1: \chapter{Event reconstruction and particle identification}
2:
3: %
4: \section{Event reconstruction}
5:
6: In this chapter we discuss the event reconstruction and the extraction of raw particle yield via the dE/dx method.
7: The main detector for the analysis is the Time Projection Chamber (TPC) of STAR.
8:
9: \subsection{Hit and cluster finding}
10:
11: Each TPC sector has 45 pad rows as shown in Fig.~\ref{fig:tpcendevent}.
12: Hence a charged particle traveling through the TPC can leave 45 possible hits.
13: Reading out the pixel information of the padrows allows the location of the hit to be determined.
14: The hit finder algorithm reads in the time ordered information from each padrow pixel, that are above a trigger threshold and marked as good channels by the DAQ (Data Acquisition System). Gain correction and timing information are also included in the hit finding. In the next step the cluster finder identifies hits in a 2D cluster in the plane of the padrow and the longitudinal direction. In the cluster a single hit as a centroid of the 2D distribution or multiple hits as local maximum in the deconvoluted distribution can be determined. Deconvolution of the hits is very important in particle identification and multiplicity measurements. The deconvoluted hits are converted to real space points in the local coordinate system of the TPC, including the calibration parameters such as drift velocity, trigger timing and geometry. Beside the space information the deposited energy is also stored for each hit.
15:
16: \subsection{Track finding}
17: The track finder, starting from the outer padrows, assigns hits to a track candidate. In the first iteration many track segments (tracklets) are created as candidates for a track. Then, these track segments are fitted by the algorithm that keeps or rejects the hits, depending on their position with respect to the fitted track. In this stage of fitting the effect of multiple Coulomb scattering and energy loss assuming pion mass are taken into account. At the end of the algorithm, the collection of the tracks is produced with their space coordinates and their 3-momenta.
18:
19: \subsection{Global and primary tracks}
20: As the final step in the event reconstruction the $global$ and $primary$ tracks are created. As mentioned in the previous subsection, tracks are reconstructed in the local coordinate system of the TPC. To perform data analysis, global information of the tracks is needed. The global track finder first re-fits the tracks in the TPC, based on a 3D helix model. After the re-fit with the knowledge of the alignment of the different subsystems, the global track finder reconstructs the global tracks from the 'local' tracks in the subdetectors. Based on the information of the global tracks in the event the primary vertex can be found. Those global tracks with a distance to the primary vertex smaller than 3 cm (distance of closest approach, hereafter: $dca$), are re-fitted including the primary vertex as an additional point in the fit. In the analysis these primary tracks are used. Primary tracks are largely the particles produced in the primary interaction. Global tracks
21: include a large number of particles from background or pile-up processes.
22:
23:
24:
25: %
26: %
27: \begin{figure}[!h]
28: %
29: \begin{center}
30: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/pid/dedx_withFits.eps}}
31: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/pid/zK.eps}}
32: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/pid/zpbar.eps}}
33: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/pid/zpi.eps}}
34: \caption{Top left figure shows the specific ionization energy loss as a function of transverse momentum measured in the TPC. The white dots represent the fitted $\pi^{-}$, $e^{-}$, $K^{-}$ and $\overline{p}$ centroids. Top right figure shows the kaon mass parameterized $z$ distribution. Left bottom plots shows the proton mass and right bottom plots shows the pion mass parameterized $z$ distributions.\label{fig:dedxandz}}
35: \end{center}
36: \end{figure}
37: %
38: %
39: %
40: %
41: %
42: %
43: \section{Particle identification}
44: As the main tracking detector of the STAR, the TPC can identify particles by measuring the mass dependent specific ionization energy loss (dE/dx) at low transverse momentum ($p_{T} < ~$1.5 GeV/c).
45: A charged particle traversing the TPC gas volume ionizes the gas atoms. In the electric field these charge clouds drift from their creation point to the two ends of the TPC, where the charges are read out on the padrows. Produced charge in each hit on a padrow is proportional to the energy loss of the particle traversing through the TPC volume. If a particle travels through the entire TPC volume, 45 dE/dx points can be measured on the 45 padrows.
46:
47: Energy loss of a charged particle for a given track length can be described with a Landau probability distribution.
48: However, the mean of the distribution is sensitive to the fluctuations in the tail of the distribution. Therefore, the highest 30$\%$ of the measured charge clusters is discarded for each track. The truncated mean is calculated from the remaining 70$\%$ and defines the average ionization energy loss used in the data analysis.
49:
50: Specific ionization in an isotropic homogeneous medium can be parameterized by the Bethe-Bloch formula:
51: \begin{equation}
52: -\frac{dE}{dx}=4\pi N_{0}r_{e}^{2}m_{e}c^{2}\frac{Z}A\rho \frac{1}{\beta^{2}}z^{2} \left[ ln \left( \frac{2m_{e}c^{2}}I\beta^{2}\gamma^{2}\right) - \beta^{2} - \frac{\delta}2\right]
53: \end{equation}
54: where $N_{0}$ is the Avogadro number, $r_{e}$ is the classical electron radius, $Z$ is the atomic number of the medium, $\rho$ is the density of the medium, $z$ is the charge of the particle traveling through the medium, $I$ is the ionization potential of the medium, and $\delta$ accounts for the density effect of the medium.
