First observation of ²⁸O

The study investigates whether the extremely neutron-rich oxygen isotopes ²⁷O and ²⁸O exist as resonances beyond the neutron drip line and examines their structure relative to shell closures, particularly the expected N = 20 closure that would render ²⁸O doubly magic in the standard shell model. Light, neutron-rich isotopes at extreme N/Z ratios are stringent tests for nuclear-structure theories, including modern ab initio approaches. Prior work established the neutron drip line up to neon and observed the most neutron-rich bound oxygen isotope at ²⁴O, which itself is doubly magic with a closed N = 16 shell, while the N = 20 closure is known to disappear in the Island of Inversion (Ne, Na, Mg) and extend to neighboring fluorine isotopes ²⁸,²⁹F. Only ²⁵O and ²⁶O had previously been observed among the heavier oxygen isotopes, with ²⁶O barely unbound. This work aims to directly observe ²⁷O and ²⁸O via their multineutron decays to ²⁴O, quantify their decay energies and widths, and test whether ²⁸O exhibits a closed N = 20 shell or instead belongs to the Island of Inversion with significant pf-shell neutron configurations.

• Neutron drip line established up to Z = 10 (neon) and the heaviest unbound nucleus observed for fluorine is ²⁸F; the tetra-neutron near-threshold structure has also been reported.
• Classical magic numbers (2, 8, 20, 28, 50, 82, 126) define spherical closed shells and doubly-magic nuclei; ²⁸O had long been anticipated as a potential doubly-magic nucleus but had not been observed.
• Doubly-magic character of neutron-rich ⁷⁸Ni (Z=28, N=50) and ¹³²Sn (Z=50, N=82) has been confirmed, while the magicity of ⁶He (Z=2, N=8) remains unresolved.
• The N = 20 shell closure disappears in neutron-rich Ne, Na, Mg isotopes (Island of Inversion), where pf-shell intruder configurations dominate ground states; the Island of Inversion has been shown to extend to ²⁸,²⁹F.
• ²⁴O is doubly magic with a new shell closure at N = 16. Among heavier O isotopes, ²⁵O (1n-unbound) and ²⁶O (2n-unbound, E ≈ 18(5) keV) had been observed; previous searches for ²⁷,²⁸O did not confirm their existence as resonances.
• Theoretical frameworks include large-scale shell-model interactions (USDB, SDPF-M, EEdf family), continuum shell models (CSM, GSM), and ab initio methods (VS-IMSRG, SCGF, coupled cluster). Three-nucleon forces are crucial to describe neutron-rich oxygen isotopes and the Z = 8 drip line at ²⁴O.

• Beam and production: Neutron-unbound ²⁷O and ²⁸O were produced via proton-induced nucleon knockout from a 235 MeV/u ²⁹F beam at the RIKEN RI Beam Factory.
• Target and tracking: ²⁹F ions were characterized and tracked onto a 151 mm thick liquid-hydrogen target surrounded by the MINOS time projection chamber to determine reaction vertices, enabling high luminosity while preserving decay-energy resolution.
• Detection systems: Forward-focused charged fragments and neutrons were measured with the SAMURAI spectrometer. Neutron detection employed large-area segmented plastic scintillator arrays (NeuLAND and NEBULA). Overall detection efficiencies at 0.5 MeV decay energy were ~2% for 3n and ~0.4% for 4n decays.
• Kinematic reconstruction: Decay energies were reconstructed using the invariant-mass method with FWHM resolution ~0.2 MeV at 0.5 MeV decay energy.
• Event identification: ²⁴O fragments were identified via magnetic rigidity, energy loss, and time-of-flight. Neutrons were identified using time-of-flight and deposited energy. Dedicated offline crosstalk-rejection procedures were applied to mitigate neutrons scattering between scintillators.
• Analysis strategy: Neutrons were ordered by increasing two-body relative energy to ²⁴O (n₁, n₂, ...). Five-body (²⁴O+4n) decay energy spectra E₀₁₂₃₄ and four-body (²⁴O+3n) spectra E₀₁₂₃ were analyzed. Residual crosstalk contributions were quantified and included via full Monte Carlo simulations to consistently describe ²⁴O+xn spectra. Partial decay energies (e.g., E₀₁₂ for ²⁴O+n₁+n₂) were used to identify sequential decay pathways via ²⁶O. Gated fits with E₀₁₂ < 0.08 MeV selected decays through the ²⁶O ground state to reduce uncertainties from higher-lying ²⁵O resonances. Line shapes incorporated experimental response and residual crosstalk from simulations.
• Cross-section and spectroscopic factor: The single-proton removal cross-section from ²⁹F populating ²⁸O was measured and analyzed with a distorted-wave impulse approximation (DWIA) to deduce the spectroscopic factor C²S. Momentum distributions were compared with expectations for removal from the 1d5/2 orbital, supporting a 5/2⁺ assignment in the parent nucleus.
• Theory comparisons: Experimental energies were compared to large-scale shell-model calculations with the χEFT-based EEdf3 interaction (including pf-shell), continuum shell models (CSM, GSM), SDPF-M and USDB interactions, and ab initio approaches (VS-IMSRG, SCGF, coupled cluster). Coupled-cluster calculations were combined with a Bayesian history matching statistical framework to explore 17 χEFT low-energy constants, generating posterior predictive distributions for ²⁷,²⁸O energies.

