1: \chapter{Analysis method of identified charged particle spectra}
2:
3: In this section the analysis technique of identified charged particle spectra measurements of $\pi^{\pm}$, $K^{\pm}$, $\overline{p}$ and p are reported for 200 GeV pp, 200 GeV dAu and 62.4 GeV Au-Au collisions.
4:
5: \section{General procedure of data analysis}
6:
7: Before the detailed discussion, a general overview is given to provide a conceptual framework for the data analysis. Our goal is to extract the corrected particle spectra and their properties for identified pions, kaons and protons/antiprotons. Steps of the analysis leading to the fully corrected identified particle spectra are listed below:
8: \begin{enumerate}
9: \item {Good events are selected from data on tape, satisfying trigger and vertex requirements. Event wise variables such as the uncorrected charged particle multiplicity are corrected for vertex inefficiencies upon selecting the good events in pp and in minimum bias and peripheral dAu collisions.}
10: \item{Once a good event is identified, good tracks are selected based on the analysis specific quality cuts.
11: In the case of kaon or proton/antiproton tracks, each track is corrected for energy loss upon selection.}
12:
13: \item{At this point selected data includes event and track corrections, which is followed by the extraction of raw yield from the multi-Gaussian fits described in Sec.~\ref{4Gauss}.}
14: \item{The extracted raw yield is corrected for tracking efficiency and acceptance depending on particle type, multiplicity and/or centrality.}
15: \begin{itemize}
16: \item{Raw pion yield is further corrected for weak decay and detector background contamination.}
17: \item{Raw proton yield is corrected for background contribution from detector material.}
18: \item{In the case of minimum bias pp and dAu and peripheral dAu collisions, a fake vertex correction is applied for all particle types.}
19: \end{itemize}
20: \item{Finally point-to-point systematic errors are assigned to each spectrum point.}
21: \end{enumerate}
22:
23: At the end of this procedure the fully corrected identified particle spectra are obtained and one can proceed to extract the bulk properties of the collisions which will be discussed in the Result section.
24:
25:
26: \section{Data sets and trigger}
27:
28: Data presented here are collected in three different RHIC runs: pp collisions at $\sqrt{s_{NN}}$ = 200 GeV in 2002, dAu collisions at $\sqrt{s_{NN}}$ = 200 GeV in 2003, and Au-Au collisions at $\sqrt{s_{NN}}$ = 62.4 GeV in 2004.
29:
30: Various combinations of the trigger detectors (BBC - CTB, ZDC - CTB) are utilized to measure charged particle and neutral particle multiplicity.
31: In pp collisions the minimum bias events are selected by the coincidence of the two BBCs measuring charged particle multiplicity near beam rapidity. In dAu collisions the minimum bias events are obtained from the combination of BBC and ZDC coincidence. In Au-Au collisions the minimum bias events are selected from the CTB-ZDC charged-neutral multiplicity correlation. In each run the magnetic field strength is set at 0.5 Tesla.
32:
33: \section{Event selection}
34:
35: The position of the collision vertices are distributed around the center of the detector. To select events with approximately uniform detector acceptance in pseudorapidity, the primary vertex position has to be limited. Selection on the $z$ component of the primary vertex is specific to the colliding species. Additionally, events have to satisfy the following requirements: $\left|v_{x}\right|< $ 3.5 cm and $\left|v_{y}\right|< $ 3.5 cm.
36:
37: To experimentally vary the impact parameter/centrality of the collision, cuts on the uncorrected reference multiplicity (or charged particle multiplicity) are applied. The uncorrected reference multiplicity is defined as the number of charged primary tracks in pseudo-rapidity ($\eta)$: -0.5 $<$ $\eta$ $<$ 0.5.
38:
39: The specific $z$ vertex and multiplicity/centrality selection is presented below.
40:
41: \subsection{Proton - Proton collisions}
42:
43: The $z$ component of the primary vertex in each minimum bias event has to satisfy the following condition: $\left|v_{z}\right|< $ 30.0 cm. With this vertex cut and minimum bias trigger 3.9 M good minimum bias events are selected. Figure~\ref{fig:pp_mult} shows the uncorrected reference multiplicity distribution in minimum bias pp collisions. To gain more insight, we will investigate the bulk properties not only in minimum bias pp collisions but also as a function of charged particle multiplicity. In pp collisions five multiplicity classes are chosen as summarized in Table~\ref{tab:auau_coll_prop}.
44: %
45: %
46: %
47: \begin{figure}[!h]
48: \begin{center}
49: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/analysis/pp_multiplicity3.eps}}
50: \caption{Uncorrected reference multiplicity distribution in 200 GeV minimum bias pp collisions. \label{fig:pp_mult}}
51: \end{center}
52: \end{figure}
53: %
54:
55: \subsection{Deuteron - Gold collisions}
56: %
57: %
58: %
59: \begin{figure}[!h]
60: \begin{center}
61: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/analysis/dAuMult.eps}}
62: \caption{Uncorrected charged particle multiplicity measured in the East FTPC (on the outgoing Au side) in 200 GeV dAu collisions. \label{fig:dau_mult1}}
63: \end{center}
64: \end{figure}
65: %
66: %
67: %
68: \begin{figure}[!h]
69: \begin{center}
70: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/analysis/dAu_TPC_FTPC_mult.eps}}
71: \caption{Uncorrected charged particle multiplicity measured in the {\bf East FTPC} vs. measured in the {\bf TPC} in 200 GeV dAu collisions. \label{fig:dau_mult2}}
72: \end{center}
73: \end{figure}
74: %
75: %
76: %
77: \begin{figure}[!h]
78: \begin{center}
79: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/analysis/dAuMult2.eps}}
80: \caption{Uncorrected charged particle multiplicity measured in the TPC in 200 GeV dAu collisions. Color coding refers to the centrality selection as shown in Fig.~\ref{fig:dau_mult1}. \label{fig:dau_mult3}}
81: \end{center}
82: \end{figure}
83: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
84: %
85: %
86: %
87: In dAu collisions the z component of the primary vertex in each minimum bias event has to satisfy the following condition: $\left|v_{z}\right|< $ 50.0 cm. With this vertex cut and minimum bias trigger, 8.8 M good events are selected. Broader vertex distribution is used in dAu than in pp collisions because of the asymmetric bunch timing.
88: In dAu collisions the uncorrected reference multiplicity is defined in the East FTPC, as shown in Fig.~\ref{fig:dau_mult1}, (situated on the outgoing Au side) as the number of charged primary tracks in the pseudo-rapidity range of: -3.8 $< \eta <$ -2.8. Three centrality classes are selected based on the East FTPC, which represent 0-20$\%$, 20-40$\%$, 40-100$\%$ of the geometrical cross-section. Figure~\ref{fig:dau_mult2} shows the uncorrected charged particle multiplicity measured in the East FTPC as a function the TPC multiplicity. Collisions selected in a FTPC multiplicity window correspond to a broad range of multiplicites in the TPC. Figure~\ref{fig:dau_mult3} shows the multiplicity distributions of the corresponding FTPC centrality selection.
89:
90: Collision properties for pp and dAu collisions are summarized in Table~\ref{tab:auau_coll_prop}.
91:
92:
93: %
94: %
95: %
96: \begin{figure}[!h]
97: \begin{center}
98: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/analysis/AuAuMult.eps}}
99: \caption{Uncorrected charged particle multiplicity measured in the TPC in {\bf 62.4} GeV Au-Au collisions (front) and in {\bf 200} GeV Au-Au collisions (back). \label{fig:auau62_mult}}
100: \end{center}
101: \end{figure}
102: %
103: %
104: %
105: \subsection{Gold - Gold collisions}
106: %
107: %
108: %
109: Events collected in 62.4 GeV Au-Au collisions are required to have a $z$ vertex component in $\left|v_{z}\right|< $ 30 cm. With this vertex cut and minimum bias trigger selection 6.3 M good events are selected. Nine centrality classes are defined based on the charged particle multiplicity measured in -0.5 $<$ $\eta$ $<$ 0.5. Figure~\ref{fig:auau62_mult} shows the centrality selection of the uncorrected charged particle multiplicity in 62.4 GeV collisions and the corresponding multiplicity and centrality selection in 200 GeV collisions.
110: The nine centrality bins correspond to the fraction of the total geometrical cross-section: 0 - 5\%, 5 - 10\%, 10 - 20\%, 20 - 30\%, 20 - 30\%, 30 - 40\%, 40 - 50\%, 50 - 60\%, 60 - 70\%, 70 - 80\% as shown in Table~\ref{tab:auau_coll_prop}. The last centrality bin 80-100\% is not used in data analysis due to significant trigger bias.
