Here we discuss weatherB.

Zwiebach
February 9, 2016

Chapter 1: Key Features of Quantum

Mechanics
Quantum mechanics is now almost one-hundred years old, but we are still
discovering some of its surprising features and it remains the subject of much
investigation and speculation. The framework of quantum mechanics is a rich and
elegant extension of the framework of classical physics. It is also counterintuitive and
almost paradoxical.

Quantum physics has replaced classical physics as the correct

fundamental description of our physical universe. It is used routinely
to describe most phenomena that occur at short distances. Quantum
physics is the result of applying the framework of quantum mechanics
to different physical phenomena. We thus have Quantum
Electrodynamics, when quantum mechanics is applied to
electromagnetism, Quantum Optics, when it is applied to light and
optical devices, or Quantum Gravity, when it is applied to gravitation.
Quantum mechanics indeed provides a remarkably coherent and
elegant framework. The era of quantum physics begins in 1925, with
the discoveries of Schr¨odinger and Heisenberg. The seeds for these
discoveries were planted by Planck, Einstein, Bohr, de Broglie, and
others. It is a tribute to human imagination that we have been able to
discover the counterintuitive and abstract set of rules that define
quantum mechanics. Here we aim to explain and provide some
perspective on the main features of this framework.

We will begin by discussing the property of linearity, which

quantum mechanics shares with electromagnetic theory. This
property tells us what kind of theory quantum mechanics is and why,
it could be argued, it is simpler than classical mechanics. We then
turn to photons, the particles of light. We use photons and polarizers
to explain why quantum physics is not deterministic and, in contrast
with classical physics, the results of some experiments cannot be
predicted. Quantum mechanics is a framework in which we can only
predict the probabilities for the various outcomes of any given
experiment. Our next subject is quantum superpositions, in which a
quantum object somehow manages to exist simultaneously in two

1
mutually incompatible states. A quantum light-bulb, for example,
could be in a state in which it is both on and off at the same time!

Linearity of the equations of motion

In physics a theory is usually described by a set of equations for some
quantities called the dynamical variables of the theory. After writing a
theory, the most important task is finding solutions of the equations.
A solution of the equations describes a possible reality, according to
the theory. Because an expanding universe is a solution of Albert
Einstein’s gravitational equations, for example, it follows that an
expanding universe is possible, according to this theory. A single
theory may have many solutions, each describing a possible reality.

There are linear theories and nonlinear theories. Nonlinear theories

are more complex than linear theories. In a linear theory a remarkable
fact takes place: if you have two solutions you obtain a third solution
of the theory simply by adding the two solutions. An example of a
beautiful linear theory is Maxwell’s theory of electromagnetism, a
theory that governs the behavior of electric and magnetic fields. A
field, as you probably know, is a quantity whose values may depend
on position and on time. A simple solution of this theory describes an
electromagnetic wave propagating in a given direction. Another
simple solution could describe an electromagnetic wave propagating
in a different direction. Because the theory is linear, having the two
waves propagating simultaneously, each in its own direction and
without affecting each other, is a new and consistent solution. The
sum is a solution in the sense that the electric field in the new
solution is the sum of the electric field in the first solution plus the
electric field in the second solution. The same goes for the magnetic
field: the magnetic field in the new solution is the sum of the
magnetic field in the first solution plus the magnetic field in the
second solution. In fact you can add any number of solutions to still
find a solution. Even if this sounds esoteric, you are totally familiar
with it. The air around you is full of electromagnetic waves, each one
propagating oblivious to the other ones. There are the waves of
thousands of cell phones, the waves carrying hundreds of wireless
internet messages, the waves from a plethora of radio-stations, TV
stations, and many, many more. Today, a single transatlantic cable

2
can carry simultaneously millions of telephone calls, together with
huge amounts video and internet data. All of that courtesy of linearity.

More concretely, we say that Maxwell’s equations are linear

equations. A solution of Maxwell’s equation is described by an electric
field E a magnetic field B, a charge density ρ and a current density J,
all collectively denoted as (E, B,ρ, J). This collection of fields and
sources satisfy Maxwell’s equations. Linearity implies that if (E, B,ρ, J)
is a solution so is (αE, αB,αρ, αJ), where all fields and sources have
been multiplied by the constant α. Given two solutions

(E1,B1,ρ1,J1), and (E2,B2,ρ2,J2), (1.1)

linearity also implies that we can obtain a new solution by

adding them
(E1 + E2 , B1 + B2 , ρ1 + ρ2 , J1 + J2). (1.2)
The new solution may be called the superposition of the two original
solutions.

