1. Status of CEPC HCAL Optimization Study in Simulation LIU Bing On behalf the CEPC Calorimeter working group

2. Outline
The introduction of the CEPC Calorimeter(SDHCAL, AHCAL, dual readout…)
HCAL Optimization based on simulation(Mainly done by Jifeng)
Software versions and samples
Geometry structure
Performance
Particle identification optimization in SDHCAL
BDT applied in Particle ID
Conclusion

3. PFA and Imaging Calorimeter
For future colliders, jet energy resolution will be a determinant factor of understanding high energy physics.
magnet yoke and
muon system
Simulation of W, Z reconstructed masses in hadronic mode.
60%/E
30%/E
To improve on the jet energy resolution
PFA is a promising solution to reduce the
confusion term  high granularity Calorimeters

4. Options: SDHCAL based RPC
the first technological prototype among a family of prototypes of high-granularity calorimeters
Total Size:1.0x1.0x1.4m3
Total Layers: 48
Total Channel(pads):440000
Power consumption:𝟏𝟎𝝁𝑾/𝒄𝒉𝒂𝒏𝒏𝒆𝒍
ASIC HARDROC(64 channel)
three-threshold (Semi-digital)
110fC,5pC,15pC

5. SDHCAL based on RPC GRPC advantages:
homogenous, cost-effective, negligible dead zone
allow high longitudinal and transverse segmentation

6. HCAL optimization

7. Why HCAL optimization matters
By Manqi
Both physics goals and budgets require geometry optimization.
ECAL and HCAL are two essential components for gamma/lepton/jets reconstruction and identification.

8. Software Version and Samples
Software versions,
Simulation: Mokka revised
Reconstruction: Arbor_KD_3.3 plus track-related processors
Digitization : G2CDArbor
Samples,
𝒆𝒆→𝒁𝑯→𝒍𝒍 𝒔 =𝟐𝟓𝟎 𝑮𝒆𝑽
𝒆𝒆→𝒁𝑯→𝒃 𝒔 =𝟐𝟓𝟎 𝑮𝒆𝑽
𝒔 =𝟐𝟓𝟎 𝑮𝒆𝑽

9. Geometry of HCAL Longitudinal profile Transverse profile Barrel Endcap
R min (mm)
R max (mm)
Z min (mm)
Z max (mm)
Barrel
2058
3376
2350
Endcap
226
3280
2650
3948
By default, number of layers is 48, RPC based detector.

10. Jet Energy Deposition in HCAL
𝒆𝒆→𝒁𝑯→𝒃 𝒔 =𝟐𝟓𝟎 𝑮𝒆𝑽, 1E5 Events
Detector class
<=Ecal
=HCAl
<=HCAL
Jet Energy Deposit ratio
60%
20%
92%

11. Jet Energy Deposition in HCAL
Fraction of jet energy deposition
Jet energy distribution
The energy deposition fraction arrive at a plateau~23% for jet energy>80GeV

12. Jet Energy @Higgs measurement
Higgs->gg
40 Layer for HCAL is reasonable

13. Jet Energy @Higgs decay branch ratio
𝒆𝒆→𝒁𝑯→𝒍𝒍 𝒔 =𝟐𝟓𝟎 𝑮𝒆𝑽
40 Layer is a good choice

14. Particle identification in SDHCAL

15. Particle identification using BDT in SDHCAL
BDT 6 var Input:
First layer of the shower(Begin)
Number of tracks in the shower
(TrackMultiplicity)
Ratio of shower layers over total fired layers(NInteractinglayer/Nlayers)
Shower density(Density)
Shower radius(Radius)
Maximum shower position(Length)
Density
Radius
Shower layers / fired layers
Trackmultiplicty
Shower Length
Begin

16. Strong separation power in
Training and Test
MC samples training
TMVA of root, Methods: BDT 6var
Training and Test
Signal: pion events with energy 10,20,30,40,50,60,70 and 80GeV
Background: electron events with energy 10,20,30,40,50,60,70 and 80GeV
Background:  muon events with
energy 10,20,30,40,50,60,70 and 80GeV
Mixed Background
Ntraining : Ntest=1 : 1
Strong separation power in
pi/e and pi/muon

17. Pion eff vs Bkg rejection rate
MC samples training
Good pi/e and pi/muon separation
High pion efficiency exceeding 99% with electron and muon rejection of the same level (>99%)

18. Particle identification using BDT @Beam data validation
SPS 2015
electron 10,20,30,40 and 50 GeV
Pion 10,20,30,40,50,60,70,80GeV
Muon 110 GeV
The beam data also show the performance of pi-e and pi-mu separation are good .

19. Comparison with standard selection energy reconstruction

20. Particle identification using BDT in SDHCAL

21. Conclusion HCAL Optimization one option of CEPC Calorimeter:SDHCAL
Several benchmark Higgs decay channels have been studied under different calorimeter configurations
For Higgs->bb ,Higgs->gg, Higgs->WW analysis, JER and W mass are not sensitive to the number of HCAL layer after 40 Layers
one option of CEPC Calorimeter:SDHCAL
SDHCAL provide a powerful high granularity tool for the PFA
leading to an excellent energy resolution.
PID using BDT is reliable: Good pion efficiency with high electron and muon rejection rate

22. YES, the End, Thanks for your attention

23. Backup

24. Comparison with standard selection
In the low energy, Using BDT save many events.