The STAR experiment

In an effort to better understand the future of the TPC in the high luminosity regime, several meetings were held and efforts carried through.

Present at the meeting(s) were: myself (Jerome Lauret), Alexei Lebedev, Yuri Fisyak, Jim Thomas, Howard Wiemann, Blair Stringfellow, Tonko Ljubicic, Gene van Buren, Nikolai Smirnof, Wayne Betts (If I forgot anyone, let me know).

See this presentation by Jim Thomas at the February 2005 STAR Collaboration Meeting for a more general discussion of TPC distortions at higher luminosities in the future of RHIC. I will focuse here on SpaceCharge + GridLeak.

---

As part of the QA process, I want to examine the performance of the corrections as the luminosity increases. I can do this by plotting the spread in sDCA values versus a luminosity measure, and I do so against ZDC coincidence rate in the plots below. The spread at a given luminosity is the RMS of the sDCA distribution (sDCA is measured event-wise via a fit to the mean of a central peak in track-wise sDCA's formed by primaries among the global tracks), and is shown by black error bars. The red error bars are the error on the mean, which gives you an idea of the statistics I had for that luminosity when making these plots.

Several of the high luminosity runs exhibit very high rates in the BBC yellow and blue beam backgrounds (bbcyb, bbcbb). Here one can see this tail of high background periods, and the simple cuts (shown by lines) which I have chosen to remove them (I use only the lower left corner).

Excluding the high backgrounds, excellent performance at low luminosities is followed by slightly worse, but stable, performance at mid and high luminosities. The conclusion I draw from this is that if backgrounds can be controlled, we see no evidence to indicate that we cannot handle even higher luminosities than those used in the 2004 AuAu run. It will be of high interest to see what effect the background shielding installed in the fall of 2005 will have on backgrounds in upcoming runs.

High Background Included

You can see that statistics get poor for the highest luminosities. For this plot, the only cuts are that |sDCA| < 1.0cm (to remove the pull of any poor event-wise sDCA fits) and correction mode=21 (event-by-event method was used for correction, not the fall-back to the prepass value).

High Background Excluded

I lost some statistics at high luminosities with the backgrounds cut, but it is clear that the high backgrounds are primarily responsible for the degradation in spread of sDCA seen in the high backgrounds plot at higher luminosities.

---

Gene