QM Identical Particles Problems
QM Identical Particles Problems
QM Identical Particles Problems
iI
c
i
[i (3.1)
for some constants c
i
C. In this way, an arbitrary state is a linear combination of some particular set of
states; this is called superposition. From now on well work entirely in terms of basis states.
We want a way to be able to check if some state [ has nonzero [i part, so we dene the dual object
to the ket [i, the bra i[. Bras act on kets through the basis: if i, j I, then
i[j
ij
=
_
1 i = j
0 i ,= j,
(3.2)
where
ij
is the Kronecker delta symbol. This says that the state [i has no [j part the overlap is zero;
we will assume this is true of all basis states. If the system is in the state [, the probability amplitude to
observe the state [i is i[:
i[ = i[
jI
c
j
[j =
jI
c
j
i[j = c
i
. (3.3)
We also dene the operation of switching between bras and kets, called Hermitian conjugation (denoted by
a superscripted ):
[ = [[]
=
_
iI
c
i
[i
_
iI
c
i
i[, (3.4)
where c
i
is the complex conjugate of c
i
. Similarly, [[]
i :
i I
n
, so that an arbitrary
n-lacrosse ball state [ 1
n
is
[ =
iI
n
c
i
[
i =
i1I
i2I
. . .
inI
c
i1i2...in
[i
1
, i
2
, . . . i
n
, (3.5)
2
JHMT 2014 Power round Identical Quantum Particles February 15, 2014
where Ive expanded everything out in the last step to remind you of whats going on. The action of multi-
particle bras follows straightforwardly from the single-particle case: If I want to the overlap between [x
and [y, then I need the overlaps between x
1
and y
1
, x
2
and y
2
, and so forth. The checks are independent
(because theyre not actually interfering with the state), so the probability amplitude is multiplicative:
x[y =
n
i=1
x
i
[y
i
. (3.6)
3. (a) Write out [ for the state [ in equation (3.5), in terms of the coecients c
i
.
(b) Operators like permutations also have Hermitian conjugates;
is dened as follows:
[[x]
= [[(x)]
= (x)[ = x[
. (3.7)
Determine the state [
(y) =
[y by calculating (x)[y.
4 Symmetry and Multi-Particle States
Let my physical system be a set of three boxes, and say I have two lacrosse balls; the states of my multi-particle
system will be the dierent ways I can place the lacrosse balls into the boxes. So [0, 3 denotes the state in
which ball 1 isnt in any of the boxes, and ball 2 is in box 3.
This state would be perfectly acceptable classically: if I had two lacrosse balls and three boxes, I could construct
this state. But since [3, 0 the state with ball 1 in box 3 and ball 2 outside the boxes isnt the same as
[0, 3 (the balls are dierent, and 3, 0[0, 3 = 0), the state [0, 3 is not a possible quantum state for electrons.
Using the notation weve developed, since (1, 2)[0, 3 , = [0, 3, the behavior of the system is not invariant under
particle interchanges. But I can construct a state out of [0, 3 and [3, 0 that doesnt change when I switch
particles 1 and 2:
[
+
=
1
2
([0, 3 +[3, 0). (4.1)
Here, the 1/
2 sets
+
[
+
= 1 to preserve probability and is called a normalization factor. By adding
states together, I can create symmetric combinations. Interestingly, in quantum mechanics probabilities are
actually determined by the complex modulus squared of the probability amplitude: so the probability to nd
the conguration [3, 0 in the state [
+
is actually
P
+([3, 0) = 3, 0[
+
3, 0[
+
. (4.2)
This means that the state
[
=
1
2
([0, 3 [3, 0), (4.3)
which receives an overall minus sign upon interchange of particles 1 and 2, is also acceptable! For all the
computable probabilities (i.e., the physics) depend on the square of the state and are hence unaected by
overall minus signs in the states themselves. Which sign we use depends on the kind of particle were dealing
with. Fermions particles like electrons and protons actually use the minus sign, while bosons like
photons (which make up light), helium nuclei, and the Higgs boson use the plus sign. Bosons are easier, so
well consider them rst:
If I have n Higgs bosons, then the state [x = [x
1
, x
2
, . . . , x
n
is not acceptable quantum mechanically (unless
x
1
= x
2
= = x
n
). To construct an acceptable state out of [x, I need to add to it all possible states with the
state labels interchanged, like [x
2
, x
1
, . . . , x
n
and [x
n
, x
n1
, . . . , x
1
. To take into account all possible states,
we use the symmetrizer operator S
(n)
:
S
(n)
=
Sn
, (4.4)
where the sum is over all permutations S
n
. For n = 2, the symmetrizer is just
S
(2)
=
S2
= (1, 2) + (2, 1). (4.5)
3
JHMT 2014 Power round Identical Quantum Particles February 15, 2014
4. Show that the state
S
(n)
[x =
Sn
[x (4.6)
is invariant under exchanges of particles by proving S
(n)
[x = S
(n)
[x for any S
n
.
5. Show that S
2
(n)
= C
+
S
(n)
for some constant C
+
, and nd C
+
. (From now on I will redene S
(n)
C
1/2
+
S
(n)
, so that S
2
(n)
= S
(n)
.)
6. Suppose I want to describe the state in which n
1
bosons have state label x
1
, n
2
particles have state label
x
2
, and so on (with all the x
i
I), so that I have n =
i=1
n
i
bosons that possess dierent state labels.
So a corresponding ket might be
[ = [ x
1
, x
1
. . . , x
1
. .
n1 times
, x
2
, x
2
, . . . x
2
. .
n2 times
, . . . x
, x
. . . , x
. .
n
times
. (4.7)
Determine the normalization constant N
+
R
+
such that [
+
= N
+
S
(n)
[ satises
+
[
+
= 1.
For fermions, I need each interchange of two particles to produce an overall minus sign, so that an acceptable
two-particle fermion state [ satises (2, 1)[ = [. Hence fermion states are antisymmetric. Now try to
give my two particles in the same state labels: [ = [x, x. But then (2, 1)[ = (2, 1)[x, x = [x, x, and
since multi-fermion states must be antisymmetric under interchange, we must still have (2, 1)[ = [. So
[ = [, which means [ = 0 the state doesnt exist! This is the basis for the Pauli exclusion principle,
which states that two electrons in an atom cant have the same quantum numbers.
To create antisymmetrized multi-fermion states, we need more information about permutations:
Denition 4.1 (Transposition). The transposition
ij
S
n
is the permutation that switches only i and
j; i.e.,
ij
(k) =
_
_
j k = i
i k = j
k k ,= i and k ,= j.
(4.8)
As it turns out, every permutation S
n
can be written as a product of transpositions. For example, I can
write
(3, 1, 2) =
12
23
= (2, 1, 3)(1, 3, 2). (4.9)
This statement makes intuitive sense: If I have n lacrosse balls in front of me, and I want to perform permutation
upon them, I can switch two balls at a time until Im done. Of course, the product isnt unique; I could
keep switching around lacrosse balls as much as I like, as long as I end with the right conguration. However,
it turns out that, given a particular permutation , the number of transpositions I need to achieve is either
always even or always odd. For this reason, I can create a function on S
n
: Dene
sgn =
_
1 is a product of an even number of transpositions
1 is a product of an odd number of transpositions.
(4.10)
We wont prove this is well-dened; physicists use mathematical results all the time without understanding
them. With this, we can dene the antisymmetrizer A
(n)
:
A
(n)
=
Sn
(sgn ). (4.11)
So for n = 2,
A
(2)
=
S2
(sgn ) = (1, 2) (2, 1), (4.12)
and n = 3,
A
(3)
=
S3
(sgn ) = (1, 2, 3) + (3, 1, 2) + (2, 3, 1) (2, 1, 3) (1, 3, 2) (3, 2, 1). (4.13)
4
JHMT 2014 Power round Identical Quantum Particles February 15, 2014
7. Show that the state
A
(n)
[x =
Sn
(sgn )[x (4.14)
satises A
(n)
[x = (sgn )A
(n)
[x for all S
n
, so that it has the correct behavior under particle
interchange for a multi-fermion state.
8. Show that A
2
(n)
= C
A
(n)
for some constant C
, and nd C
A
(n)
, so that A
2
(n)
= A
(n)
.)
9. (a) Show that if x
i
= x
j
for some i ,= j, A
(n)
[x = 0.
