Measurements of Bottom Anti-Bottom Azimuthal Production Correlations in Proton-Antiproton Collisions at s**(1/2)= 1.8 TeV

TL;DR

The paper probes bottom quark pair production at the Tevatron by measuring azimuthal correlations with two complementary approaches: a BVTX-based double-tag analysis and a J/psi-lepton correlation study. These methods isolate higher-order QCD processes (flavor excitation and gluon splitting) that smear Delta phi, revealing that about a quarter to a third of bbbar pairs populate small opening angles, in line with parton-shower MC and NLO predictions. The findings strengthen the Standard Model picture of bbbar production at hadron colliders and show that no beyond-Standard-Model mechanisms are needed to describe the observed correlations. The work also demonstrates robust techniques for background subtraction, detector-effect corrections, and bottom-quark level extrapolations, informing future Run II flavor-tagging and theoretical comparisons. Overall, the results corroborate the importance of higher-order contributions in heavy-flavor production at high-energy hadron colliders and validate current MC and NLO frameworks for bbbar kinematics.

Abstract

We have measured the azimuthal angular correlation of bottom anti-bottom production, using 86.5 pb-1 of data collected by Collider Detector at Fermilab (CDF) in proton anti-proton collisions at sqrt{s}=1.8 TeV during 1994-1995. In high-energy p-pbar collisions, such as at the Tevatron, bottom anti-bottom production can be schematically categorized into three mechanisms. The leading-order (LO) process is ``flavor creation,'' where both the bottom and anti-bottom quarks substantially participate in the hard scattering and result in a distinct back-to-back signal in final state. The ``flavor excitation'' and the ``gluon splitting'' processes, which appear at next-leading-order (NLO), are known to make a comparable contribution to total bottom anti-bottom cross section, while providing very different opening angle distributions from the LO process. An azimuthal opening angle between bottom and anti-bottom, Delta phi, has been used for the correlation measurement to probe the interaction creating bottom anti-bottom pairs. The Delta phi distribution has been obtained from two different methods. One method measures the Delta phi between bottom hadrons using events with two reconstructed secondary vertex tags. The other method uses b-bbar --> (J/psi X)(l X') events, where the charged lepton (l) is an electron or a muon, to measure Delta phi between bottom quarks. The bottom anti-bottom purity is determined as a function of Delta phi by fitting the decay length of the J/psi and the impact parameter of the lepton. Both methods quantify the contribution from higher-order production mechanisms by the fraction of the b-bbar pairs produced in the same azimuthal hemisphere, f_toward. The measured f_toward values are consistent with both parton shower Monte Carlo and NLO QCD predictions.

Figures (27)

• Figure 1: Example Feynman diagrams that contribute bottom production. The bottom two virtual exchange diagrams enter into the NLO calculation through interferences with leading-order terms. Interferences between the flavor creation, flavor excitation, and gluon splitting diagrams, as well as the virtual exchange diagrams, are ignored in the parton shower approximation
• Figure 2: Schematic of a quarter cross-section of the CDF Run 1b detector.
• Figure 3: The trigger electron $p_{T}$ and $E_{T}$ and the trigger muon $p_{T}$ distributions from data compared to the high ISR Pythia Monte Carlo sample. The comparisons of data to other Monte Carlo samples is similar and can be found in Ref. lannon-thesis. The muon trigger threshold is clearly visible in the lower plot. The small number of muons below the 8 GeV trigger threshold come from events containing a second muon that passes the offline selection.
• Figure 4: The secondary vertex tag distributions from electron data compared to the high ISR Pythia Monte Carlo sample. Comparisons involving muon data and comparisons to other Monte Carlo samples are similar and can be found in Ref. lannon-thesis.
• Figure 5: Comparisons between the shape of the tag $\Delta\phi$ distribution for data (points) and the tag $\Delta\phi$ distribution predicted by the various Monte Carlo samples (line). The contributions from flavor creation, flavor excitation and gluon splitting are added together according to the individual Monte Carlo cross section predictions for these contributions and only a single common normalization is varied to get the best match to data. The $\chi^{2}$ values shown in the plots account only for the data and Monte Carlo statistical errors.