hcal%pre)cd)1% november2 ,2017 · conclusionsfrom pythia%analyses% •...

47
HCAL PreCD1 Conceptual Design Review November 2 nd , 2017

Upload: others

Post on 11-Jun-2020

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

HCAL  Pre-­‐CD-­‐1  Conceptual  Design  

Review  November  2nd,  2017  

Page 2: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Conclusions  from  Pythia  Analyses  •  Response  from  different  detector  configura4ons  does  not  depend  on  R  

9/20/17   Rosi  Reed  -­‐  HCal   2  

Songkyo    Lee  

Page 3: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Conclusions  from  Pythia  Analyses  •  Response  from  different  detector  configura4ons  does  not  depend  on  R  or  strongly  on  pT  

9/20/17   Rosi  Reed  -­‐  HCal   3  

(GeV/c)T

GenJet p20 30 40 50 60

µ /

σ

0.0

0.2

0.4

0.6

0.8

1.0SIMULATIONsPHENIX

|< 0.6η R=0.4 Tower Jets, |tanti-kSteel HCALINSteel (frame) HCALINAl HCALINAl (frame) HCALIN

(GeV/c)T

Truth Jet p20 30 40 50 60

µ

0.4

0.6

0.8

1.0

1.2PYTHIA8 @ 200 GeV

T / truth jet p

Treco jet p

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 200.020.040.060.08

0.10.120.140.160.18 SIMULATIONsPHENIX

(GeV/c) < 25Ttruth15 < Jet p

|< 0.6η R=0.4 Tower Jets, |tanti-kSteel HCALINSteel (frame) HCALINAl HCALINAl (frame) HCALIN

T / truth jet p

Treco jet p

0 0.20.4 0.6 0.8 1 1.21.4 1.6 1.8 200.020.040.060.08

0.10.120.140.160.18

0.20.220.24

(GeV/c) < 60Ttruth35 < Jet p

T / truth jet p

Treco jet p

0 0.20.4 0.60.8 1 1.2 1.41.61.8 200.020.040.060.08

0.10.120.140.160.180.2

0.22

(GeV/c) < 35Ttruth25 < Jet p

PYTHIA8 pp @ 200 GeV

Raghav  Elayavalli    

Page 4: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Conclusions  from  Pythia  Analyses  •  Response  from  different  detector  configura4ons  does  not  depend  on  R  or  strongly  on  pT  

•  Response  slowly  degrades  with  removal  of  interac4on  lengths  with  ac4ve  readout  – There  is  no  cliff  that  we  fall  off  of  –  inclusive  observables  are  rela4vely  insensi4ve  

– Unfolding  will  start  to  become  difficult  – Minimal  fragmenta4on  bias  will  increase  

• Will  have  further  study  

9/20/17   Rosi  Reed  -­‐  HCal   4  

Page 5: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  in  Heavy  Ion  Collisions  •  Large,  fluctua4ng  background  in  heavy-­‐ion  collisions  complicates  jet  analyses  – Combinatorial  jets  – Changes  Jet  Energy  Scale  (JES)  – Changes  Jet  Energy  Resolu4on  (JER)  –  Jet  finding  Efficiency  

•  Solu4ons  involve  – Background  subtrac4on  – Selec4on  of  “hard”  processes  – Unfolding  

9/20/17   Rosi  Reed  -­‐  HCal   5  

Page 6: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  in  Heavy  Ion  Collisions  •  Large,  fluctua4ng  background  in  heavy-­‐ion  collisions  complicates  jet  analyses  – Combinatorial  jets  – Changes  Jet  Energy  Scale  (JES)  – Changes  Jet  Energy  ResoluKon  (JER)  –  Jet  finding  Efficiency  

•  Solu4ons  involve  – Background  subtracKon  – Selec4on  of  “hard”  processes  – Unfolding  

9/20/17   Rosi  Reed  -­‐  HCal   6  

These  are  the  key  aspects  that  pertain  specifically  to  the  HCAL  design  

Page 7: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Background  SubtracKon  

•  sPHENIX  MIE,  subtrac4on  rou4ne  à  arxiv:1203.1353  

•  We  are  con4nuing  to  work  on  this  à  2  features  not  yet  implemented:  – Refinement  of  exclusion  region  defini4on  in  first  itera4on  step  

– Flow  es4ma4on  /  modula4on  

•  It  is  important  to  note  that  the  background  is  determined  for  each  calorimeter  layer  separately!  

7  Rosi  Reed  -­‐  sPHENIX  Collabora4on  Mee4ng  -­‐  June  2017  

Page 8: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

8  Rosi  Reed  -­‐  sPHENIX  Collabora4on  Mee4ng  -­‐  June  2017  

arxiv:1203.1353    

For  now  using  R=0.2,  pTreco  >  25  GeV  jets  before  subtracKon  as  exclusion  regions”  

No  v2  modulaKon  yet  

Page 9: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Background  SubtracKon  •  Algorithm  adjusts  the  addi4on  to  the  JES  by  the  UE,  and  requires:  – Creates  0.1x0.1-­‐towerized  version  of  CEMC  – Crea4on  of  towers  from  all  3  calorimeter  subsystems  

– Default  jet  reconstruc4on  •  Es4mates  UE  contribu4on  to  towers  in  η=0.1  rings  – Separate  for  each  layer  –  Creates  new  UE-­‐subtracted  tower  containers  

•  Jet  reconstruc4on  has  been  modified  to  handle  nega4ve-­‐E  tower  inputs  

9  Rosi  Reed  -­‐  sPHENIX  Collabora4on  Mee4ng  -­‐  June  2017  

Page 10: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Response  •  Background  smears  reconstructed  momentum  –  Increases  JER  –  Depends  on  R  and  pT  

•  Can  decrease  reconstruc4on  efficiency  

truthTp/reco

Tp

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Cou

nts

0

50

100

150

200

250

300

350

400

sPHENIX G4 Sim=0-4 fmbHijing Au+Au

dijet events = 30-35 GeVtruth

TpR=0.4,

truthTp/reco

Tp

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Cou

nts

0

100

200

300

400

500

600

700sPHENIX G4 Sim

dijet eventsp+p = 30-35 GeVtruth

TpR=0.4,

pp  

AA  �  Jet  energy  Scale  (JES)  

�  Increased  by  background  �  Subtrac4on  method          necessary  �  Unfolding  requires  well          behaved  Response  Matrix  

Erin  Bossard  

Erin  Bossard  

10  

Page 11: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

ReconstrucKon  Efficiency  •  Truth  jets  are  associa4on  with  closest  reco  jet  within  ΔR  <  Ran4-­‐kT    –  If  this  fails,  the  jet  is  not  reconstructed  

(GeV)truthTp

20 25 30 35 40 45 50 55 60

Rec

onst

ruct

ion

Effic

ienc

y

0

0.2

0.4

0.6

0.8

1

1.2

sPHENIX G4 Sim dijet eventsp+p

R=0.4R=0.3R=0.2

(GeV)truthTp

20 25 30 35 40 45 50 55 60

Rec

onst

ruct

ion

Effic

ienc

y

0

0.2

0.4

0.6

0.8

1

1.2

sPHENIX G4 Sim=0-4fmbHijing Au+Au

dijet eventsR=0.4R=0.3R=0.2

�  Larger  jets  and  lower  pT  jets  can  be  obscured  by  the  fluctua4ng  background  

pp  

AA  

Erin  Bossard  

Erin  Bossard  

Rosi  Reed  -­‐  sPHENIX  Collabora4on  Mee4ng  -­‐  June  2017  

11  

Page 12: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Energy  Scale  •  For  high  momentum  jets,  the  JES  is  

similar  in  pp  and  AA  •  Defect  in  the  JES  at  low-­‐pT  in  AA  à  

suspect  cause  is  the  crude  exclusion  seed  selec4on  –  Low-­‐pT  (hard  scakering)  jets  are  

included  in  the  es4mate  of  the  background  à  over  subtrac4ng  

(GeV)truthTp

20 25 30 35 40 45 50 55 60

Jet E

nerg

y Sc

ale

00.10.20.30.40.50.60.70.80.9

1

sPHENIX G4 Sim dijet eventsp+p

R=0.4R=0.3R=0.2

(GeV)truthTp

20 25 30 35 40 45 50 55 60

Jet E

nerg

y Sc

ale

00.10.20.30.40.50.60.70.80.9

1

sPHENIX G4 Sim=0-4 fmbHijing Au+Au

dijet eventsR=0.4

R=0.3

R=0.2

pp  

AA  

�  Mean  value  of:          pT,reco  /  pT,truth  �  Determined  by  Gaussian  fit  