111:
112: \section{Track selection}
113:
114: Tracks selected for spectra analysis are required to satisfy certain quality cuts. The first criterion is the number of fit points cut. Tracks traversing through the TPC volume can leave 45 possible hits. To avoid splitting tracks we require at least 25 fit points on the track. The distance of closest approach (dca) should be less than 3 cm, which ensures that tracks come from the triggered event vertex and not from a secondary collision or interaction. These tracks are called primary tracks. To estimate the systematic errors on track selection three additional variations of quality cuts have been implemented as shown in Table~\ref{Tab:qa_cuts}. Set 1 represents the default quality cuts for the spectra analysis implemented in this work.
115: %
116: %
117: %
118: \begin{table}[!h]
119: \begin{center}
120: \caption{Collection of quality cuts, implemented for systematic studies. }
121: \begin{tabular}{|c|c|c|c|c|}
122: \hline
123: Cuts & {\bf Set 1} & Set 2& Set 3 & Set 4\\ \hline
124: $|y|<$ & {\bf 0.1} & 0.1 & 0.1 & 0.3 \\
125: Number of fit points $\geq$ & {\bf 25} & 35 & 25 & 25\\
126: dca (cm) $\leq$ & {\bf 3.0} & 3.0 & 1.0 & 3.0 \\ \hline
127:
128: \end{tabular}
129: \end{center}
130: \label{Tab:qa_cuts}
131: \end{table}
132: %
133: %
134: \section{Short description of Monte Carlo Glauber calculation}\label{sec:glauber}
135:
136: Sometimes it is desirable to connect measurements to geometrical quantities of the collisions. Typical examples are the number of participants ($N_{part}$) and the number of binary collisions ($N_{bin}$ or $N_{bin}$) or even the impact parameter (b). These parameters cannot be directly measured, but can be calculated in a geometrical model of a nucleus-nucleus collision, namely the Glauber model~\cite{GlModel}. The model is based on individual nucleon-nucleon collisions which are controlled by the elementary nucleon-nucleon cross-section. In the Monte Carlo Glauber calculation nuclei are independently generated, distributing the nucleons according to the Wood-Saxon density profile:
137: \begin{equation}
138: \rho(r) = \frac{\rho_o}{1+e^{\frac{r-r_0}{a}}} .
139: \end{equation}
140: Here $r_0 = 6.5 \pm 0.1$~fm and $a = 0.535 \pm 0.027$~fm are experimentally measured in $e$-Au scattering~\cite{Antinori:2000ph} and $\rho_0 = 0.169$~fm$^{-3}$ is fixed by the normalization.
141: Each nucleon in the nucleus is separated by a distance larger than $d_{min}$ = 0.4 fm. This cut off value is the characteristic length of the repulsive force acting on the nucleons.
142:
143: $N_{part}$ is defined as the total number of nucleons that underwent at least one collision. $N_{bin}$ is defined as the total number of interactions in the event. The nuclei generation and the nucleon-nucleon selection is repeated with random impact parameter ($b$) selection, where $b^2$ is a flat distribution. The extracted quantities can be studied as the fraction of the total geometrical cross-section.
144: The distributions of $d\sigma/db$ (and $d\sigma/dN_{part}$ and $d\sigma/dN_{bin}$) are determined. Each distribution is divided into bins corresponding to the fractions of the measured total cross-section of the used centrality bins and the mean values of $\left\langle N_{part}\right\rangle$ and $\left\langle N_{bin}\right\rangle$ are extracted for each centrality bin.
145:
146: Moreover, from the MC Glauber calculation the transverse area ($S_{Glauber}$) of the colliding nuclei can be determined from the spatial distribution of the nucleons. The $S_{Glauber}$ is defined as the average transverse area of the overlapping nucleons in a given centrality bin.
147: To make a comparison with previously published results, the overlap area ($S$) can also be calculated as:
148: \begin{equation}
149: S\ =\ \pi\cdot R^{2} =\ \pi\cdot \left( a_{0}\cdot A^{1/3}\right)^2 =\ \pi\cdot a_{0}^{2}\cdot (0.5\cdot N_{part})^{2/3}
150: \end{equation}
151: where $a_{0}$ = 1.12.
152: For detailed description of the Glauber calculation implemented in STAR, we refer the reader to~\cite{Adams:2003yh}. In the calculations the proton-proton cross-sections are obtained from the Particle Data Group~\cite{Hagiwara:2002fs}.
153:
154: The proton-proton cross-section used in the MC Glauber calculation is 36 $\pm$ 3 mb for 62.4 GeV and 41 $\pm$ 3 mb for 200 GeV. Systematic uncertainties are obtained from the variation of the proton-proton cross-section by $\pm$ 3 mb and the variation of the Wood-Saxon parameters.
155: The calculated MC Glauber quantities are listed in Table~\ref{tab:auau_coll_prop}.
156:
157:
158:
159:
160:
161: %
162: \section{Embedding}
163:
164: The correction in our analysis relies on good knowledge of the detector and its simulation. The STAR geometry has been implemented in GEANT~\cite{geant,pnevski} with detailed detector material description. Moreover, realistic simulation of the TPC pad response has been implemented~\cite{Gong:00} in the STAR simulation framework. Physical processes such as drift of the electrons in the TPC gas, the amplification of the signal at the sense/read-out wires, the induction on the readout pads, and the response of the readout electronics (ADCs) are encoded in the TPC Response Simulator (TRS).
165:
166: To obtain realistic corrections, simulated tracks (from GEANT) are embedded into a real event at the raw data (ADC) level. The traces of charged particles in the TPC are simulated, starting with the initial ionization of the TPC gas, then electron transport and multiplication in the drift field, and finally the induced signal on the TPC’s read-out pads and the response of read-out electronics (TRS). The obtained raw simulated signal is then embedded into a real event and then passed through the STAR Offline reconstruction chain. The resulting mixed events are of the same format and contain the same information as real raw data delivered by the data acquisition system. This procedure is called $embedding$, providing nearly realistic simulation of the collision environment.
167:
168: \subsection{Hit level studies}
169:
170: To calculate proper efficiencies it is important to check the quality of the embedding process. First, hit level quantities are compared from embedding and real events, such as X-Y hit distributions in the east and west TPC padrows, as shown in Fig.~\ref{fig:xyhittpc}.
171: %
172: %
173: \begin{figure}[!h]
174: \begin{center}
175: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/analysis/all_XY_hits.eps}}
176: \caption{Hit distribution in the East and West half of the TPC for real and embedded events in 62.4 GeV Au-Au collisions.}\label{fig:xyhittpc}
177: \end{center}
178: \end{figure}
179: %
180: %
181: Since the embedded MC tracks are reconstructed with real events starting at the raw hit level, the calibration database of the given run has to be propagated into the embedding as well. (Separate off-line event reconstruction chains are used for embedding and real events.) Figure~\ref{fig:xyhittpc} shows the hit distributions from real data (left panels) and embedding (right panel). The hits density is represented in the color coding. The sector structure of the TPC is clearly shown. Empty white spots in the sensitive area of the TPC represent dead sectors and the larger white areas at the 4 o'clock position represents a bad Read Out Board for this particular run. Propagation of the correct hit level calibration information is essential to calculate proper efficiencies.
182:
183: The amount of embedded tracks is $\sim$ 5$\%$ of the total number of tracks in the real event. To calculate acceptance and tracking efficiency corrections one has to use the reconstructed $associated$ $tracks$. In the reconstruction process hit information of the MC track is kept and can be compared to the hit information of the reconstructed tracks. A MC track is associated to a reconstructed track if they share at least 3 common hit points within 5 mm in $x$, $y$ and $z$ hit coordinates.
184: %
185: \begin{figure}[!h]
186: \begin{center}
187: \resizebox{.9\textwidth}{!}{\includegraphics{Plots/detectors/dipangleTIF.eps}}
188: \caption{Schematic view of the crossing angle ($\lambda$) and the dip angle ($\phi$)~\cite{dipcrosstex}.\label{fig:dipandcross}}
189: \end{center}
190: \end{figure}
191: %
192: For embedding calibration purposes, the longitudinal and transverse resolution have to be compared to real data as a function of the longitudinal distance ($z$), the crossing angle and the dip angle. In the local coordinate system of the padrow, a coordinate system can be defined as the $x$ axis is along the padrow direction and the $y$ axis is perpendicular to that, as shown in Fig.~\ref{fig:dipandcross} (left panel). The first points of the track are denoted as $x_{0}$, $y_{0}$, $z_{0}$.
193: The crossing angle is the angle enclosed by the momentum of the particle crossing the padrow and the $x$ direction. The dip angle ($\lambda$) is defined as the angle between the momentum of the particle and the momentum component perpendicular to the drift direction, as shown in Fig.~\ref{fig:dipandcross} (right panel). Hit level quantities are propagated to track finding and hence to multiplicity and spectra quantities, therefore embedding has to reproduce data reasonably well.
