Primary vertex in the Alice experiment in absence of the event reconstruction

Full text: The vertex detector (Inner Tracking System) of the Alice experiment was designed for performing several tasks: the determination of the primary and secondary vertices, the improvement of the resolution of the Time Projection Chamber and the measurement of the low momenta (below 100 MeV/c). The ITS consists in six co-axial barrels: the two most internal use Silicon Pixel Detectors (SPD), the two layers in the middle use Silicon Drift Detectors (SDD) and the two most external use Silicon Strip Detectors (SSD). The very large granularity of the SPD gives also the opportunity to determine the primary vertex position even in the absence of tracking and, therefore, without event reconstruction. In principle, in fact, it is sufficient to look at the correlation between the hits in the two pixel layers. Several tasks of the reconstruction programs can take advantage from the knowledge of the primary vertex position. In particular it is required for the reconstruction of the dimuon events, it can be used as a constraint in the tracking algorithm and it is essential for a good efficiency in determining the position of the secondary vertices. The absence of pattern recognition, however, causes uncertainty in the vertex determination. A first kind of uncertainty comes from the impossibility to distinguish the primary particles from the products of the secondary decay. Luckily, it can be shown that the contribution of the secondary tracks to the total distribution is statistically negligible (less than 10%). A second source of uncertainty is the impossibility to establish if two hits are generated by the same track. This forces, in the beginning, to consider all the possible combinations between the hits of the first and second pixel layer. The associations of hits belonging to different tracks produces, in the analysis, a combinatorial background that strongly dominates the signal from hits of the same track. Consequently it is necessary to reduce this combinatorial background as much as possible, without impoverishing the signal statistics because the resolution on the vertex position depends on the multiplicity n as 1/√ n. This was performed by using two criteria; the selection of the hits along the beam axis (z-coordinate) and the selection on their relative angle in the transverse plane. The results obtained in the vertex determination show a deviation from the true position of some tens of microns in the transversal plane and lower than ten microns along the beam axis. Practically the results do not vary with changing the magnetic field from 0.2 to 0.4 Tesla and the distance of the ion-ion collision point from the ITS central axis. (Author)