Scintillating sampling ECAL technology for the LHCb ECAL Upgrade II

The aim of the LHCb Upgrade II is to be able to operate at a luminosity of 1.5×1034 cm-2 s-1 to collect a data set of 300 fb-1. The required substantial modifications of the current LHCb electromagnetic calorimeter due to high radiation doses in the central region and increased particle densities are referred to as LHCb ECAL Upgrade II. A consolidation of the ECAL already during the long shutdown 3 will reduce the occupancy and mitigate the effects of substantial ageing in the central region after Run 3. Several scintillating sampling ECAL technologies are being investigated in an ongoing R&D campaign: Spaghetti Calorimeter (SpaCal) with garnet scintillating crystals and tungsten absorber, SpaCal with scintillating plastic fibres and tungsten or lead absorber, and Shashlik with polystyrene tiles, lead absorber and fast WLS fibres. Timing capabilities with tens of picoseconds precision for neutral electromagnetic particles and increased granularity with a denser absorber in the central region are needed for pile-up mitigation. Time resolutions of better than 20 ps at high energy were observed in test beam measurements of prototype SpaCal and Shashlik modules. Energy resolutions with sampling contributions of about 10%/√E, in line with the requirements, were observed.


Introduction
LHCb is an experiment dedicated to the study of heavy flavor physics at the Large Hadron Collider (LHC) [1].Its principal objective involves the search for indirect manifestations of new physics through the observation of CP violation and rare decays in beauty and charm hadrons.One third of the decay products from heavy flavor processes consists of  0 mesons or other neutral particles undergoing decay into photons across a broad energy spectrum ranging from a few to several hundred GeV.Then, the necessity of a high-performance electromagnetic calorimeter within the experiment is evident.The selection of a scintillating sampling calorimeter arises from a balanced compromise, encompassing factors such as energy resolution, compact representation of electromagnetic showers, rapid temporal response, and a reasonably moderate cost [2,3].

Current LHCb ECAL and motivation to upgrade
The existing ECAL is constructed with Shashlik-type modules with transverse dimensions measuring 12×12 cm 2 .These modules are built alternating layers of 2 mm thick lead tiles and 4 mm scintillator tiles.Scintillation light is gathered and conveyed by wavelength-shifting fibres to photomultiplier tubes (PMT) affixed at the rear of each module.The resultant assembly yields a Molière radius of 36 mm and a depth of 25 radiation lengthes per 415 mm, delivering a module energy resolution of ()/ = 10%/ √  + 1% in the energy range between few GeV to 100 GeV and with an instantaneous luminosity up to 2.0×10 32 cm −2 s −1 .To maintain occupancy below 10%, the calorimeter is compartmentalized into -1 -  three regions with distinct granularities.Granularity is defined by internally segmenting the scintillating tiles, and collecting the light per tower of tiles or cells to form modules with cell sizes of 4×4 cm 2 for the inner region, 6×6 cm 2 for the middle region, and 12×12 cm 2 for the outer region.
A crucial parameter of the ECAL is its radiation tolerance.The innermost ECAL region is estimated to accumulate a dose of 0.6 kGy per fb −1 at the front and rear planes and 1.2 kGy per fb −1 at the position of the shower maximum [4].These dose values have been estimated from FLUKA simulations [5], which are approximately 30% more conservative when compared to measurements obtained from strategically positioned dosimeters on the detector.In-depth studies on the radiation tolerance of Shashlik modules have been conducted through irradiation campaigns [6], concluding that 40 kGy is their operational limit.Modules in the innermost region are projected to reach this limit by the long shutdown 3 of LHC (LS3), whichi will begin in 2025.As the ECAL is expected to maintain sustained performance in Run 3, only the ECAL electronics have been replaced during the long shutdown 2 (LS2) to align with the 40 MHz readout of Upgrade I [7].
In the forthcoming High-Luminosity (HL) phase of the LHC, the LHCb is expected to work at an instantaneous luminosity of 1-2 × 10 34 cm −2 s −1 .A significant surge of particle density is anticipated particularly in the in proximity to the beam-pipe at lower angles for ECAL [8] where the radiation doses are expected to reach up to 1 MGy at 300 fb −1 (figure 1).In response to this, the existing modules necessitate a comprehensive redesign incorporating radiation-hard materials capable of withstanding these conditions.The revised design (figure 2) seeks granularity as fine as 15 mm (while respecting the outer dimensions of the current modules of 12 × 12 cm 2 ), alongside the provision of precise timing information within the range of a few tens of picoseconds to accommodate high-energy electromagnetic showers.Furthermore, the goal is to maintain the current energy resolution levels, characterized by a 10% sampling term and a 1% constant term [9,10].
The proposed solution to address these requirements is a spaghetti calorimeter (SpaCal).A SpaCal is a sampling calorimeter in which scintillating fibres are inserted into a dense absorber.The scintillating fibres convert the deposited energy into light and transport it to the photodetectors, avoiding wavelength-shifters which entail a reduction in light collection efficiency.At the same time, the electromagnetic shower dimension can be tuned by selecting absorber materials with adequate radiation length and Molière radius.
-2 -  3 Technologies for the Upgrade II

