First ATLAS Physics Paper: Charged Particle Multiplicities
Today the ATLAS Collaboration posted their first physics paper on the archive (arXiv:1003.3124):
Charged-particle multiplicities in pp interactions at √s = 900 measured with the ATLAS detector at the LHC
This is the third physics result from the LHC. The first paper was submitted by the ALICE Collaboration on 28-November (see my discussion here) and by the CMS Collaboration on 3-February (my discussion is here). All three papers concern the same measurement – the number of primary charged hadrons produced in non-single-diffractive events, and their distributions in pseudorapidity (η) and transverse momentum (pT).
The authors take care to define their suite of models for min-bias events. There are essentially two event generators available: PYTHIA and PHOJET, with different theoretical underpinnings, and there are several “tunes” of free parameters in PYTHIA which have been set using data from the Tevatron.
In this kind of analysis, the selection of tracks is crucial. The authors set a minimum pT of 500 MeV (CMS used 100 MeV and the ALICE magnet was off). The pseudorapidity range is |η|<2.5, compared to 2.4 in CMS and 1.6 in ALICE. The agreement of the detector simulation with the data is remarkable, allowing a precise control of contamination from secondary hadrons (e.g. pions produced in K-short decays). In contrast to the CMS and ALICE papers, the ATLAS paper shows clear measurements of vertex and track reconstruction efficiencies. (This is not to say that ALICE and CMS did not measure the efficiencies – they did not chose to transmit the results.) Furthermore, this paper describes in detail how corrections were made for inefficiencies, contamination and resolution. These corrections are not large, and the ATLAS methods are standard. There is a thorough investigation of possible systematic effects which all turn out to be quite small.
Corrected distributions are nicely presented:
The plot of dN/dη shows the characteristic sea-gull shape I discussed earlier (it arises from the use of pseudorapidity η in place of rapidity y). It is plain that the charged multiplicity is significantly underestimated by all models, and the discrepancy is worst in the center. The paper reports an average value of 1.333±0.003±0.040 for |η|<0.2. This cannot be directly compared to the CMS result because it is restricted to pT>0.5 GeV; nonetheless, the fact that the models underestimate the multiplicity by a few percent is nicely confirmed.
The plot of dN/dpT shows that models also fail to reproduce the distribution for pT>0.7 GeV. For the highest pT reached, around 10 GeV, the discrepancy is as large as 50%. (The CMS paper shows no such comparisons, showing instead a fit to an empirical function.) The lower plot on the right shows that the discrepancy is worst in events with many charged hadrons.
Here is a very nice composite of measurements by ATLAS, CMS and UA1:
If you look carefully at the ratios displayed in the lower part of the plot, you will see that the CMS multiplicities are systematically lower than those measured by ATLAS. The authors understand this to come from the way CMS normalize their result. The UA1 result is higher due to the definition of the trigger used. Evidently a question as simple as What is the charged multiplicity? inevitably requires choices that can be arbitrary to some degree. The ATLAS authors go the extra mile and derive a measurement that can be directly compared to the CMS result by reducing their range in |η| and applying a model-dependent correction for non-single-diffractive events as was done by the CMS authors. They find, for pT>0.5 GeV, dN/dη=1.240±0.040 to be compared to the CMS measurement 1.202±0.043; in both cases the errors are mainly systematic uncertainties. The figure above shows the CMS data stopping at 4 GeV because the CMS paper does not report results above that value.
The ATLAS analysis looks to be very solid and thoroughly done. Due to the limitation pT > 0.5 GeV, the authors cannot go further, so there is no result on the mean pT or the total charged multiplicity to be compared to earlier experiments. There is also no result in this paper from the 2.36 TeV data, for which CMS observed an increase in the charged multiplicity that is substantially higher than predicted by the models.
In short, both ATLAS and CMS agree that the models for min-bias events are inaccurate and need to be tuned in order to understand the LHC data. This has spurred lots of activity between the experimental and theoretical communities concerned with this kind of physics – in the coming months there will be lots of discussion of models for non-perturbative QCD interactions and more measurements with which to tune them.
Congratulations to the ATLAS colleagues! 🙂
Entry filed under: Particle Physics.