Choose timezone
Your profile timezone:
Powered by Indico
v3.3.2-pre
Minutes from 23/04 Meeting
Introduction
===========================
Main open points before Forum conclusion:
* signal uncertainties for Madgraph and Powheg
* validity ad completion of EW/EFTs
* ATLAS/CMS convergence on single top parameters/scan
* mono-Higgs models
Timeline:
* Distribute document summarizing signal choices only towards MC production, on Monday
* Prepare draft of full write-up by May 14th (cut-off for plot/material inclusion), arXiv submission ~2 weeks after
EFT validity - J. Alcaraz
===========================
##Talk
Basic idea for EFT validity
* quantify how wrong we are if we use the EFT naively, using e.g. fraction of valid events
* truncate to make experimental searches conservative if we're unacceptably wrong
This is not a new issue, and it appears in other BSM models. The differences are:
* in other BSM models, we focus the discussion on the unitarity violation of the x-section (in some cases cured by depletion of cross-section rather than setting cross-section set to zero), here we talk about whether EFT is an accurate description of the kinematics
* potentially invalid EFT appear in WIMP-nucleon scattering plots
First question: is the cross-section arbitrarily large? 1403.6610 contains most of the constraints for DM lagrangians.
From this paper, g cannot be larger than 5 (4 in scalar case). For V/A Q<5.1*Lambda (for mDM<<Q), which may be a lower limit wrt the one we would use in the truncations and it's lower than the usual qualitative limit of g<4pi.
Second question: validity of EFT approach and the kinematics it simulates.
No treatment of this in CMS until CMS-EXO-14-004, where limits are shown compared to the contours with the fraction of valid events: Q_tr < g_eff*Lambda. This is not very different wrt what ATLAS has been doing, but it does not work for shape analyses.
For shape analyses, one should redo the analysis dropping events with Q_tr<M(mediator). This leads to a mass-dependent analysis. However most LHC searches are mass-dependent already, and often Q<M is not really high MET.
##Discussion and summary
The main open point from this talk is whether to recommend to include the unitarity bound on the Lagrangian terms rather than the qualitative ones, and how.
EFT validity proposal for DM Forum: experimentalists only provide high-mass mediator results (following the simplified model m_DM scan) rather than EFT operator results in the cases where simplified models are available, let theorists reinterpret those if they need a contact interaction limit. Using an EFT in other regimes would not represent the kinematic of the signals we may see: where the mediator is produced resonantly the EFT limit is weak, while at low mediator masses the EFT limit would be too aggressive. This proposal is based on http://arxiv.org/abs/1308.6799 and on the experimental plots comparing EFT and simplified model limits (e.g. https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PAPERS/EXOT-2013-13/fig_11a.png): at high mediator masses, there is no difference for the s-channel between simplified models and EFT benchmarks. A reinterpretation of the various simplified models in terms of EFT operators will be possible but left to theorists. In order to find the mediator mass to be simulated for this purpose, one could use the truncation procedure and set the mediator mass to the point where EFT leads to 100% valid events. One can also provide a 2D-version of the plot in fig. 11a above, and point out the region where the EFT limit is valid.
Further agreements from the discussion:
* in the full Forum write-up, we are describing the truncation procedures, and we will recommend generator-level removal in case of shape analyses. The truncation / event removal procedures may be still useful when there are no simplified models available for a given EFT (e.g. electroweak boson models with direct DM-boson couplings). In those cases however the completion may proceed via loops, and those procedures and formulas are not directly usable as done so far.
* the comparison with ID/DD should not be done in case of EFTs, and the presentation of results will be discussed as one of the follow-up points when the Forum works on simplified models and EFT are concluded.
b-flavored simplified model - B. Penning
===========================
This model from 1404.1373 has already been used in an ATLAS analysis (1410.4031), weighted to relic density.
Suggestion for parameter scan:
* scan points in mMed, mDM, up to mDM ~ 300 GeV;
* have a grid normalized to the relic density, so that theorists don't have to deal with interpolation - it still has to be seen if shapes change wrt other points or it's just a change in cross-sections.
* perform a scan in g and test the kinematic shapes;
The cross-section increases significantly between 8 and 14 TeV - potential reasons (not yet final):
* b-quark in PDF
* relic density constraints less severe
the 8 TeV cross section can nevertheless be reproduced
Work still in progress (also working on HF+DM), will provide a parameter scan proposal as soon as possible.
Parameter scan for HF+DM models (scalar) - K. Hai Sung
===========================
Details of spin-0 mediated ttbar+DM model:
* scalar/pseudoscalar mediator phi
* Dirac fermion DM particle
* Universal gSM=gDM, multiplicative "yukawa" for gSM
Considering sensitivity as well: need o(1) events, 30 inv fb, efficiency 1%, make sure couplings are perturbative.
Kinematics explored within the sensitive region in terms of masses/coupling type/coupling strength:
* the largest kinematic differences are seen in variations of the mediator mass, especially if the mediator is off-shell.
* scalar and pseudoscalar differ in:
* cross-sections (the exponent in the width is different)
* kinematics below the ttbar threshold (however, this is coincidental and driven by the coupling of the top pair to the mediator)
* no difference when changing the coupling strength away from the resonance region, but for larger couplings the width is larger and PDF suppression will kick in.
The summary proposal for the scan is 68 grid points (still iterating to harmonize with monojet):
* Generate scalar and pseudoscalar separately
* Mmed<1500 GeV
* off-shell region: include mPhi = 2m_DM threshold, otherwise select m_DM diagonal
For the smallest m_DM value, consider the sensitivity of DD experiments: 10 GeV (current) and 1 GeV (future)
Also produce table for xsec/width to:
- demonstrate validity and limit of simple scaling
- help model reinterpretation (with MG settings)
The final version of the write-up tex will be available in the order of a few days.
Single top models and write-up - R. Madar
===========================
Two models:
- resonant: with a vertex involving SM and DM at the same time (not necessarily a mediator in the strict sense of the word)
- non-resonant: production of top+mediator, where "mediator" decays into DM
What has been included in the write-up attached:
- study of the width effects for non-resonant model.
- potential of searches for visible decay of mediator (however, outside of DMF scope)
Open points to be agreed upon between ATLAS and CMS:
* Non-resonant: agreed upon
* Resonant:
* How to set the total width? Assume only DM decays?
* Does the width impact the kinematics? No for the non-resonant model, no results yet on the resonant model. This may have to be left for a later date wrt Forum timeline.
Mono-Higgs (verbal) update - L. Kashif
===========================
Ironed out some bugs during communication between ATLAS and theorists, proposal sent to the mailing list, needs CMS feedback.
Initial observations: using the vector baryonic Z' models (higher xsec), only one parameter seems to make a difference, the coupling between H and Z'. We would like to proceed with the scan that has been already proposed in previous meetings. We need to converge on this over the course of next week, so we'll discuss / have a dedicated Forum meeting as soon as possible.