CALICE Software Data handling, prototype reconstruction, and physics analysis Niels Meyer, DESY DESY DV Seminar June 29, 2009
The ILC Well, the next kid around the block (hopefully...) Precision physics in e+e- collisions at 500 GeV and beyond Page 2
What's CALICE? Page 3
What's CALICE? ILD SID 4th Page 4
What's CALICE? ~300 physicists from 53 Institutes in 17 countries Horizontal approach: contributions to ILD and SiD R&D on 7 technologies Physics prototypes hadronisation studies ILD SID 4th Technological prototypes towards 'Module 0' Page 5
Jet Energy Resolution al H ad - C al EM -C Tr a ck er LEP-style: Measurement per subdetector and mutual weighting / correction Page 6
Particle Flow ILC requirement: unprecedented jet energy resolution of 30% / E Solution: particle flow = measurement per particle Neutral hadron Photons Charged hadrons Page 7
Combined Test Beam Physics prototypes: calorimeter chain with realistic depth for hadron shower studies ECal, HCal, tail catcher Setup allows for scan of position and impact angular Data for different beam particles in large energy ranges on tape already Page 8
Si-W ECal 30 layers, 24 X0 total depth 18x18 cm2 instrumented area per layer, 1x1 cm2 Si-pads 9720 channels Page 9
Sci-W ECal 30 layers, 24 X0 total depth 18x18 cm2 instrumentation per layer from 1x4.5 cm2 scintillator strips 2160 channels with individual MPPC readout Page 10
AHCal 38 layers, 4.5 total depth 90x90 cm2 instrumetaion per layer from 216 scintillator tiles of 3x3 to 12x12 cm2 (last 8 layers only 141 tiles) 7608 channels with individual SiPM readout Movable / rotatable stage together with ECal Page 11
TCMT 32 layers with two sampling thicknesses 1x1 m2 instrumentation from 100x5 cm2 scintillator strips 320 channels with individual SiPM readout Page 12
Data composition Almost 20,000 channels, no hardware reduction, ~40kB / evt several hundred million events recorded since 2006 Page 13
Imaging Calorimetry TCMT AHCal SiW ECal Page 14
Data Flow online Experimental site Online monitor Raw data (.bin) MySQL server Raw data (.bin) Translation and event building Conditions from DAQ stream, e.g. slow control Calibration and correction Manual written, e.g. cable map Analysis Offline extracted, e.g. calibrations and alignment Converted (.slcio) Reconstructed (.slcio) Page 15 offline GRID
ILC Core Software Raw data (.bin) Translation and event building Conditions from DAQ stream, e.g. slow control Calibration and correction Manual written, e.g. cable map Analysis Offline extracted, e.g. calibrations and alignment Converted (.slcio) Reconstructed (.slcio) Datamodel: LCIO Data processing: Marlin Conditions handling: LCCD See Steve Aplin's talk at this seminar from Nov. 2008 for more on these... Page 16
LCIO Converter Raw data (.bin) Translation and event building Conditions from DAQ stream, e.g. slow control Converted (.slcio) Each DAQ record has a sibling class implementing EVENT::LCGenericObject Converter implements marlin::datasourceprocessor and consists of one module per record type to do the translation Two more processors write event data into an.slcio file and conditions data into a MySQL database (using LCCD interfaces) Up to here: use LCIO I/O, but CALICE specific data model Page 17
Reconstruction Conditions from DAQ stream, e.g. slow control Converted (.slcio) Calibration and correction Reconstructed (.slcio) Manual written, e.g. cable map Offline extracted, e.g. calibrations and alignment Meant to include all physics-independent corrections (nonlinearity, coherent movement, calibration) Output convention: EVENT::CalorimeterHit format, common cell index encoding, geometry as hit positions All detectors formally independent in processing Page 18
AHCal Calibration SiPM: Geiger-mode diodes with ~1000 pixels Limited #pixels and non-zero dead time: SiPMs show saturation at high illumination Self-calibrating: signal per pixel can be extracted. Correction requires calibration to pixel scale plus response curve Page 19
Standard Candle Standard candle in calorimetry: response to minimum ionizing particle For AHCal, this is a 'tile+sipm' property and determined from muon data 1 Mip ~ 15 pixel, leads to statistical spread Tiles are not light-tight this affects the Mip scale and thus the total energy Correction would be best, but simulation is minimum requirement Page 20
Example: AHCal cable map calib consts setup, alignment Page 21
Non-Calorimeters Beam line instrumentation: triggers, multiplicity counter, 1D tracking chambers, Cherenkov threshold For the moment, all reconstruction based on CALICE-only code, conventions, and assumptions Page 22
Simulation Working with ILD tools: take advantage of MOKKA Every element of setup implemented as MOKKA driver, free parameters included in official MOKKA database, at least one new detector model per test beam setup Frequently changing parameters are 'steerable', e.g. detector positions and rotation angles Perfect world: no mis-alignment, no detector effects Digitization: business of detectors similar to reconstruction Page 23
Example: AHCal Page 24
Example: AHCal Page 25
Detector Understanding Benchmark new technologies with well understood physics EM showers in the AHCal: Detector is linear to good approximation Energy scale is understood Digitization not yet tuned to reproduce all smearing Page 26
Hadron Shower Studies Ready for the physics analyzes Just few examples flashed here where detector understanding and performance allow comparison of MC models Page 27
Separation Studies PFA is all about shower separation, and this can already be studied with test beam data Overlay hits from two events knowledge as in MC with realism of data In agreement with previous MC studies, assumptions used for detector studies are ok Page 28
The future... Data analysis will continue detector technologies are under control, rich spectrum of hadron shower studies are ahead relatively few high-level analysis done, yet. Better cooperation within CALICE and developments with core software people CALICE physics program will continue Still new technologies are under study digital hadron calorimetry, based on PRC, Megas, and GEMs digital ECal based in MAPS ('pixel detector inbetween absorbers') Established technologies move on towards technological prototypes integrated electronics and mechanics, studies towards 'Module 0' Page 29