Combination of ATLAS and CMS searches for Higgs boson pair production at $\sqrt{s} = 13$ TeV

This paper presents a combined ATLAS and CMS analysis of Higgs boson pair production at 13 TeV using Run 2 data, which found no significant evidence for the signal but established upper limits on the production cross-section and set constraints on the Higgs trilinear self-coupling and its coupling to vector bosons.

The Gist

Imagine the universe is a giant, complex machine built by a master architect. For decades, scientists have been trying to understand the blueprints of this machine. In 2012, they found a crucial piece of the puzzle: the Higgs boson. You can think of the Higgs boson as the "glue" that gives other particles their mass, like how a heavy backpack makes a runner move slower.

But finding the glue wasn't enough. The architects wanted to know: How strong is the glue? Does it stick to itself?

This new paper is like a massive report card from two giant teams of detectives, ATLAS and CMS, who work at the world's biggest particle accelerator, the Large Hadron Collider (LHC) in Switzerland. They spent years smashing protons together at incredible speeds to see if they could catch two Higgs bosons being born at the same time.

Here is the story of their findings, explained simply:

1. The Big Hunt: Catching a "Double Trouble"

Finding one Higgs boson is like finding a needle in a haystack. Finding two Higgs bosons (a pair) at the same time is like finding two needles in that same haystack, but the haystack is on fire and spinning!

The Higgs boson is very shy. It usually decays (breaks apart) instantly into other particles. To find a pair, the scientists had to look for specific "fingerprints" left behind, like:

• Four bottom quarks (heavy particles).
• Two bottom quarks and two photons (light particles).
• Various combinations involving electrons, muons, and other particles.

They sifted through 140 trillion collisions (that's 140,000,000,000,000!) to find just a handful of these rare double-Higgs events.

2. The "Self-Love" Test: The Trilinear Coupling

The main goal of this paper was to measure something called the trilinear self-coupling.

The Analogy: Imagine the Higgs boson is a person.

• Single Higgs: We know how this person interacts with others (like giving mass to other particles).
• Double Higgs: We are now asking, "Does this person like themselves? How strongly do they hug themselves?"

In physics terms, this "self-hug" strength is determined by a number called $\kappa_\lambda$ (kappa-lambda).

• If the number is 1, the Higgs hugs itself exactly as the "Standard Model" (the current rulebook of physics) predicts.
• If the number is 0, it doesn't hug itself at all.
• If the number is 10, it's a very intense hug!

3. The Results: "It's a Bit Shy, But Normal"

The ATLAS and CMS teams combined their data to get the most accurate picture possible. Here is what they found:

• Did they find the signal? They saw a tiny hint of double Higgs production, but it wasn't a loud "Eureka!" moment. The evidence was about 1.1 standard deviations strong. In the world of particle physics, you usually need 5 to claim a discovery. Think of it like hearing a whisper in a noisy room; you think you heard something, but you aren't 100% sure yet.
• The "Self-Hug" Strength: They measured the strength of the Higgs self-interaction. The result was 0.8 (with a big range of uncertainty).
  • The Standard Model predicts it should be 1.0.
  • Their result of 0.8 is very close to 1.0. It means the Higgs boson is behaving exactly as the rulebook says it should. It's not doing anything weird or magical.

4. Why Does This Matter?

You might ask, "If it's just normal, why write a whole paper?"

The Analogy: Imagine you are testing a new car engine. You expect it to run at 100 miles per hour.

• If it runs at 100 mph, you know the engine is working as designed.
• If it runs at 150 mph or 50 mph, you know there is a flaw or a secret feature in the engine that we didn't know about.

By proving the Higgs is "normal" (running at 100 mph), the scientists have ruled out many wild theories that predicted the Higgs would behave strangely. They have tightened the screws on the universe's blueprint.

5. The Future: Tightening the Net

The paper concludes that while they haven't found "new physics" yet, they have set the tightest net ever cast.

