Pseudo-Bosons from Landau Levels
We construct examples of pseudo-bosons in two dimensions arising from the Hamiltonian for the Landau levels. We also prove a no-go result showing that non-linear combinations of bosonic creation and annihilation operators cannot give rise to pseudo-bosons.
Saved in:
| Date: | 2010 |
|---|---|
| Main Author: | |
| Format: | Article |
| Language: | English |
| Published: |
Інститут математики НАН України
2010
|
| Series: | Symmetry, Integrability and Geometry: Methods and Applications |
| Online Access: | https://nasplib.isofts.kiev.ua/handle/123456789/146525 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Journal Title: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Cite this: | Pseudo-Bosons from Landau Levels / F. Bagarello // Symmetry, Integrability and Geometry: Methods and Applications. — 2010. — Т. 6. — Бібліогр.: 14 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-146525 |
|---|---|
| record_format |
dspace |
| spelling |
nasplib_isofts_kiev_ua-123456789-1465252025-02-09T17:45:56Z Pseudo-Bosons from Landau Levels Bagarello, F. We construct examples of pseudo-bosons in two dimensions arising from the Hamiltonian for the Landau levels. We also prove a no-go result showing that non-linear combinations of bosonic creation and annihilation operators cannot give rise to pseudo-bosons. This paper is a contribution to the Proceedings of the Workshop “Supersymmetric Quantum Mechanics and Spectral Design” (July 18–30, 2010, Benasque, Spain). The full collection is available at http://www.emis.de/journals/SIGMA/SUSYQM2010.html. The author would like to thank A. Andrianov for his kind invitation and the local people in Benasque for their warm welcome. 2010 Article Pseudo-Bosons from Landau Levels / F. Bagarello // Symmetry, Integrability and Geometry: Methods and Applications. — 2010. — Т. 6. — Бібліогр.: 14 назв. — англ. 1815-0659 2010 Mathematics Subject Classification: 81Q65; 65H17 DOI:10.3842/SIGMA.2010.093 https://nasplib.isofts.kiev.ua/handle/123456789/146525 en Symmetry, Integrability and Geometry: Methods and Applications application/pdf Інститут математики НАН України |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| language |
English |
| description |
We construct examples of pseudo-bosons in two dimensions arising from the Hamiltonian for the Landau levels. We also prove a no-go result showing that non-linear combinations of bosonic creation and annihilation operators cannot give rise to pseudo-bosons. |
| format |
Article |
| author |
Bagarello, F. |
| spellingShingle |
Bagarello, F. Pseudo-Bosons from Landau Levels Symmetry, Integrability and Geometry: Methods and Applications |
| author_facet |
Bagarello, F. |
| author_sort |
Bagarello, F. |
| title |
Pseudo-Bosons from Landau Levels |
| title_short |
Pseudo-Bosons from Landau Levels |
| title_full |
Pseudo-Bosons from Landau Levels |
| title_fullStr |
Pseudo-Bosons from Landau Levels |
| title_full_unstemmed |
Pseudo-Bosons from Landau Levels |
| title_sort |
pseudo-bosons from landau levels |
| publisher |
Інститут математики НАН України |
| publishDate |
2010 |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/146525 |
| citation_txt |
Pseudo-Bosons from Landau Levels / F. Bagarello // Symmetry, Integrability and Geometry: Methods and Applications. — 2010. — Т. 6. — Бібліогр.: 14 назв. — англ. |
| series |
Symmetry, Integrability and Geometry: Methods and Applications |
| work_keys_str_mv |
AT bagarellof pseudobosonsfromlandaulevels |
| first_indexed |
2025-11-29T00:21:21Z |
| last_indexed |
2025-11-29T00:21:21Z |
| _version_ |
1850082005190967296 |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications SIGMA 6 (2010), 093, 9 pages
Pseudo-Bosons from Landau Levels?
Fabio BAGARELLO
Dipartimento di Metodi e Modelli Matematici, Facoltà di Ingegneria,
Università di Palermo, I-90128 Palermo, Italy
E-mail: bagarell@unipa.it
URL: http://www.unipa.it/∼bagarell/
Received October 25, 2010, in final form December 02, 2010; Published online December 12, 2010
doi:10.3842/SIGMA.2010.093
Abstract. We construct examples of pseudo-bosons in two dimensions arising from the
Hamiltonian for the Landau levels. We also prove a no-go result showing that non-linear
combinations of bosonic creation and annihilation operators cannot give rise to pseudo-
bosons.
