Observation of $t\bar{t}γγ$ production at $\sqrt{s}=$13 TeV with the ATLAS detector

Using 140 fb$^{-1}$ of proton-proton collision data at $\sqrt{s}=13$ TeV collected by the ATLAS detector, this paper reports the first observation of top-quark pair production in association with two photons ($t\bar{t}\gamma\gamma$) with a significance of 5.2 standard deviations and measures its fiducial cross section to be $2.42^{+0.58}_{-0.53}$ fb.

The Gist

The Big Picture: Catching a Rare Double-Act

Imagine the Large Hadron Collider (LHC) as the world's most powerful, high-speed bowling alley. Instead of bowling balls, they smash protons together at nearly the speed of light. Usually, when these protons collide, they create a chaotic explosion of particles, like a pinball machine gone wild.

Most of the time, you get standard debris: electrons, quarks, and photons (particles of light). But sometimes, very rarely, the collision produces something spectacular: a pair of heavy "top quarks" (the heaviest known particles in the universe) that immediately burst apart, and in the process, they shoot out two bright flashes of light (photons) at the same time.

This paper is the official announcement that the ATLAS detector at CERN has finally caught this rare event. It's the first time anyone has definitively seen a top-quark pair production accompanied by two photons.

The Analogy: The "Double-Flash" Fireworks

To understand why this is a big deal, let's use an analogy:

• The Top Quarks: Think of these as two massive, heavy fireworks shells. They are so heavy that they are hard to make.
• The Photons: These are the sparks or flashes of light.
• The Standard Event ($t\bar{t}\gamma$): Usually, when these heavy shells explode, they might shoot out one bright spark. Scientists have seen this many times before. It's like a standard firework display.
• The New Discovery ($t\bar{t}\gamma\gamma$): This paper is about catching the moment when the explosion shoots out two distinct, bright sparks simultaneously.

Why is this hard? Because the universe prefers to be lazy. It's much easier to shoot out one spark than two. The odds of getting two sparks are roughly one in a thousand compared to getting just one. It's like trying to win a lottery where you have to match two specific numbers in a row, rather than just one.

How They Did It: The "Super-Sieve"

The ATLAS detector is a giant, 3D camera surrounding the collision point. Every second, it takes millions of pictures of these collisions. But most of these pictures are boring background noise (like static on an old TV).

To find the "Double-Flash" event, the scientists had to build a Super-Sieve:

1. The Filter (Event Selection): They looked for a very specific pattern in the data. They needed to see:
   • Two heavy top quarks (which decay into other particles like jets and a single electron or muon).
   • Exactly two high-energy photons.
   • No extra "noise" (like extra electrons) that would ruin the picture.
2. The Detective (The BDT): Even with the filters, there were still imposters. Some collisions look like the real thing but are actually fakes (like a sparkler that looks like a firework). To solve this, they used a Boosted Decision Tree (BDT).
   • Think of the BDT as a super-smart detective. It was trained on millions of computer simulations to look at tiny details: How far apart are the sparks? How heavy are the debris? What is the angle of the explosion?
   • The detective gives every event a "suspicion score." If the score is high enough, it's a real Double-Flash event.

The Results: A 5-Star Discovery

After sifting through 140 billion collisions (collected over four years, 2015–2018), the team found their signal.

• The Count: They found enough events to be sure it wasn't just a lucky fluke.
• The Significance: In the world of physics, you need a "5-sigma" result to claim a discovery. This is like flipping a coin and getting "Heads" 50 times in a row. The odds of this happening by random chance are about 1 in 3.5 million. The ATLAS team got a 5.2 sigma result. They are 99.9999% sure they saw it.
• The Measurement: They measured how often this happens (the "cross-section"). It turns out to be about 2.42 femtobarns.
  • Analogy: A femtobarn is a tiny unit of area, roughly the size of a proton. Saying the rate is 2.42 femtobarns is like saying, "If you shot a billion bullets at a target the size of a proton, you'd hit this specific double-flash pattern about 2.4 times."

