The paper in Nature Communications is here: go.nature.com/2vgoElD
In 1965, Gordon Moore — co-founder of Intel Corp. — predicted that the size of transistors on a microprocessor chip will shrink by a factor of two every year (he revised it to every two years in 1975). This prediction — which came to be known as "Moore’s law" — has proven to be remarkably accurate for the past 5 decades! To appreciate the effect transistor scaling has had on the working of microprocessor chips, consider this: As compared to Intel’s first generation microprocessor chip, 1971 4004, the fifth-generation Core i5 processor launched in 2015 offers 3,500 times more performance, is 90,000 times more energy efficient at about 60,000 times lower cost. (The transistor’s feature size of the fifth-generation is about 700 times smaller than that of the first generation: 14-nanometres and 10-micrometres feature sizes, respectively. As of May, 2017, IBM claimed to have built the world’s first transistor of feature size 5-nanometres.)
But transistor scaling based on Moore's law is unlikely to go-on much longer. By early 2020s, when the transistor's feature size is expected to be just 2-3-nanometres, electron behaviour will be governed by quantum uncertainties (at that scale) that will make transistors hopelessly unreliable. The demise of Moore's law will render our use of silicon-based information technology (IT) unsustainable; by most projections, more than half of the world’s energy will be consumed by information technologies within a couple of decades. Yes, silicon should be ditched in favour of new materials. What is (will be) the new material to replace silicon? Carbon nanotubes? Multiferroics? Quantum computing? Spintronics? Topological insulators?
Or, could it be molecular spintronics? Molecular spintronics seeks to exploit the intrinsic quantum mechanical nature of molecules — spin, for example — and to manipulate those properties in a more energy-efficient way; and, of course, to build conventional devices and beyond, from single or few molecules. Spin-crossover molecules (SCMs) are a case in point. SCMs can be reversibly switched between two different spin states (or magnetic states) — called high-spin and low-spin states — by using external stimuli like temperature, light, pressure, and electric field. In addition, spin-crossover molecules exhibit different electrical conductance in the two spin states: high-conductance in the high-spin state and low-conductance in the low-spin state. This makes SCMs potentially suitable for application as both active and passive components in IT.
However, the molecules’ sensitivity to external stimuli — more often than not — has proven to be a double-edged sword. SCMs are rather fragile, and when they come in contact with solid surfaces — which is a prerequisite for any single-molecule-based device realization — the spins are quenched in either one or both of the states (coexist), or only a small percentage of the molecules retain their bistability. The reasons may vary; for one, even an inert surface like gold could cause fragmentation of the molecules, while molecular distortion from the usual octahedral symmetry — responsible for the bistability — due to interaction with surfaces could also result in the loss of bistability. In fact, the coexistence of the spin states on surfaces is so common that it came to be regarded as the true thermodynamic phase.
In this work — which is a culmination of years of research on trying to gain control over the spin-crossover property on surfaces, mainly between our research group at Freie Universität Berlin and our collaborators at Christian-Albrecht-Universität zu Kiel — we showed the complete and reversible switching between the two spin states of spin-crossover molecules on a highly oriented graphite surface, with coverages ranging from submonolayers to multilayers. These results, along with the earlier works by Bernien et. al., debunked the more prominently held view of the spin-state coexistence as being an intrinsic property of spin-crossover molecules when in contact with solid surfaces, but rather a property dependent upon molecule-substrate combinations. It is also interesting to note that SCMs in reduced dimensions retain much of their behavior from the bulk, like exhibiting cooperative effects in the temperature induced, and non-cooperative effects in the light-induced spin transitions. These findings are expected to give a renewed impetus to research in this field.