Constraining the rp-process by measuring 23 Al ( d , n ) 24 Si with GRETINA and LENDA at NSCL

The 23Al(p,γ)24Si stellar reaction rate has a significant impact on the lightcurve emitted in X-ray bursts. Theoretical calculations show that the reaction rate is mainly determined by the properties of direct capture as well as low-lying 2 states and a possible 4 state in 24Si. Currently, there is little experimental information on the properties of these states. In this proceeding we will present a new experimental study to investigate this reaction, using the surrogate reaction 23Al(d,n) at 47 AMeV at the National Superconducting Cyclotron Laboratory (NSCL). We will discuss our new experimental setup which allows us to use full kinematics employing the Gamma-Ray Energy Tracking In-beam Nuclear Array (GRETINA) to detect the γ-rays following the de-excitation of excited states of the reaction products and the Low Energy Neutron Detector Array (LENDA) to detect the recoiling neutrons. The S800 was used for identification of the 24Si recoils. As a proof of principle to show the feasibility of this concept the Q-value spectrum of 22Mg(d,n)23Al is reconstructed.


Introduction
X-ray bursts are produced when a neutron star accretes H-and He-rich material from a companion star.During infall of the material onto the neutron star gravitional energy is released which produces a constant X-ray flux.Observations of accreting neutron stars show that on top of the constant flux occasionally bright peaks occur with typical duration of 10-100s and a recurrence time of hours to days.While the neutron star accretes, the material gets compressed and heats up.At a certain point the temperature of this thin layer gets high enough for the layer to become thermonuclear unstable.What follows is a thermonuclear runaway, involving (α,p) ,(α,γ) and (p,γ)-reactions and beta-decays, which produces the characteristic light curve.It has been shown that this light curve can be used to extract stellar parameters [1].In order to extract information of the underlying system, one needs to correctly simulate the X-ray bursts behavior using nuclear physics information such as masses and reaction rates.Recent extensive burst simulations show that the uncertainty in the 23 Al(p,γ) 24 Si-rate has a large impact on the overall light curve [2].The 23 Al(p,γ) 24 Si reaction has a relatively low Q-value of only 3.29 (20) MeV, and it is therefore mainly determined by resonant capture into low-lying resonances in 24 Si.Earlier attempts to reconstruct the structure of 24 Si are not conclusive.A low-lying 2 + resonance has been identified at 3.441 (10) MeV [4] and 3.410 (16) MeV in [4] and [5], respectively.According to the mirror nucleus 24 Ne and USD shell model calculations [5], another low-lying resonance between 150-200 keV might exist, which has not been observed yet and could potentially have a significant effect on the reaction rate.We therefore performed an experiment in inverse and complete kinematics to study the surrogate reaction 23 Al(d,n) 24 Si to populate the 24 Si states of interest in the astrophysically relevant 23 Al(p,g) 24 Si reaction.First preliminary results will be presented in the next sections, proving the feasibility of this novel technique.

Experimental Details
Spectroscopic information of states in 24 Si near the proton separation energy is needed for 23 Al(p,γ) rate calculations.Thus, the main challenge in the experiment is to measure these quantities for all states up to 4 MeV in 24 Si, using a transfer method to populate the sates of interest (see [3] for more information).Our new experimental concept combines the state-of-the-art γ-ray detector GRETINA [6,7] with the neutron detector LENDA [8,9].This setup gives us the opportunity to perform a measurement in full kinematics and obtain very precise level energies from the gamma data and the properties of the states from the angular distribution of the recoiling neutrons.The experiment was carried out at the National Superconducting Cyclotron Laboratory (NSCL).A primary beam of 22 Mg at 180 MeV/A impinged on a 9 Be target producing 23 Al via multi-particle exchange reactions.The 23 Al beam was then selected using the A1900 fragment separator and transported to the S800 vault.The experiment setup at the target location included GRETINA at its pivot point and LENDA at backward angles.The beam was focused on a 110.5 mg/cm 2 CD 2 target in the center of GRETINA to produce 24 Si by the desired 23 Al(d,n)-reaction.Due to the fact that half of the hemisphere was covered by the support structure of GRETINA a new double row holding structure for LENDA was used to cover an angle from 115-175 degree with respect to the beam axis.The first and second row with 12 LENDA bars each were placed at 1 and 1.1 m distance to the target, respectively.This was the best compromise between a good time of flight resolution and a high angular coverage of the 24 30x4.5x2.5cmbars for high detection efficiency.Due to reactions on carbon atoms in the CD 2 target, runs with a 78 mg/cm 2 thick C target were used to substract background.During 7.5 days of beam time we were able to collect 96 h of data with beam on the CD 2 target and 35 h of data with the carbon target.

Beam production and particle identification
Figure 1 shows the incoming particle identification plot for the runs without the CD 2 target inserted.The beam composition was dominated by 22 Mg, resulting in a purity of 23 Al of only 13 %.Using the time of flight dependency between this two scintillators in the runs with target allows us to gate only on incoming 23 Al isotopes.The right side of Figure 1 shows the outgoing particle identification diagram gated on incoming 23 Al when using the CD 2 target.During the experiment we detected 11400 24 Si isotopes when using the CD 2 target and 3400 24 Si isotopes when using the C target.

Test case: 22 Mg(d,n) 23 Al
In order to understand and test this new experimental technique, a reaction with higher statistics was investigated: 22 Mg(d,n) 23 Al.The structure of 23 Al offers other advantages: there is only one excited state at 1616 keV that decays via γ-ray emission, all other states have a 100% proton-decay branching, since the proton separation threshold is very low (140 keV) and the first excited state is observed at 550 keV.The γ data of GRETINA show that the population of the state at 1616 keV via the (d,n)reaction is negligible.All other excited states decay to 22 Mg before arriving at the spectrograph.Thus, only the ground state of 23 Al is populated through the (d,n) reaction when gating on outgoing 23 Al fragments in the S800 focal plane.Figure 2 shows the angle of the detected neutrons of the 22 Mg(d,n) 23 Al as a function of the kinetic energy which was determined using time-of-flight.For comparison the expected kinematics line range are also shown.The kinematic line range is a consequence of the thick target.This results in a large spread of different reaction energies in the target, which can not be reconstructed uniquely.Nevertheless, the good agreement with the experimental data shows that the employed technique is successful and we are able to investigate the transfer reaction with this method.The Q-value spectrum can be reconstructed by: where m N , m in and m out are the mass of the neutron, incoming and outgoing fragment. of the outgoing fragment.By using the energy E in kin , which we expect in the middle of the target, the reconstructed Q-value spectrum is rather wide (Figure 2 red line), because of the large energy loss through the target.Using event-by-event reconstruction of the isotope trajectories in the S800, a more precise determination of the energy loss in the target was obtained.This, in turn, permits the extraction of the reaction vertex in the target and results together with an approximation of the incoming beam energy by a time-of-flight measurement of each fragment between two scintillators before the target in a much more precise E in kin value event-by-event.With this approach, we managed to obtain a resolution of 1.1 MeV (Figure 2

Figure 1 .
Figure 1.Incoming particle identification (ToF between the extended focal plan scintillator of the A1900 (xfp) plotted against ToF for fragments between the object scintillator of the S800) of runs without CD 2 target inserted (left).Reactions products identified in the S800 focal plane after the 23 Al beam hit the CD 2 target (right).

3 EPJ
E in kin , E N kin are the kinetic energies of the incoming fragment and the neutron and Ex out is the excitation energy Web of Conferences 165, 01055 (2017) DOI: 10.1051/epjconf/201716501055 NPA8 2017 blue line).A GEANT4 simulation shows that this resolution is dominated by the time resolution of the time-of-flight measurement of the neutrons which is about 951 ± 12 ps (FWHM).Analysis of the 24 Si data is currently under-way.This work was supported by US National Science Foundation under Grant Numbers PHY-08-22648, and PHY-1430152 (JINA Center for the Evolution of the Elements).The research leading to these results has received funding from the European Research Council under the European Unions's Seventh Framework Programme (FP/2007-2013) / ERC Grant Agreement n. 615126

Figure 2 .
Figure 2. Left: Kinetic energy as a function of detection angle for 22 Mg(d,n) 23 Al-reaction cuts.Expected kinematics line of this reaction is shown in red.Right: Reconstructed Q-value spectra of this reaction before and after the additional correction.