For a few years, the work of the LHC community has been split between the exploitation of the LHC Run2 and first Run3 data and the preparation for the High Luminosity (HL) phase.  At the HL-LHC, the effort of searching for new physics will be rethought: with no significant collision-energy increase at hand, we will be mainly searching for new physics on the bulk of the Standard Model distributions. Precision measurements will be the main tool to put indirect bounds on new physics energy scale, e.g., with effective field theory (EFT). This transition already started with the data analysis of the full Run2, and eventually Run3, datasets.

While we are preparing for this step up in data analysis challenges, we should not dismiss the role of Run3 in our quest for new physics. At ICHEP 2022, as a side dish to the celebration of the 10th anniversary of the Higgs discovery, a few anomalies in CMS and ATLAS data were reported, which attracted the interest of part of the theory community. 

A few searches for heavy Higgs bosons reported moderate excesses for a new boson X with mass around 650 GeV. Small evidence for a signal was reported by multiple analyses in the so-called vector boson fusion category. These excesses would indicate the possibility of a new boson being produced in the fusion of two vector bosons, resulting in the associated production of the X boson with two jets with large rapidity gap. This new particle would then decay to final states like γγ and ZZ (considered by ATLAS) or WW (considered by CMS).

The extension of H → ττ search to a mass range smaller than 125 GeV resulted in a small signal hint for masses between 90 and 100 GeV, aligning to an excess seen at 95 GeV in γγ final states (using only 2016 dataset). This would potentially point to a new Y boson decaying to fermion or boson pairs. Some theorists speculated that these excesses were supporting an excess seen in H → bb searches at LEP. On the other hand, this speculation is not supported by LEP runs at higher energies, which dismissed this excess already. Nevertheless, the case for a light Y boson is intriguing. Unfortunately, one cannot combine these two results without introducing unjustified speculations on the relative branching ratio in the two final states. As a result, the overall evidence is still too low to support any strong statement. Obviously, the presence of the Z boson in the same mass range calls for extra caution before any claim can be made. The improvement of rejection techniques for Z → ττ and Z → ee backgrounds is indeed a topic of active research.

An intriguing result was reported by CMS in the search for a new heavy boson X, decaying to a light boson Y and to a Higgs boson. When reconstructing H in its clean gamma gamma final state and Y in bb, an excess was observed. Surprisingly, the excess would correspond to a X mass of ~ 650 GeV and a Y mass of ~ 95 GeV. This would align to the above excesses and potentially connect them in a single framework.

While such a list is certainly intriguing, it should be taken with some grain of salt. One cannot interpret statistically such a set of excesses without having a model behind, cherry picking results with excesses while not considering those without. In other words, all these excesses cannot be combined trivially to make any discovery claim.

According to several publications appearing recently in the literature, CMS and ATLAS should further investigate these excesses. Some of these studies pointed to specific scenarios that had already predicted the existence of a new boson at 650 GeV. Others pointed to generic extensions of the Higgs sector that could accommodate most of these excesses. Certainly, the alignment of the excesses is intriguing. On the other end, one should keep in mind that a search spanning a high-dimensional space of possibilities (the mass values for X and Y and their relative couplings to various standard model particles) would be affected by a large dilution in sensitivity, due to the so-called “look elsewhere” effect. With so many potential signals that this search could be sensitive to, any statistical fluctuation in the background could result in a false claim.  This is why the bare significance (the so-called local significance) is not solid enough to claim a discovery.

In CMS, this activity is progressing on two fronts. On one hand, our immediate goal was to use Run2 data to complete the picture. We are now  exploiting the entire Run2 data to update the other studies, for which we only have partial Run2 results (typically using only data collected in 2016). With the objective of providing independent measurements, we are giving priority to those on which ATLAS reported an excess.

A first outcome of this effort was the release of the full-Run2 search for Y → γγ, presented at Moriond EW in 2023. A few months afterward, ATLAS released their result. While an excess as large as the one seen in CMS could not be confirmed, ATLAS data show a mild excess in the very same mass range. This fact generated some speculation (and some attempts to combine the two analyses). While any conclusion is premature, the case for a Y boson at ~ 95 GeV is not yet ruled out. 

At Moriond 2024, ATLAS reported their X → Y(γγ) H(bb) result. The analysis has a sensitivity similar to the CMS one (a signal corresponding to the CMS excess should have resulted in a 2.7σ excess) but observed no excess. This result excludes the X → YH excess, while the others will require further studies. 

The CMS collaboration is fully engaged in the effort of analysing Run3 data. While the amount of data collected in the first two years only accounts for about half the Run2 data size, many improvements in our reconstruction strategy (e.g., better trigger algorithms of all-jet events) make new analyses with these data effectively as powerful as Run2 data, for certain final states. This improvement has been achieved mainly thanks to the use of cutting-edge machine learning techniques. The Artificial Intelligence revolution, a vision that our community had back in 2015, is now a mainstream reality. 

Around the time of the ICHEP conference (July 2024), we might start to see the first outcomes of this effort. New results might exclude these excesses (hence pointing to a statistical fluctuation in Run2 data) or confirm them. While we don’t expect to reach enough sensitivity to be able to make any observation claim, a positive outcome of any of these studies would enhance the attention to the 2024 and 2025 data taking, during which the Run3 dataset might increase by up to a factor 4.

There is no solid argument to justify any overstatement about the significance of these effects, which are still on the low end of global significance. Nevertheless, investigating them is our duty as experimentalists and the CMS collaboration is fully committed to this goal.

Note: In the following article, you can find a full list of the results that CMS collaboration reported in the 58^th Rencontres de Moriond 2024: https://cms.cern/news/cms-moriond-2024