Candidate event for the ZH → μμ cc process, where a Z boson and a Higgs boson decay to two muons (red tracks) and two charm-tagged jets (blue cones). (Image: ATLAS collaboration)

In the summer of 2024, the ATLAS collaboration released two new publications that provide significantly improved experimental probes of Higgs-boson interactions with the top, bottom and charm quark. The ATLAS researchers achieved the improved experimental precision by reanalysing their dataset of LHC proton-proton collisions recorded between 2015 and 2018 and implementing improved reconstruction techniques, sophisticated machine-learning algorithms and refined descriptions of the background processes. These results provide a crucial test of the mass-generation mechanisms for fermions. In the Standard Model (SM) these masses are generated via Yukawa interactions between the Higgs boson and the fermions, which implies a correlation between fermion masses and the Yukawa-coupling strength. However, the values of these coupling strengths are not predicted by the SM and have to be measured in experiments.

Heavy, Heavier, Top Quark 

The top quark is the heaviest known fundamental particle. Hence, in the SM it is expected to exhibit the strongest coupling to the Higgs boson. Since it is heavier than the Higgs boson, the Higgs-top interaction cannot be studied in Higgs-boson decays but instead is measured in the associated production of a Higgs boson and a pair of top quarks, ttH. The latest ATLAS ttH measurement targets Higgs-boson decays to bottom quarks, which is the most abundant with a probability of almost 60%. Together with the decays of the two top quarks, which yield additional b-jets, leptons and jets, this produces a very busy signature, posing challenges in the experimental reconstruction and the theoretical description of background processes. Previous measurements showed that existing simulations of top-quark pair production with additional heavy-flavour quarks did not model experimental data well. Thus, a new simulation of tt+≥1b processes was developed, improving the description of key distributions, such as the transverse momentum of the b-jet pair identified as the Higgs-boson candidate.

Diving deeper into top + heavy flavour production

In parallel to the ttH measurement, a dedicated measurement of the production of top-quark pairs in association with heavy-flavour jets was developed. Different jet-multiplicity and jet-flavour-identification categories probe different topologies and flavour compositions. The measured total cross-section was compared to various theoretical predictions (Fig. 1), and the differential cross-section was measured in many kinematic variables. These measurements confirm that the new tt+≥1b simulation used by the ttH measurement provides a better prediction of the total cross-section and distributions of crucial variables compared to previous samples. However, it also revealed that more work is needed on describing e.g. angular observables. This measurement presents an important wealth of information for future simulations of top-quark production. 

Fig. 1: Cross-section measurement (black markers + green band) of top-quark pair production in association with additional heavy flavour jets in different jet multiplicity and heavy flavour multiplicity compared to various predictions (coloured markers). The Powheg+Pythia 8 ttbb (dark purple square) is the new simulation used in the ttH measurement. The Powheg+Pythia8 sample (dark purple star) is the standard ATLAS simulation for top-quark pair production

Measuring the ttH production cross section  

The ttH measurement utilises a multi-classification transformer-based neural network to classify events as signal-like or as part of one of five background processes. This approach defines a signal-enriched phase space for extracting the ttH cross-section and multiple background-enriched phase-space regions for auxiliary measurements of the main background processes. This led to a sizable mitigation of the impact of the background prediction uncertainty compared to earlier ttH measurements. With more than two b-jets present, the b-jets from the Higgs-boson decay cannot be identified unambiguously. An additional neural network that aids in determining the two Higgs-boson-candidate jets was crucial in performing an improved measurement of the ttH production cross-section as a function of the Higgs boson’s transverse momentum, p_T^H. The cross-section is measured in six distinct p_T^H regions with a precision of up to 50% (Fig. 2). The total ttH cross-section is determined with a precision of 23% and is compatible with the SM expectation (Fig. 2). This measurement presents the most precise determination of the ttH production cross-section to date. 

Fig. 2: Left: the distribution of the machine-learning discriminant score (equivalent to “ttH-likeness”) in data (black markers) compared to the sum of the predictions of ttH production (red) and background processes (other colours). Right: the total measured ttH cross-section (“inclusive”) and the measured ttH cross-section as a function of Higgs-boson transverse momentum, p_T^H, normalised to the SM prediction (indicated by the dashed line at 1). 

