Particle detectors based on glass: toward the Multigap Resistive Plate Chambers

Abstract

Glass found a large application in particle physics, and it can be considered as a key turning point in the technology of gaseous detectors, in particular in the development that lead from parallel plate chambers to resistive plate chambers and multigap resistive plate chambers (MRPCs). One of the crucial parameters to take into account when building such a detector is the resistivity of the material used to collect the signal produced by a particle crossing the detector. In this paper a (not comprehensive) history of the gaseous detectors will be presented, in order to describe the use of glass in the MRPC, including a few examples of experiments making use of this technology.

1 Introduction

Multigap resistive plate chambers (MRPCs) are a type of particle detector used in high-energy physics experiments. They are widely used in various experiments due to their excellent time resolution, high efficiency, and low cost. Before describing the use of glass in the MRPC detectors, a brief explanation of the processes involved in the particle detection by gaseous detectors will be given [1].

1.1 Working principles of gaseous detectors

The working principle of gaseous detectors is the creation of charges by ionization and the collection of the produced signal performed in several ways depending on the type of detector used. Avalanche formation in gaseous detectors is a fundamental process that occurs when charged particles pass through a gas medium and create ionization. The avalanche formation process in gaseous detectors can be described in the following steps:

• 1. Ionization: a charged particle passes through the gas medium and interacts with the gas atoms or molecules, by losing a fraction of its energy. This interaction can lead to the ejection of electrons from the gas atoms or molecules, creating pairs of positively charged ions and free electrons;

• 2. Electric field: the gaseous detector has an electric field applied across the gas volume, typically achieved by applying a high voltage between two electrodes, whose geometry depends on the type of detector. This electric field causes the newly formed free electrons to accelerate toward the anode and the positive ions toward the cathode;

• 3. Electron multiplication: as the free electrons move through the gas, they gain enough energy to ionize additional gas molecules through collisions. Each ionization event produces more free electrons, resulting in an exponential multiplication of the electron population. This phenomenon is known as an electron avalanche;

• 4. Positive ion drift: simultaneously, the positive ions created during ionization also experience drift toward the cathode, contributing to the overall signal response;

• 5. Signal readout: the amplified electron avalanche leads to a detectable electrical signal at the anode, which can be measured and analyzed to determine the properties of the incident charged particle, such as its energy (for specific types of detector), position, and time of interaction;

The avalanche formation process is crucial in gaseous detectors because it amplifies the weak initial signal caused by the ionizing particle, making it easier to detect and measure. The size and shape of the avalanche, as well as the time it takes to develop, depend on the gas composition, pressure, and the strength of the applied electric field. Different types of gaseous detectors, such as proportional counters, drift chambers, and time-projection chambers, utilize variations of the avalanche formation process to achieve specific detection capabilities and applications. Understanding and controlling the avalanche formation is essential for optimizing the performance of gaseous detectors in various scientific and technological fields.

Ionization energy is the minimum amount of energy required to make an electron eject from an atom (X+first ionization energy $\rightarrow X^{+}+e^{-}$). With the presence of an electric field, the electrons are accelerated, and if its energy is larger than the ionization energy of the atom or molecule, bound electrons can be ejected, leaving behind a positive ion. Normally the most probable process is the creation of a singly charged ion, but a further ionization of the gas is possible depending on the transferred energy and on the charge density. The ionization cross section increases rapidly above the threshold and for most gases has a maximum at an energy above 100eV.

The process of ionization in a gas under the influence of an electric field can be described by a parameter, known as Townsend coefficient, or first Townsend ionization coefficient (or simply ionization coefficient). It is denoted by the symbol $\alpha$ and is measured in units of inverse distance (usually expressed in $\textrm{cm}^{-1}$ or $\textrm{m}^{-1}).$¹ The Townsend coefficient quantifies the probability of ionization per unit distance traveled by an electron in the gas. When an electron gains enough energy from the electric field, it can collide with a gas atom or molecule, causing ionization, i.e., the creation of electron-ion pairs. The Townsend coefficient depends on various factors, including:

• 1. Gas composition: different gases have different ionization properties, and thus, their Townsend coefficients vary. For example, the Townsend coefficients of noble gases like argon and xenon make them suitable for proportional counters and other gaseous detectors;

• 2. Electric field strength: the Townsend coefficient is not constant but depends on the strength of the electric field applied to the gas. At low electric fields, the coefficient is small, indicating a lower probability of ionization. As the electric field increases, the coefficient rises until it reaches a point of saturation;

• 3. Gas pressure: the Townsend coefficient also depends on the gas pressure. In some gases, the coefficient increases with pressure, while in others, it decreases;

• 4. Electron energy: the energy of the incident electron affects the ionization probability. High-energy electrons are more likely to cause ionization than lower-energy ones.

1.2 Townsend coefficient and gain

The Townsend coefficient represents the number of ion pairs produced per unit length of drift and it relates to the ionization cross section $\sigma _{i}$ through the expression $\alpha =N\sigma _{i}$ where N is the number of molecules per unit volume. It is essential in understanding the behavior of gas-filled detectors, such as proportional counters, Geiger–Muller counters$\ddot{u}$, and other gaseous detectors used in particle physics experiments and radiation detection applications. It helps to describe the gas amplification process in these detectors, where a primary ionization event initiated by a high-energy particle results in an avalanche of secondary ionizations, leading to a measurable electrical signal. By studying the Townsend coefficient, researchers can optimize the performance of gaseous detectors and design gas mixtures that are suitable for specific applications.