• They have narrowed down the possible values for the Higgs self-coupling to a very small range.
• This means that if there is a secret, it's hiding in a very small corner.
• The next step is to wait for the LHC to run even more data (Run 3 and the High-Luminosity LHC) to see if that whisper becomes a shout.

Summary

Think of this paper as two giant teams of scientists joining forces to take a very high-resolution photo of the Higgs boson hugging itself. The photo is a bit blurry (due to the rarity of the event), but it clearly shows the hug is happening exactly as the universe's instruction manual predicted.

The verdict: The Higgs boson is behaving itself. The Standard Model is still the champion, but the search for something new continues!

Technical Summary

Here is a detailed technical summary of the paper "Combination of ATLAS and CMS searches for Higgs boson pair production at $\sqrt{s}=13$ TeV" (CERN-EP-2026-011).

1. Problem and Motivation

The Standard Model (SM) of particle physics predicts the existence of the Higgs boson ($H$) and its associated potential, which is responsible for electroweak symmetry breaking. While the mass of the Higgs boson and its couplings to other particles have been measured with high precision, the shape of the Higgs potential remains largely unexplored. Specifically, the Higgs boson trilinear self-coupling ($\lambda_{HHH}$) and the coupling between two Higgs bosons and two vector bosons ($VVHH$) are critical parameters that define the global structure of the potential.

• The Challenge: The production of Higgs boson pairs ($HH$) is the primary process to measure these couplings directly. However, the SM cross-section for $HH$ production is extremely small ($\sim 33$ fb at 13 TeV), and the signal is buried under large backgrounds.
• The Goal: To improve the sensitivity to these rare events by combining the independent datasets and analyses from the two general-purpose detectors at the Large Hadron Collider (LHC), ATLAS and CMS, using the full Run 2 dataset ($\sqrt{s}=13$ TeV, integrated luminosity $\approx 126\text{--}140 \text{ fb}^{-1}$).

2. Methodology

The paper presents a statistical combination of searches performed by both collaborations. The methodology involves several key technical components:

• Input Analyses: The combination includes the most sensitive channels from both experiments:

  • $bbbb$: Four $b$-quark final states (highest branching fraction $\sim 34\%$, but high multijet background). Includes both resolved and merged-jet topologies (crucial for Vector Boson Fusion, VBF).
  • $bb\tau\tau$: Two $b$-quarks and two $\tau$-leptons.
  • $bb\gamma\gamma$: Two $b$-quarks and two photons (clean signature, low branching fraction $\sim 0.26\%$).
  • $bbWW$ / $bb\ell\ell + E_T^{miss}$: Semi-leptonic and dileptonic decays.
  • Multilepton: Final states involving multiple leptons ($e, \mu, \tau_h$) and photons.
  • Note: Some specific CMS analyses (e.g., $bbWW$ hadronic, $VHH$) were excluded to reduce technical complexity, with an estimated negligible impact ($\sim 1\%$) on sensitivity.

• Signal Parametrization:

  • The signal strength is defined as $\mu_{HH} = \sigma_{measured} / \sigma_{SM}$.
  • The analysis constrains two coupling modifiers normalized to the SM:
    • $\kappa_\lambda$: Modifies the trilinear self-coupling ($\lambda_{HHH}$).
    • $\kappa_{2V}$: Modifies the $VVHH$ coupling (scaling both $WWHH$ and $ZZHH$).
  • Theoretical uncertainties (QCD scales, PDFs, top mass scheme) are treated as fully correlated between experiments.

• Statistical Framework:

  • Results are derived using a profile likelihood ratio test statistic.
  • Systematic uncertainties are modeled as nuisance parameters.
  • The $CL_s$ criterion is used to set upper limits.
  • Asimov datasets are generated to compute expected sensitivities under both the background-only ($\mu_{HH}=0$) and SM signal ($\mu_{HH}=1$) hypotheses.

