Top physics


The top is a very peculiar quark, due to its very large mass, compared to that of the other quarks. It was discovered at Fermilab and the Tevatron is still the only collider that regularly produces it. This makes CDF the ideal place to study the properties of this quark and any connection it has with the electroweak symmetry breaking or new physics.

Measurement of the Ratio R=B(t → W b)/B(t → W q)

The ratio R of branching fractions B(t → W b)/B(t → W q) is almost exactly equal to 1, according to the Standard Model. Any significant deviation from unity would be an indication of new physics:
  • Either a non-standard top quark production
  • Or a non-standard background to the top production
  • Or a fourth generation of quarks
  • Or all of the above !
The ratio of branching fractions is sensitive to the relative top content in data subsets with 0, 1, and 2 identified b jets ("b tags"). Simply because the number of b-tagged jets we observe depends on the number of b quarks that are produced (which depends on R) and our ability to tag the respective b-jets (which depends on our b-tagging efficiency). The strongest constrain comes from the top quark content in the 0-tag subset. The University of New Mexico group was the only one that could measure the top content in a CDF data subset with no b-tagged jets, with the use of an artificial neural network (ANN). This lead to the first measurement of the ratio R=B(t → W b)/B(t → W q) with CDF II. We used 160 pb-1 of lepton+jets data. The ANN was trained with W+jets and t-tbar MC events. A QCD-rich dataset was passed through the trained neural network and fixed to the estimated size of it. The content of top-quark in our data before any tagging ("pretag"), and in our data with 0, 1, and 2 b-tags can be seen here:

The ANN output signal and background shapes are fitted to the ANN output for the CDF data, to give us the top quark content for all b-tag jets bins. As can be seen, the fraction of measured top events (red) increases with the number of identified b jets, as expected. The fraction of top content for the 0,1,2 b-tag bins is entered in a likelihood, the maximization of which gives us R. As seen above, the maximization of the likelihood gives an unphysical answer. This is not unexpected, given that we are really sensitive to the product of R*&epsilon, where &epsilon is the event b-tagging efficiency. We employed the Feldman-Cousins method to determine (in an unbiased way) the 95% CL limit on R, given our measurement, as seen above. Our result is R>0.61 at 95% CL. Assuming 4 generations of quarks and using the unitarity of the CKM matrix, we can translate this result as |Vtb|>0.78 at 95% CL. To learn more, please visit the public analysis web page.

We included the dilepton data and used a more precise a priori method for the 1-tag and 2-tag W+jets background. The final result was the most precise determination of R. Assuming 3 generations and given the unitarity of the CKM matrix, we also determined indirectly the Vtb matrix element. Our results were published in PRL (Phys. Rev. Lett. 95, 102002 (2005)).

W Helicity from top decays

I was a "godparent" (internal CDF reviewer) for the paper: “Measurement of the helicity fractions of W bosons from top quark decays using fully reconstructed t-tbar events with CDF II”, CDF Collaboration, Phys. Rev. D 75, 052001 (2007). I provided a method (based on the method of moments) for the extraction of the two W helicity factors simultaneously. This was used by the authors for verifying the intrinsic value of these helicities at the Monte Carlo generation level. I later expanded the method, and I proposed a way for the simultaneous measurement of the helicities of the W from top events in an environment of low-statistics, where 2-dimensional fits could be practically difficult. This method is described here.