Can theoretical chemists predict the outcome of a photochemical reaction?

Outcomes of a challenge, co-organised by the School of Chemistry's Professor Basile Curchod, to predict the future of a gas-phase molecule have been featured in an article in the magazine Chemistry World.
In the challenge, organised by Professor Curchod alongside scientists at Stanford University's SLAC National Accelerator Laboratory (SLAC), the community of theoretical photochemists was challenged to predict the photochemistry of a gas-phase molecule, cyclobutanone - before the experiment in question was conducted.
The idea for this challenge arose following a workshop where attendees discussed whether it was possible for computational photochemistry to be predictive and the results of a state-of-the-art experiment to be simulated before it is carried out.
The only information provided ahead of the challenge was that the molecule would be photoexcited at 200 nm in the gas phase, and the resulting ultrafast dynamics would be tracked with time-resolved MeV-UED (mega-electronvolt ultrafast electron diffraction) in an experiment. To be considered, each theoretical prediction had to be submitted to the Journal of Chemical Physics before January 2024, the date when the experiment would be conducted. This gave six months to the theoretical teams to come up with a prediction.
This represented a significant challenge, yet an impressive total of 15 articles were eventually submitted, with more than 70 researchers contributing overall. Two time-resolved MeV-UED experiments were published as a result in early 2025, including one from SLAC in collaboration with the School of Chemistry's Professor Andrew Orr-Ewing
All contributors met in April 2025 to critically assess the results of the prediction challenge. Some participating groups managed to predict the full experimental signal with accuracy, but the most important outcome of this challenge is that it acted as a calibration exercise for the computational photochemistry community. Research groups could understand which strategies work best and develop best practices for the field, leading to a transformative improvement in the quality of simulations in computational photochemistry.
All articles relating to the challenge were published as part of a special issue of the Journal of Chemical Physics, and the final conclusions were recently summarised in a Perspective article​.