========================================
April 18th 2018 - Journal club meeting minutes
========================================
3 papers today: R32 for strong coupling constant (CMS), a theory paper proposing how alpha-s can be used to confine BSM physics, and a paper discussing limiting factors for QCD in the future.
Rebecca chairs.
------------
Topics:
------------
* Summaries of papers: (see details below)
* Discussion plots and concepts in both papers.
Notes (CMS original R32) - 7 TeV CoM
Intro
* Any event with three or more jets in the final state originates from hard gluon radiation and other higher order QCD processes.
* The jet pT and rapidity are strongly corrected with that of the final state partons
* R32 was binned as a function of HT = sum pT of jets in ROI for the Run1 paper
* HT is less sensitive to individual detector effects (?) than pT
* JES and jet selections efficiency uncertainties cancel in R32 mostly, uncertainty on lumi cancels completely.
Selection
* As before there are exclusive trigger regions but this time for Ht
* Trigger efficiency is 100%
* Events with 2+ jets with HT > 200 GeV, pT > 50 GeV, and |y|<2.5 are selected.
* A subsample of these events has an additional 3rd jet that passes the pT and y criteria. So the 3-jet sample is a subset of the 2-jet?
Extraction of R32
* Efficiency in the formula (epsilon) is defined as n selected-events (@HT interval) / total generated events (@HT interval) - it should be 100% as this is signal MC
* Smearing correction: number of selected events (at any reconstructed HT) arising from a given, generated HT interval divided by the total number of selected events in that interval. So.. how many events get selected in the wrong HT bin?
* The uncertainty on the multiplicative correction factor is around 1%
* Figure 2: rises with increasing HT as phase space for third jet opens. Then reaches a pleat which is dependent on the pT, y, and jet selection criteria used. The plateau measured here is 0.8. In D0 it was < 0.4 because of the different CoM energy (was 1.8 TeV in D0 and is 7 TeV in this).
Uncertainties and results
* JES uncertainty here is quoted at ~1%. That’s much better than the other paper, why?
* 3-8% inefficiencies due to MC being not understood
* MG_Pythia 6 does better at low Ht. Then pythia8 - tune 26 does well.
* MG does not describe event shoes and diet angular decor relations well. However, here we care more about the probability that a third jet was emitted than as opposed to the final state of the event shape.
* MG and Alpgen do differently because of the different jet-carton matching parameters used in the MC programs.
Notes (CMS R32 for alpha-s) - 8 TeV CoM
Intro
* This time we’re biking as a function of the average pT of the two leading jets.
* This measurement is experimentally more precise than the 7 TeV one from last time (that was 1.11% and this is 0.87%), theory uncertainties are similar
* Previous measurement was 0.1171, this one is 0.115
* Within perturbative QCD , the x-section of a high-pT scattering process can be expressed a a sum of terms with increasing powers of alpha-s: (the beta-function?)
* They refer to the binning of avg. pT of two leading jets as HT/2 but it isn’t really that…
Selection
* 6 single-jet HLT, with various pT ranges.
* Triggers: the efficiency of each trigger is estimated using the lower-pT-threshold triggers.. how does this work?
* Jet Energy Corrections: reco jets corrected for nonuniformity or non-linearity in detector response. Applied as a multiplicative factor to 4-momentum jet vector. Max. Correction is for smallest pT jets, ~10% for 100 GeV. PU corrections not needed as jets have high pT.
* Events selected with 2+ kets with pT > 150 GeV and |y|<5. Then, events with the 2 leading jets are |y|<2.5 are selected.
* MET < 0.3 as event should be kinematically balanced.
* Why 0.7? And not 0.4? The future prospect paper says that one of ATLAS’s main uncertainties is due to jet size. It’s because of the pile-up range the they don’t need to worry about the cone being narrow due to boost. Neha points out that in the last paper the 0,7 was cimpared to 05 and found that 0.7 was better. Also 0.7 could help when looking for merged jets but there is no mention of trimming.
* It has to be kinematically balanced though
*
R32
* Different kinematic upper limits for the 2 and 3-jet multiplicities
* Here they quote the efficiency (epsilon) parameter as only being due to trigger or jet selection efficiencies. What about the smearing corrections for binning? They take this into consideration in MC by smearing the binning? 10-20% correlations between bins
* They smear the theoretical predictions according to the Jet Energy Resolution.
