Notes 311219

Download as pdf or txt
Download as pdf or txt
You are on page 1of 82

Part IB Quantum Mechanics, Michaelmas 2019

Tuesday and Thursday, 11.05 a.m. to 11.55 a.m., Mill Lane lecture room 3
Prof. Adrian Kent ([email protected])

I am very grateful to previous lecturers of this course, and in particular to


Jonathan Evans and Eugene Lim, who very generously shared their lecture notes
and thoughts on the material. These notes have been influenced by their presenta-
tions of many of the course topics, and sometimes draw directly on their discussions.
I am also grateful to past students of the course, whose feedback and corrections
have been very helpful. More feedback – whether noting typos or other errors or
more general comments – would be very welcome!
The completed lecture notes are intended to be a reasonably complete summary
of the course. However, material not covered in the printed notes will be added
during lectures, for instance in answer to questions (which are encouraged!) or
whenever there is time for further discussion that could be helpful. Non-examinable
sections of the lecture notes are marked by asterisks at the start and the end. Some
footnotes also contain references to results proved in textbooks or other references
but not in the notes: these proofs too are, obviously, non-examinable.
Further course material will be added from time to time on the course web page,
which is linked from www.qi.damtp.cam.ac.uk (follow the link to Undergraduate and
Masters Lecture Courses on the left menu).
If you wish to make additional notes during lectures you will probably find it
simplest to make them on separate sheets of paper, with footnote numbers to refer
them to the appropriate place in the printed notes.

Highlights of this course:

• Historical development of quantum mechanics

• The one-dimensional Schrödinger equation; solution for particles in various


potentials; probabilistic interpretation; beam scattering and tunneling.
• The basic formalism of quantum mechanics – states, operators, observables,
measurement, the uncertainty principle: a new way of treating familiar dy-
namical quantities (position, momentum, energy, angular momentum).
• Quantum mechanics in three dimensions: the 3D Schrödinger equation, an-
gular momentum, the hydrogen atom and other solutions of the 3D SE.

Version dated 31.12.19. Any updated versions will be placed on the course web
page linked from www.qi.damtp.cam.ac.uk.

1
Recommended books

(*) R. Feynman, R. Leighton and M. Sands, Feynman lectures on physics, volume


3, chapters 1-3 (Addison-Wesley, 1989).
A beautifully written and profoundly thoughtful introduction to some of the
basic ideas of quantum theory. Feynman was one of the twentieth century’s most
creative physicists. As these chapters illustrate, he also thought very deeply and
carefully about fundamental questions in physics and about the scientific process
itself. I really recommend these chapters very strongly as background reading for
the course.

(*) A. Rae and J. Napolitano, Quantum Mechanics, chapters 1-5 (IOP Publish-
ing, 2002).
A very good textbook which covers a range of elementary and more advanced
topics in quantum theory, including a short discussion of the conceptual problems
of quantum mechanics. The first five chapters form a good course text for IB QM.

(*) S. Gasiorowicz, Quantum Physics (Wiley 2003).


A very nicely and clearly written book, with good illustrations, which covers
most of the course material well.

P. Landshoff, A. Metherell and G. Rees, Essential Quantum Physics (Cambridge


University Press, 2001).
Another nicely and concisely written textbook, which also covers most of the
course material for IB QM well.

P. A. M. Dirac, The Principles of Quantum Mechanics (Oxford University Press,


1967; reprinted 2003).
A more advanced treatment. Despite its age it is still a valuable exposition of
the insights and perspectives of one of the pioneers of quantum theory. You may
want to consider looking at this if you intend to go beyond this course and pursue
Part II and Part III courses in quantum physics.

S. Brandt and H.D. Dahmen, The Picture Book of Quantum Mechanics (4th
edition; Springer, 2012). An excellent book with accompanying plots and simu-
lations. These give visual explanations that nicely illuminate the mathematics of
solutions of the Schrödinger equation, tunnelling, reflection from barriers, atomic
electron states, and other quantum phenomena.

(*) Particularly recommended.

2
Contents
1 Quantum Mechanics, science and technology 5
1.1 Quantum Mechanics and fundamental science . . . . . . . . . . . . . 5
1.2 Quantum Mechanics and technology . . . . . . . . . . . . . . . . . . 6

2 Historical development of quantum mechanics 7


2.1 Planck’s quantum hypothesis (1900) . . . . . . . . . . . . . . . . . . 7
2.2 Einstein’s explanation of the photoelectric effect (1905) . . . . . . . 8
2.3 Diffraction of single photons (1909) . . . . . . . . . . . . . . . . . . . 9
2.4 The Rutherford atom (1911) . . . . . . . . . . . . . . . . . . . . . . 10
2.5 The Bohr atom (1913) . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.6 Compton scattering (1923) . . . . . . . . . . . . . . . . . . . . . . . 14
2.7 Wave and particle models of electromagnetic radiation . . . . . . . . 15
2.8 De Broglie waves (1924) . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.9 Matter wave diffraction (1923-7) . . . . . . . . . . . . . . . . . . . . 16
2.10 Discussion of the double slit experiment . . . . . . . . . . . . . . . . 16
2.10.1 What can we conclude from double slit experiments? . . . . 19
2.11 *Ongoing tests of quantum theory . . . . . . . . . . . . . . . . . . . 21
2.12 Closing comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3 The one-dimensional Schrödinger equation 23


3.1 The 1D Schrödinger equation for a free particle . . . . . . . . . . . . 23
3.2 The momentum operator . . . . . . . . . . . . . . . . . . . . . . . . 24
3.3 The 1D Schrödinger equation for a particle in a potential . . . . . . 25
3.4 The wavefunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.4.1 What is the wavefunction? . . . . . . . . . . . . . . . . . . . 26
3.4.2 What the wavefunction definitely isn’t . . . . . . . . . . . . . 27
3.5 The superposition principle . . . . . . . . . . . . . . . . . . . . . . . 27
3.6 Probabilistic interpretation of the wavefunction: the Born rule . . . 28
3.6.1 Probability density and probability current . . . . . . . . . . 29

4 Solutions of the 1D Schrödinger equation 31


4.1 Example I: Stationary states . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Completeness of the energy eigenfunctions . . . . . . . . . . . . . . . 32
4.3 Example II: Gaussian wavepackets . . . . . . . . . . . . . . . . . . . 34
4.4 Example III: Particle in an infinite square potential well . . . . . . . 35
4.4.1 Handling discontinuities . . . . . . . . . . . . . . . . . . . . . 36
4.5 Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.6 Example IV: Particle in a finite square well . . . . . . . . . . . . . . 38
4.7 Example V: The quantum harmonic oscillator . . . . . . . . . . . . . 41

5 Tunnelling and Scattering 45


5.1 Scattering states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.2 Interpretation of plane wave scattering solutions . . . . . . . . . . . 46
5.3 Example I: The potential step . . . . . . . . . . . . . . . . . . . . . . 47
5.4 Example II: The square potential barrier . . . . . . . . . . . . . . . . 49
5.4.1 *Important examples . . . . . . . . . . . . . . . . . . . . . . . 51

6 The basic formalism of quantum mechanics 52


6.1 Spaces of functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.2 The inner product . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.2.1 Properties of the inner product . . . . . . . . . . . . . . . . . 52
6.3 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3
6.3.1 Some examples of operators . . . . . . . . . . . . . . . . . . . 53
6.4 Hermitian operators . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.4.1 Classical states and dynamical variables . . . . . . . . . . . . 53
6.4.2 Quantum states and observables . . . . . . . . . . . . . . . . 54
6.5 Some theorems about hermitian operators . . . . . . . . . . . . . . . 55
6.6 Quantum measurement postulates . . . . . . . . . . . . . . . . . . . 57
6.7 Expectation values . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
6.8 Commutation relations . . . . . . . . . . . . . . . . . . . . . . . . . . 59
6.8.1 The canonical commutation relations . . . . . . . . . . . . . . 60
6.9 Heisenberg’s Uncertainty Principle . . . . . . . . . . . . . . . . . . . 61
6.9.1 What does the uncertainty principle tell us? . . . . . . . . . . 63
6.10 Ehrenfest’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.10.1 Applications of Ehrenfest’s theorem . . . . . . . . . . . . . . 64
6.11 *The harmonic oscillator revisited . . . . . . . . . . . . . . . . . . . 66

7 The 3D Schrödinger equation 68


7.1 Quantum mechanics in three dimensions . . . . . . . . . . . . . . . . 68
7.2 Spherically symmetric potentials . . . . . . . . . . . . . . . . . . . . 69
7.3 Examples of spherically symmetric potentials . . . . . . . . . . . . . 71
7.3.1 The spherical harmonic oscillator . . . . . . . . . . . . . . . . 71
7.3.2 The spherical square well . . . . . . . . . . . . . . . . . . . . 71
7.4 Spherically symmetric bound states of the hydrogen atom . . . . . . 72
7.5 Canonical Commutation Relations in 3D . . . . . . . . . . . . . . . . 74
7.6 Orbital Angular Momentum . . . . . . . . . . . . . . . . . . . . . . . 74
7.6.1 Angular momentum commutation relations . . . . . . . . . . 75
7.6.2 Angular momentum operators in spherical polar coordinates 76
7.7 Solving the 3D Schrödinger equation for a spherically symmetric po-
tential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7.7.1 Degeneracies . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
7.8 The Hydrogen atom . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.8.1 Energy level degeneracies . . . . . . . . . . . . . . . . . . . . 81
7.9 Towards the periodic table . . . . . . . . . . . . . . . . . . . . . . . . 81

4
1 Quantum Mechanics, science and technology
1.1 Quantum Mechanics and fundamental science

Quantum mechanics is the non-relativistic version of relativistic quantum field the-


ory. Collectively, we refer to these theories as “quantum theory”. Most physicists
would agree quantum theory is the most remarkable, interesting and surprising
physical theory we have discovered.1 Although ultimately only a non-relativistic
part of a larger theory, quantum mechanics already teaches us that our universe
follows laws that involve beautiful and intricate mathematics, and whose form we
could not possibly have imagined had they not been illuminated by experiment.
Among other things, quantum mechanics explains the essential features of the
following:

• the structure of atoms and molecules and their chemical interactions; i.e.
chemistry, and biochemistry and so, in principle, biology. We will begin to
discuss this towards the end of the course, when we consider quantum mechan-
ical descriptions of the hydrogen atom and, more qualitatively, other atoms.
This description is made more precise and taken further in Part II Principles
of Quantum Mechanics.
• the structure and properties of solids (and so, in principle, much of classical
mechanics). Conductivity (some basic theory of which is introduced in Part
II Applications of Quantum Mechanics) and superconductivity.
• the thermodynamics of light and other electromagnetic radiation and also how
electromagnetic radiation interacts with matter. To describe this properly
requires relativistic quantum field theory, which isn’t covered until Part III.
• the physics of subatomic particles, radioactivity, nuclear fission and fusion.
Again, we need relativistic quantum field theory to describe these phenomena
in full detail. But, as we’ll see later, even elementary quantum mechanics
gives us some useful insights into the physics of nuclear fission and fusion. For
example, we can understand the random nature of these processes, and the
way fusion and fission rates depend on relevant potentials, as a consequence
of quantum tunnelling.

Modern cosmological models are also based on quantum theory. Since we do not
have a quantum theory of gravity, and do not know for sure whether there is one,
these cosmological models are at best incomplete. Nonetheless they give a good
qualitative fit to observational data. Many physicists hope that this project can be
completed, so that we can describe the entire universe and its evolution by quantum
theory.2 These topics are covered in detail in the Part II and Part III Cosmology
courses.
1 Einstein’s general theory of relativity is the only other contender. It is an extraordinarily

beautiful theory that transformed our understanding of space and time and their relationship to
matter, and that gave us the tools to understand the cosmos. The two theories are fundamentally
incompatible, and it is uncertain precisely which parts of which theory will survive in a future uni-
fication. Still, I find quantum theory more intriguing, because of the variety of deep mathematical
concepts it combines and because it is so different from and so much stranger than the theories
that preceded it.
2 It is also true that many thoughtful physicists suspect that the project cannot be completed,

because we will need something other than a quantum theory of gravity, or because quantum
theory will turn out not to hold on large scales. If so, comparing quantum cosmological models to
observation should eventually give us insight into new physics.

5
1.2 Quantum Mechanics and technology

Many of the revolutionary technological developments of the last hundred years rely
on quantum mechanics:

• semiconductor physics – transistors, diodes, integrated circuits, and hence the


computing and IT industries.
• laser physics
• nuclear power and the as yet unrealised dream of controllable fusion power
• tunnelling electron microscopy and atomic manipulation

• More recent inventions relying on the distinctive properties of quantum infor-


mation. These include quantum cryptography. Its best known application is
quantum key distribution, which in principle allows perfectly secure commu-
nication, and is now practical with large data transmission rates and over long
distances, including between satellite and ground stations. There are many
other applications, including quantum authentication, quantum position ver-
ification, quantum bit string commitment, quantum multi-party computation
and quantum digital signatures. Quantum information allows perfect security
for some of these tasks, and security advantages for others.
Another major development has been the invention of quantum algorithms
and various types of quantum computers. We now know that quantum com-
puters are substantially more efficient than classical computers for some im-
portant applications, including factorisation (at least compared to the best
known classical algorithms) and the simulation of quantum systems. Small
scale quantum computers have been built, and many research groups are com-
peting to build a quantum computer large enough to exploit the theoretical
advantages. In September 2019, Google claimed the first demonstration of
so-called “quantum supremacy”, a calculation on a quantum computer that
is infeasible on classical computers.3
Other fascinating applications include quantum teleportation – which in prin-
ciple gives a way of effectively deleting a physical system at A and recreating
it at B without sending it along a path from A to B – and other types of
quantum communication.
All of these topics are covered in the Part II Quantum Information and Com-
putation course. Notes for that course, and for Part III courses in this area,
along with some information about research work on these topics in DAMTP,
can be found on http://www.qi.damtp.cam.ac.uk.

3 There are nuances of definition here, and the community is still assessing the claim.

6
Figure 1: Google’s Bristlecone quantum processor. Image on left, map of qubit
connections on right.

2 Historical development of quantum mechanics

2.1 Planck’s quantum hypothesis (1900)


One of the great puzzles in late 19th century physics was the inability of clas-
sical thermodynamics and electromagnetism to predict the correct spectrum — or
indeed any sensible spectrum — for the frequency distribution of radiation from
an idealized black body. Classical thermodynamics predicted an emission spectrum
which suggested that the flux of emitted radiation tends to infinity as the frequency
tends to infinity, and hence that an infinite amount of energy is emitted per unit
time. In 1900, Max Planck showed that one could predict the experimentally ob-
served spectrum by postulating that, because of some (at that point) unknown
physics, matter can emit or absorb light of frequency ν only in discrete quanta
which have energy

E = hν = ~ω . (2.1)

Here ω = 2πν is the light’s angular frequency (the number


of radians of oscillation per second), and
h
~= ≈ 1.055 × 10−34 Joule sec . (2.2)

The constant h is a new constant of nature (Planck’s constant). For most purposes
it turns out to be notationally more convenient to use ~ rather than h, and we will
generally do so.

7
2.2 Einstein’s explanation of the photoelectric effect (1905)
Experiment shows that light hitting a metal surface in a vacuum can cause
electrons to be ejected with a range of energies. To emit any electrons, the incident
light needs to have angular frequency ω satisfying ω ≥ ωmin , where ωmin is a constant
depending on the particular metal. When ω ≥ ωmin , one finds that the maximum
energy of the emitted electrons, Emax , obeys

Emax = ~ω − ~ωmin ≡ ~ω − φ , (2.3)


where φ is the so-called work function of the metal. The average rate of electron
emission is found to be proportional to the intensity of the incident light, but
individual electrons appear to be emitted at random (and so in particular, measured
over small enough time intervals, the emission rate fluctuates).

Incoming light Emitted electrons

𝑒−
𝑒−

𝑒−
𝑒− 𝑒− 𝑒− 𝑒− 𝑒− 𝑒−

Figure 2: Schematic illustration of the photoelectric effect.

Although light had been understood as an electromagnetic wave, it was hard to


explain these effects in terms of a wave model of light. If we think of an incident
wave transmitting energy to the electrons and knocking them out of the metal, we
would expect the rate of electron emission to be constant (i.e. we would not expect
random fluctuations), and we would also expect that light of any frequency would
eventually transmit enough energy to electrons to cause some of them to be emitted.
To explain the photoelectric effect, Albert Einstein (in 1905) was led to postulate
instead that light is quantised in small packets called photons, and that a photon of
angular frequency ω has energy ~ω. He reasoned that two photons are very unlikely
to hit the same electron in a short enough time interval that their combined effect
would knock the electron out of the metal: the energy an electron acquires from
being hit by a photon is very likely to have dissipated by the time it is hit again.

8
Thus, one can explain the photoelectric effect as the result of single photons hitting
electrons near the metal surface, if one assumes that an electron needs to acquire
kinetic energy ≥ φ to overcome the binding energy of the metal. An electron which
acquires energy ~ω thus carries away energy ≤ ~ω −φ = Emax . One can also explain
the emission rate observations: the average rate of photon arrival is proportional
to the intensity of the light, and the rate of emission of electrons is proportional
to the rate of photon arrival. However, individual photons arrive, and hit electrons
so as to knock them out of the metal, at random – hence the randomly distributed
emissions of electrons.

2.3 Diffraction of single photons (1909)


In Cambridge in 1909, J.J. Thomson suggested to G.I. Taylor (who had asked for
a research project) that he investigate the interference of light waves of very low
intensity. Taylor carried out an experiment in which a light source was successively
filtered so that the energy flux was equivalent to the flux of a source sending no
more than one photon at a time through the apparatus. He then observed the
photographic image built up by diffraction of this feeble light around a needle. The
characteristic diffraction pattern – the same pattern seen for strong light sources –
was still observed.

Figure 3: Photograph of diffraction bands caused by a thin wire in feeble light.


See: Taylor, Proc. Camb. Phil. Soc., 15, 114, 1909. The exposure was 400 hours.
Source: Cavendish Laboratory. Licensed under Creative Commons.

This seems to suggest that single photons propagate through the apparatus and
nonetheless “self-interfere” in such a way that the diffraction pattern is cumulatively
built up. This is indeed how Taylor’s results were interpreted for several decades
after the development of quantum mechanics in 1926. Problems were later noticed

9
with this interpretation of this particular experiment: having the average energy flux
of a single photon does not always imply that a source is emitting single photons.
However, the experiment was subsequently repeated with genuine single photon
sources, still showing the same diffraction pattern. Qualitatively similar patterns
are seen in diffraction experiments for single electrons and other types of matter
(see below).