It is not hard to explain what is, in general, a linear equation or a

linear set of equations. Consider the equation

Lu = 0,
where, schematically, u denotes the unknown. The unknown may be a
number, or a function of time, a function of space, a function of time
and space, essentially anything unknown! In fact, u could represent a
collection of unknowns, in which case we would replace u above by
u1,u2,.... The symbol L denotes a linear operator, an object that satisfies
the following two properties

L(u1 + u2) = Lu1 + Lu2 , L(au) = aLu, (1.4)

where a is a number. Note that these conditions imply that

L(αu1 + βu2) = αLu1 + βLu2 , (1.5)
showing that if u1 is a solution ( Lu1 = 0) and u2 is a solution (Lu2 = 0)
then αu1 + βu2 is also a solution. We call αu1 +βu2 the general
superposition of the solutions u1 and u2. An example may help. Consider
the equation

,
where τ is a constant with units of time. This is, in fact, a linear
differential equation, and takes the form Lu = 0 if we define

3
Exercise 1. Verify that (1.7) satisfies the conditions for a linear
operator.

Einstein’s theory of general relativity is a nonlinear theory whose

dynamical variable is a gravitational field, the field that describes, for
example, how planets move around a star. Being a nonlinear theory,
you simply cannot add the gravitational fields of different solutions to
find a new solution. This makes Einstein’s theory rather complicated,
by all accounts much more complicated than Maxwell theory. In fact,
classical mechanics, as invented mostly by Isaac Newton, is also a
nonlinear theory! In classical mechanics the dynamical variables are
positions and velocities of particles, acted by forces. There is no
general way to use two solutions to build a third.

Indeed, consider the equation of motion for a particle on a line

under the influence of a timeindependent potential V (x), which is in
general an arbitrary function of x. The dynamical variable in this
problem is x(t), the position as a function of time. Letting V 0 denote
the derivative of V with respect to its argument, Newton’s second law
takes the form

.
The left-hand side is the mass times acceleration and the right hand
side is the force experienced by the particle in the potential. It is
probably worth to emphasize that the right hand side is the function V
0
(x) evaluated for x set equal to x(t):

While we could have used here an ordinary derivative, we wrote a

partial derivative as is commonly done for the general case of time
dependent potentials. The reason equation (1.8) is not a linear
equation is that the function V 0(x) is not linear. In general, for
arbitrary functions u and v we expect

V 0(au) 6= aV 0(u), andV 0(u + v) 6= V 0(u) + V (v).

4
As a result given a solution x(t), the scaled solution αx(t) is not
expected to be a solution. Given two solutions x1(t) and x2(t) then x1(t)
+ x2(t) is not guaranteed to be a solution either.
Exercise. What is the most general potential V (x) for which the
equation of motion for x(t) is linear?
Quantum mechanics is a linear theory. The signature equation in
this theory, the so-called Schr¨odinger equation is a linear equation
for a quantity called the wavefunction and it determines its time
evolution. The wavefunction is the dynamical variable in quantum
mechanics but, curiously, its physical interpretation was not clear to
Erwin Schr¨odinger when he wrote the equation in 1925. It was Max
Born, who months later suggested that the wavefunction encodes
probabilities. This was the correct physical interpretation, but it was
thoroughly disliked by many, including Schr¨odinger, who remained
unhappy about it for the rest of his life. The linearity of quantum
mechanics implies a profound simplicity. In some sense quantum
mechanics is simpler than classical mechanics. In quantum
mechanics solutions can be added to form new solutions.

The wavefunction Ψ depends on time and may also depend on

space. The Schr¨odinger equation (SE) is a partial differential
equation that takes the form

where the Hamiltonian (or energy operator) Hˆ is a linear operator

that can act on wavefunctions:

ˆ ˆ ˆ ˆ ˆ
H (aΨ) = aH Ψ, H (Ψ1 + Ψ2) = H (Ψ1) + H (Ψ2),

with a a constant that in fact need not be real; it can be a complex

number. Of course, Hˆ itself does not depend on the wavefunction! To
check that the Schr¨odinger equation is linear we cast it in the form
LΨ = 0 with L defined as

