Commissioning of the MEG II tracker system
M. Chiappini, A. M. Baldini, G. Cavoto, F. Cei, G. Chiarello, A. Corvaglia, M. Francesconi, L. Galli, F. Grancagnolo, M. Grassi, M. Hildebrandt, M. Meucci, A. Miccoli, D. Nicolò, M. Panareo, A. Papa, F. Raffaelli, F. Renga, P. Schwendimann, G. Signorelli, G. F. Tassielli, C. Voena
AAccepted for publication in JINST
Commissioning of the MEG II tracker system
M. Chiappini, a , A. M. Baldini, a G. Cavoto, c , d F. Cei, a , b G. Chiarello, c , d , A. Corvaglia, e M. Francesconi, a , b L. Galli, a F. Grancagnolo, e M. Grassi, a M. Hildebrandt, h M. Meucci, c , d A. Miccoli, e D. Nicolò, a , b M. Panareo, e , f A. Papa, a , b , h F. Raffaelli, a F. Renga, c , g P. Schwendimann, h , i G. Signorelli, a G. F. Tassielli, e , f C. Voena c a INFN Sezione di Pisa, Largo B. Pontecorvo 3, 56127, Pisa, Italy b Dipartimento di Fisica dell’Università di Pisa, Largo B. Pontecorvo 3, 56127, Pisa, Italy c INFN Sezione di Roma, Piazzale A. Moro 2, 00185, Roma, Italy d Dipartimento di Fisica dell’Università “Sapienza” di Roma, Piazzale A. Moro 2, 00185, Roma, Italy e INFN Sezione di Lecce, Via per Arnesano, 73100, Lecce, Italy f Dipartimento di Matematica e Fisica dell’Università del Salento, Via per Arnesano, 73100, Lecce, Italy g Laboratori Nazionali di Frascati, Via Enrico Fermi 40, 00044, Frascati, Italy h Paul Scherrer Institut (PSI), Forschungsstrasse 111, 5232, Villigen, Switzerland i ETH Zürich, Rämistrasse 101, 8092, Zürich, Switzerland
E-mail: [email protected]
Abstract: The MEG experiment at the Paul Scherrer Institut (PSI) represents the state of the artin the search for the charged Lepton Flavour Violating (cLFV) µ + → e + γ decay. With the phase 1,MEG set the new world best upper limit on the BR ( µ + → e + γ ) < . × − (90% C.L.). With thephase 2, MEG II, the experiment aims at reaching a sensitivity enhancement of about one order ofmagnitude compared to the previous MEG result. The new Cylindrical Drift CHamber (CDCH) is akey detector for MEG II. CDCH is a low-mass single volume detector with high granularity: 9 layersof 192 drift cells, few mm wide, defined by ∼ . × − X , thus minimizing the Multiple Coulomb Scattering (MCS) contribution and allowingfor a single-hit resolution < µ m and an angular and momentum resolutions of 6 mrad and 90keV/c respectively. This article presents the CDCH commissioning activities at PSI after the wiringphase at INFN Lecce and the assembly phase at INFN Pisa. The endcaps preparation, HV tests andconditioning of the chamber are described, aiming at reaching the final stable working point. Theintegration into the MEG II experimental apparatus is described, in view of the first data taking withcosmic rays and µ + beam during the 2018 and 2019 engineering runs. The first gas gain resultsare also shown. A full engineering run with all the upgraded detectors and the complete DAQelectronics is expected to start in 2020, followed by three years of physics data taking.Keywords: Tracking Detectors, Gas Detectors, Drift Chambers, MEG II Experiment Corresponding author. a r X i v : . [ phy s i c s . i n s - d e t ] M a y ontents The MEG experiment with its first phase of operation at the Paul Scherrer Institut (PSI) set themost stringent constraint on the charged Lepton Flavour Violating (cLFV) µ + → e + γ decay. Theanalysis of the 2009–2013 full data set resulted in the new best upper limit on the BR ( µ + → e + γ ) < . × − at 90% C.L. (ref. [1]), imposing one of the tightest bounds on models predicting cLFVenhancements through new physics beyond the Standard Model (refs. [2, 3]). The MEG experimenthas reached its ultimate level of sensitivity, limited by the resolutions on the measurement of thekinematic variables of the two decay products (ref. [4]). Therefore an upgrade of MEG, i.e. MEGII, was designed (ref. [5]) and is presently in the commissioning phase at PSI. MEG II aims atreaching a sensitivity level of the order of 6 × − in three years of data taking, by improvingboth, the detector resolutions and the muon stopping rate by a factor of two. The new MEG II positron tracker is a single volume drift chamber with a cylindrical symmetryalong the µ + beam axis. The length is ∼
191 cm and the radial width ranges from ∼
17 to ∼
29 cm(figure 1). The full azimuthal coverage around the µ + stopping target is guaranteed. This improvesthe geometric acceptance for signal e + and allows to use new tracking procedures capable to exploita factor of four more hits than MEG for a larger tracking efficiency ( ∼ ◦ in the innermost layer to 8.5 ◦ in the outermost one. The stereo anglehas an alternating sign, depending on layer, allowing to reconstruct the longitudinal hit coordinate.The stereo configuration gives CDCH the shape of a rotation hyperboloid. The single drift cell isquasi-square with a 20 µ m Au-plated W sense wire surrounded by 40/50 µ m Ag-plated Al fieldwires, with 5:1 field-to-sense wires ratio and a total number of 11904 wires.– 1 – igure 1 . The fully wired MEG II CDCH. The sensitive volume is filled with a low-mass He:iC H (90:10) gas mixture (ref. [6]), whichis a good compromise between high transparency (radiation length ∼ . × − X ) and single-hit resolution ( < µ m, measured on prototypes, ref. [7]). Full MC studies show angular andmomentum resolutions in agreement with the MEG II experimental requirements: 6 mrad and 90keV/c respectively. The CDCH design and construction were very challenging (refs. [8–10]). Indeed, given the high wiredensity (12 wires/cm ), the classical technique with wires anchored to endplates with feedthroughsis hard to implement. CDCH is the first drift chamber ever designed and built in a modular way.Wires were not strung directly on the final chamber, but they were soldered at both ends on thepads of two PCBs, which were then mounted on the chamber (figure 2, left). The wiring procedurewas performed at INFN Lecce, exploiting an automatic robot which fixed the wires on PCBs witha contact-less laser soldering. CDCH was then assembled at INFN Pisa by radially overlapping thewire-PCBs in the twelve 30 ◦ sectors of the helm-shaped endplates, between the spokes which actas housing for the PCBs. Each wire-PCB was placed at the proper radius through PEEK spacerswhose thickness was adjusted to have the correct radial dimension of the drift cells (figure 2, right). Figure 2 . Left: one of the PCBs where wires are soldered. The soldering pads are highlighted. Right:wire-PCBs stack with PEEK spacers between the spokes of the CDCH endplate. – 2 –t the outermost radius, a 2 mm-thick Carbon Fiber (CF) support structure encloses thesensitive volume and keeps the endplates at the correct distance, ensuring the proper mechanicaltension of the wires (figure 3, left). Twelve turnbuckles per side were connected to each individualspoke, allowing a fine tuning of the endplates distance, planarity and parallelism at a level better than100 µ m. The CDCH geometry was continuously monitored during the assembly with a coordinatemeasuring machine. At the innermost radius, a 20 µ m one-side-Al Mylar foil separates the CDCHgas volume from the He-filled target region (figure 3, right). The gas mixture tightness is achievedby using the ThreeBond 1530 glue and Stycast 2850 resin. Figure 3 . Left: the CF support structure with some turnbuckles highlighted. Right: insertion phase of the20 µ m Al Mylar foil. Aluminum wire breaking problems arose during the CDCH assembly and commissioning,despite the fact that all the operations were performed inside cleanrooms with a strict monitoring ofthe environmental conditions. The problem was deeply investigated performing optical inspectionswith microscopes, chromatography, practical tests and SEM/EDS analyses. We developed a safeprocedure to extract the broken wire pieces from the chamber. By simulating the drift cells electricfield with Garfield and ANSYS , the effect of a missing cathode wire on the e + reconstructionwas found to be totally negligible. Chemical and mechanical analyses showed that the origin of thebreaking phenomenon is the chemical corrosion of the Al core in presence of water condensationon wires from ambient humidity. Keeping the wires volume in an absolutely dry atmosphere with acontinuous flow of inert gas (Nitrogen or Helium) proved to be effective to stop the development ofcorrosion. Once the assembly was completed, CDCH was transported to PSI for the commissioningphase.More details about the CDCH design, construction and Al wires corrosion can be found inref. [11]. https://garfieldpp.web.cern.ch/garfieldpp/. – 3 – CDCH commissioning at PSI
After the CDCH sealing, the next step was the preparation of the endcap services. Two Al innerextension cylinders were connected to the endplates to couple the inner CDCH volume to the MEGII beam line. Twelve Al holders per side with grooves at the correct radii were machined to keep inposition the 216 Front-End (FE) boards per side (figure 4, left).
Figure 4 . Left: 216 FE boards connected to wire-PCBs tails (figure 2, right) and kept in position by theAl holders. One Al inner extension cylinder is visible. Right: one endcap region enclosed by the CF outerextension cylinder.
Figure 5 . Left: the 4 mm Cu pipes of the liquid cooling system directly embedded in the FE holders. Right:one 3D-printed piece used to flush dry air inside the endcaps. – 4 –he FE boards can supply the HV to the wires from one side and read out the signal at bothwire ends. In order to carry away the heat generated by the active electronics ( ∼
300 W/endplate)a chiller system is used. The 4 mm Cu pipes of the liquid cooling system are directly embeddedin the FE holders (figure 5, left). A system of twelve inlet tubes per side supplies dry air to3D-printed pieces screwed on the back side of each holder (figure 5, right). The dry air flushingsystem is used to uniform the temperature inside the endcaps and avoid water condensation on theFE electronics. Temperature and humidity sensors were added to monitor the endcaps environmentwhich is enclosed by CF outer extension cylinders (figure 4, right).
The HV Working Point (WP) for CDCH was estimated through gas gain simulations with theHe:Isobutane 90:10 mixture and typical atmospheric pressure values at PSI. Since we aim to besensitive to the single ionization cluster, the WP was defined as the HV to get a gas gain G = × .A HV tuning by 10 V/layer was considered to compensate for the variation of the cell dimensionswith radius. Furthermore, given the stereo wires geometry, the cell dimensions also vary along thechamber axis. The average HV WP as a function of the cell layers is summarized in table 1. Table 1 . HV working point as a function of the drift cell layers (L1 outermost, L9 innermost).
L1 L2 L3 L4 L5 L6 L7 L8 L91480 V 1470 V 1460 V 1450 V 1440 V 1430 V 1420 V 1410 V 1400 VFigure 6 (left) shows an example of HV conditioning of the chamber at the first power up(here at 700 V). The residual currents drawn by the HV channels to correctly polarize the dielectricmaterials of the endplates reached a value of ∼
10 nA/cell, starting with a value more than a factorof 300 higher. The characteristic time was about three hours.The CDCH working length was experimentally determined through systematic HV tests atdifferent lengths/wires elongations, adjusted through geometry survey campaigns with a lasertracker ( ∼ µ m accuracy). The final length was set to +5.6 mm of wires elongation with respectto the zero tension position, corresponding to ∼
70% of the elastic limit. This guarantees anelectrostatic stability safety margin of ∼
100 V above the WP. The map showing the HV valuereached by each drift cell is reported in figure 6 (right). The cells which did not reach the WP areconsidered inefficient. The measured cell inefficiency was 1 . e + reconstruction.CDCH was finally integrated into the MEG II experimental apparatus for the first time in2018, as shown in figure 7 (left). Complete signal/HV cabling and gas inlet/outlet connections tothe final MEG II gas system were performed. A gas analyzer at the chamber outlet monitors thecontaminants. An example for the Oxygen content as a function of time since the starting of thegas mixture flux is shown in figure 7 (right). The cooling pipes were also routed and the coolingsystem was successfully tested. After the 2018 and 2019 engineering runs the CDCH geometrywas measured and the CDCH mechanics proved to be stable and adequate to sustain a MEG II run. Electronegative impurities, like O , can affect the e − avalanche development inside the drift cell, causing gas gainfluctuations. This is the so-called electron attachment. – 5 – igure 6 . Left: example of HV conditioning of the chamber at the first power up. Right: final HV map atthe working point. The color scale ranges from 1150 V to 1480 V. Figure 7 . Left: CDCH fully integrated into the MEG II experimental apparatus. The signal cables arevisible, together with the gas/cooling system pipes. Right: O contamination vs. time at the gas outlet. After completing the beam line and starting the gas mixture flux, CDCH was ready for data taking.Only 192 DAQ channels were available for 2018 and 2019 runs. This corresponds to six layers inone sector (16 drift cells per layer) with the double-side read out. Due to the limited number ofDAQ channels a particle tracking test was not possible. Nevertheless, we studied the noise level inthe experimental environment and performed several HV scans around the WP with Cosmic Rays(CR) and with the µ + beam at different intensities: ∼ stopped µ + /s (low rate), 3 × stopped µ + /s (rate during the phase 1 of MEG), 7 × stopped µ + /s (rate planned for MEG II). CR dataallowed the first experimental gas gain studies in a clean environment. Michel e + data allowed totest the chamber response in a high rate environment and the combined data taking with the otherMEG II sub-detectors (ref. [5]).Figure 8 shows an example of a CR event display at the HV WP (left) from layer 1 (L1, outer)to layer 6 (L6, inner) and the corresponding occupancy plot integrating 25000 events. One drift cellhas a hit if the WaveForm (WF) exceeds a predefined threshold: typically × ∼ igure 8 . Left: CR event display at the HV WP. Right: occupancy plot integrating 25000 events. Figure 9 . Left: typical WFs as measured at both ends of two adjacent cells. Right: L1-L6 hit occupancy. adjacent cells (left) and the occupancy as a function of the first six layers (right). The hit scalingas expected from MC simulations was correctly observed. Another interesting plot is reported infigure 10 which shows the ∼
18% width scaling of the raw hit time distributions for L1 (left) andL6 (right). This is related to the different drift cell dimensions: 7.54 mm and 6.40 mm for L1 andL6 respectively at the CDCH center ( z =
0, figure 1). Figure 11 shows the distribution of the signalWF amplitude from CR data as a function of the same HV applied to L1, L2, L3.Distributions were fitted with a gaussian pedestal + Pólya distribution for signal. The meanamplitude and thus the separation from pedestal increase as the HV was set to higher values. Themean amplitude is higher for L3 than L1 at fixed HV, given the higher gain for inner layers (smallercells). This is the reason for the HV scaling shown in table 1. The mean amplitude was then Typical shape from the avalanche statistics. – 7 – igure 10 . Hit time distributions for L1 (left) and L6 (right). The ∼
18% width scaling is related to thedifferent drift cell dimensions.