194: %
195: %
196: \begin{figure}[!h]
197: \begin{center}
198: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/analysis/Piminus_longRes_dipAng_p2.eps}}
199: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/analysis/Piminus_longResMean_dipAng_p2.eps}}
200: \caption{Longitudinal hit resolution and mean as a function of the tangent of the dip angle.}\label{fig:longhitdipangle}
201: \end{center}
202: \end{figure}
203: %
204: %
205: %
206: %
207: \begin{figure}[!h]
208: \begin{center}
209: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/analysis/Piminus_longRes_z_p2.eps}}
210: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/analysis/Piminus_longResMean_z_p2.eps}}
211: \caption{Longitudinal hit resolution and mean as a function of the $z$ coordinate.}\label{fig:longhitz}
212: \end{center}
213: \end{figure}
214: %
215: %
216: %
217: %
218: \begin{figure}[!h]
219: \begin{center}
220: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/analysis/Piminus_transRes_crosAng_p2.eps}}
221: \resizebox{.45\textwidth}{!}{\includegraphics{Plots/analysis/Piminus_transRes_z_p2.eps}}
222: \caption{Transverse hit resolution as a function of the crossing angle and the $z$ coordinate.}\label{fig:tranhitcrosz}
223: \end{center}
224: \end{figure}
225: %
226: %
227:
228: As an example of the hit level simulation of the TPC, the comparison of longitudinal and transverse hit resolution between real data and embedding as a function of the dip angle (Fig.~\ref{fig:longhitdipangle}), $z$ vertex coordinate (Fig.~\ref{fig:longhitz}) and the crossing angle (Fig.~\ref{fig:tranhitcrosz}) are shown. Plots are generated from negative kaon embedding and real data produced from 62.4 GeV Au-Au collisions in the transverse momentum range: 400 - 500 MeV/c. The embedding can reproduce real data well, both transverse and longitudinal hit resolution is $\sim$ 10 - 12\%, and deviation form the mean is less than 2 $\%$.
229:
230: \subsection{Track level studies}
231:
232: Since the same analysis cuts are applied on the embedding and on real data, to extract the efficiencies one has to compare the track level distributions (cuts used to select tracks for identified particle spectra): $dca$ and number of fit points ($N_{fit}$).
233: %
234: %
235: %
236: \begin{figure}[!h]
237: \begin{center}
238: \resizebox{.45\textwidth}{!}{
239: \includegraphics{Plots/analysis/pp_pim_dca_in_pT_200_300_range.eps}}
240: \resizebox{.45\textwidth}{!}{
241: \includegraphics{Plots/analysis/pp_pim_dca_in_pT_600_700_range.eps}}
242: \caption{Comparison of $\pi^{-}$ $dca$ extracted from data and embedding for 200 GeV pp collisions.}
243: \label{fig:pp_embeddatacomppim1}
244: \end{center}
245: \end{figure}
246: %
247: %
248: \begin{figure}[!h]
249: \begin{center}
250: \resizebox{.45\textwidth}{!}
251: {
252: \includegraphics{Plots/analysis/pp_pim_nfit_in_pT_200_300_range.eps}}
253: \resizebox{.45\textwidth}{!}{
254: \includegraphics{Plots/analysis/pp_pim_nfit_in_pT_600_700_range.eps}}
255: \caption{Comparison of $\pi^{-}$ fit points, extracted from data and embedding for 200 GeV pp collisions.}
256: \label{fig:pp_embeddatacomppim2}
257: \end{center}
258: \end{figure}
259: %
260: %
261: %
262: \begin{figure}[!h]
263: \begin{center}
264: \resizebox{.45\textwidth}{!}{
265: \includegraphics{Plots/analysis/pp_km_dca_in_pT_300_400_range.eps}}
266: \resizebox{.45\textwidth}{!}{
267: \includegraphics{Plots/analysis/pp_km_dca_in_pT_600_700_range.eps}}
268: \caption{Comparison of $K^{-}$ $dca$ extracted from data and embedding for 200 GeV pp collisions.}
269: \label{fig:pp_embeddatacompkm1}
270: \end{center}
271: \end{figure}
272: %
273: %
274: \begin{figure}[!h]
275: \begin{center}
276: \resizebox{.45\textwidth}{!}
277: {
278: \includegraphics{Plots/analysis/pp_km_nfit_in_pT_300_400_range.eps}}
279: \resizebox{.45\textwidth}{!}{
280: \includegraphics{Plots/analysis/pp_km_nfit_in_pT_600_700_range.eps}}
281: \caption{Comparison of $K^{-}$ fit points extracted from data and embedding for 200 GeV pp collisions.}
282: \label{fig:pp_embeddatacompkm2}
283: \end{center}
284: \end{figure}
285: %
286: %
287: %
288: %
289: \begin{figure}[!h]
290: \begin{center}
291: \resizebox{.45\textwidth}{!}{
292: \includegraphics{Plots/analysis/pp_pbar_dca_in_pT_400_500_range.eps}}
293: \resizebox{.45\textwidth}{!}{
294: \includegraphics{Plots/analysis/pp_pbar_dca_in_pT_800_900_range.eps}}
295: \caption{Comparison of $\overline{p}$ $dca$ extracted from data and embedding for 200 GeV pp collisions.}
296: \label{fig:pp_embeddatacomppbar1}
297: \end{center}
298: \end{figure}
299: %
300: %
301: \begin{figure}[!h]
302: \begin{center}
303: \resizebox{.45\textwidth}{!}
304: {
305: \includegraphics{Plots/analysis/pp_pbar_nfit_in_pT_300_400_range.eps}}
306: \resizebox{.45\textwidth}{!}{
307: \includegraphics{Plots/analysis/pp_pbar_nfit_in_pT_700_800_range.eps}}
308: \caption{Comparison of $\overline{p}$ fit points extracted from data and embedding for 200 GeV pp collisions.}
309: \label{fig:pp_embeddatacomppbar2}
310: \end{center}
311: \end{figure}
312: %
313: %
314:
315:
316: %%%%%%%%%%%%%%%%%%%% AuAu 62 GeV %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
317: %
318: %
319: %
320: \begin{figure}[!h]
321: \begin{center}
322: \resizebox{.45\textwidth}{!}{
323: \includegraphics{Plots/analysis/auau62_pim_dca_in_pT_300_400_cent_0_range.eps}}
324: \resizebox{.45\textwidth}{!}{
325: \includegraphics{Plots/analysis/auau62_pim_dca_in_pT_300_400_cent_8_range.eps}}
326: \caption{Comparison of $\pi^{-}$ $dca$ extracted from data and embedding for 62.4 GeV Au-Au collisions.}
327: \label{fig:auau62_embeddatacomppim1}
328: \end{center}
329: \end{figure}
330: %
331: %
332: %
333: \begin{figure}[!h]
334: \begin{center}
335: \resizebox{.45\textwidth}{!}{
336: \includegraphics{Plots/analysis/auau62_pim_dca_in_pT_500_600_cent_0_range.eps}}
337: \resizebox{.45\textwidth}{!}{
338: \includegraphics{Plots/analysis/auau62_pim_dca_in_pT_500_600_cent_8_range.eps}}
339: \caption{Comparison of $\pi^{-}$ $dca$ extracted from data and embedding for 62.4 GeV Au-Au collisions.}
340: \label{fig:auau62_embeddatacomppim2}
341: \end{center}
342: \end{figure}
343: %
344: %
345: %
346: \begin{figure}[!h]
347: \begin{center}
348: \resizebox{.45\textwidth}{!}{
349: \includegraphics{Plots/analysis/auau62_pim_nfit_in_pT_300_400_cent_0_range.eps}}
350: \resizebox{.45\textwidth}{!}{
351: \includegraphics{Plots/analysis/auau62_pim_nfit_in_pT_300_400_cent_8_range.eps}}
352: \caption{Comparison of $\pi^{-}$ fit points extracted from data and embedding for 62.4 GeV Au-Au collisions.}
353: \label{fig:auau62_embeddatacomppim3}
354: \end{center}
355: \end{figure}
356: %
357: %
358: %
359: \begin{figure}[!h]
360: \begin{center}
361: \resizebox{.45\textwidth}{!}{
362: \includegraphics{Plots/analysis/auau62_pim_nfit_in_pT_500_600_cent_0_range.eps}}
363: \resizebox{.45\textwidth}{!}{
364: \includegraphics{Plots/analysis/auau62_pim_nfit_in_pT_500_600_cent_8_range.eps}}
365: \caption{Comparison of $\pi^{-}$ fit points extracted from data and embedding for 62.4 GeV Au-Au collisions.}
366: \label{fig:auau62_embeddatacomppim4}
367: \end{center}
368: \end{figure}
369: %
370:
371: In the presented embedding real data comparisons are repeated for each particle species, collision types and multiplicity/centrality. As one can see in the comparison plots embedding can successfully reproduce real data within 3$\sigma$ particle selection. Figure~\ref{fig:pp_embeddatacomppim1} shows the $dca_{\pi^{-}}$ distribution and Fig~\ref{fig:pp_embeddatacomppim2} shows the Nfit$_{\pi^{-}}$ distribution in 200 GeV pp collisions. Figure~\ref{fig:pp_embeddatacompkm2} and Fig.~\ref{fig:pp_embeddatacompkm2} show the same distribution for negatively charged kaons. Antiproton distributions are shown in Fig.~\ref{fig:pp_embeddatacomppbar1} and Fig.~\ref{fig:pp_embeddatacomppbar2}. Complete set of the $dca$ and $N_{fit}$ plots can be found in Appendix~\ref{app:dataembed}.
