Constraining simplified dark matter models with the LHC

Transcription

1 Constraining simplified dark matter models with the LHC Karl Nordström 1 and Thomas Jacques 2 1 University of Glasgow 2 Université de Genève December 18th, 2014

2 Introduction Some brief background: The LHC can investigate dark matter (DM) models where there is some way for the dark sector to talk to light quarks Can roughly separate searches into two types: 1. Model-dependent (SUSY searches,...) 2. Model-agnostic (typically mono-x) I will discuss the models used for settings limits on 2. and present some constraints using monojet (a single jet + ET miss ) limits in particular Karl Nordström 1 / 10

3 Introduction We will assume: Dirac fermion dark matter χ with mass m DM A vector mediator Z with mass M and pure axial-vector 1 couplings g q, g DM L MSDM q g q Z µ qγ µ γ 5 q g DM Z µ χγ µ γ 5 χ (1) This is the interaction term of our minimal simplified dark matter model (MSDM). There are four free parameters (M, m DM, g q, g DM ). 1 The LHC has little sensitivity to vector couplings compared to direct detection (spin-independent vs spin-dependent). Karl Nordström 2 / 10

4 Introduction Can expand the mediator propagator: g q g DM Q 2 M 2 g qg DM M 2 (1 + Q2 M 2 + O ( )) Q 4 M 4 (2) Let Λ = M/ g q g DM, then: L EFT 1 Λ 2 χγµ γ 5 χqγ µ γ 5 q (3) where we have integrated out the mediator using (2). This is an effective interaction term which is valid when Q M. We ve reduced the number of free parameters to two (Λ, m DM ). Karl Nordström 3 / 10

5 Introduction Figure : Example of what we are working with. Karl Nordström 4 / 10

6 Introduction Figure : Example of ATLAS limits on our EFT operator. Karl Nordström 5 / 10

7 EFTs at the LHC To facilitate the comparison to direct detection constraints 2 LHC experiments have generally interpreted model-agnostic searches using EFTs. But: Direct detection: Q O(10 kev) EFT valid always LHC: Q O(1 TeV) EFT valid? Answer ( , others): EFT is only valid for M 2.5 TeV at s = 8 TeV. 2 And also just because there are fewer parameters cheaper, easier. Karl Nordström 6 / 10

8 EFTs at the LHC Figure : Ratio of simplified model to EFT cross-section for g q, g DM = 1 (from ). Karl Nordström 7 / 10

9 Constraining simplified models Some recent studies using the CMS limits and NLOPS predictions for χ χ + 1 jet ( , ) Generally present constrains in e.g. the M m DM plane as an exclusion contour for a particular choice of g q, g DM Note that the minimum width Γ min can be calculated from the input parameters and needs to be taken into account! We have studied constraints using ATLAS limits and LOPS predictions scanning g q /g DM and g q.g DM assuming Γ M = Γ min. Figure : Example of constraints from Karl Nordström 8 / 10

10 Constraining simplified models Karl Nordström 9 / 10

11 Concluding remarks Dark matter is a Big Thing at 14 TeV LHC Need to make sure constraints are robust and can be compared to Direct Detection EFTs are of limited use Simplified Models give a consistent and robust framework at cost of more parameters Karl Nordström 10 / 10

12 Backup slides 1. Implement Lagrangian in FeynRules 2. Generate parton level events with MadGraph5 3. Match to Pythia 8 for showering 4. Perform detector simulation and analysis in ATOM+Rivet 5. Get out visible cross-section, compare to ATLAS limits Karl Nordström 11 / 10

13 Backup slides Since we assume axial-vector couplings the minimal width 3 is: Γ min = N C gdm 2 M(1 4m2 DM /M2 ) 3/2 Θ(M 2m DM ) 12π + N C gq 2 M(1 4mq/M 2 2 ) 3/2 Θ(M 2m q ) 12π q 3 Assuming no additional invisible decays. Karl Nordström 11 / 10

14 Backup slides Karl Nordström 11 / 10