⚛️ high-energy experiments

Cross-section measurements of boosted Higgs boson production in final states with pairs of hadronically decaying $\tau$ -leptons with the ATLAS experiment

Using the full Run 2 and partial Run 3 datasets from the ATLAS experiment, this paper presents the first evidence of boosted Higgs boson production in the di- $\tau$ final state with a significance of 3.8 standard deviations, achieved through dedicated reconstruction techniques for high-transverse-momentum Higgs bosons ( $p_T > 300$ GeV) and reporting both inclusive and differential cross-section measurements.

Original authors: ATLAS Collaboration

Published 2026-07-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: ATLAS Collaboration

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

Imagine the universe as a giant, cosmic kitchen where the most famous chef, the Higgs boson, is busy cooking up mass for all the other particles. We know this chef exists because we found the recipe in 2012, but now scientists want to see exactly how the chef behaves when the kitchen gets incredibly hot and chaotic. Specifically, they are looking for moments when the Higgs boson is "boosted"—meaning it's zooming through the particle collider at nearly the speed of light, carrying so much energy that its ingredients (the particles it decays into) get squished together into a tight, messy bundle.

The main character in this story is the Higgs boson, a particle that gives other things their weight. When it breaks apart, it often turns into a pair of "tau" particles, which are like heavy, unstable cousins of the electron. Usually, these tau particles fly apart in different directions, making them easy to spot. But when the Higgs is boosted, it's like a spinning top moving so fast that its arms (the tau particles) get pulled so close together they look like a single, confused blob. The challenge for physicists is to build a detector smart enough to look at that single blob and say, "Ah, I see two distinct tau particles hiding inside that mess!" This paper is about building that super-spy glass and checking if the Higgs chef is indeed cooking up these high-speed, squished meals just as the standard recipe predicts.

The High-Speed Hunt for the Squished Higgs

In this study, the ATLAS collaboration at CERN's Large Hadron Collider (LHC) decided to play a game of "find the needle in a haystack," but the needle is a Higgs boson zooming at incredible speeds, and the haystack is a mountain of particle collisions. They looked at data collected from 2015 to 2024, which includes a massive amount of proton collisions at energies of 13 TeV and 13.6 TeV. That's a lot of data—equivalent to 140 and 162 "inverse femtobarns" of collisions, which is a fancy way of saying they watched trillions of particle crashes to find just the right ones.

The problem they faced was like trying to identify two people hugging tightly in a crowded stadium. When the Higgs boson is moving slowly, its decay products (the tau particles) separate easily, and standard detectors can spot them like two distinct faces in a crowd. But when the Higgs is "boosted" with a transverse momentum (a measure of how hard it's moving sideways) above 300 GeV, those two tau particles get squished so close together that they merge into what looks like a single jet of particles. Standard tools, designed for well-separated particles, would just see one big blob and miss the Higgs entirely.

To solve this, the team invented a new set of "glasses" called a dedicated boosted di-tau reconstruction. Imagine looking at a single, large, blurry blob and using a special algorithm to zoom in and realize, "Wait, that's actually two smaller blobs hugging each other!" They used a giant net (a large-radius jet) to catch the whole mess, then used a finer mesh (smaller subjets) to separate the two tau particles inside. To make sure they weren't just seeing random noise or other particles pretending to be a Higgs, they trained a super-smart computer brain called an "OMNI tagger." This tagger uses advanced machine learning, similar to the technology that helps self-driving cars recognize pedestrians, to distinguish between a real Higgs decay and a fake one made of ordinary junk (multi-jet background).

The results of this high-speed hunt are exciting. After sifting through all that data, the team found evidence that the Higgs boson is indeed being produced in this boosted, squished state. They observed a signal that stands out from the background noise with a significance of 3.8 standard deviations. In the world of particle physics, this is like hearing a distinct "ding" in a noisy room that is loud enough to be sure it's not just a random clatter, though not quite loud enough to be a "shout" (which would require 5 standard deviations). This provides strong evidence for Higgs production in this high-energy regime.

They didn't just find the Higgs; they measured how often it happens. They calculated the "cross-section," which is basically the probability of this specific high-speed event occurring. For the combined data from their two main data-taking periods, they found a production rate that matches the predictions of the Standard Model (the rulebook of particle physics) very well. They also broke down the results into different categories, looking at how the Higgs is made (either by smashing gluons together or by fusing vector bosons) and how its speed varies. In every category, the numbers they measured were consistent with what the Standard Model predicted, meaning the Higgs boson is behaving exactly as expected, even when it's running at breakneck speeds.

So, what's the takeaway? The Higgs boson is a reliable character. Even when it's moving so fast that its parts merge into a single, confusing blob, our new high-tech tools can still spot it, and it behaves exactly as the theory says it should. This study extends our understanding of the Higgs into a new, high-energy territory, confirming that our current map of the subatomic world is still accurate, even in the most extreme corners of the particle zoo.

Technical Summary: Cross-section measurements of boosted Higgs boson production in final states with pairs of hadronically decaying $\tau$ -leptons with the ATLAS experiment

Problem and Motivation
The observation of the Higgs boson in 2012 completed the Standard Model (SM) particle content, but precision measurements of its properties remain a primary objective of the LHC physics program. The $H \to \tau\tau$ decay channel is critical as it probes the largest leptonic Yukawa coupling. While previous measurements have established the properties of the Higgs boson at low transverse momentum ( $p_T$ ), the high- $p_T$ regime ( $p_T^H > 300$ GeV) remains experimentally challenging. In this "boosted" regime, the decay products of the Higgs boson become highly collimated. Specifically, for hadronically decaying $\tau$ -leptons ( $\tau_{had}$ ), the two $\tau$ -leptons often merge into a single jet, rendering standard reconstruction algorithms, which assume well-separated decay products, inefficient. Furthermore, deviations from SM predictions in the top-quark Yukawa coupling or the gluon-gluon fusion (ggF) loop structure are expected to be most pronounced at high $p_T$ . This paper addresses the need to extend sensitivity to the $H \to \tau\tau$ process in this boosted phase space to test SM predictions and constrain potential new physics.

Methodology
The analysis utilizes the full Run 2 dataset (2015–2018, $\sqrt{s}=13$ TeV, 140 fb $^{-1}$ ) and a partial Run 3 dataset (2022–2024, $\sqrt{s}=13.6$ TeV, 162 fb $^{-1}$ ) collected by the ATLAS detector. The selection targets events where both $\tau$ -leptons decay hadronically ( $\tau_{had}$ ).

Reconstruction and Identification: To address the collimation of decay products, the analysis employs a dedicated boosted di- $\tau_{had}$ reconstruction algorithm. This method reconstructs a large-radius jet ( $R=1.0$ ) seeded by calorimeter energy clusters and identifies two subjets ( $R=0.2$ ) within it, corresponding to the individual $\tau_{had}$ decay products. A novel machine-learning-based identifier, the "OMNI tagger," utilizing a Point-Edge Transformer (PET) architecture, is used to distinguish real di- $\tau_{had}$ decays from the dominant multi-jet QCD background. The tagger achieves a background rejection factor of approximately ten compared to previous Boosted Decision Tree (BDT) methods at an 85% signal efficiency working point.
Event Selection: Events are selected requiring a di- $\tau_{had}$ candidate with $p_T > 300$ GeV and $|\eta| < 2.5$ . The analysis defines Signal Regions (SR) enriched in opposite-sign (OS) di- $\tau_{had}$ candidates passing the OMNI tagger, and Control Regions (CR) dominated by same-sign (SS) candidates or those failing the tagger score, used to estimate the misidentified $\tau_{had}$ (fake) background via a data-driven fake factor method.
Statistical Analysis: A multivariate analysis (MVA) approach is adopted. A two-class BDT separates signal from background, while a four-class BDT categorizes events into Simplified Template Cross-Section (STXS) bins: one inclusive Vector Boson Fusion (VBF) bin and three ggF bins ( $300 < p_T^H < 450$ GeV, $450 < p_T^H < 650$ GeV, and $p_T^H > 650$ GeV). Systematic uncertainties, including those related to the di- $\tau_{had}$ energy scale, identification, and theoretical modeling of signal and backgrounds, are incorporated as nuisance parameters in a profile likelihood fit.

Key Contributions

First ATLAS Measurement in Boosted Regime: This work presents the first ATLAS measurements of the $H \to \tau\tau$ cross-section for $p_T^H > 300$ GeV.
Extended Kinematic Reach: The analysis extends the lower $p_T^H$ threshold of boosted Higgs measurements from 450 GeV (as in previous $H \to b\bar{b}$ measurements) down to 300 GeV, significantly increasing the accessible phase space.
Advanced Reconstruction Techniques: The implementation of the dedicated di- $\tau_{had}$ reconstruction and the OMNI tagger demonstrates a successful strategy for recovering efficiency in the highly collimated regime where standard algorithms fail.
Comprehensive Cross-Section Measurements: The paper provides inclusive cross-section measurements for $p_T^H > 300$ GeV, differential cross-sections in three $p_T^H$ bins, and STXS measurements for both ggF and VBF production modes.

Results

Significance: Combining Run 2 and Run 3 data, the analysis observes a signal significance of 3.8 standard deviations ( $\sigma$ ) over the background-only hypothesis, with an expected significance of 3.3 $\sigma$ . This constitutes evidence for Higgs boson production in the boosted di- $\tau$ final state.
Signal Strength: The measured signal strength for $p_T^H > 300$ GeV is $\mu = 1.16^{+0.42}_{-0.25} (\text{stat}) ^{+0.23}_{-0.27} (\text{syst})$ , which is compatible with the SM expectation at the 69% confidence level.
Cross-Sections: The measured inclusive cross-sections for Run 2 and Run 3 are $9.0^{+7.0}_{-4.8} (\text{stat}) ^{+2.7}_{-3.3} (\text{syst})$ fb and $25.2^{+9.1}_{-6.3} (\text{stat}) ^{+5.5}_{-6.2} (\text{syst})$ fb, respectively. These are in good agreement with SM predictions of 14 $\pm$ 3 fb and 15.6 $\pm$ 3.5 fb.
STXS and Differential Results: All STXS and differential cross-section measurements are compatible with SM predictions within uncertainties. While a deficit is observed in the Run 2 data for the ggF bin with $p_T^H > 650$ GeV, the combined Run 2 and Run 3 results remain consistent with the SM. The results for the $450 < p_T^H < 650$ GeV interval show compatibility with the $H \to b\bar{b}$ measurement in Run 3 data.

Significance
The paper claims that these results provide the first evidence of Higgs boson production in the boosted regime for the $H \to \tau\tau$ channel using ATLAS data. By extending the sensitivity to $p_T^H > 300$ GeV and employing dedicated reconstruction techniques for merged $\tau$ -lepton pairs, the analysis successfully probes a kinematic region where new physics contributions are expected to be enhanced. The compatibility of all measurements with the Standard Model reinforces the current understanding of Higgs boson properties and Yukawa couplings, while establishing the methodology for future high-precision studies in the high- $p_T$ regime.

More like this