This action might not be possible to undo. Are you sure you want to continue?

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

Richard Fitzpatrick Professor of Physics The University of Texas at Austin

Contents

1 Introduction 1.1 Intended audience . 1.2 Major Sources . . . 1.3 Aim of Course . . . 1.4 Outline of Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 5 5 6 6 7 7 7 7 9 11 13 13 13 14 15 18 19 21 21 22 25 25 26

2 Probability Theory 2.1 Introduction . . . . . . . . . . . . . . . . 2.2 What is Probability? . . . . . . . . . . . . 2.3 Combining Probabilities . . . . . . . . . . 2.4 Mean, Variance, and Standard Deviation 2.5 Continuous Probability Distributions . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

3 Wave-Particle Duality 3.1 Introduction . . . . . . . . . . . . . . . . . . . . 3.2 Wavefunctions . . . . . . . . . . . . . . . . . . . 3.3 Plane Waves . . . . . . . . . . . . . . . . . . . . 3.4 Representation of Waves via Complex Functions 3.5 Classical Light Waves . . . . . . . . . . . . . . . 3.6 Photoelectric Effect . . . . . . . . . . . . . . . . 3.7 Quantum Theory of Light . . . . . . . . . . . . . 3.8 Classical Interference of Light Waves . . . . . . 3.9 Quantum Interference of Light . . . . . . . . . . 3.10 Classical Particles . . . . . . . . . . . . . . . . . 3.11 Quantum Particles . . . . . . . . . . . . . . . . . 3.12 Wave Packets . . . . . . . . . . . . . . . . . . .

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

2

3.13 3.14 3.15 3.16 Evolution of Wave Packets . . . . Heisenberg’s Uncertainty Principle Schr¨dinger’s Equation . . . . . . o Collapse of the Wave Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

QUANTUM MECHANICS

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 32 35 36 39 39 39 39 42 43 45 47 50 53 56 57 60 63 63 63 64 70 72 73 75 80 85 85 85 86 88 89 93 93 93 96 97 100

4 Fundamentals of Quantum Mechanics 4.1 Introduction . . . . . . . . . . . . . 4.2 Schr¨dinger’s Equation . . . . . . . o 4.3 Normalization of the Wavefunction 4.4 Expectation Values and Variances . 4.5 Ehrenfest’s Theorem . . . . . . . . . 4.6 Operators . . . . . . . . . . . . . . 4.7 Momentum Representation . . . . . 4.8 Heisenberg’s Uncertainty Principle . 4.9 Eigenstates and Eigenvalues . . . . 4.10 Measurement . . . . . . . . . . . . 4.11 Continuous Eigenvalues . . . . . . . 4.12 Stationary States . . . . . . . . . . 5 One-Dimensional Potentials 5.1 Introduction . . . . . . . . . 5.2 Inﬁnite Potential Well . . . . 5.3 Square Potential Barrier . . . 5.4 WKB Approximation . . . . 5.5 Cold Emission . . . . . . . . 5.6 Alpha Decay . . . . . . . . . 5.7 Square Potential Well . . . . 5.8 Simple Harmonic Oscillator 6 Multi-Particle Systems 6.1 Introduction . . . . . . . 6.2 Fundamental Concepts . 6.3 Non-Interacting Particles 6.4 Two-Particle Systems . . 6.5 Identical Particles . . . .

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

7 Three-Dimensional Quantum Mechanics 7.1 Introduction . . . . . . . . . . . . . 7.2 Fundamental Concepts . . . . . . . 7.3 Particle in a Box . . . . . . . . . . . 7.4 Degenerate Electron Gases . . . . . 7.5 White-Dwarf Stars . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

CONTENTS

8 Orbital Angular Momentum 8.1 Introduction . . . . . . . . . . . . . . . 8.2 Angular Momentum Operators . . . . . 8.3 Representation of Angular Momentum 8.4 Eigenstates of Angular Momentum . . . 8.5 Eigenvalues of Lz . . . . . . . . . . . . 8.6 Eigenvalues of L2 . . . . . . . . . . . . 8.7 Spherical Harmonics . . . . . . . . . . 9 Central Potentials 9.1 Introduction . . . . . . . . . . . 9.2 Derivation of Radial Equation . 9.3 Inﬁnite Spherical Potential Well 9.4 Hydrogen Atom . . . . . . . . . 9.5 Rydberg Formula . . . . . . . . 10 Spin Angular Momentum 10.1 Introduction . . . . . . 10.2 Spin Operators . . . . . 10.3 Spin Space . . . . . . . 10.4 Eigenstates of Sz and S2 10.5 Pauli Representation . 10.6 Spin Precession . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

103 103 103 105 106 107 108 110 115 115 115 118 121 126 129 129 129 130 131 133 136 141 141 141 144 147 151 151 151 153 155 156 160 163 164 169 170 175

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

11 Addition of Angular Momentum 11.1 Introduction . . . . . . . . . . . . . . . . . 11.2 General Principles . . . . . . . . . . . . . . 11.3 Angular Momentum in the Hydrogen Atom 11.4 Two Spin One-Half Particles . . . . . . . . 12 Time-Independent Perturbation Theory 12.1 Introduction . . . . . . . . . . . . . . 12.2 Improved Notation . . . . . . . . . . 12.3 Two-State System . . . . . . . . . . . 12.4 Non-Degenerate Perturbation Theory 12.5 Quadratic Stark Effect . . . . . . . . . 12.6 Degenerate Perturbation Theory . . . 12.7 Linear Stark Effect . . . . . . . . . . . 12.8 Fine Structure of Hydrogen . . . . . . 12.9 Zeeman Effect . . . . . . . . . . . . . 12.10 Hyperﬁne Structure . . . . . . . . . . 13 Time-Dependent Perturbation Theory

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

4

13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 13.9 13.10 13.11 13.12 13.13 13.14 Introduction . . . . . . . . . . . . . . Preliminary Analysis . . . . . . . . . . Two-State System . . . . . . . . . . . Spin Magnetic Resonance . . . . . . . Perturbation Expansion . . . . . . . . Harmonic Perturbations . . . . . . . . Electromagnetic Radiation . . . . . . Electric Dipole Approximation . . . . Spontaneous Emission . . . . . . . . Radiation from a Harmonic Oscillator Selection Rules . . . . . . . . . . . . 2P → 1S Transitions in Hydrogen . . Intensity Rules . . . . . . . . . . . . . Forbidden Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

QUANTUM MECHANICS

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 175 177 178 180 181 183 186 188 190 191 192 194 194 197 197 197 199 204 209 209 209 211 213 216 217 219 220

14 Variational Methods 14.1 Introduction . . . . . . 14.2 Variational Principle . . 14.3 Helium Atom . . . . . 14.4 Hydrogen Molecule Ion

15 Scattering Theory 15.1 Introduction . . . . . . . . . . 15.2 Fundamentals . . . . . . . . . 15.3 Born Approximation . . . . . . 15.4 Partial Waves . . . . . . . . . 15.5 Determination of Phase-Shifts 15.6 Hard Sphere Scattering . . . . 15.7 Low Energy Scattering . . . . 15.8 Resonances . . . . . . . . . . .

(John Wiley & Sons. Bohm. Menlo Park CA. 1974). Grifﬁths. (Dover. E. (John Wiley & Sons. In particular. 1998). Dirac. Nonclassical Physics. R. ordinary and partial differential equations. basic probability theory.Introduction 5 1 Introduction 1. Menlo Park CA. Cambridge UK. (McGraw-Hill. Oxford UK.L. 2005). Squires.1 Intended audience These lecture notes outline a single semester course on non-relativistic quantum mechanics which is primarily intended for upper-division undergraduate physics majors. and Fourier series and transforms. New York NY. P. Quantum Mechanics. The course assumes some previous knowledge of physics and mathematics.J. G. 2nd Edition. 1. (Pearson Prentice Hall. 1970). Introduction to Quantum Mechanics.2 Major Sources The textbooks which I have consulted most frequently whilst developing course material are: The Principles of Quantum Mechanics. . Quantum Physics. 1958). S. Sakurai.M. (Addison-Wesley. 4th Edition (revised). D. New York NY. 1995). the physics and mathematics of waves (including the representation of waves via complex functions). New York NY. Modern Quantum Mechanics. New York NY. vector algebra.J. J. 2nd Edition.A. (Cambridge University Press. Introduction to the Quantum Theory. 1985). Quantum Theory. 1996). 1989). Gasiorowicz. 2nd Edition. D. (Oxford University Press. (Benjamin/Cummings. elementary classical electromagnetism and special relativity. prospective students should be reasonably familiar with Newtonian dynamics. Problems in Quantum Mechanics. Merzbacher. Harris. Upper Saddle River NJ. Park. linear algebra. D. 2nd Edition.

After a brief review of probability theory. Chapter 15 contains an introduction to quantum scattering theory.e. respectively. in Chapters 5. Finally. and hyperﬁne structure. we shall start. and 7. ﬁne structure.6 QUANTUM MECHANICS 1. the concept that waves sometimes act as particles.. and a single particle in three dimensions. with a minimum of abstract mathematics. 6. and Chapter 9 to the closely related subject of particle motion in a central potential. respectively. we shall examine spin angular momentum. in Chapter 3. . Quantum mechanics is used to examine the motion of a single particle in one dimension. many particles in one dimension. Time-dependent perturbation theory is employed to study radiative transitions in the hydrogen atom in Chapter 13. Finally.3 Aim of Course The aim of this course is to develop non-relativistic quantum mechanics as a complete theory of microscopic dynamics. the Zeeman effect. We shall then proceed to investigate the rules of quantum mechanics in a more systematic fashion in Chapter 4. Chapter 14 illustrates the use of variational methods in quantum mechanics. in the hydrogen atom. 1. Chapter 8 is devoted to the investigation of orbital angular momentum. and particles as waves. in Chapters 10 and 11. The second part of this course describes selected practical applications of quantum mechanics. by examining how many of the central ideas of quantum mechanics are a direct consequence of wave-particle duality—i. in Chapter 2. time-independent perturbation theory is used to investigate the Stark effect. capable of making detailed predictions. and the addition of orbital and spin angular momentum. In Chapter 12.4 Outline of Course The ﬁrst part of the course is devoted to an in-depth exploration of the basic principles of quantum mechanics.

(2. introduction to a branch of mathematics known as probability theory. Σ. Let us determine the probability of obtaining the outcome X or Y. respectively. which we shall denote P(X | Y). This can result in any one of a number of different possible outcomes. with probabilities of occurrence P(X) and P(Y). 2. and fairly low level.2) Ω(Σ)→∞ Ω(Σ) . The probability of the outcome X is deﬁned as the ratio of the number of systems in the ensemble which exhibit this outcome to the total number of systems. which is just the French for “group. We can see that the probability P(X) must be a number between 0 and 1. Suppose that we wish to ﬁnd the probability of some general outcome. This is just another way of saying that there is no chance of the outcome X. we have to consider the system as a member of a large set.2 What is Probability? What is the scientiﬁc deﬁnition of probability? Well. Mathematicians have a fancy name for a large group of similar systems. They call such a group an ensemble. (2.3 Combining Probabilities Consider two distinct possible outcomes.” So. Σ. In order to ascribe a probability. let us consider an observation made on a general system. X. even when the number of systems goes to inﬁnity. From the basic deﬁnition of probability. Ω(X | Y) P(X | Y) = lim .Probability Theory 7 2 Probability Theory 2. S. X and Y. This is another way of saying that the outcome X is bound to occur. in the limit that the latter number tends to inﬁnity. S. We can write this symbolically as Ω(X) . The probability is zero if no systems exhibit the outcome X. of an observation made on the system S. 2. and Ω(X) the number of systems exhibiting the outcome X. of similar systems. let us consider an ensemble. The probability is unity if all systems exhibit the outcome X in the limit as the number of systems goes to inﬁnity.1) P(X) = lim Ω(Σ)→∞ Ω(Σ) where Ω(Σ) is the total number of systems in the ensemble. of similar systems.1 Introduction This section is devoted to a brief.

Suppose that we make an observation on a system picked at random from the ensemble.e. Now. Let us denote this ensemble Σ ⊗ Σ. this quantity is also equal to the sum of the probabilities of all the individual outcomes. (2. from what has just been said. In order to determine this probability. We are assuming here that the ﬁrst observation does not inﬂuence the second observation in any way. It follows.5) Furthermore.4). Let us determine the probability of obtaining any of these outcomes. P(X | Y) = P(X) + P(Y). we have to form an ensemble of all of the possible pairs of systems which we could choose from the ensemble Σ. and must be satisﬁed by any complete set of probabilities. which equals 1/3.7) The above expression is called the normalization condition. Let us denote all of the M..8 QUANTUM MECHANICS where Ω(X | Y) is the number of systems in the ensemble which exhibit either the outcome X or the outcome Y. because each of the systems in the ensemble must exhibit one of the possible outcomes. Let us determine the probability of obtaining the outcome X in the ﬁrst system and the outcome Y in the second system. For instance. i=1 (2. by (2. the number of pairs of systems in the ensemble Σ ⊗ Σ which exhibit the outcome X in the ﬁrst system and Y in the second system is simply the product of the number of systems which exhibit the outcome X and the number of systems which exhibit the outcome Y in the original ensemble.6) . so Ω(Σ ⊗ Σ) = Ω(Σ) Ω(Σ). which we shall denote P(X ⊗ Y). and then pick a second system completely independently and make another observation. with a six-sided die the probability of throwing any particular number (one to six) is 1/6. so we conclude that this sum is equal to unity: i. possible outcomes of an observation made on the system S by Xi . (2. that the probability of throwing either a one or a two is simply 1/6 + 1/6. Ω(X | Y) = Ω(X) + Ω(Y) (2. The number of pairs of systems in this new ensemble is just the square of the number of systems in the original ensemble. M P(Xi ) = 1. from the basic deﬁnition of probability. the probability of the outcome X or the outcome Y is just the sum of the individual probabilities of X and Y. This condition is equivalent to the self-evident statement that an observation of a system must deﬁnitely result in one of its possible outcomes.3) if the outcomes X and Y are mutually exclusive (which must be the case if they are two distinct outcomes). The fancy mathematical way of saying this is that the two observations are statistically independent. There is another way in which we can combine probabilities. because all of the possible outcomes are considered to be equally likely. say. But. so that Ω(X ⊗ Y) = Ω(X) Ω(Y).4) So. (2. where i runs from 1 to M. Thus. This quantity is unity.

for a general variable u.13) .10) Well. P(u2 ). etc. (2. (Actually. etc. Variance. and Standard Deviation What is meant by the mean or average of a quantity? Well. this deﬁnition is only accurate in the limit that Nstudents is very large. f(u1 ) corresponds to u1 and occurs with the probability P(u1 ). this probability is deﬁned P18 ≃ N18 Nstudents . It follows from our previous deﬁnition that the mean value of f(u) is given by M f(u) ≡ P(ui ) f(ui ). which is denoted u . for each of the M possible values of u there is a corresponding value of f(u) which occurs with the same probability. P(uM ). and so on. with corresponding probabilities P(u1 ). · · · . i=1 (2.11) u ≡ P(ui ) ui .12) Suppose that f(u) is some function of u. which can take on any one of M possible values u1 . i=1 (2. the probability of obtaining the outcomes X and Y in two statistically independent observations is the product of the individual probabilities of X and Y.8) Thus. were currently enrolled. (2. We would then write something like Average Age ≃ N18 × 18 + N19 × 19 + N20 × 20 + · · · . For instance. there is nothing special about the age distribution of students at UT Austin. Then. is deﬁned as M where Nstudents is the total number of enrolled students. N18 + N19 + N20 · · · (2. which equals 1/36. the mean or average value of u. suppose that we wished to calculate the average age of undergraduates at the University of Texas at Austin.Probability Theory It follows from the basic deﬁnition of probability that P(X ⊗ Y) = Ω(X ⊗ Y) = P(X) P(Y). Thus. Ω(Σ)→∞ Ω(Σ ⊗ Σ) lim 9 (2. uM. We could go to the central administration building and ﬁnd out how many eighteen year-olds. · · · . u2 . So. the probability of throwing a one and then a two on a six-sided die is 1/6 × 1/6. Suppose that we were to pick a student at random and then ask “What is the probability of this student being eighteen?” From what we have already discussed. nineteen year-olds.9) where N18 is the number of enrolled eighteen year-olds.) We can now see that the average age takes the form Average Age ≃ P18 × 18 + P19 × 19 + P20 × 20 + · · · . 2.4 Mean.

unless there is no scatter at all in the distribution. The standard deviation is essentially the width of the range over which u is distributed around its mean value u . so that all possible values of u correspond to the mean value u . The average value of this quantity. if c is a general constant then c f(u) = c f(u) . But. Finally. (2.15) We now know how to deﬁne the mean value of the general variable u. giving (∆u)2 = u2 − u 2 . (2. M (∆u) 2 = i=1 P(ui ) (ui − u )2 . σu = (∆u)2 1/2 (2. (2.19) is usually called the variance. in which case it is zero.22) which is usually called the standard deviation of u.10 QUANTUM MECHANICS Suppose that f(u) and g(u) are two general functions of u. It follows that M M M f(u) + g(u) = i=1 P(ui ) [f(ui ) + g(ui )] = i=1 P(ui ) f(ui ) + i=1 P(ui ) g(ui). A more useful measure of the scatter is given by the square root of the variance. how can we characterize the scatter around the mean value? We could investigate the deviation of u from its mean value u .21) The variance of u is proportional to the square of the scatter of u around its mean value. . (2. A more interesting quantity is the square of the deviation. The variance is a positive number.16) (2. this is not a particularly interesting quantity. (2. The following general relation is often useful (u − u )2 = (u2 − 2 u u + u 2 ) = u2 − 2 u u + u 2 . In fact. since its average is zero: ∆u = (u − u ) = u − u = 0.14) so f(u) + g(u) = f(u) + g(u) .18) (2. (2. which is denoted ∆u ≡ u − u .20) .17) This is another way of saying that the average deviation from the mean vanishes.

(a) What is the probability of the player still being alive after playing the game N times? (b) What is the probability of the player surviving N − 1 turns in this game. s0 where 0 ≤ s ≤ ∞. More explicitly. we expect the probability that u takes on a value in the range u to u + du to be directly proportional to du. The player then spins the drum. now. (a) Determine A in terms of s0 . and (2.24) (2.26) u (∆u)2 respectively.. aims at his/her head.Probability Theory 11 2.5 Continuous Probability Distributions Suppose. Exercises = = −∞ P(u) (u − u )2 du = u2 − u 2 . P(u) u du.25) (2. P(s) ds gives the probability that a car has a speed between s and s + ds. leaving the other ﬁve chambers of the drum empty. (2. In general. (b) What is the mean value of the speed? (c) What is the “most probable” speed: i. and pulls the trigger. 1. P(u ∈ u : u + du) = P(u) du. In the “game” of Russian roulette. and then being shot the Nth time he/she pulls the trigger? (c) What is the mean number of times the player gets to pull the trigger? 2. A and s0 are positive constants. the player inserts a single cartridge into the drum of a revolver. that the variable u can take on a continuous range of possible values. Here.e.5). Suppose that the probability density for the speed s of a car on a road is given by P(s) = A s exp − s . The earlier results (2. in the limit that du → 0. the speed for which the probability density has a maximum? (d) What is the probability that a car has a speed more than three times as large as the mean value? . (2. In other words.12).19) generalize in a straightforward manner to give ∞ 1 = −∞ ∞ P(u) du. −∞ ∞ (2.23) where P(u) is known as the probability density.

. Demonstrate that P(t) = e−w t .12 QUANTUM MECHANICS 3. What is the mean lifetime of the atom? . Let P(t) be the probability of the atom not having decayed at time t. the probability of decay in a time interval dt is w dt. An radioactive atom has a uniform decay probability per unit time w: i. given that it was created at time t = 0.e.

e. It follows that the wavefunction (3. a wave is generally represented in terms of a wavefunction: e. in quantum mechanics.1) is periodic in x with period λ = 2π/k: i. appearing in expression (3. 3. t) = ψ(x. as θ varies. is termed the phase angle. the whole wave. k. if we are considering a light wave then ψ(x. ψ(x + λ. is called its period.e. and the quantity ω the wave angular frequency. Furthermore. θ.1). t represents time. the wavefunction oscillates between the minimum and maximum values −A and +A.e. the quantity ϕ. the wavefunction is periodic in t with period T = 2π/ω: i. where v = ω/k.1) where x represents position. For instance.1 Introduction In classical mechanics. The function also oscillates between the minimum and maximum values −1 and +1. we shall examine how wave-particle duality shapes the general features of quantum mechanics. The spatial period of the wave.. cos(θ + 2π) = cos θ for all θ. is periodic in its argument. This follows because the maxima of cos(θ) occur at θ = j 2π. as x and t vary. λ. is ν = 1/T = ω/2π. the quantity k the wavenumber. the cosine function. Moreover. t+T ) = ψ(x.. t) might correspond to the pressure perturbation associated with the wave at position x and time t. with period 2π: i. Finally.Wave-Particle Duality 13 3 Wave-Particle Duality 3. waves and particles are two completely distinct types of physical entity. Analogous . The conventional wave frequency. is known as its wavelength. waves sometimes act as particles. t) = A cos(k x − ω t + ϕ). In one dimension. respectively. It follows that the maximum. and determines the exact positions of the wave maxima and minima at a given time. t) for all x and t.. propagates in the positive x-direction at the velocity ω/k. ψ(x. Finally. T . respectively. and the temporal period. and A. However. Note that a given maximum satisﬁes x = (j − ϕ/2π) λ + v t. In this chapter. On the other hand. if we are considering a sound wave then ψ(x. Note that the units of ω are radians per second. where j is an integer. As is wellknown. in cycles per second (otherwise known as hertz). In fact. the maxima are located at k x − ω t + ϕ = j 2π. by implication. the quantity A is termed the wave amplitude..2 Wavefunctions A wave is deﬁned as a disturbance in some physical system which is periodic in both space and time. and. cos(θ).g. t) might represent the wave’s transverse electric ﬁeld. ω > 0. whereas particles are discrete and have little or no spatial extent. (3. Waves are continuous and spatially extended. t) for all x and t. ψ(x. and particles sometimes act as waves—this strange behaviour is known as waveparticle duality.

which are equally spaced a distance λ = 2π/k apart. is represented by the following wavefunction: ψ(x. and propagate along the positive x-axis at the velocity v = ω/k.2) is the wavefunction of a wave of amplitude A.1: The solution of n · r = d is a plane. (3. t) = A cos(k x − ω t + ϕ).3 Plane Waves As we have just seen. (3. normal to the x-axis.3) Now. Furthermore. reasoning reveals that ψ(x. . (3. and phase angle ϕ.4) can be re-written in the form x = d. the type of wave represented above is conventionally termed a one-dimensional plane wave. Moreover.14 QUANTUM MECHANICS n r d plane origin Figure 3. (3.5) where d = (j − ϕ/2π) λ + v t. wavenumber k. which are located at k x − ω t + ϕ = j 2π. (3. it is a plane wave because the wave maxima. angular frequency ω.5) is the equation of a plane. 3. t) = A cos(−k x − ω t + ϕ) = A cos(k x + ω t − ϕ). It is one-dimensional because its associated wavefunction only depends on the single Cartesian coordinate x. wavenumber k. whose distance of closest approach to the origin is d. as is well-known. angular frequency ω. consist of a series of parallel planes. propagating in the positive x-direction. These conclusions follow because Eq. (3. a wave of amplitude A. and phase angle ϕ.4) where j is an integer. which propagates in the negative x-direction at the velocity ω/k. normal to the x-axis.

It follows. 3. x (say).6) where n = (1. it follows that if n is re-interpreted as a unit vector pointing in an arbitrary direction then (3. Now. (3.e. The wave represented above is conventionally termed a three-dimensional plane wave. and. it is a plane wave because the wave maxima are located at k · r − ω t + ϕ = j 2π. See Fig. See Fig. λ = 2π/k. 15 (3. Since there is nothing special about the x-direction. ψ(x.1. It is threedimensional because its wavefunction. x is termed the real part of z. the plane is normal to n.Wave-Particle Duality The previous equation can also be written in the coordinate-free form n · r = d. z. As before. y. and r = (x. the complex conjugate of z is deﬁned z∗ = x − i y. one of the solutions of i2 = −1. a real number. whereas its magnitude determines the wavenumber. This observation allows us to write the three-dimensional equivalent to the wavefunction (3. which are equally spaced a distance λ apart.6) can be re-interpreted as the general equation of a plane.7) where the constant vector k = (kx . In fact.. can take any value in a continuum of different values lying between −∞ and +∞.9) (3. k. z.10) where x and y are real numbers. In addition. the wavelength. depends on all three Cartesian coordinates. and which propagate in the k-direction at the velocity v.2.4 Representation of Waves via Complex Functions In mathematics. kz ) = k n is called the wavevector. Hence. normal to the wavevector.6). y. the symbol i is conventionally used to represent the square-root of minus one: i. ky . k. (3. 0) is a unit vector directed along the positive x-axis. y.3) as ψ(x. Moreover. t). (3. the direction of the wavevector speciﬁes the wave propagation direction. by comparison with Eq. Note that the wavenumber. (3. whereas the square of an imaginary number is a negative real number. an imaginary number takes the general form i y. and its distance of closest approach to the origin is d. or where λ = 2π/k. Finally. 0. On the other hand. thus. k: i.e.. n · r = (j − ϕ/2π) λ + v t. and y the imaginary part of z. is the magnitude of the wavevector. and v = ω/k. k ≡ |k|. It follows that the square of a real number is a positive real number. z) represents the vector displacement of a general point from the origin.8) 3. 3. that the wave maxima consist of a series of parallel planes. a general complex number is written z = x + i y. . where y is a real number. t) = A cos(k · r − ω t + ϕ). This is written mathematically as x = Re(z) and y = Im(z).

z. and y = r sin θ. The distance. All such representations depend ultimately on a fundamental mathematical identity. The coordinates of the point in question are the real and imaginary parts of the number: i. This idea is illustrated in Fig.e. k the wavenumber. a one-dimensional wavefunction takes the general form ψ(x.. of the representative point from the origin is termed the modulus of the corresponding complex number. Now. which takes the form e i φ ≡ cos φ + i sin φ. can be written z = r e i θ. Incidentally.16 QUANTUM MECHANICS k λ Figure 3. that the straight-line joining the representative point to the origin subtends with the real axis is termed the argument of the corresponding complex number.14) .2: Wave maxima associated with a three-dimensional plane wave. θ = tan−1 (y/x). just as we can visualize a real number as a point on an inﬁnite straight-line.11) where φ is a real number. given that z = r cos θ + i r sin θ = r (cos θ + i sin θ). (3. we can visualize a complex number as a point in an inﬁnite plane. z ≡ (x. The angle. it follows from de Moivre’s theorem that any complex number. t) = A cos(k x − ω t + ϕ). it follows that z z∗ = x2 + y2 = |z|2 . z.12) where A is the wave amplitude. ω the angular frequency. Now. r = |z| its modulus. (3. z = r cos θ + i r sin θ. Hence. known as de Moivre’s theorem. Incidentally.13) (3. where r = |z| and θ = arg(z) are real numbers. Consider the complex wavefunction ψ(x. y). r = x2 + y2 . and ϕ the phase angle. (3. This is written mathematically as |z| = x2 + y2 .3. It follows from standard trigonometry that x = r cos θ. Complex numbers are often used to represent wavefunctions. and θ = arg(z) its argument. where z is a general complex number. 3. t) = ψ0 e i (k x−ω t) . This is written mathematically as arg(z) = tan−1 (y/x). z.

where ψ0 is a complex constant. (3. (3.17) (3. (3. a general one-dimensional real wavefunction.18) The main advantage of the complex representation. and Eq. and . We can write ψ0 = A e i ϕ .15) In other words.3: Representation of a complex number as a point in a plane. where A is the modulus. t) = ψ0 e i (k x−ω t) . over the more straightforward real representation. t). A. (3. that Re ψ0 e i (k x−ω t) = A cos(k x − ω t + ϕ) = ψ(x.13). of ψ0 . For ease of notation.14).18). and our general one-dimension wavefunction is simply written ψ(x. the “take the real part” aspect of the above expression is usually omitted.Wave-Particle Duality 17 imaginary z r y θ x real Figure 3. can be represented as the real part of a complex wavefunction of the form (3.13). it follows from de Moirve’s theorem. (3.16) (3. Hence. Thus. and ϕ the argument. we deduce that Re ψ0 e i (k x−ω t) = Re A e i ϕ e i (k x−ω t) = Re A e i (k x−ω t+ϕ) = A Re e i (k x−ω t+ϕ) .13). is that the former enables us to combine the amplitude. (3.

21) where ϕ is real. ω the angular frequency. and ψ is a complex wave amplitude. equivalently. (3. k and ω are real parameters. (3. where k is the wavevector. of course. (According to standard electromagnetic theory. monochromatic. t) = Re[ψ(x.22) and (3.5 Classical Light Waves Consider a classical. i = −1. It follows that the wave maxima and minima satisfy x − c t = constant. In addition. t) = ψ0 e i (k·r−ω t) . (3. and ν = ω/2π the frequency (in hertz). or. t) = |ψ| cos (k [x − (ω/k) t] + ϕ) = |ψ| cos (k [x − c t] + ϕ) .25) Note that Ey depends on x and t only via the combination x − c t. (3. the three dimensional generalization of the above expression is ψ(r. (3.) Now.23) .24) yield ¯ ¯ Ey (x. propagating through a vacuum in the x-direction. It is convenient to characterize a light wave (which is. Suppose that ¯ ¯ ψ = |ψ| e i ϕ . of the wavefunction into a single complex amplitude. the physical electric ﬁeld is the real part of the above expression. plane light wave. ω = k c. t) = ψ e i (k x−ω t) . Finally. and ϕ the phase angle. the frequency and wavelength of light waves are related according to the well-known expression c = ν λ. a type of electromagnetic wave) by specifying its associated electric ﬁeld. where c = 3 × 108 m/s. Suppose that the wave is polarized such that this electric ﬁeld oscillates in the y-direction. the magnetic ﬁeld oscillates in the z-direction. ψ0 . It follows that the physical electric ﬁeld takes the form ¯ Ey (x. ϕ. According to standard electromagnetic theory.22) ¯ where |ψ| is the amplitude of the electric oscillation. in phase with the electric ﬁeld. linearly polarized.20) √ ¯ Here. Equations (3.26) (3. λ = 2π/k is the wavelength. with an amplitude which is that of the electric ﬁeld divided by the velocity of light in vacuum. By convention.24) (3. k the wavenumber. the electric ﬁeld can be conveniently represented in terms of a complex wavefunction: ¯ ψ(x. t)] = |ψ| cos(k x − ω t + ϕ).19) 3. (3.18 QUANTUM MECHANICS the phase angle.

**Wave-Particle Duality
**

Thus, the wave maxima and minima propagate in the x-direction at the ﬁxed velocity dx = c. dt

19

(3.27)

An expression, such as (3.24), which determines the wave angular frequency as a function of the wavenumber, is generally termed a dispersion relation. As we have already seen, and as is apparent from Eq. (3.25), the maxima and minima of a plane wave propagate at the characteristic velocity ω (3.28) vp = , k which is known as the phase velocity. Hence, the dispersion relation (3.24) is effectively saying that the phase velocity of a plane light wave propagating through a vacuum always takes the ﬁxed value c, irrespective of its wavelength or frequency. Now, from standard electromagnetic theory, the energy density (i.e., the energy per unit volume) of a light wave is 2 Ey U= , (3.29) ǫ0 where ǫ0 = 8.85 × 10−12 F/m is the permittivity of free space. Hence, it follows from Eqs. (3.20) and (3.22) that U ∝ |ψ| 2 . (3.30) Furthermore, a light wave possesses linear momentum, as well as energy. This momentum is directed along the wave’s direction of propagation, and is of density G= U . c (3.31)

3.6 Photoelectric Effect

The so-called photoelectric effect, by which a polished metal surface emits electrons when illuminated by visible and ultra-violet light, was discovered by Heinrich Hertz in 1887. The following facts regarding this effect can be established via careful observation. First, a given surface only emits electrons when the frequency of the light with which it is illuminated exceeds a certain threshold value, which is a property of the metal. Second, the current of photoelectrons, when it exists, is proportional to the intensity of the light falling on the surface. Third, the energy of the photoelectrons is independent of the light intensity, but varies linearly with the light frequency. These facts are inexplicable within the framework of classical physics. In 1905, Albert Einstein proposed a radical new theory of light in order to account for the photoelectric effect. According to this theory, light of ﬁxed frequency ν consists of a collection of indivisible discrete packages, called quanta,1 whose energy is E = h ν.

1

(3.32)

Plural of quantum: Latin neuter of quantus: how much?

20

QUANTUM MECHANICS

K

h

0 0

W/h

ν

Figure 3.4: Variation of the kinetic energy K of photoelectrons with the wave-frequency ν. Here, h = 6.6261 × 10−34 J s is a new constant of nature, known as Planck’s constant. Incidentally, h is called Planck’s constant, rather than Einstein’s constant, because Max Planck ﬁrst introduced the concept of the quantization of light, in 1900, whilst trying to account for the electromagnetic spectrum of a black body (i.e., a perfect emitter and absorber of electromagnetic radiation). Suppose that the electrons at the surface of a metal lie in a potential well of depth W. In other words, the electrons have to acquire an energy W in order to be emitted from the surface. Here, W is generally called the work function of the surface, and is a property of the metal. Suppose that an electron absorbs a single quantum of light. Its energy therefore increases by h ν. If h ν is greater than W then the electron is emitted from the surface with residual kinetic energy K = h ν − W. (3.33)

Otherwise, the electron remains trapped in the potential well, and is not emitted. Here, we are assuming that the probability of an electron simultaneously absorbing two or more light quanta is negligibly small compared to the probability of it absorbing a single light quantum (as is, indeed, the case for low intensity illumination). Incidentally, we can calculate Planck’s constant, and the work function of the metal, by simply plotting the kinetic energy of the emitted photoelectrons as a function of the wave frequency, as shown in Fig. 3.4. This plot is a straight-line whose slope is h, and whose intercept with the ν axis is W/h. Finally, the number of emitted electrons increases with the intensity of the light because the more intense the light the larger the ﬂux of light quanta onto the surface. Thus, Einstein’s quantum theory is capable of accounting for all three of the previously mentioned observational facts regarding the photoelectric effect.

Wave-Particle Duality

21

**3.7 Quantum Theory of Light
**

According to Einstein’s quantum theory of light, a monochromatic light wave of angular frequency ω, propagating through a vacuum, can be thought of as a stream of particles, called photons, of energy E = ¯ ω, h (3.34) where ¯ = h/2π = 1.0546 × 10−34 J s. Since classical light waves propagate at the ﬁxed h velocity c, it stands to reason that photons must also move at this velocity. Now, according to Einstein’s special theory of relativity, only massless particles can move at the speed of light in vacuum. Hence, photons must be massless. Special relativity also gives the following relationship between the energy E and the momentum p of a massless particle, p= E . c (3.35)

Note that the above relation is consistent with Eq. (3.31), since if light is made up of a stream of photons, for which E/p = c, then the momentum density of light must be the energy density divided by c. It follows from the previous two equations that photons carry momentum p=¯k h (3.36) along their direction of motion, since ω/c = k for a light wave [see Eq. (3.24)].

**3.8 Classical Interference of Light Waves
**

Let us now consider the classical interference of light waves. Figure 3.5 shows a standard double-slit interference experiment in which monochromatic plane light waves are normally incident on two narrow parallel slits which are a distance d apart. The light from the two slits is projected onto a screen a distance D behind them, where D ≫ d. Consider some point on the screen which is located a distance y from the centre-line, as shown in the ﬁgure. Light from the ﬁrst slit travels a distance x1 to get to this point, whereas light from the second slit travels a slightly different distance x2 . It is easily demonstrated that d (3.37) ∆x = x2 − x1 ≃ y, D provided d ≪ D. It follows from Eq. (3.20), and the well-known fact that light waves are superposible, that the wavefunction at the point in question can be written ψ(y, t) ∝ ψ1 (t) e i k x1 + ψ2 (t) e i k x2 , (3.38)

where ψ1 and ψ2 are the wavefunctions at the ﬁrst and second slits, respectively. However, ψ1 = ψ2 , (3.39)

22

QUANTUM MECHANICS

incoming wave

x1

y double slits

d x2

projection screen

D

Figure 3.5: Classical double-slit interference of light. since the two slits are assumed to be illuminated by in-phase light waves of equal amplitude. (Note that we are ignoring the difference in amplitude of the waves from the two slits at the screen, due to the slight difference between x1 and x2 , compared to the difference in their phases. This is reasonable provided D ≫ λ.) Now, the intensity (i.e., the energy ﬂux) of the light at some point on the projection screen is approximately equal to the energy density of the light at this point times the velocity of light (provided that y ≪ D). Hence, it follows from Eq. (3.30) that the light intensity on the screen a distance y from the center-line is I(y) ∝ |ψ(y, t)| 2. (3.40) Using Eqs. (3.37)–(3.40), we obtain I(y) ∝ cos2 k ∆x 2 ≃ cos2 kd y . 2D (3.41)

Figure 3.6 shows the characteristic interference pattern corresponding to the above expression. This pattern consists of equally spaced light and dark bands of characteristic width Dλ ∆y = . (3.42) d

**3.9 Quantum Interference of Light
**

Let us now consider double-slit light interference from a quantum mechanical point of view. According to quantum theory, light waves consist of a stream of massless photons

the interference pattern is built up one photon at a time: i. is if each photon has a greater probability of striking the ﬁlm at points where the classical interference pattern is bright. Now. So.e. 3. Py (y) is the probability that a given photon strikes . do we see an interference pattern which looks like that shown in Fig. Hence. N (3. in order to reconcile the classical and quantum viewpoints. where ∆y is a relatively small division. then.43) where I(y) is given in Eq. Suppose.41). Suppose. according to the above discussion. the pattern is not due to the interaction of different photons. after a great many photons have passed through the apparatus. as determined by experiment. Moreover. that we reduce the intensity of the light source illuminating the slits until the source is so weak that only a single photon is present between the slits and the projection screen at any given time. moving at the speed of light. the number of photons which strike a region of the ﬁlm in a given time interval is equivalent to the intensity of the light illuminating that region of the ﬁlm multiplied by the area of the region. given that there is only one photon in the apparatus at any given time. if we wait a sufﬁciently long time that a great many photons have passed through the slits and struck the photographic ﬁlm. Let us also replace the projection screen by a photographic ﬁlm which records the position where it is struck by each photon. Here. we expect the two slits in Fig. we need Py (y) ≡ lim N→∞ n(y) ∝ I(y) ∆y.Wave-Particle Duality 23 I(y) ∆y 0 y Figure 3. however. (3. is that we see exactly the same interference pattern.5 to be spraying photons in all directions at the same rate.6: Classical double-slit interference pattern. the point at which a given photon strikes the ﬁlm is not inﬂuenced by the points at which previous photons struck the ﬁlm. the only way in which the classical interference pattern can be reconstructed. and then develop the ﬁlm. and a lesser probability of striking the ﬁlm at points where the interference pattern is dark. and then count the number of photons which strike the recording ﬁlm between y and y + ∆y. Let us call this number n(y).. since each photon carries a ﬁxed amount of energy. Hence.6? The answer to this question. Hence. Now. that we allow N photons to pass through our apparatus. 3.

we know exactly which slit each photon passed through. (3. Note that Py ∝ ∆y. Another way of saying this is that the probability of a measurement of the photon’s distance from the centerline. In principle. and then open the slit and close off the other slit. In this second experiment. which is such that the probability of a photon striking a region of the ﬁlm of inﬁnitesimal width dy is Py (y) = P(y) dy. which is virtually identical to the ﬁrst on the individual photon level. the probability density of a photon striking a given point on the ﬁlm is proportional to the modulus squared of the wavefunction at that point. t)| 2 . yielding a result between y and y+dy is proportional to |ψ(y. However. t)| 2 dy. I(y) ∝ |ψ(y. the probability of a photon striking a region of the ﬁlm of width ∆y is directly proportional to this width.24 QUANTUM MECHANICS the ﬁlm between y and y + ∆y. we can only reconcile the wave-like and particle-like properties of light in a statistical sense. Note that. in the quantum mechanical version of the two-slit interference experiment. It follows that. It is convenient to deﬁne a quantity known as the probability density. A probability of 0 means that there is no chance of a photon striking the ﬁlm between y and y + ∆y. this is only true as long as ∆y is relatively small. Thus. which gives P(y) ∝ I(y). .43) yields Py (y) ∝ I(y) dy. it is impossible to obtain a two-slit interference pattern from a single slit. given its initial position and velocity. send N/2 photons through the apparatus.44) In other words. suppose that our original interference experiment involves sending N ≫ 1 photons through our apparatus. whereas a probability of 1 means that every photon is certain to strike the ﬁlm in this interval. Hence. we must think of each photon as essentially passing through both slits simultaneously. t)| 2. In other words. Suppose that we perform a modiﬁed interference experiment in which we close off one slit. and send N/2 photons through the apparatus. according to Eq. Now. In other words. according to wave theory. P(y). in the quantum mechanical picture. in principle. If photons behaved classically then we could. can we determine which slit a given photon passed through? Well. After all. (3. we conclude that any attempt to measure which slit each photon in our two-slit interference experiment passes through results in the destruction of the interference pattern. the wave theory of light (which we expect to agree with the quantum theory in the limit N ≫ 1) tells us that our modiﬁed interference experiment will not result in the formation of an interference pattern. This loss of determinancy in quantum mechanics is a direct consequence of wave-particle duality. It is impossible to reconcile them on the individual particle level. Eq. each photon which passes through our apparatus is equally likely to pass through one of the two slits. (3. we obtain P(y) ∝ |ψ(y. However. So.40). We know that we get an interference pattern in this experiment. solve their equations of motion and predict exactly where each photon was going to strike the ﬁlm. Actually. at the location of the ﬁlm. we can only predict the probability that a given photon strikes a given point on the ﬁlm. This probability is simply a number between 0 and 1.

) The de Broglie wavelength of a proton is λp = 2. electrons). of mass m. move classically in a straight-line with velocity p v= . it is actually quite difﬁcult to do particle interference experiments. Now. Usually the “slits” consist of crystals. h (3. (3. Equation (3. It is found that h λ= .g. it is possible to obtain a double-slit interference pattern from a stream of mono-energetic electrons passing through two closely spaced narrow slits. that of an electron is λe = 1.42)]. on the behaviour of nonrelativistic particles of non-zero mass (e.48) where the electron energy is conveniently measured in units of electron-volts (eV). it turns out that massive particles sometimes exhibit wave-like properties. Hence. the effective wavelength of the electrons can be determined by measuring the width of the light and dark bands in the interference pattern [see Eq. Note that the de Broglie wavelength is generally pretty small. (3. The above wavelength is called the de Broglie wavelength.50) . (3. in order to perform an effective interference experiment.10 Classical Particles In this course.49) Given the smallness of the de Broglie wavelengths of common particles. and so on. such particles.47) can be rearranged to give p = ¯ k.9 × 10−11 [E(eV)]−1/2 m.11 Quantum Particles Just as light waves sometimes exhibit particle-like properties.45) m and satisfy p2 . we are going to concentrate. energy E. (An electron accelerated from rest through a potential difference of 1000 V acquires an energy of 1000 eV. (3. the spacing of the slits must not be too much greater than the wavelength of the wave. In general. particle interference experiments require either very low energy particles (since λ ∝ E−1/2 ).46) E= 2m 3. (3. or very closely spaced slits. For instance. almost exclusively. For instance.47) p The same relation is found for other types of particles.2 × 10−9 [E(eV)]−1/2 m. which act a bit like diffraction gratings with a characteristic spacing of order the inter-atomic spacing (which is generally about 10−9 m).Wave-Particle Duality 25 3. In the absence of external forces. (3. after Louis de Broglie who ﬁrst suggested that particles should have wave-like properties in 1923. and momentum p..

Since the relation (3. that this is only half of the classical particle velocity. a particle plane wave propagates at vp = (3. this phase velocity is only half of the classical velocity of a massive particle. (3. However.53). It follows that the momentum of a quantum particle. the modulus squared of the wavefunction (3.53) is incorrect? Let us investigate further.56) represents a plane wave whose maxima and minima propagate in the positive x-direction with the phase velocity vp = ω/k. which depends on neither x nor t.26 QUANTUM MECHANICS which is exactly the same as the relation between momentum and wavenumber that we obtained earlier for photons [see Eq. In other words.52) for particle waves. (3. (3.53) 2m Now. then Eqs. Expression (3.36) between momentum and wavenumber applies to both photons and massive particles. and we can write E=¯ω h (3.34) between energy and wave angular frequency should also apply to both photons and particles.55) Note.56) is |ψ| 2 . this wavefunction represents a particle which is equally .56) where k = p/¯ > 0 and ω = E/¯ > 0. can be written ¯ ψ(x.45).50) yield the following dispersion relation for such waves: ¯ k2 h ω= . Here. As we have seen.36)]. h (3. vp = ω . it seems plausible that the closely related relation (3. 2m (3. hence. is always parallel to its wavevector. its velocity. Does this imply that the dispersion relation (3. ω and k are linked via the dispersion relation h h (3.51) where p is the particle’s vector momentum.46) and (3. the above relation generalizes to give p = ¯ k. t)| 2 is proportional to the probability density of ﬁnding the particle at position x at ¯ time t. t) = ψ e i (k x−ω t) . we saw earlier that a plane wave propagates at the so-called phase velocity.54) However. (3.12 Wave Packets The above discussion suggests that the wavefunction of a massive particle of momentum p and energy E. 3. From before. and k its wavevector. If this is the case. k p . from Eq. according to the above dispersion relation. and. For the case of a particle moving the three dimensions. the most reasonable physical interpretation of the wavefunction is that |ψ(x. (3. moving in the positive x-direction.

4 (∆x) 2 (3.. For instance.. (3. suppose that at t = 0 the wavefunction of our particle takes the form ψ(x. ∞ ψ(x. f(k) is known as the Fourier transform of the function f(x).e.15. and very unlikely to yield a value which differs from x0 by more than 3 ∆x.60) is the wavefunction of a particle which is initially localized around x = x0 in .e. the superposition (3.Wave-Particle Duality 27 likely to be found anywhere on the x-axis at all times..59) f(k) 2π −∞ ¯ Here.61) This particular probability distribution is called a Gaussian distribution. which states that if ∞ 1 ¯ f(x) = √ f(k) e i k x dk. Incidentally. Hence. it is possible to add two valid wave solutions to form a third valid wave solution. this is indeed the case. ψ(k) represents the complex amplitude of plane waves of wavenumber k in this combination.60) Thus. 0)| 2 ∝ exp − (x − x0 ) 2 . (3. It can be seen that a measurement of the particle’s position is most likely to yield the value x0 . We can use Fourier’s ¯ theorem to ﬁnd the k-space function ψ(k) which generates any given x-space wavefunction ψ(x) at a given time. t) = −∞ ¯ ψ(k) e i (k x−ω t) dk. 3. we are relying on the assumption that particle waves are superposable: i. 3. Hence.58) 2π −∞ then ∞ ¯ = √1 f(x) e−i k x dx. a plane wave which varies as exp[i (k x − ω t)] and has a negative k (but positive ω) propagates in the negative x-direction at the phase velocity ω/|k|. known as Fourier’s theorem.57) includes both forward and backward propagating waves. and is plotted in Fig. how can we write the wavefunction of a particle which is localized in x: i. (3. The ultimate justiﬁcation for this assumption is that particle waves satisfy a differential wave equation which is linear in ψ. there is a useful mathematical theorem.57) ¯ Here. (3. As we shall see. So. Thus.e.7. 0) ∝ exp i k0 x − (x − x0 ) 2 . In writing the above expression. Now. a particle which is more likely to be found at some positions on the x-axis than at others? It turns out that we can achieve this goal by forming a linear combination of plane waves of different wavenumbers: i. 2 (∆x) 2 (3. in Sect. the initial probability density of the particle is written |ψ(x. the fact that the maxima and minima of the wavefunction propagate at a phase velocity which does not correspond to the classical particle velocity does not have any real physical consequences.

Now. ∞ ψ(x. (3. 0) e−i k x dx.60).62) Hence. we obtain ∞ ψ(x.28 QUANTUM MECHANICS Figure 3. −∞ 2 (3. −∞ (3. The above equation can be rearranged to give ∞ 2 ¯ ψ(k) ∝ e−i k x0 −β /4 e−(y−y0 ) dy.7: A Gaussian probability distribution in x-space. this reduces to ∞ ¯ ψ(k) ∝ e−i k x0 exp −i β y − y2 dy. (3. according to Eq. This type of wavefunction is known as a wave packet. we can employ Fourier’s theorem to invert this expression to give ∞ ¯ ψ(k) ∝ Making use of Eq. 0) = −∞ ¯ ψ(k) e i k x dk. 4 (∆x)2 (3. (3. −∞ (3.65) where β = 2 (k − k0 ) ∆x. some region whose width is of order ∆x.57).64) Changing the variable of integration to y = (x − x0 )/(2 ∆x).63) ¯ ψ(k) ∝ e−i (k−k0 ) x0 exp −i (k − k0 ) (x − x0 ) − −∞ (x − x0 )2 dx.66) .

e. where 1 ∆x ∆k = . 3. See Eq. (3.. if we only combine plane waves whose wavenumbers differ by a small amount (i. 3. according to Eq.e. if ∆k is small) then the resulting wave packet will be very extended in x-space (i. in Eq.68) 2 ∆x Now. According to Eq. p). how does this wavefunction evolve in time? Well. (3..e.61) and Fig. as can easily be seen by making the change of variable z = y − y0 . The integral now just reduces to a number. a Gaussian is the only mathematical function in x-space which has the same form as its Fourier transform in k-space. (Recall that p = ¯ k. if ∆x is small) then we need to combine plane waves with a very wide range of different k-values (i. if we wish to construct a packet which is very localized in x-space (i. we obtain (k − k0 ) 2 ¯ ψ(k) ∝ exp −i k x0 − . we have ∞ ¯ ψ(x.e.61)] transforms to a Gaussian probability distribution of characteristic width ∆k in k-space [see Eq. Hence. 2 (∆k) 2 (3.67) 1 .67). Conversely. (3.69) Note that this probability distribution is a Gaussian in k-space.7. 4 (∆k)2 where (3.. (3. Namely. (3.57). if |ψ(x)| 2 is proportional to the probability density of a measurement of the par¯ ticle’s position yielding the value x then it stands to reason that |ψ(k)| 2 is proportional to the probability density of a measurement of the particle’s wavenumber yielding the value k.70) This illustrates an important property of wave packets. (3.Wave-Particle Duality 29 where y0 = −i β/2. ∆k = ¯ |ψ(k)| 2 ∝ exp − (k − k0 ) 2 . how to write the wavefunction of a particle which is initially localized in x-space. t) = ψ(k) e i φ(k) dk. and very unlikely to yield a value which differs from k0 by more than 3 ∆k.71) −∞ . We have just seen that a Gaussian probability distribution of characteristic width ∆x in x-space [see Eq.69)]. Hence. 2 (3. (3. k.60). so a measurement of the particle’s wavenumber.13 Evolution of Wave Packets We have seen. is equivalent to h a measurement of the particle’s momentum. ∆x will be large). a measurement of k is most likely to yield the value k0 . (3.. But. ∆k will be large). Incidentally.

