This action might not be possible to undo. Are you sure you want to continue?
a
r
X
i
v
:
h
e
p

t
h
/
9
4
0
5
0
2
9
v
2
5
M
a
y
1
9
9
4
LAUR94569
Supersymmetry and Quantum Mechanics
Fred Cooper
Theoretical Division, Los Alamos National Laboratory,
Los Alamos, NM 87545
Avinash Khare
Institute of Physics,
Bhubaneswar 751005, INDIA
Uday Sukhatme
Department of Physics, University of Illinois at Chicago,
Chicago, IL 60607
February 1, 2008
Abstract
In the past ten years, the ideas of supersymmetry have been prof
itably applied to many nonrelativistic quantum mechanical problems.
In particular, there is now a much deeper understanding of why cer
tain potentials are analytically solvable and an array of powerful new
approximation methods for handling potentials which are not exactly
solvable. In this report, we review the theoretical formulation of su
persymmetric quantum mechanics and discuss many applications. Ex
actly solvable potentials can be understood in terms of a few basic
ideas which include supersymmetric partner potentials, shape invari
ance and operator transformations. Familiar solvable potentials all
1
have the property of shape invariance. We describe new exactly solv
able shape invariant potentials which include the recently discovered
selfsimilar potentials as a special case. The connection between in
verse scattering, isospectral potentials and supersymmetric quantum
mechanics is discussed and multisoliton solutions of the KdV equation
are constructed. Approximation methods are also discussed within the
framework of supersymmetric quantum mechanics and in particular
it is shown that a supersymmetry inspired WKB approximation is
exact for a class of shape invariant potentials. Supersymmetry ideas
give particularly nice results for the tunneling rate in a double well
potential and for improving large N expansions. We also discuss the
problem of a charged Dirac particle in an external magnetic ﬁeld and
other potentials in terms of supersymmetric quantum mechanics. Fi
nally, we discuss structures more general than supersymmetric quan
tum mechanics such as parasupersymmetric quantum mechanics in
which there is a symmetry between a boson and a parafermion of
order p.
2
Contents
1 Introduction 5
2 Hamiltonian Formulation of Supersymmetric Quantum Me
chanics 13
2.1 State Space Structure of the SUSY Harmonic Oscillator . . . . 21
2.2 Broken Supersymmetry . . . . . . . . . . . . . . . . . . . . . . 23
3 Factorization and the Hierarchy of Hamiltonians 28
4 Shape Invariance and Solvable Potentials 31
4.1 General Formulas for Bound State Spectrum, Wave Functions
and SMatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.2 Shape Invariance in More Than One Step . . . . . . . . . . . . 34
4.3 Strategies For Categorizing Shape Invariant Potentials . . . . 36
4.3.1 Solutions Involving Translation . . . . . . . . . . . . . 36
4.3.2 Solutions Involving Scaling . . . . . . . . . . . . . . . . 42
4.3.3 Solutions of MultiStep Shape Invariance . . . . . . . . 47
4.3.4 Other Solutions . . . . . . . . . . . . . . . . . . . . . . 49
4.4 Shape Invariance and Noncentral Solvable Potentials . . . . . 51
4.5 Shape Invariance and 3Body Solvable Potentials . . . . . . . 52
5 Operator Transforms – New Solvable Potentials from Old 56
5.1 Natanzon Potentials . . . . . . . . . . . . . . . . . . . . . . . 60
5.2 Generalizations of Ginocchio and Natanzon Potentials . . . . . 62
6 Supersymmetric WKB Approximation 66
6.1 SWKB Quantization Condition for Unbroken Supersymmetry 66
6.2 Exactness of the SWKB Condition for Shape Invariant Potentials 68
6.3 Comparison of the SWKB and WKB Approaches . . . . . . . 70
6.4 SWKB Quantization Condition for Broken Supersymmetry . . 71
7 Isospectral Hamiltonians 72
7.1 One Parameter Family of Isospectral Potentials . . . . . . . . 72
7.2 Generalization to nParameter Isospectral Family . . . . . . . 75
7.3 nSoliton Solutions of the KdV Equation . . . . . . . . . . . . 77
3
8 Path Integrals and Supersymmetry 79
8.1 Superspace Formulation of Supersymmetric Quantum Mechanics 85
9 Perturbative Methods for Calculating Energy Spectra and
Wave Functions 87
9.1 Variational Approach . . . . . . . . . . . . . . . . . . . . . . . 88
9.2 δ Expansion Method . . . . . . . . . . . . . . . . . . . . . . . 91
9.3 Supersymmetry and Double Well Potentials . . . . . . . . . . 93
9.4 Supersymmetry and LargeN Expansions . . . . . . . . . . . . 97
10 Pauli Equation and Supersymmetry 99
11 Supersymmetry and the Dirac Equation 102
11.1 Dirac Equation With Lorentz Scalar Potential . . . . . . . . . 103
11.2 Supersymmetry and the Dirac Particle in a Coulomb Field . . 105
11.3 SUSY and the Dirac Particle in a Magnetic Field . . . . . . . 108
11.4 SUSY and the Euclidean Dirac Operator . . . . . . . . . . . . 109
11.5 Path Integral Formulation of the Fermion Propagator . . . . . 111
12 Singular Superpotentials 114
12.1 General Formalism, Negative Energy States and Breakdown
of the Degeneracy Theorem . . . . . . . . . . . . . . . . . . . 115
12.2 Bound States in the Continuum . . . . . . . . . . . . . . . . . 120
13 Parasupersymmetric Quantum Mechanics and Beyond 128
13.1 Parasupersymmetric Quantum Mechanics . . . . . . . . . . . . 129
13.2 Orthosupersymmetric Quantum Mechanics . . . . . . . . . . . 135
14 Omitted Topics 137
4
1 Introduction
Physicists have long strived to obtain a uniﬁed description of all basic inter
actions of nature, i.e. strong, electrowea k, and gravitational interactions.
Several ambitious attempts have been made in the last two decades, and it is
now widely felt that supersymmetry (SUSY) is a necessary ingredient in any
unifying approach. SUSY relates bosonic and fermionic degrees of freedom
and has the virtue of taming ultraviolet divergences. It was discovered in
1971 by Gel’fand and Likhtman [1], Ramond [2] and Neveu and Schwartz
[3] and later was rediscovered by several groups [4, 5, 6]. The algebra in
volved in SUSY is a graded Lie algebra which closes under a combination of
commutation and anticommutation relations. It was ﬁrst introduced in the
context of the string models to unify the bosonic and the fermionic sectors.
It was later shown by Wess and Zumino [5] how to construct a 3 +1 dimen
sional ﬁeld theory which was invariant under this symmetry and which had
very interesting properties such as a softening of ultraviolet divergences as
well as having paired fermionic and bosonic degrees of freedom. For particle
theorists, SUSY oﬀered a possible way of unifying spacetime and internal
symmetries of the Smatrix which avoided the nogo theorem of Coleman and
Mandula [7] which was based on the assumption of a Lie algebraic realization
of symmetries (graded Lie algebras being unfamiliar to particle theorists at
the time of the proof of the nogo theorem). Gravity was generalized by
incorporating SUSY to a theory called supergravity [8, 9]. In such theories,
Einstein’s general theory of relativity turns out to be a necessary consequence
of a local gauged SUSY. Thus, local SUSY theories provide a natural frame
work for the uniﬁcation of gravity with the other fundamental interactions
of nature.
Despite the beauty of all these uniﬁed theories, there has so far been
no experimental evidence of SUSY being realized in nature. One of the
important predictions of unbroken SUSY theories is the existence of SUSY
partners of quarks, leptons and gauge bosons which have the same masses
as their SUSY counterparts. The fact that no such particles have been seen
implies that SUSY must be spontaneously broken. One hopes that the scale
of this breaking is of the order of the electroweak scale of 100 GeV in order
that it can explain the hierarchy problem of mass diﬀerences. This leads
to a conceptual problem since the natural scale of symmetry breaking is
the gravitational or Planck scale which is of the order of 10
19
GeV. Various
5
schemes have been invented to try to resolve the hierarchy problem, including
the idea of nonperturbative breaking of SUSY. It was in the context of this
question that SUSY was ﬁrst studied in the simplest case of SUSY quantum
mechanics (SUSY QM) by Witten [10] and Cooper and Freedman [11]. In
a subsequent paper, a topological index was introduced (the Witten index)
by Witten [12] for studying SUSY breaking and several people studied the
possibility that instantons provide the nonperturbative mechanism for SUSY
breaking. In the work of Bender et al. [13] a new critical index was introduced
to study nonperturbatively the breakdown of SUSY in a lattice regulated
theory. Thus, in the early days, SUSY was studied in quantum mechanics as
a testing ground for the nonperturbative methods of seeing SUSY breaking
in ﬁeld theory.
Once people started studying various aspects of SUSY QM, it was soon
clear that this ﬁeld was interesting in its own right, not just as a model
for testing ﬁeld theory methods. It was realized that SUSY gives insight
into the factorization method of Infeld and Hull [14] which was the ﬁrst
method to categorize the analytically solvable potential problems. Gradually
a whole technology was evolved based on SUSY to understand the solvable
potential problems. One purpose of this article is to review some of the major
developments in this area.
Before we present a brief historical development of supersymmetric quan
tum mechanics, let us note another remarkable aspect. Over the last 10
years, the ideas of SUSY have stimulated new approaches to other branches
of physics [15]. For example, evidence has been found for a dynamical SUSY
relating eveneven and evenodd nuclei. The Langevin equation and the
method of stochastic quantization has a path integral formulation which em
bodies SUSY. There have also been applications of SUSY in atomic, con
densed matter and statistical physics [15].
In SUSY QM one is considering a simple realization of a SUSY algebra
involving the fermionic and the bosonic operators. Because of the existence
of the fermionic operators which commute with the Hamiltonian, one obtains
speciﬁc relationships between the energy eigenvalues, the eigenfunctions and
the Smatrices of the component parts of the full SUSY Hamiltonian. These
relationships will be exploited in this article to categorize analytically solv
able potential problems. Once the algebraic structure is understood, the
results follow and one never needs to return to the origin of the FermiBose
symmetry. In any case, the interpretation of SUSY QM as a degenerate Wess
6
Zumino ﬁeld theory in one dimension has not led to any further insights into
the workings of SUSY QM.
The introduction by Witten [12] of a topological invariant to study dy
namical SUSY breaking led to a ﬂurry of interest in the topological aspects of
SUSY QM. Various properties of the Witten index were studied in SUSY QM
and it was shown that in theories with discrete as well as continuous spectra,
the index could display anomalous behavior [16, 17, 18, 19, 20, 21, 22, 23].
Using the WignerKirkwood ¯ h expansion it was shown that for systems in
one and two dimensions, the ﬁrst term in the ¯ h expansion gives the exact
Witten index [24]. Further, using the methods of SUSY QM, a derivation of
the AtiyahSinger index theorem was also given [25, 26, 27, 28]. In another
development, the relationship between SUSY and the stochastic diﬀerential
equations such as the Langevin equation was elucidated and exploited by
Parisi and Sourlas [29] and Cooper and Freedman [11]. This connection,
which implicitly existed between the FokkerPlanck equation, the path inte
grals and the Langevin equation was then used to prove algorithms about
the stochastic quantization as well as to solve nonperturbatively for the
correlation functions of SUSY QM potentials using the Langevin equation.
A path integral formulation of SUSY QM was ﬁrst given by Salomonson
and van Holten [30]. Soon afterwards it was shown by using SUSY methods,
that the tunneling rate through double well barriers could be accurately
determined [31, 32, 33, 34]. At the same time, several workers extended
ideas of SUSY QM to higher dimensionsal systems as well as to systems with
large numbers of particles with a motivation to understand the potential
problems of widespread interest in nuclear, atomic, statistical and condensed
matter physics [35, 36, 37, 38, 39, 40, 41, 42, 43].
In 1983, the concept of a shape invariant potential (SIP) within the struc
ture of SUSY QM was introduced by Gendenshtein [44]. This Russian paper
remained largely unnoticed for several years. A potential is said to be shape
invariant if its SUSY partner potential has the same spatial dependence as the
original potential with possibly altered parameters. It is readily shown that
for any SIP, the energy eigenvalue spectra could be obtained algebraically
[44]. Much later, a list of SIPs was given and it was shown that the energy
eigenfunctions as well as the scattering matrix could also be obtained alge
braically for these potentials [45, 46, 47, 48]. It was soon realized that the
formalism of SUSY QM plus shape invariance (connected with translations of
parameters) was intimately connected to the factorization method of Infeld
7
and Hull [14].
It is perhaps appropriate at this point to digress a bit and talk about the
history of the factorization method. The factorization method was ﬁrst intro
duced by Schr¨odinger [49] to solve the hydrogen atom problem algebraically.
Subsequently, Infeld and Hull [14] generalized this method and obtained a
wide class of solvable potentials by considering six diﬀerent forms of factor
ization. It turns out that the factorization method as well as the methods
of SUSY QM including the concept of shape invariance (with translation
of parameters), are both reformulations [50] of Riccati’s idea of using the
equivalence between the solutions of the Riccati equation and a related sec
ond order linear diﬀerential equation. This method was supposedly used for
the ﬁrst time by Bernoulli and the history is discussed in detail by Stahlhofen
[51].
The general problem of the classiﬁcation of SIPs has not yet been solved.
A partial classiﬁcation of the SIPs involving a translation of parameters was
done by Cooper et al. [52, 53]. It turns out that in this case one only gets
all the standard analytically solvable potentials contained in the list given by
Dutt et al. [54] except for one which was later pointed out by Levai [55]. The
connection between SUSY, shape invariance and solvable potentials [56, 57]
is also discussed in the paper of Cooper et al. [52] where these authors
show that shape invariance even though suﬃcient, is not necessary for exact
solvability. Recently, a new class of SIPs has been discovered which involves
a scaling of parameters [58]. These new potentials as well as multistep SIPs
[59] have been studied, and their connection with selfsimilar potentials as
well as with qdeformations has been explored [60, 61, 62, 63].
In yet another development, several people showed that SUSY QM oﬀers
a simple way of obtaining isospectral potentials by using either the Dar
boux [64] or AbrahamMoses [65] or Pursey [66] techniques, thereby oﬀering
glimpses of the deep connection between the methods of the inverse quantum
scattering [67], and SUSY QM [68, 69, 70, 71, 72]. The intimate connection
between the soliton solutions of the KdV hierarchy and SUSY QM was also
brought out at this time [73, 74, 75, 76].
Approximate methods based on SUSY QM have also been developed.
Three of the notable ones are the 1/N expansion within SUSY QM [77], δ ex
pansion for the superpotential [78] and a SUSY inspired WKB approximation
(SWKB) in quantum mechanics for the case of unbroken SUSY [79, 80, 81].
It turns out that the SWKB approximation preserves the exact SUSY rela
8
tions between the energy eigenvalues as well as the scattering amplitudes of
the partner potentials [82]. Further, it is not only exact for large n (as any
WKB approximation is) but by construction it is also exact for the ground
state of V
1
(x). Besides it has been proved [45] that the lowest order SWKB
approximation necessarily gives the exact spectra for all SIP (with transla
tion). Subsequently a systematic higher order SWKB expansion has been
developed and it has been explicitly shown that to O(¯ h
6
) all the higher order
corrections are zero for these SIP [83]. This has subsequently been gener
alised to all orders in ¯ h [84, 85]. Energy eigenvalue spectrum has also been
obtained for several nonSIP [86, 87, 88, 89] and it turns out that in many
of the cases the SWKB does better than the usual WKB approximation.
Based on a study of these and other examples, it has been suggested that
shape invariance is not only suﬃcient but perhaps necessary for the lowest
order SWKB to give the exact bound state spectra [90]. Some attempts have
also been made to obtain the bound state eigenfunctions within the SWKB
formalism [91, 92, 93].
Recently, Inomata and Junker [94] have derived the lowest order SWKB
quantization condition (BSWKB) in case SUSY is broken. It has recently
been shown that for the cases of shape invariant three dimensional oscillator
as well as for P¨oschlTeller I and II potentials with broken SUSY, this lowest
order BSWKB calculation gives the exact spectrum [95, 97]. Recently, Dutt
et al. [96] have also developed a systematic higher order BSWKB expansion
and using it have shown that in all the three (shape invariant) cases, the
higher order corrections to O(¯ h
6
) are zero. Further, the energy eigenvalue
spectrum has also been obtained in the case of several nonSIP and it turns
out that in many cases BSWKB does as well as (if not better than) the usual
WKB approximation [96].
Recursion relations between the propagators pertaining to the SUSY part
ner potentials have been obtained and explicit expressions for propagators of
several SIP have been obtained [98, 99].
Several aspects of the Dirac equation have also been studied within SUSY
QM formalism [100, 101, 102]. In particular, it has been shown using the re
sults of SUSY QM and shape invariance that whenever there is an analytically
solvable Schr¨odinger problem in 1dimensional QM then there always exists
a corresponding Dirac problem with scalar interaction in 1+1 dimensions
which is also analytically solvable. Further, it has been shown that there is
always SUSY for massless Dirac equation in two as well as in four Euclidean
9
dimensions. The celebrated problem of the Dirac particle in a Coulomb ﬁeld
has also been solved algebraically by using the concepts of SUSY and shape
invariance [103]. The SUSY of the Dirac electron in the ﬁeld of a magnetic
monopole has also been studied [104, 105]. Also, the classic calculation of
Schwinger on pair production from strong ﬁelds can be dramatically simpli
ﬁed by exploiting SUSY.
The formalism of SUSY QM has also been recently extended and models
for parasupersymmetric QM [106, 107, 108] as well as orthosupersymmetric
QM [109] of arbitrary order have been written down. The question of singular
superpotentials has also been discussed in some detail within SUSY QM
formalism [110, 111, 112, 113, 114]. Very recently it has been shown that
SUSY QM oﬀers a systematic method [115] for constructing bound states in
the continuum [116, 117, 118].
As is clear from this (subjective) review of the ﬁeld, several aspects of
SUSY QM have been explored in great detail in the last ten years and it is
almost impossible to cover all these topics and do proper justice to them. We
have therefore, decided not to pretend to be objective but cover only those
topics which we believe to be important and which we believe have not so
far been discussed in great detail in other review articles. We have, however,
included in Sec. 14 a list of the important topics missed in this review and
given some references so that the intersested reader can trace back and study
these topics further. We have been fortunate in the sense that review articles
already exist in this ﬁeld where several of these missing topics have been
discussed [119, 120, 121, 122, 54, 123, 124, 125, 100]. We must also apologize
to several authors whose work may not have been adequately quoted in this
review article in spite of our best attempts.
The plan of the article is the following: In Sec. 2, we discuss the Hamil
tonian formalism of SUSY QM. We have deliberately kept this section at a
pedagogical level so that a graduate student should be able to understand
and work out all the essential details. The SUSY algebra is given and the
connection between the energy eigenvalues, the eigenfunctions and the S
matrics of the two SUSY partner Hamiltonians are derived. The question of
unbroken vs. broken SUSY is also introduced at a pedagogical level using
polynomial potentials of diﬀerent parity and the essential ideas of partner po
tentials are illustrated using the example of a one dimensional inﬁnite square
well. The ideas of SUSY are made more explicit through the example of one
dimensional SUSY harmonic oscillator.
10
In Sec. 3, we discuss the connection between SUSY QM and factorization
and show how one can always construct a hierarchy of p ≥ 1 Hamiltonians
with known energy eigenvalues, eigenfunctions and Smatrices by starting
from any given Hamiltonian with p bound states whose energy eigenvalues,
eigenfunctions and Smatrices (reﬂection coeﬃcient R and transmission co
eﬃcient T) are known.
Sec. 4 is in a sense the heart of the article. We ﬁrst show that if the
SUSY partner potentials satisfy an integrability condition called shape in
variance then the energy eigenvalues, the eigenfunctions and the Smatrices
for these potentials can be obtained algebraically. We then discuss satisfying
the condition of shape invariance with translations and show that in this
case the classiﬁcation of SIP can be done and the resultant list of solvable
potentials include essentially all the popular ones that are included in the
standard QM textbooks. Further, we discuss the newly discovered one and
multistep shape invariant potentials when the partner potentials are related
by a change of parameter of a scaling rather than translation type. It turns
out that in most of these cases the resultant potentials are reﬂectionless and
contain an inﬁnte number of bound states. Explicit expressions for the energy
eigenvalues, the eigenfunctions and the transmission coeﬃcients are obtained
in various cases. It is further shown that the recently discovered selfsimilar
potentials which also statisfy qSUSY, constitute a special case of the SIP.
Finally, we show that a wide class of noncentral but separable potential prob
lems are also algebraically solvable by using the results obtained for the SIP
[131]. As a by product, exact solutions of a number of three body problems
in one dimension are obtained analytically [132].
In Sec. 5, we discuss the solvable but nonshape invariant Natanzon and
Ginocchio potentials and show that using the ideas of SUSY QM, shape
invariance and operator transformations, their spectrum can be obtained
algebraically. We also show that the Natanzon potentials are not the most
general solvable potentials in nonrelativistic QM.
Sec. 6 is devoted to a discussion of the SUSY inspired WKB approxima
tion (SWKB) in quantum mechanics both when SUSY is unbroken and when
it is spontaneously broken. In the unbroken case, we ﬁrst develop a system
atic higher order ¯ h expansion for the energy eigenvalues and then show that
for the SIP with translation, the lowest order term in the ¯ hexpansion gives
the exact bound state spectrum. We also show here that even for many of
the nonSIP, the SWKB does as well as if not better than the WKB approx
11
imation. We then discuss the broken SUSY case and in that case too we
develop a systematic ¯ hexpansion (BSWKB) for the energy eigenvalues. We
show that even in the broken case the lowest order BSWKB gives the exact
bound state spectrum for the SIP with translation.
Sec. 7 contains a description of how SUSY QM can be used to con
struct multiparameter families of isospectral and strictly isospectral poten
tials. As an illustration we give plots of the one continuous parameter family
of isospectral potentials corresponding to the onedimensional harmonic os
cillator. From here we are immediately able to construct the two as well as
multisoliton solutions of the KdV equation.
ln Sec. 8, we discuss more formal aspects of SUSY QM. In particular, we
discuss the path integral formulation of SUSY QM as well as various sub
tleties associated with the Witten index [12] ∆ = (−1)
F
. We also discuss
in some detail the connection of SUSY QM with classical stochastic pro
cesses and discuss how one can develop a systematic strong coupling and δ
expansion for the Langevin equation.
Sec. 9 contains a description of several approximation schemes like the
variational method, the δexpansion, largeN expansion, energy splitting in
double well potentials within the SUSY QM framework.
In Sec. 10, we discuss the question of SUSY QM in higher dimensions.
In particular, we discuss the important problem of a charged particle in a
magnetic ﬁeld (Pauli equation) in two dimensions and show that there is
always a SUSY in the problem so long as the gyromagnetic ratio is 2.
In Sec. 11, we show that there is always a SUSY in the case of massless
Dirac equation in two or four Euclidean dimensions in the background of
external electromagnetic ﬁelds. Using the results of SUSY QM we then list
a number of problems with nonuniform magnetic ﬁeld which can be solved
analytically. We also show here that whenever a Schr¨odinger problem in 1
dimensional QM is analytically solvable, then one can always obtain an exact
solution of a corresponding Dirac problem with the scalar coupling. We also
show how the calculation of the fermion propagator in an external ﬁeld can
be simpliﬁed by exploiting SUSY.
Sec. 12 contains a comprehensive discussion of the general problem of
singular superpotentials, explicit breaking of SUSY, negative energy states
and unpaired positive energy eigenstates. We also show here how to construct
bound states in the continuum within the formalism of SUSY QM.
Quantum mechanical models relating bosons and parafermions of order
12
p are described in Sec. 13. It is shown that such models encompass p su
persymmetries. Various consequences of such models are discussed including
the connection with the hierarchy of Hamiltonians as well as with strictly
isospectral potentials. We also discuss a quantum mechanical model where
instead there is a symmetry between a boson and an orthofermion of order
p.
Finally, in Sec. 14, we give a list of topics related to SUSY QM which we
have not discussed and provide some reference s for each of these topics.
2 Hamiltonian Formulation of Supersymmet
ric Quantum Mechanics
One of the key ingredients in solving exactly for the spectrum of one dimen
sional potential problems is the connection between the bound state wave
functions and the potential. It is not usually appreciated that once one
knows the ground state wave function (or any other bound state wave func
tion) then one knows exactly the potential (up to a constant). Let us choose
the ground state energy for the moment to be zero. Then one has from the
Schr¨odinger equation that the ground state wave function ψ
0
(x) obeys
H
1
ψ
0
(x) = −
¯ h
2
2m
d
2
ψ
0
dx
2
+ V
1
(x)ψ
0
(x) = 0 , (1)
so that
V
1
(x) =
¯ h
2
2m
ψ
′′
0
(x)
ψ
0
(x)
. (2)
This allows a global reconstruction of the potential V
1
(x) from a knowledge
of its ground state wave function which has no nodes (we will discuss the
case of using the excited wave functions later in Sec. 12). Once we realize
this, it is now very simple to factorize the Hamiltonian using the following
ansatz:
H
1
= A
†
A (3)
where
A =
¯ h
√
2m
d
dx
+ W(x) , A
†
=
−¯ h
√
2m
d
dx
+ W(x) . (4)
13
This allows us to identify
V
1
(x) = W
2
(x) −
¯ h
√
2m
W
′
(x) . (5)
This equation is the wellknown Riccati equation. The quantity W(x) is
generally referred to as the “superpotential” in SUSY QM literature. The
solution for W(x) in terms of the ground state wave function is
W(x) = −
¯ h
√
2m
ψ
′
0
(x)
ψ
0
(x)
. (6)
This solution is obtained by recognizing that once we satisfy Aψ
0
= 0, we
automatically have a solution to H
1
ψ
0
= A
†
Aψ
0
= 0.
The next step in constructing the SUSY theory related to the original
Hamiltonian H
1
is to deﬁne the operator H
2
= AA
†
obtained by reversing
the order of A and A
†
. A little simpliﬁcation shows that the operator H
2
is
in fact a Hamiltonian corresponding to a new potential V
2
(x).
H
2
= −
¯ h
2
2m
d
2
dx
2
+ V
2
(x) , V
2
(x) = W
2
(x) +
¯ h
√
2m
W
′
(x) . (7)
The potentials V
1
(x) and V
2
(x) are known as supersymmetric partner poten
tials.
As we shall see, the energy eigenvalues, the wave functions and the S
matrices of H
1
and H
2
are related. To that end notice that the energy
eigenvalues of both H
1
and H
2
are positive semideﬁnite (E
(1,2)
n
≥ 0) . For
n > 0, the Schr¨odinger equation for H
1
H
1
ψ
(1)
n
= A
†
Aψ
(1)
n
= E
(1)
n
ψ
(1)
n
(8)
implies
H
2
(Aψ
(1)
n
) = AA
†
Aψ
(1)
n
= E
(1)
n
(Aψ
(1)
n
) . (9)
Similarly, the Schr¨odinger equation for H
2
H
2
ψ
(2)
n
= AA
†
ψ
(2)
n
= E
(2)
n
ψ
(2)
n
(10)
implies
H
1
(A
†
ψ
(2)
n
) = A
†
AA
†
ψ
(2)
n
= E
(2)
n
(A
†
ψ
(2)
n
) . (11)
14
From eqs. (8)(11) and the fact that E
(1)
0
= 0, it is clear that the eigenvalues
and eigenfunctions of the two Hamiltonians H
1
and H
2
are related by (n =
0, 1, 2, ...)
E
(2)
n
= E
(1)
n+1
, E
(1)
0
= 0, (12)
ψ
(2)
n
= [E
(1)
n+1
]
−1/2
Aψ
(1)
n+1
, (13)
ψ
(1)
n+1
= [E
(2)
n
]
−1/2
A
†
ψ
(2)
n
. (14)
Notice that if ψ
(1)
n+1
( ψ
(2)
n
) of H
1
(H
2
) is normalized then the wave function
ψ
(2)
n
(ψ
(1)
n+1
) in eqs. (13) and (14) is also normalized. Further, the operato r A
(A
†
) not only converts an eigenfunction of H
1
(H
2
) into an eigenfunction of
H
2
(H
1
) with the same energy, but it also destroys (creates) an extra node in
the eigenfunction. Since the ground state wave function of H
1
is annihilated
by the operator A, this state has no SUSY partner. Thus the picture we get is
that knowing all the eigenfunctions of H
1
we can determine the eigenfunctions
of H
2
using the operator A, and vice versa using A
†
we can reconstruct all
the eigenfunctions of H
1
from those of H
2
except for the ground state. This
is illustrated in Fig. 2.1.
The underlying reason for the degeneracy of the spectra of H
1
and H
2
can
be understood most easily from the properties of the SUSY algebra. That is
we can consider a matrix SUSY Hamiltonian of the form
H =
¸
H
1
0
0 H
2
(15)
which contains both H
1
and H
2
. This matrix Hamiltonian is part of a closed
algebra which contains both bosonic and fermionic operators with commu
tation and anticommutation relations. We consider the operators
Q =
¸
0 0
A 0
, (16)
Q
†
=
¸
0 A
†
0 0
(17)
in conjunction with H. The following commutation and anticommutation
relation s then describe the closed superalgebra sl(1/1):
15
[H, Q] = [H, Q
†
] = 0 ,
¦Q, Q
†
¦ = H, ¦Q, Q¦ = ¦Q
†
, Q
†
¦ = 0 . (18)
The fact that the supercharges Q and Q
†
commute with H is responsibl
e for the degeneracy. The operators Q and Q
†
can be interpreted as oper
ators which change bosonic degrees of freedom into fermionic ones and vice
versa. This will be elaborated further below using the example of the SUSY
harmonic oscillator.
Let us look at a well known potential, namely the inﬁnite square well and
determine its SUSY partner potential. Consider a particle of mass m in an
inﬁnite square well potential of width L.
V (x) = 0, 0 ≤ x ≤ L ,
= ∞, −∞< x < 0, x > L . (19)
The ground state wave function is known to be
ψ
(1)
0
= (2/L)
1/2
sin(πx/L), 0 ≤ x ≤ L , (20)
and the ground state energy is E
0
=
¯h
2
π
2
2mL
2
.
Subtracting oﬀ the ground state energy so that we can factorize the
Hamiltonian we have for H
1
= H −E
0
that the energy eigenvalues are
E
(1)
n
=
n(n + 2)
2mL
2
¯ h
2
π
2
(21)
and the eigenfunctions are
ψ
(1)
n
= (2/L)
1/2
sin
(n + 1)πx
L
, 0 ≤ x ≤ L . (22)
The superpotential for this problem is readily obtained using eq. (6)
W(x) = −
¯ h
√
2m
π
L
cot(πx/L) (23)
and hence the supersymmetric partner potential V
2
is
V
2
(x) =
¯ h
2
π
2
2mL
2
[2cosec
2
(πx/L) −1] . (24)
16
The wave functions for H
2
are obtained by applying the operator A to the
wave functions of H
1
. In particular we ﬁnd that
ψ
(2)
0
∝ sin
2
(πx/L), ψ
(2)
1
∝ sin(πx/L) sin(2πx/L) . (25)
Thus we have shown using SUSY that two rather diﬀerent potentials
corresponding to H
1
and H
2
have exactly the same spectra except for the fact
that H
2
has one fewer bound state. In Fig. 2.2 we show the supersymmetric
partner potentials V
1
and V
2
and the ﬁrst few eigenfunctions. For convenience
we have chosen L = π and ¯ h = 2m = 1.
Supersymmetry also allows one to relate the reﬂection and transmission
coeﬃcients in situations where the two partner potentials have continuum
spectra. In order for scattering to take place in both of the partner potentials,
it is necessary that the potentials V
1,2
are ﬁnite as x →−∞ or as x →+∞
or both. Deﬁne:
W(x →±∞) = W
±
. (26)
Then
V
1,2
→W
2
±
as x →±∞. (27)
Let us consider an incident plane wave e
ikx
of energy E coming from the
direction x →−∞. As a result of scattering from the potentials V
1,2
(x) one
would obtain transmitted waves T
1,2
(k)e
ik
′
x
and reﬂected waves R
1,2
(k)e
−ikx
.
Thus we have
ψ
(1,2)
(k, x →−∞) →e
ikx
+ R
1,2
e
−ikx
,
ψ
(1,2)
(k
′
, x →+∞) →T
1,2
e
ik
′
x
. (28)
SUSY connects continuum wave functions of H
1
and H
2
having the same
energy analogously to what happens in the discrete spectrum. Thus we have
the relationships:
e
ikx
+ R
1
e
−ikx
= N[(−ik + W
−
)e
ikx
+ (ik + W
−
)e
−ikx
R
2
],
T
1
e
ik
′
x
= N[(−ik
′
+ W
+
)e
ik
′
x
T
2
], (29)
where N is an overall normalization constant. On equating terms with the
same exponent and eliminating N, we ﬁnd:
R
1
(k) =
W
−
+ ik
W
−
−ik
R
2
(k),
T
1
(k) =
W
+
−ik
′
W
−
−ik
T
2
(k), (30)
17
where k and k
′
are given by
k = (E −W
2
−
)
1/2
, k
′
= (E −W
2
+
)
1/2
. (31)
A few remarks are now in order at this stage.
(1) Clearly [R
1
[
2
= [R
2
[
2
and [T
1
[
2
= [T
2
[
2
, that is the partner potentials
have identical reﬂection and transmission probabilities.
(2) R
1
(T
1
) and R
2
(T
2
) have the same poles in the complex plane except that
R
1
(T
1
) has an extra pole at k = −iW
−
. This pole is on the positive imaginary
axis only if W
−
< 0 in which case it corresponds to a zero energy bound state.
(3) In the special case that W
+
= W
−
, we have that T
1
(k) = T
2
(k).
(4) When W
−
= 0 then R
1
(k) = −R
2
(k).
It is clear from these remarks that if one of the partner potentials is a con
stant potential (i.e. a free particle), then the other partner will be of necessity
reﬂectionless. In this way we can understand the reﬂectionless potentials of
the form V (x) = Asech
2
αx which play a critical role in understanding the
soliton solutions of the KdV hierarchy. Let us consider the superpotential
W(x) = A tanh αx . (32)
The two partner potentials are
V
1
= A
2
−A (A+ α
¯ h
√
2m
)sech
2
αx ,
V
2
= A
2
−A(A−α
¯ h
√
2m
)sech
2
.αx (33)
We see that for A = α
¯h
√
2m
, V
2
(x) corresponds to a constant potential so that
the corresponding V
1
is a reﬂectionless potential. It is worth noting that V
1
is
¯ hdependent. One can in fact rigorously show,though it is not mentioned in
most text books,that the reﬂectionless potentials are necessarily ¯ hdependent.
So far we have discussed SUSY QM on the full line (−∞< x < ∞). Many
of these results have analogs for the ndimensional potentials with spherical
symmetry. For example, in three dimensions one can make a partial wave
expansion in terms of the wave functions:
ψ
nlm
(r, θ, φ) =
1
r
R
nl
(r)Y
lm
(θ, φ) . (34)
18
Then it is easily shown [126] that the reduced radial wave function R satisﬁes
the onedimensional Schr¨odinger equation (0 < r < ∞)
−
¯ h
2
2m
d
2
ψ(r)
dr
2
+ [V (r) +
l(l + 1)¯ h
2
2mr
2
]ψ(r) = Eψ(r) (35)
We notice that there is an eﬀective one dimensional potential which contains
the original potential plus an angular momentum barrier. The asymptotic
form of the radial wave function for the l
′
th partial wave is
ψ(r, l) →
1
2k
′
[S
l
(k
′
)e
ik
′
r
−(−1)
l
e
−ik
′
r
] , (36)
where S
l
is the scattering function for the l
′
th partial wave. i.e. S
l
(k) = e
iδ
l
(k)
and δ is the phase shift.
For this case we ﬁnd the relations:
S
l
1
(k
′
) =
W
+
−ik
′
W
+
+ ik
′
S
l
2
(k
′
) . (37)
Here W
+
= W(r →∞).
We thus have seen that when H
1
contained a known ground state wave
function then we could factorize the Hamiltonian and ﬁnd a SUSY partner
Hamiltonian H
2
. Now let us consider the converse problem. Suppose we are
given a superpotential W(x). In this case there are two possibilities. The
candidate ground state wave function is the ground state for H
1
(or H
2
) and
can be obtained from:
Aψ
(1)
0
(x) = 0 ⇒ψ
(1)
0
(x) = N exp
−
√
2m
¯ h
x
W(y) dy
,
A
†
ψ
(2)
0
(x) = 0 ⇒ψ
(2)
0
(x) = N exp
+
√
2m
¯ h
x
W(y) dy
. (38)
By convention, we shall always choose W in such a way that amongst
H
1
, H
2
only H
1
(if at all) will have a normalizable zero energy ground state
eigenfunction. This is ensured by choosing W such that W(x) is posi
tive(negative) for large positive(negative) x. This deﬁnes H
1
to have fermion
number zero in our later formal treatment of SUSY.
If there are no normalizable solutions of this form, then H
1
does not
have a zero eigenvalue and SUSY is broken. Let us now be more precise. A
19
symmetry of the Hamiltonian (or Lagrangian) can be spontaneously broken
if the lowest energy solution does not respect that symmetry, as for example
in a ferromagnet, where rotational invariance of the Hamiltonian is broken
by the ground state. We can deﬁne the ground state in our system by a two
dimensional column vector:
[0 >= ψ
0
(x) =
¸
ψ
(1)
0
(x)
ψ
(2)
0
(x)
¸
(39)
For SUSY to be unbroken requires
Q[0 >= Q
†
[0 >= 0[0 > (40)
Thus we have immediately from eq. (18 ) that the ground state energy
must be zero in this case. For all the cases we discussed previously, the ground
state energy was indeed zero and hence the ground state wave function for
the matrix Hamiltonian can be written:
ψ
0
(x) =
¸
ψ
(1)
0
(x)
0
¸
(41)
where ψ
(1)
0
(x) is given by eq. (38).
If we consider superpotentials of the form
W(x) = gx
n
, (42)
then for n odd and g positive one always has a normalizable ground state
wave function. However for the case n even and g arbitrary, then there is
no normalizable ground state wave function. In general when one has a
superpotential W(x) so that neither Q nor Q
†
annihilates the ground state
as given by eq. (39) then SUSY is broken and the potentials V
1
and V
2
have
degenerate positive ground state energies. Stated another way, if the ground
state energy of the matrix Hamiltonian is non zero then SUSY is broken.
For the case of broken SUSY the operators A and A
†
no longer change the
number of nodes and there is a 11 pairing of all the eigenstates of H
1
and
H
2
. The precise relations that one now obtains are:
E
(2)
n
= E
(1)
n
> 0, n = 0, 1, 2, ... (43)
ψ
(2)
n
= [E
(1)
n
]
−1/2
Aψ
(1)
n
, (44)
20
ψ
(1)
n
= [E
(2)
n
]
−1/2
A
†
ψ
(2)
n
. (45)
while the relationship between the scattering amplitudes is still given by
eqs. (30) or (37). The breaking of SUSY can be described by a topological
quantum number called the Witten index [12] which we will discuss later.
Let us however remember that in general if the sign of W(x) is opposite as
we approach inﬁnity from the positive and the negative sides, then SUSY is
unbroken, whereas in the other cases it is always broken.
2.1 State Space Structure of the SUSY Harmonic Os
cillator
For the usual quantum mechanical harmonic oscillator one can introduce a
Fock space of boson occupation numbers where we label the states by the
occupation number n. To that eﬀect one introduces instead of P and q the
creation and annihilation operators a and a
†
. The usual harmonic oscillator
Hamiltonian is
H =
P
2
2m
+
1
2
mω
2
q
2
. (46)
Let us rescale the Hamiltonian in terms of dimensionless coordinates and
momenta x and p so that we measure energy in units of ¯ hω. We put
H = H¯ hω, q = (
¯ h
2mω
)
1/2
x, P = (2m¯ hω)
1/2
p. (47)
Then
H = (p
2
+
x
2
4
), [x, p] = i. (48)
Now introduce
a = (
x
2
+ ip), a
†
= (
x
2
−ip). (49)
Then
[a, a
†
] = 1, [N, a] = −a, [N, a
†
] = a
†
,
N = a
†
a, H = N +
1
2
. (50)
The usual operator formalism for solving the harmonic oscillator potential
is to deﬁne the ground state by requiring
a[0 >= 0 , (51)
21
which leads to a ﬁrst order diﬀerential equation for the ground state wave
function. The n particle state (which is the n’th excited wave function in the
coordinate representation) is then given by:
[n
b
>=
a
† n
(n!)
[0 > , . (52)
where we have used the subscript b to refer to the boson sector as distinct from
the fermions we will introduce below. For the case of the SUSY harmonic
oscillator one can rewrite the operators Q (Q
†
) as a product of the bosonic
operator a and the fermionic operator ψ. Namely we write Q = aψ
†
and
Q
†
= a
†
ψ where the matrix fermionic creation and annihilation operators
are deﬁned via:
ψ = σ
+
=
¸
0 1
0 0
, (53)
ψ
†
= σ
−
=
¸
0 0
1 0
. (54)
Thus ψ and ψ
†
obey the usual algebra of the fermionic creation and
annihilation operators, namely,
¦ψ
†
, ψ¦ = 1, ¦ψ
†
, ψ
†
¦ = ¦ψ, ψ¦ = 0, (55)
as well as obeying the commutation relation:
[ψ, ψ
†
] = σ
3
=
¸
1 0
0 −1
. (56)
The SUSY Hamiltonian can be rewritten in the form
H = QQ
†
+ Q
†
Q = (−
d
2
dx
2
+
x
2
4
)I −
1
2
[ψ, ψ
†
]. (57)
The eﬀect of the last term is to remove the zero point energy.
The state vector can be thought of as a matrix in the Schr¨odinger picture
or as the state [n
b
, n
f
> in this Fock space picture. Since the fermionic
creation and annihilation operators obey anticommutation relations hence
the fermion number is either zero or one. As stated before, we will choose
22
the ground state of H
1
to have zero fermion number. Then we can introduce
the fermion number operator
n
F
=
1 −σ
3
2
=
1 −[ψ, ψ
†
]
2
. (58)
Because of the anticommutation relation, n
f
can only take on the values
0 and 1. The action of the operators a, a
†
, ψ, ψ
†
in this Fock space are then:
a[n
b
, n
f
> = [n
b
−1, n
f
>, ψ[n
b
, n
f
>= [n
b
, n
f
−1 >,
a
†
[n
b
, n
f
> = [n
b
+ 1, n
f
>, ψ
†
[n
b
, n
f
>= [n
b
, n
f
+ 1 > . (59)
We now see that the operator Q
†
= −iaψ
†
has the property of changing a
boson into a fermion without changing the energy of the state. This is the
bosonfermion degeneracy characteristic of all SUSY theories.
For the general case of SUSY QM, the operator a gets replaced by A
in the deﬁnition of Q, Q
†
, i.e. one writes Q = Aψ
†
and Q
†
= A
†
ψ. The
eﬀect of Q and Q
†
are now to relate the wave functions of H
1
and H
2
which
have fermion number zero and one respectively but now there is no simple
Fock space description in the bosonic sector because the interactions are non
linear. Thus in the general case, we can rewrite the SUSY Hamiltonian in
the form
H = (−
d
2
dx
2
+ W
2
)I −[ψ, ψ
†
]W
′
. (60)
This form will be useful later when we discuss the Lagrangian formulation
of SUSY QM in Sec. 8.
2.2 Broken Supersymmetry
As discussed earlier, for SUSY to be a good symmetry, the operators Q
and Q
†
must annihilate the vacuum. Thus the ground state energy of the
superHamiltonian must be zero since
H = ¦Q
†
, Q¦.
Witten [12] proposed an index to determine whether SUSY is broken in
supersymmetric ﬁeld theories. The index is deﬁned by
△ = Tr(−1)
F
, (61)
23
where the trace is over all the bound states and continuum states of the
superHamiltonian. For SUSY QM, the fermion number n
F
≡ F is deﬁned
by
1
2
[1 − σ
3
] and we can represent (−1)
F
by the matrix σ
3
. If we write the
eigenstates of H as the vector:
ψ
n
(x) =
¸
ψ
(+)
n
(x)
ψ
(−)
n
(x)
(62)
then the ± corresponds to the eigenvalues of (−1)
F
being ±1. For our con
ventions the eigenvalue +1 corresponds to H
1
and the bosonic sector and the
eigenvalue −1 corresponds to H
2
and the fermionic sector. Since the bound
states of H
1
and H
2
are paired, except for the case of unbroken SUSY where
there is an extra state in the bosonic sector with E = 0 we expect for the
quantum mechanics situation that △ = 0 for broken SUSY and △ = 1 for
unbroken SUSY. In the general ﬁeld theory case, Witten gives arguments
that in general the index measures N
+
(E = 0) − N
−
(E = 0) which is the
diﬀerence △ between the number of Bose states and Fermi states of zero
energy. In ﬁeld theories the Witten index needs to be regulated to be well
deﬁned so that one considers instead
△(β) = Tr(−1)
F
e
−βH
, (63)
which for SUSY quantum mechanics becomes
△(β) = Tr[e
−βH
1
−e
−βH
2
]. (64)
In ﬁeld theory it is quite hard to determine if SUSY is broken non
perturbatively, and thus SUSY quantum mechanics became a testing ground
for ﬁnding diﬀerent methods to understand nonperturbative SUSY break
ing. In the quantum mechanics case, the breakdown of SUSY is related to
the question of whether there is a normalizable wave function solution to the
equation Q[0 >= 0[0 > which implies
ψ
0
(x) = Ne
−
W(x)dx
. (65)
As we said before, if this candidate ground state wave function does not
fall oﬀ fast enough at ±∞then Q does not annihilate the vacuum and SUSY
is spontaneously broken. Let us show using a trivial calculation that for
24
two simple polynomial potentials the Witten index does indeed provide the
correct answer to the question of SUSY breaking. Let us consider
△(β) = Trσ
3
[
dpdx
2π
]e
−β[p
2
/2+W
2
/2−σ
3
W
′
(x)/2
. (66)
Expanding the term proportional to σ
3
in the exponent and taking the trace
we obtain
△(β) =
[
dpdx
π
]e
−β[p
2
/2+W
2
/2
sinh(βW
′
(x)/2). (67)
We are interested in the regulated index as β tends to 0, so that practically
we need to evaluate
△(β) =
[
dpdx
2π
]e
−β[p
2
/2+W
2
/2]
(βW
′
(x)/2). (68)
If we directly evaluate this integral for any potential of the form W(x) =
gx
2n+1
, which leads to a normalizable ground state wave function, then all
the integrals are gamma functions and we explictly obtain △ = 1. If in
stead W(x) = gx
2n
so that the candidate ground state wave function is not
normalizable then the integrand becomes an odd function of x and therefore
vanishes. Thus we see for these simple cases that the Witten index immedi
ately coincides with the direct method available in the quantum mechanics
case.
Next let us discuss a favorite type of regularization scheme for ﬁeld
theory– namely the heat kernel method. (Later we will discuss a path integral
formulation for the regulated Witten Index).
Following Akhoury and Comtet [16] one deﬁnes the heat kernels K
±
(x, y; β)
which satisfy
(
d
dβ
−
d
2
dx
2
+ W
2
∓W
′
)K
±
= 0. (69)
These have the following eigenfunction representation:
K
±
(x, y; β) =< y[e
−βH
±
[x >
=
¸
n
e
−βEn
ψ
±
n
(x)ψ
±
n
(x) +
dEe
−βEn
ψ
±
E
(x)ψ
±
E
(x). (70)
In terms of the heat kernels one has
△(β) =
dx[K
+
(x, x; β) −K
−
(x, x; β)]
25
, or
△(β) = N
+
(E = 0) −N
−
(E = 0) +
∞
E
D
dEe
−βE
(ρ
+
(E) −ρ
−
(E)), (71)
where ρ
±
corresponds to the density of states. What Akhoury and Comtet
were able to show, was that in cases when W(x) went to diﬀerent constants at
plus and minus inﬁnity, then the density of states factors for the continuum
did not cancel and that △(β) could depend on β and be fractional at β = 0.
We refer the interested reader to the original paper for further details.
Another nonperturbative method for studying SUSY breaking in ﬁeld
theory is to explicitly break SUSY by placing the theory on a lattice and
either evaluating the path integral numerically or via some lattice non weak
perturbative method such as the strong coupling (or high temperature) ex
pansion. This method was studied in detail in [13, 21] and we will summarize
the results here. The basic idea is to introduce a new parameter, namely the
lattice spacing a. This parameter explicitly breaks SUSY so that the ground
state energy of the system will no longer be zero, even in the unbroken case.
One hopes that as the lattice spacing is taken to zero then the ground state
energy will go to zero as a power of the lattice spacing if SUSY is unbroken,
that is we expect
E
0
(a) = ca
γ
, (72)
where γ is a critical index which if greater than zero should be easy to
measure in a Monte Carlo calculation. Measuring the ground state energy
at two diﬀerent lattice spacings one studies:
γ = ln
E
0
(a
′
)
E
0
(a)
/ln
a
′
a
(73)
in the limit a
′
and a →0. The case of broken SUSY is more diﬃcult because
then we expect γ = 0 which is a hard measurement to make numerically.
In this latter case it is easier to directly measure the ground state energy
and show that it remains nonzero as one takes the lattice spacing to zero.
To see how this works in quantum mechanics one can do a lattice strong
coupling expansion of the Langevin equation which allows one to determine
the ground state wave function of H
1
as we shall show later.
For the superpotential W(x) = gx
3
, we expect to ﬁnd a positive critical
index since here the candidate ground state wave function is proportional to
26
e
−gx
4
/4
and is normalizable. The ground state expectation values of x
n
for
the Hamiltonian H
1
can be determined by ﬁrst solving the Langevin equation
dx
dt
+ W(x) = η(t) (74)
and then averaging x(η(t)) over Gaussian noise whose width is related to ¯ h.
Since by the virial theorem
E
0
= 3g
2
< x
6
> −3g < x
2
> ,
knowing the corelation functions will allow us to calculate the ground state
energy. We ﬁrst put the Langevin equation on a time lattice (t
n
= na):
ǫ(x
n
−x
n−1
) + gx
3
= η
n
, (75)
where ǫ = 1/a, which allows a solution by strong coupling expansion for large
g. The result is
x
n
= (
η
n
g
)
1/3
+
ǫ
3g
2/3
(η
1/3
n−1
η
−2/3
n
−η
−1/3
n
) + 0(ǫ
2
). (76)
As we will demonstrate in Sec. 8, the quantum mechanical expectation values
of < x
n
(t) > are the same as the noise averaged expection values of x
n
(η)
< x
n
(t) >=
[Dη]P[η]x
n
(η(t)). (77)
On the lattice the path integral becomes a product of ordinary integrals
which can be performed with:
P[η] = Π
i
e
−aη
2
(i)/2
a
2π
. (78)
The ground state energy on the lattice regulated theory then has the form
E
0
=
√
gz
m
¸
n=0
C
n
z
n
, (79)
where
z
3
=
2
a
√
g
27
is a dimensionless length. The critical index can be determined from the
logarithmic derivative of E
0
with respect to z. Using Pad´e approximants to
extrapolate the lattice series to small lattice spacing we found that [21]
E
0
= ca
1.16
(80)
verifying that SUSY is unbroken in the continuum limit. Using the same
methods for the case W(x) = gx
2
/2 we were able to verify that the ground
state energy was not zero as we took the continuum limit. After verifying the
applicability of this method in SUSY QM, Bender et al. then successfully
used this method to study nonperturbative SUSY breaking in WessZumino
models of ﬁeld theory [13].
3 Factorization and the Hierarchy of Hamil
tonians
In the previous section we found that once we know the ground state wave
function corresponding to a Hamiltonian H
1
we can ﬁnd the superpotential
W
1
(x) from eq. (6). The resulting operators A
1
and A
†
1
obtained from eq.
(4) can be used to factorize Hamiltonian H
1
. We also know that the ground
state wave function of the partner Hamiltonian H
2
i s determined from the
ﬁrst excited state of H
1
via the application of the operator A
1
. This allows a
refactorization of the second Hamiltonian in terms of W
2
. The partner of this
refactorization is now another Hamiltonian H
3
. Each of the new Hamiltoni
ans has one fewer bound state, so that this process can be continued until
the number of bound states is exhausted. Thus if one has an exactly solv
able potential problem for H
1
, one can solve for th e energy eigenvalues and
wave functions for the entire hierarchy of Hamiltonians created by repeated
refactorizations. Conversely if we know the ground state wave functions for
all the Hamiltonians in this hierarchy, we can reconstruct th e solutions of
the original problem. Let us now be more speciﬁc.
From the last section we have seen that if the ground state energy of a
Hamiltonian H
1
is zero then it can always be written in a factorizable form
as a product of a pair of linear diﬀerential operators. It is then clear that if
the ground state energy of a Hamiltonian H
1
is E
(1)
0
with eigenfunction ψ
(1)
0
then in view of eq. (3), it can always be written in the form (unless stated
28
otherwise, from now on we set ¯ h = 2m = 1 for simplicity):
H
1
= A
†
1
A
1
+ E
(1)
0
= −
d
2
dx
2
+ V
1
(x), (81)
where
A
1
=
d
dx
+ W
1
(x) , A
†
1
= −
d
dx
+ W
1
(x) , W
1
(x) = −
d lnψ
(1)
0
dx
. (82)
The SUSY partner Hamiltonian is then given by
H
2
= A
1
A
†
1
+ E
(1)
0
= −
d
2
dx
2
+ V
2
(x), (83)
where
V
2
(x) = W
2
1
+ W
′
1
+ E
(1)
0
= V
1
(x) + 2W
′
1
= V
1
(x) −2
d
2
dx
2
lnψ
(1)
0
. (84)
We will introduce the notation that in E
(m)
n
, n denotes the energy level
and (m) refers to the m’th Hamiltonian H
m
. From Sec. 2, the energy eigen
values and eigenfunctions of the two Hamiltonians H
1
and H
2
are related
by
E
(1)
n+1
= E
(2)
n
, ψ
(2)
n
= (E
(1)
n+1
−E
(1)
0
)
−1/2
A
1
ψ
(1)
n+1
. (85)
Now starting from H
2
whose ground state energy is E
(2)
0
= E
(1)
1
one can
similarly generate a third Hamiltonian H
3
as a SUSY partner of H
2
since we
can write H
2
in the form:
H
2
= A
1
A
†
1
+ E
(1)
0
= A
†
2
A
2
+ E
(1)
1
, (86)
where
A
2
=
d
dx
+ W
2
(x) , A
†
2
= −
d
dx
+ W
2
(x) , W
2
(x) = −
d lnψ
(2)
0
dx
. (87)
Continuing in this manner we obtain
H
3
= A
2
A
†
2
+ E
(1)
1
= −
d
2
dx
2
+ V
3
(x), (88)
29
where
V
3
(x) = W
2
2
+ W
′
2
+ E
(1)
1
= V
2
(x) −2
d
2
dx
2
lnψ
(2)
0
= V
1
(x) −2
d
2
dx
2
ln(ψ
(1)
0
ψ
(2)
0
). (89)
Furthermore
E
(3)
n
= E
(2)
n+1
= E
(1)
n+2
,
ψ
(3)
n
= (E
(2)
n+1
−E
(2)
0
)
−1/2
A
2
ψ
(2)
n+1
= (E
(1)
n+2
−E
(1)
1
)
−1/2
(E
(1)
n+2
−E
(1)
0
)
−1/2
A
2
A
1
ψ
(1)
n+2
. (90)
In this way, it is clear that if the original Hamiltonian H
1
has p(≥ 1)
bound states with eigenvalues E
(1)
n
, and eigenfunctions ψ
(1)
n
with 0 ≤ n ≤ (p−
1), then we can always generate a hierarchy of (p−1) Hamiltonians H
2
, ...H
p
such that the m’th member of the hierarchy of Hamiltonians (H
m
) has the
same eigenvalue spectrum as H
1
except that the ﬁrst (m−1) eigenvalues of
H
1
are missing in H
m
. In particular, we can always write (m = 2, 3, ...p):
H
m
= A
†
m
A
m
+ E
(1)
m−1
= −
d
2
dx
2
+ V
m
(x), (91)
where
A
m
=
d
dx
+ W
m
(x) , W
m
(x) = −
d lnψ
(m)
0
dx
. (92)
One also has
E
(m)
n
= E
(m−1)
n+1
= ... = E
(1)
n+m−1
,
ψ
(m)
n
= (E
(1)
n+m−1
−E
(1)
m−2
)
−1/2
...(E
(1)
n+m−1
−E
(1)
0
)
−1/2
A
m−1
...A
1
ψ
(1)
n+m−1
V
m
(x) = V
1
(x) −2
d
2
dx
2
ln(ψ
(1)
0
...ψ
(m−1)
0
) . (93)
In this way, knowing all the eigenvalues and eigenfunctions of H
1
we im
mediately know all the energy eigenvalues and eigenfunctions of the hierarchy
of p−1 Hamiltonians. Further the reﬂection and transmission coeﬃcients (or
phase shifts) for the hierarchy of Hamiltonians can be obtained in terms of
30
R
1
, T
1
of the ﬁrst Hamiltonian H
1
by a repeated use of eq. (30). In particular
we ﬁnd
R
m
(k) =
¸
W
(1)
−
−ik
W
(1)
−
+ ik
¸
...
¸
W
(m−1)
−
−ik
W
(m−1)
−
+ ik
¸
R
1
(k),
T
m
(k) =
¸
W
(1)
−
−ik
W
(1)
+
−ik
′
¸
...
¸
W
(m−1)
−
−ik
W
(m−1)
+
−ik
′
¸
T
1
(k), (94)
where k and k
′
are given by
k = [E −(W
(1)
−
)
2
]
1/2
, k
′
= [E −(W
(1)
+
)
2
]
1/2
. (95)
4 Shape Invariance and Solvable Potentials
Most text books on quantum mechanics describe how the one dimensional
harmonic oscillator problem can be elegantly solved using the raising and
lowering operator method. Using the ideas of SUSY QM developed in Sec. 2
and an integrability condition called the shape invariance condition [44], we
now show that the operator method for the harmonic oscillator can be gen
eralized to the whole class of shape invariant potentials (SIP) which include
all the popular, analytically solvable potentials. Indeed, we shall see that for
such potentials, the generalized operator method quickly yields all the bound
state energy eigenvalues, eigenfunctions as well as the scattering matrix. It
turns out that this approach is essentially equivalent to Schr¨odinger’s method
of factorization [49, 14] although the language of SUSY is more appealing.
Let us now explain precisely what one means by shape invariance. If the
pair of SUSY partner potentials V
1,2
(x) deﬁned in Sec. 2 are similar in shape
and diﬀer only in the parameters that appear in them, then they are said to
be shape invariant. More precisely, if the partner potentials V
1,2
(x; a
1
) satisfy
the condition
V
2
(x; a
1
) = V
1
(x; a
2
) + R(a
1
), (96)
where a
1
is a set of parameters, a
2
is a function of a
1
(say a
2
= f(a
1
)) and
the remainder R(a
1
) is independent of x, then V
1
(x; a
1
) and V
2
(x; a
1
) are said
to be shape invariant. The shape invariance condition (96) is an integrability
condition. Using this condition and the hierarchy of Hamiltonians discussed
31
in Sec. 3, one can easily obtain the energy eigenvalues and eigenfunctions of
any SIP when SUSY is unbroken.
4.1 General Formulas for Bound State Spectrum, Wave
Functions and SMatrix
Let us start from the SUSY partner Hamiltonians H
1
and H
2
whose eigenval
ues and eigenfunctions are related by SUSY. Further, since SUSY is unbroken
we know that
E
(1)
0
(a
1
) = 0, ψ
(1)
0
(x; a
1
) = N exp
¸
−
x
W
1
(y; a
1
)dy
. (97)
We now show that the entire spectrum of H
1
can be very easily obtained
algebraically by using the shape invariance condition (96). To that purpose,
let us construct a series of Hamiltonians H
s
, s = 1, 2, 3... In particular,
following the discussion of the last section it is clear that if H
1
has p bound
states then one can construct p such Hamiltonians H
1
, H
2
.....H
p
and the n’th
Hamiltonian H
n
will have the same spectrum as H
1
except that the ﬁrst n−1
levels of H
1
will be absent in H
n
. On repeatedly using the shape invariance
condition (96), it is then clear that
H
s
= −
d
2
dx
2
+ V
1
(x; a
s
) +
s−1
¸
k=1
R(a
k
), (98)
where a
s
= f
s−1
(a
1
) i.e. the function f applied s −1 times. Let us compare
the spectrum of H
s
and H
s+1
. In view of eqs. (96) and (98) we have
H
s+1
= −
d
2
dx
2
+ V
1
(x; a
s+1
) +
s
¸
k=1
R(a
k
)
= −
d
2
dx
2
+ V
2
(x; a
s
) +
s−1
¸
k=1
R(a
k
). (99)
Thus H
s
and H
s+1
are SUSY partner Hamiltonians and hence have identical
bound state spectra except for the ground state of H
s
whose energy is
E
(s)
0
=
s−1
¸
k=1
R(a
k
). (100)
32
This follows from eq. (98) and the fact that E
(1)
0
= 0. On going back from
H
s
to H
s−1
etc, we would eventually reach H
2
and H
1
whose ground state
energy is zero and whose n’th level is coincident with the ground state of the
Hamiltonian H
n
. Hence the complete eigenvalue spectrum of H
1
is given by
E
−
n
(a
1
) =
n
¸
k=1
R(a
k
); E
−
0
(a
1
) = 0. (101)
We now show that, similar to the case of the one dimensional harmonic
oscillator, the bound state wave functions ψ
(1)
n
(x; a
1
) for any shape invariant
potential can also be easily obtained from its ground state wave function
ψ
(1)
0
(x; a
1
) which in turn is known in terms of the superpotential. This is
possible because the operators A and A
†
link up the eigenfunctions of the
same energy for the SUSY partner Hamiltonians H
1,2
. Let us start from the
Hamiltonian H
s
as given by eq. (98) whose ground state eigenfunction is
then given by ψ
(1)
0
(x; a
s
). On going from H
s
to H
s−1
to H
2
to H
1
and using
eq. (14) we then ﬁnd that the n’th state unnormalized, energy eigenfunction
ψ
(1)
n
(x; a
1
) for the original Hamiltonian H
1
(x; a
1
) is given by
ψ
(1)
n
(x; a
1
) ∝ A
†
(x; a
1
)A
†
(x; a
2
)...A
†
(x; a
n
)ψ
(1)
0
(x; a
n+1
), (102)
which is clearly a generalization of the operator method of constructing the
energy eigenfunctions for the one dimensional harmonic oscillator.
It is often convenient to have explicit expressions for the wave functions.
In that case, instead of using the above equation, it is far simpler to use the
identify [46]
ψ
(1)
n
(x; a
1
) = A
†
(x; a
1
)ψ
(1)
n−1
(x; a
2
). (103)
Finally, it is worth noting that in view of the shape invariance condition
(96), the relation (30) between scattering amplitudes takes a particularly
simple form
R
1
(k; a
1
) =
W
−
(a
1
) + ik
W
−
(a
1
) −ik
R
1
(k; a
2
), (104)
T
1
(k; a
1
) =
W
+
(a
1
) −ik
′
W
−
(a
1
) −ik
T
1
(k; a
2
), (105)
thereby relating the reﬂection and transmission coeﬃcients of the same Hamil
tonian H
1
at a
1
and a
2
(= f(a
1
)).
33
4.2 Shape Invariance in More Than One Step
We can expand the list of solvable potentials by extending the shape invari
ance idea to the more general concept of shape invariance in two and even
multisteps. We shall see later that in this way we will be able to go much
beyond the factorization method and obtain a huge class of new solvable
potentials [59].
Consider the unbroken SUSY case of two superpotentials W(x; a
1
) and
˜
W(x; a
1
) such that V
2
(x; a
1
) and
˜
V
1
(x; a
1
) are same up to an additive constant
i.e.
V
2
(x; a
1
) =
˜
V
1
(x; a
1
) + R(a
1
) (106)
or equivalently
W
2
(x; a
1
) + W
′
(x; a
1
) =
˜
W
2
(x; a
1
) −
˜
W
′
(x; a
1
) + R(a
1
). (107)
Shape invariance in two steps means that
˜
V
2
(x; a
1
) = V
1
(x; a
2
) +
˜
R(a
1
), (108)
that is
˜
W
2
(x; a
1
) +
˜
W
′
(x; a
1
) = W
2
(x; a
2
) −W
′
(x; a
2
) +
˜
R(a
1
). (109)
We now show that when this condition holds, the energy eigenvalues and
eigenfunctions of the potential V
1
(x; a
1
) can be obtained algebraically. First
of all, let us notice that unbroken SUSY implies zero energy ground states
for the potentials V
1
(x; a
1
) and
˜
V
1
(x; a
1
):
E
(1)
0
(a
1
) = 0,
˜
E
(1)
0
(a
1
) = 0 . (110)
The degeneracy of the energy levels for the SUSY partner potentials yields
E
(2)
n
(a
1
) = E
(1)
n+1
(a
1
);
˜
E
(2)
n
(a
1
) =
˜
E
(1)
n+1
(a
1
). (111)
From eq. (106) it follows that
E
(2)
n
(a
1
) =
˜
E
(1)
n
(a
1
) + R(a
1
), (112)
so that for n = 0, these two equations yield
E
(1)
1
(a
1
) = R(a
1
). (113)
34
Also, the shape invariance condition (108) yields
˜
E
(2)
n
(a
1
) = E
(1)
n
(a
2
) +
˜
R(a
1
). (114)
From the above equations one can then show that
E
(1)
n+1
(a
1
) = E
(1)
n−1
(a
2
) + R(a
1
) +
˜
R(a
1
). (115)
On solving these questions recursively we obtain (n = 0, 1, 2, ...)
E
(1)
2n
=
n
¸
k=1
[R(a
k
) +
˜
R(a
k
)], (116)
E
(1)
2n+1
=
n
¸
k=1
[R(a
k
) +
˜
R(a
k
)] + R(a
n+1
). (117)
We now show that, similar to the discussion of the last subsection, the
bound state wave functions ψ
(1)
n
(x; a
1
) can also be easily obtained in terms
of the ground state wave functions ψ
(1)
0
(x; a
1
) and
˜
ψ
(1)
0
(x; a
1
) which in turn
are known in terms of the superpotentials W and
˜
W. In particular from eq.
(106) it follows that
ψ
(1)
n+1
(x; a
1
) ∝ A
†
(x; a
1
)ψ
(2)
n
(x; a
1
) ∝ A
†
(x; a
1
)
˜
ψ
(1)
n
(x; a
1
), (118)
while from eq. (108) we have
˜
ψ
(1)
n+1
(x; a
1
) ∝
˜
A
†
(x; a
1
)
˜
ψ
(2)
n
(x; a
1
) ∝
˜
A
†
(x; a
1
)ψ
(1)
n
(x; a
2
). (119)
Hence on combining the two equations we have the identity
ψ
(1)
n+2
(x; a
1
) ∝ A
†
(x; a
1
)
˜
A
†
(x; a
1
)ψ
(1)
n
(x; a
2
). (120)
Recursive application of the above identity yields
ψ
(1)
2n
(x; a
1
) ∝ [A
†
(x; a
1
)
˜
A
†
(x; a
1
)]...[A
†
(x; a
n
)
˜
A
†
(x; a
n
)]ψ
(1)
0
(x; a
n+1
), (121)
ψ
(1)
2n+1
(x; a
1
) ∝ [A
†
(x; a
1
)
˜
A
†
(x; a
1
)]...[A
†
(x; a
n
)
˜
A
†
(x; a
n
)]A
†
(x; a
n+1
)
˜
ψ
(1)
0
(x; a
n+1
),
(122)
35
where we have used the fact that
ψ
(1)
1
(x; a
1
) ∝ A
†
(x; a
1
)
˜
ψ
(1)
0
(x; a
1
). (123)
Finally, it is easily shown that the relation (30) between the scattering
amplitudes takes a particularly simple form
R
1
(k; a
1
) =
W
−
(a
1
) + ik
W
−
(a
1
) −ik
˜
W
−
(a
1
) + ik
˜
W
−
(a
1
) −ik
R
1
(k; a
2
), (124)
T
1
(k; a
1
) =
W
+
(a
1
) −ik
′
W
−
(a
1
) −ik
˜
W
+
(a
1
) −ik
′
˜
W
−
(a
1
) −ik
T
1
(k; a
2
), (125)
thereby relating the reﬂection and transmission coeﬃcients of the same Hamil
tonian at a
1
and a
2
.
It is clear that this procedure can be easily generalized and one can con
sider multistep shape invariant potentials and in these cases too the spec
trum, the eigenfunctions and the scattering matrix can be obtained alge
braically.
4.3 Strategies For Categorizing Shape Invariant Po
tentials
Let us now discuss the interesting question of the classiﬁcation of various
solutions to the shape invariance condition (96). This is clearly an important
problem because once such a classiﬁcation is available, then one discovers
new SIPs which are solvable by purely algebraic methods. Although the
general problem is still unsolved, two classes of solutions have been found and
discussed. In the ﬁrst class, the parameters a
1
and a
2
are related to each other
by translation (a
2
= a
1
+ α) [52, 53]. Remarkably enough, all well known
analytically solvable potentials found in most text books on nonrelativistic
quantum mechanic s belong to this class. Last year, a second class of solutions
was discovered in which the parameters a
1
and a
2
are related by scaling
(a
2
= qa
1
) [58, 59].
4.3.1 Solutions Involving Translation
We shall now point out the key steps that go into the classiﬁcation of SIPs
in case a
2
= a
1
+ α [52]. Firstly one notices the fact that the eigenvalue
36
spectrum of the Schr¨odinger equation is always such that the n’th eigenvalue
E
n
for large n obeys the constraint [133]
1/n
2
≤ E
n
≤ n
2
, (126)
where the upper bound is saturated by the square well potential and the
lower bound is saturated by the Coulomb potential. Thus, for any SIP, the
structure of E
n
for large n is expected to be of the form
E
n
∼
¸
α
C
α
n
α
, −2 ≤ α ≤ 2. (127)
Now, since for any SIP, E
n
is given by eq. (101), it follows that if
R(a
k
) ∼
¸
γ
k
γ
(128)
then
−3 ≤ γ ≤ 1 (129)
How does one implement this constraint on R(a
k
)? While one has no
rigorous answer to this question, it is easily seen that a fairly general factor
izable form of W(x; a
1
) which produces the above kdependence in R(a
k
) is
given by
W(x; a
1
) =
m
¸
i=1
(k
i
+ c
i
)g
i
(x) + h
i
(x)/(k
i
+ c
i
) + f
i
(x) (130)
where
a
1
= (k
1
, k
2
...), a
2
= (k
1
+ α, k
2
+ β...) (131)
with c
i
, α, β being constants. Note that this ansatz excludes all potentials
leading to E
n
which contain fractional powers of n. On using the above ansatz
for W in the shape invariance condition (96) one can obtain the conditions
to be satisﬁed by the functions g
i
(x), h
i
(x), f
i
(x). One important condition
is of course that only those superpotentials W are admissible which give a
square integrable ground state wave function. It turns out that there are no
solutions in case m ≥ 3 in eq. (130), while there are only two solutions in
case m = 2 i.e. when
W(x; a
1
) = (k
1
+ c
1
)g
1
(x) + (k
2
+ c
2
)g
2
(x) + f
1
(x), (132)
37
which are given by
W(r; A, B) = Atanhαr −Bcoth αr, A > B > 0, (133)
and
W(x; A, B) = Atanαx −Bcot αx; A, B > 0, (134)
where 0 ≤ x ≤ π/2α and ψ(x = 0) = ψ(x = π/2α) = 0. For the simplest
possibility of m = 1, one has a number of solutions to the shape invariance
condition (96). In Table 4.1, we give expressions for the various shape invari
ant potentials V
1
(x), superpotentials W(x), parameters a
1
and a
2
and the
corresponding energy eigenvalues E
(1)
n
[54, 55].
As an illustration, let us consider the superpotential given in eq. (134).
The corresponding partner potentials are
V
1
(x; A, B) = −(A + B)
2
+ A(A−α) sec
2
αx + B(B −α)cosec
2
αx
V
2
(x; A, B) = −(A + B)
2
+ A(A + α) sec
2
αx + B(B + α)cosec
2
αx (135)
V
1
and V
2
are often called P¨oschlTeller I potentials in the literature. They
are shape invariant partner potentials since
V
2
(x; A, B) = V
1
(x; A + α, B + α) + (A+ B + 2α)
2
−(A+ B)
2
(136)
and in this case
¦a
1
¦ = (A, B); ¦a
2
¦ = (A+α, B+α), R(a
1
) = (A+B+2α)
2
−(A+B)
2
. (137)
In view of eq. (101), the bound state energy eigenvalues of the potential
V
1
(x; A, B) are then given by
E
(1)
n
=
n
¸
k=1
R(a
k
) = (A+ B + 2nα)
2
−(A+ B)
2
. (138)
The ground state wave function of V
1
(x; A, B) is calculated from the super
potential W as given by eq. (134). We ﬁnd
ψ
(1)
0
(x; A, B) ∝ (cos αx)
s
(sin αx)
λ
(139)
where
s = A/α; λ = B/α. (140)
38
The requirement of A, B > 0 that we have assumed in eq. (134) guaran
tees that ψ
(1)
0
(x; A, B) is well behaved and hence acceptable as x −→0, π/2α.
Using this expression for the ground state wave function and eq. (103)
one can also obtain explicit expressions for the bound state eigenfunctions
ψ
(1)
n
(x; A, B). In particular, in this case, eq. (103) takes the form
ψ
n
(x; ¦a
1
¦) = (−
d
dx
+ Atanαx −Bcot αx)ψ
n−1
(x; ¦a
2
¦). (141)
On deﬁning a new variable
y = 1 −2 sin
2
αx (142)
and factoring out the ground state state wave function
ψ
n
(y; ¦a
1
¦) = ψ
0
(y; ¦a
1
¦)R
n
(y; ¦a
1
¦) (143)
with ψ
0
being given by eq. (139), we obtain
R
n
(y; A.B) = α(1 −y
2
)
d
dy
R
n−1
(y; A+ α, B + α)
+[(A−B) −(a + B + α)y]R
n−1
(y; A+ α, B + α).(144)
It is then clear [46] that R
n
(y; A, B) is proportional to the Jacobi poly
nomial so that the unnormalized bound state energy eigenfuncti ons for this
potential are
ψ
n
(y; A, B) = (1 −y)
λ/2
(1 +y)
s/2
P
λ−1/2,s−1/2
n
(y). (145)
The procedure outlined above has been applied to all known SIPs [46, 125]
and the energy eigenfunctions ψ
(1)
n
(y) have been obtained in Table 4.1, where
we also give the variable y for each case.
Several remarks are in order at this time.
