1. Open charm measurements in NA61/SHINE
The SPS Heavy Ion and Neutrino Experiment (NA61/SHINE)  at CERN was designed for studies of the properties of the onset of deconfinement and search for the critical point of strongly interacting matter. These goals are being pursued by investigating p+p, p+A and A+A collisions at different beam momenta from 13 to 158 GeV/c for ions and up to 400 GeV/c for protons.
Heavy flavor quarks are produced at the early stage of the hadron collision and may serve as sensitive probes for properties of the QGP. However, only indirect measurements of open charm production in nucleus-nucleus collisions at the SPS energies exist, but precise data on both open and hidden charm production are required to provide information on the in-medium behavior of quarkonia in a model independent way .
For open charm measurements in nucleus-nucleus collisions at SPS energies, the experiment NA61/SHINE was upgraded with the new Small Acceptance Vertex Detector (SAVD) [3,4,5]. The low yields of mesons and their short lifetime require detectors located close to the primary vertex that are capable of providing high tracking efficiency and high primary and secondary vertex resolution.
Since mesons typically decay tens to hundreds of microns downstream of their production point, one needs to reconstruct the decay vertex with a precision of about 50 m in order to recognize daughter particles that originate from decays.
2. Small Acceptance Vertex Detector SAVD for NA61/SHINE
The requirements of high precision and resolution can only be reached by adding Vertex Detector to the NA61/SHINE set-up. The constructed SAVD is positioned between the target and the VTPC-1 (see Fig. figFigure1). It consists of two arms (called Jura and Saleve) with four planes (stations) of coordinate-sensitive detectors located at 5, 10, 15 and 20 cm distance from the target. High coordinate resolution MIMOSA-26 sensors  based on the CMOS Pixel Sensor technology were chosen as the basic detection element of the SAVD stations. The sensors provide a spatial resolution of 3.5 m, have very low material budget (50 m thickness), and their readout time is 115.2 s.
In order to reduce beam-gas interactions, the SAVD and the target were placed in a helium enclosure. The magnetic field in the SAVD volume is small and inhomogeneous (0.13 - 0.25T) because the detector is located close to the edge of the VTPC-1 magnet.
Figure figFigure2 shows the fully integrated detector in the helium vessel installed in December 2016 in the NA61/SHINE experiment for data taking with Pb+Pb collisions at 150 GeV/c . The main goal of the test was to prove precise tracking in the large track multiplicity environment, to demonstrate the ability of precise primary vertex reconstruction and eventually to reconstruct a signal from the collected data.
3. Data reconstruction
Track and vertex reconstruction
A straight line was chosen as the track model: .
The first step of track reconstruction uses a combinatorial method of track recognition based on checking of combinations of all hits from different stations. If the hits lie on a straight line according to a criterion, the combination is accepted as a reconstructed track.
From the distributions of the residuals of hits from the reconstructed tracks the spatial sensor resolution was determined to have values of = 4.7 m and = 5.0 m.
The angular distribution of the reconstructed tracks in the plane is shown in Fig. figFigure3. The distribution for each arm a has clear three-peak structure for each arm. The narrow inner most peak (small angles) was associated with particles produced far upstream and travelling parallel to the beam for a long distance. The middle peak corresponds to particles produced upstream of the target. The outer peak is created by particles produced on the Pb target; these tracks are used for further analysis.
The primary vertex is defined as the point of the closest convergence of all reconstructed tracks. Thus, the longitudinal coordinate of the primary vertex is found by minimizing the expression:
which describes the sum of the relative distances of all track pairs reconstructed in a single event at the given transverse plane defined by the longitudinal coordinate . The and coordinates are afterwards calculated as the average of and positions of tracks at .
Primary vertex reconstruction was done separately for two samples of tracks with x-slope in the interval 0.02 to 0.04 and with x-slope 0.04. By looking at the longitudinal distribution of the primary vertex for these samples of tracks (the left picture in Fig. figFigure4) one may see that indeed the tracks associated with the most outer peak in Fig. figFigure3 (blue) originate from the target which is located about 49 mm upstream the fist VD station; the primary vertices associated with tracks from the middle peak (red in Fig. figFigure3) are relatively smoothly distributed upstream of the target in the range from -1200 mm to -55 mm. At -190 mm the distribution has a sharp peak which is related to interactions in the aluminized Mylar front window of the SAVD. One can also see that between the window and the target the frequency of interaction drops due to the presence of helium gas in the SAVD vessel.
The distribution of from the interactions in the target is shown by the right panel of Fig. figFigure4. The near rectangular shape of the distribution allows to reconstruct the thickness of the target (= 1 mm) and its location with respect to the SAVD. The primary vertex resolution is = 5 m, = 1.5 m and = 30 m (for more details see ). The difference between and is caused by the presence of a vertical component of the magnetic field in the SAVD volume.
After the primary vertex is found, the next step of track reconstruction searches for 3-hit tracks using the Hough transform method (for more details see ). It uses a parametric description of the track by a set of parameters. Once a parametric description of the track in the coordinate space is given, the hit position measurements are transformed into the track parameter space. Then, in this so-called accumulation space track candidates are obtained as local maxima using the voting procedure: the most popular candidates are accepted as real tracks.
The magnetic field in the Vertex Detector volume allows momentum reconstruction of the SAVD-tracks . The following formula was used to calculate momentum:
where , are angles of the track in the first and last stations respectively. Momentum reconstruction is imprecise due to the fact that the field is weak and inhomogeneous ( = 0.04 T m). Nevertheless, the obtained values are useful for verifying track matching between SAVD-tracks and TPC-tracks.
The track matching algorithm tries all possible combinations of pairs of SAVD with TPC tracks, assigning to each SAVD track the total momentum of the paired TPC track (which is precise). Next, the SAVD tracks and TPC tracks are extrapolated to a common surface. The algorithm then searches for the best match among all combinations based on the difference in the position and direction of the tracks and uses charge and total momentum to verify the match (since the SAVD has small reconstruction power for charge and momentum). About 80% of SAVD tracks are in the acceptance of the TPCs according to GEANT4 simulations. The track matching efficiency is determined by number of matched SAVD tracks divided by number of SAVD tracks in the TPC acceptance, and is on the level of 90%.
4. Reconstruction of signal
SAVD tracks matched to TPC tracks are used to search for the signal. In the current analysis PID information was not used. Each SAVD track is paired with another SAVD track and is assumed to be either a kaon or pion. Thus each pair contributes twice in the combinatorial invariant mass distribution. The combinatorial background is several orders of magnitude higher than the signal due to the low yield of charm particles. In order to reduce the large background and maximize the signal to noise ratio (SNR) of the reconstructed peak, four cuts are applied, which were determined from the simulations [4,5]:
• A cut on the track transverse momentum GeV/c;
• A cut on the track impact parameter m;
• A cut on the longitudinal distance between the decay candidate and the interaction point m;
• A cut on the impact parameter D of the back-extrapolated candidate momentum vector m.
The left panel of Fig. figFigure5 shows the invariant mass distribution of unlike charge daughter candidates with applied cuts. One observes a small peak emerging at 1.8 GeV/ , which could be an indication of production. The invariant mass distribution was fitted (red line) using of a third order polynomial function to describe the background and a Gaussian peak for the signal contribution. The indicated errors are statistical only. From the fit one finds the width of the peak to be 40 15 MeV and the total yield to amount to 55 20 with a SNR of 3.
To test the detector capabilities and validate the peak extraction procedure, it was also attempted to reconstruct the , which is much more abundantly produced. However, one can reconstruct only a small fraction of all with the SAVD since the detector design and the tracking algorithm were optimized for decays which have much shorter decay length. A clear peak corresponding to decays is seen in the right panel of Fig. figFigure5 with a width of 18 5 MeV and a total yield of 130 30 with a SNR of about 4.5.
A small-acceptance version of the NA61/SHINE Vertex Detector (SAVD) was installed in 2016. A modest set of data of Pb+Pb collisions at 150 GeV/c was collected and analyzed. A weak indication of a signal was observed in the decay channel. However, the data set is not finally calibrated, and further optimization of tracking, track matching and analysis algorithms, as well as of cuts is still ongoing. It may be premature to draw a final conclusion from the analysis result, though the result appears promising.
The SAVD will be used in p+Pb and Xe+La data taking in the autumn of 2017. Also, high statistics measurements in Pb-Pb collisions at 150 GeV/c are planned in 2018. Looking forward, an upgraded version the so-called Large Acceptance Vertex Detector (LAVD) with more sensors is being planned; the exact design of this detector is currently under investigation.