Electromagnetic Calorimeter for the CLAS12 Forward Detector S. Stepanyan (JLAB) Collaborating institutions: Yerevan Physics Institute (Armenia) James Madison University Ohio University Norfolk State University Orsay-IPN (France) Jefferson Lab CLAS12 Detector review, April 23, 2007 , JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB From CLAS to CLAS12 Nucleon structure in the regime of strong QCD spectrum of baryon resonances transition form factors spin structure functions Meson spectroscopy NN corellations in nuclei Experiments with (polarized) electron (L=2x1034 cm-2 sec-1) and (polarized) photon (L=5x107 photons/sec) beams, using variety of (polarized) targets Most of the personal, who work on the construction of the CLAS, are still around with 10 years of experience in operation and maintenance S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

From CLAS to CLAS12 (cont.) Generalized Parton Distributions Inclusive Nucleon Structure Semi-Inclusive DIS Properties of QCD from Nuclear Medium Baryon Form Factors Baryon and meson spectroscopy Experimental program requires: large solid angle coverage, high luminosity (L=1035 cm-2 sec-1), use of variety of polarized and unpublicized targets. Most of experiments require efficient detection of showering particles and reliable identification of p0’s CLAS12 detector S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

CLAS forward electromagnetic calorimeter (EC) U - plane V - plane W - plane Lead-scintillator sandwich, 16 radiation lengths 39, 10mm thick scintillator and 38, 2.3 mm thick lead layers, each - almost an equilateral triangle, with 8 m2 coverage Three stereo readout planes (13 layers per view), each following scintillator layer have strips rotated to 120 degree relative to the orientation of the previous layer Projected geometry, 36 strips per layer, scintillator strip width 10 to 12 cm, readout from one end via bundle of clear optical fibers In the direction of the shower, readout is segmented into 2 parts: 5 layers per view for “Inner” and 8 layers “Outer” parts LG Inner bundle Inner PMT Outer bundle Outer PMT S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB CLAS EC performance EC cluster reconstruction 10 years of operational experience Calibration procedures are established Cluster reconstruction algorithm is developed and tested extensively Energy resolution DE ~ 0.1E1/2 Time resolution <0.5 ns p0 mass resolution ~13 MeV Successfully used in the physics analysis that used EC for electron, photon, and neutron detection EC was the first CLAS detector to go in Hall B. It will be the first CLAS12 detector as well. EC will be used in the CLAS12 without changes S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

CLAS EC at high energies Two problems will arise at high energies Diminishing energy resolution due to leakage from the back Separation of clusters from high energy p0’gg decay One cluster with E=Ep Both are important for successful execution of the CLAS12 physics program S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Proposal for the CLAS12 calorimetery Install a pre-shower calorimeter in front of the existing EC, with full coverage of the EC acceptance Basic structure - similar to the CLAS EC: lead-scintillator sandwich about 5-6 radiation lengths thick three stereo readout planes finer transverse segmentation of readout planes than in EC Light transport from scintillator to a photo-detector via green wave-shifting fibers, embedded in the grooves on the surface of the scintillator strips The main requirements for the pre-shower are: high detection efficiency for showering particles (e,g) good accuracy for electromagnetic shower energy reconstruction separation of two photon clusters from high energy (~10 GeV) p0 decay Adding more scintillator layers will increase neutron detection efficiency S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

CLAS12 pre-shower project [WBS 1.4.2.2.2.3] R&D (FY06-FY07): full GEANT simulation and reconstruction test of the pre-shower components (scintillators, WLS fibers, and PMTs) construction and test of prototypes determination of the light yield, light attenuation, and the time characteristics of the pre-shower (see also K. Giovanetti’s presentation) PED (FY07-FY08): design of the pre-shower module design of the mounting fixtures design of prototypes (see also D. Kashy’s presentation) S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

GEANT simulation of the PCAL N. Dashyan (YerPhI), K. Whitlow (SULI) Goals: establish design parameters of PCAL determine characteristics of electromagnetic shower and p0’gg reconstruction in the CLAS12 forward electromagnetic calorimeter Tools: GEANT simulation package for the CLAS detector, GSIM, with PCAL positioned in front of EC CLAS event reconstruction package, RECSIS. Modified EC package for cluster reconstruction S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Pre-shower in GSIM The pre-shower (PCAL) was inserted in front of EC Internal geometry of the PCAL: alternating layers of scintillators and lead 10 mm thick scintillator, 2.2 mm thick lead triangular shape layers to match the geometry of the EC three stereo readout views (U, V, and W) In the initial simulations: 15 layers of scintillators (5 per view) and 14 layers of lead the whole stack was sandwiched between two end plates, constructed from composite foam and 2mm thick SS sheets scintillator layers were sliced into 108 strips of equal width S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

PCAL+EC simulations (15 layers and 108 strips) Energy resolution for electrons thrown in the center of the of the calorimeter Efficiency of two photon cluster reconstruction from high energy p0’gg decays S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Configurations with 9, 12, and 15 layers p0 reconstruction efficiency decreases at high energies for 12 and 9 layer modules not enough radiator thickness and some of high energy photons do not convert 12 layers configuration with first 3 layers of lead with double thickness had comparable efficiency but ~10% worse energy resolution 1 module = 3 layers of scintillator-lead S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Different readout segmentations Simulations with 15 layers of scintillators and 14 layer of lead Efficiency of reconstruction of two clusters from p0’gg decay decreases for small number of readout segments For p0 energies up to 10 GeV efficiency of two cluster separation is reasonably high for 4.5 cm wide segmentation S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Readout with variable segmentation Beam All single strips U-double V&W-single W-single V-single U-single W-single, U&V-double V-single, U&W-double S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Variable segmentation with 196PMTs/sector Simulations with15 layers, 4.5 cm strip width Total of 86x4.5 cm wide strips in each readout view (U, V, and W) Finer segmentation (4.5 cm) of short U strips and long V and W strips will allow to achieve desired results with smaller number of readout channels Final number of strips per layer and arrangement of the readout will be performed after design is completed Two cluster reconstruction efficiency 86 single strip readout Single strip readout 66 U, 30 V, 30 W 58 U, 34 V, 34 W 50 U, 38 V, 38 W 42 U, 42 V, 42 W p0 momentum (GeV/c) S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Summary of simulations Initial design parameters of the pre-shower are established using a full GEANT simulations of the CLAS12 electromagnetic calorimeter system Proposed configuration for the pre-shower: 15 layers of the lead and scintillator, 2.2mm lead, 10mm scintillator three stereo readout views, UVW (5 layers per readout view) 4.5 cm segmentation of the scintillator layers in the forward region for all three UVW views. In the backward region, double-strip (9 cm) readout for U view (perpendicular to the beam direction) and single- or double-strip readout for V or W Simulations will continue - final arrangement of the readout segmentation photoelectron statistics realistic attenuation of light in the fibers integration of the PCAL+EC configuration into the CLAS12 simulation package S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Light collection system Light transport from scintillator to PMT via 1 mm diameter green wave-shifting fibers embedded in grooves on the surface of the scintillator strips Does not require high performance scintillator strips Light gets absorbed locally and transported through fibers Low cost extruded scintillator strips with grooves made in the extrusion process can be used Technique is widely used in several large detectors (MINOS, MINERvA …) S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Selection of the PCAL components Scintillator-WLSF-PMT combination with reasonably high photo-electron statistics at low cost Choice for the scintillator, WLS fiber, and PMT is : Fermi Lab extruded scintillators, 4.5x1 cm2 with 3 grooves Kuraray, 1mm diameter Y11 single clad HAMAMATSU R6095 selected with Q.E.>16% @ 500 nm Expected photo-electron yield ~11p.e./MeV for 3 fibers (yield for EC readout from the test measurements was ~8.4p.e./MeV) Details in K. Giovanetti’s presentation S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Prototyping Two main issues will be resolved: performance of the PCAL and refining the design and construction details A small scale prototype, rectangular 5x5 matrix (XXY), total of 15 layers absolute light yield and time characteristics (MIP) shower reconstruction (energy, position) and time characteristics (in the beam) Full scale prototype (not necessary to have all PMTs and readout electronics) workout design details (fiber routes, PMT holders, light guides …) techniques for the fiber gluing and the quality control work out details of the assembly procedure establish a technique for fiber-adapter-PMT assembly attenuation and light yield measurements (MIP) S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Signal readout and electronics Trigger logic Each PMT will be furnished with HV power supply, ADC, TDC, and scaler channel Anode signal of the PMT will be split with 1:3 split PCAL response will be measured in FADC (25% of the signal) Time of the signal will be measure in multi-hit TDC Fast readout of FADC will be used in the trigger logic - FPGA programming in the mainframe controller FADC 1 3 PMT Spliter Discr HV TDC Scaler Details will be discussed in C. Cuevas’s presentation S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Trigger with FADC and FPGA logic Single sided readout and exponential light attenuation lead to non uniform energy response Using FPGA programming, a simple sum of FADCs can be substituted with weighted sum Trigger logic will be used to isolate energy clusters S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

