First CMS Physics Paper!
Today the first CMS physics paper appeared on the arXiv, 1002.0621,
Transverse momentum and pseudorapidity distributions of charged hadrons in pp collisions at √s = 0.9 and 2.36 TeV.
Notice that data from the √s = 2.36 TeV data are included – these are the highest-energy data in the world at the present time, hence a kind of feather in CMS’s cap.
More soberly, the actual parton-parton energies are quite low, since the events are non-single-diffractive interactions, basically, glancing blows of the two protons and a far cry from, say, the production of a pair of top quarks or W bosons.
The measurements concern the transverse momentum pT and (pseudo-)rapidity η distributions of charged hadrons. As I discussed earlier, these distributions can be related to scaling arguments started by Feynman, and as such lie in the area of non-perturbative physics of hadron production, for which there are phenomenological models. At a minimum, these models must be tested and constrained so that they can be used for modeling underlying event structure for high-energy collisions. These measurements also serve as a baseline for heavy-ion collisions.
The event selection was as open and simple as can be imagined, demanding little more than signals in beam monitors indicating that bunches as collided at the center of the CMS detector. A very loose cut on the number of pixel detector hits was enough to eliminate beam-gas events entirely. The event had to have a reconstructed vertex, too. Selections efficiencies are high, naturally, around 86% or so.
Three methods were used to measure the rapidity distribution, dN/dη. The most primitive method simple counts reconstructed clusters in the pixel barrel detector, since the shape of a cluster already provides a good indication for a charged hadron track. The second method links such clusters to build short track segments called “tracklets” which do not provide curvature information but clearly allow for a better indication of the origin of the track. Finally, a full-blown track reconstruction, using both the pixel and the silicon strip detector, was used, which provides momentum measurements as well as direction. The point of the three methods is to demonstrate the robustness of the measurements with respect to the methods used and the performance of the detector – which was excellent in any case.
The systematic uncertainties concern the acceptance and efficiency estimates and to what they degree they depend on the phenomenological (Monte Carlo) models. The exclusion of single-diffractive events, in which the hadronic final state is typically very forward and difficult to observe, is only partially successful; the purity of the final samples is roughly 95%. This purity estimate depends again on the models, so there is a systematic uncertainty. The net uncertainty is only 3%. An additional 2-3% comes from reconstruction efficiencies, and 1% for knowledge of the tracker geometry.
The transverse momentum distribution of primary charged hadrons is shown below, for both 0.9 and 2.36 TeV. The mean pT is measured to be 0.46±0.01±0.01 GeV at 0.9 TeV, increasing slightly to 0.50±0.01±0.01 GeV at 2.36 TeV. This increase is clearly seen in the tail of the pT distribution:
The number density as a function of pseudorapidity, dN/dη, is presented for both √s = 0.9 TeV and 2.36 TeV. The results from the three methods (not shown) agree within errors. One might notice the much greater accuracy provided by the CMS measurement as compared to the early one by ALICE.
The distribution dN/dY, where Y is the rapidity, is expected to be flat for Y≈0. The pseudorapidity usually coincides with the rapidity except when the rest mass of the particle is not negligible, which is the case with pions and kaons with a transverse momentum of a couple hundred MeV. The subtle wavy effect seen in the figure comes from the numerical differences between η and Y.
It is clear from the figure that more hadrons are produced per unit of rapidity at high energies than at lower energies, which is interesting given that peripheral nature of these collisions. According to the original arguments of Feynman, the variation with center-of-mass energy should go as ln(s), but some years later experiments showed that the rise is quadratic in ln(s). It is also worth noting that the difference between p+p and p+anti-p collisions is less than a couple of percent, again underscoring the soft, peripheral nature of these collisions.
From the CMS data, dN/dη = 3.48±0.02±0.13 at 0.9 TeV and dN/dη = 4.47±0.04±0.15 at 2.36 TeV.
As stated in the paper, the increase of 28.4% is significantly more than the 18.5% predicted by a tuned version of PYTHIA, and the 14.5% predicted by PHOJET models. Interesting.
So the first paper from CMS is interesting, with a non-trivial result, and perhaps a good harbinger for the next few months, when we will surely see many measurements of hadronic event properties from the LHC experiments. 🙂
(Note: I am a member of the CMS Collaboration and my name appears on the author’s list.)
Entry filed under: Particle Physics.