75) (3. The above expression can be rearranged to give ∞ ψ(x. QUANTUM MECHANICS (3. ′ φ0 = ′′ φ0 1 ′′ ′ ¯ ψ(k) exp i φ0 + φ0 (k − k0 ) + φ0 (k − k0 ) 2 2 −∞ ∞ .63) and (3. (3.30 where φ(k) = k x − ω(k) t. (3. t) ∝ e i (k0 x−ω0 t) where e i β1 y−(1+i β2 ) y dy.83) . vg = dω(k0 ) . rearranging.69). −∞ 2 (3. according to Eq. dk2 (3. β2 = 2 α (∆k) 2 t. |ψ(k)| is strongly peaked around k = k0 .72) ¯ The function ψ(k) is obtained by Fourier transforming the wavefunction at t = 0. ∆k = 1/(2 ∆x).82) Incidentally. it is a reasonable approximation to Taylor expand φ(k) about k0 . dk d2 φ(k0 ) = = −α t. (3. we obtain ψ(x. See ¯ Eqs.78) (3.79) Substituting from Eq.80) β1 = 2 ∆k (x − x0 − vg t).73) (3. Keeping terms up to second-order in k − k0 . we get ∞ ψ(x.74) (3. dk2 with ω0 = ω(k0 ).77) (3. dk d2 ω(k0 ) α = . (3. Thus.67).67).81) (3. Now. t) ∝ where φ0 = φ(k0 ) = k0 x − ω0 t. t) ∝ e i (k0 x−ω0 t)−(1+i β2 ) β 2 /4 e−(1+i β2 ) (y−y0 ) dy.76) dφ(k0 ) = x − vg t. where ∆x is the initial width of the wave packet. (3. and then changing the variable of integration to y = (k − k0 )/(2 ∆k). −∞ 2 (3.

According to Eq.53) for particle waves that vg = p . dk (3.87) 2 σ 2 (t) Hence.84) The integral now just reduces to a number. vg = dω/dt. Hence.85).86) Note that the above wavefunction is identical to our original wavefunction (3. justiﬁes the approximation which we made earlier by Taylor expanding the phase factor φ(k) about k = k0 .60) at t = 0. vp = ω/k. (3. t)| ∝ σ (t) exp − .85) (3.90) However. (3. Now. Again changing the variable of integration to z = (1 + i β2 )1/2 (y − y0 ). as soon as we realize that individual particles must be identiﬁed with wave packets rather than plane waves. m (3.89) (3.53) turns out to be consistent with classical physics. the particle’s most likely position is given by x = x0 + vg t. This. whereas a wave packet travels at the group velocity. It can be seen that the particle effectively moves at the uniform velocity vg = dω .88) which is known as the group velocity. −∞ 2 (3. it can be seen from Eq. . Thus. it follows from the dispersion relation (3. (3.Wave-Particle Duality 31 where y0 = i β/2 and β = β1 /(1 + i β2). Now. Hence. In fact. the probability distribution is a Gaussian. of characteristic width σ. we obtain ψ(x. t) ∝ where σ2 (t) = (∆x) 2 + exp i (k0 x − ω0 t) − (x − x0 − vg t)2 {1 − i 2 α (∆k) 2 t}/(4 σ2) [1 + i 2 α (∆k) 2 t]1/2 α2 t2 . In other words. the dispersion relation (3.45) that this is identical to the classical particle velocity. which peaks at x = x0 + vg t. 4 (∆x) 2 . the most likely position of our particle coincides with the peak of the distribution function. we get ∞ ψ(x. a plane wave travels at the phase velocity. after all. t) ∝ (1 + i β2 )−1/2 e i (k0 x−ω0 t)−(1+i β2 ) β 2 /4 e−z dz. (3. the probability density of our particle as a function of time is written (x − x0 − vg t) 2 2 −1 |ψ(x. a plane wave is usually interpreted as a continuous stream of particles propagating in the same direction as the wave.

However. is equivalent to a measurement of its momentum. Hence.53) and (3.79) that the characteristic time for a wave packet of original width ∆x to double in spatial extent is t2 ∼ m (∆x)2 . (3.86). t) from Eq. It follows that light pulses propagate through a vacuum without spreading. Another property of linear dispersion relations is that the phase velocity. from the previous analysis. a particle wave packet which is initially localized in x-space with characteristic width ∆x is also localized in kspace with characteristic width ∆k = 1/(2 ∆x).. hence.53) for particle waves is not linear in k. (3. Now. (3. Note. However.e. Hence.67) at all times. the second derivative of ω with respect to k is zero. for the special case where ω is a linear function of k. our previous analysis obtained ψ(x. it follows from Eqs. Now. It follows from the above inequality that h ∆x ∆p > ∼ ¯ h . See Eqs. a measurement of a particle’s wavenumber. respectively.e.93) . (3. [After all. Thus. ∆x ∼ 10−10 m) then the doubling time is only about 10−16 s. in general. and particle wave packets also gradually disperse as time progresses. (3..71). there is no dispersion of wave packets: i. 3.86). This is why a functional relationship between ω and k is generally known as a dispersion relation: i. 2 (3. but assumed that ¯ ψ(k) was given by Eq.91) For instance. that the rate of spreading of a wave packet is ultimately governed by the second derivative of ω(k) with respect to k. and the group velocity. whilst that of the wave packet in k-space stays the same. Of course. if an electron is originally localized in a region of atomic scale (i. ¯ h (3.14 Heisenberg’s Uncertainty Principle According to the analysis contained in the previous two sections.32 QUANTUM MECHANICS According to Eq. particle wave packets (for freely moving particles) spread very rapidly. 2 (3. we can think of ∆x and ∆k as characterizing our uncertainty regarding the values of the particle’s position and wavenumber. the width of the wave packet in x-space increases. since p = ¯ k. the dispersion relation (3. and. both plane light waves and light pulses propagate through a vacuum at the characteristic speed c = 3 × 108 m/s.24) for light waves is linear in k. the width of our wave packet grows as time progresses. as time progresses. Evidently.] Hence.79) and (3. vp = ω/k. particle plane waves and particle wave packets propagate at different velocities. k. because it governs how wave packets disperse as time progresses.92) Furthermore. p. we can say that ∆x ∆k > ∼ 1 . wave packets propagate without changing shape. Indeed. vg = dω/dk. an uncertainty in k of order ∆k translates to an h uncertainty in p of order ∆p = ¯ ∆k. are identical.. the dispersion relation (3.e.

8: Heisenberg’s microscope. Figure 3. Since the initial uncertainty in the particle’s position is ∆x. the spreading of a particle wavefunction must be interpreted as an increase in our uncertainty regarding the particle’s position. (3. It can be seen from Eqs. a good knowledge of the particle’s position implies a poor knowledge of its momentum. This translates to an uncertainty in velocity of ∆v = h ¯ /(m ∆x).53). then the wavefunction will spread as time progresses. and h others at v0 − ∆v/2. (3. at large t we expect the width of the wavefunction to be ¯t h σ ∼ ∆v t ∼ . m ∆x (3. where v0 is the mean propagation velocity. it is impossible to simultaneously measure the position and momentum of a particle (exactly). Evidently.95) m ∆x which is identical to Eq. this would only be possible . if we imagine that parts of the wavefunction propagate at v0 + ∆v/2.Wave-Particle Duality D lens 33 f y x scattered photon θ α electron incoming photon Figure 3. and (3.94). Note that the uncertainty principle is a direct consequence of representing particles as waves. (3. Indeed. Suppose that we try to image an electron using a simple optical system in which the objective lens is of diameter D and focal-length f. (3. (In practice. rather than an increase in the spatial extent of the particle itself.86) that at large t a particle wavefunction of original width ∆x (at t = 0) spreads out such that its spatial extent becomes σ∼ ¯t h . This is the famous Heisenberg uncertainty principle.8 illustrates a famous thought experiment known as Heisenberg’s microscope. Indeed.79). ﬁrst proposed by Werner Heisenberg in 1927. it follows that the uncertainty in its momentum is of order ¯ /∆x. and vice versa.94) It is easily demonstrated that this spreading is a consequence of the uncertainty principle. Thus. According to this principle.

It follows that the uncertainty in the electron’s momentum is 4π ¯ α h ∆p ≃ 2 ¯ k sin α ≃ h .101) λ . the magnitude of the k-vector is assumed to be the same before and after scattering. by using short wavelength radiation. (3. which is where the minimum resolving power is achieved. and the scattered photon will still be collected by the imaging system. 3. Assuming that α is small. k cos θ). where λ is the wavelength of the light illuminating the electron. it is impossible to measure any property of a microscopic dynamical system without . First.8.e. Let the wavevector of the incoming photon have the (x. (3. See Fig. then this translates to a uncertainty in the electron’s transverse position of ∆x ≃ f However. (3. y) components (k. (3. This is exactly the opposite of what we need to do to reduce the uncertainty in the position. By momentum conservation..98) 2f so λ . we are ignoring any wavelength shift of the photon on scattering— i.96) which is essentially the uncertainty principle. 0).97) f where α is the half-angle subtended by the lens at the electron.. This translates to a change in the photon’s x-component of momentum of ∆px = ¯ k (sin θ − h 1). However. as shown in the ﬁgure. If the scattered photon subtends an angle θ with the center-line of the optical system. D (3. then its wavevector is written (k sin θ. the electron is imaged when it scatters an incoming photon towards the objective lens. the uncertainty principle follows from two facts. D/2 . and a wide-angle lens. Thus. Here. θ can range all the way from −α to +α. Let us now examine Heisenberg’s microscope from a quantum mechanical point of view.) It is a well-known result in optics that such a system has a minimum angular resolving power of λ/D.34 QUANTUM MECHANICS using extremely short wavelength light. (3. we obtain tan α = ∆x ∆p ∼ h. we can write D α≃ . According to quantum mechanics. According to Heisenberg’s microscope.e. If the electron is placed at the focus of the lens. Multiplying the previous two equations. the change in the x-component of the photon’s wavevector is ∆kx = k (sin θ−1). the electron’s x-momentum will change by an equal and opposite amount.100) λ Note that in order to reduce the uncertainty in the momentum we need to maximize the ratio λ/α.99) ∆x ≃ 2α It follows that we can reduce the uncertainty in the electron’s position by minimizing the ratio λ/α: i.

102) ¯ where ψ(k) is determined by ψ(x.104) ∞ ¯ (−k2 ) ψ(k) e i (k x−ω t) dk. Multiplying through by ¯ . t) = −∞ ¯ ψ(k) e i (k x−ω t) dk. −∞ ∞ (3. Second.103). 0). and ω(k) = Now. ω− −∞ ¯ k2 ¯ h ψ(k) e i (k x−ω t) dk = 0.108) is only applicable to freely moving particles. Eq.102).102) that ∂ψ = ∂x and ∂2 ψ = ∂x2 ∂ψ = ∂t i ∂ψ ¯ ∂2 ψ h + = ∂t 2 m ∂x2 ∞ ¯ k2 h . (3. partial differential o o equation which governs the time evolution of a particle wavefunction. It is plausible. there is a limit to how small we can make the aforementioned disturbance. we h obtain ∂ψ ¯ 2 ∂2 ψ h i¯ h =− .106) Thus.108) is .105) whereas ¯ (−i ω) ψ(k) e i (k x−ω t) dk. that we can identify k with the differential operator −i ∂/∂x. from Eq. −∞ ∞ (3. (3. second-order. Thus. −∞ (3. Hence. and is generally easier to solve than the integral equation (3. 2m (3. 2m (3.104). since it was ﬁrst introduced by Erwin o Schr¨dinger in 1925. it follows from Eq.108) ∂t 2 m ∂x2 This expression is known as Schr¨dinger’s equation. particle and light energy and momentum are quantized. (3. the differential operator on the right-hand side of Eq. Hence. (3. Fortunately. Of course.15 Schr¨ dinger’s Equation o We have seen that the wavefunction of a free particle of mass m satisﬁes ∞ ψ(x.107) where use has been made of the dispersion relation (3. (3. (3. 3. Schr¨dinger’s equation is a linear. it is fairly easy to guess the generalization of this equation for particles moving in some potential V(x). there is an irreducible uncertainty in certain measurements which is a consequence of the act of measurement itself.Wave-Particle Duality 35 disturbing the system somewhat.103) ¯ (i k) ψ(k) e i (k x−ω t) dk.

109) 2 2 m ∂x 2 m ∂x2 This leads to the general form of Schr¨dinger’s equation: o i¯ h ∂ψ ¯ 2 ∂2 ψ h =− + V(x) ψ.13. and obtain the value x0 . as soon as the wavefunction has collapsed. According to our usual interpretation.9. and has no chance of giving any other value. This is illustrated in Fig. We conclude that there are two types of time evolution of the wavefunction in quantum mechanics. If we make another measurement immediately after the ﬁrst one then what value do we expect to obtain? Well. a measurement of the particle’s position is certain to give the value x0 . Second. |ψ(x. If the wavefunction is extended then there is a wide range of likely values that this measurement could give. t). Calculate . in order to guarantee that the same result will be obtained. This implies that the wavefunction must have collapsed to some sort of “spike” function located at x = x0 . Of course. But. the particle’s energy is written p2 /(2 m) + V. h h which is just the energy of a freely moving particle. A He-Ne laser emits radiation of wavelength λ = 633 nm.16 Collapse of the Wave Function Consider an extended wavefunction ψ(x. We now know that the particle is located at x = x0 . immediately after the ﬁrst measurement. Thus. This evolution takes place between measurements. 3.36 QUANTUM MECHANICS equivalent to ¯ 2 k2 /(2 m). in the presence of a potential V(x). The above discussion illustrates an important point in quantum mechanics. that the wavefunction of a particle changes discontinuously (in time) whenever a measurement is made. 3. Suppose that we make such a measurement. the second measurement must be made reasonably quickly after the ﬁrst. since the particle cannot have shifted position appreciably in an inﬁnitesimal time interval. x0 . there o is a discontinuous evolution which takes place each time a measurement is made. However. it starts to expand again. How many photons are emitted per second by a laser with a power of 1 mW? What force does such laser exert on a body which completely absorbs its radiation? 2. Thus. First. Namely.110) 3. Exercises 1. Thus. ∂t 2 m ∂x2 (3.60 eV (1 eV is the energy acquired by an electron accelerated through a potential difference of 1 V). p = ¯ k. (3. there is a smooth evolution which is governed by Schr¨dinger’s equation. the operator is also equivalent to p2 /(2 m). as discussed in Sect. Thus. it seems reasonable to make the substitution ¯ 2 ∂2 h ¯ 2 ∂2 h − →− + V(x). The ionization energy of a hydrogen atom in its ground state is Eion = 13. common sense tells us that we must obtain the same value. t)| 2 is proportional to the probability density of a measurement of the particle’s position yielding the value x at time t.

frequently emit electrons with energies of 1–10 MeV.9: Collapse of the wavefunction upon measurement of x. and wavenumber of the electromagnetic radiation which will just ionize the atom. 4. 6. . Use the uncertainty principle to show that electrons of energy 1 MeV could not be contained in the nucleus before the decay. The maximum energy of photoelectrons from aluminium is 2. typically of size 10−14 m. where V is measured in volts. 5.29 × 10−9 V −1/2 m. Nuclei. Show that the de Broglie wavelength of an electron accelerated from rest across a potential difference V is given by λ = 1. and 0. What are the group and phase velocities of such waves as functions of ν0 and λ? 7. Use this data to calculate Planck’s constant.5 kV would be required to produce an electron diffraction pattern from the crystal.3 eV for radiation of wavelength ˚ ˚ 2000 A. The relationship between wavelength and frequency for electromagnetic waves in a waveguide is c .Wave-Particle Duality 37 BEFORE |ψ|2 → x→ AFTER |ψ|2 → x0 x→ Figure 3. λ= 2 ν2 − ν0 where c is the velocity of light in vacuum. the frequency. and the work function of aluminium. wavelength. 3. If the atoms in a regular crystal are separated by 3×10−10 m demonstrate that an accelerating voltage of about 1.90 eV for radiation of wavelength 2580 A.

h QUANTUM MECHANICS where A and a are positive real constants. For what potential function V(x) does ψ satisfy the Schr¨dinger equation? o .38 8. t) = A exp[−a (m x2 /¯ + i t)]. A particle of mass m has a wavefunction ψ(x.

According to Eq.1 Introduction The previous chapter serves as a useful introduction to many of the basic concepts of quantum mechanics.2 Schr¨ dinger’s Equation o Consider a dynamical system consisting of a single non-relativistic particle of mass m moving along the x-axis in some real potential V(x). ∂t 2 m ∂x2 (4. we shall examine these concepts in a more systematic fashion. t). t)| 2 dx.Fundamentals of Quantum Mechanics 39 4 Fundamentals of Quantum Mechanics 4. t)| 2 dx = 1. we shall concentrate on one-dimensional systems.3 Normalization of the Wavefunction Now. or ∞ |ψ(x. It follows that Px ∈ −∞:∞ = 1. −∞ (4.2). An outcome of a measurement which has a probability 0 is an impossible outcome. t)| 2 dx. whereas an outcome which has a probability 1 is a certain outcome.4) . In this chapter. Thus.2) Note that this quantity is real and positive deﬁnite.1) The wavefunction is interpreted as follows: |ψ(x. This wavefunction evolves in time according to Schr¨dinger’s equation: o i¯ h ∂ψ ¯ 2 ∂2 ψ h =− + V(x) ψ. t)| 2 is the probability density of a measurement of the particle’s displacement yielding the value x. In quantum mechanics. the instantaneous state of the system is represented by a complex wavefunction ψ(x. 4. the probability of a measurement of x yielding a result between −∞ and +∞ is ∞ Px ∈ −∞:∞ (t) = |ψ(x.3) However. the probability of a measurement of the displacement giving a result between a and b (where a < b) is b Px ∈ a:b (t) = |ψ(x. a (4. (4. 4. a probability is a real number between 0 and 1. a measurement of x must yield a value between −∞ and +∞. since the particle has to be located somewhere. For the sake of simplicity. −∞ (4.

unity) to change in time. multiplying Schr¨dinger’s equation by ψ∗ /(i ¯ ). (4..7) ∞ e−y dy = −∞ 2 √ π.8) which implies that |ψ0 | 2 = 1 . and of characteristic width σ (see Sect. suppose that we wish to normalize the wavefunction of a Gaussian wave packet. For example.11) for wavefunctions satisfying Schr¨dinger’s equation. a general normalized Gaussian wavefunction takes the form ψ(x) = eiϕ 2 2 e−(x−x0 ) /(4 σ ) . (4. we simply substitute Eq. The above equation gives o d dt ∞ ∞ ψ∗ ψ dx = −∞ −∞ ∂ψ∗ ∂ψ ψ + ψ∗ ∂t ∂t dx = 0.12) Now.9) Hence. we require that d dt ∞ |ψ(x. (4. t)| 2 dx = 0. manifestly. since it does not make sense for the probability that a measurement of x yields any possible outcome (which is. Now.5) into Eq. 2 ∂t 2m ∂x ¯ h (4.40 QUANTUM MECHANICS which is generally known as the normalization condition for the wavefunction.10) where ϕ is an arbitrary real phase-angle. (4. If this is not o the case then the probability interpretation of the wavefunction is untenable.4). centered on x = x0 .e. (4.13) .6) √ Changing the variable of integration to y = (x − x0 )/( 2 σ). we get √ |ψ0 | 2 2 σ However. −∞ (4. ∞ e−y dy = 1. Hence. to obtain ∞ |ψ0 | 2 −∞ e−(x−x0 ) 2 /(2 σ2 ) dx = 1. we obtain o h ψ∗ i ¯ ∗ ∂2 ψ h i ∂ψ = ψ − V |ψ| 2 . (2π σ2)1/2 (4. −∞ 2 (4. (4. 2 )1/4 (2π σ (4.5) In order to determine the normalization constant ψ0 . 3. ψ(x) = ψ0 e−(x−x0 ) 2 /(4 σ2 ) . it is important to demonstrate that if a wavefunction is initially normalized then it stays normalized as it evolves in time according to Schr¨dinger’s equation.12): i.

. −∞ (4.. A∗ ∗ get ∂ψ∗ i¯ h ∂2 ψ∗ i =− ψ + V |ψ| 2 (4.18) is a probability conservation equation.16) The above equation is satisﬁed provided However. and i∗ = −i]. For such wavefunctions. j(b. we are interpreting j(x. Note. t) = ψ0 e i (k x−ω t) (4. dt is deﬁned in Eq.12) and (4. t) = 0.Fundamentals of Quantum Mechanics The complex conjugate of this expression yields 41 [since (A B)∗ = A∗ B∗ .e.4) is satisﬁed at one instant in time then it is satisﬁed at all subsequent times.. a (4.2). It is also possible to demonstrate. are such that the integral in Eq. this is a necessary condition for the integral on the left-hand side of Eq. (4.4) to converge. (4. that dPx ∈ a:b + j(b.18) |ψ| → 0 as |x| → ∞. we conclude that all wavefunctions which are square-integrable [i. and. a plane wave wavefunction ψ(x. (4.e. t)| 2 dx. −ψ 2 m ∂x ∂x ∂x ∞ ∂ψ∗ ∂2 ψ ∂ψ i¯ h ∂2 ψ∗ ψ∗ ψ + ψ∗ = −ψ ∂t ∂t 2m ∂x2 ∂x2 d dt ∞ = (4.. and that out of the interval [i.4).19) is known as the probability current. (4.e. thus. unless otherwise stated. t)]. According to this equation. all wavefunctions are assumed to be square-integrable and normalized. Note that j is real. we ψ ∂ψ i¯ ∂ h ∂ψ∗ ψ∗ . ﬁnally. j(a. and j(x. t) − j(a. t) = i¯ h ∂ψ∗ ∂ψ ψ − ψ∗ 2m ∂x ∂x (4. t) as the ﬂux of probability in the +x-direction at position x and time t. cannot be normalized. (4.17) where Px ∈ a:b (4. the probability of a measurement of x lying in the interval a to b evolves in time due to the difference between the ﬂux of probability into the interval [i. via very similar analysis to the above.20) is not square-integrable. Here. t)].15) Equations (4.4) converges] have the property that if the normalization condition (4. the best we can say is that b Px ∈ a:b (t) ∝ |ψ(x. Summing the previous two equations.21) In the following. Equation (4. Hence.14) 2 ∂t 2m ∂x ¯ h = A. that not all wavefunctions can be normalized according to the scheme set out in Eq.15) can be combined to produce ∂ψ i¯ h ∂ψ∗ ψ∗ |ψ| dx = −ψ 2m ∂x ∂x −∞ 2 = 0. For instance.

from the deﬁnition of probability.25) The expectation value of x associated with this wavefunction is x =√ 1 2π σ2 ∞ x e−(x−x0 ) −∞ 2 /(2 σ2 ) dx.27) However.23) In general.24) which is known as the variance of x.e.. We generally expect the results of measurements of x to lie within a few standard deviations of the expectation value. (4. consider the normalized Gaussian wave packet [see Eq. the value of x which maximizes |ψ| 2 ). 2 )1/4 (2π σ (4. t)| 2 is the probability density of a measurement of a particle’s displacement yielding the value x at time t. Similarly the expectation value of any function of x is ∞ f(x) = −∞ f(x) |ψ| 2 dx.26) √ Let y = (x − x0 )/( 2 σ). (4. the results of the various different measurements of x will be scattered around the expectation value x . −∞ 2 (4. x is called the expectation value of x. In general. (4. is called the standard deviation of x. The square-root of this quantity.42 QUANTUM MECHANICS 4. Suppose that we made a large number of independent measurements of the displacement on an equally large number of identical quantum systems. . Hence. by symmetry.28) Evidently. (4. making use of Eq. However. the mean of all these results is simply ∞ x = −∞ x |ψ| 2 dx. σx .22) Here. It follows that x0 x =√ π ∞ e −∞ −y2 √ 2σ dy + √ π ∞ y e−y dy. the second integral on the right-hand side is zero.10)] ψ(x) = eiϕ 2 2 e−(x−x0 ) /(4 σ ) . For instance. (4.4 Expectation Values and Variances We have seen that |ψ(x.8). (4. the expectation value of x for a Gaussian wave packet is equal to the most likely value of x (i. (4. measurements made on different systems will yield different results. we obtain x = x0 . The degree of scatter is parameterized by the quantity ∞ σ2 = x −∞ (x − x ) 2 |ψ| 2 dx ≡ x2 − x 2 .

. ∞ ∞ ∂j p = −m dx = m x j dx. where j is the probability current deﬁned in Eq. (4.5 Ehrenfest’s Theorem A simple way to calculate the expectation value of momentum is to evaluate the time derivative of x . It follows that σ2 x However. Thus. −∞ 2 (4.25) is σ2 = √ x 1 2π σ2 ∞ 43 (x − x0 ) 2 e−(x−x0 ) −∞ 2 /(2 σ2 ) dx.29) √ Let y = (x − x0 )/( 2 σ). 3. (4.18)].36) −∞ ∂x −∞ where we have integrated by parts.33) 4.34) However. It follows from Eq. p =m dx d =m dt dt ∞ ∞ x |ψ| 2 dx = m −∞ −∞ x ∂|ψ| 2 dx.32) This result is consistent with our earlier interpretation of σ as a measure of the spatial extent of the wave packet (see Sect. it is easily demonstrated that ∂j ∂|ψ| 2 + =0 ∂t ∂x (4.37) .19) that p =− i¯ h 2 ∞ ψ∗ −∞ ∂ψ ∂ψ∗ − ψ dx = −i ¯ h ∂x ∂x ∞ ψ∗ −∞ ∂ψ dx. (4.Fundamentals of Quantum Mechanics The variance of x associated with the Gaussian wave packet (4. (4.e. and then multiply by the mass m: i. ∞ 2 σ2 = √ π ∞ y2 e−y dy.35) [this is just the differential form of Eq. 2 (4. ∂t (4. 2 )1/4 (2π σx (4. It follows that we can rewrite the Gaussian wave packet (4. ∂x (4.25) in the convenient form ψ(x) = eiϕ 2 2 e−(x− x ) /(4 σx ) .30) y e −∞ 2 −y2 dy = √ π . (4.12).31) giving 2 σx = σ2 . (4.19).

42) (4. Suppose that the potential V(x) is slowly varying. and o simplifying.41). (4. ∂x (4. (4.38) ψ∗ dt ∂x −∞ It follows from the above that dp dt ∞ = −i ¯ h −∞ ∞ ∂2 ψ ∂ψ∗ ∂ψ dx + ψ∗ ∂t ∂x ∂t∂x ∗ = −∞ ∂ψ i¯ h ∂t ∂ψ ∂ψ ∂ψ∗ i¯ h + ∂x ∂x ∂t dx.44) (4. This result is known as Ehrenfest’s theorem. dt dx 2 dx3 (4.40) Integration by parts yields dp =− dt ∞ −∞ dV dV |ψ| 2 dx = − dx dx . we can expand dV/dx as a Taylor series about x .39) where we have integrated by parts. we obtain dp = dt ∞ − −∞ ∂|ψ| 2 ¯ 2 ∂ ∂ψ∗ ∂ψ h + V(x) dx = 2 m ∂x ∂x ∂x ∂x ∞ V(x) −∞ ∂|ψ| 2 dx. dV dx .45) 2 since 1 = 1. according to Eqs.34) and (4. Keeping terms up to second order. (4. m dx dt dp dt = p.41) Hence. (4. the expectation values of displacement and momentum obey time evolution equations which are analogous to those of classical mechanics.1). the expectation value of the momentum can be written ∞ dx ∂ψ p =m = −i ¯ h dx.44 QUANTUM MECHANICS where we have again integrated by parts.43) yields 2 dp dV( x ) σx dV 3 ( x ) =− − . (4. (4. and (x − x ) 2 = σx . In this case. we obtain dV( x ) dV 2 ( x ) 1 dV 3 ( x ) dV(x) = + (x − x ) + (x − x ) 2 .43) = − Evidently. Substituting from Schr¨dinger’s equation (4. and x − x = 0. 2 3 dx dx dx 2 dx Substitution of the above expansion into Eq. Hence. The ﬁnal term on the righthand side of the above equation can be neglected when the spatial extent of the particle .

in general. and a ﬁnal f(x) is understood [where f(x) is a general function]. since x f(x) is a different function to f(x).42) and (4. O (say). since df(x)/dx is a different function to f(x). Now. Of course. dx (4. an operator. (4. since we know that classical mechanics gives the correct answer in this limit.43) reduce to m dx dt dp dt = p. (4. dV( x ) . some operators do commute: e. Of course. d/dx is also an operator. is much smaller than the variation length-scale of the potential.6 Operators An operator. Hence. and is fully speciﬁed once f(x) is given. dx dx x (4. we conclude that quantum mechanics corresponds to classical mechanics in the limit that the spatial extent of the wavefunction (which is typically of order the de Boglie wavelength) is negligible. (4. operators do not commute. and c a general complex number. with x playing the role of the particle displacement. 4. x This can also be written df d = (x f) . is termed linear if O(c f(x)) = c O(f(x)). Furthermore.e.48) For instance.Fundamentals of Quantum Mechanics 45 wavefunction. (4.. x x2 = x2 x. O.49) d d = x. if the spatial extent of the wavefunction is negligible then a measurement of x is almost certain to yield a result which lies very close to x . (4. . x is an operator. Eqs.. σx .50) dx dx where the operators are assumed to act on everything to their right.46) (4. This is an important result. All of the operators employed in quantum mechanics are linear.51) Finally.52) where f is a general function. O(f(x)) → g(x). In this case. and is fully speciﬁed once f(x) is given..47) = − These equations are exactly equivalent to the equations of classical mechanics. is a mathematical entity which transforms one function into another: i.e.g. The above expression illustrates an important point: i.

56) Clearly. or ∞ ∞ ψ∗ (O ψ) dx = −∞ −∞ (O ψ)∗ ψ dx.38). displacement is represented by the algebraic operator x. in quantum mechanics.57) where O∗ is the complex conjugate of O. . So. are represented in quantum mechanics by linear operators which act on the wavefunction. Second.. Finally. and momentum by the differential operator −i ¯ ∂/∂x: i. there is no difference between x p and p x. is deﬁned as follows: ∞ ∞ ψ∗ (O ψ) dx = −∞ −∞ (O† ψ)∗ ψ dx. (4.53) (4.e. h p ≡ −i ¯ h ∂ . However. ∞ QUANTUM MECHANICS x p = −∞ ∞ ψ∗ x ψ dx. (1/2) [x p + (x p)† ] is Hermitian.e. the expectation value of some dynamical variable represented by the operator O(x) is simply ∞ O = −∞ ψ∗ (x. then (A B)† = B† A† . neither of these combinations is Hermitian. classical dynamical variables.. (4. Suppose that we wish to ﬁnd the operator which corresponds to the classical dynamical variable x p.22) and (4. ∂x (4. should be choose x p or p x? Actually. First. In classical mechanics. The Hermitian conjugate. if an operator is to represent a dynamical variable which has physical signiﬁcance then its expectation value must be real. if A and B are two operators. An operator which satisﬁes the above constraint is called an Hermitian operator. h ψ∗ −i ¯ −∞ (4. In other words. of a general operator. p† = p. For a non-Hermitian operator. we have already seen that x p = p x. However. which neatly resolves our problem of which order to put x and p.55) Finally. O† . t) dx. if the operator O represents a physical variable then we require that O = O ∗ .58) The Hermitian conjugate of an Hermitian operator is the same as the operator itself: i. and that the operator O + O† is Hermitian. O. it is easily demonstrated that (O† )† = O. ∂x These expressions suggest a number of things. (4. (1/2) [x p + (x p)†] = (1/2) (x p + p† x† ) = (1/2) (x p + p x). It is easily demonstrated that x and p are both Hermitian. O (say). Moreover. such as x and p. t) O(x) ψ(x.54) = ∂ ψ dx. from Eqs.46 Now. (4.

then a reasonable guess for the corresponding operator in quantum mechanics is (1/2) [O(x. and h H = i ¯ ∂/∂t. t) = √ . (4. the time-dependent Schr¨dinger equation can be written o i¯ h ∂ψ = H ψ. ψ(x.64) (4. if O(x. −∞ where k represents wavenumber. p. t) e−i p x/¯ dx.12). −∞ ∞ (4. p. where p = −i ¯ ∂/∂x. t) = √ 2π 1 ¯ ψ(k.65) ψ(x. t) e+i p x/¯ dp.66) 2π ¯ −∞ h ∞ 1 h √ φ(p. Now. ∂t (4. and energy. t) = √ 2π ∞ ¯ ψ(k. t) = ψ(x. 2 m ∂x2 ∂t H ≡ i¯ h (4. (4. applied to one-dimensional wavefunctions. t).67) 2π ¯ −∞ h √ ¯ h where φ(p.60) However. p.62) ∂ . E) is a classical dynamical variable which is a function of displacement.55) that H≡− ¯ 2 ∂2 h + V(x). 3. (4. p = ¯ k. H)]. we can also write h ∞ 1 h φ(p. However. ∂t Thus. momentum. t) = ψ(k.Fundamentals of Quantum Mechanics 47 It is a reasonable guess that the operator corresponding to energy (which is called the Hamiltonian. H) + O† (x.61) (4. Hence. h 4. it follows from Eq. t)/ ¯ is the momentum-space equivalent to the real-space wavefunction ψ(x. 2m (4. yields 1 ψ(x. according to Schr¨dinger’s equation. (4. t) e+i k x dk.7 Momentum Representation Fourier’s theorerm (see Sect.1). 2 m ∂x2 (4. and conventionally denoted H) takes the form H≡ p2 + V(x). t) e−i k x dx. we have o − so ¯ 2 ∂2 h ∂ + V(x) = i ¯ h .63) Finally.59) Note that H is Hermitian.

it is convenient to introduce a useful function called the Dirac deltafunction. (4. Consider the integral ∞ f(x) δ(x) dx.68). (4. by analogy with Eq. The former scheme is known as the momentum representation of quantum mechanics. but its integral is well-deﬁned. was ﬁrst devised by Paul Dirac.73) (4. (4. and dynamical variables are represented by different operators. φ(p. This function. the expectation value of some operator O(p) takes the form ∞ O = −∞ φ∗ (p. −∞ 1 2π ¯ h ∞ h e+i p (x−x0 )/¯ dp. Eq. to give ∞ ∞ f(x) δ(x) dx = f(0) −∞ −∞ δ(x) dx = f(0). ψ(x.70) where use has been made of Eq. (4.74) It turns out that we can just as well formulate quantum mechanics using momentumspace wavefunctions. t). and has the following rather unusual properties: δ(x) is zero for x = 0.71) which can also be thought of as an alternative deﬁnition of a delta-function. t) dp. Furthermore. A simple generalization of this result yields ∞ f(x) δ(x − x0 ) dx = f(x0 ).71) that h e−i p x0 /¯ φ(p) = √ . t) O(p) φ(p..67) and (4. δ(p − p0 ) = 1 2π ¯ h ∞ h e+i (p−p0 ) x/¯ dx. wavefunctions are the Fourier transforms of the equivalent real-space wavefunctions. t). we can safely replace f(x) by f(0) in the above integral (assuming f(x) is well behaved at x = 0). However.72) Hence.68) The delta-function is an example of what is known as a generalized function: i. In the momentum representation. (4.69) Since δ(x) is only non-zero inﬁnitesimally close to x = 0. (4. Suppose that ψ(x) = δ(x − x0 ). −∞ (4. the singularity at x = 0 is such that ∞ δ(x) dx = 1.48 QUANTUM MECHANICS At this stage. 2π ¯ h (4. denoted δ(x). as real-space wavefunctions.75) .56). −∞ (4. It follows from Eqs.e. −∞ (4. its value is not well-deﬁned at all x.66) yields the important result δ(x − x0 ) = Similarly. and is inﬁnite at x = 0. −∞ (4.

66). t) −i ¯ h 1 2π ¯ h ∞ ∞ ∞ ∂ ψ(x. t)| 2 can be interpreted as the probability density of a measurement of momentum yielding the value p at time t.. t) p φ(p. ∂p 1 = 2π ¯ h φ∗ (p ′. However.83) . t) dx dp dp ′.81) Hence. ′ = (4. (4.Fundamentals of Quantum Mechanics Consider momentum.76) −∞ −∞ −∞ where use has been made of Eq. It follows that φ(p.78) Evidently. We can write ∞ x = −∞ ψ∗ (x.74) and (4. t) φ(p.22)] that |φ(p.71). t) dx ∞ ∞ ∞ (4. −∞ (4. t) dp = −∞ p |φ| 2 dp. We can write ∞ 49 p = −∞ ψ∗ (x. making use of Eqs. The above expression also strongly suggests [by comparison with Eq. t) p e+i (p−p ) x/¯ dx dp dp ′ . momentum is represented by the operator p in the momentum representation. t)| 2 dp = 1. (4. using Eq. ∂p (4. (4.80) ∂ ′ h e+i (p−p ) x/¯ dx dp dp ′ .4): i. (4.71). t) dx ∂x h φ∗ (p ′ . it follows from Eq. (4.79) Consider displacement.77) Hence. t) e+i (p−p ) x/¯ i ¯ −∞ −∞ −∞ ′ ∂ φ(p. we obtain ∞ ∞ p = −∞ φ∗ (p. t) −i ¯ h −∞ −∞ −∞ Integration by parts yields x = 1 2π ¯ h ∞ ∞ ∞ h h φ∗ (p ′ . ∞ |φ(p.82) Evidently. t) must satisfy an analogous normalization condition to Eq.e. t) φ(p. we obtain x = 1 2π ¯ h ∞ φ∗ (p) i ¯ h −∞ ∂ φ(p) dp. t) φ(p. t) p δ(p − p ′) dp dp ′. t) x ψ(x.74) that ∞ ∞ p = −∞ −∞ φ∗ (p ′. (4. displacement is represented by the operator x ≡ i¯ h ∂ ∂p (4. (4. ∂p (4. (4.

50

QUANTUM MECHANICS

in the momentum representation. Finally, let us consider the normalization of the momentum-space wavefunction φ(p, t). We have

∞

ψ∗ (x, t) ψ(x, t) dx =

−∞

1 2π ¯ h

∞

∞

∞ h φ∗ (p ′, t) φ(p, t) e+i (p−p ) x/¯ dx dp dp ′ . (4.84)

′

−∞ −∞ −∞

**Thus, it follows from Eqs. (4.71) and (4.74) that
**

∞ ∞

|ψ(x, t)| dx =

−∞ −∞

2

|φ(p, t)| 2 dp.

(4.85)

Hence, if ψ(x, t) is properly normalized [see Eq. (4.4)] then φ(p, t), as deﬁned in Eq. (4.67), is also properly normalized [see Eq. (4.79)]. The existence of the momentum representation illustrates an important point: i.e., that there are many different, but entirely equivalent, ways of mathematically formulating quantum mechanics. For instance, it is also possible to represent wavefunctions as row and column vectors, and dynamical variables as matrices which act upon these vectors.

**4.8 Heisenberg’s Uncertainty Principle
**

Consider a real-space Hermitian operator O(x). A straightforward generalization of Eq. (4.57) yields

∞ ∞

ψ∗ (O ψ2 ) dx = 1

−∞ −∞

(O ψ1 )∗ ψ2 dx,

(4.86)

**where ψ1 (x) and ψ2 (x) are general functions. Let f = (A − A ) ψ, where A(x) is an Hermitian operator, and ψ(x) a general wavefunction. We have
**

∞ ∞ ∞

|f| 2 dx =

−∞ −∞

f∗ f dx =

−∞

[(A − A ) ψ] ∗ [(A − A ) ψ] dx.

(4.87)

**Making use of Eq. (4.86), we obtain
**

∞ ∞

|f| dx =

−∞ −∞

2

2 ψ∗ (A − A ) 2 ψ dx = σA ,

(4.88)

2 where σA is the variance of A [see Eq. (4.24)]. Similarly, if g = (B − B ) ψ, where B is a second Hermitian operator, then ∞ 2 |g| 2 dx = σB , −∞

(4.89)

**Now, there is a standard result in mathematics, known as the Schwartz inequality, which states that
**

b 2 b b

f (x) g(x) dx

a

∗

≤

|f(x)| 2 dx

a a

|g(x)| 2 dx,

(4.90)

**Fundamentals of Quantum Mechanics
**

where f and g are two general functions. Furthermore, if z is a complex number then 1 |z| = [Re(z)] + [Im(z)] ≥ [Im(z)] = (z − z∗ ) 2i

2 2 2 2 2

51

.

(4.91)

Hence, if z =

∞ f∗ −∞

**g dx then Eqs. (4.88)–(4.91) yield
**

2 2 σA σB ≥

1 (z − z∗ ) 2i

2

.

(4.92)

However,

∞ ∞

z=

−∞

[(A − A ) ψ] [(B − B ) ψ] dx =

−∞

∗

ψ∗ (A − A ) (B − B ) ψ dx,

(4.93)

**where use has been made of Eq. (4.86). The above equation reduces to
**

∞

z=

−∞

ψ∗ A B ψ dx − A B .

(4.94)

**Furthermore, it is easily demonstrated that
**

∞

z =

−∞

∗

ψ∗ B A ψ dx − A B .

(4.95)

Hence, Eq. (4.92) gives

2 σA 2 σB

≥

1 [A, B] 2i

2

,

(4.96)

where [A, B] ≡ A B − B A. (4.97) Equation (4.96) is the general form of Heisenberg’s uncertainty principle in quantum mechanics. It states that if two dynamical variables are represented by the two Hermitian operators A and B, and these operators do not commute (i.e., A B = B A), then it is impossible to simultaneously (exactly) measure the two variables. Instead, the product of the variances in the measurements is always greater than some critical value, which depends on the extent to which the two operators do not commute. For instance, displacement and momentum are represented (in real-space) by the operators x and p ≡ −i ¯ ∂/∂x, respectively. Now, it is easily demonstrated that h [x, p] = i ¯ . h Thus, σx σp ≥ ¯ h , 2 (4.98)

(4.99)

52

QUANTUM MECHANICS

which can be recognized as the standard displacement-momentum uncertainty principle (see Sect. 3.14). It turns out that the minimum uncertainty (i.e., σx σp = ¯ /2) is only h achieved by Gaussian wave packets (see Sect. 3.12): i.e., ψ(x) = φ(p) =

h e+i p0 x/¯ −(x−x0 ) 2 /4 σx 2 e , 2 )1/4 (2π σx h e−i p x0 /¯ −(p−p0 ) 2/4 σp 2 e , 2 )1/4 (2π σp

(4.100) (4.101)

where φ(p) is the momentum-space equivalent of ψ(x). Energy and time are represented by the operators H ≡ i ¯ ∂/∂t and t, respectively. h These operators do not commute, indicating that energy and time cannot be measured simultaneously. In fact, [H, t] = i ¯ , h (4.102) so σE σt ≥ This can be written, somewhat less exactly, as ∆E ∆t > ¯ , ∼ h (4.104) ¯ h . 2 (4.103)

where ∆E and ∆t are the uncertainties in energy and time, respectively. The above expression is generally known as the energy-time uncertainty principle. For instance, suppose that a particle passes some ﬁxed point on the x-axis. Since the particle is, in reality, an extended wave packet, it takes a certain amount of time ∆t for the particle to pass. Thus, there is an uncertainty, ∆t, in the arrival time of the particle. Moreover, since E = ¯ ω, the only wavefunctions which have unique energies are those h with unique frequencies: i.e., plane waves. Since a wave packet of ﬁnite extent is made up of a combination of plane waves of different wavenumbers, and, hence, different frequencies, there will be an uncertainty ∆E in the particle’s energy which is proportional to the range of frequencies of the plane waves making up the wave packet. The more compact the wave packet (and, hence, the smaller ∆t), the larger the range of frequencies of the constituent plane waves (and, hence, the large ∆E), and vice versa. To be more exact, if ψ(t) is the wavefunction measured at the ﬁxed point as a function of time, then we can write ∞ 1 h χ(E) e−i E t/¯ dE. (4.105) ψ(t) = √ 2π ¯ −∞ h In other words, we can express ψ(t) as a linear combination of plane waves of deﬁnite energy E. Here, χ(E) is the complex amplitude of plane waves of energy E in this combination. By Fourier’s theorem, we also have χ(E) = √ 1 2π ¯ h

∞ h ψ(t) e+i E t/¯ dt. −∞

(4.106)

Fundamentals of Quantum Mechanics

53

**For instance, if ψ(t) is a Gaussian then it is easily shown that χ(E) is also a Gaussian: i.e., ψ(t) = χ(E) =
**

h e−i E0 t/¯ −(t−t0 ) 2 /4 σt2 e , (2π σt2)1/4 h e+i E t0 /¯ −(E−E0 ) 2 /4 σ 2 E, e 2 (2π σE )1/4

(4.107) (4.108)

where σE σt = ¯ /2. As before, Gaussian wave packets satisfy the minimum uncertainty h principle σE σt = ¯ /2. Conversely, non-Gaussian wave packets are characterized by σE σt > h ¯ /2. h

**4.9 Eigenstates and Eigenvalues
**

Consider a general real-space operator A(x). When this operator acts on a general wavefunction ψ(x) the result is usually a wavefunction with a completely different shape. However, there are certain special wavefunctions which are such that when A acts on them the result is just a multiple of the original wavefunction. These special wavefunctions are called eigenstates, and the multiples are called eigenvalues. Thus, if A ψa (x) = a ψa (x), (4.109)

where a is a complex number, then ψa is called an eigenstate of A corresponding to the eigenvalue a. Suppose that A is an Hermitian operator corresponding to some physical dynamical variable. Consider a particle whose wavefunction is ψa . The expectation of value A in this state is simply [see Eq. (4.56)]

∞ ∞

A =

−∞

ψ∗ a

A ψa dx = a

−∞

ψ∗ ψa dx = a, a

(4.110)

**where use has been made of Eq. (4.109) and the normalization condition (4.4). Moreover,
**

∞ ∞ ∞

A2 =

−∞

ψ∗ A2 ψa dx = a a

−∞

ψ∗ A ψa dx = a2 a

−∞

ψ∗ ψa dx = a2 , a

(4.111)

**so the variance of A is [cf., Eq. (4.24)]
**

2 σA = A2 − A 2

= a2 − a2 = 0.

(4.112)

The fact that the variance is zero implies that every measurement of A is bound to yield the same result: namely, a. Thus, the eigenstate ψa is a state which is associated with a unique value of the dynamical variable corresponding to A. This unique value is simply the associated eigenvalue.

from Eq. a ψ∗ ψa ′ dx.122) In other words. which correspond to the two different eigenvalues a and a ′ .54 QUANTUM MECHANICS It is easily demonstrated that the eigenvalues of an Hermitian operator are all real. if ψ1 = ψ2 = ψa then ∞ ∞ ψ∗ (A ψa ) dx = a −∞ −∞ (A ψa )∗ ψa dx. (4. a −∞ (4.109)] a = a∗ . ψ1 (x) and ψ2 (x). eigenstates of an Hermitian operator corresponding to different eigenvalues are automatically orthogonal. Thus. we obtain a ∞ ∞ (A ψa )∗ ψa ′ dx = a −∞ ∞ −∞ ∞ ψ∗ ψa ′ dx. we can write ∞ (a − a ) −∞ ′ ψ∗ ψa ′ dx = 0. Two wavefunctions. the left-hand sides of the above two equations are equal.113).86)] that an Hermitian operator satisﬁes ∞ ∞ ψ∗ 1 −∞ (A ψ2 ) dx = −∞ (A ψ1)∗ ψ2 dx. yielding ∞ ψ∗ ψa ′ dx = 0. respectively.115) ψ∗ ψ2 dx = 0. a = a ′ . (4. (4. A ψa = a ψa . a (4. (4. ψa and ψa ′ .117) (4.113) Hence. are said to be orthogonal if ∞ (4. Recall [from Eq. . 1 −∞ (4.116) Consider two eigenstates of A.119) (4.120) ψ∗ (A ψa ′ ) dx = a ′ a −∞ However. and the second equation by ψ∗ . Hence. A ψa ′ = a ′ ψa ′ .121) By assumption.118) Multiplying the complex conjugate of the ﬁrst equation by ψa ′ . (4. and then integrating over all x.114) which reduces to [see Eq. (4. a −∞ (4. assuming that ψa is properly normalized.

i −∞ (4. It is also possible to demonstrate that the eigenstates of an Hermitian operator form a complete set: i.124) is a properly normalized eigenstate of A.125) and (4. and ∞ (4. the proof is quite difﬁcult.Fundamentals of Quantum Mechanics 55 ′ Consider two eigenstates of A. (4.123) then it is easily demonstrated that ′′ ψa = |c| 1− |c|2 ′ ψa − c−1 ψa (4. we conclude that the eigenstates of an Hermitian operator are. given an Hermitian operator A. It is straightforward to generalize the above argument to three or more degenerate eigenstates. corresponding to the eigenvalue a. Note. In summary.127) yield |ci |2 = 1. and we shall not attempt it here. (4. if ψa and ψa are properly normalized.e.125) i where the ci are complex weights. a.127) that ∞ ci = −∞ ψ∗ ψ dx. where ai is the eigenvalue corresponding to the eigenstate ψi .126) ψ∗ ψj dx = δij . ψa and ψa . (4. however. a −∞ (4. we can always choose two linear combinations of these ′ eigenstates which are orthogonal. Moreover. even ′ if ψa and ψa are not orthogonal. ψ(x). or can be chosen to be. It follows from Eqs. Such eigenstates are termed degenerate. (4.127) Here. the expansion coefﬁcients in Eq. which correspond to the same eigenvalue. any general wavefunction. if ψ is a properly normalized wavefunction then Eqs.128) Thus.e. Hence. which is orthogonal to ψa .. The above proof of the orthogonality of different eigenstates fails for degenerate eigenstates. A ψi = ai ψi .125) and (4. Thus. can be written ψ= ci ψi .. and zero otherwise. and ∞ ′ ψ∗ ψa dx = c. δij is called the Kronecker delta-function. that any general wavefunction can be written as a linear combination of these eigenstates. mutually orthogonal. However. and the ψi are the properly normalized (and mutually orthogonal) eigenstates of A: i.125) are easily determined. (4. and takes the value unity when its two indices are equal. that any linear combination ′ of ψa and ψa is also an eigenstate of A corresponding to the eigenvalue a. i (4. For instance. given the wavefunction ψ and the eigenstates ψi .129) i .

and all of the others are zero. This result can easily be generalized to the case where there are more than two eigenstates.131) and 2 σA = A2 − A = 0. 3.130) where ψi is an eigenstate of A corresponding to the eigenvalue ai . suppose that there are only two eigenstates. we expect that if a measurement of A yields the result a then the act of measurement will cause the wavefunction to collapse to a state in which a measurement of A is bound to give the result a. we have ψ= i ci ψi . In other words. Eq.137) The only solutions are x = 0 and x = 1. (4. (4.132) Now it is easily seen that A A2 Thus. (4. if a general wavefunction is expanded as a linear combination . (4. i (4. a? Well.135) |ci |2 = 1.16. It follows that a state associated with a deﬁnite value of A is one in which one of the |ci |2 is unity. It immediately follows that the result of a measurement of A must be one of the eigenvalues of A. and (a1 − a2 )2 x (1 − x) = 0. What sort of wavefunction. expressing ψ as a linear combination of the eigenstates of A. ψ. The above two equations then reduce to |c1 |2 = x. (4.133) (4. the normalization condition yields ai |ci |2 = 0. |ci |2 ai2. the only states associated with deﬁnite values of A are the eigenstates of A. 2 (4. and |c2 |2 = 1 − x.134) = i i Furthermore. By analogy with the discussion in Sect. where 0 ≤ x ≤ 1. Moreover. is such that a measurement of A is bound to yield a certain result.56 QUANTUM MECHANICS 4. i (4. If a measurement of A is bound to yield the result a then A = a.132) gives ai2 |ci |2 − = i |ci |2 ai .10 Measurement Suppose that A is an Hermitian operator corresponding to some dynamical variable.136) For instance.