Figure 11 . Distribution of the signal WF amplitude vs. HV for L1, L2, L3 from CR data. These plots wereused to extract the first gas gain estimate (figure 12 left). converted into the effective gas gain G by means of simulations of the ionization clusters and theresponse of the FE amplification stage. The first calibrated gain curves, as extracted from CR data,are reported in figure 12 (left), showing agreement with simulation at the HV WP.As aforementioned, also µ + beam data were collected. Figure 12 (right) shows the currentdrawn by a HV channel in [ µ A] as a function of the HV applied around the WP (from WP - 30 Vup to WP + 10 V in step of 10 V) at a beam intensity of 7 × µ + /s (nominal MEG II rate).The experimental gain curve showed a nearly exponential behaviour as expected from gas gainsimulations.During the 2018 and 2019 engineering runs we experienced anomalous high current levels insome sectors and layers, starting with the µ + beam. The complete power down of the chamber wasnecessary to drive the big currents to zero. This behaviour needs to be carefully understood.– 8 – igure 12 . Experimental gain curves as extracted from CR data (left) and from a HV scan with µ + beam atthe MEG II intensity (right). At present the commissioning phase at PSI is still ongoing. A full engineering run with allthe upgraded detectors and the complete DAQ electronics is expected to start in 2020, followed bythree years of physics data taking.More details about the CDCH commissioning and first data taking can be found in ref. [11].
References [1] A. M. Baldini et al.,
Search for the lepton flavour violating decay µ + → e + γ with the full dataset ofthe MEG experiment , Eur. Phys. J. C (2016) 76:434.[2] L. Calibbi and G. Signorelli, Charged lepton flavour violation: An experimental and theoreticalintroduction , La Rivista del Nuovo Cimento, Volume 41, Serie 5, Numero 2 (2018).[3] F. Cei and D. Nicolò,
Lepton Flavour Violation Experiments , Advances in High Energy Physics,Article ID 282915 (2014).[4] A. M. Baldini et al.,
MEG Upgrade Proposal , arXiv:1301.7225 (2013).[5] A. M. Baldini et al.,
The design of the MEG II experiment , Eur. Phys. J. C (2018) 78:380.[6] A. M. Baldini et al.,
Gas distribution and monitoring for the drift chamber of the MEG II experiment ,JINST 13 (2018) P06018.[7] A. M. Baldini et al.,
Single-hit resolution measurement with MEG II drift chamber prototypes , JINST11 (2016) P07011.[8] M. Chiappini et al.,
The new drift chamber of the MEG II experiment , Nucl. Instrum. Methods Phys.Res. A, Volume 936 (2019) 501–502.[9] G. Chiarello et al.,
The construction technique of the new MEG II tracker , Nucl. Instrum. MethodsPhys. Res. A, Volume 936 (2019) 495–496.[10] A. M. Baldini et al.,
The ultra light drift chamber of the MEG II experiment , Nucl. Instrum. MethodsPhys. Res. A, Volume 958 (2020) 162152.[11] M. Chiappini PhD Thesis (2019): https://meg.web.psi.ch/docs/theses/chiappini_phd.pdf., Nucl. Instrum. MethodsPhys. Res. A, Volume 958 (2020) 162152.[11] M. Chiappini PhD Thesis (2019): https://meg.web.psi.ch/docs/theses/chiappini_phd.pdf.