�  Uses  default  simula4ons  �  Linear  tower  calibra4on  à  

assumed  sampling  frac4ons  �  no  addi4onal  jet-­‐level  

calibra4on  

Erin  Bossard  

Erin  Bossard  

Rosi  Reed  -­‐  sPHENIX  Collabora4on  Mee4ng  -­‐  June  2017  

12  

Page 13: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Energy  ResoluKon  •  In  HI  collisions  JER  increases  with  increasing  R  

•  Rapidly  increases  for  pT,truth  ~  average  background  

(GeV)truthTp

20 25 30 35 40 45 50 55 60

Jet E

nerg

y R

esol

utio

n

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35sPHENIX G4 Sim

dijet eventsp+pR=0.4R=0.3R=0.2

(GeV)truthTp

20 25 30 35 40 45 50 55 60

Jet E

nerg

y R

esol

utio

n

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35 sPHENIX G4 Sim=0-4 fmbHijing Au+Au

dijet eventsR=0.4R=0.3R=0.2

Erin  Bossard  

Erin  Bossard  

pp  

AA  

�  Width  of  pT,reco  /  pT,truth  distribu4on  �  Determined  by  Gaussian  fit    

Rosi  Reed  -­‐  sPHENIX  Collabora4on  Mee4ng  -­‐  June  2017  

13  

Page 14: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

T / truth jet p

Treco jet p

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 200.010.020.030.040.050.060.070.08 SIMULATIONsPHENIX

(GeV/c) < 25Ttruth15 < Jet p

|< 0.6η R=0.4 Tower Jets, |tanti-k

Steel HCALINAl (frame) HCALIN

T / truth jet p

Treco jet p

0 0.20.4 0.6 0.8 1 1.21.4 1.6 1.8 200.020.040.060.08

0.10.120.14

(GeV/c) < 60Ttruth35 < Jet p

T / truth jet p

Treco jet p

0 0.20.4 0.60.8 1 1.2 1.41.61.8 20

0.02

0.04

0.06

0.08

0.1 (GeV/c) < 35

Ttruth25 < Jet p

PYTHIA8 pp @ 200 GeV

Embedding  into  HIJING    (R  =0.4  jets)  

•  Comparison  of  default  and  Al  no  readout  – Versus  pT,jet  –  Inclusive  

9/20/17   Rosi  Reed  -­‐  HCal   14  

Raghav  Kunnawalkam  Elayavalli  

Page 15: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Embedding  into  HIJING  

•  Inclusive  jet  observables  in  HI  collisions  follow  the  same  trend  as  in  PYTHIA  studies  

•  For  high  pT  jets,  background  has  minimal  effect  

•  Addi4onal  smearing  due  to  UE  added  in  quadrature  with  smearing  from  reduced  IHCAL  could  reduce  analysis  window  

9/20/17   Rosi  Reed  -­‐  HCal   15  

(GeV/c)T

GenJet p20 30 40 50 60

µ /

σ

0.0

0.2

0.4

0.6

0.8

1.0SIMULATIONsPHENIX

|< 0.6η R=0.4 Tower Jets, |tanti-k

Steel HCALINAl (frame) HCALIN

(GeV/c)T

Truth Jet p20 30 40 50 60

µ

0.4

0.6

0.8

1.0

1.2PYTHIA8+HIJING 0-4fm @ 200 GeV

Page 16: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Outlook  to  CDR    •  Current  sPHENIX  design  similarity  to  LHC  experiments  has  allowed  reasonable  knowledge  transfer  

•  There  do  not  seem  to  be  any  gaping  holes  in  the  methods  needed  for  compensa4ng  for  the  background  in  HI  collisions  with  respect  to  the  design  with  any  IHCAL  configura4on  discussed  so  far  

•  However,  a  fluctua4ng  background  with  a  fluctua4ng  detector  response  can  be  complicated.  

9/20/17   Rosi  Reed  -­‐  HCal   16  

Page 17: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Outlook  to  CDR    •  New  performance  plots  with  respect  to  the  development  that  has  taken  place  – Regenerate  plots  of  JES/JER  vs  R  and  pT  with  the  latest  geometry  

•  Incorpora4on  of  JEWEL  (a  jet  quenching  model)  will  allow  us  to  look  at  quenched  versus  non-­‐quenched  jets  – Small  difference  between  quark  and  gluon  jets  indicates  this  will  not  be  a  dominate  effect  

•  Benchmark  background  subtrac4on  methods  versus  different  calorimeter  systems  

9/20/17   Rosi  Reed  -­‐  HCal   17  

Page 18: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Back  Up  

9/20/17   Rosi  Reed  -­‐  HCal   18  

Page 19: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

HCAL  Descoping  –  Jet  perspecKve  •  What  impact  does  each  inner  HCAL  configura4on  have  on  the  JES  and  JER?  – How  much  can  we  correct  in  unfolding?  