SpaCal technology for inner region
For the upcoming LS3 enhancement, we have selected to emploi the SpaCal technology.Using this technology, the Molière radius and the energy resolution can be configured by design to meet the specifications.Additionally, as the scintillating fibres also transport the light, there is no need of using a wavelength shifter, which is not available with high radiation hardness (figure 3).
Table 1 provides an overview of distinct SpaCal module types slated for application in future upgrades of the LHCb ECAL.The determination of cell sizes is contingent upon anticipated occupancies, with Molière radii tailored through the strategic selection of materials and geometry to harmonize with the specified cell sizes.This optimization yields position resolutions of the order of 1 mm.Furthermore, the lengths of the modules are configured to correspond to a minimum of 25 radiation lengths, thus ensuring a negligible longitudinal leakage effects.The transverse dimensions of all SpaCal modules mirror those of the current ECAL Shashlik modules, measuring 12 × 12 cm 2 .
The LHCb ECAL Upgrade II SpaCal technologies for inner region include two module types.The innermost modules with radiation-hard crystal scintillators and a tungsten absorber (1.5 × 1.5 cm 2 cell size), and the 40-200 kGy region with scintillating plastic fibres and lead absorber (3 × 3 cm 2 cell size) [11].

Shashlik technology
In the outer layers of the ECAL Upgrade II, approximately 3300 novel Shashlik modules will be deployed, each of them aiming to provide precise timing information at the scale of a few tens of picoseconds.The current proposal encompasses the incorporation of these modified Shashlik modules, featuring enhanced timing capabilities and a double-sided readout configuration (refer to figure 4).

LS3 enhancement
The ongoing plans for the LS3 enhancement entail the utilization of SpaCal technology with tungsten absorbers for the innermost modules (table 1).These modules are equipped with scintillating plastic fibers configured for a 2 × 2 cm 2 cell size and single-side readout.This strategic design is intended to offset the rapid degradation experienced by the innermost channels.
Given the much increased occupancy expected in LHC Run 4, the number of zones with different granularity (section 2) have been increased from three, in the present ECAL, to five (figure 2) in order to optimise the cost-to-efficiency performance.Furthermore, the modules of the two inner zones use the new SpaCal technology, thus profiting from the higher radiation tolerance as well as from a reduced transverse shower profile that has been matched to the cell size.
After LS3, assuming a luminosity value of 2×10 33 cm −2 s −1 and using occupancy data derived from a meticulous simulation, which incorporates the hadronic component, substantial occupancies are expected in the inner regions with present ECAL channel distribution (figure 5

Research and test beam results
The evaluation of different SpaCal module prototypes has been conducted through test beams at DESY and SPS, covering energy ranges of 1 to 5 GeV and 20 to 100 GeV, respectively [12].In the subsequent discussion, we will not only address the performance in terms of energy resolution, considered the most pertinent parameter for LS3 enhancement, but also delve into the timing capability to demonstrate the alignment of SpaCal technology with the given requirements for Upgrade II.

