. electrons in any crystal as the The band structure of a crystal can often be explained by the NFE model for which the band electrons are treated as perturbed only weakly by the periodic potential of the ion cores. This model answers almost all the qualitative questions about the behaviour of clectrons im metals. ‘The Bragg reflection is a characteristic feature of wave propagation in crystals. Bragg reflection of electron waves in crystals is the cause of energy gaps. In one-dimensional crystal, Bragg reflections occur at =tna/a ‘Thus, the first reflections and the first energy gap occur at k = +41/a. Other energy gaps occur for other values of the integer n. Consider a one-dimensional metal. The positive ions are separated by a distance a, the lattice parameter. In the free electron model, the cnexgy E of a conduction electron with wave vector k is given by re pant Tm (4.50) The plot of E as a function of k is a parabola, as shown in Fig. 4.26. —>n ok Fig. 426 The energy of a free electron. The interaction with the positive ions affects this relationship between E and k. It is easiest lo see this at the boundary of the first Brillouin zone, k= 2%2/a. There, two free-electron wave functions areaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Energy Bands 427 lattice. It is the locus of points in k-space that describes Bragg scattering. From the Bragg’s law 2k-G + G? = 0, we observe that for a given reciprocal lattice vector, its right bisector is the reflecting planc. One can extend the procedure for finding the reflecting planes corresponding to reciprocal lattice vectors connecting the reciprocal lattice point O (origin) with its neighbours in the reciprocal space. The volume bounded by these planes is referred to as the geometical definition of the first Brillouin zone —Figure 4.29 of reciprocal space for a ique latice, showing the lines bisecting the midpoinis of some reciprocal lanice vectors from 0. The six shortest of these vectors can be bisected to produce the first Brillouin zone about 0. Fig 429 A portion ponion gives a portion of the reciprocal space for a 2-dimensional oblique lattice showing the lines bisecting some reciprocal lattice vectors from 0. The six shortest of these vectors can be right bisected to produce the first BZ (Brillouin zone) centred on the reciprocal lattice point 0. It will be recalled that in the construction for diffraction the incident wave vector terminated on the perpendicular bisector of a pair of reciprocal lattice points and hence any wave vector which at the Brillouin zone is one which would be diffracted. The relevant diffraction condition due to Bragg is wG+G=0 where G is a reciprocal lattice vector defined by G = mb, + rgb, +1363 where the primitive translation vectors (b1, 62,53) of the reciprocal laitice are defined by ab; = 2 dy where dy is the Kronecker delta ; a,b) = aby = grb = 2 and 4y7by = "by = yb; = ay-by = a3°b, = 43'by = 0 Using ay-b;=ay:b,=0, we find that b; is perpendicular to both az and a3 . Thus, by = C(@ Xa) Putting it in a)-b, = 27, we find that c= 2s __ x. _ ay"(@,X as) Ve where V- = ay - (a2 Xa) is the volume of the primitive unit cell of the direct lattice. Thus, Fa x a3) Likewise ma by =F (asx a) Qn =F Xm) BCC Lattice : a, =3¢ f+ f+ ® a5 (i f+ ® ane =F (P+ 7- B 3 Ve =a -(ay x a3) = >aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Energy Bands Solving for k, (458) ‘Thus, the number of electron states with encrey less than E is the same as the number of states with wave number less than the value of k given by eq. (458). These states are contained within a sphere of radius k in k-space. The volume of this sphere is given by 32 Wy gak =a ml Number of Electron States in a Band. Let us next determine the number of eleciron states in an energy band. Each energy band contains states throughout the entire first Brillouin zone. We found for the free-clectron model that the density of .states in Kkspace is given by 2(k) = 2V/(22)*. This relation still holds here since the Bloch functions must satisfy the same periodic boundary conditions given in cq. (454). We must be careful to note, however, that (le) now means the density of states in k-space for a single energy band. To find the number of states, we must first find the volume Vj of the first Brillouin zone. This is simply the volume of a primitive unit cell of the reciprocal lattice. The answer is simple, If V, is the volume of a primitive unit cell of the direct lattice, then the volume Vy of a