New short-lived isotope 223 Np and the absence of the Z = 92 subshell closure near N = 126

New short-lived isotope 223Np and the absence of the Z = 92 subshell closure near N = 126 M.D. Sun a,b,c, Z. Liu a,∗, T.H. Huang a, W.Q. Zhang a, J.G. Wang a, X.Y. Liu a,b, B. Ding a, Z.G. Gan a, L. Ma a, H.B. Yang a, Z.Y. Zhang a, L. Yu a, J. Jiang a,b, K.L. Wang a,b, Y.S. Wang a, M.L. Liu a, Z.H. Li d, J. Li d, X. Wang d, H.Y. Lu a,b, C.J. Lin e, L.J. Sun e, N.R. Ma e, C.X. Yuan f, W. Zuo a, H.S. Xu a, X.H. Zhou a, G.Q. Xiao a, C. Qi g, F.S. Zhang h,i


Introduction
The evolution of proton shell structure beyond 208 Pb is of decisive importance for the shell stabilization of superheavy elements. The existence of a subshell or even shell gap at Z = 92 between the proton h 9/2 and f 7/2 orbitals has been a topic of intense theoretical debate. A substantial Z = 92 shell gap is predicted in many relativistic mean-field calculations like an early work for heavy elements [1], in most of the covariant density functionals (CDFs) [2,3] and also in some non-relativistic models [4].
Macroscopic-microscopic calculations [5] predicted a subshell gap at Z = 92. This is at variance with large-scale shell-model calculations [6], which show no sign of a shell gap at Z = 92 for N = 126 isotones and are in overall agreement with spectroscopic data on these isotones up to U [6][7][8][9][10].
The spurious shell closures of Z = 92 and others can be cured in the upgraded CDF model [11] by including ρ-tensor Fock terms, restoring the pseudo-spin symmetry which qualitatively represents the balance of nuclear forces [11][12][13]. However, the very recent experimentally observed sudden decrease of the reduced α-decay width at Z = 92 along the N = 130 isotonic chain (see Fig. 5a in [14]) cannot exclude the possibility of a subshell closure at Z = 92. The proton separation energy, ground-state spin and parity of odd-Z isotopes beyond U, e.g. Np isotopes ( Z = 93), could help clarify the absence/presence of the Z = 92 subshell closure.
Such experimental verification is necessary and valuable for testing the nuclear structure models and understanding the nature of nuclear forces as well [11][12][13].
For isotopes of elements above lead and far off the β-stability line, α decay prevails as the major radioactive decay mode and α spectroscopy is an indispensable tool to investigate the low-energy structure of heavy neutron-deficient nuclei. In the classical picture, α decay occurs through the preformation of an α particle in the nucleus and its subsequent tunneling through Coulomb and centrifugal barriers [15,16]. Above shell closure, the preformation probability of α particle and the decay energy Q α increase simultaneously, therefore the most enhanced α decays take place above doubly magic nucleus such as 208 Pb and 100 Sn [17,18].
All over the chart of nuclides, the region to the "north-east" of 208 Pb, with Z ≥ 84 and N = 128-130, hosts the shortest-lived α radioactivities, with half-lives in the range of nanoseconds to microseconds. So far the shortest-lived α emitter known with directly measured half-life is 219 Pa (T 1/2 = 53 ns) [19] with N = 128. Synthesis and detection of neutron-deficient isotopes above thorium in this region are challenging due to their low production cross sections and short half-lives. With increasing atomic number Z , the fission probability of the compound nucleus increases rapidly and the evaporation of protons and α particles is by far dominant over neutron evaporation [20]. Progress in this region has been very slow in the last three decades, the frontier in this region was only pushed forwards from Pa to U [14]. For the N = 130 isotones experimental data are available up to 222 U. Recently the semi-magic 219 Np (N = 126) was reported in [21] as the daughter of 223 Am, but the assignments of these two isotopes are in doubt because the half-life of 223 Am is expected much shorter while that of 219 Np much longer than claimed in [21]. The most neutron-deficient neptunium isotope 225 Np was discovered over 20 years ago [22]. In the present letter, we report on the first observation of the short-lived N = 130 neptunium isotope 223 Np. As its daughter nucleus 219 Pa is extremely short-lived, the α-decay signals of 223 Np and 219 Pa will pile up in the implantation detector, and even with the 223 Np implant signal if the half-life of 223 Np is in the range of microseconds. With conventional analog electronics, the shortest half-lives accessible are around tens of μs, it is extremely difficult or impossible to resolve these pileup signals in either energy or time. In recent years, digital pulse processing has been successfully applied to resolve such pileup events in the charged particle spectroscopy of short-lived nuclei [14,18,23].

