Ab initio description of monopole resonances in light- and medium-mass nuclei

Abstract

Giant Resonances are, with nuclear rotations, the most evident expression of collectivity in finite nuclei. These two categories of excitations, however, are traditionally described within different formal schemes, such that vibrational and rotational degrees of freedom are separately treated and coupling effects between those are often neglected. The present work puts forward an approach aiming at a consitent treatment of vibrations and rotations. Specifically, this paper is the last in a series of four dedicated to the investigation of the giant monopole resonance in doubly open-shell nuclei via the ab initio Projected Generator Coordinate Method (PGCM). The present focus is on the treatment and impact of angular momentum restoration within such calculations. The PGCM being based on the use of deformed mean-field states, the angular-momentum restoration is performed when solving the secular equation to extract vibrational excitations. In this context, it is shown that performing the angular momentum restoration only after solving the secular equation contaminates the monopole response with an unphysical coupling to the rotational motion, as was also shown recently for (quasi-particle) random phase approximation calculations based on a deformed reference state. Eventually, the present work based on the PGCM confirms that an a priori angular momentum restoration is necessary to handle consistently both collective motions at the same time. This further pleads in favor of implementing the full-fledged projected (quasi-particle) random phase approximation in the future.

1 Introduction

Giant Resonances (GRs) Berman and Fultz (1975); Harakeh and Woude (2001); Garg and Colò (2018); Colò (2022) are collective nuclear excitations that can best be pictured in terms of oscillations of the nuclear surface in an effective liquid-drop model. GRs are, in this respect, one of the clearest manifestations of collective motion in finite nuclei. The other most evident collective behaviour in nuclei is provided by rotational excitations Elliott (1958). These features have traditionally been described through empirical models adopting various resolutions, e.g. the macroscopic Bohr-Mottelson collective models Bohr and Mottelson (1998) and their microscopic counterparts Rowe (1985); Libert et al. (2016), the Interacting Boson Model Arima and Iachello (1976, 1978); Iachello and Arima (1987), the Fermion Dynamical Symmetry Model Wu et al. (1987), etc. In a microscopic framework, a unified and rigorous treatment of both rotations and vibrations is not amenable to simple solutions of the Schrödinger equation. For this reason, the two are often addressed separately. This is justified by the fact that nuclear rotations and vibrations typically pertain to different energy regimes. Indeed, rotations are low-energy excitations, with ground-state rotational bands typically spanning few MeV’s above the nuclear ground state. Instead, GRs typically appear above 10 MeV excitation energy with large differences depending on the system and on the multipolarity of interest.

When addressing doubly open-shell nuclei, both realms are typically approached starting from mean-field reference states breaking rotational symmetry, e.g. deformed Hartree-Fock-Bogoliubov (HFB) vacua. Rotations are then extracted by performing an angular momentum projection (AMP) acting either on one deformed HFB vacuum Sheikh and Ring (2000) or on a linear superposition of them, thus defining the projected generator coordinate method (PGCM) Bender et al. (2003); Robledo et al. (2018). Lately, rotational spectra have also been generated starting from deformed coupled cluster (CC) calculations with explicit AMP Hagen et al. (2022); Sun et al. (2024); Hu et al. (2024). As for GRs, the (quasi-particle) random phase approximation ((Q)RPA) is the usual method of choice, where nuclear vibrations are treated as harmonic fluctuations around the deformed HFB minimum Péru and Goutte (2008); Yoshida (2009); Avogadro and Nakatsukasa (2011); Beaujeault-Taudière et al. (2023), without any AMP.

In a recent work Porro et al. (2024a), an AMP in RPA calculations was attempted a posteriori, i.e. after solving the RPA secular equation based on the deformed reference state. The corresponding monopole response was shown to be contaminated with an unphysical coupling to the rotational motion and an empirical method was designed to subtract it. On a conceptually similar ground, when $U(1)$ symmetry is broken, QRPA pair transfer probabilities have been demonstrated to overestimate the exact results within the exactly solvable Richardson model Richardson and Sherman (1964). The most significant discrepancies occur near the transition from the normal to the superfluid phase Gambacurta and Lacroix (2012). On the other hand, the present series of four papers Porro et al. (2024b, c, d) addressing the giant monopole resonance (GMR) from an ab initio standpoint has demonstrated the suitability of the symmetry-conserving PGCM to address GRs. In this context, the goal of the present paper, the fourth of the series, is to investigate whether performing the AMP a posteriori on top of the GCM solutions induces the same shortcomings as those observed within the RPA framework and thus confirm that the a priori AMP is mandatory to properly handling the coupling between rotational and vibrational motions.