primitive unit cell of the reciprocal lattice is given by 4 pad es (459) Cd) 4-G a (460) This holds true for any Bravais lattice. We will demonstrate the validity of eq. (4.60) for the se lattice. If the lattice parameter of the direct lattice is a, then the lattice parameter of the reciprocal lattice is 2x/a. Thus, V, =a, and VY = (21/a)3, in agreement with eq. (4.60), 431 The number of states in the first Brillouin zone can now be calculated : 3 Vest) = Ge [axs| “ 2% Ye The ratio V/V. is the total volume of the crystal divided by the volume of each primitive unit cell, This gives us simply the number N_ of primitive unit cells in the crystal. Therefore, the number of states in the first Brillouin zone is equal to 2N,, twice the number of primitive unit cells in the crystal, This is the number of states in a single energy band, States are filled with conduction electrons just like in the free-electron model. Using Pauli’s exclusion principle, we fill the states starting with those of lowest energy until we have put in all the electrons. For example, sodium metal (Na) has a bec lattice with one atom in cach primitive unit cell. Since cach atom contributes one conduction electron, the number of these electrons must be equal to N., the number of primitive unit cells (which is the number of atoms in this case). Since each energy band contains 2N, states, there are enough conduction clectrons to fill the first energy band half way. (4.61) 4.9 Fermi Surface ‘The Fermi surface (FS) concept enables to visualise the relative fullness or occupation of the allowed empty lattice bands geometrically in k-space and thus helps us in the theoretical determination of the electronic properties of a solid—may it be a metal, semiconductor or an insulator. In fact, the purpose of the FS construction is to know about the details of the motion of an itinerant electron in three dimensions. If the crystal is composed of N atoms then each Brillouin zone contains N allowed values of k. As an electron has two spin cigenstatcs, cach zone may then contain 2V. electrons, which is equivalent to saying thataa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Energy Bands 2. The gradient of the energy function (if non-degenerate) must vanish normal to the zone ‘boundary, which makes the constant energy surfaces always come out at right angles to the boundary. This renders the surfaces quite different from spheres near the corners of the zone. 3. The free-electron energy contours are perturbed by the lattice potential such that depression of the energy just inside a zone boundary moves the energy contours out towards the boundary, and similarly the increased Fig. 4.35 (a) Perturbation (dotted curve) of a free electron ‘energy contour by a lattice potential (b) Occupied states ct T=0 for a divalent metal (in ‘two-dimensions). The FS of a ‘current carying state is shown dovted 4.35 energy just outside a zone boundary moves a constant energy contour in towards the zone boundary. Thus, zone boundaries ‘attract’ energy contours. Accordingly, the gradient of E in k-space is parallel to the boundary and contour of E is perpendicular to the boundary. Figure 435(a) shows the perturbed energy contours. For a two-dimensional lattice of Nx N atoms, we have N? states in the first zone, cach of which can be occupied by two electrons. The occupied states for T=0 for a lattice of divalent atoms are, therefore, determined by two conditions : (i) We must fill the area in k-space equal to that of the first zone; Gi) We must fil all levels below Ep. The solution to the problem is shown in Fig. 435(b), where the occupied area of k-space is shaded. Although the FS is very different from a free electron sphere, the essential point is that it does have some free area, and can be slightly displaced as indicated by dotted lines to give a current carrying state for assigning metallic properties. The FS of Fig. 4.35(b) consists of a number of disconnected pieces, broken up by zone boundaries. It is convenient to bring these Fig. 4.36 Occupied states of Fig 4.35 (b) remapped by ‘moving portions of the FS by (21/a) in kx Or hyaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Enemy Bands these electrons is approximately cqual to n(va/vs) Where n is the density of all the conduction electrons. Now the current density is equal to the density of conducting electrons (ava/vg here) times the charge ¢ of an electron times the velocity of the conducting electrons (vs here). This gives us