Conseil Scientifique et Technique du SPhN RESEARCH PROPOSAL

Transcription

1 Conseil Scientifique et Technique du SPhN RESEARCH PROPOSAL Title: MICROMEGAS CENTRAL TRACKER FOR CLAS12 Experiment carried out at: Jefferson Laboratory Spokes person(s): Contact person at SPhN: J. Ball Experimental team at SPhN: J. Ball, P. Konczykowski, B. Moreno, H. Moutarde, S. Procureur, F. Sabatié List of IRFU divisions and number of people involved: SEDI (8), SIS (3-5) List of the laboratories and/or universities in the collaboration and number of people involved: Central Tracker Collaboration : Univ. New Hampshire (3), Moscow State Univ. (1), JLab (5) SCHEDULE Possible starting date of the project and preparation time [months]: 2010 Total beam time requested: Expected data analysis duration [months]: REQUESTED BUDGET Total investment costs for the collaboration: 29 M$ for CLAS12 Share of the total investment costs for SPhN: 4% Investment for SPhN: 919 k Total travel budget for SPhN: 180 k Travel budget/year for SPhN: 36 k If already evaluated by another Scientific Committee: If approved Allocated beam time: Possible starting date: If Conditionally Approved, Differed or Rejected please provide detailed information:

2 MICROMEGAS CENTRAL TRACKER FOR CLAS12 Introduction: The electromagnetic probe has been, for many decades, the tool of choice to analyze the nucleon. The nucleon, which was the basic studied object by accessing its inner components, quarks and gluons, has become a metalaboratory in itself. In parallel to the instrumentation progress, new theoretical tools have been developed to understand the nucleon structure at deeper and deeper levels. While elastic electron scattering enabled to measure nucleon form factors which reflect the spatial shape of charge distributions, deep inelastic scattering experiments (DIS) measured parton distribution functions which determine longitudinal momentum and helicity distributions. The Generalized Parton Distributions (GPDs) formalism includes both former approaches and carry more information about the dynamical degrees of freedom inside the nucleon. GPDs [1-4] enable to access the spatial distribution of quarks and gluons inside the nucleon and also the quark orbital momentum which might be one of the keys to solve the nucleon spin issue. The sophistication of the achieved experiments has followed closely the improvement of the technology needed to produce high precision beams while the reachable energies increased over the years. The Saclay group has been strongly involved for ten years in the first GPD s related experiments through Deeply Virtual Compton Scattering (DVCS) in Hall A and B at Jefferson Laboratory (JLab). GPD s were one of the main issues which boosted the decision to increase the JLab electron accelerator energy from 6 to 12 GeV. This proposal is following upon the ending R & D Project, initiated at the end of 2006, on the feasibility of a Central Tracker for CLAS12 based on the use of bulk Micromegas detectors. CLAS12 is the upgraded version of Cebaf Large Acceptance Spectrometer (CLAS) in Hall B at JLab, it will have the required features to match the the higher energy range and improve the performances of the existing setup. We recall that all the steps of the R & D Project have been thoroughly documented and reviewed since its start by CSTS s, Irfu Project Management and JLab Internal and External Committees. Our R & D Project had been funded by an ANR grant. 1. Central Tracking in CLAS12. The features of CLAS12 will include: operation at higher luminosity (above /cm 2 /s), higher counting rates, better overall efficiency and Particle ID. In the design, the central detector will be a crucial and challenging equipment as it will be inserted in the warm bore of a 5 T superconducting solenoid around the cryogenic target and able to detect particles in the angular range 35 to 125 with an additional forward part covering the 5 to 40 range (fig. 1). The 5 T solenoid magnet provides the magnetic field necessary to analyze track, has a compensation coil to shield the forward detector from Möller-electrons and can also be used to polarize targets. A set of Time-of-Flight counters will provide the Particle ID in the detector. Eventually the trajectory of charged particles will be analyzed using a central tracking device in which the Saclay group has a main contribution.

