Total forward and differential cross sections of neutral D mesons produced in 500 GeVrc p y nucleon interactions

Transcription

1 9 September 1999 Physics Letters B otal forward and differential cross sections of neutral D mesons produced in 5 GeVrc p y nucleon interactions E791 Collaboration EM Aitala i, S Amato a, JC Anjos a, JA Appel e, D Ashery n, S Banerjee e, I Bediaga a, G Blaylock h, SB Bracker o, PR Burchat m, RA Burnstein f, Carter e, HS Carvalho a, NK Copty l, LM Cremaldi i, C Darling r, K Denisenko e, S Devmal c, A ernandez k, G ox l, P Gagnon b, C Gobel a, K Gounder i, AM Halling e, G Herrera d, G Hurvits n, C James e, PA Kasper f, S Kwan e, DC Langs l, J Leslie b, B Lundberg e, J Magnin a, S Mayal-Beck n, B Meadows c, JR de Mello Neto a, D Mihalcea g, RH Milburn p, JM de Miranda a, A Napier p, A Nguyen g, AB d Oliveira c,k, K O Shaughnessy b, KC Peng f, LP Perera c, MV Purohit l, B Quinn i, S Radeztsky q, A Rafatian i, NW Reay g, JJ Reidy i, AC dos Reis a, HA Rubin f, DA Sanders i, AKS Santha c, AS Santoro a, AJ Schwartz c, M Sheaff d,q, RA Sidwell g, AJ Slaughter r, MD Sokoloff c, J Solano a, NR Stanton g, RJ Stefanski e, K Stenson q, DJ Summers i, S akach r, K horne e, AK ripathi g, S Watanabe q, R Weiss-Babai n, J Wiener j, N Witchey g, E Wolin r, SM Yang g, D Yi i, S Yoshida g, R Zaliznyak m, C Zhang g a Centro Brasileiro de Pesquisas ısicas, Rio de Janeiro, Brazil b UniÕersity of California, Santa Cruz, CA 9564, USA c UniÕersity of Cincinnati, Cincinnati, OH 451, USA d CINVESAV, 7 Mexico City, D Mexico e ermilab, BataÕia, IL 651, USA f Illinois Institute of echnology, Chicago, IL 6616, USA g Kansas State UniÕersity, Manhattan, KS 6656, USA h UniÕersity of Massachusetts, Amherst, MA 13, USA i UniÕersity of Mississippi Oxford, UniÕersity, MS 38677, USA j Princeton UniÕersity, Princeton, NJ 8544, USA k UniÕersidad Autonoma de Puebla, Mexico l UniÕersity of South Carolina, Columbia, SC 98, USA m Stanford UniÕersity, Stanford, CA 9435, USA r99r$ - see front matter q 1999 Published by Elsevier Science BV All rights reserved PII: S