1. The P¨oschlTeller I and II superpotentials as given by eqs. (134) and
(133) respectively have not been included in Table 4.1 since they are
equivalent to the Scarf I (trigonometric) and generalized P¨oschlTeller
superpotentials
W
1
= −Atan αx + Bsec αx,
W
2
= Acoth αr −Bcosech αr, (146)
39
by appropriate redeﬁnition of the parameters [95]. For example, one
can write
W
2
= (
A+ B
2
) tanh(
αr
2
) −(
B −A
2
) coth(
αr
2
), (147)
which is just the P¨oschlTeller II superpotential of eq. (133) with re
deﬁned parameters.
2. Throughout this section we have used the convention of ¯ h = 2m = 1. It
would naively appear that if we had not put ¯ h = 1, then the shape in
variant potentials as given in Table 4.1 would all be ¯ hdependent. How
ever, it is worth noting that in each and every case, the ¯ hdependence
is only in the constant multiplying the xdependent function so that in
each case we can always redeﬁne the constant multiplying the function
and obtain an ¯ hindependent potential. For example, corresponding
to the superpotential given by eq. (134), the ¯ hdependent potential is
given by (2m = 1)
V
1
(x; A, B) = W
2
−¯ hW
′
= −(A+ B)
2
+ A(A+ ¯ hα) sec
2
αx
+ B(B + ¯ hα)cosec
2
αx. (148)
On redeﬁning
A(A+ ¯ hα) = a; B(B + ¯ hα) = b, (149)
where a, b are ¯ hindependent parameters, we then have a ¯ hindependent
potential.
3. In Table 4.1, we have given conditions (like A > 0, B > 0) for the su
perpotential (134), so that ψ
(1)
0
= N exp(−
x
W(y)dy) is an acceptable
ground state energy eigenfunction. Instead one can also write down
conditions for ψ
(2)
0
= N exp(
x
W(y)dy) to be an acceptable ground
state energy eigenfunction.
4. It may be noted that the Coulomb as well as the harmonic oscillator
potentials in ndimensions are also shape invariant potentials.
40
5. Does this classiﬁcation exhaust all shape invariant potentials? It was
believed that the answer to the question is yes [85, 134] but as we shall
see in the next subsection, the answer to the question is in fact negative.
However, it appears that this classiﬁcation has perhaps exhausted all
SIPs where a
2
and a
1
are related by translation.
6. No new solutions (apart from those in Table 4.1) have been obtained
so far in the case of multistep shape invariance and when a
2
and a
1
are related by translation.
7. What we have shown here is that shape invariance is a suﬃcient con
dition for exact solvability. But is it also a necessary condition? This
question has been discussed in detail in ref. [52] where it has been
shown that the solvable Natanzon potentials [56, 57] are in general not
shape invariant.
Before ending this subsection, we would like to remark that for the SIPs
(with translation) given in Table 4.1, the reﬂection and transmission ampli
tudes R
1
(k) and T
1
(k) (or phase shift δ
1
(k) for the threedimensional case)
can also be calculated by operator methods. Let us ﬁrst notice that since for
all the cases a
2
= a
1
+α, hence R
1
(k; a
1
) and T
1
(k; a
1
) are determined for all
values of a
1
from eqs. (104) and (105) provided they are known in a ﬁnite
strip. For example, let us consider the shape invariant superpotential
W = ntanhx, (150)
where n is positive integer (1,2,3,...). The two partner potentials
V
1
(x; n) = n
2
−n(n + 1)sech
2
x,
V
2
(x; n) = n
2
−n(n −1)sech
2
x, (151)
are clearly shape invariant with
a
1
= n , a
2
= n −1 . (152)
On going from V
1
to V
2
to V
3
etc., we will ﬁnally reach the free particle
potential which is reﬂectionless and for which T = 1. Thus we immediately
41
conclude that the series of potentials V
1
, V
2
, ... are all reﬂectionless and the
transmission coeﬃcient of the reﬂectionless potential V
1
(x; n) is given by
T
1
(k, n) =
(n −ik)(n −1 −ik)...(1 −ik)
(−n −ik)(−n + 1 −ik)...(−1 −ik)
=
Γ(−n −ik)Γ(n + 1 −ik)
Γ(−ik)Γ(1 −ik).
(153)
The scattering amplitudes of the Coulomb [47] and the potential correspond
ing to W = Atanhx + Bsech x [48] have also been obtained in this way.
There is, however, a straightforward method [48] for calculating the scat
tering amplitudes by making use of the n’th state wave functions as given in
Table 4.1. In order to impose boundary conditions appropriate to the scat
tering problem, two modiﬁcations of the bound state wave functions have to
be made: (i) instead of the parameter n labelling the number of nodes, one
must use the wave number k
′
so that the asymptotic behaviour is exp(ik
′
x)
as x →∞. (ii) the second solution of the Schr¨odinger equation must be kept
(it had been discarded for bound state problems since it diverged asymptot
ically). In this way the scattering amplitude of all the SIPs of Table 4.1 have
been calculated in ref. [48].
4.3.2 Solutions Involving Scaling
For almost nine years, it was believed that the only shape invariant potentials
are those given in Table 4.1 and that there were no more shape invariant
potentials. However, very recently we have been able to discover a huge
class of new shape invariant potentials [58, 59]. It turns out that for many of
these new shape invariant potentials, the parameters a
2
and a
1
are related by
scaling (a
2
= qa
1
, 0 < q < 1) rather than by translation, a choice motivated
by the recent interest in qdeformed Lie algebras. We shall see that many
of these potentials are reﬂectionless and have an inﬁnite number of bound
states. So far none of these potentials have been obtained in a closed form
but are obtained only in a series form.
Let us consider an expansion of the superpotential of the from
W(x; a
1
) =
∞
¸
j=0
g
j
(x)a
j
1
(154)
42
and further let
a
2
= qa
1
, 0 < q < 1 . (155)
This is slightly misleading in that a reparameterization of the form a
2
= qa
1
,
can be recast as a
′
2
= a
′
1
+α merely by taking logarithms. However, since the
choice of parameter is usually an integral part of constructing a SIP, it is in
practice part of the ansatz. For example, we will construct below potentials
by expanding in a
1
, a procedure whose legitimacy and outcome are clearly
dependent on our choice of parameter and hence reparameterization. We
shall see that, even though the construction is noninvariant, the resulting
potentials will still be invariant under redeﬁnition of a
1
. On using eqs. (154)
and (155) in the shape invariance condition (96), writting R(a
1
) in the form
R(a
1
) =
∞
¸
j=0
R
j
a
j
1
, (156)
and equating powers of a
1
yields [59, 58]
2g
′
0
(x) = R
0
; g
′
1
(x) + 2d
1
g
0
(x)g
1
(x) = r
1
d
1
, (157)
g
′
n
(x) + 2d
n
g
0
(x)g
n
(x) = r
n
d
n
−d
n
n−1
¸
j=1
g
j
(x)g
n−j
(x), (158)
where
r
n
≡ R
n
/(1 −q
n
), d
n
= (1 −q
n
)/(1 +q
n
), n = 1, 2, 3, ... (159)
This set of linear diﬀerential equations is easily solvable in succession to give
a general solution of eq. (96). Let us ﬁrst consider the special case g
0
(x) = 0,
which corresponds to R
0
= 0. The general solution of eq. (158) then turns
out to be
g
n
(x) = d
n
dx[r
n
−
n−1
¸
j=1
g
j
(x)g
n−j
(x)], n = 1, 2, ... (160)
where without loss of generality we have assumed the constants of integration
to be zero. We thus see that once a set of r
n
are chosen, then the shape
invariance condition essentially ﬁxes the g
n
(x) (and hence W(x; a
1
)) and
determines the shape invariant potential. Implicit constraints on this choice
43
are that the resulting ground state wave function be normalizable and the
spectrum be sensibly ordered which is ensured if R(q
n
a
1
) > 0.
The simplest case is r
1
> 0 and r
n
= 0, n ≥ 2. In this case the eq. (160)
takes a particularly simple form
g
n
(x) = β
n
x
2n−1
, (161)
where
β
1
= d
1
r
1
, β
n
= −
d
n
(2n −1)
n−1
¸
j=1
β
j
β
n−j
(162)
and hence
W(x; a
1
) =
∞
¸
j=1
β
j
a
j
1
x
2j−1
=
√
a
1
F(
√
a
1
x) . (163)
For a
2
= qa
1
, this gives
W(x; a
2
) =
√
qW(
√
qx, a
1
) , (164)
which corresponds to the selfsimilar W of Shabat and Spiridonov [60, 61]. It
is worth pointing out that these selfsimilar potentials can be shown to satisfy
qsupersymmetry [62]. It may be noted here that instead of choosing r
n
=
0, n ≥ 2, if any one r
n
(say r
j
) is taken to be nonzero then one again obtains
selfsimilar potentials [59, 58] and in these instances the results obtained from
shape invariance and selfsimilarity are entirely equivalent and the Shabat
Spiridonov selfsimilarity condition turns out to be a special case of the shape
invariance condition.
It must be emphasized here that shape invariance is a much more general
concept than selfsimilarity. For example, if we choose more than one r
n
to
be nonzero, then SIP are obtained which are not contained within the self
similar ansatz. Consider for example, r
n
= 0, n ≥ 3. Using eq. (160) one can
readily calculate all the g
n
(x), of which the ﬁrst three are
g
1
(x) = d
1
r
1
x, g
2
(x) = d
2
r
2
x −
1
3
d
2
1
r
2
1
x
3
,
g
3
(x) = −
2
3
d
1
r
1
d
2
r
2
d
3
x
3
+
2
15
d
3
1
r
3
1
d
2
d
3
x
5
. (165)
Notice that in this case W(x) contains only odd powers of x. This makes
the potentials V
1,2
(x) symmetric in x and also guarantees unbroken SUSY.
44
The energy eigenvalues follow immediately from eqs. (101) and (156) and
are given by (0 < q < 1)
E
(1)
n
(a
1
) = Γ
1
(1 +q)(1 −q
n
)
(1 −q)
+ Γ
2
(1 +q
2
)(1 −q
2n
)
(1 −q
2
)
, n = 0, 1, 2, .. (166)
where Γ
1
= d
1
r
1
a
1
, Γ
2
= d
2
r
2
a
2
1
while the unnormalized ground state wave
function is
ψ
(1)
0
(x; a
1
) = exp[−
x
2
2
(Γ
1
+Γ
2
) +
x
4
4
(d
2
Γ
2
1
+2d
3
Γ
1
Γ
2
+d
4
Γ
2
2
) +0(x
6
)] (167)
The wave functions for the excited states can be recursively calculated from
the relation (103).
We can also calculate the transmission coeﬃcient of this symmetric po
tential (k = k
′
) by using the relation (125) and the fact that for this SIP
a
2
= qa
1
. Repeated application of the relation (125) gives
T
1
(k; a
1
) =
[ik −W(∞, a
1
)][ik −W(∞, a
2
)]...[ik −W(∞, a
n
)]
[ik + W(∞, a
1
)][ik + W(∞, a
2
)]...[ik + W(∞, a
n
)]
T
1
(k; a
n+1
)
(168)
where
W(∞, a
j
) =
E
(1)
∞ −E
(1)
j
. (169)
Now, as n → ∞, a
n+1
= q
n
a
1
→ 0(0 < q < 1) and, since we have taken
g
0
(x) = 0, one gets W(x; a
n+1
) → 0. This corresponds to a free particle
for which the reﬂection coeﬃcient R
1
(k; a
1
) vanishes and the transmission
coeﬃcient is given by
T
1
(k; a
1
) =
∞
¸
j=1
[ik −W(∞, a
j
)]
[ik + W(∞, a
j
)]
. (170)
The above discussion keeping only r
1
, r
2
= 0 can be readily generalized
to an arbitrary number of nonzero r
j
. The energy eigenvalues for this case
are given by (Γ
j
≡ d
j
r
j
a
j
1
)
E
(1)
n
(a
1
) =
¸
j
Γ
j
(1 +q
j
)(1 −q
jn
)
(1 −q
j
)
, n = 0, 1, 2, ... (171)
45
All these potentials are also symmetric and reﬂectionless with T
1
as given
by eq. (170). The limits q → 0 and q → 1 of all these potentials are simple
and quite interesting. At q = 1, the solution of the shape invariance condition
(96) is the standard one dimensional harmonic oscillator with W(x) = Rx/2
while in the limit q → 0 the solution is the RosenMorse superpotential
corresponding to the one soliton solution given by
W(x) =
√
Rtanh(
√
Rx). (172)
Hence the general solution as obtained above with 0 < q < 1 can be regarded
as the multiparameter deformation of the hyperbolic tangent function with q
acting as the deformation parameter. It is also worth noting that the number
of bound states increase discontinuously from just one at q = 0 to inﬁnity for
q > 0. Further, whereas for q = 1 the spectrum is purely discrete, for q even
slightly less than one, we have the discrete as well as the continuous spectra.
Finally, let us consider the solution to the shape invariance condition (96)
in the case when R
0
= 0. From eq. (157) it then follows that g
0
(x) = R
0
x/2
rather than being zero. One can again solve the set of linear diﬀerential
equations (157) and (158) in succession yielding g
1
(x), g
2
(x),...Further, the
spectrum can be immediately obtained by using eqs. (101) and (156). For
example, in the case of an arbitrary number of nonzero R
j
(in addition to
R
0
), it is given by
E
(1)
n
= nR
0
+
¸
j
Γ
j
(1 +q
j
)(1 −q
nj
)
(1 −q
j
)
(173)
which is the spectrum of a qdeformed harmonic oscillator [136, 137]. It is
worth pointing out here that, unlike the usual qoscillator where the space
is noncommutative but the potential is normal (ω
2
x
2
), in our approach the
space is commutative, but the potential is deformed, giving rise to a multi
parameter deformed oscillator spectrum.
An unfortunate feature of the new SIPs obtained above is that they are
not explicitly known in terms of elementary functions but only as a Taylor
series about x = 0. Questions about series convergence naturally arise. Nu
merical solutions pose no serious problems. As a consistency check, Barclay
et al. have checked numerically that the Schr¨odinger equation solved with
numerically obtained potentials indeed has the analytical energy eigenvalues
given above. From numerical calculations one ﬁnds that the superpotential
46
and the potential are as shown in Figs. 4.1 and 4.2 [59] corresponding to the
case when r
1
= 0, r
n
= 0, n ≥ 2. These authors, however, see no evidence
of the oscillations in W and V as reported by Degasperis and Shabat [63].
A very unusual new shape invariant potential has also been obtained [59]
corresponding to r
1
= 1, r
2
= −1, r
n
= 0, n ≥ 3 (with q = 0.3 and a = 0.75)
which is shown in Fig. 4.3. In this case, whereas V
1
(x) is a double well
potential, its shape invariant partner potential V
2
(x) is a single well.
It is worth pointing out that even though the potentials are not known in
a closed form in terms of elementary functions, the fact that these are reﬂec
tionless symmetric potentials can be used to constrain them quite strongly.
This is because, if we regard them as a solution of the KdV equation at
time t = 0, then being reﬂectionless, it is well known that such solutions
as t → ±∞ will break up into an inﬁnite number of solitons of the form
2k
2
i
sech
2
k
i
x [138, 139]. On using the fact that the KdV solitons obey an
inﬁnite number of conservation laws corresponding to mass, momentum, en
ergy ..., one can immediately obtain constraints on the reﬂectionless SIPs
obtained above [59].
4.3.3 Solutions of MultiStep Shape Invariance
Having obtained potentials which are multiparameter deformations of the
RosenMorse potential corresponding to the one soliton solution, an obvious
question to ask is if one can also obtain deformations of the multisoliton
solutions. The answer is yes [59] and as an illustration we now explicitly
obtain the multiparameter deformations of the two soliton case by using the
formalism of twostep shape invariance as developed earlier in this section.
Let us take the scaling ansatz a
2
= qa
1
and expand the two superpoten
tials W and
˜
W in powers of a
1
W(x; a
1
) =
∞
¸
j=0
g
j
(x)a
j
1
;
˜
W(x; a
1
) =
∞
¸
j=0
h
j
(x)a
j
1
(174)
Further, we write R and
˜
R in the form
R(a
1
) =
∞
¸
j=0
R
j
a
j
1
;
˜
R(a
1
) =
∞
¸
j=0
˜
R
j
a
j
1
(175)
47
Using these in eqs. (107) and (109) and equating powers of a
1
yields
(n=0,1,2,...)
g
′
n
+
n
¸
j=0
g
j
g
n−j
=
n
¸
j=0
h
j
h
n−j
−h
′
n
+ R
n
(176)
h
′
n
+
n
¸
j=0
h
j
h
n−j
= q
n
∞
¸
j=0
g
j
g
n−j
−q
n
g
′
n
+
˜
R
n
(177)
This set of linear equations is easily solved in succession. For example when
R
0
=
˜
R
0
= 0 (and hence g
0
(x) = h
0
(x) = 0) and further R
n
=
˜
R
n
= 0, n ≥ 3
one can readily calculate all the g
n
(x) and h
n
(x). The ﬁrst two of each are
g
1
=
(R
1
−
˜
R
1
)
(1 −q)
x,
g
2
=
(R
2
−
˜
R
2
)
(1 −q
2
)
x +
x
3
3(1 −q)
3
[(1 −q)(
˜
R
2
1
−R
2
1
) −2(1 +q)R
1
˜
R
1
],(178)
h
1
=
(
˜
R
1
−qR
1
)
(1 −q)
x; h
2
=
R
2
x
(1 −q
2
)
−
x
3
3(1 +q)(1 −q)
2
[(1 +q)
˜
R
2
1
+ (1 +q)(1 −q
2
)R
2
1
−2q(1 −q)R
1
˜
R
1
]
(179)
The energy eigenvalues which follow from eqs. (117) and (118) are [59]
E
(1)
2n
(a
1
) =
2
¸
j=1
(R
j
+
˜
R
j
)a
j
1
(1 −q
jn
)
(1 −q
j
)
E
(1)
2n+1
(a
1
) =
2
¸
j=1
R
j
a
j
1
(1 −q
j(n+1)
)
(1 −q
j
)
+
2
¸
j=1
˜
R
2
j
a
j
1
(1 −q
jn
)
(1 −q
j
)
(180)
For the special case of R
2
=
˜
R
2
= 0 the spectrum was obtained previously
by Spiridonov from the considerations of the twostep selfsimilar potentials
[135].
The limit q → 0 of the above equation is particularly simple and yields
the RosenMorse potential corresponding to the two soliton case i.e.
W = 2
˜
Rtanh
˜
Rx,
˜
W =
˜
Rtanh
˜
Rx (181)
48
provided R = 3
˜
R. This procedure can be immediately generalized and one
can consider shape invariance with a scaling ansatz in 3, 4, .....p steps and
thereby obtain multiparameter deformations of the 3, 4, ...p soliton Rosen
Morse potential.
4.3.4 Other Solutions
So far we have obtained solutions where a
2
and a
1
are related either by scaling
or by translation. Are there shape invariant potential where a
2
and a
1
are
neither related by scaling nor by translation? It turns out that there are
other possibilities for obtaining new shape invariant potentials. Some of the
other possibilities are: a
2
= qa
p
1
with p= 2,3,... and a
2
= qa
1
/(1 +pa
1
). Let
us ﬁrst consider the case when
a
2
= qa
2
1
(182)
i.e p = 2. Generalization to arbitrary p is straightforward [59]. On using eqs.
(155) and (156) one obtains the set of equations
g
′
2m
(x) +
2m
¸
j=0
g
j
(x)g
2m−j
(x) = q
m
m
¸
j=0
g
j
(x)g
m−j
(x) −q
m
g
′
m
(x) + R
2m
, (183)
g
′
2m+1
(x) +
2m+1
¸
j=0
g
j
(x)g
2m+1−j
(x) = R
2m+1
, (184)
which can be solved in succession and one can readily calculate all the g
n
(x).
For example, when only R
1
and R
2
are nonzero, the ﬁrst three g
′
s are
g
1
(x) = R
1
x, g
2
(x) = (R
2
−qR
1
)x −
1
3
R
3
1
x
3
,
g
3
(x) =
2
3
R
1
(qR
1
−R
2
)x
3
+
2
15
R
3
1
x
5
. (185)
The corresponding spectrum turns out to be (E
(1)
0
(a
1
) = 0)
E
(1)
n
=
R
1
q
n
¸
j=1
(a
1
q)
2
j−1
+
R
2
q
2
n
¸
j=1
(a
1
q)
2
j
, n = 1, 2, ... (186)
The q → 0 limit of these equations again correspond to the RosenMorse
potential corresponding to the one soliton solution. One can also consider
49
shape invariance in multisteps along with this ansatz thereby obtaining de
formations of the multisoliton RosenMorse potential.
One can similarly consider solutions to the shape invariance condition
(96) in case
a
1
=
qa
1
1 +pa
1
(187)
when 0 < q < 1 and pa
1
<< 1 so that one can expand (1 +pa
1
)
−1
in powers
of a
1
. For example, when only R
1
and R
2
are nonzero, then one can show
that the ﬁrst two nonzero g
n
are
g
1
(x) =
R
1
x
(1 +q)
, g
2
(x) = (R
2
+
pqR
1
(1 +q)
)x −
(1 −q)
(1 +q)
2
(1 +q
2
)
x
3
3
(188)
and the energy eigenvalue spectrum is (E
(1)
0
= 0)
E
(1)
n
= R
1
n
¸
j=1
q
j−1
a
1
[1 +pa
1
(
1−q
j−1
1−q
)]
+ R
2
n
¸
j=1
(q
j−1
a
1
)
2
[1 +pa
1
(
1−q
j−1
1−q
)]
2
(189)
Generalization to the case when several R
j
are nonzero as well as shape
invariance in multisteps is straight forward.
We would like to close this subsection with several comments.
1. Just as we have obtained qdeformations of the reﬂectionless Rosen
Morse and harmonic oscillator potentials, can one also obtain deforma
tions of the other SIPs given in Table 4.1?
2. Have we exhausted the list of SIP? We now have a signiﬁcantly ex
panded list but it is clear that the possibilities are far from exhausted.
In fact it appears that there are an unusually large number of shape
invariant potentials, for all of which the whole spectrum can be ob
tained algebraically. How does one classify all these potentials? Do
these potentials include all solvable potentials [52, 53]?
3. For those SIP where a
2
and a
1
are not related by translation, the spec
trum has so far only been obtained algebrically. Can one solve the
Schr¨odinger equation for these potentials directly?
50
4. There is a fundamental diﬀerence between those shape invariant poten
tials for which a
2
and a
1
are related by translation and other choices
(like a
2
= qa
1
). In particular, whereas in the former case the poten
tials are explicitly known in a closed form in terms of simple functions,
in the other cases they are only known formally as a Taylor series.
Secondly, whereas in the latter case, all the SIP obtained so far have
inﬁnite number of bound states and are either reﬂectionless (or have
no scattering), in the former case one has also many SIP with nonzero
reﬂection coeﬃcients.
4.4 Shape Invariance and Noncentral Solvable Poten
tials
We have seen that using the ideas of SUSY and shape invariance, a number
of potential problems can be solved algebraically. Most of these potentials
are either one dimensional or are central potentials which are again essen
tially one dimensional but on the half line. It may be worthwhile to enquire
if one can also algebraically solve some noncentral but separable potential
problems. As has been shown recently [131], the answer to the question is
yes. It turns out that the problem is algebraically solvable so long as the
separated problems for each of the coordinates belong to the class of SIP.
As an illustration, let us discuss noncentral separable potentials in spherical
polar coordinates.
In spherical polar coordinates the most general potential for which the
Schr¨odinger equation is separable is given by [140]
V (r, θ, φ) = V (r) +
V (θ)
r
2
+
V (φ)
r
2
sin
2
θ
(190)
where V (r), V (θ) and V (φ) are arbitrary functions of their argument. First,
let us see why the Schr¨odinger equation with a potential of the form given
by eq. (190) is separable in the (r, θ, φ) coordinates. The equation for the
wave function ψ(r, θ, φ) is
¸
−(
∂
2
ψ
∂
2
r
+
2
r
∂ψ
∂r
) −
1
r
2
(
∂
2
ψ
∂θ
2
+ cotθ
∂ψ
∂θ
) −
1
r
2
sin
2
θ
d
2
ψ
∂φ
2
= (E −V )ψ (191)
51
It is convenient to write ψ(r, θ, φ) as
ψ(r, θ, φ) =
R(r)
r
H(θ)
(sin θ)
1/2
K(φ). (192)
Substituting eq. (192) in eq. (191) and using the standard separation of
variables procedure, one obtains the following equations for the functions
K(φ), H(θ) and R(r) :
−
d
2
K
dφ
2
+ V (φ)K(φ) = m
2
K(φ), (193)
−
d
2
H
dθ
2
+ [V (θ) + (m
2
−
1
4
) cosec
2
θ] H(θ) = l
2
H(θ), (194)
−
d
2
R
dr
2
+ [V (r) +
(l
2
−
1
4
)
r
2
] R(r) = E R(r), (195)
where m
2
and l
2
are separation constants.
As has been shown in ref. [131], the three Schr¨odinger equations given by
(193), (194) and (195) may be solved algebraically by choosing appropriate
SIPs for V (φ), V (θ) and V (r). Details can be found in ref. [131].
Generalization of this technique to noncentral but separable potentials in
other orthogonal curvilinear coordinate systems as well as in other dimensions
is quite straightforward. Further, as we show below, one could use this trick
to discover a number of new exactly solvable threebody potentials in one
dimension.
4.5 Shape Invariance and 3Body Solvable Potentials
Many years ago, in a classic paper, Calogero [141] gave the complete solution
of the Schr¨odinger equation for three particles in one dimension, interacting
via twobody harmonic and inversesquare potentials. Later, Wolfes [142]
used Calogero’s method to obtain analytical solutions of the same problem
in the presence of an added threebody potential of a special form. Attention
then shifted to the exact solutions of the manybody problem and the general
question of integrability [143, 144, 145]. Recently, there has been renewed
interest in the onedimensional manybody problem in connection with the
physics of spin chains [146, 147, 148, 149]. Also, there has been a recent
generalization of Calogero’s potential for Nparticles to SUSY QM [150].
52
The purpose of this subsection is to show that using the results for SIPs
derived above, one can discover a number of new 3body potentials for which
the 3body problem in one dimension can be solved exactly [132]. There is
a rule of thumb that if one can solve a 3body problem then one can also
solve the corresponding nbody problem. Thus hopefully one can also solve
the corresponding statistical mechanics problem.
The important point to note is that three particles in one dimension,
after the center of mass motion is eliminated, have two degrees of freedom.
This may therefore be mapped on to a onebody problem in two dimensions.
Calogero [141] considered the case where the two dimensional potential is
noncentral but separable in polar coordinates r, φ. From the above discussion,
it is clear that if the potentials in each of the coordinates r and φ are chosen
to be shape invariant, then the whole problem can be solved exactly.
Calogero’s [141] solution of the threebody problem is for the potential
V
C
= ω
2
/8
¸
i<j
(x
i
−x
j
)
2
+ g
¸
i<j
(x
i
−x
j
)
−2
, (196)
where g > −1/2 to avoid a collapse of the system. Wolfes [142] showed that
a threebody potential
V
W
= f[(x
1
+ x
2
−2x
3
)
−2
+ (x
2
+ x
3
−2x
1
)
−2
+ (x
3
+ x
1
−2x
2
)
−2
] (197)
is also solvable when it is added to V
C
, with or without the pairwise cen
trifugal term. The last two terms on the righthand side of eq. (197) are
just cyclic permutations of the ﬁrst. Henceforth, such terms occuring in
any potential are referred to as “cyclic terms”. In this subsection, we give
more examples of threebody potentials that can be solved exactly. Our ﬁrst
solvable example is the threebody potential of the form
V
1
=
√
3f
1
2r
2
¸
(x
1
+ x
2
−2x
3
)
(x
1
−x
2
)
+ cyclic terms
¸
, (198)
which is added to the Calogero potential V
C
of eq. (196). In the above
equation,
r
2
=
1
3
[(x
1
−x
2
)
2
+ (x
2
−x
3
)
2
+ (x
3
−x
1
)
2
]. (199)
To see why potentials given by eqs. (196) to (198) are solvable, deﬁne the
Jacobi coordinates
R =
1
3
(x
1
+ x
2
+ x
3
), (200)
53
x =
(x
1
−x
2
)
√
2
, y =
(x
1
+ x
2
−2x
3
)
√
6
. (201)
After elimination of the center of mass part from the Hamiltonian, only
the x and ydegrees of freedom remain, which may be mapped into polar
coordinates
x = r sin φ , y = r cos φ . (202)
Obviously, the variables r, φ have ranges 0 ≤ r ≤ ∞ and 0 ≤ φ ≤ 2π. It is
straightforward to show that
(x
1
−x
2
) =
√
2 r sin φ,
(x
2
−x
3
) =
√
2 r sin(φ + 2π/3),
(x
3
−x
1
) =
√
2 r sin(φ + 4π/3). (203)
It turns out that V
C
, V
W
as well as V
1
are all noncentral but separable po
tentials in the polar coordinates r,φ. As a result, the Schr¨odinger equation
separates cleanly in the radial and angular variables, and the wave function
can be written as
ψ
nl
(r, φ) =
R
nl
(r)
√
r
F
l
(φ). (204)
In all three cases, the radial wave function obeys the equation
−
d
2
dr
2
+
3
8
ω
2
r
2
+
(B
2
l
−
1
4
)
r
2
¸
R
nl
= E
nl
R
nl
(r), (205)
where B
2
l
is the eigenvalue of the Schr¨odinger equation in the angular vari
able. Eq. (205) corresponds to a SIP and the eigenvalues E
nl
and the eigen
functions R
nl
which follow from Table 4.1 are
E
nl
=
3/2 ω(2n + B
l
+ 1) , n, l = 0, 1, 2... (206)
R
nl
= r
B
l
exp[
−1
4
(
3/2)ωr
2
]L
B
l
n
[
1
2
(
3/2)ωr
2
], (207)
where B
l
> 0. To examine the angular part of the eigenfunction F
l
(φ), take
the potential V
C
+ V
1
. Then, in the variable φ, the Schr¨odinger equation is
−
d
2
dφ
2
+
g
2
3
¸
m=1
cosec
2
[φ + 2(m−1)
π
3
] +
3
2
f
1
3
¸
m=1
cot[φ + 2(m−1)
π
3
]
¸
F
l
(φ)
54
= B
2
l
F
l
(φ). (208)
On using the identities
3
¸
m=1
cosec
2
[φ + 2(m−1)π/3] = 9cosec
2
3φ (209)
3
¸
m=1
cot[φ + 2(m−1)π/3] = 3cot 3φ (210)
eq. (208) reduces to
−
d
2
dφ
2
+
9
2
g cosec
2
3φ +
9
2
f
1
cot 3φ
¸
F
l
= B
2
l
F
l
(φ). (211)
Now this is again a SIP and hence its eigenvalues and eigenfunctions are
B
2
l
= 9(l + a + 1/2)
2
−
9
16
f
2
1
/(l + a + 1/2)
2
, (212)
F
l
(φ) = exp(−i
π
2
˜
l)(sin 3φ)
˜
l
exp[
3
4
f
1
φ
˜
l
]P
−
˜
l−
if
1
4
˜
l
,−
˜
l+
if
1
4
˜
l
l
(i cot 3φ) (213)
where P
α,β
n
is the Jacobi polynomial and
a = 1/2(1 + 2g)
1/2
. (214)
It must be noted that g > −1/2 for meaningful solutions. As expected, we
recover the results for the Calogero potential V
C
in the limit f
1
= 0.
For distinguishable particles, a given value of φ deﬁnes a speciﬁc order
ing. For 0 ≤ φ ≤ π/3, eq. (203) implies x
1
≥ x
2
≥ x
3
, and other ranges
of φ correspond to diﬀerent orderings [141, 142]. For singular repulsive
potentials, crossing is not allowed, and F
l
(φ) of eq. (213) is zero outside
0 ≤ φ ≤ π/3. Following Calogero, the wave function for the other ranges
may be constructed. Similarly, for indistiguishable particles, symmetrized or
antisymmetrized wave functions may be constructed.
Proceeding in the same way and using the results of Table 4.1, it is easily
show n that the other exactly solvable potentials are [132]
V
2
=
1
8
ω
2
¸
i<j
(x
i
−x
j
)
2
+ 3g [(x
1
+ x
2
−2x
3
)
−2
+ cyclic terms ]
−
3
√
3
2
f
1
r
2
¸
(x
1
−x
2
)
(x
1
+ x
2
−2x
3
)
+ cyclic terms
¸
, (215)
55
V
3
=
1
8
ω
2
¸
i<j
(x
i
−x
j
)
2
+g
¸
i<j
(x
i
−x
j
)
−2
−
f
3
√
6r
¸
(x
1
+ x
2
−2x
3
)
(x
1
−x
2
)
2
+ cyclic terms
¸
,
(216)
V
4
=
1
8
ω
2
¸
i<j
(x
i
−x
j
)
2
+ 3 g [ (x
1
+ x
2
−x
3
)
−2
+ cyclic terms ]
+
1
3
√
2
f
3
r
¸
(x
1
−x
2
)
(x
1
+ x
2
−2x
3
)
2
+ cyclic terms
¸
. (217)
Notice that in all these cases one has combined the SIP in the angular variable
with the harmonic conﬁnement. The three body scattering problem has also
been studied in these cases after droping the harmonic term [132]. The
possibility of replacing the harmonic conﬁnement term with the attractive
1
r

type interaction has also been considered. Note that in this case one has both
discrete and continuous spectra. Further, in this case one can obtain exact
solutions of threebody problems for all the SIPs discussed above (i.e.V
1
,...
V
4
, V
C
, V
W
, V
C
+ V
W
) along with the attractive
1
r
potential [132].
5 Operator Transforms – New Solvable Po
tentials from Old
In 1971, Natanzon [56] wrote down (what he thought at that time to be)
the most general solvable potentials i.e. for which the Schr¨odinger equation
can be reduced to either the hypergeometric or conﬂuent hypergeometric
equation . It turns out that most of these potentials are not shape invari
ant. Further, for most of them, the energy eigenvalues and eigenfunctions
are known implicitly rather than explicitly as in the shape invariant case.
One might ask if one can obtain these solutions from the explicitly solv
able shape invariant ones. One strategy for doing this is to start with a
Schr¨odinger equation which is exactly solvable (for example one having a
SIP) and to see what happens to this equation under a point canonical coor
dinate transformation. In order for the Schr¨odinger equation to be mapped
into another Schr¨odinger equation, there are severe restrictions on the nature
of the coordinate transformation. Coordinate transformations which satisfy
these restrictions give rise to new solvable problems. When the relationship
between coordinates is implicit, then the new solutions are only implicitly
56
determined, while if the relationship is explicit then the newly found solvable
potentials are also shape invariant [127, 128, 129]. In a more speciﬁc special
application of these ideas, Kostelecky et al. [130] were able to relate, using
an explicit coordinate transformation, the Coulomb problem in d dimensions
with the ddimensional harmonic oscillator. Other explicit applications of the
coordinate transformation idea are found in the review article of Haymaker
and Rau [120].
Let us see how this works. We start from the onedimensional Schr¨odinger
equation
−
d
2
dx
2
+ [V (x) −E
n
]
¸
ψ
n
(x) = 0. (218)
Consider the coordinate transformation from x to z deﬁned by
f(z) =
dz
dx
, (219)
so that
d
dx
= f
d
dz
. (220)
The ﬁrst step in obtaining a new Schr¨odinger equation is to change coordin
ates and divide by f
2
so that we have:
−
d
2
dz
2
−
f
′
f
d
dz
+
[V −E
n
]
f
2
¸
ψ
n
= 0 (221)
To eliminate the ﬁrst derivative term, one next rescales the wave function:
ψ = f
−1/2
¯
ψ. (222)
Adding a term ǫ
n
¯
ψ to both sides of the equation yields
−
d
2
¯
ψ
n
dz
2
+ [
¯
V (E
n
) + ǫ
n
]
¯
ψ
n
= ǫ
n
¯
ψ
n
, (223)
where
¯
V (E
n
) =
V −E
n
f
2
−
¸
f
′2
4f
2
−
f
′′
2f
¸
. (224)
In order for this to be a legitimate Schr¨odinger equation, the potential
¯
V (E
n
) + ǫ
n
must be independent of n. This can be achieved if the quantity
G deﬁned by
G = V −E
n
+ ǫ
n
f
2
(225)
57
is independent of n. How can one satisfy this condition? One way is to have
f
2
and G to have the same functional dependence on x(z) as the original
potential V . This further requires that in order for
¯
V to b e independent
of n, the parameters of V must change with n so that the wave function
corresponding to the n’th energy level of the new Hamiltoni an is related to
a wave function of the old Hamiltonian with parameters which depen d on
n. This can be made clear by a simple example.
Let us consider an exactly solvable problem  the three dimensional har
monic oscillator in a given angular momentum state with angular momentum
β. The reduced ground state wave function for that angular momentum is
ψ
0
(r) = r
β+1
e
−αr
2
/2
, (226)
so that the superpotential is given by
W(r) = αr −(β + 1)/r, (227)
and H
1
is given by:
H
1
= −
d
2
dr
2
+
β(β + 1)
r
2
+ α
2
r
2
−2α(β + 3/2). (228)
By our previous argument we must choose f =
dz
dr
to be of the form:
f
2
= (
dz
dr
)
2
=
A
r
2
+ Br
2
+ C. (229)
The solution of this equation gives z = z(r) which in general is not invertible
so that one knows r = r(z) only implicitly as discussed before. However for
special cases one has an invertible function. Let us for simplicity now choose
f = r, z = r
2
/2 . (230)
As discussed earlier, the energy eigenvalues of the three dimensional harmonic
oscillator are give by
E
n
= 4αn , (231)
so that the condition we want to satify is
V −4αn + ǫ
n
f
2
= G =
D
r
2
+ Er
2
+ F. (232)
58
Equating coeﬃcients we obtain
D = β(β + 1) ,
E = ǫ
n
+ α
2
= γ ,
F = −4αn −2α(β + 3/2) = −2Ze
2
. (233)
We see that for the quantities D, E, F to be independent of n one needs
to hav e that α, which describes the strength of the oscillator potential, be
dependent on n. Explicitly, solving the above three equations and choosing
β = 2l + 1/2, we obtain the relations:
α(l, n) =
Ze
2
2(l + n + 1)
, ǫ
n
= γ −
Z
2
e
4
4(l + 1 +n)
2
. (234)
We now choose γ = α
2
(l, n = 0) so that the ground state energy is zero.
These energy levels are those of the hydrogen atom. In fact, the new Hamil
tonian written in terms of z is now
¯
H = −
d
2
dz
2
+
l(l + 1)
z
2
−
Ze
2
z
+
Z
2
e
4
4(l + 1)
2
(235)
and the ground state wave function of the hydrogen atom is obtained from
the ground state wave function of the harmonic oscillator via
¯
ψ
0
= f
1/2
ψ
0
= x
l+1
e
−α(l,n=0)x
. (236)
Higher wave functions will have values of α which depend on n so that the
diﬀerent wave functions correspond to harmonic oscillator solutions with dif
ferent strengths.
All the exactly solvable shape invariant potentials of Table 4.1 can be
interrelated by point canonical coordinate transformations [128, 129]. This
is nicely illustrated in Fig. 5.1 . In general, r cannot be explicitly found in
terms of z, and one has
dz/dr = f =
A/r
2
+ Br
2
+ C , (237)
whose solution is:
59
z =
√
A+ Cr
2
+ Br
4
2
+
√
A log(r
2
)
2
−
√
A log(2 A+ Cr
2
+ 2
√
A
√
A + Cr
2
+ Br
4
)
2
+
C log(C + 2 Br
2
+ 2
√
B
√
A+ Cr
2
+ Br
4
)
4
√
B
. (238)
This clearly is not invertible in general. If we choose this general coordinate
transformation, then the potential that one obtains is the particular class
of Natanzon potentials whose wave functions are conﬂuent hypergeometric
functions in the variable r and are thus only implicitly known in terms of the
true coordinate z. In fact, even the expression for the transformed potential
is only known in terms of r:
¯
V (z, D, E) = 1/f
2
[D/r
2
+ Er
2
+ F −f
′2
/4 +ff
′′
/2] (239)
and thus only implicitly in terms of z. Equating coeﬃcients, we get an
implicit expression for the eigenvalues:
Cǫ
n
−F
2
√
E −Bǫ
n
−
D + 1/4 −Aǫ
n
= 2n + 1 (240)
as well as the state dependence on α and β necessary for the new Hamiltonian
to be energy independent:
α
n
=
E −Bǫ
n
, β
n
= −1/2 +
D + 1/4 −Aǫ
n
. (241)
5.1 Natanzon Potentials
The more general class of Natanzon potentials whose wave functions are
hypergeometric functions can be obtained by making an operator transfor
mation of the generalized P¨oschlTeller potential whose Hamiltonian is:
H
1
= A
†
1
A
1
= −
d
2
dr
2
+
β(β −1)
sinh
2
r
−
α(α + 1)
cosh
2
r
+ (α −β)
2
(242)
This corresponds to a superpotential
W = αtanh r −βcoth r (243)
60
and a ground state wave function given by:
ψ
0
= sinh
β
r cosh
−α
r . (244)
The energy eigenvalues were discussed earlier and are
E
n
= (α −β)
2
−(α −β −2n)
2
. (245)
The most general transformation of coordinates from r to z which pre
serves the Schr¨odinger equation is described by:
f
2
=
B
sinh
2
r
−
A
cosh
2
r
+ C = (
dz
dr
)
2
(246)
From this we obtain an explicit expression for z in terms of r.
z =
√
A tan
−1
¦
−3 A+ B −C + (A + B −C) cosh 2r
2
√
A
−2 A+ 2 B −C + 2 (A+ B) cosh 2r + C cosh
2
2r
¦
−
√
B log¦−A+ 3 B −C + (A+ B + C) cosh 2r
+ 2
√
B
−2 A + 2 B −C + 2 (A+ B) cosh 2r + C cosh
2
2r¦
+
√
C log¦A+ B + C cosh 2r
+
√
C
−2 A+ 2 B −C + 2 (A+ B) cosh 2r + C cosh
2
2r¦
+
√
B log¦2 sinh
2
r¦ (247)
However the expression for the transformed potential is only known in terms
of r:
¯
V (z, γ, δ) =
1
f
2
(
δ(δ −1)
sinh
2
r
−
γ(γ + 1)
cosh
2
r
+ σ −
f
′2
4f
2
+
f
′′
2f
) (248)
and thus only implicitly in terms of z. Equating coeﬃcients, we get an
implicit expression for the eigenvalues:
[(γ +1/2)
2
−Aǫ
n
]
1/2
−[(δ −1/2)
2
−Bǫ
n
]
1/2
−(σ −Cǫ
n
)
1/2
= 2n +1 (249)
as well as the state dependence on α and β necessary for the new Hamiltonian
to be energy independent:
α
n
= [(γ + 1/2)
2
−Aǫ
n
]
1/2
−1/2 ,
β
n
= [(δ −1/2)
2
−Bǫ
n
]
1/2
+ 1/2 . (250)
61
5.2 Generalizations of Ginocchio and Natanzon Poten
tials
In ref. [52] it was explicitly shown that the Ginocchio and Natanzon po
tentials, whose wave functions are hypergeometric functions of an implicitly
determined variable are not shape invariant and thus one could construct
new exactly solvable potentials from these using the factorization method.
It is convenient to use the original approach [57, 56] to deﬁne the Ginoc
chio and Natanzon potentials and then obtain their generalization to more
complicated solvable potentials whose wave functions are sums of hyperge
ometric functions. These generalized potentials have coordinates which are
only implicitly known. The operator which allows one to construct the eigen
functions of H
2
from those of H
1
converts single hypergeometric functions of
the implicitly known coordinate to sums of hypergeometric functions. This
process yields totally new solvable potentials. The resulting potentials are
ratios of polynomials in the transformed coordinate in which the wave func
tions are sums of hypergeometric functions. The transformed coordinate is
only implicitly known in terms of the coordinate system for the Schr¨odinger
equation.
The original method of obtaining the Natanzon potentials was to ﬁnd a
coordinate transformation that mapped the Schr¨odinger equation onto the
equation for the hypergeometric functions.
If we denote by r the coordinate appearing in the Schr¨odinger equation :
[−
d
2
dr
2
+ V (r)]ψ(r) = 0 (251)
where −∞ ≤ r ≤ ∞ and z the coordinate describing the hypergeometric
function F(α, β; γ; z) appearing in the diﬀerential equation:
z(1 −z)
d
2
F
dz
2
+ [γ −(α + β + 1)z]
dF
dz
−αβF = 0, (252)
where 0 ≤ z ≤ 1, then the transformation of coordinates is given by [56]
dz
dr
=
2z(1 −z)
R
1/2
, R = az
2
+ (c
1
−c
0
−a)z + c
0
(253)
In terms of the coordinate z the most general potential V (r) which cor
respon ds to a Schr¨odinger equation getting mapped into a hypergeometric
62
function equation is given by:
V (r) = [fz(z −1) +h
0
(1 −z) + h
1
z + 1]/R+
¦a +
a + (c
1
−c
0
)(2z −1)
z(z −1)
−
5△
4R
¦
z
2
(1 −z)
2
R
2
(254)
where
△ = (a −c
0
−c
1
)
2
−4c
0
c
1
.
We note that z is only implicitly known in terms of r. The potential V (r)
is a function of six dimensionless parameters f, h
0
, h
1
, a, c
0
and c
1
. These six
parameters will be related to the energy ǫ
n
and the parameters (α, β, γ) of
the hypergeometric function below.
The class of potentials called the Ginocchio potentials is a subclass of the
Natanzon class which has only two independent variables nu and λ where
c
0
= 0, c
1
= 1/λ
4
, a = c
1
−1/λ
2
, h
0
= −3/4, h
1
= −1, and f = (ν+1/2)
2
−1.
The Ginocchio potentials are more easily discussed, however by con
sidering them as derived from coordinate transformations which map the
Schr¨odinger equation into the diﬀerential equation for the Gegenbauer poly
nomials.
The transformation is given by
dy
dr
= (1 −y
2
)[1 + (λ
2
−1)y
2
] (255)
where now −1 ≤ y ≤ 1 and the potential is given by:
V (r) = ¦−λ
2
ν(ν +1)+
1
4
(1−λ
2
)[5(1−λ
2
)y
4
−(7−λ
2
)y
2
+2]¦(1−y
2
). (256)
Although y is not known explictly in terms of r, it is implicitly known via:
rλ
2
= tanh
−1
(y) −(1 −λ
2
)
1/2
tanh
−1
[(1 −λ
2
)
1/2
y], λ < 1,
rλ
2
= tanh
−1
(y) + (λ
2
−1)
1/2
tan
−1
[(λ
2
−1)
1/2
y], λ > 1. (257)
By introducing the variables: x = λy/[g(y)]
1/2
where
g(y) = 1 + (λ
2
−1)y
2
,
and changing the variable x to an angle x = cos θ, one obtains from the
Schr¨odinger equation, the equation for the Gegenbauer polynomials:
63
[
d
2
dθ
2
+ (n + µ + 1/2)
2
−
µ
2
−1/4
sin
2
θ
]Z
1/2
ψ
n
= 0 (258)
where
Z =
[1 −y
2
[
1/2
λ
ǫ
n
= −µ
2
n
λ
4
. (259)
The corresponding energy eigenvalues and eigenfuctions are
µ
n
λ
2
= [λ
2
(ν + 1/2)
2
+ (1 −λ
2
)(n + 1/2)
2
]
1/2
−(n + 1/2), (260)
ψ
n
= (1 −y
2
)
µn/2
[g(y)]
−(2µn+1)/4
C
(µn+1/2)
n
(x). (261)
Thus the superpotential is :
W(r) = −
ψ
′
(r)
ψ(r)
=
1
2
(1 −λ
2
)y(y
2
−1) +µ
0
yλ
2
(262)
For the Natanzon potential (254) we have that the energy eigenvalues are
given by:
(2n + 1) = (1 −aǫ
n
+ f)
1/2
−(1 −c
0
ǫ
n
+ h
0
)
1/2
− (1 −c
1
ǫ
n
+ h
1
)
1/2
≡ α
n
−β
n
−δ
n
. (263)
and the corresponding (unnormalized) energy eigenfunctions are given by :
ψ
n
= R
1/4
z
βn/2
(1 −z)
δn/2
F(−n, α
n
−n; 1 +β
n
; z). (264)
It is easy to determine the superpotential from the ground state wave
function:
W(r) = [δ
0
z −(1 +β
0
)(1 −z)]/R
1/2
+ [(c
1
−c
0
−a)z + 2c
0
](1 −z)/(2R
3/2
). (265)
Once we have the superpotential and the explicit expression for the wave
functions we can determine all the wave functions of the partner potential
using the operator:
64
A
1
=
d
dr
+ W(r) =
dz
dr
d
dz
+ W(z) =
2z(1 −z)
R
1/2
d
dz
+ W(z)
for the case of the Natanzon potential and the operator:
A
1
=
d
dr
+ W(r) =
dy
dr
d
dy
+ W(y) = (1 −y
2
)[1 + (λ
2
−1)y
2
]
d
dy
+ W(y)
for the case of the Ginocchio potentials. In general, these operators take a
single hypergeometric function into a sum of two hypergeometric functions
which cannot be reexpressed as a single hypergeometric function except in
degenerate cases where one obtains the shape invariant special cases discussed
earlier. Explicitly, one has for the unnormalized eigenfunctions of H
2
, the
partner to the origina l Natanzon Hamiltonian:
ψ
(2)
n−1
= R
−1/4
z
βn/2
(1−z)
δn/2
[(1−z)[β
n
−β
0
−z(δ
n
−δ
0
)]F(−n, α
n
−n; 1+β
n
; z)
−2z(1 −z)n(n + 1 +β
n
+ δ
n
)F(−n + 1, α
n
−n + 1; 2 +β
n
; z)] (266)
and for the partner potential V
2
= W
2
+ W
′
one has
V
2
= ǫ
0
+[(β
0
+δ
0
)(β
0
+δ
0
+2)z(z −1) +(β
2
0
−1)(1 −z) +(δ
2
0
−1)z +1]/R
+[a −[c
1
(3β
0
+ δ
0
) + c
0
(β
0
+ 3δ
0
)a(1 −β
0
−δ
0
)]/[z(z −1)]
−(2z −1)[(c
1
−c
0
)(β
0
+ δ
0
+ 1) +a(β
0
−δ
0
)]/[z(z −1)]
+ 7△/(4R)]z
2
(1 −z
2
)/R
2
. (267)
By using the hierarchy of Hamiltonians one can construct in the usual manner
more and more complicated potentials which are ratios of higher and higher
order polynomials in z (as well as R) which are isospectral to the original
Natanzon potential except for the usual missing states. The wave functions
will be sums of hypergeometric functions. The arguments are identical for the
Ginocchio class with the hypergeometric functions now being restricted to
being Gegenbauer polynomials. The potentials one obtains can have multiple
local minima and several of these are displayed in [52]. In [52] it was also
shown that the ser ies of Hamiltonians and SUSY charges form the graded
Lie algebra sl(1/1) ⊗ SU(2). However this algebra did not lead to any new
insights. We shall also see in Sec.13 that the series of Hamiltonians form a
parasupersymmetry of order p if the original Hamiltonian has p bound states.
65
6 Supersymmetric WKB Approximation
The semiclassical WKB method [151] is one of the most useful approxima
tions for computing the energy eigenvalues of the Schr¨odinger equation. It
has a wider range of applicability than standard perturbation theory which is
restricted to perturbing potentials with small coupling constants. The pur
pose of this section is to describe and give applications of the supersymmetric
WKB method (henceforth called SWKB) [79, 80] which has been inspired by
supersymmetric quantum mechanics.
6.1 SWKB Quantization Condition for Unbroken Su
persymmetry
As we have seen in previous sections, for quantum mechanical problems, the
main implication of SUSY is that it relates the energy eigenvalues, eigenfunc
tions and phase shifts of two supersymmetric partner potentials V
1
(x) and
V
2
(x). Combining the ideas of SUSY with the lowest order WKB method,
Comtet, Bandrauk and Campbell [79] obtained the lowest order SWKB quan
tization condition in case SUSY is unbroken and showed that it yields energy
eigenvalues which are not only exact for large quantum numbers n (as any
WKB approximation scheme should in the semiclassical limit) but which are
also exact for the ground state (n = 0). We shall now show this in detail.
In lowest order, the WKB quantization condition for the one dimensional
potential V (x) is [151]
x
2
x
1
2m[E
n
−V (x)] dx = (n + 1/2)¯ hπ, n = 0, 1, 2, ... , (268)
where x
1
and x
2
are the classical turning points deﬁned by E
n
= V (x
1
) =
V (x
2
). For the potential V
1
(x) corresponding to the superpotential W(x),
the quantization condition (268) takes the form
x
2
x
1
2m[E
(1)
n −W
2
(x) +
¯ h
√
2m
W
′
(x)] dx = (n + 1/2)¯ hπ. (269)
Let us assume that the superpotential W(x) is formally O(¯ h
0
). Then, the
W
′
term is clearly O(¯ h). Therefore, expanding the left hand side in powers
66
of ¯ h gives
b
a
2m[E
(1)
n −W
2
(x)] dx+
¯ h
2
b
a
W
′
(x) dx
E
(1)
n −W
2
(x)
+... = (n+1/2)¯ hπ, (270)
where a and b are the turning points deﬁned by E
(1)
n
= W
2
(a) = W
2
(b). The
O(¯ h) term in eq. (270) can be integrated easily to yield
¯ h
2
sin
−1
W(x)
E
(1)
n
¸
¸
b
a
. (271)
In the case of unbroken SUSY, the superpotential W(x) has opposite signs
at the two turning points, that is
−W(a) = W(b) =
E
(1)
n . (272)
For this case, the O(¯ h) term in (271) exactly gives ¯ hπ/2, so that to leading
order in ¯ h the SWKB quantization condition when SUSY is unbroken is
[79, 80]
b
a
2m[E
(1)
n −W
2
(x)] dx = n¯ hπ, n = 0, 1, 2, ... (273)
Proceeding in the same way, the SWKB quantization condition for the po
tential V
2
(x) turns out to be
b
a
2m[E
(2)
n −W
2
(x)] dx = (n + 1)¯ hπ, n = 0, 1, 2, ... (274)
Some remarks are in order at this stage.
(i) For n = 0, the turning points a and b in eq. (273) are coincident and
E
(1)
0
= 0. Hence SWKB is exact by construction for the ground state energy
of the Hamiltonian H
1
= (−¯ h
2
/2m)d
2
/dx
2
+ V
1
(x).
(ii) On comparing eqs. (273) and (274), it follows that the lowest order
SWKB quantization condition preserves the SUSY level degeneracy i.e. the
approximate energy eigenvalues computed from the SWKB quantization con
ditions for V
1
(x) and V
2
(x) satisfy the exact degeneracy relation E
(1)
n+1
= E
(2)
n
.
(iii) Since the lowest order SWKB approximation is not only exact as
expected for large n, but is also exact by construction for n = 0, hence,
unlike the ordinary WKB approach, the SWKB eigenvalues are constrained
to be accurate at both ends of allowed values of n at least when the spectrum
is purely discrete. One can thus reasonably expect better results than the
WKB scheme.
67
6.2 Exactness of the SWKB Condition for Shape In
variant Potentials
How good is the SWKB quantization condition [eq. (273)]? To study this
question, the obvious ﬁrst attempts consisted of obtaining the SWKB bound
state spectra of several analytically solvable potentials like Coulomb, har
monic oscillator, Morse, etc. [79, 80]. In fact, it was soon shown that the
lowest order SWKB condition gives the exact eigenvalues for all SIPs [45]!
The proof of this statement follows from the facts that the SWKB con
dition preserves the level degeneracy and a vanishing ground state energy
eigenvalue. For the hierarchy of Hamiltonians H
(s)
discussed in Sec. 3, the
SWKB quantization condition takes the form
2m[E
(s)
n −
s
¸
k=1
R(a
k
) −W
2
(a
s
; x)] dx = n¯ hπ . (275)
Now, since the SWKB quantization condition is exact for the ground state
energy when SUSY is unbroken, hence
E
(s)
0
=
s
¸
k=1
R(a
k
) (276)
as given by eq. (275), must be exact for Hamiltonian H
(s)
. One can now go
back in sequential manner from H
(s)
to H
(s−1)
to H
(2)
and H
(1)
and use the
fact that the SWKB method preserves the level degeneracy E
(1)
n+1
= E
(2)
n
. On
using this relation n times , we ﬁnd that for all SIPs, the lowest order SWKB
condition gives the exact energy eigenvalues [45]. This is a very substantial
improvement over the usual WKB formula eq. (268) which is not exact
for most SIPs. Of course, one can artiﬁcially restore exactness by ad hoc
Langerlike corrections [152]. However, such modiﬁcations are unmotivated
and have diﬀerent forms for diﬀerent potentials. Besides, even with such
corrections, the higher order WKB contributions are nonzero for most of
these potentials[152, 153].
What about higher order SWKB contributions? Since the lowest order
SWKB energies are exact for shape invariant potentials, it would be nice to
check that higher order corrections vanish order by order in ¯ h. By starting
from the higher order WKB formalism, one can readily develop the higher
order SWKB formalism [83]. It has been explicitly checked for all known
68
SIPs that up to O(¯ h
6
) there are indeed no corrections.. This result can be
extended to all orders in ¯ h [84, 85]. Conditions on the superpotential which
ensure exactness of the lowest order SWKB condition have been given in ref.
[85].
It has been proved above that the lowest order SWKB approximation
reproduces the exact bound state spectrum of any SIP. This statement has
indeed been explicitly checked for all SIPs known until last year i.e. solutions
of the shape invariance condition involving a translation of parameters a
2
=
a
1
+ constant. However, it has recently been shown [154] that the above
statement is not true for the newly discovered class [58, 59, 60, 135] of SIPs
discussed in Sec. 4.2.2, for which the parameters a
2
and a
1
are related by
scaling a
2
= qa
1
. What is the special feature of these new potentials that
interferes with the proof that eq. (273) is exact for SIPs ? To understand this,
let us look again at the derivation of the lowest order SWKB quantization
condition. In the derivation, W
2
is taken as O(¯ h
0
) while ¯ hW
′
is O(¯ h) and
hence one can expand the integrand on the left hand side in powers of ¯ h. This
assumption is justiﬁed for all the standard SIPs [14, 54, 55] since W
2
is indeed
of O(¯ h
0
) while ¯ hW
′
is indeed of O(¯ h). One might object to this procedure
since the resulting potential V
1
is then ¯ hdependent. However, in all cases,
this ¯ hdependence can be absorbed into some dimensionful parameters in the
problem. For example, consider
W(x) = Atanhx , V
1
(x) = A
2
−A(A+ ¯ h)sech
2
x. (277)
Taking A such that A(A + ¯ h) is independent of ¯ h gives the desired ¯ h
independent potential (the additive constant is irrelevant and so can contain
¯ h). Such a move may appear to be of limited value since one cannot apply
SWKB directly to a superpotential W which is now ¯ hdependent. However,
because A is a free parameter, one can continue the SWKB results obtained
for A (and hence W) independent of ¯ h over to this superpotential and so
obtain an SWKB approximation for a ¯ hindependent potential.
What about the new potentials ? In the simplest of these cases, the
only free parameter in the problem (apart from q) is the combination R
1
a
1
,
on which W depends as W(x, R
1
a
1
) =
√
R
1
a
1
F(
√
R
1
a
1
x/¯ h). Incorporating
diﬀerent dependences on ¯ h in R
1
a
1
will give diﬀerent ones in W, V
1
and E
n
,
but F is a suﬃciently complicated function that there is no way of eliminating
¯ h from W
2
. This is a direct consequence of the scaling reparameterisation
69
a
2
= qa
1
not involving ¯ h. If W
2
(x; a
1
) were independent of ¯ h, so would
W
2
(x; a
2
) be and in taking the lowest order of the shape invariance condition
one would get W
2
(x; a
1
) = W
2
(x; a
2
), which corresponds to the harmonic
oscillator. Furthermore, with a
2
= qa
1
, W
2
and ¯ hW
′
are now of a similar
order in ¯ h. The basic distinction between them involved in deriving eq.
(273) is thus no longer valid and we are prevented from deriving the SWKB
condition for these new potentials.
We thus see that the SWKB quantization condition is not the correct
lowest order formula in the case of the new SIPs and hence it is not re
ally surprising that eq. (273) does not give the exact eigenvalues for these
potentials. In other words, it remains true that the lowest order SWKB
quantization condition is exact for SIPs (if the SUSY is unbroken), but only
in those cases for which the formula is applicable in the ﬁrst place. It is
thus still the case that the SIPs given in refs. [54, 55] are the only known
ones for which the lowest order SWKB formula is exact and the higher order
corrections are all zero.
6.3 Comparison of the SWKB and WKB Approaches
Let us now compare the merits of the two schemes [WKB and SWKB]. For
potentials for which the ground state wave function (and hence the super
potential W) is not known, clearly the WKB approach is preferable, since
one cannot directly make use of the SWKB quantization condition eq. (273).
On the other hand, we have already seen that for shape invariant potentials,
SWKB is clearly superior. An obvious interesting question is to compare
WKB and SWKB for potentials which are not shape invariant but for whom
the ground state wave function is known. One choice which readily springs
to mind is the Ginocchio potential given by [57]
V (x) = (1 −y
2
)
−λ
2
ν(ν + 1) +
(1 −λ
2
)
4
[2 −(7 −λ
2
)y
2
+ 5(1 −λ
2
)y
4
]
¸
(278)
where y is related to the independent variable x by
dy
dx
= (1 −y
2
)[1 −(1 −λ
2
)y
2
] . (279)
70
Here the parameters ν and λ measure the depth and shape of the potential
respectively. The corresponding superpotential is [52]
W(x) = (1 −λ
2
)y(y
2
−1)/2 +µ
0
λ
2
y (280)
where µ
n
is given by [57]
µ
n
λ
2
=
[λ
2
(ν + 1/2)
2
+ (1 −λ
2
)(n + 1/2)
2
] −(n + 1/2) (281)
and the bound state energies are
E
n
= −µ
2
n
λ
4
, n = 0, 1, 2, ... (282)
For the special case λ = 1, one has the RosenMorse potential, which is shape
invariant. The spectra of the Ginocchio potential using both the WKB and
SWKB quantization conditions have been computed [90]. The results are
shown in Table 6.1. In general, neither semiclassical method gives the exact
energy spectrum. The only exception is the shape invariant limit λ = 1,
in which case the SWKB results are exact, as expected. Also, for n = 0, 1
the SWKB values are consistently better, but there is no clear cut indica
tion that SWKB results are always better. This example, as well as other
potentials studied in ref. [86, 87, 90, 88, 89] , support the conjecture that
shape invariance is perhaps a necessary condition for the lowest order SWKB
approximation to yield the exact spectrum [90].
So far, we have concentrated our attention on the energy eigenvalues.
For completeness, let us note that several authors have also obtained the
wave functions in the SWKB approximation [79, 91, 92, 93]. As in the WKB
method, the SWKB wave functions diverge at the turning points a and b.
These divergences can be regularized either by the uniform approximation
[91, 155] or by appropriate retention of higher orders in ¯ h [92, 93].
6.4 SWKB Quantization Condition for Broken Super
symmetry
The derivation of the lowest order SWKB quantization condition for the case
of unbroken SUSY is given in Sec. 6.1 [eqs. (268) to (273)]. For the case of
broken SUSY, the same derivation applies until one examines the O(¯ h) term
in eq. (271). Here, for broken SUSY, one has
W(a) = W(b) =
E
(1)
n (283)
71
and the O(¯ h) term in (271) exactly vanishes. So, to leading order in ¯ h the
SWKB quantization condition for broken SUSY is [94, 96]
b
a
2m[E
(1)
n −W
2
(x)] dx = (n + 1/2)¯ hπ, n = 0, 1, 2, ... (284)
As before, it is easy to obtain the quantization condition which includes
higher orders in ¯ h [96] and to test how well the lowest order broken SWKB
condition works for various speciﬁc examples. As for the case of unbroken
SUSY, it is found that exact spectra are obtained for shape invariant po
tentials with broken SUSY [95]. For potentials which are not analytically
solvable, the results using eq. (284) are usually better than standard WKB
computations. Further discussion can be found in ref. [96].
7 Isospectral Hamiltonians
In this section, we will describe how one can start from any given one
dimensional potential V
1
(x) with n bound states, and use supersymmetric
quantum mechanics to construct an nparameter family of strictly isospec
tral potentials V
1
(λ
1
, λ
2
, . . . , λ
n
; x) i.e., potentials with eigenvalues, reﬂection
and transmission coeﬃcients identical to those for V
1
(x). The fact that such
families exist has been known for a long time from the inverse scattering ap
proach [67], but the GelfandLevitan approach to ﬁnding them is technically
much more complicated than the supersymmetry approach described here.
Indeed, the advent of SUSY QM has produced a revival of interest in the de
termination of isospectral potentials [65, 68, 69, 156, 157, 158] [66, 70, 72, 75].
In Sec. 7.1 we describe how a one parameter isospectral family is obtained
by ﬁrst deleting and then reinserting the ground state of V
1
(x) using the
Darboux procedure [64, 70]. The generalization to obtain an nparameter
family is described in Sec. 7.2 [72]. When applied to a reﬂectionless potential
(Sec. 7.3), the nparameter families provide surprisingly simple expressions
for the pure multisoliton solutions [75] of the Kortewegde Vries (KdV) and
other nonlinear evolution equations [159, 160, 138, 161, 162].
7.1 One Parameter Family of Isospectral Potentials
In this subsection, we describe two approaches of obtaining the oneparameter
family V
1
(λ
1
; x) of potentials isospectral to a given potential V
1
(x).
72
The ﬁrst approach follows from asking the following question: Suppose
V
2
(x) is the SUSY partner potential of the original potential V
1
(x), and let
W(x) be the superpotential such that V
2
(x) = W
2
+W
′
and V
1
(x) = W
2
−W
′
.
Then, given V
2
(x), is the original potential V
1
(x) unique i.e., for a given
V
2
(x), what are the various possible superpotentials
ˆ
W(x) and corresponding
potentials
ˆ
V
1
(x) =
ˆ
W
2
−
ˆ
W
′
? Let us assume [68] that there exists a more
general superpotential which satisﬁes
V
2
(x) =
ˆ
W
2
(x) +
ˆ
W
′
(x). (285)
Clearly,
ˆ
W = W is one of the solutions to eq. (285). To ﬁnd the most general
solution, let us set
ˆ
W(x) = W(x) + φ(x) (286)
in eq. (285). We ﬁnd then that y(x) = φ
−1
(x) satisﬁes the Bernoulli equation
y
′
(x) = 1 + 2Wy (287)
whose solution is
1
y(x)
= φ(x) =
d
dx
ln[1
1
(x) + λ
1
]. (288)
Here
1
1
(x) ≡
x
−∞
ψ
2
1
(x
′
)dx
′
. (289)
λ
1
is a constant of integration and ψ
1
(x) is the normalized ground state wave
function of V
1
(x) = W
2
(x) −W
′
(x). Thus the most general
ˆ
W(x) satisfying
eq. (285) is given by
ˆ
W(x) = W(x) +
d
dx
ln[1
1
(x) + λ
1
], (290)
so that the one parameter family of potentials
ˆ
V
1
(x) =
ˆ
W
2
(x) −
ˆ
W
′
(x) = V
1
(x) −2
d
2
dx
2
ln[1
1
(x) + λ
1
] (291)
have the same SUSY partner V
2
(x).
In the second approach, we delete the ground state ψ
1
at energy E
1
for
the potential V
1
(x). This generates the SUSY partner potential V
2
(x) =
73
V
1
− 2
d
2
dx
2
lnψ
1
, which is almost isospectral to V
1
(x) i.e., it has the same
eigenvalues as V
1
(x) except for the bound state at energy E
1
. The next step
is to reinstate a bound state at energy E
1
.
Although the potential V
2
does not have an eigenenergy E
1
, the function
1/ψ
1
satiﬁes the Schr¨odinger equation with potential V
2
and energy E
1
. The
other linearly independent solution is
x
−∞
ψ
2
1
(x
′
)dx
′
/ψ
1
. Therefore, the most
general solution of the Schr¨odinger equation for the potential V
2
at energy
E
1
is
Φ
1
(λ
1
) = (1
1
+ λ
1
)/ψ
1
. (292)
Now, starting with a potential V
2
, we can again use the standard SUSY
(Darboux) procedure to add a state at E
1
by using the general solution
Φ
1
(λ
1
),
ˆ
V
1
(λ
1
) = V
2
−2
d
2
dx
2
ln Φ
1
(λ
1
) . (293)
The function 1/Φ
1
(λ
1
) is the normalizable ground state wave function of
ˆ
V
1
(λ
1
), provided that λ
1
does not lie in the interval −1 ≤ λ
1
≤ 0. Therefore,
we ﬁnd a oneparameter family of potentials
ˆ
V
1
(λ
1
) isospectral to V
1
for
λ
1
> 0 or λ
1
< −1.
ˆ
V
1
(λ
1
) = V
1
−2
d
2
dx
2
ln(ψ
1
Φ
1
(λ
1
))
= V
1
−2
d
2
dx
2
ln(1
1
+ λ
1
) . (294)
The corresponding ground state wave functions are
ˆ
ψ
1
(λ
1
; x) = 1/Φ
1
(λ
1
) . (295)
Note that this family contains the original potential V
1
. This corresponds to
the choices λ
1
→±∞.
To elucidate this discussion, it may be worthwhile to explicitly construct
the oneparameter family of strictly isospectral potentials corresponding to
the one dimensional harmonic oscillator [70]. In this case
W(x) =
ω
2
x (296)
74
so that
V
1
(x) =
ω
2
4
x
2
−
ω
2
. (297)
The normalized ground state eigenfunction of V
1
(x) is
ψ
1
(x) =
ω
2π
1/4
exp(−ωx
2
/4) (298)
Using eq. (289) it is now straightforward to compute the corresponding 1
1
(x).
We get
1
1
(x) = 1 −
1
2
erfc
√
ω
2
x
; erfc(x) =
2
√
π
∞
x
e
−t
2
dt . (299)
Using eqs. (294) and (295), one obtains the one parameter family of
isospectral potentials and the corresponding wave functions. In Figs. 7.1
and 7.2, we have plotted some of the potentials and wave functions for the
case ω = 2. We see that as λ
1
decreases from ∞ to 0,
ˆ
V
1
starts developing
a minimum which shifts towards x = −∞. Note that as λ
1
ﬁnally becomes
zero this attractive potential well is lost and we lose a bound state. The
remaining potential is called the Pursey potential V
P
(x) [66]. The general
formula for V
P
(x) is obtained by putting λ
1
= 0 in eq. (294). An analogous
situation occurs in the limit λ
1
= −1, the remaining potential being the
AbrahamMoses potential [65].
7.2 Generalization to nParameter Isospectral Family
The second approach discussed in the previous subsection can be generalized
by ﬁrst deleting all n bound states of the original potential V
1
(x) and then
reinstating them one at a time. Since one parameter is generated every time
an eigenstate is reinstated, the ﬁnal result is a nparameter isospectral family
[72]. Recall that deleting the eigenenergy E
1
gave the potential V
2
(x). The
ground state ψ
2
for the potential V
2
is located at energy E
2
. The procedure
can be repeated “upward”, producing potentials V
3
, V
4
, . . . with ground states
ψ
3
, ψ
4
, . . . at energies E
3
, E
4
, . . . , until the top potential V
n+1
(x) holds no
bound state (see Fig. 7.3, which corresponds to n = 2).
75
In order to produce a twoparameter family of isospectral potentials, we
go from V
1
to V
2
to V
3
by successively deleting the two lowest states of V
1
and then we readd the two states at E
2
and E
1
by SUSY transformations.
The most general solutions of the Schr¨odinger equation for the potential V
3
are given by Φ
2
(λ
2
) = (1
2
+ λ
2
)/ψ
2
at energy E
2
, and A
2
Φ
1
(λ
1
) at energy
E
1
(see Fig. 7.3). The quantities 1
i
are deﬁned by
1
i
(x) ≡
x
−∞
ψ
2
i
(x
′
)dx
′
. (300)
Here the SUSY operator A
i
relates solutions of the potentials V
i
and V
i+1
,
A
i
=
d
dx
−(ln ψ
i
)
′
. (301)
Then, as before, we ﬁnd an isospectral oneparameter family
ˆ
V
2
(λ
2
),
ˆ
V
2
(λ
2
) = V
2
−2
d
2
dx
2
ln(1
2
+ λ
2
) . (302)
The solutions of the Schr¨odinger equation for potentials V
3
and
ˆ
V
2
(λ
2
) are
related by a new SUSY operator
ˆ
A
†
2
(λ
2
) = −
d
dx
+ (ln Φ
2
(λ
2
))
′
. (303)
Therefore, the solution Φ
1
(λ
1
, λ
2
) at E
1
for
ˆ
V
2
(λ
2
) is
Φ
1
(λ
1
, λ
2
) =
ˆ
A
†
2
(λ
2
)A
2
Φ
1
(λ
1
) . (304)
The normalizable function 1/Φ
1
(λ
1
, λ
2
) is the ground state at E
1
of a new
potential, which results in a twoparameter family of isospectral systems
ˆ
V
1
(λ
1
, λ
2
),
ˆ
V
1
(λ
1
, λ
2
) = V
1
−2
d
2
dx
2
ln(ψ
1
ψ
2
Φ
2
(λ
2
)Φ
1
(λ
1
, λ
2
))
= V
1
−2
d
2
dx
2
ln(ψ
1
(1
2
+ λ
2
)Φ
1
(λ
1
, λ
2
)) , (305)
for λ
i
> 0 or λ
i
< −1. A useful alternative expression is
76
ˆ
V
1
(λ
1
, λ
2
) = −
ˆ
V
2
(λ
2
) + 2(Φ
′
1
(λ
1
, λ
2
)/Φ
1
(λ
1
, λ
2
))
2
+ 2E
1
. (306)
The above procedure is best illustrated by the pyramid structure in Fig. 7.3.
It can be generalized to an nparameter family of isospectral potentials for an
initial system with n bound states. The formulas for an nparameter family
are
Φ
i
(λ
i
) = (1
i
+ λ
i
)/ψ
i
, i = 1, , n ; (307)
A
i
=
d
dx
−(ln ψ
i
)
′
; (308)
ˆ
A
†
i
(λ
i
, , λ
n
) = −
d
dx
+ (ln Φ
i
(λ
i
, , λ
n
))
′
; (309)
Φ
i
(λ
i
, λ
i+1
, , λ
n
)
=
ˆ
A
†
i+1
(λ
i+1
, λ
i+2
, , λ
n
)
ˆ
A
†
i+2
(λ
i+2
, λ
i+3
, , λ
n
)
ˆ
A
†
n
(λ
n
)
A
n
A
n−1
A
i+1
Φ
i
(λ
i
) ; (310)
ˆ
V
1
(λ
1
, , λ
n
) = V
1
−2
d
2
dx
2
ln(ψ
1
ψ
2
ψ
n
Φ
n
(λ
n
) Φ
1
(λ
1
, , λ
n
)) . (311)
The above equations [72] summarize the main results of this section.
