Resonance analysis in pp collisions with the ALICE detector

Transcription

1 IL NUOVO CIMENO Vol.?, N.?? Resonance analysis in pp collisions with the ALICE detector A Pulvirenti( 1 )( 2 )( ), for the ALICE Collaboration ( 1 ) INFN, Sezione di Catania ( 2 ) Università di Catania Summary. ALICE is the LHC experiment mainly dedicated to the study of hot and high energy density nuclear matter created in heavy-ion collisions. It has also developed a detailed proton-proton physics programme, in order to exploit its capabilities to investigate the novel energy regime made available by LHC, which turns out to be quite interesting in itself besides being important as a baseline for the heavy-ion data. Results will be presented of the first resonance measurements done on the data taken at LHC during 2009 and 20, in pp collisions at 900 GeV and 7 ev center of mass energy. PACS Nq production of quark-gluon plasma. PACS Gz resonance production in relativistic heavy-ion collisions. PACS Jn, n strangeness. 1. Introduction he study of identified particle production in pp collisions provides a fundamental baseline for properly tuning the QCD-inspired models in order to make predictions at higher energies. In the low-p region this helps especially in understanding the nature of thesoft partof the underlying event. he ALICE detector [1], thanks toits low operating magnetic field (0.5 ) and to its excellent particle identification (PID) capabilities, is well suited for the study of low-p particle production at mid-rapidity. Studying the hadronic resonances in pp collisions, besides contributing to this topic, provides also a baseline for a better understanding of heavy-ion collisions, where they are useful probes of the hot and dense matter created there [2, 3, 4, 5, 6]. Several resonances were measured in the first LHC pp runs, taken between the end of 2009 and the first half of 20, respectively at the energies of 900 GeV and 7 ev. Section 2 will briefly illustrate the most important aspects of the ALICE detector exploited for these studies. Section 3 will present in more detail the analysis carried on on the φ(20) resonance with the first 2009 runs at 900 GeV. Section 4 will show some ( ) Corresponding author. alberto.pulvirenti@ct.infn.it c Società Italiana di Fisica 1

2 2 A PULVIRENI, FOR HE ALICE COLLABORAION preliminary results in the analysis of K (892) 0, Σ(1385) and φ(20) resonances with the pp collisions at 7 ev. Finally, in section 5 some conclusions will be given. 2. Experimental setup ALICE is the LHC experiment most specifically devoted to the study of heavy-ion collisions. It is designed to guarantee a very complete reconstruction and identification of the particles produced in those collisions (for details, refer to [1]). he contribution of several detectors is used in this analysis. Charged particles are tracked in the ime Projection Chamber (PC) and in the Inner racking System (IS). he Silicon Pixel Detectors (SPD) in the two innermost IS layers also contribute to the primary vertex reconstruction. Both the PC and the IS return a de/dx measurement which can be used for particle identification (PID). At larger momenta (above 0.7 GeV/c) PID is also supported by the ime of Flight detector (OF). At the trigger level the two VZERO counters allow one to reject most of the beamgas interactions, in order to have an initial hardware-level minimum bias event selection. his is then refined offline by means of a check on trigger masks, in order to accept only the events collected by the two V0 triggers and that coming from the SPD. Moreover, all events with longitudinal primary vertex position Z V cm are rejected in order to avoid biases due to the detector acceptance. 3. φ(20) analysis with LHC commissioning runs in rack selection. he φ resonance was reconstructed through its K + K decay channel (BR 49.2%). racks used for analysis were required to have been reconstructed by both PC and IS, with at least 80 clusters in the PC and a total χ 2 (normalized to the number of PC points) smaller than 4, in order to ensure a reasonable track quality. Since resonance daughters are indistinguishable from primaries, tracks were also required to have a distance of closest approach to primary vertex smaller than 0.5 cm (3 cm) in the transverse (longitudinal) direction. APIDrequestwasalsomade,intheformofa compatibilitycut : atrackwasaccepted if its associated PC and OF signals were close to the expected one for a kaon with the same total momentum, within a fiducial window. he PC fiducial window was chosen to be 3σ PC (5σ PC ) if the track total momentum at the PC inner wall was larger (smaller) than 350 MeV/c (see fig. 1 left); the status of the art of the PC calibration for the analyzed sample allowed one to assume a resolution σ PC = 6%. he OF fiducial window on time difference vs. momentum was defined by two hyperbolas, as it is shown in figure 1 (right), to cope with the worsening resolution at small momenta (for p 260 MeV/c the OF is never used). Since the PC-OF matching efficiency is almost 60%, whenever a OF signal was absent in the track, the PC PID cut only was used Signal extraction and correction. he measurement was done in four p bins between 0.7 and 3 GeV/c and in a rapidity window y 0.6, where the efficiency was found to be almost flat. An invariant mass distribution was computed with all unlike-sign charged track pairs (see fig 2, left) and fitted with a square-root function (A m m 0 ) plus a Gaussian, to reproduce the peak plus the background. hen, the number of φ was computed from the subtraction of the background function integral from the total

3 RESONANCE ANALYSIS IN P P COLLISIONS WIH HE ALICE DEECOR 3 Fig. 1. (Left) PC signal vs. total momentum at the inner PC wall. he central continuous curve is the Bethe-Bloch estimation, and the two broken curves define the momentum-dependent fiducial compatibility window. (Right) he relative difference between OF measured time and Kaon integrated time, vs. total momentum at the vertex; the curves define the fiducial window. number of entries counted in a fixed invariant mass range, defined as m φ ± 4σ, where σ = Γ φ /2.35 and m φ and Γ φ are the PDG nominal mass and width of this resonance. o correct the measured counts, the φ reconstruction efficiency was estimated through a PYHIA simulation made with a realistic implementation of the ALICE detector (see fig. 2, right). Finally, the corrected counts were normalized to the total number of inelastic collisions. wo main sources of systematic error were identified: one was related to the choice of the function for the background subtraction, and another to PC cus (de/dx and number of clusters). hey were estimated to range between 0.4 to 5.6% in the first case and from 0.9 to 6% in the second one, depending on p. Anyway, they are smaller than the statistical error which ranges between % and 20% due to the sample size. Counts Efficiency Inv. mass (GeV/c ) p (GeV/c) Fig. 2. (Left) K + K invariant mass spectrum in the p bin between 1 and 1.5 GeV/c. he curve is the combined fit of square-root function + Gaussian to the points, and the vertical lines delimit the region where the subtraction is computed. (Right) Efficiency of reconstructed φ computed in the PYHIA sample cited in the text, in the four measured p bins.

4 4 A PULVIRENI, FOR HE ALICE COLLABORAION -1 1 N ev (GeV/c) dn dy dp -2-1 (GeV/c) dn dy dp -2 1 N ev -3 pp φ s + X = 900 GeV -3 pp φ s + X = 900 GeV -4 Levy function Exp function (p ) Stat + Syst error Statistical error -4 PYHIA PHOJE Stat + Syst error Statistical error p (GeV/c) p (GeV/c) Fig. 3. (Left) Corrected φ spectrum as a function of p. Shaded boxes represent the sum in quadrature of statistical and systematic errors, and the bars indicate the statistical errors only. he curves are the results of the fit by with the two functions described in the text. (Right) Comparison of measured spectrum with those generated in a PYHIA and in a PHOJE sample, normalized to the same integral in the measured p range (0.7 to 3 GeV/c) Results. he measured and corrected p spectrum was fitted with an exponential function depending on p and with a Levy function, which should reproduce better the power-law tail at large momenta. he full dn/dy distribution was computed by summing the integral fit functions in the GeV/c and 3 GeV/c p ranges and the integral of measured counts in the region between p = 0.7 and 3 GeV/c, which turned out to correspond to 52% of the total. In both cases the average transverse momentum was also computed. Figure 3 (left) reports the measured points and the two used functions, and table I reports the computed yields and mean p. As expected, the two functions return different values for the latter, due to the very different contribution coming from the high momentum tail, even if the values agree within errors. Figure 3 (right) reports a comparison of the measured points with PYHIA and PHOJE. Both MonteCarlo spectra were normalized to the same integral in the measured region (0.7 to 3 GeV/c), and while in this range they appear to be quite compatible, PHOJE gives a larger expectation in the extrapolation region at low p. Integral Levy % Exponential % GeV/c ± ± GeV/c 0.04 ± ± GeV/c ± ± otal ± ± p (GeV/c) 0.98 ± ± 0.06 able I. Computed dn/dy and mean p after the spectra fit (see text). First and third line come from function integration, second line comes from measured histogram integral.

5 RESONANCE ANALYSIS IN P P COLLISIONS WIH HE ALICE DEECOR 5 Fig. 4. Kπ invariant mass spectra for the extraction of the K (892) resonance in pp collisions at the CMS energy of 7 ev (p -integrated). he darker histogram (blue in colored picture) is built with all the unlike-sign pairs, where the peak is well visible, while the other is built using like-sign pairs and is used to estimate and then subtract the combinatoria background. 4. Results at 7 ev he sample of pp collision data at 7 ev has a very large statistics (up to 700 millions of events taken during 20 run). his allowed one to approach the study of resonances which oftern turns out to be difficult due to their width and signal/background ratio (e.g: the K (890) through the charged Kπ decay channel) or to the low efficiency for retrieving their daughters (e.g.: the Σ(1385) through the Λ + π decay channel). Figure 4 shows a preliminary result on the extraction of the K (892) peak from a sample of 17 millions pp minimum-bias events at 7 ev: in this case, the background was estimated from the invariant mass distribution of all like-sign pairs. Figures 5 and 6 show a preliminary result on the the Σ(1385), computed with 76 millions of minimumbias events at the same energy as above. In this case, the background was reproduced by a 3 rd order polynomial fit outside the peak range, and then subtracted. he very large statistics obtained allows one to measure the φ(20) resonance with a p binning much finer than it was possible to do with the sample at 900 GeV. A small sample (about millions of events) of 7 ev pp data has been analyzed in order to extract the φ p spectrum. he analysis was carried out using essentially the same strategy adopted for the 900 GeV sample. Moreover, in this case, the analysis included also the tracks which were reconstructed in the IS stand-alone using all clusters rejected after the global tracking. his is expected to help in recovering about % of the tracks which were lost due to PC dead zones or to their low transverse momentum (leq1 GeV/c). Figure 7 shows a preliminary performance plot where one can clearly see the resonance peak in several small p bins, down to 400 MeV/c. 5. Conclusions We have shown the measurement of φ(20) resonance in ALICE obtained with the first LHC data from pp collisions at 900 GeV and preliminary results on φ(20),

6 6 A PULVIRENI, FOR HE ALICE COLLABORAION Fig. 5. Λπ invariant mass spectrum for the extraction of the Σ(1305) resonance in pp collisions at the CMS energy of 7 ev (p -integrated). he points describe the signal pairs, where the peak is somewhat visible, while the filled histogram is a polynomial description of the background fitted outside the peak range. K (892) 0 and Σ(1385) measurements with the 20 LHC data from pp collisions at 7 ev. he 900 GeV analysis is completed and dn/dy and p were measured for the first time. QCD-inspired models reproduce quite well the φ(20) p spectrum, even if they differ in their predictions at low momenta, where the low statistic available for this analysis prevents one from drawing more detailed conclusions. All preliminary results from data at 7 ev look very promising, and demonstrate the optimum PID performances of the ALICE detector and its capabilities for resonance study. Fig. 6. Λπ invariant mass spectrum after background subtraction, fitted by means of a Breit- Wigner function.

7 RESONANCE ANALYSIS IN P P COLLISIONS WIH HE ALICE DEECOR 7 Fig. 7. K + K invariant mass distribution in several p bins in pp collisions at the CMS energy of 7 ev (bin sizes expressed in GeV/c). In all plots, the signal is well visible over the background and a fit is performed using the combination of a 3 rd order polynomial background and a Breit-Wigner function for the peak. 6. Bibliography REFERENCES [1] ALICE Collaboration, JINS 3 (2008)S08002 [2] Markert C et al, Phys. Lett. B 669 (2008), 92 and references therein [3] UA5 Collaboration, Z. Phys. C 41 (1988), 179; Nucl. Phys. B 328 (1989), 36 [4] E735 Collaboration, Z. Phys. C 67 (1995), 411 [5] AFS Collaboration, Nucl. Phys. B 203 (1982), 27 [6] SAR Collaboration, Phys. Lett. B 612 (2005), 181; Phys. Rev. C 79 (2009),