WorksPhysics and Astronomy
Uncertainty quantified three-body model applied to the two-neutron halo C
Abstract
Two-neutron halo nuclei offer a fascinating probe into the behaviour of quantum few-body systems at the limits of binding. Although few nuclei have already been clearly identified, many of their properties remain poorly constrained. C, one of the heaviest, still lacks a precise identification of its static and dynamic properties, such as its mass and dipole strength in the continuum. One main difficulty is that properties of two-neutron halo nuclei are inferred from experimental data using a theoretical model. Therefore, accurately determining the characteristics of two-neutron halo nuclei requires an accurate theoretical model and careful quantification of the uncertainties. In this work, we examine C with a three-body model, seeing C as a C core and two halo neutrons, and quantify for the first time the uncertainties associated with the C- interaction using a Bayesian approach. We propagate these uncertainties to properties of bound and scattering states of C, as well as its dipole strength. The comparison of our prediction for the matter radius to experimentally-derived values suggests that C is bound by less than 0.35~MeV and is dominated by a configuration. Our analysis of the dipole strength shows that final-state interaction needs to be included for an accurate description, the uncertainties on the strength function are about 50% and are mostly influenced by uncertainties on the ground-state properties, and partial-wave occupation of C depends on the scattering length and the resonance energy of the C- unbound system. Such sensitivity of the dipole strength to the properties of both C and C properties motivates a precise measurement of the C dipole strength function, that will allow to precisely and accurately resolve the spectroscopy of these nuclei.
Introduction and motivation
Recent experimental progress has allowed for measurements of isotopes near the driplines, exhibiting exotic features such as unexpected shell ordering [1, 2, 3] and clusterized structures [4, 5], thereby challenging traditional models of nuclear structure. Among the most extreme manifestation of clusterizing are halo nuclei [6], in which one or more loosely bound nucleons are found in spatially diffuse wavefunctions extending far beyond the core. Two-neutron halo nuclei, i.e. those composed of a core and two halo neutrons, can exhibit a Borromean structure, meaning that the three-body system forms a bound state, but none of the two-body subsystems is bound. These Borromean structures are strongly influenced by couplings to the continuum and offer an interesting probe of strongly correlated few-body systems including the dineutron [7, 8, 9, 10]. The identification of such exotic structures relies on state-of-the-art experimental techniques and robust theoretical models to interpret the data. One particularly interesting observable is the dipole strength distribution in the continuum, which exhibits a characteristic low-energy enhancement [11]. These dipole strengths are commonly inferred from Coulomb breakup measurements and have been studied for light halo nuclei, including [12, 13], [14], [15], [16] [17, 18] and [19]. From this dipole strength, one can extract, provided an accurate theoretical model, key properties of the halo systems, such as its binding energy. However, this interpretation is inherently model dependent and therefore introduces systematic uncertainties that must be carefully quantified.
Recent progress in ab initio nuclear theory has allowed structure calculations to reach the dripline, describing one-nucleon and even some two-neutron halo nuclei [20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42]. However, these approaches remain computationally demanding and have therefore only been applied to a limited number of systems. As a result, properties of two-neutron halo nuclei have typically been predicted using three-body models [7, 8, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52] which use the core and valence neutron degrees of freedom and accurately include couplings to continuum states. These models rely on effective core-neutron interaction, which are usually poorly-constrained due to the scarcity of experimental data. Provided that the three-body problem is solved accurately most of the uncertainties stem from these interactions. Because halo systems exhibit a natural separation of scales between the core and valence nucleons, they are ideal candidate for an effective field theory (EFT) description [53, 54, 55, 56, 57]. Such a model uses only a few low-energy constants, often the scattering lengths and binding energies. At leading order, only the binding energy is used
Email addresses: [email protected] (Patrick McGlynn), [email protected] (Chloë Hebborn) to constrain the EFT and the properties of the system are described by a universal behavior [58, 59, 60]. One main advantage of such approach is that the uncertainties associated with the truncation of the EFT can be quantified. Nevertheless, these models typically focus on a few low- orbitals, which makes it challenging to capture the structure of higher- shells in the nucleus. With more experimental data becoming available in the mid-mass region, halo structures are expected to be found around deformed cores, for which an accurate description of higher- orbitals will be needed.
is one of the heaviest confirmed two-neutron halos, and besides its existence, little is known for sure about it. Direct mass measurements [61] and mass evaluations [62] only set an upper bound for the two-neutron separation energy of keV. Two independent measurements of the total interaction cross section for suggested a large matter radius, but led to discrepant values [63, 64]. In terms of its shell structure, in the absence of low-energy continuum states in the system, one expects that core neutrons fill the level and the two halo neutron would be in a -wave. However, experimental studies suggest the presence of a virtual -wave state, which could enhance halo formation [65], as well as a low-energy resonance [66]. To further resolve the properties of and , it is essential to incorporate all available experimental information within a robust statistical framework, thereby improving the theory–experiment comparison used to refine their spectroscopy.
In this work, we quantify for the first time the uncertainties associated with the calibration of the effective - interaction using a Bayesian approach, and propagate those uncertainty to structure and reaction observables. We work within a framework of hyperspherical harmonics with an R-matrix approach [67], which treats consistently bound and scattering states. Using our uncertainty-quantified prediction, we investigate the competition between universal behaviour and single-particle structure of . We detail the Bayesian calibration of the - interaction in § 2. We present in § 3 our predictions for observables, and compare to the matter radius derived from experimental data [63, 64]. We also discuss how a precise measurement of dipole strength could be used to determine precisely the two-neutron separation energy and the single-particle structure of as well as the low-lying continuum states of . § 4 contains the conclusions and prospects of this work.
Calibration of two- and three-body parameters
In our three-body model, we describe as composed of an inert core in its ground state, to which two neutrons are loosely bound. We use a Hamiltonian composed of two-body interactions (- and -) and three-body interactions (more details about the formalism can be found in [67]). The - interaction used is the Minnesota potential [68, 69, 70], which reproduces low-energy - scattering. The - interaction is parametrised as a Woods-Saxon potential defined for each partial wave by a depth and global radius and diffuseness . We calibrate the core-neutron potential parameters and (the depths of the - and -wave potentials), (the strength of the spin-orbit force) and (the Woods-Saxon range), to the limited available information on low-lying spectrum using a Bayesian approach.