8. In fact. We conclude that if A and B commute then they possess simultaneous eigenstates. we need all of the eigenstates of A to also be eigenstates of B. where ci is the coefﬁcient in front of the ith eigenstate in the expansion. the probability of a measurement of A resulting in any possible answer is unity.. that these probabilities are properly normalized: i. that if A and B do not commute (i. (4. Under what circumstances is it possible to simultaneously measure these two variables (exactly)? Well. therefore. 4. It follows that (A B − B A) ψi = (A B − B ai ) ψi = (A − ai ) B ψi = 0. This suggests that the condition for simultaneous measurement is that A and B should commute. and the general deﬁnition of a mean. thus. Let us. that the probability of a measurement of A yielding the eigenvalue ai is simply |ci |2 .140) (4. so that all states associated with unique values of A are also associated with unique values of B. and vice versa. x and p—possess eigenvalues which lie in a continuous range and non-square-integrable eigenstates (in fact. respectively. (4. Thus. like in Eq.136). or B ψi = bi ψi . some operators—most notably. In other words.130). if A B = B A) then they cannot be simultaneously measured. a simultaneous eigenstate of A and B. ψi is an eigenstate of B. Now.11 Continuous Eigenvalues In the previous two sections. .Fundamentals of Quantum Mechanics 57 of the eigenstates of A.139) Thus.138) where bi is a constant of proportionality. (4. then it is clear from Eq.. we have already seen. Suppose that this is the case. and. the possible results of measurements of A and B are the eigenvalues of A and B.e. B ψi ∝ ψi . (4. and are thus simultaneously measurable (exactly). investigate the eigenstates and eigenvalues of the displacement and momentum operators. to simultaneously measure A and B (exactly) there must exist states which are simultaneous eigenstates of A and B. in Sect. Finally. from Eq. Unfortunately. B ψi is an eigenstate of A corresponding to the eigenvalue ai (though not necessarily a normalized one). Hence. Note. respectively. and that the ψi and ai are the normalized eigenstates and eigenvalues of A. (4. Consider two physical dynamical variables represented by the two Hermitian operators A and B. 4. these two properties go hand in hand). it was tacitly assumed that we were dealing with operators possessing discrete eigenvalues and square-integrable eigenstates.e. and vice versa. if a measurement of A results in the eigenvalue ai then immediately after the measurement the system will be left in the eigenstate corresponding to ai . in order for A and B to be simultaneously measurable under all circumstances.133). or A (B ψi ) = ai (B ψi ).

. x ′ ) = δ(x − x ′ ). or ∞ c(x ′) = −∞ ψ∗ (x. (4. It follows that x ψx (x. Let us make the constant of proportionality unity. x −∞ (4. x (4.149) . ψx (x. −∞ (4. (4. Now. x ′ ) is proportional to δ(x − x ′ ). x ′ ) be the eigenstate of x corresponding to the eigenvalue x ′ ..9. x ′ ) ψx (x.148) are analogous to Eqs. x ′ ) satisﬁes the orthonormality condition ∞ ψ∗ (x. x ′′ ) dx = δ(x ′ − x ′′ ). these probabilities are properly normalized provided ψ(x) is properly normalized [cf.128). (4. (4. ψx (x. Finally.e.147) where c(x ′) = ψ(x ′ ). Evidently. x ′ ).148) In other words.145) This condition is analogous to the orthonormality condition (4. Eq. respectively. so that ψx (x.125) and (4.127) satisﬁed by squareintegrable eigenstates. x ′ ) dx ′ . by deﬁnition.143) δ(x − x ′ ) δ(x − x ′′ ) dx = δ(x ′ − x ′′ ). (4. we can expand a general wavefunction ψ(x) as a linear combination of the eigenstates. Equations (4.147) and (4.129)]: i. which is equivalent to the standard result |ψ(x ′ )| 2 . ∞ ∞ |c(x ′ )| 2 dx ′ = −∞ −∞ |ψ(x ′ )| 2 dx ′ = 1. by analogy with the results in Sect. Now. x ′ ) = x ′ ψx (x. of the displacement operator.146) where f(x) is a general function. −∞ (4.58 QUANTUM MECHANICS Let ψx (x. for square-integrable eigenstates. δ(x − x ′ ) satisﬁes ∞ f(x) δ(x − x ′ ) dx = f(x ′ ). it is easily demonstrated that ∞ (4. 4. x ′ ) (4. We can thus write ∞ ψ(x) = −∞ c(x ′ ) ψx (x. ψx (x. x ′ ) ψ(x) dx. Moreover. We can write x δ(x − x ′ ) = x ′ δ(x − x ′ ).141) for all x. the probability density of a measurement of x yielding the value x ′ is |c(x ′ )| 2 . Consider the Dirac delta-function δ(x − x ′ ).142) since δ(x − x ′ ) is only non-zero inﬁnitesimally close to x = x ′ .144) Hence.

128).. p ′). The probabilities are also properly normalized provided ψ(x) is properly normalized [cf.74) that the constant of proportionality in Eq. p ) = .Fundamentals of Quantum Mechanics 59 Finally. p ′ ) ∝ e+i p ′ x/¯ h . (4.e. for square-integrable eigenstates. p ′). (4. we can expand a general wavefunction ψ(x) as a linear combination of the eigenstates. p −∞ (4. p ′) to satisfy an analogous orthonormality condition to Eq. .155) are again analogous to Eqs. (4. Likewise.85)]: i. if a measurement of x yields the value x ′ then the system is left in the corresponding displacement eigenstate. respectively. ψp (x. or ∞ c(p ′) = −∞ ψ∗ (x. p ′′) dx = δ(p ′ − p ′′ ). as discussed in Sect.152) Thus.e. (4. δ(x − x ′ ).e.156) Finally.151) Now.154) and (4. (4. x ′ ).. (4. ∞ ψ∗ (x. (4. we require ψp (x. p ′) = p ′ ψp (x. Eq. ∞ ∞ ∞ |c(p )| dp = −∞ −∞ ′ 2 ′ |φ(p )| dp = −∞ ′ 2 ′ |ψ(x ′ )| 2 dx ′ = 1.145): i. the wavefunction collapses to a “spike-function”. immediately after the measurement: i. ∞ ψ(x) = −∞ c(p ′) ψp (x. immediately after the measurement...155) In other words. which is equivalent to the standard result |φ(p ′ )| 2 .151) should be (2π ¯ )−1/2 : i.125) and (4. if a mesurement of p yields the value p ′ then the system is left in the corresponding momentum eigenstate.154) where c(p ′) = φ(p ′ ) [see Eq.150) − i¯ h ∂x It is evident that ψp (x. according to Eqs. 3. p ′) dp ′. (2π ¯ )1/2 h ′ ′ (4. (4. p ′) ψ(x) dx. Now. Equations (4.16. (4. of the momentum operator. (4.66) and (4. an eigenstate of the momentum operator p ≡ −i ¯ ∂/∂x corresponding to the h ′ eigenvalue p satisﬁes ∂ψp (x. ψp (x. it follows from Eq. p ′) ψp (x. p ′).153) Furthermore. h h e+i p x/¯ ψp (x..67). p (4. (4. ψx (x.e. the probability density of a measurement of p yielding the result p ′ is |c(p ′)| 2 .67)].

161) Moreover. and contains no speciﬁc time dependence.161) and (4. The generalization of the above results to the case where H has continuous eigenvalues is straightforward. Of course.162) ψ(x. the system is left in the corresponding energy eigenstate.157) ∂t It is evident that this equation can be solved by writing h ψE (x. i −∞ (4. then it is evident from Eqs. we can express a general wavefunction as a linear combination of energy eigenstates: h ci ψi (x) e−i Ei t/¯ . this equation o takes the form H ψi = Ei ψi . Ei) = ψi (x) e−i Ei t/¯ .60 QUANTUM MECHANICS 4. (4. (4. . Furthermore. (4. Note that a stationary state is associated with a unique value for the energy. |ci | 2 is the probability that a measurement of the energy will yield the eigenvalue Ei .12 Stationary States An eigenstate of the energy operator H ≡ i ¯ ∂/∂t corresponding to the eigenvalue Ei h satisﬁes ∂ψE (x.159) This is known as the time-independent Schr¨dinger equation. t. immediately after such a measurement.. In this sense. More generally. Ei). (4. non-time-varying) wavefunction.158) where ψi (x) is a properly normalized stationary (i. the dynamical variable in question is a constant of the motion. Ei ) i¯ h = Ei ψE (x. (4. Substitution of the above expression into Schr¨dinger’s equation o (4. t. t) = i where ci = ∞ ψ∗ (x) ψ(x. t.1) yields the equation satisﬁed by the stationary wavefunction: ¯ 2 d2 ψi h = [V(x) − Ei ] ψi . If a dynamical variable is represented by some Hermitian operator A which commutes with H (so that it has simultaneous eigenstates with H).e. 0) dx.162) that the expectation value and variance of A are time independent. since the probability density |ψE | 2 is non-time-varying. The wavefunction ψE (x.160) where H is assumed not to be an explicit function of t.163) Here. i −∞ (4. the ψi satisfy the usual orthonormality condition: ∞ ψ∗ ψj dx = δij . t. Ei ) corresponds to a so-called stationary state. 2 m dx2 (4.

Demonstrate that if a particle in a one-dimensional stationary state is bound then the expectation value of its momentum must be zero. Classically. Calculate x .Fundamentals of Quantum Mechanics Exercises 61 ˚ 1. 8. . 2 + a2 π x + a2 ) = π/a. Suppose that V(x) is complex. An operator A. If ∞ −∞ ψ∗ ψ2 dx = c. corresponding to a physical quantity α. Obtain an expression for ∂P(x. x2 . has an expectation value which is constant in time. 5. for the normalized wavefunction ψ(x) = Use these to ﬁnd σx σp . with eigenvalues b1 and b2 . and σx . If β is then measured and then α again. What is the subsequent spread in wavelengths of the no longer monochromatic for 10 light? 2. 7. has normalized eigenfunctions φ1 (x) and φ2 (x). The eigenfunctions are related via ψ1 = (2 φ1 + 3 φ2 ) ψ2 = (3 φ1 − 2 φ2 ) √ √ 13. An operator B. if a particle is not observed then the probability of ﬁnding it in a one-dimensional box of length L. with eigenvalues a1 and a2 . t) dx from Schr¨dinger’s equation. is a constant 1/L per unit length. Note that 1 2 a3 . α is measured and the value a1 is obtained. the expectation value of x2 is L2 /3. 1 where c is real. and σp . Monochromatic light with a wavelength of 6000 A passes through a fast shutter that opens −9 sec. corresponding to another physical quantity β. and the √ standard deviation of x is L/ 12. Demonstrate that p = −i ¯ ∂/∂x is an Hermitian operator. p2 . has two normalized eigenfunctions ψ1 (x) and ψ2 (x). 13. Show that the classical expectation value of x is L/2. What does this tell us about a complex V(x)? o 6. show that the probability of obtaining a1 a second time is 97/169. and contains no explicit time dependence. 4. t)/∂t and d/dt P(x. ψ1 (x) and ψ2 (x) are normalized eigenfunctions corresponding to the same eigenvalue. Demonstrate that an operator which commutes with the Hamiltonian. (b) ψ1 + ψ2 . Find the Hermitian conjugate of h a = x + i p. ∞ 2 −∞ dx/(x 3. as well as p . 9. which extends from x = 0 to x = L. ﬁnd normalized linear combinations of ψ1 and ψ2 which are orthogonal to (a) ψ1 .

For a certain system. corresponding to properly normalized eigenfunctions √ φ1 = (u1 + u2 ) 2. show that the expectation value of A at time t is a1 + a2 a1 − a2 [E1 − E2 ] t A = + cos . √ φ2 = (u1 − u2 ) 2. It has eigenvalues a1 and a2 . the operator corresponding to the physical quantity A does not commute with the Hamiltonian.62 QUANTUM MECHANICS 10. 2 2 ¯ h . where u1 and u2 are properly normalized eigenfunctions of the Hamiltonian with eigenvalues E1 and E2 . If the system is in the state ψ = φ1 at time t = 0.

(5. Evidently. k2 = (5. t) = ψ(x) e−i E t/¯ .5) . (5. ψ) is to remain ﬁnite then ψ must go to zero in regions where the potential is inﬁnite.One-Dimensional Potentials 63 5 One-Dimensional Potentials 5.3) It follows from Eq. we shall investigate the interaction of a non-relativistic particle of mass m and energy E with various one-dimensional potentials. ψ = 0 in the regions x ≤ 0 and x ≥ a. Hence. (5. otherwise (5.2) that ψ satisﬁes d2 ψ = −k2 ψ dx2 in this region. It is easily demonstrated that there are no solutions with E < 0 which are capable of satisfying the boundary conditions (5. V(x). 4. 2 dx ¯ h (5.19) would become inﬁnite (which is also unphysical). the solution.4) Furthermore.4). Since we are searching for stationary solutions with unique energies. it follows from Eq. Likewise.1 Introduction In this chapter. hence.12) h ψ(x. the solution must be continuous. ψ(x). we can write the wavefunction in the form (see Sect. 5. (5. The boundary conditions on ψ in the region 0 < x < a are ψ(0) = ψ(a) = 0.2) In general.2 Inﬁnite Potential Well Consider a particle of mass m and energy E moving in the following simple potential: V(x) = 0 ∞ for 0 ≤ x ≤ a .2) that if d2 ψ/dx2 (and.6) ¯2 h Here. (5. where 2mE . we are assuming that E > 0. otherwise the probability density |ψ| 2 would become inﬁnite (which is unphysical).1) where ψ(x) satisﬁes the time-independent Schr¨dinger equation: o 2m d2 ψ = 2 [V(x) − E] ψ. to the above equation must be ﬁnite. the problem is equivalent to that of a particle trapped in a onedimensional box of length a. otherwise the probability current (4.

(5.. In the regions to the left and to the right of the barrier. (5. According to the discussion in Sect. 0 (5.7) where the An are arbitrary (real) constants. Now. Hence. En = 2 m a2 (5. (5.15) . ψ(x) satisﬁes d2 ψ = −k2 ψ.e.6) and (5.4).5). (5. again from Sect. (5. subject to the boundary conditions (5. · · ·. 0) dx. Finally.10) 2/a.8) that the energy E is only allowed to take certain discrete values: i. 2. a (5.9) In other words.12) where cn = a ψn (x) ψ(x. 2. and kn = nπ .12.∞ h cn ψn (x) e−i En t/¯ . solutions in which |ψ| → 0 as |x| → ∞. 0 (5.12. we expect the stationary eigenfunctions ψn (x) to satisfy the orthonormality constraint a ψn (x) ψm (x) dx = δnm . provided An = ψn (x) = 2 x sin n π a a for n = 1.64 QUANTUM MECHANICS The solution to Eq. 3.e.. is ψn (x) = An sin(kn x). it can be seen from Eqs.14) 0 otherwise where V0 > 0. 4. (5. · · ·. dx2 (5. 3. the eigenvalues of the energy operator are discrete.8) for n = 1.13) 5. This is a general feature of bounded solutions: i.11) It is easily demonstrated that this is the case. n2 π2 ¯ 2 h .3 Square Potential Barrier Consider a particle of mass m and energy E > 0 interacting with the simple square potential barrier V0 for 0 ≤ x ≤ a V(x) = . t) = n=0. the general time-dependent solution can be written as a linear superposition of stationary solutions: ψ(x. 4.

We interpret the ﬁrst plane wave as an incoming particle (or. Let us adopt the following solution of the above equation to the left of the barrier (i. 0 ≤ x ≤ a). in a stationary state (i. (5.35). the general solution to Eq.15) to the right of the barrier (i. we must have jl = jr .19) Now. We interpret this as a particle (or stream of particles) transmitted through the barrier. (4.. Inside the barrier (i..e.16) This solution consists of a plane wave of unit amplitude traveling to the right [since the time-dependent wavefunction is multiplied by exp(−i ω t). The probability current in the region x > a takes the form jr = v |T | 2 . In this case. (5. Hence. which takes the form jl = v (1 − |R| 2 ). since reﬂection and transmission are the only possible outcomes for a particle incident on the barrier. (5. according to Eq. Hence. This can be seen by calculating the probability current (4. (5. (5. (5. dx2 where q2 = (5.22) ¯2 h Let us.18) This solution consists of a plane wave of complex amplitude T traveling to the right. |T | 2 is the probability of transmission. and the second as a particle (or stream of particles) reﬂected by the potential barrier.21) 2 m (E − V0 ) .20) In other words. rather. ﬁrst of all. ∂j/∂x = 0).6).19) in the region x < 0. consider the case where E > V0 .e. where ω = E/¯ > 0].e. x < 0): ψ(x) = e i k x + R e−i k x .. x > a): ψ(x) = T e i k x .e. ψ(x) satisﬁes d2 ψ = −q2 ψ. ∂|ψ| 2 /∂t = 0). or |R| 2 + |T | 2 = 1.21) inside the barrier takes the form ψ(x) = A e i q x + B e−i q x . (5. Hence. h Let us adopt the following solution to Eq.17) where v = p/m = ¯ k/m is the classical particle velocity. (5. a stream of incoming particles). as must be the case. (5. (5. the probabilities of reﬂection and transmission sum to unity. and h a plane wave of complex amplitude R traveling to the left.. |R| 2 is the probability of reﬂection.e.One-Dimensional Potentials 65 where k is given by Eq. the probability current is a spatial constant (i.23) .

. the classical probability of reﬂection is zero. the special barriers widths are integer multiples of half the de Broglie wavelength of the particle inside the barrier. there is a substantial probability that the incident particle will be reﬂected by the barrier.27) After considerable algebra.28) and (5. that the classical result. that at certain barrier widths the probability of reﬂection goes to zero.28) and |T | 2 = 4 k2 q2 .29) Note that the above two expression satisfy the constraint (5.29). 5. According to classical physics. k (1 − R) = q (A − B). · · ·. (5. It is evident.1.30) where n = 1. It turns out that this is true irrespective of the energy of the incident particle. |R| 2 = 0 and |T | 2 = 1. and the classical probability of transmission is unity. q A e i q a − B e−i q a = k T e i k a. reﬂection is impossible when V0 < E. x = 0) yields 1 + R = A + B. the above four equations yield (k2 − q2 ) 2 sin2 (q a) . |R| = 4 k2 q2 + (k2 − q2 ) 2 sin2 (q a) 2 (5. x = a) gives A e i q a + B e−i q a = T e i k a . continuity of ψ and dψ/dx at the right edge of the barrier (i.29). (5. when V0 is of order E.28).20). at x = 0 and x = a) are that ψ and dψ/dx are both continuous. 3.2.e. In other words. 5. (5. (5. from Fig. from Fig. that these special barrier widths correspond to q a = n π.1 and 5..e. according to classical physics. 2. Now. These boundary conditions ensure that the probability current (4. (5. V0 ≪ E).66 QUANTUM MECHANICS where q = 2 m (E − V0 )/¯ 2 .28) and (5.25) Likewise. However. It can be seen. respectively. There is no reﬂection at . from Eq. Continuity of ψ and dψ/dx at the left edge of the barrier (i.2. 5. 4 k2 q2 + (k2 − q2 ) 2 sin2 (q a) (5. is obtained in the limit where the height of the barrier is relatively small (i. It is instructive to compare the quantum mechanical probabilities of reﬂection and transmission—(5. In other words. the boundary conditions at the edges of the barrier (i. but is otherwise unaffected. are plotted in Figs.e. The reﬂection and transmission probabilities obtained from Eqs. It can also be seen. if a particle of energy E is incident on a potential barrier of height V0 < E then the particle slows down as it passes through the barrier.26) (5.19) remains ﬁnite and continuous across the edges of the boundary. as must be the case if it is to be a spatial constant. respectively—with those derived from classical physics..24) (5. h Now..e.

to the free-space de Broglie wavelength.One-Dimensional Potentials 67 Figure 5.2: Transmission (solid-curve) and reﬂection (dashed-curve) probabilities for a particle of energy E incident on a square potential barrier of height V0 = 0. as a function of the ratio of the width of the barrier. a.25 λ. as a function of the ratio of the height of the barrier. Figure 5. to the energy. E. .1: Transmission (solid-curve) and reﬂection (dashed-curve) probabilities for a square potential barrier of width a = 1.75 E. where λ is the free-space de Broglie wavelength. of the incident particle. λ. V0 .

from Fig.e. 5. V0 ≫ E). This phenomenon. Note that Eqs.37) 4 k2 q2 + (k2 + q2 ) 2 sinh2 (q a) These expressions can also be obtained from Eqs. the general solution to Eq. at these widths. Nevertheless.4.28) and (5. Now. It is again instructive to compare the quantum mechanical probabilities of reﬂection and transmission—(5.36) and (5. which is inexplicable within the context of classical physics. continuity of ψ and dψ/dx at the right edge of the barrier (i.21) inside the barrier takes the form ψ(x) = A e q x + B e−q x . 5. |R| 2 = 1 and |T | 2 = 0. (5.37). i k (1 − R) = q (A − B). |R| = 4 k2 q2 + (k2 + q2 ) 2 sinh2 (q a) 2 (5. transmission is impossible when V0 > E. (5. |T | 2 = . is called tunneling. After considerable algebra.36) and (5.20). respectively—with those derived from classical physics.36) and 4 k2 q2 . is obtained for relatively thin barriers (i. the backward traveling wave reﬂected from the left edge of the barrier interferes destructively with the similar wave reﬂected from the right edge of the barrier to give zero net reﬂected wave.4. It can also be seen.29) by making the substitution q → −i q. In this case. q a ≫ 1). (5.e. 5.32) (5..31) where q = 2 m (V0 − E)/¯ 2 . (5.e.68 QUANTUM MECHANICS the special barrier widths because. there is a small but ﬁnite probability that a particle incident on the barrier will be transmitted.. the above four equations yield (k2 + q2 ) 2 sinh2 (q a) .35) (5. even for very wide barriers (i. (5.3 and 5. there is a substantial probability that the incident particle will be transmitted by the barrier. and the classical probability of transmission is zero.e. (5. are plotted in Figs. that the classical result.. according to classical physics... the classical probability of reﬂection is unity. if a particle of energy E is incident on a potential barrier of height V0 > E then the particle is reﬂected. It can be seen.e. Continuity of ψ and dψ/dx at the left edge of the barrier h (i. when V0 is of order E. x = a) gives A e q a + B e−q a = T e i k a .36) and (5.37) satisfy the constraint (5. now. x = 0) yields 1 + R = A + B. respectively. However. q (A e q a − B e−q a ) = i k T e i k a . consider the case E < V0 .3. (5.33) Likewise. The reﬂection and transmission probabilities obtained from Eqs.34) (5.37). In other words. from Fig. that the transmission probability decays exponentially as the width of the barrier increases. Let us. According to classical physics. q a ∼ 1) in the limit where the height of the barrier is relatively large (i.

.5 λ. V0 . as a function of the ratio of the energy.One-Dimensional Potentials 69 Figure 5.3: Transmission (solid-curve) and reﬂection (dashed-curve) probabilities for a square potential barrier of width a = 0. λ. Figure 5.4: Transmission (solid-curve) and reﬂection (dashed-curve) probabilities for a particle of energy E incident on a square potential barrier of height V0 = (4/3) E. E. as a function of the ratio of the width of the barrier. where λ is the free-space de Broglie wavelength. to the free-space de Broglie wavelength. a. of the barrier. of the incoming particle to the height.

. It follows that h dψ(x) = i k(x) ψ(x). 1 Similarly. |ψ(x)| 2 = |ψ0 | 2 . (5. Eq. (5.38) of the form k2 (x) = x (5. which is approximately the same as the variation length-scale of V(x). and k(x) is consequently real (i.40). (5. an allowed region according to classical physics). A comparison of Eqs. ′| |k (5. Note that this solution represents a particle propagating in the positive x-direction [since the full wavefunction is multiplied by exp(−i ω t). where ω = E/¯ > 0] with the continuously varying wavenumber k(x). (5. must be much greater than the particle’s de Broglie wavelength (which is of order k−1 ). (5. dx and (5. Incidentally.A. the approximation involved in dropping the ﬁrst term on the right-hand side of Eq. Suppose. (5. (5.. or k ≫ k−1 .38) provided that the ﬁrst term on its right-hand side is negligible compared to the second. Wentzel.70 QUANTUM MECHANICS 5.e. (5.41) d2 ψ(x) = i k ′ (x) ψ(x) − k2 (x) ψ(x). Brillouin. 2 dx where 2 m [E − V(x)] . The particle’s wavefunction satisﬁes d2 ψ(x) = −k2 (x) ψ(x).44) as long as the particle moves through a region in which E > V(x).42) 2 dx where k ′ ≡ dk/dx.4 WKB Approximation Consider a particle of mass m and energy E > 0 moving through some slowly varying potential V(x).e. ¯2 h Let us try a solution to Eq.38) (5.42) reveals that Eq.40) where ψ0 is a complex constant. that the 1 After G. This yields the validity criterion |k ′ | ≪ k2 . H. however.43) In other words.38) and (5.42) is generally known as the WKB approximation.40) is termed a WKB solution.40) represents an approximate solution to Eq. and L. Let us suppose that this is the case. . According to the WKB solution (5. the variation length-scale of k(x). Kramers. the probability density remains constant: i. (5.39) ψ(x) = ψ0 exp 0 i k(x ′ ) dx ′ .

at x = x1 and x2 ).e.45).e. (5. x > x2 ).3). Let the barrier extend from x = x1 to x2 . and k(x) is consequently imaginary.43) for the validity of the WKB approximation implies that the above transmission probability is very small. It is easily demonstrated that the probability of a particle incident from the right tunneling through the barrier is the same. (5. in which the WKB approximation breaks down to the integral in Eq. It follows that the probability density at the right-hand side of the barrier (i. where 0 < x1 < x2 ..47) where |ψ1 | 2 is the probability density at the left-hand side of the barrier (i. in the region to the right of the barrier (i. By deﬁnition. the above expression for the tunneling probability is a reasonable approximation provided that the incident particle’s de Broglie wavelength is much smaller than the spatial extent of the potential barrier. x = x2 ) is |ψ2 | 2 = |ψ1 | 2 exp −2 x2 x1 |k(x ′ )| dx ′ .49) (see Sect.49) is fairly negligible. According to the WKB solution (5.43) breaks down completely at the edges of the barrier (i. a region from which the particle is excluded according to classical physics). (5.48) Note that |ψ2 | 2 < |ψ1 | 2 . we have neglected the unphysical exponentially growing solution. (5. 5.. (5. However.. x = x1 ). Note that the criterion (5..46) Here.45) where ψ1 = ψ0 exp x1 0 i k(x ′ ) dx ′ .. x |ψ(x)| 2 = |ψ1 | 2 exp −2 x1 |k(x ′ )| dx ′ . the probability density decays exponentially inside the barrier: i. Unfortunately. Hence. |ψ2 | 2 |T | = = exp −2 |ψ1 | 2 2 x2 x1 |k(x ′ )| dx ′ (5. the validity criterion (5. Of course. Hence. that a particle incident from the left will tunnel through the barrier and emerge on the other side: i. E < V(x) inside such a barrier.e.e. the WKB approximation only applies to situations in which there is very little chance of a particle tunneling through the potential barrier in question.. the probability density takes the constant value |ψ2 | 2 . |T | 2 .One-Dimensional Potentials 71 particle encounters a potential barrier (i. around x = x1 and x2 . .e.. We can interpret the ratio of the probability densities to the right and to the left of the potential barrier as the probability. it can be demonstrated that the contribution of those regions. The WKB solution inside the barrier is written x ψ(x) = ψ1 exp − x1 |k(x ′ )| dx ′ .e. since k(x) = 0 at these points.e.

say. E. the applied electric ﬁeld is shielded from the interior of the metal. The potential barrier is sketched in Fig.5.6) that electrons just below the surface of a metal can be regarded as being in a potential well of depth W. Hence. (5. that an electron just below the surface of the metal is conﬁned by a triangular potential barrier which extends from x = x1 to x2 . 5. from Fig. (5. where x1 = 0 and x2 = W/e E.51) . It can be seen. where W is called the work function of the surface. is √ 2 2 m x2 2 |T | = exp − V(x) − E dx .50) ¯ h x1 or |T | 2 = exp − 2 √ 2m ¯ h W/e E 0 √ W − e E x dx . which is directed such that it accelerates electrons away from the surface. and the surface at x = 0. Now. let the metal lie at x < 0. and consequently being emitted from the surface.5 Cold Emission Suppose that an unheated metal surface is subject to a large uniform external electric ﬁeld of strength E. The electric ﬁeld modiﬁes this to V(x)−E = W−e E x. 5. In the absence of the electric ﬁeld.5. the energy. the potential barrier just above is the surface is simply V(x)−E = W. of an electron just below the surface is unaffected by the ﬁeld. We have already seen (in Sect. Making use of the WKB approximation (see the previous subsection). the probability of such an electron tunneling through the barrier. 5. 3.72 QUANTUM MECHANICS 1111111111111 0000000000000 x→ 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 V − E = W − eE x 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 VACUUM METAL 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 E 0000000000000 1111111111111 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 1111111111111 0000000000000 Energy → x1 x2 Figure 5. Adopting a simple one-dimensional treatment of the problem.5: The potential barrier for an electron in a metal surface subject to an external electric ﬁeld.

Such a nucleus thus decays to produce a daughter nucleus of charge-number Z1 = Z − 2 and mass-number A1 = A − 4.e. 3 ¯ eE h (5. the αparticle. Now. charge-number Z. and. An STM can thus be used to construct a very accurate contour map of the surface under investigation.52) |T | = exp − The above result is known as the Fowler-Nordheim formula. In 1928. A large voltage difference is applied between the probe and the surface. hence. An STM consists of a very sharp conducting probe which is scanned over the surface of a metal (or any other solid conducting medium). and mass-number A. STMs are capable of achieving sufﬁcient resolution to image individual atoms 2 m1/2 W 3/2 .53). The cold emission of electrons from a metal surface is the basis of an important device known as a scanning tunneling microscope. 0 (5. and is emitted after tunneling through the potential barrier between itself and the daughter nucleus.53) 5. Note that the probability of emission increases exponentially as the electric ﬁeld-strength above the surface of the metal increases.54) for Z ≫ 1. In fact. is trapped in a potential well of radius R by the potential barrier Z1 Z2 e2 V(r) = (5. and inversely proportional to the spacing between the tip and the surface. and an αparticle of charge-number Z2 = 2 and mass-number A2 = 4. Let us investigate the α-decay of a particular type of atomic nucleus of radius R. The magnitude of this current is proportional to the tunneling probability (5.. or an STM. helium nucleii) of some characteristic energy.5 × 10−15 A1/3 m = 2.One-Dimensional Potentials This reduces to |T | 2 = exp −2 or 2 73 √ m1/2 W 3/2 2 ¯ eE h 4 √ 1 1 − y dy . It follows that the current is an extremely sensitive function of the surface electric ﬁeld-strength. of the spacing between the tip and the surface (assuming that the potential difference is held constant).55) 4π ǫ0 r . Let the characteristic energy of the α-particle be E. George Gamov proposed a very successful theory of α-decay. Electrons tunneling between the surface and the probe tip give rise to a weak electric current. whose energy is E. This process is know as α-decay. In other words. according to which the α-particle moves freely inside the nucleus. the surface electric ﬁeld-strength immediately below the probe tip is proportional to the applied potential difference. Incidentally.0 × 10−15 Z1 m 1/3 (5.6 Alpha Decay Many types of heavy atomic nucleus spontaneously decay to produce daughter nucleii via the emission of α-particles (i. nuclear radii are found to satisfy the empirical formula R = 1.

2/3 (5. the α-particle effectively attempts to tunnel through the potential barrier ν times a second.57) 1/2 = 0. coordinate). It follows that the particle bounces backward and forward within the nucleus at the frequency ν ≃ v/R. the α-particle moves inside the nucleus with the characteristic velocity v = 2 E/m.63) . E (5. and Ec = Z1 Z2 e2 2/3 = 1. If each of these attempts has a probability |T | 2 of succeeding. Hence.74 QUANTUM MECHANICS for r > R. (5.44 Z1 MeV 4π ǫ0 R (5. The above expression reduces to |T | 2 = exp −2 where β= is a dimensionless constant. giving ν ≃ 2 × 1028 yr−1 (5.62) |T | 2 ≃ exp −2 (5. if there are N(t) ≫ 1 undecayed nuclii at time t then there are only N + dN at time t + dt.60) when ǫ ≪ 1. (5. the probability of the α-particle tunneling through the barrier is √ 2 2 m r2 |T | 2 = exp − V(r) − E dr .59) Z1 Z2 e2 m R 4π ǫ0 ¯ 2 h √ Ec /E 2β 1 E 1 − y Ec 1/2 dy . m = 4 mp is the α-particle mass. It is easily demonstrated that 1/ǫ 1 1 −ǫ y 1/2 dy ≃ √ π 2β 2 π √ −2 2 ǫ Ec − 2 .61) for a 1 MeV α-particle trapped inside a typical heavy nucleus of radius 10−14 m. Thus. Of course. Hence.58) is the characteristic energy the α-particle would need in order to escape from the nucleus without tunneling. Here.56) ¯ h r1 where r1 = R and r2 = Z1 Z2 e2 /(4π ǫ0 E). rather than a Cartesian. where dN = −N ν |T |2 dt. then the probability of decay per unit time is ν |T |2 . Making use of the WKB approximation (and neglecting the fact that r is a radial.74 Z1 (5. E ≪ Ec . Now.

66) The half-life. when the particle encounters the well it is either reﬂected or transmitted. we obtain log10 [τ(yr)] = −C1 − C2 Z1 where C1 = 28. 5. and the α-particle energy.72) 5. Finally.83.5.70) (5.9. C2 = 1. As is easily demonstrated.64) Now.67) (5.73) 0 otherwise where V0 > 0. Thus. for atomic nucleii which undergo α-decay are indeed found to satisfy a relationship of the form (5.61. the half-life. C3 = 1. ν |T |2 (5. the daughter charge-number. 75 (5. C3 = 1. (5.6) is obtained using C1 = 28. The best ﬁt to the data (see Fig. E. if E > 0 then the particle is unbounded. the reﬂection and . τ. It follows from the above formula that τ= ln 2 . Note that these values are remarkably similar to those calculated above.71) (5.69) 2/3 + C3 Z1 E(MeV) . (5.7 Square Potential Well Consider a particle of mass m and energy E interacting with the simple square potential well −V0 for −a/2 ≤ x ≤ a/2 V(x) = . C2 = 1.66). (5. Z1 = Z − 2. is deﬁned as the time which must elapse in order for half of the nuclii originally present to decay.60. making use of the above results.One-Dimensional Potentials This expression can be integrated to give N(t) = N(0) exp(−ν |T |2 t). Now. (5.65) Note that the half-life is independent of N(0).73. τ.68) (5.

Both half-lives are measured in years. U 230. Data obtained from IAEA Nuclear Data Centre. the points correspond to the following nucleii: Rn 215. and E the characteristic energy of the emitted α-particle in MeV. U 232. U 234. Z1 = Z − 2.6: The experimentally determined half-life. Fm 250.61 Z1/ E. where Z is the charge number of the nucleus. Here. U 238.76 QUANTUM MECHANICS Figure 5. Gd 150. Po 214. Gd 152. Ac 225.9 − 1. In order of increasing half-life. of various atomic nucleii which 2/3 decay via α emission versus the best-ﬁt theoretical half-life log10 (τth ) = −28. U 236. .60 Z1 + √ 1. Po 197. Pt 190. τex . Nd 144. Po 216.

¯2 h 2 m (E + V0 ) = .e. or totally anti-symmetric [i. where E0 = 2¯2 h . ¯2 h By symmetry. that E < 0.83) (5.78) (5.. at x = ±a/2) yields q2 = k = q tan(q a/2).79) The solution inside the well (i.e.One-Dimensional Potentials transmission probabilities are given by Eqs. where Let us. In this case.73) must be either totally symmetric [i.e. ψ(−x) = −ψ(x)].28) and (5. Is is possible to ﬁnd bounded solutions of Schr¨dinger’s equation in the ﬁnite o square potential well (5.29). |ψ|2 → 0 as |x| → ∞). (5.e. the particle is bounded (i. (5.. where k2 = q2 2mE . The constraint that ψ(x) and its ﬁrst derivative be continuous at the edges of the well (i. however.e. x < −a/2).e. Let y = q a/2. respectively. the solution of Schr¨dinger’s equation which satisﬁes the boundary o condition ψ → 0 and x → −∞ is ψ(x) = A e k x .81) ¯2 h Here.76) 2 m |E| . (5. x > a/2) is k2 = ψ(x) = A e−k x . V(−x) = V(x)] potential (5. search for a totally symmetric solution. (5. Moreover.74) (5. |x| ≤ a/2) which satisﬁes the symmetry constraint ψ(−x) = ψ(x) is ψ(x) = B cos(q x). (5.2) in the o symmetric [i. In the region to the left of the well (i..77) (5. we have assumed that E > −V0 .e.82) where . ¯2 h 77 (5.e. m a2 (5.75) Suppose.. It follows that E = E0 y2 − V0 .. the solutions must satisfy the boundary condition ψ→0 as |x| → ∞. it is easily seen that independent solutions of Schr¨dinger’s equation (5..73)? Now.80) 2 m (V0 + E) . ﬁrst of all. the solution in the region to the right of the well (i.. ψ(−x) = ψ(x)].84) (5.

82) becomes λ − y2 = tan y.e. from the ﬁgure. Figure 5.7 shows these two curves plotted for a particular value of λ.5 π2. Eq. (5. it is also evident that there is always at least one totally symmetric bound state. as the well becomes deeper) there are more and more bound states. The √ latter curve takes the value 0 when y > λ.86) √ Here. However.. the solutions to Eq..e. Now.85) . In this case. E must lie in the range −V0 < E < 0. E0 (5. In the limit λ ≫ 1 (i.e. no matter how small λ becomes (i. (5.8).87) These solutions are equivalent to the odd-n inﬁnite square well solutions speciﬁed by Eq. Moreover. (5. the curves intersect twice. (5.7: The curves tan y (solid) and λ − y2 /y (dashed). it is evident.85) asymptote to the roots of tan y = ∞. (5. the limit in which the well becomes very deep).. no matter how shallow the well becomes). where j is a positive integer.78 QUANTUM MECHANICS Figure 5.e. the solutions to Eq. y must lie in the range 0 < y < λ: i.. a (5.85) correspond to the intersection of the curve λ − y2 /y with the curve tan y. y with λ= V0 . Moreover. or q= (2 j − 1) π . calculated for λ = 1. indicating the existence of two totally symmetric bound states in the well. This gives y = (2 j − 1) π/2. that as λ increases (i.

λ < (π/2)2] then there is no totally anti-symmetric bound state. This gives y = j π. or 2jπ .7. (5..8: The curves tan y (solid) and −y/ λ − y2 (dashed). the curves intersect once. In other words. However. similar analysis to the above yields − y λ − y2 = tan y.8 shows these two curves plotted for the same value of λ as that used in Fig. calculated for λ = 1.e. indicating the existence of a single totally anti-symmetric bound state in the well.e.e.88) The solutions of this equation correspond to the intersection of the curve tan y with the curve −y/ λ − y2 . For the case of a totally anti-symmetric bound state. It is. as the well becomes deeper) there are more and more bound states. 5. from the ﬁgure. a very shallow potential well always possesses a totally symmetric bound state. the limit in which the well becomes very deep).5 π2.. a q= (5. (5.One-Dimensional Potentials 79 Figure 5. In this case. Figure 5.8). that as λ increases (i. the solutions to Eq. In the limit λ ≫ 1 (i. .89) These solutions are equivalent to the even-n inﬁnite square well solutions speciﬁed by Eq. but does not generally possess a totally anti-symmetric bound state. evident.. it is also evident that when λ becomes sufﬁciently small [i. again.88) asymptote to the roots of tan y = 0. where j is a positive integer. (5.

Assuming that the quantum mechanical Hamiltonian has the same form as the classical Hamiltonian.98) . (5.94) 2 dy We need to ﬁnd solutions to the above equation which are bounded at inﬁnity: i. where ω is the oscillator’s classical angular frequency of oscillation. the timeindependent Schr¨dinger equation for a particle of mass m and energy E moving in a o simple harmonic potential becomes 2m d2 ψ = 2 2 dx ¯ h 1 K x2 − E ψ.93) ǫ= ¯ω h Equation (5.8 Simple Harmonic Oscillator The classical Hamiltonian of a simple harmonic oscillator is p2 1 H= + K x2 . in this limit the equation simpliﬁes somewhat to give d2 ψ − y2 ψ ≃ 0. Consider the behavior of the solution to Eq.80 QUANTUM MECHANICS 5. (5. (5. Furthermore.94) in the limit |y| ≫ 1. Clearly. (5. (5. if ψ(y) is to remain bounded as |y| → ∞ then we must chose the exponentially decaying solution.97) into Eq. solutions which satisfy the boundary condition ψ → 0 as |y| → ∞. 2 dy The approximate solutions to the above equation are ψ(y) ≃ A(y) e±y 2 /2 (5. we obtain dh d2 h −2y + (ǫ − 1) h = 0. Substituting Eq.91) reduces to d2 ψ − (y2 − ǫ) ψ = 0. (5. dy2 dy (5.95) . let mω y= x. function of y.94).97) where we would expect h(y) to be an algebraic. rather than an exponential. This suggests that we should write 2 ψ(y) = h(y) e−y /2 ..92) ¯ h and 2E . As is easily seen. (5. (5. 2m 2 (5.90) where K > 0 is the so-called force constant of the oscillator.96) where A(y) is a relatively slowly varying function of y.91) Let ω = K/m. 2 (5.e.

It is eash ily demonstrated that the (normalized) wavefunction of the lowest energy state takes the form 2 2 e−x /2 d √ . h (5.98). h where ω is the classical oscillation frequency. 1. This behavior is unacceptable. · · ·. (5. This is sometimes called zero-point energy. (5. the lowest energy state (n = 0) possesses the ﬁnite energy (1/2) ¯ ω. i Hence.102) It follows that ψ(y) = h(y) exp(−y2 /2) varies as exp( y2 /2) as |y| → ∞. Furthermore.100) (i + 1) (i + 2) Consider the behavior of h(y) in the limit |y| → ∞. This implies. h Let ψn (x) be an energy eigenstate of the harmonic oscillator corresponding to the eigenvalue En = (n + 1/2) ¯ ω.103) where n is a non-negative integer. we obtain E = (n + 1/2) ¯ ω. we have h(y) ∼ C j y2 j 2 ∼ C ey . j! (5.101) ci+2 ≃ ci . Finally.99) is n + 1. The only way in which we can prevent ψ from blowing up as |y| → ∞ is to demand that the power series (5.106) . at large |y|. 2. Note that the number of terms in the power series (5. h (5. (5.One-Dimensional Potentials Let us attempt a power-law solution of the form ∞ 81 h(y) = i=0 ci yi .99) Inserting this test solution into Eq. The spacing between successive energy levels is ¯ ω. we conclude that a particle moving in a harmonic potential has quantized energy levels which are equally spaced. (5.104) for n = 0. when the higher powers of y dominate. Hence. (5. that ǫ = 2 n + 1.100).105) ψ0 (x) = π1/4 d where d = ¯ /m ω. using Eq. we obtain the recursion relation (2 i − ǫ + 1) ci+2 = ci . (5.99) terminate at some ﬁnite value of i. The above recursion relation simpliﬁes to 2 (5. since it does not satisfy the boundary condition ψ → 0 as |y| → ∞. from the recursion relation (5.93). and equating the coefﬁcients of yi .

these operators satisfy the commutation relation [a+ .110) (5.115) 2 from which the result H ψn = (n + 1/2) ¯ ω ψn = En ψn h (5.113) (5.112) (5.n−1 .n+1 + n δm. and d = ¯ /m ω.111) (5. (5. and vice versa.116) is readily deduced.107). Finally. for the harmonic oscillator: i. (5. dy where x = d y. Eq.e.107) Now.113). operating on the wavefunction with a+ causes the quantum number n to increase by unity.94) can be written d2 − 2 + y2 ψn = (2n + 1) ψn .82 QUANTUM MECHANICS Assuming that the ψn are properly normalized (and real).109) (5. (5.. Eq. n ψn−1 .114) (5. we have ∞ ψn ψm dx = δnm . The above two equations imply that a+ ψn = a− ψn = √ √ n + 1 ψn+1 . and (5. (5. It is helpful to deﬁne the operators h d 1 +y .108) We conclude that a+ and a− are raising and lowering operators. 2mω .117) √ √ ¯ h m δm. (5. a± = √ ∓ dy 2 As is easily demonstrated.114) yield the useful expression d ψm x ψn dx = √ 2 −∞ = ∞ ∞ ψm (a+ + a− ) ψn dx −∞ (5. a−] = −1. or a− a+ ψn = (n + 1) ψn. Using these operators. The Hamiltonian for the harmonic oscillator can be written in the form 1 H = ¯ ω a+ a− + h . respectively. Eqs. −∞ (5. (5.108) can also be written in the forms a+ a− ψn = n ψn .

Suddenly the well expands to twice its original size. where k2 = (2m/¯ 2 ) E and q 2 = (2m/¯ 2 ) (E − W). A particle of mass m is in the ground-state of an inﬁnite one-dimensional square-well of width a. and k2 = (2m/¯ 2 ) E.e. x≥L . What now is the subsequent time evolution? Calculate the probability of ﬁnding the particle between 0 and a/2 as a function of time in each case. leaving the wavefunction momentarily undisturbed. as the right wall moves from a to 2a. What is the most probable result? What is the probability of obtaining this result? What is the next most probable result. Consider the half-inﬁnite potential well ∞ V(x) = −V0 0 x≤0 0<x<L . What is the subsequent time evolution of the wavefunction? Suppose that the initial wavefunction is ψ(x.One-Dimensional Potentials Exercises 83 1. 7. A particle of mass m moves freely in one dimension between impenetrable walls located at x = 0 and a. A particle of mass m and energy E is in one of the wells. 3. A stream of particles of mass m and energy E > 0 encounter a potential step of height W(< E): i. where β = m α/¯ 2 k. Show that the fraction reﬂected is R = β2 /(1 + β2 ). h 2. Does such a potential have a bound state? If so. Estimate the time required for the particle to tunnel to the other well. h h what is its energy? 6. h h 5.. Show that the wavefunction of a particle of mass m in an inﬁnite one-dimensional squarewell of width a returns to its original form after a quantum revival time T = 4 m a2 /π ¯ . V(x) = 0 for x < 0 and V(x) = W for x > 0 with the particles incident from −∞. 0) = 1/a sin(π x/a) [1 + 2 cos(π x/a)]. A stream of particles of mass m and energy E > 0 encounter the delta-function potential V(x) = −α δ(x). where α > 0. 0) = 2/a sin(3π x/a). Show that the fraction reﬂected is R= k−q k+q 2 . and what is its probability of occurrence? What is the expectation value of the energy? 4. Two potential wells of width a are separated by a distance L ≫ a. Its initial wavefunction is ψ(x. The energy of the particle is now measured.

Find the properly normalized ﬁrst two excited energy eigenstates of the harmonic oscillator.109). Hint: Consider the raising and lowering operators a± deﬁned in Eq. .84 QUANTUM MECHANICS where V0 > 0. as well as the expectation value of the potential energy in the nth energy eigenstate. h 8. Demonstrate that the bound-states of a particle of mass m and energy −V0 < E < 0 satisfy tan 2 m (V0 + E) L/¯ = − (V0 + E)/(−E). (5.

. 4. xN . This wavefunction is interpreted as follows. dxN . The probability of ﬁnding the particle between x and x + dx at time t is given by |ψ(x. t). . x2 . This interpretation only makes sense if the wavefunction is normalized such that ∞ |ψ(x.2 Fundamental Concepts We have already seen that the instantaneous state of a system consisting of a single nonrelativistic particle. respectively. By analh ogy. t)|2 dx = 1 −∞ (6. The probability of ﬁnding the ﬁrst particle between x1 and x1 + dx1 . xN . . one-dimensional formulation of nonrelativistic quantum mechanics introduced in the previous sections in order to investigate one-dimensional chapters containing multiple particles. .1 Introduction In this chapter. . whereas the corresponding momentum is represented by the differential operator ∂ .6). The physical signiﬁcance of this normalization requirement is that the probability of the particle being found anywhere on the x-axis must always be unity (which corresponds to certainty). In a single-particle system. . 6. labeled i = 1.1) at all times. . moving in one dimension. . t)|2 dx1 dx2 . . position is represented by the algebraic operator x.. t)|2 dx. the position of the ith particle is represented by the algebraic operator xi . of the ith particle. (6.3) pi = −i ¯ h ∂xi . Consider a system containing N non-relativistic particles. By analogy with the single-particle case. t)|2 dx1 dx2 . xN space.2) at all times. N. . whose position coordinate is x. . dxN = 1 (6. x2 . the instantaneous state of a multiparticle system is speciﬁed by a complex wavefunction ψ(x1 . .Multi-Particle Systems 85 6 Multi-Particle Systems 6. Let xi and mi be the position coordinate and mass. we shall extend the single particle. . is fully speciﬁed by a complex wavefunction ψ(x. x2 . in a multi-particle system. . at time t is given by |ψ(x1 . t). xN . where the integration is taken over all x1 x2 . etc. It follows that the wavefunction must satisfy the normalization condition |ψ(x1 . . . whereas momentum is represented by the differential operator −i ¯ ∂/∂x (see Sect. . the second particle between x2 and x2 + dx2 . .

xj ] = 0. x2 . pj ] = i ¯ δij . . . from the above commutation relations.10.4)–(6. x2 . t) satisﬁes the usual time-dependent Schr¨dinger equation o [see Eq. . (4.86 QUANTUM MECHANICS Since the xi are independent variables (i. . xN .6) Now. 6. h (6. (6. that the only restriction on measurement in a one-dimensional multi-particle system is that it is impossible to simultaneously measure the position and momentum of the same particle. we conclude that the various position and momentum operators satisfy the following commutation relations: [xi . . that two dynamical variables can only be (exactly) measured simultaneously if the operators which represent them in quantum mechanics commute with one another. t) = pi2 + V(x1 . [xi . a multi-particle state of deﬁnite energy E (i. . .160)] H ψE = E ψE . an eigenstate of the Hamiltonian with eigenvalue E) is written (see Sect. that a knowledge of the position or momentum of a given particle does not in any way preclude a similar knowledge for a different particle. .5) (6. x2 . .3 Non-Interacting Particles In general. [pi . The commutation relations (6. it is clear. 2 mi i=1. t). . if H(x1 . x2 . .10) Here. as well as the interaction of the particles with any external forces. . (6. Note. we know. . xN .63)] ∂ψ i¯ h = H ψ. the ﬁrst term on the right-hand side represents the total kinetic energy of the system.. . .7) ∂t Likewise. In this case. . .12) h ψ(x1 . from Sect.6) illustrate an important point in quantum mechanics: namely. x2 .e. pj ] = 0. . .9) Here. ∂xi /∂xj = δij ). 4. (6. in particular. . . H is assumed not to be an explicit function of t. Thus. 4.e. t) is the Hamiltonian of the system then the multi-particle wavefunction ψ(x1 . Finally. .8) where the stationary wavefunction ψE satisﬁes the time-independent Schr¨dinger equation o [see Eq. xN ) e−i E t/¯ . . we expect the Hamiltonian of a multi-particle system to take the form H(x1 .4) (6. xN . t) = ψE (x1 . . whereas the potential V speciﬁes the nature of the interaction between the various particles making up the system.N (6. xN . that operators corresponding to different degrees of freedom of a dynamical system tend to commute with one another. x2 . . xN .. (4. the different degrees of freedom correspond to the different motions of the various particles making up the system.

t) = i=1. . . ψ(x1 .7) o for a system of N non-interacting particles factorizes into N independent equations of the form ∂ψi i¯ h = Hi ψi .12) and (6.. .e. V(x1 . This probability is completely unaffected by the positions of the other particles.16) ∂t Assuming that V(x. . x2 . t). According to Eqs. for the case of non-interacting particles. (6. (6.14) illustrates an important point in quantum mechanics: namely. . ψN (xN .e. . the multi-particle Hamiltonian of the system can be written as the sum of N independent single-particle Hamiltonians. . . |ψi (xi .17) .. . t) ψ2(x2 . t) = ψ1 (x1 .11) Hence. −∞ (6. . 2 mi (6.14) Here. . It is evident that ψi (xi . This implies that each particle moves in a common potential: i. xN . t) . x2 . the time-dependent Schr¨dinger equation (6. (6. t) must satisfy the normalization constraint ∞ |ψi (xi . t) can be written as the product of N independent single-particle wavefunctions: i.N V(xi . . t)|2 dxi is the probability of ﬁnding the ith particle between xi and xi + dxi at time t. xN . (6. Here. xN . Hi represents the energy of the ith particle. (6. we expect their instantaneous positions to be completely uncorrelated with one another. t). .13) In other words.N Hi (xi . t)|2 dxi = 1. . t) = i=1. xN . . t) ≡ V(x).Multi-Particle Systems 87 Suppose that the particles do not interact with one another. we can write H(x1 . Furthermore.2) for the multi-particle wavefunction is automatically satisﬁed. This immediately implies that the multi-particle wavefunction ψ(x1 . .14). Equation (6. given that the various particles which make up the system are non-interacting. . that we can generally write the total wavefunction of a many degree of freedom system as a product of different wavefunctions corresponding to each degree of freedom. and is completely unaffected by the energies of the other particles.9) also faco torizes to give Hi ψEi = Ei ψEi . t). (6. x2 . t). the time-independent Schr¨dinger equation (6. x2 .12) where Hi = pi2 + V(xi .15) If this is the case then the normalization constraint (6.

