Quantum Entanglement in Continuous System

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/309484888
Quantum Entanglement in Continuous System
Thesis · May 2004

DOI: 10.13140/RG.2.2.24338.45764
CITATIONS READS
2 393
1 author:
Kwan Yuet Stephen Ho

National Institutes of Health
14 PUBLICATIONS 108 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Large Scale Text Clustering View project
SOCcer View project
All content following this page was uploaded by Kwan Yuet Stephen Ho on 28 October 2016.
The user has requested enhancement of the downloaded file.

Quantum Entanglement in Continuous System
by
Ho Kwan Yuet
Supervisor: Dr. Law Chi Kwong
A Final Year Project (PHY4610/PHY4620) Report

submitted to the
Department of Physics
The Chinese University of Hong Kong
May 24, 2004

Abstract
The concept of quantum entanglement arises since A. Einstein, B. Podolsky

and N. Rosen proposed their open question about the non-local quantum cor-
relation, which is now known as the EPR paradox.
In this report, we first review the famous EPR (Einstein-Podolsky-Rosen)
paradox and its connection to the property of entanglement. We discuss the
method of Schmidt decomposition to analyze the degree of entanglement using
participation ratio in general two-particle systems. We deal with the entangle-
ment of continuous systems, including the two-mode squeezed state, the com-
pound harmonic oscillators, the compound Morse oscillators and the positron-
ium. We explore the connection between the symmetry and the antisymmetry
of the wavefunction and the degeneracy in the Schmidt decomposition of the
state.
i
Acknowledgments
In this project, I must thank my supervisor, Dr. Law Chi Kwong, for his great
effort in guiding my project. He has spent so much precious time meeting with
me and giving me useful suggestions in doing the project. The meeting with
Dr. Law is very challenging. I encountered a lot of problems in the project,
and without his instruction and guidance, the problems would be very difficult.
In the project, I explore different ideas on entanglement and acquire a lot of
skills in doing research, like many programming skills, writing report using
LATEX and reading articles published in the journals. And most importantly,
I learned that I have to work very hard with the problems and be faced with
failure, and they are some of the crucial factors in solving any problems.
ii
Contents
1 EPR Paradox 1
1.1 Principle of Locality and Physical Reality . . . . . . . . . . . . 1
1.2 An Illustration using Bohm’s Spin-Singlet State . . . . . . . . . 2
1.2.1 Simultaneous Reality of Non-commutative Spins . . . . . 2
1.2.2 Mathematical Form of the Paradox . . . . . . . . . . . . 4
1.2.3 Resolving the EPR Paradox by Inspecting the Uncer-
tainty Relations . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 An Illustration using Two-Mode Squeezed States . . . . . . . . 4
2 Schmidt Decomposition and the Entangled States 8

2.1 Schmidt Decomposition of a Composite State . . . . . . . . . . 8
2.1.1 Proof of the Existence of the Schmidt Decomposition . . 9
2.1.2 Schmidt Decomposition in Continuous Systems . . . . . 11
2.1.3 Algorithm of Finding the Schmidt Decomposition of a
State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.4 Schmidt Decomposition of the Two-Mode Squeezed State 14
2.2 Properties of Entangled States . . . . . . . . . . . . . . . . . . . 18
2.2.1 Mixed Subsystems of Entangled States . . . . . . . . . . 18
2.2.2 EPR Paradox Revisited . . . . . . . . . . . . . . . . . . 20
2.2.3 Duan’s Condition for Entanglement of Continuous Systems 20
2.3 Quantifying Entanglement . . . . . . . . . . . . . . . . . . . . . 22
2.3.1 Entanglement Entropy . . . . . . . . . . . . . . . . . . . 23
2.3.2 Participation Ratio . . . . . . . . . . . . . . . . . . . . . 23
2.3.3 Entanglement under Various Transformations . . . . . . 25
2.3.4 Participation Ratio of the Two-Mode Squeezed State . . 28
3 Entanglement in Various Oscillating Systems 30

3.1 Composite Morse Oscillators . . . . . . . . . . . . . . . . . . . . 30
iii
3.1.1 Morse Potential . . . . . . . . . . . . . . . . . . . . . . . 30
3.1.2 Entanglement of Composite Morse Oscillators . . . . . . 33
3.2 Compound Harmonic Oscillators . . . . . . . . . . . . . . . . . 41
3.2.1 Decomposing the Ground State . . . . . . . . . . . . . . 42
3.2.2 Entanglement of the Ground State of Large Mass Dif-
ference . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.2.3 Entanglement of the Ground State of Equal Masses . . . 46
3.2.4 Entanglement of the Excited States of Equal Masses . . 46
3.3 Entanglement Degeneracies in Symmetric Wavefunctions . . . . 53
3.4 Asymptotic Behaviour of Entanglement in Oscillating Systems
of Equal Masses . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4 Entanglement in Three-Dimensional Systems 63

4.1 Three-Dimensional Two-mode Squeezed State . . . . . . . . . . 63
4.2 Three-Dimensional Positronium System . . . . . . . . . . . . . . 64
Bibliography 73
iv
List of Figures
1.1 Existence of the element of physical reality corresponding to Sz in

subsystem B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Existence of the element of physical reality corresponding to Sx in
subsystem B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 3D Plot of the two-mode squeezed state with b = 0.3333 . . . . . . . 5
2.1 Schmidt coefficients of the two-mode squeezed state with b = 0.256027 16

2.2 The first six Schmidt modes of the two-mode squeezed state . . . . . 17
2.3 Participation ratio K as a function b for the two-mode squeezed state 29
3.1 Plot of the Morse potential (rescaled) for ² = 24.01 . . . . . . . . . 31

3.2 Plot of the eigenstates under Morse potential (rescaled) for ² = 24.01
(j = 4.4) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3 Plot of the first 10 Schmidt coefficients of the composite Morse os-
cillator for ² = 24.01 (j = 4.4) and b = 0.1 . . . . . . . . . . . . . . 34
3.4 Schmidt modes of the composite Morse oscillator for ² = 24.01 (j =
4.4), b = 0.1 and ν = 0 . . . . . . . . . . . . . . . . . . . . . . . . 35
4.4), b = 0.1 and ν = 1 . . . . . . . . . . . . . . . . . . . . . . . . 36
4.4), b = 0.1 and ν = 2 . . . . . . . . . . . . . . . . . . . . . . . . 37
4.4), b = 0.1 and ν = 3 . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4), b = 0.1 and ν = 4 . . . . . . . . . . . . . . . . . . . . . . . . 39
3.9 Variation of the participation ratio of the composite Morse oscillator
against b for ² = 24.01 (j = 4.4) for ν = 0, 1, 2, 3, 4 . . . . . . . . . . 41
v
3.10 Schmidt coefficients of a state containing two particles of different
masses. We have adopted m1 = 1840 (the mass of a proton) and
m2 =1 (the mass of an electron) (in ratio) when b = 0.256027. . . . . 45
3.11 Variation of the participation ratio K of the state of a two particles
of different masses against b. We have adopted m1 = 1840 (the mass
of a proton) and m2 =1 (the mass of an electron) (in ratio). . . . . . 45
3.12 Plot of the first 10 Schmidt coefficients of the composite harmonic
oscillators for b = 2 for n from 0 to 5 . . . . . . . . . . . . . . . . . 47
3.13 Schmidt modes of the compound harmonic oscillator for b = 2 and
n=1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
n=2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
n=3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.16 Variation of the participation ratio of the compound harmonic oscil-
lator against b for n = 0, 1, 2, 3, 4, 5 . . . . . . . . . . . . . . . . . . 51
3.17 Plot of the wavefunction (3.42) for d = 4.0 . . . . . . . . . . . . . . 54
3.18 Plot of the wavefunction (3.42) for d = 1.0 . . . . . . . . . . . . . . 54
3.19 Variation of the participation ratio of the wavefunction (3.42) as the
parameter d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.20 The first two Schmidt coefficients of the state (3.42) and some of the
corresponding Schmidt modes . . . . . . . . . . . . . . . . . . . . 56
3.21 The wavefunction composing of six hemispheres with unit radius
centred at the marked coordinates. . . . . . . . . . . . . . . . . . . 57
3.22 First 12 Schmidt coefficients of the six-hemispherical (and six-fold
degenerate) wavefunction . . . . . . . . . . . . . . . . . . . . . . . 58
3.23 The 1st to 6th Schmidt modes of the six-hemispherical wavefunction 59
3.24 The 7th to 12th Schmidt modes of the six-hemispherical wavefunction 60
4.1 Plot of participation ratio of the three-dimensional two-mode squeezed

state against the parameter σ . . . . . . . . . . . . . . . . . . . . 64
4.2 Plot of participation ratio of the trapped positronium against the
parameter σ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3 Plot of Λnlm against n and l for σ = 0.5 . . . . . . . . . . . . . . . 70
4.4 Plot of Λnlm against n and l for σ = 2.5 . . . . . . . . . . . . . . . 70
4.5 Schmidt modes of (4.6) for l = 0 and σ = 2.0 . . . . . . . . . . . . 71
vi
vii
List of Tables
4.1 Sum of Schmidt coefficients of (4.12) in numerical calculation . . . . 68
viii
Chapter 1
EPR Paradox
The concept of entangled states arised when A. Einstein, B. Podolsky and
N. Rosen published its famous article entitled ”Can Quantum-Mechanical
Description of Physical Reality Be Considered Complete?” on the Physical
Review [1]. They used the concept of physical reality and the principle of
non-locality to argue for the incompleteness of the formalism of quantum me-
chanics. The state they used to show the incompleteness is an entangled state.
To look at quantum entanglement, we must review the EPR paradox first.
1.1 Principle of Locality and Physical Reality

In the original article written by EPR, they used a continuous wavefunction
to make their argument, but we will deal with the paradox here using another
way. The EPR paradox have the following assumptions:
1. Principle of Locality There is no action at a distance, i.e., the measure-

ment of one subsystem should have no influence on another subsystem
that is spatially apart.
2. Physical Reality “If, without in any way disturbing a system, we can

predict with certainty (i.e., with probability equal to unity) the value of a
physical quantity, then there exists an element of physical reality corre-
sponding to this physical quantity.”[1]
We consider two observables A and B that do not commute, i.e.,
[A, B] 6= 0 (1.1)
1
Chapter 1 EPR Paradox 2
Examples of a pair of non-commutative observables are the position and the

momentum operator in the same direction ([x, p] = i~), and the spins of a par-
ticle in different directions ([Sx , Sy ] = i~Sz ). Non-commutative observables do
not have compatible eigenstates. If |ψi is a set of eigenstates of A, then it is
not the set of eigenstates of B. In other words, for the state |ψi, there is an
element of physical reality corresponding to the observables A, but not to B.
From this simple illustration, we can easily conclude that there does not exist
elements of physical reality corresponding to non-commutative observables si-
multaneously (in EPR’s article, it is phrased as “non-commutative observables
does not have simultaneous reality.”[1]).
1.2 An Illustration using Bohm’s Spin-Singlet

State
1.2.1 Simultaneous Reality of Non-commutative Spins
D. Bohm suggested a spin-singlet state (total angular momentum equal to zero)
illustrating the EPR paradox [2], which is easier to follow than the example
raised in the EPR’s article. Consider a two-electron system in a spin-singlet
state, i.e., with a total spin of zero. They are then set spatially apart so that
no interaction can take place between them. The state is expressed in Dirac
notation as
1
|Ψi = √ (| ↑iA | ↓iB − | ↓iA | ↑iB ) (1.2)
2
Now, suppose the two subsystems, A and B, are separated apart such that no
interactions can take place between them.
We then measure Sz of subsystem A, and get, say, | ↑i. Then we get that
the subsystem B is in the state of | ↓i without disturbing the system, due
to the principle of locality. Then there exists an element of physical reality
corresponding to Sz in subsystem B. (figure 1.1)
Figure 1.1: Existence of the element of physical reality corresponding to Sz in

subsystem B
However, we know that the state in (1.2) can also be expressed as

1
|Ψi = √ (| ←iA | →iB − | →iA | ←iB ) (1.3)
2
Instead of measuring Sz , we measure Sx of subsystem A, and get, say, | ←i.
Then we get that the subsystem B is in the state of | →i without disturbing the
system, due to, again, the principle of locality. Then there exists an element
of physical reality corresponding to Sx in subsystem B. (figure 1.2)
Figure 1.2: Existence of the element of physical reality corresponding to Sx in

subsystem B
In the above two measurements, we conclude that in subsystem B, there

exists simultaneous reality corresponding to the two non-commutative observ-

ables Sz and Sx , which contradicts with the previous conclusion that they must
not have simultaneous reality. Thus EPR argued that quantum-mechanical de-
scription of wavefunction or reality is incomplete, and urged for a theory that
provide a complete description.
1.2.2 Mathematical Form of the Paradox

