New readout codification of large-area multi-gap timing RPCs for Muon Scattering Tomography

Contents lists available at ScienceDirect Nuclear Inst. and Methods in Physics Research, A journal homepage: www.elsevier.com/locate/nima New readout codification of large-area multi-gap timing RPCs for Muon Scattering Tomography João Saraiva ∗, Alberto Blanco LIP, Laboratory of Instrumentation and Experimental Particle Physics, Portugal A R T I C L E I N F O Keywords: Gaseous detectors Resistive Plate Chambers Readout codification Muon scattering tomography Monte Carlo simulations A B S T R A C T A new readout technique decoupling the number of FEE channels from the detector area was tested with large-area Resistive Plate Chambers (RPCs) featuring a sensitive area of 130 × 90 cm2. Despite using only 48 electronic channels to read out 888 pick-up strips, a 2D submillimetric spatial precision was achieved during a long run with cosmic rays, along with a time precision of 89 ps (𝜎). FLUKA Monte Carlo simulations were also performed to evaluate the use of RPCs with high spatial and temporal precision for scanning large volumes employing the Muon Scattering Tomography technique. 1. Introduction A novel readout system of Resistive Plate Chambers (RPCs) was designed for large surface applications requiring high spatial and temporal precision, such as the Muon Scattering Tomography (MST). Considering that the cost of RPC systems is often driven by the front-end electronics, a new codification was developed to significantly reduce the dependence of the number of electronic channels on the detector area, without substantially decreasing its performance. The new method was first tested with a double-stack multi-gap timing RPC (tRPC) with a sensitive area of 30 × 30 cm2, using 24+24 preamplifiers to read out 120+120 strips. A spatial precision better than 1 mm and time precision below 100 ps were achieved with this prototype, as reported in [1]. Subsequent measurements were carried out with a detector 13 times larger in area (130 × 90 cm2), using the same number of front-end electronics (FEE) channels (24+24) to read out a considerably higher number of strips: 360 longitudinally and 528 transversely. In this paper, we briefly describe the new readout codification and report the results obtained with the large-area tRPCs. Monte Carlo simulations are also presented, highlighting the importance of a millimeter spatial precision and very good time precision for the MST technique. 2. Readout system The new readout method involves two types of readout boards, one of which having much wider strips than the other: the thin-strip readout electrodes with a pitch of 2.54 mm and interstrip of 1 mm, and the wide-strip readout board with a 61-mm pitch and interstrip ∗Corresponding author. E-mail address: [email protected] (J. Saraiva). of 2 mm (see Fig. 1). The thin-strip readout electrode provides, via charge interpolation, the fine position of events in one dimension. Two of them were used, with the strips oriented in orthogonal directions, to achieve a fine 2D position. The wide-strip readout electrode, in turn, simultaneously provides the time and an additional 2D coarse position of interactions, as described further. These pick-up electrodes are used in combination with a double-stack tRPC, with the thin-strip boards positioned at the top and bottom of the stack, and the wide-strip board placed at the center, between the tRPCs (refer to the layer diagram in the next section for more details). To reduce the number of electronic channels, each channel reads out N thin strips previously connected in parallel by the Signal Merging PCB (SMPCB), as the one depicted in Fig. 1. The SMPCB routes signals from N strips into a single track, transmitting them to the respective electronic channel without any processing. To protect the FEE, all strips on the SMPCB were grounded through 20 MΩ resistors, preventing electrical floating. With this novel approach, the sensitive area of a detector can be significantly increased while maintaining the same number of preamplifiers. This is accomplished by simply modifying the number of thin strips connected in parallel (N). To determine in which thin strip the signal was in fact induced, the wide-strip readout electrode must be added