General MUON finding in ZEUS using ORANGE
This page was created by Achim Geiser.
Last update 2/05/07
Preamble
As illustrated in the high mass dimuon event picture above, muons can be
identified in ZEUS from many different subdetectors.
The muon momentum is usually measured by the CTD, potentially supplemented
by information from the FTD/RTD, and, for post-upgrade data, the MVD and STT.
Next, the muon passes the calorimeter, where it leaves a characteristic
Minimum Ionizing Particle (MIP) signature. Finally, if its momentum is
sufficiently large, the muon traverses the muon spectrometer consisting
of inner muon chambers (inner B/RMUON, FMUON for barrel, rear, or forward),
the backing calorimeter (BAC), and the outer muon chambers (outer B/R/FMUON).
In the very forward region, the FMUON system is extended by an additional
spectrometer which allows a standalone muon momentum measurement without the
need for a CTD track.
For a general introduction, see also the lecture on Muon Finding in ZEUS.
1. The general GMUON implementation
The General MUON finding (GMUON) package in ORANGE is based on a combination
of the GLOMU, BREMAT, MAMMA, MPMATCH, MUFO, MIP, MV, and MUBAC muon finders
described elsewhere.
A rudimentary BAC match and FMUON standalone information based on the MFCTS
tables are provided in addition. Further extensions are forseen.
It is available from standard release 2003a.2 onwards. The current version is
2005a.1. For potential new test
implementations see below.
The algorithm finds prompt muons based on an "OR" of all these packages, and
combines the
information into a single entry per muon candidate as much as possible. The
basic format of this entry is independent of the respective finder.
Cross references to more detailed
information for each finder and cross-correlations between the different
finders are provided.
In the case of ambiguities, alternative reconstructions of the same
muon are sometimes listed.
A muon quality flag is provided which should help the non-expert user to
select good quality muon candidates from this list as a starting point for an
analysis. Since muon quality is process specific, for critical applications
it is recommended that the user redefine his/her own quality criteria based
on the information provided.
Feedback on 'features',
bugs, missing information and/or possible improvements
is greatly appreciated.
2. Description of variables
For a brief description of the variables of the GMUON block, see gmuontup.inc.
See also the general ORANGE variable documentation on the ORANGE page.
A more detailed description will be provided.
3. Description of muon-related ORANGE cards
Card default description
DOMUON OFF enables muon finding in general
DOMUON1 ON enable old (pre-2003) simple muon finders
DOMUON2 ON enable new sophisticated muon finders
MUZUFO OFF use muons for ZUFO corrections
MUJETA OFF calculate Muon association to jet type A
NTMUON OFF enable old MUON ntuple block
NTGMUON OFF enable new GMUON ntuple block
NTGMUJET OFF include muon-jet information into GMUON
NTMUMC OFF include muon MC information into GMUON
NTMUQU 2 only store muons with quality flag >= value (not yet implemented)
DOBAC OFF enable muon finding based on BAC (J. Sztuk)
NTBAC OFF store detailed MUBAC information (J. Sztuk)
NTBREMAT OFF store detailed BREMAT information (G. Abbiendi)
NTMPMAT OFF store detailed MPMATCH/MFRTZ/MFCTS information (M. Corradi)
NTMV OFF store detailed MV information (V. Kuzmin)
In addition there are cards for invidual packages and for GEANE initialization
for expert use.
*** For the moment, whenever DOMUON2 is ON, a set of GEANT and BREMAT cards needs to be added. ***
These are needed for the proper functioning of BREMAT, MUFO, and MPMATCH.
Solutions to replace this by a default initialization are under study.
Examples for use of cards:
The minimum setting for non-expert users which do not need backwards
compatibility with pre-2003 versions of the ntuple is
DOMUON ON
NTGMUON ON
DOBAC ON
add the GEANT and BREMAT cards
set the approriate trigger date for the corresponding year
(see GEANT cards)
This will switch on all essential features, and yield a GMUON block
containing all the necessary basic information.
If information on muon isolation is wished the card
ORANGE-NTBLOCKS GMU_ISOL
will make this information available in the GMUON block
If information on muon-jet association is wished, and some jet finder is
enabled the cards
ORANGE-NTBLOCKS GMU_JET
or NTGMUJET ON
will make this information available in the GMUON block, separately for
jet blocks A, B, C, and D.
If simulated true information on muons is wished the card
ORANGE-NTBLOCKS GMU_MC
or NTMUMC ON
will make this information available in the GMUON block
For experts wishing more detailed information, a selection from the cards
NTMUON ON (for GLOMU, MAMMA, MIP)
NTBAC ON
NTBREMAT ON
NTMPMAT ON
NTMV ON
will switch on additional expert blocks from the respective finders
If zufos are enabled, and are to be corrected for the muon mip deposit,
the card
MUZUFO ON
should be enabled.
To enable the zufo dead material correction (NOT to be used in conjunction
with any other jet corrections), use
ORANGE-ZUFODMCO ON
Be careful with the use of ZUFORMVE ON (electron removal) since this removes
all tracks in the vicinity of an electron from the zufo list, even if they
belong to a muon! (best solution: DON'T use this flag).
To switch on all (or a selection of) GMUON and related ntuple blocks, you
can also use the new NTUPLE BLOCK notation
ORANGE-NTBLOCKS GMUON GMU_ISOL GMU_JET GMU_MC MUON BREMAT FMU_MPM FMUmfrtz MV BAC
For full backwards compatibility with the pre-2003 version of the ntuple
(now obsolete!), use
DOMUON ON
DOMUON2 OFF
NTMUON ON
This will switch off all new features, and yield a muon block identical to the
old one.
The setting to profit from the new GMUON block, with simultaneous
access to the "old" information for backwards compatibility, should be
DOMUON ON
NTMUON ON
NTGMUON ON
DOBAC ON
do not forget to add the GEANT and BREMAT cards
4. How to use the GEANT and BREMAT cards
The GEANE extrapolation for the BREMAT, MPMATCH and MUFO finders needs a
special set of cards defining the detector geometry. Unfortunately this
set needs to be different for different years, especially since the recent
changes of the CAL geometry description. These card files must be added
into your standard orange cards file for both data and MC, depending on the
year.
for 1995 data/MC, use GEANT95.cards
for 1996/97 data/MC, use GEANT96.cards
for 1998/00 data/MC, use GEANT00.cards
for 2003/04 pre-reprocessing data/MC, use GEANT.cards
for 2003/04 post-reprocessing data/MC, use GEANT04.cards
for 2005 pre-reprocessing data/MC, use GEANT.cards
for 2005 post-reprocessing (and 06) data/MC, use GEANT05.cards
Up to 2004, the files differ only by the so called trigger target
data (ZGEN-TRGTDATE). Other dates from the same running period can be
used if needed for e.g. a specific trigger date for the BRMUON dead LST
corrections. From 2005 reprocessed data onwards, it is necessary to
disable the beam pipe geometry to avoid an inconsistency in the Mozart
geometry logic (will be fixed in next release). From the 2006a release onwards,
a few of the BREMAT-related card settings are not needed any more, but they do not
harm, so they are kept for backwards compatibility.
An effort will be made to get rid of these card files completely by one of the next
releases. Meanwhile, apologies for the inconvenience.
5. Additional FAQ's
Please ask your questions! The answers will be posted here.
6. Some basic performance numbers and plots
Some plots will be posted here:
Muon quality for prompt and fake muons from an inclusive bbbar sample,
separately for each finder. There are significant overlaps.
Correlations between the efficiencies for each finder
Dimuon invariant mass distribution from a single data run in 2000.
Ntrack (ctd) <=2 and muon quality >=2 and sum of muon qualities >=6.
=> can be used for DQM.
Same distribution from an inclusive J/psi sample.
Muon detection threshold as "measured" from an inclusive J/psi MC sample
from a comparison of different finders.
(mu,nu) transverse mass from inclusive W -> mu nu
(and W -> tau nu, tau -> mu) MC sample. Muon quality >= 5. pt(mu) > 5 GeV.
Missing pt straight from muon corrected zufos.
Expected efficiencies and event samples: Numbers do NOT yet include trigger
and other selection efficiencies.
Beauty (from small inclusive PYTHIA bbbar sample):
Assume total bbbar cross section of ~10 nb.
-> 10^6 events for 100 pb^-1.
~ 200000 events with >= 1 prompt muon p>1 GeV and pT>0.5 GeV
~ 110000 events with reconstructed muon candidate, muon quality >=4
~ 90000 events with reconstructed prompt muon, muon quality >=4
~ 30000 events with rec. prompt muon qual>=4 ass. to jet pt>4 GeV (ptrel)
~ 2400 events with rec. prompt muon 5dof + 2 jets pt>6,7 GeV +p/pt cut
Beauty -> D*+muon: Selection with muon quality >= 4 doubles event sample
(currently from BREMAT 4dof only) while keeping same signal/bg ratio.
W->munu: Either significant increase in efficiency (~30%) for similar purity,
or significant increase in purity for same efficiency (=> loosen other cuts?).
7. Acknowledgements
Most of the work which made this GMUON package possible was invested by the
authors of the individual muon finders.
Many thanks to them for making their
code available. The zufo correction and muon-jet association has been
implemented by Oliver Gutsche based on code provided by Massimo Corradi.
In addition I owe great thanks to the following people who have explicitly
contributed with additional pieces of code, tests, or expert advice:
Lorenzo Bellagamba, Alessandro Bertolin, Ingo Bloch, Graziano Bruni,
Vincenzo Chiochia, Massimo Corradi, Oliver Gutsche, Valentin Kuzmin,
Rainer Mankel, Bruce Straub.