1. Introduction
A new era of nuclear spectroscopy with heavy-ion beams has begun by the systematic use of heavy-ion single-charge-exchange (SCE) and, very recently, double-charge-exchange (DCE) reactions. While SCE studies reach back to the last century, DCE investigations are a rather new application of heavy-ion beams for spectroscopic studies. As reviewed in recent articles [
1,
2], over the last decade, DCE research has evolved rapidly, both on the experimental and the theoretical sides. Yet, compared to the status of SCE reaction physics, heavy-ion DCE research is still in its infancy. Going back in history, one finds that early studies of heavy-ion-induced DCE reactions were not pursued further after the first measurements at Los Alamos [
3], GANIL at Caen [
4] and NSCL at Michigan State University [
5] led to poor yields and conflicting results. The large differences in the reported cross sections have never been explored systematically: neither experimentally nor theoretically. Not in the least, the lack of an adequate theoretical framework for nuclear DCE reactions inhibited a clear interpretation of the data, leaving the question unsolved as to what degree the final DCE channels were indeed produced by sequential multi-nucleon transfer processes, which were the theoretically favored explanations for such reactions at the time of the early DCE data [
6,
7,
8,
9], leaving open the question of whether there are other competing reaction mechanisms.
From our recent theoretical investigations summarized in [
1,
2,
10], a more complete and complex picture of heavy-ion DCE reactions emerged. In general, heavy-ion DCE reactions are determined by three competing reaction mechanisms with different dynamical origins and spectroscopic content. Multi-step transfer reactions are always involved but can be suppressed to a negligible level by an appropriate choice of projectile and target nuclei and by performing experiments at incident energies sufficiently high above the Coulomb barrier. Transfer processes are maintained by the nuclear mean-field. As such, they are probing, in the first place, single-particle and nucleon-pair dynamics without directly accessing the isovector response of nuclei and the intrinsic isospin properties of nucleons as of interest for spectroscopic work and nuclear beta decay.
The nucleonic degrees of freedom beyond the mean-field are probed only in collisional NN interactions involving explicitly isovector NN interactions. In fact, nuclear-charge-exchange reactions will always contain collisional NN contributions. As in heavy-ion SCE reactions, the conditions of DCE reactions may be chosen as to enhance the collisional components. However, in practice, that will require researchers to always study the transfer route and the collisional route to the final DCE channel in order to keep as much as possible full control over the reaction. Of high importance for reliable reaction calculations are high-quality optical potentials, which always requires in parallel measurements of elastic scattering cross sections, at least for the incident channel configuration. The multi-method approach pursued by the NUMEN collaboration is based on such a far-reaching scheme and experimentally and theoretically investigates elastic scattering and the various intermediate transfer and collisional SCE reactions that possibly contribute as intermediate states to a DCE reaction.
As depicted in
Figure 1, in a heavy-ion DCE reaction, the isospin structure of hadrons and the isospin response of the nuclear medium can be probed in two ways: either by sequential action of the NN isovector interactions as a double single-charge-exchange (DSCE) reaction or in a virtual meson–nucleon double-charge-exchange reaction as a Majorana DCE (MDCE) reaction [
10,
11]. The MDCE mechanism is dominated by pion DCE. Utilizing the meson-exchange picture of nuclear interactions, the two collisional DCE mechanisms have in common that charge conversion proceeds by the exchange of virtual charged mesons between the reacting nuclei. An MDCE reaction is of first-order in an isotensor interaction dynamically generated by a pair of virtual
reactions. The peculiarities of MDCE reactions will be discussed in a separate paper. In this work, DSCE reactions will be considered. They are second-order reactions in which the final DCE channel is populated in a two-step manner by acting twice with the NN isovector T-matrix.