2.4 The Rutherford atom (1911)


After discovering the electron in 1897, J.J. Thomson proposed a model of the
atom as a sort of “plum pudding” with Z pointlike electrons of charge −e embedded
in a sphere of positive charge +Ze.
Geiger and Marsden’s famous experiment, carried out at Rutherford’s sugges-
tion, tested for large angle scattering of a beam of α-particles directed at gold foil.
One would not expect significant scattering from a loosely distributed low charge
density “plum pudding” atom, and Rutherford thought it unlikely anything interest-
ing would be observed. But, in fact, some α-particles were observed to be scattered
through angles of up to 180◦ . In Rutherford’s famous phrase,
“It was as if you fired a 15-inch shell at tissue paper and it came
back and hit you.”
The scattering suggests a high density positive charge within the atom. Rutherford
thus postulated a new model of the atom, with a heavy nearly pointlike nucleus, of
charge +Ze, surrounded by Z electrons in orbit.
(A short popular account can be found at
http://physicsopenlab.org/2017/04/11/the-rutherford-geiger-marsden-experiment/
.)

2.5 The Bohr atom (1913)


Although the Rutherford atom was more compatible with the Geiger-Marsden
scattering data than was the “plum pudding” model, it had a number of theoretical
defects.
First, according to Maxwell’s electrodynamics, electrons in orbit around a nu-
cleus would radiate, since they are continually undergoing acceleration. This would
cause them to lose energy and fall in towards the nucleus. Stationary electrons
would also fall into the nucleus because of electrostatic attraction. This would
suggest that atoms must be unstable, which they generally are not.
Second, the model fails to explain why atoms have characteristic line spectra cor-
responding to discrete frequencies at which they absorb or emit light. For example,
hydrogen has frequencies given by the Rydberg formula (Rydberg, 1890):
1 1
ωmn = 2πcR0 ( − 2) for m > n , (2.4)
n2 m
where the Rydberg constant

R0 ≈ 1.097 × 107 m−1 . (2.5)

Third, it fails to explain why atoms belong to a finite number of chemical species,
with all members of the same species behaving identically. For instance, if a hydro-
gen atom can have an electron in any type of orbit around its nucleus, one would
expect infinitely many different types of hydrogen atom, corresponding to the in-
finitely many different possible orbits, and one would expect the atoms to have
different physical and chemical properties, depending on the details of the orbit.

10
Figure 4: Schematic illustration of the Geiger-Marsden experiment, from the web
page cited above.

Niels Bohr, in 1913, observed that these problems could be resolved in a way
consistent with Planck’s and Einstein’s earlier postulates, if we suppose that the
electron orbits of hydrogen atoms are quantised so that the orbital angular momen-
tum takes one of a discrete set of values

L = n~ , (2.6)

where n is a positive integer.


Thus, if we take an electron e moving with velocity v in a circular orbit of radius
r about a proton p, F = me a gives us that the Coulomb force

e2 me v 2
= . (2.7)
4π0 r2 r
If we also have
L = me vr = n~ (2.8)

then
n2 ~2 e2
= (2.9)
r3 me 4π0 r2
and hence
~2 4π0
2
r=n ( 2
) = n 2 a0 , (2.10)
me e
11
Figure 5: Spectra for the hydrogen atom. The figure shows three horizontal lines
at small distances from each other. Between the two lower lines, the Lyman series,
with four vertical red bands in compact form, is shown. This has nf = 1 and ni ≥ 2,
and wavelengths in the range 91 − 100 nanometers. The Balmer series is shown to
the right side of this series. This has nf = 2 and ni ≥ 3, and wavelengths in the
range 365 − 656 nanometers. At the right side of this, the Paschen series bands
are shown. This has nf = 3 and ni ≥ 4, and wavelengths in the range 820 − 1875
nanometers. The Rydberg formula is obtained by taking nf = n, ni = m.
Source for figure and caption: https://opentextbc.ca/physicstestbook2
Image licenced under Creative Commons.

where
4π0 ~2
a0 = ≈ 0.53 × 10−10 m (2.11)
m e e2
is the Bohr radius.
We can then obtain the energy of the n-th Bohr orbit from (2.7) and the Coulomb
potential:

1 e2 e2 e2 e4 me 1 E1
En = me v 2 − =− =− = − = ,
2 4π0 r 8π0 r 8π0 n2 a0 32π 2 20 ~2 n2 n2
(2.12)
where
e4 m e
E1 = − ≈ −13.6 eV . (2.13)
32π 2 20 ~2
Thus we have n = 1 with energy E = E1 defining the lowest possible energy
state, or ground state, of the Bohr atom. The higher energy excited states, so called
because they can be created by “exciting” the ground state atom with radiation,
correspond to n > 1. These can decay to the ground state: the ground state has no
lower energy state to decay to, and so is stable. (The Bohr model does not allow a
state with zero orbital angular momentum, which would correspond to n = 0, r = 0
and E = −∞.)

12
The energy emitted for a transition from the m-th to the n-th Bohr orbital is
Emn = Em − En . Using Emn = ~ωmn , where ωmn is the angular frequency of the
emitted photon, we have

1 1
ωmn = 2πR0 c( 2
− 2) , (2.14)
n m
where
me c e2
R0 = ( )2 , (2.15)
2~ 4π0 ~c
which agrees well with the experimentally determined value of the Rydberg con-
stant.

Figure 6: The orbits of Bohr’s planetary model of an atom; five concentric circles
are shown. The radii of the circles increase from innermost to outermost circles.
On the circles, labels E1 , E2 , up to Ei are marked. Source for figure and caption:
https://opentextbc.ca/physicstestbook2
Licenced under Creative Commons.

Bohr’s model of the atom was rather more successful than its predecessors. It
predicts the energy levels of the hydrogen atom and the spectrum of photons emitted
and absorbed. It also accounted for spectroscopic data for ionised helium (He+ ) and
for some emission and absorption spectra for other atoms. (We now understand that
these are the spectra produced by the innermost electrons, which can be modelled
in a way qualitatively similar to the electron in the hydrogen atom.)
However, as Bohr himself stressed, the model offered no explanation of atomic
physics. For example, as Rutherford commented, it’s quite mysterious that an
electron which jumps from the m-th to the n-th orbit seems to know in advance what
frequency to radiate at during the transition. Moreover, the Bohr model is quite

13
wrong about the details of electron orbits, even for the hydrogen atom. Nonetheless,
it was an important stepping stone on the path to quantum mechanics, suggesting
some link between Planck’s constant, atomic spectra and atomic structure, and the
quantisation of angular momentum and other dynamical quantities.

2.6 Compton scattering (1923)


In 1923, Arthur Holly Compton observed the scattering of X-rays by electrons
associated with atoms in a crystal. Because the X-ray energies were much larger
than the electron binding energies, the electrons can effectively be modelled by
free electrons. Indeed, we also directly observe that if an electron beam and an
X-ray beam converge, some electrons and some X-rays are deflected. This is very
difficult to reconcile with a pure wave model of electromagnetic radiation, because
the energy and momentum transfer for individual scatterings does not depend on
the intensity of the X-ray beam.
A simple alternative explanation is that the scattering results from collisions
between a single photon in one beam and a single electron in the other, in which
energy and momentum are transferred between the photon and the electron. (A
relativistic treatment of this scattering process was given in the IA Dynamics and
Relativity course.) This explanation is consistent with the observed scattering data
and with conservation of (relativistic) energy and momentum, provided we assume
that a photon of angular frequency ω has a definite momentum

p = ~k , (2.16)
where k is the wave vector of the corresponding electromagnetic wave, so that

|p| = = ~|k| . (2.17)
c

14
2.7 Wave and particle models of electromagnetic radiation
We thus see the emergence of two different models of light and other electro-
magnetic radiation.
Sometimes (classical electromagnetism, diffraction experiments with a strong
light source, . . .) it is useful to model light in terms of waves:
ei(k.x−ωt) , (2.18)
ω
where k is the wave vector, ω the angular frequency, c = the speed of light in a
|k|
2πc 2π
vacuum, and the wavelength λ = ω = |k| .
Sometimes (photoelectric effect, spectroscopy in the Bohr model of the hydrogen
atom, Compton scattering, . . .) it is useful to adopt a particle model, in which light
is made up of photons with energy and momentum
E = ~ω , p = ~k . (2.19)
The word “model” is chosen deliberately here. A model can be useful (as the
wave and particle models of light are, in the appropriate contexts) without be-
ing completely correct. Indeed, G.I. Taylor’s 1909 demonstration of single photon
diffraction already gave an example of a single experiment for which neither the
wave nor the particle model of light appeared to be adequate. A simple particle
model would not predict the observed diffraction pattern, while a simple wave model
cannot explain the observation of single photons recorded on the photographic film.4

2.8 De Broglie waves (1924)


Louis de Broglie reexamined and extended Einstein’s photon hypothesis. If,
he argued, Einstein was right that light waves can be considered as composed of
particles – photons – might it not equally be the case that objects like electrons,
which were thought of as particles, could exhibit wave-like behaviour?
As he pointed out in his 1924 PhD thesis, this would make the Bohr angular
momentum quantization condition
L = pr = n~ (2.20)
at least somewhat less mysterious. If we suppose that an electron of momentum p
can (somehow) be thought of as a wave with de Broglie wavelength
2π~
λ= , (2.21)
p
then the electron in the n-th Bohr orbit would be in a standing wave pattern with
n wavelengths:

2πn~
2πr = = nλ . (2.22)
p
De Broglie hypothesised that the frequency and wavevectors of the relevant wave
were related to the energy and momentum of the particle as for photons: that is, we
have E = ~ω, p = ~k even for particles with nonzero mass. (In fact, he speculated –
incorrectly, according to our current understanding – that photons might also have
a small nonzero mass.)
Einstein wrote, on learning of de Broglie’s hypothesis, that
4 See earlier comment: to justify this fully we need to consider versions of the experiment with

a guaranteed single photon source.

15
De Broglie’s model of the atom. Electrons occupy orbits
corresponding to a integer multiple of their de Broglie wavelength.

Figure 7: Schematic illustration of the de Broglie model of electron orbits in an


atom.

“I believe it is the first feeble ray of light on this worst of our physics
enigmas.”
It was.

2.9 Matter wave diffraction (1923-7)


We know that the wave model of light predicts, correctly, that light should form
interference and diffraction patterns. De Broglie’s hypothesis suggests the same
should be true of electrons and other massive particles. This was first confirmed in
experiments carried out (from 1923-7) by Davisson and Germer and (independently
in 1927) by G.P. Thomson, who observed diffractive scattering of electrons from
metallic crystals, with diffraction patterns consistent with the de Broglie wavelength
λ = 2π~
p .
G.P. Thomson was the son of J.J. Thomson, who in 1897 discovered the electron,
in experiments in which it behaves as (and so was then understood to be) a particle.
It is a pleasing historical quirk that G.P. Thomson was the co-discoverer of the
wave-like behaviour of electrons in diffraction experiments.
Many diffraction experiments with electrons, neutrons and other particles have
since been carried out, all confirming de Broglie’s prediction.

2.10 Discussion of the double slit experiment


(Cf Feynman volume III chapters I-III)
A nice version of the double slit experiment with electrons was carried out by
Akira Tonomura. It is described at
http://www.hitachi.com/rd/portal/highlight/quantum/doubleslit/index.html
Like the other diffraction experiments mentioned above, this shows that electrons
and other massive particles can produce interference patterns in the same sort of
way as light and water waves. At the same time, it gives very vivid evidence of
electrons being detected as individual particles. We stress again that both the wave
description and the particle description are just models that are sometimes useful
but, as this experiment again illustrates, not fundamentally correct. To analyse this

16
Figure 8: Single electron detection events building up an interference pattern in
Tonomura’s experiment. Copyright in these documents published on Hitachi World-
Wide Web Server is owned by Hitachi, Ltd.

conclusion in more detail, let’s follow Feynman in considering an idealised version


of the double slit experiment, in which we assume we have perfect detectors that
can register the passage of an electron without affecting it.
The observed interference pattern in a double slit experiment agrees with that
predicted by a wave model (and disagrees with that predicted by a particle model).
But the electrons arrive individually at the detector, which registers each time one
arrives — as a particle model (but not a wave model) would suggest. The same is
true of other massive particles, and also of photons. Even if we reduce the intensity
of the source so that only one electron on average is between the source and the
screen at any given time, we still see individual electrons detected in a pattern that
cumulatively reproduces the distribution predicted by the wave model.
It’s tempting to think that, when electrons leave the source, they behave like
bullets from a gun – i.e. like particles coming from a well-defined small region.
Certainly if we place a detector near the source this is what the detections suggest
(although if we do this the electrons don’t continue into the rest of the apparatus).
It’s also tempting to think that, since the electrons (etc.) arrive at the detector and
are detected there as particles, with a definite or nearly definite position, they must
have behaved like classical particles throughout, following some definite path from
the source, through one hole or the other, to the detector. But, tempting though
this last intuition may be, it’s hard to reconcile with the observed interference
pattern. If the electrons behaved like bullets throughout, we’d expect something
like a superposition of two Gaussian distributions from the two slits, instead of the
pattern of minima and maxima we observe.

17
Quantum weirdness: the double slit experiment

The observed patterns for one and two slits fit


very well with a wave model of light, and seem to
refute a model in which light is made up of particles.

Figure 9: The double slit experiment.

There seems, in any case, a simple way to investigate more closely. Suppose
that we have ideal detectors, which register that an electron has passed through a
region, but don’t obstruct its path. We can set up a double slit experiment with
one of these ideal detectors adjacent to each slit (for definiteness, let us say they are
on the far side, between the slit and the screen). Now we don’t need to speculate:
we can observe which slit the electron goes through. But when we do this, the
interference pattern changes: we see a superposition of two Gaussians, as predicted
by a particle model, rather than the pattern of maxima and minima predicted by
the wave model and observed in the previous experiment.

18
What is really going on in the double slit
experiment?
Why not just look and see which slit the photon goes through?
What if we include passive
detectors registering which
slit the photon goes through?

Light By introducing new detectors, we changed


source the experiment.

Quantum theory predicts, and experiment


confirms, that this changes the outcome.
Absorbing filter
Screen
with two slits

Figure 10: Observing which slit the electron goes through in the double slit exper-
iment.

2.10.1 What can we conclude from double slit experiments?

1. As we already stressed, the wave model and the particle model are just that —
models. Neither of them is adequate to explain the behaviour of electrons, photons,
or other objects. Each of them can sometimes give a partial explanation of our
observations, but that explanation is not consistent with all the data.
2. In particular, the type of reasoning about electron paths that would apply
within a particle model does not generally apply in Nature. We can’t assume that
the electron follows a definite path through one slit or the other, and we can’t
assume (as we could with a classical particle) that observing which slit it goes
through makes no difference to the interference pattern.
3. Some textbooks summarize the state of affairs described in point 1 by saying
that electrons (photons, etc.) exhibit something called “wave-particle duality”.
This term can mislead, if it is interpreted as a sort of explanation of what is going
on rather than just a catchy name for it.
The fact is that our classical wave and particle models are fundamentally inad-
equate descriptions. It isn’t correct to say that an electron (or photon, etc.) is both
a wave and a particle in the classical senses of either of those words. The electron
is something different again, though it has some features in common with both. To
go further, we need a new physical model: quantum mechanics.
4. We saw that the electron interference pattern builds up over time, yet the
points at which individual electrons hit the screen do not appear to be precisely
predictable: they seem to arrive at random. It might seem natural to speculate
that this apparent randomness might be explained by the fact that we don’t have
complete data about the experiment. Perhaps, for example, the electrons leave the

19
source in slightly different directions, or perhaps they have some sort of internal
structure that we haven’t yet discussed (and perhaps hasn’t yet been discovered),
and perhaps this determines precisely where they hit the screen.
As we’ll see, according to quantum mechanics this is not the case. Quantum
mechanics, unlike classical mechanics, is an intrinsically probabilistic theory, and it
tells us that there is simply no way to predict precisely where the electron will hit
the screen, even if we have a precise and complete description of its physical state
when it leaves the source.
Now, of course, quantum mechanics might not be the final theory of nature.
It’s possible that some as yet unknown and more complete theory could underlie
quantum mechanics, and it’s logically possible that this theory (if there is one)
could be deterministic. However, there are very strong reasons to doubt that any
theory underlying quantum mechanics can be deterministic. In particular, it can be
proved (given a few very natural assumptions) that any such theory would be incon-
sistent with special relativity. This follows from Bell’s theorem and experimental
tests thereof: it is discussed further in the Part II course “Principles of Quantum
Mechanics” and in Part III courses.)

20
2.11 *Ongoing tests of quantum theory
Although later developments are beyond the scope of this course, it would be
wrong to leave the impression that the historical development of quantum mechan-
ics ended in the first part of last century. Indeed, the basic principles of quantum
mechanics are still being questioned and tested by some theorists and experimen-
talists today. Quantum theory is very well confirmed as a theory of microscopic
physics. However, the case for believing that it applies universally to physics on all
scales is much weaker.
There is a good scientific motivation for testing any scientific theory in new
domains, which is that theories developed to explain phenomena in one domain may
not necessarily apply in other domains. For example, quantum theory itself shows
us the limits of validity of Newtonian mechanics and of classical electrodynamics.
Similarly, Einstein’s general theory of relativity shows us the limits of validity of
Newton’s theory of gravity.
There is also a specific motivation for wanting to test quantum theory for macro-
scopic systems. This is that the problems in making sense of quantum theory as
an explanation of natural phenomena seem to stem from the fact that the classical
physics we see on macroscopic scales appears to emerge from quantum theory in a
way that, despite many attempts, remains hard to pin down. Many theorists be-
lieve it remains fundamentally unexplained. Many others believe it is explained or
explainable, but there is no real consensus among them about the right explanation.
Interestingly, we know there are consistent (non-relativistic) theories that agree
very precisely with quantum mechanical predictions for microscopic (small mass)
particles, but disagree for macroscopic (large mass) ones.
In the past few years, experimental technology has advanced far enough to
demonstrate diffraction of quite complex molecules. (Some descriptions of these
experiments can be found at
https://vcq.quantum.at/; see in particular the work of the Arndt and Aspelmayer
research groups.) Attempts continue to demonstrate interferometry for larger and
larger objects, motivated not only by the technological challenge but also by the
possibility of testing the validity of quantum mechanics in new domains. In Octo-
ber 2019, Fein et al. reported interference experiments for molecules of weight
larger than 25kDa (See https://www.nature.com/articles/s41567-019-0663-9 and
the rather amusing Q and A summary
https://www.quantumnano.at/detailview-news/news/facts-fiction-in-reports-on-high-
mass-interference/ ).
It should be stressed that there is to date (October 2019) no experimental evi-
dence for any deviation from quantum mechanics, which has been confirmed in an
impressive array of experiments investigating many different physical regimes. *

2.12 Closing comments


1. As we’ve seen, the photon hypothesis played a key role in the development of
quantum mechanics. We’ve also seen that photons (which are massless) and massive
particles (such as electrons) produce qualitatively similar interference and diffraction
patterns. However, we won’t have much more to say about photons in this course. It
turns out that we can develop quantum mechanics for the electron and other massive
particles using relatively simple equations. We can build up a good intuition about
how quantum systems behave in experiments and in nature from these equations. A
fully satisfactory quantum treatment of photons or other massless particles requires
a relativistic quantum theory of fields, which requires more sophisticated concepts
(and is much harder to make mathematically rigorous). Quantum electrodynamics,
which is a relativistic quantum field theory incorporating photons, is discussed in

21
Figure 11: Interference of molecules larger than 25KDa, from the Fein et al. paper
cited above.

Part III theoretical physics courses, along with other relevant quantum field theories.
2. Although we’ve already seen the classical particle model is inadequate, we
still need some collective name for electrons, protons, neutrons and so on. Perhaps
physicists should have invented another term, but we still call these “particles”. We
will follow this tradition, so that we might say that the electron is an elementary
particle, talk about quantum mechanics applied to an abstract particle of mass m,
and so forth — always keeping in mind that the classical particle model does not
actually apply.

