Contemporary Quantum Physics 2017/18

Topic 4: The quantum harmonic oscillator, Coherent states and Schrödinger’s cat states

Dimitris Kakofengitis E-mail: [email protected]

March 28, 2018

4.1 Introduction
In this workshop we will revisit the quantum harmonic oscillator and the creation and annihilation operators.
In the first part of the assignment you will use the creation operator to produce the excited states of the
quantum harmonic oscillator. In the second part you are asked to produce the coherent state, also known as
the Glauber state of the quantum harmonic oscillator. By superposing two coherent states, you are asked to
produce the Schrödinger cat state both in position space and in phase space.

4.2 The Quantum Harmonic Oscillator

The quantum harmonic oscillator makes use of the classical harmonic oscillator potential of the form
k 2
V (x) = x . (1)
2
The dynamics of the quantum harmonic oscillator are described by the time-dependent Schrödinger equation
of the form (with state index n = 0, 1, 2, . . .),

~2 ∂ 2
 
k 2 ∂Ψn (x, t)
ĤΨn (x, t) = − + x Ψn (x, t) = i~ , (2)
2M ∂x2 2 ∂t

where Ĥ is the quantum harmonic oscillator’s Hamiltonian. The quantum harmonic oscillator’s eigenfunctions
take the form
i
Ψn (x, t) = ψn (x) · e− ~ En t (3)
where ψn and En are its corresponding eigenfunctions and eigenenergies, respectively (Hn are the Hermite
polynomials of degree n),
!
M ω 1/4 − M ωx2
r
n 
   
1 Mω n x 2 d −x 2
 1
ψn = √ · e 2~ Hn x , Hn (x) = (−1) e e , En = ~ω n + . (4)
2n n! π~ ~ dxn 2

4.2.1 Creation and Annihilation Operators

The excited states of the quantum harmonic oscillator can be produced via the creation and annihilation
operators, where in position space x̂ = x and p̂ = −i~ ∂∂x , are the position and momentum operators
r   r  
† Mω i Mω i
a = x̂ − p̂ , a = x̂ + p̂ (5)
2~ Mω 2~ Mω
Using the following equation we can produce the nth excited state of the quantum harmonic oscillator, given
prior knowledge of the (n − 1)th state OR the (n + 1)th state,

α† ψn−1 aψn+1
ψn = √ =√ . (6)
n n+1

1
 M ωx2
Therefore, starting from the groundstate ψ0 = N0 · e− 2~ , the nth state can be defined as

n
α † ψ0
ψn = √ , (7)
n!
where N0 is the normalisation constant for the groundstate.
Exercise 1: Generate a plot of the probability distributions of the quantum harmonic oscillator potential
for a spring constant of k = 1.
(a) Using the creation operator produce the first thirty (30) excited states of the quantum harmonic
oscillator and their respective energy levels.
[2 mark]
(b) For the groundstate and the excited states of the quantum harmonic oscillator produced in part (a),
check that they fulfill:
(i) Schrödinger’s equation,
[1 mark]
(ii) the normalisation condition.
[1 mark]
(c) On the same graph plot the harmonic oscillator potential and the probability distributions of its
ground and excited states produced in part (a). All the probability distributions must be displayed at the
corresponding energy level.
[3 marks]
(d) Produce a check which shows that the groundstate of the quantum harmonic oscillator and its excited
states produced in part (a), are orthogonal to each other.
[2 mark]
[Total for Exercise 1: 9 marks]

4.3 Coherent States

Coherent states, also known as Glauber states, play a central role in quantum optics and specifically in
quantum electrodynamics. In quantum mechanics, the coherent state is the best example resembling the
classical motion of a particle in the harmonic oscillator potential. It has a continuous distribution of its
constituent parts as its coefficients have the form
|α2 |
e− 2 αn
√ (8)
n!
where α can be any complex number. The coherent state has the form

α2 Ψ2 (x, t) α3 Ψ3 (x, t)
 
|α2 |
Ψcoh (x, t, α) = e− 2 Ψ0 (x, t) + αΨ1 (x, t) + √ + √ + ··· (9)
2! 3!
∞
|α2 | X αn
= e− 2 √ Ψn (x, t) , (10)
n=0 n!

where the time-dependent eigenfunctions Ψn (x, t), of the quantum harmonic oscillator, are defined in Eq. (3).