(b) Given the state [x in which x
i
I for all i and x
i
,= x
j
for all i, j, nd the constant N
R
+
such
that [
= N
A
(n)
[x satises
= 1.
5 Fock Spaces and Creation and Annihilation Operators
Now that we know how to create multi-particle states with the correct symmetry properties, our notation looks
very cumbersome: Acceptable states are huge sums of kets, and the states [
i B
n
the building blocks of all
possible states arent even realizable, because they dont have the correct symmetry. We ought only to deal
with realizable states, like [
+
from problem 6. So let
[n
1
, x
1
; n
2
, x
2
; . . . ; n
, x
= N
(n)
[ = N
(n)
[ x
1
, x
1
. . . , x
1
. .
n1 times
, x
2
, x
2
, . . . x
2
. .
n2 times
, . . . x
, x
. . . , x
. .
n
times
(5.1)
denote the state with n =
l
i=1
n
i
particles, in which n
i
particles have label x
i
. The subscript on the ket
denotes whether the particles are bosons (+) or fermions (), so S
+
(n)
is the symmetrizer S
(n)
and S
(n)
is the
antisymmetrizer A
(n)
. (Of course, for fermions this state is automatically zero if any of the n
i
s is greater than
one.) For a set number of particles n, this state lives in the Hilbert space S
(n)
1
n
, where 1 is the single-particle
Hilbert space. But sometimes physical process create or destroy particles, and that means we have to consider a
space that contains all possible states corresponding to all possible particle numbers. This is the innite tensor
product
T
n=0
S
(n)
1
n
and is called the Fock space of the system. Note that there is one special state that contains no particles; we
call it the vacuum and denote it by [0. To add particles to the vacuum to create multi-particle states we
introduce the annihilation operator a:
Denition 5.1 (Annihilation operator). The annihilation operator a
i
removes a particle with state label
i from the system. So a
i
[1, i = [0, and we dene a
i
[0 = 0.
Denition 5.2 (Creation operator). The creation operator a
i
the Hermitian conjugate of a
i
adds
a particle with state label i to the system. So a
i
[0 = [1, i.
Notice that, since a
i
is the Hermitian conjugate of a
i
, 0[a
i
= [a
i
[0]
j
] =
ij
and [b
i
, b
j
] = 0 (5.4)
while fermion creation and annihilation operators for which well use the letter f satisfy anticommutation
relations:
f
i
, f
j
=
ij
and f
i
, f
j
= 0. (5.5)
(Physicists usually use a for both fermions and bosons, which is less transparent.) Since [b
i
, b
j
] = 0, b
i
b
j
= b
j
b
i
.
This says we can shue around all boson annihilation operators (or creation operators) amongst each other and
nothing changes. For fermions, though, we pick up minus signs, since f
i
f
j
= f
j
f
i
. This is exactly what we
want to happen when we interchange two particles: the boson state remains the same, and the fermion state
picks up a minus sign.
Commuting annihilation operators past creation operators is more complicated: for example,
b
i
(b
i
)
2
[0 = (b
i
b
i
)b
i
[0 = (b
i
b
i
+ [b
i
, b
i
])b
i
[0 = (b
i
b
i
+ 1)b
i
[0 = b
i
(b
i
b
i
+ [b
i
, b
i
])[0 + b
i
[0
= 2b
i
[0.
(5.6)
Manipulations like these which put creation and annihilation operators into normal order allow us to build
up the entire Fock space by acting with an ordered product of creation operators on the vacuum. For the
moment, set b
i
= b and b
i
= b
, b) = (b
)
n
b
m
is the normal ordering of B
n,m
.
If I evaluate a normal-ordered product on 0[ or [0, I always get zero (unless the products just a constant).
And for any product of operators B
n,m
, I can use the commutation relations to commute all the bs to the left
of the b
s, so that B
n,m
must be equal to some sum of products of operators in normal order. For example,
b
2
(b
)
2
= 4b
b + 2.
10. Write b(b
)
n
as a sum of operators in normal order. Then use the fact that, for operators p and q,
(pq)
= q
to write b
n
b
)
n
[0 (5.7)
and use problem 10 to nd c
n
such that m[n =
mn
. Use this to nd b
)
n
as a sum of normal-ordered products; that is, show that
b
n
(b
)
n
=
n
k=0
c
n,k
(b
)
nk
b
nk
(5.8)
and nd c
n,k
for all n N.