The present paper, denoted as Paper IV, is organized as follows. First, the different strategies to perform a symmetry restoration within the (P)GCM are formally introduced in Sec. 2 before discussing the impact on the strength function in Sec. 3. In Sec. 4, the empirical method used to remove the spurious coupling due to the a posteriori AMP is detailed. Numerical results are eventually presented in Sec. 5 whereas conclusions are drawn in Sec. 6 .

2 Formalism

The (P)GCM formalism was introduced in detail in Paper I Porro et al. (2024b). The essential elements needed to discuss the effects of the symmetry breaking and restoration in strength functions are only briefly recalled below.

2.1 The generator coordinate method

The original GCM formulation is introduced first. The GCM wave-function ansatz Hill and Wheeler (1953); Griffin and Wheeler (1957) is a general superposition of so-called generating functions reading as

$$\ket{\Psi_{\nu}}\equiv\sum_{q}f_{\nu}(q)\ket{\Phi(q)}\,,$$ (1)

where $q$ denotes a set of variables referred to as the generator coordinates. The index $\nu$ refers to a principal quantum number while $f_{\nu}(q)$ is a weight function to be determined. The ensemble $\{|\Phi(q),\,q\in[q_{0},q_{1}]\rangle\}$ denotes a set of non-orthogonal Bogoliubov states typically obtained as solutions of constrained HFB calculations requiring that the solution satisfies

$$\langle\Phi(q)|Q|\Phi(q)\rangle =q\,,$$ (2)

where $Q$ denotes a set of operators defining the collective coordinates.

The unknown coefficients $f_{\nu}(q)$ are determined variationally based on Ritz’ variational principle, namely by minimising the energy associated with $\ket{\Psi_{\nu}}$

$$\delta\frac{\braket{\Psi_{\nu}}{H}{\Psi_{\nu}}}{\braket{\Psi_{\nu}}{\Psi_{\nu}}}=0.$$ (3)

The variation with respect to the weights $f_{\nu}^{*}(q)$ eventually leads to a generalised eigenvalue problem known as the Hill-Wheeler-Griffin (HWG) secular equation Griffin and Wheeler (1957) reading as

$$\sum_{q}\Big{[}\mathcal{H}(p,q)-E_{\nu}\mathcal{N}(p,q)\Big{]}f_{\nu}(q)=0\,,$$ (4)

where the so-called Hamiltonian and norm kernels are respectively defined as

$$\mathcal{H}(p,q) \equiv\braket{\Phi(p)}{H}{\Phi(q)}\,,$$ (5a)

$$\mathcal{N}(p,q) \equiv\braket{\Phi(p)}{\Phi(q)}\,.$$ (5b)

2.2 The projected generator coordinate method

Constrained HFB solutions typically break symmetries of the Hamiltonian, so that restoring those symmetries is mandatory to obtain approximations to exact eigentates carrying good symmetry quantum numbers. The symmetry restoration is achieved within the PGCM by adding a projection operator onto good symmetry quantum numbers to the symmetry-breaking GCM state. While presently focusing on rotational symmetry associated with angular-momentum conservation, the approach is general and consists of modifying Eq. (1) according to

$$\ket{\Psi_{\nu}^{\sigma}}=\sum_{q}\,f_{\nu}^{\sigma}(q)P^{\sigma}\ket{\Phi(q)}\,,$$ (6)

where $P^{\sigma}$ is the projection operator associated with the (a) symmetry (sub)group $\mathcal{G}$ of the Hamiltonian. The projection operator selects the components of each $\ket{\Phi(q)}$ carrying the good symmetry quantum numbers $\sigma\equiv(JM\Pi NZ)$, i.e. the PGCM ansatz in Eq. (6) has good total angular momentum $J$ and angular momentum projection $M$, parity $\Pi=\pm 1$ as well as neutron $N$ and proton $Z$ numbers. The projector $P^{\sigma}$ can be generically written as

$$P^{\sigma}=\int\text{d}\varphi\,g^{\sigma}(\varphi)R(\varphi)\,,$$ (7)