EPR made their argument using the concept of reality, but it can also be put
in a more mathematical form. Uncertainty relation requires that
~
∆Sz ∆Sx ≥ (1.4)
2
but for the spin-singlet state, in the subsystem B, it is evident from the above
argument that
∆Sz ∆Sx = 0 (1.5)
The contradiction between (1.4) and (1.5) gives the mathematical form of the
EPR paradox.
1.2.3 Resolving the EPR Paradox by Inspecting the Un-

certainty Relations
There are several ways of resolving the EPR paradox. Mathematically, the
EPR paradox comes out from the contradiction with the uncertainty relation.
However, the variances (h(∆Sz )2 i and h(∆Sx )2 i) in the uncertainty relation
(1.4) refers to the same state, but in the relation (1.5), the two variances refer to
different states, since they are the different states after different measurements.
There is another way to look at the EPR paradox, using the concept of
mixed state density operator, which will be discussed in Chapter 2.
1.3 An Illustration using Two-Mode Squeezed

States
The above spin-singlet state is easy to digest, but the original EPR state is a
continuous state. [1] However, this state is not normalizable. Cohen suggested
a modified EPR state for discussion. [3] But we are now looking at another
continuous state - the two-mode squeezed state, or the double gaussian state.
A two-mode squeezed state is in general given by
r
b £ ¤
Ψ(x1 , x2 ) = 2 exp −b2 (x1 − x2 )2 − (x1 + x2 )2 (1.6)
π
Its plot for b = 0.3333 is shown in figure 1.3. [4] This state has the property
that for b → 0, the product in the uncertainty relation ∆x∆k approaches the
smallest possible value. It is seen from the plot that there is some correlation
between x1 and x2 , just like the spin-singlet state. Using this state, the EPR
paradox can be demonstrated by considering the position and momentum of
the second particle. We assume b ≥ 1 without loss of generality.
Figure 1.3: 3D Plot of the two-mode squeezed state with b = 0.3333
We first measure the position of the particle 1, and suppose we get ξ. Due
to this measurement, the state of the second system is changed to
· ¸ 14 · ¸
2(1 + b2 ) 2 2 2 (1 − b2 )2 2
ψ(x2 ) = exp −(1 + b )x2 + 2ξ(b − 1)x2 − ξ
π 1 + b2
The variance of the position of this state can be calculated:

Z ∞
|ξ||b2 − 1|
hx2 i = hψ|x2 |ψi = dx2 · ψ ∗ (x2 )x2 ψ(x2 ) = −
1 + b2
Z−∞∞
hx2 2 i = hψ|x2 2 |ψi = dx2 · ψ ∗ (x2 )x2 2 ψ(x2 )
−∞
(1 + b2 ) + 4ξ 2 (b2 − 1)2
= −
4(1 + b2 )2
1
h(∆x2 )2 i = hx2 2 i − hx2 i2 =
1 + b2
However, by the principle of locality, the measurement on particle 1 does not
influence the particle 2. Using EPR’s language, we say that there is an element
of physical reality corresponding to the position of the particle 2.
Now, we consider the momenta of the particles. We must first find the
Fourier transform of the state:
Z ∞ Z ∞
1
Φ(k1 , k2 ) = dx1 dx2 · Ψ(x1 , x2 ) exp(−ik1 x1 ) exp(−ik2 x2 )
2π −∞ −∞
· ¸
1 1 + b2 2 2 2(1 − b2 )
= √ exp − (k1 + k2 ) + k1 k2 (1.7)
4πb 16b2 16b2
We measure the momentum of the particle 1, and suppose we get η. Due to
this measurement, the state of the second system is changed to
µ ¶ · ¸
1 + b2 (1 − b2 )2 2 1 + b2 2 2(1 − b2 )
φ(k2 ) = exp − 2 η − k2 + ηk2
8b2 π 16b (1 + b2 ) 16b2 16b2
Then the variance of momentum of this state can be calculated:
Z ∞
|1 − b2 ||η|
hk2 i = hφ|k2 |φi = dk2 · φ∗ (k2 )k2 φ(k2 ) = −
1 + b2
Z−∞∞
hk2 2 i = hφ|k2 2 |φi = dk2 · φ∗ (k2 )k2 2 φ(k2 )
−∞
4b2 (1 + b2 ) + (1 − b2 )2 η 2
= −
(1 + b2 )2
4b2
h(∆k2 )2 i = hk2 2 i − hk2 i2 =
1 + b2
Again, by the principle of locality, the measurement on particle 1 does not
influence the particle 2, or we say that there is an element of physical reality
corresponding to the momentum of the particle 2.
We see that there exists simultaneous reality of the position and momen-
tum of the particle 2, where position and momentum are non-commutative
observables in the formalism of quantum mechanics. This again illustrates the

”incompleteness” of the formalism. Or we put it in the mathematical form
µ ¶2
2 2 2b
h(∆x2 ) ih(∆k2 ) i = ≤1 (1.8)
1 + b2
but by the uncertainty principle, for two non-commutative observables,

1 1
h(∆x)2 ih(∆k)2 i ≥ |h[x, k]i|2 =
4 4
And it exhibits the EPR paradox in the mathematical form.
Chapter 2
Schmidt Decomposition and the

Entangled States
The EPR paradox discussed in chapter 1 has made the concept of entangled
states more renowned. Actually, all the examples raised to illustrate the EPR
paradox are entangled states. In this chapter, we first review the Schmidt
decomposition, which is a mathematical technique to express a bipartite state
into a special form. We will define entangled states, and review how we mea-
sure the amount of entanglement in terms of entanglement entropy and the
participation ratio.
2.1 Schmidt Decomposition of a Composite State

The state of a composite system can be generally represented by
XX
|ΨiAB = cij |ai iA |bj iB (2.1)
i j
It can be proved that any composite state of the form of (2.1) can be repre-
sented as
Xp
|Ψi = λi |ui iA |vi iB (2.2)
i
where {|ui i, i = 1, 2, 3, . . .} and {|vi i, i = 1, 2, 3, . . .} are the orthonormal basis

√
for subsystems A and B respectively. λi are real coefficients. The repre-
sentation of the composite state in the form of (2.2) is called the Schmidt
decomposition. [5] In the Schmidt decomposition of the state, there is only one
summation sign. The state |ui iA of the subsystem A must correspond to the
state |vi iB of the subsystem B.
8
Chapter 2 Schmidt Decomposition and the Entangled States 9
The state expressed in the form of (2.2) has its significant meaning. Note
that the subsystems A and B may be spatially apart. If there are only one
terms in the Schmidt decomposition, i.e., say λ1 = 1 and other λ’s are zero,
such that |Ψi = |u1 iA |v1 iB , then it is called a separable state [6]. It is sim-
ply the direct tensor product of two states (probably spatially apart) in two
different Hilbert spaces. A separable state is operationally a state that can
be made from a pure product state by local operations and classical commu-
nication (LOCC). [7] [8]. Local operations include unitary transformations,
measurements and throwing away part of the system, each performed by one
party on his or her subsystem.
However, many composite states cannot be simply stated by only one single
term in the Schmidt decomposition. Such states are called entangled states [6].
Entangled states cannot be prepared using LOCC [8], which may be spatially
apart. Operations involving two subsystems at the same time are needed in
preparing an entangled state. Some correlations between the two subsystems
exhibit, as given by the representation of state given by (2.2). Let us consider
the Bohm’s spin singlet state given by (1.2) as an example:
1
|Ψi = √ (| ↑iA | ↓iB − | ↓iA | ↑iB )
2
It is obviously an entangled state because the eigenkets (| ↑iA and | ↓iA , and
| ↑iB and | ↓iB ) in each subsystems form an orthonormal basis, and there
are two (more than one) terms in the Schmidt decomposition. This state
is not prepared locally when the two spin- 12 particles are spatially apart; it is
supposed to be prepared at the same position, and is then brought apart. Some
information about the state is stored as the non-local correlation between the
two particles.
2.1.1 Proof of the Existence of the Schmidt Decompo-

sition
It can be shown that any composite states have its own Schmidt decomposition
given by (2.2). The proof is shown below:
Proof: Suppose for subsystem A there is a base {|ai i, i = 1, 2, 3, . . .} such
that the reduced density operator for system A, denoted by ρA , is diagonal.
The orthonormal bases for subsystem B is arbitrarily chosen to be {|b0j i, j =
1, 2, 3, . . .}. Then any state |Ψi can be written as

XX
|Ψi = cij |ai iA |b0j iB (2.3)
i j
and the corresponding density operator of this pure state is given by
ρ = |ΨihΨ|
XXXX
= cij c∗kl |ai iA |b0j iB A hak |B hb0l |
i j k l
The reduced density operator for subsystem A is given by
ρA = trB (ρ)
XXX
= cij c∗kj |ai iAA hak |
i j k
Since we have supposed that ρA is diagonal, there must exist some complex
numbers gi (i=1,2,3,. . . ) such that
X
cij c∗kj = |gi ||gk |δik
j
= |gi |2 δik (2.4)
Obviously, we can construct a set of orthonormal bases in subsystem B

X cij
|bi i = |b0j i (2.5)
j
|gi |
for each |gi | 6= 0. It is evident that |bi i, i = 1, 2, 3, . . . form a set of orthonormal

bases:
X X c∗ij ckl
hbi |bk i = hb0j |b0l i
j l
|gi ||gk |
XX
= δik δjl
j l
= δik
where equation (2.4) has been applied. Putting equation (2.5) in equation
(2.3) gives
XX
|ΨiAB = cij |ai iA |b0j iB
i j
X X cij X
= |gi ||ai iA |b0j iB = |gi ||ai iA |bi iB
i j
|gi | i
Xp
= λi |ai iA |bi iB
i
√
where in the last line we replace |gi | = λi because it is real. Hence we get
the representation in the form of equation (2.2). ¶
2.1.2 Schmidt Decomposition in Continuous Systems

There are many quantum states that can be described by projecting them in
continuous variables. The most common examples of the projected variables
are the positions and the momenta of the particles. In wave mechanics, states
are denoted by wavefunction, which is merely the state projected onto the
position. There are other continuous system like the Fock-Bargmann repre-
sentation. In this report, we will mainly deal with continuous system. Hence
we must study the validity of the existence of Schmidt decomposition in con-
tinuous systems.
A continuous system can be generally denoted by
Z Z
|ΨiAB = dx dy · ψ(x, y)|xiA |yiB (2.6)
The reduced density operator of the subsystem A is

Z Z
ρA = dx dx0 · ρA (x, x0 )|xiAA hx0 | (2.7)
where Z
0
ρA (x, x ) = dy · ψ(x, y)ψ ∗ (x0 , y) (2.8)
In the proof given by the previous section, we have assumed ρA is diago-

nalized. To find the Schmidt decomposition of (2.6), we must first diagonalize
ρA . The eigenvalue equation is
ρA |φi iA = λi |φi iA (2.9)
and it is equivalent to
Z
dx0 · ρA (x, x0 )φi (x0 ) = λi φi (x) (2.10)
where i is any integer. ρA (x, x0 ) in (2.8) is called the kernel of the integral equa-
tion (2.10). λi and φi (x) are the eigenvalues and eigenfunctions respectively.
However, the existence of solutions of (2.10) is not always guaranteed. There
are necessary conditions that there are existence of the solution of the integral
equations (2.10). As long as the kernel ρA (x, x0 ) in the integral equation is

quadratically integrable, i.e.,
Z Z
dx dx0 · ρA (x, x0 )
converges, there exist an orthonormal set of eigenstates such that

Z
hφi |φj i = dx · φ∗i (x)φj (x) = δij
[9] Then the kernel in the integral equation can be expressed as

X
ρA (x, x0 ) = λi φi (x)φ∗i (x0 )
i
and the reduced density operator of A is

Z Z X X
ρA = dx dx0 λi φi (x)φ∗i (x0 )|xiAA hx0 | = λi |φi ihφi |
i i
which is diagonalized. The diagonalizability of ρA means the existence of the

Schmidt decomposition of (2.6) in terms of (2.2).
2.1.3 Algorithm of Finding the Schmidt Decomposition

of a State
It is not easy to compute the Schmidt decompositions, especially for the con-
tinuous states. Unless there exists a specific formula, we usually have to find
the decomposition numerically.
We know that for continuous variables basis, say, the position of a particle,
hx|x0 i = δ(x − x0 )
where δ(x − x0 ) is the Dirac delta function. It is analogous to the discrete

eigenkets basis
hψi |ψj i = δij
where δij is the Kronecker delta symbol. Therefore, it is possible to discretize

the continuous variable to be the basis.
The steps of doing Schmidt decomposition are described as follow:
1. Discretize the wavefunction. Suppose the interval is discretized into

