The conceptual design of the miniBeBe detector proposed for NICA-MPD
Ramón Acevedo Kado, Mauricio Alvarado Hernández, Alejandro Ayala, Marco Alberto Ayala Torres, Wolfgang Bietenholz, Dario Chaires, Eleazar Cuautle, Isabel Domínguez, Alejandro Guirado, Ivonne Maldonado, Julio Maldonado, Eduardo Moreno-Barbosa, P. A. Nieto-Marín, Miguel Enrique Patiño Salazar, Lucio Rebolledo, Mario Rodríguez-Cahuantzi, D. Rodríguez-Figueroa, Valeria Z. Reyna-Ortiz, Guillermo Tejeda-Muñoz, María Elena Tejeda-Yeomans, Luis Valenzuela-Cázares, C. H. Zepeda Fernández
TThe conceptual design of the miniBeBe detector proposed for NICA-MPD
Ram´on Acevedo Kado , Mauricio Alvarado Hern´andez , Alejandro Ayala , , Marco Alberto Ayala Torres ,Wolfgang Bietenholz , Dario Chaires , Eleazar Cuautle , Isabel Dom´ınguez , Alejandro Guirado , IvonneMaldonado , Julio Maldonado , Eduardo Moreno-Barbosa , P. A. Nieto-Mar´ın , Miguel Enrique Pati˜no Salazar ,Lucio Rebolledo , Mario Rodr´ıguez-Cahuantzi , ∗ , D. Rodr´ıguez-Figueroa , Valeria Z. Reyna-Ortiz , GuillermoTejeda-Mu˜noz , Mar´ıa Elena Tejeda-Yeomans , Luis Valenzuela-C´azares , and C. H. Zepeda Fern´andez , Instituto de Ciencias Nucleares,Universidad Nacional Aut´onoma de M´exico,Apartado Postal 70-543, CdMx 04510, Mexico Centre for Theoretical and Mathematical Physics,and Department of Physics, University of Cape Town,Rondebosch 7700, South Africa Centro de Investigaci´on y Estudios Avanzados del IPN,Apartado Postal 14-740, CdMx 07000, Mexico Facultad de Ciencias - CUICBAS,Universidad de Colima,Bernal D´ıaz del Castillo No. 340,Col. Villas San Sebasti´an, 28045 Colima, Mexico Facultad de Ciencias F´ısico-Matem´aticas,Universidad Aut´onoma de Sinaloa,Avenida de las Am´ericas y Boulevard Universitarios,Ciudad Universitaria, C.P. 80000,Culiac´an, Sinaloa, Mexico Departamento de Investigaci´on en F´ısica,Universidad de Sonora,Boulevard Luis Encinas J. y Rosales,Colonia Centro, Hermosillo, Sonora 83000, Mexico Facultad de Inform´atica,Universidad Aut´onoma de Sinaloa,Avenida de las Am´ericas y Boulevard Universitarios,Ciudad Universitaria, C.P. 80000,Culiac´an, Sinaloa, Mexico Facultad de Ciencias F´ısico Matem´aticas,Benem´erita Universidad Aut´onoma de Puebla,Av. San Claudio y 18 Sur, Edif. EMA3-231,Ciudad Universitaria 72570, Puebla, Mexico Universidad de las Am´ericas, Puebla, MexicoEx-Hacienda Sta. Catarina M´artir San Andr´es Cholula,Puebla C.P. 72820, Mexico C´atedra CONACyT, CdMx 03940, Mexico a Corresponding author: [email protected] submitted to Journal of Instrumentation (JINST) (Dated: July 24, 2020)We present the conceptual design for the miniBeBe detector proposed to be installed as a level-0trigger for the TOF of the NICA-MPD. We discuss the design and the geometrical array of itssensitive parts, the read-out electronics as well as the mechanical support that is envisioned. Wealso present simulation results for a wide range of multiplicities obtained in p + p and A + Acollisions to study its capabilities, both as a level-0 trigger for the TOF as well as to serve as abeam-gas interaction veto and to locate the beam-beam interaction vertex.
I. INTRODUCTION
MiniBeBe is a detector designed to provide a wake-uptrigger signal for events ranging from low to high mul-tiplicities, for the Time-Of-Flight (TOF) of the MPD(Multi-Purpose Detector) at the NICA (Nuclotron-basedIon Collider fAcility) complex. The detector name stems from the acronym of “Beam-Beam” counter. Given thatits dimensions make it to be overall small, the name hasbeen supplemented with the prefix “mini”.In order to reliably separate pions, kaons and protonsin a wide range of momenta, the TOF is expected tohave an overall time resolution better than 100 ps. Thisrequires the trigger signal to be optimized. The nominalMPD element designed to provide this trigger is the Fast a r X i v : . [ phy s i c s . i n s - d e t ] J u l Forward Detector (FFD) [1], which — in simulations —has proven to be very efficient for central and semi-centralA + A collisions, although its efficiency decreases for lowmultiplicity events. To improve the trigger, the miniBeBeis required to be efficient for low multiplicity p + p, p +A and A + A events as well as to have a fast response.The target for the miniBeBe time resolution has been setto 30 ps at most. Given that the miniBeBe is proposedto be located near the beam pipe around the interactionpoint, another requirement for its design is a low materialbudget.The sensitive detector elements consist of SiliconPhoto Multipliers (SiPMs) coupled to plastic scintillator.SiPMs have been widely used during the past two decadesin different areas, not only in high energy physics [2, 3],but also in medical instruments, for example, devicesbased on the Positron Emission Tomography (PET) tech-nique [4], where — depending on the particular applica-tion — one replaces the use of Photo Multiplier Tubes(PMTs) [5], given that their voltage operation, cost andsmall size make them more versatile than PMTs.Since the year 2013, the SensL corporation has devel-oped SiPMs with two kinds of outputs: standard andfast, as described in Ref. [6]. The main characteris-tic of the fast output is a short pulse width, rangingfrom ps to some ns depending on the model. The de-velopment of this recent technology prompted several re-search works [7, 8], particularly for TOF measurementsof charged particles requiring a fast detector attached toscintillator materials, for instance in PET instrumenta-tion [4, 9]. The configuration scintillator-SiPM has al-ready been studied and tested before [10] to replace thePMTs with SiPMs. Recent research reports also on theuse of plastic scintillators coupled to SiPMs for PET ap-plications [11–14].In high-energy physics, particle detectors also makeuse of SiPMs, as described for example in Ref. [3]. Thistechnology has become particularly useful for CERN ex-periments, such as ALICE, to achieve a fast integrationtrigger [15] and for the first calibrations of the beamlosses [16]. For the NICA project, SiPMs have alreadybeen considered for the design of the ElectromagneticCalorimeter, as reported in Refs. [2, 17].In general, most of the applications based on SiPMs usethe standard output to estimate the charge deposited bythe photo current acquired in the anode, which is relatedto the deposited energy [18, 19]. Only until recently,the means to process the fast output signal have been re-leased. As described in Ref. [20], the pulse shape discrim-ination for fast neutrons and gamma rays was performedusing the fast signal output from 6 × SensL MicroCSiPMs. In addition, Ref. [21] reports a time resolutioncomparison between fast and slow signals for 3 × SensL SiPMs, showing an equivalent coincidence resolv-ing time for the fast and standard outputs connected toan external shaping filter, optimized to match the fastpulse shapes before the timing comparator.In this work we incorporate these developments into the design of the miniBeBe detector. We describe indetail the concept for the miniBeBe design and report onthe progress of the construction of its parts, including thearray of sensitive elements, the read-out electronics andthe mechanical support. We also present results fromsimulations to explore its performance as a trigger underdifferent multiplicity environments.This work is organized as follows: in Sec. II we describethe overall detector concept. In Sec. III we present thedetails of the front-end electronics and in Sec. IV themechanical structure designed to support the sensitiveelements and the electronics. In Secs. V– VIII we showour simulation results to study the time resolution capa-bilities using p + p and A + A collisions. Sec. IX presentsresults of Monte Carlo simulations to estimate the intrin-sic time resolution for a basic cell. We summarize andconclude in Sec. X.
II. OVERALL DESCRIPTION
In order to achieve the fast trigger signal and low mate-rial budget requirements, the proposed geometry for theminiBeBe consists of 16 strips of length 600 mm. Eachstrip is made of an array of 20 squared plastic scintilla-tor cells with dimensions 20 × × . There are4 SiPMs coupled to each cell. The strips are supportedby a cylindrical, lightweight shell, whose inner and outerradii are 220 and 260 mm, respectively, measured fromthe symmetry axis of the beam pipe. The lengths of theinner and outer shell radii can vary depending on the fi-nal design of the shell that will be surrounding the beampipe during the first phase of the MPD. In total, theminiBeBe is made of 320 squared plastic scintillator cellsand 1,280 SiPMs covering an effective sensitive area of128,000 mm and a pseudo-rapidity range of | η | < . and a light decay constant of 1.8 ns. Thiskind of plastic scintillator can be used in vacuum envi-ronments. The wavelength of maximum emission is 408nm. Its softening point is at 70 ◦ C. The SiPMs make useof a recent technology for silicon semiconductors. Unlikeprevious semiconductor-based models, these devices havethe ability to resolve even a single photon. The selectedmodel for the miniBeBE detector is the MicroFC-60035SensL SiPM with dimensions 6 × manufactured bySensL Technologies, Ltd., with a cell length of 35 µ m, fora total of 18,980 cells distributed over its 6 × sur-face. This model has a high-gain and an ultra-fast signalfor timing applications with a rise time of 1 ns and a pulsewidth of 3.2 ns. The maximum of the Photon DetectionEfficiency (PDE) is typically reached for a wavelength of420 nm, and the gain range is from 20% to 41% [25]. III. FRONT-END ELECTRONICS
The main goal for the design and implementation ofthe front-end electronics is to generate trigger pulses forthe TOF, based on the detection of fast moving particles.For the output, only the fast signal is used since it has abetter timing response compared with the standard sig-nal. As recommended in Ref. [26], a voltage higher thanthe breaking voltage V br , given by V br + 5 V, was usedto maximize the SiPM PDE. The fast signal must have aload resistance in order to generate a current path to thereference ground. Hence, a 50 Ω resistor is used as anoutput load with an analog output signal. The schemat-ics of the 2 by 2 SiPMs array with a parallel configura-tion, and with a front-end electronics in a Printed CircuitBoard (PCB), is shown in Fig. 3, where V f stands for thevoltage of the fast output signal and I f for the outputcurrent flowing through the 50 Ω load resistor [27]. Asingle BC404 plastic scintillator is attached to the SiPMsand the PCB, resulting in a Scintillator and Front-EndDetector (SFED) for radiation detection. The PCB isdesigned on a Flame Retardant 4 (FR-4) material (thatcomplies with the NEMA UL94V standards) with a di-electric constant of 4.34 at 1 GHz. The output analogsignal is transformed into a digital differential signal byusing the analog comparator HMC674 [28] with an inputbandwidth of 9 . FIG. 1. Schematic illustration of the mechanical structure ofthe miniBeBe detector. The structure holds sixteen 600 mmlong strips mounted on a cylinder placed around the beampipe. The length of the cylinder is designed to detect colli-sions when the interaction does not take place at the nominalposition. Each strip consists of 20 squared plastic scintillatorswith dimensions 20 × × , with four SiPMs coupledto each cell. The length is 600 mm and the inner and outerradii are 220 and 260 mm, respectively. FIG. 2. MiniBeBe dimensions. The total length along thesymmetry axis (71 . (RSPECL) standard, described in Table I [28].Each SFED card is attached to a ribbon Rigid-Flexcard where the analog-to-digital conversion is performedvia one HMC674 per SFED card. In this way, the ribboncard consists of 10 analog comparators HMC674 witha corresponding miniature mezzanine connector, locatedin the rigid part of the strip card. Each pair of triggersignals is sent through this ribbon card to avoid cableand material excess. The SFED PCB and ribbon cardsare shown in Fig. 4.The ribbon card length is 600 mm, designed to support TABLE I. Voltage levels for the RSPECL standardParameter Symbol Min. Typ. Max. UnitsHigh level V OH V OL Ribbon PCBSiPM Card (SFED)FIG. 4. Illustration of the basic elements (ribbon PCB andSiPM card) of the front-end electronics. up to 10 SFED cards, each attached to a mezzanine microconnector and separated 30 mm from each other over alength of 500 mm, ending in a mini-D connector.In anticipation of the scenario that the length of theminiBeBe could be increased from 600 to 1000 mm, tworibbon cards are used, covering 1200 mm of total length.Therefore, a detector strip is composed of two ribboncards, containing 20 SFEDs and thus, generating 20 dif- ferential pair trigger signals available in two mini-D con-nectors, located on each side of the strip. The area cov-ered by the rigid sections of the ribbon card is 8,000 mm (20 SFEDs with an area of 20 ×
20 mm ).Since the miniBeBe consists of 16 strips, 320 radi-ation SFED detectors will give the corresponding dif-ferential pair signals. The area covered by the SFEDcard is 128,000 mm , corresponding to 15% of the totalminiBeBe cylinder area.All the trigger signals are collected using a TRB3 FieldProgrammable Gate Array (FPGA) card, controlled froma Linux computer to acquire and store up to 264 inputchannels of information in a data center [29]. Part ofthe signal processing task will be developed inside theFPGA card. Thus a single trigger for the TOF sensorwill be generated inside this FPGA card, achieving themain objective of this front-end. As described by thegeneral schematics shown in Fig. 5, a power supply bankis required with low ripple, high pass filtering and goodgrounding system to avoid interference and noise induc-tion to all the front-end design. The voltage and currentrequirements are specified in Table II. IV. MECHANICAL STRUCTURE ANDMATERIAL BUDGET
The mechanical structure consists of the main supportfor the plastic scintillators cells and for the readout elec-tronics. This is schematically shown in Fig. 6.