inva r= |4) ev =nevg (4.70) YE which is the same as in the classical case. In the quantum model, the meaning of vq is the displacement of the Fermi surface, as given by eq. (4.69), and not an average “drift velocity,” which was the original meaning of vg in the classical model. This displacement of the Fermi surface is governed by how far the electrons near the Fermi surface can be accelerated by the electric field before they are scattered by some kind of collision. This acceleration pushes these electrons into new energy states outside the original Ferni swface. All the other electrons inside the Fermi surface are also accelerated by the field, but they cannot go any further than the electrons near the Fermi surface, due to Pauli’s exclusion principle. The electrons near the Fermi surface are in the way, occupying states that the other electrons might otherwise be accelerated into. Thus we see that the displacement Ak is limited by the scattering of the electrons near the Fenni surface. Using the same reasoning as we did for the classical model, we can easily show that the extra velocity picked up by the clectrons near the Fermi surface is given by 47) Here ty is the average time between collisions of electrons near the Fermi surface. Putting this into Eq, (4.70) and remembering that electrical 439 conductivity is defined by the relation J = oE, we have ne? nm (472) From experimental values of o, we can calculate Tee 4.11 Metals, Semiconductors and Insulators From an clectrical viewpoint a material may be classificd as a metal, semiconductor or an insulator by the value of its resistivity. At room temperature metals have a resistivity about 107° ohm.m, semiconductors have values in the range 107* ~ 10° ohm and insulators have resistivities greater than about 10° ohm.m. The main factors which decide this classification for a material are : (@® the valencies of the constituent atoms, and (ii) the magnitude of the energy gaps between the bands. General experimental conclusions can be summarised as follows : @ odd numbers of electrons per primitive unit cell give a metal; Gi) even numbers of electrons per fate unit cell give a metal if there is a id overlap, a semiconductor if there is a small band gap (< 1eV), and an insulator if there is a large band gap. It would be demonstrated that no electric current could arise either from empty bands or from completely filled bands. Any electrical conductivity exhibited by a crystal, due to the motion of free electrons, must therefore arise from the motion of electrons in energy bands which are only partially filled. This observation forms the basis of the distinction between insulators, ‘ metallic conductors and conductors. semi-aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.5.10 Solid State Physics and Devices The transition from the superconducting state to the normal state is observed to be a second order phase transition. In such a transition, there is no latent heat involved, but there is a discontinuity in the heat capacity. The evidence for a gap in the energy spectrum also suggests the possibility that electrons in the energy states below the gap may have different properties from those above and perhaps be responsible for _ super- conductivity. If thermal excitation across the gap is to be small, kT must be less than 2A( 1074 eV) indicating that temperatures of about 1 K are required. Also the specific heat measurements indicate that the superconducting state has a lower entropy than the normal state and so is some way more ordered. Since the superconducting—normal transition —_only involves changes in electronic properties, other properties being little affected, it is reasonable to suppose that the greater degree of order is associated with the electrons. A picture emerges therefore in which ordered electrons below the energy gap are in some way responsible for superconductivity while those above the gap possess normal properties. 5.8 Energy Gap Dependent Properties (a) Microwave and infrared properties : The existence of an energy gap means that photon of cnergy less than the gap energy arc not absorbed. Nearly all the photons incident are reflected as for any metal because of the impedance mismatch at the boundary between vacuum and metal, but for a very thin (~30 A) film more photons are transmitted in the superconducting state than in the normal state, For photon energies less than energy gap, the resistivity of a superconductor vanishes at absolute zero, At T<

