This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:
IP Address: 36.88.252.88
This content was downloaded on 15/03/2014 at 10:02
Please note that terms and conditions apply.
Is quantum theory intrinsically nonlinear?
View the table of contents for this issue, or go to the journal homepage for more
2013 Phys. Scr. 87 038117
(http://iopscience.iop.org/1402-4896/87/3/038117)
Home Search Collections Journals About Contact us My IOPscience
IOP PUBLISHING PHYSICA SCRIPTA
Phys. Scr. 87 (2013) 038117 (10pp) doi:10.1088/0031-8949/87/03/038117
Is quantum theory intrinsically
nonlinear?
Dieter Schuch
Institut f ur Theoretische Physik, J W Goethe-Universit at Frankfurt am Main, Max-von-Laue-Strae 1,
D-60438 Frankfurt am Main, Germany
E-mail: Schuch@em.uni-frankfurt.de
Received 2 November 2012
Accepted for publication 4 December 2012
Published 11 February 2013
Online at stacks.iop.org/PhysScr/87/038117
Abstract
In contrast with classical physics, complex quantities have a fundamental physical meaning in
quantum physics and action, being essentially the quantized entity, should be given more
attention instead of focusing mainly on Hamiltonians or Lagrangians that have the dimension
of energy. Phase and amplitude of the complex quantities in (time-dependent and
time-independent) quantum mechanics are not independent of each other but coupled via some
conservation law. This coupling can be understood if the systems are described in terms of
complex nonlinear Riccati equations. These equations not only enable a connection to the
Pythagorean triples, probably the oldest and most abstract quantization problem, but also
lead to dynamical invariants with the dimension of action. Factorization of the corresponding
operator provides generalized creation and annihilation operators, which is also possible for
dissipative systems where no conventional Hamiltonian formalism exists. Formal similarities
with other elds, particularly with nonlinear dynamics, are shown.
PACS numbers: 03.65.Ta, 03.65.w, 03.65.Yz
1. Introduction
On 4 July 2012, CERN announced the discovery of a new
elementary particle that is probably the Higgs boson which
is needed to complete the standard model of elementary
particles. If it turns out to be true, is this the completion
of theoretical physics? Denitely not! Not only is the
unication of the electromagnetic, weak and strong forces
with gravity still an unsolved puzzle, there are even more
serious shortcomings of the present state of theoretical physics
as, e.g. stated by Rothman in a recent paper [1]. He claims
that the building of theoretical physics resembles a dilapidated
Tower of Babel with many cracks. He correctly points out
that in the so-called fundamental equations of motion there is
no distinction between past and future, no direction of time
and phenomena like friction producing heat and increasing
entropy in irreversible processes do not occur. Besides, he
also criticizes the large number of adjustable parameters and
thus the lack of beauty in the standard model and the
fact that even rather simple systems like a double-pendulum
obey nonlinear (NL) equations of motion that may lead to
chaotic time-evolution. Also many other phenomena in our
surrounding world (like the weather) obey NL evolution
equations. On the other hand, the most fundamental and
successful theory, quantum theory, is supposedly a linear
theory. Is it possible to unite the missing aspects of non-
linearity, irreversibility and dissipation with the existing
fundamental theories, particularly with quantum mechanics?
In this paper, it shall be shown that a positive answer can
be given to this question if certain prejudices are abandoned.
This will lead to a NL formulation of quantum mechanics
that still allows for a superposition principle and is based
on a complex, quadratically-nonlinear Riccati equation and
the replacement of energy by action as the fundamental
quantity in physics. This formulation is advantageous because
it is also to be found in the same mathematical form
in other elds of physics, like optics, quantum optics,
BoseEinstein condensates (BEC), supersymmetric (SUSY)
quantum mechanics, NL dynamics, cosmology and others.
2. The roots of quantization
Considering quantum mechanics, properties come to mind
like linear wave equation (and thus superposition principle),
wave-particle duality, uncertainty principle, quantized energy
levels etc. But what are really the essential differences
between classical and quantum physics?
0031-8949/13/038117+10$33.00 Printed in the UK & the USA 1 2013 The Royal Swedish Academy of Sciences
Phys. Scr. 87 (2013) 038117 D Schuch
The era of quantum physics began around 1900 [2] with
Max Plancks introduction of h, the quantum of action, that
he needed to quantize the energy of the harmonic oscillator
(HO). So the actual quantized quantity is action, not energy;
the quantized energy of the HO is only a consequence of the
constant frequency
0
of the oscillator; for a time-dependent
(TD) frequency (t ) it looks quite different, as will be
discussed later.
In classical as well as in quantum mechanics, the
Hamiltonian, representing the energy of the system and
containing the potential from which the force can be derived,
plays the dominant role. For the HO with TD frequency, the
Hamiltonian is no longer a constant of motion, but an invariant
still exists with the dimension of an action in this case. In the
Riccati equations that occur in SUSY quantum mechanics and
in our NL version of quantum mechanics, the potential plays
only the role of an inhomogeneity that essentially enters in a
particular solution of the corresponding differential equation.
Another fundamental difference between classical and
quantum physics, as stated by Yang in his talk square root
of minus one, complex phase and Erwin Schr odinger [3], is
that with quantum mechanics, for the rst time, the imaginary
unit enters physics in a fundamental way and complex
numbers became a conceptual element of the very foundations
of physics. He continues that the very meaning of the
fundamental equations of matrix mechanics (Heisenbergs
commutation relation) and of wave mechanics (the TD
Schr odinger equation (SE)) would be totally destroyed if one
tries to get rid of i by writing them in terms of real and
imaginary parts.
I totally agree with this statement but want to go a step
further. The reason it is not sufcient to simply write these
complex equations in terms of two real ones is due to the
fact that real and imaginary parts, or phase and amplitude,
respectively, are not independent of each other but coupled
via some kind of conservation law, as will be shown below.
Where does the idea of quantization actually originate?
And what is it that should be quantized? In the western culture
one might rst think of the Greek philosopher Democritus
(about 460about 371 BC) and his idea of dividing our world
into minute components that are not further divisible, leading
to the term atom that is still used for the building blocks
of our physical world which, for some time, seemed further
indivisible.
A similar idea was formulated in a more abstract way
by another equally famous Greek philosopher living around
the same time. In his work Timaios, Plato (428/27348/47
BC) gives his view of how the world is built up in terms
of right-angled triangles, essentially his quanta. Werner
Heisenberg, equally fascinated with, and puzzled by, this text,
summarizes this idea [4] in the general sense as follows:
Matter is made up of right-angled triangles which, after
being paired to form isosceles triangles or squares, are joined
together to build the regular bodies of stereometry: cube,
tetrahedron, octahedron and icosahedron. These four solids
are then supposed to be the basic units of the four elements
earth, re, air and water.
So, geometric objects like right-angled triangles should
be the quanta of nature. But one could go even further and
ask if particular right-angled triangles might play a special
role in quantization. Yet another Greek philosopher enters the
scene. We remember Pythagoras theorem from school, i.e.
a
2
+b
2
= c
2
where a and b are the two shorter sides (catheti)
and c is the longest side (hypotenuse) of a right-angled
triangle. Pythagoras (around 570500 BC) and his disciples
were well-known for their dogma everything is number,
with number meaning integer. So let us look for right-angled
triangles where the lengths of all three sides are integers
(a kind of second quantization) fullling the Pythagorean
theorem. The most common example for such a Pythagorean
triple is (3, 4, 5) with 9 +16 = 25. But, asked for a few more
examples of the kind, even mathematically-afliated persons
have difculties nding anyalthough an innite number of
triples exists! Moreover, there is even a rather simple rule of
nding these triples. This rule was probably already known
to the Babylonians more than 3500 years ago [5] but, at
least, Diophantus of Alexandria (around 250 AD) knew of it.
Why do I mention this here? What does this have to do with
quantum mechanics? In the following it will be shown that
from the dynamics of Gaussian wave packet (WP) solutions
of the TDSE a complex NL evolution equation (Riccati
equation) can be obtained that also contains the key for the
answer to the above question of obtaining the Pythagorean
triple. This complex Riccati equation will lead to a dynamical
invariant that has the dimension of an action and still exists
for cases where the Hamiltonian is no longer invariant; e.g.
for the HO with TD frequency or certain dissipative systems.
This will also allow for the construction of generalized
creation/annihilation operators and corresponding coherent
states (CS).
The same type of NL Riccati equations also occur in
time-independent (TI) SE, as well as in BECs, cosmology and
other elds mentioned before. More details will be given in
sections 4 and 6.
In section 3, the WP solutions of the TDSE and the
corresponding complex Riccati equation will be discussed.
A generalization to include irreversibility and dissipation
leading to new qualitative properties like bifurcations will
be outlined in section 5 and a comparison with systems
showing relaxation, scale-invariance or both, as they occur
in NL dynamics, will be given in section 6. A summary and
perspectives conclude the paper.
3. Time-dependent Schr odinger equation (SE),
complex Riccati equation and dynamical invariant
In the following, one-dimensional problems with exact
analytic solutions of the TDSE in the form of Gaussian WPs
will be considered, particularly the free motion (potential
V(x) = 0) and the HO (V =
m
2

