Vladimir Litvin, Harvey Newman, Sergei Shevchenko

1
RS-1 graviton diphoton decay study Status and
plans

• Vladimir Litvin, Harvey Newman, Sergei Shevchenko
• Caltech CMS
• Jim Branson, Marco Pieri
• UCSD CMS
• Marie-Claude Lemaire
• SACLAY
• Mikhail Dubinin
• MSU

2
Outline

• Performance comparison of single photons with
  cmsim and OSCAR for the RS-1 graviton study
• Electronics saturation study
• Signal production request
• Background production request
• Plans

3
Status

• Performance comparison of single photons with
  cmsim and OSCAR for the RS-1 graviton study
• Resolutions are slightly worse with OSCAR_3_4_0
  ORCA_8_4_0 than with CMSIM133 ORCA_8_1_3 for
  both single photons and Randall-Sundrum Graviton
• The double peak in eta resolution in endcap is
  understood now - it was a wrong parameter for the
  shower depth in the endcap
• Need to redo the exercice with OSCAR_3_6_5 and
  ORCA_8_7_1 as they are the official versions for
  theTDR
• (From M.-C. Lemaire presentation on egamma group
  meeting Jan, 12 http//cmsdoc.cern.ch/Physics/ega
  mma/transparencies/m132-3.pdf)
• Requested single particle datasets are not ready
  yet

4
Status

• Electronics saturation study
• Simple saturation corrections developed for the
  di-electron channel (CMS Note 2004/024) lead to
  promising results for the di-photon channel
• CTEQ6L LHAPDF vs CTEQ5L was studied
• Simple Calorimetry and Tracker isolation was
  included
• Calorimeter isolation criteria For each SC, the
  energy in a cone of DR 0.5 (excluding the SC)
  should be lt 0.02 ET(SC)
• E(HCAL)/ E(ECAL) lt 0.1
• Tracker isolation criteria For each track
  associated with a Super-Cluster, required that
  the number of tracks of pTgt2.5 GeV in a cone DR
  0.3 (excluding the matched track) should be 0
• (From M.-C. Lemaire presentation on BSM meeting
  Jan, 11 http//agenda.cern.ch/askArchive.php?base
  agendacatega05246ida05246s0t112Ftransparenci
  es2Flemaire050111.pdf)
• Requested single particle datasets are not ready
  yet

5
Status

• Points of interest for RS-1 Graviton

6
Status

• Five different sources of background
• Born (MSEL0, MSUB18)
• Box (MSEL0, MSUB114)
• Brem (MSEL0,MSUB14,29,115)
• QCD (MSEL1)
• Drell Yan(MSEL0,MSUB1, MSTP 433, MSTJ 41 1)
• Signal for different masses and coupling
  constants
• Background will be produced in 5 different mass
  ranges and we are try to keep 50-100 events in
  mass windows in a 3G around each invariant mass
  (G - graviton width).

7
Channel independent datacards

• MSTP 51 7 ! CTEQ 5L
• MSTU 21 1 ! Check on possible errors
  during program execution
• MSTJ 11 3 ! Choice of the fragmentation
  function
• MSTJ 22 2 ! Decay those unstable
  particles
• PARJ 71 10. ! for which ctau 10 mm
• MSTP 2 1 ! which order running alphaS
• MSTP 330 ! K factor use parametrized by
  PARP(33)
• MSTP 811 ! multiple parton interactions
  1 is Pythia default
• MSTP 82 4 ! D1 structure of
  multiple interactions
• PARP 82 1.9 ! D2.1 regularization
  scale of transverse momentum
• PARP 84 0.4 ! D0.2 core radius
• PARP 89 1000. ! D1000 power of the
  energy rescaling term

8
Channel independent datacards

• PARP 83 0.5
• PARP 87 0.7
• PARP 88 0.5
• PMAS 5,1 4.8
• PMAS 6,1 175.0
• PARP 67 1. ! DC04 default
• PARP 85 0.33 ! DC04 default
• PARP 86 0.66 ! DC04 default
• PARP 90 0.16 ! DC04 default
• PARP 91 1. ! DC04 default

9
Signal

• We are asking about 1k of events per point and
  10k of events for the (M3500 GeV, c0.1) and
  (M2000 GeV, c0.01) to study angular
  distributions. While the di-electron graviton
  study does require a lot of statistics to be able
  to distinguish RS-1 graviton from Z, it is not
  the case in our study. This is why we need 1k of
  events per point. Meanwhile, we would like to
  study the angular properties of RS-1 diphoton
  decay and this is why we requesting only two
  points with enhanced statistics.
• 52k of events in total.
• Very fast to produce, but if it will be needed we
  can produce PYTHIA Ntuples by ourselves and
  submit in official production. Datacards are
  ready and can be submitted at any time.

10
Born

• Preselection (MSEL0,MSUB18)
• generate PYTHIA event
• find all particles with KF111,221,331,223,22,11
  with pT gt 20 GeV and h lt 2.5
• make all invariant masses and require at least
  one candidate with mass greater than
  CKIN(1)-50GeV
• Additional datacards
• CKIN(3) 100.
• CKIN(1) and CKIN(2) are according to the table on
  the next slide

11
Born background
12
Box

• Preselection (MSEL0,MSUB114)
• generate PYTHIA event
• find all particles with KF111,221,331,223,22,11
  with pT gt 20 GeV and h lt 2.5
• make all invariant masses and require at least
  one candidate with mass greater than
  CKIN(1)-50GeV
• Additional datacards
• CKIN(3) 100.
• CKIN(1) and CKIN(2) are according to the table on
  the next slide

13
Box background
14
Drell Yan

• Preselection (MSEL0,MSUB1,MSTP 433)
• generate PYTHIA event
• switch off the PYTHIA radiation (MSTJ 41 1)
• using PHOTOS to produce radiation for the Z boson
  
• Additional datacards
• CKIN(1) and CKIN(2) are according to the table on
  the next slide

15
Drell Yan background
16
QCD and brem backgrounds

• This is preliminary study for QCD and brem
  background. We shall improve the preselection
  during next couple of weeks using additional
  track isolation on PYTHIA level. Unfortunately we
  cannot use CKIN(1)-50 GeV mass invariant cut here
  because too many events will be lost. We will put
  constant invariant mass cut 550 GeV for every
  dataset (preliminary, will be updated)
• In this case more than one dataset will deposit
  events into specified mass regions. In a tables
  below you can find a total number of events in a
  mass region from all datasets
• CMS Internal Note will be prepare to describe the
  preselection routine for this case as well as for
  the Higgs diphoton study

17
Brem

• Preselection (MSEL0, MSUB14,29,115)
• generate PYTHIA event
• find all particles with KF111,221,331,223,22,11
  with pT gt 20 GeV and h lt 2.5
• make all invariant masses and require at least
  one candidate with mass greater than 550 GeV for
  EACH dataset
• Additional datacards
• CKIN(3) 100.
• CKIN(1) and CKIN(2) are according to the table on
  the next slide

18
Brem background
19
QCD

• Preselection (MSEL1)
• generate PYTHIA event
• find all particles with KF111,221,331,223,22,11
  with E??? gt 150 GeV, E??? gt 100 GeV and h lt 2.5
• make all invariant masses and require at least
  one candidate with mass greater than 550 GeV for
  EACH dataset
• Additional datacards
• CKIN(3) 100.
• CKIN(1) and CKIN(2) are according to the table on
  the next slide

20
QCD background
21
Plans

• Submit official requests for the full MC
  simulation and reconstruction of all datasets. To
  improve preselection for brem and QCD by means of
  using loose tracker isolation criteria on PYTHIA
  level. CMS Internal Note will be published.
• To analyze single particle datasets which will be
  done (hopefully) soon
• To produce small subsets of our data recently at
  Caltech to check the cuts
• To simulate the tails of background distributions
  with another generator (MadGraph or CompHEP)
• To check and tune FAMOS based on fully simulated
  data and to use it ONLY if it will be suitable to
  produce more background. We need very tiny
  fraction of events and it might be done with full
  simulation easily
• To perform study about possible systematic
  uncertainties
• difference between two models of multiple
  interactions MSTP 82 1 and MSTP 82 4
• to use different LHAPDF structure functions to
  understand more deeply this source of systematic
  uncertainty
• difference between two generators
• To analyze the whole datasets when they will be
  produced