Why Does This Matter?

You might ask, "So what? We just saw two sparks. Who cares?"

Here is why physicists are excited:

1. Testing the Rules of the Universe: The Standard Model is our rulebook for how particles interact. It predicts exactly how often these double-sparks should happen. By measuring it, scientists are stress-testing the rulebook. If the number they measured is different from the prediction, it could mean there are new, unknown forces or particles hiding in the shadows.
2. The "Background" Problem: This process is actually a major "background noise" for finding the Higgs boson. The Higgs boson sometimes decays into two photons. To find the Higgs, you have to subtract the "Double-Flash" noise. Now that we know exactly how loud that noise is, we can find the Higgs (and potentially new physics) more clearly.
3. The Ratio: They also measured the ratio of "Double-Flash" events to "Single-Flash" events. This is like measuring the ratio of double-yolk eggs to single-yolk eggs. It helps refine our understanding of how the top quark interacts with light.

The Conclusion

In simple terms: The ATLAS team successfully found a needle in a haystack the size of a mountain.

They proved that top quarks can indeed produce two photons at once, just as the laws of physics predicted, but at a rate so low it took four years of data and a super-smart computer detective to find it. This discovery doesn't break the laws of physics; instead, it confirms them with incredible precision and gives scientists a sharper tool to look for the next big discovery.

Technical Summary

1. Problem and Motivation

• Scientific Context: Measurements of top-quark pair production ($t\bar{t}$) in association with neutral vector bosons ($\gamma, Z$) provide a direct probe of the electroweak couplings of the top quark. While $t\bar{t}\gamma$ production has been extensively studied, the associated production of top-quark pairs with two photons ($t\bar{t}\gamma\gamma$) remains largely unexplored.
• Theoretical Gap: The $t\bar{t}\gamma\gamma$ process is a rare occurrence (expected cross-section at the per-mille level relative to $t\bar{t}\gamma$). It serves as a critical irreducible background for Higgs boson production in association with top quarks ($t\bar{t}H$) where the Higgs decays to two photons ($H \to \gamma\gamma$). Furthermore, it offers sensitivity to the top-quark's electric charge, electroweak dipole moments, and anomalous chromomagnetic/chromoelectric dipole moments.
• Simulation Challenges: Current Monte Carlo (MC) generators cannot simulate $t\bar{t}\gamma\gamma$ (and $t\bar{t}\gamma$) with high-order accuracy when including photons radiated from all charged particles (initial state, top quarks, and decay products) simultaneously. Existing theoretical calculations often assume stable top quarks or parton-level decays, making direct comparison with experimental data difficult.
• Objective: The primary goal is to provide the first reference measurement of the $t\bar{t}\gamma\gamma$ fiducial cross-section and the ratio of $t\bar{t}\gamma\gamma$ to $t\bar{t}\gamma$ cross-sections to guide future theoretical developments, rather than focusing on immediate comparison with existing simulations.