It is now a simple matter to verify that L is a linear operator. Physically

this means that if Ψ1 and Ψ2 are solutions to the Schr¨odinger

5
equation, then so is the superposition αΨ1 +βΨ2, where α and β are
both complex numbers, i.e. (α,β ∈ C)

Complex Numbers are Essential

Quantum mechanics is the first physics theory that truly makes use of
complex numbers. The numbers most of us use for daily life (integers,
fractions, decimals) are real numbers. The set of complex numbers is
denoted by C and the set of real numbers is denoted by R. Complex
numbers appear when we combine real numbers with the imaginary
unit i, defined to be equal to the square root of
√
minus one: i ≡ −1. Being the square root of minus one, it means that i
squared must give minus one: i2 = −1. Complex numbers are
fundamental in mathematics. An equation like x2 = −4, for an unknown
x cannot be solved if x has to be real. No real number squared gives
you minus one. But if we allow for complex numbers, we have the
solutions x = ±2i. Mathematicians have shown that all polynomial
equations can be solved in terms of complex numbers.

A complex number z, in all generality, is a number of the form

z = a + ib ∈ C, a,b ∈ R.

Here a and b are real numbers, and ib denotes the product of i with b.
The number a is called the real part of z and b is called the imaginary
part of z:

Rez = a, Imz = b. (2.

2)
The complex conjugate z∗ of z is defined by
(2.
z = a − ib.
∗
3)
You can quickly verify that a complex number z is real if z∗ = z and it is
purely imaginary if z∗ = −z. For any complex number z = a+ib one can
define the norm |z| of the complex number to be a positive, real number
given by
p
|z| = a2 + b2 .
You can quickly check that

|z|2 = zz∗ ,

6
where z∗ ≡ a − ib is called the complex conjugate of z = a + ib. Complex
numbers are represented as vectors in a two dimensional “complex
plane”. The real part of the complex number is the x component of
the vector and the imaginary part of the complex number is the y
component. If you consider the unit length vector in the complex
plane making an angle θ with the x axis has x component cosθ and y
component sinθ. The vector is therefore the complex number cos θ +
isinθ. Euler’s identity relates this to the exponential of iθ:
eiθ = cosθ + isinθ.

A complex number of the form eiχ, with χ real is called a pure phase.
While complex numbers are sometimes useful in classical
mechanics or Maxwell theory, they are not strictly needed. None of
the dynamical variables, which correspond to measurable quantities,
is a complex number. In fact, complex numbers can’t be measured at
all: all measurements in physics result in real numbers. In quantum
mechanics, however, complex numbers are fundamental. The
Schr¨odinger equation involves complex numbers. Even more, the
wavefunction, the dynamical variable of quantum mechanics it itself a
complex number:

Ψ ∈ C.

Since complex numbers cannot be measured the relation between the

wavefunction and a measurable quantity must be somewhat indirect.
Born’s idea to identify probabilities, which are always positive real
numbers, with the square of the norm of the wavefuntion was very
natural. If we write the wavefunction of our quantum system as Ψ, the
probabilities for possible events are computed from |Ψ|2. The
mathematical framework required to express the laws of quantum
mechanics consists of complex vector spaces. In any vector space we
have objects called vectors that can be added together. In a complex
vector space a vector multiplied by a complex number is still a vector.
As we will see in our study of quantum mechanics it is many times
useful to think of the wavefunction Ψ as a vector in some complex
vector space.

7
Loss of Determinism
Maxwell’s crowning achievement was the realization that his
equations of electromagnetism allowed for the existence of
propagating waves. In particular, in 1865 he conjectured that light
was an electromagnetic wave, a propagating fluctuation of electric
and magnetic fields. He was proven right in subsequent experiments.
Towards the end of the nineteenth century physicists were convinced
that light was a wave. The certainty, however, did not last too long.
Experiments on blackbody radiation and on the photo-emission of
electrons suggested that the behavior of light had to be more
complicated than that of a simple wave. Max Planck and Albert
Einstein were the most prominent contributors to the resolution of the
puzzles raised by those experiments.

In order to explain the features of the photoelectric effect, Einstein

postulated (1905) that in a light beam the energy comes in quanta –
the beam is composed of packets of energy. Einstein essentially
implied that light was made up of particles, each carrying a fixed
amount of energy. He himself found this idea disturbing, convinced
like most other contemporaries that, as Maxwell had shown, light was
a wave. He anticipated that a physical entity, like light, that could
behave both as a particle and as a wave could bring about the demise
of classical physics and would require a completely new physical
theory. He was in fact right. Though he never quite liked quantum
mechanics, his ideas about particles of light, later given the name
photons, helped construct this theory.

