Search for single production of a vector-like B' quark decaying to a top quark and a W boson in the single-lepton final state in proton-proton collisions at $\sqrt{s}$ = 13 TeV

The CMS collaboration presents a search for the single production of a vector-like B' quark decaying to a top quark and a W boson in proton-proton collisions at 13 TeV using 138 fb$^{-1}$ of data, which sets the most stringent limits to date on the production of narrow-width singlet B' quarks with masses between 0.8 and 1.23 TeV.

The Gist

Imagine the Large Hadron Collider (LHC) as the world's most powerful "particle smasher." It shoots tiny protons at each other at nearly the speed of light, creating a chaotic explosion of debris. Physicists at the CMS experiment (one of the detectors at the LHC) are like detectives sifting through this debris, looking for a very specific, rare piece of evidence that shouldn't exist according to our current rulebook of physics, known as the Standard Model.

This paper is about a search for a "ghost" particle called a Vector-Like B' Quark.

The Mystery: Why look for this particle?

Our current rulebook (the Standard Model) works great, but it has a glitch. It requires some very delicate, unnatural adjustments to explain why the Higgs boson (a particle that gives other particles mass) has the weight it does. Physicists suspect there are "hidden helpers" in nature that fix this glitch. One of these helpers could be a heavy, "vector-like" quark.

Think of the Standard Model quarks as a team of players where some are left-handed and some are right-handed. A "vector-like" quark is a new kind of player who is ambidextrous (both left and right-handed at the same time). If these exist, they are likely very heavy and hard to spot.

The Hunt: How did they look?

The scientists collected data from 2016 to 2018, smashing protons together 138 times (in terms of "luminosity," which is a measure of how many collisions they saw). They were looking for a specific scenario:

1. A heavy B' quark is created.
2. It immediately falls apart (decays) into a Top quark and a W boson.
3. The Top quark and W boson then break down further. One of them produces a lepton (an electron or a muon, which are like heavy versions of electrons), a missing energy (carried away by invisible neutrinos), and some jets (sprays of particles).

Because the B' quark is so heavy, its decay products fly out with incredible speed, like a firework exploding. The scientists built a "reconstruction kit" to piece these flying pieces back together to see if they formed a B' quark.

The Challenge: Finding a needle in a haystack

The problem is that the Standard Model produces billions of "fake" events that look almost exactly like the signal they are looking for. It's like trying to find a specific rare coin in a pile of billions of identical-looking coins.

To solve this, the scientists used a clever trick called ABCDnn.

• The Analogy: Imagine you are trying to predict how many people will buy a specific rare item in a store (the Signal Region). You can't just guess; you need data. So, you look at four different aisles in the store (Control Regions A, B, C, and D) where you know the item doesn't sell, but where the customer behavior is similar.
• The AI Twist: Instead of just doing simple math, they used a sophisticated Neural Network (a type of AI) to learn the complex patterns of how the "fake" background events behave across these different aisles. The AI learned to transform the data from the aisles where they knew the answer into a prediction for the aisle where they were looking for the mystery particle. This allowed them to predict the background with incredible precision.

The Results: What did they find?

After analyzing the data with their AI tools, they looked at the "reconstructed mass" of the particles they found.

• The Verdict: They did not find the B' quark. The data matched the "Standard Model" prediction perfectly. There was no sign of the heavy, ambidextrous quark.
• The Exclusion: Because they didn't find it, they can now say with 95% confidence that if this particle does exist, it cannot be too light. They ruled out B' quarks with masses between 0.8 and 1.23 TeV (about 800 to 1,230 times the mass of a proton) if they have a specific "narrow width" (a measure of how quickly they decay).

Why is this important?

This is the most sensitive search for this specific type of particle ever done.

• Narrow Widths: Previous searches were good at finding particles that decay quickly (broad width), but this search was the first to be sensitive enough to find particles that decay very slowly (narrow width).
• New Limits: Even though they didn't find the particle, they drew a "Do Not Enter" line on the map of physics. They told theorists: "If you want to build a theory with a B' quark, it must be heavier than 1.23 TeV (or have different properties)."

Summary

The CMS team used a massive dataset and a smart AI system to look for a heavy, exotic particle that could fix a flaw in our understanding of the universe. They didn't find it, but by proving it doesn't exist in the mass range they searched, they have narrowed down the possibilities for what new physics might look like. It's a bit like searching a whole city for a specific person and, while not finding them, proving they aren't hiding in any of the houses you checked.