An appealing aspect of DCE physics is the possibility to probe the same nuclear configurations as encountered in double-beta-decay (DBD). That topic is gaining new and wide attention: see, e.g., [
12]. A comprehensive introduction and overview on DBD physics was given by Tomoda [
13]. The present status of DBD research was reviewed by Ejiri et al. [
14] and Agostini et al. [
15]. Cutting the meson lines vertically in the DSCE diagram in
Figure 1, a striking diagrammatical similarity of DSCE reactions to two-neutrino (
) DBD is discovered. However, dynamically, the two charge-converting processes are fundamentally different although they proceed by the same kind of spin–isospin operators. The origins and strengths of the underlying dynamics and the intrinsic structure of the interaction vertices are quite different, and the nuclear and weak coupling constants differ by orders of magnitude. But it is worthwhile to recall the established, fruitful connection between nuclear SCE reactions and single-neutrino beta decay (SBD) [
16,
17,
18,
19]. Light-ion SCE reactions were, in fact, used to indirectly collect information on DBD nuclear matrix elements [
14,
20,
21,
22,
23]. In combination, nuclear SCE and DCE reactions are the only known processes that allow the simultaneous study of the rank-1 isovector and the rank-2 isotensor responses of nuclei in laboratory experiments under reproducible conditions. It is worthwhile to mention that both kinds of collisional DCE reactions proceed by effective four-body isotensor interactions acting as two-body interactions in the participating nuclei. Heavy-ion DCE reactions are unique for providing for the first time the opportunity to investigate such high-rank operators.
Although in about the past two decades a lot of experience has been collected on multi-particle-hole dynamics in low-lying nuclear excitations—see e.g., [
24,
25,
26,
27] for experiments and theories on double-photon emission—only a little work has been spent on systematic investigations of nuclear DCE spectroscopy. Possible relations between DBD and double-photon processes were investigated quite recently by Jokiniemi and Menendez [
28]. For such comparative studies, heavy-ion DCE reactions offer new opportunities by giving access to another class of double-excitation processes driven by nuclear interactions.
In a previous article [
11], we have already discussed the theoretical methods required to convert the second-order DSCE reaction results, for which a description in a t-channel formalism is the natural approach, to second-order nuclear matrix elements, which are s-channel objects. As an independent additional issue, we present in this article an alternative approach by performing that transformation directly on the operator level. That method has the advantage of obtaining deeper insight into the DCE dynamics and the interplay of reaction and structure physics. As an important result, we derive the effective isotensor interactions that are generated in a DSCE reaction dynamically in a cooperative manner by the colliding ions.
The mentioned aspects will be discussed here for the DSCE reaction mechanism, but the ISI/FSI complex and the obtained results are of general relevance for all nuclear reactions of second-order distorted wave character. Similar features are also encountered in reactions proceeding by the MDCE mechanism; however, this is deserving of separate consideration because of the vastly different reaction dynamics. The MDCE scenario will be addressed separately in a follow-up paper.
The central goal of this paper is the “proof of principle” of how to extract from a heavy-ion DCE reaction the spectroscopic information under the conditions of strong ISI and FSI. Anticipating the result, such a separation is indeed possible in momentum representation where the second-order DSCE reaction amplitude is obtained as a folding integral of nuclear transition form factors and an ISI/FSI distortion kernel. Further progress is made by the closure approximation, for which an estimate of the neglected terms by sum rules is derived. Investigations of the intermediate propagator reveal that short-range correlations are induced between the two pairs of SCE vertices in the interacting nuclei, thus indicating that DSCE transitions are in fact two-nucleon processes. An important step for the proper definition of nuclear matrix elements is the transformation of the NN isovector interactions from the t-channel formulation used in reaction theory to the s-channel formalism appropriate for nuclear structure investigations. As a relatively easy to handle case, the formalism is used to evaluate the DSCE reaction amplitude in the Black Sphere (BS) strong absorption limit, which is appropriate for heavy-ion reactions.
In the following, we address two key questions related to DSCE theory:
Both issues are of vital importance for the aim of using heavy-ion DCE reactions as a spectroscopic tool, not the least as surrogate reactions for double-beta decay. Both questions are investigated in the context of double single-charge-exchange reactions treated as second-order processes of the NN isovector T-matrix.
In
Section 2, second-order DW reaction theory is recapitulated, and the DSCE reaction amplitude is presented in standard second-order form. The main result of
Section 2, however, is the introduction of the closure approximation for the DSCE amplitude. In
Section 3, an alternative, new formulation is presented for the DSCE reaction amplitude utilizing the momentum representation. The most important aspect is the separation of the nuclear structure and reaction dynamics.
