Journal of Physics: Conference Series

PAPER • OPEN ACCESS

Symmetry Analysis of the Square Well Potential

To cite this article: E Suárez et al 2023 J. Phys.: Conf. Ser. 2448 012008

View the article online for updates and enhancements.

This content was downloaded from IP address 189.203.118.57 on 25/02/2023 at 21:33

Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

Symmetry Analysis of the Square Well Potential

E Suárez1 , R D Santiago-Acosta2 and R Lemus1
1
Instituto de Ciencias Nucleares, UNAM
A. P. 70-543, Circuito Exterior, C. U., 04510 Mexico, CDMX, Mexico
2
Tecnológico de Monterrey, Escuela de Ingenierı́a y Ciencias, Carr. al Lago de Guadalupe Km.
3.5, 52926 Edo. de Mex., Mexico
E-mail: [email protected]

Abstract. Symmetry considerations are taken into account when a particle in a square well
potential is studied. This system may display natural degeneracy, accidental degeneracy or
systematic accidental degeneracy depending on the depth of the potential. In order to obtain
the solutions associated with an arbitrary potential an algebraic discrete variable representation
approach based on Pöschl-Teller functions is proposed. It is proved that the geometrical group
C4v is the symmetry group of the system for the case of a finite potential barrier. A similar
analysis is carried out for the rectangular square well potential with commensurate sides. In
both cases the symmetry projection is crucial to simplify the calculations.

1. Introduction
Given the solutions of the 1D quantum box, the generalization to 2D and 3D Cartesian infinite
square well potentials is straightforward [1–4]. This is possible as long as an infinite barrier
is considered. The 2D confined particle has only bound states with expected degeneracy
pattern corresponding to the point group C4v associated with the geometry of square, in case
the symmetry group was correctly identified. This situation is present for semirigid molecules
for instance. However in this system a systematic degeneracy appears which means that the
geometrical group is just a subgroup of the true symmetry group. This fact suggests the search
of a new group containing the geometrical group as a subgroup [5]. The 2D confined particle
presents an analogy with the non-relativistic hydrogen atom where the geometrical symmetry
group is SO(3) while the true symmetry corresponds to the SO(4) group [6–8]. For the square
well potential for infinite walls the new group corresponds to the semidirect group G = T ∧ C4v ,
where T stands for a one-dimensional compact continuous group [9]. The construction of its
irreducible representations (irreps) has a close analogy with the construction of irreps for the
space groups [10]. A confined particle in a cubic box presents similar properties [11].
When the potential depth becomes finite the solutions cannot be obtained analytically and
consequently it is not possible to identify the nature of the degeneracy. The analysis of the square
impenetrable well potential has already been presented [9, 11, 12], but results for finite wells has
not been obtained. The reason may be twofold. On one hand the solutions cannot be obtained in
a straightforward way and on the other hand their study requires group representation theory. In
this contribution we present the numerical solutions using a complete basis associated with the
Pöschl-Teller potential. In this scenery we show that the spectrum renders normal degeneracy
in the framework of the symmetry group C4v .

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

A convenient approach to obtain the solutions of piecewise potentials is the discrete variable
representation approach (DVR). This method is based on the use of a discrete basis [13–23], and
it has been independently developed with different names: quadrature discretisation method
[24–29], Lagrange mesh method [30–33] and configuration localised states [34–36]. The basis
of these methods is the use of orthonormal polynomials with the Gaussian quadrature method
to establish a grid. These methods were successfully developed for 2D and 3D cases. In the
present situation a convenient method to consider 2D systems from 1D results, consists in
taking direct products, like in the description of the vibrational degrees of freedom in terms
of internal coordinates [37]. The direct product basis, although simple to be constructed, has
the disadvantage of involving large basis dimensions. This problem has been overcome by using
the Lanczos algorithm [38–40]. In this contribution we simplify our description by using the
machinery of group representation theory.
Recently an alternative DVR methods based on the algebraic description of different systems
have been proposed. One of them is based on the dynamical groups U (n + 1) for nD systems,
closely related with the harmonic oscillator basis. These approaches has been called unitary
group approach (UGA) [41–46]. An additional algebraic DVR approach, called HO-DVR
approach, is derived by taking a subspace of finite dimension. The resulting discrete basis
provides a DVR basis which corresponds to the zeros of the polynomials associated with the
solutions of the harmonic oscillator. Both the U (n + 1)-UGA and HO-DVR methods takes an
harmonic oscillator basis but they provide different results [47, 48].
The harmonic oscillator functions have been proved to represent a useful basis [49], but it may
be not appropriate for some potentials. For 1D systems alternative algebraic DVR approaches
associated with the specific potentials of Morse (M-DVR) and Pöschl-Teller (PT-DVR) have
been proposed [48]. The application depends of the symmetry of the system to be considered,
M-DVR for asymmetric potentials and the PT-DVR for symmetric potentials.
In this work a study of one particle under a 2D square well potential is presented. We
are interested in convergence since piecewise potentials do not provide exponential convergence
[50]. The algebraic DVR approach provides a useful method to study the symmetry breaking
T ∧ C4v ↓ C4v . Our analysis shows that this system exemplifies the concepts of natural and
accidental degeneracy.
This article is organized as follows. In Section 2 a summary of the 1D algebraic DVR approach
is presented. Section 3 is devoted to the theoretical analysis needed to describe a particle in a
square well potential. In Section 4 the results are presented. Finally, in Section 5 conclusions
are given.

2. PT-DVR method
The bound states of the hyperbolic version of the Pöschl-Teller potential do not form a complete
set of states in the Hilbert space [51–53]. A complete basis is provided by [54]
σ
Φσn (u) = Aσn (1 − u2 ) 2 Cnσ−1/2 (u); n = 0, 1, 2, . . . , (1)
σ−1/2
where Cn (u) are the Gegenbauer polynomials [55] with argument u = tanh(αq), σ is a
parameter taken to be σ = 1 in order to decouple the bound from the continuum spectrum [54],
and Aσn is the normalization constant given by
s
αn!(σ + n − 1/2)(Γ[σ − 1/2])2
Aσn = . (2)
π22−2σ Γ[2σ + n − 1]
Using the factorization method it is possible to obtain ladder operators {K̂± , K̂0 } with the
following effect [54]
K̂+ Φσn (u) = k+ Φσn+1 (u), (3)

2
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

K̂− Φσn (u) = k− Φσn−1 (u), (4)

K̂0 Φσn (u) = k0 Φσn (u), (5)

with
q
k+ = (n + 1)(2σ + n − 1), (6)
q
k− = n(2σ + n − 2), (7)
k0 = (n + σ − 1/2). (8)

Based on these results, the momentum and the natural variable u take the following form [54]
ih̄α 1
p̂ = [B̂+ − B̂− ]; û = [Â+ + Â− ]; (9)
2 2
where the new operators are defined by the following action over the basis:
s
(n + 1)(2σ + n − 1)
Â+ Φσn (u) = Φσ (u), (10)
(n + σ − 1/2)(n + σ + 1/2) n+1
s
n(2σ + n − 2)
Â− Φσn (u) = Φσ (u), (11)
(n + σ − 1/2)(n + σ − 3/2) n−1
B̂+ Φσn (u) = (σ + n)Â+ Φσn (u), (12)
B̂− Φσn (u) = (σ + n − 1)Â− Φσn (u). (13)

Using these expressions we can obtain the matrix elements of the coordinate and momentum,
which diagonalization in the subspace
σ
LPT
N = {|Φn ⟩, n = 0, 1, . . . , N − 1}, (14)

gives rise to the eigenvectors

N
X −1
|ui ⟩ = ⟨Φσn |ui ⟩|Φσn ⟩, (15)
n=0
N
X −1
|pi ⟩ = ⟨Φσn |pi ⟩|Φσn ⟩, (16)
n=0

providing the momentum and coordinate representations characterized by

û|ui ⟩ = ui |ui ⟩, p̂|pi ⟩ = pi |pi ⟩. (17)

For the coordinate we have the relation u = tanh(αq) and consequently

⟨uj |V (q)|ui ⟩ = V [qi (ui )]δij , ⟨pj |G(p)|pi ⟩ = G(pi )δij , (18)

with
1
arctanh(ui ).
qi = (19)
α
Hence the factorization method provides the tool to obtain the discrete representation bases (15)
and (16) that allows the DVR approach to be implemented. Let us now consider the general 1D
Hamiltonian
1 2
Ĥ = p̂ + V (q), (20)
2µ

3
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

with p̂ = −ih̄d/dq. Because of the property (18) The matrix representation of a Hamiltonian
takes the form
H = W† Λ(p) W + T† Λ(q) T, (21)
where the diagonal matrices are given by ||Λ(p) || = (p2i /2µ)δij and ||Λ(q) || = V (qi )δij in the
momentum and coordinate representations, respectively. Here the matrices W and T correspond
to the transformation coefficients T = ||⟨Φσn |ui ⟩|| and W = ||⟨Φσn |pi ⟩||, associated with the
coordinate and momentum representations, respectively. In this approach for a given basis
dimension N we obtain the transformation coefficients which allow us to obtain the solutions
for any potential V (q). This approach will be applied to the square well potential.

3. Square well potential

The discussion of the square 2D potential is seldom discussed in textbooks. However it presents
fundamental problem from the point of view of symmetry [9]. The solutions for the confined
case are easly obtained with energies

h̄2 π 2 2
En1 n2 = (n + n22 ). (22)
2µL2 1

This expression presents a degeneracy pattern not explained assuming the symmetry group C4v
because of the presence of accidental degeneracy, as depicted in Figure 1, where the energy units
were taken to be Ēn1 n2 = En1 n2 /(h̄2 π 2 /2µL2 ). The degeneracy is explained by identifying the
symmetry group G = T ∧ C4v , where the apparent accidental degeneracy renders normal [9].
This peculiar behavior is particular to the confined system. When the walls become finite we
shall prove that the degeneracy is either partially or completely removed. However, the general
description of the 2D system is a non trivial task that will be considered in this work using the
algebraic PT-DVR approach.

Figure 1: Symmetry labels of a particle confined in a square well potential. Systematic accidental
degeneracy with irreps (A1 , B1 ) and (A2 , B2 ) has been remarked (see Ref. [9]).

First we notice that this problem can be solved in terms of the direct product of 1D bases
corresponding to the x and y subspaces:

|Φσnx Φσny ⟩ = |Φσnx ⟩ ⊗ |Φσny ⟩. (23)

4
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

In order to generate the DVR basis we have to consider the discrete basis associated with both
the x and y coordinates. For the x component we have
N
X −1
|xi ⟩ = ⟨Φσnx |xi ⟩|Φσnx ⟩, (24)
nx =0
N
X −1
|pxi ⟩ = ⟨Φσnx |pxi ⟩|Φσnx ⟩, (25)
nx =0

while for the y degree of freedom

N
X −1
|yi ⟩ = ⟨Φσny |yi ⟩|Φσny ⟩, (26)
ny =0
N
X −1
|pyi ⟩ = ⟨Φσny |pyi ⟩|Φσny ⟩. (27)
ny =0

Introducing the discrete bases, the direct product (23) takes the form
N X
X N
|Φσnx Φσny ⟩ = ⟨xi |Φσnx ⟩⟨yj |Φσny ⟩ |xi ⟩ ⊗ |yj ⟩. (28)
i=1 j=1

For the sake of convenience we next introduce the following notation for the matrix elements
involved in (28):

Ai,nx = ⟨xi |Φσnx ⟩, (29)

Bj,ny = ⟨yj |Φσny ⟩, (30)
Tij,nx ny = ⟨xi |Φσnx ⟩⟨yj |Φσny ⟩, (31)

with the following relation

T = A ⊗ B. (32)
Following the same procedure for the momentum representation we define

Ci,nx = ⟨pxi |Φσnx ⟩, (33)

Dj,ny = ⟨pyj |Φσny ⟩, (34)
Wij,nx ny = ⟨pxi |Φσnx ⟩⟨pyj |Φσny ⟩, (35)

with
W = C ⊗ D. (36)
In terms direct product basis (23) the Hamiltonian matrix representation takes the form

H = W† Λ(p) W + T† Λ(V ) T, (37)

where the matrix elements of Λ(p) in the momentum representations are

1 2
||Λ(p) || = (p + p2yj )δii′ δjj ′ , (38)
2µ xi
while for the potential in the coordinate representation

||Λ(V ) || = V (xi , yj )δii′ δjj ′ . (39)

5
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

Although the direct product basis is simple, the basis dimension increases as N 2 and it may
be necessary large basis dimension for convergence. In order to simplify our analysis we shall
introduce a symmetry adapted basis. In this scheme the projected basis takes the form
X
|q ϕ(Γ)
γ ⟩= Snx ny ;qΓγ |Φσnx Φσny ⟩, (40)
nx ,ny

where Γ and γ are the irrep and component respectively of the group C4v . The index q = q(Γ)
refers to the multiplicity of the irreps and it provides the dimension of the matrix representation
to be diagonalized. In this context the Hamiltonian matrix has the form

Hs = S† [W† Λ(p) W + T† Λ(V ) T]S, (41)

which is in blocked form: X

Hs = ⊕HΓs , (42)
Γ
where each blocked is characterized by the one irrep of the group. The dimension of each block is
given by q(Γ)2 . It is important to emphasize that beyond the fact that the symmetry projection
considerably simplifies the calculations, it is a compulsory task because we are interested in
identifying the degeneracy associated with the symmetry. We next present the approach we
have followed in order to obtain the functions (40).

3.1. Symmetry adapted basis

Here we present the application of the eigenfunction method to project the direct product basis
(28), which consists in diagonalizing a complete set of commuting operators (CSCO) defined in
terms of classes [56–59].

Figure 2: Square well potential and symmetry elements associated with the geometrical
symmetry C4v . Two reference systems are shown: (a) origin located at the left corner used
to obtain simple analytical solutions for the confined particle and (b) origin located at the
center of the square preferred in this work.

The CSCO used to project the basis is defined in accordance with a suitable canonical
subgroup chain, which in our case has been chosen to be

C4v ⊃ Csd , (43)

where
Csd = {E, σda }, (44)

6
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

with symmetry elements depicted in panel (b) of Figure 2. Assigning the i-th class of the group
C4v as Ki and the classes of the subgroup Csd as kj , we have for the CSCO, called CII , the
following linear combination
Cd
ĈII = ĈIC4v + ĈI s , (45)
where
Cd
ĈIC4v = K̂2 + 3K̂5 ; ĈI s = k̂2 , (46)
with the classes
K̂2 = Ĉ4 + Ĉ43 ; K̂5 = σ̂da + σ̂db ; k̂2 = σ̂da , (47)
in accordance with the standard class numbering in the character tables. The operators ĈIC4v
Cd
and ĈI s distinguish the irreps of the groups C4v and Csd , respectively, while the operator ĈII
distinguishes irreps and components of the group C4v . Their simultaneous diagonalization leads
to the eigenkets |µν⟩ defined by the eigensystem

Cd
ĈIC4v |µν⟩ = µ|µν⟩; ĈI s |µν⟩ = ν|µν⟩, (48)

where µ = µ(Γ) and ν = ν(γ) are functions of the group and the subgroup characters:
(Γ) (Γ)
2χ2 3χ (γ)
µ= + 5 ; ν = χ2 , (49)
nΓ nΓ

where here γ stands for the irrep of the subgroup Csd , nΓ refers to the dimension of the Γ-th
irrep and γ ∈ D(Γ) (C4v ) ↓ Csd . Therefore the symmetry projected functions are equally expressed
either with |Γγ⟩ or |µν⟩.
In order to establish the effect of the operators involved in (45) it is enough to consider the
operators Ĉ4 and σ̂da since they are generators of the group. Formally we start establishing the
effect of these operators over the basis

Ĉ4 |x⟩ = |C4 x⟩ = |y⟩; σ̂da |x⟩ = |σda x⟩ = |y⟩,

Ĉ4 |y⟩ = |C4 y⟩ = | − x⟩; σ̂da |y⟩ = |σda y⟩ = |x⟩. (50)

Hence for the rotation C4 we have

Z
⟨Φσn′x Φσn′y |Ĉ4 |Φσnx Φσny ⟩ = dxdydx′ dy ′ ⟨Φσn′x Φσn′y |x′ y ′ ⟩⟨x′ y ′ |C4 |xy⟩⟨xy|Φσnx Φσny ⟩
Z4
= dxdydx′ dy ′ ⟨Φσn′x Φσn′y |x′ y ′ ⟩⟨xy|Φσnx Φσny ⟩δ(y − x′ )δ(y ′ + x)
4
= (−1)nx δn′y nx δn′x ny . (51)

In similar form we obtain

⟨Φσn′x Φσn′y |Ĉ43 |Φσnx Φσny ⟩ = (−1)ny δn′y nx δn′x ny , (52)

⟨Φσn′x Φσn′y |σ̂da |Φσnx Φσny ⟩ = δn′y nx δn′x ny , (53)
⟨Φσn′x Φσn′y |σ̂db |Φσnx Φσny ⟩ = (−1)(nx +ny ) δn′x ny δn′y nx . (54)

Gathering all this matrix element in the operator ĈII , we obtain

⟨Φσn′x Φσn′y |ĈII |Φσnx Φσny ⟩ = [(−1)nx + (−1)ny + 4 + 3(−1)nx +ny ]δn′x ny δn′y nx . (55)

7
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

This matrix representation is block diagonal with N one-dimensional matrices and a number of
N (N − 1)/2 two-dimensional matrices for N even. Hence a Hamiltonian matrix of dimension
N 2 given by (37) reduces to the diagonalization of five matrices of dimension aΓ in accordance
to the reduction of the representation ∆(red) (CII ):

S−1 ∆(red) (CII )S =

X
⊕aΓ DΓ (CII ). (56)
µ

We now consider this projection to describe the square well potential through the
diagonalization of the representation matrix (41).

4. Results and discussion

In this section we consider the case of a confined particle with exact energies given by (22). The
parameters were chosen to be: a = 4 Å , V0 = 108 cm−1 and µ = 1.672 621 9 × 10−27 kg. In
the PT-DVR method the parameter α is free to be fitted. The results are given in Table 1,
where the criterion for the basis dimension N 2 = 3600 was given by obtaining errors of 1 %.
To evaluate the quality of the wave functions we have projected the eigenvectors to the position
representation. The eigenvectors |ψαΓγ ⟩ of the Hamiltonian (41) are given in terms of a linear
combination of the symmetry adapted basis (40)
X
|ψαΓ,γ ⟩ = Vqα |q ϕ(Γ)
γ ⟩. (57)
q

The wave functions ψαΓ,γ (x, y) obtained by the projection ψαΓ,γ (x, y) = ⟨xy|ψαΓ,γ ⟩, with
X
⟨xy|q ϕ(Γ)
γ ⟩= Sn1 n2 ;qΓγ Φσnx (x)Φσny (y), (58)
n1 ,n2

where Φσnx (x) and Φσny (y) are 1D PT given by (1). In Figure 3 a selected set of eigenstates
has been depicted. At the level of this resolution these wave functions are identical to the ones
obtained by projecting the exact functions

ψn1 n2 (x, y) = ψn1 (x)ψn2 (y), (59)

where r r
2 nx πx 2 ny πy
   
ψnx (x) = sin , ψny (y) = sin , (60)
L L L L
are the solutions taking the origin in accordance with (a) in Figure 2. In order to magnify the
differences between the calculated and exact wave functions in the same Figure 2 at the right
column we have plotted the absolute value of the difference of these wave functions, a result
that gives confidence in our approach.
We now pay attention to the N -dependence of the mean square root deviation displayed in
Figure 4. In this approach the potential is evaluated at the discrete points. From the total
points N 2 some of them, say Nin , fall inside the box. We thus define
Nin
%P = × 100, (61)
N2
corresponding to the percentage of points inside the box. The importance of this quantity is
that % P is correlated to the rms, as shown in the lower panel of Figure 4 with dashed lines.
This is a before remarkable result that provides us with a criterion to select the appropriate
basis dimension in the process of convergence. This is possible because Nin can be obtained

8
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

Table 1: Energies for one particle confined in a 2D square well potential using PT-DVR method.
The fitted parameter was chosen to be α = 4.5 × 109 m−1 . The basis dimension was N 2 = 3600.
In this case Ec is the energy calculated.

Irrep (n1 , n2 ) Exact Ec (cm−1 ) % Error

A1 (1,1) 20.65 20.64 0.04
E (2,1) 51.62 51.48 0.27
B2 (2,2) 82.59 82.32 0.33
A1 (3,1) 103.24 103.20 0.04
B1 (3,1) 103.24 103.20 0.04
E (3,2) 134.22 134.04 0.13
E (4,1) 175.51 175.94 0.24
A1 (3,3) 185.84 185.77 0.04
A2 (4,2) 206.48 205.78 0.34
B2 (4,2) 206.48 205.78 0.34

before the calculation is carried out, since the discrete points are associated with the zeros of
the Gegenbauer polynomials.
We now proceed to consider the square well with a finite potential. In this case the spectrum
is expected to be modified in both energy and degeneracy, as it will be shown.
Our system is characterized by the existence of the degenerate levels (A1 , B1 ) and (A2 , B2 ).
When the potentials walls are finite both couple of levels split, as shown in Figure 5. This
means that the symmetry group becomes the geometrical group C4v , a fact technically given
by the subduction G ↓ C4v , as displayed in the same Figure 5. Notice that the levels (A2 , B2 )
split sooner that the levels (A1 , B1 ), which is explained by the fact that the states {|ϕA B2
42 ⟩, ϕ42 ⟩}
2

have a higher energy. While the splitting of these levels manifest around V0 = 450, the splitting
of the states {|ϕA B1
31 ⟩, ϕ31 ⟩} is clearly shown until V0 = 250. This means that at some values of
1

the potential V0 the systematic accidental degeneracy disappears, although it may happen that
some near-accidental degeneracy remains at the low lying region of the spectrum. To measure
the splitting it is useful the parameter ζ defined as [60, 61]

2 (E1 − E2 )
 
ζ= arctan , (62)
π (E1 + E2 )/2

which measures the relative splitting between the levels E1 and E2 . For complete degeneracy
ζ = 0, while as the splitting increases ζ → 1. From Figure 5 we see that as the potential depth
decreases the first parameter ζA2 B2 starts increasing before ζA1 B1 . It is until V0 = 350 that
ζA1 B1 shows a splitting, establishing the approximate limit of accidental degeneracy. We have
selected the potentials V0 = 108 , 350, 250, 180, to illustrate the order of the induced splitting.
For potential depth V0 = 350 the levels A1 and B1 are nearly degenerate, while the splitting of
the levels A2 and B2 becomes manifest. This pattern was displayed with the parameters V0 and
wide a, but for different parameters lower bound states A1 , B1 may remain degenerate, showing
the existence of an accidental degeneracy.

5. Conclusions
In this work we have studied in detail one particle under a square well potential. The spectrum
for the confined particle is characterized by the presence of systematic accidental degeneracy

9
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

Φ3,1(A1)
2

(a)

y (Å)
0

-1

-2
-2 -1 0 1 2
x (Å)

Φ3,1(B1)
2

(b)
y (Å)

-1

-2
-2 -1 0 1 2
x (Å)

Φ4,1(E, A')
2

(c)
y (Å)

-1

-2
-2 -1 0 1 2
x (Å)

Φ4,1(E, A'')
2

(d)
y (Å)

-1

-2
-2 -1 0 1 2
x (Å)

Figure 3: In the first column selected symmetry projected wave functions are displayed. Exact
and and calculated wave functions provide the same plots. The wave functions are distributed
in accordance with accidental degeneracy with irreps (A1 , B1 ) in (a) and (b), respectively,
while natural degeneracy with components (EA′ , EA′′ ) in (c) and (d), respectively. At the right
column the absolute value of the differences between the exact and calculated wave functions
are displayed.

with respect to the geometrical point group C4v . This degeneracy is explained by identifying
the symmetry group T ∧ C4v Ref. [9]. When the group T ∧ C4v is considered the systematic
accidental degeneracy becomes normal. We show that for finite potential dimension a symmetry
breaking is present, formally described by the subduction T ∧ C4v ↓ C4v . A partial symmetry
breaking may be manifested depending of the depth of the potential and width a. Regarding
the convergence, a correlation exists between the rms and the number of points % P falling
inside the square box. The number of points inside the box can be calculated with the zeros
of Gegenbauer polynomials, which means that the basis dimension can be determined before
carrying out the calculation. This fact compensate the lack of exponential convergence. The
square well potential although simple, it represents a system where all types of the degeneracy
may be present, a feature that may be used to exemplify the concept of symmetry group and
the consequences of not having such group appropriately identified. We stress that during
our analysis the importance of the symmetry projection to identify the degeneracy but also to
simplify the calculations. To accomplish this task we have used the eigenfunction method, a
quite efficient approach particularly useful for this system. Finally we should mention that this

10
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

30

% of points
28

26

24

35 22

30 N 2 = 3600
15
RMS (cm-1)

25

RMS (cm-1)
20 10
15
10 5
5
0
0 90 95 100 105 110 115 120
20 40 60 80 100
N N

Figure 4: At the left panel the root mean square deviation as a function of the basis dimension
N . The basis dimension N 2 is highlighted with an arrow. In the same panel the average is
included taking into account seven point to the right and to the left when possible. In the right
panel a zoom is shown including the percentage of points % P defined in Eq. (61), where it is
clear the correlation with the rms. The basis dimension was chosen to correspond to the first
absolute minimum in rms.

analysis can be applied to the case of a rectangular well potential with commensurate sides. The
qualitative behavior is similar to the square well potential.
Hence with this example we have shown that the algebraic PT-DVR approach is suitable to
deal with piecewise potentials, even though the lack of exponential convergence.

11
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

Figure 5: (a) Correlation diagram of the energy levels as a function of the potential depth for
the levels {A1 , B1 } and {A2 , B2 }. In (b) the associated parameters ζ for these couple of levels
are shown.

Acknowledgments
This work is partially supported by DGAPA-UNAM, Mexico, under project IN-212020.

References
[1] C. Cohen-Tannoudji, B. Diu and F. Laloë 1977 Quantum Mechanics. Vol I. John Wiley and Sons. New York.
[2] F.S. Levin 2002 An Introduction to Quantum Theory. Cambridge University Press. U.K.
[3] W. Greiner 1989 Quantum Mechanics. An introduction. Springer. Frankfurt am Main. Germany.
[4] D.C. Harris and M.D. Bertolucci 1989 Symmetry and Spectroscopy. An Introduction to Vibrational and
Electronic Spectroscopy. Dover Publications, INC., New York.
[5] H.V. MacIntosh 1971 Symmetry and Degeneracy. Group Theory and its Application. Vol.2. Ed. E.M. Loebel
(New York: Academic) pp 75-144.
[6] B.G. Wybourne 1974 Classical Groups for Physicists. New York, Wiley.
[7] V. Fock 1935 Z. Phys. 98 145.
[8] V. Bargmann 1936 Z. Phys. 98 576.
[9] F. Leybraz, A. Frank, R. Lemus and M.V. Andrés 1997 Am. J. Phys. 65 1087.
[10] S.L. Altmann 1977 Induced Representations in Crystals and Molecules. Point, space and non-rigid molecular
groups. New York: Academic.
[11] A.O. Hernández-Castillo and R. Lemus 2013 J. Phys. A: Math. Theor. 46 465201.
[12] R. Lemus, A. Frank, M.V. Andrés and F. Leybraz 1998 Am. J. Phys. 66 629.

12
Quantum Fest 2021 IOP Publishing
Journal of Physics: Conference Series 2448 (2023) 012008 doi:10.1088/1742-6596/2448/1/012008

