Ks and Lambda skims
Analysis Algorithm
The skimming routine for Ks and Lambdas in e831/FOCUS is
veeskim.sf.
Vee candidates satisfying the analysis cuts are flagged with non-zero
values of the two arrays goodks(vee#)
and goodlb(vee#) , contained in the
common block
veeskim.inc.
Besides imposing analysis cuts for skimming, the routine performs the
following tasks:
- Arbitration : SSD vees (type 9) are chosen over
MIC vees (type 8) and M1 vees (types 1,4,5); MIC vees
are chosen over M1 vees.
- Cerenkov identification : is performed again using
the vee fitted momenta for the two prongs and the resulting
istatp id's are stored into the array
vistatp(1:2,vee#) .
The routine veeskim.sf is called by
drvanal.sf by setting the flag RECVEE=2 (for vee reconstruction).
Analysis Cuts
First vee candidates are selected based on the following conditions:
- M1 vees (types 1,4,5):
- constrained fit succeded ( qdv(vee#).ne.0 )
- Z of vee vertex between last SSD plane and P0Y
(3 sigma tollerance)
- 3-chamber tracks: at most 2 missing hits in P0, P1, P2
- 5-chamber tracks: at most 3 missing hits in P0, P1, P2
- TT vees (type 4): minimum opening angle between prongs
in the XZ view
- MIC vees (type 8): accept all
- SSD vees (type 9): accept all
- 1-LINK SSD vees (type 10):
- momentum of unlinked ssd track < 10 GeV.
- L/sigma>30.
- RECON vees (types 2,7): reject all
- 1-LINK PWC vees (types 11,14,15): reject all
Candidates which further satisfy the following mass cuts are flagged:
- 0.40 < M(Ks) < 0.60 GeV --> goodks(vee#)=1
- 1.08 < M(Lb) < 1.16 GeV --> goodlb(vee#)=1
In Figure 1 and
Figure 2 we show
the Ks and Lambda samples identified by goodks>=1 and goodlb>=1 (respectively),
for different vee types.
Next Lambda candidates with goodlb(vee#)=1 are tested for the Cerenkov
identification of the fast prong, and in case are promoted to
goodlb(vee#)=2 :
- if pfast < 20 GeV --> require 8 <= istatp <= 14
- if 20 GeV < pfast < 110 GeV --> require istatp == 8,12
- if 110 GeV < pfast --> require 8 <= istatp <= 15
The Lambda sample identified by goodlb>=2 is shown in
Figure 3 .
The normalized mass for Ks and Lambdas of different types is
shown in Figure 7 and
Figure 8 . It can be noticed that the
normalized mass is well centered around 0 for Ks, while it is not for
Lambdas, probabily due to the presence of higher background under the
signal. For this reason, we do not use the normalized Lambda mass as
part of the skimming criteria.
Ks candidates with goodks(vee#)=1 are skimmed out if
they satisfy a normalized mass cut; Lambda candidates with
goodlb(vee#)=2 are skimmed out if they satisfy a tight straight mass cut:
- Ks : abs( (M(Ks)-0.497672)/sigma(M(Ks) ) < cut -->
goodks(vee#)=2 --> keepks=1
(where "cut" depends on the vee type)
- Lb : 1.09 GeV < M(lb) < 1.14 GeV -->
goodlb(vee#)=3 --> keeplb=1
In Figure 4 and
Figure 5 we show
the Ks and Lambda samples identified by goodks>=2 and goodlb>=3 (respectively),
i.e. the Ks/Lambda skimmed samples.
In Figure 6 we show once again the
effect of the normalized mass cut on the Ks sample: the plots in the first
line are the Ks invariant mass distributions for Ks types 1,4,5 and 9
after the cut goodks>=2 has been applied;
the plots in the second line are the cooresponding normalized mass
distributions for the same samples;
finally, the plots in the last line are again the Ks invariant mass
distributions after Cerenkov conditions on the two Ks prongs have been
applied.
Update: Citadl cuts
The previous Cerenkov requirement on the proton of the Lambda
has been modified to include tracks identified as proton by the
Citadl algorithm. Specifically, Lambda candidates satisfying
thev
Summary
- Ks :
- goodks(vee#)=1 : flags a clean Ks
sample without tight mass cuts
- goodks(vee#)=2 : flags the
subsample satisfying the normalized mass cut required
by the skim
- Lb :
- goodlb(vee#)=1 : flags a clean Lb
sample without Cerenkov requirements and
without tight mass cuts
- goodlb(vee#)=2 : flags the Lb
candidates further satisfying the Cerenkov requirement
on the fast prong
- goodlb(vee#)=3 : flags the Lb
subsample satisfying the mass cut required by the skim
Author: Luca Cinquini (
cinquini@pizero.colorado.edu)
Last update: 1 May 1997
