Quantum Mechanics
Quantum Mechanics
Quantum Mechanics
Richard Fitzpatrick
Professor of Physics
The University of Texas at Austin
Contents
1 Introduction 7
1.1 Intended audience . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2 Major Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Aim of Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Outline of Course . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
I Fundamentals 10
2 Probability Theory 11
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 What is Probability? . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Combining Probabilities . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 The Mean, Variance, and Standard Deviation . . . . . . . . . . . . 14
2.5 Continuous Probability Distributions . . . . . . . . . . . . . . . . . 16
3 Wave-Particle Duality 18
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Classical Light-Waves . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 The Photoelectric Effect . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Quantum Theory of Light . . . . . . . . . . . . . . . . . . . . . . . 22
3.5 Classical Interference of Light-Waves . . . . . . . . . . . . . . . . . 23
3.6 Quantum Interference of Light . . . . . . . . . . . . . . . . . . . . 25
3.7 Classical Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.8 Quantum Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.9 Wave-Packets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.10 Evolution of Wave-Packets . . . . . . . . . . . . . . . . . . . . . . . 33
3.11 Heisenberg’s Uncertainty Principle . . . . . . . . . . . . . . . . . . 36
3.12 Schrödinger’s Equation . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.13 Collapse of the Wave-Function . . . . . . . . . . . . . . . . . . . . 41
2
4.3 Normalization of the Wave-Function . . . . . . . . . . . . . . . . . 44
4.4 Expectation Values and Variances . . . . . . . . . . . . . . . . . . . 47
4.5 Ehrenfest’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.6 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.7 The Momentum Representation . . . . . . . . . . . . . . . . . . . . 54
4.8 The Uncertainty Principle . . . . . . . . . . . . . . . . . . . . . . . 58
4.9 Eigenstates and Eigenvalues . . . . . . . . . . . . . . . . . . . . . . 61
4.10 Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.11 Continuous Eigenvalues . . . . . . . . . . . . . . . . . . . . . . . . 67
4.12 Stationary States . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
5 One-Dimensional Potentials 73
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.2 The Infinite Potential Well . . . . . . . . . . . . . . . . . . . . . . . 73
5.3 The Square Potential Barrier . . . . . . . . . . . . . . . . . . . . . . 75
5.4 The WKB Approximation . . . . . . . . . . . . . . . . . . . . . . . 81
5.5 Cold Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.6 α-Decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5.7 The Square Potential Well . . . . . . . . . . . . . . . . . . . . . . . 90
5.8 The Simple Harmonic Oscillator . . . . . . . . . . . . . . . . . . . . 94
6 Multi-Particle Systems 99
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.2 Fundamental Concepts . . . . . . . . . . . . . . . . . . . . . . . . . 99
6.3 Non-Interacting Particles . . . . . . . . . . . . . . . . . . . . . . . . 101
6.4 Two-Particle Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 103
6.5 Identical Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
3
8 Orbital Angular Momentum 121
8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
8.2 Angular Momentum Operators . . . . . . . . . . . . . . . . . . . . 121
8.3 Representation of Angular Momentum . . . . . . . . . . . . . . . . 123
8.4 Eigenstates of Angular Momentum . . . . . . . . . . . . . . . . . . 125
8.5 Eigenvalues of Lz . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
8.6 Eigenvalues of L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
8.7 Spherical Harmonics . . . . . . . . . . . . . . . . . . . . . . . . . . 130
II Applications 179
4
12.3 The Two-State System . . . . . . . . . . . . . . . . . . . . . . . . . 183
12.4 Non-Degenerate Perturbation Theory . . . . . . . . . . . . . . . . . 185
12.5 The Quadratic Stark Effect . . . . . . . . . . . . . . . . . . . . . . . 187
12.6 Degenerate Perturbation Theory . . . . . . . . . . . . . . . . . . . 192
12.7 The Linear Stark Effect . . . . . . . . . . . . . . . . . . . . . . . . . 194
12.8 The Fine Structure of Hydrogen . . . . . . . . . . . . . . . . . . . . 196
12.9 The Zeeman Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
12.10 Hyperfine Structure . . . . . . . . . . . . . . . . . . . . . . . . . . 205
5
15.4 Partial Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
15.5 Determination of Phase-Shifts . . . . . . . . . . . . . . . . . . . . . 259
15.6 Hard Sphere Scattering . . . . . . . . . . . . . . . . . . . . . . . . 261
15.7 Low Energy Scattering . . . . . . . . . . . . . . . . . . . . . . . . . 263
15.8 Resonances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
6
1 INTRODUCTION
1 Introduction
The textbooks which I have consulted most frequently whilst developing course
material are:
The principles of quantum mechanics, P.A.M. Dirac, 4th Edition (revised), (Oxford
University Press, Oxford UK, 1958).
Quantum mechanics, E. Merzbacher, 2nd Edition, (John Wiley & Sons, New York
NY, 1970).
Introduction to the quantum theory, D. Park, 2nd Edition, (McGraw-Hill, New
York NY, 1974).
Modern quantum mechanics, J.J. Sakurai, (Benjamin/Cummings, Menlo Park CA,
1985).
Quantum theory, D. Bohm, (Dover, New York NY, 1989).
Problems in quantum mechanics, G.L. Squires, (Cambridge University Press, Cam-
bridge UK, 1995).
7
1.3 Aim of Course 1 INTRODUCTION
Quantum physics, S. Gasiorowicz, 2nd Edition, (John Wiley & Sons, New York
NY, 1996).
Nonclassical physics, R. Harris, (Addison-Wesley, Menlo Park CA, 1998).
Introduction to quantum mechanics, D.J. Griffiths, 2nd Edition, (Pearson Prentice
Hall, Upper Saddle River NJ, 2005).
8
1.4 Outline of Course 1 INTRODUCTION
9
Part I
Fundamentals
10
2 PROBABILITY THEORY
2 Probability Theory
2.1 Introduction
This first section is devoted to a brief, and fairly low level, introduction to a
branch of Mathematics known as probability theory.
where Ω(Σ) is the total number of systems in the ensemble, and Ω(X) the number
of systems exhibiting the outcome X. We can see that the probability P(X) must
be a number between 0 and 1. The probability is zero if no systems exhibit the
outcome X, even when the number of systems goes to infinity. This is just another
way of saying that there is no chance of the outcome X. The probability is unity
if all systems exhibit the outcome X in the limit as the number of systems goes to
infinity. This is another way of saying that the outcome X is bound to occur.
11
2.3 Combining Probabilities 2 PROBABILITY THEORY
where Ω(X | Y) is the number of systems in the ensemble which exhibit either
the outcome X or the outcome Y. It is clear that
if the outcomes X and Y are mutually exclusive (which must be the case if they
are two distinct outcomes). Thus,
So, the probability of the outcome X or the outcome Y is just the sum of the indi-
vidual probabilities of X and Y. For instance, with a six-sided die the probability
of throwing any particular number (one to six) is 1/6, because all of the possible
outcomes are considered to be equally likely. It follows, from what has just been
said, that the probability of throwing either a one or a two is simply 1/6 + 1/6,
which equals 1/3.
12
2.3 Combining Probabilities 2 PROBABILITY THEORY
which is called the normalization condition, and must be satisfied by any complete
set of probabilities. This condition is equivalent to the self-evident statement that
an observation of a system must definitely result in one of its possible outcomes.
It is also fairly obvious that the number of pairs of states in the ensemble Σ ⊗ Σ
which exhibit the outcome X in the first state and Y in the second state is just the
product of the number of states which exhibit the outcome X and the number of
states which exhibit the outcome Y in the original ensemble, so
Thus, the probability of obtaining the outcomes X and Y in two statistically inde-
pendent observations is just the product of the individual probabilities of X and
Y. For instance, the probability of throwing a one and then a two on a six-sided
die is 1/6 × 1/6, which equals 1/36.
13
2.4 The Mean, Variance, and Standard Deviation 2 PROBABILITY THEORY
Suppose that f(u) is some function of u. Then, for each of the M possible
values of u, there is a corresponding value of f(u) which occurs with the same
probability. Thus, f(u1 ) corresponds to u1 and occurs with the probability P(u1 ),
and so on. It follows from our previous definition that the mean value of f(u) is
given by
XM
hf(u)i ≡ P(ui ) f(ui ). (2.13)
i=1
14
2.4 The Mean, Variance, and Standard Deviation 2 PROBABILITY THEORY
Suppose that f(u) and g(u) are two general functions of u. It follows that
M
X M
X M
X
hf(u) + g(u)i = P(ui ) [f(ui ) + g(ui )] = P(ui ) f(ui ) + P(ui ) g(ui ),
i=1 i=1 i=1
(2.14)
so
hf(u) + g(u)i = hf(u)i + hg(u)i. (2.15)
Finally, if c is a general constant then it is clear that
We now know how to define the mean value of the general variable u. But,
how can we characterize the scatter around the mean value? We could investigate
the deviation of u from its mean value hui, which is denoted
∆u ≡ u − hui. (2.17)
In fact, this is not a particularly interesting quantity, since its average is obviously
zero:
h∆ui = h(u − hui)i = hui − hui = 0. (2.18)
This is another way of saying that the average deviation from the mean vanishes.
A more interesting quantity is the square of the deviation. The average value of
this quantity,
XM
2
P(ui ) (ui − hui)2 ,
D E
(∆u) = (2.19)
i=1
is usually called the variance. The variance is clearly a positive number, unless
there is no scatter at all in the distribution, so that all possible values of u cor-
respond to the mean value hui, in which case it is zero. The following general
relation is often useful
giving
(u − hui)2 = u2 − hui2 .
D E D E
(2.21)
15
2.5 Continuous Probability Distributions 2 PROBABILITY THEORY
Suppose, now, that the variable u can take on a continuous range of possible
values. In general, we expect the probability that u takes on a value in the range
u to u + du to be directly proportional to du, in the limit that du → 0. In other
words,
P(u ∈ u : u + du) = P(u) du, (2.23)
where P(u) is known as the probability density. The earlier results (2.5), (2.12),
and (2.19) generalize in a straight-forward manner to give
Z∞
1 = P(u) du, (2.24)
−∞
Z∞
hui = P(u) u du, (2.25)
−∞
Z∞
(∆u)2 P(u) (u − hui)2 du = u2 − hui2 ,
D E D E
= (2.26)
−∞
respectively.
Problems
1. In the “game” of Russian roulette, the player inserts a single cartridge into the drum of a revolver,
leaving the other five chambers of the drum empty. The player then spins the drum, aims at his/her
head, and pulls the trigger.
(a) What is the probability of the player still being alive after playing the game N times?
16
2.5 Continuous Probability Distributions 2 PROBABILITY THEORY
(b) What is the probability of the player surviving N − 1 turns in this game, 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. Suppose that the probability density for the speed s of a car on a road is given by
s
P(s) = A s exp − ,
s0
where 0 ≤ s ≤ ∞. Here, A and s0 are positive constants. More explicitly, P(s) ds gives the probability
that a car has a speed between s and s + ds.
3. An radioactive atom has a uniform decay probability per unit time w: i.e., the probability of decay
in a time interval dt is w dt. Let P(t) be the probability of the atom not having decayed at time t,
given that it was created at time t = 0. Demonstrate that
P(t) = e−wt.
17
3 WAVE-PARTICLE DUALITY
3 Wave-Particle Duality
3.1 Introduction
In classical mechanics, waves and particles are two completely different types of
physical entity. Waves are continuous and spatially extended, whereas particles
are discrete and have little or no spatial extent. However, in quantum mechan-
ics, waves sometimes act as particles, and particles sometimes act as waves—this
strange behaviour is known as wave-particle duality. In this section, we shall
examine how wave-particle duality shapes the general features of quantum me-
chanics.
18
3.2 Classical Light-Waves 3 WAVE-PARTICLE DUALITY
Finally, from standard electromagnetic theory, the energy density (i.e., the en-
ergy per unit volume) of a light-wave is
Ey2
U= , (3.12)
ǫ0
where ǫ0 = 8.85 × 10−12 F/m is the permittivity of free space. Hence, it follows
from Eqs. (3.1) and (3.3) that
U ∝ |ψ| 2 . (3.13)
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
U
G= . (3.14)
c
The so-called photoelectric effect, by which a polished metal surface emits elec-
trons when illuminated by visible and ultra-violet light, was discovered by Hein-
rich Hertz in 1887. The following facts regarding this effect can be established
via careful observation. First, a given surface only emits electrons when the fre-
quency 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 lin-
early with the light frequency. These facts are inexplicable within the framework
of classical physics.
20
3.3 The Photoelectric Effect 3 WAVE-PARTICLE DUALITY
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. (3.16)
Otherwise, the electron remains trapped in the potential well, and is not emitted.
Here, we are assuming that the probability of an electron absorbing two or more
light quanta is negligibly small compared to the probability of 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. 1. 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.
1
Plural of quantum: Latin neuter of quantus: how much.
21
3.4 Quantum Theory of Light 3 WAVE-PARTICLE DUALITY
0
0 W/h ν
E = h̄ ω, (3.17)
22
3.5 Classical Interference of Light-Waves 3 WAVE-PARTICLE DUALITY
incoming wave
x1
y
double slits x2
d
projection
screen
D
Figure 2: Classical double-slit interference of light.
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
23
3.5 Classical Interference of Light-Waves 3 WAVE-PARTICLE DUALITY
to get to this point, whereas light from the second slit travels a slightly different
distance x2 . It is easily demonstrated that
d
∆x = x2 − x1 ≃ y, (3.20)
D
provided d ≪ D. It follows from Eq. (3.1), and the well-known fact that light-
waves are superposible, that the wave-function at the point in question can be
written
ψ(y, t) ∝ ψ1 (t) e i k x1 + ψ2 (t) e i k x2 , (3.21)
where ψ1 and ψ2 are the wave-functions at the first and second slits, respectively.
However,
ψ1 = ψ2 , (3.22)
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. (3.13) that the light intensity on the screen a distance y from the center-line
is
I(y) ∝ |ψ(y, t)| 2 . (3.23)
Using Eqs. (3.20)–(3.23), we obtain
k ∆x kd
! !
2
I(y) ∝ cos ≃ cos2 y . (3.24)
2 2D
Figure 3 shows the characteristic interference pattern corresponding to the above
expression. This pattern consists of equally spaced light and dark bands of char-
acteristic width
Dλ
∆y = . (3.25)
d
24
3.6 Quantum Interference of Light 3 WAVE-PARTICLE DUALITY
I(y) ∆y
0
y
Figure 3: Classical double-slit interference pattern.
It follows, from the above discussion that the interference pattern is built up
one photon at a time: i.e., the pattern is not due to the interaction of different
photons. Moreover, the point at which a given photon strikes the film is clearly
not influenced by the points that previous photon struck the film, given that there
25
3.6 Quantum Interference of Light 3 WAVE-PARTICLE DUALITY
is only one photon in the apparatus at any given time. Hence, the only way
in which the classical interference pattern can be reconstructed, after a great
many photons have passed through the apparatus, is if each photon has a greater
probability of striking the film at points where the classical interference pattern is
bright, and a lesser probability of striking the film at points where the interference
pattern is dark.
Suppose, then, that we allow N photons to pass through our apparatus, and
then count the number of photons which strike the recording film between y and
y + ∆y, where ∆y is a relatively small division. Let us call this number n(y).
Now, 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, since each photon carries a fixed amount
of energy. Hence, in order to reconcile the classical and quantum viewpoints, we
need
n(y)
Py (y) ≡ lim ∝ I(y) ∆y, (3.26)
N→∞ N
where I(y) is given in Eq. (3.24). Here, Py (y) is the probability that a given
photon strikes the film between y and y+∆y. This probability is simply a number
between 0 and 1. A probability of 0 means that there is no chance of a photon
striking the film between y and y + ∆y, whereas a probability of 1 means that
every photon is certain to strike the film in this interval. Note that Py ∝ ∆y. In
other words, the probability of a photon striking a region of the film of width ∆y
is directly proportional to this width. Obviously, this is only true as long as ∆y
is relatively small. It is convenient to define a quantity known as the probability
density, P(y), 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. Now, Eq. (3.26) yields
Py (y) ∝ I(y) dy, which gives P(y) ∝ I(y). However, according to Eq. (3.23),
I(y) ∝ |ψ(y)| 2 . Thus, we obtain
In other words, the probability density of a photon striking a given point on the
film is proportional to the modulus squared of the wave-function at that point.
Another way of saying this is that the probability of a measurement of the pho-
26
3.6 Quantum Interference of Light 3 WAVE-PARTICLE DUALITY
ton’s distance from the centerline, at the location of the film, yielding a result
between y and y + dy is proportional to |ψ(y)| 2 dy.
Note that, in the quantum mechanical picture, we can only predict the prob-
ability that a given photon strikes a given point on the film. If photons behaved
classically then we could, in principle, solve their equations of motion and pre-
dict exactly where each photon was going to strike the film, given its initial po-
sition and velocity. This loss of determinancy in quantum mechanics is a direct
consequence of wave-particle duality. In other words, we can only reconcile the
wave-like and particle-like properties of light in a statistical sense. It is impossible
to reconcile them on the individual particle level.
In principle, each photon which passes through our apparatus is equally likely
to pass through one of the two slits. So, can we determine which slit a given
photon passed through? Well, suppose that our original interference experiment
involves sending N ≫ 1 photons through our apparatus. We know that we get
an interference pattern in this experiment. Suppose that we perform a modi-
fied interference experiment in which we close off one slit, send N/2 photons
through the apparatus, and then open the slit and close off the other slit, and
send N/2 photons through the apparatus. In this second experiment, which is
virtually identical to the first on the individual photon level, we know exactly
which slit each photon passed through. However, it is clear from the wave theory
of light (which we expect to agree with the quantum theory in the limit N ≫ 1)
that our modified interference experiment will not result in the formation of an
interference pattern. After all, according to wave theory, it is impossible to ob-
tain a two-slit interference pattern from a single slit. Hence, we conclude that
any attempt to measure which slit each photon in our two-slit interference ex-
periment passes through results in the destruction of the interference pattern. It
follows that, in the quantum mechanical version of the two-slit interference ex-
periment, we must think of each photon as essentially passing through both slits
simultaneously.
27
3.7 Classical Particles 3 WAVE-PARTICLE DUALITY
28
3.8 Quantum Particles 3 WAVE-PARTICLE DUALITY
29
3.9 Wave-Packets 3 WAVE-PARTICLE DUALITY
3.9 Wave-Packets
where k = p/h̄ and ω = E/h̄. Here, ω and k are linked via the dispersion
relation (3.35). Expression (3.38) represents a plane-wave which propagates in
the x-direction with the phase-velocity vp = ω/k. As we have seen, this phase-
velocity is only half of the classical velocity of a massive particle.
30
3.9 Wave-Packets 3 WAVE-PARTICLE DUALITY
For instance, suppose that at t = 0 the wave-function of our particle takes the
form
2
(x − x )
0
ψ(x, 0) ∝ exp i k0 x − . (3.42)
4 (∆x) 2
Thus, the initial probability density of the particle is written
(x − x0 ) 2
2
|ψ(x, 0)| ∝ exp −
. (3.43)
2 (∆x) 2
This particular probability distribution is called a Gaussian distribution, and is
plotted in Fig. 4. It can be seen that a measurement of the particle’s position is
most likely to yield the value x0 , and very unlikely to yield a value which differs
from x0 by more than 3 ∆x. Thus, (3.42) is the wave-function of a particle which
is initially localized around x = x0 in some region whose width is of order ∆x.
This type of wave-function is known as a wave-packet.
where y0 = −i β/2. The integral now just reduces to a number, as can easily be
seen by making the change of variable z = y − y0 . Hence, we obtain
(k − k0 ) 2
ψ̄(k) ∝ exp −i k x0 −
, (3.49)
4 (∆k)2
where
1
∆k = . (3.50)
2 ∆x
32
3.10 Evolution of Wave-Packets 3 WAVE-PARTICLE DUALITY
We have seen, in Eq. (3.42), how to write the wave-function of a particle which
is initially localized in x-space. But, how does this wave-function evolve in time?
Well, according to Eq. (3.39), we have
Z∞
ψ(x, t) = ψ̄(k) e i φ(k) dk, (3.53)
−∞
where
φ(k) = k x − ω(k) t. (3.54)
33
3.10 Evolution of Wave-Packets 3 WAVE-PARTICLE DUALITY
where
β1 = 2 ∆k (x − x0 − vg t), (3.63)
β2 = 2 α (∆k) 2 t, (3.64)
The above expression can be rearranged to give
Z∞
2 2
ψ(x, t) ∝ e i (k0 x−ω0 t)−(1+i β2 ) β /4 e−(1+i β2 ) (y−y0 ) dy, (3.65)
−∞
34
3.10 Evolution of Wave-Packets 3 WAVE-PARTICLE DUALITY
where
2 α2 t2 2
σ (t) = (∆x) + . (3.68)
4 (∆x) 2
Note that the above wave-function is identical to our original wave-function
(3.42) at t = 0. This, justifies the approximation which we made earlier by
Taylor expanding the phase factor φ(k) about k = k0 .
However, it can be seen from Eq. (3.28) that this is identical to the classical parti-
cle velocity. Hence, the dispersion relation (3.35) turns out to be consistent with
classical physics, after all, as soon as we realize that particles must be identified
with wave-packets rather than plane-waves.
According to Eq. (3.68), the width of our wave-packet grows as time pro-
gresses. It follows from Eqs. (3.35) and (3.61) that the characteristic time for a
wave-packet of original width ∆x to double in spatial extent is
m (∆x)2
t2 ∼ . (3.73)
h̄
So, if an electron is originally localized in a region of atomic scale (i.e., ∆x ∼
10−10 m) then the doubling time is only about 10−16 s. Clearly, particle wave-
packets (for freely moving particles) spread very rapidly.
Note, from the previous analysis, that the rate of spreading of a wave-packet
is ultimately governed by the second derivative of ω(k) with respect to k. This
is why a functional relationship between ω and k is generally known as a dis-
persion relation: i.e., because it governs how wave-packets disperse as time pro-
gresses. However, for the special case where ω is a linear function of k, the
second derivative of ω with respect to k is zero, and, hence, there is no dis-
persion of wave-packets: i.e., wave-packets propagate without changing shape.
Now, the dispersion relation (3.7) for light-waves is linear in k. It follows that
light-pulses propagate through a vacuum without spreading. Another property
of linear dispersion relations is that the phase-velocity, vp = ω/k, and the group-
velocity, vg = dω/dk, are identical. Thus, both plane light-waves and light-pulses
propagate through a vacuum at the characteristic speed c = 3 × 108 m/s. Of
course, the dispersion relation (3.35) for particle waves is not linear in k. Hence,
particle plane-waves and particle wave-packets propagate at different velocities,
and particle wave-packets also gradually disperse as time progresses.
36
3.11 Heisenberg’s Uncertainty Principle 3 WAVE-PARTICLE DUALITY
It can be seen from Eqs. (3.35), (3.61), and (3.68) that at large t a particle
wave-function of original width ∆x (at t = 0) spreads out such that its spatial
extent becomes
h̄ t
σ∼ . (3.76)
m ∆x
It is easily demonstrated that this spreading is a consequence of the uncertainty
principle. Since the initial uncertainty in the particle’s position is ∆x, it follows
that the uncertainty in its momentum is of order h̄/∆x. This translates to an
uncertainty in velocity of ∆v = h̄/(m ∆x). Thus, if we imagine that parts of the
wave-function propagate at v0 + ∆v/2, and others at v0 − ∆v/2, where v0 is the
mean propagation velocity, then the wave-function will clearly spread as time
37
3.11 Heisenberg’s Uncertainty Principle 3 WAVE-PARTICLE DUALITY
D
lens
f scattered photon
y θ α
x
electron incoming photon
However,
D/2
tan α = , (3.79)
f
where α is the half-angle subtended by the lens at the electron. Assuming that α
is small, we can write
D
α≃ . (3.80)
2f
Hence,
λ
∆x ≃ . (3.81)
2α
It follows that we can reduce the uncertainty in the electron’s position by mini-
mizing the ratio λ/α: i.e., by using short wave-length radiation, and a wide-angle
lens.
39
3.12 Schrödinger’s Equation 3 WAVE-PARTICLE DUALITY
and Z∞
∂2 ψ
= (−k2 ) ψ̄(k) e i (k x−ω t) dk, (3.87)
∂x2 −∞
whereas Z∞
∂ψ
= (−i ω) ψ̄(k) e i (k x−ω t) dk. (3.88)
∂t −∞
Thus,
Z∞
h̄ ∂2 ψ h̄ k2
∂ψ
i + = ω − ψ̄(k) e i (k x−ω t) dk = 0, (3.89)
∂t 2 m ∂x2 −∞ 2m
40
3.13 Collapse of the Wave-Function 3 WAVE-PARTICLE DUALITY
where use has been made of the dispersion relation (3.85). Multiplying through
by h̄, we obtain
∂ψ h̄2 ∂2 ψ
i h̄ =− . (3.90)
∂t 2 m ∂x2
This is known as Schrödinger’s equation, since it was first introduced by Erwin
Schrödinger in 1925. Schrödinger’s equation is a linear, second-order, differential
equation which governs the time evolution of a particle wave-function, and is
generally easier to solve than the integral equation (3.84).
h̄2 ∂2 h̄2 ∂2
− →− + V(x). (3.91)
2 m ∂x2 2 m ∂x2
This leads to the general form of Schrödinger’s equation:
∂ψ h̄2 ∂2 ψ
i h̄ =− + V(x) ψ. (3.92)
∂t 2 m ∂x2
41
3.13 Collapse of the Wave-Function 3 WAVE-PARTICLE DUALITY
BEFORE
ψ→
x→
AFTER
ψ→
x0 x→
common sense tells us that we must obtain the same value, x0 , since the particle
cannot have shifted position appreciably in an infinitesimal time interval. Thus,
immediately after the first measurement, a measurement of the particle’s posi-
tion is certain to give the value x0 , and has no chance of giving any other value.
This implies that the wave-function must have collapsed to some sort of “spike”
function located at x = x0 . This is illustrated in Fig. 6. Of course, as soon as the
wave-function has collapsed, it starts to expand again, as discussed in Sect. 3.10.
Thus, the second measurement must be made reasonably quickly after the first,
in order to guarantee that the same result will be obtained.
42
3.13 Collapse of the Wave-Function 3 WAVE-PARTICLE DUALITY
Problems
1. A He-Ne laser emits radiation of wavelength λ = 633 nm. 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? [modified from Squires]
2. The ionization energy of the hydrogen atom in its ground state is Eion = 13.60 eV (1 eV is the energy
acquired by an electron accelerated through a potential difference of 1 V). Calculate the frequency,
wavelength, and wave-number of the electromagnetic radiation which will just ionize the atom.
[from Squires]
3. The maximum energy of photoelectrons from aluminium is 2.3 eV for radiation of wavelength
2000 Å, and 0.90 eV for radiation of wavelength 2580 Å. Use this data to calculate Planck’s con-
stant, and the work function of aluminium. [from Gasiorowicz]
4. The relationship between wavelength and frequency for electromagnetic waves in a waveguide is
c
λ= q .
ν2 − ν02
What are the group and phase velocities of such waves as functions of ν0 and λ? [modified from
Gasiorowicz]
5. Nuclei, typically of size 10−14 m, frequently emit electrons with energies of 1–10 MeV. Use the
uncertainty principle to show that electrons of energy 1 MeV could not be contained in the nucleus
before the decay. [from Gasiorowicz]
where A and a are positive real constants. For what potential function V(x) does ψ satisfy the
Schrödinger equation? [from Griffiths]
43
4 FUNDAMENTALS OF QUANTUM MECHANICS
4.1 Introduction
The previous section served as a useful introduction to many of the basic concepts
of quantum mechanics. In this section, we shall examine these concepts in a more
systematic and mathematically rigorous fashion. For the sake of simplicity, we
shall concentrate on one-dimensional systems.
44
4.3 Normalization of the Wave-Function 4 FUNDAMENTALS OF QUANTUM MECHANICS
which has a probability 1 is a certain outcome. According to Eq. (4.2), the prob-
ability of a measurement of x yielding a result between −∞ and +∞ is
Z∞
Px ∈ −∞:∞ (t) = |ψ(x, t)| 2 dx. (4.3)
−∞
However, a measurement of x must yield a value between −∞ and +∞, since the
particle must be located somewhere. It follows that Px ∈ −∞:∞ = 1, or
Z∞
|ψ(x, t)| 2 dx = 1, (4.4)
−∞
However, Z∞
2 √
e−y dy = π, (4.8)
−∞
which implies that
1
|ψ0 | 2 = . (4.9)
(2π σ2 )1/2
Hence, a general normalized Gaussian wave-function takes the form
eiϕ −(x−x0 ) 2 /(4 σ2 )
ψ(x) = e , (4.10)
(2π σ2 )1/4
45
4.3 Normalization of the Wave-Function 4 FUNDAMENTALS OF QUANTUM MECHANICS
46
4.4 Expectation Values and Variances 4 FUNDAMENTALS OF QUANTUM MECHANICS
However, this is a necessary condition for the integral on the left-hand side of
Eq. (4.4) to converge. Hence, we conclude that all wave-functions which are
square-integrable [i.e., are such that the integral in Eq. (4.4) converges] have the
property that if the normalization condition (4.4) is satisfied at one instant in
time then it is satisfied at all subsequent times.
