Last night the ATLAS Collaboration released its latest search for dark matter and other beyond the standard model theories  based on the full dataset from the LHC Run I (2010-2012).
By looking for proton-proton collisions where jets of hadronic particles are produced only in one direction (Figure 1), violating conservation of momentum only in appearance, we use ATLAS to search for weakly interacting massive particles (WIMPs), such as dark matter particles. Because they are weakly interacting, the WIMPs escape ATLAS undetected and lead to what looks like missing momentum.
Figure 1: A proton-proton collision recorded by ATLAS in November 2012, viewed in the transverse plane perpendicular to the beam axis. It shows a 1.2 TeV jet of hadronic particles (tracks and green bars) on one side, one can note the absence of activity on the other side, mimicking a violation of momentum conservation by as much as 1.2 TeV.
Neutrinos do exist and are often produced in LHC proton-proton collisions. These collisions look exactly like dark matter candidates and make up most of the background. We need to precisely predict the expected number of background collisions with outgoing neutrinos and compare that with the observed number of events in ATLAS data. Only a significant excess over the expected background allows a claim for a new physics signal.
The new ATLAS result  explores collisions with very high missing momentum, beyond any previous LHC search and we observe a good agreement between the number of events actually detected and the background alone prediction.
How well do we know the backgrounds? This is the most central question of this work. The sensitivity to dark matter is dominated by systematic uncertainties on the background. This is why at Stockholm University we decided to focus on the background calculations.
In the fall of 2014 Olof Lundberg, PhD student at the Stockholm university physics department, defended his licentiate thesis  where in particular he presented the development of a new technique to compute the backgrounds. All the gory details of the calculations are in his licentiat, all the work that went into defining and understanding our new background control region and the mechanics to extrapolate to the signal regions. This 1.5 year long effort has really made a difference. With our new control region we were able to significantly increase the sensitivity to dark matter and other exotic signals. Figure 2 shows our new limits on the WIMP-WIMP annihilation cross section as function of the WIMP mass for various explored models.
Figure 2: Exclusion upper limits on WIMP-WIMP annihilation cross section as function of the WIMP mass, from the ATLAS monojet analysis  in various signal scenarios. D5 corresponds to an effective field theory where the WIMP dark matter is a Dirac particle interacting via a massive spin-1 vector particle. D8 corresponds to a scenario where the WIMP would interact via an axial-vector interaction. This graph also illustrates the complementarity with astrophysical searches for WIMP-WIMP annihilation with HESS and FERMI-LAT.
The paper came out last night, but to tell the truth the data analysis has been ready and unblinded for almost six months, albeit embargoed at the time of Olofs licentiat thesis. The theoretical interpretation took a long time. To be able to translate ATLAS absence of new physics signal into WIMP-WIMP annihilation cross sections a specific model has to be used. Several models have been investigated, Figure 2 is based on an effective field theory which is only valid under certain assumptions. The validity of the approach depends on the momenta of the initial partons involved in the proton-proton collisions and the exchanged momentum. In the end we chose a very conservative approach and for Figure 2 we simply assumed that we had zero sensitivity to dark matter signals when the validity limit was broken. This weakened our limits but at least the limits feel more robust.
We are of course disappointed we did not find anything new. As Figure 2 shows, a WIMP with a mass below 20 GeV and interacting via a vector particle could no longer explain on its own the whole relic dark matter density provided to us by fits to cosmological data.
On the other hand our little Stockholm monojet analysis team: Gabriele Bertoli (grad student), Olof Lundberg (grad student), Valerio Rossetti (postdoc), Christophe Clément (faculty) is stronger than ever and we are already working on new developments for the upcoming 2015 ATLAS data. This spring LHC is restarting stronger than before, with a much higher center of mass energy and much more data. So we are certainly looking forward a very exciting second ride with the ATLAS data.
– Christophe Clément
 ATLAS Collaboration, Search for new phenomena in final states with an energetic jet and large missing transverse momentum in pp collisions at √s=8 TeV with the ATLAS detector; http://arxiv.org/abs/1502.01518 (submitted to Eur. Phys. J. C).
 Olof Lundberg, Searches for Dark Matter and Extra Dimensions in Monojet Final States with the ATLAS Experiment, Licentiat thesis Stockholm University, October 2014.