Beautiful and Charming Higgs-Boson Interactions

The bottom and the charm quark are next in line in the quark-mass hierarchy. Their couplings to the Higgs boson are studied in Higgs-boson decays to bottom or charm quarks. To reduce multi-jet background, researchers study these decays in the associated production of a Higgs boson with a W or Z boson decaying to leptons (VH production, with V=W or V=Z). For the first time, the two couplings were extracted simultaneously, making use of the similarity of the two signatures. A new jet-identification algorithm and custom scheme sorts jets into c-jet-like and b-jet-like (Fig. 3), which is used to define H → bb and H → cc enriched phase spaces. The H → bb phase space is further split into a “resolved” category, where the two b-jets from the Higgs-boson decay are reconstructed as individual jets, and a “boosted” category, where the Higgs-boson decay products are reconstructed as a single large-radius jet. 

Fig. 3: Jet-flavour classification based on two multivariate-discriminant scores (“b-tag score” and “c-tag score”) to classify events into b-jet and c-jet-like to define H → bb and H → cc enriched phase space regions. For each jet-classification category, the efficiency for a jet of a given flavour to be sorted into it is given. 

Machine-learning-based signal-background classifiers, previously only used in the resolved H → bb regime, were extended to H → cc and boosted H → bb. The design of a vast amount of “control regions” to provide auxiliary measurements of the background processes was crucial to reduce the dependence on theory assumptions, thus significantly reducing the measurement uncertainty. These are only a few of the many improvements that led to an increase in sensitivity of 15%, 50% and a factor of 3 for the resolved H → bb, boosted H → bb and H → cc regimes, respectively, compared to previous measurements. All experimental techniques were successfully verified by measuring VZ,Z → bb and VZ,Z → cc processes, presenting one of the most precise measurements of these processes. Notably, the VZ,Z → cc process was observed with a 5.2 sigma significance above the background-only hypothesis.

Enhanced Precision to Higgs-boson Decays to Bottom and Charm Quarks

The implemented improvements led to an unprecedented precision of 16% in measuring H → bb decays. The precision for H → cc decays is lower due to a smaller branching ratio, larger background cross-sections and challenges in c-jet identification. However, researchers established the most stringent experimental limit on H → cc decays of 12 times the SM expectation at the 95% confidence level (Fig. 4), which translates to the exclusion of Higgs-charm coupling enhancement factors of larger than 4.2 times the SM expectation. The combined study of H → bb and H → cc enables the extraction of the ratio of the Higgs-boson coupling to charm quarks compared to the bottom-quark coupling without external assumptions. This confirmed with >99% confidence that the Higgs-charm coupling is weaker than the Higgs-bottom coupling. Additionally, the excellent precision in the H → bb channel enabled detailed measurements of WH and ZH cross-sections across 13 phase space regions (Fig. 4). Compared to previous WH/ZH cross-section measurements, the sensitivity in almost all bins was improved. New measurements at high transverse-momentum and splitting of the ZH measurement into different jet multiplicities provide additional knowledge on potential modifications from new physics and higher-order processes. This results in the most precise and comprehensive WH/ZH cross-section measurement to date.

Fig. 4: Left: Expected (dashed vertical line) and observed (solid vertical line) limit, at the 95% confidence level, on H → cc in the three channels based on the expected W/Z boson decay and their combination. Right: Measurement (black markers) of the production cross section of a Higgs boson in association with a W or Z boson as a function of transverse momentum and, in the case of ZH, also the jet multiplicity compared to the SM expectation (red).

ATLAS Reaches New Heights in Probe of Higgs-Boson Couplings to Quarks

The new publications by the ATLAS collaboration broaden our understanding of Higgs-boson couplings. They showcase enhanced precision through improved experimental techniques and emphasize the need for accurate knowledge of background processes. These new Higgs-boson studies hone in on the coupling strength of the Higgs boson to top, bottom and charm quarks. Additionally, they provide the most precise measurements of ttH, WH and ZH production cross-sections. This represents a significant advancement in the exploration of the Higgs sector.

Read more

ATLAS ttH,H → bb : https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PAPERS/HIGG-2020-24/

ATLAS VH,H → bb,cc: https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PAPERS/HIGG-2020-20/  

ATLAS tt+heavy flavour: https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PAPERS/TOPQ-2019-03/

CERN LHC seminar: https://indico.cern.ch/event/1441581/