372:
373: These distributions are also plotted for central (0-5\%) and peripheral (70-80\%) 62.4 GeV Au-Au collisions. The $dca_{\pi^{-}}$ is shown in Fig.~\ref{fig:auau62_embeddatacomppim1} and Fig.~\ref{fig:auau62_embeddatacomppim2}. The $Nfit_{\pi^{-}}$ is shown in Fig.~\ref{fig:auau62_embeddatacomppim3} and Fig.~\ref{fig:auau62_embeddatacomppim4}.
374:
375: $Dca$ extracted from real data shows wider a distribution compared to embedding, especially at low transverse momentum. This is due to secondary contaminations, especially at low momentum. The secondary contaminations is most pronounced in the real proton $dca$ distribution at low momentum.
376:
377: The number of fit points cut is important to avoid merging and splitting tracks in charged multiplicity (number of fit points $\geq$ 15) and spectra (number of fit points $\geq$ 25) measurements. For each colliding set the number of fit points distributions extracted from embedding and real data agree well for number of fit points 10 and higher.
378:
379: The overall agreement of the embedding and real data ensures that corrections extracted from embedding reflect realistic calculations.
380:
381:
382: \subsection{Corrections}
383:
384: Raw spectra are corrected for detector acceptance, tracking inefficiency, hadronic interactions and resonance particle decays.
385: The following subsections provide detailed overviews of these corrections.
386:
387: Since detector parameters (gas pressure in the TPC, temperature) can change over
388: the run, a minimum uncertainty ($\sim$ 5\%) is assigned to the obtained correction factors.
389: Errors on efficiencies are binomial and calculated as:
390: \begin{equation}
391: Err(p)=\sqrt{\frac{p(1-p)}{N}}
392: \end{equation}
393: where $p$ is the efficiency in a given bin and $N$ is the number of entries in the bin.
394:
395:
396:
397:
398: %==================================================================================================
399: \subsection{Energy loss correction}
400:
401: Low momentum particles lose a significant amount of energy traveling
402: through the detector material. The track reconstruction algorithm
403: takes into account the Coulomb scattering and the energy loss, but
404: assumes $pion$ $mass$ for each particle. Therefore, the reconstructed momentum for heavier particles (in our case: kaons and protons/antiprotons) is biased, lower than the original momentum.
405:
406: The correction is obtained from embedding by comparing the input MC and the reconstructed momentum: p$_{\normalfont{reconstructed}}$-p$_{\normalfont{MC}}$ as a function of p$_{\normalfont{reconstructed}}$ as shown in Fig.~\ref{fig:energyloss}.
407:
408:
409: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
410: \begin{figure}[thbp]
411: \centering
412: \includegraphics[width=0.9\textwidth]{Plots/analysis/eloss_pp.eps}
413: \includegraphics[width=0.9\textwidth]{Plots/analysis/eloss_dau.eps}
414: \includegraphics[width=0.9\textwidth]{Plots/analysis/eloss_auau62cent.eps}
415: \includegraphics[width=0.9\textwidth]{Plots/analysis/eloss_auau62per.eps}
416: \caption{Energy loss correction for $\pi^{\pm}$, $K^{\pm}$, p and $\overline{p}$
417: as a function of momentum in 200 GeV pp collisions (first row), 200 GeV dAu collisions (second row), central (0-5\%) (third row) and peripheral (70-80\%) (fourth row) 62.4 GeV Au-Au collisions.}
418: \label{fig:energyloss}
419: \end{figure}
420: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
421: Increasing bias can be observed with increasing particle mass at low momentum.
422: Furthermore, Fig.~\ref{fig:energyloss} also shows the extracted correction for particles: $\pi^{\pm}$, $K^{\pm}$, $\overline{p}$ and p for 200 GeV pp, 200 GeV dAu and 62.4 GeV Au-Au collisions. At low transverse momenta the difference for protons is $\sim$ 100 - 120 MeV/c and decreases to $<$ 10 MeV at $p_{T}$ = 1 GeV/c. This limits our low $p_{T}$ cut off for protons/antiprotons.
423:
424: The pion transverse momentum difference is flat through the measured $p_{T}$ range and the correction is smaller than 0.3$\%$ at any $p_{T}$. This is because $\pi$ energy loss is corrected in reconstruction and the remaining small effect is negligible. However, kaons and protons/antiprotons show larger discrepancy between the MC and the reconstructed transverse momentum at low momentum and the deviation from MC input is the same for particles and antiparticles.
425:
426:
427: Energy loss for a specific particle type is independent of collision type (pp - dAu - peripheral Au-Au) as expected.
428: The slight difference between collision types is due to the changing detector setup between different runs (pp: Run II., dAu: Run III., and Au -Au: Run IV.) and only controlled by the amount of absorbing material in the detector itself. Between the runs, as already mentioned, the SVT supporting frame has been modified and a new silicon layer has been installed: SSD. The GEANT description of the SVT and SSD becomes more refined through the runs and the corresponding simulation packages, but it also introduces uncertainty on the calculated efficiencies and corrections which is included in the overall 5\% uncertainty.
429:
430: The energy loss correction for kaons and protons/antiprotons can be parameterized as:
431: %
432: \begin{equation}
433: {p_{T}}_{corrected}={p_{T}}_{measured}+c_{1}+c_{2}\left(1+\frac{m^{2}}{p^{2}}\right)^{c_{3}}
434: \end{equation}
435: %
436: %
437: %
438: where $m$ is the mass of the particle and $c_{i}$ (i=1, 2, 3) are the parameters extracted from the fit to the energy loss curve.
439: The change in the fit parameters is negligible between the collision systems and centralities, therefore the characteristic numbers can be quoted:
440: $c_1 = 0.006 (0.013)$~GeV/$c$, $c_2 = -0.0038 (-0.0081)$, and $c_3 = 1.10 (0.03)$ for kaons (protons/antiprotons), respectively.
441:
442: The energy loss correction seems to show a small dependence on centrality in dAu and in Au-Au collisions. Weak dependence can be observed with varying quality cuts. The change in the energy loss correction due to different rapidity selection in $|y| < 0.5$ is negligible.
443:
444: The energy loss correction is applied off-line to the raw data upon selecting tracks from the dE/dx distribution to be used for spectra analysis.
445: Since individual particle identification is not possible, the energy loss correction of the particular specie of interest is applied to all tracks, eg. when analysing kaons each track (even from the pion and proton/antiproton bands) are corrected for kaon energy loss.
446: This method does not introduce artificial bias on the extracted raw particle yield, since raw yield is only extracted for a particular particle at one time and it only changes the scale in the direction of the transverse momentum, but leaves the magnitude of the dE/dx unchanged.
447:
448: %==================================================================================================
449: \subsection{Vertex correction}
450:
451: In pp and dAu collisions the average number of tracks per event
452: is small compared to Au-Au collisions, and the event rate is high.
453: Multiple bunch crossing (pile-up) within the same read-out window (complete drift of the triggered event to the read out electronics) is a significant problem of drift detectors, which leads to high background rate in the triggered events.
454: In higher multiplicity collisions a larger number of tracks defines the vertex more precisely, but in low multiplicity
455: collisions the vertex finder (ppLMV: proton - proton Low Multiplicity Vertex Finder implemented in STAR) is sensitive to pile-up events. Pile-up can shift the position of the reconstructed primary vertex. In very low multiplicity events ppLMV can fail to find the vertex.
456:
457: To correct for these inefficiencies, the possible problems are identified. In a second step a physical quantity is identified which is insensitive to the pile-up rate and accessible from data (since the pile-up level is not known from data). In the third step data is corrected using the quantity mentioned above.
458: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
459: \begin{figure}[!h]
460: \centering
461: \includegraphics[width=0.45\textwidth]{Plots/analysis/vtx_efficiencies_pp.eps}
462: \includegraphics[width=0.45\textwidth]{Plots/analysis/vtx_efficiencies_dau.eps}
463:
464:
465: \caption{Vertex inefficiencies in minimum bias pp and dAu collisions at 200 GeV.}
466: \label{fig:vtxineff}
467: \end{figure}
468: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
469:
470: The ppLMV can fail to find the primary vertex or it can find the vertex at a wrong place.
