3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico

Page created by Warren Silva
 
CONTINUE READING
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
3D Crystal Calorimeter: R&D status
Yong Liu (Institute of High Energy Physics, CAS),
on behalf of the CEPC Calorimeter Working Group

HKUST IAS High Energy Physics Conference
Jan. 19-21, 2021

 Yong Liu (liuyong@ihep.ac.cn)
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
Motivations (reminder)
• Background: future lepton colliders (e.g. CEPC)
 • Precision measurements with Higgs and Z/W
• Why crystal calorimeter?
 • Homogeneous structure
 • Optimal intrinsic energy resolution: ~3%/ ۩ ~1%
 • Energy recovery of electrons: to improve Higgs recoil mass
 • Corrections to the Bremsstrahlung of electrons
 • High sensitivity to detect low-energy photons
 • Capability to trigger single photons
 • Flavour physics at Z-pole, potentials in search of new physics, …
• Fine segmentation
 • PFA capability for precision measurements of jets

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 2
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
3D crystal ECAL: past workshops

• Ideas firstly proposed: CEPC calorimetry workshop (March 2019)
• Follow-up workshop: Mini-workshop on a detector concept with a crystal ECAL
 • R&D efforts targeting key issues and technical challenges
 Virtual mini-workshop on a detector concept with a crystal ECAL,
 July 22-23, 2020, https://indico.ihep.ac.cn/event/11938/

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 3
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
R&D efforts targeting key issues and technical challenges
• Key issues: performance studies and optimization
 • Segmentation: in longitudinal and lateral directions
 • Performance: single particles and jets with PFA -> separation, energy splitting
 • Impacts from dead materials: upstream tracker, services (cabling, cooling)
 • Fine timing: e.g. for positioning
 • Dual-gated or dual-readout techniques (to improve hadronic energy resolution)
• Critical technical questions/challenges
 • Detector unit: crystal options (BGO, PWO, etc.), SiPMs (HPK, NDL, etc.)
 • Front-end electronics: cornerstone for instrumentation of high-granularity calorimetry
 • Multi-channel ASIC: high signal-noise ratio, wide dynamic range, continuous
 working mode, minimal dead time, etc.
 • Cooling and supporting mechanics design
 • Calibration schemes and monitoring systems: SiPMs, crystals and ASICs
 • System integration: scalable detector design (modules), mass assembly, QA/QC

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 4
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
3D crystal ECAL: 2 major designs
 Design 1 Design 2

 Crystal bars SiPM FE+PCB Cooling + Support

Incident
particles
 Incident
 particles
 • Long bars: 1×40cm, double-sided readout
 • Fine segmentation: longitudinal & transverse
 • Super cell: 40×40cm (24X0 depth)
 • Single-ended readout with SiPM
 • Crossed arrangement in adjacent layers
 • Potentials with PFA
 • Significant reduction of #channels
 • Timing at two sides: positioning along bar

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 5
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
Longitudinal segmentation optimisation
Boson Mass Resolution vs #Layer in ECAL PFA Fast Simulation
 Yuexin Wang (IHEP)

 Full simulation with
 SiW ECAL (CDR), by
 Yukun Shi (USTC)

 • Full simulation with SiW-ECAL via the benchmark Higgs to 2 gluons
 • 10 longitudinal layers or more in ECAL can help achieve better than 4% of BMR
 • Expect small impact from ECAL intrinsic energy resolution (PFA fast simulation)
 • Guidance for the longitudinal segmentation
 • Will perform more benchmark studies for crystal ECAL in the CEPC detector simulation

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 6
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
Longitudinal segmentation: impact from services
• Energy resolution with different numbers of sampling layers
 • 24X0 total depth for crystals: fixed in all scenarios
 • Used copper to model the inter-layer services (e.g. cooling)
 • Light materials will be considered for realistic cooling designs: Al, carbon-fibre…

 Note: energy fluctuations and leakages dominate; impacts of digitization in the next page

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 7
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
Longitudinal segmentation: with digitisation
 • Digitisation tool
 • Photon statistics (crystal and SiPM): reasonably high light yield
 • Electronics resolution for single photons: taken from the existing ASIC
 Note: for copper, X0=14.36mm, 1mm Cu = 0.07X0;
 6 longitudinal layers in depth: 4X0 crystal/layer for Aluminum, 0.07X0 = 6.2mm (X0=88.97mm)

 Cu: 0mm/layer Cu: 1mm/layer Cu: 2mm/layer

• 0.84X0 copper in total will degrade the stochastic • Require minimum dead materials from mechanics,
term to ~2.5% (aim of 3%) cooling, etc.

