⚛️ high-energy experiments

Simultaneous measurement of the Higgs boson decay rates to WW and ZZ in fully hadronic final states at FCC-ee

This paper presents a study of simultaneous measurements of Higgs boson decay rates to WW and ZZ in fully hadronic six-jet final states at FCC-ee, demonstrating that with an integrated luminosity of 10.8 ab⁻¹ at 240 GeV, the production cross section times branching fraction can be determined with relative uncertainties of 1.48% and 8.20% for the H→WW and H→ZZ channels, respectively.

Original authors: Jan Eysermans, Mila Pandurovic, Christoph Paus, Michele Selvaggi, Aman Desai

Published 2026-10-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jan Eysermans, Mila Pandurovic, Christoph Paus, Michele Selvaggi, Aman Desai

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Higgs boson is a fundamental particle that gives mass to other particles, a discovery that completed the standard model of physics. However, knowing the particle exists is only the beginning; scientists must now measure exactly how it interacts with other particles to see if our current understanding of the universe is complete. One of the most important ways to test this is by measuring how often the Higgs boson decays, or breaks apart, into pairs of force-carrying particles called W and Z bosons. These measurements are crucial because they allow physicists to calculate the total width of the Higgs boson, a property that reveals how quickly it disappears into other forms of matter. If the measurements differ from predictions, it could signal the existence of new, hidden physics.

To achieve the extreme precision required for these tests, researchers are planning a future electron-positron collider known as the FCC-ee. In a new study, a team of physicists simulated how this machine would perform a specific, difficult measurement: observing the Higgs boson decaying into W and Z bosons when all of those particles subsequently break down into streams of quarks, which appear in detectors as sprays of particles called jets. This "fully hadronic" approach is attractive because the Higgs boson decays into W bosons very frequently, and Z bosons often decay into quarks as well, offering a large number of events to study. However, it is also notoriously messy. The final result is a collision with six jets of particles, a configuration that is hard to distinguish from background noise and where it is difficult to tell which jets came from a W boson and which came from a Z boson.

The researchers focused on a scenario where the collider operates at an energy of 240 gigaelectronvolts, a setting designed to produce Higgs bosons in association with a Z boson. They simulated a massive amount of data, equivalent to running the collider for a long time to collect a specific amount of collisions, to see if they could successfully separate the signal from the background. The primary challenge was that the W and Z bosons have similar masses, and when they decay into jets, the resulting patterns look very much alike. Furthermore, the decay into two Z bosons happens much less often than the decay into two W bosons, making the Z signal easy to lose in the shadow of the more common W events.

To solve this, the team developed a strategy that treats the problem as a simultaneous puzzle. Instead of trying to find the W events and Z events separately, they built a computer program that looks at both possibilities at once. First, they filtered the data to keep only events with exactly six jets, discarding anything that didn't fit this specific shape. Then, they used a sophisticated algorithm to group the jets into pairs, trying to reconstruct the original W and Z bosons. A key insight was that W bosons are made of different types of quarks, while Z bosons are made of matching pairs. By using a machine-learning tool to identify the "flavor" or type of quark inside each jet, the researchers could better distinguish between the two types of decays. They trained a classifier to sort events into two groups: those that look like they came from a W boson and those that look like they came from a Z boson.

After sorting the events, the team applied another layer of analysis to remove the remaining background noise, which mostly comes from other types of particle collisions that mimic the six-jet signature. They used statistical methods to fit the data, accounting for the fact that some W events might be misidentified as Z events and vice versa. The result of this simulation was a clear path to measuring both decay rates with high precision. The study predicts that with the planned amount of data, the uncertainty on the measurement of the W boson decay rate would be about 1.48 percent, and the uncertainty for the Z boson decay rate would be about 8.20 percent. These numbers represent a significant improvement over what is currently possible, showing that even the messy, fully hadronic channel can provide clean, high-quality data.

The authors note that while they did not include every possible source of error, such as imperfections in the detector hardware, the abundance of control data available at the future collider should allow them to constrain these effects. The study concludes that this approach is viable and will play a major role in the future program of the FCC-ee. By successfully measuring these decay rates, physicists will be able to determine the total width of the Higgs boson without relying on theoretical assumptions, providing a model-independent check on the Standard Model. This work demonstrates that with the right tools and a careful strategy, even the most complex particle collisions can be untangled to reveal the fundamental properties of the universe.

Technical Summary: Simultaneous Measurement of Higgs Boson Decay Rates to WW and ZZ in Fully Hadronic Final States at FCC-ee**

Problem Statement
The precise determination of Higgs boson couplings to electroweak gauge bosons ( $g_W$ and $g_Z$ ) is a primary objective of the Future Circular Collider (FCC-ee) physics program. These measurements are critical for the model-independent determination of the total Higgs boson width ( $\Gamma_H$ ). While the Higgsstrahlung process ( $e^+e^- \to ZH$ ) at $\sqrt{s} = 240$ GeV offers a clean environment for such studies, the fully hadronic decay channel ( $H \to WW^*$ and $H \to ZZ^*$ , where the associated $Z$ and the vector bosons from the Higgs decay all decay to quarks) presents significant experimental challenges. These include large backgrounds from diboson and fermion-pair production, combinatorial ambiguities in assigning six jets to their parent bosons, and substantial kinematic overlap between the $H \to WW^*$ and $H \to ZZ^*$ signal modes. The $H \to ZZ^*$ mode is particularly difficult to isolate due to its branching fraction being approximately eight times smaller than that of $H \to WW^*$ , making it susceptible to contamination from misidentified $WW^*$ events.

Methodology
The study utilizes Monte Carlo simulations generated at $\sqrt{s} = 240$ GeV with an integrated luminosity of $10.8 \text{ ab}^{-1}$ , assuming four interaction points and an IDEA-like detector configuration simulated via DELPHES. The analysis strategy proceeds through the following stages:

Event Generation and Preselection: Signal samples ( $ZH \to Z(WW^*)$ and $ZH \to Z(ZZ^*)$ ) and background samples ( $WW$ , $ZZ$ , and $Z/\gamma^* \to q\bar{q}$ ) are generated using WHIZARD3 and PYTHIA8. Events are required to have fully hadronic activity with no identified leptons ( $p > 10$ GeV) and low missing momentum ( $p_{miss} < 20$ GeV).
Jet Clustering and Topology Selection: Events are clustered into exactly six jets using the Durham $k_T$ algorithm. To exploit the six-parton nature of the signal versus the lower-multiplicity backgrounds, specific jet-transition scales ( $\sqrt{d_{ij}}$ ) are imposed ( $\sqrt{d_{23}} > 50$ GeV, $\sqrt{d_{34}} > 25$ GeV, $\sqrt{d_{45}} > 15$ GeV, $\sqrt{d_{56}} > 10$ GeV).
Jet Pairing and Mass Reconstruction: The six jets are grouped into three pairs (associated $Z$ , on-shell vector boson, off-shell vector boson) by minimizing a $\chi^2$ -like quantity based on invariant masses. This is performed independently under both $H \to WW^*$ and $H \to ZZ^*$ hypotheses.
Signal Categorization: A Boosted Decision Tree (BDT) classifier separates events into mutually exclusive $H \to WW^*$ -like and $H \to ZZ^*$ -like categories. The classifier utilizes reconstructed boson kinematics and jet-flavor information derived from a Graph Neural Network (GNN) tagger. The flavor information is crucial, as hadronic $W$ decays involve different-flavor quarks, while $Z$ decays can produce same-flavor pairs (e.g., $b\bar{b}$ , $c\bar{c}$ ).
Background Suppression: Within each category, a second dedicated BDT is trained to suppress remaining backgrounds (dominated by $WW$ production) using kinematic variables and flavor tagging.
Statistical Interpretation: A simultaneous binned likelihood fit is performed on the output distributions of the background-suppression BDTs in both categories to extract the signal strengths ( $\mu_{WW^*}$ and $\mu_{ZZ^*}$ ) simultaneously, accounting for cross-contamination and background normalization uncertainties.

Key Contributions

Simultaneous Extraction: The paper demonstrates a method to simultaneously extract signal strengths for $H \to WW^*$ and $H \to ZZ^*$ in the challenging fully hadronic channel, addressing the kinematic overlap that typically hinders such measurements.
Flavor-Tagging Integration: The analysis explicitly incorporates jet-flavor information (via GNN taggers) into the multivariate classifiers to distinguish between $W$ and $Z$ boson decays, leveraging the distinct heavy-flavor composition of $Z \to b\bar{b}/c\bar{c}$ versus $W \to q\bar{q}'$ .
Optimization Strategy: The study details an optimization of jet-transition scales and categorization thresholds that maximizes the combined significance of both signal modes, acknowledging the trade-off where $H \to ZZ^*$ efficiency is slightly compromised to favor the higher-yield $H \to WW^*$ topology.

Results
Assuming the baseline FCC-ee scenario ( $\sqrt{s} = 240$ GeV, $10.8 \text{ ab}^{-1}$ ), the analysis yields the following expected relative uncertainties on the production cross section times branching fraction ( $\sigma_{ZH} \times \text{BR}$ ):

$H \to WW^*$ : 1.48%
$H \to ZZ^*$ : 8.20%

These results include a 1% normalization uncertainty on each background process. The two signal strength parameters exhibit a correlation of $-28.4\%$ . The analysis notes that if background normalization uncertainties were removed, the uncertainties would improve to 1.32% and 7.78%, respectively. The study finds that the analysis remains predominantly statistically limited, as the impact of background normalization uncertainties is modest.

Significance and Claims
The paper claims that the fully hadronic final state, despite its experimental challenges, provides significant sensitivity to both $H \to WW^*$ and $H \to ZZ^*$ decay modes. By achieving percent-level precision on $H \to WW^*$ and sub-10% precision on $H \to ZZ^*$ , this channel is positioned to make an important contribution to the precision determination of Higgs couplings to electroweak gauge bosons. Ultimately, these measurements are essential for the model-independent determination of the total Higgs boson width at the FCC-ee. The authors emphasize that while detector and theoretical systematic uncertainties were not explicitly evaluated in this study, the availability of abundant control samples at FCC-ee suggests that the analysis can effectively constrain these effects, maintaining its statistical dominance.

More like this