FIG. 5. General schematics of the miniBeBe front-end electronics.
The mechanical structure has been designed account-ing for the requirement of a low material budget, whichwas translated into a lightweight but at the same timefirm structure. This design considers an eventual 3Dprinting consisting of removable pieces allowing to even-tually assemble the essential parts, make decisions and adjust rigidity and precision for the overall structure.Figure 7 shows the estimated weight of the mechanicalsupport as a function of different density percentages of3D printing materials [30].The structure has been developed having in mind thePlug&Play concept and the possibility to replace the rails
TABLE II. Power supply requirementsParameter Symbol Min. Typ. Max. Units1 SFED voltage V SFED I SFED
80 100 120 mA1 Analog comparator power – – 140 – mW1 Analog comparator voltage V Acomp − . V TRB3 – 48 980 VTRB3 current I TRB3 – 10 – AFIG. 6. 3D model for the miniBeBe mechanical structureexhibiting its main components. that support the electronics and plastic scintillators atwill, without having to disassemble the whole structure.Deformation simulations of the structure parts wereperformed using finite element analysis with the
AutodeskInventor software, to approximate the behavior of thestructure under extreme conditions of temperature vari-ations and of differential pressure. Table III shows thevolume corresponding to each of the structure parts, asan indicator for finite element simulations. The designconsiders the mass of each integrated element within theminiBeBe structure.Figure 8 shows the “sensor rail” supporting one of theplastic scintillator strips. The rail is designed to hold astrip consisting of 20 cells of dimensions 20 × × ,each connected to its corresponding readout electronics.The rails are to be screwed to the support cross bars toprovide support, rigidity and stability. An example of arail support is shown in Fig. 9. The design also considerssimulations carried out within the MPDRoot [31] framefor a 16 strips cylindrical geometry. The whole structureis designed so that the detector cells are located 250 mm TABLE III. Volume of the miniBeBe support structure com-ponents used in our simulations.External flanges 1373094.2 mm Inner rings 556121.1 mm Cross bars 115860.1 mm Sensor rails 77380.0 mm Top cover for rails 44776.0 mm from the beam axis. Each rail is separated by 22.5 ◦ inthe transverse plane. The support has an external radiusof 260 mm and an internal radius of 220 mm. The lattercorresponds to the ring that supports the cross bars. Thecaps have a 60 mm internal radius, so that direct con-tact with the beam pipe is avoided. The whole cylinderconsists of two sections with a semicircular cross sectionon the transverse plane that can be clamped togetheraround the beam pipe. A schematic representation of oneof the cylinder halves, viewed from the transverse plane,is shown in Fig. 10, where the dimensions described abovecan also be seen.For tolerance tests and structural alignment of thecylinder, 3D prints were made at a density of 10% in Poly-lactic Acid (PLA) and NylonX (nylon reinforced withcarbon fiber) in order to obtain a prototype for manufac-turing in 100% carbon fiber using additive manufacturingtechnologies.The structure is designed for easy assembly. Each ofthe strips is individually assembled over the support railsand then placed on the cross bars to be later screwedtogether. This makes maintenance and replacement ofparts quick and easy.In order to estimate the possible effect on the energyof particles passing through the detector material, wehave also performed studies of the energy loss of primaryparticles (pions and muons) in the range of 5 MeV to5 GeV. To asses the effect of the different detector ma-terials, we perform the analysis both for the DetectorElement (DE) as well as for the blind area (BA). Theformer consists of the BC404 plastic scintillator to whichthe SiPMs are attached together with the Polyvinyl Chlo-ride (PVC) where the electronic circuits are printed. Thelatter is considered to be made of polyacrinolitrile. Thesimulation studies were made using the Geant4 software.The BA is taken to have a thickness of 6.56 mm whereasthe DE has a thickness of 4 mm. We find that there FIG. 7. Estimated weight as a function of the print densityfor Polylactic Acid (PLA), Acrylonitrile Butadiene Styrene(ABS) and NylonX (Carbon Fiber Reinforced Nylon) fila-ments.