It took physicists until 1925 to accept that light could behave like a
particle. The experiments of Arthur Compton (1923) eventually
convinced most skeptics. Nowadays, particles of light, or photons, are
routinely manipulated in laboratories around the world. Even if
mysterious, we have grown accustomed to them. Each photon of
visible light carries very little energy – a small laser pulse can contain
many billions of photons. Our eye, however, is a very good photon
detector: in total darkness, we are able to see light when as little as
ten photons hit upon our retina. When we say that light behaves like a
particle we mean a quantum mechanical particle: a packet of energy
and momentum that is not composed of smaller packets. We do not
mean a classical point particle or Newtonian corpuscle, which is a
zero-size object with definite position and velocity.

8
 As it turns out, the energy of a photon depends only on the color of
the light. As Einstein discovered the energy E and frequency ν for a
photon are related by

E = hν

The frequency of a photon determines the wavelength λ of the light

through the relation νλ = c, where c is the speed of light. All green
photons, for example, have the same energy. To increase the energy
in a light beam while keeping the same color, one simply needs more
photons.

As we now explain, the existence of photons implies that Quantum

Mechanics is not deterministic. By this we mean that the result of an
experiment cannot be determined, as it would in classical physics, by
the conditions that are under the control of the experimenter.

Consider a polarizer whose preferential direction is aligned along

the xˆ direction, as shown in Figure 1. Light that is linearly polarized
along the xˆ direction namely, light whose electric field points in this
direction, goes through the polarizer. If the incident light polarization
is orthogonal to the xˆ direction the light will not go through at all.
Thus light linearly polarized in the yˆ direction will be totally absorbed
by the polarizer. Now consider light polarized along a direction
forming an angle α with the x-axis, as shown in Figure 2. What
happens?

Figure 1: A polarizer that transmits light linearly polarized along the xˆ

direction.

9
 Figure 2: Light linearly polarized along the direction at an angle α
hitting the polarizer.

Thinking of the light as a propagating wave, the incident electric

field Eα makes an angle α with the x-axis and therefore takes the form

Eα = E0 cosα xˆ + E0 sinα yˆ .

This is an electric field of magnitude E0. In here we are ignoring the

time and space dependence of the wave; they are not relevant to our
discussion. When this electric field hits the polarizer, the component
along xˆ goes through and the component along yˆ is absorbed. Thus

Beyond the polarizer: E = E0 cosα xˆ .

You probably recall that the energy in an electromagnetic wave is

proportional to the square of the magnitude of the electric field. This
means that the fraction of the beam’s energy that goes through the
polarizer is (cosα)2. It is also well known that the light emerging from
the polarizer has the same frequency as the incident light.

So far so good. But now, let us try to understand this result by

thinking about the photons that make up the incident light. The
premise here is that all photons in the incident beam are identical.
Moreover the photons do not interact with each other. We could even
imagine sending the whole energy of the incident light beam one
photon at a time. Since all the light that emerges from the polarizer
has the same frequency as the incident light, and thus the same
frequency, we must conclude that each individual photon either goes
through or is absorbed. If a fraction of a photon went through it would
be a photon of lower energy and thus lower frequency, which is
something that does not happen.

10
 But now we have a problem. As we know from the wave analysis,
roughly a fraction (cosα)2 of the photons must go through, since that
is the fraction of the energy that is transmitted. Consequently a
fraction 1 − (cosα)2 of the photons must be absorbed. But if all the
photons are identical, why is it that what happens to one photon does
not happen to all of them?

The answer in quantum mechanics is that there is indeed a loss of

determinism. No one can predict if a photon will go through or will get
absorbed. The best anyone can do is to predict probabilities. In this
case there would be a probability (cosα)2 of going through and a
probability 1 − (cosα)2 of failing to go through.

Two escape routes suggest themselves. Perhaps the polarizer is not

really a homogeneous object and depending exactly on where the
photon his it either gets absorbed or goes through. Experiments show
this is not the case. A more intriguing possibility was suggested by
Einstein and others. A possible way out, they claimed, was the
existence of hidden variables. The photons, while apparently identical,
would have other hidden properties, not currently understood, that
would determine with certainty which photon goes through and which
photon gets absorbed. Hidden variable theories would seem to be
untestable, but surprisingly they can be tested. Through the work of
John Bell and others, physicists have devised clever experiments that
rule out most versions of hidden variable theories. No one has figured
out how to restore determinism to quantum mechanics. It seems to
be an impossible task.

When we try to describe photons quantum mechanically we could

use wavefunctions, or equivalently the language of states. A photon
polarized along the xˆ direction is not represented using an electric
field, but rather we just give a name for its state:

(3.4) photon;x .

We will learn the rules needed to manipulate such objects, but for the
time being you could think of it like a vector in some space yet to be
defined. Another state of a photon, or vector is

photon;y,

11
representing a photon polarized along yˆ. These states are the
wavefunctions that represent the photon. We now claim that the
photons in the beam that is polarized along the direction α are in a
state photon;α that can be written as a superposition of the above
two states:

photon; photon; photon;y .

This equation should be compared with (3.2). While there are some
similarities –both are superpositions– one refers to electric fields and
the other to “states” of a single photon. Any photon that emerges from
the polarizer will necessarily be polarized in the xˆ direction and
therefore it will be in the state

Beyond the polarizer: photon;x.

This can be compared with (3.3) which with the factor cos α carries
information about the amplitude of the wave. Here, for a single
photon, there is no room for such a factor.
In the famous Fifth Solvay International Conference of 1927 the
world’s most notable physicists gathered to discuss the newly
formulated quantum theory. Seventeen out of the twenty nine
attendees were or became Nobel Prize winners. Einstein, unhappy
with the uncertainty in quantum mechanics stated the nowadays
famous quote: “God does not play dice”, to which Niels Bohr is said to
have answered: “Einstein, stop telling God what to do.” Bohr was
willing to accept the loss of determinism, Einstein was not.

Quantum Superpositions
We have already discussed the concept of linearity; the idea that the
sum of two solutions representing physical realities represents a new,
allowed, physical reality. This superposition of solutions has a
straightforward meaning in classical physics. In the case of
electromagnetism, for example, if we have two solutions, each with
its own electric and magnetic field, the “sum” solution is simply
understood: its electric field is the sum of the electric fields of the two
solutions and its magnetic field is the sum of the magnetic fields of
the two solutions. In quantum mechanics, as we have explained,

12
linearity holds. The interpretation of a superposition, however, is very
surprising.

One interesting example is provided by a Mach-Zehnder

interferometer; an arrangement of beam splitters, mirrors, and
detectors used by Ernst Mach and Ludwig Zehnder in the 1890’s to
study interference between two beams of light.

A beam splitter, as its name indicates, splits an incident beam into

two beams, one that is reflected from the splitter and one that goes
through the splitter. Our beam-splitters will be balanced: they split a
given beam into two beams of equal intensity (Figure 3). The light
that bounces off is called the reflected beam, the light that goes
through is called the transmitted beam. The incident beam can hit the
beam splitter from the top or from the bottom.

The Mach-Zehnder configuration, shown in Figure 4, has a left

beam splitter (BS1) and a right beam splitter (BS2). In between we
have the two mirrors, M1 on the top and M2 on the bottom. An
incoming beam from the left is split by BS1 into two beams, each of
which hits a mirror and is then sent into BS2. At BS2 the beams are
recombined and sent into two outgoing beams that go into photon
detectors D0 and D1.

It is relatively simple to arrange the beam-splitters so that the

incoming beam, upon splitting at BS1 and recombination at BS2
emerges in the top beam which goes into D0. In this arrangement no
light at all goes into D1. This requires a precise interference effect at
BS2. Note that we have

Figure 3: An incident beam hitting a beam-splitter results in a reflected beam and

a transmitted beam. Left: incident beam coming from the top. Right: incident beam
coming from the bottom.

13
Figure 4: A Mach-Zehnder interferometer consists of two beam splitters BS1 and
BS2, two mirrors M1 and M2, and two detectors D0 and D1. An incident beam will
be split into two beams by BS1. One beam goes through the upper branch, which
contains M1, the other beam goes through the lower branch, which contains M2.
The beams on the two branches recombine at BS2 and are then sent into the
detectors. The configuration is prepared to produce an interference so that all
incident photons end at the detector D0, with none at D1.

two beams incident upon BS2; the top beam is called ‘a’ and the
lower beam is called ‘b’. Two contributions go towards D0: the
reflection of ‘a’ at BS2 and the transmission from ‘b’ at BS2. These
two contributions interfere constructively to give a beam going into
D0. Two contributions also go towards D1: the transmission from ‘a’ at
BS2 and the reflection from ‘b’ at BS2. These two can indeed be
arranged to interfere destructively to give no beam going into D1.
It is instructive to think of the incoming beam as a sequence of
photons that we send into the interferometer, one photon at a time.
This shows that, at the level of photons, the interference is not
interference of one photon with another photon. Each photon must
interfere with itself to give the result. Indeed interference between two
photons is not possible: destructive interference, for example, would
require that two photons end up giving no photon, which is impossible
by energy conservation.

Therefore, each photon does the very strange thing of going

through both branches of the interferometer! Each photon is in a
superposition of two states: a state in which the photon is in the top
beam or upper branch, added to a state in which the photon is in the
bottom beam or lower branch. Thus the state of the photon in the

14
interferometer is a funny state in which the photon seems to be doing
two incompatible things at the same time.

Equation (3.6) is another example of a quantum superposition. The

photon state has a component along an x-polarized photon and a
component along a y-polarized photon.

When we speak of a wavefunction, we also sometimes call it a

state, because the wavefunction specifies the “state” of our quantum
system. We also sometimes refer to states as vectors. A quantum
state may not be a vector like the familiar vectors in three-
dimensional space but it is a vector nonetheless because it makes
sense to add states and to multiply states by numbers. Just like
vectors can be added, linearity guarantees that adding wavefunctions
or states is a sensible thing to do. Just like any vector can be written
as a sum of other vectors in many different ways, we will do the same
with our states. By writing our physical state as sums of other states
we can learn about the properties of our state.
Consider now two states and . Assume, in addition, that when
measuring some property Q in the state the answer is always a,
and when measuring the same property Q in the state the answer is
always b. Suppose now that our physical state |Ψi is the superposition

What happens now if we measure property Q in the system described

by the state |Ψi? It may seem reasonable that one gets some
intermediate value between a and b, but this is not what happens. A
measurement of Q will yield either a or b. There is no certain answer,
classical determinism is lost, but the answer is always one of these
two values and not an intermediate one. The coefficients α and β in
the above superposition affect the probabilities with which we may
obtain the two possible values. In fact, the probabilities to obtain a or
b

Probability(a) ∼ |α|2 ,Probability(b) ∼ |β|2 .

Since the only two possibilities are to measure a or b, the actual

probabilities must sum to one and therefore they are given by
2

Probability(, Probability(.

15
 If we obtain the value a, immediate repeated measurements would
still give a, so the state after the measurement must be |Ai. The same
happens for b, so we have

After measuring a the state becomes |Ψi = |Ai,

(4.4)
After measuring b the state becomes |Ψi = |Bi.

In quantum mechanics one makes the following assumption:

Superposing a state with itself doesn’t chance the physics, nor does it change
the state in a non-trivial way. Since superimposing a state with itself
simply changes the overall number multiplying it, we have that Ψ and
αΨ represent the same physics for any complex number α different
∼
from zero. Thus, letting = represent physical equivalence

This assumption is necessary to verify that the polarization of a

photon state has the expected number of degrees of freedom. The
polarization of a plane wave, as one studies in electromagnetism, is
described by two real numbers. For this consider an elliptically
polarized wave, as shown in Figure 5. At any given point, the electric
field vector traces an ellipse whose shape is encoded by the ratio a/b
of the semi-major axes (the first real parameter) and a tilt encoded by
the angle θ (the second real parameter). Consider for this a general
photon state formed by superposition of the two independent
polarization states |photon;xi and |photon;yi:

α|photon;xi + β|photon;yi, α,β ∈ C.

At first sight it looks as if we have two complex parameters α and β,

or equivalently, four real parameters. But since the overall factor
does not matter we can multiply this state by 1/α to get the
equivalent state that encodes all the physics

|photon; photon;yi,

showing that we really have one complex parameter, the ratio β/α.
This is equivalent to two real parameters, as expected.

16
 Figure 5: Parameters that define an elliptically polarized state.

Let us do a further example of superposition using electrons.

