⚛️ quantum physics

Sampling methods to describe superradiance in large ensembles of quantum emitters

This paper introduces and benchmarks two approximate numerical sampling methods, enhanced by offset corrections, to accurately calculate the photon statistics of superradiance in large quantum emitter ensembles where exact calculations are intractable due to exponential Hilbert space scaling.

Original authors: Daniel Eyles, Emmanuel Lassalle, Adam Stokes, Ramón Paniagua-Domínguez, Ahsan Nazir

Published 2026-01-30

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel Eyles, Emmanuel Lassalle, Adam Stokes, Ramón Paniagua-Domínguez, Ahsan Nazir

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a massive choir of singers (quantum emitters) standing in a grid. When they all sing together, they don't just make a louder noise; they can synchronize perfectly to create a beam of sound that shoots out in a specific direction. In physics, this phenomenon is called superradiance.

The scientists in this paper want to predict exactly how these singers behave, specifically looking at the "statistics" of their notes (how likely it is to hear two notes at once). This measurement is called $g^{(2)}$ .

The Problem: The Math is Too Hard
If you have just a few singers, you can calculate exactly how they interact. But if you have 64, 100, or 1,000 singers, the math becomes impossible. The number of possible ways they can interact grows so fast (exponentially) that even the world's fastest supercomputers would take longer than the age of the universe to solve it exactly.

The Solution: The "Sampling" Strategy
Since they can't solve the whole choir at once, the authors developed a clever trick: Sampling. Instead of listening to the whole choir, they listen to small, random groups of singers, calculate how those small groups behave, and then average the results to guess what the whole choir is doing.

They tested two different ways of doing this sampling:

1. The "Pairwise" Method (The "Duet" Approach)

How it works: You pick random pairs of singers, calculate how they sing together, and ignore how they sing alone. You do this thousands of times and average the results.
The Flaw: By ignoring the solo singers, this method tends to overestimate the excitement. It's like assuming every time two people high-five, the whole room is going wild, even if the rest of the room is quiet.
When it works best: It works well when the choir is huge (many emitters).

2. The "m-wise" Method (The "Small Group" Approach)

How it works: Instead of just pairs, you pick random groups of $m$ singers (where $m$ could be 3, 4, 5, etc.). You calculate how that specific group behaves, including their solo moments, and average the results.
The Flaw: Because you are counting solo moments multiple times as you shuffle through different groups, this method tends to underestimate the excitement. It's like being so focused on individual singers that you miss the energy of the crowd.
When it works best: It works well when the choir is smaller (or when you can afford to pick larger groups).

The "Offset" Fix

The authors realized these methods weren't perfect. The "Duet" method was too high, and the "Small Group" method was too low.

They discovered a mathematical "correction factor" (an offset) they could add to the results.
Think of it like a scale that always reads 5 pounds too light. You just add 5 pounds to the final number to get the truth.
By applying these corrections, they made both methods much more accurate.

The Golden Rule: Which Method to Use?

The paper found a simple rule for which method to choose based on the size of the choir ( $N$ ) and the size of your sample group ( $m$ ):

If the choir is small (specifically, if $N < 2m$ ): Use the m-wise (Small Group) method.
If the choir is large (specifically, if $N > 2m$ ): Use the Pairwise (Duet) method.

The "Safety Net"

The most powerful part of their discovery is that these two methods act as bookends.

The "Pairwise" method gives you an upper limit (the maximum possible excitement).
The "m-wise" method gives you a lower limit (the minimum possible excitement).
By using both, you create a "window" that is guaranteed to contain the true answer, even if you can't calculate the exact number.

Summary

The paper doesn't invent new physics; it invents a new calculator for complex quantum systems. It shows that by taking random samples of a large group and applying a simple mathematical "tweak" (the offset), scientists can accurately predict how large groups of quantum emitters will behave. This allows them to study superradiance in systems that were previously too big to understand.