4.3.1 Schrödinger’s cat state

The Schrödinger’s Cat state is a superposition of two coherent states with opposite phase

Ψcoh (x, t, α) + Ψcoh (x, t, −α)

Ψcat = √ (11)
2

2
 Exercise 2: Generate a Schrödinger cat state with α = 2.6 for a harmonic oscillator potential with spring
constant k = 0.65.
(a) Determine a reasonable cut-off for the number of eigenfunctions you include in the sum that constitutes
the coherent state (N=30 is unnecessarily large).
[1 mark]
Hint: For a reasonable cut-off for the maximum number “N” of wave functions you include into the state’s
implementation, check the sum of the coefficients’ magnitude and be satisfied if that sum is above 0.99 ≈ 99%.
(b) Create a 3-D plot of the probability distribution of Schrödinger’s cat state.
[2 marks]
(c) Investigate the time development. In particular the period of the oscillation. Comment on your
observation.
[1 mark]
(d) Determine the times when the interference is most pronounced. Comment on your observations.
[1 mark] √
(e) Reproduce part (c) for α = 2.6i, where i = −1. Comment on your observations.
[1 mark]
(f ) How does the fringe spacing of the Schrödinger cat state’s interference fringes depend on the magnitude
of the amplitudes |α|. State your answer qualitatively.
[1 mark]
(g) Try to find a quantitative answer for how the fringes’ distances depend on the strength of the potential
and the magnitudes of the amplitudes.
[2 mark]
Hint: Think of the harmonic oscillator as a classical oscillator. What is the momentum of the oscillator
at the origin? What deBroglie wavelength do you expect it to have? Now, derive an expression for the
interference pattern.
[Total for Exercise 2: 9 marks]

4.4 The Wigner function

According to Heisenberg’s uncertainty principle [1], ∆x∆px ≥ 12 ~, where ∆x and ∆p are the standard devia-
tions of the probability distributions P = |Ψ|2 and P̃ = |Ψ̃|2 , we cannot simultaneously measure the precise
position and momentum of a quantum particle. Therefore we cannot assign a probability distribution to it
referring to its exact position and momentum. But there does exist a quantum phase space function, the
Wigner function, first introduced by Eugene Wigner in 1932 [2]. The Wigner function is set apart from
other quantum phase space functions [3] by the fact that it simultaneously yields the projections (probability
distributions P and P̃ ), while maintaining its form in Eq. (12) when evolved in time. It therefore joins P
and P̃ , into phase space. “Of course, [the Wigner function] cannot be really interpreted as the simultaneous
probability for coordinates and momenta” [2], due to the uncertainty principle (discussed above) and also since
it can assume negative values. Also, only the Wigner function’s averages and uncertainties evolve momentarily
classically [4, 5]. This is why we consider the Wigner function as the “closest quantum analogue of the classi-
cal phase-space distribution” [6], and is therefore also referred to as a quantum phase space quasi -probability
distribution.
The Wigner function W has the form
Z ∞ Z ∞
1 ∗ 2i 1 2i
W (x, p, t) = py
Ψ (x + y)Ψ(x − y)e ~ dy = Ψ̃∗ (p + s)Ψ̃(p − s)e− ~ xs ds , (12)
π~ −∞ π~ −∞

where asterisk (*) denotes complex conjugation, rendering it time-independent for eigenfunctions.