Another crucial parameter for gaseous detectors is the so-called multiplication factor (or gas amplification or gain); it is related to the process of successive ionizations resulting in charge amplification in proportional counters. An electron released in a gas volume where a uniform electric field is applied will produced a new electron-ion pair after a mean free path $\lambda$, making two electrons drifting and creating two more pairs after another mean free path. This process continues till the energy of the electrons is enough to produce new pairs and creates a charge avalanche in the active volume of the detector. Given is the number of electrons in a given position n, after a path dx the number of electrons will be $dn=n\alpha dx$. Integrating over a path length x, $n=n_{0}e^{\alpha x}$. The gain is defined as the ratio between n and $n_{0}$:

$$\begin{aligned} G=\frac{N}{N_{0}}=e^{\int _{x_{0}}^{x}{\alpha (x)\textrm{d}x}} \end{aligned}$$

(1)

where where x and $x_{0}$ are the initial and final coordinates of the multiplication path and $\alpha$ usually deepens on the electric field (so that $\alpha =\alpha (x)$). If alpha does not depend on x the gain is $G=e^{\alpha x}$. Since there is a very large difference in mobility and diffusion of electrons and ions, the process of charge multiplication results in a characteristic drop-like avalanche, with all electrons at the front and a slow ions trail. There is a limit on the gain due to secondary ionization processes in the gas and on the walls, that cause charge spread over the gas volume. Due to this effect, the electric field strongly increases in front of and behind the multiplying charge and space charge-induced distortions occur. So, a large number of secondary avalanches give rise to transition of the proportional avalanche to a streamer and eventually to a spark breakdown, effects that usually should be avoided. Indeed if not damped by the detector geometry or by a reduction of the electric field, the streamer can propagate through the whole gas gap, leading to a spark breakdown. The so called Raether’s limit ($\alpha x<20$ or a total avalanche size of $\sim 10^{8}$) is an empirical condition that sets the limit to the transition described above. In order to build a gaseous detector many factors have to be taken into account, including controlling the avalanche formation and development, the value of the applied electric field and the geometry. The choice of these and many other factors determine the specific type of gaseous detector. In particular, in the next sessions the use of glass inside gaseous detectors will be described.

It was mentioned that a key aspect of gaseous detectors is the collection of the produced charge. It occurs via the mechanism of the charge induction on conductors. Electrons and ions created in the gas volume drift, respectively, toward anode and cathode under the effect of the applied electric field; the final outcome is the collection at the electrodes of all charges. During avalanche drift, induced signals appear on the electrodes, with polarity and time structure that depend on the detector geometry, field strength and charge mobility. Once defined the way the charge is collected on the electrodes, it is possible to vary some parameters in order to build a certain type of detector based on ionization. There is a dependence of the involved processes on the gas amplification, on the applied voltage, on the geometry and on the gas mixture. The total charge collected is a function of the applied voltage, with a specific dependence depending on the voltage range. It is possible to define different operation regimes, each of them defining a certain type of detection strategy (i.e., type of detector).

1.3 Gaseous detector regimes

Gaseous detectors can operate in different regimes depending on the strength of the electric field (applied voltage) and the gas properties. These regimes describe the behavior of the ionization and charge multiplication processes within the gas volume. The main gaseous detector regimes are:

• 1. Recombination Region ($G<1$): at zero voltage no charge is collected and the ion-electron pairs recombine; as the voltage is raised the recombination forces are overcome, and the current begins to increase as more and more of the pairs are collected before they can recombine;

• 2. Ionization Chamber Region ($G\sim 1$): at some point all created pairs are collected and further increases in voltage show no effect (saturation in the flat region); this operation mode is suited for measurements of particle fluxes (dosimeters) but not for the detection of single particles since the signal is too small (order of 0.01fC/cm for MIP). Detectors working in this region are called ionization chambers;

• 3. Proportional Region ($G\approx 10^{3}-10^{5}$): the electrons gain sufficient energy between two collisions to produce secondary ionization and a charge avalanche (or cascade) develops. In this region with increasing operation voltage, the amplified charge is over a wide voltage range proportional to the primary particle. The detectors here are called proportional chambers. A charge measurement in combination with a momentum measurement can be used for Particle IDentification (PID). In the final part of this range, we find the limited proportionality region ($G\approx 10^{5}-10^{8}$): with the increasing in the voltage, the proportionality of the output signal to the primary ionization is limited by space charge effects, and the electric field get distorted. This effect is due to the fact that the ions slowly drift away and this reduce the electric field. This region is reached earlier the higher the primary ionization;

• 4. Saturation and Geiger Region ($G\approx 10^{8}$): with further increasing gas amplification, the output signal finally becomes independent of the primary ionization; the positive and negative charges of an avalanche, that drift away from each other, start to neutralize in the inner region of the avalanche because of the higher recombination probability at high charge density. This recombination leads to the emission of photons which can produce further electrons by the photoelectric effect and thus initiate new avalanches. The output current becomes completely saturated, always giving the same amplitude regardless of the energy of the initial event. In order to stop the discharge, a quenching gas must be present in the medium to absorb photons. Detectors working in this region are called Geiger–Mu$\ddot{u}$ller counters;

• 5. Discharge or Streamer Region ($G\approx 10^{8}-10^{9}$): at very high voltages self-sustaining discharges occur, either spontaneously or triggered by ionizing particles (breakdown); in a discharge both the electrodes are connected through a conducting plasma and before the plasma reaches the electrodes a limited streamer builds up in the region between them. The transitions saturated avalanche$\rightarrow$streamer$\rightarrow$discharges are strongly dependent on the electrode geometry, gas properties (composition, pressure, temperature, humidity) and on the resistance of the electrical current circuit. This regime can be interesting for particle detection because very high signal can be obtained. However with the modern electronics, this is used in special cases only, because on a long-time scale the lifetime of the electrodes can be a problem.

