RPC EcoGas@GIF++ Collaboration

• RPC EcoGas@GIF++ Collaboration is a multi-experiment program dedicated to qualifying eco-friendly gas mixtures for RPC detectors under LHC conditions.
• It integrates diverse detector technologies and standardized beam tests at CERN’s GIF++ facility to assess performance, rate capability, and ageing under irradiation.
• The initiative addresses environmental challenges by comparing HFO-based mixtures with standard gases to reduce GWP while maintaining operational stability.

The RPC EcoGas@GIF++ Collaboration is a multi-experiment research and development program devoted to the qualification of low-environmental-impact gas mixtures for Resistive Plate Chambers under realistic LHC operating conditions. Centered on the CERN Gamma Irradiation Facility (GIF++), it brings together RPC detectors from ALICE, ATLAS, CMS, LHCb/SHiP, and the CERN EP-DT group in a common framework for beam tests, irradiation campaigns, and long-term ageing studies, with the specific objective of replacing or reducing the use of high-GWP fluorinated gases while preserving avalanche-mode performance, rate capability, and operational stability (Quaglia, 2022, Quaglia et al., 2023, Ahmadi et al., 6 Oct 2025).

1. Origins and institutional scope

The collaboration emerged from an RPC eco-gas effort that started in 2018 between the ALICE, ATLAS, and CMS RPC groups and the CERN gas team, with SHiP joining in 2019; later publications describe the RPC EcoGas@GIF++ Collaboration as created in 2019 (Ferretti, 2021, Quaglia et al., 4 Feb 2025). Its institutional structure is explicitly cross-experimental: ALICE, ATLAS, CMS, CERN EP-DT, and SHiP/LHCb contribute detectors, electronics, and gas-system expertise, while the program is also framed within the European AIDAinnova project, Work Package 7.2, on multigap RPCs for fast timing and eco-friendly gas mixtures for RPCs (Quaglia et al., 2023).

Its mandate is both practical and comparative. Collaboration papers define it as a common effort to identify and characterize eco-friendly RPC gas mixtures, perform shared beam and irradiation tests under controlled conditions, and evaluate performance and ageing under LHC-relevant backgrounds, with particular emphasis on mixtures based on hydrofluoro-olefins and diluents such as $CO_2$ (Quaglia et al., 2022). This common structure is essential because the participating detectors are technologically heterogeneous—single-gap and double-gap layouts, 1 mm to 2 mm gas gaps, different bakelite or HPL implementations, and different front-end chains—and the collaboration’s purpose is not restricted to one detector family but to the broader RPC ecosystem used at the LHC and in future experiments (Quaglia, 2022, Costa et al., 19 May 2025).

A distinctive feature of the collaboration is that it treats eco-gas qualification as a systems problem rather than as a mixture scan in isolation. The gas mixture, electrode resistivity, front-end threshold, rate-induced voltage drop, industrial reproducibility, and accumulated charge are all measured in a common environment. This suggests a methodological shift from experiment-specific gas studies toward a shared reference infrastructure for RPC sustainability and longevity.

2. Environmental and regulatory rationale

The collaboration is rooted in the environmental burden of the standard RPC gases. In the ALICE forward muon spectrometer, the 72 RPCs are operated in avalanche mode with a mixture of $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$, while the standard ATLAS/CMS reference mixture used in many GIF++ studies is $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$ (Quaglia et al., 2022). In these mixtures, $C_{2}H_{2}F_{4}$ acts as the main electronegative bulk gas, $i\text{-}C_{4}H_{10}$ as a quencher, and $SF_6$ as a strong electron-attaching additive for streamer suppression and field stabilization (Quaglia et al., 2022).

The environmental problem arises from the high Global Warming Potential of the fluorinated components. Collaboration papers quote $\mathrm{GWP}(C_{2}H_{2}F_{4}) \simeq 1400$ and $\mathrm{GWP}(SF_{6}) \simeq 22800$, with $C_{2}H_{2}F_{4}$ providing the dominant contribution to the mixture GWP because of its large fraction; by contrast, the candidate HFO gas $C_{3}H_{2}F_{4}$ is quoted with GWP around $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$0 to $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$1 (Quaglia et al., 2022, Quaglia et al., 4 Feb 2025). In ALICE-specific accounting, the effective GWP of the standard MID mixture is 1351, and during Run 1 the ALICE RPCs alone accounted for 12.5% of CERN’s total gaseous-detector greenhouse-gas emissions; when similar RPC systems in CMS and ATLAS are included, the fraction rises to 75%, with more than 90% of the contribution coming from R134a (Ferretti, 2021).