•  Do  the  configura4ons  increase  the  bias  with  respect  to  fragmenta4on?  – Quark  vs  Gluon  jets  – High  z  jets  – Dis4nguishing  modifica4on  of  the  fragmenta4on  

•  How  does  this  effect  the  background  removal  techniques  in  HI  collisions?  

•  We  want  to  push  the  kinema4c  reach  to  create  jet  probes  which  overlap  with  LHC  kinema4cs  

9/20/17   Rosi  Reed  -­‐  HCal   19  

Page 20: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Inclusive  Jet  ResoluKon  vs  R  

Different  HCAL  configura4ons  have  a  minimal  dependence  on  R  for  inclusive  jets  

9/20/17   Rosi  Reed  -­‐  HCal   20  

Songkyo  Lee  

Page 21: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

•  d  

9/20/17   Rosi  Reed  -­‐  HCal   21  

Songkyo    Lee  

We  can  study  effects/configura4ons  with  one  or  two  R  choices  

Inclusive  Jet  ResoluKon  vs  R  

Page 22: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

FragmentaKon  

•  Inclusive  jet  performance  shows  minimal  difference  however,  the  impact  on  high-­‐z  or  other  “special”  jets  can  be  substan4al  – Modifica4on  of  the  fragmenta4on  func4on  is  a  key  sPHENIX  measure  

– The  ability  to  dis4nguish  jet  quenching  effects  from  quark  vs  gluon  will  be  important  

•  sPHENIX  is  a  next  genera4on  detector  and  should  be  able  to  make  next  genera4on  jet  measurements  

9/20/17   Rosi  Reed  -­‐  HCal   22  

Page 23: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Highest  z  Charged  Hadron  (π,  K,  p)  

9/20/17   Rosi  Reed  -­‐  HCal   23  

Default  

JES  à  1.29  corr  JER  à  13.0%  

Al  IHCal  

More  punch  through?  

Al  IHCal  NO  READOUT  

100%/√(50)  =  14.1%  

Red        -­‐    all  jets  Black  -­‐    0.2  <  z  <  0.4  Blue      -­‐    z  >  0.6  

JES  à  1.29  corr  JER  à  13.1%  

JES  à  1.36  corr  JER  à  15.1%   Jamie  

Nagle  

Correc4on  of  1.29  indicates  a  Calo  scale  issue  

Page 24: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

emfrac0 0.2 0.4 0.6 0.8 1

reco

10

20

30

40

50

60

70

reco:emfrac {truth>50.0}tp

Entries 13082Mean 0.2838Mean y 41.42Std Dev 0.2153Std Dev y 6.095

JES/EM  fracKon  (SS310)  JES  Correc4on  of  1.29  •  CEMC  energy  scale  set  for  EM  

showers!  For  a  jet  with  a  high  z  π0  •  Ereco  ~  Etruth  As  EM  frac4on  decreases,  Ereco  decreases  •  IHCal  inves4ga4ons  as  a  

func4on  of  leading  z  hadron  à  walking  up  and  down  this  curve  

High  peak  in  CEMC  energy  frac4on  à  “miscalculated”  by  using  the  EM  energy  scale  to  calculate  the  CEMC  tower  energy      •  Suppresses  the  correla4on  with  

reco  jet  energy  

9/20/17   Rosi  Reed  -­‐  HCal   24  

Jamie  Nagle  

John  Lajoie  

Page 25: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Electron  IdenKficaKon  

•  An4protonsà  main  source  of  fake  electrons  at  low  pT  

•  The  worse  case:  low  pT  an4-­‐protons  steel  frame  example  – Hadron  rejec4on  for  90%  electron  ID  –  Just  E/p  cut:  5.11+/-­‐  0.26  –  2D  cut:  5.05  +/-­‐  0.25  