22
3 The one-dimensional Schrödinger equation
3.1 The 1D Schrödinger equation for a free particle
We are first going to develop quantum mechanics in one space (and one time)
dimension. We can solve the equations for simple physical models more easily in
1D than in 3D and, happily, it turns out that 1D solutions give a good qualitative
feel for a range of interesting 3D physical phenomena.
In 1924, Schrödinger developed de Broglie’s ideas further, into what became a
standard way of framing the laws of quantum mechanics.5
Recall that de Broglie postulated that matter is described by waves, and that
the energy and momentum are related to the angular frequency and wave vector by
E = ~ω and p = ~k, or in one dimension p = ~k. We can express this by associating
to a particle of energy E and momentum p a wave of the form

ψ(x, t) = exp(i(kx − ωt)) . (3.1)


2
p
Now, for a mass m particle, we have E = 21 mv 2 = 2m , and so

i p2
ψ(x, t) = exp(i(kx − ωt)) = exp( (px − t)) . (3.2)
~ 2m
Notice that we have taken ψ(x, t) to be complex. Using complex numbers to
represent waves is familiar in classical electromagnetism, where it allows us to com-
bine the electric and magnetic fields in a single equation. In that context, it’s
mathematically convenient, but the real and imaginary parts each have a simple
physical interpretation. We will see that complex-valued solutions to (generalised)
wave equations play a more essential role in quantum mechanics.
The simplest wave equation to which the de Broglie wave is the general solution
is the time-dependent 1D Schrödinger equation for a free particle:

1 ∂ ∂
(−i~ )2 ψ(x, t) = i~ ψ(x, t) . (3.3)
2m ∂x ∂t
(By a free particle we mean a particle not subject to external forces, i.e. one moving
in a potential V (x) = 0.)
5 There is an equivalent alternative way of describing quantum mechanics, developed by Heisen-

berg, Born and Jordan. But Schrödinger’s formulation is easier to work with and gives a more
intuitive physical picture – so we will follow his approach. Note that “more intuitive” here is a
relative statement. As we will see, many of the predictions of quantum mechanics are counter-
intuitive. Also, some of the intuitions Schrödinger’s picture suggests may be helpful to us in some
contexts but are not fundamentally justifiable.

23
3.2 The momentum operator
We define the momentum operator


p̂ = −i~ . (3.4)
∂x
2
p
So, for the de Broglie wave ψ(x, t) = exp( ~i (px − 2m t)), we have p̂ψ = pψ. In other
words, acting with the momentum operator on the de Broglie wave is equivalent to
multiplying by the wave momentum.6 We can rewrite (3.3) as

1 2 ∂ψ
p̂ ψ = i~ . (3.5)
2m ∂t
This is our first example of a general feature of quantum mechanics. Physically
significant quantities (in this case momentum) are represented by operators. These
operators act on functions that represent physical states (in this case the idealized
state defining the de Broglie wave).
Formally, we define an operator Ô to be a linear map from a space of functions7
to itself, i.e. any map such that

Ô(a1 ψ1 + a2 ψ2 ) = a1 Ôψ1 + a2 Ôψ2 (3.6)


for all complex numbers a1 , a2 and all functions ψ1 , ψ2 in the relevant space.
6 This tells us that the de Broglie wave is an eigenfunction of the momentum operator with

eigenvalue p: we will define these terms more generally later.


7 We will not be too precise about which space of functions we are working with, but will

assume that they are suitably “well behaved”. For example, and depending on the context, we
might want to consider the space of infinitely differentiable functions, C ∞ (R), or the space of
“square integrable” functions, i.e. those satisfying Eqn. (3.11).

24
3.3 The 1D Schrödinger equation for a particle in a potential
We want to consider particles subject to a potential V (x) as well as free particles.
Examples: alpha rays scattering from a nucleus, electrons diffracting from a
crystal, buckyballs going through an interferometer, neutrons or larger massive
particles moving in a gravitational field.

To do this, we replace the kinetic energy term in (3.5) by an operator corre-


sponding to the hamiltonian or total energy:

p2
H= + V (x) , (3.7)
2m
namely

p̂2
Ĥ = + V (x) , (3.8)
2m
where the second operator corresponds to multiplication by V (x).

This gives the general form of the time-dependent 1D Schrödinger equation for
a single particle:
∂ψ
Ĥψ(x, t) = i~ . (3.9)
∂t
Or, more explicitly:

~2 ∂ 2 ψ ∂ψ
− + V (x)ψ = i~ . (3.10)
2m ∂x2 ∂t
Note that there is no way to prove mathematically that Eqns. (3.9, 3.10) are
physically relevant, although we have given some motivation for them in the light
of previous physical models and experimental results. As with any new physical
theory, the only real test is experiment. Since it is not yet obvious what the complex-
valued solutions to (3.9, 3.10) have to do with experimentally observable quantities,
we will first need to give rules for interpreting them physically, and then test these
predictions.

25
3.4 The wavefunction
We call a complex valued function ψ(x, t) that is a solution to (3.9) or (3.10) a
wavefunction. We say the wavefunction ψ(x, t) is normalisable (at time t) if

Z ∞
0< |ψ(x, t)|2 dx < ∞ . (3.11)
−∞
Note that for any complex valued ψ(x, t) the integral is real and non-negative.

We say ψ(x, t) is normalised if

Z ∞
|ψ(x, t)|2 dx = 1 . (3.12)
−∞
R∞
So, given a normalisable ψ(x, t) with −∞ |ψ(x, t)|2 dx = C, so that 0 < C < ∞,
we can define a normalised wavefunction C −1/2 ψ(x, t).

3.4.1 What is the wavefunction?


As we will explain in following lectures, the wavefunction ψ(x, t) of a particle is
a mathematical object that allows us to calculate the probability of detecting the
particle at any given position if we set up a detector there. More generally, it
allows us to calculate the probability of any given outcome for the measurement of
any observable quantity (for example, energy or momentum) associated with the
particle.
Sometimes in the course of your studies you may suspect that lecturers are
temporarily keeping the full truth from you.8 Sometimes you would be right, but
not here. We really don’t have a better and more intuitively comprehensible story
about the wavefunction.9
8 If charitable, you may also suspect there may be good pedagogical reasons for this.
9 At least, not one that is generally agreed.

26
3.4.2 What the wavefunction definitely isn’t
Schrödinger initially hoped to interpret the wave function as describing a dispersed
cloud of physical material that somehow corresponds to a “smeared-out” particle.
This looks a natural interpretation at first sight, but proved untenable and was
abandoned.
One problem with this interpretation is that if a charged particle is really a
dispersed cloud of charge, we would expect to be able to detect bits of the cloud
carrying fractions of the charge of the electron. However, we always find that
charged objects carry a charge that is some integer multiple of the electron charge.
Classical electrodynamics also suggests that a dispersed charge of cloud should in-
teract repulsively with itself via the Coulomb force, and thus tend to be additionally
dispersed, in a way that we do not observe.
Even if these objections could somehow be overcome, another problem remains.
No matter how widely the electron’s wavefunction is spread out in space, when we
look for it by setting up detectors we always find an apparently pointlike particle
in a definite location. If the wavefunction really represented a dispersed cloud, this
cloud would have to suddenly coalesce into a particle at a single point when we
carry out a measurement. This would mean that parts of the cloud would have to
travel extremely fast — often much faster than light speed. This is inconsistent
with special relativity.

3.5 The superposition principle


Exercises 1. The Schrödinger equation

~2 ∂ 2 ∂
− 2
ψ(x, t) + V (x)ψ(x, t) = i~ ψ(x, t)
2m ∂x ∂t
is linear in ψ(x, t): if ψ1 and ψ2 are solutions then a1 ψ1 + a2 ψ2 is a solution too,
for any complex a1 , a2 .

2. If ψ1 and ψ2 are normalisable and a1 ψ1 + a2 ψ2 is nonzero 10 , then it is also


normalisable.
3. Show that it is not generally true that if ψ1 and ψ2 are normalised then
a1 ψ1 + a2 ψ2 is, even if |a1 |2 + |a2 |2 = 1.
The linearity of the Schrödinger equation implies the so-called superposition
principle: there is a physical solution corresponding to any linear combination of
two (or more) physical solutions.
The superposition principle is an essential feature of quantum mechanics, which
does not generally apply in classical physics. It makes no sense in Newtonian
mechanics to add a linear combination of two orbits of a planet around the Sun:
this doesn’t define another possible orbit. But in quantum theory, taking sums of
physical wavefunctions, for example those of an electron orbiting the nucleus of a
hydrogen atom, gives us another wavefunction that has a direct physical meaning.
We will see shortly that we need to normalise a wavefunction to obtain a sensible
probability distribution from it and make physical predictions. So to make physical
predictions from a superposition, we generally need to normalise the sum ψ =
a1 ψ1 + a2 ψ2 . As we’ve just seen, this is always possible if ψ1 , ψ2 are normalisable
and ψ is nonzero.
10 I.e. not the zero function; it does not vanish everywhere.

27
3.6 Probabilistic interpretation of the wavefunction: the Born
rule
Max Born in 1926 explained the essential connection between the wavefunction and
experiment, via the so-called Born rule:
If we carry out an experiment to detect the position of a particle described
by a normalised wavefunction ψ(x, t), the probability of finding the particle in the
interval [x, x + dx] at time t is

Z x+dx
|ψ(y, t)|2 dy ≈ dx|ψ(x, t)|2
x
= dxρ(x, t) , (3.13)

where we write ρ(x, t) = |ψ(x, t)|2 (see below).


More generally, the probability of finding the particle in any interval [a, b] is

Z b
|ψ(y, t)|2 dy . (3.14)
a
Intuitively, it may seem that (3.14) should follow from (3.13). Certainly, it would
be peculiar if the probability of finding a particle in a given interval depended on
how the interval was sub-divided (i.e. on how precise our position measurements
were). But we have already seen some apparently peculiar predictions of quantum
mechanics, which show it is not safe to rely on intuition. We should rather under-
stand (3.14) as a general postulate from which (3.13) follows as a special case. We
will see later (section 6.6) that (3.14) itself is a special case of the general quantum
measurement postulates, which apply to measurements of any physical quantity
(not only position).

28
3.6.1 Probability density and probability current
The following mathematical quantities give very useful insights into the behaviour
of solutions to the Schrödinger equation:
The probability density

ρ(x, t) = |ψ(x, t)|2 . (3.15)

We see that the Born rule justifies the interpretation of ρ(x, t) as a probability
density. If we measure the position of the particle at time t, the probability of
finding the particle in the interval [x, x + dx] is ρ(x, t)dx.
The probability current

i~ ∗ ∂
J(x, t) = − {ψ (x, t) ψ(x, t)−
2m ∂x
∂ ∗
( ψ (x, t))ψ(x, t)} . (3.16)
∂x
It is easy to verify from (3.10) that

∂J ∂ρ
+ = 0. (3.17)
∂x ∂t
Thus ρ and J do indeed satisfy a conservation equation, with ρ behaving as a
density whose total integral is conserved, and J as a current describing the density
flux.
The key point here is that ∂ρ ∂t can be written as a spatial derivative of some
quantity. This means that we can calculate the time derivative of the probability
of finding the particle in a region [a, b]:

Z b Z b
d 2 ∂
|ψ(x, t)| dx = − J(x, t)dx = J(a, t) − J(b, t) .
dt a a ∂x
(3.18)
The last term describes the probability density flux across the endpoints of the
interval – the “net flow of probability out of (or in to) the interval”.
Now if ψ is normalised, i.e. Eqn (3.12) holds, then ψ(x) → 0 as x → ±∞.

Thus J(x) → 0 as x → ±∞, assuming (as we will here) that ∂x ψ(x) is bounded as
x → ∞. Thus

Z ∞
d
|ψ(x, t)|2 dx = lim J(a, t) − lim J(b, t) = 0 . (3.19)
dt −∞ a→−∞ b→∞

TheR∞total probability of finding the particle in −∞ < x < ∞ is thus constant over
time: −∞ |ψ(y, t)|2 dy = 1 for any time t. So, the Born probabilistic interpretation
is consistent: whenever we look for the particle, we will definitely find it somewhere,
and only in one place.

Notes:

29
• We will consider measurements of quantities other than position later.

• The Born rule says nothing about where the particle is if we do not measure
its position. According to the standard understanding of quantum mechanics,
this is a question with no well-defined answer: the particle’s position is not
defined unless we measure it.
As we’ll see, according to quantum mechanics, we can generally only calcu-
late the probabilities of possible measurement results: we can’t predict with
certainty which result will occur. Moreover, the theory only allows us to pre-
dict probabilities for the possible results of measurements that actually take
place in a given experiment. We cannot consistently combine these predic-
tions with those for other measurements that could have been made had we
done a different experiment instead.11
We’ll see when we discuss the general measurement postulates of quantum
mechanics in section 6.6 that measuring the position of a particle generally
changes its wavefunction. Recall the earlier discussion of the 2-slit experiment.
We found no definite answer to the question “which slit did the particle go
through?” – unless we put detectors beside the slits, and this changed the
experiment and changed the interference pattern.

11 See again the analysis of the double slit experiment above, and (for example) the relevant

chapters of Feynman’s lecture notes, for further discussion of this.

30
4 Solutions of the 1D Schrödinger equation
We now look at various examples of solutions to the 1D SE for one particle:

4.1 Example I: Stationary states


~2 ∂ 2 ∂
− ψ(x, t) + V (x)ψ(x, t) = i~ ψ(x, t) . (4.1)
2m ∂x2 ∂t
Assume V is time-independent. We can then use the method of separation of
variables to find solutions, by trying the ansatz

ψ(x, t) = ψ(x)T (t) . (4.2)

This gives us


(Ĥψ(x))T (t) = ψ(x)(i~ T (t)) , (4.3)
∂t
and so
1 1 ∂
Ĥψ(x) = (i~ T (t)) . (4.4)
ψ(x) T (t) ∂t
Since the left hand side depends only on x and the right hand side only on t, both
must equal a constant, which we call E. We thus have

−i
T (t) = exp( Et) . (4.5)
~
and the time-independent Schrödinger equation

~2 d2
− ψ(x) + V (x)ψ(x) = Eψ(x) . (4.6)
2m dx2
Solutions to the SE of the form

−i
ψ(x) exp( Et),
~
where ψ(x) is a solution to (4.6) are called stationary states.
We say ψ(x) is an eigenfunction of the hamiltonian operator Ĥ which corre-
sponds to a physical state whose energy is given by the eigenvalue E. When V = 0,
(4.6) reduces to

~2 d2
− ψ(x) = Eψ(x) . (4.7)
2m dx2

More generally, if
Aψ(x) = aψ(x) (4.8)

31
for some operator A and complex number a, then we say ψ(x) is an eigenfunc-
tion of A with eigenvalue a. The terminology is a natural extension to infinite-
dimensional spaces (of functions) of the definitions of eigenvector and eigenvalue
for finite-dimensional matrices: compare the eigenequation Av = av for matrix A,
vector v and complex number a.
As we will see, the eigenvalues of physically significant operators are observable
physical quantities. They turn out always to be real – as we would expect, since
we know of no natural way to directly observe a complex number in nature. We
will see the mathematical reason for this when we discuss the general formalism of
quantum mechanics.

Notes
2 2
k
• Equation (4.7) has solution ei(kx−ωt) , where E = ~2m = ~ω. This is the
de Broglie wave we originally considered, so our discussion is at least self-
consistent.
• This de Broglie wave solution to the 1D SE for a free particle is not normalis-
able, and so does not have a well-defined probabilistic interpretation via the
Born rule. The quantum mechanical free particle is a very useful idealisa-
tion, but not a physical solution. We never actually have a uniformly zero
potential throughout space, nor a single particle whose wavefunction is spread
uniformly over all of space. There could not be any consistent way of assign-
ing probabilities to finding such a particle in finite regions, since there is no
well-defined uniform probability distribution on the 1D real line. (The same
is true in 3D, of course.)
iEt
• Any stationary state ψ(x, t) = ψ(x)e− ~ has probability density

ρ(x, t) = |ψ(x, t)|2 = |ψ(x)|2 .

In other words, its probability density is time-independent. More generally,


we will see the probability of measuring any given outcome for any dynamical
quantity is time-independent. Hence the name stationary state.

4.2 Completeness of the energy eigenfunctions

An important fact is that the general solution to the SE with t-independent potential
V (x) is a superposition of stationary states:

iEn t
X
ψ(x, t) = an ψn (x)e− ~ . (4.9)
n=1

This follows from the fact that any wavefunction at a given time, say t = 0, can
be written as a superposition of energy eigenfunctions, so that,

X
ψ(x, 0) = an ψn (x) , (4.10)
n=1

32
and from the linearity of the SE.
We will see later (from theorem 6.3) that Eqn. (4.10) is a special case of the
more general result that any wavefunction can be written as a superposition of
eigenfunctions for any operator corresponding to a physically observable quantity.
If the set of energy eigenvalues is continuous, we need to replace the sum by
an integral. If there are continuous and discrete subsets, we need both sum and
integral. Thus if there is a discrete set of normalised energy eigenfunctions {ψi }N
i=1

(we include N = ∞ as a possibility) and a continuous set of eigenfunctions {ψα }α∈∆ ,


where the indexing set ∆ is some sub-interval of the real numbers, we can write

N Z
− iE~n t iEα t
X
ψ(x, t) = an ψn (x)e + aα ψα (x)e− ~ dα , (4.11)
n=1 ∆

33
4.3 Example II: Gaussian wavepackets

As an example of a continuous superposition of free particle stationary states,


consider
Z ∞ 2
− σ2 (k−k0 )2 i(kx− ~k
2m t)
ψ(x, t) = C e e dk . (4.12)
−∞

The exponent here is − 12 (σ + i~t 2 σ 2


m )k + (k0 σ + ix)k − 2 k0 , which we can write in
1 b 2 b2 σ 2 i~t
the form − 2 a(k − a ) + ( 2a − 2 k0 ) , where a = σ + m , b = k0 σ + ix.
Now we have Z r
b 2
− 12 a(k− a ) 2π
e dk = . (4.13)
a

This gives us
s
0 2π (k0 σ + ix)2
ψ(x, t) = C exp( ) (4.14)
σ + i~t
m 2(σ + i~t
m )
s
2 00 1 (k0 σ + ix)2 (k0 σ − ix)2
|ψ(x, t)| = C 2 2 exp( + )
σ 2 + ~mt2 2(σ + i~t m ) 2(σ − i~t
m )
s 2k σx~t
1 σ 3 k02 − σx2 + 0m
= C 00 2 2 exp( 2 2 )
σ 2 + ~mt2 σ 2 + ~mt2
s
000 1 −σ(x − km 0 ~t 2
)
=C 2 2 exp( 2 2 ) , (4.15)
σ 2 + ~mt2 σ 2 + ~mt2
where C 0 , C 00 and C 000 are constants.