2 6 ( ) EM Aitala et alrphysics Letters B n el AÕiÕ UniÕersity, el AÕiÕ 69978, Israel o Box 19, Enderby, British Columbia VE 1V, Canada p ufts UniÕersity, Medford, MA 155, USA q UniÕersity of Wisconsin, Madison, WI 5376, USA r Yale UniÕersity, New HaÕen, C 6511, USA Received 9 June 1999; accepted 6 July 1999 Editor: L Montanet Abstract We measure the neutral D total forward cross section and the differential cross sections as functions of eynman-x x and transverse momentum squared for 5 GeVrc p y nucleon interactions he results are obtained from 88 99"46 reconstructed neutral D mesons from ermilab experiment E791 using the decay channels D K y p q and D y q y K p p p q and charge conjugates We extract fit parameters from the differential cross sections and provide the first direct measurement of the turnover point in the x distribution, 131" 38 We measure an absolute D qd q18 x ) cross section of 154 mbarnrnucleon assuming a linear A dependence y3 he differential and total forward cross sections are compared to theoretical predictions and to results of previous experiments q 1999 Published by Elsevier Science BV All rights reserved PACS: 1387Ce; 144Lb; 136Le; 58Hp Charm hadroproduction is a convolution of short range processes that can be calculated in perturbative quantum chromodynamics QCD and long range processes that cannot be treated perturbatively and thus must be modeled using experimental measurements he large theoretical uncertainties from both contributions are reflected in the relatively large number of input parameters that can be adjusted when comparing models to the results of experiments A single measurement, no matter how precise, cannot unambiguously determine these parameters However, the results of high statistics measurements like the ones reported here, when combined with other measurements of similar precision, can constrain such parameters as the charm quark mass, the intrinsic transverse momentum of the partons in the incoming hadrons, and the effective factorization and renormalization scales used in theoretical calculations We report here measurements of the differential cross sections versus the kinematic variables eynman-x x and transverse momentum squared p, as well as the total forward cross section for the hadroproduction of neutral D mesons he relatively high pion beam momentum, 5 GeVrc, coupled with the good geometric acceptance of the agged Photon Laboratory PL spectrometer, allows us to investigate a wide kinematic region that includes points at negative x We are able to measure the shape of the differential cross section versus x with sufficient precision to confirm, for the first time, that the turnover in the cross section does occur at x ), as expected for incident pions wx 1 Combining data from two D decay modes, D Kp and D Kppp, 1 we extract a sample of 88 99" "43 at x ) fully reconstructed charm decays to use for these measurements In addition to the greater statistical significance, the use of two modes provides a means to better understand the systematic errors associated with the reconstruction of the decay products of these fully-charged decays he data were accumulated during the 1991r199 ermilab fixed-target run of experiment E791 w,3 x he experiment utilized the spectrometer built by the previous PL experiments, E516 wx 4, E691 wx 5, and 1 Charge conjugates are always implied We use D to repre- sent the sum of D and D Similarly, Kp Kppp includes y q y q y K p K p p p q and charge conjugate