In the -wave, we consider the existence of a virtual state1, as suggested in [65]. In particular we use for the scattering length of this virtual state fm, using as mean value the one determined in [65]. Since the original work did not provide any uncertainty, we take a conservative approach and assign a large error to this scattering length. For the -wave, we consider the state to be bound by MeV, consistent with the peak of the - relative energy distribution seen in [66], and following their interpretation of that peak. We treat the state as completely occupied by neutrons in the core, and the to have no occupation by core neutrons. Since they do not identify another peak, the resonance is assumed not to be seen below MeV, which we also use to constrain our calibration. Finally, we require that there be exactly one bound state in the , and waves and no -wave resonance below MeV to match the structure of .
We use wide Gaussian priors for the potential parameters: both and use a prior centred at MeV with a width of MeV, consistent with the range of similar model calculations [46]. The geometry of the potential was informed by fm, giving a central value of fm and a width of fm. We fix the diffuseness to fm so that only the range controls the long-range behaviour. Since no relevant two-body data exists to constrain the -wave, was taken as the average of and . Finally the spin-orbit strength lacks an obvious choice of prior, so the value used in [46] was used to give the central value of MeV and a width of MeV was chosen. More details about the Bayesian calibration are given in Appendix A.
We propagate the uncertainties of the - interaction to the observables by performing three-body calculations for samples of our posterior distributions. As explained in [67], the Pauli-forbidden states of the core-neutron potential are removed using a projection operator and we use a three-body force to fix the two-neutron separation energy . This quantity is not well known for , with the two values MeV (as evaluated in AME2020 [62]) or MeV (as obtained from a direct mass measurement [61]), both being consistent with a range between and MeV. Previous calculations [52, 60, 71] agree with this range of separation energies, and tend to favour a smaller separation energy. To reflect this uncertainty and determine its effect on observables, we use five values of : , , , and MeV. These are controlled by modifying the strength of the three-body force. Since the three-body force only applies in the channel, the scattering state is unchanged. In the following section, we analyse the credible intervals for observables, that are computed with the model space described in Ref. [67].
Prediction for properties
To study how uncertainties in the -n interaction influence the halo character of , we naturally start by analyzing root-mean-squared (rms) matter radius. In a three-body model of , the matter radius can be separated into two components [46, 51]
where is the rms matter radius of the core and is the rms hyperradius of the three-body +n+n system. We show in Figure 1 our predictions for the rms hyperradius for various separation energies, with each violin shape obtained from the 315 samples of the -n interaction. As expected, the radius decreases when the separation energy increases. Interestingly, for all values of , two “modes” appear in the distribution of the rms radii: the majority of samples favour a larger hyperradius, while the minority show a preference for a much lower hyperradius, with less dependence on separation energy. Further investigation (detailed in Appendix B) shows that that the larger hyperradii corresponds to s-wave dominated samples, i.e. parameters in which the -n scattering length is large or the resonance energy is high, and the tail to smaller hyperradii correspond to the to d-wave dominated samples. The presence of these two modes demonstrates that rms radii are sensitive to the single-particle structure of halo systems.
Figure 1. Root-mean-square hyperradius as a function of two-neutron separation energy. Violin plots show a smoothed probability density function computed from 315 samples for each energy. Horizontal lines indicate the mean and intervals. The interval from Ref. [64] is shown. Note the cutoff of the probability density function at the extreme values does not imply it is impossible to exceed these values.
We now compare our predicted rms hyperradii with matter radii derived from measurements of interaction cross sections on a hydrogen target fm [63] and
on a carbon target fm [64]. Because the reported uncertainties do not include contributions from theoretical modelling in the analysis of the reaction data, they are likely underestimated, which could explain the observed discrepancies [72]. Using the experimentally-derived value of rms radius ( fm [73]), these two values correspond respectively to fm and fm. None of our calculations in Figure 1 are compatible with the value derived in [63] unless the core is assumed to have a much larger size, and only calculations reproducing MeV that are s-wave dominated are consistent with the value of Togano et al. [64]. This is consistent with previous predictions rms radius of [51, 52]. This suggests that the ground state has a s-wave dominated single-particle structure and is bound by less than 350 keV. Nevertheless, it would be worth revisiting this analysis in the future including the uncertainties associated with the description of the reaction with the target.
As previously mentioned, another interesting observable for halo nuclei is the dipole response in the continuum. This observable describes the excitation from the ground state to a state in the continuum, it is hence computed from both bound and scattering wavefunctions. Using our three-body framework of hyperspherical harmonics with an -matrix approach, both wavefunctions are computed consistently and with the correct boundary condition. Figure 2 shows our prediction for the strength function obtained with the 315 samples of the -n interaction and for different values (colours). As expected for a halo system [59, 60], we see a peak at low energy, with the energy of that peak increasing as the system becomes more bound. As shown in the inset Figure 2, the peak energy depends approximately linearly on the three-body binding energy, as expected from a universal picture.3 Nevertheless, the large errors on the slope of the line indicates deviation from this universal behavior. In terms of magnitude, the overall dipole strength decreases as the separation energy increases. This can be understood considering centers of mass and charge are closer together for more bound systems, resulting in a smaller dipole strength. In terms of uncertainties, the overall scale of the credible interval is large for all separation energies and corresponds approximately to a constant 50% error. This indicates a dependence on the details of the -n potential, again suggesting deviations from a truly universal picture.
Figure 2. 68% credible intervals for the dipole strength function of as a function of the relative energy of the three-body system. The different colours correspond to different two neutron separation energies. (Inset) Energy of the peak of the curve as a function of with error bars indicating 68% credible intervals. Points are computed using both unapproximated scattering waves and a plane wave approximation in the scattering channel to determine the effect of final state interactions (FSI).