It is crucial to remind that for detector operating in proportional regime, it is needed to control the growing of the avalanche to continue to work in the same regime; in the discharge regime, it is needed to control the termination of the discharge to protect the detector against destructive sparking. So in both cases, a mechanism to control the related processes is crucial. According to the previous arguments (charge diffusion and operation regimes), the main criteria for the gas/gas-mixture choice are: high ionization density, little charge loss, low voltage at the required amplification, stable operation (safety against spark discharge), proportionality between ionization and output signal, low diffusion (in particular for detectors with position resolution), suitable drift velocity, good rate capability, and hence low or minor dead time (low space charge accumulation), radiation tolerance, low costs, eco-friendly (since the last few years).

2 From the proportional chamber (PC) to the multi-wire proportional chamber (MWPC)

The basis for most detector types measuring the ionization with gas amplification is counting tubes with cylindrical geometry and a landmark development was the Geiger–Mu$\ddot{u}$ller counter (Geiger mode) about a century ago. The signal processing started to be more precise with the first important developments in electronics; this made the counting tubes more and more used in particle physics. The working principle of a proportional counting tube are:

• In a cylindrical tube filled with a gas, a thin wire is stretched along the cylinder axis;

• A high voltage is applied between the wire (anode) and the cylinder wall (cathode);

• The electric field ($\sim 1/r$) is weak far from the wire and the electrons/ions drift in the gas volume;

• The multiplication occurs only near the anode.

The time development of an avalanche near the wire occurs in the following steps:

• A ionizing particle produces an electron-ion pair and the two charges drift to their respective electrodes (the electron proceeds toward the wire anode and the ion toward the wall cathode);

• In the region of increasingly high electric field, the electron starts a secondary ionization, and the avalanche multiplication starts to take place;

• The charges generated in the avalanche are separated by the electric field;

• Because of the lateral diffusion of the electrons, which is much stronger than that of the ions, a drop-like avalanche develops, and it spreads out around the whole circumference of the anode wire;

• The electrons are quickly collected on the anode ($\sim ns$) while the ions drift over a longer period (up to ms) toward the cathode.

By taking into account that the electric field corresponds to that of a cylinder capacitor, it is easy to derive that in this kind of detector the integrated charge signal is dominated by the contribution from the ions generated in the avalanche. Because of the dominance of the ion signal, the time evolution of the current is dominated by the contribution of the ion movement. The ions start near the anode (where the field strength is large) so the initial velocity of the ions is very high. Considering the ion drift velocity and using the Shockley–Ramo theorem [2], it is easy to derive the time an ion needs to arrive at the cathode, and it is found to be $\approx ms$. In case of a measurement rates $>1$ kHz, this time is too long and produce signal pile-up (to avoid this effect, an external differentiating signal circuit can be coupled to the detector). Another limitation of a proportional counter is that it does not give a position measurement of the incident particle.

Multi-wire proportional chambers (MWPCs) are done with planar arrangement of proportional counters without separation walls. They were invented by G. Charpak et al. in 1968 (Nobel Prize in 1992), and they were a breakthrough application in particle physics. MWPCs are composed by a set of thin, parallel, and equally spaced anode wires, symmetrically placed between two cathode planes, the gas gaps being typically larger than wire spacing.

As compared to a cylindrical counter, the electric field geometry is quite different:

• Nearby the anode wires there is a radial electric field, but far away it is homogeneous (parallel-plate capacitor);

• Symmetric negative potentials are applied to the cathodes;

• Electrons created by an ionizing particle drift along the field lines, approaching the high field region (close to the anode wires) where avalanche multiplication can take place;

• Every anode wire is read out separately and spatial information can be determined;

• Achievable coordinate resolution is of the order of 1 mm.

High mechanical precision, both for geometry and wire tension are needed as even small geometric displacement of a single wire can affect the field quality. Signal generation (as in proportional counters) is due to electrons that drift to the closest anode wire and (mainly) slow ions. The performance of the MWPC can be significantly increased if induced signals on cathode are also recorded, providing a second spatial coordinate. A useful geometry of the detector consists in segmented cathode with strips readout, so that it is possible to use charge sharing between adjacent strips to determine the center of gravity of signals on cathode strips; this leads to a precision of 50–$300\,\upmu \textrm{m}$. A time resolution of 10ns is possible with the “OR” of all channels. An alternative geometry could consist in using pads instead of strips; this would improve the resolution but the number of readout channels would explode.

3 From the drift chamber to the to the time projection chamber (TPC)

The drift chambers are a further development of the MWPCs. In a drift chamber, the position where a particle crosses the detector is determined by the drift time of the electrons measured from the generation of ionization at the time $t_{\textrm{start}}$ to the arrival at the anode at the time $t_{\textrm{stop}}$. The basic formula is an integral over the drift velocity which in general may vary in time

$$\begin{aligned} x=\int _{t_{\textrm{start}}}^{t_{\textrm{stop}}} v_{D} \,\textrm{d}t \end{aligned}$$

So for a drift chamber, the position resolution is no longer directly dependent on the spacing of the signal electrodes but rather on the precision of the time measurement of the electron drift until the arrival at the anode. The drift distances can be much larger than the wire distance in the MWPC and in general the number of the readout channels and the material per area that a particle has to cross are reduced. The position resolution improves as compared to MWPC, reaching values of $\sim 20\mu m$. However, for large volume detectors (where there are cubic meters as in experiments at colliders), the typical values are in the range 100–$200\,\upmu \textrm{m}$.