3 ( ) EM Aitala et alrphysics Letters B E769 wx 6, with significant improvements he experiment employed a 5 GeVrc p y beam tracked by eight planes of proportional wire chambers PWC s and six planes of silicon microstrip detectors SMD s he beam impinged on one 5 mm thick platinum foil 16 cm in diameter followed by four 156 mm thick diamond foils 14 cm in diameter, each foil center separated from the next by an average of 153 cm, allowing most charm particles to decay in air he downstream spectrometer consisted of 17 planes of SMD s for vertexing and tracking along with 35 planes of drift chambers, PWC planes, and analysis magnets bending in the same direction for track and momentum measurement wo multicell threshold Cerenkov ˇ counters, an electromagnetic calorimeter, a hadronic calorimeter, and a wall of scintillation counters for muon detection provided particle identification he trigger was generated using signals from scintillation counters as well as the electromagnetic and hadronic calorimeters he beam scintillation counters included a beam counter 13 cm in diameter 14 cm upstream of the first target and a large beam-halo veto counter with a 1 cm hole 8 cm upstream of the first target he interaction counter was located cm downstream of the last target and 6 cm upstream of the first SMD plane he first-level trigger required a signal corresponding to at least 1r of that expected for a minimum ionizing particle MIP in the beam counter, no signal greater than 1r of a MIP in the beam halo counter, and a signal corresponding to greater than ; 45 MIP s in the interaction counter consistent with a hadronic interaction in one of the targets he second-level trigger required more than 3 GeV of transverse energy in the calorimeters Additional requirements eliminated events with multiple beam particles A fast data acquisition system wx 7 collected data at rates up to 3 Mbyters with 5 msrevent deadtime Over =1 1 events were written to 4 8mm magnetic tapes during a six-month period he raw data were reconstructed and filtered w3,8x to keep events with at least two separated vertices, consistent with a primary interaction and a charm particle decay ollowing the event reconstruction and filtering, selection criteria for the D candidates were determined by maximizing Sr' S q B where S is the normalized number of signal events resulting from a Monte Carlo simulation and B is the number of background events appearing in the data sidebands of the reconstructed Kp or Kppp mass distribution Only selection variables that are well modeled by the Monte Carlo simulation were used he final selection criteria varied by decay type Kp and Kppp and by x region he full range of a cut variation is given in the descriptions below o eliminate generic hadronic interaction backgrounds as well as secondary interactions, the secondary vertex was required to be longitudinally separated from the primary vertex by more than 8-11 times the measurement uncertainty on the longitudinal separation ; 4 mm and to lie outside of the target foils Backgrounds from the primary interaction were also reduced by requiring that the candidate decay tracks miss the primary vertex by at least -4 mm o ensure a correctly reconstructed charm particle and primary vertex, the momentum vector of the D candidate was required to point back to within 35-6 mm of the primary vertex and to have a momentum component perpendicular to the line connecting the primary and secondary vertices of less than MeVrc inally, the sum of the squares of the transverse momenta of the decay tracks relative to the candidate D momentum vector was required to be greater than 4 GeVrc 15 GeVrc for the Kp Kppp candidates to favor the decay of a high-mass particle All primary vertices were required to occur in the diamond targets hus, our results come from a light, isoscalar target he Čerenkov information is not used in this analysis; all particle-identification combinations are tried he inclusive Kp and Kppp signals are shown in ig 1 he reconstructed data were split into bins of x, integrating over all p, and bins of p, integrating over x ) o combine data from varied conditions eg, using particles that pass through one and two magnets, the normalized mass m n is constructed for each candidate using its calculated mass and error m and s m and the measured mean my m mass m : m ' D D n Using the binned maximum s m likelihood method, the normalized mass distributions were fit to a simple Gaussian for the signal and linear or quadratic polynomials for the background he acceptance can be factorized into trigger efficiency e and reconstruction efficiency e trig rec

4 8 ( ) EM Aitala et alrphysics Letters B ig 1 he Kp and Kppp signals for data with y15-x -8 Most of the trigger inefficiency is due to vetoes on multiple beam particles hese resulted in a 73" 11"4 % trigger efficiency, where the first error is statistical and the second error is systematic he interaction and transverse energy requirements were greater than 99% efficient for reconstructable hadronic charm decays Writing and reading the data tapes was 976" 1 % efficient Combining these efficiencies gives e s 683" 44 trig %, where the error is dominated by the systematic error he reconstruction efficiency is obtained from a Monte Carlo simulation he Monte Carlo simulation used PYHIArJESE wx 9 as a physics generator and models the effects of resolution, geometry, magnetic fields, and detector efficiencies as well as all analysis cuts he efficiencies were separately modeled for five evenly spaced temporal periods during the experiment his was motivated by a highly inefficient region of slowly increasing size in the center of the drift chambers caused by the MHz pion beam he Monte Carlo events were weighted to match the observed data distributions of x, p, and the summed p of all two-magnet charged tracks in the event other than those from the candidate D meson he resulting reconstruction efficiencies as a function of x and p are shown in ig rom the number of reconstructed D candidates, the reconstruction efficiency, the trigger efficiency, ig D Kp and D Kppp reconstruction efficiencies versus x left and p right

