PHY 314: Introduction to Quantum Mechanics, Varsha 2013

Lecture 2
Anil Shaji
School of Physics, IISER TVM
The Stern-Gerlach experiment. The spin quantum number (to be revisited later). The
Hilbert space and some general properties of linear vector spaces. Rays and vectors in Hilbert
space. Normalization. Basis vectors.

We have met the quantum state vector formally and also talked about the so-called quantum two
state system that really has an infinity of allowed states. Now let us talk about a real experiment
in which a quantum two state system appears and what it does.

I.

THE STERN-GERLACH EXPERIMENT

A schematic diagram of the experiment is given below (Taken from J. J .Sakurais book)

A stream of silver atoms emanating from an oven go through an inhomogeneous magnetic field
oriented along the z-direction. Silver atoms have a nucleus and 47 electrons. Now both the electrons
and nucleus have what is called spin or intrinsic angular momentum. This is in addition to orbital
angular momentum. The 46 electrons together have no net angular momentum. For our purposes
we can imagine that the nucleus does not have any angular momentum either. The only angular
momentum for the whole atom comes from the intrinsic angular momentum of the 47th electron.
This makes it a relatively simple system to analyze and that was why silver atoms were chosen in

2
the original experiment. The magnetic moment of the whole atom is also proportional to the spin
angular momentum:
~

~ S
The interaction energy of the atom in the presence of the magnetic field is
~
E = ~
B
If the magnetic field is oriented along the z-axis and if there is a magnetic field gradient then the
force along the z-axis felt by each atom is
Fz =

E
Bz
= z
z
z

Since the atom is rather heavy we can assume that its trajectory in the magnetic field is very
much like that of a classical particle. The initial orientation of the magnetic moment is fairly
random on coming out of the oven. So each atom has a different value for the z-component of the
magnetic moment, z and should therefore feel different forces along the z-axis. If Sz is positive,
the atoms experience a positive force and if Sz is negative they experience a One would therefore
expect the positions of the atoms hitting the collection plate to be anywhere along a line with most
atoms concentrated at the center.
When Stern and Gerlach actually did the experiment, what they fount, much to their surprise
was two spots on the collection plane. The atoms that could have been states with arbitrary
orientation for the angular momentum, chose to be in one of the two states |upi and |downi. Or
in a notation more familiar to us, they chose to be in the states |0i and |1i.
This experiment shows that two state quantum systems do exist. One can also compute the
value of the spin angular momentum for the atoms knowing the strength of the magnetic field and
the final location of the atoms on the collection plate. It turns out that
h

Sz = ,
2

h =

h
,
2

h = 6.626 1034 J s

Now we encounter a new and important concept in quantum mechanics.

Classically the z-

~ and
component of the angular momentum of the atom could have had any value between |S|
~ but quantum mechanically only a discrete set of values (in this case two) are observed. We
|S|
say that the angular momentum is quantized along the z-direction. The scale of the quantization
is set by the constant h that we now know as Plancks constant. Max Planck being the first person
to come across this constant in a different context thereby starting the quantum revolution.

