Nuclear Instruments and Methods in Physics Research A

Nuclear Instruments and Methods in Physics Research A 699 (2013) 134 138 Contents lists available at SciVerse ScienceDirect Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima Instrumentation at the LHC The silicon tracking systems Manfred Krammer Institute of High Energy Physics, Austrian Academy of Sciences, Vienna, Austria article info Available online 24 May 2012 Keywords: LHC Tracking Silicon detectors abstract For the experiments at the Large Hadron Collider at CERN the detectors in the vertex region and the tracking systems are very critical components for the success of the projects. Due to the harsh environment at the high intensity hadron collider these detectors are also the most challenging to construct. The design of the experiments started about 25 years ago. At this time silicon detectors as particle detectors for high energy physics experiments were still in their infancy. But, because of the huge progress made in developing this technology all LHC experiments finally employed silicon detectors and some experiments even constructed very large silicon systems. The LHC experiments went into operation in 2009, and after two years of operation the performance of the detectors can now be evaluated. This paper gives an overview of the different silicon detector systems at LHC and summarizes the installed silicon detector types and dimensions. & 2012 Elsevier B.V. All rights reserved. 1. Historical remarks In the mid-1980s the project of a Large Hadron Collider at CERN utilizing the existing LEP tunnel with 27 km circumference was proposed. In the first presentations this project aimed for a centre of mass energy of 16 TeV and a luminosity of about 10 33 cm 2 s 1. At the same time, the construction of the SSC in the USA had started an even larger project to build a proton proton collider with a centre of mass energy of 40 TeV and a similar luminosity. The LHC project gained momentum and was soon seen as a competitor project to the SSC. A first optimistic time schedule of the LHC foresaw its completion well before the start-up of the SSC. It became however clear that in order to beat the SSC for example on the search for the Higgs Boson the lower collision energy of the LHC had to be compensated by a higher luminosity. This was the birth of the high luminosity LHC. A collider with a peak luminosity of 10 34 cm 2 s 1 seemed feasible, but it was unclear how one could exploit such a high intensity machine with the detector technologies of those days. The first experimental concepts focused on calorimetry and on the detection of muons. In these schemes the inner region of the experiments consisted of either absorbers (for a beam dump type experiment) or of transition radiation detectors and calorimeters. Outside of these inner detectors or absorbers, a muon spectrometer was envisaged to detect and measure muons. See Fig. 1 for an example of a detector concept discussed at that time. It was a widespread belief by many physicists at these days that tracking or vertexing at such particle intensities would not be E-mail address: manfred.krammer@oeaw.ac.at possible. Silicon detectors just started to be developed for high energy experiments. The first ever silicon microstrip strip detector, a surface barrier sensor, was tested at CERN a few years ago [2]. The first silicon detectors using the planar technology and implanted strips were installed in the NA11 experiment in 1983 [3]. Nevertheless, in the same conference (Como 1988) where beam dump experiments and calorimetry were discussed for the LHC, some foresighted colleagues proposed silicon tracking systems utilizing this new technology. In their paper [4] a multilayer silicon strip tracker was proposed with a projected silicon surface of 40 m 2 a design surprisingly close to some of the finally adopted detector layouts. Between these early discussions on possible detector concepts and the start of the LHC experiments lie 30 years of remarkable progress in silicon detector technology, but also of substantial advances in the read-out and connectivity technology. Size and complexity of the detectors increased and silicon systems became indispensable components for high energy physics experiments. The employment of strip and pixel sensors in so-called vertex detectors opened up the window to heavy flavour physics. Meanwhile, no experiment in particle physics is designed without silicon sensors. Also more and more detector physicists in the field turned to silicon applications, thus making this detector community the fastest growing. Fig. 2 compares the number of institutes participating in the construction of silicon detectors during various time periods. The early microstrip detectors for fixed target experiments were built by only one institution (CERN; NA1 experiment, eight physicists) and by three institutions (CERN, MPI Munich, TU Munich; NA11 experiment). The experiment Mark II at SLAC and the LEP experiments (ALEPH, DELPHI, L3, and OPAL) at CERN used silicon vertex detectors for 0168-9002/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nima.2012.05.057

M. Krammer / Nuclear Instruments and Methods in Physics Research A 699 (2013) 134 138 135 2. ATLAS experiment Fig. 1. Conceptual design for a non-magnetic detector at LHC or SSC [1]. The ATLAS experiment (A Toroidal LHC ApparatuS) is a multipurpose experiment placed inside a huge underground cavern [5]. The inner tracking system of ATLAS consists of a silicon hybridpixel detector in its centre, followed by the silicon strip detector, called SemiConductor Tracker (SCT), and the gas straw-tube Transition Radiation Tracker (TRT) surrounding the SCT. The ATLAS pixel detector [5 18] consists of three barrel layers (1456 modules) and of three end cap discs (288 modules) on each side. The innermost pixel layer is placed at a radius of 50.5 mm from the beam axis. The pixel dimensions are 50 400 mm 2. The total number of read-out channels of this device is 80 million this makes the ATLAS pixel detector the largest pixel detector installed at the LHC. Surrounding the pixel detector is the ATLAS SCT [7] consisting of four barrel layers (2112 modules) and nine end cap discs on both sides (1976 modules). The SCT has a radial coverage of 30 cmor o52 cm, and extends up to 9Z9 ¼ 2:5 in pseudorapidity. The silicon area of the SCT is 61 m 2 with a total of 6.3 million read-out channels. A charged track originating from the vertex region traverses first three layers of the pixel detector, then eight silicon strip layers before it enters the transition radiation tracker, where it generates hits in 36 straw tubes. The performance of these detector systems was excellent during the first years of operation with results close to the design values. For details about the performance of the ATLAS pixel detector and of the SCT see for example Refs. [8,9]. 3. ALICE experiment Fig. 2. Number of high energy physics institutes involved in the construction and operation of silicon detector systems: for the first detectors ever, for the LEP experiments, and for the LHC experiments. the first time in a collider environment. The first vertex detectors, followed by several upgrades, were already designed and built by about 55 institutes. At present, for the LHC experiments about 165 institutes are developing, constructing and operating the various silicon detector systems. At the time of the technical proposals for the LHC experiments, the acquired knowledge and the obvious advantages of high performance track reconstruction for the quality of the physics results led to the proposal of complex silicon tracking systems. The designs of the central trackers were optimized to reconstruct isolated high p t tracks and high p t tracks within jets with high efficiency over a large rapidity range. The momentum resolution for isolated charged leptons in the central rapidity region was aimed to be Dp t =p t 0:1p t (p t in TeV) in order to also allow the determination of the lepton charge up to about 2 TeV. For important discoveries the ability of b-tagging at highest luminosities was thought to be crucial. This requirement led to the proposal of layers of pixel detectors with a projected impact parameter resolution of the order of 20 mm for high p t tracks. The LHC finally started regular operations in 2009. During the first years, the LHC produced proton proton collisions with a centre of mass energy of 7 TeV at luminosities up to the design value of 10 34 cm 2 s 1. Two running periods with colliding lead ions took place at the end of 2010 and 2011. In order to exploit the physics potential of this machine the four large experiments (ALICE, ATLAS, CMS and LHCb) and a few small experiments (TOTEM, LHCf) went into operation to record and analyse the collisions. The main silicon systems of these experiments will be explained in the following chapters. The ALICE Experiment (A Large Ion Collider Experiment) [10] is dedicated to the exploitation of the heavy ion mode of operation of the LHC. Compared to proton proton collisions a much lower luminosity of about 10 27 cm 2 s 1 is anticipated during heavy ion collisions. However, due to the nature of these collisions much larger charged multiplicities are produced per event. The ALICE experiment was designed to cope with a maximum charged particle multiplicity of up to 8000 per unit of rapidity. Obviously, these different experimental conditions led to a different detector layout compared to the other LHC experiments. The main tracking detector of ALICE is a huge time projection chamber the biggest ever built. The TPC has a length of 5 m, an inner radius of 0.85 m, and an outer radius of 2.5 m resulting in a gas volume of 88 m 3. Positioned at radii below 0.5 m, there are three silicon subdetectors. The innermost detector is a two-layer Silicon hybrid Pixel Detector (SPD) with pixel dimensions of 50 425 mm 2, and 9.8 million pixels [11]. Surrounding the pixel detector are two layers of Silicon Drift Detectors (SDD) [12]. The ALICE SDD is the only silicon drift detector used in the LHC experiments. It covers a surface of 1:31 m 2 with (only) 133,000 read-out channels. It is a detector offering a compromise between precisely measured two-dimensional points and a reduced number of read-out channels. To reach the required precision an elaborate alignment and correction procedure is however required. As an example, the effect of the correction for the non-uniformity of the drift velocity on the alignment of the ALICE SDD is shown in Fig. 3. The third and outermost silicon detector consists of two layers of Silicon Strip Detectors (SSD). The SSD is built out of doublesided strip detectors, covering a surface of 5 m 2, and consists of 2.6 million read-out channels.

136 M. Krammer / Nuclear Instruments and Methods in Physics Research A 699 (2013) 134 138 Fig. 3. Sigma of the track-to-point residuals in the bending plane, before and after the correction for the non-uniform drift velocity for both ALICE SDD layers [13]. While the operation of the silicon detectors in ALICE was not without problems (see Ref. [13] for details) the physics design goals of ALICE could be reached. Due to the excellent particle identification, using de/dx from four layers of silicon (SDD, SSD), hadron identification and separation is achieved for particle momenta of as low as 100 MeV/c. Fig. 4. Sketch illustrating the rf geometry of the VELO sensors. For clarity, only a portion of the strips is illustrated. In the f-sensor, the strips on two adjacent modules are indicated to highlight the stereo angle [14]. made several measurements in the heavy flavour sector improving the precision of important quantities. 4. LHCb experiment The LHCb experiment [14] is a dedicated experiment to study CP violation and the decays of B hadrons. The experimental detector is a single arm forward spectrometer, hence with a very different configuration compared to the other LHC experiments. Excellent tracking capability and precise reconstruction of primary and secondary vertices are essential requirements for LHCb. Several tracking detector stations and a bending magnet are used to reconstruct the trajectories of charged particles. The tracking stations in front and behind the magnet are composed of silicon strip detectors. These two sub-detector systems, the Tracker Turicensis (TT) and the Inner Tracker (IT), have four and, respectively, three detection layers covering a total silicon area of about 12 m 2 [15]. The most important detector component of LHCb, however, is the silicon vertex detector, called VELO (VErtex LOcator). The VELO surrounds the vertex region and is made of two detector halves, each one composed of 21 modules. A module is equipped with two semi-circular micro-strip silicon sensors. One of the sensors has radial strips and the second has f-strips. Fig. 4 shows a sketch illustrating the two different sensor geometries. The silicon sensors of the VELO have to be positioned as close as 7 mm to the beam axis. However, this is too close for safe operation of the detector during beam injection and beam steering. Therefore the VELO system allows to move the two half shells in safe positions (30 mm away from the beam axis) until stable beams are declared. This is achieved with a precise motion system capable of positioning the VELO with an accuracy of about 10 mm. The full system is operated in a secondary vacuum separated from the LHC vacuum by 300 mm thick aluminium foils. Like the other LHC experiments, also the LHCb detectors fulfill their performance specifications [16]. Already now, LHCb has 5. CMS experiment The CMS experiment (Compact Muon Solenoid) [17] is performed with the smaller of the two multipurpose detectors at LHC. Concerning the use of silicon sensors in the experiment, the CMS collaboration had decided to construct a full silicon tracker, unlike the other experiments where silicon detectors are used in combination with other types of tracking devices. As a consequence CMS has designed and installed the largest silicon tracking detector built up to now. The CMS tracker is geometrically divided into several sub-structures (see Fig. 5): the pixel detector very close to the interaction point and the Silicon Strip Tracker (SST) consisting of the inner barrel detector (TIB), the inner discs (TID), the outer barrel (TOB), and of the two end cap detector systems (TEC). The overall length of the tracker is 5.4 m and its outer diameter 2.4 m. The pixel detector modules are built as hybrid pixel assemblies with a pixel cell size of 100 150 mm 2. The 1440 pixel modules (66 million pixels) in three barrel layers and in two times two vertical discs cover an area of 1 m 2 [18]. A special feature of the pixel system is the fact that it is installed on rails, and that within a modest time interval, it can be extracted from and re-inserted into the CMS detector if maintenance requires it. This operation was already successfully executed during the winter stop of 2010. The very large Silicon Strip Tracker (SST) contains a total area of 200 m 2 (!) equipped with silicon strip detectors. It surrounds the pixel detector and adds 10 tracking layers (4 TIB, 6 TOB) in the central region. In addition, three small and nine large detector discs (TID and TEC) are located on either side [19]. The 24,244 silicon sensors in 15 different sensor designs form the barrel and end cap layers [20]. The inner parts of the SST consist of modules with one sensor, while the outer parts consist of modules with two sensors (daisy-chained). The 15,148 detector modules have

M. Krammer / Nuclear Instruments and Methods in Physics Research A 699 (2013) 134 138 137 Fig. 5. Schematic cross-section through the CMS tracker. Each line represents a detector module. Double lines indicate back-to-back modules which deliver stereo hits. Fig. 6. The x y maps of the reconstructed material distribution in the CMS tracker extracted from data: (left) from photon conversions, and (right) from nuclear interactions. The x y bin size is 0:5 0:5 cm 2. been produced in an industrial-type production chain involving many institutes worldwide. A last figure worth to be mentioned is the number of wire bonds, which sums up to about 25 millions. For these very large detector systems with the complicated mechanical structures and the associated service systems the description and modelling of the detector materials is very important. The overall tracker performance is severely affected by the material of the tracker itself. The tracker material modifies the trajectories of charged particles through bremsstrahlung, photon conversion, nuclear interactions, multiple scattering, and energy loss. It is therefore of utmost importance to establish a precise description of the material distribution in order to correctly treat all these effects in the detector simulation. Fig. 6 shows the vertices of photon conversions and nuclear interactions in the CMS tracker, reconstructed from data [21]. In order to check the consistency between the tracker simulation and the real material distribution, a quantitative comparison of simulation and data was performed. In general, the observed relative agreement was of the order of 10%, except for a few larger discrepancies in very localized regions. Also the other LHC experiments have executed similar exercises with comparable results, and all experiments are continuing to improve their understanding of the material distribution. Despite some problems with the C 6 F 14 cooling system which forced CMS to shut off a few cooling loops, the tracker operated efficiently and stably. Out of the large number of SST read-out channels 97.7% were fully operational and stable over time. The CMS tracker provided data with high quality for a robust reconstruction of tracks and vertices, and herewith delivered an excellent input for the CMS physics analyses [22]. 6. TOTEM experiment Last but not least the silicon detectors of the TOTEM experiment are worth to be described. TOTEM (TOTal Elastic and diffractive crosssection Measurement) is an experiment studying very forward physics at the LHC, for example measuring the total cross-section, elastic scattering, and diffraction dissociation [23]. TOTEM is a small system compared to the other LHC experiments. It consists of several detector stations located at 10.5 m, 14 m, 147 m, and at 220 m on either side of the point 5 interaction region where the CMS experiment is located. The stations close to the collision region contain cathode strip chambers and triple GEMs, while in the stations at 147 m and at 220 m silicon sensors are used inside Roman Pots. Similar to the system described for LHCb, also these sensors are

138 M. Krammer / Nuclear Instruments and Methods in Physics Research A 699 (2013) 134 138 Table 1 Summary of area and number of read-out channels for the installed silicon systems at LHC. Experiment detector Silicon area (m 2 ) Number of channels (million) ATLAS Pixel 2.2 80 ATLAS SCT 61 6.27 ALICE SPD 0.21 9.8 ALICE SDD 1.31 0.133 ALICE SSD 5 2.6 LHCb VELO 0.0055 0.172 LHCb TT 8.2 0.143 LHCb IT 4.2 0.129 CMS Pixel 1 66 CMS SST 200 9,6 CMS Preshower 16 0.137 TOTEM 0.294 0.123 LHC total 300 175 Jet-Energy Resolution 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 placed in a parking position during beam injection and are brought close to the beam during stable beam conditions. As a peculiarity, the TOTEM silicon sensors are designed as edgeless sensors with the active strips only 50 mm away from the physical edge of the sensor. For the operation these sensors are moved close to the beam axis approaching the beam as close as 10s. First results from the operation of the TOTEM silicon system may be found in Ref. [24]. 7. Summary CMS Preliminary 10 2 p T [GeV/c] Corrected Calo-Jets Particle-Flow Jets 0 < η < 1.5 Fig. 7. Jet energy resolution in CMS as a function of p T for corrected calorimeterjets (open squares) and for particle-flow jets (upwards triangles) in the central region of 9Z9o1; 5 [25]. The LHC experiments have all installed large to very large silicon tracking systems. The silicon areas covered, the numbers of read-out channels, and the sophistication of the systems did probably exceed the expectations of even the most optimistic physicists discussing and proposing silicon detectors for the LHC experiments in the mid- 1980s. Table 1 summarizes the silicon detectors of the four large LHC experiments (ATLAS, ALICE, LHCb, CMS) and of the smaller experiment TOTEM. In total, these experiments have installed a silicon area of 300 m 2 with a total of 