Searches for Dark Matter at the ATLAS experiment

. Searches for strongly produced dark matters in events with jets, photons, heavy-ﬂavor quarks or massive gauge bosons recoiling against large missing transverse momentum in ATLAS are presented. These " E missT + X" signatures provide powerful probes to dark matter production at the LHC, allowing us to interpret results in terms of e ﬀ ective ﬁeld theory and / or simpliﬁed models with pair production of Weakly Inter-acting Massive Particles. Recent ATLAS results on dark matter searches at LHC Run 1 and the connection to astroparticle physics are discussed.


Introduction
Dark Matter existence in the Universe is well established through numerous astrophysical and cosmological observations, as documented in Refs.[1][2][3][4], however little is known of its particle nature or its non-gravitational interactions.At the Large Hadron Colliders, one can search for a weakly interacting massive particle (WIMP), denoted by χ, and for interactions between χ and Standard Model particles, see Ref. [5].Searches conducted at the Large Hadron Collider are especially sensitive at low Dark Matter masses (m χ ≤ 10 GeV), and therefore provide results complementary to direct Dark Matter searches Refs.[6][7][8][9] .
Interaction of particles mediating between Dark Matter and Standard Model particles can be described by contact operators in the framework of an effective field theory (EFT) Refs.[10][11][12] in cases where they are too heavy to be produced directly in the experiment.In the absence of signal, limits can be placed in terms of the effective mass scale of the interaction, M * and of the χ−nucleon cross-section, σ χ−N , as a function of m χ .In addition to the investigation with the EFT operators, pair production of WIMPs is also investigated within so-called simplified models, where a pair of WIMPs couples to a pair of Standard Models particles explicitly via a new mediator particle, e.g a new vector boson Z .In this case, limits on M * and/or m χ are placed as a function of the mediator mass M med .
A number of dedicated approaches to finding evidence for Dark Matter have been carried out with the ATLAS detector experiment Ref. [13] during Run 1, using typically 20.

Mono-X Seaches
Tagging events using a variety of recoil objects, mostly stemming from initial state radiation (ISR), gives access to a broad range of EFT operators, as the respective sensitivity depends on the tag object in question.All cases require large amounts of E miss T , coming from the Dark Matter particle, and tag objects include single jets or photons as well as electroweak bosons Z and W .

Mono-jet+ E miss T
As introduced in Ref. [14], using a single jet as recoil object gives sensitivity to six EFT operators (D1, D5, D8, D9, D11 and D5).To enhance the expected signal, events are required to contain at least one central jet with transverse momentum p T , larger than both 120 GeV and half of E miss T .To ensure that the jet is in fact recoiling against the Dark Matter particles, the angle between the jet and the missing transverse momentum in the events is required to be above one.To further suppress background, which is mainly comprised of Z(νν)+jets and W( ν)+jets events, events containing leptons or high-p T isolated tracks are vetoed.As all measurements are consistent with Standard Model expectations, the most sensitive out of nine signal regions, defined by requirements on E miss T ranging from 150 GeV to 700 GeV, is used to place limits on σ χ−N for each of the operators under investigation.As an example result of this analysis, inferred 90% confidence level (CL) limits on σ χ−N as a function of m χ for the spin-independent as well as spin-dependent case for different operators are shown in Fig. 1  and top-right).To ensure the validity of the EFT approach, the results are also shown after applying a truncation procedure as described in Ref. [14].Limits on M * as a function of M med in the context of a simplified model are shown in Fig. 1 (bottom).
[GeV]  for a coupling strength √ g f g χ of unity or the maximum value (4π) that keeps the model within its perturbative regime.The truncation procedure is applied for both cases.Observed lower limits at 95% CL on M * as a function of M med (bottom), for a Z -like mediator with vector interactions.For an m χ of 50 GeV or 400 GeV, results are shown for different values of the mediator total decay width Σ and compared to the EFT observed limit results for a D5 (vector) interaction.M * vs m V contours for an overall coupling √ g f g χ = 0.1, 0.2, 0.

Mono-photon+ E miss T
A search using a photon as tag object, allow to access three EFT operators (D5, D8 and D9), has been shown in Ref. [31].Using events with a single highly energetic photon, large E miss T , no leptons and at most one jet; the main backgrounds in this analysis remain Z(νν)+γ, diboson, Wγ and Zγ with lost leptons as well as W and Z production with leptons misidentified as photons.As for the mono-jet search, all measurements are consistent with Standard Model expectations and lower (upper) limits on M * (σ χ−N ) are presented both with and without applying the EFT truncation procedure mentioned above, as shown in examples in Fig. 2 (top-left and top-right).In addition limits on M * as a function of M med , as shown in Fig. 2 (bottom), are derived in the context of a simplified model.