.24) (6. . when expressed in terms of the new variables. the Hamiltonian of the system is written ¯ 2 ∂2 h ¯ 2 ∂2 h H(x1 . and E= i=1.88 QUANTUM MECHANICS where ψi (xi . H(x .25) Hence.20) where ψE (x1 . ψEN (xN ). x2 ) = − − + V(x1 − x2 ). . x ′ and X. (6. According to Eqs. t) = ψEi (xi ) exp(−i Ei t/¯ ). . . xN ) e−i E t/¯ . 2 2 2 m1 ∂x1 2 m2 ∂x2 Let x ′ = x1 − x2 be the particles’ relative position. . for the case of non-interacting particles. 6.4 Two-Particle Systems Consider a system consisting of two particles. mass m1 and m2 . .18) (6. x2 . . . ∂x1 m1 + m2 ∂X ∂x ∂ m2 ∂ ∂ = − ′.19) (6. and X= m1 x1 + m2 x2 m1 + m2 (6. t) = ψE (x1 . Hence. the Hamiltonian becomes ¯ 2 ∂2 h ¯ 2 ∂2 h − + V(x ′ ).27) .10). . x2 . x2 . a h multi-particle state of deﬁnite energy E has a wavefunction of the form h ψ(x1 . It is easily demonstrated that ∂ m1 ∂ ∂ = + ′. and Ei is the energy of the ith particle. X) = − 2 ′2 2 M ∂X 2 µ ∂x ′ (6. the energy of the whole system is simply the sum of the energies of the component particles.23) (6. Clearly. . (6.N Ei . .21) the position of the center of mass.26) where M = m1 + m2 (6.22) (6. . ∂x2 m1 + m2 ∂X ∂x (6.3) and (6. interacting via the potential V(x1 − x2 ) which only depends on the relative positions of the particles. xN ) = ψE1 (x1 ) ψE2 (x2 ) . . xn .

ψ(x1 . moving in the potential V(x1 − x2 ). from Eqs.32) X/¯ −E ′ t/¯ ) h h . Note that the total momentum of the system can be written P = −i ¯ h ∂ ∂ + ∂x1 ∂x2 = −i ¯ h ∂ . (6. the time-dependent Schr¨dinger equation (6. It is clear. moving in the potential V(x). . x2 .e. X) = Hx ′ (x ′ ) + HX (X)] suggests. As before. It follows that ψX = ψ0 .5 Identical Particles Consider a system consisting of two identical particles of mass m.34) In other words.31) (6. H(x ′ . and E ′ = P ′ 2 /2 M are constants. Suppose that we work in the centre of mass frame of the system.Multi-Particle Systems is the total mass of the system. (6. the instantaneous state of the system is speciﬁed by the complex wavefunction ψ(x1 . which is characterized by P ′ = 0.. that the wavefunction can be factorized: i.29).33) where ψ0 . x2 . This is a familiar result from classical dynamics.e. i¯ h ∂t 2 M ∂X2 The above equation can be solved to give ψX (X. are equivalent to a single particle of mass µ. where x = x1 − x2 . t). t) = ψx ′ (x ′ . momentum is conserved. ∂t 2 µ ∂x 2 (6.26) is separable when expressed in terms of the new coordinates [i.33). t) = ψ0 e i (P ′ (6. ∂t 2 µ ∂x ′ 2 (6.e. two particles of mass m1 and m2 . ∂X (6. x2 . and (6. that the total momentum of the system takes the constant value P ′ : i.30) ¯ 2 ∂2 ψX h ∂ψX =− . P ′ . t) = ψx ′ (x ′ . However. 6. and µ= m1 m2 m1 + m2 89 (6. In this case. Hence. where ∂ψ ¯ 2 ∂2 ψ h i¯ h =− + V(x) ψ.28) the so-called reduced mass..29) The fact that the Hamiltonian (6.7) also factorizes to give o i¯ h and ∂ψx ′ ¯ 2 ∂2 ψx ′ h =− + V(x ′ ) ψx ′ . t). (6. t) ψ0 ≡ ψ(x1 − x2 . t).. we can write the wavefunction of the system in the form ψ(x1 . in the center of mass frame. by analogy with the analysis in the previous section. t) ψX(X.30).

if we swap the labels on particles 1 and 2 (which are.39) (6. x1 . x2 . stationary wavefunction corresponding to a state of deﬁnite energy E. Ea ) ψ(x2 .90 QUANTUM MECHANICS the only thing which this wavefunction tells us is that the probability of ﬁnding the ﬁrst particle between x1 and x1 + dx1 . t)|2 dx1 dx2 . and repeat the argument. t) = −ψ(x1 . Particles with wavefunctions which are symmetric under label interchange are said to obey Bose-Einstein statistics. (6. we infer that for a system consisting of two identical particles. at time t (since. Particles with wavefunctions which are anti-symmetric under label interchange are said to obey Fermi-Dirac statistics. this must be the same as the probability of ﬁnding the ﬁrst particle between x2 and x2 +dx2 . x1 . and are called fermions—for instance. since the particles are . t) = e i ϕ ψ(x1 . Eb) + ψ(x2 . x1 . Consider a system containing two identical and non-interacting bosons. However. and are called bosons—for instance. x2 . x2 ) = √ [ψ(x1 . and neutrons are fermions. It turns out that whether the wavefunction of a system containing many identical particles is symmetric or anti-symmetric under interchange of the labels on any two particles is determined by the nature of the particles themselves. Thus. arbitrary for identical particles).37) (6. x2 .e. electrons. t) = ψ(x1 . t). the physical outcome of the measurement is exactly the same). Hence. t)|2. The stationary wavefunction of the whole system is written 1 ψE boson (x1 . x2 . t) = e i ϕ ψ(x2 . (6. e 2 i ϕ = 1. after all. This expression automatically satisﬁes the symmetry requirement on the wavefunction. t). or ψ(x1 . Hence. x1 .35) when the energies of the two particles are Ea and Eb . since the particles are identical. t).38) The only solutions to the above equation are ϕ = 0 and ϕ = π.41) (6. at time t is |ψ(x1 . and the second between x1 and x1 + dx1 . E) be a properly normalized. photons are bosons. t). Ea ) ψ(x1. x1 . and the second between x2 and x2 + dx2 .36) where ϕ is a real constant.40) The above argument can easily be extended to systems containing more than two identical particles. x2 . we also conclude that ψ(x2 . in both cases. we conclude that |ψ(x1 . or ψ(x2 . 2 (6. single-particle. protons. However. the wavefunction must be either symmetric or anti-symmetric under interchange of particle labels: i. Let ψ(x.. t)|2 = |ψ(x2 . either ψ(x2 . x2 . Incidentally. Eb )] . (6.

(6.45) For the case of two identical bosons. using the above three wavefunctions. Finally. so that ∞ ψ∗ (x.b ) dx. Eb )] . on average. and the other Eb . Eb) dx = 0. in quantum mechanics. Ea. between the two particles. Ea ) ψ(x2 .b ) xn ψ(x.49) Equation (6. Eb) dx.42) Again. (6. It is easily demonstrated that if the particles are distinguishable then (x1 − x2 )2 dist = x2 a + x2 b − 2 x a x b . 2 (6. Now. (6. we have assumed that Ea = Eb .b ∞ = −∞ ψ∗ (x. Ea ) ψ(x. the stationary wavefunction of the whole system takes the form 1 ψE fermion (x1 . on average. (6. Ea.44) where xn a. for the case of two identical fermions. We thus conclude that it is impossible for the two fermions in our system to occupy the same single-particle stationary state.43) Let us evaluate the variance of the distance. (6. For a system consisting of two identical and non-interacting fermions. we cannot be sure which particle has energy Ea . (6. Ea ) ψ(x2 . we ﬁnd (x1 − x2 )2 where x ab boson = (x1 − x2 )2 dist − 2| x 2 ab | .48) Finally.46) ∞ = −∞ ψ∗ (x. Note that if Ea = Eb then the total wavefunction becomes zero everywhere. x2 ) = ψ(x1 . Eb ) − ψ(x2 . Ea ) x ψ(x. this expression automatically satisﬁes the symmetry requirement on the wavefunction. and which has energy Eb —only that one particle has energy Ea . Ea ) ψ(x1 . . a null wavefunction corresponds to the absence of a state. we obtain (x1 − x2 )2 fermion = (x1 − x2 )2 dist + 2| x 2 ab | .Multi-Particle Systems 91 identical. −∞ (6.46) shows that the symmetry requirement on the total wavefunction of two identical bosons forces the particles to be. Conversely. closer together than two similar distinguishable particles.49) shows that the symmetry requirement on the total wavefunction of two identical fermions forces the particles to be. x2 ) = √ [ψ(x1 . Eb ). x1 − x2 . if the two particles are somehow distinguishable then the stationary wavefunction of the system is simply ψE dist (x1 . Eq. (6.47) Here.

If one particle is in the ground-state. What are the four lowest energies of the system? What are the degeneracies of these energies if the two particles are (a) distinguishable. 3. the energies of the particles are E1 .) 1. ψa (x). ψ(xN . . Consider a system consisting of two non-interacting particles. . E2 ) . are in a one-dimensional box of length a.50) This expression is known as the Slater determinant. that if these two wavefunctions do not overlap to any great extent then identical bosons or fermions will act very much like distinguishable particles. . (b) indistinguishable bosons. . Ea ) and ψ(x. How many different two-particle states can be constructed if the particles are (a) distinguishable. . or (c) indistingishable fermions? 4. E2 ) ψ(x2 . (b) indistinguishable bosons. or (c) indistinguishable fermions. E2. and ψc (x). . .e. EN. It is evident. again. and the other in the ﬁrst excited state. calculate (x1 − x2 )2 assuming that the particles are (a) distinguishable. Neglect spin in the following questions. E1 ) . . . . ψ(xN . that if any two particles in the system have the same energy (i. . Write the properly normalized wavefunctions if the particles are (a) distinguishable. . (b) indistinguishable bosons. x2 . Two non-interacting particles. E2 ) 1 ψE (x1 . and three one-particle states. For a system containing N identical and non-interacting fermions. E1 ) ψ(x1 . . Two particles in a one-dimensional box of length a occupy the n = 4 and n ′ = 3 states. . . Consider two non-interacting particles. EN ) ψ(x2 . Exercises (N. . . xN ) = √ . . or (c) indistinguishable fermions. . .B. . N! . This important result is known as the Pauli exclusion principle. each of mass m. or (c) indistinguishable fermions? 2. EN ) . the strength of this effect depends on square of the magnitude of x ab . with the same mass m. Eb ). ψ(x1 . (b) indistinguishable bosons. and automatically satisﬁes the symmetry requirements on the wavefunction. However. We conclude that it is impossible for any two identical fermions in a multi-particle system to occupy the same single-particle stationary state. Note. E1 ) ψ(x2 . EN ) (6. Here. ψb (x). in a one-dimensional harmonic oscillator potential of classical oscillation frequency ω. the anti-symmetric stationary wavefunction of the system is written ψ(x1 .. . which measures the overlap between the wavefunctions ψ(x. . . ψ(xN . . if Ei = Ej for some i = j) then the total wavefunction is null. then.92 QUANTUM MECHANICS further apart than two similar distinguishable particles.

we shall extend our previous one-dimensional formulation of non-relativistic quantum mechanics to produce a fully three-dimensional theory.5) ∂j ∂|ψ|2 + = 0. the probability of ﬁnding the particle at time t between x and x + dx. −∞ −∞ −∞ (7. t) dx dy dz. t)|2 . (7. t) = |ψ(x. z. z. ψ(x. t). By analogy with the one-dimensional case. ∂t ∂x where j= ∂ψ∗ ∂ψ i¯ h ψ − ψ∗ 2m ∂x ∂x (7. where P(x. ψ(x.3) As usual.5) (7.2) at all times. we can write the probability conservation equation (see Sect.2 Fundamental Concepts We have seen that in one dimension the instantaneous state of a single non-relativistic particle is fully speciﬁed by a complex wavefunction. where P(x. t)|2 dx dy dz = 1. t)|2 dx = 1 −∞ (7. y. z. In three dimensions. y. y. 4.4) This normalization constraint ensures that the probability of ﬁnding the particle anywhere is space is always unity. the instantaneous state of a single particle is also fully speciﬁed by a complex wavefunction. this interpretation of the wavefunction only makes sense if the wavefunction is normalized such that ∞ ∞ ∞ |ψ(x. y. y.1 Introduction In this chapter.6) . the wavefunction is normalized such that ∞ (7. 7. z.1) |ψ(x. t)|2. t) = |ψ(x. t). between y and y + dx. The probability of ﬁnding the particle at time t between x and x + dx is P(x. In one dimension. and between z and z + dz. z. is P(x. Moreover. t) dx.Three-Dimensional Quantum Mechanics 93 7 Three-Dimensional Quantum Mechanics 7.

respectively. have the following representations: px ≡ −i ¯ h py pz ∂ .94 QUANTUM MECHANICS is the ﬂux of probability along the x-axis. and z are represented by the algebraic operators x.12) (7. Integrating Eq.13) (7. By analogy.10) the ﬂux of probability along the y-axis. and making use of the fact that ψ → 0 as |r| → ∞ if ψ is to be square-integrable. Integrating Eq. Since the xi are independent variables (i. ∂xi /∂xj = δij ). by analogy with the one dimensional case. we obtain d dt ∞ ∞ ∞ |ψ(x. z. etc.6). the normalization of the wavefunction is again preserved as time progresses. This is a necessary criterion for the viability of our basic interpretation of |ψ|2 as a probability density. and z. y. and p1 = px . ∂x ∂ ≡ −i ¯ h . if the wavefunction is initially normalized then it stays normalized as time progresses. as must be the case if |ψ|2 is to be interpreted as a probability density. ∂z (7. ∂y ∂ ≡ −i ¯ h . (7. x3 = z. whereas the three Cartesian components of momentum. whereas momentum is represented by the differential operator −i ¯ ∂/∂x (see Sect. In one dimension.8) ∂t ∂x ∂y ∂z Here. we obtain d dt ∞ |ψ(x.e. and ∂ψ∗ ∂ψ i¯ h ψ − ψ∗ jy = 2m ∂y ∂y (7. and making use of the fact that ψ → 0 as |x| → ∞ if ψ is to be square-integrable. (7. ∂ψ∗ ∂ψ i¯ h ψ (7. In three dimensions. y. we conclude that the various position and momentum operators satisfy the .9) − ψ∗ jx = 2m ∂x ∂x is the ﬂux of probability along the x-axis. (7. t)|2 dx = 0. t)|2 dx dy dz = 0. the probability conservation equation becomes ∂|ψ|2 ∂jx ∂jy ∂jz + + + = 0. (7.14) Let x1 = x. py .5) over all space. and pz . etc. 4.8) over all space.11) −∞ −∞ −∞ Thus. y.. px . the Cartesian coordinates x. position is represented by the algebraic operator x. (7. in h three dimensions. x2 = y.7) −∞ In other words.

21).10. (7. The commutation relations (7. The same equation governs the time evolution of the wavefunction in three dimensions. [xi . this expression generalizes to i¯ h 2 2 px + py + pz2 H= + V(x.17) again illustrate the point that quantum mechanical operators corresponding to different degrees of freedom of a dynamical system (in this case. Note.12)–(7.17) Now.15)–(7. or two different components of the momentum.2). making use of Eqs. pj ] = i ¯ δij . motion in different directions) tend to commute with one another (see Sect.19) H= 2m where V(x) is the potential energy. h 95 (7. that the only restriction on measurement in a system consisting of a single particle moving in three dimensions is that it is impossible to simultaneously measure a given position coordinate and the corresponding component of momentum. the differential operator ∂ψ ¯2 2 h =− ∇ ψ + V ψ.1)] o i¯ h Here. from the above commutation relations. however. 4. ∂t 2m ∂2 ∂2 ∂2 + 2+ 2 ∂x2 ∂y ∂z (7. t). (7.21) ∇2 ≡ (7. we know. the probability conservation equation (7. 6.14) and (7.22) is known as the Laplacian. pj ] = 0. In three dimensions. t).63)] ∂ψ = H ψ. that two dynamical variables can only be (exactly) measured simultaneously if the operators which represent them in quantum mechanics commute with one another. the Hamiltonian of a non-relativistic particle of mass m takes the form 2 px + V(x.18).20) 2m Hence.8) is easily derivable from Eq. from Sect. z. Thus. it is clear. (4. that it is perfectly possible to simultaneously measure two different positions coordinates. (7. Incidentally. An eigenstate of the Hamiltonian corresponding to the eigenvalue E satisﬁes H ψ = E ψ.15) (7. the three-dimensional version of the time-dependent Schr¨ndiger equation becomes [see Eq. y. in one dimension. In one dimension.Three-Dimensional Quantum Mechanics following commutation relations: [xi . [pi . Now. the time evolution of the wavefunction is given by [see Eq. (7. (7.16) (7. (4. (7.18) ∂t where H is the Hamiltonian.23) . xj ] = 0.

The particle’s stationary wavefunction. (4.30) (7. ψ(x. ky . t) = ψ(x.26) where E is the particle energy. spatial constants. y. (7. QUANTUM MECHANICS (7. The solutions to the above . (7.12) h ψ(x. because it would not otherwise be possible to satisfy the boundary conditions. and Z(0) = Z(a) = 0. and rearranging. Substituting (7. Y(0) = Y(a) = 0. 7. satisﬁes ∂2 ∂2 ∂2 2m + 2 + 2 ψ = − 2 E ψ.2). The wavefunction satisﬁes the boundary condition that it must be zero at the edges of the box. X Y ′′ 2 = −ky . (7.28) X Y Z ¯ h where ′ denotes a derivative with respect to argument. z). Z (7. z. ∂x2 ∂y ∂z ¯ h (7. y. and kz2 are spatial constants.18) that (see Sect. ¯2 h (7. y. we obtain X ′′ Y ′′ Z ′′ 2m + + = − 2 E. 5. Y Z ′′ = −kz2 . y. 4.159)]: o ∇2 ψ = 2m (V − E) ψ. z) = X(x) Y(y) Z(z). rather than positive.27) The factors of the wavefunction satisfy the boundary conditions X(0) = X(a) = 0. y.96 It follows from Eq. Note that the right-hand sides of the above equations must contain negative.31) 2 2 where kx . (7.25) where V is assumed not to depend explicitly on t. z) e−i E t/¯ .3 Particle in a Box Consider a particle of mass m trapped inside a cubic box of dimension a (see Sect. Let us search for a separable solution to the above equation of the form ψ(x.24) where the stationary wavefunction ψ(x. z) satisﬁes the three-dimensional version of the time-independent Schr¨ndiger equation [see Eq.27) into Eq.26).29) (7. It is evident that the only way in which the above equation can be satisﬁed at all points within the box is if X ′′ 2 = −kx .

35) (7. which can take one of two different values. Let us treat the electrons as essentially non-interacting particles. lx .9)] l2 π2 ¯ 2 h E= . Furthermore.28)–(7. ly . and lz . 10). The spin states of an electron are governed by an additional quantum number. and lz .36) (7. = a (7. a lz π . because electrons possess an intrinsic angular momentum called spin (see Cha.38) 2 m a2 where 2 2 l2 = lx + ly + lz2 .33) (7. are [see Eq. It turns out that this is not quite true.39) 7. The exclusion principle states that no two electrons in our system can occupy the same single-particle energy level. Thus. Now. ly . a (7. and satisfy the boundary conditions. (7.Three-Dimensional Quantum Mechanics 97 equations which are properly normalized. the single-particle energy levels for a particle in a box are characterized by the three quantum numbers lx .32) (7. the energy is written [see Eq. and lz are positive integers. when spin is taken into account. (7. ly . (5. the total energy of a system consisting of many non-interacting particles is simply the sum of the single-particle energies of the individual particles.5). Thus.37) 2 sin(kx x). 6. a 2 sin(kz z). from Eqs. 6.34) Here. from the previous section. Hence.11)] X(x) = Y(y) = Z(z) = where kx = ky kz lx π . we conclude that a maximum of two electrons (with different spin quantum . electrons are subject to the Pauli exclusion principle (see Sect. since they are indistinguishable fermions. (5.31). a ly π = . a 2 sin(ky y).3. we conclude that no two electrons in our system can have the same set of values of lx .4 Degenerate Electron Gases Consider N electrons trapped in a cubic box of dimension a. (7. According to Sect.

the distance between nearest neighbour energy levels is unity. ly . we would expect them to ﬁll the lowest single-particle energy levels available to them. from Eqs. .38). and lz can only take positive values. and lz . N. lF = 3N π 1/3 . and lz can only take positive values. and the factor 1/8 is to take account of the fact that lx . We can imagine the single-particle energy levels as existing in a sort of threedimensional quantum number space whose Cartesian coordinates are lx . N. In this case. This implies that the number of energy levels per unit volume is also unity. The number of energy levels contained in a sphere of radius l is approximately equal to the volume of the sphere— since the number of energy levels per unit volume is unity. it follows that if the number of electrons. Note. Note that the Fermi energy only depends on the number density of the conﬁned electrons. we can write 1 4π 3 N=2× × l .39). from the origin. (7. they have comparatively little thermal energy. (7.43) where n = N/a3 is the number of electrons per unit volume (in real space). (7. the factor 2 is to take into account the two spin states of an electron. the energy levels are uniformly distributed in this space on a cubic lattice. l2 = lx + ly + lz2 . (7. It turns out that this is not quite correct. Hence. Thus. Thus.e. Moreover. (7. Since we expect cold electrons to occupy the lowest energy levels available to them. but only two electrons can occupy a given energy level. ly . and lz . the energy of a given energy level is 2 2 proportional to its distance. is very large then the ﬁlled energy levels will be approximately distributed in a sphere centered on the origin of quantum number space. ly . ly .38) and (7. because we have forgotten that the quantum numbers lx .40) 8 3 F Here. that the associated particle energy 2 2 is proportional to l2 = lx + ly + lz2 . (7. Finally. The radius lF of the octant of ﬁlled energy levels in quantum number space can be calculated by equating the number of energy levels it contains to the number of electrons. This can also be written as π2 ¯ 2 h EF = 2 me 3n π 2/3 . Thus.. the energy of the most energetic electrons—which is known as the Fermi energy—is given by 2 π2 ¯ 2 h lF π2 ¯ 2 h = EF = 2 me a2 2 m a2 3N π 2/3 .41) According to Eq.42) where me is the electron mass. Suppose that our electrons are cold: i.98 QUANTUM MECHANICS numbers) can occupy a single-particle energy level corresponding to a particular set of values of lx . the ﬁlled energy levels actually only occupy one octant of a sphere.

if kB T ≫ EF then the electrons are hot. In turns out that the conduction (i. the pressure predicted by classical physics is P = n kB T . The total energy of a degenerate electron gas is ¯ 3 Etotal = N E = N EF . 5 Hence. and kB the Boltzmann constant.42) and (7.47) where λ is the de Broglie wavelength. Thus. P ∝ V −5/3 [see Eqs. Note that. and are essentially governed by classical physics— electrons in this state are termed non-degenerate. (7. and becomes degenerate as the mean spacing approaches the de Broglie wavelength. which is deﬁned B = −V ∂P ∂V 5/3 (7. Indeed. (7. most metals are hard to compress as a direct consequence of the high degeneracy pressure of their conduction electrons.44) since E ∝ l2 . me E p (7. To be more exact. free) electrons inside metals are highly degenerate (since the number of electrons per unit volume is very large. Now. and is due to the fact that identical fermions cannot get signiﬁcantly closer together than a de Broglie wavelength without violating the Pauli exclusion principle. the electron energy is much larger than that predicted by classical physics—electrons in this state are termed degenerate.. a degenerate electron gas has a much higher pressure than that which would be predicted by classical physics. Note that. Hence.Three-Dimensional Quantum Mechanics The mean energy of the electrons is given by ¯ E = EF lF 0 99 l2 4π l2 dl 4 3 5 π lF = EF . Thus. (7. and EF ∝ n2/3 ). 5 π3 ¯ 2 h B= P= 3 9m 3n π .46) since EF ∝ a−2 = V −2/3 [see Eq.e. the mean spacing between degenerate electrons is d ∼ n−1/3 ∼ √ h h ∼ ∼ λ.48) Now. Now. ∂V 5 (7. an electron gas is non-degenerate when the mean spacing between the electrons is much greater than the de Broglie wavelength. according to classical physics. the gas pressure takes the form P=− 2 ∂Etotal = n EF .43). Thus. where T is the electron temperature.42)]. and the energy levels are uniformly distributed in quantum number space inside an octant of radius lF .46)].45) (7.49) . resistance to compression is usually measured in terms of a quantity known as the bulk modulus. 3 5 (7. the mean thermal energy of the electrons is (3/2) kB T . in this case. On the other hand. for a ﬁxed number of electrons. (7. This is an entirely quantum mechanical effect. according to Eq. if kB T ≪ EF then our original assumption that the electrons are cold is valid.

The total energy of a white-dwarf star can be written E = K + U. (7. Now. and electrons. and the thermal pressure tending to make it expand. such as the Sun. (7. and U is the gravitational potential energy. This leads to the following estimate for the bulk modulus: B ∼ 6. even at zero temperature. What is the ultimate fate of such a star? A burnt-out star is basically a gas of electrons and ions.53) . V the volume of the star. as it radiates away its remaining thermal energy. the thermal energy of the star is generated by nuclear reactions occurring deep inside its core. is maintained in equilibrium via the balance of the gravitational attraction tending to make it collapse. so the ion gas remains non-degenerate. start to collapse. and. because the Pauli exclusion principle prevents the mean electron separation from becoming significantly smaller than the typical de Broglie wavelength (see previous section). for the sake of simplicity. Eventually.50) where K is the kinetic energy of the degenerate electrons (the kinetic energy of the ion is negligible). M is the stellar mass. and the electron gas becomes degenerate. From the previous subsection. the mean separation becomes of order the de Broglie wavelength of the electrons. The interior of a white-dwarf star is composed of atoms like C12 and O16 which contain equal numbers of protons. Let us investigate the physics of white-dwarfs in more detail. it is possible for a burnt-out star to maintain itself against complete collapse under gravity via the degeneracy pressure of its constituent electrons.100 QUANTUM MECHANICS For example. the star will run out of burnable fuel.5 White-Dwarf Stars A main-sequence hydrogen-burning star. the gravitational potential energy takes the form 3 G M2 U=− .6 × 1028 m−3 . (7. (7. Eventually.51) 5 R where G is the gravitational constant. Note. 7. and me the electron mass. M = 2 N mp. however. therefore.5 × 1010 N m−2 . that the density of the star is uniform. In this case. Let us assume.52) where N is the number of electrons. Thus. neutrons. Thus. the kinetic energy of a degenerate electron gas is simply 2 2 h ¯ = 3 N EF = 3 N π ¯ K = NE 5 5 2 me 3N πV 2/3 . its density increases. The actual bulk modulus is B = 4. a degenerate electron gas exerts a substantial pressure. Such stars are termed white-dwarfs. As the star collapses. the number density of free electrons in magnesium is n ∼ 8. that the de Broglie wavelength of the ions is much smaller than that of the electrons.4 × 1010 N m−2 . and R is the stellar radius. and so the mean separation between its constituent particles decreases. Of course.

and M⊙ = 2 × 1030 kg the solar mass.59) where R⊙ = 7 × 105 km is the solar radius. 5/3 8 G me mp M1/3 (7. R∗ . In fact.58) The above formula can also be written M⊙ R∗ = 0. Equations (7. and thus is unable to balance the gravitational pressure [which also scales as R−1 —see Eq. known as the Chandrasekhar limit. which turns out to be about 1. (7.54)]. (7. the electrons can become relativistic. about the same as the radius of the Earth.55) (7. for a sufﬁciently massive white dwarf. 2 R R (7.59) that E ∝ M4/3 .53) can be combined to give E= where 3 A = 20 B = 9π 8 2/3 101 A B − . (7. thousands of white-dwarfs have been observed.Three-Dimensional Quantum Mechanics where mp is the proton mass. all with properties similar to those described above. (7. rather that R−2 . it is easily demonstrated that 2A R∗ = . It turns out that the degeneracy pressure for relativistic electrons only scales as R−1 . Stars whose mass exceeds the Chandrasekhar limit inevitably collapse to produce extremely compact objects. It follows that electron degeneracy pressure is only able to halt the collapse of a burnt-out star provided that the stellar mass does not exceed some critical value. In other words.4 times the mass of the Sun. is that which minimizes the total energy E. 5 The equilibrium radius of the star. ¯ Note from Eqs. Hence.53).. and (7. (7. . (7. such as neutron stars (which are held up by the degeneracy pressure of their constituent neutrons).54) ¯2 h me M mp 5/3 . the mean energy of the electrons inside a white dwarf increases as the stellar mass increases. It follows that the radius of a typical solar mass white-dwarf is about 7000 km: i.56) 3 G M2 .50)–(7. Nowadays.52).57) B which yields R∗ = ¯2 h (9π)2/3 .e. The ﬁrst white-dwarf to be discovered (in 1862) was the companion of Sirius. or black holes.010 R⊙ M 1/3 .

where ρ is the density of states. This number can be written dN = ρ(E) dE. Consider a particle of mass m moving in a three-dimensional isotropic harmonic oscillator potential of force constant k. Repeat the above calculation for a two-dimensional electron gas in a square box of area L2 . Repeat the calculation of the Fermi energy of a gas of fermions by assuming that the fermions are massless. 4. calculate the Fermi energy in electron volts. 6. Given that the number density of free electrons in copper is 8. 3. At what mass does the Fermi energy equal the rest mass energy? . calculate the number of electron states in the interval E to E + dE. 5. bearing in mind that there are two electrons per energy state. Calculate the density of states of an electron gas in a cubic box of volume L3 .102 Exercises QUANTUM MECHANICS 1. Solve the problem via the separation of variables. so that the energy-momentum relation is E = p c. In other words. and the velocity of an electron whose kinetic energy is equal to the Fermi energy.5 × 1028 m−3 . Obtain an expression for the Fermi energy (in eV) of an electron in a white dwarf star as a function of the stellar mass (in solar masses). and obtain an expression for the allowed values of the total energy of the system (in a stationary state). 2.

5) where use has been made of the deﬁnitions of the Li [see Eqs. h h (8. etc. Ly ] = [y pz − z py . and commutation relations (7. ﬁrst of all.15)–(7. given that the xi and the pi are Hermitian operators. (8. L. 4.) correspond to the appropriate quantum mechanical position and momentum operators. x2 ≡ y. it is easily seen that the Li are also Hermitian. The ﬁrst point to note is that expressions (8. (8. thus. z] + x py [z. consider whether it is possible to use the above expressions as the deﬁnitions of the operators corresponding to the components of angular momentum in quantum mechanics.3) (8. 8.6).4) are plausible deﬁnitions for the quantum mechanical operators which represent the components of angular momentum. since the various position and momentum operators appearing in them all commute with one another [see Eqs.4) (8.Orbital Angular Momentum 103 8 Orbital Angular Momentum 8.1 Introduction As is well-known. z px − x pz ] = y px [pz . pz] = i ¯ (x py − y px ) = i ¯ Lz .2)–(8.4) are unambiguous with respect to the order of the terms in multiplicative factors. angular momentum plays a vitally important role in the classical description of three-dimensional motion. (7.17)].2 Angular Momentum Operators In classical mechanics. the vector angular momentum. [Lx . It follows that Lx = y pz − z py .2)–(8. conclude that Eqs.4)].2)–(8. Let us now investigate the role of angular momentum in the quantum mechanical description of such motion. of a particle of position vector r and linear momentum p is deﬁned as L = r × p. Lz = x py − y px . For instance. We. Moreover. Let us now derive the commutation relations for the Li . There are two similar commutation . assuming that the xi and pi (where x1 ≡ x. p1 ≡ px . since only Hermitian operators can represent physical variables in quantum mechanics (see Sect. (8. Ly = z px − x pz . This is important.17) for the xi and pi .2) (8.1) Let us.

10. 4. but are the Hermitian conjugates of one another (see Sect. the commutation relations (8. Finally. B] A..15) . Lx ] = [Ly . Lz] = 0. L2 . [L2. it is easily demonstrated that if A and B are two general operators then [A2 .6): i. it is helpful to deﬁne the operators L± = Lx ± i Ly . (8. Likewise. which represents the magnitude squared of the angular momentum vector.12) Recall. Ly ] = Lx + Ly + ¯ Lz h = L2 − Lz2 + ¯ Lz .11) where use has been made of Eqs. Lz . h [Ly . together with. the operators which represent them in quantum mechanics must commute with one another. Lx ] Lz = i ¯ (−Ly Lz − Lz Ly + Lz Ly + Ly Lz ) = 0. 2 [L2 . it is easily demonstrated that L2 also commutes with Ly . Thus. (L± )† = L∓ . from Sect. By convention. h (8.14) Moreover. is deﬁned 2 2 L2 = Lx + Ly + Lz2 . Lx ] + [Lz2.e. Lx ] = [L2 .13) Note that L+ and L− are not Hermitian operators. h [Lz . and one for Lz and Lx . Lx ] = i ¯ Ly .7) (8.10) = Ly [Ly . (8. Lx ] Ly + Lz [Lz . the operator L2 . (8.8) and (8. Hence. and with Lz .12) imply that we can only simultaneously measure the magnitude squared of the angular momentum vector. it is easily seen that 2 2 2 2 L+ L− = (Lx + i Ly ) (Lx − i Ly ) = Lx + Ly − i [Lx .104 QUANTUM MECHANICS relations: one for Ly and Lz .6)–(8. L2 commutes with Lx .8). B] = A [A. Lx ] + [Ly . h (8.8) By analogy with classical mechanics.6) (8. In other words. at most. Lx ] + [Lz . 4. h (8. Ly ] = i ¯ Lz . Lx ] (8. (8. Hence. that in order for two physical quantities to be (exactly) measured simultaneously. B] + [A.9) Now. Lz ] = i ¯ Lx . Collecting all of these commutation relations together.6)–(8. (8. we shall always choose to measure the z-component. one of its Cartesian components. Ly] = [L2 . we obtain [Lx .

the fundamental representation (7.28) (8.Orbital Angular Momentum Likewise. (8. It is most convenient to perform our investigation using conventional spherical polar coordinates: i.19) 8. similarly. θ.2. [L− . 7. h h h and.9). L− L+ = L2 − Lz2 − ¯ Lz .4).20) (8. and φ. after some tedious analysis. (8.22) ∂ ∂ . h 105 (8. It follows.e.21) (8.25). we ﬁnally obtain Lx = −i ¯ − sin φ h Ly = −i ¯ cos φ h Lz = −i ¯ h ∂ . that the operators.25) ∂z ∂r r ∂θ Making use of the deﬁnitions (8. in Sect. and a great deal of tedious algebra. Let us now investigate whether angular momentum h operators can similarly be represented as spatial differential operators. pi . the Eqs. Lz] = [Lx . L−] = 2 ¯ Lz .23) ∂x ∂r r ∂θ r sin θ ∂φ ∂ cos θ sin φ ∂ cos φ ∂ ∂ = sin θ sin φ + + .26) (8. that ∂ ∂ cos θ cos φ ∂ sin φ ∂ = sin θ cos φ + − . − sin φ cot θ ∂θ ∂φ . Lz] + i [Ly .24) ∂y ∂r r ∂θ r sin θ ∂φ ∂ ∂ sin θ ∂ = cos θ − . which represent the Cartesian components of linear momentum in quantum mechanics. h We also have [L+ .3 Representation of Angular Momentum Now. ∂φ ∂ ∂ .18) (8. we saw earlier. r. Lz] = −i ¯ Ly − ¯ Lx = −¯ L+ .17) (8.16) (8. (8.14) of the pi operators as spatial differential operators. z = r cos θ. and (8. h giving [L+ .2)–(8. can be represented as the spatial differential operators −i ¯ ∂/∂xi ..20)– (8. These are deﬁned with respect to our usual Cartesian coordinates as follows: x = r sin θ cos φ. − cos φ cot θ ∂θ ∂φ (8.13). (8.27) (8. (8. y = r sin θ sin φ.12)–(7. Lz] = ¯ L− .

the Yl.m (θ. 8.m (θ. ′ (8. Hence.m ) = (m − 1) ¯ (L− Yl. h (8.30) We.m (θ. 4. so that Yl∗. we can write Lz Yl. but not involving the radial coordinate. whereas the dimensionless quantities m and l parameterize the eigenvalues of Lz and L2 . it stands to reason that their eigenstates only depend on the angular coordinates θ and φ. r. We.35) .m to be both mutually orthogonal and properly normalized (see Sect. reh h spectively. thus.m ′ (θ. we expect the Yl. conclude that when the operator L+ operates on an eigenstate of Lz corresponding to the eigenvalue m ¯ it converts it to an h eigenstate corresponding to the eigenvalue (m + 1) ¯ . It is also easily demonstrated that Lz (L− Yl.32) Here.m = (L+ Lz + ¯ L+ ) Yl.34) where use has been made of Eq. φ) are the eigenstates in question. φ). Since both of these operators can be represented as purely angular differential operators. L+ is known as the raising h operator (for Lz ). θ and φ. h L2 Yl.9). φ) = m ¯ Yl. φ) = l (l + 1) ¯ 2 Yl. L+]) Yl. conclude that all of our angular momentum operators can be represented as differential operators involving the angular spherical coordinates.m ).m ).m (θ.m (θ. Thus.m h = (m + 1) ¯ (L+ Yl.18). which are m ¯ and l (l + 1) ¯ 2 . Lz (L+ Yl. + i cot θ ∂θ ∂φ QUANTUM MECHANICS ∂ 1 ∂2 1 ∂ sin θ + .29) (8.4 Eigenstates of Angular Momentum Let us ﬁnd the simultaneous eigenstates of the angular momentum operators Lz and L2 .m ) = (L+ Lz + [Lz . φ) Yl. thus. h (8.33) where dΩ = sin θ dθ dφ is an element of solid angle. (8. φ). and the integral is over all solid angle. sin θ ∂θ ∂θ sin2 θ ∂φ2 (8.m (θ. Now. Of course. h (8. φ) dΩ = δll ′ δmm ′ .31) (8.106 as well as L2 = −¯ 2 h and L± = ¯ e±i φ ± h ∂ ∂ .

m (θ. h L− Yl. where use has been made of Eq. Writing L+ Yl.m+1 Yl. It therefore follows from Eqs. when L− operates on an eigenstate of Lz corresponding to the eigenvalue m ¯ it converts it to an eigenstate corresponding to the eigenvalue (m − 1) ¯ .m .m = c+ Yl.m = c− Yl. (8.37) (8. and (8. − c+ h2 l.m+1 = [l (l + 1) − m (m + 1)] ¯ .5 Eigenvalues of Lz It seems reasonable to attempt to write the eigenstate Yl.m = [l (l + 1) − m (m − 1)] ¯ .m (θ) sin θ dθ = δll ′ .46) (8.45) that − i¯ h dΦm = m ¯ Φm .42) 8. h l. It follows that c+ c− h2 l. φ) in the separable form Yl. (8. φ) = Θl.m (θ. (8.41) These equations are satisﬁed when c± = [l (l + l) − m (m ± 1)]1/2 ¯ . h dφ (8. L− is h h known as the lowering operator (for Lz ).43) (8.28).31).47) Note.48) . and is independent of the polar angle θ.m Yl.m = c+ c− h2 l.m . 0 (8. l.m = [l (l + 1) − m (m + 1)] ¯ Yl. that the differential operator which represents Lz only depends on the azimuthal angle φ. (8.Orbital Angular Momentum 107 In other words.16). h (8.45) Θl∗′ .m Hence.m+1 .m (θ) Φm (φ).33) provided that π (8. from Eq.m = c+ h2 l.m = [l (l + 1) − m (m − 1)]1/2 ¯ Yl.m−1 cl.m = [l (l + 1) − m (m + 1)]1/2 ¯ Yl.m l.38) (8. Likewise.m ′ (θ) Θl.m we obtain L− L+ Yl.m−1 cl.m+1 . −π 2π ∗ Φm ′ (φ) Φm (φ) dφ = δmm ′ . Hence. (8. We can satisfy the orthonormality constraint (8. (8.m L− Yl.15).36) (8.m l.m−1 . we can write L+ Yl.39) where use has been made of Eq.m = [l (l + 1) − m (m − 1)] ¯ Yl. − L+ L− Yl. l.m−1 . (8.40) (8.28).44) (8.

14). φ) = L+ Yl.52) It follows from Eqs. making use of Eqs. since the proof of hermiticity involves an integration by parts in φ that has canceling contributions from φ = 0 and φ = 2π.m ) dΩ = = ∗ Yl.49) by making use of the orthonormality constraint (8. where m is an integer. and (8.108 The solution to this equation is Φm (φ) ∼ e i m φ . the probability density at a given point would not. (8. the symbol ∼ just means that we are neglecting multiplicative constants. which does not make physical sense..47). h (8.58) and (8. (4. in general. Hence.51) since ψ∗ ψ ≡ |ψ|2 is a positive-deﬁnite real quantity. φ).31)–(8. we demand that the wavefunction (8.e.m ) dΩ ∗ Yl.33) that ∗ Yl. Φm (φ + 2 π) = Φm (φ) for all φ.m ¯ 2 [l (l + 1) − m (m + 1)] Yl. we ﬁnd that (L+ Yl.53) .6 Eigenvalues of L2 Consider the angular wavefunction ψ(θ. and take the values m ¯ .49) be single-valued: i. h 8.m (L2 − Lz2 − ¯ Lz ) Yl. We know that ψ∗ (θ. m can only take integer values.m Yl. φ) ψ(θ.m dΩ = h ∗ Yl. In fact.m (L+ )† (L+ Yl.m dΩ = ¯ 2 [l (l + 1) − m (m + 1)] h = ¯ 2 [l (l + 1) − m (m + 1)] ≥ 0.m dΩ ≥ 0.m )∗ (L+ Yl. This immediately implies that the quantity m is quantized. we conclude that the eigenvalues of Lz are also quantized. QUANTUM MECHANICS (8.] Finally.m L− L+ Yl.m dΩ h ∗ Yl. 2π (8. have a unique value.49) Here. φ) dΩ ≥ 0. Hence.50) This is the properly normalized eigenstate of Lz corresponding to the eigenvalue m ¯ . (8.m (θ. our basic interpretation of a wavefunction as a quantity whose modulus squared represents the probability density of ﬁnding a particle at a particular point in space suggests that a physical wavefunction must be single-valued in space. Otherwise. Thus. We obtain eimφ Φm (φ) = √ . Now. [A more h rigorous argument is that Φm (φ) must be continuous in order to ensure that Lz is an Hermitian operator. we can easily normalize the eigenstate (8. (8.16).

32).58) . (8. If l is non-negative then the constraints (8. (8. According to Eq.Orbital Angular Momentum We. h L2 Yl.63) (8. Well. and let Yl. ∗ Yl.60) (8.58). (8.54) and (8.m− = (L+ L− + Lz2 − ¯ Lz ) Yl. only take non-negative values. Suppose we act on this eigenstate with the lowering operator L− . from a physical standpoint. thus. therefore.m )∗ (L− Yl. h and should. (8. no such state exists.m ) dΩ = leads to a second constraint: l (l + 1) ≥ m (m − 1).m− = 0.35).m− .m− . the solution to the above equation is m− = −l. Thus. from Eq. It follows that l (l + 1) = m− (m− − 1). the inequality (L− Yl.56) Without loss of generality. Now.15).55) (8.56) are equivalent to the following constraint: − l ≤ m ≤ l. A non-existent state is represented in quantum mechanics by the null wavefunction.54) (8. since l (l + 1) ¯ 2 is supposed to represent the magnitude squared of something. Likewise.m− = m− (m− − 1) Yl.61) (8.59) (8.31). thus.m dΩ ≥ 0 109 (8.m L+ L− Yl. or l (l + 1) Yl. (8. Assuming that m− is negative. ψ = 0. we can assume that l ≥ 0.62) (8. This is reasonable. this will have the effect of converting the eigenstate into that of a state with a lower value of m. However. conclude that the quantum number m can only take a restricted range of integer values. and (8.57) We. we must have L− Yl. Let us call this special value m− .m− be the corresponding eigenstate. L2 = L+ L− + Lz2 − ¯ Lz h Hence. where use has been made of (8. if m can only take a restricted range of integer values then there must exist a lowest possible value it can take. obtain the constraint l (l + 1) ≥ m (m + 1).

l (θ. Writing Yl. we obtain ¯ eiφ h This equation yields dΘl. φ) ∼ (sin θ)l e i l φ .67) (8.l (θ) e i l φ = 0. Hence. we know that L+ Yl. φ) = 0.71) (8. We can now formulate the rules which determine the allowed values of the quantum numbers l and m. dθ which can easily be solved to give Θl. Let us investigate their functional form.72) (8. it is easy to demonstrate that Yl.m (θ. +1.65) Thus. if l = 0 then m can only take the value 0. + i cot θ ∂θ ∂φ (8. of L2 and Lz are known as the spherical harmonics.−l (θ. Once l is given.68) (8.7 Spherical Harmonics The simultaneous eigenstates. and making use of Eq. if l = 2 then m can take the values −2. 2. and so on. the quantum number m can take any integer value in the range − l. The quantum number l takes the non-negative integer values 0. (8.64) The above two results imply that l is an integer. (8.l = 0.110 We can similarly show that the largest possible value of m is m+ = +l.l (θ. (8. +1. 0.70) . · · · .49)]. 8. we conclude that Yl.66) since there is no state for which m has a larger value than +l.69) ∂ ∂ Θl. l − 1. · · · 0. 3. −1. 0. (8.l (θ) e i l φ [see Eqs. Now. φ) ∼ (sin θ)l e−i l φ . −l + 1. QUANTUM MECHANICS (8.l ∼ (sin θ)l . l. (8. 1. φ) = Θl. if l = 1 then m can take the values −1.30). φ). Likewise.l (θ. Yl. +2.45) and (8.l − l cot θ Θl. · · ·. since m− and m+ are both constrained to be integers.

φ) ∼ L+ Yl.l with the lowering operator L− . The above equation yields Yl.l−1 with the lowering operator.74) We can now obtain Yl.30). and (8.82) . d 1 d (sin θ)l f(θ) . (8. Thus. sin θ dθ (8.73) + i cot θ ∂θ ∂φ where use has been made of Eq. we can obtain Yl.−l+1 e−i (l−2) φ ∼ (sin θ)l−2 1 d sin θ dθ 2 1 d sin θ dθ 2 (sin θ)2 l .76).l−1 ∼ e−i φ − which reduces to Yl. sin θ dθ (8.l−2 (θ. making use of Eq.l−1 (θ. φ) ∼ L+ Yl. sin θ dθ (8.m (θ. we can write Yl. we can show that Yl.l−1 by operating on Yl. (8. (8. φ) ∼ (sin θ)m 1 d sin θ dθ l−m (sin θ)2 l .80) (sin θ)2 l .l−2 ∼ e−i (l−2) φ d 1 + (l − 1) cot θ dθ (sin θ)l−1 1 d (sin θ)2 l . sin θ dθ (8.78) Finally.l−1 ∼ e i (l−1) φ Now. We get Yl. φ) ∼ Likewise. (8.−l+2 (θ.76) (8. (8.75) d + l cot θ (sin θ)l . dθ (8. φ) ∼ (sin θ)l−2 Likewise. (8.75). (8.77) e i (l−1) φ (sin θ)l−1 1 d (sin θ)2 l .80) reveals the general functional form of the spherical harmonics: eimφ Yl. Hence. we can show that Yl.Orbital Angular Momentum 111 Once we know Yl. (8. we obtain e i (l−2) φ Yl. ∂ ∂ Yl.l−1 ∼ L− Yl.−l+1 (θ.81) A comparison of Eqs.l−2 ∼ L− Yl.79) ∂ ∂ + i cot θ ∂θ ∂φ e i (l−1) φ (sin θ)l−1 1 d (sin θ)2 l .l ∼ e−i φ − (sin θ)l e i l φ .l−2 by operating on Yl.−l ∼ e−i (l−1) φ (sin θ)l−1 1 d (sin θ)2 l . + l cot θ f(θ) ≡ dθ (sin θ)l dθ where f(θ) is a general function.l .71).