3
 Exercise 3:
(a) Show that the Wigner function for eigenfunctions, i.e. not in a superposition, is time-independent.
[2 mark]
(b) Show that the two expressions of W in Eq. (12) are equal.
[3 marks]
Hint: You can do this by substituting for Ψ̃ the Fourier transform of Ψ,
Z ∞
1 i
Ψ̃(p, t) = √ Ψ(x, t)e− ~ px dx , (13)
2π~ −∞
and then performing a change of variables, x1 = X + y and x2 = X − y. Do not forget the absolute value of
the Jacobian.

4.4.1 The marginal projections of the Wigner function:

The main property of the Wigner function distinguishing it from other quantum phase space functions [3,
7],
R ∞ is its projections along p andR ∞x, which are the probability distributions of position and momentum,
−∞ W (x, p, t) dp = P (x, t) and −∞ W (x, p, t) dx = P̃ (p, t), respectively. It follows that for normalised wave
R∞ R∞
functions, the Wigner function is also normalised, since −∞ P (x, t) dx = −∞ P̃ (p, t) dp = 1.
Exercise 3: R∞ R∞
(c) Show that the marginal projections of the Wigner function, i.e. −∞ W dp and −∞ W dx, are equal
to the probability distributions P = |Ψ|2 and P̃ = |Ψ̃|2 , respectively.
[3 marks]
Hint: You can make use of both expressions of W and you will also need the Dirac δ function
Z ∞ Z ∞
1 i 1 2i
δ(x − x0 ) = dp e ~ p(x−x0 ) = dp e ~ p(x−x0 ) (14)
2π~ −∞ π~ −∞

Some of its properties which you might find useful, are:

δ(x − x0 ) = δ(x0 − x) (15)

Z ∞
dδ(x − x0 ) 2i 2i
= dp pe ~ p(x−x0 ) (16)
dx π~2 −∞
Z ∞ Z ∞
dδ(x) df (x)
dx f (x) =− δ(x) (17)
−∞ dx −∞ dx
Z ∞
dx f (x)δ(x − x0 ) = f (x0 ) (18)
−∞

4.4.2 Numerical Computation of the Wigner Function

It is sometimes impossible to determine an analytical expression of the Wigner function. Thus, W in Eq. (12)
is determined using, for example, the midpoint Riemann sum with integration limits [−Y, Y ],
Ny
∆Y X ∗ 2i nπ 1
W (x, p, t) ≈ Ψ (x + yr )Ψ(x − yr )e ~ 2Y yr , for yr = −Y + (r − )∆Y , (19)
π~ 2
r=1

2Y nπ
where Ny is an even number of partitions of width ∆Y = N y
. The substitution p = 2Y follows from
n o
N N
the Nyquist–Shannon sampling theorem [8] with n ∈ Z : − 2y ≤ n ≤ 2y . Thus, the p-axis domain of the
h i
πN πN
numerical integration in Eq. (19) is p ∈ − 4Yy , 4Yy . Note that yr in Eq. (19), ranges from y1 = −Y + NYy
Y
to yNy = Y − Ny . Therefore, the values of yr represent the values of y in Eq. (12) and appear within the

4
integration limits [−Y, Y ]. We also make sure that Ψ∗ (x + y)Ψ(x − y) ≈ 0 at the integration limits for all
values of x. For a sufficiently high value of Ny , the numerically integrated Wigner function in Eq. (19), must
satisfy the Wigner function projections, together with the normalisation condition,
ZZ ∞
dp dx W (x, p, t) = 1 . (20)
−∞

Exercise 3:
(d) Numerically determine the Wigner function W in Eq. (19) for the Schrödinger cat state in Eq. (11).
For this you will need to determine a sufficiently large integration limit Y and number of partitions Ny . The
Maple command you will need to use is ApproximateInt. While the Wigner function is evolved in time,
which direction does it rotate? Clockwise or counterclockwise? Does your observation agree with the second
animation in this link: Cat state?
[4 marks]
[Total for Exercise 3: 12 marks]