5 ( ) EM Aitala et alrphysics Letters B and the PDG branching fractions w1 x, we obtain the number of D mesons produced in our experiment during the experiment lifetime, N prod he cross sec- tion as a function of each variable z where z s x or p, is: N D ; z y s p N D X; z s 1 N y prod N is the number of nucleons per area in the target calculated from the target thickness and is 14 y6 " 4 = 1 nucleonsrmb N y p is the number of incident p y particles during the experiment lifetime his is obtained directly from a scaler which counted clean beam particles ) 1r MIP signals in the beam and interaction counters and no signal greater than 1r MIP in the beam-halo veto counter N p during the experiment lifetime hese are the only beam particles which could cause a first-level trig- ger Using Eq 1, we obtained the D qd differential cross sections versus x and p shown in ig 3 and 4 he systematic errors are divided into two categories and incorporated in two stages he uncorrelated systematic errors are determined individually for the Kp and Kppp results hese systematic errors include uncertainties in the Monte Carlo modeling of the selection criteria, the background functions, and the widths used in the Gaussian signal functions he correlated systematic errors are calculated for the combined D result, obtained from adding the Kp and Kppp samples together, weighted by the inverse-square of the combined ig 3 its to the D qd x differential cross section with the functions given in Eq dashed and 3 dotted Error bars do not include ay11 q1 % normalization uncertainty

6 3 ( ) EM Aitala et alrphysics Letters B ig 4 its to the D qd p differential cross section with the functions given in Eq 4 dashed, top, 5 dashed, bottom, and 6 dotted, top and bottom he top plot shows the range -p -4 GeVrc while the bottom plot shows the full range, -p -18 GeVrc Error bars do not include a q11 % normalization uncertainty y15 statistical and uncorrelated systematic errors he correlated errors are associated with uncertainties in the D lifetime, the Monte Carlo production model, the Monte Carlo weighting procedure, and the run period weighting procedure inally, we compare our measured Kp to Kppp branching ratio to the PDG w1x value to estimate the residual tracking and vertexing efficiency modeling error In addition to these errors, which can affect both the normalization and the shape of the differential cross sections, there are two errors which affect only the normalization: the uncertainties in the trigger efficiency and target thickness or the differential cross sections shown in ig 3 and 4, the systematic errors are factorized into shape and normalization parts he error bars in the figures show the sum, in quadrature, of the statistical and all systematic errors after factoring out the normalization component Although the relative importance varies bin-by-bin, the most important systematic errors generally come from uncertainties in the signal width and the Monte Carlo efficiency modeling In all cases, the systematic error dominates or the total forward cross section, all of the errors are summarized and summed in able 1 In the past, x distributions have been fit with ds n sa 1y< x < dx

7 ( ) EM Aitala et alrphysics Letters B able 1 Sources and values of the uncertainties on the total forward D qd cross section measurement he total systematic error comes from the quadratic sum of the D systematic errors Although some of the systematic errors are subdivided, each D systematic error is obtained directly including correlations and is not the quadratic sum of the components Error type Error % Kp Kppp D q47 q73 q47 Statistical & uncorrelated systematic errors y9 y11 y76 Statistics "1 "8 q1 q1 Selection criteria efficiency modeling y y q q MC background function y55 y6 q4 q39 D signal width in fits y41 y44 PDG w1x branching ratio "3 "53 q69 Correlated systematic errors y99 q6 D lifetime y5 q Monte Carlo kinematic weighting y3 q59 Monte Carlo production model y5 1 q1 ime dependent efficiency modeling y racking and vertex finding "46 rigger efficiency "64 arget thickness "3 q115 otal y148 itting Eq in the range 5-x -5, we find ns461"19, as shown in ig 3 his function does not provide a complete representation of our data Although the x rdof is small 3, the value of n is quite dependent on the range fitted and on the errors on the data points Another function, which can be extended into the negative x region, is an extension of Eq which uses a power-law function in the tail region and a Gaussian in the central region; that is, ds dx n < < X A 1y x yx, < x yx < c c )x b s~ X 1 A exp y x yx rs, < x yx < -x c c b 3 Requiring continuous functions and derivatives allows us to write Eq 3 with one normalization parameter and three shape parameters: n X gives the shape in the tail region, xc is the turnover point, and x b is the boundary between the Gaussian and power- law function he fit parameters from this function are nearly independent of the fit range itting our data in the range y15-x -5 gives n X s 468"1, xc s131"38, and x b s6 "13 with a x rdofs4, as shown in ig 3 his is the first measurement of the turnover point xc in the charm sector he fact that it is signifi- cantly greater than zero is consistent with a harder gluon distribution in the beam pions than in the target nucleons he functions which have been used in the past to fit the p distribution are: ds ybp sae 4 dp at low p p - 4 GeVrc for this analysis, ds X yb p sae 5 dp at high p p ) 1 GeVrc for this analysis, and b ds A s 6 dp a m qp c over all p with m set to 15 GeVrc w11 x c he results of fitting these equations to the data are

