🔭 astrophysics

High-statistics simulations of NewAthena WFI background using Geant4

This paper presents high-statistics Geant4 simulations of the NewAthena WFI background using a dual-approach of detailed and simplified geometry models, leveraging High Performance Computing to optimize throughput and provide critical insights for mitigating cosmic ray-induced background in X-ray astronomy.

Original authors: Matthew K. Heine, Catherine E. Grant, Marshall W. Bautz, Beverly J. LaMarr, Eric D. Miller, Michael W. J. Hubbard, David Hall, Joan Requena, Emanuele Perinati, Steven W. Allen, Artem Poliszczuk, Dan W

Published 2026-07-30

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Matthew K. Heine, Catherine E. Grant, Marshall W. Bautz, Beverly J. LaMarr, Eric D. Miller, Michael W. J. Hubbard, David Hall, Joan Requena, Emanuele Perinati, Steven W. Allen, Artem Poliszczuk, Dan Wilkins, Fabio Gastaldello, Silvano Molendi, Ralph P. Kraft, Gerrit Schellenberger, Arnab Sarkar

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 are trying to take a photograph of a faint, glowing cloud of hot gas floating in deep space. To do this, you need a camera so sensitive it can detect a single whisper of light. But there's a problem: your camera is sitting in a storm of invisible, high-speed particles zipping through the universe. When these cosmic particles hit your camera, they create a chaotic static noise that looks exactly like the faint signal you are trying to find. This is the challenge facing astronomers with the new "NewAthena" space telescope. They want to see the universe's hottest structures, but their view is constantly blurred by this "background noise" caused by cosmic rays. To fix this, scientists need to understand exactly how these particles interact with their camera's tiny, intricate parts. They can't just build the camera and test it in space; they have to build a perfect digital twin of the camera and run millions of virtual experiments to see where the noise comes from and how to stop it.

This paper describes how a team of scientists used a super-powerful computer to run these virtual experiments for the Wide Field Imager (WFI), a key camera on the NewAthena telescope. Think of the WFI as a high-tech digital eye made of millions of tiny sensors. The team used a software tool called Geant4 to simulate how cosmic rays crash into this digital eye. They didn't just run one simulation; they ran two different types of "virtual cameras." The first was a "Shell Model," a super-simplified version of the camera that looks like a few nested balls. This was like using a sketch to quickly test a new idea. The second was a "Detailed Mass Model," a hyper-realistic 3D blueprint containing about 1,500 different parts, including tiny bolts and screws. This was like building a full-scale, working prototype in a video game.

To make these simulations happen, the team needed a massive amount of computing power. They used a supercomputer called SuperCloud, which has hundreds of processors working together. Imagine a single computer taking three years to finish a job; by using 1,152 processors at once, they finished the same job in a single day. This allowed them to simulate billions of particle interactions, a number so huge it would be impossible to count by hand.

The results of these simulations were like a detective story. By looking at where the "noise" was created in their virtual camera, the scientists found something surprising. They discovered that tiny, seemingly unimportant parts, like the small metal bolts holding the camera together, were actually major sources of background noise. When cosmic rays hit these bolts, they created a specific type of glowing signal (fluorescence) that confused the camera. The simulations showed that if you removed a specific gold-colored part inside a "light trap" (a shield designed to block stray light), you could get rid of a strong gold signal. However, the simulation also revealed a tricky trade-off: removing that gold part actually made a nickel signal much worse because the gold part was accidentally shielding the nickel from other particles. This taught the team that you can't just remove parts to fix noise; you have to understand how every piece protects or exposes the others.

The team also used their simplified "Shell Model" to test the software itself. They noticed that when they upgraded their simulation software to a newer version, the background noise suddenly jumped up. By running quick tests on the simple model, they figured out that a new feature in the software called the "General Neutron Process" was causing this error. They found that this feature was a bug that shouldn't be used for space missions. Thanks to their simulations, they now know to turn this feature off to get accurate results.

In short, this paper shows that by using supercomputers to run both quick, rough sketches and incredibly detailed, slow-motion movies of cosmic rays hitting a camera, scientists can figure out exactly how to design better instruments. They proved that tiny details matter, that changing one part can have unexpected side effects, and that even the software used to simulate the universe needs to be checked for bugs. These findings will help ensure that when the NewAthena telescope launches in the late 2030s, it can see the faintest whispers of the universe without being drowned out by the static of cosmic rays.