Since
s s
∞ 2 2
2(σ 2 + ~mt2 )π
Z
1
|ψ(x, t)|2 dx = C 000 2 2
−∞ 2σ (σ 2 + ~mt2 )
r
π
= C 000 (4.16)
σ
the normalisation condition gives us
r
000 σ
C = . (4.17)
π
Notes:
• The Gaussian wavepacket describes an approximately localised particle, almost
all of whose probability density is within a finite region of size ≈ N (σ 2 +

34
~2 t2 12 − 21
m2 ) σ at time t, for some smallish positive number N .12 This follows since
the probability density function is a Gaussian curve with standard deviation
2 2 1 1
(σ 2 + ~mt2 ) 2 σ − 2 .
• Recall that we obtained the Gaussian wavepacket as a superposition of sta-
tionary states. Considering its evolution, we thus see that a superposition of
stationary states is not necessarily stationary, since the Gaussian wavepacket’s
probability density function varies with time.
• From Eqn. (4.15), we see the speed of the Gaussian wavepacket’s centre is
≈ ~k
m . Since its standard deviation is
0

~2 t2 1 − 1 ~t
(σ 2 + 2
)2 σ 2 ≈ 1 (4.18)
m mσ 2
for large t, the speed at which it spreads in width is ≈ ~ 1 . Note that,
mσ 2
mathematically speaking, neither is necessarily bounded by the speed of light
c. This illustrates that the Schrödinger equation is non-relativistic – i.e. in-
compatible with special relativity. If it was possible to produce Gaussian
wavepackets for any values of k0 and σ, and if the Schrödinger equation pre-
cisely described their evolution, we could use the wavepackets to send signals
faster than light.13
As this suggests, the SE is only approximately valid. 14 We cannot use Gaus-
sian wavepackets (or, unless we find experimental evidence against special
relativity, anything else) to send signals faster than light. However, the SE
is a good enough approximation to allow us to understand a great deal of
physics and chemistry.
• Pictures of an evolving Gaussian wave packet can be found at (for example)
http://demonstrations.wolfram.com/EvolutionOfAGaussianWavePacket/

4.4 Example III: Particle in an infinite square potential well


Consider a particle confined by a 1D box with impermeable walls. We can
model this by a potential


0 |x| < a ,
V (x) = (4.19)
∞ |x| > a .

We look for stationary states obeying

~2 d2
− ψ(x) + V (x)ψ(x) = Eψ(x) , (4.20)
2m dx2
12 For example, 99.7% of the probability density is within the region defined by N = 3.
13 We can also see that the SE is non-relativistic by noting that it is second order in x and first
order in t. A relativistically invariant equation must be of the same order in x and t, because
Lorentz transformations map x and t to linear combinations.
14 Quantum mechanics, which is the topic of this course, is a non-relativistic theory, which is

approximately correct when relativistic corrections are small. Physicists tend to think of quantum
theory as defining a mathematical framework, and non-relativistic quantum mechanics as a par-
ticular theory within this framework. The framework of quantum theory also includes relativistic
quantum field theories, which (as the name suggests) are consistent with special relativity. In
particular, there is a relativistically invariant equation for the evolution of an electron’s wavefunc-
tion, from which the SE emerges as a non-relativistic limit. We won’t worry about relativistic
corrections or extensions of quantum mechanics in this course; they are discussed in Part III.

35
which gives us

~2 d2

Eψ(x) |x| < a ,
− ψ(x) = (4.21)
2m dx2 (E − ∞)ψ(x) |x| > a .
Here the second alternative is a useful but unrigorous shorthand, and should per-
haps be written in quotation marks to emphasize this. The next paragraph explains
how we handle the infinity.

4.4.1 Handling discontinuities


We need a consistent treatment of discontinuities in V (x) and (hence perhaps)
in ψ(x), dψ
dx , . . .. One method is to take a discontinuous V (x) as the limiting case
of a set of continuous potentials, and consider the limits of the solutions ψ(x). But
we can address the SE with discontinuous V (x) directly using the following rules:

A step function discontinuity in ψ would create a delta function derivative in


d2 ψ
dx2 on the LHS of the SE, which would not be cancelled by other terms. So we
expect ψ to be continuous everywhere.
d2 ψ
A step function discontinuity in dψ
dx would create a delta function in dx2 on the
LHS of the SE; we can allow this only at points where V (x) is infinitely discontin-
uous. 2
A step function discontinuity in ddxψ2 gives a step function discontinuity on the
LHS of the SE; we can allow this only at points where V (x) has a finite step function
discontinuity so that the two discontinuities on the LHS cancel.

So, for the infinite square well we expect ψ to be continuous everywhere, and

dx to be continuous except at |x| = a. Since V = ∞ for |x| > a, we need ψ = 0 in
this region for consistency of the SE. We have

r
2mE
ψ(x) = A cos(kx)+B sin(kx) , k= in |x| < a .
~2
(4.22)
Now ψ = 0 at x = ±a gives

A cos(ka) ± B sin(ka) = 0 (4.23)

and hence
A cos(ka) = 0 and B sin(ka) = 0 . (4.24)

36
Hence

either A = 0 , sin(ka) = 0 ,

ψ(x) = B sin(kx) , k= for n ≥ 2 even .
2a
or B = 0 , cos(ka) = 0 ,

ψ(x) = A cos(kx) , k= for n ≥ 1 odd .
2a
(4.25)

This gives us solutions for

~2 π 2 n2
E = En = for n = 1, 2, 3, . . . (4.26)
8ma2

Notes:

• We have quantised (i.e. discrete) energies for the stationary states of the infi-
nite square well. Cf. the Bohr atom – it begins to look plausible that quantum
mechanics could explain the apparent quantisation of physical states. Perhaps
the allowed electron orbits in the Bohr atom correspond to the stationary
states in a 3D Coulomb potential?15

• The solutions fall into two classes:



ψ(−x) even parity ,
ψ(x) = (4.27)
−ψ(−x) odd parity .

15 We will see later that the relationship is a bit more complicated, but the intuition is along the

right lines.

37
4.5 Parity

Let us consider this last phenomenon more generally. Suppose that V (x) =
V (−x) and that ψ(x) is a solution of the time-independent SE. We have

~2 d2
− ψ(x) + V (x)ψ(x) = Eψ(x) ,
2m dx2
~2 d2
− ψ(−x) + V (x)ψ(−x) = Eψ(−x) ,
2m dx2
obtaining the second equation by substituting y = −x in the first. Hence ψ(−x) is
also a solution of the time-independent SE.

There are two possibilities:


(i) ψ(x) and ψ(−x) are linearly independent. Then ψ(x) + ψ(−x) and ψ(x) −
ψ(−x) are also linearly independent, and have parity +1 and −1 respectively.
(ii) ψ(x) and ψ(−x) are linearly dependent: ψ(−x) = Aψ(x). Then
ψ(x) = Aψ(−x) = A2 ψ(x) (4.28)
so we have
A = ±1 . (4.29)
A = 1 implies ψ(x) = ψ(−x), so ψ has even parity. A = −1 implies ψ(x) =
−ψ(−x), so ψ has odd parity.

Conclusion Any solution of the t-independent SE with a symmetric potential


is a linear combination of solutions of definite parity.16

4.6 Example IV: Particle in a finite square well

A slightly more realistic model of a confined particle is given by the finite square
well potential 
−U |x| < a ,
V (x) = (4.30)
0 |x| > a .
– a crude model of the Coulomb potential of a charged nucleus. We will look
for bound state solutions – normalisable solutions – as opposed to unnormalisable
scattering solutions, whose physical interpretation we will discuss later.

A solution of energy E obeys

~2 d2

(E + U )ψ(x) |x| < a ,
− ψ(x) = (4.31)
2m dx2 Eψ(x) |x| > a .
If E > 0, the solution for |x| > a takes the form ψ = Aeikx + Be−ikx , where
k 2 = 2mE
~ . This is unnormalisable.
16 One can also prove that the ground state (lowest energy solution), if one exists, is unique and

has even parity. In fact, one can prove the stronger result that the ground state wavefunction is
unique, real and strictly positive, which implies that it must have even parity. See for example,
volume 4 of Reed and Simon, “Methods of Modern Mathematical Physics”, Theorem XIII.4.6. A
proof is also given in Eugene Lim’s lecture notes for an earlier version of this course, available
online.

38
So we require E < 0. We will assume that E > −U , i.e. that |E| < U , since we
do not expect there to be solutions with energy lower than the minimum potential.

Exercise: Verify that indeed there are no such solutions.

Let
2m|E|
k2 = , (4.32)
~2
2m(U − |E|)
l2 = , (4.33)
~2
2mU
so that k 2 + l2 = 2 . (4.34)
~

For |x| > a we have

ψ(x) = Ae−kx + Bekx , (4.35)

and normalisability implies that

Ae−kx x > a ,

ψ(x) = (4.36)
Bekx x < −a .
For |x| < a we have

ψ(x) = C cos(lx) + D sin(lx) . (4.37)

Since we know that the states of definite parity span the space of all bound
states, we can solve separately for even and odd parity states to get a complete
spanning set of solutions.
Consider the even parity solutions. If ψ(x) = ψ(−x) we have:

 Ae−kx

x > a,
kx
ψ(x) = Ae x < −a , (4.38)
C cos(lx) |x| < a.

 −kAe−kx

x > a,
0 kx
ψ (x) = Ake x < −a , (4.39)
−lC sin(lx) |x| < a.

As ψ and ψ 0 are continuous at x = ±a, we find

C cos(la) = A exp(−ka) , (4.40)

−lC sin(la) = −kA exp(−ka) . (4.41)

39
This gives
k = l tan(la)
or
(ka) = (la) tan(la) (4.42)

and from (4.34) we have

2 2mU a2
2
(ka) + (la) = . (4.43)
~2
We can solve these last two equations graphically.
The odd parity bound states (if any exist) can be similarly obtained.

The finite square well illustrates a general feature of quantum potentials V (x)
such that V (x) → 0 as x → ±∞. The time-independent SE has two linearly
independent solutions.
For E > 0, these take the form of scattering solutions

ψ(x) = A exp(ikx) + B exp(−ikx) as x → −∞


C(A, B) exp(ikx) + D(A, B) exp(−ikx) as x → ∞ .
For E < 0, they have the form

ψ(x) = A exp(kx) + B exp(−kx) as x → −∞


C(A, B) exp(kx) + D(A, B) exp(−kx) as x → ∞ .
Here C(A, B) and D(A, B) are linear in A and B, and depend also on k and on
the parameters defining V (x).

Only for special values of E < 0 can we find solutions such that B = 0 and
C(A, B) = 0, as required for normalisability. For the remaining values of E < 0,
the solutions blow up exponentially as x → −∞ or as x → ∞ (or both) and are
unphysical: neither bound states nor scattering solutions.
[These asymptotically exponential functions are not physically meaningful: they
are not normalisable and do not define eigenfunctions of Ĥ.]

40
4.7 Example V: The quantum harmonic oscillator

has potential
1
V (x) = mω 2 x2 . (4.44)
2
This is a particularly important example for several reasons.
First, it is not only solvable but (as we will see later) has a very elegant solution
that explains some important features of quantum theory.
Second, generic symmetric potentials in 1D are at least approximately described
by Eqn. (4.44). To see this, consider a potential V (x) taking a minimum value Vmin
at xmin and symmetric about xmin . Writing y = x − xmin , we have

y 2 00 y3
V (y) = V (0) + yV 0 (0) + V (0) + V 000 (0) + . . .
2! 3!
y 2 00
≈ Vmin + V (0) + O(y 4 ) .
2!
Renormalising the potential so that Vmin = 0, and taking V 00 (0) = mω 2 , we recover
(4.44). (If V 00 (0) = 0 this argument does not hold, but even in this case the harmonic
oscillator potential can be a good qualitative fit. See for example comments below
on approximating a finite square well potential by a harmonic oscillator potential.)
In fact, we can show something more general. Suppose we have a system with n
degrees of freedom that has a unique stable equilibrium. Then the small oscillations
about equilibrium can be approximated by n independent harmonic oscillators. This
means that solving the harmonic oscillator allows us to give a quantum description
of molecules excited by radiation (and hence understand the greenhouse effect and
many other phenomena) and of the behaviour of crystals and other solids.
Third, the harmonic oscillator plays an even more fundamental role in quantum
field theory, where we understand particles as essentially harmonic oscillator exci-
tations of fields. Our entire understanding of matter and radiation is thus based on
the quantum harmonic oscillator!
To obtain the stationary states of the quantum harmonic oscillator we need to
solve
~2 d2 1
− 2
ψ(x) + mω 2 x2 ψ(x) = Eψ(x) . (4.45)
2m dx 2
We define r
mω 2E
y= x and α = . (4.46)
~ ~ω
We have
d2 ψ
2
− y 2 ψ = −αψ . (4.47)
dy

If ψ ∼ exp(− 21 y 2 ), then ψ 0 ∼ −y exp(− 21 y 2 ) and ψ 00 ∼ y 2 exp(− 12 y 2 ) to leading


order, which means that ψ 00 − y 2 ψ ∼ (y 2 − y 2 ) exp(− 21 y 2 ) satisfies (4.47) to leading
order. Similarly ψ ∼ y n exp(− 12 y 2 ) satisfies (4.47) to leading order. This suggests
considering power series solutions of the form

41
2
− y2
ψ(y) = H(y)e , (4.48)

X∞
H(y) = am y m , (4.49)
m=0
and trying to match the coefficients am to produce an exact solution.

From (4.47) we get


ψ 00 + (α − y 2 )ψ = 0 (4.50)

and substituting from (4.48, 4.50) we get

d2 y2 2 y2
(H exp(− )) + (α − y )H exp(− ) = 0 (4.51)
dy 2 2 2

00 0 2 y2 2
(H − 2yH + Hy − H + (α − y )H) exp(− ) = 0 (4.52)
2
00 0 y2
(H − 2yH + (α − 1)H) exp(− ) = 0 (4.53)
2

We have

X
H 0 (y) = mam y m−1 , (4.54)
m=0
X∞
H 00 (y) = m(m − 1)am y m−2
m=0
X∞
= am+2 (m + 2)(m + 1)y m . (4.55)
m=0

So from (4.53) we get



X
(am+2 (m + 2)(m + 1) − am (2m + 1 − α))y m = 0 . (4.56)
m=0
This must be true for all y, so each coefficient of y m must vanish:

(2m + 1 − α)
am+2 = am for m = 0, 1, 2, . . . . (4.57)
(m + 2)(m + 1)
which determines the am for m ≥ 2 in terms of a0 and a1 .

42
2
For large m, (4.57) gives am+2 ≈ m am , which asymptotically describes the
2
coefficients of ey :
∞ ∞
y2
X 1 2n X
e = y = bn y n (4.58)
n=0
n! n=0
where for n odd we have bn = 0 and for n even we have

1 2 2
bn+2 = n bn = bn ≈ bn . (4.59)
2 +1 n+2 n
So, if the power series is infinite, we expect
2
H(y) ≈ Cey
H(y)
for some nonzero constant C: i.e. that Cey2
→ 1 as y → ∞. This implies that

y2 1 2
ψ(y) ≈ H(y)e− 2 ≈ Ce 2 y , (4.60)

which is not normalisable.


We can justify this rigorously by noting that, for any  > 0, there exists some
integer m0 such that for all m ≥ m0 we have
am+2 2(1 − )
> . (4.61)
am m
Hence we have that 2
H(y) > Ce(1−)y − P (y) , (4.62)
where C = am0 6= 0 and P (y) is a polynomial of degree ≤ m0 . This implies that
there exists a y0 such that for all y ≥ y0 we have
C (1−)y2
|H(y)| > e , (4.63)
2
and hence
C ( 1 −)y2
|ψ(y)| >
e 2 . (4.64)
2
So indeed ψ(y) diverges as y → ∞, and in particular ψ is not normalisable.

We thus need the power series to be truncated to a polynomial in order to find


normalisable physical solutions. Again, we can consider fixed (even and odd) parity
solutions separately:

• Even parity: a0 6= 0, a1 = 0, α = 2m + 1 for some even


m.
• Odd parity: a0 = 0, a1 6= 0, α = 2m + 1 for some odd
m.
So the physical solutions are given by

2E
α= = 2m + 1 (m = 0, 1, 2, . . .) (4.65)

43
i.e.
1
E = (m + )~ω (m = 0, 1, 2, . . .) . (4.66)
2
The harmonic oscillator has energy levels equally spaced, separated by ~ω, with
minimum (or zero-point) energy 21 ~ω.

The polynomials Hn (y) corresponding to the physical solutions with energy


(n+ 12 )~ω are the Hermite polynomials of degree n. We can obtain them explicitly17
from (4.65), using conventional normalisations:

H0 = 1 , H1 = 2y , H2 = 4y 2 − 2 , H3 = 8y 3 − 12y , . . . . (4.67)

The corresponding wavefunctions are thus (up to normalisation)


1 2 1 2
ψ0 = e− 2 y , ψ1 = 2ye− 2 y , . . . (4.68)

17 In fact one can derive a general expression (see e.g. Schiff, “Quantum Mechanics”, 3rd edition):
2 dn −y2
Hn (y) = (−1)n ey e .
dy n

44
5 Tunnelling and Scattering
Let us reconsider the bound state solutions for the finite square well potential:
 −kx
 e x > a,
ψ(x) ∼ ekx x < −a , (5.1)
cos(lx) or sin(lx) |x| < a.

Although ψ tends to zero rapidly outside the well, it is non-zero there. In


particular Z
|ψ(x)|2 dx > 0 (5.2)
|x|>a
2 2
k
for a bound state solution, although the energy E = − ~2m < 0.