8 3 ( ) EM Aitala et alrphysics Letters B shown in ig 4 or the ranges given above, the fit results are: Ø bs83" with x rdof s 8, Ø b X s41"3 with x rdof s 17, and Ø as36"3 GeVrc y and bs594"39 with x rdof s 3 Eq 4 does not provide a good fit even over the very limited range to which it is applied While the x rdof 17 of the fit to Eq 5 is not good, it appears to be a reasonable fit to the data Eq 6 provides a very good fit to the data over the entire range of p Unfortunately, using two free parameters in addition to the normalization makes it more difficult to compare to other experiments and theory since the parameters in this fit are highly correlated his is reflected in the large 7-1% errors on a X and b compared to the error on b 1%, as shown above igs 5 and 6 show a comparison of our x and p distributions to theoretical predictions for charm quark and D meson production Although the data come from D mesons, the theoretical predictions for charm quark production are included for completeness he theoretical curves are normalized to obtain the best fit lowest x rdof to our data he theoretical predictions come from a next-to-leading order NLO calculation by Mangano, Nason, and Ridolfi MNR w1x and the PYHIArJESE w9x event generator he MNR NLO charm quark calculation uses SMRS w13x HMRSB w14x NLO parton distri- ig 5 he D qd x differential cross section compared to various theoretical predictions described in the text he curves are normalized to obtain the best fit to the data in each case Error bars do not include a q1 y11% normalization uncertainty

9 ( ) EM Aitala et alrphysics Letters B ig 6 he D qd p differential cross section compared to various theoretical predictions described in the text he curves are normalized to obtain the best fit to the data in each case he top plot shows the range -p -4 GeVrc while the bottom plot shows the full range, - p - 18 GeVrc Error bars do not include a q11 % normalization uncertainty y15 bution functions for the pion nucleon, a charm quark mass of 15 GeVrc, and an average intrinsic transverse momentum of the incoming partons ² : ² ( k of 1 GeVrc he value for ( k : t t was suggested by Mangano w15x and is independently motivated by the study of azimuthal angle correlations between two charm particles in the same event w11 x he D meson results are obtained by convoluting the charm quark results with the Peterson fragw16x with e s 1 he low value mentation function for e was also suggested by Mangano w15x in response to a reanalysis of D fragmentation in e q e y collisions w17 x he PYHIArJESE event generator uses leading order DO w18x CEQL w19x parton distribution functions for the pion nucleon, a charm quark mass of 135 GeVrc, (² k : t of 44 GeVrc, and the Lund string fragmentation scheme to obtain D results ables and 3 show a comparison of our x and p fit results to theoretical predictions and to recent high-statistics charm experiments which used pion beams he evident energy dependence of the shape parameters in ables and 3 are consistent with theoretical predictions w11 x We obtain the total forward cross section by summing the x differential cross section for x ) and assuming the cross section for 8-x -1 is half that of the cross section for 6-x -8 but with the same error Assuming a linear dependence

10 34 ( ) EM Aitala et alrphysics Letters B able Comparison of x shape parameters to recent high-statistics pion-beam charm production experiments and to theory E791 results are for q D mesons; E769 WA9 results are from a combined sample of D, D, and D D and D q s mesons he theoretical results are X obtained using the default parameters described in the text he fit range for n Eq is in the table he fit range for n and x c Eq 3 is y15- x - 5 Experiment Energy GeV x range n n x E " " 1 131" 38 WA9 wx " 11 E769 w1x " 18 MNR NLO c MNR NLO D Pythia c Pythia D X c on the atomic number w x, we obtain the neutral D total forward cross section, s D q D ; x ) s 154 q18 y3 mbarnrnucleon o obtain the total charm cross section, s cc, we multiply our D qd cross section by 17 his accounts for three multiplicative effects: the relative production of charm quarks com- pared to D mesons 1, the conversion from x ) to all x 16, and the conversion to the cc cross section from the sum of charm plus anticharm cross sections 5 w3 x We compare our total charm cross section to other experiments and to the NLO predictions as a function of pion-beam energy in ig 7 All experimental results are obtained by multiply- ing the D qd cross section by 17 he rise of the charm production cross section with energy is modeled reasonably well by the NLO theory, although the absolute value at any point depends greatly on the input parameters to the theory In this paper we have presented the total forward cross section and differential cross sections versus x and p for D mesons from ermilab experiment E791 data his analysis represents the first measurement of the D cross section for a 5 GeVrc pion beam he high statistics allows us to clearly observe a turnover point greater than zero xcs 131" 38 in the eynman-x distribution, providing evidence for a harder gluon distribution in the pion than in the nucleon We have compared our differential cross section results to predictions from the next-to-leading order calculation by Mangano, Nason, and Ridolfi w1x and to the Monte Carlo event generator PYHIA by Sjostrand et al wx 9 With suitable choices for the intrinsic k of the partons and the Peterson fragment tation function parameter, the NLO D meson calculation provides a good match to the p spectra and a fair match to the x distribution he string fragmen- tation scheme in PYHIA softens the original charm quark p distribution too much, and hardens the x spectra too much in both directions However, the able 3 Comparison of p shape parameters to recent high-statistics pion-beam charm production experiments and to theory Data samples are as described in able he fit range for b Eq 4 is -p -4 GeVrc except for WA9 which is -p-7 GeVrc he functions X used to extract b Eq 5 and a,b Eq 6 are fit in the range p ) 1 GeVrc and p - 18 GeVrc, respectively Experiment Energy b b a b GeV GeVrc y GeVrc y1 GeVrc E " 41" 3 36" 3 594" 39 WA9 wx 3589" E769 w1x 5 18" 5 74" 9 14" 3 5" 6 MNR NLO c MNR NLO D Pythia c Pythia D X

11 ( ) EM Aitala et alrphysics Letters B ig 7 heoretical and experimental results for the p N cc cross section versus energy he theoretical curves are obtained from MNR NLO predictions w3 x he three bands solid, dashed, and dotted correspond to three different charm quark masses 15, 1, and 18 GeVrc he variation within the bands comes only from varying the renormalization scale he factorization scale is kept fixed All experimental data points are 17 times the single inclusive D qd x ) cross sections he E653, NA7, WA9, NA3 GeV, NA3 3 GeV, and E769 data were obtained from Refs w4,5,,6 8 x, respectively D result does predict the flattening of the x cross section at high x he many adjustable parameters in the theoretical models allow one to obtain distributions which are quite consistent with these data Unfortunately, a given set of parameters is neither unique, nor does it necessarily provide a good match to other data In conjunction with other charm production results from this and other recent high-statistics experiments, however, it may be possible to find a unique set of parameters hese results come from experiments with a variety of beam energies and types, and include measurements of differential cross sections w,8,31 x, production asymmetries w,8 33 x, and correlations between two charm particles in the same event w3,34 36 x Unlike the uncertainties in the theoretical calculations of the differential cross sections, the uncertainties in the theoretical calculation of the total cross section come mostly from the perturbative calculation he relatively large uncertainties are due to the low mass of the charm quark, which results in a large unknown contribution from higher-order terms he total forward D q D cross section mea- q18 sured by E791 is s D qd ; x ) s154y3 mbarnrnucleon, assuming a linear atomic number dependence he cross section is consistent with the MNR NLO prediction Acknowledgements We express our special thanks to S rixione, ML Mangano, P Nason, and G Ridolfi for the use, and help in the use, of their NLO QCD software We

