Various energy sources can be used to induce chemical reactions — often heat, but also electricity, light and even ultrasound. But there is another option that has been much less explored: mechanical energy from, for example, grinding reactants together or applying pressure to solids. Mechanochemical reactions have been intensively researched in the past few years, but remain poorly understood at the atomic level. In a paper in Nature, Yan et al.1 provide some much-needed insight into these processes.
Mechanochemistry has a long history and is currently undergoing a renaissance2. The ancient Greek philosopher Theophrastus described how cinnabar (mercury sulfide) can be converted into elemental mercury by grinding it in a copper mortar and pestle with vinegar. In the nineteenth century, Michael Faraday studied mechanochemical reactions3, and the chemist M. Carey Lea is often credited with demonstrating that mechanically induced reactions can yield different products from thermally induced ones4. Polymer chains have also been observed to snap under mechanical stress, and mechanochemical reactions can be expected wherever materials are degraded by mechanical forces5.
Although mechanochemistry was neglected for many years as a viable method for manufacturing chemicals, in the past decade it has attracted much interest in this regard. This is partly because mechanochemical reactions normally need little or even no solvent, and can therefore save on waste and cost compared with conventional, solvent-based synthetic methods6. But despite its growing importance, mechanochemistry remains something of an enigma — exactly how is mechanical energy translated into chemical reactivity?
Yan et al. shed some light on this intriguing question. Their work centres on the effect of mechanical force on molecules that consist of carbon-based organic parts connected to an inorganic core of copper and sulfur atoms (Fig. 1). The copper–sulfur core is mechanically ‘soft’ and susceptible to rupture when a mechanical force is applied; such mechanically sensitive regions of molecules are generally known as mechanophores. By contrast, the organic parts are mechanically ‘hard’.
The authors applied pressure of up to 12 gigapascals to crystals of these molecules and observed a reaction in which nanometre-scale particles of metallic copper were produced, as well as organic by-products that contain sulfur–sulfur bonds. In this mechanochemical reaction, electrons are transferred from sulfur to copper, chemically reducing the copper from copper(i) to metallic copper(0). The outcome of the mechanochemical reaction was different from that of heating, which instead produced copper(i) sulfide (Cu2S) — that is, the mechanochemical and thermochemical reactions broke and formed different chemical bonds. This brings us back to the question of precisely how the mechanochemical process occurs at the atomic level.
Yan and colleagues used computational modelling to show that the application of pressure to the crystals results in ‘non-isotropic’ compression of the molecules — the mechanically hard organic groups squash the soft copper–sulfur core, dramatically distorting the bond angles between the copper and sulfur atoms. The modelling also showed that, simultaneously with the bond-angle changes, the copper–sulfur bonds become weaker and electron density transfers from the sulfur to the copper atoms. These effects neatly explain the mechanochemical process that leads to the formation of copper-metal particles.
The authors went on to study a variant of their molecule that also contains hard units attached to a soft core but that does not undergo the mechanochemical reaction. Their modelling suggested that the difference in reactivity between the two molecules is due to subtle differences in their structures and crystal packing — for the mechanochemical reaction to occur, the hard organic groups need to have sufficient wiggle room in the crystal to exert force on the copper–sulfur core.
The work is interesting in several further regards. Although a range of mechanophores is known, they are mainly based on organic chemistry and have been explored in the context of polymers. In particular, the rupture of polymer chains caused by stretching can be controlled so that it occurs at specific mechanophore sites5. By contrast, Yan and colleagues’ mechanophores are inorganic units that undergo redox processes, which occur in response to compression rather than stretching. Moreover, the authors have successfully applied the mechanophore concept to crystals of small molecules, rather than to long-chain polymers. Compressive mechanochemistry such as this might offer a general way to produce nanoparticles.
However, mechanochemical reactions are extremely diverse, and the mechanism put forward in this work helps to explain only a relatively small subset of them. Nevertheless, atomic-level insight into how mechanochemical reactions can occur is invaluable, and Yan and colleagues’ work will no doubt stimulate further efforts to reach a similar level of understanding for other reactions, such as those relevant to the synthesis of commercial materials and pharmaceuticals. The authors’ use of modelling is particularly effective, and gets around the practical difficulties of directly observing mechanochemical reactions at atomic resolution using microscopy.
Finally, the work points to a future in which an informed understanding of mechanochemistry will improve the prediction and design of reactions. This would provide a more rational basis for the field and facilitate the application of mechanochemical reactions as a mainstream synthetic method.
Nature 554, 468-469 (2018)