Technical Summary: Sampling Methods for Superradiance in Large Ensembles

Problem Statement
Superradiance, a quantum phenomenon where coherence between emitters leads to enhanced and directional radiation, is characterized by the second-order quantum correlation function, $g^{(2)}(t, \tau)$ . While $g^{(2)}$ is essential for understanding photon statistics (distinguishing between bunching/superradiance where $g^{(2)} > 1$ , and antibunching/subradiance where $g^{(2)} < 1$ ), its exact calculation becomes intractable for large ensembles of $N$ emitters. The dimension of the emitter Hilbert space scales exponentially as $2^N$ , rendering exact numerical solutions impossible for large $N$ . Consequently, researchers must rely on approximate approaches, such as mean-field theory, which may fail to capture complex collective effects.

Methodology
The authors propose and benchmark two numerical sampling methods (SMs) designed to approximate $g^{(2)}(t, 0)$ for large ensembles embedded in electromagnetic environments. Both methods rely on averaging results over many realizations of randomly selected subsets of emitters:

Pairwise Sampling Method: Adapted from existing literature, this approach selects random pairs of emitters ( $m=2$ ). It computes the exact dynamics for these pairs, including all mutual interactions, but explicitly neglects single-emitter contributions to avoid overcounting during averaging. Theoretically, this neglect leads to a systematic overestimation of correlations.
$m$ -wise Sampling Method: A complementary approach introduced in this work, where random samples of size $m$ (where $1 < m < N$ ) are selected. Unlike the pairwise method, this approach retains single-emitter terms and captures higher-order clusters. Due to the unavoidable overcounting of single-emitter terms during the sampling process, this method yields a systematic underestimation of the normalized correlation function.

To address the systematic biases of both methods, the authors introduce offset corrections. These are constant offsets derived from the limit of infinite emitter separation ( $d \to \infty$ ), where emitters behave independently.

For the $m$ -wise method, the offset is $1/m - 1/N$ .
For the pairwise method, the offset is $1/N$ .
These corrections are designed to improve quantitative accuracy while preserving the bounding character of the methods (i.e., ensuring the $m$ -wise method remains a lower bound).

Key Results and Benchmarks
The methods were benchmarked against exact solutions in two regimes:

Fully Inverted Arrays (Undriven): An $8 \times 8$ $8 \times 8$ square lattice of 64 emitters and a 1D chain of 9 emitters. Exact closed-form expressions or numerical solutions were available.
- The uncorrected $m$ -wise method with $m=2$ failed to predict superradiance ( $g^{(2)} > 1$ ) for inverted states, a limitation of the sample size rather than the physics. Superradiance was recovered only when $m > 2$ .
- The uncorrected pairwise method significantly outperformed the uncorrected $m$ -wise method for the 64-emitter array.
- Upon applying offset corrections, the accuracy of both methods improved substantially, reducing errors from ~10% to ~0.5% in specific regimes.
Coherently Driven Arrays: A steady-state analysis of driven 1D chains ( $N=8$ $N = 8$ ) and square lattices ( $N=64$ $N = 64$ ).
- For $N < 2m$ , the $m$ -wise method provided qualitatively superior results even without corrections.
- For $N > 2m$ , the pairwise method became more accurate.
- The authors identify a critical transition point at $N = 2m$ , where the optimal choice of method switches. Below this threshold, the $m$ -wise method is preferred; above it, the pairwise method is preferred.

Significance and Claims
The paper claims to provide new theoretical tools for studying superradiance in large ensembles where exact calculations are infeasible. The primary contributions are:

Bounding Strategy: By employing both methods simultaneously, researchers can establish upper and lower bounds for $g^{(2)}(t, 0)$ , offering a "bounded window" of possible values rather than a single uncertain estimate.
Regime Optimization: The identification of the $N = 2m$ crossover allows for the selection of the most accurate method based on the total number of emitters and available computational resources (which limit the maximum feasible $m$ ).
Quantitative Improvement: The introduction of offset corrections significantly enhances the quantitative accuracy of both sampling methods while maintaining their theoretical bounding properties.

The authors conclude that these combined approaches offer a practical and accurate means to estimate photon statistics and characterize collective emission regimes in systems ranging from quantum networks to superradiant lasers, without requiring the exponential computational cost of exact diagonalization.