To understand the influence of the scattering states on the dipole strength, we repeated these calculations considering no final-state interactions (FSI), i.e., approximating the scattering solutions with plane waves, rather than using the full solutions to the three-body problem. The plane-wave predictions lead to dipole strength distribution with a peak at a significantly higher energy than the full calculations (compare the empty squares and the filled circles in the inset of Figure 2). This shift in the energy of the peak is more pronounced for more deeply-bound states. That can also be intuitively understood considering the spatial extension of the two-neutron halo nucleus: a more bound system will be less spatially extended, and hence each cluster (core or halo neutron) is more sensitive to the interactions with the other clusters. Moreover, we show in Appendix § 8 that although including FSI affects the overall shape of the , it does not change the uncertainties on its magnitude. This indicates that the large uncertainties in stem from the bound-state description. This large sensitivity of the to the details of the bound state wavefunction confirms that this observable is a powerful tool to study the ground states of two-neutron halo systems. Nevertheless, using a model without FSI to extract properties of a two-neutron halo, such as its binding energy, from a measurement of would lead to a skewed result.
Similarly to the study on the rms hyperradius, we now investigate the sensitivity of to the single-particle structure of the ground state. Figure 3 shows predictions for the that corresponds to the -wave dominated states (colored bands). Relative to Figure 2, obtained from all - and -wave samples, credible intervals for are now much narrower and the upper limits of are similar across all energies. This reduction of uncertainties occurs because the relative error from the bound states in the -wave samples is much smaller than in the -wave samples (see Appendix § 8). The smaller amplitude of the -wave dominated calculation is explained partly by the more compact shape (seen in Figure 1 and caused by a larger centrifugal barrier), and partly by the fact that the E1 transition between the scattering state and ground state proceeds mostly to the -wave component of the bound state. The -wave curves are more consistent with the universal picture [9, 59] of a weakly-bound halo nucleus, i.e., the credible intervals in Figure 3 can be well described by simple functions of the binding energy.
Figure 3. Same as Figure 2 using only -wave dominated samples. (Inset) 68% credible ellipses for the peak energy and the value of at the peak, separated into - (colored ellipse) and -wave (unfilled ellipse) dominated samples.
The dipole strength shape and magnitude are strongly sensitive to both the two-neutron separation energy and the single-particle structure of ; a precise measurement of this observable would enable to more accurately determine properties.
In the inset of Figure 3, we show how a comparison with an experiment could provide two simultaneous pieces of information: the separation energy of from the position of the peak, and the relative contribution of - and -waves in the ground state from the peak height. Moreover, using our Bayesian framework, this theory-experiment comparison would also help clarify the properties of the low-lying spectrum of .
Conclusions
In this work, we have presented the first uncertainty-quantified three-body prediction of a two-neutron halo nucleus, using a Bayesian calibration of the core-neutron interaction. Focusing on , the properties of which remain poorly constrained, we employ a Bayesian approach to calibrate -n interaction on available information on system and propagate these uncertainties to properties of . This framework allows us to accurately describe both the universal features and the signatures of single-particle properties in the two-neutron halo observables.
Our study focuses on two key observables for halo systems: the rms matter radius and the dipole strength in the continuum. We find that both observables are strongly sensitive to the two-neutron separation energy and the partial-wave content of the ground state wavefunction. Comparing our model’s predictions to a matter radius derived from interaction cross-section measurements [64], our results suggest that has a two-neutron separation energy below 0.35 MeV and a ground state dominated by a configuration for the halo neutrons. In future work, we plan to revisit this analysis including the uncertainties associated with the analysis of the interaction cross sections, that were not included here.
We then turn to the dipole strength, and we highlight the importance of including final-state interactions for an accurate interpretation of experimental data. Our study also shows that dipole strengths corresponding to -wave dominated ground states exhibit the expected universal behaviour, i.e. the peak position depends linearly on the binding energy of the three-body system. This universality is lost for -wave dominated ground states. Moreover, our Bayesian approach would enable backpropagation of these data to the properties of , which remain poorly known, including the scattering length and the energy of the resonance. Such a study would clarify the shell structure of the levels in this region of the nuclear chart and would provide useful insights not only into this system, but also into the behaviour of loosely-bound neutron-rich nuclei in general.
Finally, this work demonstrates the viability of performing Bayesian uncertainty quantification of three-body calculations without approximating the three-body continuum. Applying this framework to other two-neutron halo systems would enhance theory-experiment comparisons and refine the properties of these systems inferred from data. Future works include studying other two-neutron halo nuclei and embedding our predictions in a reaction framework to enable direct comparison with experimental data, similarly to Ref. [46], rather than with derived values. This will allow more accurate comparisons with experiments on two-neutron halo systems.
Acknowledgements
We are grateful to Miguel Marques, Nigel Orr, Julien Gibelin and Daniel Phillips for insightful discussions related to this work. We also thank the few-body group at MSU for regular discussions and support. This project also received financial support from the CNRS through the AIQI-IN2P3 project. Calculations were performed using the High-Performance Computing Center at MSU’s Institute for Cyber-Enabled Research.
Calculation of Bayesian posterior
In this appendix, we provide more details about the Bayesian calibration of the - potential parameters and show the corner plot of the posterior distributions (Fig. A.4). The posterior distributions are obtained via a Monte-Carlo sampling process using the python package emcee [74] and are the result of 20000 steps using 32 walkers, after discarding the first 1000 steps as burn-in. The corner plot is obtained from 10000 samples selected randomly from the posterior distribution to build the cornerplot. The 315 samples used for the three-body model are drawn from this subset of 10000.
Fig. A.4 shows the posterior (histograms) alongside the prior (solid red line) distributions. Except for , none of the posterior distributions are prior-dominated, indicating that the calibration data efficiently informs the parameters. The influence of the prior for is unsurprising, given that the spin-orbit splitting can only be suitably constrained with information about both levels of a spin-orbit pair, or indirectly with enough information about resonance energies and widths in some cases. Since such information is not available, the spin-orbit parameter remains poorly constrained and would benefit from more precise information on the - system.