3 Figure 1: CAD view of the central detector of CLAS12 in the current stage of design. The central tracking device will mostly detect hadrons in the momentum range from 300 to 1500 MeV/c. In the case of DVCS and other hard exclusive reactions, this corresponds to the detection of the recoil protons. In order to match the momentum resolution of the forward CLAS12 detector, the design value for the fractional momentum resolution at a momentum of 1 GeV/c is 5%. In the same manner, angular resolution should be of the order of 10 mrad or less to match the forward detector. The Jefferson Lab group working on CLAS12 and the Saclay group came to an agreement that the baseline design for the Central Barrel Tracker would consist in a 6 layers hybrid Silicon- Micromegas cylindrical structure which benefits from the specificities of both technologies as foreseen through simulations. From the target outward, the first three layers will consist in the Silicon Vertex Tracker (SVT) displayed in polyhedral arrangement and layers three to six will consist in the cylindrical Micromegas Vertex Tracker (MVT). All six layers are actually regions and contain either two types of U and V stereo strips for SVT or Y and Z strips for the MVT. The central tracker is also to include a central forward part in order to tag all forward tracks (5-40 o ) and improve the vertex and angle resolutions for these tracks. This forward part will consist in 3 disk shaped flat bulk Micromegas double layers (X,Y). 2. Summary of the R & D Project A Micromegas [5] is a micro pattern gaseous detector based on a parallel plate electrode structure and a set of micro-strips for readout, (see Fig.2).

4 Figure 2 : Principle of the Micromegas detector. An incident particle ionizes the gas in the conversion region, and the resulting electrons drift down to the mesh and produce avalanches in the amplification region. The mesh allows a very fast collection of the ions created in the avalanche. The bulk Micromegas [6] is an upgrade of the regular Micromegas. The detailed features can be found in the R & D Feasibility report [7]. The principle of bulk MM is to embed a metallic woven mesh on a PCB. For CLAS12, the foreseen detectors will use thin PCB boards (100 µm) with 5 µm thick copper strips, typically 300 µm wide, as the anode plane.. The whole detector is nearly built in one process as the drift plane has eventually to be glued above the mesh to determine the conversion gap. Being thin the bulk MM is flexible and likely to be used in a cylindrical structure. The R & D Project established first that a cylindrical bulk MM could be used to build a working detector and reached the same spatial resolution (~70 µm - see below-) as a classic flat and thick MM [8]. Prototypes were produced with protocols established to qualify them in terms of quality. They went through characterization measurements followed by procedure tests to validate simulations. The crucial issue concerned the behaviour of the barrel detector in the 5 T magnetic field environment transverse to the intrinsic electric field of the detector which would in a regular MM set up disable the detector, the Lorentz angle of primary electron deviation being too large. A modification of the MM working conditions to overcome this difficult situation was worked out through simulations and could be successfully tested and validated at moderate ( 1.5 T) and at high field (4.7 T). Eventually these studies were published in two papers [7-8]. Micromegas detectors are robust and reliable but they do spark with a probability depending on the hadron content of the particles hitting the detectors and of its rate. Tests were performed in Fall 2009 at CERN, in the framework of RD51. Measurements were taken using the SPS 150 GeV muon and pion beams with a rate of 10 6 particles per spill and they showed that bulk MM sparked with a probability equivalent to regular MM (see Fig.3)

5 Figure 3: Spark probability vs. Gain for various Bulk Micromegas prototypes and regular Micromegas (black dots). Measurements performed at CERN with SPS 150 GeV muon beam. In the foreseen CLAS12 conditions, the working gain would be around 3000 which would yield a sparking probability lower than 10-5, which is acceptable at JLab. The same tests runs showed that as far as spatial resolutions were concerned, bulk and regular ( classic ) MM were found to be equivalent, the average reachable resolution being around 70 µm (Fig.4). Figure 4: Spatial resolution as a function of the mesh high voltage for various Bulk Micromegas prototypes and regular Micromegas (black dots). Measurements performed at CERN with SPS 150 GeV muon beam.

6 Final tests will be done this Summer at JLab to check, in realistic beam and magnetic field environments, the behaviour, towards simulations, of Forward flat bulk MM prototypes. They will be followed by additional tests at CERN PS with low momentum hadrons (pions and protons) to signalize the spark studies on the barrel detector. Eventually, the tracking performances foreseen with the MM tracker were simulated using a GEANT4 code. Tracks were reconstructed in the presence of background and results indicate high efficiency and no loss of resolution [7]. 3. The Proposal: Micromegas tracker for CLAS12 Our proposal, as mentioned earlier, concerns the central tracking device for CLAS12. The Micromegas barrel part associated with Silicon detectors will become part of the upgraded design while the forward part will be an alternate for the basic design. The most recent solution (February 2010) following JLab and Saclay discussions on the available room to set the tracker, considering the acceptance requirements (35 to 125 for central-barrel and 5 to 40 for centralforward), is displayed on Fig.5. Figure 5: Central Tracker Layout with 3 double layers of SVT (in black) and 3 double layers of MVT (in red) as well as the 6-layer FMT (in red). 3.1 Central Barrel Tracker The SVT group suggested using a specific three-layer geometry (Regions CR1 to CR3) inorder to accommodate the Micromegas detectors, with 8, 12 and 16 sectors therefore fixing the radii of the three silicon layers to , and mm for regions 1, 2 and 3 respectively. CLAS12 requires a 35 acceptance even for the forward events of a 30mm long

7 target, therefore the longitudinal location of the detectors right at the 35 mark had to be adjusted, the set of coordinates for the Barrel Silicon tracker is presented in Table 1. Sides Radius (mm) Detector Length (mm) Z coordinate of detector center (mm) CR CR CR Z coordinate of detector end (mm) Table 1: Silicon detectors layout, also presented in Figure 5. Using the Silicon layout as a starting point, the Micromegas layout consists in 6 detectors covering as much radial distance as possible. They are staged in 3 regions (CR4 to CR6) from radii 120mm to 225mm, leaving respectively 15mm clearance to the Silicon detectors and 25mm clearance to the CTOF. Between CR3 (Si) and CR4 (MM), we must introduce a thin opaque cylinder to keep the silicon isolated from light. This cylinder may also be conductive in order to provide some EMI shielding to protect the Micromegas from the silicon electronics. Note that CR4 may have to be moved slightly further away from the Silicon if necessary, but no more than 5mm for a total of 20mm. In order to provide the maximum lever arm for the measurement of the polar angle θ (provided by the Y detector), which constitutes the main advantage of Micromegas, the very first and last of the 6 detectors are foreseen to be Y detectors. In order to minimize the number of electronics channels, we decided to use the following sequence from inner to outer: Y, Z, Z, Y, Z, Y. Garfield simulations indicated that the spatial resolution for Z layers is dominated by statistical fluctuations in the ionization process, in the presence of a Lorentz angle. Therefore, the spatial resolution depends only weakly on the pitch, as illustrated in Figure 2 (left). The situation is slightly different for Y layers, where a smaller pitch can be beneficial for the resolution, asseen in Figure 6 (right). Figure 6: Spatial resolution of Micromegas detectors in the Z (left) and Y (right) configurations, as a function of the pitch, and for different Argon based gas mixtures. We therefore decided to adopt the following pitches for our Micromegas detectors: Pitch Y = 270µm, Pitch Z = 540µm.

8 This low Y-pitch needs to be confirmed on a prototype, the decrease of the pitch is not linear with the extension of the price when laser serigraphy and drilling is needed) Table 2 details the location of the 6 Micromegas detectors constituting Central Regions 4 through 6: Radius (mm) Detector Full Length (mm) Detector Active Length (mm) Z coordinate of detector center (mm) Z coordinate of detector (mm) CR6-Y CR6-Z CR5-Y CR5-Z CR4-Z CR4-Y end Table 2: Micromegas detectors layout, also presented in Figure 5. Note that the full length of the Micromegas detectors do not have to be used in order to accommodate the 125 acceptance in the back direction. The back-ends of the detectors have been aligned as a way to simplify the mechanics. In order to calculate the number of electronics channels actually needed, only the active length of the detectors will be used (fourth column in Table 2). The CR4 mechanics will most likely extend downstream up to CR5 and CR6 where they will meet their Rohacell fixation. 3.2 Forward Vertex Tracker The central tracker working group decided to abandon the possibility to use Silicon tracker in the forward direction, as it is both impractical due to cooling requirements, but also precludes from using Micromegas detectors for CR4 to 6. The Forward Micromegas Tracker (FMT) is composed of 3 regions of 2 detectors each. For now, we envision using a X-Y configuration for each region, rotated by 60 from one another. Each detector would then be a 215mm radius disk, and all 6 detectors would be equally spaced in Z from 290 to 340mm from target center. Note that the HTCC detector is expected to start at 380mm from target center in the current design. In order to accommodate the beam through the FMT, a 29.75mm radius disk would be cut-off in all the detectors. The 40mm free space between the last FMT and the HTCC should be kept free for FMT until a prototype is made to validate a 10 mm thick layer including mechanics and connectors. Realistic simulation and reconstruction have shown that a Forward Tracker largely improves the vertex resolution provided by the Drift Chambers (up to a factor of 5-10 for the impact parameter). However, because of the very limited lever arm of the Forward Tracker, the vertex resolution is mainly determined by the FMT distance to the target and the error propagation from the Drift Chambers. This is illustrated in Figure 7, where resolutions and track matching efficiency are shown as a function of the FMT pitch. We see in particular that the resolution on the impact parameter (the most important one for secondary vertices determination) is degraded

9 only by 35% if the pitch is increased from 300 to 500µm. The resulting resolution still meeting the physics and tracking requirements, a 500µm pitch has been chosen for the FMT. Figure 7: Forward resolutions (using FMT in combination with DC) in φ angle (top left), the z position of the vertex (top right), the impact parameter (bottom left) and the matching efficiency between DC and FMT tracks (bottom right). Table 3 summarizes the location of the 6 detectors constituting the FMT. Inner radius (mm) Outer radius (mm) FR3-Y FR3-X FR2-Y FR2-X FR1-Y FR1-X Z coordinate of detector center (mm) Table 3: Forward Micromegas Tracker layout. FR stands for Forward Region (1 through 3). Each region is rotated by 60 from one another in the current design. Figure 8 has been produced from a simplified CAD file and illustrates the location of the Micromegas detectors in the central tracker (without the SVT).

