Performance of the Cylindrical-GEM prototype for the Inner Tracker of KLOE-2 G. Bencivenni, S. Cerioni, D. Domenici, G. Felici, M. Gatta, M. Pistilli Laboratori Nazionali di Frascati - INFN Via Enrico Fermi 40, I-00044 Frascati, Italy Abstract—We developed a low mass, fully cylindrical and dead-zone-free GEM detector as inner tracker for the KLOE experiment upgrade at the DAFNE Φ-factory. The proposed detector, that opens the way for a new and competitive category of ultra-light, full sensitive vertex detectors for high energy physics experiments, will play a crucial role in the study of the KS and η rare decays and in the measurement of the neutral kaon interferometry. The main physics requirements are: good spatial resolutions, σ(rφ) = 200 μm and σ(z) = 500 μm and a very low material budget, 2% of X0 for the whole detector. The inner tracker will be composed by five layers of cylindrical triple- GEM detectors (CGEM), covering the space from the beam pipe to the inner wall of the KLOE Drift Chamber (from 150 mm to 250 mm radius). Each CGEM is realized inserting one into the other the required five cylindrical structures made of thin (50 μ m) polyimide foils: the cathode, the three GEMs and the anode readout. In order to avoid the use of support frames inside the sensitive volume, the cylindrical GEMs are mechanically stretched from their ends where annular fiberglass frames are glued. The final result is a very light detector: only 0.2% of X0 per layer inside the active area. A full scale prototype (300 mm diameter, 360 mm length) of the first layer of the inner tracker has been successfully built and characterized under different experimental conditions. After a brief description of the construction procedure, the results of the extensive tests are presented. I. KLOE-2 AT DAFNE I N 2010 the KLOE experiment will start a new data taking with the upgraded DAFNE machine delivering a factor of 5 more luminosity. The KLOE apparatus will be upgraded with a new Inner Tracker (IT) placed inside the present Drift Chamber, aimed to improve the reconstruction of low momentum tracks near the Interaction Region, mainly K S and η decay products. The IT will be composed by five independent tracking layers of Cylindrical-GEM detectors (CGEM). The diameters will range from 150 mm (closer to the beam pipe) to 250 mm (closer to the inner wall of the Drift Chamber), with an active length of 700 mm. The idea is to realize the very large GEM foils required (up to 1600x700 mm 2 ) as a join of three smaller foils (about 550x700 mm 2 ). The requirements of the experiment are a spatial resolution of σ(rφ) = 200 μm and σ(z) = 500 μm and an overall material budget below 2% of X 0 to minimize the multiple scattering before the Drift Chamber. The readout will be performed with a XV pattern of readout strips engraved on the anode foil [1]. II. THE CGEM PROTOTYPE In 2007 a CGEM prototype has been built with the same diameter of the innermost IT layer (300 mm) and a reduced length of 360 mm. For sake of simplicity the anode has been segmented only with longitudinal strips, with a pitch of 650 μm, providing the rφ coordinate. About 320 strips over the 1500 have been connected to FEE and readout. Each CGEM is a Triple-GEM detector with a geometrical configuration of the gaps of 3/2/2/2 mm, respectively for drift/transfer1/transfer2/induction. The cathode is inner to the anode. A. Prototype construction At first three GEM foils are glued together to obtain the single large foil needed to make a cylindrical electrode. We used an epoxy (Araldite), distributed along an edge of the GEM, on a 3 mm wide region. Then the foil is rolled on an aluminum mould coated with a machined Teflon film providing a non-stick, low friction surface. The mould is then enveloped in a vacuum bag and vacuum is applied with a Venturi system, resulting in a uniform pressure of 0.8 Kg/cm 2 throughout the whole surface. At this step two fiberglass annular rings are also glued at the edges of the electrode, representing all the mechanical frames needed to support the detector. After the curing cycle of the glue the cylindrical electrode is extracted from the mould. Cathode and anode are obtained with the same procedure as well. At the end the five electrodes are inserted one into the other and the detector is sealed with epoxy on both sides [2]. B. X-ray test The prototype has been flushed with a Ar/iC 4 H 10 /CF 4 = 65/7/28 and tested in current mode with a 6 keV X-ray gun. A 10x10 cm 2 planar GEM has been placed in the same gas line and used as a reference in order to normalize the gain for changes of atmospheric variables. The gas gain has been measured up to a value of 2x10 4 and no discharge has been observed. The electron transparencies have also measured as a function of the various fields, resulting in a good agreement with the measurements found in literature. The fluctuations of the gain throughout the 940 mm of circumference were within 9%, showing a good uniformity on such a large surface.