A resistive plate chamber using diamond-like carbon electrodes with current evacuation patterns

1 Introduction

An idea involving diamond-like carbon (DLC)-coated resistive plate chambers (RPCs) has been developed to improve tolerance to high rates (Wang et al., 2021; Bencivenni et al., 2023). The DLC coating represents a different approach to high-rate-capable RPC technologies than thin float glass (Weiping et al., 2014) and low-resistive glass developed for the CBM TOF system (Liu et al., 2020), as it enables a further reduction in surface resistivity and thus faster charge neutralization.

The authors of (Ieki et al., 2024) have been developing a high-rate-capable DLC-coated resistive plate chamber (DLC-RPC) with an ultra-low material budget for the MEG II experiment (Baldini et al., 2018). The detector is required to detect minimum-ionizing particle (MIP) positrons from muon decay with an efficiency of 90% in a high-intensity (4 MHz/cm²) and low-momentum (28 MeV/c) muon beam. Therefore, both an ultra-low material budget (below 0.1%X ₀) and high-rate tolerance (up to 4 MHz/cm²) are required. In terms of material budget, the concept also draws on a design developed at Kobe University ¹ that employed only DLC-coated polyimide foil as a low-material-budget electrode structure. Up to five polyimide foils can be stacked within 0.1% X.₀, corresponding to four active RPC gas gaps.

A 2 cm × 2 cm prototype DLC-RPC with a DLC layer of 50 MΩ/sq to 70 MΩ/sq and spacing pillars defining a gas gap (400 µm diameter, 384 µm thickness, and 2.5 mm pitch) was tested in a muon beam at the Paul Scherrer Institute (Ieki et al., 2024). The detector was operated with a single-gap configuration and a gas mixture of 94% R-134a (C₂H₂F₄), 1% SF₆, and 5% isobutane. This gas mixture was similar to those used in RPC systems of ATLAS (Boscherini, 2023) and CMS (Shopova et al., 2025). The detection efficiency for MIP positrons from muon decay was measured to be (55±3)% at a pulse-height threshold of 10 mV. Furthermore, the efficiency decreased to 42% in a 1 MHz/cm² muon beam with a momentum of 28 MeV/c.

The development toward the MEG II experiment represented the most demanding application of the DLC-RPC technology, and the application was ultimately abandoned because of concerns regarding detector stability in high-particle-flux environments. We are therefore extending the application of the DLC-RPC to beam-monitoring detectors that can be installed in a beam to measure the beam profile. Even for this application, achieving scalability and stable operation of the DLC-RPC remains challenging.

In high-particle-flux environments, large avalanche charges flowing through resistive electrodes cause a voltage drop, thereby causing a reduction in gas gain (Abbrescia, 2016; Aielli et al., 2016; Carboni et al., 2003). Therefore, the voltage drop must be suppressed to achieve adequate performance under such conditions. The voltage drop of the DLC-RPC followswhere V d r o p ( x , y ) is the magnitude of the voltage drop at position ( x , y ), Q m e a n is the average avalanche charge per incident particle, f ( x , y ) is the position-dependent particle rate per unit area, and ρ S is the surface resistivity of the electrode. The authors of (Ieki et al., 2024) analyzed the voltage-drop effect based on Equation 1 and demonstrated that this equation explains performance degradation in high-particle-flux environments.

∂ 2 ∂ x 2 + ∂ 2 ∂ y 2 V d r o p x , y = Q m e a n × f x , y × ρ S , (1)

The boundary conditions of Equation 1 are defined by conductive electrodes connected to the high-voltage supply, where there is no voltage drop. This implies that the voltage drop can be reduced by shortening the distance to the high-voltage supply. To achieve this in large-area DLC-RPCs, a segmented high-voltage supply is required. One approach is to implement strip-shaped current-evacuation patterns on the DLC layer, as proposed in (Bencivenni et al., 2023; Ieki et al., 2024).

We attempted to implement a strip-shaped current-evacuation pattern. In an initial attempt (Yamamoto et al., 2023), DLC-RPC electrodes with such a pattern were fabricated for the first time; however, poor-quality spacing pillars prevented stable operation. In a subsequent study (Takahashi et al., 2025), electrodes with higher-quality pillars were successfully fabricated by changing the pillar material. Measurements of the discharge-quenching capability demonstrated stable operation at a hit rate of O 1 k H z / m m 2. However, these measurements were limited to relatively low particle rates and short operating periods, leaving the detector stability at higher particle fluxes and over longer operating times unaddressed.

In this study, we extend these investigations by evaluating the discharge behavior and operational stability of additional DLC-RPC samples incorporating strip-shaped current-evacuation patterns at hit rates of up to O 10 k H z / m m 2. The goal is to establish empirical design guidelines for the current-evacuation pattern and to clarify its impact on discharge quenching (Section 4.1) and long-term operation (Section 4.2). Finally, possible directions for future development toward the intended beam-monitoring application are discussed in Section 4.3.

2 Designs of the DLC-RPC

2.1 Design concept

This section presents the design concept of the DLC-RPC. Details of the fabrication processes and design parameters for each component are provided in Section 2.2.

Figures 1a,b show a schematic and a cross-sectional view of the DLC-RPC design, respectively. The electrodes are composed of thin polyimide foils with sputtered DLC resistive layers, minimizing the overall material budget. Strip-shaped current-evacuation patterns and spacing pillars supporting the gas gap are formed on the DLC layer. Commercially available aluminum-coated polyimide foils are used for the readout electrodes by forming a strip shape through an etching process, while the DLC-RPC electrodes are simultaneously formed on the opposite surface. This design allows the outermost electrodes to integrate the readout and DLC-RPC layers without requiring a printed circuit board over the active area, further reducing the material budget.

Table 1 summarizes the material budget for each component of the active region of the DLC-RPC. For a single-layer DLC-RPC configuration, consisting of two polyimide foils, two DLC layers, and two aluminum readout pads, the material budget in the active area is approximately 0.034% X 0. In regions where spacing pillars and insulating covers are implemented, the material budget increases further.

TABLE 1. Material budget for each component of the DLC-RPC (Navas et al., 2024). The solder mask is not included because sufficient information on its material composition is not publicly available.

2.2 Prototype with the current-evacuation pattern

In 2024, a prototype DLC-RPC electrode with a current-evacuation pattern, as shown in Figure 2, was fabricated at the Micro-Pattern Technology workshop at CERN (Takahashi et al., 2025). The electrode had an active area of 3 cm × 3 cm and included a single strip-shaped current-evacuation pattern. The aluminum readout layer described in Section 2.1 was not implemented in this prototype; instead, an aluminum-coated polyimide foil was cut and placed on the electrode, as described in Section 3.1.

2.2.1 DLC layer and conductive layer

The DLC and conductive layers were formed by magnetron sputtering (Kelly and Arnell, 2000). DLC, chromium, and copper were sputtered onto the entire polyimide foil in the order shown in Figure 3a. This sandwich structure is motivated by the good adhesion of chromium to DLC. The typical thicknesses of the DLC, chromium, and copper layers are 0.1 µm, 0.05 µm and 10 μm, respectively.

The conductive layer (chromium and copper) was chemically etched and patterned into strips 0.1 mm wide, corresponding to the minimum achievable width, as shown in Figure 3b. Subsequently, the DLC layer was removed using a jet-scrubbing process, a form of abrasive blast cleaning (Raykowski et al., 2001), resulting in a DLC-coated area of 3.5 cm × 3.5 cm. The surface resistivity varied across samples, ranging from 6 MΩ/sq to 15 MΩ/sq.

2.2.2 Insulating cover

Photolithographic techniques (Okoroanyanwu, 2024) were used to form a protective cover over the conductive layer and the edges of the DLC (Figures 2, 3) using an insulating dry-film solder mask (Eternal Tech. Corp., DYNAMASK®). The solder mask was laminated onto a polyimide foil with the DLC/Cr/Cu electrode, and photolithography was subsequently performed. After development, only UV-exposed regions, which are the DLC/Cr/Cu pattern and the edges of the DLC, remained as the insulating cover, shown as green areas in Figure 2. This protection is required because electric-field distortions in these regions increase the probability of discharges, and bare conductors cannot suppress excessive avalanches.