S-ar putea să vă placă și

• Muni Raju

  

  Document1 pagină

  Muni Raju

  Sri Pudu

  Încă nu există evaluări
• Materials Characterization - Video Course - S.sankaran - IITM - NPTEL

  

  Document5 pagini

  Materials Characterization - Video Course - S.sankaran - IITM - NPTEL

  Sri Pudu

  Încă nu există evaluări
• Srinivasulu P

  

  Document4 pagini

  Srinivasulu P

  Sri Pudu

  Încă nu există evaluări
• • Reviste
  • Podcast-uri
  • Partitură
• Books and Authors - Questions and Answers Test GKBA3

  

  Document3 pagini

  Books and Authors - Questions and Answers Test GKBA3

  Sri Pudu

  Încă nu există evaluări
• Principles of Metallurgical Thermodynamics - CRC Press Book

  

  Document1 pagină

  Principles of Metallurgical Thermodynamics - CRC Press Book

  Sri Pudu

  0% (3)
• Books and Authors - Questions and Answers Test GKBA1

  

  Document3 pagini

  Books and Authors - Questions and Answers Test GKBA1

  Sri Pudu

  Încă nu există evaluări
• Books and Authors - Questions and Answers Test GKBA2

  

  Document3 pagini

  Books and Authors - Questions and Answers Test GKBA2

  Sri Pudu

  Încă nu există evaluări
• IISC Bangalore - Materials Engineering Courses-2013-14

  

  Document8 pagini

  IISC Bangalore - Materials Engineering Courses-2013-14

  Sri Pudu

  Încă nu există evaluări
• Courses For The January-2014 Semester - Materials Engineering

  

  Document4 pagini

  Courses For The January-2014 Semester - Materials Engineering

  Sri Pudu

  Încă nu există evaluări
• InterviewCallLetter 14104707

  

  Document1 pagină

  InterviewCallLetter 14104707

  Sri Pudu

  Încă nu există evaluări
• Nptel Phase 2 - Courses

  

  Document2 pagini

  Nptel Phase 2 - Courses

  Sri Pudu

  Încă nu există evaluări
• HH

  

  Document260 pagini

  HH

  Sri Pudu

  Încă nu există evaluări
• A Mind Dump of Mathematics..

  

  Document11 pagini

  A Mind Dump of Mathematics..

  Sri Pudu

  Încă nu există evaluări
• Lecture 1

  

  Document9 pagini

  Lecture 1

  Richa Tripathi

  Încă nu există evaluări
• Cover Letters To Get You Hired - Fastweb

  

  Document4 pagini

  Cover Letters To Get You Hired - Fastweb

  Sri Pudu

  Încă nu există evaluări
• Best Books For SBI PO Exam 2014 - Toppers Choice - Government Jobs - Recruitment 2014-2015 - IBPS - SSC - UPSC - Railway Etc

  

  Document6 pagini

  Best Books For SBI PO Exam 2014 - Toppers Choice - Government Jobs - Recruitment 2014-2015 - IBPS - SSC - UPSC - Railway Etc

  Sri Pudu

  Încă nu există evaluări
• Cover Letter NIT Warangal

  

  Document2 pagini

  Cover Letter NIT Warangal

  Sri Pudu

  0% (1)
• (Paper) SBI PO SOLVED EXAM PAPER 2008 - General Awareness & Computer Knowledge - UPSCPORTAL - India's Largest Community For IAS, CSAT, Civil Services Exam Aspirants

  

  Document14 pagini

  (Paper) SBI PO SOLVED EXAM PAPER 2008 - General Awareness & Computer Knowledge - UPSCPORTAL - India's Largest Community For IAS, CSAT, Civil Services Exam Aspirants

  Sri Pudu

  Încă nu există evaluări
• Notes On Quantum Mechanics

  

  Document397 pagini

  Notes On Quantum Mechanics

  Bruno Skiba

  100% (3)
• CVs and Cover Letters - Example CVs - Prospects - Ac

  

  Document3 pagini

  CVs and Cover Letters - Example CVs - Prospects - Ac

  Sri Pudu

  Încă nu există evaluări
• Books Verification

  

  Document2 pagini

  Books Verification

  Sri Pudu

  Încă nu există evaluări
• Combined Geo-Scientist 2014 - Physics Syllabus

  

  Document8 pagini

  Combined Geo-Scientist 2014 - Physics Syllabus

  Srinivasulu Pudu

  Încă nu există evaluări
• Phase Diagrams

  

  Document3 pagini

  Phase Diagrams

  Sri Pudu

  Încă nu există evaluări
• PHD Brouchre

  

  Document5 pagini

  PHD Brouchre

  ilurohit

  Încă nu există evaluări
• Appendix A

  

  Document10 pagini

  Appendix A

  Sri Pudu

  Încă nu există evaluări
• CSIR Physical Sciences June 2009 Paper 2

  

  Document2 pagini

  CSIR Physical Sciences June 2009 Paper 2

  Sri Pudu

  Încă nu există evaluări
• Melvin Seeman

  

