Long lived charged particles are allowed by many models of physics beyond the standard model (SM). In general, different models have different characteristics. However, long lived charged particles have a common characteristic in many models; when produced at the LHC some of them would travel with velocity significantly lower than the speed of light. Although the momentum (and therefore β) spectrum of those particles is model dependent, the common signal is a heavy long-lived charged particle with velocity smaller than the speed of light, β<1. This unique signature makes the search for it model independent.

The new particles which have β values close to unity are indistinguishable from ordinary muons. The particles with β significantly lower than 1 can be identified and their mass measured. Such a search is model independent since the velocity and mass of the new particle are measured directly, without reference to other features of the model. We will present first results of this search, based on 45pb-1 integrated luminosity collected with the ATLAS detector.

Our search uses custom trigger and reconstruction methods which we developed in order to avoid the loss of slowly moving particles. These methods are now part of the standard ATLAS software.  
 
Our goal is to find a long lived charged particle, if it exists. This requires an accurate measurement of the signal mass, accurate modeling of the background mass distribution and optimizing the signal to background ratio. The errors of the above should be estimated and included in any statistical test we perform on the data.

The accuracy of the β and mass measurement depends on the calibration of the ATLAS detector. This is an ongoing effort of the ATLAS collaboration. We perform an ad-hoc calibration to improve the β resolution. The data resolution is reproduced in the monte carlo (MC) using a similar method.

The background for this search is dominated by high pT muons with miss-measured β. Its mass distribution and quantity are estimated using a method we developed for this purpose. The method uses only the data and does not depend on MC. It relies on the premises that the muon β distribution is independent of the muon source and kinematics, and that the signal to background ratio before mass measurement is small.

We consider other event characteristics to reject background and improve the signal to background ratio.  Such considerations may be model dependent and the results should be interpreted accordingly.

In absence of discovery, the interpretation of our results would be a limit on the mass and cross-section of a long lived charged particles. Due to the dependence of the identification efficiency on parameters such as the expected β spectrum, production mechanism, decay chains, hadronization and interaction models, any limits would be model dependent. 

*Recipient of the 2010 IPS Fraenkel Prize for a Graduate Student in Particle Physics, Nuclear Physics, and Astrophysics