to the setup. With both extremities of the wide strips connected to current-sensitive amplifiers, the additional readout board provides a 2D coarse position of the interactions in the detector, needed to resolve the ambiguity raised by adding thin strips in parallel. It also enables measuring the time of events, useful for Time Of Flight (TOF) applications. https://doi.org/10.1016/j.nima.2025.170466 Received 29 January 2025; Received in revised form 17 March 2025; Accepted 24 March 2025 Nuclear Instruments and Methods in Physics Research A 1076 (2025) 170466 Available online 2 April 2025 0168-9002/© 2025 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies. J. Saraiva and A. Blanco Fig. 1. (a) Example of pick-up electrodes required for the new readout codification and tested with a 30 × 30 cm2 prototype [1]. Left: thin-strip PCB with 120 strips measuring 36 cm in length; right: wide-strip PCB with 5 strips 38 cm long. (b) Signal Merging PCB (SMPCB) with the bottom view showing both sides of the board and how strips are wired in parallel. The wide-strip pick-up electrode must be designed in such a way that the strip pitch corresponds approximately to the width of the group of thin strips connected to separated electronic channels: 24 thin strips for the tests here reported.1 Moreover, while the number of FEE channels connected to the thin strips remains unchanged when increasing the detector area, new channels must be added to read out the additional wide strips present in the larger detector. However, because of the significant width of these strips, the increase in the number of electronic channels can be considered residual. 3. Experimental setup 3.1. Large-area detector As initially described in [1], the new readout codification based on the SMPCB was first tested with a 30 × 30 cm2 prototype. The same 24channel SMPCB was afterwards used with the 130 × 90 cm2 detector shown in Fig. 2. Both detectors have an identical layered structure, namely: •a stack of two multi-gap tRPCs each with 6 gas gaps, 300 μm wide; •resistive electrodes made of float glass, 1.1 mm thick; •two thin-strip readout PCBs with strips oriented orthogonally to achieve a 2D fine position of ionizing events, one located at the bottom and the other at the top of the stack; •a wide-strip readout PCB at the center of the stack, for an additional 2D coarse position of events and providing also the respective time of events. With the 30 × 30 cm2 prototype, 24 electronic channels were needed to read out 24 groups of N= 5 thin strips wired in parallel.2 124 thin strips × 2.54 mm/strip = ∼61 mm, corresponding to the pitch of the wide-strip readout PCB. 2A total of 24 × 5 = 120 strips per thin-strip readout board. Fig. 2. Partial view of the large-area detector (active area of 130 × 90 cm2) used with the novel readout technique. At the bottom of the stack: three SMPCBs connected in a daisy-chain configuration, creating 24 groups of 15 thin strips wired in parallel (360 longitudinal thin strips in total); the wide-strip readout board is also visible just below the top tRPC. The thin-strip readout board located at the top of the stack was removed for better visibility of the underlying layers. Fig. 3. Left: layer diagram of the large-area detector (1.17 m2). Right: detailed view of the stacked layers (top thin-strip readout board removed). The daisy-chain link between two SMPCBs is also visible (black jumpers at the top of the figure). For the 130 × 90 cm2 detector, in turn, N increased from 5 to 15 for the thin-strip readout board with longitudinal strips (X) and from 5 to 22 for the board with transverse strips (Y). In this way, despite the substantial increase in area, the same 24 channels are employed for each dimension (X, Y) to read out the same 24 groups of thin strips, though each has additional strips connected in parallel.3 As previously mentioned, a residual increase of electronic channels occurs due to the additional wide strips required with the larger detector: only 20 additional channels were needed (from 58 to 78 channels,4 i.e. a 1.3-fold increase), while the detector area was scaled up by a factor of 13 (from 0.09 to 1.17 m2) and the number of strips, required for fine spatial measurements, increased from 240 to 888. A close-up view of the SMPCBs used with the wide-area detector and its layer diagram can be seen in Fig. 3. The detector was operated, with a gas mixture of 99% of R-134a and 1% of SF6, during weeks using a coincidence trigger generated externally by plastic scintillators5 (Bicron BC420) located few centimeters below and above the detector. The reduced electric field was set to approximately 360 Td (around 2.7 kV/gap, 90 kV/cm) leading to an efficiency close to 95%. 