Experiment and results
The isotope 223 Np was produced in the 187 Re( 40 Ar, 4n) reaction channel with isotopically enriched (98.6%) 187 Re targets. The 40 Ar beam was accelerated to 188 MeV by the Sector-Focusing Cyclotron (SFC) of the Heavy Ion Research Facility in Lanzhou (HIRFL). The beam intensity on target, monitored via Faraday Cups upstream and downstream of the target chamber, was around 320 pnA in average during the entire experiment of 110 hours, with an uncertainty of up to a factor of 2. The targets were 460 μg/cm 2 thick, sputtered on carbon foils of thickness 80 μg/cm 2 , with carbon foils facing the beam. After the experiment, target thicknesses were measured to be basically the same as before the experiment within a precision of ∼15%. In the center of the target, the excitation energy of the compound nucleus 227 Np is estimated to be 44 MeV, close to that expected for the maximum cross-section for 223 Np by the HIVAP code [24], with other main reaction channels being 223 U(p3n evaporation channel), 220 Pa(α3n evaporation channel) and 220 Th(αp2n evaporation channel).
Evaporation residues were separated from the primary beam by the recoil separator SHANS [25] filled with ∼0.6 mbar helium gas and implanted into a 300-μm double-sided silicon strip detector (DSSSD). The average charge state, q, of the evaporation residues was simulated to be q = 6.9 [26]. For optimum transmission, the magnetic rigidity of SHANS was set to Bρ = 1.785 Tm. The DSSSD had 48 horizontal and 128 vertical strips of 1 mm width, forming a total of 6144 pixels. A multiwire proportional chamber (MWPC) was mounted in front of the DSSSD detector and was used to distinguish between implantation and decay events. To minimize the interference from scattered light ions in the DSSSD, three Si detectors of 50 mm × 50 mm size and 300 μm thick were placed side by side behind the DSSSD detector and used as veto detectors. Typical MWPC and DSSSD implantation detector rates were less than 100/s during beam-on periods, indicating a very good primary beam and transfer background suppression performance of SHANS.
In this experiment, very short-lived nuclei with N = 128-130 were produced either as ERs or as decay products of ERs. In order to resolve pileup signals, a data acquisition system based on fast digital pulse processing (DPP) was used. Signals from all the preamplifiers of the DSSSD strips, MWPC and veto detectors were digitized directly by using the 14-bit, 100-MS/s fADCs from CAEN S.p.A [27]. The digitizers allow for dead-time free acquisition, and all the channels are able to generate triggers independently. The timing is based on a so-called RC-CR 2 filter. In analogy with the constant-fraction discrimination, the RC-CR 2 signal is bipolar and its zero crossing corresponds to the trigger time-stamp. The preamplifier signals and RC-CR 2 signals were sampled simultaneously at the same frequency of 20 ns a sampling point and waveforms of 15 μs length were recorded for offline analysis. Energy calibrations were performed with 175 Lu, 186 W and 187 Re targets at the same beam energy, covering a range of 6-19 MeV, specifically 6.3-9.4 MeV for single α energy and up to 19 MeV for double α sum energy. For non-pileup traces of long-lived α radioactivities, a trapezoidal filter with rising time 5 μs and flat top 3 μs was used to extract the full pulse height [28]. The energy resolution (FWHM) obtained with all vertical (horizontal) strips summed up is 22 (30) keV at α-particle energy of ∼7000 keV.
For pileup events, depending on the time difference T between the overlapping signals, the energies of individual signals were extracted using different algorithms. For overlapping signals with T = 0.5-15 μs, a trapezoidal filter with rising time 200 ns and flat top 200 ns was used. For signals with T = 200-500 ns, the pulse-height of individual signal was obtained from the difference between the average of about six data points in the plateau area after the leading edge and that before the leading edge (average difference algorithm). The energy resolution of vertical strips for α decays recorded in double/multiple pulse traces with T down to ∼0.5 μs and ∼0.2 μs are around 55 keV and 70 keV, respectively. In the interval T = 100-200 ns, the average difference algorithm was applied but with smaller number of data points, the energy resolution obtained is around 140 keV. It is worth noting that this is the shortest time difference between two overlapping pulses which have been analyzed in α/p spectroscopy, thanks to the very fast rising times of the signals from the DSSSD preamplifiers which are typically 40-60 ns in the present work. For even shorter time difference, T < 100 ns, the boundary between the two α pulses is difficult/impossible to determine. In some of such cases, the individual α energy may be extracted using the pulse height of each α, but the results will be rather arbitrary and unreliable.
For α pileup signals with T < 200 ns, where the two α signals are difficult/impossible to be separated, the sum am-  Table 1. As examples, the traces corresponding to events 1, 4 and 6 are plotted in Fig. 2.
The α sum energies of 223 Np and 219 Pa in five (events [1][2][3][4][5] out of the ten decay chains are very close (within 50 keV) and much larger than the values in the rest, implying that only one α  (37) keV, in agreement with the previous value of 9900(50) keV [19] within the error bar. In [19] the ERs were stopped in a catcher foil behind the target and α particles emitted from the stopper were detected with an ionization chamber, the energy resolution of which was poor (FWHM ∼ 100 keV  Table 1 Decay chains attributed to the new isotope 223 Np. α i represents α particle from 223 Np, 219 Pa, 215 Ac and 211 Fr, for i = 1, 2, 3 and 4, respectively. The units are keV for the implantation energy of ER, α particle energies and standard deviations (σ ). The column (E α1 + E α2 ) lists the sum energy of two overlapping α signals. similar to the value reported in [25] where the same reaction at beam energy of 177 MeV was used. Taking