3. Key Contributions

• First Global Combination: This is the first time ATLAS and CMS have combined their $HH$ search results, significantly reducing statistical uncertainties and improving constraints on the Higgs potential.
• Unified Treatment of Systematics: The paper rigorously handles correlations between the two experiments, particularly for theoretical uncertainties in signal production and background modeling (e.g., single Higgs processes).
• Refined Signal Modeling: The combination incorporates updated signal normalizations based on the PDF4LHC21 parton distribution functions and corrections to the two-loop amplitude in the POWHEG simulation for the gluon fusion (ggF) process.
• Comprehensive Channel Coverage: By combining resolved and merged-jet topologies, the analysis covers both the dominant ggF production mode and the VBF mode, which is sensitive to anomalous $\kappa_{2V}$ couplings at high transverse momentum.

4. Results

Signal Strength ($\mu_{HH}$)

• Observed Best Fit: $\hat{\mu}_{HH} = 0.8^{+0.9}_{-0.7}$.
• Significance: The observed significance of the $HH$ signal is 1.1 standard deviations ($\sigma$), compared to an expected significance of 1.3 $\sigma$ assuming the SM.
• Upper Limits (95% CL):
  • Assuming no $HH$ production ($\mu_{HH}=0$): The observed upper limit is 2.5 (expected: 1.7).
  • Assuming SM production ($\mu_{HH}=1$): The expected upper limit is 2.8.
  • In terms of cross-section: The observed limit is 73 fb (expected: 50 fb).

Coupling Constraints ($\kappa_\lambda$ and $\kappa_{2V}$)

The combination provides the tightest constraints to date on the Higgs self-coupling and the $VVHH$ coupling:

• $\kappa_\lambda$ (Trilinear Coupling):

  • Observed 95% CL interval: $-0.71 < \kappa_\lambda < 6.1$.
  • Expected 95% CL interval (assuming SM): $-1.3 < \kappa_\lambda < 6.7$.
  • Best-fit value: $\hat{\kappa}_\lambda = 1.8^{+2.8}_{-1.5}$.

• $\kappa_{2V}$ (Vector Boson Coupling):

  • Observed 95% CL interval: $0.73 < \kappa_{2V} < 1.3$.
  • Expected 95% CL interval (assuming SM): $0.66 < \kappa_{2V} < 1.4$.
  • Best-fit value: $\hat{\kappa}_{2V} = 1.02^{+0.14}_{-0.15}$.

• Sensitivity Drivers:

  • The $\kappa_\lambda$ constraint is primarily driven by the $bb\gamma\gamma$ and $bb\tau\tau$ channels.
  • The $\kappa_{2V}$ constraint is driven by the $bbbb$ channel in Lorentz-boosted topologies (merged jets), where anomalous couplings enhance the signal at high $m_{HH}$ and high $p_T$. The observed interval is slightly tighter than expected due to a deficit of events in the ATLAS $bbbb$ boosted analysis, pulling the fit toward the SM value.

5. Significance

• Validation of the SM: All results are consistent with the Standard Model predictions within uncertainties. No significant deviation indicating new physics in the Higgs sector has been observed.
• Improved Precision: The combination improves the expected sensitivity on $\kappa_\lambda$ by 10% and on $\kappa_{2V}$ by 8% compared to the best individual results from ATLAS or CMS alone.
• Future Outlook: These results represent the most comprehensive constraints on Higgs pair production to date. They set the stage for the High-Luminosity LHC (HL-LHC) era, where significantly larger datasets will be required to precisely measure the trilinear coupling and probe the shape of the Higgs potential with the precision needed to distinguish between the SM and various Beyond the Standard Model (BSM) scenarios.

In summary, this paper marks a milestone in Higgs physics by demonstrating the power of combining independent experimental efforts to probe the rarest Standard Model processes, providing the most stringent limits yet on the Higgs boson's self-interactions.