where $g^{\sigma}(\varphi)$ represents the irreducible representations of $\mathcal{G}$ whereas $R(\varphi)$ denotes the unitary symmetry transformation operator changing the orientation of a state by an angle $\varphi$. The PGCM ansatz can thus be expanded as

$$\ket{\Psi_{\nu}^{\sigma}} =\sum_{q}\,f_{\nu}^{\sigma}(q)\int\text{d}\varphi\,g^{\sigma}(\varphi)R(\varphi)\ket{\Phi(q)}$$

$$\equiv\sum_{q}\,f_{\nu}^{\sigma}(q)\int\text{d}\varphi\,g^{\sigma}(\varphi)\ket{\Phi(q,\varphi)}\,,$$ (8)

where the $\varphi$-rotated Bogoliubov state $\ket{\Phi(q,\varphi)}\equiv~{}R(\varphi)\ket{\Phi(q)}$ has been introduced. Applying the variational procedure based on the PGCM ansatz (Eq. (2.2)) leads now to a set of $\sigma$-dependent HWG equations

$$\sum_{q}\,\Big{[}\mathcal{H}^{\sigma}(p,q)-E^{\sigma}_{\nu}\mathcal{N}^{\sigma}(p,q)\Big{]}f_{\nu}^{\sigma}(q)=0\,,$$ (9)

where the so-called symmetry-restored Hamiltonian and norm kernels are defined as

$$\mathcal{H}^{\sigma}(p,q) \equiv\braket{\Phi(p)}{HP^{\sigma}}{\Phi(q)}$$

$$=\int\text{d}\varphi\,g^{\sigma}(\varphi)\braket{\Phi(p)}{H}{\Phi(q,\varphi)}\,,$$ (10a)

$$\mathcal{N}^{\sigma}(p,q) \equiv\braket{\Phi(p)}{P^{\sigma}}{\Phi(q)}$$

$$=\int\text{d}\varphi\,g^{\sigma}(\varphi)\braket{\Phi(p)}{\Phi(q,\varphi)}\,.$$ (10b)

2.3 PAV-GCM

In the PGCM described above the secular equation (Eq. (9)) is solved in presence of the symmetry projection, i.e. the variational minimization is restricted to each irreducible representation of the symmetry group. For this reason, this scheme is presently denoted as the variation after projection GCM (VAP-GCM).

In between the GCM and the VAP-GCM, a scheme can be considered in which the symmetry projection is performed only after the GCM solution based on a symmetry breaking ansatz has been obtained. Such an intermediate approach is naturally denoted as the projection after variation GCM (PAV-GCM) scheme.

Projecting the GCM states solution of Eqs. (1)-(4), one works with the set of projected states

$$\ket{\tilde{\Psi}^{\sigma}_{\nu}} \equiv N_{\nu}^{\sigma}P^{\sigma}\ket{\Psi_{\nu}}=N_{\nu}^{\sigma}\sum_{q}f_{\nu}(q)P^{\sigma}\ket{\Phi(q)}\,,$$ (11)

where $N_{\nu}^{\sigma}$ is a normalising factor provided by the condition

$$1 =\braket{\tilde{\Psi}_{\nu}^{\sigma}}{\tilde{\Psi}_{\nu}^{\sigma}}\,,$$ (12)

such that

$$(N_{\nu}^{\sigma})^{-2}=\sum_{pq}f_{\nu}^{*}(p)f_{\nu}(q)\mathcal{N}^{\sigma}(p,q)\,.$$ (13)

2.4 Discussion

c:c Symmetry breaking
GCM (Q)RPA
      Large-amplitude superposition       of deformed HF(B) states       Harmonic fluctuations       around a deformed HF(B) state
Status: available Status: available
PAV GCM PAV (Q)RPA
      Angular-momentum projection       of deformed GCM states       Angular-momentum projection       of deformed (Q)RPA states
Status: developed in this work Status: developed in Refs. Erler (2012); Porro (2023); Porro et al. (2024a).
PGCM P(Q)RPA
      Proper treatment of rotation-vibration       coupling within the GCM       Proper treatment of rotation-vibration       coupling within the (Q)RPA
Status: available Status: formalism available Federschmidt and Ring (1985); Tsuchimochi and Van Voorhis (2015)
Symmetry conserving

Table 1: Schematic representation of different projection levels based on symmetry-breaking GCM and (Q)RPA.

