Construction of the Global $\chi^2$ Function for the Simultaneous Fitting of Correlated Energy-Dependent Cross Sections

This paper constructs a global $\chi^2$ function designed for the simultaneous fitting of correlated energy-dependent cross sections by incorporating correlations between different processes and energy points, as well as uncertainties from integrated luminosity and center-of-mass energy measurements.

The Gist

Imagine you are a detective trying to solve a mystery about a very specific, elusive suspect: a subatomic particle called the J/ψ meson. To catch this suspect, you need to know their exact "weight" (mass) and how long they usually stick around before disappearing (total width).

In the world of high-energy physics, you don't just look at the suspect once. You run thousands of experiments at different speeds (energies) to see how often they appear. This is like measuring the cross-section (the probability of the particle appearing) at various energy levels.

The Problem: The Messy Data

Usually, scientists measure these probabilities at different energy points. However, real-world data is messy.

1. Correlations: If you measure the suspect at Energy A and Energy B, the errors in your measurements aren't totally independent. Maybe your ruler was slightly bent, or your lighting was dim. If you made a mistake at Energy A, you likely made a similar mistake at Energy B.
2. The "Lighting" Issue (Luminosity): To know how many suspects you saw, you need to know how bright your "flashlight" (the collider's beam) was. If your flashlight flickers, it affects all your measurements at once.
3. The "Speedometer" Issue (Energy): You also need to know exactly how fast the particles were moving. If your speedometer is slightly off, it throws off your calculation of the probability.

If you try to fit your data using a simple method that ignores these connections, you might get a "best guess" that looks good but is actually wrong because you didn't account for how the errors were linked.

The Solution: The "Global Chi-Square" Function

The authors of this paper (Linquan Shao and colleagues) built a new, super-powered mathematical tool called a Global $\chi^2$ (Chi-Square) Function.

Think of the standard $\chi^2$ function as a simple scorecard. It asks: "How far off is my theory from the actual data?" The lower the score, the better the fit.

But the authors realized that for two different processes (let's call them Process A and Process B) happening at the same time, a simple scorecard isn't enough. They needed a Master Scorecard that understands the whole story.

Here is how their new tool works, using a creative analogy:

1. The "Double-Book" Ledger

Imagine you are keeping a ledger for two different bank accounts (Process A and Process B) that are run by the same bank manager.

• Standard Method: You check Account A and Account B separately. You see Account A has a $10 error and Account B has a $10 error. You think, "Okay, total error is $20."
• The Global Method: You realize both accounts share the same bank manager (the Integrated Luminosity). If the manager made a mistake in the ledger, both accounts are wrong by the same amount. The Global Method says, "Wait! These errors are linked. We can't just add them up; we have to treat them as a single, connected system."

2. The "Ruler" Effect (Center-of-Mass Energy)

Imagine you are measuring the height of two different trees (Process A and Process B) using a ruler.

• If your ruler is slightly stretched (an error in Energy Measurement), every tree you measure will appear taller than it really is.
• The authors' formula calculates exactly how much that "stretched ruler" affects the final result. It adds a specific "correction factor" to the scorecard that says: "Because our speedometer was slightly off, we need to adjust the probability calculation for both trees simultaneously."

3. The "Big Matrix"

To do this mathematically, the authors created a giant grid (a Covariance Matrix).

• Think of this grid as a massive spreadsheet where every cell tells you how much one measurement influences another.
• If you measured Process A at 10:00 AM and Process B at 10:05 AM, the spreadsheet knows they are related because they happened under the same "lighting" conditions.
• The formula takes this giant grid, flips it upside down (mathematically inverting it), and multiplies it by the differences between the theory and the data. This gives them a "Global Score" that is much more accurate than any individual score.

Why Does This Matter?

Before this paper, scientists might have had to choose: "Do I fit Process A first, then Process B?" or "Do I ignore the fact that my energy measurements are correlated?"

This new method allows them to fit everything at once.

• It treats the data like a complex web rather than a pile of separate threads.
• It accounts for the fact that if your "flashlight" (luminosity) flickers, it affects all your data points together.
• It accounts for the fact that if your "speedometer" (energy) is off, it shifts the whole curve.

The Bottom Line

The authors have built a universal translator for experimental physics. They took a messy situation where data points are tangled together (correlated) and created a single, elegant equation that untangles them.

This equation has already been used to measure the properties of the J/ψ meson with incredible precision at the BESIII experiment. It's like upgrading from a magnifying glass to a high-definition microscope, allowing scientists to see the true nature of the subatomic world without the blur of unaccounted-for errors.

And the best part? While they built it for two processes, the "blueprint" is so good that it can easily be expanded to handle three, four, or even more processes simultaneously. It's a flexible tool for the future of discovery.

Technical Summary

1. Problem Statement

In high-energy physics experiments, extracting resonant parameters (such as mass and total width) for hadrons (e.g., the $J/\psi$ meson) often requires the simultaneous fitting of energy-dependent cross sections from multiple physical processes.

• The Challenge: Standard fitting methods often treat data points as independent or fail to fully account for complex correlations.
• Specific Correlations: The authors identify three critical sources of correlation that must be handled simultaneously:
  1. Cross-process correlations: Correlations between the measured cross sections of different physical processes.
  2. Energy-point correlations: Correlations between measurements taken at different center-of-mass (CM) energy points.
  3. Systematic uncertainties: Correlations arising from shared systematic errors in integrated luminosity ($L$) and CM energy ($W$) measurements.
• Goal: To construct a rigorous global $\chi^2$ function that incorporates all these correlations to improve the precision of parameter extraction.