471: To correct for these inefficiencies in pp and dAu collisions, MC (HIJING) events were embedded to abort gap events and were
472: reconstructed in the full reconstruction chain. (Events triggered and reconstructed at empty bunch crossings are called abort gap events.)
473: In every MC event there is a well defined primary vertex with well defined coordinates. After reconstructing the embedded events with the MC information in hand the vertex reconstruction efficiency can be studied.
474:
475: In data analysis, only those events are taken which satisfy certain quality cuts. The first one is the cut on the primary vertex position. The $x$ and $y$ positions are well defined (and also restricted by the beam pipe), however the $z$ position can vary along the beam direction over a wide range due to the difference in bunch timing of the two collider rings. Good events with $z$ vertex position are selected from $\pm$ 30.0 cm in pp and $\pm$ 50.0 cm in dAu for data analysis. The overall vertex (in)efficiency can be determined as the ratio of the number of reconstructed good events with respect to the number of good MC events satisfying the vertex cut, as shown by Eq.~\ref{eq:vtxineff}.
476: \begin{equation}
477: Efficiency\ =\ \frac{number\ of\ good\ reconstructed\ events}{number\ of\ good\ MC\ events}
478: \label{eq:vtxineff}
479: \end{equation}
480: This overall correction is applied as a multiplication factor to the extracted raw yield.
481: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
482: \begin{figure}[!h]
483: \centering
484: \includegraphics[width=0.45\textwidth]{Plots/analysis/multimap_pp.eps}
485: \includegraphics[width=0.45\textwidth]{Plots/analysis/multimap_dau.eps}
486:
487: \caption{Map of number of good primary tracks vs. number of good global tracks.}
488: \label{fig:vtxmapping}
489: \end{figure}
490: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
491:
492: To characterize the vertex inefficiency a parameter should be
493: chosen that can be measured in the data and is not effected
494: by the pile-up. In order to study the pile-up effect two simulated
495: files of pp events are mixed at the raw data level and reconstructed
496: in the full event reconstruction chain. The first set is considered
497: the real event and the other set is used as the pile-up background event.
498: The pile-up range was varied from 0 - 100 $\%$, where 100$\%$ means
499: each real event has a pile-up event in it. After mixing, events
500: were reconstructed and the number of good global and primary tracks
501: were examined as a function of the pile-up level. (A global or primary track is called good if
502: its distance of closest approach is smaller than 3cm and it has at least 15 fit points out of 45 possible).
503: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
504: \begin{figure}[!h]
505: \centering
506: \includegraphics[width=0.85\textwidth]{Plots/analysis/pp_dau_vtx_eff.eps}
507: \caption{Vertex finding efficiency as a function of good primary tracks in 200 GeV $\bf{pp}$ and $\bf{dAu}$ collisions.}
508: \label{fig:ppvtxeff}
509: \end{figure}
510: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
511: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
512: \begin{figure}[!h]
513: \centering
514: \includegraphics[width=0.85\textwidth]{Plots/analysis/fake_spectra_pp.eps}
515: \caption{Transverse momentum spectra of good and fake events extracted from embedding in 200 GeV $\bf{pp}$ collisions. }
516: \label{fig:ptspectrafake}
517: \end{figure}
518: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
519:
520: The number of good primary tracks were chosen to characterize the vertex inefficiency
521: since the number of good primary tracks is independent of the
522: degree of pile-up. The number of good global tracks increases with increasing
523: pile-up rate in the event. The vertex inefficiency cannot be described by the
524: number of good primary tracks directly. Hence, the correction is performed
525: in two steps. First, the number of lost events the number of fake events are obtained as a function of good global tracks
526: (which are required to have at least 15 hit points).
527: This is shown in Fig.~\ref{fig:vtxineff}.
528: %
529: %
530: %
531: \begin{figure}[!h]
532: \centering
533: \includegraphics[width=0.85\textwidth]{Plots/analysis/pp_dAu_fake_mb.eps}
534: \caption{Fake vertex correction as a function of $p_{T}$ in 200 GeV $\bf{pp}$ and $\bf{dAu}$ collisions. }
535: \label{fig:FakeVertexCorrection}
536: \end{figure}
537: %
538: %
539: %
540: The number of good global track depends on the pile-up as well.
541: In the second step, the efficiency distributions are converted to the function of the number of good primary tracks through
542: the mapping of the good primary good global track distribution.
543: The mapping is shown in Fig.~\ref{fig:vtxmapping}.
544: For each good primary track bin the lost and fake distributions are convoluted with
545: the good global distribution. Finally, the vertex correction is given as a
546: function of the number of good primary tracks as shown in Fig.~\ref{fig:ppvtxeff}
547: and applied in each event to the raw particle spectra. Each event and each track is weighted by
548: the inverse of the vertex efficiency in pp, minimum bias dAu and peripheral dAu events.
549:
550: %==================================================================================================
551: \subsection{Fake vertex correction}
552:
553: As we mentioned in the previous section pile-up can affect the vertex reconstruction. One typical problem is the shifted vertex.
554:
555: In the low multiplicity events the reconstruction software can be biased by the pile-up and the vertex may be reconstructed away from the real vertex. This can be studied via embedding. If the reconstructed vertex in the embedding is farther than 2 cm from the corresponding MC vertex, the reconstructed vertex is labeled as a fake vertex.
556: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
557: \begin{figure}[!h]
558: \centering
559: \includegraphics[width=0.45\textwidth]{Plots/analysis/pp_tracking_eff_neg.eps}
560: \includegraphics[width=0.45\textwidth]{Plots/analysis/pp_tracking_eff_pos.eps}
561: \caption{Tracking efficiency of $\pi^{-}$, $K^{-}$, $\overline{p}$ (left panels) and $\pi^{+}$, $K^{+}$ and $p$ (right panels) in 200 GeV minimum bias $\bf{pp}$ collisions as a function of transverse momentum.}
562: \label{fig:treffpp}
563: \end{figure}
564: \begin{figure}[!h]
565: \centering
566: \includegraphics[width=0.45\textwidth]{Plots/analysis/dau_tracking_eff_neg.eps}
567: \includegraphics[width=0.45\textwidth]{Plots/analysis/dau_tracking_eff_pos.eps}
568: \caption{Tracking efficiency of $\pi^{-}$, $K^{-}$, $\overline{p}$ (left panels) and $\pi^{+}$, $K^{+}$ and $p$ (right panels) in 200 GeV $\bf{dAu}$ collisions as a function of transverse momentum.}
569: \label{fig:treffdau}
570: \end{figure}
571: \begin{figure}[!t]
572: \centering
573: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff05_neg.eps}
574: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff05_pos.eps}
575:
576: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff510_neg.eps}
577: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff510_pos.eps}
578:
579: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff1020_neg.eps}
580: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff1020_pos.eps}
581:
582: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff2030_neg.eps}
583: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff2030_pos.eps}
584:
585: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff3040_neg.eps}
586: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff3040_pos.eps}
587:
588:
589: \caption{Tracking efficiency of $\pi^{-}$, $K^{-}$, $\overline{p}$ (left panels) and $\pi^{+}$, $K^{+}$ and $p$ (right panels) in 62.4 GeV $\bf{Au-Au}$ collisions as a function of transverse momentum and centrality.}
590: \label{fig:treffauau1}
591: \end{figure}
592: %
593: In the real data it is not possible to determine whether a vertex is fake or not. Therefore,
594: one has to apply a different approach. It is found that the transverse momentum spectrum of the
595: fake events is different than the spectrum of good events as shown in Fig.~\ref{fig:ptspectrafake}.
596: Spectrum from fake events is flatter than from good events, presumably due to the refitting of tracks including the wrong vertex position.
597:
598: Figure~\ref{fig:FakeVertexCorrection} shows the ratio of charged hadron $p_{T}$ spectrum in good vertex events to that in all events with a reconstructed vertex (i.e. sum of good and fake vertex events) for minimum bias pp and d-Au collisions. The spectra are normalized per event before the ratio is taken. Therefore, a single function:
599: \begin{equation}
600: \frac{p_{T}^{good vertex}}{p_{T}^{all vertex}}\ =\ c_{0}\cdot e^{c_{1}\cdot p_{T}^{ c_{2}} }-1
601: \label{eq:fakecorr}
602: \end{equation}
603: is used for the $p_{T}$-dependence of the fake vertex correction. The correction is done by multiplying the $p_{T}$ spectra by the function described in Eq.~\ref{eq:fakecorr}.
604: In low multiplicity and minimum bias pp events and in peripheral
605: and minimum bias dAu events the correction is $\sim$ 3 - 5 \% and vanishes around $p_{T} \sim$ 1 GeV as shown in Fig.~\ref{fig:FakeVertexCorrection}.