It is also possible to demonstrate, via very similar analysis to the above, that
dPx ∈ a:b
+ j(b, t) − j(a, t) = 0, (4.18)
dt
where Px ∈ a:b is defined in Eq. (4.2), and
i h̄ ∂ψ∗ ∗ ∂ψ
!
j(x, t) = ψ −ψ (4.19)
2m ∂x ∂x
is known as the probability current. Note that j is real. Equation (4.18) is a
probability conservation equation. According to this equation, 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.e., j(a, t)], and that out of the
interval [i.e., j(b, t)]. Here, we are interpreting j(x, t) as the flux of probability in
the +x-direction at position x and time t.
Note, finally, that not all wave-functions can be normalized according to the
scheme set out in Eq. (4.4). For instance, a plane-wave wave-function
ψ(x, t) = ψ0 e i (k x−ω t) (4.20)
is obviously not square-integrable, and, thus, cannot be normalized. For such
wave-functions, the best we can say is that
Zb
Px ∈ a:b (t) ∝ |ψ(x, t)| 2 dx. (4.21)
a
In the following, all wave-functions are assumed to be square-integrable and
normalized, unless otherwise stated.
47
4.4 Expectation Values and Variances 4 FUNDAMENTALS OF QUANTUM MECHANICS
Here, hxi is called the expectation value of x. Similarly the expectation value of
any function of x is Z∞
hf(x)i = f(x) |ψ| 2 dx. (4.23)
−∞
For instance, consider the normalized Gaussian wave-packet [see Eq. (4.10)]
eiϕ 2 2
ψ(x) = 2 1/4
e−(x−x0 ) /(4 σ ) . (4.25)
(2π σ )
The expectation value of x associated with this wave-function is
Z∞
1 2 2
hxi = √ x e−(x−x0 ) /(2 σ ) dx. (4.26)
2π σ2 −∞
√
Let y = (x − x0 )/( 2 σ). It follows that
Z √ Z
x0 ∞ −y2 2σ ∞ 2
hxi = √ e dy + √ y e−y dy. (4.27)
π −∞ π −∞
48
4.5 Ehrenfest’s Theorem 4 FUNDAMENTALS OF QUANTUM MECHANICS
However, the second integral on the right-hand side is zero, by symmetry. Hence,
making use of Eq. (4.8), we obtain
hxi = x0 . (4.28)
Clearly, the expectation value of x for a Gaussian wave-packet is equal to the most
likely value of x (i.e., the value of x which maximizes |ψ| 2 ).
50
4.5 Ehrenfest’s Theorem 4 FUNDAMENTALS OF QUANTUM MECHANICS
Suppose that the potential V(x) is slowly varying. In this case, we can expand
dV/dx as a Taylor series about hxi. Keeping terms up to second order, we obtain
dV(x) dV(hxi) dV 2 (hxi) 1 dV 3 (hxi)
= + 2
(x − hxi) + 3
(x − hxi) 2 . (4.44)
dx dhxi dhxi 2 dhxi
Substitution of the above expansion into Eq. (4.43) yields
dhpi dV(hxi) σx2 dV 3 (hxi)
=− − , (4.45)
dt dhxi 2 dhxi3
since h1i = 1, and hx − hxii = 0, and h(x − hxi) 2 i = σx2 . The final term on the
right-hand side of the above equation can be neglected when the spatial extent
of the particle wave-function, σx , is much smaller than the variation length-scale
of the potential. In this case, Eqs. (4.42) and (4.43) reduce to
dhxi
m = hpi, (4.46)
dt
dhpi dV(hxi)
= − . (4.47)
dt dhxi
These equations are exactly equivalent to the equations of classical mechanics,
with hxi playing the role of the particle displacement. Of course, if the spatial ex-
tent of the wave-function is negligible then a measurement of x is almost certain
to yield a result which lies very close to hxi. Hence, we conclude that quantum
mechanics corresponds to classical mechanics in the limit that the spatial extent
of the wave-function (which is typically of order the de Boglie wave-length) is
negligible. This is an important result, since we know that classical mechanics
gives the correct answer in this limit.
51
4.6 Operators 4 FUNDAMENTALS OF QUANTUM MECHANICS
4.6 Operators
x x2 = x2 x. (4.51)
where f is a general function, and c a general complex number. All of the opera-
tors employed in quantum mechanics are linear.
52
4.6 Operators 4 FUNDAMENTALS OF QUANTUM MECHANICS
Suppose that we wish to find the operator which corresponds to the classical
dynamical variable x p. In classical mechanics, there is no difference between
x p and p x. However, in quantum mechanics, we have already seen that x p 6=
p x. So, should be choose x p or p x? Actually, neither of these combinations
is Hermitian. However, (1/2) [x p + (x p)† ] is Hermitian. Moreover, (1/2) [x p +
(x p)† ] = (1/2) (x p + p† x† ) = (1/2) (x p + p x), which neatly resolves our problem
of which order to put x and p.
53
4.7 The Momentum Representation 4 FUNDAMENTALS OF QUANTUM MECHANICS
h̄2 ∂2
H≡− + V(x). (4.60)
2 m ∂x2
However, according to Schrödinger’s equation, (4.1), we have
h̄2 ∂2 ∂
− + V(x) = i h̄ . (4.61)
2 m ∂x2 ∂t
Hence,
∂
H ≡ i h̄
. (4.62)
∂t
Thus, the time-dependent Schrödinger equation can be written
∂ψ
i h̄ = H ψ. (4.63)
∂t
54
4.7 The Momentum Representation 4 FUNDAMENTALS OF QUANTUM MECHANICS
Suppose that ψ(x) = δ(x − x0 ). It follows from Eqs. (4.67) and (4.71) that
e−i p x0 /h̄
φ(p) = √ . (4.72)
2 π h̄
55
4.7 The Momentum Representation 4 FUNDAMENTALS OF QUANTUM MECHANICS
It turns out that we can just as well formulate quantum mechanics using
momentum-space wave-functions, φ(p, t), as real-space wave-functions, ψ(x, t).
The former scheme is known as the momentum representation of quantum me-
chanics. In the momentum representation, wave-functions are the Fourier trans-
forms of the equivalent real-space wave-functions, and dynamical variables are
represented by different operators. Furthermore, by analogy with Eq. (4.56), the
expectation value of some operator O(p) takes the form
Z∞
hOi = φ∗ (p, t) O(p) φ(p, t) dp. (4.75)
−∞
56
4.7 The Momentum Representation 4 FUNDAMENTALS OF QUANTUM MECHANICS
Eq. (4.22)] that |φ(p, t)| 2 can be interpreted as the probability density of a mea-
surement of momentum yielding the value p at time t. It follows that φ(p, t)
must satisfy an analogous normalization condition to Eq. (4.4): i.e.,
Z∞
|φ(p, t)| 2 dp = 1. (4.79)
−∞
57
4.8 The Uncertainty Principle 4 FUNDAMENTALS OF QUANTUM MECHANICS
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)].
58
4.8 The Uncertainty Principle 4 FUNDAMENTALS OF QUANTUM MECHANICS
where use has been made of Eq. (4.86). The above equation reduces to
Z∞
z= ψ∗ A B ψ dx − hAi hBi. (4.94)
−∞
59
4.8 The Uncertainty Principle 4 FUNDAMENTALS OF QUANTUM MECHANICS
Energy and time are represented by the operators H ≡ i h̄ ∂/∂t and t, respec-
tively. These operators do not commute, indicating that energy and time cannot
be measured simultaneously. In fact,
[H, t] = i h̄, (4.102)
so
h̄
σE σt ≥ . (4.103)
2
This can be written, somewhat less exactly, as
∆E ∆t >∼ h̄, (4.104)
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 ar-
rival time of the particle. Moreover, since E = h̄ ω, the only wave-functions
60
4.9 Eigenstates and Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
which have unique energies are those 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 wave-numbers, and, hence, different frequencies, there will be an
uncertainty ∆E in the particle’s energy which is proportional to the range of fre-
quencies 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 wave-function measured at the fixed point as a function
of time, then we can write
Z∞
1
ψ(t) = √ χ(E) e−i E t/h̄ dE. (4.105)
2π h̄ −∞
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
Z∞
1
χ(E) = √ ψ(t) e i E t/h̄ dt. (4.106)
2π h̄ −∞
For instance, if ψ(t) is a Gaussian then it is easily shown that χ(E) is also a
Gaussian: i.e.,
e−i E0 t/h̄ −(t−t0 ) 2 /4 σt2
ψ(t) = e , (4.107)
(2π σt2 )1/4
e+i E t0 /h̄ −(E−E0 ) 2 /4 σE2
χ(E) = e , (4.108)
(2π σE2 )1/4
where σE σt = h̄/2. As before, Gaussian wave-packets satisfy the minimum un-
certainty principle σE σt = h̄/2. Conversely, non-Gaussian wave-packets are char-
acterized by σE σt > h̄/2.
Consider a general real-space operator A(x). When this operator acts on a gen-
eral wave-function ψ(x) the result is usually a wave-function with a completely
61
4.9 Eigenstates and Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
different shape. However, there are certain special wave-functions which are
such that when A acts on them the result is just a multiple of the original wave-
function. These special wave-functions 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.
where use has been made of Eq. (4.109) and the normalization condition (4.4).
Moreover,
Z∞ Z∞ Z∞
∗ ∗
2
hA i = 2
ψa A ψa dx = a ψa A ψa dx = a 2
ψ∗a ψa dx = a2 , (4.111)
−∞ −∞ −∞
Hence, if ψ1 = ψ2 = ψa then
Z∞ Z∞
ψ∗a (A ψa ) dx = (A ψa )∗ ψa dx, (4.114)
−∞ −∞
62
4.9 Eigenstates and Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
However, from Eq. (4.113), the left-hand sides of the above two equations are
equal. Hence, we can write
Z∞
′
(a − a ) ψ∗a ψa ′ dx = 0. (4.121)
−∞
By assumption, a 6= a ′ , yielding
Z∞
ψ∗a ψa ′ dx = 0. (4.122)
−∞
Consider two eigenstates of A, ψa and ψa′ , which correspond to the same eigen-
value, a. Such eigenstates are termed degenerate. The above proof of the orthog-
onality of different eigenstates fails for degenerate eigenstates. Note, however,
63
4.9 Eigenstates and Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
A ψi = ai ψi , (4.126)
Here, δij is called the Kronecker delta-function, and takes the value unity when its
two indices are equal, and zero otherwise.
64
4.10 Measurement 4 FUNDAMENTALS OF QUANTUM MECHANICS
Thus, the expansion coefficients in Eq. (4.125) are easily determined, given the
wave-function ψ and the eigenstates ψi . Moreover, if ψ is a properly normalized
wave-function then Eqs. (4.125) and (4.127) yield
X
|ci |2 = 1. (4.129)
i
4.10 Measurement
65
4.10 Measurement 4 FUNDAMENTALS OF QUANTUM MECHANICS
For instance, suppose that there are only two eigenstates. The above two
equations then reduce to |c1 |2 = x, and |c2 |2 = 1 − x, where 0 ≤ x ≤ 1, and
(a1 − a2 )2 x (1 − x) = 0. (4.137)
Clearly, the solutions are x = 0 and x = 1. This result can easily be generalized
to the case where there are more than two eigenstates. It follows that a state
associated with a definite value of A is one in which one of the |ci |2 is unity, and
all of the others are zero. In other words, the only states associated with defi-
nite values of A are the eigenstates of A. It immediately follows that the result of
a measurement of A must be one of the eigenvalues of A. Moreover, if a general
wave-function is expanded as a linear combination of the eigenstates of A, like
in Eq. (4.130), then it is clear from Eq. (4.133), and the general definition of a
mean, that the probability of a measurement of A yielding the eigenvalue ai is
simply |ci |2 , where ci is the coefficient in front of the ith eigenstate in the expan-
sion. Note, from Eq. (4.136), that these probabilities are properly normalized:
i.e., the probability of a measurement of A resulting in any possible answer is
unity. Finally, 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 .
66
4.11 Continuous Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
Now, we have already seen, in Sect. 4.8, that if A and B do not commute
(i.e., if A B 6= B A) then they cannot be simultaneously measured. This suggests
that the condition for simultaneous measurement is that A and B should com-
mute. Suppose that this is the case, and that the ψi and ai are the normalized
eigenstates and eigenvalues of A, respectively. It follows that
(A B − B A) ψi = (A B − B ai ) ψi = (A − ai ) B ψi = 0, (4.138)
or
A (B ψi ) = ai (B ψi ). (4.139)
Thus, B ψi is an eigenstate of A corresponding to the eigenvalue ai (though not
necessarily a normalized one). In other words, B ψi ∝ ψi , or
B ψi = bi ψi , (4.140)
In the previous two subsections, it was tacitly assumed that we were dealing with
operators possessing discrete eigenvalues and square-integrable eigenstates. Un-
fortunately, some operators—most notably, x and p—possess eigenvalues which
lie in a continuous range and non-square-integrable eigenstates (in fact, these
two properties go hand in hand). Let us, therefore, investigate the eigenstates
and eigenvalues of the displacement and momentum operators.
67
4.11 Continuous Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
68
4.11 Continuous Eigenvalues 4 FUNDAMENTALS OF QUANTUM MECHANICS
(4.147) and (4.148) are analogous to Eqs. (4.125) and (4.128), respectively, for
square-integrable eigenstates. Finally, by analogy with the results in Sect. 4.9,
the probability density of a measurement of x yielding the value x ′ is |c(x ′ )| 2 ,
which is equivalent to the standard result |ψ(x ′ )| 2 . Moreover, these probabilities
are properly normalized provided ψ(x) is properly normalized [cf., Eq. (4.129)]:
i.e., Z Z
∞ ∞
′ ′
2
|c(x )| dx = |ψ(x ′ )| 2 dx ′ = 1. (4.149)
−∞ −∞
Finally, if a measurement of x yields the value x ′ then the system is left in the
corresponding displacement eigenstate, ψx (x, x ′ ), immediately after the measure-
ment: i.e., the wavefunction collapses to a “spike-function”, δ(x−x ′ ), as discussed
in Sect. 3.13.
Thus, it follows from Eq. (4.74) that the constant of proportionality in Eq. (4.151)
should be (2 π h̄)−1/2 : i.e.,
e i p x/h̄
′
′
ψp (x, p ) = . (4.153)
(2π h̄)1/2
Furthermore, according to Eqs. (4.66) and (4.67),
Z∞
ψ(x) = c(p ′ ) ψp (x, p ′ ) dp ′ , (4.154)
−∞
69
4.12 Stationary States 4 FUNDAMENTALS OF QUANTUM MECHANICS
Problems
1. Monochromatic light with a wavelength of 6000 Å passes through a fast shutter that opens for 10−9
sec. What is the subsequent spread in wavelengths of the no longer monochromatic light? [from
Gaziorowicz]
2. Calculate hxi, hx2i, and σx, as well as hpi, hp2i, and σp, for the normalized wavefunction
s
2 a3 1
ψ(x) = .
π x + a2
2
R∞
Use these to find σx σp. Note that −∞ dx/(x2 + a2) = π/a. [modified from Gaziorowicz]
71
4.12 Stationary States 4 FUNDAMENTALS OF QUANTUM MECHANICS
3. Classically, if a particle is not observed then the probability of finding it in a one-dimensional box
of length L, which extends from x = 0 to x = L, is a constant 1/L per unit length. Show that the
classical expectation
√ value of x is L/2, the expectation value of x2 is L2/3, and the standard deviation
of x is L/ 12. [from Harris]
4. Demonstrate that if a particle in a one-dimensional stationary state is bound then the expectation
value of its momentum must be zero. [from Harris]
R
5. Suppose that V(x) is complex. Obtain an expression for ∂P(x, t)/∂t and d/dt P(x, t) dx from
Schrödinger’s equation. What does this tell us about a complex V(x)? [from Gasiorowicz]
6. ψ1(x) and ψ2(x) are normalized eigenfunctions corresponding to the same eigenvalue. If
Z∞
ψ∗1 ψ2 dx = c,
−∞
where c is real, find normalized linear combinations of ψ1 and ψ2 which are orthogonal to (a) ψ1,
(b) ψ1 + ψ2. [from Squires]
7. Demonstrate that p = −i h̄ ∂/∂x is an Hermitian operator. Find the Hermitian conjugate of a = x+i p.
α is measured and the value a1 is obtained. If β is then measured and then α again, show that the
probability of obtaining a1 a second time is 97/169. [from Squires].
9. Demonstrate that an operator which commutes with the Hamiltonian, and contains no explicit time
dependence, has an expectation value which is constant in time.
10. For a certain system, the operator corresponding to the physical quantity A does not commute with
the Hamiltonian. It has eigenvalues a1 and a2, corresponding to properly normalized eigenfunctions
.√
φ1 = (u1 + u2) 2,
.√
φ2 = (u1 − u2) 2,
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, show that the expectation value of A at
time t is
a1 + a2 a1 − a2 [E1 − E2] t
hAi = + cos .
2 2 h̄
[from Squires]
72
5 ONE-DIMENSIONAL POTENTIALS
5 One-Dimensional Potentials
5.1 Introduction
Consider a particle of mass m and energy E moving in the following simple po-
tential:
0 for 0 ≤ x ≤ a
V(x) = . (5.3)
∞ otherwise
It is clear, from Eq. (5.2), that if d2 ψ/dx2 (and, hence, ψ) is to remain finite then
ψ must go to zero in regions where the potential is infinite. Hence, ψ = 0 in the
regions x ≤ 0 and x ≥ a. Clearly, the problem is equivalent to that of a particle
trapped in a one-dimensional box of length a. The boundary conditions on ψ in
the region 0 < x < a are
ψ(0) = ψ(a) = 0. (5.4)
73
5.2 The Infinite Potential Well 5 ONE-DIMENSIONAL POTENTIALS
74
5.3 The Square Potential Barrier 5 ONE-DIMENSIONAL POTENTIALS
Finally, again from Sect. 4.12, the general time-dependent solution can be writ-
ten as a linear superposition of stationary solutions:
X
ψ(x, t) = cn ψn (x) e−i En t/h̄ , (5.12)
n=0,∞
where Za
cn = ψn (x) ψ(x, 0) dx. (5.13)
0
Consider a particle of mass m and energy E > 0 interacting with the simple
square potential barrier
V0 for 0 ≤ x ≤ a
V(x) = , (5.14)
0 otherwise
where V0 > 0. In the regions to the left and to the right of the barrier, ψ(x)
satisfies
d2 ψ
2
= −k2 ψ, (5.15)
dx
where k is given by Eq. (5.6).
Let us adopt the following solution of the above equation to the left of the
barrier (i.e., x < 0):
ψ(x) = e i k x + R e−i k x . (5.16)
This solution consists of a plane-wave of unit amplitude traveling to the right
[since the time-dependent wave-function is multiplied by a factor exp(−i E t/h̄)],
and a plane-wave of complex amplitude R traveling to the left. We interpret the
first plane-wave as an incoming particle, and the second as a particle reflected by
the potential barrier. Hence, |R| 2 is the probability of reflection. This can be seen
by calculating the probability current (4.19) in the region x < 0, which takes the
form
jl = v (1 − |R| 2 ), (5.17)
75
5.3 The Square Potential Barrier 5 ONE-DIMENSIONAL POTENTIALS
Let us adopt the following solution to Eq. (5.15) to the right of the barrier (i.e.
x > a):
ψ(x) = T e i k x . (5.18)
This solution consists of a plane-wave of complex amplitude T traveling to the
right. We interpret this as a particle transmitted through the barrier. Hence, |T | 2
is the probability of transmission. The probability current in the region x > a
takes the form
jr = v |T | 2 . (5.19)
Now, according to Eq. (4.35), in a stationary state (i.e., ∂|ψ| 2 /∂t = 0), the prob-
ability current is a spatial constant (i.e., ∂j/∂x = 0). Hence, we must have jl = jr ,
or
|R| 2 + |T | 2 = 1. (5.20)
In other words, the probabilities of reflection and transmission sum to unity, as
must be the case, since reflection and transmission are the only possible outcomes
for a particle incident on the barrier.
Let us, first of all, consider the case where E > V0 . In this case, the general
solution to Eq. (5.21) inside the barrier takes the form
Now, the boundary conditions at the edges of the barrier (i.e., at x = 0 and
x = a) are that ψ and dψ/dx are both continuous. These boundary conditions
76
5.3 The Square Potential Barrier 5 ONE-DIMENSIONAL POTENTIALS
ensure that the probability current (4.19) remains finite and continuous across
the edges of the boundary, as must be the case if it is to be a spatial constant.
Continuity of ψ and dψ/dx at the left edge of the barrier (i.e., x = 0) yields
1 + R = A + B, (5.24)
k (1 − R) = q (A − B). (5.25)
Likewise, continuity of ψ and dψ/dx at the right edge of the barrier (i.e., x = a)
gives
A e i q a + B e−i q a = T e i k a , (5.26)
q A e i q a − B e−i q a = k T e i k a.
(5.27)
After considerable algebra, the above four equations yield
2 (k2 − q2 ) 2 sin2 (q a)
|R| = , (5.28)
4 k2 q2 + (k2 − q2 ) 2 sin2 (q a)
and
2 4 k2 q2
|T | = . (5.29)
4 k2 q2 + (k2 − q2 ) 2 sin2 (q a)
Note that the above two expression satisfy the constraint (5.20).
The reflection and transmission probabilities obtained from Eqs. (5.28) and
(5.29), respectively, are plotted in Figs. 7 and 8. It can be seen, from Fig. 7, that
the classical result, |R| 2 = 0 and |T | 2 = 1, is obtained in the limit where the height
of the barrier is relatively small (i.e., V0 ≪ E). However, when V0 is of order E,
there is a substantial probability that the incident particle will be reflected by the
barrier. According to classical physics, reflection is impossible when V0 < E.
77
5.3 The Square Potential Barrier 5 ONE-DIMENSIONAL POTENTIALS
Figure 7: Transmission (solid-curve) and reflection (dashed-curve) probabilities for a square poten-
tial barrier of width a = 1.25 λ, where λ is the free-space de Broglie wave-length, as a function of the
ratio of the height of the barrier, V0, to the energy, E, of the incident particle.
78
5.3 The Square Potential Barrier 5 ONE-DIMENSIONAL POTENTIALS
It can also be seen, from Fig. 8, that at certain barrier widths the probability of
reflection goes to zero. It turns out that this is true irrespective of the energy of
the incident particle. It is clear, from Eq. (5.28), that these special barrier widths
correspond to
q a = n π, (5.30)
where n = 1, 2, 3, · · ·. In other words, the special barriers widths are integer
multiples of half the de Broglie wave-length of the particle inside the barrier.
There is no reflection at the special barrier widths because, at these widths, 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.
Let us, now, consider the case E < V0 . In this case, the general solution to
Eq. (5.21) inside the barrier takes the form
1 + R = A + B, (5.32)
i k (1 − R) = q (A − B). (5.33)
Likewise, continuity of ψ and dψ/dx at the right edge of the barrier (i.e., x = a)
gives
A e q a + B e−q a = T e i k a , (5.34)
q A e q a − B e−q a = i k T e i k a.
(5.35)
2 (k2 + q2 ) 2 sinh2 (q a)
|R| = , (5.36)
4 k2 q2 + (k2 + q2 ) 2 sinh2 (q a)
and
2 4 k2 q2
|T | = . (5.37)
4 k2 q2 + (k2 + q2 ) 2 sinh2 (q a)
79
5.3 The Square Potential Barrier 5 ONE-DIMENSIONAL POTENTIALS
Figure 9: Transmission (solid-curve) and reflection (dashed-curve) probabilities for a square poten-
tial barrier of width a = 0.5 λ, where λ is the free-space de Broglie wave-length, as a function of the
ratio of the energy, E, of the incoming particle to the height, V0, of the barrier.
These expressions can also be obtained from Eqs. (5.28) and (5.29) by making
the substitution q → −i q. Note that Eqs. (5.36) and (5.37) satisfy the constraint
(5.20).
The reflection and transmission probabilities obtained from Eqs. (5.36) and
(5.37), respectively, are plotted in Figs. 9 and 10. It can be seen, from Fig. 9, that
the classical result, |R| 2 = 1 and |T | 2 = 0, is obtained for relatively thin barriers
(i.e., q a ∼ 1) in the limit where the height of the barrier is relatively large (i.e.,
V0 ≫ E). However, when V0 is of order E, there is a substantial probability that
the incident particle will be transmitted by the barrier. According to classical
physics, transmission is impossible when V0 > E.
80
5.4 The WKB Approximation 5 ONE-DIMENSIONAL POTENTIALS
Figure 10: 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, as a function of the ratio of
the width of the barrier, a, to the free-space de Broglie wave-length, λ.
It can also be seen, from Fig. 10, that the transmission probability decays ex-
ponentially as the width of the barrier increases. Nevertheless, even for very wide
barriers (i.e., q a ≫ 1), there is a small but finite probability that a particle inci-
dent on the barrier will be transmitted. This phenomenon, which is inexplicable
within the context of classical physics, is called tunneling.
Consider a particle of mass m and energy E > 0 moving through some slowly
varying potential V(x). The particle’s wave-function satisfies
d2 ψ(x)
2
= −k2 (x) ψ(x), (5.38)
dx
where
2 m [E − V(x)]
k2 (x) = . (5.39)
h̄2
81
5.4 The WKB Approximation 5 ONE-DIMENSIONAL POTENTIALS
According to the WKB solution (5.40), the probability density remains con-
stant: i.e.,
|ψ(x)| 2 = |ψ0 | 2 , (5.44)
as long as the particle moves through a region in which E > V(x), and k(x) is con-
sequently real (i.e., an allowed region according to classical physics). Suppose,
however, that the particle encounters a potential barrier (i.e., a region from which
the particle is excluded according to classical physics). By definition, E < V(x)
inside such a barrier, and k(x) is consequently imaginary. Let the barrier extend
2
After G. Wentzel, H.A. Kramers, and L. Brillouin.
82
5.4 The WKB Approximation 5 ONE-DIMENSIONAL POTENTIALS
from x = x1 to x2 , where 0 < x1 < x2 . The WKB solution inside the barrier is
written Z
x
According to the WKB solution (5.45), the probability density decays exponen-
tially inside the barrier: i.e.,
Zx
2 2 ′ ′
|ψ(x)| = |ψ1 | exp −2 |k(x )| dx , (5.47)
x1
where |ψ1 | 2 is the probability density at the left-hand side of the barrier (i.e.,
x = x1 ). It follows that the probability density at the right-hand side of the
barrier (i.e., x = x2 ) is
Z x2
Note that |ψ2 | 2 < |ψ1 | 2 . Of course, in the region to the right of the barrier (i.e.,
x > x2 ), the probability density takes the constant value |ψ2 | 2 .
We can interpret the ratio of the probability densities to the right and to the
left of the potential barrier as the probability, |T | 2 , that a particle incident from
the left will tunnel through the barrier and emerge on the other side: i.e.,
2
Z x2
|ψ 2 |
|T | 2 = 2
= exp −2 |k(x ′ )| dx ′ (5.49)
|ψ1 | x1
(see Sect. 5.3). It is easily demonstrated that the probability of a particle incident
from the right tunneling through the barrier is the same.
Note that the criterion (5.43) for the validity of the WKB approximation im-
plies that the above transmission probability is very small. Hence, the WKB ap-
proximation only applies to situations in which there is very little chance of a
83
5.5 Cold Emission 5 ONE-DIMENSIONAL POTENTIALS
It can be seen, from Fig. 11, that an electron just below the surface of the
metal is confined by a triangular potential barrier which extends from x = x1 to
x2 , where x1 = 0 and x2 = W/e E. Making use of the WKB approximation (see
the previous subsection), the probability of such an electron tunneling through
the barrier, and consequently being emitted from the surface, is
√ Z x2 q
2 2 m
|T | 2 = exp − V(x) − E dx , (5.50)
h̄ x1
or √ Z W/e E
√
2 2 m
|T | 2 = exp − W − e E x dx . (5.51)
h̄ 0
84
5.5 Cold Emission 5 ONE-DIMENSIONAL POTENTIALS
1111111111111
0000000000000 x→
0000000000000
1111111111111
Energy →
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111 V − E = W − eE x
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
METAL VACUUM
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
E 1111111111111
0000000000000
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111
0000000000000
1111111111111 x1 x2
Figure 11: The potential barrier for an electron in a metal surface subject to an external electric field.
This reduces to
√ m1/2 W 3/2 Z 1 q
2
|T | = exp −2 2 1 − y dy , (5.52)
h̄ e E 0
or √ 1/2 3/2
4 2 m W
|T | 2 = exp − . (5.53)
3 h̄ e E
The above result is known as the Fowler-Nordheim formula. Note that the prob-
ability of emission increases exponentially as the electric field-strength above the
surface of the metal increases.
The cold emission of electrons from a metal surface is the basis of an important
device known as a scanning tunneling microscope, or an STM. An STM consists of a
very sharp conducting probe which is scanned over the surface of a metal (or any
other solid conducting medium). A large voltage difference is applied between
the probe and the surface. Now, the surface electric field-strength immediately
below the probe tip is proportional to the applied potential difference, and in-
versely proportional to the spacing between the tip and the surface. Electrons
tunneling between the surface and the probe tip give rise to a weak electric cur-
rent. The magnitude of this current is proportional to the tunneling probability
85
5.6 α-Decay 5 ONE-DIMENSIONAL POTENTIALS
(5.53). It follows that the current is an extremely sensitive function of the surface
electric field-strength, and, hence, of the spacing between the tip and the surface
(assuming that the potential difference is held constant). An STM can thus be
used to construct a very accurate contour map of the surface under investigation.