Figure A.4. Two-parameter corner plots of posterior distribution of - potential parameters. Red lines show the prior distribution used in the Bayesian calibration.
Fig. A.4 also shows that several of the posterior distributions are non-Gaussian, especially the distribution of , confirming the suitability of using a Bayesian approach over simpler frequentist methods [76] [75]. The apparent correlation between and wave depths and arises from the imposition of the same geometry in both partial waves. In all future calculations, unless stated otherwise, a random selection of 315 such samples is used, with that number being driven by the need for convergence of the three-body credible intervals.
Separation of calculations into s-wave and d-wave dominated samples
As shown in Figure 1, part of the three-body calculations predicts larger rms hyperradii, while another part yields smaller hyperradii; this behavior is observed across all separation energies. To better understand the origin of these two groups, we focus on the calculations reproducing MeV. We divide the samples into two categories: the ones that lead to rms hyperradii above 6.5 fm (246 out of 315 samples considered in this work) and the remaining samples, producing smaller rms hyperradii. The threshold of 6.5 fm corresponds approximately to the minimum of the probability distribution between the two modes. Interestingly, calculations from each group lead to ground states that have similar partial-wave decompositions.
Before presenting in details the partial-wave decomposition obtained for each group, we clarify the model employed in this work. Within our three-body model of two-neutron halo nucleus, two Jacobi sets are relevant [67]. The first, which we refer to as the Y-basis{4} (shell-model like), is defined by the core-neutron relative coordinate and the coordinate from the other neutron to the center of mass of the core-neutron subsystem . The second set, the T-basis, is defined by neutron-neutron and the core-dineutron relative coordinates and , respectively. Each basis carries its own set of orbital angular momenta: in the Y-basis and in the T-basis. The same state with the total angular momentum and parity has a different decomposition in terms of partial waves in each basis. In the specific case of C, modeled as a C core in its ground state with two halo neutrons, we predict a ground state and partial waves with and are populated.
Figure B.5. Occupation number of C ground state bound by 0.5 MeV per partial waves of the core-neutron system , . We separate the samples dominated by a (blue) and (red) configurations. The varying opacity of the bar correspond to the distribution of occupation numbers, while points with error bars indicate the mean occupation and 68% credible intervals.
Figure B.6. Same as Fig. B.5 but for the neutron-neutron angular momentum and the core-dineutron angular momentum, with .

Figure B.7. Scatter plot of resonance energy and -wave scattering length in C, coloured by whether the resultant three-body state has an rms hyperradius greater (resp. lower) than 6.5 fm, which is referred to as - (resp. -) wave dominated samples.
We first analyze the partial-wave decomposition of C ground states obtained in each group in the Y-basis in terms of and which couples the spin of one neutron to its orbital angular momentum . Fig. B.5 shows the occupation for the group leading to fm (blue histograms) and fm (red histograms). The first group consists of calculations which are almost entirely with only minor contributions from other partial waves. The second group corresponds to -wave dominated calculations which are characterized by configurations with admixtures ranging from 4 and 40% -wave. No samples have significant contributions from any odd parities or from , which is unsurprising since the bound states in both the and -waves are Pauli-forbidden. Such different occupation motivates the name "s-wave dominated" and "d-wave dominated" samples for calculations with fm and fm, respectively.
A complementary partial-wave decomposition of the ground state can be performed in the T-basis in Fig. B.6, which is more natural for interpreting C as a dineutron coupled to a C core. Regardless of whether the configurations are - or -wave dominated in the Y-basis, all samples exhibit at least , corresponding to an -wave dineutron. The main difference is in the different admixture: the -wave dominated samples tend to see roughly equal mixture of this configuration with the -wave dineutron component, while the -wave dominated samples tend to exhibit -wave dineutron spectroscopic factors around 30%. We also verified that the dependence of these partial-wave decompositions in both the Y- and T-basis on the binding energy is small.
Finally, we investigate how these two groups relate to the properties of the C system. Fig. B.7 shows all samples (blue and red points correspond to - and -wave configurations, respectively) as a function of the energy position of the resonance in C and the -wave scattering length. This figure shows that -wave (resp. -wave) dominated samples are associated with higher (resp. lower) resonance energies and larger (resp. smaller) scattering lengths. This analysis confirms that observables such as rms radii and dipole strengths in the continuum are closely related to the partial-wave decomposition of the ground state of two-neutron halo nuclei, as well as to properties of the core-neutron system.
Sensitivity of to scattering states and its uncertainty budget
In this appendix, we investigate the sensitivity of to the description of the three-body scattering states. First, we analyze the importance of FSI in Fig. C.8 by comparing dipole strength for a ground state bound by 0.2 MeV that are obtained with the calculated three-body scattering waves along with similar calculations using plane waves, i.e., by using zero -n and n-n potential to generate the scattering states. As already mentioned in § 3, FSI influence both the position of the peak and the shape of the distribution. This indicates that predictions for using plane waves are not accurate. Interestingly, the credible intervals are similar in both cases, indicating that most uncertainties in the dipole strength arises from the description of the ground state while uncertainties in the scattering states are comparatively smaller.
Figure C.8. 68% credible intervals for the strength function calculated with MeV using a plane wave approximation (gray band) and full scattering waves (red band, also shown in Figure 2).
We further investigate the uncertainty budget of (that includes FSI) in Fig. C.9, which shows the relative contribution of bound and scattering states to the total error. The uncertainties associated with the description of ground state dominate the total uncertainty, accounting for (solid black line) out of approximately a total error (see Figure 2), while those associated with the scattering states are significantly smaller, around (dashed black line). To assess the influence of the single-particle structure on the uncertainty budget of , we perform a similar analysis separately for samples which have a - and -wave dominated ground states (blue and red lines in Fig. C.9). In both cases, the error on the is dominated by the ground-state description, while contribution from the scattering-sate uncertainties remains small.
