STUDY OF THE PROCESS e^{+}e^{-}\to 3(\pi^{+}\pi^{-}) IN THE C.M.ENERGY RANGE 1.5–2.0 GEV WITH THE CMD-3 DETECTOR

Study of the Process in the Range 1.5–2.0 Gev With the Cmd-3 Detector

R.R.Akhmetshin A.V.Anisenkov S.A.Anokhin V.M.Aulchenko V.S.Banzarov L.M.Barkov N.S.Bashtovoy D.E.Berkaev A.E.Bondar A.V.Bragin S.I.Eidelman D.A.Epifanov L.B.Epshteyn G.V.Fedotovich S.E.Gayazov A.A.Grebenuk D.N.Grigoriev E.N.Gromov F.V.Ignatov S.V.Karpov V.F.Kazanin B.I.Khazin I.A.Koop A.N.Kozyrev P.P.Krokovny A.E.Kuzmenko A.S.Kuzmin I.B.Logashenko A.P.Lysenko P.A.Lukin K.Yu.Mikhailov Yu.N.Pestov E.A.Perevedentsev S.A.Pirogov S.G.Pivovarov A.S.Popov Yu.S.Popov S.I.Redin Yu.A.Rogovsky A.L.Romanov A.A.Ruban N.M.Ryskulov A.E.Ryzhenenkov V.E.Shebalin D.N.Shemyakin B.A.Shwartz D.B.Shwartz A.L.Sibidanov P.Yu.Shatunov Yu.M.Shatunov I.G.Snopkov E.P.Solodov1 V.M.Titov A.A.Talyshev A.I.Vorobiov Yu.V.Yudin A.S.Zaytsev Budker Institute of Nuclear Physics, SB RAS, Novosibirsk, 630090, Russia Novosibirsk State University, Novosibirsk, 630090, Russia Novosibirsk State Technical University, Novosibirsk, 630092, Russia University of Sydney, School of Physics, Falkiner High Energy Physics, NSW 2006, Sydney, Australia Brookhaven National Laboratory, P.O. Box 5000 Upton, NY 11973-5000, USA University of Tokyo, Department of Physics, 7-3-1 Hongo Bunkyo-ku Tokyo, 113-0033, Japan
September 23, 2019

The cross section of the process has been measured using 22 pb of integrated luminosity collected with the CMD-3 detector at the VEPP-2000 collider in the c.m. energy range 1.5 – 2.0 GeV. The measured cross section exhibits a sharp drop near the threshold. A first study of dynamics of six-pion production has been performed.

1 Introduction

Production of six pions in annihilation was studied at DM2 6pidm2 and with much larger effective integrated luminosity at BaBar isr6pi , using Initial-State Radiation (ISR) events. The DM2 experiment observed a “dip” in the cross section at about 1.9 GeV, confirmed later by the FOCUS Collaboration in the photoproduction focus ; focus1 and by the BaBar Collaboration, where this structure was also observed in the final state isr6pi . The origin of the “dip” remains unclear, but the most popular explanation suggests a presence of the under-threshold proton-antiproton () resonance. This hypothesis is supported by the fast increase of the form factor to the threshold, recently confirmed by the high-statistics BaBar study isrppbar , and discussed in many theoretical papers (see, e.g., Ref. ppbartheory ). Even earlier, a narrow structure near the proton-antiproton threshold has been also observed in the total cross section of annihilation into hadrons in the FENICE experiment fenice .

The cross section is also used in the calculation of the hadronic contribution to the muon anomalous magnetic moment g-2 . The detailed study of the production dynamics can further improve the accuracy of these calculations and help in explaining the cross section anomaly.

In this paper we report the analysis of the data sample based on 33 pb of integrated luminosity collected at the CMD-3 detector in the 1.0-2.0 GeV center-of-mass energy range. We observe only a few candidate events below 1.5 GeV. Since their number is consistent with background, we present our results for the 1.5-2.0 GeV center-of-mass energy range, corresponding to 22 pb of integrated luminosity. These data were collected in three energy scans performed at the VEPP-2000 collider vepp .

