⚛️ nuclear experiments

Modernization in processing and dissemination of experimental cross section data in EXFOR

This paper introduces X4VIEW and a preliminary version of X4TOCX, tools designed to convert the legacy, card-based EXFOR format into modern HTML and CX4 formats to facilitate easier access, tabulation, and visualization of experimental cross-section data for users utilizing contemporary programming languages like Python.

Original authors: Naohiko Otuka, Vidya Devi

Published 2026-09-16

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Naohiko Otuka, Vidya Devi

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

In the world of modern medicine, doctors rely on tiny, unstable atoms called radionuclides to see inside the human body and treat diseases. These atoms are not found in nature in useful quantities; they must be created in laboratories by firing beams of charged particles, like protons, at specific target materials. To make these medical isotopes efficiently and safely, scientists need to know exactly how likely a particle is to hit a target atom and create the desired result. This likelihood is called a cross section. It acts like a map, showing how the probability of creating a specific atom changes as the energy of the incoming beam changes. Without this map, researchers cannot predict how much medicine they will produce or how much unwanted radioactive waste they might create. For decades, the global community of scientists has stored these experimental maps in a massive digital library called EXFOR. However, this library was built for a different era of computing, designed to be read by machines using a format that looks like a list of numbers on a grid, much like the old punch cards used in early computers. While experts who built the library can navigate it, modern scientists who want to use this data for medical planning often find the format difficult to read and impossible to use with the flexible software tools they prefer today.

A team of researchers at the International Atomic Energy Agency has developed a new way to bridge this gap, turning the rigid, old-fashioned data files into formats that are easy for both humans and modern computers to understand. The scientists, Naohiko Otuka and Vidya Devi, recognized that while the original data format is excellent for storing information between data centers, it creates a barrier for the people who actually need to use that information to design medical treatments. To solve this, they created two new digital tools that act as translators. The first tool, called X4VIEW, takes the cryptic data files and converts them into clear, readable web pages. Imagine opening a document where every number is accompanied by a plain-language explanation of what it means, rather than a string of codes that requires a manual to decipher. This allows researchers to instantly see the details of an experiment, such as the specific type of radiation emitted or the exact conditions under which the data was collected, without needing to be an expert in the old file format.

The second tool, known as X4TOCX, goes a step further by organizing the data into neat, simple tables that are ready for analysis and graphing. In the original files, data for different types of atoms produced in the same experiment can be tangled together in a single block of text, making it hard to isolate just the information needed for a specific medical isotope. The new tool untangles these threads, separating the data into clean columns of energy and probability. It even performs complex conversions automatically. For instance, if an experiment measured the energy of particles in a way that is difficult to compare with other studies, the tool can mathematically adjust those numbers into a standard format that matches what other scientists use. This means a researcher can take data from a decades-old experiment and immediately compare it with new results, or combine them to create a more accurate picture of how to produce a life-saving drug.

The researchers tested these tools using real-world examples, such as the production of isotopes used for brain imaging and cancer treatment. They showed how the new system could take a raw data file, identify the specific reaction needed, and present the numbers in a way that makes it easy to spot errors or inconsistencies. In one case, they demonstrated how the tool could help update old measurements to match the latest scientific standards, a task that previously required tedious manual copying and pasting. While the tool for creating the simple tables is still being refined to handle every possible type of data, the system for creating readable web pages is already available for anyone to use. By lowering the technical barrier to entry, these tools ensure that the vast history of nuclear experiments is not locked away in an old format, but is instead accessible to the modern scientists who need it to improve medical care. The work represents a quiet but significant shift in how scientific data is shared, moving from a system designed for storage to one designed for understanding.

Technical Summary: Modernization in Processing and Dissemination of Experimental Cross Section Data in EXFOR

Problem Statement
The EXFOR library serves as the primary database for experimental nuclear reaction data, including radionuclide production cross sections essential for medical isotope optimization. However, the EXFOR format, originally designed in the 1970s for 80-column punched cards and Fortran processing, presents significant barriers for modern users and software engineers. While the format remains efficient for compilers and data exchange between centers, it is difficult for users to parse using modern programming languages (e.g., Python) without developing complex custom parsers.