9/20/17   Rosi  Reed  -­‐  HCal   25  

Sasha  Lebedev  

Page 26: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Set  up  –  Jet  Shape  Observables  •  Analysis  code  à  developed  to  study  jets  with  forward  instrumenta4on:  

•  Same  PYTHIA  8  jets  (R  =  0.4,  pT,jet  >  50  GeV)  •  3  Different  reconstructed  jets  collec4ons  –  Primary  Par4cle  Jets  -­‐  No  muons,  neutrinos  –  Track  Jets  -­‐  Tracks  require  ndf>60,  χ2/ndf<1.5,  DCA2D<0.1cm  

–  Calorimeter  Tower  Jets  -­‐  Require  tower  energy  >  100  MeV  •  3  condi4ons  –  SS310,  Al,  and  Al  without  readout  •  “Matched”  jets  require  track  and  tower  jets  match  the  primary  jet  within  ΔR<0.3  

9/20/17   Rosi  Reed  -­‐  HCal   26  

Page 27: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Shape  in  HCAL  

9/20/17   Rosi  Reed  -­‐  HCal   27  

Matched  tower  jets  à  Etower  with  respect  to  the  jet  axis    

Inner  HCAL   Outer  HCAL  

Less  energy  in  the  inner  HCAL  and  more  in  the  outer  HCAL  with  Al  •  Look  at  normalized  distribu4ons  to  see  whether  shape  changes    

John  Lajoie  

Page 28: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Shape  in  HCAL  

9/20/17   Rosi  Reed  -­‐  HCal   28  

Matched  tower  jets  à  Etower  with  respect  to  the  jet  axis    •  Normalize  by  area  and  integral  to  look  at  the  shape  

Small  difference  in  the  shape  in  the  inner  HCAL  between  SS  and  Al  à  A  likle  broader,  at  a  scale  <  R  

Inner  HCAL   Outer  HCAL  

John  Lajoie  

Page 29: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Combined  Jets  •  Combine  the  tower  and  track  jets  to  use  the  best  informa4on  from  both  –  Similar  to  par4cle  flow  à  uses  truth  informa4on  instead  of  a  real  clustering  algorithm  

•  Point  each  track  jet  cons4tuent  into  the  calorimeter,  sum  up  the  energy  it  contributed  to  the  tower  jet  in  each  calorimeter  segment  

•  Combine  the  tower  and  track  jet  informa4on:  –  Expected  E/p  resolu4on  calculated  using  measured  quan44es  –  If  the  ptrack  resolu4on  is  beker  than  Etower  resolu4on  à  add  track  to  jet  

–  If  the  Etower  resolu4on  is  beker,  rescale  ptrack  to  Etower,  add  track  to  jet    

–  Remaining  Etower  à“neutral  energy”  à  add  to  jet  9/20/17   Rosi  Reed  -­‐  HCal   29  

Page 30: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

9/20/17   Rosi  Reed  -­‐  HCal   30  

Tower  JES/JER  –  degrada4on  is  visible  but  small  compared  to  the  SS310  baseline:    

Tower  Jets  

Combined  Jets  

JES   ResoluKon  

SS310   0.744   9.3%  

Al   0.724   9.8%  

Al    (no  IHCAL)  

0.691   10.7%  

<neutral  frac4on>  >~0.33,  due  to  CEMC  energy  scale?    

Jet  Energy  Scale  

John  Lajoie  

Page 31: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

9/20/17   Rosi  Reed  -­‐  HCal   31  

Use  the  “black  hole”  to  add  back  E  within  R=0.4  of  the  combined  jet  axis    

Shape  of  the  low  side  largely  remains  the  same?  •  Low  side  tail  not  due  to  leakage?  •  Imperfec4ons  in  combina4on  algorithm?  

•  Jet  axis  resolu4on?  

AccounKng  for  Leakage  

John  Lajoie  

Default  

Al  IHCal  

Al  IHCal  NO  READOUT  

Page 32: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Flavor/FragmentaKon  Bias  I  

9/20/17   Rosi  Reed  -­‐  HCal   32  

light  quarks  

gluons  

Light  Quarks   Gluons  JES   Resolu4on   JES   Resolu4on  

SS310   0.749   9.4%   0.722   8.0%  

Al   0.727   9.7%   0.702   9.0%  

Al  (no  IHCAL)  

0.694   10.7%   0.671   10.0%  

•  Separate  response  for  light  quark              (u+d)  and  gluon  jets  •  Require  Eparton  >50%  Ejet    

•  JES  is  a  few  %  lower  for  gluon  jets  than  for  quark  jets  à  JER  is  a  bit  beker    

•  The  difference  between  q/g  jets  is  ~  to  the  difference  between  SS310  and  Al.    