Technical Summary: High-statistics simulations of NewAthena WFI background using Geant4

Problem Statement
The Wide Field Imager (WFI) on ESA's NewAthena X-ray observatory is designed to observe faint, diffuse hot gas structures. However, the measurement sensitivity of these sources is limited by instrumental background generated by cosmic ray particle interactions within the detector. To achieve science goals, it is critical to understand, characterize, and reduce this background. The challenge lies in the computational complexity required to simulate these interactions with sufficient fidelity and statistical power to resolve subtle design dependencies and specific background components, such as fluorescence lines.

Methodology
The authors employ a two-pronged simulation strategy using the Geant4 toolkit (version 11.4.0) to model particle interactions within the WFI instrument, leveraging High Performance Computing (HPC) resources on the MIT SuperCloud system.

Dual Mass Model Approach:
- Detailed Mass Model: A high-fidelity geometry derived from instrument CAD drawings, containing approximately 1,500 physical components (e.g., DEPFETs, bolts, molybdenum structures). This model is used for accurate, detailed simulations to uncover subtle sensitivities of background to specific design features. It requires high statistics (up to $10^9$ primary particles) to resolve low-yield events like fluorescence peaks.
- Shell Model: A simplified, spherically symmetric geometry consisting of a silicon slab with coatings, encased in concentric shells of Aluminum, Molybdenum, and Kapton. This model contains only six physical components. It is computationally inexpensive, allowing for rapid iteration, extraction of general trends, and preliminary "what-if" scenario testing.
HPC Implementation:
- Simulations are executed on Intel Xeon Platinum 8260 CPU nodes (48 cores, 192 GB RAM).
- The workflow utilizes a perfectly parallel approach where primary particles are divided into chunks processed by different CPU cores.
- The multi-threaded version of Geant4 is employed to optimize memory usage within nodes.
- Post-processing is performed in parallel on the same chunks before aggregation.
- This architecture allows for massive speedups: a "burst" allocation of 1,152 cores can complete in one day what would take a single core over three years.
Physics Configuration:
- Simulations utilize the QBBC_EMZ physics lists, enabling electromagnetic modeling, atomic relaxation, fluorescence, Auger processes, and particle-induced X-ray emission (PIXE).
- Primary particles are drawn from probability distributions for Galactic Cosmic Ray (GCR) protons, alpha particles, electrons, and the Cosmic X-ray Background (CXB).

Key Results
The study demonstrates the utility of combining high-statistics detailed simulations with rapid simplified modeling:

Background Composition: GCR protons are identified as the dominant contributor to the instrumental background. The total background spectrum is a sum of contributions from protons, alphas, electrons, and CXB, with distinct fluorescence lines visible.
Spatial Origins of Fluorescence: High-statistics simulations of the detailed mass model revealed that X-ray "hot spots" (creation locations of in-band X-rays) are concentrated near the detectors. Specifically, cylindrical regions corresponding to the exterior of bolts were identified as significant sources of Chromium and Iron fluorescence lines.
Design Trade-offs: A case study on the "light trap" component demonstrated complex trade-offs. Removing a specific gold component within the light trap successfully eliminated a strong gold fluorescence peak. However, the simulation also revealed that this removal significantly increased the Nickel fluorescence peak because the removed component had been partially shielding Nickel sources elsewhere in the trap. This highlights the necessity of high-fidelity simulations to predict unintended consequences of design changes.
Software Versioning: The simplified shell model was used to investigate a significant increase in simulated background observed when upgrading Geant4 from version 10.6.3 to 11.2.2. The investigation isolated the "General Neutron Process" as the cause. Consequently, the authors recommend disabling this process in space applications using Geant4 versions prior to 11.4, where it is disabled by default.

Significance and Claims
The paper claims that this simulation pipeline is essential for the NewAthena WFI mission to optimize instrument design and readout strategies. The authors emphasize that:

HPC is a Necessity: High-statistics, high-fidelity simulations are computationally expensive but necessary to resolve specific background sources (like minor component fluorescence) that lower-statistics runs cannot detect.
Complementary Utility: The combination of detailed and simplified models allows for both deep physical insight and rapid design iteration.
General Applicability: While focused on the WFI, the described pipeline and methodology are applicable to other X-ray missions.
Software Vigilance: The study underscores the importance of rigorously testing new Geant4 physics lists and versions, as updates can introduce significant, non-physical artifacts (such as the General Neutron Process issue) that affect background predictions.

The work is presented as an ongoing effort within the NewAthena WFI Background Working Group, providing a framework for understanding and reducing instrumental background through computationally intensive, high-fidelity modeling.

More like this