2. Methodology

The authors employ the standard covariance matrix method based on the least-squares principle. The construction involves the following steps:

A. Data Definition

The analysis considers $n$ energy points where two physical processes (Channel 1 and Channel 2) are measured. For each point, the following are known:

• Experimental cross sections: $\sigma_{ch1}^{exp}(i)$ and $\sigma_{ch2}^{exp}(i)$.
• Integrated luminosities: $L(i)$.
• CM energies: $W(i)$ with uncertainties $\Delta W(i)$.
• Covariance Matrices: $n \times n$ matrices are available for cross sections ($V_{ch1}, V_{ch2}$), luminosities ($V_L$), and a diagonal matrix for CM energies ($V_W$).

B. The Global $\chi^2$ Formulation

The global $\chi^2$ is defined as:
$$\chi^2 = \Delta\sigma^T \cdot V^{-1} \cdot \Delta\sigma$$
Where:

• $\Delta\sigma$ is a $2n$-dimensional vector containing the differences between experimental and theoretical cross sections for both channels across all $n$ energy points.
• $V$ is the global $2n \times 2n$ covariance matrix.

C. Handling Energy Uncertainties

A key methodological innovation is treating the theoretical cross sections ($\sigma^{the}$) as "semi-experimental" when CM energy uncertainties ($\Delta W$) are non-negligible. The authors define the covariance element $V(i, j)$ as the variance of the difference $(\sigma^{exp} - \sigma^{the})$.
Using error propagation, the covariance matrix elements are derived by summing contributions from:

1. Experimental cross-section covariance: Direct measurement correlations.
2. Luminosity correlations: Propagated through the relation $\sigma \propto N/L$.
3. Energy correlations: Propagated through the partial derivatives of the theoretical cross section with respect to energy ($\frac{\partial \sigma^{the}}{\partial W}$).

3. Key Contributions: The Global Covariance Matrix

The core contribution is the explicit derivation of the $2n \times 2n$ covariance matrix $V(i, j)$, which is divided into four distinct regions based on the indices $i$ and $j$ (representing the channel and energy point).

Let $i, j$ range from $1$ to $2n$. The first $n$ indices correspond to Channel 1, and $n+1$ to $2n$ correspond to Channel 2.

1. Diagonal Block (Channel 1 vs. Channel 1):
For $i, j \in [1, n]$:
$$V(i, j) = V_{ch1}(i, j) + \delta(i, j) \left( \frac{\partial \sigma_{ch1}^{the}}{\partial W(i)} \Delta W(i) \right)^2$$

• Term 1: Experimental covariance of Channel 1.
• Term 2: Uncertainty contribution from CM energy measurement (diagonal only).

2. Off-Diagonal Block (Channel 1 vs. Channel 2):
For $i \in [1, n]$ and $j \in [n+1, 2n]$ (where $j-n$ is the energy index for Channel 2):
$$V(i, j) = \frac{\sigma_{ch1}^{exp}(i) \sigma_{ch2}^{exp}(j-n)}{L(i) L(j-n)} V_L(i, j-n) + \delta(i, j-n) \frac{\partial \sigma_{ch1}^{the}}{\partial W(i)} \frac{\partial \sigma_{ch2}^{the}}{\partial W(i)} (\Delta W(i))^2$$

• Term 1: Correlation due to integrated luminosity measurement (shared $L$).
• Term 2: Correlation due to CM energy measurement (only if $i$ and $j-n$ refer to the same energy point).

3. Off-Diagonal Block (Channel 2 vs. Channel 1):
Symmetric to Case 2, swapping indices and channel labels.

4. Diagonal Block (Channel 2 vs. Channel 2):
For $i, j \in [n+1, 2n]$:
$$V(i, j) = V_{ch2}(i-n, j-n) + \delta(i, j) \left( \frac{\partial \sigma_{ch2}^{the}}{\partial W(i-n)} \Delta W(i-n) \right)^2$$

• Term 1: Experimental covariance of Channel 2.
• Term 2: Uncertainty contribution from CM energy measurement.

4. Results

• General Formula: The paper provides the complete analytical expression (Eq. 19) for the global covariance matrix elements, covering all combinations of channels and energy points.
• Validation: The method has been successfully applied to the measurement of the total and leptonic decay widths of the $J/\psi$ resonance at the BESIII experiment.
• Scalability: While derived for two channels, the authors note that the method can be easily generalized to three or more channels using the same covariance matrix construction logic.

5. Significance

• Precision Improvement: By rigorously accounting for correlations in luminosity and energy measurements, this method prevents the underestimation of uncertainties and the overestimation of statistical significance in simultaneous fits.
• Standardization: It provides a standardized framework for handling correlated data in high-energy physics, moving beyond simplified $\chi^2$ definitions that ignore systematic correlations.
• Applicability: The approach is critical for modern experiments (like BESIII, Belle II, LHCb) where high-precision resonance parameter extraction relies on combining multiple datasets with shared systematic errors.

In summary, this paper offers a mathematically rigorous solution for constructing a global $\chi^2$ function that fully integrates experimental correlations, luminosity uncertainties, and energy measurement errors, significantly enhancing the reliability of resonance parameter extraction in particle physics.