Diffraction & Forward Physics in CMS: results and perspectives

. In this plenary talk given at ISMD2014 the interesting CMS analyses on di ﬀ ractive proton-proton cross section and Mueller-Navelet dijet azimuthal decorrelations performed at √ s = 7 TeV were presented.

. The composition of the total pp cross section is where "SD" means Single-Diffractive, "DD" means Double Diffractive, "CD" means Central Diffractive and "ND" means Non-Diffractive, all belonging to the inelastic cross section σ INEL . At LHC the elastic cross section, σ EL ≈ 25mb, is about 20% of the σ tot [2] and the inelastic, σ INEL ≈ 69mb, is responsible for about 80% of σ tot [3].
Moreover, hadronic interactions with LRG are mediated by Pomeron exchange, based on the trajectory of Regge theory α(t) = 1 + + α t, with t = (p 1 − p 2 ) 2 . The first CMS analysis presented in this talk [4] has used the PYTHIA8-MBR (Minimum Bias Rockefeller) [5] and PYTHIA8-4C (with diffraction from Schuler & Sjostrand) [6] for comparison, with α = 0.25GeV −2 . The data were collected in 2010 using a low pileup scenario which is the most suitable for diffractive event selection using a LRG signature. Signal in both BPTX (Beam Pick-up Timing Experiment) detectors and activity in any of the BSC (3.2 < |η| < 4.7) were required for the presence of two bunches along with minimal activity in the central CMS detector (Minimum Bias trigger). In addition, Single Diffractive (SD) and Double Diffractive (DD) events were separated with CASTOR (6.6 < |η| < 5.2).
To accept low to moderate diffractive masses (12 < M X < 100GeV), no vertex was required. The diffractive offline selection was based on LRG within |η| < 4.7 (HF was not used, therefore limiting central CMS detector coverage), the proton momentum loss ξ was reconstructed from particles in |η| < 4.7 according expressions 1 where M X is the mas of the diffractive system. Three experimental topologies based on the position of the LRG were analysed according to Fig. 3, as well as the distributions for the Mininum Bias data samples, following the aforementioned experimental topologies, compared with PYTHIA8-MBR predictions. The cuts to select the samples for SD1, SD2 and DD topologies with a central LRG signature were η max < 1, η min > −1 and Δη 0 > 3, respectively, where η max(min) is the highest (lowest) η of a particle candidate within |η| < 4.7. In Fig. 4 the SD and DD contribution from the SD2 event sample (three samples tagged according CASTOR) are best described by Pythia8-MBR tune. 1 Detector level; "i" runs over all Particle Flow objects measured in the central detector; dissociated system occurs on the ±z side of the detector.
One can see the SD and DD contributions from SD2 event sample in the top of the Fig. 5. Comparison with PYTHIA8-MBR, PYTHIA8-4C, and PYTHIA6 [7] MC simulations were done with two values of Pomeron intercept in the MBR model. It is shown that PYTHIA8-4C does not reproduce the falling behavior of SD and in the -5.5 < log 10 ξ < -2.5 region the total SD cross section reaches σ S D Total = 4.27 ± 0.04(stat.) +0.65 −0.58 (syst.) mb.  In addition, the DD contribution from DD event sample is shown in the right side of the Fig. 5. Over the region Δη > 3, M X > 10GeV and M Y > 10GeV, the cross section value approaches σ DD = 0.93 ± 0.01(stat.) +0.26 −0.22 (syst.) mb. An alternative approach to the study of diffractive events is to measure the differential cross section of the forward rapidity gap "Δη F ". In Fig. 6 we can notice the exponential falling of the ND contribution and the diffractive plateau at Δη F > 3 with mixtures of SD and DD events for all stable final-state particles with pT > 200 MeV in |η| < 4.7. A good description using PYTHIA8-MBR with

Forward Jets at CMS: the azimuthal angle decorrelations of Mueller-Navelet jets at √ s =7 TeV
In the hard processes regime ( √ s ≥ pT Λ QCD ), data are well-described by perturbative QCD within the DGLAP evolution equation (mainly pT ordered). On the other hand, for the semi-hard processes ( √ s pT Δ QCD ) the asymptotic region ( √ s → ∞) is described by the BFKL approach, that is the number of emitted partons increases with an increasing rapidity interval Δη = |η 1 − η 2 | between the Muller-Navelet (MN) jets and, hence, the MN jets are no longer back-to-back (Δφ = π) in azimuthal direction, leading to a decorrelation. The observation of such decorrelations, see Fig. 7 expressions, indicates the presence of BFKL contributions [8].
The measurement conditions in the second CMS analysis of this talk [8] were Δη MN Jets < 9.4, two jets with pT > 35 GeV and |η| < 4.7 using the anti-kT algorithm with jet size R = 0.5. Several observables, like the azimuthal, Δφ, distribution as a function of Δη; the average cosines C n = cos(n(π − Δφ)) with n=1,2,3; and the ratios C2/C1 and C3/C2 were analysed. In Fig. 8 the Δφ distributions in two rapidity intervals (Δη < 3.0 and 6.0 < Δη < 9.4) are shown. We can notice azimuthal decorrelation raising with increasing |Δη| as expected. Concerning the MC generators, the HERWIG++ shows the best description in all bins, the PYTHIA, SHERPA and CASCADE present too large decorrelation. In the comparison plots for the average cosine C n of Fig.9 the measured average cosines decrease with increasing Δη indicating increasing decorrelation of jets in azimuthal angle. The MC generators PYTHIA and HER-WIG++ agree mostly with the data, while SHERPA, CAS-CADE and NLL BFKL do not, being either above or below them 2 . It is important to notice that, at large values of Δη, additional emissions are becoming kinematically suppressed due to energy-momentum conservation near the phase-space boundary. Thus, coplanarity of MN jets results in an increase of average cosines.
The azimuthal angle decorrelation depends also on multiple particle interactions (MPI), producing additional jets not correlated with those from the primary interaction. The Figs.10 and 12 show the polar angle ordering (AO) effects in the parton showering and multiparton interactions (MPI), and indeed the average cosines C n are very sensitive. A good description of the data requires polar angle ordering in parton showering.   The ratios of C n are expected to be more sensitive to BFKL effects (See Fig. 11). PYTHIA and HERWIG++ show good agreement at low Δη, but discrepancies in the large Δη region. On the other hand SHERPA is above the data while CASCADE is far below the data. NLL BFKL calculation describes the ratios quite well, especially the C 2 /C 1 one.

Summary
Two interesting CMS results related with inclusive Single Diffractive and Double Diffractive cross section measurement at 7 TeV and BFKL effects through azimuthal decorrelations in MN jets were presented.
The forward and diffractive physics show as a perfect testing ground for models and theories [10].
We notice the importance of Monte Carlo generators comparisons and tuning for better data matching. In addition, several 8 TeV analyses, with 2012 data and integrated luminosity ≈20fb −1 are in progress.