Performance of the BIS78 RPC detectors: a new concept of electronics and detector integration for high-rate and fast timing large size RPCs

The reduction of the average charge per count in the gas along with the capability to discriminate very small signals from noise can allow an efficient and long-term Resistive Plate Chamber detector operation in high radiation background environment. This goal has been reached during the R&D program of the BIS78 upgrade project of the ATLAS detector at LHC through the deep integration of a fast (100 ps peaking time) and sensitive (as small as 100 `V threshold) Front- End electronics (FE) with a very large size detector structure. This innovative RPC integration concept pivots on a newly conceived faraday cage, embedding the readout strips and the FE, tightly wrapped around a 1 mm gas gap RPC with 1.2 mm thick electrodes, as a fully independent singlet structure. We studied the performance of BIS78 production triplet chambers, made of 3 independent singlets of 2<2, each providing a 2D space and time information, showing a minimum threshold achievable of 2 pC of average charge per count produced inside the gas gap. We show that these chambers grant a record combined performance of better


BIS78 ATLAS RPC
The BIS78 project is the pilot project [1] for the barrel inner (BI) plane of the muon spectrometer upgrade scheduled for the Phase-2 ATLAS upgrade. These chambers are designed to be compatible with the high luminosity LHC (HL-LHC) conditions and the BI RPCs will inherit most of the design. The BIS78 RPC chambers will be deployed in the Phase-1 upgrade, providing a new system which integrates small-diameter Muon Drift Tube (sMDT) + RPC detectors to be installed in the barrel-endcap transition region at 1.0 < |η| < 1.3, which is the region characterized by the highest background rates in the barrel. The estimated total rate of the Level-1 single-muon trigger with p T > 20 GeV is expected to rise during Run-3 up to 57.6 kHz, while the allocable is 25 kHz for muon triggers out of a total bandwidth of 100 kHz. In order to address this problem the pseudorapidity region 1.3 < |η| < 2.5 will be covered by the New Small Wheels [1] that will solve the problems of fake triggers mainly coming from secondary charged particles generated inside the cryostats and will improve the muon selectivity and the discrimination of low momentum muons. The region at 1.0 < |η| < 1.3, which corresponds to the barrel-endcap transition region, will be covered with the BIS78 RPC chambers, leading to a total coverage of the trigger of ≈ 83.5% of the transition region. The BIS78 stations are formed by two triplets of RPCs coupled to a multilayer sMDT in the same envelope, without making any electrical contact between the two systems preserving their electrical and mechanical integrity. The BIS78 RPC system consists of 16 BIS7 and 16 BIS8 triplets. The total area covered by this system is 94 m 2 . The new RPCs have a gas gap width of 1 mm. This choice allows to reduce the total thickness of the system and leads to other benefits: the RPC time resolution scales with the gas gap width. Thus this reduction leads to an improvement of the time resolution up to ≈ 0.4 ns. The resistive electrodes thickness also has been reduced to 1.2 mm. This reduction takes the advantage of increasing the charge fraction transferred to the pick-up electrodes, resulting in an improvement of the Signal-to-Noise ratio. Each detector is readout on both sides by two panels of orthogonal strips, with strip pitches of 24-26 mm depending on the chamber type, providing η and φ coordinates. The challenge of making the chambers compatible with operation at higher hit rates is addressed by a proportional reduction of charge per count, while simultaneously increasing the sensitivity and signal-to-noise ratio of the Front-End electronics without degrading the fast timing features of the RPCs. In order to improve the rate capability a new Front End electronics has been realized. Also a new Faraday cage design, suitable for low-threshold operation, is being developed, allowing a better shielding of the more sensitive Front-End electronics. All these inter-dependent requirements have led to a full re-design of the RPC, optimizing it in all aspects, from materials, to readout electronics, up to chamber layout. The new-generation RPCs will increase the rate capability by an order of magnitude, decrease the total chamber weight and thickness, operate at near half the working voltage. In Table 1 are reported the main features of the BIS78-type RPCs compared to the ATLAS Standard one.  [3] rate capability is mainly limited by the current that can be driven by the high resistivity electrodes. It can be improved by modifying the highly interconnected parameters which define the voltage drop on the electrodes. These parameters may be derived by applying the Ohm's law: Where V gas is the voltage applied across the gas, V a is the total voltage supplied to the detector, ρ is the resistivity of the electrodes, d is the electrodes thickness, Q is the average charge per count produced inside the gas and Φ particles represents the flux of ionizing particles. This equation shows several ways to increase the detectable particle flux, however, the approach chosen by this project is to reduce the average charge per count Q , allowing to increase the rate capability while operating the detector at fixed current. The drawback of this approach is that reducing the average charge per count implies a reduction of the injected charge inside the Front End, hence the signal which needs to be detected will be much smaller. For this reason, a very sensitive Front End electronics is mandatory in order to detect such small signal. Moreover an high suppression of the noise induced inside the detector by the electronics and by external sources is required, leading to the chamber structure optimization as a Faraday cage. The developed Front End electronics board, reported in Figure 1, is composed of eight channels of a new preamplifier coupled with two full-custom ASIC discriminators with four channels each, and LVDS transmitters integrated directly inside the board. The overall features of the preamplifier and of the ASIC discriminator are reported in the Table 2.

BIS78 RPC performance with cosmic rays
The BIS78 RPCs performance have been tested by means of cosmic rays. The detector tested is composed of 3 independent mono-gap singlet equipped with the new Front-End electronics. The gas mixture used is 95% TFE, 4.7% I-C 4 H 10 , 0.3% SF6. The trigger system has been realized by the usage of two layers of scintillators, and the tested chamber has been put between them. The data acquisition has been realized by using a CAEN TDC with 100 ps of time resolution. In Figure  2 the experimental setup is reported. The main parameters monitored in order to study the detector performance with cosmic rays are the efficiency, the cluster size, the channels noise and the time resolution.

Efficiency
The efficiency parameter has been studied in order to verify the overall behaviour of the detector. Two measurements have been performed checking this parameter: the efficiency value as function of the applied High Voltage and the efficiency map at the knee point. The Figure 3a shows the expected efficiency curve with a 95% of efficiency at 5.6 kV. The Figure 3b represents the efficiency map taken at the knee point of the efficiency curve (5.5 kV), in order to be more sensitive to any efficiency variation. This map has been obtained by dividing the RPC in 8 different zones (positions) and for each zone the efficiency has been measured. The efficiency variations reported in the map are within the statistical fluctuations, hence the efficiency is homogeneous through all the detector surface. The Cluster Size parameter has been studied mainly to monitor the self-induced noise and the crosstalk effects, since those effects would increase the cluster size to anomalous values. The cluster size mapping has been obtained by dividing the chamber in 8 positions and for each position the size of clusters has been monitored. This map allows to verify the absence of any hot spot and to check the homogeneity of this variable. The cluster size map is reported in Figure 4.

Cluster Size
The map shows an homogeneous behaviour of the cluster size all over the detector surface with a value completely consistent with the charge induction phenomenon, implying that there are neither crosstalk nor self-induced noise effects.

Channels noise
The noise map allows to highlight problems related to disturbance and broken channels. This measurement has been performed by using a random trigger and by counting for each channel the spurious events. The channels counters have been renormalized according to the whole time acquisition window and to the strip area. The noise map reported in Figure 5 shows the noise rate in units of Hz/cm 2 of both the readout layers η and φ. The η readout has an almost homogeneous behaviour with an average noise of 0.3 Hz/cm 2 . The φ layer has an average noise rate of 0.5 Hz/cm 2 . The φ layer is more susceptible to the Low Voltage connection point, which is located near the channel 64, exhibiting an higher noise rate near this region.
However the overall noise rate for both readout layers is within the requirements being below 1 Hz/cm 2 .

Time resolution
The time resolution of the BIS78 RPC detectors has been tested by using the Time-Of-Flight method between 2 singlets out of 3 (all the combinations have been checked leading to the same result) and by correcting the discriminator timewalk effect with the Time-Over-Threshold measurement perfomed by the Front-End electronics itself [5] (see Figure 6). Moreover, for each readout channel all the systematic effects have been taken into account and corrected for. In order to correct the timewalk effect, the function which correlates the amplitude of the signals with the time in which the signal surpasses the threshold has been found and corrected for event by event. In Figure 7 the time resolution of the BIS78 RPC is reported with and without the time walk correction, achieved at the working point (5.6 kV).
The raw BIS78 time resolution is ∼ 400 ps, while applying the timewalk correction leads to a slight improvement obtaining a record time resolution of ∼ 330 ps with a monogap RPC. The  improvement achieved with the timewalk correction is small due to the already very fast Front-End electronics that makes the timewalk very small [5].

BIS78 RPC performance in high radiation environment
The performance of the RPC detector equipped with the newly developed Front-End electronics under high irradiation has been the crucial point of the GIF++ testbeam at CERN. Moreover, this measurement has been fundamental in order to verify the improvement in the rate capability achieved by the RPC detector by means of this new FE electronics. However the FE version used in this test is not the final one and it shows some instabilities in terms of crosstalk and self induced noise, causing multiple channels to fire simulataneously or to fire multiple times on the same signal or increasing the discriminated signal width. Those problems have been solved within the final version.
This test has been carried on by setting two different sets of FE electronics parameters. One set is the most performing in terms of efficiency and minimum achievable threshold, lacking, however, in system stability. The other set is more conservative in terms of minimum achievable threshold granting on the other hand high system stability. The goal of this measurement has been the study of the BIS78 RPC performance under high irradiation; on one hand testing the detector standard working condition when the conservative set is used and on the other hand checking the highest performance achievable when the performing parameters are set. The results achieved by both FE settings and by irradiating the detector with the 14 T Bq 137 Cs source with various shielding factors and by means of a muon beam have been reported. The detector has been positioned at ∼ 5 m from the source, its area is 1.8x1.1 m 2 and the whole experimental setup is reported in Figure 8.
The detector I-V curve as function of the absorption factor has been monitored and it is reported in Figure 9. The current is defined as a function of the average charge per count Q and the counting rate C photons on the detector as: The current I is normalized with respect to the detector active area, hence the detector current density (µA/cm 2 ) is considered. Therefore, measuring the rate of converted photons C photons and the current driven inside the detector with the various source shielding, the average charge per count Q per converted photon can be estimated. Since C photons is also the number of photons measured by the FE-RPC system, the threshold applied by the Front-End electronics can be expressed in this way in terms of the average charge per count delivered inside the detector.
This measurement has been perfomed, considering the current density in the first efficiency plateau point after the knee with efficiency ≥ 95%. The Front-End thresholds applied on the charge per count are evaluated by means of the Equation 3.1.
This study has been performed on both η and φ readouts for different absorption factors [5] and for both performing and conservative threshold settings. By means of the FE conservative setting an average charge per count produced inside the gas Q of 5 ± 1 pC is achieved.
The efficiency curves for muons with the conservative settings and for different absorption factors is reported in the Figure 10a. The efficiency curves does not show any deviation from the source-off behaviour. Operating the detector in these conditions leads to no efficiency losses for rate of converted photons up to 2 kHz/cm 2 .
Taking into account the FE performing setting, the estimated FE threshold in terms of average charge per count on both η and φ readouts is ∼ 3 ± 1 pC. The big uncertainty is due to the difficulty to correctly estimate the occupancy and the measured rate, due to system instabilities described before.
The efficiency curves achieved with the perfoming threshold and with different absorption factors are reported in Figure 10b. It must be underlined how those efficiency values, especially for very low absorption factors (< 10Abs), represent only the lower limit to the real detector efficiency which is higher, since the high level of occupancy is the main cause for the flat efficiency losses. However with the performing threshold set a rate capability up to 9 kHz/cm 2 has been achieved, confirming how the Front-End electronics achieved a ∼ 3 pC charge threshold on the average charge per count produced within the gas gap.

Conclusion
The BIS78 project is the pilot project for the ATLAS detector upgrade of the barrel muon spectrometer foreseen to meet the requirements of the HL-LHC running conditions. The BIS78 RPCs detectors grant a record combined performance of better than 95% single gap efficiency, time resolution of ∼ 330 ps and up to 9 kHz/cm 2 of rate capability. Moreover, the succesfulness of the integration of the FE electronics within a large area detector has been demonstrated, reaching high performance along with a good homogeneity overall behaviour through all the BIS78 modules produced.