Key words: non-hermitian Hamiltonians; pseudo-bosons
2010 Mathematics Subject Classification: 81Q65; 65H17
1 Introduction
In a series of recent papers [1, 2, 3, 4, 5, 6], we have investigated some mathematical aspects
of the so-called pseudo-bosons, originally introduced by Trifonov1 in [8]. They arise from the
canonical commutation relation [a, a†] = 11 upon replacing a† by another (unbounded) opera-
tor b not (in general) related to a: [a, b] = 11. We have shown that, under suitable assumptions,
N = ba and N † = a†b† can be both diagonalized, and that their spectra coincide with the set of
natural numbers (including 0), N0. However the sets of related eigenvectors are not orthonormal
(o.n.) bases but, nevertheless, they are automatically biorthonormal. In most of the examples
considered so far, they are bases of the Hilbert space of the system, H, and, in some cases, they
turn out to be Riesz bases.
In [9] and [10] some physical examples arising from quantum mechanics have been discussed.
In particular, these examples have suggested the introduction of a difference between what we
have called regular pseudo-bosons and pseudo-bosons, to better focus on what we believe are the
mathematical or on the physical aspects of these particles. Indeed all the examples of regular
pseudo-bosons considered so far arise from Riesz bases [4], with a rather mathematical construc-
tion, while pseudo-bosons are those which one can find when starting with the Hamiltonian of
some realistic quantum system.
In this paper, after a short review of the general framework, we discuss a two-dimensional
example arising from the Hamiltonian of the Landau levels. It should be stressed that this
example is of a completely different kind than those considered in [10], where a modified version
of the Landau levels have been considered.
We close the paper with a no-go result, suggesting that non-linear combinations of ordinary
bosonic creation and annihilation operators, even if they produce pseudo-bosonic commutation
rules, cannot satisfy the Assumptions of our construction, see Section 2.
?This paper is a contribution to the Proceedings of the Workshop “Supersymmetric Quantum Me-
chanics and Spectral Design” (July 18–30, 2010, Benasque, Spain). The full collection is available at
http://www.emis.de/journals/SIGMA/SUSYQM2010.html
1It should be mentioned that pseudo-bosons already appeared in [7] but with a different meaning.
mailto:bagarell@unipa.it
http://www.unipa.it/~bagarell/
http://dx.doi.org/10.3842/SIGMA.2010.093
http://www.emis.de/journals/SIGMA/SUSYQM2010.html
2 F. Bagarello
2 The commutation rules
In this section we will review a d-dimensional version of what originally proposed in [1, 6].
Let H be a given Hilbert space with scalar product 〈·, ·〉 and related norm ‖ · ‖. We intro-
duce d pairs of operators, aj and bj , j = 1, 2, . . . , d, acting on H and satisfying the following
commutation rules
[aj , bk] = δj,k11, (1)
j, k = 1, 2, . . . , d. Of course, these collapse to the CCR’s for d independent modes if bj = a†j ,
j = 1, 2, . . . , d. It is well known that aj and bj are unbounded operators, so they cannot be
defined on all of H. Following [1], and writing D∞(X) := ∩p≥0D(Xp) (the common domain of
all the powers of the operator X), we consider the following:
Assumption 1. There exists a non-zero ϕ0 ∈ H such that ajϕ0 = 0, j = 1, 2, . . . , d, and
ϕ0 ∈ D∞(b1) ∩D∞(b2) ∩ · · · ∩D∞(bd).
Assumption 2. There exists a non-zero Ψ0 ∈ H such that b†jΨ0 = 0, j = 1, 2, . . . , d, and
Ψ0 ∈ D∞(a†1) ∩D∞(a†2) ∩ · · · ∩D∞(a†d).
Under these assumptions we can introduce the following vectors in H:
ϕn := ϕn1,n2,...,nd
=
1√
n1!n2! · · ·nd!
bn1
1 bn2
2 · · · bnd
d ϕ0,
Ψn := Ψn1,n2,...,nd
=
1√
n1!n2! · · ·nd!
a†1
n1
a†2
n2 · · · a†d
nd Ψ0,
nj = 0, 1, 2, . . . for all j = 1, 2, . . . , d. Let us now define the unbounded operators Nj := bjaj
and Nj := N †
j = a†jb
†
j , j = 1, 2, . . . , d. Each ϕn belongs to the domain of Nj , D(Nj), and
Ψn ∈ D(Nj), for all possible n. Moreover,
Njϕn = njϕn, NjΨn = njΨn.
Under the above assumptions, and if we chose the normalization of Ψ0 and ϕ0 in such a way
that 〈Ψ0, ϕ0〉 = 1, we find that
〈Ψn, ϕm〉 = δn,m =
d∏
j=1
δnj ,mj .
This means that the sets FΨ = {Ψn} and Fϕ = {ϕn} are biorthonormal and, because of this,
the vectors of each set are linearly independent. If we now call Dϕ and DΨ respectively the
linear span of Fϕ and FΨ, and Hϕ and HΨ their closures, then
f =
∑
n
〈Ψn, f〉ϕn, ∀ f ∈ Hϕ, h =
∑
n
〈ϕn, h〉Ψn, ∀h ∈ HΨ.
