Feature - Single and playing hard to get

Cecilia Gerber of the University of Illinois at Chicago presented the DZero result at a seminar at Fermilab, followed by Rainer Wallny of UCLA who presented the CDF result. Image courtesy of Fermilab.

Scientists from the CDF and DZero collaborations at Fermilab announced last week the discovery of the top quark produced as a single particle. This comes almost 14 years to the day since these same teams first observed top quark pairs in 1995 – the production of single top quarks proved much harder to identify experimentally. The analysis tools and techniques that yielded this discovery also significantly advance the scientists’ goal of observing the Higgs particle at Fermilab.

To make the single top discovery, physicists of the CDF and DZero collaborations spent years combing independently through the results of proton-antiproton collisions recorded by their respective experiments at the Tevatron.

“Already 25 years ago theorists suggested that top quarks should be produced singly. It took us experimentalists until today,” said Rainer Wallny of UCLA who announced the CDF result to a standing-room-only crowd at Fermilab. “The challenges of finding the single top are similar to those of finding the Higgs boson—we must extract an extremely small signal from a very large background.”

Schematic views (called Feynman Diagrams) of the processes that produce a single top quark. Protons and antiprotons comprise quarks, antiquarks and gluons. A proton-antiproton collision can produce a single top quark in two different ways: (a) A quark and an antiquark can create a W boson, which then decays into a top quark and an anti-bottom quark. (b) Or a gluon and a quark interact, with the quark emitting a W boson and the gluon creating a bottom quark and anti-bottom quark. The W boson then interacts with the bottom quark and produces a single top quark. In both cases, the top quark is short-lived and decays, for example, into a bottom quark, a lepton (such as a muon) and a neutrino.  Image courtesy of the DZero collaboration

This analysis has been the single most computing-intensive effort at Fermilab, according to Cecilia Gerber, a DZero scientist from the University of Illinois at Chicago who announced her experiment’s result. She said it would not have been possible as recently as five years ago.

The analysis required comparing the results to simulations, which both groups have run on Open Science Grid resources.

“We used around 200 CPU years to generate simulated events for single top production,” said Dmitri Denisov, one of the co-spokespersons for DZero. “This represents around 85 million fully reconstructed events. These simulations are critical for calculating which signals to accept for analysis and which to reject as ‘background’.”

CDF estimated having consumed over 400 CPU years over the last six calendar years for calculations devoted to the single top search, but they didn’t keep track precisely.

“Our big grid resources mean that we can focus more on how much real time it takes to get a job done and not on how much CPU time we use,” said Thomas Junk of CDF.

Each team identified several thousand collision events that met their expectations of a single top signal.

“The signature of single top events is easily mimicked by other subatomic processes that occur at much higher rates,” said Gerber.  “To stand a chance of observing the single top quark signal, we had to develop sophisticated data analysis techniques to isolate the expected 200 or so single top events from the billions of particle collisions that occurred at the Tevatron.”

The scientists pared the set down to a few hundred events of the real thing – where “real” is determined by statistics. For both teams, the probability that background events have faked the signal is now only one in nearly four million. Both experiments consider this result solid.

—Anne Heavey, iSGTW