55:
56: %
57:
58: As shown in Fig.~\ref{fig:dedxandz} the dE/dx bands for particles with different mass can be separated.
59: Figure~\ref{fig:dedxandz} also shows the kinematic range for particle identification: antiprotons can be identified in 0.3 - 1.2 GeV/c, kaons can be measured in 0.2 - 0.7 GeV/c and the pions can be measured in 0.2 - 0.7 GeV/c. To extract the raw yield of the particles one can introduce the so called $z$ variable:
60: \begin{equation}
61: z = \log\left(\frac{dE/dx_{measured}}{dE/dx_{parameterized}}\right)
62: \end{equation}
63: where $(\frac{dE}{dx})_{measured}$ is the measured energy loss and $(\frac{dE}{dx})_{parameterized}$ is the parameterized form of the energy loss~\cite{Aguilar-Benitez:1991yy}. The $z$ variable has the advantage that each particle has a Gaussian distribution around the expected Bethe-Bloch value. In our analysis a single parameter approximation of the Bethe-Bloch formula is used:
64: \begin{equation}
65: dE/dx_{measured}\ =\ A\cdot\left(1+\frac{m^{2}}{p^{2}}\right)
66: \end{equation}
67: where $A$ is a constant, $m$ is the mass of a given particle and $p$ is the total momentum of the particle.
68: Figure~\ref{fig:dedxandz} (top left panel) shows the measured energy loss as a function of transverse momentum in minimum bias pp collisions at 200 GeV. The white dots represent the centroid positions of $\pi^{-}$, $e^{-}$, $K^{-}$ and $\overline{p}$ from the multi Gaussian fits to the energy loss corrected $z_{K^{-}}$ distribution. Figure~\ref{fig:dedxandz} also shows the $z$ distributions of $K^{-}$, $\pi^{-}$ and $\overline{p}$. The energy loss band of the particle of interest is centered around zero and the other bands are well separated.
69:
70:
71:
72:
73: \section{Extraction of raw particle yields}\label{4Gauss}
74:
75: As noted above the specific energy loss is parameterized with the Bethe-Bloch formula for each particle to calculate the $z$ variable which is centered around 0. The parameterization is slightly dependent on the multiplicity/centrality and adjusted for each data production due to calibration.
76: To extract the raw particle yields, the $z$ distributions/peaks are simultaneously fitted by multiple Gaussian functions, as demonstrated in Fig.~\ref{fig:ppGaussianFitsKaonSample}.
77:
78: \begin{figure}[!h]
79: \begin{center}
80: \resizebox{.7\textwidth}{!}{\includegraphics{Plots/pid/SpFits/km_Nch_1_100_pT_0.70_0.75.eps}}
81: \caption{Multi-Gaussian fit to the $z_{K^{-}}$ distribution in 200 GeV pp collisions.\label{fig:ppGaussianFitsKaonSample}}
82: \end{center}
83: \end{figure}
84:
85: Charged pions can be separated in the transverse momentum region 200 - 800 MeV/c. Figure~\ref{fig:ppGaussianFitsPion} shows the multi Gaussian fits to the $z_{\pi^{-}}$ distribution in these momentum slices. Black dots represent the measured $z_{\pi^{-}}$ distribution and the colored lines represent the Gaussian fits to $\pi^{-}$, $e^{-}$, $K^{-}$ and the combined fit.
86: %
87: The raw kaon yield can be extracted in the transverse momentum region $\sim$200 - 800 MeV/c. The extraction of the raw yield is more complicated since the electron and kaon peaks start to merge at $\sim$ 450 MeV/c. Therefore, the raw electron yield is extracted in the $p_{T} <$ 450 MeV region and extrapolated in the merged bins to obtain the raw kaon yields. The measured raw electron yield is fitted to an exponential function (inspired by MC studies) in the momentum range 200 - 450 MeV/c. The fit result is fed to the multi-Gaussian fit in the large $p_{T}$ region and the electron yield is either fixed or left to vary within a reasonable range around the fitted value.
88: %
89: %
90: The fit results are shown in Figure~\ref{fig:ppGaussianFitsKaon}.
91:
92: As shown in Fig.~\ref{fig:dedxandz}, electrons are also merged into the $\pi$ band, although to a less degree than the kaon band.
93: In the merged region the same procedure is applied as for pions. The raw electron yield is estimated from an exponential fit over the momentum range where electrons are well separated.
94:
95: Protons/antiprotons are well separated from the rest of the particles in the momentum region $\sim$ 300 - 1200 MeV/c, as shown in Fig.~\ref{fig:dedxandz}, and can be fitted to a single Gaussian up to $\sim$ 850 MeV/c, as shown in Fig.~\ref{fig:ppGaussianFitsPbar}.
96: In the following bins the electron contamination is estimated the procedure is the same as for kaons.
97:
98: Figure~\ref{fig:ppGaussianFitsPion}, Fig.~\ref{fig:ppGaussianFitsKaon} and Fig.~\ref{fig:ppGaussianFitsPbar} show the typical fits to the negative identified particles in 200 GeV pp collisions. Fits to positive particles and to 200 GeV dAu and 62.4 GeV Au-Au data are similar.
99: The momentum range can vary due to the changing resolution of the dE/dx bands in the different datasets.
100:
101: %
102: