Chapter 9

TIME-DEPENDENT PERTURBATION THEORY

Time dependent Schrödinger equation

∂Ψ
HΨ = i~
∂t
can be solved by separation of variables, (when V does not depend on time.)

Ψ(r, t) = ψ(r)eiEt/~

where ψ is a solution of the time-independent equation.

Hψ = Eψ

If we want to allow for transitions (quantum jumps, as they are sometimes called)
between one energy level and another, we must introduce a time-dependent potential
(quantum dynamics). There are precious few exactly solvable problems in quantum
dynamics. However, if the time-dependent portion of the Hamiltonian is small
compared to the time-independent part, it can be treated as a perturbation. My
purpose in this chapter is to develop time-dependent perturbation theory, and study its
most important application: the emission or absorption of radiation by an atom.
 9.1 TWO-LEVEL SYSTEMS

Suppose there are two states of the (unperturbed) system, ψa and ψb

H 0 ψa = Ea ψa , and H 0 ψb = Eb ψb (1)

orthogonality:
hψa |ψb i = δab (2)
An initial state at t = 0
Ψ(0) = ca ψa + cb ψb (3)
For any t
Ψ(t) = ca ψa e−iEa t/~ + cb ψb e−iEb t/~ (4)
Probability
|ca |2 + |cb |2 = 1 (5)

9.1.1 The perturbed system

Now turn on a time-dependent perturbation H 0 (t). The coefficients ca and cb become
functions of time
Ψ(t) = ca (t)ψa e−iEa t/~ + cb (t)ψb e−iEb t/~ (6)

We solve ca (t) and cb (t)

∂Ψ
HΨ = i~ , where H = H 0 + H 0 (t) (7)
 ca [H 0 ψa ]e−iEa t/~ + cb [H 0 ψb ]e−iEb t/~ + ca [H 0 ψa ]e−iEa t/~ + cb [H 0 ψb ]e−iEb t/~
h
= i~ ċa ψa e−iEa t/~ + ċb ψb e−iEb t/~
   
iEa iEb i
+ ca ψa − e−iEa t/~ + cb ψb − e−iEb t/~
~ ~

h i
ca [H 0 ψa ]e−iEa t/~ + cb [H 0 ψb ]e−iEb t/~ = i~ ċa ψa e−iEa t/~ + ċb ψb e−iEb t/~ (8)

ca hψa |H 0 |ψa ie−iEa t/~ + cb hψa |H 0 |ψb ie−iEb t/~ = i~ċa e−iEb t/~
For a short hand notation we define

Hij0 ≡ hψi |H 0 |ψj i (9)

we get
i h 0 0
i
ċa = − ca Haa + cb Hab e−i(Eb −Ea )t/~ (10)
~
In a parallel way

ca hψb |H 0 |ψa ie−iEa t/~ + cb hψb |H 0 |ψb ie−iEb t/~ = i~ċb e−iEb t/~
Hence
i h 0 0 i(Eb −Ea )t/~
i
ċb = − cb Hbb + ca Hba e (11)
~
In many cases the diagonal component vanishes:
0 0
Haa = Hbb =0 (12)

Then
i 0 −iω0 t i 0 iω0 t
ċa = − Hab e cb , ċb = − Hba e ca (13)
~ ~
where
Eb − Ea
ω0 ≡ (14)
~

—————————————————————-
Problem 9.1 A hydrogen atom is placed in a (time-dependent) electric field E = E(t)k̂ .
Calculate all four matrix elements Hij0 of the perturbation H 0 = eEz between the ground
state (n = 1) and the (quadruply degenerate) first excited states (n = 2). Also show
that Hii0 = 0 for all five states. Note: There is only one integral to be done here, if you
exploit oddness with respect to z; only one of the n = 2 states is “accessible” from the
ground state by a perturbation of this form, and therefore the system functions as a
two-state configuration–assuming transitions to higher excited states can be ignored.
—————————————————————–
Problem9.2 Solve Equation 9.13 for the case of a time-independent perturbation,
assuming that ca (0) = 1 and cb (0) = 0. Check that |ca (t)|2 + |cb (t)|2 = 1. Comment:
Ostensibly, this system oscillates between “pure ψa and “some ψb . Doesn’t this
contradict my general assertion that no transitions occur for time- independent
perturbations? No, but the reason is rather subtle: In this case ψa and ψb are not, and
never were, eigenstates of the Hamiltonian–a measurement of the energy never yields
Ea or Eb . In time-dependent perturbation theory we typically contemplate turning on
the perturbation for a while, and then tuming it off again, in order to examine the
system. At the beginning, and at the end, ψa and ψb are eigenstates of the exact
Hamiltonian, and only in this context does it make sense to say that the system
underwent a transition from one to the other. For the present problem, then, assume
that the perturbation was tumed on at time t = 0, and off again at time t–this doesn’t
affect the calculations, but it allows for a more sensible interpretation of the result.
—————————————————
Problem 9.3 Suppose the perturbation takes the form of a delta function (in time):

