Optimisation of the vertex detector and measurement of Higgs decays to second-generation quarks at the CEPC

This study utilizes an AI-driven Jet Origin Identification framework to demonstrate that optimizing the vertex detector's inner radius and spatial resolution at the CEPC significantly enhances the precision of measuring Higgs decays to second-generation quarks, particularly $H \to s\bar{s}$.

The Gist

Imagine the CEPC (Circular Electron-Positron Collider) as a massive, ultra-precise particle factory. Its main job is to smash electrons and positrons together to create Higgs bosons, the famous "God particle" that gives other particles mass. Once created, these Higgs bosons instantly decay (break apart) into other particles.

The scientists in this paper are trying to catch a very specific, rare type of decay: the Higgs turning into strange quarks (like a ghost) or charm quarks (like a shadow). These are the "second-generation" particles, and catching them is like finding a needle in a haystack made of other, much more common needles.

To do this, they need a super-sensitive camera called a Vertex Detector. Think of this detector as a high-speed, 3D motion tracker that watches exactly where particles are born.

The Problem: The "Inner Radius" and "Pixel Sharpness"

The paper asks a simple question: How should we build this camera to get the best results?

They focused on two main settings:

1. The Inner Radius: How close the camera's first layer is to the center of the collision (the beam pipe). Imagine a camera lens; the question is, "How close can the glass get to the action without getting in the way?"
2. Spatial Resolution: How sharp the camera's pixels are. Is it a blurry 1080p camera, or a crystal-clear 8K camera?

The Experiment: Turning the Dials

The researchers used a powerful computer simulation (like a video game engine for physics) to test different camera designs. They used an AI (Artificial Intelligence) system called "Jet Origin Identification" (JOI).

• The Analogy: Imagine you are trying to identify which of two people threw a ball.
  • If the ball is thrown from far away, it's hard to tell who threw it.
  • If the ball is thrown from right next to you, you can see the hand motion clearly.
  • The Inner Radius is about how close the camera gets to the "thrower" (the collision point).
  • The Spatial Resolution is about how clearly the camera sees the "hand motion."

The Findings: Proximity Wins

The study found that getting closer matters much more than having a sharper lens.

• Halving the distance (Inner Radius): When they moved the first layer of the detector twice as close to the center, the camera's ability to track the particles improved dramatically. It was like moving from the back row of a concert to the front row; suddenly, you could see exactly who was doing what.
  • Result: This improved the measurement of the rare "Charm" decay by 4% and the "Strange" decay by 8%.
• Doubling the distance: If they moved the camera twice as far away, the performance got significantly worse.
• Changing the sharpness (Resolution): Tweaking the pixel sharpness (making the lens twice as sharp or twice as blurry) had a very small effect. It was like having a slightly sharper lens when you are already sitting in the front row; it helps a tiny bit, but it doesn't change the view as much as moving your seat.

Why This Matters

The Higgs boson decaying into strange quarks is a "Holy Grail" measurement. It's incredibly rare (only about 1 in 4,000 Higgs bosons do this).

• The "Ghost" Hunt: The paper suggests that by optimizing the detector to be as close as possible to the collision point, we can increase our chances of spotting this rare "ghost" decay.
• The AI Advantage: The AI used in the study acts like a super-smart detective. It looks at the tiny tracks left by particles and says, "I'm 99% sure this came from a strange quark, not a background noise." The better the camera (the closer it is), the better the AI can do its job.

The Bottom Line

The paper concludes that for the future CEPC collider, designers should prioritize getting the detector layers as close to the beam as physically possible. While making the pixels sharper is nice, it's not the game-changer. Getting closer to the action is the key to unlocking the secrets of the Higgs boson's rarest behaviors.

In short: Don't just buy a better camera; move the camera closer to the stage.

Technical Summary

Technical Summary: Optimisation of the Vertex Detector and Measurement of Higgs Decays to Second-Generation Quarks at the CEPC

Problem Statement
Precision measurements of the Higgs boson at future electron-positron colliders, specifically the Circular Electron–Positron Collider (CEPC), are critical for probing physics beyond the Standard Model (SM). A key objective is the direct measurement of Yukawa couplings to second-generation quarks via the decay modes $H \to c\bar{c}$ and $H \to s\bar{s}$. While $H \to c\bar{c}$ is expected to be measured with percent-level accuracy, the observation of $H \to s\bar{s}$ (with a branching ratio of $\sim 2.3 \times 10^{-4}$) remains a significant challenge. The ability to distinguish these rare decays from dominant backgrounds (such as $H \to b\bar{b}$, $gg$, and light quarks) relies heavily on Jet Origin Identification (JOI). The performance of JOI is fundamentally limited by the geometric configuration and resolution of the vertex detector, particularly the innermost radius and spatial resolution of the silicon pixel sensors. This study addresses the need to quantify how variations in these detector parameters translate into physics sensitivity for second-generation quark measurements.

Methodology
The authors perform a conceptual optimization study using the $\nu\bar{\nu}H$ production channel (where $Z \to \nu\bar{\nu}$ or via $WW$-fusion) as a benchmark, corresponding to an integrated luminosity of $20 \text{ ab}^{-1}$ and approximately $0.9 \times 10^6$ Higgs bosons.

1. Simulation Framework: The study utilizes a fast simulation framework based on Delphes, validated against full Geant4-based simulations. The validation shows agreement within 5% for impact parameter resolution and less than 3% difference in jet-flavor migration matrix diagonal elements. Beam-induced backgrounds are neglected based on CEPC Technical Design Report (TDR) findings indicating their impact is under control.
2. Parameter Scans: The study varies two key geometric scaling factors relative to the CEPC baseline design:
   • $R_{rad}$: The scaling factor for the innermost layer radius ($r_{inner}$).
   • $R_{res}$: The scaling factor for the spatial resolution ($\sigma_{hit}$).
     Scans are performed with factors ranging from 0.5 (ambitious upgrade) to 2.0 (conservative degradation).
3. Analysis Chain:
   • Track Reconstruction: Impact parameter resolution ($\Delta d_0, \Delta z_0$) is quantified via the root-mean-square (RMS) of distributions.
   • Flavor Tagging: An AI-driven JOI algorithm (based on the ParticleNet architecture) processes kinematic variables, particle-type information, and impact parameters to assign likelihoods to 11 jet-flavor categories ($b, c, s, u, d, g$ and their antiparticles). Performance is evaluated using migration matrices and the trace of the matrix ($Tr(M)$).
   • Physics Benchmarks: Event-level Bayesian likelihoods are constructed to calculate the posterior probability $S_{f\bar{f}}$ for specific Higgs decay hypotheses. The optimal working point is determined by maximizing the product of efficiency ($\epsilon$) and purity ($P$).
   • Metrics: Statistical uncertainty for $H \to c\bar{c}$ and signal significance (Asimov significance) for $H \to s\bar{s}$ are calculated.

Key Contributions and Results

• Impact Parameter Resolution: The study establishes that the inner radius ($R_{rad}$) is the dominant factor in determining impact parameter resolution. Halving the inner radius improves the resolution by approximately a factor of two, whereas doubling it degrades performance similarly. In contrast, variations in spatial resolution ($R_{res}$) have a marginal effect, altering resolution by only about 10%.
• Flavor Tagging Performance:
  • Reducing the inner radius significantly lowers misidentification rates, particularly $b \to s$ and $c \to s$, which are critical for isolating the $H \to s\bar{s}$ signal.
  • The trace of the migration matrix ($Tr(M)$), representing total correct classification probability, shows a linear dependence on $\log_2(R_{rad})$ that is five times more sensitive than changes in spatial resolution.
  • The JOI algorithm (ParticleNet) demonstrates superior performance compared to traditional XGBoost methods, achieving misidentification rates one to three orders of magnitude lower in the relevant efficiency range.
• Physics Sensitivity:
  • $H \to c\bar{c}$: Halving the inner radius ($R_{rad}=0.5$) increases the signal yield and suppresses the $b\bar{b}$ background, reducing the relative statistical uncertainty by approximately 4% compared to the baseline. Doubling the radius increases uncertainty by a similar margin. Spatial resolution changes have negligible impact (<1%).
  • $H \to s\bar{s}$: The $H \to s\bar{s}$ measurement is highly sensitive to the inner radius. Halving the radius suppresses the $gg$ background and enhances the signal significance by 8% (from $4.42\sigma$ to $4.76\sigma$). Conversely, doubling the radius reduces significance to $4.23\sigma$. The study notes that the $b\bar{b}$ background is fully suppressed in all configurations due to the high performance of the JOI algorithm.
  • Overall Trend: The maximum $\epsilon \times P$ for $H \to c\bar{c}$ shifts by 8% per factor-of-two radius change, while $H \to s\bar{s}$ varies by 12%.

Significance and Claims
The paper claims that the inner radius of the vertex detector is the primary geometric parameter governing the sensitivity to rare Higgs decays to second-generation quarks at the CEPC. While spatial resolution is important, its impact is secondary compared to the inner radius.

The authors project that with the baseline CEPC design, the statistical precision for $H \to c\bar{c}$ will be in the range of 0.66–0.71%, and the significance for $H \to s\bar{s}$ will exceed $4\sigma$. The study emphasizes that optimizing the inner radius can improve these metrics by less than 10%, whereas degrading it has a comparable negative effect. The authors position these results as "upper-limit estimates" achievable under idealized background conditions, acknowledging that realistic beam-induced backgrounds and other Standard Model processes would further dilute sensitivity.

The work provides concrete guidance for detector design, suggesting that efforts to minimize the innermost layer radius are more impactful than marginal improvements in spatial resolution for the specific goal of identifying $H \to s\bar{s}$. The authors conclude that with an optimized vertex detector and continued advancements in analysis techniques (including better hadron identification and expanded production channels), the first unambiguous observation of $H \to s\bar{s}$ and a stringent test of second-generation Yukawa couplings are within reach.