ATLAS silicon microstrip detector system (SCT)

Transcription

1 Nuclear Instruments and Methods in Physics Research A 511 (2003) ATLAS silicon microstrip detector system (SCT) Y. Unno* Institute for Particle and Nuclear Studies, High Energy Accelerator Research Organization (KEK), 1-1 Oho, Tsukuba City, Ibaraki-ken , Japan On behalf of the SCTCollaboration 1 Abstract The SCT together with the pixel and the transition radiation tracker systems and with a central solenoid forms the central tracking system of the ATLAS detector at LHC. Series production of SCT Silicon microstrip sensors is near completion. The sensors have been shown to be robust against high voltage operation to the 500 V required after fluences of protons=cm 2 : SCTbarrel modules are in series production. A low-noise CCD camera has been used to debug the onset of leakage currents. r 2003 Elsevier B.V. All rights reserved. PACS: Ka; Wk; s Keywords: LHC; ATLAS; Tracking detector; Silicon microstrip; Radiation tolerance; Microdischarge *Tel.: ; fax: address: unno@post.kek.jp (Y. Unno). 1 Institute representatives: B. Stugu, Univ. Bergen; J. Dowell, Univ. Birmingham; J. Carter, Univ. Cambridge; S. Roe, CERN; P. Malecki, INP, Crakow; W. Dabrowski, Univ. Mining and Metal., Crakow; J. Ludwig, A-L-Univ. Freiburg; A. Clark, Univ. Geneve; K. Smith, Univ. Glasgow; T. Ohsugi, Hiroshima Univ.; Y. Unno, KEK; R. Takashima, Kyoto Univ. Edu.; P. Ratoff, Lancaster Univ.; M. Gilchriese, LBNL; J. Jackson, Univ. Liverpool; M. Mikuz, Univ. Ljubljana; A. Carter, QMW; M. Ibbotson, Univ. Manchester; G. Taylor, Univ. Melbourne; M. Merkine, Moscow State Univ.; H.-G. Moser, MPI Munich; F. Hartjes, NIKHEF; I. Nakano, Okayama Univ.; S. Stapnes, Univ. Oslo; R. Nickerson, Univ. Oxford; A. Vorobiev, IHEP Protovino; J. Bohm, Acad. Scie. Czech, Prague; I. Wilhelm, Charles Univ. Prague; S. Pospisil, Czech Tech. Univ. Prague; M. Tyndel, RAL; T. Jones, UCL; A. Seiden, Univ. California, Santa Cruz; C. Buttar, Sheffield Univ.; L. Peak, Univ. Sydney; S.C. Lee, Acad. Sinica, Taipei; K. Hara, Univ. Tsukuba; T. Ekelof, Univ. Uppsala; J. Fuster, Univ. Valencia CSIS; R. Jared, Univ. Wisconsin. 1. Introduction The Large Hadron Collider (LHC) [1] is under construction at CERN to investigate a fundamental question of the Standard Model, the generation of mass. A general-purpose detector at the LHC, ATLAS [2], is also under construction. The central tracking system of the ATLAS, the inner detector (ID), is made of three detector subsystems based, respectively, on silicon pixels (PIXEL), silicon microstrips (SCT), and straw drift chambers with transition radiation function (TRT) [3]. A superconducting solenoid provides a 2 Tmagnetic field for analysing the momenta of charged particles [4]. The three subsystems are all under construction. For the SCT, silicon microstrip sensors, modules, and support cylinders and optoharnesses, are in series production. The SCT is scheduled to be /03/$ - see front matter r 2003 Elsevier B.V. All rights reserved. doi: /s (03)