606:
607: %==================================================================================================
608:
609:
610:
611: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
612: \subsection{Tracking efficiency and acceptance}
613:
614: Raw particle spectra have to be corrected for detector and tracking efficiency. The corrections are obtained from MC embedding. The obtained correction includes the net effect of detector acceptance, tracking efficiency interaction losses, decays, etc.
615:
616: For each investigated particle specie ($\pi^{\pm}, K^{\pm}, p$ and $\overline{p}$) particles are embedded with flat distributions in $p_{T}$ and $y$ to have uniform statistics.
617: \begin{figure}[!h]
618: \centering
619:
620: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff4050_neg.eps}
621: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff4050_pos.eps}
622: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff5060_neg.eps}
623: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff5060_pos.eps}
624: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff6070_neg.eps}
625: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff6070_pos.eps}
626: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff7080_neg.eps}
627: \includegraphics[width=0.45\textwidth]{Plots/analysis/auau62_treff7080_pos.eps}
628:
629:
630: \caption{Tracking efficiency of n$\pi^{-}$, $K^{-}$, $\overline{p}$ (left panels) and $\pi^{+}$, $K^{+}$ and $p$ (right panels) in 62.4 GeV $\bf{Au-Au}$ collisions as a function of transverse momentum and centrality.}
631: \label{fig:treffauau2}
632: \end{figure}
633:
634: Tracking efficiency for 200 GeV pp collisions is shown in Fig.~\ref{fig:treffpp}, for 200 GeV dAu collisions in Fig.~\ref{fig:treffdau} and for 62.4 GeV Au-Au collisions in Fig.~\ref{fig:treffauau1} and Fig.~\ref{fig:treffauau2}.
635: %
636:
637: %
638: \begin{table}
639: \scriptsize
640: \begin{center}
641: \caption{Tracking efficiency parameterization for pions.\label{tab:treffpion}}
642: \begin{scriptsize}
643: \begin{tabular}{|c|c|c|c|c|c|c|}
644: \hline
645: Collision & \multicolumn{3}{|c|}{$\pi^{-}$ ($c_{0,1,2} $)} & \multicolumn{3}{|c|}{$\pi^{+}$ ($c_{0,1,2}$)}\\ \hline
646: pp MB & 0.89905 & 0.06974 & 1.73860 & 0.90280 & 0.06526 & 1.60725 \\ \hline
647: Au-Au 62.4 GeV 70-80\% & 0.84839 & 0.08926 & 2.55107 & 0.83915 & 0.12901 & 4.70611 \\ \hline
648: Au-Au 62.4 GeV 0-5\% & 0.76833 & 0.06332 & 1.21189 & 0.76969 & 0.06244 & 1.18241\\ \hline
649: \end{tabular}
650: \end{scriptsize}
651: \end{center}
652: \end{table}
653: %
654: %
655:
656: %
657: %
658: \begin{table}
659: \scriptsize
660: \begin{center}
661: \caption{Tracking efficiency parameterization for kaons.\label{tab:treffkaon}}
662: \begin{scriptsize}
663: \begin{tabular}{|c|c|c|c|c|c|c|}
664: \hline
665: Collision & \multicolumn{3}{|c|}{$K^{-}$ ($c_{0,1,2} $)} & \multicolumn{3}{|c|}{$K^{+}$ ($c_{0,1,2}$)}\\ \hline
666: pp MB & 0.73584 & 0.24398 & 2.49097 & 0.75943 & 0.23179 & 2.03419\\ \hline
667: Au-Au 62.4 GeV 70-80\% & 0.60802 & 0.23787 & 2.42484 & 0.51912 & 0.24281 & 5.39325\\ \hline
668: Au-Au 62.4 GeV 0-5\% & 0.45012 & 0.22891 & 3.92516 & 0.45835 & 0.23128 & 4.02922\\ \hline
669: \end{tabular}
670: \end{scriptsize}
671: \end{center}
672: \end{table}
673: %
674:
675: %
676: %
677: \begin{table}
678: \scriptsize
679: \begin{center}
680: \caption{Tracking efficiency parameterization for protons/antiprotons.\label{tab:treffproton}}
681: \begin{scriptsize}
682: \begin{tabular}{|c|c|c|c|c|c|c|}
683: \hline
684: Collision: $\overline{p}$ and $p$ & \multicolumn{2}{|c|}{pp MB } & \multicolumn{2}{|c|}{Au-Au 62.4 GeV 70-80\%} & \multicolumn{2}{|c|}{Au-Au 62.4 GeV 0-5\%} \\ \hline
685: $p_{0}$ & 0.76733 & 0.930071 & 7.76198 & 32.7826& 0.35058 & 0.33582 \\ \hline
686: $p_{1}$ & 0.28580 & 0.290520 & 0.30947 & 0.32498 & 0.30266 & 0.33119 \\ \hline
687: $p_{2}$ & 13.08348 & 10.57888 & 20.65861 & 16.3731& 11.2653 & 9.44440 \\ \hline
688: $p_{3}$ & 0.11409 & -0.00805 & 1.71647 & 5.65292 & 0.02679 & 0.01084 \\ \hline
689: $p_{4}$ & - & - & 40.61263 & 363.1 & 0.53201 & 0.53312 \\ \hline
690: $p_{5}$ & - & - & 3.29003 & 732.8 & 0.09304 & 0.22337 \\ \hline
691: $p_{6}$ & - & - & 0.13745 & 0.09921 & 1.11354 & 2.66341 \\ \hline
692:
693: \end{tabular}
694: \end{scriptsize}
695: \end{center}
696:
697: \end{table}
698: %
699: %
700:
701: Tracking efficiencies for $\pi^{+}$ and $\pi^{-}$ are similar. Pion efficiency quickly rises from $p_{T}$ = 0.1 GeV/c to $p_{T} \approx$ 0.3 GeV/c and levels off at $\sim$ 90\%. Small difference can be observed at low transverse momenta for $K^{+}$ and $K^{-}$ and for $p$ and $\overline{p}$ due to the absorption effect in the detecor material. Kaon efficiencies rise monotonicaly with increasing transevse momentum. Proton/antiproton efficiencies rise sharply around $p_{T}$ = 0.3 GeV/c and flatten out with an increasing damping in mid-central and central Au-Au collisions.
702:
703: Calculated tracking efficiencies depend on the measured particle
704: multiplicity, which is clearly shown in Au-Au collisions, comparing the most central (0-5 \%) and the peripheral bins (eg. 60-70 \%). The change in the pion efficiency at $p_{T} = 400$ MeV between most central and most peripheral Au-Au collisions is $\sim$ 20 \%, which correspondes to a change of 350 between the average charged particle multiplicity of the two centralities.
705: In pp and dAu collisions the variation of event multiplicity is small (in the highest multiplicity pp bin the average multiplicity is $\sim$ 10, and $\sim$ 19 in central dAu collisions), therefore tracking efficiency has no dependence on event classes.
706:
707: Tracking efficiency also depends on the quality cuts as expected. The final corrected spectra should be the same for different quality cuts.
708: Systematic uncertainties in efficiency can therefore be assessed by different quality cuts.
709: %
710: \begin{figure}[!h]
711: \centering
712: \includegraphics[width=0.45\textwidth]{Plots/analysis/PiBG_pp_mb.eps}
713: \includegraphics[width=0.45\textwidth]{Plots/analysis/PipBG_pp_mb.eps}
714: \caption{Correction to pion spectra for weak decays ($\Lambda$,$K^{0}_{S}$ and muons) as a function of $p_{T}$ in 200 GeV $\bf{pp}$ collisions. }
715: \label{fig:pibgpp}
716: \end{figure}
717: %
718: \begin{figure}[!h]
719: \centering
720: \includegraphics[width=0.45\textwidth]{Plots/analysis/PiBG_dAu_mb.eps}
721: \includegraphics[width=0.45\textwidth]{Plots/analysis/PipBG_dAu_mb.eps}
722: \caption{Correction to pion spectra for weak decays ($\Lambda$,$K^{0}_{S}$ and muons) as a function of $p_{T}$ in 200 GeV $\bf{dAu}$ collisions. }
723: \label{fig:pibgdAu}
724: \end{figure}
725: %
726: %
727: %
728: \begin{figure}[!h]
729: \centering
730: \includegraphics[width=0.45\textwidth]{Plots/analysis/PiBG_AuAu62.eps}
731: \includegraphics[width=0.45\textwidth]{Plots/analysis/PipBG_AuAu62.eps}
732: \caption{Averaged correction to pion spectra for weak decays ($\Lambda$,$K^{0}_{S}$ and muons) as a function of $p_{T}$ in 62.4 GeV $\bf{Au-Au}$ collisions. }
733: \label{fig:pibgAuAu62}
734: \end{figure}
735: %
736:
737: The extracted raw identified particle spectra are corrected for tracking efficiencies through the following parameterizations.