A further development is the time projection chamber (TPC), that is an “electronic bubble chamber” where it is possible to perform a full 3-D measurement of charged particle tracks [3]. There are two types of TPC:

• Gas filled, for accelerator experiments, where a low material budget is required;

• Liquid (usually a noble-element) filled, for non-accelerator experiments (neutrino or dark matter detection, i.e., rare events with very low cross-sections), where high-density and large active mass are required.

Because of the long drift time, TPC is not a fast detector; therefore, it is not suitable for high rates, but it can measure thousands of tracks in single events. Typically, the gas-filled TPC is used as a central tracking detector in a solenoidal magnetic field (in order to measure the track momentum) whose axis is parallel to that of the electric field, so that the Lorentz angle is zero. It is made by a (mostly) cylindrical detector with a central high voltage electrode at negative voltage and two “end-cap” anodes. The electrons produced by ionization in the gas drift in the homogenous electric field toward these end-caps and are detected by, e.g., MWPC or other detectors. The detectors at the end-caps measure the arrival and position. In particular, the measurements of coordinate is performed as follows:

• z coordinated (long side) by drift time;

• x coordinate by charge sharing of cathode pads (MWPC);

• y coordinate by wire/pad number (MWPC).

TPC allows a good PID by energy loss (induced charge on pads). Typical resolutions are of the order of mm for the z and y coordinate and 150–$300\,\upmu \textrm{m}$ for the x coordinate. The main challenges when using a TPC are long drift path, that increases the probability of attachment and diffusion; large volume, that requires high mechanical precision and very high voltage, that can cause discharges.

A further step in developing gaseous detectors was the deployment of the parallel plate chambers (PPC), that do not make use of wires. The PPC is a single-gap gaseous detector operated in avalanche mode; it consists normally of two planar electrodes made of metal, or metallized ceramic or plastic, kept apart at a fixed distance of 0.5–2 mm using precise spacers, with 5–$10\,\upmu \textrm{m}$ accuracy. The inner surface of the electrodes are carefully polished so that the electric field produced has excellent uniformity. The electrode size ranges from $10\times 10\,\textrm{mm}^{2}$ to $100\times 100\,\textrm{mm}^{2}$ or even larger.

Since the electric field inside a PPC is high and uniform over the whole sensitive gaseous volume of the detector, any ionization produced by a through-going charged particle will immediately start a gas avalanche process and it will eventually generate the observed signals on the pick-up electrodes. There is no time drift associated with the movement of the electrons to a region of high electric field, and the time jitter of these devices is only caused by the fluctuations in the growth of the avalanche.

The total PPC thickness is usually a few millimeters, depending on the design and the materials. The high electric field of 2–6 kV/mm between the electrodes allows the immediate Townsend avalanche amplification of the primary ionization, at any point within the sensitive volume. The motion of the electrons in the avalanche induces a very fast signal (rise time less than 1ns) on the electrodes. This fast signal is followed by a slow signal that lasts for several microseconds, which corresponds to the motion of positive ions. Depending on the gas filling, a gain of $10^{3}$–$10^{4}$ can be reached with a very low spark probability.

The basic comparison between wired-based detectors and PPC can be summarized as follows: in wired-based detectors, there is a statistical distribution of the primary ionization clusters and dispersions due to diffusion in the time needed to drift and amplify the ionization charge at the anode; this leads to an intrinsic limitation to the achievable time resolution. In PPC instead, the electric field is high and uniform over the whole gas volume and multiplication occurs immediately after the release of primary electrons in the gas; so, charge signals is detected on electrodes without a delay and the time resolution is subsequently much better.

The advantages of the PPC technique include a very fast response and a high rate capability of $10^{7}\textrm{Hz}/\textrm{cm}^{2}$. The use of PPCs as efficient detectors for MIPs (with the added possibility of high resolution timing) was not under investigation for two main reasons: the PPC time resolution must be of the order of 1–2ns for a 1 mm gas gap, caused by the intrinsic jitter of the avalanche; the PPC signals are small (about 1–10 fC) and this made detection of MIPs rather inefficient. Typically, with a gas mixture containing $CO_{2}$, $i-C_{4}H_{10}$ and $CH_{4}$, the maximum efficiency was 50–$80\%$, without any observable plateau.

4 High resistivity electrodes and spark counters

Small signal requires detecting the induced signals at the very beginning of the avalanche development; good results can only be obtained with large gains, obtainable at very high fields. This implies critical operation in a standard PPC with conductive electrodes, as discharges and their propagation throughout the whole counter easily occur. A structure capable of achieving the required high gains while limiting the discharge propagation, named the Spark Counter (Pestov Counters), employs one or two high resistivity electrodes. In this way, the large current increase generated by an avalanche results in a localized field drop, thus dumping the discharge [4]. Operation near to or in the discharge mode is then possible and provides fast and large signals. Localization can be conveniently performed with external readout strips or pads in contact with the outer face of the resistive electrodes (by the way fluctuations related to the transition from an avalanche to a streamer have to be taken into account).

Gas mixture designed to absorb photons in a wide energy range avoids propagation and ensures the locality of the discharge. A narrow gap (100–$300\,\upmu \textrm{m}$) requires high fields for multiplication; this reduces the statistical fluctuations in the avalanche development and provides the best time resolution. In order to achieve this working condition, constraints to the quality and uniformity of the electrodes have to be taken into account. Energy loss probability in a narrow gap reduces the efficiency of MIP detection for this reason the Spark Counter has to be operated at high pressures.