Figure C.9. Relative error, defined as the ratio of the standard deviation of the sample to their mean, for calculations varying only bound (solid lines) or scattering states (dashed lines). We consider all 315 samples (black lines), only the -wave dominated (blue lines) and only the -wave dominates ones (red lines).
For completeness, we provide in Fig. C.10 the 68% and 95% credible intervals for the largest eigen-phaseshift of the scattering waves (top panel) and the relative sizes of these intervals with respect to the mean (bottom panel). The relative error associated with 68% credible interval is indeed relatively small, i.e., around 5% across all energies.
Figure C.10. (Top) 68 and 95% credible intervals for the largest eigenphase of the three-body scattering wave. (Bottom) relative error of the same.
References
- [1]L. Fortunato, J. Casal, W. Horiuchi, et al., The ²⁹F nucleus as a lighthouse on the coast of the island of inversion, Commun. Phys. 3 (2020) 132. doi:10.1038/s42005-020-00402-5. URL https://www.nature.com/articles/s42005-020-00402-5DOI
- [2]J. Singh, J. Casal, W. Horiuchi, et al., Prediction of two-neutron halos in the N=28 isotones ⁴⁰Mg and ³⁹Na, Phys. Lett. B 853 (2024) 138694. doi:10.1016/j.physletb.2024.138694. URL https://www.sciencedirect.com/science/article/pii/S0370269324002521DOI
- [3]A. Poves, Shell model spectroscopy far from stability, J. Phys. G: Nucl. Part. Phys. 44 (8) (2017). doi:10.1088/1361-6471/aa7789. URL https://iopscience.iop.org/article/10.1088/1361-6471/aa7789DOI
- [4]D. Bazin, K. Becker, F. Bonaiti, et al., Perspectives on Few-Body Clusters in Exotic Nuclei, Few-Body Syst. 64 (2023) 25. doi:10.1007/s00601-023-01794-0. URL https://doi.org/10.1007/s00601-023-01794-0
- [5]M. Freer, H. Horiuchi, Y. Kanada-En’yo, et al., Microscopic clustering in light nuclei, Rev. Mod. Phys. 90 (2018) 035004. doi:10.1103/RevModPhys.90.035004. URL https://link.aps.org/doi/10.1103/RevModPhys.90.035004DOI
- [6]I. Tanihata, H. Savajols, R. Kanungo, Recent experimental progress in nuclear halo structure studies, Prog. Part. Nucl. Phys. 68 (2013) 215. doi:10.1016/j.ppnp.2012.07.001. URL https://www.sciencedirect.com/science/article/pii/S0146641012001081DOI
- [7]J. Casal, J. Gómez-Camacho, Identifying structures in the continuum: Application to ¹⁶Be, Phys. Rev. C 99 (2019) 014604. doi:10.1103/PhysRevC.99.014604. URL https://link.aps.org/doi/10.1103/PhysRevC.99.014604DOI
- [8]A. E. Lovell, F. M. Nunes, I. J. Thompson, Three-body model for the two-neutron emission of ¹⁶Be, Phys. Rev. C 95 (2017) 034605. doi:10.1103/PhysRevC.95.034605. URL https://link.aps.org/doi/10.1103/PhysRevC.95.034605DOI
- [9]D. B. Costa, M. Hongo, D. T. Son, Effective field theory for weakly bound two-neutron halo nuclei: Corrections from neutron-neutron effective range, Phys. Rev. C 112 (2025) 014001. doi:10.1103/1ds3-g3tp. URL https://link.aps.org/doi/10.1103/1ds3-g3tpDOI
- [10]B. Monteagudo, F. M. Marqués, J. Gibelin, et al., Mass, Spectroscopy, and Two-Neutron Decay of ¹⁶Be, Phys. Rev. Lett. 132 (2024) 082501. doi:10.1103/PhysRevLett.132.082501. URL https://link.aps.org/doi/10.1103/PhysRevLett.132.082501DOI
- [11]T. Aumann, T. Nakamura, The electric dipole response of exotic nuclei, Phys. Scr. T152 (2013) 014012. doi:10.1088/0031-8949/2013/T152/014012. URL https://iopscience.iop.org/article/10.1088/0031-8949/2013/T152/014012DOI
- [12]T. Aumann, D. Aleksandrov, L. Axelsson, et al., Continuum excitations in ⁶He, Phys. Rev. C 59 (1999) 1252. doi:10.1103/PhysRevC.59.1252. URL https://link.aps.org/doi/10.1103/PhysRevC.59.1252DOI
- [13]J. Wang, A. Galonsky, J. J. Kruse, et al., Dissociation of ⁶He, Phys. Rev. C 65 (2002) 034306. doi:10.1103/PhysRevC.65.034306. URL https://link.aps.org/doi/10.1103/PhysRevC.65.034306DOI
- [14]T. Nakamura, A. M. Vinodkumar, T. Sugimoto, et al., Observation of strong low-lying E1 strength in the two-neutron halo nucleus ¹¹Li, Phys. Rev. Lett. 96 (2006) 252502. doi:10.1103/PhysRevLett.96.252502. URL https://link.aps.org/doi/10.1103/PhysRevLett.96.252502DOI
- [15]M. Labiche, Halo Structure of ¹⁴Be, Phys. Rev. Lett. 86 (2001) 600. doi:10.1103/PhysRevLett.86.600. URL https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.86.600
- [16]K. J. Cook, T. Nakamura, Y. Kondo, other, Halo structure of the neutron-dripline nucleus ¹⁹B, Phys. Rev. Lett. 124 (2020) 212503. doi:10.1103/PhysRevLett.124.212503. URL https://link.aps.org/doi/10.1103/PhysRevLett.124.212503DOI
- [17]N. Kobayashi, T. Nakamura, J. A. Tostevin, et al., One- and two-neutron removal reactions from the most neutron-rich carbon isotopes, Phys. Rev. C 86 (2012) 054604. doi:10.1103/PhysRevC.86.054604. URL https://link.aps.org/doi/10.1103/PhysRevC.86.054604DOI
- [18]T. Nagahisa, W. Horiuchi, Examination of the ²²C radius determination with interaction cross sections, Phys. Rev. C 97 (2018) 054614. doi:10.1103/PhysRevC.97.054614. URL https://link.aps.org/doi/10.1103/PhysRevC.97.054614DOI