The Born rule (3.13) thus tells us that the probability of finding the particle
outside the well is non-zero, even though the particle’s energy is less than the
potential height in this region (V = 0 in |x| > a by our convention).
Clearly, a classical particle with E < 0 would never make its way outside the
well. We have here a suggestive argument for the existence of an intrinsically
quantum phenomenon – the possibility of tunnelling through a potential barrier
into a classically forbidden state.
However, the physical interpretation of this calculation is complicated by the fact
that we can only assign meaning in quantum mechanics to things we can detect,
and we do not have any way to detect negative energy particles. So, we cannot
actually observe a bound state particle outside the region |x| < a unless we alter
the potential.
What we can and do observe are quantum particles tunnelling through finite
width potential barriers through which a classical particle of the same energy could
not pass.

5.1 Scattering states


To understand tunnelling we need a general treatment of scattering: the transmis-
sion or reflection of particles by potential barriers (or potential wells).
In principle, we could take as initial state an incoming wavepacket (e.g. an
approximately Gaussian packet) far from the potential barrier, solve the time-
dependent Schrödinger equation, and obtain a solution that at asymptotic late
times takes the form of a superposition of two wavepackets, one reflected by the
barrier and one transmitted through it:

ψ ∼ ψr + ψt . (5.3)

We could then calculate the reflection probability


Z 0
2
|ψr | ≈ |ψ(x, t)|2 dx for large t (5.4)

and transmission probability
Z ∞
2
|ψt | ≈ |ψ(x, t)|2 dx for large t . (5.5)
0

45
Some pictures of this process can be found in Schiff (3rd edition, pp 106-9), and
in Brandt and Dahmen, “Picture Book of Quantum Mechanics”.

However, studying wavepacket scattering directly is mathematically compli-


cated. We can model it computationally, and this is often helpful and illuminating,
but it is not so easy to prove simple analytic results.
A simpler way to obtain scattering probabilities is to consider unnormalised
stationary state solutions that asymptotically take the form of superpositions of
plane waves as x → ±∞. Thus:

ψ = exp(ikx) + R exp(−ikx) as x → −∞
ψ = T exp(ik 0 x) as x → ∞ .

If V = 0 as x → ±∞ then k = k 0 . We are also interested in examples where V


has different limits as x → ±∞, so allow k 0 6= k.
Here we follow the convention that incoming plane waves arrive from −∞ but
not ∞. Now exp(ikx − ωt) is rightward travelling, and exp(−ikx − ωt) is leftward
travelling. We consider a stationary solution at (for simplicity) t = 0. We thus
allow a component of the rightward travelling incoming wave exp(ikx) as x → −∞
but do not allow any component of a leftward travelling wave exp(−ik 0 x) as x → ∞,
which would represent a wave incoming from ∞.
We expect in general a reflected plane wave of form

R exp(−ikx) as x → −∞ ,
, and a transmitted plane wave of form

T exp(ik 0 x) as x → ∞ .

5.2 Interpretation of plane wave scattering solutions


We can justify taking these unnormalised stationary state solutions as representing:
(i) an approximation to the behaviour of a 1-particle incoming wavepacket with
wave vector sharply peaked about k.
(ii) an approximation to the behaviour of a beam of particles (whose interactions
with one another are negligible) with approximate wave vector k.
(iii) a mathematical calculation of the behaviour of the wave vector k Fourier
component of (i) or (ii).

The beam picture is perhaps the most intuitive, and we will use it, considering
beam scattering from various potentials. We can interpret |R|2 as the density of
particles in the reflected beam, |T |2 as the density in the transmitted beam. Recall
p
that we have p = ~k and so the speed v = m = ~k
m . The particles in the incoming,
reflected and transmitted beams thus have speeds

~k ~k ~k 0
,− , (5.6)
m m m
46
respectively.

We define the particle flux in the beams to be the number of particles per second
in the beam passing a fixed point. We have
 ~k
 m incoming ,
2
flux = velocity × density = m 0|R|
~k
reflected ,
2
m |T |
~k
transmitted.

(5.7)
Particle conservation – no particles are destroyed or created in the scattering process
– thus implies that

~k ~k 2 ~k 0 2
= |R| + |T | . (5.8)
m m m
We will verify this in particular examples.

5.3 Example I: The potential step


Consider the potential

0 x < 0,
V (x) = (5.9)
U x > 0.
A solution of energy E obeys

~2 d2

(E − U )ψ(x) x > 0 ,
− ψ(x) = (5.10)
2m dx2 Eψ(x) x < 0.
For the moment we consider U > 0 (a step rather than a drop).

We want solutions of the form exp(ikx) + R exp(−ikx) for x < 0; these have
2 2
k
E = ~2m > 0.
Case 1: E < U Define
√ p
2mE 2m(U − E)
k= , l= . (5.11)
~ ~

We have
d2 l2 ψ(x)

x > 0,
ψ(x) = 2 (5.12)
dx2 −k ψ(x) x < 0 .

So

A exp(−lx) x > 0,
ψ(x) = (5.13)
exp(ikx) + R exp(−ikx) x < 0 .
We see there is no transmitted plane wave in this case.

47
Continuity of ψ and ψ 0 at x = 0 gives

1+R=A (5.14)

ik(1 − R) = −lA (5.15)

Hence

2k
A= (5.16)
k + il
k − il
R= . (5.17)
k + il
In particular |R|2 = 1: the reflected flux equals the incoming flux, and thus the
reflection probability for any given incoming particle is one.

Case 2: E > U Now we have


p
2m(E − U )
l= (5.18)
~

and

T exp(ilx) x > 0,
ψ(x) = (5.19)
exp(ikx) + R exp(−ikx) x < 0 .
Continuity of ψ, ψ 0 at x = 0 gives

T =1+R (5.20)
l
T = 1−R. (5.21)
k
Hence

2k
T = , (5.22)
k+l
k−l
R= . (5.23)
k+l

The incoming, reflected and transmitted fluxes are respectively

~k
FI = (5.24)
m
~k 2 ~k k − l 2
FR = |R| = ( ) (5.25)
m m k+l
~l ~l 4k 2
FT = |T |2 = ( ). (5.26)
m m (k + l)2

48
We see that

~ k(k − l)2 + 4k 2 l
F R + FT =
m (k + l)2
~ k(k + l)2
=
m (k + l)2
~k
=
m
= FI (5.27)

as particle conservation requires.

Comments 1. Case 1 (E < U ) accords at least roughly with classical intu-


ition: a particle with insufficient energy to climb the step is (eventually) reflected
back.
2. The calculations for case 2 (E > U ) apply for U < 0 – a downward step – as
well as for U > 0. In the classical limit in which the step is negligible compared to
the particle energy, E  |U |, we have

k ≈ l , FR ≈ 0 , FT ≈ 1 .
This implies near-perfect transmission, again as classical intuition would suggest.

3. But the results in general do not accord with classical intuition. Consider for
example the case U < 0 and E  |U |. Here we have

l  k , |R|2 ≈ 1 , |T |2 ≈ 0 .
In other words, we find near perfect reflection from a downward step, precisely
the opposite result to that indicated by classical intuition.

5.4 Example II: The square potential barrier


Consider an incoming beam of particles with E < U scattering from the po-
tential

 0 for x < 0 ,
V (x) = U for 0 < x < a (5.28)
0 for x > a .

We have

 T exp(ikx) for x > a ,
ψ(x) = A exp(−lx) + B exp(lx) for 0 < x < a ,
exp(ikx) + R exp(−ikx) for x < 0 ,

(5.29)

49
where
~2 k 2 ~2 l2
E= , U −E = . (5.30)
2m 2m

Continuity of ψ and ψ 0 at x = 0 gives

1+R =A+B, (5.31)

ik(1 − R) = l(B − A) , (5.32)

Continuity of ψ and ψ 0 at x = a gives

A exp(−la) + B exp(la) = T exp(ika) , (5.33)

−lA exp(−la) + lB exp(la) = ikT exp(ika) . (5.34)

Exercise Solve the algebra, to obtain

4ikle−ika
T = . (5.35)
(2ikl + l2 − k 2 )e−la + (2ikl − l2 + k 2 )ela

This gives

16k 2 l2
|T |2 = . (5.36)
4k 2 l2 (ela + e−la )2 + (l2 − k 2 )2 (ela − e−la )2

In the case la  1 we can neglect the exp(−la) terms, obtaining

16k 2 l2
|T |2 ≈ . (5.37)
4k 2 l2 exp(2la) + (l2 − k 2 )2 exp(2la)
16k 2 l2
≈ exp(−2la) . (5.38)
(k 2 + l2 )2

2m(U −E)
Since l = ~ , we find

2a p
|T |2 ≈ exp(− 2m(U − E)) . (5.39)
~

Comments 1. The most immediately striking result is the fact that T is


nonzero: quantum particles can tunnel through classically impenetrable potentials.
2. If la  1 the tunnelling probability is small; we approach the classical limit
as la → ∞. The tunnelling probability
√ in this
√ regime depends inverse exponentially
on the barrier width a and on U − E ≈ U (if E  U ). This is a general feature
of tunnelling, true for a wide class of barrier potentials.

50
5.4.1 *Important examples

• Nuclear fission.
• Nuclear fusion.
• Muon-catalysed nuclear fusion.

• Scanning tunnelling electron microscopy. *

51
6 The basic formalism of quantum mechanics

6.1 Spaces of functions


We will not be too precise about the space of functions to which physically
relevant wavefunctions belong in any given context. One possible example is the
set of infinitely differentiable complex-valued functions ψ(x) of one real variable x.
This forms a complex vector space which we denote by C ∞ (R). Another is the space
of normalisable wavefunctions, which again forms a complex vector space, L2 (R).
Exercise Check these are indeed vector spaces.

6.2 The inner product

We can define a natural inner product on the space of normalisable wavefunctions


by Z ∞
(ψ1 , ψ2 ) = ψ1∗ (x)ψ2 (x)dx . (6.1)
−∞
Exercise: Show that this is well-defined whenever ψ1 and ψ2 are normalisable.
[Hint: look ahead to Thm. 6.6 below.]
We extend the definition of the inner product ( , ) to any ψ1 and ψ2 for which
the integral is well-defined, whether or not ψ1 or ψ2 is normalisable.

6.2.1 Properties of the inner product


1. (ψ1 , ψ2 ) = (ψ2 , ψ1 )∗

2. The inner product is anti-linear in the first entry and linear in the second:

(a1 ψ1 + a2 ψ2 , φ) = a∗1 (ψ1 , φ) + a∗2 (ψ2 , φ) ,


(ψ, a1 φ1 + a2 φ2 ) = a1 (ψ, φ1 ) + a2 (ψ, φ2 ) .

3. The inner product is positive definite on continuous wavefunctions, i.e.

(ψ, ψ) ≥ 0 and (ψ, ψ) = 0 if and only if ψ = 0 .

To see this, note that


Z ∞ Z ∞

(ψ, ψ) = ψ (x)ψ(x)dx = |ψ(x)|2 dx .
−∞ −∞

52
6.3 Operators

Recall that we defined an operator Ô to be any linear map from a space of functions18
(for example C ∞ (R) or L2 (R)) to itself, i.e. any map such that

Ô(a1 ψ1 + a2 ψ2 ) = a1 Ôψ1 + a2 Ôψ2 (6.2)

for all complex numbers a1 , a2 and all ψ1 , ψ2 ∈ C ∞ (R).

6.3.1 Some examples of operators


∂i
Pn
• Finite differential operators i=0 pi (x) ∂x i , where the pi (x) are polynomials

in x. (This class includes p̂ and x̂, defined below.)


• The translation operator

Sa : ψ(x) → ψa (x) = ψ(x − a) .

• The parity operator

P : ψ(x) → ψP (x) = ψ(−x) .

6.4 Hermitian operators

We define the hermitian conjugate A† of an operator A to be the operator such


that
(A† ψ1 , ψ2 ) = (ψ1 , Aψ2 ) (6.3)

for all normalisable wavefunctions ψ1 , ψ2 . An operator is hermitian if A = A .
Exercise Verify the identities

(a1 A1 + a2 A2 )† = a∗1 A†1 + a∗2 A†2 ,


(AB)† = B † A† .

6.4.1 Classical states and dynamical variables


By the state of a physical system we mean its mathematical representation in
a given theory. In classical mechanics, the states are points in a 2n-dimensional
space, phase space. For example, a system of n point particles in 1D is described by
2n coordinates: (x1 , . . . , xn , p1 , . . . , pn ), where xi is the position of the ith particle
18 *Non-examinable comment: In our discussions below, we generally take the space of
functions to be the space L2 (R) of normalisable physical wavefunctions satisfying Eqn. (3.11).
Strictly speaking, a completely rigorous discussion ought to take account of subtleties which arise
from this choice. For example, the momentum operator is not actually defined on all functions
in L2 (R), but only those that are differentiable. Similarly, the position operator is not defined
on all functions in L2 (R), but only those normalisable functions ψ(x) for which xψ(x) is also
normalisable.
However, a fully rigorous treatment would go well beyond the scope of this course. We shall
follow standard practice in introductory treatments of quantum mechanics and simply assume,
whenever we have an equation in which the position, momentum or other operators act on physical
wavefunctions, that the wavefunctions are chosen such that the action of the operators is well
defined. [You should make the same assumption in tackling problem sheets or exam questions.] *

53
positions and pi its momentum. Classical dynamical variables – for instance the
energy
n
X p2i
E= + V (x1 , . . . , xn ) (6.4)
i=1
2mi
– are defined by functions on phase space. Note that the classical state of a finite
number of particles can be described by a finite number of parameters.
In principle, the classical state and (hence) the value of all classical dynamical
variables can be measured with arbitrary precision.

6.4.2 Quantum states and observables


The possible quantum mechanical states of a physical system at any given time
t are normalised wavefunctions. In particular,
R ∞ the state of a single particle in 1D is
given by a wavefunction ψ(x, t) such that −∞ |ψ(x, t)|2 dx = 1. We need infinitely
many parameters to specify a normalised wavefunction, and thus to specify the
state of a quantum system – even a single particle in 1D.

All quantum dynamical variables or observables – quantities we can measure


– are represented by hermitian operators, and vice versa. Examples of hermitian
operators defining quantum observables are the position, momentum and energy
operators


p̂ = −i~ (6.5)
∂x
x̂ = x (i.e. multiplication by x) (6.6)

~2 ∂ 2
Ĥ = − + V (x) (6.7)
2m ∂x2
Another example is the parity operator

P : ψ(x) → ψP (x) = ψ(−x) . (6.8)

Exercise (important!): Verify that each of the operators (6.5-6.8) is


hermitian.

54
6.5 Some theorems about hermitian operators
Theorem 6.1. The eigenvalues of a hermitian operator are real.

Proof. Let A be hermitian and ψ be a normalised eigenfunction with eigenvalue


a: Aψ = aψ. We have

(ψ, Aψ) = (ψ, aψ) = a(ψ, ψ) = a ,


(ψ, Aψ) = (A† ψ, ψ) = (Aψ, ψ)
= (aψ, ψ) = a∗ (ψ, ψ) = a∗ .
Hence a = a∗ . 

Theorem 6.2. Let A be a hermitian operator and ψ1 , ψ2 be normalised eigen-


functions with different eigenvalues a1 , a2 . Then ψ1 and ψ2 are orthogonal.

Proof. We have Aψ1 = a1 ψ1 and Aψ2 = a2 ψ2 , and by theorem 6.1 we have


that a1 , a2 are real.

a1 (ψ1 , ψ2 ) = a∗1 (ψ1 , ψ2 ) = (a1 ψ1 , ψ2 )


= (Aψ1 , ψ2 ) = (A† ψ1 , ψ2 )
= (ψ1 , Aψ2 ) = (ψ1 , a2 ψ2 )
= a2 (ψ1 , ψ2 ) .
Hence (a1 − a2 )(ψ1 , ψ2 ) = 0, and since a1 6= a2 we have (ψ1 , ψ2 ) = 0. 

Our discussion is complicated by the fact that the eigenfunctions of hermitian


operators (i.e. of quantum observables) are not necessarily all normalisable. For
example:

• The momentum operator p̂ = −i~ ∂x has eigenfunctions exp( ~i px) with eigen-
value p.
• The position operator x̂ has eigenfunctions δ(x − q) with eigenvalue q: we
have x̂δ(x − q) = xδ(x − q) = qδ(x − q).
1 2
• The energy operator Ĥ = 2m p̂ + V (x), for a finite square well potential
V (x), has normalisable eigenfunctions (corresponding to bound states) and
unnormalisable eigenfunctions (corresponding to scattering solutions).

Theorems 6.1 and 6.2 extend to these non-normalisable eigenfunctions of p̂, x̂


and Ĥ: these are the only operators with continuous sets of eigenfunctions that we
consider in this course.
In general, a hermitian operator may have both a set of normalisable eigenfunc-
tions with discrete eigenvalues and a set of non-normalisable eigenfunctions with
continuous eigenvalues, and the theorems apply to both sets.19
19 *Non-examinable technical note: it is well beyond our scope here but, in fact, it turns out

one can find a more general notion of normalisability which covers both sets of eigenfunctions,
and more general versions of the theorems can be precisely framed in terms of this condition.
This definition includes bound states and scattering solutions to the time-independent SE, but
not solutions which blow up exponentially. A discussion can be found in, for example, Messiah,
“Quantum Mechanics”, vol. 1, chap V.9.*

55
Theorem 6.3. The discrete and continuous sets of eigenfunctions of any her-
mitian operator together form a complete orthogonal basis of the physical wavefunc-
tions, i.e. of the normalisable complex-valued functions ψ(x) of one real variable
x.
Proof. This is quite hard to prove in complete generality. We will assume it without
proof in this course. 

Note: We say the hermitian operator A has a degenerate eigenvalue λ if


it has more than one linearly independent eigenfunction with eigenvalue λ. If A
has degenerate eigenvalues, we define an orthogonal basis of its eigenfunctions by
choosing orthogonal bases for the eigenfunctions associated with each degenerate
eigenvalue λ.

Corollary 6.3.1. Let A be a hermitian operator with a discrete set of nor-


malised eigenfunctions {ψi }N
i=1 (we include N = ∞ as a possibility) and a contin-

uous set of eigenfunctions {ψα }α∈∆ , where the indexing set ∆ is some sub-interval
of the real numbers. Then any wavefunction ψ can be written as

N
X Z
ψ= ai ψi + aα ψα dα , (6.9)
i=1 ∆

where ai = (ψi , ψ) and aα are complex numbers.


Comment In fact, by appropriately normalising, we can also ensure that
aα = (ψα , ψ).

Corollary 6.3.2. Let A be a hermitian operator with (only) a discrete set of


orthonormalised eigenfunctions {ψi }∞
i=1 . Then any wavefunction ψ can be written

as

X
ψ= (ψi , ψ)ψi . (6.10)
i=1

Proof. Theorem 6.3 implies that



X
ψ= ai ψ i (6.11)
i=1
for some complex numbers ai . Taking the inner product with ψj , and applying
theorem 6.2, we see (ψj , ψ) = aj . 

