Sealed (zero gas flow) RPCs — a new operational paradigm for gaseous detectors

1 Introduction

Resistive Plate Chambers (RPCs) Santonico and Cardarelli (1981) are a well-established class of gaseous detectors in High Energy Physics (HEP), known for their good time resolution, high detection efficiency, scalability to large areas and affordability Abbrescia et al. (2018). Depending on their design and operating conditions, RPCs can be broadly classified into trigger RPCs, among others, ATLAS Collaboration (2021); Collaboration et al. (2019); ALICE Collaboration et al. (2012), generally characterized by gas gaps of a few millimeters (typically ∼ 1–2 mm) and time resolutions of the order of ∼ 0.5–1 ns, and timing RPCs, such as, Carnesecchi (2019); Belver et al. (2009), which employ narrow gas gaps (typically ∼ 0.2–0.3 mm) and high electric fields to achieve time resolutions better than 100 ps. These characteristics have led to their widespread use in collider experiments and to their exploration in other fields such as medical imaging Fonte et al. (2023), neutron detection Margato et al. (2023), and cosmic-ray physics Aielli et al. (2006); Lopes et al. (2023).

The operation of RPCs depends on the purity of the gas mixture, which is mainly based on hydrofluorocarbons (HFCs), predominantly C 2 H 2 F 4, with small fractions of SF 6. Gas contamination arises primarily from leaks and material permeability, which allow atmospheric gases and humidity to enter the system, as well as from chemical by-products generated during avalanche processes. To maintain gas purity, RPC systems rely on continuous gas circulation schemes involving recirculation, purification, and monitoring stages Rigoletti et al. (2023). These systems are inherently complex and costly, particularly in large-area detectors. Even with advanced purification systems, the continuous injection of significant amounts of fresh gas is unavoidable, due to leaks and permeability in detectors, gas lines, and associated components Mandelli et al. (2022). This ultimately results in substantial gas consumption. The current European regulatory framework governing the use of HFCs European Parliament and Council of the European Union (2024), where their release is prohibited and their progressive phase-out has been mandated due to their high Global Warming Potential (GWP), imposes increasingly stringent constraints on existing systems and represents a major limitation for future detectors, which may no longer be designed under the same operational paradigm.

This situation has motivated an intense effort within the community to develop alternative gas mixtures with reduced environmental impact, the so-called eco-friendly gases Quaglia et al. (2024); Mandelli et al. (2022); Guida et al. (2020). Among them, C 3 H 2 F 4, HFO-1234ze, with a GWP100 of approximately one according to the IPCC AR6 assessment Masson-Delmotte et al. (2021), has emerged as a promising replacement for C 2 H 2 F 4, with a GWP100 of about 1530 Masson-Delmotte et al. (2021). While these solutions have shown encouraging results for trigger RPCs, their application to timing RPCs still represents an important challenge. In particular, identifying a suitable alternative to SF 6, with a GWP100 of approximately 24,300 Masson-Delmotte et al. (2021), which plays a crucial role in controlling avalanche stability at high electric fields Camarri et al. (1998), remains an active area of research. Recent studies have reported promising developments regarding eco-gas mixtures for timing Garillot et al. (2024), although further work is still required to demonstrate comparable long-term performance and stability. Furthermore, even in the case of successful gas replacement, the intrinsic complexity and cost of gas handling systems would remain essentially unchanged.

A conceptually different approach consists in the development of RPCs capable of operating without any external gas supply. In this scheme, the detector is filled with gas during assembly and subsequently hermetically sealed, similar to the Geiger–Müller tubes. These devices, referred to as sealed RPCs (sRPCs) Lopes et al. (2020), drastically reduce gas consumption and therefore mitigate the environmental impact associated with HFC usage. At the same time, they eliminate the need for complex gas infrastructures, significantly reducing system cost and operational complexity. In addition, the absence of gas systems enables new deployment scenarios, including portable devices and large-area detector arrays in remote or inaccessible locations, thereby opening new opportunities for the expansion of RPC technology beyond traditional laboratory environments. Importantly, this approach is intrinsically compatible with timing RPC operation, as it does not require modifications to the gas mixture composition.

