Probing the Rare Four-Bottom Higgs Decay $H\to b\bar b b\bar b$ at the HL-LHC and ILC

This paper proposes the rare Standard Model Higgs decay $H\to b\bar b b\bar b$ as a probe of Higgs interactions, calculating its branching ratio of approximately $1.6\times10^{-3}$ and demonstrating that it can be observed with high significance at both the High-Luminosity LHC and the ILC using multivariate analysis techniques.

The Gist

Imagine the Higgs boson as a very shy, rare celebrity who usually keeps to themselves. When this celebrity does "decay" (break apart), they almost always split into two heavy particles called bottom quarks. Physicists have already seen this happen. But this paper asks a much harder question: What if the Higgs boson splits into four bottom quarks at once?

This is like asking if our celebrity celebrity could suddenly split into four identical twins instead of just two. It's incredibly rare, but if we can catch it, it tells us a lot about how the celebrity interacts with the world.

Here is a simple breakdown of what the authors did and found:

1. The Mystery of the Four Twins

The authors calculated the odds of this "four-twin" event happening. They found it's rare (about 1 in 600 times), but not impossible.

They discovered that this event doesn't happen in just one way. It's like a magic trick that can be performed using three different methods:

• Method A (The Gluon Split): The Higgs splits into a bottom pair and a "gluon" (a particle that carries the strong force), and that gluon then splits into another bottom pair. This is the most common way (about 68% of the time).
• Method B (The Z-Boson Bridge): The Higgs briefly turns into two Z-bosons (another type of particle), which then turn into the four bottom quarks. This happens about 30% of the time.
• Method C (The Loop): A more complex, loop-based process that happens very rarely (about 2%).

The Interference Twist:
Here is the tricky part. When these three methods happen at the same time, they don't just add up like numbers. They interfere with each other, like two waves in a pond crashing together. Sometimes they cancel each other out. The authors found that these waves mostly cancel each other out, making the final event slightly less likely than if you just added the three methods together. This "destructive interference" is a crucial detail they calculated for the first time with high precision.

2. The Hunt at the Large Hadron Collider (HL-LHC)

The authors tried to figure out how to find these four twins at the HL-LHC (a massive particle collider in Switzerland that smashes protons together).

• The Problem: Imagine trying to find four specific, rare coins dropped in a stadium filled with millions of other coins. The "noise" (background events where random particles just happen to look like four bottom quarks) is huge. For every one real signal, there are about 160 background "fake" signals.
• The Solution: They used a "smart filter" called a Boosted Decision Tree (BDT). Think of this as a super-smart AI detective. Instead of just looking at one thing (like the weight of the coins), the AI looks at 20 different clues at once: the energy of the particles, their angles, how they are grouped, and how they move.
• The Result: Even with the AI, it's a tough fight. At the HL-LHC, they estimate they might see about 3.5 "sigma" of evidence. In science, 3 sigma is a strong hint ("we think we see it!"), but not quite a full discovery (which requires 5 sigma). However, if they combine data from all the detectors, they might just cross that line.
• The Catch: Even if they find it, the "noise" is so loud that they can't measure the details very precisely. It's like hearing a whisper in a rock concert; you know someone is talking, but you can't make out the words.

3. The Hunt at the International Linear Collider (ILC)

To get a clear picture, the authors looked at a future machine called the ILC (a proposed electron-positron collider).

• The Advantage: Imagine the HL-LHC is a chaotic rock concert, but the ILC is a silent library. Because electrons and positrons are "cleaner" particles than protons, there is almost no background noise.
• The Result: In this quiet environment, the "four twins" signal stands out clearly. The AI filter can separate the signal from the background almost perfectly.
• The Payoff: At the ILC, they could find this event with 5.5 sigma (a confirmed discovery) with just a small amount of data. More importantly, because the background is so low, they could measure the exact rate of this decay with 5% to 6% precision. This turns the event from a "maybe we saw it" into a "we know exactly how it works."