  Document10 pagini

  Melvin Seeman

  Sri Pudu

  Încă nu există evaluări
• References Final 1

  

  Document8 pagini

  References Final 1

  Sri Pudu

  Încă nu există evaluări
• Appendix B Indexed XRD Data

  

  Document1 pagină

  Appendix B Indexed XRD Data

  Sri Pudu

  Încă nu există evaluări
• The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  

  De la Everand

  The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  Mark Manson

  Evaluare: 4 din 5 stele

  4/5 (5794)
• The Little Book of Hygge: Danish Secrets to Happy Living

  

  De la Everand

  The Little Book of Hygge: Danish Secrets to Happy Living

  Meik Wiking

  Evaluare: 3.5 din 5 stele

  3.5/5 (400)
• Shoe Dog: A Memoir by the Creator of Nike

  

  De la Everand

  Shoe Dog: A Memoir by the Creator of Nike

  Phil Knight

  Evaluare: 4.5 din 5 stele

  4.5/5 (537)
• Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  

  De la Everand

  Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  Margot Lee Shetterly

  Evaluare: 4 din 5 stele

  4/5 (895)
• The Yellow House: A Memoir (2019 National Book Award Winner)

  

  De la Everand

  The Yellow House: A Memoir (2019 National Book Award Winner)

  Sarah M. Broom

  Evaluare: 4 din 5 stele

  4/5 (98)
• The Emperor of All Maladies: A Biography of Cancer

  

  De la Everand

  The Emperor of All Maladies: A Biography of Cancer

  Siddhartha Mukherjee

  Evaluare: 4.5 din 5 stele

  4.5/5 (271)
• A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  

  De la Everand

  A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  Dave Eggers

  Evaluare: 3.5 din 5 stele

  3.5/5 (231)
• Never Split the Difference: Negotiating As If Your Life Depended On It

  

  De la Everand

  Never Split the Difference: Negotiating As If Your Life Depended On It

  Chris Voss

  Evaluare: 4.5 din 5 stele

  4.5/5 (838)
• Grit: The Power of Passion and Perseverance

  

  De la Everand

  Grit: The Power of Passion and Perseverance

  Angela Duckworth

  Evaluare: 4 din 5 stele

  4/5 (588)
• On Fire: The (Burning) Case for a Green New Deal

  

  De la Everand

  On Fire: The (Burning) Case for a Green New Deal

  Naomi Klein

  Evaluare: 4 din 5 stele

  4/5 (74)
• Yes Please

  

  De la Everand

  Yes Please

  Amy Poehler

  Evaluare: 4 din 5 stele

  4/5 (1891)
• Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  

  De la Everand

  Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  Ashlee Vance

  Evaluare: 4.5 din 5 stele

  4.5/5 (474)
• Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  

  De la Everand

  Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  Gilbert King

  Evaluare: 4.5 din 5 stele

  4.5/5 (266)
• The Unwinding: An Inner History of the New America

  

  De la Everand

  The Unwinding: An Inner History of the New America

  George Packer

  Evaluare: 4 din 5 stele

  4/5 (45)
• Fear: Trump in the White House

  

  De la Everand

  Fear: Trump in the White House

  Bob Woodward

  Evaluare: 3.5 din 5 stele

  3.5/5 (738)
• Team of Rivals: The Political Genius of Abraham Lincoln

  

  De la Everand

  Team of Rivals: The Political Genius of Abraham Lincoln

  Doris Kearns Goodwin

  Evaluare: 4.5 din 5 stele

  4.5/5 (234)
• Principles: Life and Work

  

  De la Everand

  Principles: Life and Work

  Ray Dalio

  Evaluare: 4 din 5 stele

  4/5 (599)
• Angela's Ashes: A Memoir

  

  De la Everand

  Angela's Ashes: A Memoir

  Frank McCourt

  Evaluare: 4.5 din 5 stele

  4.5/5 (440)
• The World Is Flat 3.0: A Brief History of the Twenty-first Century

  