FIG. 8. Sensor rail to hold 20 plastic scintillators with dimen-sions 20 × × .FIG. 9. Schematic representation of one of the support crossbars for the sensor rails. is no distinction between the energy deposit of the twoconsidered primary particles. For the EA, the energy de-posited per particle is in the range of 0.49 to 0.94 MeV,and for the BA it is in the range of 1.35 to 2.58 MeV.These findings are summarized in Table IV. V. ARRIVAL TIME DISTRIBUTION ATGENERATOR LEVEL USING FIRST ANDSECOND ARRIVALS IN P+P COLLISIONS
We use PYTHIA 8 with the SoftQCD module for p +p collisions at √ s NN = 10 and 11 GeV to generate 10 events. With this sample we study the time differencebetween first and second arrivals at transverse and longi-tudinal distances from the point of interaction given by r = 25 cm and | z | = 30 cm, respectively.In Fig. 11 we show the p T and η distribution for FIG. 10. Transverse plane view of the beam support for thesensor rails. The labels A – E refer to the elements depictedin Fig. 6. charged pions. We characterize the complete p + p sam-ples using distributions in p T and η for all charged parti-cles and focus on two charged particle sets — all andpions only — for the next part of the analysis (“all”charged particles are composed of about 82% pions, 5.5%kaons and 11% protons, with the remaining 1.5% beingleptons). This allows us to understand the characteristicenergy scales of the particles we will detect, as well asto visualize their expected η distributions. With this inmind, we have optimized the η coverage of the miniBeBe.When we discuss the charged particles produced in thep + p collisions, we characterize the samples with thedistributions of charged particles arrival time for differentgeometrical cuts. This has been used to optimize the sizeof the miniBeBe. In Fig. 12 we show the normalized yieldof charged pions with respect to their arrival time, in timebins of 10 ps. From top to bottom we show: the first andsecond arrival time, t and t , and their event-by-eventdifference ∆ t = t − t for charged pions produced inthe two different Center-of-Mass (CoM) energies in p +p collisions.The arrival time ( (cid:104) t (cid:105) ± σ ) of the first and secondcharged particles, as well as their difference for 10 GeVand 11 GeV, are summarized in Table V. For a 10 pstime resolution we have less than 10% standard deviation(spread) on the mean arrival time. We find that themean time for first arrivals for all charged particles isbetween 870 ps and 959 ps for the collision energies underconsideration. For all charged particles the mean ∆ t isin the range of 34.5 to 34.8 ps but the spread in thisdifference is up to 57.9 ps. Therefore, time differencesfor the all-charged selection are, at best, resolved in therange of 91.4 to 92.4 ps, depending on the collision energy.When we select charged pions, the mean time for firstand second arrivals is similar and stable with respect tothe all-charged selection; now the mean ∆ t is in the range45–46 ps and the time spread in this difference is up to70.0 ps. Therefore, time differences for the charged-pionselection are, at best, resolved in the range of 114.0 to115.8 ps, depending on the collision energy.To summarize, in order to produce a trigger signalwithin a time ∼
10 - 30 ps, it is necessary to use the
IE (GeV) E loss in DE (MeV) E loss in BA (MeV)0.05 0 . ± .
01 2 . ± . . ± .
07 1 . ± .
181 0 . ± .
01 1 . ± .
153 0 . ± .
06 1 . ± .
155 0 . ± .
06 1 . ± . loss ) of primary particles (pions andmuons) with a given Incident Energy (IE) over the DetectorElements (DE) and the Blind Area (BA) of the miniBeBe.The energy loss is negligible for the considered range of inci-dent energy and thus we expect that the material budget willnot affect the particle properties while passing through thedetector. √ s NN t [ps] t [ps] ∆ t [ps]10 GeV 958 . ± . . ± . . ± . . ± . . ± . . ± .
910 GeV 967 . ± . . ± . . ± . . ± . . ± . . ± . (cid:104) t (cid:105) ± σ ) of the first and sec-ond charged particles for p + p collisions, as well as theirdifference, obtained from PYTHIA 8 and SoftQCD module.The difference in time is exceeded by the standard deviation(spread) σ of individual measurements of arrivals. √ s NN t [ps] t [ps] ∆ t [ps]9 GeV 892 . ± . . ± . . ± . . ± . . ± . . ± .
99 GeV 869 . ± . . ± . . ± . . ± . . ± . . ± . (cid:104) t (cid:105) ± σ ) of the first and secondcharged particles for Au + Au collisions, as well as their dif-ference, obtained from UrQMD v.3.4. The difference in timeis exceeded by the standard deviation (spread) σ of individualmeasurements of arrivals. first arriving charged particles.We add that we have carried out the same analysis us-ing UrQMD as the event generator. The event-by-eventtime distribution shows slight differences compared tothe case hereby discussed using PYTHIA 8 as the eventgenerator. This suggests a systematic uncertainty wellbeyond the statistical errors, which are displayed for thepionic case in Fig. 12. We therefore emphasize that it be-comes important to carry out further systematic studiesusing different event generators to have a more completepicture of the kind of time distributions that can be ex-pected as these energies. VI. ARRIVAL TIME DISTRIBUTION ATGENERATOR LEVEL USING FIRST ANDSECOND ARRIVALS IN AU + AU COLLISIONS
We use UrQMD v3.4 for Au + Au collisions at √ s NN = 9 and 11 GeV to generate 10 events and weanalyze the time difference between first and second ar-rivals at transverse and longitudinal distances from thepoint of interaction given by r = 25 cm and | z | = 30 cm,respectively.In Fig. 13 we characterize the complete Au + Au sam-ples using charged pion distributions in p T and η . Again,this allows us to understand the characteristic energyscales of the pions we will detect, as well as to visual-ize the expected η distribution. This is one of the inputswe have used to optimize the η coverage of the miniBeBe.Figure 14 shows the normalized yield of charged pionswith respect to their arrival time at r = 25 cm and | z | =30 cm, in time bins of 10 ps. From top to bottom weshow: the first and second arrival time, t and t , and FIG. 11. Charged pions p T and η distributions in p + pcollisions at √ s NN = 10 and 11 GeV for 10 events generatedwith PYTHIA 8 with the SoftQCD module. their event-by-event difference ∆ t = t − t , for chargedpions generated at two different CoM energies.We find that the mean time for first arrivals for allcharged particles is between 885 ps and 893 ps in Au +Au collisions for different CoM energies. The mean valuesof the arrival times at both energies under considerationare summarized in Table VI.When we select charged pions, the mean time for firstand second arrivals is similar to the all-charged sample;now the mean ∆ t is in the range 11–29 ps and the timespread in this difference is up to 61.8 ps. Therefore, timedifferences for the charged pion selection are, at best,resolved in the range of 56.2 to 90.8 ps, depending on thecollision energy.To summarize, in Au + Au collisions at √ s NN = 9and 11 GeV, for a wake-up detector to produce a signal FIG. 12. Charged pions normalized yield in time bins of 10 psusing PYTHIA 8 with the SoftQCD module for p + p withgeometry cuts implemented as | z | = 30 cm and r = 25 cm.From top to bottom we show: the first and second arrivaltime, t and t , and their difference ∆ t = t − t for chargedpions produced in p + p collisions at √ s NN = 10 and 11 GeV. FIG. 13. Charged pions p T and η distributions for Au + Aucollisions at √ s NN = 9 and 11 GeV for 10 events generatedwith UrQMD v.3.4. within a time ∼
10 - 30 ps, it is necessary to use thefirst arriving charged particles, as is the case for p + pcollisions.We now proceed to perform this analysis, using the im-plementation of the miniBeBe in the MPDRoot frame-work, where the conditions of the experiment (magneticfield, occupancy, etc.) will significantly enlarge the timedifferences between first and second arrivals. This is re-ported in the next section.
FIG. 14. Charged pions normalized yield in time bins of 10 psusing UrQMD v3.4 for Au + Au with geometry cuts imple-mented as | z | = 0 . r = 0 .