10 Figure 8: Simplified CAD picture of the Micromegas portion of the Central Tracker, both barrel and forward (green). The baseline solution for the CTOF (light blue) is used in this CAD model. The 5 T solenoid section is drawn in pink. 4. Readout Electronics 4.1 Channel count The channel count for the MM detectors has been re-evaluated for the baseline configuration of the detector taking into account new dimensions and adjusted pitch sizes of the barrel and forward stations. The obtained channel count is as follows: Central Barrel Regions Z Type Pitch Strips (µm) CR4Z 1457 CR5Z CR6Z Central Barrel Regions Y Type Pitch Strips (µm) CR4Y 3422 CR5Y CR6Y Forward Regions Type Pitch (µm) FR Total 24262

11 4.2 Front-end electronics Assuming a 64-channel ASIC, a 360-channel board is the most optimal front-end unit (FEU): 5 ASICs per board. The needed number of ASICs and boards are as follows: Central Barrel Regions Z Type FEU ASIC CR4Z 6 30 CR5Z 6 30 CR6Z Central Barrel Regions Y Type FEU ASIC CR4Y CR5Y CR6Y Forward Regions Type FEU ASIC FR Grand Total Micromegas data The amount of MM data has been estimated for the baseline configuration. The physics background rates of 10 and 20 MHz have been assumed for the barrel and forward regions respectively. The hit multiplicity of 3 was used for all detector stations. The trigger rate of 20 khz has been taken. A zero suppression mechanism with subtraction of a common noise and application of a threshold is assumed. The results are summarized in the table below. Detector Ghost hit Data Throughput khz Byte Mbyte/s CR4X CR5X CR6X CR4Y CR5Y CR6Y FR Total Table 4 Micromegas event size and data bandwidth. The total event size and the data bandwidth are modest. The relatively high ghost hit rate of 62.5 khz in the forward region may result into a pile-up discrimination inefficiency of 1.6% assuming 250 µs shaped signals. 4.4 Back-end Electronics The JLab Hall D Electronics Working Group develops and produces various electronics boards for the Hall D Trigger System. The electronics will be widely used at CLAS12 as well. It seems

12 possible to adapt some of the developments for the backend electronics of the MM read-out system. The back-end electronics would be housed in a single VXS crate and would consist of a Trigger Interface (TI) module, a Signal Distribution (SD) module, a number of Sub-System Processor (SSP) modules and a crate controller CPU, as shown on Figure 9. Figure 9: Various boards needed to implement Micromegas back-end electronics. The interfaces of the back-end electronics are shown on Figure 10, below. Figure 10: Interfaces of the back-end electronics. The TI receives a low jitter system clock and fixed latency trigger signals from the CLAS12 trigger supervisor. It also delivers to the trigger supervisor combined status information (e.g. Busy) of the MM read-out system. The physical layer interface is based on a parallel optic technology. The clock and trigger signals are delivered to the SD board over the VXS backplane. The SD board conveys properly delayed and aligned clock and trigger signals to the SSP boards. It also gathers their status information, combines and sends it to the TI board. These communications happen over the VXS backplane.

