Discussion on particle identification with the Rich
Hartwig Albrecht
Wed, June 14, 2000
Participants: H. Albrecht, I. Arinyo Ros, D. Dujmic, R. Eckmann,
W. Hulsbergen, P.Krizan, M. Schmelling
There exist two competing principal approaches to perform particle
identification with the Rich:
(a) rely on the track parameters determined in other detector
components (Main Tracker). This gives the centre of
a Cerenkov ring and allows, e.g. by an iterative method,
to determine the Cerenkov angle.
(b) Perform a pattern recognition in the Rich without information
from other detector components. The result are the angles
tx and ty of tracks and their Cerenkov angle.
For the time being, the accuracy of determining tx and ty in the
Main Tracker is worse than determining it in the Rich. The
covariance matrices are not really understood. It is
unclear how far the resolution in the Main tracker can be
improved in future.
Method (a) suffers from the poor resolution in the Main
Tracker. The efficiency of method (b) depends stronger
on track momentum and Rich occupancy than method (a). On the other
hand, tracks which have not been found in the Main Tracker
can be found with method (b), using input from some other detector
component (e.g. SVD).
Denis proposed the following procedure (method c):
* find rings in the Rich and, based thereupon, tracks in other
tracking devices (= method b);
* find more tracks in the Main Tracker and try to associate
Cerenkov rings to them, now using the track parameters as
seed for the ring search.
Numbers for efficiency and selectivity were presented for
all methods but they turned out to be difficult to compare,
causing a controversary discussion. Most (all?) participants
agreed that a good comparison of the different methods
can be based on the analysis of mass peaks in Kshort -> pi+ pi-,
Lambda -> p pi-, phi -> K+ K- (?), converted photons,
K* -> K+ pi-. A sample of events with K0 decays exists.
Conclusions:
============
Decision on the method:
Provide plots on efficiency and selectivity for all pid methods,
based on the K0 -> pi+ pi- mass peak. The decision has to be
made AS SOON AS POSSIBLE.
Demands on the Main Tracker:
* try to understand the chi2 and covariance matrix behaviour
before and after refit in the Main Tracker (W.H. and
S. Spiridonov)
* improvements of the Main Tracker software and hardware are
essential for the Rich analysis:
- spatial resolution
- efficiency
- removal of bad hits (outliers)
- reducing the number of ghosts (e.g. a clean-up procedure
after the Main Tracker reconstruction)
The Rich is a veritable tracking device:
* use the Rich rings (+ SVD or else) as seed for tracking in the
Main Tracker. This certainly will improve the efficiency.
The progam exists. D.D. will take care of.
* use the high precision of the determination of tx and ty
in the Rich and store the parameters from combined Main Tracker
and Rich fit into the track table RTRA. CAVEAT: this procedure
improves the resolution. In case of wrong assignments of rings
to tracks, however, it may reduce the efficiency and introduce
ghosts. This behaviour has to be tested, using K0 -> pi+ pi-
or similar. R.E. will investigate this.
Hartwig Albrecht
Email: albr...@ifh.de
Mail: DESY-F15
Notkestr. 85
D 22603 Hamburg
Tel: +49 40 8998 3188
Fax: +49 40 8998 4018