SpaCal-W with crystal fibres for LS4
The innermost region of the ECAL in the Upgrade II will be equipped with modules composed of scintillating crystal fibres and tungsten absorber.A SpaCal prototype module consisting of 9 cells -4 -of 1.5 × 1.5 cm 2 of pure tungsten absorber with a density of 19 g/cm 3 , with two segmented sections 4 and 10 cm long (offering a 7 and 18 radiation lengthes and a Molière radius of 1.45 cm [13]) was used for testing at DESY.Energy resolution was found to be better for large incident angles.The specific results for tilting it with respect to the beam by 3 • in both the horizontal and vertical planes (azimuthal and polar angles, 3 • +3 • ) and a beam of energies ranging from 1 to 5 GeV, provide a sampling term of (10.2 ± 0.1)% and a constant term of 1-2%, as shown in the figure 6 (left).Similarly, the time measurements were performed with the double-side readout channel equipped with GFAG crystals, and the time stamps were obtained using constant fraction discrimination (CFD).The resulting time resolution was better than 20 ps for energies equal or higher than 4 GeV (figure 6, right).It is expected to be similar or even lower for higher energies, as it will be shown in figure 9.

3D printing of absorber and "Module 0"
The application of 3D printing utilizing pure tungsten powder has been identified as a scalable technology for absorber production.It is imperative to maintain a smooth surface to prevent damage to the fibres during insertion, and a very good roughness with an average profile height mean deviations of 5 μm has been successfully achieved.
An extensive research and development campaign in collaboration with EOS in Germany has been undertaken.This initiative involved the fabrication of initial 1.5 × 1.5 cm 2 cells with lengths of up to 10 cm, followed by the production of 4.5 × 4.5 cm 2 pieces (figure 7).Subsequently, larger pieces measuring 12.1 × 12.1 cm 2 were produced and employed in the creation of the named Module 0. In parallel, Laser Add Technology Co. in China has recently produced module-sized pieces, including two 12.1 × 12.1 cm 2 components in the year 2023.

SpaCal-W with polystyrene fibres for LS3
The full-scale Module 0, featuring a tungsten absorber, was assembled at CERN with a 3D-printed tungsten absorber.This module was filled with single-cladded organic scintillating fibres (1×1 mm 2 , Kuraray SCSF-78), and a well-established procedure involving gluing and polishing was employed during assembly.A quartz fibre was inserted in one hole per cell for the calibration system using LEDs.
-5 -  The PMT holder featured radiation-hard hollow light guides, and the entire module (figure 8) underwent extensive test beam characterization at DESY in May 2023, as well as at CERN in June and August-September 2023 [14].The Module 0 energy resolution was found to be 9.88% for the sampling term and 1.127% for the constant term (figure 9 left) after the noise contribution was subtracted, which is in very good agreement with simulations.
The timing measurements were studied using the prototype with 2 × 2 cells, hollow light guides and the multi-anode Hamamatsu R7600-M4 (instead of the R14755U-100 used for the energy resolution).It resulted in a time resolution better than 20 ps for energies above 40 GeV.

SpaCal with lead absorber
The SpaCal with lead absorber prototype is integrated with polystyrene fibres, specifically of 1 mm diameter utilizing SCSF-78M.This configuration comprises 9 cells, each measuring 3 × 3 cm 2 , with an approximate Molière radius of 3 cm.The cell is divided in two longitudinal sections with dimensions of 8+21 cm (7+18 radiation length).A reflective mirror is strategically positioned between these sections to optimize the time resolution.
The test beam study presents findings derived from experimental runs conducted at both DESY and the CERN SPS, providing a comprehensive assessment of the performance of the configured detector.The configurations scrutinized include the channel to PMT direct contact, the use of PMMA light guides and bundling the fibres outside the module for a direct contact on the PMT.An important aspect of the investigation involved the comparison of different readout configurations in order to search the one with the best time resolution, in this case the one in the direct contact to the PMT, and use it as the reference.The derived energy resolution after the noise contribution was subtracted results in a sampling term of 10.04% and a constant term of 1.156% (figure 10 left).A comparative analysis with simulation results reveals a commendable level of agreement between the experimental data and the simulated expectations.
Employing a double-side readout configuration, and a PMT in direct contact with the cell, the time resolution reaches 20 ps for energies higher than 20 GeV (figure 10 right).

Shashlik with fast WLS fibres
The existing Shashlik modules in the LHCb experiment exhibit favorable time properties, and there is a pursuit of further enhancement through the replacement of wavelength-shifting (WLS) fibres with faster alternatives.The current LHCb configuration employs Y11 fibres with a decay time of 7 ns.To advance the temporal characteristics, alternative fibres have been considered, including YS2 with a 3 ns decay time and YS4 with an even shorter decay time of 1.1 ns.
To systematically evaluate the impact of these fibre replacements, comprehensive measurements have been conducted at both DESY and the SPS.The study involved the use of both current (R7899-20) and faster (R7600-20) photomultiplier tubes from Hamamatsu in a single-side readout configuration.Figure 11 shows the different time resolutions of a Shashlik module using the different optical fibres.The results show the time resolution of 20 ps is achieved for energies higher than 40 GeV when a YS4 fibre is used together with a R7600-20 PMT.

Summary and conclusion
The impending replacement of the innermost 176 modules within the LHCb ECAL during LS3 is needed to mitigate the effects of the radiation damage.The proposed solution involves the adoption of SpaCal technology, incorporating tungsten and lead absorbers, as it satisfactorily meets the prescribed requirements for this specific region.
The subsequent Upgrade II during LS4 will introduce enhanced timing capabilities at the tens of picosecond level, accompanied by heightened radiation hardness criteria.The attainment of timing precision surpassing 20 ps will be realized through the utilization of Shashlik and SpaCal technologies.In the central region, crystal fibres will be integrated to augment performance.
Concurrently, a comprehensive research initiative is underway which encompasses test beam measurements utilizing prototypes, detailed Monte Carlo simulations, exploration of novel absorber production techniques, examination of suitable photomultiplier tubes (PMTs), and the development of advanced readout electronics.Additionally, the investigation extends to the exploration of new radiation-hard and fast scintillators.

Figure 1 .
Figure 1.Expected accumulated radiation dose (Gy) on the ECAL after an integrated luminosity of 300 fb −1 .

Figure 3 .
Figure 3. Scheme of a SpaCal cell with longitudinal scintillating crystals inserted in an absorber.

Figure 4 .
Figure 4. Scheme of a Shashlik cell, in which the light is captured with WLS fibres and transported to the photomultiplier tubes at the ends.
left), posing challenges, particularly in the reconstruction of neutral pions.Post LS3 enhancement, considering the new SpaCal technology, an examination of the occupancy map reveals a comparatively uniform distribution across the regions of interest (figure 5 right).

Figure 5 .
Figure 5. Granularity effect on the occupancy with the present ECAL channel distribution (left) and after LS3 enhancement (right).

Figure 6 .
Figure 6.Energy resolution to electrons (left) and time resolution (right) for the SpaCal-W with crystal fibres prototype.

Figure 8 .
Figure 8.The fully assembled SpaCal prototype used at DESY and CERN test beams in 2023.It includes the W-SpaCal with plystyrene fibres, as well as the PMT holder and the hollow light guides, PMTs and cables.

Figure 9 .
Figure 9. Energy resolution to electrons (left) and time resolution (right) for the SpaCal-W 3D printed "Module-0" with polystyrene fibres and "hollow" light guides prototype (in LS3 enhancement configuration).

Figure 10 .
Figure 10.Energy resolution (left) and time resolution (right) for the SpaCal lead with polystyrene fibres.

Figure 11 .
Figure 11.Time resolution for the Shashlik technology using different optical guides.

Table 1 .
Basic SpaCal module technology types to be installed in LHCb during LS3 and LS4.