In fact, STMs are capable of achieving sufficient resolution to image individual
atoms
5.6 α-Decay
for Z ≫ 1.
Making use of the WKB approximation (and neglecting the fact that r is a
radial, rather than a Cartesian, coordinate), the probability of the α-particle tun-
86
5.6 α-Decay 5 ONE-DIMENSIONAL POTENTIALS
where 1/2
Z1 Z2 e2 m R
2/3
β= = 0.74 Z1 (5.58)
4π ǫ0 h̄2
is a dimensionless constant, and
Z1 Z2 e2 2/3
Ec = = 1.44 Z1 MeV (5.59)
4π ǫ0 R
is the characteristic energy the α-particle would need in order to escape from the
nucleus without tunneling. Of course, E ≪ Ec . It is easily demonstrated that
Z 1/ǫ " #1/2
1 π
−ǫ dy ≃ √ − 2 (5.60)
1 y 2 ǫ
when ǫ ≪ 1. Hence.
v
√ π t Ec
u
|T | 2 ≃ exp
u
−2 2β − 2
. (5.61)
2 E
Now,
q
the α-particle moves inside the nucleus with the characteristic velocity
v = 2 E/m. It follows that the particle bounces backward and forward within
the nucleus at the frequency ν ≃ v/R, giving
for a 1 MeV α-particle trapped inside a typical heavy nucleus of radius 10−14 m.
Thus, the α-particle effectively attempts to tunnel through the potential barrier
87
5.6 α-Decay 5 ONE-DIMENSIONAL POTENTIALS
88
5.6 α-Decay 5 ONE-DIMENSIONAL POTENTIALS
Figure 12: The experimentally determined half-life, τex, of various atomic nucleii which decay
√ via α
2/3
emission versus the best-fit theoretical half-life log10(τth) = −28.9 − 1.60 Z1 + 1.61 Z1/ E. Both
half-lives are measured in years. Here, Z1 = Z − 2, where Z is the charge number of the nucleus,
and E the characteristic energy of the emitted α-particle in MeV. In order of increasing half-life, the
points correspond to the following nucleii: Rn 215, Po 214, Po 216, Po 197, Fm 250, Ac 225, U 230,
U 232, U 234, Gd 150, U 236, U 238, Pt 190, Gd 152, Nd 144. Data obtained from IAEA Nuclear
Data Centre.
89
5.7 The Square Potential Well 5 ONE-DIMENSIONAL POTENTIALS
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) = , (5.73)
0 otherwise
where V0 > 0.
Now, if E > 0 then the particle is unbounded. Thus, when the particle en-
counters the well it is either reflected or transmitted. As is easily demonstrated,
the reflection and transmission probabilities are given by Eqs. (5.28) and (5.29),
respectively, where
2mE
k2 = , (5.74)
h̄2
2 m (E + V0 )
q2 = . (5.75)
h̄2
Suppose, however, that E < 0. In this case, the particle is bounded (i.e.,
2
|ψ| → 0 as |x| → ∞). Is is possible to find bounded solutions of Schrödinger’s
equation in the finite square potential well (5.73)?
Let us, first of all, search for a totally symmetric solution. In the region to the
left of the well (i.e. x < −a/2), the solution of Schrödinger’s equation which
satisfies the boundary condition ψ → 0 and x → −∞ is
ψ(x) = A e k x , (5.77)
90
5.7 The Square Potential Well 5 ONE-DIMENSIONAL POTENTIALS
where
2 m |E|
k2 = . (5.78)
h̄2
By symmetry, the solution in the region to the right of the well (i.e., x > a/2) is
ψ(x) = A x−k x . (5.79)
The solution inside the well (i.e., |x| ≤ a/2) which satisfies the symmetry con-
straint ψ(−x) = ψ(x) is
ψ(x) = B cos(q x), (5.80)
where
2 m (V0 + E)
q2 = . (5.81)
h̄2
Here, we have assumed that E > −V0 . The constraint that ψ(x) and its first
derivative be continuous at the edges of the well (i.e., at x = ±a/2) yields
k = q tan(q a/2). (5.82)
q Now, the solutions to Eq. (5.85) correspond to the intersection of the curve
λ − y2 /y with the curve tan y. Figure 13 shows these two curves plotted for
91
5.7 The Square Potential Well 5 ONE-DIMENSIONAL POTENTIALS
q
Figure 13: The curves tan y (solid) and λ − y2/y (dashed), calculated for λ = 1.5 π2. The latter
√
curve takes the value 0 when y > λ.
a particular value of λ. In this case, the curves intersect twice, indicating the
existence of two totally symmetric bound states in the well. Moreover, it is clear,
from the figure, that as λ increases (i.e., as the well becomes deeper), there are
more and more bound states. However, it is also clear that there is always at
least one totally symmetric bound state, no matter how small λ becomes (i.e., no
matter how shallow the well becomes). In the limit λ ≫ 1 (i.e., the limit in which
the well becomes very deep), the solutions to Eq. (5.85) asymptote to the roots
of tan y = ∞. This gives y = (2 j − 1) π/2, where j is a positive integer, or
(2 j − 1) π
q= . (5.87)
a
These solutions are equivalent to the odd-n infinite square well solutions specified
by Eq. (5.8).
For the case of a totally anti-symmetric bound state, similar analysis to the
above yields
y
−q = tan y. (5.88)
λ − y2
The solutions of thisqequation correspond to the intersection of the curve tan y
with the curve −y/ λ − y2 . Figure 14 shows these two curves plotted for the
92
5.7 The Square Potential Well 5 ONE-DIMENSIONAL POTENTIALS
q
Figure 14: The curves tan y (solid) and −y/ λ − y2 (dashed), calculated for λ = 1.5 π2.
same value of λ as that used in Fig. 13. In this case, the curves intersect once,
indicating the existence of a single totally anti-symmetric bound state in the well.
It is, again, clear, from the figure, that as λ increases (i.e., as the well becomes
deeper), there are more and more bound states. However, it is also clear that
when λ becomes sufficiently small [i.e., λ < (π/2)2 ] then there is no totally anti-
symmetric bound state. In other words, a very shallow potential well always
possesses a totally symmetric bound state, but does not generally possess a totally
anti-symmetric bound state. In the limit λ ≫ 1 (i.e., the limit in which the well
becomes very deep), the solutions to Eq. (5.88) asymptote to the roots of tan y =
0. This gives y = j π, where j is a positive integer, or
2jπ
q= . (5.89)
a
These solutions are equivalent to the even-n infinite square well solutions speci-
fied by Eq. (5.8).
93
5.8 The Simple Harmonic Oscillator 5 ONE-DIMENSIONAL POTENTIALS
Consider the behavior of the solution to Eq. (5.94) in the limit |y| ≫ 1. As is
easily seen, in this limit, the equation simplifies somewhat to give
d2 ψ
2
− y2 ψ ≃ 0. (5.95)
dy
The approximate solutions to the above equation are
2
ψ(y) ≃ A(y) e±y /2
, (5.96)
94
5.8 The Simple Harmonic Oscillator 5 ONE-DIMENSIONAL POTENTIALS
Inserting this test solution into Eq. (5.98), and equating the coefficients of yi , we
obtain the recursion relation
(2 i − ǫ + 1)
ai+2 = ai . (5.100)
(i + 1) (i + 2)
Consider the behavior of h(y) in the limit |y| → ∞. The above recursion relation
simplifies to
2
ai+2 ≃ ai . (5.101)
i
Hence, at large |y|, when the higher powers of y dominate, we have
X y2 j 2
h(y) ∼ C ∼ C ey . (5.102)
j
j!
It follows that ψ(y) = h(y) exp(−y2 /2) varies as exp( y2 /2) as |y| → ∞. This
behavior is unacceptable, since it does not satisfy the boundary condition ψ → 0
as |y| → ∞. The only way in which we can prevent ψ from blowing up as |y| → ∞
is to demand that the power series (5.99) terminate at some finite value of i. This
implies, from the recursion relation (5.100), that
ǫ = 2 n + 1, (5.103)
95
5.8 The Simple Harmonic Oscillator 5 ONE-DIMENSIONAL POTENTIALS
where n is a non-negative integer. Note that the number of terms in the power
series (5.99) is n + 1. Finally, using Eq. (5.93), we obtain
E = (n + 1/2) h̄ ω, (5.104)
for n = 0, 1, 2, · · ·.
96
5.8 The Simple Harmonic Oscillator 5 ONE-DIMENSIONAL POTENTIALS
Using these operators, Eq. (5.108) can also be written in the forms
a + a − ψn = n ψ n , (5.111)
or
a− a+ ψn = (n + 1) ψn . (5.112)
The above two equations imply that
√
a + ψn = n + 1 ψn+1 , (5.113)
√
a − ψn = n ψn−1 . (5.114)
We conclude that a+ and a− are raising and lowering operators, respectively, for
the harmonic oscillator: i.e., operating on the wave-function with a+ causes the
quantum number n to increase by unity, and vice versa. The Hamiltonian for the
harmonic oscillator can be written in the form
1
!
H = h̄ ω a+ a− + , (5.115)
2
from which the result
H ψn = (n + 1/2) h̄ ω ψn = En ψn (5.116)
is readily deduced. Finally, Eqs. (5.107), (5.113), and (5.114) yield the useful
expression
Z∞ Z
a ∞
ψm x ψn dx = √ ψm (a+ + a− ) ψn dx
−∞ 2 −∞
v
h̄ √ √
u
u
= t
m δm,n+1 + n δm,n−1 . (5.117)
2mω
Problems
2. A particle of mass m moves freely in one dimension between impenetrable walls located at x = 0
and a. Its initial wavefunction is
q
ψ(x, 0) = 2/a sin(3π x/a).
97
5.8 The Simple Harmonic Oscillator 5 ONE-DIMENSIONAL POTENTIALS
What is the subsequent time evolution of the wavefunction? Suppose that the initial wavefunction is
q
ψ(x, 0) = 1/a sin(π x/a) [1 + 2 cos(π x/a)].
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.
4. A stream of particles of mass m and energy E > 0 encounter a potential step of height W(< E): i.e.,
V(x) = 0 for x < 0 and V(x) = W for x > 0 with the particles incident from −∞. Show that the
fraction reflected is
k−q 2
R= ,
k+q
where k2 = (2m/h̄2) E and q 2 = (2m/h̄2) (E − W). [from Squires]
5. A stream of particles of mass m and energy E > 0 encounter the delta-function potential V(x) =
−α δ(x), where α > 0. Show that the fraction reflected is
R = β2/(1 + β2),
where β = m α/h̄2 k, and k2 = (2m/h̄2) E. Does such a potential have a bound state? If so, what is
its energy? [modified from Griffiths]
6. Two potential wells of width a are separated by a distance L ≫ a. A particle of mass m and energy
E is in one of the wells. Estimate the time required for the particle to tunnel to the other well.
where V0 > 0. Demonstrate that the bound-states of a particle of mass m and energy −V0 < E < 0
satisfy q q
tan 2 m (V0 + E) L/h̄ = − (V0 + E)/(−E).
[from Harris]
8. Find the properly normalized first two excited energy eigenstates of the harmonic oscillator, as well
as the expectation value of the potential energy in the nth energy eigenstate. Hint: Consider the
raising and lowering operators a± defined in Eq. (5.109).
98
6 MULTI-PARTICLE SYSTEMS
6 Multi-Particle Systems
6.1 Introduction
99
6.2 Fundamental Concepts 6 MULTI-PARTICLE SYSTEMS
Since the xi are independent variables (i.e., ∂xi /∂xj = δij ), we conclude that
the various position and momentum operators satisfy the following commutation
relations:
[xi , xj ] = 0, (6.4)
[pi , pj ] = 0, (6.5)
[xi , pj ] = i h̄ δij . (6.6)
Now, we know, from Sect. 4.10, that two dynamical variables can only be (ex-
actly) measured simultaneously if the operators which represent them in quantum
mechanics commute with one another. Thus, it is clear, from the above commu-
tation relations, that the only restriction on measurement in a one-dimensional
multi-particle system is that it is impossible to simultaneously measure the posi-
tion and momentum of the same particle. Note, in particular, 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. The commutation relations (6.4)–(6.6)
illustrate an important point in quantum mechanics: namely, that operators cor-
responding to different degrees of freedom of a dynamical system tend to commute
with one another. In this case, the different degrees of freedom correspond to the
different motions of the various particles making up the system.
Here, the first term on the right-hand side represents the total kinetic energy
of the system, whereas the potential V specifies the nature of the interaction
between the various particles making up the system, as well as the interaction of
the particles with any external forces.
Suppose that the particles do not interact with one another. This implies that
each particle moves in a common potential: i.e.,
X
V(x1 , x2 , . . . , xN , t) = V(xi , t). (6.11)
i=1,N
where
pi2
Hi =+ V(xi , t). (6.13)
2 mi
In other words, for the case of non-interacting particles, the multi-particle Hamil-
tonian of the system can be written as the sum of N independent single-particle
101
6.3 Non-Interacting Particles 6 MULTI-PARTICLE SYSTEMS
Hamiltonians. Here, Hi represents the energy of the ith particle, and is com-
pletely unaffected by the energies of the other particles. Furthermore, given that
the various particles which make up the system are non-interacting, we expect
their instantaneous positions to be completely uncorrelated with one another. This
immediately implies that the multi-particle wave-function ψ(x1 , x2 , . . . xN , t) can
be written as the product of N independent single-particle wave-functions: i.e.,
Here, |ψi (xi , t)|2 dxi is the probability of finding the ith particle between xi and
xi + dxi at time t. This probability is completely unaffected by the positions
of the other particles. It is evident that ψi (xi , t) must satisfy the normalization
constraint Z∞
|ψi (xi , t)|2 dxi = 1. (6.15)
−∞
If this is the case, then the normalization constraint (6.2) for the multi-particle
wave-function is automatically satisfied. Equation (6.14) illustrates an important
point in quantum mechanics: namely, that we can generally write the total wave-
function of a many degree of freedom system as a product of different wave-
functions corresponding to each degree of freedom.
where
ψE (x1 , x2 , . . . , xN ) = ψE1 (x1 ) ψE2 (x2 ) . . . ψEN (xN ), (6.19)
102
6.4 Two-Particle Systems 6 MULTI-PARTICLE SYSTEMS
and X
E= Ei . (6.20)
i=1,N
Clearly, for the case of non-interacting particles, the energy of the whole system
is simply the sum of the energies of the component particles.
Consider a system consisting of two particles, mass m1 and m2 , interacting via the
potential V(x1 − x2 ) which only depends on the relative positions of the particles.
According to Eqs. (6.3) and (6.10), the Hamiltonian of the system is written
h̄2 ∂2 h̄2 ∂2
H(x1 , x2 ) = − − + V(x1 − x2 ). (6.21)
2 m1 ∂x12 2 m2 ∂x22
Let
x ′ = x1 − x2 (6.22)
be the particles’ relative position, and
m1 x1 + m2 x2
X= (6.23)
m1 + m2
the position of the center of mass. It is easily demonstrated that
∂ m1 ∂ ∂
= + ′, (6.24)
∂x1 m1 + m2 ∂X ∂x
∂ m2 ∂ ∂
= − ′. (6.25)
∂x2 m1 + m2 ∂X ∂x
Hence, when expressed in terms of the new variables, x ′ and X, the Hamiltonian
becomes
′ h̄2 ∂2 h̄2 ∂2
H(x , X) = − 2
− ′ 2
+ V(x ′ ), (6.26)
2 M ∂X 2 µ ∂x
where
M = m1 + m2 (6.27)
103
6.4 Two-Particle Systems 6 MULTI-PARTICLE SYSTEMS
The fact that the Hamiltonian (6.26) is separable when expressed in terms of
the new coordinates [i.e., H(x ′ , X) = Hx ′ (x ′ ) + HX (X)] suggests, by analogy with
the analysis in the previous subsection, that the wave-function can be factorized:
i.e.,
ψ(x1 , x2 , t) = ψx ′ (x ′ , t) ψX (X, t). (6.30)
Hence, the time-dependent Schrödinger equation (6.7) also factorizes to give
∂ψx ′ h̄2 ∂2 ψx ′
i h̄ =− ′ 2
+ V(x ′ ) ψx ′ , (6.31)
∂t 2 µ ∂x
and
∂ψX h̄2 ∂2 ψX
i h̄ =− . (6.32)
∂t 2 M ∂X2
The above equation can be solved to give
ψX (X, t) = ψ0 e i (P
′
X/h̄−E ′ t/h̄)
, (6.33)
where ψ0 , P ′ , and E ′ = P ′ 2 /2 M are constants. It is clear, from Eqs. (6.29), (6.30),
and (6.33), that the total momentum of the system takes the constant value P ′ :
i.e., momentum is conserved.
Suppose that we work in the centre of mass frame of the system, which is
characterized by P ′ = 0. It follows that ψX = ψ0 . In this case, we can write the
wave-function of the system in the form ψ(x1 , x2 , t) = ψx ′ (x ′ , t) ψ0 ≡ ψ(x1 −x2 , t),
where
∂ψ h̄2 ∂2 ψ
i h̄ =− + V(x) ψ. (6.34)
∂t 2 µ ∂x 2
In other words, in the center of mass frame, two particles of mass m1 and m2 ,
moving in the potential V(x1 − x2 ), are equivalent to a single particle of mass µ,
104
6.5 Identical Particles 6 MULTI-PARTICLE SYSTEMS
or
ψ(x1 , x2 , t) = e i ϕ ψ(x2 , x1 , t), (6.36)
where ϕ is a real constant. However, if we swap the labels on particles 1 and 2
(which are, after all, arbitrary for identical particles), and repeat the argument,
we also conclude that
Hence,
e 2 i ϕ = 1. (6.38)
The only solutions to the above equation are ϕ = 0 and ϕ = π. Thus, we infer
that for a system consisting of two identical particles, the wave-function must
be either symmetric or anti-symmetric under interchange of particle labels: i.e.,
either
ψ(x2 , x1 , t) = ψ(x1 , x2 , t), (6.39)
or
ψ(x2 , x1 , t) = −ψ(x1 , x2 , t). (6.40)
105
6.5 Identical Particles 6 MULTI-PARTICLE SYSTEMS
The above argument can easily be extended to systems containing more than two
identical particles.
It turns out that whether the wave-function of a system containing many iden-
tical particles is symmetric or anti-symmetric under interchange of the labels on
any two particles is determined by the nature of the particles themselves. Par-
ticles with wave-functions which are symmetric under label interchange are said
to obey Bose-Einstein statistics, and are called bosons—for instance, photons are
bosons. Particles with wave-functions which are anti-symmetric under label in-
terchange are said to obey Fermi-Dirac statistics, and are called fermions—for
instance, electrons, protons, and neutrons are fermions.
For a system consisting of two identical and non-interacting fermions, the sta-
tionary wave-function of the whole system takes the form
1
ψE fermion (x1 , x2 ) = √ [ψ(x1 , Ea ) ψ(x2 , Eb ) − ψ(x2 , Ea ) ψ(x1 , Eb )] , (6.42)
2
Again, this expression automatically satisfies the symmetry requirement on the
wave-function. Note that if Ea = Eb then the total wave-function becomes zero
everywhere. Now, in quantum mechanics, a null wave-function corresponds to
the absence of a state. We thus conclude that it is impossible for the two fermions
in our system to occupy the same single-particle stationary state.
Finally, if the two particles are somehow distinguishable then the stationary
106
6.5 Identical Particles 6 MULTI-PARTICLE SYSTEMS
Let us evaluate the variance of the distance, x1 − x2 , between the two parti-
cles, using the above three wave-functions. It is easily demonstrated that if the
particles are distinguishable then
where Z∞
n
hx ia,b = ψ∗ (x, Ea,b ) xn ψ(x, Ea,b ) dx. (6.45)
−∞
For the case of two identical bosons, we find
where Z∞
hxiab = ψ∗ (x, Ea ) x ψ(x, Eb ) dx. (6.47)
−∞
Here, we have assumed that Ea 6= Eb , so that
Z∞
ψ∗ (x, Ea ) ψ(x, Eb ) dx = 0. (6.48)
−∞
Equation (6.46) shows that the symmetry requirement on the total wave-function
of two identical bosons forces the particles to be, on average, closer together than
two similar distinguishable particles. Conversely, Eq. (6.49) shows that the sym-
metry requirement on the total wave-function of two identical fermions forces the
particles to be, on average, further apart than two similar distinguishable parti-
cles. However, the strength of this effect depends on square of the magnitude
of hxiab , which measures the overlap between the wave-functions ψ(x, Ea ) and
ψ(x, Eb ). It is evident, then, that if these two wave-functions do not overlap to
107
6.5 Identical Particles 6 MULTI-PARTICLE SYSTEMS
any great extent then identical bosons or fermions will act very much like distin-
guishable particles.
This expression is known as the Slater determinant, and automatically satisfies the
symmetry requirements on the wave-function. Here, the energies of the particles
are E1 , E2 , . . . , EN . Note, again, that if any two particles in the system have the
same energy (i.e., if Ei = Ej for some i 6= j) then the total wave-function is null.
We conclude that it is impossible for any two identical fermions in a multi-particle
system to occupy the same single-particle stationary state. This important result
is known as the Pauli exclusion principle.
1. Consider a system consisting of two non-interacting particles, and three one-particle states, ψa(x),
ψb(x), and ψc(x). How many different two-particle states can be constructed if the particles are
(a) distinguishable, (b) indistinguishable bosons, or (c) indistinguishable fermions? [modified from
Griffiths]
3. Two non-interacting particles, with the same mass m, are in a one-dimensional box of length a.
What are the four lowest energies of the system? What are the degeneracies of these energies if the
two particles are (a) distinguishable, (b) indistinguishable bosons, or (c) indistingishable fermions?
[modified from Squires]
4. Two particles in a one-dimensional box of length a occupy the n = 4 and n ′ = 3 states. Write
the properly normalized wave-functions if the particles are (a) distinguishable, (b) indistinguishable
bosons, or (c) indistinguishable fermions. [modified from Harris]
108
7 THREE-DIMENSIONAL QUANTUM MECHANICS
7.1 Introduction
We have seen that in one dimension the instantaneous state of a single non-
relativistic particle is fully specified by a complex wave-function, ψ(x, t). The
probability of finding the particle at time t between x and x + dx is P(x, t) dx,
where
P(x, t) = |ψ(x, t)|2 . (7.1)
Moreover, the wave-function is normalized such that
Z∞
|ψ(x, t)|2 dx = 1 (7.2)
−∞
at all times.
This normalization constraint ensures that the probability of finding the particle
anywhere is space is always unity.
109
7.2 Fundamental Concepts 7 THREE-DIMENSIONAL QUANTUM MECHANICS
In three dimensions, by analogy with the one dimensional case, the probability
conservation equation becomes
∂|ψ|2 ∂jx ∂jy ∂jz
+ + + = 0. (7.8)
∂t ∂x ∂y ∂z
Here,
i h̄ ∂ψ∗ ∂ψ
!
jx = ψ − ψ∗ (7.9)
2m ∂x ∂x
is the flux of probability along the x-axis, and
i h̄ ∂ψ∗ ∂ψ
!
jy = ψ − ψ∗ (7.10)
2m ∂y ∂y
the flux of probability along the y-axis, etc. Integrating Eq. (7.8) over all space,
and making use of the fact that ψ → 0 as |r| → ∞ if ψ is to be square-integrable,
we obtain Z Z Z
d ∞ ∞ ∞
|ψ(x, y, z, t)|2 dx dy dz = 0. (7.11)
dt −∞ −∞ −∞
Thus, the normalization of the wave-function is again preserved as time pro-
gresses, as must be the case if |ψ|2 is to be interpreted as a probability density.
110
7.2 Fundamental Concepts 7 THREE-DIMENSIONAL QUANTUM MECHANICS
[xi , xj ] = 0, (7.15)
[pi , pj ] = 0, (7.16)
[xi , pj ] = i h̄ δij . (7.17)
Now, we know, from Sect. 4.10, that two dynamical variables can only be (ex-
actly) measured simultaneously if the operators which represent them in quantum
mechanics commute with one another. Thus, it is clear, from the above commuta-
tion relations, that the only restriction on measurement in a system consisting of
a single particle moving in three dimensions is that it is impossible to simultane-
ously measure a given position coordinate and the corresponding component of
momentum. Note, however, that it is perfectly possible to simultaneously mea-
sure two different positions coordinates, or two different components of the mo-
mentum. The commutation relations (7.15)–(7.17) again illustrate the point that
quantum mechanical operators corresponding to different degrees of freedom of
a dynamical system (in this case, motion in different directions) tend to commute
with one another (see Sect. 6.2).
111
7.2 Fundamental Concepts 7 THREE-DIMENSIONAL QUANTUM MECHANICS
112
7.3 Particle in a Box 7 THREE-DIMENSIONAL QUANTUM MECHANICS
Let us search for a separable solution to the above equation of the form
ψ(x, y, z) = X(x) Y(y) Z(z). (7.27)
The factors of the wave-function satisfy the boundary conditions X(0) = X(a) =
0, Y(0) = Y(a) = 0, and Z(0) = Z(a) = 0. Substituting (7.27) into Eq. (7.26),
and rearranging, we obtain
X ′′ Y ′′ Z ′′ 2m
+ + = − 2 E, (7.28)
X Y Z h̄
where ′ denotes a derivative with respect to argument. It is evident that the only
way in which the above equation can be satisfied at all points within the box is if
X ′′
= −kx2 , (7.29)
X
Y ′′
= −ky2 , (7.30)
Y
Z ′′
= −kz2 , (7.31)
Z
where kx2 , ky2 , and kz2 are spatial constants. Note that the right-hand sides of the
above equations must contain negative, rather than positive, spatial constants,
because it would not otherwise be possible to satisfy the boundary conditions.
The solutions to the above equations which are properly normalized, and satisfy
the boundary conditions, are [see Eq. (5.11)]
v
u2
u
X(x) = t
sin(kx x), (7.32)
a
113
7.4 Degenerate Electron Gases 7 THREE-DIMENSIONAL QUANTUM MECHANICS
v
u2
u
Y(y) = t
sin(ky y), (7.33)
a
v
u2
u
Z(z) = t
sin(kz z), (7.34)
a
where
lx π
kx = , (7.35)
a
ly π
ky = , (7.36)
a
lz π
kz = . (7.37)
a
Here, lx , ly , and lz are positive integers. Thus, from Eqs. (7.28)–(7.31), the energy
is written [see Eq. (5.9)]
l2 π2 h̄2
E= . (7.38)
2 m a2
where
l2 = lx2 + ly2 + lz2 . (7.39)
114
7.4 Degenerate Electron Gases 7 THREE-DIMENSIONAL QUANTUM MECHANICS
momentum called spin (see Sect. 10). The spin states of an electron are governed
by an additional quantum number, which can take one of two different values.
Hence, when spin is taken into account, we conclude that a maximum of two
electrons (with different spin quantum numbers) can occupy a single-particle
energy level corresponding to a particular set of values of lx , ly , and lz . Note,
from Eqs. (7.38) and (7.39), that the associated particle energy is proportional
to l2 = lx2 + ly2 + lz2 .
Suppose that our electrons are cold: i.e., they have comparatively little thermal
energy. In this case, we would expect them to fill the lowest single-particle energy
levels available to them. We can imagine the single-particle energy levels as
existing in a sort of three-dimensional quantum number space whose Cartesian
coordinates are lx , ly , and lz . Thus, the energy levels are uniformly distributed in
this space on a cubic lattice. Moreover, the distance between nearest neighbour
energy levels is unity. This implies that the number of energy levels per unit
volume is also unity. Finally, the energy of a given energy level is proportional to
its distance, l2 = lx2 + ly2 + lz2 , from the origin.
Since we expect cold electrons to occupy the lowest energy levels available
to them, but only two electrons can occupy a given energy level, it follows that
if the number of electrons, N, is very large then the filled energy levels will be
approximately distributed in a sphere centered on the origin of quantum number
space. The number of energy levels contained in a sphere of radius l is approxi-
mately equal to the volume of the sphere—since the number of energy levels per
unit volume is unity. It turns out that this is not quite correct, because we have
forgotten that the quantum numbers lx , ly , and lz can only take positive values.
Hence, the filled energy levels actually only occupy one octant of a sphere. 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, N. Thus, we can write
1 4π 3
N=2× × l . (7.40)
8 3 F
Here, the factor 2 is to take into account the two spin states of an electron, and
the factor 1/8 is to take account of the fact that lx , ly , and lz can only take positive
115
7.4 Degenerate Electron Gases 7 THREE-DIMENSIONAL QUANTUM MECHANICS
values. Thus,
3 N 1/3
!
lF = . (7.41)
π
According to Eq. (7.38), the energy of the most energetic electrons—which is
known as the Fermi energy—is given by
lF2 π2 h̄2 π2 h̄2 3 N 2/3
!
EF = = , (7.42)
2 me a2 2 m a2 π
where me is the electron mass. This can also be written as
π2 h̄2 3 n
!2/3
EF = , (7.43)
2 me π
where n = N/a3 is the number of electrons per unit volume (in real space).