4.4.3 Expectation Values

The position expectation value is of the form
Z ∞
hxi = Ψ∗ (x, t) x Ψ(x, t) dx (21)
−∞

and the momentum expectation value is equivalently expressed as

Z ∞
hpi = Ψ̃∗ (p, t) p Ψ̃(p, t) dp (22)
−∞

Exercise 4:
(a) Show that the position expectation value can be written in terms of W as
ZZ ∞
hxi = x W (x, p, t) dp dx . (23)
−∞

[2 marks]
(b) Similarly show that the momentum expectation value can be written in terms of W as
ZZ ∞
hpi = p W (x, p, t) dp dx . (24)
−∞

[2 marks]
(c) Show that the momentum expectation value can also be written as
Z ∞
∂Ψ(x, t)
hpi = −i~Ψ∗ (x, t) dx . (25)
−∞ ∂x

[2 marks] Hint: Make use of the Dirac δ function identities in Eqs. (16) and (17). Alternatively, substitute
Eq. (13) into Eq. (22).
The expectation values of x2 and p2 can also be written as
ZZ ∞

2
x = x2 W (x, p, t) dp dx (26)
−∞

and ZZ ∞

2
p = p2 W (x, p, t) dp dx (27)
−∞

5
The standard deviation in position and momentum can be written as
q
∆x = hx2 i − hxi2 (28)

and q
∆p = hp2 i − hpi2 (29)

Heisenberg’s uncertainty principle is then defined using the above standard deviations as ∆x∆p ≥ 21 ~.
Exercise 4:
(d) Plot the time evolution of the product of the standard deviation of position and momentum (∆x∆p)
of the cat state produced in Exercise 3, over one period time and comment on your result with reference to
Heisenberg’s uncertainty principle.
[4 marks]
[Total for Exercise 4: 10 marks]

4.5 Submission
The deadline for this assignment is on the 18th of April at 1500hrs

Submit your handwritten answers as a single PDF document and for the exercises which require Maple,
submit a single Maple spreadsheet after you have removed the output (using Alt-E-V-W). The file name
of both the PDF document and Maple spreadsheet must start with your name and/or student number in the
file name. The total marks for this assignment is 40. The following is some advice on what you can do to
avoid losing your work:

• Save your work frequently. In the context of Maple, before saving your spreadsheet, first remove all the
output by using the following keys: Alt-E-V-W. Then press Ctrl-S to save and submit the code only.
Remember that I will then do the reverse: Alt-E-E-W, so please check that no errors will come up

• Do not leave your submission until five minutes before 1500hrs

• Back up your work: http://www.pgbovine.net/unison_guide.htm

• Switch from Math to Maple format using Ctrl-M or do it globally: Tools, Options, Display. Set “In-
put display” to “Maple Notation”. Then click “Apply Globally”. http://www.mapleprimes.com/
questions/37703-Text-Mode-Input

• Add comments to make your work and reasoning legible: Remove clutter from your program, add
comments using the # key or using the print command. To add comments in text format, first create a
new execution group with either Ctrl-J or Ctrl-K and then using Ctrl-T you can start writting in text
format

• The following keys are useful:

Always end a command with semicolon ; or colon :
Ctrl-Delete deletes an execution group
F3 splits an execution group and F4 merges an execution group with the one below
Shift-Enter gets you into the next line in the same execution group
Always start your spreadsheet with the restart command. It is similar to “clear all” in MatLab.

References
[1] W. Heisenberg, Zeitschrift fur Physik 43, 172 (1927).

[2] E. Wigner, Phys. Rev. 40, 749 (1932).

6
[3] M. Hillery, R. F. O’Connell, M. O. Scully, and E. P. Wigner, Phys. Rep. 106, 121 (1984).

[4] A. Royer, Found. Phys. 22, 727 (1992).

[5] L. E. Ballentine, Y. Yang, and J. P. Zibin, Phys. Rev. A 50, 2854 (1994).

[6] W. H. Zurek, Nature 412, 712 (2001).

[7] W. P. Schleich, Quantum Optics in Phase Space (WILEY-VCH Verlag GmbH, 2001).

[8] C. E. Shannon, Proc. IEEE 86, 447 (1998).