738: Pion tracking efficiency is fitted to the following function:
739: %cicc
740: %
741: \begin{equation}
742: Efficiency\ (\pi)=c_{0}\cdot e^{-\left(\frac{c_{1}}{p_{T}}\right)^{c_{2}}}
743: \end{equation}
744: %
745: %
746: where $c_{0,1,2}$ are the parameters.
747: To account for the flat rising part of the kaon efficiency the following function can be used:
748: %
749: \begin{equation}
750: Efficiency\ (K)=c_{0}\cdot e^{-\left(\frac{c_{1}}{p_{T}}\right)^{c_{2}}}+c_{3}\cdot p_{T}
751: \label{eq:kaoneff}
752: \end{equation}
753: %
754: where $c_{0,1,2,3}$ are the parameters.
755: To characterize the proton/antiproton efficiency two different functions are implemented. For saturating efficiency in pp and dAu collisions a similar form can be used as for kaons.
756: To account for the small decrease in the proton/antiproton efficiency a simple linear combination of the previous functions can be used in mid-central and central Au-Au collisions. Table~\ref{tab:treffpion}, Tab.~\ref{tab:treffkaon} and Tab.~\ref{tab:treffproton} show a representative set of the extracted parameters.
757:
758:
759:
760: %cicc
761:
762:
763: As Figs.~\ref{fig:treffpp},~\ref{fig:treffdau},~\ref{fig:treffauau1},~\ref{fig:treffauau2} show, these functions describe the tracking efficiencies well.
764:
765: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
766: \subsection{Pion background corrections}
767:
768: %
769: The pion spectra are corrected for weak decays, muon contamination and background pions from the detector materials.
770: Corrections are extracted from HIJING and PYTHIA simulations propagated through the STAR geometry and reconstructed as real data.
771: For each simulated particle, the origin, the parent particle and the decay particle type are known. From this information, we can select pions created in the simulated collision (primary particles) from the ones created in the detector material or produced from resonance decay.
772:
773: The weak-decay daughter pions are mainly from $K_{0}$ and $\Lambda$ and are identified by the parent particle information accessible from the simulation. In real data, pion decay muons can be mis-identified as primordial pions because of the similar masses of muon and pion. By selecting the parent particle information in the simulation, the muon contamination can be extracted.
774:
775: Once these selections are applied, the amount of background contamination can be extracted for each transverse momentum bin as shown in Fig.~\ref{fig:pibgpp} for 200 GeV pp and in Fig.~\ref{fig:pibgdAu} for 200 GeV dAu collisions.
776: %
777: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
778: %
779: The magnitude of the background pion contamination falls steeply. At low transverse momentum ($p_{T}$ = $\sim$ 0.3 GeV/c) it is in the order of $\sim$ 15\% and decreases to $\sim$ 5\% at $p_{T}$ = 1 GeV/c.
780:
781: The pion background is independent of event multiplicity in 200 GeV $pp$ and d-Au collisions; therefore, a single correction is applied. In 62.4 GeV Au-Au collisions the multiplicity dependence of the pion background is week, within 1.5\% over the entire centrality range. Therefore, a single averaged correction is applied to all centralities, similarly to~\cite{Adams:2003xp}. The correction is shown in Fig.~\ref{fig:pibgAuAu62}.
782:
783:
784:
785:
786:
787:
788:
789: \subsection{Proton background corrections}
790: %
791:
792: %
793: A particle, created in the primary collision, has to travel through the beam pipe, the SVT (and SSD) layers and the support structure and finally the TPC Inner Field Cage (IFC) before it can be detected in the TPC active drift volume. Particles traversing through the detector material create secondary particles due to the interaction with the nuclei of the detector material~\cite{Ashery:1987nt}. Radiation lengths are small for the subsystems (SVT: $<$ 6\%~\cite{Bellwied:2002ag}, SSD: 1\%~\cite{Arnold:2002wx} and TPC IFC: 0.5\%~\cite{Anderson:2003ur}) but the level of background protons is considerable at low $p_{T}$.
794: %
795: %
796: %
797: \begin{figure}[!h]
798: \begin{center}
799: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_0_5_pT_400_450.eps}
800: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_0_5_pT_600_650.eps}
801: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_0_5_pT_900_950.eps}
802: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_70_80_pT_400_450.eps}
803: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_70_80_pT_600_650.eps}
804: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_70_80_pT_900_950.eps}
805:
806:
807: \caption{Sample proton/antiproton dca distributions for various transverse momenta in central (0-5\%) and in peripheral (70-80\%) 62.4 GeV Au-Au collisions.
808: \label{fig:pbg_auau62_bins}}
809: \end{center}
810: \end{figure}
811:
812: \begin{figure}[!h]
813: \begin{center}
814:
815: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_1_100_pT_40_45.eps}
816: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_1_100_pT_60_65.eps}
817: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_1_100_pT_90_95.eps}
818:
819:
820: \caption{Sample proton/antiproton dca as a function of transverse momentum in Minimum Bias 200 GeV pp collisions.}
821: \label{fig:pbg_pp_mb}
822: \end{center}
823:
824: \end{figure}
825: %
826: %
827: %
828: \begin{figure}[!h]
829: \begin{center}
830:
831: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_1_2_pT_40_45.eps}
832: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_1_2_pT_60_65.eps}
833: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_1_2_pT_90_95.eps}
834:
835: \caption{Sample proton/antiproton dca as a function of transverse momentum in $Nch_{0-2}$ 200 GeV pp collisions.\label{fig:pbg_pp_nch02}}
836: \end{center}
837:
838: \end{figure}
839:
840: To correct for background protons, the $dca$ distributions of protons and antiprotons are extracted and compared from real data. Since the $dca$ distribution of the background protons cannot be obtained from real data, the $dca$ distribution of the background protons is obtained from embedding. The method presented here is the same as can be found in earlier STAR publications~\cite{Adler:2001bp,Adler:2001aq}.
841:
842: Due to the geometry of the detector structure, secondary protons are created far from the primary vertex (couple cm away, mainly in the beam pipe), hence their global $dca$ will be larger than for primary protons. Since antiprotons do not create secondaries their $dca$ distribution should be the same as that primary protons, hence the background contribution can be extracted as follows:
843: \begin{equation}
844: dca_{proton}\ =\ p_{1}\cdot dca_{background\ protons}\ +\ p_{2}\cdot dca_{antiprotons}\ +\ p_{3}\cdot dca_{primary\ protons}
845: \label{eq:protonbggeneral}
846: \end{equation}
847: %
848: %
849: \begin{figure}[!h]
850: \begin{center}
851:
852: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_9_100_pT_40_45.eps}
853: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_9_100_pT_60_65.eps}
854: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_9_100_pT_90_95.eps}
855:
856: \caption{Sample proton/antiproton dca as a function of transverse momentum in $Nch_{9-...}$ 200 GeV pp collisions. \label{fig:pbg_pp_nch9100}}
857: \end{center}
858:
859: \end{figure}
860: %
861: %
862: \begin{figure}[!h]
863: \begin{center}
864:
865: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_0_200_pT_40_45.eps}
866: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_0_200_pT_60_65.eps}
867: \includegraphics[width=0.3\textwidth]{Plots/analysis/protonBG_nch_0_200_pT_90_95.eps}
868:
869: \caption{Sample proton/antiproton dca as a function of transverse momentum in minimum bias 200 GeV dAu collisions.}
870: \end{center}
871: \label{fig:pbg_dau_mb}
872: \end{figure}
873: %
874: %
875: \begin{figure}[!h]
876: \begin{center}
877: \includegraphics[width=0.45\textwidth]{Plots/analysis/pp_proton_bg_fractions.eps}
878: \includegraphics[width=0.45\textwidth]{Plots/analysis/dAuAuAu_proton_bg_fractions.eps}
879:
880:
881: \caption{Fraction of primary protons as a function of transverse momentum in 200 GeV pp collisions (left panel) and in 200 GeV dAu and 62.4 GeV Au-Au collisions (right panels).\label{fig:pbg_corr_curves}}
882: \end{center}
883:
884: \end{figure}
885: %
886: %
887: \begin{figure}[!h]
888: \end{figure}
889:
890: Primary tracks are selected within 3$\sigma$ of the Bethe-Bloch parameterization of the protons/anitproton energy loss bands. The $dca$ of primary tracks is defined up to 3 cm, therefore to extract proper corrections, one has to access the dca region up to 10 cm or so.
891: This is achieved through the mapping of global tracks to primary tracks.