H 0 = Uδ(t)

assume that Uaa = Ubb = 0, and let Uab = Uba ∗ ≡ α. If c (−∞) = 1 and c (−∞) = 0,
a b
find ca (t) and cb(t), and check that |ca (t)| + |cb (t)|2 = 1. What is the net probability
2

(Pa→b for t → ∞) that a transition occurs? Hint: You might want to treat the delta
function as the limit of a sequence of rectangles. Answer:

Pa→b = sin2 (|α|/h).

———————————————-
9.1.2 Time-Dependent Perturbation Theory
So far, everything is exact: We have made no assumption about the size of the per-
turbation. But if H 0 is “small,” we can solve Equation 9.13 by a process of succes- sive
approximations, as follows. Suppose the particle starts out in the lower state:

ca (0) = 1, cb (0) = 0 (15)

Zeroth Order:
(0) (0)
ca (t) = 1, cb (t) = 0 (16)

First Order
(1)
dca (1)
dt
= 0 ⇒ ca (t) = 1
(17)
(1)
dcb (1) Rt 0 iω0 t 0 dt 0
dt
= − ~i Hba
0 eiω0 t ⇒ cb = − ~i 0
0 Hba (t )e

Second Order
(2)  R
dca
= − ~i Hab
0 e−iω0 t − i t 0 0 iω0 t 0 dt 0 ⇒
dt ~ 0 Hba (t )e
(18)
t0
hR i
(2) 0 −iω0 t 0 0 (t 00 )eiω0 t 00 dt 00
1
Rt 0
ca (t) =1− ~2 0 Hab (t )e 0 Hba dt 0

(2) (1) (2)

while cb is unchanged (cb (t) = cb (t)). (Notice that ca (t) includes the zeroth-order
term; the second-order correction would be the integral part alone.)
—————————————
Problem 9.4 Suppose you don’t assume Haa 0 = H 0 = 0.
bb
(a) Find ca (t) and cb (t) in first-order perturbation theory, for the case ca (0) =
 (1) (1)
1, cb (0) = 0. Show that |ca (t)|2 + |cb (t)|2 = 1, to first order in H 0 .
(b) There is a nicer way to handle this problem. Let
0
(t 0 )dt 0 0
(t 0 )dt 0
i Rt i Rt
Haa Hbb
da ≡ e ~ 0 ca , db ≡ e ~ 0 cb (19)

Show that
i 0 i
ḋa = − eiφ Hab e−iω0 t db ; ḋb = − e−iφ Hba
0 iω0 t
e da (20)
~ ~
where
1
Z t
0
φ(t) ≡ [Haa (t 0 ) − Hbb
0
(t 0 )]dt 0 (21)
~ 0

So the equations for da and db are identical in stmcture to Equation 9.13 (with an extra
factor eiφ tacked onto H 0 ).
(c) Use the method in part (b) to obtain ca (t) and cb(t) in first-order perturbation
theory, and compare your answer to (a). Comment on any discrepancies.
—————————————————–
Problem 9.5 Solve Equation 9.13 to second order in perturbation theory, for the general
case ca (0) = a, cb(0) = b.
———————————————————-
Problem9.6 Calculate ca (t) and cb (t), to second order, for a time-independent
perturbation (Problem 9.2). Compare your answer with the exact result.
 9.1.3 Sinusodial Perturbation
Suppose the perturbation has sinusodial time dependence:
H 0 (r, t) = V (r) cos(ωt) (22)
so that
0
Hab = Vab cos(ωt) (23)
where
Vab ≡ hψa |V |ψb i (24)

To first order, we have (Equation 9.17); (suppresing the superscript)

Z t
i iVba t h i(ω0 +ω)t 0
Z
0 0
i
cb (t) ∼
= − Vba cos(ωt 0 )eiω0 t dt 0 = − e + ei(ω0 −ω)t dt 0 (25)
~ 0 2~ 0
" #
Vba ei(ω0 +ω)t − 1 ei(ω0 −ω)t − 1
=− +
2~ ω0 + ω ω0 − ω
This is the it answer. For the simlicity, let us assmue
ω0 + ω  |ω0 − ω| (26)
Dropping the first term, we have
Vba ei(ω0 −ω)t/2 h i(ω0 −ω)t/2 i
cb (t) ∼
=− e − e−i(ω0 −ω)t/2 (27)
2~ ω0 − ω
Vba sin[(ω0 − ω)t/2] i(ω0 −ω)t/2
= −i e
~ ω0 − ω
Transition probability

|Vab |2 sin2 [(ω0 − ω)t/2]

Pa→b = |cb (t)|2 ∼
= (28)
~2 (ω0 − ω)2

The most remarkable feature of this result is that, as a function of time, the transition
probability oscillates sinusoidally (Figure 9.1). After rising to a maxi- mum of
|Vab |2 /h2 (ω0 − ω)2 –necessarily much less than 1, else the assumption that the
perturbation is “small” would be invalid–it drops back down to zero!

As I noted earlier, the probability of a transition is greatest when the driving frequency
is close to the “natural” frequency, ω0. This is illustrated in Figure 9.2, where Pa→b is
plotted as a function of ω. The peak has a height of (|Vab |t/2})2 and a width 4π/t;
evidently it gets higher and narrower as time goes on. (Ostensibly, the maximum
increases without limit. However, the perturbation assumption breaks down before it
gets close to 1, so we can believe the result only for relatively small t.
————————————-
Problem 9.7 The first term in Equation 9.25 comes from the eiωt /2 part of cos(ωt), and
the second from e−iωt /2. Thus dropping the first term is formally equivalent to writing
H 0 = (V /2)e−iωt , which is to say,

0 Vba −iωt 0 Vab iωt

Hba = e , Hab = e (29)
2 2
(The latter is required to make the Hamiltonian matrix hermitian–or, if you pre- fer, to
pick out the dominant term in the formula analogous to Equation 9.25 for ca (t).) Rabi
noticed that if you make this so-called rotating wave approxi- mation at the beginning of
the calculation, Equation 9.13 can be solved exactly, with no need for perturbation
theory, and no assumption about the strength of the field.
(a) Solve Equation 9.13 in the rotating wave approximation (Equation 9.29), for the
usual initial conditions: ca (0) = 1, cb (0) = 0. Express your results (ca (t) and cb(t)) in
terms of the Rabi flopping frequency,
1
q
ωr ≡ (ω − ω0 )2 + (|Vab |/~)2 (30)
(b) Determine the transition probability, Pa→b (t), and show that it never exceeds 1.
Confirm that |ca (t)|2 + |cb (t)|2 = 1.
(c) Check that Pa→b (t) reduces to the perturbation theory result (Equation 9.28) when
the perturbation is “small,” and state precisely what small means in this context, as a
constraint on V .
(d) At what time does the system first retum to its initial state?
 9.2 EMISSION AND ABSORPTION OF RADIATION
9.2.1 Electromagnetic waves

An atom, in the presence of a passing light wave, responds primarily to the electric
component. If the wavelength is long (compared to the size of the

E = E0 cos(ωt)k̂ (31)

The perturbing Hamiltonian is

H 0 = −qE0 z cos(ωt) (32)

where q is the charge of the electron. Evidently,

0
Hba = −℘E0 cos(ωt), where ℘ ≡ qhψb |z|ψa i (33)
We may identify
Vba = −℘E0 (34)

9.2.2 Absorption, Stimulated Emission, and Spontaneous Emission

The lower state ψa : upper state ψb
|℘|E0 2 sin2 [(ω0 − ω)t/2]
 
Pa→b (t) = (35)
~ (ω0 − ω)2
with the absorption energy Eb − Ea = ~ω0 from the EM wave.
We may start from the upper state ψb (ca (0) = 0 and cb (0) = 1), then
|℘|E0 2 sin2 [(ω0 − ω)t/2]
 
Pb→a (t) = (36)
~ (ω0 − ω)2
(It has to come out this way–all we’re doing is switching a ↔ b, which substitutes −ω0
for ω0 .
If the particle is in the upper state, and you shine light on it, it can make a transition to
the lower state, and in fact the probability of such a transition is exactly the same as for
a transition upward from the lower state. This process, which was first predicted by
Einstein, is called stimulated emission.

(a) Absorption (b) Stimulated emission (c) Spontaneous emission

In the case of stimulated emission the electromagnetic field gains energy }ω0 from
the atom; we say that one photon went in and two photons came out–the original one
that caused the transition plus another one from the transition itself (Figure 9.4(b)).
This raises the possibility of ampliffication, for if I had a bottle of atoms, all in the upper
state, and triggered it with a single incident photon, a chain reaction would occur, with
the first photon producing 2, these 2 producing 4, and so on. We’d have an enormous
number of photons coming out, all with the same frequency and at virtually the same
instant. This is, of course, the principle behind the laser (light amplification by
stimulated emission of radiation). Note that it is essential (for laser action) to get a
majority of the atoms into the upper state (a so-called population inversion), because
absorption (which costs one photon) competes with stimulated emission (which
produces one); if you started with an even mixture of the two states, you d get no
amplification at all.
There is a third mechanism (in addition to absorption and stimulated emission) by
which radiation interacts with matter; it is called spontaneous emission. Here an
atom in the excited state makes a transition downward, with the release of a photon,
but without any applied electromagnetic field to initiate the process (Figure 9.4(c)).

9.2.3 Incoherent Perturbation

The energy density in an electromagnetic wave is

0 2
u= E (37)
2 0

So the transition prob- ability (Equation 9.36) is (not surprisingly) proportional to the
energy density of the fields:

2u sin2 [(ω0 − ω)t/2]

Pb→a (t) = 2
|℘|2 (38)
0 ~ (ω0 − ω)2

But this is for a monochromatic wave, at a single frequency ω. In many applica- tions
the system is exposed to electromagnetic waves at a whole range of fre- quencies; in
that case u → ρ(ω)dω, where ρ(ω)dω is the energy density in the frequency range dω,
and the net transition probability takes the form of an integral:
( )
2 ∞ sin2 [(ω0 − ω)t/2]
Z
Pb→a (t) = |℘|2 ρ(ω) dω (39)
0 ~2 0 (ω0 − ω)2

we may as well replace ρ(ω) by ρ(ω0), and take it outside the integral:

2|℘|2 ∞ sin2 [(ω0 − ω)t/2]

Z
Pb→a (t) ∼
= ρ(ω0 ) dω (40)
0 ~2 0 (ω0 − ω)2

Since
∞ sin2 x
Z
= dx = π (41)
−∞ x2
we get
π|℘|2
Pb→a (t) ∼
= ρ(ω0 )t (42)
 ~2
the transition rate R = dP/dt is a constant:

π
Rb→a (t) = |℘|2 ρ(ω0 ) (43)
0 ~2

So far, we have assumed that the perturbing wave is coming in along the y direction
(Figure 9.3), and polarized in the z direction. But we are interested in the case of an
atom bathed in radiation coming from all directions, and with all possible polarizations;
the energy in the fields (ρ(ω)) is shared equally among these different modes. What
we need, in place of |℘|2 , is the average of |℘ · n̂|2 , where

℘ ≡ qhψb |r|ψa i (44)

the average is over all polarizations and all incident directions.

The averaging can be carried out as follows: Choose spherical coordinates such that
the direction of propagation (k̂ ) is along x, the polarization (n̂) is along z, and the
vector µ defines the spherical angles θ and φ (Figure 9.5). Actually ℘ is fixed here, and
we are averaging over all k̂ and n̂ consistent with k̂ ⊥ n̂—which is to say, over all θ and
φ. But it’s really the coordinate system, not the vector ℘, that is changing.) Then

n̂ = cos φî + sin φĵ, ℘ = ℘ sin θĵ + ℘ cos θk̂ (45)

and
℘ · n̂ = ℘ sin θ sin φ

Conclusion: The transition rate for stimulated emission from state b to state a, under
the influence of incoherent, unpolarized light incident from all directions, is

1
Z
|℘ · n̂|2ave = |℘|2 sin2 θ sin2 φ sin θdθdφ (46)
4π
|℘|2 π
Z Z 2π 1
= sin3 θdθ sin2 φdφ = |℘|2
4π 0 0 3
π
Rb→a = |℘|2 ρ(ω0 ) (47)
30 ~2
where ℘ is the matrix element of the electric dipole moment between the two states
(Equation 9.44), and ρ(ω0) is the energy density in the fields, per unit frequency,
evaluated at ω0 = (Eb − Ea )/~
 9.3. SPONTANEOUS EMISSION
9.3.1 Einstein’s A and B Coefficents

Picture a container of atoms, Na of them in the lower state (ψa ), and Nb of them in the
upper state (ψb ). Let A be the spontaneous emission rate, so that the number of
particles leaving the upper state by this process, per unit time, is Nb A The transition
rate for stimulated emission, as we have seen (Equation 9.47), is propor- tional to the
energy density of the electromagnetic field: Bba ρ(ω0 ); the number of particles leaving
the upper s tate by this mechanism, per unit time, is Nb Bba ρ(ω0 ). The absorption rate
is likewise proportional to ρ(ω0)–call it Bab p(ω0); the num- ber of particles per unit
time joining the upper level is therefore Na Bab ρ(ω0 ). All told, then,

dNb
= −Nb A − dNb Bba ρ(ω0 ) + Na Bab ρ(ω0 ) (48)
dt
Assume the thermal equilibrium. In that case dN/dt = 0. So

A
ρ(ω0 ) = (49)
(Na /Nb )Bab − Bba

From the Boltzman factor exp(−E/kB T ), we get, on the other hand,

Na e−Ea /kB T
= −E /k T = e~ω0 /kB T (50)
Nb e b B
hence
A
ρ(ω0 ) = (51)
e~ω0 /kB T B −B
Planck’s blackbody formula tells us

~ ω3
ρ(ω) = (52)
π 2 c 3 e~ω/kB T − 1

We may conclude that

Bab = Bba (53)
and
ω03 ~
A= Bba (54)
π2 c 3
We also know that
π
Bba = |℘|2 (55)
30 ~2
The spontaneous emission rate is

ω03 |℘|2
A= (56)
3π0 ~c 3

————————————————
Problem 9.8 As a mechanism for downward transitions, spontaneous emission
competes with thermally stimulated emission (stimulated emission for which black-
body radiation is the source). Show that at room temperature (T = 300K) thermal
stimulation dominates for frequencies well below 5 × 1012 Hz, whereas spontaneous
emission dominates for frequencies well above 5 ×1012 Hz. Which mechanism
dominates for visible light?
————————————————-
Problem 9.9 You could derive the spontaneous emission rate (Equation 9.56) with- out
the detour through Einstein’s A and B coefficients if you knew the ground-state energy
density of the electromagnetic field, p0 (ω), for then it would simply be a case of
stimulated emission (Equation 9.47). To do this honestly would require quantum
electrodynamics, but if you are prepared to believe that the ground state consists of
one photon in each mode, then the derivation is very simple:
(a) Replace Equation 5.111 by Nω = dk , and deduce ρ0 (ω). (Presumably this formula
breaks down at high frequency, else the total “vacuum energy” would be inf finite. . .
but that’s a story for a different day.)
(b) Use your result, together with Equation 9.47, to obtain the spontaneous emission
rate. Compare Equation 9.56.
—————————————-

9.3.2 The Lifetime of an Excited State

Suppose there a some amount of excited states. The spontaneus emission rate tells us
the population change
dNb = −ANb dt (57)
in the time interval dt. The solution is

Nb (t) = Nb (0)e−At (58)

with a time constant

1
τ = (59)
A
which is called the lifetime of the state.
I have assumed all along that there are only two states for the system, bm this was
just for notational simplicity–the spontaneous emission formula (Equation 9.56) gives
the transition rate for ψb → ψa regardless of what other states may be accessible (see
Problem 9.15). Typically, an excited atom has many different decay modes (that is: ψb
can decay to a large number of different lower- energy states, ψa1 , ψa2 , ψa3 , . . .). In
that case the transition rates add, and the net lifetime is
1
τ = (60)
A1 + A2 + A3 + · · ·

