Measurements of Cross Sections for e + e - m + m - at Energies from 3.8 to 4.6 GeV

Transcription

1 Measurements of Cross Sections for e + e - m + m - at Energies from 3.8 to 4.6 GeV G. Rong, Y. Fu, J.Y. Zhao (Institute of High Energy Physics, CAS, Beijing, China) BESIII Charmonium Physics Meeting, 12 June., 2018

2 Motivation Summary Outline Measurements of cross sections Data & Monte Carlo Events Measuring cross sections Analysis of cross sections Br(y s (4040, 4160, 4416 )m + m - ) Phase angle of y s (4040, 4160, 4416 )m + m - Evidence for structure S(4230)m + m -

3 Motivation Measurements of cross sections for e + e - m + m - could give some important information about vector meson production and decays Up to now, no branching fraction of heavy y s [y(4040), y(4160) and y(4415)] electromagnetic di-muon decays is available in Review of Particle Physics or in Particle Physics BookLet Measure the branching fractions of the heavy y s electromagnetic di-muon decays. Measure the phase angles of the heavy y s relative to the continuum e + e - m + m - process. Search for some new structure(s) in the decay of Sm + m These improve our knowledge for better understanding of the vector meson production in e + e - annihilation, as well as for better understanding of the vector meson decays

4 Data Data Samples and Software Data taken at energies from 3.8~4.4 GeV in 2013 Data taken at energies from 3.85~4.6 GeV in 2013 Data taken at energies from 4.4~4.6 GeV in 2014 Data taken at GeV ( pb -1 ) Software BOSS version of p01 Monte Carlo events are generated with Babayaga Other Monte Carlo simulated events are generated with the KKMC + BesEvtGen

5 Charged track cos < 0.8 V xy < 1 cm V z < 10 cm N Good = 2, Selection of m +/- T 1 -T 2 <2 ns (T 1 and T 2 are the time from TOF) (p m+ +p m- )>0.9E cm 4C-Kinematic fit Event Selection i Q = <E EMC /p<0.40 c 2 4C-Fit <30 (to reject KK and pp backgrounds)

6 No. of candidates for e + e - m + m - To separate e + e - m + m - from e + e - K + K - and e + e - pp, we examine the energy distributions of the accepted events satisfying the selection criteria. If a K or p is misidentified as a m, the energy of the charged track would be lower than that expected. This allows us to separate m/k/p. We defined a quantity E measured /E cm to examine the energy distributions of the selected events from data samples collected at different energies. The number of candidates for e + e - m + m - is obtained by fitting the distributions of the normalized energies (E measured /E cm ) of the final states satisfying the selection criteria.

7 Numbers of candidates for e + e - m + m - E measured /E cm Signal shape: Double Gaussian function Background shape: Polynomial (More distributions of the ratio and fits are shown in the backup slides)

8 Background Estimate No. of background events The events satisfying the selection criteria still contain some backgrounds. The number of these background events can be subtracted from the events with b N i N Lσ η i i( KK, eemm ) L : luminosity N b : the number of background s i : cross section of i th Bkg. source : mis-identification rate Possible backgrounds e + e - m + m - is mis-identified from e + e - + -, KK, pp e + e - m + m - is mis-identified from e + e e + e - m + m - is mis-identified from e + e - e + e - X

9 No. Events Main Backgrounds Possible backgrounds The distribution of the M 2 missing of the events satisfying the selection criteria shows that no the e + e event is satisfied with the selection criteria We study possible backgrounds with all kinds of Monte Carlo simulated event samples generated at GeV and other E cm E measured /E cm The Monte Carlo studies shows that the e + e is the mainly backgrounds from the continuum e + e - qq events

10 Number of Backgrounds Cross sections for e + e Misidentification rate Using this cross-section shape and shape, we can estimate the number of background events at energies from 3.8 to 4.6 GeV Number of the background events N b = L σ e+ e π + π η

11 Observed Cross Sections Cross Sections s obs e e m m - N obs - N L b where N obs is the number of candidates for e + e - m + m - observed, N b is the number of background events, L is the integrated luminosity, is the detection efficiency Luminosity Measured with e + e - (g)e + e -, or e + e - gg Efficiency Determined with Babayaga, and/or with KKMC+BesEvtGen

12 MC Events & Selection Efficiency Monte Carlo simulated signal events At each energy point of the R scan data taken at energies from 3.85~4.6 GeV in 2013, we generated Monte Carlo signal events for e + e - m + m - ; We generated Monte Carlo signal events at the energies from 3.8~4.6 GeV for the XYZ scan data taken in 2013 in Selection efficiencies Analyzing these Monte Carlo simulated signal events yields selection efficiencies at each of these energies.