N bases, denoted by {x11 , x12 , x13 , . . . , x1N } and {x21 , x22 , x23 , . . . , x2N }.
Then a continuous state in the form of (2.6) is discretized as follow:
Z Z
|ΨiAB = dx1 dx2 · ψ(x1 , x2 )|x1 iA |x2 iB
XX
≈ ψij |x1i iA |x2j iB (2.11)
i j
2. Find the reduced density operator of the subsystem B. We can

find the reduced density operator of the subsystem B from (2.11), which
is given by Ã !
XX X
ρB ≈ ψij ψil∗ |x2j iB B hx2l | (2.12)
j l i
3. Evaluate the eigenvalues and eigenvectors of ρB . This step can be

done with routines provided in some mathematical libraries. In general,
the eigenkets can be expressed as
X
|φi iB = tij |x2j iB (2.13)
j
4. Find the inverse of the transformation matrix. We find the in-

verse of the matrix tij . It can be done, again, with routines in some
mathematical libraries. In general,
X
|x2j iB = (t−1 )lj |φl iB (2.14)
l
5. Find the new eigenkets in subsystem A. Putting (2.14) in (2.11)

gives
Ã !
X XX
|ΨiAB ≈ ψij (t−1 )lj |x1i iA |φl iB
l i j
Then the unnormalized eigenkets in subsystem A is given by

XX
|ψ̃l iA = ψij (t−1 )lj |x1i iA (2.15)
i j
6. Normalize the unnormalized eigenket of subsystem A and get

the Schmidt coefficients We normalize the kets in (2.15) and we get
the Schmidt coefficients:
λi = A hψ̃l |ψ̃l iA (2.16)

|ψ̃l iA |ψ̃l iA
|ψl i = q = √ (2.17)
λi
A hψ̃l |ψ̃l iA
Then, after these steps, the Schmidt coefficients and the eigenkets in both
subsystems A and B are found.
2.1.4 Schmidt Decomposition of the Two-Mode Squeezed

State
A two-mode squeezed state is the state that is given in (1.6):
r
b £ ¤
Ψ(x1 , x2 ) = 2 exp −b2 (x1 − x2 )2 − (x1 + x2 )2
π
It can be expressed in terms of Schmidt decomposition in the form of
∞ p
X
Ψ(x1 , x2 ) = λn fn (x1 )fn (x2 ) (2.18)
n=0
We expect the two Schmidt modes in two subsystems are the same because if
we exchange x1 and x2 , the total state is unchanged, i.e.,
Ψ(x1 , x2 ) = Ψ(x2 , x1 ) (2.19)
Of course we can find the Schmidt decomposition of the two-mode squeezed

state using the algorithm described in the previous section. However, we would
like to do it analytically by using the Mehler’s formula for Hermite polynomial
[10]:
∞
X zn 1 2xyz − z 2 (x2 + y 2 )
Hn (x)Hn (y) = √ exp[ ] (2.20)
n=0
2n n! 1 − z2 1 − z2
where Hn (y) are Hermite polynomials of order n, where n can be any non-
negative integers.
We have to employ the eigenfunctions of harmonic oscillator for the sake

of normalization. The eigenfunctions are given by
s √
2 b √
un (x) = 1 Hn (2 bx) exp(−2bx2 ) (2.21)
π 2 n!2n
We know that {un (x), n = 0, 1, 2, ...} forms an orthonormal set of functions,
i.e., Z ∞
dx · um (x)un (x) = δmn (2.22)
−∞
Then, we can now perform the Schmidt decomposition by employing (2.20)

and (2.21). Firstly we should expand the terms inside the exponential function
Ψ(x1 , x2 )
r
b
= 2 exp[−b2 (x1 − x2 )2 − (x1 + x2 )2 ]
rπ
b
= 2 exp[−(b2 + 1)(x1 2 + x2 2 ) + 2(b2 − 1)x1 x2 ]
π
Here comes with one of the tricky steps here - splitting the exponential into
two parts:
r
b
= 2 exp[−2b(x1 2 + x2 2 )] exp[−(b − 1)2 (x1 2 + x2 2 ) + 2(b2 − 1)x1 x2 ]
π
Another step here is to write the expression into a form so that (2.20) can be
used:
r ½ · ¸¾
b 2 2 2 b−1 b−1 2 2 2
= 2 exp[−2b(x1 + x2 )] exp (b + 1) 2 x1 x2 − ( ) (x1 + x2 )
π b+1 b+1
r ( " #)
b−1 2 2 2 b−1
b −( b+1
) (x 1 + x 2 ) + 2 b+1
x 1 x 2
= 2 exp[−2b(x1 2 + x2 2 )] exp 4b
π 1 − ( b−1 b+1
)2
r (" √ √ √ √ #)
b−1 2 2 2 b−1
b −( b+1
) ((2 bx 1 ) + (2 bx 2 ) ) + 2 b+1
(2 bx 1 )(2 bx 2 )
= 2 exp[−2b(x1 2 + x2 2 )] exp
π 1 − ( b−1
b+1
)2
Now apply the Mehler’s formula (2.20):

r s µ ¶2 ∞
b b−1 X √ √ | b−1 |n
= 2 2 2
exp[−2b(x1 + x2 )] 1 − Hn (2 bx1 )Hn (2 bx2 ) b+1
π b + 1 n=0 2n n!
Then using the normalized function (2.21) to group the constants properly and
we will get
Ψ(x1 , x2 )
s
X∞
4b(b − 1)2n
=
n=0
(b + 1)2(n+1)
s √  s √ 
2 b √ 2 b √
· 1 Hn (2 bx1 ) exp(−2bx1 2 )  1 Hn (2 bx2 ) exp(−2bx2 2 )
π 2 2n n! π 2 2n n!
∞ p
X
= λn fn (x1 )fn (x2 )
n=0
where
s
p 4b(b − 1)2n
λn = (2.23)
(b + 1)2(n+1)
s √
2 b √
fn (x) = 1 Hn (2 bx) exp(−2bx2 ) (2.24)
π 2 2n n!
We see that fn (x) are eigenstates of quantum harmonic oscillator. The Schmidt
coefficients of the state with b = 0.256027 is plotted as shown in Fig. 2.1. Their
Schmidt modes are plotted as shown in figure 2.2. [4]
Figure 2.1: Schmidt coefficients of the two-mode squeezed state with b = 0.256027
1.5 1.5
1 1
0.5 0.5
x x
-2 -1 1 2 -2 -1 1 2
-0.5 -0.5
-1 -1
-1.5 -1.5
i=0 i=1
1.5 1.5
1 1
0.5 0.5
x x
-2 -1 1 2 -2 -1 1 2
-0.5 -0.5
-1 -1
-1.5 -1.5
i=2 i=3
1.5 1.5
1 1
0.5 0.5
x x
-2 -1 1 2 -2 -1 1 2
-0.5 -0.5
-1 -1
-1.5 -1.5
i=4 i=5
Figure 2.2: The first six Schmidt modes of the two-mode squeezed state
We are now going to look at the entanglement of the two-mode squeezed

state. There are two cases:
Case I: If b = 1, then
"µ ¶ 1 # "µ ¶ 1 #
4 4 4 4
Ψ(x1 , x2 ) = exp(−2x1 2 ) · exp(−2x2 2 )
π π
Therefore, the state is a separable state, since it got only one term in
Schmidt decomposition, or because the state can be factorized into the
product of a function of x1 and another function of x2 .
Case II: If b 6= 1, then

∞ p
X
Ψ(x1 , x2 ) = λn fn (x1 )fn (x2 )
n=0
Hence it is an entangled state, because there is an infinite number of

terms in the Schmidt decomposition.
2.2 Properties of Entangled States

With the technique of Schmidt decomposition, we are now able to further
study the entangled states. We know that either subsystems of an entangled
state is a mixed state, rather than a pure state.
2.2.1 Mixed Subsystems of Entangled States

We are going to see that an entangled state a state with its subsystems being a
mixed state. A state which is not an entangled state is then a separable state.
The proof is shown below:
Proof: Suppose the state |Ψi is an entangled state, with more than one
term in the Schmidt decomposition, i.e.,
Xp
|Ψi = λi |ui iA |vi iB
i
where more than one λi ’s have non-zero values and all λi < 1.
Then the reduced density operator of, say, subsystem A is given by
X
ρA = λi |ui iAA hui |
i
Hence, its square is simply given by

X
ρA 2 = λi 2 |ui iAA hui |
i
X
6= λi |ui iAA hui | = ρA
i
It is seen that the reduced density operator of subsystem A is not idempotent,

and thus subsystem A is a mixed state. By similar argument, it can be shown
that the subsystem B is also a mixed state. ¶
Therefore, to see whether a state is an entangled state, we can see whether
the reduced density operator is idempotent.
First, we take the Bohm’s spin singlet state given by (1.2) as an example:
1
|Ψi = √ (| ↑iA | ↓iB − | ↓iA | ↑iB )
2
It is entangled because its reduced density operator of subsystem A (say) is
1
ρA = (| ↓iB B h↓ | + | ↑iB B h↑ |)
2
and it is obviously a density operator of mixed state since it is not idempotent:

1 1
ρA 2 = (| ↓iB B h↓ | + | ↑iB B h↑ |) = ρA 6= ρA
4 2
Therefore, the spin-singlet state is an entangled state.
Let’s turn our attention to two-mode squeezed state. In the previous sec-
tion, we show, from the Schmidt decomposition of the state, that for b = 1,
the state is separable; otherwise, the state is entangled. The reduced density
operator of subsystem A is given by (2.7):
Z ∞ Z ∞
ρA = dx1 dx01 · ρA (x1 , x01 )|x1 iAA hx01 |
−∞ −∞
where the kernel of the two-mode squeezed state is given by
Z ∞
0
ρA (x1 , x1 ) = dy · ψ(x1 , x2 )ψ ∗ (x01 , x2 )
−∞
√ · 4 ¸
2 2b b + 6b2 + 1 2 02 (b2 − 1)2 0
= p exp − (x1 + x1 ) + 2 x1 x1
π(b2 + 1) 2(b2 + 1) b +1
The squared density operator is then given by
Z ∞ Z ∞
2 (2)
ρA = dx1 dx01 · ρA (x1 , x01 )|x1 iAA hx01 |
−∞ −∞
where
Z ∞
(2)
ρA (x1 , x01 ) = dz · ρA (x1 , z)ρA (z, x01 )
−∞
8b2
= p
π(b2 + 1)(b4 + 6b2 + 1)
· 4 ¸
b + 6b2 + 1 2 02 (b2 − 1)4 (x1 + x01 )2
· exp − (x1 + x1 ) +
2(b2 + 1) 4(b2 + 1)(b4 + 6b2 + 1)
There are two cases:
Case I: If b = 1, then
2 h i
(2) 2
ρA (x1 , x01 ) = ρA (x1 , x01 ) = √ exp −2(x1 2 + x01 )
π
Therefore, the state is a separable state.
Case II: If b 6= 1, then

(2)
ρA (x1 , x01 ) 6= ρA (x1 , x01 )
Hence it is an entangled state.
Later we will go back to these two examples when we are talking another
equivalent definition of entangled states.
2.2.2 EPR Paradox Revisited

In fact the states used to illustrate the EPR paradox are entangled state. Take
the spin-singlet state (1.2) as an example. As shown above, the reduced density
operators of A and B of the state are those of mixed states. Therefore, the
variances of the spins in z and x directions in subsystem B should be given by
£ ¤
h(∆SzB )2 i = tr ρB (∆SzB )2
£ ¤
h(∆SxB )2 i = tr ρB (∆SxB )2
But in the EPR paradox, EPR neglected the fact that any subsystems of the
state are actually mixed states, and mistakenly claim that the two variances
above are merely zero. With the use of reduced density operators, we resolve
the paradox.
2.2.3 Duan’s Condition for Entanglement of Continuous