========================================================
Example 9.1 Suppose a charge q is attached to a spring and constrained to oscil- late
along the x-axis. Say it starts out in the state |n} (Equation 2.61), and decays by
spontaneous emission to state |n0 }. From Equation 9.44 we have

℘ = qhn|x|n0 iî

You calculated the matrix elements of x back in Problem 3.33:

~ √ 0
r
0
√ 
hn|x|n i = n δn,n0 −1 + nδn0 ,n−1
2mω
where ω is the natural frequency of the oscillator. (I no longer need this letter for the
frequency of the stimulating radiation.) But we’re talking about emission, so n0 must be
lower than n; for our purposes, then,
r
n~
℘=q δn0 ,n−1 î (61)
Evidently transitions occur only to states one step lower on the ’‘ladder,” and the
frequency of the photon emitted is

En − En0 (n + 1/2)~ω − (n0 + 1/2)~ω

ω0 = = = (n − n0 )ω = ω (62)
~ ~
The transition rate is
nq 2 ω 2
A= (63)
6π0 mc 3
the lifetime of n-th state is
6π0 mc 3
τn = (64)
nq 2 ω 2
Meanwhile, each radiated photon carries an energy hω, so the power radiated is

q 2 ω2
P= (n~ω)
6π0 mc 3

since the energy of an oscillator is E = (n + 1/2)~ω,

q 2 ω2
 
1
P= E − ~ω (65)
6π0 mc 3 2
This is the average power radiated.
From classical electrodynamics, the power is gvie by the Larmor formula.

q 2 a2
P= (66)
6π c 3
For SHO x(t) = x0 cos(ωt) and a = −ω02 x0 cos(ωt). After averaging in time, we get

q 2 x02 ω 4
P=
12π0 c 3

But the energy of the oscillator is E = (1/2)mω 2 x02 , so x02 = 2E/mω 2 , and hence

q 2 ω2
P= E (67)
6π0 mc 3

=====================================================
Problem 9.10 The half-life (t1/2 ) of an excited state is the time it would take for half the
atoms in a large s ample to make a transition. Find the relation between t1/2 and τ (the
“lifetime” of the state).
———————————————–
Problem 9.11 Calculate the lifetime (in seconds) for each of the four n = 2 states of
hydrogen. Hint: You’ll need to evaluate matrix elements of the form
{ψ100 |x|ψ200 }, {ψ100 |y |ψ211 ), and so on. Remember that x = r sin θ cos φ, y =
r sinθ sin φ, and z = r cos θ. Most of these integrals are zero, so scan them before you
start calculating. Answer: 1.60 ×10−9 seconds for all except ψ200 , which is infinite.
 9.3.3 Selection Rules
We need to evaluate the matrix elments
hψb |r|ψa i

these quantities are very often zero, and it would be helpful to know in advance when
this is going to happen, so we don’t waste a lot of time evaluating unnecessary
integrals.
Suppos we have a spherical symmetry, like in the hydgrogen atom. The matrix
elements are
hn0 l 0 m0 |r|nlmi
Clever exploitation of the angular momentum commutation relations and the hermiticity
of the angular momentum operators yields a set of powerful constraints on this quantity.
===================================================
Selection rules involving m and m0 : Consider first the commutators of Lz with x, y ,
and z, which we worked out in Chapter 4 (see Equation 4.122):
[Lz , x] = i~y , [Lz , y ] = −i~x, [Lz , z] = 0 (68)
From the third,
0 = hn0 l 0 m0 |[Lz , z]|nlmi = hn0 l 0 m0 |(Lz z − zLz )|nlmi
= hn0 l 0 m0 |[(m0 ~)z − z(m~)]|nlmi = (m0 − m)~hn0 l 0 m0 |z|nlmi
Conclusion:
Either m0 = m, or else hn0 l 0 m0 |z|nlmi = 0 (69)
So unless m0 = m, the matrix elements of z are always zero.
Meanwhile, from the commutator of Lz with x we get

hn0 l 0 m0 |[Lz , x]|nlmi = hn0 l 0 m0 |(Lz x − xLz )|nlmi

= (m0 − m)~hn0 l 0 m0 |x|nlmi = i~hn0 l 0 m0 |y |nlmi

Conclusion:
(m0 − m)~hn0 l 0 m0 |x|nlmi = ihn0 l 0 m0 |y |nlmi (70)