25 m . From top to bottomwe show: the first and second arrival time, t and t , andtheir difference ∆ t = t − t , for charged pions produced inAu + Au collisions at √ s NN = 9 and 11 GeV. FIG. 15. Geometry of the miniBeBe as simulated within MP-DRoot and rendered by the Event Display. Sixteen stripsare arranged surrounding the interaction point of the MPD.Each strip consists of 20 squared plastic scintillators of size20 × × made of BC404. The length of the simulatedsensitive area is 60 cm and its diameter is 50 cm. VII. IMPLEMENTATION OF THE MINIBEBEGEOMETRY IN THE MPDROOT FRAMEWORK:HITS, ENERGY DEPOSIT ANDTIME-OF-FLIGHT
Using the official offline framework of the MPD, MP-DRoot, we simulated the miniBeBe under the specifi-cations described in Sec. II. Figure 15 shows the EventDisplay of the miniBeBe so that we confirm that MPD-Root has the geometry implemented as per design. Inorder to test the implementation of the miniBeBe inthe MPDRoot framework, we performed simulations of950,000 Minimum Bias (MB) events (impact parameter b = 0 − . √ s NN = 9GeV and another 950,000 MB events ( b = 0 − . √ s NN = 11 GeV, usingUrQMD [32, 33].First we report on the tracks selected in the geometri-cal acceptance of the miniBeBe and study the energy ofparticles hitting the detector cells, in order to comparewith energy deposited when we include the material. Weperform a geometrical selection of the miniBeBe cells oftracks (MCTracks within MPDRoot) as shown for Bi +Bi collisions at 9 GeV in Fig. 16 with the hits in space(top) and with the η distribution of all charged particlesand primaries (bottom), where we can verify that in-deed the acceptance of the miniBeBe occurs at | η | < . FIG. 16. Geometrical selection of the miniBeBe cells of tracks(MCTracks within MPDRoot) using 5000 events for Bi + Bicollisions at 9 GeV shown as hits in space (top) and the η distribution of all charged particles and primaries (bot-tom), where we can verify that indeed the acceptance of theminiBeBe occurs at | η | < . axis. Note that there is a band regularity correspondingto the cells per strip that is reflected in the next part ofthis analysis. Note also that if we compare the energyscale of charged particles given by Fig. 17 and the scalefor energy deposit in the miniBeBe in Fig. 18, we cansee that most charged particles leave far less than 1% oftheir energy in the miniBeBe. FIG. 17. Scatter plot distribution of particles with respectto the energy they carry at generation level within MCTrackswhen they reach the miniBeBe (top) and identified particledistributions (bottom), normalized to the number of eventsfor the Bi + Bi at 9 GeV sample.
The scatter plots serve as a test of the coverage of thecells in a strip and shows the uniformity of the coverage.Now we can extract useful information to characterizethe miniBeBe using strip-averages for the relevant quan-tities. We perform a strip-average where we quantify theaverage number of hits, the energy deposit and the time-of-flight of all hits averaged per cell in a strip. Since eachstrip has 20 cells, we use the notation for evenly-spaced1
FIG. 18. Scatter plots for the hits in the miniBeBe for theMB sample of Bi+Bi at 9 GeV. The upper panel shows theenergy deposited and lower panel the time-of-flight for all hits.Given our convention to label the cells, the maximum of theenergy deposit and the minimum time-of-flight happen forcells labeled by multiples of 10. cells 1 through 20 to refer to their location from z = − z = +30 cm.Figure 19 shows the strip-average number of hits percell along a miniBeBe strip for both Bi+Bi at 9 GeVand Au+Au at 11 GeV. Both panels include the MB andthe centrality classes results. We notice that for bothsamples, we have on average almost 3 hits per strip inthe most central collisions, down to 1 hit per strip inthe semi-central collisions and no-hit on average for theperipheral collisions. Considering that the miniBeBe has16 strips, we expect the highest miniBeBe efficiency at FIG. 19. Strip-average of the number of hits per cell for theminiBeBe in Bi+Bi collisions at 9 GeV (top) and Au+Au at11 GeV (bottom). We show results for the MB ( b = 0 − . around 48 hits per event for central collisions, irrespectiveof the species or energy of the collision.Figure 20 shows the strip-average energy deposited byhits per cell along a miniBeBe strip for both collisions,Bi+Bi at 9 GeV and Au+Au at 11 GeV. Both panelsshow the MB and the centrality classes results. We no-tice that for both samples, we have an average energydeposited per miniBeBe cell of at most 0.8 MeV for allcentrality classes. So we expect the miniBeBe to with-stand, on average, 16 MeV of energy deposited per strip.Figure 21 shows the average time-of-flight per hit percell along a miniBeBe strip, again for both collisions.Both panels show the MB and the centrality classes re-sults. We notice that for the central miniBeBe cells(around z = 0) we have an average below 1.3 ns, forall the analyzed samples. For the Bi+Bi at 9 GeV sam-ple we can reach time-of-flight averages of (slightly) lessthan 1.1 ns. This sets the benchmark analysis for thetrigger capabilities of the miniBeBe in the next section,where we compare leading time vs. average time results.Notice also that peripheral heavy-ion collisions shouldbe comparable to p + p collisions. For completeness,Fig. 22 shows the average hits, energy loss and time-2 FIG. 20. Strip-average of the energy deposit per cell for theminiBeBe in Bi+Bi at 9 GeV (top) and Au+Au at 11 GeV(bottom) collisions. We show results for the MB ( b = 0 − . (cid:104) Hits (cid:105) strips 0-20% 80-100%per 16 0.2294 - 0.3248 0.0042 - 0.0047Bi + Bi cell 32 0.4588 - 0.6501 0.0083 - 0.00949 GeV complete 16 73.40 - 103.94 1.34 - 1.50detector 32 293.63 - 416.06 5.31 - 6.02UrQMD (cid:104)
Hits (cid:105) strips 4 GeV 11 GeVper 16 0.00043 - 0.00055 0.00100 - 0.00122p + p cell 32 0.00084 - 0.00106 0.00199 - 0.00245complete 16 0.138 - 0.176 0.320 - 0.390detector 32 0.538 - 0.678 1.274 - 1.568TABLE VII. Overview of average number of hits in theminiBeBe as shown in Figs. 19, 22, 23 and 24. For Bi+ Bi at √ s NN = 9 GeV and p + p at √ s NN = 4 and 11GeV, we report the range of average number of hits per celland for the complete detector. We show both the 16 and 32miniBeBe geometry results and note that, as expected, thelatter doubles the average number of hits per cell. Since eachstrip has 20 cells, the complete detector average hit rangeis obtained with a factor of 20 ×
16 and 20 ×
32, for eachgeometry. FIG. 21. Strip-average of the time-of-flight per cell forthe miniBeBe in Bi+Bi at 9 GeV (top) and Au+Au at11 GeV (bottom) collisions. We show results for the MB( b = 0 − . of-flight in the miniBeBe using 950,000 p + p collisionevents at √ s = 4 , ,
11 GeV that we generated usingUrQMD and MPDRoot. We notice that even though theaverage number of charged particles in p + p is well be-low that of A + A collisions, they deposit more energyin the detector. Overall, we have a similar scale of en-ergy deposit per cell in p + p and in A + A collisions,so our findings are summarized as follows: for both sim-ulations using Bi + Bi at √ s NN = 9 GeV and Au +Au at √ s NN = 11 GeV, we have shown that the averagenumber of hits, the average energy deposited and the av-erage time-of-flight per design geometry of the miniBeBe,happens within an average time-of-flight between 1.1 and1.6 ns. Moreover the length of the detector covers the re-gion with the highest average hits per event with no morethan 16 MeV of energy deposited per strip. We have alsoverified that the miniBeBe has a small occupancy withenergy deposit of charged particles.To conclude this section, we comment on possible andimmediate improvements for the miniBeBe design, thatstill conform to current space availability in MPD, butthat are contingent upon further financial support.3 FIG. 22. Strip-average of the number of hits (top), energydeposit (middle) and time-of-flight (bottom) per cell for theminiBeBe in p + p collisions at 4, 9 or 11 GeV.