7.3 nSoliton Solutions of the KdV Equation
As an application of isospectral potential families, we consider reﬂectionless
potentials of the form
V
1
= −n(n + 1)sech
2
x , (312)
where n is an integer, since these potentials are of special physical inter
est. V
1
holds n bound states, and we may form a nparameter family
of isospectral potentials. We start with the simplest case n = 1. We
77
have V
1
= −2sech
2
x, E
1
= −1 and ψ
1
= 2
−
1
2
sechx. The corresponding 1
parameter family is
ˆ
V
1
(λ
1
) = −2sech
2
(x +
1
2
ln(1 +
1
λ
1
)) . (313)
Clearly, varying the parameter λ
1
corresponds to translations of V
1
(x). As
λ
1
approaches the limits 0
+
(Pursey limit) and −1
−
(AbrahamMoses limit),
the minimum of the potential moves to −∞ and +∞ respectively.
For the case n = 2, V
1
= −6sech
2
x and there are two bound states at
E
1
= −4 and E
2
= −1. The SUSY partner potential is V
2
= −2sech
2
x.
The ground state wave functions of V
1
and V
2
are ψ
1
=
√
3
2
sech
2
x and ψ
2
=
1
√
2
sechx. Also, 1
1
=
1
4
(3 tanhx −tanh
3
x +2) and 1
2
=
1
2
(tanh x +1). After
some algebraic work, we obtain the 2parameter family
ˆ
V
1
(λ
1
, λ
2
) = −12
[3 + 4 cosh(2x −2δ
2
) + cosh(4x −2δ
1
)]
[cosh(3x −δ
2
−δ
1
) + 3 cosh(x + δ
2
−δ
1
)]
2
,
δ
i
≡ −
1
2
ln(1 +
1
λ
i
) , i = 1, 2 .
As we let λ
1
→ −1, a well with one bound state at E
1
will move in the +x
direction leaving behind a shallow well with one bound state at E
2
. The
movement of the shallow well is essentially controlled by the parameter λ
2
.
Thus, we have the freedom to move either one of the wells.
It is tedious but straightforward to obtain the result for arbitrary n and
get
ˆ
V
1
(λ
1
, λ
2
, , λ
n
, x). It is well known that oneparameter (t) families of
isospectral potentials can also be obtained as solutions of a certain class of
nonlinear evolution equations [159, 160, 138, 162]. These equations have the
form (q = 0, 1, 2, )
−u
t
= (L
u
)
q
u
x
(314)
where the operator L
u
is deﬁned by
L
u
f(x) = f
xx
−4uf + 2u
x
∞
x
dyf(y) (315)
and u is chosen to vanish at inﬁnity. [For q = 0 we simply get −u
t
= u
x
,
while for q = 1 we obtain the well studied Kortewegde Vries (KdV) equa
tion]. These equations are also known to possess pure (i.e., reﬂectionless)
78
multisoliton solutions. It is possible to show that by suitably choosing the
parameters λ
i
as functions of t in the nparameter SUSY isospectral family of
a symmetric reﬂectionless potential holding n bound states, we can obtain an
explicit analytic formula for the pure nsoliton solution of each of the above
evolution equations [75]. These expressions for the multisoliton solutions of
eq. (314) are much simpler than any previously obtained using other proce
dures. Nevertheless, rather than displaying the explicit algebraic expressions
here, we shall simply illustrate the 3soliton solution of the KdV equation.
The potentials shown in Fig. 7.4 are all isospectral and reﬂectionless holding
bound states at E
1
= 25/16, E
2
= −1, E
3
= 16/25 . As t increases, note
the clear emergence of three independent solitons.
In this section, we have found nparameter isospectral families by repeat
edly using the supersymmetric Darboux procedure for removing and inserting
bound states. However, as brieﬂy mentioned in Sec. 7.1, there are two other
closely related, well established procedures for deleting and adding bound
states. These are the AbrahamMoses procedure [65] and the Pursey proce
dure [66]. If these alternative procedures are used, one gets new potential
families all having the same bound state energies but diﬀerent reﬂection and
transmission coeﬃcients. Details can be found in reference [71].
8 Path Integrals and Supersymmetry
In this section, we will describe the Lagrangian formulation of SUSY QM
and discuss three related path integrals: one for the generating functional
of correlation functions, one for the Witten index  a topological quantity
which determines whether SUSY is broken, and one for a related “classical”
stochastic diﬀerential equation, namely the Langevin equation. We will also
brieﬂy discuss the superspace formalism for SUSY QM.
Starting from the matrix SUSY Hamiltonian which is 1/2 of our previous
H [eq. (60)] for convenience:
H =
1
2
p
2
+
1
2
W
2
(x)I −
1
2
[ψ, ψ
†
]W
′
(x),
we obtain the Lagrangian
L =
1
2
˙ x
2
+ iψ
†
∂
t
ψ −
1
2
W
2
(x) +
1
2
[ψ, ψ
†
]W
′
(x). (316)
79
It is most useful to consider the generating functional of correlation functions
in Euclidean space. We rotate t → iτ and obtain for the Euclidean path
integral:
Z[j, η, η
∗
] =
[dx][dψ][dψ
∗
] exp[−S
E
+
jx + ηψ
∗
+ η
∗
ψ], (317)
where
S
E
=
τ
0
dτ
1
2
x
2
τ
+
1
2
W
2
(x) −ψ
∗
[∂
τ
−W
′
(x)]ψ
and ψ and ψ
∗
are now elements of a Grassman algebra:
¦ψ
∗
, ψ¦ = ¦ψ, ψ¦ = ¦ψ
∗
, ψ
∗
¦ = 0, (318)
and
x
τ
=
dx
dτ
. The Euclidean action is invariant under the following SUSY transformations
[163, 29] which mix bosonic and fermionic degrees of freedom:
δx = ǫ
∗
ψ + ψ
∗
ǫ,
δψ
∗
= −ǫ
∗
(∂
τ
x + W(x)) ,
δψ = −ǫ (−∂
τ
x + W(x)) , (319)
where ǫ and ǫ
∗
are two inﬁnitesimal anticommuting parameters. These trans
formations correspond to N = 2 supersymmetry.
The path integral over the fermions can now be explicitly performed using
a cutoﬀ lattice which is periodic in the the coordinate x but antiperiodic in
the fermionic degrees of freedom at τ = 0 and τ = T. Namely we evaluate
the fermionic path integral:
[dψ][dψ
∗
]e
T
0
dτψ
∗
[∂τ −W
′
(x)]ψ
(320)
by calculating the determinant of the operator [∂
τ
−W
′
(x)] using eigenvectors
which are antiperiodic.
We have, following Gildener and Patrascioiu [164], that
det[∂
τ
−W
′
(x)] =
¸
m
λ
m
,
80
where
[∂
τ
−W
′
(x)]ψ
m
= λ
m
ψ
m
,
so that
ψ
m
(τ) = C
m
exp[
τ
0
dτ
′
[λ
m
+ W
′
]. (321)
Imposing the antiperiodic boundary conditions:
ψ
m
(T) = −ψ
m
(0)
yields:
λ
m
=
i(2m + 1)π
T
−
1
T
T
0
dτW
′
(x). (322)
Regulating the determinant by dividing by the determinant for the case where
the potential is zero we obtain:
det[
∂
τ
−W
′
(x)
∂
τ
] = cosh
T
0
dτ
W
′
(x)
2
. (323)
Rewriting the cosh as a sum of two exponentials we ﬁnd, as expected that Z
is the sum of the partition functions for the two pieces of the supersymmetric
Hamiltonian. Namely when the external sources are zero:
Tre
−H
1
T
+ Tre
−H
2
T
≡ Z
−
+ Z
+
. (324)
For the case when SUSY is unbroken, only the ground state of H
1
con
tributes as T →∞ We also have:
Z
±
=
[dx]exp[−S
±
E
] (325)
where
S
±
E
=
T
0
dτ
˙ x
2
2
+
W
2
(x)
2
±
W
′
(x)
2
.
A related path integral is obtained for the noise averaged correlation
functions coming from a classical stochastic equation, the Langevin equation.
If we have the stochastic diﬀerential equation
˙ x = W(x(τ)) +η(τ) (326)
where η(τ) is a random stirring forcing obeying Gaussian statistics, then the
correlation functions of x are exactly the same as the correlation functions
81
obtained from the Euclidean quantum mechanics related to the Hamiltonian
H
1
. To see this we realize that Gaussian noise is described by a probability
functional:
P[η] = Nexp[−
1
2
T
0
dτ
η
2
(τ)
F
0
] (327)
normalized so that:
DηP[η] = 1,
DηP[η]f(τ) = 0,
DηP[η]f(τ)f(τ
′
) = F
0
δ(τ −τ
′
).
The correlation functions averaged over the noise are:
< x(τ
1
)x(τ
2
)... >=
DηP[η]x(τ
1
)x(τ
2
)... (328)
where we have in mind ﬁrst solving the Langevin equation explicitly for
x(η(τ)) and then averaging over the noise as discussed in Sec. 2.2. An
other way to calculate the correlation function is to change variables in the
functional integral from η to x.
< x(τ
1
)x(τ
2
)... >=
D[x] P[η] Det [
dη(τ)
dx(τ
′
)
[ x(τ
1
)x(τ
2
)... (329)
This involves calculating the functional determinant,
Det[
dη(τ)
dx(τ
′
)
[ (330)
subject to the boundary condition that the Green’s function obey causality,
so one has retarded boundary conditions. One has
Det[
dη
dx
[ = exp
dt Tr ln
[
d
dτ
−W
′
(x(τ))]δ(τ −τ
′
)
dτ. (331)
When there are no interactions the retarded boundary conditions yield
G
0
(τ −τ
′
) = θ(τ −τ
′
). (332)
82
Expanding ln(1 −G
0
W
′
) one ﬁnds because of the retarded boundary condi
tions that only the ﬁrst term in the expansion contributes so that
Det[
dη(τ)
dx(τ
′
)
[ = exp[
1
2
T
0
dτW
′
(x)]. (333)
Choosing F
0
= ¯ h so that
P[η] = N exp[−
1
2
T
0
dτ
η
2
(τ)
F
0
]
= N exp[−
1
2¯ h
T
0
dτ( ˙ x −W(x))
2
]
= N exp[−
1
2¯ h
T
0
dτ( ˙ x
2
+ W
2
(x))], (334)
we ﬁnd that the generating functional for the correlation functions is exactly
the generating functional for the correlation functions for Euclidean quantum
mechanics corresponding to the Hamiltonian H
1
:
Z[j] = N
D[x] exp[−
1
2
T
0
dτ
˙ x
2
+ W
2
(x) −W
′
(x) −2j(τ)x(τ)
] (335)
Thus we see that we can determine the correlation functions of x for the
Hamiltonian H
1
by either evaluating the path integral or solving the Langevin
equation and averaging over Gaussian noise.
An equation related to the Langevin equation is the FokkerPlanck equa
tion, which deﬁnes the classical probability function P
c
for the equal time
correlation functions of H
1
. Deﬁning the noise average:
P
c
(z) =< δ(z −x(t) >
η
=
Dηδ(z −x(t))P[η] (336)
one obviously has:
dzz
n
P
c
(z, t) =
Dη[x(t)]
n
P[η] =< x
n
> .
One can show [165] that P
c
obeys the FokkerPlanck equation:
∂P
∂t
=
1
2
F
0
∂
2
P
∂z
2
+
∂
∂z
[W(z)P(z, t))] (337)
83
For an equilibrium distribution to exist at long times t one requires that
P(z, t) →
ˆ
P(z)
and
ˆ
P(z)dz = 1.
Setting
∂P
∂t
= 0 in the FokkerPlanck equation, we obtain
ˆ
P(z) = Ne
−2
z
0
W(y)dy
= ψ
0
(z)
2
. (338)
Thus at long times only the ground state wave function contributes (we are in
Euclidean space) and the probability function is just the usual ground state
wave function squared. We see from this that when SUSY is broken, one
cannot deﬁne an equilibrium distribution for the classical stochastic system.
A third path integral for SUSY QM is related to the Witten index. As
we discussed before, one can introduce a “fermion” number operator via
n
F
=
1 −σ
3
2
=
1 −[ψ, ψ
†
]
2
. (339)
Thus
(−1)
F
= [ψ, ψ
†
] = σ
3
. (340)
The Witten index is given by △ = Tr(−1)
F
. As we discussed earlier, the
Witten index needs to be regulated and the regulated index is deﬁned as:
△(β) = Tr(−1)
F
e
−βH
= Tr(e
−βH
1
−e
−βH
2
) . (341)
In Sec 2.2 we showed how to determine △(β) using heat kernel methods and
how it was useful in discussing nonperturbative breaking of SUSY. Here
we will show that the Witten index can also be obtained using the path
integral representation of the generating functional of SUSY QM where the
fermion determinant is now evaluated using periodic boundary conditions to
incorporate the factor (−1)
F
. It is easy to verify a posteriori that this is the
case. Consider the path integral:
△(β) =
[dx][dΨ][dΨ
∗
]e
β
0
L
E
(x,Ψ,Ψ
∗
)dτ
, (342)
84
where
L
E
=
1
2
x
2
τ
+
1
2
W
2
−Ψ
∗
[∂
τ
−W
′
(x)]Ψ.
To incorporate the (−1)
F
in the trace, one changes the boundary conditions
for evaluating the fermion determinant at τ = 0, β to periodic ones:
x(0) = x(β) Ψ(0) = Ψ(β)
The path integral over the fermions can again be explicitly performed using
a cutoﬀ lattice which is periodic in the fermionic degrees of freedom at τ = 0
and τ = β. We now impose these boundary conditions on the determinant
of the operator [∂
τ
−W
′
(x)] using eigenvectors which are periodic.
We again have
det[∂
τ
−W
′
(x)] =
¸
m
λ
m
.
Imposing periodic boundary conditions:
λ
m
=
i(2m)π
β
−
1
β
β
0
dτW
′
(x). (343)
Regulating the determinant by dividing by the determinant for the case where
the potential is zero we obtain:
det
¸
∂
τ
−W
′
(x)
∂
τ
¸
= sinh
β
0
dτ
W
′
(x)
2
. (344)
Again rewriting the sinh as a sum of two exponentials we ﬁnd, as expected
that we obtain the regulated Witten index:
△(β) = Z
−
−Z
+
= Tre
−βH
1
−Tre
−βH
2
. (345)
8.1 Superspace Formulation of Supersymmetric Quan
tum Mechanics
One can think of SUSY QM as a degenerate case of supersymmetric ﬁeld
theory in d = 1 in the superspace formalism of Salam and Strathdee [167]
(This idea is originally found in unpublished lecture notes of S. Raby [166]).
superﬁelds are deﬁned on the space (x
n
; θ
a
) where x is the space coordinate
and θ
a
are anticommuting spinors. In the degenerate case d = 1 the ﬁeld
85
is replaced by x(t) so that the only coordinate is time. The anticommuting
variables are θ and θ
∗
where
¦θ, θ
∗
¦ = ¦θ, θ¦ = [θ, t] = 0.
Consider the following SUSY transformation:
t
′
= t −i(θ
∗
ǫ −ǫ
∗
θ), θ
′
= θ + ǫ, θ
∗′
= θ
∗
+ ǫ
∗
. (346)
If we assume that ﬁnite SUSY transformations can be parametrized by
L = e
i(ǫ
∗
Q
∗
+Qǫ)
,
then from
δA = i[ǫ
∗
Q
∗
+ Qǫ, A] (347)
we infer that the operators Q and Q
∗
are given by:
Q = i∂
θ
−θ
∗
∂
t
, Q
∗
= −i∂
θ
∗ −θ∂
t
. (348)
Now these charges obey the familiar SUSY QM algebra:
¦Q, Q
∗
¦ = 2i∂
t
= 2H, [Q, H] = 0. (349)
The Lagrangian in superspace is determined as follows. A superﬁeld made
up of x and θ and θ
∗
can at most be a bilinear in the Grassman variables:
φ(x, θ, θ
∗
) = x(t) + iθψ(t) −iψ
∗
θ
∗
+ θθ
∗
D(t). (350)
Under a SUSY transformation, the following derivatives are invariant:
D
θ
= ∂
θ
−iθ
∗
∂
t
or in component form:
D
θ
φ = iψ −θ
∗
D −iθ
∗
˙ x + θ
∗
θ
˙
ψ, (351)
and
D
θ
∗ = ∂
θ
∗ −iθ∂
t
or in component form:
[D
θ
φ]
∗
= −iψ
∗
−θD + iθ ˙ x + θ
∗
θ
˙
ψ
∗
. (352)
86
The most general invariant action is:
S =
dtdθ
∗
dθ
1
2
[D
θ
φ[
2
−f(φ)
(353)
Again the expansion in terms of the Grassman variables causes a Taylor
expansion of f to truncate at the second derivative level. Integrating over
the Grassman degrees of freedom using the usual path integral rules for
Grassman variables:
θdθ =
θ
∗
dθ
∗
= 1,
dθ =
dθ
∗
= 0,
one obtains
S =
dt(
1
2
˙ x
2
+ ψ
∗
[∂
t
−f
′′
(x)]ψ +
1
2
D
2
+ Df
′
(x)). (354)
Eliminating the constraint variable D = −f
′
(x) = W(x) we obtain our
previous result for the action (now in Minkowski space):
S =
dt
1
2
˙ x
2
+ ψ
∗
[∂
t
−W
′
(x)]ψ −
1
2
W
2
(355)
A more complete discussion of this can be found in ref. [11].
9 Perturbative Methods for Calculating En
ergy Spectra and Wave Functions
The framework of supersymmetric quantum mechanics has been very useful
in generating several new perturbative methods for calculating the energy
spectra and wave functions for one dimensional potentials. Four such meth
ods are described in this section.
In Secs. 9.1 and 9.2, we discuss two approximation methods (the vari
ational method and the δ− expansion) for determining the wave functions
and energ y eigenvalues of the anharmonic oscillator making use of SUSY
QM. Sec. 9.3 contains a description of a SUSY QM calculation of the energy
splitting and rate of tunneling in a double well potential. The result is a
rapidly converging series which is substantially better than the usual WKB
tunneling formula. Finally, in Sec. 9.4, we describe how the large N expan
sion (N = number of spatial dimensions) used in quantum mechanics can be
further improved by incorporating SUSY.
87
9.1 Variational Approach
The anharmonic oscillator potential V (x) = gx
4
is not exactly solvable. To
determine the superpotential one has to ﬁrst subtract the ground state energy
E
0
and solve the Riccati equation for W(x):
V
1
(x) = gx
4
−E
0
≡ W
2
−W
′
, (356)
Once the ground state energy and the superpotential is known to some order
of accuracy, one can then determine the partner potential and its ground
state wave function approximately. Then, using the SUSY operator
d
dx
−W(x)
one can construct the ﬁrst excited state of the anharmonic oscillator in the
usual manner. Using the hierarchy of Hamiltonians discussed in Sec. 3,
one can construct from the approximate ground state wave functions of the
hierarchy and the approximate superpotentials W
n
all the excited states of
the anharmonic oscillator approximately.
First let us see how this works using a simple variational approach. For
the original potential, we can determine the optimal Gaussian wave function
quite easily. Assuming a trial wave function of the form
ψ
0
= (
2β
π
)
1/4
e
−βx
2
, (357)
we obtain
< H >=<
p
2
2
+ gx
4
>=
β
2
+
3g
16β
2
. (358)
(In this subsection, we are taking m = 1 in order to make contact with pub
lished numerical results). Minimizing the expectation value of the Hamiltonia
n with respect to the parameter β yields
E
0
= (
3
4
)
4/3
g
1/3
, β = (
3
4
)
1/3
g
1/3
.
This is rather good for this crude approximation since the exact ground state
energy of the anharmonic oscillator determined numerically is E
0
= .668g
1/3
88
whereas (
3
4
)
4/3
= .681 . The approximate potential W resulting from this
variation calculation is
W(x) = −
dlogψ
0
dx
= 2βx, (359)
which leads to a Gaussian approximation to the potential
V
1G
= 4βx
2
−2β. (360)
The approximate supersymmetric partner potential is now
V
2G
= 4βx
2
+ 2β. (361)
Since V
2G
diﬀers from V
1G
by a constant, the approximate ground state wave
function for V
2
is given by eq. (357). The approximate ground state energy
of the second potential is now
< H
2
>=
v
< 0[
p
2
2
+ V
2G
[0 >
v
= H
1
+ 4β. (362)
Thus we have approximately that the energy diﬀerence between the ground
state and ﬁrst excited state of the anharmonic oscillator is
E
1
−E
0
= 4β = 4(
3
4
)
1/3
g
1/3
.
The approximate (unnormalized) ﬁrst excited state wave function is
ψ
(1)
1
= [
d
dx
−2βx]ψ
0v
∝ −4βxe
−βx
2
. (363)
This method can be used to ﬁnd all the excited state wave functions and
energy levels of the anharmonic oscillator by using the methods discussed in
ref. [168].
Let us now look at a more general class of trial wave functions. If we
choose for the trial ground state wave functions of the hierarchy of Hamilto
nians,
ψ
(k)
0
= N
k
exp
−
1
2
x
2
ρ
k
n
k
, N
k
=
2
√
ρ
k
Γ(1 +
1
2n
k
)
−1/2
.
(364)
89
we obtain much better agreement for the low lying eigenvalues and eigen
functions. It is convenient in this case to ﬁrst scale the Hamiltonian for the
anharmonic oscillator,
H = −
1
2
d
2
dx
2
+ gx
4
, (365)
by letting x → x/g
1/6
and H → g
1/3
H. Then we ﬁnd the ground state
energy of the anharmonic oscillator and the variational parameters ρ
1
and
n
1
by forming the functional
E
0
(ρ
1
, n
1
) = ' ψ
0
[ −
1
2
d
2
dx
2
+ x
4
[ψ
0
`. (366)
Thus we ﬁrst determine ρ
1
and n
1
by requiring
∂E
0
∂ρ
1
= 0 ,
∂E
0
∂n
1
= 0 . (367)
The equation for the energy functional for the anharmonic oscillator is
E
0
(ρ
1
, n
1
) =
n
2
1
2ρ
1
Γ(2 −
1
2n
1
)
Γ(
1
2n
1
)
+ ρ
2
1
Γ(
5
2n
1
)
Γ(
1
2n
1
)
. (368)
Minimizing this expression, we obtain the following variational result:
E
0
= 0.66933, n
1
= 1.18346, ρ
1
= 0.666721 . (369)
This ground state energy is to be compared with a numerical evaluation
which yie lds 0.667986 . Since the trial wave function for all ground states is
given by eq. (364), the variational superpotential for all k is
W
kv
= n
k
[x[
2n
k
−1
(ρ
k
)
−n
k
. (370)
Since we are interested in the energy diﬀerences E
n
−E
n−1
of the anharmonic
oscillator, we consider the variational Hamiltonian
¯
H
vk+1
=
1
2
A
kv
A
†
kv
(371)
which approximately determines these energy diﬀerences. We obtain the
approximate energy splittings by minimizing the energy functional
δE
k
(ρ
k
, n
k
) =
1
2
' ψ
(vk+1)
0
[ −
d
2
dx
2
+ W
2
vk
+ W
′
vk
[ψ
(vk+1)
0
` . (372)
90
Performing the integrals one obtains the simple recursion relation:
δE
k
(ρ
k
, n
k
) =
n
2
k
2ρ
k
Γ
2 −
1
2n
k
Γ
1
2n
k
+
n
2
k−1
2ρ
k
ρ
k
ρ
k−1
2n
k−1
Γ
4n
k−1
−1
2n
k
Γ
1
2n
k
+
n
k−1
2ρ
k
(2n
k−1
−1)
ρ
k
ρ
k−1
n
k−1
Γ
2n
k−1
−1
2n
k
Γ
1
2n
k
. (373)
One can perform the minimization in ρ analytically leaving one minimization
to perform numerically.
The results for the variational parameters and for the energy diﬀerences
are presented in Table 9.1 for the ﬁrst three energy eigenvalues and compared
with a numerical calculation, based on a shooting method.
9.2 δ Expansion Method
In this method [78] we consider the anharmonic oscillator as an analytic
continuation from the harmonic oscillator in the paramater controlling the
anharmonicity. That is we consider simultaneously potentials of the form
V
1
(x) = M
2+δ
x
2+2δ
−C(δ) ≡ W
2
(x, δ) −W
′
, (374)
where M is a scale parameter, δ measures the anharmonicity, and C is the
ground state energy of the anharmonic oscilator. C is subtracted as usual
from the potential so it can be factorized. The standard anharmonic oscil
lator corresponds to δ = 1 and M = (2g)
1/3
. To approximately determine
W(x) from V
1
(x) we assume that both W(x) and V
1
(x) have a Taylor series
expansion in δ. Thus we write:
V
1
(x) = M
2
x
2
∞
¸
n=0
δ
n
[ln(Mx
2
)]
n
n!
−
∞
¸
n=0
2E
n
δ
n
, (375)
where E
n
corresponds to the order Taylor expansion of the dependence of
the ground state energy on the parameter δ. We assume
W(x) =
∞
¸
n=0
δ
n
W
(n)
(x), (376)
91
and insert these expressions in eq. (374) and match terms order by order.
At lowest order in δ the problem reduces to the supersymmetric harmonic
oscillator. We have:
W
2
0
−W
′
0
= M
2
x
2
−2E
0
, (377)
whose solution is
W
0
(x) = Mx, E
0
=
1
2
M (378)
To next order we have the diﬀerential equation:
dW
1
dx
−2W
1
W
0
= −M
2
x
2
ln(Mx
2
) + 2E
1
(379)
which is to be solved with the boundary conditionW
n
(0) = 0. The order δ
contribution to the energy eigenvalue E
1
is determined by requiring that the
ground state wave function be square integrable. Solving for W
1
we obtain
W
1
(x) = −e
Mx
2
x
0
dye
−My
2
[M
2
y
2
ln(My
2
) −2E
1
]. (380)
To ﬁrst order in δ the ground state wave function is now :
ψ
0
(x) = e
−Mx
2
(1 −δ
x
0
dyW
1
(y))
Imposing the condition that ψ
0
vanishes at inﬁnity, we obtain:
E
1
=
1
4
Mψ(3/2) , ψ(x) = Γ
′
(x)/Γ(x). (381)
Writing M = (2g)
1/3
, we ﬁnd that the ﬁrst two terms in the δ expansion for
the ground state energy are
E =
1
2
(2g)
1/3
[1 +
1
2
ψ(3/2)δ]. (382)
At δ = 1, we get
E = .6415g
1/3
. (383)
A more accurate determination of the ground state energy can be obtained
by calculating up to order δ
2
and then analytically continuing in δ using Pad´e
approximants. This is discussed in ref. [169].
92
9.3 Supersymmetry and Double Well Potentials
Supersymmetric quantum mechanics has been proﬁtably used to obtain a
novel perturbation expansion for the probabililty of tunneling in a double
well potential [34]. Since double wells are widely used in many areas of
physics and chemistry, this expansion has found many applications ranging
from condensed matter physics to the computation of chemical reaction rates
[170, 171, 172, 173, 174, 175, 176] ,[177, 178, 31, 32, 33, 179, 180] . In what
follows, we shall restrict our attention to symmetric double wells, although
an extension to asymmetric double wells is relatively straightforward [181].
Usually, in most applications the quantity of interest is the energy diﬀer
ence t ≡ E
1
−E
0
between the lowest two eigenstates, and corresponds to the
tunneling rate through the doublewell barrier. The quantity t is often small
and diﬃcult to calculate numerically, especially when the potential barrier
between the two wells is large. Here, we show how SUSY facilitates the
evaluation of t. Indeed, using the supersymmetric partner potential V
2
(x),
we obtain a systematic, highly convergent perturbation expansion for the
energy diﬀerence t. The leading term is more accurate than the standard
WKB tunneling formula, and the magnitude of the nonleading terms gives a
reliable handle on the accuracy of the result.
First, we brieﬂy review the standard approach for determining t in the
case of a symmetric, onedimensional doublewell potential, V
1
(x), whose
minima are located x = ±x
0
. We deﬁne the depth, D, of V
1
(x) by D ≡
V
1
(0) − V
1
(x
0
). An example of such a potential is shown in Fig. 9.1. For
suﬃciently deep wells, the doublewell structure produces closely spaced pairs
of energy levels lying below V
1
(0). The number of such pairs, n, can be
crudely estimated from the standard WKB boundstate formula applied to
V
1
(x) for x > 0:
nπ =
xc
0
[V
1
(0) −V
1
(x)]
1/2
dx, (384)
where x
c
is the classical turning point corresponding to energy V
1
(0) and we
have chosen units where ¯ h = 2m = 1. We shall call a doublewell potential
“shallow” if it can hold at most one pair of bound states, i.e., n ≤ 1. In
contrast, a “deep” potential refers to n ≥ 2.
The energy splitting t of the lowestlying pair of states can be obtained
by a standard argument [126]. Let χ(x) be the normalized eigenfunction for
93
a particle moving in a single well whose structure is the same as the right
hand well of V
1
(x) (i.e., x > 0). If the probability of barrier penetration is
small, the lowest two eigenfunctions of the doublewell potential V
1
(x) are
well approximated by
ψ
(1)
0,1
(x) = [χ(x) ±χ(−x)]/
√
2 . (385)
By integration of Schr¨odinger’s equation for the above eigenfunctions, it can
be shown that [126]
t ≡ E
1
−E
0
= 4χ(0)χ
′
(0) , (386)
where the prime denotes diﬀerentiation with respect to x. This result is
accurate for “deep” potentials, but becomes progressively worse as the depth
decreases. Use of WKB wave functions in eq. (386) yields the standard
result:
t
WKB
= ¦[2V
′′
1
(x
0
)]
1/2
/π¦ exp
−2
x
0
0
[V
1
(x) −V
1
(x
0
)]
1/2
dx
. (387)
The same result can also be obtained via instanton techniques [182].
Using the supersymmetric formulation of quantum mechanics for a given
Hamiltonian, H
1
= −d
2
/dx
2
+ V
1
(x), and its zeroenergy ground state wave
function ψ
0
(x), we know that the supersymmetric partner potential V
2
(x) is
given by
V
2
(x) = V
1
(x) −2(d/dx)(ψ
′
0
/ψ
0
)
= −V
1
(x) + 2(ψ
′
0
/ψ
0
)
2
. (388)
Alternatively, in terms of the superpotential W(x) given by W(x) =
−ψ
′
0
/ψ
0
we can write
V
2,1
(x) = W
2
(x) ±dW/dx . (389)
From the discussion of unbroken SUSY in previous sections, we know that
the energy spectra of the potentials V
2
and V
1
are identical, except for the
ground state of V
1
which is missing from the spectrum of V
2
[10]. Hence,
for the doublewell problem, we see that if V
1
(x) is “shallow” (i.e., only the
lowest two states are paired), then the spectrum of V
2
is well separated. In
this case, V
2
is relatively structureless and simpler than V
1
. Previous papers
[31, 32, 33] have implicitly treated just the case of shallow potentials, and,
94
not surprisingly, have found that the use of SUSY simpliﬁes the evaluation
of the energy diﬀerence t. In contrast, let us now consider the case of a
deep double well as shown in Fig. 9.1 . Here, the spectrum of V
2
has a
single unpaired ground state followed by paired excited states. In order to
produce this spectrum, V
2
has a doublewell structure together with a sharp
“δ− function like” dip at x = 0. This central dip produces the unpaired
ground state, and becomes sharper as the potential V
1
(x) becomes deeper.
As a concrete example, we consider the class of potentials whose ground
state wave function is the sum of two Gaussians, centered around ±x
0
,
ψ
0
(x) ∼ e
−(x−x
0
)
2
+ e
−(x+x
0
)
2
. (390)
The variables x and x
0
have been chosen to be dimensionless. The corre
sponding superpotential W(x), and the two supersymmetric partner poten
tials V
1
(x) and V
2
(x), are given respectively by
W(x) = 2[x −x
0
tanh(2xx
0
)], (391)
V
2,1
(x) = 4[x −x
0
tanh(2xx
0
)]
2
±2[1 −2x
2
0
sech
2
(2xx
0
)] . (392)
The minima of V
1
(x) are located near ±x
0
and the well depth (in the
limit of large x
0
) is D ≃ 4x
2
0
. We illustrate the potentials V
1
(x) and V
2
(x) in
Fig. 9.2 for the two choices x
0
= 1.0 and x
0
= 2.5. We see that in the limit
of large x
0
, for both V
1
(x) and V
2
(x), the wells become widely separated and
deep and that V
2
(x) develops a strong central dip.
The asymptotic behavior of the energy splitting, t, in the limit x
0
→ ∞
can be calculated from eq. (386), with χ(x) given by one of the (normalized)
Gaussians in eq. (390). We ﬁnd that
t →8x
0
(2/π)
1/2
e
−2x
2
0
. (393)
The same result can be obtained by observing that V
1
(x) → 4([ x [ −x
0
)
2
as x
0
→ ∞. This potential has a well known [183] analytic solution, which
involves solving the parabolic cylindrical diﬀerential equation. After carefully
handling the boundary conditions, one obtains the separation of the lowest
two energy levels to be 8x
0
(2/π)
1/2
exp(−2x
2
0
), in agreement with eq. (393).
95
We now turn to the evaluation of t via the ground state energy of the
supersymmetric partner potential V
2
(x). In general, since V
2
(x) is not ana
lytically solvable, we must solve an approximate problem and calculate the
corrections perturbatively. The use of SUSY, coupled with the observation
that the magnitude of t is in general small, allows us to construct a suit
able unperturbed problem. Consider the Schr¨odinger equation for V
2
(x) and
E = 0. From supersymmetry [eq. (388)] we see immediately that 1/ψ
0
is a
solution. Since t is small, we expect this solution to be an excellent approx
imation to the correct eigenfunction for small values of x. However, 1/ψ
0
is not normalizable and hence is not acceptable as a starting point for per
turbation theory. One possibility is to regularize the behavior artiﬁcially at
large [ x [ [31]. This procedure is cumbersome and results in perturbation
corrections to the leading term which are substantial. Instead, we choose
for our unperturbed problem the second linearly independent solution of the
Schr¨odinger equation given by [184]
φ(x) =
1
ψ
0
∞
x
ψ
2
0
(x
′
)dx
′
, x > 0 , (394)
and φ(x) = φ(−x) for x < 0. Clearly, φ(x) is well behaved at x = ±∞ and
closely approximates 1/ψ
0
at small x; thus we expect it to be an excellent
approximation of the exact ground state wave function of V
2
(x) for all values
of x. The derivative of φ(x) is continuous except at the origin, where, unlike
the exact solution, it has a discontinuity φ [
0+
−φ [
0−
= −2ψ
0
(0). Hence φ(x)
is actually a zeroenergy solution of the Schr¨odinger equation for a potential
V
0
(x) given by
V
0
(x) = V
2
(x) −4ψ
2
0
(0)δ(x) , (395)
where we have assumed that ψ
0
(x) is normalized. We calculate the perturba
tive corrections to the ground state energy using ∆V = +4ψ
2
0
(0)δ(x) as the
perturbation. Note that the coeﬃcient multiplying the δfunction is quite
small so that we expect our perturbation series to converge rapidly.
For the case of a symmetric potential such as V
2
(x), the perturbative
corrections to the energy arising from ∆V can be most simply calcualted
by use of the logarithmic perturbationtheory [185] formulation of the usual
RayleighSchr¨odinger series. The ﬁrst and second order corrections to the
unperturbed energy E = 0 are
96
E
(1)
=
1
2ξ(0)
, ξ(x) ≡
∞
x
φ
2
(x
′
)dx, E
(2)
= −2
∞
0
¸
E
(1)
ξ(x
′
)
φ(x
′
)
¸
2
dx
′
.
(396)
For our example, we numerically evaluate these corrections in order to
obtain an estimate of t. The results are shown in Fig. 9.3 for values of
x
0
≤ 2. Estimates of t correct to ﬁrst, second, and third order calculated
from logarithmic perturbation theory are compared with the exact result for
V
2
, obtained by the RungeKutta method. The asymptotic behavior of t
given by eq. (393) is also shown. This asymptotic form can also be recovered
from eq. (396) by a suitable approximation of the integrand in the largex
0
limit. Even for values of x
0
≤ 1/
√
2, in which case V
1
(x) does not exhibit a
doublewell structure, the approximation technique is surprisingly good. The
thirdorder perturbative result and the exact result are indistinguishable for
all values of x
0
.
In conclusion, we have demonstrated how SUSY can be used to calculate
t, the energy splitting for a doublewell potential. Rather than calculating
this splitting as a diﬀerence between the lowest lying two states of V
1
(x),
one can instead develop a perturbation series for the ground state energy t
of the partner potential V
2
(x). By choosing as an unperturbed problem the
potential whose solution is the normalizable zeroenergy solution of V
2
(x),
we obtain a very simple δ−function perturbation which produces a rapidly
convergent series for t [34]. The procedure is quite general and is applicable
to any arbitrary doublewell potential, including asymmetric ones [181]. The
numerical results are very accurate for both deep and shallow potentials.
9.4 Supersymmetry and LargeN Expansions
The largeN method, where N is the number of spatial dimensions, is a pow
erful technique for analytically determining the eigenstates of the Schr¨odinger
equation, even for potentials which have no small coupling constant and hence
not amenable to treatment by standard perturbation theory [186, 187, 188,
190, 191, 192, 193, 194, 195, 200]. A slightly modiﬁed, physically motivated
approach, called the “shifted largeN method” [196, 197, 198, 199] incorpo
rates exactly known analytic results into 1/N expansions, greatly enhancing
their accuracy, simplicity and range of applicability. In this subsection, we
97
will descibe how the rate of convergence of shifted 1/N expansions can be
still further improved by using the ideas of SUSY QM [77].
The basic idea in obtaining a 1/N expansion in quantum mechanics con
sists of solving the Schr¨odinger equation in N spatial dimensions, assuming
N to be large, and taking 1/N as an “artiﬁcially created” expansion parame
ter for doing standard perturbation theory. At the end of the calculation, one
sets N = 3 to get results for problems of physical interest in three dimensions.
For an arbitrary spherically symmetric potential V (r) in N dimensions,
the radial Schr¨odinger equation contains the eﬀective potential
V
eﬀ
(r) = V (r) +
(k −1)(k −3)¯ h
2
8mr
2
, k = N + 2l . (397)
It is important to note that N and l always appear together in the combi
nation k = N +2l. This means that the eigenstates, which could in principle
have depended on the three quantities N, l, n, in fact only depend on k and
n, where n is the radial quantum number which can take values 0,1,2,... One
now makes a systematic expansion of eigenstates in the parameter 1/k, where
k = k − a. Of course, for very large values of N, the two choices k and k
are equivalent. However, for N = 3 dimensions, a properly chosen shift a
produces great improvements in accuracy and simplicity. At small values of
r, the n = 0 wave function ψ
0
(r) has the behavior r
(k−1)/2
. If one sets
ψ
0
(r) = r
(k−1)/2
Φ
0
(r) (398)
where Φ
0
(r) is ﬁnite at the origin, then eq. (398) readily gives the supersym
metric partner potential of V
eﬀ
(r) to be
V
2
(r) = V (r) +
(k + 1)(k −1)¯ h
2
8mr
2
−
¯ h
2
m
d
2
dr
2
lnΦ
0
(r) (399)
V
2
(r) and V
eﬀ
(r) have the same energy values [except for the ground
state]. However, largeN expansions with the partner potential V
2
(r) are
considerably better since the angular momentum barrier in eq. (399) is given
by (k
′
−1)(k
′
−3)¯ h
2
/8mr
2
, where k
′
= k + 2. So, eﬀectively, one is working
in two extra spatial dimensions! Thus, for example, in order to calculate the
energy of the state with quantum numbers k, n of V
eﬀ
(r) one can equally
well use k
′
= k + 2, n −1 with V
2
(r). To demonstrate this procedure, let us
give an explicit example. Using the usual choice of units ¯ h = 2m = 1, the
98
swave Hulthen eﬀective potential in three dimensions and its ground state
wave function are:
V
H
eﬀ
(r) = −
2δe
−δr
1 −e
−δr
+
(2 −δ)
2
4
, ψ
0
(r) ∼ (1 −e
−δr
) e
−
(2−δ)r
2
, (400)
where the parameter δ is restricted to be less than 2. The supersymmetric
partner potential is
V
H
2
= V
H
1
+
2δ
2
e
−δr
(1 −e
−δr)
2
. (401)
As r tends to zero, V
H
2
goes like 2r
−2
, which as mentioned above, corre
sponds to the angular momentum barrier (k
′
−1)(k
′
−3)¯ h
2
/8mr
2
for k
′
= 5
(N = 5, l = 0). Let us compute the energy of the ﬁrst excited state of
V
H
eﬀ
(r). For the choice δ = 0.05, the exact answer is known to be 0.748125
[201]. The results upto leading, second and third order using a shifted 1/N
expansion for V
H
eﬀ
are 0.747713, 0.748127 and 0.748125 . The corresponding
values using the supersymmetric partner potential V
H
2
are all 0.748125 ! It is
clear that although excellent results are obtained with the use of the shifted
1/N expansion for the original potential V
H
eﬀ
(r) in three dimensions, even
faster convergence is obtained by using the supersymmetric partner poten
tial, since we are now eﬀectively working in ﬁve dimensions instead of three.
Thus, SUSY has played an important role in making a very good expansion
even better [77]. In fact, for many applications, considerable analytic simpli
ﬁcation occurs since it is suﬃcient to just use the leading term in the shifted
1/N expansion for V
2
(r). Other examples can be found in ref. [77].
10 Pauli Equation and Supersymmetry
So far, we have discussed the concept of SUSY for Schr¨odinger Hamiltonians
in onedimension and for central potentials in higher dimensions which are
essentially again onedimensional problems. In this section we shall show that
the concept of SUSY can also be applied to some geniune twodimensional
problems. In particular we show that the Pauli Hamiltonian which deals with
the problem of a charged particle in external magnetic ﬁeld can always be put
in SUSY form provided the gyromagnetic ratio is equal to two. It is worth
emphasizing here that not only the uniform magnetic ﬁeld problem (i.e. the
famous Landau level problem) but the nonuniform magnetic ﬁeld problems
99
can also be put in SUSY form. Using the concepts of SUSY and shape
invariance we show that some of the nonuniform magnetic ﬁeld problems can
be solved analytically.
The Pauli Hamiltonian for the motion of a charged particle in external
magnetic ﬁeld in two dimensions is given by (¯ h = 2m = e = 1)
H = (p
x
+ A
x
)
2
+ (p
y
+ A
y
)
2
+
g
2
(
∇
A)
z
σ
z
(402)
It is easily seen that this H along with the supercharges Q
1
and Q
2
deﬁned
by [35, 39]
Q
1
=
1
√
2
[−(p
y
+ A
y
)σ
x
+ (p
x
+ A
x
)σ
y
]
Q
2
=
1
√
2
[(p
x
+ A
x
)σ
x
+ (p
y
+ A
y
)σ
y
] (403)
satisfy the N = 1 supersymmetry algebra provided the gyromagnetic ratio g
is two
¦Q
α
, Q
β
¦ = Hδ
αβ
, [H, Q
α
] = 0; α, β = 1, 2 (404)
It is interesting to note that SUSY ﬁxes the value of g. The above Hamil
tonian has an additional 0(2) ⊗0(2) symmetry coming from σ
z
and an 0(2)
rotation in the A
1
, A
2
plane (A
1
≡ p
x
+ A
x
; A
2
≡ p
y
+ A
y
). Let us ﬁrst
consider the problem in an asymmetric gauge i.e.
A
y
(x, y) = 0, A
x
(x, y) = W(y) (405)
where W(y) is an arbitrary function of y. In this case the Pauli Hamiltonian
takes the form
H = (p
x
+ W(y))
2
+ p
2
y
−W
′
(y)σ
z
(406)
Since this H does not depend on x, hence the eigenfunction
˜
ψ can be factor
ized as
˜
ψ(x, y) = e
ikx
ψ(y) (407)
where k is the eigenvalue of the operator p
x
(−∞≤ k ≤ ∞). The Schr¨odinger
equation for ψ(y) then takes the form
[−
d
2
dy
2
+ (W(y) + k)
2
−W
′
(y)σ]ψ(y) = Eψ(y) (408)
100
where σ(= ±1) is the eigenvalue of the operator σ
z
. Thus we have reduced
the problem to that of SUSY in one dimension with superpotential W(y) +k
where W(y) must be independent of k. This constraint on W(y) strongly
restricts the allowed forms of shape invariant W(y) for which the spectrum
can be written down algebraically. In particular from Table 4.1 we ﬁnd that
the only allowed forms are (i) W(y) = ω
c
y + c
1
(ii) W(y) = a tanhy + c
1
(iii) W(y) = a tany + c
1
(iv) W(y) = c
1
− c
2
exp (−y) for which W(y) can
b e written in terms of simple functions and for which the spectrum can be
written down algebraically [202, 101]. In particular when
W(y) = ω
c
y + c
1
(409)
which corresponds to the uniform magnetic ﬁeld, then the energy eigenvalues
known as Landau levels are given by
E
n
= (2n + 1 +σ)ω
c
; n = 0, 1, 2... (410)
Note that the ground state and all excited states are inﬁnitefold degenerate
since E
n
does not depend on k which assumes a continuous sequence of value
(−∞≤ k ≤ ∞).
The magnetic ﬁeld corresponding to the other choices of W are (ii) B =
−asech
2
y (iii)B = −a sec
2
y(
−π
2
≤ y ≤
π
2
) (iv) B = +c
2
exp (−y) and as
mentioned above, all these problems can be solved algebrically by using the
results of Sec. 4.
Let us now consider the same problem in the symmetric gauge. We choose
A
x
= ω
c
yf(ρ), A
y
= −ω
c
xf(ρ) (411)
where ρ
2
= x
2
+ y
2
and ω
c
is a constant. The corresponding magnetic ﬁeld
B
z
is then given by
B
z
(x, y) ≡ ∂
x
A
y
−∂
y
A
x
= −2ω
c
f(ρ) −ω
c
ρf
′
(ρ) (412)
In this case the Hamiltonian (402) can be shown to take the form
H = −(
d
2
dx
2
+
d
2
dy
2
) + ω
2
c
ρ
2
f
2
−2ω
c
fL
z
−(2ω
c
f + ω
c
ρf
′
)σ
z
(413)
101
where L
z
is the zcomponent of the orbital angular momentum operator.
Clearly the corresponding Schr¨odinger problem can be solved in the cylindri
cal coordinates ρ, φ. In this case, the eigenfunction ψ(ρ, φ) can be factorized
as
ψ(ρ, φ) = R(ρ)e
imφ
/
√
ρ (414)
where m = 0, ±1, ±2, .. is the eigenvalue of L
z
. In this case the Schr¨odinger
equation for R(ρ) takes the form
[−
d
2
dρ
2
+ ω
2
c
ρ
2
f
2
−2ω
c
fm−(2ω
c
f + ω
c
ρf
′
(ρ))σ +
m
2
−1/4
ρ
2
]R(ρ) = ER(ρ)
(415)
where σ(= ±1) is the eigenvalue of the operator σ
z
. On comparing with
Table 4.1, it is easily checked that there is only one shape invariant potential
(f(ρ) = 1) for which the spectrum can be written down algebraically. This
case again corresponds to the famous Landau level problem i.e. it corresponds
to the motion of a charged particle in the x − y plane and subjected to a
uniform magnetic ﬁeld (in the symmetric gauge) in the zdirection. The
energy eigenvalues are
E
n
= 2(n + m+ [ m [)ω
c
; n = 0, 1, 2... (416)
so that all the states are again inﬁnitefold degenerate. It is worth noting
that the nonuniform magnetic ﬁelds can also give this equispaced spectrum
[203]. However, they do so only for one particular value of m while for other
values of m, the spectrum is in general not equispaced.
The fact that in this example there are inﬁnite number of degenerate
ground states with zero energy can be understood from the AharonovCasher
theorem [204] which states that if the total ﬂux deﬁned by Φ =
B
z
dxdy =
n +ǫ(0 ≤ ǫ < 1) then there are precisely n −1 zero energy states. Note that
in our case Φ is inﬁnite.
11 Supersymmetry and the Dirac Equation
There have been many applications of SUSY QM in the context of the Dirac
equation. In view of the limitations of space, we shall concentrate on only a
few of these applications [100]. In particular, we discuss the supersymmetric
structure of the Dirac Hamiltonian (H
D
) and show how the methods used
102
to obtain analytical solutions of the Schr¨odinger equation can be extended
to the Dirac case. First of all, we consider the Dirac equation in 1+1 di
mensions with Lorentz scalar potential φ(x). We show that whenever the
onedimensional Schr¨odinger equation is analytically solvable for a potential
V (x), then there always exists a corresponding Dirac scalar potential prob
lem which is also analytically solvable [101]. It turns out that, on the one
hand, φ(x) is essentially the superpotential of the Schr¨odinger problem and
on the other hand, it can be looked upon as the kink solution of a scalar
ﬁeld theory in 1+1 dimensions. Next we discuss the celebrated problem of
the Dirac particle in a Coulomb ﬁeld [103] and show that its eigenvalues and
eigenfunctions can be simply obtained by using the concept of SUSY and
shape invariance as developed in Secs. 2 and 4. We also discuss the problem
of the Dirac equation in an external magnetic ﬁeld in two dimensions and
show that there is always a supersymmetry in the problem in the massless
case. We also classify a number of magnetic ﬁeld problems whose solutions
can be algebraically obtained by using the concepts of SUSY and shape in
variance [101]. In addition, we show that the Euclidean Dirac operator in
four dimensions, in the background of gauge ﬁelds, can always be cast in
the language of SUSY QM. Finally, we discuss the path integral formulation
of the fermion propagator in an external ﬁeld and show how the previously
known results for the constant external ﬁeld can be very easily obtained by
using the ideas of SUSY [101].
11.1 Dirac Equation With Lorentz Scalar Potential
The Dirac Lagrangian in 1+1 dimensions with a Lorentz scalar potential
φ(x) is given by
L = iψγ
µ
∂
µ
ψ −ψψφ. (417)
The scalar potential φ(x) can be looked upon as the static, ﬁnite energy, kink
solution corresponding to the scalar ﬁeld Lagrangian
L
φ
=
1
2
∂
µ
φ(x)∂
µ
φ(x) −V (φ). (418)
Such models have proved quite useful in the context of the phenomenon
of fermion number fractionalization [205, 206, 207] which has been seen in
certain polymers like polyacetylene. Further, a variant of this model is also
103
relevant in the context of supersymmetric ﬁeld theories in 1+1 dimensions
[208, 209]. Note that the coupling constant in eqs. (417) and (418) has been
absorbed in φ and V (φ) respectively.
The Dirac equation following from eq. (417) is
iγ
µ
∂
µ
ψ(x, t) −φ(x)ψ(x, t) = 0. (419)
Let
ψ(x, t) = exp(−iωt)ψ(x) (420)
so that the Dirac equation reduces to
γ
0
ωψ(x) + iγ
1
dψ(x)
dx
−φ(x)ψ(x) = 0. (421)
We choose
γ
0
= σ
1
=
0 1
1 0
, γ
1
= iσ
3
=
i 0
0 −i
, ψ(x) =
ψ
1
(x)
ψ
2
(x)
(422)
so that we have the coupled equations
Aψ
1
(x) = ωψ
2
(x)
A
†
ψ
2
(x) = ωψ
1
(x), (423)
where
A =
d
dx
+ φ(x), A
†
= −
d
dx
+ φ(x). (424)
We can now easily decouple these equations. We get
A
†
Aψ
1
= ω
2
ψ
1
, AA
†
ψ
2
= ω
2
ψ
2
. (425)
On comparing with the formalism of Sec. 2, we see that there is a supersym
metry in the problem and φ(x) is just the superpotential of the Schr¨odinger
formalism. Further ψ
1
and ψ
2
are the eigenfunctions of the Hamiltonians
H
−
≡ A
+
A and H
+
≡ AA
†
respectively with the corresponding potentials
being V
∓
(x) ≡ φ
2
(x) ∓φ
′
(x). The spectrum of the two Hamiltonians is thus
degenerate except that H
−
(H
+
) has an extra state at zero energy so long
as φ(x → ±∞) have opposite signs and φ(x → +∞) > 0(< 0). Using the
results of the Secs. 2 and 4 we then conclude that for every SIP given in
104
Table 4.1 there exists an analytically solvable Dirac problem with the corre
sponding scalar potential φ(x) being the superpotential of the Schr¨odinger
problem. In particular, using the reﬂectionless superpotential given by
W(x) = ntanh x (426)
one can immediately construct perfectly transparent Dirac potentials with
n bound states [210]. Further, using the results for the SIP with scaling
ansatz (a
2
= qa
1
) [58, 59], one can also construct perfectly transparent Dirac
potentials with an inﬁnite number of bound states.
11.2 Supersymmetry and the Dirac Particle in a Coulomb
Field
The Dirac equation for a charged particle in an electromagnetic ﬁeld is given
by (e = ¯ h = c = 1)
[iγ
µ
(∂
µ
+ iA
µ
) −m]ψ = 0 (427)
For a central ﬁeld i.e.
A = 0 and A
0
(x, t) = V (r), this equation can be
written as [211]
i
∂ψ
∂t
= Hψ = ( α p + βm+ V )ψ (428)
where
α
i
=
0 σ
i
σ
i
0
, β =
1 0
0 −1
(429)
with σ
i
being the Pauli matrices. For central ﬁelds, this Dirac equation can
be separated in spherical coordinates and ﬁnally for such purposes as the
computation of the energy levels one only needs to concentrate on the radial
equations which are given by [211]
G
′
(r) +
kG
r
−(α
1
−V )F = 0
F
′
(r) −
kF
r
−(α
2
+ V )G = 0 (430)
where
α
1
= m+ E, α
2
= m−E (431)
105
and G
k
is the “large” component in the nonrelativistic limit. Of course
the radial functions G
k
and F
k
must be multiplied by the appropriate two
component angular eigenfunctions to make up the full fourcomponent solu
tions of the Dirac equation [211]. These coupled equations are in general not
analytically solvable; one of the few exceptions being the case of the Dirac
particle in a Coulomb ﬁeld for which
V (r) = −
γ
r
, γ = Ze
2
(432)
We now show that the Coulomb problem can be solved algebraically by
using the ideas of SUSY and shape invariance. To that end, we ﬁrst note
that in the case of the Coulomb potential, the coupled equations (430) can
be written in a matrix form as
G
′
k
(r)
F
′
k
(r)
+
1
r
k −γ
γ −k
G
k
F
k
=
0 α
1
α
2
0
G
k
F
k
(433)
where k is an eigenvalue of the operator −(σ
L + 1) with the allowed values
k = ±1, ±2, ±3..., and satisﬁes [ k [= J+
1
2
. Following Sukumar [103], we now
notice that the matrix multiplying 1/r can be diagonalized by multiplying it
by a matrix D from the left and D
−1
from the right where
D =
k + s −γ
−γ k + s
, s =
k
2
−γ
2
. (434)
On multiplying eq. (433) from the left by the matrix D and introducing the
new variable ρ = Er leads to the pair of equations
A
˜
F = (
m
E
−
k
s
)
˜
G ,
A
†
˜
G = −(
m
E
+
k
S
)
˜
F (435)
where
˜
G
˜
F
= D
G
F
(436)
and
A =
d
dρ
−
s
ρ
+
γ
s
, A
†
= −
d
dρ
−
s
ρ
+
γ
s
. (437)
106
Thus we can easily decouple the equations for
˜
F and
˜
G thereby obtaining
H
−
˜
F ≡ A
+
A
˜
F = (
k
2
s
2
−
m
2
E
2
)
˜
F,
H
+
˜
G ≡ AA
+
˜
G = (
k
2
s
2
−
m
2
E
2
)
˜
G. (438)
We thus see that there is a supersymmetry in the problem and H
±
are shape
invariant supersymmetric partner potentials since
H
+
(ρ; s, γ) = H
−
(ρ; s + 1, γ) +
γ
2
s
2
−
γ
2
(s + 1)
2
. (439)
On comparing with the formalism of Sec. 4 it is then clear that in this case
a
2
= s + 1, a
1
= s, R(a
2
) =
γ
2
a
2
1
−
γ
2
a
2
2
(440)
so that the energy eigenvalues of H
−
are given by
(
k
s
)
2
−(
m
2
E
2
n
) ≡ E
(−)
n
=
n+1
¸
k=2
R(a
k
) = γ
2
(
1
s
2
−
1
(s + n)
2
). (441)
Thus the Coulomb bound state energy eigenvalues E
n
are given by
E
n
=
m
[1 +
γ
2
(s+n)
2
]
1/2
, n = 0, 1, 2, ... (442)
It should be noted that every eigenvalue of H
−
is also an eigenvalue of H
+
except for the ground state of H
−
which satisﬁes
A
˜
F = 0 =⇒
˜
F
0
(ρ) = ρ
s
exp(−γρ/s) (443)
Using the formalism for the SIP as developed in Sec. 4, one can also alge
braically obtain all the eigenfunctions of
˜
F and
˜
G.
Notice that the spectrum as given by eq. (442) only depends on [ k [
leading to a doublet of states corresponding to k =[ k [ and k = − [ k [ for
all positive n. However, for n = 0, only the negative value of k is allowed
and hence this is a singlet state.
107
11.3 SUSY and the Dirac Particle in a Magnetic Field
Let us again consider the Dirac equation in an electromagnetic ﬁeld as given
by eq. (427) but now consider the other case when the vector potential is
nonzero but the scalar potential is zero i.e. A
0
= 0,
A = 0. Now as shown
by Feynman and GellMann [212] and Brown [213], the solution of the four
component Dirac equation in the presence of an external electromagnetic
ﬁeld can be generated from the solution of a two component relativistically
invariant equation. In particular, if ψ obeys the two component equation
[(
P +
A)
2
+ m
2
+σ(
B + i
E)]ψ = (E + A
0
)
2
ψ (444)
then the four component spinors that are solutions of the massive Dirac
equation are generated from the two component ψ via
ψ
D
=
(σ(
P +
A) + E −A
0
+ m)ψ
(σ(
P +
A) + E −A
0
−m)ψ
. (445)
Thus, in order to solve the Dirac equation, it is suﬃcient to solve the much
simpler twocomponent eq.(444) and then generate the corresponding Dirac
solutions by the use of eq. (445). In the special case when the scalar po
tential A
0
(and hence
E) vanishes, the twocomponent equation then has
the canonical form of the Pauli equation describing the motion of a charged
particle in an external magnetic ﬁeld. If further, m = 0 and the motion is
conﬁned to two dimensions, then the Pauli eq. (444) exactly reduces to the
eq. (402) of the last section. Further, since
(H
D
)
2
= [ α(
P +
A)]
2
= H
Pauli
(446)
hence, there is a supersymmetry in the massless Dirac problem in external
magnetic ﬁelds in two dimensions since H
2
D
, Q
1
and Q
2
(see eq. (403) satisfy
the SUSY algebra as given by eq. (404). Clearly, this supersymmetry will
also be there in two Euclidean dimensions. We can now immediately borrow
all the results of the last section. In particular, it follows that if the total ﬂux
Φ(=
B
z
dxdy)=n + ǫ(0 ≤ ǫ < 1) then there are precisely n −1 zero modes
of the massless Dirac equation in two dimensions in the background of the
external magnetic ﬁeld B (B ≡ B
z
) [204, 214]. Further, in view of eqs. (444)
and (445) we can immediately write down the exact solution of the massless
Dirac equation in an external magnetic ﬁeld in two dimensions in all the four
108
situations discussed in the last section when the gauge potential depended
on only one coordinate (say y). Further using the results of that section,
one can also algebraically obtain the exact solution of the Dirac equation in
an uniform magnetic ﬁeld in the symmetric gauge when the gauge potential
depends on both x and y.
Even though there is no SUSY, exact solutions of the Pauli and hence
the Dirac equation are also possible in the massive case. On comparing
the equations as given by (444) (with A
0
= 0,
E = 0) and (402) it is clear
that the exact solutions in the massive case are simply obtained from the
massless case by replacing E
2
by E
2
− m
2
. Summarizing, we conclude that
the exact solutions of the massive (as well as the massless) Dirac equation
in external magnetic ﬁeld in two dimensions can be obtained algebraically
in case the magnetic ﬁeld B(≡ B
Z
) has any one of the following four forms
(i) B = constant (ii) B = −a sech
2
y (iii) B = −a sec
2
y(−π/2 ≤ y ≤ π/2)
(iv) B = −c
2
exp(−y). Further, in the uniform magnetic ﬁeld case, the
solution can be obtained either in the asymmetric or in the symmetric gauge
[202, 101].
11.4 SUSY and the Euclidean Dirac Operator
In the last few years there has been a renewed interest in understanding
the zero modes and the complete energy spectrum of the square of the Dirac
operator for a Euclidean massless fermionic theory interacting with the back
ground gauge ﬁelds [25, 28, 215, 216]. Let us ﬁrst show that there is always
a supersymmetry in the problem of the Euclidean massless Dirac operator in
the background of the gauge ﬁelds. On deﬁning
Q
±
=
1
2
(1 ±γ
5
)γ
µ
D
µ
(447)
where
D
µ
= ∂
µ
+ iA
µ
, (448)
one ﬁnds that the Dirac operator can be written as
γ
µ
D
µ
= Q
+
+ Q
−
(449)
Let us take the following representation of the Euclidean γ matrices
γ
0
=
0 i
−i 0
, γ
i
=
0 σ
i
σ
i
0
, γ
5
=
1 0
0 −1
(450)
109
where σ
i
are the usual Pauli matrices. It is then easily seen that the operators
Q
+
, Q
−
and H ≡ −(γ
µ
D
µ
)
2
satisfy the usual N = 1 SUSY algebra
H = −(γ
µ
D
µ
)
2
= ¦Q
+
, Q
−
¦, [H, Q
+
] = 0 = [H, Q
−
] (451)
This supersymmetry is popularly known as chiral supersymmetry. In view
of the above representation of the γ matrices, we ﬁnd
γ
µ
D
µ
=
0 iD
0
+σ
D
−iD
0
+σ
D 0
=
0 L
L
†
0
(452)
and hence
−H = (γ
µ
D
µ
)
2
= D
µ
D
µ
+
σ(
B +
E) 0
0 σ(
B −
E)
=
LL
†
0
0 L
†
L
(453)
where we have used the convention
F
µν
= ∂
µ
A
ν
−∂
ν
A
µ
, B
i
=
1
2
ǫ
ijk
F
jk
, E
i
= F
0i
(454)
Using the techniques of Secs. 2 and 4, one can algebraically obtain the
eigenvalues and the eigenfunctions of H for few special cases [101]. Further,
once we have an eigenfunction of H with a nonzero eigenvalue E
1
, then one
can easily obtain the eigenfunctions φ
±
of γ
µ
D
µ
with eigenvalues ±(E
1
)
1/2
by the construction
φ
±
= (γ
µ
D
µ
±(E
1
)
1/2
)ψ (455)
Finally, it is worth pointing out that in case the gauge potential A
µ
has
the special form
A
µ
= f
µν
∂
ν
χ (456)
where the constant 4 by 4 matrix f
µν
has the properties:
f
µν
= −f
νµ
, f
µν
f
να
= −δ
µα
(457)
then there exists another breakup of the Dirac operator γ
µ
D
µ
which leads to
the so called complex supersymmetry [101].
110
11.5 Path Integral Formulation of the Fermion Prop
agator
We have seen that there is always a supersymmetry associated with the
Euclidean Dirac operator in four as well as two dimensions in the background
of the gauge ﬁelds. This SUSY was ﬁrst successfully exploited in the context
of the study of the chiral anomalies [217]. In a nut shell, the solvability of the
Dirac operator is related to the ability to integrate the path integral. The
quantities one wishes to calculate are the Green’s functions
G(x, x; τ) =< x [ e
−Hτ
[ x
′
>[
x=x
′ (458)
and
G
5
(x, x; τ) =< x [ Tr(γ
5
e
−Hτ
) [ x
′
>[
x=x
′ (459)
By using Schwinger’s proper time formalism we can determine S(x, x; A)
from G. Similarly, the index Z
5
which in the limit τ = 0 is related to the
chiral anomaly, is just the spatial integral over G
5
.
Here we would like to show that by introducing the fermionic degrees
of freedom, the path integral for the Dirac operator, which is initially a
path ordered integral gets reduced to an ordinary path integral. This trick
was ﬁrst introduced by Rajeev [218] who was interested in reformulating
quantum electrodynamics as a supersymmetric theory of loops. Once we
introduce the fermionic variables then G(G
5
) is determined by choosing the
antiperiodic (periodic) boundary conditions for the fermions. By integrating
over the fermionic degrees of freedom, a purely bosonic path integral is then
obtained. For the case of a constant external ﬁeld strength F
µν
, the bosonic
path integral is a Gaussian, which allows one to trivially obtain G and G
5
.
Let us consider the square of the Dirac operator i.e. (γ
µ
D
µ
)
2
. It can also
be written as
H = (P + A)
2
−
1
2
σ
µν
F
µν
(460)
where
σ
µν
=
1
2i
[γ
µ
, γ
ν
]. (461)
If we do not introduce auxiliary fermions then the related matrix valued
Lagrangian would be
L(x, ˙ x) =
1
4
˙ x
µ
˙ x
µ
−iA
µ
˙ x
µ
+
1
2
σ
µν
F
µν
(462)
111
and we would obtain for Feynman’s path integral representation
< x [ e
−Hτ
[ x >=
Tx
µ
(τ)P exp[−
τ
0
dτ
′
L(x, ˙ x)] (463)
where P denotes path ordering. Now, taking analogy from SUSY quantum
mechanics, it has been noticed [218, 101] that one can introduce the Grass
man variables ψ
µ
via
ψ
µ
=
1
√
2
γ
µ
, ¦ψ
µ
, ψ
ν
¦ = δ
µν
. (464)
Then H can be written as
H = (P + A)
2
+
i
2
ψ
µ
F
µν
ψ
ν
(465)
and hence the Lagrangian now becomes
L
ss
=
1
4
˙ x
µ
˙ x
µ
−iA
µ
˙ x
µ
−
i
2
ψ
µ
(∂
τ
δ
µν
+ F
µν
)ψ
ν
(466)
which is invariant under the SUSY transformations
δx
µ
= −iǫψ
µ
; δψ
µ
= ǫ ˙ x
µ
. (467)
We now obtain for the path integral
G(x, x; τ) =
Tx
µ
(t)Tψ
µ
exp[−
τ
0
dτ
′
L
ss
(x, ˙ x)], (468)
where we impose antiperiodic boundary conditions on the fermion at 0 and τ.
Since the fermionic path integral is quadratic, we can perform the functional
integral over the fermionic degrees of freedom exactly for arbitrary F
µν
.
The result of the fermionic path integral is
Det
1/2
(i∂
τ
δ
µν
+ F
µν
)
Det
′
1/2
(i∂
τ
δ
µν
)
, (469)
where the prime denotes the omission of the zero mode. To evaluate the
determinant one puts F
µν
in skew diagonal form:
F
µν
=
0 x
1
0 0
−x
1
0 0 0
0 0 0 x
2
0 0 −x
2
0
¸
¸
¸
¸
¸
. (470)
112
and imposes either periodic or antiperiodic boundary conditions. One obtains
for the fermion determinant using antiperiodic boundary conditions relevant
for G
¸
i
cosh
1
2
τ
0
x
i
dτ. (471)
and for the fermion determinant with periodic boundary conditions relevant
for the determination of the index:
(
2
τ
)
n
¸
i
[−sinh
1
2
τ
0
x
i
dτ
′
]. (472)
Thus for any arbitrary external ﬁeld one can explicitly perform the fermionic
path integral and is left with the purely bosonic path integral which, for ex
ample, could by performed numerically by Monte Carlo techniques. For the
particular case of a constant external ﬁeld one can further explicitly perform
the resulting bosonic path integral. When
A
µ
= −
1
2
F
µν
x
ν
, (473)
with F
µν
being a constant matrix, one obtains for the remaining bosonic path
integral:
¸
i
1
2
x
i
sinh
x
i
τ
2
(474)
G and G
5
are easy to calculate in two and four Euclidean dimensions for
the constant external ﬁeld case [101] using this method. In two dimensions,
where only F
01
= B exists one has that (x
1
= B) [101]
G = Btanh(Bτ); G
5
= B =
1
2
.ǫ
µν
F
µν
(475)
It is worth remarking here that the same result has been obtained with
more diﬃculty by Akhoury and Comtet [16]. In four dimensions one has
a simpler derivation of the result of Schwinger [219] rewritten in Euclidean
space. One ﬁnds that
x
1
= [F + (F
2
−G
2
)
1/2
]
1/2
, x
2
= [F −(F
2
−G
2
)
1/2
]
1/2
, (476)
where
F =
1
4
F
µν
F
µν
113
G =
1
4
F
µν
F
∗
µν
so that
¸
i
x
i
=
1
4
F
µν
F
∗
µν
. (477)
From this we obtain that:
G(x, x; τ) = F
µν
F
∗
µν
2
¸
i=1
cosh
x
i
τ
2
sinh
x
i
τ
2
G
5
(x, x; τ) =
1
4
F
µν
F
∗
µν
. (478)
Once one has obtained the Green’s function it is easy to reconstruct the
eﬀective action which allows one to obtain the rate of pair production from
a strong external electric ﬁeld. The eﬀective action is given by
S
eff
=
dxL(x), L(x) =
∞
ǫ
dss
−1
e
−m
2
s
G(s) (479)
12 Singular Superpotentials
So far, we have only considered nonsingular superpotentials W(x) which give
rise to the supersymmetric partner potentials V
1
(x) and V
2
(x). This choice
of the superpotential was based on the ground state wave function ψ
0
(x) of
V
1
(x) by the relation W(x) = −ψ
′
0
(x)/ψ
0
(x). In this section, we describe a
general procedure for constructing all possible superpotentials which yield a
given potential
˜
V (x) upto an additive constant [114]. This general procedure
is based on any arbitrary solution φ(x) of the Schr¨odinger equation for
˜
V (x),
rather than just the ground state wave function. We shall see that singular
superpotentials are given by W
φ
= −φ
′
/φ and the singularities are located
at the zeros of excited wave functions. Such singularities produce interesting
properties for the potentials V
1(φ)
(x) and V
2(φ)
(x) obtained via the Riccati
equations V
1(φ)
(x) = W
2
(x)−W
′
(x) and V
2(φ)
(x) = W
2
(x)+W
′
(x). Singular
superpotentials are responsible for negative energy eigenstates for V
1(φ)
(x)
and often give rise to a breakdown of the degeneracy of energy levels for
V
1(φ)
(x) and V
2(φ)
(x) [110, 112, 113, 220, 114]. Another interesting physical
phenomenon occurs if one considers a singular superpotential resulting from
a solution φ(x) for an energy E in the classical energy continuum. [114] The
114
isospectral family of V
1(φ)
(x) is then found to have bound states (normalised
eigenfunctions) at energy E. Thus SUSY QM provides a systematic pro
cedure for generating potentials possessing the purely quantum mechanical
phenomenon of bound states in the continuum[116, 117].
12.1 General Formalism, Negative Energy States and
Breakdown of the Degeneracy Theorem
Given any nonsingular potential
˜
V (x) with eigenfunctions ψ
n
(x) and eigen
values E
n
(n = 0, 1, 2, ...), let us now enquire how one can ﬁnd the most
general superpotential W(x) which will give
˜
V (x) upto an additive constant
[114]. To answer this question consider the Schr¨odinger equation for
˜
V (x):
−φ
′′
+
˜
V (x)φ = ǫφ (480)
where ǫ is a constant energy to be chosen later. For convenience, and without
loss of generality, we will always choose a solution φ(x) of eq. (480) which
vanishes at x = −∞. Note that whenever ǫ corresponds to one of the eigen
values E
n
, the solution φ(x) is the eigenfunction ψ
n
(x). If one deﬁnes the
quantity W
φ
= −φ
′
/φ and takes it to be the superpotential, then clearly the
partner potentials generated by W
φ
are
V
2(φ)
= W
2
φ
+ W
′
φ
, V
1(φ)
= W
2
φ
−W
′
φ
=
φ
′′
φ
=
˜
V (x) −ǫ , (481)
where we have used eq. (480) for the last step. The eigenvalues of V
1(φ)
are
therefore given by
E
n(φ)
= E
n
−ǫ. (482)
One usually takes ǫ to be the ground state energy E
0
and φ to be the ground
state wave function ψ
0
(x), which makes E
0(0)
= 0 and gives the familiar case
of unbroken SUSY. With this choice, the superpotential W
0
(x) = −ψ
′
0
/ψ
0
is nonsingular, since ψ
0
(x) is normalizable and has no zeros. The partner
potential V
2(φ)
has no eigenstate at zero energy since A
0
ψ
0
(x) = [d/dx +
W
0
(x)]ψ
0
(x) = 0; however, the remaining eigenvalues of V
2(φ)
are degenerate
with those of V
1(φ)
.
Let us now consider what happens for other choices of ǫ, both below and
above the ground state energy E
0
. For ǫ < E
0
, the solution φ(x) has no
115
nodes, and has the same sign for the entire range −∞ < x < +∞. The
corresponding superpotential W
φ
(x) is nonsingular. Hence the eigenvalue
spectra of V
1(φ)
and V
2(φ)
are completely degenerate and the energy eigenval
ues are given by eq. (482). In particular, E
0(φ)
= E
0
−ǫ is positive. Here, W
φ
has the same sign at z = ±∞, and we have the wellstudied case of broken
SUSY [10].
If the constant ǫ is chosen in the range E
0
< ǫ < E
1
, then the solution
φ(x) of eq. (482) will have only one node, say at the point x = x
s
. Near the
point x = x
s
, the node of φ(x) makes the superpotential W(x) singular
φ = a(x −x
s
), φ
′
= a, W(x) = −(x −x
s
)
−1
, (483)
and the partner potentials have the behavior [using eq. (481)]
V
1(φ)
= 0, V
2(φ)
= 2(x −x
s
)
−2
. (484)
It is wellknown [221, 222, 223] that for any singular potential V (x) =
λ(x −x
s
)
−2
, the behavior of the wave function at x = x
s
is governed by the
value of λ. For values λ >
3
4
, the potential V (x) has a “strong” singularity
which forces ψ(x
s
) = 0 and the range of x is eﬀectively broken into two
disjoint pieces x < x
s
and x > x
s
with no communication between them. For
−
1
4
< λ <
3
4
, the potential has a singularity of “intermediate” strength. It is
not strong enough to make ψ(x
s
) vanish and in fact the two regions x < x
s
and x > x
s
do communicate and one has the full range −∞ < x < +∞.