2 Y. Unno / Nuclear Instruments and Methods in Physics Research A 511 (2003) installed into the inner detector complex in early 2005 and the ATLAS detector to be completed at the end of 2006 with the first beam collision in February Central Tracking System 2.1. Central solenoid The superconducting solenoid has been fabricated under the supervision of KEK and the ATLAS-Japan group [5]. In order to reduce the amount of material and space in front of the Liquid Argon electromagnetic calorimeter, the superconducting coil and the calorimeter share a common vacuum vessel. The solenoid was completed and tested in Japan in January 2001, transported to CERN in October 2001, and is being integrated into the vessel Inner detector (ID) A quadrant view of the ID is shown in Fig. 1. A major change since the ID TDR has been to incorporate an insertion tube to facilitate installation of the PIXEL subsystem. This tube required an increase to the inner radius of the SCTendcap regions, the inner edge of the endcap silicon sensors rising to 270 mm from 259 mm in the TDR. To cope with the reduction of radial coverage, the location of the endcap disks was rearranged and the inner modules of disk 1 moved to disk 2 to optimise the average number of hits Semiconductor tracker (SCT) The SCT system consists of a barrel made of four cylinders and two endcaps each of nine disks. The cylinders together carry 2112 detector units, the barrel modules described in Section 3. The disks carry in total 1976 endcap modules. The major geometrical parameters of the barrel and endcap regions are summarised in Table 1. A total of 8448 barrel and 7104 endcap microstrip sensors are required, all being fabricated from 4-in. silicon wafers Radiation level The ID volume will be subject to a fluence of charged and neutral particles from the collision point and from back-scattered neutrons from the calorimeters. An estimated fluence at the innermost of the SCT, in neutron equivalent units, is B neutrons=cm 2 in 10 years of operation. Fig. 1. A quadrant view of the inner detector consisted of PIXEL, SCT, and TRT detector systems. The detector boxes indicate the envelopes of active elements. The figure is based on the engineering drawing of TB P.

3 60 Y. Unno / Nuclear Instruments and Methods in Physics Research A 511 (2003) Table 1 SCTparameters in the barrel and in the endcap regions Barrel cylinder r (mm) Tilt angle (deg) Modules B B B B total 2112 Endcap disk Inner r (mm) Outer r (mm) Modules 1, ,3,4,5, total 1976 Table 2 Fig. 2. ATLAS SCT barrel module. Barrel module parameters In proton equivalent units, the fluence is B protons=cm 2 : The SCT has been specified to be able to withstand these fluences. 3. SCT barrel modules The barrel cylinders of SCT all carry detector units of an identical design, the barrel modules. A 3D view of the module is shown in Fig. 2. The module components include four silicon microstrip sensors, a baseboard, and an electronic hybrid wrapped around near the centre of the module [6]. The major design parameters are listed in Table 2. The sensors were designed for the ATLAS SCT specification [7]. The baseboard is made of thermal pyrolytic graphite (TPG), providing a mechanical structure, a high thermal conductivity heat path, and an electrical connection to the backside of the sensors [8]. The hybrid is made of four layers of Cu/Polyimide flexible printed circuit, reinforced mechanically, thermally, and electrically with Carbon Carbon substrate [9], and carrying 12 readout ASICs [10]. Electrical connections are made by Al wire-bonds between the sensor pairs, between the sensors and the hybrids, and between the ASICs and the hybrids. There are about 5400 wire-bonds in a module. The endcap modules are described elsewhere [11]. Detection planes Sensors Strips Strip directions Operating temperature Total chip power Thermal runaway heat flux Mechanical precisions back-to-back Fixation point Radiation length 1:2% X 0 4. Silicon microstrip sensors 4.1. Series production Two with small stereo angle crossing 63:56 63:96 mm 2 =sensor single-sided p-in-n Si wafer pair of sensors top and bottom side 80 mm pitch 126 mm length (2 mm dead in middle) þ= 20 mrad 7 C 6:0 W nominal 8:1 W max. > 240 mw=mm 2 at 0 C o5 mm (in-plane lateral) o10 mm (in-plane longitudinal) o50 mm (out-of-plane) o30 mm (in-plane) The SCT requires 15,552 silicon microstrip sensors for the experiment, and is producing with spares B19; 000: Japan, UK, and Norway share the responsibility for the B10; 600 barrel sensors produced by Hamamatsu Photonics; the UK,