  De la Everand

  The World Is Flat 3.0: A Brief History of the Twenty-first Century

  Thomas L. Friedman

  Evaluare: 3.5 din 5 stele

  3.5/5 (2259)
• The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  

  De la Everand

  The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  Brene Brown

  Evaluare: 4 din 5 stele

  4/5 (1090)
• Rise of ISIS: A Threat We Can't Ignore

  

  De la Everand

  Rise of ISIS: A Threat We Can't Ignore

  Jay Sekulow

  Evaluare: 3.5 din 5 stele

  3.5/5 (137)
• The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  

  De la Everand

  The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  Ben Horowitz

  Evaluare: 4.5 din 5 stele

  4.5/5 (344)
• Bad Feminist: Essays

  

  De la Everand

  Bad Feminist: Essays

  Roxane Gay

  Evaluare: 4 din 5 stele

  4/5 (1015)
• Steve Jobs

  

  De la Everand

  Steve Jobs

  Walter Isaacson

  Evaluare: 4.5 din 5 stele

  4.5/5 (806)
• John Adams

  

  De la Everand

  John Adams

  David McCullough

  Evaluare: 4.5 din 5 stele

  4.5/5 (2409)
• The Glass Castle: A Memoir

  

  De la Everand

  The Glass Castle: A Memoir

  Jeannette Walls

  Evaluare: 4.5 din 5 stele

  4.5/5 (1712)
• The Outsider: A Novel

  

  De la Everand

  The Outsider: A Novel

  Stephen King

  Evaluare: 4 din 5 stele

  4/5 (1839)
• The Light Between Oceans: A Novel

  

  De la Everand

  The Light Between Oceans: A Novel

  M.L. Stedman

  Evaluare: 4.5 din 5 stele

  4.5/5 (789)
• The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  

  De la Everand

  The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  Viet Thanh Nguyen

  Evaluare: 4.5 din 5 stele

  4.5/5 (121)
• Brooklyn: A Novel

  

  De la Everand

  Brooklyn: A Novel

  Colm Toibin

  Evaluare: 3.5 din 5 stele

  3.5/5 (1937)
• The Woman in Cabin 10

  

  De la Everand

  The Woman in Cabin 10

  Ruth Ware

  Evaluare: 3.5 din 5 stele

  3.5/5 (2322)
• A Man Called Ove: A Novel

  

  De la Everand

  A Man Called Ove: A Novel

  Fredrik Backman

  Evaluare: 4.5 din 5 stele

  4.5/5 (4609)
• Little Women

  

  De la Everand

  Little Women

  Louisa May Alcott

  Evaluare: 4 din 5 stele

  4/5 (104)
• Manhattan Beach: A Novel

  

  De la Everand

  Manhattan Beach: A Novel

  Jennifer Egan

  Evaluare: 3.5 din 5 stele

  3.5/5 (792)
• Wolf Hall: A Novel

  

  De la Everand

  Wolf Hall: A Novel

  Hilary Mantel

  Evaluare: 4 din 5 stele

  4/5 (3811)
• The Perks of Being a Wallflower

  

  De la Everand

  The Perks of Being a Wallflower

  Stephen Chbosky

  Evaluare: 4.5 din 5 stele

  4.5/5 (2103)
• The Art of Racing in the Rain: A Novel

  

  De la Everand

  The Art of Racing in the Rain: A Novel

  Garth Stein

  Evaluare: 4 din 5 stele

  4/5 (4200)
• A Tree Grows in Brooklyn

  

  De la Everand

  A Tree Grows in Brooklyn

  Betty Smith

  Evaluare: 4.5 din 5 stele

  4.5/5 (1929)
• Sing, Unburied, Sing: A Novel

  

  De la Everand

  Sing, Unburied, Sing: A Novel

  Jesmyn Ward

  Evaluare: 4 din 5 stele

  4/5 (1103)
• Her Body and Other Parties: Stories

  

  De la Everand

  Her Body and Other Parties: Stories

  Carmen Maria Machado

  Evaluare: 4 din 5 stele

  4/5 (821)
• The Constant Gardener: A Novel

  

  De la Everand

  The Constant Gardener: A Novel

  John le Carré

  Evaluare: 3.5 din 5 stele

  3.5/5 (104)