CERN Accelerating science

Exploring dark sector with NA64: First results from the combined analysis of 2016-2018 runs

Searches for dark matter (DM) particles, an essential experimental goal for particle physics, are entering in an interesting stage. Despite the strong astrophysical evidence for dark matter, little is still known about the origin and dynamics of this substance of our Universe.

The significant efforts on both direct and indirect searches at the LHC and other dedicated experiments around the world has ruled  out many DM models, but  have not yet yielded unambiguous DM signals. Perhaps, the main difficulty that hinders progress, keeping the mystery around DM, is that it can only be studied through its gravitational interaction with ordinary matter. Therefore, the idea of introducing a new way of interaction between dark and visible matter, in addition to gravity, opens a rather exciting path for exploration.

Several theoretical models  motivate the extension of the SM by introducing the concept of  dark sectors consisting of a set of particles and fields which is similar to that of the SM, but transform under new gauge symmetries and do not interact with the SM.  In analogy with QED, for which the massless photon mediate the force between charged particles, there could be a dark QED with a so-called dark photon that mediates a force between dark particles.  Such a dark photon (A ´) could have  a mass mA ´≲ 1 GeV and it could  couple to the SM through the kinetic mixing with the ordinary photon parameterized by the mixing strength ε.

In the presence of light dark matter particles, with masses < m, the dark photon  could predominantly decay invisibly into those particles. Models introducing such invisible A´ decays offer new intriguing possibilities to explain hints on astrophysical signals of dark matter. They assume the existence of  a dark photon with a mass in the sub-GeV region and a coupling strength to ordinary photons in the range of ε ≃10 −5-10−3.

This assumption has motivated a wide range of ideas and proposals that were widely discussed during the Physics Beyond Colliders annual workshop at CERN. Proposals include ideas for dark forces and other portals between the visible and dark sectors which might be probed at low energy, high-precision experiments at CERN’s SPS.

One of the experiments that attracted noticeable attention during the PBC meeting was NA64. It was designed for a broad search for dark sector physics by looking for missing energy events using the active beam dump approach.  NA64 was approved in 2016 and its primary goal was to search for invisible decays of dark photons, covering a large part of the parameter space for exploring the observed  muon g-2 anomaly.

The basic idea is the following,  if dark photons (A’) exist they can be produced via the kinetic mixing with bremsstrahlung photons through the reaction eZ → eZA´ of high-energy electrons scattering off nuclei (Z) from an active target. This will be followed immediately by the decay of A´  → χ χ into dark matter particles χ. The χ would penetrate a hermetic detector located downstream the target without interacting and would carry away a fraction of the beam energy.  Therefore, observing an excess of  events with large missing  energy above small backgrounds would signal the presence of dark photons. Moreover, to ensure that any excess is due to the 100 GeV electrons, the emitted synchrotron radiation from the beam is used to tag each electron individually and beyond any doubt.

The technique of observing missing energy in the products of high-energy interactions  currently being explored by NA64 on an electron beam was put forward during the Physics Beyond Colliders kick-off workshop in 2016. This approach complements classical beam dump experiments and provides much better sensitivity for a specific parameter space. In the case of beam dump experiments, the products of dark photon decays ( i.e. other dark sector particles) could  penetrate the dump and have observable interactions in a far detector.

Figure 1. Na64 limits on the γ - A’ mixing from the combined 2016-2018 runs.



NA64 began operation in July 2016 for a period of two weeks and completed its physics programme with a six-week run in 2018.  The July results were published in Phys. Rev. Lett. (2017) and exclude values of the dark photon (Aʹ) coupling that would make dark photons relevant for explaining the muon g-2 anomaly.  While not ruling out the existence of dark photons, these results serve as an important demonstration of the feasibility of the NA64 approach and give guidance for the further upgrade of the detector. Significantly more data accumulated in 2017-2018 with an improved apparatus  allowed to search for A´s as a mediator of dark matter production in invisible decay mode and other invisible decays of dark-sector particles. The combined NA64 bounds from 2016-2018 runs on the γ - A’ mixing are shown in Fig. 1, while new limits  on the parameter Y are shown in Fig. 2. This variable characterizes the cross section for the DM <-> SM annihilation and is convenient  for comparison of results from other experiments, also shown in Fig. 2. The black curves calculated for the scalar, Majorana and Pseudo-Dirac DM relic abundance represents the Y parameter space, which is an explicit target for NA64.  