2
x
2
) with constant frequency,
=
0
, or TD frequency, = (t ). In these cases, the
solution of the TDSE (here for the HO, the case V = 0, in
the following, is always obtained in the limit 0)
i

h

t
(x, t ) =
_

h
2
2m

2
x
2
+
m
2

2
x
2
_
(x, t ), (1)
(where

h =
h
2
) can be written as
(x, t ) = N(t ) exp
_
i
_
y(t ) x
2
+
p

h
x + K(t )
__
(2)
2
Phys. Scr. 87 (2013) 038117 D Schuch
with the shifted coordinate x = x x = x (t ), where
the mean value x =
_
+

x dx = (t ) corresponds to
the classical trajectory, p = m represents the classical
momentum and the coefcient of the quadratic term in the
exponent, y(t ) = y
R
(t ) +iy
I
(t ), is a complex function of time.
The (possibly TD) normalization factor N(t ) and the purely
TD function K(t ) in the exponent are not relevant to the
following discussion in this section.
The equations of motion for (t ) and y(t ), or
_
2 h
m
y
_
,
that are obtained by inserting WP (2) into equation (1), are
important for our purpose and given by
+
2
= 0, (3)
and
_
2

h
m
y
_
+
_
2

h
m
y
_
2
+
2
= 0, (4)
where overdots denote derivatives with respect to time. The
Newtonian equation (3) simply means that the maximum
of the WP, located at x = x = (t ), follows the classical
trajectory. The equation for the quantity
2 h
m
y(t ) has the
form a of a complex NL Riccati equation and describes
the time-dependence of the WP width that is related with
the position uncertainty via y
I
=
1
4 x
2

with x
2
= x
2

x
2
being the mean square deviation of position. This
quadratically NL complex equation will be the link to our
Pythagorean quantization as well as to an invariant with the
dimension of an action.
To show this, a new (real) variable (t ) is introduced
via
_
2 h
m
y
I
_
=
1

2
. Inserting this into the imaginary part of
equation (4) allows one to determine the real part of the
variable as
_
2 h
m
y
R
_
=

, which, when inserted into the real

part of (4) together with the above denition of
_
2 h
m
y
I
_
, nally
turns the complex Riccati equation into the real NL so-called
Ermakov equation
1
for (t ),
+
2
=
1

3
. (5)
It had been shown by Ermakov [9] in 1880, i.e. 45 years
before quantum mechanics was formulated by Schr odinger
and Heisenberg, that from the pair of equations (3) and (5),
coupled via
2
, by eliminating
2
from the equations, a
dynamical invariant, the Ermakov-invariant
I
L
=
1
2
_
( )
2
+
_

_
2
_
= const (6)
can be obtained (this invariant was rediscovered by several
authors, also in a quantum mechanical context; see,
e.g. [1012]).
This invariant has (at least) two remarkable properties:
(i) it is also a constant of motion for = (t ), in the
case where the corresponding Hamiltonian does not have
this property; (ii) apart from a missing constant factor m,
1
The author is grateful to the referee for bringing to his attention that this
equation had been studied already in 1874 by Steen [6]. However, Steens
work was ignored by mathematicians and physicists for more than a century,
because it was published in Danish in a journal usually not containing many
articles on mathematics. An English translation of the original paper [7] and
generalizations can be found in [8].
i.e. mass of the system, it has the dimension of an action,
not of an energy. The missing factor m is due to the fact
that Ermakov used the mathematical equation (3), whereas
in a physical context, Newtons equation of motion, i.e.
equation (3) multiplied by m, is relevant.
Furthermore, as will be shown below, an invariant of this
type also exists for certain dissipative systems, i.e. systems
for which a conventional Hamiltonian does not even exist.
Also, factorization of the corresponding operator leads to
generalized creation and annihilation operators (see end of
this section). In this context, the complex Riccati equation (4)
again plays the central role.
There are different ways of treating the (inhomogeneous)
Riccati equation. Instead of transforming it into the (real)
NL Ermakov equation (5), it can be solved directly by
transforming it into a homogeneous NL (complex) Bernoulli
equation if a particular solution
_
2 h
m
y
_
of the Riccati equation
is known. The general solution of equation (4) is then given
by
2 h
m
y =
2 h
m
y +
2 h
m
v(t ) where
2 h
m
v(t ) fulls the Bernoulli
equation
_
2

h
m
v
_
+
_
4

h
m
y
__
2

h
m
v
_
+
_
2

h
m
v
_
2
= 0. (7)
The coefcient A = 2(
2 h
m
y) of the linear term depends
on the particular solution. Equation (7) can be linearized via
2 h
m
v =
1
w(t )
to yield
w Aw = 1, (8)
which can be solved straightforwardly. For constant A, w(t )
can be expressed in terms of exponential or hyperbolic
functions (for real A) or trigonometric functions (for
imaginary A). For A being TD, w(t ) and hence
2 h
m
v(t ) can be
expressed in terms of I(t ) =
_
t
dt