Mono-W/Z+ E miss T
An analysis using W and Z bosons, respectively their decay products, as tag objects gives sensitivity to four EFT operators (C1, D1, D5 and D9) and has been presented in Refs.[37][38][39], for both hadronic and leptonic decay of the vector bosons.In the analysis aiming at hadronic decays, events are required to contain at least one high-p T large-radius jet with reasonably balanced sub-jets, originating from the vector boson; at most one additional regular jet; and no leptons and photons.The background yield in the two signal regions, defined by requirements on E miss T of 350 GeV and 500 GeV, is dominated by Z(νν)+jets as well as Z( )+jets and W( ν)+jets with lost leptons.Looking at the leptonic decays, the event selection differs for W or Z bosons.In the first case, events are required to contain exactly one high-p T lepton and a transverse mass of the W boson candidate incompatible with direct production; while in the latter case, events have to contain two leptons giving an invariant mass close to Z peak and no additional lepton or jets.Both cases require large values of E miss T and the main background contributions are coming from diboson events in the Z case and in addition W( ν) and Z( ) with lost leptons in the W case.All yields are consistent with Standard Model expectations and limits on M * as a function of m χ and σ χ−N , are presented both for spin-independent and spin-dependent EFT operators, as shown in Fig. 3.    [6-9, 29, 30, 32, 34, 35, 37, 38].Plots taken from Refs.[37][38][39]

Heavy-quarks Searches
A search for Dark Matter pair production in association with bottom or top quarks has been presented in Ref. [40].Aside from being sensitive to three EFT operators (C1, D1 and D9), this analysis also places constraints on the mass of a coloured mediator suitable to explain a possible signal of annihilating Dark Matter, using a simplified model model approach.Several signal regions are defined requiring combinations of increasing jet and b-jet multiplicity, 0 or 1 lepton and values of E miss T above 200 GeV to 300 GeV in the events.Applying additional kinematic cuts, the main backgrounds is still coming form t t-events as well as single-top and W/Z+jets events.All measurements are consistent with Standard Model expectations and lower (upper) limits on M * (σ χ−N ) are presented for three EFT operators (C1, D1 and D9), as shown in Fig. 4 for D1operator (top-left and top-right).Due to the proportionality of the scalar operator to the quark mass, limits for D1 are in fact better than those obtained by the above mentioned mono-jet analysis.Constraints on b-flavoured Dark Matter models, using a simplified model, are also presented and shown in Fig. 4 (bottom).For a Dark Matter particle of about 35 GeV, as suggested by an interpretation in Ref. [41] of data recorded by the Fermi-LAT collaboration see Ref. [42], mediator masses between roughly 300 GeV and 500 GeV are excluded at 95% CL.
[GeV]  .Solid lines and markers indicate the validity range of the effective field theory assuming couplings g q g χ < 4π, the dashed lines and hollow makers represent the full collider constraints.Upper limits at 90% CL on σ χ−N for the scalar operator D1 as a function of m χ (top-right) compared to other results.The coupling is assumed to be g q g χ = 4π.Exclusion contour at 95% CL for the b-flavoured Dark Matter model (bottom) from combined results of two signal regions.The expected limit is given by the solid red line.The region beneath the curve indicating the observed limit is excluded.For details see Refs.[6,7,40,43].Plots taken from Ref. [40]

Conclusions
The ATLAS Collaboration has performed a broad variety of searches for Dark Matter signatures, using √ s = 8 TeV Run 1 data and with tag objects ranging from single jets, photons, to W/Z bosons and as well as heavy quarks .No signs of Dark Matter have been observed, and stringent limits have set on the different benchmark models, emphasising the complementary nature of collider searches to direct and indirect detection experiments, especially at low Dark Matter masses and for spin-dependent EFT operators.
3 fb −1 of data collected at a centre-of-mass energy of √ s = 8 TeV.So-called mono-X searches take advantage of a variety of different tag objects, X, together with large absolute values of missing transverse momentum, E miss T , in the final state to constitute a clean and distinctive signature.Heavy-quark searches use events with large E miss T in association with high-momentum jets of which one or more are identified as jets containing b or top-quarks.a e-mail: habreu@cern.ch,

Figure 2 .
Figure 2. Limits at 90% CL on M * as a function of m χ (top-left), for the vector operator D5 .Results where EFT truncation is applied are also shown, assuming coupling values √ g f g χ = 1, 4π.Upper limits at 90% CL on σ χ−N as a function of m χ spin-independent (top-right-left) and spin-dependent (top-right-right) interactions,for a coupling strength √ g f g χ of unity or the maximum value (4π) that keeps the model within its perturbative regime.The truncation procedure is applied for both cases.Observed lower limits at 95% CL on M * as a function of M med (bottom), for a Z -like mediator with vector interactions.For an m χ of 50 GeV or 400 GeV, results are shown for different values of the mediator total decay width Σ and compared to the EFT observed limit results for a D5 (vector) interaction.M * vs m V contours for an overall coupling √ g f g χ = 0.1, 0.2, 0.5, 1, 2, 5, 4π

Figure 4 .
Figure 4. Lower limits on M * at 90% CL for different signal regions as a function of m χ (top-left) for the operator D1.Solid lines and markers indicate the validity range of the effective field theory assuming couplings g q g χ < 4π, the dashed lines and hollow makers represent the full collider constraints.Upper limits at 90% CL on σ χ−N for the scalar operator D1 as a function of m χ (top-right) compared to other results.The coupling is assumed to be g q g χ = 4π.Exclusion contour at 95% CL for the b-flavoured Dark Matter model (bottom) from combined results of two signal regions.The expected limit is given by the solid red line.The region beneath the curve indicating the observed limit is excluded.For details see Refs.[6,7,40,43].Plots taken from Ref.[40]