Systems
When we are dealing with mixed state, we would rather adopt another defini-
tion for entanglement. A state is non-entangled (separable) if and only if its
density operator can be expressed in the following form:
X
ρ= pi ρiA ⊗ ρiB (2.25)
i
[11] It can be reduced to the case of a pure state where only one pi is unity
while all others are zero. It is the same definition using the density operator
approach in the previous section. Using this definition, we can derive another
sufficient criterion for the entanglement of a state.
Suppose there are a pair of non-commutative observables for each of the
two subsystems A and B, satisfying the following commutation relation:
[xi , kj ] = iδij (2.26)
They are just like positions and momenta of different particles. Define the
following type of EPR-like operators:
(
u = |a|x1 + a1 x2
(2.27)
v = |a|k1 − a1 k2
For a non-entangled mixed state described by the density operator ρ, the

total variance of the EPR-like operators must satisfy the following inequality
[11]:
1
h(∆u)2 i + h(∆v)2 i ≥ a2 + 2 (2.28)
a
Proof : By (2.25),
X
h(∆u)2 i + h(∆v)2 i = pi (hu2 ii + hv2 ii ) − hui2 − hvi2
i
and by (2.27),
X µ ¶
2 2 1 2 2 2 1 2
= pi a hx1 ii + 2 hx2 ii + a hp1 ii + 2 hp2 ii
i
a a
Ã !
a X X
+2 pi hx1 ii hx2 ii − pi hp1 ii hp2 ii − hui2 − hvi2
|a| i i
X µ 1 1
¶
2 2 2 2 2 2
= pi a h(∆x1 ) ii + 2 h(∆x2 ) ii + a h(∆p1 ) ii + 2 h(∆p2 ) ii
i
a a
Ã !2 Ã !2
X X X X
+ pi huii − pi huii i + pi hvii − pi hvii i
i i i i
Because
h(∆xj )2 ii + h(∆pj )2 ii
q
≥ 2 h(∆xj )2 ii h(∆pj )2
q
≥ |h[xj , pj ]i|2
= 1
and
Ã !Ã ! Ã !2
X X X
2
pi pi huii ≥ pi |huii |
i i i
Hence the left hand side of the inequality (2.28) is greater than a2 + a12 .¶
On the other hand, if the inequality (2.28) is violated, the state is entangled.
(Note: but the converse is not necessarily true. Therefore, there are cases that
even if the state is entangled, the inequality is not violated.) If we are dealing
with Gaussian states, it can be proved that the inequality (2.28) is satisfied if
and only if the state is separable (non-entangled). [11]
We take the two-mode squeezed state as an example. We set a = 1 in

(2.27). Then
hui = hx1 i + hx2 i = 0

1
hu2 i = hx1 2 i + 2hx1 x2 i + hx2 2 i =
4
1
h(∆u)2 i =
4
and
hvi = hk1 i + hk2 i = 0

hv2 i = hk1 2 i + 2hk1 k2 i + hk2 2 i = 4b2
h(∆v)2 i = 4b2
Therefore,
1
h(∆u)2 i + h(∆v)2 i = + 4b2 (2.29)
4
√
If b < 47 , then the inequality (2.29) violates (2.28). Then we know that in
this range of b, the two-mode squeezed state is an entangled state.
2.3 Quantifying Entanglement

The degree of entanglement can be measured by the several kinds of quantities.
Let us denote such quantity by E. Such quantities are artificially defined, but
must satisfy the requirements [7]:
1. E vanishes if the state is separable (non-entangled).
2. E is invariant under local unitary transformations.
3. E does not increase under LOCC.
4. It should be additive for tensor products of independent states, shared

among the same set of observers.
In this section, we are just talking about the entanglement entropy of a

pure state.
2.3.1 Entanglement Entropy

To quantify entanglement of a pure state, say, denoted by |Ψi, we use the von
Neumann entropy of one of the partial density operators of the state. In other
fields in physics, entropy is the measurement of ”amount of chaos” of a system,
or the degree of lacking information of a system. In entangled states, some
”hidden information” is not ”stored” in either subsystems by local operations.
Thus, an entangled state must exhibit some lacking of information. [12] En-
tropy, often used in statistical mechanics, is a good measure of entanglement
for a state. The von Neumann entropy S of a square matrix ρ is defined as
the following trace:
S(ρ) = −tr[ρ log2 ρ] (2.30)
It bears resemblance to the entropy in statistical mechanics. [13] The entan-
glement entropy of the state |Ψi is given by
S = (S ¦ trB )(|ΨihΨ|) (2.31)
If the Schmidt decomposition of the state |Ψi is given by (2.2):

Xp
i
, then the entanglement entropy of the state is then given by

X
S=− λi log2 λi (2.32)
i
The entanglement entropy would be the maximum if all the Schmidt terms
are equally probable, i.e., λ1 = λ2 = . . . = λN .
If the state is separable (non-entangled), then S = 0.
Obviously, if we perform any local unitary transformation on either subsys-
tems, the value of the entanglement entropy is not changed, since from (2.32),
the entropy depends on the Schmidt coefficients λi ’s.
2.3.2 Participation Ratio

In the calculation of the entropy, the logarithm of the Schmidt coefficients is
involved. Calculations involving logarithm are rather computationally expen-
sive. Because of this, we define the participation ratio of the state given by
(2.2):
1
K=P 2 (2.33)
i λi
It is also a measurement of the amount of entanglement of a state. In fact, the

participation ratio is related to the entanglement entropy. By the definition of
entanglement entropy given by (2.32),
X X
S = − λi log2 λi = λi log2 [1 + (λi − 1)]
i i
X
≈ − λi (λi − 1)
i
X 1
= 1− λi 2 = 1 −
i
K
where we have assumed λi ¿ 1 for all i. The above expression gives the clue
that K increases/decreases as S does however the state is changed. Hence, K
can be used to measure the amount of entanglement as S can. A more rigorous
proof is given in the following. Suppose there are two sets of Schmidt coeffi-
cients, denoted by λi (1) and λi (2) . Using the fact that log2 x is an increasing
function in x for 0 < x < 1. Then
λi (1) > λi (2) equivalent to log2 λi (1) > log2 λi (2)
2 2
=⇒ λi (1) > λi (2) equivalent to λi (1) log2 λi (1) > λi (2) log2 λi (2)
2 2
X 2 X 2 X X
i i i i
1 1 X X
=⇒ P <P equivalent to − λi (1) log2 λi (1) < − λi (2) log2 λi (2)
(1) 2 (2) 2
i λi i λi i i
It ensures that S (1) < S (2) is equivalent to K (1) < K (2) . Hence, K in-
creases/decreases as S does however the state is changed.
If the state is separable (non-entangled), then K = 1, which is the minimum
value of the participation ratio.
Obviously, the calculation of the participation ratio is much more easier
than the entanglement entropy, since we just need to add up the squared
Schmidt coefficients. There is another advantage of the use of participation
ratio in the computational aspect. From the Schmidt decomposition of the
state |Ψi given by (2.2), the reduced density operator of subsystem A, say, is
given by
X
ρA = λi |ai iAA hai |
i
and its square is

X
ρA 2 = λi 2 |ai iAA hai |
i
Hence, its trace is

X 1
tr(ρA 2 ) = λi 2 = (2.34)
i
K
But we know that the trace is base-independent. This means we can get the
value of participation ratio by directly finding the reduced density operator
of the state |Ψi without doing the Schmidt decomposition. However, the
calculation of the participation ratio by tracing the squared reduced density
operator would be very time-taking since we have to deal with several levels
of summation.
2.3.3 Entanglement under Various Transformations

The degree of entanglement would be unchanged under various local unitary
transformations [14]. We can see this from the Schmidt decomposition. Sup-
pose the Schmidt decomposition is given in the form of (2.2):
Xp
i
We denote the two local unitary transformation by LA and LB , where

(
LA |ui iA = |ũi iA
(2.35)
LB |vi iB = |ṽi iB
and
LA † LA = LA LA † = 1
LB † LB = LB LB † = 1
Suppose we transform the state |Ψi by the operator L = LA ⊗ LB to |Φi.

Then the Schmidt decomposition of |Φi can be found at
|Φi = L|Ψi
Xp
= λi (LA |ui iA ) (LB |vi iB )
i
Xp
= λi |ũi iA |ṽi iA
i
The entanglement entropy and the participation ratio of |Φi are just equal
to that of |Ψi. Local operations on the composite states do not affect the
entanglement of the states.
One of these examples is the translation of the coordinates, which is done
by the unitary operator exp[− ip~x x ]. Suppose the system is given originally as
Xp
Ψ(x1 , x2 ) = λn φn (x1 )ψn (x2 )
n
where the Schmidt modes φn (x1 ) and ψn (x2 ) form two orthonormal sets, i.e.,
Z
dx1 · un (x1 )um (x1 ) = δmn
Z
dx2 · vn (x2 )vm (x2 ) = δmn
Now the system is changed in a way that the reference point (0, 0) is translated
into another location, or quantitatively, the coordinates (x1 , x2 ) is transformed
to another set of coordinates (x01 , x02 ) by
(
x01 = x1 + x01
x02 = x2 + x02
The entanglement does not change because the Schmidt coefficients does not
change:
Ψ0 (x01 , x02 ) = Ψ(x01 − x01 , x02 − x02 )

Xp
= λn · φn (x01 − x01 ) · ψn (x02 − x02 )
n
and the Schmidt coefficients of the initial and final states are not changed.
Hence the entanglement entropy and the participation ratio are not changed
too.
Unlike local operations, non-local operations is highly likely to change the

degree of entanglement of the system. One of the examples is the ”rotation”.
Suppose the system is changed in a way that the reference point is unchanged,
but on the x1 -x2 plane, the axes rotated by an angle θ in an anti-clockwise
direction, or quantitatively
Ã ! Ã !Ã !
x01 cos α sin α x1
=
x02 − sin α cos α x2
After rotation,
Ψ0 (x01 , x02 ) = Ψ(x01 cos α − x02 sin α, x01 sin α + x02 cos α)
Xp
= λn φn (x01 cos α − x02 sin α)ψn (x01 sin α + x02 cos α)
n
Ã !Ã !
Xp X√ Xp
= λn ηk fk (x01 )gk (x02 ) ξl hl (x01 )pl (x02 )
n k l
Xp
= λ0kl φ0kl (x01 )ψkl
0
(x02 )
kl
where
 P √
 0
 λkl = n λn ηk ξl
φ0kl (x01 ) = fk (x01 )hl (x02 )

 0 0
ψkl (x2 ) = gk (x01 )pl (x02 )
The entanglement entropy of this state is given by

X
S0 = − λ0kl log2 λ0kl
kl
which is in general not equal to (2.32). After rotation, it is not always possible
to reserve the entanglement of the state. A vivid example is the two-mode
squeezed state given by (1.6):
r
b £ ¤
Ψ(x1 , x2 ) = 2 exp −b2 (x1 − x2 )2 − (x1 + x2 )2
π
Suppose there is a non-local operation, denoted by =, transforms the system
by the rotation
(
u1 = x1√+x
2
2
x1√
−x2
u2 = 2
Then the new state is given by
Φ(u1 , u2 ) = =Ψ(x1 , x2 )
r
b ¡ ¢
= 2 exp −2u1 2 − 2b2 u2 2
π
which is certainly a separable state, although the original state is entangled
for b 6= 1. This non-local operation on the state alters its entanglement.
2.3.4 Participation Ratio of the Two-Mode Squeezed

State
The two-mode squeezed state given by (1.6) has the Schmidt decomposition
given by (2.23) and (2.24). Using the result of the Schmidt decomposition,
1 1
K(b) = (b + ) (2.36)
2 b
The variation of K as a function of b is plotted as shown in figure 2.3 [4]. Note
that when b = 1, K = 1, and it means the state is separable (non-entangled)
here. This conclusion complies with the previous illustrations about the en-
tangled states that when b = 1, the two-mode squeezed state is a separable
state.
We can observe the reciprocal symmetry:
µ ¶
1
K(b) = K (2.37)
b
It is expected because if b is replaced by 1b , the dominant part of the correlation

is just changed from the x1 − x2 to x1 + x2 part.
Figure 2.3: Participation ratio K as a function b for the two-mode squeezed state
Chapter 3
Entanglement in Various
Oscillating Systems
In chapter 2, we have reviewed the basic formalism of quantum entanglement.
In this chapter, we are going to explore how entanglement behaves in various
oscillating systems. These oscillating systems are described in continuous vari-
ables. We explore how entanglement behaves as the parameters of the system
(e.g., the width / standard deviation of the centre of mass) changes. Schmidt
modes (the eigenfunctions in the subsystems in the Schmidt decomposition)
are also found.
3.1 Composite Morse Oscillators

3.1.1 Morse Potential
Intermolecular potential are often approximated by the Morse potential given
by [15]
£ ¤
V (x) = D (1 − e−ax )2 − 1 (3.1)
where D gives the dissociation energy between the two atoms, and the pa-
rameter a gives the range of the potential. Suppose the particles are trapped
under this potential. Its wavefunction can be solved by the following time-
independent Schrödinger equation:
~2 d2 ψ(x)
− + V (x)ψ(x) = Eψ(x) (3.2)
2m dx2
where m is the reduced mass of the diatomic molecule and E is the energy
eigenvalues of the molecule. We introduce the following transformation of
30
Chapter 3 Entanglement in Various Oscillating Systems 31
variables: 

 χ = ax

 T = ~a2 t
2m
(3.3)