0.112 QUANTUM MECHANICS Figure 8. respectively. The solid. (8. φ) = (−1) m 2 l + 1 (l − m)! 4π (l + m)! 1/2 Pl. and longdashed curves correspond to l.86) for m ≥ 0. m = 0.m are known as associated Legendre polynomials.81) that ∗ Yl.84) We now need to normalize our spherical harmonic functions so as to ensure that |Yl. m is assumed to be non-negative. Alternatively. φ)| 2 plotted as a functions of θ.83) Yl. Making the substitution u = cos θ. where the Pl. Here. 0. (8.m .88) .m (u) = (−1) 2l l! l l+m (l + m)! (1 − u2 )−m/2 (l − m)! 2l l! d du l+m l−m (1 − u2 )l (8. and 1. (1 − u2 )m/2 Pl. it is clear from Eqs.m (θ.m (θ.m (θ. short-dashed.m (u) = (−1) for m ≥ 0. (8.85) After a great deal of tedious analysis. (8.1: The |Yl. the normalized spherical harmonic functions are found to take the form Yl. φ) ∼ e (1 − u ) du Finally. (8. and are written Pl.77).−m ∼ Yl.87) d du (1 − u2 )l . φ)|2 dΩ = 1. and (8. we can also write l−m d imφ 2 −m/2 (1 − u2 )l . ±1.m (cos θ) e i m φ (8. and 1. (8.72).m (u.

The spherical harmonics characterized by m < 0 can be calculated from those characterized by m > 0 via the identity ∗ Yl.m (θ.91) (8. l = 1.m ′ Yl. The solid. and longdashed curves correspond to l. m = 2..m .2: The |Yl. (8.0 = 3 cos θ. 16π (8.92) (8. any function of θ and φ can be represented as a superposition of spherical harmonics.Orbital Angular Momentum 113 Figure 8. 4π Y1. and 2.−m = (−1)m Yl. short-dashed.m dΩ = δll ′ δmm ′ .0 = Y1. and l = 2 spherical harmonics are listed below: 1 Y0. φ)| 2 plotted as a functions of θ. and most importantly. h h All of the l = 0. respectively. 8π 5 (3 cos2 θ − 1). the spherical harmonics are the simultaneous eigenstates of Lz and L2 corresponding to the eigenvalues m ¯ and l (l + 1) ¯ 2 . In other words.89) The spherical harmonics are orthonormal: i. for m ≥ 0. ±1. ′ (8. 0.±1 = ∓ Y2. respectively. and 2. 4π 3 sin θ e±i φ .93) (8.e.94) .90) and also form a complete set. Yl∗. ±2. Finally.0 = √ .

A measurement of Lx yields the result ¯ . 32π QUANTUM MECHANICS (8. Consider a beam of particles with l = 1.±1 = ∓ Y2.114 Y2. 2. The Hamiltonian for an axially symmetric rotator is given by H= What are the eigenvalues of H? 2 2 Lx + Ly L2 + z .±2 = 15 sin θ cos θ e±i φ . Calculate Lx and Lx . 8. φ).95) (8. 8π 15 sin2 θ e±2 i φ . 3.96) The θ variation of these functions is illustrated in Figs. and with what probabilities? Repeat the calculation for the cases in which the measurement of Lx yields the results 0 and −¯ . h 4. 2 I1 2 I2 .2. What h values will be obtained by a subsequent measurement of Lz . Find the eigenvalues and eigenfunctions (in terms of the angles θ and φ) of Lx . A system is in the state ψ = Yl. Exercises 2 1.1 and 8.m (θ.

x3 ≡ z.2) 9. 8. k all run from 1 to 3) are elements of the so-called totally antisymmetric tensor. 3. we have seen that the Cartesian components of the momentum.1 Introduction In this chapter.Central Potentials 115 9 Central Potentials 9. φ). k not all different 0 1 if i. L.20)–(8. j. k are cyclic permutation of 1. V(r).1)] L = r × p. most convenient to work in spherical polar coordinates—r.2 Derivation of Radial Equation Now. 2. 7.25)]. we shall be searching for stationary wavefunctions. where r = x2 + y2 + z2 is the radial distance from the origin. 2. (9. 2. where x1 ≡ x. Likewise. from the above expressions. j. p. it is easily demonstrated. x3 ). 4. (9. the ǫijk (where i. k are anti-cyclic permutation of 1.5) Here. It is.7) −1 if i. and r ≡ (x1 . that the radial component of the momentum can be represented as p·r ∂ pr ≡ = −i ¯ h . is deﬁned [see Eq. φ—during such an investigation (see Sect. (9. we shall investigate the interaction of a non-relativistic particle of mass m and energy E with various so-called central potentials. 3 . θ. can be represented as (see Sect. of course.6) (9.3) pi = −i ¯ h ∂xi for i = 1. j. x2 ≡ y.1) (9. 2m (9. where the Hamiltonian takes the standard non-relativistic form p2 H= + V(r). j. (8.12) o H ψ = E ψ. This expression can also be written in the following form: Li = ǫijk xj pk .2) ∂ (9. (8. The values of the various elements of this tensor are determined via a simple rule: if i.3).4) r ∂r Recall that the angular momentum vector. θ. 3 ǫijk = . x2 . which satisfy the time-independent Schr¨dinger equation (see Sect. and the basic deﬁnitions of the spherical polar coordinates [see Eqs. ψ(r. Thus.

12).7).16)].10) It follows from Eqs. (7. pj ]) − 2 i ¯ xi pi . ǫ123 = ǫ231 = 1. h (9. we obtain L2 = ǫijk xj pk ǫilm xl pm = ǫijk ǫilm xj pk xl pm . (9. pi ]) = xi xi pj pj − i ¯ δij xi pj − xi pj pi xj − i ¯ δij xi pj h h = xi xi pj pj − xi pi pj xj − 2 i ¯ xi pi . We can achieve this by making use of the fundamental commutation relation for the xi and the pi [see Eq.11) Here. h Now. h (9. ǫ321 = ǫ132 = −1. (9. it is easily demonstrated that ǫijk ǫilm ≡ δjl δkm − δjm δkl . we have made use of the fairly self-evident result that δij ai bj ≡ ai bi . since the pi commute with one another [see Eq. therefore. (9.5) and (9. (9. Next. according to (9. ai bi ≡ a1 b1 + a2 b2 + a3 b3 . Let us calculate the value of L2 using Eq. etc.9) that L2 = xi pj xi pj − xi pj xj pi . p2 ] + [x3 . According to our new notation.13) Here. p3 ] = 3 i ¯ . whose elements are determined according to the rule δij = 1 0 if i and j the same .15) (9.12) Thus.17)]: [xi . p1 ] + [x2 . it is easily seen that Eqs. Equation (9. and ǫ112 = ǫ131 = 0. L2 = xi xi pj pj − xi pi (xj pj − [xj . L2 is the same as Li Li . h (9. if i and j different (9. Making use of this convention. We have also been careful to preserve the order of the various terms on the right-hand side of the above expression. For instance.6). (9. [xj . (7. Thus.11). pj ] = i ¯ δij . we have made use of the fact that pj pi = pi pj . L2 = xi (xi pj − [xi . and. pj ] ≡ [x1 .8) and (9. commute with all of the xi and the pi .6) are indeed equivalent. since the xi and the pi do not necessarily commute with one another.9) Here δij is the usual Kronecker delta. Now.116 QUANTUM MECHANICS Thus. (9. We now need to rearrange the order of the terms on the right-hand side of Eq. as well as Eq. according to which repeated indices are summed (from 1 to 3).6) also makes use of the Einstein summation convention. pj ]) pj − xi pj (pi xj + [xj . (9.8) Note that we are able to shift the position of ǫilm because its elements are just numbers.14) . (9.

Yl. we obtain r · p = r pr = −i ¯ r h p2 = −¯ 2 h 1 ∂ ∂ 1 ∂ L2 r + − 2 2 . therefore. will automatically commute with L2 and Lz .10. θ. This observation leads us to try the following separable form for the stationary wavefunction: ψ(r. 4. then we would have missed the ﬁnal term on the righthand side. This term originates from the lack of commutation between the xi and pi operators in quantum mechanics. Hence. ∂r where use has been made of Eq. r ∂r ∂r r ∂r ¯ r h (9. ∂ .2) to give the following expression for the Hamiltonian: H=− ¯2 h 2m 2 ∂ L2 ∂2 + − 2 2 ∂r2 r ∂r ¯ r h + V(r). discussed in Sect. φ) = R(r) Yl. We thus conclude that for a particle moving in a central potential the eigenstates of the Hamiltonian are simultaneous eigenstates of Lz and L2 . from Sect. θ and φ. φ). h 117 (9. (9. Now. Now. Moreover. standard vector analysis assumes that all terms commute with one another.5). and with Lz (see Sect.22) . (9. the above equation can be combined with Eq. Thus. 8. h (9. Equation (9.16) When expressed in more conventional vector notation.Central Potentials Hence. clear that the above Hamiltonian commutes with both Lz and L2 . the above expression becomes L2 = r2 p2 − (r · p)2 + i ¯ r · p. according to Sect.4). (9.17) can be rearranged to give h p2 = r−2 (r · p)2 − i ¯ r · p + L2 .2). It is.m (θ.3.20) Finally. 8.18) (9. using standard vector identities. and do not contain the radial polar coordinate. φ). r. It follows that the spherical harmonics are also eigenstates of the Hamiltonian.17) Note that if we had attempted to derive the above expression directly from Eq.7. Recall. (9. we obtain L2 = xi xi pj pj − xi pi xj pj + i ¯ xi pi .m (θ. any function of r.19) (9. or any differential operator involving r (but not θ and φ). L2 commutes both with itself. we have already found the simultaneous eigenstates of Lz and L2 —they are the spherical harmonics. (9. that Lz and L2 are represented as differential operators which depend solely on the angular spherical polar coordinates. Now. if two operators commute with one another then they possess simultaneous eigenstates.21) Let us now consider whether the above Hamiltonian commutes with the angular momentum operators Lz and L2 . 8. Of course.

we have labeled the function R(r) by two quantum numbers. we obtain the following differential equation which determines the radial variation of the stationary wavefunction: − ¯2 h 2m d2 2 d l (l + 1) + − Rn.31) and (8. Writing the wavefunction in the standard form ψ(r.32).l = E Rn.118 QUANTUM MECHANICS It immediately follows. making use of Eqs.27) we deduce (see previous section) that the radial function Rn. (9.l + + 1− Rn. n and l.l = 0 2 dr r dr r2 in the region 0 ≤ r ≤ a. is.l . 9.2). 2 dr r dr r2 (9.26) ∞ otherwise Clearly.l l (l + 1) + + k2 − Rn. does not appear in the above equation.24) Recall that the quantum numbers m and l are restricted to take certain integer values. squareintegrable) at r = 0.24).29) ¯2 h Deﬁning the scaled radial variable z = k r.23) (9. (9. n. does not inﬂuence either the function R(r) or the energy. and (9. The second quantum number. that Lz ψ = m ¯ ψ. h L2 ψ = l (l + 1) ¯ 2 ψ. where k2 = (9.e. related to the eigenvalue of L2 . (9.. and.3 Inﬁnite Spherical Potential Well Consider a particle of mass m and energy E > 0 moving in the following simple central potential: 0 for 0 ≤ r ≤ a V(r) = . and that it be zero at r = a (see Sect. 5. is determined by the constraint that the radial wavefunction be square-integrable. Within this region. [Note that the azimuthal quantum number. the ﬁrst quantum number.l 2 dRn. the wavefunction ψ is only non-zero in the region 0 ≤ r ≤ a. it is subject to the physical boundary conditions that it be well behaved (i. φ) = Rn.30) .25) Here. from (8.] As we shall see. of course. l. 8.l = 0.l + V Rn. Finally.l 2 dRn. (9. E.l (r) satisﬁes d2 Rn. 2 dz z dz z2 (9.l (r) Yl. θ.m (θ.6. m. φ).21). the above differential equation can be transformed into the standard form l (l + 1) d2 Rn. (9.28) 2mE . as explained in Sect. h (9.1). therefore. and the fact that Lz and L2 both obviously commute with R(r).

Sect. It can be seen that the spherical Bessel functions are oscillatory in nature.1. whereas the jl (z) M. 10. y1 (z) = − 2 − z z j0 (z) = (9. j1 (z) = z2 z cos z y0 (z) = − . and I. the yl (z) functions are badly behaved (i.31) (9. 1965). z Thus. The solid. j1 (z). the ﬁrst few spherical Bessel functions take the form sin z . However..1 and can be written jl (z) = z l 1 d − z dz l l sin z .e.34) (9. respectively. z l (9. and y1 (z).36) These functions are also plotted in Fig.1: The ﬁrst few spherical Bessel functions. New York NY.32) yl (z) = −z 1 d − z dz cos z . y0 (z). long-dashed. passing through zero many times. they are not square-integrable) at z = 0. 1 . short-dashed. z sin z cos z − .A. Stegun. Handbook of Mathematical Functions (Dover.33) (9.1. The two independent solutions to this well-known second-order differential equation are called spherical Bessel functions. 9.35) (9. and dot-dashed curves show j0 (z). z cos z sin z . Abramowitz.Central Potentials 119 Figure 9.

120

l=0 l=1 l=2 l=3 l=4 n=1 n=2 3.142 6.283 4.493 7.725 5.763 9.095 6.988 10.417 8.183 11.705 n=3 9.425 10.904 12.323 13.698 15.040

QUANTUM MECHANICS

n=4 12.566 14.066 15.515 16.924 18.301

Table 9.1: The ﬁrst few zeros of the spherical Bessel function jl (z). functions are well behaved everywhere. It follows from our boundary condition at r = 0 that the yl (z) are unphysical, and that the radial wavefunction Rn,l (r) is thus proportional to jl (k r) only. In order to satisfy the boundary condition at r = a [i.e., Rn,l (a) = 0], the value of k must be chosen such that z = k a corresponds to one of the zeros of jl (z). Let us denote the nth zero of jl (z) as zn,l . It follows that k a = zn,l , for n = 1, 2, 3, . . .. Hence, from (9.29), the allowed energy levels are ¯2 h . (9.38) 2 m a2 are listed in Table 9.1. It can be seen that zn,l is an increasing

2 En,l = zn,l

(9.37)

The ﬁrst few values of zn,l function of both n and l. We are now in a position to interpret the three quantum numbers—n, l, and m—which determine the form of the wavefunction speciﬁed in Eq. (9.27). As is clear from Sect. 8, the azimuthal quantum number m determines the number of nodes in the wavefunction as the azimuthal angle φ varies between 0 and 2π. Thus, m = 0 corresponds to no nodes, m = 1 to a single node, m = 2 to two nodes, etc. Likewise, the polar quantum number l determines the number of nodes in the wavefunction as the polar angle θ varies between 0 and π. Again, l = 0 corresponds to no nodes, l = 1 to a single node, etc. Finally, the radial quantum number n determines the number of nodes in the wavefunction as the radial variable r varies between 0 and a (not counting any nodes at r = 0 or r = a). Thus, n = 1 corresponds to no nodes, n = 2 to a single node, n = 3 to two nodes, etc. Note that, for the case of an inﬁnite potential well, the only restrictions on the values that the various quantum numbers can take are that n must be a positive integer, l must be a non-negative integer, and m must be an integer lying between −l and l. Note, further, that the allowed energy levels (9.38) only depend on the values of the quantum numbers n and l. Finally, it is easily demonstrated that the spherical Bessel functions are mutually orthogonal: i.e.,

a

jl (zn,l r/a) jl (zn ′ ,l r/a) r2 dr = 0

0

(9.39)

when n = n ′ . Given that the Yl,m (θ, φ) are mutually orthogonal (see Sect. 8), this ensures that wavefunctions (9.27) corresponding to distinct sets of values of the quantum numbers n, l, and m are mutually orthogonal.

Central Potentials

121

9.4 Hydrogen Atom

A hydrogen atom consists of an electron, of charge −e and mass me , and a proton, of charge +e and mass mp , moving in the Coulomb potential V(r) = − e2 , 4π ǫ0 |r| (9.40)

where r is the position vector of the electron with respect to the proton. Now, according to the analysis in Sect. 6.4, this two-body problem can be converted into an equivalent one-body problem. In the latter problem, a particle of mass µ= moves in the central potential e2 . (9.42) 4π ǫ0 r Note, however, that since me /mp ≃ 1/1836 the difference between me and µ is very small. Hence, in the following, we shall write neglect this difference entirely. Writing the wavefunction in the usual form, V(r) = − ψ(r, θ, φ) = Rn,l (r) Yl,m (θ, φ), it follows from Sect. 9.2 that the radial function Rn,l (r) satisﬁes − Let r = a z, with a= ¯2 h = 2 me (−E) E0 a0 , E (9.45) ¯2 h 2 me d2 2 d l (l + 1) e2 Rn,l − + − + E Rn,l = 0. dr2 r dr r2 4π ǫ0 r (9.44) (9.43) me mp me + mp (9.41)

where E0 and a0 are deﬁned in Eqs. (9.57) and (9.58), respectively. Here, it is assumed that E < 0, since we are only interested in bound-states of the hydrogen atom. The above differential equation transforms to d2 2 d l (l + 1) ζ + − + − 1 Rn,l = 0, 2 dz z dz z2 z where E0 2 me a e2 . 2 = 2 E 4π ǫ0 ¯ h Suppose that Rn,l (r) = Z(r/a) exp(−r/a)/(r/a). It follows that ζ= d l (l + 1) ζ d2 Z = 0. −2 − + 2 dz dz z2 z (9.47) (9.46)

(9.48)

122

QUANTUM MECHANICS

We now need to solve the above differential equation in the domain z = 0 to z = ∞, subject to the constraint that Rn,l (r) be square-integrable. Let us look for a power-law solution of the form Z(z) =

k

c k zk .

(9.49)

**Substituting this solution into Eq. (9.48), we obtain ck k (k − 1) zk−2 − 2 k zk−1 − l (l + 1) zk−2 + ζ zk−1 = 0.
**

k

(9.50)

Equating the coefﬁcients of zk−2 gives the recursion relation ck [k (k − 1) − l (l + 1)] = ck−1 [2 (k − 1) − ζ] . (9.51)

Now, the power series (9.49) must terminate at small k, at some positive value of k, otherwise Z(z) behaves unphysically as z → 0 [i.e., it yields an Rn,l (r) that is not squareintegrable as r → 0]. From the above recursion relation, this is only possible if [kmin (kmin − 1) − l (l + 1)] = 0, where the ﬁrst term in the series is ckmin zkmin . There are two possibilities: kmin = −l or kmin = l + 1. However, the former possibility predicts unphysical behaviour of Z(z) at z = 0. Thus, we conclude that kmin = l + 1. Note that, since Rn,l (r) ≃ Z(r/a)/(r/a) ≃ (r/a)l at small r, there is a ﬁnite probability of ﬁnding the electron at the nucleus for an l = 0 state, whereas there is zero probability of ﬁnding the electron at the nucleus for an l > 0 state [i.e., |ψ|2 = 0 at r = 0, except when l = 0]. For large values of z, the ratio of successive coefﬁcients in the power series (9.49) is ck 2 = , ck−1 k (9.52)

according to Eq. (9.51). This is the same as the ratio of successive coefﬁcients in the power series (2 z)k , (9.53) k! k which converges to exp(2 z). We conclude that Z(z) → exp(2 z) as z → ∞. It thus follows that Rn,l (r) ∼ Z(r/a) exp(−r/a)/(r/a) → exp(r/a)/(r/a) as r → ∞. This does not correspond to physically acceptable behaviour of the wavefunction, since |ψ|2 dV must be ﬁnite. The only way in which we can avoid this unphysical behaviour is if the power series (9.49) terminates at some maximum value of k. According to the recursion relation (9.51), this is only possible if ζ = n, (9.54) 2 where n is an integer, and the last term in the series is cn zn . Since the ﬁrst term in the series is cl+1 zl+1 , it follows that n must be greater than l, otherwise there are no terms in the series at all. Finally, it is clear from Eqs. (9.45), (9.47), and (9.54) that E= E0 n2 (9.55)

Central Potentials

and a = n a0 , where E0 = − and me e4 e2 = −13.6 eV, =− 8π ǫ0 a0 2 (4π ǫ0)2 ¯ 2 h 4π ǫ0 ¯ 2 h a0 = = 5.3 × 10−11 m. 2 me e

123

(9.56) (9.57)

(9.58)

Here, E0 is the energy of so-called ground-state (or lowest energy state) of the hydrogen atom, and the length a0 is known as the Bohr radius. Note that |E0 | ∼ α2 me c2 , where α = e2 /(4π ǫ0 ¯ c) ≃ 1/137 is the dimensionless ﬁne-structure constant. The fact that h |E0 | ≪ me c2 is the ultimate justiﬁcation for our non-relativistic treatment of the hydrogen atom. We conclude that the wavefunction of a hydrogen atom takes the form ψn,l,m (r, θ, φ) = Rn,l (r) Yl,m (θ, φ). (9.59)

Here, the Yl,m (θ, φ) are the spherical harmonics (see Sect 8.7), and Rn,l (z = r/a) is the solution of l (l + 1) 2 n 1 d 2 d z − + − 1 Rn,l = 0 (9.60) 2 dz z dz z2 z which varies as zl at small z. Furthermore, the quantum numbers n, l, and m can only take values which satisfy the inequality |m| ≤ l < n, (9.61)

where n is a positive integer, l a non-negative integer, and m an integer. Now, we expect the stationary states of the hydrogen atom to be orthonormal: i.e., ψ∗ ′ ,l ′ ,m ′ ψn,l,m dV = δnn ′ δll ′ δmm ′ , n (9.62)

where dV is a volume element, and the integral is over all space. Of course, dV = r2 dr dΩ, where dΩ is an element of solid angle. Moreover, we already know that the spherical harmonics are orthonormal [see Eq. (8.90)]: i.e., Yl∗,m ′ Yl,m dΩ = δll ′ δmm ′ . ′ It, thus, follows that the radial wavefunction satisﬁes the orthonormality constraint

∞

(9.63)

R∗ ′ ,l Rn,l r2 dr = δnn ′ . n

0

(9.64)

59). short-dashed. R3.66) (9.l (r)| 2 plotted as a functions of r/a0 . 1. plus our interpretation of |ψ|2 as a probability density. we can calculate ∞ rk = 0 r2+k |Rn. a0 (9.2 and 9.1 (r) = 3/2 a 9 (3 a0) 6 a0 3 a0 0 √ 2 2 r r 2 √ R3. and 2. 1− + 2 3/2 (3 a0) 3 a0 27 a0 3 a0 √ r r r 4 2 1− exp − .124 QUANTUM MECHANICS Figure 9. l = 1. and long-dashed curves correspond to n.69) (9.2 (r) = . (9. and 2.71) . 0. exp − 3 a0 27 5 (3 a0)3/2 a0 These functions are illustrated in Figs.l (r)| 2 dr. The ﬁrst few radial wavefunctions for the hydrogen atom are listed below: R1.68) (9.0 (r) = R2. 1− 3/2 (2 a0) 2 a0 2 a0 1 r r exp − . Given the (properly normalized) hydrogen wavefunction (9.0 (r) = 2 r 2r 2 r2 exp − .3.65) (9.0 (r) = 2 3/2 a0 exp − r .67) (9.2: The a0 r2 |Rn.1 (r) = √ R3. 0. 2 a0 3 (2 a0)3/2 a0 R2.70) 2 r r exp − . respectively. The solid. 9.

and m are integers).61) (and the fact that n.72) (9. 2 (l + 1/2) n3 a0 1 .3). 2. implies that the energy spectrum of a hydrogen atom is highly degenerate: i. l − 1. According to the inequality (9. For instance. l). where the angle-brackets denote an expectation value. · · · . V(r). there are 2 l + 1 different allowed values of m (i. respectively.55). there are many different states which possess the same energy. there . l = 3.e. the quantized energy levels of a bound-state depend on both n and l (see Sect. (9. −l.e.. for a given value of n.Central Potentials 125 Figure 9. It turns out that this is a special property of a 1/r potential. 1. Likewise. short-dashed. and long-dashed curves correspond to n. it can be demonstrated (after much tedious algebra) that r2 r 1 r 1 r2 1 r3 2 a0 n2 [5 n2 + 1 − 3 l (l + 1)].73) (9. The solid. 0. for a given value of l. 9.76) = = 1 . 2a n 0 = (9.74) (9. l. The fact that the energy levels of a hydrogen atom only depend on n.l (r)| 2 plotted as a functions of r/a0 . For a general central potential.3: The a0 r2 |Rn.. 2 a0 [3 n2 − l (l + 1)]. the energy levels of the bound-states of a hydrogen atom only depend on the radial quantum number n. and 3. and 3. = 2 1 = . 3 l (l + 1/2) (l + 1) n3 a0 According to Eq. −l + 1. and not on l and m.75) (9.

g. the second excited state (n = 3) is nine-fold degenerate. · · · . Equation (9. n − 1).e. Now.55). Likewise.32)].e. Note that the emission spectrum of hydrogen is quantized: i. should. the possible wavelengths of the photons emitted by a hydrogen atom as its electron makes transitions between different energy levels are 1 1 1 − 2 . 1. According to Eq. the ﬁrst excited state (n = 2) is four-fold degenerate. if ∆E is positive then the electron must absorb a photon of energy ν = ∆E/h.79) is known as the Rydberg formula. a hydrogen atom can only absorb photons which have the same ﬁxed set of wavelengths. Suppose that an electron in a hydrogen atom makes a transition from an initial state whose radial quantum number is ni to a ﬁnal state whose radial quantum number is nf .] 9. Hence. (3. characterized by the quantum numbers n.e.80) Here.. R is called the Rydberg constant. and m. and was ﬁrst explained theoretically by Bohr in 1913 using a primitive . (9. the energy of the electron will change by ∆E = E0 1 1 − 2 .126 QUANTUM MECHANICS are n different allowed values of l (i. 10). Likewise. 3 ǫ2 ¯ 3 c hc (4π) 0 h (9. they are degenerate).5 Rydberg Formula An electron in a given stationary state of a hydrogen atom. etc. and can be observed as “spectral lines” using a spectroscope.. when we take into account the two spin states of an electron (see Sect. if the state is slightly perturbed—e.097 × 107 m−1 . l. the total number of degenerate states corresponding to a given value of n is 1 + 3 + 5 + · · · + 2 (n − 1) + 1 = n2 . (9..79) =R 2 λ nf ni where R= −E0 me e4 = = 1. In practice. all states possessing the same value of n have the same energy (i. This set of wavelengths constitutes the characteristic emission/absorption spectrum of the hydrogen atom.. by interacting with a photon—then the electron can make a transition to another stationary state with different quantum numbers. The Rydberg formula was actually discovered empirically in the nineteenth century by spectroscopists. the degeneracy of the nth energy level becomes 2 n2 .78) If ∆E is negative then we would expect the electron to emit a photon of frequency ν = −∆E/h [see Eq. it is assumed that nf < ni .77) Thus. in principle. the ground-state (n = 1) is not degenerate. remain in that state indeﬁnitely. [Actually. a hydrogen atom can only emit photons with certain ﬁxed set of wavelengths. Likewise. 0. Given that λ−1 = ν/c. (9. 2 nf ni (9.

(f) The most probable value of r is a0 /Z. l = 0 transition in hydrogen. and so on. Transitions to the ground-state (nf = 1) give rise to spectral lines in the ultraviolet band—this set of lines is called the Lyman series. h (d) The expectation values of the potential and kinetic energies are 2 E and −E. Find a hydrogen atom’s recoil energy when it emits a photon in an n = 2 to n = 1 transition. which means that not all of the energy made available in the downward jump goes to the photon. 6. He+ . To conserve linear momentum. and r is the distance between the nucleus and the electron. A particle of mass m is placed in a ﬁnite spherical well: V(r) = −V0 0 for r ≤ a . What is the probability that the ion will be left in an l > 0 state? 5. and. via the emission of a fast electron which leaves the atom without perturbing the extranuclear electron. deuterium.Central Potentials 127 version of quantum mechanics. Find the ground-state by solving the radial equation with l = 0. Show the following: (a) A2 = β3 /π. l = 0 state. where a0 = (¯ 2 /me ) (4π ǫ0 /e2 ). h 2. Transitions to the ﬁrst excited state (nf = 2) give rise to spectral lines in the visible band—this set of lines is called the Balmer series. An atom of tritium is in its ground-state. Li++ . The normalized wavefunction for the ground-state of a hydrogen-like atom (neutral hydrogen. What fraction of the transition energy is the recoil energy? .) with nuclear charge Z e has the form ψ = A exp(−β r). Find the probability that it will be left in a n = 2. Consider a particle of mass m in the three-dimensional harmonic oscillator potential V(r) = (1/2) m ω2 r2 . Suddenly the nucleus decays into a helium nucleus. (e) The expectation value of r is (3/2) (a0 /Z). Show that there is no ground-state if V0 a2 < π2 ¯ 2 /8 m. Find the probability that the resulting He+ ion will be left in an n = 1. (b) β = Z/a0 . where A and β are constants. for r > a with V0 > 0 and a > 0. respectively. hence. etc. l = 1 to n = 1. an atom emitting a photon must recoil. 4. Calculate the wavelengths of the photons emitted from the n = 2. and positronium. h (c) The energy is E = −Z2 E0 where E0 = (me /2 ¯ 2 ) (e2 /4π ǫ0 )2 . Solve the problem by separation of variables in spherical polar coordinates. Exercises 1. 3. determine the energy eigenvalues of the system. Transitions to the second excited state (nf = 3) give rise to spectral lines in the infrared band—this set of lines is called the Paschen series. l = 0 state.

128 QUANTUM MECHANICS .

By convention. in quantum mechanics. Sx ] = [S2 . Sz ] = i ¯ Sx .2.2. and Sz —which represent the three Cartesian components of spin angular momentum.Spin Angular Momentum 129 10 Spin Angular Momentum 10.1) (10. 4.4) By analogy with the analysis in Sect. the Earth) can possess two types of angular momentum. Lx . by analogy with Sect. Moreover. In fact. Sy] = [S2 . it is reasonable to suppose that it possesses similar properties to orbital angular momentum. Sz] = 0. In other words.8)]. and Lz [see Eqs. it is plausible that these operators possess analogous commutation relations to the three corresponding orbital angular momentum operators.6)–(8. h (10. the analogy with classical extended objects is not entirely accurate. 8. (10. we would expect to be able to deﬁne three operators—Sx . . Sx ] = i ¯ Sy .2) (10.3) We can represent the magnitude squared of the spin angular momentum vector by the operator 2 2 S2 = Sx + Sy + Sz2.5) We thus conclude (see Sect. it is best to think of spin angular momentum as a kind of intrinsic angular momentum possessed by particles. [Sx . we shall always choose to measure the z-component. Sy . the internal motion consists of spinning about an axis passing through the center of mass).g. together with. h [Sy. just as each type has a characteristic charge and mass. Actually. for instance. quantum particles can possess both orbital angular momentum due to their motion through space (see Cha. Sy ] = i ¯ Sz . 8. since electrons.10) that we can simultaneously measure the magnitude squared of the spin angular momentum vector. 10. h [Sz . Thus. The ﬁrst type is due to the rotation of the object’s center of mass about some ﬁxed external point (e. (10.. It turns out that each type of elementary particle has a characteristic spin angular momentum. (8. one Cartesian component. the Sun)—this is generally known as orbital angular momentum.g. for a rigid object. are structureless point particles. it is easily demonstrated that [S2 . a classical extended object (e.. Sz . Ly . The second type is due to the object’s internal motion—this is generally known as spin angular momentum (since.1 Introduction Broadly speaking. 8).2 Spin Operators Since spin is a type of angular momentum. By analogy. and spin angular momentum due to their internal motion. at most.

8) (10. Thus. and Sz are Hermitian operators. (8. the element of the second space which corresponds to c χ is c∗ χ† .10) (10.11) Sz2 − ¯ Sz .e.9) (10. the spin angular momentum operators cannot be represented as differential operators in real space. Note that only the directions of our vectors have any physical signiﬁcance (just as only the shape of a regular wavefunction has any physical signiﬁcance). h 10. of χ.7) (10. it is easily demonstrated that S+ S− = S2 − Sz2 + ¯ Sz . Sz] = +¯ S− .3 Spin Space We now have to discuss the wavefunctions upon which the previously introduced spin operators act. Thus.2. if the vector χ corresponds to a particular internal state then c χ corresponds to the same state. if χ1 and χ2 are two vectors corresponding to two different internal states then c1 χ1 + c2 χ2 is another vector corresponding to the state obtained by superposing c1 times state 1 with c2 times state 2 (where c1 and c2 are complex numbers). We can do this by introducing a second. then S± are the Hermitian conjugates of one another: i. or norm.12) . as must be the case if they are to represent physical quantities. Now.130 QUANTUM MECHANICS If Sx . vector space whose elements are in one to one correspondence with the elements of our ﬁrst space. Likewise. It follows that the vectors making up our vector space must also be superposable.6) (10.. we have χ† χ ≥ 0 (10. Sz] = −¯ S+ . h S− S+ = S − 2 By analogy with Eq. The different members of this space correspond to the different internal conﬁgurations of the particle under investigation.. Let the element of the second space which corresponds to the element χ of the ﬁrst space be called χ† . Finally. We now need to deﬁne the length of our vectors. since the superposability of states is one of the fundamental assumptions of quantum mechanics.13). Thus. 8. Let us denote this number χ† χ. (10. h [S− . by analogy with Sect. Sy .e. (S± )† = S∓ . Finally. we expect the internal states of our particle to be superposable. h [S+ . Unlike regular wavefunctions. Moreover. we need to think of spin wavefunctions as existing in an abstract (complex) vector space. We shall assume that it is possible to combine χ and χ† in a multiplicative fashion to generate a real positive-deﬁnite number which we interpret as the length. Instead. the number of linearly independent internal states of the particle under investigation). or dual. spin wavefunctions do not exist in real space. where c is a complex number. we can deﬁne raising and lowering operators for spin angular momentum: S± = Sx ± i Sy . the dimensionality of our vector space is simply the number of linearly independent vectors required to span it (i.

4. (10.ms = ms ¯ χs.20) s.ms ) = (ms + 1) ¯ (S+ χs. We expect the χa to be mutually orthogonal. The dual of this state is (A χ)† ≡ χ† A† . ca χa .14) χ† χa ′ = δaa ′ . 4.. Now. Let these eigenstates take the form [see Eqs.4).17) Now. if χ† χ ′ = 0.ms ). respectively. if A corresponds to a physical variable then a measurement of A will result in one of its eigenvalues (see Sect. 10. that (10. from the commutation relations (10.e.ms = s (s + 1) ¯ 2 χs.9). (10.11).13) As before.10). In order to ensure that these eigenvalues are all real. h S2 χs..15) χ= a A measurement of χ will then yield the result a with probability |ca |2 . (10.32)]: h Sz χs. A. The eigenstates of Sz and S2 are assumed to be orthonormal: i.16) (10.18) h Sz (S+ χs.ms = δss ′ δms ms . (10. for spin angular momentum (see Sect. We shall also assume that it is possible to combine unlike states in an analogous multiplicative fashion to produce complex numbers.. a general spin state can be written as a superposition of the normalized eigenstates of A: i. they must possess simultaneous eigenstates (see Sect.ms ).10). 4.m .ms ) = (ms − 1) ¯ (S− χs. ′ ′ χ† s χs ′ . (10.4 Eigenstates of Sz and S2 Since the operators Sz and S2 commute.Spin Angular Momentum 131 for all χ. a Finally.ms . S+ and S− are indeed the raising and lowering operators. where A† is the Hermitian conjugate of A (this is the deﬁnition of an Hermitian conjugate in spin space). An eigenstate of A corresponding to the eigenvalue a satisﬁes A χa = a χa .19) Thus.10) and (10. operates on a general spin-state. Two states χ and χ ′ are said to be mutually orthogonal. when a general spin operator. or independent. and h Sz (S− χs. (10. In other words. χ. (8. 8. it coverts it into a different spin-state which we shall denote A χ.31) and (8. it is easily demonstrated. A must be Hermitian: i. The product of two unlike states χ and χ ′ is denoted χ† χ ′ . We can also normalize them such that they all have unit length.ms .e. A† = A (see Sect.e.

9).7).27) (10. Now. (10.25) . there is both a maximum and a minimum possible value that ms can take.20) yield s (s + 1) ≥ ms (ms + 1). S− χs. Likewise. S2 = S+ S− + Sz2 − ¯ Sz . this equation yields ms min = −s. we obtain a multiple of χs. (10.16).17). S2 χs. Assuming that s ≥ 0. (10. (10. after acting upon χs. if χ = S− χs.−s . It follows that (see Sect.23) Hence.s .m + s. converts it into some multiple of χs.ms ) = χ† s S† S+ χs. the above two inequalities imply that − s ≤ ms ≤ s. Equations (10. Employing the raising operator a second time. (10. at ﬁxed s.−s = S+ χs. However. (10. Assuming that ms min < 0. acting upon χs.22) Likewise.21) where use has been made of Eq.ms = χ† s S− S+ χs. Indeed. (10.ms min . 8.28) (10. from Eq. the raising operator S+ . that χ† χ ≥ 0. so that employing the raising operator one more time (10. s.12).ms )† (S+ χs. Since we know.ms min = 0. h Hence.31) Now.8). and (10.−s+2 .30) (10. it follows that (S+ χs.ms .−s+1 . we must obtain a multiple of χs. this process cannot continue indeﬁnitely. Let ms min be the minimum possible value of ms .24) (10.26) (10.ms ≥ 0. since there is a maximum possible value of ms . from Eq.29) (10.−s a sufﬁcient number of times with the raising operator S+ . it is easily demonstrated that ms max = +s.ms then we obtain s (s + 1) ≥ ms (ms − 1).6) S− χs.ms min = (S+ S− + Sz2 − ¯ Sz ) χs. (10. Moreover.132 QUANTUM MECHANICS Consider the wavefunction χ = S+ χs.m (10. h giving s (s + 1) = ms min (ms min − 1).s = 0.

Up to now. the origin of these restrictions is the representation of the orbital angular momentum operators as differential operators in real space (see Sect. our analysis has been very similar to that which we used earlier to investigate orbital angular momentum (see Sect. There is no equivalent representation of the corresponding spin angular momentum operators.5). In 1940. from Eqs. It thus follows. including electrons and protons. 8. This implies that the quantum number l. If this is not the case then we will inevitably obtain eigenstates of Sz corresponding to ms > s. whereas all bosons possess integer spin (i. as well as integer. all presently known fermions. Hence. (10. According to this theorem. h 2 3 2 ¯ χ± . Wolfgang Pauli proved the so-called spin-statistics theorem using relativistic quantum mechanics. values.. However. possess spin onehalf.33) Note that χ+ corresponds to an electron whose spin angular momentum vector has a positive component along the z-axis. Recall. we conclude that there is no reason why the quantum number s cannot take half-integer.e.5 Pauli Representation Let us denote the two independent spin eigenstates of an electron as χ± ≡ χ1/2.. a half-integer value of s). Hence.±1/2 .Spin Angular Momentum 133 leads to the null state [see Eq. 8. an integer value of s). Likewise.36) . the quantum number s can either take positive integer or positive half-integer values. In other words. 10.31)]. h S2 χ± = 4 (10. all fermions possess half-integer spin (i.32) where k is a positive integer.3). is also restricted to take integer values. that for orbital angular momentum the quantum number m. It follows. that ms max − ms min = 2 s = k. 8). from the above argument. In fact.e. electrons and protons are characterized by s = 1/2 and ms = ±1/2.16) and (10. is restricted to take integer values (see Cha. we could say that the spin vector points in the +z-direction (or its spin is “up”).17). that 1 Sz χ± = ± ¯ χ± . which is analogous to ms .34) (10. + − (10. which we have already demonstrated is impossible. (10. (10.35) (10. These two eigenstates satisfy the orthonormality requirements χ† χ+ = χ† χ− = 1. χ− corresponds to an electron whose spin points in the −z-direction (or whose spin is “down”). which is analogous to s. Loosely speaking.

(10. Up to now.37) A general spin state can be represented as a linear combination of χ+ and χ− : i. (10.40) Furthermore.42) Note that this particular representation of spin space is in complete accordance with the discussion in Sect. c+ χ≡ .. This socalled Pauli representation allows us to visualize spin space. we have discussed spin space in rather abstract terms. A general spin operator A is represented as a 2 × 2 matrix which operates on a spinor: i.3. χ† χ ′ = (c∗ . (10. A22 c− As is easily demonstrated. For obvious reasons. χ = c+ χ+ + c− χ− .e. c∗ ).41) Likewise. In the following. χ† χ = (c∗ . the product χ† χ ′ of two different spin states is also obtained from the rules of matrix multiplication: i. A∗ 22 12 . and also facilitates calculations involving spin.e. a vector used to represent a spin state is generally known as spinor. the Hermitian conjugate of A is represented by the transposed complex conjugate of the matrix used to represent A: i.e. (10.44) .e. + − (10. Let us attempt to represent a general spin state as a complex column vector in some two-dimensional space: i.. A12 c+ Aχ ≡ .e..e. A∗ 11 21 A∗ .38) It is thus evident that electron spin space is two-dimensional. χ† ≡ (c∗ . A† ≡ A∗ . 10.e. + − (10.134 and χ† χ− = 0. + QUANTUM MECHANICS (10.. c∗ ) + − c+ c− = c∗ c+ + c∗ c− = |c+ |2 + |c− |2 ≥ 0.39) c− The corresponding dual vector is represented as a row vector: i. we shall describe a particular representation of electron spin space due to Pauli. the product χ† χ is obtained according to the ordinary rules of matrix multiplication: i.. A11 . c∗ ) + − ′ c+ ′ c− ′ ′ = c∗ c+ + c∗ c− ...43) A21 . + − (10.

49) (10.56) c+ c− .37). the σi satisfy the commutation relations [σx . 2.36) and (10. 3 corresponds to x. . (10. 1 0 . from the above expressions. (10. It is convenient to write the spin operators Si (where i = 1. −i i. Finally.45) (10.55) In this case. Furthermore. 0 0. σy ] = 2 i σz. z) as ¯ h Si = σi . −1 . y. 0 1. a general spinor takes the form χ = c+ χ+ + c− χ− = If the spinor is properly normalized then χ† χ = |c+ |2 + |c− |2 = 1. σx ] = 2 i σy .51) (10. Note that these forms automatically satisfy the orthonormality constraints (10.46) respectively. (10. (10. σz ] = 2 i σx . h . 1 1. . According to Eqs.48) (10. these matrices are generally known as the Pauli matrices. (10.1)–(10.47) 2 Here. (10. (10. and |c− |2 as the probability that an observation of Sz will yield the result h −¯ /2.34) yields It is easily demonstrated.53) (10. 135 (10.52) (10.3).Spin Angular Momentum Let us represent the spin eigenstates χ+ and χ− as χ+ ≡ and χ− ≡ 0 1 . we can interpret |c+ |2 as the probability that an observation of Sz will yield the result +¯ /2.54) σz χ± = ±χ± . Eq. [σz . [σy . that the σi are represented by the following matrices: σx ≡ σy ≡ σz ≡ 0. the σi are dimensionless 2 × 2 matrices. 0 0.50) Incidentally.

0023192. (10. we are neglecting the electron’s translational degrees of freedom. and me the electron mass.62) Here. g=2 1+ α + · · · = 2.63)] o i¯ h ∂χ = H χ.61) (10. Furthermore. The energy of a classical magnetic moment µ in a uniform magnetic ﬁeld B is H = −µ · B. a small current loop possesses a magnetic moment of magnitude µ = I A. (4. Classically. 2 me (10. Consider a small current loop consisting of an electron in uniform circular motion. In fact. Schr¨dinger’s equation can be written [see Eq. The fact that the gyromagnetic ratio is (almost) twice that expected from classical physics is only explicable using relativistic quantum mechanics. where Ω= geB . 2π (10. 2 me (10. We can write µ=− ge S.58) where g is called the gyromagnetic ratio.136 QUANTUM MECHANICS 10. the small corrections to the relativistic result g = 2 come from quantum ﬁeld theory. It is easily demonstrated that the electron’s orbital angular momentum L is related to the magnetic moment µ of the loop via e µ=− L.57) 2 me where e is the magnitude of the electron charge.60) Assuming that the above expression also holds good in quantum mechanics. we would expect g = 1. The direction of the magnetic moment is conventionally taken to be normal to the plane of the loop. the Hamiltonian of an electron in a z-directed magnetic ﬁeld of magnitude B takes the form H = Ω Sz .63) . ∂t (10. in the sense given by a standard right-hand circulation rule. for the sake of simplicity. and A the area of the loop.59) where α = e2 /(2 ǫ0 h c) ≃ 1/137 is the so-called ﬁne-structure constant. (10. The above expression suggests that there may be a similar relationship between magnetic moment and spin angular momentum.6 Spin Precession According to classical physics. where I is the current circulating around the loop.

54).64) that i¯ h where ˙ ≡ d/dt. Here. (10.56). (4. 0 0. (10.47).72) ¯ h cos α (10. 0 c+ c− .74) . (10. Hence. and |c− |2 the probability of observing the spin-down state.73) 2 with the help of Eqs.Spin Angular Momentum 137 where the spin state of the electron is characterized by the spinor χ. 1 1.70) We can most easily visualize the effect of the time dependence in the above expressions for c± by calculating the expectation values of the three Cartesian components of the electron’s spin angular momentum. and (10.65) (10.69) (10. c ) 2 + − 0. (10.67) and (10. c− (t) = sin(α/2) exp(+i Ω t/2). c± = ∓i ˙ Let c+ (0) = cos(α/2). the expectation value of Sz is Sz = which reduces to ¯ ∗ ∗ h (c . subject to the initial conditions (10. the expectation value of Sx is Sz = Sx = ¯ ∗ ∗ h (c . (10.64) χ= c− (t) where |c+ |2 + |c− |2 = 1.68). (10.69) and (10. |c+ |2 is the probability of observing the spin-up state.70). −1 c+ c− . (10. Likewise. It follows from Eqs. (10. (10.61).66) The signiﬁcance of the angle α will become apparent presently.63). Adopting the Pauli representation. we obtain c+ (t) .66). we obtain c+ (t) = cos(α/2) exp(−i Ω t/2). Solving Eq. (10. (10. c− (0) = sin(α/2). −1 Ω c± . 2 c+ c− . (10.71) Hence. 0 0. By analogy with Eq. c ) 2 + − 1. (10.67) (10. the expectation value of a general spin operator A is simply A = χ† A χ.68) c+ ˙ c− ˙ = Ω¯ h 2 1.

or Sz will always yield either +¯ /2 or −¯ /2. Sy . An electron is at rest in an oscillating magnetic ﬁeld B = B0 cos(ω t) ez . Find the Pauli representations of the normalized eigenstates of Sx and Sy for a spin-1/2 particle. Note. Exercises 1. Demonstrate that a measurement of Sz yields ¯ /2 with probability h cos2 (θ/2).77) me This behaviour is actually equivalent to that predicted by classical physics. 2 ¯ h sin α cos(Ω t). where B0 and ω are real positive constants. however.73).76) According to Eqs. What is the probability that a measurement of Sy yields −¯ /2? h 6. that a measurement of Sx . . the expectation value of the spin angular momentum vector subtends a constant angle α with the z-axis. and precesses about this axis at the frequency eB Ω≃ . and Sz for a spin-1 particle.75). Determine the constant A by normalizing χ. h 4. and (10. and with what probabilities? What is the expectation value of Sz ? Repeat the above calculations for Sx and Sy . what values will be obtained.138 which reduces to Sx = Finally. (10. and −¯ /2 with probability sin2 (θ/2). Sy . making an angle θ with the z-axis. Suppose that a spin-1/2 particle has a spin vector which lies in the x-z plane. the expectation value of Sy is Sy = ¯ h sin α sin(Ω t). 2 QUANTUM MECHANICS (10. It is the relative h h probabilities of obtaining these two results which varies as the expectation value of a given component of the spin varies. where α and β are real.76). Find the Pauli representations of Sx . 3. An electron is in the spin-state χ=A 1 − 2i 2 in the Pauli representation. Consider a spin-1/2 system represented by the normalized spinor χ= cos α sin α exp( i β) in the Pauli representation. If a measurement of Sz is made. (10. 5. 2.75) (10. (10.

**Spin Angular Momentum
**

(a) Find the Hamiltonian of the system.

139

(b) If the electron starts in the spin-up state with respect to the x-axis, determine the spinor χ(t) which represents the state of the system in the Pauli representation at all subsequent times. (c) Find the probability that a measurement of Sx yields the result −¯ /2 as a function of h time. (d) What is the minimum value of B0 required to force a complete ﬂip in Sx ?

140

QUANTUM MECHANICS

Addition of Angular Momentum

141

11 Addition of Angular Momentum

11.1 Introduction

Consider an electron in a hydrogen atom. As we have already seen, the electron’s motion through space is parameterized by the three quantum numbers n, l, and m (see Sect. 9.4). To these we must now add the two quantum numbers s and ms which parameterize the electron’s internal motion (see the previous chapter). Now, the quantum numbers l and m specify the electron’s orbital angular momentum vector, L, (as much as it can be speciﬁed) whereas the quantum numbers s and ms specify its spin angular momentum vector, S. But, if the electron possesses both orbital and spin angular momentum then what is its total angular momentum?