In this context, the present work explores the feasibility of sealed trigger RPC technology through the study of two detector configurations: a medium-area module of approximately 0.1 m 2 and a large-area module of 1 m 2. These detectors have been characterized and operated over extended periods of time in order to assess their stability and viability.

2 Sealed RPC concept and early developments

The development of sealed RPC technology has followed an iterative process, in which several approaches were explored with varying degrees of success. Some of these early attempts and exploratory prototypes are discussed in the following References Lopes et al., 2023, Lopes et al., 2020). These initial studies were essential to identify the guiding principles and main limitations for the implementation of stable sealed detectors.

Two major challenges were identified at an early stage: the sealing of the gas gaps and the stability of the high-voltage (HV) electrodes. In particular, several configurations based on O-ring sealing were investigated in an attempt to define and isolate the gas volume. However, none of these approaches provided satisfactory results, as rapid gas degradation was consistently observed, indicating insufficient sealing and/or material compatibility issues. In parallel, the initial implementations of the HV resistive layer exhibited a strong sensitivity to ambient humidity due to the lack of proper electrical insulation, leading to unstable operating conditions.

The strategy that ultimately proved successful, and which is adopted in this work, is based on the exclusive use of glass as the material in contact with the gas. In this approach, the gas volume is entirely confined within a structure defined by soda-lime glass elements, thereby minimizing permeability and external contamination. The resistive electrodes are made of soda-lime glass plates, while the gas gaps are defined using glass spacers also made of soda-lime. Soda-lime glass was selected for the spacers for the present prototype primarily due to material availability at the laboratory during detector construction. While adequate for the detector operation reported here, its lower bulk resistivity compared to borosilicate glass may not be optimal for long-term sealed operation. Future implementations will consider borosilicate glass as a potentially more suitable alternative due to its higher resistivity. Each gap is formed by two glass electrodes, with the perimeter enclosed by a continuous peripheral glass strip. In the central region of the active area, the gap thickness is maintained by circular glass spacers glued to the electrodes using an epoxy glue (Araldite™). During assembly, the multi-gap structure is maintained under controlled pressure, and all peripheral surfaces are sealed with the same epoxy to ensure gas tightness and provide mechanical strength. Gas inlets and outlets (one per gas gap) are implemented using standard plastic dispensing needles (Loctite™), inserted through the peripheral spacer region outside the HV area. After gas flushing, the channels are hermetically sealed by mechanically crushing the needle. Although functional, this solution makes reopening the gas channels technically challenging due to their very small diameter. This aspect will be revisited in future developments. The HV electrodes consist of a semiconductive layer based on an acrylic paint (Badger Air-Opaque Paint), with a surface resistivity of approximately 100 M Ω / □, airbrushed onto the outer surfaces of the external glass plates while avoiding the regions corresponding to the spacers. These electrodes are covered with Mylar™ and Kapton™ layers for electrical insulation, see Figure 1a.

The conditioning of sealed RPC detectors follows a well-defined sequence of steps aimed at ensuring surface cleanliness, gas purity and long-term stability of the detector performance.

Prior to assembly, all glass components must be cleaned as thoroughly as possible in order to minimize the presence of dust particles, grease, or any surface contaminants that could act as localized defects or “hot spots” during operation. An exhaustive cleaning procedure, ideally performed under controlled conditions such as cleanroom environments, is of critical importance for achieving uniform and stable detector behavior. Immediately after assembly, the gas gaps are flushed with argon for several days. This step is primarily intended to remove residual humidity from the internal surfaces, particularly from the glass electrodes, as water molecules can significantly affect both the electrical properties of the detector and the gas quality. Following this initial flushing, the detector is subjected to a sustained discharge in argon. During this phase, the HV is gradually increased until a stable discharge is established across the full gap area. This process promotes a cleaning of the internal surfaces through a mild plasma etching mechanism, in which contaminants are progressively removed or neutralized. At the same time, this step allows for the verification of the global resistive behavior of the detector. A uniform and stable current over time is a strong indicator of good surface quality and gap uniformity, whereas fluctuations or localized variations in the current may reveal structural defects or insufficient cleaning. This conditioning stage is typically maintained for a period of 2 weeks.

After completion of the conditioning process, the gas gaps are flushed for several days with the nominal counting gas mixture, consisting of 97.5% C 2 H 2 F 4 and 2.5% SF 6, at a flow rate of approximately 4 cm 3/min. This ensures a proper replacement of the argon and stabilization of the gas composition. Finally, the gas inlets and outlets are hermetically sealed, and the detector is ready for operation in sealed mode.

3 Experimental setup

3.1 Sealed RPCs

Two types of sealed RPC modules have been built to study the influence of size on this type of detector: one of medium size, with an active area of 300 × 300 mm², hereafter referred to as sRPC 300, and another with an active area ten times larger, 950 × 950 mm², called sRPC 950. Each of them is described below.

3.1.1 s R P C 300 modules

The sRPC 300 modules are a multi-gap structure Zeballos et al. (1996) featuring two 1 mm gas gaps. These are defined by three 2 mm thick soda-lime glass electrodes, with a bulk resistivity of approximately 5 × 1 0 12 Ωcm at 25 °C and dimensions of 350 × 350 mm 2. Each gap includes a circular spacer, consisting of a soda-lime glass disk (10 mm in diameter) glued at the center of the active area, and a peripheral strip spacer (25 mm wide), also made of soda-lime glass. The HV electrodes are applied to the external glass plates over an area of 300 × 300 mm 2.

The sRPC 300 modules are read out using a strip plane consisting of a 1.6 mm thick Flame Retardant 4 (FR4) printed circuit board (PCB). One side of the PCB is patterned with sixteen copper strips, each 18 mm wide, with a pitch of 19 mm and a length of 400 mm, positioned above the module. A ground plane on the opposite side completes the readout structure. The entire assembly is enclosed in an aluminum box that provides electromagnetic shielding and mechanical rigidity. Each strip is connected at both ends to the inner conductor of a coaxial cable, while the outer conductor is connected to the ground plane. The coaxial cables are terminated with micro-miniature coaxial (MMCX) RF feedthrough connectors, as shown in Figure 1b.

3.1.2 s R P C 950 module

The sRPC 950 module is similar to the sRPC 300, but features an active area of approximately an order of magnitude larger, 950 × 950 mm 2, and a spacer matrix to define the gas gaps. In this case, the multi-gap structure is formed by glass electrodes with dimensions of 1000 × 1000 mm 2 and each gas gap includes a 5 × 5 circular spacer matrix distributed over the central region of the active area, as shown in Figure 2a. The HV semiconductive layers are airbrushed over an area of 950 × 950 mm 2, as shown in Figure 2b.

The module is read out by a strip plane patterned on one side with thirty-two copper strips, each 30 mm wide, with a pitch of 31 mm and a length of 1050 mm, positioned above the module. The remaining details of the readout system are identical to those of the sRPC 300 module.

3.2 Plastic scintillators

To trigger on atmospheric muons and evaluate the sealed RPC response, small plastic scintillators were used. In the case of the sRPC 300 module, a set of four 40 × 40 × 10 mm 3 plastic scintillators, each read out on one of the smaller faces by a silicon photomultiplier (SiPM, MICROFC-60035-SMT), are stacked vertically to form a mini-telescope. The scintillators are arranged such that the solid angle defined by the outermost detectors selects a region fully contained within the active area of the sRPC 300 (see Figure 3a). For the sRPC 950 module, two similar plastic scintillators with an active area of 100 × 100 mm 2 are used, positioned upstream and downstream of the detector, as shown in Figure 3b.

3.3 Electronics

All signals generated by the sealed RPC modules and the scintillators are fed into fast Front-End Electronics (fFEE) Belver et al. (2010) also used for the HADES RPC-TOF detector Belver et al. (2009). The fFEE includes a fast amplification stage based on a BGM1013 amplifier, providing a gain of approximately 35.5 dB at 1 GHz, followed by an ultra-high-speed MAX9600 comparator (in the case of the scintillators, the amplification stage is bypassed). The applied effective discriminator threshold corresponds to an input charge of approximately 25 fC before amplification. In the case of sRPC 300 2 strips are not readout to allow for the reading of the SiPM signals. The output of this fFEE, a digital pulse in low-voltage differential signal (LVDS) format, encodes the timing information of the signal in the leading edge, while the pulse width corresponds to its charge. It should be noted that the measured charge corresponds to the fast component of the signal, due to the integration time of the fFEE, of the order of 40 ns. The resulting digital signals are input to the TRB3 boards Neiser et al. (2013), which are equipped with 32 multi-hit Time to Digital Converters (TDCs) using TDC-in-FPGA technology, offering a time precision better than 20 ps and a multipurpose logic unit. This logic unit enables triggering either on muons passing through the plastic scintillators or on all signals coming from the sealed RPC modules, identified from now on as muon-trigger and self-trigger, respectively.

With this arrangement, we can provide information on:

• Charge, Q, as the sum of the induced charge on the strips (or the mean charge, < Q >, as the mean of the Q distribution), and streamer probability, defined here, as the percentage of events with a charge larger than 1.2 pC. This threshold was empirically chosen based on charge spectra observed in similar RPCs operating with pure C 2 H 2 F 4, where streamer signals appear clearly separated from the avalanche regime.
• Longitudinal position along the strips, Y, computed as the difference of times measured at both ends of the strip with maximum charge multiplied by the signal propagation velocity on the strip, evaluated to 165 mm/ns.
• Transversal position to the strips, X, evaluated as the position of the strip with maximum charge. Both, longitudinal and transversal positions are determined with a resolution better than 10 mm (measured in a detector with similar characteristics and the same electronics Soneira-Landín et al. (2025)).
• Efficiency of the sealed RPC modules defined as the ratio of the number of muons detected by the sealed RPC to the number of muons identified by the scintillator-based system.

4 Results

4.1 s R P C 300

4.1.1 s R P C 300 operated in the lab

Figure 4 shows the efficiency, mean charge, streamer probability and count rate (“dark” count rate) as a function of the applied voltage per gap for the sRPC 300 module. The efficiency curve exhibits the characteristic behavior expected for a standard RPC with similar parameters (operated in continuous gas flow), reaching a plateau above 5800 V with an average value of 97.9 %. The mean charge displays the expected trend, reaching values of about 600 fC and higher in the plateau region, while the streamer probability remains below 5% at 6300 V. Based on these results, the HV working point was set to 6000 V and kept stable throughout the measurements.

Figure 5 shows the 2D plots of occupancy, mean charge and streamer probability, as a function of the X and Y coordinates of the hits, for a HV of 6000 V. The red square indicates the active area of the sRPC 300 module. The left column, Figures 5a–c, correspond to muon-trigger events (selected by the plastic scintillators). It is possible to clearly identify the spot in the bottom-right quadrant of the sRPC 300 module created by the atmospheric muons characterized by a uniform mean charge and a very low, almost non-existent, level of streamer probability. The right column, Figures 5d–f, correspond to self-trigger events (all signals generated by the sealed RPC). Figure 5d shows localized regions with an increased accumulation of hits in the central and right-side areas of the detector, corresponding to regions with a higher background event rate. Geometrically, these regions are located close to detector spacers. Although their origin cannot be established with certainty at this stage, they could be related to local imperfections in the spacer region, such as excess glue or small construction defects affecting the local electric field. Figure 5e shows that the mean charge has no noticeable correlation with position; the streamer probability over the detector surface is shown in Figure 5f. No significant variation in the detector response was observed when changing its position relative to the ground, suggesting the absence of measurable effects related to gas stratification within the detector volume.

Figure 6 shows the trends of the quantities mentioned above as a function of time for a period of 542 days. Each point corresponds to approximately 24 h of data taking and about 5k recorded events. Figure 6a shows the efficiency for cosmic muons and count rate (“dark” count rate). Figure 6b shows the mean charge both for the muon-trigger and self-trigger events, Figure 6c shows the streamer probability for the same triggers, while Figure 6d shows the laboratory temperature and atmospheric pressure. All these plots are shown as a function of time, covering a total of 542 days of operation. However, the module has been operating since May 2024, including a period of approximately 1 month of high exposure to a radioactive source (see below), accumulating nearly 2 years of operation. It should be noted that this module was filled with gas during its assembly in May 2024 and has not received any new gas since then. Figure 6 shows only the last 542 days, as this period corresponds to stable operating conditions. During this period, the efficiency remains highly stable over time, with an average value of 98.2%, while the other parameters, namely, the count rate, mean charge, and streamer probability, show a slight dependence on laboratory temperature and atmospheric pressure.

Figure 7 explicitly shows the dependence of the mean charge for muon-trigger events with laboratory temperature and atmospheric pressure. Compared to conventional RPC detectors operating in continuous gas flow at atmospheric pressure, sealed RPC detectors exhibit an inverse dependence on temperature and pressure, due to their sealed configuration and sensitivity to mechanical deformations. Consequently, an increase in atmospheric pressure or a decrease in temperature (leading to a reduction in internal pressure) results in greater compression of the gas gaps and, therefore, an increase in the mean charge. Conversely, a decrease in atmospheric pressure or an increase in temperature (leading to an increase in internal pressure) reduces the compression of the gas gaps, resulting in a decrease in the mean charge.

4.1.2 s R P C 300 operated under strong irradiation

The same module shown on Figure 6 was exposed, in August 2024 (i.e., prior to the period shown in Figure 6), to an intense radiation source for approximately 1 month. This test was performed to investigate, in an accelerated manner, possible gas degradation effects induced by electron multiplication processes that could compromise detector performance. No noticeable changes in efficiency, background count rate, mean charge, or streamer probability were observed during the irradiation period. This exposure corresponds, based on the integrated charge, to approximately 1435 days of operation, for example, under natural atmospheric cosmic muon flux. Further details on this study can be found in Blanco et al. (2025a).

4.1.3 s R P C 300 operated at SND@LHC

Using four sRPC 300 modules, a small telescope consisting of four detection planes was assembled. The system, read out with the same electronics and data acquisition system described in Section 3.3, was installed in the TI18 tunnel of the Large Hadron Collider (LHC) at CERN (Switzerland), directly in front of the SND@LHC Acampora et al. (2024) detector. Its main purpose is the characterization of the muon flux originating from Interaction Point 1 (ATLAS). A detailed laboratory characterization of the modules prior to their transport to CERN can be found in Blanco et al. (2025b).

Unfortunately, after commissioning at CERN, a localized central region exhibiting reduced efficiency and lower mean charge was observed in all four detector planes, with efficiency values decreasing to below 20% in the affected area. Since the detector is operated as part of a four-plane telescope, the performance of each module can be independently evaluated through track reconstruction using the remaining three planes, confirming the localized nature of this effect. One possible interpretation is a local modification of the gas gap in this region. While this hypothesis has not been experimentally verified, it could be associated with the transportation of the already sealed RPC modules from Portugal to Switzerland. The significant temperature and pressure variations experienced during transport, together with the previously discussed sensitivity of these modules to internal pressure changes, may have affected the mechanical coupling of the central circular spacer to the glass electrodes in all four modules, locally altering the detector geometry.

This effect introduces large fluctuations in the measured efficiency and mean charge, primarily driven by pressure variations, as the temperature in the tunnel remains highly stable. Nevertheless, it is still relevant to present the count rate and mean charge for one of the sRPC 300 modules (the others show similar behavior) since its installation in February 2024, i.e., over more than 2 years of operation.

Figure 8a shows the rate measured in one of the telescope planes, revealing two clearly differentiated vertical regions. A lower level, below approximately 1 Hz/cm², corresponds to periods without LHC interactions, while a second region around 2–3 Hz/cm² is associated with periods in which LHC collisions are present. The winter shutdown periods, during which the LHC is not operating, are also clearly visible, showing only the lower counting-rate regime. Figure 8b presents the corresponding evolution of the mean charge over the same period. After an initial phase involving some operational adjustments until approximately June 2024, the mean charge remains relatively stable, with values typically between 600 and 700 fC.

Despite being modulated by variations in the particle flux present in the tunnel and, to a lesser extent, by the pressure-related effect discussed above, both the counting rate and the mean charge remain stable over time, indicating the robustness and long-term stability of the system.

4.2 s R P C 950

The sRPC 950 follows the same detector concept and construction principles as the sRPC 300, differing only in its active area. During the measurements reported here, the detector was operated at a voltage of the order of 6.2 kV, corresponding to stable operation conditions.

Figure 9 shows the 2D plots of occupancy, mean charge and streamer probability, as a function of the X and Y coordinates of the hits. The left column, Figure 9a, b and c, correspond to muon-trigger events (generated by the plastic scintillators). It is possible to clearly identify the spot in the bottom-left quadrant of the sRPC 950 module created by the atmospheric muons characterized by a uniform mean charge and a very low, almost non-existent, level of streamer probability. A response very similar to that of the sRPC 300 module, see Figure 5. The right column, Figures 9d–f correspond to self-trigger events (all signals generated by the sealed RPC). No significant structures are observed, except for a few hot spots visible in all cases, which are identified as the positions of some of the spacers, both those corresponding to the 5 × 5 matrix and the peripheral spacer.

Figure 10 shows the trends of the quantities mentioned above as a function of time for a period of 49 days. Each point corresponds to approximately 6 h of data taking and about 5k recorded events. Figure 10a shows the efficiency for cosmic muons and count rate of the sRPC 950 module. Figure 10b shows the mean charge both for the muon-trigger and self-trigger events, Figure 10c shows the streamer probability for the same triggers, while Figure 10d shows the laboratory temperature and atmospheric pressure. All these plots are shown as a function of time, covering a total of 49 days of operation. For the sRPC 950 module, this interval corresponds to the only period with complete monitoring data and stable operating conditions. Outside this period, either the monitoring was incomplete or the detector was operated under varying conditions, including changes in high voltage or modifications to the trigger system.

After this period, an attempt was made to evaluate the spatial uniformity of the detector response, which is particularly relevant given the large active area of the module. The results showed significant non-uniformities across the detector surface. These variations are currently interpreted as arising from mechanical deformations of the structure, likely induced by internal pressure changes. Such effects may have led to the partial detachment of one or more spacers, resulting in local variations of the gas gap and, consequently, non-uniform detector response. The nature and magnitude of the observed effect are comparable to those identified in the sRPC 300 operated at SND@LHC, suggesting that the same underlying mechanism may be responsible in both cases. After approximately 1 year of operation, the detector ceased functioning, exhibiting a rather abrupt increase in both dark counting rate and operating current and becoming non-operational within approximately 2 days. Although the exact cause could not be conclusively determined, this behavior may be related to the aforementioned structural effects, possibly resulting in a loss of gas tightness, substantial air ingress into the detector, and the depletion of the operating gas mixture.

5 Discussion and outlook

The results presented in this work clearly indicate that sealed trigger RPC technology can achieve performance comparable to that of conventional RPCs operated in continuous gas flow, namely, stable operation over long periods, detection efficiencies above 95%, induced charges of the order of 1 pC, and streamer probabilities at the level of a few percent. In particular, key detector parameters such as efficiency, mean charge, streamer probability and dark count rate exhibit the expected behavior for similar RPCs operating under continuous gas flow, as shown in Figures 4, 5. This demonstrates that the absence of continuous gas renewal does not intrinsically compromise detector performance under the conditions explored.

