Transcript PPTX-presentation

What are we doing?
What are we doing?
Large-scale ab initio No-core Shell Model
calculations
What are we doing?
Large-scale ab initio No-core Shell Model
calculations
+ new realistic NN interaction JISP
Ab initio:
Ab initio:
• No model assumptions (shell model with inert core,
cluster model, etc., are not ab initio)
Ab initio:
• No model assumptions (shell model with inert core,
cluster model, etc., are not ab initio)
• Ab initio approaches:
•
Faddeev (A  4)
•
hyperspherical (A  6)
•
•
•
Green function’s Monte Carlo (A  13)
no-core shell model (A < 20)
coupled-cluster approach (around closed shells)
Modern NN interaction models:
Modern NN interaction models:
• Realistic (phenomenological) meson-exchange NN
potentials (Nijmegen, Bonn, Argonne)
+ NNN phenomenological potentials
Modern NN interaction models:
• Realistic (phenomenological) meson-exchange NN
potentials (Nijmegen, Bonn, Argonne)
+ NNN phenomenological potentials
• EFT (ChPT) NN potentials
+ NNN EFT (ChPT) potentials
Modern NN interaction models:
• Realistic (phenomenological) meson-exchange NN
potentials (Nijmegen, Bonn, Argonne)
+ NNN phenomenological potentials
• EFT (ChPT) NN potentials
+ NNN EFT (ChPT) potentials
• JISP16 NN interaction
no NNN interaction
fitted to light nuclei
Why would be nice to avoid NNN forces?
Role of NNN force?
•
W. Polyzou and W. Glöckle theorem (Few-body Syst. 9, 97
H=T+Vij  H’=T+V’ij+Vijk,
(1990)):
where Vij and V’ij are phase-equivalent, H and H’ are isospectral.
Hope:
H’=T+V’ij+Vijk
 H=T+Vij
with (approximately) isospectral H and H’ .
JISP type interaction seems to be NN interaction minimizing NNN force.
Without NNN force calculations are simpler, calculations are faster, larger model spaces
become available; hence predictions are more reliable.
JISP
= J-matrix inverse scattering potential
J-matrix formalism:
scattering in the oscillator basis
JISP NN interaction
• NN interaction is a small matrix of the in the
oscillator basis:
9ћΩ truncation, ћΩ = 40 MeV
fast convergence of shell model calculations
• Good description of NN data
JISP16 properties
• 1992 np data base (2514 data): χ2/datum = 1.03
• 1999 np data base (3058 data): χ2/datum = 1.05
PETs
Ambiguity of JISP interaction
• Any unitary transformation of NN Hamiltonian H
generates a Phase-equivalent transformation
(PET).
• Simplest PETs with continuous parameters are
used to fit properties of light nuclei in
No-core Shell Model (NCSM) calculations.
JISP NN interaction
• A. M. Shirokov, A. I. Mazur, S. A. Zaytsev, J. P. Vary,
T. A. Weber, Phys. Rev. C 70, 044005 (2004): A ≤ 4
• A. M. Shirokov, J. P. Vary, A. I. Mazur, S. A. Zaytsev,
T. A. Weber, Phys. Lett. B 621, 96 (2005): A ≤ 6 — JISP6
• A. M. Shirokov, J. P. Vary, A. I. Mazur, T. A. Weber,
Phys. Lett. B 644, 33 (2007): A ≤ 16 — JISP16
Modern NN interaction models:
• Meson-exchange NN potentials (Nijmegen, Bonn, Argonne)
and EFT (ChPT) NN potentials
+ NNN phenomenological or EFT (ChPT) potentials
bad convergence;
effective interaction is needed
• JISP16 NN
interaction
good enough convergence with bare interaction
Our initial approach
From effective interactions to
no-core full configuration calculations
• Extrapolation:
Egs(Nmax) = ae-bNmax + Egs(∞)
• Works with bare interaction only (e.g., JISP16)
• Example:
P. Maris, J. P. Vary, A. M. Shirokov, Phys.
Rev. C 79, 014308 (2009)
Successful prediction:
14F
• 1,990,061,078 basis states
• each ħΩ point requires 2 to 3 hours on 7,626
quad-core compute nodes (30,504 processors
in total) at the Jaguar supercomputer at ORNL
Successful prediction: 14F spectrum
• Deficiency of JISP16 revealed by NCFC
extrapolations
How it looked initially:
How it looks now:
Improved interaction JISP162010
• Obtained by a more accurate fit
to nuclear data using NCFC
Nuclear matter with JISP16
Nuclear matter
• JISP162010 improves NM properties.
• Strong dependence on high partial waves
makes it possible to fit NM to
phenomenological data without violating
description of light nuclei.
Conclusions
• Ab initio NCFC approach based on NCSM is able to describe
light nuclei with A < 20.
• JISP16 provides a good description of NN data and binding
energies, spectra, EM transitions in light nuclei, etc., without
NNN forces.
• An improved version JISP162010 providing a more accurate
description of nuclei is available.
Later this version will be additionally fitted to nuclear matter
too.
• Further development: description of other observables, e.g.,
rms radii in heavy enough nuclei, description of heavier
nuclei, design of charge-dependent version of the interaction.
Thank you!