Summary

This paper proposes a new way to study the Higgs boson by looking for a very rare decay into four bottom quarks.

• At the HL-LHC: It's a difficult, noisy hunt. They might find enough evidence to say "Yes, it exists," but the background noise makes it hard to study the details.
• At the ILC: It's a clean, precise measurement. They could not only confirm it exists but measure its properties with high accuracy.

The authors conclude that while the HL-LHC might be able to spot this rare event, the ILC is the perfect tool to truly understand it. This study sets the stage for future experiments to look for this specific decay, which could also help scientists spot signs of "New Physics" if the real world behaves differently than their calculations.

Technical Summary

Technical Summary: Probing the Rare Four-Bottom Higgs Decay $H \to b\bar{b}b\bar{b}$ at the HL-LHC and ILC

Problem and Motivation
The discovery of the Higgs boson and the subsequent confirmation of its Standard Model (SM) properties have completed the SM particle pattern. While dominant decay modes ($\gamma\gamma$, $ZZ^*$, $WW^*$, $\tau\tau$) and the dominant $b\bar{b}$ mode have been observed, the rare decay into four bottom quarks ($H \to b\bar{b}b\bar{b}$, or $H4B$) remains unexplored. This decay is motivated by several factors:

1. Theoretical Relevance: The partial width is non-negligible ($\sim 10^{-3}$ branching ratio) and proportional to the bottom-quark Yukawa coupling. It contributes to the Next-to-Leading-Order (NNLO) correction of the total Higgs width.
2. Unique Topology: Unlike four-light-fermion decays dominated by $H \to ZZ^*/WW^*$, the $H \to b\bar{b}b\bar{b}$ final state is dominated by the $H \to b\bar{b}g \to b\bar{b}b\bar{b}$ topology due to the large bottom Yukawa coupling. This provides direct access to Higgs interactions with down-type quarks and the internal structure of the decay.
3. New Physics Baseline: Establishing the SM prediction and its interference structure is essential to define an irreducible baseline for searching for Beyond Standard Model (BSM) physics, such as light resonances coupled to $b$ quarks or modified Higgs couplings.

Previous studies of $H \to 4b$ were limited to Next-to-Leading-Order (NLO) calculations of the decay width itself, lacking a comprehensive signal-versus-background analysis or a correct treatment of interference effects between different diagram topologies.

Methodology
The study investigates the $H \to b\bar{b}b\bar{b}$ decay in two production channels: associated production with a $W$ boson ($pp \to WH$) at the High-Luminosity LHC (HL-LHC) and associated production with a $Z$ boson ($e^+e^- \to ZH$) at the International Linear Collider (ILC).

• Theoretical Framework: The authors compute the leading contributions to the decay width using Feynman diagrams representing three primary topologies: $H \to b\bar{b}g \to b\bar{b}b\bar{b}$, $H \to ZZ^* \to b\bar{b}b\bar{b}$, and the loop-induced $H \to gg \to b\bar{b}b\bar{b}$. The analysis explicitly calculates the interference terms between these amplitudes using CompHEP/CalcHEP.
• Kinematic Analysis: To distinguish signal from background and separate the different decay topologies, the authors identify specific kinematic observables. These include invariant masses of $b$-quark pairs (ordered by energy in the Higgs rest frame, e.g., $M_{12}$ for the two hardest, $M_{34}$ for the two softest), angular correlations ($\cos \theta_{ij}$), and transverse momenta. Two-dimensional and three-dimensional distributions are utilized to exploit correlations that are washed out in one-dimensional projections.
• Simulation and Analysis:
  • HL-LHC: Simulations are performed at $\sqrt{s} = 14$ TeV using CalcHEP for matrix elements, PYTHIA for parton showering/hadronization, and Delphes for fast detector simulation (CMS configuration). The dominant irreducible background is QCD production of $W + 4b$. A multivariate analysis (MVA) using Boosted Decision Trees (BDTG) is employed to separate the signal from the large background.
  • ILC: Simulations are performed at $\sqrt{s} = 250$ GeV. The analysis utilizes the cleaner environment of an $e^+e^-$ collider, where the dominant background is the continuum $e^+e^- \to Zb\bar{b}b\bar{b}$. The same BDTG strategy is applied to the reconstructed kinematic variables.