The PAV strategy assumes that rotational and vibrational degrees of freedom are strictly decoupled, i.e. intrinsically deformed solutions are first obtained assuming that rotational degrees of freedom are frozen before adding the rotational motion to each vibrational state thus obtained.

From a physical standpoint, such a decoupling presuppose that vibrations and rotations relate to very different time scales, such that they can be addressed separately in a Born-Oppenheimer-like approximation. Specifically, rotations are assumed to be infinitely slower than nuclear vibrations, which is a direct consequence of the rotation being ideally associated with a zero-energy (Goldstone) mode.

However, nuclear rotations happen in fact at finite frequencies, such that they cannot be decoupled a priori from vibrational modes Bohr and Mottelson (1998). Thus, the variational/diagonalisation process at play to determine physical states should be performed in a Hilbert subspace simultaneously accounting for vibrational and rotational degrees of freedom. This is achieved in the VAP scheme that is the method of choice to consistently treat the coupling effects between nuclear vibrations and rotations. A schematic summary of the different levels of symmetry breaking and restoration in the GCM and the (Q)RPA is displayed in Table 2.4.

3 Transition strength

Generically, the ground-state strength function associated with an arbitrary excitation operator $O$ reads as

$$S_{F}(\omega)\equiv\sum_{\nu\sigma^{\prime}}|\braket{\Theta^{\sigma_{0}}_{0}}{O}{\Theta^{\sigma^{\prime}}_{\nu}}|^{2}\,\delta(E^{\sigma^{\prime}}_{\nu}-E^{\sigma_{0}}_{0}-\omega)\,,$$ (14)

where $|\Theta^{\sigma}_{\nu}\rangle$ ($E^{\sigma}_{\nu}$) denotes an eigenstate (eigenenergy) of the nuclear Hamiltonian.

The present work focuses on the monopole response and on the $K=0$ component of the quadrupole response, respectively associated with the excitation operators

$$O \equiv r^{2}\equiv\sum_{i=1}^{\text{A}}r_{i}^{2}\,.$$ (15)

and

$$O \equiv Q_{20}\equiv\sum_{i=1}^{A}r_{i}^{2}\,Y_{20}(\vartheta_{i},\phi_{i})\,,$$ (16)

where $(\vartheta_{i},\phi_{i})$ denote spherical angular coordinates. Below, the ingredients entering Eq. (14) in GCM, PAV-GCM and VAP-GCM calculations are specified.

3.1 GCM

The unprojected transition amplitude between the GCM ground state and a GCM excited state reads as

$$\braket{\Psi_{0}}{O}{\Psi_{\nu}} =\sum_{pq}f_{0}^{*}(p)f_{\nu}(q)\braket{\Phi(p)}{O}{\Phi(q)}$$

$$\equiv\sum_{pq}f_{0}^{*}(p)O(p,q)f_{\nu}(q)\,.$$ (17)

Using more compact notations, the transition matrix element between any two GCM states can be written as

$$\braket{\Psi_{\mu}}{O}{\Psi_{\nu}} \equiv\sum_{pq}f^{*}_{\mu p}O_{pq}f_{q\nu}\equiv(\textbf{f}^{\dagger}\cdot\textbf{O}\cdot\textbf{f})_{\mu\nu}\,,$$ (18)

where the indices of the matrix O associated with the operator kernel $O(p,q)$ run over the generator coordinates whereas the linear coefficient matrix f indices run on the generator coordinates for the lines and on the GCM states for the columns. Naturally, the energies entering Eq. (14) are the GCM energies delivered by Eq. (4).

3.2 VAP-GCM

In PGCM calculations, the transition amplitude reads as

$$\braket{\Psi_{0}^{\sigma_{0}}}{O}{\Psi_{\nu}^{\sigma}} =\sum_{pq}f_{0}^{\sigma_{0}*}(p)f_{\nu}^{\sigma}(q)\braket{\Phi(p)}{P^{\sigma_{0}\dagger}OP^{\sigma}}{\Phi(q)}$$

$$\equiv\sum_{pq}f_{0}^{\sigma_{0}*}(p)O^{\sigma_{0}\sigma}(p,q)f_{\nu}^{\sigma}(q)$$

$$\equiv\sum_{pq}f_{0p}^{\sigma_{0}*}O_{pq}^{\sigma_{0}\sigma}f_{q\nu}^{\sigma}$$

$$\equiv(\textbf{f}^{\sigma_{0}\dagger}\cdot\textbf{O}^{\sigma_{0}\sigma}\cdot\textbf{f}^{\sigma})_{0\nu}\,,$$ (19)

where the matrix $\textbf{O}^{\sigma\sigma^{\prime}}$ associated with the projected kernel $O^{\sigma\sigma^{\prime}}(p,q)$ carries the symmetry quantum numbers of both the bra and ket. Indeed, and contrary to the Hamiltonian, the operator $O$ is not necessarily a scalar under symmetry transformations. Naturally, the energies entering Eq. (14) are the PGCM energies delivered by Eq. (9).

