Performance Reconstruction of Eco-Friendly Gas Mixtures for Improved Resistive Plate Chambers at GIF++ Using Geant4

A macroscopic reconstruction is developed to infer iRPC performance using Geant4 observables and one experimental anchor. The Geant4 energy deposition is used to estimate the primary ionization yield, while the efficiency turn-on is modeled through an induced-charge description encoded in an effective gain G(E). The absolute scale is fixed by calibrating the standard CMS mixture to its GIF++ efficiency curve and extracting macroscopic Townsend parameters (A,B). The same procedure is propagated to four alternative mixtures, including two HFO and CO2 eco-friendly blends, to reconstruct efficiency curves and working points, enabling detector mixture screening without microscopic transport inputs.

💡 Research Summary

The paper introduces a fully macroscopic method to predict the performance of improved Resistive Plate Chambers (iRPCs) without relying on detailed microscopic gas‑transport simulations. The approach is built on Geant4 simulations of the GIF++ irradiation environment (a 6 GeV muon beam combined with a 662 keV gamma field from a ¹³⁷Cs source) and on a single experimental reference: the efficiency curve of the standard CMS gas mixture (95 % C₂H₂F₄ + 4.5 % iC₄H₁₀ + 0.3 % SF₆) measured at GIF++.

From Geant4 the average energy deposited in each gas gap, ⟨E_dep⟩, is obtained. By dividing this value by an effective ionization energy W_mix (the volume‑weighted average of the ionization energies of the individual gas components) the primary ionization yield N₀(E)=⟨E_dep⟩/W_mix is calculated. Because W_mix depends only on mixture composition, N₀ is essentially independent of the applied electric field. Consequently, the field dependence of the detector response must be encoded in an effective gain G(E).

Using the Shockley‑Ramo theorem, the induced charge on the readout strips is expressed as Q_ind = e f_geom N₀ G(E), where f_geom is a purely geometrical coupling factor fixed by the double‑gap iRPC geometry. The experimentally measured efficiency ε_std(E) is first converted into a sensitivity S_std(E)=N_hit/N_tot, which together with N₀ yields Q_ind and thus G_std(E). This step provides an absolute gain scale anchored to the standard mixture.

The gain is transformed to an effective first Townsend coefficient α_eff(E) = (1/d_gap) ln G(E). The macroscopic Townsend law α_eff(E)=A p exp(−B p/E) is then fitted to the reconstructed α_eff, giving the parameters A and B for each gas mixture. For the standard mixture the fit yields A≈270 cm⁻¹ atm⁻¹ and B≈63 kV/(cm·atm), establishing the reference.

The same reconstruction pipeline is applied to four additional mixtures: two CO₂‑based blends (Mix I and Mix II) and two eco‑friendly HFO1234ze/CO₂ blends (ECO1 and ECO2). The CO₂ blends show A and B values very close to the standard mixture, resulting in similar gain curves and working voltages (≈7.0 kV, electric field ≈50 kV/cm for 95 % efficiency). In contrast, the eco‑friendly blends exhibit significantly larger B values (≈67–71 kV/(cm·atm)), which translates into a delayed gain onset and higher required operating fields (≈58.5 kV/cm for ECO1 and ≈55 kV/cm for ECO2). Consequently, their working voltages shift upward to ≈8.2 kV (ECO1) and ≈7.7 kV (ECO2).

The reconstructed sensitivities S(E) and efficiencies ε(HV) for all mixtures match the measured GIF++ data for the standard gas and provide reliable predictions for the eco‑friendly blends. The method therefore enables rapid screening of new gas mixtures without needing detailed electron transport data (e.g., drift velocities, diffusion coefficients, attachment rates), which are often unavailable for novel HFO‑based gases.

Key advantages of the approach are: (1) it requires only one experimental efficiency curve (the standard mixture) to calibrate the absolute gain scale; (2) it leverages Geant4’s realistic modeling of the mixed radiation field and detector geometry; (3) it yields macroscopic Townsend parameters that can be propagated to any mixture sharing the same geometry. Limitations include the assumption of field‑independent primary ionization, the use of a fixed charge threshold to define detection (ignoring pulse shape, time‑over‑threshold, and cluster size), and the fact that the macroscopic A and B parameters cannot capture all microscopic processes (e.g., electron‑ion recombination, space‑charge effects). The authors suggest that future work could incorporate microscopic codes such as HEED or GARFIELD++ to refine N₀(E) and to model threshold behavior more accurately.

In summary, the paper presents a novel macroscopic reconstruction framework that combines Geant4‑derived primary ionization with experimentally anchored gain and Townsend parametrization. This framework successfully predicts the efficiency turn‑on and working points of both conventional and eco‑friendly gas mixtures for iRPCs, offering a practical tool for the design and optimisation of RPC detectors in the high‑luminosity LHC era.