Section 4 is devoted to the DSCE-NME. The t-channel operator structure of the product of T-matrices is recast into s-channel two-body operators acting in the projectile and target nucleus. We reconsider our previous formulation of the problem in [
11] now on the level of operators rather than on the level of matrix elements. The various aspects of the formalism are illustrated in
Section 5. The discussed issues are summarized, conclusions are drawn and an outlook to future work is given in
Section 6. The multipole structure of the effective induced isotensor interaction is presented in a number of the appendices.
5. Illustrative Applications for DSCE Reactions
5.1. The Reaction
The approach discussed in the previous sections has been applied to the DCE reaction
O
Ne
Ar at
MeV measured at LNS Catania [
32]. Following the approach discussed in [
10], the absorption radius
fm was derived from the total reaction cross section
b in the incident channel. The latter was obtained by an optical model calculation that included partial wave equations up to angular momentum
. A double folding potential was used and was calculated with Hartree–Fock–Bogolyubov (HFB) ground state densities of the A=18 and A=40 nuclei and a NN T-matrix [
2] that was newly derived for NN energies in the region
MeV in Love–Franey parametrization [
33,
34,
35].
For that reaction, the propagator contains a real-valued pole at
MeV/c corresponding to a kinetic energy
MeV if in the intermediate channel excitation energies are neglected. Adding excitation energy will move the pole to lower values of
, but qualitatively, the same results are obtained. As seen in
Figure 3, the propagator
is narrowly peaked at small distances
x, confirming the mentioned strong localization around
.
5.2. The Black Sphere Limit
Explicit calculations for strongly absorbing systems like interacting heavy ions show that the shape of the reduced distortion form factors
resembles a Heaviside distribution
, where
describes the excluded volume of the absorptive overlap region, known as the
black sphere (BS) approximation. In momentum space, the BS reaction kernels are obtained by the Fourier–Bessel transform of
, leading to the form factor
,
, where
and
is a spherical Riccati–Bessel function. In BS approximation, the reduced absorption form factor is given by
The function is strongly peaked at with a width . Since increases with the mass numbers, the width of decreases with increasing , approaching as the limiting case a delta distribution.
We also note that in the black sphere limit, the second-order form factor becomes
For
, the kernel attains an intriguing simple form:
The minus sign indicates the sizable reduction in the DSCE cross section by several orders of magnitude due to cancellation of the plane-wave part by the absorption exerted by the imaginary part of the ion–ion optical potential. In the following case studies, the BS approximation will be used, mainly because of its especially transparent structure and easy reproducibility.
5.3. Induced Correlations
An important ingredient of DSCE dynamics is the previously introduced correlation functions
. Since they are controlling the transition form factors, their properties are decisive for the observed NME and, moreover, for the whole reaction. As seen in
Figure 3, the propagator encountered in the reaction
O+
at
MeV has a maximum at
. At larger distances, the oscillatory pattern declines as
. Averaging the correlation functions over the orientations of
amounts to evaluating
where
is the zeroth-order Riccati–Bessel function, which oscillates with a period determined by the value of half the sum of the incident and exit channel momenta
.
For the above reaction, one finds in the measured angular range MeV/c, which is almost identical to the momentum MeV/c by which the propagator evolves. Hence, both functions oscillate with periods fm. In total, one finds that are strongly peaked at small values of . The short-range character is confirmed by the root mean square (rms) radius of the correlation function fm evaluated over a volume with a radius of half the distance between the centers of and . The rms value is in surprisingly good agreement with the two-nucleon correlation fm induced by the exchange of the Majorana neutrinos, as found in DBD studies. Thus, favors spatial configurations with , implying that the pairs of SCE vertices in the projectile and target are arranged in almost the same manner. The correlations, however, are of a rather fragile character. In addition to the incident energy and the nuclear masses, they also depend on the scattering angle through and on the (mean) intermediate excitation energy contained in .