3.3 PAV-GCM

Given the PAV-GCM states introduced in Eq. (11), the corresponding transition amplitude reads as

$$\braket{\tilde{\Psi}^{\sigma_{0}}_{0}}{O}{\tilde{\Psi}^{\sigma}_{\nu}} =N^{\sigma_{0}}_{0}N^{\sigma}_{\nu}\sum_{pq}f_{0}(p)f_{\nu}(q)\braket{\Phi(p)}{P^{\sigma_{0}\dagger}OP^{\sigma}}{\Phi(q)}$$

$$=N^{\sigma_{0}}_{0}N^{\sigma}_{\nu}(\textbf{f}^{\dagger}\cdot\textbf{O}^{\sigma_{0}\sigma}\cdot\textbf{f})_{0\nu}\,.$$ (20)

Up to a normalising factor, the PAV-GCM transition amplitude combines the projected kernels introduced in Eq. (19) with the mixing coefficients of the two involved GCM states.

In this scheme, the energies entering Eq. (14) are the PAV-GCM energies delivered by

$$\tilde{E}^{\sigma}_{\nu} \equiv\frac{\braket{\tilde{\Psi}^{\sigma}_{\nu}}{H}{\tilde{\Psi}^{\sigma}_{\nu}}}{\braket{\tilde{\Psi}^{\sigma}_{\nu}}{\tilde{\Psi}^{\sigma}_{\nu}}}=|N^{\sigma}_{\nu}|^{2}(\textbf{f}^{\dagger}\cdot\textbf{H}^{\sigma\sigma}\cdot\textbf{f})_{\nu\nu}\,,$$ (21)

where in fact a single projector is sufficient to compute the kernel given that the Hamiltonian is a scalar under rotation.

4 Spurious coupling to rotational motion

In this section the concept of spurious coupling to the rotational motion in theories breaking angular-momentum conservation is introduced. For the reasons mentioned in Sec. 2.2, vibrational GCM excitations obtained without AMP may be non-orthogonal to rotational states. This feature is considered spurious given that the neglect of the rotational degrees of freedom in the HWG equation precisely assumes that the intrinsic GCM states are fully decoupled from them. Based on this consideration, a method is now designed to subtract a posteriori the spurious coupling between excited GCM states and a pure rotational motion of the corresponding ground state.

4.1 Subtracted GCM

Given the overlap between an arbitrary GCM excited state $\ket{\Psi_{\nu}}$ and the PAV-GCM ground state

$$a_{\nu} \equiv\braket{\tilde{\Psi}^{\sigma_{0}}_{0}}{\Psi_{\nu}}\,,$$ (22)

the spurious coupling can be subtracted by redefining the excited state as Faessler and Nojarov (1988)

$$\ket{\breve{\Psi}_{\nu}}\equiv N_{\breve{\nu}}\big{[}\ket{\Psi_{\nu}}-a_{\nu}\ket{\tilde{\Psi}^{\sigma_{0}}_{0}}\big{]}\,.$$ (23)

It is immediate to check that the orthogonalisation condition

$$\braket{\tilde{\Psi}^{\sigma_{0}}_{0}}{\breve{\Psi}_{\nu}}=0\,$$ (24)

is satisfied and that the normalisation constant $N_{\breve{\nu}}$ is given by

$$(N_{\breve{\nu}})^{-2}=1-|a_{\nu}|^{2}\,.$$ (25)

Replacing $\ket{\Psi_{\nu}}$ by $\ket{\breve{\Psi}_{\nu}}$ in Eq. (17), the subtracted GCM (sub-GCM) transition strength can be computed.

4.2 Subtracted PAV-GCM

The subtraction method is extended to the PAV-GCM via the introduction of

$$\ket{\breve{\tilde{\Psi}}_{\nu}^{\sigma}}\equiv N_{\breve{\nu}}^{\sigma}P^{\sigma}\ket{\breve{\Psi}_{\nu}}\,,$$ (26)

where the coefficients $a_{\nu}$ are now defined to fulfill the orthogonalisation condition

$$\braket{\tilde{\Psi}^{\sigma_{0}}_{0}}{\breve{\tilde{\Psi}}_{\nu}^{\sigma}}=0\,.$$ (27)

In fact, $\ket{\breve{\tilde{\Psi}}_{\nu}^{\sigma}}$ differ from the PAV-GCM state $\ket{\tilde{\Psi}_{\nu}^{\sigma}}$ only if it carries the same symmetry quantum numbers as the ground state, i.e. $\sigma=\sigma_{0}$, which in the case of present interest corresponds to $J=0$ states. Correspondingly, the coefficient $a_{\nu}$ satisfies Eq. (22) and the normalising factor $N_{\breve{\nu}}^{\sigma_{0}}$ reads

$$(N_{\breve{\nu}}^{\sigma_{0}})^{-2} =\braket{\Psi_{\nu}}{P^{\sigma_{0}}}{\Psi_{\nu}}-|a_{\nu}|^{2}=(N_{\nu}^{\sigma_{0}})^{-2}-|a_{\nu}|^{2}\,.$$ (28)

Replacing $\ket{\tilde{\Psi}^{\sigma}_{\nu}}$ by $\ket{\breve{\tilde{\Psi}}^{\sigma}_{\nu}}$ in Eq. (20), the subtracted PAV-GCM (sub-PAV-GCM) transition strength can be computed.

5 Applications

The above considerations are illustrated below in the case of ²⁸Si that acts as a typical example. The general conclusions reached below have been checked to be valid in all the nuclei studied in Papers I, II and III.

5.1 Numerical setting

Calculations whose results are presented below are realised employing a spherical harmonic oscillator basis characterized by $\hbar\omega=12$ MeV and $\;e_{\!{}_{\;\text{max}}}=10$ and the chiral effective field theory ($\chi$EFT) Hamiltonian of Ref. Hüther et al. (2020) built at next-to-next-to-next-to-leading order (N³LO). Two-dimensional (P)GCM calculations are performed in axial symmetry using $(r,\beta_{2})$ as generator coordinates. All details of the calculations can be found in Paper II Porro et al. (2024c). Specifically, here particle-number projection is always included such that GCM calculations only omit the AMP, which is focus of the present work.

5.2 Levels of calculation

The different levels of calculations discussed in the present work are

• GCM: no AMP.

• PAV-GCM: AMP performed a posteriori on GCM states.

• VAP-GCM: full PGCM with AMP a priori.

• sub-GCM: GCM with a posteriori subtraction of the rotational coupling.

• sub-PAV-GCM: AMP performed a posteriori on sub-GCM states.

5.3 Results

Figure 1: Monopole (top) and quadrupole (bottom) responses for GCM and VAP-GCM calculations in ²⁸Si.

The monopole (quadrupole) GCM and VAP-GCM responses in ²⁸Si are compared in the upper (lower) panel of Fig. 1. The GCM monopole response is fragmented among four peaks in the interval $[11,23]$ MeV, the two dominant ones being located around $18$ MeV. These four peaks are fully correlated with those appearing in the quadrupole response, even though their relative weights are different in the two cases. This correlation is the fingerprint of the coupling between both modes due to the intrinsic (oblate) deformation of ²⁸Si. This topic was discussed at length in Paper II Porro et al. (2024c).

The inclusion of the AMP in VAP-GCM calculations impacts the responses in two ways. First, the excitation energy of the four dominant peaks are shifted up by $1.4$, $1.0$, $1.5$ and $0.7$ MeV, respectively, in the $J=0$ (monopole) channel and by $2.4$, $2.1$, $2.5$ and $1.7$ MeV in the $J=2$ (quadrupole) channel, i.e. the $J=0$ and $J=2$ states originating from the same intrinsic state are no longer strictly degenerate. Second, the intensity of the peaks is modified. The quadrupole response is significantly suppressed, i.e. while the strength of the first two peaks is divided by about a factor of three, the third peak has entirely disappeared and the fourth peak has been severely shrunk. The intensity of the peaks in the monopole response remains overall unchanged except that the relative weight of the two main contributions near $18$ MeV is strongly modified. Overall, the a priori inclusion of the AMP impacts the monopole and quadrupole responses non negligibly but it does so in a way that maintains a close connection to their intrinsic GCM counterparts.

Figure 2: Monopole (top) and quadrupole (bottom) responses for VAP-GCM and PAV-GCM calculations in ²⁸Si. For the monopole response, the overlap between excited GCM states and the PAV-GCM ground state is also displayed, the corresponding scale being shown on the right y-axis.

Next, VAP-GCM and PAV-GCM results are compared in Fig. 2. The quadrupole responses are very similar except for a shift up by about $1$ MeV for the $J=2$ VAP-GCM excitation energies compared to the PAV-GCM ones. Contrarily, the monopole responses differ very notably. As a matter of fact, the a posteriori AMP impacts the monopole amplitudes much more significantly than in the VAP-GCM calculation. This is particularly true beyond the first three peaks where the originally subleading fourth peak is very strongly enhanced, along with many significant peaks appearing at even higher energies where no GCM strength was visible in the first place. The anomalously large impact of the a posteriori AMP compared to the VAP-GCM results makes the validity of the PAV-GCM monopole response dubious.

In order to analyse the content of these results, the overlap $a_{\nu}$ between excited GCM states and the PAV-GCM ground-state is also displayed in Fig. 2. The PAV-GCM monopole strength happens to be anomalously large for GCM excited states that are strongly coupled to the PAV-GCM ground state, even though no associated strength was originally present in the GCM response.

Figure 3: Monopole (top) and quadrupole (bottom) responses for GCM and sub-GCM calculations in ²⁸Si.

Figure 4: Monopole response for VAP-GCM and sub-PAV-GCM calculations in ²⁸Si.

As discussed in Sec. 4, the spurious coupling to the rotational motion can be subtracted a posteriori from GCM and PAV-GCM results. Figure 3 demonstrates that, even though intrinsically-deformed GCM states carry such a spurious coupling, it has no impact on the associated monopole strength function. In particular, while the full AMP accomplished via VAP-GCM eventually improve over GCM results the latter still deliver a meaningful approximation of the former. Contrarily, the subtraction of the spurious coupling strongly corrects the PAV-GCM strength function as seen in Fig. 4 such that the sub-PAV-GCM monopole response becomes consistent with the VAP-GCM one. In fact, the sub-PAV-GCM monopole response remains very close to the original GCM one, i.e. once the spurious component is removed, the effect of the AMP is underestimated when performed a posteriori.

6 Conclusions

The impact of angular momentum projection (AMP) on the monopole and quadrupole responses of doubly open-shell nuclei has been investigated within the frame of the (projected) generator coordinate method ((P)GCM) based on intrinsically deformed mean-field states. More specifically, the objective was to investigate whether the AMP can be safely performed a posteriori, i.e. solving the secular equation for intrinsic states within the GCM and projecting the solutions on good angular momentum only afterwards. To do so, results were confronted with results from full PGCM calculations where the AMP is performed a priori, i.e. where the secular equation is solved directly for good-symmetry states.

Using ²⁸Si as typical example, the angular momentum projection was shown to have a non-negligible impact on both the monopole and quadrupole responses in full PGCM calculations. First, the position of the dominant peaks are shifted up by about $1$ MeV ($2$ MeV) in the $J=0$ ($J=2$) channel. Second, while only the relative weight of certain monopole transitions are modified, quadrupole transitions are strongly suppressed.

Next, the a posteriori angular-momentum restoration was shown to contaminate the monopole response with an unphysical coupling to the rotational motion, a result that is fully consistent with the one recently observed in (quasi-particle) random phase approximation calculations based on a deformed reference state Porro et al. (2024a). Eventually, the present work based on the PGCM confirms that an a priori angular momentum restoration is necessary to handle consistently rotational and vibrational collective motions at the same time, which further pleads in favor of implementing the full-fledged projected (quasi-particle) random phase approximation in the future.

Acknowledgements

Calculations were performed by using HPC resources from GENCI-TGCC (Contract No. A0130513012). A.P. was supported by the CEA NUMERICS program, which has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie grant agreement No 800945. A.P. and R.R. are supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – Projektnummer 279384907 – SFB 1245. R.R. acknowledges support though the BMBF Verbundprojekt 05P2021 (ErUM-FSP T07, Contract No. 05P21RDFNB). M.F. is supported by the CEA-SINET project.

Data Availability Statement

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