3
Now the z-direction is nothing special. If Stern and Gerlach had oriented their magnetic field
along some other direction they they would have found two spots on the collecting plate along
a line parallel to that direction. We also have seen previously that according to the principle of
superposition, if there are two allowed states for a quantum system, then any state of the form
|i = cos |0i + sin |1i
is also an allowed state. How can we find these states in the silver atom experiment?
We can find these states by extending the experiment. Suppose we block one of the beams of
atoms (say the one with Sz =
h/2 and then taking the remaining atoms with Sz = h
/2 and pass
them through another inhomogeneous magnetic field that is oriented again along the z-axis. We
find that there is only one spot in the collection plate as expected. So we conclude that since we
picked out only those atoms in the state |Sz = h
/2i, all of them are deflected equally by the next
magnetic field, again along the z-axis.
Things get more interesting if we now orient the second magnetic field along the x-direction.
We again start getting two beams. Does that mean that the initial beam is made of half the atoms
in the state |Sz = h/2; Sx = h/2i and half the atoms in the state |Sz = h/2; Sx = h/2i? This
picture runs into difficulty real soon. Suppose we pick one of the beams that come out of the
x-oriented magnetic field and apply a third magnetic field along the z-direction we should then
expect only one beam at the output. In reality two beams appear again. So how do we reconcile
all this. To do so we have to write
1
|Sx = h
/2i = (|Sz = h
/2i + |Sz = h/2i)
2
1
|Sx =
h/2i = (|Sz = h
/2i |Sz = h/2i)
2
This automatically means that
1
|Sz = h
/2i = (|Sx = h
/2i + |Sx = h/2i)
2
1
|Sz =
h/2i = (|Sx = h
/2i |Sx = h/2i)
2
In fact
|S =
h/2i = cos |Sz = h
/2i sin |Sz = h/2i

4
The Stern-Gerlach experiment also introduces a new quirk of quantum mechanics. Namely, what
decides which one of the two paths the atom takes? The answer is that the quantum mechanics
only allows you to compute the probability of each atom taking one or the other of the two paths.
It does not tell you which path a particular atom might choose to follow. It does, however, let
you compute the fraction of atoms that end up in one path or the other if we started with a large
group or ensemble of atoms. So we see that quantum mechanics makes probabilistic predictions
and is in its very nature a probabilistic theory. We will see in the next lecture how the different
probabilities are computed.

II.

THE HILBERT SPACE

The state vectors live in Hilbert space and Hilbert space is a complex vector space. What does
it mean to say a complex vector space? Operationally it means that if we have two vectors |1 i
and |2 i then we can add them together to form a third vector,
|3 i = c1 |1 i + c2 |2 i
with c1 and c2 being complex number rather than being real numbers. The dimensionality of the
vector space depends on the property of the system that were interested in.
The basic properties of the ket vector are
1. Kets can be added together
2. If we multiply a ket by a complex number the resulting ket is also a ket (c|psii).
3. If c is zero then the resulting ket is called the null ket or null state
4. A physical postulate is the |psi1 i and c|1 i represent the same state. Mathematically one
would say that states are rays (and not vectors) in Hilbert space in that only the direction
of the state vector matters
We also postulate that there is a dual space of vectors with a one to one correspondence with the
ket vectors. In other words for every ket |i we identify a unique dual, h|. The dual correspondence
is such that
|1 i + |2 i h1 | + h2 |
c1 |1 i + c2 |2 i c1 h1 | + c2 h2 |

5
Now we can define an inner product h1 |2 i just as one would define an inner (dot) product
of two ordinary vectors. The inner product is usually a complex number with h1 |2 i = h2 |1 i .
This also means h|i is real and we also require it to be greater than or equal to zero (positive
semi-definite).
Two kets are orthogonal if h1 |2 i = 0.
Since the magnitude of the vectors do not matter and only the direction matters we can make

the length of each state vector one. Given an arbitrary ket |i we can define a normalized ket |i
p
|i

1
|i.
h|i

It is easy to verify that

h|i is the norm of the ket |i. Making the norm unity is normalization.

We can identify a minimal set of ket vectors |1 i, |2 i . . . |n i in terms of which any other vector
in the Hilbert space can be written. This set of vectors can be called a basis for the Hilbert space.
It is often desirable that the basis vectors be normalized and also that they be orthogonal to each
other hi |j i = 0 if i 6= j and hi |i i = 1
We can now explore other ways of writing states. If we choose our basis in the Stern-Gerlach
experiment as being the two states along the z-axis, namely |Sz = h/2i then we can, for instance
write |Sx =
h/2i as

1 1
|Sx = h
/2i = ,
2 1
and so on.

1 1
|Sx = h/2i =

2 1