892:
893: In real data there exists a map between primary and global tracks, so the momentum of global tracks can be associated to the primary one (global and primary momentum can be different for the same track due to refitting) and can be corrected for energy loss of protons/antiprotons (which is obtained for primary tracks). By mapping the global tracks to the energy loss corrected primary tracks, the global dca of global tracks corrected global momentum map is created. This map is obtained for each multiplicity/centrality bin. By this mapping, proton and antiproton $dca$ distributions can be compared up to $dca \sim$ 20 cm, in each $p_{T}$ bin, as shown in Fig.~\ref{fig:pbg_auau62_bins}.
894:
895: The long, nearly flat $dca$ tail in the proton distribution comes mainly from knock-out background proton, due to interactions of produced particles with detector materials. Antiprotons do not have knock-out background and the flat $dca$ tail is absent from their dca distribution. In order to correct for the knock-out background protons, the $dca$ dependence at $dca<3$~cm is needed for the knock-out protons. Such a dependence is obtained from MC simulations, and is found to be of the form~\cite{Adler:2001bp}:
896: %
897: %
898: \begin{equation}
899: p_{knocked-out\ proton}(dca) = 1-e^{-\frac{dca}{dca_0}}.
900: \label{eq:dca_knockout}
901: \end{equation}
902: %
903: %
904: This background contribution is indicated by the green curve in Fig.~\ref{fig:pbg_auau62_bins} and for the rest of the proton background plots in this subsection.
905:
906: Assuming that the background subtracted proton $dca$ distribution is identical in shape to the anti-proton $dca$ distribution, Eq.~\ref{eq:protonbggeneral} can be written as:
907: %
908: \begin{equation}
909: p(dca) = \overline{p}(dca)/r_{\overline{p}/p} + A\cdot p_{knocked-out proton}(dca).
910: \label{eq:dca_fit}
911: \end{equation}
912: %
913: Data can be fitted treating the magnitude of the background protons $A$, the parameter $dca_0$, and the antiproton-to-proton ratio $r_{\overline{p}/p}$ as free parameters.
914:
915: This assumption is, however, not strictly valid because the weak decay contributions to proton and anti-proton are in principle different and the $dca$ distribution of the weak decay products differs from that of the primordial (anti-)protons. However, the measured anti-lambda/lambda ratio is close to the anti-proton/proton ratio~\cite{Adler:2002uv}. The difference in $dca$ distributions between protons and anti-protons arising from weak decay contaminations is small.
916:
917: The $dca$ distributions of protons and antiprotons in Au-Au collisions are fitted with Eq.~(\ref{eq:dca_fit}) in each $p_{T}$ and centrality bin, as shown in Fig.~\ref{fig:pbg_auau62_bins}.
918: The obtained fraction of knock-out background protons is approximately 60\% at $p_{T} = 0.35$~GeV/c and less than 5\% at $p_{T} = 1$~GeV/c. The amount of knock-out background protons depends directly on the total particle multiplicity and their kinetic energies produced in the collisions. Since the proton multiplicity over total particle multiplicty varies somewhat with centrality and the particle kinematics change with centrality, the background fraction depends on centrality. The variation from peripheral to central collisions is on the order of 10\% in Au-Au collisions.
919:
920: In pp collisions, the proton background strongly depends on the multiplicity, as shown in Fig.~\ref{fig:pbg_pp_mb} and Fig.~\ref{fig:pbg_pp_nch02}, Fig.~\ref{fig:pbg_pp_nch9100}. In large multiplicity events the evolution of the dca distributions are the same as observed in Au-Au collisions, but in low multiplicity events, even in minimum bias pp collisions, significant background excess develops. This background excess is smaller in minimum bias and peripheral dAu collisions, than in pp collisions and disappears in central dAu collisions as shown in Fig.~\ref{fig:pbg_dau_mb}. Looking at pp and peripheral and minimum bias dAu collisions in detail, on top of the long, flat $dca$ tail due to knock-out protons, there are two bumps in the proton dca distribution, one at $dca<4$~cm and the other at $4<dca<10$~cm. These bumps, absent from the antiproton dca distributions, come from other sources of proton background. The sharp drop at 4~cm suggests that they come from interactions in the beam pipe which is at radius of 4~cm. For straight tracks originating from the beam pipe and uniform in azimuth, the distribution in $dca$ (from the primary vertex of the real event) would be of the form of $1/\sqrt{1-(dca/4 {\rm cm})^2}$. This form gives not an unreasonable description of the bump at $dca<4$~cm considering the finite $dca$ resolution and curving of low momentum particles in the magnetic field.
921:
922: One possible source of such interactions is those between beam protons and the beryllium beam pipe. Such interactions are asymmetric and have a nucleon-nucleon center-of-mass energy of $\sqrt{S_{NN}}\approx 14$~GeV, which is smaller than that at the SPS. The proton-antiproton pair production rate is small compared to the number of protons transported from the initial baryons; the antiproton/proton ratio at mid-rapidity ($y\approx 2.5$) should be smaller than that at the SPS, which is on the order of a few percent. While the produced particles (pions, antiprotons, etc.) are symmetric in rapidity, the proton rapidity distributions should be peaked significantly toward the target rapidity ($y=0$) because a large fraction of the protons come from the beryllium target. The number of these protons may not be small compared to protons produced in 200 GeV pp collisions though the relative magnitudes of the bump and produced protons in a $pp$ collision depend on the rate of the background interaction per pp event. At $y=0$, the antiproton/proton ratio should be significantly smaller than that at $y\approx 2.5$, which could explain that such a background is not observable in the antiproton $dca$ distributions.
923:
924: In principle, such background interactions should also be present in Au-Au running, depending on the quality of beam focusing. However, the center-of-mass of Au-beryllium interactions is shifted significantly to the Au beam rapidities. The proton (and antiproton) yield at rapidity $y=0$ should be very small. This is especially so when compared to the large multiplicity in Au-Au collisions.
925:
926: The bump at $4<dca<10$~cm in the proton distribution in $pp$ could be due to interactions between the beam and the SVT and SSD materials which are located outside the beam pipe. The two-bump structure is also observed in minimum bias d-Au and peripheral d-Au collisions. The effect is not significant in the other two centralities of d-Au.
927:
928:
929: In order to correct for those background protons, the bumps are parameterized by
930: \begin{equation}
931: p_{bump,\ 0<dca<4\ cm} =\left(1.2+1.8\frac{dca}{4{ cm}}\right)\left(1-\exp \left[-7.2\left(1-\frac{dca}{4{ cm}}\right)\right]\right) \\
932: \end{equation}
933: \begin{equation}
934: p_{bump,\ 4<dca<10\ cm} =0.54\left[1-\left(\frac{dca}{3{ cm}}-\frac{7}{3}\right)^2\right]
935: \label{eq:dca_bumps}
936: \end{equation}
937: %
938: %
939: %
940: The $dca$ distributions from $pp$ (minimum-bias and various multiplicity classes) and d-Au (minimum bias and the peripheral centrality) are fitted with
941: %
942: %
943: \begin{equation}
944: p(dca) = \overline{p}(dca)/r_{\overline{p}/p} + A\cdot p_{knock-out} + B\cdot p_{bump},
945: \label{eq:dca_fit_pp}
946: \end{equation}
947: %
948: %
949: treating $A$, $dca_0$, $B$, and $r_{\overline{p}/p}$ as free parameters. For the two non-peripheral d-Au centralities, the same fit procedure using Eq.~\ref{eq:dca_fit}, as for the Au-Au data, is used.
950: Different curves in Fig.~\ref{fig:pbg_pp_mb} represent the various background contributions. The green curve is the knock-out proton background described by Eq.~\ref{eq:dca_knockout}. The pink curve is the additional proton background described by Eq.~\ref{eq:dca_bumps}. The black histogram is the fit results by Eq.~\ref{eq:dca_fit_pp}. The fitted background protons are subtracted from the proton data.
951:
952: The knock-out proton background should scale with the event multiplicity, as the fit results indicate. If the beam-material interaction is responsible for the two-bump structure, then the magnitude of such background should be independent of the event multiplicity. This seems to be supported by the fit results.
953: The parameter $dca_0$ is found to be approximately 2.0 in pp, d-Au and varies between 2.5 - 1.4 in Au-Au collisions with increasing $p_{T}$ and shows almost no (weak) centrality dependence.
954:
955: %
956:
957: Figure~\ref{fig:pbg_corr_curves} shows the fits to the obtained primary proton fractions as a function of $p_{T}$ in minimum bias pp and dAu, and for two multiplicities/centralities in pp and Au-Au. In higher multiplicity and minimum bias collisions the primary proton fraction rises quickly from $\sim$ 30-40 \% to 100 \% in the measured proton spectra range. The lowest multiplicity pp bin carries significant proton background even at larger transeverse momentum. In dAu and Au-Au collisions the primary proton fraction is larger and steeply increases to 100 \%. Corrections depend on multiplicity/centrality therefore the primary proton fraction is calculated for each multiplicity/centrality selection.
958:
959:
960:
961: %
962:
963:
964:
965:
966:
967:
968:
969:
970: