Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
11institutetext: IRFU, CEA, Université Paris-Saclay, 91191 Gif-sur-Yvette, France 22institutetext: KU Leuven, Department of Physics and Astronomy, Instituut voor Kern- en Stralingsfysica, 3001 Leuven, Belgium 33institutetext: CEA, DES, IRESNE, DER, SPRC, LEPh, 13115 Saint-Paul-lez-Durance, France 44institutetext: Technische Universität Darmstadt, Department of Physics, 64289 Darmstadt, Germany 55institutetext: ExtreMe Matter Institute EMMI, GSI Helmholtzzentrum für Schwerionenforschung GmbH, 64291 Darmstadt, Germany 66institutetext: Max-Planck-Institut für Kernphysik, 69117 Heidelberg, Germany

Impact of correlations on nuclear binding energies

Ab initio calculations of singly and doubly open-shell nuclei
A. Scalesi\thanksrefad:saclay    T. Duguet\thanksrefad:saclay,ad:kul    P. Demol\thanksrefad:kul    M. Frosini\thanksrefad:des    V. Somà\thanksrefad:saclay    A. Tichai\thanksrefad:ger1,ad:ger2,ad:ger3
Abstract

A strong effort will be dedicated in the coming years to extend the reach of ab initio nuclear-structure calculations to heavy doubly open-shell nuclei. In order to do so, the most efficient strategies to incorporate dominant many-body correlations at play in such nuclei must be identified. With this motivation in mind, the present work pedagogically analyses the inclusion of many-body correlations and their impact on binding energies of Calcium and Chromium isotopes.

Employing an empirically-optimal Hamiltonian built from chiral effective field theory, binding energies along both isotopic chains are studied via a hierarchy of approximations based on polynomially-scaling expansion many-body methods. More specifically, calculations are performed based on (i) the spherical Hartree-Fock-Bogoliubov mean-field approximation plus correlations from second-order Bogoliubov many-body perturbation theory or Bogoliubov coupled cluster with singles and doubles on top of it, along with (ii) the axially-deformed Hartree-Fock-Bogoliubov mean-field approximation plus correlations from second-order Bogoliubov many-body perturbation theory built on it. The corresponding results are compared to experimental data and to those obtained via valence-space in-medium similarity renormalization group calculations at the normal-ordered two-body level that act as a reference in the present study.

The spherical mean-field approximation is shown to display specific shortcomings in Ca isotopes that can be understood analytically and that are efficiently corrected via the consistent addition of low-order dynamical correlations on top of it. While the same setting cannot appropriately reproduce binding energies in doubly open-shell Cr isotopes, allowing the unperturbed mean-field state to break rotational symmetry permits to efficiently capture the static correlations responsible for the phenomenological differences observed between the two isotopic chains.

Eventually, the present work demonstrates in a pedagogical way that polynomially-scaling expansion methods based on unperturbed states that possibly break (and restore) symmetries constitute an optimal route to extend ab initio calculations to heavy closed- and open-shell nuclei.

1 Introduction

Predictions based on ab initio nuclear structure calculations are currently moving to heavier systems Hu et al. (2021); Miyagi et al. (2022); Hebeler et al. (2023); Tichai et al. (2024); Arthuis et al. (2024) and/or doubly open-shell nuclei Novario et al. (2020); Hagen et al. (2022); Frosini et al. (2022a, 2024). One ambition of such developments is to efficiently capture the dominant many-body correlations at play. Qualitatively speaking, many-body correlations separate into two different categories. The first category concerns so-called dynamical correlations carried by all nucleons and delivering the bulk of the correlation energy. Dynamical correlations are well captured by a sum of many low-rank elementary, e.g. particle-hole, excitations out of a well-chosen unperturbed state. The second category concerns so-called static correlations that strongly impact the ground-state of open-shell nuclei and are driven by the valence nucleons. While being largely subleading, static correlations vary quickly with the number of valence nucleons and, as such, strongly impact differential quantities as well as spectroscopic observables. Such correlations can be efficiently captured via an optimal choice of the unperturbed state Tichai et al. (2020); Frosini et al. (2022b).

In this context, the present work wishes to pedagogically analyse the impact of many-body correlations on binding energies and associated differential quantities, i.e. first- and second-order derivatives with respect to the (even) neutron number. To do so, the study is conducted along neighboring Calcium (Z=20𝑍20Z=20italic_Z = 20) and Chromium (Z=24𝑍24Z=24italic_Z = 24) isotopic chains spanning a large range of (even) neutron numbers from N=12𝑁12N=12italic_N = 12 till N=50𝑁50N=50italic_N = 50. Most of Ca isotopes are of singly open-shell character whereas most of Cr isotopes are of doubly open-shell character. Comparing the behavior of binding energies along these isotopic chains allows one to illustrate the roles played by static and dynamical correlations in the two types of nuclei and the capacity of ab initio many-body methods to efficiently capture them by employing an optimal formulation. In order to control how some of the identified features depend on the nuclear mass, additional calculations are performed along the Tin (Z=50𝑍50Z=50italic_Z = 50) isotopic chain from N=50𝑁50N=50italic_N = 50 till N=82𝑁82N=82italic_N = 82.

The paper is organized as follows. Section 2 briefly characterises the numerical calculations performed in the present study. In Sec. 3, the results obtained at the spherical mean-field level are analysed, pointing to specific deficiencies that need to be remedied by the addition of correlations. In Sec. 4, low-order dynamical correlations on top of the spherical mean-field are proven to correct all such shortcomings to a high degree in Ca isotopes. In Sec. 5, the inclusion of static correlations either via a complete diagonalization in the valence space or via the explicit breaking of rotational symmetry is shown to be critical to obtain an equally good description of Cr isotopes. The paper is complemented by an appendix in which semi-analytical formulae are derived to provide a more intuitive understanding of the numerical results.

Refer to caption
Figure 1: Systematics along Ca (left) and Cr (right) isotopic chains. First line: sHFB absolute binding energies against experimental data. Second (third) line: sHFB and sHF-EFA two-neutron separation energy (two-neutron shell gap) against experimental data. Between 42Ca and 48Ca (1f7/2 shell) as well as between 62Ca and 70Ca (1g9/2 shell), sHFB-ZP semi-analytical results are also shown. Fourth line: sHFB neutron-number variance against the minimal possible variance in sHFB calculations Duguet et al. (2020). The sequence of the underlying neutron canonical shells are also displayed.

2 Numerical calculations

Ab initio many-body calculations are carried out employing a one-body spherical harmonic oscillator basis characterized by the frequency ω=12Planck-constant-over-2-pi𝜔12\hbar\omega=12roman_ℏ italic_ω = 12 MeV. All states up to emaxmax(2n+l)=12subscript𝑒maxmax2𝑛𝑙12e_{\!{}_{\;\text{max}}}\equiv\text{max}(2n+l)=12italic_e start_POSTSUBSCRIPT start_FLOATSUBSCRIPT max end_FLOATSUBSCRIPT end_POSTSUBSCRIPT ≡ max ( 2 italic_n + italic_l ) = 12 are included, with n𝑛nitalic_n the principal quantum number and l𝑙litalic_l the orbital angular momentum. The representation of three-body operators is further restricted by only employing three-body states up to e3max=18(24)subscript𝑒3max1824e_{\!{}_{\;\text{3max}}}=18\,(24)italic_e start_POSTSUBSCRIPT start_FLOATSUBSCRIPT 3max end_FLOATSUBSCRIPT end_POSTSUBSCRIPT = 18 ( 24 ) in Ca and Cr (Sn) isotopes.

Calculations are performed using the EM 1.8/2.0 Hamiltonian from Ref. Hebeler et al. (2011) containing two-nucleon (2N) and three-nucleon (3N) interactions originating from chiral effective field theory (χ𝜒\chiitalic_χEFT). The 3N interaction is approximated via the rank-reduction method developed in Ref. Frosini et al. (2021). This particular Hamiltonian is employed because it is empirically known to give an excellent reproduction of binding energies in the mid-mass region Stroberg et al. (2021).

The present study is based on three complementary expansion many-body methods. First, the spherical Hartree-Fock Bogoliubov (sHFB) mean-field approximation plus second-order Bogoliubov many-body perturbation theory (sBMBPT(2)) correction Tichai et al. (2018); Arthuis et al. (2019) is employed. As a non-perturbative complement to sBMBPT(2), spherical Bogoliubov coupled cluster with singles and doubles (sBCCSD) Signoracci et al. (2015); Tichai et al. (2024) calculations are also carried out. Third comes the axially-deformed Hartree-Fock Bogoliubov (dHFB) mean-field approximation plus second-order Bogoliubov many-body perturbation theory (dBMBPT(2)) correction Frosini et al. (2021, 2024) .

Available valence-space in-medium similarity renormalization group (VS-IMSRG(2)) results in Ca and Cr isotopes Stroberg et al. (2021) based on the same Hamiltonian111The numerical setting is slightly different given that these calculations employ emax=14subscript𝑒max14e_{\!{}_{\;\text{max}}}=14italic_e start_POSTSUBSCRIPT start_FLOATSUBSCRIPT max end_FLOATSUBSCRIPT end_POSTSUBSCRIPT = 14 and e3max=16subscript𝑒3max16e_{\!{}_{\;\text{3max}}}=16italic_e start_POSTSUBSCRIPT start_FLOATSUBSCRIPT 3max end_FLOATSUBSCRIPT end_POSTSUBSCRIPT = 16 while extrapolations in emaxsubscript𝑒maxe_{\!{}_{\;\text{max}}}italic_e start_POSTSUBSCRIPT start_FLOATSUBSCRIPT max end_FLOATSUBSCRIPT end_POSTSUBSCRIPT are performed to obtain infrared convergence Furnstahl et al. (2015). The effects of 3N interactions between valence nucleons is captured via the ensemble normal-ordering of Ref. Stroberg et al. (2017). are presently employed as a reference given that static and dynamical correlations generated within the valence space are accounted for to all orders via the diagonalization of the associated effective Hamiltonian. Notice that calculations along complete Ca and Cr isotopic chains require a reset of the valence space below N=20𝑁20N=20italic_N = 20 and above N=40𝑁40N=40italic_N = 40. The data presently employed correspond to the choice of valence spaces delivering the most optimal results Stroberg et al. (2021).

3 Spherical mean-field approximation

The baseline of more advanced treatment based on a many-body expansion is given by the mean-field approximation restricted to spherical symmetry Hergert (2020). Because the present study targets open-shell systems, the minimal version presently considered is given by sHFB that can naturally capture pairing correlations via the breaking of U(1) symmetry associated with particle-number conservation Ring and Schuck (1980).

3.1 Ca chain

Systematic sHFB results along the Ca isotopic chain are displayed in the left panels of Fig. 1. In the first line, one observes that sHFB calculations significantly underbind experimental data, e.g. by more than 100100100100 MeV in 48Ca, in a way that increases with neutron excess. Such a quantitative defect is expected from a mean-field approximation in the context of ab initio calculations. Indeed, while static neutron-neutron pairing correlations are incorporated within sHFB, one is missing dynamical correlations whose inclusion account for a significant fraction of the binding energy Hergert (2020); Tichai et al. (2020); Frosini et al. (2022a); Duguet et al. (2023).

The evolution of binding energies can be scrutinized via the two-neutron separation energy

S2n(N,Z)E(N2,Z)E(N,Z)subscript𝑆2𝑛𝑁𝑍𝐸𝑁2𝑍𝐸𝑁𝑍\displaystyle S_{2n}(N,Z)\equiv E(N-2,Z)-E(N,Z)italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) ≡ italic_E ( italic_N - 2 , italic_Z ) - italic_E ( italic_N , italic_Z ) (1)

displayed in the second line of Fig. 1. Because S2n(N,Z)subscript𝑆2𝑛𝑁𝑍S_{2n}(N,Z)italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) is a first derivative of the binding energy E(N,Z)𝐸𝑁𝑍E(N,Z)italic_E ( italic_N , italic_Z ) with respect to (even) neutron number, the large offset seen in the first line has disappeared. Eventually, the S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT from sHFB slightly underestimate experimental data overall such that adding dynamical correlations is expected to correct for this quantitative discrepancy.

The main characteristics of the experimental S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT, i.e. the sudden drops at N=20𝑁20N=20italic_N = 20 and 28282828, and to a lesser extent at N=32𝑁32N=32italic_N = 32 and 34343434, as well as the smooth evolution in between, are well accounted for by sHFB results. However, crucial differences are revealed upon closer inspection. First, the amplitude of the drops at N=20𝑁20N=20italic_N = 20 and 28282828 is too large and the trend in between, i.e. while filling the 1f7/2 shell, is qualitatively wrong. Correlated with the too large drop at N=20𝑁20N=20italic_N = 20, the S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT value in 42Ca is significantly too low. Further adding neutrons, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT increases linearly throughout the 1f7/2 shell instead of decreasing linearly as for experimental data222While the same patterns are at play when going through 48Ca, the size of the 2p3/2 shell is too small to make the rising slope of sHFB results really visible. The highly degenerate 1g9/2 shell between 62Ca and 70Ca is more favorable in this respect even though the slope of the sHFB results is actually zero in this case. These nuclei are anyway predicted to be unbound and there is no experimental data yet to be confronted with..

Given that S2n(N,Z)subscript𝑆2𝑛𝑁𝑍S_{2n}(N,Z)italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) is the first derivative of the binding energy, the patterns identified above relate to specific features of the binding energies that could not be fully appreciated in the first line of Fig. 1 due to the large scale employed. The fact that S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT evolves linearly with the number avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT of nucleons in the valence shell for both sHFB results and experiment data implies that E(N,Z)𝐸𝑁𝑍E(N,Z)italic_E ( italic_N , italic_Z ) is essentially quadratic in between two closed-shell isotopes. The fact that S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT starts from a too low value in sHFB calculations in the open-shell relates to the fact that the linear decrease of E(N,Z)𝐸𝑁𝑍E(N,Z)italic_E ( italic_N , italic_Z ) is not pronounced enough such that the difference to the data increases throughout the shell. Finally, the fact S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is rising linearly instead of decreasing linearly indicates that the sHFB energy is concave instead of being convex.

Refer to caption
Figure 2: Energy curvature of Ca (left) and Cr (right) isotopes between N=20𝑁20N=20italic_N = 20 and N=28𝑁28N=28italic_N = 28 (1f7/2 shell). Upper panels: energies rescaled to N=20𝑁20N=20italic_N = 20 and N=28𝑁28N=28italic_N = 28 (see text for details). Bottom panels: two-neutron shell gap Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT. Experimental data are compared to sHFB results. For Ca isotopes, results from sHF-EFA and sHFB-ZP semi-analytical results are also shown.

These characteristics can be pinned down quantitatively by looking at the third line of Fig. 1 displaying the so-called two-neutron shell gap

Δ2n(N,Z)S2n(N,Z)S2n(N+2,Z).subscriptΔ2𝑛𝑁𝑍subscript𝑆2𝑛𝑁𝑍subscript𝑆2𝑛𝑁2𝑍\displaystyle\Delta_{2n}(N,Z)\equiv S_{2n}(N,Z)-S_{2n}(N+2,Z)\,.roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) ≡ italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) - italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N + 2 , italic_Z ) . (2)

Whenever Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT displays a sudden increase, the amplitude of the spike provides an empirical measure of the extra stability associated with a mean-field picture of a closed-shell nucleus displaying a large Fermi gap. Otherwise, Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is linked to the second derivative, i.e. the curvature, of the smoothly evolving binding energy (see B for details).

The left panel displaying Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT in Fig. 1 confirms the two patterns identified above. First, the amplitude of the spikes at N=20𝑁20N=20italic_N = 20 and N=28𝑁28N=28italic_N = 28 are too large by 4.04.04.04.0 and 2.12.12.12.1 MeV, respectively333Contrarily, the sudden increase is correctly reproduced for N=32𝑁32N=32italic_N = 32 and N=34𝑁34N=34italic_N = 34. It seems that the larger the over-stability in the data, i.e. the more pronounced the magic character of the isotope, the larger the sHFB exaggeration.. Second, the essentially constant character of the experimental Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT between 42Ca and 48Ca is well captured by sHFB calculations but the associated value is negative instead of positive, i.e. to a very good approximation the sHFB energy is indeed quadratic with the number of valence nucleons avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT but it is concave instead of being convex444Between 62Ca and 70Ca, where there is no experimental data, Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is constant but actually null such that the sHFB energy is rather linear with avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT..

Eventually, the issue associated with the curvature of the energy can be even better appreciated from the left panels of Fig. 2 focusing on the isotopes between N=20𝑁20N=20italic_N = 20 and N=28𝑁28N=28italic_N = 28. While the bottom panel shows Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT, the upper panel displays the total energy rescaled to N=20𝑁20N=20italic_N = 20 and rotated around that point such that the value at N=28𝑁28N=28italic_N = 28 is aligned with it. This effectively removes the overall shift between the different curves along with the linear trend between the two closed-shell isotopes. Both panels make clear that, while experimental energies of Ca isotopes are essentially quadratic and convex between two closed-shell isotopes, sHFB calculations generate a quadratic dependence of energies whose curvature carries the wrong sign.

3.2 Analytical investigation

The wrong qualitative behavior of the sHFB energy along semi-magic chains was already visible in past calculations Tichai et al. (2018); Somà et al. (2021); Tichai et al. (2024) based on different chiral Hamiltonians. It seems to indicate that this behavior is deeply rooted into the spherical mean-field approximation based on realistic nuclear Hamiltonians. This expectation can in fact be confirmed analytically as demonstrated below.

In order to proceed, one must first make a crucial observation thanks to the results shown on the last line of Fig. 1 comparing the neutron-number variance in the sHFB calculation to the minimal variance obtained in the zero-pairing limit of sHFB theory (sHFB-ZP) Duguet et al. (2020). As already noticed Somà et al. (2021), chiral Hamiltonians typically generate only little static pairing at the mean-field level555In an ab initio setting, pairing properties such as the odd-even mass staggering are expected to largely originate from (i.e. required to account for) higher-order processes associated with the exchange of collective medium fluctuations between paired particles Barranco et al. (2004); Gori et al. (2005); Pastore et al. (2008); Idini et al. (2011). Achieving a quantitative description of pairing properties from first principles constitutes a major challenge for ab initio nuclear structure theory Somà et al. (2021)., i.e. the computed neutron-number variance is indeed very close to the minimal variance in most open-shell isotopes, except in 56,58Ca and for nuclei in the continuum. As visible from the left panels of Fig. 1, this is confirmed by the proximity of sHFB results to those obtained from spherical Hartree-Fock calculations performed within the equal-filling approximation Perez-Martin and Robledo (2008) (sHF-EFA) that do not include pairing correlations by construction. Results are indeed very close overall, except in 56,58Ca (1f5/2 shell) where sHFB better reproduce experimental values for S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT. As for the curvature within open-shells, the left panels of Fig. 2 reveals that the curvature of sHF-EFA results also carries the wrong sign but is such that the concavity is even more pronounced than for sHFB, i.e. the weak pairing correlations present within the 1f7/2 shell in sHFB do improve the situation compared to the case where pairing would indeed be strictly zero.

Based on this observation, the sHFB energy of an open-shell nucleus relative to the closed-shell (CS) core666The following considerations can be meaningfully applied only to singly open-shell nuclei, since they rely on the existence of a spherical core and highly-degenerate shells on top of it. can be, to a good approximation, expressed analytically as a function of avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT and of specific 2N and 3N interaction matrix elements within sHFB-ZP and sHF-EFA. Both cases are worked out in details in A. Since both variants provide almost identical numerical results, only the simpler sHF-EFA expressions are reported here whereas the complete set of formulae valid in sHFB-ZP can be found in A.

Canonical single-particle states k(nk,lk,jk,mk,τk)𝑘subscript𝑛𝑘subscript𝑙𝑘subscript𝑗𝑘subscript𝑚𝑘subscript𝜏𝑘k\equiv(n_{k},l_{k},j_{k},m_{k},\tau_{k})italic_k ≡ ( italic_n start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_l start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_j start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_m start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_τ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ) diagonalizing the one-body density matrix ρsHF-EFAsuperscript𝜌sHF-EFA\rho^{\text{sHF-EFA}}italic_ρ start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT gather in shells carrying degeneracy dk2jk+1subscript𝑑𝑘2subscript𝑗𝑘1d_{k}\equiv 2j_{k}+1italic_d start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ≡ 2 italic_j start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT + 1 characterized by the single-particle energies ϵk=ϵk˘subscriptitalic-ϵ𝑘subscriptitalic-ϵ˘𝑘\epsilon_{k}=\epsilon_{\breve{k}}italic_ϵ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT = italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT (see Eq. 5 below) where k˘(nk,lk,jk,τk)˘𝑘subscript𝑛𝑘subscript𝑙𝑘subscript𝑗𝑘subscript𝜏𝑘\breve{k}\equiv(n_{k},l_{k},j_{k},\tau_{k})over˘ start_ARG italic_k end_ARG ≡ ( italic_n start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_l start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_j start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_τ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ). For a system with A (even) nucleons, these shells separate into three categories in sHF-EFA

  1. 1.

    ϵh˘subscriptitalic-ϵ˘\epsilon_{\breve{h}}italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_h end_ARG end_POSTSUBSCRIPT denoting “hole states”,

  2. 2.

    ϵv˘subscriptitalic-ϵ˘𝑣\epsilon_{\breve{v}}italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT denoting “valence states”,

  3. 3.

    ϵp˘subscriptitalic-ϵ˘𝑝\epsilon_{\breve{p}}italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_p end_ARG end_POSTSUBSCRIPT denoting “particle states” ,

such that Aav𝐴subscript𝑎𝑣A-a_{v}italic_A - italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT nucleons fill the hole states whereas 0<avdv0subscript𝑎𝑣subscript𝑑𝑣0<a_{v}\leq d_{v}0 < italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ≤ italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT nucleons occupy the valence shell.

Given this setting, one eventually obtains the total energy of an open-shell nucleus relative to the CS core along with the corresponding two-neutron separation energy and two-neutron shell gap as

ΔEsHF-EFA(av)Δsuperscript𝐸sHF-EFAsubscript𝑎𝑣\displaystyle\Delta E^{\text{sHF-EFA}}(a_{v})roman_Δ italic_E start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) EsHF-EFA(av)EsHF-EFA(0)absentsuperscript𝐸sHF-EFAsubscript𝑎𝑣superscript𝐸sHF-EFA0\displaystyle\equiv E^{\text{sHF-EFA}}(a_{v})-E^{\text{sHF-EFA}}(0)≡ italic_E start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) - italic_E start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT ( 0 )
=αv˘av+βv˘2av2,absentsubscript𝛼˘𝑣subscript𝑎𝑣subscript𝛽˘𝑣2subscriptsuperscript𝑎2𝑣\displaystyle=\alpha_{\breve{v}}a_{v}+\frac{\beta_{\breve{v}}}{2}a^{2}_{v}\,,= italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT + divide start_ARG italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG start_ARG 2 end_ARG italic_a start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT , (3a)
S2nsHF-EFA(av)subscriptsuperscript𝑆sHF-EFA2𝑛subscript𝑎𝑣\displaystyle S^{\text{sHF-EFA}}_{2n}(a_{v})italic_S start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =2αv˘2βv˘(av1),absent2subscript𝛼˘𝑣2subscript𝛽˘𝑣subscript𝑎𝑣1\displaystyle=-2\alpha_{\breve{v}}-2\beta_{\breve{v}}(a_{v}-1)\,,= - 2 italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - 2 italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT - 1 ) , (3b)
Δ2nsHF-EFA(av)subscriptsuperscriptΔsHF-EFA2𝑛subscript𝑎𝑣\displaystyle\Delta^{\text{sHF-EFA}}_{2n}(a_{v})roman_Δ start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =4βv˘.absent4subscript𝛽˘𝑣\displaystyle=4\beta_{\breve{v}}\,.= 4 italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT . (3c)

with

αv˘subscript𝛼˘𝑣\displaystyle\alpha_{\breve{v}}italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT =ϵv˘CSabsentsubscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle=\epsilon^{\text{CS}}_{\breve{v}}= italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT
tvv+hv¯vhvh+12hhw¯vhhvhh,absentsubscript𝑡𝑣𝑣subscriptsubscript¯𝑣𝑣𝑣12subscriptsuperscriptsubscript¯𝑤𝑣superscript𝑣superscript\displaystyle\equiv t_{vv}+\sum_{h}\overline{v}_{vhvh}+\frac{1}{2}\sum_{hh^{% \prime}}\overline{w}_{vhh^{\prime}vhh^{\prime}}\,,≡ italic_t start_POSTSUBSCRIPT italic_v italic_v end_POSTSUBSCRIPT + ∑ start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_v italic_h italic_v italic_h end_POSTSUBSCRIPT + divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (4a)
βv˘subscript𝛽˘𝑣\displaystyle\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT =1dvmvdv(v¯vvvv+hw¯vvhvvh)absent1subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝑣𝑣superscript𝑣𝑣superscript𝑣subscriptsubscript¯𝑤𝑣superscript𝑣𝑣superscript𝑣\displaystyle=\frac{1}{d_{v}}\sum_{m_{v^{\prime}}}^{d_{v}}\left(\overline{v}_{% vv^{\prime}vv^{\prime}}+\sum_{h}\overline{w}_{vv^{\prime}hvv^{\prime}h}\right)= divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ( over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + ∑ start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h end_POSTSUBSCRIPT )
1dvmvdv𝕧¯vvvv.absent1subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\displaystyle\equiv\frac{1}{d_{v}}\sum_{m_{v^{\prime}}}^{d_{v}}\overline{% \mathbb{v}}_{vv^{\prime}vv^{\prime}}\,.≡ divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT . (4b)

Equation (3a) proves that the sHF-EFA energy is indeed quadratic777As shown in A, the 3N interaction actually induces the presence of a cubic term in the energy. However, present numerical applications demonstrate that it is negligible for all nuclei under consideration such that it can be dropped altogether in the present discussion. in the number of valence nucleons throughout any given open-shell. The coefficient αv˘subscript𝛼˘𝑣\alpha_{\breve{v}}italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT of the linear term is nothing but the mean-field single-particle energy of the valence shell computed in the CS core ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT, whose interaction energy contributions are displayed diagrammatically in Fig. 3. The coefficient βv˘subscript𝛽˘𝑣\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT of the quadratic term, i.e. the curvature, of the energy is given by the average over the valence magnetic substates of the diagonal valence-shell two-body matrix elements888As seen from Eq. (4b), 𝕧¯vvvvsubscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\overline{\mathbb{v}}_{vv^{\prime}vv^{\prime}}over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT includes the effective contribution obtained by averaging the 3N interaction over the CS core. 𝕧¯vvvvsubscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\overline{\mathbb{v}}_{vv^{\prime}vv^{\prime}}over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT displayed diagrammatically in Fig. 4. Such an averaging corresponds to the monopole valence-shell matrix element per valence state. As visible from Eq. (3b), 2ϵv˘CS2subscriptsuperscriptitalic-ϵCS˘𝑣-2\epsilon^{\text{CS}}_{\breve{v}}- 2 italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT sets the initial value of S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT999As seen from Tab. 1, the relation |αv˘||βv˘|much-greater-thansubscript𝛼˘𝑣subscript𝛽˘𝑣|\alpha_{\breve{v}}|\gg|\beta_{\breve{v}}|| italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT | ≫ | italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT | holds in practice such that the starting value of S2nsHF-EFAsubscriptsuperscript𝑆sHF-EFA2𝑛S^{\text{sHF-EFA}}_{2n}italic_S start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT (av=2subscript𝑎𝑣2a_{v}=2italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT = 2) in the open shell is essentially dictated by ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT. whereas 2βv˘2subscript𝛽˘𝑣-2\beta_{\breve{v}}- 2 italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT drives its linear evolution throughout the open-shell. Eventually, Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT extracts 4βv˘4subscript𝛽˘𝑣4\beta_{\breve{v}}4 italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT.

Refer to caption
Figure 3: First-order interaction energy contributions to the valence-shell single-particle energy ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT computed in the closed-shell core.
Refer to caption
Figure 4: First-order contributions to the valence-shell effective two-body matrix elements 𝕧¯vvvvsubscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\overline{\mathbb{v}}_{vv^{\prime}vv^{\prime}}over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT.
Refer to caption
Figure 5: Neutron canonical single-particle energies ϵk˘subscriptitalic-ϵ˘𝑘\epsilon_{\breve{k}}italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT from sHFB calculations along the Ca isotopic chain. Semi-analytical sHFB-ZP results for ϵ1f7/2subscriptitalic-ϵ1subscriptf72\epsilon_{1\text{f}_{7/2}}italic_ϵ start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT are also shown between 40Ca and 48Ca.

Extracting ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT and 𝕧¯vvvvsubscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\overline{\mathbb{v}}_{vv^{\prime}vv^{\prime}}over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT numerically from the presently employed chiral Hamiltonian (see Tab. 1), the semi-analytical results from Eqs. (3)-(4) (in fact of their sHFB-ZP counterparts; see A) are superimposed on the left panels of Fig. 1 between 42Ca and 48Ca (1f7/2 valence shell) as well as between 62Ca and 70Ca (1g9/2 valence shell). The results perfectly match the numerical sHF-EFA curves, that are themselves very close to sHFB results. Looking at the left panels of Fig. 2, one indeed sees the fully quantitative agreement between sHF-EFA and the semi-analytical results.

Open shell αv˘subscript𝛼˘𝑣\alpha_{\breve{v}}italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT (MeV) βv˘subscript𝛽˘𝑣\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT (MeV)
1f7/2 -6.005 -0.290
1g9/2 2.976 -0.270
Table 1: Coefficient of the linear and quadratic term of the HFB-ZP energy (Eq. (17)-(18)) extracted numerically for two neutron valence shells along the Ca isotopic chain using the EM 1.8/2.0 Hamiltonian Hebeler et al. (2011). The coefficient of the cubic term γv˘subscript𝛾˘𝑣\gamma_{\breve{v}}italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT is numerically zero in all cases.

The semi-analytical results first clarify that, in an ab initio setting, the reason why in a given open shell

  1. 1.

    EsHFBsuperscript𝐸sHFBE^{\text{sHFB}}italic_E start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT looses energy relatively to experiment,

  2. 2.

    S2nsHFBsubscriptsuperscript𝑆sHFB2𝑛S^{\text{sHFB}}_{2n}italic_S start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT starts from a too low value,

relates directly to the fact that the mean-field valence-shell single-particle energy in the CS core ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT delivered by χ𝜒\chiitalic_χEFT interactions is systematically too small in absolute, i.e. non negative enough. This is accompanied with the fact that the effective mass is too low at the mean-field level, as testified by the too large (value) decrease of S2nsHFBsubscriptsuperscript𝑆sHFB2𝑛S^{\text{sHFB}}_{2n}italic_S start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT (Δ2nsHFBsubscriptsuperscriptΔsHFB2𝑛\Delta^{\text{sHFB}}_{2n}roman_Δ start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT), which is actually a key reason why pairing correlations are so weak. Second, the fact that

  1. 1.

    EsHFBsuperscript𝐸sHFBE^{\text{sHFB}}italic_E start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT is concave,

  2. 2.

    S2nsHFBsubscriptsuperscript𝑆sHFB2𝑛S^{\text{sHFB}}_{2n}italic_S start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is rising,

  3. 3.

    Δ2nsHFBsubscriptsuperscriptΔsHFB2𝑛\Delta^{\text{sHFB}}_{2n}roman_Δ start_POSTSUPERSCRIPT sHFB end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is negative,

throughout open shells, in opposition to experimental data, relates to the attractive character of the monopole valence-shell matrix element delivered by χ𝜒\chiitalic_χEFT interactions.

Interestingly, the above features are typically not displayed by sHFB calculations based on effective and empirical energy density functionals (EDF), see e.g. Burrello et al. (2021). Indeed, EDFs are tailored via a fit to empirical data to implicitly incorporate the dominant effect of dynamical correlations. In practice, this generally results into a significantly larger effective mass101010Even though Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is traditionally left to overestimate experimental data at shell closures in EDF calculations to leave some room for additional correlations, it does so on a much smaller scale than in present sHFB ab initio calculations that overestimate Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT at, e.g. N=20𝑁20N=20italic_N = 20 by 4444  MeV. and into much stronger pairing correlations since the pairing part of the functional is typically adjusted to reproduce experimental pairing gaps at the sHFB level. At the same time, it is striking that EDF parametrizations only tailored to reproduce many-body calculations of infinite nuclear matter and employed in finite nuclei at the strict mean-field level, i.e. without an explicit account of dynamical correlations on a nucleus-by-nucleus basis, do display the features identified above Burrello et al. (2021).

Refer to caption
Figure 6: Results of sHFB, sBMBPT(2), sBCCSD and VS-IMSRG(2) calculations against experimental data along Ca (left) and Cr (right) isotopic chains. Upper panels: absolute binding energies. Middle panels: two-neutron separation energy. Lower panels: two-neutron shell gap.

The above observations are also consistent with the evolution of canonical single-particle energies throughout an open shell, and more specifically of the valence-shell single-particle energy itself. In the sHF-EFA approximation, it can easily be shown that its evolution with avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT is linear

ϵv˘sHF-EFA(av)=subscriptsuperscriptitalic-ϵsHF-EFA˘𝑣subscript𝑎𝑣absent\displaystyle\epsilon^{\text{sHF-EFA}}_{\breve{v}}(a_{v})=italic_ϵ start_POSTSUPERSCRIPT sHF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) = ϵv˘CS+βv~av,subscriptsuperscriptitalic-ϵCS˘𝑣subscript𝛽~𝑣subscript𝑎𝑣\displaystyle\epsilon^{\text{CS}}_{\breve{v}}+\beta_{\tilde{v}}a_{v}\,,italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + italic_β start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT , (5)

the coefficient of the slope being given by βv~subscript𝛽~𝑣\beta_{\tilde{v}}italic_β start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT. As visible in Fig. 5, neutron canonical single-particle energies do evolve linearly within a given open-shell. In particular, the evolution of ϵ1f7/2subscriptitalic-ϵ1subscriptf72\epsilon_{1\text{f}_{7/2}}italic_ϵ start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT between 40Ca and 48Ca is perfectly reproduced using Eq. (5) (in fact its sHFB-ZP counterpart; see A). Eventually, this linear down-slopping evolution is fully correlated with the concavity of the binding energy.

3.3 Cr chain

Having characterized sHFB results along the semi-magic Ca isotopic chain, the focus turns now to doubly open-shell Cr isotopes.

As seen in the upper-right panel of Fig. 1, the global trend of sHFB binding energies is similar, relative to the data, than for Ca isotopes. Looking through the magnifying glass of S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT, experimental data do not however display the characteristic patterns identified along the Ca chain. In particular, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT decreases more gradually such that the sudden drops (sudden spikes in Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT), e.g. at N=20𝑁20N=20italic_N = 20 and 28282828, have all disappeared. Contrarily, a small bump (spike) is now visible in S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT (Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT) for N=24𝑁24N=24italic_N = 24, i.e. in 44Cr located in the middle of the 1f7/2 shell. These changes are not at all accounted for by sHFB results that closely follow those obtained previously. Indeed, in addition to displaying the defects identified along the Ca isotopic chain, sHFB results further fail to capture the qualitative modifications seen in the data, i.e. sHFB results keep a strong memory of the underlying spherical shell structure whose fingerprints are no longer visible along the Cr isotopic chain.

4 Spherical beyond mean-field corrections

Based on the previous analysis, the goal is now to assess whether consistently adding dynamical correlations via sBMBPT(2), sBCCSD or VS-IMSRG(2) can correct for the shortcomings identified at the sHFB level.

Refer to caption
Figure 7: Same as Fig. 2 for sHFB, sBMBPT(2), sBCCSD and VS-IMSRG(2).

4.1 Ca chain

As seen in the upper-left panel of Fig. 6, dynamical correlations compensate for the underbinding observed at the sHFB level such that all three methods reproduce well experimental binding energies along the Ca isotopic chain with the presently employed Hamiltonian. This is particularly true for VS-IMSRG(2) whose root-mean-square error to the data is equal to 1.91.91.91.9 MeV, while it is equal to 7.37.37.37.3 and 8.88.88.88.8 MeV for BMBPT(2) and BCCSD, respectively. In particular, the increasing underbinding of sHFB results as a function of neutron excess is essentially compensated for.

The improvement goes indeed beyond a plain shift as can be inferred from the middle-left panel of Fig. 6. Indeed, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT are systematically improved against experimental data for all three methods. First, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT are globally increased by up to about 5555 MeV. Second, the amplitudes of the sudden drops at magic numbers are reduced. As visible from the bottom-left panel, the two-neutron shell gap at N=20𝑁20N=20italic_N = 20 is reduced from 13131313 MeV in sHFB to 8.68.68.68.6 MeV in VS-IMSRG(2), which is comparable to the experimental value of 9.19.19.19.1 MeV. In sBMBPT(2) and sBCCSD the reduction is not pronounced enough, the Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT being equal to 12.412.412.412.4 in both cases, thus showing that low-rank elementary excitations are not enough to produce a fully quantitative picture of the N=20𝑁20N=20italic_N = 20 magicity. While sBCCSD is third-order-complete, it is of interest to investigate how much including genuine fourth-order triple excitations, e.g. by going to (approximate) BCCSDT, can help in this respect Vernik et al. (2024).

In spite of the N=20𝑁20N=20italic_N = 20 two-neutron shell gap being still overestimated in sBMBPT(2) and sBCCSD, the S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT at the beginning of each open shell is increased to be in much better agreement with experimental data. For example, dynamical correlations bring S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT in 42Ca from 13.313.313.313.3 MeV in sHFB to 18.018.018.018.0 and 18.118.118.118.1 MeV in BMBPT(2) and BCCSD, respectively, as well as to 18.618.618.618.6 MeV in VS-IMSRG(2), which compares favorably with the experimental value of 19.819.819.819.8 MeV. Third, the wrong linear increase throughout any given open shell is corrected for, as can be seen for example between 42Ca and 48Ca. This reflects the improvement of the curvature of the energy throughout open shells that can be better appreciated from the left-panels of Fig. 7 that focuses on the 1f7/2 shell. Dynamical correlations turn the energy from beyond concave at the sHFB level to being convex, in a way that is essentially identical with the three employed methods.

Eventually, the agreement with data for S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT along the Ca chain is qualitatively and quantitatively satisfying for all three methods even though the N=20𝑁20N=20italic_N = 20 magicity is still exaggerated in BMBPT(2) and BCCSD and the convexity throughout the 1f7/2 shell is not pronounced enough compared to experimental data for all three methods, which points to yet missing correlations. It will be interesting to investigate in the future whether the lack of convexity in the energy is correlated with the inability of presently employed ab initio methods to correctly reproduce the (infamous) evolution of charge radii between 40Ca and 48Ca Caurier et al. (2001).

Refer to caption
Figure 8: Second-order (on-shell) diagonal self-energy correction Σv˘(2)subscriptsuperscriptΣ2˘𝑣\Sigma^{(2)}_{\breve{v}}roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT to the valence-shell single-particle energy computed in the closed-shell core. Left: 1-particle/2-hole diagram. Right: 2-particle/1-hole diagram.
Refer to caption
Figure 9: Second-order (on-shell) correction v¯vvvv(2)subscriptsuperscript¯𝑣2𝑣superscript𝑣𝑣superscript𝑣\bar{v}^{(2)}_{vv^{\prime}vv^{\prime}}over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT to the diagonal valence-shell effective two-body matrix elements. Left: hole-hole diagram. Right: particle-particle diagram.

4.2 Analytical investigation

As demonstrated in Sec. 3.2, the deficiencies of sHFB can be understood via a semi-analytical analysis performed in the zero-pairing limit. The capacity of dynamical correlations to correct for those shortcomings is now analyzed in a similar manner within the frame of sMBPT(2). As demonstrated in A.4, the mean-field result of Eq. (3) can be extended, for av2subscript𝑎𝑣2a_{v}\geq 2italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ≥ 2, to

S2n(2)(av)subscriptsuperscript𝑆(2)2𝑛subscript𝑎𝑣\displaystyle S^{\text{(2)}}_{2n}(a_{v})italic_S start_POSTSUPERSCRIPT (2) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =2ϵv˘CS(2)2βv˘(2)(av1),absent2subscriptsuperscriptitalic-ϵCS2˘𝑣2subscriptsuperscript𝛽2˘𝑣subscript𝑎𝑣1\displaystyle=-2\epsilon^{\text{CS}(2)}_{\breve{v}}-2\beta^{(2)}_{\breve{v}}(a% _{v}-1)\,,= - 2 italic_ϵ start_POSTSUPERSCRIPT CS ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - 2 italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT - 1 ) , (6a)
Δ2n(2)(av)subscriptsuperscriptΔ(2)2𝑛subscript𝑎𝑣\displaystyle\Delta^{\text{(2)}}_{2n}(a_{v})roman_Δ start_POSTSUPERSCRIPT (2) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =4βv˘(2),absent4subscriptsuperscript𝛽2˘𝑣\displaystyle=4\beta^{(2)}_{\breve{v}}\,,= 4 italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT , (6b)

where the second-order (on-shell) valence-shell single-particle energy and averaged valence-shell interaction computed in the CS core

ϵv˘CS(2)subscriptsuperscriptitalic-ϵCS2˘𝑣\displaystyle\epsilon^{\text{CS}(2)}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ϵv˘CS+Σv˘(2)(ϵv˘CS),absentsubscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptΣ2˘𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\equiv\epsilon^{\text{CS}}_{\breve{v}}+\Sigma^{(2)}_{\breve{v}}(% \epsilon^{\text{CS}}_{\breve{v}})\,,≡ italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) , (7a)
βv~(2)subscriptsuperscript𝛽2~𝑣\displaystyle\beta^{(2)}_{\tilde{v}}italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT 1dvmvdv˘(𝕧¯vvvv+𝕧¯vvvv(2)(ϵv˘CS)).absent1subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑˘𝑣subscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscript¯𝕧2𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\equiv\frac{1}{d_{v}}\sum_{m_{v^{\prime}}}^{d_{\breve{v}}}\left(% \overline{\mathbb{v}}_{vv^{\prime}vv^{\prime}}+\overline{\mathbb{v}}^{(2)}_{vv% ^{\prime}vv^{\prime}}(\epsilon^{\text{CS}}_{\breve{v}})\right)\,.≡ divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ( over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + over¯ start_ARG blackboard_v end_ARG start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) ) . (7b)

involve the (on-shell) valence-shell self-energy and two-body effective interaction corrections

Σv˘(2)(ϵv˘CS)subscriptsuperscriptΣ2˘𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\Sigma^{(2)}_{\breve{v}}(\epsilon^{\text{CS}}_{\breve{v}})roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) =+12hhp|𝕧¯hhvp|2ϵpCS+ϵv˘CSϵhCSϵhCSabsent12subscriptsuperscript𝑝superscriptsubscript¯𝕧superscript𝑣𝑝2subscriptsuperscriptitalic-ϵCS𝑝subscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptitalic-ϵCSsubscriptsuperscriptitalic-ϵCSsuperscript\displaystyle=+\frac{1}{2}\sum_{hh^{\prime}p}\frac{|\overline{\mathbb{v}}_{hh^% {\prime}vp}|^{2}}{\epsilon^{\text{CS}}_{p}+\epsilon^{\text{CS}}_{\breve{v}}-% \epsilon^{\text{CS}}_{h}-\epsilon^{\text{CS}}_{h^{\prime}}}= + divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_p end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_p end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT + italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_ARG
12pph|𝕧¯vhpp|2ϵpCS+ϵpCSϵhCSϵv˘CS12subscript𝑝superscript𝑝superscriptsubscript¯𝕧𝑣𝑝superscript𝑝2subscriptsuperscriptitalic-ϵCS𝑝subscriptsuperscriptitalic-ϵCSsuperscript𝑝subscriptsuperscriptitalic-ϵCSsubscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\hskip 11.38092pt-\frac{1}{2}\sum_{pp^{\prime}h}\frac{|\overline{% \mathbb{v}}_{vhpp^{\prime}}|^{2}}{\epsilon^{\text{CS}}_{p}+\epsilon^{\text{CS}% }_{p^{\prime}}-\epsilon^{\text{CS}}_{h}-\epsilon^{\text{CS}}_{\breve{v}}}- divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_h italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT + italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG (8a)
𝕧¯vvvv(2)(ϵv˘CS)subscriptsuperscript¯𝕧2𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\overline{\mathbb{v}}^{(2)}_{vv^{\prime}vv^{\prime}}(\epsilon^{% \text{CS}}_{\breve{v}})over¯ start_ARG blackboard_v end_ARG start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) =+12hh|𝕧¯hhvv|22ϵv˘CSϵhCSϵhCSabsent12subscriptsuperscriptsuperscriptsubscript¯𝕧superscript𝑣superscript𝑣22subscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptitalic-ϵCSsubscriptsuperscriptitalic-ϵCSsuperscript\displaystyle=+\frac{1}{2}\sum_{hh^{\prime}}\frac{|\overline{\mathbb{v}}_{hh^{% \prime}vv^{\prime}}|^{2}}{2\epsilon^{\text{CS}}_{\breve{v}}-\epsilon^{\text{CS% }}_{h}-\epsilon^{\text{CS}}_{h^{\prime}}}= + divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG 2 italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_ARG
12pp|𝕧¯vvpp|2ϵpCS+ϵpCS2ϵv˘CS,12subscript𝑝superscript𝑝superscriptsubscript¯𝕧𝑣superscript𝑣𝑝superscript𝑝2subscriptsuperscriptitalic-ϵCS𝑝subscriptsuperscriptitalic-ϵCSsuperscript𝑝2subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\hskip 11.38092pt-\frac{1}{2}\sum_{pp^{\prime}}\frac{|\overline{% \mathbb{v}}_{vv^{\prime}pp^{\prime}}|^{2}}{\epsilon^{\text{CS}}_{p}+\epsilon^{% \text{CS}}_{p^{\prime}}-2\epsilon^{\text{CS}}_{\breve{v}}}\,,- divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT + italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT - 2 italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG , (8b)

displayed diagrammatically in Figs. 8 and 9, respectively. The self-energy correction collects a positive (2-hole/1-particle) contribution and a negative (1-hole/2-particle) contribution. Similarly, the valence-shell interaction correction collects a positive (hole-hole) contribution and a negative (particle-particle) contribution.

As seen from Eqs. (6)-(8), and in agreement with the results shown in the middle-left panel of Fig. 6 and analyzed in the present section, dynamical correlations modify both the starting value and the slope of S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT in the valence-shell. For example, the negative second-order self-energy correction Σ1f7/2(2)subscriptsuperscriptΣ21subscriptf72\Sigma^{(2)}_{1\text{f}_{7/2}}roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT lowers ϵ1f7/2CS(2)subscriptsuperscriptitalic-ϵCS21subscriptf72\epsilon^{\text{CS}(2)}_{1\text{f}_{7/2}}italic_ϵ start_POSTSUPERSCRIPT CS ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT in such a way that S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT computed in sBMBPT(2) increases from 13.2813.2813.2813.28 to 18.0218.0218.0218.02 MeV in 42Ca to almost match the experimental value (19.8419.8419.8419.84 MeV). This effect relates to the coupling of a propagating nucleon to 1-particle/2-hole and 2-particle/1-hole configurations as represented in Fig 8, the latter winning over the former111111The lowering of ϵ1f7/2subscriptitalic-ϵ1subscriptf72\epsilon_{1\text{f}_{7/2}}italic_ϵ start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT is not accompanied by a decrease of Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT in 40Ca in sBMBPT(2) and sBCCSD, contrary to sVS-IMSRG(2), i.e. ϵ1d3/2subscriptitalic-ϵ1subscriptd32\epsilon_{1\text{d}_{3/2}}italic_ϵ start_POSTSUBSCRIPT 1 d start_POSTSUBSCRIPT 3 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT is lowered as much as ϵ1f7/2subscriptitalic-ϵ1subscriptf72\epsilon_{1\text{f}_{7/2}}italic_ϵ start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT. Thus, the needed increase of the effective mass associated with the compression of on-shell single-particle energies is not accounted for by low-order corrections to sHFB.. Consistently, the second-order correction to the average 1f7/2 valence-shell effective interaction is repulsive, with the hole-hole contribution winning over the particle-particle one. In the present calculation, such a correction is larger in absolute value than the mean-field contribution and manages to turn the total energy from being concave to being convex, i.e. it makes S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT decrease linearly between 42Ca and 48Ca as for experimental data121212The amount by which S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is increased at the start of the open-shell and the fact that its slope is actually inverted depend on the Hamiltonian under use; see Refs. Tichai et al. (2018); Somà et al. (2021) for examples where the qualitative defects of the sHFB results are not actually corrected via the inclusion of low-order dynamical correlations.. Still, and as can be seen from the bottom-left panel of Fig. 7, the positive curvature β1f7/2(2)=25subscriptsuperscript𝛽21subscriptf7225\beta^{(2)}_{1\text{f}_{7/2}}=25italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 1 f start_POSTSUBSCRIPT 7 / 2 end_POSTSUBSCRIPT end_POSTSUBSCRIPT = 25 keV131313This value is essentially constant throughout the valence shell. is not large enough141414The same is true for sBCCSD and VS-IMSRG(2) calculations as can be inferred from the bottom-left panel of Fig. 7. compared to experimental data (Δ2n/4220subscriptΔ2𝑛4220\Delta_{2n}/4\approx 220roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT / 4 ≈ 220 keV in 42-46Ca), thus pointing to yet missing many-body correlations as discussed earlier on.

4.3 Cr chain

While the deficiencies observed at the sHFB were shown to be qualitatively and quantitatively corrected via the consistent addition of dynamical correlations in Ca isotopes, it remains to be seen to which extent this is the case along the Cr isotopic chain.

As seen in the upper-right panel of Fig. 6, correlations brought by sBMBPT(2), sBCCSD and VS-IMSRG(2) provide the bulk of the missing binding along the Cr chain as well, even though the end values are globally further away from experimental data than for Ca isotopes. While the rms error to the data is 1.91.91.91.9, 7.37.37.37.3 and 8.88.88.88.8 MeV for VS-IMSRG(2), sBMBPT(2) and sBCCSD in Ca isotopes, it becomes 4.04.04.04.0, 10.610.610.610.6 and 14.714.714.714.7 MeV in Cr isotopes, respectively; i.e. the deterioration is more pronounced for sBMBPT(2) and sBCCSD.

Looking at the middle- and bottom-right panels of Fig. 6, sBMBPT(2) and sBCCSD are seen to improve the reproduction of experimental S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT compared to sHFB. Still, the level of agreement is neither on the same level as in Ca isotopes nor on the same level as for VS-IMSRG(2) in those Cr isotopes. The large spikes of Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT seen at N=20,28𝑁2028N=20,28italic_N = 20 , 28 and 40404040 for sHFB are only slightly diminished in sBCCSD calculations, thus wrongly keeping the imprint of the spherical magic numbers. Even if the behavior throughout the 1f7/2 shell is improved, as can also be appreciated from the left panels of Fig. 7, it remains quite remote from experimental data. As for sBMBPT(2) results, Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT bear little resemblance to experimental data and are clearly not credible.

Contrarily, the S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT predicted by VS-IMSRG(2) are both in qualitative and quantitative agreement with experimental data151515The slight degradation observed in the vicinity on N=20𝑁20N=20italic_N = 20 and 40404040 is attributable to the need to reset the valence space.. Indeed, the disappearance of the spikes at N=20,28𝑁2028N=20,28italic_N = 20 , 28 and 40404040, as well as the appearance of a new one for N=24𝑁24N=24italic_N = 24, are perfectly reproduced. This demonstrates that the exact diagonalization of the effective Hamiltonian within the fp𝑓𝑝fpitalic_f italic_p shell is able to capture crucial static correlations that are not accounted for by low-rank excitations on top of a spherical mean field via sBMBPT(2) and sBCCSD.

5 Deformed unperturbed state

Even if challenges remain to be overcome to reach high accuracy or the description of specific observables impacted by collective fluctuations (e.g. superfluidity, radii between 40Ca and 48Ca…), the discussion above demonstrates that polynomially-scaling expansion methods built on top of a spherical Bogoliubov reference state and implemented to rather low truncation order deliver a good account of mid-mass doubly closed-shell and singly open-shell nuclear ground states. Contrarily, doubly open-shell nuclei require the inclusion of specific static correlations that can hardly be incorporated following this strategy, i.e. they require a full diagonalization of the effective Hamiltonian in an appropriate valence space, thus compromising with the polynomial scaling that will eventually become crucial in heavy nuclei.

Refer to caption
Figure 10: Systematic dHFB, dBMBPT(2) and VS-IMSRG(2) calculations against experimental data along Ca (left) and Cr (right) isotopic chains. First line: absolute binding energies. Second line: two-neutron separation energy. Third line: two-neutron shell gap. Fourth line: intrinsic axial quadrupole deformation of the dHFB solution.
Refer to caption
Figure 11: Same as Fig. 2 for dHFB, dBMBPT(2) and VS-IMSRG(2).

On a principle level, the solution delivered by expansion many-body methods is eventually independent of the unperturbed state whenever all terms in the expansion series are summed up – provided that the expansion series actually converges Leininger et al. (2000); Roth and Langhammer (2010); Langhammer et al. (2012); Tichai et al. (2016); Demol et al. (2021, 2020). In practice however, the interesting question relates to how close to the exact solution one can be at the most economical cost. In this context, it is believed that dominant static correlations can be efficiently captured in doubly open-shell nuclei via an appropriate redefinition of the unperturbed state, at the price of breaking Novario et al. (2020); Frosini et al. (2024) (and eventually restoring Duguet (2015); Yao et al. (2020); Hagen et al. (2022); Frosini et al. (2022a); Sun et al. (2024)) rotational symmetry associated with angular-momentum conservation. The present section wishes to pedagogically illustrate that a quantitative description of doubly open-shell nuclei can indeed be achieved at (low) polynomial cost via dBMBPT(2) calculations performed on top of a deformed HFB unperturbed state.

5.1 Ca chain

Results of systematic dHFB, dBMBPT(2), as well as VS-IMSRG(2) calculations of Ca isotopes are displayed on the left-hand panels of Fig. 10. Comparing those to the results shown before on the left-hand panels of Fig. 6, it is clear that allowing the mean-field solution to deform does not lead to any significant modification along the Ca isotopic chain. Indeed, and as demonstrated by the lower panel of Fig. 10, almost all Ca isotopes do not take advantage of this possibility at the mean-field level161616The few isotopes that do deform, i.e. 32,44,46,68,70Ca, only acquire a small intrinsic deformation.. The fact that static correlations associated with quadrupolar deformations are not emerging from the calculation is consistent with the fact sBMBPT(2) and sBCCSD results were already satisfactory as discussed extensively in Sec. 4.

5.2 Cr chain

As the comparison of the right-hand panels of Figs. 6 and 10 illustrate, the energetic of doubly open-shell Cr isotopes is instead strongly impacted by the breaking of rotational symmetry. Indeed, most Cr isotopes do acquire a large intrinsic deformation171717Interestingly, neutron deficient isotopes 34-42 are predicted to display a strong oblate-prolate oscillation. Isotopes between N=20𝑁20N=20italic_N = 20 and N=28𝑁28N=28italic_N = 28 all display a large prolate deformation, which slowly fades away towards N=40𝑁40N=40italic_N = 40. Eventually, the prolate deformation suddenly increases again going across N=40𝑁40N=40italic_N = 40 and stays large until the predicted neutron drip line at N=48𝑁48N=48italic_N = 48. as seen in the lower-right panel of Fig. 10. While the overall rms error of total binding energies remains similar in sBMBPT(2) and dBMBPT(2), the evolution with N𝑁Nitalic_N is strongly impacted as can be inferred from the behavior of S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT.

As a matter of fact, the qualitative (quantitative) reproduction of S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT (Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT) is already excellent at the deformed mean-field level, i.e. all deficiencies identified in sHFB results are already corrected by dHFB. In particular, the fictitious shell closures at N=20,28𝑁2028N=20,28italic_N = 20 , 28 and 40404040 have disappeared in dHFB results. Eventually, dynamical correlations added on top of dHFB via dBMBPT(2) increase S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT systematically to reach an excellent agreement with both VS-IMSRG(2) results and experimental data. While the rms error to experimental S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT was 2.92.92.92.9 MeV for sBMBPT(2) (5.85.85.85.8 MeV for sHFB), it is 0.90.90.90.9 MeV for dBMBPT(2) (4.44.44.44.4 MeV for dHFB), which is to be compared to 2.22.22.22.2 MeV for VS-IMSRG(2).

Focusing on the 1f7/2 shell, the right panels of Fig. 11 show that the curvature of the energy is already very well captured at the dHFB level, while it was qualitatively wrong for both sHFB and sBMBPT(2), and becomes essentially as good as with VS-IMSRG(2) for dBMBPT(2).

These results demonstrate that static correlations in doubly open-shell nuclei can be qualitatively and quantitatively seized via polynomially-scaling expansion methods built on top of a deformed reference state and implemented to rather low truncation order.

6 Sn chain

Refer to caption
Figure 12: Results from sHFB, sBMBPT(2), sBCCSD, dHFB and dBMBPT(2) calculations against experimental data along the Sn isotopic chain. First line: two-neutron separation energy. Second line: two-neutron shell gap. Third line: intrinsic axial quadrupole deformation of sHFB and dHFB solutions.

As a last step, the discussion is extended to semi-magic Sn isotopes between 100Sn and 132Sn, i.e. going through the sub-shell closures at N=58,64,66𝑁586466N=58,64,66italic_N = 58 , 64 , 66 and 70707070 located between the N=50𝑁50N=50italic_N = 50 and 82828282 major shell closures. In Fig. 12, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT computed from mean-field and beyond-mean-field calculations with and without breaking rotational symmetry are displayed.

It is clear that experimental data do not show any fingerprint of the sub-shell closures, i.e. S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT decreases linearly between N=52𝑁52N=52italic_N = 52 and 82828282 such that Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is flat. Contrarily, sHFB results strongly reflect the presence of those sub-shell closures in a way that is consistent with the behavior seen in Ca isotopes, i.e. S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT are too low overall and rise linearly throughout open-shells, especially along the highly degenerate 1g7/2 and 1h11/2 shells.

Dynamical correlations brought on top of sHFB via sBMBPT(2) and sBCCSD largely ameliorate the situation, i.e. S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT are increased overall and the behavior throughout open-shells are corrected. However, the imprint of the sub-shell closures remain visible.

The larger mass combined with the weak pairing correlations induced by χ𝜒\chiitalic_χEFT interactions at the mean-field level makes several semi-magic Sn isotopes take advantage of deformation if authorized to do so181818This is again at variance with mean-field calculations based on effective EDFs. Indeed, the strong built-in pairing typically constrains all Sn isotopes to remain spherical between N=50𝑁50N=50italic_N = 50 and N=82𝑁82N=82italic_N = 82 in that case. as can be seen from the lower panel of Fig. 12. Still, the axial quadrupole deformation parameter remains small in all cases. As in Ca isotopes, the wrong trend of S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT with N𝑁Nitalic_N observed at the sHFB level is thus not corrected by dHFB calculations and dBMBPT(2) eventually deliver very similar results to sBMBPT(2).

7 Conclusions

In order to extend the reach of ab initio calculations to heavy doubly open-shell nuclei in the future, the most efficient strategy to incorporate dominant many-body correlations at play in (heavy) nuclei must be identified. With this in mind, the present work analyzed in details the impact of many-body correlations on binding energies of Calcium and Chromium isotopes with an (even) neutron number ranging from N=12𝑁12N=12italic_N = 12 to N=50𝑁50N=50italic_N = 50.

Using an empirically-optimal (soft) χ𝜒\chiitalic_χEFT-based Hamiltonian, binding energies computed in the spherical mean-field approximation were first shown to display specific shortcomings in semi-magic Ca isotopes. In addition to being associated (as expected) to a significant underbinding, the corresponding energy was shown to evolve qualitatively incorrectly throughout (highly degenerate) open shells, i.e. whereas the linear decrease with the number of valence nucleons is too slow, the quadratic term makes the energy concave instead of being convex. Relying on the observation that χ𝜒\chiitalic_χEFT-based interactions generate very little pairing at the spherical mean-field level, these two features could be related analytically to the fact that (i) single-particle energies are not enough bound and that (ii) the monopole valence-shell two-body matrix elements is attractive.

Next, the consistent addition of dynamical correlations at polynomial cost via, e.g., low-order perturbation theory was shown to correct the deficiencies identified at the spherical mean-field level. This decisive improvement could also be understood analytically. Eventually, it is possible to reach a description of semi-magic Ca isotopes on essentially the same quantitative level as valence-space in-medium similarity renormalization group calculations, which rely on the diagonalization of the effective Hamiltonian in the fp𝑓𝑝fpitalic_f italic_p valence space. Either way, some yet missing correlation energy was identified between 40Ca and 48Ca that could be correlated with the (infamous) difficulty to describe the evolution of the charge radius between those two isotopes.

Moving to doubly open-shell Cr isotopes, calculations based on a spherical mean-field unperturbed state could not appropriately reproduce the binding energy evolution. However, allowing this unperturbed mean-field state to break rotational symmetry proved to be sufficient to capture the static correlations responsible for the phenomenological modifications observed between the two isotopic chains and that otherwise need the diagonalization of the effective Hamiltonian in large valence spaces.

Semi-magic Sn isotopes behave similarly to lighter Ca isotopes with a spherical mean-field delivering qualitatively wrong patterns that are corrected by the consistent addition of low-order dynamical correlations.

Eventually, the present work demonstrates in a pedagogical way that polynomially-scaling expansion methods based on unperturbed states possibly breaking (and restoring) symmetries constitute an optimal route to extend ab initio calculations to heavy closed- and open-shell nuclei.

Acknowledgements

The authors thank H. Hergert and T. Miyagi for providing the interaction matrix elements used in the numerical simulations. The work of A.S. was supported by the European Union’s Horizon 2020 research and innovation program under grant agreement No 800945 - NUMERICS - H2020-MSCA-COFUND-2017. The work of P.D. was supported by the Research Foundation Flanders (FWO, Belgium, grant 11G5123N). The work of A.T. was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 101020842). The work of M.F. was carried out in the framework of the SINET project funded by the CEA. Calculations were performed by using HPC resources from GENCI-TGCC (Contract No. A0150513012).

Data Availability Statement

This manuscript has no associated data or the data will not be deposited.

Appendix A Zero-pairing description

The present work shows that χ𝜒\chiitalic_χEFT-based interactions typically lead to a mean-field approximation displaying very weak pairing correlations in open-shell nuclei. This property is intimately linked to the fact that the total HFB energy is concave rather than convex throughout long (enough) degenerate spherical shells. This connection can be validated analytically by considering that the system is in the extreme zero-pairing limit.

The zero-pairing mean-field description of an open-shell system can be meaningfully achieved on the basis of two different many-body formalisms, i.e. the Hartree-Fock-Bogoliubov theory in the zero-pairing limit (HFB-ZP) Duguet et al. (2020) or the Hartree-Fock theory in the equal filling approximation (HF-EFA) Perez-Martin and Robledo (2008). The two cases are worked out analytically below to validate the results obtained through realistic sHFB calculations in the body of the text.

A.1 Hartree-Fock Bogoliubov

The fully-paired HFB vacuum associated with a time-reversal symmetric system is written in its canonical, i.e., BCS-like, form as Ring and Schuck (1980)

|Φk>0[uk+vkakak¯]|0.ketΦsubscriptproduct𝑘0delimited-[]subscript𝑢𝑘subscript𝑣𝑘subscriptsuperscript𝑎𝑘subscriptsuperscript𝑎¯𝑘ket0|\Phi\rangle\equiv\prod_{k>0}\left[u_{k}+v_{k}a^{\dagger}_{k}a^{\dagger}_{\bar% {k}}\right]|0\rangle\,.| roman_Φ ⟩ ≡ ∏ start_POSTSUBSCRIPT italic_k > 0 end_POSTSUBSCRIPT [ italic_u start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT + italic_v start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over¯ start_ARG italic_k end_ARG end_POSTSUBSCRIPT ] | 0 ⟩ . (9)

Operators {ak,ak}subscriptsuperscript𝑎𝑘subscript𝑎𝑘\{a^{\dagger}_{k},a_{k}\}{ italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT , italic_a start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT } characterize the so-called canonical one-body basis in which pairs of conjugate states (k,k¯)𝑘¯𝑘(k,\bar{k})( italic_k , over¯ start_ARG italic_k end_ARG ) are singled out by the Bogoliubov transformation via a quantum number mksubscript𝑚𝑘m_{k}italic_m start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT such that k(k˘,mk)𝑘˘𝑘subscript𝑚𝑘k\equiv(\breve{k},m_{k})italic_k ≡ ( over˘ start_ARG italic_k end_ARG , italic_m start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ) and k¯(k˘,mk)¯𝑘˘𝑘subscript𝑚𝑘\bar{k}\equiv(\breve{k},-m_{k})over¯ start_ARG italic_k end_ARG ≡ ( over˘ start_ARG italic_k end_ARG , - italic_m start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ). The state k¯¯𝑘\bar{k}over¯ start_ARG italic_k end_ARG (k𝑘kitalic_k) corresponds to the time-reversal state of k𝑘kitalic_k (k¯¯𝑘\bar{k}over¯ start_ARG italic_k end_ARG) up to a sign ηksubscript𝜂𝑘\eta_{k}italic_η start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT (ηk¯subscript𝜂¯𝑘\eta_{\bar{k}}italic_η start_POSTSUBSCRIPT over¯ start_ARG italic_k end_ARG end_POSTSUBSCRIPT) such that ηk¯ηk=1subscript𝜂¯𝑘subscript𝜂𝑘1\eta_{\bar{k}}\eta_{k}=-1italic_η start_POSTSUBSCRIPT over¯ start_ARG italic_k end_ARG end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT = - 1.

The BCS-like occupation numbers ukuk˘subscript𝑢𝑘subscript𝑢˘𝑘u_{k}\equiv u_{\breve{k}}italic_u start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT ≡ italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT and vk=ηkvk˘subscript𝑣𝑘subscript𝜂𝑘subscript𝑣˘𝑘v_{k}=\eta_{k}v_{\breve{k}}italic_v start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT = italic_η start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT fulfilling uk2+vk2=1superscriptsubscript𝑢𝑘2superscriptsubscript𝑣𝑘21u_{k}^{2}+v_{k}^{2}=1italic_u start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + italic_v start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT = 1 are expressed in terms of the positive mksubscript𝑚𝑘m_{k}italic_m start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT-independent coefficients (uk˘,vk˘)subscript𝑢˘𝑘subscript𝑣˘𝑘(u_{\breve{k}},v_{\breve{k}})( italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT , italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT ). Employing the latter, the non-zero elements of the normal and anomalous density matrices read in the canonical basis as

ρkk=subscript𝜌𝑘𝑘absent\displaystyle\rho_{kk}=italic_ρ start_POSTSUBSCRIPT italic_k italic_k end_POSTSUBSCRIPT = ρk¯k¯=vk˘2,subscript𝜌¯𝑘¯𝑘subscriptsuperscript𝑣2˘𝑘\displaystyle\rho_{\bar{k}\bar{k}}=v^{2}_{\breve{k}}\,,italic_ρ start_POSTSUBSCRIPT over¯ start_ARG italic_k end_ARG over¯ start_ARG italic_k end_ARG end_POSTSUBSCRIPT = italic_v start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT , (10)
κkk¯=subscript𝜅𝑘¯𝑘absent\displaystyle\kappa_{k\bar{k}}=italic_κ start_POSTSUBSCRIPT italic_k over¯ start_ARG italic_k end_ARG end_POSTSUBSCRIPT = κk¯k=ηkuk˘vk˘.subscript𝜅¯𝑘𝑘subscript𝜂𝑘subscript𝑢˘𝑘subscript𝑣˘𝑘\displaystyle-\kappa_{\bar{k}k}=\eta_{k}u_{\breve{k}}v_{\breve{k}}\,.- italic_κ start_POSTSUBSCRIPT over¯ start_ARG italic_k end_ARG italic_k end_POSTSUBSCRIPT = italic_η start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT . (11)

Based on the above and given the nuclear Hamiltonian

H𝐻absent\displaystyle H\equivitalic_H ≡ ijtijaiajsubscript𝑖𝑗subscript𝑡𝑖𝑗subscriptsuperscript𝑎𝑖subscript𝑎𝑗\displaystyle\sum_{ij}t_{ij}a^{\dagger}_{i}a_{j}∑ start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT italic_t start_POSTSUBSCRIPT italic_i italic_j end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT
+1(2!)2ijklv¯ijklaiajalak1superscript22subscript𝑖𝑗𝑘𝑙subscript¯𝑣𝑖𝑗𝑘𝑙subscriptsuperscript𝑎𝑖subscriptsuperscript𝑎𝑗subscript𝑎𝑙subscript𝑎𝑘\displaystyle+\frac{1}{(2!)^{2}}\sum_{ijkl}\overline{v}_{ijkl}a^{\dagger}_{i}a% ^{\dagger}_{j}a_{l}a_{k}+ divide start_ARG 1 end_ARG start_ARG ( 2 ! ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_i italic_j italic_k italic_l end_POSTSUBSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_i italic_j italic_k italic_l end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_l end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT
+1(3!)2ijklmnw¯ijklmnaiajakanamal,1superscript32subscript𝑖𝑗𝑘𝑙𝑚𝑛subscript¯𝑤𝑖𝑗𝑘𝑙𝑚𝑛subscriptsuperscript𝑎𝑖subscriptsuperscript𝑎𝑗subscriptsuperscript𝑎𝑘subscript𝑎𝑛subscript𝑎𝑚subscript𝑎𝑙\displaystyle+\frac{1}{(3!)^{2}}\sum_{ijklmn}\overline{w}_{ijklmn}a^{\dagger}_% {i}a^{\dagger}_{j}a^{\dagger}_{k}a_{n}a_{m}a_{l}\,,+ divide start_ARG 1 end_ARG start_ARG ( 3 ! ) start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_i italic_j italic_k italic_l italic_m italic_n end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_i italic_j italic_k italic_l italic_m italic_n end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_i end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_j end_POSTSUBSCRIPT italic_a start_POSTSUPERSCRIPT † end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_n end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_m end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_l end_POSTSUBSCRIPT , (12)

the total HFB energy reads in the canonical basis as

EHFBsubscript𝐸HFBabsent\displaystyle E_{\text{HFB}}\equivitalic_E start_POSTSUBSCRIPT HFB end_POSTSUBSCRIPT ≡ Φ|H|Φquantum-operator-productΦ𝐻Φ\displaystyle\langle\Phi|H|\Phi\rangle⟨ roman_Φ | italic_H | roman_Φ ⟩
\displaystyle\equiv E|Φkin+E|ΦHF+E|ΦBsubscriptsuperscript𝐸kinketΦsubscriptsuperscript𝐸HFketΦsubscriptsuperscript𝐸BketΦ\displaystyle E^{\text{kin}}_{|\Phi\rangle}+E^{\text{HF}}_{|\Phi\rangle}+E^{% \text{B}}_{|\Phi\rangle}italic_E start_POSTSUPERSCRIPT kin end_POSTSUPERSCRIPT start_POSTSUBSCRIPT | roman_Φ ⟩ end_POSTSUBSCRIPT + italic_E start_POSTSUPERSCRIPT HF end_POSTSUPERSCRIPT start_POSTSUBSCRIPT | roman_Φ ⟩ end_POSTSUBSCRIPT + italic_E start_POSTSUPERSCRIPT B end_POSTSUPERSCRIPT start_POSTSUBSCRIPT | roman_Φ ⟩ end_POSTSUBSCRIPT
=\displaystyle== ktkkvk˘2subscript𝑘subscript𝑡𝑘𝑘subscriptsuperscript𝑣2˘𝑘\displaystyle\sum_{k}t_{kk}\,v^{2}_{\breve{k}}∑ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_t start_POSTSUBSCRIPT italic_k italic_k end_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT
+12kkv¯kkkkvk˘2vk˘212subscript𝑘superscript𝑘subscript¯𝑣𝑘superscript𝑘𝑘superscript𝑘superscriptsubscript𝑣˘𝑘2superscriptsubscript𝑣superscript˘𝑘2\displaystyle+\frac{1}{2}\sum_{kk^{\prime}}\overline{v}_{kk^{\prime}kk^{\prime% }}\,v_{\breve{k}}^{2}\,v_{\breve{k}^{\prime}}^{2}+ divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT
+14kkv¯kk¯kk¯ηkηkuk˘vk˘uk˘vk˘14subscript𝑘superscript𝑘subscript¯𝑣𝑘¯𝑘superscript𝑘superscript¯𝑘subscript𝜂𝑘subscript𝜂superscript𝑘subscript𝑢˘𝑘subscript𝑣˘𝑘subscript𝑢superscript˘𝑘subscript𝑣superscript˘𝑘\displaystyle+\frac{1}{4}\sum_{kk^{\prime}}\overline{v}_{k\bar{k}k^{\prime}% \bar{k}^{\prime}}\,\eta_{k}\eta_{k^{\prime}}\,u_{\breve{k}}v_{\breve{k}}\,u_{% \breve{k}^{\prime}}v_{\breve{k}^{\prime}}+ divide start_ARG 1 end_ARG start_ARG 4 end_ARG ∑ start_POSTSUBSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_k over¯ start_ARG italic_k end_ARG italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT
+16kkk′′w¯kkk′′kkk′′vk˘2vk˘2vk˘′′216subscript𝑘superscript𝑘superscript𝑘′′subscript¯𝑤𝑘superscript𝑘superscript𝑘′′𝑘superscript𝑘superscript𝑘′′superscriptsubscript𝑣˘𝑘2superscriptsubscript𝑣superscript˘𝑘2superscriptsubscript𝑣superscript˘𝑘′′2\displaystyle+\frac{1}{6}\sum_{kk^{\prime}k^{\prime\prime}}\overline{w}_{kk^{% \prime}k^{\prime\prime}kk^{\prime}k^{\prime\prime}}\,v_{\breve{k}}^{2}\,v_{% \breve{k}^{\prime}}^{2}\,v_{\breve{k}^{\prime\prime}}^{2}+ divide start_ARG 1 end_ARG start_ARG 6 end_ARG ∑ start_POSTSUBSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_k start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_k start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_k start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT
+14kkk′′w¯kk¯k′′kk¯k′′ηkηkuk˘vk˘uk˘vk˘vk˘′′2.14subscript𝑘superscript𝑘subscriptsuperscript𝑘′′subscript¯𝑤𝑘¯𝑘superscript𝑘′′superscript𝑘superscript¯𝑘superscript𝑘′′subscript𝜂𝑘subscript𝜂superscript𝑘subscript𝑢˘𝑘subscript𝑣˘𝑘subscript𝑢superscript˘𝑘subscript𝑣superscript˘𝑘superscriptsubscript𝑣superscript˘𝑘′′2\displaystyle+\frac{1}{4}\sum_{kk^{\prime}}\sum_{k^{\prime\prime}}\overline{w}% _{k\bar{k}k^{\prime\prime}k^{\prime}\bar{k}^{\prime}k^{\prime\prime}}\,\eta_{k% }\eta_{k^{\prime}}\,u_{\breve{k}}v_{\breve{k}}\,u_{\breve{k}^{\prime}}v_{% \breve{k}^{\prime}}\,v_{\breve{k}^{\prime\prime}}^{2}\,.+ divide start_ARG 1 end_ARG start_ARG 4 end_ARG ∑ start_POSTSUBSCRIPT italic_k italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT ∑ start_POSTSUBSCRIPT italic_k start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_k over¯ start_ARG italic_k end_ARG italic_k start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_k start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_k start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT . (13)

Canonical single-particle states further gather in degenerate shells. All states belonging to a given shell share the same set of quantum numbers k˘˘𝑘\breve{k}over˘ start_ARG italic_k end_ARG and only differ by the value of mksubscript𝑚𝑘m_{k}italic_m start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT such that the single-particle energy defining the shell is independent of it, i.e. ϵk=ϵk˘subscriptitalic-ϵ𝑘subscriptitalic-ϵ˘𝑘\epsilon_{k}=\epsilon_{\breve{k}}italic_ϵ start_POSTSUBSCRIPT italic_k end_POSTSUBSCRIPT = italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_k end_ARG end_POSTSUBSCRIPT.

In the zero-pairing limit Duguet et al. (2020), states belonging to three categories of shells need to be distinguished according to

  1. 1.

    ϵh˘λ<0subscriptitalic-ϵ˘𝜆0\epsilon_{\breve{h}}-\lambda<0italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_h end_ARG end_POSTSUBSCRIPT - italic_λ < 0, casually denoted as “hole states”,

  2. 2.

    ϵv˘λ=0subscriptitalic-ϵ˘𝑣𝜆0\epsilon_{\breve{v}}-\lambda=0italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_λ = 0, casually denoted as “valence states”,

  3. 3.

    ϵp˘λ>0subscriptitalic-ϵ˘𝑝𝜆0\epsilon_{\breve{p}}-\lambda>0italic_ϵ start_POSTSUBSCRIPT over˘ start_ARG italic_p end_ARG end_POSTSUBSCRIPT - italic_λ > 0, casually denoted as “particle states” ,

where λ𝜆\lambdaitalic_λ denotes the chemical potential. Accordingly, it can be shown that canonical states display the following average occupations

  1. 1.

    Hole state: vh˘2=1subscriptsuperscript𝑣2˘1v^{2}_{\breve{h}}=1italic_v start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_h end_ARG end_POSTSUBSCRIPT = 1,

  2. 2.

    Valence state: 0<vv˘210subscriptsuperscript𝑣2˘𝑣10<v^{2}_{\breve{v}}\leq 10 < italic_v start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ≤ 1,

  3. 3.

    Particle state: vp˘2=0subscriptsuperscript𝑣2˘𝑝0v^{2}_{\breve{p}}=0italic_v start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_p end_ARG end_POSTSUBSCRIPT = 0.

The valence shell gathers pv=dv/2subscript𝑝𝑣subscript𝑑𝑣2p_{v}=d_{v}/2italic_p start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT = italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT / 2 pairs of conjugated states such that the number of valence states dvsubscript𝑑𝑣d_{v}italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT (pairs pvsubscript𝑝𝑣p_{v}italic_p start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT) is equal to the number of mvsubscript𝑚𝑣m_{v}italic_m start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT (|mv|subscript𝑚𝑣|m_{v}|| italic_m start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT |) different values. Consequently, the A nucleons making up the system are exhausted in such a way that 0avdv0subscript𝑎𝑣subscript𝑑𝑣0\leq a_{v}\leq d_{v}0 ≤ italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ≤ italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT of them sit in the valence shell whereas Aav𝐴subscript𝑎𝑣A-a_{v}italic_A - italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT occupy the hole states. Consequently, the occupation of each of the dvsubscript𝑑𝑣d_{v}italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT valence states is

vv˘2ov˘=avdv,subscriptsuperscript𝑣2˘𝑣subscript𝑜˘𝑣subscript𝑎𝑣subscript𝑑𝑣v^{2}_{\breve{v}}\equiv o_{\breve{v}}=\frac{a_{v}}{d_{v}}\,,italic_v start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ≡ italic_o start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = divide start_ARG italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG , (14)

thus leading to

uv˘vv˘=ov˘(1ov˘).subscript𝑢˘𝑣subscript𝑣˘𝑣subscript𝑜˘𝑣1subscript𝑜˘𝑣u_{\breve{v}}v_{\breve{v}}=\sqrt{o_{\breve{v}}(1-o_{\breve{v}})}\,.italic_u start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_v start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = square-root start_ARG italic_o start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( 1 - italic_o start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) end_ARG . (15)

Based on the above, the HFB energy (Eq. 13) of an open-shell system with avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT nucleons in the valence shell can be computed relatively to the CS core in the zero-pairing limit. After a lengthy but straightforward derivation, one obtains

ΔEHFB-ZP(av)Δsuperscript𝐸HFB-ZPsubscript𝑎𝑣\displaystyle\Delta E^{\text{HFB-ZP}}(a_{v})roman_Δ italic_E start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) EHFB-ZP(av)EHFB-ZP(0)absentsuperscript𝐸HFB-ZPsubscript𝑎𝑣superscript𝐸HFB-ZP0\displaystyle\equiv E^{\text{HFB-ZP}}(a_{v})-E^{\text{HFB-ZP}}(0)≡ italic_E start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) - italic_E start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT ( 0 ) (16)
=αv˘av+βv˘2av2+γv˘6av3,absentsubscript𝛼˘𝑣subscript𝑎𝑣subscript𝛽˘𝑣2superscriptsubscript𝑎𝑣2subscript𝛾˘𝑣6superscriptsubscript𝑎𝑣3\displaystyle=\alpha_{\breve{v}}a_{v}+\frac{\beta_{\breve{v}}}{2}a_{v}^{2}+% \frac{\gamma_{\breve{v}}}{6}a_{v}^{3}\,,= italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT + divide start_ARG italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG start_ARG 2 end_ARG italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT + divide start_ARG italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG start_ARG 6 end_ARG italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 3 end_POSTSUPERSCRIPT , (17)

where

αv˘subscript𝛼˘𝑣absent\displaystyle\alpha_{\breve{v}}\equivitalic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ≡ ϵv˘CS+Δv˘4,subscriptsuperscriptitalic-ϵCS˘𝑣subscriptΔ˘𝑣4\displaystyle\epsilon^{\text{CS}}_{\breve{v}}+\frac{\Delta_{\breve{v}}}{4}\,,italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + divide start_ARG roman_Δ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG start_ARG 4 end_ARG , (18a)
βv˘subscript𝛽˘𝑣absent\displaystyle\beta_{\breve{v}}\equivitalic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ≡ Uv˘dv12dv(Δv˘Zv˘),subscript𝑈˘𝑣subscript𝑑𝑣12subscript𝑑𝑣subscriptΔ˘𝑣subscript𝑍˘𝑣\displaystyle\frac{U_{\breve{v}}}{d_{v}}-\frac{1}{2d_{v}}\left(\Delta_{\breve{% v}}-Z_{\breve{v}}\right)\,,divide start_ARG italic_U start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG - divide start_ARG 1 end_ARG start_ARG 2 italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ( roman_Δ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_Z start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) , (18b)
γv˘subscript𝛾˘𝑣absent\displaystyle\gamma_{\breve{v}}\equivitalic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ≡ 1dv2(Xv˘32Zv˘),1subscriptsuperscript𝑑2𝑣subscript𝑋˘𝑣32subscript𝑍˘𝑣\displaystyle\frac{1}{d^{2}_{v}}\left(X_{\breve{v}}-\frac{3}{2}Z_{\breve{v}}% \right)\,,divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ( italic_X start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - divide start_ARG 3 end_ARG start_ARG 2 end_ARG italic_Z start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) , (18c)

where the valence-shell single-particle energy computed in the CS core

ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT tvv+hv¯vhvh+12hhw¯vhhvhh,absentsubscript𝑡𝑣𝑣subscriptsubscript¯𝑣𝑣𝑣12subscriptsuperscriptsubscript¯𝑤𝑣superscript𝑣superscript\displaystyle\equiv t_{vv}+\sum_{h}\overline{v}_{vhvh}+\frac{1}{2}\sum_{hh^{% \prime}}\overline{w}_{vhh^{\prime}vhh^{\prime}}\,,≡ italic_t start_POSTSUBSCRIPT italic_v italic_v end_POSTSUBSCRIPT + ∑ start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_v italic_h italic_v italic_h end_POSTSUBSCRIPT + divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (19)

and the mvsubscript𝑚𝑣m_{v}italic_m start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT-independent quantities191919The quantities introduced in Eq. 20 being independent of mvsubscript𝑚𝑣m_{v}italic_m start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT, an additional sum over mvsubscript𝑚𝑣m_{v}italic_m start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT simply delivers a factor dvsubscript𝑑𝑣d_{v}italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT.

Uv˘subscript𝑈˘𝑣\displaystyle U_{\breve{v}}italic_U start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT mvdv(v¯vvvv+hw¯vvhvvh)absentsuperscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝑣𝑣superscript𝑣𝑣superscript𝑣subscriptsubscript¯𝑤𝑣superscript𝑣𝑣superscript𝑣\displaystyle\equiv\sum_{m_{v^{\prime}}}^{d_{v}}\left(\overline{v}_{vv^{\prime% }vv^{\prime}}+\sum_{h}\overline{w}_{vv^{\prime}hvv^{\prime}h}\right)≡ ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ( over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + ∑ start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h end_POSTSUBSCRIPT )
mvdv𝕧¯vvvv,absentsuperscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\displaystyle\equiv\sum_{m_{v^{\prime}}}^{d_{v}}\overline{\mathbb{v}}_{vv^{% \prime}vv^{\prime}}\,,≡ ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (20a)
Δv˘subscriptΔ˘𝑣\displaystyle\Delta_{\breve{v}}roman_Δ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ηvmvdv(v¯vv¯vv¯+hw¯vv¯hvv¯h)ηvabsentsubscript𝜂𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝑣𝑣¯𝑣superscript𝑣superscript¯𝑣subscriptsubscript¯𝑤𝑣¯𝑣superscript𝑣superscript¯𝑣subscript𝜂superscript𝑣\displaystyle\equiv\eta_{v}\sum_{m_{v^{\prime}}}^{d_{v}}\left(\overline{v}_{v% \bar{v}v^{\prime}\bar{v}^{\prime}}+\sum_{h}\overline{w}_{v\bar{v}hv^{\prime}% \bar{v}^{\prime}h}\right)\eta_{v^{\prime}}≡ italic_η start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ( over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_v over¯ start_ARG italic_v end_ARG italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + ∑ start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v over¯ start_ARG italic_v end_ARG italic_h italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h end_POSTSUBSCRIPT ) italic_η start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT
ηvmvdv𝕧¯vv¯vv¯ηv,absentsubscript𝜂𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝕧𝑣¯𝑣superscript𝑣superscript¯𝑣subscript𝜂superscript𝑣\displaystyle\equiv\eta_{v}\sum_{m_{v^{\prime}}}^{d_{v}}\overline{\mathbb{v}}_% {v\bar{v}v^{\prime}\bar{v}^{\prime}}\eta_{v^{\prime}}\,,≡ italic_η start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v over¯ start_ARG italic_v end_ARG italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (20b)
Xv˘subscript𝑋˘𝑣\displaystyle X_{\breve{v}}italic_X start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT mvdvmv′′dvw¯vvv′′vvv′′,absentsuperscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣′′subscript𝑑𝑣subscript¯𝑤𝑣superscript𝑣superscript𝑣′′𝑣superscript𝑣superscript𝑣′′\displaystyle\equiv\sum_{m_{v^{\prime}}}^{d_{v}}\sum_{m_{v^{\prime\prime}}}^{d% _{v}}\overline{w}_{vv^{\prime}v^{\prime\prime}vv^{\prime}v^{\prime\prime}}\,,≡ ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (20c)
Yv˘subscript𝑌˘𝑣\displaystyle Y_{\breve{v}}italic_Y start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT mvdvmv′′dvw¯vvv¯vv′′v¯′′ηvηv′′,absentsuperscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣′′subscript𝑑𝑣subscript¯𝑤𝑣superscript𝑣superscript¯𝑣𝑣superscript𝑣′′superscript¯𝑣′′subscript𝜂superscript𝑣subscript𝜂superscript𝑣′′\displaystyle\equiv\sum_{m_{v^{\prime}}}^{d_{v}}\sum_{m_{v^{\prime\prime}}}^{d% _{v}}\overline{w}_{vv^{\prime}\bar{v}^{\prime}vv^{\prime\prime}\bar{v}^{\prime% \prime}}\,\eta_{v^{\prime}}\eta_{v^{\prime\prime}}\,,≡ ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (20d)
Zv˘subscript𝑍˘𝑣\displaystyle Z_{\breve{v}}italic_Z start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ηvmvdvmv′′dvw¯vv¯v′′vv¯v′′ηv,absentsubscript𝜂𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣′′subscript𝑑𝑣subscript¯𝑤𝑣¯𝑣superscript𝑣′′superscript𝑣superscript¯𝑣superscript𝑣′′subscript𝜂superscript𝑣\displaystyle\equiv\eta_{v}\sum_{m_{v^{\prime}}}^{d_{v}}\sum_{m_{v^{\prime% \prime}}}^{d_{v}}\overline{w}_{v\bar{v}v^{\prime\prime}v^{\prime}\bar{v}^{% \prime}v^{\prime\prime}}\,\eta_{v^{\prime}}\,,≡ italic_η start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v over¯ start_ARG italic_v end_ARG italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT over¯ start_ARG italic_v end_ARG start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT italic_η start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT , (20e)

have been introduced to express the results in a compact way. Equations (20a) and (20b) make use of the effective valence-shell two-body matrix elements

𝕧¯vvv′′v′′′subscript¯𝕧𝑣superscript𝑣superscript𝑣′′superscript𝑣′′′\displaystyle\overline{\mathbb{v}}_{vv^{\prime}v^{\prime\prime}v^{\prime\prime% \prime}}over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT v¯vvv′′v′′′+hw¯vvhv′′v′′′h,absentsubscript¯𝑣𝑣superscript𝑣superscript𝑣′′superscript𝑣′′′subscriptsubscript¯𝑤𝑣superscript𝑣superscript𝑣′′superscript𝑣′′′\displaystyle\equiv\overline{v}_{vv^{\prime}v^{\prime\prime}v^{\prime\prime% \prime}}+\sum_{h}\overline{w}_{vv^{\prime}hv^{\prime\prime}v^{\prime\prime% \prime}h}\,,≡ over¯ start_ARG italic_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + ∑ start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT over¯ start_ARG italic_w end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h italic_v start_POSTSUPERSCRIPT ′ ′ end_POSTSUPERSCRIPT italic_v start_POSTSUPERSCRIPT ′ ′ ′ end_POSTSUPERSCRIPT italic_h end_POSTSUBSCRIPT , (21)

incorporating the contribution from the initial three-body interaction associated with an averaging over the CS core.

As demonstrated by Eqs. 17-18, the HFB-ZP energy is manifestly202020Equation 17 displays the explicit dependence of ΔEHFB-ZPΔsuperscript𝐸HFB-ZP\Delta E^{\text{HFB-ZP}}roman_Δ italic_E start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT on avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT. However, additional implicit dependences are in fact at play in Eq. 17. First, the two-body part of the center-of-mass kinetic energy correction included in the two-body interaction matrix elements actually depends on A. Second, all matrix elements at play carry an implicit dependence on avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT through the nature of their indices. Indeed, canonical single-particle states are nucleus-dependent and thus evolve as the valence shell is being filled, i.e. with avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT. However, it was checked numerically that both effects are largely subleading. cubic with the number of valence nucleons. While the cubic term originates entirely from the three-nucleon interaction, the curvature βv˘subscript𝛽˘𝑣\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT of the HFB-ZP energy relates to a specific linear combination of two- and three-body matrix elements that can be extracted from actual HFB calculations. The sign of this combination of matrix elements determines the convexity or concavity character through the open shell, under the assumption that the cubic term is subleading, which can also be directly checked from a subset of three-body matrix elements.

The two-neutron separation energy of open-shell nuclei is given, for av2subscript𝑎𝑣2a_{v}\geq 2italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ≥ 2, by

S2nHFB-ZP(av)subscriptsuperscript𝑆HFB-ZP2𝑛subscript𝑎𝑣absent\displaystyle S^{\text{HFB-ZP}}_{2n}(a_{v})\equivitalic_S start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) ≡ ΔEHFB-ZP(av2)ΔEHFB-ZP(av)Δsuperscript𝐸HFB-ZPsubscript𝑎𝑣2Δsuperscript𝐸HFB-ZPsubscript𝑎𝑣\displaystyle\Delta E^{\text{HFB-ZP}}(a_{v}\!-\!2)\!-\!\Delta E^{\text{HFB-ZP}% }(a_{v})roman_Δ italic_E start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT - 2 ) - roman_Δ italic_E start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT )
=\displaystyle== 2(αv˘βv˘+23γv˘)2(βv˘γv˘)avγv˘av2.2subscript𝛼˘𝑣subscript𝛽˘𝑣23subscript𝛾˘𝑣2subscript𝛽˘𝑣subscript𝛾˘𝑣subscript𝑎𝑣subscript𝛾˘𝑣superscriptsubscript𝑎𝑣2\displaystyle\!-\!2(\alpha_{\breve{v}}\!-\!\beta_{\breve{v}}\!+\!\frac{2}{3}% \gamma_{\breve{v}})\!-\!2(\beta_{\breve{v}}\!-\!\gamma_{\breve{v}})a_{v}\!-\!% \gamma_{\breve{v}}\,a_{v}^{2}\,.- 2 ( italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + divide start_ARG 2 end_ARG start_ARG 3 end_ARG italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) - 2 ( italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT - italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT . (22)

Under the (realistic) assumption that |αv˘||βv˘||γv˘|much-greater-thansubscript𝛼˘𝑣subscript𝛽˘𝑣much-greater-thansubscript𝛾˘𝑣|\alpha_{\breve{v}}|\gg|\beta_{\breve{v}}|\gg|\gamma_{\breve{v}}|| italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT | ≫ | italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT | ≫ | italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT |, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT starts at 2αv˘2subscript𝛼˘𝑣-2\alpha_{\breve{v}}- 2 italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT and evolves linearly throughout the open shell with a negative (positive) slope 2βv˘2subscript𝛽˘𝑣-2\beta_{\breve{v}}- 2 italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT when the energy is convex (concave).

Following Eq. (A.1), the two-neutron shell gap is given, for av2subscript𝑎𝑣2a_{v}\geq 2italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ≥ 2, by

Δ2nHFB-ZP(av)subscriptsuperscriptΔHFB-ZP2𝑛subscript𝑎𝑣absent\displaystyle\Delta^{\text{HFB-ZP}}_{2n}(a_{v})\equivroman_Δ start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) ≡ S2nHFB-ZP(av)S2nHFB-ZP(av)(av+2)subscriptsuperscript𝑆HFB-ZP2𝑛subscript𝑎𝑣subscriptsuperscript𝑆HFB-ZP2𝑛subscript𝑎𝑣subscript𝑎𝑣2\displaystyle S^{\text{HFB-ZP}}_{2n}(a_{v})\!-S^{\text{HFB-ZP}}_{2n}(a_{v})(a_% {v}+2)italic_S start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) - italic_S start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT + 2 )
=\displaystyle== 4βv˘+4γv˘av.4subscript𝛽˘𝑣4subscript𝛾˘𝑣subscript𝑎𝑣\displaystyle 4\beta_{\breve{v}}+4\gamma_{\breve{v}}a_{v}\,.4 italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + 4 italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT . (23)

Under the (realistic) assumption that |βv˘||γv˘|much-greater-thansubscript𝛽˘𝑣subscript𝛾˘𝑣|\beta_{\breve{v}}|\gg|\gamma_{\breve{v}}|| italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT | ≫ | italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT |, Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT is constant throughout the open shell with a positive (negative) value when the energy is convex (concave).

Eventually, the evolution of the valence-shell single-particle energy as a function of avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT is given by

ϵv˘HFB-ZP(av)=subscriptsuperscriptitalic-ϵHFB-ZP˘𝑣subscript𝑎𝑣absent\displaystyle\epsilon^{\text{HFB-ZP}}_{\breve{v}}(a_{v})=italic_ϵ start_POSTSUPERSCRIPT HFB-ZP end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) = ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT
+1dv(Uv˘+14Yv˘)av1subscript𝑑𝑣subscript𝑈˘𝑣14subscript𝑌˘𝑣subscript𝑎𝑣\displaystyle+\frac{1}{d_{v}}\left(U_{\breve{v}}+\frac{1}{4}Y_{\breve{v}}% \right)a_{v}+ divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ( italic_U start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + divide start_ARG 1 end_ARG start_ARG 4 end_ARG italic_Y start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT
+12dv2(Xv˘12Yv˘)av2.12subscriptsuperscript𝑑2𝑣subscript𝑋˘𝑣12subscript𝑌˘𝑣subscriptsuperscript𝑎2𝑣\displaystyle+\frac{1}{2d^{2}_{v}}\left(X_{\breve{v}}-\frac{1}{2}Y_{\breve{v}}% \right)a^{2}_{v}\,.+ divide start_ARG 1 end_ARG start_ARG 2 italic_d start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ( italic_X start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - divide start_ARG 1 end_ARG start_ARG 2 end_ARG italic_Y start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) italic_a start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT . (24)

The valence-shell single-particle energy contains linear and quadratic contributions in avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT, the coefficient of the former (latter) being closely related to the curvature (cubic coefficient) of the HFB-ZP energy.

A.2 Equal-filling approximation

While the previous section provides analytical expressions derived within the frame of the HFB formalism in the zero-pairing limit Duguet et al. (2020), a simpler mean-field treatment of open-shell systems in absence of pairing correlations is provided by the HF theory in the equal filling approximation. While their results are closely related, the two formalisms are fundamentally different. Indeed, while HFB describes the system via a pure quantum state, the EFA is formulated within the frame of statistical quantum mechanics, i.e. the system is described in terms of a statistical density operator Perez-Martin and Robledo (2008).

Effectively, EFA results can be trivially obtained by setting Δv˘=Yv˘=Zv˘=0subscriptΔ˘𝑣subscript𝑌˘𝑣subscript𝑍˘𝑣0\Delta_{\breve{v}}=Y_{\breve{v}}=Z_{\breve{v}}=0roman_Δ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = italic_Y start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = italic_Z start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = 0 in the HFB-ZP formulae. Thus, Eqs. (17), (A.1) and (24) apply, but with the modified coefficients

αv˘subscript𝛼˘𝑣\displaystyle\alpha_{\breve{v}}italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT =ϵv˘CS,absentsubscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle=\epsilon^{\text{CS}}_{\breve{v}}\,,= italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT , (25a)
βv˘subscript𝛽˘𝑣\displaystyle\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT =1dvUv˘,absent1subscript𝑑𝑣subscript𝑈˘𝑣\displaystyle=\frac{1}{d_{v}}U_{\breve{v}}\,,= divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG italic_U start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT , (25b)
γv˘subscript𝛾˘𝑣\displaystyle\gamma_{\breve{v}}italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT =1dv2Xv˘.absent1subscriptsuperscript𝑑2𝑣subscript𝑋˘𝑣\displaystyle=\frac{1}{d^{2}_{v}}X_{\breve{v}}\,.= divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG italic_X start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT . (25c)

A.3 Discussion

As already mentioned, numerical applications deliver γv˘=0subscript𝛾˘𝑣0\gamma_{\breve{v}}=0italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = 0 in all cases under scrutiny. Furthermore, the pairing contributions to αv˘subscript𝛼˘𝑣\alpha_{\breve{v}}italic_α start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT and βv˘subscript𝛽˘𝑣\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT are also negligible such that the HF-EFA results for γv˘=0subscript𝛾˘𝑣0\gamma_{\breve{v}}=0italic_γ start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT = 0 give an excellent account of HFB-ZP under the form

ΔEHF-EFA(av)Δsuperscript𝐸HF-EFAsubscript𝑎𝑣\displaystyle\Delta E^{\text{HF-EFA}}(a_{v})roman_Δ italic_E start_POSTSUPERSCRIPT HF-EFA end_POSTSUPERSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =ϵv˘CSav+βv~2av2,absentsubscriptsuperscriptitalic-ϵCS˘𝑣subscript𝑎𝑣subscript𝛽~𝑣2subscriptsuperscript𝑎2𝑣\displaystyle=\epsilon^{\text{CS}}_{\breve{v}}a_{v}+\frac{\beta_{\tilde{v}}}{2% }a^{2}_{v}\,,= italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT + divide start_ARG italic_β start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG start_ARG 2 end_ARG italic_a start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT , (26a)
S2nHF-EFA(av)subscriptsuperscript𝑆HF-EFA2𝑛subscript𝑎𝑣\displaystyle S^{\text{HF-EFA}}_{2n}(a_{v})italic_S start_POSTSUPERSCRIPT HF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =2ϵv˘CS2βv~(av1),absent2subscriptsuperscriptitalic-ϵCS˘𝑣2subscript𝛽~𝑣subscript𝑎𝑣1\displaystyle=-2\epsilon^{\text{CS}}_{\breve{v}}-2\beta_{\tilde{v}}(a_{v}-1)\,,= - 2 italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - 2 italic_β start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT - 1 ) , (26b)
ϵv~HF-EFA(av)subscriptsuperscriptitalic-ϵHF-EFA~𝑣subscript𝑎𝑣\displaystyle\epsilon^{\text{HF-EFA}}_{\tilde{v}}(a_{v})italic_ϵ start_POSTSUPERSCRIPT HF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =ϵv˘CS+βv~av,absentsubscriptsuperscriptitalic-ϵCS˘𝑣subscript𝛽~𝑣subscript𝑎𝑣\displaystyle=\epsilon^{\text{CS}}_{\breve{v}}+\beta_{\tilde{v}}a_{v}\,,= italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + italic_β start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT , (26c)
Δv~HF-EFA(av)subscriptsuperscriptΔHF-EFA~𝑣subscript𝑎𝑣\displaystyle\Delta^{\text{HF-EFA}}_{\tilde{v}}(a_{v})roman_Δ start_POSTSUPERSCRIPT HF-EFA end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =4βv~.absent4subscript𝛽~𝑣\displaystyle=4\beta_{\tilde{v}}\,.= 4 italic_β start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT . (26d)

The evolutions of the total binding energy, the two-nucleon separation and the valence-shell single-particle energy, as one fills the valence shell, are strictly correlated and entirely driven by the valence-shell single-particle energy computed in the core ϵv˘CSsubscriptsuperscriptitalic-ϵCS˘𝑣\epsilon^{\text{CS}}_{\breve{v}}italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT (diagrammatically represented in Fig. 3) and by βv˘subscript𝛽˘𝑣\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT that is nothing but the average diagonal matrix elements of the effective valence-shell two-body interaction (diagrammatically represented in Fig. 4)

βv˘subscript𝛽˘𝑣\displaystyle\beta_{\breve{v}}italic_β start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT =1dvmvdv𝕧¯vvvv.absent1subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣\displaystyle=\frac{1}{d_{v}}\sum_{m_{v^{\prime}}}^{d_{v}}\overline{\mathbb{v}% }_{vv^{\prime}vv^{\prime}}\,.= divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT . (27)

More specifically, while the total energy is quadratic in avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT, the two-nucleon separation energy and the valence-shell single-particle energy are linear. The coefficient of the linear (quadratic) term in the former drives the initial value (slope) of the latter, knowing that the slopes of the two-nucleon separation energy and of single-particle energy are opposite.

A.4 Second-order MBPT

Having semi-analytical expressions as a function of avsubscript𝑎𝑣a_{v}italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT for the mean-field results in the zero-pairing limit, it is now relevant to investigate the addition of dynamical correlations.

This is presently done by evaluating the MBPT(2) corrections to the valence-shell single-particle energy computed in the CS core and to the valence-shell effective two-body interaction. To do so, the valence shell is taken to be doubly degenerate (v,v)𝑣superscript𝑣(v,v^{\prime})( italic_v , italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT )212121This setting is mandatory to compute the energy of two successive even isotopes via MBPT, i.e. to avoid actually dealing with open-shell systems given that pairing was shown to be negligible for the present discussion and given that no perturbation theory based on a HF-EFA statistical operator is available to date. and the total energy is computed at the MBPT(2) level for both the CS core and the system with two more particles in order to compute the two-nucleon separation energy.

After a lengthy but straightforward derivation, the separation energy between the two even isotopes is obtained as

S2n(2)(2)subscriptsuperscript𝑆(2)2𝑛2\displaystyle S^{\text{(2)}}_{2n}(2)italic_S start_POSTSUPERSCRIPT (2) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( 2 ) =E(2)(0)E(2)(2)absentsuperscript𝐸20superscript𝐸22\displaystyle=E^{(2)}(0)-E^{(2)}(2)= italic_E start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT ( 0 ) - italic_E start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT ( 2 )
=2(ϵv˘CS+Σv˘(2)(ϵv˘CS))absent2subscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptΣ2˘𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle=-2\left(\epsilon^{\text{CS}}_{\breve{v}}+\Sigma^{(2)}_{\breve{v}% }(\epsilon^{\text{CS}}_{\breve{v}})\right)= - 2 ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) )
(𝕧¯vvvv+𝕧¯vvvv(2)(ϵv˘CS)),subscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscript¯𝕧2𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\hskip 11.38092pt-\left(\overline{\mathbb{v}}_{vv^{\prime}vv^{% \prime}}+\overline{\mathbb{v}}^{(2)}_{vv^{\prime}vv^{\prime}}(\epsilon^{\text{% CS}}_{\breve{v}})\right)\,,- ( over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + over¯ start_ARG blackboard_v end_ARG start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) ) , (28)

where the (on-shell) valence-shell self-energy and two-body effective interaction corrections are given by

Σv˘(2)(ϵv˘CS)subscriptsuperscriptΣ2˘𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\Sigma^{(2)}_{\breve{v}}(\epsilon^{\text{CS}}_{\breve{v}})roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) =+12hhp|𝕧¯hhvp|2ϵpCS+ϵv˘CSϵhCSϵhCSabsent12subscriptsuperscript𝑝superscriptsubscript¯𝕧superscript𝑣𝑝2subscriptsuperscriptitalic-ϵCS𝑝subscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptitalic-ϵCSsubscriptsuperscriptitalic-ϵCSsuperscript\displaystyle=+\frac{1}{2}\sum_{hh^{\prime}p}\frac{|\overline{\mathbb{v}}_{hh^% {\prime}vp}|^{2}}{\epsilon^{\text{CS}}_{p}+\epsilon^{\text{CS}}_{\breve{v}}-% \epsilon^{\text{CS}}_{h}-\epsilon^{\text{CS}}_{h^{\prime}}}= + divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_p end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_p end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT + italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_ARG
12pph|𝕧¯vhpp|2ϵpCS+ϵpCSϵhCSϵv˘CS,12subscript𝑝superscript𝑝superscriptsubscript¯𝕧𝑣𝑝superscript𝑝2subscriptsuperscriptitalic-ϵCS𝑝subscriptsuperscriptitalic-ϵCSsuperscript𝑝subscriptsuperscriptitalic-ϵCSsubscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\hskip 11.38092pt-\frac{1}{2}\sum_{pp^{\prime}h}\frac{|\overline{% \mathbb{v}}_{vhpp^{\prime}}|^{2}}{\epsilon^{\text{CS}}_{p}+\epsilon^{\text{CS}% }_{p^{\prime}}-\epsilon^{\text{CS}}_{h}-\epsilon^{\text{CS}}_{\breve{v}}}\,,- divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_h end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_h italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT + italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG , (29a)
𝕧¯vvvv(2)(ϵv˘CS)subscriptsuperscript¯𝕧2𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\overline{\mathbb{v}}^{(2)}_{vv^{\prime}vv^{\prime}}(\epsilon^{% \text{CS}}_{\breve{v}})over¯ start_ARG blackboard_v end_ARG start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) =+12hh|𝕧¯hhvv|22ϵv˘CSϵhCSϵhCSabsent12subscriptsuperscriptsuperscriptsubscript¯𝕧superscript𝑣superscript𝑣22subscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptitalic-ϵCSsubscriptsuperscriptitalic-ϵCSsuperscript\displaystyle=+\frac{1}{2}\sum_{hh^{\prime}}\frac{|\overline{\mathbb{v}}_{hh^{% \prime}vv^{\prime}}|^{2}}{2\epsilon^{\text{CS}}_{\breve{v}}-\epsilon^{\text{CS% }}_{h}-\epsilon^{\text{CS}}_{h^{\prime}}}= + divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_h italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG 2 italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h end_POSTSUBSCRIPT - italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_h start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_ARG
12pp|𝕧¯vvpp|2ϵpCS+ϵpCS2ϵv˘CS,12subscript𝑝superscript𝑝superscriptsubscript¯𝕧𝑣superscript𝑣𝑝superscript𝑝2subscriptsuperscriptitalic-ϵCS𝑝subscriptsuperscriptitalic-ϵCSsuperscript𝑝2subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\hskip 11.38092pt-\frac{1}{2}\sum_{pp^{\prime}}\frac{|\overline{% \mathbb{v}}_{vv^{\prime}pp^{\prime}}|^{2}}{\epsilon^{\text{CS}}_{p}+\epsilon^{% \text{CS}}_{p^{\prime}}-2\epsilon^{\text{CS}}_{\breve{v}}}\,,- divide start_ARG 1 end_ARG start_ARG 2 end_ARG ∑ start_POSTSUBSCRIPT italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT divide start_ARG | over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_p italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT | start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG start_ARG italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p end_POSTSUBSCRIPT + italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_p start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT - 2 italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT end_ARG , (29b)

and displayed diagrammatically in Figs 8 and 9, respectively. The second-order corrections to the total binding energy translate for the two-neutron separation energy into a correction of the mean-field valence-shell single-particle energy and of the effective valence-shell two-body interaction.

Extending candidly the situation to a dvsubscript𝑑𝑣d_{v}italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT-fold degenerate valence-shell in a EFA-like spirit, S2nsubscript𝑆2𝑛S_{2n}italic_S start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT and Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT evolve for av2subscript𝑎𝑣2a_{v}\geq 2italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ≥ 2 as

S2n(2)(av)subscriptsuperscript𝑆(2)2𝑛subscript𝑎𝑣\displaystyle S^{\text{(2)}}_{2n}(a_{v})italic_S start_POSTSUPERSCRIPT (2) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =2αv~(2)2βv~(2)(av1),absent2subscriptsuperscript𝛼2~𝑣2subscriptsuperscript𝛽2~𝑣subscript𝑎𝑣1\displaystyle=-2\alpha^{(2)}_{\tilde{v}}-2\beta^{(2)}_{\tilde{v}}(a_{v}-1)\,,= - 2 italic_α start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT - 2 italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT - 1 ) , (30a)
Δ2n(2)(av)subscriptsuperscriptΔ(2)2𝑛subscript𝑎𝑣\displaystyle\Delta^{\text{(2)}}_{2n}(a_{v})roman_Δ start_POSTSUPERSCRIPT (2) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_a start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT ) =4βv~(2),absent4subscriptsuperscript𝛽2~𝑣\displaystyle=4\beta^{(2)}_{\tilde{v}}\,,= 4 italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT , (30b)

with

αv~(2)subscriptsuperscript𝛼2~𝑣\displaystyle\alpha^{(2)}_{\tilde{v}}italic_α start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ϵv˘CS+Σv˘(2)(ϵv˘CS),absentsubscriptsuperscriptitalic-ϵCS˘𝑣subscriptsuperscriptΣ2˘𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\equiv\epsilon^{\text{CS}}_{\breve{v}}+\Sigma^{(2)}_{\breve{v}}(% \epsilon^{\text{CS}}_{\breve{v}})\,,≡ italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT + roman_Σ start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) , (31a)
βv~(2)subscriptsuperscript𝛽2~𝑣\displaystyle\beta^{(2)}_{\tilde{v}}italic_β start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over~ start_ARG italic_v end_ARG end_POSTSUBSCRIPT 1dvmvdv(𝕧¯vvvv+𝕧¯vvvv(2)(ϵv˘CS)),absent1subscript𝑑𝑣superscriptsubscriptsubscript𝑚superscript𝑣subscript𝑑𝑣subscript¯𝕧𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscript¯𝕧2𝑣superscript𝑣𝑣superscript𝑣subscriptsuperscriptitalic-ϵCS˘𝑣\displaystyle\equiv\frac{1}{d_{v}}\sum_{m_{v^{\prime}}}^{d_{v}}\left(\overline% {\mathbb{v}}_{vv^{\prime}vv^{\prime}}+\overline{\mathbb{v}}^{(2)}_{vv^{\prime}% vv^{\prime}}(\epsilon^{\text{CS}}_{\breve{v}})\right)\,,≡ divide start_ARG 1 end_ARG start_ARG italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_ARG ∑ start_POSTSUBSCRIPT italic_m start_POSTSUBSCRIPT italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT end_POSTSUBSCRIPT start_POSTSUPERSCRIPT italic_d start_POSTSUBSCRIPT italic_v end_POSTSUBSCRIPT end_POSTSUPERSCRIPT ( over¯ start_ARG blackboard_v end_ARG start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT + over¯ start_ARG blackboard_v end_ARG start_POSTSUPERSCRIPT ( 2 ) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT italic_v italic_v start_POSTSUPERSCRIPT ′ end_POSTSUPERSCRIPT end_POSTSUBSCRIPT ( italic_ϵ start_POSTSUPERSCRIPT CS end_POSTSUPERSCRIPT start_POSTSUBSCRIPT over˘ start_ARG italic_v end_ARG end_POSTSUBSCRIPT ) ) , (31b)

the latter being the averaged valence-shell interaction at second order in perturbation theory.

As seen in Eq. (30) and (31), dynamical correlations modify both the starting value of the S2n(2))subscriptsuperscript𝑆(2))2𝑛S^{\text{(2))}}_{2n}italic_S start_POSTSUPERSCRIPT (2)) end_POSTSUPERSCRIPT start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT in the valence-shell and the slope governing its evolution, i.e. the self-energy correction impacts the former whereas the correction to the valence-shell interaction modifies the latter.

Appendix B Δ2nsubscriptΔ2𝑛\Delta_{2n}roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT

The two-neutron shell gap defined in Eq. (2) explicitly reads as

Δ2n(N,Z)=E(N2,Z)2E(N,Z)+E(N+2,Z).subscriptΔ2𝑛𝑁𝑍𝐸𝑁2𝑍2𝐸𝑁𝑍𝐸𝑁2𝑍\Delta_{2n}(N,Z)=E(N-2,Z)-2E(N,Z)+E(N+2,Z).roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) = italic_E ( italic_N - 2 , italic_Z ) - 2 italic_E ( italic_N , italic_Z ) + italic_E ( italic_N + 2 , italic_Z ) . (32)

The second derivative of the total energy centered around N𝑁Nitalic_N can be written through finite difference coefficients as

2E(N,Z)N2=14(E(N2,Z)2E(N,Z)+E(N+2,Z)),superscript2𝐸𝑁𝑍superscript𝑁214𝐸𝑁2𝑍2𝐸𝑁𝑍𝐸𝑁2𝑍\dfrac{\partial^{2}E(N,Z)}{\partial N^{2}}=\dfrac{1}{4}(E(N-2,Z)-2E(N,Z)+E(N+2% ,Z))\,,divide start_ARG ∂ start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_E ( italic_N , italic_Z ) end_ARG start_ARG ∂ italic_N start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG = divide start_ARG 1 end_ARG start_ARG 4 end_ARG ( italic_E ( italic_N - 2 , italic_Z ) - 2 italic_E ( italic_N , italic_Z ) + italic_E ( italic_N + 2 , italic_Z ) ) , (33)

which proves that

2E(N,Z)N2=Δ2n(N,Z)4.superscript2𝐸𝑁𝑍superscript𝑁2subscriptΔ2𝑛𝑁𝑍4\dfrac{\partial^{2}E(N,Z)}{\partial N^{2}}=\dfrac{\Delta_{2n}(N,Z)}{4}\,.divide start_ARG ∂ start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT italic_E ( italic_N , italic_Z ) end_ARG start_ARG ∂ italic_N start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG = divide start_ARG roman_Δ start_POSTSUBSCRIPT 2 italic_n end_POSTSUBSCRIPT ( italic_N , italic_Z ) end_ARG start_ARG 4 end_ARG . (34)

References