Key Contributions

1. Interference Calculation: The paper provides an accurate treatment of interference effects between the leading diagrams ($H \to b\bar{b}g$, $H \to ZZ^*$, $H \to gg$). The authors find a net destructive interference of approximately $-6.4\%$, correcting previous literature which reported a positive interference for the $b\bar{b}g \times gg$ term.
2. Branching Ratio: The calculated branching ratio for $H \to b\bar{b}b\bar{b}$ is $\text{Br} \approx 1.66 \times 10^{-3}$. The dominant contribution ($\sim 68\%$) comes from $H \to b\bar{b}g$, followed by $H \to ZZ^*$ ($\sim 30\%$) and $H \to gg$ ($\sim 2\%$).
3. Kinematic Discriminants: The study identifies that the signal is characterized by correlated patterns in energies, invariant masses, and angles, rather than a single variable. Specifically, the $H \to b\bar{b}g$ topology exhibits a "horizontal" distribution in 2D mass plots (hard $M_{12}$, soft $M_{34}$), while $H \to ZZ^*$ shows a "vertical" distribution (both pairs near $M_Z$).
4. MVA Strategy: The authors demonstrate that a multivariate approach using 20 input variables (including $p_T$, energies in the Higgs rest frame, invariant masses, and angular correlations) is necessary to achieve observable significance, as simple cut-based analyses are insufficient due to the large backgrounds.

Results

• HL-LHC ($pp \to WH \to Wb\bar{b}b\bar{b}$):

  • At an integrated luminosity of $3000 \text{ fb}^{-1}$, the signal-to-background ratio is initially very low ($\sim 1:160$ after loose preselection).
  • Using the BDTG classifier, the statistical significance reaches $\alpha \approx 3.5\sigma$ at the optimal working point ($S/B \approx 3.1\%$).
  • A tighter, high-purity working point ($BDTG > 0.76$) yields $S/B \approx 5.2\%$ with a significance close to $3\sigma$.
  • The authors project that a combined dataset from ATLAS and CMS (approx. $9 \text{ ab}^{-1}$) could potentially reach the $5\sigma$ discovery level, though the large surviving background limits precision measurements of the branching ratio or decay components.

• ILC ($e^+e^- \to ZH \to Zb\bar{b}b\bar{b}$):

  • The cleaner environment results in a significantly higher signal purity. At $300 \text{ fb}^{-1}$, the BDTG-selected sample achieves a statistical significance of $\alpha = 5.46$ with $S/B \approx 9.8$.
  • The reconstructed four-$b$ invariant mass remains sharply peaked around the Higgs mass, unlike the smeared distribution at the LHC.
  • The statistical precision on the branching ratio is estimated to be $\sim 18\%$ at $300 \text{ fb}^{-1}$, improving to $\sim 5-6\%$ at integrated luminosities of $3-4 \text{ ab}^{-1}$.

Significance and Claims
The paper concludes that the $H \to b\bar{b}b\bar{b}$ decay is a rare but realistic target for future Higgs studies.

• Observability: The HL-LHC has the potential to establish the existence of this decay mode through associated production, moving from theoretical prediction to experimental evidence.
• Precision Measurement: The ILC is presented as the ideal machine for precision studies of this channel. Its ability to produce a high-purity sample allows for the measurement of the branching ratio with percent-level precision and the potential to disentangle the individual decay components ($b\bar{b}g$ vs. $ZZ^*$).
• BSM Sensitivity: By establishing the SM baseline and its interference structure, this work provides a necessary foundation for searching for new physics that enhances or distorts the four-bottom final state. The analysis strategy developed serves as a direct tool for probing non-SM contributions in this challenging but motivated Higgs decay topology.