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 8
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
3D crystal ECAL: 2 major designs
 Design 1: short bars Design 2: long bars (current focus)
 Crystal bars SiPM FE+PCB Cooling + Support

Incident
particles
 Incident
 particles
 • Advantages
 • Design optimizations • Fine longitudinal granularity: 24 layers; minimum dead materials
 • Save #channels: ~15 times less
 • Transverse: separation power • 3D calorimeter: timing for hit positions for transverse granularity
 • Longitudinal: impact from dead material • Key issues
 • Ambiguity: multiple incident particles within one super cell
 • Neutral pion reconstruction • Separation of nearby showers
 • Impact on the Jet Energy Resolution (JER)

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 9
3D Crystal Calorimeter: R&D status - HKUST IAS High Energy Physics Conference Jan. 19-21, 2021 - CERN Indico
Studies on physics requirements: reminder Yuexin Wang (IHEP)
• Estimate the multiplicity level of jets: fast simulation
 • Detailed studies with incident particles (from a jet) hitting the hottest tower

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 10
Jets in Event Display Di-jet events from → 

• Impressions of topology of EM/hadron showers within jets
• Intuitive guidance for the reconstruction development
• Current strategy: first studies with (close-by) EM showers, then hadron showers
 (due to the intrinsic complexity)

 Multiple gammas and a charged pion Multiple gammas and hadrons

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 11
Reconstruction with 2 incident particles
 Patterns in event display: 2 photons Shower profiles: 2 photons
 Longitudinal Profile

 3 cm Lateral Profile: timing

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 12
Shower profiles studies in simulation
• Key question: how to separate two close-by EM showers
• EM shower profiles in 3D with highly granular cells: ongoing studies
 • Input to the weights for energy splitting
 Lateral shower profile:
 100GeV electrons

 1 ( , )
 = + 1 − ( )
 ( ) 

 Longitudinal profile:
 100GeV electrons

 Distance along radial direction from
 shower axis (scaled in Moliere radius)

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 13
Separation studies: 2 close-by gammas
 • Key question: how to separate two close-by EM showers
 • Benefits of fine granularity
 • Energy weighted positions (near shower maximum): useful to separate close-by showers
 2 gammas separated by
 40mm in both X and Y

 2 gammas
 (10 GeV)

 Side View: fine granularity
Dashed lines as directions
of incident gammas 20mm distance in X,Y 10mm distance in X,Y

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 14
Latest progress in the new CEPC software Fangyi Guo, Baohua Qi (IHEP)

 • Crystal calorimeter in CEPCSW (common framework)
 • Geometry of long crystal bars: implementation in DD4HEP (done)
 • Digitisation: based on the Geant4 full simulation of a single crystal bar (done)
 • Reconstruction with single particles and two close-by showers: ongoing

Crystal ECAL geometry (barrel) in CEPCSW 400×10×10 mm3 BGO Crystal
 SiPM muon ESR wrapping SiPM

 Time stamps of
 Time stamps of
 1st photons
 detected photons

 Geant4 full simulation established

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 15
Latest progress in CEPCSW Fangyi Guo (IHEP)

• Barrel ECAL implemented
 Inner radius = 1.8 
 • BGO crystal bars: 1cm × 1cm × ~40cm Length along beam line = 4.6 
 • Every 2 layers of vertically intersected bars Total depth = 28 
 • Readout at two ends

• General information BGO: 0 = 1.12 , = 2.23 
 • 8 identical staves (trapezoids)
 • Avoid projectile cracks pointing to the IP − 
 • Ideal layout: excluding electronics and
 mechanics
 • Gaps identified Gaps in − and
 z-x planes
 Minimum Δ = 0.85°
• DD4Hep used for geometry construction z− Minimum Δ = 0.89°

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 16
Reconstruction algorithm in CEPCSW Fangyi Guo (IHEP)

 General flow-chart: work in progress

 • Sort out information to feed PFA

 • Used the knowledge on the
 traditional crystal calorimeter
 (e.g. BESIII EMC)

 • Working on the implementation
 in CEPCSW
 • Performance check to be done

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 17
Critical questions for 3D crystal ECAL: technical part
• Detector unit: crystal options (BGO, PWO, etc.), SiPMs (HPK, NDL, etc.)
• Front-end electronics
 • Cornerstone for successful instrumentation of high-granularity calorimetry:
 e.g. CALICE prototypes, CMS HGCAL project
 • Multi-channel ASIC: high signal-noise ratio, wide dynamic range, continuous
 working mode, minimal dead time, etc.
• Cooling and supporting mechanics design
 • Power consumption (solid inputs from electronics)
 • Impacts of cooling structure to performance
• Calibration schemes and monitoring systems
 • For SiPMs, crystals and ASICs in the long term
• System integration: scalable detector design (modules), mass assembly, QA/QC

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 18
Test stand for crystal-SiPM readout
 • Key questions
 • To quantify performance of crystal bars and SiPMs (esp. timing)
 • Validation of Geant4 full simulation, used for digitization
 • Infrastructure mostly ready, mechanics for long crystals in production
 • To test with different SiPMs (types, vendors) and wrapping foils

BGO crystals

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 19
First studies NDL-SiPM: new generation Light yield studies

 NDL-SiPMs Latest prototype
 11-3030C-S
 Parameters NDL 22-1313-15S
 Breakdown Voltage 27.5 V 19 V
 Pixel Pitch 10 µm 15 µm
 Peak PDE 31% @420nm 45% @400nm
 Pixels 90k 7.4k ~16.6 p.e./MIP with
 Sensitive Area 3×3 mm² 2.6×2.6mm² 2.6x2.6mm²SiPM

 MIP response with
11-3030C-S 19.5 p.e./MIP 16.6 p.e./MIP
 10x10x45 mm³PWO bar
 • Tests made for a NDL-SiPM prototype of the latest generation
 • Many improvements: lower dark-count noise, higher PDE,…
 • Cosmic muon tests: light yield to be improved
 • Further tests planned with new NDL-SiPMs
 PbWO crystal (SIC),
 • High density: 3×3 mm², 6µm, 245k pixels, PDE~30% (e.g. for BGO) 10x10x45 mm³
 • Large area: 6×6 mm², 15µm, 170k pixels, PDE~40% (e.g. for PWO)

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 20
Crystal studies: near-future plans
• BGO and PWO crystal samples: varying dimensions, surface treatment
 • With a major focus on timing performance: e.g. Cherenkov photons
 Two long BGO bars
 ready (1×1 ×40 cm³)

 Short BGO bars with
 different dimensions
 and surface treatment

 PWO crystals (SIC)
 also ready

 BGO crystal samples already delivered to IHEP (photo by courtesy of Junfeng Chen, SIC-CAS)

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 21
Front-end electronics for SiPM readout U. Heidelberg, IHEP

• ASIC “KLauS”: developed within the CALICE collaboration
 • Designed by U. Heidelberg (KIP), originally for CALICE AHCAL
 • Promising candidate: 36-channel, low-power
 • Excellent S/N ratio: stringently required by high-dynamic SiPMs (small pixels)
 • Continuous working mode: crucial for circular colliders (no power pulsing)
 • Need to quantitatively verify its performance and power consumption

 Wire-bonded
 Klaus5 chip Klaus5 in BGA

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 22
Klaus5 tests with NDL-SiPM Single photon calibration

• NDL-SiPM features: high pixel density (>10k pixels/mm²) and high PDE
 • Requires high S/N ratio in electronics to resolve single photons, due to the low intrinsic
 gain of small pixel pitches
• Klaus5 tested and proved to be able to resolve the single photons (32fC/p.e.)
 • Benefits from its high S/N ratio and high resolution
 Single photon spectrum in 12-bit ADC mode: after corrections

 NDL-SiPM: nominal gain
 2 × 105 with 10µm pixels

 Single photon spectrum in 10-bit
 ADC mode: can not be resolved

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 23
Klaus5 dynamic range: charge injection Dynamic Range

 • Testing of all 36 channels
 • Different working modes (high gain and low gain)
 • Dynamic range: ~550pC as the maximum charge (preliminary results)

 With 5nF capacitor

 With 1pF capacitor

Adapter PCB to inject charge
pulses injection to 36 channels
 ADC after pedestal subtraction ADC after pedestal subtraction
 (mid High Gain mode) (ultra Low Gain mode)

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 24
Klaus5 dead time measurements Potential for continuous mode
 Injection Pulses (from an Arbitrary Function Generator)

 Time Interval
 X: 200ns/div
 Y: 100mV/div

 Pulses from ASIC probe: after preamp + shaper, before ADC

 Tested at IHEP (at 20 ℃)
 X: 200ns/div
 Y: 200mV/div Results also
 at arXiv:2005.08745

• Varying time interval between 2 injection pulses: 100ns - 10µs
• When time interval > 500ns, 100% efficiency of separating the two pulses
 • Promising feature for 100% duty cycle ( required by circular colliders)
 Note: tests were made for a single channel;
 further tests with all 36 channels will be performed

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 25
Summary
• 3D crystal ECAL: highly segmented crystals
 • Aim to achieve optimal energy resolution and PFA performance
 • Key issues for optimization and technical challenges identified
 • R&D progress ongoing
 • Simulation studies on the detector layout with long crystal bars
 • Geometry and digitization implemented in CEPCSW
 • Developing reconstruction algorithm, shower studies in Geant4
 • Technical developments
 • SiPMs and crystals
 • Characterisations of SiPM-dedicated low-power readout ASIC (KLauS)
• Welcome broader collaborations: synergies expected
 • Early R&D stage, many open questions/issues
 • Work in the common software framework (Gaudi, Key4HEP, DD4HEP, …)

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 26
Backup slides

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 27
Studies on physics requirements Yuexin Wang (IHEP)
• Estimate the multiplicity level of jets: fast simulation Hottest 40cm×40cm tower
 • Mean ~4 particles within the hottest tower

 Particle E > 0.2GeV

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 28
Shower lateral profile: layer-wise
 • EM shower lateral profiles
 • Histograms stored for each
 longitudinal layer
 • “Normalised” to the shower maximum
 • Need to locate the shower axis
 beforehand
 • Assign weights for energy splitting in
 the same bar
 • Other trials
 • Tried to use models to fit the curves

 Average radial energy profiles

2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 29
Crystal calorimeter: simulation and digitization Fangyi Guo (IHEP)

• Simulation performed with Geant4 in CEPCSW
 • Electromagnetic interactions (current focus)
 i

• Digitization for each long crystal bar { − , − } { + , + }
 • Readout at two ends: charge Q and time T
 • Contribution from the i-th G4step
 L
 ±zi
 • Qi± = E0 ∙ exp(− 2
 ); T±i = T0 + Gaus(z±
 i
 /v, σT )
 LAtt
 • For a full bar:
 • Q ± = σstep Qi±
 • T± = T±k | σki=1 Qi± > ϵQtot
 ± , ϵ = 5%.
 • Considering a simplified scenario: no light
 attenuation along bars (LAtt = ∞), Q ± ∝ Etot

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 30
Hit reconstruction in CEPCSW Fangyi Guo (IHEP)

• Hit reconstruction: locating hits in 2 crossed bars.
 ( + )
 • Position: ( , , )
 2
 • Energy: use energy distributed in crossed bars as a energy fraction:
 
 • = × + × , where = , =
 Σ 

• MC-Truth level hits: merge G4steps in each 1*1*1 cm3 cube as a truth hit.
 1− 2
• Time information: = + .
 2
 • Match with to veto ghost hits.

 Reconstruction

 2 layers of 1 × 1 × 40 cm3 1 layer of 1 × 1 × 2 3
 crystal bars (40+40) 3 crystal cells (40 × 40 )

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 31
Geant4 full simulation for a single crystal bar Baohua Qi (IHEP)

 SiPM1 SiPM2
• Information extracted from G4 BGO crystal
 Muon
 • Energy deposition: mean ~10 MeV/MIP, determined by crystal thickness
 • #scintillation photons
 • #detected photons at either SiPM
 • Time stamp of each detected photon
 • T0: shooting of the primary particle (muon)
 • Included: scintillation time (~hundreds ns), propagation time (a few ns) within the
 crystal bar
 • Excluded: timing uncertainties from SiPMs and electronics
• Digitization
 • Timing: Choose the time stamp of the 1st photon detected at each SiPM
 • #detected photons : proportional to energy deposition

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 32
EM shower studies with crystal bars in Geant4
 • Two sets of reconstructed positions
 • Hits (energy weights) and timing difference

 • Reconstructed positions from hits
 • Fine granularity: 10mm
 • Even layers: Y
 • Odd layers: X
 • Reconstructed positions from timing
 • Complementary to hit positions
Gammas • Even layers: X
(10 GeV)
 • Odd layers: Y
 • Constraints from timing resolution

 2 gammas separated by
 40mm in both X and Y Side View
 (one quarter )

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 33
Studies with PWO crystal bar and NDL-SiPM
• Firsts tests with a PWO crystal using cosmics PbWO Crystal
 • Read out with a 3x3mm²SiPM, 90k pixels (not ideal sensor) SiPM
 • SiPM designed by Novel Device Lab (NDL) in Bejiing Normal University Bottom Trigger
 muons

 #events
 Example of pre-cut Tyvek paper
 MIP response: ~20pe/MIP
 with 3x3mm²SiPM

Wrapped with Tyvek paper and ESR
PbWO crystal (produced by SIC), 10x10x45 mm³
 #p.e. at SiPM
 NDL-SiPM 3x3mm²with 10um pixels
 Note: a larger SiPM (e.g. 6x6mm²or larger) can be used for better light collection efficiency

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 34
Geant4 full simulation for a single crystal bar Baohua Qi (IHEP)

• To quantify the timing resolution at different hitting positions
• To provide input for the digitization model in CEPCSW

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 35
Patterns with jets in Event Display
 3. Reconstructed hits with
• Patterns for first impression, but still complex… energy information
• Need further studies on positioning and energy splitting

 2.Reconstructed hits using timing

 Red: high E → Purple: low E

 4. Reconstructed hits with
 energy > 4MIPs

 2021/1/20 Yong Liu (liuyong@ihep.ac.cn) HKUST IAS High Energy Physics Conference 36
You can also read