Getting the most out of the 09 data

9 February 2010

A possible kaon to charged pion decay in an event from December 6th last year.



ATLAS collected over one million 900 GeV events at the end of 2009, and even recorded a brief flash of 2.36 TeV collisions; not bad for the first few weeks of LHC action. Right now, the pressure is on to squeeze out every last drop of information from that data, in order to better understand the detector ahead of this year’s high-energy regime.

“With these low-energy collisions, we can learn the way our detector operates, we can test our reconstruction software, and we can test the simulation,” says Deputy Physics Coordinator Aleandro Nisati, adding: “At this stage the most important thing is to check the way the central tracker operates, as well as the calorimeter.”

The key things to look for in order to verify the performance of the Inner Detector tracking are charged pions. “Actually the way it turns out is that about 90 percent of the tracks we see are from pions,” says PhD student Heather Gray, who is part of a group working on the tracking efficiency for the Minumum Bias analysis of the first data.

Comparing these tracks to 900 GeV simulations, the Minimum Bias group are working to understand and improve the tracking efficiency of the reconstruction software. In order to do that though, they need to know how good the simulation that they’re comparing it to is; how accurate its image of the detector is – inactive material, dead channels and all.

If that sounds a little like they’re using one uncertainty to investigate another, well, that’s because they are. “This is why it’s a bit of a balancing act,” explains Heather. “With what we’ve got right now, you can’t come up with the perfect procedure … you can’t use the easy methods that give you nice results. What you do is you check it more and more until you have confidence in it.”

Using around 300,000 of the 900 GeV events – those from when all the sub-detectors were on and beams were colliding well – the Minimum Bias group has a preliminary estimate of the systematic uncertainty in the amount of material in the detector, an important factor in the final tracking efficiency estimate.

“Right now, we have a preliminary estimate, but it needs further study to back it up. The biggest question is estimating how well we know the material in the Inner Detector,” says Heather. The preliminary results show that the simulation models the barrel contents well, but that there are differences between the data and the simulation of the endcaps that aren’t yet understood. “Once we have new data though, we’ll have new techniques that we can use to refine the results.”

Whilst the Minimum Bias group looks at all the tracks seen in the detector, others have been using the reconstruction software to single out interesting events and produce resonance plots to identify ATLAS’s first candidates from the “particle zoo”. At the energies seen so far, collaborators have been able to identify π0, K0s, Λ0, η, and Ω particles, and have been comparing the size, shape, and position of these resonance peaks with simulations, to evaluate the level of understanding of detector performance.

“Already at this initial phase, the agreement is almost astonishing,” reports Aleandro. This bodes well. “Although clearly we have to look more in the details. We had a very good start, but now we have to make a systematic effort to check again all the aspects of our simulation and verify the accuracy.”

The events last year also gave ATLAS a first opportunity to work with particles originating from the centre of the detector, and to make alignment studies and hone the trigger timing for collisions. Smooth trigger functioning will be crucial once events start to occur at high rate in a few weeks’ time.

In the meanwhile, even though the limit for extractable physics objects and system-performance checks has been reached, there is still more to be gleaned from the 2009 data. Rather than identifying particles and defining their physics parameters, the focus should now be on looking at the kinematics of the events and measuring the way particles were actually produced, according to Aleandro: “The next step in my opinion is to turn now all the data we have available into physics measurements… this is the only way to deeply understand your system. For example, for the K0s, we can detect an extremely nice signal with very low background. So we have a good opportunity here to measure the rate of production of this particle in Minimum Bias events.“

Look out for the March 8th edition of e-News, where we’ll hear direct from Heather and others involved in handling the first full analysis.

 

Ceri Perkins

ATLAS e-News