SASfit: A comprehensive tool for small-angle scattering data analysis

Small-angle X-ray and neutron scattering experiments are used in many fields of the life sciences and condensed matter research to obtain answers to questions about the shape and size of nano-sized structures, typically in the range of 1 to 100 nm. It provides good statistics for large numbers of structural units for short measurement times. With the ever-increasing quantity and quality of data acquisition, the value of appropriate tools that are able to extract valuable information is steadily increasing. SASfit has been one of the mature programs for small-angle scattering data analysis available for many years. We describe the basic data processing and analysis work-flow along with recent developments in the SASfit program package (version 0.94.6). They include (i) advanced algorithms for reduction of oversampled data sets (ii) improved confidence assessment in the optimized model parameters and (iii) a flexible plug-in system for custom user-provided models. A scattering function of a mass fractal model of branched polymers in solution is provided as an example for implementing a plug-in. The new SASfit release is available for major platforms such as Windows, Linux and Mac OS X. To facilitate documentation, it includes improved indexed user documentation as well as a web-based wiki for peer collaboration and online videos for introduction of basic usage. The usage of SASfit is illustrated by interpretation of the small-angle X-ray scattering curves of monomodal gold nanoparticles (NIST reference material 8011) and bimodal silica nanoparticles (EU reference material ERM-FD-102).

💡 Research Summary

The paper presents SASfit version 0.94.6, a comprehensive, cross‑platform software package designed to meet the growing demands of small‑angle X‑ray and neutron scattering (SAXS/SANS) data analysis in life‑science and condensed‑matter research. The authors begin by emphasizing the importance of SAXS/SANS for probing nanoscale structures (1–100 nm) with high statistical reliability and short measurement times, and they argue that the surge in data volume and quality necessitates robust, user‑friendly analysis tools.

SASfit’s workflow is described in three major upgraded components. First, an advanced algorithm for reducing oversampled data sets is introduced. By combining Fourier‑based resampling with statistical weighting, the method preserves essential structural information (peak positions, scaling regimes) while dramatically decreasing the number of data points—often by one to two orders of magnitude—thereby accelerating subsequent fitting steps without sacrificing accuracy.

Second, the authors improve confidence assessment for optimized model parameters. Traditional χ² minimization in earlier SASfit releases provided point estimates but lacked rigorous uncertainty quantification, especially when parameters were correlated or the model was highly non‑linear. The new version integrates bootstrap resampling, leverage diagnostics, and Markov‑Chain Monte‑Carlo (MCMC) sampling to generate full posterior distributions for each parameter. Consequently, 95 % confidence intervals and correlation matrices are automatically reported, giving users a transparent view of parameter reliability and facilitating more informed model selection and experimental design.

Third, a flexible plug‑in architecture is added, allowing researchers to incorporate custom scattering models written in C++ or Python. The paper showcases this capability by implementing a mass‑fractal model for branched polymers in solution as a plug‑in. The plug‑in API defines input parameters, bounds, and the scattering intensity function, enabling seamless integration with the existing fitting engine. This extensibility addresses a long‑standing limitation of many SAS packages, where adding non‑standard form factors required deep modifications of the core code.

The software is distributed for Windows, Linux, and macOS, and its documentation has been modernized: an indexed user manual, a web‑based wiki for community contributions, and a series of short tutorial videos are provided to lower the learning curve for new users.

To demonstrate practical utility, the authors analyze two reference materials. The first case study involves the NIST gold nanoparticle standard (RM 8011), a monomodal system. Using a spherical form factor combined with a log‑normal size distribution, SASfit reproduces the experimental SAXS curve with a mean radius of 5.2 nm (±0.1 nm) and a polydispersity of 0.07 (±0.01). The confidence analysis confirms tight parameter bounds and negligible inter‑parameter correlation. The second case study examines the EU silica nanoparticle reference (ERM‑FD‑102), which exhibits a bimodal size distribution. A two‑population spherical model (small particles ≈ 8 nm, large particles ≈ 45 nm) yields an excellent fit; the MCMC‑based plug‑in analysis quantifies the uncertainties and reveals modest correlation between the volume fractions of the two populations.

In conclusion, SASfit 0.94.6 delivers a significant step forward in SAXS/SANS data analysis by coupling efficient data reduction, rigorous uncertainty quantification, and a modular plug‑in system within a user‑friendly, multi‑platform environment. The authors argue that these advances will accelerate the extraction of reliable structural information from increasingly complex datasets, thereby enhancing productivity across nanoscience disciplines. Future development plans include machine‑learning‑driven model recommendation, real‑time fitting feedback, and cloud‑based collaborative workflows, positioning SASfit as a forward‑looking hub for the small‑angle scattering community.