Forward Spectrometer
The Forward Spectrometer provides full particle identification for charged particles in the polar angle range 1° - 11°. This is achieved by tracking the particle through a magnetic field, which gives momentum information, and measuring the time-of-flight. From both information together the mass of the charged particle can be determined.
Fig. 1: Overview track reconstruction. Pink squares indicate clusters in different detectors. The dashed lines represent extrapolated track in front of and behind the Open-Dipole magnet. A first estimation of the track assuming a uniform box shaped magnetic field is shown in red. A more realistic representation of the magnetic field finally results in the blue trajectory. The difference is exaggerated to make it visible.
Track reconstruction
The tracks of charged particles in the forward spectrometer are reconstructed in two halfs, the front tracking between target and Open-Dipole Magnet, and the rear tracking behind the magnet.
Front tracking is done with two scintillating fibre detectors MOMO and SciFi. A straight line through both cluster positions, indicated by purple squares in the figure, gives a first estimate of the front track (indicated by a dashed line). Since MOMO has a slightly lower efficiency it is also possible to use the target center and only a SciFi cluster for the front tracking. By default MOMO is required.
The rear track is reconstructed from driftchamber and ToF clusters. The front track is extrapolated to the ToF-Walls, in the vertical plane is no Lorentz force and a straight line can be assumed. If the extrapolated track can be associated to atleast one cluster in the ToF-Walls the information are combined. From this an estimate of the deflection and the particles momentum can already be made (red line). However this method assumes a box shaped uniform magnetic field and neglects fringe fields. To account for non-uniformities and fringe fields a momentum dependent correction is applied (blue line). The size of this effect is exaggerated in the figure to make it visible.
Additionally all driftchamber clusters in proximity to this line are also included in the track. Two linear fits are then applied. The first in the vertical (y-z)-plane includes MOMO, SciFi, all Y-orientated drift chambers and the ToF clusters, using both the position and error information. The second fit is performed in the horizontal (x-z)-plane and includes only the driftchamber and ToF-clusters. Combined these fits describe the particles trajectory behind the magnet (the second dashed line). Due to non-uniformities and fringe fields the extrapolated lines of front and rear track do not necessarily meet in the center of the magnet.
To get an accurate description of the particles trajectory and momentum the particle is "stepped" through the forward spectrometer starting from the reaction vertex in the target. The particle is assumed to have the initial momentum estimation, in small intervals the Lorentz force and the expected energy loss are determined and the particles momentum is altered accordingly until the particle reaches the ToF-Walls. Finally a minimization technique compares the calculated trajectory to the error weighted cluster positions of the track and minimizes the deviations. This is repeated iterativly, changing the initial momentum, to improve the agreement between the calculated and fitted trajectories. Since the energy loss is different for different particle masses this procedure is performed assuming proton, kaon and pion mass separately.
The velocity β of the particles is determined from the track length and the time the particle took from the target to the respective ToF-Wall. If more than one ToF-Wall was hit a weighted average is used.
Fig. 2: Reconstructed β versus reconstructed momentum. Here selected events are shown to suppress background from non hadronic reaction and enhance the amount of charged kaons.
Fig. 3: Mass calculated from β and momentum reconstructed in the forward spectrometer. Again selected events are shown to suppress background and enhance the kaon contribution. Additionally only positively charged particles are shown.
Performance
When plotting the reconstructed β versus the reconstructed momentum characteristic banana shaped lines appear as can be seen in figure 2. Each line corresponds to a certain particle. At β close to 1 background from e⁺/e⁻ pairs is visible. Below lines corresponding to charged pions, kaons and protons can be seen.
In figure 3 the calculated masses of the particles are shown. Distinct peaks for charged pions, kaons and protons are visible. The separation of the different particle types is excellent over a large momentum range. It varies with momentum between 3σ and 7σ.
Using simulated data a polar angle resolution (in σ) of 0.3° was determined. At the maximum magnetic field strength in the open dipole magnet , the momentum has a constant resolution (in σ) of 3%. The β resolution is approximately 2.4% close to β = 1.