Reconstructing rare particle source by femtoscopic correlations

This paper introduces a novel Statistical Reconstruction method that bypasses conventional Gaussian assumptions to directly extract single-particle emission sources from rare particle yields via event-by-event analysis, successfully demonstrating its application to reconstructing the $J/\psi$ source in $pp$ collisions with a systematic uncertainty of approximately 13%.

The Gist

The Big Picture: Taking a "Femto-Photo" of the Invisible

Imagine you are trying to figure out the shape of a tiny, invisible balloon floating in a dark room. You can't see the balloon itself, but you can watch how two other objects (let's say, two tiny marbles) bounce off each other when they fly past it.

In the world of high-energy physics, scientists do something similar. They smash particles together at nearly the speed of light. When these particles fly apart, they leave behind a "fingerprint" called a correlation. By studying how these particles pair up, scientists try to reconstruct the shape and size of the "source" (the balloon) where they were born. This field is called femtoscopy (because it measures distances as small as a femtometer, which is one-quadrillionth of a meter).

The Problem: The "Rare Guest" Dilemma

For a long time, scientists have had a reliable way to guess the shape of these sources, but it only works well for very common particles (like pions or protons). They assume the source looks like a perfect, smooth Gaussian bell curve (like a classic hill).

However, the paper focuses on rare particles, specifically the $J/\psi$ (a heavy particle made of a charm quark and an anti-charm quark).

• The Issue: Because $J/\psi$ particles are so rare, you can't get enough data to build a perfect "bell curve" picture.
• The Old Way: Traditional methods try to measure the "pair" (the relationship between two particles). But for rare particles, we really want to know about the single particle source. The old methods are like trying to guess the shape of a single person's shadow by looking at a blurry photo of two people standing together. It's an indirect guess, and for rare particles, it often fails or relies on the wrong assumptions (like assuming the source is a perfect hill).

The Solution: A New Statistical "Reconstruction" Tool

The authors, led by Liang Zhang and colleagues, invented a new method called Statistical Reconstruction.

The Analogy: The Detective and the Echo
Imagine you are in a canyon (the particle source). You shout a word (the correlation), and it echoes back.

• The Old Way: You assume the canyon is a perfect circle, so you calculate how the echo should sound based on that assumption.
• The New Way: The authors say, "Let's not guess the shape. Let's listen to the echo particle-by-particle."

They treat the correlation data not as a single blurry image, but as a collection of individual clues.

1. The Reference: They use a "known" particle (protons) as a reference. Think of this as having a map of the canyon walls that we already know well.
2. The Kernel (The Clue): They calculate a mathematical "kernel" for every single rare particle. This kernel is like a unique "echo signature" that tells you how that specific rare particle interacted with the reference particles.
3. The Reconstruction: Instead of guessing the shape, they statistically reverse-engineer the source. They ask: "If the source looked like this, what would the collection of these individual echoes look like?" Then they adjust the source shape until the echoes match the real data.

The Experiment: Testing the Tool

To prove this works, they didn't just guess; they ran a massive simulation using a supercomputer program called EPOS4HQ.

• The Setup: They simulated 100,000 proton-proton collisions at the energy levels of the Large Hadron Collider (LHC).
• The Test: They "hid" the true shape of the $J/\psi$ source in the simulation. Then, they used their new method to try and find it, using the known proton source and theoretical physics (from something called HAL QCD) as their guide.

The Results: It Works!

• Success: The new method successfully reconstructed the shape of the $J/\psi$ source.
• Key Finding: The $J/\psi$ source turned out to be much more compact (smaller and tighter) than the proton source. This makes sense because $J/\psi$ particles are created very early in the collision, while protons are created later and spread out more.
• Accuracy: The method was very precise. When they compared their reconstructed source back to the original simulation, the error (uncertainty) was only about 13%.

Why This Matters (According to the Paper)

The paper claims this is a breakthrough because:

1. No More "Bell Curve" Assumptions: You don't have to assume the source is a perfect hill anymore. You can find out what it actually looks like.
2. Rare Particles: It finally allows scientists to study the "birthplaces" of rare, exotic particles that were previously too difficult to measure directly.
3. Direct Measurement: It moves from inferring a "pair" source to directly reconstructing the "single" particle source.

In short: The authors built a new statistical camera that can take a clear picture of the tiny, invisible "birth room" of rare particles, without needing to guess what the room looks like beforehand. They tested it in a computer simulation, and it worked with high accuracy.

Technical Summary

Technical Summary: Reconstructing Rare Particle Source by Femtoscopic Correlations

Problem Statement
Femtoscopy in high-energy nuclear collisions traditionally relies on measuring two-particle correlation functions to infer the space-time structure of particle-emission sources. Conventional analyses typically assume Gaussian parameterizations for these sources. While effective for abundant particles, this approach faces significant challenges when applied to rare particles (e.g., charm hadrons like $J/\psi$).

1. Indirect Inference: Standard femtoscopy accesses source information only through two-particle (pair) observables. Consequently, the single-particle emission source of a rare particle is implicitly folded into pair observables and cannot be directly measured.
2. Model Dependence: The assumption of Gaussian source shapes is not a priori valid, particularly for particles emitted during non-equilibrium states or early collision stages.
3. Inverse Problem: Reconstructing a source from correlation data is an ill-posed inverse problem. Existing advanced methods (e.g., machine learning, optical deblurring) operate at the level of two-particle sources, leaving the direct extraction of single-particle sources for rare particles as an open problem.