Tower  Jets  John  Lajoie  

Page 33: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Embedding  into  HIJING  

•  Chris  has  produced  HIJING  Files  for  embedding:  –  0-­‐4  fm:  /sphenix/data/data02/rescope-­‐2017-­‐09-­‐08/inner_hcal_al/sHijing/fm_0-­‐4  

– Other  b  ranges  (0-­‐12,  5-­‐9,    9-­‐11):  /sphenix/data/data02/rescope-­‐2017-­‐09-­‐08/inner_hcal_al/sHijing/fm_<b  range>/  

•  JER  is  degraded  due  to  fluctua4ng  background  •  JES  correc4on  via  background  subtrac4on  is  necessary  – Atlas  itera4ve  rou4ne  subtracts  layer  by  layer  à  How  do  descoping  op4ons  effect  this?  

9/20/17   Rosi  Reed  -­‐  HCal   33  

Page 34: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Conclusions  •  The  differences  between  SS,  Al  and  Al  without  read-­‐out  – Not  large  for  inclusive  jet  observables  

•  Including  first  look  at  embedded  jets  – Does  not  have  much  R  dependence  – Can  maker  for  fragmenta4on/flavor  analyses  

•  JES  difference  for  q/g  ~  SS310/Al  difference  •  Need  to  to  properly  set  the  CEMC  energy  scale  •  Combined  jets  show  a  proper  PF  algo  might  get  the  JES  close  to  1  – Leakage  is  s4ll  an  issue  

9/20/17   Rosi  Reed  -­‐  HCal   34  

Page 35: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Concerns/Thoughts  •  We  should  combine  85%  EMCal  +  descoped  HCAL  – Could  simply  cut  on  |ηjet|<0.85-­‐R  – We  have  an  increase  in  sta4s4cs  due  to  addi4onal  years,  but  will  lose  on  the  acceptance  

•  Need  clear  communica4on  over  default  and  other  poten4al  configura4ons  – Understanding  Calo/Tracking  even  for  the  default  

•  Person-­‐power  as  always  •  Jet-­‐trees  will  be  produced  for  pp/AA  – Help  with  person-­‐power  – Keep  details  consistent  

9/20/17   Rosi  Reed  -­‐  HCal   35  

Page 36: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Back-­‐up  

9/20/17   Rosi  Reed  -­‐  HCal   36  

Page 37: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Some  thoughts  

•  Simula4ons  have  become  increasingly  realis4c  – Great  progress,  and  indica4ons  are  that  simpler  simula4ons  were  reasonable  

•  However,  addi4onal  development  both  in  the  simula4ons  and  our    understanding  is  vital  – Even  without  descoping  – We  also  do  not  know  how  more      realis4c  simula4ons  will  affect    the  response  9/20/17   Rosi  Reed  -­‐  HCal   37  

Page 38: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

JES/EM  fracKon  (SS310)  Reco  jet  energy  vs.  reco  jet  EM  fracKon  

 

9/20/17   Rosi  Reed  -­‐  HCal   38  

The  CEMC  energy  frac4on  is  peaked  highà  “miscalculated”  by  using  the  EM  energy    scale  to  calculate  the  CEMC  tower  energy.      •  Suppresses  the  correla4on  with  reco  jet  energy.      

John  Lajoie  

Page 39: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Response  vs  z  (π+,π-­‐)  

•       

•  Mean  +/-­‐  RMS  •  Shiz  in  response  visible  here,  some  z  dependence  

•  But  this  not  the  full  story  

9/20/17   Rosi  Reed  -­‐  HCal   39  

z =pTruthT ,particle

pTruthT , jet

Page 40: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Highest  z  Charged  Hadron  (π,  K,  p)  R  =  0.4  

9/20/17   Rosi  Reed  -­‐  HCal   40  Inclusive  jets,  jets  with  0.2  <  z  <  0.4,  and  jets  with  z  >  0.6.  

Jamie  Nagle  

Page 41: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Combined  Jets  •  Combine  the  tower  and  track  jets  to  use  the  best  informa4on  from  both:  

–  Analysis  code  includes  bookeeping  of  par4cle  energy  contribu4on  to  calorimeter  tower  energy  •  This  uses  truth  informa4on  (think  of  it  as  perfect  clustering/cluster  spli{ng)  

–  Point  each  track  jet  cons4tuent  into  the  calorimeter,  sum  up  the  energy  it  contributed  to  the  tower  jet  in  each  calorimeter  segment  

–  Combine  the  tower  and  track  jet  informa4on:  •  Expected  energy/momentum  resolu4on  calculated  using  measured  quan44es:  

–  Track  momentum  resolu4on  δpT  /pT  =  0.005  +  (0.001*pT)  –  Sort  EM/hadronic  par4cles  by  CEMC  tower  energy  frac4on  >  0.9  –  CEMC  resolu4on    0.12/sqrt(E)  for  EM  par4cles,      

CEMC+HCAL  resolu4on  0.15  +  0.7/sqrt(E)  for  hadrons    •  If  the  track  momentum  resolu4on  is  beker  than  the  tower  resolu4on,  add  the  track  to  

the  combined  jet  •  If  the  calorimeter  energy  resolu4on  is  beker,  rescale  the  track  total  momentum  to  match  

the  tower  energy.    –  Improves  energy  resolu4on  but  keeps  improved  poin4ng  resolu4on  of  tracking  (eta,phi)  

•  Remaining  tower  energy  azer  all  par4cles    is  “neutral  energy”  and  is  added  to  the  combined  jet.    

–  This  is  similar  to  a  par4cle  flow  algorithm,  with  the  excep4on  that  it  uses  truth  informa4on  instead  of  a  real  clustering/cluster  spli{ng  algorithm  

–  Gives  an  idea  of  what  is  the  best  you  could  possibly  do,  or  how  much  informa4on  is  in  principle  available  for  you  to  take  advantage  of  •  Leakage  out  the  back  is  a  loss  of  informa4on.    

9/20/17   Rosi  Reed  -­‐  HCal   41  

Page 42: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Flavor/FragmentaKon  Bias  II  

9/20/17   Rosi  Reed  -­‐  HCal   42  

light  quarks  

gluons  

Combined  jets  show  a  slightly  different  response  for  quarks  and  gluons  (missing  soz  hadron  tracks?)    

Combined  Jets  

Page 43: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Comparison  with  “old”  results  

•  JES  has  shized  downwards  with  new  default  – Old:  0.83+/-­‐0.09  – New:  0.78  +/-­‐  0.10  

9/20/17   Rosi  Reed  -­‐  HCal   43  

Page 44: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Response  •  For  high  pT  jets,  it  seem  the  default  and  the  Al  are  not  so  different  – On  average  

•  Looks  similar  to  Ragav’s  trend  

•  Compare:  – Default:  0.78  +/-­‐  0.10  

– Al:  0.78+/-­‐0.10  – Al  no  readout:  7.4  +/-­‐0.11  

9/20/17   Rosi  Reed  -­‐  HCal   44  

Page 45: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Jet  Shape  II  

9/20/17   Rosi  Reed  -­‐  HCal   45  

For  matched  tower  jets,  look  at  the  distribu4on  of  tower  energies  with  respect  to  the  jet  axis.  Normalize  the  Al  distribu4on  to  the  SS310  integral  to  look  at  the  shape.      

Inner  HCAL   Outer  HCAL  

There  is  a  small  difference  in  shape  for  the  inner  HCAL,  energy  distribu4on  pushed  out  a  likle,  but  at  a  scale  that  is  smaller  than  the  jet  reconstruc4on  radius.  

Page 46: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

Larger  Jet  Radius  

9/20/17   Rosi  Reed  -­‐  HCal   46  

Combined  jets,  re-­‐run  with  R=0.8    for  comparison  –  likle  change.    

Combined   Combined  +  BH  

With  the  R=0.8  the  BH  seems  to  pick  up  a  likle  more  energy.    

Page 47: HCAL%Pre)CD)1% November2 ,2017 · Conclusionsfrom Pythia%Analyses% • Response’from’differentdetector’ configuraons’does’notdepend’on’Ror’ strongly’on’p

η/φ  ResoluKon  

9/20/17   Rosi  Reed  -­‐  HCal   47  

Nothing  special  here,  this  just  shows  the  improved  η/φ  resolu4on  of  the  combined  just  by  using  the  tracking  direc4on  and  the  calorimeter  energy.