What is not in general ensured is that Hϕ = HΨ = H. Indeed, we can only state that Hϕ ⊆ H
and HΨ ⊆ H. However, motivated by the examples discussed so far in the literature, we consider
Assumption 3. The above Hilbert spaces all coincide: Hϕ = HΨ = H.
This means, in particular, that both Fϕ and FΨ are bases of H. The resolution of the identity
in the bra-ket formalism looks like∑
n
|ϕn〉〈Ψn| =
∑
n
|Ψn〉〈ϕn| = 11.
Pseudo-Bosons from Landau Levels 3
Let us now introduce the operators Sϕ and SΨ via their action respectively on FΨ and Fϕ:
SϕΨn = ϕn, SΨϕn = Ψn,
for all n, which also imply that Ψn = (SΨ Sϕ)Ψn and ϕn = (Sϕ SΨ)ϕn, for all n. Hence
SΨSϕ = SϕSΨ = 11 ⇒ SΨ = S−1
ϕ .
In other words, both SΨ and Sϕ are invertible and one is the inverse of the other. Furthermore,
we can also check that they are both positive, well defined and symmetric [1]. Moreover, it is
possible to write these operators as
Sϕ =
∑
n
|ϕn〉〈ϕn|, SΨ =
∑
n
|Ψn〉〈Ψn|.
These expressions are only formal, at this stage, since the series may not converge in the uniform
topology and the operators Sϕ and SΨ could be unbounded. Indeed we know [11], that two
biorthonormal bases are related by a bounded operator, with bounded inverse, if and only if
they are Riesz bases2. This is why in [1] we have also considerered
Assumption 4. Fϕ and FΨ are both Riesz bases.
Therefore, as already stated, Sϕ and SΨ are bounded operators and their domains can be
taken to be all of H. While Assumptions 1, 2 and 3 are quite often satisfied, [12], it is quite
difficult to find physical examples satisfying also Assumption 4. On the other hand, it is rather
easy to find mathematical examples satisfying all the assumptions, see [1, 6]. This is why
in [9] we have introduced a difference in the notation: we have called pseudo-bosons (PB) those
satisfying the first three assumptions, while, if they also satisfy Assumption 4, they are called
regular pseudo-bosons (RPB).
As already discussed in our previous papers, these d-dimensional pseudo-bosons give rise
to interesting intertwining relations among non self-adjoint operators, see in particular [3] and
references therein. For instance, it is easy to check that
SΨNj = NjSΨ and NjSϕ = SϕNj ,
j = 1, 2, . . . , d. This is related to the fact that the spectra of, say, N1 and N1, coincide and that
their eigenvectors are related by the operators Sϕ and SΨ, in agreement with the literature on
intertwining operators [13, 14].
3 The example
In this section we will consider an example arising from a quantum mechanical system, i.e.
a single electron moving on a two-dimensional plane and subject to a uniform magnetic field
along the z-direction. Taking ~ = m = eB
c = 1, the Hamiltonian of the electron is given by the
operator
H1 =
1
2
(
p−A(r)
)2 =
1
2
(
px +
y
2
)2
+
1
2
(
py −
x
2
)2
, (2)
where we have used minimal coupling and the symmetric gauge ~A = 1
2(−y, x, 0). The Hilbert
space of the system is H = L2(R2).
2Recall that a set of vectors φ1, φ2, φ3, . . ., is a Riesz basis of a Hilbert space H, if there exists a bounded
operator V , with bounded inverse, on H, and an o.n. basis of H, ϕ1, ϕ2, ϕ3, . . ., such that φj = V ϕj , for all
j = 1, 2, 3, . . .
4 F. Bagarello
The spectrum of this Hamiltonian is easily obtained by first introducing the new variables
Q1 = px + y/2, P1 = py − x/2. (3)
In terms of P1 and Q1 the single electron Hamiltonian, H1, can be rewritten as
H1 =
1
2
(Q2
1 + P 2
1 ).
The transformation (3) is part of a canonical map from the variables (x, y, px, py) to (Q1, Q2, P1,
P2), where
Q2 = py + x/2, P2 = px − y/2,
which can be used to construct a second Hamiltonian H2 = 1
2(Q2
2+P 2
2 ). Since [x, px] = [y, py] = i,
[x, py] = [y, px] = [x, y] = [px, py] = 0, we deduce that
[Q1, P1] = [Q2, P2] = i, [Q1, P2] = [Q2, P1] = [Q1, Q2] = [P1, P2] = 0,
so that [H1,H2] = 0. The two Hamiltonians correspond to two opposite magnetic fields, respec-
tively along +k̂ and −k̂. Let us now introduce the operators
Ak =
1√
2
(Qk + iPk) ,
k = 1, 2, together with their adjoints. Then [Ak, A
†
l ] = δk,l11, the other commutators being zero.