This rigidity creates specific challenges in the evaluation of radionuclide production cross sections:

Readability: Critical information, such as decay data and monitor reaction cross sections, is often encoded in cryptic formats (e.g., specific numeric codes under DECAY-DATA), making it difficult for evaluators to identify reference values needed for renormalization without consulting the EXFOR Formats Manual.
Data Manipulation: Extracting numerical data for mathematical operations (e.g., converting isotopic to elemental cross sections, applying renormalization factors based on updated monitor values) typically requires manual copying and pasting into spreadsheets, which is inefficient and prone to error.
Visualization Limitations: Existing web retrieval systems often display data in ways that obscure discrepancies between datasets (e.g., mixing elemental and isotopic cross sections) or fail to provide user-friendly table formats suitable for plotting.

Methodology
To address these issues, the authors developed a modern processing pipeline based on the J4 format, a JSON representation of EXFOR data. This pipeline utilizes the ForEXy code package to convert raw EXFOR files into two distinct, more accessible formats:

X4VIEW (Human-Readable Conversion):
- Pipeline: X4TOJ4 (EXFOR to JSON) $\rightarrow$ POIPOI (removes EXFOR "pointers" to isolate specific data subsets) $\rightarrow$ J4VIEW (JSON to HTML).
- Function: Converts EXFOR entries into self-explanatory HTML files. It expands cryptic keywords (e.g., decoding DECAY-DATA into readable gamma energies and intensities) and preserves all original information.
- Features: Supports all EXFOR keywords, allows CSS customization, and can hyperlink references to DOIs.
X4TOCX (Tabulation and Plotting Conversion):
- Pipeline: X4TOJ4 $\rightarrow$ POIPOI $\rightarrow$ J4TOC6 (JSON to C6 format) $\rightarrow$ C6TOCX (C6 to CX4 format).
- Function: Converts EXFOR data into CX4 (Compact EXFOR) files.
- Output Structure: Generates a four-column table format: Incident Energy ( $E$ ), Energy Uncertainty ( $\Delta E$ ), Cross Section ( $\sigma$ ), and Cross Section Uncertainty ( $\Delta \sigma$ ).
- Capabilities: Handles complex data structures by resolving pointers (e.g., isolating specific production channels like $^{65}$ Zn from a multi-product entry) and performing necessary unit conversions (e.g., converting Astrophysical S-factors and center-of-mass energies to laboratory cross sections).

Key Contributions and Results

X4VIEW Implementation: Released in May 2026 as part of the ForEXy package (available on PyPI). The tool successfully transforms raw EXFOR entries into readable HTML. The authors report that this format has improved the proofreading process for EXFOR compilers and authors, as the HTML output makes the data self-explanatory, reducing ambiguity in author reviews.
X4TOCX Development: A preliminary version of the tool was developed to generate CX4 files. The authors demonstrated its utility by converting complex datasets, such as:
- Multi-channel production data (e.g., $nat $Cu(p,x)$ ^{62,63,65}$Zn) into isolated, plot-ready tables for specific isotopes.
- Astrophysical S-factor data (e.g., $^{17}$ O(p, $\gamma$ ) $^{18}$ F) into standard cross-section tables, automatically handling the conversion from center-of-mass energy to laboratory energy.
Performance Metrics: Resource estimation for converting the entire EXFOR library (version 2026-08-20) on an Intel Core i7-9700 system showed:
- EXFOR to J4: ~6000 seconds, ~1 GB peak memory.
- J4 to C6: ~8000 seconds, ~450 MB peak memory.
- C6 to CX4: ~1000 seconds, ~25 MB peak memory.
- The authors note that approximately 20 entries failed to convert due to compilation errors (e.g., integrated cross sections with angular/energy dependencies not supported in CX4), which data centers are currently addressing.

Significance and Claims
The paper claims that these tools lower the barrier for modern software engineers and nuclear data evaluators to access and process EXFOR data. By moving away from the rigid 80-column format toward JSON and structured text representations, the authors aim to:

Enhance the accuracy and efficiency of data extraction for cross-section evaluations, particularly for renormalization tasks.
Facilitate the creation of machine-readable inputs for fitting tools and covariance matrix generation (via the MAKCOV module).
Improve the quality of EXFOR entries through better author interaction during the proofreading phase.

The authors maintain a modest stance regarding X4TOCX, noting it is still under development and currently supports basic quantities. They plan to extend the module to support additional formats (e.g., CSV) and develop a dedicated plotting module to visually verify the correctness of the conversion process. The work represents a step toward modernizing the nuclear data infrastructure without abandoning the established EXFOR exchange standard.

More like this