Only six of 79 controlled molecular collisions produced a chemical reaction — and all of them occurred within a narrow geometric window that researchers at the University of Graz have now mapped in real space at the single-molecule level. The work, published in Science by physicist Leonhard Grill and post-doctoral researcher Matthew Timm, defines what the team calls a "cone of reaction": the set of collision geometries — contact point, impact parameter and relative orientation — within which a coupling reaction can actually proceed.

The finding matters for anyone modelling or engineering surface reactions, because it quantifies just how unforgiving the geometry is. Reactions occurred mainly when a difluorocarbene (CF2) molecule approached a BTFyl radical target along the same copper row on a single-crystal surface to which the target was anchored. When the orientation of the two reactants deviated by more than 15°, reactions did not happen at all. The successful collisions correspond to a narrow range of impact parameters — the distance of the collision line from the reactants' centre of mass — a variable that has long resisted experimental control.

The experiment hinged on a scanning tunnelling microscope (STM) set-up that let the researchers launch CF2 molecules in straight lines along different atomic rows of the copper surface, effectively selecting the impact parameter. They also controlled the orientation of the BTFyl target by rotating it around its anchor point. This is the technical core of the result: the impact parameter, historically dubbed the "forbidden fruit" of reaction dynamics, became a tunable experimental variable rather than something averaged over in molecular beams.

"For the first time, we have been able to directly observe in real space and with single molecules which collision geometry must be met for a successful reaction," Grill explains. "Importantly, our experiment covers all geometric aspects of a reaction, namely, the precise location at which the reactants 'touch' each other upon collision, the 'impact parameter' (that is, how far the collision is from the reactants' centre of mass) and the orientation of the reactants."

For R&D managers in catalysis, surface science and molecular electronics, the immediate relevance is methodological. STM-based single-molecule manipulation on well-defined single crystals gives a way to interrogate reaction dynamics variables that ensemble experiments smear out. The caveat is scale: the Graz team worked with one reactant pair on one surface system, with 79 observed collisions — a small sample by any statistical standard. The behaviour may not transfer to other molecule–surface combinations, a limitation the researchers themselves are now testing by studying similar systems to assess whether the observed geometry rules generalize.

The technical difficulty was not trivial even for this system. "Controlling the impact parameter was for a long time considered to be the 'forbidden fruit' of reaction dynamics because of the technical difficulties to control it," Grill notes. "The main challenge for us was how to do such experiments with relatively large molecules, which can have many adsorption orientations on the surface."

The conceptual payoff is a sharpened picture of what a productive collision looks like. "The new picture we have gleaned concerns not only the impact parameter, so where one reactant has to collide with another, but also the orientation of the involved compounds," Grill says. "If the reactants become larger and also have various side groups, understanding the cone of reaction might become even more important to understand, control and predict reaction rates."

In a related Science article, Jonas Björk of Linköping University in Sweden frames the forward agenda. Further measurements, he suggests, could reveal how molecular structure, reactive sites and environment shape reaction pathways, enabling more predictive control of surface chemical transformations. "They could also enable single-molecule studies of how chirality influences chemical reactivity by directly controlling the handedness and orientation of reactants," he adds.

That points toward the practical stakes: if the cone-of-reaction framework generalizes, it could feed into rational design of on-surface synthesis routes, where orientation-dependent coupling yields currently force researchers to rely on empirical screening. For now, the Graz results are a measured proof of principle on one system, with the generalization studies already under way.