56
6.6 Quantum measurement postulates
• Postulate 1 Every quantum observable O is represented by a hermitian op-
erator Ô.
• Postulate 2 The possible outcomes of a measurement of O are the eigenvalues
of Ô.

• Postulate 3 If Ô has (only) a discrete set of normalised eigenfunctions


{ψi }∞
i=1 with corresponding distinct eigenvalues {λi }, and a measurement of

O is carried out on a particle with normalised wavefunction



X
ψ= ai ψi ,
i=1

then the probability of outcome λi is |ai |2 .


• Postulate 4 (the projection postulate) If a measurement of the observable O
is carried out on a particle with normalised wavefunction ψ(x, t) at time t and
the outcome λi is obtained, the wavefunction instantaneously after the mea-
surement becomes ψi (x). [This is sometimes referred to as the “collapse of the
wavefunction”.] The wavefunction then evolves according to the Schrödinger
equation, with initial state ψi (x) at time t, until the next measurement.

Notes

• It follows from postulates 2 and 3 that the total probability of all possible
outcomes is
X X X
|ai |2 = (ai ψi , ai ψi ) = (ai ψi , aj ψj ) = (ψ, ψ) = 1 .
i i ij
(6.12)

So the postulates are consistent: the sum of the probabilities of all possible
outcomes is 1, and so if you carry out a measurement you will certainly get
some outcome and you will only get one outcome. (We already verified this
in the case of the Born rule for position measurements.)

• It follows from postulates 3 and 4 that if an observable O is measured twice,


with infinitesimal time separating the two measurements, then if the first
outcome is λi the second will, with probability one, also be λi .

Exercise: Check this.

This has the important consequence that quantum measurements resemble


classical measurements in at least one sense: they establish a property of the
system that can be repeatedly verified. If we measure something, and then
quickly measure it again, we get the same answer. If this was not true, it
would be hard to find any good reason for the use of the term “measurement”
in quantum mechanics.

57
• If the wavefunction ψ is an eigenfunction ψi of Ô, the measurement outcome
will be λi with probability one. For example, a stationary state obeying
Ĥψ = Eψ will always give outcome E if the energy is measured.
• But unless the wavefunction ψ is an eigenfunction of the measured observable,
the measurement outcome is not definitely predictable. In contrast to classical
mechanics, a quantum observable does not generally have a definite value on
a quantum state.
• We can extendP postulates 3 and 4 to the case when Ô has degenerate eigen-
values. If ψ = i ai ψi is measured, where {ψi }∞i=1 are orthonormalised eigen-
functions of Ô and {ψi }i∈I are a complete set of orthonormalised P
eigenfunc-
tions with the same eigenvalue λ, the probability of outcome λ is i∈I |ai |2 ,
and the state
P resulting after a measurement with outcome λ is (up to normal-
isation) i∈I ai ψi .

• The projection postulate is so called because it implies that the the post-
measurement wavefunction ψi (x, t) is obtained from the pre-measurement
wavefunction ψ(x, t) by the action of the projection operator Pi defined by
Pi : ψ → (ψi , ψ)ψi , up to normalisation. We call Pi a projection since it maps
any state onto its component in a particular linear subspace, namely the sub-
space spanned by ψi – an action analogous to, for instance, the projection of
a 3D vector (x, y, z) onto its x-component (x, 0, 0).

6.7 Expectation values

Consider a measurement of the observable A on the state ψ. If the correspond-


ing hermitian operator  has only a discrete set of normalisable eigenfunctions
{ψi }, the possible outcomes are the corresponding eigenvalues λi , and the outcome
probabilities are pi = |(ψ, ψi )|2 . The expectation value of the measured outcome, in
the standard statistical sense of the term, is thus
X X
pi λi = |(ψ, ψi )|2 λi
i i
X X
= ( (ψ, ψi )ψi , λj (ψ, ψj )ψj )
i j

= (ψ, Âψ) . (6.13)

We write hÂiψ for the expectation value of a measurement of A on the state ψ.


Thus we have
hÂiψ = (ψ, Âψ) . (6.14)

We can similarly justify this definition of expectation value for the position op-
erator x̂ from the Born rule. Recall that the probability of obtaining a position
measurement outcome in the interval [x, x + dx] is given by |ψ(x)|2 dx. The expec-
tation value of a position measurement is thus

58
Z ∞ Z ∞
2
x|ψ(x)| dx = ψ ∗ (x)xψ(x)dx = (ψ, x̂ψ) .
−∞ ∞

This encourages us to take Eqn. (6.14) as a general definition of expectation


value for any observable, whether its eigenvalues are discrete, continuous or a com-
bination. This general definition too can be similarly justified, given a careful
treatment of the eigenfunctions of general observables and their normalisation: this
is beyond our scope here, but it is important to note that the definition turns out
to agree with the statistical definition of expectation in all cases. In particular, the
definitions

hp̂iψ = (ψ, p̂ψ) , hĤiψ = (ψ, Ĥψ) ,


often give simple ways of calculating the statistical expectation of measurements
of momentum and energy for general wavefunctions ψ, since the right hand sides
are often given by simple integrals.

Note that the expectation value is linear with respect to real scalars: i.e.

ha + bB̂iψ = ahÂiψ + bhB̂iψ , (6.15)

for any hermitian operators A, B and any real numbers a, b. We restrict to a, b real
here because the interpretation of hÂiψ as an expectation value of an observable
requires that A is hermitian, since observables are always represented by hermitian
operators. A complex multiple of a hermitian operator is not generally hermitian:
if A is hermitian then (aA)† = a∗ A† .

6.8 Commutation relations

We define the commutator of two operators by

[A, B] = AB − BA . (6.16)

It is easy to verify the identities

[A, A] = 0 ,
[A, B] = −[B, A] ,
[A, BC] = [A, B]C + B[A, C]
[AB, C] = A[B, C] + [A, C]B . (6.17)

Note that the commutator [A, B] is a sum of products of operators, and thus
itself an operator. Note also that it depends linearly on both entries.

59
The commutator plays a crucial role in describing symmetries of quantum me-
chanical systems, as we will see when we consider angular momentum. It also
gives a way of calibrating how close two operators are to having simultaneously
determinable eigenvalues: see the following note and the later discussion of the
uncertainty principle.

Definition We say two hermitian operators A and B are simultaneously


diagonalisable if the space of normalisable wavefunctions has a complete basis of
joint eigenfunctions {ψi }, i.e. of eigenfunctions such that Aψi = ai ψi and Bψi =
bi ψi for some real numbers ai , bi .
Theorem 6.4. Two hermitian operators A and B are simultaneously diagonalisable
if and only if [A, B] = 0.

Proof. If A and B are simultaneously diagonalisable, the space of normalisable


wavefunctions has a complete basis of joint eigenfunctions {ψi }. Now for any such
eigenfunction

[A, B]ψi = ABψi − BAψi = (ai bi − bi ai )ψi = 0 . (6.18)


P
If the basis {ψi } is complete, any wavefunction ψ can be written as ψ = i ci ψi ,

and we have that


X X
[A, B]ψ = [A, B] ci ψi = ci [A, B]ψi = 0 . (6.19)
i i
So we have the operator equation [A, B] = 0, where 0 is the zero operator, which
maps any wavefunction to the zero function.
Conversely, if [A, B] = 0 and Aψi = ai ψi , then 0 = [A, B]ψi = A(Bψi )−ai (Bψi ),
so that Bψi is also an eigenfunction of A with eigenvalue ai . Thus B maps the
eigenspace of A with eigenvalue ai to itself. If we write E for this eigenspace, and
B|E for the operator B restricted to E, then clearly B|E is a hermitian operator on
E. Thus, by theorem 6.3, we can find a basis of E in which B acts diagonally. Since
this is true for all eigenspaces of A, we can find a complete basis of simultaneous
eigenfunctions of A and B. 

6.8.1 The canonical commutation relations



Recall that the position and momentum operators are p̂ = −i~ ∂x and x̂ = x
(multiplication by x). We can work out their commutator by considering their
action on a general wavefunction ψ:

∂ψ
x̂p̂ψ = −ix~
∂x

p̂x̂ψ = −i~ (xψ)
∂x
∂ψ
= −i~ψ − ix~
∂x
so [x̂, p̂]ψ = x̂p̂ψ − p̂x̂ψ = i~ψ

60
and as this is true for all ψ we have

[x̂, p̂] = i~ . (6.20)

It can be shown (though not in this course: see Part II Principles of Quantum
Mechanics) that these commutation relations essentially characterise the position
and momentum operators. That is, any pair of operators satisfying these relations
is equivalent (in a sense that can be made precise) to position and momentum.

6.9 Heisenberg’s Uncertainty Principle

We define the uncertainty ∆ψ A in a measurement of A on the state ψ by

(∆ψ A)2 = h(A − hAiψ )2 iψ


= hA2 iψ − (hAiψ )2 . (6.21)

Note that Theorem 6.1 implies that the expectation value and the uncertainty
are always real, as we would expect if they are physically meaningful.
Exercise Verify that (∆ψ A)2 is the statistical variance of the probability
distribution for the possible outcomes of the measurement of A on ψ, and ∆ψ A is
the distribution’s standard deviation.
Lemma 6.5. The uncertainty ∆ψ A ≥ 0 and ∆ψ A = 0 if and only if ψ is an
eigenfunction of A.

Proof. We can write

(∆ψ A)2 = h(A − hAiψ )2 iψ


= ((A − hAiψ )ψ, (A − hAiψ )ψ) .
Write φ = (A − hAiψ )ψ. We have (φ, φ) ≥ 0, with equality only if φ = 0, i.e.

Aψ = hAiψ ψ , (6.22)

which implies that ψ is an eigenfunction of A. Conversely, if ψ is an eigenfunction


of A then (ψ , Aψ) = hAiψ , and so Aψ = hAiψ ψ. In other words A has eigenvalue
hAiψ , and so ∆ψ A = 0. 
Theorem 6.6 (Schwarz’s inequality). If φ, ψ are any two normalisable wave-
functions, then |(φ, ψ)|2 ≤ (φ, φ)(ψ, ψ). We have equality if and only if φ = aψ for
some complex number a.
(ψ,φ)
Proof. For any a, 0 ≤ (φ − aψ, φ − aψ). If we take a = (ψ,ψ) , we see

|(ψ, φ)|2 |(ψ, φ)|2 |(ψ, φ)|2


0≤ (φ, φ) − − +
(ψ, ψ) (ψ, ψ) (ψ, ψ)
|(ψ, φ)|2
= (φ, φ) − .
(ψ, ψ)

Multiplying through by (ψ, ψ) we obtain the inequality. We have equality only if


φ − aψ = 0, for the value of a specified above. 

61
Theorem 6.7 (the generalised uncertainty relations). If A and B are any two
observables, and ψ is any state, then
1
∆ψ A∆ψ B ≥ |(ψ, [A, B]ψ)| . (6.23)
2
Proof. We have

(∆ψ A)2 = h(A − hAiψ )2 iψ


= ((A − hAiψ )ψ, (A − hAiψ )ψ)
2
(∆ψ B) = ((B − hBiψ )ψ, (B − hBiψ )ψ) .

So, writing A0 = A − hAiψ , B 0 = B − hBiψ , we have

(∆ψ A)2 (∆ψ B)2 = (A0 ψ, A0 ψ)(B 0 ψ, B 0 ψ)


≥ |(A0 ψ, B 0 ψ)|2 ( from Thm. 6.6)
(6.24)

= |(ψ, A0 B 0 ψ)|2 (as A0 is hermitian) .

Now: (1) A0 B 0 = 12 ([A0 , B 0 ] + {A0 , B 0 }), where the anti-commutator {A0 , B 0 } =


A B 0 + B 0 A0 ;
0

(2) [A0 , B 0 ] = [A, B];


(3) (ψ, {A0 , B 0 }ψ) is real, since

(ψ, {A0 , B 0 }ψ) = (({A0 , B 0 })† ψ, ψ) = (({A0 , B 0 })ψ, ψ) = (ψ, {A0 , B 0 }ψ)∗ .
(6.25)
A similar argument shows that (ψ, [A0 , B 0 ]ψ) is imaginary.
So
1
|(ψ, A0 B 0 ψ)|2 = (|(ψ, {A0 , B 0 }ψ)|2 + |(ψ, [A0 , B 0 ]ψ)|2 ) . (6.26)
4
Combining (6.24) and (6.26), we have that
1
(∆ψ A)2 (∆ψ B)2 ≥ |(ψ, [A, B]ψ)|2 . (6.27)
4
Taking the square root gives (6.23). 

Corollary 6.7.1. (the Heisenberg uncertainty principle for position and momen-
tum).
1
(∆ψ x)(∆ψ p) ≥ ~ (6.28)
2
Proof. Taking A = x̂ and B = p̂, we have [A, B] = i~, and the result follows from
Thm. 6.23. 
Thus, the smaller the uncertainty in position, ∆ψ x, the greater the minimum
possible uncertainty in momentum, ∆ψ p, and vice versa.

62
Lemma 6.8. If
x̂ψ = iap̂ψ (6.29)
for some real parameter a, then (∆ψ x)(∆ψ p) = 12 ~.

Proof. If x̂ψ = iap̂ψ then we have

(ψ, {x̂, p̂}ψ) = (ψ, x̂p̂ψ) + (ψ, p̂x̂ψ)


= (x̂ψ, p̂ψ) + (p̂ψ, x̂ψ)
= (iap̂ψ, p̂ψ) + (p̂ψ, iap̂ψ)
= (−ia + ia)(p̂ψ, p̂ψ) = 0 ,

which is the condition for the first term on the RHS of (6.26) to vanish. We also
have that hx̂iψ = iahp̂iψ and, since both expectations are real, this implies that
hx̂iψ = hp̂iψ = 0. Hence

(x̂ − hx̂iψ )ψ = ia(p̂ − hp̂iψ )ψ ,

which means we have equality in the Schwarz’s inequality (6.24) used to derive
(6.28). 
Lemma 6.9. The condition (6.29) holds if and only if ψ(x) = C exp(−bx2 ) for
some constants b, C.

Proof. If x̂ψ = iap̂ψ for some real a, we have that xψ = a~ ∂x ψ and so ψ(x) =
2 1
C exp(−bx ) for some real b = − 2a~ , and because we have equality in (6.28) we know
the uncertainty is minimised. Conversely, any wavefunction of the form ψ(x) =
C exp(−bx2 ) satisfies x̂ψ = iap̂ψ for some real a. 

Note: For the wavefunction to be normalisable, we require b > 0 and C 6= 0.


We can take C = |C| > 0 by multiplying ψ(x) by a phase factor (which does not
alter any physical quantity: the probabilities of outcomes for any measurement are
unaffected).
Exercise: Show that the condition (6.29) for minimum uncertainty is nec-
essary as well as sufficient. Hence show that the normalisable minimum uncertainty
states are precisely the wavefunctions defined by Gaussian functions.

6.9.1 What does the uncertainty principle tell us?


The uncertainty principle is a mathematical statement relating the uncertainties of
x and p (or more generally A and B), which are quantities defined for a given state
ψ. We can say, for example, that for any state ψ with ∆ψ x = δ, we have ∆ψ p ≥ 2δ ~
.
Heisenberg originally suggested that the uncertainty principle can be understood
simply by observing that a measurement of A creates uncertainty by disturbing the
value of any observable B that does not commute with A. This is not a valid
argument! There are two problems with it.
First, according to quantum mechanics, it is not generally the case that there is a
definite fixed pre-measurement value of either A or B, which is somehow disturbed
in the process of our attempting to measure it. Unless the wavefunction is an
eigenfunction of A or B, the result of any measurement is indeterminate.
Second, the mathematical derivation of the uncertainty principle does not re-
quire us to consider measurements of A or B actually taking place. The quantity
∆ψ A is mathematically defined whether or not we choose to carry out a measure-
ment of A. Even if we choose to interpret it as referring to a possible measurement,
we cannot interpret the mathematics as referring to successive measurements of A

63
and B. (∆ψ A) and (∆ψ B) are the standard deviations for the outcomes of mea-
surements of A and B, but these hypothetical measurements are alternative possible
measurements on the same state ψ, not actual measurements carried out one after
the other. If we measured, say, first A and then B, the first measurement would
collapse the wavefunction onto an eigenfunction of A, and the second measurement
would hence not generally be a measurement on the original state ψ.

6.10 Ehrenfest’s theorem

Theorem 6.10 (Ehrenfest’s theorem). The expectation value hAiψ of an op-


erator A in the state ψ evolves by
d i ∂A
hAiψ = h[Ĥ, A]iψ + h iψ . (6.30)
dt ~ ∂t

Proof. We have
d ∞ ∗
Z
dhAiψ
= ψ Aψdx
dt dt −∞
Z ∞
∂ψ ∗ ∂A ∂ψ
= ( Aψ + ψ ∗ ψ + ψ ∗ A )dx
−∞ ∂t ∂t ∂t
Z ∞
∂A i
= h iψ + ((Ĥψ)∗ Aψ − ψ ∗ A(Ĥψ))dx
∂t ~ −∞
i ∞ ∗
Z
∂A
= h iψ + (ψ ĤAψ − ψ ∗ A(Ĥψ))dx
∂t ~ −∞
i ∂A
= h[Ĥ, A]iψ + h iψ . (6.31)
~ ∂t


6.10.1 Applications of Ehrenfest’s theorem


p̂2
For Ĥ = 2m + V (x), we have

[Ĥ, p̂] = [V (x), p̂]



= [V (x), −i~ ]
∂x
dV
= i~ (6.32)
dx
p̂2
[Ĥ, x̂] = [ , x̂]
2m
1 −i~
= 2[p̂, x̂]p̂ = p̂ (6.33)
2m m
[Ĥ, Ĥ] = 0 . (6.34)

64
∂ Ĥ
Since none of these operators is explicitly time-dependent, we have that ∂t =
∂ x̂ ∂ p̂
∂t = ∂t = 0 and so the h ∂A
∂t iψ term on the RHS of (6.30) vanishes in each case,

giving

d dV
hp̂iψ = −h iψ ,
dt dx
d 1
hx̂iψ = hp̂iψ ,
dt m
d
hĤiψ = 0 . (6.35)
dt
d 1
These are quantum versions of the classical laws dt x= m p (which follows from
d dV d
p = mv), dt p = − dx (which follows from F = ma), and dt E = 0 (conservation of
total energy).