2. Methodology

• Dataset: The analysis utilizes proton-proton collision data collected by the ATLAS detector at the Large Hadron Collider (LHC) during Run 2 (2015–2018) at a center-of-mass energy of $\sqrt{s} = 13$ TeV. The integrated luminosity is 140 fb$^{-1}$.
• Event Selection (Single-Lepton Channel):
  • Triggers: Single-electron or single-muon triggers with $p_T$ thresholds between 20 and 27 GeV.
  • Object Requirements:
    • Exactly two isolated photons with $p_T > 20$ GeV and $|\eta| < 2.37$ (excluding the transition region $1.37 < |\eta| < 1.52$).
    • Exactly one isolated electron or muon with $p_T > 25$ GeV.
    • At least four jets with $p_T > 25$ GeV, including at least one $b$-tagged jet.
  • Kinematic Cuts: Invariant mass of the lepton and any photon is vetoed in the $Z$-boson mass window (86.19–96.19 GeV) to suppress $Z \to ee$ background.
• Background Estimation:
  • Prompt Photons: Estimated using MC simulations (e.g., $V\gamma\gamma$, $t\bar{t}H$, $t\bar{t}\gamma$).
  • Fake Photons: Estimated using data-driven methods:
    • e-fake: Electrons misidentified as photons (estimated using $Z \to e^+e^-$ control regions).
    • h-fake: Hadrons misidentified as photons (estimated using the "ABCD" method).
  • After selection, background processes account for approximately 50% of the total events.
• Signal Extraction:
  • A Boosted Decision Tree (BDT) classifier (XGBoost) is trained to separate the $t\bar{t}\gamma\gamma$ signal from background processes using 19 input variables (photon properties, angular separations, invariant masses, $E_T^{miss}$, and jet properties).
  • The cross-section is measured using a profile likelihood fit to the BDT output distribution in a fiducial phase space defined at the particle level.
  • A simultaneous fit is performed to extract the ratio $R = \sigma(t\bar{t}\gamma\gamma) / \sigma(t\bar{t}\gamma)$, utilizing inputs from a previous $t\bar{t}\gamma$ analysis (Ref. [13]) to cancel normalization uncertainties.

3. Key Contributions

• First Observation: This paper reports the first observation of $t\bar{t}\gamma\gamma$ production.
• Fiducial Cross-Section Measurement: Provides a precise measurement of the fiducial cross-section at the particle level, defined by specific kinematic criteria for leptons, photons, and jets.
• Ratio Measurement: Determines the ratio of the $t\bar{t}\gamma\gamma$ to $t\bar{t}\gamma$ cross-sections, a key observable for constraining anomalous dipole moments.
• Reference for Theory: Establishes a benchmark dataset for future theoretical calculations, as current high-order simulations for this specific multi-photon final state are not yet available for direct comparison.

4. Results

• Observed Significance: The signal is observed with a significance of 5.2 standard deviations (expected significance: 3.8$\sigma$ for LO simulation, 5.7$\sigma$ with an approximate K-factor of 1.7).
• Fiducial Cross-Section:
  $$\sigma_{t\bar{t}\gamma\gamma} = 2.42^{+0.58}_{-0.53} \text{ fb}$$
  • Statistical uncertainty: $\pm 0.46$ fb.
  • Systematic uncertainty: $^{+0.35}_{-0.38}$ fb.
  • Total relative uncertainty: ~23% (dominated by statistical uncertainty, ~17%).
• Cross-Section Ratio:
  $$R_{t\bar{t}\gamma\gamma/t\bar{t}\gamma} = (3.30^{+0.70}_{-0.65}) \times 10^{-3}$$
  • Total relative uncertainty: ~20%.
  • The ratio measurement benefits from the cancellation of systematic uncertainties related to jets, $b$-tagging, photons, and luminosity.
• Comparison with Theory: The measured cross-section is higher than the Leading Order (LO) prediction from MadGraph5_aMC@NLO + Pythia 8 ($1.53^{+0.40}_{-0.52}$ fb), consistent with expectations that LO simulations underestimate the cross-section (indicated by a K-factor of ~1.7).

5. Significance

• Validation of Standard Model: The observation confirms the existence of a rare Standard Model process involving multiple electroweak bosons and heavy quarks.
• Background for Higgs Physics: The precise measurement of $t\bar{t}\gamma\gamma$ is crucial for reducing systematic uncertainties in the search for and measurement of the $t\bar{t}H$ process with $H \to \gamma\gamma$, a key channel for studying the top-Higgs Yukawa coupling.
• New Physics Constraints: The results provide new constraints on the top-quark's electroweak dipole moments and anomalous couplings. The ratio measurement is particularly powerful as it minimizes experimental systematic errors, offering a clean probe for Beyond the Standard Model (BSM) physics.
• Theoretical Development: By providing the first data-driven reference, this work stimulates the development of higher-order theoretical calculations (NLO/NNLO) for $t\bar{t}\gamma\gamma$ production, which are currently lacking.