The general purpose detector CMD-3 has been described in detail elsewhere sndcmd3 . Its tracking system consists of a cylindrical drift chamber (DC) dc and double-layer multiwire proportional Z-chamber, both also used for a trigger, and both inside a thin (0.2 X) superconducting solenoid with a field of 1.3 T. The liquid xenon (LXe) barrel calorimeter with 5.4 X thickness has fine electrode structure, providing good spatial resolution lxe , and shares the cryostat vacuum volume with the superconducting solenoid. The barrel CsI crystal calorimeter with thickness of 8.1 X is placed outside the LXe calorimeter, and the end-cap BGO calorimeter with a thickness of 13.4 X is placed inside the solenoid cal . The luminosity is measured using events of Bhabha scattering at large angles lum .

2 Selection of events

Candidates for the process under study are required to have five and more charged-particle tracks with the following “good” track definition:

  • A track contains more than five hits in the DC.

  • A track momentum is larger than 40 MeV/c.

  • A minimum distance from a track to the beam axis in the transverse plane is less than 0.5 cm.

  • A minimum distance from a track to the center of the interaction region along the beam axis Z is less than 10 cm.

  • A track has a polar angle large enough to cross half of the DC radius.

The number of events with seven or more selected tracks is found to be less than 1%. Reconstructed momenta and angles of the tracks for six-track and five-track events were used for further selection.

Figure 1: (a) Scatter plot of the difference () between the total energy and c.m. energy versus total momentum for six-track events. The line shows the boundary of the applied selection; (b) Projection plot of (a) after selection. The histogram shows the normalised MC-simulated distribution; (c) Scatter plot of difference () of total energy and c.m. energy versus total momentum for five-track events; (d) Difference between the total energy of five-tracks plus missing track energy and c.m. energy (points). The histogram shows the distribution for the MC simulated background events (see text).

For six- or five-track candidates we calculate the total energy and total momentum assuming all tracks to be pions:

Figure 1(a) shows a scatter plot of the difference between the total energy and c.m. energy E=Etot–Ec.m. versus total momentum for six-track candidates. The histograms combine events from three highest energy points. A clear signal of six-pion events is seen as a cluster of dots near zero. Events with a radiative photon have non-zero total momentum and total energy which is always smaller than the nominal one. A momentum of any pion incorrectly reconstructed due to interaction with detector material or DC resolution leads to momentum-energy correlated “tails” in both directions.

We select events with total momentum less than 150 MeV/c and show the difference E in Fig. 1(b). The experimental points are in good agreement with the corresponding Monte Carlo (MC) simulated distribution shown by the histogram. We require -200E100 MeV to determine the number of six-pion events. Six-track events have practically no background: we estimate it from MC simulation of the major background processes and (one of the photon from the decay converts to a pair at the vacuum pipe), and found a contribution of less than 1%. We use this value as an estimate of the corresponding systematic uncertainty.

To determine the number of six-pion events with one missing track, a sample with five selected tracks is used. A track can be lost if it flies at small polar angles outside the efficient DC region, decays in flight, due to incorrect reconstruction, nuclear interactions or by overlapping with another track. Figure 1(c) shows a scatter plot of the difference E between the total energy and c.m. energy versus total momentum for five-track events. Six-pion candidates in the five-track sample have energy deficit correlated with the total momentum. This sample has some admixture of background events from multihadron processes mentioned above with photons from the decays. We apply an additional requirement on the “neutral” (not associated with charged tracks) energy in the calorimeter to be less than 300 MeV. This requirement reduces the background by a factor of two and removes less than 2% of signal events estimated using MC simulation.

The direction and momentum of a missing pion can be calculated assuming a six-pion final state. We add energy of a missing pion to the energy of five detected pions and show the difference E in Figure 1(d) by points. A corresponding background distribution from the MC simulation of the and events is shown in Figure 1(d) by the histogram: background events contribute less or about 10% to the signal region after applying a requirement 300 MeV.