• First direct observation of ²⁸O and ²⁷O as narrow, low-lying resonances decaying to ²⁴O + 4n and ²⁴O + 3n, respectively.
• ²⁸O decay energy: E₀₁₂₃₄ = 0.46 ± 0.04(stat) ± 0.02(syst) MeV; resonance width upper limit ≤ 0.7 MeV (68% CI).
• ²⁷O decay energy: E ≈ 1.09 ± 0.04(stat) ± 0.02(syst) MeV (consistent with the peak at E₀₁₂₃ ≈ 1 MeV and the measured ΔE(²⁷,²⁸O) = 0.63 ± 0.06(stat) ± 0.03(syst) MeV). Width upper limit ≤ 0.18 MeV (68% CI). Tentative J^π assignment 3/2⁺ or 7/2⁺ based on width constraints.
• Decay pathways: Both ²⁸O and ²⁷O decay sequentially via ²⁶O g.s.; the three-body partial decay-energy spectrum exhibits a sharp near-threshold peak consistent with the known ²⁶O decay energy of 18(5) keV.
• Production cross-section: Single-proton removal from ²⁹F to ²⁸O resonance: σ = 1.36 ± 0(stat) ± 0.13(syst) mb (systematic dominated by neutron-detection efficiency).
• Spectroscopic factor: From DWIA analysis, C²S = 0.48 ± 0.05(stat) ± 0.05(syst). This sizable value indicates that the neutron configuration of ²⁸O resembles that of ²⁹F and is inconsistent with a rigid N = 20 closed shell. EEdf3 predicts C²S ≈ 0.68 (after c.m. correction), with a known reduction factor typical of knockout and (e,e′p) reactions; if neutrons were confined to the sd shell, C²S would be ~0.13, contrary to experiment.
• Shell-model structure: EEdf3 predicts significant pf-shell occupancy, with summed neutron pf occupation numbers of 2.5 (²⁸O) and 1.4 (²⁷O) and 1d occupancy of 2.0 (²⁸O) and 2.1 (²⁷O), consistent with collapse of the N = 20 shell closure.
• Theory comparison: Most models (USDB, SDPF-M, VS-IMSRG, SCGF, Λ-CCSD(T)) predict higher ²⁷O, ²⁸O energies (underbinding compared to data); EEdf3 also overpredicts by ~1 MeV. GSM with continuum couplings shows better agreement.
• Coupled-cluster with history matching: Predicted correlated ppd yields median and 68% credible intervals for ΔE(²⁷,²⁸O) = 0.11^{+0.39}{-0.05} MeV and ΔE(²⁸,²⁴O) = 2.11^{+0.03}{-0.12} MeV; experimental values ΔE(²⁷,²⁸O) = 0.63 ± 0.06 ± 0.03 MeV and ΔE(²⁸,²⁴O) = 0.46^{+0.05}_{-0.04} ± 0.02 MeV lie at the edge of the 68% credible regions, indicating the new data strongly constrain χEFT interactions.
• Implication for magicity: Cross-section and spectroscopic factor indicate disappearance of the N = 20 shell closure in ²⁸O; the Island of Inversion extends into the oxygen isotopes; ²⁸O is not doubly magic.