Here, in principle, one can have singular wave functions which require self
adjoint extensions [221]; in what follows, we will deal with regular solutions.
Finally, for λ < −
1
4
, the Hamiltonian is unbounded from below.
From eq. (484), it is clear that V
2(φ)
has a “strong” singularity which
makes ψ(x
s
) = 0. Thus, the problem of ﬁnding the eigenstates of V
2(φ)
is
really two separate problems; one in the range −∞ < x
s
and the other in
the range x
s
< x < +∞. Clearly, this is very diﬀerent from the range of
V
1(φ)
which is the whole real axis −∞ < x < +∞. Hence, in general, the
degeneracy theorem obtained from SUSY for the spectra of V
2(φ)
and V
1(φ)
is not valid. The above discussion can be readily extended to all values of
the constant ǫ  the only diﬀerence being the number of poles present. A
summary of results is given in Table 12.1.
So far we have considered superpotentials (both nonsingular and singular)
which give rise to a nonsingular potential V
1(φ)
. However, if the potential
116
V (x) itself has singularities, then so must all superpotentials which produce
it. Consider the case of a simple pole singularity at x = x
0
. Then, near
x = x
0
, the singular superpotential W(x) = g(x − x
0
)
−1
gives the following
behavior to the corresponding partner potentials:
V
1(φ)
=
g(g + 1)
(x −x
0
)
2
, V
2(φ)
=
g(g −1)
(x −x
0
)
2
. (485)
Then clearly, for g >
3
2
or g < −
3
2
, both V
1(φ)
and V
2(φ)
have “strong”
singularities and both have two disjoint regions x < x
0
and x < x
0
. Here,
one expects degenerate energy levels (corresponding to broken or unbroken
SUSY). Similarly, for −
1
2
< g < +
1
2
, both V
1(φ)
and V
2(φ)
have “intermediate”
strength singularities and the whole region −∞ < x < +∞ is valid for
both. Here again, one obtains degeneracy. However, for the two regions
−
3
2
< g < −
1
2
and
1
2
< g <
3
2
only one of the partner potentials has a
“strong” singularity, whereas the other has a singularity of “intermediate”
strength. Therefore, the two potentials have diﬀerent Hilbert spaces and,
in general, degeneracy is not there. The above discussion is borne out by
the work of Jevicki and Rodrigues [110] who consider a superpotential of the
form W(x) = g/x −x.
The general discussion of singular superpotentials, negative energy states
and breakdown of the degeneracy theorem is best illustrated with some spe
ciﬁc examples.
Consider the harmonic oscillator potential
˜
V (x) = x
2
. The energy eigen
values are E
n
= 2n and the ﬁrst few eigenstates are
ψ
0
(x) =
e
−x
2
/2
π
1/4
, ψ
1
(x) =
√
2 x e
−x
2
/2
π
1/4
, ψ
2
(x) =
(2x
2
−1)e
−x
2
/2
√
2 π
1/4
. (486)
Using the solution φ = ψ
0
(x), one gets the usual nonsingular superpotential
W
0
= −ψ
0
′/ψ
0
= x, leading to the following partner potentials and eigenval
ues:
V
1(0)
= x
2
−1, V
2(0)
= x
2
+ 1, E
(1)
n(0)
= 2n, E
(2)
n(0)
= 2n + 2, n = 0, 1, 2, ...
(487)
This is standard unbroken SUSY. The harmonic oscillator potential can also
be obtained if one starts from the solution φ = ψ
1
(x). One gets the singular
117
superpotential W
1
= −ψ
′
1
/ψ
1
= x − 1/x. It has a pole at x = 0 and is a
special case of the form discussed in Ref. [110]. The partner potentials are
V
1(1)
= x
2
−3, V
2(1)
= x
2
+
2
x
2
−1, (488)
and both are exactly solvable [54, 44]. The two eigenvalue spectra are
E
(1)
n(1)
= 2n −2, E
(2)
n(1)
= 4n + 4, n = 0, 1, 2, ... (489)
There occurs a negative energy state in V
1(1)
at E
(1)
0(1)
= −2. This is expected
since we chose φ to be the ﬁrst excited state and E
(1)
1(1)
= 0, which pushes
the ground state to a negative energy. Proceeding along the same lines, we
can get the harmonic oscillator potential using yet another superpotential.
Taking φ = ψ
2
(x), one gets W
2
= −ψ
′
2
/ψ
2
= x − 4x/(2x
2
− 1), which has
poles at x = ±1/
√
2. The corresponding potentials are
V
1(2)
= x
2
−5, V
2(2)
= x
2
−3 +
8(2x
2
+ 1)
(2x
2
−1)
2
. (490)
The eigenvalue spectrum of V
1(2)
is E
(1)
n(2)
= 2n − 4 indicating two negative
energy states at −4 and −2, which was anticipated. The potential V
2(2)
is
not analytically solvable. The “strong” singularities at x = ±1/
√
2 break
the xaxis into three disjoint regions −∞ < x < −1/
√
2, −1/
√
2 < x <
1/
√
2, 1/
√
2 < x < +∞ and the degeneracy theorem breaks down.
In summary, the potentials V
1(0)
and V
2(0)
give a realization of SUSY QM
with a nondegenerate zero energy ground state and pairing of excited states.
For the case of V
1(1)
and V
2(1)
the eigenvalues are given by eq. (489). There
is partial degeneracy of the spectrum [see Fig. 12.1] due to the fact that the
potentials are symmetric. The states which are missing in V
2(1)
are the even
parity states, since the 2/x
2
barrier requires the vanishing of wave functions
at x = 0. Finally, for the partner potentials V
1(2)
and V
2(2)
, degeneracy is
completely absent.
Our second example is the Morse potential
˜
V (x) = A
2
+B
2
e
−2αx
−B(α+
2A)e
−αx
. The superpotential based on the ground state is W
0
(x) = A −
Be
−αx
. For concreteness, we shall take A = 4, α = 1 and B = 1. The
corresponding energy eigenvalues are E
(1)
n(0)
= 16 − (4 − n)
2
; there are four
118
bound states with eigenvalues 0, 7, 12 and 15 and a continuum starting above
16. The two lowest eigenfunctions are [54]
ψ
0
(x) ∼ e
−(4x+e
−x
)
, ψ
1
(x) ∼ e
−(3x+e
−x
)
(7 −2e
−x
). (491)
The supersymmetric partner potentials constructed from the nonsingular su
perpotential W
0
(x) = 4 −e
−x
are
V
1(0)
= 16 −9e
−x
+ e
−2x
, V
2(0)
= 16 −7e
−x
+ e
−2x
. (492)
The spectrum of V
2(0)
is identical to that of V
1(0)
, except that there is no
state at zero energy. An alternative singular superpotential W
1
(x) which
also yields the Morse potential is
W
1
(x) = −
ψ
′
1
ψ
1
=
21 −15e
−x
+ 2e
−2x
7 −2e
−x
. (493)
There is a simple pole at x
s
= −ln(3.5). The partner potentials are
V
1(1)
= 9 −9e
−x
+ e
−2x
, V
2(1)
=
441 −567e
−x
+ 281e
2x
−56e
−3x
+ 4e
−4x
(7 −2e
−x
)
2
(494)
By construction, the potential V
1(1)
is the Morse potential and V
2(1)
has a
singularly at x
s
= −ln(3.5). Expanding V
2(1)
about the singular point gives
V
2(1)
∼ 2/(x − x
s
)
2
, which requires the wave function to vanish at x = x
s
.
This eﬀectively breaks the potential and the real axis into two parts: V
2(1)
(left) for −∞ < x < x
s
and V
2(1)
(right) for x
s
< x < +∞. The potentials
and their eigenstates are plotted in Figure 12.2. The potential V
1(1)
has
energy levels located at 7, 0, 5, 8 and a continuum above 9. As expected
from Table 12.1, there is a negative energy state at 7. We have calculated the
energy levels of V
2(1)
(right) and V
2(1)
(left) numerically. They are E
(2)
(1)
(right)
= 6.08, 8.65 with a continuum above 9 and E
(2)
(1)
(left) = 22.96, 50.69, 82.16,
116.8,.... The potential V
2(1)
(left) has an inﬁnite number of bound states
and as is obvious from the spectra, the degeneracy between V
2(1)
and V
1(1)
is
completely broken by the “strong” singularity in W
1
(x). Note that since the
Morse potential is asymmetric, no partial degeneracy remains. This is unlike
our ﬁrst example (harmonic oscillator) which had symmetric potentials and
V
2(1)
and V
1(1)
had a partial degeneracy.
119
Our last example is the reﬂectionless potential
˜
V (x) = A
2
− A(A +
1)sech
2
x. The motivation for considering this example is to make contact
with singular superpotentials considered by Casahorran and Nam [112]. The
ground state wave function is given by ψ
0
= (sechx)
A
. The nonsingular su
perpotential W
0
= Atanhx comes from ψ
0
. Using the property of shape
invariance [54, 44] one readily obtains ψ
1
(x) = tanh x(sechx)
A−1
as the wave
function for the ﬁrst excited state. If one takes φ = ψ
1
, the resulting su
perpotential is W
1
= −ψ
′
1
/ψ
1
= −2Acosech 2x + (A − 1) cothx, which
agrees with eq. (3.14) of ref. [112]. The energy eigenvalues for V
1(0)
are
E
(1)
n(0)
= A
2
−(A−n)
2
and one has unbroken SUSY. The potentials generated
by W
1
(x) are V
1(1)
= (A −1)
2
−A(A + 1)sech
2
x, V
2(1)
= (A −1)
2
− A(A −
1)sech
2
x+2cosech
2
x. These potentials are shape invariant [54, 44] and their
eigenvalues are E
(1)
n(1)
= (A−1)
2
−(A−n)
2
, E
(2)
n(1)
= (A−1)
2
−(A−2n−3)
2
, in
dicating partial degeneracy, as expected for symmetric potentials. Similarly,
our previous experience indicates that the partner potentials constructed
from W
2
(x) will have no degeneracy. Thus, we have not only reproduced
some families of singular superpotentials previously considered in the litera
ture, but given the general method to construct new ones.
In conclusion, we have shown how excited state wave functions can be
used to construct singular superpotentials in SUSY QM. Our method pro
vides a complete and uniﬁed picture of the origin of negative energy states
and the presence or absence of degeneracy. Although our technique is per
fectly general, our examples were taken from the class of shape invariant
potentials, since these are analytically solvable.
12.2 Bound States in the Continuum
In 1929, Von Neumann and Wigner [116] realized that it was possible to
construct potentials which have quantum mechanical bound states embed
ded in the classical energy continuum (BICs). Further developments, by
many authors [117, 224, 225, 226] have produced more examples and a bet
ter understanding of the kind of potential that can have such bound states,
although there is not as yet a fully systematic approach. These authors
have also suggested possible applications to atoms and molecules. Capasso
et al. [227] have recently reported direct evidence for BICs by constructing
suitable potentials using semiconductor heterostructures grown by molecular
120
beam epitaxy. Finally, it is interesting to note that BICs have found their
way into a recently written text [118].
In this subsection, we show how one can start from a potential with
a continuum of energy eigenstates, and use the methods of SUSY QM to
generate families of potentials with bound states in the continuum [BICs]
[115]. Basically, one is using the technique of generating isospectral potentials
(discussed in Sec. 7) but this time starting from states in the continuum. The
method preserves the spectrum of the original potential except it adds these
discrete BICs at selected energies. As illustrative examples, we compute and
graph potentials which have bound states in the continuum starting from a
null potential representing a free particle and the Coulomb potential.
(a) One Parameter Family of BICs
Consider any spherically symmetric potential V (r) which vanishes as r →
∞. The radial swave Schr¨odinger equation for the reduced wave function
u(r) (in units where ¯ h = 2m = 1) is
−u
′′
+ V (r) u(r) = E u(r), (495)
where we have scaled the energy and radial variables such that all quantities
are dimensionless. Eq. (495) has a classical continuum of positive energy
solutions which are clearly not normalizable.
As we have seen in Sec. 7, the Darboux [64] procedure for deleting and
then reinstating the ground state u
0
(r) of a potential V (r), generates a fam
ily of potentials
ˆ
V (r; λ) which have the same eigenvalues as V (r). These
isospectral potentials are labeled by a real parameter λ in the ranges λ > 0
or λ < −1. The isospectral potential
ˆ
V (r; λ) is given in terms of the orig
inal potential V (r) and the original ground state wave function u
0
(r) by
[68, 70, 72]
ˆ
V (r; λ) = V (r) −2[ln(I
0
+ λ)]
′′
= V (r) −
4u
0
u
′
0
I
0
+ λ
+
2u
4
0
(I
0
+ λ)
2
, (496)
where
I
0
(r) ≡
r
0
u
2
0
(r
′
)dr
′
. (497)
Except in this section, u
0
was taken to be the nodeless, normalizable
ground state wave function of the starting potential V (r). However, it is easy
to generalize the above equations to the case where u
0
(r) is any solution of
121
eq. (495) with arbitrary energy E
0
. If u
0
(r) has nodes, this leads to singular
superpotentials and to singularities in the partner potential V
2
(r). However,
when the original state at E
0
is reinserted, the resulting family of potentials
ˆ
V (r; λ) is free of singularities [114]. Our results are best summarized in the
following statement:
Theorem: Let u
0
(r) and u
1
(r) be any two nonsingular solutions of the
Schr¨odinger equation for the potential V (r) corresponding to arbitrarily se
lected energies E
0
and E
1
respectively. Construct a new potential
ˆ
V (r; λ) as
prescribed by eq. (496). Then, the two functions
ˆ u
0
(r; λ) =
u
0
(r)
I
0
+ λ
, (498)
and
ˆ u
1
(r; λ) = (E
1
−E
0
)u
1
+ ˆ u
0
W(u
0
, u
1
), (499)
[where W denotes the Wronskian, W(u
0
, u
1
) ≡ u
0
u
′
1
−u
1
u
′
0
] are solutions of
the Schr¨odinger equation for the new potential
ˆ
V (r; λ) corresponding to the
same energies E
0
and E
1
.
While the new potential in eq. (496) and the new wave functions in eq.
(498) were originally inspired by SUSY QM, the easiest proof of the above
theorem is by direct substitution. One simply computes −ˆ u
′′
i
+
ˆ
V (r; λ)ˆ u
i
(i=0,1), with the wave functions ˆ u
i
given in the theorem. After straightfor
ward but tedious algebraic manipulations, one gets E
i
ˆ u
i
, thus establishing
the theorem. The algebra is considerably simpliﬁed by using the following
identity for the Wronskian of two solutions of the Schr¨odinger equation:
d
dr
W(u
0
, u
1
) = (E
0
−E
1
)u
0
u
1
. (500)
Let us now take u
0
to be a scattering solution at a positive energy E
0
= k
2
of a potential V (r) which vanishes at r=∞. Taking u
0
(r = 0) = 0 satisﬁes
one of the required boundary conditions, but clearly u
0
oscillates as r →∞
and has an amplitude which does not decrease. Consequently, the integral
I
0
(r) in eq. (497) now grows like r at large r and ˆ u
0
is now square integrable
for λ > 0, while the original wave function u
0
was not. Negative values of λ
are no longer allowed. Therefore, we see that all the potentials
ˆ
V (r; λ) have
a BIC with energy E
0
. Note from eq. (498) that ˆ u
0
has the same zeros as
the original u
0
. At zeros of u
0
,
ˆ
V (r; λ) and V (r) are equal. All the other
122
oscillatory solutions of the Schr¨odinger equation with V (r) get transformed
into oscillatory solutions to the new Schr¨odinger equation with
ˆ
V (r; λ) with
the same energy. In particular, note that ˆ u
1
(r; λ) remains a nonnormalizable
scattering solution of the corresponding Schr¨odinger equation.
We note that the new potential
ˆ
V (r; λ) in eq. (496) and the BIC at energy
E
0
are formed using the corresponding wave function u
0
(r). Any other state,
say u
1
(r), is transformed into a solution of the new Schr¨odinger equation
by the operation given in eq. (499) which involves both u
0
and u
1
. The
central column of Table 12.2 gives a convenient overview of the relationship
of the potentials V and
ˆ
V and the solutions of the corresponding Schr¨odinger
equations.
We now give two examples to explicitly illustrate how one applies the
above procedure to obtain potentials possessing one BIC.
Example 1: Free Particle on the Half Line. Consider a free particle
on the half line (V ≡ 0 for 0 ≤ r < ∞). We choose u
0
= sin kr, the spherical
wave solution, corresponding to energy E
0
= k
2
> 0, which vanishes at r = 0.
The integral I
0
given in eq.(497) becomes
I
0
= [2kr −sin(2kr)]/(4k). (501)
We observe that I
0
→r/2 as r →∞.
The potential family
ˆ
V , deﬁned in eq.(496) becomes
ˆ
V (r; λ) =
32 k
2
sin
4
kr
D
2
0
−
8 k
2
sin(2kr)
D
0
(502)
with
D
0
(r; λ) = 2kr −sin(2kr) + 4kλ. (503)
ˆ
V has a BIC at energy E
0
= k
2
with wave function
ˆ u
0
(λ) = 4k sin kr/D
0
. (504)
For special values of the parameters k and λ, the potential
ˆ
V and its
BIC wave functions are shown in Figs. 12.3a and 12.3b. The original null
potential has now become an oscillatory potential which asymptotically has
a 1/r envelope. The new wave function at E
0
= k
2
also has an additional
damping factor of 1/r which makes it square integrable. As u
0
appears in the
numerator of
ˆ
V , eq. (496), every node of ˆ u
0
is associated with a node of
ˆ
V
123
but not every node of
ˆ
V produces a node of ˆ u
0
. The value of the eigenenergy
E
0
is clearly above the asymptotic value, zero, of the potential. Evidently,
the many oscillations of this potential, none of them able to hold a bound
state, conspire in such a way as to keep the particle trapped.
The parameter λ which appears in the denominator function D
0
(r; λ)
plays the role of a damping distance; its magnitude indicates the value of r
at which the monotonically growing integral I
0
becomes a signiﬁcant damping
factor, both for the new potential and for the new wave function. This is
illustrated graphically in Figs. 12.3a and 12.3b which are drawn for very
diﬀerent values of λ. [Note that the wave functions shown in the ﬁgures are
not normalized]. The parameter λ must be restricted to values greater than
zero in order to avoid inﬁnities in
ˆ
V and in the wave functions. In the limit
λ →∞,
ˆ
V becomes identical to V .
Example 2: Coulomb Potential. Here V = Z/r, and the unbound,
reduced l = 0 wave function satisﬁes the Schr¨odinger equation eq.(495), which
can be written in standard form
u
′′
0
+ (1 −2
˜ η
ρ
)u
0
= 0 (505)
with ρ =
√
Er and ˜ η = Z/2
√
E.
The solutions involve conﬂuent hypergeometric functions which in the
asymptotic limit approach sine waves phaseshifted by a logarithmic term.
Useful expressions for these solutions in the regions near and far from the
origin are available in the literature [228, 229]. Stillinger and Herrick [117],
following the method of Von Neumann and Wigner [116], have constructed
BIC potentials and wave functions for the case of the repulsive Coulomb
potential. Here we use our theorem to construct a oneparameter family of
isospectral potentials containing a BIC. The procedure is the same for both
positive and negative Z; the only diﬀerence being in the sign of ˜ η. The formal
expressions for the BIC potentials and wave functions have been given above,
eqs.(496) and (498), in terms of u
0
.
The positive energy solution of eq.(505) can be written in the usual form
[228, 229, 117] as the real function
u
0
(ρ) = C
0
(˜ η) e
−iρ
M(1 −i˜ η, 2, 2iρ), (506)
where
C
0
(˜ η) = (e
−π˜ η/2
) [ Γ(1 +i˜ η) [ (507)
124
and M(a, b; z) is Kummer’s function. Using tabulated expressions for the
Coulomb wave functions [229] and doing the integral for I
0
numerically, we
have obtained the BIC wave functions for representative values of λ. The
corresponding oneparameter family of potentials obtained by the SUSY pro
cedure is given in eq. (496) with V
0
= Z/r.
The results are displayed in Fig. 12.4. Fig. 12.4a shows the BIC partner
to the attractive Coulomb potential for λ = 1, k = 1, and Z = 2. Fig. 12.4b
shows the (unnormalized) wave function of the bound state in the continuum
for this potential at E
0
= k
2
. For comparison the original Coulomb potential
and wave function are also shown dotted. It is seen that the potential which
holds a bound state of positive energy shows an oscillatory behavior about
the Coulomb potential, V
C
, as is also evident from the form of eq. (496) for
ˆ
V . Since the oscillating component vanishes whenever u
0
vanishes, we have
ˆ
V = V at each node of u
0
. Compared to the original, unnormalizable wave
function, the BIC wave function in both cases shows a damped behavior due
to the denominator function. This is also seen in the ﬁgures.
A similar behavior is also expected for other spherically symmetric po
tentials with a continuous spectrum of positive eigenvalues. For arbitrary
onedimensional potentials, where the range extends from −∞ to +∞, the
situation is not so clear cut. Our method works for the Morse potential which
is steeply rising on the negative xaxis with correspondingly damped wave
functions. It also works for the case of a particle in a constant electric ﬁeld
for similar reasons. For potentials, such as V (x) = −V
0
sech
2
x, the integral
I
0
in eq. (497) is not convergent if the starting point is chosen at −∞, and it
gets negative contributions if the starting point is selected at ﬁnite xvalues.
This leads to a vanishing denominator function in the expressions for some
wave functions which makes them unacceptable.
(b) Two Parameter Family of Potentials
In the previous section (a), we have seen how to generate a oneparameter
family of potentials with one BIC. We now show how this procedure can be
extended to construct twoparameter families which contain two BICs.
In constructing the new wave functions for the oneparameter family, eq.
(496), we observe that the denominator function given in eq. (498) was all
that was needed to create the BIC, while the operation in eq. (499) ensured
that the wave functions for all the other states, there represented by ˆ u
1
, are
a solution to the new potential. Note again, there is nothing special about
the ordering of the two energy values nor the relative magnitude of E
0
and
125
E
1
, therefore we can repeat this procedure by applying the theorem to the
wave functions and the potential of the oneparameter family, but this time
we transform the state at E
1
into a BIC. The state at E
0
, which already is
a BIC, is transformed in the step of eq. (499), suitably modiﬁed, to become
a solution to the new potential. In this way we obtain the two parameter
family of potentials
ˆ
ˆ
V (r; λ, λ
1
) =
ˆ
V −2[ln(
ˆ
I
1
+ λ
1
)]
′′
=
ˆ
V −
4ˆ u
1
ˆ u
′
1
ˆ
I
1
+ λ
1
+
2ˆ u
4
1
(
ˆ
I
1
+ λ
1
)
2
(508)
with the solutions of the corresponding Schr¨odinger equation
ˆ
ˆ u
0
= (E
0
−E
1
)ˆ u
0
+
ˆ
ˆ u
1
W(ˆ u
1
, ˆ u
0
), (509)
ˆ
ˆ u
1
=
ˆ u
1
ˆ
I
1
+ λ
1
, (510)
and
ˆ
I
1
≡
r
0
ˆ u
2
1
(r
′
)dr
′
. (511)
The precise relationship of the new potential and its wave functions, which
are now both BICs, is illustrated in the last column of Table 12.2.
While the compact form of Eqs. (508  510) explicitly shows the method
of construction, it is useful to observe that the integral
ˆ
I
1
can be conveniently
recast into a simpler form which contains integrals of the form
I
i
=
r
0
u
2
i
(r
′
)dr
′
, (512)
involving the original wave functions only. Making use of eq. (499) for ˆ u
1
,
we get
ˆ
I
1
=
r
0
¸
(E
1
−E
0
)
2
u
2
1
+
u
2
0
W
2
(I
0
+ λ)
2
+ 2(E
1
−E
0
)
u
0
u
1
(I
0
+ λ)
W
¸
dr
′
. (513)
The second term is integrated by parts as
r
0
u
2
0
(I
0
+ λ)
2
W
2
(r
′
)dr
′
=
−W
2
I
0
+ λ
r
0
+
r
0
2WW
′
(I
0
+ λ)
dr
′
. (514)
126
We now use eq. (500) for the derivative of a Wronskian of two solutions of the
Schr¨odinger equation to rewrite the second term and observe, that it exactly
cancels the last term in eq. (513). We therefore have
ˆ
I
1
(r) =
−W
2
(r)
I
0
+ λ
+ (E
1
−E
0
)
2
I
1
(r). (515)
Here we have made use of the fact that our boundary conditions imply that
W(0) = 0.
As an example, we evaluate the twoparameter potential
ˆ
ˆ
V = V −2
¸
ln
(I
0
+ λ)[(E
1
−E
0
)
2
I
1
−
W
2
(r)
I
0
+ λ
+ λ
1
]
¸¸
′′
. (516)
The argument of the logarithm can be rewritten as
(E
1
−E
0
)
2
I
0
I
1
−W
2
(r) + λλ
1
+ λ(E
1
−E
0
)
2
I
1
+ λ
1
I
0
. (517)
We happen to have transformed ﬁrst the state at energy E
0
into a BIC and
then, in the second step, the state at E
1
, which introduced the parameters
λ and λ
1
. Let us now consider applying our procedure in the reverse order,
that is let us ﬁrst transform the state at energy E
1
into a BIC and then the
state at energy E
0
, producing the parameters µ and µ
1
. For this situation,
the argument corresponding to eq. (517) is
(E
1
−E
0
)
2
I
0
I
1
−W
2
(r) + µµ
1
+ µ
1
(E
1
−E
0
)
2
I
0
+ µI
1
. (518)
Clearly, one expects symmetry. This is guaranteed if the parameters are
related by
µ = λ(E
1
−E
0
)
2
, µ
1
= λ
1
/(E
1
−E
0
)
2
. (519)
This also leads to the same twoparameter wave functions. We also note that
transforming any state twice by eq. (498) does not create a second denomi
nator or anything else new, but simply changes the value of the parameter λ
as shown in ref. [70]. Finally, relation (509) ensures that all other eigenstates
will be solutions to the new potentials.
Shown in Fig. 12.5 is a potential with two BICs at energies E
0
= 1,
E
1
= 4. Clearly, the above procedure can be readily extended to obtain
multiparameter families with multiple BICs at arbitrarily selected energies.
127
Our discussion of BICs has been restricted to eﬀectively one dimensional
problems, and as stated before there is now some experimental evidence for
the existence of BICs under appropriately chosen conditions [227]. Recently,
there has also been a computation which claims the existence of BICs in QED
in three dimensions [230]. These authors have also speculated that the BIC
energies they compute in e
+
e
−
scattering in QED are in fact in reasonable
agreement with unusual peaks observed in recent heavy ion experiments [231].
13 Parasupersymmetric Quantum Mechanics
and Beyond
In the last few years, exotic quantum statistics have been widely discussed
in the literature. For example, in two space dimensions, one can have a one
parameter family of statistics interpolating between Bose and Fermi statistics
[232]. On the other hand, in three and higher space dimensions parafermi and
parabose statistics [233, 234, 235, 236] are the natural extensions of the usual
Fermi and Bose statistics. In particular, whereas Fermi and Bose statistics
describe the two onedimensional representations of the permutation group,
parafermi and parabose statistics describe the higher dimensional represen
tations of the same group. In view of the fact that the SUSY has provided
us with an elegant symmetry between fermions and bosons, it is natural to
enquire if there exists a generalization which includes the above exotic statis
tics. Such a question was raised many years ago in the context of parastring
models [237], but the speciﬁc symmetry algebra was still the usual SUSY
one.
In this section, we study the possibility of having a symmetry between
bosons and parafermions. We shall construct parasupersymmetric quantum
mechanics (PSQM) of a boson and a parafermion of order p(= 1, 2, 3...). It
turns out that whereas in the usual SUSY QM, the symmetry generators obey
structure relations that involve bilinear products, in PSQM of order p, the
structure relations involve products of (p + 1) parasupersymmetry (PARA
SUSY) charges. Various consequences of this algebra are also discussed.
It is worth adding here that historically the PSQM of order 2 was in
troduced ﬁrst [106] and its various consequences were discussed [238, 239,
240, 241, 242]. Initially it was felt that the generalization to order p was not
128
possible in the sense that the PSQM of order p cannot be characterized with
one universal algebraic relation [243]. However, later on, one was indeed able
to construct such a PSQM of order p [107].
Very recently, new forms of quantum statistics called orthofermi and or
thobose statistics, have been constructed [244]. It is then natural to construct
orthosupersymmetric quantum mechanics (OSQM) where there is symmetry
between a boson and an orthofermion of order p [109]. Unlike the PSQM
of order p, one ﬁnds that the structure relations now involve only bilinear
products of the symmetry generators. Various consequences of SUSY QM,
PSQM and OSQM are also discussed in this section. In particular, it is worth
pointing out that whereas in SUSY QM and OSQM of order p, the energy
eigenvalues are necessarily nonnegative, in PSQM of order p they need not
be so.
13.1 Parasupersymmetric Quantum Mechanics
In order to motivate the algebra of PSQM, let us recall that in SUSY QM
the symmetry between a boson and a fermion is characterized by the algebra
Q
2
= 0 = Q
†2
, [H, Q] = 0, QQ
†
+ Q
†
Q = 2H. (520)
Note that there is an extra factor of 2 on the right hand side compared to the
algebra given in Sec. 2. This results from assuming m = 1 in this section,
in conformity with the notation followed by the various authors in this ﬁeld.
The SUSY QM algebra is easily motivated by recalling that the fermionic
operators a, a
†
satisfy the algebra
a
2
= 0 = a
†2
, ¦a, a
†
¦ = 1. (521)
A useful representation of a and a
†
is in terms of the 2 2 matrices
a =
0 1
0 0
, a
†
=
0 0
1 0
. (522)
Let us now consider the parafermi operators b, b
†
of order p (=1,2,... )
which are known to satisfy the algebra [236]
(b)
p+1
= 0 = (b
†
)
p+1
, (523)
129
[[b
†
, b], b] = −2b, [[b
†
, b], b
†
] = 2b
†
. (524)
On letting
J
+
= b
†
, J
−
= b, J
3
=
1
2
[b
†
, b], (525)
it immediately follows from eqs. (523) and (524) that the operators J
±
and
J
3
satisfy the SU(2) algebra
[J
+
, J
−
] = 2J
3
, [J
3
, J
±
] = ±J
±
. (526)
Let us now choose J
3
to represent the third component of the spin
p
2
representation of the SU(2) group as given by
J
3
= diag(
p
2
,
p
2
−1, ... −
p
2
+ 1, −
p
2
). (527)
It is now easily seen [107] that the operators b and b
†
can be represented
by the following (p + 1) (p + 1) matrices [α, β = 1, 2..., (p + 1)]
(b)
αβ
= C
β
δ
α,β+1
; (b
†
)
αβ
= C
β
δ
α+1,β
(528)
where
C
β
=
β(p −β + 1) = C
p−β+1
. (529)
It is easily checked that the operators b and b
†
indeed satisfy the algebra
as given by eqs. (523) and (524). One can now ask as to what multilinear
relation is satisﬁed by b and b
†
apart from the one given by eq. (523)? It
turns out that the nontrivial relation is
b
p
b
†
+ b
p−1
b
†
b + ... + bb
†
b
p−1
+ b
†
b
p
=
1
6
p(p + 1)(p + 2)b
p−1
, (530)
where one has (p + 1) terms on the left hand side. As expected, for p = 1
this reduces to the bilinear relation for the fermionic operators given in eq.
(521).
This relation between b and b
†
strongly suggests that one may have an
analogous multilinear relation in the algebra of PSQM of order p. To that
purpose, let us choose the PARASUSY charges Q
1
and Q
†
1
as (p+1)(p+1)
matrices as given by
(Q
1
)
αβ
= (P −iW
β
)δ
α,β+1
; (Q
†
1
)
αβ
= (P + iW
α
)δ
α+1,β
(531)
130
where α, β = 1, 2, ..., (p + 1), so that Q
1
and Q
†
1
automatically satisfy
Q
p+1
1
= 0 = (Q
†
1
)
p+1
. (532)
Further, it is easily shown that the Hamiltonian (¯ h = m = 1)
(H)
αβ
= H
α
δ
αβ
, (533)
where (r = 1, 2, ..., p)
H
r
=
P
2
2
+
1
2
(W
2
r
−W
′
r
) +
1
2
C
r
H
p+1
=
P
2
2
+
1
2
(W
2
p
+ W
′
p
) +
1
2
C
p
(534)
commutes with the PARASUSY charges Q
1
and Q
†
1
provided (s = 2, 3..., p)
W
2
s−1
+ W
′
s−1
+ C
s−1
= W
2
s
−W
′
s
+ C
s
. (535)
Here C
1
, C
2
, ..., C
p
are arbitrary constants with the dimension of energy. It
turns out that the nontrivial relation between Q
1
, Q
†
1
and H is given by [107]
Q
p
1
Q
†
1
+ Q
p−1
1
Q
†
1
Q
1
+ ... + Q
1
Q
†
1
Q
p−1
1
+ Q
†
1
Q
p
1
= 2pQ
p−1
1
H, (536)
and Hermitianconjugated relations (which we shall not write explicitly) pro
vided
C
1
+ C
2
+ ... + C
p
= 0. (537)
An example is in order at this stage to illustrate the structure of PSQM
of orde r p. If one chooses
W
1
= W
2
= ... = W
p
= ωx, (538)
then it follows from eq. (535) that in this case (r = 1, 2, ..., p)
C
r+1
−C
r
= 2ω, (539)
and the Hamiltonian (533) takes a very simple form given by
H =
P
2
2
+
1
2
ω
2
x
2
−J
3
ω, (540)
131
where J
3
is as given by eq. (527). This H describes the motion of a particle
with spin p/2 in an oscillator potential and a uniform magnetic ﬁeld. The
spectrum of this Hamiltonian is
E
n,m
= (n +
1
2
−m)ω, (541)
where n = 0, 1, 2, ... and m = p/2, p/2 −1, ..., −p/2 so that the ground state
energy of the system is negative unlike in the usual SUSY QM. It is also clear
from here that whereas the ground state is nondegenerate, the ﬁrst excited
state is twofold degenerate, etc. and ﬁnally the p’th and higher excited
states are (p + 1)fold degenerate. Of course for p = 1 we recover the well
known results of SUSY QM.
Several comments are in order at this stage:
1. For arbitrary W
r
too one can show that the spectrum of the PSQM
Hamiltonian is (p + 1)fold degenerate at least starting from the p’th
and higher excited states. The nature of the ground and the ﬁrst (p−1)
excited states would however depend on the speciﬁc form of W
r
.
2. With the PSQM Hamiltonian as given by eqs. (533) and (534), one
can associate p supersymmetries characterized by the corresponding p
superpotentials W
r
(r = 1, 2, ..., p) and the corresponding p SUSY QM
Hamiltonians are given by
H
(r)
SUSY
=
H
r
−
1
2
C
r
0
0 H
r+1
−
1
2
C
r
, (542)
where H
r
is given by eq. (534).
3. Apart from Q
1
, there are (p −1) other conserved, independent PARA
SUSY charges Q
2
, Q
3
, ..., Q
p
deﬁned by (s = 2, 3, ...p)
Q
s
= (P −iW
β
)δ
α,β+1
if β = s,
= −(P −iW
β
)δ
α,β+1
if β = s, (543)
all of which commute with the Hamiltonian (534) and also satisfy the
PSQM algebra as given by eqs. (532) and (536) [107].
132
4. There is an interesting application of PSQM in nonrelativistic quantum
mechanics. As we have seen in Sec. 3, given any Hamiltonian H
1
with
p bound states with energies E
1
, E
2
, ..., E
p
and the corresponding eigen
functions ψ
1
, ψ
2
, ..., ψ
p
, one can always generate p other Hamiltonians
H
2
, H
3
, ..., H
(p+1)
with the same spectrum as H
1
except that 1, 2, ..., p
levels respectively are missing from them. Further in that case there
are p supersymmetries with the corresponding Hamiltonians being pre
cisely given by eq. (542). Besides, the p constants C
1
, C
2
, ..., C
p
are
related to the energy eigenvalues by (r = 1, 2, ..., p)
C
r
=
2
p
[E
p
+ E
p+1
+ ... + E
r+1
−(p −1)E
r
+ E
r−1
+ ... + E
1
]. (544)
Thus instead of associating the symmetry algebra sl(1/1) ⊗ SU(2) or
U(1) ⊗ SU(2) as we did in Sec. 3, one can also associate parasuper
symmetry of order p to the hierarchy of Hamiltonians H
1
, H
2
, ..., H
p
[107].
5. What is the most general solution of the relation (535) which must be
satisﬁed in order to have the PSQM of order p? Treating W
r
± W
r−1
as the two variables, eq. (535) can be reduced to a simple nonlinear
equation which can be immediately solved [106]. This is however not
very useful as it gives us W
r
+ W
r−1
for a given W
r
− W
r−1
. Ideally,
we would like to know the most general solution for W
2
for a given W
1
and then using this W
2
, one would recursively obtain W
3
, W
4
, ..., W
p
.
Unfortunately, this problem is still unsolved. Of course shape invariant
potentials satisfy eq. (535) but shape invariance is clearly not necessary.
As discussed above, given any H with p bound states, one can always
construct W
1
, W
2
, ....W
p
which will satisfy eq. (535).
One unsatisfactory feature of the PSQM of order p is that except for p = 1,
H cannot be directly expressed in terms of the PARASUSY charges Q
1
and
Q
†
1
. This is because, in eq. (536), H is multiplied by Q
p−1
1
whose inverse does
not exist. Another unsatisfactory feature is that unlike in SUSY QM, the
energy eigenvalues are not necessarily nonnegative and there is no connection
between the nonzero (zero) ground state energy and the broken (unbroken)
PARASUSY. It turns out that in case all the constants C
1
, C
2
....C
p
are chosen
to the zero, then one can take care of both of these unsatisfactory features
133
[108]. In particular, in that case H as given by eq.(534) can be expressed in
terms of any one of the p PARASUSY charges Q
s
by (s = 1, 2, ..., p)
H =
1
2
[(Q
†
s
Q
s
−Q
s
Q
†
s
)
2
+ Q
†
s
Q
s
Q
s
Q
†
s
]
1/2
(545)
It is immediately clear from here that all the energy eigenvalues are nec
essarily nonnegative and that the ground state energy being 0 (> 0) corre
sponds to unbroken (broken) PARASUSY. Further, one can also show that
in this case all the excited states are always (p +1)fold degenerate while the
nature of the ground state will depend on the speciﬁc form of W
s
[108]. In
the limit of p →∞ it has been shown that the relation (545) reduces to
H =
1
2
Q
s
Q
†
s
, (546)
which can be termed as the PSQM of inﬁnite order whose algebra corresponds
to that of Greenberg’s inﬁnite statistics [245].
There is an interesting application of this version of PSQM to the strictly
isospectral Hamiltonians discussed in Sec. 7. In particular, one can show that
any potential V
1
≡ W
2
1
−W
′
1
with at least one bound state, forms PSQM of
order p along with its SUSY partner potential V
2
= W
2
1
+ W
′
1
and with the
strictly isospectral potential families V
2
(x, λ
2
), V
2
(x, λ
3
), ..., V
2
(x, λ
p
) where
λ
2
, λ
3
, ..., λ
p
are arbitrary parameters which are either > 0 or < −1 [203].
Eq. (545) shows an explicit expression for the Hamiltonian H which
involves just one charge Q
s
but an overall square root. There is an alternative
expression for H where it is not necessary to take the square root, but which
involves all the p PARASUSY charges [108]:
2H = Q
r
Q
†
r
+ Q
†
r
Q
r
+
1
4
p
¸
s=1
(Q
†
r
Q
s
+ Q
†
s
Q
r
−2Q
†
r
Q
r
) (547)
where, r, s = 1, 2, ...p and s = r . Further, H and Q
r
also satisfy the simpler
relation
Q
r
Q
†
r
Q
r
= 2Q
r
H. (548)
In case one chooses W
α
to be of form (α = 1, 2..., p)
W
α
= −
λ + α −1
x
, (549)
134
then one has a model for conformally invariant PSQM of order p. In this case
one can show that the dilatation operator D and the conformal operator K
deﬁned by
D = −
1
4
(xP + Px), K =
1
2
x
2
(550)
satisfy relations analogous to those given in eqs. (545), (547) and (536) in
terms of the PARASUSY charges Q
α
and parasuperconformal charges S
α
deﬁned by [α = 1, 2, ..., p; i, j = 1, 2, ..., (p + 1); r = 2, 3, ..., p]
(S
1
)
ij
= −xδ
i+1,j
(551)
(S
r
)
ij
= −xδ
i+1,j
if i = r + 1,
= xδ
i+1,j
if i = r + 1. (552)
13.2 Orthosupersymmetric Quantum Mechanics
Recently, Mishra and Rajasekaran [244] have introduced new forms of quan
tum statistics called orthofermi and orthobose statistics. Orthofermi statis
tics contain a new exclusion principle which is more stringent than the Pauli
exclusion principle: an orbital state shall not contain more than one particle,
whatever be the spin direction. The wave function is thus antisymmetric in
spatial indices alone with the order of the spin indices frozen. In an analo
gous way, one can also deﬁne orthobose statistics. All these properties follow
provided the corresponding creation and annihilation operators C
†
and C
satisfy
C
kα
C
†
mβ
±δ
αβ
p
¸
γ=1
C
†
m gamma
C
kγ
= δ
km
δ
αβ
, (553)
C
kα
C
mβ
±C
mα
C
kβ
= 0, (554)
where the upper and the lower signs lead to the orthofermi and the ortho
bose cases respectively and the Latin indices k, m, ... and the Greek indices
α, β, γ, ... correspond to space and spin indices respectively.
For constructing the quantum mechanics of a boson and an orthofermion,
we ignore the spatial indices in eqs. (553) and (554) and obtain
C
α
C
†
β
+ δ
αβ
p
¸
γ=1
C
†
γ
C
γ
= δ
αβ
, (555)
135
C
α
C
β
= 0. (556)
Eq. (555) implies that
C
1
C
†
1
= C
2
C
†
2
= ... = C
p
C
†
p
. (557)
Following the discussion of the last subsection it is easy to see that a
useful representation of these operators is in terms of the (p + 1) (p + 1)
matrices deﬁned by (α, β = 1, 2, ..., p; r, s = 1, 2, ..., (p + 1))
(C
α
)
rs
= δ
r,1
δ
s,α+1
; (C
†
α
)
rs
= δ
s,1
δ
r,α+1
(558)
Let us now try to write down the algebra for the OSQM of order p where
there is a symmetry between a boson and an orthofermion of order p [109].
On comparing the algebra for the fermionic and the orthofermionic opera
tors as given by eqs. (521), (555) and (556) and remembering the SUSY
QM algebra as given by eq.(520) it is easy to convince oneself that the p
ORTHOSUSY charges Q
α
, Q
†
α
and the Hamiltonian must satisfy the algebra
(α, β = 1, 2, ..., p)
Q
α
Q
†
β
+ δ
αβ
p
¸
γ=1
Q
†
γ
Q
γ
= 2δ
αβ
H, (559)
Q
α
Q
β
= 0, [H, Q
α
] = 0. (560)
Note that for p = 1 we recover the usual SUSY QM algebra. It is easily
checked that if we choose the p ORTHOSUSY charges Q
α
as (p+1) (p+1)
matrices as given by (r = 1, 2, ..., (p + 1))
(Q
α
)
rs
= (P −iW
α
)δ
r,1
δ
s,α+1
; (Q
†
α
)
rs
= (P + iW
α
)δ
s,1
δ
r,α+1
(561)
and the Hamiltonian H as
(H)
rs
= H
r
δ
rs
, (562)
where (r = 1, 2, ..., p)
H
1
=
P
2
2
+
1
2
(W
2
1
+ W
′
1
)
H
r+1
=
P
2
2
+
1
2
(W
2
1
−W
′
1
) (563)
136
then the OSQM algebra as given by eqs. (559) and (560) is indeed satisﬁed
provided
W
2
α
+ W
′
α
= W
2
β
+ W
′
β
. (564)
Note that this condition directly follows from the OSQM relations
Q
1
Q
†
1
= Q
2
Q
†
2
... = Q
p
Q
†
p
, (565)
which follow from the OSQM relation (559). Thus, unlike PSQM, the con
stants C
1
, C
2
, ... are not allowed in OSQM. The various consequences of the
OSQM have been discussed in detail in [109]. At this point, it may be worth
while to make a relative comparison of SUSY QM with PSQM and OSQM.
1. First of all, the close similarity in structure between OSQM and PSQM
must be noted. For order p, both are based on (p+1)(p+1) matrices
and the structure of H in the two cases is very similar. The chief
diﬀerence between the two is the absence of the constants C
1
, C
2
..., C
p
in the former while in the latter they may or may not be zero.
2. Whereas the ground state energy is zero (> 0) in case the SUSY or the
ORTHOSUSY is unbroken (spontaneously broken), in general, there is
no such restriction in the PSQM case and the energy eigenvalues can
even be negative. However, in the special case when C
1
, C
2
, ..., C
p
are
all zero then the PSQM has similar prediction as the other two.
3. Whereas in the SUSY QM, all the excited states are necessarily two
fold degenerate, in OSQM of order p they are necessarily (p + 1)fold
degenerate. On the other hand, in the PSQM of order p, all the levels
starting from the p’th excited state (and above) are necessarily (p+1)
fold degenerate except when all the constants are zero in which case all
the excited states are necessarily (p + 1)fold degenerate.
14 Omitted Topics
So much work has been done in the area of SUSY QM in the last 12 years
that it is almost impossible to cover all the topics in such a review. We have
therefore decided to give a brief description of some of the omitted topics.
137
For each topic, a few references are provided, so that interested readers can
go back and trace other references and get a good idea of the developments.
(i) Supermathematics [246].
(ii)SUSY in Atomic Physics [41, 42]
In a series of papers, Kostelecky et al. [41] have discussed the relationship
between the physical spectra of diﬀerent atoms and ions using SUSY QM. In
particular, they have suggested that the helium and hydrogen spectra come
from SUSY partner potentials. This connection has been commented upon
by Rau [42].
(iii)SUSY in Condensed Matter and Statistical Physics [163, 248, 249,
250, 251, 252]
Ideas from SUSY have been used to give insight into random magnetic
ﬁelds in Isinglike models, polymers, electron localization in disordered media
and ferromagnets. These topics are covered in a set of review talks found in
the Proceedings of the Conference on Supersymmetry held at CNLS in 1983
[15].
(iv)Index Theorem and SUSY [253, 25, 28, 26, 27]
The AtiyahSinger index theorem can be related to understanding the
index of the Dirac operator on suitably deﬁned spaces. From our previous
discussion of SUSY and the Dirac equation, it is not surprising that the index
of the Dirac operator is related to the Witten index of the related SUSY QM.
Using techniques similar to calculating the fermion propagator in an external
ﬁeld, it has been possible to give a proof of the AtiyahSinger index theorem
using the supersymmetric representation of the index theorem. This work
is best described in lectures of AlvarezGaume given at the Bonn Summer
School [27]. The question of axial anamolies are best phrased using these
methods.
(v)Factorization Method and Solvable Potentials [49, 14, 51]
The method of factorization which can be traced as far back as Bernoulli(1702)
and Cauchy(1827) can be shown to be exactly equivalent to solving poten
tials by SUSY and shape invariance with translation. The reader interested
in the history as well as an excellent presentation of the factorization method
and its connection with SUSY is referred to the work of Stahlhofen [51].
(vi)Group Theory Method and Solvable Potentials [254]
In the doctoral thesis of Jainshi Wu, it is proven that there is a one to
one correspondence between using the diﬀerential realizations of the S0(2, 1)
potential group and the factorization method of Infeld and Hull [14]. Thus
138
all the solvable potentials found by Infeld and Hull can be also obtained
using group theory methods. By extending the symmetry group to SO(3, 1)
and SU(3, 1) the Yale group was also able to study several three dimensional
scattering problems.
(vii)QuasiSolvable Potentials [255, 256]
These potentials, for which a ﬁnite number of states can be determined
analytically, are intermediate between nonsolvable and analytically solvable
potentials discussed in this review. Quasisolvable potentials have been dis
cussed by Shifman and Turbiner [255, 256] and using the techniques of SUSY
QM a large number of new quasisolvable problems have been discovered
[257].
(viii)SUSY Breaking and Instantons [12, 30, 11, 258, 259]
One of the least understood problems in supersymmetric ﬁeld theories is
that of the origin of SUSY breaking. One suggestion for SUSY breaking is
that it is of dynamical origin and that instantons are responsible for it. As a
testing ground of these ideas, the role of instantons in the dynamical breaking
of SUSY has been studied extensively in various quantum mechanical models.
(ix)Propagators for SUSY Partner Potentials [98, 99]
Since the propagator for a system with a given potential is determined
from the Hamiltonian, and the various Hamiltonians in a heirarchy of SIP
are related by
Q(a
s
)H
(s)
= H
(s+1)
Q(a
s
), (566)
one can derive recursion relations for the propagators of the heirarchy. If
one member the heirarchy has a known propagator (such as the harmonic
oscillator or the free particle) then it is easy to use these recursion relations
to derive the propagators for the heirarchy. This also allows one to calculate
path integrals for SIP.
(x) SUSY and Nbody Problem [260, 150]
A novel algebraic structure is found for 1/r
2
family of many body prob
lems which provide a novel link between SUSY and quantum integrability
[260]. Also of considerable interest is a supersymmetric generalization of a N
particle quantum mechanical model with combined harmonic and repulsive
forces [150].
(xi) Coherent states for SIP [261]
Recently a Lie algebraic treatment of the SIP has been given and using
it coherent states have been constructed for these potentials.
139
(xii) SUSY and New Soliton Solutions [262]
It is well known that the stability equation for the kink solutions in scalar
ﬁeld theories in 1+1 dimensions is a Schr¨odingerlike equation. It has been
shown that new scalar ﬁeld theories with kink solutions can be obtained by
considering the isospectral deformation of this problem.
(xiii) SWKB and Tunneling [263, 82]
Expression for transmission coeﬃcient T through potential barriers has
been obtained within the SWKB approximation and it has been shown that
the analytic continuation of T for the inverted SIP (with translation) leads
to the exact bound state spectrum.
(xiv) TimeDependent Pauli Equation [264]
Kostelecky et al. [264] have succeeded in factorizing the time dependent
Pauli Equation and utilizing the SUSY to solve the Pauli Equation for a time
dependent spatially uniform magnetic induction.
(xv) SUSY and Nuclear Physics [247, 157, 265].
Relations between the spectra of eveneven and neighbouring evenodd
nuclei has been obtained by using the concepts of SUSY [247]. Also, Baye
[157] has shown that the deep and shallow nucleusnucleus potentials which
have been successfully used in the past are in fact SUSY partner potentials.
It has also been suggested that the relationships seen between the energy
levels of adjacent superdeformed nuclei can be understood in terms of SUSY
[265]
Acknowledgments It is a pleasure to thank all of our many collaborators
without whose eﬀorts this work would not have been accomplished. We would
especially like to thank Joe Ginocchio for carefully reading the manuscript.
We would like to thank Los Alamos National Laboratory for its hospitality
during the writ ing of the manuscript. We gratefully acknowledge the U.S.
Department of Energy for partial support.
140
References
[1] Y.A. Gel’fand and E.P. Likhtman, JETP Lett. 13 (1971) 323.
[2] P. Ramond, Phys. Rev. D3 (1971) 2415.
[3] A. Neveu and J. Schwarz, Nucl. Phys. B31 (1971) 86.
[4] D. Volkov and V. Akulov, Phys. Lett. B46 (1973) 109.
[5] J. Wess and B. Zumino, Nucl. Phys. B70 (1974) 39; ibid. B78 (1974)
1.
[6] For more details and other references in this ﬁeld see, for example, M.F.
Sohnius, Phys. Rep. 128 (1985) 39.
[7] S. Coleman and J. Mandula, Phys. Rev. 159 (1967) 1251.
[8] S.Ferrara, D. Freedman and P. van Nieuwenhuizen, Phys. Rev. D13
(1976) 3214.
[9] S. Deser and B. Zumino, Phys. Lett. B62 (1976) 335.
[10] E. Witten, Nucl. Phys. B188 (1981) 513.
[11] F. Cooper and B. Freedman, Ann. Phys. 146 (1983) 262.
[12] E. Witten, Nucl. Phys. B202 (1982) 253.
[13] C. Bender, F. Cooper and A. Das, Phys. Rev. D28 (1983) 1473.
[14] L. Infeld and T.E. Hull, Rev. Mod. Phys. 23 (1951) 21.
[15] See for example, Supersymmetry in Physics, editors: V.A. Kostelecky
and D.K. Campbell, North Holland (1985).
[16] R. Akhoury and A. Comtet, Nucl. Phys. B245 (1984) 213.
[17] A. Niemi and L.C.R. Wijerwardhana, Phys. Lett. B138 (1984) 389.
[18] A. Khare and J. Maharana, Phys. Lett. B145 (1984) 77.
[19] S. Cecotti and L. Girardello, Ann. Phys. 145 (1983) 81.
141
[20] D. Boyanovsky and R. Blankenbecler, Phys. Rev. D30 (1984) 1821.
[21] B. Freedman and F. Cooper, Physica D15 (1985) 138.
[22] L. Girardello, C. Imbimbo and S. Mukhi, Phys. Lett. B132 (1983) 69.
[23] A. Kihlberg, P. Salomonson and B.S. Skagerstam, Zeit. Phys. C28
(1985) 203.
[24] R.K. Bhaduri, A. Khare and M.V.N. Murthy, Mod. Phys. Lett. A4
(1989) 2701.
[25] L. AlvarezGaum´e, Jour. Phys. A16 (1983) 4177.
[26] L. AlvarezGaum´e and E. Witten, Nucl. Phys. B234 (1984) 269.
[27] L. AlvarezGaum´e, Bonn Summer School Lectures (1984), unpublished.
[28] D. Friedan and P. Windey, Physica D15 (1985) 71.
[29] G. Parisi and N. Sourlas, Nucl. Phys. B206 (1982) 321.
[30] P. Salomonson and J. van Holten, Nucl. Phys. B196 (1982) 509.
[31] M. Bernstein and L. Brown, Phys. Rev. Lett. 52 (1984) 1933.
[32] P. Kumar, M. RuizAltaba and B.S. Thomas, Phys. Rev. Lett. 57 (1986)
2749.
[33] F. Marchesoni, P. Sodano and M. Zannetti, Phys. Rev. Lett. 61 (1988)
1143.
[34] W.Y. Keung, E. Kovacs and U. Sukhatme, Phys. Rev. Lett. 60 (1988)
41.
[35] M. de Crumbrugghe and V. Rittenberg, Ann. Phys. 151 (1983) 99.
[36] L.F. Urrutia and E. Hernandez, Phys. Rev. Lett. 51 (1983) 755.
[37] H. Ui, Prog. Theor. Phys. 72 (1984) 192, 813; H. Ui and A. Takeda,
ibid. 72 (1984) 266.
142
[38] A. Andrianov, N. Borisov and M. Ioﬀe, Phys. Lett. A105 (1984) 19,
Phys. Lett. B181 (1986) 141; A. Andrianov, N. Borisov, M. Eides and
M. Ioﬀe, Phys. Lett. A109 (1985) 143.
[39] A. Khare and J. Maharana, Nucl. Phys. B244 (1984) 409.
[40] A.B. Balantekin, Ann. Phys. 164 (1985) 277.
[41] A. Kostelecky and M.M. Nieto, Phys. Rev. Lett. 53 (1984) 2285, Phys.
Rev. A32 (1985) 1293,3243, Phys. Rev. Lett. 56 (1986) 96.
[42] A. R. P. Rau, Phys. Rev. Lett. 56 (1986) 95.
[43] C. Blockley and G. Stedman, Euro. Jour. Phys. 6 (1985) 218.
[44] L. Gendenshtein, JETP Lett. 38 (1983) 356.
[45] R. Dutt, A. Khare and U. Sukhatme, Phys. Lett. B181 (1986) 295.
[46] J. Dabrowska, A. Khare and U. Sukhatme, Jour. Phys. A21 (1988)
L195.
[47] F. Cooper, J.N. Ginocchio and A. Wipf, Phys. Lett, A129 (1988) 145.
[48] A. Khare and U. Sukhatme, Jour. Phys. A21 (1988) L501.
[49] E. Schr¨odinger, Proc. Roy. Irish Acad. A46 (1940) 9.
[50] W.T. Reid, Riccati Diﬀerential Equations, Math. in Sci. and Eng., Aca
demic Press, N.Y. (1972).
[51] A. Stahlhofen, Duke Univ. preprint (1989).
[52] F. Cooper, J.N. Ginocchio and A. Khare, Phys. Rev. D36 (1987) 2458.
[53] C. Chuan, Jour. Phys. A 24 (1991) L1165.
[54] R. Dutt, A. Khare and U. Sukhatme, Am. Jour. Phys. 56 (1988) 163.
[55] G. Levai, Jour. Phys. A22 (1989) 689.
[56] G.A. Natanzon, Vestnik Leningrad Univ. 10 (1971) 22, Teoret. Mat.
Fiz. 38 (1979) 146.
143
[57] J.N. Ginocchio, Ann. Phys. 152 (1984) 203; ibid. 159 (1985) 467.
[58] A. Khare and U. Sukhatme, Jour. Phys. A26 (1993) L901.
[59] D. Barclay, R. Dutt, A. Gangopadhyaya, A. Khare, A. Pagnamenta and
U. Sukhatme, Phys. Rev. A48 (1993) 2786.
[60] A. Shabat, Inv. Prob. 8 (1992) 303.
[61] V. Spiridonov, Phys. Rev. Lett. 69 (1992) 298.
[62] V. Spiridonov, Univ. de Montreal preprint UdeMLPNTH9492.
[63] A. Degasperis and A. Shabat, Univ. de Roma preprint No. 919 (1992)
[64] G. Darboux, C.R. Academy Sc. (Paris) 94 (1882) 1456.
[65] P. Abraham and H. Moses, Phys. Rev. A22 (1980) 1333.
[66] D. Pursey, Phys. Rev. D33 (1986) 1098, 1103, 2267; M. Luban and D.
Pursey, ibid. D33 (1986) 431.
[67] K. Chadan and P.C. Sabatier, Inverse Problems in Quantum Scattering
Theory, Springer Verlag (1977).
[68] M.M. Nieto, Phys. Lett. B145 (1984) 208.
[69] C.V. Sukumar, Jour. Phys. A18 (1985) 2917, 2937.
[70] A. Khare and U. Sukhatme, Jour. Phys. A22 (1989) 2847.
[71] A. Khare and U. Sukhatme, Phys. Rev. A40 (1989) 6185.
[72] W.Y. Keung, U. Sukhatme, Q. Wang and T. Imbo, Jour. Phys. A22
(1989) 987.
[73] W. Kwong and J. Rosner, Prog. Theor. Phys. Supp. 86 (1986) 366.
[74] C.V. Sukumar, Jour. Phys. A19 (1986) 2297; ibid. A20 (1987) 2461.
[75] Q. Wang, U. Sukhatme, W.Y. Keung and T. Imbo, Mod. Phys. Lett.
A5 (1990) 525.
144
[76] W. Kwong, H. Riggs, J. Rosner and H.B. Thacker, Phys. Rev. D39
(1989) 1242.
[77] T. Imbo and U. Sukhatme, Phys. Rev. Lett. 54 (1985) 2184.
[78] F. Cooper and P. Roy, Phys. Lett. A143 (1990) 202.
[79] A. Comtet, A. Bandrauk and D.K. Campbell, Phys. Lett. B150 (1985)
159.
[80] A. Khare, Phys. Lett. B161 (1985) 131.
[81] For a good review on this ﬁeld and references therein see, R. Dutt, A.
Khare and U. Sukhatme, Am. Jour. Phys. 59 (1991) 723.
[82] R. Dutt, R. De, R. Adhikari and A. Comtet, Phys. Lett. A152 (1991)
381.
[83] R. Adhikari, R. Dutt, A. Khare and U. Sukhatme, Phys. Rev. A38
(1986) 1679.
[84] K. Raghunathan, M. Seetharaman and S. Vasan, Phys. Lett. B188
(1987) 351.
[85] D. Barclay and C.J. Maxwell, Phys. Lett. A157 (1991) 351.
[86] R. Dutt, A. Khare and Y.P. Varshni, Phys. Lett. A123 (1987) 375.
[87] P. Roy, R. Roychoudhury and Y.P. Varshni, Jour. Phys. A21 (1988)
1589.
[88] D. DeLaney and M.M. Nieto, Phys. Lett. B247 (1990) 301.
[89] Y.P. Varshni, Jour. Phys. A25 (1992) 5761.
[90] A. Khare and Y.P. Varshni, Phys. Lett. A142 (1989) 1.
[91] S. Fricke, A.B. Balantekin, P. Hatchell and T. Uzer, Phys. Rev. A37
(1988) 1686.
[92] Y. Murayama, Phys. Lett. A136 (1989) 455.
145
[93] U. Sukhatme and A. Pagnamenta, Phys. Lett. A151 (1990) 7.
[94] A. Inomata and G. Junker, Lectures on Path Integration, editors: H.
Cerdeira et al, World Scientiﬁc, Singapore (1992); Symposium on Adv.
Topics in Quantum Mechanics, editors: J.Q. Liang et al, Science Press,
Beijing (1992).
[95] R. Dutt, A. Gangopadhyaya, A. Khare, A. Pagnamenta and U.
Sukhatme, Phys. Lett. A174 (1993) 363.
[96] R. Dutt, A. Gangopadhyaya, A. Khare, A. Pagnamenta and U.
Sukhatme, Phys. Rev. A48 (1993) 1845.
[97] A. Inomata, G. Junker and A. Suparmi, Jour. Phys. A26 (1993) 2261.
[98] A. Das and W. Huang, Phys. Rev. D41 (1990) 3241.
[99] A. Jayannavar and A. Khare, Phys. Rev. D47 (1993) 4796.
[100] B. Thaller, The Dirac Equation, Springer Verlag (1992).
[101] F. Cooper, A. Khare, R. Musto, and A. Wipf, Ann. Phys. 187 (1988)
1.
[102] Y. Nogami and F. Toyama, Phys. Rev. A47 (1993) 1708.
[103] C.V. Sukumar, Jour. Phys. A18 (1985) L697.
[104] E. D’Hoker and L. Vinet, Phys. Lett. B137 (1984) 72.
[105] H. Yamagishi, Phys. Rev. D29 (1984) 2975.
[106] V. Rubakov and V. Spiridonov, Mod. Phys. Lett. A3 (1993) 1337.
[107] A. Khare, Jour. Phys. A25 (1992) L749, Jour. Math. Phys. 34 (1993)
1277.
[108] A. Khare, A.K. Mishra and G. Rajasekaran, Mod. Phys. Lett. A8
(1993) 107.
[109] A. Khare, A.K. Mishra and G. Rajasekaran, Int. Jour. Mod. Phys. A8
(1993) 1245.
146
[110] A. Jevicki and J. Rodrigues, Phys. Lett. B146 (1989) 55.
[111] M. Shifman, A. Smilga and A. Vainshtein, Nucl. Phys. B249 (1988)
79.
[112] J. Casahorran and S. Nam, Int. Jour. Mod. Phys. A6 (1991) 2729; J.
Casahorran, Phys. Lett. B156 (1991) 1925: P. Roy, R.Roychoudhury
and Y.P. Varshni, Jour. Phys. A21 (1988) 3673.
[113] P. Roy and R. Roychoudhury, Phys. Rev. D32 (1985) 1597.
[114] P. Panigrahi and U. Sukhatme, Phys. Lett. A178 (1993) 251.
[115] J. Pappademos, U. Sukhatme and A. Pagnamenta, Phys. Rev. A48
(1993) 3525.
[116] J. von Neumann and E. Wigner, Zeit. Phys. 30 (1929) 465.
[117] F. Stillinger and D. Herrick, Phys. Rev. A11 (1975) 446.
[118] L.E. Ballentine, Quantum Mechanics, Prentice Hall, PP. 205 (1990).
[119] L.E. Gendenshtein and I.V. Krive, Sov. Phys. Usp. 28 (1985) 695.
[120] R. Haymaker and A.R.P. Rau, Am. Jour. Phys. 54 (1986) 928.
[121] D. Lancaster, Nuovo Cimento A79 (1984) 28.
[122] G. Stedman, Euro. Jour. Phys. 6 (1985) 225.
[123] A. Lahiri, P. Roy and B. Bagchi, Int. Jour. Mod. Phys. A5 (1990) 1383.
[124] O. de Lange and R. Raab, Operator Methods in Quantum Mechanics,
Oxford Univ. Press (1991).
[125] A. Khare and U. Sukhatme, Physics News (India) 22 (1991) 35.
[126] L. Landau and E. Lifshitz, Nonrelativistic Quantum Mechanics 3rd
Volume, Pergamon Press, N.Y. (1977).
[127] G. Junker, Jour. Phys. A23 (1990) L881.
[128] R. De, R. Dutt and U. Sukhatme, Jour. Phys. A25 (1992) L843.
147
[129] A. Gangopadhyaya, P. Panigrahi and U. Sukhatme, Univ. of Montreal
preprint UdeMLPNTH93172 (1993).
[130] V.A. Kostelecky, M.M. Nieto, D.Truax, Phys. Rev. D32 (1985) 2627.
[131] A. Khare and R.K. Bhaduri, Inst. of Physics, Bhubaneswar preprint
IPBBSR/9365; hepth/9310104.
[132] A. Khare and R.K. Bhaduri, Jour. Phys. A27 (1994) in press.
[133] M.M. Nieto, Am. Jour. Phys. 47 (1979) 1067; M.M. Nieto and L.M.
Simmons,ibid. 47 (1979) 634.
[134] R. Montemayor and L.D. Salem, Phys. Rev. A40 (1989) 2170
[135] V. Spiridonov, Mod. Phys. Lett. A7 (1992) 1241.
[136] L.C. Biedenharn, Jour. Phys. A22 (1989) L873.
[137] A.M. Macfarlane, Jour. Phys. A22 (1989) 4581.
[138] G.L. Lamb, Elements of Soliton Theory, Wiley, N.Y. (1980).
[139] A. Das, Integrable Models, World Scientiﬁc, Singapore (1989).
[140] See for example, P.M. Morse and H. Feshbach, Methods of Theoretical
Physics, Vol.I, Chapter 5, McGraw Hill, N.Y. (1953).
[141] F. Calogero, Jour. Math. Phys. 10 (1969) 2191.
[142] J. Wolfes, Jour. Math. Phys. 15 (1974) 1420.
[143] B. Sutherland, Jour. Math. Phys. 12 (1971) 246.
[144] F. Calogero, Jour. Math. Phys. 12 (1971) 419.
[145] For a list of solvable potentials and other references in the ﬁeld see
M. Olshanetsky and A. Perelomov, Phys. Rep. 71 (1981) 314; ibid. 94
(1983) 6.
[146] F. Haldane, Phys. Rev. Lett. 60 (1988) 635.
[147] B. Shastry, Phys. Rev. Lett. 60 (1988) 639.
148
[148] H. Frahm, Jour. Phys. A26 (1993) L473.
[149] A. Polychronakos, Phys. Rev. Lett. 69 (1992) 703.
[150] D. Z. Freedman and P. F. Mende, Nucl. Phys. B344 (1990) 317.
[151] N. Froman and P. O. Froman, JWKB Approximation, Contributions
to the Theory, North Holland, Amsterdam (1965).
[152] J. B. Krieger and C. Rosenzweig, Phys. Rev. 164 (1967) 171; Jour.
Chem. Phys. 47 (1967) 2942.
[153] M. Seetharaman and S. Vasan, Jour. Phys. A17 (1984) 2485, 2493;
ibid. A18 (1985) 1041.
[154] D. Barclay, A. Khare and U. Sukhatme, Phys. Lett. A183 (1993) 263.
[155] S. Miller and R. Good, Jr., Phys. Rev. 91 (1953) 174.
[156] S. Chaturvedi and R. Raghunathan, Jour. Phys. A19 (1986) 1775.
[157] D. Baye, Phys. Rev. Lett. 58 (1987) 2738.
[158] R. Amado, Phys. Rev. A37 (1988) 2277.
[159] C. Gardner, J. Greene, M. Kruskal and R. Miura, Comm. Pure App.
Math. 27 (1974) 97.
[160] W. Eckhaus and A. Van Harten, The Inverse Scattering Transforma
tion and the Theory of Solitons, NorthHolland, Amsterdam (1981).
[161] S. Chern and C. Peng, Manuscr. Math. 28 (1979) 207.
[162] F. Calogero and A. Degasperis, Spectral Transforms and Solitons,
NorthHolland, Amsterdam, (1982).
[163] N. Sourlas, Physica 15D (1985) 115.
[164] E. Gildner and A. Patrascioiu, Phys. Rev. D16 (1977) 1802.
[165] S. Ma and G. Mazenko, Phys. Rev. B11 (1975) 4077.
[166] S. Raby, Unpublished Notes and Lectures, Los Alamos (1981).
149
[167] A. Salam and J. Strathdee, Phys. Rev. D11 (1975) 1521.
[168] E. Gozzi, M. Reuter and W. Thacker, Phys. Lett. A 183 (1993) 29/
[169] C. Bender et al., Phys. Rev. D37 (1988) 1472.
[170] A. Caldeira and A. Leggett, Phys. Rev. Lett. 46 (1981) 211; Ann. Phys.
(N.Y.) 149 (1983) 374.
[171] S. Chakravarty, Phys. Rev. Lett. 49 (1982) 681.
[172] A. Bray and M. Moore, Phys. Rev. Lett. 49 (1982) 1545.
[173] S. Chakravarty and A. Leggett, Phys. Rev. Lett. 52 (1984) 5.
[174] G. Schon and A. Zaikin, Phys. Rep. 198 (1990) 237.
[175] S. Glasstone, K. Laidler and H. Eyring, Theory of Rate Processes,
McGrawHill, N.Y. (1941).
[176] W. Miller, Jour. Chem. Phys. 61 (1974) 1823; ibid. 62 (1975) 1899.
[177] N. van Kampen, Jour. Stat. Phys. 17 (1977) 71.
[178] I. Aﬄeck, Phys. Rev. Lett. 46 (1981) 388.
[179] D. Boyanovsky, R. Willey and R. Holman, Nucl. Phys. B376 (1992)
599.
[180] K. Sch¨onhammer, Zeit. Phys. B78 (1990) 63.
[181] A. Gangopadhyaya, P. Panigrahi and U. Sukhatme, Phys. Rev. A47
(1993) 2720.
[182] S. Coleman, Aspects of Symmetry, Cambridge Univ. Press, N.Y. (1985).
[183] E. Merzbacher, Quantum Mechanics, Wiley, N.Y. (1970).
[184] F. Hildebrand, Advanced Calculus for Applications, PrenticeHall, En
glewood Cliﬀs (1976).
[185] T. Imbo and U. Sukhatme, Am. Jour. Phys. 52 (1984) 140.
150
[186] L. Mlodinow and N. Papanicolaou, Ann. Phys. 128 (1980) 314; ibid.
131 (1981) 1.
[187] E. Witten in Recent Developments in Gauge Theories, ed. G.’ tHooft
et al., Plenum, N.Y. (1980).
[188] C. Bender, L. Mlodinow and N. Papanicolaou, Phys. Rev. A25 (1982)
1305.
[189] L. Mlodinow and M. Shatz, Jour. Math. Phys. 25 (1984) 943.
[190] J. Ader, Phys. Lett. A97 (1983) 178.
[191] R. Gangopadhyay, G. Ghosh and B. DuttaRoy, Phys. Rev. A30 (1984)
594.
[192] L. Yaﬀe, Rev. Mod. Phys. 54 (1982) 407.
[193] M. SinhaRoy, R. Gangopadhyay and B. DuttaRoy, Jour. Phys. A17
(1984) L687.
[194] J. Miramontes and C. Pajares, Nuovo Cimento 84 (1984) 10.
[195] S. Hikami and E. Brezin, Jour. Phys. A12 (1979) 759.
[196] U. Sukhatme and T. Imbo, Phys. Rev. D28 (1983) 418.
[197] T. Imbo, A. Pagnamenta and U. Sukhatme, Phys. Rev. D29 (1984)
1669.
[198] T. Imbo, A. Pagnamenta and U. Sukhatme, Phys. Lett. A105 (1984)
183.
[199] T. Imbo and U. Sukhatme, Phys. Rev. D31 (1985) 2655.
[200] R. Dutt, U. Mukherji and Y.P. Varshni, Jour. Phys. B18 (1985) 3311.
[201] C. S. Lam and Y.P. Varshni, Phys. Rev. A4 (1971) 1875.
[202] G. Stanciu, Jour. Math. Phys. 8 (1967) 2043.
[203] A. Khare and C.N. Kumar, Mod. Phys. Lett. A8 (1993) 523.
151
[204] Y. Aharonov and A. Casher, Phys. Rev. A19 (1978) 2461.
[205] R. Jackiw and C. Rebbi, Phys. Rev. D13 (1976) 3358.
[206] R. Jackiw and J.R. Schriﬀer, Nucl. Phys. B190 (1981) 253.
[207] D.K. Campbell and A. Bishop, Nucl. Phys. B200 (1982) 293.
[208] E. Witten and D. Olive, Phys. Lett. B78 (1978) 97.
[209] A. d’Adda and P. di Vecchia, Phys. Lett. B73 (1978) 162.
[210] F.M. Toyama, Y. Nogami and Z. Zhao, Phys. Rev. A47 (1993) 897.
[211] J. Bjorken and S. Drell, Relativistic Quantum Mechanics, McGraw Hill
(1964) Chapter 4.
[212] R. Feynman and M. GellMann, Phys. Rev. 109 (1958) 193.
[213] L.M. Brown, Phys. Rev. 111 (1958) 193, in Lectures in Theoretical
Physics, Boulder, Interscience Publishers, Inc., N.Y. 4 (1962) 324.
[214] R. Jackiw, Phys. Rev. D29 (1984) 2375.
[215] R. Musto, L. O’Raifeartaigh and A. Wipf, Phys. Lett. B175 (1986)
433.
[216] R. Hughes, V.A. Kostelecky and M.M. Nieto, Phys. Rev. D34 (1986)
1100, Phys. Lett. B171 (1986) 226.
[217] K. Fujikawa, Phys. Rev. D21 (1987) 2848.
[218] S.G. Rajeev, Ann. Phys. (N.Y.) 173 (1987) 249.
[219] J. Schwinger, Phys. Rev. 82 (1951) 664.
[220] T. Imbo and U. Sukhatme, Phys. Rev. D33 (1986) 3147.
[221] W. Frank, D. Land and R. Spector, Rev. Mod. Phys. 43 (1971) 36.
[222] H. Narnhofer, Acta Physica Austriaca 40 (1974) 306.
[223] L. Lathouwers, Jour. Math. Phys. 16 (1975) 1393.
152
[224] N. MeyerVernet, Am. Jour. Phys. 50 (1982) 354.
