Measurement of the open beauty and charm production cross sections in two photon collisions with DELPHI

SUPPLEMENTS ELSEVIER

Nuclear Physics B (Proc. Suppl.) 126 (2004) 185-190

Measurement of the Open Beauty and Charm Production in Two Photon Collisions with DELPHI W. Da Silva”*

“LPNHE

Cross Sections

t

CNRS/IN2P3

Universit&

P6 and P7, Paris, France

Heavy b and c quark production in yy collisions has been measured through semileptonic DELPHI detector at LEP. The 413 pb-lof data were collected at center-of-mass energies from GeV. The corresponding extracted cross sections a(e+e-+ e+e-bbX) and a(e+e-+ efe-cZX) to NLO perturbative QCD calculations. The cross section for b production is found to exceed by a factor of about 3. K-lepton double tagging, used for the first time in yy physics, confirms

decays with the 189 GeV to 209 are

compared

QCD predictions this excess.

1. Introduction

Heavy quark production has already been measured in many experiments [l] where an excess of rates, as compared to QCD predictions, has been observed for the production of bb pairs. This is indeed a challenging problem for perturbative QCD since the latter wa.s expected to be fully predictive thanks to the larger quark mass involved [2], and since next-to-leading order corrections do not appear to sufficiently reduce the observed discrepancy that remains at the level of a factor of about 2 to 3, at least at HERA or LEP. At LEP, the L3 and OPAL collaborations have already reported on b& pair production from yy collisions [l]. We here present a new analysis of this process, performed at the DELPHI detector at LEP. The main mechanisms for heavy quark production in yy collisions are expected to be the direct and the resolved processes depicted in fig. 1, which are of the same order at LEP energies. b and c quarks are selected through the muons coming from their semileptonic decays. The distribution of transverse momenta with respect to the closest jet of the muons provides a clean way of tagging. As a premihre in two-photon physics, in our analysis we also made use of the K-lepton charge correlation,

with

kaons

being

identified

with

the RICH

detector. This allowed us to enhance the b signal and to strengthen the fact that the observed ex*On behalf of the DELPHI tN. Arteaga, C. Carimalo thanked.

Collaboration. and F. Kapusta

0920-5632&T - see front matter doi:10.1016/j.nuc1pl~ysbps.2003.11.031

0 2004 Published

b, c

Figure 1. Direct and single resolved processes

cess really comes from b quark production anisms. 2.

Monte

Carlo

samples

Data

used were

collected

with

the DELPHI

de-

tector during the period 1998-2000, with LEP center-of-mass energies from 189 GeV to 209 GeV. The luminosity sample after some run quality selection is about 463 pb-’ . The luminosity weighted average e+e- center-of-mass energy is approximately

6

= 198 GeV.

PYTHIA6.143

[6] was used as the reference Monte-Carlo generator for describing two-photon events at Leading Order. Light quarks (u&) were generwithout

mass,

using

the

full

default

y*y*

PYTHIA machinery. Masses of heavy quarks (b,c) were taken into account only for the di-

warmly

by Elsevier

and

The DELPHI detector and its performances have been described in detail elsewhere [3,4].

ated are

Data

mech-

t3.V

W! Du Silva/Nuclear

186

Physics

B (Proc.

rect processes. Dedicated samples for b and c quarks were produced for the direct and the single resolved components according to each year energies and integrated luminosity. The single resolved cross section was taken to be equal to the direct one [7]. Samples of simulated background to antitagged yy processes were used : e+e- -+ r+r-, e+e- 4 e+e-r+t,-, e+e- + Zy 7 e+e- -+ W+W-, e+e- -+ ZZ and a DISey component. All these processes were generated respectively according to BDKRC [8], KORALZ [9], WPHACT [lo] and PYTHIA6.159 [6]. 3. Event

Suppl.)

I26

(2004)

185-190

requirement [5]. As a consequence the trigger efficiency was 98 %. Anti-tagging selection required no neutral with energy greater than 40 GeV even from the STIC detector at very small angle. A total of 56416 events satisfied these cuts. The invariant mass distribution of fig. 2 shows the contributions of the various physical processes.