Throughout this paper, the term “cover width” denotes the total width of the insulating cover, including the 0.1 mm-wide DLC/Cr/Cu strip underneath, rather than the distance from the strip edge to the cover edge. Widths of 0.2 mm, 0.3 mm, 0.4 mm, 0.6 mm and 1.0 mm were investigated in the present study. Microscopic observations indicated that the variation in cover width among the tested samples was within approximately 10 µm. For the 3 cm × 3 cm active region used in this study, the occupancy of the strip-shaped inactive region is approximately 0.7% for a cover width of 0.2 mm and 3.3% for 1.0 mm.

2.2.3 Spacing pillars

Spacing pillars were fabricated using the same techniques and materials as the insulating cover. Five 75 µm-thick solder-mask layers were stacked to form the pillars (Figure 4a). UV exposure was performed layer by layer to avoid undercuring due to the low transparency of the solder mask.

The fabricated spacing pillars had a diameter of 0.6 mm and a thickness of (365±5) µm (Figure 4b). They were arranged with a pitch of 2.5 mm (Figure 2). The corresponding area fraction for the 3 cm × 3 cm active region was approximately 4.9%.

3 Performance of the prototype

In this study, the detection efficiency and stability of the prototype DLC-RPC were measured using β rays from a ⁹⁰Sr source at a rate of O 1 k H z / m m 2 and O 10 k H z / m m 2. The higher hit rate corresponds to approximately 1 MHz/cm² in the locally irradiated region and exceeds the 10 kHz/cm² to 100 kHz/cm² range commonly investigated in other high-rate RPC studies (Liu et al., 2020).

3.1 Experimental setup

Figure 5 shows the experimental setup, which is very similar to that described in (Takahashi et al., 2025). The chamber has a window made of DLC-sputtered polyimide foil with a ground connection. A DLC-RPC gas gap was formed by placing the prototype electrode on the window. The chamber was filled with a gas mixture of 94% R-134a (C₂H₂F₄), 1% SF₆, and 5% isobutane. This gas mixture was adopted to allow a direct comparison with our previous DLC-RPC measurements.

In contrast to (Takahashi et al., 2025), strip-shaped conductors were implemented on the cathode-side electrode (Figure 6). They were manually fabricated using copper tape covered with Kapton tape and arranged parallel to the strip-shaped conductor on the prototype electrode. As described in Section 3.3, the pitch of the copper-tape conductors was varied for each measurement.

The signal readout system was the same as that described in (Takahashi et al., 2025). Inside the chamber, an approximately 3 cm × 3 cm aluminized polyimide foil was placed directly outside the anode-side polyimide foil of the RPC gap and connected to ground. Outside the chamber, a strip-shaped aluminized polyimide foil was installed on the cathode side as the signal-readout electrode. A trigger counter (a 5 mm × 5 mm × 5 mm plastic scintillator coupled to a SiPM) was installed downstream of the DLC-RPC, opposite to the ⁹⁰Sr source. The counter was positioned approximately 35 mm downstream of the active gap and aligned with the center of the active area.

Signals from the DLC-RPC were induced in the aluminized polyimide, amplified by a 38 dB amplifier (Cattaneo et al., 2014), and then fed into a DRS4 waveform digitizer (Ritt, 2004). Signals from the trigger counter were fed directly into the DRS4. The trigger threshold was set to 20 mV. No thermal-noise signals were observed on an oscilloscope at this threshold.

3.2 Efficiency measurement

First, the detection efficiency of the prototype DLC-RPC was measured using β rays from a ⁹⁰Sr source at a rate of O 10 k H z / m m 2. The β rays were collimated to a 2 mm-diameter beam and directed away from the current-evacuation pattern to avoid associated inefficiencies.

The pulse-height spectra of triggered events are shown in Figure 7a; the measurements were performed at 970 hPa. Events near 0 V arise from the inefficiency of the DLC-RPC. As the applied voltage increased, events with higher pulse heights became more frequent, and a plateau developed in the spectrum.

Figure 7b shows the efficiency curves at hit rates of O 10 k H z / m m 2 (this study) and O 1 k H z / m m 2 (Takahashi et al., 2025). For the former, the detection efficiency is defined as the fraction of triggered events with signals exceeding 30 mV, corresponding to an induced charge of approximately 36 fC on the readout strip. This threshold was determined by analyzing noise heights in an off-timing region. At this threshold, the noise fraction was estimated to be approximately 2.5%, although it includes a sub-percent contribution from signal pulses due to pile-up at the high hit rate. The applied voltages for the latter have been corrected for differences in atmospheric pressure, as the measurements were conducted in different environments.

Both curves reached approximately 60% at 2.63 kV, consistent with measurements using electrodes with 384 µm-thick pillars and no conductive pattern. No significant difference was observed between the two hit rates. The measured currents at 2.63 kV were 12 nA and 650 nA at O 1 k H z / m m 2 and O 10 k H z / m m 2, respectively, corresponding to an estimated voltage drop of less than 5 V based on (Aielli et al., 2016), indicating negligible voltage drop.

Although β rays from a ⁹⁰Sr source are not equivalent to MIPs and exhibit a broader energy spectrum (Dixon et al., 2016), the measured pulse-height spectra are comparable to those previously measured for MIP positrons from muon decay (Ieki et al., 2024). Therefore, while the present measurements do not directly determine the MIP detection efficiency, they provide a reasonable indication of the performance expected for MIPs in a detector with a similar gas gap configuration.

3.3 Discharge-quenching capability

Next, we focused on discharges at the edge of the insulating cover on the anode side (Figure 8) and investigated the quenching capability in this region by irradiating it with β rays from a ⁹⁰Sr source at a hit rate of O 10 k H z / m m 2.

The irradiation position was intentionally chosen near the anode-side strip structure to evaluate the discharge-quenching capability associated with the anode-side conductor. The cathode-side strip-shaped conductors were manually assembled using copper tape as described in Section 3.1, and operational instabilities caused by imperfections such as tape wrinkles were occasionally observed. Therefore, this measurement focused on the anode-side structure.

If a discharge occurs around the anode-side strip structure, the resistances under the insulating cover ( R a n o d e ) and between its edge and the cathode-side conductor ( R c a t h o d e ) are expected to contribute to quenching, as shown in Figure 8. In this study, R a n o d e and R c a t h o d e are defined as the “discharge-quenching capability” and can be approximated aswhere ρ S a n o d e ( c a t h o d e ) denotes the surface resistivity on the anode (cathode) side, and d a n o d e ( c a t h o d e ) is the distance between the corresponding conductor and the discharge location. In particular, d c a t h o d e is given by l p i t c h − d a n o d e, where l p i t c h is the conductor pitch (Figure 8).

R a n o d e ( c a t h o d e ) = ρ S a n o d e ( c a t h o d e ) × d a n o d e ( c a t h o d e ) , (2)

Table 2 summarizes the configurations for each measurement, and we determined which configurations could achieve the target operating voltage. As described in Section 2.2.2, prototype electrodes with different insulating-cover widths were used in the measurements. Electrodes with a greater resistivity than approximately 10 MΩ/sq were selected to ensure stable operation. The distance l p i t c h was set to 12.5 mm and approximately 14 mm by repositioning the cathode-side copper tape conductors. These two pitches were chosen to investigate the effect of the cathode-side current path length. The DLC-sputtered polyimide window installed on the cathode side (as described in Section 3.1), which had also been used in previous measurements (Takahashi et al., 2025), was replaced during the tests due to discharge-induced damage. The surface resistivities of the respective DLC layers were measured to be 10 MΩ/sq and 14 MΩ/sq.

TABLE 2. Configurations used in the discharge-quenching-capability study and the corresponding maximum applied voltages. Cover width, surface resistivity ρ S, and conductor pitch l p i t c h are measured values. R a n o d e and R c a t h o d e are calculated from the equation Equation 2.

Maximum applied voltages for each configuration are also shown in Table 2. Based on the efficiency curves shown in Figure 7B, all configurations reached operating voltages corresponding to efficiencies exceeding 50%. For cover widths of 0.2 mm and 0.3 mm, operation at the maximum applied voltage was sustained for only a few minutes due to discharges, whereas operation was maintained for approximately one hour for cover widths of 0.4 mm or wider.

For the 0.2 mm cover-width configuration, R c a t h o d e varied from 138 MΩ mm to 195 MΩ mm as the conductor pitch l p i t c h was changed; however, no significant improvement in the maximum applied voltage was observed. In contrast, increasing the cover width (corresponding to R a n o d e ≳ 2.2 MΩ mm) enabled longer-term operation at higher voltages.

3.4 Long-term stability

For configurations with a width of 0.4 mm or wider, stability during long-term operation was evaluated. Figure 9 shows the time evolution of the applied voltage and detector current for the 0.4 mm-width configuration, with β rays collimated to a diameter of 2 mm and irradiated onto the strip-shaped current-evacuation pattern at a hit rate of O 10 k H z / m m 2. The detector was operated at 2.625 kV and 2.63 kV and exhibited discharges after approximately 1.5 h, preventing further operation. The other configurations (with widths of 0.6 mm and 1.0 mm) also exhibited the same behavior after approximately 1 h–2 h at each maximum applied voltage.

Additional long-term operations were performed at a lower hit rate of O 1 k H z / m m 2 and reduced operating voltages (Figure 10). In these measurements, β rays with an approximate diameter of 1 cm were irradiated over a region including the strip structure. The applied voltages in Figure 10 were corrected for atmospheric pressure to match the atmospheric conditions of the test shown in Figure 9.

Figure 10a shows operation with a cover width of 0.4 mm at 2.55 kV. The first discharge occurred after 16 h, and continued operation became impossible after approximately 32 h. Figure 10b shows operation with a cover width of 0.6 mm at 2.45 kV and 2.48 kV. At 2.45 kV, the detector operated for 72 h without discharges. The voltage was then increased to 2.48 kV, after which discharges prevented further operation within 12 h. Operation was terminated by damage to the spacing pillars or the edges of the insulating cover, as shown in Figure 11, with the damage localized at the β-ray irradiation point.

4 Discussion

4.1 Discharge-quenching capability

As mentioned in Section 3.3, a clear difference in operating time was observed between samples less than 0.3 mm wide and those 0.4 mm wide or wider. On the other hand, for samples with a width of 0.4 mm or wider, no significant differences were observed in the maximum applied voltage or the maximum operating time.

A comparison with the results reported in (Takahashi et al., 2025) suggests that higher-particle-flux environments require greater discharge-quenching capability. In (Takahashi et al., 2025), higher operating voltages were achieved using an electrode with a cover width of 0.2 mm (corresponding to R a n o d e ≈0.6 MΩ mm) at a hit rate of O 1 k H z / m m 2. In contrast, in this study, increasing the hit rate to O 10 k H z / m m 2 rendered the same configuration insufficient to sustain the operating voltage.

These observations suggest that the discharge-quenching capability is more strongly influenced by the anode-side resistance R a n o d e than by the cathode-side resistance R c a t h o d e. This behavior can be understood from the difference in charge transport between avalanche electrons and ions: electrons, owing to their high mobility, contribute to fast processes such as discharges, whereas ions, with much lower mobility, are expected to have a smaller effect.

It should be noted that clear sample-to-sample reproducibility could not be confirmed in the present study. This is likely because the experimental setup was assembled manually, and variations in assembly quality affected detector stability. Therefore, the suggested cover width should be regarded as an empirical design guideline rather than a strict threshold.

Based on the available results, for DLC-RPC electrodes with a surface resistivity of 10 MΩ/sq and a strip-shaped conductor width of 0.1 mm, a cover width of approximately 0.4 mm is suggested to sustain operation at a hit rate of O 10 k H z / m m 2. This value should be adjusted according to the particle flux in the operating environment.

4.2 Long-term stability

As presented in Section 3.4, at a hit rate of O 10 k H z / m m 2 and an operating voltage of approximately 2.63 kV (60% efficiency), operation was sustained for only 1 h–2 h. At a hit rate of O 1 k H z / m m 2, operation at 2.55 kV was terminated by discharges after approximately 32 h, whereas operation at 2.45 kV continued for 72 h without discharges. The observed failure mode was localized discharge-induced damage to spacing pillars or the edges of the insulating cover on the anode-side strip-shaped current-evacuation pattern.

For comparison, the study reported in (Takahashi et al., 2024) demonstrated stable operation for approximately 1 week under X-ray irradiation at 5.7 × 10⁹ counts/sec and at an operating voltage corresponding to a detection efficiency of approximately 60%. The electrode used in that study did not incorporate a current-evacuation pattern, had a higher surface resistivity of 40 MΩ/sq, and used spacing pillars made of a different material (DuPont, Pyralux®). These differences indicate that several aspects of the detector design may affect long-term stability, although the mechanism responsible for the observed instabilities remains unclear. Several possible explanations for the observed instabilities were therefore considered, including ionic contamination, gradual material degradation due to accumulated damage from large avalanche charges, and accidental excessive avalanches.

Ionic contamination is typically accompanied by an increase in dark current and/or a degradation of detector performance over time (Petriş et al., 2024; Takahashi et al., 2024). Since neither of these signatures was clearly observed in the present study, ionic contamination is unlikely to have been the dominant cause of the observed instabilities.

The available data are insufficient to determine whether gradual material degradation, accidental excessive avalanches, or another mechanism is primarily responsible for the observed instabilities. Additional measurements under a wider range of operating conditions will therefore be required.

Nevertheless, the results presented in Section 3.4 indicate that lower operating voltages are associated with longer operating times and lower discharge rates. Reducing the operating voltage is therefore a possible approach to improving long-term stability.

The implications of these observations for the intended beam-monitoring application remain to be established, as the required operational lifetime has not yet been fully determined. Dedicated long-term tests under realistic operating conditions will therefore be required.

4.3 Future development toward beam-monitoring applications

The present study provided empirical guidelines for the electrode structure and operational conditions of DLC-RPCs with a current-evacuation pattern, as discussed in Sections 4.1, 4.2, and identified operational limitations that must be addressed before the suitability of the detector concept for beam-monitoring applications can be established. This section discusses possible directions for such future developments.

4.3.1 Surface resistivity

An increase in the surface resistivity to around 30 MΩ/sq is one possible approach to improving long-term stability, based on our previous experience with stable operation at an anode surface resistivity in the range of 20 MΩ/sq to 30 MΩ/sq (Ieki et al., 2024). According to Equation 1, the voltage drop is proportional to the surface resistivity. Therefore, increasing the surface resistivity is expected to increase the voltage drop, thereby reducing high-rate capability.

In this study, the voltage drop was estimated to be approximately 5 V at a hit rate of O 10 k H z / m m 2, as described in Section 3.2. Therefore, increasing the surface resistivity from 10 MΩ/sq to 30 MΩ/sq would increase the voltage drop to approximately 15 V, which is expected to have only a limited impact on detector performance under the conditions investigated in this study. However, the reduced rate capability associated with increased surface resistivity should be taken into account when considering applications in substantially higher-rate environments.

For the intended beam-monitoring applications, particle counts measured by the DLC-RPC can be corrected for detector efficiency, provided that the efficiency remains stable and quantitatively understood. The reduction in rate capability associated with higher surface resistivity may therefore be acceptable, depending on the operational requirements of the application.

4.3.2 Operating gas mixture

The operating gas mixture of 94% R-134a (C₂H₂F₄), 1% SF₆, and 5% isobutane was not optimized in terms of detector stability. Isobutane acts as a quencher of UV photons, and SF₆ suppresses the streamer transition (Camarri et al., 1998). Therefore, the influence of the operating gas mixture on detector stability should be investigated systematically.

We have performed a limited number of exploratory tests with increased isobutane or SF₆ concentrations of up to 15% and 10%, respectively. However, no clear improvement in detector stability was observed with these gas mixtures under the limited conditions investigated. Since these tests were not conducted systematically, the available data are insufficient to determine an optimal gas mixture. A systematic investigation of the gas composition therefore remains a topic for future study.

4.3.3 Operation with multi-layer configuration

Operation with multiple gas gaps is being considered as a possible future option, although its adoption will depend on both detector requirements and technical feasibility.

A multi-layer configuration could compensate for the reduction in detection efficiency associated with low-voltage operation by increasing the number of active gas gaps. However, stacking thin-film electrodes is technically challenging. In fact, the previous study (Ieki et al., 2024) reported a limitation in operating voltage due to discharges associated with the non-flatness of the thin-film electrodes. Furthermore, the currently intended beam-monitoring application does not require a very high absolute detection efficiency, since particle counts can be corrected for detector efficiency when the detector response is stable and well characterized.

For these reasons, it is currently unclear whether the final detector concept should employ a single-gap or multi-gap configuration, and further studies will be required to evaluate both the practical benefits and the technical feasibility of multi-layer operation.

5 Conclusion

The DLC-RPC, a resistive plate chamber with diamond-like carbon electrodes, has been developed for applications requiring a high-rate capability and an ultra-low material budget and is currently being developed for the intended beam-monitoring application. A previous study (Takahashi et al., 2025) demonstrated the feasibility of a DLC-RPC incorporating a current-evacuation pattern and its operation under β rays at a hit rate of O 1 k H z / m m 2. The present work extends this study by investigating discharge behavior and operational stability at hit rates up to O 10 k H z / m m 2, corresponding to approximately 1 MHz/cm2 in the locally irradiated region, using additional electrode configurations.

The tested DLC-RPC electrodes served as the anode-side electrodes and employed a 0.1 mm-wide strip-shaped current-evacuation pattern covered by insulating layers with cover widths of 0.2 mm–1.0 mm. Additional strip-shaped conductors were manually installed on the cathode side at distances of approximately 12.5 mm and 14 mm from the anode-side strip. The measurements were performed with a (365 ± 5) µm single-gap DLC-RPC and a gas mixture of 94% R-134a (C₂H₂F₄), 1% SF₆, and 5% isobutane. A detection efficiency was measured to be 60% for β rays from a ⁹⁰Sr source at a hit rate of O 10 k H z / m m 2, consistent with the performance measured at a hit rate of O 1 k H z / m m 2. Although the present measurement was performed using a ⁹⁰Sr β source rather than MIPs, the measured efficiency was comparable to that previously obtained with a single-gap DLC-RPC having a similar gas-gap thickness for MIP positrons from muon decay (Ieki et al., 2024).

The discharge-quenching capability was investigated by measuring the maximum applied voltage using electrodes with different cover widths. At a hit rate of O 10 k H z / m m 2, longer operating times were obtained for cover widths of 0.4 mm or wider with a surface resistivity of approximately 10 MΩ/sq, whereas no significant correlation was observed between the maximum applied voltage and cover width. These results suggest that the discharge-quenching capability is more strongly influenced by the anode-side resistance R a n o d e than by the cathode-side resistance R c a t h o d e, and that larger cover widths are required at higher particle fluxes.

Long-term operation was limited by sudden discharges, with damage localized at the β-ray irradiation point. At a hit rate of O 10 k H z / m m 2 and an operating voltage of 2.63 kV, corresponding to a detection efficiency of 60%, operation continued for 1 h–2 h. At a lower hit rate of O 1 k H z / m m 2, operation at 2.55 kV was terminated by discharges after approximately 32 h, and operation at 2.48 kV was terminated within 12 h after the voltage increase. In contrast, no discharges were observed during 72 h of operation at 2.45 kV, corresponding to a detection efficiency of approximately 45%. These results suggest that reducing the operating voltage is a possible approach to achieving long-term stability under irradiation conditions comparable to O 1 k H z / m m 2.

The present study identified several operational limitations that must be addressed before the suitability of the detector concept for the intended beam-monitoring application can be established. Dedicated long-term tests at lower operating voltages, as suggested by the present measurements, will be required under realistic beam-monitoring conditions. In addition, there remains room for optimization of the operating gas mixture. The final detector configuration has not yet been established. In particular, the adoption of a multi-layer configuration will require further evaluation of both its potential to compensate for the efficiency reduction at lower operating voltages and the technical challenges associated with stacking thin-film electrodes.

In memoriam

The authors wish to acknowledge the late Atsuhiko Ochi (Department of Physics at Kobe University) for his contributions as a collaborator in this research project and as the originator of the DLC-RPC concept. He made significant contributions to the acquisition of funding and to discussions concerning the prototype fabrication described in this work.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. ChatGPT, GPT-5.3 was used to improve the grammar and wording of the manuscript written by the authors.

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: Diego Gonzalez-Diaz, University of Santiago de Compostela, Spain

Reviewed by: Gabriella Pugliese, National Institute for Nuclear Physics of Bari, Italy

Marco Poli Lener, Istituto Nazionale di Fisica Nucleare di Perugia, Italy

Alessandro Paoloni, National Laboratory of Frascati (INFN), Italy

Footnotes

K. Ogawa, presentation at Mini-Work Shop and Seminar on Resistive Plate Chamber at the University of Tsukuba, 2017 (materials no longer available online). This project involved developing a fast-timing RPC with a thin gas gap, taking advantage of the surface flatness and low surface resistivity of DLC-coated polyimide foil.