2mD
² = ~2 a2


ρ = 2mE
~2 a2
and the Schrödinger equation in (3.2) becomes
d2 ψ £ −χ 2
¤
− + ² (1 − e ) − 1 ψ = ρψ (3.4)
dχ2
ρ gives the information about the energy eigenvalues and ² gives the informa-
tion about the dissociation energy. Suppose we have a diatomic molecule with
reduced mass m = 0.5u (u: atomic mass unit) under the Morse potential with
its range given by a = 3.732
a0
(a0 : Bohr radius) and the dissociation energy given
by D = 5.0eV . Then after variable transformation, the parameter concerning
the dissociation energy becomes ² = 24.01. The plot of the Morse potential is
given in figure 3.1. [4]
Figure 3.1: Plot of the Morse potential (rescaled) for ² = 24.01
The eigenstates and energy eigenvalues of (3.4) are given by [16]

"µ ¶ µ ¶2 #
√ 1 1
ρν = 2 ² ν + − ν+ (3.5)
2 2
ξ
ψjν = Njν e− 2 ξ j−ν L2j−2ν
ν (ξ) (3.6)
where L2j−2ν
ν (ξ) are the associated Laguerre polynomials. The normalization
constant Njν is given by
2ν!(j − ν)
Njν 2 = (3.7)
Γ(2j − ν + 1)
and the Morse variable is given by
ξ = (2j + 1)e−χ (3.8)
j is a parameter given by
√ 1
j= ²− (3.9)
2
In (3.5) and (3.6), ν is an integer ranging from 0 and [j] (the largest integral
part of j). The allowed bound eigenstates are given by these values of ν, and
thus there are only [j] + 1 bound states for the Morse system. In the set of
parameters stated above, j = 4.4 and therefore there are only five bound eigen-
states, as plotted in figure 3.2. [4] Note that unlike the harmonic oscillators,
the eigenstates of Morse oscillators are neither odd or even functions.
y y
1 1
0.75 0.75
0.5 0.5
0.25 0.25
c c
2 4 6 8 10 2 4 6 8 10
-0.25 -0.25
-0.5 -0.5
-0.75 ν=0 -0.75 ν=1
y y
1 1
0.75 0.75
0.5 0.5
0.25 0.25
c c
2 4 6 8 10 2 4 6 8 10
-0.25 -0.25
-0.5 -0.5
-0.75
ν=2 -0.75
ν=3
1
0.75
0.5
0.25
c
2 4 6 8 10
-0.25
-0.5
-0.75
ν=4
Figure 3.2: Plot of the eigenstates under Morse potential (rescaled) for ² = 24.01
(j = 4.4)
3.1.2 Entanglement of Composite Morse Oscillators

In the previous section we considered the oscillations of the reduced mass under
the Morse potential. In this section we are going to discuss about a more
realistic case - two particles are connected together by the Morse potential,
while they are at the same time trapped under a harmonic potential. The
Hamiltonian of such system should be given by
H = HCM + HMorse (3.10)
where HCM is the Hamiltonian of the centre of mass, and in this case it
is the Hamiltonian of a harmonic oscillator, and HMorse is the interaction
Hamiltonian, which is essentially the Hamiltonian of the Morse oscillator. If
the centre of mass is just at the ground state of the harmonic Hamiltonian,
the state of this composite system of two particles (assuming equal masses) is
given by
µ 2 ¶ 14
8b 2 2
Ψjν (x1 , x2 ) = e−b (x1 +x2 ) ψjν (x1 − x2 ) (3.11)
π
where b is related to the uncertainty of the centre of mass of the composite sys-
tem. This is not a symmetric state. Unlike the two-mode squeezed state, there
are no formulae like the Mehler’s formula for doing the Schmidt decomposition.
Hence, we must perform Schmidt decomposition and study its entanglement
property numerically.
The Schmidt decomposition for ² = 24.01 is done for ν = 0, 1, 2, 3, 4. The
Schmidt coefficients are evaluated numerically and plotted as shown in fig-
ure 3.3. Since the state (3.11) is not symmetric, the Schmidt modes are not
identical for the two subsystems, as shown in figures 3.4, 3.5, 3.6, 3.7 and 3.8.
Schmidt Coefficients of Compound Morse Oscillator
1 ν
0
0.9
1
0.8 2
3
0.7
4
b=0.1
λi 0.6
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
i
Figure 3.3: Plot of the first 10 Schmidt coefficients of the composite Morse oscillator
for ² = 24.01 (j = 4.4) and b = 0.1
Schmidt Modes of Morse Oscillator b=0.1 ν=0

1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=1 i=2 i=3
1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B
0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=4 i=5 i=6
Figure 3.4: Schmidt modes of the composite Morse oscillator for ² = 24.01 (j = 4.4),
b = 0.1 and ν = 0

1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=1 i=2 i=3
1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B
0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=4 i=5 i=6
b = 0.1 and ν = 1

1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=1 i=2 i=3
1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B
0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=4 i=5 i=6
b = 0.1 and ν = 2
1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=1 i=2 i=3
1 1 0.5
0.5 0.5
0
0 0
A 0.5
0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5
0.5
B 0 0
0
0.5 0.5
1 1 0.5
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=4 i=5 i=6
b = 0.1 and ν = 3
Schmidt Modes of Morse Oscillator b=0.1 n=4

1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=1 i=2 i=3
1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
1 1 1
0.5 0.5 0.5
B
0 0 0
0.5 0.5 0.5
1 1 1
20 10 0 10 20 20 10 0 10 20 20 10 0 10 20
x x x
i=4 i=5 i=6
b = 0.1 and ν = 4
We can observe in figures 3.4, 3.5, 3.6, 3.7 and 3.8 that the Schmidt modes
must have the following characteristics:
φi (x) = ±ψi (−x) (3.12)
The Schmidt modes of the same states are mirror images of each other. In fact,
it is true for any non-degenerate wavefunction given in the form of (3.11) where
the eigenstates of the reduced mass ψjν can be taken in any form. Now, we
give a mathematical explanation of this. Suppose the Schmidt decomposition
takes the form
µ 2 ¶ 14 X∞ p
8b 2 2
Ψ(x1 , x2 ) = e−b (x1 +x2 ) ψjν (x1 − x2 ) = λi φi (x1 )ψi (x2 )
π i=0
We interchange the position of x1 and x2 , getting

µ 2 ¶ 14 X∞ p
8b −b2 (x1 +x2 )2
Ψ(x2 , x1 ) = e ψjν (x2 − x1 ) = λi ψi (x1 )φi (x2 )
π i=0
but it can also be written in the form

µ 2 ¶ 14 X∞ p
8b −b2 (x1 +x2 )2
Ψ(x2 , x1 ) = e ψjν ((−x1 ) − (−x2 )) = λi φi (−x1 )ψi (−x2 )
π i=0
Obviously, comparing the above two forms of Ψ(x2 , x1 ), we get

∞ p
X ∞ p
X
λi ψi (x1 )φi (x2 ) = λi φi (−x1 )ψi (−x2 )
i=1 i=1
If the Schmidt decomposition of this composite wavefunction is non-degenerate,

then by the completeness relation, we know that
ψi (x1 )φi (x2 ) = φi (−x1 )ψi (−x2 )
and by separation of variables,

ψi (x1 ) ψi (−x2 )
= =C (3.13)
φi (−x1 ) φi (x2 )
where C is a constant that does not depend on x1 and x2 . Without loss of
generality, consider ψi (x1 ) = Cφi (−x1 ) and the normalizability of the Schmidt
modes,
Z ∞ Z ∞
∗ 2
dx1 · ψi (x1 )ψi (x1 ) = |C| dx1 · φ∗i (−x1 )φi (−x1 )
−∞ −∞
|C|2 = 1
C = ±1
Putting C = ±1 back to (3.13), we will resume the relation (3.12) about the
Schmidt modes of the composite Morse oscillator (3.11) given by the figures
3.4, 3.5, 3.6, 3.7 and 3.8.
The participation ratio of the composite Morse oscillator given by (3.11)

against the parameter b is plotted as shown in figure 3.9. It is seen from
the figure that for large b (small uncertainty for the position of the centre of
mass of the two molecules), the more excited the system is (at a state with
larger ν, the more entangled the system is (the larger the participation ratio).
However, when b decreases to smaller values (uncertainty in position of the
centre of mass increases to a larger value), this rule does not hold.
Participation Ratio of Compound Morse Oscillator
15
10
5 ν
4
3
2
1
0
0
0 1 2
b
Figure 3.9: Variation of the participation ratio of the composite Morse oscillator
against b for ² = 24.01 (j = 4.4) for ν = 0, 1, 2, 3, 4
3.2 Compound Harmonic Oscillators

In this section, we are going to study the entanglement of the general harmonic
oscillators, where both the centre of masses and the relative positions of the
two particles are both eigenstates of the harmonic oscillators. The Hamiltonian
is given by
H = HCM + Hrel (3.14)
where HCM is the Hamiltonian of the centre of mass and Hrel is that of the
relative position. Both Hamiltonians are that of harmonic osccillators. We
are only interested in the case that the centre of mass is at the ground state.
Therefore, in general, the state of the two particles is given by
· ¸ · ¸
2 m 1 x1 + m 2 x2 2 1 2
Ψ(x1 , x2 ) ∼ exp −b ( ) exp − (x1 − x2 ) Hn (x1 − x2 )
m1 + m2 2
(3.15)
where b is a parameter related to the standard deviation of the centre of mass,
and n is any non-negative integer indicating the state of the relative position.
Hn are the Hermite polynomials.
3.2.1 Decomposing the Ground State

If the relative position is at the ground state, the state can be decomposed
analytically. Its normalized wavefunction is given by
µ ¶ 12 " µ ¶2 #
2b m x
1 1 + m x
2 2 £ ¤
Ψ(x1 , x2 ) = exp −b2 exp − (x1 − x2 )2 (3.16)
π m1 + m2
The unit of length is taken to be the Bohr radius aB which has the value 0.53Å.
This state can be reduced to the two-mode squeezed state by setting m1 = m2 .
To find the Schmidt decomposition of (3.16), we must adopt the Mehler’s
Hermite Polynomial Formula [10] given in (2.20).
Let
(
x1 = c1 x01
x2 = c2 x02
and (3.16) becomes

µ
¶1 ½ · ¸ · ¸
2b 2 b2 m1 2 2 02 b2 m2 2 2
Ψ(x1 , x2 ) = exp − 1 + c 1 x1 − 1 + c2 2 x02
π (m1 + m2 )2 (m1 + m2 )2
· ¸ ¾
b2 m1 m2 0 0
+2 1 − c1 c2 x1 x2
(m1 + m2 )2
µ ¶ 12 µ 02 ¶ ½ · ¸
2b x1 x02 2 2 1 b2 m1 2 c1 2 2
= exp − − exp − c1 − + 2
x01
π 2 2 2 (m1 + m2 )
· 2 2 2
¸ · ¸ ¾
2 1 b m2 c2 02 b2 m1 m2 0 0
− c2 − + x +2 1− c 1 c 2 x1 x2
2 (m1 + m2 )2 2 (m1 + m2 )2
In order to make this to be favourable for doing Schmidt decomposition using

(2.20), we have to solve the following system of equations
 2
 w 2 1 b2 m1 2 c1 2
 2 = c1 − 2 + (m +m )2
 21−w 1 2
w 2 1 b2 m2 2 c2 2
= c2 − 2 + (m1 +m2 )2
 1−w2 h i

 w b2 m1 m2
2 = 1 −
1−w 2
(m1 +m2 )
c1 c2
for w, c1 and c2 . Assuming b is much less than 1, the system of equations gives
[4]
 √
 −b(m1 +m2 )2 + [(1+b2 )m1 2 +2m1 m2 +m2 2 ][m1 2 +2m1 m2 +(1+b2 )m2 2 ]

 w=

 r q m1 2 +(2−b2 )m1 m2 +m2 2

1 m1 2 +2m1 m2 +(1+b2 )m2 2
c1 = 2b (1+b2 )m1 2 +2m1 m2 +m2 2 (3.17)

 r q



 c2 = 2b 1 (1+b2 )m1 2 +2m1 m2 +m2 2
m1 2 +2m1 m2 +(1+b2 )m2 2
1
Note that c1 c2 = 2b . We can do the Schmidt decomposition of (3.16) using
(3.17) and (2.20):
µ ¶ 12 µ 02 ¶ · ¸
2b x1 x02 2 −w2 (x01 2 + x02 2 ) + 2wx01 x02
Ψ(x1 , x2 ) = exp − − exp
π 2 2 1 − w2
µ ¶ 12 µ 02 ¶ X∞
2b x1 x02 2 √ wn
= exp − − 1−w 2 Hn (x01 )Hn (x02 ) n
π 2 2 n=0
2 n!
(by (2.20))
We see that the Schmidt modes are eigenstates of harmonic oscillators. The
eigenstates of harmonic oscillators are given by [17]
r µ 2 2¶
a ax
un (x) = 1 Hn (ax) exp −
π 2 n!2n 2
Arranging coefficients give the Schmidt decomposition of (3.16) in the form
given by (2.2) with the Schmidt coefficients
λn = (1 − w2 )w2n (3.18)
and Schmidt modes of the two subsystems

µ ¶ µ ¶
1 x1 2 x1
fn (x1 ) = q exp − 2 Hn (3.19)
1 2c1 c1
π 2 n!2n c1
µ ¶ µ ¶
1 x2 2 x2
gn (x2 ) = q exp − 2 Hn (3.20)
1 2c2 c2
π 2 n!2n c2
where n are non-negative integer. We can easily check that

∞
X
λn = 1
n=0
Using the Schmidt coefficients found in (3.18), the participation ratio of

the different-mass oscillators state (3.16) is
1 + w2
K= (3.21)
1 − w2
3.2.2 Entanglement of the Ground State of Large Mass

Difference
There are often cases for m2 ¿ m1 , such as a hydrogen atom, which consists
of one proton and one electron, with the former much more massive than the
latter. Under this limit, the constants in (3.17) is given by
√
w ≈ −b
s + 1 + b2
1
c1 ≈ √
2b 1 + b2
s√
1 + b2
c2 ≈
2b
and the Schmidt coefficients become

√
λn ≈ 2b(−b + 1 + b2 )2n+1 (3.22)
Then by (3.21), the participation ratio is

√
1 + b2
K≈ (3.23)
b
The Schmidt coefficients λn for non-negative integral n is given graphically
in figure 3.10. [4] The variation of K against b for a hydrogen atom is given
graphically in figure 3.11. [4] As expected from (3.23), when b → ∞, K → 1,
and it is also shown in the figure 3.11. In other words, when the centre of mass
of the system has a very well-defined position (like at high temperature), then
the system is almost non-entangled.
Figure 3.10: Schmidt coefficients of a state containing two particles of different

masses. We have adopted m1 = 1840 (the mass of a proton) and m2 =1 (the mass
of an electron) (in ratio) when b = 0.256027.
Figure 3.11: Variation of the participation ratio K of the state of a two particles
of different masses against b. We have adopted m1 = 1840 (the mass of a proton)
and m2 =1 (the mass of an electron) (in ratio).
3.2.3 Entanglement of the Ground State of Equal Masses

There are often cases with a pair of particles of the same mass, i.e.,
m1 = m2 = m (3.24)
Then the state in (3.16) essentially becomes the two-mode squeezed state.
Then the constants in (3.17) becomes
2−b
w=
2+b
1
c1 = c2 = √ (3.25)
2
and the Schmidt coefficients become
µ ¶2n
2b 2−b
λn = (3.26)
2+b 2+b
Then by (3.21), the participation ratio is
4 + b2
K= (3.27)
4b
When b = 2, the states become separable; when b gets away from 2 further,
the participation ratio increases. That means the two particles of the same
mass connected with a spring is more entangled when either the translational
motion outweighs their oscillation, or vice versa.
3.2.4 Entanglement of the Excited States of Equal Masses

If the state is given by the Hamiltonian in (3.14), the relative position can take
any eigenstates of ordinary harmonic oscillators (so does the centre mass, but
we will not consider that). Then, if the masses of the two particles are equal,
such normalized wavefunction can be written as
Ã √ ! 12 · ¸
8b £ 2 2
¤ 1 2
Ψn (x1 , x2 ) = exp −b (x1 + x2 ) exp − (x1 − x2 ) Hn (x1 − x2 )
π2n n! 2
(3.28)
where Hn are Hermite polynomials for non-negative integers n, and b is the
parameter related to the uncertainty of the centre of mass of the composite
system. The case of n = 0 is essentially the two-mode squeezed state. We have
not done the Schmidt decomposition of (3.28) of the excited states analytically
but numerically. Denote the Schmidt decomposition of (3.28) by
Xp
Ψn (x1 , x2 ) = λj φj (x1 )ψj (x2 ) (3.29)
j
The Schmidt coefficients of (3.28) for b = 2 and n = 0, 1, 2, 3, 4, 5 is plotted

as shown in figure 3.12. Note that there is an interesting features concerning
to the oscillators with n being an odd integer (they are the states which are
antisymmetric, Ψn (x1 , x2 ) = −Ψn (x2 , x1 )): the Schmidt coefficients comes in
pair ! In other words, there are double degeneracy for the Schmidt coefficients.
Schmidt Coefficients of Compound Harmonic Oscillator
1
n
0.9 0
1
0.8 2
3
0.7
4
0.6 5
λj
0.5 b=2
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
j
Figure 3.12: Plot of the first 10 Schmidt coefficients of the composite harmonic
oscillators for b = 2 for n from 0 to 5
Schmidt modes of (3.28) for n = 1, 2 and 3 are also evaluated and plotted as
shown in figures 3.13, 3.14 and 3.15 respectively. Note that for those oscillators
with n being an odd integer (those with degenerate Schmidt coefficients), the
Schmidt modes of the degenerate states must have the relationships that:
(
φj (x) = ψj+1 (x)
(3.30)
φj+1 (x) = −ψj (x)
For those oscillators with n being even, the Schmidt modes of the two subsys-
tems are the same up to a sign.
Schmidt Modes of Compound Harmonic Oscillator b=2 n=1

1.5 1.5 1.5
0.5 0.5 0.5
A 0.5 0.5 0.5
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
1.5 1.5 1.5
0.5 0.5 0.5
B
0.5 0.5 0.5
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
x x x
j=1 j=2 j=3
1.5 1.5 1.5
0.5 0.5 0.5
A
0.5 0.5 0.5
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
1.5 1.5 1.5
0.5 0.5 0.5

B
0.5 0.5 0.5
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
x x x
j=4 j=5 j=6
Figure 3.13: Schmidt modes of the compound harmonic oscillator for b = 2 and
n=1

1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
A 0.5 0.5 0.5
1 1 1
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
x x x
j=1 j=2 j=3
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
A
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5

B
0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5

4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
x x x
j=4 j=5 j=6
n=2

2 2 2
1 1 1
0 0 0
A
1 1 1
2 2 2
4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
2 2 2
1 1 1
B 0 0 0
1 1 1
2 2 2
4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
x x x
j=1 j=2 j=3
2 2 2
1 1 1
0 0 0
A
1 1 1
2 2 2
4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
2 2 2
1 1 1
B
0 0 0
1 1 1
2 2 2
4 2 0 2 4 4 2 0 2 4 4 2 0 2 4
x x x
j=4 j=5 j=6
n=3
The participation ratio of the compound harmonic oscillator given by (3.28)

against the parameter b is plotted as shown in figure 3.16. It is seen that
in general, the more excited the state is, the more entangled the state is.
Moreover, the entanglements of the state are of the less degree in around b ≈ 1
(order) and they increase as they approaches further from this value of b.
Participation Ratio of Compound Harmonic Oscillator
20
15
n
4
K
10
3
5
2
1
0
0
0 0.5 1 1.5
b
Figure 3.16: Variation of the participation ratio of the compound harmonic oscil-
lator against b for n = 0, 1, 2, 3, 4, 5
Degeneracy in Antisymmetric Wavefunction
The special properties to the Schmidt coefficients and the Schmidt modes to
those odd antisymmetric excited states are actually the general properties of
the antisymmetric wavefunction. The wavefunction can be expressed in the
form Xp
Ψ(z1 , z2 ) = λj φj (z1 )ψj (z2 ) (3.31)
j
where {ψj (z)} and {φj (z)} are two orthonormal complete set of basis, i.e.,
Z
dz · φ∗j (z)φj 0 (z) = δjj 0
Z
dz · ψj∗ (z)ψj 0 (z) = δjj 0 (3.32)
And the following integral equations are also true:

Z
p
dz1 · Ψ(z1 , z2 )φ∗j (z1 ) = λj ψj (z2 ) (3.33)
Z p
dz2 · Ψ(z1 , z2 )ψj∗ (z2 ) = λj φj (z1 ) (3.34)
Antisymmetric wavefunctions are those wavefunctions that satisfy the fol-

lowing condition
Ψ(z1 , z2 ) = −Ψ(z2 , z1 ) (3.35)
Then given φj (z1 )ψj (z2 ) being a pair of Schmidt modes in the Schmidt
p
decomposition with eigenvalue given by λj , is −ψj (z1 )φj (z2 ) another pair of
Schmidt mode? We can evaluate
Z
dz1 · Ψ(z1 , z2 )ψj∗ (z1 )
Z
= − dz1 · Ψ(z2 , z1 )ψj∗ (z1 )
p
= − λj φj (z2 ) (3.36)
and similarly, Z
p
dz2 · Ψ(z1 , z2 )φ∗j (z2 ) = − λj ψj (z1 ) (3.37)
Therefore, −ψj (z1 )φj (z2 ) is a pair of Schmidt mode with the same eigenvalue
p
λj .
These Schmidt modes are of the same eigenvalues. Are they the same or
they are different? By (3.34) and (3.37), it is found that
Z Z
p ∗
p
− λj dz · ψj 0 (z)φj (z) = λj 0 dz · ψj∗ (z)φj 0 (z) (3.38)
and by (3.33) and (3.36), it is found that

Z Z
p p
− λj dz · φj 0 (z)ψj (z) = λj 0 dz · φ∗j (z)ψj 0 (z)
∗
(3.39)
[Note: in proving (3.38) and (3.39), we have applied:

XXp
Ψ∗ (z1 , z2 )Ψ(z1 , z20 ) = λk λl · φ∗k (z1 )φl (z1 )ψk∗ (z2 )φl (z20 )
k l
XXp
∗
Ψ (z1 , z2 )Ψ(z10 , z2 ) = λk λl · φ∗k (z1 )φl (z10 )ψk∗ (z2 )φl (z2 )
k l
which can be easily verified using the Schmidt expansion (3.31) ]

Putting j = j 0 in either (3.38) or (3.39), it is proved that
Z
dz · φ∗j (z)ψj (z) = 0 (3.40)
We can immediately see that φj (z1 )ψj (z2 ) and −ψj (z1 )φj (z2 ) are orthogonal,
i.e., they are different pairs of Schmidt modes, but they have the same eigen-
p
value λj . Hence there are degeneracy.
For those state given by (3.28) for odd n’s, it is antisymmetric as given by
(3.35). Then those states exhibit the degeneracy in Schmidt coefficients and
symmetric properties to the Schmidt modes as given by figure 3.13 and 3.15.
3.3 Entanglement Degeneracies in Symmetric

Wavefunctions
In the previous section, we know that the Schmidt decomposition of an anti-
symmetric wavefunction exhibits double degeneracy. In this section, we would
show that some of the symmetric wavefunctions, i.e., those wavefunctions sat-
isfying the following property:
Ψ(x1 , x2 ) = Ψ(x2 , x1 ) (3.41)
would exhibit this kind of degeneracy too. The symmetric compound harmonic
states, i.e., states in the form of (3.28) with even n’s, have not shown this kind
of degeneracy. However, in figure 3.12, it can be observed that the symmetric
states have some pairs of eigenvalues that are very close in magnitudes. In this
case, we would consider the following normalized wavefunction:
1 h i
−(x1 +d)2 −(x2 −d)2 −(x1 −d)2 −(x2 +d)2
Ψ(x1 , x2 ) = p e + e (3.42)
π(1 + e−4d2 )
where d is a parameter that describes how far the two Gaussian peaks are
from the origin (0, 0). More exactly, the Gaussian peaks in (3.42) are located
at (−d, d) and (d, −d). The wavefunction (3.42) for d = 4.0 and d = 1.0 are
plotted as shown in figure 3.17 and 3.18 [4] respectively. We can see that the
peaks of the former one is separated far apart, but that of the latter one is
very close. They exhibit different entanglement property.
Figure 3.17: Plot of the wavefunction (3.42) for d = 4.0
Figure 3.18: Plot of the wavefunction (3.42) for d = 1.0
The participation ratio of the state (3.42) is found by tracing the squared
reduced density operator of one of the subsystems, i.e.,
·Z ∞ ¸−1
(2) 2
K= dy · ρA (y, y) = (3.43)
−∞ 1 + cosh21(2d2 )
where the kernel of the reduced density operator of subsystem A and its square
is respectively given by
Z ∞
ρA (x1 , y1 ) = dx2 · Ψ(x1 , x2 )Ψ(y1 , x2 ) (3.44)
Z−∞
∞
(2)
ρA (x1 , y1 ) = dz · ρA (x1 , z)ρA (z, y) (3.45)
−∞
The variation of the participation ratio K in 3.43 with the parameter d is

plotted as shown in figure 3.19 [4]. It can be seen that when the Gaussian peaks
are very far apart, they are entangled, where K approaches the asymptotic
limit 2. This limit value implies there are special values taken by the Schmidt
coefficients. When the peaks become closer together, i.e., d approaches 0, they
tend to become separable, i.e., K approaches the values 1.
Figure 3.19: Variation of the participation ratio of the wavefunction (3.42) as the
parameter d
Numerical Schmidt decomposition has been done to the state (3.42) for
several values of the parameter d. In figure 3.20, the first two Schmidt coeffi-
cients of the state (3.42) and the corresponding modes of d = 0.5 and d = 4.0
are shown. For large d, we see that the two Schmidt coefficients are nearly
coincident, i.e., double degenerate. The Schmidt modes are identical up to a
sign. It is corresponding to the participation ratio given by the limiting value
2 in figure 3.19. As d approaches smaller values, the two Schmidt coefficients
detach from the degenerate value 0.5 and become non-degenerate. As the two
values become unequal, the amount of entanglement decreases according to

figure 3.19. The participation ratio approaches 1, i.e., non-degenerate, and it
is also evident from figure 3.20 that one of the Schmidt coefficients approach
one while another approach zero. It is just equivalent to becoming a separable
state.
Schmidt Decomposition of the Wavefunction
[ 2 2
] [
Ψ (x1 , x 2 ) ~ exp − (x1 + d ) − (x 2 − d ) + exp − (x1 − d ) − (x 2 + d )
2 2
]
Schmidt Coefficients vs Position of Peaks
λ1
0.9
λ2
0.8
0.7
Schmidt Coeffients
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Schmidt modes Schmidt modes d