4 Y. Unno / Nuclear Instruments and Methods in Physics Research A 511 (2003) Switzerland, and Spain for the B5300 endcap sensors produced by Hamamatsu Photonics, and Germany for the B3400 endcap sensors produced by CiS. Pre-series production started in late 1999 and series production is near completion as of November Radiation tolerance After extensive studies of the radiation tolerance of the prototype sensors, the SCTadopted a design of p strips in n-type Silicon wafers (p-in-n sensors). A number of pre-series sensors and modules were irradiated to the fluence of protons=cm 2 : Beamtests of non-irradiated and irradiated modules to measure charge collection and signal-to-noise ratios showed the bias voltage for the required performance was 150 V for the non-irradiated and over 350 V and as high as 500 V for the irradiated sensors [12]. The leakage current behaviour of the sensors is of interest in operating the modules. A number of irradiated pre-series sensors have been analysed as summarised in Table 3. The sensor shapes are a square type for the barrel modules (B), and five wedge types for the endcap modules (W). The SCT studied both 111 and 100 wafer orientations, qualified both, but chose 111 due to easier availability. Fig. 3 shows the leakage currents per unit volume pre- (bottom figure, four individual sensors) and post- (top figure, averages of the numbers of sensors listed in Table 3) irradiation. The level of leakage currents before irradiation was excellent, less than 200 na=cm 3 up to 500 V at room temperature. Following irradiation the sensors were annealed for 7 days at þ25 C; equivalent to 10 years of operation in the experiment. The leakage currents at 18 C were 150 ma=cm 3 at 500 V; three order of magnitude larger than pre-irradiation, but with no breakdown observed up to 500 V: In series production testing we have occasionally seen microdischarge, an onset of high leakage current, above 350 V: The series production sensors are specified to have no microdischarge below 350 V: We selected a number of out-ofspecification samples that had microdischarge or larger leakage current as shown in the lower part of Fig. 4. The leakage current behaviours of the same samples after irradiation are shown in the upper part of Fig. 4. No onset of large leakage current was observed up to 500 V: Table 3 Irradiated sensors sampled from Hamamatsu products Type Listed Dimensions ðw L TÞ (mm) Volume ðcmþ 3 B/111S 15 63:56 63:96 0: B/100S 5 63:56 63:96 0: W12 4 ð55:488 þ 45:735Þ=2 61:060 0: W21 3 ð66:130 þ 55:734Þ=2 65:085 0: W22 2 ð74:847 þ 66:152Þ=2 54:435 0: W31 3 ð64:636 þ 56:475Þ=2 65:540 0: W32 6 ð71:814 þ 64:653Þ=2 57:515 0: Fig. 3. Bias voltage dependence of the leakage current per unit volume of the ATLAS SCT silicon microstrip sensors fabricated by Hamamatsu: pre-irradiation (bottom) and postirradiation after protons=cm 2 (top). The pre-irradiation is of four individual sensors and the post-irradiation the averages of samples (number after /S bracket) in different shapes and crystal orientations in the inset box.

5 62 Y. Unno / Nuclear Instruments and Methods in Physics Research A 511 (2003) Fig. 5. Long-term behavior of leakage currents of the modules with onsets of high leakage current. The discontinuities were caused by re-starts of the DAQ system. Fig. 4. Bias voltage dependence of the leakage currents per unit volume of the ATLAS SCT silicon microstrip sensors fabricated by Hamamatsu: pre-irradiation (bottom) and postirradiation after protons=cm 2 (top). Those that have onset voltages of microdischarge below 350 V are out-ofspecification sensors specifically irradiated for testing purposes. The onset of microdischarge is caused by high electric field that exceeds the avalanche breakdown voltage in Silicon. Because of the type inversion and increase of the avalanche breakdown voltage of the irradiated Silicon bulk we expected and confirmed that the onset of the microdischarge was well above 500 V for irradiated SCTsensors. 5. Diagnosis of modules with microdischarge The barrel modules are in series production [13]. About 10% fraction of the modules exhibited onset of high leakage currents above 350 V and were subjected to a long-term test at 500 V: The decay of leakage currents is shown in Fig. 5. After 100 min most of the currents had decreased to typical levels without microdischarge except one sample, module 100. The fast decay of current is a Fig. 6. The hot spot identified in module 099 was at one unboded end of one of the sensors (top). An enlarged view revealed the spot was at the edge of strip and pad, with no visible external scratch at the region(bottom). typical characteristic of microdischarge localised to one spot. Using a low-noise CCD camera [14], one can take an image of hot spot caused by discharge. Fig. 6 shows one such image of module 099. The spot identified was typical of a microdischarge at one location and where no visible damage could be observed. Although bias voltages more than 200 V are not required initially in the experiment, such decay of the leakage currents is further insurance against later high voltage operation. In module 100, a hot sport was found at one of wire-bonds between sensors, as shown in Fig. 7.

6 Y. Unno / Nuclear Instruments and Methods in Physics Research A 511 (2003) microdischarge at a localised spot. A low-noise CCD camera confirmed such a hot spot, but also revealed that a hot spot that had a much longer current decay time was caused by foreign debris near a wire-bond. The barrel support cylinders and optoharnesses are also in series production, with the completed SCTscheduled for installation into the inner detector in early Fig. 7. The hot spot identified in module 100 was where wirebonds were (top). An enlarged view (bottom) revealed a coiled shaving of Al wire (GND potential) that may have discharged to the open metal area surrounding the edge of the sensor (HV). With a larger magnification, we found a coiled shaving of Al wire that may have caused a discharge to the open metal around the edge of the sensor. The much longer decay time was an indication of a cause other than microdischarge in the Silicon. 6. Summary ATLAS is scheduled to be completed at the end of 2006 with the first beam collision in February The superconducting coil of the central solenoid is completed and being installed into the vacuum vessel of the Liquid Argon calorimeter at CERN. Series production of B19000 SCTSilicon microstrip sensors is near completion. The sensors have been shown robust against high voltage operation up to the 500 V required after fluence of protons=cm 2 : Sensors that have onset of microdischarge as low as 250 V before irradiation show no onset up to 500 V after irradiation. The SCT barrel modules are in series production. About 10% of the modules constructed so far have shown onset of high leakage current above 350 V: A long-term test at 500 V showed rather fast decay of the leakage currents, typical of References [1] The LHC Conceptual Design Report The Yellow Book, CERN/AC/95-05(LHC), [2] ATLAS Technical Proposal for a General-Purpose pp Experiment at the Large Hadron Collider at CERN, CERN/LHCC/94-43, [3] ATLAS Inner Detector Technical Design Report, Vol. 1, 2, ATLAS TDR 4 and 5, CERN/LHCC and 97-17, [4] ATLAS Central Solenoid Technical Design Report, ATLAS TDR 9, CERN/LHCC [5] S. Mizumaki, Y. Makida, T. Kobayashi, H. Yamaoka, Y. Kondo, M. Kawai, Y. Doi, T. Haruyama, S. Mine, H. Takano, A. Yamamoto, T. Kondo, H. ten Kate, IEEE Trans. Appl. Superconductivity 12 (2002) 416. [6] T. Kondo, et al., Nucl. Instr. and Meth. A 485 (2002) 27. [7] Y. Unno, Nucl. Instr. and Meth. A 453 (2000) 109; D. Robinson, et al., Nucl. Instr. and Meth. A 485 (2002) 84. [8] A. Carter, R. de Oliveira, A. Gandi, Novel thermal management structures and their applications in new hybrid technologies and feed-through structures, CERN Report (ISBN ). Also patent World International Property Organisation: WO 00/03567A1, January [9] T. Kondo, et al., IEEE Trans. Nucl. Sci. NS-49 (2002) [10] W. Dabrowski, et al., Proceedings of the Seventh Workshop for Electronics for LHC Exp., pp , CERN , and references therein. [11] L. Field, Nucl. Instr. and Meth. A (2003) in these proceedings. [12] Y. Unno, et al., IEEE Trans. Nucl. Sci. NS-49 (2002) [13] Y. Ikegami, et al., Proceedings of the Eighth Workshop on Electronics for LHC exp., pp , CERN [14] Digital CCD camera series, C , Hamamatsu Photonics, Co. Ltd., Japan.