Prospects for New Physics at the LHC Michael

Prospects for New Physics at the LHC Michael Schmitt Northwestern University PANIC 08 Eilat, ISRAEL 10 – Nov – 2008 10 -Nov-2008 New Physics at the LHC (Schmitt) 1

This is a vast topic. And I have just 20 minutes! I will discuss just three analyses, and - hopefully - give some real content… intro: CMS and ATLAS • Z' → m+m • H → gg • jets and missing energy concluding remarks 10 -Nov-2008 New Physics at the LHC (Schmitt) 2

CMS ALICE LHC-b ATLAS Drs. Funk and Morettini will have discussed this in the prior session. 10 -Nov-2008 New Physics at the LHC (Schmitt) 3

CMS 10 -Nov-2008 New Physics at the LHC (Schmitt) 4

ATLAS 10 -Nov-2008 New Physics at the LHC (Schmitt) 5

Z' → m+m. Many theories of new physics predict a new gauge boson which will decay to fermion pairs, like the usual Z boson. Also, in extra-dimensional theories, spin-2 excited states of SM gauge bosons can decay to fermion pairs. Masses, couplings, cross sections, branching ratios are all unknown a priori (there are some theory constraints). We can observe these new bosons as resonances in di-fermion invariant mass distributions. The best result comes when the peak is narrow. This means electrons and muons are best. 10 -Nov-2008 New Physics at the LHC (Schmitt) 6

Elements of the Analysis: ATLAS & CMS have endcap muon systems – good acceptance. • select a pair of high-p. T opposite-charge leptons • make sure kinematics are well measured Drell-Yan continuum, boson pairs, top quarks, fake leptons in jets • none of these gives a peak in di-lepton invariant mass • fake leptons can be strongly suppressed by demanding that the selected leptons not fall inside a jet – “isolation” Main background will be continuum di-lepton pairs. Look for a peak on top of a smooth background. 10 -Nov-2008 New Physics at the LHC (Schmitt) CMS 7

Results / Sensitivity : Assume resonance is relatively narrow. Test: are the data compatible with a peak at an arbitrary mass? Since couplings are indeterminate, take some benchmarks. CMS What can we see with 100 pb-1? 10 -Nov-2008 Surpass the Tevatron… New Physics at the LHC (Schmitt) 8

Discussion : Assume resonance is relatively narrow. → Then the width of the bump comes from detector resolution; a narrower peak means better signal/background. linear For muons, this means primarily alignment: point errors translate into curvature, ie, p. T errors. “Prospects for Discovery” → getting a good alignment early! 10 -Nov-2008 New Physics at the LHC (Schmitt) 9

Alignment Methodology: 0) use photogrammetry to get excellent starting point. 1) use lasers and sensors to refine this, and to track movements of gross structures 2) use reconstructed tracks for ultimate precision cosmic rays ATLAS 10 -Nov-2008 New Physics at the LHC (Schmitt) 10

H → gg This is a SM process – that has never been observed… Theoretically novel – decay proceeds only through a loop, which could reveal new particles indirectly (in principle). In the SM, the rate is firmly predicted (as function of MH). And in SUSY, variations are small – lightest Higgs is SM-like. Main production is from gluons (another loop!) with important contributions from VBF and tt. H production. BR, however, is very small, typically 0. 002. (MH < 140 Ge. V) eff. cross-section = (30 – 50 pb) x 0. 002 = 60 – 100 fb. 10 -Nov-2008 New Physics at the LHC (Schmitt) 11

Elements of the Analysis: As before, look for a narrow peak on top of a continuum. • two well-reconstructed photons, high ET > 25, 40 Ge. V The background is huge!! irreducible – direct production of 2 photons w/o Higgs reducible – one of the photons is from p 0 or is fake Use isolation and shower shape to suppress bad photons. photon ATLAS jet fraction outside core 10 -Nov-2008 strip width New Physics at the LHC (Schmitt) 12

Interesting Extension: Really try to remove all background – pull out all the stops! Combined kinematic quantities in an artificial neural network: • polar decay angle • scaled photon energies • difference in h • longitudinal momentum of di-photon pair Standard “cut-based” analysis simply looks for a peak; - hence, is lett model-dependent; This approach assumes the model and exploits all features wisely. → sensitivity improved by about 30% The two approaches are complimentary. 10 -Nov-2008 New Physics at the LHC (Schmitt) 13

Results / Sensitivity : Although signal is fairly well “known” the background is not, so discovery potential estimates comes with a lot of uncertainty. CMS NN Fortunately, sidebands are a ready source of background events → directly measure the rate, and kinematic characteristics 10 -Nov-2008 New Physics at the LHC (Schmitt) 14

Energy Resolution : Again – we want as narrow a peak as possible. CMS comparison test-beam and CR: 1. 5% (need 0. 5%) 1) EM calorimeter must be well-calibrated. • basic calibration done with test beam • inter-channel calibration done with inclusive p 0’s and muons (cosmic rays!) • absolute scale will be set from W → en and Z peak 2) The amount of material in front of calorimeter is crucial • more than one X 0 ! • 40% of the photons in H → gg convert. • quality degrades when shower is spread out laterally • recuperation is possible – not fully developed yet • divide samples according to photon quality. 10 -Nov-2008 New Physics at the LHC (Schmitt) 15

Jets and Missing Energy Supersymmetry = theorist’s best guess about BSM. LHC might turn out to be the machine that finds SUSY. Characteristic feature: LSP is weakly-interacting & neutral. Best place to find SUSY is probably in a sample of Jets+MET. Signatures vary widely, difficult to specify and hence most analyses pick one possibility and optimize for it: • two jets and MET • three or more jets, and MET • jets and one lepton, and MET • etc. 10 -Nov-2008 New Physics at the LHC (Schmitt) 16

Elements of the Analysis: Reconstruct jets from calorimeter energy clusters. • demand at least 3: ET > 30 Ge. V and | h | < 3 • MET > 200 Ge. V • ET 1 > 180 Ge. V, ET 2 > 110 Ge. V, HT > 500 Ge. V Event Cleanup: • vertex reconstructed from charged tracks CMS cosmic ray data! • fractions of “electromagnetic” and “charged” energy not too large. 10 -Nov-2008 New Physics at the LHC (Schmitt) 17

Veto events with leptons (loose criteria). QCD 3 -jet events have large MET usually due to one badly mis-measured jet: • MET vector should not point close to any jet. CMS QCD 10 -Nov-2008 SUSY New Physics at the LHC (Schmitt) 18

Results / Sensitivity : Even though the selection is fairly straight-forward, the expected signal varies greatly. Here are two “benchmarks scenarios” : 10 -Nov-2008 New Physics at the LHC (Schmitt) 19

Interesting Model-Independent Approach : Promising ATLAS study* looks for statistical excess anywhere in a predefined kinematic space. Kinematic quantities should be well suited to the generic channel. Construct a sphere around a given data point and compare number of data events to expectation (simul’n). No use of signal simulation at all. Performance appears to be as good as “targeted” analyses, yet much less model-dependent. 10 -Nov-2008 New Physics at the LHC (Schmitt) (* E. Duchovni et al. , Weizmann Institute) 20

Concluding Remarks You have seen a brief overview of three selected analysis – they are already quite advanced; others, too. Prospects for discoveries will be good since so much is being done in advance of collisions… 10 -Nov-2008 New Physics at the LHC (Schmitt) 21

10 -Nov-2008 New Physics at the LHC (Schmitt) 22