Schmidt modes Schmidt modes
1 1 1 1
0.5 0.5 0.5 0.5
0 0 0 0
0.5 0.5 0.5 0.5
x1 x1 x1 x1
1 1 1 1
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
1 1 1 1
0.5 0.5 0.5 0.5
0 0 0 0
0.5 0.5 0.5 0.5
x2 x2 x2 x2
1 1 1 1
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
Figure 3.20: The first two Schmidt coefficients of the state (3.42) and some of the
corresponding Schmidt modes
We can generate n-fold degeneracy wavefunction artificially. Figure 3.21

illustrates a 6-fold degenerate wavefunction. This wavefunction is composed

of six hemispheres with unit radius centred at the marked coordinates. The
spheres are separated remarkably. Doing the Schmidt decomposition numeri-
cally will give six-fold degenerate Schmidt coefficients as shown in figure 3.22,
with the first 12 Schmidt modes plotted as shown in figures 3.23 and 3.24.
x2
(-6,6)
(-4,4)
(-2,2)
x1
(2,-2)
(4,-4)
(6,-6)
Figure 3.21: The wavefunction composing of six hemispheres with unit radius cen-
tred at the marked coordinates.
Schmidt Coefficients
0.9
0.8
0.7
0.6
λi
0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10 11 12
i
Figure 3.22: First 12 Schmidt coefficients of the six-hemispherical (and six-fold

degenerate) wavefunction
Schmidt Modes
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
A 0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
x x x
j=1 j=2 j=3
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
A 0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
2 2 2
1.5 1.5 1.5
1 1 1
B 0.5 0.5 0.5

0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
x x x
j=4 j=5 j=6
Figure 3.23: The 1st to 6th Schmidt modes of the six-hemispherical wavefunction
Schmidt Modes
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
0.5 0.5 0.5
A
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
B 0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
x x x
j=7 j=8 j=9
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
0 0 0
A0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
2 2 2
1.5 1.5 1.5
1 1 1
0.5 0.5 0.5
B
0 0 0
0.5 0.5 0.5
1 1 1
1.5 1.5 1.5
2 2 2
10 5 0 5 10 10 5 0 5 10 10 5 0 5 10
x x x
j=10 j=11 j=12
Figure 3.24: The 7th to 12th Schmidt modes of the six-hemispherical wavefunction
3.4 Asymptotic Behaviour of Entanglement in

Oscillating Systems of Equal Masses
In the previous calculations, it is seen that in composite oscillating systems,
if the masses of the two particles are equal, then the entanglement in general
increases for extremely large and small b’s. (See figures 3.9 and 3.16) We are
going to explore this behaviour further.
For the sake of our analysis, we assume the wavefunction of the oscillating
systems have the following form:
µ ¶ 41
β2
Ψ(x1 , x2 ) = exp[−β 2 (x1 + x2 )2 ] · ψ[α(x1 − x2 )] (3.46)
2π
where ψ is the eigenfunction of the interaction between the two particles. It

should be a normalized bound state. α and β are related to the variances of
the centre of mass and the relative coordinates. In fact, we can transform the
coordinates so that b = αβ , restoring the previous forms like (3.11) and (3.28)
with the parameter b. Large b’s refer to the case β À α while small b’s refer
to α À β.
Let’s look at b → ∞, or its equivalent β À α. We do the following
transformation:
(
x˜1 = αx1
x˜2 = αx2
Then the wavefunction in (3.46) restore the previous forms

µ ¶ 14
b2
Ψ(x˜1 , x˜2 ) = exp[−b2 (x˜1 + x˜2 )2 ] · ψ(x˜1 − x˜2 ) (3.47)
2π
When b → ∞, the interaction ψ becomes less significant, and the exponential

terms become more dominant. The wavefunction behaves like the Dirac delta
function, given by
Ψ(x1 , x2 ) ∼ δ(x˜1 + x˜2 ) ∼ δ(x1 + x2 ) (3.48)
because the Dirac delta function can be expressed as [18]

r
α2
δ(x) = lim exp(−α2 x2 ) (3.49)
α→∞ π
In bra-ket form, (3.48) can be expressed as

Z ∞ Z ∞
|Ψi ∼ dx1 dx2 · δ(x1 + x2 )|x1 i|x2 i
−∞ −∞
Z ∞
= dx · |xi| − xi
−∞
with equal values of Schmidt coefficients. It implies the maximal entanglement.

A Dirac delta function exhibit great non-local correlation between x1 and x2 .
It looks like the illustrative examples for the original EPR state. [1] Thus, for
b → ∞, the entanglement increases to a very large value, as shown in 3.9 and
3.16.
Now, let’s consider b → 0, or its equivalent α À β. We do the following
transformation
(
x˜1 = βx1
x˜2 = βx2
Then the wavefunction is transformed to

µ 2 ¶ 14 · ¸
b 2 1
Ψ(x˜1 , x˜2 ) = exp[−(x˜1 + x˜2 ) ] · ψ (x˜1 − x˜2 ) (3.50)
2π b
For b → 0, 1b → ∞. The interaction factor ψ becomes more important.

Situation is much more complicated here because we just know that ψ is a
normalized bound state. However, we expect it bears the Gaussian shape, or
if it is oscillating, it bears the Gaussian envelop. It is in general quite true for
the states we have used, as shown in figures 2.2 and 3.2. Therefore, similar to
the case b → ∞, now as b → 0,
Ψ(x1 , x2 ) ∼ δ(x˜1 − x˜2 ) ∼ δ(x1 − x2 ) (3.51)
and again
Z ∞ Z ∞
|Ψi ∼ dx1 dx2 · δ(x1 − x2 )|x1 i|x2 i
Z−∞
∞
−∞
= dx · |xi|xi
−∞
which is another kind of non-local correlation of x1 and x2 . Thus, for b → 0,

the entanglement increases to a very large value, as shown in 3.9 and 3.16.
Chapter 4
Entanglement in
Three-Dimensional Systems
In chapter 3, we discussed the entanglement in one-dimensional composite
systems. In this chapter, we are going to discuss about the entanglement
of three-dimensional systems. We are going to discuss the entanglement of
a three-dimensional two-mode squeezed state and the positronium (a system
consisting of an electron and a positron orbiting about each other).
4.1 Three-Dimensional Two-mode Squeezed State

We know that the one-dimensional two-mode squeezed state given by (1.6) have
the Schmidt decomposition given by (2.23) and (2.24). Now, if we consider a
three-dimensional version of (1.6), i.e.,
8 (r1 + r2 )2
Ψ(r1 , r2 ) = 3 exp[− ] exp[−(r1 − r2 )2 ] (4.1)
(8πσ 2 ) 2 8σ 2
Obviously, its Schmidt decomposition can be found by separating it into three

components, since each of its component is a two-mode squeezed state:
Ψ(r1 , r2 ) = Ψx (x1 , x2 )Ψy (y1 , y2 )Ψz (z1 , z2 )

X∞ X ∞ X ∞
p
= λm λn λp [fm (x1 )fn (y1 )fp (z1 )][fm (x2 )fn (y2 )fp (z2 )]
m=0 n=0 p=0
where the λ’s and f (x) are given by (2.23) and (2.24) respectively. The Schmidt
coefficients are then given by
Λmnp = λm λn λp (4.2)
63
Chapter 4 Entanglement in Three-Dimensional Systems 64
The participation ratio of (4.1) is then given by

1 X
= Λmnp 2
K mnp
∞
X ∞
X ∞
X
2 2
= λm · λn · λp 2
m=0 n=0 p=0
· µ ¶¸−3
1 2 1
= 8σ + 2
2 8σ
where (2.36) has been applied, taking b2 = 8σ1 2 . Then
µ ¶3
1 2 1
K= 8σ + 2 (4.3)
8 8σ
The participation ratio in (4.3) is plotted as shown in the figure 4.1. [4]
Participation Ratio of Three-Dimensional

K Two-Mode Squeezed State against σ
100
80
60
40
20
s
0.5 1 1.5 2 2.5 3
Figure 4.1: Plot of participation ratio of the three-dimensional two-mode squeezed
state against the parameter σ
4.2 Three-Dimensional Positronium System

A positronium comprises of an electron and positron orbiting about each other.
The interaction between them is just Coulomb potential. This implies that the
relative coordinates have the eigenstates given by those similar to the hydrogen
atom. Suppose the whole system is trapped in a harmonic potential, and the
interaction between the electron and the positron is at the ground state. Then
the state is given by
µ ¶ µ ¶
1 (r1 + r2 )2 |r1 − r2 |
Ψ(r1 , r2 ) = 5 3 3 exp − exp − (4.4)
π 4 σ 2 a0 2 8σ 2 a0
where a0 is the “Bohr radius” of the positronium, given by
8π²0 ~2
a0 = 2 (4.5)
e me
where e and me are the electronic charge and mass respectively. Note that the
electron and the positron have the same mass. σ is the standard deviation of
the position of the centre of mass. Because of its simplicity, we can express the
state (4.4) in terms of the radial distance of the two particles from the origin
(equilibrium point of the harmonic potential) and the included angle between
the two radial vectors of the two particles from the origin, denoted by r1 , r2
and γ respectively. Then the state (4.4) can be rewritten as
µ ¶
1 r1 2 + r2 2 + 2r1 r2 cos γ
Ψ(r1 , r2 , cos γ) = 5 3 3 exp −
π 4 σ 2 a0 2 8σ 2
Ã p !
r1 2 + r2 2 − 2r1 r2 cos γ
· exp − (4.6)
a0
In the following analysis, the unit of length is taken to be the “Bohr radius”
of the positronium, i.e., a0 = 1.
To study the entanglement of the state given by (4.6), we must find its
Schmidt decomposition. To do the Schmidt decomposition, we are eager to
see the contribution of the angular momenta to the two-particle state (4.6),
and therefore we should find the Legendre representation of (4.6) [19] first:
∞
X
Ψ(r1 , r2 , cos γ) = αl (r1 , r2 )Pl (cos γ) (4.7)
l=0
where Z π
2l + 1
αl (r1 , r2 ) = dγ · sin γ · Ψ(r1 , r2 , cos γ)Pl (cos γ) (4.8)
2 0
We can find the expansion of αl (r1 , r2 ), given by the form
Xp
αl (r1 , r2 ) = λnl φnl (r1 )ψnl (r2 ) (4.9)
n
where the expansion eigenfunctions must satisfy the following orthonomality

condition: ( R∞
dr · r 2 φ (r )φ 0 (r ) = δnn0
R0∞ 1 1 2 nl 1 n l 1 (4.10)
0
dr2 · r 2 ψ nl (r2 )ψ n 0 l (r2 ) = δnn0
It is not easy to directly find the functions satisfying the orthogonality con-
dition (4.10). To do this expansion, we should find the Schmidt decomposition
of r1 r2 αl (r1 , r2 ) first and we can get the decomposition in the form of
Xp
r1 r2 αl (r1 , r2 ) = λnl [φñl (r1 )][ψñl (r2 )]
n
By usual Schmidt decomposition, the eigenfunctions above satisfy the orthono-

mality condition without any weighting function:
( R∞
dr · φ̃ (r )φ̃ 0 (r ) = δnn0
R0∞ 1 nl 1 n l 1
0
dr2 · ψ̃nl (r2 )ψ̃n0 l (r2 ) = δnn0
We can get back the desired Schmidt modes in (4.9) by dividing φñl (r1 ) and
ψñl (r2 ) by r1 and r2 respectively:
(
φnl (r1 ) = φ̃nlr(r
1
1)
ψ̃nl (r2 )
ψnl (r2 ) = r2
Finally, we employ the spherical harmonic addition theorem [20]:

l
X 4π ∗
Pl (cos γ) = Ylm (θ1 , φ1 )Ylm (θ2 , φ2 ) (4.11)
m=−l
2l + 1
Obviously, the two particles are pointing in the opposite directions, implying
no net angular momentum of the whole system. Then the expression in (4.7)
can be further expressed as
X∞ X ∞ X l µp ¶
4π
Ψ(r1 , r2 , cos γ) = λnl
n=1 l=0 m=−l
2l + 1
∗
· [φnl (r1 )Ylm (θ1 , φ1 )] [ψnl (r2 )Ylm (θ2 , φ2 )]
Or we can say that (4.6) can be expressed in the form of Schmidt decomposi-
tion:
∞ X
X ∞ X
l p
Ψ(r1 , r2 , cos γ) = Λnlm [φnl (r1 )Yl−m (θ1 , φ1 )] [ψnl (r2 )Ylm (θ2 , φ2 )]
n=1 l=0 m=−l
(4.12)
where the Schmidt coefficients Λnlm are given as

µ ¶2
4π
Λnlm = λnl (4.13)
2l + 1
∗
and we have used Ylm = Yl−m . It is expected that the Schmidt coefficients
should be normalized:
X∞ X ∞ X l
Λnlm = 1 (4.14)
n=1 l=0 m=−l
The participation ratio should be given by
X∞ X ∞ X l
1
= Λnlm 2 (4.15)
K n=1 l=0 m=−l
From the final Schmidt form of the ground state positronium (4.12), we
see that the positron and the electron must be in the same energy eigenstate
(same n), having the same total angular momenta (same l) but in opposite
direction (m’s having opposite sign).
To study the entanglement of (4.6), we have to find the Schmidt coefficients
(4.13) and the participation ratio (4.15). Besides, we are also interested to find
the Schmidt modes in the Schmidt decomposition (4.12). All of these have to
be done numerically. In the numerical computation, it is very computationally
expensive to consider all possible Schmidt coefficients. In our calculation, for
each value of the parameter σ, we considered l from 0 to 10 and for each l
we considered n from 1 to 10. To see whether it is a good approximation,
we can check all the Schmidt coefficients considered satisfy the normalization
condition (4.14). Actually, such approximation is quite good for large σ, as
shown in the table 4.1 [4]. For σ = 0.25, the approximation is not quite good
because the wavefunction is highly peaked and it seems more values of l are
needed to add up to make this highly peaked functions.
P
σ (unit: a0 ) Computed nlm Λnlm Percentage Error
0.25 0.799435 20.1%
0.5 0.989149 1.09%
0.75 0.999465 0.0535%
1.0 0.999931 0.00690%
1.25 0.999888 0.0112%
1.5 0.999740 0.0260%
1.75 0.999468 0.0532%
2.0 0.999025 0.0975%
2.25 0.998355 0.165%
2.5 0.997405 0.260%
Table 4.1: Sum of Schmidt coefficients of (4.12) in numerical calculation
Using the approximation described, we can find approximately how the

participation ratio changes with the parameter σ, as shown in figure 4.2. It
has the shape of the participation ratio of the three-dimensional two-mode
squeezed state (4.1) in figure 4.1, but the order of magnitudes between them
differ a lot. Like the two-mode squeezed state, there probably exists a value
of σ that the participation ratio equals 1, which corresponds to a separable
(non-entangled) state. Going away from this value, the state becomes more
and more entangled.
Participation Ratio of Positronium against σ

8
K
4
0
0 1 2 3
σ (Unit: a0)
Figure 4.2: Plot of participation ratio of the trapped positronium against the pa-
rameter σ
The Schmidt coefficients Λnlm against n and l for σ = 0.5 and σ = 2.5 are
plotted as shown in figures 4.3 and 4.4 respectively. Note that Λnlm does not
depend on m. At σ = 0.5, it is just close to the non-entangled state: only one
of the Λnlm is dominant, and it is the state of zero total angular momentum
(l = 0). At σ = 2.5, it is more like an entangled state, but the zero angular
momenta states are still more dominant, while some others corresponding to
l = 1, 2 are not unimportant.
n 3
4
2
1
0
1
0.75
Λnlm 0.5
0.25
0
0
1
2
3
l
Figure 4.3: Plot of Λnlm against n and l for σ = 0.5
n 3
4
2
1
0
1
0.75
Λnlm 0.5
0.25
0
0
1
2
3
l
Figure 4.4: Plot of Λnlm against n and l for σ = 2.5
The Schmidt modes of (4.6) with σ = 2.0 for l = 0, 1 and 2 are plotted
as shown in figures 4.5, 4.6 and 4.7. The modes of the two subsystems are
expected to be identical.
Schmidt Modes of Positronium (l=0)

n=1 n=2 n=3 n=4 n=5
3 3 3 3 3
2 2 2 2 2
1 1 1 1 1
0 0 0 0 0
x1 x1 x1 x1 x1
1 1 1 1 1
0 4 8 0 4 8 0 4 8 0 4 8 0 4 8
n=6 n=7 n=8 n=9 n=10
3 3 3 3 3
2 2 2 2 2
1 1 1 1 1
0 0 0 0 0
x1 x1 x1 x1 x1
1 1 1 1 1
0 4 8 0 4 8 0 4 8 0 4 8 0 4 8
Figure 4.5: Schmidt modes of (4.6) for l = 0 and σ = 2.0


2
n=1 2
n=2 2
n=3 2
n=4 2
n=5
1.5 1.5 1.5 1.5 1.5
1 1 1 1 1
0.5 0.5 0.5 0.5 0.5
0 0 0 0 0
0.5 0.5 0.5 0.5 0.5
1
x1 1
x1 1
x1 1
x1 1 x1
0 4 8 0 4 8 0 4 8 0 4 8 0 4 8
n=6 n=7 n=8 n=9 n=10
2 2 2 2 2
1.5 1.5 1.5 1.5 1.5
1 1 1 1 1
0.5 0.5 0.5 0.5 0.5
0 0 0 0 0
0.5 0.5 0.5 0.5 0.5
1 x1 1 x1 1 x1 1 x1 1
x1
0 4 8 0 4 8 0 4 8 0 4 8 0 4 8

1.5
n=1 1.5
n=2 1.5
n=3 1.5
n=4 1.5
n=5
1 1 1 1 1
0.5 0.5 0.5 0.5 0.5
0 0 0 0 0
0.5 0.5 0.5 0.5 0.5
x1 x1 x1 x1 x1
1 1 1 1 1
0 4 8 0 4 8 0 4 8 0 4 8 0 4 8
n=6 n=7 n=8 n=9 n=10
1 1 1 1 1
0 0 0 0 0
x1 x1 x1 x1 x1
1 1 1 1 1
0 4 8 0 4 8 0 4 8 0 4 8 0 4 8

Bibliography
[1] A. Einstein, B. Podolsky, and N. Rosen, Phys. Rev. 47, 777 (1935).
[2] D. Bohm, Quantum Theory, Prentice Hall, Englewood Cliffs, NJ.
[3] O. Cohen, Phys. Rev. A 56, 3484 (1997).
[4] Mathematica 5.0, Wolfram Research.
[5] A. Ekert and P. L. Knight, Am. J. Phys. 63, 5 (1995).
[6] J. Preskill, Lecture Notes for Physics 229: Quantum Information and
Computation, California Institute of Technology, CA.
[7] C. H. Bennett, S. Popescu, D. Rohrlich, J. A. Smolin, and A. V. Thapliyal,

Phys. Rev. A 63, 012307 (2000).
[8] J. Eisert and M. B. Pleino, quant-ph , 0312071 (2003).
[9] S. Parker, S. Bose, and M. B. Plenio, Phys. Rev. A 61, 032305 (2000).
[10] E. W. Weisstein, Mehler’s Hermite Polynomial Formula, MathWorld–A

Wolfram Web Resource.
[11] L. M. Duan, G. Giedke, J. I. Cirac, and P. Zoller, Phys. Rev. Lett. 84,
2722 (2000).
[12] A. Wehrl, Rev. Mod. Phys. 50, 221 (1978).
[13] F. Reif, Fundamentals of Statistical and Thermal Physics, McGraw Hill.
[14] C. H. Bennett, H. J. Bernstein, S. Popescu, and B. Schumacher, Phys.

Rev. A 53, 2046 (1996).
[15] P. M. Morse, Phys. Rev. 34, 57 (1929).
73
View publication stats
[16] A. Frank, A. L. Rivera, and K. B. Wolf, Phys. Rev. A 61, 054102 (2000).
[17] E. W. Weisstein, Hermite Polynomials, MathWorld–A Wolfram Web

Resource.
[18] G. B. Arfken and H. J. Weber, Mathematical Methods for Physicists,

Harcourt, 2001.
[19] E. W. Weisstein, Legendre Polynomials, MathWorld–A Wolfram Web

Resource.
[20] E. W. Weisstein, Spherical Harmonic Addition Theorem, MathWorld–A

Wolfram Web Resource.
74

Quantum Entanglement in Continuous System

Uploaded by

Copyright:

Available Formats

Quantum Entanglement in Continuous System

Uploaded by

Document Information

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

Quantum Entanglement in Continuous System

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Quantum Entanglement in Continuous System

Thesis · May 2004

Kwan Yuet Stephen Ho

Large Scale Text Clustering View project

SOCcer View project

The user has requested enhancement of the downloaded file.

Supervisor: Dr. Law Chi Kwong

A Final Year Project (PHY4610/PHY4620) Report

May 24, 2004

The concept of quantum entanglement arises since A. Einstein, B. Podolsky

2 Schmidt Decomposition and the Entangled States 8

3 Entanglement in Various Oscillating Systems 30

4 Entanglement in Three-Dimensional Systems 63

1.1 Existence of the element of physical reality corresponding to Sz in

2.1 Schmidt coefficients of the two-mode squeezed state with b = 0.256027 16

3.1 Plot of the Morse potential (rescaled) for ² = 24.01 . . . . . . . . . 31

4.1 Plot of participation ratio of the three-dimensional two-mode squeezed

4.1 Sum of Schmidt coefficients of (4.12) in numerical calculation . . . . 68

1.1 Principle of Locality and Physical Reality

1. Principle of Locality There is no action at a distance, i.e., the measure-

2. Physical Reality “If, without in any way disturbing a system, we can

We consider two observables A and B that do not commute, i.e.,

Examples of a pair of non-commutative observables are the position and the

1.2 An Illustration using Bohm’s Spin-Singlet

Figure 1.1: Existence of the element of physical reality corresponding to Sz in

However, we know that the state in (1.2) can also be expressed as

Figure 1.2: Existence of the element of physical reality corresponding to Sx in

In the above two measurements, we conclude that in subsystem B, there

exists simultaneous reality corresponding to the two non-commutative observ-

1.2.2 Mathematical Form of the Paradox

1.2.3 Resolving the EPR Paradox by Inspecting the Un-

1.3 An Illustration using Two-Mode Squeezed

Figure 1.3: 3D Plot of the two-mode squeezed state with b = 0.3333

The variance of the position of this state can be calculated:

observables in the formalism of quantum mechanics. This again illustrates the

but by the uncertainty principle, for two non-commutative observables,

Schmidt Decomposition and the

2.1 Schmidt Decomposition of a Composite State

where {|ui i, i = 1, 2, 3, . . .} and {|vi i, i = 1, 2, 3, . . .} are the orthonormal basis

2.1.1 Proof of the Existence of the Schmidt Decompo-

1, 2, 3, . . .}. Then any state |Ψi can be written as

and the corresponding density operator of this pure state is given by

The reduced density operator for subsystem A is given by

= |gi |2 δik (2.4)

Obviously, we can construct a set of orthonormal bases in subsystem B

for each |gi | 6= 0. It is evident that |bi i, i = 1, 2, 3, . . . form a set of orthonormal

2.1.2 Schmidt Decomposition in Continuous Systems

The reduced density operator of the subsystem A is

In the proof given by the previous section, we have assumed ρA is diago-

ρA |φi iA = λi |φi iA (2.9)

equations (2.10). As long as the kernel ρA (x, x0 ) in the integral equation is

converges, there exist an orthonormal set of eigenstates such that

[9] Then the kernel in the integral equation can be expressed as

and the reduced density operator of A is

which is diagonalized. The diagonalizability of ρA means the existence of the

2.1.3 Algorithm of Finding the Schmidt Decomposition

where δ(x − x0 ) is the Dirac delta function. It is analogous to the discrete

hψi |ψj i = δij

where δij is the Kronecker delta symbol. Therefore, it is possible to discretize