The Geometry of the PCAL Sc layers The last layer of PCAL scintillators projected from the EC will be: Y bU 15.24 hU=360.3 cm, bU=368.97 cm Height for V/W hV=328.43 cm The same size layers and 4.5 cm scintillator strips – Area S=66477.4 cm2 Strip length for a layer S/4.5=14773 cm Total length for a sector 2216 meters The total length of Sc strips: 13296 meters The total length of fibers: 41040 meters (3x13296+1152x5x0.2) yh a hV hU X PCAL EC yl Design details will be discussed in D. Kashy’s presentation S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

PCAL Construction: cost and manpower S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB HAMAMATSU R6095 HAMAMATSU will select PMTs with SK>100 mA/lm Minimal requirements for gain variations QE and Gain performance of the first batch of 10 PMTs was in the acceptable range S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

FNAL-NICADD extruded scintillators Initial test measurements were done using the scintillator strip extruded for MINOS Far detector: 4x1 cm2 scintillator strips with 1 groove in the center New die was ordered for strips 4.5x1 cm2 with 3 grooves S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Construction Schedule If funding and resources are available on time and no delays in the procurement and delivery of components: Requires three air-conditioned, semi-clean rooms for fiber gluing, stacking, and assembly of PMTs Climate-controlled area is needed for the final testing S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Contingency analysis (2007): procurement Cost estimates for the main components of the pre-shower (scintillators, fibers, PMTs, cables, PMT housing) are based on estimates received from vendors and on catalog prices. Possible source of cost increase vendor or product unavailability when construction starts. In this respect the critical elements are scintillators, fibers, and PMTs. If selected components are not available than using backup solutions will increase overall cost at most by ~25%. Component First choice Backup Performance difference Cost difference Total cost difference Scintillator FNAL Kharkov 10% less p.e. 100% increase 14% increase Fiber KURARAY Bicron PMT HAMAMATSU R6095 HAMAMATSU R1450 HAMAMATSU R7899EG PHOTONIS XP2802 15% less p.e. 25% more p.e. 5% less p.e. 55% increase 11% increase S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Contingency analysis (2007): labor Estimate of the labor for the construction of PCAL modules is based on the previous experience with construction of EC. In the PCAL fibers will be glued on the scintillators before stacking. Although we do not anticipate any problems with this arrangement, it may require additional manpower if some of components will be delayed or some of processes will take longer than expected. Better estimates can be made after prototypes are built. In any case we do not anticipate significant increases and estimated contingency on the labor for the construction is ~30%. S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Risk Analysis There is no technological or design risk associated with this project. Proposed detector concept for the pre-shower, light transmission from scintillator to PMT via wave-shifting fibers, is well studied and widely used Vendors for major components of the PCAL are identified. Work with vendors, particularly with FNAL-NICADD and HAMAMATSU, is a big benefit by reducing the over all cost without compromise in the performance. We do not expect any problems with manufacturing of required components However, there might be some risk associated with delay in production and delivery of some of components. Strategy to avoid construction delays are: have backup solution for obtaining needed product have overlap between different stages of the construction increase manpower S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB EH&S issues There are no major hazards in use of materials, high power PS, machine tools or radioactive materials. Assembly of the calorimeter will include work with lead sheets, adhesives, HV power supplies for PMTs. Most of personnel involved in the construction are trained and well familiar with these types of works. There will be some heavy lifting. Each PCAL module will weight ~5 T. Moving of the PCAL modules will be done by qualified personnel using appropriate lifting and moving tools and procedures. One of the requirements for work area is access to overhead crane for these purposes. S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Summary Pre-shower R&D and PED are in progress Main design parameters are established using the full GEANT simulations Components of the PCAL (scintillator-fiber-PMT) are selected More measurements of the light yield and the light attenuation will be done using the new extruded scintillators from Fermi Lab Design and construction of a small prototype is in progress Design and construction details can be worked out with building of a full scale prototype Cost estimate for whole project is completed Contingency and risk analysis are performed Construction stages for the main detector are identified S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB Backups S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB GSIM12 – GEANT 3 package for the CLAS12 TOF-PD1B TOF-PD2 EC TOF-PD1F PCAL S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB

Electron/p- separation e-/p- with P>6 GeV/c Energy detected in the scintillators With DE vs DEPCAL cut P – momentum > S. Stepanyan, CLAS12 detector review, April 23, 2007, JLAB