Electrons are particles with spin. Classically, we imagine them as tiny
balls spinning around an axis that goes through the particle itself.
Once an axis is fixed, the electron has two and only two options: its
rotation may be clockwise or counterclockwise about the axis, but in
both cases it spins at the same fixed rate. These opposite ways of
spinning are called spin up and spin down along the axis (see Figure 6).
The up and down refer to the direction of the angular momentum
associated with the rotation, and it is indicated by an arrow. According
to quantum mechanics, and as verified by multiple experiments, the
same possibilities, up or down, arise whatever axis we use to measure
the spin of the electron.

Physicists usually set up coordinate systems in space by choosing

three orthogonal directions, the directions of the x, y, and z axes. Let
us choose to describe our spinning electrons using the z axis. One
possible state of an electron is to be spin up along the z axis. Such a
state is described as | ↑;zi, with an arrow pointing up, and the label z
indicating that the spin arrow points along the increasing z direction.
Another possible state of an electron is spin down along the z axis.
Such a state is described as | ↓;zi, with an arrow pointing down,
meaning this time that the spin points along the decreasing z
direction. If these two are possible realities, so it would be the state |
Ψi representing the sum

|Ψi = | ↑;zi + | ↓;zi.

The state |Ψi is in a superposition of a spin up and a spin down state.

What kind of physics does this sum |Ψi represent? It represents a
state in which a measurement of the spin along the z axis would result
in two possible outcomes with equal probabilities: an electron with

17
spin up or an electron with spin down. Since we can only speak of
probabilities, any experiment must involve repetition until
probabilities can be determined. Suppose we had a large ensemble of
such electrons, all of them in the above state |Ψi. As we measured
their spin along z, one at a time, we would find about half of them
spinning up along z and the other half spinning down along z. There is
no way to predict which option will be realized as we measure each
electron. It is not easy to imagine superposition, but one may try as
follows. An electron in the above state is in a different kind of
existence in which it is able to both be spinning up along z and
spinning down along z simultaneously! It is in such a ghostly, eerie
state, doing incompatible things simultaneously, until its spin is
measured. Once measured, the electron must immediately choose
one of the two options; we always find electrons either spinning up or
spinning down.

Figure 6: An electron with spin along the z axis. Left: the electron is said to have
spin up along z. Right: the electron is said to have spin down along z. The up and
down arrows represent the direction of the angular momentum associated with the
spinning electron.

A critic of quantum mechanics could suggest a simpler explanation

for the above observations. He or she would claim that the following
simpler ensemble results in identical experimental results. In the
critic’s ensemble we have a large number of electrons with 50% of
them in the state | ↑;zi and 50% of them in the state | ↓;zi. He or she
would then state, correctly, that such an ensemble would yield the
same measurements of spins along z as the ensemble of those
esoteric |Ψi states. The new ensemble could provide a simpler

18
explanation of the result without having to invoke quantum
superpositions.

Quantum mechanics, however, allows for further experiments that

can distinguish between the ensemble of our friendly critic and the
ensemble of |Ψi states. While it would take us too far afield to explain
this, if we measured the spin of the electrons in the x direction,
instead of z direction, the results would be different in the two
ensembles. In the ensemble of our critic we would find 50% of the
electrons up along x and 50% of the electrons down along x. In our
ensemble of |Ψi states, however, we would find a very simple result:
all states pointing up along x. The critic’s ensemble is not equivalent
to our quantum mechanical ensemble. The critic is thus shown wrong
in his or her attempt to show that quantum mechanical
superpositions are not required.

Entanglement
When we consider superposition of states of two particles we can get
the remarkable phenomenon called quantum mechanical entanglement.
Entangled states of two particles are those in which we can’t speak
separately of the state of each particle. The particles are bound
together in a common state in which they are entangled with each
other.
Let us consider two non-interacting particles. Particle 1 could be in
any of the states

while particle 2 could b in any of the statese

It may seem reasonable to conclude that the state of the full system,
including particle 1 and particle 2 would be specified by stating the
state of particle 1 and the state of particle 2. If that would be the case
the possible states would be written as

for some specific choice of i and j that specify the state of particle one
and particle two, respectively. Here we have used the symbol ⊗,

19
which means tensor product, to combine the two states into a single
state for the whole system. We will study ⊗ later, but for the time
being we can think of it as a kind of product that distributes over
addition and obeys simple rules, as follows

(5.4)
.

The numbers can be moved across the ⊗ but the order of the states
must be preserved. The state on the left-hand side –expanded out on
the right-hand side– is still of the type where we combine a state of
the first particle ( ) with a state of the second particle (
). Just like any one of the states listed in (5.3) this state is
not entangled.

Using the states in (5.3), however, we can construct more

intriguing superpositions. Consider the following one

.
A state of two particles is said to be entangled if it cannot be written in
the factorized form (···)⊗(···) which allows us to describe the state by
simply stating the state of each particle. We can easily see that the
state (5.5) cannot be factorized. If it could it would have to be with a
product as indicated in (5.4). Clearly, involving states like |u3i or |v3i
that do not appear in (5.5) would not help. To determine the constants
α1,α2,β1,β2 we compare the right hand side of (5.4) with our state and
conclude that we need

α1β1 = 1, α1β2 = 0, α2β1 = 0, α2β2 = 1.

It is clear that there is no solution here. The second equation, for

example, requires either α1 or β2 to be zero. Having α1 = 0 contradicts
the first equation, and having β2 = 0 contradicts the last equation.
This confirms that the state (5.5) is indeed an entangled state. There
is no way to describe the state by specifying a state for each of the
particles.

Let us illustrate the above discussion using electrons and their spin
states. Consider a state of two electrons denoted as | ↑i ⊗ | ↓i. As the

20
notation indicates, the first electron, described by the first arrow, is
up along z while the second electron, described by the second arrow,
is down along z (we omit the label z on the state for brevity). This is
not an entangled state. Another possible state is one where they are
doing exactly the opposite: in | ↓i ⊗ | ↑i the first electron is down and
the second is up. This second state is also not entangled. It now
follows that by superposition we can consider the state
| ↑i ⊗ | ↓i + | ↓i ⊗ | ↑i.
This is a entangled state of the pair of electrons.
Exercise. Show that the above state cannot be factorized and thus is
indeed entangled.

In the state (5.7) the first electron is up along z if the second

electron is down along z (first term), or the first electron is down along
z if the second electron is up along z (second term). There is a
correlation between the spins of the two particles; they always point
in opposite directions. Imagine that the two entangled electrons are
very far away from each other: Alice has one electron of the pair on
planet earth and Bob has the other electron on the moon. Nothing we
know is connecting these particles but nevertheless the states of the
electrons are linked. Measurements we do on the separate particles
exhibit correlations. Suppose Alice measures the spin of the electron
on earth. If she finds it up along z, it means that the first summand in
the above superposition is realized, because in that summand the
first particle is up. As discussed before, the state of the two particles
immediately becomes that of the first summand. This means that the
electron on the moon will instantaneously go into the spin down-along-z
configuration, something that could be confirmed by Bob, who is
sitting in the moon with that particle in his lab. This effect on Bob’s
electron happens before a message, carried with the speed of light,
could reach the moon telling him that a measurement has been done
by Alice on the earth particle and the result was spin up. Of course,
experiments must be done with an ensemble that contains many
pairs of particles, each pair in the same entangled state above. Half
of the times the electron on earth will be found up, with the electron
on the moon down and the other half of the times the electron on
earth will be found down, with the electron on the moon up.

Our friendly critic could now say, correctly, that such correlations
between the measurements of spins along z could have been

21
produced by preparing a conventional ensemble in which 50% of the
pairs are in the state | ↑i ⊗ | ↓i and the other 50% of the pairs are in the
state | ↓i ⊗ | ↑i. Such objections were dealt with conclusively in 1964 by
John Bell, who showed that if Alice and Bob are able to measure spin
in three arbitrary directions, the correlations predicted by the quantum
entangled state are different from the classical correlations of any
conceivable conventional ensemble. Quantum correlations in
entangled states are very subtle and it takes sophisticated
experiments to show they are not reproducible as classical
correlations. Indeed, experiments with entangled states have
confirmed the existence of quantum correlations. The kind of
instantaneous action at a distance associated with measurements on
well-separated entangled particles does not lead to paradoxes nor, as
it may seem, to contradictions with the ideas of special relativity. You
cannot use quantum mechanical entangled states to send information
faster than the speed of light.

Sarah Geller transcribed Zwiebach’s handwritten notes to create the first LaTeX
version of this document.

22
MIT OpenCourseWare
https://ocw.mit.edu

8.04 Quantum Physics I

Spring 2016

For information about citing these materials or our Terms of Use, visit: https://ocw.mit.edu/terms.

23