Note that the Fermi energy only depends on the number density of the confined
electrons.
since EF ∝ a−2 = V −2/3 [see Eq. (7.42)]. Now, the pressure predicted by classical
physics is P = n kB T . Thus, a degenerate electron gas has a much higher pressure
than that which would be predicted by classical physics. This is an entirely quan-
tum mechanical effect, and is due to the fact that identical fermions cannot get
significantly closer together than a de Broglie wave-length without violating the
Pauli exclusion principle. Note that, according to Eq. (7.43), the mean spacing
between degenerate electrons is
h h
d ∼ n−1/3 ∼ √ ∼ ∼ λ, (7.47)
me E p
where λ is the de Broglie wave-length. Thus, an electron gas is non-degenerate
when the mean spacing between the electrons is much greater than the de Broglie
wave-length, and becomes degenerate as the mean spacing approaches the de
Broglie wave-length.
In turns out that the conduction (i.e., free) electrons inside metals are highly
degenerate (since the number of electrons per unit volume is very large, and
EF ∝ n2/3 ). Indeed, most metals are hard to compress as a direct consequence
of the high degeneracy pressure of their conduction electrons. To be more exact,
resistance to compression is usually measured in terms of a quantity known as
the bulk modulus, which is defined
∂P
B = −V (7.48)
∂V
Now, for a fixed number of electrons, P ∝ V −5/3 [see Eqs. (7.42) and (7.46)].
Hence,
5 π3 h̄2 3 n 5/3
!
B= P= . (7.49)
3 9m π
For example, the number density of free electrons in magnesium is n ∼ 8.6 ×
1028 m−3 . This leads to the following estimate for the bulk modulus: B ∼ 6.4 ×
1010 N m−3 . The actual bulk modulus is B = 4.5 × 1010 N m−3 .
117
7.5 White-Dwarf Stars 7 THREE-DIMENSIONAL QUANTUM MECHANICS
A burnt-out star is basically a gas of electrons and ions. As the star collapses,
its density increases, and so the mean separation between its constituent parti-
cles decreases. Eventually, the mean separation becomes of order the de Broglie
wave-length of the electrons, and the electron gas becomes degenerate. Note,
that the de Broglie wave-length of the ions is much smaller than that of the elec-
trons, so the ion gas remains non-degenerate. Now, even at zero temperature,
a degenerate electron gas exerts a substantial pressure, because the Pauli exclu-
sion principle prevents the mean electron separation from becoming significantly
smaller than the typical de Broglie wave-length (see the previous subsection).
Thus, it is possible for a burnt-out star to maintain itself against complete col-
lapse under gravity via the degeneracy pressure of its constituent electrons. Such
stars are termed white-dwarfs. Let us investigate the physics of white-dwarfs in
more detail.
118
7.5 White-Dwarf Stars 7 THREE-DIMENSIONAL QUANTUM MECHANICS
From the previous subsection, the kinetic energy of a degenerate electron gas
is simply
3 3 π2 h̄2 3 N 2/3
!
K = N Ē = N EF = N , (7.52)
5 5 2 me π V
where N is the number of electrons, V the volume of the star, and me the electron
mass.
The interior of a white-dwarf star is composed of atoms like C12 and O16 which
contain equal numbers of protons, neutrons, and electrons. Thus,
M = 2 N mp , (7.53)
where R⊙ = 7 × 105 km is the solar radius, and M⊙ = 2 × 1030 kg the solar mass. It
follows that the radius of a typical solar mass white-dwarf is about 7000 km: i.e.,
about the same as the radius of the Earth. The first white-dwarf to be discovered
(in 1862) was the companion of Sirius. Nowadays, thousands of white-dwarfs
have been observed, all with properties similar to those described above.
Note from Eqs. (7.52), (7.53), and (7.59) that Ē ∝ M4/3 . In other words, the
mean energy of the electrons inside a white dwarf increases as the stellar mass in-
creases. Hence, for a sufficiently massive white dwarf, the electrons can become
relativistic. It turns out that the degeneracy pressure for relativistic electrons only
scales as R−1 , rather that R−2 , and thus is unable to balance the gravitational pres-
sure [which also scales as R−1 —see Eq. (7.54)]. It follows that electron degener-
acy pressure is only able to halt the collapse of a burnt-out star provided that the
stellar mass does not exceed some critical value, known as the Chandrasekhar
limit, which turns out to be about 1.4 times the mass of the Sun. Stars whose
mass exceeds the Chandrasekhar limit inevitably collapse to produce extremely
compact objects, such as neutron stars (which are held up by the degeneracy
pressure of their constituent neutrons), or black holes.
Problems
120
8 ORBITAL ANGULAR MOMENTUM
8.1 Introduction
121
8.2 Angular Momentum Operators 8 ORBITAL ANGULAR MOMENTUM
Let us now derive the commutation relations for the Li . For instance,
where use has been made of the definitions of the Li [see Eqs. (8.2)–(8.4)], and
commutation relations (7.15)–(7.17) for the xi and pi . There are two similar
commutation relations: one for Ly and Lz , and one for Lz and Lx . Collecting all
of these commutation relations together, we obtain
[Lx , Ly ] = i h̄ Lz , (8.6)
[Ly , Lz ] = i h̄ Lx , (8.7)
[Lz , Lx ] = i h̄ Ly . (8.8)
Now, it is easily demonstrated that if A and B are two general operators then
Hence,
where use has been made of Eqs. (8.6)–(8.8). In other words, L2 commutes with
Lx . Likewise, it is easily demonstrated that L2 also commutes with Ly , and with
Lz . Thus,
[L2 , Lx ] = [L2 , Ly ] = [L2 , Lz ] = 0. (8.12)
Recall, from Sect. 4.10, that in order for two physical quantities to be (ex-
actly) measured simultaneously, the operators which represent them in quantum
122
8.3 Representation of Angular Momentum 8 ORBITAL ANGULAR MOMENTUM
mechanics must commute with one another. Hence, the commutation relations
(8.6)–(8.8) and (8.12) imply that we can only simultaneously measure the mag-
nitude squared of the angular momentum vector, L2 , together with, at most, one
of its cartesian components. By convention, we shall always choose to measure
the z-component, Lz .
Now, we saw earlier, in Sect. 7.2, that the operators, pi , which represent the carte-
sian components of linear momentum in quantum mechanics, can be represented
123
8.3 Representation of Angular Momentum 8 ORBITAL ANGULAR MOMENTUM
Making use of the definitions (8.2)–(8.4), (8.9), and (8.13), the fundamental
representation (7.12)–(7.14) of the pi operators as spatial differential operators,
the definitions (8.20)–(8.22), and a great deal of tedious analysis, we finally
obtain
∂ ∂
!
Lx = −i h̄ − sin φ − cos φ cot θ , (8.23)
∂θ ∂φ
∂ ∂
!
Ly = −i h̄ cos φ − sin φ cot θ , (8.24)
∂θ ∂φ
∂
Lz = −i h̄ , (8.25)
∂φ
as well as
1 ∂2
2 1 ∂ ∂
!
2
L = −h̄ sin θ + , (8.26)
sin θ ∂θ ∂θ sin2 θ ∂φ2
and
∂ ∂
!
±i φ
L± = h̄ e ± + i cot θ . (8.27)
∂θ ∂φ
We, thus, conclude that all of our angular momentum operators can be repre-
sented as differential operators involving the angular spherical coordinates, θ
and φ, but not involving the radial coordinate, r.
124
8.4 Eigenstates of Angular Momentum 8 ORBITAL ANGULAR MOMENTUM
Here, the Yl,m (θ, φ) are the eigenstates in question, whereas the dimensionless
quantities m and l parameterize the eigenvalues of Lz and L2 , which are m h̄
and l (l + 1) h̄2 , respectively. Of course, we expect the Yl,m to be both mutually
orthogonal and properly normalized (see Sect. 4.9), so that
I
Yl∗′ ,m ′ (θ, φ) Yl,m (θ, φ) dΩ = δll ′ δmm ′ , (8.30)
where dΩ = sin θ dθ dφ is an element of solid angle, and the integral is over all
solid angle.
Now,
where use has been made of Eq. (8.18). We, thus, conclude that when the op-
erator L+ operates on an eigenstate of Lz corresponding to the eigenvalue m h̄ it
converts it to an eigenstate corresponding to the eigenvalue (m + 1) h̄. Hence, L+
is known as the raising operator (for Lz ). It is also easily demonstrated that
125
8.5 Eigenvalues of Lz 8 ORBITAL ANGULAR MOMENTUM
Writing
L+ Yl,m = c+
l,m Yl,m+1 , (8.33)
L− Yl,m = c−
l,m Yl,m−1 , (8.34)
we obtain
L− L+ Yl,m = c+ − 2
l,m cl,m+1 Yl,m = [l (l + 1) − m (m + 1)] h̄ Yl,m , (8.35)
where use has been made of Eq. (8.16). Likewise,
L+ L− Yl,m = c+ − 2
l,m−1 cl,m Yl,m = [l (l + 1) − m (m − 1)] h̄ Yl,m , (8.36)
where use has been made of Eq. (8.15). It follows that
c+ − 2
l,m cl,m+1 = [l (l + 1) − m (m + 1)] h̄ , (8.37)
c+ − 2
l,m−1 cl,m = [l (l + 1) − m (m − 1)] h̄ . (8.38)
These equations are satisfied when
c±
l,m = [l (l + l) − m (m ± 1)]
1/2
h̄. (8.39)
Hence, we can write
L+ Yl,m = [l (l + 1) − m (m + 1)]1/2 h̄ Yl,m+1 , (8.40)
L− Yl,m = [l (l + 1) − m (m − 1)]1/2 h̄ Yl,m−1 . (8.41)
8.5 Eigenvalues of Lz
It seems reasonable to attempt to write the eigenstate Yl,m (θ, φ) in the separable
form
Yl,m (θ, φ) = Θl,m (θ) Φm (φ). (8.42)
We can satisfy the orthonormality constraint (8.30) provided that
Zπ
Θl∗′ ,m ′ (θ) Θl,m (θ) sin θ dθ = δll ′ , (8.43)
−π
Z 2π
∗
Φm ′ (φ) Φm (φ) dφ = δmm ′ . (8.44)
0
126
8.6 Eigenvalues of L2 8 ORBITAL ANGULAR MOMENTUM
Note, from Eq. (8.25), that the differential operator which represents Lz only
depends on the azimuthal angle φ, and is independent of the polar angle θ. It
therefore follows from Eqs. (8.25), (8.28), and (8.42) that
dΦm
− i h̄ = m h̄ Φm . (8.45)
dφ
The solution to this equation is obvious:
Φm (φ) ∼ e i m φ . (8.46)
Here, the symbol ∼ just means that we are neglecting multiplicative constants.
Finally, we can easily normalize the eigenstate (8.46) by making use of the
orthonormality constraint (8.44). We obtain
eimφ
Φm (φ) = √ . (8.47)
2π
This is the properly normalized eigenstate of Lz corresponding to the eigenvalue
m h̄.
8.6 Eigenvalues of L2
Consider the angular wave-function ψ(θ, φ) = L+ Yl,m (θ, φ). We know that
I
ψ∗ (θ, φ) ψ(θ, φ) dΩ ≥ 0, (8.48)
127
8.6 Eigenvalues of L2 8 ORBITAL ANGULAR MOMENTUM
l (l + 1) ≥ m (m + 1). (8.51)
l (l + 1) ≥ m (m − 1). (8.53)
128
8.6 Eigenvalues of L2 8 ORBITAL ANGULAR MOMENTUM
Well, if m can only take a restricted range of integer values, then there must
exist a lowest possible value it can take. Let us call this special value m− , and
let Yl,m− be the corresponding eigenstate. Suppose we act on this eigenstate with
the lowering operator L− . According to Eq. (8.32), this will have the effect of
converting the eigenstate into that of a state with a lower value of m. However,
no such state exists. A non-existent state is represented in quantum mechanics
by the null wave-function ψ = 0. Thus, we must have
L− Yl,m− = 0. (8.55)
Now, from Eq. (8.15),
L2 = L+ L− + Lz2 − h̄ Lz (8.56)
Hence,
L2 Yl,m− = (L+ L− + Lz2 − h̄ Lz ) Yl,m− , (8.57)
or
l (l + 1) Yl,m− = m− (m− − 1) Yl,m− , (8.58)
where use has been made of (8.28), (8.29), and (8.55). It follows that
l (l + 1) = m− (m− − 1). (8.59)
Assuming that m− is negative, the solution to the above equation is
m− = −l. (8.60)
We can similarly show that the largest possible value of m is
m+ = +l. (8.61)
The above two results imply that l is an integer, since m− and m+ are both con-
strained to be integers.
We can now formulate the rules which determine the allowed values of the
quantum numbers l and m. The quantum number l takes the non-negative in-
teger values 0, 1, 2, 3, · · ·. Once l is given, the quantum number m can take any
integer value in the range
− l, −l + 1, · · · 0, · · · , l − 1, l. (8.62)
Thus, if l = 0 then m can only take the value 0, if l = 1 then m can take the
values −1, 0, +1, if l = 2 then m can take the values −2, −1, 0, +1, +2, and so on.
129
8.7 Spherical Harmonics 8 ORBITAL ANGULAR MOMENTUM
The simultaneous eigenstates, Yl,m (θ, φ), of L2 and Lz are known as the spherical
harmonics. Let us investigate their functional form.
Once we know Yl,l , we can obtain Yl,l−1 by operating on Yl,l with the lowering
operator L− . Thus,
∂ ∂
!
Yl,l−1 ∼ L− Yl,l ∼ e −i φ
− + i cot θ (sin θ)l e i l φ , (8.70)
∂θ ∂φ
where use has been made of Eq. (8.27). The above equation yields
d
!
i (l−1) φ
Yl,l−1 ∼ e + l cot θ (sin θ)l . (8.71)
dθ
130
8.7 Spherical Harmonics 8 ORBITAL ANGULAR MOMENTUM
Now,
d 1 d h
!
l
i
+ l cot θ f(θ) ≡ (sin θ) f(θ) , (8.72)
dθ (sin θ)l dθ
where f(θ) is a general function. Hence, we can write
e i (l−1) φ 1 d
!
Yl,l−1 (θ, φ) ∼ l−1
(sin θ)2 l . (8.73)
(sin θ) sin θ dθ
Likewise, we can show that
e−i (l−1) φ 1 d
!
Yl,−l+1 (θ, φ) ∼ L+ Yl,−l ∼ l−1
(sin θ)2 l . (8.74)
(sin θ) sin θ dθ
We can now obtain Yl,l−2 by operating on Yl,l−1 with the lowering operator. We
get
∂ ∂ e i (l−1) φ 1 d
! !
−i φ
Yl,l−2 ∼ L− Yl,l−1 ∼ e − + i cot θ l−1
(sin θ)2 l ,
∂θ ∂φ (sin θ) sin θ dθ
(8.75)
which reduces to
d 1 1 d
" # !
−i (l−2) φ
Yl,l−2 ∼ e + (l − 1) cot θ l−1
(sin θ)2 l . (8.76)
dθ (sin θ) sin θ dθ
Finally, making use of Eq. (8.72), we obtain
!2
e i (l−2) φ 1 d
Yl,l−2 (θ, φ) ∼ (sin θ)2 l . (8.77)
(sin θ)l−2 sin θ dθ
Likewise, we can show that
!2
e−i (l−2) φ 1 d
Yl,−l+2 (θ, φ) ∼ L+ Yl,−l+1 ∼ (sin θ)2 l . (8.78)
(sin θ)l−2 sin θ dθ
A comparison of Eqs. (8.68), (8.73), and (8.77) reveals the general functional
form of the spherical harmonics:
!l−m
eimφ 1 d
Yl,m (θ, φ) ∼ (sin θ)2 l . (8.79)
(sin θ)m sin θ dθ
131
8.7 Spherical Harmonics 8 ORBITAL ANGULAR MOMENTUM
After a great deal of tedious analysis, the normalized spherical harmonic func-
tions are found to take the form
1/2
2 l + 1 (l − m)!
Yl,m (θ, φ) = (−1)m Pl,m (cos θ) e i m φ (8.83)
4 π (l + m)!
for m ≥ 0, where the Pl,m are known as associated Legendre polynomials, and are
written
!l−m
l+m (l + m)! (1 − u2 )−m/2 d
Pl,m (u) = (−1) (1 − u2 )l (8.84)
(l − m)! 2l l! du
for m ≥ 0. Alternatively,
!l+m
(1 − u2 )m/2 d
l
Pl,m (u) = (−1) (1 − u2 )l , (8.85)
2l l! du
for m ≥ 0. The spherical harmonics characterized by m < 0 can be calculated
from those characterized by m > 0 via the identity
∗
Yl,−m = (−1)m Yl,m . (8.86)
132
8.7 Spherical Harmonics 8 ORBITAL ANGULAR MOMENTUM
Figure 15: The |Yl,m(θ, φ)| 2 plotted as a functions of θ. The solid, short-dashed, and long-dashed
curves correspond to l, m = 0, 0, and 1, 0, and 1, ±1, respectively.
133
8.7 Spherical Harmonics 8 ORBITAL ANGULAR MOMENTUM
and also form a complete set. In other words, any function of θ and φ can be
represented as a superposition of spherical harmonics. Finally, and most im-
portantly, the spherical harmonics are the simultaneous eigenstates of Lz and L2
corresponding to the eigenvalues m h̄ and l (l + 1) h̄2 , respectively.
Problems
1. A system is in the state ψ = Yl,m(θ, φ). Calculate hLxi and hLx2i. [from Squires]
2. Find the eigenvalues and eigenfunctions (in terms of the angles θ and φ) of Lx. [from Squires]
3. Consider a beam of particles with l = 1. A measurement of Lx yields the result h̄. What values will
be obtained by a subsequent measurement of Lz, and with what probabilities? Repeat the calculation
for the cases in which the measurement of Lx yields the results 0 and −h̄. [modified from Squires]
4. The Hamiltonian for an axially symmetric rotator is given by
Lx2 + Ly2 L2
H= + z .
2 I1 2 I2
What are the eigenvalues of H? [from Gaziorowicz].
134
8.7 Spherical Harmonics 8 ORBITAL ANGULAR MOMENTUM
Figure 16: The |Yl,m(θ, φ)| 2 plotted as a functions of θ. The solid, short-dashed, and long-dashed
curves correspond to l, m = 2, 0, and 2, ±1, and 2, ±2, respectively.
135
9 CENTRAL POTENTIALS
9 Central Potentials
9.1 Introduction
Now, we have seen that the cartesian components of the momentum, p, can be
represented as (see Sect. 7.2)
∂
pi = −i h̄ (9.3)
∂xi
for i = 1, 2, 3, where x1 ≡ x, x2 ≡ y, x3 ≡ z, and r ≡ (x1 , x2 , x3 ). Likewise, it is
easily demonstrated, from the above expressions, and the basic definitions of the
spherical polar coordinates [see Eqs. (8.20)–(8.22)], that the radial component
of the momentum can be represented as
p·r ∂
pr ≡ = −i h̄ . (9.4)
r ∂r
Recall that the angular momentum vector, L, is defined [see Eq. (8.1)]
L = r × p. (9.5)
136
9.2 Derivation of Radial Equation 9 CENTRAL POTENTIALS
Let us calculate the value of L2 using Eq. (9.6). According to our new notation,
L2 is the same as Li Li . Thus, we obtain
L2 = ǫijk xj pk ǫilm xl pm = ǫijk ǫilm xj pk xl pm . (9.8)
Note that we are able to shift the position of ǫilm because its elements are just
numbers, and, therefore, commute with all of the xi and the pi . Now, it is easily
demonstrated that
ǫijk ǫilm ≡ δjl δkm − δjm δkl . (9.9)
Here δij is the usual Kronecker delta, whose elements are determined according
to the rule
1 if i and j the same
δij = . (9.10)
0 if i and j different
It follows from Eqs. (9.8) and (9.9) that
L2 = xi pj xi pj − xi pj xj pi . (9.11)
Here, we have made use of the fairly self-evident result that δij ai bj ≡ ai bi . We
have also been careful to preserve the order of the various terms on the right-
hand side of the above expression, since the xi and the pi do not necessarily
commute with one another.
137
9.2 Derivation of Radial Equation 9 CENTRAL POTENTIALS
We now need to rearrange the order of the terms on the right-hand side of
Eq. (9.11). We can achieve this by making use of the fundamental commutation
relation for the xi and the pi [see Eq. (7.17)]:
[xi , pj ] = i h̄ δij . (9.12)
Thus,
L2 = xi (xi pj − [xi , pj ]) pj − xi pj (pi xj + [xj , pi ])
= xi xi pj pj − i h̄ δij xi pj − xi pj pi xj − i h̄ δij xi pj
= xi xi pj pj − xi pi pj xj − 2 i h̄ xi pi . (9.13)
Here, we have made use of the fact that pj pi = pi pj , since the pi commute with
one another [see Eq. (7.16)]. Next,
L2 = xi xi pj pj − xi pi (xj pj − [xj , pj ]) − 2 i h̄ xi pi . (9.14)
Now, according to (9.12),
[xj , pj ] ≡ [x1 , p1 ] + [x2 , p2 ] + [x3 , p3 ] = 3 i h̄. (9.15)
Hence, we obtain
L2 = xi xi pj pj − xi pi xj pj + i h̄ xi pi . (9.16)
When expressed in more conventional vector notation, the above expression be-
comes
L2 = r2 p2 − (r · p)2 + i h̄ r · p. (9.17)
Note that if we had attempted to derive the above expression directly from Eq.
(9.5), using standard vector identities, then we would have missed the final term
on the right-hand side. This term originates from the lack of commutation be-
tween the xi and pi operators in quantum mechanics. Of course, standard vector
analysis assumes that all terms commute with one another.
Let us now consider whether the above Hamiltonian commutes with the an-
gular momentum operators Lz and L2 . Recall, from Sect. 8.3, that Lz and L2 are
represented as differential operators which depend solely on the angular spheri-
cal polar coordinates, θ and φ, and do not contain the radial polar coordinate, r.
Thus, any function of r, or any differential operator involving r (but not θ and φ),
will automatically commute with L2 and Lz . Moreover, L2 commutes both with it-
self, and with Lz (see Sect. 8.2). It is, therefore, clear that the above Hamiltonian
commutes with both Lz and L2 .
Now, according to Sect. 4.10, if two operators commute with one another
then they possess simultaneous eigenstates. We thus conclude that for a particle
moving in a central potential, the eigenstates of the Hamiltonian are simultaneous
eigenstates of Lz and L2 . Now, we have already found the simultaneous eigenstates
of Lz and L2 —they are the spherical harmonics, Yl,m (θ, φ), discussed in Sect. 8.7.
It follows that the spherical harmonics are also eigenstates of the Hamiltonian.
This observation leads us to try the following separable form for the stationary
wave-function:
ψ(r, θ, φ) = R(r) Yl,m (θ, φ). (9.22)
It immediately follows, from (8.28) and (8.29), and the fact that Lz and L2 both
obviously commute with R(r), that
Lz ψ = m h̄ ψ, (9.23)
L2 ψ = l (l + 1) h̄2 ψ. (9.24)
Recall that the quantum numbers m and l are restricted to take certain integer
values, as explained in Sect. 8.6.
139
9.3 The Infinite Potential Well 9 CENTRAL POTENTIALS
Finally, making use of Eqs. (9.1), (9.21), and (9.24), we obtain the follow-
ing differential equation which determines the radial variation of the stationary
wave-function:
h̄2 d2
2d l (l + 1)
− + − Rn,l + V Rn,l = E Rn,l . (9.25)
2 m dr2 r dr r2
Here, we have labeled the function R(r) by two quantum numbers, n and l. The
second l is, of course, related to the eigenvalue of L2 . [Note that the azimuthal
quantum number, m, does not appear in the above equation, and, therefore,
does not influence either the function R(r) or the energy, E.] As we shall see, the
first, n, is determined by the constraint that the radial wave-function be square-
integrable.
Consider a particle of mass m and energy E > 0 moving in the following simple
central potential:
0 for 0 ≤ r ≤ a
V(r) = . (9.26)
∞ otherwise
Obviously, the wave-function ψ is only non-zero in the region 0 ≤ r ≤ a. Within
this region, it is subject to the physical boundary conditions that it be well-
behaved (i.e., square-integrable) at r = 0, and that it be zero at r = a (see
Sect. 5.2). Writing the wave-function in the standard form
+ + k − Rn,l = 0 (9.28)
dr2 r dr r2
in the region 0 ≤ r ≤ a, where
2mE
k2 = . (9.29)
h̄2
140
9.3 The Infinite Potential Well 9 CENTRAL POTENTIALS
Figure 17: The first few spherical Bessel functions. The solid, short-dashed, long-dashed, and dot-
dashed curves show j0(z), j1(z), y0(z), and y1(z), respectively.
141
9.3 The Infinite Potential Well 9 CENTRAL POTENTIALS
Defining the scaled radial variable z = k r, the above differential equation can
be transformed into the standard form
d2 Rn,l 2 dRn,l
l (l + 1)
+ + 1 − Rn,l = 0. (9.30)
dz2 r dz z2
The two independent solutions to this well-known second-order differential equa-
tion are called spherical Bessel functions,3 and can be written
!l
1 d sin z
!
l
jl (z) = z − , (9.31)
z dz z
!l
1 d cos z
!
l
yl (z) = −z − . (9.32)
z dz z
Thus, the first few spherical Bessel functions take the form
sin z
j0 (z) = , (9.33)
z
sin z cos z
j1 (z) = − , (9.34)
z2 z
cos z
y0 (z) = − , (9.35)
z
cos z sin z
y1 (z) = − 2 − . (9.36)
z z
These functions are also plotted in Fig. 17. It can be seen that the spherical Bessel
functions are oscillatory in nature, passing through zero many times. However,
the yl (z) functions are badly behaved (i.e., they are not square-integrable) at
z = 0, whereas the 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
wave-function 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 , (9.37)
3
M. Abramowitz, and I.A. Stegun, Handbook of mathematical functions (Dover, New York NY, 1965), Sect. 10.1.
142
9.3 The Infinite Potential Well 9 CENTRAL POTENTIALS
Table 1: The first few zeros of the spherical Bessel function jl(z).
2 h̄2
En,l = zn,l . (9.38)
2 m a2
The first few values of zn,l are listed in Table 1. It can be seen that zn,l is an
increasing function of both n and l.
143
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
when n 6= n ′ . Given that the Yl,m (θ, φ) are mutually orthogonal (see Sect. 8),
this ensures that wave-functions (9.27) corresponding to distinct sets of values of
the quantum numbers n, l, and m are mutually orthogonal.
144
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
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, (9.46)
dz2 z dz z2 z
where v
2
2 me a e u E0
u
ζ= =2 t
. (9.47)
4π ǫ0 h̄2 E
Suppose that Rn,l (r) = Z(r/a) exp(−r/a)/(r/a). It follows that
d2
d l (l + 1) ζ
− 2 − + Z = 0. (9.48)
dz2 dz z2 z
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.
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 square-integrable 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
145
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
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
= , (9.52)
ck−1 k
according to Eq. (9.51). This is the same as the ratio of successive coefficients in
the power series
X (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
R not correspond to physically acceptable behaviour of the wave-function,
2
since |ψ| 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
E0
E= 2 (9.55)
n
and
a = n a0 , (9.56)
where
me e4 e2
E0 = − =− = −13.6 eV, (9.57)
2 (4π ǫ0 )2 h̄2 8π ǫ0 a0
and
4π ǫ0 h̄2
a0 = 2
= 5.3 × 10−11 m. (9.58)
me e
146
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
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 h̄ c) ≃ 1/137 is the dimensionless fine-structure
constant. The fact that |E0 | ≪ me c2 is the ultimate justification for our non-
relativistic treatment of the hydrogen atom.
Here, the Yl,m (θ, φ) are the spherical harmonics (see Sect 8.7), and Rn,l (z = r/a)
is the solution of
1 d 2 d l (l + 1) 2 n
2
z − 2
+ − 1 Rn,l = 0 (9.60)
z dz dz z z
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. (8.87)]: i.e.,
I
Yl∗′ ,m ′ Yl,m dΩ = δll ′ δmm ′ . (9.63)
It, thus, follows that the radial wave-function satisfies the orthonormality con-
straint Z∞
R∗n ′ ,l Rn,l r2 dr = δnn ′ . (9.64)
0
147
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
The first few radial wave-functions for the hydrogen atom are listed below:
2 r
!
R1,0 (r) = 3/2
exp − , (9.65)
a0 a0
2 r r
! !
R2,0 (r) = 1 − exp − , (9.66)
(2 a0 )3/2 2 a0 2 a0
1 r r
!
R2,1 (r) = √ exp − , (9.67)
3 (2 a0 )3/2 a0 2 a0
2 r2
2 2r r
!
R3,0 (r) = 1 − + exp − , (9.68)
(3 a0 )3/2 3 a0 27 a02 3 a0
√
4 2 r r r
! !
R3,1 (r) = 1− exp − , (9.69)
9 (3 a0 )3/2 a0 6 a0 3 a0
√
2 2 r 2 r
! !
R3,2 (r) = √ exp − . (9.70)
27 5 (3 a0 )3/2 a0 3 a0
These functions are illustrated in Figs. 18 and 19.
Given the (properly normalized) hydrogen wave-function (9.59), plus our in-
terpretation of |ψ|2 as a probability density, we can calculate
Z∞
k
hr i = r2+k |Rn,l (r)| 2 dr, (9.71)
0
where the angle-brackets denote an expectation value. For instance, it can be
demonstrated (after much tedious algebra) that
2 a02 n2
hr i = [5 n2 + 1 − 3 l (l + 1)], (9.72)
2
a0
hri = [3 n2 − l (l + 1)], (9.73)
2
1 1
* +
= , (9.74)
r n 2 a0
1 1
* +
= , (9.75)
r2 (l + 1/2) n3 a02
1 1
* +
= . (9.76)
r3 l (l + 1/2) (l + 1) n3 a03
148
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
Figure 18: The a0 r2 |Rn,l(r)| 2 plotted as a functions of r/a0. The solid, short-dashed, and long-dashed
curves correspond to n, l = 1, 0, and 2, 0, and 2, 1, respectively.