175 million read-out channels. In addition to what silicon tracking detectors were designed for, they today not only provide efficient tracking information but they also revolutionize the measurement of particle parameters usually belonging to the field of calorimetric methods. With the development of the particle flow method, not an option considered in the early design stages of the experiments, silicon tracking detector information is used to substantially improve e.g. the measurement of the jet energy, the tau identification, and the determination of the missing energy. One of the many examples where the particle flow method improves the measurement is given in Fig. 7. There the expected jet energy resolution in CMS using the standard calorimeter method is shown in comparison to the resolution achieved with the particle flow method. The silicon detector systems deployed in the LHC experiments operate very successfully and with remarkable stability. In spite of some problems, most notably with some of the cooling systems, the number of non-working elements is very small in the order of a few percent only. The physics performance of the detector systems has reached or even exceeded the respective design values already after a short period of operation. The research and development to upgrade the experiments in view of the second phase of LHC with an even higher luminosity has already started in all collaborations. For these upgraded versions of the experiments, silicon detectors will again be the preferred choice for many of the detector systems. References [1] R. Donaldson, M.G.D. Gilchriese (Eds.), Experiments, Detectors and Experimental Areas for the Supercollider, World Scientific, Singapore, 1988. [2] E.H.M. Heijne, et al., Nuclear Instruments and Methods in Physics Research 178 (1980) 331. [3] B. Hyams, et al., Nuclear Instruments and Methods in Physics Research 205 (1983) 99. [4] H.F.-W. Sadrozinski, A. Seiden, A.J. Weinstein, Nuclear Instruments and Methods in Physics Research Section A 279 (1989) 223. [5] ATLAS Collaboration, The ATLAS experiment at the CERN large hadron collider, Journal of Instrumentation 3 (2008) S08003. [6] G. Aad, et al., ATLAS pixel detector electronics and sensors, Journal of Instrumentation 3 (2008) P07007. [7] A. Abdesselam, et al., The integration and engineering of the ATLAS semiconductor tracker barrel, Journal of Instrumentation 3 (2008) P10006. [8] B. Di Girolamo, The ATLAS pixel detector, in: Proceedings of the International Workshop on Vertex Detectors VERTEX2011, Rust, Austria, 19 24 June 2011, PoS (Vertex 2011) 006. [9] P. Haefner, Operation and performance of the atlas silicon microstrip tracker, in: Proceedings of the International Workshop on Vertex Detectors VER- TEX2011, Rust, Austria, 19 24 June 2011, PoS (Vertex 2011) 007. [10] ALICE collaboration, The ALICE experiment at the CERN LHC, Journal of Instrumentation 3 (2008) S08002. [11] R. Turrisi, et al., ALICE Pixel Detector Operations and Performance, PoS (Vertex 2010) 007, 2010. [12] S. Beole, et al., Nuclear Instruments and Methods in Physics Research Section A 582 (2007) 733. [13] R. Santoro, Performance of the ALICE silicon detectors, in: Proceedings of the International Workshop on Vertex Detectors VERTEX2011, Rust, Austria, 19 24 June 2011, PoS (Vertex 2011) 015. [14] LHCb Collaboration, The LHCb Detector at the LHC, Journal of Instrumentation 3 (2008) S08005. [15] M. Tobin, et al., Performance of the LHCb Silicon Tracker in pp Collisions at the LHC, LHCb-PROC-2010-054. [16] S. Borghi, Tracking and alignment performance of the LHCb silicon detectors, in: Proceedings of the International Workshop on Vertex Detectors VER- TEX2011, Rust, Austria, 19 24 June 2011, PoS (Vertex 2011) 016. [17] CMS Collaboration, The CMS Experiment at the CERN LHC, Journal of Instrumentation 3 (2008) S08004. [18] A. Dominguez, Nuclear Instruments and Methods in Physics Research Section A 581 (2007) 343. [19] G.H. Dirkes, Nuclear Instruments and Methods in Physics Research Section A 581 (2007) 299. [20] M. Krammer, Nuclear Instruments and Methods in Physics Research Section A 531 (2004) 238. [21] CMS Collaboration, Studies of Tracker Material, CMS PAS TRK-10-003, 2010. [22] Performance of the CMS Silicon Tracker, G. Sguazzoni, in: Proceedings of Science, Proceedings of the 20th Workshop on Vertex Detectors, Rust, Austria, June 19 24, 2011, PoS (Vertex 2011) 013, CMS CR-2011/187. [23] TOTEM Collaboration, The TOTEM experiment at the CERN large hadron collider, Journal of Instrumentation 3 (2008) S08007. [24] G. Ruggiero, The TOTEM roman pot detector system at the LHC: status, operation and performance, in: Proceedings of the International Workshop on Vertex Detectors VERTEX2011, Rust, Austria, 19 24 June 2011, PoS (Vertex 2011) 017. [25] CMS Collaboration, Particle Flow Event Reconstruction in CMS and Performance for Jets, Taus, and E miss T, CMS PAS PFT-09/001.