Higgs production in association with a W boson is considered to be one of the most sensitive processes for Higgs searches at low mass ("110–150 GeV/c2). At the Tevatron, the relatively large Higgs boson production cross section for WH, combined with the decay of the Higgs to b¯b, makes WH to lepton and neutrino + bb an ideal channel for the Higgs search at low mass. Although the gluon fusion production cross section is about ten times larger than that of WH, the Higgs decay to a b quark and anti b quark channel is not viable due to the large bb background from SM QCD processes. In the WH channel, the presence of a lepton from the decay of the associated W boson reduces the huge bb background rate.

The CDF WH/Single Top analysis group, recently published two papers on searches for WH to lepton and neutrino + bbar at sqrt(s) = 1.96 TeV in Physical Review D (using an integrated luminosity of 1.9/fb) and Physical Review Letters (2.7/fb). These publications present searches based on improved analysis techniques that rely on a secondary vertex b-tagging algorithm (SECVTX), together with a b-tagging neural network filter, to distinguish b-quark jets from lightflavor and gluon jets. Features of the latest result in PRD include (1) an extension of the double-tag sample to use not only the SECVTX tight b-tagger, but also a jet probability btagger, (2) an increase in signal acceptance by including electrons in the forward region of the detector, and (3) the introduction of a neural-network-based multivariate discriminant technique (NN) to reduce the large background contamination. The publication in PRL uses a larger data sample and combines the NN technique with a second multivariate strategy—a boosted decision tree technique derived from a matrix-element approach (MEBDT)—to improve our sensitivity.

Since the submission of these publications, we have rapidly pursued improvements that extend the signal acceptance and increase the signal-to- background separation. Frank is one of a small handful of graduate students at CDF working intensely on these improvements. He singlehandedly reorganized the NN software into a new WHNN package, which became the basis for an analysis framework used by several postdocs and graduate students. Frank has systematically reproduced previous versions of the analysis and has gained tremendous expertise in understanding the QCD backgrounds, which are created using templates from fake leptons in data, as well as other backgrounds modeled using Monte Carlo simulations. Frank is now responsible for nearly all aspects of the NN analysis at CDF. Our group is working closely with another graduate student, Adrian Buzatu (McGill University), to include a new non-triggered lepton category known as “isotracks.”

Over the past two years, Frank has presented the CDF WH analysis at three conferences, including the 2010 APS April Meeting. This summer and fall, Frank has been focused on searching for additional sensitivity in WH events with three jets, a new analysis direction for CDF. Previous versions of the analysis have considered only events with two energetic jets (presumably from the b and anti b). Frank is investigating WH Monte Carlo simulations with jet clustering methods to understand the processes that are responsible for the third jet, and we expect that these studies will provide an effective way to discriminate signal from background. Preliminary results indicate that the inclusion of three-jet events will improve the sensitivity by 5–8% for low-mass Higgs. The goal is to have a new set of results ready in time for the 2011 winter conferences.