ab initio method: in-medium no-core shell...
TRANSCRIPT
Ab Initio Method: In-Medium No-Core Shell Model
E. Gebrerufael1 K. Vobig1 H. Hergert2 R. Roth1
1 Institut für Kernphysik, TU Darmstadt 2 NSCL/FRIB Laboratory and Department of Physics & Astronomy, MSU
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 1
12C
accepted on PRL, arXiv:1610.05254
Overview
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 2
No-Core Shell Model (NCSM)
In-Medium Similarity Renormalization Group (IM-SRG)
In-Medium No-Core Shell Model
Results
o Evolution of Ground-State Energy
o Evolution of Excitation Energies
o Spectra
Summary and Outlook
No-Core Shell Model Basics
construct matrix representation of Hamiltonian using basis of HO/HF Slater determinants truncated w.r.t. excitation quanta Nmax
solve large-scale eigenvalue problem for a few smallest eigenvalues
range of applicability limited by factorial growth of basis with Nmax & A
adaptive importance-truncation extends the range of NCSM
… is one of the most powerful exact ab initio methods
for the p- and lower sd-shell
Barrett, Vary, Navratil, ...
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 3
In-Medium Similarity Renormalization Group Basics
flow equation for Hamiltonian:
… uses flow equation for normal-ordered Hamiltonian to decouple
the reference state from its excitations
Tsukiyama, Bogner, Schwenk, Hergert,..
flow parameter generator designed to decouple the reference state
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 4
In-Medium Similarity Renormalization Group Basics
flow equation for Hamiltonian:
… uses flow equation for normal-ordered Hamiltonian to decouple
the reference state from its excitations
H in multi-reference normal-ordered form w.r.t. [Kutzelnigg,Mukherjee]
Tsukiyama, Bogner, Schwenk, Hergert,..
flow parameter generator designed to decouple the reference state
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 4
In-Medium Similarity Renormalization Group Basics
flow equation for Hamiltonian:
… uses flow equation for normal-ordered Hamiltonian to decouple
the reference state from its excitations
H in multi-reference normal-ordered form w.r.t. [Kutzelnigg,Mukherjee]
Tsukiyama, Bogner, Schwenk, Hergert,..
flow parameter generator designed to decouple the reference state
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 4
In-Medium Similarity Renormalization Group Basics
flow equation for Hamiltonian:
… uses flow equation for normal-ordered Hamiltonian to decouple
the reference state from its excitations
H in multi-reference normal-ordered form w.r.t. [Kutzelnigg,Mukherjee]
Tsukiyama, Bogner, Schwenk, Hergert,..
flow parameter generator designed to decouple the reference state
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 4
In-Medium Similarity Renormalization Group Basics
flow equation for Hamiltonian:
… uses flow equation for normal-ordered Hamiltonian to decouple
the reference state from its excitations
H in multi-reference normal-ordered form w.r.t. [Kutzelnigg,Mukherjee]
Tsukiyama, Bogner, Schwenk, Hergert,..
flow parameter generator designed to decouple the reference state
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 4
In-Medium Similarity Renormalization Group Basics
flow equation for Hamiltonian:
… uses flow equation for normal-ordered Hamiltonian to decouple
the reference state from its excitations
H in multi-reference normal-ordered form w.r.t. [Kutzelnigg,Mukherjee]
Tsukiyama, Bogner, Schwenk, Hergert,..
flow parameter generator designed to decouple the reference state
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 4
In-Medium No-Core Shell Model Why should we merge…
+ easy access to heavy nuclei
+ soft computational scaling with A
+ decoupling in A-body space
IM- NCSM
- limited to light nuclei
- factorial growth of model space
- difficult to obtain model-space
convergence
SRG
- not exact method
- only for ground state
- spectroscopy not straight forward
+ exact method
+ easy access to excited states
+ spectroscopy for free
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 5
In-Medium No-Core Shell Model Why should we merge…
+ easy access to heavy nuclei
+ soft computational scaling with A
+ decoupling in A-body space
IM- NCSM
- limited to light nuclei
- factorial growth of model space
- difficult to obtain model-space
convergence
- not exact method
- only for ground state
- spectroscopy not straight forward
+ exact method
+ easy access to excited states
+ spectroscopy for free
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 5
diagonalize Hamiltonian in small model space
ground state defines reference state
In-Medium No-Core Shell Model How should we merge…
evolve Hamiltonian and other operators
pre-diagonalization in A-body space
NCSM define
reference state
IM-SRG evolve
operators
NCSM extract
observables
diagonalize IM-SRG evolved Hamiltonian
obtain eigenstates and extract observables
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 6
In-Medium No-Core Shell Model IM-NCSM is different from …
IM-NCSM is different from IM-SRG for valence-space interactions:
• build on explicit multi-reference formulation
• full no-core approach
• all model-space truncations are converged
NCSM define
reference state
IM-SRG evolve
operators
NCSM extract
observables
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 7
Nmax=0 Slater determinants
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
for sufficiently large flow parameter eigenvalues in Nmax=0, 2 and 4 equal
Nmax=0 eigenstates Slater determinants
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
for sufficiently large flow parameter eigenvalues in Nmax=0, 2 and 4 equal
Nmax=0 eigenstates Slater determinants
IM-SRG decouples reference state from Nmax =2 and 4 spaces
Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamiltonian Matrix in A-Body Basis: 12C
for sufficiently large flow parameter eigenvalues in Nmax=0, 2 and 4 equal
Nmax=0 eigenstates Slater determinants
IM-SRG decouples reference state from Nmax =2 and 4 spaces
Why do E(s) and Nmax=0 eigenvalue differ? Nmax=2 Nmax=4
Nm
ax=4
N
max
=2
Nm
ax=0
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 8
In-Medium No-Core Shell Model Hamilton Matrix in A-Body Basis: 12C
Nmax=0 states couple to
reference state
E(s) and Nmax=0 eigenvalue
not identical
diagonalization of evolved Hamiltonian
necessary
first basis state = reference state
Nmax=0 eigenstates
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 9
Results Evolution of Ground-State Energy
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
drastically enhanced model-space convergence for IM-NCSM
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
drastically enhanced model-space convergence for IM-NCSM
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
drastically enhanced model-space convergence for IM-NCSM
NCSM with explicit 3N
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
drastically enhanced model-space convergence for IM-NCSM
NO2B approximation + induced many-body contribution = 4.0 MeV (≈ 5 %)
NCSM with explicit 3N
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
drastically enhanced model-space convergence for IM-NCSM
NO2B approximation + induced many-body contribution = 4.0 MeV (≈ 5 %)
for s > 0.3 MeV-1 induced many-body contribution becomes significant
NCSM with explicit 3N
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
drastically enhanced model-space convergence for IM-NCSM
NO2B approximation + induced many-body contribution = 4.0 MeV (≈ 5 %)
for s > 0.3 MeV-1 induced many-body contribution becomes significant
NCSM with explicit 3N
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 10
arXiv:1610.05254
Results Evolution of Ground-State Energy
E(s) more robust than in 12C case
NO2B approximation + induced many-body contribution = 2.3 MeV (< 2 %)
NCSM with explicit 3N
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 11
Results NCSM vs. IM-NCSM vs. MR-IM-SRG
NCSM @ Nmax extrap. IM-NCSM @ Nmax=4 MR-IM-SRG (HFB, White) Experiment
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 12
10% <1% 3% 5% 4%
10% largest deviation in 12C
best case 14C
very good agreement
largest deviation in 26O
<2%
arXiv:1610.05254
Results Evolution of Excitation Energies
E* of 2+ increases abruptly at the end due to kink in ground-state energy
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 13
Results Evolution of Excitation Energies
E* of 2+ increases abruptly at the end due to kink in ground-state energy
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 13
Results Evolution of Excitation Energies
E* of 2+ increases abruptly at the end due to kink in ground-state energy
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 13
Results Evolution of Excitation Energies
E* of 2+ increases abruptly at the end due to kink in ground-state energy
Nmax convergence from above in decoupled regime variational principle
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 13
Results Evolution of Excitation Energies
E* of 2+ increases abruptly at the end due to kink in ground-state energy
Nmax convergence from above in decoupled regime variational principle
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 13
Results Evolution of Excitation Energies
E* of 2+ increases abruptly at the end due to kink in ground-state energy
Nmax convergence from above in decoupled regime variational principle
first excited 0 + behaves differently and drops by ≈5 MeV Hoyle state?
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 13
Results Signatures of Hoyle State in 12C
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 14
trends are compatible with Hoyle-state interpretation
experiment
exp. value for Hoyle state 0.4 fm larger than the ground state [Phys. Rev. C, 80:054603]
need better control of induced many-body terms for quantitative statements
exp. value 5.5 e fm2 [Nuclear Physics A, 506(1)]
Results Evolution of Excitation Energies
large dependence on s in Nmax=0
dependence on s reduces with increasing Nmax
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 15
Results Evolution of Excitation Energies
large dependence on s in Nmax=0
dependence on s reduces with increasing Nmax
E* converges monotonically from above for evolved Hamiltonian
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 15
Results Evolution of Excitation Energies
analyze E* as function of Nmax large dependence on s in Nmax=0
dependence on s reduces with increasing Nmax
E* converges monotonically from above for evolved Hamiltonian
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 15
Results Spectra
IM-NCSM
uncertainty band due to flow-parameter variation between smax/2 and smax
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 16
Results Spectra
NCSM IM-NCSM
uncertainty band due to flow-parameter variation between smax/2 and smax
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 16
Results Spectra
NCSM IM-NCSM
uncertainty band due to flow-parameter variation between smax/2 and smax
2+ and 1+ in IM-NCSM and NCSM in good agreement
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 16
Results Spectra
NCSM IM-NCSM
uncertainty band due to flow-parameter variation between smax/2 and smax
2+ and 1+ in IM-NCSM and NCSM in good agreement
second 0+ in NCSM (Hoyle?) slow convergence, IM-NCSM closer to experiment
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 16
Results Spectra
NCSM IM-NCSM
uncertainty band due to flow-parameter variation between smax/2 and smax
2+ and 1+ in IM-NCSM and NCSM in good agreement
second 0+ in NCSM (Hoyle?) slow convergence, IM-NCSM closer to experiment
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 16
Results Spectra
first 2+ and 4+ robust and well converged in IM-NCSM
higher-lying states show small flow-parameter dependence
1+ not yet observed experimentally theoretical prediction
NCSM IM-NCSM
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 17
Summary
established a many-body technique IM-NCSM = IM-SRG + NCSM
IM-SRG decouples reference state from higher Nmax
extremely enhanced Nmax convergence for subsequent NCSM
Nmax≤4 sufficient to extract converged ground-state energies
variational principle becomes valid for excitation energies
since ground-state energy converged
preliminary ab initio studies regarding Hoyle state in 12C
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 18
Outlook
o more detailed analysis of the Hoyle state in 12C
o study of exotic nuclei: island-of-inversion physics, …
o evolve vector operator, for instance E1 transition operators
o extend applicability of IM-NCSM to odd nuclei
using particle-attached particle-removed formalism
o …
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 19
Thanks to my group & collaborator
• S. Alexa, T. Hüther, L. Mertes, R. Roth, S. Schulz,
H. Spielvogel, C. Stumpf, A. Tichai, K. Vobig, R. Wirth
Institut für Kernphysik, TU Darmstadt
• H. Hergert
NSCL/FRIB, Michigan State University
Thank You For Your Attention
JURECA LOEWE-CSC LICHTENBERG
COMPUTING TIME
Eskendr Gebrerufael - TU Darmstadt - Apr. 2017 20