With this, I can now equate
[n
1
, x
1
; n
2
, x
2
; . . . ; n
, x
+
=
_
i=1
c
ni,ni
(b
i
)
ni
_
[0, (5.9)
where the c
ni,ni
are the coecients from problem 12, and for operators whose order matters I dene
i=1
q
i
= q
q
1
q
1
. (5.10)
6
JHMT 2014 Power round Identical Quantum Particles February 15, 2014
For bosons all the creation operators commute, so order doesnt mater; for fermions it will matter. Moving to
fermions, obviously the state [n, i doesnt make sense unless n = 0 or 1, because
(f
i
)
2
= f
i
, f
i
= f
i
, f
i
= 0 (5.11)
for all f
i
. The notation automatically prohibits fermion states in which two or more particles have the same
label. So I can write
[1, x
1
; 1, x
2
; . . . ; 1, x
=
_
i=1
(f
i
)
ni
_
[0. (5.12)
More generally, physicists tend to consider the state labels x
i
themselves as variables, so we would write b
xi
instead of b
i
, and [b
xi
, b
yj
] =
xiyj
. For example, say Im considering the scattering of two Higgs bosons; then
the important state labels are the bosons momenta p
1
and p
2
. So my initial state is [1, p
1
; 1, p
2
, which would
be
b
p2
b
p1
[0. (5.13)
Then, I act on it with the scattering operator o (which is a very complicated object) and check the overlap
between some nal two-Higgs state [1, p
3
; 1, p
4
:
1, p
3
; 1, p
4
[o[1, p
1
; 1, p
2
. (5.14)
As long as momentum is conserved ( p
3
+ p
4
= p
1
+ p
2
), p
3
and p
4
can be anything, so the correct object to
evaluate is actually
p3
p4
K( p
1
, p
2
, p
3
, p
4
)1, p
3
; 1, p
4
[o[1, p
1
; 1, p
2
1, p
3
; 1, p
4
[o[1, p
1
; 1, p
2
, (5.15)
where K( p
1
, p
2
, p
3
, p
4
) is some function determined by physics independent of o. Now were doing quantum
mechanics. Anyway, we need to be able to calculate things like
1, y
1
; 1, y
2
; . . . ; 1, y
n
[1, x
1
; 1, x
2
; . . . ; 1, x
m
, (5.16)
where its possible that y
i
= y
j
for some i ,= j. We can actually do this for both fermions and bosons at the
same time:
13. Show that
b
y
n
i=1
b
xi
[0 =
n
i=1
yxi
n1
j=1
b
i(xj)
[0, (5.17)
where
i
(x
j
) =
_
x
j
j < i
x
j+1
j i,
(5.18)
and nd a corresponding formula for
f
y
n
i=1
f
xi
[0. (5.19)
14. Show that
P
n,m
(y, x)
1, y
1
; 1, y
2
; . . . ; 1, y
n
[1, x
1
; 1, x
2
; . . . ; 1, x
m
=
nm
Sn
s
()
n
i=1
xiy
(i)
, (5.20)
where s
+
() = 1 and s
() = sgn .
Formula (5.20) says that the state with particles with state labels y
i
overlaps with the state with m particles
with state labels x
i
if (i) = m so the states contain the same number of particles and (ii) the y
i
s are
some permutation of the x
i
s. For fermions, the order of the state labels in the bra and ket matter, so we need
to multiply the term corresponding to by sgn .
7
JHMT 2014 Power round Identical Quantum Particles February 15, 2014
In a way, this formula is as complicated as it gets: If I were to compute something like
1, y
1
; 1, y
2
; . . . ; 1, y
[o[1, x
1
; 1, x
2
; . . . ; 1, x
m
(5.21)
for some operator o, the smart thing to do is to write o in terms of the creation and annihilation operators
themselves. Say I want to know the mass M of a state; if I know the mass corresponding to each state label x,
then the mass operator is just
M =
x
m
x
a
x
a
x
. (5.22)
This actually counts the number of particles with state label x and adds m
x
for each of them. In this way
and with a few tweaks and embellishments physicists can reduce problems with arbitrary numbers of many
kinds of particles to a more manageable form.
8