Variable coherence model for free-electron laser pulses

Imagine you are trying to take a perfect, high-resolution photograph of a tiny, fast-moving atom. To do this, you need a super-bright flash of light. In the world of science, the best "flash" we have is a Free-Electron Laser (FEL).

However, there's a catch. The most powerful type of these lasers (called SASE) doesn't produce a clean, steady beam like a laser pointer. Instead, it's more like a chaotic storm of light.

The Problem: The "Static" on the Light

Think of a standard laser beam as a smooth, calm river. A SASE FEL pulse, on the other hand, is like a river that has been churned up by a storm. It's full of random spikes, bumps, and "static."

In scientific terms, this light is stochastic (random). If you look at the light pulse, it's not one smooth wave; it's a jumbled mess of tiny, random "sub-pulses" (little spikes of energy) fighting each other. This randomness makes it hard for scientists to predict exactly how the light will interact with the atoms they are studying. It's like trying to bake a cake when your oven temperature jumps up and down randomly every second.

The Solution: The "Variable Coherence Model" (VCM)

The authors of this paper, Austin, Nils, and François, invented a new computer simulation tool called the Variable Coherence Model (VCM).

Think of the VCM as a "Randomness Dial" or a "Chaos Knob" on a mixing board.

Turn the knob to the left (Zero Coherence): You get the most chaotic, messy light possible. It looks like static on an old TV. This is what current real-world FELs mostly produce.
Turn the knob to the right (High Coherence): The static disappears. The light becomes smooth, organized, and perfect, like a calm river. This is what scientists wish they had to get the clearest pictures.
Turn the knob anywhere in between: You can create light that is "somewhat" messy. This allows scientists to test exactly how much chaos their experiments can handle.

How They Tested It

The researchers didn't just guess; they ran thousands of computer simulations to see what happens when they turn this "Chaos Knob."

Counting the Spikes: They looked at how many "sub-pulses" (the little spikes of light) appeared in a single flash.
- Analogy: Imagine a crowd of people shouting.
- Low Coherence: It sounds like a chaotic riot where 20 different people are shouting at once.
- High Coherence: It sounds like one person speaking clearly.
- The Finding: As they turned the knob to increase coherence, the "riot" of 20 voices slowly quieted down until only one clear voice remained.
Time vs. Frequency: They checked if the chaos looked different depending on whether they watched the light over time (how long the flash lasts) or frequency (the colors of the light).
- They found that the light behaves differently in these two views, kind of like how a spinning coin looks like a blur from the side but a clear circle from the top. The "Chaos Knob" affects these two views at different speeds.

Why Does This Matter? (The Absorption Experiment)

The most important part of the paper is what happens when they shine this light on an atom.

They simulated a model atom and shone their "random" light on it to see how the atom absorbed the energy.

The Result: When the light was very chaotic (low coherence), the atom reacted differently than when the light was smooth. The "peak" of the reaction was weaker and shifted slightly.
The Lesson: If scientists use the old, messy models to predict how atoms react, they might get the wrong answer. By using the VCM, they can now simulate the exact amount of messiness their real laser has and get a much more accurate prediction of how the atom will behave.

The Big Picture

This paper is like giving scientists a calibration tool.

Before this, they had to assume their laser light was either "perfectly smooth" (which it isn't) or "completely random" (which is too messy to work with). Now, they have a tool that lets them dial in the exact level of imperfection their real laser has.

This means:

Better Experiments: They can design experiments that account for the laser's "jitters."
Better Data: They can interpret their results more accurately, knowing exactly how the randomness of the light affected the atom.
Future Tech: It helps them understand how to eventually build lasers that are smoother and more controlled.

In short, the authors built a virtual laboratory where they can control the "noise" of a laser beam, helping real-world scientists see the universe more clearly through the static.

Here is a detailed technical summary of the paper "Variable coherence model for free-electron laser pulses."

1. Problem Statement

Free-electron lasers (FELs) operating in the Self-Amplified Spontaneous Emission (SASE) regime generate high-intensity, ultra-short X-ray pulses essential for attosecond science. However, SASE pulses are inherently stochastic, characterized by random intensity spikes (sub-pulses) and limited temporal/spectral coherence. This shot-to-shot variability complicates the interpretation of experimental data, particularly in nonlinear spectroscopy and pump-probe schemes.

Existing simulation models, such as the partial coherence model (Pfeifer et al., 2010), typically generate pulses with a fixed, minimal coherence width (effectively $\Omega = 0$ ). This limits the ability to simulate the full spectrum of FEL behaviors, ranging from highly stochastic SASE pulses to fully coherent, Fourier-limited pulses. There is a need for a flexible simulation framework that allows for the continuous control of coherence while maintaining fixed average pulse parameters (central frequency, bandwidth, and duration) to systematically study the impact of stochasticity on physical processes.

2. Methodology

The authors introduce the Variable Coherence Model (VCM), a phenomenological approach to simulate SASE pulses with a tunable coherence width ( $\Omega$ ).

Mathematical Formulation:
The electric field $E(t)$ is constructed via an inverse Fourier transform of a frequency-domain window function $W_F(\omega)$ modulated by a stochastic phase $\phi(\omega)$ :
$E(t) = N W_T(t-t_0) \mathcal{F}^{-1} [W_F(\omega-\omega_0) e^{i\phi(\omega)}]$
The phase is decomposed into a coherent part ( $\phi_c = -\omega t_0$ ) and a stochastic part ( $\phi_v$ ).
Stochastic Phase Generation:
- Zero Coherence Limit ( $\Omega = 0$ ): $\phi_v$ is sampled from a uniform random distribution $[0, 2\pi]$ , recovering the standard partial coherence model.
- Variable Coherence ( $\Omega > 0$ ): $\phi_v$ is generated using a Lévy process (specifically Brownian motion in this study). The phase difference between consecutive frequency samples is governed by:
  $\phi_v(\omega_{k+1}) = \phi_v(\omega_k) + \pi \sqrt{\frac{\omega_{k+1} - \omega_k}{\Omega}} \Phi$
  where $\Phi$ is a random variable. A smaller $\Omega$ increases phase randomness, while a larger $\Omega$ enforces phase correlation across the spectrum, approaching a constant phase (fully coherent) as $\Omega \to \infty$ .
Simulation Parameters:
The model was tested across three distinct FEL parameter sets corresponding to LCLS-I (long soft X-ray), LCLS-II (short soft X-ray), and FLASH (XUV), varying in central frequency ( $\omega_0$ ), bandwidth ( $F$ ), and pulse duration ( $T$ ).
Statistical Analysis:
The authors simulated 10,000 pulses per parameter set and coherence width. They developed an algorithm to detect "sub-pulses" (local intensity maxima) in both time and frequency domains, applying prominence thresholds to filter noise.
Absorption Simulation:
To illustrate physical implications, the authors performed Time-Dependent Density Functional Theory (TDDFT) simulations on a 1D atomic model (4 active electrons) to calculate nonlinear absorption cross-sections under varying VCM pulse conditions.

3. Key Contributions

Development of VCM: A novel simulation framework that bridges the gap between minimally coherent SASE pulses and fully coherent Fourier-limited pulses via a single tunable parameter ( $\Omega$ ).
Systematic Statistical Characterization: A comprehensive analysis of how coherence width affects sub-pulse statistics (intensity and number) in both time and frequency domains.
Cross-Domain Correlation Analysis: Investigation of the joint probability distributions of sub-pulse numbers and intensities between time and frequency domains.
Nonlinear Spectroscopy Impact: Demonstration of how pulse stochasticity (coherence width) alters nonlinear absorption features, highlighting the necessity of accurate pulse modeling in theoretical spectroscopy.

4. Key Results

A. Sub-pulse Statistics (Intensity and Number)

Sub-pulse Count: As $\Omega$ $Ω$ increases, the average number of sub-pulses decreases monotonically from a maximum at $\Omega=0$ $Ω = 0$ to one (the Fourier limit) at high $\Omega$ $Ω$ .
- In the frequency domain, the number of sub-pulses converges to 1 relatively quickly (when $\Omega \approx 5-6 \times$ bandwidth).
- In the time domain, convergence is slower (requiring $\Omega \approx 4-5 \times$ bandwidth) because increasing coherence narrows the central peak but leaves low-intensity "pedestals" that are counted as sub-pulses until $\Omega$ is very large.
Sub-pulse Intensity:
- Frequency Domain: At low $\Omega$ , intensity distributions are broad. As $\Omega$ increases, the distribution narrows and converges to the Fourier-limited intensity.
- Time Domain: At low $\Omega$ , intensities are concentrated at lower values. As $\Omega$ increases, the most probable intensity rises toward the asymptotic Fourier limit.
- LCLS-II Anomaly: For the LCLS-II set (very short pulses), the average sub-pulse intensity is maximized at $\Omega=0$ and declines as coherence increases, unlike the other regimes. This is attributed to the short pulse duration relative to the bandwidth.

B. Time-Frequency Correlations

Sub-pulse Number: There is a weak positive correlation between the number of sub-pulses in time and frequency at low coherence. As coherence increases, the correlation strengthens, and the distributions concentrate along the diagonal (fewer sub-pulses in both domains).
Sub-pulse Intensity: There is generally no strong one-to-one correlation between specific time-domain and frequency-domain sub-pulse intensities, especially at low coherence. The joint probability distributions highlight that time and frequency domains approach their Fourier-limited limits at different rates.

C. Nonlinear Absorption Effects

Convergence: Simulations of a model atom showed that at zero coherence ( $\Omega=0$ ), a massive number of pulses (thousands) are required to converge the average absorption response function due to shot-to-shot fluctuations. As $\Omega$ increases, convergence is achieved with significantly fewer pulses.
Spectral Reshaping: The average absorption response function for low-coherence pulses differs significantly from the fully coherent case. Specifically, the peak of the response function is suppressed (~10% lower) and shifted compared to the Fourier-limited reference. This demonstrates that assuming fully coherent pulses in theoretical models can lead to significant errors in predicting nonlinear X-ray interactions.

5. Significance

The Variable Coherence Model provides a critical tool for the X-ray free-electron laser community. By enabling the continuous tuning of coherence, the VCM allows researchers to:

Benchmark Experiments: Systematically isolate the effects of pulse stochasticity from other experimental variables.
Improve Theoretical Accuracy: Highlight that neglecting the stochastic nature of SASE pulses in nonlinear spectroscopy simulations can yield incorrect predictions regarding absorption features and response magnitudes.
Optimize Pulse Shaping: Serve as a foundation for simulating advanced pulse shaping techniques (e.g., chirped pulses) where controlling the degree of coherence is essential.

The study concludes that while SASE pulses are often treated as random noise, their statistical properties are highly dependent on the coherence width, and accurate modeling of this parameter is vital for interpreting ultrafast X-ray science experiments.