Physics Letters B 735 (2014) 149–156
Contents lists available at ScienceDirect
Physics Letters B
www.elsevier.com/locate/physletb
The infrared limit of the SRG evolution and Levinson’s theorem
E. Ruiz Arriola
a,∗
, S. Szpigel
b
, V.S. Timóteo
c
a
Departamento de Física Atómica, Molecular y Nuclear and Instituto Carlos I de Fisica Teórica y Computacional, Universidad de Granada, E-18071 Granada, Spain
b
Centro de Rádio-Astronomia e Astrofísica Mackenzie, Escola de Engenharia, Universidade Presbiteriana Mackenzie, Brazil
c
Grupo de Óptica e Modelagem Numérica – GOMNI, Faculdade de Tecnologia – FT, Universidade Estadual de Campinas – UNICAMP, Brazil
a r t i c l e i n f o a b s t r a c t
Article history:
Received
23 April 2014
Received
in revised form 10 June 2014
Accepted
12 June 2014
Available
online 17 June 2014
Editor:
W. Haxton
On a finite momentum grid with N integration points p
n
and weights w
n
(n = 1,...,N) the Similarity
Renormalization Group (SRG) with a given generator G unitarily evolves an initial interaction with a
cutoff λ on energy differences, steadily driving the starting Hamiltonian in momentum space H
0
n
,m
=
p
2
n
δ
n,m
+ V
n,m
to a diagonal form in the infrared limit (λ → 0), H
G,λ→0
n
,m
= E
π(n)
δ
n,m
, where π (n) is
a permutation of the eigenvalues E
n
which depends on G. Levinson’s theorem establishes a relation
between phase-shifts δ(p
n
) and the number of bound-states, n
B
, and reads δ(p
1
) − δ(p
N
) = n
B
π. We
show that unitarily equivalent Hamiltonians on the grid generate reaction matrices which are compatible
with Levinson’s theorem but are phase-inequivalent along the SRG trajectory. An isospectral definition
of the phase-shift in terms of an energy-shift is possible but requires in addition a proper ordering of
states on a momentum grid such as to fulfill Levinson’s theorem. We show how the SRG with different
generators G induces different isospectral flows in the presence of bound-states, leading to distinct
orderings in the infrared limit. While the Wilson generator induces an ascending ordering incompatible
with Levinson’s theorem, the Wegner generator provides a much better ordering, although not the
optimal one. We illustrate the discussion with the nucleon–nucleon (NN) interaction in the
1
S
0
and
3
S
1
channels.
© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/3.0/). Funded by SCOAP
3
.
1. Introduction
During the last decade the renormalization group equations
have advantageously been used as a technique to simplify micro-
scopic
large scale calculations in Nuclear Structure and Reactions.
More specifically, the Similarity Renormalization Group (SRG) has
been intensively applied to handle multinucleon forces in order to
soften the short-distance core [1,2] with a rather universal pat-
tern
for nuclear symmetries [3,4] and interactions [5]. The ba-
sic
strategy underlying the SRG method is to evolve a starting
(bare) interaction H
0
which has been fitted to nucleon–nucleon
(NN) scattering data via a continuous unitary transformation that
runs a cutoff λ on energy differences. Such a transformation gen-
erates
a family of unitarily equivalent smooth interactions H
λ
=
U
λ
H
0
U
†
λ
with a band-diagonal structure of a prescribed width
roughly given by the SRG cutoff λ. For most cases of interest a
finite momentum grid with N integration points p
n
and resolu-
tion
weights p
n
= w
n
(n = 1,...,N) is needed to solve the SRG
flow equations numerically, and for such a finite basis the SRG
*
Corresponding author.
E-mail
addresses: earriola@ugr.es (E.R. Arriola), szpigel@mackenzie.br
(S. Szpigel),
varese@ft.unicamp.br (V.S. Timóteo).
transformation corresponds to a continuum generalization of the
well-known Gauss reduction method of a matrix to the diagonal
form. Typical calculations have taken w
n
∼ 0.01 fm
−1
, λ ∼ 2fm
−1
and N ∼ 200 [1,2]. For reasons to be elaborated below we will ana-
lyze
here a quite different regime, namely smaller SRG cut-offs and
grids, so our discussion will have no implications for the many ex-
isting
calculations.
Unfortunately
the NN force is not yet known from first prin-
ciples
and most information on the NN interaction is strongly
constrained by the abundant np and pp scattering data (see e.g.
Ref. [6] for a recent upgrade and references therein). Roughly
speaking this is equivalent to knowing the phase-shifts with their
uncertainties at some center-of-mass (CM) momenta and in a given
range, 0 <p ≡ p
1
< ··· < p
N
≡ Λ, and in fact a common prac-
tice
has been to tabulate the phase-shifts at given discrete set of
energy values. The implicit assumption underlying this practice is
that one expects this discrete information to encode and summa-
rize
sufficient details on the interaction, in full harmony with the
need of solving SRG flow equations on a finite grid. The computa-
tional
advantages of using properly chosen few discrete variables
for finite volume systems such as nuclei [7] have been emphasized
as the number of states gets drastically reduced.
http://dx.doi.org/10.1016/j.physletb.2014.06.032
0370-2693/
© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). Funded by
SCOAP
3
.