The regulatory trigger is the European F-gas phase-down. EU rules impose a progressive reduction in the production and use of fluorinated greenhouse gases, with the explicit objective of reducing overall F-gas emissions by about two thirds in 2030 with respect to 2014; even where research activities remain formally exempt, the collaboration’s papers emphasize the indirect effects on price, supply, and institutional policy at CERN (Quaglia et al., 2022). This is why the collaboration does not treat eco-gas work as optional optimization but as a strategic requirement for long-term detector operation.

The term “eco-friendly,” as used in the collaboration, is therefore relative rather than absolute. ECO2 and ECO3 reduce the mixture GWP strongly, but they are not completely F-gas-free because they still retain $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$2 (Quaglia et al., 2022, Abbrescia et al., 2023). Later summaries also note the debated question of HFO-1234ze degradation to trifluoroacetic acid in the atmosphere, while stating that current evaluations suggest negligible impact at the expected usage scale (Abbrescia et al., 2023). A common misconception is thus that low-GWP automatically implies a complete elimination of fluorinated components or of ageing risk; the collaboration’s results do not support that simplification.

3. GIF++ infrastructure and common methodology

GIF++ is the collaboration’s central experimental environment. In EcoGas@GIF++ papers it is described as combining a high-activity $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$3Cs gamma source—typically quoted as 12.5 TBq in the collaboration’s beam-test literature—with adjustable attenuation filters and dedicated SPS muon-beam periods, allowing controlled superposition of background irradiation and tracking-quality muon triggers (Quaglia et al., 2022, Quaglia et al., 2024). The source intensity can be varied through lead filters, and the facility is instrumented so that detectors can be installed at different distances, notably around 3 m and 6 m from the source, giving access to experiment-relevant background gradients (Quaglia, 2022, Abbrescia et al., 2023).

This setup supports the collaboration’s two-step methodology: first, determine baseline detector performance for each gas mixture; second, expose detectors to sustained irradiation and follow the evolution of current, efficiency, and related observables with integrated charge (Quaglia et al., 2022). In practice, beam periods are used for efficiency, cluster-size, timing, and charge-spectrum measurements, while source-on runs accelerate accumulated charge and reveal rate-capability limitations.

Across the collaboration, high voltage is normalized to environmental conditions through the effective-voltage correction

$$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$4

with reference values depending on the setup and experiment (Gul et al., 2016, Costa et al., 19 May 2025). Efficiency curves are then typically parameterized with a sigmoid,

$$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$5

from which a knee voltage and working point are derived (Gul et al., 2016, Costa et al., 19 May 2025). This common parametrization is important because it lets the collaboration compare chambers with different geometries and front-end thresholds on a consistent basis.

Long-duration operation also required a reusable slow-control layer. In the CMS GIF++ program, which became an important infrastructure reference for later EcoGas studies, a dedicated Detector Control System was built on WinCC-OA 3.11 with the JCOP framework, using CAEN SY1527 and EASY hardware, OPC communication for HV/LV and ADCs, and DIP links to the central GIF++ systems (Gul et al., 2016). This provided channel-level monitoring of HV, current, gas flow, temperature, pressure, humidity, source state, and attenuator settings, and it enabled unattended HV scans and stability runs over long periods. For EcoGas@GIF++, this kind of architecture functions as the slow-control backbone of comparative irradiation campaigns rather than as an auxiliary subsystem.

4. Detector portfolio and gas-mixture programme

The collaboration’s detector ensemble is deliberately heterogeneous. Collaboration papers list an ALICE single-gap 2 mm bakelite RPC, an ATLAS single-gap 2 mm prototype, the CMS RE11 double-gap 2 mm + 2 mm chamber, an EP-DT single-gap 2 mm RPC, an LHCb/SHiP single-gap 1.6 mm RPC, and, in later studies, a BARI 1 mm single-gap chamber and industrial 1 mm-gap HPL chambers built for HL-LHC-grade qualification (Quaglia, 2022, Quaglia et al., 2023, Costa et al., 19 May 2025). The readout ecosystem is similarly diverse: FEERIC front-end boards and TDCs for ALICE and SHiP/LHCb, CMS standard front-end chains for RE11, and waveform digitizers for ATLAS and EP-DT (Quaglia et al., 2024, Ahmadi et al., 6 Oct 2025).

The gas strategy centers on replacing $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$6 with HFO-1234ze diluted by $$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$7. In the ALICE-focused GIF++ studies, the standard reference mixture was

$$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$8

while the two main collaboration mixtures were

$$89.7\%\ C_{2}H_{2}F_{4} + 10\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$9

and

$$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$0

both with total GWP of about 230 (Quaglia et al., 2022). In broader collaboration scans, these same compositions also appear as MIX5 and MIX3, respectively, within a larger family ranging from 0% to 40% HFO and 95% to 55% $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$1 (Quaglia et al., 4 Feb 2025).

The logic of this program is explicit in the papers. HFO-1234ze is chemically similar to $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$2 but has a smaller effective first Townsend coefficient; a one-to-one substitution would push avalanche working voltages above 15 kV, which is incompatible with existing infrastructure and with detector breakdown limits (Quaglia et al., 2022, Ferretti, 2021). The role of $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$3 is therefore to dilute HFO and lower the working voltage, while small fractions of $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$4 and $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$5 are adjusted to control quenching and streamer probability (Quaglia et al., 2022, Terlizzi, 2022).

Before ECO2 and ECO3 were selected for collaborative GIF++ studies, the ALICE cosmic-ray program had already identified two useful benchmarks that illustrate the trade-off structure: $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$6, with GWP $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$7 and a working-voltage increase of about 1 kV but higher streamer fraction; and $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$8, with GWP $$95.2\%\ C_{2}H_{2}F_{4} + 4.5\%\ i\text{-}C_{4}H_{10} + 0.3\%\ SF_{6}$$9, a larger HV shift, and streamer probability closer to the ALICE standard (Ferretti, 2021). This trade-off between GWP, working voltage, and large-signal suppression recurs throughout the EcoGas@GIF++ programme.

5. Performance findings under beam and irradiation

The collaboration’s performance studies establish a consistent qualitative picture across detectors. In the ALICE-like 50 $C_{2}H_{2}F_{4}$0 50 cm$C_{2}H_{2}F_{4}$1 GIF++ chamber, source-off efficiency curves for ECO2 and ECO3 are shifted to higher voltages relative to the standard mixture, with the working-point shift reported as about 1.2 kV for ECO2 and about 0.4 kV for ECO3 (Quaglia et al., 2022). Under $C_{2}H_{2}F_{4}$2 backgrounds of 100 and 300 Hz/cm$C_{2}H_{2}F_{4}$3, the efficiency curves shift further to higher voltages and the plateau efficiency decreases slightly, reflecting rate-induced voltage loss in the resistive electrodes; at 100 Hz/cm$C_{2}H_{2}F_{4}$4, which is close to the expected hit rate on ALICE RPCs with a safety factor of two, the performance remains acceptable (Quaglia et al., 2022). A central quantitative concern is that the absorbed current density for ECO2 and ECO3 is roughly doubled with respect to the standard gas at comparable efficiency (Quaglia et al., 2022).

Signal-shape studies sharpen that picture. In waveform-based measurements, the working point increases by roughly 1 kV for every additional 10% of HFO in the mixture, while higher HFO fractions improve charge control and reduce the fraction of large signals (Quaglia et al., 4 Feb 2025). At their working points, HFO-based mixtures show average total charge per gamma hit typically about 1.5 to 2 times higher than the standard mixture, a result that directly links eco-gas operation to faster integrated-charge accumulation (Quaglia et al., 4 Feb 2025). In ATLAS digitizer data, the eco-mixtures also show broader charge distributions than the standard gas, which collaboration papers relate to the larger $C_{2}H_{2}F_{4}$5 content (Ahmadi et al., 6 Oct 2025).

Cross-experiment high-rate tests confirm that these are not ALICE-specific features. In the multi-detector GIF++ scans, eco-mixtures generally preserve plateau efficiencies above 95% without irradiation and remain above 90% under backgrounds reaching the $C_{2}H_{2}F_{4}$6 scale, though with a larger loss than the standard gas (Abbrescia et al., 2023). Current densities at the efficiency knee increase roughly by a factor of two for single-gap chambers operated with ECO2 or ECO3 compared with the standard mixture, whereas cluster sizes remain close to the standard-gas values and tend to decrease slightly as irradiation increases (Ahmadi et al., 6 Oct 2025, Abbrescia et al., 2023). When efficiency curves are replotted against the effective gas voltage $C_{2}H_{2}F_{4}$7, the source-on curves largely collapse onto a common trend, indicating that the dominant rate effect is the ohmic drop across the electrodes rather than a change in intrinsic gas amplification (Abbrescia et al., 2023).

A frequent misunderstanding is that “basically similar” performance means a drop-in replacement. The collaboration’s measurements do not support that interpretation. Eco-mixtures can reproduce the essential operational envelope of standard RPC gases, but they do so with higher working voltages, higher current densities, and a narrower safety margin against large-signal growth above the working point (Quaglia et al., 2022, Quaglia et al., 4 Feb 2025).

6. Ageing programme, open issues, and later extensions

Ageing is the collaboration’s decisive qualification axis. Its long-term methodology combines continuous irradiation at fixed corrected voltage with periodic source-off scans, using the integrated charge density as the primary ageing variable:

$C_{2}H_{2}F_{4}$8

or, equivalently, $C_{2}H_{2}F_{4}$9 when expressed in current density (Quaglia et al., 2022, Ahmadi et al., 6 Oct 2025). Over time, the collaboration monitors efficiency, cluster size, dark current, noise rate, and the shift of the working point under a fixed background (Quaglia et al., 2022).

Early ALICE-specific ageing results already showed why this programme was necessary. Under the ECO1 mixture, an ALICE RPC accumulated 22 mC/cm$i\text{-}C_{4}H_{10}$0 and then exhibited a significant increase of dark current at the estimated working point, together with an approximately linear sub-threshold current component that was explicitly interpreted as inconsistent with a simple localized “tip” effect and suggestive of a more complex ageing mechanism (Ferretti, 2021). The same study concluded that in-depth ageing investigations were mandatory (Ferretti, 2021).

Later cross-experiment campaigns substantially expanded the evidence base. In the long-term irradiation program running from July 2022 to July 2025, mainly with ECO2, collaboration papers report that plateau efficiencies often remained above 95% while dark and ohmic currents showed a characteristic evolution: after integrated charges of about 50–100 mC/cm$i\text{-}C_{4}H_{10}$1, dark currents usually started to increase, and working points shifted to higher voltages for several detectors (Ahmadi et al., 6 Oct 2025). For ALICE, comparisons between 2022 and 2023 showed stable plateau efficiency but a clear shift of the efficiency curve to higher voltage, both with source off and under irradiation, after about 80 mC/cm$i\text{-}C_{4}H_{10}$2 (Ahmadi et al., 6 Oct 2025). These observations are consistent with the resistive nature of the electrodes: higher currents imply larger voltage drops and thus higher required applied voltages (Ahmadi et al., 6 Oct 2025).

The collaboration’s open questions are correspondingly specific. They include long-term stability and ageing under the higher currents drawn by HFO/$i\text{-}C_{4}H_{10}$3 mixtures, gas-material compatibility of bakelite surfaces and linseed-oil coatings, rate capability at several hundred Hz/cm$i\text{-}C_{4}H_{10}$4 and above, CERN safety and flammability constraints, and the industrial supply chain for HFO mixtures at detector scale (Quaglia et al., 2022, Ahmadi et al., 6 Oct 2025). The papers do not report catastrophic degradation as a general outcome, but neither do they present eco-mixtures as fully qualified for all HL-LHC operating horizons.

Two later developments broadened the collaboration’s scope. First, industrialization work at the Max Planck Institute for Physics, in partnership with MIRION and PTS, established a reproducible 1 mm-gap HPL RPC production and certification chain, culminating in a year-long GIF++ irradiation campaign for full-scale modules; this provided a well-characterized HL-LHC-grade benchmark platform for future eco-gas comparisons (Costa et al., 19 May 2025). Second, a Geant4-based macroscopic reconstruction framework was proposed to infer effective gain curves, macroscopic Townsend parameters, efficiency curves, and working points for alternative mixtures from simulated energy deposition plus one experimental anchor, thereby offering a screening tool for candidate mixtures before committing to extensive GIF++ beam time (Ramirez-Beltran et al., 15 Dec 2025).

Taken together, these results define the RPC EcoGas@GIF++ Collaboration not simply as a gas-optimization effort, but as the shared experimental and methodological infrastructure through which the LHC RPC community is attempting to reconcile regulatory pressure, greenhouse-gas reduction, detector performance, and multi-year operational stability.