Discussion
The by γ transition. The internal conversion coefficient for such a transition is smaller than 0.6. Taking into account the detection efficiency of conversion electron within one pixel, the chance for the energy summing of α with conversion electron is negligible. So the measured charged-particle energy comes from α only.
If 223 Np has a 9/2 − ground state, as predicted by the shellmodel calculations presented below, the α decay to the 9/2 − g.s.
of 219 Pa is expected to be dominant, consistent with the fact that only one α line is observed. If 223 Np has a 7/2 − ( f 7/2 ) ground state, the 7/2 − → 9/2 − g.s. to g.s. transition is strongly hindered due to the spin flip between the initial and final states, it will decay to the 7/2 − excited state in 219 Pa with α energy 9477 keV, followed by γ transition.
Detailed information on nuclear structure can be obtained from the α-particle preformation probability inside the nucleus [32], which microscopically quantifies the stability against α decay. Conventionally, an equivalent variable, the reduced width for α decay δ 2 [33], which takes into account the angular momentum of the emitted α particle, is used. For the two possible α-decay paths above, where spins and parities of initial and final states are identical, the reduced decay width is calculated to be 0.17 ( 8 4 ) MeV using the Q α of 9687(45) keV and T 1/2 obtained for 223 Np in this work, comparable to those of neighboring N = 130 isotones with Z = 86-91 as shown in Fig. 3.
In order to understand the structure/spin-parity of 223   The spins and parities of the first few states obtained for the N = 126 isotones using such truncated model spaces are found to be the same with those obtained using the full model space in [6], giving us confidence in the present calculations in the truncated model space. The spins and parities of 223 Np and 219 Pa are calculated to be both 9/2 − . In the previous shell-model calculations [6], the ground state of the semi-magic 219 Np was predicted to be 9/2 − as well.
Based on the large-scale shell-model calculations above, spin and parity of 9/2 − are tentatively proposed for the ground state of 223 Np, supporting the absence of a subshell closure at Z = 92 and N = 130. However, an α-γ coincidence experiment with much higher statistics than in the present work is needed to exclude/confirm the alternative 7/2 − spin and parity.
It should be noted however that the reduced decay width of 222 U is anomalously small compared with those of other N = 130 isotones, even smaller than those of the neighboring odd-Z 221 Pa and 223 Np. As the present results for 223 Np and the previous data for N = 126 isotones do not support the existence of a Z = 92 subshell closure around N = 126, the reason remains unclear and this anomaly calls for further study.

Summary
In summary, we report on the discovery of the new short-lived isotope 223 Np, which was synthesized in the fusion reaction 40 Ar + 187 Re and identified through temporal and spatial correlations with subsequent α decays in the decay chain starting from 223 Np.
The half-life and energy were extracted to be T 1/2 = 2.15( 100 52 ) μs and E α = 9477(44) keV from pileup traces by using modern digital pulse processing techniques. The energy of individual α in pileup trace with time difference between overlapping signals down to ∼100 ns was extracted, the shortest analyzed so far using this method. The trend in proton separation energy shows no sign of a Z = 92 subshell closure. The spin and parity of 223 Np are proposed to be 9/2 − by combining the reduced α-decay width and largescale shell-model calculations in truncated model space, negating the presence of a h 9/2 subshell closure at Z = 92 near N = 126.
The decay chain of 219 Pa, the shortest-lived α emitter known with directly measured half-life, was established for the first time.