Technical Summary

Technical Summary: Search for Single Production of a Vector-Like B′ Quark in CMS

Problem and Motivation
This paper presents a search for the single production of a narrow-width vector-like B′ quark, a hypothetical heavy fermion proposed in various Beyond the Standard Model (SM) scenarios to address the "naturalness" problem regarding the Higgs boson mass. Unlike chiral SM quarks, vector-like quarks (VLQs) have left- and right-handed components that transform identically under SM gauge groups. The B′ quark is expected to mix predominantly with third-generation quarks, decaying into a top quark ($t$) and a $W$ boson. While previous searches have focused on broader decay widths or pair production, this analysis targets the single production of B′ quarks with narrow decay widths ($\Gamma/m_{B'} = 1\%$ and $5\%$). These narrow widths correspond to smaller production cross sections and weaker electroweak couplings, which were previously difficult to probe due to limited sensitivity and significant Standard Model interference effects.

Methodology
The analysis utilizes proton-proton collision data collected by the CMS experiment at $\sqrt{s} = 13$ TeV, corresponding to an integrated luminosity of 138 fb$^{-1}$ from 2016 to 2018.

• Event Selection: The search targets the single-lepton final state where the B′ quark decays to $tW$, and one $W$ boson decays leptonically ($e$ or $\mu$). Events are required to contain exactly one high-transverse momentum ($p_T > 55$ GeV) isolated lepton, missing transverse momentum ($p_T^{miss} > 60$ GeV), and at least one large-radius jet ($p_T > 200$ GeV). A small-radius jet is included if the top quark decays semileptonically.
• Reconstruction and Categorization: A B′ candidate is reconstructed from the lepton, $p_T^{miss}$, a large-radius jet, and potentially a $b$-tagged small-radius jet. Events are categorized into four cases based on the PARTICLENET algorithm's identification of the large-radius jet (as a top or $W$ candidate) and the presence of a $b$-tagged jet:
  1. Leptonic $W$ + $t$-tagged jet.
  2. Semileptonic $t$ candidate + $W$-tagged jet.
  3. Semileptonic $t$ candidate + untagged jet.
  4. Leptonic $W$ + untagged jet.
     Cases 1 and 2 offer the highest signal sensitivity.
• Background Estimation (ABCDnn): The dominant Standard Model backgrounds (primarily $t\bar{t}$, $W$+jets, and QCD multijets) are modeled using a novel data-driven technique called "ABCDnn." This method extends the traditional ABCD method by employing neural autoregressive flows to learn a transformation between five control regions (CRs) and the signal region (SR). The network maps simulated distributions to data distributions in the CRs and extrapolates this transformation to the SR, using the reconstructed B′ mass ($m_{tW}$) and the scalar sum of transverse momenta ($S_T$) as transform variables. Minor backgrounds are estimated via simulation.
• Statistical Analysis: A binned maximum likelihood fit is performed on the $m_{tW}$ distribution in the signal region. Systematic uncertainties, including those from jet energy scale, $b$-tagging, and the ABCDnn training accuracy, are incorporated as nuisance parameters.

Key Contributions

• Novel Background Modeling: The application of the ABCDnn technique, utilizing neural autoregressive flows, represents a significant advancement in modeling complex background shapes in high-mass resonance searches, allowing for data-driven predictions that account for correlations between observables without relying on strict independence assumptions.
• Enhanced Sensitivity to Narrow Widths: By combining the full Run 2 dataset (138 fb$^{-1}$) with advanced jet identification (PARTICLENET) and the ABCDnn background estimation, this analysis achieves the highest sensitivity to date for single B′ production with narrow decay widths.
• Comprehensive Coupling Limits: The study provides limits not only on the production cross section but also on the coupling factor ($\kappa$) of the B′ quark to electroweak bosons for both singlet and doublet scenarios.

Results
No significant excess of events over the Standard Model background prediction was observed in any reconstruction category. The background-only fit yielded a $p$-value of 0.15.

• Exclusion Limits:
  • For singlet B′ quarks produced in association with a $b$ quark ($bqB'$) with a 5% relative decay width, masses between 0.8 and 1.23 TeV are excluded at 95% confidence level (CL). The expected exclusion limit was 1.35 TeV.
  • For $t$-associated production ($tqB'$), the observed limits exclude masses of 0.9 and 1.0 TeV for both the 5%-width singlet and 1%-width doublet scenarios.
  • Upper limits on the production cross section for $tqB'$ production range from 230 fb at 0.8 TeV to 10 fb at 2.0 TeV.
• Coupling Factor Limits:
  • For singlet $bqB'$ production, coupling factors $\kappa > 0.30$ are excluded for masses between 0.8 and 1.8 TeV.
  • For $tqB'$ production, $\kappa > 0.41$ is excluded for singlet masses (0.8–1.2 TeV), and $\kappa > 0.29$ is excluded for doublet masses (0.8–2.0 TeV).

Significance
The paper claims that this search provides the most stringent sensitivity to date for the single production of narrow-width B′ quarks. It successfully probes smaller cross sections and coupling factor values than previous CMS searches, effectively extending the reach of VLQ searches into the regime of narrow widths and weak couplings. The results serve as a model-independent complement to finite-width studies, constraining the parameter space for vector-like quarks that mix with third-generation SM quarks.