Plastic scintillator fiber detectors for heavy ion trajectory reconstruction for the Super-FRS at FAIR

At the FAIR facility, currently under construction at GSI (Darmstadt), a 1.5 AGeV uranium beam with intensities up to 2.5 × 1011 238U/spill will impinge on a graphite target at the entrance of the Super-FRS for the production of a wide range of rare isotopes by projectile fission and fragmentation. The next generation in-flight magnetic separator Super-FRS, operated up to a magnetic rigidity of 20 Tm with a large angular acceptance (Δθ = ± 40 mrad, Δϕ = ± 20 mrad) and momentum acceptance (Δ p/p = ± 2.5%), requires a new generation of tracking detectors with a position resolution of 0.2 mm (σx ) over large detector areas reaching up to 570 cm2. Besides gas detectors, planar detectors made of scintillating fibers are an option worth investigating not only because of the comparable material budget but especially for the fast response and high-rate capability. A one-dimensional prototype consisting of 128 fibers with active area of 25.6 × 100 mm2 coupled to Multi-Pixel-Photon Counters (MPPCs) and readout by FPGA TDC is described together with some recent 197Au beam test results.


Introduction
The success of the current in-flight Fragment Separator (FRS) [1], which has enabled new nuclear and atomic research, including the discovery of more than 150 new isotopes, has motivated the construction of the next-generation exotic nuclear beam facilities worldwide such as FAIR (Facility for Antiproton and Ion Research).Larger momentum and angular acceptance in the FRS successor, the Superconducting Fragment Separator (Super-FRS) [2], preserves the spatial separation principle based on the -Δ- method, and in fact it doubles it (2-stage separation).The planned Super-FRS at FAIR is shown in figure 1.
In the first phase of the FAIR operation, radioactive ion beams will be produced by fission shooting a 238 U primary ion beam produced by the synchrotron SIS18 on a target located at the entrance of the Super-FRS.More stringent requirements on position resolution (  = 0.2 mm) and rate capability (≈ 2 kHz/mm 2 ) over the active area of the Super-FRS tracking detectors are crucial for achieving a momentum resolution of Δ/ ≈ 10 −4 and get an unambiguous identification of the accepted fragments [3].The isotope identification in mass over charge ratio (A/Q) is obtained from the velocity and momentum measurement as: where  is the magnetic rigidity of the fragment,  is its velocity and  is the Lorenz factor.The magnetic rigidity of ions at first order can be reconstructed from the position measurements of the horizontal position at mid-focal plane (FMF2) and at the last focal plane (FHF1) of the Main-Separator as: where  0 is the reference magnetic rigidity,  is the dispersion and  is the magnification between FMF2 and FHF1.Additionally,  and  position measurements at the Pre-Separator and at all other focal planes are also needed to steer the primary beam from section to section as well as for calibration purpose.In particular, seven focal planes are required to be equipped with tracking detectors at the commissioning phase.
-1 - The GSI is currently able to design planar tracking detectors consisting of scintillating fibers (SciFib) 0.2 mm thick with a maximum active area of 2500 cm 2 for the R 3 B experiments [4].Thus, they are suitable to provide the 2D-position of the transmitted ions event-by-event via two perpendicularly arranged layers of fibers with minimal thickness of the traversed material as low as 0.4 mm per tracking detector.
A one-dimensional SciFib prototype built at GSI with an active area of 25.6 × 100 mm 2 was tested at the FRS using for the first time a 850 MeV/nucleon 197 Au beam.In this proceeding, after a description of the detector prototype, the experimental setup, the method to reconstruct the position of the ions, and some preliminary results of the detector test are presented.

Detector prototype
The SciFib prototype (see figure 2) consisted of 128 fibers with a quadratic cross section of 0.2×0.2mm 2 , coupled to two MPPC arrays for light collection and read out by a MPPC ReadOut Board (ROB) developed at GSI.The material selected was SCSF-78 scintillation fibers manufactured by Kuraray, having a square cross section of  = 0.2 mm, with a polystylene (PS) core and polymethylmethacrylate (PMMA) -2 -cladding whose thickness is just 2% of .The fibers are characterised by blue light emission with a peak at 450 nm, a decay time of 2.8 ns, and a long attenuation length (> 4.0 m).The photons emitted inside a single fiber at an angle smaller than 20.4 • with respect to the fiber axis are internally reflected, corresponding to 4.2% of trapping efficiency [5].
At the GSI Detector Laboratory, fibers are laid into a ribbon using a winding machine and validated by an optical inspection.Figure 3a shows the cross-section of the ribbon clearly visible at the microscope, used to verify that none of the fibers were broken or misaligned.In figure 3b a part of the one-layer prototype detector with 128 fibers divided into two bunches and sorted into two 8 × 8 hole arrays is shown.Each fiber passed through the housing is connected to the photon counter.A 3D-printed housing with a hole matrix was used for guiding and supporting each ribbon.The light collection is performed by two custom designed 8 × 8 MPPC S13360-1350PE arrays from Hamamatsu [6], as shown in figure 4a.Only one side of the fibers was connected to a 128 channel MPPC ROB (see figure 4b).The ROB board is equipped with multihit FPGA TDCs able to provide Time over Treshold (ToT) measurements with a time resolution of 200 ps ().In addition, each board can supply the MPPC with power and provide a first stage amplification [4].-3 -

JINST 19 C06008 3 In-beam detector test
The experiment was running at the mid-focal plane of the FRS [1] on June 2022.A 850 MeV/nucleon 197 Au 79+ beam was delivered by the SIS18.A beam intensity up to 10 5 ions/spill with 1.6 s spill duration was provided.A scheme of the experimental setup is show in figure 5.The SciFib prototype was placed in between two FRS TPC [7], used as reference detectors.A 1 mm thick plastic scintillator located behind the first TPC acted as a trigger and provided the start signal for the TDC measurement.
The -slits mounted in front of the first TPC were used to select the fully stripped 197 Au ions transmitted up to the setup, thanks to the presence of a Cu (90 mg/cm 2 ) target at the entrance of the separator.Figure 6a shows the SciFib mounted in front of the TPC2.Because of the presence of a solenoid not fully dismounted from the beam line, it was not possible to rearrange the configuration of the detectors in order to minimize the energy straggling of the travelling ions.ATIMA [8] simulations give a calculated angular straggling in front of the SciFib Δ ≈ 3 mrad.The TPC time signals were readout with the multi-hit TDC CAEN V1190.The calibration of the TPC detectors in the horizontal position was obtained by selecting the control-sum events according to the formula [7]: where   and   are the times extracted from left and right side of the delay line,   is the electron drift time.For the events satisfying the (3.1) condition, the ion positions in mm at each delay line were determined based on the formula: where  is gain factor (mm/ns) and  is the position offset (mm).The final position was determined as the average of the positions from the two delay lines.The count rates of each detector were collected by the VULOM4B GSI scaler module.The efficiency of the tracking detectors was estimated normalizing over the count rate measured by the plastic scintillator detector.The efficiency of the TPC1 and TPC2 detectors at rate equal to 10 4 ions/spill was 99% and 97%, respectively.Figure 6b shows the position correlation between the SciFib detector and the TPC2 at 10 4 ions/spill.In this case, the ToT method (see section 4) was used to reconstruct the horizontal position distribution at the SciFib detector.Qualitative and quantitative insight into the effectiveness of the reconstruction method used are reported in the next section.

Analysis method
The Au ions traversing the SciFib detector produced a large amount of light ( loss ≈ 300 MeV).A large amount of -rays produced from the ions traversing the material in front the SciFib prototype was observed.The effects produced by the collected -ray signals were clearly observed.As a result, eight fibers fired on average for each (triggered) event recorded, represented by a cluster.In order to identify the fired fibers (hit) at the SciFib detector, the ToT values of the pulses recorded on each fiber were used.For each event, a cluster made by adjacent strips (channels) with ToT above the discrimination threshold was identified and selected off-line.Two reconstructions method of the hits were adopted and compared.The first method is the center of gravity method.The second one consists in assigning the position of the ion to the strip with the highest ToT measurement.Both methods are graphically illustrated in the figure 8a.
The ToT histogram (see figure 7) of the hit fibers is characterized by two regions: the one at higher ToT values, where a peak corresponding to the 197 Au beam signals is expected, and the region at lower ToT values, which contains -ray and noise signals.In the ToT projection spectrum of all fibers at a rate of 10 4 ions/spill, shown in figure 7a, the two regions are well separated.The small 197 Au peak was fitted by a Gaussian distribution, resulting centered at ToT = 31.7 ns and having a  = 4.8 ns.A clear identification of the 197 Au peak was obtained after a ToT calibration by matching the fitted peak at higher ToT value for each fiber placed in the beam region.As a result of the calibration, a maximum ToT offset equal to 17 ns was found.
In figure 7b the calibrated ToT of each fiber is plotted as a function of the fiber channels at a rate of 10 4 ions/spill.The fibers hit by the Au ions are clearly identified for 55 < channel < 110.One can also notice the absence of data at channel 57 corresponding to a broken fiber, which was excluded from further analysis.Figure 7c shows that the distribution of only those channels with maximum ToT per event is characterized by a pronounced peak in the 197 Au area.The fitted peak has a centroid at about 39 ns and a width  ≈ 2.6 ns.Similar results of the fit were obtained also analyzing the data collected at higher intensities.
The results of the two position reconstruction methods are shown in figure 8b.The horizontal position distribution obtained by the nearest TPC2 (see figure 5) without offset correction is taken for comparison.The shape of the distributions obtained with the two methods are very similar.For a large charge deposited in a fiber, like in the case of the Au ions, the multiplicity of the reconstructed clusters could be very large (up to 14, with 8-channel clusters being the most common), depending on the applied ToT thresholds.For most of the reconstructed clusters, only one fiber has a very high ToT value, as expected.All other firing channels have ToT values lower and far below the 197 Au peak region shown in figure 7c.However, the cluster distributions were mostly asymmetric -6 -

JINST 19 C06008
with respect to the channel with the maximum ToT value.This could explain the presence of count fluctuations in the distribution obtained with the center of gravity method, as visible in figure 8b.
The efficiency of the SciFib detector choosing the plastic scintillator detector as a reference was close to 1 for both methods.It remained close to 100% with the increasing of the beam intensity, whereas the efficiency of the TPC detectors showed a slight decreasing and require further analysis investigations.

Position resolution results
Due to the large distance between TPC1 and TPC2 (over 2 m according to figure 5) and the direct proximity of the SciFib detector to the TPC2, it was decided that the reference measurement would be that from TPC2 instead of calculating an interpolated position based on data from both TPCs detectors.The spatial resolution can be defined as the standard deviation  Δ of the quantity  Despite the efficient reconstruction of the clusters, this result, surprisingly, is a factor 2.25 larger than the width of the single fiber.Preliminary investigations on the position dependence of the width of the quantity calculated by equation (5.1), provide a smaller value  Δ ≈ 0.5 mm when stringent conditions on the TPC2 position, e.g.0 <  < 4 mm on TPC2, are applied.We cannot exclude that the present results will not improve after performing a more accurate TPC position calibration.
At higher rates, the width of the position distribution of the SciFib prototype remains constant, as expected.On the contrary, at rate higher than 10 4 ions/spill, the position distribution of the reference -7 -detector showed some unexpected changes of the width.This effect, until now not understood, prevents to be conclusive on the results obtained by a direct comparison.

Summary
The position distribution of the SciFib prototype with an active area of 25.6 × 100 mm 2 built at the GSI Detector Laboratory was tested at GSI.The distribution was reconstructed by using two methods, which showed similar results.Up to a rate of 10 4 197 Au/spill, the obtained widths of the two reconstructed horizontal position distributions reproduced well the width measured with the standard FRS TPC detector.Improvements in the accuracy of the measured central position, presently shifted by few mm, are still possible.Currently, the present measured position resolution of the SciFib prototype over the detector area is about 0.45 mm.A slightly better spatial resolution was found in the region 0 <  TPC2 < 4 mm.
A more accurate position calibration of the reference detectors is mandatory to conclude whether a spatial resolution comparable to the width of the single fiber is achievable or not.

Figure 1 .
Figure 1.The FAIR facility with the planned Super-FRS.

Figure 2 .
Figure 2. The SciFib prototype installed at the FRS beam line.

Figure 3 .
Figure 3. (a) Fiber ribbon monitored at the microscope; (b) SciFib detector with sorted fibers glued in a hole matrix on the bottom panel to which the MPPC arrays were attached.

Figure 5 .
Figure 5. Scheme of the experimental setup with distances between detectors.Transverse and longitudinal dimensions are not in scale.

Figure 6 .
Figure 6.(a) SciFib mounted in front of the second TPC and SciFib prototype out of the beam position; (b) Position correlation between the TPC2 and the SciFib detectors at 10 4 ions/spill.

Figure 7 .Figure 8 .
Figure 7. (a) Raw ToT projection spectrum of all fired channels for intensity of 10 4 ions/spill; (b) Calibrated ToT projection of all fibers as a function of the fiber channel; (c) ToT distribution selecting only the fiber with maximum ToT within one event.
.1)    with  TPC2 the position obtained from the measurements at the TPC2 detector and  SciFib the reconstructed position at the SciFib detector.The Δx distributions calculated using both reconstruction methods are plotted in figure9at an intensity of 10 4 ions/spill.A quantitative analysis was done to determine the spatial resolution based on equation (5.1).The Gaussian fits of the two distributions are also shown in figure9.The results of the two fits exhibit the same dispersion value with  Δ ≈ 0.6 mm.After taking into account the angular straggling contribution ( strag ≈ 0.4 mm), a spatial resolution of   ≈ 0.45 mm is obtained.