**11.2 General Principles
**

The three basic orbital angular momentum operators, Lx , Ly , and Lz , obey the commutation relations (8.6)–(8.8), which can be written in the convenient vector form: L × L = i ¯ L. h (11.1)

Likewise, the three basic spin angular momentum operators, Sx , Sy , and Sz , obey the commutation relations (10.1)–(10.3), which can also be written in vector form: i.e., S × S = i ¯ S. h (11.2)

Now, since the orbital angular momentum operators are associated with the electron’s motion through space, whilst the spin angular momentum operators are associated with its internal motion, and these two types of motion are completely unrelated (i.e., they correspond to different degrees of freedom—see Sect. 6.2), it is reasonable to suppose that the two sets of operators commute with one another: i.e., [Li , Sj ] = 0, where i, j = 1, 2, 3 corresponds to x, y, z. Let us now consider the electron’s total angular momentum vector J = L + S. We have J × J = (L + S) × (L + S) = i¯ L+i¯ S h h (11.4) (11.3)

= L×L+S×S+L×S+S×L=L×L+S×S = i ¯ J. h (11.5)

142

In other words, J × J = i ¯ J. h

QUANTUM MECHANICS

(11.6)

It is thus evident that the three basic total angular momentum operators, Jx , Jy , and Jz , obey analogous commutation relations to the corresponding orbital and spin angular momentum operators. It therefore follows that the total angular momentum has similar properties to the orbital and spin angular momenta. For instance, it is only possible to simultaneously measure the magnitude squared of the total angular momentum vector,

2 2 J2 = Jx + Jy + Jz2 ,

(11.7)

together with a single Cartesian component. By convention, we shall always choose to measure Jz . A simultaneous eigenstate of Jz and J2 satisﬁes Jz ψj,mj = mj ¯ ψj,mj , h J2 ψj,mj = j (j + 1) ¯ 2 ψj,mj , h (11.8) (11.9)

where the quantum number j can take positive integer, or half-integer, values, and the quantum number mj is restricted to the following range of values: − j, −j + 1, · · · , j − 1, j. Now J2 = (L + S) · (L + S) = L2 + S2 + 2 L · S, which can also be written as J2 = L2 + S2 + 2 Lz Sz + L+ S− + L− S+. (11.12) (11.11) (11.10)

We know that the operator L2 commutes with itself, with all of the Cartesian components of L (and, hence, with the raising and lowering operators L± ), and with all of the spin angular momentum operators (see Sect. 8.2). It is therefore clear that [J2 , L2] = 0. A similar argument allows us to also conclude that [J2 , S2] = 0. (11.14) (11.13)

Now, the operator Lz commutes with itself, with L2 , with all of the spin angular momentum operators, but not with the raising and lowering operators L± (see Sect. 8.2). It follows that [J2 , Lz] = 0. (11.15) Likewise, we can also show that [J2 , Sz] = 0. (11.16)

mj .j.s.s.ms .m.Addition of Angular Momentum Finally.s.s.mj = l (l + 1) ¯ 2 ψl.m. h Jz ψl.mj = mj ¯ ψl. The ﬁrst set correspond to the operators L2 . Lz .s. (11.j. the quantum numbers controlling the z-components of the various angular momentum vectors can simply be added algebraically. and Jz corresponding to the following eigenvalues: L2 ψl.ms = m ¯ ψl.mj represent a simultaneous eigenstate of L2 . J2 . It is easily seen that h Jz ψl. together with the z-component of the total angular momentum vector.ms . mj = m + ms . we have Jz = Lz + Sz . and Jz . Sz .j. h Hence. In addition. S2 .m.s. S2 .17) where [Jz . J2 . S2 . and Jz .s.m. Lz] = [Jz .ms represent a simultaneous eigenstate of L2 .ms = (Lz + Sz ) ψl.s. The second set correspond to the operators L2 .ms .j. S2 .m.mj = s (s + 1) ¯ 2 ψl.ms . h h S2 ψl. we can always measure the magnitude squared of the orbital and spin angular momentum vectors.26) (11. h (2) (2) (2) (2) (2) (2) (2) (2) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1) (11. it follows.s.s.s.m.21) (11. Sz] = 0. and Sz corresponding to the following eigenvalues: L2 ψl.m.m. h Sz ψl.ms = s (s + 1) ¯ 2 ψl. h J2 ψl. from the above discussion.j.j.ms = ms ¯ ψl.m.s.ms = l (l + 1) ¯ 2 ψl.ms = mj ¯ ψl.m.s.24) (11.20) (11.mj . 143 (11.22) (11.23) In other words.j.s.s.27) .m. h Lz ψl. In other words.s. Recalling that only commuting operators correspond to physical quantities which can be simultaneously measured (see Sect.s.s. that there are two alternative sets of physical variables associated with angular momentum which we can measure simultaneously. (2) Finally.s.mj . let ψl. 4.mj = j (j + 1) ¯ 2 ψl.ms .ms = (m + ms ) ¯ ψl. h S2 ψl. (1) Let ψl.10).18) (11.j.s. or the magnitude squared of the total angular momentum vector.19) (11.mj .s.m.25) (11.s. Lz . we can either choose to measure the z-components of the orbital and spin angular momentum vectors.j.m.

(11.1/2 + β ψl.m.m. For instance.35) (2) (2) (2) (11.m+1/2 = α ψl.m+1 χ− .1/2. and Sz can be written in the separable form ψl. we obtain h S+ χs. and the fact that ms can only take the values ±1/2. (8.43) and (8. and the internal state of the electron is speciﬁed by the spinors χ± (see Sect.m χ+ + β Yl. χs. (1) Since the eigenstates ψl.3 Angular Momentum in the Hydrogen Atom In a hydrogen atom. (1) (11. Note that the number of ψ(1) states which can appear on the right-hand side of the above expression is limited to two by the constraint that mj = m + ms [see Eq.1/2.−1/2 .m+1/2 .j.36) (11. S± . h where [see Eq.32) (11. 10.j.26) that J2 ψl.37) . (11.12)] J2 = L2 + S2 + 2 Lz Sz + L+ S− + L− S+. h (11.m+1/2 = α Yl. Yl.m (θ. Hence. (2) (1) (1) (2) (11.ms = [s (s + 1) − ms (ms − 1)]1/2 ¯ χs.±1/2 are (presumably) orthonormal.1/2. as yet. it follows from Eq. 8.m+1. the simultaneous eigenstates of L2 . the wavefunction of an electron in a simultaneous eigenstate of L2 and Lz has an angular dependence speciﬁed by the spherical harmonic Yl.ms .44). (11.31) (11.m. act on a general spinor.29) where α and β are. respectively.144 QUANTUM MECHANICS 11. and form a complete set.m+1/2 = j (j + 1) ¯ 2 ψl. whereas spin angular momentum operators act on the spinors.5).28) and (11.1/2.ms −1 . If the electron is also in an eigenstate of S2 and Sz then the quantum numbers s and ms take the values 1/2 and ±1/2. that L+ Yl.34) (11. (11.m−1 . when the spin raising and lowering operators. (11. unknown coefﬁcients.23)].1/2. h L− Yl. we can express the eigenstates ψl.1/2. Assuming that the ψ(2) eigenstates are properly normalized. Recall. ψl. h S− χs. from Eqs. Now. we can write ψl.j.33) By analogy. φ) (see Sect.m+1 .1/2.ms = [s (s + 1) − ms (ms + 1)]1/2 ¯ χs.28) Here.29).mj as linear combinations of them. Lz . according to Eqs.j. Moreover.±1/2 = Yl. χ± . it is understood that orbital angular momentum operators act on the spherical harmonic functions.m = [l (l + 1) − m (m + 1)]1/2 ¯ Yl.30) (11.ms +1 .7).m = [l (l + 1) − m (m − 1)]1/2 ¯ Yl. S2 .1/2.j. we have α2 + β2 = 1.m χ± .1/2.m .

ms = ±1/2). Eqs. (11. and [(l − m) (l + m + 1)]1/2 α = . s = 1/2. h Hence.39) (11.m+1 χ− . Once x is speciﬁed. (1) (11.41) (11. respectively.Addition of Angular Momentum 145 For the special case of spin one-half spinors (i.40) (11.38) and S± χ∓ = ¯ χ± .m+1/2 = l−m 2l+ 1 (1) ψm.43) can be solved to give x (x + 1) = l (l + 1).45) (11.m+1/2 = l+m+1 2l+ 1 1/2 1/2 (1) ψm.m χ+ h +[l (l + 1) − m (m + 1)]1/2 ¯ 2 Yl. (11.48) . −[l (l + 1) − m (m + 1)] where x = j (j + 1) − l (l + 1) − 3/4..1/2 l+m+1 − 2l+ 1 1/2 ψm+1.42) (11.43) α + (x + m + 1) β = 0.e.m χ+ . h It follows from Eqs.46) (11.42) and (11.31) and (11.33) yield (x − m) α − [l (l + 1) − m (m + 1)]1/2 β = 0.1/2 l−m + 2l+ 1 1/2 ψm+1.−1/2 .32) and (11. (1) (11. which corresponds to j = l + 1/2 or j = l − 1/2.m+1 χ− = [l (l + 1) + 3/4 − m − 1] ¯ 2 Yl. h and J2 Yl. It follows that x = l or x = −l − 1. (11.46) can be solved for α and β.44) 1/2 (11.m χ+ = [l (l + 1) + 3/4 + m] ¯ 2 Yl.39) that J2 Yl. (11.−1/2 . β x−m (11. Eqs.30) and (11.m+1 χ− h +[l (l + 1) − m (m + 1)]1/2 ¯ 2 Yl. Equations (11. We obtain (2) ψl+1/2. the above expressions reduce to S+ χ+ = S− χ− = 0.34)–(11.47) and (2) ψl−1/2.

146 m. 3 0.1/2 1 (1) ψ + 3 −1.e. ψ3/2.−1/2 2 (1) ψ . (2) The information contained in Eqs. 3 0.m+1/2 l−m + 2l+ 1 1/2 ψl−1/2.−1/2 2 (1) ψ . whereas Eq. (11.m+1/2 . −1/2 √ (l−m)/(2 l+1) √ m..±1/2 .±1/2 . and ψ1/2.l+1/2.53) (11. Lz . (11. whereas ms can m. 1/2 (2) (11. there are clearly six such states: i. m + 1/2 l − 1/2.±3/2 = ψ±1.1/2 = = l+m+1 2l+ 1 l−m 2l+ 1 1/2 1/2 (2) ψl+1/2.1.52) (11.m (1) (1) (1) take the values ±1/2. (11.1/2 (11. The eigenstates of 2 L .m+1/2 ≡ (2) (2) ψl.e. The eigenstates of L2 .−1/2 (2) ψl+1/2.55) 2 (1) ψ + 3 0. J2 .. the various different eigenstates are interrelated as follows: ψ3/2.51) 1 (1) ψ . are denoted ψj.m+1/2 .−1/2 = ψ3/2. and Jz . ψ3/2.−1/2 1 (1) ψ .±3/2 .mj .m+1/2 l+m+1 − 2l+ 1 ψl−1/2. m + 1/2 j.54) (11. 0. let us consider the l = 1 states of a hydrogen atom.1. ms l + 1/2.±1/2 .47) is obtained by reading the ﬁrst row of this table.1/2.1/2 1 (1) ψ − 3 0.47)–(11.m+1/2 .±1/2 . are denoted ψ(1) s .1/2 = ψ1/2. and Sz .1/2 2 (1) ψ − 3 −1.1: Clebsch-Gordon coefﬁcients for adding spin one-half to spin l. (1) (2) (2) (2) (2) (2) (2) (1) (2) (11.50) is obtained by reading the second column. we have neglected the common subscripts l. According to Table 11. ψ3/2. Eq.56) . S2 . and ψ−1.±3/2 .49) (11.±1/2 = ψ3/2.e.±1/2 .. S2 .1/2 = ψ1/2. The above equations can easily be inverted to give the ψ(1) eigenstates in terms of the ψ(2) eigenstates: (1) ψm. mj − (l+m+1)/(2 l+1) Table 11. As an example.−1/2 (11. For instance. 3 1. Since l = 1 and s = 1/2 can be combined together to form either j = 3/2 or j = 1/2 (see earlier).50) is neatly summarized in Table 11. ψ0. 3 1. 1/2 for the sake of clarity: i.±1/2 . 1. ψl+1/2. The coefﬁcients in this type of table are generally known as Clebsch-Gordon coefﬁcients. there are also six such (2) (2) (2) states: i. Since m can take the values −1.−1/2 = and ψ±1. 1/2 √ (l+m+1)/(2 l+1) √ (l−m)/(2 l+1) QUANTUM MECHANICS m + 1. ψ1. etc.50) (1) ψm+1. Here.

the state represented by ψ0. 11.52) that a subsequent measurement of Lz and Sz will yield m = 0. −1/2 3/2.58) (11. mj = 1/2 with probability 2/3.. a measurement of the total angular momentum will yield j = 3/2. Let S1 and S2 be the spin angular momentum . and m = 1.59) (11. 3/2 j. 3 1/2. ms 3/2. and also that the different rows and different columns are mutually orthogonal.59) is neatly summed up in Table 11. ms = 1/2 with probability 2/3. 3 1/2. −1/2 1. the electron (2) is thrown into the corresponding eigenstate. and obtain the result j = 3/2.−1/2 ψ1.−1/2 1 (2) ψ + 3 3/2. 1/2 m. It thus follows from Eq.−1/2 = ψ0. ms = −1/2 with probability 1/3.e.57) (11. and j = 1/2. Of course.1/2 . −3/2 3/2.Addition of Angular Momentum −1. mj = 1/2 with probability 1/3.−1/2 2 (2) ψ .2: Clebsch-Gordon coefﬁcients for adding spin one-half to spin one. 3 1/2. if we know that an electron in a hydrogen atom is in an l = 1 state characterized by (1) m = 0 and ms = 1/2 [i.1/2 ] then. mj = 1/2.1/2 1 (2) ψ .1/2 1 (2) ψ . −1/2 −1.2.1/2 2 (2) ψ + 3 3/2. 1/2 1/2. As a result of the measurement. ψ3/2.−1/2 2 (2) ψ . The information contained in Eqs.1/2 = ψ0.51)–(11. (11. 1/2 3/2.1/2 = (1) (1) (1) (1) (11. Suppose that we make such a measurement. −1/2 0. (11. (11. 3 1/2. mj 1 147 √ 1/3 √ 2/3 − √ 2/3 √ 1/3 √ 2/3 √ 1/3 − √ 1/3 √ 2/3 1 Table 11.60) Thus. 1 (2) ψ − 3 3/2. Suppose that the system does not possess any orbital angular momentum.−1/2 = ψ−1. according to Eq. −1/2 1/2. 1/2 0.4 Two Spin One-Half Particles Consider a system consisting of two spin one-half particles. 1/2 1.58).1/2 2 (2) ψ − 3 3/2. Note that each row and column of this table has unit norm. Only non-zero coefﬁcients are shown. this is because the ψ(1) and ψ(2) eigenstates are orthonormal.

s. or S1 . or a state with overall spin s = 0. then 2 S1 χ(2). To be more exact. h (2) s1 Of course. ms = ms1 + ms2 . and one possible s = 0 state (corresponding to ms = 0).69) (11. we either obtain a state with overall spin s = 1. h (2) s1 2 h (2) S2 χ(2).67) (11.−1.63) (11.s. s1 Sz χ(2).s.ms2 = s1 (s1 + 1) ¯ 2 χs1 .ms2 is a simultaneous eigenstate of S1 . whereas the s = 0 state is known as the singlet state.s2 .s2 . when two spin one-half particles are combined.72) .ms2 . we saw. S1z .ms s1 1 .s2 . if the spinor (1) 2 2 χs1 . there are three possible s = 1 states (corresponding to ms = −1.s. 2 2 Likewise. s.s.ms s1 2 S2 χ(1).71) Now.65) (11.ms = s (s + 1) ¯ 2 χs1 . since both particles have spin one-half.3.s2 . S2 .ms2 . S2 .s.−1 = χ−1/2. 1).ms2 . By analogy with the previous analysis. s1 (11. Furthermore. . In other words.ms . S2 .1 (with l = 1/2). ms2 .ms1 .ms = s2 (s2 + 1) ¯ 2 χs1 . S2z . and Sz . The three s = 1 states are generally known as the triplet states.2 . respectively.ms is a simultaneous eigenstate of S1 .ms .ms2 . Let the quantum numbers associated with measurements of S1 . 1 1 1 . that when spin l is added to spin one-half then the possible values of the total angular momentum quantum number are j = l ± 1/2.62) (11. and Sz .ms s1 Sz χ(1).61) be the total spin angular momentum operator.s2 .64) (11.ms h (1) = ms ¯ χs1 .ms .ms h (1) 1 1 1 .s.s2 .ms h (1) = ms2 ¯ χs1 . and S2z .68) (11. respectively. 0.s2 .ms2 . s2z = ±1/2. and are listed in Table 11.s2 .ms = ms ¯ χs1 . in the previous section.ms2 . and let S = S1 + S2 QUANTUM MECHANICS (11.s2 . s2 . then 2 S1 χ(1). if the spinor χs1 .ms s1 S2z χ(1).s. S2 . 2 2 S1z . by analogy with previous analysis. It follows from this table that the three triplet states are: χ1. . s1 = s2 = 1/2. S2 .ms s1 (2) 1 .s2 .s2 . and s1z .s2 .66) 1 1 S1z χ(1). and Sz . The Clebsch-Gordon coefﬁcients for adding spin one-half to spin one-half can easily be inferred from Table 11.s2 . and ms . (2) (1) (11. and Sz be s1 .ms2 .s2 . ms1 .ms = s1 (s1 + 1) ¯ 2 χs1 .s2 .148 operators of the ﬁrst and second particles.ms2 .s2 . (11. In other words.ms . S2z .70) h (2) S2 χ(2). we 2 2 2 2 conclude that it is possible to simultaneously measure either S1 . when spin one-half is added to spin one-half then the possible values of the total spin quantum number are s = 1/2 ± 1/2. S2 .ms h (1) = ms1 ¯ χs1 . By analogy. S2 .ms2 . (11.s2 . S1z .s2 .s2 .ms h (1) = s2 (s2 + 1) ¯ 2 χs1 .s. S2 .s2 .

where χ(2) s is shorthand for χs1 . (g) What is the probability density for ﬁnding the electron at r. and with what probabilities? (b) Same for Lz .Addition of Angular Momentum −1/2. (2) (2) (1) (11. (e) Same for J2 . (d) Same for Sz . ms 1 √ 1/ 2 √ 1/ 2 √ 1/ 2 √ −1/ 2 1 149 Table 11. φ? (h) What is the probability density for ﬁnding the electron in the spin up state (with respect to the z-axis) at radius r? 2. etc.−1/2 . Calculate the potential energy for the neutron-proton system: . 1/2 ms1 . r2 where σ1 denotes the vector of the Pauli matrices of the neutron. In a low energy neutron-proton system (with zero orbital angular momentum) the potential energy is given by V(r) = V1 (r) + V2 (r) 3 (σ1 · r) (σ2 · r) − σ1 · σ2 + V3 (r) σ1 · σ2 . 1 s.74) (11. 2 Exercises 1.0 = √ χ−1/2. 2 χ1.3: Clebsch-Gordon coefﬁcients for adding spin one-half to spin one-half. 0 0. 1 (2) (1) (1) χ1. 0 1. θ. (c) Same for S2 .s.1/2 .0 = √ χ−1/2. ms2 1.73) (11.1 χ− . the singlet state is written: s. An electron in a hydrogen atom occupies the combined spin and position state R2.75) (a) What values would a measurement of L2 yield. Likewise. −1/2 1/2. Only nonzero coefﬁcients are shown.1/2 + χ1/2. (f) Same for Jz .1/2 − χ1/2.m 1 (1) (1) (2) χ0.1 = χ1/2. and σ2 denotes the vector of the Pauli matrices of the proton.−1/2 .0 χ+ + 2/3 Y1. −1 1.1 1/3 Y1. −1/2 −1/2. 1/2 1/2.s2 .ms .

what is the probability that a measurement of the x-component of the spin h of the other electron yields Sx = −¯ /2? h Finally. and electron 2 is in a spin state described by cos α2 χ+ + sin α2 e i β2 χ− . what is the probability that a measurement of the z-component of the spin of h the other electron yields Sz = ¯ /2? h (b) If a measurement of the spin of one of the electrons shows that it is in the state with Sy = ¯ /2.150 (a) In the spin singlet state. QUANTUM MECHANICS (a) If a measurement of the spin of one of the electrons shows that it is in the state with Sz = ¯ /2. 3. (b) In the spin triplet state. what is the probability that the two-electron spin state is a triplet state? . if electron 1 is in a spin state described by cos α1 χ+ + sin α1 e i β1 χ− . Consider two electrons in a spin singlet state.

3) i −∞ In three dimensions (see Cha. be regarded as being small compared to H0 . It is also assumed that the unperturbed Hamiltonian. Incidentally. H0 . by performing some sort of perturbation expansion about the eigenvalues and eigenstates of the original Hamiltonian. in which the modiﬁcation to the Hamiltonian.5) The generalization to the case where ψ is a product of a regular wavefunction and a spinor is fairly obvious. H ψi = Ei ψi . (12. (12.Time-Independent Perturbation Theory 151 12 Time-Independent Perturbation Theory 12. corresponding to the eigenvalues Ei : i. i −∞ −∞ −∞ (12. i (12.e. is time-independent. The Hamiltonian of a quantum mechanical system is written H = H0 + H1 . it is helpful to introduce some improved notation.. 12. (12. We can represent all of the above possibilities by writing ψi |ψj ≡ i|j = δij .2 Improved Notation Before commencing our investigation.1) Here.9). 7). in this chapter. 4. but is sufﬁciently complicated that when we add it to H0 we can no longer ﬁnd the exact energy eigenvalues and eigenstates. the above expression generalizes to ∞ ∞ ∞ ψ∗ ψj dx dy dz = δij . H0 is a simple Hamiltonian whose eigenvalues and eigenstates are known exactly.2) Now. H1 introduces some interesting additional physics into the problem. Can we ﬁnd approximate eigenvalues and eigenstates of the modiﬁed Hamiltonian. In one dimension. However. (12. H. we shall only discuss so-called time-independent perturbation theory. 10) then we have ψ† ψj = δij . if the ψi are spinors (see Cha.1 Introduction Consider the following very commonly occurring problem. in some sense (which we shall specify more precisely later on). H0 ? Let us investigate. has no explicit dependence on time. H1 can. this implies that ∞ ψ∗ ψj dx = δij .6) . H0 + H1 .4) Finally. we expect the ψi to be orthonormal (see Sect. H1 . Let the ψi be a complete set of eigenstates of the Hamiltonian.

(12. the term in angle brackets represents the integrals in Eqs. i (12.11) ψi |ψa ψi ≡ i|a ∗ i|a ψi .9) A = i. (12.9) Finally. 4. i (12.14) Here. (12. i We can represent all of the above possibilities by writing ci = ψi |ψa ≡ i|a . and the spinor product (12.e.. The expansion (12. ψi . Expanding a general wavefunction.15) whereas in three dimensions we get ∞ ∞ ∞ Aij = −∞ −∞ −∞ ψ∗ A ψj dx dy dz.7) thus becomes ψa = i (12.7) i In one dimension. In one dimension. i (12. if A is a general operator. 4. the Aij are unsurprisingly known as the matrix elements of A.10) (12. respectively. the expansion coefﬁcients take the form (see Sect. in terms of the energy eigenstates. three-dimensional wavefunctions.4) in oneand three-dimensional regular space. we obtain ψa = ci ψi .8) whereas in three dimensions we get ∞ ∞ ∞ ci = −∞ −∞ −∞ ψ∗ ψa dx dy dz. it follows that = a|i . i (12.3) and (12.13) Finally. etc.5) in spinspace. i (12.16) .j c∗ cj Aij .7). spinors.9) ∞ ci = −∞ ψ∗ ψa dx. ψa . if ψ is a spinor then we have ci = ψ† ψa . and the wavefunction ψa is expanded in the manner shown in Eq. The advantage of our new notation is its great generality: i. it can deal with one-dimensional wavefunctions.152 QUANTUM MECHANICS Here. then the expectation value of A is written (see Sect.12) Incidentally. the matrix elements take the form ∞ Aij = −∞ ψ∗ A ψj dx. i (12. (12.

In such a system. (12. (12. there are only two independent eigenstates of the unperturbed Hamiltonian: i. solve this problem exactly.19) that |i i| ≡ 1.Time-Independent Perturbation Theory Finally.24) that i|H0 + H1 |E = E i|E . (12. it is clear from Eq.14) thus becomes A ≡ a|A|a = Incidentally. Since the eigenstates of H0 form a complete set.17) (12.18) a|i i|A|j j|a . and to form a complete set.19) = j|A† |i . (12.e.22) (12.58)] i|A|j Finally.. (12.26) (12. H0 ψ2 = E2 ψ2 . we can write [see Eq.20) (12. i We can represent all of the above possibilities by writing Aij = ψi |A|ψj ≡ i|A|j . if ψ is a spinor then we have Aij = ψ† A ψj . The expansion (12. We also expect the states to be orthonormal. are known.23) It is assumed that these states. 12. i. ∗ 153 (12.21) i where the ψi are a complete set of eigenstates.25) .3 Two-State System Consider the simplest possible non-trivial quantum mechanical system.j (12. in fact. (12. and 1 is the identity operator. and their associated eigenvalues.12)] ψE = 1|E ψ1 + 2|E ψ2 . It follows from (12. (4.24) We can. H0 ψ1 = E1 ψ1 . Let us now try to solve the modiﬁed energy eigenvalue problem (H0 + H1 ) ψE = E ψE . it follows that [see Eq.

22). (12.34) 1 1 (E1 + E2 ) ± (E1 − E2 )(1 + 2 ǫ2 + · · ·).28) Here.31) 1|E 2|E = 0 0 . 2|H1 |2 . E1 − E2 (12. |E1 − E2 | (12. and the orthonormality condition i|j = δij . (12. E1 − E2 e12 = ψ2 − ψ1 + · · · .25).37) |e12 |2 = E2 − + ···.27) yield two coupled equations which can be written in matrix form: E1 − E + e11 e12 e∗ E2 − E + e22 12 where e11 = e22 = e12 = 1|H1 |1 . It is easily demonstrated that the modiﬁed eigenstates take the form ′ ψ1 = ψ1 + ′ ψ2 e∗ 12 ψ2 + · · · . use has been made of the fact that H1 is an Hermitian operator. (12. (12.26).32) The solution of Eq. (12. Consider the special (but not uncommon) case of a perturbing Hamiltonian whose diagonal matrix elements are zero. Equations (12. so that e11 = e22 = 0.23). (12. E1 − E2 (12.35) 2 2 The above expression yields the modiﬁcation of the energy eigenvalues due to the perturbing Hamiltonian: E≃ ′ E1 = E1 + ′ E2 |e12 |2 + ···.28) (obtained by setting the determinant of the matrix to zero) is E= (E1 + E2 ) ± (E1 − E2 )2 + 4 |e12|2 . 1|H1 |2 = 2|H1 |1 ∗ . E1 − E2 Note that H1 causes the upper eigenvalue to rise. and the lower to fall.38) (12.30) (12. (12. (12.33) 2 Let us expand in the supposedly small parameter ǫ= We obtain |e12 | . (12.29) (12.36) (12.39) . (12.154 QUANTUM MECHANICS where i = 1 or 2.

when we say that H1 needs to be small compared to H0 . Consider a system in which the energy eigenstates of the unperturbed Hamiltonian.43) If follows that m|H0 + H1 |E = E m|E . (12.46) where emk = m|H1 |k . (12. so that m|n = δnm . We assume that emk ∼ O(ǫ) Em − Ek (12. We can combine the above equations to give (Em − E + emm ) m|E + k=m emk k|E = 0. we can express ψE as a linear superposition of the unperturbed energy eigenstates: ψE = k k|E ψk . (12.Time-Independent Perturbation Theory 155 Thus. H0 .40) In other words.48) (12. The eigenstates are assumed to be orthonormal. what we are really saying is that the above inequality must be satisﬁed.45) where k runs from 1 to N. are denoted H0 ψn = En ψn .4 Non-Degenerate Perturbation Theory Let us now generalize our perturbation analysis to deal with systems possessing more than two energy eigenstates. Now our expansion procedure is only valid when ǫ ≪ 1. plus a slight admixture of the other. 12. (12. (12. the modiﬁed energy eigenstates consist of one of the unperturbed eigenstates.47) . Let us now develop our perturbation expansion. (12. Now. This suggests that the condition for the validity of the perturbation method as a whole is |e12 | ≪ |E1 − E2 |. Let us now try to solve the energy eigenvalue problem for the perturbed Hamiltonian: (H0 + H1 ) ψE = E ψE .42) and to form a complete set.41) where n runs from 1 to N.44) where m can take any value from 1 to N. (12.

12. The Hamiltonian of the system can be split into two parts. We ﬁnd that (Em − En ) m|E + emn ≃ 0. Let us search for a modiﬁed version of the nth unperturbed energy eigenstate for which E = En + O(ǫ).50) and n|E m|E = 1.46). e2 p2 − . = O(ǫ) (12. H0 = 2 me 4πǫ0 r (12. (12.54) Em − En Substituting the above expression into Eq.49) for all m. neglecting terms which are O(ǫ2 ) according to our expansion scheme.58) .56) and a wavefunction ′ ψn = ψn + k=n ekn ψk + O(ǫ2 ). Namely. (12. and neglecting O(ǫ3 ) terms. En − Ek (12. giving m|E ≃ − (12. the modiﬁed nth energy eigenstate possesses an eigenvalue ′ En = En + enn + k=n |enk |2 + O(ǫ3 ). (12.55) (En − E + enn ) − Ek − En k=n Thus. Suppose that we write out Eq.5 Quadratic Stark Effect Suppose that a hydrogen atom is subject to a uniform external electric ﬁeld. En − Ek (12. where ǫ ≪ 1 is our expansion parameter. we obtain |enk |2 ≃ 0.57) Incidentally. (12. (12.51) (12. of magnitude |E|.46) for m = n.53) emn . We also assume that emm ∼ O(ǫ) Em (12. evaluated for m = n.156 QUANTUM MECHANICS for all m = k. directed along the z-axis.52) for m = n. the unperturbed Hamiltonian. it is easily demonstrated that the modiﬁed eigenstates remain orthonormal to O(ǫ2 ).

63) (12. h Hence. by deﬁnition. Let us try to ﬁnd a set of rules which determine when these matrix elements are non-zero. Thus. l ′ . 157 (12. m ′ = n. These rules are usually referred to as the selection rules for the problem in hand. z]|n ′ . z] (12.17)] [Lz .61) it follows that [see Eqs. m|[Lz. m ′ |2 . l. n.l ′ . (8.m ′ =n. l ′ . m|z|n ′ .4)] Lz = x py − y px . (7. 9).59) Note that the electron spin is irrelevant to this problem (since the spin operators all commute with H1 ). m|Lz z − z Lz |n ′ . and let their corresponding energy eigenvalues be the Enlm . We have 2 2 [L2. n. This follows because the matrix elements n.60) This energy-shift is known as the Stark effect. The sum on the right-hand side of the above equation seems very complicated. l. since [see Eq. l. However. it is clear. the energy eigenstates of the unperturbed Hamiltonian are characterized by three quantum numbers—the radial quantum number n. l.62) = ¯ (m − m ′ ) n. l ′. the change in energy of the eigenstate characterized by the quantum numbers n. l. (12.64) . so we can ignore the spin degrees of freedom of the system. According to the analysis in the previous section. l. it turns out that most of the terms in this sum are zero. m|z|n ′ . Hence. l ′ . m +e2 |E|2 | n. m|z|n ′ . and the two angular quantum numbers l and m (see Cha. Let us now determine the selection rule for l. z] = [Lx . l.Time-Independent Perturbation Theory and the perturbing Hamiltonian H1 = e |E| z. that one of the selection rules is that the matrix element n. z] = 0. m ′ are zero for virtually all choices of the two sets of quantum number. from the above equation. m ′ is zero unless m ′ = m. l ′ . m ′ = 0. Now. l. l. l. l ′ . m and n ′ . l ′. h since ψnlm is. m|z|n ′ . m ′ (12. m ′.15)–(7. z] + [Ly . m|z|n. m in the presence of a small electric ﬁeld is given by ∆Enlm = e |E| n.l. Enlm − En ′ l ′ m ′ n ′ . Let us denote these states as the ψnlm . an eigenstate of Lz corresponding to the eigenvalue m ¯ .m (12.

4). z]] = 2 ¯ 2 (L2 z + z L2 ).70) implies that n.67) Since. the above expression expands to give L4 z − 2 L2 z L2 + z L4 − 2 ¯ 2 (L2 z + z L2) = 0.158 QUANTUM MECHANICS = Lx [Lx .70) (12. ψnlm is an eigenstate of L2 corresponding to the eigenvalue l (l+1) ¯ 2 . z]] = 2 i ¯ L2 . Hence. Ly x − Lx y + i ¯ z h h = 2 i ¯ Ly [L2. h However. z] Ly = i ¯ (−Lx y − y Lx + Ly x + x Ly ) h = 2 i ¯ (Ly x − Lx y + i ¯ z) h h = 2 i ¯ (Ly x − y Lx ) = 2 i ¯ (x Ly − Lx y).66) +2 (L2 z − z L2) = −¯ 2 4 (Lx x + Ly y + Lz z) Lz − 2 (L2 z + z L2) . h this expression yields l2 (l + 1)2 − 2 l (l + 1) l ′ (l ′ + 1) + l ′2 (l ′ + 1)2 −2 l (l + 1) − 2 l ′ (l ′ + 1)} n. h which reduces to 2 2 [L2. m|z|n ′. [L2. (12. z]] = −¯ 2 4 (Lx x + Ly y + Lz z) Lz − 4 (Lx + Ly + Lz2) z h (12. m|L4 z − 2 L2 z L2 + z L4 − 2 ¯ 2 (L2 z + z L2)|n ′ .71) (12. [L2. it is clear from Eqs. y] + i ¯ [L2. l.4) that Lx x + Ly y + Lz z = 0.73) = 0. (8. l. m = 0. [L2. m which reduces to (l + l ′ + 2) (l + l ′ ) (l − l ′ + 1) (l − l ′ − 1) n. l ′ .68) (12.69) (12.10). Thus. (12. h Equation (12. (7. and (8. l. by deﬁnition. z] Lx + Ly [Ly . z] + [Ly . [L2 . l ′ .65) where use has been made of Eqs. x] − Lx [L2 . we obtain [L2. m|z|n ′.17). l ′ . z] + [Lx . z] h h = −4 ¯ 2 Ly (y Lz − Ly z) + 4 ¯ 2 Lx (Lx z − x Lz ) h h −2 ¯ 2 (L2 z − z L2).2)–(8.72) . h h (12. h (12.2)–(8. m = 0. (8.15)–(7. h Finally.

Hence. follows.75) (12.74) to Eq.e. we conclude that Eqs. 0. all of the terms in Eq. m = 0 and Enlm = En ′ l ′ m . l. For this reason.75) and (12. l ′ . m which couples two degenerate unperturbed energy eigenstates: i. (12. 0 |2 . 1. (12. l. according to the selection rules. In conclusion. m|z|n ′ . l. there is one fairly obvious problem with Eq.77) Unfortunately. (12. Now. l ′ . 9. however.Time-Independent Perturbation Theory 159 According to the above formula. Only those terms which vary quadratically with the ﬁeld-strength survive. l ′ . if n. Application of the selection rules (12. Enlm − En ′ l ′ m ′ . m |2 . m|z|n ′ . m|z|n ′ . according to non-degenerate perturbation theory. Hence. l. l ′ .e. m |2 . that the matrix element n. since l and l ′ can never be negative. m is zero unless l ′ = l ± 1.60) which vary linearly with the electric ﬁeld-strength vanish. our perturbation method breaks down completely in this situation. = 2e En ′ l ′ m − Enlm n ′ .64) and (12.60) yields ∆Enlm = e2 |E|2 | n.76) 2 Hence. n = 1) of a hydrogen atom is given by α = 2 e2 n>1 | 1. can never be zero. it predicts an inﬁnite energy-shift if there exists some non-zero matrix element n. m vanishes unless l = l ′ = 0 or l ′ = l ± 1.74) Note that. Namely.l ′ =l±1 n (12. the unperturbed eigenstates of a hydrogen atom have energies which only depend on the radial quantum number n (see Cha. that an l = 0 wavefunction is spherically symmetric. m|z|n ′ . the matrix element n. Now. (12. the electric polarizability of an atom is deﬁned in terms of the energy-shift of the atomic state as follows: 1 ∆E = − α |E|2 . the factor l + l ′ + 2. m is zero when l = l ′ = 0. l ′ . Clearly. 0|z|n.l ′ =l±1 2 (12. It follows that we can only apply the above results to the n = 1 eigenstate (since for n > 1 there will be coupling to degenerate eigenstates with the same value of n but different values of l).78) . m|z|n ′ .] Recall. the type of perturbation theory employed here is known as non-degenerate perturbation theory.75). we can write αnlm | n. 9).77) are only applicable to cases where the coupled eigenstates are non-degenerate. l. l. It. therefore. l.. m|z|n ′ . from symmetry. l ′. in the above equation. m|z|n ′ . Thus. [Of course. this type of energy-shift of an atomic state in the presence of a small electric ﬁeld is known as the quadratic Stark effect. the selection rule for l is that the matrix element n. the polarizability of the groundstate (i. (12. l ′ . from Cha. En00 − E100 (12..

l ′ . 0. 0 n. 0 |2 = n>1 n>1 e2 . (12.4) En00 = − where a0 = is the Bohr radius.81) 16 4πǫ0 a0 3 | 1. 1. 0. we conclude that 16 3 3 4πǫ0 a0 ≃ 5. 0|r2|1.72) that 2 1. 0.160 QUANTUM MECHANICS Here. We can write H0 ψnlm = En ψnlm . 3 (12. 9. [see Eq.83) where we have made use of the selection rules. 0 = 1 1. 8π ǫ0 a0 n2 4πǫ0 ¯ 2 h me e2 e2 3 . 0.87) .84) Hence. m ′ |z|1.l ′ . l ′ . 0.79) (12. 0|z2|1. 0|r2|1. 1. 0 1.85) 3 The exact result (which can be obtained by solving Schr¨dinger’s equation in parabolic o coordinates) is 9 3 3 α = 4πǫ0 a0 = 4. 0 . 1.e. (12. 0. 0. 4 8π ǫ0 a0 (12.5 4πǫ0 a0 . rather naively. 0.m ′ form a complete set. n>1 (12. (9. l ′ . 0|z|n. it follows from Eq. 0. 1. m ′ n ′. n > 1) state of the hydrogen atom using standard non-degenerate perturbation theory.86) 2 α< 12. investigate the Stark effect in an excited (i.21)] | 1. (12. (12. we have made use of the fact that En10 = En00 .80) (12. Finally.. The sum in the above expression can be evaluated approximately by noting that (see Sect. the fact that the ψn ′ . 0 n ′ . 0 = 3 a0 . 0|z|n. and the fact the the ground-state of hydrogen is spherically symmetric.m ′ = = 1. 0 |2. 0|z|1. 0. (12. 0.3 4πǫ0 a0 .6 Degenerate Perturbation Theory Let us. 0|z|n. 0. we can write En00 − E100 ≥ E200 − E100 = which implies that α< However.82) 1. 0|z|n ′. Hence.

89) would reduce to zero. non-degenerate perturbation theory yields the following expressions for the perturbed energy levels and eigenstates [see Eqs.89) where en ′ l ′ nl = n ′ . there is no n = 1 state with l ′ = 1. k.e. the selection rule l ′ = l ± 1. Suppose that there are Nn coupled eigenstates belonging to the eigenvalue En . The ψnlm are chosen in such a manner that they are also eigenstates of the perturbing Hamiltonian. m ψnkm . H1 ψnlm = λnl ψnlm . (12. enl ′ nl .l ′ =l±1 n (12.Time-Independent Perturbation Theory 161 since the energy eigenstates of the unperturbed Hamiltonian only depend on the quantum number n. m|H1 |n.64) and (12. the above equation is not satisﬁed. all of the singular terms in Eqs. m|H1 |n. l. m|n. corresponding to the eigenvalue En . Hence.l ′ =l±1 n (12. if n > 1 then there are multiple coupled states corresponding to the eigenvalue En . because there exist non-zero matrix elements. only allow l ′ to take the single value 1.92) Note that these new states are also degenerate energy eigenstates of the unperturbed (1) Hamiltonian. l ′. Fortunately. (12.74).91) is satisﬁed.56) and (12.Nn (1) n. m = λnl δll ′ .e.57)]: ′ Enl = En + enlnl + |en ′ l ′ nl |2 . Unfortunately. Of course. In this case. l(1) . En − En ′ ′ .88) and (12. l. we can always redeﬁne the unperturbed eigenstates corresponding to the eigenvalue En in such a manner that Eq. in this case.90) Unfortunately. Unfortunately. H0 . m . (12.88) and ′ ψnlm = ψnlm + en ′ l ′ nl ψn ′ l ′ m . H1 : i.e. there exist non-zero matrix elements which couple states with the same value of n. This does not occur if n = 1 because. but different values of l. Thus. (12. Making use of the selection rules (12. if (12. En − En ′ ′ . there is only one coupled state corresponding to the eigenvalue E1 .91) n.93) . but different values of l.. and the fact that l = 0 (since 0 ≤ l < n)... (1) (1) (12. l ′ . they are simultaneous eigenstates of H0 and H1 . Note that our problem would disappear if the matrix elements of the perturbed Hamiltonian corresponding to the same value of n. These particular matrix elements give rise to singular factors 1/(En − En ) in the summations. Let us deﬁne Nn new states which are linear combinations of our Nn original degenerate eigenstates: ψnlm = k=1. which couple degenerate eigenstates: i. were all zero: i. (12. if n > 1 then the summations in the above expressions are not well-deﬁned.

e. k. degenerate. (12. l. (12. m n. l ′′ . k. m = λnl n.Nn where 1 denotes the identity operator in the sub-space of all coupled unperturbed eigenstates corresponding to the eigenvalue En . l ′′. with Nn corresponding eigenvectors xnl .98). (12. l ′(1) . l ′.100) for k = 1 to Nn . . (12. (12.. then we can employ Eqs. m .98) |en ′ l ′ nl |2 . It follows that n. E − En ′ n ′ =n.162 The ψnlm are also chosen so as to be orthonormal: i. (xnl )k = n. m|H1|n.e.88) and (12. n.99) Provided that the determinant of U is non-zero. l(1) .101) and ψnlm = ψnlm + (1) ′ (1) en ′ l ′ nl ψn ′ l ′ m . (1) QUANTUM MECHANICS (12. since the summations are over n ′ = n: i. m|H1 |n. l ′(1) . m|n. if we use the new eigenstates. l(1) . Using this completeness relation. (12. since all of the singular terms vanish. m|n.102) There are no singular terms in these expressions.e. m n.93) can be transformed into a straightforward matrix equation: n.l ′ =l±1 n (12. l(1) . m|H1 |n. m . m = λnl δll ′ . l. Now [see Eq. Eqs. m = δll ′ . m| ≡ 1.95) Thus. m|n. m . Note that the ﬁrst-order energy shifts are equivalent to the eigenvalues of the matrix equation (12. the eigenvalue equation (12. Eq. they speciﬁcally exclude the problematic.89) directly. l ′′ =1. l(1) . l ′ . In our new scheme.96) l=1. (12.Nn (12. m|n.98) can always be solved to give Nn eigenvalues λnl (for l = 1 to Nn ).88) and (12. where the elements of the Nn × Nn Hermitian matrix U are Ujk = n.. The normalized eigenvectors specify the weights of the new eigenstates in terms of the original eigenstates: i..21)] |n. E − En ′ n ′ =n. instead of the old ones. j. The only remaining difﬁculty is to determine the new eigenstates in terms of the original ones. unperturbed energy eigenstates corresponding to the eigenvalue En .97) This can be written more transparently as U x = λ x.94) (12. l(1) .l ′ =l±1 n (12.89) yield ′ Enl = En + λnl + (12.

1). we have to solve the matrix eigenvalue equation U x = λ x. 0. Hence. we have again made use of the selection rules. that 2.e. 0. 0. 1. E200 = −e2 /(32π ǫ0 a0 ). Here. (12. 0 = 3 a0 . As before. and l ′ = l ± 1). In order to apply perturbation theory to the ψ200 and ψ210 states. The corresponding normalized eigenvectors are √ 1/ 2 x1 = √ . 0|z|2.103) According to the previously determined selection rules (i.104) where U is the matrix of the matrix elements of H1 between these states. Thus. 0|z|2. that the eigenvalues of U are λ1 = 3 e a0 |E| and λ2 = −3 e a0 |E|.105) where the rows and columns correspond to ψ200 and ψ210 . There are four such states: an l = 0 state. m ′ = m. 0 = 2. It is easily demonstrated. respectively. and three l = 1 states (with m = −1. from the exact forms of the 2S and 2P wavefunctions. 2 ψ200 − ψ210 √ . since these are not coupled to any other n = 2 states by the perturbing Hamiltonian. usually referred to as 2P. non-degenerate perturbation theory works ﬁne for the ψ211 and ψ21−1 states. which tell us that the matrix element of z between two hydrogen atom states is zero unless the states possess l quantum numbers which differ by unity. 1. usually referred to as 2S. by inspection.110) It follows that the simultaneous eigenstates of H0 and H1 take the form . x2 = (12.7 Linear Stark Effect Returning to the Stark effect. On the other hand. (12. (12.106) It can be seen. let us examine the effect of an external electric ﬁeld on the energy levels of the n = 2 states of a hydrogen atom.. 0|z|2.107) 1/ 2 √ 1/ 2 √ . 0. 0|z|2.109) (12.Time-Independent Perturbation Theory 163 12. non-degenerate perturbation theory breaks down when applied to these two states. 0 2. = 2 (12.108) −1/ 2 ψ1 = ψ2 ψ200 + ψ210 √ . U = e |E| 0 2. (12. the perturbing Hamiltonian is H1 = e |E| z. 1. All of these states possess the same unperturbed energy. 0. this Hamiltonian couples ψ200 and ψ210 . 1. 0 0 . (12.

58). the unperturbed Hamiltonian is given by Eq. by an amount 3 e a0 |E|. 12. 12. we recognize the ﬁrst term on the right-hand side of this equation as the standard non-relativistic expression for the kinetic energy. (12. m . Note.117) . The ﬁrst-order energy shifts induced by an external electric ﬁeld are given by ∆E1 = +3 e a0 |E|.113) In the non-relativistic limit p ≪ m c. l.164 QUANTUM MECHANICS In the absence of an external electric ﬁeld. These states are orthogonal linear combinations of the original ψ200 and ψ210 states. so this effect—which is known as the linear Stark effect—is much larger than the quadratic effect described in Sect.4). the energies of states 1 and 2 are shifted upwards and downwards. according to standard ﬁrst-order perturbation theory (see Sect. and m is given by 1 n.111) (12. Of course. l. p2 p4 T≃ − .115) 2 m 8 m3 c2 Of course. (12. that the energies of the ψ211 and ψ21−1 states are not affected by the electric ﬁeld to ﬁrst-order. m ∆Enlm = n. and the perturbing Hamiltonian takes the form H1 = − p4 . 12.112) Thus. l. (12. ∆E2 = −3 e a0 |E|. to second-order the energies of these states are shifted by an amount which depends on the square of the electric ﬁeld-strength (see Sect. 3 8 me c2 (12. the difference between the total energy and the rest mass energy) of a particle of rest mass m and momentum p is T= p2 c2 + m2 c4 − m c2.116) Now. l.114) +O 2m 4 mc mc Hence. m = − 3 c2 8 me = − 1 n. the kinetic energy (i. (12. we can expand the square-root in the above expression to give p2 1 p 2 p 4 T= 1− . m|p2 p2 |n.8 Fine Structure of Hydrogen According to special relativity. m|p4|n. 3 8 me c2 (12. l. l. Let us consider the effect of this type of correction on the energy levels of a hydrogen atom.5).e. respectively. E200 . 12. So.5. also. the lowestorder relativistic correction to the energy of a hydrogen atom state characterized by the standard quantum numbers n. Note that the energy shifts are linear in the electric ﬁeld-strength. The second term is the lowest-order relativistic correction to this energy. l. m|H1|n. both of these states possess the same energy. (12. in the presence of an electric ﬁeld..

58). (9. m|(En − V)2 |n.57). m 2 me c2 1 E 2 − 2 En n. Note that the above derivation implicitly assumes that p4 is an Hermitian operator. using the ψnlm eigenstates. However.124) B=− 2 . It turns out that this is not the case for l = 0 states. It follows that there is no coupling between states with different l and m quantum numbers. it follows that ∆Enlm = − 1 n.74) and (9. (9. according to electromagnetic theory. l. also. m = − 2 me c2 n 1 e2 E 2 + 2 En = − 2 me c2 n 4πǫ0 165 (12. our calculation still gives the correct answer when l = 0. that we are able to use non-degenerate perturbation theory in the above calculation. Note. However.120) Finally.75) that ∆Enlm 1 .l. m|V|n. 2 (l + 1/2) n3 a0 (12.121) 1 e2 ≃ (12. l. all coupled states have different n quantum numbers. Since p2 is an Hermitian operator. the above expression reduces to ∆Enlm = En where α2 n2 3 n .m . l.123) due to the charge on the nucleus.m = 2 me (En − V) ψn. and (9. because the perturbing Hamiltonian commutes with both L2 and Lz . Schr¨dinger’s equation for a unperturbed hydrogen atom can be written o p2 ψn.118) 1 e2 + r 4πǫ0 2 It follows from Eqs. m|V 2 |n.119) 1 e2 E 2 + 2 En = − 2 me c2 n 4πǫ0 1 e2 + n2 a0 4πǫ0 2 1 . a nonrelativistic particle moving in a electric ﬁeld E with velocity v also experiences an effective magnetic ﬁeld v×E (12. l. r2 (12.l.55).122) 4πǫ0 ¯ c h 137 is the dimensionless ﬁne structure constant.Time-Independent Perturbation Theory However. where V = −e2 /(4πǫ0 r). Hence. l. making use of Eqs. − l + 1/2 4 (12. m + n. somewhat fortuitously. c . an electron in a hydrogen atom experiences an electric ﬁeld α= E= er 4πǫ0 r3 (12. l. Now. and therefore have different energies. (9.

126) where L = me r × v is the electron’s orbital angular momentum. by a factor 2. and J2 discussed in Sect.129) (12.j. therefore. (12. using the above expression as the perturbing Hamiltonian.2.125) due to its spin angular momentum. the simultaneous eigenstates of the unperturbed Hamiltonian (12. (10. according to Sect. Hence.3. 11.j.127) are the same as the simultaneous eigenstates of L2 . It is important to know this since. the true spin-orbit correction to the Hamiltonian is H1 = e2 L · S.s.58) and (10.58) and the perturbing Hamiltonian (12.j. Hence. we can only safely apply perturbation theory to the simultaneous eigenstates of the unperturbed and perturbing Hamiltonians.mj be a simultaneous eigenstate of L2 .127) Let us now apply perturbation theory to the hydrogen atom.s. We.130) 2 Recall.166 QUANTUM MECHANICS Recall. that whilst J2 commutes with both L2 and S2 . (10. but does not commute with either Lz or Sz . from Sect. 2 8π ǫ0 me c2 r3 (12.mj = l (l + 1) ¯ 2 ψl.3. 11. J2 = L2 + S2 + 2 L · S. that an electron possesses a magnetic moment [see Eqs. J2 . 2 4πǫ0 me c2 r3 (12.mj . (2) Adopting the notation introduced in Sect. S.127) also commutes with both L2 and S2 .6. S2 . Hence. giving 1 2 (J − L2 − S2 ). expect an additional contribution to the Hamiltonian of a hydrogen atom of the form [see Eq. Now J=L+S (12.s. It follows that the perturbing Hamiltonian (12. and Jz corresponding to the eigenvalues L·S = L2 ψl. S2 .131) .59)] µ=− e S me (12. 12. h (2) (2) (12. This effect is known as spin-orbit coupling. It turns out that the above expression is too large. let ψl. it does not commute with either Lz or Sz .60)] H1 = −µ · B e2 = − v×r·S 4πǫ0 me c2 r3 = e2 L · S.128) is the total angular momentum of the system. 11. due to an obscure relativistic effect known as Thomas precession.