[13] D.O. Harris, G.G. Engerholm and W.D. Gwinn 1965 J.Chem.Phys. 43 151.
[14] A.S. Dickinson and P.R. Certain 1968 J.Chem.Phys. 49 4209.
[15] J.V. Lill, G.A. Parker and J.C. Light 1982 Chem.Phys. Lett. 89 483.
[16] J.C. Light, I.P. Hamilton and J.V. Lill 1985 J. Chem. Phys. 82 1400.
[17] Z. Bacic and J.C. Light 1986 J.Chem.Phys. 85 4594.
[18] J.V. Lill, G.A. Parker and J.C. Light 1986 J.Chem.Phys. 85 900.
[19] J.C. Light and Z.Bacic 1987 J.Chem.Phys. 87 4008.
[20] Z. Bacic and J.C. Light 1987 J.Chem.Phys. 86 3065.
[21] S.E. Choi and J.C. Light 1989 J.Chem.Phys. 90 3065.
[22] J. Tennison and J.R. Henderson 1989 J.Chem.Phys. 91 3815.
[23] J.C. Light and T. Carrington Jr. 2000 Adv. Chem. Phys. 114 263.
[24] B.D. Shizgal and R. Blackmore 1984 J. Comput. Phys. 55 313.
[25] B.D. Shizgal and H. Chen 1996 J. Chem. Phys. 104 4137.
[26] B.D. Shizgal 2015 Spectral Methods in Chemistry and physics. Applications to Kinetic Theory and Quantum
Mechanics. Springer. New York.
[27] B.D. Shizgal 2016 Comp. and Theor. Chem. 1084 51.
[28] B.D. Shizgal 2017 Comp. and Theor. Chem. 1114 25.
[29] J. Bao and B.D. Shizgal 2019 Comp. and Theor. Chem. 1149 49.
[30] D. Baye and P.H. Heenen 1986 J. Phys. A 19.11 2041.
[31] M. Vincke, L. Malegat and D. Baye 1993 J. Phys. B 26.11 811.
[32] D. Baye 2006 Phys. Status Solidi B 243 1095.
[33] D. Baye 2015 Physics Reports 565 1.
[34] M. Carvajal, J.M. Arias and J. Gómez-Camacho 1999 Phys. Rev. A 59 1852.
[35] M. Carvajal, J.M. Arias and J. Gómez-Camacho 1999 Phys. Rev. A 59 3462.
[36] F. Pérez-Bernal, J.M. Arias, M. Carvajal and J. Gómez-Camacho 2000 Phys. Rev. A 61 042504.
[37] H. Wei and T. Carrington 1992 J. Chem. Phys. 97 3029.
[38] M.J. Bramley and T. Carrington 1993 J. Chem. Phys. 99 8519.
[39] M.J. Bramley , J.W. Tromp, T.Carrington and G.C.Corey 1994 J. Chem. Phys. 100 6175.
[40] Xiao-Gang and T. Carrington Jr. 2001 J. Phys. Chem. A 105 2575.
[41] R. Lemus 2019 Mol. Phys. 117 167.
[42] R. Lemus 2019 J. Phys. Comm. 3.2 025012.
[43] M.M. Estévez-Fregoso, J.M. Arias, J. Gómez-Camacho and R. Lemus 2018 Mol. Phys. 116 2254.
[44] M.M. Estévez-Fregoso and R. Lemus 2018 Mol. Phys. 116 2374.
[45] M. Rodrı́guez-Arcos and R. Lemus 2018 Chem. Phys. Lett. 713 266.
[46] M. Rodrı́guez-Arcos, R. Lemus, J.M. Arias and J.Gómez-Camacho 2019 Mol. Phys. 118
[47] M. Bermúdez-Montaña, M. Rodrı́guez-Arcos, R. Lemus, J.M.Arias and J.Gómez-Camacho and E. Orgaz
2020 Symmetry 12 1719
[48] M. Rodrı́guez-Arcos, M. Bermúdez-Montaña, J.M. Arias, J. Gómez-Camacho, E. Orgaz and R. Lemus 2021
Mol. Phys. 119 e1876264.
[49] M. Moshinsky 1969 The harmonic oscillator in modern physics: from atoms to quarks, Gordon and Breach,
New York.
[50] R. G. Littlejohn, M. Cargo, T. Carrington, K. Mitchell and B. Poirier 2002 J. Chem. Phys. 116 8691.
[51] G. Pöschl and E. Teller 1933 Z. Phys. 83 143.
[52] N. Rosen and P.M. Morse 1932 Phys. Rev. 42 210.
[53] L.D. Landau and E.M. Lifshitz 1997 Quantum Mechanics, non relativistic theory. 3th edition, Pergamon.
Oxford.
[54] J.M. Arias, J. Gómez-Camacho and R. Lemus 2004 J. Phys. A; Math. Gen. 37 877.
[55] I.S. Gradshtein and I.M. Rizhic 1994 Tables of Integrals, Series and Product 5th edition, Academic Press.
New York.
[56] j-Q. Chen 1989 Group representation theory for physicists. World Scientific. New York.
[57] R. Lemus 2003 Mol. Phys. 101 2511.
[58] O. Álvarez-Bajo, R. Lemus, M. Carvajal and F. Pérez-Bernal 2011 Mol. Phys. 109 797.
[59] R. Lemus 2012 Symmetry 4 667.
[60] M. Bermúdez-Montaña, R. Lemus and O. Castaños 2016 EPL. 116 13001.
[61] M. Bermúdez-Montaña, R. Lemus and O. Castaños 2017 Mol. Phys. 115 3076.

13