A time resolution down to 30 ps has been demonstrated in the early 1980s [4], by using the following configuration:

• Custom-made cathode with a copper film sputtered on standard floating glass;

• Semi-conductive anode, optically polished on both sides, with a resistivity of $10^{9}$–$10^{10}\Omega \textrm{cm}$

• $100\mu m$ gap;

• A complex mixture of hydrocarbons at 12bar.

The difficult procurement of these types of glass and the high pressures of operation, requiring heavy gas containment vessels, have, however, discouraged the use of this technology. Nevertheless the main findings of the development constituted the basis for further developments of parallel plate counters.

5 From the resistive plate chambers (RPCs) to the multigap resistive plate chambers (MRPC)

The resistive plate chamber (RPC) is a parallel plate chamber with resistive electrodes in place of the metallic plates used in PPCs. Aiming at the realization of very large detection areas, the RPCs employ high-resistivity electrodes made of phenolic polymer laminates, (usually Bakelite), with a resistivity between $10^{9}$–$10^{10}\Omega \textrm{cm}$, that are cheap and easily available. The sensitive gas gaps of RPCs are a few mm thick and full detection efficiency and time resolutions of the order of ns for fast particles are assured [5]. Further developments allowed to consolidate the technology and build the very large systems, hundreds of square meters in area, needed for muon detection in high-energy physics experiments. The high resistivity electrode plates are assembled with insulating support frames, providing tightness and gas distribution; for large detector sizes, regularly spaced insulating pillars between the electrodes ensure the uniformity of the gap. One important aspect is the need to provide operating voltage over large detection areas and avoid potential drops caused by the signal currents. Plates are then coated on the outer side with a thin graphite layer having a moderate surface resistivity ($\sim 200$–$300k\Omega /\square$); it distributes the voltage while allowing detection of the signals induced by the avalanche multiplication on sets of external strips, insulated from the electrodes. A schematic cross section of a standard single-gap RPC is shown in Fig. 1.

Fig. 1

Schematic cross section of a standard single-gap RPC

5.1 Avalanche formation and operation regime for RPCs

For a few thick gap RPC, the process of the formation of the avalanche takes 10ns or less; when the avalanche size exceeds $10^{6}$–$10^{7}$ electrons, the exponential growth is damped due to the field reduction induced by the growing space charge. This leads to a saturated avalanche regime. The convenient large signals obtained in this working condition made this operating mode considered for the first prototypes, but it causes a low rate capability of the detector caused by the large local voltage drop, only slowly recovering through the high resistivity of the electrodes.

In [6], comparison between the detection efficiency in the avalanche and streamer modes as a function of particle flux is shown: it is clear that in the streamer mode the maximum flux for efficient detection is reduced by several orders of magnitude.

To achieve high and stable gains, RPCs are operated with heavily quenched, photon absorbing gases, as freon ($CF_{3}Br$) or mixtures of tetrafluoroethane ($C_{2}H_{2}F_{4}$) and isobutane (i–$C_{4}H_{10}$) [7, 8]. In order to reduce the voltage needed for operation, heavily quenched argon-based mixtures are used [9,10,11]. While the streamer mode offers a substantial advantage in terms of signal charge, for many applications an operation in the saturated avalanche mode is preferred to avoid efficiency losses in the high flux areas of the detector. It is mandatory to ensure that the onset of the streamer mode appears at a voltage sufficiently above the one required for normal operation. Addition of small percentage of an electro-negative gas to the mixture considerably increases the gap between the avalanche and streamer modes, as shown in [12], where several percentages of sulphur hexsafluoride ($SF_{6}$) added to a $C_{2}H_{2}F_{4}$ $i-C_{4}H_{10}$ mixture are described. The efficiency plateau slowly moves toward higher voltages by adding more sulphur hexsafluoride, but the advantage of moving the onset of the streamer mode at a higher voltage is clear.

Even if phenolic resin electrodes allowing for the realization of very large detectors are cheap, the usual modest surface quality of the materials can cause local discharges and consequent noise in the detector. This issue was solved by wetting the electrodes with linseed oil dissolved in alcohol, having the property to form a thin, high resistivity polymer coating when drying, covering the surface imperfections [13].

5.2 Glass as electrodes for RPCs

An alternative to the use of phenolic laminates as resistive electrodes, is using plates of floating glass as electrodes. They are available in a wide range of sizes and thickness with a resistivity of $10^{11}$–$10^{12}\Omega \textrm{cm}$. The operating voltage is also provided by a lower resistivity layer, typically of about $1M\Omega /\square$ deposited on the outer surface of the glass [14,15,16]. A very important aspect is that at the application of the external electric field, free charges within the glass (presumably ions) migrate and rearrange, thus creating an internal field opposing the applied field and making the plates equipotential.

A clear advantage of glass, compared to Bakelite, is its very good surface smoothness, therefore not requiring special conditioning. Moreover while in phenolic-based RPC a small percentage of water in the gas mixture is needed in order to avoid resistivity increase (usually observed after about 30 days of gas flowing), this is not the case for glass; this avoid the risk to increase HF production. The drawback, due to the high resistivity, is a limited rate capability (however compensated by a more stable operation with time [17, 18]). It is worth to note that sub–mm localization accuracy can be achieved with the use of thin electrodes and small pitch readout strips.

5.3 RPC in a nutshell

In order to introduce the MRPC detector, the main features and aspects of RPC are summarized in this section. In the simplest configuration, the RPC is composed by two parallel resistive plates (electrodes), typically Bakelite (volumic resistivity $10^{9}$–$10^{10}\Omega \textrm{cm}$) and works in streamer mode with the signals induced on the pickup cells placed externally. The uniform electric field applied at the edges of the sensitive volume is of the order of tens of kV/cm. The formation and pick-up of the signal is shown in Fig. 2: the fast signal is the one induced by the electrons (moving with a drift velocity $v_{d}^{-}\sim 5\cdot 10^{6}\,\textrm{cm}/\textrm{s}$); the ions induce a slow signal (drift velocity $v_{d}^{+}$ a few order of magnitude less). The avalanche develops and the signal is picked up: the fast signal passes the threshold after a certain number of steps and this is independent from the position of the avalanche starting point. A charged particle crossing the gas volume where an electric field is applied creates clusters of electrons and positive ions and after a $\Delta t$ the avalanche is large enough to modify the electric field between the electrodes. The electrons reach the anode and they locally modify the charge on the electrode. The positive ions are much slower and they take longer to reach the cathode. Locally the electric field is “reversed” and it takes some time to dissipate the charge on the electrodes (RPC $\tau \sim 10$ ms).

Fig. 2

Scheme of signal formation in a gas gap

The electron avalanche grows according to Townsend law and only avalanches that traverse full gas gap will produce detectable signals; so only clusters of ionization produced close to the cathode are important for the signal generation. In a 2 mm gap RPC only ionizing clusters within a few hundred of $\upmu \textrm{m}$ from the cathode can make avalanches big enough to generate signal. It is worth to note that if the electric field is increased so that all clusters can generate detectable signal, sparks would occur. Finally in measurements done with RPC, the time jitter due to how avalanche develops initially has to be taken into account.

5.4 Multigap Resistive Plate Chambers (MRPC)

It is interesting to investigate if the gas gain can be increased such that avalanche produces detectable signal immediately. The answer is yes, given that very high gas gain is ensured and a way of stopping growth of avalanches to avoid streamers/sparks is in place. The idea to achieve this configuration was to add boundaries that stop avalanche development and that must be invisible to the fast induced signal, induced on external pickup cells.

Experimentally good results were obtained by the RPC; however, the LAA project at CERN has studied RPCs, over a few years, with the goal of improving their operation characteristics so they could be used in hadron colliders. An important step forward implied the following points:

• Increase the rate capability limit (from few tens to few hundreds of $\textrm{Hz}/\textrm{cm}^{2}$);

• Improve the time resolution;

• Increase the gain but at the same time find a way to stop the growth of the avalanche;

• Reduce the current across the gas.

The study of this configuration is described in [19]. Before the invention of MRPCs, many studies were carried out. One of the earlier attempts to improve the 2 mm gap RPC was to increase the size of the gas gap. Wide gap RPC (gas gap of 6 to 8 mm) had many advantageous features when compared to the conventional 2 mm gap chamber:

• Gap tolerance less critical compared with smaller gaps (easier the construction);

• A “light” gas can be used (high fraction of Argon) $\Rightarrow$ much lower electronegativity compared to the freon-rich mixtures used in 2 mm gap chambers $\Rightarrow$ smaller ionic current (i.e., much smaller charge produced in the gas gap) $\Rightarrow$ better rate capability.

The only disadvantage of the wide gap RPC was a significant worsening of the time resolution.

At that step it was clear that there was a need to keep the advantages of the wide gap but, at the same time, dramatically improve the time resolution (ensured by smaller gas gaps). The clever idea to achieve this result was to have one sensitive volume composed of a series of small gas gaps, with one set of read-out strips reading out the analogue sum of all gas gaps.

The first results of this study were obtained by comparing two prototypes: a wide gap RPC and a 3-gap multigap design; in both cases the total gas gap is similar ($\sim 9\,\textrm{mm}$), but the multigap is split into 3 sub-gaps of 3 mm each.

The intermediate plates are electrically floating and achieve the correct voltage by electrostatic effect. They are kept at the correct voltage by the flow of electrons and positive ions, created by avalanches in the gas gaps. In fact, there is a strong feedback mechanism that forces the gas gain to be equal in all subgaps, the principle of the “floating electrode” being very stable. The final result was that a time resolution similar to a single gap is obtained with this $3\times 3\,\textrm{mm}$ multigap.

There is a further advantage in subdividing the gas gap into many small gaps. A through-going particle in a single gap chamber produces individual and separate clusters of primary ionization; each of them will start an avalanche and the final signal will be the sum of all avalanches, that are not independent, so that the resultant signal is as if there was only one large “merged” avalanche (instead of the sum of many small discrete avalanches). The conclusion is that inherent fluctuations within the avalanche mechanism dominate for a single gap. For the multigap configuration instead, the signal is truly the sum of independent avalanches (since the avalanches occur in independent subgaps).

The nature of the resultant signal of MRPC has, as a direct consequence, an influence on the shape of the charge spectrum. The charge spectrum of a single wide gap RPC has an exponential shape, while a MRPC has a $\Gamma$ function shaped charge spectrum, similar to a Landau distribution. An exponential-shaped charge spectrum makes life very difficult for detector builders, since one is forced to keep the threshold as low as possible, to avoid a large reduction in efficiency. The charge spectrum from the MRPC has instead a peak well separated from zero, and thus the setting of the threshold is far less critical with respect to RPC. This improvement of the shape of charge spectrum further improves the high rate capability of this detector.

