1 Heisenberg Uncertainty Principle

1 Heisenberg Uncertainty Principle
In this section, we give a brief derivation and discussion of Heisenberg s uncertainty

principle. The derivation follows closely that given in Von Neumanns Mathematical
Foundations of Quantum Mechanics pp 230-232. The system we deal with is one di-
mensional with coordinate X ranging , +. The uncertainty principle is a direct
consequence of the commutation rule
[X, P] = i h (1)
While we need this operator equation to derive a concrete result, the general idea is
present in any system where there are plane waves. A physically realizable wave is
always in the form of a wave packet which is nite in extent. A wave packet is built up
by superposing waves with denite wave number. By simple Fourier analysis, a highly
localized packet will require a wide spread of wave vectors, whereas a packet with a large
spacial extent can be composed of wave numbers quite close to a specic value.
We will carry out the derivation, following Von Neumann, using standard Hilbert
space notation instead of Dirac notation. For expected values we write
< P >

P = (, P), and < X >

X = (, X).
Recall that the length of a vector in Hilbert space is dened by
||||
(, )
We will also need the uncertainty in X and P. For any observable O, the standard
denition of uncertainty is
O
(, (O
2

O
2
)),
which is a measure of the spread of values of O around its mean or expected value.
To make progress in the derivation, we need an expression which can be manipulated
into one involving [X, P]. The expression used by Von Neumann is (X, P). Using the
fact that X and P are self-adjoint, can write
2iIm[(X, P)] = (X, P) (P, X) (2)
Now we using self-adjointness again, we can move all operators to the right slot, giving
(X, P) (P, X) = (, XP) (, PX).
Next we use the X, P commutator and we have
(X, P) (P, X) = (, [X, P]) = i h(, ) = i h, (3)
1
where in the last equality we have assumed a normalizable state with (, ) = 1. Com-
paring Eqs.(2) and (3) we now have
Im[(X, P)] =
h
2
(4)
So far we have made no approximations. Eq.(4) is exact. To get the uncertainty relation,
we now make use of some inequalities. First we make use of the fact that the imaginary
part of a complex number is less than the absolute value of the complex number. For us
this says
Im[(X, P)] |(X, P)|,
which implies using Eq.(4) that
|(X, P)|
h
2
(5)
We now have an inequality involving X and P which came directly from the commutator
Eq.(1), but it involves X and P in the same matrix element. To separate them, we now
use the Schwarz inequality, which states
|(, )| |||| ||||,
which is the Hilbert space version of the familiar statement that the dot product of two
vectors is the product of their lengths. For us, the vectors we will apply the Schwarz
inequality to are X, and P. We have
|(X, P)| ||X|| ||P||,
so we now have
||X|| ||P||
h
2
(6)
This is almost what we want, but makes no reference to the expected values

X and

P,
both of which come into their respective uncertainties. To remedy this we dene
X
= X

XI, P
= P

PI.
These operators satisfy the basic commutation rule,
[X
, P
] = i h.
All the steps in our derivation hold if we replace X by X
, and P by P
, This allows us
to write
||X
|| ||P
||
h
2
. (7)
This really is the uncertainty principle. To see this, we write
||X
||
2
= (X
, X
) = (, X
) = (, (X
2
2X

X +

X
2
))
2
= (, (X
2

X
2
)) = (X)
2
,
and likewise for ||P
||
2
. Using these results we nally have
X P
h
2
This completes the mathematical derivation of the uncertainty principle. There is a very
nice physical discussion of it in Heisenbergs book The Physical Principles of Quantum
Theory around page 20. As a historical note, Heisenberg only proved his principle for
gaussian wave functions. The full Hilbert Space derivation given above was apparently
rst done by Hermann Weyl. Below is a photo of Heisenberg in his prime.
Figure 1: Young Heisenberg
Heisenbergs Measurement Uncertainty The above discussion does not say a word
about measurements. It is instead a limit on the possible properties of a given state.
Roughly, if < x| > has a narrow distribution, then < p| > will have a spread out
distribution and vice versa. Heisenbergs original paper on uncertainty concerned a much
more physical picture. The example he used was that of determining the location of an
electron with an uncertainty x, by having the electron interact with X-ray light. For an
X-ray of wavelength , the best that can be done is
x .
But if an X-ray photon scatters from an electron, it will disturb the electrons momentum
by an amount p. We expect p to be of order the X-ray photons momentum, i.e.
3
p h/. The net result is that the product (x)(p) should be of order Plancks constant,
or
(x)(p) h.
(We are not following factors of 2 or at this point.)
A question that has been studied ever since Heisenbergs original paper is whether a
precise inequality can be established for the case where actual measurements are made
on a quantum system. As of October, 2013, this has not been denitively settled, but we
can discuss the general framework presently being used to address the issue. We imagine
that we have a quantum system S. We take S to have a coordinate X
s
and conjugate
momentum P
s
, which satisfy
[X
s
, P
p
] = i h.
The measurement of say X
s
is done by coupling the quantum systemS to another system,
the probe P. In real experiments, P will have many degrees of freedom. The nal result
of the interaction of P with S will be a record or reading. This nal stage is classical,
no issue of non-commuting operators or quantum uctuations is involved at the end, i.e.
the meter reads a certain number, period.
Theoretical analysis of measurement uncertainty involving a system S and a probe P
have so far not included the case where P has many degrees of freedom, the last stages
of which are classical. Instead the probe is modeled as a second quantum system, whose
quantum variables are somehow accessible to observation, and which can give information
about the quantum variables of S. This can be illustrated by denoting the S quantum
variables by X
s
and P
s
, and the P variables by X
p
, P
p
.
The interaction of system and probe is described by an interaction or coupling Hamil-
tonian, H
c
, which acts over a short time interval, T. Since the evolution of the total system
is unitary, all operators at time T can be expressed in terms of the operators at time
0, the time at which H
c
is turned on. The interaction Hamiltonian, H
c
is chosen so
< X
p
(T) >=< X
s
(0) >, i.e. by making a measurement of X
p
at time T, we can deduce
information about the operator X
s
at time 0, i.e just before the coupling is turned on.
There is inevitably some noise or measurement error introduced, and the so-called noise
operator is dened for this case as
N = X
p
(T) X
s
(0).
As a measure of the noise, the uncertainty or standard deviation of N is used,
(N) =
< (X
p
(T) X
s
(0))
2
>
In general, the momentum P
s
of the system will be modied or disturbed by the inter-
action of S with P. The disturbance operator is dened by
D = P
s
(T) P
s
(0).
The measure of disturbance is then
(D) =
< (P
s
(T) P
s
(0))
2
>.
4
The question then becomes, is
((N))((D))
h
2
?
No denitive answer is available at present. The discussion can be traced by looking at
the article Proof mooted for Quantum Uncertainty, in Nature vol. 498, p 419, June
27, 2013.
Even if this controversy is nally resolved, it is clear that the above description of
the probe is too simple. The case of a probe with many degrees of freedom, the nal
stages of which are classical or macroscopic, remains to be discussed in a satisfying way.
Heisenbergs uncertainty is likely to remain an alive subject for some time.
5

1 Heisenberg Uncertainty Principle

Uploaded by

Copyright:

Available Formats

1 Heisenberg Uncertainty Principle

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 Heisenberg Uncertainty Principle

Uploaded by

Copyright:

Available Formats

1 Heisenberg Uncertainty Principle

In this section, we give a brief derivation and discussion of Heisenberg s uncertainty

You might also like