Dr A.D. Dinga Wonanke

Chemistry and materials problems that need computational modelling, scientific AI or research software. Typical examples are explaining why a material behaves as it does, screening candidates before synthesis, predicting plausible synthesis conditions, and turning a research workflow into reliable, documented software. The consulting services describe each area.

With a short conversation about the scientific problem, your data and your constraints. I then assess which computational, AI or software methods fit, and we agree on scope, deliverables, milestones and timeline before any work begins. An engagement can be a focused technical review, a research and development project, or a software and AI implementation.

Yes. Confidentiality, data handling, intellectual property and deliverables are agreed before work begins. Please keep confidential details out of a first enquiry; a brief overview is enough to start.

No. My independent consulting is separate from my academic appointment at Nottingham Trent University and from my role as founder of StructSub Ltd. It does not imply sponsorship or endorsement by either.

Bring an experimental challenge and we turn it into a testable computational question. That might mean proposing synthesis conditions, checking structures against powder X-ray diffraction (PXRD), predicting properties or explaining an unexpected result.

StructSub Ltd is the company I founded to build a browser-based scientific workspace for materials design, simulation and analysis. It brings structure building, modification, simulation and analysis into one place, so researchers spend less time writing scripts and moving between disconnected tools. Visit StructSub.

It grew from a recurring frustration in my own research: hours spent writing one-off scripts and moving structures between tools before any science could happen. Researchers should be able to modify molecules and materials without that overhead. Read the full story.

Many promising materials never reach practical use because the knowledge of how to make them is fragmented, poorly documented or hard to reproduce. I work to make synthesis knowledge explicit, computable and transferable, so researchers can reproduce materials with fewer trial-and-error experiments and a clearer path to scale-up.

Synthesis prediction uses AI and, where relevant, physics-based chemistry to estimate how to make a target material: reagents, solvents, temperature, reaction time, pH, modulators and processing steps. My focus is structure-to-synthesis, which starts from a 3D structure, such as a CIF file, and predicts plausible conditions for making it.

They are probabilistic, not definitive, and they do not replace chemical intuition. Their value is in narrowing the experimental search: suggesting plausible conditions and flagging constraints or likely failure modes before time and materials are spent in the lab.

Less trial and error means shorter development timelines and lower costs when a new material moves towards industrial production and scale-up.

For metal–organic frameworks, yes. My open-source package fairmofsyncondition generates a synthesis-condition report from a CIF file. Install it with pip install fairmofsyncondition, then run fairmofsyncondition_syncon my_mof.cif -o my_mof_report.txt.

Mainly porous materials, including metal–organic frameworks (MOFs), covalent organic frameworks (COFs) and molecular cages, for gas storage, CO₂ capture, separations, chemical sensors and battery applications.

Density functional theory (DFT) for stability, geometry, reactivity and electronic structure, including band structures and projected density of states. Binding energy calculations to assess how strongly guest molecules interact with a material, for example in chemical sensors. Energy decomposition analysis (EDA) to understand the interactions within a system by separating them into physical contributions, such as electrostatics, dispersion and charge transfer. Atoms-in-Molecules (AIM) analysis for charge and bonding, for example to locate redox-active sites. Grand canonical Monte Carlo (GCMC) for gas adsorption, such as CO₂ and hydrogen uptake in porous frameworks. Molecular dynamics for ion migration and diffusivity in battery materials. Graph neural networks that learn structure–property relationships across large materials databases.

Yes. My tools include mofstructure for analysing porous frameworks, cage_isomer_builder for molecular cages, and the FAIR-MOFs dataset. I build them with tests, documentation and reproducible workflows. See all tools.