First in-beam studies of a Resistive-Plate WELL gaseous multiplier

The THGEM electrode used in this work was 10$\times$10 cm² in size, manufactured by 0.5 mm diameter hole-drilling in a 0.8 mm thick FR4 plate, copper-clad on one side only. The holes were arranged in a square lattice with a pitch of 0.96 mm; 0.1 mm wide rims were chemically etched around the holes. The hole-diameter, electrode thickness and rim width were chosen based on previous optimization studies [34]. Based on the results in [32], the single-sided electrode was coupled to an 8$\times$8 matrix of 1$\times$1 cm² pads through a 10$\times$10 cm² Semitron ESD225¹¹1http://www.quadrantplastics.com static dissipative plastic plate machined to a thickness of 0.4 mm. The bulk resistivity of this material was measured in [32] to be 2$\times$10⁹ $\Omega$cm. Good electrical contact between the resistive plate and the readout pads is essential for efficient clearance of the avalanche electrons from the bottom of the resistive material. This was ensured as follows: the bottom side of the plastic was first painted uniformly with conductive silver paint (Demetron Leitsilber 200²²2http://www.tedpella.com/technote_html/16062_TN.pdf), then, a grid matching the readout pad structure was milled over its surface to avoid charge spread to neighboring pads. Each anode readout pad was glued with conductive epoxy (Circuit Works³³3https://www.chemtronics.com/descriptions/document/Cw2400tds.pdf) to the corresponding painted silver-pad on the resistive plate. Sections of the readout pad electrode and of the matching conductive pads painted on the resistive plate are shown in figure 2-a and  2-b respectively. On the basis of our previous experience with THGEM detector operation in neon mixtures, offering high attainable gas gain at relatively low operation potentials [35, 4], the detector was operated in $\mathrm{Ne/(5\%CH_{4})}$, at a gas flow of 25-50 cc/min. At atmospheric pressure and room temperature, the most probable value (MPV) of the Landau distribution for cosmic muons corresponds to a total number of $\sim$14 ionization electrons along their track in the present 5 mm drift gap; this has been estimated in the laboratory from the ratio between the mean value of the peak of ⁵⁵Fe x-rays and the MPV measured for cosmics, and is in accordance with the literature [36]. The electrodes were biased with individual HV power-supply CAEN A1833P and A1821N boards, remotely controlled with a CAEN SY2527 unit. The voltage and current in each channel were monitored and stored. All HV inputs were connected through low-pass filters. While the RPWELL bias with respect to the grounded anode ($\Delta$V_RPWELL) was varied throughout the experiment, the drift voltage was kept constant ($\Delta$V_drift= 250 V corresponding to a drift field of $\sim$0.5 kV/cm).

The external trigger system used for event selection is described in [37]; it comprised three 10$\times$10 cm² scintillators arranged in coincidence. The particle-tracking system, covering a total area of 6$\times$6 cm², comprised three MICROMEGAS detectors. The RPWELL chamber was placed between two tracker elements. The tracker signals, as well as those recorded from the 64 pads of the RPWELL detector were read out by an SRS readout system  [38] equipped with a single Front-end Concentrator and front-end APV25 hybrids [39]. The SRS data acquisition was triggered by the scintillators coincidence signal. For each triggered event, the amplitude of the analog waveform (sampled in 20 bins of 25 ns each) in each channel was stored; this occurred only for signals crossing a threshold set by a zero-order suppression requirement, as described in section 2.3. The mmDAQ⁴⁴4Developed by M.Z.D. Byszewski (marcin.byszewski@cern.ch) online data acquisition software was used to store the synchronized data on a PC for further analysis.

The APV25 chip [40], originally designed for the silicon tracking detectors in CMS, has high rate capability and low noise. It was operated with a 75 ns shaping time. Due to the low noise level and high sensitivity of the chip, gas gains of the order of few times 10³ are typically sufficient for efficient operation of the detector. As shown in [31, 32] the rise-time of a typical signal of single-sided THGEM detectors in a WELL configuration is $\sim$1-2 $\mu$s. It comprises a fast component (up to 100 ns in $\mathrm{Ne/(5\%CH_{4})}$) and a slow one arising from the motion of avalanche ions inside the hole; the latter carries typically about 80$\%$ of the total amplitude. Since the APV25 was operated with 75 ns shaping time, less than 20$\%$ of the total induced signal was collected, requiring the operation of the detector with gas gains of the order of 10⁴. In what follows we relate to the effective gain, which includes the effect of the short shaping time. This effective gain, estimated from the Landau MPV (section 2.1), is roughly an order of magnitude smaller than the gas gain.

The analysis framework, including the noise suppression and the clustering algorithms are described in details in [25]. A common zero-order suppression factor (ZSF) sets the threshold of all APV25 channels. For each event the signal in an individual channel is stored only if the following condition is fulfilled:

Here S is the sum of the ADC counts in all of the n_bins time bins after pedestal subtraction, and P_σ is the pedestal standard deviation.

The global detector efficiency was defined as the fraction of tracks matched to a cluster found not more than W [mm] away from the track trajectory in both x and y directions. The single-event multiplicity was defined as the number of pads in the matched cluster. The average pad multiplicity was defined as the average of the single-event multiplicity. A ZSF of 2 and cluster-track matching parameter W value of 15 mm were found optimal for both muons and pions. We emphasize that in the case of muons a value of W= 8 mm is sufficient for optimal performance. This point, as well as the optimization procedure are described in section 2.4. The detector’s discharge probability was defined as the number of discharges divided by the number of hits in the active region of the detector (i.e., in the total area covered by the crossing beam). The number of discharges was extracted directly from the power supply log files by counting the resulting spikes in the current monitor. The pion beam was narrower (a maximum area of 4$\times$3 cm²) than both the acceptance of the scintillators and the tracker detectors. Therefore, the number of pion hits in the active region of the detector was estimated as the number of events where the three scintillators fired in coincidence. Due to the low rate of the muon beam, only pion runs were used to estimate the discharge probability. One should note that pions are prone to induce highly-ionizing secondary events; thus our study yielded an upper limit of MIP-induced discharge probability.

The working point was adjusted to optimize the detector performance, targeting high detection efficiency at low average pad multiplicity. The latter is a requirement for particle counting, e.g. in a potential application of the RPWELL as a sampling element in Digital Hadron Calorimetry (DHCAL) [23]. The optimization was done in three steps using a set of measurements with $\sim$500 Hz/cm² wide (5$\times$5 cm²) muon beam and a $\sim$13000 Hz/cm² narrow (2$\times$2 cm²) pion beam. In both beams, only tracks hitting the detector in a 4$\times$4 cm² area around the beam center were considered.

Under a low-rate ($\sim$500 Hz/cm²) muon beam and nominal operation conditions, a global detection efficiency greater than 99$\%$ at an average pad multiplicity of 1.21 was recorded (figure 5-a). Under higher rate ($\sim$13000 Hz/cm²) pion beam and the same nominal operation conditions, the performance of the detector was similar; a global detection efficiency greater than 98$\%$ at an average pad multiplicity of 1.23 (figure 5-b). In figure 6 the detection efficiency (figure 6-a) and the average pad multiplicity (figure 6-b) as a function of the distance from the pad boundaries along one axis on the detector plane are shown for operation with $\sim$500 Hz/cm² muons. The detector was operated at the nominal operation conditions. While the local efficiency is uniform, higher multiplicity is seen to originate from muons traversing the detector close to the pad boundaries resulting in charge sharing between neighboring pads.

In [32], it was shown that the gain dependence on the counting rate was moderate for an RPWELL configuration similar to the one investigated here (0.6 mm thick ESD225 layer, bulk resistivity of $\sim$10⁹ $\Omega$cm). A gain drop of $\sim$30$\%$ was observed at a counting rate spanning over 4 orders of magnitude. This study was repeated here with a 5$\times$5 cm² muon beam, with the lowest flux of $\sim$500 Hz/cm², and with a narrow (2$\times$3 to 4$\times$3 cm²) pion beam, reaching 35000 Hz/cm². The Landau spectra measured at three different incoming particle fluxes are shown in figure 7. The detector was operated at the nominal operation voltage, $\Delta$V_RPWELL= 880 V. The spectra recorded at higher incoming particle flux are shifted towards lower charge values. The measured charge - estimated from the Landau MPV - is shown in figure 8-a as a function of the incoming particle flux. The global detection efficiency as a function of the incoming particle flux is shown, for the same set of measurements, in figure 8-b. A gain drop of $\sim$23$\%$ is noticeable over a counting-rate spanning over more than 2 orders of magnitude; this is in reasonable agreement with the gain drop reported in [32] at a similar gain for the same flux of soft x-rays. Over the same range of incoming particle flux, the global detection efficiency dropped by $\sim$5$\%$ from over 99$\%$ to $\sim$95$\%$ by increasing the flux from $\sim$500 Hz/cm² to 35000 Hz/cm² (figure 8-b). This drop in the detection efficiency could be avoided by setting the nominal operation voltage slightly higher than 880 V, still maintaining a discharge-free operation as shown in section 3.3 and 3.4.

The detected charge as a function of time is shown in figure 9 for measurements carried out with low-rate ($\sim$500 Hz/cm²) muons and high rate ($\sim$15000 Hz/cm²) pions. The detector was operated under the nominal operation voltage with muons while with pions the voltage was increased to $\Delta$V_RPWELL= 900 V (corresponding to an effective gain of $\sim$3000) to compensate for the slight gain drop with the high-rate pions (section 3.2). No significant gain variations were observed along the six hours of operation under the muon beam (figure 9-a) (RMS= 0.2 fC), and the two hours of operation under the pion beam (figure 9-b) (RMS= 0.1 fC). The measurement with muons showed a global detection efficiency of 99$\%$ at an average pad multiplicity of 1.2, while the pion beam showed a global efficiency of more than 98$\%$ at an average multiplicity of 1.25. As explained earlier, this small efficiency loss could be partially recovered selecting a higher matching parameter W, as also indicated by the slightly higher multiplicity.

As mentioned in section 2.3, a discharge was defined as an abrupt increase in the current supplied to the detector. Due to the low rate of the muon beam, the discharge probability was measured with pions. As an example, during the 1.5 hours of measurement presented in figure 9-b, no current activity was recorded on any of the detector electrodes. At an incoming particle flux of $\sim$15000 Hz/cm², this corresponds to discharge-free operation with over 10⁸ pions resulting in a discharge probability of less than 10^-8. It is worthwhile mentioning that under these conditions the detector had over 98$\%$ global detection efficiency. One should also note that pions are prone to induce highly-ionizing secondary events; thus this study yields an upper limit of discharge probability with MIPs. Similar behavior was recorded throughout longer periods, supporting the claim that the RPWELL is a discharge-free detector.