The original LAA-multigap design (3 gaps of 3 mm) had a time resolution $\sigma _{t}\sim 2ns$. A further step of the LAA group was to build a MRPC with 4 gaps of $0.7\,\textrm{mm}$, operating with a heavy freon gas mixture at atmospheric pressure. The result was a time resolution of less than 1ns and this is exactly the time resolution needed to build a muon trigger and to allow to distinguish the direction of the particle passing through a typical LHC experiment.

In summary:

• MRPC allows the simultaneous readout of many (sub) gaps;

• Each gap is separated from its neighbor by a sheet of resistive material;

• These highly resistive sheets are transparent to the fast signals generated by the avalanches;

• They are electrically floating $\Rightarrow$ strong feedback mechanism that keeps equal gain in all subgaps;

• Allow to stay in the avalanche operation mode (even if the high voltage would be in the streamer region).

Good timing needs small gas gaps, so very small gaps are needed (gaps of $250\,\upmu \textrm{m}$ for the ALICE TOF system have been used). Gas gaps of small size need to be constructed with very tight mechanical tolerance to have uniform field (indeed a $25\,\upmu \textrm{m}$ bump in the electrode would cause a $10\%$ local increase in the electric field for a $250\,\upmu \textrm{m}$ gap, compared with $1\%$ for a 2 mm gap).

The mechanism of the electric field adjustment, described before, keep the average total charge in a MRPC at low values ($\sim 2pC$). This makes possible a long free-streamer plateau and good rate capability, with the further advantage that small variation of high voltage have no macro-effects on the MRPC.

6 MRPC technology in particle physics experiments

RPCs and MRPCs are widely used in particle physics, both at colliders and in cosmic ray arrays. Among several experimental setups exploiting the MRPC technology, two specific applications are described in this section, with the aim to show the excellent performance in terms of time-of-flight/PID measurements and tracking: A Large Ion Collider Experiment (ALICE) at LHC and Extreme Energy Events (EEE). MRPCs are indeed used in ALICE as Time Of Flight system² and in the cosmic muon array Extreme Energy Events (EEE) as “muon telescopes”.

6.1 Advanced time of flight detector: ALICE at LHC

A Large Ion Collider Experiment (ALICE) [20] at the LHC at CERN is a dedicated heavy-ion experiment, whole main physics case is the study of the Quark–gluon plasma (QGP), a state of mattern in which the elementary particles that make up the hadrons of baryonic matter are deconfined [21]. ALICE is designed to cope with very high multiplicities and is composed with many sub-detectors (tracking system, time of flight system, calorimeters). It uses all known PID techniques, by exploiting different detectors on this purpose. ALICE is equipped with the largest area and most performing Time Of Flight (TOF) system, that makes use of the MRPC technology. The requirements of this system are:

• Large coverage ($150\textrm{m}^{2}$);

• High efficiency ($>95\%$);

• Good global time resolution $\sim 100$ ps;

• High granularity $\sim 10^{5}$ channels (occupancy $\le 15\%$ in Pb–Pb with a few thousands of primary charged particles produced per unit of rapidity).

Given these requirements, an appropriate solution for the detector to be used is MRPC (in double-stack configuration), that can ensure an efficiency $\epsilon \sim 99\%$ and an intrinsic time resolution $\sigma _{int}<50$ ps. In order to improve time resolution and efficiency, the following scheme (shown in Fig. 3) has been designed:

• Smaller gas gaps and higher Townsend coefficient;

• Lower applied voltage;

• Increase in the number of gas gaps;

• Smaller charge foot-print on the readout pad (smaller distance between the electrodes).

Fig. 3

Double stack configuration of MRPC used in the ALICE TOF system

In particular the final geometry of MRPCs used for the TOF system in ALICE at LHC is:

• Two stacks of resistive plates;

• $10\times 250\,\upmu \textrm{m}$ gas gaps;

• $120\times 7.4\,\textrm{cm}^{2}$ active area;

• Highly-segmented readout pads (96 pickup pads with $3.5\times 2.5\,\textrm{cm}^{2}$ area).

The whole TOF system design [20] is:

• Cylindrical surface;

• 3.7m from the beam pipe;

• $2\pi$ full azimuthal acceptance;

• $|\eta |<0.9$ polar acceptance;

• Divided in 18 sectors, 5 modules each;

• 91 MRPCs per sector;

• 1593 MRPCs in total;

• 153,000 channels.

All MRPCs of ALICE TOF system were built and tested in Bologna INFN laboratories (University of Bologna, University of Salerno, University of Gangeung, ITEP Moscow), with a mass production started in 2004 and completed in 2006. The MRPCs were tested and validated with beam-tests and subsequently assembled in 18 “supermodules”, installed in ALICE in 2008 and ready to take data in 2009 for the LHC start.

During the mass production many tests were performed [22]; among them, in this context, it is worth to mention:

• A careful choice and (resistivity) test of the acrylic paint used for the external (0.55 mm thick) glasses (industrially produced);

• Cleaning of all glasses in an ultrasonic bath before assembly. Measurement of resistive coating of the thicker glass after painting them (example of the measured mean value and spread ($\sigma$/mean) shown in Fig. 4);

• Measurement of gap (created with 0.4 mm thick internal glass plates) widths by means of a CCD camera for each MRPC. Pictures were taken in 8 positions (4 per stack), and the gaps were measured in a semiautomated way (an example of results shown in Fig. 5.

Fig. 4

Measurement of average surface resistivity over 5 points (left) and $\sigma$/mean (right) for a sample of the mass production of the ALICE TOF-MRPC glasses with resistive coating

From the very first signal seen with cosmic rays Summer 2008 to the full detector exploitation for Physics at the LHC during the following years, the excellent performance of the detector (in terms of efficiency, resolution, stability) has been verified and used in a large amount of physics analysis.

Fig. 5

Left: typical picture of the gaps of ALICE TOF-MRPC taken with the CCD camera. Right: gap measurements for external (left) and internal (right) glasses

6.2 Muon telescopes: extreme energy events (EEE)

The EEE Project [23, 24] is an array of cosmic muon telescopes, exploiting the MRPC technology, with a double goal: research and education (addressed to High School students). The main features of the EEE array are:

• MRPC telescopes covering an overall area of $\sim 3\times 10^{5}\,\textrm{km}^{2}$ (Fig. 6);

• Clusters and standalone stations hosted in Italian Secondary Schools, INFN labs and CERN;

• Each station (62) made of 3 MRPC chambers;

• Long-living MRPC-based system ($>15$ years);

• $>120$ billion tracks currently collected;

• Unconventional working sites (mainly school buildings, non-professional electrical lines, non-controlled environmental parameters, heterogeneous maintenance conditions).

Fig. 6

Map of the EEE array, showing schools and laboratories hosting muon telescopes

The configuration chosen for the EEE MRPCs ensured an efficiency of $\sim 90\%$, a time resolution of $\sim 250$ ps, a spatial resolution of the order of 1 cm and a good stability during the years. The performance of the timing and tracking system of the EEE detectors allowed the study of several physics cases.

The main features of the EEE telescopes (Fig. 7) are:

• $6\times 300\,\upmu \textrm{m}$ ($250\,\upmu \textrm{m}$ for the new chambers built during the EEE upgrade [25]);

• 2 external glass sheets (anode and cathode) - $160\,\textrm{cm}\times 85$ cm, 1.9 mm thick (resistive paint 5–$20\,M\Omega /\square$;

• 5 intermediate (electrically floating) glass sheets - $158\,\textrm{cm}\times 82$ cm, 1.1 mm thick;

• 24 copper strips (anode and cathode) to pick up the signal - $158\,\textrm{cm}\times 25\,\textrm{cm}$, spaced by 7 mm;

• Honeycomb panels to ensure mechanical stability - $182\,\textrm{cm}\times 90\,\textrm{cm}$;

• Gas-tight aluminum box - $200\,\textrm{cm}\times 100\,\textrm{cm}$.

The front–end is composed by 6 FEA cards (2 for each chamber), that incorporate the ultrafast and low power NINO ASIC amplifier/discriminator specifically designed for MRPC operation [26]. The determination of the position of the charged particle crossing the telescope is done as follows:

• The y coordinate is determined by the strip on which the signal is induced;

• The x coordinate is determined by measuring the difference between the arrival time of the signal at the two ends of the strip.

Three MRPC chambers assembled in a telescope are shown in Fig. 7.

Fig. 7

EEE telescope, showing details of MRPC boxes and electronics

The topology of the EEE network allows to measure time-correlated events at distances never addressed before. Telescopes placed in the same city can detect individual EAS [27], whereas telescopes located hundreds of kilometers apart can detect the coincidence between two different correlated air showers, for which a few interesting events were found [28]. A few details (glass, fishing line as spacer to create gas gaps, vetronite panel, 24 copper strips to collect the signal) of MRPCs used in EEE are shown in Figs. 8 and 9.

Fig. 8

Glasses, $300\,\upmu \textrm{m}$ fishing line as spacer to create gas gaps and vetronite panel of the MRPCs used in EEE

Fig. 9

EEE MRPCs: 24 copper strips to collect the signal

7 Conclusions

In order to detect a particle, it must interact with the material of the detector and transfer energy in some recognizable fashion and create a signal. Detection of particles happens via their energy loss in the material they traverses, in several ways according to the type of the interaction they undergo. One of the mechanisms to detect particles is based on ionization; many types of detectors using this feature have been designed in the last 50 years. Ionization can be detected by measuring electric currents in an electric field or light from atomic excitation. In the first case, the detector technology includes gaseous detectors, cryogenic detectors working with liquid noble gases and semiconductor detectors; in the last case, inorganic and organic scintillators can be considered. The (not complete) history of the gaseous detectors is described in this paper with a focus on detectors using glass as a key component.

The RPCs were built in the 1980s to substitute resistive electrodes for the metallic electrodes of the PPC detector. The choice of resistive electrode materials is crucial:

• The resistance of the resistive plate has to be relatively high (the material is not supposed to affect the generation of induced signals due to the movement of charge);

• On the other hand, the counting rate of the detector is related to the time that the resistive electrode needs to recover and charge up again, which depends on the resistivity of the electrode; excessive resistivity will increase the detector’s dead time and decrease the detector’s counting rate capability.

RPC detectors generally adopt glass or Bakelite as the resistive material, whose volume resistivity is usually within the range of $10^{9}$–$10^{13}\Omega \textrm{cm}$. The use of glass in MRPC, by exploiting the subdivision in gas gaps, is crucial in order to better control the avalanche footprint, avoid a too large signal, minimize dead time, improve intrinsic time resolution and rate capability, that all by keeping en excellent efficiency. RPCs and MRCPs can operate, as trigger and/or tracking and timing devices, at around 95% efficiency with a time resolution of sub–ns and at a counting rate of hundreds of Hz/cm$^{2}$. Their design and use in different experiments (at colliders or to detect cosmic radiation) has been described.