In terms of these operators we can write Hk = A†kAk + 1
211, k = 1, 2, whose eigenvectors are
Φ(k)
n = 1√
n!
(A†k)
nΦ(k)
0 , where k = 1, 2, n = 0, 1, 2, . . . and Φ(k)
0 is the vacuum of Ak: AkΦ
(k)
0 = 0.
Furthermore we have 〈Φ(k)
n ,Φ(k)
m 〉 = δn,m and HkΦ
(k)
n =
(
n + 1
2
)
Φ(k)
n , for k = 1, 2. It is natural
to introduce the sets Fk :=
{
Φ(k)
n , n ≥ 0
}
, k = 1, 2, and the closures of their linear span, H1
and H2. Hence, by construction, Fk is an o.n. basis of Hk. Moreover, we can also introduce an
o.n. basis of H as the set FΦ whose vectors are defined as follows:
Φn,m :=
1√
n!m!
(A†1)
n(A†2)
mΦ0,0,
where Φ0,0 := Φ(1)
0 ⊗ Φ(2)
0 is such that A1Φ0,0 = A2Φ0,0 = 0. It is clear that Φn,m = Φ(1)
n ⊗ Φ(2)
m
and that H = H1 ⊗H2.
3.1 Pseudo-bosons in H1
Let us now define the following operators: A1(α) = A1 and B1(α) = A†1 + 2αA1, where α is
a fixed complex number. It is clear that, for α 6= 0, A1(α)† 6= B1(α). Moreover, [A1(α), B1(α)] =
11, ∀α. Hence, we recover (1) for d = 1 in H1. We want to show that A1(α) and B1(α) generate
PB in H1 which are not regular.
To begin with, we define ϕ
(1)
0 (α) := Φ(1)
0 . This non zero vector of H1 satisfies Assumption 1:
A1(α)ϕ(1)
0 (α) = 0, clearly, and ϕ
(1)
0 (α) ∈ D∞(B1(α)). This follows from the fact that, since
B1(α) = A†1 + 2αA1, B1(α)nϕ
(1)
0 (α) is a finite linear combination of the vectors Φ(1)
0 ,Φ(1)
1 , . . .,
Φ(1)
n , which is clearly a vector of H1.
Before considering Assumption 2, it is convenient to observe that, introducing the following
invertible and densely defined operator U1(α) := eαA2
1 , we can write
A1(α) = U1(α)A1U1(α)−1, B1(α) = U1(α)A†1U1(α)−1,
Pseudo-Bosons from Landau Levels 5
ϕ(1)
n (α) :=
1√
n!
B1(α)nϕ
(1)
0 (α) = U1(α)Φ(1)
n , (4)
for all n ≥ 0. Of course, ϕ
(1)
n (α) is well defined for all n ≥ 0 since, as we have seen, B1(α)nϕ
(1)
0 (α)
is well defined for all complex α. Now, if we define (at least formally, at this stage)
Ψ(1)
0 (α) :=
(
U1(α)†
)−1Φ(1)
0 , (5)
it is possible to show that, if |α| < 1
2 : (i) Ψ(1)
0 (α) is well defined in H1, and is different from
zero; (ii) B1(α)†Ψ(1)
0 (α) = 0; (iii) Ψ(1)
0 (α) ∈ D∞(A†1). It is furthermore possible to check that,
for the same values of α,
Ψ(1)
n (α) :=
1√
n!
(
A1(α)†
)nΨ(1)
0 (α) =
(
U1(α)†
)−1Φ(1)
n . (6)
Let us prove point (iii) above. We have, for all n ≥ 0,
(
A†1
)n
e−αA†
1
2
Φ(1)
0 =
∞∑
k=0
(−α)k
k!
√
(2k + n)!Φ(1)
2k+n,
so that∥∥(A†1)ne−αA†
1
2
Φ(1)
0
∥∥2 =
∞∑
k=0
|α|2k
(k!)2
(2k + n)!,
which converges inside the disk |α| < 1
2 . In particular, if n = 0, this implies the statement in (i)
above. The proof of (ii) is trivial and the last equality in (6) can be deduced using (4) and (5) in
the definition Ψ(1)
n (α) := 1√
n!