The long-term operation of sealed RPCs also appears to be feasible, even with the relatively simple construction approach adopted in this work, based on glued glass electrodes. The sRPC 300 module, operated in laboratory conditions, shows stable operation over periods of almost 2 years when exposed to low particle fluxes, such as atmospheric cosmic rays (see Figure 6). The modules operated at SND@LHC, exposed to very low muon fluxes of around 2 Hz/c m 2 (see Figure 8). a, showed a broadly similar behavior, at least to some extent, although structural issues affected their performance. These results suggest that sealed RPCs can sustain long-term operation under low particle rate conditions, even with non-optimized construction techniques.

The influence of environmental parameters, namely, atmospheric pressure and temperature, is well understood in the context of sealed operation. In contrast to conventional RPCs, the observed dependence is inverted, reflecting the impact of internal pressure variations on the mechanical structure of the detector and, consequently, on the gas gap thickness. While this effect does not pose a significant limitation for medium-size detectors operated under controlled laboratory conditions, it becomes increasingly relevant for large-area detectors and during transportation of already sealed modules, where significant pressure variations may occur.

From an outlook perspective, several aspects require further investigation and development. Although the construction technique based on glued glass plates has proven sufficient to demonstrate the feasibility of sealed RPCs, significant improvements are expected from more advanced assembly methods, such as glass sealing or bonding techniques, which could enhance both mechanical robustness and long-term stability.

In addition, strategies to mitigate or eliminate the sensitivity to internal pressure variations should be explored. Possible approaches include the implementation of expansion volumes or some type of mechanical stabilization, with the aim of minimizing mechanical stress and preventing structural degradation under large internal pressure excursions.

While the results presented here indicate that glass-based structures provide an effective barrier against external contamination from atmospheric gases and humidity, the role of internal gas degradation mechanisms remains an open question. In particular, the accumulation of by-products generated during avalanche processes could become a limiting factor. While such effects were not observed in the explored modules, they may become relevant under high particle flux conditions, which have not been explored in this work.

An additional aspect that deserves further investigation concerns the long-term evolution of the gas quality inside sealed RPCs. Although no significant degradation has been observed within the time scales explored in this work, it remains important to assess whether detector performance could eventually deteriorate due to internal gas aging processes. In such a scenario, it would be crucial to determine whether the detector can be refilled with fresh gas, restoring its initial performance for a new operational cycle. The feasibility of such a refill procedure would represent a highly valuable feature, particularly if intrinsic limits to the gas lifetime are identified.

Another key objective for future research is the extension of the sealed concept to timing RPCs. This would require demonstrating stable operation at high electric fields in closed volumes, preserving the excellent timing performance characteristic of these detectors.

Finally, while operation outside laboratory conditions has already been demonstrated for RPC systems, it remains to be fully assessed whether sealed RPCs can operate reliably in less controlled environments. The results obtained so far provide encouraging indications in this direction. This is particularly relevant for applications such as large-scale outdoor detector arrays for astroparticle physics, where the simplicity, robustness, and portability of sealed systems could enable a significant expansion of RPC technology.

6 Conclusion

In this work, the feasibility of sealed trigger Resistive Plate Chambers has been demonstrated through the development and characterization of medium- and large-area detectors operated without any external gas supply. The results show that sealed trigger RPCs can achieve performance comparable to conventional trigger RPCs in terms of efficiency, mean charge, streamer probability and dark count rate. In particular, the medium-area detector exhibits stable operation over extended periods approaching 2 years under exposure to natural cosmic radiation, with no observable performance degradation.

These findings validate the sealed approach as a viable alternative to gas-flow RPC systems, for low particle rate environments, significantly reducing system complexity and gas consumption. Despite the challenges identified in large-area implementations, particularly related to mechanical stability and environmental sensitivity, the results presented here establish a solid basis for further development of this technology towards robust, scalable, and environmentally sustainable gaseous detector systems.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Paul Colas, Commissariat à l'Energie Atomique et aux Energies Alternatives (CEA), France

Reviewed by: Alessandra Pastore, National Institute for Nuclear Physics of Bari, Italy

Luca Quaglia, University of Turin, Italy