12 36 ( ) EM Aitala et alrphysics Letters B gratefully acknowledge the assistance from ermilab and other participating institutions his work was supported by the Brazilian Conselho Nacional de Desenvolvimento Cientıfico e echnologico, CONA- Cy Mexico, the Israeli Academy of Sciences and Humanities, the US Department of Energy, the US- Israel Binational Science oundation, and the US National Science oundation References wx 1 RK Ellis, C Quigg, ERMILAB-N-445 January 1987 wx JA Appel, Ann Rev Nucl Part Sci ; DJ Summers et al, Proceedings of the XXVII th Rencontre de Moriond, Electroweak Interactions and Unified heories, Les Arcs, rance 15- March, wx 3 E791 Collaboration, EM Aitala et al, submitted to Eur Phys J C, ermilab-pub e, hep-exr9899 Sep- tember 1998 wx 4 E516 Collaboration, K Sliwa, Phys Rev D wx 5 E691 Collaboration, JR Raab, Phys Rev D wx 6 E769 Collaboration, GA Alves, Phys Rev Lett wx 7 S Amato, Nucl Instr and Meth A wx 8 S Bracker et al, IEEE rans Nucl Sci ; Rinaldo, S Wolbers, Comput Phys wx 9 H-U Bengtsson, Sjostrand, Comput Phys Commun ; Sjostrand, Pythia 57 and Jetset 74 Physics and Manual, CERN-H711r93, 1995 w1x Particle Data Group, Eur Phys J C w11x S rixione, ML Mangano, P Nason, G Ridolfi, Nucl Phys B w1x ML Mangano, P Nason, G Ridolfi, Nucl Phys B w13x PJ Sutton, Phys Rev D w14x PN Harriman, Phys Rev D w15x ML Mangano, hep-phr ; ML Mangano, private communication April 1999 w16x C Peterson, Phys Rev D w17x M Cacciari, M Greco, Phys Rev D w18x DW Duke, J Owens, Phys Rev D w19x J Botts, Phys Lett B wx BEARICE Collaboration, M Adamovich, Nucl Phys B w1x E769 Collaboration, GA Alves, Phys Rev Lett wx E769 Collaboration, GA Alves, Phys Rev Lett w3x S rixione, ML Mangano, P Nason, G Ridolfi, Heavy lavours II, Advanced Series on Directions in High Energy Physics, 1997, hep-phr9787 w4x E653 Collaboration, K Kodama, Phys Lett B w5x LEBC-EHS Collaboration, M Aguilar-Benitez et al, Phys Lett B w6x ACCMOR Collaboration, S Barlag, Zeit Phys C w7x ACCMOR Collaboration, S Barlag, Zeit Phys C w8x E769 Collaboration, GA Alves, Phys Rev Lett w9x E791 Collaboration, EM Aitala et al, Phys Lett B ; Phys Lett B w3x E769 Collaboration, GA Alves, Phys Rev Lett w31x SELEX Collaboration, G Garcia, SY Jun et al, to be published in the proceedings of American Physical Society APS Meeting of the Division of Particles and ields DP 99, Los Angeles, CA, 5-9 Jan 1999, hep-exr9953 w3x WA89 Collaboration, MI Adamovich et al, submitted to Eur Phys J C, CERN-EPr98-41, hep-exr9831 March, 1998 w33x E687 Collaboration, PL rabetti, Phys Lett B w34x WA8 Collaboration, M Adamovich, Phys Lett B w35x BEARICE Collaboration, M Adamovich et al, Phys Lett B ; Nucl Phys B ; Phys Lett B w36x E687 Collaboration, PL rabetti, Phys Lett B