+ DELPHI

selection

data

The strategy for open beauty event selection proceeded in three steps : selection of antitagged two-photon events, muon selection and jet reconstruction. 3.1. Anti-tag events Charged particles used for event selection were required to satisfy the following criteria : momentum greater than 0.2 GeV/c, polar angle f3 with respect to the beam axis between loo and 170°, measured track length greater than 30 cm, radial projection of the impact parameter relative to the interaction point less than 3 cm, projection of the impact parameter along the beam direction less than 7 cm, relative error on the energy measurement less than 100 %. All calorimetric informations were used. The minimum momentum required for the neutral depends on the detector used. A neutral was accepted if measured by HCAL (momentum greater than 1 GeV/c ), HPC (energy greater than 250 MeV) or FEMC (energy greater than 250 MeV). All these selected particles entered the event discriminating quantities used to select two-photon candidates. The number of charged tracks had to be greater than or equal to 6 (this cut reduced strongly the background coming from r decay), total energy of charged tracks and neutral below 70 GeV, visible invariant mass greater than 3 GeV/c2 and lower than 35 GeV/c2. A track having a transverse momentum with respect to the beam axis greater than 1.2 GeV/c and 1COST] lower than 0.8 was required in order to satisfy the trigger condition

0

20

40

._

60

80

100 120 146 160 180 264

wyisible ^^^.^.-.~._.

GeV

1 ^.

Figure 2. Visible invariant mass distribution

3.2. Muon Identification To identify a charged particle with momentum greater than 2 GeV/c as a muon candidate, its track was extrapolated to each of the layers of the muon chambers, taking into account multiple scattering in the material and the propagation of track reconstruction errors. A fit was then made between the track extrapolation and the position and direction of the hits in the muon chambers. Ambiguities with muon chamber hits associated to more than one extrapolated track were resolved by selecting the track with the best

n! Da SilualNuclear

Physics B (Proc. Suppl.) 126 (2004) 155-190

fit. The charged particle was then identified as a muon if the fit was sufficiently good and if hits were found outside the return iron yoke [ll]. In order to remove muons coming from annihilation processes, muon candidates with momentum greater than 20 GeV/c were discarded. To exclude regions with poor geometrical acceptance, charged particles were accepted if their polar angle 0 satisfied either 0.03 < 1cos8) < 0.57 or 0.71 < 1cos81 < 0.94, which defines the barrel region and the forward region, respectively. For isolated muons, the tagging efficiency was found to be 0.82iO.10 [ll]. 3.3. Jet reconstruction and transverse momentum of muons The sensitive variable for beauty events chosen in this analysis is the transverse momentum of the lepton with respect to the direction of the jet to which it belongs, hereafter denoted by PT. The jet algorithm is defined in the subroutine LUCLUS inside the LUND Monte Carlo program [la], with default parameter Djoin equal to 2.5 GeV. The jet analysis was performed on charged tracks and neutral particles. The latter should have an energy greater than 800 MeV in order to suppress spurious fluctuations in the jet direction. The muon candidate was included in the jet finding algorithm, but the lepton momentum was not included in the determination of the jet direction. The transverse energy of the jet with respect to the beam axis, computed in summing all transverse energy particles in the jet excluding the muon, must be greater than 500 MeV and lower than 8 GeV. The lower cut reduces background from non-perturbative two-photon interactions and the upper cut reduces strongly the background from annihilation processes. After this final cut 651 events remained. 4. Extraction

of the b cross section

The pT-distribution of candidate muons is shown in fig.3 The corresponding distributions from b and c quark have well separated shapes. The visible b contribution was fitted by minimizing the binned x2 :

x2= c (Nil- %J2/(% +$3

(1)

!87

* DELPHI

0

1

2

3

PT of muon

4

candidate

data

5

6 GeVk

7:

1

Figure 3. Inclusive muon transverse momentum

with nP = abni +a,+~ +a,dsn~ds +nF”gd where NV is the number of muon candidates in a bin and np is the sum of the various contributions in that bin. The quantities nb,, n;, nEds and nfkgd are respectively the number of expected beauty, charm, light quarks and background events from the Monte-Carlo normalized to the luminosity of the sample. nr is the variance of the expected events and ob, a,, &da are the corresponding renormalization coefficients. The coefficient QI, was fixed so as to reproduce the LEP average total cross section for open charm production a(e+e- -+ e+e-c?X) = 984 f 128 pb [13-151. This procedure reduces the impact of the charm cross section on the beauty cross section measurement, and limits the charm cross section propagation error. The coefficient a,& was determined following the above procedure, selecting hadrons (7~, K, p) instead of muons, using the powerful &Y/da; particle identification [16]. The uds component is dominant (79 f 0.7 (stat) %) as shown in fig. 4. Q&s is found to be 1.13 f O.Ol(stat).

188

K Du Silva/Nucleur

Physics

B (Proc.

Suppl.)

126 (2004)

185-190

%.

b systematics : source of uncertainty Event selection : Nch, p$ack,Wv+, pp ,cos 8,, Data/MC discrepancy Muon : efficiencv, misidentification ”

I

O?LEP>

Jet reconstruction : leading particle discrepancy %ds : stat, Data/MC on Nch, p~dron, Wvis Br(b + p) model Ratio direct/resolved 1:2 to 2:l Total

hbb

(%)

16.5 9.8 8.3 7.8 3.5 3.4 3.1 23.0

Table 1

Figure 4. Inclusive 7r, K, p transverse momentum and visible invariant mass

Using another variable like the visible invariant mass gives a similar result. The obtained value of o&s was taken as an input for the ob fit which led to Ndbata = 118 & 26(stat) events and to the corresponding fractions of the various components : fb = 18 f 1 (stat) %, fc “_ 41 %, f& fi 39% and fbc,+gd _” 2%. The total b?i event selection efficiency was found to be about 1.7 % . The corresponding x2/Nd0, is 1.3. This results in ae+e--te+e-bbX

= 14.9 f 3.3(stat)pb

(2)

The following sources of systematics displayed in table 1 were taken into account : trigger effects, selection cut variables and their discrepancy between data and Monte Carlo, muon efficiency and an average misidentification of 12.5 % [ll], the charm cross section and the jet algorithm using the leading particle. Systematical and statistical errors on Q&s were estimated using the discrepancy between data and Monte-Carlo selection variables distributions, leading to an error of 4.8

It should be noted that the systematical and statistical errors on ozlrks resealing were taken into account but these errors are correlated to muon misidentification. The error of 3.4 % on the b -+ p branching ratio was taken from [ll]. In order to take into account the uncertainties of the parametrization of the photon partonic density, the ratio of the direct to resolved cross sections was taken either as 0.5 or 2. Each time the fit was redone, half of the difference between the results was used as systematic uncertainty. All these errors were quadratically summed giving the result ge+e---teSe-b6X = 14.9 f 3.3(stat) f 3.4(syst)pb 5. K-lepton

correlations

The use of charge correlation between kaon and lepton arising from the semileptonic decay of heavy quarks shown in fig. 5 is a useful tool to increase the purity of b and c events. Extracting cross sections in particular provides a check on the internal consistency of the measurement in case of b quark production. The muon pTdistribution was again used as described in Sec. 4. In addition, a kaon had to be identified in the muon jet using the specific ionisation dEJdx measured in the TPC detector and the Cherenkov radiation detected in the barrel and forward RICH

(3)

W Da Siluu/Nucleur

Physics B (Pvoc. Suppl.) I26 (2004) 185-190

I h

189

DELPHI preliminary K’ L- correlatiai3

. DELPHI data

Figure 5. K lepton correlation duction

for b and c pro-

[16]. The informations from liquid and gas radiators are combined with the dE/dx to yield a probability criterium [16]. After these criteria selection we obtained 171 events in the K*!r sample and 149 in the K*tl* one. In order to obtain the charm cross section from the K&CT sample, the procedure described in Sec. 4 was applied, except that o!b was fixed to its previously obtained value while oc was fitted. The range of low transverse momenta of muons (pi < 400 MeV) was excluded from the fit because data are not well reproduced by the Monte Carlo generator in this domain, as can be seen from fig. 6. The corresponding fractions of the various components were found to be : ftids N 35%, fc N 48 %, fb N 16% and fb&,j ” 1%. The corresponding x2/iVdo, is 0.37. The result is

u e+e-=&e-cEX

= 937 f 19l(stat)

f 206(syst)pb

c systematics : source of uncertainty Event selection Kaon: efficiency, fragmentation Jet reconstruction Ratio direct /resolved %ds

Muon : efficiency,misidentification Obi;

JWc -+ P) Total Table 2

Au@

(4)

(%)

12.5

I

Figure 6. Inclusive muon transverse momentum with K-lepton tagging

11.0 10.9 5.6 4.2 4.0 3.5 3.2 22.0

Several sources of systematic errors come in addition with those already described in Sec. 4 : kaon selection efficiency, beauty cross section and b + c branching ratio [ll]. All these errors are displayed in table 2. Using the charge correlation between K* and !*, the fractions of the various components were found to be : f& -” 30%, fc 21 38 %, fb 21 30% and fb&gd ” 2%. The corresponding x2/Ndop is 0.29. The beauty cross section is then u e+e---f e+e-bb

= 11.4 & 4.5 pb

which is quite compatible

with result (3).

6. Conclusions The measurement of the cross section of beauty and charm quarks in two photon collisions at LEP2 average center of mass energy of 198 GeV has been performed. The results are respectively #+e-+e+e-bbX = 14.9 f

190

K I3u SilvalNuclear

Physics

B (Proc.

33(stat) f 3.4(syst)pb and ae’e-+e+e-cEX = 937f lSl(stat) f 206(syst)pb. These values agree with L3 an OPAL measurements. All the published and preliminary results on the total e+ecross section for b and c production in yy collisions are shown in fig 7. They are compared to

Strppl.)

2. 3. 4. 5. 6.

7. 8.

9. 10. 11. 12. Figure 7. World results 13. the Drees, Kramer, Zunft and Zerwas NLO QCD calculation [7]. For the charm production a significant gluon content in the photon is required. The DELPHI extracted b cross section is in excess of about 2 standard deviations over the predicted value.

REFERENCES 1. M. Acciari et al. (L3 Collab.), Phys. Lett. B 453 (1999) 83. M. Acciari et al. (L3 Collab.), Phys. Lett. B 467 (1999) 137. A. Bohrer and 0. Krasel (ALEPH Collab.) ALEPH-2000-031. A. Csilling (OPAL Collab.), hep-ex/0010060. M. Acciari et al. (L3 Collab.), Phys. Lett. B 514 (2001) 19. M. Acciari et al. (L3 Collab.), Phys. Lett. B

14. 15. 16.

126 (2004)

185-190

503 (2001) 10. P. Achard et al. (L3 Collab.), Phys. Lett. B 535 (2002) 59. S. Frixione et al., J. Phys. G 26 (2000) 723. P. Aarnio et al. (DELPHI Collab.), Nucl. Inst. Meth. A303 (1991) 233. P. Abreu et al. (DELPHI Collab.), Nucl. Inst. Meth. A378 (1996) 57. V. Canale et al. DELPHI/99-7 DAS 188. T. Sjijstrand, Comp. Phys. Comm. 82(1994) 74. C. Friberg and T. Sjdstrand, Eur. Phys. J. C13(2000) 151 (hep-ph/9907245). M. Drees, M. Kramer, J. Zunft and P.M. Zerwas, Phys. Lett. B 306 (1993) 371. F.A Berends, P.H. Daverveldt, R. Kleiss, Comp. Phys. Comm 4086271, Comp. Phys. Comm 4086285, Comp. Phys. Comm 40863009. S. Jadach and Z. Was, Comp. Phys. Comm 7994503. E.Accomando, A. Ballestrero and E. Maina, hep-ex/0204052. P. Abreu et al. (DELPHI Collab.) Eur. Phys. J.C20 (2001) 455. T. Sjiistrand, Comp. Phys. Comm. 39 347 T. Sjijstrand, PYTHIA 5.6 and JETSET 7.3, CERN-TH/6488-92. OPAL Coll., A. Csilling, Proc of PHOTON2000, AIP conf. proceeding, v. 571 (2000) 276. L3 Coll., S. Saremi, Proc of PHOTON2000, AIP conf. proceeding, v. 571 (2000) 266. DELPHI Collab., A. Sokolov, Proc of PHOTON2001, (2001) 109. M. Battaglia, P.M. Kluit, DELPHI/96-133 W. Adam et al. , DELPHI/94-112