To obtain the number of six-pion events from the five-track sample, we fit the distribution shown in Fig 1(d) with a sum of functions describing a signal peak and background. The signal line shape is taken from the MC simulation of the six-pion process and is well described by a sum of two Gaussian distributions. The photon emission by initial electrons and positrons is taken into account in the MC simulation and gives a small asymmetry observed in the distributions of Figs. 1 (b,d). We describe this asymmetry by an admixture of a third Gaussian function. All parameter ratios of the signal function are fixed except for the number of events and main Gaussian resolution. The third-order polynomial is used to describe the background distribution.

To estimate a systematic uncertainty of the background subtraction procedure, we compare the MC simulated background distribution with the experimental events with an MeV requirement, and found reasonable agreement with the histogram shown in Fig 1(d). A variation of the polynomial fit parameters for the experimental and MC simulated background distributions leads to about 3% uncertainty on the number of signal events.

We found 2887 six-track events and 5069 five-track events corresponding to the process . The numbers of six- () and five-track () events determined at each energy point are listed in Table 1.

3 First study of the production dynamics

To obtain a detection efficiency, we simulate six-pion production in a primary generator, pass simulated events through the CMD-3 detector using the GEANT4 geant4 package, and reconstruct them with the same reconstruction software as experimental data. In our experiment, the acceptance of the DC for the charged tracks is not 100%, and the detection efficiency depends on the production dynamics of six pions. The dynamics of the process was not studied previously in detail. The BaBar Collaboration isr6pi reported the observation of only one from all invariant mass combinations and no structures in any other (three-, four-pion) invariant mass combinations.

We investigate a few production mechanisms, and compare simulated angular and invariant mass distributions with those in data. All studied distributions strongly contradict to a phase space model, which assumes all pions to be completely independent. We exclude the phase space model from further consideration. In this paper we illustrate our study with three models, all with one per event. To conserve the initial state quantum numbers, six pions must have .

Figure 2: Cosines of the relative angle of two pions with opposite-sign charge (a), and of two pions with same-sign charge (b) for experimental events (dots) and MC simulation for (dotted histogram), (solid histogram) and (dashed histogram).

In the model #1 we use the following decay chain: . This model uses dominant decays and  cmd2_4pi , and naturally includes the decay to describe the process with one charged  isr6pi . We use PDG values pdg for the resonance parameters and the model allows to introduce a form factor in each decay vertex.

Another studied model (#2) was simpler: it includes the production of one and four pions in S-wave. We try two options: the four pions are distributed according to the phase space or forming a scalar resonances or .

And finally, the model (#3) assumes with a tensor resonance in the four-pion final state.

MC simulation should reproduce experimental angular distributions of the pions to obtain correct detection efficiency. Figure 2 shows (by points) the cosines of open angles between pions for opposite-sign (a) and same-sign (b) pion pairs for data.

We compare distributions of Fig. 2 with the MC simulated distributions for the model #1 (dotted histogram), model #2 (solid histogram) and model #3 (dashed histrogram), and the best agreement was found with the model #2.

Note, that variation of the resonance parameters in the models does not significantly affect these angular distributions. For example, model #2 with production of one exhibits the same angular distributions both in the case, when the remaining four pions are distributed according to phase space or form a scalar resonance ( or ).

Figure 3: (a) Polar angle distribution for six-pion events with six detected tracks for data (points) and MC simulation (histogram); (b) Polar angle distribution for six-pion events with five detected tracks for data (circles) and MC simulation (solid histogram). The polar angle distribution for a missing track is shown by squares (data) and the dashed histogram (MC simulation).
Figure 4: Experimental invariant mass distributions (from left to right) for two, three and four pions for (top to bottom) 1600, 1800 and 2000 MeV c.m. energies in comparison with simulation of one with the remaining four pions in S-wave, and distributed according to the phase space (solid histogram) or form a scalar resonance (dashed histogram).

Figure 3(a) presents the polar angle () distribution for six-pion events with all detected tracks. The requirement for a track to cross half of the DC radius effectively determines a cut on this parameter. The result of the MC simulation in model #2, presented by the histogram, well describes the observed distribution. Figure 3(b) presents the polar angle distribution for five detected tracks (circles for data, the solid histogram for the MC simulation) after background subtraction. The polar angle distribution for the missing track is shown by squares (data) and the dashed histogram (MC). With our “effective” DC acceptance we have almost two times more six-pion events with one missing track than events with all tracks detected.

We calculate invariant masses for the combinations of two, four (total charge zero), and three (total charge ) pions for the different c.m. energies and show them in Fig. 4. We compare the obtained distributions with model #2 (), and observe good agreement with experiment at c.m. energies 1600 MeV and 2000 MeV, if four pions are distributed according to phase space (solid histogram). But at the c.m. energy of 1800 MeV the experimental data are better decsribed by the same model with four pions forming . Note that invariant mass distributions for models #1 and #3 do not describe data in any mass interval, but some admixture of these channels cannot be excluded.

From the study of the mass distributions in Fig. 4 we conclude that production dynamics of six charged pions changes in the relatively narrow energy region (1700-1900 MeV). This phenomenon demands a further investigation.

4 Detection efficiency

We calculate the detection efficiency from the MC simulated events as a ratio of events after selections described in Sec. 2 to the total number of generated events. With the limited DC acceptance, incorrect simulation of the pion angular distribution leads to a systematic error in the efficiency calculation and thus in the cross section measurement.

Figure 5: (a) Ratio of events with five and six detected tracks for data (points with errors) and MC simulation for model #1 (dotted line), model #2 (solid line) and model #3 (dashed line); (b) Double ratio / versus energy. The line shows a fit with a constant.

In the five-track sample, about 15-17% of events have a missing track due to the DC reconstruction inefficiency, well reproduced by the MC simulation. The remaining events migrate from the six- to the five-track sample due to the limited DC acceptance (see Fig. 3). It makes the ratio very sensitive to the pion angular distribution, and we study it to validate the model used for the efficiency calculation.

Figure 5 (a) shows the ratio versus energy for data (points with errors) and for three models, discussed in Sec. 3. The experimental average value = 1.740.03 is in good agreement with = 1.76 for the model #2 (solid line), but inconsistent with model #1 ( = 1.92, dotted line) and model #3 ( = 1.30, dashed line). A “naive” phase space model for the six-pion production (all tracks uncorrelated) gives = 2.1.

To estimate a model-dependent systematic error, we compare the experimental number of six- and five-track events after normalisation to the MC simulated acceptance. We calculate a double ratio for each energy point for the model #2, and show it in Fig. 5 (b). The average value 0.9840.018 (=56/35) is in good agreement with the prediction of model #2 in the studied energy interval, so that a maximum systematic deviation from unity does not exceed 3.4%. However, a relatively large value can be an indication of the additional systematic uncertainty, and we conservatively take 4% as an estimate of a systematic error on the detection efficiency using as a scale factor.

Figure 6: (a) Efficiency calculated from the MC simulation for six-track events (squares) and for a sum of five- and six-track events (circles); (b) Radiative correction.

The detection efficiency thus obtained with model #2 is shown in Fig. 6(a) for events with six detected tracks (squares) and for a sum of five- and six-track events (circles), icreasing efficiency by factor 2.5. Note that if a sum of six- and five-track events () is taken for the detection efficiency calculation, the data-MC inconsistencies in the description of the DC inefficiency and (partly) in the model-dependent angular distributions are significantly reduced.

5 Cross Section Calculation

At each energy the cross section is calculated as

where is the integrated luminosity for this energy point, is the detection efficiency (Fig. 6(a)), and is the radiative correction calculated according to kur_fad and shown in Fig. 6 (b). The energy dependence of the radiative correction reflects a sharp dip in the cross section. To calculate the correction we use BaBar data isr6pi as a first approximation and then use our cross section data for iterations.

The integrated luminosity, the number of six and five-track events, detection efficiency, radiative correction and obtained cross section for each energy point are listed in Table 1.

Figure 7: The cross section measured with the CMD-3 detector at VEPP-2000 (dots). The results of the BaBar measurement isr6pi are shown by open circles. The line shows the threshold.
Ec.m. , MeV L, nb , nb
2000 474.7 88 166.014.8 0.480 0.905 1.280.09
1975 516.5 95 168.414.3 0.484 0.906 1.200.08
1950 458.8 91 124.813.2 0.488 0.913 1.090.08
1925 582.2 110 179.415.0 0.492 0.934 1.120.07
1900 495.6 104 155.113.5 0.496 0.964 1.130.07
1850 431.8 94 156.915.3 0.504 0.892 1.340.10
1800 440.1 86 168.615.0 0.513 0.883 1.330.09
1750 541.8 54 126.218.9 0.513 0.877 0.770.09
1700 486.1 38 72.510.0 0.513 0.865 0.530.06
1650 463.3 21 42.37.5 0.513 0.873 0.320.04
1600 441.9 9 10.55.5 0.513 0.900 0.0990.032
1550 521.1 9 12.14.0 0.505 0.914 0.0910.013
1500 554.6 3 5.94.1 0.497 0.921 0.0370.018
1890 521.5 95 137.413.7 0.498 0.984 0.940.07
1870 663.4 163 259.135.9 0.501 0.891 1.480.13
1825 500.8 113 179.116.5 0.509 0.885 1.340.09
1775 550.7 85 139.713.5 0.513 0.878 0.940.07
1725 523.0 70 104.611.7 0.513 0.867 0.780.06
1675 561.4 32 63.49.8 0.513 0.865 0.400.05
1625 508.5 16 32.46.1 0.513 0.888 0.220.03
1575 522.2 7 10.23.5 0.509 0.907 0.0740.011
1525 530.9 3 7.53.3 0.501 0.920 0.0450.016
1980 602.2 111 217.916.5 0.484 0.905 1.290.08
1960 680.1 117 214.616.7 0.487 0.910 1.140.07
1940 988.7 173 322.420.2 0.490 0.923 1.150.06
1920 491.5 90 171.814.0 0.493 0.934 1.200.08
1900 883.3 145 257.117.7 0.496 0.964 0.990.05
1872 845.6 193 340.020.2 0.501 0.891 1.460.07
1840 952.1 197 390.722.4 0.506 0.892 1.420.06
1800 972.1 157 332.620.6 0.513 0.883 1.150.06
1760 950.4 153 252.218.7 0.513 0.878 0.980.05
1720 797.4 95 126.515.3 0.513 0.867 0.650.05
1680 879.2 58 79.712.0 0.513 0.865 0.370.04
1600 812.7 10 32.46.5 0.513 0.900 0.1170.020
1520 825.3 2 8.93.6 0.500 0.920 0.0300.011
Table 1: Luminosity, Number of events, Detection efficiency, Rad. correction and Cross section for each c.m. energy point. Horizontal lines separate three energy scans.

6 Systematic errors

The following sources of systematic uncertainties are considered.

  • The model dependence of the acceptance is determined using the angular distributions, which are specific for each particular model. As shown in Sec. 4, a model with one and remaining pions in S-wave (phase space or ) gives good overall agreement with the observed angular distributions. Using the ratio of six- and five-track events we estimate a systematic uncertainty on the detection efficiency as 4%.

  • Since only one charged track is sufficient for a trigger (99-98% efficiency), we assume that for the multi-track events, considered in this analysis, the trigger inefficiency gives a negligible contribution to the systematic error.

  • A systematic error due to the selection criteria is studied by varying the cuts described previously and doesn’t exceed 3%.

  • The uncertainty on the determination of the integrated luminosity comes from the selection criteria of Bhabha events, radiative corrections and calibrations of DC and CsI and does not exceed 2% lum .

  • The admixture of the background events not subtracted from the six-track sample is estimated as 1%.

  • The accuracy of background subtraction for five-track events is studied by the variation of functions used for a background description in Fig. 1(d) and is estimated as 3%.

  • A possible uncertainty on the beam energy is studied using the momentum distribution of Bhabha events and total energy of four-pion events. The uncertainty at the level of is not excluded and because of the cross section variation it can result in a 1% change of the cross section.

  • A radiative correction uncertainty is estimated as about 1% mainly due to the uncertainty on the maximum allowed energy of the emitted photon, as well as from the uncertainty on the cross section.

The above systematic uncertainties summed in quadrature give an overall systematic error of about 6%.

The obtained cross section is in overall agreement with the results of the most precise measurement performed by the BaBar Collaboration isr6pi shown in Fig. 7 by open circles.


The total cross section of the process has been measured using 22 pb of integrated luminosity collected by the CMD-3 detector at the VEPP-2000 collider in the 1.5-2.0 GeV c.m. energy range. The five- and six-track events are used to estimate the model-dependent uncertainty in the acceptance calculation. From our study we can conclude that the observed production mechanism can be described by the production of one with four remaining pions in S-wave and distributed according to phase space. We also observe that the production dynamics changes in the 1700-1900 MeV range and demands further investigation. A detailed analysis of the production dynamics will be performed in the combined analysis of the processes and .

The measured cross section is in good agreement with all previous experiments in the energy range studied, and exhibits a sharp dip near the threshold.


The authors are grateful to A.I. Milstein and Z.K. Silagadze for their help with a theoretical interpretation and development of the models. We thank the VEPP-2000 team for excellent machine operation.

This work is supported in part by the Russian Education and Science Ministry, by FEDERAL TARGET PROGRAM ”Scientific and scientific-pedagogical personnel of innovative Russia in 2009-2013”, by agreement 14.B37.21.07777, by the Russian Fund for Basic Research grants RFBR 10-02-00695-a, RFBR 10-02-00253-a, RFBR 11-02-00328-a, RFBR 11-02-00112-a, RFBR 12-02-31501-a, RFBR 12-02-31499-a, RFBR 12-02-31498-a, and RFBR 12-02-01032-a.


  • (1) R. Baldini et al., reported at the “Fenice” Workshop, Frascati (1988). A. B. Clegg and A. Donnachie, Z. Phys. C45, 677 (1990). M.R. Whalley, J. Phys. G29, A1 (2003).
  • (2) B. Aubert et al. (BaBar Collaboration), Phys. Rev. D73, 052003 (2006).
  • (3) P.L. Frabetti et al. (FOCUS Collaboration), Phys. Lett. B514, 240 (2001).
  • (4) P.L. Frabetti et al. (FOCUS Collaboration), Phys. Lett. B578, 290 (2004).
  • (5) B. Aubert et al. (BaBar Collaboration), Phys. Rev. D73, 012005 (2006).
  • (6) A. Sibirtsev and J. Haidenbauer, Phys. Rev. D71, 054010 (2005).
  • (7) A. Antonelli et al. (FENICE Collaboration), Phys. Lett. B365, 427 (1996).
  • (8) M. Davier, S. Eidelman, A. Höcker, Z. Zhang, Eur. Phys. J. C31, 503 (2003).
  • (9) V.V. Danilov et al., Proceedings EPAC96, Barcelona, p.1593, (1996). I.A.Koop, Nucl. Phys. B (Proc. Suppl.), 181-182, 371 (2008).
  • (10) B.I.Khazin, Nucl. Phys. B (Proc. Suppl.), 181-182, 376 (2008).
  • (11) F. Grancagnolo et al., Nucl. Instr. and Meth. A, 623, 114 (2010).
  • (12) A.V.Anisyonkov et al., Nucl. Instr. and Meth. A, 598, 266 (2009).
  • (13) D. Epifanov (CMD-3 Collaboration), J. Phys. Conf. Ser. 293, 012009, (2011).
  • (14) R.R. Akhmetshin et al.,Nucl. Phys. B (Proc. Suppl.), 225-227, 69 (2012).
  • (15) S. Agostinelli et al. (GEANT4 Collaboration), Nucl. Instr. and Meth. A, 506, 250 (2003).
  • (16) R.R. Akhmetshin et al. Preprint BudkerINP 98-83, Novosibirsk 1998; Phys. Lett. B466, 392 (1999).
  • (17) J. Beringer et al. (Particle Data Group), Phys. Rev. D86, 1 (2012).
  • (18) E.A. Kuraev and V.S. Fadin, Sov. J. Nucl. Phys. 41, 466 (1985).
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description