149
9.4 The Hydrogen Atom 9 CENTRAL POTENTIALS
Figure 19: The a0 r2 |Rn,l(r)| 2 plotted as a functions of r/a0. The solid, short-dashed, and long-dashed
curves correspond to n, l = 3, 0, and 3, 1, and 3, 2, respectively.
150
9.5 The Rydberg Formula 9 CENTRAL POTENTIALS
The fact that the energy levels of a hydrogen atom only depend on n, and
not on l and m, implies that the energy spectrum of a hydrogen atom is highly
degenerate: i.e., there are many different states which possess the same energy.
According to the inequality (9.61) (and the fact that n, l, and m are integers),
for a given value of l, there are 2 l + 1 different allowed values of m (i.e., −l, −l +
1, · · · , l − 1, l). Likewise, for a given value of n, there are n different allowed
values of l (i.e., 0, 1, · · · , n − 1). Now, all states possessing the same value of n
have the same energy (i.e., they are degenerate). Hence, the total number of
degenerate states corresponding to a given value of n is
1 + 3 + 5 + · · · + 2 (n − 1) + 1 = n2 . (9.77)
Thus, the ground-state (n = 1) is not degenerate, the first excited state (n = 2)
is four-fold degenerate, the second excited state (n = 3) is nine-fold degenerate,
etc. [Actually, when we take into account the two spin states of an electron (see
Sect. 10), the degeneracy of the nth energy level becomes 2 n2 .]
Problems
152
9.5 The Rydberg Formula 9 CENTRAL POTENTIALS
2. Consider a particle of mass m in the three-dimensional harmonic oscillator potential V(r) = (1/2) m ω2 r2.
Solve the problem by separation of variables in spherical polar coordinates, and, hence, determine
the energy eigenvalues of the system. [from Griffiths]
3. The normalized wave-function for the ground-state of a hydrogen-like atom (neutral hydrogen, He+,
Li++, etc.) with nuclear charge Z e has the form
ψ = A exp(−β r),
where A and β are constants, and r is the distance between the nucleus and the electron. Show the
following:
(a) A2 = β3/π.
(b) β = Z/a0, where a0 = (h̄2/me) (4π ǫ0/e2).
(c) The energy is E = −Z2 E0 where E0 = (me/2 h̄2) (e2/4π ǫ0)2.
(d) The expectation values of the potential and kinetic energies are 2 E and −E, respectively.
(e) The expectation value of r is (3/2) (a0/Z).
(f) The most probable value of r is a0/Z.
[from Squires]
4. An atom of tritium is in its ground-state. Suddenly the nucleus decays into a helium nucleus, via the
emission of a fast electron which leaves the atom without perturbing the extranuclear electron, Find
the probability that the resulting He+ ion will be left in an n = 1, l = 0 state. Find the probability
that it will be left in a n = 2, l = 0 state. What is the probability that the ion will be left in an l > 0
state? [from Squires]
6. To conserve linear momentum, an atom emitting a photon must recoil, which means that not all of
the energy made available in the downward jump goes to the photon. Find a hydrogen atom’s recoil
energy when it emits a photon in an n = 2 to n = 1 transition. What fraction of the transition energy
is the recoil energy? [from Harris]
153
10 SPIN ANGULAR MOMENTUM
10.1 Introduction
Broadly speaking, a classical extended object (e.g., the Earth) can possess two
types of angular momentum. The first type is due to the rotation of the object’s
center of mass about some fixed external point (e.g., the Sun)—this is generally
known as orbital angular momentum. The second type is due to the object’s
internal motion—this is generally known as spin angular momentum (since, for
a rigid object, the internal motion consists of spinning about an axis passing
through the center of mass). By analogy, quantum particles can possess both
orbital angular momentum due to their motion through space (see Sect. 8), and
spin angular momentum due to their internal motion. Actually, the analogy with
classical extended objects is not entirely accurate, since electrons, for instance,
are structureless point particles. In fact, in quantum mechanics, it is best to think
of spin angular momentum as a kind of intrinsic angular momentum possessed
by particles. It turns out that each type of elementary particle has a characteristic
spin angular momentum, just as each type has a characteristic charge and mass.
[Sx , Sy ] = i h̄ Sz , (10.1)
[Sy , Sz ] = i h̄ Sx , (10.2)
[Sz , Sx ] = i h̄ Sy . (10.3)
154
10.3 Spin Space 10 SPIN ANGULAR MOMENTUM
We can represent the magnitude squared of the spin angular momentum vector
by the operator
S2 = Sx2 + Sy2 + Sz2 . (10.4)
By analogy with the analysis in Sect. 8.2, it is easily demonstrated that
We thus conclude (see Sect. 4.10) that we can simultaneously measure the mag-
nitude squared of the spin angular momentum vector, together with, at most,
one cartesian component. By convention, we shall always choose to measure the
z-component, Sz .
By analogy with Eq. (8.13), we can define raising and lowering operators for
spin angular momentum:
S ± = Sx ± i S y . (10.6)
If Sx , Sy , and Sz are hermitian operators, as must be the case if they are to rep-
resent physical quantities, then S± are the hermitian conjugates of one another:
i.e.,
(S± )† = S∓ . (10.7)
Finally, by analogy with Sect. 8.2, it is easily demonstrated that
S+ S− = S2 − Sz2 + h̄ Sz , (10.8)
S− S+ = S2 − Sz2 − h̄ Sz , (10.9)
[S+ , Sz ] = −h̄ S+ , (10.10)
[S− , Sz ] = +h̄ S− . (10.11)
We now have to discuss the wave-functions upon which the previously intro-
duced spin operators act. Unlike regular wave-functions, spin wave-functions do
not exist in real space. Likewise, the spin angular momentum operators cannot
be represented as differential operators in real space. Instead, we need to think
155
10.3 Spin Space 10 SPIN ANGULAR MOMENTUM
We now need to define the length of our vectors. We can do this by introducing
a second, or dual, vector space whose elements are in one to one correspondence
with the elements of our first space. Let the element of the second space which
corresponds to the element χ of the first space be called χ† . Moreover, the element
of the second space which corresponds to c χ is c∗ χ† . We 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, or norm, of χ. Let us denote this
number χ† χ. Thus, we have
χ† χ ≥ 0 (10.12)
for all χ. We also assume that it is possible to combine unlike states in an anal-
ogous multiplicative fashion to produce complex numbers. The product of two
unlike states χ and χ ′ is denoted χ† χ ′ . Two states χ and χ ′ are said to be mutually
orthogonal, or independent, if χ† χ ′ = 0.
156
10.4 Eigenstates of Sz and S2 10 SPIN ANGULAR MOMENTUM
Since the operators Sz and S2 commute, they must possess simultaneous eigen-
states (see Sect. 4.10). Let these eigenstates take the form [see Eqs. (8.28) and
(8.29)]:
Sz χs,ms = ms h̄ χs,ms , (10.16)
S2 χs,ms = s (s + 1) h̄ 2 χs,ms . (10.17)
157
10.4 Eigenstates of Sz and S2 10 SPIN ANGULAR MOMENTUM
Let ms min be the minimum possible value of ms . It follows that (see Sect. 8.6)
S− χs,ms min = 0. (10.25)
Now, from Eq. (10.8),
S2 = S+ S− + Sz2 − h̄ Sz . (10.26)
Hence,
S2 χs,ms min = (S+ S− + Sz2 − h̄ Sz ) χs,ms min , (10.27)
giving
s (s + 1) = ms min (ms min − 1). (10.28)
Assuming that ms min < 0, this equation yields
ms min = −s. (10.29)
Likewise, it is easily demonstrated that
ms max = +s. (10.30)
Moreover,
S− χs,−s = S+ χs,s = 0. (10.31)
158
10.4 Eigenstates of Sz and S2 10 SPIN ANGULAR MOMENTUM
Now, the raising operator S+ , acting upon χs,−s , converts it into some multiple
of χs,−s+1 . Employing the raising operator a second time, we obtain a multiple
of χs,−s+2 . However, this process cannot continue indefinitely, since there is a
maximum possible value of ms . Indeed, after acting upon χs,−s a sufficient num-
ber of times with the raising operator S+ , we must obtain a multiple of χs,s , so
that employing the raising operator one more time leads to the null state [see
Eq. (10.31)]. If this is not the case, then we will inevitably obtain eigenstates of
Sz corresponding to ms > s, which we have already demonstrated is impossible.
where k is a positive integer. Hence, the quantum number s can either take pos-
itive integer or positive half-integer values. Up to now, our analysis has been very
similar to that which we used earlier to investigate orbital angular momentum
(see Sect. 8). Recall, that for orbital angular momentum the quantum number
m, which is analogous to ms , is restricted to take integer values (see Sect. 8.5).
This implies that the quantum number l, which is analogous to s, is also restricted
to take integer values. However, the origin of these restrictions is the representa-
tion of the orbital angular momentum operators as differential operators in real
space (see Sect. 8.3). There is no equivalent representation of the corresponding
spin angular momentum operators. Hence, we conclude that there is no reason
why the quantum number s cannot take half-integer, as well as integer, values.
In 1940, Wolfgang Pauli proved the so-called spin-statistics theorem using rel-
ativistic quantum mechanics. According to this theorem, all fermions possess
half-integer spin (i.e., a half-integer value of s), whereas all bosons possess integer
spin (i.e., an integer value of s). In fact, all presently known fermions, includ-
ing electrons and protons, possess spin one-half. In other words, electrons and
protons are characterized by s = 1/2 and ms = ±1/2.
159
10.5 The Pauli Representation 10 SPIN ANGULAR MOMENTUM
160
10.5 The Pauli Representation 10 SPIN ANGULAR MOMENTUM
Likewise, the product χ† χ ′ of two different spin states is also obtained from the
rules of matrix multiplication: i.e.,
c+′
A† ≡ . (10.44)
A∗12 , A∗22
and
0
χ− ≡ , (10.46)
1
161
10.5 The Pauli Representation 10 SPIN ANGULAR MOMENTUM
respectively. Note that these forms automatically satisfy the orthonormality con-
straints (10.36) and (10.37). It is convenient to write the spin operators Si
(where i = 1, 2, 3 corresponds to x, y, z) as
h̄
Si = σ i . (10.47)
2
Here, the σi are dimensionless 2 × 2 matrices. According to Eqs. (10.1)–(10.3),
the σi satisfy the commutation relations
[σx , σy ] = 2 i σz , (10.48)
[σy , σz ] = 2 i σx , (10.49)
[σz , σx ] = 2 i σy . (10.50)
Furthermore, Eq. (10.34) yields
σz χ± = ±χ± . (10.51)
It is easily demonstrated, from the above expressions, that the σi are represented
by the following matrices:
0, 1
σx ≡ , (10.52)
1, 0
0, −i
σy ≡ , (10.53)
i, 0
1, 0
σz ≡ . (10.54)
0, −1
Incidentally, these matrices are generally known as the Pauli matrices.
162
10.6 Spin Precession 10 SPIN ANGULAR MOMENTUM
The above expression suggests that there may be a similar relationship be-
tween magnetic moment and spin angular momentum. We can write
ge
µ=− S, (10.58)
2 me
where g is called the gyromagnetic ratio. Classically, we would expect g = 1. In
fact,
α
!
g=2 1+ + · · · = 2.0023192, (10.59)
2π
where α = e2 /(2 ǫ0 h c) ≃ 1/137 is the so-called fine-structure constant. The fact
that the gyromagnetic ratio is (almost) twice that expected from classical physics
is only explicable using relativistic quantum mechanics. Furthermore, the small
corrections to the relativistic result g = 2 come from quantum field theory.
H = −µ · B. (10.60)
Assuming that the above expression also holds good in quantum mechanics, the
Hamiltonian of an electron in a z-directed magnetic field of magnitude B takes
the form
H = Ω Sz , (10.61)
163
10.6 Spin Precession 10 SPIN ANGULAR MOMENTUM
where
geB
Ω=
. (10.62)
2 me
Here, for the sake of simplicity, we are neglecting the electron’s translational
degrees of freedom.
where |c+ |2 + |c− |2 = 1. Here, |c+ |2 is the probability of observing the spin-up
state, and |c− |2 the probability of observing the spin-down state. It follows from
Eqs. (10.47), (10.54), (10.61), (10.63), and (10.64) that
ċ+ Ω h̄ 1, 0 c+
i h̄ = , (10.65)
ċ− 2 0, −1 c−
The significance of the angle α will become apparent presently. Solving Eq. (10.66),
subject to the initial conditions (10.67) and (10.68), we obtain
164
10.6 Spin Precession 10 SPIN ANGULAR MOMENTUM
We can most easily visualize the effect of the time dependence in the above ex-
pressions for c± by calculating the expectation values of the three cartesian com-
ponents of the electron’s spin angular momentum. By analogy with Eq. (4.56),
the expectation value of a general spin operator A is simply
hAi = χ† A χ. (10.71)
which reduces to
h̄
hSz i = cos α (10.73)
2
with the help of Eqs. (10.69) and (10.70). Likewise, the expectation value of Sx
is
h̄ ∗ ∗ 0, 1 c+
hSx i = c ,c , (10.74)
2 + − 1, 0 c−
which reduces to
h̄
hSx i = sin α cos(Ω t). (10.75)
2
Finally, the expectation value of Sy is
h̄
hSy i = sin α sin(Ω t). (10.76)
2
According to Eqs. (10.73), (10.75), and (10.76), 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≃ . (10.77)
me
This behaviour is actually equivalent to that predicted by classical physics. Note,
however, that a measurement of Sx , Sy , or Sz will always yield either +h̄/2 or
−h̄/2. It is the relative probabilities of obtaining these two results which varies as
the expectation value of a given component of the spin varies.
165
10.6 Spin Precession 10 SPIN ANGULAR MOMENTUM
Problems
1. Find the Pauli representations of Sx, Sy, and Sz for a spin-1 particle.
2. Find the Pauli representations of the normalized eigenstates of Sx and Sy for a spin-1/2 particle.
3. Suppose that a spin-1/2 particle has a spin vector which lies in the x-z plane, making an angle θ with
the z-axis. Demonstrate that a measurement of Sz yields h̄/2 with probability cos2(θ/2), and −h̄/2
with probability sin2(θ/2).
in the Pauli representation, where α and β are real. What is the probability that a measurement of
Sy yields −h̄/2? [from Gaziorowicz]
B = B0 cos(ω t) ^z,
[from Griffiths]
166
11 ADDITION OF ANGULAR MOMENTUM
11.1 Introduction
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 h̄ L. (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 h̄ S. (11.2)
Now, since the orbital angular momentum operators are associated with the elec-
tron’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 an-
other: i.e.,
[Li , Sj ] = 0, (11.3)
where i, j = 1, 2, 3 corresponds to x, y, z.
167
11.2 General Principles 11 ADDITION OF ANGULAR MOMENTUM
J = L + S. (11.4)
We have
J × J = (L + S) × (L + S)
= L×L+S×S+L×S+S×L=L×L+S×S
= i h̄ L + i h̄ S
= i h̄ J. (11.5)
In other words,
J × J = i h̄ J. (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,
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. (11.10)
Now
J2 = (L + S) · (L + S) = L2 + S2 + 2 L · S, (11.11)
168
11.2 General Principles 11 ADDITION OF ANGULAR MOMENTUM
J2 = L2 + S2 + 2 Lz Sz + L+ S− + L− S+ . (11.12)
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. (11.13)
A similar argument allows us to also conclude that
[J2 , S2 ] = 0. (11.14)
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 , Jz ] 6= 0. (11.15)
Likewise, we can also show that
[J2 , Sz ] 6= 0. (11.16)
Finally, we have
Jz = Lz + Sz . (11.17)
169
11.3 Angular Momentum in the Hydrogen Atom 11 ADDITION OF ANGULAR MOMENTUM
(1)
Let ψl,s;m,ms represent a simultaneous eigenstate of L2 , S2 , Lz , and Sz corre-
sponding to the following eigenvalues:
(1) (1)
L2 ψl,s;m,ms = l (l + 1) h̄2 ψl,s;m,ms , (11.18)
(1) (1)
S2 ψl,s;m,ms = s (s + 1) h̄2 ψl,s;m,ms , (11.19)
(1) (1)
Lz ψl,s;m,ms = m h̄ ψl,s;m,ms , (11.20)
(1) (1)
Sz ψl,s;m,ms = ms h̄ ψl,s;m,ms . (11.21)
Hence,
mj = m + ms . (11.23)
In other words, the quantum numbers controlling the z-components of the vari-
ous angular momentum vectors can simply be added algebraically.
(2)
Finally, let ψl,s;j,mj represent a simultaneous eigenstate of L2 , S2 , J2 , and Jz
corresponding to the following eigenvalues:
(2) (2)
L2 ψl,s;j,mj = l (l + 1) h̄2 ψl,s;j,mj , (11.24)
(2) (2)
S2 ψl,s;j,mj = s (s + 1) h̄2 ψl,s;j,mj , (11.25)
(2) (2)
J2 ψl,s;j,mj = j (j + 1) h̄2 ψl,s;j,mj , (11.26)
(2) (2)
Jz ψl,s;j,mj = mj h̄ ψl,s;j,mj . (11.27)
170
11.3 Angular Momentum in the Hydrogen Atom 11 ADDITION OF ANGULAR MOMENTUM
the internal state of the electron is specified by the spinors χ± (see Sect. 10.5).
Hence, the simultaneous eigenstates of L2 , S2 , Lz , and Sz can be written in the
separable form
(1)
ψl,1/2;m,±1/2 = Yl,m χ± . (11.28)
Here, it is understood that orbital angular momentum operators act on the spher-
ical harmonic functions, Yl,m , whereas spin angular momentum operators act on
the spinors, χ± .
(1)
Since the eigenstates ψl,1/2;m,±1/2 are (presumably) orthonormal, and form a
(2)
complete set, we can express the eigenstates ψl,1/2;j,mj as linear combinations of
them. For instance,
(2) (1) (1)
ψl,1/2;j,m+1/2 = α ψl,1/2;m,1/2 + β ψl,1/2;m+1,−1/2 , (11.29)
where α and β are, as yet, unknown coefficients. 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. (11.23)], and the fact that ms
can only take the values ±1/2. Assuming that the ψ(2) eigenstates are properly
normalized, we have
α2 + β2 = 1. (11.30)
J2 = L2 + S2 + 2 Lz Sz + L+ S− + L− S+ . (11.32)
171
11.3 Angular Momentum in the Hydrogen Atom 11 ADDITION OF ANGULAR MOMENTUM
By analogy, when the spin raising and lowering operators, S± , act on a general
spinor, χs,ms , we obtain
For the special case of spin one-half spinors (i.e., s = 1/2, ms = ±1/2), the above
expressions reduce to
S+ χ+ = S− χ− = 0, (11.38)
and
S± χ∓ = h̄ χ± . (11.39)
and
(x − m) α − [l (l + 1) − m (m + 1)]1/2 β = 0, (11.42)
−[l (l + 1) − m (m + 1)]1/2 α + (x + m + 1) β = 0, (11.43)
where
x = j (j + 1) − l (l + 1) − 3/4. (11.44)
Equations (11.42) and (11.43) can be solved to give
x (x + 1) = l (l + 1), (11.45)
and
α [(l − m) (l + m + 1)]1/2
= . (11.46)
β x−m
172
11.3 Angular Momentum in the Hydrogen Atom 11 ADDITION OF ANGULAR MOMENTUM
m, 1/2 m + 1, −1/2 m, ms
√ √
l + 1/2, m + 1/2 (l+m+1)/(2l+1) (l−m)/(2l+1)
√ √
l − 1/2, m + 1/2 (l−m)/(2l+1) − (l+m+1)/(2l+1)
j, mj
173
11.3 Angular Momentum in the Hydrogen Atom 11 ADDITION OF ANGULAR MOMENTUM
(2) (2)
or j = 1/2 (see earlier), there are also six such states: i.e., ψ3/2,±3/2 , ψ3/2,±1/2 , and
(2)
ψ1/2,±1/2 . According to Tab. 2, the various different eigenstates are interrelated
as follows:
(2) (1)
ψ3/2,±3/2 = ψ±1,±1/2 , (11.51)
v v
u2 u1
u u
(2) (1) (1)
ψ3/2,1/2 = t
ψ0,1/2 + t
ψ1,−1/2 , (11.52)
3 3
v v
u1 u2
u u
(2) (1) (1)
ψ1/2,1/2 = t
ψ0,1/2 − t
ψ1,−1/2 , (11.53)
3 3
v v
u1 u2
u u
(2) (1) (1)
ψ3/2,−1/2 = t
ψ−1,1/2 + t
ψ0,−1/2 , (11.54)
3 3
v v
u2 u1
u u
(2) (1) (1)
ψ1/2,−1/2 = t
ψ−1,1/2 − t
ψ0,−1/2 , (11.55)
3 3
and
(1) (2)
ψ±1,±1/2 = ψ3/2,±3/2 , (11.56)
v v
u2 u1
u u
(1) (2) (2)
ψ0,1/2 = t
ψ3/2,1/2 + t
ψ1/2,1/2 , (11.57)
3 3
v v
u1 u2
u u
(1) (2) (2)
ψ1,−1/2 = t
ψ3/2,1/2 − t
ψ1/2,1/2 , (11.58)
3 3
v v
u1 u2
u u
(1) (2) (2)
ψ−1,1/2 = t
ψ3/2,−1/2 + t
ψ1/2,−1/2 , (11.59)
3 3
v v
u2 u1
u u
(1) (2) (2)
ψ0,−1/2 = t
ψ3/2,−1/2 − t
ψ1/2,−1/2 , (11.60)
3 3
Thus, if we know that an electron in a hydrogen atom is in an l = 1 state char-
(1)
acterized by m = 0 and ms = 1/2 [i.e., the state represented by ψ0,1/2 ] then,
according to Eq. (11.57), a measurement of the total angular momentum will
yield j = 3/2, mj = 1/2 with probability 2/3, and j = 1/2, mj = 1/2 with prob-
ability 1/3. Suppose that we make such a measurement, and obtain the result
j = 3/2, mj = 1/2. As a result of the measurement, the electron is thrown into
174
11.4 Two Spin One-Half Particles 11 ADDITION OF ANGULAR MOMENTUM
3/2, 3/2 1
j, mj
Table 3: Clebsch-Gordon coefficients for adding spin one-half to spin one. Only non-zero coefficients
are shown.
(2)
the corresponding eigenstate, ψ3/2,1/2 . It thus follows from Eq. (11.52) that a sub-
sequent measurement of Lz and Sz will yield m = 0, ms = 1/2 with probability
2/3, and m = 1, ms = −1/2 with probability 1/3.
Consider a system consisting of two spin one-half particles. Suppose that the
system does not possess any orbital angular momentum. Let S1 and S2 be the
spin angular momentum operators of the first and second particles, respectively,
and let
S = S1 + S2 (11.61)
be the total spin angular momentum operator. By analogy with the previous
analysis, we conclude that it is possible to simultaneously measure either S12 ,
S22 , S2 , and Sz , or S12 , S22 , S1z , S2z , and Sz . Let the quantum numbers associated
with measurements of S12 , S1z , S22 , S2z , S2 , and Sz be s1 , ms1 , s2 , ms2 , s, and ms ,
(1)
respectively. In other words, if the spinor χs1 ,s2 ;ms1 ,ms2 is a simultaneous eigenstate
175
11.4 Two Spin One-Half Particles 11 ADDITION OF ANGULAR MOMENTUM
S12 χs(1)
1 ,s2 ;ms ,ms2 = s1 (s1 + 1) h̄2 χs(1)
1 ,s2 ;ms ,ms2 , (11.62)
1 1
S22 χs(1)
1 ,s2 ;ms ,ms2 = s2 (s2 + 1) h̄2 χs(1)
1 ,s2 ;ms ,ms2 , (11.63)
1 1
S1z χs(1)
1 ,s2 ;ms ,ms2 = ms1 h̄ χs(1)
1 ,s2 ;ms ,ms2 , (11.64)
1 1
S2z χs(1)
1 ,s2 ;ms ,ms2 = ms2 h̄ χs(1)
1 ,s2 ;ms ,ms2 , (11.65)
1 1
Sz χs(1)
1 ,s2 ;ms ,ms2 = ms h̄ χs(1)
1 ,s2 ;ms ,ms2 . (11.66)
1 1
(2)
Likewise, if the spinor χs1 ,s2 ;s,ms is a simultaneous eigenstate of S12 , S22 , S2 , and Sz ,
then
S12 χs(2)
1 ,s2 ;s,ms
= s1 (s1 + 1) h̄2 χs(2)
1 ,s2 ;s,ms
, (11.67)
S22 χs(2)
1 ,s2 ;s,ms
= s2 (s2 + 1) h̄2 χs(2)
1 ,s2 ;s,ms
, (11.68)
S2 χs(2)
1 ,s2 ;s,ms
= s (s + 1) h̄2 χs(2)
1 ,s2 ;s,ms
, (11.69)
Sz χs(2)
1 ,s2 ;s,ms
= ms h̄ χs(2)
1 ,s2 ;s,ms
. (11.70)
Of course, since both particles have spin one-half, s1 = s2 = 1/2, and s1z , s2z =
±1/2. Furthermore, by analogy with previous analysis,
Now, we saw, in the previous subsection, 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. By analogy, 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. In other
words, when two spin one-half particles are combined, we either obtain a state
with overall spin s = 1, or a state with overall spin s = 0. To be more exact,
there are three possible s = 1 states (corresponding to ms = −1, 0, 1), and
one possible s = 0 state (corresponding to ms = 0). The three s = 1 states are
generally known as the triplet states, whereas the s = 0 state is known as the
singlet state.
176
11.4 Two Spin One-Half Particles 11 ADDITION OF ANGULAR MOMENTUM
−1/2, −1/2 −1/2, 1/2 1/2, −1/2 1/2, 1/2 ms1 , ms2
1, −1 1
√ √
1, 0 1/ 2 1/ 2
√ √
0, 0 1/ 2 −1/ 2
1, 1 1
s, ms
Table 4: Clebsch-Gordon coefficients for adding spin one-half to spin one-half. Only non-zero coeffi-
cients are shown.
The Clebsch-Gordon coefficients for adding spin one-half to spin one-half can
easily be inferred from Tab. 2 (with l = 1/2), and are listed in Tab. 4. It follows
from this table that the three triplet states are:
(2) (1)
χ1,−1 = χ−1/2,−1.2 , (11.72)
(2) 1 (1) (1)
χ1,0 = √ χ + χ1/2,−1/2 , (11.73)
2 −1/2,1/2
(2) (1)
χ1,1 = χ1/2,1/2 , (11.74)
(2) (2)
where χs,m s
is shorthand for χs1 ,s2 ;s,ms , etc. Likewise, the singlet state is written:
Problems
1. An electron in a hydrogen atom occupies the combined spin and position state
q q
R2,1 1/3 Y1,0 χ+ + 2/3 Y1,1 χ− .
(a) What values would a measurement of L2 yield, and with what probabilities?
(b) Same for Lz.
(c) Same for S2.
(d) Same for Sz.
(e) Same for J2.
(f) Same for Jz.
177
11.4 Two Spin One-Half Particles 11 ADDITION OF ANGULAR MOMENTUM
[from Griffiths]
2. In a low energy neutron-proton system (with zero orbital angular momentum) the potential energy
is given by
(σ1 · r) (σ2 · r)
V(r) = V1(r) + V2(r) 3 − σ1 · σ2 + V3(r) σ1 · σ2,
r2
where σ1 denotes the vector of the Pauli matrices of the neutron, and σ2 denotes the vector of the
Pauli matrices of the proton. Calculate the potential energy for the neutron-proton system:
[from Gaziorowicz]
(a) If a measurement of the spin of one of the electrons shows that it is in the state with Sz = h̄/2,
what is the probability that a measurement of the z-component of the spin of the other electron
yields Sz = h̄/2?
(b) If a measurement of the spin of one of the electrons shows that it is in the state with Sy = h̄/2,
what is the probability that a measurement of the x-component of the spin of the other electron
yields Sx = −h̄/2?
Finally, if electron 1 is in a spin state described by cos α1 χ+ + sin α1 e i β1 χ−, and electron 2 is in a
spin state described by cos α2 χ+ + sin α2 e i β2 χ−, what is the probability that the two-electron spin
state is a triplet state? [from Gaziorowicz]
178
Part II
Applications
179
12 TIME-INDEPENDENT PERTURBATION THEORY
12.1 Introduction
180
12.2 Improved Notation 12 TIME-INDEPENDENT PERTURBATION THEORY
181
12.2 Improved Notation 12 TIME-INDEPENDENT PERTURBATION THEORY
182
12.3 The Two-State System 12 TIME-INDEPENDENT PERTURBATION THEORY
183
12.3 The Two-State System 12 TIME-INDEPENDENT PERTURBATION THEORY
Here, use has been made of the fact that H1 is an Hermitian operator.
184
12.4 Non-Degenerate Perturbation Theory12 TIME-INDEPENDENT PERTURBATION THEORY
Let us now generalize our perturbation analysis to deal with systems possess-
ing more than two energy eigenstates. Consider a system in which the energy
eigenstates of the unperturbed Hamiltonian, H0 , are denoted
H0 ψn = En ψn , (12.41)
where n runs from 1 to N. The eigenstates are assumed to be orthonormal, so
that
hm|ni = δnm , (12.42)
and to form a complete set. Let us now try to solve the energy eigenvalue problem
for the perturbed Hamiltonian:
(H0 + H1 ) ψE = E ψE . (12.43)
If follows that
hm|H0 + H1 |Ei = E hm|Ei, (12.44)
where m can take any value from 1 to N. Now, we can express ψE as a linear
superposition of the unperturbed energy eigenstates:
X
ψE = hk|Ei ψk , (12.45)
k
where
emk = hm|H1 |ki. (12.47)
hn|Ei = 1, (12.51)
hm|Ei = O(ǫ) (12.52)
for m 6= n. Suppose that we write out Eq. (12.46) for m 6= n, neglecting terms
which are O(ǫ2 ) according to our expansion scheme. We find that
giving
emn
hm|Ei ≃ − . (12.54)
Em − En
Substituting the above expression into Eq. (12.46), evaluated for m = n, and
neglecting O(ǫ3 ) terms, we obtain
X |enk |2
(En − E + enn ) − ≃ 0. (12.55)
k6=n
Ek − En
and a wave-function
X ekn
ψn′ = ψn + ψk + O(ǫ2 ). (12.57)
E − Ek
k6=n n
186
12.5 The Quadratic Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
H1 = e |E| z. (12.59)
Note that the electron spin is irrelevant to this problem (since the spin oper-
ators all commute with H1 ), so we can ignore the spin degrees of freedom of
the system. Hence, the energy eigenstates of the unperturbed Hamiltonian are
characterized by three quantum numbers—the radial quantum number n, and
the two angular quantum numbers l and m (see Sect. 9). Let us denote these
states as the ψnlm , and let their corresponding energy eigenvalues be the Enlm .
According to the analysis in the previous subsection, the change in energy of the
eigenstate characterized by the quantum numbers n, l, m in the presence of a
small electric field is given by
The sum on the right-hand side of the above equation seems very complicated.
However, it turns out that most of the terms in this sum are zero. This follows
because the matrix elements hn, l, m|z|n ′ , l ′ , m ′ i are zero for virtually all choices
of the two sets of quantum number, n, l, m and n ′ , l ′ , m ′ . Let us try to find a set
of rules which determine when these matrix elements are non-zero. These rules
are usually referred to as the selection rules for the problem in hand.
187
12.5 The Quadratic Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
Lz = x py − y px , (12.61)
[Lz , z] = 0. (12.62)
Thus,
m ′ = m. (12.64)
where use has been made of Eqs. (7.15)–(7.17), (8.2)–(8.4), and (8.10). Thus,
188
12.5 The Quadratic Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
which reduces to
[L2 , [L2 , z]] = −h̄2 4 (Lx x + Ly y + Lz z) Lz − 4 (Lx2 + Ly2 + Lz2 ) z
+2 (L2 z − z L2 )
= −h̄2 4 (Lx x + Ly y + Lz z) Lz − 2 (L2 z + z L2 ) . (12.67)
However, it is clear from Eqs. (8.2)–(8.4) that
Lx x + Ly y + Lz z = 0. (12.68)
Hence, we obtain
[L2 , [L2 , z]] = 2 h̄2 (L2 z + z L2 ). (12.69)
Finally, the above expression expands to give
L4 z − 2 L2 z L2 + z L4 − 2 h̄2 (L2 z + z L2 ) = 0. (12.70)
189
12.5 The Quadratic Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
Application of the selection rules (12.64) and (12.74) to Eq. (12.60) yields
X |hn, l, m|z|n ′ , l ′ , mi|2
2 2
∆Enlm = e |E| . (12.75)
n ,l =l±1
′ ′
Enlm − E n ′l ′m
Note that, according to the selection rules, all of the terms in Eq. (12.60) which
vary linearly with the electric field-strength vanish. Only those terms which vary
quadratically with the field-strength survive. Hence, this type of energy-shift of
an atomic state in the presence of a small electric field is known as the quadratic
Stark effect. Now, 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 . (12.76)
2
Hence, we can write
2
X |hn, l, m|z|n ′ , l ′ , mi|2
αnlm = 2e . (12.77)
E n ′ l ′ m − Enlm
n ,l =l±1
′ ′
Unfortunately, there is one fairly obvious problem with Eq. (12.75). Namely,
it predicts an infinite energy-shift if there exists some non-zero matrix element
hn, l, m|z|n ′ , l ′ , mi which couples two degenerate unperturbed energy eigenstates:
i.e., if hn, l, m|z|n ′ , l ′ , mi =
6 0 and Enlm = En ′ l ′ m . Clearly, our perturbation method
breaks down completely in this situation. Hence, we conclude that Eqs. (12.75)
and (12.77) are only applicable to cases where the coupled eigenstates are non-
degenerate. For this reason, the type of perturbation theory employed here is
known as non-degenerate perturbation theory. Now, the unperturbed eigenstates
of a hydrogen atom have energies which only depend on the radial quantum
number n (see Sect. 9). 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).
190
12.5 The Quadratic Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
Here, we have made use of the fact that En10 = En00 . The sum in the above
expression can be evaluated approximately by noting that (see Sect. 9.4)
e2
En00 =− , (12.79)
8π ǫ0 a0 n2
where
4π ǫ0 h̄2
a0 = (12.80)
me e2
is the Bohr radius. Hence, we can write
3 e2
En00 − E100 ≥ E200 − E100 = , (12.81)
4 8π ǫ0 a0
which implies that
16 X
α< 4π ǫ0 a0 |h1, 0, 0|z|n, 1, 0i|2 . (12.82)
3 n>1
However, [see Eq. (12.21)]
X X
2
|h1, 0, 0|z|n, 1, 0i| = h1, 0, 0|z|n, 1, 0i hn, 1, 0|z|1, 0, 0i
n>1 n>1
X
= h1, 0, 0|z|n ′ , l ′ , m ′ i hn ′ , l ′ , m ′ |z|1, 0, 0i
n ′ ,l ′ ,m ′
1
= h1, 0, 0|z2 |1, 0, 0i =
h1, 0, 0|r2 |1, 0, 0i, (12.83)
3
where we have made use of the selection rules, the fact that the ψn ′ ,l ′ ,m ′ form a
complete set, and the fact the the ground-state of hydrogen is spherically sym-
metric. Finally, it follows from Eq. (9.72) that
h1, 0, 0|r2 |1, 0, 0i = 3 a02 . (12.84)
Hence, we conclude that
16
α< 4π ǫ0 a03 ≃ 5.3 4π ǫ0 a03 . (12.85)
3
The exact result (which can be obtained by solving Schrödinger’s equation in
parabolic coordinates) is
9
α = 4π ǫ0 a03 = 4.5 4π ǫ0 a03 . (12.86)
2
191
12.6 Degenerate Perturbation Theory 12 TIME-INDEPENDENT PERTURBATION THEORY
Let us, rather naively, investigate the Stark effect in an excited (i.e., n > 1) state
of the hydrogen atom using standard non-degenerate perturbation theory. We
can write
H0 ψnlm = En ψnlm , (12.87)
since the energy eigenstates of the unperturbed Hamiltonian only depend on the
quantum number n. Making use of the selection rules (12.64) and (12.74), non-
degenerate perturbation theory yields the following expressions for the perturbed
energy levels and eigenstates [see Eqs. (12.56) and (12.57)]:
′
X |en ′ l ′ nl |2
Enl = En + enlnl + , (12.88)
n ′ ,l ′ =l±1
En − E n ′
and X
′ en ′ l ′ nl
ψnlm = ψnlm + ψn ′ l ′ m , (12.89)
n ′ ,l ′ =l±1
En − E n ′
where
en ′ l ′ nl = hn ′ , l ′ , m|H1 |n, l, mi. (12.90)
Unfortunately, if n > 1 then the summations in the above expressions are not
well-defined, because there exist non-zero matrix elements, enl ′ nl , which couple
degenerate eigenstates: i.e., there exist non-zero matrix elements which couple
states with the same value of n, but different values of l. These particular matrix
elements give rise to singular factors 1/(En − En ) in the summations. This does
not occur if n = 1 because, in this case, the selection rule l ′ = l ± 1, and the
fact that l = 0 (since 0 ≤ l < n), only allow l ′ to take the single value 1. Of
course, there is no n = 1 state with l ′ = 1. Hence, there is only one coupled
state corresponding to the eigenvalue E1 . Unfortunately, if n > 1 then there are
multiple coupled states corresponding to the eigenvalue En .
Note that our problem would disappear if the matrix elements of the perturbed
Hamiltonian corresponding to the same value of n, but different values of l, were
all zero: i.e., if
hn, l ′ , m|H1 |n, l, mi = λnl δll ′ . (12.91)
192
12.6 Degenerate Perturbation Theory 12 TIME-INDEPENDENT PERTURBATION THEORY
In this case, all of the singular terms in Eqs. (12.88) and (12.89) would reduce
to zero. Unfortunately, the above equation is not satisfied. Fortunately, we can
always redefine the unperturbed eigenstates corresponding to the eigenvalue En
in such a manner that Eq. (12.91) is satisfied. Suppose that there are Nn coupled
eigenstates belonging to the eigenvalue En . Let us define Nn new states which
are linear combinations of our Nn original degenerate eigenstates:
(1)
X
ψnlm = hn, k, m|n, l(1) , mi ψnkm . (12.92)
k=1,Nn
Note that these new states are also degenerate energy eigenstates of the unper-
(1)
turbed Hamiltonian, H0 , corresponding to the eigenvalue En . The ψnlm are cho-
sen in such a manner that they are also eigenstates of the perturbing Hamiltonian,
H1 : i.e., they are simultaneous eigenstates of H0 and H1 . Thus,
(1) (1)
H1 ψnlm = λnl ψnlm . (12.93)
(1)
The ψnlm are also chosen so as to be orthonormal: i.e.,
hn, l ′(1) , m|n, l(1) , mi = δll ′ . (12.94)
It follows that
hn, l ′(1) , m|H1 |n, l(1) , mi = λnl δll ′ . (12.95)
Thus, if we use the new eigenstates, instead of the old ones, then we can employ
Eqs. (12.88) and (12.89) directly, since all of the singular terms vanish. The only
remaining difficulty is to determine the new eigenstates in terms of the original
ones.
where 1 denotes the identity operator in the sub-space of all coupled unperturbed
eigenstates corresponding to the eigenvalue En . Using this completeness relation,
the eigenvalue equation (12.93) can be transformed into a straightforward matrix
equation:
X
hn, l ′ , m|H1 |n, l ′′ , mi hn, l ′′ , m|n, l(1) , mi = λnl hn, l ′ , m|n, l(1) , mi.
l ′′ =1,Nn
(12.97)
193
12.7 The Linear Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
′
X |en ′ l ′ nl |2
Enl = En + λnl + , (12.101)
n 6=n,l =l±1
′ ′
En − E n ′
and X
(1) ′ (1) en ′ l ′ nl
ψnlm = ψnlm + ψn ′ l ′ m . (12.102)
n ′ 6=n,l ′ =l±1
En − E n ′
There are no singular terms in these expressions, since the summations are over
n ′ 6= n: i.e., they specifically exclude the problematic, degenerate, unperturbed
energy eigenstates corresponding to the eigenvalue En . Note that the first-order
energy shifts are equivalent to the eigenvalues of the matrix equation (12.98).
Returning to the Stark effect, let us examine the effect of an external electric
field on the energy levels of the n = 2 states of a hydrogen atom. There are
four such states: an l = 0 state, usually referred to as 2S, and three l = 1
states (with m = −1, 0, 1), usually referred to as 2P. All of these states possess
the same unperturbed energy, E200 = −e2 /(32π ǫ0 a0 ). As before, the perturbing
Hamiltonian is
H1 = e |E| z. (12.103)
194
12.7 The Linear Stark Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
In order to apply perturbation theory to the ψ200 and ψ210 states, we have to
solve the matrix eigenvalue equation
U x = λ x, (12.104)
where U is the matrix of the matrix elements of H1 between these states. Thus,
0 h2, 0, 0|z|2, 1, 0i
U = e |E| , (12.105)
h2, 1, 0|z|2, 0, 0i 0
where the rows and columns correspond to ψ200 and ψ210 , respectively. Here, we
have again made use of the selection rules, 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. It is easily demonstrated, from the exact forms of
the 2S and 2P wave-functions, that
h2, 0, 0|z|2, 1, 0i = h2, 1, 0|z|2, 0, 0i = 3 a0 . (12.106)
In the absence of an external electric field, both of these states possess the same
energy, E200 . The first-order energy shifts induced by an external electric field are
given by
Thus, in the presence of an electric field, the energies of states 1 and 2 are shifted
upwards and downwards, respectively, by an amount 3 e a0 |E|. These states are
orthogonal linear combinations of the original ψ200 and ψ210 states. Note that
the energy shifts are linear in the electric field-strength, so this effect—which is
known as the linear Stark effect—is much larger than the quadratic effect de-
scribed in Sect. 12.5. Note, also, that the energies of the ψ211 and ψ21−1 states
are not affected by the electric field to first-order. Of course, 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. 12.5).
According to special relativity, the kinetic energy (i.e., the difference between the
total energy and the rest mass energy) of a particle of rest mass m and momentum
p is q
T = p2 c2 + m2 c4 − m c2 . (12.113)
In the non-relativistic limit p ≪ m c, we can expand the square-root in the above
expression to give
p2
!2 !4
1 p p
T= 1− +O . (12.114)
2m 4 mc mc
Hence,
p2 p4
T≃ − . (12.115)
2 m 8 m3 c2
Of course, we recognize the first term on the right-hand side of this equation
as the standard non-relativistic expression for the kinetic energy. The second
196
12.8 The Fine Structure of Hydrogen 12 TIME-INDEPENDENT PERTURBATION THEORY
where
e2 1
α= ≃ (12.122)
4π ǫ0 h̄ c 137
is the dimensionless fine structure constant.
Note that the above derivation implicitly assumes that p4 is an Hermitian op-
erator. It turns out that this is not the case for l = 0 states. However, somewhat
fortuitously, out calculation still gives the correct answer when l = 0. Note, also,
that we are able to use non-degenerate perturbation theory in the above calcula-
tion, using the ψnlm eigenstates, because the perturbing Hamiltonian commutes
with both L2 and Lz . It follows that there is no coupling between states with
different l and m quantum numbers. Hence, all coupled states have different n
quantum numbers, and therefore have different energies.
H1 = −µ · B
e2
= − v×r·S
4π ǫ0 me c2 r3
e2
= L · S, (12.126)
4π ǫ0 me2 c2 r3
198
12.8 The Fine Structure of Hydrogen 12 TIME-INDEPENDENT PERTURBATION THEORY
Now
J=L+S (12.128)
is the total angular momentum of the system. Hence,
J2 = L2 + S2 + 2 L · S, (12.129)
giving
1 2
L·S= (J − L2 − S2 ). (12.130)
2
Recall, from Sect. 11.2, that whilst J2 commutes with both L2 and S2 , it does
not commute with either Lz or Sz . It follows that the perturbing Hamiltonian
(12.127) also commutes with both L2 and S2 , but does not commute with either
Lz or Sz . Hence, the simultaneous eigenstates of the unperturbed Hamiltonian
(12.58) and the perturbing Hamiltonian (12.127) are the same as the simultane-
ous eigenstates of L2 , S2 , and J2 discussed in Sect. 11.3. It is important to know
this since, according to Sect. 12.6, we can only safely apply perturbation theory
to the simultaneous eigenstates of the unperturbed and perturbing Hamiltonians.
(2)
Adopting the notation introduced in Sect. 11.3, let ψl,s;j,mj be a simultaneous
eigenstate of L2 , S2 , J2 , and Jz corresponding to the eigenvalues
(2) (2)
L2 ψl,s;j,mj = l (l + 1) h̄2 ψl,s;j,mj , (12.131)
(2) (2)
S2 ψl,s;j,mj = s (s + 1) h̄2 ψl,s;j,mj , (12.132)
(2) (2)
J2 ψl,s;j,mj = j (j + 1) h̄2 ψl,s;j,mj , (12.133)
(2) (2)
Jz ψl,s;j,mj = mj h̄ ψl,s;j,mj . (12.134)
199
12.8 The Fine Structure of Hydrogen 12 TIME-INDEPENDENT PERTURBATION THEORY
e2 h̄2 1
* +
= 2 2
[j (j + 1) − l (l + 1) − 3/4] 3 . (12.135)
16π ǫ0 me c r
Here, we have made use of the fact that s = 1/2 for an electron. It follows from
Eq. (9.76) that
e2 h̄2
j (j + 1) − l (l + 1) − 3/4
∆El,1/2;j,mj = 2 2 3
, (12.136)
16π ǫ0 me c a0 l (l + 1/2) (l + 1) n3
where n is the radial quantum number. Finally, making use of Eqs. (9.55), (9.57),
and (9.58), the above expression reduces to
α2 n {3/4 + l (l + 1) − j (j + 1)}
∆El,1/2;j,mj = En 2 , (12.137)
n 2 l (l + 1/2) (l + 1)
where α is the fine structure constant. A comparison of this expression with
Eq. (12.121) 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. We can add these two corrections together (making use of
the fact that j = l ± 1/2 for a hydrogen atom—see Sect. 11.3) to obtain a net
energy-shift of
α2 n
3
∆El,1/2;j,mj = En 2 − . (12.138)
n j + 1/2 4
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.
200
12.9 The Zeeman Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
For n = 1, in the absence of fine structure, there are two degenerate 1S1/2
states. According to Eq. (12.138), the fine structure induced energy-shifts of these
two states are the same. Hence, fine structure does not break the degeneracy of
the two 1S1/2 states of hydrogen.
For n = 2, in the absence of fine structure, there are two 2S1/2 states, two
2P1/2 states, and four 2P3/2 states, all of which are degenerate. According to
Eq. (12.138), the fine structure induced energy-shifts of the 2S1/2 and 2P1/2 states
are the same as one another, but are different from the induced energy-shift of the
2P3/2 states. Hence, fine structure does not break the degeneracy of the 2S1/2 and
2P1/2 states of hydrogen, but does break the degeneracy of these states relative to
the 2P3/2 states.
For n = 3, in the absence of fine structure, there are two 3S1/2 states, two
3P1/2 states, four 3P3/2 states, four 3D3/2 states, and six 3D5/2 states, all of which
are degenerate. According to Eq. (12.138), fine structure breaks these states into
three groups: the 3S1/2 and 3P1/2 states, the 3P3/2 and 3D3/2 states, and the 3D5/2
states.
Note, finally, that although expression (12.137) does not have a well-defined
value for l = 0, when added to expression (12.121) it, somewhat fortuitously,
gives rise to an expression (12.138) which is both well-defined and correct when
l = 0.
3D5/2
3P3/2 3D3/2
3S1/2 3P1/2
3D5/2
3P3/2 3D3/2
3S1/2 3P1/2
2P3/2
2S1/2 2P1/2
2P3/2
2S1/2 2P1/2
1S1/2
1S1/2
Figure 20: Effect of the fine structure energy-shift on the n = 1, 2 and 3 states of a hydrogen atom.
Not to scale.
is the total electron magnetic moment, including both orbital and spin contribu-
tions [see Eqs. (10.57)–(10.59)]. Thus,
eB
H1 = (Lz + 2 Sz ). (12.141)
2 me
Suppose that the applied magnetic field is much weaker than the atom’s inter-
nal magnetic field (12.124). Since the magnitude of the internal field is about 25
tesla, this is a fairly reasonable assumption. In this situation, we can treat H1 as
a small perturbation acting on the simultaneous eigenstates of the unperturbed
Hamiltonian and the fine structure Hamiltonian. Of course, these states are the
simultaneous eigenstates of L2 , S2 , J2 , and Jz (see previous subsection). Hence,
from standard perturbation theory, the first-order energy-shift induced by a weak
external magnetic field is
∆El,1/2;j,mj = hl, 1/2; j, mj |H1 |l, 1/2; j, mj i
eB
= (mj h̄ + hl, 1/2; j, mj |Sz |l, 1/2; j, mj i) , (12.142)
2 me
202
12.9 The Zeeman Effect 12 TIME-INDEPENDENT PERTURBATION THEORY
2P3/2 (4/3)ǫ
(4/3)ǫ
(4/3)ǫ
1S1/2
2ǫ
Figure 21: The Zeeman effect for the n = 1 and 2 states of a hydrogen atom. Here, ǫ = µB B. Not to
scale.
In conclusion, in the presence of a weak external magnetic field, the two de-
generate 1S1/2 states of the hydrogen atom are split by 2 µB B. Likewise, the four
degenerate 2S1/2 and 2P1/2 states are split by (2/3) µB B, whereas the four de-
generate 2P3/2 states are split by (4/3) µB B. This is illustrated in Fig. 21. Note,
(2)
finally, that since the ψl,mj are not simultaneous eigenstates of the unperturbed
and perturbing Hamiltonians, Eqs. (12.149) and (12.150) can only be regarded
as the expectation values of the magnetic-field induced energy-shifts. However,
as long as the external magnetic field is much weaker than the internal magnetic
field, these expectation values are almost identical to the actual measured values
of the energy-shifts.
204
12.10 Hyperfine Structure 12 TIME-INDEPENDENT PERTURBATION THEORY
The proton in a hydrogen atom is a spin one-half charged particle, and therefore
possesses a magnetic moment. By analogy with Eq. (10.58), we can write
gp e
µp = Sp , (12.151)
2 mp
where µp is the proton magnetic moment, Sp is the proton spin, and the proton
gyromagnetic ratio gp is found experimentally to take that value 5.59. Note that
the magnetic moment of a proton is much smaller (by a factor of order me /mp )
than that of an electron. According to classical electromagnetism, the proton’s
magnetic moment generates a magnetic field of the form
µ0 h 2 µ0
µp δ3 (r),
i
B= 3
3 (µ p · ^
r) ^
r − µ p + (12.152)
4π r 3
where ^r = r/r. 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. All magnetic field-lines generated by the loop must pass through the loop.
Hence, if the size of the loop goes to zero, then the field will be infinite at the
origin, and this contribution is what is reflected by the delta-function term. Now,
the Hamiltonian of the electron in the magnetic field generated by the proton is
simply
H1 = −µe · B, (12.153)
where
e
Se .
µe = − (12.154)
me
Here, µe is the electron magnetic moment [see Eqs. (10.58) and (10.59)], and Se
the electron spin. Thus, the perturbing Hamiltonian is written
µ0 gp e2 3 (Sp · ^r) (Se · ^r) − Sp · Se µ0 gp e2
H1 = 3
+ Sp · Se δ3 (r). (12.155)
8π mp me r 3 mp me
Note that, since we have neglected coupling between the proton spin and the
magnetic field generated by the electron’s orbital motion, the above expression is
only valid for l = 0 states.
205
12.10 Hyperfine Structure 12 TIME-INDEPENDENT PERTURBATION THEORY
Let
S = Se + Sp (12.158)
be the total spin. We can show that
1 2
Sp · Se = (S − Se2 − Sp2 ). (12.159)
2
Thus, the simultaneous eigenstates of the perturbing Hamiltonian and the main
Hamiltonian are the simultaneous eigenstates of Se2 , Sp2 , and S2 . However, both
the proton and the electron are spin one-half particles. According to Sect. 11.4,
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, there are three spin
1 states, known as triplet states, and a single spin 0 state, known as the singlet
state. For all states, the eigenvalues of Se2 and Sp2 are (3/4) h̄2 . The eigenvalue of
S2 is 0 for the singlet state, and 2 h̄2 for the triplet states. Hence,
3
hSp · Se i = − h̄2 (12.160)
4
for the singlet state, and
1 2
hSp · Se i = h̄ (12.161)
4
for the triplet states.
206
12.10 Hyperfine Structure 12 TIME-INDEPENDENT PERTURBATION THEORY
It follows, from the above analysis, that spin-spin coupling breaks the degen-
eracy of the two 1S1/2 states in hydrogen, lifting the energy of the triplet con-
figuration, and lowering that of the singlet. This splitting is known as hyperfine
structure. The net energy difference between the singlet and the triplet states is
8 me 2
∆E = gp α E0 = 5.88 × 10−6 eV, (12.162)
3 mp
where E0 = 13.6 eV is the (magnitude of the) ground-state energy. Note that the
hyperfine energy-shift is much smaller, by a factor me /mp , than a typical fine
structure energy-shift. If we convert the above energy into a wave-length then
we obtain
λ = 21.1 cm. (12.163)
This is the wave-length 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. 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.
207
13 TIME-DEPENDENT PERTURBATION THEORY
13.1 Introduction
H0 ψm = Em ψm . (13.2)
We know (see Sect. 4.12) that if the system is in one of these eigenstates then, in
the absence of an external perturbation, it remains in this state for ever. However,
the presence of a small time-dependent perturbation can, in principle, give rise to
a finite probability that if the system is initially in some eigenstate ψn of the un-
perturbed 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). In other
words, a time-dependent perturbation can cause the system to make transitions
between its unperturbed energy eigenstates. Let us investigate this effect.
where the cm are complex numbers. Thus, the initial state is some linear su-
perposition of the unperturbed energy eigenstates. In the absence of the time-
dependent perturbation, the time evolution of the system is simply (see Sect. 4.12)
X
ψ(t) = cm exp (−i Em t/h̄) ψm . (13.4)
m
208
13.2 Preliminary Analysis 13 TIME-DEPENDENT PERTURBATION THEORY
where Pn (t) = |cn (t)|2 . Here, we have carefully separated the fast phase oscilla-
tion of the eigenstates, which depends on the unperturbed Hamiltonian, from the
slow variation of the amplitudes cn (t), which depends entirely on the perturba-
tion (i.e., cn is constant in time if H1 = 0). Note that in Eq. (13.7) the eigenstates
ψm are time-independent (they are actually the eigenstates of H0 evaluated at the
initial time, t = 0).
We also have
∂ψ X dcm
!
i h̄ = i h̄ + cm Em exp (−i Em t/h̄) ψm , (13.10)
∂t m dt
since the ψm are time-independent. According to Eq. (13.8), we can equate the
right-hand sides of the previous two equations to obtain
X dcm X
i h̄ exp (−i Em t/h̄) ψm = cm exp (−i Em t/h̄) H1 ψm . (13.11)
m dt m
209
13.3 The Two-State System 13 TIME-DEPENDENT PERTURBATION THEORY
where
Hnm (t) = hn|H1 (t)|mi, (13.13)
and
En − Em
ωnm = . (13.14)
h̄
Suppose that there are N linearly independent eigenstates of the unperturbed
Hamiltonian. According to Eqs. (13.12), the time-dependence of the set of N co-
efficients cn , which specify the probabilities of finding the system in these eigen-
states at time t, is determined by N coupled first-order differential equations.
Note that Eqs. (13.12) are exact—we have made no approximations at this stage.
Unfortunately, we cannot generally find exact solutions to these equations. In-
stead, we have to obtain approximate solutions via suitable expansions in small
quantities. However, for the particuilarly simple case of a two-state system (i.e.,
N = 2), it is actually possible to solve Eqs. (13.12) without approximation. This
solution is of great practical importance.
H0 ψ1 = E1 ψ1 , (13.15)
H0 ψ2 = E2 ψ2 . (13.16)
Suppose, for the sake of simplicity, that the diagonal elements of the interaction
Hamiltonian, H1 , are zero: i.e.,
210
13.3 The Two-State System 13 TIME-DEPENDENT PERTURBATION THEORY
Now, the probability of finding the system in state 1 at time t is simply P1 (t) =
|c1 (t)|2 . Likewise, the probability of finding the system in state 2 at time t is
211
13.3 The Two-State System 13 TIME-DEPENDENT PERTURBATION THEORY
According to the above result, the system starts off in state 1 at t = 0. After a
time interval π/(2 γ) it is certain to be in state 2. After a further time interval
π/(2 γ) it is certain to be in state 1 again, and so on. Thus, the system periodi-
cally flip-flops between states 1 and 2 under the influence of the time-dependent
perturbation. This implies that the system alternatively absorbs and emits energy
from the source of the perturbation.
The absorption-emission cycle also takes place away from the resonance, when
ω 6= ω21 . However, the amplitude of the oscillation in the coefficient c2 is re-
duced. This means that the maximum value of P2 (t) is no longer unity, nor is the
minimum of P1 (t) zero. In fact, if we plot the maximum value of P2 (t) as a func-
tion of the applied frequency, ω, we obtain a resonance curve whose maximum
(unity) lies at the resonance, and whose full-width half-maximum (in frequency)
is 4 γ. Thus, if the applied frequency differs from the resonant frequency by sub-
stantially more than 2 γ then the probability of the system jumping from state 1
to state 2 is always very small. In other words, 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. Clearly, the weaker the
perturbation (i.e., the smaller γ becomes), the narrower the resonance.
212
13.4 Spin Magnetic Resonance 13 TIME-DEPENDENT PERTURBATION THEORY
The eigenstates of the unperturbed Hamiltonian are the “spin up” and “spin
down” states, denoted χ+ and χ− , respectively. Of course, these states are the
eigenstates of Sz corresponding to the eigenvalues +h̄/2 and −h̄/2 respectively
(see Sect. 10). Thus, we have
g e h̄ B0
H0 χ± = ∓ χ± . (13.33)
4m
The time-dependent Hamiltonian can be written
g e B1
H1 = − [exp( i ω t) S− + exp(−i ω t) S+ ] , (13.34)
4m
where S+ and S− are the conventional raising and lowering operators for spin
angular momentum (see Sect. 10). It follows that
h+|H1 |+i = h−|H1 |−i = 0, (13.35)
213
13.4 Spin Magnetic Resonance 13 TIME-DEPENDENT PERTURBATION THEORY
and
g e B1
h−|H1 |+i = h+|H1 |−i∗ = − exp( i ω t). (13.36)
4m
It can be seen that this system is exactly the same as the two-state system
discussed in the previous subsection, provided that the make the following in-
dentifications:
ψ1 → χ+ , (13.37)
ψ2 → χ− , (13.38)
g e B0
ω21 → , (13.39)
2m
g e B1
γ → − . (13.40)
4m
The resonant frequency, ω21 , is simply the spin precession frequency in a uniform
magnetic field of strength B0 (see Sect. 10.6). In the absence of the perturbation,
the expectation values of Sx and Sy oscillate because of the spin precession, but
the expectation value of Sz remains invariant. If we now apply a magnetic pertur-
bation rotating at the resonant frequency then, according to the analysis of the
previous subsection, the system undergoes a succession of spin flips, χ+ ↔ χ− ,
in addition to the spin precession. 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 res-
onance (in frequency) is determined by the strength of the oscillating magnetic
perturbation. Experimentalists are able to measure the gyromagnetic ratios of
spin one-half particles to a high degree of accuracy by placing the particles in a
uniform magnetic field of known strength, and then subjecting them to an oscil-
lating magnetic field whose frequency is gradually scanned. By determining the
resonant frequency (i.e., the frequency at which the particles absorb energy from
the oscillating field), it is possible to determine the gyromagnetic ratio (assuming
that the mass is known).
214
13.5 Perturbation Expansion 13 TIME-DEPENDENT PERTURBATION THEORY
Let us recall the analysis of Sect. 13.2. The ψn are the stationary orthonor-
mal eigenstates of the time-independent unperturbed Hamiltonian, H0 . Thus,
H0 ψn = En ψn , where the En are the unperturbed energy levels, and hn|mi =
δnm . Now, in the presence of a small time-dependent perturbation to the Hamil-
tonian, H1 (t), the wave-function of the system takes the form
X
ψ(t) = cn (t) exp(−i ωn t) ψn , (13.41)
n
where Hnm (t) = hn|H1 (t)|mi and ωnm = (En − Em )/h̄. Finally, the probability of
finding the system in the nth eigenstate at time t is simply
Pn (t) = |cn (t)|2 (13.43)
P 2
(assuing that, initially, n |cn | = 1).
215
13.6 Harmonic Perturbations 13 TIME-DEPENDENT PERTURBATION THEORY
subject to the boundary condition cn(1) (0) = 0. The solution to the above equation
is Z
i t
(1)
cn = − Hni (t ′ ) exp( i ωni t ′ ) dt ′ . (13.47)
h̄ 0
It follows that, up to first-order in our perturbation expansion,
Z
i t
cn (t) = δni − Hni (t ′ ) exp( i ωni t ′ ) dt ′ . (13.48)
h̄ 0
Hence, the probability of finding the system in some final energy eigenstate la-
beled f at time t, given that it is definitely in a different initial energy eigenstate
labeled i at time t = 0, is
Zt 2
2
i ′ ′
′
Pi→f (t) = |cf (t)| = − Hfi (t ) exp( i ωfi t ) dt . (13.49)
h̄ 0
Note, finally, that our perturbative solution is clearly only valid provided
Pi→f (t) ≪ 1. (13.50)
216
13.6 Harmonic Perturbations 13 TIME-DEPENDENT PERTURBATION THEORY
Figure 22: The functions sinc(x) (dashed curve) and sinc2(x) (solid curve). The vertical dotted lines
denote the region |x| ≤ π.
where
sin x
sinc x ≡ . (13.56)
x
Now, the function sinc(x) takes its largest values when |x| <∼ π, and is fairly
negligible when |x| ≫ π (see Fig. 22). Thus, the first and second terms on the
right-hand side of Eq. (13.55) are only non-negligible when
2π
|ω + ωfi | <∼ , (13.57)
t
and
2π
|ω − ωfi | <∼
, (13.58)
t
respectively. Clearly, as t increases, the ranges in ω over which these two terms
are non-negligible gradually shrink in size. Eventually, when t ≫ 2π/|ωfi |, these
217
13.6 Harmonic Perturbations 13 TIME-DEPENDENT PERTURBATION THEORY
two ranges become strongly non-overlapping. Hence, in this limit, Pi→f = |cf |2
yields
t2
2 2 † 2 2
Pi→f (t) = 2 |Vfi | sinc [(ω + ωfi ) t/2] + |Vfi | sinc [(ω − ωfi ) t/2] . (13.59)
h̄
Now, the function sinc2 (x) is very strongly peaked at x = 0, and is completely
negligible for |x| >∼ π (see Fig. 22). It follows that the above expression exhibits
a resonant response to the applied perturbation at the frequencies ω = ±ωfi .
Moreover, the widths of these resonances decease linearly as time increases. At
each of the resonances (i.e., at ω = ±ωfi ), the transition probability Pi→f (t)
varies as t2 [since sinh(0) = 1]. This behaviour is entirely consistent with our
earlier result (13.28), for the two-state system, in the limit γ t ≪ 1 (recall that
our perturbative solution is only valid as long as Pi→f ≪ 1).
Ef − Ei = −h̄ ω. (13.60)
This implies that the system loses energy h̄ ω to the perturbing field, whilst mak-
ing a transition to a final state whose energy is less than the initial state by h̄ ω.
This process is known as stimulated emission. The resonance at ω = ωfi corre-
sponds to
Ef − Ei = h̄ ω. (13.61)
This implies that the system gains energy h̄ ω from the perturbing field, whilst
making a transition to a final state whose energy is greater than that of the initial
state by h̄ ω. This process is known as absorption.
218
13.7 Electromagnetic Radiation 13 TIME-DEPENDENT PERTURBATION THEORY
where we have made use of the facts that ωif = −ωfi > 0, and |Vfi |2 = |Vif† |2 .
Likewise, the transition probability for absorption is
abs t2 † 2 2
Pi→f (t) = 2 |Vfi | sinc [(ω − ωfi ) t/2] . (13.63)
h̄
Let us use the above results to investigate the interaction of an atomic electron
with classical (i.e., non-quantized) electromagnetic radiation.
p → p + q A, (13.65)
H → H − q φ, (13.66)
where A(r) is the vector potential, and φ(r) the scalar potential. Note that
∂A
E = −∇φ − , (13.67)
∂t
B = ∇ × A. (13.68)
(p − e A)2
H= + e φ + V0 (r), (13.69)
2 me
where A and φ are functions of the position operators. The above equation can
be written
p2 − e A·p − e p·A + e2 A2
H= + e φ + V0 (r). (13.70)
2 me
219
13.7 Electromagnetic Radiation 13 TIME-DEPENDENT PERTURBATION THEORY
Now,
p·A = A·p, (13.71)
provided that we adopt the gauge ∇·A = 0. Hence,
p2 e A·p e2 A2
H= − + + e φ + V0 (r). (13.72)
2 me me 2 me
It follows from Eqs. (13.53), (13.63), and (13.79) that the transition probabil-
ity for radiation induced absorption is
t2 e2 |A0 |2
abs
Pi→f (t) = 2 2
|hf|ǫ·p exp( i k·r)|ii| 2 sinc2 [(ω − ωfi ) t/2]. (13.80)
h̄ 4 me
220
13.7 Electromagnetic Radiation 13 TIME-DEPENDENT PERTURBATION THEORY
u= + = ǫ0 |E0 |2 , (13.81)
2 2 2 µ0 2
where E0 = A0 ω and B0 = E0 /c are the peak electric and magnetic field-
strengths, respectively. It thus follows that
t2 e2
abs
Pi→f (t) = 2 2 2
|hf|ǫ·p exp( i k·r)|ii| 2 u sinc2 [(ω − ωfi ) t/2]. (13.82)
2 ǫ0 h̄ me ω
Thus, not surprisingly, the transition probability for radiation induced absorp-
tion (or stimulated emission) is directly proportional to the energy density of the
incident radiation.
Suppose that the incident radiation is not monochromatic, but instead extends
over a range of frequencies. We can write
Z∞
u= ρ(ω) dω, (13.83)
−∞
where ρ(ω) dω is the energy density of radiation whose frequencies lie between
ω and ω + dω. Equation (13.82) generalizes to
Z∞
t2 e2
abs
Pi→f (t) = 2 2 2
|hf|ǫ·p exp( i k·r)|ii| 2 ρ(ω) sinc2 [(ω − ωfi ) t/2] dω.
−∞ 2 ǫ0 h̄ me ω
(13.84)
Note, however, that the above expression is only valid provided the radiation in
question is incoherent: i.e., there are no phase correlations between waves of
different frequencies. This follows because it is permissible to add the intensities
of incoherent radiation, whereas we must always add the amplitudes of coherent
radiation. Given that the function sinc2 [(ω−ωfi ) t/2] is very strongly peaked (see
Fig. 22) about ω = ωfi (assuming that t ≫ 2π/ωfi ), and
Z∞
sinc2 (x) dx = π, (13.85)
−∞
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. (13.82)] to one which is proportional to t. As has already been explained,
abs
the above expression is only valid when Pi→f ≪ 1. However, the result that
abs
dPi→f π e2 ρ(ωfi ) 2
wabs
i→f ≡ = 2 2
|hf|ǫ·p exp( i k·r)|ii| (13.87)
dt ǫ0 h̄ me2 ωfi
is constant in time is universally valid. Here, wabs
i→f is the transition probability
per unit time interval, otherwise known as the transition rate. Given that the
transition rate is constant, we can write (see Sect. 2)
abs abs abs
(t) wabs
h i
Pi→f (t + dt) − Pi→f (t) = 1 − Pi→f i→f dt : (13.88)
i.e., 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
abs
time interval dt—the probabilities of these two events are 1−Pi→f (t) and wabs
i→f dt,
respectively. It follows that
abs
dPi→f
+ wabs abs abs
i→f Pi→f = wi→f , (13.89)
dt
abs
with the initial condition Pi→f (0) = 0. The above equation can be solved to give
abs
(t) = 1 − exp wabs
Pi→f i→f t . (13.90)
Using similar arguments to the above, the transition probability for stimulated
emission can be shown to take the form
stm
(t) = 1 − exp wstm
Pi→f i→f t , (13.91)
222
13.8 The Electric Dipole Approximation 13 TIME-DEPENDENT PERTURBATION THEORY
can be approximated by its first term, unity. This approach is known as the electric
dipole approximation. It follows that
Equations (13.97) and (13.98) give the transition rates for absorption and
stimulated emission, respectively, induced by a linearly polarized plane-wave.
Actually, we are more interested in the transition rates induced by unpolarized
isotropic radiation. To obtain these we must average Eqs. (13.97) and (13.98)
223
13.8 The Electric Dipole Approximation 13 TIME-DEPENDENT PERTURBATION THEORY
over all possible polarizations and propagation directions of the wave. To fa-
cilitate this process, we can define a set of cartesian coordinates such that the
wave-vector k, which specifies the direction of wave propagation, points along
the z-axis, and the vector dif , which specifies the direction of the atomic dipole
moment, lies in the x-z plane. It follows that the vector ǫ, which specifies the
direction of wave polarization, must lie in the x-y plane, since it has to be orthog-
onal to k. Thus, we can write
Hence, the transition rates for absorption and stimulated emission induced by
unpolarized isotropic radiation are
π
wabs
i→f =
2
2 dif ρ(ωfi ), (13.107)
3 ǫ0 h̄
π
wstm
i→f = 2
2 dif ρ(ωif ), (13.108)
3 ǫ0 h̄
respectively.
224
13.9 Spontaneous Emission 13 TIME-DEPENDENT PERTURBATION THEORY
So far, we have calculated the rates of radiation induced transitions between two
atomic states. This process is known as absorption when the energy of the final
state exceeds that of the initial state, and stimulated emission when the energy
of the final state is less than that of the initial state. Now, in the absence of any
external radiation, we would not expect an atom in a given state to spontaneously
jump into an state with a higher energy. On the other hand, 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. This process is known as spontaneous emission.
Consider two atomic states, labeled i and f, with Ei > Ef . One of the tenants
of statistical thermodynamics is that in thermal equilibrium we have so-called
detailed balance. This means that, irrespective of any other atomic states, 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. The former rate (i.e., number of transitions per unit time in the ensemble)
is written
Wi→f = Ni (wspn stm
i→f + wi→f ), (13.110)
where wspn
i→f is the rate of spontaneous emission (for a single atom) between states
225
13.9 Spontaneous Emission 13 TIME-DEPENDENT PERTURBATION THEORY
i and f, and Ni is the number of atoms in the ensemble in state i. Likewise, the
latter rate takes the form
Wf→i = Nf wabs
f→i , (13.111)
where Nf is the number of atoms in the ensemble in state f. The above expres-
sions describe how atoms in the ensemble make transitions from state i to state
f due to a combination of spontaneous and stimulated emission, and make the
opposite transition as a consequence of absorption. In thermal equilibrium, we
have Wi→f = Wf→i , which gives
Nf abs
wspn
i→f = wf→i − wstm
i→f . (13.112)
Ni
According to Eqs. (13.107) and (13.108), we can also write
Nf π
!
wspn
i→f = −1 2
2 dif ρ(ωif ). (13.113)
Ni 3 ǫ0 h̄
Now, another famous result in statistical thermodynamics is that in thermal equi-
librium the number of atoms in an ensemble occupying a state of energy E is
proportional to exp(−E/kB T ). This implies that
Nf exp(−Ef /kB T )
= = exp( h̄ ωif /kB T ). (13.114)
Ni exp(−Ei /kB T )
Thus, it follows from Eq. (13.109), (13.113), and (13.114) that the rate of spon-
taneous emission between states i and f takes the form
ωif3 dif2
wspn
i→f = . (13.115)
3π ǫ0 h̄ c3
Note, that, although the above result has been derived for an atom in a radiation-
filled cavity, it remains correct even in the absence of radiation. Finally, the cor-
responding absorption and stimulated emission rates for an atom in a radiation-
filled cavity are
ωfi3 dif2 1
wabs
i→f = , (13.116)
3π ǫ0 h̄ c3 exp(h̄ ωfi /kB T ) − 1
ωif3 dif2 1
wstm
i→f = , (13.117)
3π ǫ0 h̄ c3 exp(h̄ ωif /kB T ) − 1
226
13.10 Radiation from a Harmonic Oscillator 13 TIME-DEPENDENT PERTURBATION THEORY
respectively.
Let us estimate the typical value of the spontaneous emission rate for a hydro-
gen atom. We expect the dipole moment dif to be of order e a0 , where a0 is the
Bohr radius [see Eq. (9.58)]. We also expect ωif to be of order |E0 |/h̄, where E0 is
the energy of the ground-state [see Eq. (9.57)]. It thus follows from Eq. (13.115)
that
wspn 3
i→f ∼ α ωif , (13.118)
where α = e2 /(4π ǫ0 h̄ c) ≃ 1/137 is the fine-structure constant. This is an impor-
tant result, since our perturbation expansion is based on the assumption that the
transition rate between different energy eigenstates is much slower than the fre-
quency of phase oscillation of these states: i.e., that wspn
i→f ≪ ωif (see Sect. 13.2).
This is indeed the case.
Suppose that the electron is initially in an excited state: i.e., n > 0. In princi-
ple, the electron can decay to a lower energy state via the spontaneous emission
of a photon of the appropriate frequency. Let us investigate this effect. Now, ac-
cording to Eq. (13.115), 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
Z∞
′
(dx )n,n ′ = hn|e x|n i = e ψn (x) x ψn ′ (x) dx (13.120)
−∞
227
13.10 Radiation from a Harmonic Oscillator 13 TIME-DEPENDENT PERTURBATION THEORY
2
is non-zero [since dif ≡ (dx )n,n ′ for the case in hand]. However, according to
Eq. (5.117),
Z∞ v
√ √
h̄
u
′
u
ψn x ψn ′ dx = t
n δn,n ′ +1 + n δn,n ′ −1 . (13.121)
−∞ 2 me ω0
Since we are dealing with emission, we must have n > n ′ . Hence, we obtain
v
h̄ n
u
u
(dx )n,n ′ = e t
δn,n ′ +1 . (13.122)
2 me ω0
It is clear that (in the electric dipole approximation) we can only have sponta-
neous emission between states whose quantum numbers differ by unity. Thus,
the frequency of the photon emitted when the nth excited state decays is
En − En−1
ωn,n−1 = = ω0 . (13.123)
h̄
Hence, we conclude that, no matter which state decays, the emitted photon al-
ways has the same frequency as the classical oscillator.
According to Eq. (13.115), the decay rate of the nth excited state is given by
3 2
ωn,n−1 (dx )n,n−1
wn = . (13.124)
3π ǫ0 h̄ c3
It follows that
n e2 ω02
wn = . (13.125)
6π ǫ0 me c3
The mean radiated power is simply
e2 ω02
Pn = h̄ ω0 wn = [En − (1/2) h̄ ω0 ]. (13.126)
6π ǫ0 me c3
Classically, an electron in a one-dimensional oscillator potential radiates at the
oscillation frequency ω0 with the mean power
e2 ω02
P= E, (13.127)
6π ǫ0 me c3
where E is the oscillator energy. It can be seen that a quantum oscillator radiates
in an almost exactly analogous manner to the equivalent classical oscillator. The
only difference is the factor (1/2) h̄ ω0 in Eq. (13.126)—this is needed to ensure
that the ground-state of the quantum oscillator does not radiate.
228
13.11 Selection Rules 13 TIME-DEPENDENT PERTURBATION THEORY
Let us now consider spontaneous transitions between the different energy levels
of a hydrogen atom. Since the perturbing Hamiltonian (13.77) does not contain
any spin operators, we can neglect electron spin in our analysis. Thus, according
to Sect. 9.4, the various energy eigenstates of the hydrogen atom are labeled by
the familiar quantum numbers n, l, and m.
According to Eqs. (13.106) and (13.115), a hydrogen atom can only make
a spontaneous transition from an energy state corresponding to the quantum
numbers n, l, m to one corresponding to the quantum numbers n ′ , l ′ , m ′ if the
modulus squared of the associated electric dipole moment
d2 = |hn, l, m|e x|n ′ , l ′ , m ′ i|2 + |hn, l, m|e y|n ′ , l ′ , m ′ i|2 + |hn, l, m|e z|n ′ , l ′ , m ′ i|2
(13.128)
is non-zero. Now, we have already seen, in Sect. 12.5, that the matrix element
hn, l, m|z|n ′ , l ′ , m ′ i is only non-zero provided that m ′ = m and l ′ = l ± 1. It
turns out that the proof that this matrix element is zero unless l ′ = l ± 1 can, via
a trivial modification, also be used to demonstrate that hn, l, m|x|n ′ , l ′ , m ′ i and
hn, l, m|y|n ′ , l ′ , m ′ i are also zero unless l ′ = l ± 1. Consider
x± = x + i y. (13.129)
[Lz , x± ] = ± h̄ x± . (13.130)
Hence,
229
13.12 2P → 1S Transitions in Hydrogen 13 TIME-DEPENDENT PERTURBATION THEORY
are obviously both zero if hn, l, m|x+ |n ′ , l ′ , m ′ i and hn, l, m|x− |n ′ , l ′ , m ′ i are both
zero. Hence, we conclude that hn, l, m|x|n ′ , l ′ , m ′ i and hn, l, m|y|n ′ , l ′ , m ′ i are
only non-zero if m ′ = m ± 1.
l ′ = l ± 1, (13.133)
m ′ = m, m ± 1. (13.134)
These are termed the selection rules for electric dipole transitions (i.e., transitions
calculated using the electric dipole approximation). Note, finally, that since the
perturbing Hamiltonian does not contain any spin operators, the spin quantum
number ms cannot change during a transition. Hence, we have the additional
selection rule that ms′ = ms .
Let us calculate the rate of spontaneous emission between the first excited state
(i.e., n = 2) and the ground-state (i.e., n ′ = 1) of a hydrogen atom. Now the
ground-state is characterized by l ′ = m ′ = 0. Hence, in order to satisfy the se-
lection rules (13.133) and (13.134), the excited state must have the quantum
numbers l = 1 and m = 0, ±1. Thus, we are dealing with a spontaneous tran-
sition from a 2P to a 1S state. Note, incidentally, that a spontaneous transition
from a 2S to a 1S state is forbidden by our selection rules.
According to Sect. 9.4, the wave-function of a hydrogen atom takes the form
where the radial functions Rn,l are given in Sect. 9.4, and the spherical harmonics
Yl,m are given in Sect. 8.7. Some straight-forward, but tedious, integration reveals
that
27
h1, 0, 0|x|2, 1, ±1i = ± 5 a0 , (13.136)
3
230
13.12 2P → 1S Transitions in Hydrogen 13 TIME-DEPENDENT PERTURBATION THEORY
27
h1, 0, 0|y|2, 1, ±1i = i 5 a0 , (13.137)
3
√ 27
h1, 0, 0|z|2, 1, 0i = 2 5 a0 , (13.138)
3
where a0 is the Bohr radius specified in Eq. (9.58). All of the other possible
2P → 1S matrix elements are zero because of the selection rules. If follows from
Eq. (13.128) that the modulus squared of the dipole moment for the 2P → 1S
transition takes the same value
215
d = 10 (e a0 )2
2
(13.139)
3
for m = 0, 1, or −1. Clearly, the transition rate is independent of the quantum
number m. It turns out that this is a general result.
Now, the energy of the eigenstate of the hydrogen atom characterized by the
quantum numbers n, l, m is E = E0 /n2 , where the ground-state energy E0 is
specified in Eq. (9.57). Hence, the energy of the photon emitted during a 2P → 1S
transition is
E0 /4 − E0 3
h̄ ω = = − E0 = 10.2 eV. (13.140)
h̄ 4
This corresponds to a wave-length of 1.215 × 10−7 m.
231
13.13 Intensity Rules 13 TIME-DEPENDENT PERTURBATION THEORY
Incidentally, since the 2P state only has a finite life-time, it follows from the
energy-time uncertainty relation that the energy of this state is uncertain by an
amount
h̄
∆E2P ∼ ∼ 4 × 10−7 eV. (13.144)
τ2P
This uncertainty gives rise to a finite width of the spectral line associated with the
2P → 1S transition. This natural line-width is of order
∆λ ∆E2P
∼ ∼ 4 × 10−8 . (13.145)
λ h̄ ω
Now, we know, from Sect. 12.8, 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. In fact, these states are divided into two groups with slightly different
energies. There are four states characterized by the overall angular momentum
quantum number j = 3/2—these are called the 2P3/2 states. The remaining two
states are characterized by j = 1/2, and are thus called the 2P1/2 states. The
energy of the 2P3/2 states is slightly higher than that of the 2P1/2 states. In fact,
the energy difference is
α2
∆E = − E0 = 4.53 × 10−5 eV. (13.146)
16
Thus, the wave-length of the spectral line associated with the 2P → 1S transition
in hydrogen is split by a relative amount
∆λ ∆E
= = 4.4 × 10−6 . (13.147)
λ h̄ ω
Note that this splitting is much greater than the natural line-width estimated in
Eq. (13.145), so there really are two spectral lines. How does all of this affect the
rate of the 2P → 1S transition?
Well, we have seen that the transition rate is independent of spin, and hence
of the spin quantum number ms , and is also independent of the quantum number
232
13.14 Forbidden Transitions 13 TIME-DEPENDENT PERTURBATION THEORY
Atomic transitions which are forbidden by the electric dipole selection rules (13.133)
and (13.134) are unsurprisingly known as forbidden transitions. It is clear from
the analysis in Sect. 13.8 that a forbidden transition is one for which the matrix
element hf|ǫ · p|ii is zero. However, this matrix element is only an approxima-
tion to the true matrix element for radiative transitions, which takes the form
hf|ǫ·p exp( i k·r)|ii. Expanding exp( i k·r), and keeping the first two terms, the
matrix element for a forbidden transition becomes
hf|ǫ·p exp( i k·r)|ii ≃ i hf|(ǫ·p) (k·r)|ii. (13.148)
Hence, if the residual matrix element on the right-hand side of the above expres-
sion is non-zero then a “forbidden” transition can take place, allbeit at a much
reduced rate. In fact, in Sect. 13.9, we calculated that the typical rate of an
electric dipole transition is
wi→f ∼ α3 ωif . (13.149)
Since the transition rate is proportional to the square of the radiative matrix
element, it is clear that the transition rate for a forbidden transition enabled by
the residual matrix element (13.148) is smaller than that of an electric dipole
transition by a factor (k r)2 . Estimating r as the Bohr radius, and k as the wave-
number of a typical spectral line of hydrogen, it is easily demonstrated that
wi→f ∼ α5 ωif (13.150)
233
13.14 Forbidden Transitions 13 TIME-DEPENDENT PERTURBATION THEORY
for such a transition. Of course, there are some transitions (in particular, the
2S → 1S transition) for which the true radiative matrix element hf|ǫ·p exp( i k·r)|ii
is zero. Such transitions are absolutely forbidden.
Finally, it is fairly obvious that excited states which decay via forbidden tran-
sitions have much longer life-times than those which decay via electric dipole
transitions. 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 associ-
ated with forbidden transitions are generally much sharper than those associated
with electric dipole transitions.
234
14 VARIATIONAL METHODS
14 Variational Methods
14.1 Introduction
We have seen, in Sect. 9.4, that we can solve Schrödinger’s equation exactly to
find the stationary eigenstates of a hydrogen atom. Unfortunately, it is not possi-
ble to find exact solutions of Schrödinger’s equation for atoms more complicated
than hydrogen, or for molecules. In such systems, the best that we can do is to
find approximate solutions. Most of the methods which have been developed for
finding such solutions employ the so-called variational principle discussed below.
H ψ = E ψ, (14.1)
E0 ≤ hψ|H|ψi. (14.2)
Let us prove the variational principle. Suppose that the ψn and the En are the
true eigenstates and eigenvalues of H: i.e.,
H ψn = En ψn . (14.3)
Furthermore, let
E0 < E1 < E2 < · · · , (14.4)
235
14.2 The Variational Principle 14 VARIATIONAL METHODS
so that ψ0 is the ground-state, ψ1 the first excited state, etc. The ψn are assumed
to be orthonormal: i.e.,
hψn |ψm i = δnm . (14.5)
If our trial wave-function ψ is properly normalized then we can write
X
ψ= cn ψn , (14.6)
n
where X
|cn | 2 = 1. (14.7)
n
Now, the expectation value of H, calculated with ψ, takes the form
*X
X + X
hψ|H|ψi = cn ψn H cm ψm = cn∗ cm hψn |H|ψm i
n m n,m
X X
= cn∗ cm Em hψn |ψm i = En |cn | 2 , (14.8)
n n
where use has been made of Eqs. (14.3) and (14.5). So, we can write
X
hψ|H|ψi = |c0 | 2 E0 + |cn | 2 En . (14.9)
n>0
Now, the second term on the right-hand side of the above expression is positive
definite, since En − E0 > 0 for all n > 0 [see (14.4)]. Hence, we obtain the desired
result
hψ|H|ψi ≥ E0 . (14.12)
236
14.3 The Helium Atom 14 VARIATIONAL METHODS
ψ̃0 (i.e., hψ|ψ̃0 i = 0) then, by repeating the above analysis, we can easily demon-
strate that
hψ|H|ψi ≥ E1 . (14.13)
Thus, by varying ψ until the expectation value of H is minimized, we can obtain
an approximation to the wave-function and energy of the first excited state. Ob-
viously, we can continue this process until we have approximations to all of the
stationary eigenstates. Note, however, that the errors are clearly cumulative in
this method, so that any approximations to highly excited states are unlikely to
be very accurate. For this reason, the variational method is generally only used
to calculate the ground-state and first few excited states of complicated quantum
systems.
Let the nucleus lie at the origin of our coordinate system, and let the position
vectors of the two electrons be r1 and r2 , respectively. The Hamiltonian of the
system thus takes the form
h̄2 2 e2 2
2
2 1
H=− ∇1 + ∇2 − + − , (14.14)
2 me 4π ǫ0 r1 r2 |r2 − r1 |
where we have neglected any reduced mass effects. The terms in the above
expression represent the kinetic energy of the first electron, the kinetic energy of
the second electron, the electrostatic attraction between the nucleus and the first
electron, the electrostatic attraction between the nucleus and the second electron,
and the electrostatic repulsion between the two electrons, respectively. It is the
final term which causes all of the difficulties. Indeed, if this term is neglected
then we can write
H = H1 + H2 , (14.15)
where
h̄2 2 2 e2
H1,2 =− ∇ − . (14.16)
2 me 1,2 4π ǫ0 r1,2
237
14.3 The Helium Atom 14 VARIATIONAL METHODS
In other words, the Hamiltonian just becomes the sum of separate Hamiltonians
for each electron. In this case, we would expect the wave-function to be separa-
ble: i.e.,
ψ(r1 , r2 ) = ψ1 (r1 ) ψ2 (r2 ). (14.17)
Hence, Schrödinger’s equation
Hψ = Eψ (14.18)
reduces to
H1,2 ψ1,2 = E1,2 ψ1,2 , (14.19)
where
E = E1 + E2 . (14.20)
Of course, Eq. (14.19) is the Schrödinger equation of a hydrogen atom whose
nuclear charge is +2 e, instead of +e. It follows, from Sect. 9.4 (making the
substitution e2 → 2 e2 ), that if both electrons are in their lowest energy states
then
ψ1 (r1 ) = ψ0 (r1 ), (14.21)
ψ2 (r2 ) = ψ0 (r2 ), (14.22)
where
4 2r
!