57).1/2.mj = s (s + 1) ¯ 2 ψl. − j + 1/2 4 (12.mj . (12.s. According to Eq. in the absence of ﬁne structure. (9. mj|H1 |l. it is conventional to refer to the energy eigenstates of a hydrogen atom which are also simultaneous eigenstates of J2 as nLj states. We can add these two corrections together (making use of the fact that j = l ± 1/2 for a hydrogen atom—see Sect. For n = 1. · · ·) as l = (0. L = (S. (9. where n is the radial quantum number.138) on these eigenstates for n = 1.Time-Independent Perturbation Theory S2 ψl.132) (12.j. there are two degenerate 1S1/2 states.s. the energy-shift induced in such a state by spin-orbit coupling is given by ∆El. 2 and 3. · · ·).s. h (2) (2) (2) (2) (2) (2) 167 (12. Finally.s. A comparison of this expression with Eq.1/2.1/2. (12.138).mj . mj 2 16π ǫ0 me c2 r3 = e2 ¯ 2 h [j (j + 1) − l (l + 1) − 3/4] 2 16π ǫ0 me c2 1 .3) to obtain a net energy-shift of ∆El. and j is the total angular momentum quantum number. 11. P. j. mj l.134) According to standard ﬁrst-order perturbation theory. r3 (12.136) 2 c2 a 3 16π ǫ0 me l (l + 1/2) (l + 1) n3 0 where n is the radial quantum number.j.1/2. h J2 ψl.j. 2. j. the ﬁne structure induced energy-shifts of these two states are .55). D.j. 1/2. 1/2.j. h Jz ψl. 1/2. 3.j.mj . Let us examine the effect of the ﬁne structure energy-shift (12.133) (12. mj e2 J2 − L2 − S2 = 1.s. j.j. (9. j. 1/2. (12. we have made use of the fact that s = 1/2 for an electron. It follows from Eq.mj = l. Now.135) Here.mj = En α2 n2 3 n .121) reveals that the energy-shift due to spin-orbit coupling is of the same order of magnitude as that due to the lowest-order relativistic correction to the Hamiltonian.j.s.mj = mj ¯ ψl.mj = .j. F.138) This modiﬁcation of the energy levels of a hydrogen atom due to a combination of relativity and spin-orbit coupling is known as ﬁne structure.mj = j (j + 1) ¯ 2 ψl.j. 1. the above expression reduces to ∆El. = En 2 n 2 l (l + 1/2) (l + 1) (12.76) that e2 ¯ 2 h j (j + 1) − l (l + 1) − 3/4 ∆El. and (9. making use of Eqs.137) where α is the ﬁne structure constant.58).mj α2 n {3/4 + l (l + 1) − j (j + 1)} .

gives rise to an expression (12. when added to expression (12. ﬁne structure does not break the degeneracy of the two 1S1/2 states of hydrogen. there are two 2S1/2 states. but are different from the induced energy-shift of the 2P3/2 states. For n = 3. and the 3D5/2 states. four 3P3/2 states. all of which are degenerate.1.138) which is both well-deﬁned and correct when l = 0.138). . there are two 3S1/2 states.137) does not have a well-deﬁned value for l = 0. and 3 energy states of a hydrogen atom is illustrated in Fig. 2. the 3P3/2 and 3D3/2 states. two 2P1/2 states. two 3P1/2 states. and six 3D5/2 states.168 3D5/2 3P3/2 3S1/2 3D3/2 3P1/2 QUANTUM MECHANICS 3D5/2 3P3/2 3S1/2 2P3/2 2S1/2 3D3/2 3P1/2 2P1/2 2P3/2 2S1/2 2P1/2 1S1/2 1S1/2 unperturbed + ﬁne structure Figure 12.121) it. in the absence of ﬁne structure. ﬁne structure breaks these states into three groups: the 3S1/2 and 3P1/2 states. Hence. somewhat fortuitously.1: Effect of the ﬁne structure energy-shift on the n = 1. in the absence of ﬁne structure. Not to scale. the same. but does break the degeneracy of these states relative to the 2P3/2 states. four 3D3/2 states. Hence. Note. 12. The effect of the ﬁne structure energy-shift on the n = 1. all of which are degenerate. and four 2P3/2 states. ﬁnally. that although expression (12. (12. For n = 2. 2 and 3 states of a hydrogen atom. ﬁne structure does not break the degeneracy of the 2S1/2 and 2P1/2 states of hydrogen. According to Eq. According to Eq. (12. the ﬁne structure induced energy-shifts of the 2S1/2 and 2P1/2 states are the same as one another.138).

h 2 me (12. (12. In this situation. 1/2. ψj. Here.145) It follows from Eqs.−1/2 (1) (12.139) e (L + 2 S) (12.mj = = l. we can treat H1 as a small perturbation acting on the simultaneous eigenstates of the unperturbed Hamiltonian and the ﬁne structure Hamiltonian. 1/2.1/2.142) since Jz = Lz + Sz . Hence.143)–(12. 1/2.mj are the simultaneous eigenstates of L2 . Since the magnitude of the internal ﬁeld is about 25 tesla.1/2 j + 1 + mj − 2l+ 1 1/2 ψmj +1/2. and the orthormality of the ψ(1) . Sz ψm. mj|Sz |l.57)–(10. S2 . including both orbital and spin contributions [see Eqs. j. Thus. mj|Sz |l. S2 . 1/2. according to Eqs. 1/2.m (2) whereas the ψj. j.±1/2 (12. S2 . The modiﬁcation to the Hamiltonian of the system is H1 = −µ · B. the ψ(1) s are the simultaneous eigenstates of L2 . Lz . j.±1/2 = ± (1) ¯ (1) h ψ . Now.124). (11. where µ=− (12. mj|H1 |l.1/2 + (1) j − mj 2l+1 1/2 ψmj +1/2.140) 2 me is the total electron magnetic moment. and Jz . from standard perturbation theory.−1/2 (1) (12.141) 2 me Suppose that the applied magnetic ﬁeld is much weaker than the atom’s internal magnetic ﬁeld (12. In particular. J2 . mj = ± mj ¯ h 2l +1 (12.Time-Independent Perturbation Theory 169 12. that l.mj = (2) j + mj 2l+1 1/2 ψmj −1/2. 2 m. eB H1 = (Lz + 2 Sz).144) when j = l − 1/2.145). this is a fairly reasonable assumption. the ﬁrst-order energy-shift induced by a weak external magnetic ﬁeld is ∆El.143) when j = l + 1/2.47) and (11. (10.mj = j + 1 − mj 2l+ 1 1/2 (1) ψmj −1/2.146) .48).9 Zeeman Effect Consider a hydrogen atom placed in a uniform z-directed external magnetic ﬁeld of strength B. m.59)]. mj eB (mj ¯ + l. these states are the simultaneous eigenstates of L2 . and (2) ψj. and Sz . j. 1/2. Of course.j. and Jz (see previous section). J2 . j. mj ) . j. (12.

the four degenerate 2S1/2 and 2P1/2 states are split by (2/3) µB B. Sp is the proton spin.148) 1 2l+ 1 (12. Thus. and therefore possesses a magnetic moment. that since the ψl. It.1/2. (10. µB = e¯ h = 5. thus.149) ∆Ej=l+1/2 = µB B 2l+ 1 Likewise. Note. whereas the four degenerate 2P3/2 states are split (2) by (4/3) µB B.149) and (12. Note that the magnetic . By analogy with Eq.2. 2 mp (12. the two degenerate 1S1/2 states of the hydrogen atom are split by 2 µB B. Here. This is illustrated in Fig.147) that the Zeeman effect splits degenerate states characterized by j = l + 1/2 into 2 j + 1 equally spaced states of interstate spacing 2l+ 2 .151) where µp is the proton magnetic moment.10 Hyperﬁne Structure The proton in a hydrogen atom is a spin one-half charged particle. 2l+ 1 (12. (12. follows from Eq.mj are not simultaneous eigenstates of the unperturbed and perturbing Hamiltonians.j.147) is known as the Bohr magnetron.58). the induced energy-shift when a hydrogen atom is placed in an external magnetic ﬁeld—which is known as the Zeeman effect—becomes ∆El.788 × 10−5 eV/T 2 me (12.150) can only be regarded as the expectation values of the magnetic-ﬁeld induced energy-shifts. we can write µp = gp e Sp . However. Eqs.170 QUANTUM MECHANICS when j = l ± 1/2.150) In conclusion. the Zeeman effect splits degenerate states characterized by j = l − 1/2 into 2 j + 1 equally spaced states of interstate spacing ∆Ej=l−1/2 = µB B 2l . in the presence of a weak external magnetic ﬁeld. 12. 12. (12. (12. the quantum number mj takes values differing by unity in the range −j to j. and the proton gyromagnetic ratio gp is found experimentally to take that value 5. Of course.59.mj = µB B mj 1 ± where the ± signs correspond to j = l ± 1/2. ﬁnally. as long as the external magnetic ﬁeld is much weaker than the internal magnetic ﬁeld. Likewise. these expectation values are almost identical to the actual measured values of the energy-shifts.

8π mp me r3 3 mp me (12. moment of a proton is much smaller (by a factor of order me /mp ) than that of an electron.2: The Zeeman effect for the n = 1 and 2 states of a hydrogen atom.155) . Here.152) where er = r/r. if the size of the loop goes to zero then the ﬁeld will be inﬁnite at the origin. the proton’s magnetic moment generates a magnetic ﬁeld of the form B= 2 µ0 µ0 µ δ3 (r).154) me Here. Not to scale.58) and (10. the perturbing Hamiltonian is written H1 = µ 0 g p e2 µ0 gp e2 3 (Sp · er ) (Se · er ) − Sp · Se + Sp · Se δ3 (r). the Hamiltonian of the electron in the magnetic ﬁeld generated by the proton is simply H1 = −µe · B. 3 (µp · er ) er − µp + 3 4π r 3 p (12. ǫ = µB B.153) e Se . µe is the electron magnetic moment [see Eqs. We can understand the origin of the delta-function term in the above expression by thinking of the proton as a tiny current loop centred on the origin. where µe = − (12. (12. and Se the electron spin. Thus.Time-Independent Perturbation Theory 2P3/2 (4/3)ǫ (4/3)ǫ (4/3)ǫ 2P1/2 (2/3)ǫ (2/3)ǫ (2/3)ǫ 171 2S1/2 1S1/2 2ǫ unperturbed + ﬁne structure + Zeeman Figure 12. (10. All magnetic ﬁeld-lines generated by the loop must pass through the loop. and this contribution is what is reﬂected by the delta-function term. Hence. Now. According to classical electromagnetism.59)].

156) For the ground-state of hydrogen. It follows. 2 (12. The net energy difference between the singlet and the triplet states is ∆E = me 2 8 gp α E0 = 5. 3 mp (12. µ 0 g p e2 ∆E = 8π mp me 3 (Sp · er ) (Se · er ) − Sp · Se r3 µ 0 g p e2 + Sp · Se |ψ(0)|2 . the energy-shift induced by spinspin coupling between the proton and the electron is the expectation value of the perturbing Hamiltonian. it is easily demonstrated that |ψ000 (0)|2 = 3 1/(π a0 ). there are three spin 1 states. For all states. 3 mp me (12.172 QUANTUM MECHANICS Note that. h h Hence.162) . According to Sect. This splitting is known as hyperﬁne structure. We can show that Sp · Se = 1 2 2 2 (S − Se − Sp ). that spin-spin coupling breaks the degeneracy of the two 1S1/2 states in hydrogen.4. (12. Sp . and S2 .158) Thus. known as triplet states. However. since we have neglected coupling between the proton spin and the magnetic ﬁeld generated by the electron’s orbital motion. and lowering that of the singlet. known as the singlet state. and 1 Sp · Se = ¯ 2 h (12. when two spin one-half particles are combined (in the absence of orbital angular momentum) the net state has either spin 1 or spin 0. In fact. and a 2 2 single spin 0 state. Hence. we obtain µ 0 g p e2 ∆E = Sp · Se . which is spherically symmetric. the ﬁrst term in the above expression vanishes by symmetry. According to standard ﬁrst-order perturbation theory. The eigenvalue of S2 is 0 for the singlet state. 3 h (12. and 2 ¯ 2 for the triplet states. Moreover. both the proton and the electron are spin one-half particles. the above expression is only valid for l = 0 states.157) 3 3π mp me a0 Let S = Se + Sp be the total spin.160) Sp · Se = − ¯ 2 4 for the singlet state. lifting the energy of the triplet conﬁguration.159) (12.161) 4 for the triplet states. 11.88 × 10−6 eV. from the above analysis. the eigenvalues of Se and Sp are (3/4) ¯ 2 . the simultaneous eigenstates of the perturbing Hamiltonian and the main Hamilto2 2 nian are the simultaneous eigenstates of Se . Thus.

Time-Independent Perturbation Theory 173 where E0 = 13. If we convert the above energy into a wavelength then we obtain λ = 21.6 eV is the (magnitude of the) ground-state energy. The 21 cm line is famous in radio astronomy because it was used to map out the spiral structure of our galaxy in the 1950’s. by a factor me /mp . . (12.163) This is the wavelength of the radiation emitted by a hydrogen atom which is collisionally excited from the singlet to the triplet state. and then decays back to the lower energy singlet state.1 cm. Note that the hyperﬁne energy-shift is much smaller. than a typical ﬁne structure energy-shift.

174 QUANTUM MECHANICS .

e.12) that if the system is in one of these eigenstates then. in principle.12) ψ(t) = m cm exp (−i Em t/¯ ) ψm . (13.1 Introduction Consider a system whose Hamiltonian can be written H(t) = H0 + H1 (t). 4.. H1 now represents a small time-dependent external perturbation.1) Here. In other words. 4. h since the unperturbed eigenstates are assummed to be orthonormal: i. (13. H0 is again a simple time-independent Hamiltonian whose eigenvalues and eigenstates are known exactly. Let the eigenstates of H0 take the form H0 ψm = Em ψm . it remains in this state for ever. the probability of ﬁnding the system in state n at time t is Pn (t) = | ψn |ψ |2 = |cn exp (−i En t/¯ )|2 = |cn |2 = Pn (0). However. Let us investigate this effect.6) (13.3) where the cm are complex numbers.4) Now. n|m = δnm . the presence of a small time-dependent perturbation can. However. (13. h (13. the initial state is some linear superposition of the unperturbed energy eigenstates. In the absence of the time-dependent perturbation. 13. the time evolution of the system is simply (see Sect.2) We know (see Sect.2 Preliminary Analysis Suppose that at t = 0 the state of the system is represented by ψ(0) = m cm ψm . a time-dependent perturbation can cause the system to make transitions between its unperturbed energy eigenstates.Time-Dependent Perturbation Theory 175 13 Time-Dependent Perturbation Theory 13. (13. Thus. in the absence of an external perturbation.5) . give rise to a ﬁnite probability that if the system is initially in some eigenstate ψn of the unperturbed Hamiltonian then it is found in some other eigenstate at a subsequent time (since ψn is no longer an exact eigenstate of the total Hamiltonian).

(4.11) Projecting out the component of the above equation which is proportional to ψn . The time-dependent Schr¨dinger equation [see Eq. from the slow variation of the amplitudes cn (t). we obtain i¯ h where Hnm (t) = n|H1 (t)|m .7) the eigenstates ψm are time-independent (they are actually the eigenstates of H0 evaluated at the initial time. According to Eq. h (13.e. h m (13. which depends entirely on the perturbation (i. which specify the probabilities of ﬁnding the system in these eigenstates at time t. (13.176 QUANTUM MECHANICS Clearly. ∂t (13. which depends on the unperturbed Hamiltonian. (13. h (13. the probability of ﬁnding the system in state ψn at time t is exactly the same as the probability of ﬁnding the system in this state at the initial time.8) 2 Now.7) that (H0 + H1 ) ψ = m cm exp (−i Em t/¯ ) (Em + H1 ) ψm . we can equate the right-hand sides of the previous two equations to obtain i¯ h m dcm exp (−i Em t/¯ ) ψm = h dt cm exp (−i Em t/¯ ) H1 ψm . Thus.63)] yields o i¯ h ∂ψ(t) = H(t) ψ(t) = [H0 + H1 (t)] ψ(t). (13. t = 0. m (13. t = 0). (13. and ωnm = (13. it follows from Eq. with H1 = 0. hence. Here. (13. is determined .8).13) dcn (t) = dt Hnm (t) exp ( i ωnm t) cm (t). (13. According to Eqs. we can write ψ(t) = cm (t) exp (−i Em t/¯ ) ψm . with H1 = 0. we have carefully separated the fast phase oscillation of the eigenstates. using Eq.6).12).7) m where Pn (t) = |cn (t)| .9) We also have i¯ h ∂ψ = ∂t i¯ h m dcm + cm Em exp (−i Em t/¯ ) ψm ..12) En − Em . Note that in Eq.14) ¯ h Suppose that there are N linearly independent eigenstates of the unperturbed Hamiltonian. h dt (13. cn is constant in time if H1 = 0).10) since the ψm are time-independent. cn—to vary with time. we expect Pn —and. the time-dependence of the set of N coefﬁcients cn . However.

we cannot generally ﬁnd exact solutions to these equations.15) (13. (13.20) are c2 (t) = −i (ω − ω21 ) t −i γ exp sin(Ω t).16) Suppose.21) Once we have solved for c2. Thus.e. d2 c2 dc2 + i (ω − ω21 ) + γ2 c2 = 0. Note that Eqs. thus. are zero: i. 2 dt dt (13. dt i (13.e. 1|H1 |1 = 2|H1 |2 = 0. The above two equations can be combined to give a second-order h differential equation for the time-variation of the amplitude c2 : i. (13. 1|H1 |2 = 2|H1 |1 ∗ = γ ¯ exp(i ω t). we have to obtain approximate solutions via suitable expansions in small quantities. for the sake of simplicity.20) where ω21 = (E2 −E1 )/¯ .3 Two-State System Consider a system in which the time-independent Hamiltonian possesses two eigenstates. Unfortunately. we can use Eq. (13. denoted H0 ψ1 = E1 ψ1 .e. Instead. Let us search for a solution in which the system is certain to be in state 1 (and. has no chance of being in state 2) at time t = 0. It is easily demonstrated that the appropriate solutions to (13.19) (13.. Eq.12) without approximation. for the particuilarly simple case of a two-state system (i. (13.18) where γ and ω are real. that the diagonal elements of the interaction Hamiltonian. dt dc2 i = γ exp [−i (ω − ω21 ) t] c1 .Time-Dependent Perturbation Theory 177 by N coupled ﬁrst-order differential equations. (13. 13. our initial conditions are c1 (0) = 1 and c2 (0) = 0.. For a two-state system. Ω 2 (13. Note that it is only the off-diagonal matrix elements which give rise to the effect which we are interested in—namely.20) to obtain the amplitude c1 . However. h (13. it is actually possible to solve Eqs.22) . This solution is of great practical importance.12) reduces to dc1 = γ exp [+i (ω − ω21 ) t] c2 .. H1 . H0 ψ2 = E2 ψ2 .12) are exact—we have made no approximations at this stage. transitions between states 1 and 2. N = 2). (13.21) and (13..e.17) The off-diagonal elements are assumed to oscillate sinusoidally at some frequency ω: i.

we ﬁnd that P1 (t) = cos2 (γ t). The absorption-emission cycle also takes place away from the resonance. Equation (13. the system periodically ﬂip-ﬂops between states 1 and 2 under the inﬂuence of the time-dependent perturbation. In other words. if we plot the maximum value of P2 (t) as a function of the applied frequency. Thus. At resonance. nor is the minimum of P1 (t) zero.. a bound electron) placed in a uniform z-directed magnetic ﬁeld. the probability of ﬁnding the system in state 2 at time t is P2 (t) = |c2 (t)|2.27) (13. This implies that the system alternatively absorbs and emits energy from the source of the perturbation. the time-dependent perturbation is only effective at causing transitions between states 1 and 2 if its frequency of oscillation lies in the approximate range ω21 ± 2 γ. 2 (13. ω. the weaker the perturbation (i. ω. when the oscillation frequency of the perturbation. This means that the maximum value of P2 (t) is no longer unity. we obtain a resonance curve whose maximum (unity) lies at the resonance. Likewise. the amplitude of the oscillation in the coefﬁcient c2 is reduced.178 c1 (t) = exp − where Ω= γ2 + (ω − ω21 )2 /4. 2 /4 + (ω − ω21 ) 2 (13. i (ω − ω21 ) t cos(Ω t) 2 QUANTUM MECHANICS i (ω − ω21 ) t i (ω − ω21 ) exp sin(Ω t). if the applied frequency differs from the resonant frequency by substantially more than 2 γ then the probability of the system jumping from state 1 to state 2 is always very small. and then subject to . the system starts off in state 1 at t = 0.4 Spin Magnetic Resonance Consider a system consisting of a spin one-half particle with no orbital angular momentum (e. and so on. P2 (t) = sin (γ t).g. After a time interval π/(2 γ) it is certain to be in state 2. the probability of ﬁnding the system in state 1 at time t is simply P1 (t) = |c1 (t)|2.e.23) (13. 13. It follows that P1 (t) = 1 − P2 (t).24) Now. After a further time interval π/(2 γ) it is certain to be in state 1 again. 2Ω 2 (13. when ω = ω21 .26) exhibits all the features of a classic resonance. and whose full-width half-maximum (in frequency) is 4 γ..28) According to the above result.26) This result is known as Rabi’s formula. In fact. Clearly. However.25) (13. the smaller γ becomes). matches the frequency ω21 . the narrower the resonance. Thus. P2 (t) = γ2 γ sin2 (Ω t).

The eigenstates of the unperturbed Hamiltonian are the “spin up” and “spin down” states. 10). (13.38) (13. where H0 = − and H1 = − g e B0 Sz . The Hamiltonian is written H = −µ · B = H0 + H1 . respectively. the electric component of the wave has no effect.29) where B0 and B1 are constants. 10).32) 2m Here. 2m g e B1 . respectively. with B1 ≪ B0 .37) (13.36) 4m It can be seen that this system is exactly the same as the two-state system discussed in the previous subsection.40) ψ2 → χ− .35) g e B1 exp( i ω t). we h h have g e ¯ B0 h H0 χ± = ∓ χ± . (13. provided that the make the following indentiﬁcations: =− ψ1 → χ+ . g and m are the gyromagnetic ratio [see Eq.39) (13. γ → − 4m . g e B0 ω21 → . The rotating magnetic ﬁeld usually represents the magnetic component of an electromagnetic wave propagating along the z-axis.33) 4m The time-dependent Hamiltonian can be written H1 = − g e B1 [exp( i ω t) S− + exp(−i ω t) S+] . Thus. 2m (13.31) g e B1 [cos(ω t) Sx + sin(ω t) Sy] . Thus.30) (13.151)] and mass of the particle in question. denoted χ+ and χ− .Time-Dependent Perturbation Theory 179 a small time-dependent magnetic ﬁeld rotating in the x-y plane at the angular frequency ω. these states are the eigenstates of Sz corresponding to the eigenvalues +¯ /2 and −¯ /2 respectively (see Sect. Of course. B = B0 ez + B1 [cos(ω t) ex + sin(ω t) ey] . 4m (13.34) where S+ and S− are the conventional raising and lowering operators for spin angular momentum (see Sect. and −|H1 |+ = +|H1 |− ∗ (13. (13. In this system. (12. It follows that +|H1 |+ = −|H1 |− = 0. (13. (13.

Thus. subject to the initial condition cn (0) = δni . If we now apply a magnetic perturbation rotating at the resonant frequency then. 13. m (13. H0 ψn = En ψn . according to (13. thus. where the En are the unperturbed energy levels. the expectation values of Sx and Sy oscillate because of the spin precession. the cn h . according to the analysis of the previous subsection.44) Let us attempt a perturbative solution of Eq. H0 . (13. in the presence of a small time-dependent perturbation to the Hamiltonian. n |cn |2 = 1). initially.42). in addition to the spin precession. the system undergoes a succession of spin ﬂips. the frequency at which the particles absorb energy from the oscillating ﬁeld). the wavefunction of the system takes the form ψ(t) = cn (t) exp(−i ωn t) ψn.42) where Hnm (t) = n|H1 (t)|m and ωnm = (En − Em )/¯ .2. Experimentalists are able to measure the gyromagnetic ratios of spin onehalf particles to a high degree of accuracy by placing the particles in a uniform magnetic ﬁeld of known strength.42) is. Now. and n|m = δnm . the probability of ﬁnding h the system in the nth eigenstate at time t is simply Pn (t) = |cn (t)|2 (13. The ψn are the stationary orthonormal eigenstates of the time-independent unperturbed Hamiltonian. Now. Equation (13. We also know that if the oscillation frequency of the applied ﬁeld is very different from the resonant frequency then there is virtually zero probability of the ﬁeld triggering a spin ﬂip.. The width of the resonance (in frequency) is determined by the strength of the oscillating magnetic perturbation. Suppose that at t = 0 the system is in some initial energy eigenstate labeled i. Finally. (13. In the absence of the perturbation.6). χ+ ↔ χ− . 10.41) n where ωn = En /¯ .42) using the ratio of H1 to H0 (or Hnm to ¯ ωnm . but the expectation value of Sz remains invariant.180 QUANTUM MECHANICS The resonant frequency. is simply the spin precession frequency in a uniform magnetic ﬁeld of strength B0 (see Sect. By determining the resonant frequency (i.5 Perturbation Expansion Let us recall the analysis of Sect. The amplitudes cn (t) satisfy h i¯ h dcn = dt Hnm exp( i ωnm t) cm . ω21 . H1 (t). and then subjecting them to an oscillating magnetic ﬁeld whose frequency is gradually scanned. it is possible to determine the gyromagnetic ratio (assuming that the mass is known). 13.43) (assuing that. (13. to be more exact) as our expansion parameter.e.

i cn (t) = δni − ¯ h t Hni (t ′) exp( i ωni t ′ ) dt ′.48) and (13.50) 13.45) n The ﬁrst-order solution is obtained. (13. the zeroth-order solution is simply c(0) (t) = δni . cf (t) = − where Vfi = † Vfi = (13. and spin operators. (13.Time-Dependent Perturbation Theory 181 are constant in time in the absence of the perturbation. that our perturbative solution is clearly only valid provided Pi→f (t) ≪ 1.6 Harmonic Perturbations Consider a (Hermitian) perturbation which oscillates sinusoidally in time. (13. is 2 i t (13.46) subject to the boundary condition c(1) (0) = 0. to ﬁrst-order. ﬁnally. Thus. in general. we obtain i¯ h dc(1) n = dt Hnm exp( i ωnm t) c(0) = Hni exp( i ωni t).48) Hence.47) It follows that. Hence. the probability of ﬁnding the system in some ﬁnal energy eigenstate labeled f at time t. via iteration. a function of position. It follows from Eqs. where V is. (13. 0 (13.51) that.51) i ¯ h t 0 † Vfi exp( i ω t ′) + Vfi exp(−i ω t ′) exp( i ωfi t ′ ) dt ′. given that it is deﬁnitely in a different initial energy eigenstate labeled i at time t = 0. (13.49) Hfi (t ′) exp( i ωfi t ′ ) dt ′ . m m (13. This is usually termed a harmonic perturbation. by substituting the zeroth-order solution into the right-hand side of Eq.52) f|V|i .53) (13. The solution to the above equation is n c(1) = − n i ¯ h t Hni (t ′ ) exp( i ωni t ′) dt ′ . (13. f|V † |i = i|V|f ∗ .54) . 0 (13. Such a perturbation takes the form H1 (t) = V exp( i ω t) + V † exp(−i ω t).42). up to ﬁrst-order in our perturbation expansion. Pi→f (t) = |cf (t)|2 = − ¯ 0 h Note. momentum.

56) x Now. Eventually. Integration with respect to t ′ yields cf (t) = − it (Vfi exp [ i (ω + ωfi ) t/2] sinc [(ω + ωfi ) t/2] ¯ h (13. (13. t (13. 13. where sin x . Hence.57) 2π . (13. as t increases.1: The functions sinc(x) (dashed curve) and sinc2 (x) (solid curve).1).55) † +Vfi exp [−i (ω − ωfi ) t/2] sinc [(ω − ωfi ) t/2] . and is fairly negligible ∼ when |x| ≫ π (see Fig. Pi→f = |cf |2 yields |ω − ωfi | < ∼ Pi→f (t) = t2 † 2 2 2 2 2 |Vfi | sinc [(ω + ωfi ) t/2] + |Vfi | sinc [(ω − ωfi ) t/2] . in this limit.58) t respectively.182 QUANTUM MECHANICS Figure 13. The vertical dotted lines denote the region |x| ≤ π. the ﬁrst and second terms on the right-hand side of Eq. ¯ h (13. (13. Thus.55) are only non-negligible when sinc x ≡ |ω + ωfi | < ∼ and 2π . the function sinc(x) takes its largest values when |x| < π. when t ≫ 2π/|ωfi |. the ranges in ω over which these two terms are nonnegligible gradually shrink in size. Clearly.59) . these two ranges become strongly non-overlapping.

This process is h known as stimulated emission.66) H → H − q φ.e. The resonance at ω = −ωfi corresponds to Ef − Ei = −¯ ω. Stimulated emission and absorption are mutually exclusive processes.28). since the ﬁrst requires ωfi < 0. ¯ h (13. This behaviour is entirely consistent with our earlier result (13. h (13. Hence.61) This implies that the system gains energy ¯ ω from the perturbing ﬁeld. we can write the transition probabilities for both processes separately. the transition probability Pi→f (t) varies as t2 [since sinh(0) = 1]. whereas the second requires ωfi > 0.. (13. It follows that the above expression exhibits a resonant response ∼ to the applied perturbation at the frequencies ω = ±ωfi . in the limit γ t ≪ 1 (recall that our perturbative solution is only valid as long as Pi→f ≪ 1). for the two-state system.Time-Dependent Perturbation Theory 183 Now. at ω = ±ωfi ). The unperturbed Hamiltonian of the system is H0 = p2 + V0 (r). 2 me (13. This h process is known as absorption. . non-quantized) electromagnetic radiation. whilst making a h transition to a ﬁnal state whose energy is less than the initial state by ¯ ω. the widths of these resonances decease linearly as time increases. the function sinc2 (x) is very strongly peaked at x = 0.65) (13. (13. Thus. from (13..e.64) Now.7 Electromagnetic Radiation Let us use the above results to investigate the interaction of an atomic electron with classical (i. h (13. and is completely negligible for |x| > π (see Fig. and |Vfi |2 = |Vif |2 . Likewise. the transition probability for stimulated emission is t2 † stm Pi→f (t) = 2 |Vif | 2 sinc2 [(ω − ωif ) t/2] . 13.59). the standard classical prescription for obtaining the Hamiltonian of a particle of charge q in the presence of an electromagnetic ﬁeld is p → p + q A. the transition probability for absorption is abs Pi→f (t) t2 † 2 = 2 |Vfi | sinc2 [(ω − ωfi ) t/2] . whilst making a h transition to a ﬁnal state whose energy is greater than that of the initial state by ¯ ω.1).63) 13.62) ¯ h † where we have made use of the facts that ωif = −ωfi > 0.60) This implies that the system loses energy ¯ ω to the perturbing ﬁeld. The resonance at ω = ωfi corresponds to Ef − Ei = ¯ ω. Moreover. At each of the resonances (i.

(13.75) with H0 = and H1 ≃ − p2 + V0 (r). 2 me e A·p . The Hamiltonian becomes H = H0 + H1 (t). 2 me (13. monochromatic. the Hamiltonian of an atomic electron placed in an electromagnetic ﬁeld is (p − e A)2 H= + e φ + V0 (r). has been neglected.72) (13.68) This prescription also works in quantum mechanics. A = A0 ǫ cos(k·r − ωt) . In this case.69) where A and φ are functions of the position operators. ∂t (13..71) p2 − e A·p − e p·A + e2 A2 2 me + e φ + V0 (r). and φ(r) the scalar potential. p·A = A·p. ∂A .78) .77) where the A2 term. provided that we adopt the gauge ∇·A = 0. Note that E = −∇φ − B = ∇ × A.e.70) Suppose that the perturbation corresponds to a linearly polarized. Thus. H= e A·p e2 A2 p2 − + + e φ + V0 (r). 2 me (13.76) (13.73) (13.184 QUANTUM MECHANICS where A(r) is the vector potential. The perturbing Hamiltonian can be written H1 = − e A0 ǫ·p [exp( i k·r − i ωt) + exp(−i k·r + i ωt)] . plane-wave. (13. The above equation can be written H= Now. Hence. Note that ǫ · k = 0.67) (13. (13. and ǫ a unit vector which speciﬁes the direction of polarization (i. me (13.74) where k is the wavevector (note that ω = k c). the direction of E). φ = 0. 2 me me 2 me (13. which is second order in A0 .

(13.86) . respectively. that the above expression is only valid provided the radiation in question is incoherent: i. but instead extends over a range of frequencies. not surprisingly. (13. It thus follows that abs Pi→f (t) = t 2 e2 | f|ǫ·p exp( i k·r)|i | 2 u sinc2 [(ω − ωfi ) t/2]. Suppose that the incident radiation is not monochromatic. Given that the function sinc2 [(ω − ωfi ) t/2] is very strongly peaked (see Fig. (13.79) It follows from Eqs.81) where E0 = A0 ω and B0 = E0 /c are the peak electric and magnetic ﬁeld-strengths. Equation (13. 2 2 ǫ0 ¯ 2 me ωfi h (13.82) generalizes to ∞ abs Pi→f (t) = −∞ t 2 e2 | f|ǫ·p exp( i k·r)|i | 2 ρ(ω) sinc2 [(ω − ωfi ) t/2] dω.80) Now. 13.83) where ρ(ω) dω is the energy density of radiation whose frequencies lie between ω and ω + dω. and ∞ sinc2 (x) dx = π. the mean energy density of an electromagnetic wave is u= 1 2 = 1 ǫ0 |E0 |2 . (13. however.53).e.79) that the transition probability for radiation induced absorption is abs Pi→f (t) = t2 e2 |A0 |2 | f|ǫ·p exp( i k·r)|i | 2 sinc2 [(ω − ωfi ) t/2]. there are no phase correlations between waves of different frequencies.1) about ω = ωfi (assuming that t ≫ 2π/ωfi ). (13.51).Time-Dependent Perturbation Theory This has the same form as Eq. (13.85) −∞ the above equation reduces to abs Pi→f (t) = π e2 ρ(ωfi ) | f|ǫ·p exp( i k·r)|i | 2 t. This follows because it is permissible to add the intensities of incoherent radiation. the transition probability for radiation induced absorption (or stimulated emission) is directly proportional to the energy density of the incident radiation. whereas we must always add the amplitudes of coherent radiation.63). and (13. We can write ∞ u= −∞ ρ(ω) dω.. provided that V† = − e A0 ǫ·p exp( i k·r ). 2 me 185 (13.82) Thus.84) 2 2 ω2 2 ǫ0 ¯ me h Note. 2 2 ¯ 4 me h ǫ0 |E0 |2 |B0 |2 + 2 2 µ0 (13. 2 (13. 2 2 ω2 2 ǫ0 ¯ me h (13.

e.82)] to one which is proportional to t. Here. the transition probability for stimulated emission can be shown to take the form stm Pi→f (t) = 1 − exp −wstm t . wabs is the transition probability per unit time i→f interval.e. i→f i→f dt abs with the initial condition Pi→f (0) = 0. It follows that i→f abs dPi→f abs + wabs Pi→f = wabs .8 Electric Dipole Approximation In general.186 QUANTUM MECHANICS Note that in integrating over the frequencies of the incoherent radiation we have transformed a transition probability which is basically proportional to t2 [see Eq.92) 13. i→f Using similar arguments to the above. respectively.. the wavelength of the type of electromagnetic radiation which induces.87) is constant in time is universally valid. the above expression is only abs valid when Pi→f ≪ 1. (13.86) provided wabs t ≪ 1: i. However.93) . (13. Thus. 2) abs abs abs Pi→f (t + dt) − Pi→f (t) = 1 − Pi→f (t) wabs dt : i→f (13. i→f (13. the result that wabs ≡ i→f abs dPi→f π e2 ρ(ωfi ) | f|ǫ·p exp( i k·r)|i | 2 = 2 2 ω2 dt ǫ0 ¯ me fi h (13.88) i.90) abs This result is consistent with Eq. (13. transitions between different atomic energy levels is much larger than the typical size of an atom. As has already been explained. 2 2 ω2 ǫ0 ¯ me if h (13.91) where the corresponding transition rate is written wstm i→f π e2 ρ(ωif ) = | i|ǫ·p exp( i k·r)|f | 2 . the probability that the system makes a transition from state i to state f between times t and t + dt is equivalent to the probability that the system does not make a transition between times 0 and t and then makes a transition in a time interval dt—the probabilities abs of these two events are 1 − Pi→f (t) and wabs dt. otherwise known as the transition rate. The above equation can be solved to give abs Pi→f (t) = 1 − exp −wabs t .. provided that Pi→f ≪ 1.89) (13. i→f (13. exp( i k·r) = 1 + i k·r + · · · . we can write (see Cha. Given that the transition rate is constant. or is emitted during.

It follows that the vector ǫ. if sin2 θ cos2 φ dΩ . (13. we can deﬁne a set of Cartesian coordinates such that the wavevector k. ǫ0 ¯ h dif = f|e r|i (13. we are more interested in the transition rates induced by unpolarized isotropic radiation. lies in the x-z plane.98) over all possible polarizations and propagation directions of the wave. It follows that f|ǫ·p exp( i k·r)|i ≃ ǫ· f|p|i . since it has to be orthogonal to k. ǫ0 ¯ h π 2 wstm = (13. |ǫ·dif | 2 2 = dif (13.95) (13. we can write k = (0. Equations (13. (13.101) (13. Thus. 0. 0.96) ¯ h Thus. unity.94) me f|[r. To facilitate this process. induced by a linearly polarized plane-wave. Here.98) give the transition rates for absorption and stimulated emission.98) i→f 2 |ǫ·dif | ρ(ωif ). our previous expressions for the transition rates for radiation induced absorption and stimulated emission reduce to π 2 (13. me (13.102) av (13. respectively. points along the z-axis. it is readily demonstrated that [r. 0). H0 ] = so f|p|i = −i i¯ p h .99) respectively. which speciﬁes the direction of wave propagation.97) and (13. is the effective electric dipole moment of the atom when making a transition from state i to state f. Now.103) We must now average the above quantity over all possible values of θ and φ. which speciﬁes the direction of the atomic dipole moment.97) wabs = i→f 2 |ǫ·dif | ρ(ωfi ).104) . Actually. To obtain these we must average Eqs. sin φ. and the vector dif . must lie in the x-y plane. k). ǫ = (cos φ. H0]|i = i me ωfi f|r|i .100) (13. This approach is known as the electric dipole approximation.97) and (13.Time-Dependent Perturbation Theory 187 can be approximated by its ﬁrst term. which implies that |ǫ·dif | 2 = d2 sin2 θ cos2 φ. dif cos θ). dif = (dif sin θ. 4π (13. which speciﬁes the direction of wave polarization. Thus.

3 ǫ0 ¯ h (13. we have calculated the rates of radiation induced transitions between two atomic states. One of the tenants of statistical thermodynamics is that in thermal equilibrium we have so-called detailed balance. with Ei > Ef . 3 ǫ0 ¯ 2 h π 2 = 2 dif ρ(ωif ). Consider two atomic states.106) Hence. and the integral is taken over all solid angle.105) av 3 2 Here. It is easily demonstrated that 2 dif 2 |ǫ·dif | = . it should be possible for such an atom to spontaneously jump into an state with a lower energy via the emission of a photon whose energy is equal to the difference between the energies of the initial and ﬁnal states. According to statistical thermodynamics. Let the system have attained thermal equilibrium. This process is known as absorption when the energy of the ﬁnal state exceeds that of the initial state. This well-known result was ﬁrst obtained by Max Planck in 1900. and stimulated emission when the energy of the ﬁnal state is less than that of the initial state. π 2 dif ρ(ωfi ). Now. π2 c3 exp(¯ ω/kB T ) − 1 h (13. labeled i and f.108) 13.107) (13. dif stands for 2 dif = | f|e x|i | 2 + | f|e y|i | 2 + | f|e z|i | 2 .9 Spontaneous Emission So far. we would not expect an atom in a given state to spontaneously jump into an state with a higher energy. the transition rates for absorption and stimulated emission induced by unpolarized isotropic radiation are wabs = i→f wstm i→f respectively. Consider a very large ensemble of similar atoms placed inside a closed cavity whose walls (which are assumed to be perfect emitters and absorbers of radiation) are held at the constant temperature T . (13. in the absence of any external radiation. This . On the other hand. (13.109) where kB is the Boltzmann constant. This process is known as spontaneous emission.188 QUANTUM MECHANICS where dΩ = sin θ dθ dφ. It is possible to derive the rate of spontaneous emission between two atomic states from a knowledge of the corresponding absorption and stimulated emission rates using a famous thermodynamic argument due to Einstein. the cavity is ﬁlled with so-called “black-body” electromagnetic radiation whose energy spectrum is ρ(ω) = ¯ h ω3 .

we have Wi→f = Wf→i .110) where wspn is the rate of spontaneous emission (for a single atom) between states i and i→f f.108).114) Thus. which gives wspn = i→f Nf abs wf→i − wstm . it remains correct even in the absence of radiation. Likewise. h Ni exp(−Ei /kB T ) (13.107) and (13.115) Note. the corresponding absorption and stimulated emission rates for an atom in a radiation-ﬁlled cavity are wabs = i→f wstm = i→f 3 2 ωfi dif 1 . (13. The former rate (i. that. irrespective of any other atomic states. and Ni is the number of atoms in the ensemble in state i.109). Finally. the latter rate takes the form Wf→i = Nf wabs .e.112) According to Eqs. This implies that exp(−Ef /kB T ) Nf = = exp( ¯ ωif /kB T ). 3π ǫ0 ¯ c3 h (13. i→f Ni (13.113) Now.114) that the rate of spontaneous emission between states i and f takes the form wspn = i→f 3 2 ωif dif .Time-Dependent Perturbation Theory 189 means that. and make the opposite transition as a consequence of absorption. and (13. the rate at which atoms in the ensemble leave state i due to transitions to state f is exactly balanced by the rate at which atoms enter state i due to transitions from state f. (13.113). it follows from Eq. The above expressions describe how atoms in the ensemble make transitions from state i to state f due to a combination of spontaneous and stimulated emission. In thermal equilibrium. we can also write wspn = i→f π Nf 2 −1 2 dif ρ(ωif ). number of transitions per unit time in the ensemble) is written Wi→f = Ni (wspn + wstm ).111) f→i where Nf is the number of atoms in the ensemble in state f. although the above result has been derived for an atom in a radiation-ﬁlled cavity. 3 exp(¯ ω /k T ) − 1 3π ǫ0 ¯ c h h fi B 3 2 ωif dif 1 . (13. (13.. i→f i→f (13.116) (13. Ni 3 ǫ0 ¯ h (13. another famous result in statistical thermodynamics is that in thermal equilibrium the number of atoms in an ensemble occupying a state of energy E is proportional to exp(−E/kB T ).117) . 3 exp(¯ ω /k T ) − 1 3π ǫ0 ¯ c h h if B (13.

However. Let us investigate this effect.118) where α = e2 /(4π ǫ0 ¯ c) ≃ 1/137 is the ﬁne-structure constant.119) where ω0 is the frequency of the corresponding classical oscillator.120) 2 is non-zero [since dif ≡ (dx )n.e. (9. according to Eq.e. In principle. according to Eq. This is an important h result. Suppose that the electron is initially in an excited state: i.57)]. the system can only make a spontaneous transition from an energy state corresponding to the quantum number n to one corresponding to the quantum number n ′ if the associated electric dipole moment ∞ (dx ) n. (13. Here. (13. Now.n ′ −1 . i→f (13. h (13. 5. (9. the unperturbed energy eigenvalues of the system are En = (n + 1/2) ¯ ω0 . 2 me ω0 (13. i→f 13. Let us estimate the typical value of the spontaneous emission rate for a hydrogen atom.115) that wspn ∼ α3 ωif . Hence. where a0 is the Bohr radius [see Eq.121) Since we are dealing with emission. the electron can decay to a lower energy state via the spontaneous emission of a photon of the appropriate frequency.n ′ = n|e x|n = e −∞ ′ ψn (x) x ψn ′ (x) dx (13. where E0 is the energy of the groundh state [see Eq. 13.190 QUANTUM MECHANICS respectively..10 Radiation from a Harmonic Oscillator Consider an electron in a one-dimensional harmonic oscillator potential aligned along the x-axis. · · ·.117).8. since our perturbation expansion is based on the assumption that the transition rate between different energy eigenstates is much slower than the frequency of phase oscillation of these states: i. ∞ ψn x ψn ′ dx = −∞ √ √ ¯ h n δn. that wspn ≪ ωif (see Sect. n > 0.n ′ for the case in hand]. We expect the dipole moment dif to be of order e a0. We also expect ωif to be of order |E0 |/¯ .122) .n ′ = e ¯n h δn. 1. the quantum number n takes the values 0.2).58)]. we must have n > n ′ . This is indeed the case. (5.n ′ +1 + n ′ δn. we obtain (dx )n. It thus follows from Eq. Let the ψn (x) be the (real) properly normalized unperturbed eigenstates of the system. 2.115).n ′ +1 . According to Sect.. 2 me ω0 (13.

The only difference is the factor (1/2) ¯ ω0 in Eq. 9. l.124) (13. According to Eqs. the emitted photon always has the same frequency as the classical oscillator. Since the perturbing Hamiltonian (13.106) and (13.n−1 = En − En−1 = ω0 . according to Sect.n−1 (dx )n. m ′ |2 + | n. no matter which state decays. the decay rate of the nth excited state is given by wn = It follows that wn = The mean radiated power is simply Pn = ¯ ω0 wn = h 2 e 2 ω0 [En − (1/2) ¯ ω0 ].Time-Dependent Perturbation Theory 191 It is clear that (in the electric dipole approximation) we can only have spontaneous emission between states whose quantum numbers differ by unity. m to one corresponding to the quantum numbers n ′ . 13. 6π ǫ0 me c3 (13.126) Classically.77) does not contain any spin operators. (13.125) (13. l. m ′ if the modulus squared of the associated electric dipole moment d2 = | n. Thus. m ′ |2 + | n. l. a hydrogen atom can only make a spontaneous transition from an energy state corresponding to the quantum numbers n.115).123) Hence. According to Eq. and m. m|e x|n ′ . (13. m|e y|n ′ .126)—this is needed to ensure that the ground-state of h the quantum oscillator does not radiate. we conclude that.n−1 . l ′ . h 6π ǫ0 me c3 3 2 ωn. It can be seen that a quantum oscillator radiates in an almost exactly analogous manner to the equivalent classical oscillator. l ′ . l ′ . Thus. the frequency of the photon emitted when the nth excited state decays is ωn. l ′ . an electron in a one-dimensional oscillator potential radiates at the oscillation frequency ω0 with the mean power P= 2 e 2 ω0 E. m|e z|n ′ . l. 6π ǫ0 me c3 (13. the various energy eigenstates of the hydrogen atom are labeled by the familiar quantum numbers n.127) where E is the oscillator energy. ¯ h (13. we can neglect electron spin in our analysis. 3π ǫ0 ¯ c3 h 2 n e 2 ω0 .11 Selection Rules Let us now consider spontaneous transitions between the different energy levels of a hydrogen atom. m ′ |2 (13. l. (13.115).128) .4.

133) and (13. l. Hence. l ′ . 12. φ).l (r) Yl. m|x− |n ′ . l ′ . l. the excited state must have the quantum numbers l = 1 and m = 0. x+ ] − ¯ x+ |n ′ .135) 2P → 1S Transitions in Hydrogen .e.129) It is easily demonstrated that [Lz . Thus.5. l. m|x− |n ′ . l ′ . l ′ . m ′ is zero unless m ′ = m − 1. l ′ .133) (13. x− ] + ¯ x− |n ′ . l. h Hence. in Sect. ±1.m (r. m ′ is only non-zero provided that m ′ = m and l ′ = l ± 1. m ± 1. m|x− |n ′ .130) Clearly. l. m ′ and n. m ′ are obviously both zero if n. m|y|n ′ . Now the ground-state is characterized by l ′ = m ′ = 0. that since the perturbing Hamiltonian does not contain any spin operators. l ′ . m ′ and n. (13. Now. θ.. m ′ and n. we have the additional selection rule that ms = ms .134) m ′ = m.. l. in order to satisfy the selection rules (13.l. l. we are dealing with a spontaneous transition from a 2P to a 1S state. m ′ is zero unless m ′ = m + 1. m|y|n ′ . m|[Lz. m|[Lz. m|x|n ′ . The above arguments demonstrate that spontaneous transitions between different energy levels of a hydrogen atom are only possible provided l ′ = l ± 1.131) (13. the wavefunction of a hydrogen atom takes the form ψn..4.134). m ′ are only non-zero if m ′ = m ± 1. l.e. m|x+ |n ′ . n = 2) and the ground-state (i. (13.12 Let us calculate the rate of spontaneous emission between the ﬁrst excited state (i. l ′. m|y|n ′ . n ′ = 1) of a hydrogen atom. m ′ = ¯ (m − m ′ − 1) n. l ′ . l ′ . we conclude that n. transitions calculated using the electric dipole approximation). It turns out that the proof that this matrix element is zero unless l ′ = l ± 1 can. m ′ are both zero. m|x+ |n ′ . 9. (13. incidentally. m|x|n ′ . we have already seen. the spin quantum number ms cannot change ′ during a transition. According to Sect.132) (13. l ′ . m ′ = ¯ (m − m ′ + 1) n. Hence. ﬁnally. m ′ are also zero unless l ′ = l±1. l ′ . Now. l ′ . l ′ . m|x|n ′ . l. These are termed the selection rules for electric dipole transitions (i. n. m ′ = 0. l. l ′ . that a spontaneous transition from a 2S to a 1S state is forbidden by our selection rules. m|x+|n ′ . m ′ = 0. via a trivial modiﬁcation. l. n. Hence. l ′ . l. l. Note. φ) = Rn. h h (13. Note. that the matrix element n.m (θ. m ′ and n. l.e. 13. l. h h and n.192 QUANTUM MECHANICS is non-zero. and n. Consider x± = x + i y. x± ] = ± ¯ x± . m|z|n ′ . also be used to demonstrate that n. n.

but tedious. the 2P → 1S transition rate is written 1. ±1 = i 5 a0 . 1.140) ¯ ω = E0 /4 − E0 = − E0 = 10.2 eV.58). 8. (13.141) which reduces to w2P→1S = 2 me c2 = 6. it follows from the energy-time uncertainty relation that the energy of this state is uncertain by an amount ¯ h ∼ 4 × 10−7 eV.137) 3 √ 27 1. 0. (13. 0|z|2. or −1. If follows from Eq. the mean life-time of a hydrogen 2P state is τ2P = (w2P→1S )−1 = 1. l. the energy of the eigenstate of the hydrogen atom characterized by the quantum numbers n. 0.138) 2 5 a0 . Here. 1.136) 35 27 1. α = 1/137 is the ﬁne-structure constant. and the spherical harmonics Yl. (13. the energy of the photon emitted during a 2P → 1S transition is 3 (13. ±1 = ± w2P→1S 8 ω 3 d2 = .144) ∆E2P ∼ τ2P This uncertainty gives rise to a ﬁnite width of the spectral line associated with the 2P → 1S transition.145) λ ¯ω h .Time-Dependent Perturbation Theory 193 where the radial functions Rn. Now. where the ground-state energy E0 is speciﬁed in Eq.143) Incidentally. All of the other possible 2P → 1S matrix elements are zero because of the selection rules. This natural line-width is of order ∆λ ∆E2P ∼ ∼ 4 × 10−8 . Some straight-forward.7. (13. (13.6 ns. since the 2P state only has a ﬁnite life-time. Clearly. (9.4.27 × 108 s−1 (13.215 × 10−7 m. integration reveals that 27 a0 . the transition rate is independent of the quantum number m. (13. (9. (13. 9. 0 = (13.140). Finally. according to Eq.115). It turns out that this is a general result. (13. 1.l are given in Sect. Hence. 0.142) α5 3 ¯ h with the aid of Eqs.m are given in Sect. Hence.139) and (13. 3π ǫ0 ¯ c3 h (13. h 4 This corresponds to a wavelength of 1. 3 where a0 is the Bohr radius speciﬁed in Eq. m is E = E0 /n2 .139) d2 = 10 (e a0)2 3 for m = 0. 0|y|2.128) that the modulus squared of the dipole moment for the 2P → 1S transition takes the same value 215 (13. 1. 0|x|2.57).

there are four 2P3/2 states and only two 2P1/2 states.13 Intensity Rules Now. we know. since they have almost the same energies—and since they decay to the 1S state at the same rate. λ ¯ω h (13. we have seen that the transition rate is independent of spin.8 that a forbidden transition is one for which the matrix element f|ǫ · p|i is zero. 12. 13. 16 (13. However. and are thus called the 2P1/2 states. Expanding exp( i k·r).4 × 10−6 . The energy of the 2P3/2 states is slightly higher than that of the 2P1/2 states. and hence of the spin quantum number ms . Hence. this matrix element is only an approximation to the true matrix element for radiative transitions. that when we take electron spin and spin-orbit coupling into account the degeneracy of the six 2P states of the hydrogen atom is broken.145). so there really are two spectral lines.53 × 10−5 eV. (13. In fact. the matrix element for a forbidden transition becomes f|ǫ·p exp( i k·r)|i ≃ i f|(ǫ·p) (k·r)|i . (13. then the transition rate is plainly also independent of the total angular momentum quantum number j. The remaining two states are characterized by j = 1/2.194 QUANTUM MECHANICS 13. 13. How does all of this affect the rate of the 2P → 1S transition? Well.148) .133) and (13.147) Note that this splitting is much greater than the natural line-width estimated in Eq.8. However. these states are divided into two groups with slightly different energies. However. In fact. It follows that the transition rate is independent of the z-component of total angular momentum quantum number mj = m + ms . which takes the form f|ǫ·p exp( i k·r)|i .146) Thus. the energy difference is ∆E = − α2 E0 = 4. from Sect. If these states are equally populated—which we would certainly expect to be the case in thermal equilibrium.14 Forbidden Transitions Atomic transitions which are forbidden by the electric dipole selection rules (13. if this is the case. we expect the 2P3/2 → 1S and 2P1/2 → 1S transition rates to be the same. the wavelength of the spectral line associated with the 2P → 1S transition in hydrogen is split by a relative amount ∆E ∆λ = = 4. There are four states characterized by the overall angular momentum quantum number j = 3/2—these are called the 2P3/2 states.134) are unsurprisingly known as forbidden transitions. It is clear from the analysis in Sect. and is also independent of the quantum number m. and keeping the ﬁrst two terms. it stands to reason that the spectral line associated with the 2P3/2 → 1S transition is twice as bright as that associated with the 2P1/2 → 1S transition.