5.4. The DSCE Cross Section in DW and BS Approximation
In
Figure 4, the results obtained with the second-order black sphere model are displayed and compared to the full second-order DW results of [
29] and the measured angular distribution. In [
29], second-order DW calculations were performed in a fully microscopic approach. The optical potentials were calculated in a double folding approach using HFB proton and neutron ground state densities and the isoscalar and isovector parts of the NN T-matrix. Also the ion–ion Coulomb potentials were obtained microscopically by double folding the nuclear charge densities with the two-body Coulomb interaction, including contributions due to anti-symmetrization; see, e.g., [
30,
36,
37]. The ion–ion potentials were checked against elastic scattering data. The optical model wave equations were integrated numerically for partial waves up to
, ensuring convergence for elastic and SCE and DCE angular distributions. The second-order DSCE reaction amplitude was constructed according to Equation (
7) but in pole approximation. The first-order SCE-type reaction amplitudes were calculated in partial wave representation and covered nuclear transitions with multipolarities
in the projectile and target. The nuclear transition form factors and response functions were obtained in QRPA calculations using the Giessen energy density functional [
26,
38,
39]. QRPA spectral distributions are found in [
10]. It is worth mentioning that the measured angular distribution is reproduced in absolute terms without additional adjustments. Note that the angular range covered by the data corresponds to a remarkable range of momentum transfers up to
MeV/c.
Following Equation (
41), the BS reaction amplitude is constructed by superimposing the plane-wave amplitude, which contains the DSCE-NME without ISI/FSI, and the distortion amplitude, which accounts for ISI/FSI. In [
29], second-order PW and second-order DW amplitudes were compared; the study impressively showed the anticipated strong suppression of the strength by about five orders of magnitude. In BS approximation, the same effect is produced by the interfering distortion amplitude. The DCE reaction is a
transition in both nuclei that constrains the total angular momentum transfer to
. However, as discussed in
Section 4, this is compatible to two combinations of total orbital and spin transfers,
and
, where
and
are the properly coupled total angular momentum and spin transfers in the first and the second step of the DCE reaction. Hence, the NME is given by an
and an
amplitude, where the monopole component dominates by far at small scattering angles. The PW amplitude, i.e., the ISI/FSI-free NME, is well-described by a zeroth and second-order Riccati–Bessel function with the same radius parameter
fm. For the ISI/FSI amplitude, we find
fm. The two kinds of amplitudes interfere with the phase
that was fitted to the data. Hence, the PW and ISI/FSI contributions interfere almost destructively, which explains the strong suppression of the transition strength.
As anticipated in
Section 3.2, the ISI and FSI terms not only induce strong suppression of the DSCE amplitude with respect to the plane-wave limit but also imprint on the DCE angular distributions their own diffraction structure caused by ISI and FSI and reflecting the size of the excluded overlap volume of strong absorption. The suppression effect and the modification of the diffraction structure by ISI and FSI were already illustrated in [
29], where second-order PWA and DWA angular distributions were compared. Thus, the diffraction pattern observed for heavy-ion DCE cross sections is mainly determined by the optical model potentials and only to a lesser degree by the transition form factors. The multipolarity of the form factors, however, determines the overall shape of an angular distribution, especially their behavior during small momentum transfers. That effect was already discussed in [
10] for SCE reactions, but in the DCE case, the ISI/FSI suppression is considerably enhanced.
In view of the simplicity of the black sphere model, the overall agreement between the two model calculations and the data is remarkable. At forward angles, i.e., small momentum transfers, the two theoretical approaches give almost identical results. The deviations, which develop with increasing scattering angles, indicate—as to be expected—the remaining differences of the schematic, semi-classical black sphere model to the quantal second-order DW calculation.
The question may arise as to what extent the BS results depend on the choice of the functional form of the distortion form factors. That point was checked by alternatively using form factors with a Fermi-function shape with a diffuse surface: thus modeling an opaque gray sphere. However, to reproduce with satisfactory accuracy the distortion coefficients obtained with the distorted waves from optical model calculations together with the DCE angular distribution, a rather small diffusivity parameter fm is required. With such a small value, the Fermi-distribution converges with very good approximation to a Heaviside distribution.
6. Summary and Discussion
Starting from a fully microscopic and quantal approach to a heavy-ion DSCE reaction, the role of initial state (ISI) and final state (FSI) interactions and the induced rank-2 isotensor interaction were investigated. The DSCE amplitude, being an, in principle, well known distorted wave two-step structure, was reconsidered in a momentum space approach that allowed the separation of the ISI and FSI contributions from nuclear matrix elements. Since heavy-ion DCE reactions are peripheral direct reactions, the elastic ion–ion interactions underlying ISI and FSI are well described by optical model potentials: either of empirical or theoretical origin. The accuracy achieved in the description of DCE data—as in all other kinds of direct nuclear reactions—depends, of course, on the quality of the optical potentials. In order to control that part of the theory, elastic scattering data for at least the incident system are indispensable as a probe for the optical potentials and their optimization.
The critical parts for understanding the induced suppression of transition strengths are the imaginary potentials. They are acting as sinks for the probability flux of the scattering waves by redirecting a large fraction of elastic flux into other reaction channels in a
never come back manner. Distortions and absorption effects were arranged into reaction kernels. An important aspect of ISI and FSI is that they affect the scattering wave functions. Hence, the relation to the optical potentials is of a complex nature that is mathematically defined by the underlying differential equation. The highly non-linear relation between the potential and wave amplitude is the reason that the absorption amplitudes are determined by the ion–ion total reaction cross sections [
10,
40]. Total reaction cross sections are the observable consequences of flux absorption by coupled channel dynamics.
In a nuclear DCE reaction, the transition form factors are the relevant nuclear structure objects. Nuclear matrix elements are obtained in the limit of vanishing momentum transfer, known as the long wavelength limit. By approximating the intermediate Green’s function by an average, channel-independent propagator, the DCE form factors can be investigated in closure approximation. Different to two-neutrino DBD, the closure approach is well justified for heavy-ion DCE reactions because nuclear interactions excite a large spectrum of intermediate states and multipolarities. detailed study of the interplay between the nuclear structure and reaction effects led to the interesting result that the intermediate propagation induces correlations between the first and second SCE events by imposing constraints on the vertices. ISI and FSI were discussed in the black sphere limit as a simple to handle but realistic approximation. The comparison of BS cross section results to full second-order DSCE calculations and DCE data led to surprisingly good agreement for the angular distribution.
A section was devoted to derive from the pair of two-body NN T-matrices the effective DSCE interaction. That goal was achieved by transforming the reaction-theoretical t-channel interactions into the s-channel representation required for nuclear structure studies. At the end, the products of t-channel operators were rearranged and recoupled into products of an s-channel operator. The s-channel interaction is given by tensorial dyadic products of two-body operators describing the and the associated complementary DCE transitions of two particle-two hole character in the projectile and target.
The BS model, albeit having convincing simplicity and surprising success, is not meant to replace a full microscopic description of DCE reactions. First of all, an optical model calculation is still needed for the determination of the absorption parameters. At sufficiently high energies—typically reached at about AMeV—the distortion and absorption effects can be treated safely by eikonal theory, which will simplify that step considerably. When extracting NMEs, it remains to understand their spectroscopic content. In general, that task will be confronted by disentangling a superposition of nuclear spin scalar and spin vector multipoles in the projectile and target. Thus, nuclear structure calculations remain an indispensable tool for identifying the spectroscopic content. Moreover, what we learn from the BS case studies is that the proper treatment of ISI/FSI effects is essential for a realistic description of the magnitude and the diffraction pattern of heavy-ion DCE angular distributions.
As a closing remark, we emphasize that modern nuclear reaction and nuclear structure theory, as used, e.g., in [
29], provides the proper theoretical tools and numerical methods for the quantitative description of processes as complex as a DCE reaction. The scope of this article was to clarify the role of ISI/FSI dynamics and their interplay with residual ion–ion interaction and to elucidate their cooperation for inducing correlations and effective isotensor interactions. It should be noted that the effective DSCE isotensor interaction is a four-body operator. Heavy-ion DCE reactions provide, for the first time, the environment appropriate to investigate such high-rank operators on a data-driven basis.