Finally, the commutator of Lz with y yields

hn0 l 0 m0 |[Lz , y ]|nlmi = hn0 l 0 m0 |(Lz y − yLz )|nlmi

= (m0 − m)~hn0 l 0 m0 |y |nlmi = −i~hn0 l 0 m0 |x|nlmi

Conclusion:
(m0 − m)hn0 l 0 m0 |y |nlmi = −ihn0 l 0 m0 |x|nlmi (71)

And

(m0 − m)2 hn0 l 0 m0 |x|nlmi = i(m0 − m)hn0 l 0 m0 |y |nlmi = hn0 l 0 m0 |x|nlmi

Either (m0 − m)2 = 1, or else hn0 l 0 m0 |x|nlmi = hn0 l 0 m0 |y |nlmi = 0 (72)

the selection rule for m

No transitions occur unless ∆m = ±1 or 0 (73)

======================================================
Selection rules involving l and l 0 : In Problem 9.12 you are asked to derive the
following commutation relation:
h i
L2 , [L2 , r] = 2~2 (rL2 + L2 r) (74)

We sandwich this commutator between {n0 l 0 m0 | and |nlm} to derive the selection rule:
h i
hn0 l 0 m0 | L2 , [L2 , r] |nlmi = 2~2 hn0 l 0 m0 |(rL2 + L2 r)|nlmi (75)

= 2~4 [l(l + 1) + l 0 (l 0 + 1)]hn0 l 0 m0 |r|nlmi = hn0 l 0 m0 |(L2 [L2 , r] − [L2 , r]L2 )|nlmi
= ~2 [l 0 (l 0 + 1) − l(l + 1)]hn0 l 0 m0 |[L2 , r]|nlmi
= ~2 [l 0 (l 0 + 1)−l(l + 1)]hn0 l 0 m0 |(L2 r − rL2 )|nlmi
= ~4 [l 0 (l 0 +1) − l(l + 1)]2 hn0 l 0 m0 |r|nlmi

Conclusion:

Either 2[l(l + 1) + l 0 (l 0 + 1)] = [l 0 (l 0 + 1) − l(l + 1)]2 (76)

0 0 0
or else hn l m |r|nlmi = 0
But
[l 0 (l 0 + 1) − l(l + 1)] = (l 0 + l + 1)(l 0 − l)
and
2[l(l + 1) + l 0 (l 0 + 1)] = (l 0 + l + 1)2 + (l 0 − l)2 − 1
so
[(l 0 + l + 1)2 − 1][(l 0 − l)2 − 1] = 0 (77)

l =0 l =1 l=2 l =3

slection rule for l:

No transitions occur unless ∆l = ±1 (78)
There exists a so-called metastable state (like a 2S state), and its lifetime is indeed
much longer than that of, for example, the 2P states (ψ211 , ψ210 , and ψ21−1 ).
Metastable states do eventually decay, by collisions, or by what are (misleadingly)
called forbidden transitions (Problem 9.21), or by multiphoton emission.
=================================================
Problem 9.12 Prove the commutation relation in Equation 9.74. Hint: First show that

[L2 , z] = 2i~(xLy − yLx − i~z)

Use this, and the fact that r · L = r · (r × p) = 0, to demonstrate that

[L2 , [L2 , z]] = 2~2 (zL2 + L2 z)

The generalization from z to r is trivial.

———————————————–
Problem 9.13 Close the “loophole” in Equation 9.78 by showing that if l 0 = 1 = 0 then
{n0 l 0 m0 |r|nlm} = 0.
—————————————————
Problem 9.14 An electron in the n = 3, l = 0, m = 0 state of hydrogen decays by a
sequence of (electric dipole) tran sitions to the ground state.
(a) What decay routes are open to it? Speci f y them in the following way:

|300i → |nlmi → |n0 l 0 m0 i → · · · → |100i

(b) If you had a bottle full of atoms in this state, what fraction of them would decay via
each route?
(c) What is the lifetime of this state? Hint: Once it’s made the first transition, it’s no
longer in the state |300i, so only the first step in each sequence is relevant in
computing the lifetime. When there is more than one decay route open, the transition
rates add.