In Figs. 23 and 24 we show the expected improvementof the average number of hits in the miniBeBe when dou-bling the number of strips. We use 5 × events for Bi +Bi collisions at √ s NN = 9 GeV generated with UrQMDand for p + p at √ s NN = 4, 9, 11 GeV, transported withMPDRoot through an upgraded miniBeBe that now has32 strips. In Table VII we summarize our findings forthe average number of hits per cell and for the completedetector, in comparison with the 16-strip design. As ex-pected, the average number of hits per cell doubles whenproceeding from the 16-strip to the 32-strip design. Sinceeach strip has 20 cells, the complete detector average hitrange is obtained with a factor of 20 ×
16 and 20 × FIG. 23. Strip-average for the upgraded geometry with 32strips, of the number of hits (top), energy deposit (middle)and time-of-flight (bottom) per cell for the miniBeBe in Bi+Bicollisions at √ s NN = 9 GeV. VIII. SIMULATIONS FOR THE MINIBEBE:TRIGGER CAPABILITIES
We used UrQMD [32, 33] for Bi+Bi collisions andbeam-gas interactions. For Bi+Bi collisions a sample of9,000 MB events with a centrality range between 0 and90% was generated. For beam-gas interactions we sim-ulated p+O collisions at √ s NN = 9 GeV with a vertexposition at ±
19 m along the z − axis and a width of ± . . c and c .The simulation was done to evaluate the trigger capa-bilities of the miniBeBe for heavy-ion collisions as well4 FIG. 24. Strip-average for the upgraded geometry with 32strips, of the number of hits (top), energy deposit (middle)and time-of-flight (bottom) per cell for the miniBeBe in p+pcollisions at 4, 9 and 11 GeV. as to be used as a beam-gas interactions veto. Triggerefficiencies for miniBeBe have been obtained for Bi+Bicollisions at √ s NN = 9 GeV. Figure 25 shows the trig-ger efficiency considering that at least one charged pionhits the miniBeBe. For low charged particle multiplic-ity events ( (cid:46)
60 charged particles), the miniBeBe triggerefficiency is less than 60%. This behavior is due to theforward events that UrQMD generates, with few chargedpions produced in the central barrel region. If we consideronly events with charged particles within the miniBeBedetector acceptance ( | η | < . (cid:39) FIG. 25. MiniBeBe trigger efficiency as a function of thecharged particle multiplicity (top) and pseudo-rapidity (bot-tom).
A. Multiplicity
At this level of development of the miniBeBe in theMPDRoot frame, we decided to use the information ofthe physical interaction of particles produced in heavy-ion collisions at NICA energies using the volume of theminiBeBe which is sensitive to hits. Hits in the miniBeBeare produced when a Monte Carlo track enters into theactive sensitive volume, without any deposited energy re-striction. This is the standard definition of a hit in MPD-Root. The simplest information that we can extract fromminiBeBe simulations is the number of hits per event andits corresponding time information. In this case, we as-sume that the number of hits in the miniBeBe can betaken as a raw multiplicity.Figure 26 shows a (roughly linear) relation between thenumber of hits produced in the miniBeBe and the num-ber of generated charged particles. This result is useful ifwe intend to produce an online centrality trigger with theminiBeBe. As shown in Fig. 27, the miniBeBe raw mul-tiplicity varies with respect to different centrality ranges.This behavior has been reported at higher energies inRef. [34] where it is explained in terms of the geomet-rical properties of a heavy-ion collisions. Some eventsmay be assigned to a wrong centrality range. This effectcan be corrected offline during the data analysis or datareconstruction.5
FIG. 26. Number of charged particles that hit the miniBeBevs. the generated number of charged particles.FIG. 27. MiniBeBe multiplicity per centrality range.
