Spark-proof GEM gives higher gain

30 April 2007

A team from CERN and INFN has developed the resistive-electrode thick GEM, or RETGEM, which offers a high gain while being intrinsically protected against sparks.


Un nouveau GEM pour gain plus élevé

Le détecteur GEM (multiplicateur d’électrons à gaz), mis au point au CERN par Fabio Sauli, est particulièrement utile dans le cadre des grandes expériences de physique des hautes énergies. Cependant, bien que sa mise au point ait considérablement progressé, le GEM reste un détecteur relativement fragile et peut facilement être endommagé par les étincelles, qui sont presque inévitables à gain élevé. Une petite équipe du CERN et de l’INFN a maintenant développé le RETGEM, un multiplicateur GEM dont les électrodes métalliques ont été remplacées par des électrodes en matériau résistif. RETGEM offre un gain maximal qui est dix fois plus élevé; il est aussi intrinsèquement protégé contre les étincelles et est donc très robuste.

The gas electron multiplier (GEM) detector developed at CERN by Fabio Sauli has several unique features (CERN Courier December 1998 p19). For example, it can operate at relatively high gains in pure noble gases, and can be combined with other devices of the same kind to operate in a cascade mode. Indeed, cascaded GEM structures now feature in several large-scale high-energy physics experiments, such as COMPASS, TOTEM and LHCb at CERN. The basic device consists of a metallized polymer foil chemically pierced to form a dense array of microscopic holes. Applying a voltage across the foil creates a high electric field in the holes which then act as tiny proportional counters, amplifying ionization charge. However, despite great progress in its development and optimization, the GEM is still a rather fragile detector. It requires very clean and dust-free conditions during its manufacture and assembly and it can be easily damaged by sparks, which are almost unavoidable when operating at high gain.

To try to overcome these problems, a few years ago a team of physicists from CERN and the Royal Institute of Technology in Stockholm developed a more robust version of the GEM, which was further improved by a team at the Weizman Institute of Science in Rehovot. Called the thick GEM (TGEM), it is based on printed circuit boards (PCBs) metallized on both sides, with an array of tiny holes drilled through (figure 1). Typically 0.5–1.0 mm thick, it is manufactured using the standard industrial PCB processing techniques for precise drilling and etching. The TGEM has excellent rate characteristics and can operate at higher gains than the GEM, but it can still be damaged by sparks.

Now a small team from CERN and INFN has developed a new, more spark-resistant version of the GEM in which the metallic electrode layers are replaced with electrodes of resistive material. We built the first prototypes from a standard PCB 0.4 mm thick. We glued sheets of resistive kapton (100XC10E5) 50 μm thick onto both surfaces of the PCB to form resistive electrode structures, and drilled holes 0.3 mm in diameter with a pitch of 0.6 mm using a CNC machine. The surface resistivity of the material created in this way varied from 500 to 800 kΩ/square, depending on the particular sample. After the drilling was finished, the copper foils were etched from the active area of the detector (30 mm × 30 mm), leaving only a copper frame for the connection of the high-voltage wires in the circular part of the detector (figure 2). We call this the resistive-electrode thick GEM (RETGEM).

The detector operates in the following way. When a high voltage is applied to the copper frames, the kapton electrodes act as equipotential layers, owing to their finite resistivity, and the same electric field forms inside and outside of the holes as occurs in the TGEM with the metallic electrodes. So at low counting rates the detector should operate as a conventional TGEM, while at high counting rates and in the case of discharges the detector’s behaviour should be more like that of a resistive-plate chamber. The RETGEM is only seven times thicker than the conventional GEM structures and could easily be bent to form a semi-cylindrical shape, as is preferred in some cases, such as in the future NA49 experiment at CERN.

We have made systematic studies and further developments of the RETGEM in collaboration with the High Momentum Particle Identification (HMPID) group of the ALICE Collaboration and the ICARUS research group from INFN Padova. These investigations show that the maximum achievable gain before sparks appear in the RETGEM is at least 10 times higher than in the case of the conventional GEM (figure 3). Moreover, when sparks do appear at higher gains, the current in these discharges is of an order of magnitude less than in the case of the TGEMs, so they do not damage either the detector or the front-end electronics.

We have since manufactured RETGEMs 1 and 2 mm thick with active areas of 30 mm × 30 mm and 70 mm × 70 mm in the TS/DEM/PMT workshop at CERN and successfully tested the devices. The maximum gain achieved was 2–3 times higher than with the device that was only about 0.4 mm thick, reaching a value of close to 105; as before, sparks did not damage the detector. The RETGEMs could operate at up to 10 kHz/cm2 without a noticeable drop in the signal amplitude, while at higher counting rates the signal amplitude began dropping, as happens with resistive-plate chambers. We also found that double RETGEMs can operate stably in a cascade mode; we observed no charging-up effect despite the high resistivity of the electrodes and achieved gains close to 106 with the double-step RETGEMs.

The most interesting discovery was that if we coat the cathode of the RETGEM with a caesium iodide (CsI) photosensitive layer, the detector acquires high sensitivity to ultraviolet light – an approach that has already been used with the conventional GEM with metallic electrodes. In contrast to these earlier attempts, however, in our case, the CsI was deposited directly onto the dielectric layer, that is, there was no metallic substrate present. Surprisingly enough, this detector worked very stably in the pulse-counting mode, easily achieving gains of 6 × 105 in double-step operation. The measured quantum efficiency was 34% at a wavelength of 120 nm, which is sufficient for some applications such as ring imaging Cherenkov detectors (RICH) or for the detection of the scintillation light from the noble liquids.

These studies have shown that RETGEMs can compete with the GEM in many applications that do not require very fine position resolution. Indeed the RETGEM offers a maximum achievable gain that is 10 times higher, is intrinsically protected against sparks and is thus very robust, can be assembled in ordinary laboratory conditions without using a clean room, and can operate in poorly quenched gases and gas mixtures. Other resistive coatings could also be used and the resistivity optimized for each application.

We believe that the new detector will have a great future and will find a wide range of applications in many areas. In high-energy physics it can be used, for example in RICH, muon detectors, calorimetry and noble-liquid time projection chambers.

• The RETGEM team comprises Rui de Oliveira (CERN TS/DEM/PMT workshop), Paolo Martinengo (ALICE HMPID group), Vladimir Peskov (ALICE HMPID group), Francesco Pietropaolo (INFN Padova) and Pio Picchi (INFN Frascati).

Further reading

R Olivera et al. 2007 to be published in Nucl. Inst. Meth. A doi:10.1016/j.nima.2007.03.010.

V Peskov et al. 2007 Vienna Conference on Instrumentation,

bright-rec iop pub iop-science physcis connect