3A total of 24 × 15 = 360 longitudinal thin strips and 24 × 22 = 528 transverse thin strips. 4Number of electronic channels for the 30 × 30 cm2 prototype: 24+24+5+5 = 58, namely: 24 channels for the longitudinal thin strips (X), 24 for the transverse ones (Y) and 5+5 for both ends of 5 wide strips. For the 130 × 90 cm2 detector: 24+24+15+15 = 78, i.e. only 10+10 additional channels were needed to read out the extra 10 wide strips. 5Two parallelepipeds of 8 × 3 × 2 cm3, coupled to photomultiplier tubes Hamamatsu H6533. Nuclear Inst. and Methods in Physics Research, A 1076 (2025) 170466 2 J. Saraiva and A. Blanco Table 1 Detailed information on the type of preamplifiers connected to the readout electrodes, as well as the measured and extracted quantities with the respective FEE. Charge-sensitive amp Current-sensitive amp Pick-up electrode Thin-strip PCB Wide-strip PCB Strip readout Single-ended Double-endeda Measured quantity Charge Charge, Time (T) Extracted quantity 1D fine position; with two orthogonal boards: (X𝑓𝑖𝑛𝑒 , Y𝑓𝑖𝑛𝑒 )b 2D coarse position: X𝑐𝑜𝑎𝑟𝑠𝑒b, Y𝑐𝑜𝑎𝑟𝑠𝑒 = (T𝑓−T𝑏)/2 event time: T = (T𝑓+T𝑏)/2 a Front (f) & back (b) sides of the strips. b Via charge interpolation. 3.2. Electronics & data acquisition system Two types of FEE were used with the new readout method here described: (i) charge-sensitive amplifiers connected to the thin-strip readout electrodes and used to integrate the induced signals over few microseconds, corresponding to the time needed for both components of the generated ion-electron pairs to reach the respective resistive electrodes; (ii) current-sensitive amplifiers for the readout of the widestrip pick-up electrode, providing time and charge measurements using only the electron-dominated part of the induced signals. The Measured and derived quantities achieved with this setup are summarized in Table 1. For the thin strips, a custom-designed FEE was chosen relying on the ultralow-noise operational amplifier AD8599 configured as integrator. Featuring a 10 MHz bandwidth, the preamp is therefore insensitive to signal reflections that might occur by the fact of having instrumented only one end of the thin strips. With an integration time constant set to 100 μs, the charge induced by both the electronic and ionic components of the signal is collected, maximizing in this way the Signal Noise Ratio (SNR) and therefore enhancing the precision computing the event position. The FEE of the HADES-RPC TOF experiment (GSI, Darmstadt) was used for the readout of the wide strips. It includes a fast 2.1 GHz BGM1013 preamp feeding a high-speed MAX9601 comparator that outputs a digital signal providing the event time with a precision below 35 ps (𝜎) [2]. The preamplified signal is also routed through a second branch and then integrated by a OPA690 amplifier with an RC time constant of 40 ns, allowing for a Time over Threshold (ToT) measurement of the integrated signal. In this way, the digital signal generated by the FEE includes both the time and charge information, encoded in the leading edge and the width of the signal, respectively. The data acquisition system is based on the multi-purpose Trigger Readout Board - v3 (TRB3) [3], an FPGA-based platform with four peripheral FPGAs with a 32-channel Time-to-Digital Converter (TDC) implemented in each of them, featuring a time precision below 20 ps [4]. An Analog-to-Digital Converter (ADC) expansion card6 was connected to one of the peripheral FPGAs to digitize the integrated signals from the charge-sensitive amplifiers. 4. Experimental results As previously mentioned, the main goal of the readout technique described in this work, is to decouple the number of FEE channels from the readout area, allowing for an increase of the surface of the detector without substantial change in the number of electronic channels. To assess the impact of this method on detector performance, two measurements were compared between the 30 × 30 cm2 prototype and the 130 × 90 cm2 tRPCs here presented: (i) the time precision of 6Equipped with twelve 4-channel AD9219, featuring a resolution of 10 bits and sampling rate of 40 MHz. Fig. 4. Time precision of the 130 × 90 cm2 tRPCs, decreasing from 109 to 89 ps (𝜎) after removal of the scintillator contribution. both detectors, inferred by subtracting the measured time of the tRPCs from that of the scintillators used for the coincidence trigger; (ii) the details obtained in the reconstructed image of the scintillators projected on the tRPCs. The time precision of the large-area tRPCs, achieved over a 50-day acquisition period, is presented in Fig. 4. After applying the time walk correction and removing the scintillator contribution,7 a final precision of 89 ps (𝜎) was achieved, a few tens of picoseconds above the value obtained with the first prototype [1]. The slight degradation in time precision observed with the larger detector might simply be attributed to the initial conditioning of the tRPCs, as the precision systematically improved over time, reaching 73 ps (𝜎) when considering only the last 10 days of acquisition. The projected shadow of the scintillators on the tRPCs is shown in Fig. 5. The presence of 300 μm-diameter fishing lines between the glass electrodes of the tRPCs results in the complete inhibition of gas ionization at this location. The absence of events at the spacer location in the reconstructed image of the scintillators was therefore expected, indicating that the spatial precision of the tRPCs should be submillimetric. This observation leads us to conclude that the increase in electronic noise, resulting directly from adding strips in parallel with SMPCBs,8 was not significant enough to compromise the detection of the 300 μm spacers. 5. Monte Carlo simulations applied to muon tomography 5.1. Muon Scattering Tomography (MST) Materials of high atomic number (Z) concealed in medium-sized volumes can be identified by the Muon Scattering Tomography (MST) technique, as first reported in 2003 [5]. Based on Multiple Coulomb Scattering (MCS) of naturally occurring cosmic-ray muons, this technique enables mapping the internal structure of the scanned object. To reconstruct the scattering events within the fiducial region, a minimum of two points along the muon path must be measured, both upstream and downstream of the analyzed object. A stack of four detectors is therefore required, each with sensitive areas ranging from a few to tens of square meters, resulting inevitably in a large number of FEE channels. 7Obtained by computing the time differences between detectors (tRPC and scintillators 1 and 2 (SC1, SC2)) and solving the following system of equations: (1) 𝜎2 𝑡𝑅𝑃 𝐶 +𝜎2 𝑆𝐶1=𝜎2 𝑡𝑅𝑃 𝐶−𝑆𝐶1,(2) 𝜎2 𝑡𝑅𝑃 𝐶 +𝜎2 𝑆𝐶2=𝜎2 𝑡𝑅𝑃 𝐶−𝑆𝐶2,(3) 𝜎2 𝑆𝐶1+𝜎2 𝑆𝐶2= 𝜎2 𝑆𝐶1−𝑆𝐶2. 8Up to 22 strips of 90 cm linked in parallel, corresponding to a total strip length of almost 20 m. Nuclear Inst. and Methods in Physics Research, A 1076 (2025) 170466 3 J. Saraiva and A. Blanco Fig. 5. (a) 2D position map of the 8 × 3 × 2 cm3 scintillators. Positioned with the 3 cm side aligned vertically, the scintillators project a shadow of approximately 8 cm × 2 cm on the tRPCs. The red arrow identifies the location of the nylon monofilament spacers installed longitudinally between the resistive electrodes. (b) projection of the 2D event map onto the X axis, showing a clear reduction in the number of events at the location of the 300-μm gas gap spacers. RPCs with almost 2 m2 were used in 2021 to infer the presence of small blocks of aluminum, iron and tungsten within a scanned volume slightly below 1 m3 [6]. At that time, despite the detector not being specifically designed for the MST technique, a 5 cm-thick tungsten block was identified in 10 min. The detector was equipped with 512 electronic channels, featuring a spatial precision at the centimeter level. With the readout method presented in this communication, a spatial precision on the submillimeter scale could have been achieved with a considerably lower number of FEE channels. 5.2. Monte Carlo simulations To assess the impact of high spatial and temporal precision on the MST method, several Monte Carlo simulations were conducted using the following methodology: •generate the muon flux at sea level using two alternative methods: –the FLUKA atmospheric model composed of 100 spherical layers from 0 to 70 km above sea level, each with different air densities [7,8]; –a muon generator developed to feature a 𝑐𝑜𝑠2(𝜃) zenith angle distribution and a differential energy spectrum proportional to 𝐸−2.7 (flat below 2 GeV); •after validation,9 any of the muon fluxes could be loaded as source term and transported in second step simulations using the FLUKA code. Two different geometries were created10 for this purpose (see Fig. 6): –a mid-sized geometry (fiducial region of ∼0.9 m3), equivalent to the muon telescope used in 2021 [6]; –an extensive geometry (fiducial region of ∼154 m3), corresponding to a full-size truck carrying a large sea container; •for each simulated scenario, the energy and coordinates of every muon crossing a detector plane were dumped using a customized mgdraw.f FLUKA subroutine. Several simulations were carried out with different geometry variants, namely: –the full geometry: including a 10 × 10 × 10 cm3 block of high-Z material,11 4 RPCs, the surrounding air, and in case of the large geometry: the floor, the truck platform and the container situated above it; 9Being, for instance, in very good agreement with the muon energy spectrum from [9]. 10 Using the FLAIR graphical interface [10]. 11 Several material were tested: uranium, tungsten, iron, etc. The results obtained with tungsten are here reported. Fig. 6. Geometries used to evaluate the effect of the material budget and the spatial and temporal precision on the MST technique. Left: medium-sized geometry including 4 RPCs, each measuring 1.6 m × 1.2 m, and a block of high-Z material inside the volume between the two inner planes (fiducial region). Right: significantly larger fiducial region (also with 2 detector planes on the top and bottom sides), including a 15-cm-thick concrete floor, a truck platform with a thickness of 10 cm and steel walls 2 mm thick, and a shipping containers also made of steel, with 2 mm thick walls and of dimensions: 13.7 m × 2.5 m × 2.9 m. A 10 × 10 × 10 cm3 high-Z block is also present inside the container. –equivalent to the previous geometry, but without the block of high atomic number. The muon scatterings generated from this simulation can be considered as ‘noise’ since related to all the materials except the one of interest; –as opposed to the last configuration, a geometry including only the high-Z object generates scatterings exclusively related to this material, thus corresponding to the expected ‘signal’; –another variant also considered, included only the surrounding air, to estimate its contribution to the total angular distribution of the muons after crossing the fiducial region. •the collected data from the Monte Carlo calculations could then be used to: –perform scattering angle distributions of muons within the fiducial region, as shown in Fig. 7; –reconstruct the 3D image of the scanned object plotting directly the midpoint of the common perpendicular (also known as point of the closest approach (POCA)) between incident and exiting trajectories, as illustrated in Fig. 8; –estimate the exposure time required to identify the high-Z material based on the SNR achieved after populating the voxelized fiducial region with the POCAs, as depicted in Fig. 9. 5.3. Simulation outcomes 5.3.1. Material budget effect on angular distribution Comparing scattering angle distributions of simulations where components are progressively removed from the geometry, allowed us to assess the contribution of each material to the total scattering of the muons after crossing the fiducial region. For the simulations corresponding to the muon telescope used in 2021 (mid-sized geometry of Fig. 6), it was found that the impact of the RPCs12 on the cumulative scattering is indeed considerable. In fact, as shown in Fig. 7 (left), the MCS on the tungsten block only start to be dominant for large deflections (above ∼12◦), mainly because of the presence of the RPCs. This result corroborated the experimental observations of 2021, where the detection of the tungsten block could be improved discarding angular deflections up to 11◦ [6], i.e. removing a substantial number of events, and consequently increasing the exposure time needed for the material identification. 12 Each detector plane, consisted of a stack of 2 multi-gap RPCs with 2 gas gaps each (1 mm wide) and glass electrodes 2 mm thick, hence a total of 2 × 3 × 2 mm of glass plus 2 × 2 mm for the aluminum box. Nuclear Inst. and Methods in Physics Research, A 1076 (2025) 170466 4 J. Saraiva and A. Blanco Fig. 7. (a) Scattering angle distributions as a function of material budget for the midsized geometry. (b) Same distributions but removing muons with energy up to 500 MeV. Fig. 8. 3D reconstruction of the scanned truck, container and tungsten block (10 × 10 × 10 cm3), for an exposure time corresponding to a duration of 55 min. The POCAs were plotted directly in the 3D volume, applying the following restrictions: only scattering angles above 5.5◦ (left), and removing also muons with energy below 1 GeV (right). An equivalent reconstruction image for the mid-sized geometry can be found in [6]. The simulations also showed that, by rejecting low-energy muons (a few hundred of MeV), the scatterings from the tungsten block become dominant over almost the entire scattering range (Fig. 7 (right)). Since muons of low energy are highly scattered, even in the surrounding air, removing them improves not only the reconstructed image but also the exposure time required to detect high-Z materials, as further discussed. For the muon telescope used in 2021, it would have considerably reduced the time necessary for the tungsten identification. 5.3.2. TOF for low-energy muon rejection The TOF technique can be employed for the aforementioned purpose of removing low-energy muons. For instance, with a time precision of 100 ps (similar to the precision achieved with the new readout method here presented), muons with energy up to 500 MeV can be rejected for the medium-sized geometry of Fig. 6 and up to 1 GeV in case of the truck geometry, where the distance between outer planes exceeds 5 m. The larger distance between detectors also improves relative momentum resolution. Fig. 8 shows the reconstructed image of the truck and its container with the tungsten block inside, after more than 50 min of acquisition. As previously stated, removing muons with energy below 1 GeV (Fig. 8 (right)) significantly improves the image of the scanned object, making it possible to identify the different elements of the geometry: the concrete floor, above it the truck deck supporting the container, the walls of the container (better observed from a frontal perspective) and the tungsten block at the center. 5.3.3. Time required to identify the tungsten block To estimate the exposure time necessary to identify the tungsten block in both tested geometries, the following approach was established: (i) discretize the fiducial region into voxels: 5 × 5 × 5 cm3 for the mid-sized geometry and 10 × 10 × 10 cm3 for the extensive one; (ii) add a Gaussian noise to the muon positions to emulate the detector’s spatial precision; (iii) compute the POCAs and allocate them to the corresponding voxels; Fig. 9. SNR (ratio between the highest values of the normalized number of POCAs per voxel, inside and outside the correct location of the high-Z material) vs. exposure time in minutes: (a) calculated for the mid-sized geometry, with spatial precisions up to 1.0 cm; (b) obtained for the truck geometry using the same detector’s spatial precisions. An additional curve was added showing the significant SNR decrease when no restrictions (in scatter and energy) are applied. (iv) calculate the average number of POCAs per voxel to identify the voxels with significantly higher concentrations of POCAs; (v) reduce the exposure time and repeat all the above steps. Outlier voxels with a number of POCAs exceeding a normalized threshold can then be plotted to identify their location within the fiducial region. For the performed simulations, where the position of the high-Z material is known, it is possible to assess whether the plotted voxels are accurately located, and compute the SNR by comparing the highest outlier values inside and outside the correct position. By doing so, it was observed that: •for all the simulations presented in this study, and based on the obtained SNR (see Fig. 9) but also on the density of outliers in a particular region, one minute of exposure time is enough to reliably identify the 10 × 10 × 10 cm3 tungsten block; •for both considered geometries, a detector’s spatial precision at the centimeter level is not ideal for the MST technique, as the SNR remains close to one for all exposure times, requiring for instance a higher scattering angle restriction; •for the truck simulations: –restrictions on scattering deflection and muon energy must be applied to improve the SNR, reinforcing the importance of very good temporal precision to use the TOF technique; –a submillimeter spatial resolution is not of paramount importance for the achieved SNR. 6. Conclusion A novel readout technique was developed to decouple the number of electronic channels from the detector area, without substantial degradation of its performance. The approach relies on the Signal Merging PCB that connects several pick-up strips in parallel before being fed to a single FEE channel. The new method was tested with multi-gap tRPCs, 130 × 90 cm2 in area, where up to 22 strips (each ∼90 cm long) were wired together and routed to the same electronic channel. The tRPCs were operated in coincidence trigger with scintillators, measuring the cosmic ray flux during weeks. A time precision of 89 ps (𝜎) was achieved, corresponding to a slight degradation in comparison to a previously tested small prototype (30 × 30 cm2) that used the same number of FEE channels. This time precision improved systematically over time, indicating that the observed degradation might be related to the initial conditioning of the tRPCs. Moreover, despite the significant number of strips connected in parallel with this technique, a 2D submillimetric spatial precision was achieved. FLUKA Monte Carlo calculations applied to the Muon Scattering Tomography were also performed to assess the feasibility and relevance Nuclear Inst. and Methods in Physics Research, A 1076 (2025) 170466 5 J. Saraiva and A. Blanco of using RPCs, with very good spatial and temporal precision, to scan large volumes (above 150 m3). The simulations revealed that: (i) the material budget of the RPCs has an appreciable impact on the multiple scatterings suffered by the muons, especially the low-energy ones; (ii) for very large geometries, where the single-scattering approximation shows its limitations, a millimeter spatial precision is enough to quickly identify high-Z materials, while the submillimeter precision level does not improve significantly this capacity; (iii) a time precision on the order of 100 ps is essential to overcome the limitations mentioned in the previous points, allowing for the rejection of highly scattered lowenergy muons via the TOF technique. Finally, based on the simulation results and assuming millimeter spatial precision and time precision around 100 ps, an exposure time of one minute was sufficient to detect a 10 × 10 × 10 cm3 tungsten block in the tested geometries. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The authors acknowledge the GSI and LIP, with special thanks to the members of the detector laboratory and the mechanical workshop of LIP for their support. This work was supported by the Foundation for Science and Technology (Portugal) (CERN/FIS-INS/0006/2021) and the European Union’s Horizon 2020 Research and Innovation programme under Grant Agreement AIDAinnova n.◦ 101004761. References [1] J. Saraiva, A. Blanco, NIMA 1068 (2024) 169803, http://dx.doi.org/10.1016/j. nima.2024.169803. [2] D. Belver, et al., IEEE Trans. Nucl. Sci. 57 (2010) 2848–2856, http://dx.doi.org/ 10.1109/TNS.2010.2056928. [3] TRB-family, trb.gsi.de. [4] A. Neiser, et al., JINST 8 (C) (2013) 12043, http://dx.doi.org/10.1088/17480221/8/12/C12043. [5] K. Borozdin, et al., Nature 422 (2003) 277, http://dx.doi.org/10.1038/422277a. [6] J. Saraiva, et al., NIMA 1050 (2023) 168183, http://dx.doi.org/10.1016/j.nima. 2023.168183. [7] G. Battistoni, et al., Ann. Nucl. Energy 82 (2015) 10–18. [8] C. Ahdida, et al., Front. Phys. 9 (2022) 788253, http://dx.doi.org/10.3389/fphy. 2021.788253. [9] PDG, Rev. Part. Phys. Prog. Theor. Exp. Phys. 083C01 (2020) http://dx.doi.org/ 10.1093/ptep/ptaa104. [10] V. Vlachoudis, M & C International Conference, ISBN: 9781615673490, 2009, pp. 790–800. Nuclear Inst. and Methods in Physics Research, A 1076 (2025) 170466 6