Discrete Non-Orthogonal Shell Model

Updated 6 July 2026

Discrete Non-Orthogonal Shell Model is a variational shell-model approach that expands low-lying nuclear states using a compact set of non-orthogonal, projected intrinsic configurations.
It employs projection techniques to restore angular momentum and parity, drastically reducing the configuration space compared to traditional shell-model diagonalization.
The framework addresses the shell-model secular problem, offering efficient and precise solutions for low-energy nuclear states and facilitating studies of heavy and exotic nuclei.

Discrete Non-Orthogonal Shell Model (DNO-SM, or DNOSM) denotes a variational shell-model framework in which low-lying nuclear states are expanded in a discrete set of non-orthogonal Slater determinants, usually with angular-momentum and parity restoration by projection, instead of the very large orthonormal configuration-interaction basis of conventional shell-model diagonalization. In its recent formulations, the method is presented as a way to address the shell-model secular problem by replacing a huge orthonormal valence-space basis with a compact set of physically optimized intrinsic configurations, while retaining a shell-model Hamiltonian in a shell-model valence space (Dao et al., 2022, Dao et al., 11 Jul 2025).

1. Conceptual definition and scope

DNO-SM is a shell-model method because it works with a shell-model effective Hamiltonian in a chosen valence space, but it departs from standard configuration interaction in the representation of the many-body wave function. Instead of expanding the state in the full orthonormal basis of spherical Slater determinants, it uses a small discrete set of non-orthogonal intrinsic states. These are then projected onto good quantum numbers and mixed through a generalized eigenvalue problem. The method is therefore positioned at the intersection of shell-model configuration interaction, symmetry restoration, projected Hartree-Fock-type methods, and Hill-Wheeler-Griffin configuration mixing (Dao et al., 25 Nov 2025).

A central formal motivation is the Broeckhove–Deumens theorem, invoked in DNO-SM work as stating the existence of a discrete set of non-orthogonal wavefunctions spanning the relevant shell-model space. In the finite-dimensional shell-model context, this implies that one should be able to find a finite non-orthogonal set that spans the same valence-space Hilbert space used in conventional shell-model calculations. DNO-SM is explicitly presented as a numerical realization of that idea for low-lying states of interest (Dao et al., 11 Jul 2025).

The method has been developed in stages. “Nuclear Structure with Discrete Non-Orthogonal Shell-Model : new frontiers” (Dao et al., 2022) formulated DNO-SM as diagonalization of shell-model Hamiltonians in a discrete non-orthogonal basis, emphasized basis-state selection optimization, and described its implementation in CARINA. “Exact solutions of the nuclear shell-model secular problem: Discrete Non-Orthogonal Shell Model within a Variation After Projection approach” (Dao et al., 11 Jul 2025) sharpened the claim to exact shell-model solutions for low-lying states in benchmark cases using Variation After Projection (VAP). “Discrete non-orthogonal shell model for nuclear structure: Towards heavy elements” (Dao et al., 25 Nov 2025) extended the same framework toward proton-rich and superheavy nuclei.

2. Variational formalism and generalized eigenvalue problem

In DNO-SM, the starting Hamiltonian is the standard shell-model Hamiltonian

$\hat H = \sum_{ij} t_{ij} c^\dagger_i c_j + \frac 1 4 \sum_{ijkl} V_{ijkl} c^\dagger_i c^\dagger_j c_l c_k ,$

with antisymmetrized two-body matrix elements $V_{ijkl}$ (Dao et al., 11 Jul 2025).

The many-body state is written as a projected non-orthogonal expansion. In the projection-after-variation formulation,

$\begin{aligned} | {\psi^{\pi JM}_n} \rangle = \sum_{q,K}C^{\pi J}_{n,qK} \mathcal P^J_{MK}\:P^\pi | {\phi_q}\rangle + \sum_{q,K,N}C^{\pi J}_{n,qK,N} \mathcal P^J_{MK}\:P^\pi | {\phi_q(N\mathrm{p}N\mathrm{h})}\rangle , \end{aligned}$

where $|\phi_q\rangle$ are intrinsic Slater determinants, $|\phi_q(N\mathrm{p}N\mathrm{h})\rangle$ are explicit particle-hole excitations built on them, $\mathcal P^J_{MK}$ is the angular-momentum projector, and $P^\pi$ is the parity projector. In the variation-after-projection formulation, the central ansatz is the simpler projected determinant expansion

$| {\psi^{\pi JM}_n} \rangle = \sum_{q,K}C^{\pi J}_{n,qK} \mathcal P^J_{MK}\:P^\pi | {\phi_q}\rangle .$

The coefficients are determined by a generalized secular equation,

$\mathcal H^{\pi J} C_n^{\pi J} = E_n^{\pi J}\,\mathcal N^{\pi J} C_n^{\pi J},$

with projected Hamiltonian and norm kernels

$\mathcal H^{\pi J}_{qK,q'K'}= \langle \phi_q| \hat H\, \mathcal P^J_{KK'} P^\pi|\phi_{q'}\rangle, \qquad \mathcal N^{\pi J}_{qK,q'K'}= \langle \phi_q| \mathcal P^J_{KK'} P^\pi|\phi_{q'}\rangle .$

This is the characteristic non-orthogonal shell-model structure: $V_{ijkl}$ 0, and the diagonalization is performed in a basis with a nontrivial overlap metric (Dao et al., 11 Jul 2025).