* Ehrenfest’s theorem and the harmonic oscillator. *

65
6.11 *The harmonic oscillator revisited

By considering commutation relations, we can give a much nicer and more illumi-
nating derivation of the energy spectrum of the harmonic oscillator. This derivation
forms part of the material for the Part II Principles of Quantum Mechanics course.
It is non-examinable material, in the sense that if you are asked to derive the energy
spectrum (without any method being stipulated) then the derivation given earlier is
a perfectly adequate answer. However, the derivation below is simpler and slicker,
and of course it also may be used in this context.
Recall that the harmonic oscillator hamiltonian is
p̂2 1
Ĥ = + mω 2 x̂2
2m 2
1 iω
= (p̂ + imωx̂)(p̂ − imωx̂) + [p̂, x̂]
2m 2
1 ~ω
= (p̂ + imωx̂)(p̂ − imωx̂) + (6.36)
2m 2

Define the operator a = √ 1 (p̂ − imωx̂). Since p̂ and x̂ are hermitian, we have
2m
a† = √ 1 (p̂
2m
+ imωx̂), and

1
Ĥ = a† a + ~ω . (6.37)
2

We have the following commutation relations:


1
[a, a† ] = (−imω)2[x̂, p̂] = ~ω , (6.38)
2m
[Ĥ, a] = [a† a, a] = [a† , a]a = −~ωa , (6.39)
† † † † † †
[Ĥ, a ] = [a a, a ] = a [a, a ] = ~ωa . (6.40)

Suppose now that ψ is a harmonic oscillator eigenfunction of energy E:

Ĥψ = Eψ .

We then have

Ĥaψ = [Ĥ, a]ψ + aĤψ = (E − ~ω)aψ (6.41)

Ĥa† ψ = [Ĥ, a† ]ψ + a† Ĥψ = (E + ~ω)a† ψ , (6.42)

so that aψ and a† ψ are eigenfunctions of energy (E − ~ω) and (E + ~ω).


We can use this to prove by induction that an ψ and (a† )n ψ are eigenfunctions
of energy (E − n~ω) and (E + n~ω). For example,

Ĥan ψ = Ĥa(an−1 ψ) = (En−1 − ~ω)an ψ , (6.43)

where Er is the energy eigenvalue of ar ψ. Since E0 = E, it follows by induction


that En = (E − n~ω).

66
In particular, if it were true that an ψ 6= 0 for all n, there would be eigenfunctions
of arbitrarily low energy, and so there would be no ground state.
However, given any physical wavefunction ψ, we have that
Z ∞
−~2 d2 ψ 1
hĤiψ = ψ∗ ( + mω 2 x2 ψ)dx
−∞ 2m dx2 2
≥ 0,

since both terms are non-negative. (Important note: this argument can obviously
be generalised to show that, if we have any potential V such that V (x) ≥ 0 for all
x, then hĤiψ ≥ 0 for all states ψ.)

So there cannot be negative energy eigenfunctions. Thus there must be a lowest


energy (i.e. ground state) eigenfunction ψ0 such that
1
0 = aψ0 = √ (p̂ − imωx̂)ψ0 , (6.44)
2m
which implies
dψ0
−i~ = imωxψ0 (6.45)
dx
and hence
mωx2
ψ0 (x) = C exp(− ), (6.46)
2~
which is indeed the ground state wavefunction we previously obtained. (This time,
though, we derived it much more simply, by introducing the operators a, a† and
considering their commutation relations.)

Since Ĥ = a† a+ ~ω
2 and aψ0 = 0, we have Ĥψ0 =

2 ψ0 , giving us the previously

obtained value of 2 for the ground state energy. We have also obtained a closed
form expression (6.46) for the ground state and hence for the excited states,

1 mωx2
(a† )n ψ0 = C( √ (p̂ + imωx̂))n exp(− ), (6.47)
2m 2~
and we see immediately that their energies are (n + 21 )~ω, as previously obtained
by a less direct argument.
We can similarly show that there cannot be eigenfunctions ψ 0 with energies
taking values other than (n + 12 )~ω). If there were, then am ψ 0 cannot vanish for
any m, since ψ0 is the unique wavefunction annihilated by a. So there would be
negative energy eigenfunctions, which contradicts the result shown above.
With a little more thought, we can similarly also show that the eigenspaces
must all be non-degenerate: i.e. there is (up to scalar multiplication) only one
eigenfunction of each energy.
The derivation of the harmonic oscillator spectrum in this subsection illustrates
an important general feature: symmetries or regularities in a quantum mechanical
spectrum (such as the regular spacing of the harmonic oscillator energy levels)
suggest the existence of a set of operators whose commutation relations define the
symmetry or explain the regularity (in this case, the operators Ĥ, a and a† ). *

67
7 The 3D Schrödinger equation
7.1 Quantum mechanics in three dimensions

We can develop quantum mechanics in three dimensions following the analogy with
classical mechanics that we used to obtain the 1D Schrödinger equation. The clas-
sical state of a single particle is described by six dynamical variables: its position
P3 P3
x = (x1 , x2 , x3 ) = i=1 xi ei and momentum p = (p1 , p2 , p3 ) = i=1 pi ei , where ei
are the standard orthonormal basis vectors. The particle’s energy is
p.p
H= + V (x) . (7.1)
2M
(Throughout this section, we use M to denote mass, to avoid confusion with the
L3 angular momentum eigenvalue denoted by m which we introduce below.)

Proceeding by analogy with the 1D case, we can introduce operators


x̂i = xi ( i.e. multiplication by xi ) ,

pˆi = −i~ ,
∂xi
or in vector form x̂ = x (7.2)
p̂ = −i~∇ (7.3)
p̂.p̂
and Ĥ = + V (x)
2M
~2 2
=− ∇ + V (x) . (7.4)
2M

Here we define a vector operator A to be a triple of operators (A1 , A2 , A3 ) such


that Aψ(x) = (A1 ψ(x), A2 ψ(x), A3 ψ(x)) is a vector for all wavefunctions ψ(x) and
all x.
We also take the wavefunction ψ to depend on 3 space and 1 time coordinates:
ψ ≡ ψ(x, t). The 3D normalisation condition is
Z
|ψ(x, t)|2 d3 x = 1 . (7.5)

Following the 1D analogy gives us the 3D time-dependent Schrödinger equation



Ĥψ(x, t) = i~ ψ(x, t) , (7.6)
∂t
or more explicitly, for a time-independent potential V ,

~2 2 ∂
− ∇ ψ(x, t) + V (x)ψ(x, t) = i~ ψ(x, t) . (7.7)
2M ∂t
Using the method of separation of variables, as before, we can derive the 3D time-
independent Schrödinger equation
~2 2
− ∇ ψ(x) + V (x)ψ(x) = Eψ(x) . (7.8)
2M

68
We can define the probability density and current

ρ(x, t) = |ψ(x, t)|2 , (7.9)


−i~ ∗
J(x, t) = (ψ (x, t)∇ψ(x, t) − (∇ψ ∗ (x, t))ψ(x, t)) ,
2m
(7.10)

and as in the 1D case (cf (3.17)) we can show that they obey a conservation equation

∂ρ
+ ∇.J = 0 . (7.11)
∂t

Notice that the 3D Schrödinger equation, like the 1D SE, is linear and the
superposition principle thus applies to its solutions: there is a physical solution
corresponding to any linear combination of two (or more) physical solutions.
The Born rule naturally extends to the 3D case: the probability of finding a
particle in a small 3D volume V which contains a point x0 is
Z
|ψ(x, t)|2 d3 x ≈ V |ψ(x0 , t)|2 (7.12)
V
= V ρ(x0 , t) .

We define the inner product of 3D wavefunctions by


Z
(ψ1 , ψ2 ) = ψ1∗ (x)ψ2 (x)d3 x .

The definition of the expectation value of an observable  in terms of the cor-


responding hermitian operator A also naturally extends to 3D wavefunctions:
Z
hÂiψ = ψ ∗ (x, t)Aψ(x, t)d3 x = (ψ, Aψ) . (7.13)

We can thus define the uncertainty ∆ψ A as in (6.21), using the definition (7.13) for
expectation values.
The proofs of theorems 6.1 and 6.2 apply to hermitian operators on 3D wave-
functions just as to hermitian operators on 1D wavefunctions.
Theorem 6.3 also extends to hermitian operators on 3D wavefunctions: the dis-
crete and continuous sets of eigenfunctions of a hermitian operator form a complete
orthogonal basis of th normalisable complex-valued functions ψ(x) of 3D vectors x.

7.2 Spherically symmetric potentials


The 3D time-independent Schrödinger equation simplifies considerably when
the potential V (x) is spherically symmetric about the origin.20 We call this a
central potential.
20 Of course, if V is spherically symmetric about any given point, we can redefine our coordinates

by translation to make that point the origin.

69
It is convenient to use spherical polar coordinates

x1 = r sin θ cos φ , x2 = r sin θ sin φ , x3 = r cos θ , (7.14)

which have ranges 0 ≤ r < ∞, 0 ≤ φ < 2π, 0 ≤ θ ≤ π.

In spherical polars, a central potential V (x) ≡ V (r, θ, φ) depends only on r = |x|.


∂2 ∂2 ∂2
Recall that in Cartesian coordinates we have ∇2 = ∂x 2 + ∂x2 + ∂x2 . In spherical
1 2 3
polars this becomes21
1 ∂ 2 ∂ 1 ∂ ∂ 1 ∂2
∇2 = 2
(r )+ 2 (sin θ ) + 2 2 (7.15)
r ∂r ∂r r sin θ ∂θ ∂θ r sin θ ∂φ2

For spherically symmetric stationary states ψ(r), the time-independent SE sim-


∂ψ ∂ψ
plifies. We have ∂θ = ∂φ = 0 and so

1 d 2d
∇2 ψ = (r )ψ
r2 dr dr
d2 ψ 2 dψ
= 2 +
dr r dr
2
1 d
= (rψ) . (7.16)
r dr2
so we have
~2 1 d2
− (rψ(r)) + V (r)ψ(r) = Eψ(r) , (7.17)
2M r dr2
which we can rewrite as
~2 d 2
− (rψ(r)) + V (r)(rψ(r)) = E(rψ(r)) . (7.18)
2M dr2

Notice that (7.18) is the 1D time-independent SE for φ(r) = rψ(r), on the


interval 0 ≤ r < ∞.
We require φ(r) → 0 as r → 0, otherwise ψ(r) ≈ O(1/r) as r → 0 and so is
singular at r = 0. It can be shown that this means that the 3D Schrödinger equation
(7.7) fails to hold there.22

Any solution to (7.18) with φ(r) → 0 as r → 0 can be extended to an odd parity


dφ̃
solution φ̃(r) of the 1D SE in −∞ < r < ∞ of the same energy, with φ̃(r) and dr

continuously defined at r = 0, by defining



φ(r) r ≥ 0,
φ̃(r) = (7.19)
−φ(−r) r < 0 .

Conversely, any odd parity solution of the 1D SE for −∞ < r < ∞ defines a
solution to (7.18) with φ(r) → 0 as r → 0 and dφdr finite at r = 0. Provided that
21 See IA Vector Calculus, or e.g. Collinson “Introductory Vector Analysis”, Chap. 12.
22 * See Dirac, “The Principles of Quantum Mechanics” (4th edition), Chap. VI for a full
discussion. *

70
V (r) is finite and continuous at r = 0, these are the correct continuity conditions
for solutions of the 3D SE: they imply that ψ and ψ 0 are continuous at the origin.
Solving (7.18) thus becomes equivalent to finding odd parity solutions to the 1D
SE for −∞ < r < ∞.

Comments We will show later (see Thm. 7.2) that the ground state (the
lowest energy bound state, if there is one) of a 3D quantum system with spherically
symmetric potential is itself spherically symmetric. (Cf. the 1D result that the
ground state of a symmetric potential always has even parity.) Hence we can always
use the method above to obtain the ground state.
One might wonder whether there might not exist even parity solutions φ+ (r) of
the 1D SE with the property that φ+ (0) = dφ +
dr (0) = 0, which would also define
solutions to (7.18) for 0 ≤ r < ∞ with the appropriate properties. The following
lemma rules this out.
Lemma 7.1. There are no even parity solutions φ+ (r) of the 1D SE with the
property that φ+ (0) = dφ +
dr (0) = 0.
Proof. If such a solution φ+ were to exist, we could define a continuous odd parity
solution φ− (r) by 
φ+ (r) r ≥ 0,
φ− (r) = (7.20)
−φ+ (−r) r < 0 .
Then, by the superposition principle, φ(r) = φ+ (r) − φ− (r) would also be a
solution. But we have φ(r) = 0 for r > 0, so that all derivatives of φ vanish for
r ≥ 0. The Schrödinger equation has no non-trivial solutions with this property:
hence φ(r) = 0 for all r. Hence φ+ (r) = φ− (r) = 0 for all r, so in particular
the hypothesised even parity solution φ+ is not a physical solution, as it vanishes
everywhere. 

7.3 Examples of spherically symmetric potentials


7.3.1 The spherical harmonic oscillator

has potential
1
V (r) = M ω 2 r2 .
2
The general method we have just given for constructing spherically symmetric
stationary states shows that its spherically symmetric stationary states are related
by (7.19) to the odd parity bound states of the 1D harmonic oscillator, and have
the same energies. Thus the lowest energy spherically symmetric stationary state –
i.e. the ground state – has energy 32 ~ω, and the higher energy (excited) spherically
symmetric states have energies (2n + 32 )~ω for positive integer n.

7.3.2 The spherical square well

has potential 
−U r < a ,
V (r) = (7.21)
0 r > a.
By the above argument, spherically symmetric stationary states correspond to odd
parity bound states of the 1D square well potential

−U |x| < a ,
V (x) = (7.22)
0 |x| > a .

71
These, if they exist, can be obtained by the graphical method used earlier to
obtain 1D square well potential bound states. In particular, one can show (cf.
section 4 above and example sheet I, question 10) that there exists an odd parity
bound state if and only if r
2M U π
2
≥ . (7.23)
~ 2a
So, if this condition is not satisfied, the 3D spherical square well does not have a
spherically symmetric stationary state: i.e. it does not have a ground state, and
thus does not have any bound states.
As this illustrates, 3D potential wells (continuous potentials with V (x) ≤ 0 for
all x, V (x) < 0 for some x, and V (x) = 0 for |x| > a, for some finite a) do not
necessarily have bound states. This is in contrast to the 1D case:
Exercise (important!): Show that all 1D potential wells have at least
one bound state.

7.4 Spherically symmetric bound states of the hydrogen atom


We model the hydrogen atom by treating the proton as infinitely massive and at
rest.23
We seek spherically symmetric bound state wavefunctions ψ(r) for the electron
e2
orbiting in a Coulomb potential V (r) = − 4π 0r
:

~2 d2 ψ 2 dψ e2
− ( 2 + )− ψ(r) = Eψ(r) , (7.24)
2M dr r dr 4π0 r

e2 M −2M E
for some E < 0. Writing a = 2π0 ~2 , b= ~ , we have

d2 ψ 2 dψ a
+ + ψ − b2 ψ = 0 . (7.25)
dr2 r dr r
If we try the ansatz

ψ(r) ≈ exp(−br) , (7.26)

we see the first and fourth terms dominate the other two for large r, and cancel one
another precisely.
P∞ This suggests trying an ansatz of the form ψ(r) = f (r) exp(−br),
with f (r) = n=0 an rn , in the hope of finding values of the coefficients an such
that the four terms cancel precisely to all orders. (Cf. our first solution to the
harmonic oscillator.)

Our previous discussion assumed that V (r) is nonsingular as r → 0. Since here


V (r) diverges as r → 0, we cannot use the same justification as previously to argue
that that φ(r) = rψ(r) → 0 as r → 0. However, we still require ψ(r) to define a
normalisable 3D wavefunction, so that
Z ∞
0< r2 |ψ(r)|2 dr < ∞ .
r=0

This implies that ψ(r) = O(r−1 ) as r → 0, i.e. that ψ can at worst have a singularity
of order r−1 at zero. We also require that ψ should correspond to a continuous
wavefunction. This excludes a singularity of order r−1 , so we require that ψ is
regular — i.e. has a finite limit — as r → 0.
23 *This can be shown to be an excellent approximation (see Part II Principles of Quantum

Mechanics): it gives the correct energy levels up to an overall constant factor (the same for each
energy level) of order 1 + (me /mp ) ≈ 1.0005.*

72
We have
d2 f 2 df 1
2
+ ( − 2b) + (a − 2b)f (r) = 0 . (7.27)
dr r dr r
Hence

X
(an n(n − 1)rn−2 + 2an nrn−2 − 2ban nrn−1 + (a − 2b)an rn−1 ) = 0 (7.28)
n=0

and taking the coefficient of rn−2 we have

an n(n − 1) + 2an n − 2ban−1 (n − 1) + (a − 2b)an−1 = 0 for n ≥ 1 . (7.29)

This gives

(2b(n − 1) − (a − 2b))
an = an−1
n(n − 1) + 2n
2bn − a
= an−1 . (7.30)
n(n + 1)
We thus have that an → 2b
n an−1 for large n. If the coefficients do not vanish for large
n, this means they have the asymptotic behaviour of the coefficients of exp(2br),
i.e. f (r) ≈ C exp(2br). This would give ψ(r) ≈ C exp(2br) exp(−br) = exp(br),
leading to an unnormalisable and thus unphysical wavefunction. So there must be
some integer N ≥ 1 for which aN = 0, and we can take N to be the smallest such
integer.

Then aN −1 6= 0, so that aN = 0 implies 2bN = a or b = a/2N , and so

~2 a2
E=−
8M N 2
M e4
=− , (7.31)
32π 2 20 ~2 N 2
which is precisely the energy spectrum of the Bohr orbits, but now derived from
quantum mechanics (though still with an assumption of spherical symmetry, which
we will need to relax to obtain the general orbital wavefunction).

From
2bn − a
a = 2bN and an = an−1 (7.32)
n(n + 1)
we obtain
n−N
an = an−1 2b . (7.33)
n(n + 1)
This gives solutions of the form

 1 N = 1,
f (r) = (1 − br) N = 2, (7.34)
(1 − 2br + 23 (br)2 ) N = 3 ,

73
and generally f (r) = L1N −1 (2br) where L1N −1 is one of the associated Laguerre
polynomials.24
The corresponding wavefunctions are ψ(r) = CL1N −1 (2br) exp(−br), where the
constant C is determined by normalisation.

7.5 Canonical Commutation Relations in 3D


From (7.2) we have

x̂i = xi (multiplication by xi ) , (7.35)



pˆi = −i~ , (7.36)
∂xi

and so, by calculating the action on a general wavefunction as before, we obtain

[x̂i , xˆj ] = 0 = [p̂i , pˆj ] (7.37)

[x̂i , pˆj ] = i~δij . (7.38)



(The prefactors i on the right hand sides are −1.)

7.6 Orbital Angular Momentum


Recall that in classical mechanics we define the angular momentum vector

L = x ∧ p, Li = ijk xj pk , (7.39)

and that L is conserved in a spherically symmetric potential V (r).