[225] H. Friedrich and D. Wintgen, Phys. Rev. A32 (1985) 3231.
[226] M. Robnik, Jour. Phys. A12 (1979) 1175, ibid. 19 (1986) 3845.
[227] F. Capasso et al., Nature 358 (1992) 5656.
[228] Handbook of Mathematical Functions, edited by M. Abramowitz, Irene
A. Stegun, National Bureau of Standards Applied Math. Series, Nr. 55
(1964).
[229] Tables of Coulomb Wave Functions, National Bureau of Standards Ap
plied Mathematics Series, Vol. 17 (1952).
[230] J. Spence and J. Vary, Phys. Lett. B254 (1991) 1.
[231] T. Cowan et al., Phys. Rev. Lett. 56 (1986) 444; P. Salabura et al.,
Phys. Lett. B245 (1990) 153; W. Koenig et al., Phys. Lett. B218 (1989)
12.
[232] J. Leinaas and J. Myrheim, Nuovo Cimento B37 (1977) 1.
[233] H. Green, Phys. Rev. 90 (1953) 270.
[234] D. Volkov, Zh. Eksp. Teor. Fiz 38 (1960) 519.
[235] O. Greenberg and A. Messiah, Phys. Rev. B138 (1965) 1155.
[236] Y. Ohnuki and S. Kamaefuchi, Quantum Field Theory and Parastatis
tics, Tokyo Univ. Press (1982).
[237] F. Ardalan and F. Mansouri, Phys. Rev. D9 (1974) 334.
[238] J. Beckers and N. Debergh, Mod. Phys. Lett. A4 (1989) 1209, 2289;
Jour. Math. Phys. A31 (1990) 1523 ; Jour. Phys. A23 (1990) L751,
L1073 ; Nucl. Phys. B340 (1990) 777.
[239] S. Durand, R. Floreanini, M. Mayrand and L. Vinet, Phys. Lett. B233
(1989) 158 ; S. Durand and L. Vinet, Jour. Phys. A23 (1990) 3661.
[240] V. Spiridonov, Jour. Phys. A24 (1991) L529.
153
[241] A. Andrianov and M. Ioﬀe, Phys. Lett. B255 (1991) 543 ; A. Andri
anov, M. Ioﬀe, V. Spiridonov and L. Vinet, ibid. B272 (1991) 297.
[242] V. Merkel, Mod. Phys. Lett. A5 (1990) 2555; ibid. A6 (1991) 199.
[243] S. Durand, M. Mayrand, V. Spiridonov and L. Vinet, Mod. Phys. Lett.
A6 (1991) 3163.
[244] A.K. Mishra and G. Rajasekaran, Pramana (Jour. Phys.) 36 (1991)
537; ibid. 37 (1991) 455(E).
[245] O. Greenberg, Phys. Rev. Lett. 64 (1990) 705.
[246] F. Berezin, The Method of Second Quantization, Nauka, Moscow
(1965).
[247] F. Iachello, Phys. Rev. Lett. 44 (1980) 772; Physica D15 (1985) 85
and references therein; R. Casten, Physica D15 (1985) 99.
[248] J. Cardy, Physica D15 (1985) 123 and references therein.
[249] Y. Shapir, Physica D15 (1985) 129 and references therein.
[250] Y. Nambu, Physica D15 (1985) 147 and references therein.
[251] K. Efetov, Sov. Phy. JETP 55 (1985) 514; Adv. Phys. 32 (1983) 53.
[252] E. Brezin, D. Gross and C. Itzykson, Nucl. Phys. B235 (1984) 24.
[253] M. Atiyah and I.M. Singer, Bull. Am. Math. Soc. 69 (1963) 422.
[254] Y. Alhassid, F. Gursey and F. Iachello, Phys. Rev. Lett. 50 (1983) 873;
Ann. of Phys. 148 (1983) 336; J. Wu, Ph.D. thesis, Yale Univ. (1985),
unpublished.
[255] M. Shifman, Int. Jour. Mod. Phys. A4 (1989) 2897.
[256] A. Turbiner and A. Usheridze, Phys. Lett. A126 (1987) 181, A. Tur
biner, Sov. Phys. JETP 67 (1988) 230 ; Comm. Math. Phys. 118 (1988)
467.
[257] D.P. Jatkar, C.N. Kumar and A. Khare, Phys. Lett. A142 (1989) 200.
154
[258] R.B. Abbott and W.J. Zakrzewski, Z. Phys. C20 (1983) 227.
[259] A. Khare and J. Maharana, Zeit. Phys. C3 (1984) 191.
[260] B.S. Shastry and B. Sutherland, Phys. Rev. Lett. 70 (1993) 4029, B.S.
Shastry, ibid. 69 (1992) 164.
[261] T. Fukui and N. Aizawa, Phys. Lett. A180 (1993) 308; T. Fukui,
Yukawa Institute, Kyoto preprint (1993).
[262] C.N. Kumar, Jour. Phys. A20 (1989) 5347.
[263] T. Sil, A. Mukherjee, R. Dutt and Y.P. Varshni, Phys. Lett. A184
(1993) in press.
[264] V.A. Kostelecky, V.I. Man’ko, M.M. Nieto, and D. Truax, Phys. Rev.
A48 (1993) 951.
[265] R. D. Amado, R. Bijker, F. Cannata and J. Dedonder, Phys. Rev. Lett.
67 (1991) 2777.
155
FIGURE CAPTIONS
Figure 2.1: The energy levels of two supersymmetric partner potentials V
1
and V
2
. The ﬁgure corresponds to unbroken SUSY. The energy levels are
degenerate except that V
1
has an extra state at zero energy E
(1)
0
= 0. The
action of the operators A and A
†
in connecting eigenfunctions is shown.
Figure 2.2: The inﬁnite square well potential V = 0 of width L = π and
its supersymmetric partner potential 2cosec
2
x in units ¯ h = 2m = 1. The
ground state of the inﬁnite square well has energy 1. Note the degenerate
higher energy levels at energies 2
2
, 3
2
, 4
2
, ...
Figure 4.1: Selfsimilar superpotentials W(x) for various values of the de
formation parameter q. The curve labeled H.O. (harmonic oscillator) cor
responds to the limiting case of q = 1. Note that only the range x ≥ 0 is
plotted since the superpotentials are antisymmetric W(x) = −W(−x).
Figure 4.2: Selfsimilar potentials V
1
(x) (symmetric about x = 0) corre
sponding to the superpotentials shown in Fig. 4.1 . The curves are taken
from ref. [59].
Figure 4.3: A double well potential V
1
(x) (solid line) and its single well
supersymmetric partner V
2
(x) (dotted line). Note that these two potentials
are shape invariant [59] with a scaling change of parameters. The energy
levels of V
1
(x) are clearly marked.
Figure 5.1: Diagram showing how all the shape invariant potentials of Table
4.1 are interrelated by point canonical coordinate transformations. Poten
tials on the outer hexagon have eigenfunctions which are hypergeometric
functions whereas those on the inner triangle have eigenfunctions which are
conﬂuent hypergeometric functions. These two types are related by suitable
limiting procedures. The diagram is from ref. [129].
Figure 7.1: Selected members of the family of potentials with energy spectra
identical to the one dimensional harmonic oscillator with ω = 2. The choice
of units is ¯ h = 2m = 1. The curves are labeled by the value λ
1
, and cover
156
the range 0 < λ
1
≤ ∞. The curve λ
1
= ∞ is the one dimensional harmonic
oscillator. The curve marked λ
1
= 0 is known as the Pursey potential [66]
and has one bound state less than the oscillator.
Figure 7.2: Ground state wave functions for all the potentials shown in Fig.
7.1, except the Pursey potential.
Figure 7.3: A schematic diagram showing how SUSY transformations are
used for deleting the two lowest states of a potential V
1
(x) and then re
inserting them, thus producing a twoparameter (λ
1
, λ
2
) family of potentials
isospectral to V
1
(x).
Figure 7.4: The pure threesoliton solution of the KdV equation as a function
of position (x) and time (t). This solution results from constructing the
isospectral potential family starting from a reﬂectionless, symmetric potential
with bound states at energies E
1
= −25/16, E
2
= −1, E
3
= −16/25. Further
details are given in the text and ref. [75].
Figure 9.1: A “deep” symmetric double well potential V
1
(x) with minima at
x = ±x
0
and its supersymmetric partner potential V
2
(x).
Figure 9.2: Supersymmetric partner potentials V
1
(x) and V
2
(x) correspond
ing to two choices of the parameter x
0
for the potentials given in eq. (392).
For a detailed discussion, see the text and ref. [34].
Figure 9.3: The energy splitting t = E
1
−E
0
as a function of the separation
2x
0
of the superposed Gaussians in the ground state wave function ψ
0
(x).
This ﬁgure, taken from ref. [34], shows the remarkable accuracy of the SUSY
based energy splitting computations for a double well potential.
Figure 12.1: The harmonic oscillator potential V
1(1)
and its partner potential
V
2(1)
as given by eq. (488). These potentials arise from a singular super
potential W
1
= x − 1/x corresponding to the choice φ = ψ
1
(x), ǫ = E
1
in
Table 12.1 . The energy levels for both potentials are shown. Notice the
negative energy state of V
1(1)
at energy −2, and the partial degeneracy of the
eigenvalue spectra coming from the even parity of the potentials [114].
157
Figure 12.2: The Morse and its supersymmetric partner potential [eq. (494)]
coming from the singular superpotential W
1
given in eq. (493). The singu
larity at x = −ln 3.5 breaks V
2(1)
into two disjoint pieces, and the eigenvalue
spectra have no degeneracy [114].
Figure 12.3: Potentials
ˆ
V (r) (solid lines) possessing one bound state in the
continuum (BIC) obtained by starting from a free particle and constructing
a oneparameter (λ) family. The BIC wave function ˆ u
0
(r) is at energy E = 1,
and is shown by the dashed line. Fig. (a) corresponds to λ = 0.5 and ﬁg.
(b) corresponds to a much larger value λ = 5 .
Figure 12.4: (a) A potential (solid line) with one BIC at energy E = 0.25
obtained by starting from an attractive Coulomb potential (dotte d line). (b)
A plot of the wave functions corresponding to the potentials in part (a).
Figure 12.5:(a) An example of a potential with two BICs at energies E
0
= 1
and E
1
= 4. (b) The wave functions at energies E
0
= 1 (dashed line) and
E
1
= 4 (dotted line) [115].
TABLE CAPTIONS
Table 4.1: All known shape invariant potentials in which the parameters a
2
and a
1
are related by a translation (a
2
= a
1
+α). The energy eigenvalues and
eigenfunctions are given in units ¯ h = 2m = 1. The constants A, B, α, ω, l are
all taken ≥ 0. Unless otherwise stated, the range of potentials is −∞≤ x ≤
∞, 0 ≤ r ≤ ∞. For spherically symmetric potentials, the full wave function
is ψ
nlm
(r, θ, φ) = ψ
nl
(r)Y
lm
(θ, φ). Note that the wave functions for the ﬁrst
four potentials (Hermite and Laguerre polynomials) are special cases of the
conﬂuent hypergeometric function while the rest (Jacobi polynomials) are
special cases of the hypergeometric function. Fig. 5.1, taken from ref. [129],
shows the interrelations between all the SIPs in the table via point canonical
coordinate transformations.
158
Table 6.1: Comparison of the lowest order WKB and SWKB predictions
for the bound state spectrum of the Ginocchio potential for diﬀerent values
of the parameters λ, ν and several values of the quantum number n. The
exact answer is also given. Units corresponding to ¯ h = 2m = 1 are used
throughout.
Table 9.1: Comparison of the three lowest energy eigenvalues obtained by a
variational method with the exact results.
Table 12.1: Choice of the constant ǫ and the corresponding solution φ(x) in
eq. (480) determines the choice of the superpotential W
φ
which produces any
given nonsingular potential
˜
V (x) [eq. (481)]. The table shows how certain
choices of ǫ give rise to singular superpotentials, negative energy eigenstates
and a breakdown of the degeneracy theorem. We have taken φ(x = −∞) = 0
for convenience [114].
Table 12.2: One and two parameter families of potentials with bound states
in the continuum. These families are generated by applying the theorem
described in the text to scattering states u
0
and u
1
at energies E
0
and E
1
in
succession [115].
159
This figure "fig11.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig21.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig31.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig41.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig51.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig61.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig71.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig12.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig22.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig32.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig42.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig52.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig62.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig72.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig13.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig23.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig33.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig43.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig53.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig63.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
This figure "fig73.png" is available in "png"format from:
http://arXiv.org/ps/hepth/9405029v2
have the property of shape invariance. We describe new exactly solvable shape invariant potentials which include the recently discovered selfsimilar potentials as a special case. The connection between inverse scattering, isospectral potentials and supersymmetric quantum mechanics is discussed and multisoliton solutions of the KdV equation are constructed. Approximation methods are also discussed within the framework of supersymmetric quantum mechanics and in particular it is shown that a supersymmetry inspired WKB approximation is exact for a class of shape invariant potentials. Supersymmetry ideas give particularly nice results for the tunneling rate in a double well potential and for improving large N expansions. We also discuss the problem of a charged Dirac particle in an external magnetic ﬁeld and other potentials in terms of supersymmetric quantum mechanics. Finally, we discuss structures more general than supersymmetric quantum mechanics such as parasupersymmetric quantum mechanics in which there is a symmetry between a boson and a parafermion of order p.
2
Contents
1 Introduction 5 2 Hamiltonian Formulation of Supersymmetric Quantum Mechanics 13 2.1 State Space Structure of the SUSY Harmonic Oscillator . . . . 21 2.2 Broken Supersymmetry . . . . . . . . . . . . . . . . . . . . . . 23 3 Factorization and the Hierarchy of Hamiltonians 4 Shape Invariance and Solvable Potentials 4.1 General Formulas for Bound State Spectrum, Wave Functions and SMatrix . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Shape Invariance in More Than One Step . . . . . . . . . . . 4.3 Strategies For Categorizing Shape Invariant Potentials . . . 4.3.1 Solutions Involving Translation . . . . . . . . . . . . 4.3.2 Solutions Involving Scaling . . . . . . . . . . . . . . . 4.3.3 Solutions of MultiStep Shape Invariance . . . . . . . 4.3.4 Other Solutions . . . . . . . . . . . . . . . . . . . . . 4.4 Shape Invariance and Noncentral Solvable Potentials . . . . 4.5 Shape Invariance and 3Body Solvable Potentials . . . . . . 28 31 . . . . . . . . . 32 34 36 36 42 47 49 51 52
5 Operator Transforms – New Solvable Potentials from Old 56 5.1 Natanzon Potentials . . . . . . . . . . . . . . . . . . . . . . . 60 5.2 Generalizations of Ginocchio and Natanzon Potentials . . . . . 62 6 Supersymmetric WKB Approximation 6.1 SWKB Quantization Condition for Unbroken Supersymmetry 6.2 Exactness of the SWKB Condition for Shape Invariant Potentials 6.3 Comparison of the SWKB and WKB Approaches . . . . . . . 6.4 SWKB Quantization Condition for Broken Supersymmetry . . 66 66 68 70 71
7 Isospectral Hamiltonians 72 7.1 One Parameter Family of Isospectral Potentials . . . . . . . . 72 7.2 Generalization to nParameter Isospectral Family . . . . . . . 75 7.3 nSoliton Solutions of the KdV Equation . . . . . . . . . . . . 77
3
11. . . . . . . . . . . 10 Pauli Equation and Supersymmetry 11 Supersymmetry and the Dirac Equation 11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 SUSY and the Euclidean Dirac Operator . . . . . . . 9. . . .2 Orthosupersymmetric Quantum Mechanics . . Spectra and .5 Path Integral Formulation of the Fermion Propagator . . . . . . . .3 SUSY and the Dirac Particle in a Magnetic Field . .1 Dirac Equation With Lorentz Scalar Potential .1 Parasupersymmetric Quantum Mechanics . . . . . .1 Variational Approach . . . .2 δ Expansion Method . . 9. .1 General Formalism. . . . . . . 11. . . .2 Supersymmetry and the Dirac Particle in a Coulomb Field 11. . . .2 Bound States in the Continuum . . . . 129 13. . . . . 11. 9. . . . 87 88 91 93 97 99 102 103 105 108 109 111 12 Singular Superpotentials 114 12. . . . . . . . . 120 13 Parasupersymmetric Quantum Mechanics and Beyond 128 13. . . . . . . . . . . . . . . 115 12. . . . . . . . .3 Supersymmetry and Double Well Potentials . . . . .1 Superspace Formulation of Supersymmetric Quantum Mechanics 85 9 Perturbative Methods for Calculating Energy Wave Functions 9. . . . . . . . . . . . . . Negative Energy States and Breakdown of the Degeneracy Theorem . .8 Path Integrals and Supersymmetry 79 8. . . . .4 Supersymmetry and LargeN Expansions . . . 135 14 Omitted Topics 137 4 . . . .
This leads to a conceptual problem since the natural scale of symmetry breaking is the gravitational or Planck scale which is of the order of 1019 GeV. The algebra involved in SUSY is a graded Lie algebra which closes under a combination of commutation and anticommutation relations. One hopes that the scale of this breaking is of the order of the electroweak scale of 100 GeV in order that it can explain the hierarchy problem of mass diﬀerences. leptons and gauge bosons which have the same masses as their SUSY counterparts. One of the important predictions of unbroken SUSY theories is the existence of SUSY partners of quarks. Despite the beauty of all these uniﬁed theories. Several ambitious attempts have been made in the last two decades. 6]. and it is now widely felt that supersymmetry (SUSY) is a necessary ingredient in any unifying approach. For particle theorists. It was discovered in 1971 by Gel’fand and Likhtman [1]. In such theories. Thus. electrowea k. The fact that no such particles have been seen implies that SUSY must be spontaneously broken. Einstein’s general theory of relativity turns out to be a necessary consequence of a local gauged SUSY. 9].e. local SUSY theories provide a natural framework for the uniﬁcation of gravity with the other fundamental interactions of nature. strong. Various 5 . and gravitational interactions. there has so far been no experimental evidence of SUSY being realized in nature. 5. It was ﬁrst introduced in the context of the string models to unify the bosonic and the fermionic sectors. Gravity was generalized by incorporating SUSY to a theory called supergravity [8. i. It was later shown by Wess and Zumino [5] how to construct a 3 + 1 dimensional ﬁeld theory which was invariant under this symmetry and which had very interesting properties such as a softening of ultraviolet divergences as well as having paired fermionic and bosonic degrees of freedom.1 Introduction Physicists have long strived to obtain a uniﬁed description of all basic interactions of nature. SUSY oﬀered a possible way of unifying spacetime and internal symmetries of the Smatrix which avoided the nogo theorem of Coleman and Mandula [7] which was based on the assumption of a Lie algebraic realization of symmetries (graded Lie algebras being unfamiliar to particle theorists at the time of the proof of the nogo theorem). SUSY relates bosonic and fermionic degrees of freedom and has the virtue of taming ultraviolet divergences. Ramond [2] and Neveu and Schwartz [3] and later was rediscovered by several groups [4.
Because of the existence of the fermionic operators which commute with the Hamiltonian. Once the algebraic structure is understood. In any case. a topological index was introduced (the Witten index) by Witten [12] for studying SUSY breaking and several people studied the possibility that instantons provide the nonperturbative mechanism for SUSY breaking. evidence has been found for a dynamical SUSY relating eveneven and evenodd nuclei. There have also been applications of SUSY in atomic. It was in the context of this question that SUSY was ﬁrst studied in the simplest case of SUSY quantum mechanics (SUSY QM) by Witten [10] and Cooper and Freedman [11]. not just as a model for testing ﬁeld theory methods. the ideas of SUSY have stimulated new approaches to other branches of physics [15]. For example. the eigenfunctions and the Smatrices of the component parts of the full SUSY Hamiltonian. it was soon clear that this ﬁeld was interesting in its own right. In a subsequent paper. Once people started studying various aspects of SUSY QM. Over the last 10 years. In the work of Bender et al. Gradually a whole technology was evolved based on SUSY to understand the solvable potential problems. One purpose of this article is to review some of the major developments in this area. in the early days. let us note another remarkable aspect. the results follow and one never needs to return to the origin of the FermiBose symmetry. The Langevin equation and the method of stochastic quantization has a path integral formulation which embodies SUSY. SUSY was studied in quantum mechanics as a testing ground for the nonperturbative methods of seeing SUSY breaking in ﬁeld theory. In SUSY QM one is considering a simple realization of a SUSY algebra involving the fermionic and the bosonic operators.schemes have been invented to try to resolve the hierarchy problem. Before we present a brief historical development of supersymmetric quantum mechanics. the interpretation of SUSY QM as a degenerate Wess6 . These relationships will be exploited in this article to categorize analytically solvable potential problems. one obtains speciﬁc relationships between the energy eigenvalues. [13] a new critical index was introduced to study nonperturbatively the breakdown of SUSY in a lattice regulated theory. It was realized that SUSY gives insight into the factorization method of Infeld and Hull [14] which was the ﬁrst method to categorize the analytically solvable potential problems. Thus. including the idea of nonperturbative breaking of SUSY. condensed matter and statistical physics [15].
At the same time. This connection. 40. 37. 22. 38. 47. 20. Soon afterwards it was shown by using SUSY methods. Much later. 43]. which implicitly existed between the FokkerPlanck equation. atomic. The introduction by Witten [12] of a topological invariant to study dynamical SUSY breaking led to a ﬂurry of interest in the topological aspects of SUSY QM. This Russian paper remained largely unnoticed for several years. 36. statistical and condensed matter physics [35. the ﬁrst term in the h expansion gives the exact ¯ Witten index [24]. that the tunneling rate through double well barriers could be accurately determined [31. In another development. A path integral formulation of SUSY QM was ﬁrst given by Salomonson and van Holten [30]. the energy eigenvalue spectra could be obtained algebraically [44]. 23]. 27.Zumino ﬁeld theory in one dimension has not led to any further insights into the workings of SUSY QM. It was soon realized that the formalism of SUSY QM plus shape invariance (connected with translations of parameters) was intimately connected to the factorization method of Infeld 7 . 42. 39. the path integrals and the Langevin equation was then used to prove algorithms about the stochastic quantization as well as to solve nonperturbatively for the correlation functions of SUSY QM potentials using the Langevin equation. several workers extended ideas of SUSY QM to higher dimensionsal systems as well as to systems with large numbers of particles with a motivation to understand the potential problems of widespread interest in nuclear. A potential is said to be shape invariant if its SUSY partner potential has the same spatial dependence as the original potential with possibly altered parameters. 17. Using the WignerKirkwood h expansion it was shown that for systems in ¯ one and two dimensions. the index could display anomalous behavior [16. It is readily shown that for any SIP. Further. 26. 41. 28]. 21. the relationship between SUSY and the stochastic diﬀerential equations such as the Langevin equation was elucidated and exploited by Parisi and Sourlas [29] and Cooper and Freedman [11]. 48]. In 1983. a derivation of the AtiyahSinger index theorem was also given [25. using the methods of SUSY QM. 18. 32. 19. 34]. Various properties of the Witten index were studied in SUSY QM and it was shown that in theories with discrete as well as continuous spectra. 46. a list of SIPs was given and it was shown that the energy eigenfunctions as well as the scattering matrix could also be obtained algebraically for these potentials [45. 33. the concept of a shape invariant potential (SIP) within the structure of SUSY QM was introduced by Gendenshtein [44].
57] is also discussed in the paper of Cooper et al.and Hull [14]. is not necessary for exact solvability. 53]. [54] except for one which was later pointed out by Levai [55]. It turns out that the SWKB approximation preserves the exact SUSY rela8 . Infeld and Hull [14] generalized this method and obtained a wide class of solvable potentials by considering six diﬀerent forms of factorization. 61. 72]. and SUSY QM [68. thereby oﬀering glimpses of the deep connection between the methods of the inverse quantum scattering [67]. The intimate connection between the soliton solutions of the KdV hierarchy and SUSY QM was also brought out at this time [73. shape invariance and solvable potentials [56. δ expansion for the superpotential [78] and a SUSY inspired WKB approximation (SWKB) in quantum mechanics for the case of unbroken SUSY [79. These new potentials as well as multistep SIPs [59] have been studied. It is perhaps appropriate at this point to digress a bit and talk about the history of the factorization method. 76]. 63]. are both reformulations [50] of Riccati’s idea of using the equivalence between the solutions of the Riccati equation and a related second order linear diﬀerential equation. [52] where these authors show that shape invariance even though suﬃcient. 74. [52. This method was supposedly used for the ﬁrst time by Bernoulli and the history is discussed in detail by Stahlhofen [51]. The connection between SUSY. 81]. The factorization method was ﬁrst introduced by Schr¨dinger [49] to solve the hydrogen atom problem algebraically. 75. and their connection with selfsimilar potentials as well as with qdeformations has been explored [60. 70. o Subsequently. A partial classiﬁcation of the SIPs involving a translation of parameters was done by Cooper et al. 71. Approximate methods based on SUSY QM have also been developed. In yet another development. The general problem of the classiﬁcation of SIPs has not yet been solved. It turns out that the factorization method as well as the methods of SUSY QM including the concept of shape invariance (with translation of parameters). Recently. a new class of SIPs has been discovered which involves a scaling of parameters [58]. 80. several people showed that SUSY QM oﬀers a simple way of obtaining isospectral potentials by using either the Darboux [64] or AbrahamMoses [65] or Pursey [66] techniques. Three of the notable ones are the 1/N expansion within SUSY QM [77]. 62. 69. It turns out that in this case one only gets all the standard analytically solvable potentials contained in the list given by Dutt et al.
Dutt et al. Further. [96] have also developed a systematic higher order BSWKB expansion and using it have shown that in all the three (shape invariant) cases. 93]. This has subsequently been generalised to all orders in h [84. Further. 87. Recently. Further. In particular. 89] and it turns out that in many of the cases the SWKB does better than the usual WKB approximation. the higher order corrections to O(¯ 6 ) are zero. the energy eigenvalue h spectrum has also been obtained in the case of several nonSIP and it turns out that in many cases BSWKB does as well as (if not better than) the usual WKB approximation [96]. it is not only exact for large n (as any WKB approximation is) but by construction it is also exact for the ground state of V1 (x). It has recently been shown that for the cases of shape invariant three dimensional oscillator as well as for P¨schlTeller I and II potentials with broken SUSY. Some attempts have also been made to obtain the bound state eigenfunctions within the SWKB formalism [91. 85]. Energy eigenvalue spectrum has also been ¯ obtained for several nonSIP [86. 99]. Recently. Subsequently a systematic higher order SWKB expansion has been developed and it has been explicitly shown that to O(¯ 6 ) all the higher order h corrections are zero for these SIP [83]. this lowest o order BSWKB calculation gives the exact spectrum [95. 88. Several aspects of the Dirac equation have also been studied within SUSY QM formalism [100. 92.tions between the energy eigenvalues as well as the scattering amplitudes of the partner potentials [82]. Inomata and Junker [94] have derived the lowest order SWKB quantization condition (BSWKB) in case SUSY is broken. 101. it has been shown using the results of SUSY QM and shape invariance that whenever there is an analytically solvable Schr¨dinger problem in 1dimensional QM then there always exists o a corresponding Dirac problem with scalar interaction in 1+1 dimensions which is also analytically solvable. Based on a study of these and other examples. 97]. 102]. Besides it has been proved [45] that the lowest order SWKB approximation necessarily gives the exact spectra for all SIP (with translation). Recursion relations between the propagators pertaining to the SUSY partner potentials have been obtained and explicit expressions for propagators of several SIP have been obtained [98. it has been suggested that shape invariance is not only suﬃcient but perhaps necessary for the lowest order SWKB to give the exact bound state spectra [90]. it has been shown that there is always SUSY for massless Dirac equation in two as well as in four Euclidean 9 .
The ideas of SUSY are made more explicit through the example of one dimensional SUSY harmonic oscillator. The question of unbroken vs. we discuss the Hamiltonian formalism of SUSY QM. 114].dimensions. The SUSY algebra is given and the connection between the energy eigenvalues. 105]. As is clear from this (subjective) review of the ﬁeld. included in Sec. The question of singular superpotentials has also been discussed in some detail within SUSY QM formalism [110. 117. 54. 14 a list of the important topics missed in this review and given some references so that the intersested reader can trace back and study these topics further. 118]. decided not to pretend to be objective but cover only those topics which we believe to be important and which we believe have not so far been discussed in great detail in other review articles. 2. 124. the eigenfunctions and the Smatrics of the two SUSY partner Hamiltonians are derived. Very recently it has been shown that SUSY QM oﬀers a systematic method [115] for constructing bound states in the continuum [116. 107. 125. The SUSY of the Dirac electron in the ﬁeld of a magnetic monopole has also been studied [104. broken SUSY is also introduced at a pedagogical level using polynomial potentials of diﬀerent parity and the essential ideas of partner potentials are illustrated using the example of a one dimensional inﬁnite square well. We have been fortunate in the sense that review articles already exist in this ﬁeld where several of these missing topics have been discussed [119. The plan of the article is the following: In Sec. The formalism of SUSY QM has also been recently extended and models for parasupersymmetric QM [106. 10 . several aspects of SUSY QM have been explored in great detail in the last ten years and it is almost impossible to cover all these topics and do proper justice to them. 120. 113. We have therefore. The celebrated problem of the Dirac particle in a Coulomb ﬁeld has also been solved algebraically by using the concepts of SUSY and shape invariance [103]. 122. the classic calculation of Schwinger on pair production from strong ﬁelds can be dramatically simpliﬁed by exploiting SUSY. We have. We must also apologize to several authors whose work may not have been adequately quoted in this review article in spite of our best attempts. 123. 111. We have deliberately kept this section at a pedagogical level so that a graduate student should be able to understand and work out all the essential details. 100]. 108] as well as orthosupersymmetric QM [109] of arbitrary order have been written down. Also. however. 112. 121.
we discuss the connection between SUSY QM and factorization and show how one can always construct a hierarchy of p ≥ 1 Hamiltonians with known energy eigenvalues. We ﬁrst show that if the SUSY partner potentials satisfy an integrability condition called shape invariance then the energy eigenvalues. exact solutions of a number of three body problems in one dimension are obtained analytically [132].In Sec. eigenfunctions and Smatrices by starting from any given Hamiltonian with p bound states whose energy eigenvalues. the eigenfunctions and the Smatrices for these potentials can be obtained algebraically. Explicit expressions for the energy eigenvalues. 6 is devoted to a discussion of the SUSY inspired WKB approximation (SWKB) in quantum mechanics both when SUSY is unbroken and when it is spontaneously broken. 5. constitute a special case of the SIP. We then discuss satisfying the condition of shape invariance with translations and show that in this case the classiﬁcation of SIP can be done and the resultant list of solvable potentials include essentially all the popular ones that are included in the standard QM textbooks. eigenfunctions and Smatrices (reﬂection coeﬃcient R and transmission coeﬃcient T ) are known. Sec. we show that a wide class of noncentral but separable potential problems are also algebraically solvable by using the results obtained for the SIP [131]. It turns out that in most of these cases the resultant potentials are reﬂectionless and contain an inﬁnte number of bound states. In the unbroken case. the SWKB does as well as if not better than the WKB approx11 . we discuss the newly discovered one and multistep shape invariant potentials when the partner potentials are related by a change of parameter of a scaling rather than translation type. Further. As a by product. It is further shown that the recently discovered selfsimilar potentials which also statisfy qSUSY. their spectrum can be obtained algebraically. we discuss the solvable but nonshape invariant Natanzon and Ginocchio potentials and show that using the ideas of SUSY QM. the eigenfunctions and the transmission coeﬃcients are obtained in various cases. We also show here that even for many of the nonSIP. In Sec. Sec. 3. we ﬁrst develop a systematic higher order h expansion for the energy eigenvalues and then show that ¯ for the SIP with translation. We also show that the Natanzon potentials are not the most general solvable potentials in nonrelativistic QM. 4 is in a sense the heart of the article. the lowest order term in the hexpansion gives ¯ the exact bound state spectrum. Finally. shape invariance and operator transformations.
negative energy states and unpaired positive energy eigenstates.imation. In Sec. We also show here that whenever a Schr¨dinger problem in 1o dimensional QM is analytically solvable. Quantum mechanical models relating bosons and parafermions of order 12 . the δexpansion. we show that there is always a SUSY in the case of massless Dirac equation in two or four Euclidean dimensions in the background of external electromagnetic ﬁelds. In particular. We also discuss in some detail the connection of SUSY QM with classical stochastic processes and discuss how one can develop a systematic strong coupling and δ expansion for the Langevin equation. From here we are immediately able to construct the two as well as multisoliton solutions of the KdV equation. In particular. we discuss the important problem of a charged particle in a magnetic ﬁeld (Pauli equation) in two dimensions and show that there is always a SUSY in the problem so long as the gyromagnetic ratio is 2. We also show how the calculation of the fermion propagator in an external ﬁeld can be simpliﬁed by exploiting SUSY. Sec. we discuss the path integral formulation of SUSY QM as well as various subtleties associated with the Witten index [12] ∆ = (−1)F . Using the results of SUSY QM we then list a number of problems with nonuniform magnetic ﬁeld which can be solved analytically. 10. As an illustration we give plots of the one continuous parameter family of isospectral potentials corresponding to the onedimensional harmonic oscillator. we discuss more formal aspects of SUSY QM. then one can always obtain an exact solution of a corresponding Dirac problem with the scalar coupling. ln Sec. explicit breaking of SUSY. Sec. We then discuss the broken SUSY case and in that case too we develop a systematic hexpansion (BSWKB) for the energy eigenvalues. In Sec. 8. we discuss the question of SUSY QM in higher dimensions. largeN expansion. We ¯ show that even in the broken case the lowest order BSWKB gives the exact bound state spectrum for the SIP with translation. Sec. 12 contains a comprehensive discussion of the general problem of singular superpotentials. We also show here how to construct bound states in the continuum within the formalism of SUSY QM. 7 contains a description of how SUSY QM can be used to construct multiparameter families of isospectral and strictly isospectral potentials. energy splitting in double well potentials within the SUSY QM framework. 9 contains a description of several approximation schemes like the variational method. 11.
p are described in Sec. 2m dx 2m dx 13 (4) . 12). It is not usually appreciated that once one knows the ground state wave function (or any other bound state wave function) then one knows exactly the potential (up to a constant). Once we realize this. Various consequences of such models are discussed including the connection with the hierarchy of Hamiltonians as well as with strictly isospectral potentials. 2m dx2 ′′ h2 ψ0 (x) ¯ . we give a list of topics related to SUSY QM which we have not discussed and provide some reference s for each of these topics. 14. We also discuss a quantum mechanical model where instead there is a symmetry between a boson and an orthofermion of order p. 2 Hamiltonian Formulation of Supersymmetric Quantum Mechanics One of the key ingredients in solving exactly for the spectrum of one dimensional potential problems is the connection between the bound state wave functions and the potential. A† = √ + W (x) . It is shown that such models encompass p supersymmetries. 2m ψ0 (x) (1) (2) This allows a global reconstruction of the potential V1 (x) from a knowledge of its ground state wave function which has no nodes (we will discuss the case of using the excited wave functions later in Sec. in Sec. Then one has from the Schr¨dinger equation that the ground state wave function ψ0 (x) obeys o H1 ψ0 (x) = − so that V1 (x) = h2 d2 ψ0 ¯ + V1 (x)ψ0 (x) = 0 . Finally. Let us choose the ground state energy for the moment to be zero. 13. it is now very simple to factorize the Hamiltonian using the following ansatz: H1 = A† A (3) where h d ¯ −¯ d h A= √ + W (x) .
As we shall see. the energy eigenvalues. V2 (x) = W 2 (x) + √ W ′ (x) . To that end notice that the energy (1. Similarly. W (x) = − 2m ψ0 (x) (6) This solution is obtained by recognizing that once we satisfy Aψ0 = 0. H2 = − h2 d2 ¯ h ¯ + V2 (x) . the Schr¨dinger equation for H1 o (1) (1) (1) (1) H1 ψn = A† Aψn = En ψn (8) (9) implies (1) (1) (1) (1) H2 (Aψn ) = AA† Aψn = En (Aψn ) .2) eigenvalues of both H1 and H2 are positive semideﬁnite (En ≥ 0) .This allows us to identify h ¯ V1 (x) = W 2 (x) − √ W ′ (x) . The next step in constructing the SUSY theory related to the original Hamiltonian H1 is to deﬁne the operator H2 = AA† obtained by reversing the order of A and A† . 14 . we automatically have a solution to H1 ψ0 = A† Aψ0 = 0. 2 2m dx 2m (7) The potentials V1 (x) and V2 (x) are known as supersymmetric partner potentials. For n > 0. the Schr¨dinger equation for H2 o (2) (2) (2) (2) H2 ψn = AA† ψn = En ψn (10) (11) implies (2) (2) (2) (2) H1 (A† ψn ) = A† AA† ψn = En (A† ψn ) . The quantity W (x) is generally referred to as the “superpotential” in SUSY QM literature. the wave functions and the Smatrices of H1 and H2 are related. The solution for W (x) in terms of the ground state wave function is ′ h ψ0 (x) ¯ √ . A little simpliﬁcation shows that the operator H2 is in fact a Hamiltonian corresponding to a new potential V2 (x). 2m (5) This equation is the wellknown Riccati equation.
From eqs. (8)(11) and the fact that E0 = 0, it is clear that the eigenvalues and eigenfunctions of the two Hamiltonians H1 and H2 are related by (n = 0, 1, 2, ...)
(2) En = En+1 , (1) (1)
(1)
E0 = 0,
(1)
(1)
(12) (13) (14)
(2) ψn = [En+1 ]−1/2 Aψn+1 , (2) (2) ψn+1 = [En ]−1/2 A† ψn . (1) (1)
(2) Notice that if ψn+1 ( ψn ) of H1 (H2 ) is normalized then the wave function (1) (2) ψn (ψn+1 ) in eqs. (13) and (14) is also normalized. Further, the operato r A (A† ) not only converts an eigenfunction of H1 (H2 ) into an eigenfunction of H2 (H1 ) with the same energy, but it also destroys (creates) an extra node in the eigenfunction. Since the ground state wave function of H1 is annihilated by the operator A, this state has no SUSY partner. Thus the picture we get is that knowing all the eigenfunctions of H1 we can determine the eigenfunctions of H2 using the operator A, and vice versa using A† we can reconstruct all the eigenfunctions of H1 from those of H2 except for the ground state. This is illustrated in Fig. 2.1. The underlying reason for the degeneracy of the spectra of H1 and H2 can be understood most easily from the properties of the SUSY algebra. That is we can consider a matrix SUSY Hamiltonian of the form
H=
H1 0
0 H2
(15)
which contains both H1 and H2 . This matrix Hamiltonian is part of a closed algebra which contains both bosonic and fermionic operators with commutation and anticommutation relations. We consider the operators Q= Q† = 0 0 , A 0 0 A† 0 0 (16) (17)
in conjunction with H. The following commutation and anticommutation relation s then describe the closed superalgebra sl(1/1):
15
[H, Q] = [H, Q† ] = 0 , {Q, Q† } = H, {Q, Q} = {Q† , Q† } = 0 .
(18)
The fact that the supercharges Q and Q† commute with H is responsibl e for the degeneracy. The operators Q and Q† can be interpreted as operators which change bosonic degrees of freedom into fermionic ones and vice versa. This will be elaborated further below using the example of the SUSY harmonic oscillator. Let us look at a well known potential, namely the inﬁnite square well and determine its SUSY partner potential. Consider a particle of mass m in an inﬁnite square well potential of width L. V (x) = 0, = ∞,
(1)
0≤x≤L, −∞ < x < 0, x > L . 0≤x≤L,
(19)
The ground state wave function is known to be ψ0 = (2/L)1/2 sin(πx/L),
2 2
(20)
h ¯ π and the ground state energy is E0 = 2mL2 . Subtracting oﬀ the ground state energy so that we can factorize the Hamiltonian we have for H1 = H − E0 that the energy eigenvalues are (1) En =
n(n + 2) 2 2 hπ ¯ 2mL2
(21)
and the eigenfunctions are
(1) ψn = (2/L)1/2 sin
(n + 1)πx , L
0≤x≤L.
(22)
The superpotential for this problem is readily obtained using eq. (6) h π ¯ W (x) = − √ cot(πx/L) 2m L and hence the supersymmetric partner potential V2 is V2 (x) = h2 π 2 ¯ [2cosec2 (πx/L) − 1] . 2mL2 16 (23)
(24)
The wave functions for H2 are obtained by applying the operator A to the wave functions of H1 . In particular we ﬁnd that Thus we have shown using SUSY that two rather diﬀerent potentials corresponding to H1 and H2 have exactly the same spectra except for the fact that H2 has one fewer bound state. In Fig. 2.2 we show the supersymmetric partner potentials V1 and V2 and the ﬁrst few eigenfunctions. For convenience we have chosen L = π and h = 2m = 1. ¯ Supersymmetry also allows one to relate the reﬂection and transmission coeﬃcients in situations where the two partner potentials have continuum spectra. In order for scattering to take place in both of the partner potentials, it is necessary that the potentials V1,2 are ﬁnite as x → −∞ or as x → +∞ or both. Deﬁne: W (x → ±∞) = W± . (26) Then Let us consider an incident plane wave eikx of energy E coming from the direction x → −∞. As a result of scattering from the potentials V1,2 (x) one ′ would obtain transmitted waves T1,2 (k)eik x and reﬂected waves R1,2 (k)e−ikx . Thus we have ψ (1,2) (k, x → −∞) ψ (1,2) (k ′ , x → +∞) → eikx + R1,2 e−ikx , ′ → T1,2 eik x . (28)
2 V1,2 → W±
ψ0 ∝ sin2 (πx/L),
(2)
ψ1 ∝ sin(πx/L) sin(2πx/L) .
(2)
(25)
as x → ±∞.
(27)
SUSY connects continuum wave functions of H1 and H2 having the same energy analogously to what happens in the discrete spectrum. Thus we have the relationships: eikx + R1 e−ikx = N[(−ik + W− )eikx + (ik + W− )e−ikx R2 ], ′ ′ T1 eik x = N[(−ik ′ + W+ )eik x T2 ], (29)
where N is an overall normalization constant. On equating terms with the same exponent and eliminating N, we ﬁnd: R1 (k) = T1 (k) = W− + ik R2 (k), W− − ik W+ − ik ′ T2 (k), W− − ik 17
(30)
2m h ¯ V2 = A2 − A(A − α √ )sech2 . The two partner potentials are h ¯ V1 = A2 − A (A + α √ )sech2 αx . we have that T1 (k) = T2 (k). then the other partner will be of necessity reﬂectionless. This pole is on the positive imaginary axis only if W− < 0 in which case it corresponds to a zero energy bound state. ¯ So far we have discussed SUSY QM on the full line (−∞ < x < ∞). a free particle).that the reﬂectionless potentials are necessarily hdependent. V2 (x) corresponds to a constant potential so that 2m the corresponding V1 is a reﬂectionless potential.e.though it is not mentioned in ¯ most text books. that is the partner potentials have identical reﬂection and transmission probabilities. Many of these results have analogs for the ndimensional potentials with spherical symmetry. 2 k ′ = (E − W+ )1/2 . It is clear from these remarks that if one of the partner potentials is a constant potential (i. (31) A few remarks are now in order at this stage. φ) . (2) R1 (T1 ) and R2 (T2 ) have the same poles in the complex plane except that R1 (T1 ) has an extra pole at k = −iW− . in three dimensions one can make a partial wave expansion in terms of the wave functions: 1 ψnlm (r.where k and k ′ are given by 2 k = (E − W− )1/2 . For example. In this way we can understand the reﬂectionless potentials of the form V (x) = Asech2 αx which play a critical role in understanding the soliton solutions of the KdV hierarchy. (3) In the special case that W+ = W− . θ. One can in fact rigorously show. r 18 (34) . It is worth noting that V1 is hdependent. Let us consider the superpotential W (x) = A tanh αx . (4) When W− = 0 then R1 (k) = −R2 (k). φ) = Rnl (r)Ylm(θ.αx 2m (32) (33) ¯ We see that for A = α √h . (1) Clearly R1 2 = R2 2 and T1 2 = T2 2 .
We thus have seen that when H1 contained a known ground state wave function then we could factorize the Hamiltonian and ﬁnd a SUSY partner Hamiltonian H2 . l) → 1 l ′ ik′ r ′ [S (k )e − (−1)l e−ik r ] . A 19 . Now let us consider the converse problem. H2 only H1 (if at all) will have a normalizable zero energy ground state eigenfunction. we shall always choose W in such a way that amongst H1 . For this case we ﬁnd the relations: l S1 (k ′ ) = W+ − ik ′ l S2 (k ′ ) . This is ensured by choosing W such that W (x) is positive(negative) for large positive(negative) x. i. This deﬁnes H1 to have fermion number zero in our later formal treatment of SUSY. W+ + ik ′ (37) Here W+ = W (r → ∞). In this case there are two possibilities. The asymptotic form of the radial wave function for the l′ th partial wave is ψ(r. If there are no normalizable solutions of this form.e. Suppose we are given a superpotential W (x).Then it is easily shown [126] that the reduced radial wave function R satisﬁes the onedimensional Schr¨dinger equation (0 < r < ∞) o − l(l + 1)¯ 2 h h2 d2 ψ(r) ¯ + [V (r) + ]ψ(r) = Eψ(r) 2 2 2m dr 2mr (35) We notice that there is an eﬀective one dimensional potential which contains the original potential plus an angular momentum barrier. (38) A† ψ0 (x) = 0 ⇒ ψ0 (x) = N exp + h ¯ By convention. 2k ′ (36) where S l is the scattering function for the l′ th partial wave. h ¯ √ 2m x (2) (2) W (y) dy . Let us now be more precise. then H1 does not have a zero eigenvalue and SUSY is broken. S l (k) = eiδl (k) and δ is the phase shift. The candidate ground state wave function is the ground state for H1 (or H2 ) and can be obtained from: √ 2m x (1) (1) Aψ0 (x) = 0 ⇒ ψ0 (x) = N exp − W (y) dy .
symmetry of the Hamiltonian (or Lagrangian) can be spontaneously broken if the lowest energy solution does not respect that symmetry. as for example in a ferromagnet.. Stated another way. In general when one has a superpotential W (x) so that neither Q nor Q† annihilates the ground state as given by eq. (38). For the case of broken SUSY the operators A and A† no longer change the number of nodes and there is a 11 pairing of all the eigenstates of H1 and H2 . (42) then for n odd and g positive one always has a normalizable ground state wave function. 2. . 20 . (18 ) that the ground state energy must be zero in this case. then there is no normalizable ground state wave function. the ground state energy was indeed zero and hence the ground state wave function for the matrix Hamiltonian can be written: ψ0 (x) = (1) ψ0 (x) 0 (1) (41) where ψ0 (x) is given by eq. (39) then SUSY is broken and the potentials V1 and V2 have degenerate positive ground state energies. The precise relations that one now obtains are: (2) (1) En = En > 0. if the ground state energy of the matrix Hamiltonian is non zero then SUSY is broken.. We can deﬁne the ground state in our system by a two dimensional column vector: 0 >= ψ0 (x) = For SUSY to be unbroken requires Q0 >= Q† 0 >= 00 > (40) ψ0 (x) (2) ψ0 (x) (1) (39) Thus we have immediately from eq. n = 0. 1. where rotational invariance of the Hamiltonian is broken by the ground state. However for the case n even and g arbitrary. If we consider superpotentials of the form W (x) = gxn . (43) (44) (1) (2) (1) ψn = [En ]−1/2 Aψn . For all the cases we discussed previously.
whereas in the other cases it is always broken. a] = −a. a0 >= 0 . 21 (51) [a. 2 . 2mω x2 ). To that eﬀect one introduces instead of P and q the creation and annihilation operators a and a† . a† ] = a† .1 State Space Structure of the SUSY Harmonic Oscillator For the usual quantum mechanical harmonic oscillator one can introduce a Fock space of boson occupation numbers where we label the states by the occupation number n. 2 x a† = ( − ip). The usual harmonic oscillator Hamiltonian is 1 P2 + mω 2 q 2 . 2. p] = i. (45) while the relationship between the scattering amplitudes is still given by eqs. a† ] = 1. (50) N = a† a. h (47) [x. We put ¯ H = H¯ ω. q=( h 1/2 ¯ ) x. (30) or (37). h Then H = (p2 + Now introduce Then [N. The breaking of SUSY can be described by a topological quantum number called the Witten index [12] which we will discuss later. 4 P = (2m¯ ω)1/2 p. Let us however remember that in general if the sign of W (x) is opposite as we approach inﬁnity from the positive and the negative sides. H=N+ 2 The usual operator formalism for solving the harmonic oscillator potential is to deﬁne the ground state by requiring [N. then SUSY is unbroken.(1) (2) (2) ψn = [En ]−1/2 A† ψn . (46) H= 2m 2 Let us rescale the Hamiltonian in terms of dimensionless coordinates and momenta x and p so that we measure energy in units of hω. 1 . (48) (49) x a = ( + ip).
The n particle state (which is the n’th excited wave function in the coordinate representation) is then given by: nb >= a† n (n!) 0 > . we will choose 22 . ψ † } = {ψ. (52) where we have used the subscript b to refer to the boson sector as distinct from the fermions we will introduce below. nf > in this Fock space picture. . dx2 4 2 (57) The eﬀect of the last term is to remove the zero point energy. namely. Namely we write Q = aψ † and Q† = a† ψ where the matrix fermionic creation and annihilation operators are deﬁned via: ψ = σ+ = ψ † = σ− = 0 1 . {ψ † . Since the fermionic creation and annihilation operators obey anticommutation relations hence the fermion number is either zero or one.which leads to a ﬁrst order diﬀerential equation for the ground state wave function. ψ} = 0. As stated before. {ψ † . ψ † ]. The state vector can be thought of as a matrix in the Schr¨dinger picture o or as the state nb . ψ † ] = σ3 = 1 0 . For the case of the SUSY harmonic oscillator one can rewrite the operators Q (Q† ) as a product of the bosonic operator a and the fermionic operator ψ. ψ} = 1. 1 0 (53) (54) Thus ψ and ψ † obey the usual algebra of the fermionic creation and annihilation operators. 0 0 0 0 . 0 −1 (56) The SUSY Hamiltonian can be rewritten in the form H = QQ† + Q† Q = (− d2 x2 1 + )I − [ψ. (55) as well as obeying the commutation relation: [ψ.
The eﬀect of Q and Q† are now to relate the wave functions of H1 and H2 which have fermion number zero and one respectively but now there is no simple Fock space description in the bosonic sector because the interactions are nonlinear. Witten [12] proposed an index to determine whether SUSY is broken in supersymmetric ﬁeld theories. nf >. nf >. nf > = nb − 1. This is the bosonfermion degeneracy characteristic of all SUSY theories. a† nb . Q}. Thus in the general case. 8. i.e. ψ † nb . the operator a gets replaced by A in the deﬁnition of Q. 23 (61) . Then we can introduce the fermion number operator nF = 1 − σ3 1 − [ψ. Thus the ground state energy of the superHamiltonian must be zero since H = {Q† . ψnb . For the general case of SUSY QM. nf > = nb + 1. 2 2 (58) Because of the anticommutation relation. The index is deﬁned by △ = Tr(−1)F . for SUSY to be a good symmetry. Q† .2 Broken Supersymmetry As discussed earlier.the ground state of H1 to have zero fermion number. ψ † ] = . nf + 1 > . H = (− 2. ψ. (59) We now see that the operator Q† = −iaψ † has the property of changing a boson into a fermion without changing the energy of the state. nf >= nb . The action of the operators a. (60) 2 dx This form will be useful later when we discuss the Lagrangian formulation of SUSY QM in Sec. a† . nf can only take on the values 0 and 1. ψ † in this Fock space are then: anb . ψ † ]W ′ . we can rewrite the SUSY Hamiltonian in the form d2 + W 2 )I − [ψ. the operators Q and Q† must annihilate the vacuum. one writes Q = Aψ † and Q† = A† ψ. nf >= nb . nf − 1 >.
the fermion number nF ≡ F is deﬁned by 1 [1 − σ3 ] and we can represent (−1)F by the matrix σ3 . Let us show using a trivial calculation that for 24 . Since the bound states of H1 and H2 are paired. (65) As we said before. if this candidate ground state wave function does not fall oﬀ fast enough at ±∞ then Q does not annihilate the vacuum and SUSY is spontaneously broken. the breakdown of SUSY is related to the question of whether there is a normalizable wave function solution to the equation Q0 >= 00 > which implies ψ0 (x) = Ne− W (x)dx . (64) (63) In ﬁeld theory it is quite hard to determine if SUSY is broken nonperturbatively. which for SUSY quantum mechanics becomes △(β) = Tr[e−βH1 − e−βH2 ]. In ﬁeld theories the Witten index needs to be regulated to be well deﬁned so that one considers instead △(β) = Tr(−1)F e−βH . For our conventions the eigenvalue +1 corresponds to H1 and the bosonic sector and the eigenvalue −1 corresponds to H2 and the fermionic sector. Witten gives arguments that in general the index measures N+ (E = 0) − N− (E = 0) which is the diﬀerence △ between the number of Bose states and Fermi states of zero energy. and thus SUSY quantum mechanics became a testing ground for ﬁnding diﬀerent methods to understand nonperturbative SUSY breaking. In the general ﬁeld theory case.where the trace is over all the bound states and continuum states of the superHamiltonian. For SUSY QM. except for the case of unbroken SUSY where there is an extra state in the bosonic sector with E = 0 we expect for the quantum mechanics situation that △ = 0 for broken SUSY and △ = 1 for unbroken SUSY. In the quantum mechanics case. If we write the 2 eigenstates of H as the vector: ψn (x) = (+) ψn (x) (−) ψn (x) (62) then the ± corresponds to the eigenvalues of (−1)F being ±1.
β) =< ye−βH± x > = n ± ± e−βEn ψn (x)ψn (x) + ± ± dEe−βEn ψE (x)ψE (x). (67) △(β) = [ π We are interested in the regulated index as β tends to 0. x. y. 2π (66) Expanding the term proportional to σ3 in the exponent and taking the trace we obtain dpdx −β[p2 /2+W 2 /2 ]e sinh(βW ′(x)/2). Let us consider △(β) = T rσ3 [ dpdx −β[p2 /2+W 2 /2−σ3 W ′ (x)/2 ]e . then all the integrals are gamma functions and we explictly obtain △ = 1. 2π (68) If we directly evaluate this integral for any potential of the form W (x) = gx2n+1 . Following Akhoury and Comtet [16] one deﬁnes the heat kernels K± (x. Thus we see for these simple cases that the Witten index immediately coincides with the direct method available in the quantum mechanics case. y. β) which satisfy d d2 ( − 2 + W 2 ∓ W ′)K± = 0. β)] 25 . (Later we will discuss a path integral formulation for the regulated Witten Index). so that practically we need to evaluate △(β) = [ dpdx −β[p2 /2+W 2 /2] ]e (βW ′(x)/2). (69) dβ dx These have the following eigenfunction representation: K± (x. (70) In terms of the heat kernels one has △(β) = dx[K+ (x. Next let us discuss a favorite type of regularization scheme for ﬁeld theory– namely the heat kernel method. x. If instead W (x) = gx2n so that the candidate ground state wave function is not normalizable then the integrand becomes an odd function of x and therefore vanishes. β) − K− (x. which leads to a normalizable ground state wave function.two simple polynomial potentials the Witten index does indeed provide the correct answer to the question of SUSY breaking.
, or △(β) = N+ (E = 0) − N− (E = 0) +
∞ ED
dEe−βE (ρ+ (E) − ρ− (E)),
(71)
where ρ± corresponds to the density of states. What Akhoury and Comtet were able to show, was that in cases when W (x) went to diﬀerent constants at plus and minus inﬁnity, then the density of states factors for the continuum did not cancel and that △(β) could depend on β and be fractional at β = 0. We refer the interested reader to the original paper for further details. Another nonperturbative method for studying SUSY breaking in ﬁeld theory is to explicitly break SUSY by placing the theory on a lattice and either evaluating the path integral numerically or via some lattice non weakperturbative method such as the strong coupling (or high temperature) expansion. This method was studied in detail in [13, 21] and we will summarize the results here. The basic idea is to introduce a new parameter, namely the lattice spacing a. This parameter explicitly breaks SUSY so that the ground state energy of the system will no longer be zero, even in the unbroken case. One hopes that as the lattice spacing is taken to zero then the ground state energy will go to zero as a power of the lattice spacing if SUSY is unbroken, that is we expect E0 (a) = caγ , (72) where γ is a critical index which if greater than zero should be easy to measure in a Monte Carlo calculation. Measuring the ground state energy at two diﬀerent lattice spacings one studies: γ = ln a′ E0 (a′ ) /ln E0 (a) a (73)
in the limit a′ and a → 0. The case of broken SUSY is more diﬃcult because then we expect γ = 0 which is a hard measurement to make numerically. In this latter case it is easier to directly measure the ground state energy and show that it remains nonzero as one takes the lattice spacing to zero. To see how this works in quantum mechanics one can do a lattice strong coupling expansion of the Langevin equation which allows one to determine the ground state wave function of H1 as we shall show later. For the superpotential W (x) = gx3 , we expect to ﬁnd a positive critical index since here the candidate ground state wave function is proportional to 26
e−gx /4 and is normalizable. The ground state expectation values of xn for the Hamiltonian H1 can be determined by ﬁrst solving the Langevin equation dx + W (x) = η(t) dt (74)
4
and then averaging x(η(t)) over Gaussian noise whose width is related to h. ¯ Since by the virial theorem E0 = 3g 2 < x6 > −3g < x2 > , knowing the corelation functions will allow us to calculate the ground state energy. We ﬁrst put the Langevin equation on a time lattice (tn = na): ǫ(xn − xn−1 ) + gx3 = ηn , (75)
where ǫ = 1/a, which allows a solution by strong coupling expansion for large g. The result is xn = ( ηn 1/3 ǫ 1/3 −2/3 −1/3 ) + 2/3 (ηn−1 ηn − ηn ) + 0(ǫ2 ). g 3g (76)
As we will demonstrate in Sec. 8, the quantum mechanical expectation values of < xn (t) > are the same as the noise averaged expection values of xn (η) < xn (t) >= [Dη]P [η]xn (η(t)). (77)
On the lattice the path integral becomes a product of ordinary integrals which can be performed with: P [η] = Πi e−aη
2 (i)/2
a . 2π
(78)
The ground state energy on the lattice regulated theory then has the form E0 = where √
m
gz
n=0
Cn z n ,
(79)
2 z3 = √ a g 27
is a dimensionless length. The critical index can be determined from the logarithmic derivative of E0 with respect to z. Using Pad´ approximants to e extrapolate the lattice series to small lattice spacing we found that [21] E0 = ca1.16 (80)
verifying that SUSY is unbroken in the continuum limit. Using the same methods for the case W (x) = gx2 /2 we were able to verify that the ground state energy was not zero as we took the continuum limit. After verifying the applicability of this method in SUSY QM, Bender et al. then successfully used this method to study nonperturbative SUSY breaking in WessZumino models of ﬁeld theory [13].
3
Factorization and the Hierarchy of Hamiltonians
In the previous section we found that once we know the ground state wave function corresponding to a Hamiltonian H1 we can ﬁnd the superpotential W1 (x) from eq. (6). The resulting operators A1 and A† obtained from eq. 1 (4) can be used to factorize Hamiltonian H1 . We also know that the ground state wave function of the partner Hamiltonian H2 i s determined from the ﬁrst excited state of H1 via the application of the operator A1 . This allows a refactorization of the second Hamiltonian in terms of W2 . The partner of this refactorization is now another Hamiltonian H3 . Each of the new Hamiltonians has one fewer bound state, so that this process can be continued until the number of bound states is exhausted. Thus if one has an exactly solvable potential problem for H1 , one can solve for th e energy eigenvalues and wave functions for the entire hierarchy of Hamiltonians created by repeated refactorizations. Conversely if we know the ground state wave functions for all the Hamiltonians in this hierarchy, we can reconstruct th e solutions of the original problem. Let us now be more speciﬁc. From the last section we have seen that if the ground state energy of a Hamiltonian H1 is zero then it can always be written in a factorizable form as a product of a pair of linear diﬀerential operators. It is then clear that if (1) (1) the ground state energy of a Hamiltonian H1 is E0 with eigenfunction ψ0 then in view of eq. (3), it can always be written in the form (unless stated 28
2. the energy eigenvalues and eigenfunctions of the two Hamiltonians H1 and H2 are related by (1) (1) (1) (1) (2) (2) En+1 = En . 1 dx dx dx (1) (1) d2 + V1 (x). From Sec. dx2 (81) (82) The SUSY partner Hamiltonian is then given by H2 = A1 A† + E0 = − 1 where 2 ′ ′ V2 (x) = W1 + W1 + E0 = V1 (x) + 2W1 = V1 (x) − 2 (1) (1) d2 + V2 (x). A† = − + W1 (x) . 2 dx dx dx Continuing in this manner we obtain H3 = A2 A† + E1 = − 2 29 (1) (2) (1) (1) (2) (1) (86) (87) d2 + V3 (x). 1 2 where d d lnψ0 d A2 = + W2 (x) . W1 (x) = − . n denotes the energy level and (m) refers to the m’th Hamiltonian Hm . from now on we set h = 2m = 1 for simplicity): ¯ H1 = A† A1 + E0 = − 1 where A1 = d d d lnψ0 + W1 (x) . dx2 (83) d2 (1) lnψ0 . dx2 (88) . A† = − + W2 (x) . ψn = (En+1 − E0 )−1/2 A1 ψn+1 . W2 (x) = − .otherwise. (85) Now starting from H2 whose ground state energy is E0 = E1 one can similarly generate a third Hamiltonian H3 as a SUSY partner of H2 since we can write H2 in the form: H2 = A1 A† + E0 = A† A2 + E1 . 2 dx (84) (m) We will introduce the notation that in En .
... Further the reﬂection and transmission coeﬃcients (or phase shifts) for the hierarchy of Hamiltonians can be obtained in terms of 30 . (1) (1) (1) (1) (m) ψn = (En+m−1 − Em−2 )−1/2 .(En+m−1 − E0 )−1/2 Am−1 . In particular. (1) (1) (90) In this way.where 2 ′ V3 (x) = W2 + W2 + E1 = V2 (x) − 2 (1) d2 (2) lnψ0 dx2 (89) = V1 (x) − 2 Furthermore (3) En = En+1 = En+2 .Hp such that the m’th member of the hierarchy of Hamiltonians (Hm ) has the same eigenvalue spectrum as H1 except that the ﬁrst (m − 1) eigenvalues of H1 are missing in Hm . dx2 (m) (91) d d lnψ0 Am = + Wm (x) ... (2) (1) d2 (1) (2) ln(ψ0 ψ0 )..A1 ψn+m−1 d2 (1) (m−1) ). (93) Vm (x) = V1 (x) − 2 2 ln(ψ0 . it is clear that if the original Hamiltonian H1 has p(≥ 1) (1) (1) bound states with eigenvalues En . 2 dx (3) ψn = (En+1 − E0 )−1/2 A2 ψn+1 (1) (1) (1) (2) (2) (2) = (En+2 − E1 )−1/2 (En+2 − E0 )−1/2 A2 A1 ψn+2 . we can always write (m = 2..p): Hm = A† Am + Em−1 = − m where (1) d2 + Vm (x). Wm (x) = − dx dx One also has (m) En = En+1 (m−1) (1) (1) .. . = En+m−1 .. 3. and eigenfunctions ψn with 0 ≤ n ≤ (p− 1). then we can always generate a hierarchy of (p − 1) Hamiltonians H2 ... (92) = .. . knowing all the eigenvalues and eigenfunctions of H1 we immediately know all the energy eigenvalues and eigenfunctions of the hierarchy of p−1 Hamiltonians.ψ0 dx In this way.
(1) (m−1) ′ W+ − ik W+ − ik ′ (m−1) ik W− . analytically solvable potentials. (m−1) ik W− − ik + ik R1 (k). T1 of the ﬁrst Hamiltonian H1 by a repeated use of eq. More precisely. a2 ) + R(a1 ).2 (x) deﬁned in Sec. 14] although the language of SUSY is more appealing.2 (x.. the generalized operator method quickly yields all the bound state energy eigenvalues.R1 . we now show that the operator method for the harmonic oscillator can be generalized to the whole class of shape invariant potentials (SIP) which include all the popular. eigenfunctions as well as the scattering matrix. a2 is a function of a1 (say a2 = f (a1 )) and the remainder R(a1 ) is independent of x. Using the ideas of SUSY QM developed in Sec. Let us now explain precisely what one means by shape invariance.. then V1 (x. 2 are similar in shape and diﬀer only in the parameters that appear in them.. then they are said to be shape invariant. The shape invariance condition (96) is an integrability condition. we shall see that for such potentials. (1) (95) 4 Shape Invariance and Solvable Potentials Most text books on quantum mechanics describe how the one dimensional harmonic oscillator problem can be elegantly solved using the raising and lowering operator method. a1 ) satisfy the condition V2 (x. a1 ) and V2 (x. T1 (k). (1) k ′ = [E − (W+ )2 ]1/2 . If the pair of SUSY partner potentials V1. In particular we ﬁnd (1) W − (1) W− Rm (k) = Tm (k) = − + where k and k ′ are given by (1) (m−1) W − ik W− − ik − . a1 ) are said to be shape invariant. (94) k = [E − (W− )2 ]1/2 . Indeed. a1 ) = V1 (x. if the partner potentials V1. (30). 2 and an integrability condition called the shape invariance condition [44]. It turns out that this approach is essentially equivalent to Schr¨dinger’s method o of factorization [49. (96) where a1 is a set of parameters. Using this condition and the hierarchy of Hamiltonians discussed 31 ..
In particular.. one can easily obtain the energy eigenvalues and eigenfunctions of any SIP when SUSY is unbroken. 4. 32 (100) . as+1 ) + R(ak ) dx k=1 s−1 d2 = − 2 + V2 (x.in Sec..1 General Formulas for Bound State Spectrum. as ) + R(ak ). 3. Further. s = 1... a1 ) = N exp − (1) x W1 (y. (1) ψ0 (x. In view of eqs. let us construct a series of Hamiltonians Hs . following the discussion of the last section it is clear that if H1 has p bound states then one can construct p such Hamiltonians H1 . 2. it is then clear that Hs = − s−1 d2 + V1 (x. a1)dy .e. To that purpose. dx k=1 (99) Thus Hs and Hs+1 are SUSY partner Hamiltonians and hence have identical bound state spectra except for the ground state of Hs whose energy is E0 = k=1 (s) s−1 R(ak ). On repeatedly using the shape invariance condition (96).Hp and the n’th Hamiltonian Hn will have the same spectrum as H1 except that the ﬁrst n−1 levels of H1 will be absent in Hn . dx2 k=1 (98) where as = f s−1(a1 ) i. H2 . (96) and (98) we have Hs+1 s d2 = − 2 + V1 (x. 3. since SUSY is unbroken we know that E0 (a1 ) = 0.. (97) We now show that the entire spectrum of H1 can be very easily obtained algebraically by using the shape invariance condition (96). Let us compare the spectrum of Hs and Hs+1 . as ) + R(ak ).. Wave Functions and SMatrix Let us start from the SUSY partner Hamiltonians H1 and H2 whose eigenvalues and eigenfunctions are related by SUSY. the function f applied s − 1 times.
This follows from eq. (98) and the fact that E0 = 0. On going back from Hs to Hs−1 etc, we would eventually reach H2 and H1 whose ground state energy is zero and whose n’th level is coincident with the ground state of the Hamiltonian Hn . Hence the complete eigenvalue spectrum of H1 is given by
n − En (a1 ) = k=1
(1)
R(ak );
− E0 (a1 ) = 0.
(101)
We now show that, similar to the case of the one dimensional harmonic (1) oscillator, the bound state wave functions ψn (x; a1 ) for any shape invariant potential can also be easily obtained from its ground state wave function (1) ψ0 (x; a1 ) which in turn is known in terms of the superpotential. This is possible because the operators A and A† link up the eigenfunctions of the same energy for the SUSY partner Hamiltonians H1,2 . Let us start from the Hamiltonian Hs as given by eq. (98) whose ground state eigenfunction is (1) then given by ψ0 (x; as ). On going from Hs to Hs−1 to H2 to H1 and using eq. (14) we then ﬁnd that the n’th state unnormalized, energy eigenfunction (1) ψn (x; a1 ) for the original Hamiltonian H1 (x; a1 ) is given by
(1) ψn (x; a1 ) ∝ A† (x; a1 )A† (x; a2 )...A† (x; an )ψ0 (x; an+1 ), (1)
(102)
which is clearly a generalization of the operator method of constructing the energy eigenfunctions for the one dimensional harmonic oscillator. It is often convenient to have explicit expressions for the wave functions. In that case, instead of using the above equation, it is far simpler to use the identify [46] (1) (1) ψn (x; a1 ) = A† (x; a1 )ψn−1 (x; a2 ). (103) Finally, it is worth noting that in view of the shape invariance condition (96), the relation (30) between scattering amplitudes takes a particularly simple form W− (a1 ) + ik R1 (k; a1 ) = R1 (k; a2 ), (104) W− (a1 ) − ik T1 (k; a1 ) = W+ (a1 ) − ik ′ T1 (k; a2 ), W− (a1 ) − ik (105)
thereby relating the reﬂection and transmission coeﬃcients of the same Hamiltonian H1 at a1 and a2 (= f (a1 )). 33
4.2
Shape Invariance in More Than One Step
We can expand the list of solvable potentials by extending the shape invariance idea to the more general concept of shape invariance in two and even multisteps. We shall see later that in this way we will be able to go much beyond the factorization method and obtain a huge class of new solvable potentials [59]. Consider the unbroken SUSY case of two superpotentials W (x; a1 ) and ˜ (x; a1 ) such that V2 (x; a1 ) and V1 (x; a1 ) are same up to an additive constant ˜ W i.e. ˜ V2 (x; a1 ) = V1 (x; a1 ) + R(a1 ) (106) or equivalently ˜ ˜ W 2 (x; a1 ) + W ′ (x; a1 ) = W 2 (x; a1 ) − W ′ (x; a1 ) + R(a1 ). Shape invariance in two steps means that ˜ ˜ V2 (x; a1 ) = V1 (x; a2 ) + R(a1 ), that is ˜ ˜ ˜ W 2 (x; a1 ) + W ′ (x; a1 ) = W 2 (x; a2 ) − W ′ (x; a2 ) + R(a1 ). (109) (108) (107)
We now show that when this condition holds, the energy eigenvalues and eigenfunctions of the potential V1 (x; a1 ) can be obtained algebraically. First of all, let us notice that unbroken SUSY implies zero energy ground states ˜ for the potentials V1 (x; a1 ) and V1 (x; a1 ): E0 (a1 ) = 0,
(1)
˜ (1) E0 (a1 ) = 0 .
(110)
The degeneracy of the energy levels for the SUSY partner potentials yields
(1) (2) ˜ (2) ˜ (1) En (a1 ) = En+1 (a1 ); En (a1 ) = En+1 (a1 ).
(111)
From eq. (106) it follows that
(2) ˜ (1) En (a1 ) = En (a1 ) + R(a1 ),
(112)
so that for n = 0, these two equations yield E1 (a1 ) = R(a1 ). 34
(1)
(113)
Also, the shape invariance condition (108) yields
(1) ˜ (2) ˜ En (a1 ) = En (a2 ) + R(a1 ).
(114)
From the above equations one can then show that
(1) (1) ˜ En+1 (a1 ) = En−1 (a2 ) + R(a1 ) + R(a1 ).
(115)
On solving these questions recursively we obtain (n = 0, 1, 2, ...) E2n =
(1) n (1) n
˜ [R(ak ) + R(ak )],
(116)
k=1
E2n+1 =
˜ [R(ak ) + R(ak )] + R(an+1 ).
(117)
k=1
We now show that, similar to the discussion of the last subsection, the (1) bound state wave functions ψn (x; a1 ) can also be easily obtained in terms (1) ˜(1) of the ground state wave functions ψ0 (x; a1 ) and ψ0 (x; a1 ) which in turn ˜ are known in terms of the superpotentials W and W . In particular from eq. (106) it follows that
(2) ˜(1) ψn+1 (x; a1 ) ∝ A† (x; a1 )ψn (x; a1 ) ∝ A† (x; a1 )ψn (x; a1 ), (1)
(118)
while from eq. (108) we have
(1) ˜(1) ˜(2) ˜ ˜ ψn+1 (x; a1 ) ∝ A† (x; a1 )ψn (x; a1 ) ∝ A† (x; a1 )ψn (x; a2 ).
(119)
Hence on combining the two equations we have the identity
(1) (1) ˜ ψn+2 (x; a1 ) ∝ A† (x; a1 )A† (x; a1 )ψn (x; a2 ).
(120)
Recursive application of the above identity yields ˜ ˜ ψ2n (x; a1 ) ∝ [A† (x; a1 )A† (x; a1 )]...[A† (x; an )A† (x; an )]ψ0 (x; an+1 ),
(1) (1) (1) (1)
(121)
˜ ˜ ˜ ψ2n+1 (x; a1 ) ∝ [A† (x; a1 )A† (x; a1 )]...[A† (x; an )A† (x; an )]A† (x; an+1 )ψ0 (x; an+1 ), (122) 35
3. a second class of solutions was discovered in which the parameters a1 and a2 are related by scaling (a2 = qa1 ) [58. 59]. Firstly one notices the fact that the eigenvalue 36 . a1 ) ∝ A† (x. ˜ W− (a1 ) − ik (124) (125) thereby relating the reﬂection and transmission coeﬃcients of the same Hamiltonian at a1 and a2 . Remarkably enough.1 Solutions Involving Translation We shall now point out the key steps that go into the classiﬁcation of SIPs in case a2 = a1 + α [52]. 4. a1 ) = T1 (k. 53]. a1 ) = W− (a1 ) + ik W− (a1 ) − ik W+ (a1 ) − ik ′ W− (a1 ) − ik ˜ W− (a1 ) + ik R1 (k. Although the general problem is still unsolved. a1 )ψ0 (x. the parameters a1 and a2 are related to each other by translation (a2 = a1 + α) [52.3 Strategies For Categorizing Shape Invariant Potentials Let us now discuss the interesting question of the classiﬁcation of various solutions to the shape invariance condition (96). This is clearly an important problem because once such a classiﬁcation is available. In the ﬁrst class. it is easily shown that the relation (30) between the scattering amplitudes takes a particularly simple form R1 (k.where we have used the fact that ˜ ψ1 (x. ˜ W− (a1 ) − ik ˜ W+ (a1 ) − ik ′ T1 (k. It is clear that this procedure can be easily generalized and one can consider multistep shape invariant potentials and in these cases too the spectrum. the eigenfunctions and the scattering matrix can be obtained algebraically. then one discovers new SIPs which are solvable by purely algebraic methods. 4. two classes of solutions have been found and discussed. a1 ). a2 ). (1) (1) (123) Finally. a2 ). Last year. all well known analytically solvable potentials found in most text books on nonrelativistic quantum mechanic s belong to this class.
37 (132) . En is given by eq. It turns out that there are no solutions in case m ≥ 3 in eq. Thus. k2. β being constants. (101). since for any SIP. while there are only two solutions in case m = 2 i. α − 2 ≤ α ≤ 2. (130).) (131) with ci .. a1 ) = (k1 + c1 )g1 (x) + (k2 + c2 )g2 (x) + f1 (x). for any SIP. (126) where the upper bound is saturated by the square well potential and the lower bound is saturated by the Coulomb potential. (127) Now. it is easily seen that a fairly general factorizable form of W (x. k2 + β. a1 ) which produces the above kdependence in R(ak ) is given by m kγ γ (128) W (x. a1 ) = i=1 (ki + ci )gi (x) + hi (x)/(ki + ci ) + fi (x) (130) where a1 = (k1 .e.spectrum of the Schr¨dinger equation is always such that the n’th eigenvalue o En for large n obeys the constraint [133] 1/n2 ≤ En ≤ n2 . Note that this ansatz excludes all potentials leading to En which contain fractional powers of n. it follows that if R(ak ) ∼ then −3≤γ ≤1 (129) How does one implement this constraint on R(ak )? While one has no rigorous answer to this question.). the structure of En for large n is expected to be of the form En ∼ Cα nα . fi (x). One important condition is of course that only those superpotentials W are admissible which give a square integrable ground state wave function. On using the above ansatz for W in the shape invariance condition (96) one can obtain the conditions to be satisﬁed by the functions gi (x). a2 = (k1 + α. α. when W (x. hi (x)....
A. (134) where 0 ≤ x ≤ π/2α and ψ(x = 0) = ψ(x = π/2α) = 0. and W (x. (138) The ground state wave function of V1 (x. (134). (101). As an illustration. (137) In view of eq. B). A. (140) (1) (139) . B) is calculated from the superpotential W as given by eq. {a2 } = (A+α. (133) V2 (x. B) = A tan αx − B cot αx. A. A + α. B) = A tanh αr − B coth αr. we give expressions for the various shape invariant potentials V1 (x).which are given by W (r. A. B) ∝ (cos αx)s (sin αx)λ where s = A/α. parameters a1 and a2 and the (1) corresponding energy eigenvalues En [54. let us consider the superpotential given in eq. superpotentials W (x). B) = −(A + B)2 + A(A − α) sec2 αx + B(B − α)cosec2 αx V1 and V2 are often called P¨schlTeller I potentials in the literature. B+α). 55]. (134). B) are then given by (1) En n A > B > 0. B + α) + (A + B + 2α)2 − (A + B)2 and in this case {a1 } = (A. B) = V1 (x. A. R(a1 ) = (A+B+2α)2 −(A+B)2 . A. The corresponding partner potentials are V1 (x. B) = −(A + B)2 + A(A + α) sec2 αx + B(B + α)cosec2 αx (135) (136) = k=1 R(ak ) = (A + B + 2nα)2 − (A + B)2 .1. 38 λ = B/α. For the simplest possibility of m = 1. They o are shape invariant partner potentials since V2 (x. A. A. the bound state energy eigenvalues of the potential V1 (x. B > 0. A. In Table 4. We ﬁnd ψ0 (x. one has a number of solutions to the shape invariance condition (96).
B) is well behaved and hence acceptable as x −→ 0. {a1 }) = (− d + A tan αx − B cot αx)ψn−1 (x. 144) ( (143) (141) On deﬁning a new variable (142) It is then clear [46] that Rn (y. eq. W2 = A coth αr − Bcosech αr.B) = α(1 − y 2 ) d Rn−1 (y. (103) takes the form ψn (x. (134) and o (133) respectively have not been included in Table 4. Several remarks are in order at this time.1 since they are equivalent to the Scarf I (trigonometric) and generalized P¨schlTeller o superpotentials W1 = −A tan αx + Bsec αx. (103) one can also obtain explicit expressions for the bound state eigenfunctions (1) ψn (x. A. B + α) dy +[(A − B) − (a + B + α)y]Rn−1(y. A.The requirement of A. {a1}) with ψ0 being given by eq. B + α).1. 125] (1) and the energy eigenfunctions ψn (y) have been obtained in Table 4. (145) The procedure outlined above has been applied to all known SIPs [46. {a1})Rn (y. B) = (1 − y)λ/2 (1 + y)s/2 Pn (y).s−1/2 ψn (y. In particular. B > 0 that we have assumed in eq. π/2α. (139). A. A. dx y = 1 − 2 sin2 αx and factoring out the ground state state wave function ψn (y. we obtain Rn (y. A. B). The P¨schlTeller I and II superpotentials as given by eqs. where we also give the variable y for each case. in this case. A + α. {a1}) = ψ0 (y. A + α. (134) guaran(1) tees that ψ0 (x. 39 (146) . 1. {a2 }). Using this expression for the ground state wave function and eq. B) is proportional to the Jacobi polynomial so that the unnormalized bound state energy eigenfuncti ons for this potential are λ−1/2.
It ¯ would naively appear that if we had not put h = 1. A. ¯ (148) On redeﬁning A(A + hα) = a. Instead one can also write down (2) conditions for ψ0 = N exp( x W (y)dy) to be an acceptable ground state energy eigenfunction. b are hindependent parameters.1 would all be hdependent. In Table 4. 2 2 2 2 (147) which is just the P¨schlTeller II superpotential of eq. so that ψ0 = N exp(− x W (y)dy) is an acceptable ground state energy eigenfunction.1. the hdependent potential is ¯ given by (2m = 1) V1 (x. it is worth noting that in each and every case. Throughout this section we have used the convention of h = 2m = 1. 40 . It may be noted that the Coulomb as well as the harmonic oscillator potentials in ndimensions are also shape invariant potentials. (133) with reo deﬁned parameters. (134). B) = W 2 − hW ′ = −(A + B)2 + A(A + hα) sec2 αx ¯ ¯ 2 + B(B + hα)cosec αx. B > 0) for the su(1) perpotential (134).by appropriate redeﬁnition of the parameters [95]. For example. corresponding ¯ to the superpotential given by eq. we have given conditions (like A > 0. 2. the hdependence ¯ is only in the constant multiplying the xdependent function so that in each case we can always redeﬁne the constant multiplying the function and obtain an hindependent potential. 3. one can write W2 = ( αr B−A αr A+B ) tanh( ) − ( ) coth( ). ¯ (149) where a. then the shape in¯ variant potentials as given in Table 4. we then have a hindependent ¯ ¯ potential. For example. ¯ B(B + hα) = b. 4. How¯ ever.
V2 (x.. Before ending this subsection.5. 57] are in general not shape invariant.1) have been obtained so far in the case of multistep shape invariance and when a2 and a1 are related by translation. let us consider the shape invariant superpotential W = n tanh x. Thus we immediately 41 . (104) and (105) provided they are known in a ﬁnite strip.2.. No new solutions (apart from those in Table 4. a1 ) and T1 (k. Does this classiﬁcation exhaust all shape invariant potentials? It was believed that the answer to the question is yes [85. n) = n2 − n(n − 1)sech2 x. Let us ﬁrst notice that since for all the cases a2 = a1 + α. the reﬂection and transmission amplitudes R1 (k) and T1 (k) (or phase shift δ1 (k) for the threedimensional case) can also be calculated by operator methods. But is it also a necessary condition? This question has been discussed in detail in ref. a1 ) are determined for all values of a1 from eqs. we would like to remark that for the SIPs (with translation) given in Table 4. 6. it appears that this classiﬁcation has perhaps exhausted all SIPs where a2 and a1 are related by translation. n) = n2 − n(n + 1)sech2 x.1. What we have shown here is that shape invariance is a suﬃcient condition for exact solvability. [52] where it has been shown that the solvable Natanzon potentials [56. 7. (152) (150) (151) On going from V1 to V2 to V3 etc. 134] but as we shall see in the next subsection. the answer to the question is in fact negative. For example. we will ﬁnally reach the free particle potential which is reﬂectionless and for which T = 1. The two partner potentials V1 (x. are clearly shape invariant with a1 = n ..3. However..). where n is positive integer (1. a2 = n − 1 . hence R1 (k.
a choice motivated by the recent interest in qdeformed Lie algebras. In order to impose boundary conditions appropriate to the scattering problem.. 4. 0 < q < 1) rather than by translation. a straightforward method [48] for calculating the scattering amplitudes by making use of the n’th state wave functions as given in Table 4. one must use the wave number k ′ so that the asymptotic behaviour is exp(ik ′ x) as x → ∞.(−1 − ik) Γ(−n − ik)Γ(n + 1 − ik) = Γ(−ik)Γ(1 − ik). So far none of these potentials have been obtained in a closed form but are obtained only in a series form. Let us consider an expansion of the superpotential of the from W (x. . however. two modiﬁcations of the bound state wave functions have to be made: (i) instead of the parameter n labelling the number of nodes. [48]. However.(1 − ik) (−n − ik)(−n + 1 − ik). n) = (n − ik)(n − 1 − ik).1 and that there were no more shape invariant potentials. a1 ) = ∞ j=0 gj (x)aj 1 (154) 42 . 59].. n) is given by T1 (k. We shall see that many of these potentials are reﬂectionless and have an inﬁnite number of bound states. In this way the scattering amplitude of all the SIPs of Table 4.. very recently we have been able to discover a huge class of new shape invariant potentials [58. It turns out that for many of these new shape invariant potentials.. (ii) the second solution of the Schr¨dinger equation must be kept o (it had been discarded for bound state problems since it diverged asymptotically).3. the parameters a2 and a1 are related by scaling (a2 = qa1 . V2 .conclude that the series of potentials V1 . (153) The scattering amplitudes of the Coulomb [47] and the potential corresponding to W = A tanh x + Bsech x [48] have also been obtained in this way.1 have been calculated in ref. it was believed that the only shape invariant potentials are those given in Table 4. are all reﬂectionless and the transmission coeﬃcient of the reﬂectionless potential V1 (x.. There is..2 Solutions Involving Scaling For almost nine years.1.
(154) and (155) in the shape invariance condition (96). even though the construction is noninvariant.. we will construct below potentials by expanding in a1 . since the choice of parameter is usually an integral part of constructing a SIP. a procedure whose legitimacy and outcome are clearly dependent on our choice of parameter and hence reparameterization. which corresponds to R0 = 0. Implicit constraints on this choice 43 . 2. However. ′ g1 (x) + 2d1 g0 (x)g1 (x) = r1 d1 . then the shape invariance condition essentially ﬁxes the gn (x) (and hence W (x. 3. n = 1. it is in practice part of the ansatz.. dn = (1 − q n )/(1 + q n ).and further let a2 = qa1 . For example. . We shall see that. Let us ﬁrst consider the special case g0 (x) = 0.. The general solution of eq. the resulting potentials will still be invariant under redeﬁnition of a1 . n = 1. j=1 where rn ≡ Rn /(1 − q n ). 2.. . ′ can be recast as a2 = a′1 + α merely by taking logarithms. writting R(a1 ) in the form R(a1 ) = ∞ j=0 Rj aj . 0<q<1 . (96). n−1 ′ gn (x) + 2dn g0 (x)gn (x) = rn dn − dn (157) (158) gj (x)gn−j (x). a1 )) and determines the shape invariant potential. (159) This set of linear diﬀerential equations is easily solvable in succession to give a general solution of eq. (158) then turns out to be n−1 gn (x) = dn dx[rn − gj (x)gn−j (x)]. j=1 (160) where without loss of generality we have assumed the constants of integration to be zero. 1 (156) and equating powers of a1 yields [59. (155) This is slightly misleading in that a reparameterization of the form a2 = qa1 . We thus see that once a set of rn are chosen. 58] ′ 2g0 (x) = R0 . On using eqs.
58] and in these instances the results obtained from shape invariance and selfsimilarity are entirely equivalent and the ShabatSpiridonov selfsimilarity condition turns out to be a special case of the shape invariance condition. rn = 0. a2 ) = √ √ qW ( qx. a1 ) . if any one rn (say rj ) is taken to be nonzero then one again obtains selfsimilar potentials [59. 61]. Consider for example. (163) which corresponds to the selfsimilar W of Shabat and Spiridonov [60. It must be emphasized here that shape invariance is a much more general concept than selfsimilarity. where β1 = d1 r1 . 1 3 15 g1 (x) = d1 r1 x. It is worth pointing out that these selfsimilar potentials can be shown to satisfy qsupersymmetry [62]. n ≥ 3. This makes the potentials V1. if we choose more than one rn to be nonzero. of which the ﬁrst three are 1 2 g2 (x) = d2 r2 x − d2 r1 x3 . In this case the eq. (164) ∞ j=1 n−1 dn βj βn−j (2n − 1) j=1 (161) βn = − (162) βj aj x2j−1 = 1 √ √ a1 F ( a1 x) . 3 1 2 2 3 g3 (x) = − d1 r1 d2 r2 d3 x3 + d3 r1 d2 d3 x5 . n ≥ 2. The simplest case is r1 > 0 and rn = 0. Using eq. (160) takes a particularly simple form gn (x) = βn x2n−1 . this gives W (x. (165) Notice that in this case W (x) contains only odd powers of x. For example.are that the resulting ground state wave function be normalizable and the spectrum be sensibly ordered which is ensured if R(q n a1 ) > 0. and hence W (x. (160) one can readily calculate all the gn (x). then SIP are obtained which are not contained within the selfsimilar ansatz. It may be noted here that instead of choosing rn = 0. a1 ) = For a2 = qa1 .2 (x) symmetric in x and also guarantees unbroken SUSY. 44 . n ≥ 2.
an+1 ) → 0. a1 )][ik − W (∞. a1 ) = where W (∞. We can also calculate the transmission coeﬃcient of this symmetric potential (k = k ′ ) by using the relation (125) and the fact that for this SIP a2 = qa1 .. j=1 [ik + W (∞. n = 0. one gets W (x. a1 ) = [ik − W (∞. 2. since we have taken g0 (x) = 0. (1 − q) (1 − q 2 ) (166) where Γ1 = d1 r1 a1 . a1 ) vanishes and the transmission coeﬃcient is given by T1 (k. an+1 ) [ik + W (∞. a2 )].. an+1 = q n a1 → 0(0 < q < 1) and. 2. as n → ∞. (1 − q j ) 45 (171) . (101) and (156) and are given by (0 < q < 1) (1) En (a1 ) = Γ1 (1 + q 2 )(1 − q 2n ) (1 + q)(1 − q n ) + Γ2 . Γ2 = d2 r2 a2 while the unnormalized ground state wave 1 function is ψ0 (x. a2 )]. a1 ) = exp[− (1) x4 x2 (Γ1 + Γ2 ) + (d2 Γ2 + 2d3 Γ1 Γ2 + d4 Γ2 ) + 0(x6 )] (167) 1 2 2 4 The wave functions for the excited states can be recursively calculated from the relation (103). r2 = 0 can be readily generalized to an arbitrary number of nonzero rj .. an )] (168) . The energy eigenvalues for this case are given by (Γj ≡ dj rj aj ) 1 (1) En (a1 ) = j Γj (1 + q j )(1 − q jn ) .[ik − W (∞. aj )] ∞ (170) The above discussion keeping only r1 . n = 0. a1 )][ik + W (∞. aj )] ... This corresponds to a free particle for which the reﬂection coeﬃcient R1 (k. 1. . aj ) = E∞ − Ej (1) (1) [ik − W (∞.[ik + W (∞.. Repeated application of the relation (125) gives T1 (k.The energy eigenvalues follow immediately from eqs. an )] T1 (k. 1.. (169) Now. .
the spectrum can be immediately obtained by using eqs. giving rise to a multiparameter deformed oscillator spectrum. As a consistency check. Numerical solutions pose no serious problems. in the case of an arbitrary number of nonzero Rj (in addition to R0 ). It is worth pointing out here that. (170). it is given by (1) En = nR0 + j Γj (1 + q j )(1 − q nj ) (1 − q j ) (173) which is the spectrum of a qdeformed harmonic oscillator [136. (101) and (156). the solution of the shape invariance condition (96) is the standard one dimensional harmonic oscillator with W (x) = Rx/2 while in the limit q → 0 the solution is the RosenMorse superpotential corresponding to the one soliton solution given by √ √ (172) W (x) = R tanh( Rx). have checked numerically that the Schr¨dinger equation solved with o numerically obtained potentials indeed has the analytical energy eigenvalues given above. we have the discrete as well as the continuous spectra. An unfortunate feature of the new SIPs obtained above is that they are not explicitly known in terms of elementary functions but only as a Taylor series about x = 0.Further. unlike the usual qoscillator where the space is noncommutative but the potential is normal (ω 2 x2 ). for q even slightly less than one..All these potentials are also symmetric and reﬂectionless with T1 as given by eq. g2 (x). but the potential is deformed. It is also worth noting that the number of bound states increase discontinuously from just one at q = 0 to inﬁnity for q > 0. whereas for q = 1 the spectrum is purely discrete. For example. (157) it then follows that g0 (x) = R0 x/2 rather than being zero.. Hence the general solution as obtained above with 0 < q < 1 can be regarded as the multiparameter deformation of the hyperbolic tangent function with q acting as the deformation parameter. At q = 1. From eq. The limits q → 0 and q → 1 of all these potentials are simple and quite interesting. From numerical calculations one ﬁnds that the superpotential 46 .. in our approach the space is commutative. Barclay et al. One can again solve the set of linear diﬀerential equations (157) and (158) in succession yielding g1 (x). 137]. Questions about series convergence naturally arise. Further. Finally. let us consider the solution to the shape invariance condition (96) in the case when R0 = 0.
4. A very unusual new shape invariant potential has also been obtained [59] corresponding to r1 = 1. whereas V1 (x) is a double well potential. n ≥ 3 (with q = 0. 1 ˜ W (x. then being reﬂectionless. rn = 0.and the potential are as shown in Figs. its shape invariant partner potential V2 (x) is a single well. a1 ) = ∞ j=0 gj (x)aj .3. one can immediately obtain constraints on the reﬂectionless SIPs obtained above [59]. The answer is yes [59] and as an illustration we now explicitly obtain the multiparameter deformations of the two soliton case by using the formalism of twostep shape invariance as developed earlier in this section. momentum. if we regard them as a solution of the KdV equation at time t = 0.3. rn = 0.1 and 4. 4. the fact that these are reﬂectionless symmetric potentials can be used to constrain them quite strongly. however. an obvious question to ask is if one can also obtain deformations of the multisoliton solutions.2 [59] corresponding to the case when r1 = 0.. see no evidence of the oscillations in W and V as reported by Degasperis and Shabat [63]. This is because. we write R and R in the form R(a1 ) = ∞ j=0 Rj aj .. n ≥ 2.3 Solutions of MultiStep Shape Invariance Having obtained potentials which are multiparameter deformations of the RosenMorse potential corresponding to the one soliton solution. 4. In this case.75) which is shown in Fig. energy . It is worth pointing out that even though the potentials are not known in a closed form in terms of elementary functions. a1 ) = ∞ j=0 hj (x)aj 1 (174) ˜ Further.. it is well known that such solutions as t → ±∞ will break up into an inﬁnite number of solitons of the form 2 2ki sech2 ki x [138. 1 ˜ R(a1 ) = ∞ j=0 ˜ 1 Rj aj (175) 47 .3 and a = 0. r2 = −1. These authors. 139]. On using the fact that the KdV solitons obey an inﬁnite number of conservation laws corresponding to mass. Let us take the scaling ansatz a2 = qa1 and expand the two superpoten˜ tials W and W in powers of a1 W (x.
˜ W = 2 R tanh ˜ ˜ Rx. The ﬁrst two of each are g1 = g2 ˜ (R1 − R1 ) x...e. n ≥ 3 one can readily calculate all the gn (x) and hn (x).. (107) and (109) and equating powers of a1 yields (n=0.2.Using these in eqs. (117) and (118) are [59] (1) E2n (a1 ) = ˜ 1 (1 − q ) (Rj + Rj )aj (1 − q j ) j=1 2 2 jn E2n+1 (a1 ) = j=1 (1) Rj aj 1 2 jn (1 − q j(n+1) ) ˜ 2 j (1 − q ) Rj a1 + (1 − q j ) (1 − q j ) j=1 (180) ˜ For the special case of R2 = R2 = 0 the spectrum was obtained previously by Spiridonov from the considerations of the twostep selfsimilar potentials [135].1. The limit q → 0 of the above equation is particularly simple and yields the RosenMorse potential corresponding to the two soliton case i. For example when ˜ ˜ R0 = R0 = 0 (and hence g0 (x) = h0 (x) = 0) and further Rn = Rn = 0. h2 = (1 − q) (1 − q 2 ) 3 x 2 ˜2 ˜ [(1 + q)R1 + (1 + q)(1 − q 2 )R1 − 2q(1 − q)R1 R1 ] − 2 3(1 + q)(1 − q) (179) h1 = The energy eigenvalues which follow from eqs. 178) (1 − q 2 ) 3(1 − q)3 ˜ R2 x (R1 − qR1 ) x. (1 − q) ˜ (R2 − R2 ) x3 2 ˜2 ˜ ( = x+ [(1 − q)(R1 − R1 ) − 2(1 + q)R1 R1 ]. W = 48 ˜ ˜ R tanh Rx (181) .) n n ′ gn + j=0 n gj gn−j = j=0 ∞ hj hn−j − h′n + Rn ′ ˜ gj gn−j − q n gn + Rn (176) (177) h′n + j=0 hj hn−j = q n j=0 This set of linear equations is easily solved in succession.
2. the ﬁrst three g ′s are 1 3 g2 (x) = (R2 − qR1 )x − R1 x3 .. . (183) 2m+1 ′ g2m+1 (x) + j=0 gj (x)g2m+1−j (x) = R2m+1 .. 3 2 3 5 2 R1 (qR1 − R2 )x3 + R1 x . Generalization to arbitrary p is straightforward [59]. One can also consider 49 . 4.. when only R1 and R2 are nonzero. Some of the other possibilities are: a2 = qap with p= 2.. j (186) The q → 0 limit of these equations again correspond to the RosenMorse potential corresponding to the one soliton solution. The corresponding spectrum turns out to be (E0 (a1 ) = 0) (1) En = (1) (185) R1 q n j=1 (a1 q)2 j−1 + R2 q2 n j=1 (a1 q)2 .4 Other Solutions So far we have obtained solutions where a2 and a1 are related either by scaling or by translation.3..p steps and thereby obtain multiparameter deformations of the 3.3. g3 (x) = 3 15 g1 (x) = R1 x. and a2 = qa1 /(1 + pa1 ).e p = 2..˜ provided R = 3R. . On using eqs. 4. (184) which can be solved in succession and one can readily calculate all the gn (x). (155) and (156) one obtains the set of equations 2m ′ g2m (x) + j=0 gj (x)g2m−j (x) = q m m j=0 ′ gj (x)gm−j (x) − q m gm (x) + R2m ... This procedure can be immediately generalized and one can consider shape invariance with a scaling ansatz in 3. .p soliton RosenMorse potential. Let 1 us ﬁrst consider the case when a2 = qa2 1 (182) i.. Are there shape invariant potential where a2 and a1 are neither related by scaling nor by translation? It turns out that there are other possibilities for obtaining new shape invariant potentials. For example. n = 1... 4.
can one also obtain deformations of the other SIPs given in Table 4. Can one solve the Schr¨dinger equation for these potentials directly? o 50 . One can similarly consider solutions to the shape invariance condition (96) in case qa1 (187) a1 = 1 + pa1 when 0 < q < 1 and pa1 << 1 so that one can expand (1 + pa1 )−1 in powers of a1 . Have we exhausted the list of SIP? We now have a signiﬁcantly expanded list but it is clear that the possibilities are far from exhausted. 53]? 3.shape invariance in multisteps along with this ansatz thereby obtaining deformations of the multisoliton RosenMorse potential. the spectrum has so far only been obtained algebrically. In fact it appears that there are an unusually large number of shape invariant potentials. Just as we have obtained qdeformations of the reﬂectionless RosenMorse and harmonic oscillator potentials. For example.1? 2. when only R1 and R2 are nonzero. We would like to close this subsection with several comments. How does one classify all these potentials? Do these potentials include all solvable potentials [52. For those SIP where a2 and a1 are not related by translation. for all of which the whole spectrum can be obtained algebraically. then one can show that the ﬁrst two nonzero gn are g1 (x) = R1 x . 1. (1 + q) g2 (x) = (R2 + pqR1 (1 − q) x3 )x − (1 + q) (1 + q)2 (1 + q 2 ) 3 (1) (188) and the energy eigenvalue spectrum is (E0 = 0) (1) En = R1 n q j−1a1 (q j−1 a1 )2 + R2 1−q j−1 1−q j−1 2 j=1 [1 + pa1 ( 1−q )] j=1 [1 + pa1 ( 1−q )] n (189) Generalization to the case when several Rj are nonzero as well as shape invariance in multisteps is straight forward.
in the other cases they are only known formally as a Taylor series. As an illustration. all the SIP obtained so far have inﬁnite number of bound states and are either reﬂectionless (or have no scattering). let us discuss noncentral separable potentials in spherical polar coordinates. φ) = V (r) + V (θ) V (φ) + 2 2 2 r r sin θ (190) where V (r). the answer to the question is yes. Most of these potentials are either one dimensional or are central potentials which are again essentially one dimensional but on the half line. in the former case one has also many SIP with nonzero reﬂection coeﬃcients. The equation for the wave function ψ(r. whereas in the latter case. whereas in the former case the potentials are explicitly known in a closed form in terms of simple functions. As has been shown recently [131].4. In particular. It may be worthwhile to enquire if one can also algebraically solve some noncentral but separable potential problems. let us see why the Schr¨dinger equation with a potential of the form given o by eq. There is a fundamental diﬀerence between those shape invariant potentials for which a2 and a1 are related by translation and other choices (like a2 = qa1 ). θ. a number of potential problems can be solved algebraically. First. θ. It turns out that the problem is algebraically solvable so long as the separated problems for each of the coordinates belong to the class of SIP. V (θ) and V (φ) are arbitrary functions of their argument. In spherical polar coordinates the most general potential for which the Schr¨dinger equation is separable is given by [140] o V (r.4 Shape Invariance and Noncentral Solvable Potentials We have seen that using the ideas of SUSY and shape invariance. φ) is −( ∂ 2 ψ 2 ∂ψ 1 ∂2ψ ∂ψ 1 d2 ψ = (E − V )ψ (191) + ) − 2 ( 2 + cotθ ) − 2 2 ∂2r r ∂r r ∂θ ∂θ r sin θ ∂φ2 51 . θ. φ) coordinates. Secondly. (190) is separable in the (r. 4.
there has been renewed interest in the onedimensional manybody problem in connection with the physics of spin chains [146. as we show below. 4. φ) as ψ(r. (194) and (195) may be solved algebraically by choosing appropriate SIPs for V (φ). As has been shown in ref. φ) = R(r) H(θ) K(φ). 2 dφ (193) d2 H 1 + [V (θ) + (m2 − ) cosec2 θ] H(θ) = l2 H(θ). Further.5 Shape Invariance and 3Body Solvable Potentials Many years ago. in a classic paper. 149]. one could use this trick to discover a number of new exactly solvable threebody potentials in one dimension. 145]. [131]. Wolfes [142] used Calogero’s method to obtain analytical solutions of the same problem in the presence of an added threebody potential of a special form. Also. one obtains the following equations for the functions K(φ). the three Schr¨dinger equations given by o (193). Later. [131]. θ. H(θ) and R(r) : − − d2 K + V (φ)K(φ) = m2 K(φ). Generalization of this technique to noncentral but separable potentials in other orthogonal curvilinear coordinate systems as well as in other dimensions is quite straightforward.It is convenient to write ψ(r. Details can be found in ref. (195) − 2 + [V (r) + dr r2 where m2 and l2 are separation constants. 144. 52 . (192) in eq. θ. 147. 148. r (sin θ)1/2 (192) Substituting eq. there has been a recent generalization of Calogero’s potential for Nparticles to SUSY QM [150]. Attention then shifted to the exact solutions of the manybody problem and the general question of integrability [143. (194) dθ2 4 1 (l2 − 4 ) d2 R ] R(r) = E R(r). (191) and using the standard separation of variables procedure. Calogero [141] gave the complete solution of the Schr¨dinger equation for three particles in one dimension. Recently. V (θ) and V (r). interacting o via twobody harmonic and inversesquare potentials.
we give more examples of threebody potentials that can be solved exactly. This may therefore be mapped on to a onebody problem in two dimensions. φ. one can discover a number of new 3body potentials for which the 3body problem in one dimension can be solved exactly [132]. The last two terms on the righthand side of eq. (199) 3 To see why potentials given by eqs. (197) are just cyclic permutations of the ﬁrst. From the above discussion. In the above equation. 1 r 2 = [(x1 − x2 )2 + (x2 − x3 )2 + (x3 − x1 )2 ]. Henceforth. have two degrees of freedom. Wolfes [142] showed that a threebody potential VW = f [(x1 + x2 − 2x3 )−2 + (x2 + x3 − 2x1 )−2 + (x3 + x1 − 2x2 )−2 ] (197) is also solvable when it is added to VC .The purpose of this subsection is to show that using the results for SIPs derived above. The important point to note is that three particles in one dimension. (200) 3 53 . with or without the pairwise centrifugal term. There is a rule of thumb that if one can solve a 3body problem then one can also solve the corresponding nbody problem. Calogero [141] considered the case where the two dimensional potential is noncentral but separable in polar coordinates r. (196) to (198) are solvable. deﬁne the Jacobi coordinates 1 R = (x1 + x2 + x3 ). (196) where g > −1/2 to avoid a collapse of the system. after the center of mass motion is eliminated. Thus hopefully one can also solve the corresponding statistical mechanics problem. (196). then the whole problem can be solved exactly. In this subsection. such terms occuring in any potential are referred to as “cyclic terms”. Our ﬁrst solvable example is the threebody potential of the form √ 3f1 (x1 + x2 − 2x3 ) V1 = + cyclic terms . it is clear that if the potentials in each of the coordinates r and φ are chosen to be shape invariant. Calogero’s [141] solution of the threebody problem is for the potential VC = ω 2 /8 i<j (xi − xj )2 + g i<j (xi − xj )−2 . (198) 2r 2 (x1 − x2 ) which is added to the Calogero potential VC of eq.
VW as well as V1 are all noncentral but separable potentials in the polar coordinates r. r In all three cases. (202) x= Obviously. n. (201) 2 6 After elimination of the center of mass part from the Hamiltonian. take the potential VC + V1 .φ. which may be mapped into polar coordinates x = r sin φ .and ydegrees of freedom remain. (x2 − x3 ) = √ 2 r sin(φ + 4π/3). in the variable φ.1 are Enl = 3/2 ω(2n + Bl + 1) . To examine the angular part of the eigenfunction Fl (φ). and the wave function can be written as Rnl (r) (204) ψnl (r. (206) 1 −1 (207) ( 3/2)ωr 2]LBl [ ( 3/2)ωr 2 ]. φ) = √ Fl (φ). (203) (x3 − x1 ) = It turns out that VC . φ have ranges 0 ≤ r ≤ ∞ and 0 ≤ φ ≤ 2π. As a result. 1. the radial wave function obeys the equation d2 3 2 2 (Bl2 − 1 ) 4 − 2+ ω r + Rnl = Enl Rnl (r). n 4 2 where Bl > 0. Then.. 2 dr 8 r (205) where Bl2 is the eigenvalue of the Schr¨dinger equation in the angular vario able. (x1 − x2 ) = √ 2 r sin(φ + 2π/3). It is straightforward to show that √ 2 r sin φ. y= . only the x. Eq. (205) corresponds to a SIP and the eigenvalues Enl and the eigenfunctions Rnl which follow from Table 4. y = r cos φ . the Schr¨dinger equation is o Rnl = r Bl exp[ − 3 π d2 π g 3 3 cosec2 [φ + 2(m − 1) ] + f1 cot[φ + 2(m − 1) ] Fl (φ) + 2 dφ 2 m=1 3 2 m=1 3 54 .. 2. the variables r.(x1 − x2 ) (x1 + x2 − 2x3 ) √ √ . the Schr¨dinger equation o separates cleanly in the radial and angular variables. l = 0.
a given value of φ deﬁnes a speciﬁc ordering. (208) reduces to − 9 9 d2 + g cosec2 3φ + f1 cot 3φ 2 dφ 2 2 Fl = Bl2 Fl (φ). for indistiguishable particles.= Bl2 Fl (φ). crossing is not allowed. (211) Now this is again a SIP and hence its eigenvalues and eigenfunctions are 9 2 f1 /(l + a + 1/2)2 . On using the identities 3 m=1 (208) cosec2 [φ + 2(m − 1)π/3] = 9cosec2 3φ 3 (209) (210) m=1 cot[φ + 2(m − 1)π/3] = 3cot 3φ eq. For 0 ≤ φ ≤ π/3.β where Pn is the Jacobi polynomial and Bl2 = 9(l + a + 1/2)2 − a = 1/2(1 + 2g)1/2 . For distinguishable particles. (213) is zero outside 0 ≤ φ ≤ π/3. symmetrized or antisymmetrized wave functions may be constructed. eq. (212) (213) (214) It must be noted that g > −1/2 for meaningful solutions. As expected.1. 142]. (203) implies x1 ≥ x2 ≥ x3 . it is easily show n that the other exactly solvable potentials are [132] 1 V2 = ω 2 (xi − xj )2 + 3g [(x1 + x2 − 2x3 )−2 + cyclic terms ] 8 i<j √ 3 3 f1 (x1 − x2 ) − + cyclic terms .−˜ if˜ π ˜ 4l 4l l)(sin 3φ)l exp[ Fl (φ) = exp(−i ˜ (i cot 3φ) ]Pl 2 4 ˜ l α. Following Calogero. 16 l− 1 l+ 1 3 f1 φ −˜ if˜ . the wave function for the other ranges may be constructed. and other ranges of φ correspond to diﬀerent orderings [141. and Fl (φ) of eq. Proceeding in the same way and using the results of Table 4. For singular repulsive potentials. Similarly. we recover the results for the Calogero potential VC in the limit f1 = 0. 2 r 2 (x1 + x2 − 2x3 ) 55 (215) .
VW .e. Natanzon [56] wrote down (what he thought at that time to be) the most general solvable potentials i. V4 . the energy eigenvalues and eigenfunctions are known implicitly rather than explicitly as in the shape invariant case. for most of them. in this case one can obtain exact solutions of threebody problems for all the SIPs discussed above (i. V3 = ω 2 (xi −xj )2 +g (xi −xj )−2 − √ 8 i<j (x1 − x2 )2 6r i<j (216) V4 = 1 2 ω (xi − xj )2 + 3 g [ (x1 + x2 − x3 )−2 + cyclic terms ] 8 i<j 1 f3 + √ 3 2 r (x1 − x2 ) + cyclic terms .e.f3 (x1 + x2 − 2x3 ) 1 + cyclic terms . VC .. (x1 + x2 − 2x3 )2 (217) Notice that in all these cases one has combined the SIP in the angular variable with the harmonic conﬁnement. VC + VW ) along with the attractive 1 potential [132]. Coordinate transformations which satisfy these restrictions give rise to new solvable problems.. The three body scattering problem has also been studied in these cases after droping the harmonic term [132]. When the relationship between coordinates is implicit. It turns out that most of these potentials are not shape invariant. The possibility of replacing the harmonic conﬁnement term with the attractive 1 r type interaction has also been considered. r 5 Operator Transforms – New Solvable Potentials from Old In 1971. for which the Schr¨dinger equation o can be reduced to either the hypergeometric or conﬂuent hypergeometric equation .. One might ask if one can obtain these solutions from the explicitly solvable shape invariant ones. One strategy for doing this is to start with a Schr¨dinger equation which is exactly solvable (for example one having a o SIP) and to see what happens to this equation under a point canonical coordinate transformation. Further. In order for the Schr¨dinger equation to be mapped o into another Schr¨dinger equation.V1 . then the new solutions are only implicitly 56 . Further. there are severe restrictions on the nature o of the coordinate transformation. Note that in this case one has both discrete and continuous spectra.
We start from the onedimensional Schr¨dinger o equation d2 (218) − 2 + [V (x) − En ] ψn (x) = 0. while if the relationship is explicit then the newly found solvable potentials are also shape invariant [127. 129]. dz 2 (223) f ′′ V − En f ′2 ¯ . (220) dx dz The ﬁrst step in obtaining a new Schr¨dinger equation is to change coordin o ates and divide by f 2 so that we have: − d2 f′ d [V − En ] ψn = 0 − + 2 dz f dz f2 ¯ ψ = f −1/2 ψ. − − V (En ) = f2 4f 2 2f (224) In order for this to be a legitimate Schr¨dinger equation. (221) To eliminate the ﬁrst derivative term. dx Consider the coordinate transformation from x to z deﬁned by f (z) = so that dz . Kostelecky et al. one next rescales the wave function: (222) ¯ Adding a term ǫn ψ to both sides of the equation yields − where ¯ d2 ψn ¯ ¯ ¯ + [V (En ) + ǫn ]ψn = ǫn ψn . 128. using an explicit coordinate transformation. the Coulomb problem in d dimensions with the ddimensional harmonic oscillator. the potential o ¯ V (En ) + ǫn must be independent of n. [130] were able to relate. This can be achieved if the quantity G deﬁned by G = V − En + ǫn f 2 (225) 57 .determined. In a more speciﬁc special application of these ideas. Let us see how this works. dx (219) d d =f . Other explicit applications of the coordinate transformation idea are found in the review article of Haymaker and Rau [120].
is independent of n. This can be made clear by a simple example. (226) (227) (228) By our previous argument we must choose f = f2 = ( to be of the form: (229) dz 2 A ) = 2 + Br 2 + C. This further requires that in order for V to b e independent of n. dr 2 r2 dz dr 2 /2 . the energy eigenvalues of the three dimensional harmonic oscillator are give by En = 4αn . dr r The solution of this equation gives z = z(r) which in general is not invertible so that one knows r = r(z) only implicitly as discussed before.the three dimensional harmonic oscillator in a given angular momentum state with angular momentum β. (231) so that the condition we want to satify is V − 4αn + ǫn f 2 = G = 58 D + Er 2 + F. z = r 2 /2 . How can one satisfy this condition? One way is to have f 2 and G to have the same functional dependence on x(z) as the original ¯ potential V . The reduced ground state wave function for that angular momentum is ψ0 (r) = r β+1 e−αr so that the superpotential is given by W (r) = αr − (β + 1)/r. However for special cases one has an invertible function. r2 (232) . the parameters of V must change with n so that the wave function corresponding to the n’th energy level of the new Hamiltoni an is related to a wave function of the old Hamiltonian with parameters which depen d on n. and H1 is given by: H1 = − β(β + 1) d2 + + α2 r 2 − 2α(β + 3/2). Let us for simplicity now choose f = r. Let us consider an exactly solvable problem . (230) As discussed earlier.
In fact. which describes the strength of the oscillator potential. 5. α(l. This is nicely illustrated in Fig. In general. n) = 2(l + n + 1) 4(l + 1 + n)2 (234) We now choose γ = α2 (l. r cannot be explicitly found in terms of z. (233) We see that for the quantities D. be dependent on n. (236) Higher wave functions will have values of α which depend on n so that the diﬀerent wave functions correspond to harmonic oscillator solutions with different strengths.1 can be interrelated by point canonical coordinate transformations [128. we obtain the relations: Z 2 e4 Ze2 . (237) 59 . the new Hamiltonian written in terms of z is now d2 l(l + 1) Ze2 Z 2 e4 ¯ H =− 2 + − + dz z2 z 4(l + 1)2 (235) and the ground state wave function of the hydrogen atom is obtained from the ground state wave function of the harmonic oscillator via ¯ ψ0 = f 1/2 ψ0 = xl+1 e−α(l. F = −4αn − 2α(β + 3/2) = −2Ze2 .Equating coeﬃcients we obtain D = β(β + 1) . n = 0) so that the ground state energy is zero. 129]. All the exactly solvable shape invariant potentials of Table 4.1 . These energy levels are those of the hydrogen atom. E = ǫn + α2 = γ . solving the above three equations and choosing β = 2l + 1/2. E. Explicitly. ǫn = γ − .n=0)x . F to be independent of n one needs to hav e that α. and one has dz/dr = f = whose solution is: A/r 2 + Br 2 + C .
D.1 Natanzon Potentials The more general class of Natanzon potentials whose wave functions are hypergeometric functions can be obtained by making an operator transformation of the generalized P¨schlTeller potential whose Hamiltonian is: o d2 β(β − 1) α(α + 1) − + (α − β)2 + 2 2 2 dr sinh r cosh r This corresponds to a superpotential H1 = A† A1 = − 1 W = α tanh r − βcoth r 60 (242) (243) .√ A + C r2 + B r4 A log(r 2 ) + z = 2 √ 2√ √ 2 A log(2 A + C r + 2 A A + C r 2 + B r 4 ) − 2 √ √ C log(C + 2 B r 2 + 2 B A + C r 2 + B r 4 ) √ + . βn = −1/2 + D + 1/4 − Aǫn . then the potential that one obtains is the particular class of Natanzon potentials whose wave functions are conﬂuent hypergeometric functions in the variable r and are thus only implicitly known in terms of the true coordinate z. Equating coeﬃcients. (241) 5. we get an implicit expression for the eigenvalues: Cǫ − F √ n − 2 E − Bǫn D + 1/4 − Aǫn = 2n + 1 (240) as well as the state dependence on α and β necessary for the new Hamiltonian to be energy independent: αn = E − Bǫn . If we choose this general coordinate transformation. even the expression for the transformed potential is only known in terms of r: ¯ V (z. 4 B √ (238) This clearly is not invertible in general. E) = 1/f 2 [D/r 2 + Er 2 + F − f ′2 /4 + f f ′′/2] (239) and thus only implicitly in terms of z. In fact.
δ) = 2 ( ) (248) − +σ− 2 + f sinh2 r 4f 2f cosh2 r and thus only implicitly in terms of z.and a ground state wave function given by: ψ0 = sinhβ r cosh−α r . 61 (250) . γ. Equating coeﬃcients. we get an implicit expression for the eigenvalues: [(γ + 1/2)2 − Aǫn ]1/2 − [(δ − 1/2)2 − Bǫn ]1/2 − (σ − Cǫn )1/2 = 2n + 1 (249) as well as the state dependence on α and β necessary for the new Hamiltonian to be energy independent: αn = [(γ + 1/2)2 − Aǫn ]1/2 − 1/2 . z = − + + + + √ √ A tan−1 { √ 2 A −2 A + 2 B − C + 2 (A + B) cosh 2r + C cosh2 2r } B log{−A + 3 B − C + (A + B + C) cosh 2r √ 2 B −2 A + 2 B − C + 2 (A + B) cosh 2r + C cosh2 2r} √ C log{A + B + C cosh 2r √ C −2 A + 2 B − C + 2 (A + B) cosh 2r + C cosh2 2r} √ B log{2 sinh2 r} (247) However the expression for the transformed potential is only known in terms of r: 1 δ(δ − 1) γ(γ + 1) f ′′ f ′2 ¯ V (z. (245) (244) The most general transformation of coordinates from r to z which preserves the Schr¨dinger equation is described by: o f2 = A dz B + C = ( )2 2 − 2 dr sinh r cosh r −3 A + B − C + (A + B − C) cosh 2r (246) From this we obtain an explicit expression for z in terms of r. βn = [(δ − 1/2)2 − Bǫn ]1/2 + 1/2 . The energy eigenvalues were discussed earlier and are En = (α − β)2 − (α − β − 2n)2 .
It is convenient to use the original approach [57. 56] to deﬁne the Ginocchio and Natanzon potentials and then obtain their generalization to more complicated solvable potentials whose wave functions are sums of hypergeometric functions. [52] it was explicitly shown that the Ginocchio and Natanzon potentials. β. If we denote by r the coordinate appearing in the Schr¨dinger equation : o [− d2 + V (r)]ψ(r) = 0 dr 2 (251) where −∞ ≤ r ≤ ∞ and z the coordinate describing the hypergeometric function F (α.5. The operator which allows one to construct the eigenfunctions of H2 from those of H1 converts single hypergeometric functions of the implicitly known coordinate to sums of hypergeometric functions. The resulting potentials are ratios of polynomials in the transformed coordinate in which the wave functions are sums of hypergeometric functions. z) appearing in the diﬀerential equation: z(1 − z) d2 F dF + [γ − (α + β + 1)z] − αβF = 0. The original method of obtaining the Natanzon potentials was to ﬁnd a coordinate transformation that mapped the Schr¨dinger equation onto the o equation for the hypergeometric functions. 2 dz dz (252) where 0 ≤ z ≤ 1. then the transformation of coordinates is given by [56] 2z(1 − z) dz = . The transformed coordinate is only implicitly known in terms of the coordinate system for the Schr¨dinger o equation. This process yields totally new solvable potentials. whose wave functions are hypergeometric functions of an implicitly determined variable are not shape invariant and thus one could construct new exactly solvable potentials from these using the factorization method.2 Generalizations of Ginocchio and Natanzon Potentials In ref. dr R1/2 R = az 2 + (c1 − c0 − a)z + c0 (253) In terms of the coordinate z the most general potential V (r) which correspon ds to a Schr¨dinger equation getting mapped into a hypergeometric o 62 . γ. These generalized potentials have coordinates which are only implicitly known.
. The potential V (r) is a function of six dimensionless parameters f. (254) λ > 1.function equation is given by: V (r) = [f z(z − 1) + h0 (1 − z) + h1 z + 1]/R + a + (c1 − c0 )(2z − 1) 5△ z 2 (1 − z)2 − } {a + z(z − 1) 4R R2 where We note that z is only implicitly known in terms of r. The Ginocchio potentials are more easily discussed. and f = (ν +1/2)2 −1. β. γ) of the hypergeometric function below. h1 = −1. The class of potentials called the Ginocchio potentials is a subclass of the Natanzon class which has only two independent variables nu and λ where c0 = 0. a. λ < 1. The transformation is given by dy = (1 − y 2 )[1 + (λ2 − 1)y 2] dr where now −1 ≤ y ≤ 1 and the potential is given by: 1 V (r) = {−λ2 ν(ν + 1) + (1 −λ2 )[5(1 −λ2 )y 4 −(7 −λ2 )y 2 + 2]}(1 −y 2). (256) 4 Although y is not known explictly in terms of r. h1 . c1 = 1/λ4 . These six parameters will be related to the energy ǫn and the parameters (α. one obtains from the Schr¨dinger equation. (257) (255) △ = (a − c0 − c1 )2 − 4c0 c1 . however by considering them as derived from coordinate transformations which map the Schr¨dinger equation into the diﬀerential equation for the Gegenbauer polyo nomials. the equation for the Gegenbauer polynomials: o 63 rλ2 = tanh−1 (y) + (λ2 − 1)1/2 tan−1 [(λ2 − 1)1/2 y]. c0 and c1 . By introducing the variables: x = λy/[g(y)]1/2 where g(y) = 1 + (λ2 − 1)y 2 . a = c1 −1/λ2 . h0 = −3/4. it is implicitly known via: rλ2 = tanh−1 (y) − (1 − λ2 )1/2 tanh−1 [(1 − λ2 )1/2 y]. h0 . and changing the variable x to an angle x = cos θ.
(260) (261) Thus the superpotential is : ψ ′ (r) 1 W (r) = − = (1 − λ2 )y(y 2 − 1) + µ0 yλ2 ψ(r) 2 (262) For the Natanzon potential (254) we have that the energy eigenvalues are given by: (2n + 1) = (1 − aǫn + f )1/2 − (1 − c0 ǫn + h0 )1/2 − (1 − c1 ǫn + h1 )1/2 ≡ αn − βn − δn . n (258) (259) The corresponding energy eigenvalues and eigenfuctions are µn λ2 = [λ2 (ν + 1/2)2 + (1 − λ2 )(n + 1/2)2]1/2 − (n + 1/2). (µ ψn = (1 − y 2)µn /2 [g(y)]−(2µn +1)/4 Cn n +1/2) (x). z). αn − n. 1 + βn . (265) Once we have the superpotential and the explicit expression for the wave functions we can determine all the wave functions of the partner potential using the operator: 64 .[ where µ2 − 1/4 1/2 d2 + (n + µ + 1/2)2 − ]Z ψn = 0 dθ2 sin2 θ Z= 1 − y 2 1/2 λ ǫn = −µ2 λ4 . (263) and the corresponding (unnormalized) energy eigenfunctions are given by : (264) It is easy to determine the superpotential from the ground state wave function: W (r) = [δ0 z − (1 + β0 )(1 − z)]/R1/2 + [(c1 − c0 − a)z + 2c0 ](1 − z)/(2R3/2 ). ψn = R1/4 z βn /2 (1 − z)δn /2 F (−n.
2 + βn .13 that the series of Hamiltonians form a parasupersymmetry of order p if the original Hamiltonian has p bound states.A1 = dz d 2z(1 − z) d d + W (r) = + W (z) = + W (z) dr dr dz R1/2 dz for the case of the Natanzon potential and the operator: A1 = dy d d d + W (r) = + W (y) = (1 − y 2 )[1 + (λ2 − 1)y 2] + W (y) dr dr dy dy for the case of the Ginocchio potentials. The arguments are identical for the Ginocchio class with the hypergeometric functions now being restricted to being Gegenbauer polynomials. z) and for the partner potential V2 = W 2 + W ′ one has − 2z(1 − z)n(n + 1 + βn + δn )F (−n + 1. z)] (266) (2) 2 2 V2 = ǫ0 + [(β0 + δ0 )(β0 + δ0 + 2)z(z − 1) + (β0 − 1)(1 − z) + (δ0 − 1)z + 1]/R +[a − [c1 (3β0 + δ0 ) + c0 (β0 + 3δ0 )a(1 − β0 − δ0 )]/[z(z − 1)] −(2z − 1)[(c1 − c0 )(β0 + δ0 + 1) + a(β0 − δ0 )]/[z(z − 1)] + 7△/(4R)]z 2 (1 − z 2 )/R2 . the partner to the origina l Natanzon Hamiltonian: ψn−1 = R−1/4 z βn /2 (1−z)δn /2 [(1−z)[βn −β0 −z(δn −δ0 )]F (−n. these operators take a single hypergeometric function into a sum of two hypergeometric functions which cannot be reexpressed as a single hypergeometric function except in degenerate cases where one obtains the shape invariant special cases discussed earlier. 1+βn . However this algebra did not lead to any new insights. We shall also see in Sec. Explicitly. The potentials one obtains can have multiple local minima and several of these are displayed in [52]. (267) By using the hierarchy of Hamiltonians one can construct in the usual manner more and more complicated potentials which are ratios of higher and higher order polynomials in z (as well as R) which are isospectral to the original Natanzon potential except for the usual missing states. In [52] it was also shown that the ser ies of Hamiltonians and SUSY charges form the graded Lie algebra sl(1/1) ⊗ SU(2). 65 . αn − n + 1. In general. The wave functions will be sums of hypergeometric functions. one has for the unnormalized eigenfunctions of H2 . αn −n.
the WKB quantization condition for the one dimensional potential V (x) is [151] x2 x1 2m[En − V (x)] dx = (n + 1/2)¯ π.1 SWKB Quantization Condition for Unbroken Supersymmetry As we have seen in previous sections. Bandrauk and Campbell [79] obtained the lowest order SWKB quantization condition in case SUSY is unbroken and showed that it yields energy eigenvalues which are not only exact for large quantum numbers n (as any WKB approximation scheme should in the semiclassical limit) but which are also exact for the ground state (n = 0). h (268) where x1 and x2 are the classical turning points deﬁned by En = V (x1 ) = V (x2 ). the quantization condition (268) takes the form x2 x1 h ¯ (1) 2m[En − W 2 (x) + √ W ′ (x)] dx = (n + 1/2)¯ π. 6. The purpose of this section is to describe and give applications of the supersymmetric WKB method (henceforth called SWKB) [79.6 Supersymmetric WKB Approximation The semiclassical WKB method [151] is one of the most useful approximations for computing the energy eigenvalues of the Schr¨dinger equation. For the potential V1 (x) corresponding to the superpotential W (x). Then. expanding the left hand side in powers h 66 . eigenfunctions and phase shifts of two supersymmetric partner potentials V1 (x) and V2 (x).. We shall now show this in detail.. Combining the ideas of SUSY with the lowest order WKB method. 2. Comtet. for quantum mechanical problems. It o has a wider range of applicability than standard perturbation theory which is restricted to perturbing potentials with small coupling constants. Therefore. In lowest order. 1. n = 0. h 2m (269) Let us assume that the superpotential W (x) is formally O(¯ 0 ). . . the main implication of SUSY is that it relates the energy eigenvalues. 80] which has been inspired by supersymmetric quantum mechanics. the h W ′ term is clearly O(¯ ).
the SWKB quantization condition for the potential V2 (x) turns out to be b a 2m[En − W 2 (x)] dx = (n + 1)¯ π. the SWKB eigenvalues are constrained to be accurate at both ends of allowed values of n at least when the spectrum is purely discrete. (273) and (274). that is − W (a) = W (b) = En . the superpotential W (x) has opposite signs at the two turning points. . 2..of h gives ¯ b a 2m[En − W 2 (x)] dx+ (1) h ¯ 2 b a W ′ (x) dx En − W 2 (x) (1) +. n = 0. (i) For n = 0. h (2) (274) Some remarks are in order at this stage. (iii) Since the lowest order SWKB approximation is not only exact as expected for large n. Hence SWKB is exact by construction for the ground state energy of the Hamiltonian H1 = (−¯ 2 /2m)d2 /dx2 + V1 (x)... 67 . the O(¯ ) term in (271) exactly gives hπ/2. it follows that the lowest order SWKB quantization condition preserves the SUSY level degeneracy i. . the approximate energy eigenvalues computed from the SWKB quantization con(1) (2) ditions for V1 (x) and V2 (x) satisfy the exact degeneracy relation En+1 = En . 1.. (1) 2 E n a b (271) (272) For this case.e. n = 0. so that to leading h ¯ order in h the SWKB quantization condition when SUSY is unbroken is ¯ [79. but is also exact by construction for n = 0.. 1. (270) can be integrated easily to yield h In the case of unbroken SUSY. One can thus reasonably expect better results than the WKB scheme. 80] b a h 2m[En − W 2 (x)] dx = n¯ π. = (n+1/2)¯ π. the turning points a and b in eq.. (1) (273) Proceeding in the same way. (273) are coincident and (1) E0 = 0. (1) h −1 W (x) ¯ sin . unlike the ordinary WKB approach. h (ii) On comparing eqs. (270) h (1) where a and b are the turning points deﬁned by En = W 2 (a) = W 2 (b). The O(¯ ) term in eq. hence. 2.
(273)]? To study this question. h (275) Now. even with such corrections. Of course. the obvious ﬁrst attempts consisted of obtaining the SWKB bound state spectra of several analytically solvable potentials like Coulomb.2 Exactness of the SWKB Condition for Shape Invariant Potentials How good is the SWKB quantization condition [eq. However. it would be nice to check that higher order corrections vanish order by order in h. hence (s) E0 s = k=1 R(ak ) (276) as given by eq. Besides. Morse. [79. we ﬁnd that for all SIPs. What about higher order SWKB contributions? Since the lowest order SWKB energies are exact for shape invariant potentials. x)] dx = n¯ π . In fact. On using this relation n times .6. (275). 80]. such modiﬁcations are unmotivated and have diﬀerent forms for diﬀerent potentials. must be exact for Hamiltonian H (s) . harmonic oscillator. It has been explicitly checked for all known 68 . 3. For the hierarchy of Hamiltonians H (s) discussed in Sec. one can readily develop the higher order SWKB formalism [83]. one can artiﬁcially restore exactness by ad hoc Langerlike corrections [152]. One can now go back in sequential manner from H (s) to H (s−1) to H (2) and H (1) and use the (1) (2) fact that the SWKB method preserves the level degeneracy En+1 = En . the SWKB quantization condition takes the form 2m[En − (s) s k=1 R(ak ) − W 2 (as . it was soon shown that the lowest order SWKB condition gives the exact eigenvalues for all SIPs [45]! The proof of this statement follows from the facts that the SWKB condition preserves the level degeneracy and a vanishing ground state energy eigenvalue. since the SWKB quantization condition is exact for the ground state energy when SUSY is unbroken. 153]. the higher order WKB contributions are nonzero for most of these potentials[152. This is a very substantial improvement over the usual WKB formula eq. the lowest order SWKB condition gives the exact energy eigenvalues [45]. By starting ¯ from the higher order WKB formalism. (268) which is not exact for most SIPs. etc.
one can continue the SWKB results obtained for A (and hence W ) independent of h over to this superpotential and so ¯ obtain an SWKB approximation for a hindependent potential. One might object to this procedure h h ¯ since the resulting potential V1 is then hdependent. For example.e. It has been proved above that the lowest order SWKB approximation reproduces the exact bound state spectrum of any SIP. 4. (273) is exact for SIPs ? To understand this. 135] of SIPs discussed in Sec. 59. for which the parameters a2 and a1 are related by scaling a2 = qa1 . in all cases. Such a move may appear to be of limited value since one cannot apply ¯ SWKB directly to a superpotential W which is now hdependent.. Conditions on the superpotential which ¯ ensure exactness of the lowest order SWKB condition have been given in ref. R1 a1 ) = R1 a1 F ( R1 a1 x/¯ ). 85]. 60. What is the special feature of these new potentials that interferes with the proof that eq. ¯ because A is a free parameter.2. In the derivation. However. This ¯ assumption is justiﬁed for all the standard SIPs [14. ¯ this hdependence can be absorbed into some dimensionful parameters in the ¯ problem. 54. This statement has indeed been explicitly checked for all SIPs known until last year i. the only free parameter in the problem (apart from q) is the combination R1 a1 . it has recently been shown [154] that the above statement is not true for the newly discovered class [58. [85]. ¯ (277) Taking A such that A(A + h) is independent of h gives the desired h¯ ¯ ¯ independent potential (the additive constant is irrelevant and so can contain h). solutions of the shape invariance condition involving a translation of parameters a2 = a1 + constant.SIPs that up to O(¯ 6 ) there are indeed no corrections. √ √ on which W depends as W (x. This result can be h extended to all orders in h [84.2. ¯ but F is a suﬃciently complicated function that there is no way of eliminating h from W 2 . V1 and En . Incorporating h diﬀerent dependences on h in R1 a1 will give diﬀerent ones in W . W 2 is taken as O(¯ 0 ) while hW ′ is O(¯ ) and ¯ h h hence one can expand the integrand on the left hand side in powers of h. This is a direct consequence of the scaling reparameterisation ¯ 69 . ¯ What about the new potentials ? In the simplest of these cases. consider W (x) = A tanh x . V1 (x) = A2 − A(A + h)sech2 x. let us look again at the derivation of the lowest order SWKB quantization condition. However. However. 55] since W 2 is indeed of O(¯ 0 ) while hW ′ is indeed of O(¯ ).
The basic distinction between them involved in deriving eq. but only in those cases for which the formula is applicable in the ﬁrst place. since one cannot directly make use of the SWKB quantization condition eq. One choice which readily springs to mind is the Ginocchio potential given by [57] (1 − λ2 ) [2 − (7 − λ2 )y 2 + 5(1 − λ2 )y 4] 4 (278) where y is related to the independent variable x by V (x) = (1 − y 2) −λ2 ν(ν + 1) + dy = (1 − y 2)[1 − (1 − λ2 )y 2] . 55] are the only known ones for which the lowest order SWKB formula is exact and the higher order corrections are all zero. In other words. a1 ) = W 2 (x. it remains true that the lowest order SWKB quantization condition is exact for SIPs (if the SUSY is unbroken). ¯ (273) is thus no longer valid and we are prevented from deriving the SWKB condition for these new potentials. so would ¯ ¯ 2 W (x. 6. a1 ) were independent of h. a2 ). An obvious interesting question is to compare WKB and SWKB for potentials which are not shape invariant but for whom the ground state wave function is known. For potentials for which the ground state wave function (and hence the superpotential W ) is not known. W 2 and hW ′ are now of a similar ¯ order in h. a2 ) be and in taking the lowest order of the shape invariance condition one would get W 2 (x. It is thus still the case that the SIPs given in refs.a2 = qa1 not involving h. Furthermore. we have already seen that for shape invariant potentials. (273). which corresponds to the harmonic oscillator. clearly the WKB approach is preferable.3 Comparison of the SWKB and WKB Approaches Let us now compare the merits of the two schemes [WKB and SWKB]. We thus see that the SWKB quantization condition is not the correct lowest order formula in the case of the new SIPs and hence it is not really surprising that eq. [54. (273) does not give the exact eigenvalues for these potentials. dx (279) 70 . SWKB is clearly superior. On the other hand. If W 2 (x. with a2 = qa1 .
. In general. let us note that several authors have also obtained the wave functions in the SWKB approximation [79. These divergences can be regularized either by the uniform approximation [91. 89] .. which is shape invariant. As in the WKB method. but there is no clear cut indication that SWKB results are always better. The corresponding superpotential is [52] W (x) = (1 − λ2 )y(y 2 − 1)/2 + µ0 λ2 y where µn is given by [57] µ n λ2 = [λ2 (ν + 1/2)2 + (1 − λ2 )(n + 1/2)2 ] − (n + 1/2) En = −µ2 λ4 . The spectra of the Ginocchio potential using both the WKB and SWKB quantization conditions have been computed [90]. neither semiclassical method gives the exact energy spectrum.Here the parameters ν and λ measure the depth and shape of the potential respectively. one has the RosenMorse potential. 92. The only exception is the shape invariant limit λ = 1. 6. support the conjecture that shape invariance is perhaps a necessary condition for the lowest order SWKB approximation to yield the exact spectrum [90]. This example. (268) to (273)]. The results are shown in Table 6. Here. as well as other potentials studied in ref. one has W (a) = W (b) = 71 En (1) (283) . . 88. 1 the SWKB values are consistently better. [86. 155] or by appropriate retention of higher orders in h [92.1 [eqs. ¯ 6. For the case of h broken SUSY. For completeness. So far. 1. as expected. Also. (271). we have concentrated our attention on the energy eigenvalues. for n = 0. in which case the SWKB results are exact.4 SWKB Quantization Condition for Broken Supersymmetry The derivation of the lowest order SWKB quantization condition for the case of unbroken SUSY is given in Sec. 87. the SWKB wave functions diverge at the turning points a and b. 90. n = 0. 91. 93]. for broken SUSY. the same derivation applies until one examines the O(¯ ) term in eq.1. 2. n (281) (280) and the bound state energies are (282) For the special case λ = 1. 93].
The generalization to obtain an nparameter family is described in Sec. When applied to a reﬂectionless potential (Sec.e. we will describe how one can start from any given onedimensional potential V1 (x) with n bound states. 7.and the O(¯ ) term in (271) exactly vanishes. 72. As for the case of unbroken SUSY. .. 7. 7. . h (1) (284) As before. n = 0.. 69.1 One Parameter Family of Isospectral Potentials In this subsection. but the GelfandLevitan approach to ﬁnding them is technically much more complicated than the supersymmetry approach described here. 162]. 75].2 [72]. . to leading order in h the h ¯ SWKB quantization condition for broken SUSY is [94. 157. it is easy to obtain the quantization condition which includes higher orders in h [96] and to test how well the lowest order broken SWKB ¯ condition works for various speciﬁc examples. the results using eq. 156. the nparameter families provide surprisingly simple expressions for the pure multisoliton solutions [75] of the Kortewegde Vries (KdV) and other nonlinear evolution equations [159. In Sec. x) i. So. Further discussion can be found in ref. . 160. The fact that such families exist has been known for a long time from the inverse scattering approach [67].3). [96]. λn . . 1. we describe two approaches of obtaining the oneparameter family V1 (λ1 . 138. 2. Indeed..1 we describe how a one parameter isospectral family is obtained by ﬁrst deleting and then reinserting the ground state of V1 (x) using the Darboux procedure [64. 68. 158] [66. it is found that exact spectra are obtained for shape invariant potentials with broken SUSY [95]. For potentials which are not analytically solvable. x) of potentials isospectral to a given potential V1 (x). 7 Isospectral Hamiltonians In this section. 70]. 96] b a 2m[En − W 2 (x)] dx = (n + 1/2)¯ π. 72 . the advent of SUSY QM has produced a revival of interest in the determination of isospectral potentials [65. reﬂection and transmission coeﬃcients identical to those for V1 (x). 70. and use supersymmetric quantum mechanics to construct an nparameter family of strictly isospectral potentials V1 (λ1 . λ2 . (284) are usually better than standard WKB computations. 161. potentials with eigenvalues. 7.
what are the various possible superpotentials W (x) and corresponding 2 ′ ˆ ˆ ˆ potentials V1 (x) = W − W ? Let us assume [68] that there exists a more general superpotential which satisﬁes ˆ ˆ V2 (x) = W 2 (x) + W ′ (x).e. In the second approach. W (x) = W (x) + dx so that the one parameter family of potentials (290) d2 ˆ ˆ ˆ V1 (x) = W 2 (x) − W ′ (x) = V1 (x) − 2 2 ln[I1 (x) + λ1 ] (291) dx have the same SUSY partner V2 (x). let us set ˆ W (x) = W (x) + φ(x) (286) in eq. To ﬁnd the most general solution. (285). y(x) dx Here I1 (x) ≡ x −∞ 2 ψ1 (x′ )dx′ .The ﬁrst approach follows from asking the following question: Suppose V2 (x) is the SUSY partner potential of the original potential V1 (x). Then. (285). (287) (288) (289) λ1 is a constant of integration and ψ1 (x) is the normalized ground state wave ˆ function of V1 (x) = W 2 (x) − W ′ (x). (285) ˆ Clearly. This generates the SUSY partner potential V2 (x) = 73 . W = W is one of the solutions to eq. and let W (x) be the superpotential such that V2 (x) = W 2 +W ′ and V1 (x) = W 2 −W ′ . We ﬁnd then that y(x) = φ−1 (x) satisﬁes the Bernoulli equation y ′ (x) = 1 + 2W y whose solution is d 1 = φ(x) = ln[I1 (x) + λ1 ]. for a given ˆ V2 (x).. Thus the most general W (x) satisfying eq. (285) is given by d ˆ ln[I1 (x) + λ1 ]. given V2 (x). is the original potential V1 (x) unique i. we delete the ground state ψ1 at energy E1 for the potential V1 (x).
In this case W (x) = 74 ω x 2 (296) . which is almost isospectral to V1 (x) i.e. d2 ˆ V1 (λ1 ) = V1 − 2 2 ln(ψ1 Φ1 (λ1 )) dx d2 = V1 − 2 2 ln(I1 + λ1 ) .d V1 − 2 dx2 ln ψ1 . it has the same eigenvalues as V1 (x) except for the bound state at energy E1 . (292) Now. ˆ we ﬁnd a oneparameter family of potentials V1 (λ1 ) isospectral to V1 for λ1 > 0 or λ1 < −1. d2 ˆ (293) V1 (λ1 ) = V2 − 2 2 ln Φ1 (λ1 ) . This corresponds to the choices λ1 → ±∞. provided that λ1 does not lie in the interval −1 ≤ λ1 ≤ 0. we can again use the standard SUSY (Darboux) procedure to add a state at E1 by using the general solution Φ1 (λ1 ). starting with a potential V2 . To elucidate this discussion. Therefore. it may be worthwhile to explicitly construct the oneparameter family of strictly isospectral potentials corresponding to the one dimensional harmonic oscillator [70]. dx The function 1/Φ1 (λ1 ) is the normalizable ground state wave function of ˆ V1 (λ1 ). the function 1/ψ1 satiﬁes the Schr¨dinger equation with potential V2 and energy E1 . Although the potential V2 does not have an eigenenergy E1 .. x) = 1/Φ1 (λ1 ) . The o x 2 other linearly independent solution is −∞ ψ1 (x′ )dx′ /ψ1 . (295) (294) Note that this family contains the original potential V1 . Therefore. dx The corresponding ground state wave functions are ˆ ψ1 (λ1 . the most general solution of the Schr¨dinger equation for the potential V2 at energy o E1 is 2 Φ1 (λ1 ) = (I1 + λ1 )/ψ1 . The next step is to reinstate a bound state at energy E1 .
7. Since one parameter is generated every time an eigenstate is reinstated.so that ω2 2 ω x − . 7.3. The remaining potential is called the Pursey potential VP (x) [66]. In Figs. . . . . ψ4 . erfc(x) = √ π ∞ x e−t dt . Recall that deleting the eigenenergy E1 gave the potential V2 (x). . which corresponds to n = 2). (294). (294) and (295). The ground state ψ2 for the potential V2 is located at energy E2 . We get 1 I1 (x) = 1 − erfc 2 √ ω x 2 2 . 2 (299) Using eqs. with ground states ψ3 . one obtains the one parameter family of isospectral potentials and the corresponding wave functions. 7. V4 . . An analogous situation occurs in the limit λ1 = −1. We see that as λ1 decreases from ∞ to 0. . producing potentials V3 .1 and 7. E4 . the ﬁnal result is a nparameter isospectral family [72]. until the top potential Vn+1 (x) holds no bound state (see Fig. (289) it is now straightforward to compute the corresponding I1 (x). 75 . The procedure can be repeated “upward”. . the remaining potential being the AbrahamMoses potential [65].2. The general formula for VP (x) is obtained by putting λ1 = 0 in eq. at energies E3 . Note that as λ1 ﬁnally becomes zero this attractive potential well is lost and we lose a bound state. . 4 2 The normalized ground state eigenfunction of V1 (x) is V1 (x) = ψ1 (x) = (297) ω 1/4 exp(−ωx2 /4) (298) 2π Using eq.2 Generalization to nParameter Isospectral Family The second approach discussed in the previous subsection can be generalized by ﬁrst deleting all n bound states of the original potential V1 (x) and then reinstating them one at a time. we have plotted some of the potentials and wave functions for the ˆ case ω = 2. . V1 starts developing a minimum which shifts towards x = −∞.
7. dx (302) ˆ The solutions of the Schr¨dinger equation for potentials V3 and V2 (λ2 ) are o related by a new SUSY operator d ˆ2 A† (λ2 ) = − + (ln Φ2 (λ2 ))′ . (300) Here the SUSY operator Ai relates solutions of the potentials Vi and Vi+1 . A useful alternative expression is 76 (305) . we ﬁnd an isospectral oneparameter family V2 (λ2 ). as before. λ2 )) dx d2 = V1 − 2 2 ln(ψ1 (I2 + λ2 )Φ1 (λ1 .In order to produce a twoparameter family of isospectral potentials. d2 ˆ V1 (λ1 . λ2 ) at E1 for V2 (λ2 ) is ˆ2 Φ1 (λ1 . d2 ˆ V2 (λ2 ) = V2 − 2 2 ln(I2 + λ2 ) . λ2 ). dx (301) ˆ Then. (303) (304) The normalizable function 1/Φ1 (λ1 . which results in a twoparameter family of isospectral systems ˆ V1 (λ1 . we go from V1 to V2 to V3 by successively deleting the two lowest states of V1 and then we readd the two states at E2 and E1 by SUSY transformations. The quantities Ii are deﬁned by Ii (x) ≡ x −∞ 2 ψi (x′ )dx′ .3). dx ˆ Therefore. λ2 ) = A† (λ2 )A2 Φ1 (λ1 ) . dx for λi > 0 or λi < −1. Ai = d − (ln ψi )′ . λ2 ) = V1 − 2 2 ln(ψ1 ψ2 Φ2 (λ2 )Φ1 (λ1 . the solution Φ1 (λ1 . The most general solutions of the Schr¨dinger equation for the potential V3 o are given by Φ2 (λ2 ) = (I2 + λ2 )/ψ2 at energy E2 . λ2 )) . and A2 Φ1 (λ1 ) at energy E1 (see Fig. λ2 ) is the ground state at E1 of a new potential.
7. · · · . We start with the simplest case n = 1.3 nSoliton Solutions of the KdV Equation As an application of isospectral potential families. dx (307) (308) (309) d ˆi A† (λi . n . (306) The above procedure is best illustrated by the pyramid structure in Fig. · · · . λn ) · · · A† (λn ) n × An An−1 · · · Ai+1 Φi (λi ) . and we may form a nparameter family of isospectral potentials. Ai = d − (ln ψi )′ . (312) where n is an integer. λi+1 . V1 holds n bound states. · · · . λi+3 . (310) d2 ˆ V1 (λ1 . λ2 )/Φ1 (λ1 . we consider reﬂectionless potentials of the form V1 = −n(n + 1)sech2 x .3. λn ) = − + (ln Φi (λi . λ2 ))2 + 2E1 . · · · . λn ) = V1 − 2 2 ln(ψ1 ψ2 · · · ψn Φn (λn ) · · · Φ1 (λ1 . It can be generalized to an nparameter family of isospectral potentials for an initial system with n bound states. λn )) . λn )A† (λi+2 . · · · . dx Φi (λi . · · · . We 77 . The formulas for an nparameter family are Φi (λi ) = (Ii + λi )/ψi . 7. i = 1. λ2 ) = −V2 (λ2 ) + 2(Φ′1 (λ1 . λn ))′ . since these potentials are of special physical interest. (311) dx The above equations [72] summarize the main results of this section. · · · .ˆ ˆ V1 (λ1 . · · · . λi+2 . λn ) ˆi+1 ˆi+2 ˆ = A† (λi+1 .
i = 1. 138. The corresponding 1parameter family is 1 1 ˆ V1 (λ1 ) = −2sech2 (x + ln(1 + )) . E1 = −1 and ψ1 = 2− 2 sechx. · · ·) where the operator Lu is deﬁned by − ut = (Lu )q ux ∞ x 1 (314) Lu f (x) = fxx − 4uf + 2ux dyf (y) (315) and u is chosen to vanish at inﬁnity. a well with one bound state at E1 will move in the +x direction leaving behind a shallow well with one bound state at E2 . varying the parameter λ1 corresponds to translations of V1 (x). 1. reﬂectionless) 78 .. The SUSY partner potential √ V2 = −2sech2 x. As λ1 approaches the limits 0+ (Pursey limit) and −1− (AbrahamMoses limit). For the case n = 2. · · · . It is well known that oneparameter (t) families of isospectral potentials can also be obtained as solutions of a certain class of nonlinear evolution equations [159. These equations are also known to possess pure (i. [For q = 0 we simply get −ut = ux . V1 (λ1 .e. 2 . 160. These equations have the form (q = 0. 162]. Thus. 2 λi As we let λ1 → −1. Also. V1 = −6sech2 x and there are two bound states at E1 = −4 and E2 = −1. I1 = 1 (3 tanh x − tanh3 x + 2) and I2 = 1 (tanh x + 1).have V1 = −2sech2 x. The movement of the shallow well is essentially controlled by the parameter λ2 . λn . It is tedious but straightforward to obtain the result for arbitrary n and ˆ get V1 (λ1 . we have the freedom to move either one of the wells. (313) 2 λ1 Clearly. After 4 2 2 some algebraic work. x). λ2 ) = − 12 [cosh(3x − δ2 − δ1 ) + 3 cosh(x + δ2 − δ1 )]2 1 1 δi ≡ − ln(1 + ) . 2. is The ground state wave functions of V1 and V2 are ψ1 = 23 sech2 x and ψ2 = 1 √ sechx. we obtain the 2parameter family [3 + 4 cosh(2x − 2δ2 ) + cosh(4x − 2δ1 )] ˆ . the minimum of the potential moves to −∞ and +∞ respectively. while for q = 1 we obtain the well studied Kortewegde Vries (KdV) equation]. λ2 .
We will also brieﬂy discuss the superspace formalism for SUSY QM.1. ψ † ]W ′ (x). 7. E3 = 16/25 . The potentials shown in Fig. 2 2 2 we obtain the Lagrangian 1 1 1 L = x2 + iψ † ∂t ψ − W 2 (x) + [ψ.4 are all isospectral and reﬂectionless holding bound states at E1 = 25/16. 8 Path Integrals and Supersymmetry In this section. ψ † ]W ′ (x). These are the AbrahamMoses procedure [65] and the Pursey procedure [66]. namely the Langevin equation. we have found nparameter isospectral families by repeatedly using the supersymmetric Darboux procedure for removing and inserting bound states. E2 = −1.a topological quantity which determines whether SUSY is broken. rather than displaying the explicit algebraic expressions here. one for the Witten index . and one for a related “classical” stochastic diﬀerential equation. If these alternative procedures are used. (314) are much simpler than any previously obtained using other procedures. (60)] for convenience: 1 1 1 H = p2 + W 2 (x)I − [ψ.multisoliton solutions. Details can be found in reference [71]. 7. well established procedures for deleting and adding bound states. Starting from the matrix SUSY Hamiltonian which is 1/2 of our previous H [eq. one gets new potential families all having the same bound state energies but diﬀerent reﬂection and transmission coeﬃcients. as brieﬂy mentioned in Sec. we will describe the Lagrangian formulation of SUSY QM and discuss three related path integrals: one for the generating functional of correlation functions. Nevertheless. These expressions for the multisoliton solutions of eq. we shall simply illustrate the 3soliton solution of the KdV equation. we can obtain an explicit analytic formula for the pure nsoliton solution of each of the above evolution equations [75]. note the clear emergence of three independent solitons. However. It is possible to show that by suitably choosing the parameters λi as functions of t in the nparameter SUSY isospectral family of a symmetric reﬂectionless potential holding n bound states. In this section. there are two other closely related. As t increases. ˙ 2 2 2 79 (316) .
ψ ∗ } = 0. We rotate t → iτ and obtain for the Euclidean path integral: Z[j. that det[∂τ − W ′ (x)] = 80 λm . We have. Namely we evaluate the fermionic path integral: [dψ][dψ ∗ ]e T 0 dτ ψ∗ [∂τ −W ′ (x)]ψ (320) by calculating the determinant of the operator [∂τ −W ′ (x)] using eigenvectors which are antiperiodic. (317) 1 2 1 2 x + W (x) − ψ ∗ [∂τ − W ′ (x)]ψ 2 τ 2 0 and ψ and ψ ∗ are now elements of a Grassman algebra: SE = dτ {ψ ∗ . δψ ∗ = −ǫ∗ (∂τ x + W (x)) . These transformations correspond to N = 2 supersymmetry. following Gildener and Patrascioiu [164]. The path integral over the fermions can now be explicitly performed using a cutoﬀ lattice which is periodic in the the coordinate x but antiperiodic in the fermionic degrees of freedom at τ = 0 and τ = T . ψ} = {ψ ∗ . η ∗] = where [dx][dψ][dψ ∗ ] exp[−SE + τ jx + ηψ ∗ + η ∗ ψ].It is most useful to consider the generating functional of correlation functions in Euclidean space. ψ} = {ψ. m . 29] which mix bosonic and fermionic degrees of freedom: δx = ǫ∗ ψ + ψ ∗ ǫ. δψ = −ǫ (−∂τ x + W (x)) . The Euclidean action is invariant under the following SUSY transformations [163. η. and xτ = (318) dx dτ . (319) where ǫ and ǫ∗ are two inﬁnitesimal anticommuting parameters.
only the ground state of H1 contributes as T → ∞ We also have: Z± = where ± SE = 0 T ± [dx]exp[−SE ] (325) dτ x2 W 2 (x) W ′ (x) ˙ .where so that [∂τ − W ′ (x)]ψm = λm ψm . (321) Imposing the antiperiodic boundary conditions: ψm (T ) = −ψm (0) yields: 1 T i(2m + 1)π − dτ W ′ (x). (324) For the case when SUSY is unbroken. as expected that Z is the sum of the partition functions for the two pieces of the supersymmetric Hamiltonian. + ± 2 2 2 A related path integral is obtained for the noise averaged correlation functions coming from a classical stochastic equation. 2 (323) Rewriting the cosh as a sum of two exponentials we ﬁnd. ψm (τ ) = Cm exp[ τ 0 dτ ′ [λm + W ′ ]. (322) T T 0 Regulating the determinant by dividing by the determinant for the case where the potential is zero we obtain: λm = det[ ∂τ − W ′ (x) ] = cosh ∂τ T 0 dτ W ′ (x) . If we have the stochastic diﬀerential equation x = W (x(τ )) + η(τ ) ˙ (326) where η(τ ) is a random stirring forcing obeying Gaussian statistics. Namely when the external sources are zero: Tre−H1 T + Tre−H2 T ≡ Z− + Z+ . the Langevin equation. then the correlation functions of x are exactly the same as the correlation functions 81 .
.. The correlation functions averaged over the noise are: < x(τ1 )x(τ2 ). One has Det dη  = exp dx dt Tr ln [ d − W ′ (x(τ ))]δ(τ − τ ′ ) dτ. 2.. DηP [η]f (τ ) = 0. Det dη(τ )  dx(τ ′ ) (330) subject to the boundary condition that the Green’s function obey causality... dx(τ ′ ) (329) This involves calculating the functional determinant. >= DηP [η]x(τ1 )x(τ2 ). so one has retarded boundary conditions. 82 (332) . (328) where we have in mind ﬁrst solving the Langevin equation explicitly for x(η(τ )) and then averaging over the noise as discussed in Sec.. dτ (331) When there are no interactions the retarded boundary conditions yield G0 (τ − τ ′ ) = θ(τ − τ ′ ).. DηP [η]f (τ )f (τ ′) = F0 δ(τ − τ ′ ). < x(τ1 )x(τ2 ).obtained from the Euclidean quantum mechanics related to the Hamiltonian H1 .2. Another way to calculate the correlation function is to change variables in the functional integral from η to x. >= D[x] P [η] Det  dη(τ )  x(τ1 )x(τ2 ). To see this we realize that Gaussian noise is described by a probability functional: 1 T η 2 (τ ) ] (327) dτ P [η] = Nexp[− 2 0 F0 normalized so that: DηP [η] = 1..
(333) (334) we ﬁnd that the generating functional for the correlation functions is exactly the generating functional for the correlation functions for Euclidean quantum mechanics corresponding to the Hamiltonian H1 : Z[j] = N D[x] exp[− 1 2 T 0 dτ x2 + W 2 (x) − W ′ (x) − 2j(τ )x(τ ) ] (335) ˙ Thus we see that we can determine the correlation functions of x for the Hamiltonian H1 by either evaluating the path integral or solving the Langevin equation and averaging over Gaussian noise. Deﬁning the noise average: Pc (z) =< δ(z − x(t) >η = one obviously has: dzz n Pc (z. ˙ 2¯ 0 h 1 dη(τ )  = exp[ ′) dx(τ 2 T 0 dτ W ′ (x)]. An equation related to the Langevin equation is the FokkerPlanck equation.Expanding ln(1 − G0 W ′ ) one ﬁnds because of the retarded boundary conditions that only the ﬁrst term in the expansion contributes so that Det Choosing F0 = h so that ¯ P [η] = N exp[− 1 T η 2 (τ ) dτ ] 2 0 F0 1 T = N exp[− dτ (x − W (x))2 ] ˙ 2¯ 0 h 1 T = N exp[− dτ (x2 + W 2 (x))]. t) = Dη[x(t)]n P [η] =< xn > . Dηδ(z − x(t))P [η] (336) One can show [165] that Pc obeys the FokkerPlanck equation: ∂P 1 ∂2P ∂ = F0 2 + [W (z)P (z. t))] ∂t 2 ∂z ∂z 83 (337) . which deﬁnes the classical probability function Pc for the equal time correlation functions of H1 .
As we discussed before.Ψ∗ )dτ . t) → P (z) and Setting ∂P ∂t ˆ P (z)dz = 1. one can introduce a “fermion” number operator via nF = Thus (−1)F = [ψ. we obtain ˆ P (z) = Ne−2 z 0 W (y)dy = ψ0 (z)2 . ψ † ] 1 − σ3 = . A third path integral for SUSY QM is related to the Witten index. Here we will show that the Witten index can also be obtained using the path integral representation of the generating functional of SUSY QM where the fermion determinant is now evaluated using periodic boundary conditions to incorporate the factor (−1)F . (342) 84 . (338) Thus at long times only the ground state wave function contributes (we are in Euclidean space) and the probability function is just the usual ground state wave function squared.2 we showed how to determine △(β) using heat kernel methods and how it was useful in discussing nonperturbative breaking of SUSY. the Witten index needs to be regulated and the regulated index is deﬁned as: △(β) = Tr(−1)F e−βH = Tr(e−βH1 − e−βH2 ) . = 0 in the FokkerPlanck equation. one cannot deﬁne an equilibrium distribution for the classical stochastic system. As we discussed earlier. (341) 1 − [ψ. (340) The Witten index is given by △ = Tr(−1)F .Ψ. Consider the path integral: △(β) = [dx][dΨ][dΨ∗ ]e β 0 LE (x. 2 2 (339) In Sec 2. We see from this that when SUSY is broken. ψ † ] = σ3 . It is easy to verify a posteriori that this is the case.For an equilibrium distribution to exist at long times t one requires that ˆ P (z.
one changes the boundary conditions for evaluating the fermion determinant at τ = 0. In the degenerate case d = 1 the ﬁeld 85 . superﬁelds are deﬁned on the space (xn . as expected that we obtain the regulated Witten index: △(β) = Z− − Z+ = Tre−βH1 − Tre−βH2 . Raby [166]). τ 2 2 F To incorporate the (−1) in the trace. 2 (344) Again rewriting the sinh as a sum of two exponentials we ﬁnd. We again have det[∂τ − W ′ (x)] = λm .1 Superspace Formulation of Supersymmetric Quantum Mechanics One can think of SUSY QM as a degenerate case of supersymmetric ﬁeld theory in d = 1 in the superspace formalism of Salam and Strathdee [167] (This idea is originally found in unpublished lecture notes of S. m Imposing periodic boundary conditions: λm = i(2m)π 1 − β β β 0 dτ W ′ (x). We now impose these boundary conditions on the determinant of the operator [∂τ − W ′ (x)] using eigenvectors which are periodic. (345) 8.1 1 LE = x2 + W 2 − Ψ∗ [∂τ − W ′ (x)]Ψ. θa ) where x is the space coordinate and θa are anticommuting spinors. (343) Regulating the determinant by dividing by the determinant for the case where the potential is zero we obtain: det ∂τ − W ′ (x) = sinh ∂τ β 0 dτ W′ (x) . β to periodic ones: x(0) = x(β) Ψ(0) = Ψ(β) where The path integral over the fermions can again be explicitly performed using a cutoﬀ lattice which is periodic in the fermionic degrees of freedom at τ = 0 and τ = β.
Q∗ } = 2i∂t = 2H. Now these charges obey the familiar SUSY QM algebra: {Q. θ. Consider the following SUSY transformation: t′ = t − i(θ∗ ǫ − ǫ∗ θ). the following derivatives are invariant: Dθ = ∂θ − iθ∗ ∂t or in component form: ˙ Dθ φ = iψ − θ∗ D − iθ∗ x + θ∗ θψ. L = ei(ǫ then from δA = i[ǫ∗ Q∗ + Qǫ. θ} = [θ.is replaced by x(t) so that the only coordinate is time. θ∗ ) = x(t) + iθψ(t) − iψ ∗ θ∗ + θθ∗ D(t). θ∗ } = {θ. A superﬁeld made up of x and θ and θ∗ can at most be a bilinear in the Grassman variables: φ(x. A] we infer that the operators Q and Q∗ are given by: Q = i∂θ − θ∗ ∂t . The Lagrangian in superspace is determined as follows. θ′ = θ + ǫ. [Q. (349) (348) (347) ∗ Q∗ +Qǫ) (346) If we assume that ﬁnite SUSY transformations can be parametrized by . Under a SUSY transformation. θ∗′ = θ∗ + ǫ∗ . Q∗ = −i∂θ∗ − θ∂t . H] = 0. The anticommuting variables are θ and θ∗ where {θ. ˙ 86 . t] = 0. ˙ and or in component form: Dθ∗ = ∂θ∗ − iθ∂t (352) (351) (350) ˙ [Dθ φ]∗ = −iψ ∗ − θD + iθx + θ∗ θψ ∗ .
A more complete discussion of this can be found in ref. Finally. 87 .2. The result is a rapidly converging series which is substantially better than the usual WKB tunneling formula. Sec. 9.3 contains a description of a SUSY QM calculation of the energy splitting and rate of tunneling in a double well potential. Integrating over the Grassman degrees of freedom using the usual path integral rules for Grassman variables: S= dtdθ∗ dθ θdθ = one obtains 1 1 dt( x2 + ψ ∗ [∂t − f ′′ (x)]ψ + D 2 + Df ′ (x)). 9.The most general invariant action is: 1 Dθ φ2 − f (φ) (353) 2 Again the expansion in terms of the Grassman variables causes a Taylor expansion of f to truncate at the second derivative level. ˙ (354) 2 2 Eliminating the constraint variable D = −f ′ (x) = W (x) we obtain our previous result for the action (now in Minkowski space): S= S= dt 1 1 2 x + ψ ∗ [∂t − W ′ (x)]ψ − W 2 ˙ 2 2 (355) θ∗ dθ∗ = 1. we describe how the large N expansion (N = number of spatial dimensions) used in quantum mechanics can be further improved by incorporating SUSY. 9 Perturbative Methods for Calculating Energy Spectra and Wave Functions The framework of supersymmetric quantum mechanics has been very useful in generating several new perturbative methods for calculating the energy spectra and wave functions for one dimensional potentials. in Sec. 9. [11]. Four such methods are described in this section. dθ = dθ∗ = 0.1 and 9. we discuss two approximation methods (the variational method and the δ− expansion) for determining the wave functions and energ y eigenvalues of the anharmonic oscillator making use of SUSY QM.4. In Secs.
Minimizing the expectation value of the Hamiltonia n with respect to the parameter β yields 3 E0 = ( )4/3 g 1/3 . Assuming a trial wave function of the form ψ0 = ( we obtain < H >=< 2β 1/4 −βx2 ) e . For the original potential. 3. 4 3 β = ( )1/3 g 1/3 . 2 2 16β 2 (358) (In this subsection.1 Variational Approach The anharmonic oscillator potential V (x) = gx4 is not exactly solvable. one can then determine the partner potential and its ground state wave function approximately. To determine the superpotential one has to ﬁrst subtract the ground state energy E0 and solve the Riccati equation for W(x): V1 (x) = gx4 − E0 ≡ W 2 − W ′ . π (357) p2 β 3g + gx4 >= + . Using the hierarchy of Hamiltonians discussed in Sec. we can determine the optimal Gaussian wave function quite easily. we are taking m = 1 in order to make contact with published numerical results).668g 1/3 88 . First let us see how this works using a simple variational approach. (356) Once the ground state energy and the superpotential is known to some order of accuracy.9. Then. one can construct from the approximate ground state wave functions of the hierarchy and the approximate superpotentials Wn all the excited states of the anharmonic oscillator approximately. using the SUSY operator d − W (x) dx one can construct the ﬁrst excited state of the anharmonic oscillator in the usual manner. 4 This is rather good for this crude approximation since the exact ground state energy of the anharmonic oscillator determined numerically is E0 = .
681 . [168]. dx (359) which leads to a Gaussian approximation to the potential V1G = 4βx2 − 2β. 2 (362) Thus we have approximately that the energy diﬀerence between the ground state and ﬁrst excited state of the anharmonic oscillator is 3 E1 − E0 = 4β = 4( )1/3 g 1/3 . Let us now look at a more general class of trial wave functions. (357). The approximate ground state energy of the second potential is now < H2 >=v < 0 p2 + V2G 0 >v = H1 + 4β. 4 The approximate (unnormalized) ﬁrst excited state wave function is ψ1 = [ (1) d 2 − 2βx]ψ0v ∝ −4βxe−βx . If we choose for the trial ground state wave functions of the hierarchy of Hamiltonians. dx (363) This method can be used to ﬁnd all the excited state wave functions and energy levels of the anharmonic oscillator by using the methods discussed in ref.whereas ( 3 )4/3 = . 1 ψ0 = Nk exp − 2 (k) x2 ρk nk . √ Nk = 2 ρk Γ(1 + 1 ) 2nk −1/2 . (364) 89 . The approximate potential W resulting from this 4 variation calculation is W (x) = − dlogψ0 = 2βx. (361) (360) Since V2G diﬀers from V1G by a constant. the approximate ground state wave function for V2 is given by eq. The approximate supersymmetric partner potential is now V2G = 4βx2 + 2β.
It is convenient in this case to ﬁrst scale the Hamiltonian for the anharmonic oscillator. n1 ) = ψ0  − ∂E0 ∂ρ1 1 d2 + x4 ψ0 . ρ1 = 0.66933.667986 . =0. we consider the variational Hamiltonian 1 ¯ Hvk+1 = Akv A† kv 2 (371) which approximately determines these energy diﬀerences. (364). nk ) = 1 (vk+1) d2 (vk+1) 2 ′ ψ0  − 2 + Wvk + Wvk ψ0 . n1 = 1. 2 dx 90 (372) . Then we ﬁnd the ground state energy of the anharmonic oscillator and the variational parameters ρ1 and n1 by forming the functional E0 (ρ1 . (369) This ground state energy is to be compared with a numerical evaluation which yie lds 0. Since the trial wave function for all ground states is given by eq. (370) Since we are interested in the energy diﬀerences En −En−1 of the anharmonic oscillator. We obtain the approximate energy splittings by minimizing the energy functional δEk (ρk . (365) H=− 2 dx2 by letting x → x/g 1/6 and H → g 1/3 H.we obtain much better agreement for the low lying eigenvalues and eigenfunctions.18346.666721 . (367) The equation for the energy functional for the anharmonic oscillator is 1 5 Γ( 2n1 ) n2 Γ(2 − 2n1 ) 1 2 + ρ1 E0 (ρ1 . n1 ) = 1 1 . 1 d2 + gx4 . 2 dx2 ∂E0 ∂n1 (366) Thus we ﬁrst determine ρ1 and n1 by requiring = 0. the variational superpotential for all k is Wkv = nk x2nk −1 (ρk )−nk . 2ρ1 Γ( 2n1 ) Γ( 2n1 ) (368) Minimizing this expression. we obtain the following variational result: E0 = 0.
δ) − W ′ . δ measures the anharmonicity. (373) One can perform the minimization in ρ analytically leaving one minimization to perform numerically.Performing the integrals one obtains the simple recursion relation: 1 n2 Γ 2 − 2nk n2 k δEk (ρk . Thus we write: V1 (x) = M 2 x2 ∞ δ n [ln(Mx2 )]n 2En δ n . (374) where M is a scale parameter. 9. That is we consider simultaneously potentials of the form V1 (x) = M 2+δ x2+2δ − C(δ) ≡ W 2 (x. nk ) = + k−1 2ρk Γ 1 2ρk 2nk ρk ρk−1 Γ 2nk−1 Γ 4nk−1 −1 2nk 1 Γ 2nk ρk nk−1 (2nk−1 − 1) + 2ρk ρk−1 nk−1 2nk−1 −1 2nk 1 Γ 2nk . We assume W (x) = ∞ n=0 δ n W(n) (x).1 for the ﬁrst three energy eigenvalues and compared with a numerical calculation. The results for the variational parameters and for the energy diﬀerences are presented in Table 9. (376) 91 . and C is the ground state energy of the anharmonic oscilator. based on a shooting method. − n! n=0 n=0 ∞ (375) where En corresponds to the order Taylor expansion of the dependence of the ground state energy on the parameter δ. C is subtracted as usual from the potential so it can be factorized. The standard anharmonic oscillator corresponds to δ = 1 and M = (2g)1/3 .2 δ Expansion Method In this method [78] we consider the anharmonic oscillator as an analytic continuation from the harmonic oscillator in the paramater controlling the anharmonicity. To approximately determine W (x) from V1 (x) we assume that both W (x) and V1 (x) have a Taylor series expansion in δ.
6415g 1/3. ψ(x) = Γ′ (x)/Γ(x). We have: 2 ′ W0 − W0 = M 2 x2 − 2E0 . (383) A more accurate determination of the ground state energy can be obtained by calculating up to order δ 2 and then analytically continuing in δ using Pad´ e approximants. we ﬁnd that the ﬁrst two terms in the δ expansion for the ground state energy are 1 1 E = (2g)1/3 [1 + ψ(3/2)δ]. Solving for W1 we obtain W1 (x) = −eM x 2 x 0 dye−M y [M 2 y 2 ln(My 2 ) − 2E1 ]. [169]. 92 (382) . The order δ contribution to the energy eigenvalue E1 is determined by requiring that the ground state wave function be square integrable. (374) and match terms order by order.and insert these expressions in eq. dW1 − 2W1 W0 = −M 2 x2 ln(Mx2 ) + 2E1 dx (378) (379) which is to be solved with the boundary conditionWn (0) = 0. 2 (380) To ﬁrst order in δ the ground state wave function is now : ψ0 (x) = e−M x (1 − δ 2 x 0 dyW1(y)) Imposing the condition that ψ0 vanishes at inﬁnity. we obtain: 1 E1 = Mψ(3/2) . At lowest order in δ the problem reduces to the supersymmetric harmonic oscillator. (381) 4 Writing M = (2g)1/3 . we get E = . (377) whose solution is 1 E0 = M 2 To next order we have the diﬀerential equation: W0 (x) = Mx. This is discussed in ref. 2 2 At δ = 1.
Let χ(x) be the normalized eigenfunction for 93 . Usually. D. n ≤ 1.1. n. and corresponds to the tunneling rate through the doublewell barrier. The number of such pairs. a “deep” potential refers to n ≥ 2. 179. 178.. 174. 9. can be crudely estimated from the standard WKB boundstate formula applied to V1 (x) for x > 0: nπ = xc 0 [V1 (0) − V1 (x)]1/2 dx.9. 172. We deﬁne the depth. The energy splitting t of the lowestlying pair of states can be obtained by a standard argument [126]. An example of such a potential is shown in Fig. The quantity t is often small and diﬃcult to calculate numerically. this expansion has found many applications ranging from condensed matter physics to the computation of chemical reaction rates [170.3 Supersymmetry and Double Well Potentials Supersymmetric quantum mechanics has been proﬁtably used to obtain a novel perturbation expansion for the probabililty of tunneling in a double well potential [34]. First. highly convergent perturbation expansion for the energy diﬀerence t. 171. Since double wells are widely used in many areas of physics and chemistry. Here. we brieﬂy review the standard approach for determining t in the case of a symmetric. we obtain a systematic. 31. (384) where xc is the classical turning point corresponding to energy V1 (0) and we have chosen units where h = 2m = 1. in most applications the quantity of interest is the energy diﬀerence t ≡ E1 − E0 between the lowest two eigenstates. 180] . the doublewell structure produces closely spaced pairs of energy levels lying below V1 (0).[177. 33. onedimensional doublewell potential. 176] . In contrast. 173. of V1 (x) by D ≡ V1 (0) − V1 (x0 ).e. we shall restrict our attention to symmetric double wells. and the magnitude of the nonleading terms gives a reliable handle on the accuracy of the result. 32. For suﬃciently deep wells. we show how SUSY facilitates the evaluation of t. Indeed. using the supersymmetric partner potential V2 (x). especially when the potential barrier between the two wells is large. 175. although an extension to asymmetric double wells is relatively straightforward [181]. whose minima are located x = ±x0 . i. The leading term is more accurate than the standard WKB tunneling formula. In what follows. We shall call a doublewell potential ¯ “shallow” if it can hold at most one pair of bound states. V1 (x).
This result is accurate for “deep” potentials. In this case. only the lowest two states are paired). we know that the supersymmetric partner potential V2 (x) is given by ′ V2 (x) = V1 (x) − 2(d/dx)(ψ0 /ψ0 ) ′ = −V1 (x) + 2(ψ0 /ψ0 )2 . (388) Alternatively. (387) The same result can also be obtained via instanton techniques [182]. Previous papers [31. (389) From the discussion of unbroken SUSY in previous sections. Hence. for the doublewell problem. and its zeroenergy ground state wave function ψ0 (x). H1 = −d2 /dx2 + V1 (x). except for the ground state of V1 which is missing from the spectrum of V2 [10]. 32. but becomes progressively worse as the depth decreases. then the spectrum of V2 is well separated.a particle moving in a single well whose structure is the same as the righthand well of V1 (x) (i. If the probability of barrier penetration is small.. and. (386) where the prime denotes diﬀerentiation with respect to x.e. Using the supersymmetric formulation of quantum mechanics for a given Hamiltonian. (385) By integration of Schr¨dinger’s equation for the above eigenfunctions. x > 0). it can o be shown that [126] t ≡ E1 − E0 = 4χ(0)χ′ (0) . 94 .. we see that if V1 (x) is “shallow” (i. the lowest two eigenfunctions of the doublewell potential V1 (x) are well approximated by √ (1) ψ0. (386) yields the standard result: tWKB = {[2V1′′ (x0 )]1/2 /π} exp −2 x0 0 [V1 (x) − V1 (x0 )]1/2 dx . V2 is relatively structureless and simpler than V1 . Use of WKB wave functions in eq.e.1 (x) = [χ(x) ± χ(−x)]/ 2 .1 (x) = W 2 (x) ± dW/dx . 33] have implicitly treated just the case of shallow potentials. in terms of the superpotential W (x) given by W (x) = ′ −ψ0 /ψ0 we can write V2. we know that the energy spectra of the potentials V2 and V1 are identical.
ψ0 (x) ∼ e−(x−x0 ) + e−(x+x0 ) . 0 95 . 0 (391) (392) The minima of V1 (x) are located near ±x0 and the well depth (in the 2 limit of large x0 ) is D ≃ 4x0 . which involves solving the parabolic cylindrical diﬀerential equation.0 and x0 = 2.not surprisingly. for both V1 (x) and V2 (x). In contrast. 2 2 (390) The variables x and x0 have been chosen to be dimensionless. In order to produce this spectrum. centered around ±x0 . We see that in the limit of large x0 . 9. We illustrate the potentials V1 (x) and V2 (x) in Fig. (386). V2. V2 has a doublewell structure together with a sharp “δ− function like” dip at x = 0. let us now consider the case of a deep double well as shown in Fig. This potential has a well known [183] analytic solution. The corresponding superpotential W (x).2 for the two choices x0 = 1. are given respectively by W (x) = 2[x − x0 tanh(2xx0 )]. one obtains the separation of the lowest two energy levels to be 8x0 (2/π)1/2 exp(−2x2 ). the wells become widely separated and deep and that V2 (x) develops a strong central dip. and becomes sharper as the potential V1 (x) becomes deeper. After carefully handling the boundary conditions. As a concrete example. with χ(x) given by one of the (normalized) Gaussians in eq.1 (x) = 4[x − x0 tanh(2xx0 )]2 ± 2[1 − 2x2 sech2 (2xx0 )] .1 . have found that the use of SUSY simpliﬁes the evaluation of the energy diﬀerence t. in agreement with eq. t. The asymptotic behavior of the energy splitting. we consider the class of potentials whose ground state wave function is the sum of two Gaussians. 9. We ﬁnd that t → 8x0 (2/π)1/2 e−2x0 .5. (390). (393). in the limit x0 → ∞ can be calculated from eq. 2 (393) The same result can be obtained by observing that V1 (x) → 4( x  −x0 )2 as x0 → ∞. the spectrum of V2 has a single unpaired ground state followed by paired excited states. Here. This central dip produces the unpaired ground state. and the two supersymmetric partner potentials V1 (x) and V2 (x).
coupled with the observation that the magnitude of t is in general small. we must solve an approximate problem and calculate the corrections perturbatively. However. The ﬁrst and second order corrections to the o unperturbed energy E = 0 are 96 . thus we expect it to be an excellent approximation of the exact ground state wave function of V2 (x) for all values of x. From supersymmetry [eq. where. (388)] we see immediately that 1/ψ0 is a solution. allows us to construct a suitable unperturbed problem. since V2 (x) is not analytically solvable. The derivative of φ(x) is continuous except at the origin. This procedure is cumbersome and results in perturbation corrections to the leading term which are substantial. For the case of a symmetric potential such as V2 (x). we expect this solution to be an excellent approximation to the correct eigenfunction for small values of x. (395) where we have assumed that ψ0 (x) is normalized. Instead. Hence φ(x) is actually a zeroenergy solution of the Schr¨dinger equation for a potential o V0 (x) given by 2 V0 (x) = V2 (x) − 4ψ0 (0)δ(x) . it has a discontinuity φ 0+ −φ 0− = −2ψ0 (0). Note that the coeﬃcient multiplying the δfunction is quite small so that we expect our perturbation series to converge rapidly. unlike the exact solution. Since t is small. the perturbative corrections to the energy arising from ∆V can be most simply calcualted by use of the logarithmic perturbationtheory [185] formulation of the usual RayleighSchr¨dinger series.We now turn to the evaluation of t via the ground state energy of the supersymmetric partner potential V2 (x). we choose for our unperturbed problem the second linearly independent solution of the Schr¨dinger equation given by [184] o φ(x) = 1 ψ0 ∞ x 2 ψ0 (x′ )dx′ . In general. Clearly. 1/ψ0 is not normalizable and hence is not acceptable as a starting point for perturbation theory. (394) and φ(x) = φ(−x) for x < 0. Consider the Schr¨dinger equation for V2 (x) and o E = 0. One possibility is to regularize the behavior artiﬁcially at large  x  [31]. We calculate the perturba2 tive corrections to the ground state energy using ∆V = +4ψ0 (0)δ(x) as the perturbation. The use of SUSY. φ(x) is well behaved at x = ±∞ and closely approximates 1/ψ0 at small x. x > 0 .
In this subsection. In conclusion. 199] incorporates exactly known analytic results into 1/N expansions. 9. This asymptotic form can also be recovered from eq.E (1) 1 = . By choosing as an unperturbed problem the potential whose solution is the normalizable zeroenergy solution of V2 (x). Even for values of x0 ≤ 1/ 2. The thirdorder perturbative result and the exact result are indistinguishable for all values of x0 .lying two states of V1 (x). we 97 . The procedure is quite general and is applicable to any arbitrary doublewell potential. 198. The asymptotic behavior of t given by eq. (396) For our example.3 for values of x0 ≤ 2. A slightly modiﬁed. 194. the approximation technique is surprisingly good. The results are shown in Fig. physically motivated approach. greatly enhancing their accuracy. we numerically evaluate these corrections in order to obtain an estimate of t. 193. and third order calculated from logarithmic perturbation theory are compared with the exact result for V2 . where N is the number of spatial dimensions. 195. 9. in which case V1 (x) does not exhibit a doublewell structure. one can instead develop a perturbation series for the ground state energy t of the partner potential V2 (x). second. 191. 200]. 192. we obtain a very simple δ−function perturbation which produces a rapidly convergent series for t [34]. (393) is also shown. E 2 ′ (2) = −2 ∞ 0 E (1) ξ(x′ ) φ(x′ ) 2 dx′ . Rather than calculating this splitting as a diﬀerence between the lowest. 188. called the “shifted largeN method” [196. (396) by a suitable approximation of the integrand in the largex0 √ limit. simplicity and range of applicability. 187. The numerical results are very accurate for both deep and shallow potentials. the energy splitting for a doublewell potential. obtained by the RungeKutta method. including asymmetric ones [181]. 197.4 Supersymmetry and LargeN Expansions The largeN method. is a powerful technique for analytically determining the eigenstates of the Schr¨dinger o equation. ξ(x) ≡ 2ξ(0) ∞ x φ (x )dx. 190. we have demonstrated how SUSY can be used to calculate t. even for potentials which have no small coupling constant and hence not amenable to treatment by standard perturbation theory [186. Estimates of t correct to ﬁrst.
the n = 0 wave function ψ0 (r) has the behavior r (k−1)/2 . for example. then eq. Of course. the two choices k and k are equivalent. n. So. assuming o N to be large.. At the end of the calculation. n of Veﬀ (r) one can equally well use k ′ = k + 2. a properly chosen shift a produces great improvements in accuracy and simplicity. for very large values of N. k = N + 2l . n − 1 with V2 (r). one sets N = 3 to get results for problems of physical interest in three dimensions.will descibe how the rate of convergence of shifted 1/N expansions can be still further improved by using the ideas of SUSY QM [77]. For an arbitrary spherically symmetric potential V (r) in N dimensions. To demonstrate this procedure. (399) is given by (k ′ − 1)(k ′ − 3)¯ 2 /8mr 2 . the ¯ 98 . in order to calculate the energy of the state with quantum numbers k. At small values of r. If one sets ψ0 (r) = r (k−1)/2 Φ0 (r) (398) where Φ0 (r) is ﬁnite at the origin. (398) readily gives the supersymmetric partner potential of Veﬀ (r) to be (k + 1)(k − 1)¯ 2 h2 d2 h ¯ V2 (r) = V (r) + − lnΦ0 (r) 2 8mr m dr 2 (399) V2 (r) and Veﬀ (r) have the same energy values [except for the ground state].. and taking 1/N as an “artiﬁcially created” expansion parameter for doing standard perturbation theory. one is working h in two extra spatial dimensions! Thus. This means that the eigenstates. where k ′ = k + 2. l. in fact only depend on k and n. which could in principle have depended on the three quantities N. for N = 3 dimensions. (397) Veﬀ (r) = V (r) + 2 8mr It is important to note that N and l always appear together in the combination k = N + 2l. let us give an explicit example. where k = k − a.1. However. However. One now makes a systematic expansion of eigenstates in the parameter 1/k. the radial Schr¨dinger equation contains the eﬀective potential o (k − 1)(k − 3)¯ 2 h . where n is the radial quantum number which can take values 0..2. Using the usual choice of units h = 2m = 1. eﬀectively. The basic idea in obtaining a 1/N expansion in quantum mechanics consists of solving the Schr¨dinger equation in N spatial dimensions. largeN expansions with the partner potential V2 (r) are considerably better since the angular momentum barrier in eq.
The supersymmetric partner potential is 2δ 2 e−δr H H . the famous Landau level problem) but the nonuniform magnetic ﬁeld problems 99 . 1 − e−δr 4 (400) As r tends to zero. V2H goes like 2r −2 . The results upto leading. the exact answer is known to be 0. The corresponding values using the supersymmetric partner potential V2H are all 0. ψ0 (r) ∼ (1 − e−δr ) e− 2 . [77]. Let us compute the energy of the ﬁrst excited state of H Veﬀ (r). For the choice δ = 0.e.748125 .748125 ! It is clear that although excellent results are obtained with the use of the shifted H 1/N expansion for the original potential Veﬀ (r) in three dimensions. since we are now eﬀectively working in ﬁve dimensions instead of three. Thus. In this section we shall show that the concept of SUSY can also be applied to some geniune twodimensional problems. It is worth emphasizing here that not only the uniform magnetic ﬁeld problem (i. even faster convergence is obtained by using the supersymmetric partner potential. SUSY has played an important role in making a very good expansion even better [77]. In particular we show that the Pauli Hamiltonian which deals with the problem of a charged particle in external magnetic ﬁeld can always be put in SUSY form provided the gyromagnetic ratio is equal to two. In fact. for many applications.747713. Other examples can be found in ref. we have discussed the concept of SUSY for Schr¨dinger Hamiltonians o in onedimension and for central potentials in higher dimensions which are essentially again onedimensional problems. which as mentioned above. (401) V2 = V1 + (1 − e−δr)2 10 Pauli Equation and Supersymmetry So far.05.748127 and 0. where the parameter δ is restricted to be less than 2. l = 0). second and third order using a shifted 1/N H expansion for Veﬀ are 0.748125 [201]. corresponds to the angular momentum barrier (k ′ − 1)(k ′ − 3)¯ 2 /8mr 2 for k ′ = 5 h (N = 5. considerable analytic simpliﬁcation occurs since it is suﬃcient to just use the leading term in the shifted 1/N expansion for V2 (r).swave Hulthen eﬀective potential in three dimensions and its ground state wave function are: H Veﬀ (r) = − (2−δ)r (2 − δ)2 2δe−δr + . 0.
hence the eigenfunction ψ can be factorized as ˜ ψ(x. A2 plane (A1 ≡ px + Ax .can also be put in SUSY form. Using the concepts of SUSY and shape invariance we show that some of the nonuniform magnetic ﬁeld problems can be solved analytically. y) = 0. Ay (x. Ax (x. 39] 1 Q1 = √ [−(py + Ay )σx + (px + Ax )σy ] 2 1 Q2 = √ [(px + Ax )σx + (py + Ay )σy ] 2 (403) satisfy the N = 1 supersymmetry algebra provided the gyromagnetic ratio g is two {Qα . 2 (404) It is interesting to note that SUSY ﬁxes the value of g. α. [H. A2 ≡ py + Ay ). y) = W (y) (405) where W (y) is an arbitrary function of y. Let us ﬁrst consider the problem in an asymmetric gauge i. In this case the Pauli Hamiltonian takes the form H = (px + W (y))2 + p2 − W ′ (y)σz (406) y ˜ Since this H does not depend on x. The Schr¨dinger o equation for ψ(y) then takes the form [− d2 + (W (y) + k)2 − W ′ (y)σ]ψ(y) = Eψ(y) dy 2 100 (408) .e. The above Hamiltonian has an additional 0(2) ⊗ 0(2) symmetry coming from σz and an 0(2) rotation in the A1 . The Pauli Hamiltonian for the motion of a charged particle in external h magnetic ﬁeld in two dimensions is given by (¯ = 2m = e = 1) g H = (px + Ax )2 + (py + Ay )2 + (∇ × A)z σz 2 (402) It is easily seen that this H along with the supercharges Q1 and Q2 deﬁned by [35. Qβ } = Hδ αβ . Qα ] = 0. y) = eikx ψ(y) (407) where k is the eigenvalue of the operator px (−∞ ≤ k ≤ ∞). β = 1.
101]. Thus we have reduced the problem to that of SUSY in one dimension with superpotential W (y) + k where W (y) must be independent of k. In particular from Table 4. In particular when W (y) = ωc y + c1 (409) which corresponds to the uniform magnetic ﬁeld.. This constraint on W (y) strongly restricts the allowed forms of shape invariant W (y) for which the spectrum can be written down algebraically. then the energy eigenvalues known as Landau levels are given by En = (2n + 1 + σ)ωc . We choose Ax = ωc yf (ρ). all these problems can be solved algebrically by using the results of Sec. n = 0. (410) Note that the ground state and all excited states are inﬁnitefold degenerate since En does not depend on k which assumes a continuous sequence of value (−∞ ≤ k ≤ ∞). The magnetic ﬁeld corresponding to the other choices of W are (ii) B = −asech2 y (iii)B = −a sec2 y( −π ≤ y ≤ π ) (iv) B = +c2 exp (−y) and as 2 2 mentioned above.1 we ﬁnd that the only allowed forms are (i) W (y) = ωc y + c1 (ii) W (y) = a tanh y + c1 (iii) W (y) = a tan y + c1 (iv) W (y) = c1 − c2 exp (−y) for which W (y) can b e written in terms of simple functions and for which the spectrum can be written down algebraically [202. 4. y) ≡ ∂x Ay − ∂y Ax = −2ωc f (ρ) − ωc ρf ′ (ρ) In this case the Hamiltonian (402) can be shown to take the form H = −( d2 d2 2 + 2 ) + ωc ρ2 f 2 − 2ωc f Lz − (2ωc f + ωc ρf ′ )σz dx2 dy (413) (412) 101 . Let us now consider the same problem in the symmetric gauge. The corresponding magnetic ﬁeld Bz is then given by Bz (x..where σ(= ±1) is the eigenvalue of the operator σz . Ay = −ωc xf (ρ) (411) where ρ2 = x2 + y 2 and ωc is a constant. 1. 2.
they do so only for one particular value of m while for other values of m. is the eigenvalue of Lz . (416) so that all the states are again inﬁnitefold degenerate. Clearly the corresponding Schr¨dinger problem can be solved in the cylindrio cal coordinates ρ. it corresponds to the motion of a charged particle in the x − y plane and subjected to a uniform magnetic ﬁeld (in the symmetric gauge) in the zdirection.. 2. In this case. In particular. This case again corresponds to the famous Landau level problem i.1. In this case the Schr¨dinger o equation for R(ρ) takes the form m2 − 1/4 d2 2 + ωc ρ2 f 2 − 2ωc f m − (2ωc f + ωc ρf ′ (ρ))σ + ]R(ρ) = ER(ρ) dρ2 ρ2 (415) where σ(= ±1) is the eigenvalue of the operator σz . 11 Supersymmetry and the Dirac Equation There have been many applications of SUSY QM in the context of the Dirac equation. we shall concentrate on only a few of these applications [100]. it is easily checked that there is only one shape invariant potential (f (ρ) = 1) for which the spectrum can be written down algebraically. The fact that in this example there are inﬁnite number of degenerate ground states with zero energy can be understood from the AharonovCasher theorem [204] which states that if the total ﬂux deﬁned by Φ = Bz dxdy = n + ǫ(0 ≤ ǫ < 1) then there are precisely n − 1 zero energy states.e. φ) can be factorized as √ ψ(ρ. However.where Lz is the zcomponent of the orbital angular momentum operator. we discuss the supersymmetric structure of the Dirac Hamiltonian (HD ) and show how the methods used 102 . In view of the limitations of space. φ.. the eigenfunction ψ(ρ. . the spectrum is in general not equispaced. 1. n = 0. ±2. ±1. The energy eigenvalues are [− En = 2(n + m+  m )ωc . On comparing with Table 4.. φ) = R(ρ)eimφ / ρ (414) where m = 0. Note that in our case Φ is inﬁnite. It is worth noting that the nonuniform magnetic ﬁelds can also give this equispaced spectrum [203].
ﬁnite energy. on the one hand. it can be looked upon as the kink solution of a scalar ﬁeld theory in 1+1 dimensions. 2 and 4. We also classify a number of magnetic ﬁeld problems whose solutions can be algebraically obtained by using the concepts of SUSY and shape invariance [101].1 Dirac Equation With Lorentz Scalar Potential The Dirac Lagrangian in 1+1 dimensions with a Lorentz scalar potential φ(x) is given by L = iψγ µ ∂µ ψ − ψψφ. kink solution corresponding to the scalar ﬁeld Lagrangian 1 Lφ = ∂µ φ(x)∂ µ φ(x) − V (φ). Next we discuss the celebrated problem of the Dirac particle in a Coulomb ﬁeld [103] and show that its eigenvalues and eigenfunctions can be simply obtained by using the concept of SUSY and shape invariance as developed in Secs. (417) The scalar potential φ(x) can be looked upon as the static. We show that whenever the onedimensional Schr¨dinger equation is analytically solvable for a potential o V (x).to obtain analytical solutions of the Schr¨dinger equation can be extended o to the Dirac case. We also discuss the problem of the Dirac equation in an external magnetic ﬁeld in two dimensions and show that there is always a supersymmetry in the problem in the massless case. 2 (418) Such models have proved quite useful in the context of the phenomenon of fermion number fractionalization [205. we discuss the path integral formulation of the fermion propagator in an external ﬁeld and show how the previously known results for the constant external ﬁeld can be very easily obtained by using the ideas of SUSY [101]. a variant of this model is also 103 . In addition. we show that the Euclidean Dirac operator in four dimensions. then there always exists a corresponding Dirac scalar potential problem which is also analytically solvable [101]. It turns out that. Finally. in the background of gauge ﬁelds. we consider the Dirac equation in 1+1 dimensions with Lorentz scalar potential φ(x). can always be cast in the language of SUSY QM. φ(x) is essentially the superpotential of the Schr¨dinger problem and o on the other hand. 11. 206. Further. 207] which has been seen in certain polymers like polyacetylene. First of all.
2 and 4 we then conclude that for every SIP given in 104 . ψ(x) = 0 −i ψ1 (x) ψ2 (x) (422) dψ(x) − φ(x)ψ(x) = 0. We get A= A† Aψ1 = ω 2 ψ1 . (417) and (418) has been absorbed in φ and V (φ) respectively. γ 1 = iσ 3 = 1 0 i 0 . Using the results of the Secs. we see that there is a supersymmetry in the problem and φ(x) is just the superpotential of the Schr¨dinger o formalism. The Dirac equation following from eq. A† = − + φ(x). Let ψ(x. dx dx We can now easily decouple these equations. d d + φ(x). AA† ψ2 = ω 2 ψ2 . t) = exp(−iωt)ψ(x) so that the Dirac equation reduces to γ 0 ωψ(x) + iγ 1 We choose γ 0 = σ1 = 0 1 . t) − φ(x)ψ(x. (417) is iγ µ ∂µ ψ(x. Note that the coupling constant in eqs. The spectrum of the two Hamiltonians is thus degenerate except that H− (H+ ) has an extra state at zero energy so long as φ(x → ±∞) have opposite signs and φ(x → +∞) > 0(< 0). Further ψ1 and ψ2 are the eigenfunctions of the Hamiltonians H− ≡ A+ A and H+ ≡ AA† respectively with the corresponding potentials being V∓ (x) ≡ φ2 (x) ∓ φ′ (x). t) = 0. 209].relevant in the context of supersymmetric ﬁeld theories in 1+1 dimensions [208. 2. dx (421) (420) (419) so that we have the coupled equations Aψ1 (x) = ωψ2 (x) A† ψ2 (x) = ωψ1 (x). where (423) (424) (425) On comparing with the formalism of Sec.
one can also construct perfectly transparent Dirac potentials with an inﬁnite number of bound states. using the results for the SIP with scaling ansatz (a2 = qa1 ) [58. In particular. using the reﬂectionless superpotential given by W (x) = n tanh x (426) one can immediately construct perfectly transparent Dirac potentials with n bound states [210]. 59].e.2 Supersymmetry and the Dirac Particle in a Coulomb Field The Dirac equation for a charged particle in an electromagnetic ﬁeld is given by (e = h = c = 1) ¯ [iγ µ (∂µ + iAµ ) − m]ψ = 0 (427) For a central ﬁeld i. Further. 11. α2 = m − E 105 (431) (430) .β = (429) i σ 0 0 −1 with σ i being the Pauli matrices. t) = V (r). this equation can be written as [211] ∂ψ i = Hψ = (α·p + βm + V )ψ (428) ∂t where 0 σi 1 0 αi = .Table 4.1 there exists an analytically solvable Dirac problem with the corresponding scalar potential φ(x) being the superpotential of the Schr¨dinger o problem. this Dirac equation can be separated in spherical coordinates and ﬁnally for such purposes as the computation of the energy levels one only needs to concentrate on the radial equations which are given by [211] kG − (α1 − V )F = 0 r kF − (α2 + V )G = 0 F ′ (r) − r G′ (r) + where α1 = m + E. A = 0 and A0 (x. For central ﬁelds.
and satisﬁes  k = J + 1 . γ = Ze2 r (432) We now show that the Coulomb problem can be solved algebraically by using the ideas of SUSY and shape invariance. Of course the radial functions Gk and Fk must be multiplied by the appropriate two component angular eigenfunctions to make up the full fourcomponent solutions of the Dirac equation [211].. ±3. (434) On multiplying eq... (433) from the left by the matrix D and introducing the new variable ρ = Er leads to the pair of equations m k ˜ − )G . To that end. one of the few exceptions being the case of the Dirac particle in a Coulomb ﬁeld for which γ V (r) = − . we ﬁrst note that in the case of the Coulomb potential. These coupled equations are in general not analytically solvable. Following Sukumar [103]. ±2. we now 2 notice that the matrix multiplying 1/r can be diagonalized by multiplying it by a matrix D from the left and D −1 from the right where D= k + s −γ .and Gk is the “large” component in the nonrelativistic limit. A† = − − + . E s m k ˜ ˜ A† G = −( + )F E S ˜ AF = ( where ˜ G ˜ F A= =D G F (435) (436) and d s γ d s γ − + . dρ ρ s dρ ρ s 106 (437) . the coupled equations (430) can be written in a matrix form as G′k (r) ′ Fk (r) + 1 r k γ −γ −k Gk Fk = 0 α2 α1 0 Gk Fk (433) where k is an eigenvalue of the operator −(σ·L + 1) with the allowed values k = ±1. s= −γ k + s k2 − γ 2 .
4. R(a2 ) = 2 − 2 a1 a2 so that the energy eigenvalues of H− are given by n+1 k 1 m2 1 (−) ( )2 − ( 2 ) ≡ En = R(ak ) = γ 2 ( 2 − ). 2... γ) = H− (ρ. s E k 2 m2 ˜ ˜ ˜ H+ G ≡ AA+ G = ( 2 − 2 )G. s. γ) + γ2 γ2 − . s + 1. for n = 0.˜ ˜ Thus we can easily decouple the equations for F and G thereby obtaining k 2 m2 ˜ ˜ ˜ H− F ≡ A+ AF = ( 2 − 2 )F . Notice that the spectrum as given by eq. 1. one can also alge˜ ˜ braically obtain all the eigenfunctions of F and G. s E (438) We thus see that there is a supersymmetry in the problem and H± are shape invariant supersymmetric partner potentials since H+ (ρ. only the negative value of k is allowed and hence this is a singlet state. s En s (s + n)2 k=2 (440) (441) Thus the Coulomb bound state energy eigenvalues En are given by En = m [1 + γ2 ]1/2 (s+n)2 . n = 0. However. (442) only depends on  k  leading to a doublet of states corresponding to k = k  and k = −  k  for all positive n. 107 . . s2 (s + 1)2 (439) On comparing with the formalism of Sec. 4 it is then clear that in this case γ2 γ2 a2 = s + 1. (442) It should be noted that every eigenvalue of H− is also an eigenvalue of H+ except for the ground state of H− which satisﬁes ˜ ˜ AF = 0 =⇒ F0 (ρ) = ρs exp(−γρ/s) (443) Using the formalism for the SIP as developed in Sec. a1 = s.
(404). Further. (403) satisfy the SUSY algebra as given by eq. A = 0. (445). in order to solve the Dirac equation.3 SUSY and the Dirac Particle in a Magnetic Field Let us again consider the Dirac equation in an electromagnetic ﬁeld as given by eq.e.(444) and then generate the corresponding Dirac solutions by the use of eq. Q1 and Q2 (see eq. In particular. (445) Thus. In the special case when the scalar potential A0 (and hence E) vanishes. then the Pauli eq. Now as shown by Feynman and GellMann [212] and Brown [213]. the twocomponent equation then has the canonical form of the Pauli equation describing the motion of a charged particle in an external magnetic ﬁeld. 214]. it is suﬃcient to solve the much simpler twocomponent eq. (444) and (445) we can immediately write down the exact solution of the massless Dirac equation in an external magnetic ﬁeld in two dimensions in all the four 108 . since (HD )2 = [α·(P + A)]2 = HP auli (446) hence. there is a supersymmetry in the massless Dirac problem in external 2 magnetic ﬁelds in two dimensions since HD . We can now immediately borrow all the results of the last section. (427) but now consider the other case when the vector potential is nonzero but the scalar potential is zero i. In particular. m = 0 and the motion is conﬁned to two dimensions. (444) exactly reduces to the eq. Clearly.11. this supersymmetry will also be there in two Euclidean dimensions. in view of eqs. if ψ obeys the two component equation [(P + A)2 + m2 + σ·(B + iE)]ψ = (E + A0 )2 ψ (444) then the four component spinors that are solutions of the massive Dirac equation are generated from the two component ψ via ψD = (σ·(P + A) + E − A0 + m)ψ (σ·(P + A) + E − A0 − m)ψ . the solution of the four component Dirac equation in the presence of an external electromagnetic ﬁeld can be generated from the solution of a two component relativistically invariant equation. Further. it follows that if the total ﬂux Φ(= Bz dxdy)=n + ǫ(0 ≤ ǫ < 1) then there are precisely n − 1 zero modes of the massless Dirac equation in two dimensions in the background of the external magnetic ﬁeld B (B ≡ Bz ) [204. If further. (402) of the last section. A0 = 0.
28. γ5 = 0 1 0 0 −1 (449) (448) (447) Let us take the following representation of the Euclidean γ matrices (450) .4 SUSY and the Euclidean Dirac Operator In the last few years there has been a renewed interest in understanding the zero modes and the complete energy spectrum of the square of the Dirac operator for a Euclidean massless fermionic theory interacting with the background gauge ﬁelds [25. 101]. one can also algebraically obtain the exact solution of the Dirac equation in an uniform magnetic ﬁeld in the symmetric gauge when the gauge potential depends on both x and y. γi = −i 0 0 σi 109 σi . On comparing the equations as given by (444) (with A0 = 0. 11. Further using the results of that section. Further. On deﬁning 1 Q± = (1 ± γ5 )γµ Dµ 2 where Dµ = ∂µ + iAµ . in the uniform magnetic ﬁeld case. 216]. exact solutions of the Pauli and hence the Dirac equation are also possible in the massive case. the solution can be obtained either in the asymmetric or in the symmetric gauge [202.situations discussed in the last section when the gauge potential depended on only one coordinate (say y). Summarizing. we conclude that the exact solutions of the massive (as well as the massless) Dirac equation in external magnetic ﬁeld in two dimensions can be obtained algebraically in case the magnetic ﬁeld B(≡ BZ ) has any one of the following four forms (i) B = constant (ii) B = −a sech2 y (iii) B = −a sec2 y(−π/2 ≤ y ≤ π/2) (iv) B = −c2 exp(−y). 215. one ﬁnds that the Dirac operator can be written as γµ Dµ = Q+ + Q− γ0 = 0 i . Let us ﬁrst show that there is always a supersymmetry in the problem of the Euclidean massless Dirac operator in the background of the gauge ﬁelds. E = 0) and (402) it is clear that the exact solutions in the massive case are simply obtained from the 2 2 massless case by replacing E by E − m2 . Even though there is no SUSY.
we ﬁnd γ µ Dµ = and hence − H = (γµ Dµ )2 = Dµ Dµ + σ·(B + E) 0 0 σ·(B − E) = LL† 0 0 L† L (453) 0 −iD0 + σ·D iD0 + σ·D 0 = 0 L† L 0 (452) where we have used the convention 1 Fµν = ∂µ Aν − ∂ν Aµ . Further. once we have an eigenfunction of H with a nonzero eigenvalue E1 .where σ i are the usual Pauli matrices. it is worth pointing out that in case the gauge potential Aµ has the special form Aµ = fµν ∂ν χ (456) where the constant 4 by 4 matrix fµν has the properties: fµν = −fνµ . Ei = F0i 2 (454) Using the techniques of Secs. one can algebraically obtain the eigenvalues and the eigenfunctions of H for few special cases [101]. [H. Q+ ] = 0 = [H. Bi = ǫijk Fjk . Q− ] (451) This supersymmetry is popularly known as chiral supersymmetry. 2 and 4. 110 . Q− }. then one can easily obtain the eigenfunctions φ± of γµ Dµ with eigenvalues ±(E1 )1/2 by the construction φ± = (γµ Dµ ± (E1 )1/2 )ψ (455) Finally. It is then easily seen that the operators Q+ . fµν fνα = −δµα (457) then there exists another breakup of the Dirac operator γµ Dµ which leads to the so called complex supersymmetry [101]. Q− and H ≡ −(γµ Dµ )2 satisfy the usual N = 1 SUSY algebra H = −(γµ Dµ )2 = {Q+ . In view of the above representation of the γ matrices.
In a nut shell. x. Here we would like to show that by introducing the fermionic degrees of freedom. This trick was ﬁrst introduced by Rajeev [218] who was interested in reformulating quantum electrodynamics as a supersymmetric theory of loops. (γµ Dµ )2 . x) = xµ xµ − iAµ xµ + σ µν Fµν ˙ ˙ ˙ ˙ 4 2 111 (462) G5 (x. For the case of a constant external ﬁeld strength Fµν . (461) 2i If we do not introduce auxiliary fermions then the related matrix valued Lagrangian would be 1 1 L(x. which allows one to trivially obtain G and G5 . τ ) =< x  e−Hτ  x′ >x=x′ and By using Schwinger’s proper time formalism we can determine S(x. Similarly. the bosonic path integral is a Gaussian. Let us consider the square of the Dirac operator i.5 Path Integral Formulation of the Fermion Propagator We have seen that there is always a supersymmetry associated with the Euclidean Dirac operator in four as well as two dimensions in the background of the gauge ﬁelds. which is initially a path ordered integral gets reduced to an ordinary path integral. γ ν ]. a purely bosonic path integral is then obtained.e. the path integral for the Dirac operator. τ ) =< x  T r(γ5e−Hτ )  x′ >x=x′ (459) (458) . the index Z5 which in the limit τ = 0 is related to the chiral anomaly. It can also be written as 1 (460) H = (P + A)2 − σ µν Fµν 2 where 1 σ µν = [γ µ .11. This SUSY was ﬁrst successfully exploited in the context of the study of the chiral anomalies [217]. By integrating over the fermionic degrees of freedom. The quantities one wishes to calculate are the Green’s functions G(x. Once we introduce the fermionic variables then G(G5 ) is determined by choosing the antiperiodic (periodic) boundary conditions for the fermions. the solvability of the Dirac operator is related to the ability to integrate the path integral. A) from G. x. x. is just the spatial integral over G5 .
Since the fermionic path integral is quadratic. To evaluate the determinant one puts Fµν in skew diagonal form: 0 −x = 1 0 0 Fµν x1 0 0 0 112 0 0 0 −x2 0 0 . Now. τ ) = Dxµ (t)Dψµ exp[− τ 0 {ψµ . 2 Then H can be written as i H = (P + A)2 + ψµ Fµν ψν 2 and hence the Lagrangian now becomes i 1 ˙ ˙ ˙ Lss = xµ xµ − iAµ xµ − ψµ (∂τ δµν + Fµν )ψν 4 2 which is invariant under the SUSY transformations δxµ = −iǫψµ . The result of the fermionic path integral is Det1/2 (i∂τ δµν + Fµν ) .and we would obtain for Feynman’s path integral representation < x  e−Hτ  x >= Dxµ (τ )P exp[− τ 0 dτ ′ L(x. (464) (465) (466) (467) dτ ′ Lss (x. x. ˙ We now obtain for the path integral G(x. 101] that one can introduce the Grassman variables ψµ via 1 ψµ = √ γµ . x2 0 (470) . x)]. taking analogy from SUSY quantum mechanics. ψν } = δµν . Det′ 1/2 (i∂τ δµν ) (469) where the prime denotes the omission of the zero mode. it has been noticed [218. δψµ = ǫxµ . we can perform the functional integral over the fermionic degrees of freedom exactly for arbitrary Fµν . ˙ (468) where we impose antiperiodic boundary conditions on the fermion at 0 and τ . x)] ˙ (463) where P denotes path ordering.
ǫµν Fµν 2 (475) It is worth remarking here that the same result has been obtained with more diﬃculty by Akhoury and Comtet [16]. In two dimensions. (471) 2 0 i and for the fermion determinant with periodic boundary conditions relevant for the determination of the index: 2 ( )n τ [− sinh i 1 2 τ 0 xi dτ ′ ]. where 1 F = Fµν Fµν 4 113 (476) . For the particular case of a constant external ﬁeld one can further explicitly perform the resulting bosonic path integral. could by performed numerically by Monte Carlo techniques. (472) Thus for any arbitrary external ﬁeld one can explicitly perform the fermionic path integral and is left with the purely bosonic path integral which. 1 G5 = B = . In four dimensions one has a simpler derivation of the result of Schwinger [219] rewritten in Euclidean space. When 1 Aµ = − Fµν xν . x2 = [F − (F2 − G2 )1/2 ]1/2 . one obtains for the remaining bosonic path integral: 1 xi (474) xi τ i 2 sinh 2 G and G5 are easy to calculate in two and four Euclidean dimensions for the constant external ﬁeld case [101] using this method.and imposes either periodic or antiperiodic boundary conditions. where only F01 = B exists one has that (x1 = B) [101] G = B tanh(Bτ ). One ﬁnds that x1 = [F + (F2 − G2 )1/2 ]1/2 . One obtains for the fermion determinant using antiperiodic boundary conditions relevant for G 1 τ cosh xi dτ. for example. 2 (473) with Fµν being a constant matrix.
1 ∗ G = Fµν Fµν 4 so that i 1 ∗ xi = Fµν Fµν . This general procedure ˜ is based on any arbitrary solution φ(x) of the Schr¨dinger equation for V (x). L(x) = ∞ ǫ dss−1e−m s G(s) 2 (479) 12 Singular Superpotentials So far. 114]. (478) 4 Once one has obtained the Green’s function it is easy to reconstruct the eﬀective action which allows one to obtain the rate of pair production from a strong external electric ﬁeld. we have only considered nonsingular superpotentials W (x) which give rise to the supersymmetric partner potentials V1 (x) and V2 (x). [114] The 114 . We shall see that singular superpotentials are given by Wφ = −φ′ /φ and the singularities are located at the zeros of excited wave functions. This choice of the superpotential was based on the ground state wave function ψ0 (x) of ′ V1 (x) by the relation W (x) = −ψ0 (x)/ψ0 (x). 4 i cosh x2τ xi τ i=1 sinh 2 (477) From this we obtain that: 2 ∗ G(x. τ ) = Fµν Fµν 1 ∗ G5 (x. 220. In this section. τ ) = Fµν Fµν . o rather than just the ground state wave function. Such singularities produce interesting properties for the potentials V1(φ) (x) and V2(φ) (x) obtained via the Riccati equations V1(φ) (x) = W 2 (x) −W ′ (x) and V2(φ) (x) = W 2 (x) + W ′ (x). Another interesting physical phenomenon occurs if one considers a singular superpotential resulting from a solution φ(x) for an energy E in the classical energy continuum. x. The eﬀective action is given by Sef f = dxL(x). 112. x. 113. Singular superpotentials are responsible for negative energy eigenstates for V1(φ) (x) and often give rise to a breakdown of the degeneracy of energy levels for V1(φ) (x) and V2(φ) (x) [110. we describe a general procedure for constructing all possible superpotentials which yield a ˜ given potential V (x) upto an additive constant [114].
the solution φ(x) has no 115 . If one deﬁnes the quantity Wφ = −φ′ /φ and takes it to be the superpotential.). which makes E0(0) = 0 and gives the familiar case ′ of unbroken SUSY. Negative Energy States and Breakdown of the Degeneracy Theorem ˜ Given any nonsingular potential V (x) with eigenfunctions ψn (x) and eigenvalues En (n = 0. let us now enquire how one can ﬁnd the most ˜ general superpotential W (x) which will give V (x) upto an additive constant ˜ [114]. 117]. 2. (482) One usually takes ǫ to be the ground state energy E0 and φ to be the ground state wave function ψ0 (x).isospectral family of V1(φ) (x) is then found to have bound states (normalised eigenfunctions) at energy E. the solution φ(x) is the eigenfunction ψn (x). since ψ0 (x) is normalizable and has no zeros. For convenience. For ǫ < E0 . however. 2 ′ V1(φ) = Wφ − Wφ = φ′′ ˜ = V (x) − ǫ . Note that whenever ǫ corresponds to one of the eigenvalues En . The eigenvalues of V1(φ) are therefore given by En(φ) = En − ǫ. both below and above the ground state energy E0 .. the remaining eigenvalues of V2(φ) are degenerate with those of V1(φ) .. (480) for the last step. 12. The partner potential V2(φ) has no eigenstate at zero energy since A0 ψ0 (x) = [d/dx + W0 (x)]ψ0 (x) = 0. Let us now consider what happens for other choices of ǫ.1 General Formalism. then clearly the partner potentials generated by Wφ are 2 ′ V2(φ) = Wφ + Wφ . (480) which vanishes at x = −∞. the superpotential W0 (x) = −ψ0 /ψ0 is nonsingular. With this choice. . we will always choose a solution φ(x) of eq. Thus SUSY QM provides a systematic procedure for generating potentials possessing the purely quantum mechanical phenomenon of bound states in the continuum[116. and without loss of generality. φ (481) where we have used eq. To answer this question consider the Schr¨dinger equation for V (x): o ˜ − φ′′ + V (x)φ = ǫφ (480) where ǫ is a constant energy to be chosen later. 1.
the potential has a singularity of “intermediate” strength. W (x) = −(x − xs )−1 . φ′ = a. Here. 4 From eq. for λ < − 1 . (482). Clearly. in general. say at the point x = xs . in principle. A summary of results is given in Table 12. Near the point x = xs . (482) will have only one node. the potential V (x) has a “strong” singularity 4 which forces ψ(xs ) = 0 and the range of x is eﬀectively broken into two disjoint pieces x < xs and x > xs with no communication between them.1. the problem of ﬁnding the eigenstates of V2(φ) is really two separate problems. V2(φ) = 2(x − xs )−2 .the only diﬀerence being the number of poles present. (481)] V1(φ) = 0. we will deal with regular solutions. Hence. Finally. 223] that for any singular potential V (x) = λ(x − xs )−2 . one in the range −∞ < xs and the other in the range xs < x < +∞. in what follows. E0(φ) = E0 − ǫ is positive. 222. Here. the behavior of the wave function at x = xs is governed by the value of λ. For values λ > 3 . and has the same sign for the entire range −∞ < x < +∞. The corresponding superpotential Wφ (x) is nonsingular. It is 4 4 not strong enough to make ψ(xs ) vanish and in fact the two regions x < xs and x > xs do communicate and one has the full range −∞ < x < +∞. (483) and the partner potentials have the behavior [using eq. it is clear that V2(φ) has a “strong” singularity which makes ψ(xs ) = 0. (484) It is wellknown [221. the Hamiltonian is unbounded from below. The above discussion can be readily extended to all values of the constant ǫ . Hence the eigenvalue spectra of V1(φ) and V2(φ) are completely degenerate and the energy eigenvalues are given by eq. if the potential 116 . Wφ has the same sign at z = ±∞. the node of φ(x) makes the superpotential W (x) singular φ = a(x − xs ). (484). and we have the wellstudied case of broken SUSY [10]. So far we have considered superpotentials (both nonsingular and singular) which give rise to a nonsingular potential V1(φ) . However. In particular. one can have singular wave functions which require selfadjoint extensions [221].nodes. then the solution φ(x) of eq. this is very diﬀerent from the range of V1(φ) which is the whole real axis −∞ < x < +∞. If the constant ǫ is chosen in the range E0 < ǫ < E1 . For − 1 < λ < 3 . the degeneracy theorem obtained from SUSY for the spectra of V2(φ) and V1(φ) is not valid. Thus.
n = 0. . Here again. ˜ Consider the harmonic oscillator potential V (x) = x2 . V2(0) = x2 + 1. one gets the usual nonsingular superpotential W0 = −ψ0 ′/ψ0 = x. The harmonic oscillator potential can also be obtained if one starts from the solution φ = ψ1 (x). ψ2 (x) = . The energy eigenvalues are En = 2n and the ﬁrst few eigenstates are √ 2 2 2 e−x /2 2 x e−x /2 (2x2 − 1)e−x /2 √ ψ0 (x) = . ψ1 (x) = . Consider the case of a simple pole singularity at x = x0 . for the two regions 1 3 − 3 < g < − 2 and 1 < g < 2 only one of the partner potentials has a 2 2 “strong” singularity. 2. negative energy states and breakdown of the degeneracy theorem is best illustrated with some speciﬁc examples. one expects degenerate energy levels (corresponding to broken or unbroken 1 SUSY). whereas the other has a singularity of “intermediate” strength. Therefore. one obtains degeneracy. then so must all superpotentials which produce it. 1.. En(0) = 2n + 2. Similarly.. The above discussion is borne out by the work of Jevicki and Rodrigues [110] who consider a superpotential of the form W (x) = g/x − x. both V1(φ) and V2(φ) have “intermediate” 2 strength singularities and the whole region −∞ < x < +∞ is valid for both. in general. the singular superpotential W (x) = g(x − x0 )−1 gives the following behavior to the corresponding partner potentials: V1(φ) = g(g + 1) (x − x0 )2 . Then. V2(φ) = g(g − 1) (x − x0 )2 . the two potentials have diﬀerent Hilbert spaces and. (485) 3 Then clearly. One gets the singular (1) (2) 117 . (486) π 1/4 π 1/4 2 π 1/4 Using the solution φ = ψ0 (x).V (x) itself has singularities. Here. both V1(φ) and V2(φ) have “strong” 2 singularities and both have two disjoint regions x < x0 and x < x0 . leading to the following partner potentials and eigenvalues: V1(0) = x2 − 1. En(0) = 2n. for − 1 < g < + 2 . for g > 2 or g < − 3 . degeneracy is not there. However. near x = x0 . (487) This is standard unbroken SUSY. The general discussion of singular superpotentials.
−1/ 2 < x < √ √ 1/ 2. 2. 44].. (2) n = 0. [110]. The superpotential based on the ground state is W0 (x) = A − Be−αx . The partner potentials are V1(1) = x2 − 3. ′ Taking φ = ψ2 (x). which was anticipated. 1. one gets W2 = −ψ2 /ψ2 = x − 4x/(2x2 − 1). The potential V2(2) is √ not analytically solvable. for the partner potentials V1(2) and V2(2) .1] due to the fact that the potentials are symmetric. Proceeding along the same lines. V2(1) = x2 + 2 − 1. The (1) corresponding energy eigenvalues are En(0) = 16 − (4 − n)2 . which pushes the ground state to a negative energy. The “strong” singularities at √ = ±1/ 2 break x √ the xaxis into three disjoint regions −∞ < x < −1/ 2. The states which are missing in V2(1) are the even parity states. since the 2/x2 barrier requires the vanishing of wave functions at x = 0. (489). .′ superpotential W1 = −ψ1 /ψ1 = x − 1/x. α = 1 and B = 1. x2 (488) and both are exactly solvable [54. The corresponding potentials are V1(2) = x − 5. ˜ Our second example is the Morse potential V (x) = A2 + B 2 e−2αx − B(α + −αx 2A)e . degeneracy is completely absent. In summary. 2 There occurs a negative energy state in V1(1) at E0(1) = −2. This is expected (1) V2(2) 8(2x2 + 1) .. For concreteness. the potentials V1(0) and V2(0) give a realization of SUSY QM with a nondegenerate zero energy ground state and pairing of excited states. There is partial degeneracy of the spectrum [see Fig. which has √ poles at x = ±1/ 2. there are four 118 . =x −3+ (2x2 − 1)2 2 (1) (490) The eigenvalue spectrum of V1(2) is En(2) = 2n − 4 indicating two negative energy states at −4 and −2. we can get the harmonic oscillator potential using yet another superpotential. The two eigenvalue spectra are En(1) = 2n − 2. Finally. (1) (489) since we chose φ to be the ﬁrst excited state and E1(1) = 0. It has a pole at x = 0 and is a special case of the form discussed in Ref. 12. we shall take A = 4. 1/ 2 < x < +∞ and the degeneracy theorem breaks down. For the case of V1(1) and V2(1) the eigenvalues are given by eq. (1) En(1) = 4n + 4.
5). Note that since the Morse potential is asymmetric.5). The potential V2(1) (left) has an inﬁnite number of bound states and as is obvious from the spectra.2.65 with a continuum above 9 and E(1) (left) = 22. The partner potentials are 441 − 567e−x + 281e2x − 56e−3x + 4e−4x (7 − 2e−x )2 (494) By construction. 8.1. 50. the potential V1(1) is the Morse potential and V2(1) has a singularly at xs = − ln(3. V2(1) = = 6. there is a negative energy state at 7. 119 (2) .08. V2(0) = 16 − 7e−x + e−2x . ψ1 7 − 2e−x (493) There is a simple pole at xs = − ln(3. 7.96. except that there is no state at zero energy..16.bound states with eigenvalues 0. the degeneracy between V2(1) and V1(1) is completely broken by the “strong” singularity in W1 (x). We have calculated the (2) energy levels of V2(1) (right) and V2(1) (left) numerically. 116. (492) The spectrum of V2(0) is identical to that of V1(0) . (491) The supersymmetric partner potentials constructed from the nonsingular superpotential W0 (x) = 4 − e−x are V1(0) = 16 − 9e−x + e−2x . 8 and a continuum above 9. They are E(1) (right) V1(1) = 9 − 9e−x + e−2x .. ψ1 (x) ∼ e−(3x+e −x ) (7 − 2e−x ).69. This is unlike our ﬁrst example (harmonic oscillator) which had symmetric potentials and V2(1) and V1(1) had a partial degeneracy. 5. Expanding V2(1) about the singular point gives V2(1) ∼ 2/(x − xs )2 . As expected from Table 12.. which requires the wave function to vanish at x = xs . 12 and 15 and a continuum starting above 16. The potential V1(1) has energy levels located at 7. The potentials and their eigenstates are plotted in Figure 12.. no partial degeneracy remains. 0. An alternative singular superpotential W1 (x) which also yields the Morse potential is W1 (x) = − ′ 21 − 15e−x + 2e−2x ψ1 = . 82.8. The two lowest eigenfunctions are [54] ψ0 (x) ∼ e−(4x+e −x ) . This eﬀectively breaks the potential and the real axis into two parts: V2(1) (left) for −∞ < x < xs and V2(1) (right) for xs < x < +∞.
we have shown how excited state wave functions can be used to construct singular superpotentials in SUSY QM.14) of ref. Thus. Our method provides a complete and uniﬁed picture of the origin of negative energy states and the presence or absence of degeneracy. V2(1) = (A − 1)2 − A(A − 1)sech2 x + 2cosech2 x. 226] have produced more examples and a better understanding of the kind of potential that can have such bound states. [112]. The energy eigenvalues for V1(0) are (1) En(0) = A2 − (A − n)2 and one has unbroken SUSY. (3. by many authors [117. En(1) = (A−1)2 −(A−2n−3)2 . but given the general method to construct new ones. our examples were taken from the class of shape invariant potentials. If one takes φ = ψ1 . These authors have also suggested possible applications to atoms and molecules. [227] have recently reported direct evidence for BICs by constructing suitable potentials using semiconductor heterostructures grown by molecular 120 . These potentials are shape invariant [54. Von Neumann and Wigner [116] realized that it was possible to construct potentials which have quantum mechanical bound states embedded in the classical energy continuum (BICs). which agrees with eq. 12. The motivation for considering this example is to make contact with singular superpotentials considered by Casahorran and Nam [112]. 44] one readily obtains ψ1 (x) = tanh x(sechx)A−1 as the wave function for the ﬁrst excited state. In conclusion. The ground state wave function is given by ψ0 = (sechx)A . 44] and their (1) (2) eigenvalues are En(1) = (A−1)2 −(A−n)2 . as expected for symmetric potentials. indicating partial degeneracy. the resulting su′ perpotential is W1 = −ψ1 /ψ1 = −2Acosech 2x + (A − 1) coth x. our previous experience indicates that the partner potentials constructed from W2 (x) will have no degeneracy. Similarly. Capasso et al. Using the property of shape invariance [54. Although our technique is perfectly general. since these are analytically solvable.˜ Our last example is the reﬂectionless potential V (x) = A2 − A(A + 2 1)sech x. 224.2 Bound States in the Continuum In 1929. 225. Further developments. The potentials generated by W1 (x) are V1(1) = (A − 1)2 − A(A + 1)sech2 x. we have not only reproduced some families of singular superpotentials previously considered in the literature. although there is not as yet a fully systematic approach. The nonsingular superpotential W0 = A tanh x comes from ψ0 .
(495) has a classical continuum of positive energy solutions which are clearly not normalizable.beam epitaxy. 7) but this time starting from states in the continuum. and use the methods of SUSY QM to generate families of potentials with bound states in the continuum [BICs] [115]. 72] 4u0 u′0 2u4 0 ˆ V (r. The method preserves the spectrum of the original potential except it adds these discrete BICs at selected energies. λ) which have the same eigenvalues as V (r). we show how one can start from a potential with a continuum of energy eigenstates. (495) where we have scaled the energy and radial variables such that all quantities are dimensionless. The radial swave Schr¨dinger equation for the reduced wave function o u(r) (in units where h = 2m = 1) is ¯ − u′′ + V (r) u(r) = E u(r). However. it is interesting to note that BICs have found their way into a recently written text [118]. In this subsection. I0 + λ (I0 + λ)2 where I0 (r) ≡ 0 r (496) u2 (r ′ )dr ′ . As illustrative examples. it is easy to generalize the above equations to the case where u0 (r) is any solution of 121 . Finally. λ) is given in terms of the original potential V (r) and the original ground state wave function u0 (r) by [68. u0 was taken to be the nodeless. 0 (497) Except in this section. The isospectral potential V (r. (a) One Parameter Family of BICs Consider any spherically symmetric potential V (r) which vanishes as r → ∞. These isospectral potentials are labeled by a real parameter λ in the ranges λ > 0 ˆ or λ < −1. As we have seen in Sec. the Darboux [64] procedure for deleting and then reinstating the ground state u0 (r) of a potential V (r). one is using the technique of generating isospectral potentials (discussed in Sec. Basically. we compute and graph potentials which have bound states in the continuum starting from a null potential representing a free particle and the Coulomb potential. Eq. generates a famˆ ily of potentials V (r. λ) = V (r) − 2[ln(I0 + λ)]′′ = V (r) − + . normalizable ground state wave function of the starting potential V (r). 70. 7.
(498) were originally inspired by SUSY QM. the integral I0 (r) in eq. Therefore. this leads to singular superpotentials and to singularities in the partner potential V2 (r). the resulting family of potentials ˆ V (r. u1 ) = (E0 − E1 )u0 u1 . After straightforˆ ward but tedious algebraic manipulations. λ) corresponding to the o same energies E0 and E1 . one gets Ei ui. but clearly u0 oscillates as r → ∞ and has an amplitude which does not decrease. One simply computes −ˆ′′ + V (r. u1). Our results are best summarized in the following statement: Theorem: Let u0(r) and u1 (r) be any two nonsingular solutions of the Schr¨dinger equation for the potential V (r) corresponding to arbitrarily seo ˆ lected energies E0 and E1 respectively. λ) = ˆ and [where W denotes the Wronskian. (495) with arbitrary energy E0 . While the new potential in eq. If u0 (r) has nodes. thus establishing ˆ the theorem. λ) = (E1 − E0 )u1 + u0 W (u0 . V (r. (496). (498) that u0 has the same zeros as ˆ ˆ the original u0 . At zeros of u0. λ) and V (r) are equal.eq. u1) ≡ u0 u′1 − u1 u′0 ] are solutions of ˆ the Schr¨dinger equation for the new potential V (r. Construct a new potential V (r. when the original state at E0 is reinserted. Consequently. λ) have a BIC with energy E0 .1). The algebra is considerably simpliﬁed by using the following identity for the Wronskian of two solutions of the Schr¨dinger equation: o d W (u0 . However. λ) as prescribed by eq. ˆ ˆ (499) u0 (r) . the easiest proof of the above ˆ theorem is by direct substitution. I0 + λ (498) Let us now take u0 to be a scattering solution at a positive energy E0 = k 2 of a potential V (r) which vanishes at r=∞. λ) is free of singularities [114]. All the other 122 . λ)ˆi ui u (i=0. Taking u0 (r = 0) = 0 satisﬁes one of the required boundary conditions. we see that all the potentials V (r. (497) now grows like r at large r and u0 is now square integrable ˆ for λ > 0. Note from eq. the two functions u0 (r. with the wave functions ui given in the theorem. dr (500) u1 (r. W (u0 . Negative values of λ ˆ are no longer allowed. (496) and the new wave functions in eq. Then. while the original wave function u0 was not.
The new wave function at E0 = k 2 also has an additional damping factor of 1/r which makes it square integrable. The integral I0 given in eq. We observe that I0 → r/2 as r → ∞.(497) becomes I0 = [2kr − sin(2kr)]/(4k). As u0 appears in the ˆ ˆ numerator of V . We now give two examples to explicitly illustrate how one applies the above procedure to obtain potentials possessing one BIC. We choose u0 = sin kr. is transformed into a solution of the new Schr¨dinger equation o by the operation given in eq.oscillatory solutions of the Schr¨dinger equation with V (r) get transformed o ˆ into oscillatory solutions to the new Schr¨dinger equation with V (r. The central column of Table 12.2 gives a convenient overview of the relationship ˆ of the potentials V and V and the solutions of the corresponding Schr¨dinger o equations. deﬁned in eq. the potential V and its BIC wave functions are shown in Figs. In particular. the spherical wave solution. which vanishes at r = 0. The original null potential has now become an oscillatory potential which asymptotically has a 1/r envelope. ˆ The potential family V .(496) becomes 32 k 2 sin4 kr 8 k 2 sin(2kr) ˆ V (r. λ) = 2kr − sin(2kr) + 4kλ. say u1 (r). corresponding to energy E0 = k 2 > 0.3a and 12. Consider a free particle on the half line (V ≡ 0 for 0 ≤ r < ∞). (496). Example 1: Free Particle on the Half Line. (496) and the BIC at energy We note that the new potential V E0 are formed using the corresponding wave function u0 (r).3b. (499) which involves both u0 and u1 . λ) with o the same energy. λ) = − 2 D0 D0 with ˆ V has a BIC at energy E0 = k 2 with wave function u0 (λ) = 4k sin kr/D0 . (503) (502) (501) (504) ˆ For special values of the parameters k and λ. eq. ˆ D0 (r. 12. note that u1 (r. Any other state. λ) in eq. every node of u0 is associated with a node of V ˆ 123 . λ) remains a nonnormalizable ˆ scattering solution of the corresponding Schr¨dinger equation. o ˆ (r.
Evidently. the only diﬀerence being in the sign of η . λ) plays the role of a damping distance. both for the new potential and for the new wave function. V becomes identical to V . [Note that the wave functions shown in the ﬁgures are not normalized]. Here we use our theorem to construct a oneparameter family of isospectral potentials containing a BIC. The value of the eigenenergy ˆ E0 is clearly above the asymptotic value. Useful expressions for these solutions in the regions near and far from the origin are available in the literature [228. reduced l = 0 wave function satisﬁes the Schr¨dinger equation eq. its magnitude indicates the value of r at which the monotonically growing integral I0 becomes a signiﬁcant damping factor. in terms of u0 . The procedure is the same for both positive and negative Z. 229]. η η where ˜ C0 (˜) = (e−πη/2 )  Γ(1 + i˜)  η η (506) (507) 124 .3b which are drawn for very diﬀerent values of λ.3a and 12. 229. In the limit ˆ λ → ∞. eqs.ˆ but not every node of V produces a node of u0 . The formal ˜ expressions for the BIC potentials and wave functions have been given above. The parameter λ which appears in the denominator function D0 (r. zero. the many oscillations of this potential.(495). conspire in such a way as to keep the particle trapped. and the unbound. following the method of Von Neumann and Wigner [116]. The parameter λ must be restricted to values greater than ˆ zero in order to avoid inﬁnities in V and in the wave functions. 12. Here V = Z/r. The solutions involve conﬂuent hypergeometric functions which in the asymptotic limit approach sine waves phaseshifted by a logarithmic term. 2iρ). 2. none of them able to hold a bound state. have constructed BIC potentials and wave functions for the case of the repulsive Coulomb potential. This is illustrated graphically in Figs.(496) and (498). Stillinger and Herrick [117]. The positive energy solution of eq.(505) can be written in the usual form [228. of the potential. 117] as the real function u0 (ρ) = C0 (˜) e−iρ M(1 − i˜. Example 2: Coulomb Potential. which o can be written in standard form η ˜ u′′ + (1 − 2 )u0 = 0 (505) 0 ρ √ √ ˜ with ρ = Er and η = Z/2 E.
Since the oscillating component vanishes whenever u0 vanishes. eq. (496) with V0 = Z/r. Our method works for the Morse potential which is steeply rising on the negative xaxis with correspondingly damped wave functions. (497) is not convergent if the starting point is chosen at −∞. the integral I0 in eq. Fig. while the operation in eq. For potentials. The results are displayed in Fig. k = 1. we observe that the denominator function given in eq. the situation is not so clear cut. unnormalizable wave function. (b) Two Parameter Family of Potentials In the previous section (a). It is seen that the potential which holds a bound state of positive energy shows an oscillatory behavior about the Coulomb potential. and Z = 2. (496) for ˆ V . we have ˆ V = V at each node of u0 . Note again. such as V (x) = −V0 sech2 x. the BIC wave function in both cases shows a damped behavior due to the denominator function. This is also seen in the ﬁgures. Fig. 12. For comparison the original Coulomb potential and wave function are also shown dotted. as is also evident from the form of eq. VC . where the range extends from −∞ to +∞. The corresponding oneparameter family of potentials obtained by the SUSY procedure is given in eq.and M(a.4. This leads to a vanishing denominator function in the expressions for some wave functions which makes them unacceptable. 12. (496). and it gets negative contributions if the starting point is selected at ﬁnite xvalues. It also works for the case of a particle in a constant electric ﬁeld for similar reasons. (498) was all that was needed to create the BIC. we have seen how to generate a oneparameter family of potentials with one BIC.4a shows the BIC partner to the attractive Coulomb potential for λ = 1. 12. there represented by u1 . are ˆ a solution to the new potential. A similar behavior is also expected for other spherically symmetric potentials with a continuous spectrum of positive eigenvalues. We now show how this procedure can be extended to construct twoparameter families which contain two BICs.4b shows the (unnormalized) wave function of the bound state in the continuum for this potential at E0 = k 2 . there is nothing special about the ordering of the two energy values nor the relative magnitude of E0 and 125 . we have obtained the BIC wave functions for representative values of λ. b. In constructing the new wave functions for the oneparameter family. Compared to the original. For arbitrary onedimensional potentials. (499) ensured that the wave functions for all the other states. z) is Kummer’s function. Using tabulated expressions for the Coulomb wave functions [229] and doing the integral for I0 numerically.
E1 . is illustrated in the last column of Table 12. but this time we transform the state at E1 into a BIC. (499). ˆ I1 + λ1 r 0 (508) (509) (510) (511) u2 (r ′ )dr ′. suitably modiﬁed. which already is a BIC. therefore we can repeat this procedure by applying the theorem to the wave functions and the potential of the oneparameter family. (508 . i (512) involving the original wave functions only. λ. (I0 + λ) (514) . which are now both BICs. The state at E0 . Making use of eq.510) explicitly shows the method ˆ of construction. (499) for u1 . is transformed in the step of eq. λ1) = V − 2[ln(I1 + λ1 )]′′ = V − ˆ1 + λ1 (I1 + λ1 )2 ˆ I with the solutions of the corresponding Schr¨dinger equation o ˆ ˆ u0 = (E0 − E1 )ˆ0 + u1 W (ˆ1. ˆ1 The precise relationship of the new potential and its wave functions.2. In this way we obtain the two parameter family of potentials 2ˆ4 u1 4ˆ1 u′1 u ˆ ˆ ˆ ˆ ˆ ˆ + V (r. ˆ we get ˆ I1 = r 0 (E1 − E0 )2 u2 + 1 u0 u1 u2 W 2 0 + 2(E1 − E0 ) W dr ′. ˆ u ˆ u ˆ ˆ u1 = ˆ and ˆ I1 ≡ u1 ˆ . While the compact form of Eqs. it is useful to observe that the integral I1 can be conveniently recast into a simpler form which contains integrals of the form Ii = r 0 u2 (r ′)dr ′ . to become a solution to the new potential. u0 ). 2 (I0 + λ) (I0 + λ) (513) The second term is integrated by parts as r 0 u2 −W 2 0 + W 2 (r ′ )dr ′ = (I0 + λ)2 I0 + λ 0 126 r r 0 2W W ′ ′ dr .
I1 (r) = I0 + λ (515) Here we have made use of the fact that our boundary conditions imply that W (0) = 0. 127 . As an example. we evaluate the twoparameter potential W 2 (r) ˆ ˆ + λ1 ] V = V − 2 ln (I0 + λ)[(E1 − E0 )2 I1 − I0 + λ The argument of the logarithm can be rewritten as (E1 − E0 )2 I0 I1 − W 2 (r) + λλ1 + λ(E1 − E0 )2 I1 + λ1 I0 . (516) We happen to have transformed ﬁrst the state at energy E0 into a BIC and then. We also note that transforming any state twice by eq. (518) Clearly. producing the parameters µ and µ1 . one expects symmetry. (517) is (E1 − E0 )2 I0 I1 − W 2 (r) + µµ1 + µ1 (E1 − E0 )2 I0 + µI1 . Let us now consider applying our procedure in the reverse order. Shown in Fig.We now use eq. the state at E1 . Finally. (517) ′′ .5 is a potential with two BICs at energies E0 = 1. [70]. We therefore have −W 2 (r) ˆ + (E1 − E0 )2 I1 (r). but simply changes the value of the parameter λ as shown in ref. in the second step. (500) for the derivative of a Wronskian of two solutions of the Schr¨dinger equation to rewrite the second term and observe. which introduced the parameters λ and λ1 . (513). that is let us ﬁrst transform the state at energy E1 into a BIC and then the state at energy E0 . µ1 = λ1 /(E1 − E0 )2 . (498) does not create a second denominator or anything else new. (519) This also leads to the same twoparameter wave functions. the argument corresponding to eq. 12. Clearly. the above procedure can be readily extended to obtain multiparameter families with multiple BICs at arbitrarily selected energies. that it exactly o cancels the last term in eq. This is guaranteed if the parameters are related by µ = λ(E1 − E0 )2 . relation (509) ensures that all other eigenstates will be solutions to the new potentials. For this situation. E1 = 4.
the structure relations involve products of (p + 1) parasupersymmetry (PARASUSY) charges. On the other hand. These authors have also speculated that the BIC energies they compute in e+ e− scattering in QED are in fact in reasonable agreement with unusual peaks observed in recent heavy ion experiments [231]. For example. but the speciﬁc symmetry algebra was still the usual SUSY one.). and as stated before there is now some experimental evidence for the existence of BICs under appropriately chosen conditions [227]. In this section. In view of the fact that the SUSY has provided us with an elegant symmetry between fermions and bosons. 234. there has also been a computation which claims the existence of BICs in QED in three dimensions [230]. Such a question was raised many years ago in the context of parastring models [237]. 13 Parasupersymmetric Quantum Mechanics and Beyond In the last few years. in three and higher space dimensions parafermi and parabose statistics [233. the symmetry generators obey structure relations that involve bilinear products. 241. It is worth adding here that historically the PSQM of order 2 was introduced ﬁrst [106] and its various consequences were discussed [238. We shall construct parasupersymmetric quantum mechanics (PSQM) of a boson and a parafermion of order p(= 1. 2. it is natural to enquire if there exists a generalization which includes the above exotic statistics. we study the possibility of having a symmetry between bosons and parafermions. 239.. In particular. Recently. in PSQM of order p. parafermi and parabose statistics describe the higher dimensional representations of the same group. one can have a oneparameter family of statistics interpolating between Bose and Fermi statistics [232]. 242].. in two space dimensions.Our discussion of BICs has been restricted to eﬀectively one dimensional problems. 235. 236] are the natural extensions of the usual Fermi and Bose statistics. Initially it was felt that the generalization to order p was not 128 . exotic quantum statistics have been widely discussed in the literature. 240. Various consequences of this algebra are also discussed. It turns out that whereas in the usual SUSY QM. 3. whereas Fermi and Bose statistics describe the two onedimensional representations of the permutation group.
the energy eigenvalues are necessarily nonnegative. 2.. QQ† + Q† Q = 2H. This results from assuming m = 1 in this section. However. a† = 0 0 0 0 . 13. Unlike the PSQM of order p. b† of order p (=1. a† } = 1.. The SUSY QM algebra is easily motivated by recalling that the fermionic operators a. in conformity with the notation followed by the various authors in this ﬁeld. new forms of quantum statistics called orthofermi and orthobose statistics. Various consequences of SUSY QM. a† satisfy the algebra a2 = 0 = a†2 . {a. [H. Q] = 0. ) which are known to satisfy the algebra [236] (b)p+1 = 0 = (b† )p+1 . 1 0 (522) (521) Let us now consider the parafermi operators b.possible in the sense that the PSQM of order p cannot be characterized with one universal algebraic relation [243]. A useful representation of a and a† is in terms of the 2 × 2 matrices a= 0 1 . Very recently.2.. one ﬁnds that the structure relations now involve only bilinear products of the symmetry generators.1 Parasupersymmetric Quantum Mechanics In order to motivate the algebra of PSQM. one was indeed able to construct such a PSQM of order p [107]. have been constructed [244]. 129 (523) . let us recall that in SUSY QM the symmetry between a boson and a fermion is characterized by the algebra Q2 = 0 = Q†2 . in PSQM of order p they need not be so. (520) Note that there is an extra factor of 2 on the right hand side compared to the algebra given in Sec. In particular. PSQM and OSQM are also discussed in this section. it is worth pointing out that whereas in SUSY QM and OSQM of order p. It is then natural to construct orthosupersymmetric quantum mechanics (OSQM) where there is symmetry between a boson and an orthofermion of order p [109]. later on.
β where Cβ = β(p − β + 1) = Cp−β+1 . (529) It is easily checked that the operators b and b† indeed satisfy the algebra as given by eqs. (b† )αβ = Cβ δα+1. − 1.β+1 . − + 1. (523)? It turns out that the nontrivial relation is 1 bp b† + bp−1 b† b + .. [[b† . for p = 1 this reduces to the bilinear relation for the fermionic operators given in eq. (p + 1)] (b)αβ = Cβ δα. As expected. This relation between b and b† strongly suggests that one may have an analogous multilinear relation in the algebra of PSQM of order p. (523) and (524).. To that purpose. . One can now ask as to what multilinear relation is satisﬁed by b and b† apart from the one given by eq. J± ] = ±J± ... J− = b. (Q† )αβ = (P + iWα )δα+1. (525) 2 it immediately follows from eqs. let us choose the PARASUSY charges Q1 and Q† as (p + 1) × (p + 1) 1 matrices as given by (Q1 )αβ = (P − iWβ )δα.[[b† ..β+1 . J− ] = 2J3 . b]. J3 = [b† . b].. + bb† bp−1 + b† bp = p(p + 1)(p + 2)bp−1 . (523) and (524) that the operators J± and J3 satisfy the SU(2) algebra [J+ . − ). (526) p 2 Let us now choose J3 to represent the third component of the spin representation of the SU(2) group as given by p p p p J3 = diag( . b] = −2b. 2. β = 1.β 1 130 (531) .. 2 2 2 2 (527) It is now easily seen [107] that the operators b and b† can be represented by the following (p + 1) × (p + 1) matrices [α. 6 (530) (528) where one has (p + 1) terms on the left hand side. On letting (524) 1 J+ = b† . [J3 . (521). b† ] = 2b† . b].
p) Hr = Hp+1 P2 1 + (Wr2 − Wr′ ) + 2 2 2 1 P 2 ′ + (Wp + Wp ) + = 2 2 1 Cr 2 1 Cp 2 (533) (532) (534) commutes with the PARASUSY charges Q1 and Q† provided (s = 2.... + Cp = 0..... p) Cr+1 − Cr = 2ω. . . 2. 2.. 2 2 131 (540) (539) (538) . β = 1. 1 1 Further.. Q† and H is given by [107] 1 Qp Q† + Qp−1 Q† Q1 + . so that Q1 and Q† automatically satisfy 1 Qp+1 = 0 = (Q† )p+1 . (535) that in this case (r = 1. If one chooses W1 = W2 = . Cp are arbitrary constants with the dimension of energy.. . (p + 1)..... It turns out that the nontrivial relation between Q1 . + Q1 Q† Qp−1 + Q† Qp = 2pQp−1 H.. C2. where (r = 1. it is easily shown that the Hamiltonian (¯ = m = 1) h (H)αβ = Hα δαβ . and the Hamiltonian (533) takes a very simple form given by H= P2 1 2 2 + ω x − J3 ω. p) 1 2 ′ Ws−1 + Ws−1 + Cs−1 = Ws2 − Ws′ + Cs .. (537) An example is in order at this stage to illustrate the structure of PSQM of orde r p. . then it follows from eq. (535) Here C1 ... = Wp = ωx.. 2. 3..where α. 1 1 1 1 1 1 1 1 1 (536) and Hermitianconjugated relations (which we shall not write explicitly) provided C1 + C2 + ..
132 .. and ﬁnally the p’th and higher excited states are (p + 1)fold degenerate. (543) all of which commute with the Hamiltonian (534) and also satisfy the PSQM algebra as given by eqs. independent PARASUSY charges Q2 .. Apart from Q1 . 3.. (532) and (536) [107]. The spectrum of this Hamiltonian is En.. 2 (541) where n = 0. 1... the ﬁrst excited state is twofold degenerate. Q3 . 2. (542) where Hr is given by eq. It is also clear from here that whereas the ground state is nondegenerate. etc.m = (n + 1 − m)ω. For arbitrary Wr too one can show that the spectrum of the PSQM Hamiltonian is (p + 1)fold degenerate at least starting from the p’th and higher excited states. . (534). . (533) and (534)... there are (p − 1) other conserved. The nature of the ground and the ﬁrst (p−1) excited states would however depend on the speciﬁc form of Wr .. This H describes the motion of a particle with spin p/2 in an oscillator potential and a uniform magnetic ﬁeld. With the PSQM Hamiltonian as given by eqs. 3. Qp deﬁned by (s = 2.. and m = p/2.β+1 if β = s.where J3 is as given by eq. p) and the corresponding p SUSY QM Hamiltonians are given by HSU SY = (r) 1 Hr − 2 Cr 0 0 Hr+1 − 1 Cr 2 . Several comments are in order at this stage: 1. Of course for p = 1 we recover the well known results of SUSY QM.. −p/2 so that the ground state energy of the system is negative unlike in the usual SUSY QM.p) Qs = (P − iWβ )δα.. . . 2.. one can associate p supersymmetries characterized by the corresponding p superpotentials Wr (r = 1. (527). . = −(P − iWβ )δα.β+1 if β = s. p/2 − 1. 2.
+ E1 ]. This is because. As discussed above.. ψp ..Cp are chosen to the zero. 3. Hp [107]. H3 . H cannot be directly expressed in terms of the PARASUSY charges Q1 and Q† . Of course shape invariant potentials satisfy eq.. Besides. . 5.. What is the most general solution of the relation (535) which must be satisﬁed in order to have the PSQM of order p? Treating Wr ± Wr−1 as the two variables. H is multiplied by Qp−1 whose inverse does 1 1 not exist. eq. 3. Cp are related to the energy eigenvalues by (r = 1. 2.... Wp ... It turns out that in case all the constants C1 . one can always generate p other Hamiltonians H2 . . . one can also associate parasupersymmetry of order p to the hierarchy of Hamiltonians H1 . One unsatisfactory feature of the PSQM of order p is that except for p = 1. Ep and the corresponding eigenfunctions ψ1 .... one would recursively obtain W3 . .... one can always construct W1 . As we have seen in Sec. Unfortunately. 2. we would like to know the most general solution for W2 for a given W1 and then using this W2 ..... the p constants C1 . Further in that case there are p supersymmetries with the corresponding Hamiltonians being precisely given by eq. There is an interesting application of PSQM in nonrelativistic quantum mechanics. C2 .... W2 . the energy eigenvalues are not necessarily nonnegative and there is no connection between the nonzero (zero) ground state energy and the broken (unbroken) PARASUSY.. C2 . W4 . in eq. Another unsatisfactory feature is that unlike in SUSY QM. (544) p Thus instead of associating the symmetry algebra sl(1/1) ⊗ SU(2) or U(1) ⊗ SU(2) as we did in Sec. Ideally.. . ψ2 .Wp which will satisfy eq. . This is however not very useful as it gives us Wr + Wr−1 for a given Wr − Wr−1 .. . (535) but shape invariance is clearly not necessary. (536)..... this problem is still unsolved. . p levels respectively are missing from them. H(p+1) with the same spectrum as H1 except that 1.. (535) can be reduced to a simple nonlinear equation which can be immediately solved [106]. (542).4. (535). given any H with p bound states... . p) 2 Cr = [Ep + Ep+1 + . H2 . given any Hamiltonian H1 with p bound states with energies E1 .. E2 . then one can take care of both of these unsatisfactory features 133 . + Er+1 − (p − 1)Er + Er−1 + ..
2. in that case H as given by eq.. (545) shows an explicit expression for the Hamiltonian H which involves just one charge Qs but an overall square root. forms PSQM of 2 ′ order p along with its SUSY partner potential V2 = W1 + W1 and with the strictly isospectral potential families V2 (x. p) 1 † H = [(Q† Qs − Qs Q† )2 + Q† Qs Qs Qs ]1/2 s s s 2 (545) It is immediately clear from here that all the energy eigenvalues are necessarily nonnegative and that the ground state energy being 0 (> 0) corresponds to unbroken (broken) PARASUSY. but which involves all the p PARASUSY charges [108]: 2H = Qr Q† r + Q† Qr r 1 p + (Q† Qs + Q† Qr − 2Q† Qr ) s r 4 s=1 r (547) where. In particular. λ3 ). 2. V2 (x. Further. . λ3 . ... (548) r In case one chooses Wα to be of form (α = 1.. . s = 1. one can show that 2 ′ any potential V1 ≡ W1 − W1 with at least one bound state.. .. r.. λ2 ). H and Qr also satisfy the simpler relation Qr Q† Qr = 2Qr H.. 7. Eq.. There is an interesting application of this version of PSQM to the strictly isospectral Hamiltonians discussed in Sec. λp ) where λ2 . 2.[108].p and s = r . p) Wα = − λ+α−1 . Further. In particular. There is an alternative expression for H where it is not necessary to take the square root.(534) can be expressed in terms of any one of the p PARASUSY charges Qs by (s = 1.. In the limit of p → ∞ it has been shown that the relation (545) reduces to 1 H = Qs Q† .. λp are arbitrary parameters which are either > 0 or < −1 [203]. s 2 (546) which can be termed as the PSQM of inﬁnite order whose algebra corresponds to that of Greenberg’s inﬁnite statistics [245].. x (549) 134 .. one can also show that in this case all the excited states are always (p + 1)fold degenerate while the nature of the ground state will depend on the speciﬁc form of Ws [108]. V2 (x..
we ignore the spatial indices in eqs. correspond to space and spin indices respectively.. one can also deﬁne orthobose statistics... . j = 1...j (Sr )ij = −xδi+1.. The wave function is thus antisymmetric in spatial indices alone with the order of the spin indices frozen. K = x2 4 2 satisfy relations analogous to those given in eqs.then one has a model for conformally invariant PSQM of order p. . (p + 1).j if i = r + 1. . 2. In this case one can show that the dilatation operator D and the conformal operator K deﬁned by 1 1 (550) D = − (xP + P x). β. Orthofermi statistics contain a new exclusion principle which is more stringent than the Pauli exclusion principle: an orbital state shall not contain more than one particle. In an analogous way. m. 2.j if i = r + 1. 3. = xδi+1. (545). r = 2. p] (S1 )ij = −xδi+1... and the Greek indices α. whatever be the spin direction. For constructing the quantum mechanics of a boson and an orthofermion..2 Orthosupersymmetric Quantum Mechanics Recently... i. Mishra and Rajasekaran [244] have introduced new forms of quantum statistics called orthofermi and orthobose statistics.. All these properties follow provided the corresponding creation and annihilation operators C † and C satisfy p † Ckα Cmβ ± δαβ † Cm γ=1 gamma Ckγ = δkm δαβ . (555) 135 . p. (553) and (554) and obtain p † Cα Cβ Ckα Cmβ ± Cmα Ckβ = 0. (553) (554) where the upper and the lower signs lead to the orthofermi and the orthobose cases respectively and the Latin indices k. γ. + δαβ γ=1 † Cγ Cγ = δαβ . .. (547) and (536) in terms of the PARASUSY charges Qα and parasuperconformal charges Sα deﬁned by [α = 1. (551) (552) 13. .
Qα ] = 0. (556) (557) Following the discussion of the last subsection it is easy to see that a useful representation of these operators is in terms of the (p + 1) × (p + 1) matrices deﬁned by (α. [H. . 2.. 2.α+1 α and the Hamiltonian H as (H)rs = Hr δrs ... γ (559) (560) Qα Qβ = 0. = Cp Cp . On comparing the algebra for the fermionic and the orthofermionic operators as given by eqs. .. 2.Cα Cβ = 0. . (Q† )rs = (P + iWα )δs. r.. 2. (p + 1)) † (Cα )rs = δr.1 δr. p.. Q† and the Hamiltonian must satisfy the algebra α (α... Note that for p = 1 we recover the usual SUSY QM algebra. s = 1.(520) it is easy to convince oneself that the p ORTHOSUSY charges Qα . Eq.1 δs.. (555) implies that † † † C1 C1 = C2 C2 = .1 δr. It is easily checked that if we choose the p ORTHOSUSY charges Qα as (p + 1) × (p + 1) matrices as given by (r = 1.. 2.α+1 . (555) and (556) and remembering the SUSY QM algebra as given by eq. .1 δs...α+1 . (Cα )rs = δs. . (521). where (r = 1.α+1 (558) Let us now try to write down the algebra for the OSQM of order p where there is a symmetry between a boson and an orthofermion of order p [109]. p) p Qα Q† + δαβ β γ=1 Q† Qγ = 2δαβ H. β = 1.. (p + 1)) (Qα )rs = (P − iWα )δr.. p) H1 = Hr+1 P2 1 2 ′ + (W1 + W1 ) 2 2 P2 1 2 ′ + (W1 − W1 ) = 2 2 136 (562) (561) (563) .. β = 1...
then the OSQM algebra as given by eqs. (559) and (560) is indeed satisﬁed provided 2 ′ 2 ′ Wα + Wα = Wβ + Wβ . (564) Note that this condition directly follows from the OSQM relations Q1 Q† = Q2 Q† ... = Qp Q† , 1 2 p (565)
which follow from the OSQM relation (559). Thus, unlike PSQM, the constants C1 , C2 , ... are not allowed in OSQM. The various consequences of the OSQM have been discussed in detail in [109]. At this point, it may be worthwhile to make a relative comparison of SUSY QM with PSQM and OSQM. 1. First of all, the close similarity in structure between OSQM and PSQM must be noted. For order p, both are based on (p + 1) × (p + 1) matrices and the structure of H in the two cases is very similar. The chief diﬀerence between the two is the absence of the constants C1 , C2 ..., Cp in the former while in the latter they may or may not be zero. 2. Whereas the ground state energy is zero (> 0) in case the SUSY or the ORTHOSUSY is unbroken (spontaneously broken), in general, there is no such restriction in the PSQM case and the energy eigenvalues can even be negative. However, in the special case when C1 , C2 , ..., Cp are all zero then the PSQM has similar prediction as the other two. 3. Whereas in the SUSY QM, all the excited states are necessarily twofold degenerate, in OSQM of order p they are necessarily (p + 1)fold degenerate. On the other hand, in the PSQM of order p, all the levels starting from the p’th excited state (and above) are necessarily (p + 1)fold degenerate except when all the constants are zero in which case all the excited states are necessarily (p + 1)fold degenerate.
14
Omitted Topics
So much work has been done in the area of SUSY QM in the last 12 years that it is almost impossible to cover all the topics in such a review. We have therefore decided to give a brief description of some of the omitted topics.
137
For each topic, a few references are provided, so that interested readers can go back and trace other references and get a good idea of the developments. (i) Supermathematics [246]. (ii)SUSY in Atomic Physics [41, 42] In a series of papers, Kostelecky et al. [41] have discussed the relationship between the physical spectra of diﬀerent atoms and ions using SUSY QM. In particular, they have suggested that the helium and hydrogen spectra come from SUSY partner potentials. This connection has been commented upon by Rau [42]. (iii)SUSY in Condensed Matter and Statistical Physics [163, 248, 249, 250, 251, 252] Ideas from SUSY have been used to give insight into random magnetic ﬁelds in Isinglike models, polymers, electron localization in disordered media and ferromagnets. These topics are covered in a set of review talks found in the Proceedings of the Conference on Supersymmetry held at CNLS in 1983 [15]. (iv)Index Theorem and SUSY [253, 25, 28, 26, 27] The AtiyahSinger index theorem can be related to understanding the index of the Dirac operator on suitably deﬁned spaces. From our previous discussion of SUSY and the Dirac equation, it is not surprising that the index of the Dirac operator is related to the Witten index of the related SUSY QM. Using techniques similar to calculating the fermion propagator in an external ﬁeld, it has been possible to give a proof of the AtiyahSinger index theorem using the supersymmetric representation of the index theorem. This work is best described in lectures of AlvarezGaume given at the Bonn Summer School [27]. The question of axial anamolies are best phrased using these methods. (v)Factorization Method and Solvable Potentials [49, 14, 51] The method of factorization which can be traced as far back as Bernoulli(1702) and Cauchy(1827) can be shown to be exactly equivalent to solving potentials by SUSY and shape invariance with translation. The reader interested in the history as well as an excellent presentation of the factorization method and its connection with SUSY is referred to the work of Stahlhofen [51]. (vi)Group Theory Method and Solvable Potentials [254] In the doctoral thesis of Jainshi Wu, it is proven that there is a one to one correspondence between using the diﬀerential realizations of the S0(2, 1) potential group and the factorization method of Infeld and Hull [14]. Thus 138
all the solvable potentials found by Infeld and Hull can be also obtained using group theory methods. By extending the symmetry group to SO(3, 1) and SU(3, 1) the Yale group was also able to study several three dimensional scattering problems. (vii)QuasiSolvable Potentials [255, 256] These potentials, for which a ﬁnite number of states can be determined analytically, are intermediate between nonsolvable and analytically solvable potentials discussed in this review. Quasisolvable potentials have been discussed by Shifman and Turbiner [255, 256] and using the techniques of SUSY QM a large number of new quasisolvable problems have been discovered [257]. (viii)SUSY Breaking and Instantons [12, 30, 11, 258, 259] One of the least understood problems in supersymmetric ﬁeld theories is that of the origin of SUSY breaking. One suggestion for SUSY breaking is that it is of dynamical origin and that instantons are responsible for it. As a testing ground of these ideas, the role of instantons in the dynamical breaking of SUSY has been studied extensively in various quantum mechanical models. (ix)Propagators for SUSY Partner Potentials [98, 99] Since the propagator for a system with a given potential is determined from the Hamiltonian, and the various Hamiltonians in a heirarchy of SIP are related by Q(as )H (s) = H (s+1) Q(as ), (566) one can derive recursion relations for the propagators of the heirarchy. If one member the heirarchy has a known propagator (such as the harmonic oscillator or the free particle) then it is easy to use these recursion relations to derive the propagators for the heirarchy. This also allows one to calculate path integrals for SIP. (x) SUSY and Nbody Problem [260, 150] A novel algebraic structure is found for 1/r 2 family of many body problems which provide a novel link between SUSY and quantum integrability [260]. Also of considerable interest is a supersymmetric generalization of a Nparticle quantum mechanical model with combined harmonic and repulsive forces [150]. (xi) Coherent states for SIP [261] Recently a Lie algebraic treatment of the SIP has been given and using it coherent states have been constructed for these potentials. 139
265]. (xv) SUSY and Nuclear Physics [247. 140 . We would especially like to thank Joe Ginocchio for carefully reading the manuscript. Also. Relations between the spectra of eveneven and neighbouring evenodd nuclei has been obtained by using the concepts of SUSY [247]. Department of Energy for partial support. It has also been suggested that the relationships seen between the energy levels of adjacent superdeformed nuclei can be understood in terms of SUSY [265] Acknowledgments It is a pleasure to thank all of our many collaborators without whose eﬀorts this work would not have been accomplished. [264] have succeeded in factorizing the time dependent Pauli Equation and utilizing the SUSY to solve the Pauli Equation for a time dependent spatially uniform magnetic induction. It has been o shown that new scalar ﬁeld theories with kink solutions can be obtained by considering the isospectral deformation of this problem. (xiii) SWKB and Tunneling [263. 82] Expression for transmission coeﬃcient T through potential barriers has been obtained within the SWKB approximation and it has been shown that the analytic continuation of T for the inverted SIP (with translation) leads to the exact bound state spectrum. We gratefully acknowledge the U. We would like to thank Los Alamos National Laboratory for its hospitality during the writ ing of the manuscript.S. (xiv) TimeDependent Pauli Equation [264] Kostelecky et al. 157. Baye [157] has shown that the deep and shallow nucleusnucleus potentials which have been successfully used in the past are in fact SUSY partner potentials.(xii) SUSY and New Soliton Solutions [262] It is well known that the stability equation for the kink solutions in scalar ﬁeld theories in 1+1 dimensions is a Schr¨dingerlike equation.
Akulov. Cooper and A. D3 (1971) 2415. Sohnius. Ann. Phys. Lett. Hull. Neveu and J. editors: V. Maharana.R. Girardello. 128 (1985) 39. [8] S. Mod. Rev. Phys. Phys. Kostelecky and D.F. Akhoury and A. Nucl. Das. M. B145 (1984) 77. Rev. Phys. D28 (1983) 1473. Phys. Volkov and V. Phys. F. Wijerwardhana. B202 (1982) 253. [10] E. [13] C. Khare and J. Ramond. Phys. Supersymmetry in Physics. Phys. B138 (1984) 389. Phys. [5] J. Deser and B. van Nieuwenhuizen. [2] P. North Holland (1985). Niemi and L. Nucl. [17] A. [18] A. Likhtman. Campbell. [16] R. Lett.A. [15] See for example. Nucl. 159 (1967) 1251.E. [11] F.C. Zumino. 141 . Phys. [14] L. Phys. B70 (1974) 39. Bender. Lett. Phys. B245 (1984) 213. [3] A. ibid. B188 (1981) 513. Infeld and T. Phys. D13 (1976) 3214. Ann. Cooper and B. Rep.A. Freedman and P. [4] D. Wess and B. 13 (1971) 323. Nucl. 146 (1983) 262. B62 (1976) 335. Witten. Rev. Freedman.References [1] Y. Witten. Lett. 23 (1951) 21. Zumino. [6] For more details and other references in this ﬁeld see. Rev. B78 (1974) 1. Schwarz. [12] E. Phys. Comtet. Gel’fand and E. Nucl. B31 (1971) 86. Coleman and J. [9] S. JETP Lett.Ferrara. Phys. D. B46 (1973) 109. [19] S. 145 (1983) 81. Phys. Mandula.P. Cecotti and L. [7] S. for example. Phys. Rev.K.
M. Cooper. Marchesoni. 51 (1983) 755. Lett. Physica D15 (1985) 71. Kovacs and U. Lett. Hernandez. Ann. [37] H. Mod. Prog. Phys. 813. 142 . [33] F. Rittenberg. Phys. A4 (1989) 2701. C28 (1985) 203. Phys. Phys. unpublished. 57 (1986) 2749. Lett. B196 (1982) 509. [32] P. Lett. Salomonson and B. Lett. Phys. Rev. Phys. Sourlas. E. Kihlberg. 72 (1984) 266. Girardello. [36] L. Bhaduri. Physica D15 (1985) 138. D30 (1984) 1821. Khare and M. Zeit. Kumar. Bernstein and L. Blankenbecler. Phys. A16 (1983) 4177. Mukhi. C. B132 (1983) 69. Salomonson and J. B234 (1984) 269. P. Phys. e [26] L.N. Zannetti. Sodano and M. Nucl.S. Keung. Takeda. Rev. Murthy. [22] L. Ui. Boyanovsky and R. de Crumbrugghe and V. [23] A. Sukhatme. Urrutia and E.[20] D. [31] M. Windey. Rev. 72 (1984) 192. Phys. Jour. Friedan and P. Lett. Witten. Brown. 52 (1984) 1933. Phys. AlvarezGaum´. [21] B. Ui and A.Y. Parisi and N. [30] P. Nucl. Bonn Summer School Lectures (1984). Phys.V. Rev. [34] W. [24] R. Phys.S. 60 (1988) 41. AlvarezGaum´ and E. 61 (1988) 1143. Phys. ibid. 151 (1983) 99. A. RuizAltaba and B. Phys.F. e [28] D. e [27] L. Freedman and F. AlvarezGaum´. Theor.K. [35] M. Imbimbo and S. Skagerstam. Phys. Rev. P. [25] L. Lett. Nucl. [29] G. Thomas. van Holten. H. Rev. B206 (1982) 321.
Sukhatme. 143 . [40] A. Kostelecky and M.A. Phys. Phys. Andrianov. Rev. A129 (1988) 145.N. and Eng. Khare and U. B181 (1986) 141. Levai. Khare and J. Phys. Riccati Diﬀerential Equations. in Sci. N. Stedman. Jour. Jour. Nieto. Andrianov. [46] J. [49] E. o [50] W. Phys. Jour. Duke Univ. Rev. N. A21 (1988) L501. A21 (1988) L195. preprint (1989). M. Phys. Wipf. Proc. A. A32 (1985) 1293. A.B. Jour. Lett. R. Lett. B181 (1986) 295. [45] R. Phys. Reid. Chuan. Maharana. Lett. Sukhatme. Phys. A 24 (1991) L1165. Euro. Khare and U. Lett. Rev. Cooper. Lett. Borisov and M. Ginocchio and A. Rev. J. Phys. Phys. Phys. Phys. Phys. Vestnik Leningrad Univ. Mat. Blockley and G. Jour.[38] A. Natanzon. 164 (1985) 277. 38 (1979) 146. 56 (1986) 96. Dutt. Phys. (1972). A. Eides and M. [52] F. [43] C. Phys. [41] A. N. Irish Acad. Lett.M. JETP Lett. Schr¨dinger. Phys. Khare and U. Sukhatme. Balantekin.N. Rau. Academic Press.T. Nucl. Teoret. Dabrowska.. Roy. Sukhatme. A105 (1984) 19. Cooper. [47] F. Lett. [55] G. A22 (1989) 689. Math. Rev. Gendenshtein. Am.3243. Borisov. Phys. Stahlhofen. J. Ioﬀe. 38 (1983) 356. 6 (1985) 218. Fiz. Ioﬀe.Y. 56 (1986) 95. Jour. A. 56 (1988) 163. [51] A. Khare. P. B244 (1984) 409. A46 (1940) 9. [39] A. [42] A. D36 (1987) 2458. [53] C. Ann. Ginocchio and A. Lett. 53 (1984) 2285. A109 (1985) 143. Phys. [54] R. [48] A. 10 (1971) 22. [56] G. [44] L. Dutt. Khare and U. Phys.
Moses. Phys. Phys. Abraham and H. R. Sukhatme. Spiridonov. C. Academy Sc.R. 8 (1992) 303. Kwong and J. [67] K. Prog. 2267. M. Pursey. Ginocchio. Phys. Supp. Khare and U. [60] A. Sukumar. ibid.V. Sukhatme. Ann. Barclay. Phys. Wang. [66] D.N. [70] A. Khare and U. Q. Lett. Phys. Phys. Chadan and P. Pagnamenta and U. 1103. B145 (1984) 208. Keung and T. Phys. [61] V. Phys. Keung. Theor. A20 (1987) 2461. Darboux. (Paris) 94 (1882) 1456. Jour. Phys. Pursey. [72] W. Inv. Phys. [69] C. A5 (1990) 525. A22 (1989) 2847. Rosner. Sabatier. Degasperis and A. [71] A. Phys.C. A22 (1989) 987. Rev. Dutt. [58] A. Jour. Univ. [59] D. Springer Verlag (1977). A19 (1986) 2297. Spiridonov. Phys. Wang and T. Shabat. ibid. D33 (1986) 431.Y. A40 (1989) 6185. [73] W. U. A48 (1993) 2786. 86 (1986) 366. U. W. Jour. Rev. A22 (1980) 1333. Sukhatme. Imbo. [65] P. Lett. Jour. [68] M. Sukumar.Y. Univ. Inverse Problems in Quantum Scattering Theory. Phys. Lett. 159 (1985) 467. A18 (1985) 2917. [75] Q. A. Gangopadhyaya. Luban and D. 144 . A26 (1993) L901. Khare and U. 919 (1992) [64] G.V. A. [63] A. D33 (1986) 1098. [62] V. Rev. Jour. ibid. Shabat. Rev. Imbo. Phys. Sukhatme. Sukhatme.[57] J.M. 69 (1992) 298. 152 (1984) 203. Mod. Nieto. Khare. de Roma preprint No. de Montreal preprint UdeMLPNTH9492. Sukhatme. Prob. 2937. Rev. A. [74] C.
R. [80] A. A157 (1991) 351.M. 145 . B188 (1987) 351. A. Phys. Rev. Fricke. D39 (1989) 1242. Jour.[76] W. DeLaney and M. Phys. Barclay and C. [84] K. A38 (1986) 1679. Khare. Imbo and U. R. Phys. Uzer. [86] R. 54 (1985) 2184. Phys. Rev. Lett. B161 (1985) 131. [82] R. Phys. [78] F. Dutt. A136 (1989) 455. A. Phys. Jour. B150 (1985) 159. Phys. R. Lett. Lett. [85] D.P. Dutt. Khare and U. Phys. Phys. Varshni. [92] Y. A123 (1987) 375. M. Phys. Lett. Lett. Murayama. Lett. Phys. A. [91] S. Jour.P. Varshni. Balantekin. Phys. Kwong. Dutt. Phys. [77] T. Comtet. [88] D.B. A37 (1988) 1686. Maxwell. J. 59 (1991) 723. Lett. A. R. Riggs. A142 (1989) 1. Lett. Raghunathan. Phys.B. Sukhatme. Am. Phys. [81] For a good review on this ﬁeld and references therein see. Lett. Dutt. Seetharaman and S. A. A25 (1992) 5761. Bandrauk and D. [89] Y. A21 (1988) 1589. B247 (1990) 301. Hatchell and T. Campbell. Cooper and P. Roy. Adhikari. Comtet. Nieto. Rosner and H. Varshni. Vasan.K. Varshni. Rev. A152 (1991) 381. [90] A. Khare and Y. Adhikari and A.P. Khare and U.J. [87] P. Thacker.P. [83] R. Sukhatme. Khare and Y. Roy. A143 (1990) 202. Phys. De. R. P. Lett. Lett. Sukhatme. Roychoudhury and Y. Rev. Phys. H. [79] A.
[97] A. Mishra and G. Phys. G. Lett. Junker and A. editors: J. Int. Lett. A8 (1993) 1245. Spiridonov. A. D47 (1993) 4796. [101] F. Dutt. Inomata. A. R. Khare. [98] A. Jayannavar and A. [103] C. [94] A. Phys. Pagnamenta. Rev. Sukhatme. Phys. Phys. [105] H. Suparmi. Cerdeira et al. [106] V. World Scientiﬁc. Phys. A. Pagnamenta and U. A47 (1993) 1708. Gangopadhyaya. Science Press. Phys. Thaller. A25 (1992) L749. A. Lett. Lett. A. Toyama. Phys. Lectures on Path Integration. Junker. Rev. Jour. A151 (1990) 7. [104] E. Mod. Huang. Topics in Quantum Mechanics. [96] R. 187 (1988) 1. [95] R. [108] A. D41 (1990) 3241. Math. Liang et al. 146 . Khare. Beijing (1992). [107] A. [102] Y. A26 (1993) 2261. Jour. Lett. [100] B. A48 (1993) 1845. Phys. Mod. Khare. Sukhatme. [99] A. Jour. A8 (1993) 107. Jour. Nogami and F. Sukumar. Khare. A18 (1985) L697. A174 (1993) 363. editors: H. Rubakov and V. A. Gangopadhyaya. Springer Verlag (1992). Rajasekaran. [109] A. Phys. A. Cooper.V. Vinet. Mishra and G. 34 (1993) 1277. Khare. Symposium on Adv. Inomata and G. and A. D29 (1984) 2975. Musto. Khare. A. A. Phys. Khare. Rev. Phys. Jour. Sukhatme and A. Mod. Das and W. Rev.Q. A3 (1993) 1337. Ann. Phys. Dutt. Phys. Phys.K. Rajasekaran.K. B137 (1984) 72. The Dirac Equation. Singapore (1992). Pagnamenta and U.[93] U. Yamagishi. Rev. D’Hoker and L. Phys. Wipf. Phys.
Roy and R. Vainshtein.[110] A.V. Mod. Dutt and U. Euro. U. Prentice Hall. Varshni. Am. Wigner. A178 (1993) 251. Rev. Jour. Nonrelativistic Quantum Mechanics 3rd Volume. von Neumann and E. Khare and U. A21 (1988) 3673. Jour. Quantum Mechanics. [126] L. Nuovo Cimento A79 (1984) 28. P. B249 (1988) 79. Roy and B. De. Nucl. [124] O. [115] J. A25 (1992) L843. Physics News (India) 22 (1991) 35. B146 (1989) 55. Phys.Roychoudhury and Y. Phys. Phys. Gendenshtein and I. Rev.P. Jour. Bagchi. 6 (1985) 225. Sukhatme. Sukhatme and A. Haymaker and A. [120] R. Stedman. 30 (1929) 465. Phys. Oxford Univ. Sov. [112] J. Zeit. A48 (1993) 3525. [119] L. 54 (1986) 928. Lett. Krive. Casahorran and S.R. Nam. [127] G. Lifshitz. Mod. A11 (1975) 446. Casahorran. Jevicki and J. Phys. Sukhatme. [123] A. J. Int. Jour. Landau and E. 28 (1985) 695. Panigrahi and U. Phys. Roychoudhury.P. Pagnamenta. R. Phys. R. Herrick. A23 (1990) L881. (1977). [117] F. Lahiri. Phys. [118] L. Usp. Phys. Shifman. [114] P. D32 (1985) 1597. Phys. Rodrigues. Phys. [122] G. de Lange and R.E. N. Operator Methods in Quantum Mechanics. Stillinger and D. 205 (1990). Smilga and A. A5 (1990) 1383. Rev. Phys. 147 . B156 (1991) 1925: P. Lett. Press (1991). A6 (1991) 2729. Jour. [116] J. Phys. Lett. Pappademos. Junker. Jour. [113] P. Rau.E.Y. Phys. [121] D. Jour. Lancaster. [111] M. Sukhatme. Raab. Ballentine. Phys. PP. Roy. [128] R. A. Pergamon Press. [125] A. Phys. Int.
Shastry. Nieto. Biedenharn. Lett. Elements of Soliton Theory. Bhubaneswar preprint IPBBSR/9365. D. Jour. Rev. Phys. Calogero. Lett. Rev. Jour.D. Panigrahi and U. Chapter 5. Phys. Rev. Inst. Math. Simmons. Phys. Singapore (1989). [137] A.Truax. Calogero. P. Jour.C. [131] A.M.M. 12 (1971) 419. Nieto and L. Math. Bhaduri. Phys. 12 (1971) 246. 60 (1988) 639. A22 (1989) 4581. Vol. Macfarlane.M. Mod. Jour. of Physics.ibid. 47 (1979) 1067. Phys. Morse and H. World Scientiﬁc. Phys. M. 10 (1969) 2191.L.[129] A. Kostelecky. [147] B. Rep.M. M. Khare and R. Jour. A40 (1989) 2170 [135] V. 47 (1979) 634. Phys. Phys. Phys.M. 60 (1988) 635. [140] See for example. Perelomov. N. Methods of Theoretical Physics.I. 94 (1983) 6. [138] G. [144] F. hepth/9310104.K. [141] F. ibid. Lamb. Bhaduri. Math. Integrable Models. Spiridonov. Montemayor and L. [143] B. McGraw Hill. Phys. A27 (1994) in press. [145] For a list of solvable potentials and other references in the ﬁeld see M. A7 (1992) 1241. Gangopadhyaya. Rev.Y. [136] L. Phys. Wiley. [132] A. [134] R. Phys. Feshbach. [139] A. Nieto. Am. Phys. Sutherland. A22 (1989) L873. Das. Lett. of Montreal preprint UdeMLPNTH93172 (1993). [133] M. 71 (1981) 314. N. (1980). P. 15 (1974) 1420. Salem. Sukhatme. D32 (1985) 2627. Phys. Wolfes.K. [130] V. Khare and R. Jour. 148 .A.Y. Math. (1953). Jour. [146] F. Jour. Univ. Olshanetsky and A.M. Haldane. [142] J.
Rev. Phys. Miura. Rev. Amsterdam. Jour. Raby. [166] S. A17 (1984) 2485. [149] A. Kruskal and R. [165] S. D16 (1977) 1802. [152] J. Math. [151] N. (1982). Rev. [162] F. Phys. Jour. Phys. 2493. [156] S. Krieger and C. Peng. Gardner. Polychronakos. A26 (1993) L473. Chern and C. Frahm. A19 (1986) 1775. Freedman and P. The Inverse Scattering Transformation and the Theory of Solitons. 164 (1967) 171. Spectral Transforms and Solitons. Vasan. Manuscr. Rev. Phys. Unpublished Notes and Lectures. M. Phys. 27 (1974) 97. Barclay. Greene. Eckhaus and A. Froman and P. F. Los Alamos (1981). Miller and R. Gildner and A. Phys. Phys. Rosenzweig. Nucl. Mazenko. ibid. A18 (1985) 1041. [160] W. A183 (1993) 263. [163] N. Phys. Pure App. Degasperis.[148] H. Rev. 28 (1979) 207. Lett. Amsterdam (1981). A. [161] S. Jour. Phys. [164] E. Phys. Chaturvedi and R. J. Z. [157] D. [154] D. Contributions to the Theory. Jr. Rev. B. JWKB Approximation. Baye. Physica 15D (1985) 115. Good. Seetharaman and S.. B344 (1990) 317. Math. [150] D. Rev. 91 (1953) 174. Sourlas. Khare and U. Froman. Calogero and A. North Holland. Mende. Raghunathan. Amsterdam (1965). Lett. B11 (1975) 4077. 47 (1967) 2942. NorthHolland. Ma and G. Phys. Patrascioiu. Comm. 69 (1992) 703. NorthHolland. Chem. [159] C. Lett. [153] M. Sukhatme. 58 (1987) 2738. [158] R. Phys. Jour. Phys. O. [155] S. A37 (1988) 2277. Van Harten. 149 . Amado.
[185] T. K. Rev. [174] G. 46 (1981) 388. van Kampen. [176] W. 150 . Caldeira and A. D11 (1975) 1521. ibid. [170] A. Jour. Panigrahi and U. Gangopadhyaya. Phys. [175] S. [177] N. Merzbacher. Rev. B376 (1992) 599. Aspects of Symmetry. Leggett. 62 (1975) 1899. [172] A. Rev. Laidler and H. Wiley. Lett. Reuter and W. Lett. Boyanovsky. Jour. Rev. Gozzi. A47 (1993) 2720. Press. Glasstone. M. N. Aﬄeck. 52 (1984) 5. 49 (1982) 681. [171] S. (N. Lett. Willey and R. Bender et al. o [181] A. Thacker. Zeit. Rep. Phys. Chakravarty and A.[167] A. McGrawHill. Rev. Phys. [179] D. Bray and M. [184] F.Y.Y. 52 (1984) 140.. Phys. PrenticeHall. Lett. Phys. Lett. Phys. Am. (1941). B78 (1990) 63. Sukhatme. Phys. Sch¨nhammer. Miller. Leggett. [182] S. Holman. Coleman. Imbo and U. D37 (1988) 1472. Chem. Quantum Mechanics. Eyring. Phys. Englewood Cliﬀs (1976). (1970). [183] E. Phys. [178] I. R. Chakravarty. Nucl. Phys. Salam and J. Cambridge Univ. Rev. Moore. 46 (1981) 211. Advanced Calculus for Applications.Y. Zaikin. N. Phys. [180] K. Ann. P. A 183 (1993) 29/ [169] C. 61 (1974) 1823.) 149 (1983) 374. Phys. Lett. Rev. [173] S. Phys. [168] E. Sukhatme.Y. Rev. Stat. Theory of Rate Processes. Schon and A. Strathdee. Hildebrand. (1985). Phys. 49 (1982) 1545. Jour. 17 (1977) 71. N. Phys. 198 (1990) 237. Phys.
Varshni. G. Imbo. [191] R. Rev. Ader. [188] C. Mlodinow and N. Bender. [187] E. Lett. Jour. [198] T. [202] G.. [195] S. 128 (1980) 314. [192] L.P.[186] L. Phys. [193] M. Math. A97 (1983) 178. Rev.Y. Phys. DuttaRoy. (1980). Rev. Sukhatme. Lett. U. A12 (1979) 759. Brezin. Shatz. Phys. A4 (1971) 1875. Phys. A30 (1984) 594. Sukhatme. [197] T. SinhaRoy. Lam and Y. N. 131 (1981) 1. Papanicolaou. Rev. Miramontes and C. [190] J. Rev. Yaﬀe. A17 (1984) L687. Rev. Papanicolaou. Hikami and E. D31 (1985) 2655.P. Jour. Stanciu. L.N. Imbo and U. Pagnamenta and U. Phys. A25 (1982) 1305. B18 (1985) 3311. Rev. Phys. A8 (1993) 523. Kumar. 151 . Ann. Sukhatme. DuttaRoy. Phys. Math. G. Lett. ibid. Phys. Mukherji and Y. Imbo. D28 (1983) 418. Phys. Phys. Plenum. Phys. [201] C.’ tHooft et al. Pajares. 25 (1984) 943. A. Phys. 54 (1982) 407. [203] A. ed. Phys. A. R. Varshni. Jour. Witten in Recent Developments in Gauge Theories. [200] R. S. D29 (1984) 1669. Imbo. Gangopadhyay and B. Jour. [194] J. Mlodinow and N. Ghosh and B. Mod. A105 (1984) 183. Nuovo Cimento 84 (1984) 10. Pagnamenta and U. Phys. [196] U. Dutt. Sukhatme and T. 8 (1967) 2043. [189] L. Mlodinow and M. [199] T. Jour. Phys. Phys. Mod. Gangopadhyay. Khare and C.
Rev. [206] R. GellMann. Mod. Bjorken and S. 43 (1971) 36. D34 (1986) 1100. [213] L. Fujikawa. Phys. 16 (1975) 1393. Phys. [214] R. [218] S. Lett. Phys. di Vecchia.[204] Y. Rev. Land and R. Rev. [212] R. Rev. Nogami and Z. Imbo and U. Bishop. Rev.Y. 152 . Phys. Phys.M. [217] K. Phys. Wipf. Relativistic Quantum Mechanics. Inc. Olive. (N. Nucl.M. Phys. Phys.A. Jackiw and J. 109 (1958) 193. Kostelecky and M. Witten and D. [209] A. [208] E. Phys. Rajeev. B78 (1978) 97. Phys. V. B171 (1986) 226. [216] R. Casher. Schriﬀer. A19 (1978) 2461. Phys. Rev. Rebbi. B190 (1981) 253. Phys..R. Jackiw and C. Interscience Publishers. Ann.M. [223] L. L. Brown. Aharonov and A. Lathouwers. Zhao. A47 (1993) 897. D33 (1986) 3147.G. B175 (1986) 433. Toyama. [221] W. Y. Rev. 82 (1951) 664. Phys. Nucl. Rev. B200 (1982) 293. Boulder. 111 (1958) 193. D. Rev. Drell. Schwinger. McGraw Hill (1964) Chapter 4. O’Raifeartaigh and A. Sukhatme. Acta Physica Austriaca 40 (1974) 306. Jackiw. [219] J. Hughes. Lett.Y. [222] H. Phys. Lett. in Lectures in Theoretical Physics. Musto. [220] T. Phys. Phys. Phys. [207] D. Phys. Math. Frank. Campbell and A. 4 (1962) 324. Phys. Feynman and M. Jour. D21 (1987) 2848. B73 (1978) 162. [205] R.K. Narnhofer. N. [210] F. [215] R. Lett.) 173 (1987) 249. Rev. Nieto. Rev. Spector. d’Adda and P. D13 (1976) 3358. D29 (1984) 2375. [211] J.
Series. B218 (1989) 12. B138 (1965) 1155. [240] V. Kamaefuchi. 153 . B254 (1991) 1. [230] J. [234] D. Lett. A4 (1989) 1209. Messiah. Lett. edited by M. Abramowitz. Phys. Press (1982). Vinet. A24 (1991) L529. [239] S. Rev. Cowan et al. Rev. Teor. Phys.. P. Green. Robnik. Phys. Phys. Lett. Math. Phys. 2289. Phys. A23 (1990) L751.. ibid. Vinet. Mayrand and L. Nuovo Cimento B37 (1977) 1. Phys. Volkov.. Spence and J. Capasso et al. [228] Handbook of Mathematical Functions. Phys. National Bureau of Standards Applied Mathematics Series. A23 (1990) 3661. Phys. Irene A. B233 (1989) 158 . 50 (1982) 354. 90 (1953) 270. A12 (1979) 1175. National Bureau of Standards Applied Math. Jour. Eksp. [237] F.. [233] H. Zh. [231] T. Floreanini. Phys. [226] M. B340 (1990) 777. M. Ardalan and F. Mansouri. [229] Tables of Coulomb Wave Functions. Jour. S. Stegun. Lett. Phys. Am. R. Durand and L. [227] F. Jour. [225] H. Vol.[224] N. 55 (1964). Greenberg and A. Phys. Rev. Nr. [235] O. 17 (1952). D9 (1974) 334. Phys. Phys. Spiridonov. Friedrich and D. Leinaas and J. Vary. Salabura et al. Nature 358 (1992) 5656. Durand. [238] J. Phys. Debergh. Myrheim. Jour. Phys. Phys. Jour. A31 (1990) 1523 . A32 (1985) 3231. Ohnuki and S. Wintgen. MeyerVernet. Beckers and N. Rev. Fiz 38 (1960) 519. [236] Y. Jour. W. L1073 . Mod. Nucl. B245 (1990) 153. 56 (1986) 444. [232] J. 19 (1986) 3845. Koenig et al. Lett. Tokyo Univ. Lett. Quantum Field Theory and Parastatistics. Rev.
J. Vinet. Ph. Merkel. Rev. [251] K. 148 (1983) 336. Sov. Spiridonov and L. Kumar and A. Phys. V. Lett. Greenberg. A5 (1990) 2555. Moscow (1965). Turbiner. Nauka. F. [244] A. M. A. [247] F. Rev. D. Physica D15 (1985) 129 and references therein. Phys. of Phys.[241] A. Lett. Ioﬀe. 32 (1983) 53. Atiyah and I. Nambu. ibid. Brezin. Shapir. R. Ioﬀe. Phys. Efetov. C. [255] M. [245] O. Casten. 64 (1990) 705. Nucl.N. M. Gursey and F. Soc. Jour. ibid. Lett. [246] F. [250] Y. Ann. [253] M. Phys. Lett. Phys. Andrianov. Iachello. unpublished. Lett. Rev. Phys. Rajasekaran. Shifman. V. JETP 67 (1988) 230 . B255 (1991) 543 . Phys. A6 (1991) 199. Mod. Phys. A4 (1989) 2897. Lett. 37 (1991) 455(E). [257] D. A126 (1987) 181. 50 (1983) 873. ibid. Mod.) 36 (1991) 537. Physica D15 (1985) 147 and references therein. Yale Univ. [252] E. 154 . Turbiner and A. Phys. Pramana (Jour. [254] Y. (1985). Phys. Math. A. Phys. Phys. thesis. A6 (1991) 3163. Phy. Usheridze.M. Comm. Berezin. [243] S. Math. Alhassid. Physica D15 (1985) 85 and references therein. Adv. Khare. Itzykson. Mishra and G. Spiridonov and L. Int. Bull. Gross and C. Wu. [249] Y.D. [256] A. 69 (1963) 422. Andrianov and M. A142 (1989) 200. Phys. Vinet. Sov. [248] J. B235 (1984) 24.K. Cardy. [242] V. Am. Physica D15 (1985) 99. The Method of Second Quantization. Lett. 118 (1988) 467. Physica D15 (1985) 123 and references therein. Lett. 44 (1980) 772. Mayrand.P. Jatkar. Mod. Singer. JETP 55 (1985) 514. Durand. Phys. B272 (1991) 297. Iachello.
N. ibid. Mukherjee. Phys. 155 . Abbott and W. Phys. Jour. 69 (1992) 164.A. Lett. [259] A.B. C20 (1983) 227. Z.S. Dedonder. Cannata and J. D.P. Man’ko.I. Phys. Phys. Aizawa. Rev. C3 (1984) 191. Maharana. F. Shastry. A48 (1993) 951. Phys. Phys. M. T. Lett. Rev. [264] V. [261] T. A184 (1993) in press. Khare and J. Sutherland. Fukui. Dutt and Y. Phys. A.S. Nieto. V. [263] T. Fukui and N. Kostelecky. 67 (1991) 2777. Rev. [262] C.[258] R. [260] B. Zakrzewski. Kumar.J. [265] R. Phys. Sil.M. B. and D. Kyoto preprint (1993). Bijker. Truax. Lett. Zeit. R. Yukawa Institute. Lett. R. Varshni. Shastry and B. A180 (1993) 308. A20 (1989) 5347. 70 (1993) 4029. Amado.
1: The energy levels of two supersymmetric partner potentials V1 and V2 . The ﬁgure corresponds to unbroken SUSY.. The energy levels of V1 (x) are clearly marked. The ¯ ground state of the inﬁnite square well has energy 1. Figure 4.O. The choice of units is h = 2m = 1. The action of the operators A and A† in connecting eigenfunctions is shown. The energy levels are (1) degenerate except that V1 has an extra state at zero energy E0 = 0.. Figure 5. Note the degenerate higher energy levels at energies 22 . The diagram is from ref.3: A double well potential V1 (x) (solid line) and its single well supersymmetric partner V2 (x) (dotted line). 42 . Note that only the range x ≥ 0 is plotted since the superpotentials are antisymmetric W (x) = −W (−x). Figure 4. and cover ¯ 156 .2: The inﬁnite square well potential V = 0 of width L = π and its supersymmetric partner potential 2cosec2 x in units h = 2m = 1.1 . [129]. 32 . 4. These two types are related by suitable limiting procedures.FIGURE CAPTIONS Figure 2. .1 are interrelated by point canonical coordinate transformations. The curves are taken from ref. (harmonic oscillator) corresponds to the limiting case of q = 1.1: Selected members of the family of potentials with energy spectra identical to the one dimensional harmonic oscillator with ω = 2.1: Selfsimilar superpotentials W (x) for various values of the deformation parameter q.1: Diagram showing how all the shape invariant potentials of Table 4. Figure 4. The curves are labeled by the value λ1 . Note that these two potentials are shape invariant [59] with a scaling change of parameters. Potentials on the outer hexagon have eigenfunctions which are hypergeometric functions whereas those on the inner triangle have eigenfunctions which are conﬂuent hypergeometric functions.2: Selfsimilar potentials V1 (x) (symmetric about x = 0) corresponding to the superpotentials shown in Fig. Figure 7. The curve labeled H. [59]. Figure 2.
λ2 ) family of potentials isospectral to V1 (x).1: A “deep” symmetric double well potential V1 (x) with minima at x = ±x0 and its supersymmetric partner potential V2 (x). E3 = −16/25. This solution results from constructing the isospectral potential family starting from a reﬂectionless. (392).3: The energy splitting t = E1 − E0 as a function of the separation 2x0 of the superposed Gaussians in the ground state wave function ψ0 (x). The curve λ1 = ∞ is the one dimensional harmonic oscillator. For a detailed discussion. [34].3: A schematic diagram showing how SUSY transformations are used for deleting the two lowest states of a potential V1 (x) and then reinserting them. The curve marked λ1 = 0 is known as the Pursey potential [66] and has one bound state less than the oscillator.1. taken from ref. The energy levels for both potentials are shown. Figure 7.the range 0 < λ1 ≤ ∞.1: The harmonic oscillator potential V1(1) and its partner potential V2(1) as given by eq. see the text and ref. 7.1 . Figure 7. and the partial degeneracy of the eigenvalue spectra coming from the even parity of the potentials [114]. Figure 12. This ﬁgure. Notice the negative energy state of V1(1) at energy −2. shows the remarkable accuracy of the SUSYbased energy splitting computations for a double well potential. ǫ = E1 in Table 12. [34]. 157 . Figure 7. Figure 9. E2 = −1. [75]. symmetric potential with bound states at energies E1 = −25/16. except the Pursey potential. Figure 9.2: Ground state wave functions for all the potentials shown in Fig. Figure 9. (488). Further details are given in the text and ref.4: The pure threesoliton solution of the KdV equation as a function of position (x) and time (t). thus producing a twoparameter (λ1 .2: Supersymmetric partner potentials V1 (x) and V2 (x) corresponding to two choices of the parameter x0 for the potentials given in eq. These potentials arise from a singular superpotential W1 = x − 1/x corresponding to the choice φ = ψ1 (x).
Fig. Unless otherwise stated.3: Potentials V (r) (solid lines) possessing one bound state in the continuum (BIC) obtained by starting from a free particle and constructing a oneparameter (λ) family. the full wave function is ψnlm (r. (b) The wave functions at energies E0 = 1 (dashed line) and E1 = 4 (dotted line) [115]. (494)] coming from the singular superpotential W1 given in eq.Figure 12. (b) A plot of the wave functions corresponding to the potentials in part (a). (493). ˆ and is shown by the dashed line.5:(a) An example of a potential with two BICs at energies E0 = 1 and E1 = 4.5 breaks V2(1) into two disjoint pieces. φ).4: (a) A potential (solid line) with one BIC at energy E = 0. Fig. Note that the wave functions for the ﬁrst four potentials (Hermite and Laguerre polynomials) are special cases of the conﬂuent hypergeometric function while the rest (Jacobi polynomials) are special cases of the hypergeometric function.5 and ﬁg. 0 ≤ r ≤ ∞. ˆ Figure 12. 5.1: All known shape invariant potentials in which the parameters a2 and a1 are related by a translation (a2 = a1 + α). α. The energy eigenvalues and eigenfunctions are given in units h = 2m = 1.25 obtained by starting from an attractive Coulomb potential (dotte d line). shows the interrelations between all the SIPs in the table via point canonical coordinate transformations. Figure 12. For spherically symmetric potentials. and the eigenvalue spectra have no degeneracy [114]. TABLE CAPTIONS Table 4.2: The Morse and its supersymmetric partner potential [eq. The BIC wave function u0 (r) is at energy E = 1. φ) = ψnl (r)Ylm (θ. 158 . taken from ref. (b) corresponds to a much larger value λ = 5 . θ. [129]. the range of potentials is −∞ ≤ x ≤ ∞. l are ¯ all taken ≥ 0. The singularity at x = −ln 3. Figure 12. B. ω. The constants A. (a) corresponds to λ = 0.1.
Table 6. 159 . (480) determines the choice of the superpotential Wφ which produces any ˜ given nonsingular potential V (x) [eq. The table shows how certain choices of ǫ give rise to singular superpotentials. We have taken φ(x = −∞) = 0 for convenience [114]. negative energy eigenstates and a breakdown of the degeneracy theorem. Units corresponding to h = 2m = 1 are used ¯ throughout. ν and several values of the quantum number n. These families are generated by applying the theorem described in the text to scattering states u0 and u1 at energies E0 and E1 in succession [115].1: Comparison of the lowest order WKB and SWKB predictions for the bound state spectrum of the Ginocchio potential for diﬀerent values of the parameters λ.1: Comparison of the three lowest energy eigenvalues obtained by a variational method with the exact results.1: Choice of the constant ǫ and the corresponding solution φ(x) in eq.2: One and two parameter families of potentials with bound states in the continuum. Table 12. Table 9. Table 12. (481)]. The exact answer is also given.
org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.This figure "fig11.
png" is available in "png" format from: http://arXiv.This figure "fig21.org/ps/hepth/9405029v2 .
org/ps/hepth/9405029v2 .This figure "fig31.png" is available in "png" format from: http://arXiv.
png" is available in "png" format from: http://arXiv.org/ps/hepth/9405029v2 .This figure "fig41.
This figure "fig51.org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.
This figure "fig61.org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.
png" is available in "png" format from: http://arXiv.org/ps/hepth/9405029v2 .This figure "fig71.
png" is available in "png" format from: http://arXiv.org/ps/hepth/9405029v2 .This figure "fig12.
png" is available in "png" format from: http://arXiv.This figure "fig22.org/ps/hepth/9405029v2 .
png" is available in "png" format from: http://arXiv.org/ps/hepth/9405029v2 .This figure "fig32.
org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.This figure "fig42.
png" is available in "png" format from: http://arXiv.This figure "fig52.org/ps/hepth/9405029v2 .
org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.This figure "fig62.
org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.This figure "fig72.
This figure "fig13.png" is available in "png" format from: http://arXiv.org/ps/hepth/9405029v2 .
This figure "fig23.org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.
org/ps/hepth/9405029v2 .png" is available in "png" format from: http://arXiv.This figure "fig33.
png" is available in "png" format from: http://arXiv.This figure "fig43.org/ps/hepth/9405029v2 .
png" is available in "png" format from: http://arXiv.This figure "fig53.org/ps/hepth/9405029v2 .
png" is available in "png" format from: http://arXiv.org/ps/hepth/9405029v2 .This figure "fig63.
org/ps/hepth/9405029v2 .This figure "fig73.png" is available in "png" format from: http://arXiv.