ψ0 (r) = √ 3/2
exp − . (14.23)
2 π a0 a0
Here, a0 is the Bohr radius [see Eq. (9.58)]. Note that ψ0 is properly normalized.
Furthermore,
E1 = E2 = 4 E0 , (14.24)
where E0 = −13.6 eV is the hydrogen ground-state energy [see Eq. (9.57)]. Thus,
our crude estimate for the ground-state energy of helium becomes
E = 4 E0 + 4 E0 = 8 E0 = −108.8 eV. (14.25)
Unfortunately, this estimate is significantly different from the experimentally de-
termined value, which is −78.98 eV. This fact demonstrates that the neglected
electron-electron repulsion term makes a large contribution to the helium ground-
state energy. Fortunately, however, we can use the variational principle to esti-
mate this contribution.
238
14.3 The Helium Atom 14 VARIATIONAL METHODS
Let us employ the separable wave-function discussed above as our trial solu-
tion. Thus,
8 2 [r1 + r2 ]
ψ(r1 , r2 ) = ψ0 (r1 ) ψ0 (r2 ) = exp − . (14.26)
π a03 a0
The expectation value of the Hamiltonian (14.14) thus becomes
where
Z
e2 e2 |ψ(r1 , r2 )| 2 3
* +
hVee i = ψ ψ = d r1 d3 r2 . (14.28)
4π ǫ0 |r2 − r1 | 4π ǫ0 |r2 − r1 |
The variation principle only guarantees that (14.27) yields an upper bound on the
ground-state energy. In reality, we hope that it will give a reasonably accurate
estimate of this energy.
239
14.3 The Helium Atom 14 VARIATIONAL METHODS
Our first task is to evaluate the function I(r2 ). Let (r1 , θ1 , φ1 ) be a set of spher-
ical polar coordinates in r1 space whose axis of symmetry runs in the direction of
r2 . It follows that θ = θ1 . Hence,
Z ∞ Z π Z 2π
e−2 r1
I(r2 ) = q
2 2
r12 dr1 sin θ1 dθ1 dφ1 , (14.33)
0 0 0 r1 + r2 − 2 r1 r2 cos θ1
which trivially reduces to
Z∞ Zπ
e−2 r1
I(r2 ) = 2 π q
2 2
r12 dr1 sin θ1 dθ1 . (14.34)
0 0 r1 + r2 − 2 r1 r2 cos θ1
Making the substitution µ = cos θ1 , we can see that
Zπ Z1
1 dµ
q
2 + r 2 − 2 r r cos θ
sin θ1 dθ1 = q
2 + r2 − 2 r r µ
. (14.35)
0 r 1 2 1 2 1 −1 r 1 2 1 2
Now,
Z1 q −1
dµ r12 + r22− 2 r1 r2 µ
q =
−1 r12 + r22 − 2 r1 r2 µ r1 r2
+1
(r1 + r2 ) − |r1 − r2 |
=
r1 r2
2/r1 for r1 > r2
= , (14.36)
2/r2 for r1 < r2
giving Z r2 Z∞
1 −2 r1 −2 r1
I(r2 ) = 4π e r12 dr1 + e r1 dr1 . (14.37)
r2 0 r2
But,
Z
−β x e−β x
e x dx = − 2 (1 + β x), (14.38)
β
Z
e−β x
e−β x x2 dx = − 3
(2 + 2 β x + β2 x2 ), (14.39)
β
yielding
πh
1 − e−2 r2 (1 + r2 ) .
i
I(r2 ) = (14.40)
r2
240
14.3 The Helium Atom 14 VARIATIONAL METHODS
Since the function I(r2 ) only depends on the magnitude of r2 , the integral
(14.31) reduces to
Z
16 E0 ∞ −2 r2
hVee i = − e I(r2 ) r22 dr2 , (14.41)
π 0
which yields
Z∞
5
e−2 r2 1 − e−2 r2 (1 + r2 ) r2 dr2 = − E0 .
h i
hVee i = −16 E0 (14.42)
0 2
Hence, from (14.27), our estimate for the ground-state energy of helium is
5 11
hHi = 8 E0 − E0 = E0 = −74.8 eV. (14.43)
2 2
This is remarkably close to the correct result.
We can actually refine our estimate further. The trial wave-function (14.26)
essentially treats the two electrons as non-interacting particles. In reality, we
would expect one electron to partially shield the nuclear charge from the other,
and vice versa. Hence, a better trial wave-function might be
Z3
Z [r1 + r2 ]
ψ(r1 , r2 ) = exp − , (14.44)
π a03 a0
where Z < 2 is effective nuclear charge number seen by each electron. Let us
recalculate the ground-state energy of helium as a function of Z, using the above
trial wave-function, and then minimize the result with respect to Z. According
to the variational principle, this should give us an even better estimate for the
ground-state energy.
We can rewrite the expression (14.14) for the Hamiltonian of the helium atom
in the form
H = H1 (Z) + H2 (Z) + Vee + U(Z), (14.45)
where
h̄2 2 Z e2
H1,2 (Z) = − ∇ − (14.46)
2 me 1,2 4π ǫ0 r1,2
241
14.3 The Helium Atom 14 VARIATIONAL METHODS
U(Z) = + . (14.48)
4π ǫ0 r1 r2
It follows that
hHi(Z) = 2 E0 (Z) + hVee i(Z) + hUi(Z), (14.49)
where E0 (Z) = Z2 E0 is the ground-state energy of a hydrogen atom with nu-
clear charge +Z e, hVee i(Z) = −(5 Z/4) E0 is the value of the electron-electron
repulsion term when recalculated with the wave-function (14.44) [actually, all
we need to do is to make the substitution a0 → (2/Z) a0 ], and
e2 1
* +
hUi(Z) = 2 (Z − 2) . (14.50)
4π ǫ0 r
Here, h1/ri is the expectation value of 1/r calculated for a hydrogen atom with
nuclear charge +Z e. It follows from Eq. (9.74) [with n = 1, and making the
substitution a0 → a0 /Z] that
1 Z
* +
= . (14.51)
r a0
Hence,
hUi(Z) = −4 Z (Z − 2) E0 , (14.52)
since E0 = −e2 /(8π ǫ0 a0 ). Collecting the various terms, our new expression for
the expectation value of the Hamiltonian becomes
5 27
" # " #
2
hHi(Z) = 2 Z − Z − 4 Z (Z − 2) E0 = −2 Z2 + Z E0 . (14.53)
4 4
The value of Z which minimizes this expression is the root of
dhHi 27
" #
= −4 Z + E0 = 0. (14.54)
dZ 4
242
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
It follows that
27
Z= = 1.69. (14.55)
16
The fact that Z < 2 confirms our earlier conjecture that the electrons partially
shield the nuclear charge from one another. Our new estimate for the ground-
state energy of helium is
!6
1 3
hHi(1.69) = E0 = −77.5 eV. (14.56)
2 2
This is clearly an improvement on our previous estimate (14.43) [recall that the
correct result is −78.98 eV].
Obviously, we could get even closer to the correct value of the helium ground-
state energy by using a more complicated trial wave-function with more ad-
justable parameters.
Note, finally, that since the two electrons in a helium atom are indistinguish-
able fermions, the overall wave-function must be anti-symmetric with respect to
exchange of particles (see Sect. 6). Now, the overall wave-function is the prod-
uct of the spatial wave-function and the spinor representing the spin-state. Our
spatial wave-function (14.44) is obviously symmetric with respect to exchange of
particles. This means that the spinor must be anti-symmetric. It is clear, from
Sect. 11.4, that if the spin-state of an l = 0 system consisting of two spin one-
half particles (i.e., 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.
Hence, the ground-state of helium has overall electron spin zero.
The hydrogen molecule ion consists of an electron orbiting about two protons,
and is the simplest imaginable molecule. Let us investigate whether or not this
molecule possesses a bound state: i.e., whether or not it possesses a ground-state
whose energy is less than that of a hydrogen atom and a free proton. According
to the variation principle, we can deduce that the H+
2 ion has a bound state if we
243
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
z-axis
proton
z=R
r2
electron
r1
z=0 proton
can find any trial wave-function for which the total Hamiltonian of the system
has an expectation value less than that of a hydrogen atom and a free proton.
Suppose that the two protons are separated by a distance R. In fact, let them
lie on the z-axis, with the first at the origin, and the second at z = R (see Fig. 23).
In the following, we shall treat the protons as essentially stationary. This is rea-
sonable, since the electron moves far more rapidly than the protons.
Let us try
ψ(r)± = A [ψ0 (r1 ) ± ψ0 (r2 )] (14.57)
as our trial wave-function, where
1
ψ0 (r) = √ 3/2
e−r/a0 (14.58)
π a0
is a normalized hydrogen ground-state wave-function centered on the origin, and
r1,2 are the position vectors of the electron with respect to each of the protons (see
Fig. 23). Obviously, this is a very simplistic wave-function, since it is just a linear
combination of hydrogen ground-state wave-functions centered on each proton.
244
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
Note, however, that the wave-function respects the obvious symmetries in the
problem.
It follows that
I = 2 (1 ± J), (14.61)
with Z
J= ψ0 (r1 ) ψ0 (r2 ) d3 r. (14.62)
Let us employ the standard spherical polar coordinates (r, θ, φ). Now, it is
easily seen that r1 = r and r2 = (r2 + R2 − 2 r R cos θ)1/2 . Hence,
Z∞ Zπ
exp −x − (x2 + X2 − 2 x X cos θ)1/2 x2 dx sin θ dθ,
h i
J=2 (14.63)
0 0
where X = R/a0 . Here, we have already performed the trivial φ integral. Let
y = (x2 + X2 − 2 x X cos θ)1/2 . It follows that d(y2 ) = 2 y dy = 2 x X sin θ dθ,
giving
Zπ Z
(x2 +X2 −2 x X cos θ)1/2 1 x+X −y
e sin θ dθ = e y dy (14.64)
0 x X |x−X|
1 h −(x+X)
(1 + x + X) − e−|x−X| (1 + |x − X|) .
i
= − e
xX
Thus,
Z
2 −X X h −2 x i
J = − e e (1 + X + x) − (1 + X − x) x dx
X 0
Z∞
2
e−2 x e−X (1 + X + x) − eX (1 − X + x) x dx,
h i
− (14.65)
X X
245
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
which evaluates to
X3
−X
J=e 1+X+ . (14.66)
3
= E0 ψ − A ± . (14.69)
4π ǫ0 r2 r1
Hence,
hHi = E0 + 4 A2 (D ± E) E0 , (14.70)
where
a
* +
0
D = ψ0 (r1 ) ψ0 (r1 ) , (14.71)
r2
a
* +
0
E = ψ0 (r1 ) ψ0 (r2 ) . (14.72)
r1
Now,
Z∞ Zπ
e−2 x
D=2 2 2 1/2
x2 dx sin θ dθ, (14.73)
0 0 (x + X − 2 x X cos θ)
which reduces to
ZX Z∞
4
D= e−2 x x2 dx + 4 e−2 x x dx, (14.74)
X 0 X
246
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
giving
1
1 − [1 + X] e−2 X .
D= (14.75)
X
Furthermore,
Z∞ Zπ
exp −x − (x2 + X2 − 2 x X cos θ)1/2 x dx sin θ dθ,
h i
E=2 (14.76)
0 0
which reduces to
Z
2 −X X h −2 x i
E = − e e (1 + X + x) − (1 + X − x) dx
X 0
Z∞
2
e−2 x e−X (1 + X + x) − eX (1 − X + x) dx,
h i
− (14.77)
X X
yielding
E = (1 + X) e−X . (14.78)
F± (X) = −1 + . (14.82)
X 1 ± (1 + X + X2 /3) e−X
The functions F+ (X) and F− (X) are both plotted in Fig. 24. Recall that in order for
the H+2 ion to be in a bound state it must have a lower energy than a hydrogen
247
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
atom and a free proton: i.e., Etotal < E0 . It follows from Eq. (14.81) that a
bound state corresponds to F± < −1. Clearly, the even trial wave-function ψ+
possesses a bound state, whereas the odd trial wave-function ψ− does not [see
Eq. (14.57)]. This is hardly surprising, since the even wave-function maximizes
the electron probability density between the two protons, thereby reducing their
mutual electrostatic repulsion. On the other hand, the odd wave-function does
exactly the opposite. The binding energy of the H+
2 ion is defined as the difference
between its energy and that of a hydrogen atom and a free proton: i.e.,
According to the variational principle, the binding energy is less than or equal to
the minimum binding energy which can be inferred from Fig. 24. This minimum
occurs when X ≃ 2.5 and F+ ≃ −1.13. Thus, our estimates for the separation
between the two protons, and the binding energy, for the H+ 2 ion are R = 2.5 a0 =
−10
1.33 × 10 m and Ebind = 0.13 E0 = −1.77 eV, respectively. The experimentally
determined values are R = 1.06 × 10−10 m, and Ebind = −2.8 eV, respectively.
Clearly, our estimates are not particularly accurate. However, our calculation
does establish, beyond any doubt, the existence of a bound state of the H+ 2 ion,
which is all that we set out to achieve.
248
14.4 The Hydrogen Molecule Ion 14 VARIATIONAL METHODS
Figure 24: The functions F+(X) (solid curve) and F−(X) (dashed curve).
249
15 SCATTERING THEORY
15 Scattering Theory
15.1 Introduction
Historically, data regarding quantum phenomena has been obtained from two
main sources. Firstly, from the study of spectroscopic lines, and, secondly, from
scattering experiments. We have already developed theories which account for
some aspects of the spectrum of hydrogen, and hydrogen-like, atoms. Let us now
examine the quantum theory of scattering.
15.2 Fundamentals
H = H0 + V(r), (15.1)
where
p2 h̄2 2
H0 = ≡− ∇ (15.2)
2m 2m
is the Hamiltonian of a free particle of mass m, and V(r) the scattering potential.
This potential is assumed to only be non-zero in a fairly localized region close to
the origin. Let √
ψ0 (r) = n e i k·r (15.3)
represent an incident beam of particles, of number density n, and velocity v =
h̄ k/m. Of course,
H0 ψ0 = E ψ0 , (15.4)
where E = h̄2 k2 /2 m is the particle energy. Schrödinger’s equation for the scat-
tering problem is
(H0 + V) ψ = E ψ, (15.5)
subject to the boundary condition ψ → ψ0 as V → 0.
250
15.2 Fundamentals 15 SCATTERING THEORY
Let us calculate the value of the wave-function ψ(r) well outside the scattering
region. Now, if r ≫ r ′ then
|r − r ′ | ≃ r − ^r · r ′ (15.10)
to first-order in r ′ /r, where ^r/r is a unit vector which points from the scattering
region to the observation point. It is helpful to define k ′ = k ^r. This is the
wave-vector for particles with the same energy as the incoming particles (i.e.,
k ′ = k) which propagate from the scattering region to the observation point.
Equation (15.9) reduces to
√ eikr
251
15.3 The Born Approximation 15 SCATTERING THEORY
j = n v, (15.14)
where v = h̄ k/m is the velocity of the incident particles. Likewise, the particle
flux associated with the scattered wave-function ψ − ψ0 is
|f(k, k ′ )|2 ′
j′ = n v, (15.15)
r2
where v ′ = h̄ k ′ /m is the velocity of the scattered particles. Now,
dσ r2 dΩ |j ′ |
dΩ = , (15.16)
dΩ |j|
which yields
dσ
= |f(k, k ′ )|2 . (15.17)
dΩ
Thus, |f(k, k ′ )|2 gives the differential cross-section for particles with incident ve-
locity v = h̄ k/m to be scattered such that their final velocities are directed into
a range of solid angles dΩ about v ′ = h̄ k ′ /m. Note that the scattering conserves
energy, so that |v ′ | = |v| and |k ′ | = |k|.
252
15.3 The Born Approximation 15 SCATTERING THEORY
The Born approximation is valid provided that ψ(r) is not too different from
ψ0 (r) in the scattering region. It follows, from Eq. (15.9), that the condition for
ψ(r) ≃ ψ0 (r) in the vicinity of r = 0 is
Z ′
m
exp( i k r ) ′
3 ′
V(r ) d r ≪ 1. (15.29)
2π h̄2 r′
Consider the special case of the Yukawa potential. At low energies, (i.e., k ≪ µ)
we can replace exp( i k r ′ ) by unity, giving
2 m |V0 |
≪1 (15.30)
h̄2 µ2
as the condition for the validity of the Born approximation. The condition for the
Yukawa potential to develop a bound state is
2 m |V0 |
≥ 2.7, (15.31)
h̄2 µ2
254
15.4 Partial Waves 15 SCATTERING THEORY
Outside the range of the scattering potential, both ψ0 (r) and ψ(r) satisfy the
free space Schrödinger equation
(∇2 + k2 ) ψ = 0. (15.36)
255
15.4 Partial Waves 15 SCATTERING THEORY
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
X
ψ(r, θ) = Rl (r) Pl (cos θ), (15.37)
l
since the Legendre functions Pl (cos θ) form a complete set in θ-space. The Legen-
dre functions are related to the spherical harmonics, introduced in Sect. 8, via
v
4π
u
u
Pl (cos θ) = t
Yl,0 (θ, ϕ). (15.38)
2l + 1
Equations (15.36) and (15.37) can be combined to give
2 d2 Rl dRl
r 2
+ 2 r + [k2 r2 − l (l + 1)]Rl = 0. (15.39)
dr dr
The two independent solutions to this equation are the spherical Bessel functions,
jl (k r) and yl (k r), introduced in Sect. 9.3. Recall that
!l
1 d sin z
!
l
jl (z) = z − , (15.40)
z dz z
1 d l cos z
! !
l
yl (z) = −z − . (15.41)
z dz z
Note that the jl (z) are well-behaved in the limit z → 0 , whereas the yl (z) become
singular. The asymptotic behaviour of these functions in the limit z → ∞ is
sin(z − l π/2)
jl (z) → , (15.42)
z
cos(z − l π/2)
yl (z) → − . (15.43)
z
We can write
X
exp( i k r cos θ) = al jl (k r) Pl (cos θ), (15.44)
l
where the al are constants. Note there are no yl (k r) functions in this expres-
sion, because they are not well-behaved as r → 0. The Legendre functions are
orthonormal, Z1
δnm
Pn (µ) Pm (µ) dµ = , (15.45)
−1 n + 1/2
256
15.4 Partial Waves 15 SCATTERING THEORY
It is well-known that
Z1
(−i)l
jl (y) = exp( i y µ) Pl (µ) dµ, (15.47)
2 −1
al = i l (2 l + 1), (15.48)
giving
√ √ X l
ψ0 (r) = n exp( i k r cos θ) = n i (2 l + 1) jl (k r) Pl (cos θ). (15.49)
l
The above expression tells us how to decompose the incident plane-wave into a
series of spherical waves. These waves are usually termed “partial waves”.
The most general expression for the total wave-function outside the scattering
region is
√ X
ψ(r) = n [Al jl (k r) + Bl yl (k r)] Pl (cos θ), (15.50)
l
where the Al and Bl are constants. Note that the yl (k r) functions are allowed
to appear in this expansion, because its region of validity does not include the
origin. In the large-r limit, the total wave-function reduces to
√ X
sin(k r − l π/2) cos(k r − l π/2)
ψ(r) ≃ n A
l − Bl Pl (cos θ), (15.51)
l
k r k r
where use has been made of Eqs. (15.42) and (15.43). The above expression can
also be written
√ X sin(k r − l π/2 + δl )
ψ(r) ≃ n Cl Pl (cos θ), (15.52)
l
k r
257
15.4 Partial Waves 15 SCATTERING THEORY
where the sine and cosine functions have been combined to give a sine function
which is phase-shifted by δl . Note that Al = Cl cos δl and Bl = −Cl sin δl .
which contains both incoming and outgoing spherical waves. What is the source
of the incoming waves? Obviously, they must be part of the large-r asymptotic ex-
pansion of the incident wave-function. In fact, it is easily seen from Eqs. (15.42)
and (15.49) that
√ X l e i (k r−l π/2) − e−i (k r−l π/2)
Clearly, determining the scattering amplitude f(θ) via a decomposition into par-
tial waves (i.e., spherical waves) is equivalent to determining the phase-shifts
δl .
258
15.5 Determination of Phase-Shifts 15 SCATTERING THEORY
is thus given by
Z
σtotal = |f(θ)|2 dΩ
I Z1 XX
1
= 2 dφ dµ (2 l + 1) (2 l ′ + 1) exp[ i (δl − δl ′ )]
k −1 l l′
× sin δl sin δl ′ Pl (µ) Pl ′ (µ), (15.58)
Let us now consider how the phase-shifts δl in Eq. (15.57) can be evaluated.
Consider a spherically symmetric potential V(r) which vanishes for r > a, where
a is termed the range of the potential. In the region r > a, the wave-function ψ(r)
satisfies the free-space Schrödinger equation (15.36). The most general solution
which is consistent with no incoming spherical-waves is
∞
√ X
ψ(r) = n il (2 l + 1) Rl (r) Pl (cos θ), (15.60)
l=0
where
Rl (r) = exp( i δl ) [cos δl jl (k r) − sin δl yl (k r)] . (15.61)
Note that yl (k r) functions are allowed to appear in the above expression, because
its region of validity does not include the origin (where V 6= 0). The logarithmic
derivative of the lth radial wave-function, Rl (r), just outside the range of the
potential is given by
cos δl jl′ (k a) − sin δl yl′ (k a)
βl+ = ka , (15.62)
cos δl jl (k a) − sin δl yl (k a)
259
15.5 Determination of Phase-Shifts 15 SCATTERING THEORY
where jl′ (x) denotes djl (x)/dx, etc. The above equation can be inverted to give
k a jl′ (k a) − βl+ jl (k a)
tan δl = . (15.63)
k a yl′ (k a) − βl+ yl (k a)
Thus, the problem of determining the phase-shift δl is equivalent to that of ob-
taining βl+ .
The most general solution to Schrödinger’s equation inside the range of the
potential (r < a) which does not depend on the azimuthal angle φ is
∞
√ X
ψ(r) = n i l (2 l + 1) Rl (r) Pl (cos θ), (15.64)
l=0
where
ul (r)
Rl (r) = , (15.65)
r
and
d2 ul 2 l (l + 1) 2 m
+ k − − 2 V ul = 0. (15.66)
dr2 r2 h̄
The boundary condition
ul (0) = 0 (15.67)
ensures that the radial wave-function is well-behaved at the origin. We can
launch a well-behaved solution of the above equation from r = 0, integrate out
to r = a, and form the logarithmic derivative
1 d(ul /r)
βl− = . (15.68)
(ul /r) dr r=a
Since ψ(r) and its first derivatives are necessarily continuous for physically ac-
ceptible wave-functions, it follows that
260
15.6 Hard Sphere Scattering 15 SCATTERING THEORY
Let us test out this scheme using a particularly simple example. Consider scat-
tering by a hard sphere, for which the potential is infinite for r < a, and zero for
r > a. It follows that ψ(r) is zero in the region r < a, which implies that ul = 0
for all l. Thus,
βl− = βl+ = ∞, (15.70)
for all l. Equation (15.63) thus gives
jl (k a)
tan δl = . (15.71)
yl (k a)
δ0 = −k a. (15.73)
261
15.6 Hard Sphere Scattering 15 SCATTERING THEORY
Let us examine the low and high energy asymptotic limits of tan δl . Low energy
implies that k a ≪ 1. In this regime, the spherical Bessel functions reduce to:
(k r)l
jl (k r) ≃ , (15.76)
(2 l + 1)!!
(2 l − 1)!!
yl (k r) ≃ − , (15.77)
(k r)l+1
where n!! = n (n − 2) (n − 4) · · · 1. It follows that
−(k a)2 l+1
tan δl = . (15.78)
(2 l + 1) [(2 l − 1)!!] 2
It is clear that we can neglect δl , with l > 0, with respect to δ0 . In other words,
at low energy only S-wave scattering (i.e., spherically symmetric scattering) is
important. It follows from Eqs. (15.17), (15.57), and (15.73) that
dσ sin2 k a
= 2
≃ a2 (15.79)
dΩ k
for k a ≪ 1. Note that the total cross-section
Z
dσ
σtotal = dΩ = 4π a2 (15.80)
dΩ
is four times the geometric cross-section π a2 (i.e., the cross-section for classical
particles bouncing off a hard sphere of radius a). However, low energy scatter-
ing implies relatively long wave-lengths, so we would not expect to obtain the
classical result in this limit.
Consider the high energy limit k a ≫ 1. At high energies, all partial waves
up to lmax = k a contribute significantly to the scattering cross-section. It follows
from Eq. (15.59) that
l
max
4π X
σtotal ≃ 2 (2 l + 1) sin2 δl . (15.81)
k l=0
262
15.7 Low Energy Scattering 15 SCATTERING THEORY
This is twice the classical result, which is somewhat surprizing, since we might ex-
pect to obtain the classical result in the short wave-length limit. For hard sphere
scattering, incident waves with impact parameters less than a must be deflected.
However, in order to produce a “shadow” behind the sphere, there must also be
some scattering in the forward direction in order to produce destructive inter-
ference with the incident plane-wave. In fact, the interference is not completely
destructive, and the shadow has a bright spot (the so-called “Poisson spot”) in
the forward direction. The effective cross-section associated with this bright spot
is π a2 which, when combined with the cross-section for classical reflection, π a2 ,
gives the actual cross-section of 2π a2 .
In general, at low energies (i.e., when 1/k is much larger than the range of the
potential) partial waves with l > 0 make a negligible contribution to the scatter-
ing cross-section. It follows that, at these energies, with a finite range potential,
only S-wave scattering is important.
Note that Eq. (15.85) only applies when E > V0 . For E < V0 , we have
sinh(κ r)
R0 (r) = B , (15.87)
r
where
h̄2 κ2
V0 − E = . (15.88)
2m
Matching R0 (r), and its radial derivative, at r = a yields
k
tan(k a + δ0 ) = ′
tan(k ′ a) (15.89)
k
for E > V0 , and
k
tan(k a + δ0 ) = tanh(κ a) (15.90)
κ
for E < V0 .
Consider an attractive potential, for which E > V0 . Suppose that |V0 | ≫ E (i.e.,
the depth of the potential well is much larger than the energy of the incident
particles), so that k ′ ≫ k. We can see from Eq. (15.89) that, unless tan(k ′ a)
becomes extremely large, the right-hand side is much less that unity, so replacing
the tangent of a small quantity with the quantity itself, we obtain
k
k a + δ0 ≃ ′
tan(k ′ a). (15.91)
k
This yields
tan(k ′ a)
δ0 ≃ k a − 1 . (15.92)
k′ a
According to Eq. (15.81), the scattering cross-section is given by
2
′
4π 2 2 tan(k a)
σtotal ≃ 2 sin δ0 = 4π a − 1 . (15.93)
k k′ a
Now v
2 m |V0 | a2
u
k′ a =
u
t 2 2
k a + , (15.94)
h̄2
264
15.8 Resonances 15 SCATTERING THEORY
Note that there are values of k ′ a (e.g., k ′ a ≃ 4.49) at which δ0 → π, and the
scattering cross-section (15.93) vanishes, despite the very strong attraction of the
potential. In reality, the cross-section is not exactly zero, because of contributions
from l > 0 partial waves. But, at low incident energies, these contributions are
small. It follows that there are certain values of V0 and k which give rise to almost
perfect transmission of the incident wave. This is called the Ramsauer-Townsend
effect, and has been observed experimentally.
15.8 Resonances
The origin of this rather strange behaviour is quite simple. The condition
v
u 2 m |V0 | a2 π
u
t
2 = (15.97)
h̄ 2
265
15.8 Resonances 15 SCATTERING THEORY
We have seen that there is a resonant effect when the phase-shift of the S-wave
takes the value π/2. There is nothing special about the l = 0 partial wave, so it
is reasonable to assume that there is a similar resonance when the phase-shift of
the lth partial wave is π/2. Suppose that δl attains the value π/2 at the incident
energy E0 , so that
π
δl (E0 ) = . (15.98)
2
Let us expand cot δl in the vicinity of the resonant energy:
d cot δl
!
cot δl (E) = cot δl (E0 ) + (E − E0 ) + · · · (15.99)
dE E=E0
1 dδl
= − 2 (E − E0 ) + · · · . (15.100)
sin δl dE E=E0
Defining
dδl (E) 2
= , (15.101)
dE E=E0 Γ
we obtain
2
cot δl (E) = − (E − E0 ) + · · · . (15.102)
Γ
Recall, from Eq. (15.59), that the contribution of the lth partial wave to the
scattering cross-section is
4π 2 4π 1
σl = (2 l + 1) sin δ l = (2 l + 1) . (15.103)
k2 k2 1 + cot2 δl
266
15.8 Resonances 15 SCATTERING THEORY
Thus,
4π Γ 2 /4
σl ≃ (2 l + 1) . (15.104)
k2 (E − E0 )2 + Γ 2 /4
This is the famous Breit-Wigner formula. The variation of the partial cross-section
σl with the incident energy has the form of a classical resonance curve. The quan-
tity Γ is the width of the resonance (in energy). We can interpret the Breit-Wigner
formula as describing the absorption of an incident particle to form a metastable
state, of energy E0 , and lifetime τ = h̄/Γ .
267