- [19]S. Bagchi, R. Kanungo, Y. K. Tanaka, et al., Two-neutron halo is unveiled in ²⁹F, Phys. Rev. Lett. 124 (2020) 222504. doi:10.1103/PhysRevLett.124.222504. URL https://link.aps.org/doi/10.1103/PhysRevLett.124.222504DOI
- [20]P. Navrátil, R. Roth, S. Quaglioni, Ab initio many-body calculation of the ⁷Be(p,γ)⁸B radiative capture, Phys. Lett. B 704 (2011) 379. doi:10.1016/j.physletb.2011.09.079. URL https://www.sciencedirect.com/science/article/pii/S0370269311011646DOI
- [21]A. Calci, P. Navrátil, R. Roth, et al., Can ab initio theory explain the phenomenon of parity inversion in ¹¹Be?, Phys. Rev. Lett. 117 (2016) 242501. doi:10.1103/PhysRevLett.117.242501. URL https://link.aps.org/doi/10.1103/PhysRevLett.117.242501DOI
- [22]P. Navrátil, S. Quaglioni, G. Hupin, et al., Unified ab initio approaches to nuclear structure and reactions, Phys. Scr. 91 (2016) 053002. doi:10.1088/0031-8949/91/5/053002. URL https://iopscience.iop.org/article/10.1088/0031-8949/91/5/053002DOI
- [23]S. Quaglioni, C. Romero-Redondo, P. Navrátil, et al., Three-cluster dynamics within the ab initio no-core model with continuum: How many-body correlations and α-clustering shape ⁶He, Phys. Rev. C 97 (2018). doi:10.1103/PhysRevC.97.034332. URL https://link.aps.org/doi/10.1103/PhysRevC.97.034332DOI
- [24]C. Romero-Redondo, S. Quaglioni, P. Navrátil, et al., How many-body correlations and α clustering shape ⁶He, Phys. Rev. Lett. 117 (2016) 222501. doi:10.1103/PhysRevLett.117.222501. URL https://link.aps.org/doi/10.1103/PhysRevLett.117.222501
- [25]K. Kravvaris, P. Navrátil, S. Quaglioni, et al., Ab initio informed evaluation of the radiative capture of protons on ⁷Be, Phys. Lett. B (2023) 138156 doi:10.1016/j.physletb.2023.138156. URL https://www.sciencedirect.com/science/article/pii/S0370269323004902?via%3DihubDOI
- [26]P. Navrátil, S. Quaglioni, G. Hupin, et al., Halo Nuclei from Ab Initio Nuclear Theory, arXiv:2604.02612 [nucl-th] (2026). doi:10.48550/arXiv.2604.02612. URL http://arxiv.org/abs/2604.02612
- [27]S. Elhatisari, Ab initio calculations of the isotopic dependence of nuclear clustering, Phys. Rev. Lett. 119 (2017). doi:10.1103/PhysRevLett.119.222505. URL https://link.aps.org/doi/10.1103/PhysRevLett.119.222505
- [28]Y.-H. Song, M. Kim, Y. Kim, et al., Ab initio calculations of the carbon and oxygen isotopes: Energies, correlations, and superfluid pairing, Phys. Lett. B 872 (2026) 140086. doi:10.1016/j.physletb.2025.140086. URL https://www.sciencedirect.com/science/article/pii/S0370269325008445DOI
- [29]S. Shen, S. Elhatisari, D. Lee, et al., Ab Initio Study of the Beryllium Isotopes ⁷Be to ¹²Be, Phys. Rev. Lett. 134 (2025) 162503. doi:10.1103/PhysRevLett.134.162503. URL https://link.aps.org/doi/10.1103/PhysRevLett.134.162503DOI
- [30]F. Bonaiti, S. Bacca, G. Hagen, Ab initio coupled-cluster calculations of ground and dipole excited states in ⁸He, Phys. Rev. C 105 (3) (2022) 034313. doi:10.1103/PhysRevC.105.034313. URL https://journals.aps.org/prc/abstract/10.1103/PhysRevC.105.034313DOI
- [31]G. Hagen, T. Papenbrock, M. Hjorth-Jensen, Ab Initio Computation of the 17F Proton Halo State and Resonances in A = 17 Nuclei, Phys. Rev. Lett. 104 (18) (2010) 182501. doi:10.1103/PhysRevLett.104.182501. URL https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.104.182501DOI
- [32]P. Hagen, H. W. Hammer, L. Platter, Charge form factors of two-neutron halo nuclei in halo EFT, Eur. Phys. J. A 49 (9) (2013) 118. doi:10.1140/epja/i2013-13118-4. URL https://link.springer.com/article/10.1140/epja/i2013-13118-4DOI
- [33]S. Shen, S. Elhatisari, D. Lee, U.-G. Meißner, Z. Ren, Ab Initio Study on the Halo Structure in ¹¹Be, Particles 9 (1) (2026) 25. doi:10.3390/particles9010025. URL https://www.mdpi.com/2571-712X/9/1/25DOI
- [34]H. H. Li, Unveiling potential neutron halos in intermediate-mass nuclei: An ab initio study, Phys. Rev. C 109 (6) (2024). doi:10.1103/PhysRevC.109.L061304. URL https://journals.aps.org/prc/abstract/10.1103/PhysRevC.109.L061304DOI
- [35]D. M. Rodkin, Yu. M. Tchuvil’sky, Detailed theoretical study of the decay properties of states in the ⁷He nucleus within an ab initio approach, Phys. Rev. C 104 (4) (2021) 044323. doi:10.1103/PhysRevC.104.044323. URL https://journals.aps.org/prc/abstract/10.1103/PhysRevC.104.044323DOI
- [36]R. J. Charity, G. H. Sargsyan, K. D. Launey, T. Webb, K. Brown, L. G. Sobotka, Comparing invariant-mass spectroscopy of ⁸B with ab initio predictions, Phys. Rev. C 113 (2026) 024322. doi:10.1103/PhysRevC.113.024322. URL https://link.aps.org/doi/10.1103/PhysRevC.113.024322DOI
- [37]G. Hagen, P. Hagen, H.-W. Hammer, L. Platter, Efimov physics around the neutron-rich ⁶⁰Ca isotope, Phys. Rev. Lett. 111 (2013) 132501. doi:10.1103/PhysRevLett.111.132501. URL https://link.aps.org/doi/10.1103/PhysRevLett.111.132501DOI
- [38]M. Huang, T. Frederico, P. Yin, R. A. M. Basili, P. J. Fasano, J. P. Vary, Halo structure of ⁶He from ab initio two-nucleon spatial correlations, arXiv:2512.24123 (2025). URL https://arxiv.org/abs/2512.24123
- [39]Anonymous, Halo structure of 6 He from ab initio two-nucleon spatial correlations, Phys. Rev. C (May 2026). doi:10.1103/hy1r-4vzg.
- [40]G. Hagen, P. Hagen, H.-W. Hammer, L. Platter, Efimov Physics Around the Neutron-Rich $^{40}\mathrm{Ca}$ Isotope, Phys. Rev. Lett. 111 (13) (2013) 132501. doi:10.1103/PhysRevLett.111.132501.
- [41]E. Hiyama, R. Lazauskas, J. Carbonell, T. Frederico, Scaling of the B 19 two-neutron halo properties close to unitarity, Phys. Rev. C 106 (6) (2022) 064001. doi:10.1103/PhysRevC.106.064001.DOI
- [42]E. Hiyama, R. Lazauskas, F. M. Marqués, J. Carbonell, Modeling B 19 as a B 17 - n - n three-body system in the unitary limit, Phys. Rev. C 100 (1) (2019) 011603. doi:10.1103/PhysRevC.100.011603.DOI
- [43]F. M. Nunes, J. A. Christley, I. J. Thompson, et al., Core excitation in three-body systems: Application to ¹²Be, Nucl. Phys. A 609 (1996) 43. doi:10.1016/0375-9474(96)00288-9. URL https://www.sciencedirect.com/science/article/pii/0375947496002849DOI
- [44]J. A. Tostevin, F. M. Nunes, I. J. Thompson, Calculations of three-body observables in ⁸B breakup, Phys. Rev. C 63 (2001) 024617. doi:10.1103/PhysRevC.63.024617. URL https://link.aps.org/doi/10.1103/PhysRevC.63.024617DOI
- [45]I. J. Thompson, F. M. Nunes, B. V. Danilin, FaCE: a tool for three body Faddeev calculations with core excitation, Comp. Phys. Commun. 161 (2004) 87. doi:10.1016/j.cpc.2004.03.007. URL https://www.sciencedirect.com/science/article/pii/S0010465504002140DOI
- [46]E. C. Pinilla, P. Descouvemont, Coulomb breakup of ²²C in a four-body model, Phys. Rev. C 94 (2016) 024620. doi:10.1103/PhysRevC.94.024620. URL https://link.aps.org/doi/10.1103/PhysRevC.94.024620DOI
- [47]E. C. Pinilla, W. Leideman, G. Orlandini, other, Three-body model of ⁶He with nonlocal halo effective field theory potentials, Phys. Rev. C 112 (2025) 024003. doi:10.1103/kskt-7p8g. URL https://link.aps.org/doi/10.1103/kskt-7p8g
- [48]P. Descouvemont, C. Daniel, D. Baye, Three-body systems with Lagrange-mesh techniques in hyperspherical coordinates, Phys. Rev. C 67 (2003) 044309. doi:10.1103/PhysRevC.67.044309. URL https://link.aps.org/doi/10.1103/PhysRevC.67.044309DOI
- [49]P. Descouvemont, E. Tursunov, D. Baye, Three-body continuum states on a Lagrange mesh, Nucl. Phys. A 765 (2006) 370. doi:10.1016/j.nuclphysa.2005.11.010. URL https://www.sciencedirect.com/science/article/pii/S037594740501198XDOI
- [50]J. Casal, E. Garrido, Three-body structure of ¹⁹B: Finite-range effects in two-neutron halo nuclei, Phys. Rev. C 102 (2020) 051304. doi:10.1103/PhysRevC.102.051304. URL https://link.aps.org/doi/10.1103/PhysRevC.102.051304DOI
- [51]S. N. Ershov, Binding energy constraint on matter radius and soft dipole excitations of ²²C, Phys. Rev. C 86 (2012). doi:10.1103/PhysRevC.86.034331. URL https://link.aps.org/doi/10.1103/PhysRevC.86.034331DOI
- [52]W. Horiuchi, Y. Suzuki, ²²C : An s-wave two-neutron halo nucleus, Phys. Rev. C 74 (2006) 034311. doi:10.1103/PhysRevC.74.034311. URL https://link.aps.org/doi/10.1103/PhysRevC.74.034311DOI
- [53]C. A. Bertulani, H. W. Hammer, U. van Kolck, Effective field theory for halo nuclei: Shallow p-wave states, Nucl. Phys. A 712 (1) (2002) 37–58. doi:10.1016/S0375-9474(02)01270-8. URL https://www.sciencedirect.com/science/article/pii/S0375947402012708?via%3DihubDOI
- [54]H.-W. Hammer, D. R. Phillips, C. Ji, Effective field theory description of halo nuclei, J. Phys. G: Nucl. Part. Phys. 44 (2017) 103002. doi:10.1088/1361-6471/aa83db. URL https://iopscience.iop.org/article/10.1088/1361-6471/aa83dbDOI
- [55]H.-W. Hammer, S. König, U. van Kolck, Nuclear effective field theory: Status and perspectives, Rev. Mod. Phys. 92 (2) (2020) 025004. doi:10.1103/RevModPhys.92.025004. URL https://journals.aps.org/rmp/abstract/10.1103/RevModPhys.92.025004DOI
- [56][56*] P. Capel, Combining Halo-EFT Descriptions of Nuclei and Precise Models of Nuclear Reactions, Few-Body Systems 63 (1) (2021) 14. doi:10.1007/s00601-021-01718-w. URL https://doi.org/10.1007/s00601-021-01718-w
- [57]J. Braun, H.-W. Hammer, L. Platter, Halo structure of 17C, Eur. Phys. J. A 54 (2018). doi:10.1140/epja/i2018-12630-3. URL http://dx.doi.org/10.1140/epja/i2018-12630-3
- [58]M. Göbel, H.-W. Hammer, D. R. Phillips, Universality of n distributions of s-wave 2h halo nuclei and the unitary limit, Phys. Rev. C 110 (2024) 024003. doi:10.1103/PhysRevC.110.024003. URL https://link.aps.org/doi/10.1103/PhysRevC.110.024003DOI