_
t

dt

A(t

)
. So the general
solution of equation (4) can be written as
2

h
m
y(t ) =
2

h
m
y +
d
dt
ln [w
0
+I(t )], (9)
dening a one-parameter family of solutions depending on the
(complex) initial value of w
0
=
_
2 h
m
v
0
_
1
as parameter.
Comparison with SUSY quantum mechanics [13, 14]
shows that this solution is formally identical to the most
general superpotential W(x), fullling a real Riccati equation
and leading to a one-parameter family of isospectral potentials
that have the same SUSY partner potential (see, e.g. [1517]).
A major difference between the SUSY situation and the one
in our TD case (apart from replacing the spatial variable
by a temporal one) is the fact that the variables of the NL
equations (4) and (7) are complex, whereas W(x) is real. Also,
the parameter w
0
=
_
2 h
m
v
0
_
1
in our case is generally complex
and determines the initial conditions.
Another property of the Riccati equation, particularly
interesting in a quantum mechanical context, is the existence
of a superposition principle for this NL differential
equation [1820]. This is related to the fact that the Riccati
equation can always be linearized. In our case, this can be
achieved using the ansatz
_
2

h
m
y
_
=

, (10)
3
Phys. Scr. 87 (2013) 038117 D Schuch
with complex (t ), leading to

+
2
(t ) = 0, (11)
which has the form of the Newton-type equation (3) of the
corresponding problem, but now for a complex variable.
First, a kind of geometric interpretation of the motion
of in the complex plane shall be given. Expressed in
Cartesian coordinates, can be written as = u +iz, or in
polar coordinates as = e
i
. Inserting the polar form into
equation (10) leads to
_
2

h
m
y
_
=

+i , (12)
where the real part is already identical to
2 h
m
y
R
, as dened
above.
The quantity dened above in
2 h
m
y
I
as being
proportional to the position uncertainty is identical to the
absolute value of if it can be shown that
=
1

2
. (13)
This, however, can be proven by simply inserting real and
imaginary parts of (12) into the imaginary part of the Riccati
equation (4). Comparing relation (13), that can also be written
in the form
zu uz =
2
= 1, (14)
with the motion of a particle under the inuence of a
central force in two-dimensional physical space, it shows
that this relation corresponds to the conservation of angular
momentum, but here for the motion in the complex plane!
Relation (14) also shows, that real and imaginary parts,
or phase and amplitude, respectively, of the complex quantity
are not independent of each other but uniquely coupled. This
coupling, which is, as mentioned in the introduction, typical
for quantum systems (but not only for these) is due to the
quadratic nonlinearity in the Riccati equation. We will nd an
analogous situation also in the TI case, discussed in the next
section.
It should also be mentioned that the real part of the
variable
_
2 h
m
y
_
, as given in equation (12), does not depend on
the actual size of since only the relative change in time
matters; so this quantity is invariant on different scales.
The variables , and allow one to express the
uncertainties of position and momentum in a way that the
corresponding contribution to the energy of the WP can be
written in Lagrangian or Hamiltonian form. The uncertainties
then read
x
2

L
=

h
2m

2
L
, p
2

L
=

hm
2
_

2
L
+
2

2
L
_
,
[ x, p]
+

L
=

h
L

L
, U
L
=

h
2
4
_
1 +(
L

L
)
2
_
,
(15)
where U
L
= x
2

L
p
2

L
, p = p p and [. . . , . . .]
+
denotes
the anti-commutator (the subscript L has been added to
distinguish between quantities corresponding to the linear
SE and those related to its NL modication, described in
section 5).
The energy of the WP can be written as
E = H
L
= E
cl
+

E (16)
with the classical energy E
cl
=
m
2

2
+
m
2

2
=
1
2m
p
2
+
m
2

2
x
2
and the contribution from the uncertainties
(corresponding to the ground-state energy

E = E
0
=
h
2
) in
the form

E =
p
2

2m
+
m
2

2
x
2
=

h
4
{
2
+
2

2
+
2

2
}. (17)
In search of a Hamiltonian formalismfor the uncertainties
similar to the classical Hamiltonian formalism, one can try
using the difference

L =
1
2m
p
2

m
2

2
x
2
, expressed in
terms of , and , as a kind of Lagrangian to determine
the corresponding canonical momenta, leading to

L

= p

=

h
2
and


L

= p

=

h
2

2
=

h
2
. (18)
So,

E can nally be written as a Hamiltonian in the form

H(, p

, , p

) =
p
2

h
+
p
2

h
2
+

h
4

2
, (19)
which yields the correct Hamiltonian equation of motion
equivalent to the Ermakov equation (5) for (t ).
Let us now try to establish the relation to the Pythagorean
triples, introduced in section 2. For this purpose we consider
the complex quantity C =R+iI =
2 h
m
y =

+i , where R=

and I = =
1

2
can be considered the catheti and the
absolute value |
2 h
m
y| =
_
_

_
2
+
2
the hypothenuse of a
right-angled triangle in the complex plane.
Allow for a short intermezzo before proceeding to the
triples. If we multiply all three sides by
h
2

2
we obtain as new
catheti a =
h
2
, b =
h
2
= p

and thus Pythagoras theorem

gives us the uncertainty relation expressed as products of
the canonical variables we introduced above, i.e. a
2
+b
2
= c
2
yields
p
2

+(p

)
2
=

h
2
4
_
1 +( )
2
_
= U
L
. (20)
Now let us return to the time evolution of our complex
quantity C =
2 h
m
y. In the following, only the case V = 0 (i.e.
= 0) will be considered explicitly. As shown above, with a
particular solution of the Riccati equation its inhomogeneity
can always be removed. The resulting additional linear term
(at least for constant coefcient A) can also be removed. So
we are dealing with a complex equation of the form
d
dt
C +C
2
= 0. (21)
Then
d
dt
C is also a complex quantity, C
2
, where its real
and imaginary parts as well as its absolute value again dene
a right-angled triangle (in the complex plane) and each side
contains contributions from R and I, i.e. {C
2
} =R
2
I
2
,
{C
2
} = 2 RI and |C|
2
=R
2
+I
2
.
If we now assume that R and I are integers (with
R > I), all three sides of the right-angled triangle created
by C
2
in the complex plane are also integers. As examples, we
4
Phys. Scr. 87 (2013) 038117 D Schuch
choose: (a) R= 2, I = 1: R
2
I
2
= 3, 2 RI = 4, R
2
+I
2
=
5 with 9 +16 = 25; (b) R= 3, I = 2: R
2
I
2
= 5, 2 RI =
12 and R
2
+I
2
= 13 with 25 +144 = 169.
All possible Pythagorean triples can be obtained in this
way
2
just by applying all integers R and I with R>I. In a
physical context this means that whenever a physical quantity
obeys a complex Riccati equation and this quantity can be
quantized in the sense that its real and imaginary parts can
be expressed as multiples of some units, its evolution (in time
or space, depending on the respective derivative) can also
be expressed in terms of the same units. An example of a
Riccati equation where the complex variable depends on space
(instead of time) will be given in the next section.
To conclude the discussion based on WP solutions of
the TDSE which (particularly in the context of quantum
optics) can also be considered as CSs, it shall also be shown
how the complex Riccati variable can be used to dene
generalized creation/annihilation operators. These can be
used to construct CSs with TD width that are no minimum
uncertainty WPs but full the Schr odingerRobertson
uncertainty relation [22, 23].
WPs with TD width are not only known from the TDSE
for the free motion. They also occur for the HO with constant
frequency
0
if the initial state is not the ground state (leading
to oscillating width), for the HO with TD frequency (t )
and for effective Hamiltonians describing open dissipative
quantum systems.
The standard creation/annihilation operators can be
obtained by factorizing the Hamiltonian H
op
of the HO
[24, 25] or an operator related to it via

H
op
=
H
op

h
0
= (a
+
a +
1
2
), (22)
where a
+
a is the so-called number operator and the creation
and annihilation operators are dened by
a
+
=i
_
m
2

h
0
_
p
op
m
+i
0
x
_
=
1
2

h
0
_

x
+

m
0
x
_
,
(23)
a =i
_
m
2

h
0
_
p
op
m
i
0
x
_
=
1
2

h
0
_

h

x
+

m
0
x
_
,
(24)
where p
op
=
h
i

x
and a is the adjoint operator of a
+
.
The number that is the eigenvalue of a
+
a is the number
of quanta of the action

h since
H
op

0
has the dimension of an
action! With the help of a, the ground state wave function
can be obtained and from this, by successive application of
a
+
, the excited states can be created. Via superposition of
all these states, Schr odinger obtained a stable Gaussian WP
(with constant width) [26]. Generalizations of Schr odingers
approach were achieved for the description of coherent
2
The Pythagorean triples can also be understood in terms of rational
number points on the unit circle. This can be seen dividing both sides of
the Pythagorean theorem by c
2
leading to
_
a
c
_
2
+
_
b
c
_
2
= 1. Therefore, a
correspondence between points on the unit circle with rational coordinates
and Pythagorean triples exists. The abovementioned algorithm can then be
derived by trigonometric methods or by stereographic projection. For further
details, see e.g. [21].
light beams emitted by lasers in terms of what is now
called CS.
One of at least three different denitions of CSs is
that these are eigenstates of the annihilation operator a with
(complex) eigenvalue z, a|z = z|z. Comparing the CS |z
for the HO with the minimum uncertainty WP solution in
the form of equation (2), it shows that
0
=
2 h
m
y
I
=
1

2
0
. So,
in denitions (23) and (24), i
0
can be replaced by i
2 h
m
y
I
.
Therefore, for the more general case of WPs or CSs with TD
width,
_
2 h
m
y
R
= 0
_
, i
0
must be replaced by the full complex
quantity
_
2 h
m
y
_
in a and by
_
2 h
m
y

_
in the adjoint operator a
+
.
If one then substitutes
1

0
=
0
in front of the brackets with
(t ), the generalized creation and annihilation operators take
the form
a
+
(t ) = i
_
m
2

h
(t )
_
p
op
m

_
2

h
m
y

_
x
_
(25)
a(t ) = i
_
m
2

h
(t )
_
p
op
m

_
2

h
m
y
_
x
_
. (26)
These operators can even be turned into constants of
motion if an additional phase factor is taken into account.
But in the case of CSs, as discussed here, this factor can
be absorbed into the phase of the CS and will therefore be
omitted in the following (for further details see [27]).
Employing the above denition of the CS, but now
with our generalized annihilation operator, i.e. a(t )|z = z|z,
the CS (in position representation, x|z =
z
(x, t )) can be
obtained as

z
(x, t ) =
_
m

h
_
1/4

1/2
exp
_
i
_
y x
2
+
p x

h
+
px
2

h
__
,
(27)
which is in complete agreement with our WP denition (2)
(here, also N(t ) and K(t ) are now specied).
From the mean values of position and momentum, x
z
=
and p
z
= m , calculated with these CSs, the real and
imaginary parts of the complex eigenvalue z = z
R
+iz
I
can be
determined to be
z
R
=
_
m
2

h
_

_
, z
I
=
_
m
2

h
( ) , (28)
which looks familiar when compared with the Ermakov
invariant (6). Indeed, the absolute square of z is, up to a
constant factor, identical to I
L
,
I
L
=

h
m
_
z
2
I
+ z
2
R
_
=

h
m
zz

=

h
m
|z|
2
. (29)
An operator, corresponding to I
L
can then be written in
analogy to

H
op
=
H
op
h
0
as
m

h
I
L,op
= [a
+
(t ) a(t ) +
1
2
]. (30)
Factorization of this operator was also used [28] to nd
generalized creation and annihilation operators for the HO
with TD frequency but these operators were expressed in
terms of and or in terms of a complex variable [29]
5
Phys. Scr. 87 (2013) 038117 D Schuch
corresponding to (t ) which fullls the linear equation (11)
instead of
2 h
m
y fullling the NL equation (4). The Riccati
equation (4) for
2 h
m
y can be solved directly, as shown above.
The sensitivity of the solutions to the initial conditions
becomes obvious immediately and the Riccati form is still
valid in cases where a dissipative environment is taken into
account (see below); advantages the other approaches are
missing.
To conclude this section dealing with the TDSE, it
should be mentioned that a corresponding Riccati equation
for the WP width also exists in momentum space. There the
variable is just the inverse of the one in position space, i.e.
_
2 h
m
y
_
1
=

. Furthermore, the Ermakov invariant is directly

proportional to the exponent of the Wigner function which
is closest to the classical phase-space picture of the
corresponding problem (for details, see [30]).
4. Complex Riccati equations and the
time-independent SE
We have seen in the TD case that the real and imaginary
parts, or phase and amplitude of the complex variable
(t ) = e
i
which fulls the linear equation (11), obtained
via equation (10) from the Riccati equation (4), are not
independent of each other but coupled via the conservation
law (14). A similar situation exists when considering the
TISE, but now in the space-dependent case.
This can be shown using Madelungs hydrodynamic
formulation of quantum mechanics [31] where the wave
function is written in polar form as
(r, t ) =
1/2
(r, t ) exp
_
i

h
S(r, t )
_
(31)
with the square root of the probability density =

as
amplitude and
1
h
S as phase (r is the position vector in three
dimensions).
Inserting this form into the TDSE (1) (now in three
dimensions), and replacing

x
by the nabla operator ), leads
to a modied HamiltonJacobi equation for the phase,

t
S +
1
2m
(S)
2
+ V

h
2
2m

1/2

1/2
= 0, (32)
and a continuity equation for the amplitude,

t
+
1
m
(S) = 0. (33)
Already here, the coupling of phase and amplitude can be
seen clearly since the HamiltonJacobi equation for the phase
S contains a term (misleadingly called quantum potential,
V
qu
) depending on , and the continuity equation for the
density contains S. It can be shown that also in the TI
case this coupling is not arbitrary but related to a conservation
law.
In 1994, Reinisch [32] presented a NL formulation of
TI quantum mechanics. Since in this case

t
= 0 and

t
S = E are valid, the continuity equation (33) (we now use
the notation
1/2
= || = a) turns into
(a
2
S) = 0 (34)
and the modied HamiltonJacobi equation into

h
2
2m
a +(V E) a =
1
2m
(S)
2
a. (35)
Equation (34) is denitely fullled for S = 0, turning
(35) into the usual TISE for the real wave function a = ||
with position-independent phase S. (NB: the kinetic energy
term divided by a is just identical to V
qu
!)
However, equation (34) can also be fullled for S = 0
if only the conservation law
S =
C
a
2
(36)
is fullled with constant (or, at least, position-independent) C.
This relation now shows explicitly the coupling between
phase and amplitude of the wave function and is equivalent
to equation (13) in the TD case. Inserting (36) into the rhs of
equation (35) changes this into the Ermakov equation
a +
2m

h
2
(E V) a =
_
1

h
S
_
2
a =
_
C

h
_
2
1
a
3
, (37)
equivalent to equation (5) in the TD case.
A similar formulation of the TISE in terms of this
equation, but within a different context and different
applications has also been given in [33]. In another paper [34],
the relation between the Ermakov equation (37) and the
TISE has been extended to also include magnetic eld
effects. The NL differential equation (37) has also been used
to obtain numerical solutions of the TISE for single and
double-minimum potentials as well as for complex energy
resonance states; for details see [35, 36].
Returning to the method described in [32], so far the
energy E occurring in equation (37) is still a free parameter
that can take any value. However, solving this equation
numerically for arbitrary values of E leads, in general, to
solutions a that diverge for increasing x. Only if the energy
E is appropriately tuned to any eigenvalue E
n
of the TISE
(see equation (39), below) this divergence disappears and
normalizable solutions can be found. So, the quantization
condition that is usually obtained from the requirement of the
truncation of an innite series in order to avoid divergence
of the wave function is, in this case, obtained from the
requirement of nondiverging solutions of the NL Ermakov
equation (37) by variation of the parameter E. This has been
numerically veried in the case of the one-dimensional HO
and the Coulomb problem and there is the conjecture that this
property is universal in the sense that it does not depend on
the potential V (see [32, 37]).
The corresponding complex Riccati equation is now
given by

_
+
_

_
2
+
2m

h
2
(E V) = 0 (38)
with the complex variable
_

_
=
a
a
+i
1
h
S which corres-
ponds to
_
2 h
m
y
_
=

+i in the TD problem.
It is possible to show straightforwardly that equation (38)
can be linearized to yield the usual TISE

h
2
2m
+ V = E, (39)
6
Phys. Scr. 87 (2013) 038117 D Schuch
but in this form, the information on the coupling of phase and
amplitude, expressed by equation (36) and originating from
the quadratic NL term in equation (38), gets lost.
5. Dissipative systems with effective Hamiltonians
The conventional way of treating open dissipative systems
uses the system-plus reservoir approach, i.e. the system
of interest is coupled to some (in the limit innitely
many) environmental degrees of freedom (often HOs) and
system plus reservoir together are considered a closed
Hamiltonian system. Taking certain limits and applying
averaging processes nally leads to an irreversible dissipative
equation of motion for the system of interest. One of the
most often quoted approaches of that kind is the one of
Caldeira and Leggett [38, 39]. A similar idea, but with the
most minimalistic environment, namely only one additional
position variable (plus the corresponding momentum) is
behind the Bateman Hamiltonian [40] that represents a
constant of motion and provides an irreversible equation of
motion for the system,
x + x +
1
m

x
V = 0, (40)
i.e. a Newtonian equation with an additional linear velocity (or
momentum) dependent friction force (with friction coefcient
); actually the Langevin equation without stochastic force.
Since the environmental degrees of freedom are anyway
eliminated or ignored in the end, several approaches
exist where only the effect of the environment on the
observable system is taken into account without considering
the individual environmental degrees of freedom. This can
lead to modications of the classical Lagrange/Hamilton
formalism where the corresponding modied (linear) SE
is obtained via subsequent canonical quantization. The
canonical variables of these approaches are related with the
physical position and momentum variables via non-canonical
transformations in the classical case, corresponding to
non-unitary transformations in the quantum mechanical case
(for further details see [41, 42]). The most frequently applied
approach of that kind is the one of Caldirola [43] and
Kanai [44] which is uniquely related to one using an
exponentially expanding coordinate system [45], leading to
a Hamiltonian that is still a constant of motion. These
approaches can be directly linked to the aforementioned
ones. Using standard methods to eliminate the environmental
degrees of freedom, Yu and Sun [46, 47] have shown how
the conventional approach of CaldeiraLeggett leads directly
to the Hamiltonian operator of CaldirolaKanai. It is also
possible to eliminate the second set of variables of the
Bateman approach by imposing TD constraints [48] to get
to the Hamiltonian of the expanding system. Furthermore,
this approach and the one of CaldirolaKanai are connected
via an explicitly TD canonical transformation. In our context
it is interesting that for these two approaches also an exact
Ermakov invariant exists. In the quantized version, Gaussian
WP solutions can be obtained in the same cases as in
the conventional reversible theory, but now the maximum
follows a damped motion according to equation (40) and the
time-dependence of the width is determined by a modied
complex Riccati equation that can again be transformed into a
(real) Ermakov-type equation.
Another type of effective approaches starts already on the
quantum level by adding some friction terms W(x, p
op
, t ; )
to the Hamiltonian operator. This usually leads to NL
Hamiltonians, H
NL
= H
L
+ W, where quite different forms of
nonlinearities are considered in the literature (some are NL
only because some mean-value occurs in W) [4960]
3
.
Of these, an exact invariant was found [61, 62] for only two
approaches [58, 59].
In the following, only those NLSEs possessing an
Ermakov invariant shall be discussed explicitly since it can
be shown that the canonical approaches are unambiguously
related to these by a non-unitary transformation [41, 42].
In particular, the equations of motion for the WP maximum
and width can be uniquely transformed into each other [42].
The approach of Hasse [58] uses a combination of products
of position and momentum operators and their mean
values. The other one [59] is based on an irreversible
FokkerPlanck-type equation for the probability density that
is obtained from the usual continuity equation by adding a
time-symmetry-breaking diffusion term. Following a method
by Madelung and Mrowka [63, 64] this (real) so-called
Smoluchowski equation can be separated into two complex
equations: namely a modied SE for the wave function and
its complex conjugate

, provided the separation condition

D

2
x
2

= (ln ln ) (41)
with diffusion coefcient D is fullled (for details see,
e.g. [59, 66]).
This leads to the NLSE
i

h

t

NL
(x, t ) =
_
H
L
+

h
i
(ln
NL
ln
NL
)
_

NL
(x, t )
(42)
with a complex logarithmic nonlinearity.
The additional NL term (W
SCH
) can be written as real and
imaginary contributions in the form
W
SCH
= W
R
+iW
I
=

2

h
i
_
ln

NL

NL

_
ln

NL

NL
__
+

2

h
i
(ln
NL
ln
NL
) , (43)
where the real part only depends on the phase of the wave
function and provides the friction force in the averaged
equation of motion. The imaginary part does not contribute to
dissipation but introduces irreversibility into the evolution of
the wave function. It corresponds to the diffusion term in the
Smoluchowski equation, but still allows for normalizability
due to the subtraction of the mean value of ln . Comparison
with the afore-mentioned approaches shows that the real part
is just identical to Kostins term [54] and the imaginary part
corresponds to Berettas term [5153] introduced to describe
non-equilibrium systems (without dissipation).
The imaginary part breaks the time-reversal symmetry on
the level of the probability density, introduces a non-unitary
time evolution and turns the Hamiltonian into a non-Hermitian
3
S ussmann [57] is quoted in [58].
7
Phys. Scr. 87 (2013) 038117 D Schuch
one while still guaranteeing normalizable wave functions and
real energy mean values since its mean value vanishes.
An interesting interpretation of W
I
can be found if one
identies, according to Gr ossing et al [65], the Einstein
diffusion coefcient with the quantum mechanical one (if
the SE is considered a diffusion equation with imaginary
diffusion coefcient), i.e. D =
kT
m
=
h
2m
with temperature T
and Boltzmanns constant k. Then, W
I
turns into
W
I
= iTk (ln
NL
ln
NL
) , (44)
where kln = k
_
+

ln dx has a form like the

denition of entropy, S. So, the mean value of the linear
Hamiltonian that still represents the energy of the system,
H
L
= E, together with the second term of (44) would look
like E iTS, i.e. it has similarity with an expression for
the free energy only that here, again, the imaginary unit i
turns up in the quantum mechanical context. This point still
needs further investigation. In the context of our logarithmic
NLSE (42),
NL
is the probability density corresponding to
the Gaussian WP, which can be considered as a pure state,
so a direct comparison with the von Neumann entropy in the
usual way seems problematic. However, in Berettas work W
I
is discussed in the context of density operators and direct links
to an interpretation in terms of entropy are given [5153].
Also from the real part of W
SCH
no additional term
to the energy mean value occurs, so this is still given by
the mean value of the operators of kinetic and potential
energies. This real part is however not arbitrary but is uniquely
determined by the separation condition and provides the
correct dissipative friction forces in the equation of motion for
the mean values. Besides, the ratio of energy dissipation (for
the classical contribution) is in agreement with the classical
counterpart and arises because the mean values are calculated
with
NL
(the solution of equation (42)) instead of
L
.
The real part, by itself, would provide dissipation
but retain a unitary time-evolution of the wave function,
whereas the imaginary part, on its own, would provide
irreversibility via a non-unitary time-evolution but no
dissipation. Consequently only the combination of real and
imaginary parts provides all the desired properties of the
quantumsystemunder consideration. The reason for this is the
coupling of phase and amplitude of the wave function since
W
R
depends on the phase and W
I
on the amplitude.
The relation between the two NL approaches is discussed
in detail in [42] and can be traced back to a modication of
the Riccati equation (38) by adding a linear term. The two
NLSEs have the same WP solutions where, in both cases, the
maximum (t ) follows an equation of motion, like (40), with
a linear friction force and the WP width is determined by the
modied Riccati equation
_
2

h
m
y
_
NL
+
_
2

h
m
y
_
NL
+
_
2

h
m
y
_
2
NL
+
2
(t ) = 0 (45)
with an additional linear term depending on .
As in the conservative case,
_
2 h
m
y
I
_
NL
=
h
2 m x
2

NL
=
1

2
NL
is
valid but the real part of the complex Riccati variable now
takes the modied form
_
2

h
m
y
R
_
NL
=

NL

NL

2
. (46)
The corresponding Ermakov equation and invariant are
given now by

NL
+
_

2
4
_

NL
=
1

3
NL
(47)
and
I
NL
=
1
2

2
NL
__

_

NL

NL

2
_

_
2
+
_
1

2
NL

_
2
_
e
t
= const. (48)
From this it is obvious that, apart from the factor e
t
, I
NL
can be written in exactly the same form as in the conservative
case if expressed in terms of and
_
2 h
m
y
_
instead of and
, i.e.
I
NL
=
1
2

2
NL
_
_

_
2

h
m
y
R
_
NL

_
2
+
__
2

h
m
y
I
_
NL

_
2
_
e
t
,
(49)
which again shows the more universal validity of relations
when expressed in terms of the Riccati variable. Also, in
this dissipative case, the invariant (without the exponential
factor) can be factorized to yield generalized creation and
annihilation operators where the CSs obtained as eigenstates
of the annihilation operator are identical to the WP solutions
of the NLSEs (for details, see [27]).
6. Similarities with nonlinear dynamics and other
elds of physics
In NL dynamics, an important phenomenon is the Hopf
bifurcation as it can be the rst step on a route to turbulence
and chaos [67]. A system displaying this property can be
described by the NL evolution equation
r +r +r
3
= 0, (50)
which has the solution
r
2
(t ) =
r
2
0
e
2t
r
2
0
(1 e
2t
) +
. (51)
For 0, the trajectory approaches a xed point (the
origin); however, for < 0, it spirals towards a limit cycle
with radius r

= ||
1/2
[67]. The same type of differential
equation is also discussed by Gromann with respect to
self-similarity and scale-invariance (see [68]).
The relation to our Riccati equations (4) or (45) is easily
seen by multiplying equation (50) by 4r and introducing a new
variable R = 2r
2
, leading to

R +2R + R
2
= 0. (52)
This is exactly the form of the Bernoulli equation (7) that
can be obtained if a particular solution
_
2 h
m
y
_
of the Riccati
equation is known. The coefcient 2 of the linear term in
equation (52), corresponds to A = 2
_
2 h
m
y
_
in equation (7) and,
in the dissipative case, is simply replaced by A = 2
_
2 h
m
y
_
+ .
It has indeed been shown that, in the dissipative case, this
8
Phys. Scr. 87 (2013) 038117 D Schuch
bifurcation occurs and one obtains two different WP solutions
with different spreading behaviour of the WP width, different
uncertainties and different energies (for details, see [69]).
Another system with cubic nonlinearity is the Gross
Pitaevskii equation
i

h

t
=
_

h
2
2m
+ V(r, t ) + g||
2
_
, (53)
which is used in a mean eld approximation to describe
the macroscopic WP of a BEC where V(r, t ) can be
given by V(r, t ) =
m
2

2
(t )r
2
, i.e. a HO with TD frequency;
g parameterizes the strength of the atomic interaction.
Although equation (53) cannot be solved analytically, the
dynamics of the BEC characterized by this equation can be
described in terms of so-called moments M
n
(n = 1 4) (for
details, see e.g. [70]), where M
1
represents the norm, M
2
the
width, M
3
the radial momentum and M
4
the energy of the WP.
It can be shown that these moments satisfy a set of coupled
rst-order differential equations (where
d
dt
M
1
corresponds to
the conservation of probability or particle number). This set
can be reduced to a single equation for M
2
which can be
expressed, using a new variable X =

M
2
, in the form of an
Ermakov equation,

X +
2
(t )X =
k
X
3
, (54)
which, as shown in section 3, is equivalent to a complex
Riccati equation.
To include dissipative effects, one could add another
NL term like the logarithmic one from equation (42) to
the GrossPitaevskii equation which would correspond to
adding a linear term to the Riccati equation. So, one
simply has to solve this modied Riccati equation (or the
corresponding Ermakov equation) to obtain all moments M
n
for the dissipative BEC [71].
This treatment of the BEC is also interesting for
another reason. It has been shown by Lidsey [72] that a
correspondence can be established between positively-curved
isotropic, perfect uid cosmologies and the two-dimensional
harmonically-trapped BEC by mapping the equations of
motion for both systems onto a one-dimensional Ermakov
equation. The moments M
n
dened above can be identied
in the cosmological context with M
2
= scale factor, M
3
=
Hubble expansion parameter and M
4
= energy density of the
Universe. So the expanding Universe can be represented as an
Ermakov or complex Riccati system.
Without going into details, it should be noted that a
complex Riccati equation occurs also in the context of
problems related to quantum gravity (see, e.g. [73, 74]).
More examples could be mentioned from elds like elec-
trodynamics, optics, quantum optics, NL dynamics, super-
symmetry and others, but further details would go beyond the
scope of this paper.
7. Conclusions and perspectives
In classical physics only real quantities have any physical
signicance and energy in the form of Hamiltonians or
Lagrangians plays the dominant role. In quantum physics
however, action, i.e. the product of energy and time (or
position and momentum), is essentially the quantized entity.
The appearance of i =

1, and hence the use of complex

quantities in quantum mechanics, is not just a matter of
mathematical convenience but has fundamental physical
meaning. In the TISE, the wave function is such a complex
quantity that fulls a linear differential equation. We have
seen that this TISE is actually a linearized form of a
complex NL Riccati equation. Why should one bother with
a more complicated NL equation if there is a simpler
linear version at hand for which such nice properties like
a superposition principle exist? Because, in the linear form,
it is not obvious that real and imaginary parts, or phase
and amplitude, respectively, of the complex wave function
are not independent of each other but uniquely coupled via
a kind of conservation law. This coupling can be traced
back to the quadratic nonlinearity in the Riccati equation
and always occurs in systems that can be described by
complex Riccati equations. This complex quadratic term is
actually also the key to the quantization problem that goes
back to Plato and Pythagoras, namely, the search for an
algorithm that supplies the so-called Pythagorean triples. This
most abstract quantization problem in terms of numbers,
particularly integers, can therefore be related to physical
quantization problems whenever the evolution (in time or
space) of the physical system can be described by a complex
Riccati equation.
We found another example of that kind in the TDSE
where the time-evolution of the quantum uncertainties obeys
such an equation. The linearized version of this Riccati
equation is just a complex Newtonian equation of motion for
a quantity (t ) where the coupling of phase and amplitude
of this quantity corresponds to the conservation of angular
momentum for the motion of in the complex plane!
The complex TD Riccati equation (or its transformed
version, the real NL Ermakov equation) together with the
classical Newtonian equation for the system, lead to a
dynamical invariant with the dimension of action. This
Ermakov invariant not only essentially determines the Wigner
function of the system but, when the corresponding operator is
factorized, one obtains generalized creation and annihilation
operators that also apply in cases where the corresponding
Hamiltonian is no longer invariant. Specically, this is also
valid for certain dissipative systems when the Ermakov
invariant is expressed in terms of the complex Riccati variable.
This has been shown using some effective models for the
description of such open systems.
Finally, an initial link to NL dynamics was made
where properties like scale-invariance, bifurcations as a
route to chaos and other similar properties already emerge
when real Riccati equations occur. The relations to our NL
version of quantum mechanics, in particular the effect of
complexication will be further investigated. In addition,
formal similarities to elds like SUSY quantum mechanics,
quantum optics and cosmology shall be explored.
Acknowledgments
The author would like to dedicate this paper to Professor
Francesco Iachello on the occasion of his 70th birthday in
2012.
9
Phys. Scr. 87 (2013) 038117 D Schuch
References
[1] Rothman T 2012 Spektrum der Wissenschaft Februar 61
[2] Planck M 1900 Ann. Phys. 1 69
[3] Yang C N 1987 Schr odingerCentenary Celebration of a
Polymath ed C W Kilmister (Cambridge: Cambridge
University Press) p 53
[4] Heisenberg W 1973 Der Teil und das Ganze (M unchen:
Deutscher Taschenbuch Verlag) p 17
[5] Friberg J 1981 Historia Mathematica 8 277
[6] Steen A 1874 Overs. over d. K. Dan. Vidensk. Selsk. Forh. 1
[7] Redheffer R 1999 Aequationes Math. 58 60
[8] Redheffer R and Redheffer I 2001 Aequationes Math.
61 131
[9] Ermakov V P 1880 Univ. Jzv. Kiev 20 1
[10] Milne W E 1930 Phys. Rev. 35 863
[11] Pinney E 1950 Proc. Am. Math. Soc. 1 681
[12] Lewis H R 1967 Phys. Rev. Lett. 18 510
[13] Mielnik B and Rosas-Ortiz O 2004 J. Phys. A: Math. Gen.
37 10007
[14] Andrianov A A and Cannata F 2004 J. Phys. A: Math. Gen.
37 10297
[15] David J and Fern andez C 2010 AIP Conf. Proc. 1287 3
[16] Cooper F, Khare A and Sukhatme U 2001 Supersymmetry in
Quantum Mechanics (Singapore: World Scientic)
[17] Khare A and Sukhatme U 1989 J. Phys. A: Math. Gen.
22 2847
Keung W-Y, Sukhatme U P, Wang Q and Imbo T D 1989
J. Phys. A: Math. Gen. 22 L 987
[18] Hartley J G and Ray J R 1981 Phys. Rev. A 24 2873
[19] Sarlet W and Cantrijn F 1982 Phys. Lett. A 88 383
[20] Kevrekidis P G and Drossinos Y 2007 Math. Comput. Simul.
74 196
[21] Courant R and Robbins H 1941 What is Mathematics?
(London: Oxford University Press) (Suppl) chapter I, 3
[22] Schr odinger E 1930 Sitzungsber. Preuss. Akad. Wiss. (Berlin:
Springer) p 296
[23] Robertson H P 1929 Phys. Rev. 34 163
Robertson H P 1930 Phys. Rev. 35 667
Robertson H P 1934 Phys. Rev. 46 794
[24] Schr odinger E 1940 Proc. R. Irish Acad. 46 A 9
[25] Dirac P A M 1935 The Principles of Quantum Mechanics
(Oxford: Oxford University Press) chapter 34
[26] Schr odinger E 1926 Ann. Phys. 79 361
[27] Casta nos O, Schuch D and Rosas-Ortiz O 2013 J. Phys. A:
Math. Gen. at press (arXiv:1211.5109)
[28] Hartley J G and Ray J R 1982 Phys. Rev. D 25 382
[29] Malkin I A, Manko V I and Trifonov D A 1970 Phys. Rev. D
2 1371
[30] Schuch D 2005 Phys. Lett. A 338 225
[31] Madelung E 1926 Z. Phys. 40 322
[32] Reinisch G 1994 Physica A 206 229
[33] Lee R A 1982 J. Phys. A: Math. Gen. 15 2761
[34] Kaushal R S 2001 Int. J. Theor. Phys. 40 835
[35] Korsch H J and Laurent H 1981 J. Phys. B: At. Mol. Phys.
14 4213
[36] Korsch H J, Laurent H and M ohlenkamp R 1982 J. Phys. B:
At. Mol. Phys. 15 1
[37] Reinisch G 1997 Phys. Rev. A 56 3409
[38] Caldeira A O and Leggett A J 1981 Phys. Rev. Lett. 46 211
[39] Caldeira A O and Leggett A J 1983 Ann. Phys., NY 149 374
Caldeira A O and Leggett A J 1983 Ann. Phys., NY 153 445
(Erratum)
[40] Bateman H 1931 Phys. Rev. 38 815
[41] Schuch D 1997 Phys. Rev. A 55 935
[42] Schuch D 2012 J. Phys.: Conf. Ser. 380 012009
[43] Caldirola P 1941 Nuovo Cimento 18 393
[44] Kanai E 1948 Prog. Theor. Phys. 3 440
[45] Schuch D 1999 Int. J. Quantum Chem. 72 537
[46] Yu L H and Sun C P 1994 Phys. Rev. A 49 592
[47] Sun C P and Yu L H 1994 Phys. Rev. A 51 1845
[48] Schuch D, Guerrero J, L opez-Ruiz F F and Aldaya V
2013 in preparation
[49] Gisin N 1981 J. Phys. A: Math. Gen. 14 2259
[50] Gisin N 1982 Physica A 111 364
[51] Beretta G P 1987 Found. Phys. 17 365
[52] Beretta G P 2006 Phys. Rev. E 73 026113
[53] Beretta G P 2010 J. Phys.: Conf. Ser. 237 012004
[54] Kostin M D 1972 J. Chem. Phys. 57 3589
[55] Cho B R 1980 Suhak Gwa Mulri (Korean) (Math. and Phys.)
3 37
[56] Albrecht K 1975 Phys. Lett. B 56 127
[57] S ussmann G 1973 private communication
[58] Hasse R W 1975 J. Math. Phys. 16 2005
[59] Schuch D, Chung K-M and Hartmann H 1983 J. Math. Phys.
24 1652
[60] Doebner H-D and Goldin G A 1992 Phys. Lett. A 162 397
[61] Nassar A B 1986 J. Math. Phys. 27 755
[62] Nassar A B 1986 J. Math. Phys. 27 2949
[63] Madelung E 1950 Die Mathematischen Hilfsmittel des
Physikers (Berlin: Springer) p 432
[64] Mrowka B 1951 Z. Phys. 130 164
[65] Gr ossing G, Mesa Pasasio J and Schwabl H 2011 Found. Phys.
41 1437
[66] Schuch D and Chung K-M 1986 Int. J. Quantum Chem.
29 1561
[67] Schuster H G 1984 Deterministic Chaos An Introduction
(Weinheim: Physik Verlag) p 112
[68] Gromann S 1989 Ordnung und Chaos ed W Gerok (Stuttgart:
Wissenschaftliche Verlagsgesellschaft) p 101
[69] Schuch D 2002 J. Phys. A: Math. Gen. 35 8615
[70] Garcia-Ripoll J J, Perez-Garcia V M and Torres P 1999 Phys.
Rev. Lett. 83 1715
[71] Schuch D and Kaushal R S 2011 J. Phys.: Conf. Ser.
306 012032
[72] Lidsey J E 2004 Class. Quantum Grav. 21 777
[73] Kiefer C 1992 Phys. Rev. D 46 1658
[74] Kiefer C 2012 Phys. Rev. Lett. 108 021301
10