(A1(α)†)nΨ(1)
0 (α). This, as we have seen, is well defined if |α| < 1
2 ,
while, for |α| > 1
2 all the procedure makes no sense, since the vectors we are using do not belong
to the Hilbert space. Now, biorthonormality of the two sets Fϕ(1) := {ϕ(1)
n (α), n ≥ 0} and
FΨ(1) := {Ψ(1)
n (α), n ≥ 0} follows directly from their definitions:
〈ϕ(1)
n (α),Ψ(1)
m (α)〉 = 〈U1(α)Φ(1)
n ,
(
U1(α)†
)−1Φ(1)
m 〉 = 〈Φ(1)
n ,Φ(1)
m 〉 = δn,m.
The proof of Assumption 3 goes as follows:
First of all, as we have already stated, it is possible to check that for all n ≥ 0 we have
ϕ
(1)
n (α) = Φ(1)
n +
∑n−1
k=0 dkΦ
(1)
k , for some constants {dk, k = 0, 1, . . . , n− 1}.
Secondly, using induction on n and this simple remark we can prove that, if f ∈ H1 is such
that 〈f, ϕ
(1)
k (α)〉 = 0 for k = 0, 1, . . . , n, then 〈f,Φ(1)
k 〉 = 0 for k = 0, 1, . . . , n as well. Therefore,
if f is orthogonal to all the ϕ
(1)
k (α)’s, it is also orthogonal to all the Φ(1)
k ’s, whose set is complete
in H1. Hence f = 0, so that Fϕ(1) is also complete in H1.
As a consequence, being the vectors of Fϕ(1) linearly independent and complete in H1, they
are a basis of H1. In particular we find that, for all f ∈ H1, the following expansion holds true:
f =
∑∞
k=0〈Ψ
(1)
n (α), f〉ϕ(1)
n (α). Then, for all f, g ∈ H1,
〈g, f〉 =
〈
g,
∞∑
k=0
〈Ψ(1)
n (α), f〉ϕ(1)
n (α)
〉
=
〈 ∞∑
k=0
〈ϕ(1)
n (α), g〉Ψ(1)
n (α), f
〉
,
which, since f could be any vector in H1, implies that g =
∑∞
k=0〈ϕ
(1)
n (α), g〉Ψ(1)
n (α): FΨ(1) is
a basis of H1 as well, and Assumption 3 is satisfied. Finally, Assumption 4 is not satisfied since,
for instance, the operator
(
U1(α)†
)−1 is unbounded [11].
6 F. Bagarello
Remark 1. It might be worth stressing that, while it is quite easy to check that the set Fϕ(1)
is complete in D(U(α)†), it is not trivial at all to check that it is also complete in H1. This is
the reason why we have used the above procedure.
It is not hard to deduce the expression of two non self-adjoint operators which admit ϕ
(1)
n (α)
and Ψ(1)
n (α) as eigenstates. For that we define first h1(α) := U1(α)H1U1(α)−1 = B1(α)A1(α) +
1
211, which, in coordinate representation, looks like
h1(α) =
(
1
2
+ α
)(
px +
y
2
)2
+
(
1
2
− α
)(
py −
x
2
)2
+ 2iα
(
px +
y
2
)(
py −
x
2
)
+ α11.
We can also introduce h1(α)†, which is clearly different from h1(α). Now, as expected from
general facts in the theory of intertwining operators [13], we see that
h1(α)ϕ(1)
n (α) = (n + 1/2)ϕ(1)
n (α), h1(α)†Ψ(1)
n (α) = (n + 1/2)Ψ(1)
n (α),
for all n ≥ 0.
3.2 Pseudo-bosons in H2
In this subsection we will consider an analogous construction in H2, i.e. in the Hilbert space
related to the uniform magnetic field along −k̂. To make the situation more interesting, and
to avoid repeating essentially the same procedure considered above, instead of introducing an
operator like eβA2
2 we consider
U2(β) := eβA†
2
2
,
with β ∈ C. Then we define
A2(β) := U2(β)A2U2(β)−1 = A2 − 2βA†2,
B2(β) := U2(β)A†2U2(β)−1 = A†2. (7)
These are pseudo-bosonic operators in H2: [A2(β), B2(β)] = 11, and A2(β)† 6= B2(β), for β 6= 0.
Then, once again, it may be interesting to consider Assumptions 1–4.
If |β| < 1
2 Assumption 1 is satisfied: let us define (formally, for the moment) ϕ
(2)
0 (β) =
U2(β)Φ(2)
0 . Then A2(β)ϕ(2)
0 (β) = 0. Moreover, since [B2(β), U2(β)] = 0, B2(β)nϕ
(2)
0 (β) =
U2(β)A†2
n
Φ(2)
0 , which implies in particular that
ϕ(2)
n (β) :=
1√
n!
B2(β)nϕ
(2)
0 (β) = U2(β)Φ(2)
n .
Of course we have now to check that ϕ
(2)
n (β) is a well defined vector of H2 for all n ≥ 0. This
would make the above formal definition rigorous. The computation of ‖U2(β)Φ(2)
n ‖ follows the
same steps as that for ‖U1(α)†−1Φ(1)
n ‖ of the previous section, and we get the same conclusion:
the power series obtained for ‖U2(β)Φ(2)
n ‖2 converges if |β| < 1
2 , so that Φ(2)
n ∈ D(U2(β)) for all
n ≥ 0, inside this disk.
As for Assumption 2, this is also satisfied: to prove this it is enough to take Ψ(2)
0 (β) = Φ(2)
0 .
Then B2(β)†Ψ(2)
0 (β) = A2Φ
(2)
0 = 0. Also, since U2(β)†Φ(2)
0 = Φ(2)
0 , formula (7) implies that
1√
n!
(A†2(β))nΨ(2)
0 (β) = e−β A2
2Φ(2)
n , which is clearly a vector in H2 since it is a finite linear
combination of Φ(2)
0 ,Φ(2)
1 , . . . ,Φ(2)
n . This means that the vectors
Ψ(2)
n (β) :=
1√
n!
A2(β)†
n
Ψ(2)
0 (β) =
(
U2(β)†
)−1Φ(2)
n
Pseudo-Bosons from Landau Levels 7
are well defined in H2 for all n, independently of β. Once again we deduce that the vectors
constructed here are biorthonormal,
〈ϕ(2)
n (β),Ψ(2)
m (β)〉 = δn,m,
and that they are eigenstates of two operators which are the adjoint one of the other, and which
are related to H2 by a similarity transformation:
h2(β) := U2(β)H2U2(β)−1 = B2(β)A2(β) +
1
2
11,
which in coordinate representation looks like
h2(β) =
(
1
2
− β
)(
py +
x
2
)2
+
(
1
2
+ β
)(
px −
y
2
)2
+ 2iβ
(
py +
x
2
)(
px −
y
2
)
+ β11.
In particular we find that
h2(β)ϕ(2)
n (β) = (n + 1/2)ϕ(2)
n (β), h2(β)†Ψ(2)
n (β) = (n + 1/2)Ψ(2)
n (β),
for all n ≥ 0. The same arguments used previously prove that Fϕ(2) := {ϕ(2)
n (β), n ≥ 0} and
FΨ(2) := {Ψ(2)
n (β), n ≥ 0} are both complete in H2. More than this: they are biorthonormal
bases but not Riesz bases.
3.3 Pseudo-bosons in H
We begin this section with the following remark: none of the above sets of functions is complete
in H. Hence we could try to find a different set of vectors, also labeled by a single quantum
number, which is complete in H. It is not hard to check that this is not possible, in general. Let
us introduce, for instance, the following pseudo-bosonic operators: Xα,β := 1√
2
(A1(α) + A2(β))
and Yα,β := 1√
2
(B1(α) + B2(β)). Then [Xα,β, Yα,β] = 11, Xα,β 6= Y †α,β in general and the vectors
ϕ0,0(α, β) := ϕ
(1)
0 (α)⊗ ϕ
(2)
0 (β) and Ψ0,0(α, β) := Ψ(1)
0 (α)⊗Ψ(2)
0 (β) satisfy Assumptions 1 and 2
of Section 2. However, it is not hard to check that the vectors ηn(α, β) := 1√
n!
Y n
α,βϕ0,0(α, β),
n ≥ 0, are not complete in H: for that it is enough to consider the non zero vector f =
Ψ(1)
1 (α)⊗Ψ(2)
0 (β)−Ψ(1)
0 (α)⊗Ψ(2)
1 (β), which is non zero and orthogonal to all the ηn(α, β)’s.
This is not surprising: in Section 2, in fact, we have proposed a different way to produce two
biorthonormal bases of H, in dimension larger than 1. For instance, in d = 2 we expect that
the vectors of these bases depend on two quantum numbers rather than just one. So we may
proceed as follows: let T (α, β) be the following unbounded operator:
T (α, β) := U1(α)U2(β) = eαA2
1+βA†
2
2
.
Then the vectors ϕ0,0(α, β) and Ψ0,0(α, β) introduced above can be defined as ϕ0,0(α, β) =
T (α, β)Φ0,0 and Ψ0,0(α, β) =
(
T (α, β)†
)−1 Φ0,0. For what we have seen in the previous sec-
tions, these two vectors satisfy Assumptions 1 and 2: A1(α)ϕ0,0(α, β) = A2(β)ϕ0,0(α, β) = 0,
B1(α)†Ψ0,0(α, β) = B2(β)†Ψ0,0(α, β) = 0, and ϕ0,0(α, β) ∈ D∞(B1(α))∩D∞(B2(β)), Ψ0,0(α, β) ∈
D∞(A1(α)†) ∩D∞(A2(β)†).
Furthermore, sinceH = H1⊗H2, the sets Fϕ :={ϕn,m(α, β) := 1√
n!m!
B1(α)nB2(β)mϕ0,0(α, β)}
and FΨ := {Ψn,m(α, β) := 1√
n!m!
A1(α)†nA2(β)†mΨ0,0(α, β)} are complete in H, so that As-
sumption 3 is also satisfied. Finally, Assumption 4 is not verified, so that we have found
PB which are not regular. This is because T (α, β) is unbounded and since we can write
ϕn,m(α, β) = T (α, β)Φn,m and Ψn,m(α, β) =
(
T (α, β)†
)−1 Φn,m, for all n and m, [11].
8 F. Bagarello
Remark 2. The procedure outlined in this section clearly applies to any pair of uncoupled
harmonic oscillators h1 = a†1a1 and h2 = a†2a2, [ai, a
†
j ] = δi,j 11, i, j = 1, 2, changing properly the
definitions of the operators involved.
Remark 3. Bi-coherent states like those in [1] can be easily constructed from the ones for A1
and A2 using the operators U1(α) and U2(β).
4 A no-go result
We devote this short section to prove the following general no-go result: suppose a and a† are two
operators acting on H and satisfying [a, a†] = 11. Then, for all α 6= 0, the operators A := a−αa†
2
and B := a† are such that [A,B] = 11, A† 6= B, but they do not satisfy Assumption 1.
In fact, if such a non zero vector ϕ0 ∈ H exists, then it could be expanded in terms of the
eigenvectors Φn := a†
n
√
n!
Φ0, aΦ0 = 0, of the number operator N = a†a: ϕ0 =
∑∞
n=0 cnΦn, for
some sequence {cn, n ≥ 0} such that
∑∞
n=0 |cn|2 < ∞. Condition Aϕ0 = 0 can be rewritten as
aϕ0 = αa†
2
ϕ0. Now, inserting in both sides of this equality the expansion for ϕ0, and recalling
that a†Φn =
√
n + 1Φn+1 and aΦn =
√
nΦn−1, n ≥ 0, we deduce the following relations between
the coefficients cn: c1 = c2 = 0 and cn+1
√
n + 1 = αcn−2
√
(n− 1)n, for all n ≥ 2. The solution
of this recurrence relation is the following:
c3 = αc0
√
3!
3
, c6 = α2c0
√
6!
3 · 6
, c9 = α3c0
√
9!
3 · 6 · 9
, c12 = α4c0
√
12!
3 · 6 · 9 · 12
,
and so on. Then
ϕ0 = c0
(
Φ0 +
∞∑
k=1
αk
√
(3k)!
1 · 3 · · · 3k
Φ3k
)
.
However, computing ‖ϕ0‖ we deduce that this series only converge if α = 0, i.e. if A coincides
with a and B with a†.
A similar results can be obtained considering the operators A := a−αa†
n and B := a†−β11,
n ≥ 2, α, β ∈ C. Again we find [A,B] = 11, A† 6= B, and again, with similar techniques, we
deduce that they do not satisfy Assumption 1. In the same way, if we define A := a − α11
and B := a† − βam, m ≥ 2, α, β ∈ C, we find that, in general, [A,B] = 11, A† 6= B, but they
do not satisfy Assumption 2. This suggests that if we try to define, starting from a and a†,
new operators A = a + f(a, a†) and B = a† + g(a, a†), only very special choices of f and g are
compatible with the pseudo-bosonic structure.
5 Conclusions
We have seen how a non trivial example of two-dimensional PB arises from the Hamiltonian
of the Landau levels. We want to stress once again that this is deeply different from what we
have done in [10], where the starting point was a generalized Hamiltonian obtained with a smart
extension of that in (2) with the introduction of two related superpotentials. Among the other
differences, while the procedure outlined here works in any Hilbert space, the one in [10] works
only in L2(R2). The fact that both here and in [10] we get pseudo-bosons which are not regular
is still another indication of the mathematical nature of RPB.
It is not difficult to modify or to generalize the results in Section 3, for instance changing the
role of the operators U1 and U2, or modifying a bit their definitions. Maybe more interesting is
to try to extend the no-go result of Section 4 to other possible combinations of a and a†: this is
part of our work in progress.
Pseudo-Bosons from Landau Levels 9
Acknowledgements
The author would like to thank A. Andrianov for his kind invitation and the local people in
Benasque for their warm welcome.
References
[1] Bagarello F., Pseudobosons, Riesz bases and coherent states, J. Math. Phys. 51 (2010), 023531, 10 pages,
arXiv:1001.1136.
[2] Bagarello F., Construction of pseudobosons systems, J. Math. Phys. 51 (2010), 053508, 12 pages,
arXiv:1003.4620.
[3] Bagarello F., Mathematical aspects of intertwining operators: the role of Riesz bases, J. Phys. A: Math.
Theor. 43 (2010), 175203, 11 pages, arXiv:1003.2288.
[4] Bagarello F., Calabrese F., Pseudo-bosons arising from Riesz bases, Boll. Dip. Met. Mod. Mat. Palermo 2
(2010), 15–26.
[5] Bagarello F., Two-parameters pseudo-bosons, Internat. J. Theoret. Phys., to appear, arXiv:1010.0204.
[6] Bagarello F., (Regular) pseudo-bosons versus bosons, J. Phys. A: Math. Theor. 44 (2011), 015205, 11 pages.
[7] Besnard F., Number operator algebras and deformations of ε-Poisson algebras, Lett. Math. Phys. 55 (2001),
113–125, math-ph/0006012.
[8] Trifonov D.A., Pseudo-boson coherent and Fock states, in Proceedings of the 9th International Workshop
on “Complex Structures, Integrability and Vector Fields” (August 25–29, 2008, Sofia), Editors K. Sekigawa,
V. Gerdjikov and S. Dimiev, World Scientific Publishing Co., 2009, 241–250, arXiv:0902.3744.
[9] Bagarello F., Examples of Pseudo-bosons in quantum mechanics, Phys. Lett. A 374 (2010), 3823–3827,
arXiv:1007.4349.
[10] Ali S.T., Bagarello F., Gazeau J.-P., Modified Landau levels, damped harmonic oscillator and two-
dimensional pseudo-bosons, J. Math. Phys. 51 (2010), 123502, 11 pages, arXiv:1010.4221.
[11] Young R.M., An introduction to nonharmonic Fourier series, Academic Press, New York, 1980.
[12] Bagarello F., Pseudo-bosons, so far, submitted.
[13] Kuru S., Tegmen A., Vercin A., Intertwined isospectral potentials in an arbitrary dimension, J. Math. Phys.
42 (2001), 3344–3360, quant-ph/0111034.
Kuru S., Demircioglu B., Onder M., Vercin A., Two families of superintegrable and isospectral potentials in
two dimensions, J. Math. Phys. 43 (2002), 2133–2150, quant-ph/0201099.
Aghababaei Samani K., Zarei M., Intertwined Hamiltonians in two-dimensional curved spaces, Ann. Physics
316 (2005), 466–482, math-ph/0411030.
[14] Bagarello F., Extended SUSY quantum mechanics, intertwining operators and coherent states, Phys. Lett. A
372 (2008), 6226–6231, arXiv:0904.0199.
Bagarello F., Vector coherent states and intertwining operators, J. Phys. A: Math. Theor. 42 (2009), 075302,
11 pages, arXiv:0904.0201.
Bagarello F., Intertwining operators between different Hilbert spaces: connection with frames, J. Math.
Phys. 50 (2009), 043509, 13 pages, arXiv:0904.0203.
http://dx.doi.org/10.1063/1.3300804
http://arxiv.org/abs/1001.1136
http://dx.doi.org/10.1063/1.3397408
http://arxiv.org/abs/1003.4620
http://dx.doi.org/10.1088/1751-8113/43/17/175203
http://dx.doi.org/10.1088/1751-8113/43/17/175203
http://arxiv.org/abs/1003.2288
http://dx.doi.org/10.1007/s10773-010-0525-9
http://arxiv.org/abs/1010.0204
http://dx.doi.org/10.1088/1751-8113/44/1/015205
http://dx.doi.org/10.1023/A:1010907630787
http://arxiv.org/abs/math-ph/0006012
http://arxiv.org/abs/0902.3744
http://arxiv.org/abs/1007.4349
http://dx.doi.org/10.1063/1.3514196
http://arxiv.org/abs/1010.4221
http://dx.doi.org/10.1063/1.1383787
http://arxiv.org/abs/quant-ph/0111034
http://dx.doi.org/10.1063/1.1463217
http://arxiv.org/abs/quant-ph/0201099
http://dx.doi.org/10.1016/j.aop.2004.11.002
http://arxiv.org/abs/math-ph/0411030
http://dx.doi.org/10.1016/j.physleta.2008.08.047
http://arxiv.org/abs/0904.0199
http://dx.doi.org/10.1088/1751-8113/42/7/075302
http://arxiv.org/abs/0904.0201
http://dx.doi.org/10.1063/1.3094758
http://dx.doi.org/10.1063/1.3094758
http://arxiv.org/abs/0904.0203
1 Introduction
2 The commutation rules
3 The example
3.1 Pseudo-bosons in H1
3.2 Pseudo-bosons in H2
3.3 Pseudo-bosons in H
4 A no-go result
5 Conclusions
References
|