- [59]M. Hongo, D. T. Son, Universal properties of weakly bound two-neutron halo nuclei, Phys. Rev. Lett. 128 (2022) 212501. doi:10.1103/PhysRevLett.128.212501. URL https://link.aps.org/doi/10.1103/PhysRevLett.128.212501DOI
- [60]B. Acharya, C. Ji, D. R. Phillips, Implications of a matter-radius measurement for the structure of Carbon-22, Phys. Lett. B 723 (2013) 196. doi:10.1016/j.physletb.2013.04.055. URL http://arxiv.org/abs/1303.6720
- [61]L. Gaudefroy, W. Mittig, N. A. Orr, et al., Direct mass measurements of 19B, 22C, 29F, 31Ne, 34Na and other light exotic nuclei, Phys. Rev. Lett. 109 (2012) 202503. doi:10.1103/PhysRevLett.109.202503. URL https://link.aps.org/doi/10.1103/PhysRevLett.109.202503DOI
- [62]M. Wang, W. Huang, F. Kondev, et al., The AME 2020 atomic mass evaluation (II). Tables, graphs and references*, Chin. Phys. C 45 (2021) 030003. doi:10.1088/1674-1137/abddaf. URL https://doi.org/10.1088/1674-1137/abddaf
- [63]K. Tanaka, T. Yamaguchi, T. Suzuki, et al., Observation of a large reaction cross section in the drip-line nucleus 22C, Phys. Rev. Lett. 104 (2010) 062701. doi:10.1103/PhysRevLett.104.062701. URL https://link.aps.org/doi/10.1103/PhysRevLett.104.062701DOI
- [64]Y. Togano, T. Nakamura, Y. Kondo, et al., Interaction cross section study of the two-neutron halo nucleus 22C, Phys. Lett. B 761 (2016) 412. doi:10.1016/j.physletb.2016.08.062. URL https://www.sciencedirect.com/science/article/pii/S0370269316304890DOI
- [66]S. Leblond, Structure des isotopes de bore et de carbone riches en neutrons aux limites de la stabilité, Theses, Normandie Université, France (2015). URL https://theses.hal.science/tel-01289381
- [67]P. McGlynn, C. Hebborn, Development of an accurate formalism to predict properties of two-neutron halo nuclei: case study of 22C, arXiv:2602.15765 [nucl-th] (2026). URL http://arxiv.org/abs/2602.15765
- [68]D. R. Thompson, M. Lemere, Y. C. Tang, Systematic investigation of scattering problems with the resonating-group method, Nucl. Phys. A 286 (1977) 53. doi:10.1016/0375-9474(77)90007-0. URL https://www.sciencedirect.com/science/article/pii/0375947477900070DOI
- [69]K. Varga, Y. Suzuki, Precise solution of few-body problems with the stochastic variational method on a correlated Gaussian basis, Phys. Rev. C 52 (1995) 2885. doi:10.1103/PhysRevC.52.2885. URL https://link.aps.org/doi/10.1103/PhysRevC.52.2885DOI
- [70]S. K. Bogner, R. J. Furnstahl, H. Hergert, et al., Testing the density matrix expansion against ab initio calculations of trapped neutron drops, Phys. Rev. C 84 (2011) 044306. doi:10.1103/PhysRevC.84.044306. URL https://link.aps.org/doi/10.1103/PhysRevC.84.044306DOI
- [71]M. T. Yamashita, R. S. M. de Carvalho, T. Frederico, et al., Constraints on two-neutron separation energy in the Borromean 22C nucleus, Phys. Lett. B 697 (2011) 90. doi:10.1016/j.physletb.2011.01.040. URL https://www.sciencedirect.com/science/article/pii/S0370269311000773DOI
- [72]A. J. Smith, K. Godbey, C. Hebborn, et al., Matter radii from interaction cross sections using microscopic nuclear densities, arXiv:2603.18862 (2026). URL https://arxiv.org/abs/2603.18862
- [73]A. Ozawa, O. Bochkarev, L. Chulkov, et al., Measurements of interaction cross sections for light neutron-rich nuclei at relativistic energies and determination of effective matter radii, Nucl. Phys. A 691 (2001) 599. doi:10.1016/S0375-9474(01)00563-2. URL https://www.sciencedirect.com/science/article/pii/S0375947401005632DOI
- [74]D. Foreman-Mackey, D. W. Hogg, et al., emcee: The MCMC Hammer, Publ. Astron. Soc. Pac. 125 (2013). doi: 10.1086/670067. D. R. Phillips, R. J. Furnstahl, U. Heinz, et al., Get on the BAND wagon: a Bayesian framework for quantifying model uncertainties in nuclear dynamics, J. Phys. G: Nucl. Part. Phys. 48 (2021) 072001. doi:10.1088/1361-6471/abf1df. URL https://iopscience.iop.org/article/10.1088/1361-6471/abf1df
- [76]C. D. Pruitt, A. E. Lovell, C. Hebborn, et al., Role of the likelihood for elastic scattering uncertainty quantification, Phys. Rev. C 110 (2024) 064606. doi:10.1103/PhysRevC.110.064606. URL https://link.aps.org/doi/10.1103/PhysRevC.110.064606DOI