13 Comparison of E EMC /p between Data & MC The ratio of E EMC /p between the data and the Monte Carlo simulated events at energies above 4.45 GeV are difference. These should be corrected in determination of the cross sections

14 Correction Factor f EMC to Correction factor f EMC to selection efficiency The difference in ratio of E EMC /p between the data and the Monte Carlo simulated events varies with energy above 4.4 GeV, resulting under estimate of the selection efficiency. By comparing the distributions of the E EMC /p from both the data and the Monte Carlo simulated events, we obtain an energy dependent correction factor f EMC to the selection efficiencies. f EMC Observed cross section s obs e e m m N obs - L N b f EMC

15 Compare Data and MC c 2 distributions from both the data and the MC events at a few energies (as an example)

16 Comparison of Data and MC Distributions of cos, f and momentum of m for both the data and the MC simulated events

17 Comparison of Data and MC Distributions of E EMC /p, T m+ -T m- and E measured /E cm of m + (m - ) for both the data and the MC simulated events

18 Luminosities and Their Corrections The published luminosities of the data samples are used in the analysis Yifan Yang, Guangyi Tang, and Changzheng Yuan found that the published luminosities of the data samples suffer from some problems due to missing readout some crystal in a reconstructed cluster. They gave a set of correction factors of the published luminosities. [see their talk given at the BES-III collaboration meeting in December 3, 2017] Using these correction factors, we obtained the right luminosities of the data samples collected at energies above 4.1 GeV in 2013 and 2014,which are f Crr Lum L = (1.0 + f Crr Lum) L Published Lum

19 The Data Collected in 2013 and 2014 As an example, we list the published and corrected luminosities of the data collected at 10 energy points for XYZ scan Ecm (GeV) Lum (nb -1 ) factor Lum_Cor (nb -1 ) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± L Published Lum L

20 The Data Collected in 2013 As an example, we list the published and corrected luminosities of the data collected at 10 energy points for R scan Ecm (GeV) Lum (nb -1 ) factor Lum_Cor (nb -1 ) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± L Published Lum L

21 The Data Collected in 2013, 2014 Observed cross sections at 10 energy points σ obs = Nobs N b L ε f EMC E cm (GeV) L published lum (nb -1 ) N obs N b N net s (nb) ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

22 The Data Collected in 2013 Observed cross sections at 10 energy points E cm [GeV] L published Lum [nb -1 ] N obs N b N net s [nb] ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± Similarly, we determine the observed cross sections at all other energies

23 Observed Cross Sections R scan data XYZ scan data e + e - m + m - e + e - m + m -

24 Systematic Uncertainties cos < 0.80 cut We select y(3686)j/y + - with J/ym + m - samples, and generate Monte Carlo simulated events. By comparing the distribution of cos between the data and the MC events, we find that the difference in cos is (-0.15 ± 0.04)%, which is assigned as the systematic uncertainty E EMC /p > 0.05 and E EMC /p < 0.4 cut By comparing the distribution of E EMC /p between the data and the MC events, we find that the difference in the E EMC /p is (-0.12 ± 0.01)%, which is assigned as the systematic uncertainty

25 Systematic Uncertainties DTOF < 2.0 ns cut At present we measure the difference in the cross section at GeV with this cut and without this cut. We take the difference as the estimated uncertainty due to this cut, which is -0.48%. (P m+ + P m- ) > 0.9E cm cut We also measure the difference in the cross section at GeV with this cut and without this cut, and take the difference as the estimated uncertainty due to this cut, which is -0.59%. 4-C kinematic fit We select/generate y(3686)j/y + - with J/ym + m - samples, and compare the number of events satisfying K.F. requirement. The difference between the data and the MC is (0.97±0.26)%, which is assigned as the systematic uncertainty due to the K.F.

26 Systematic Uncertainties Fit to E mm /E cm (1) Bin size : GeV/bin GeV/bin; (2) Fit range : [0.92, 1.08] [0.96, 1.04]; (3) Background PDF: 1-order polynomial 0-order polynomial; (4) Signal PDF : Double-Gaussian function MC shape; Bin size Fit range Background PDF Signal PDF Total s sys -0.59% -0.65% -0.59% -0.63% 1.23% Tracking efficiency 1.0% per track Luminosity 1.0% ( M.Ablikim et al.(besiii Collaboration), Phys Rev.Lett.110,252001(2013) ) ( Chin. Phys. C 39, (2015) ) ( M. Ablikim et al. (BESIII Collaboration), Chin. Phys. C 41, (2017) ) Generator 1.0% [

27 Systematic Uncertainties Source Systematic uncertainty (%) cos < E EMC /p > 0.05 and E EMC /p < DTof < (P m+ + P m- ) > 0.9E cm C kinematic fit 0.97 Fit to E mm /E cm 1.23 Tracking efficiency 2.0 Luminosity 1.0 Generator 1.0 Sum 3.01 The total systematic uncertainty in the measured cross section is 3.01%

28 Analysis of the Cross Sections A c 2 fit is performed to the observed cross sections Objective c 2 function 2 c n i1 ( s obs i -s i 2 D Where s obs i is the observed cross section, s th i is the expected cross section, and D i is the uncertainty of the s obs i at ith energy point. The effects of correlation between point and point on the fit parameters are estimated by off-set method (see M. Botje, J. Phys. G (2002)] for detail) Dressed cross section σ dress con s obs i expect A + e A + e A + e A +... con 2 cm iφ1 iφ2 iφ ee uu f 12π Γ j Γ j Acon Aj, (j 1,2,...n) 2 2 tot E Ecm - M j + iγ j M j A j M 2 j 12π - E Γ 2 cm ee j Γ - iγ mm j tot j M j ) 2 e (if using this form of the B-W amplitude, the phase angle would be shifted by 180 ) iφ n A n 2 This definition is consistent with f=0, where f is the angle between continuum e + e - m + m - and y(3686)m + m -

29 s s x ' 1- Effective c.m. energy max )) ( ( ), ( ), ( ) ( ' ' x e e x s s x F dx s s ds G s 0 dress 0 expect 1 s s m m ' ' BEPC-II BEPC-II 2 ) ( exp 2 1 ), ( E E s s s s G s s Kuraev & Fadin H S V x s x F ), ( m e s ln ln e S V m s H H H x H x x x x x x H 6 ) ln(1 ) 3(1 1 1 )ln 4( Nominal c.m. energy Analysis of the Cross Sections Observed cross section

30 Measuring Br., Phase Angle and Searching for New Structure Two hypotheses We fit these cross sections with two hypotheses 1. Cnt + y(3686) +Rs(3770)+y(4040)+y(4160)+y(4415) 2. Cnt + y(3686) +Rs(3770)+y(4040)+y(4160)+y(4415)+S 1 +(S 2 ) S1 (S2) is any other structure(s) lying in the energy range from 3.8 to 4.6 GeV In the fit, the branching fractions for y(3686) and Rs(3770) decays to m + m - are fixed at values obtained by analyzing other data samples Find the better hypothesis By comparing the fit c 2 obtained with these hypotheses, we find the better fit to these cross sections and search for new structure(s)

31 Fit to the Observed Cross Sections Cnt + y(3686)+rs(3770) +y(4040)+y(4160)+y(4415) Parameters M y(4040) [MeV] Solution# (fixed) G tot y(4040) [MeV] 80.0 (fixed) G ee y(4040) [kev] Br[y(4040)m + m - ] [ 10-4 ] 0.86 (fixed) 0.088±0.098±D sys F y(4040) (-229.1±25.2±D sys ) M y(4160) [MeV] G tot y(4160) [MeV] G ee y(4160) [kev] Br[y(4160)m + m - ] [ 10-4 ] (fixed) 70 (fixed) 0.48 (fixed) 1.479±0.032±Dsys F y(4160) (-270.6±7.4±D sys ) M y(4415) [MeV] (fixed) G tot y(4415) [MeV] 62.0 (fixed) G ee y(4415) [kev] Br[y(4415)m + m - ] [ 10-4 ] 0.58 (fixed) 0.080±0.042±D sys F y(4415) (-55.2±20.9±D sys ) s(e + e - m + m - ) [nb] s(e + e - m + m - ) [nb] E cm [GeV] c 2 = c 2 /ndf = points Continuum e + e - m + m - and y(3686)m + m - subtracted

32 s(e + e - m + m - ) [nb] s(e + e - m + m - ) [nb] Fit to the Observed Cross Sections Cnt + y(3686)+rs(3770) +y(4040) +y(4160)+y(4415) + S 1 c 2 /ndf =1.252 Continuum e + e - m + m - and y(3686)m + m - subtracted E cm [GeV] c 2 = points S(4230) signal significance is more than 4s Parameters M y(4040) [MeV] Solution# (fixed) G tot y(4040) [MeV] 80.0 (fixed) G ee y(4040) [kev] Br[y(4040)m + m - ] [ 10-4 ] 0.86 (fixed) 0.167±0.128±D sys F y(4040) (-28.1±25.7±D sys ) M y(4160) [MeV] G tot y(4160) [MeV] G ee y(4160) [kev] Br[y(4160)m + m - ] [ 10-4 ] (fixed) 70 (fixed) 0.48 (fixed) 0.882±0.953±Dsys F y(4160) (-166.3±32.9±D sys ) M y(4415) [MeV] (fixed) G tot y(4415) [MeV] 62.0 (fixed) G ee y(4415) [kev] Br[y(4415)m + m - ] [ 10-4 ] 0.58 (fixed) 0.038±0.030±D sys F y(4415) (-100.4±34.9±D sys ) M y(4230) [MeV] G tot y(4230) [MeV] G ee y(4230) B y(4230) [ev] ±9.1±D sys 42.3±15.2±D sys 0.052±0.038±Dsys F y(4230) (-105.5±33.6±D sys )

