Technology Choices for the sPHENIX Calorimeter Systems C.Woody For the PHENIX Collaboration CALOR12 Santa Fe, NM June 5, 2012
Dec 28, 2015
Technology Choices for the sPHENIX Calorimeter Systems
C.WoodyFor the PHENIX Collaboration
CALOR12Santa Fe, NMJune 5, 2012
3
Detector Requirements that Determine Technology Choices
Technology Choices:• EMCAL → Tungsten Scintillating Fiber Accordion • HCAL → Iron Scintillating Tile with WLS Fiber • Readout → SiPMs
Detector Requirements • Large solid angle coverage (± 1.1 in h, 2p in f)• Moderate energy resolution
• EMCAL ~ 15%/√E • HCAL ~ 50-100 %/√E (single particle)
• Compact (for EMCAL small RM, short X0)• Hermetic• Projective (approximately)• Readout works in a magnetic field• Low cost
C.Woody, CALOR12, 6/5/12
C.Woody, CALOR12, 6/5/12 4
Optical Accordion
• Volume increases with radius
• Scintillator thickness doesn’t increase with radius, so either tungsten thickness must increase or the amplitude of the oscillation must increase, or both
• Plate thickness cannot be totally uniform due to the undulations
• Small amplitude oscillations minimize both of these problems
Layered accordion of tungsten plates and scintillating fibers
Accordion design similar to ATLAS Liquid Argon Calorimeter
Want to be projective in both r-f and h
C.Woody, CALOR12, 6/5/12 5
Sintered Tungsten Plates
Variable thickness W-platesScintillating fibers in between
Density r ~ 17.5 g/cm3
Problem is that cannot make sintered plates larger than ~ 20 cm
Phase I SBIR with
C.Woody, CALOR12, 6/5/12 6
Tungsten-SciFi Epoxy Sandwich
Scintillating fibers 1.0 mm
Pure tungsten metal sheet (r ~ 19.3 g/cm3) Thickness: 2x1.0 mm
Tungsten powder epoxy ( r ~ 10-11 g/cm3) 0.08-0.2 mm
Uniform thickness, thin pure tungsten metal sheets with wedge shaped SciFi + tungsten powder epoxy layer in between
Can be made into larger modules (> 1m)Fabricate in industry
2 W plates/layer 0.6 X0 sampling
X0 = 5.3 mm RM = 15.4 mm
C.Woody, CALOR12, 6/5/12 7
Effect of Glue on Light Yield
• Gluing fibers reduces light output due to loss of cladding light
• Depends on glue
• Does not seem to depend on whether glue contains tungsten powder or not
glued 6x6 fiber bundle
adc channels
0 100 200 300 400
coun
ts
0
100
200
300
400
500
bundle on pmtfit (mean=202ch ~ 100 pe/MeV)
With a sampling fraction of 4% and 100 p.e./MeV in scintillator
4000 p.e./GeV in the calorimeter
100 p.e./MeV
Direct fiber readout on PMT
C.Woody, CALOR12, 6/5/12 8
Light Yield and Readout DevicesWant to have small photostatistics contribution to the energy resolution
Need sufficient light output from fibers to allow randomizing and collecting the light onto a small readout device
SiPM
ScintillatingFibers
Possible Readout Schemes
Small reflecting cavity or wavelength shifting block
Need to match ~126 1 mm diameter fibers onto a 3x3 mm2 SiPM with good efficiency and uniformity (earea ~ 2%)
21 mm
21 mm
C.Woody, CALOR12, 6/5/12 9
Module Construction
Tungsten-SciFi “sandwiches” are cast together in 6 layers to form a module ~ 2 cm “tower” in r-f, ~ 9.5 cm depth, ~ 1.4 m long (L/2)
Modules assembled in groups of 4 to form sectors (~ 400 lb each)64 sectors arranged azimuthally to cover 2p (x2 for both sides)
C.Woody, CALOR12, 6/5/12 10
Hadron Calorimeter