. it is fairly obvious that excited states which decay via forbidden transitions have much longer life-times than those which decay via electric dipole transitions.9. Such transitions are absolutely forbidden. we calculated that the typical rate of an electric dipole transition is wi→f ∼ α3 ωif . and k as the wavenumber of a typical spectral line of hydrogen. there are some transitions (in particular.150) for such a transition. it is clear that the transition rate for a forbidden transition enabled by the residual matrix element (13. Estimating r as the Bohr radius. allbeit at a much reduced rate. the 2S → 1S transition) for which the true radiative matrix element f|ǫ· p exp( i k· r)|i is zero. 13. in Sect. (13. if the residual matrix element on the right-hand side of the above expression is non-zero then a “forbidden” transition can take place. it is easily demonstrated that wi→f ∼ α5 ωif (13.Time-Dependent Perturbation Theory 195 Hence.148) is smaller than that of an electric dipole transition by a factor (k r)2 . Since the natural width of a spectral line is inversely proportional to the life-time of the associated decaying state. it follows that spectral lines associated with forbidden transitions are generally much sharper than those associated with electric dipole transitions. Finally. Of course. In fact.149) Since the transition rate is proportional to the square of the radiative matrix element.

196 QUANTUM MECHANICS .

.3) cn ψn .2 Variational Principle Suppose that we wish to solve the time-independent Schr¨dinger equation o H ψ = E ψ. E0 ≤ ψ|H|ψ .5) If our trial wavefunction ψ is properly normalized then we can write ψ= n (14. that the ground-state energy.4. (14.. (14.. E0 . quite simply. In such systems. The variational principle states. Unfortunately.1) where H is a known (presumably complicated) time-independent Hamiltonian. Let ψ be a normalized trial solution to the above equation.1 Introduction We have seen.2) Thus. ψ1 the ﬁrst excited state. (14.Variational Methods 197 14 Variational Methods 14. Most of the methods which have been developed for ﬁnding such solutions employ the so-called variational principle discussed below. etc. 9.4) so that ψ0 is the ground-state. or for o molecules. ψn |ψm = δnm . 14. that we can solve Schr¨dinger’s equation exactly to ﬁnd the o stationary eigenstates of a hydrogen atom. is always less than or equal to the expectation value of H calculated with the trial wavefunction: i. we can obtain an approximation to the wavefunction and energy of the ground-state. Suppose that the ψn and the En are the true eigenstates and eigenvalues of H: i. it is not possible to ﬁnd exact solutions of Schr¨dinger’s equation for atoms more complicated than hydrogen. (14. H ψn = En ψn .e. in Sect. Let us prove the variational principle. by varying ψ until the expectation value of H is minimized. the best that we can do is to ﬁnd approximate solutions.e. let E0 < E1 < E2 < · · · . The ψn are assumed to be orthonormal: i.6) . (14. Furthermore.e.

11) Now.12) ˜ Suppose that we have found a good approximation. (14.8) where use has been made of Eqs. the second term on the right-hand side of the above expression is positive deﬁnite. so that any approximations to highly excited states are unlikely to be very accurate. (14.10) Combining the previous two equations. n QUANTUM MECHANICS (14.5)..m ∗ cn cm ψn |H|ψm = n ∗ cn cm Em ψn |ψm = n En |cn | 2 . (14. since En − E0 > 0 for all n > 0 [see (14.7) Now. by repeating the above analysis. ψ|ψ0 = 0) then. by varying ψ until the expectation value of H is minimized. that the errors are clearly cumulative in this method. So. we obtain ψ|H|ψ = E0 + n>0 |cn | 2 (En − E0 ). we obtain the desired result ψ|H|ψ ≥ E0 .198 where |cn | 2 = 1. to the ground-state wavefunc˜ ˜ tion. (14.e. (14. Note.4)]. we can obtain an approximation to the wavefunction and energy of the ﬁrst excited state.9) However. we can easily demonstrate that ψ|H|ψ ≥ E1 . calculated with ψ. . the variational method is generally only used to calculate the ground-state and ﬁrst few excited states of complicated quantum systems. (14. however. Obviously.3) and (14. we can continue this process until we have approximations to all of the stationary eigenstates.13) Thus. ψ0 . takes the form ψ|H|ψ = n cn ψn H m cm ψm = n. Hence.7) can be rearranged to give |c0 | 2 = 1 − n>0 |cn | 2 . For this reason. the expectation value of H. If ψ is a normalized trial wavefunction which is orthogonal to ψ0 (i. Eq. we can write ψ|H|ψ = |c0 | 2 E0 + n>0 |cn | 2 En . (14. (14.

and let the position vectors of the two electrons be r1 and r2 .2 ψ1. (14.2 .19) is the Schr¨dinger equation of a hydrogen atom whose nuclear o charge is +2 e. where E = E1 + E2 . The terms in the above expression represent the kinetic energy of the ﬁrst electron. Indeed. (14.2 = E1. Let us attempt to calculate its ground-state energy.15) (14.23) .Variational Methods 199 14. where H1. respectively..3 Helium Atom A helium atom consists of a nucleus of charge +2 e surrounded by two electrons. Eq. the electrostatic attraction between the nucleus and the second electron. if this term is neglected then we can write H = H1 + H2 . 9. It follows. the kinetic energy of the second electron. (14. Schr¨dinger’s equation o Hψ = Eψ reduces to H1. ψ(r1 . instead of +e. (14.2 ψ1. the electrostatic attraction between the nucleus and the ﬁrst electron. Let the nucleus lie at the origin of our coordinate system.e. In this case. respectively. It is the ﬁnal term which causes all of the difﬁculties.22) 2r .2 (14. a0 exp − (14. that if both electrons are in their lowest energy states then ψ1 (r1 ) = ψ0 (r1 ).21) ψ2 (r2 ) = ψ0 (r2 ). and the electrostatic repulsion between the two electrons. r2) = ψ1 (r1 ) ψ2 (r2 ).18) (14. 2 me 1.4 (making the substitution e2 → 2 e2 ).19) (14. we would expect the wavefunction to be separable: i. Hence.16) In other words.17) (14. the Hamiltonian just becomes the sum of separate Hamiltonians for each electron. The Hamiltonian of the system thus takes the form ¯2 h e2 2 2 1 2 2 H=− ∇1 + ∇2 − . from Sect.20) Of course.2 4π ǫ0 r1. where ψ0 (r) = √ 4 3/2 2 π a0 (14.2 ¯2 h 2 e2 2 =− ∇ − .14) + − 2 me 4π ǫ0 r1 r2 |r2 − r1 | where we have neglected any reduced mass effects.

(9.30) Vee = − 2 2 2 π r + r − 2 r1 r2 cos θ 1 2 where θ is the angle subtended between vectors r1 and r2 . our crude estimate for the ground-state energy of helium becomes E = 4 E0 + 4 E0 = 8 E0 = −108. In reality.27) yields an upper bound on the groundstate energy. This fact demonstrates that the neglected electron-electron repulsion term makes a large contribution to the helium ground-state energy.31) .26) and (14. the above expression can r also be written e−2 (r1 +r2 ) 4 E0 d3 r1 d3 r2 . (14.98 eV. (9. a0 is the Bohr radius [see Eq. r2)| 2 3 d r1 d3 r2 . r2) = ψ0 (r1 ) ψ0 (r2 ) = 8 2 [r1 + r2 ] exp − . (9. (14.58)].8 eV.29) where ^1. ψ(r1 .57).28) that Vee = − 4 E0 π2 r r e−2 (^1 +^2 ) 3 d ^1 d3^2 .6 eV is the hydrogen ground-state energy [see Eq. Note that ψ0 is properly normalized. Thus.28) (14. which is −78.24) where E0 = −13.26) The expectation value of the Hamiltonian (14. (14. we hope that it will give a reasonably accurate estimate of this energy. for the sake of clarity. Neglecting the hats. |r2 − r1 | (14.2 /a0 .2 = 2 r1. where Vee = ψ e2 e2 ψ = 4π ǫ0 |r2 − r1 | 4π ǫ0 |ψ(r1 . (14. this estimate is signiﬁcantly different from the experimentally determined value. (14. we can use the variational principle to estimate this contribution. E1 = E2 = 4 E0 . If we perform the integral in r1 space before that in r2 space then Vee = − 4 E0 π2 e−2 r2 I(r2 ) d3r2 . r r |^1 − ^2 | r r (14. It follows from Eqs. 3 π a0 a0 (14.27) The variation principle only guarantees that (14.14) thus becomes H = 8 E0 + Vee . Fortunately. Let us employ the separable wavefunction discussed above as our trial solution. Thus.25) Unfortunately.200 QUANTUM MECHANICS Here. Furthermore. however.57)].

37) e−β x x dx = − 2 (1 + β x). (14. (14. 1 −1 2 2 r1 + r2 − 2 r1 r2 µ dµ = 2 2 r1 r2 r1 + r2 − 2 r1 r2 µ +1 −1 = = giving 1 I(r2 ) = 4π r2 But.33) which trivially reduces to ∞ π 0 2 r1 I(r2 ) = 2π 0 e−2 r1 + 2 r2 − 2 r1 r2 cos θ1 2 r1 dr1 sin θ1 dθ1 . e −β x r2 (r1 + r2 ) − |r1 − r2 | r1 r2 2/r1 2/r2 ∞ for r1 > r2 . θ1 . Hence. (14. ∞ π 0 2π 0 2 r1 I(r2 ) = 0 e−2 r1 + 2 r2 − 2 r1 r2 cos θ1 2 r1 dr1 sin θ1 dθ1 dφ1 . for r1 < r2 (14. β3 (14.40) .32) Our ﬁrst task is to evaluate the function I(r2 ).35) Now. It follows that θ = θ1 . (14. β e−β x (2 + 2 β x + β2 x2 ).39) e−β x x2 dx = − yielding I(r2 ) = π 1 − e−2 r2 (1 + r2 ) . Let (r1 .36) e 0 −2 r1 2 r1 dr1 + r2 e−2 r1 r1 dr1 .Variational Methods where I(r2 ) = e−2 r1 2 2 r1 + r2 − 2 r1 r2 cos θ 201 d3 r1 .38) (14. φ1 ) be a set of spherical polar coordinates in r1 space whose axis of symmetry runs in the direction of r2 . (14. r2 (14. we can see that π 0 2 r1 1 + 2 r2 1 − 2 r1 r2 cos θ1 sin θ1 dθ1 = −1 dµ 2 r1 + 2 r2 − 2 r1 r2 µ .34) Making the substitution µ = cos θ1 .

2 2 (14.42) Hence. from (14. and vice versa. (14. 3 π a0 a0 (14. We can actually reﬁne our estimate further.45) ¯2 h Z e2 2 H1. r1 r2 (14.8 eV. In reality. and then minimize the result with respect to Z.49) e2 4π ǫ0 [Z − 2] [Z − 2] + .202 QUANTUM MECHANICS Since the function I(r2 ) only depends on the magnitude of r2 .27). Let us recalculate the ground-state energy of helium as a function of Z.14) for the Hamiltonian of the helium atom in the form H = H1 (Z) + H2 (Z) + Vee + U(Z).46) (14. using the above trial wavefunction. this should give us an even better estimate for the ground-state energy. According to the variational principle.43) This is remarkably close to the correct result. the integral (14. (14.41) π 0 which yields ∞ Vee = −16 E0 0 5 e−2 r2 1 − e−2 r2 (1 + r2 ) r2 dr2 = − E0 . We can rewrite the expression (14. and U(Z) = It follows that H (Z) = 2 E0 (Z) + Vee (Z) + U (Z).44) where Z < 2 is effective nuclear charge number seen by each electron. The trial wavefunction (14. (14.47) is the electron-electron repulsion term. Vee = e2 1 4π ǫ0 |r2 − r1 | where (14. Hence.31) reduces to 16 E0 ∞ −2 r2 2 Vee = − e I(r2 ) r2 dr2 . a better trial wavefunction might be ψ(r1 .2 is the Hamiltonian of a hydrogen atom with nuclear charge +Z e. our estimate for the ground-state energy of helium is H = 8 E0 − 11 5 E0 = E0 = −74.2 (Z) = − ∇ − 2 me 1.26) essentially treats the two electrons as non-interacting particles.48) . r2) = Z3 Z [r1 + r2 ] exp − . 2 (14.2 4π ǫ0 r1. we would expect one electron to partially shield the nuclear charge from the other.

50) Here. Our spatial wavefunction (14.69. 6).54) 27 = 1.44) is obviously symmetric with respect to exchange of particles.44) [actually. Collecting the various terms. and making the substitution a0 → a0 /Z] that 1 Z = . This means that the spinor must be anti-symmetric. all we need to do is to make the substitution a0 → (2/Z) a0 ].74) [with n = 1. (14. Vee (Z) = −(5 Z/4) E0 is the value of the electron-electron repulsion term when recalculated with the wavefunction (14. (14.55) 16 The fact that Z < 2 conﬁrms our earlier conjecture that the electrons partially shield the nuclear charge from one another.53) The value of Z which minimizes this expression is the root of dH 27 E0 = 0. we could get even closer to the correct value of the helium ground-state energy by using a more complicated trial wavefunction with more adjustable parameters.52) since E0 = −e2 /(8π ǫ0 a0 ). ﬁnally.51) r a0 Hence. Note. (14.69) = 2 2 Z= This is clearly an improvement on our previous estimate (14. Obviously.5 eV. Our new estimate for the ground-state energy of helium is 6 1 3 E0 = −77.43) [recall that the correct result is −78. U (Z) = −4 Z (Z − 2) E0 . 1/r is the expectation value of 1/r calculated for a hydrogen atom with nuclear charge +Z e. and U (Z) = 2 (Z − 2) e2 4π ǫ0 1 .98 eV]. our new expression for the expectation value of the Hamiltonian becomes H (Z) = 2 Z2 − 5 27 Z − 4 Z (Z − 2) E0 = −2 Z2 + Z E0 . r (14.Variational Methods 203 where E0 (Z) = Z2 E0 is the ground-state energy of a hydrogen atom with nuclear charge +Z e. = −4 Z + dZ 4 It follows that (14. It follows from Eq. that since the two electrons in a helium atom are indistinguishable fermions. . Now.56) H (1. (9. the overall wavefunction is the product of the spatial wavefunction and the spinor representing the spin-state. the overall wavefunction must be anti-symmetric with respect to exchange of particles (see Sect. 4 4 (14. (14.

Let us investigate whether or not this molecule possesses a bound state: i. two electrons) is anti-symmetric with respect to interchange of particles then the system is in the so-called singlet state with overall spin zero. whether or not it possesses a ground-state whose energy is less than that of a hydrogen atom and a free proton. from Sect. Hence. that if the spin-state of an l = 0 system consisting of two spin one-half particles (i.4.204 z-axis QUANTUM MECHANICS proton z=R r2 electron r1 z=0 proton Figure 14. 14. 14.e.1).4 Hydrogen Molecule Ion The hydrogen molecule ion consists of an electron orbiting about two protons.57) . Suppose that the two protons are separated by a distance R..e. and is the simplest imaginable molecule. and the second at z = R (see Fig.. Let us try ψ(r)± = A [ψ0 (r1 ) ± ψ0 (r2 )] (14. the ground-state of helium has overall electron spin zero.1: The hydrogen molecule ion. According to the variation principle. we can deduce that the H+ ion has a bound state if we can ﬁnd any trial wavefunction for which the total 2 Hamiltonian of the system has an expectation value less than that of a hydrogen atom and a free proton. since the electron moves far more rapidly than the protons. In the following. let them lie on the z-axis. with the ﬁrst at the origin. we shall treat the protons as essentially stationary. In fact. 11. It is clear. This is reasonable.

1 e−(x+X) (1 + x + X) − e−|x−X| (1 + |x − X|) . Note. θ. that the wavefunction respects the obvious symmetries in the problem. this is a very simplistic wavefunction. and r1. Hence. it is easily seen that r1 = r and r2 = (r2 + R2 − 2 r R cos θ)1/2 . ∞ π J=2 0 0 exp −x − (x2 + X2 − 2 x X cos θ)1/2 x2 dx sin θ dθ. (14. we have already performed the trivial φ integral.Variational Methods as our trial wavefunction.58) is a normalized hydrogen ground-state wavefunction centered on the origin. Here. ψ0 (r1 ) ψ0 (r2 ) d3 r. Our ﬁrst task is to normalize our trial wavefunction.57). It follows that d(y2 ) = 2 y dy = 2 x X sin θ dθ.65) .2 are the position vectors of the electron with respect to each of the protons (see Fig. φ). Obviously. however. We require that |ψ± |2 d3 r = 1. since it is just a linear combination of hydrogen ground-state wavefunctions centered on each proton. giving π e (x 0 2 +X2 −2 x X cos θ)1/2 sin θ dθ = 1 xX x+X e−y y dy |x−X| (14.62) |ψ0 (r1 )|2 + |ψ0 (r2 )|2 ± 2 ψ0 (r1 ) ψ(r2) d3 r. where ψ0 (r) = √ 1 3/2 π a0 205 e−r/a0 (14.63) where X = R/a0 .59) Let us employ the standard spherical polar coordinates (r.64) = − Thus. A = I−1/2 . Hence.1).60) (14. xX 2 −X X −2 x J = − e e (1 + X + x) − (1 + X − x) x dx X 0 2 ∞ −2 x −X e (1 + X + x) − eX (1 − X + x) x dx. 14. Let y = (x2 + X2 − 2 x X cos θ)1/2 . (14.61) (14. from (14. Now. e − X X (14. (14. where I= It follows that with J= I = 2 (1 ± J).

70) ψ0 (r1 ) a0 ψ0 (r1 ) . r1 r2 X3 . ψ0 (r1 ) ψ0 (r2 ) .68) since ψ0 (r1.72) D=2 0 e−2 x x2 dx sin θ dθ. X (14. that − ¯2 h e2 ∇2 − 2 me 4π ǫ0 1 1 + .2 (14.73) which reduces to 4 D= X giving e 0 −2 x x dx + 4 X 2 e−2 x x dx.71) (14. r1 (14. (14. ∇2 − 2 me 4π ǫ0 r1.75) E=2 0 0 exp −x − (x2 + X2 − 2 x X cos θ)1/2 x dx sin θ dθ. ∞ π 1 1 − [1 + X] e−2 X . 3 QUANTUM MECHANICS (14. (14.66) (14.76) . the Hamiltonian of the electron is written H=− Note.74) D= Furthermore.206 which evaluates to J = e−X 1 + X + Now.69) = E0 ψ − A Hence. ± r2 r1 H = E0 + 4 A2 (D ± E) E0 . however. ∞ π 0 (14.2 ). (x2 + X2 − 2 x X cos θ)1/2 X ∞ (14.67) ¯2 h e2 ψ0 (r1.2 ) are hydrogen ground-state wavefunctions. r2 a0 ψ0 (r1 ) ψ0 (r2 ) . where D = E = Now.2 ) = E0 ψ0 (r1. It follows that H ψ± = A − ¯2 h e2 ∇2 − 2 me 4π ǫ0 e2 4π ǫ0 1 1 + r1 r2 [ψ0 (r1 ) ± ψ0 (r2 )] (14.

5 a0 = 1. (14. The experimentally determined values are R = 1. for the H+ ion are R = 2. Recall that in order for the H+ ion to be in a bound state it must have a lower energy than a hydrogen atom and a 2 free proton: i.78) (14. we can write Etotal = H + Etotal = −F± (R/a0 ) E0 . Etotal < E0 .e. the existence of a bound state of the H+ ion.8 eV. Clearly. and the binding energy.79) where J. our estimates for the separation between the two protons. and F± (X) = −1 + 2 (1 + X) e−2 X ± (1 − 2 X2 /3) e−X . our estimates are not particularly accurate.13 E0 = −1. Clearly. the odd wavefunction does exactly the opposite. However. Thus.81) The functions F+ (X) and F− (X) are both plotted in Fig. Ebind = Etotal − E0 = −(F+ + 1) E0 . 14. Our expression for the expectation value of the electron Hamiltonian is H = 1+2 (D ± E) E0 . whereas the odd trial wavefunction ψ− does not [see Eq. (14.80) (14. This minimum occurs when X ≃ 2.66). The binding energy of the H+ ion is deﬁned as the difference 2 between its energy and that of a hydrogen atom and a free proton: i. since the even wavefunction maximizes the electron probability density between the two protons. (14.2. In order to obtain the total energy of the molecule. and E are speciﬁed as functions of X = R/a0 in Eqs. 4π ǫ0 R X 2 since E0 = −e /(8π ǫ0 a0 ).78). respectively. (14. On the other hand.13. respectively.Variational Methods which reduces to E = − − yielding E = (1 + X) e−X . (1 ± J) 2 −X e X 2 X ∞ X X 207 e−2 x (1 + X + x) − (1 + X − x) dx 0 e−2 x e−X (1 + X + x) − eX (1 − X + x) dx. It follows from Eq. our calculation does establish.75). 14.5 and F+ ≃ −1. we must add to this the potential energy of the two protons. This is hardly surprising.83) According to the variational principle. X 1 ± (1 + X + X2 /3) e−X (14.33 × 10−10 m 2 and Ebind = 0.81) that a bound state corresponds to F± < −1. the even trial wavefunction ψ+ possesses a bound state. respectively.2..57)].77 eV.82) (14. (14. and Ebind = −2. thereby reducing their mutual electrostatic repulsion. where E0 is the hydrogen ground-state energy.06 × 10−10 m.. the binding energy is less than or equal to the minimum binding energy which can be inferred from Fig.77) (14. and (14. 2 . (14. which is all that we set out to achieve.e. Hence. beyond any doubt. e2 2 = H − E0 . D. Thus.

.2: The functions F+ (X) (solid curve) and F− (X) (dashed curve).208 QUANTUM MECHANICS Figure 14.

4) where E = ¯ 2 k2 /2 m is the particle energy. Firstly. H0 ψ0 = E ψ0 . Sect. of number density n. We have already developed theories which account for some aspects of the spectrum of hydrogen. (15. data regarding quantum phenomena has been obtained from two main sources.Scattering Theory 209 15 Scattering Theory 15. (∇2 + k2 ) u(r) = ρ(r) is known as the Helmholtz equation. . and V(r) the scattering potential. energy conserving scattering in which the Hamiltonian of the system is written H = H0 + V(r). and velocity v = ¯ k/m. 4π |r − r ′ | ′ (15. Schr¨dinger’s equation for the scattering h o problem is (H0 + V) ψ = E ψ.2) H0 = 2m 2m is the Hamiltonian of a free particle of mass m. This potential is assumed to only be non-zero in a fairly localized region close to the origin. secondly. 11.1 Introduction Historically. Let us now examine the quantum theory of scattering. and hydrogen-like. atoms.7) 1 e i k |r−r | ρ(r ′) d3 r ′ . and.2 Fundamentals Consider time-independent. (15.6) (15.3) ψ0 (r) = n e i k·r represent an incident beam of particles. ¯2 h (15. The above equation can be rearranged to give (∇2 + k2 ) ψ = Now. (15.1) ¯2 2 h p2 ≡− ∇ (15. Let √ (15.5) subject to the boundary condition ψ → ψ0 as V → 0. The solution to this equation is well-known: u(r) = u0 (r) − 1 where 2m V ψ. from the study of spectroscopic lines. 15. Of h course. from scattering experiments.8) See Grifﬁths.4.

9) to ﬁrst-order in r ′ /r.14) where v = ¯ k/m is the velocity of the incident particles. h dσ r2 dΩ |j ′ | dΩ = . r2 (15.. whereas the second term represents an outgoing spherical wave of scattered particles.12) The ﬁrst term on the right-hand side of Eq. if r ≫ r ′ then |r − r ′ | ≃ r − ^ · r ′ r (15. r ′ ′ Let us calculate the value of the wavefunction ψ(r) well outside the scattering region. k ′) . k ′) = − √ n e i k·r + eikr f(k. Hence. the particle ﬂux) associated with a wavefunction ψ is ¯ h Im(ψ∗ ∇ψ). the particle ﬂux associated with the incident wavefunction ψ0 is j = n v. dΩ (15.6) can be inverted. dΩ |j| which yields dσ = |f(k. Likewise. u0 (r) is any solution of (∇2 + k2 ) u0 = 0. k ′)|2 .10) (15. k ′ )|2 ′ v. Equation (15. Eq.210 QUANTUM MECHANICS Here.15) where v ′ = ¯ k ′ /m is the velocity of the scattered particles. Now.e.16) . to give 2m ψ(r) = ψ0 (r) − 2 ¯ h e i k |r−r | V(r ′ ) ψ(r ′) d3 r ′ . Now. ′| 4π |r − r ′ (15. the particle ﬂux h associated with the scattered wavefunction ψ − ψ0 is j′ = n |f(k. (15. subject to the boundary condition ψ → ψ0 as V → 0.. k ′ = k) which propagate from the scattering region to the observation point. (15.2. (15.11) m √ 2π n ¯ 2 h e−i k ·r V(r ′ ) ψ(r ′) d3 r ′ .11) represents the incident particle beam. where ^/r is a unit vector which points from the scattering region to r the observation point.9) reduces to ψ(r) ≃ where f(k. divided by the incident particle ﬂux.e. It is helpful to deﬁne k ′ = k ^. This is the wavevector for particles r with the same energy as the incoming particles (i.13) j= m Thus. (15. The differential scattering cross-section dσ/dΩ is deﬁned as the number of particles per unit time scattered into an element of solid angle dΩ.17) (15. 7. the probability ﬂux (i. From Sect. (15.

(15. It follows from Eq. as yet. 15. In other words. (15. Thus. and 1/µ measures the “range” of the potential.3 Born Approximation Equation (15.12). This procedure is called the Born approximation.18) Thus.17) is not particularly useful. (15. k ′) depends on the.20) that 2 m V0 1 f(θ) = − 2 .Scattering Theory 211 Thus. k ′) is proportional to the Fourier transform of the scattering potential V(r) with respect to the wavevector q = k − k ′ . For a spherically symmetric potential. unknown wavefunction ψ(r) [see Eq. µr (15.21) where V0 is a constant. ψ0 (r). ψ(r). f(k. It is easily demonstrated that q ≡ |k − k ′ | = 2 k sin(θ/2). Consider scattering by a Yukawa potential V(r) = V0 exp(−µ r) .22) Note that f(k ′ . (15. as it stands.23) 2 + µ2 ¯ µ q h . f(k ′ . we can obtain an expression for f(k. (15. k) is just a function of q for a spherically symmetric potential. it is reasonable to suppose that the total wavefunction. Note that the scattering conserves energy. ′ ′ (15. because the quantity f(k. k ′ )|2 gives the differential cross-section for particles with incident velocity v = ¯ k/m to be scattered such that their ﬁnal velocities are directed into a range of solid h angles dΩ about v ′ = ¯ k ′ /m.12)]. k ′) by making the substitution √ ψ(r) → ψ0 (r) = n exp( i k · r) in Eq. In this case. k) ≃ − giving 2m f(k . θ is the scattering angle. (15. Suppose. Recall that the vectors k and k ′ have the same length.19) r ′ V(r ′ ) sin(q r ′ ) dr ′ . k) ≃ − 2 ¯ q h ′ ∞ m 2π ¯ 2 h exp( i q r ′ cos θ ′ ) V(r ′ ) r ′ 2 dr ′ sin θ ′ dθ ′ dφ ′ .20) where θ is the angle subtended between the vectors k and k ′ . 0 (15. |f(k. k ′) ≃ m 2π ¯ 2 h e i (k−k )·r V(r ′ ) d3 r ′ . The Born approximation yields f(k. that the scattering is not particularly strong. so that |v ′ | = |v| h ′ and |k | = |k|. via energy conservation. however. does not differ substantially from the incident wavefunction.

implying that the Born approximation is more accurate at high incident particle energies. Thus.26) exp(−µ r ′ ) sin(q r ′ ) dr ′ = The Yukawa potential reduces to the familiar Coulomb potential as µ → 0. (15.7. Eq. the differential cross-section for scattering by a Yukawa potential is 2 dσ 2 m V0 1 ≃ .29) yields 2 m |V0 | ≪ 1.212 since ∞ QUANTUM MECHANICS q . giving 2 m |V0 | ≪1 (15. (15.9). The Born approximation is valid provided that ψ(r) is not too different from ψ0 (r) in the scattering region. Eq. in the Born approximation.28) is the famous Rutherford scattering cross-section formula. It follows. (15. 2 r′ 2π ¯ h Consider the special case of the Yukawa potential. k ≪ µ) we can replace exp( i k r ′) by unity. sin (θ/2) 4 (15. . In the high-k limit.27) Recall that ¯ k is equivalent to |p|. The condition for the Yukawa potential to develop a bound state is 2 m |V0 | ≥ 2.29) V(r ′ ) d3 r ′ ≪ 1.. (15.e. so the above equation can be rewritten h dσ ≃ dΩ Z Z ′ e2 16π ǫ0 E 2 1 . At low energies. (15.30) ¯ 2 µ2 h as the condition for the validity of the Born approximation. provided that V0 /µ → Z Z ′ e2 /4π ǫ0 . Of course.24) q2 + µ2 0 Thus.31) where V0 is negative. 16 k4 sin4 (θ/2) (15.28) where E = p2 /2 m is the kinetic energy of the incident particles. if the potential is strong enough to form a bound state then the Born approximation is likely to break down.32) This inequality becomes progressively easier to satisfy as k increases. (15.25) 2 dΩ [2 k2 (1 − cos θ) + µ2] 2 ¯ µ h given that q2 = 4 k2 sin2 (θ/2) = 2 k2 (1 − cos θ). from Eq. ¯ 2 µ2 h (15. the Born differential cross-section becomes dσ ≃ dΩ 2 m Z Z ′ e2 4π ǫ0 ¯ 2 h 2 1 . that the condition for ψ(r) ≃ ψ0 (r) in the vicinity of r = 0 is exp( i k r ′) m (15. (i. In this limit. ¯2 µk h (15.

ϕ). 2l+1 (15. The direction of k ′ is speciﬁed by the polar angle θ (i.. the angle subtended between the two wavevectors).37) can be combined to give d2 Rl dRl +2r + [k2 r2 − l (l + 1)]Rl = 0. (15. jl (k r) and yl (k r).36) What is the most general solution to this equation in spherical polar coordinates which does not depend on the azimuthal angle φ? Separation of variables yields ψ(r.35) (15.e. Recall that r2 jl (z) = zl − yl (z) = −z l 1 d z dz 1 d − z dz l sin z . V(r) = V(r)] the scattering amplitude is a function of θ only: i.40) (15.3.. Equations (15. both ψ0 (r) and ψ(r) satisfy the free space Schr¨dinger equation o (∇2 + k2 ) ψ = 0. (15. via Pl (cos θ) = 4π Yl.39) 2 dr dr The two independent solutions to this equation are the spherical Bessel functions. and an azimuthal angle φ about the z-axis.e. The Legendre functions are related to the spherical harmonics. z l (15. Outside the range of the scattering potential. introduced in Cha.33) (15. points in a different direction. φ) = f(θ). in general.20) and (15.Scattering Theory 213 15. introduced in Sect.. 8.34) depend on the azimuthal angle φ.21) strongly suggest that for a spherically symmetric scattering potential [i. The scattered wavefunction is characterized by a wavevector k ′ which has the same magnitude as k.38) Equations (15. z . f(θ.4 Partial Waves We can assume. ψ(r) = √ n exp( i k r cos θ) + exp( i k r) f(θ) .41) cos z . without loss of generality.37) since the Legendre functions Pl (cos θ) form a complete set in θ-space. nor the large-r form of the total wavefunction. √ √ ψ0 (r) = n exp( i k z) = n exp( i k r cos θ). θ) = l Rl (r) Pl(cos θ). but. (15.0 (θ. that the incident wavefunction is characterized by a wavevector k which is aligned parallel to the z-axis. 9. r (15.e.36) and (15. It follows that neither the incident wavefunction.

These waves are usually termed “partial waves”. Abramowitz and I. Note there are no yl (k r) functions in this expression. because they are not well-behaved as r → 0. (15.46) It is well-known that jl (y) = (−i)l 2 1 exp( i y µ) Pl(µ) dµ.214 QUANTUM MECHANICS Note that the jl (z) are well-behaved in the limit z → 0 . (Dover.14]. n + 1/2 (15.42) (15. 10. New York NY.1. Stegun.48) i l (2 l + 1) jl(k r) Pl (cos θ). 1. − Bl kr kr (15. where the al are constants. 1 Pn (µ) Pm (µ) dµ = −1 δnm . al = i l (2 l + 1). Eq. z jl (z) → l (15.A. 2. whereas the yl (z) become singular. (15. z cos(z − l π/2) yl (z) → − . Note that the yl (k r) functions are allowed to appear in this expansion. The asymptotic behaviour of these functions in the limit z → ∞ is sin(z − l π/2) . giving ψ0 (r) = √ n exp( i k r cos θ) = √ n l (15. The Legendre functions are orthonormal. (15.47) where l = 0. Thus.44) We can write exp( i k r cos θ) = al jl (k r) Pl (cos θ). because its region of validity does not include the origin.50) l where the Al and Bl are constants. the total wavefunction reduces to ψ(r) ≃ √ n l Al sin(k r − l π/2) cos(k r − l π/2) Pl (cos θ).51) . Handbook of mathematical functions.43) (15. −1 (15.45) so we can invert the above expansion to give 1 al jl (k r) = (l + 1/2) −1 exp( i k r µ) Pl(µ) dµ.49) The above expression tells us how to decompose the incident plane-wave into a series of spherical waves. 1965). · · · [see M. In the large-r limit. The most general expression for the total wavefunction outside the scattering region is √ ψ(r) = n [Al jl (k r) + Bl yl (k r)] Pl (cos θ).

2ikr (15. Eqs.e.53) which contains both incoming and outgoing spherical waves.53)–(15. the differential scattering cross-section dσ/dΩ is simply the modulus squared of the scattering amplitude f(θ) [see Eq. .43). (15.54) in the large-r limit. it is easily seen from Eqs. The total cross-section is thus given by σtotal = = |f(θ)|2 dΩ 1 k2 1 dφ −1 dµ l l′ (2 l + 1) (2 l ′ + 1) exp[ i (δl − δl ′ )] (15. What is the source of the incoming waves? Obviously. This implies that the coefﬁcients of the incoming spherical waves in the large-r expansions of ψ(r) and ψ0 (r) must be the same. (15.35) give exp( i k r) ψ(r) − ψ0 (r) √ = f(θ).53) and (15. Now. Note that Al = Cl cos δl and Bl = −Cl sin δl . Equation (15. Eqs.42) and (15. they must be part of the large-r asymptotic expansion of the incident wavefunction. (15.42) and (15. Thus. n r (15.56) f(θ) = l=0 (2 l + 1) exp( i δl) sin δl Pl (cos θ). spherical waves) is equivalent to determining the phase-shifts δl .54) that Cl = (2 l + 1) exp[ i (δl + l π/2)].58) × sin δl sin δl ′ Pl (µ) Pl ′ (µ).55) yield ∞ (15. In fact.Scattering Theory 215 where use has been made of Eqs.. It follows from Eqs. (15. Now.55) Note that the right-hand side consists of an outgoing spherical wave only.49) that ψ0 (r) ≃ √ n l i l (2l + 1) e i (k r−l π/2) − e−i (k r−l π/2) Pl (cos θ) 2ikr (15. The above expression can also be written √ sin(k r − l π/2 + δl ) ψ(r) ≃ n Pl (cos θ). k (15. determining the scattering amplitude f(θ) via a decomposition into partial waves (i. (15.57) Clearly.52) Cl kr l where the sine and cosine functions have been combined to give a sine function which is phase-shifted by δl . (15.52) yields ψ(r) ≃ √ n l Cl e i (k r−l π/2+δl ) − e−i (k r−l π/2+δl ) Pl (cos θ).34) and (15. (15.17)].

because its region of validity does not include the origin (where V = 0).65) l (l + 1) 2 m d2 ul + k2 − − 2 V ul = 0. (15.61) √ ∞ n l=0 il (2 l + 1) Rl (r) Pl (cos θ).60) where jl′ (x) denotes djl (x)/dx.66) . (15.59) where use has been made of Eq. r (15.62) Rl (r) = exp( i δl) [cos δl jl (k r) − sin δl yl (k r)] . cos δl jl (k a) − sin δl yl (k a) (15. Rl (r). The logarithmic derivative of the lth radial wavefunction.57) can be evaluated. l QUANTUM MECHANICS (15.45). k a yl′ (k a) − βl+ yl (k a) (15.216 where µ = cos θ. ul (r) . 2 dr r2 ¯ h (15.36). In the region r > a. just outside the range of the potential is given by βl+ = k a cos δl jl′ (k a) − sin δl yl′ (k a) . (15. the wavefunction ψ(r) satisﬁes the free-space Schr¨dinger equation (15. 15.64) (15. The above equation can be inverted to give tan δl = k a jl′ (k a) − βl+ jl (k a) . Consider a spherically symmetric potential V(r) which vanishes for r > a. The most general solution which is consistent with no o incoming spherical-waves is ψ(r) = where Note that yl (k r) functions are allowed to appear in the above expression. the problem of determining the phase-shift δl is equivalent to that of obtaining βl+ .5 Determination of Phase-Shifts Let us now consider how the phase-shifts δl in Eq. etc. (15. It follows that σtotal = 4π k2 (2 l + 1) sin2 δl .63) Thus. The most general solution to Schr¨dinger’s equation inside the range of the potential o (r < a) which does not depend on the azimuthal angle φ is ψ(r) = where Rl (r) = and √ ∞ n l=0 i l (2 l + 1) Rl (r) Pl (cos θ). where a is termed the range of the potential.

kr (15. 15.74) The corresponding radial wavefunction for the incident wave takes the form [see Eq.71) yields sin(k a)/k a tan δ0 = = − tan(k a). which implies that ul = 0 for all l. yl (k a) (15. it follows that βl+ = βl− . integrate out to r = a. and zero for r > a.72) − cos(k a)/ka where use has been made of Eqs. The S-wave radial wave function is [see Eq. which is usually referred to as the S-wave. It follows that δ0 = −k a. for which the potential is inﬁnite for r < a. Consider scattering by a hard sphere. (15.73) (15.6 Hard Sphere Scattering Let us test out this scheme using a particularly simple example.63).41). We can launch a wellbehaved solution of the above equation from r = 0. and form the logarithmic derivative 1 d(ul /r) βl− = . (15.71) Consider the l = 0 partial wave. βl− = βl+ = ∞. (15.40) and (15. R0 (r) = kr (15. tan δl = jl (k a) .63) thus gives (15.49)] sin(k r) ˜ .67) ensures that the radial wavefunction is well-behaved at the origin. Equation (15. (15.70) for all l. Equation (15.69) The phase-shift δl is then obtainable from Eq.75) .61)] R0 (r) = exp(−i k a) [cos(k a) sin(k r) − sin(k a) cos(k r)] kr sin[k (r − a)] = exp(−i k a) .Scattering Theory The boundary condition ul (0) = 0 217 (15. (15. (15. It follows that ψ(r) is zero in the region r < a.68) (ul /r) dr r=a Since ψ(r) and its ﬁrst derivatives are necessarily continuous for physically acceptible wavefunctions. Thus. (15.

with l > 0. and (15. (15.57). all partial waves up to lmax = k a contribute signiﬁcantly to the scattering cross-section. l=0 (15. Let us examine the low and high energy asymptotic limits of tan δl . It follows from Eq.e. the cross-section for classical particles bouncing off a hard sphere of radius a). However. (15. (15. there must also be some scattering in the forward . so we would not expect to obtain the classical result in this limit.. (k r)l+1 (15. Thus.77) yl (k r) ≃ − where n!! = n (n − 2) (n − 4) · · · 1. However. which is somewhat surprizing. spherically symmetric scattering) is important. ka 2π (2 l + 1) ≃ 2π a2. In other words. Consider the high energy limit k a ≫ 1. low energy scattering implies relatively long wavelengths. In this regime. Note that the total cross-section σtotal = dσ dΩ = 4π a2 dΩ (15.81) With so many l values contributing. It follows that tan δl = −(k a)2 l+1 ..e. (15. For hard sphere scattering.78) It is clear that we can neglect δl . except that it is phase-shifted by k a. It follows from Eqs. At high energies.82) σtotal ≃ k2 l=0 This is twice the classical result.73) that sin2 k a dσ = ≃ a2 2 dΩ k for k a ≪ 1. in order to produce a “shadow” behind the sphere.218 QUANTUM MECHANICS Thus.79) is four times the geometric cross-section π a2 (i. (2 l + 1)!! (2 l − 1)!! . it is legitimate to replace sin2 δl by its average value 1/2. incident waves with impact parameters less than a must be deﬂected.76) (15.80) (15.17). the spherical Bessel functions reduce to: jl (k r) ≃ (k r)l . the actual l = 0 radial wavefunction is similar to the incident l = 0 wavefunction. since we might expect to obtain the classical result in the short wavelength limit.59) that σtotal 4π ≃ 2 k lmax (2 l + 1) sin2 δl . (2 l + 1) [(2 l − 1)!!] 2 (15. at low energy only S-wave scattering (i. Low energy implies that k a ≪ 1. with respect to δ0 .

67). gives the actual cross-section of 2π a2.e. The potential is repulsive for V0 > 0. Here.85) ¯ 2 κ2 h . at these energies.87) (15.Scattering Theory 219 direction in order to produce destructive interference with the incident plane-wave. (15. Here. (15. at low energies (i.. with a ﬁnite range potential. kr (15. 2m Note that Eq. V0 is a constant. (15. We obtain sin(k ′ r) . we have R0 (r) = B where sinh(κ r) .88) (15. (15.84) R0 (r) = B r where use has been made of the boundary condition (15. the interference is not completely destructive. and tan(k a + δ0 ) = k tan(k ′ a) k′ k tanh(κ a) κ (15. (15. B is a constant. The effective cross-section associated with this bright spot is π a2 which. and attractive for V0 < 0. characterized by V = V0 for r < a. The inside wavefunction follows from Eq. For E < V0 . and its radial derivative. let us consider scattering by a ﬁnite potential well. It follows that.66). 15.89) .86) (15. π a2. when combined with the cross-section for classical reﬂection.61)] R0 (r) = exp( i δ0) [cos δ0 j0 (k r) − sin δ0 y0 (k r)] = exp( i δ0) sin(k r + δ0 ) . As a speciﬁc example.83) where use has been made of Eqs. The outside wavefunction is given by [see Eq. In fact. and V = 0 for r ≥ a. only S-wave scattering is important. r (15.41). at r = a yields V0 − E = tan(k a + δ0 ) = for E > V0 . and the shadow has a bright spot (the so-called “Poisson spot”) in the forward direction. when 1/k is much larger than the range of the potential) partial waves with l > 0 make a negligible contribution to the scattering cross-section. 2m Matching R0 (r).84) only applies when E > V0 .7 Low Energy Scattering In general.40) and (15. and ¯ 2 k′ 2 h E − V0 = .

In reality. (15. Consider an attractive potential. It follows that there are certain values of V0 and k which give rise to almost perfect transmission of the incident wave. for which E > V0 ..81). can reach the value π/2. and has been observed experimentally. Suppose that the quantity 2 m |V0 | a2 /¯ 2 is slightly less than π/2. As the h incident energy increases. k′ a ≃ 2 m |V0 | a2 .92) vanishes. which is given by Eq. the right-hand side is much less that unity. it follows from Eq. In this case.88) that. Suppose that |V0 | ≫ E (i. (15. ¯2 h (15. unless tan(k ′ a) becomes extremely large. k′ (15.92) It follows that the total (S-wave) scattering cross-section is independent of the energy of the incident particles (provided that this energy is sufﬁciently small). despite the very strong attraction of the potential.8 Resonances There is a signiﬁcant exception to the independence of the cross-section on energy mentioned above. so that k ′ ≫ k.90) (15. δ0 ≃ k a k′ a According to Eq. We can see from Eq. This is called the Ramsauer-Townsend effect. because of contributions from l > 0 partial waves. (15. (15. tan(k ′ a) becomes inﬁnite.93) 4π tan(k ′ a) ≃ 2 sin2 δ0 = 4π a2 −1 k k′ a 2 k tan(k ′ a). But. 15.91) . the cross-section is not exactly zero. and the scattering cross-section (15. Note that there are values of k ′ a (e.220 QUANTUM MECHANICS for E < V0 . the scattering cross-section is given by σtotal Now k′ a = so for sufﬁciently small values of k a. (15.93). so we can no longer assume that the right-hand side of Eq.g.49) at which δ0 → π. so replacing the tangent of a small quantity with the quantity itself. (15. we obtain k a + δ0 ≃ This yields tan(k ′ a) −1 . at low incident energies. these contributions are small. k ′ a. In fact.88) that at the value of the incident .88) is small..94) k2 a2 + 2 m |V0 | a2 .e. ¯2 h (15. k ′ a ≃ 4. the depth of the potential well is much larger than the energy of the incident particles).

However.97) Let us expand cot δl in the vicinity of the resonant energy: cot δl (E) = cot δl (E0 ) + = − Deﬁning dδl (E) dE we obtain = E=E0 d cot δl dE E=E0 (E − E0 ) + · · · (15. (15.92) for k ′ a = π/2 (since k a ≪ 1). the energy of the scattering system is essentially the same as the energy of the bound state. an incident particle would like to form a bound state in the potential well. Nevertheless. This implies that σtotal = 1 4π sin2 δ0 = 4π a2 .100) . The condition 2 m |V0 | a2 π = 2 2 ¯ h (15. The origin of this rather strange behaviour is quite simple. The cross-section for resonance scattering is generally much larger than that for non-resonance scattering. Furthermore. or δ0 ≃ π/2 (since we are assuming that k a ≪ 1).Scattering Theory 221 energy when k ′ a = π/2 then we also have k a + δ0 = π/2. Suppose that δl attains the value π/2 at the incident energy E0 . so it is reasonable to assume that there is a similar resonance when the phase-shift of the lth partial wave is π/2. Γ (15. the magnitude of the cross-section is much larger than that given in Eq. for a potential well which satisﬁes the above equation. 2 . so that δl (E0 ) = π . and the subsequent decay of the bound state and release of the particle. In this situation. this sort of resonance scattering is best understood as the capture of an incident particle to form a metastable bound state. There is nothing special about the l = 0 partial wave. 2 2 a2 k k (15. Thus.96) is equivalent to the condition that a spherical well of depth V0 possesses a bound state at zero energy.95) Note that the cross-section now depends on the energy. Γ (15. since the system has a small positive energy.98) 1 dδl sin2 δl dE E=E0 (E − E0 ) + · · · . the bound state is not stable. 2 (15.99) 2 cot δl (E) = − (E − E0 ) + · · · . We have seen that there is a resonant effect when the phase-shift of the S-wave takes the value π/2.

The quantity Γ is the width of the resonance (in energy). from Eq. We can interpret the Breit-Wigner formula as describing the absorption of an incident particle to form a metastable state.102) This is the famous Breit-Wigner formula. and lifetime τ = ¯ /Γ . h . of energy E0 .222 QUANTUM MECHANICS Recall. The variation of the partial cross-section σl with the incident energy has the form of a classical resonance curve. that the contribution of the lth partial wave to the scattering cross-section is 4π 1 4π . k2 (E − E0 )2 + Γ 2 /4 (15. (15.59). σl ≃ 4π Γ 2 /4 (2 l + 1) . (15.101) σl = 2 (2 l + 1) sin2 δl = 2 (2 l + 1) k k 1 + cot2 δl Thus.

- Direct Simulation Monte Carlo Schemes for Coulomb Interactions in Plasmas
- Griffiths Solution of quantum mechanics
- nonlinearsystems
- Re Led
- PhysRevE.80.066120
- Quantum-Mechanics
- Sem_Course_Curriculum-ECE_DTU_3rd Sem
- Relativistic Quantum Mechanics
- A Practical Introduction to the Lattice Boltzmann Method
- b9wq5.Exploring.quantum.mechanics.a.collection.of.700.Solved.problems.for.Students.lecturers.and.Researchers.pdf
- SPE-23575-MS - Slug Test Method in Pressure Sensitive Reservoirs
- Sakurai - Modern Quantum Mechanics - Solutions Manual
- Alice P. A. Urbano et al- Vibrational partition functions for atom-diatom and atom-triatom van der Waals systems
- Quantum Mechanics
- Eos Zenozcme Jcp2013-Libre
- Quantum mechanics
- Oscillation Theory for Second Order Dynamic Equations
- Principles of Quantum Mechanic ed 2 ~ shankar
- 2007 (Artigo) Identification of Nonlinear Systems - MISO Modeling
- Quantum Mechanics
- FlexuralTorsional Coupled ISIK
- Spiegel SchaumsTheoryAndProblemsOfTheoreticalMechanics Text
- 01.07020402.Papakanellos
- Classical_Electrodynamics_Jackson.pdf
- F. Devreux and J. P. Boucher- Solitons in Ising-like quantum spin chains in a magnetic field
- Classical Electrodynamics
- Black Holes Dwarfs Neutrons t 6
- Shankar - Principles of quantum mechanics
- An Energy-conserving Co-rotational Procedure for the Dynamics of Planar Beam Structures
- Schaum's Outlines-Modern Physics

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd