Chapter 1

Historical overview of the

developments of quantum
mechanics

1.1 Quantum Ideas Course Overview

Course synopsis: The overall purpose of this course is to introduce you all
to the core concepts that underlie quantum physics, the key experimental
and theoretical developments in the advent of quantum mechanics, the basic
mathematical formalism, and give you a flavour of current research in quan-
tum physics. Many of you will have likely been introduced to one or more of
the topics that we will be covering, but perhaps not in the same way or in
such detail. The aim is to ensure that everyone has the same introduction to
this exciting area of physics that underpins nearly all physical phenomena
that are observed.

Syllabus and resources: I have put together a course syllabus and website
located at (http://www.physics.ox.ac.uk/users/smithb/qid.html). There will
be 12 lectures over four weeks (three lectures per week). The lectures will
take place in the Martin Wood Lecture Theatre in the Clarendon labora-
tory from 9:00-10:00am on Wednesday, Thursday and Friday of Trinity Term
weeks 1, 2, 4, and 5.

Books: There are four recommended books and six supplementary books
for your reading pleasure.
Recommended Books:

QED: The strange Theory of Light and Matter, R. P. Feynman (Prince-

ton University Press, 2006)

Quantum Theory: A Very Short Introduction, J. C. Polkinghorne (Ox-

1
 ford University Press, 2002)

The New Quantum Universe, T. Hey and P. Walters (Cambridge Uni-

versity Press, 2003)

The Strange World of Quantum Mechanics, D. F. Styer (Cambridge

University Press, 2000)

Supplementary Reading:

Thirty Years That Shook Physics: Story of Quantum Theory, G. Gamow

(Dover, 1985)

Beyond Measure: Modern Physics, Philosophy and the Meaning of

Quantum Theory, J. Baggott (Oxford University Press, 2003)

Theoretical Concepts in Physics, 2nd ed., M. Longair (Cambridge Uni-

versity Press, 2003)

Modern Physics, K. Krane (Wiley, 2012)

Quantum Generations: A History of Physics in the Twentieth Cen-

tury, H. Kragh (Princeton University Press, 2002)

Feynman Lectures on Physics, vol. 3, R. P. Feynman, R. B. Leighton,

and M. Sands (Addison Wesley, 1971)

Personally, I find Polkinghornes book to be useful for the core concepts in

that it is quite compact, but covers the topics. However, for calculations, I
suggested Kranes book, which covers much more than the course.

Lecture notes: The lectures will follow closely Dr. Axel Kuhns lecture
notes, available at
http://webnix.physics.ox.ac.uk/atomphoton/index.php?dish=lectures
and I will be writing additional notes as I work through the course. My
notes will be posted on the course website as they are completed. Be sure
you refresh your web browser when looking for new notes!

Topics covered: We will break down the material into four different topics:

Particle-like properties of light: The success of classical physics, mea-

surements in classical physics. The nature of light, the ultraviolet
catastrophe, the photoelectric effect and the quantization of radiation.

Wave-like properties of matter: Interference of atomic beams, dis-

cussion of two-slit interference, Bragg diffraction of atoms, quantum
eraser experiments. Interferometry with atoms and large molecules
(de Broglie wavelength). The Uncertainty principle by considering a

2
 microscope and the momentum of photons, zero point energy, stability
and size of atoms. Atomic spectral lines and the discrete energy levels
of electrons in atoms, the Frank-Hertz experiment and the Bohr model
of an atom.

Introductory mathematical formalism: Schrodinger equation and bound-

ary conditions. Free-space, potential barrier, and tunneling solutions.
Solution for a particle in an infinite potential well, to obtain discrete
energy levels and wave functions. Superposition of energy-eigenstates
and time evolution. Measurements in quantum physics: Expectation
values. Dirac notation.

Modern quantum physics: Measurements in quantum physics. Schrodingers

cat and the many-worlds interpretation of quantum mechanics. Mag-
netic dipoles in homogeneous and inhomogeneous magnetic fields and
the Stern-Gerlach experiment showing the quantization of the mag-
netic moment. The impossibility of measuring two orthogonal compo-
nents of magnetic moments. A glimpse of quantum engineering and
quantum computing: The EPR paradox, entanglement, hidden vari-
ables, non-locality and Aspect experiment, quantum cryptography and
the BB84 protocol.

Lecture Plan: See the course syllabus for the lecture plan.

1.2 Historical development of quantum physics

1.2.1 Introduction
Scientific enquiry has a long and interesting history. Being a scientist, I
cannot help but find the story of how this discipline of human activity has
come to be. However, aside from being of purely personal interest I find
that it can often be extremely insightful and helpful to understand how cer-
tain concepts were developed by the pioneers of science. By understanding
the approaches that were taken to conceive, refine, and develop a theory, I
believe we can better our own approach to perform research today. For ex-
ample, knowledge of how quantum physics was developed at the beginning
of the 1900s, including the dead ends that did not lead to success, has helped
me formulate the theory of photon wave mechanics [?]. Now just because
I find the history of a subject to be interesting and useful for research or
insight does not imply that a historical perspective is the most suitable for
teaching and learning a subject for the first time. This is, in my opinion,
the case for quantum mechanics. Thus, I will only outline the key historical
events that led to the development of quantum physics. The goal of this
is to understand how this subject arose from the interplay of increasingly

3
accurate experimental observation and theoretical developments to explain
newly observed phenomena.

1.2.2 Summary and success of classical physics

The three pillars of classical physics: mechanics, thermodynamics, and elec-
tromagnetism, were well understood at the end of the 1800s. The backbone
of classical physics can be summarized by the following basic sub-disciplines

Mechanics: PBased upon Newtons laws, generalized by Lagrange and

Hamilton: F = ma (Works well enough to send man to the moon
and describe micron size objects move).

(a) Isaac Newton (b) William R. Hamilton (c) Joseph Louis Lagrange

Electromagnetism: A crowning unification of three previously inde-

pendent concepts (electricity, magnetism and optics) largely due to
the work of Oersted, Faraday and Maxwell. Light is an electromag-
netic wave.

(a) Hans Christian rsted (b) Michael Faraday (c) James Clerk Maxwell

Thermodynamics: Energy conservation and equilibriation as set out

by Carnot, Clausius, and Lord Kelvin.

4
 (a) Nicolas Leonard Sadi (b) Rudolf Clausius (c) William Thomson
Carnot

Classical physics has two key foundational differences from quantum theory
(as we will see):

Measurement without disturbance: In classical physics a core assump-

tion is that a measurement on a system could reveal information without
any disturbance to the system. All that is required to do this is to turn down
the interaction strength of the measurement. For example, if we wanted to
determine the position and velocity of a particle all we must do is use a weak
light source turned on and off quickly to take two snapshots of the particle.
The first snapshot gives the initial position xi , and the second, taken at a
time t later, gives its final position xf . We can thus determine its velocity
as v = (xf xi )/t. According to classical physics one can measure pre-
cisely both the position and velocity to arbitrary accuracy. As we will see
this contradicts what is observed and predicted for quantum systems. An-
other perhaps philosophical aside associated with this classical assumption
is that the position and velocity are well defined for a system, simply await-
ing our observation. In quantum theory, these interrelated ideas do not hold.

Determinism: Classical physics is based upon deterministic differential

equations. If one knows precisely the initial values of the position and ve-
locity of all particles in a system at one moment in time, then all future
behaviour of the system can be predicted. The universe was likened to a
clock, which when set about running deterministically evolves for all time.
This concept is nicely summarized by Laplaces demon

We may regard the present state of the universe as the effect

of its past and the cause of its future. An intellect which at a
certain moment would know all forces that set nature in motion,
and all positions of all items of which nature is composed, if this

5
 intellect were also vast enough to submit these data to analysis, it
would embrace in a single formula the movements of the greatest
bodies of the universe and those of the tiniest atom; for such an
intellect nothing would be uncertain and the future just like the
past would be present before its eyes.
- Pierre Simon Laplace, A Philosophical Essay on Probabili-
ties, (1814)

Figure 1.1: Pierre Simon Laplace

We see that uncertainty in classical physics was only due to our own igno-
rance of the position and velocity of all particles at a given instant in time.
The universe was simply one massive initial-value problem and if we knew
the initial conditions at one instant in time, we could know the outcome of
everything in the future, as well as look back in the past and know all that
has been!

1.2.3 Cracks in the foundations

The foundations of classical physics started to show cracks well before the
advent of quantum theory. In the 1880s, Michelson and Morley performed
a series of experiments that demonstrated the earth does not travel through
the luminiferous ether that was thought to permeate the universe. Recall
that if light were a wave, then it must be an excitation of some medium and
as the earth travelled through the ether, it should have a different velocity
depending on the relative velocity of the earth and ether. No difference was
observed, and the resulting theory that describes this is none other than
Einsteins special theory of relativity.

It is a common misconception that physicists of the 1890s believed that

the fundamental laws of nature had been discovered and all that was left to
do was work out the consequences of these. In fact, this was a tumultuous
period in which several fundamental unresolved problems occupied some of

6
the greatest minds of the time. For example, the kinetic theory of gases and
the equipartition theorem put forth by Clausius, Maxwell, and Boltzmann
was not widely accepted. The atomic and molecular theory of the structure
of matter came under attack, as the origin of the resonances in molecules,
which were assumed to be the origin of spectral lines, was unknown. You
must recall that at that time there was no direct evidence for fundamental
particles until 1897, when J. J. Thomson discovered the electron.

These unexplained phenomena came about because of increasingly pre-

cise experimental results, which had slipped past previous generations be-
cause the observations of earlier times were not as precise. This does not
limit the applicability of classical physics, which works quite well in its realm
of applicability. For example, the new physics that arose at the outset of the
1900s (relativity and quantum physics) are not required for a transatlantic
flight or phone call (at least a phone call when I was born).1 There were
several observations that led to the development of quantum theory, which
I briefly describe below.

Spectral lines: Starting with a Swiss schoolmaster by the name of Johann

Balmer in 1885, who was trying to understand the spectral lines observed
in emission from hydrogen.

Figure 1.2: Johann Jakob Balmer

He noticed that there were regularities in the wavelengths of the emitted

1
It is useful to note here the relationship between increased measurement precision,
technological advance, and discovery of new physical phenomena. When an experimen-
talist develops a method to measure some quantity with significantly increased precision
beyond what was capable before, this new capability enables one to measure and manipu-
late systems on a more precise scale. This allows for the development of smaller, faster and
more complex technologies that we use in our daily lives. Further more, these technolo-
gies can be used to probe new areas of physics previously unexplored. New discoveries
from these experiments can then feed back into the development of new measurement
techniques and technologies.

7
lines and found that he could determine the wavelengths with the following
formula
1 1
 
1
= 0 , (1.1)
4 n2
where n is an integer greater than two, and 0 is a constant length of 364.56
nm.
In 1900, Swedish physicist Johannes Rydberg refined the expression for
hydrogen emission to describe all observed lines, which gave
 1  1
R 1 1
= 2 , (1.2)
hc m n

Figure 1.3: Johannes Robert Rydberg

where m and n are integers (m < n), R is known as the Rydberg constant
(R = 13.6 eV), h is Plancks constant (6.626 1034 Js) and c is the speed
of light in vacuum. Although a concise formula for predicting the emission
wavelengths for hydrogen were known, there was no physical description
for the origin of these discrete lines. The leading theory of the day was
that atoms and molecules had certain resonance frequencies at which they
would emit, but there was no satisfactory description of the physical origins
of these resonances. Furthermore, there were no other closed formulae to
predict the emission spectral lines of other, more complex, materials.

Photoelectric effect: When a metal surface is illuminated by light, elec-

trons can be emitted from the surface. This phenomenon is known as the
photoelectric effect, and was first discovered by Heinrich Rudolph Hertz in
1887 while investigating electromagnetic radiation. In 1902 Phillipp Lenard
observed that the maximum photoelecton kinetic energies are independent
of intensity but depend on frequency, which could not be explained by a
wave theory of light. Einstein extends the work of Planck and applies it to
describe the photoelectric effect.

8
Heat capacity of solids: Another of the minor details that needed fixing,
according to the wisdom of classical physics, was the theory of heat capacity
for solids. The specific heat capacity of a material system C, is defined as
the amount of heat Q, required to raise the temperature of a set amount of
material (typically one mole) by a given amount T
C = Q/T. (1.3)
The SI unit of specific heat capacity is a J/mol-K. Recall that accord-
ing to the equipartition theorem of classical statistical mechanics, the total
energy contained in an assembly of a large number of individual particles ex-
changing energy amongst themselves through mutual interactions is shared
equally on average by all the particles. In other words, at temperature T ,
each atom has an energy of kB T /2 per degree of freedom for both the kinetic
and potential energies. For an atom in a crystalline solid, there are three de-
grees of freedom (associated with the three directions they can wiggle about
their equilibrium positions), and thus they have kinetic energy K = 3/2kB T ,
and potential energy U = 3/2kB T , giving total thermal energy stored in the
system of E = 3kB T . Thus the amount of heat required to increase the
temperature a one mole of atoms by T is given by the difference between
the final and initial energies Q = 3kB (T + T ) 3kB T = 3kB T . Thus the
heat capacity is given by C = 3kB , which is independent of temperature.
This is known as the Dulong-Petit law after its French discoverers (1819),
and derived theoretically by Boltzmann in 1876.
This prediction of classical physics agrees fairly well with experimental
observations for most materials near room temperature. However, this tem-
perature independent behaviour was not observed at low temperatures for
certain materials, particularly diamond (carbon), boron and silicon. The
carbon anomaly had been known since 1841. In experiments published in
1905 it was shown by the Scottish chemist James Dewar that the heat ca-
pacity of diamond essentially vanishes near 20 K, and as the temperature of
any material approaches absolute zero, the heat capacity should approach
zero as well. The solution to this problem was partially solved by Einstein in
1907 by extending the ideas of Max Planck, by assuming that the atoms are
constrained to oscillated about their equilibrium positions in a lattice at fre-
quency , can oscillate with only discrete energies given by integer multiples
of h, where h is Plancks constant. Einsteins theory was further refined
and gives excellent agreement with experiments by the Dutch physicist P.
Debye in 1912. The results are shown in Fig. 1.4.

Blackbody radiation and the ultraviolet catastrophe: The previously

mentioned phenomena all played a role in the development of the quantum
theory we know today. However, there was one key unresolved problem that
led to the discovery of quantum physics and contributed to the resolution
of many of the other problems of the time: the origin of the spectrum of

9
Figure 1.4: Classical, Einstein, and Debeye treatment of heat capacity of
solids.

blackbody radiation. We will go through this in detail shortly so I delay our

discussion until then.

The beginnings of quantum physics: 1900-1932

1900 Blackbody radiation: Quantum physics was introduced to the world

by Max Planck, speaking at the meeting of the German Physical So-
ciety on 14 December 1900, when he put forth his derivation of the
blackbody spectrum. The key assumption was that the energy ab-
sorbed and emitted by the oscillators that make up the blackbody is
proportional to the frequency of oscillator.

1905 Photoelectric effect: The success of Plancks theory to describe the

observed spectral emission for blackbody radiation by constraining
the absorption and emission of radiation to discrete energy values
prompted a young patent clerk working in Bern Switzerland to apply
this idea to an outstanding problem of the time, namely the photoelec-
tric effect. Albert Einstein, one of the few scientists to take Plancks
ideas seriously, proposed a quantum of light (the photon), which be-
haves like a particle.

1907 Heat capacity of solids: Einstein extended the ideas of Max Planck,
by assuming that the atoms in a solid are which are constrained to
oscillated about their equilibrium positions in a lattice at frequency ,

10
 Figure 1.5: Peter Joseph William Debye

can only oscillate with only discrete energies given by integer multiples
of h, where h is Plancks constant.

1911 Nuclear model of atom: Ernest Rutherford infers the nucleus as the re-
sult of the alpha-scattering experiment performed by Hans Geiger and
Ernest Marsden and proposes the nuclear model of atom, superseding
Thomsons plumb-pudding model.

1913 Bohrs atom: Niels Bohr succeeds in constructing a theory of atomic

structure based on Rutherfords nuclear planetary model of the atom
and the quantum ideas of Planck and Einstein. The key insight was
that there were only discrete energies that the system could have. The
electrons were said to occupy stationary states at these energies, which
do not radiate electromagnetic energy.

1914 Franck-Hertz experiment: James Franck and Gustav Hertz confirm

the existence of stationary states through an electron-scattering ex-
periment.

1923 X-ray - electron scattering: Arthur Compton discovers the quantum

(particle) nature of x rays, thus confirming photons as particles.

1924 de Broglie waves: Louis de Broglie proposes that matter has wave
properties.

1924 Bosons: Satyendra Nath Bose and Albert Einstein find a new way
to count quantum particles, later called Bose-Einstein statistics, and
they predict that extremely cold atoms should condense into a single
quantum state, later known as a Bose-Einstein condensate.

1925 Matrix mechanics: Werner Heisenberg, Max Born, and Pascual Jor-
dan develop matrix mechanics, the first complete version of quantum

11
 mechanics, and make an initial step toward quantum field theory.

1925 Exclusion principle: Wolfgang Pauli formulates the exclusion principle

for electrons in an atom.

1926 Wave mechanics: Erwin Schrodinger develops wave mechanics by try-

ing to determine the equations of motion that describe de Broglies
waves. Max Born gives a probability interpretation of quantum me-
chanics. G.N. Lewis proposes the name photon for a light quantum.

1926 Fermions: Enrico Fermi and Paul A.M. Dirac find that quantum me-
chanics requires a second way to count particles, Fermi-Dirac statistics,
opening the way to solid-state physics.

1926 Quantum theory of light: Dirac publishes seminal paper on the quan-
tization of electromagnetism and quantum field theory is born.

1927 Heisenberg Uncertainty Principle: Werner Heisenberg formulates the

uncertainty principle: the more you know about a particles energy,
the less you know about the time of the energy (and vice versa.) The
same uncertainty applies to momenta and coordinates.

1928 Paul Dirac combines quantum mechanics and special relativity to de-
scribe the electron.

1931 Paul Dirac realizes that the positively-charged particles required by

his equation are new objects (he calls them positrons which he mis-
takenly believes is the proton). They are exactly like electrons, but
positively charged. This is the first example of antiparticles.

1932 Carl David Anderson discovers antimatter, an antielectron called the

positron.

Further developments: 1932-1995 Nuclear physics, quantum field the-

ory, superconductivity, and spooky action at a distance

1934 Enrico Fermi puts forth a theory of beta decay that introduces the
weak interaction. This is the first theory to explicitly use neutrinos
and particle flavor changes.

1934 Hideki Yukawa combines relativity and quantum theory to describe

nuclear interactions by an exchange of new particles (mesons called
pions) between protons and neutrons. From the size of the nucleus,
Yukawa concludes that the mass of the conjectured particles (mesons)
is about 200 electron masses. This is the beginning of the meson theory
of nuclear forces.

12
 1935 Albert Einstein, Boris Podolsky, and Nathan Rosen raise concerns
about the consequences of quantum theory for correlated quantum
systems and put forth the EPR paradox.

1942 Richard Feynman puts forth his path integral formulation of quantum
mechanics in his PhD thesis.

1946-48 Experiments by Isidor Rabi, Willis Lamb, and Polykarp Kusch reveal
discrepancies in the Dirac theory of hydrogen.

1947 Richard Feynman, Julian Schwinger, and Sin-Itiro Tomonaga develop

the first complete theory of the interaction of photons and electrons,
quantum electrodynamics, which accounts for the discrepancies in the
Dirac theory, giving procedures to calculate electromagnetic proper-
ties of electrons, positrons, and photons. Introduction of Feynman
diagrams.

1951 David Bohm introduced a simplified version of the EPR paradox con-
sisting of correlated spins.

1957 Julian Schwinger writes a paper proposing unification of weak and

electromagnetic interactions.

1957 John Bardeen, Leon Cooper, and J. Robert Schrieffer show that elec-
trons can form pairs whose quantum properties allow them to travel
without resistance, providing an explanation for the zero electrical re-
sistance of superconductors. This theory was later termed the BCS
theory (after the surname initials of the three physicists).

1959 Yakir Aharonov and David Bohm predict that a magnetic field affects
the quantum properties of an electron in a way that is forbidden by
classical physics. The Aharonov-Bohm effect is observed in 1960 and
hints at a wealth of unexpected macroscopic effects.

1964 John S. Bell proposes an experimental test, Bells inequalities, of

whether quantum mechanics provides the most complete possible de-
scription of a system.

1982 Alain Aspect carries out an experimental test of Bells inequalities and
confirms the completeness of quantum mechanics.

1995 Eric Cornell, Carl Wieman, and Wolfgang Ketterle trap clouds of
metallic atoms cooled to less than a millionth of a degree above abso-
lute zero, producing Bose-Einstein condensates, which were first pre-
dicted 70 years earlier. This accomplishment leads to the creation of
the atom laser and superfluid gases.

13
 Figure 1.6: George Gamow

Further reading:
For a well-written, entertaining, and lucid historical perspective on the de-
velopment of quantum physics, I highly recommend Gamows book, Thirty
Years That Shook Physics: Story of Quantum Theory. Not only did he
experience this transitional period firsthand, but he also contributed signifi-
cantly to broad developments in physics from nuclear physics to cosmology.
Another interesting view can be found in Longairs, Theoretical Concepts
in Physics, which gives a useful description of not only the development of
quantum physics, but also the other foundations of modern physics. Not
only does it present the historical background, but also the physics. Finally,
for those interested in the historical development of physics during the last
century, Kraghs book, Quantum Generations: A History of Physics in the
Twentieth Century, is an invaluable resource.

14
Chapter 2

Particle-like behavior of light

2.1 Photoelectric effect

(a) Heinrich Rudolf Hertz (b) Albert Einstein (c) Robert Andrews Mil-
likan

When a metal surface is illuminated by light, electrons can be emitted

from the surface. This phenomenon is known as the photoelectric effect, and
was first discovered by Heinrich Rudolph Hertz in 1887 while investigating
electromagnetic radiation.
The emitted electrons are known as photoelectrons. An example exper-
imental setup to observe this phenomenon is shown in Fig. 2.1.
Light incident on the metal surface (the emitter) can release electrons (e)
from the surface. These photoelectrons can travel towards a second metal
plate (the collector). A variable external voltage (Vext ) is applied between
the emitter and collector, measured by a voltmeter (V ), while an ammeter
(A) registers the current i flowing between the emitter and collector. The
current measured reflects the number of photoelectrons emitted. The ex-
periment is performed in a vacuum tube so that the photoelectrons do not

15
Figure 2.1: Schematic of experimental setup to observe the photoelectric
effect. Light is incident on a metal surface (emitter) which can release an
electron. The released photoelectron can travel toward a collector, held at
potential Vext above the emitter. If they electrons have sufficient energy,
they reach the collector and contribute to the current i measured by the
ammeter (A). A voltmeter (V ) monitors the potential difference between
the emitter and collector.

lose any energy from collisions with molecules of the air. This setup allows
the rate of photoelectron emission to be measured, which is proportional to
the current i, as well as the maximum kinetic energy of the photoelectrons,
Kmax . This is determined by applying a negative potential to the collector
that is just sufficient to repel the most energetic photoelectrons, which do
not have sufficient energy to climb the potential barrier. This potential,
known as the stopping potential Vs , is determined by increasing the magni-
tude of the voltage until the current measured drops to zero. At this point
the maximum kinetic energy is given by Kmax = eVs , where e is the mag-
nitude of the electron charge. Typical values of Vs are on the order of 1
V.

2.1.1 Classical analysis

In the classical treatment, the metal surface is illuminated with light of in-
tensity I (W/m2 ). An electron on the surface absorbs energy continuously
from the wave until the binding energy of the electron to the metal is ex-
ceeded, and then the electron is released. The minimum value of energy
required to free an electron from the surface of a metal is called the work
function or binding energy Ebind . Typical binding energies are on the order
of a few electron volts (eV). Classical theory predicts

16
 1. The maximum kinetic energy of the photoelectrons should be propor-
tional to the intensity of incident radiation. As the brightness of the
light is increased, more energy is delivered to the surface and the emit-
ted photoelectrons should be release with more energy. Another way to
think about this is the incident electric field is larger (I = 0 c |E|2 /2),
which implies an increased force on the electron F = eE, and thus
increased energy when it is emitted from the surface.

2. The photoelectric effect should occur for light of any frequency. Ac-
cording to the wave theory of light, as long as the intensity of the
incident radiation is large enough photoelectrons should be able to be
emitted.

3. The first photoelectrons should be emitted a finite amount of time after

the light hits the surface (on the order of seconds). To estimate the
time required to emit a photoelectron, assume that light of intensity
I = 120 W/m2 (about that of a helium-neon laser) is incident on
aluminium (Ebind = 4 eV). Furthermore, assume the atoms to occupy
an area of radius equal to that of an aluminium atom (r = 0.1 nm).
The time taken for the electrons in the atoms to absorb at least the
binding energy, and thus be freed, is then
Ebind
t = , (2.1)
Ir2
which gives approximately 0.2 seconds.

2.1.2 Experimental observations

The experimental characteristics of the photoelectric effect were well known
by 1902. However, these differed significantly from the classical predictions.

1. The maximum kinetic energy as determined by the stopping poten-

tial is completely independent of the intensity of the incident radia-
tion. Figure 2.2 shows the observed behaviour of the measured current.
Doubling the intensity does not change the stopping potential, indicat-
ing no change in the kinetic energy of the electrons. The experimental
result conflicts with the predictions of wave theory, which predicts the
maximum kinetic energy should depend on the intensity of light.

2. Photoelectrons are not emitted for light of any frequency. The light
must have a minimum frequency for the photoelectric effect to be ob-
served. The value of this frequency, called the cutoff frequency c , is
a characteristic of the material. For light with frequency above c ,
any intensity will produce photoelectrons. However, for light with fre-
quency below c , no amount of intensity will enable photoelectrons to

17
 be emitted. This experimental result again disagrees with the classical
theory.

3. The first photoelectrons are emitted nearly instantaneously (within a

few nanoseconds) of the light hitting the surface. Again, the experi-
ment does not fit with classical predictions.

Figure 2.2: Photoelectric current as a function of measured potential dif-

ference between emitter and collector for two different values of incident
radiation intensity. When the intensity is doubled from I1 to I2 , the cur-
rent is doubled (twice as many photoelectrons emitted), but the stopping
potential, Vs remains constant.

2.1.3 Einsteins treatment based upon Plancks idea of light

quanta
In 1905, a satisfactory description of the photoelectric effect was given by
Einstein.1 Einstein based his theory on the work of Max Planck, who five
years earlier developed a theory to explain the spectral distribution of light
emitted by hot, glowing objects (known as blackbody radiation). We will
discuss blackbody radiation and Plancks theory in the next lecture. Einstein
proposed that the energy of the light wave was not continuously distributed
across the wavefront, but instead concentrated in localized bundles, (light
quanta), which were later called photons. The energy of a photon associated
1
Although it is often said that the photoelectric effect requires the electromagnetic
field to be quantized, i.e. comprised of photons, this is not true. A so-called semi-
classical approach in which the electromagnetic field is treated as a classical wave, while
the atoms that make up the material are treated quantum mechanically. The quantized
atoms, which can only absorb and emit light at discrete energies, is sufficient to describe
the photoelectric effect. This fact should not detract from Einsteins insight at the time,
since the quantum theory of atomic systems was still eight years away.

18
with an electromagnetic wave of frequency is
E = h, (2.2)
where h is known as Plancks constant. Note this can also be written in
terms of the wavelength
hc
E= . (2.3)

In Einsteins treatment, a photoelectron is release as a result of the absorp-
tion of a single photon. The entire energy of the photon is instantaneously
transferred to a single photoelectron. If the photon energy h is greater
than the work function (binding energy) Ebind , of the material then the
photoelectron will be released. However, if the photon energy is less than
the binding energy, then no photoelectron can be emitted. This accounts
for two of the failures of the classical theory: the existence of the cutoff
frequency and lack of any measurable time delay in emission.

If the photon energy exceeds the work function, the excess is taken up in
the electron kinetic energy
Kmax = h Ebind . (2.4)
The intensity of light does not appear here. Doubling the intensity of light
means that twice as many photons are incident on the surface and thus
twice as many photoelectrons should be emitted, but they have the same
maximum kinetic energy.

The photon theory given by Einstein gave a satisfactory description of all the
observed features of the photoelectric effect. The most detailed experimental
test of the theory was performed by the American physicist Robert Millikan
in 1915.
He measured the maximum kinetic energy (the stopping potential times
the electron charge: Kmax = eVs ,), for different frequencies of light. Plotting
Vs versus frequency, we expect
Vs = h/e Ebind /e. (2.5)
The slope of the line should give h/e, as show in Fig. 3. The intercept
should give the cutoff frequency and thus the binding energy. Millikan had
obtained a value for the electron charge (e = 1.592 1019 C, which is
within 1% of the currently accepted value) from his famous oil drop ex-
periments, which he used to determine Plancks constant. Millikan obtained
a value of h = 6.57 1034 Js, which is quite close to the presently accepted
value of approximately 6.626 1034 Js. In part for his careful experiments
on the photoelectric effect, Millikan was awarded the 1923 Nobel prize in
physics. Einstein was awarded the Nobel price in physics for his theory of
the photoelectric effect.

19
Figure 2.3: Experimental results of photoelectric effect. Stopping voltage Vs
as a function of light frequency is linear. The slope of the line should be
h/e. With previous knowledge of the electron charge from his earlier experi-
ments, Millikan was able to measure the slope of this line and experimentally
determine a value for Plancks constant.

2.2 Blackbody radiation

At the turn of the 20th century the tip of the quantum iceberg was revealed
by the seemingly everyday phenomenon of thermal radiation, that is the
spectral intensity distribution of electromagnetic radiation emitted by an
object because of its temperature.

2.2.1 Introduction: Blackbody radiation

We see ordinary objects because light is reflected from their surface from
another source and is reflected to our eyes. At room temperature, most ther-
mal radiation is emitted in the infrared region of the spectrum (max 10
m), where our eye is not sensitive. However as the temperature of an object
is increased, for instance by passing a current through a metal as in an elec-
tric filament, the material begins to emit visible light. Unfortunately, the
amount of radiation emitted by everyday materials depends not only on the
temperature, but also other properties of the material such as the reflectiv-
ity of the surface. To simplify these difficulties in calculating the spectrum
of heated objects, we consider an idealized object called a blackbody, whose
surface is completely absorbing. A blackbody absorbs all incident radiation
regardless of frequency or angle of incidence.

A blackbody can be modeled by considering a large cubic cavity, which con-

sists of perfectly reflecting mirrors for the internal walls, with a small hole

20
cut in one wall as depicted in Fig. 2.4. The hole acts as a blackbody, in that
any radiation from outside the cavity that is incident on the hole, will be-
come trapped inside the cavity, bouncing around indefinitely. Thus the hole
effectively has perfect absorption, regardless of incidence angle or frequency.
Since radiation emerging from the hole is just a sample of the radiation in-
side the cavity, understanding the radiation inside the cavity allows us to
determine the emitted spectra from the hole and thus the blackbody.

Figure 2.4: An optical cavity with perfect internal mirrors acts as a simplified
blackbody.

We must now define what we mean by blackbody spectrum. For our pur-
poses, we are concerned with the energy of radiation per unit frequency
per unit volume inside the blackbody cavity with walls at temperature T ,
called the spectral energy density, (, T ). This represents the energy per
unit frequency per unit volume inside our blackbody cavity and has units of
J Hz1 m3 . The interpretation of (, T )d is that the amount of electro-
magnetic energy inside the box, per unit volume in the frequency interval
(, + d), is (, T )d. The radiation emitted from the hole in the cavity
is proportional to this energy density, and thus (, T ) determines the spec-
trum measured by an observer outside the cavity.

The calculation to determine the spectral energy density inside the black-
body cavity can be broken into two parts. First, due to the fact that the
radiation inside the blackbody cavity is confined to a finite region of space
by the walls, electromagnetic boundary conditions will restrict the number
of electromagnetic standing waves (called electromagnetic modes) with fre-
quencies between and + d. The number of modes inside the cavity,
per unit volume, per unit frequency is denoted by g() and known as the

21
density of states. A state here refers to a mode of the electromagnetic field,
which is determined by the polarization and allowed wavelength, as set by
the boundary conditions imposed by the cavity. The total number of modes
per unit volume inside the cavity is thus g()d. Each of these modes is
then assumed to be in thermal equilibrium with the cavity walls, which are
held at temperature T .

The second part of the calculation involves the determination of the aver-
age energy of the electromagnetic field at a given frequency of oscillation,
hE(, T )i, in thermal equilibrium at temperature T . The average energy at
a given frequency is
Z
hE(, T )i = EP (E, T )dE, (2.6)

where P (E, T ) is the probability density of energy at temperature T . To

determine this quantity, we must rely upon statistical mechanics, which we
will discuss shortly.

Thus to calculate the spectral energy density, we multiply the density of

states g() by the average energy of the electromagnetic field at a given
frequency hE(, T )i, giving

(, T ) = g()hE(, T )i. (2.7)

2.2.2 Density of electromagnetic states for a cavity

1D cavity
To calculate the density of states for a cubic cavity of side L, we begin
by considering a 1D cavity and then generalize this result. If we consider
electromagnetic waves confined to oscillate between two perfectly reflecting
mirrors positioned at x = 0 and x = L, the boundary conditions imply
that the field must go to zero at the mirrors. This leads to standing wave
solutions such that the an integer number of wavelengths should equal the
twice the cavity length
n = 2L. (2.8)
The factor of 2 comes from the fact that you need only half integer values
of wavelength to satisfy the boundary conditions, as depicted in Fig. ??.
This equation thus determines the allowed wavelengths of radiation in
the cavity n = 2L/n, where n is a mode label, running from 0 to . The
frequency associated with a wavelength is simply given by = c/, which
combined with the constraints on wavelength imposed by the cavity gives
the allowed frequencies of the cavity
c
n = n. (2.9)
2L

22
 Figure 2.5: One-dimensional cavity with the first two allowed states.

Thus, the electric field amplitude can be written as

X
E(x, t) = An sin(kn x n t), (2.10)
n

where the sum is taken over mode labels n. Here An is the amplitude,
kn = 2/n is the wave vector, and n = 2c/n = is the angular fre-
quency of each mode labeled by n.

The number of modes with frequency in the range (, + d) can be calcu-

lated by counting the number of mode indices that give frequencies in this
range (and multiplying by 2 for the two possible polarizations associated
with each wavelength). The number of mode indices from 0 to is given by
/ where = c/2L is the frequency spacing between subsequent modes.
Similarly, the number of mode indices from 0 to + d is ( + d)/. Thus,
the number of mode indices between and + d is given by the difference
+ d 2L
= d. (2.11)
c
Thus, the density of states for a 1D cavity is given by the number of modes
divided by the cavity volume, which for a 1D cavity is just L
1 2L 4
g1D ()d = 2 d = d, (2.12)
L c c

23
where the factor of 2 out front arises from the two possible polarizations for
each mode.
Another way to think of this is in terms of the wave vector k = 2/ =
2/c. The allowed wave vectors, kn = 2/n = n/L are on a lattice of
equal spacing k = /L. The number of modes with frequency in the range
(, + d) is equal to the number of modes with wave vectors that have
length between k and k + dk, where k = 2/c and dk = 2d/c. The
number of modes is given by the volume of wave vector space in the shell
with wave vector magnitude in the range (k, k + dk) divided by the volume
spacing between allowed wave vectors k = /L. The volume of the shell
in 1D k-space with wave vector magnitude in the range (k, k + dk) for a 1D
cavity is given by
2
k + dk k = dk = d. (2.13)
c
The spacing between allowed wave vectors is just k = n/L, so that the
number of modes with wave vector in the range (k, k + dk) is
dk dk 2L
= = d, (2.14)
k /L c
which agrees with Eq. (2.11) above.

2D cavity
This can be generalized to a 2D square cavity of sides L in a similar manner.
The allowed wave vectors now have two components, an x- and y-component,
each with a mode label nx and ny , respectively. The allowed wave vectors
are given by  
knx ,ny = nx , ny , (2.15)
L L
where we assume the cavity lengths are the same in both directions. The
allowed modes lie on a lattice in k-space, with each lattice point labeled by
(nx , ny ), as depicted in Fig. 2.6 below.
The number of modes with frequencies between and + d can be cal-
culated in a manner similar to that for the 1D case, by considering the
volume of a shell in the 2D k-space. The radius of the shell should be
between k and k + dk, where k = 2/c and dk = 2d/c. The number
of modes in this shell is given by the volume of the shell divided by the
volume associated with each mode, which is equivalent to the mode spacing.
The volume of the shell in the 2D k-space is the difference the areas of the
circles of radii k + dk and k, divided by 4
1
kdk
(k + dk)2 k 2 , (2.16)
4 2
where we have only kept terms to leading order in dk. The factor of 1/4
arises from the fact that we are the allowed values of nx and ny are positive,

24
Figure 2.6: Geometry of wave vector space to calculate the density of optical
states for a 1D, 2D and 3D cavity with sides of equal length L. The boundary
conditions on the electromagnetic field impose constraints on the allowed
modes as depicted by the various lattices.

and thus we must only consider the upper right quadrant of the 2D k-space.
This gives the 2D volume of the allowed shell. To determine the number
of allowed standing-wave modes, we take this volume and divide by the
lattice spacing, which is given by k 2 = (/L)2 for our 2D lattice with equal
spacing, and multiply by 2 for each polarization
kdk/2 L2
2 = kdk. (2.17)
(/L)2
This can be written in terms of frequency by recalling that k = 2/c and
dk = 2d/c, giving the number of allowed modes

L2 4L2
kdk = 2 d. (2.18)
c
Thus, the density of states is given by the number of allowed modes in our
cavity divided by its volume, which for the 2D cavity of sides L is L2 , giving
4
g2D ()d = d. (2.19)
c2

25
3D cavity
Finally, to determine the density of states for a 3D cavity of sides L, we
can generalize the above analysis by considering the volume of a shell in 3D
k-space with radii between k and k + dk as depicted in Fig. ??. The volume
of this shell is given by the difference in volumes of the two spheres of radii
k + dk and k
4 4
(k + dk)3 k 3 4k 2 dk, (2.20)
3 3
where we again only keep the leading order in dk. Note we have used the
expansion
(x + y)3 = x3 + 3x2 y + 3xy 2 + y 3 ,
to simplify the expression. The number of allowed modes with wave vector
magnitudes in the range (k, k + dk), is then given by the volume of this shell
in 3D k-space given in Eq. (2.20), divided by the mode volume k 3 = (/L)3 ,
divided by 8 since the allowed values of mode labels (nx , ny , nz ) are positive
only, and thus we only take the positive octant of the spherical shell, and
multiplied by 2 for the possible polarizations associated with each wave
vector
1 4k 2 dk L3 2
 
2 = k dk. (2.21)
8 k 3 2
This can also be expressed in terms of the frequency using k = 2/c and
dk = 2d/c, giving
L3 2 8L3 2
k dk = d. (2.22)
2 c3
Thus the density of states for the 3D cavity is given by the number of modes
in Eq. (2.22) divided by its volume L3
8 2
g3D ()d = d. (2.23)
c3

2.2.3 Average energy for an electromagnetic oscillator

We now have half the information we need to determine the spectral energy
density in Eq. (2.7). All that is needed is to determine hE(, T )i, the
average energy of the electromagnetic field at a given frequency when in
thermal equilibrium with the cavity walls held at temperature T . This can
be accomplished by using a statistical mechanics approach. We will discuss
the two key methods that were taken in 1900 to explain the blackbody
spectrum in the following two sections.

Classical Rayleigh-Jeans treatment and the ultraviolet catastrophe

In 1900, Lord Rayleigh published a paper that attempted to describe the
spectrum of blackbody radiation using a classical theory of waves. He fol-
lowed the same line of reasoning that led to Eq. (2.7) above. He obtained

26
 (a) Lord Rayleigh (John William (b) James Hopwood Jeans
Strutt)

the same expression for the density of states as we have in Eq. (2.23) above.
To determine the average energy per frequency mode of the electromagnetic
field, Rayleigh relied upon the results of Boltzmann and Maxwell, known as
the equipartition theorem. The equipartition theorem essentially says that
for a system of particles in thermal equilibrium at temperature T each have
mean energy of kB T /2 per degree of freedom, i.e. per quadratic term in the
Hamiltonian, where
kB = 1.381 1023 J/K,
is the Boltzmann constant. For the radiation in the blackbody cavity, each
mode is in thermal equilibrium with the atoms in the walls that oscillate at
the radiation frequency . At temperature T each atom has average kinetic
energy K = kB T /2 and average potential energy U = kB T /2 so that the
total thermal energy associated with a frequency mode is simply

hE(, T )i = kB T. (2.24)

Note that this is completely independent of frequency. Each frequency mode

of oscillation, of which there are an infinite number, has the same amount
of thermal energy. This is a key issue with the present treatment.

An equivalent way to determine the average energy per frequency mode is by

using Boltzmanns theory of statistical mechanics, in which the probability
for a system to have energy between E and E+dE is given by the Boltzmann
probability density
 
1 E
PB (E, T ) = exp , (2.25)
kB T kB T
as shown in Fig. pe.

27
 Figure 2.7: Boltzmann probability distribution.

The average energy is thus given by the integral in Eq. (2.6), with the
Boltzmann probability density. This integral is equal to the area under the
curve EPB (E, T ), shown in Fig. 2.8.

Figure 2.8: Energy distribution EPB (E, T ) for classical continuous energy.

 
E
Z
1
hE(, T )i = E exp dE = kB T, (2.26)
0 kB T kB T
where I have used the following integral
Z  
x
x exp dx = a2 .
0 a
Combining the expressions for the density of states and average energy per
frequency mode, we obtain the spectral energy density predicted by Rayleigh
8 2
RJ (, T ) = kB T. (2.27)
c3

28
Figures 2.9 and 2.10 show the behaviour of the Rayleigh-Jeans theory along
with observed spectral energy density for T = 6000 K. At short wavelengths
in the infrared and into the visible region of the spectrum the theory works
well. However, as the frequency is increased (thus decreasing wavelength),
the theoretical predictions deviate from the observations. Upon inspection
of this result, you should note that as the frequency increases towards the
ultraviolet part of the spectrum (away from the visible), so does the spectral
energy density of the blackbody radiation according to this model. However,
this is not what is observed as depicted in Figs. 2.9 and 2.10. In fact,
Eq. (2.27) predicts that the spectral energy density diverges to infinity as
the frequency increases implying that we should observe an ever-increasing
amount of energy at higher frequencies. The failure of the Rayleigh-Jeans
formula at high frequencies is known as the ultraviolet catastrophe.

Figure 2.9: Normalized frequency distribution at 6000 K.

Figure 2.10: Normalized wavelength distribution at 6000 K.

29
 Figure 2.11: Max Karl Ernst Ludwig Planck

Planck treatment and the introduction of light quanta

In the closing months of 1900, a successful theory to describe the spectral
energy density of blackbody radiation was developed by the German the-
orist, Max Planck. The method by which Planck arrived at the equation
that now bears his name, the Planck distribution, is a complex story and
for a detailed exposition I suggest the text by M. Longair. In the treatment
below, we will not follow Plancks original formulation, but rather a more
modern description.
According to Eq. (2.27), the ultraviolet catastrophe results from too
much energy being distributed at high frequencies. What is needed to fix
this is a way to make (, T ) 0 as . In Plancks analysis, he
assumed that the reflections at the walls of the blackbody cavity resulted
from radiation being absorbed and then quickly reemitted by the atoms of
the wall. During this absorption-reemission time the atoms would oscillate
at the frequency of the radiation. Since the energy of an oscillating system
depends on the frequency, 2 Planck tried to find a way to reduce the number
of high-frequency oscillators in the cavity walls, since the radiation field was
in thermal equilibrium with the walls of the cavity. He did this by suggesting
that an oscillating atom can absorb or reemit energy only in discrete bundles,
or quanta, later called photons. If the energy of the quanta were proportional
to the frequency of the radiation, then as the frequencies became large the
energy also became increasingly large. In his theory, the energy absorbed or
2
Recall that the total energy for a 1D harmonic oscillator is

p2x + 2 x2
E= ,
2m
where = 2 is the resonant frequency of the oscillator.

30
reemitted by an atom oscillating with frequency is an integer multiple of
the energy of a single quantum

E(, n) = nh, (2.28)

where the constant of proportionality,

h = 6.6260755 1034 J s, (2.29)

is now known as the Planck constant and has units of J s or action. The key
difference between the classical and quantum treatment is that in
the quantum treatment any physical quantity that oscillates with
frequency in time, has total energy E satisfying Eq. (2.28).
Whereas classically, any continuum of values for the total energy could be
obtained, as depicted in Fig. 2.12

Figure 2.12: Allowed classical and quantized energy levels.

To show how this assumption fixes the ultraviolet catastrophe, let us cal-
culate the average energy for electromagnetic modes of frequency . The
atoms of the wall that oscillate at frequency are in thermal equilibrium
with the corresponding electromagnetic frequency modes. Thus, we use the
Boltzmann density of Eq. (2.25) to describe the probability of having en-
ergy between E and E + dE at thermal equilibrium. However, instead of
integrating over a continuum of allowed energies, as in Eq. (2.26), we have a
discrete sum over the allowed energies in Eq. (2.28). This gives the following
expression for the average energy at frequency
P
E(, n)PB (E(, n), T )
hE(, T )i = n=0
P . (2.30)
n=0 PB (E(, n), T )

31
Figure 2.13: Energy distribution En PB (En , T ) for quantized energy values.
Note that for small quantization energies, the quantized sum is nearly the
same as the continuous classical case in Fig. 2.8. However, the larger the
quantization energy, the larger the deviation from the classical continuous
situation.

This argument of this sum En P (En , T ) is show in Fig. 2.13.

To evaluate the sums we first simplify the notation, but introducing
the dimensionless quantity = h/(kB T ), which allows us to write the
numerator and denominator as

X
X
E(, n)PB (E(, n), T ) = kB T nen (2.31)
n=0 n=0

and

X
X
X
n
Z= PB (E(, n), T ) = e = (e )n , (2.32)
n=0 n=0 n=0

respectively. Here we have used the Boltzmann factor for the probability
 
E(, n)
PB (E(, n), T ) = exp . (2.33)
kB T

32
Note that the denominator can be expressed as a geometric series

X 1
xn = , (2.34)
1x
n=0

where we identify
x = e ,
which gives the denominator
1 1
Z= = . (2.35)
1 e

1 exp kh
BT

To determine the numerator, we note that it is proportional to a derivative

of the denominator, which can be seen by noting that

dZ X
= n(en ). (2.36)
d
n=0

From this we see that the numerator can be expressed as

X dZ
kB T nen = kB T . (2.37)
d
n=0

To evaluate this, we take the derivative of Eq. (2.35) with respect to

dZ d  1  2
= 1 e = e 1 e = e Z 2 , (2.38)
d d
which leads to the following numerator

X
kB T nen = kB T e Z 2 . (2.39)
n=0

Thus we see that the average energy at frequency is now given by

kB T e Z 2 h
hE(, T )i = = , (2.40)
Z e 1
where we have used the fact that kB T = h. Combining this with the
density of states we determined in the previous section, Eq. (2.23), enables
us to calculate the spectral energy density for a blackbody following Plancks
quantized oscillator hypothesis
8 2 h
P (, T ) = g3D ()hE(, T )i = 3
h/(k T ) . (2.41)
c e B 1
This is Plancks major result. By assuming that the oscillators in the wall
with which the electromagnetic field modes are in thermal equilibrium can

33
only absorb and reemit radiation with energy that is an integer multiple
of the oscillator (and thus radiation) frequency, Planck was able to derive
the spectral energy density that he had previously obtained from empirical
data. With this mathematical trick Planck ushered in the quantum era.

Stefans law and total energy density: There are a few things to note
about the blackbody spectrum. The first is the total energy density emitted
by a blackbody at temperature T is given by
Z
8h 3
Z
(T ) = P (, T )d = 3 d, (2.42)
0 c 0 eh/(kB T ) 1
which can be calculated using the following definite integral and change of
variables Z
x3 4
d = . (2.43)
0 ex 1 15
The resulting total energy density is
8 5 kB
4
4
(T ) = T 4 = T 4 , (2.44)
15(hc)3 c
where is known as the Stefan-Boltzmann constant. This relationship be-
tween total energy density emitted from a blackbody and its temperature
was known by from careful experiments and is known as Stefans law.

Weins displacement law: The frequency at which a blackbody emits the

peak energy can be determined from the Planck distribution, Eq. (2.41), by
differentiating with respect to frequency, setting the result to 0 and solving
for the frequency. This gives
 
dP (, T ) 8h h
= 3 3 2 (e 1) 3 e /(e 1)2 = 0, (2.45)
d c kB T
where I have made use of the definition of = h/(kB T ). This can be
rearranged to give the following transcendental equation

= 1 e . (2.46)
3
This equation has no analytic solution, but can be solved numerically, which
gives max 2.82144, or max 2.82144kB T /h.

The spectral energy density per unit wavelength, as opposed to frequency,

does not peak at the same energy owing to the fact that the wavelength
is inversely proportional to frequency. Thus to determine the peak of the
wavelength distribution we have to write
 
 d  c
P (, T )d = P (c/, T )   d = P (c/, T ) 2 d, (2.47)
d

34
where P (, T ) is the spectral energy density per unit wavelength. Note
the additional factor of c/2 . Using the expression for the spectral energy
density per unit frequency, this gives
8 hc/
P (, T ) = 4 hc/(k
. (2.48)
e BT ) 1

The peak of the wavelength distribution can be found in a similar manner,

which leads to a slightly different transcendental equation
x
= ex 1, (2.49)
5
where x = hc/kB T is the wavelength equivalent of . This equation can
be solved numerically and gives xmax 4.965. The peak wavelength is thus
given by
hc
max = . (2.50)
xmax kB T
This is known as Weins displacement law, which was known from careful
experiments on blackbody radiation. The displacement refers to the way
in which the peak wavelength changes with temperature of the blackbody.
The product max T is a constant, as shown in Fig. 2.14.

Figure 2.14: Blackbody spectrum as a function of wavelength for four dif-

ferent temperatures. The wavelength at which the peak intensity is emitted
scales as 1/T , which is known as Weins displacement law.

2.3 Shot noise

Further evidence of the quantized nature of light can be found in the shot-
noise associated with fluctuations in the optical intensity of the electromag-

35
netic field owing to the discrete and independent nature of photons. The
term shot noise can refer to either the random, time-dependent fluctuations
in an electrical current due to the discreteness of the electron charge or sim-
ilar random, time-dependent fluctuations in the intensity of electromagnetic
radiation (observed on a detector) due to the quantization of the electro-
magnetic field energy E = nh, where n takes on positive integer values.
This is illustrated by considering image formation on a photosensitive film
when illuminated at extremely low intensity (i.e. very few photons per unit
time) as depicted in Fig. 2.15 below.

Figure 2.15: Visualization of the shot noise associated with discretization of

the electromagnetic field. The image formed on the film is built up over time
by low photon-flux illumination. In the first three time windows (top), the
shot noise dominates and there is essential no information about the image.
The SNR is approximate 1 in this case. As more photons are collected, the
SNR improves and the image starts to take shape (middle). Finally, after
sufficient illumination, there is little information gain for the collection of
additional photons (bottom).

The noise associated with counting a discrete number of objects is not a

unique phenomenon to quantum physics though, and pervades many faucets
of statistical analysis. For example, the number of rain drops that fall per
unit area or the number of nuclear decay events per unit time are also
governed by the same statistical description, called Poisson statistics. The
events are assumed to be independent and randomly occur with an average
of hni per unit time, or space, etc. (Note that the brackets imply an ensemble

36
average.) Then the average uncertainty in the number counted, quantified
by the square root of the variance var(n) = h(n hni)2 i = hn2 i hni2 ,
otherwise called the standard deviation, is just given by
p p p
n = var(n) = hn2 i hni2 = hni. (2.51)

This is characteristic of Poisson processes, that is, the standard deviation

from the mean is just the square root of the mean. The signal-to-noise
ratio (SNR) is defined as the average signal strength, which in the case of
counting photons is the average number of photons detected hni, divided by
the standard deviation (the uncertainty
por noise in our measurement), which
in the case of photon counting is just hni. Thus for a Poisson process, the
signal-to-noise ratio is
p p
SNR = hni/ hni = hni. (2.52)

As the number of photons detected increases, the better we can determine

the actual value of the average intensity (or mean photon number). This
concept is nicely demonstrated by considering the formation of an image on
a photoreactive film when illuminated at low light levels (much less than
one photon per pixel per unit time), as depicted in Fig. 2.15.

2.4 Compton Scattering

Figure 2.16: Arthur H. Compton

When a light beam encounters a single free electron, some of the radia-
tion can be scattered at an angle that deviates from the initial path. Part
of the radiation energy is exchanged for kinetic energy of the electron, and
the remainder is reemitted as electromagnetic radiation. In the classical

37
wave picture of light, the scattered radiation has less energy than the initial
beam, but the same frequency (since the only oscillating frequency is that
of the incident radiation). However, the photon concept leads to a different
conclusion as we will see.
We begin by analyzing the scattering process as an elastic collision be-
tween a single photon and an electron, which we assume to be at rest in the
laboratory frame. The geometry of the collision is depicted in Fig. 2.17.

Figure 2.17: Schematic of Compton scattering experimental setup. A source

of X rays is collimated into a beam with wavelength by two slits and
directed toward the scattering target. The scattered X rays are detected at
various angles by a moveable, spectrally-resolving detector.

The incident photon has energy E given by

hc
E = h = , (2.53)

where h is Plancks constant,c is the vacuum speed of light, is the fre-
quency of the electromagnetic field, and is corresponding wavelength.
From Einsteins relativistic formula relating energy, momentum and mass,
E 2 = c2 p2 + (mc2 )2 , we note that since the photon has zero rest mass, it
will carry linear momentum
E h
p= = . (2.54)
c
The electron at rest has energy mc2 , where m is the electron rest mass.
After scattering, the photon has energy E 0 , momentum p0 , and moves at an
angle with respect to the direction of the incident beam. The electron has
total energy Ee , momentum pe , and travels at an angle with respect to
the incident photon. We use the relativistic relations to account for pos-
sible high-energy incident photons. Next we apply energy and momentum
conservation to give

Ei = Ef
E + mc2 = E 0 + Ee , (2.55)

38
for energy conservation and component-wise for momentum

(px )i = (px )f
p = pe cos() + p0 cos(), (2.56)

and

(py )i = (py )f
0 = pe sin() p0 sin(). (2.57)

There are in general three equations and four unknowns (, , Ee , E 0 ; pe

and p0 are not independent since they are related to the others via the
Einstein energy-moment-mass equations for each particle. Thus we cannot
uniquely solve the equations. However, we can eliminate any two of the
four unknowns. Since we are interested in what happens to the photon after
scattering, we will eliminate the electron energy Ee , and scattering angle
. The angle is eliminated by solving the momentum equations for the
electron contributions, and summing the square

pe cos() = p p0 cos()
pe sin() = p0 sin().

Squaring and then adding gives

p2e = p2 2pp0 cos() + p02 . (2.58)

The electron energy can be eliminated by using the relativistic energy rela-
tion to solve for Ee in terms of pe , giving

Ee2 = c2 p2e + m2 c4 .

From Eq. (2.55), we isolate Ee by subtracting E 0 from both sides, and then
squaring to give

2
E + mc2 E 0 = Ee2 = c2 p2e + m2 c4 ,

which we can easily solve for p2e . Substituting Eq. (2.58) in for p2e gives

2
E + mc2 E 0 = Ee2 = c2 p2 2pp0 cos() + p02 + m2 c4 .


(2.59)

Using the relativistic energy formula for the initial and final photon energies,
E = cp and E 0 = cp0 in Eq. (2.59), we find after a bit of algebra
1 1 1
0
= (1 cos()) . (2.60)
E E mc2

39
This is the Compton scattering formula relating the incident photon energy
to the scattered photon energy and scattering angle. We can rewrite this
equation in terms of the photon wavelengths
h
0 = (1 cos()) , (2.61)
mc
where is the wavelength of the incident photon and 0 is the wavelength
of the scattered photon. The quantity h/mc is known as the Compton
wavelength of the electron, which has a value of 0.0002426 nm. However, keep
in mind this is not a true wavelength, but rather a change of wavelength.
The Compton formulae above give the change in energy (or wavelength) as
a function of scattering angle of the photon. Since the quantity on the
right-hand side is greater than or equal to zero, E 0 is always less than E, the
scattered photon has less energy than the incident photon. The difference
E E 0 is just the kinetic energy given to the electron, Ee mc2 . Similarly,
0 . The change in wavelength ranges from 0 at = 0 to twice the
Compton wavelength at = 180 .
The first experimental observation of this type of scattering was per-
formed by Arthur Compton in 1923. A sketch of the experimental setup is
shown in Fig. 2.18. A beam of X rays is incident on a scattering target.
Compton used carbon, and although no scattering target contains actual
free electrons, the outer-shell (valence) electrons in many materials are
weakly bound, thus behaving like nearly-free electrons. This assumes the
binding energy is negligible compared with the kinetic energy gained from
the scattering process. A moveable detector capable of measuring the scat-
tered photon wavelength (energy) is positioned at various observation angles
.

Figure 2.18: Schematic of Compton scattering experimental setup. A source

of X rays is collimated into a beam with wavelength by two slits and
directed toward the scattering target. The scattered X rays are detected at
various angles by a moveable, spectrally-resolving detector.

The results from Comptons experiments are illustrated in Fig. 2.19. At

each scattering angle, two peaks are observed, corresponding to scattered X-

40
ray photons with two different energies. The wavelength of one peak does not
change as the angle is varied. This peak corresponds to X rays scattered off
inner-shell electrons that are so tightly bound to the atom that they cannot
be knocked loose by the incident X rays. The X rays scattered from these
inner-shell electrons do not lose any energy. However, the wavelength of the
other peak varies strongly with scattering angle and is accurately described
by the Compton formula.

Figure 2.19: Sketch of the results from Comptons original X-ray scattering
experiments.

41