• Steel plates with scintillating tiles parallel to beam direction Tapered steel plates sampling fraction changes with depth• Divided into two longitudinal sections Measure longitudinal center of gravity correct for longitudinal fluctuations • Plates tilted ± 5o in opposite directions to avoid channeling • Iron in steel serves as flux return
4 inner and 4 outer plates joined together to
form one section
60 cm3.5 labs
30 cm1.5 labs
C.Woody, CALOR12, 6/5/12 11
HCAL ReadoutSimilar to T2K: Scintillating tiles with WLS fibers embedded in serpentine grooves (Scint:extruded polystyrene made in Russia, WLS:Kuraray Y11)
Fibers embedded in grooves on both sides of tile Expect ~ 12 p.e./MIP/tile (T2K) ~ 400 p.e./GeV in HCAL
T
2x11 segments in h ( Dh =0.1)64 segments in f ( Df =0.1) 1408 x 2(inner,outer) = 2816 towers
2x11 scintillator tile shapes
Inner
Outer
Inner readout(~10x10 cm2)
Outer readout
SiPMs + mixers
8 readout fibers per tower
C.Woody, CALOR12, 6/5/12 12
Testing HCAL Scintillator ComponentsBNL University of Colorado
Russia
Mixer and SiPM readout
Groved tiles
Extruded scintillator with WLS readout
13
SiPM Readout
C.Woody, CALOR12, 6/5/12
Common readout for both EMCAL and HCAL
EMCAL segmentation ~ .025 x .025 (h x f) ~ 25K channels HCAL segmentation ~ 0.1 x 0.1 ~ 3K channels
Considering various devices: Hamamatsu, AdvanSiD, Zecotek,…
Also considering APDs
• Want dynamic range ~ few x 103
10 MeV – 20 GeV per tower
• Due to saturation, must tune light levels to give ~5 – 10,000 p.e.
• Avoid noise a 1 p.e. level
• Requires temperature compensation and control dVbr/dT ~50 mV/°C dG/dT ~ 10% /°C
S10362-33-25C
Zecotek MAPD-3N3x3 mm2, 135K pixels
3x3 mm2,14.4K pixels (25 mm)G ~ 2 x 105, peak PDE ~ 25% @ 440 nm
C.Woody, CALOR12, 6/5/12 14
Readout ElectronicsLarge signals don’t need ultra low noise electronics Can use a conventional voltage/current amplifierDynamic range essentially determined by SiPM
SiPM preamp with differential output
SiPM preamp impulse response (SPICE)
Readout will use either a derivative of an existing PHENIX ADC system from the Hadron Blind Detector or the CERN SRS system interfaced through the Beetle chip
VB Adj
ORNL
C.Woody, CALOR12, 6/5/12 15
Summary
• The upgrade from PHENIX to sPHENIX will require the design and construction of two new major calorimeter systems:
• W-SciFi Optical Accordion EMCAL• Iron Scintillating Tile WLS Fiber HCAL
• Both will implement new technologies made possible through the development of new materials, photodetectors and construction techniques.
• Both calorimeters will build on existing and proven designs, but will also incorporate several new and novel designs features that need to be thoroughly understood and tested.
• We feel we have a good, sound basic design concept, but we have a lot of work to do to insure that these detectors will work as we hope.
C.Woody, CALOR12, 6/5/12 18
EMCAL Energy Resolution
Accordion Layers:• Tungsten metal: 2x1 mm• Scintillating fiber: 1 mm• Tungsten-epoxy: 0.08 – 0.20 mm
Sampling frequency: 0.6 X0 Sampling fraction: 4.2%Total layers: 30Total depth: 17.4 X0
Total thickness: 9.5 cm
sE/E ~ 14%/√E
GEANT4 Simulation
C.Woody, CALOR12, 6/5/12 19
Single Particle Energy Resolution
No CoG weighting
No CoG or E/H weighting
C.Woody, CALOR12, 6/5/12 20
Total Calorimeter Jet Resolution
Central Au+Au events (HIJING) w/o detector resolution but including underlying events
PYTHIA + Full GEANT4 simulation Recon: FastJet anti-kT, R=0.2
Single jet resolution in p+p collisions
PYTHIA + FastJet + jet energy smearing