* Uncertainties cancelling: due to lumi and residual effects. Can clearly see this in figure 2.
* Figure 2: above ~ 1 TeV stats dominate in both measurements. Below, JEC dominates.
Comparing to theory
* Figures 6 and 7
* There is a spread of values even across the same PDF which demonstrates the potential strength of rations Rmn with m-n>1.
Fits
* Chi-squared fit is the same as before to get the coupling constant
* ‘Partial cancellations between other sources of uncertainty are taken into account in the fit’ ??
* The uncertainty on the knowledge of the parameters in each pDF are taken into consideration by varying theme and seeing what the effect is.
* Some interesting notes: on page 17:
* They fit the 2-jet section, then 3-jet, then use that to get alpha s. They also fit the ratio directly. In the 2 and 3 jet case there is a very small chi squared/ndf indicating they’ve overestimated their errors. These values are really small: min 1/6, max 1/2!
* This doesn’t happen for the ratio fit which accounts for the correlations between the cros—sections.
* They speculate that this is vecyas elf the residual uncorrelated errors not actuallycancelling… but they Dion’t explain.
* THEN they just extend the bin ranges from 1 TeV to 1.68 TeV which improves the chi-square.
* They also show table 5 which includes EWK or not corrections and claim the smaller chi-squared WITH is indicative of a better agreement: but it’s still much less than 1!!
* Table 9 shows the final values for this and it’s not too bad
* Table 6 shows the overall range fit
* Table 7 - they barely reference in the text (top of page 18) - what is mu? What os this table?
* Table 8 is like table 9 but showing the two generator tunes.
* Compatible with world average
Becciolini ‘constraining colored matter’ paper (2015)
Intro
* WE can constrain the presence of new color states using a measurement that probes QCD at harder scales.
* This can be used for placing numerical bounds on new physics.
* New colored particles would not affect PDFs at least when discussing taking the ratio of cross-sections.
* It’s the running of alpha-s that’s interesting. It only depends on the mass of the new states, their color representation, and their number.
R32
* In referencing the Run1 CMS R32 paper they warn that the factorization and renormalization scales are identified with the average pT of the two leading jets in the event but that what they’re actually measuring is also 3-jet events that crosses broaden scales and so the assignment based on the pay not represent the dynamics appropriately to ‘allow a straightforward interpretation of the experimental data as a measurement of alpha-s’.
* They talk about the k-factor which is a ratio of the NLO to LO corrections? i.e. the closer to 1 the better.
* Q: How is plot 2 normalized? Why doesn’t it sum to 1? Sea quarks? But anyway, at high pT the quark-subprocesses dominate.
R32 in new physics
* R32 can be changed by a modification of the running of alpha-s, by the PDFs, and by additional contributions to the parsonic cross-section.
* They argue that only the running constant is important and that the R32 and strong relationship for establishing this is reliable.
* First: new physics in hard scattering:
* Tree level contributions will cancel in a ratio measurement
* Threshold effects might not cancel as they have been studied yet in 3-jet observables but have a very small effect on the 2-jet.
* Second running of alpha-s:
* They mention the beta-function which is basically the scale dependence of the strong coupling constant where mu is the mass scale.
* They show how it cane modified to add a number of unknown fermions and group behaviors
* They note that asymptotic freedom is lost when there are more than 10 additional unknown Dirac fermions.
* Figure 4 shows hw alpha-s varies with ‘jet matching scale’ Q
* So we can provide model-independent bounds on new physics parameterized by only number of new fermions and mass scale .
* Third: PDF’s modified:
* In R32 modifications to the PDF can be neglected
* The evolution of the PDF of new fermions above the mass threshold is driven by the gluon pdf but their contribution remains small c omapred to valence quarks and gluons.
* The effect of the presence of new fermions is therefore to reduce the gluon DF by a factor proportional to neff.
* But the quark pDF is less effected and we’ve seen from Figure 2 - that’s the one that counts. The gluon one is identical.
Bounds on new Physics
* Figure 7 essentially gives a light-weight summary of their idea in the hope of inspiring experiemntatlists to do this.
* They note that the scale uncertainties are a problem.
* There are 11 cites for this paper (from 2015) and a few are by ATLAS measuring the strong coupling constant..