We define the quantum mechanical operators


L̂ = −i~x̂ ∧ ∇ , L̂i = −i~ijk xˆj , (7.40)
∂xk
and the total angular momentum

L̂2 = L̂.L̂ = L̂21 + L̂22 + L̂23 . (7.41)

24 *There is a simple expression for the associated Laguerre polynomials:


1 x −k dN
LkN (x) = e x (xN +k e−x ) .
N! dxN
Some plots of L0N for small N and some other information about the Laguerre polynomials (LkN
for k = 0) and the associated Laguerre polynomials can be found
at mathworld.wolfram.com/LaguerrePolynomial.html
and at mathworld.wolfram.com/AssociatedLaguerrePolynomial.html .*

74
7.6.1 Angular momentum commutation relations

∂ ∂
[L̂i , L̂j ] = −~2 ilm jnp [x̂l , xˆn ]
∂xm ∂xp
∂ ∂ ∂ ∂
= −~2 ilm jnp ([x̂l , xˆn ] + xˆn [x̂l , ])
∂xm ∂xp ∂xm ∂xp
∂ ∂ ∂ ∂
= −~2 ilm jnp (x̂l [ , xˆn ] + xˆn [x̂l , ] )
∂xm ∂xp ∂xp ∂xm
∂ ∂
= −~2 ilm jnp (x̂l δmn − xˆn δlp )
∂xp ∂xm
∂ ∂
= −~2 mil mpj x̂l + ~2 pjn pmi xˆn
∂xp ∂xm
∂ ∂
= −~2 (δip δlj − δij δlp )x̂l + ~2 (δjm δni − δji δnm )(xˆn )
∂xp ∂xm
∂ ∂ ∂ ∂
= −~2 (xˆj − δij x̂l − x̂i + δij x̂l )
∂xi ∂xl ∂xj ∂xl
= i~ijk L̂k .
=

[L̂2 , L̂i ] = [L̂j L̂j , L̂i ]


= [L̂j , L̂i ]L̂j + L̂j [L̂j , L̂i ]

= i~(jik (L̂k L̂j + L̂j L̂k ))


= 0. (7.42)

Since the L̂i do not commute, they are not simultaneously diagonalisable. However,
L̂2 and any one of the L̂i can be simultaneously diagonalised, since [L̂2 , L̂i ] = 0.

We also have

[L̂i , x̂j ] = i~ijk x̂k , (7.43)


[L̂i , p̂j ] = i~ijk p̂k , (7.44)
X
[L̂i , x̂2j ] = 2i~ijk x̂j x̂k = 0 , (7.45)
j

X
[L̂i , p̂2j ] = 2i~ijk p̂j p̂k = 0 . (7.46)
j

qP
2 P 2
Now we have that r̂ = j xˆj . We also have that [L̂i , j xˆj ] = 0. One can
show directly from this (see part II Principles of Quantum Mechanics), or check
by calculation (Exercise), that [L̂i , r̂] = 0. More generally, one can show that
[L̂i , V (r)] = 0 for any spherically symmetric potential V (r). We also have that

p̂.p̂ 1 X 2
[L̂i , ] = [L̂i , p ] = 0. (7.47)
2m 2m j j

75
So, for any spherically symmetric potential V (r), we have that

~2 2
[L̂i , Ĥ] = [L̂i , − ∇ + V (r)]
2M
= 0, (7.48)

[L̂2 , Ĥ] = 0 . (7.49)

In other words, Ĥ, L̂i and L̂2 all commute with one another.

Comment This is an important and powerful result. Given any 3D quan-


tum system, we can find a basis of simultaneous eigenfunctions of Ĥ, L̂2 and L̂3 .
(The standard convention is to use L̂3 . Of course, as the operators are related by
rotational symmetry, L̂1 or L̂2 would work equally well.) As we will see, working
in such a basis greatly simplifies the solution of the time-independent SE.

7.6.2 Angular momentum operators in spherical polar coordinates


We can translate the definitions of L̂i to spherical polars. We have

x1 = r sin θ cos φ , x2 = r sin θ sin φ , x3 = r cos θ . (7.50)

Thus
∂ X ∂xi ∂
=
∂θ i
∂θ ∂xi
∂ ∂ ∂
= r cos θ cos φ + r cos θ sin φ − r sin θ , (7.51)
∂x1 ∂x2 ∂x3
∂ X ∂xi ∂
=
∂φ i
∂φ ∂xi
∂ ∂
= −r sin θ sin φ + r sin θ cos φ . (7.52)
∂x1 ∂x2

We thus obtain
∂ ∂ ∂ ∂
i~(cos φ cot θ + sin φ ) = −i~(x2 − x3 )
∂φ ∂θ ∂x3 ∂x2
= L̂1 , (7.53)
∂ ∂ ∂ ∂
i~(sin φ cot θ + cos φ ) = −i~(x3 − x1 )
∂φ ∂θ ∂x1 ∂x3
= L̂2 , (7.54)
∂ ∂ ∂
−i~ = −i~(x1 − x2 )
∂φ ∂x2 ∂x1
= L̂3 . (7.55)

Notes 1. We could obtain these identities more straightforwardly but long-


∂ ∂ ∂ ∂
windedly by first expressing ∂x i
in terms of r, θ, φ, ∂r , ∂θ , ∂φ and thus converting
the definitions of the L̂i into spherical polars.

2. Observe in particular that the L̂i are all independent of ∂r , as we expect
from the commutation relation [L̂i , r] = 0.

76
We can also obtain
X
L̂2 = L̂2i = (L̂1 + iL̂2 )(L̂1 − iL̂2 ) + i[L̂1 , L̂2 ] + L̂23
i

= (L̂1 + iL̂2 )(L̂1 − iL̂2 ) − ~L̂3 + L̂23


∂ ∂ ∂ ∂
= −~2 (cot θeiφ − ieiφ )(cot θe−iφ + ie−iφ )
∂φ ∂θ ∂φ ∂θ
∂ ∂2
+ i~2 − ~2 2
∂φ ∂φ
2
∂ ∂ 1 ∂2
= −~2 ( 2 + cot θ + )
∂θ ∂θ sin2 θ ∂φ2
1 ∂ ∂ 1 ∂2
= −~2 ( (sin θ ) + ). (7.56)
sin θ ∂θ ∂θ sin2 θ ∂φ2

Recall that [L̂2 , L̂3 ] = 0. We have


1 ∂ ∂ 1 ∂2
L̂2 = −~2 ( (sin θ ) + ),
sin θ ∂θ ∂θ sin2 θ ∂φ2

L̂3 = −i~ . (7.57)
∂φ
We can thus seek simultaneous eigenfunctions of the form Y (θ) exp(imφ), since
L̂3 exp(imφ) = ~m exp(imφ). As φ is defined modulo 2π, we need eim(φ+2π) = eimφ ,
so ei2mπ = 1 and m is an integer.
This leaves us with an eigenvalue equation for L̂2 :

2 1 d d m2
−~ ( (sin θ ) − )Y (θ) = λY (θ) . (7.58)
sin θ dθ dθ sin2 θ
From a physics perspective, the key fact about this equation is that we can show it
has non-singular solutions if and only if λ = ~2 l(l + 1) for some integers l ≥ 0 and
m such that m is in the range −l ≤ m ≤ l.
Less crucial for now are the details, although it is interesting to see them. The
solutions are called the associated Legendre functions Pl,m (θ). They can be obtained
by reducing the equation to a standard form, using the substitution w = cos θ. Since
θ is in the range 0 ≤ θ ≤ π, we have −1 ≤ w ≤ 1. We obtain the equation
d dY m2
−~2 ((1 − w2 ) ) − (λ − )Y = 0 . (7.59)
dw dw 1 − w2
For m = 0 and λ = ~2 l(l + 1) this is Legendre’s differential equation for functions
of degree l, which has solution Pl (w). For general m it’s an associated Legendre
differential equation.25
The associated Legendre functions can be obtained from the Legendre polyno-
mials Pl by
d|m|
Pl,m (θ) = (sin θ)|m| Pl (cos θ) . (7.60)
d(cos θ)|m|
We thus have the overall solution given by the spherical harmonic with total
angular momentum quantum number l and L̂3 quantum number m:

Yl,m (θ, φ) = Pl,m (θ) exp(imφ) ,


25 Details can be found in Whittaker and Watson, “A course of modern analysis”, 4th edition,

(C.U.P., 1996).

77
an eigenfunction of L̂2 and L̂3 with eigenvalues ~2 l(l + 1) and ~m respectively.
(For plots of some spherical harmonics see e.g.
mathworld.wolfram.com/SphericalHarmonic.html.)

7.7 Solving the 3D Schrödinger equation for a spherically


symmetric potential
The time-independent SE is
~2 2
− ∇ ψ + V (r)ψ = Eψ . (7.61)
2M
Recall that in spherical polar coordinates
1 ∂ 2 ∂ 1 ∂ ∂ 1 ∂2
∇2 = 2
(r )+ 2 (sin θ ) + 2 2
r ∂r ∂r r sin θ ∂θ ∂θ r sin θ ∂φ2
1 ∂ ∂ 1 ∂2
L̂2 = −~2 ( (sin θ ) + ). (7.62)
sin θ ∂θ ∂θ sin2 θ ∂φ2
So we have

2 2 ∂2 2 ∂ 2 1
−~ ∇ = −~ ( 2 + ) + 2 L̂2 . (7.63)
∂r r ∂r r
We can thus rewrite the SE as
~2 ∂ 2 2 ∂ 1
− ( + )+ L̂2 )ψ(r, θ, φ) + V (r)ψ(r, θ, φ) = Eψ(r, θ, φ) . (7.64)
2M ∂r2 r ∂r 2M r2

If we separate variables, writing ψ(r, θ, φ) = ψ(r)Yl,m (θ, φ), this gives

~2 d2 2 d ~2
− ( + )ψ(r)+( l(l+1))+V (r))ψ(r) = Eψ(r) .
2M dr2 r dr 2M r2
(7.65)
So, we have a standard 1D radial Schrödinger equation for ψ(r), with the modified
potential
~2 l(l + 1)
V (r) + .
2M r2

Comments:
• We have seen that we can find a basis of simultaneous eigenfunctions of Ĥ, L̂2
and L̂3 , with eigenvalues E, l and m respectively. Since the modified potential
depends on l but not m, if an energy eigenspace with eigenvalue E contains
any state with L̂2 eigenvalue l, it must contain states with all the associated
L̂3 eigenvalues: m = −l, −l + 1, . . . , l. This greatly simplifies the analysis of
orbital angular momentum eigenstates associated with a given energy level.
• If the angular momentum l = 0 then also m = 0, and the function Y00 (θ, φ) is
constant. Thus all zero angular momentum states are spherically symmetric.
Conversely, since the Ylm for l 6= 0 are orthogonal to Y00 , all spherically
symmetric states have zero angular momentum. This makes sense physically,
since a state ψ with hLiψ 6= 0 by definition has a nonzero vector associated
with it, which breaks spherical symmetry.

78
• *The fact that, in quantum mechanics, we can express −~2 ∇2 in terms of a
differential operator involving only r together with a term proportional to the
operator L2 is a consequence of the fact that, in classical mechanics, kinetic
energy can be expressed as a sum of terms proportional to the radial momen-
tum squared and the angular momentum squared. However, to derive the
first result from the second requires addressing one or two subtleties beyond
our scope. (A discussion can be found in Dirac’s book “The Principles of
Quantum Mechanics”, 4th edition, Section 38.)*
• *We can understand qualitatively why we should expect the radial SE to de-
pend on the angular momentum l, by noting that the “extra potential energy”
2
term l(l+1)~
2M r 2 corresponds to the potential needed to produce the centripetal
L2
force mr 3 which would keep a classical particle of angular momentum L in a
p
circular orbit, if we set L = ~ l(l + 1).*

7.7.1 Degeneracies
As noted above, the values of E for which this equation is solvable clearly may
depend on l but not on m. As there are (2l + 1) possible values of m = −l, −l +
1, . . . , l, each energy level would have degeneracy (2l + 1), assuming there are no
further degeneracies.

The ground state We can now prove a result we stated earlier.


Theorem 7.2. The ground state (i.e. lowest energy bound state) solution of the
3D Schrödinger equation for a spherically symmetric potential must have l = m = 0
and is thus spherically symmetric.

Proof. The proof is by contradiction. Suppose that ψ(r, θ, φ) = ψ(r)Ylm (θ, φ), for
some l > 0, is the lowest energy solution and has energy E. We have that

~2 d2 2 d ~2
− ( + )ψ(r)+( l(l+1))+V (r))ψ(r) = Eψ(r) .
2M dr2 r dr 2M r2
(7.66)

Now as Ĥ, L̂2 and L̂3 are commuting hermitian operators, the space of wave-
functions is spanned by their simultaneous eigenstates. In particular, the space
of zero angular momentum wavefunctions is spanned by orthonormal eigenstates
ψi (r, θ, φ) of Ĥ with E = Ei and l = m = 0, which have the form

ψi (r, θ, φ) = ψi (r)Y00 (θ, φ) = ψi (r) . (7.67)

(In other words, the eigenstates ψi are all spherically symmetricPsolutions.) We


can thus write ψ(r) = i ci ψi (r) for some constants ci such that i |ci |2 = 1.
P

79

~2 d2
Z
2 d
E= ψ ∗ (r)(− ( 2+ )ψ(r)
r=0 2M dr r dr
~2
+( l(l + 1)) + V (r))ψ(r)) (7.68)
2M r2
Z ∞
~2 d2 2 d
= ψ ∗ (r)(− ( 2+ + V (r)))ψ(r)
r=0 2M dr r dr
Z ∞
~2
+ ψ ∗ (r)( l(l + 1))ψ(r) (7.69)
r=0 2M r2
Z ∞
~2 d2 2 d
> ψ ∗ (r)(− ( 2+ + V (r)))ψ(r)
r=0 2M dr r dr
Z ∞X
~2 d2 2 d X
= ci ψi∗ (r)(− ( 2+ + V (r))) cj ψj (r) . (7.70)
r=0 i 2M dr r dr j

Now this last term is


X X
c∗i ci Ei = |ci |2 Ei . (7.71)
i i

Since we have that E > i |ci |2 Ei and that i |ci |2 = 1, we must have that
P P
E > Ei for at least one value of i. Hence E is not the lowest energy eigenvalue, in
contradiction to our original assumption. 

7.8 The Hydrogen atom


We can now obtain the general bound state solution for particles in the potential
e2
V (r) = − 4π 0r
.
As we did in √obtaining spherically symmetric solutions, we define the quantities
e2 M −2M E
a = 2π 0~
2, b = ~ . We obtain from Eqn. (7.65) the equation

d2 2 d 1 a
( 2
+ )ψ(r) + (− 2 l(l + 1)) + )ψ(r) = b2 ψ(r) . (7.72)
dr r dr r r
As we saw in discussing Eqn. (7.26), we see that the ansatz ψ(r) ≈ exp(−br) means
that the two terms which are largest asymptotically (the first term on the LHS and
the term on the RHS) cancel. This again suggests trying an ansatz of the form
ψ(r) = f (r) exp(−br), for a power series f (r).

However, the new singular term ( r12 l(l + 1)) means that the previously obtained
solutions are not generally valid.
P∞It turns out to be convenient to write the power series in the form f (r) =
n+σ
a
n=0 n r , where σ, a constant to be determined, is chosen so that a0 6= 0: i.e.
the power series begins with a term proportional to rσ .
Considering the coefficient of rσ−2 , we have −(σ(σ − 1) + 2σ) + l(l + 1) = 0 or
σ(σ + 1) = l(l + 1), a quadratic equation with roots σ = l and σ = −(l + 1). As
l ≥ 0, we choose σ = l to avoid a divergence at r = 0.

We now have
(n + l)2b − a
an = an−1 for n ≥ 1 . (7.73)
n(n + 2l + 1)
80
2b
As before, if the power series does not terminate this reduces to an ≈ n an−1 for
large n, which would give us f (r) ≈ exp(2br) and
ψ(r) ≈ exp(2br) exp(−br) ≈ exp(br) ,
a divergent and unnormalisable wavefunction, which is physically unacceptable.

The power series must thus terminate, so we have a = 2b(n + l), for some n ≥ 1.
a
Rewriting, we have b = 2N for some N ≥ l + 1, giving the same overall set of
solutions for b, and thus the same energy levels (i.e. the Bohr energy levels), as the
spherically symmetric case with l = 0 we considered earlier:
M e4 1
E=− 2 .
32π 0 ~ N 2
2 2

7.8.1 Energy level degeneracies


Each value of N is consistent with
l = 0, 1, . . . (N − 1) ; (7.74)
each value of l is consistent with
m = −l, −(l − 1), . . . , l . (7.75)

(The first of these degeneracies occurs only for a Coulomb force law; the second, as
we have seen, holds for any central potential.) The total number of values of (m, l)
consistent with N is thus
N −1 l N −1
X X X 1
1= (2l + 1) = 2( N (N − 1)) + N = N 2 . (7.76)
2
l=0 m=−l l=0

In fact, the true degeneracy of the N th energy level of the hydrogen atom in
a full non-relativistic quantum mechanical treatment is 2N 2 : the extra factor of 2
arises from an intrinsically quantum mechanical degree of freedom, the electron spin,
which has no direct classical analogue. (This is covered in the Part II Principles of
Quantum Mechanics course.)

7.9 Towards the periodic table


We could try to generalize this discussion to atoms other than hydrogen. These
have a nucleus with charge +Ze, orbited by Z independent electrons, where the
atomic number Z is an integer greater than one.
If we take the nucleus to be fixed, as we did with hydrogen, this means we need
to solve the Schrödinger equation for Z independent electrons in a central Coulomb
potential. This is not so simple, since the electrons also interact with each other.
If we ignore this temporarily, we can obtain solutions of the form
ψ(x1 , . . . , xZ ) = ψ1 (x1 ) . . . ψZ (xZ ) , (7.77)
where the ψj are rescaled solutions for the hydrogen atom (the rescaling is because
the nucleus has charge +Ze instead of +e). The energy is just the sum
Z
X
E= Ei . (7.78)
i=1

81
It turns out that for relatively small atoms this gives qualitatively the right
form, with corrections arising from the electron-electron interactions that can be
calculated using perturbation theory.26 However, we also need to allow for the
Pauli exclusion principle, which implies that no two electrons in the same atom can
be in the same state. So the lowest overall energy state is given by filling up the
energy levels in order of increasing energy, starting with the lowest. Allowing for
the twofold degeneracy arising from spin, as above, we have 2N 2 states in the N th
energy level. This gives us an atom with a full energy level with Z = 2, 10 = 8+2, . . .
for N = 1, 2, . . .; these are the elements helium, neon, . . .. The elements with outer
electrons in the 1st and 2nd energy levels fill out the corresponding first two rows of
the periodic table. The analysis gets more complicated as atoms get larger, because
electron-electron interactions become more important, and this qualitative picture
is not adequate for the third and higher rows of the periodic table.
We can understand that helium and neon are chemically inert (unreactive) as
a consequence of the fact that they have full energy levels, which turns out to be
a very stable state that does not easily undergo transitions by capturing or losing
electrons.

26 See Part II Principles of Quantum Mechanics for details.

82

You might also like