Figure 2. Na64 constraints on sub-GeV DM from the combined 2016-2018 runs.

One  can see that NA64 provides  the most stringent constraints on both variables, ε  and Y,  for the A’ and DM masses in the sub-GeV regime. Thus, demonstrating the power of the active beam dump approach for the dark matter search compared to the classical beam dump experiments such as LSND, E137, and MiniBooNE.  The further goal of NA64 is to accumulate up to ~1013 electrons on target (EOT) after LS2 in order to probe  the remaining DM parameter space shown in Fig. 2.


The search for dark photons  is one of many approaches for trying to detect dark matter. Today we have a rapidly growing variety of models suggesting new particles weakly coupled to lepton and/or quarks solving the DM and other problems in particle physics. In this unusual situation, NA64 has proven great flexibility providing a quick response in testing different models.


For example, in 2016  the experiment of Krasznahorkay et al. in ATOMKI has reported observation of a 6.8 σ excess of events in the invariant mass distributions of e+e− pairs produced in the 8Beexited state nuclear transitions to its ground state accompanied by an emission of an e+evia internal pair creation.  It has been shown that this anomaly can be interpreted as an emission of a new gauge boson (X) followed by its prompt decay X → e+e and provide a particle physics explanation of the anomaly that is consistent with all existing constraints based on the assumption that its coupling to electrons is in the range of 2×10 -4< ε  < 1.4×10 -3 and mass mX= 16.7 MeV. The models predict relatively large charged lepton couplings εe ≃ 0.001 that can also resolve the  muon g-2 anomaly.


In  2017-18 after a  detector reconfiguration  NA64 had a physics runs to obtain data for searching for the X and A´ → e+e decays  with 100-150  GeV electrons at the H4 line in CERN’s North Area. The preliminary combined results excluded a significant part of the X boson parameter space are shown in Fig. 3. The SPSC recommended the continuation of A´  and X searches with data from the 2021 run.


Another example is related to the muon (g-2)μ  anomaly, which is under more accurate study by the currently running experiment E989 at FNAL. While the A´ explanation of the muon g-2 anomaly has been ruled out by NA64 and BaBar, there is still one remaining interesting explanation provided by the Lμ – Lτ model.



Figure 3. Na64 preliminary combined results from the 2017-2018 runs on the search for the 16.7 X boson from the 8Be anomaly and A´ -> e+e decays of dark photons.

In this model  the (g-2)μ   discrepancy is explained by the possible existence of  a new leptophobic sub-GeV Z´ boson, which is weakly coupled predominantly to the second and third lepton generation leptons. The surprising fact is that the SM extension which includes such Z´ is still renormalizable and no new fermions are required.  In addition, the invisible dark Z´ boson could serve as a portal between the visible and dark sectors offer new intriguing possibilities to explain hints on dark matter.

These results motivated the NA64 Collaboration to  propose a new experiment called NA64μ aiming to search for Z´ with the M2 muon beam at the SPS. Interestingly, NA64μ could also probe the  dark photon (A´) invisible decay similar to NA64 running in the electron mode (NA64e).  The combined NA64e and NA64μ results  with ~1013 EOT and  a few 1013 MOT, respectively,  will allow to cover  almost fully the parameter space of the most interesting  sub-GeV DM models shown in Fig. 2 . This makes NA64e and NA64μ  extremely complementary to each other and  greatly increases the exploratory power of sub-GeV DM searches with CERN’s SPS.


Further Reading

NA64 Collaboration,  Phys.Rev.Lett. 118 (2017), 011802

NA64 Collaboration,  Phys.Rev.Lett. 120 (2018), 231802

NA64 Collaboration,  Phys.Rev. D97 (2018), 072002

NA64 Collaboration,  CERN-SPSC-2018-024 / SPSC-P-348-ADD-3,  13/10/2018