Methodology: Statistical Reconstruction
The authors introduce a Statistical Reconstruction method that reformulates femtoscopy as an inverse problem specifically for single-particle emission sources, bypassing the need for Gaussian assumptions.

• Theoretical Framework:
  The two-particle correlation function $C(k^*)$ is re-expressed not as a direct fit to a source distribution, but as an ensemble average over a single-particle-conditioned correlation kernel, $\tilde{C}_{k^*}(r_1)$.
  $$C(k^*) = \int dr_1 S_1(r_1) \tilde{C}_{k^*}(r_1)$$
  Here, $S_1(r_1)$ is the target single-particle source, and $\tilde{C}_{k^*}(r_1)$ represents the correlation between a particle at $r_1$ and the ensemble of the second particle.
  $$\tilde{C}_{k^*}(r_1) = \int d^3r_2 S_2(r_2) |\psi_{k^*}(r_1 - r_2)|^2$$
  The distribution of the kernel values, $y = \tilde{C}_{k^*}(r_1)$, encodes the underlying source $S_1(r_1)$.

• Experimental Accessibility:
  For rare particles where the yield is low (often one per event), the particle-by-particle distribution of the kernel can be transformed into an event-by-event extraction. The conditioned kernel for a target particle in event $i$ is defined as:
  $$\tilde{C}_{k^*;ij} = N \frac{N_{tot}^T}{N_{mixed}(k^*)} n_{ij}^{same}(k^*)$$
  This allows the experimental access to the statistical input required for reconstruction without inferring an effective pair source.

• Reconstruction Strategy:

  1. Inputs: The method requires (i) a well-characterized reference source for an abundant particle (e.g., protons) and (ii) a constrained interaction potential (e.g., from Lattice QCD/HAL QCD).
  2. Kernel Calculation: The analytical form of $\tilde{C}_{k^*}(r)$ is computed using the known reference source and interaction potential.
  3. Inversion: The source $S_1(r)$ is reconstructed by inverting the mapping between the source distribution and the kernel distribution.
  4. Stabilization: Since $\tilde{C}_{k^*}(r)$ can be non-monotonic (violating bijectivity), the method employs multi-momentum averaging. Reconstructions are performed across multiple relative momentum bins ($k^*$) and averaged to suppress artifacts and spurious peaks, yielding a robust estimate of the source.

Key Contributions and Results
The paper validates this method through two primary studies:

1. Controlled Numerical Test (Gaussian Sources):
   Using predefined Gaussian sources ($R_1=0.5$ fm, $R_2=1.4$ fm) and a simplified wavefunction, the authors demonstrated that the method can accurately reconstruct the source profile. They showed that while single-momentum reconstructions suffer from artifacts due to non-monotonicity, averaging over multiple $k^*$ bins recovers the ground-truth distribution with high fidelity.

2. Application to $J/\psi$ in $pp$ Collisions:
   The method was applied to reconstruct the $J/\psi$ emission source in $\sqrt{s} = 13.6$ TeV $pp$ collisions using EPOS4HQ simulations.

   • Setup: Protons with $p < 400$ MeV/c were used as the reference source (fitted with a Gaussian core + Cauchy tail model). The interaction potential ($N-J/\psi$) was derived from HAL QCD lattice QCD calculations.
   • Findings:
     • The reconstructed $J/\psi$ source is significantly more compact than the proton source, consistent with the expectation that charm hadrons are produced at the early, dense stage of the collision.
     • The most probable radius of the reconstructed $J/\psi$ source ($\sim 0.37$ fm) aligns better with the original EPOS4HQ simulation than a simple Gaussian fit to the pair correlation would suggest.
     • The reconstructed source preserves key characteristics (compactness and radius) despite the broader spread and enhanced tail introduced by the reconstruction process.
   • Systematic Uncertainty: By comparing the correlation function derived from the reconstructed source against the direct EPOS4HQ simulation, the authors estimate a systematic uncertainty of approximately 13%. This uncertainty arises primarily from the discretization of the kernel values and the model dependence of the reference proton source.

Significance and Claims
The paper claims that this Statistical Reconstruction method offers a novel pathway to extract single-particle emission sources for rare particles without relying on a priori Gaussian assumptions.

• It shifts the paradigm from inferring effective two-particle sources to directly reconstructing the single-particle source distribution.
• It leverages the statistical properties of rare particle production (where event-by-event fluctuations in multiplicity are suppressed) to make the problem experimentally accessible.
• The method provides quantitative constraints on rare-particle source characteristics, offering deeper insights into the intermediate processes of both $pp$ and heavy-ion collisions.
• The authors position this as a useful tool for future femtoscopy analyses, particularly for studying short-lived charm hadrons and exotic states where standard Gaussian parameterizations may be insufficient.

The study concludes that while the method introduces systematic uncertainties (estimated at ~13% in this simulation), it successfully demonstrates the feasibility of reconstructing compact, non-Gaussian sources for rare particles, addressing a fundamental limitation in current femtoscopic techniques.