33 Comparison with the PDG2016 Parameters Solution#1 PDG2016 M y(4040) [MeV] (fixed) 4039±1 G tot y(4040) [MeV] 80.0 (fixed) 80±10 G ee y(4040) [kev] 0.86 (fixed) 0.86±0.07 Br[y(4040)m + m - ] [ 10-4 ] 0.167±0.128±D sys 0.107±0.016 [for y(4040)e + e - ] F y(4040) (-28.1±25.7±D sys ) N/A G ee y(4160) B y(4160) [ev] M y(4160) [MeV] (fixed) 4191±5 G tot y(4160) [MeV] 70 (fixed) 70±10 G ee y(4160) [kev] 0.48 (fixed) 0.48±0.22 Br[y(4160)m + m - ] [ 10-4 ] 0.882±0.953±Dsys 0.069±0.033 [for y(4160)e + e - ] F y(4160) (-166.3±32.9±D sys ) N/A G ee S1 B S1 [ev] M y(4415) [MeV] (fixed) 4421±4 G tot y(4415) [MeV] 62.0 (fixed) 62±20 G ee y(4415) [kev] 0.58 (fixed) 0.58±0.07 Br[y(4415)m + m - ] [ 10-4 ] 0.038±0.030±D sys 0.094±0.032 [for y(4415)e + e - ] F y(4415) (-100.4±34.9±D sys ) N/A

34 Summary We measured the observed cross sections for e + e - m + m - at energies from 3.8 to 4.6 GeV (with Yang s correction factors to the published luminosities of the data samples). We made the first measurements of the branching fractions of heavy y s [y(4040), y(4160), and y(4415)] electromagnetic di-muon decays. We made the first measurements of the electromagnetic phase angles of the heavy y s relative to the continuum e + e - m + m -. We may observe an evidence for S(4230)m + m - with a signal significance of more than 4s by analyzing these observed cross sections.

35 Evidence for S(3760)m + m - (+ + - ) ~3s significance One of two solutions: Br[y(3686)m + m ] =(7.3±0.3±Dsys) 10-3 f =(7.8±3.7±D sys ) M s1 =(3762.6±2.8±D sys ) MeV G s1 =(8.5±3.5±D sys ) MeV G ee S1 B S1 = ± ev F 1 =(148.5±42.9±D sys ) M y(3770) = MeV (fixed) G y(3770) =27.3 MeV (fixed) G ee y(3770) = 265 ev (fixed) B y(3770) =(2.31±1.32 ±D sys ) 10-4 F 2 =(23.6±18.1±D sys )

36 Distributions of Emeasured/Ecm for e+e- m+m-

37 Distributions of Emeasured/Ecm for e+e- m+m-

38 Distributions of Emeasured/Ecm for e+e- m+m-

39 Distributions of Emeasured/Ecm for e+e- m+m-

40 Distributions of Emeasured/Ecm for e+e- m+m-

41 Distributions of Emeasured/Ecm for e+e- m+m-

42 Distributions of Emeasured/Ecm for e+e- m+m-

43 Distributions of Emeasured/Ecm for e+e- m+m-

44 Distributions of Emeasured/Ecm for e+e- m+m-

45 Distributions of Emeasured/Ecm for e+e- m+m-

46 Distributions of Emeasured/Ecm for e+e- m+m-

47 Distributions of Emeasured/Ecm for e+e- m+m-

48 Distributions of Emeasured/Ecm for e+e- m+m-

49 Correlations The systematic error sources are correlated among the different energy points, so these cannot be considered directly in the fit. But their effects on the final results can be estimated by the offset method [see M. Botje, J. Phys. G (2002)] When constructing the c 2 to fit the cross section, we have considered the correlation between different energy points. At first, we fit the cross sections only considering the statistical uncertainties. At this stage, the uncertainties of the parameter values from the fits are only due to the statistical. To estimate the systematic uncertainties of the parameter values of the structures, we shifted the observed cross sections by +1D sys or - 1D sys, where D sys is the correlated systematic error, then we fitted these cross sections again. The effects of the systematic uncertainty (correlation between energy points) on the parameters are still in progress.

50

51