B. Time information
The arrival time of the produced charged particles atindividual cells was taken from the generated hit man-aged by the
MbbPoint class available in MPDRoot. Fromthe time information of the miniBeBe hits per event,we estimated the average hit time and the time-of-flightof the first charged particle reaching miniBeBe (leadingtime) for z > t right , and for z < t left . The root meansquare (RMS) of the ∆ t = t right − t left distribution pro-vides an indication of the target for best time resolutionof the miniBeBe. Figure 28 shows the RMS of the ∆ t distribution as a function of several time windows: 3 ns,10 ns, 20 ns, 35 ns and 70 ns where in each case we as-sumed that both the average and leading times, for z > z <
0, are less than these time windows.As example, Fig. 29 shows the ∆ t distribution for theaverage and leading time of the miniBeBe. It can benoted from Figs. 28 and 29 that the minimum RMS valuefor the ∆ t distributions is obtained using the leading timefor particles reaching miniBeBe.The RMS value of the ∆ t distribution depends also on FIG. 28. MiniBeBe RMS time difference t right − t left as afunction of the time window. the collision impact parameter b . The lowest RMS valueof the ∆ t distribution is obtained for central collisions,while for larger values of b the RMS value is 0.815 ns,as can be seen in Fig. 30. Thus, a time resolution of atleast 0.026 ns is mandatory for the miniBeBe to generatea proper beam-beam trigger signal based on the leadingtime measured by the miniBeBe data acquisition system.Using the leading time of miniBeBe hits, t right and t left for z > z <
0, respectively, we can determine withthe miniBeBe the collision vertex along the z -axis as V ertexM bb = t right − t left × c .To estimate the resolution of the vertex determinationof the miniBeBe, we computed the RMS of the ∆ vtx = V ertexGen − V ertexM bb distribution, where
V ertexGen is the generated position of the collision vertex given bythe UrQMD generator. Figure 31 shows the ∆ vtx distri-
FIG. 29. Number of entries vs. ∆ t for a time window of 3ns calculated using the leading time (continuous line) and theaverage time (dotted line). FIG. 30. RMS of the ∆ t distribution as a function of theimpact parameter b of the collision. bution. The best time resolution for the vertex determi-nation using the miniBeBe is 24 cm/ c = 0 . C. Beam-gas
Beam-gas interactions are a background originated at acertain distance from the interaction point due to the in-teraction of the circulating particles in the beam with theresidual gas in the beam pipe. This background dependson the NICA nominal bunch crossing. To simulate beam-gas events we generated p+O collisions with UrQMD at √ s NN = 9 GeV with the collision vertex located at +19m from the nominal interaction point, with a width of3.5 m.In order to evaluate the miniBeBe capability to sepa-rate beam-gas interaction events from beam-beam colli-sions, we used the leading time distribution t right + t left FIG. 31. The distribution of ∆ vtx , defined at the end ofSec. VIII C. We show the difference between the generatedvertex and the vertex determined with the leading time of theminiBeBe detector. FIG. 32. The distribution of t right + t left . The sum of theleading time of the miniBeBe detector for z > z < for beam-beam and beam-gas generated events. If thebeam-gas interaction vertex events is located 19 m awayfrom the interaction point, the miniBeBe may be ableto discriminate beam-gas interactions from beam-beamcollisions. Some beam-beam events at the tail of the t right + t left distribution may be mistaken with beam-gasinteractions and vice versa. As the location of beam-gas events is moved closer to the interaction point, theminiBeBe decreases its capability to veto beam-gas in-teractions, see Fig. 32. (No correction due to fine tuningcabling delay, neither time spread of the collision nor indi-vidual plastic scintillator cell time resolution, was appliedto this analysis.) D. Summary of findings from MC simulations
The results shown in this section can be summarizedas follows: • The miniBeBe can generate a trigger signal forbeam-beam collision events with a (cid:39) • The miniBeBe leading time is optimal to generatetrigger signals. • A miniBeBe time resolution of 26 ps is needed totrigger central collision events. For non-central col-lisions, a not so stringent time resolution of only 85ps is required. The miniBeBe will be able to pro-vide a trigger signal with these requirements. • The miniBeBe will be able to distinguish beam-gas interactions from beam-beam collisions if thevertex location of beam-gas events is far from the7interaction point ( (cid:38)
19 m). If the location of beam-gas vertex interactions is closer to the interactionpoint, the miniBeBe will become less efficient to setproper trigger flags to distinguish beam-beam frombeam-gas events. • The number of hits in the miniBeBe seems to besensitive to the centrality of the collision. This in-formation may be useful to generate online central-ity trigger classes.
IX. SIMULATIONS TO ESTIMATE THEINTRINSIC TIME RESOLUTION FOR A BASICCELL
In order to study the intrinsic time resolution for thebasic elements of the miniBeBe, we performed simula-tions using Geant4 v10.06p01 [35]. The configurationswe study consist of arrays of one, two, three and fourSiPMs of size 3 × distributed on the surface of theplastic scintillator cells. The intrinsic time resolution isstudied without including the contribution from the elec-tronic output. The different configurations we considerare depicted in Fig. 33, where the black squares (scor-ers) represent the SiPMs. The goal was to explore theconfiguration that provides the minimal time resolution.This is carried out considering also two kinds of plasticscintillators: BC404 and BC422.We simulated 1000 π + -mesons, arriving one by oneat the cell where they hit the full frontal area, on theopposite face of the one where the scorers are located.The π + are given an average kinetic energy of 0.5 GeV,which corresponds to their typical energy for A + A col-lisions at NICA energies. For each event, we recorded thelowest mean of the Landau time of flight distribution ob-tained in any one of the scorers. This time represents thefirst pulse. For the BC404 plastic scintillator, our resultsimply an intrinsic time resolution of 7 . ± .
87 ps and9 . ± .
67 ps, for one and four scorers, respectively. Forthe BC422 plastic scintillator we obtained 7 . ± .
87 psand 9 . ± .
75 ps, for one and four scorers, respectively.However, these differences of up to 2 ps are not signifi-cant in light of the fact that the electronics has only atime resolution of about 20 ps [29]. In this sense, thetime resolution is equivalent for all scorer configurationsand both scintillator materials.Figure 34 shows the distribution for the case of 4 scor-ers. The two peaks are due to the randomly distributedincidences all over the cell area; the same pattern is ob-served when working with the other configurations. Tounderstand this effect, we performed further simulations,in which the beam hits one specific point of the scintil-lator. Figure 35 illustrates this scenario for the exampleof the time of flight distribution for the interaction inone point on top of the frontal scintillator area. Fig-ure 36 shows the corresponding distribution for the casewhen only one scorer is simulated. The interaction in theperimeter leads to a time resolution around 2.6 ps, and
FIG. 33. Illustration of the four scorer configurations that wesimulated in order to identify the one with the optimal timeresolution. the central interaction yields a time resolution around26 ps. We repeated this analysis for the other configu-rations which also led to approximately Gaussian peaks.Again, the interval of the time resolution is equivalent forall cases, due to the significantly coarser resolution of theelectronics. These results suggest that central interac-tions are inappropriate to obtain a lower time resolution.We conclude that all the configurations and both ma-terials are equivalent with an average value around 8 psfor interactions all over the frontal area. Albeit this timeis expected to be sensitive to the location of the inter-action point. We do not observe appreciable differencesbetween the time resolution obtained for each configura-tion. The difference is visible, however, when consideringthe photon arrival time: for the case of one scorer thistime is between 60–192 ps, decreasing to the interval 30–60 ps for the case of four scorers. Therefore, we inferthat the configuration with 4 scorers provides the bestintrinsic time resolution.We also notice that if use was made of a SiPM with alarger effective area, for example one with a 6 × area, the intrinsic time resolution would remain essen-tially the same. Any possible improvement would beof the order of a few ps. Hence our results for variousarrangements of scorers would not show any significantimprovement for the cell intrinsic time resolution. X. SUMMARY AND CONCLUSIONS
In this work we have presented the conceptual designfor the miniBeBe detector that is proposed to be installedin NICA-MPD to serve as a level-0 trigger for the TOF.We have described the detector sensitive elements andthe read-out electronics. We have performed simulations8
FIG. 34. Time of flight distribution for photons produced bythe plastic scintillator. We show the results for configurationD in Fig. 33. to show that the design is capable to provide an effi-cient trigger for low and high multiplicity events. TheminiBeBe capabilities to additionally serve as a beam-gas veto as well as to determine the beam-beam vertexare also shown. The prototype of some of its parts iscurrently being developed and will soon be tested in aradiation hard environment.It is important to mention that, as it usually happenswith any other detector concept, the current design isevolving to better suit the needs of the MPD as a whole.These needs are now being discussed within the collabo-ration which may result in a scaling up of the design. Themodifications include a larger longitudinal dimension aswell as an increase of the number of sensitive elements inthe azimuthal direction. Nevertheless, it is important tobare in mind that all the simulations that were performedfor the dimensions hereby discussed still stand and thata larger number of sensitive elements can only increasethe detector capabilities. Also, the mechanical integra-tion with the support is being actively explored as wellas the integration with other MPD subsystems. More-over, some of the capabilities of the miniBeBe could be
FIG. 35. Time of flight distribution for photons producedby the plastic scintillator when the beam hits a specific pointlocated on top of the frontal scintillator area. We show resultsfor configuration D in Fig. 33. FIG. 36. Time of flight distribution for the top interactionpoint, as in Fig. 35, but with only one scorer. enhanced if used together with the BeBe detector that wehave also proposed to be considered as a forward beam-beam counter [36]. The technical design for the detectorwill be reported in a more detailed document elsewhere.
ACKNOWLEDGEMENTS
The authors thank Adam Kisiel, Marcin Bielewicz,Slava Golovatyuk and Itzhak Tserruya for very usefulcomments and suggestions. We also acknowledge theMPD Collaboration for facilitating the use of the MPD-Root framework to carry out the simulations and for thecontinuous interaction and feedback. I.M. thanks theICN-UNAM faculty and staff for support and kind hos-pitality provided during the development of part of thiswork. L.R. thanks the BUAP Medical Physics and El-ementary Particles Labs and their faculty for their kindhospitality and support during the development of part ofthis work. The authors are in debt to Luciano D´ıaz andEnrique Murrieta for their technical support. Financialsupport for this work has been received from UNAM-DGAPA-PAPIIT grant number IG100219 and from theConsejo Nacional de Ciencia y Tecnolog´ıa (CONACyT),grant numbers A1-S-7655 and A1-S-16215. I.M. acknowl-edges support from a postdoctoral fellowship granted byCONACyT. M.R.C. thankfully acknowledges the permis-9sion to use computer resources, the technical advise andthe support provided by Laboratorio Nacional de Su- perc´omputo del Sureste de M´exico (LNS), a member ofthe CONACyT national network of laboratories, with re-sources from grant number 53/2017. [1] The Technical Design Report for the MPD sys-tems, in particular for the FFD, can be found in:http://mpd.jinr.ru/doc/mpd-tdr/[2] F. Simon, Nucl. Instrum. Meth. A C10003–C10003 (2011).[4] M. Pizzichemi, A. Polesel, G. Stringhini, S. Gundacker,P. Lecoq, S. Tavernier, M. Paganoni, and E. Auffray,Phys. Med. Biol. P10016 (2016).[8] T. Cervi, M. Babicz, M. Bonesini, A. Falcone, A. Mene-golli, G. Raselli, M. Rossella, and M. Torti, Nucl. In-strum. Meth. A, , 209–212 (2018).[9] E. Lamprou, A. J. Gonzalez, F. Sanchez, and J. M. Ben-lloch, Phys. Medica , 10–18 (2020).[10] N. Otte, B. Dolgoshein, J. Hose, S. Klemin, E. Lorenz,R. Mirzoyan, E. Popova, and M. Teshima, Nucl. Phys. B(Proc. Suppl.) , 417–420 (2006).[11] L. Raczy´nski, “37th Annual International Conference ofthe IEEE Engineering in Medicine and Biology Society(EMBC)” 2784–2787 (2015).[12] W. Krzemien, D. Alfs, T. Bednarski, P. Bia(cid:32)las, E. Cz-erwi´nski, K. Dulski et al. , “IEEE Nuclear Science Sym-posium and Medical Imaging Conference (NSS/MIC)”,1–2 (2015).[13] E. E. Ermis and C. Celiktas, Pramana,
1, (2020).[14] A. Wieczorek, K. Dulski, S. Nied´zwiecki, D. Alfs, P.Bia(cid:32)las et al. , PLOS ONE , e0186728 (2017).[15] W. H. Trzaska, Nucl. Instrum. Meth. A et al. ,Nucl. Instrum. Meth. A
P08014–P08014 (2018). [19] K. E. Kuper, V. P. Oleynikov, V. V. Porosev, G. A. Savi-nov, and W. Drozdowski, J. Instrum.
P01001–P01001 (2017).[20] J. Yu, Z. Wei, M. Fang, Z. Zhang, C. Cheng et al. , Nucl.Instrum. Meth. A et al ., IEEE Transactions on NuclearScience
5, 2657–2666 (2014).[24] R. Preston, J. Jakubek, D. Prokopovich, and J. Uher, J.Instrum. et al. , Nucl. Instrum. Meth. A :C12043–C12043 (2013).[30] Estimate calculated with ideaMaker
3D printing software3.5.2.[31] https://git.jinr.ru/nica/mpdroot[32] S. A. Bass, M. Belkacem, M. Bleicher, M. Brandstetter,L. Bravina et al. , Prog. Part. Nucl. Phys. , 225-370(1998).[33] M. Bleicher, E. Zabrodin, C. Spieles, S.A. Bass, C. Ernst et al. , J. Phys. G: Nucl. Part. Phys. , 1859-1896 (1999).[34] B. Abelev et al. (ALICE Collaboration), Phys. Rev. C , 044909 (2013).[35] J. Allison et al. , IEEE Transactions on Nuclear Science,
1, 270-278 ( 2006).[36] M. Alvarado, A. Ayala, M. A. Ayala-Torres, W. Bieten-holz, I. Dom´ınguez et al. , Nucl. Instrum. Meth. A953