The central question was whether ²⁸O exhibits a robust N = 20 shell closure (doubly magic character) or belongs to the Island of Inversion with substantial intruder pf-shell configurations. Direct detection of ²⁷O and ²⁸O multineutron decays established their existence as low-lying resonances that decay sequentially via ²⁶O, with accurately reconstructed decay energies. The measured single-proton removal cross-section from ²⁹F and the deduced spectroscopic factor C²S ≈ 0.48, together with shell-model predictions of significant pf-shell occupancy, demonstrate that ²⁸O does not possess a closed N = 20 shell. Instead, its neutron structure closely resembles that of ²⁹F, placing ²⁸O within the Island of Inversion. Comparisons with a wide range of theoretical models show that most underbind ²⁷O and ²⁸O, highlighting deficiencies in present interactions or many-body treatments near the extreme neutron-rich edge. The coupled-cluster calculations with Bayesian history matching show the experimental energies sit near the boundary of the 68% credible regions, implying only finely tuned χEFT interactions can reproduce both energies simultaneously; thus, the new data provide valuable constraints for refining ab initio nuclear forces. Overall, the findings resolve the long-standing question of ²⁸O’s magicity, demonstrate the feasibility of multineutron-decay spectroscopy, and set benchmarks for theory at the neutron drip line.

• This work reports the first observation of the neutron-unbound isotopes ²⁷O and ²⁸O as low-lying resonances decaying to ²⁴O + 3n and ²⁴O + 4n, respectively, enabled by a high-luminosity LH₂ target with vertex tracking and high-efficiency multineutron detection.
• Measured decay energies (²⁸O: 0.46 MeV; ²⁷O: ~1.09 MeV) and narrow widths, plus sequential decay via ²⁶O, provide precise experimental anchors for theory.
• The measured cross-section and deduced C²S demonstrate that ²⁸O does not exhibit a closed N = 20 neutron shell; instead, pf-shell intruder configurations dominate, extending the Island of Inversion into the oxygen isotopic chain.
• Comparisons to shell-model (including continuum) and ab initio approaches show general underbinding of ²⁷,²⁸O; coupled-cluster with statistical exploration of χEFT parameter space indicates the new data strongly constrain interactions.
• Future work: determine the excitation energy of the first 2⁺ state in ²⁸O (predicted ~2.1–2.5 MeV), quantify pf-shell admixtures in detail, and employ complementary probes of the sd–pf gap (e.g., parity splittings in ²⁴O as seen in ²⁸F). The demonstrated multineutron-decay spectroscopy opens avenues to study other nuclei beyond the neutron drip line and multineutron correlations.

• Neutron detection efficiencies are low for multineutron events (~2% for 3n and ~0.4% for 4n at 0.5 MeV decay energy), limiting statistics.
• Residual neutron crosstalk, though mitigated by dedicated rejection algorithms and modeled in simulations, contributes broad background and systematic uncertainties.
• Energy resolution (FWHM ~0.2 MeV at 0.5 MeV decay energy) limits precise width extraction; reported widths are upper limits.
• Limited efficiency precluded clear identification of higher-lying ²⁵O resonances, necessitating gating (E₀₁₂ < 0.08 MeV) to isolate sequential decays via ²⁶O.
• Theoretical comparisons carry sizable uncertainties, particularly in ab initio χEFT predictions; credible regions for ²⁷,²⁸O energies relative to ²⁴O are broad, reflecting sensitivity to low-energy constants.