13 The SSP modules have been primarily designed to be a part of the pipeline trigger generation electronics. However, it seems that it is versatile enough to be adapted to the needs of the MM backend electronics. It can communicate with up to 32 frontend units (FEU) over bidirectional optical links. Downstream fixed latency communication link will transfer to the frontend units the clock and trigger information (clock will be encoded in data). Over the upstream information the SSP board will receive event data from the FEUs. Three SSPs (3x32=96 channels) should be enough for the entire MM read-out system. The read-out of event data is done by a JLab standard crate controller CPU. High rate 2eSST DMA would be needed to retrieve the data from the SSP boards. The data are sent to the CLAS12 DAQ event builder over the gigabit Ethernet interface of the CPU. The CPU is also used for control and monitoring purposes of the read-out system. This organization of the MM back-end electronics has several advantages. It provides straightforward interfaces with the CLAS12 Trigger and DAQ system. It minimizes development and maintenance efforts. It allows for better management of a pool of spare modules. And it is scalable enough to be extended for the SVT read-out electronics. The back-end modules are in different phases of the design, prototyping and production processes. For what concerns the MM read-out, at this stage we have to verify that the SSP hardware has all needed features to be adapted to our needs and require firmware modifications only. If after the verifications the described organization of the MM backend electronics will be adopted, a minimal set with at least one SSP board will be needed presumably in the second half of 2011 for prototyping purposes. 5. Schedule and Resource The second semester of 2010 will hopefully see the beginning of the Micromegas tracker Project and the end of the R & D phase. This new step is scheduled to reach 2014 with the tracker fully operational as JLab will be commissioning the equipments in the 12 GeV configuration Expected Cost of the Project The tracker consists of a barrel tracker of 3x2 cylindrical layers (X and Y) of about 3 m 2 corresponding to about 19 k electronics channels and a forward detector of 3x2 disks, 1.5 m 2, with 6 k electronics channels. The estimated costs, including spare parts and taking into account contingency at the level of 120%, are summarized below: End of Tech. R & D k Barrel Tracker 149 k Electronics Barrel Tracker 122 k Forward Tracker 96 k Electronics Forward Tracker 149 k Flex cables 74 k Holding structure 25 k General (Gas system, High Voltage) 169 k Missions 60 k TOTAL 919 k (1139 k$)

14 The full CLAS12 Project costs is estimated to be 29 M$, the Saclay contribution would then be around 4% of the total amount Schedule of the Project Goals Segmentation 1 XY Detector Detector Production ASIC design DAQ proposal Drawings + Calculations ASIC design Slow Control ASIC Production Detailed drawings 2 nd DAQ proposal Slow Control Steps PDR ASIC Prod. FDR Integration Test1 Integration Test2 5.Overview of the SPhN/CLAS group activities The group performed in 2008 two large- scale DVCS experiments with CLAS: the continuation of the beam spin asymmetry measurements, with the aim of doubling the statistics, and the high-statistics measurement of target spin asymmetries, the second part of the dedicated DVCS experiment which ran up to September The data collected completes the set from the first part of the experiment, which led to three publications [11-13]. A PhD student from our group (P. Konczykowski) is working on the analysis of the new data collection. The group is also actively participating in efforts with theoreticians toward a global phenomenological analysis of DVCS data in order to extract GPDs in a model independent way; a first paper illustrates the start of these studies [14], further strategies will be investigated like the creation of an interactive GPD fitter web page. Beyond the ensuing analyses, the activities of the group will be focused on the 12 GeV upgrade. The main focus is the measurement of DVCS beam and target asymmetries with CLAS12, F. Sabatie is co-spokesman of approved experiment PR The group also cosigned a proposal (PR ) on the hard electroproduction of pseudoscalar mesons (using the same equipment). Approved proposals on the use of a transversely polarized target in CLAS and CLAS12 have been signed as well. The group does not exclude other investments in CLAS12, beyond the CLAS12 central tracker. The group is then very active and has achieved a very high level of recognition in the Jefferson Lab collaboration and in the hadronic physics community at large. Every two year, the CLAS Collaboration Meeting is held outside of Jefferson Laboratory, we were asked to organize the 2011 meeting which will be focused on CLAS12 in Paris.

15 References [1] D. Müller et al., Fortschr. Phys. 42 (101) [2] A.V. Radyushkin,, Phys. Rev. D56 (5524) [3] K. Goeke, M.V. Polyakov and M. Vanderhaegen, Prog. Part. Nucl. Phys. 47 (401) [4] X. Ji, Phys. Rev. Lett. 78 (610) [5] I. Giomataris et al., Nucl. Instr. Meth. A 376 (1996) 29. [6] I. Giomataris et al., Nucl. Instr. Meth. A 560 (2006) 405. [7] S. Aune et al, Micromegas Vertex Tracker Feasibility Annex to Status Report, R & D for a Central Tracker, submitted to SPhN CSTS June 2009 [8] J. Ball et al., Status Report, R & D for a Central Tracker, submitted to SPhN CSTS June 2009 [9] S. Aune et al., Nucl. Instr. Meth. A 604 (2009) 53. [10] P. Konczykowski et al., Nucl. Instr. Meth. A 612 (2010) 274. [11] FX. Girod et al., P.R.L. 100 (2008), [12] R. de Masi et al., P.R.C. 77 (2008), [13] S. Morrow et al., E.P.J.A. 39 (2009),5. [14] H. Moutarde, Phys. Rev. D.79 (2009)