How to unlock more value from neutron science with software
Neutron diffraction is one of the most powerful tools in the world for studying the structure of materials, and is used by scientists from chemistry, physics, biology and materials science. By firing a beam of neutrons at just a tiny sample of material – which could include anything from proteins and DNA, to lava and cement – the minute details of the atomic structure are revealed by analysing the intensities and dispersion angles of the neutron beams scattered by the sample.
While such experiments may sound simple enough to conduct, the resulting data is vast and extremely complex to unravel. The potential insights gained about the material can transform our understanding of processes such as the melting of DNA fibres or how lava flows under the sea, but the journey from raw data to extraordinary insights can be a challenge for neutron users.
There are over 40 instruments at Institut Laue-Langevin (ILL), the world’s most powerful neutron source, with more than a dozen tailored to use beams from the reactor in diffraction experiments. Each of these instruments is unique in the world, with generations of scientists having developed and refined the apparatus to improve our methods for studying the structure of materials. Often, the rich and expansive data from the experiments requires extensive knowledge about the technique and computing power to extract information. The knowledge can be limited to a select number of people, including the scientists responsible for each instrument, and the computing software needed to unravel the data needs to have kept pace with the technological changes we are seeing across data collection methods in science. New software capabilities at ILL are transforming how data is processed for these complex experiments.
Transforming post-experiment data
Processing diffraction data post-experiment requires complex numerical calculations to be coded by scientists into computer programs, relying on existing knowledge. For a long time, the software consisted of a series of independent pieces of code that required a high expertise for its correct use. Consequently, most users have depended on instrument scientists to analyse the data for them. Even for the scientists that did not develop the software, this was almost like a black box. This can make it hard to make the most of the capabilities of an instrument over time.
Fortunately, this situation is changing thanks to the extraordinary growth of open source software. The movement has produced a range of useful tools that can be easily adapted to the needs of the instrument software, improving performance and creating new utilities. In the diffraction field, where data visualisation is key to data analysis, this is an invaluable aid not only because it allows the specialist a more accurate and efficient analysis of the data, but also because it permits the creation of friendly graphical user interfaces (GUIs) that make calculations accessible to non-specialists, whether from academia, industry or elsewhere.
A postdoctoral position granted by the FILL2030 project has supported updates to an array of instrument programs: rewritten in a modern way, extended with new capabilities and integrated in a single application called Int3D. In addition, the use of open source software like Qt and VTK has made it possible to create user-friendly interfaces and incorporate visualisation tools in both 2D and 3D, as well as tools for direct interaction with the data.
For example, how we handle data from the D19 single crystal diffractometer at the ILL has changed substantially over the years. D19 is used specifically to determine the structure of small molecules inside a crystal matrix, as well as to study biological molecules such as haemoglobin and cellulose, industrial polymers and minerals. The original software used with this instrument was extremely important for science, but difficult to handle – now we can integrate all the required tasks for data processing and powerful visualisation and interaction tools, allowing for a better user interface.
The work of this programme will not only help to greatly relieve the pressure on the ILL’s instrument scientists, but can be taken to other neutron facilities too. It will help to save substantial time and effort both for the ILL scientists and the visiting researchers, but will also improve the quality of results – users can now almost immediately visualise their measurements in reciprocal space in both 2D and 3D after inputting raw data, and analyze them in a variety of ways. Interacting with data like this will enhance the insights that can be gained from the experiment. This unlocks the full potential of the neutron diffraction, and may result in better and more comprehensive understanding of materials at a greater speed.
One of many technological advances for big science
It is crucial for scientific progress that big data from neutron science is fully analysed and exploited, with advances in computer-aided resources helping to both interpret raw data and connect the vast network of previous experiments. ILL has been at the forefront of new data management processes for science, through the unique application of the DOI for data (Digital Object Identifier), and as scientific instruments continue to be upgraded, computer programmes developed via the FILL2030 project will help to plug gaps between the instrument and the value of the resulting data, and help gain most from existing instruments and techniques.
In light of the COVID-19 pandemic, we are seeing new challenges and opportunities for technology in big science settings, as many scientists are forced to work away from facilities. Innovations such as the Virtual Infrastructure for Scientific Analysis (VISA) are helping to unlock direct access to experimental data at ILL from a computer at home, with no need to install specialist software. Ensuring improved user interfaces will remain key to unlocking scientific discovery throughout unprecedented events, to ensure that experimental data is being utilised as much as possible.
