1. “displaced” secondary vertex
Search for Events with Displaced Vertices (DVE) from exotic phenomena
( see Neal talk URL: )
Some extensions of the Standard Model features heavy long-life unstable
particles (NPDV)(χ)...
Minimal Supergravity with Bilinear R-parity violation
Anomaly Mediated Susy “ “ “
Gauge Mediated Supersimmetry Breaking models
Hidden Valley models (Strassler and Zurek)
Minimal Supergravity with non unified gauginos masses, B and R violation (Carpenter, Kaplan, Rhee)
mχ = 35 − 250 GeV/c**2
Displaced vertex : Lχ = from µm to ∞ => large radial displacement
In general quite many tracks associated to the vertices.
Chosen model for preliminary studies:
MSSM with R-parity (B) violation (hep-ph/ ) b
h0 
bbar -
“displaced” secondary vertex p p 
Primary vertex 
sbar s
O(1) dm

2. First samples of MC data were made with « tons » of home-made fixes.
A great deal of time has been spent to make Gauss Susy-compliant... Hopefully in the next release we will have:
- new interface to Pythia
- new interface to Geant4
- definition of some Susy particles in Geant4
- option files in DecFiles for some Susy models.
MC samples are planned to be generated for different Susy parameters, χ masses and lifetime.

3. Expected number of DVE events for one year is:
20k to 200k following the model's parameters
(essentially the χ mass).
¼ of the events have at least one χ in LHCb.
Our global trigger efficiency is about 85%.
Background is small.
1 LHCb year
Generated mass
Monte-Carlo
* Detailled study of the pathologies of DVEs, has shown the need of a dedicated vertexing algorithm (similar to VV algorithm used by SM-Higgs, Cedric et al.). Status of the algorithm: reliable, but needs optimisation => will be released asap.
But, see next slide...
* The feared Enemy for this analysis:
superposition of heavy flavour decay and multiple interactions.
There’s nothing more similar to a χ than a PV from a second p-p interaction.
Problem under study.
* It is possible to reduce MinB retention to 0 Hz (tested on a sample of one million events) by …
* A HLT2 selection is being written, if needed…
...cutting on the total transverse energy of the event, the transverse displacement, the reconstructed mass and the min number of tracks of the reconstructed DV.

4. Unexpected loss of efficiency discovered while studying my vertex finder, for large decay length of χ
(presented at T-rec on 06/10/2008).
The observed inefficiency seems to come
from the combination of two facts:
1) In principle hits in 3 VELO stations are enough for a track to be reconstructible. More than 3 are needed in the case of large density of hits.
The PatVeloGeneralTracking algorithm used for Velo tracks reconstruction sems unable to deal with a high density of r, phi measurements at large z, which is often the case for DVEs.
2) Downstream tracks are reconstructed with the assumption that they come from
the decay in flight of a strange particle → A Pt cut is applied (Pt wrt the line
of flight of their mother) to better select those events. This corresponds to a max impact parameter at z=0. The consequence is that for large decay lengths of the χ, a lot of daughter tracks are lost.
May be this situation could be partially corrected by tuning the cuts, ...
But we suspect that a dedicated reconstruction setup will be needed to reach the max efficiency.