Chemical reaction networks that transform out-of-equilibrium 'fuel' to 'waste' are the engines that power the biomolecular machinery of the cell. Inspired by such systems, autonomous artificial molecular machinery is being developed that functions by catalysing the decomposition of chemical fuels, exploiting kinetic asymmetry to harness energy released from the fuel-to-waste reaction to drive non-equilibrium structures and dynamics. Different aspects of chemical fuels profoundly influence their ability to power molecular machines. Here we consider the structure and properties of the fuels biology has evolved and compare their features to those of the rudimentary synthetic chemical fuels that have been used to date to drive autonomous nonequilibrium molecular-level dynamics. We identify desirable, but context-specific, traits for chemical fuels together with challenges and opportunities for the design and invention of new chemical fuels to power synthetic molecular machinery and other nanoscale processes. MainFuels are consumed to provide the energy that devices and processes require to perform useful work. [1][2][3] The free energy available from a chemical reaction can be harnessed by molecular machines and dissipated to offset work performed, thus preserving the Second Law of Thermodynamics when tasks are carried out through stochastic molecular-level dynamics. 4,5 In this way chemical engines (Fig. 1) 6 transduce energy from chemical fuels and have the potential to power synthetic molecular nanotechnology 7-13 by driving and sustaining processes out-of-equilibrium. [7][8][9]14 While some synthetic molecular machines use light, 15,16 electrochemistry 17 or transmembrane gradients, 18 chemical fuels provide an attractive alternative energy source (Fig. 2). 19 Although the waste generated in chemical fuel-to-waste reactions must be dealt with (either recycled, as happens with ADP in the cell, or removed, as occurs for water and CO2 with aerobic respiration), chemical fuels are unencumbered by many of the issues faced by powering processes with other forms of energy. Photo-and electrochemistry often produce unstable intermediates, reactive radicals or cause photobleaching, all of which can limit the number of cycles that complex molecules survive for. Furthermore, in contrast to photons, chemical fuels have the potential to be stored and transported, enabling flexible and responsive systems that operate autonomously when and where they are needed. 19,20 Indeed, active organisms that require a high-density energy supply (such as animals) tend to rely exclusively on chemical energy sources, whereas photosynthesising organisms cannot harvest enough energy from sunlight to support powerintensive behaviours, such as rapid movement. The same trend is apparent in technology; lightpowered flight, for example, remains a major engineering challenge. For these and other reasons, while the energy input to biology largely originates from light (via photosynthesis), chemical fuels (such as adenosine triphosphate (ATP), Fig. 2A)...

Biology operates through autonomous chemically fuelled molecular machinery, 1 including rotary motors such as ATP synthase 2 and the bacterial flagellar motor. 3 Chemists have long sought to create analogous molecular structures with chemically powered, directionally rotating, components. 4-17 However, synthetic motor-molecules capable of autonomous 360 directional rotation about a single bond have proved elusive, previous designs lacking either autonomous fuelling 7,10,12 or directionality. 6 Here we show that 1-phenylpyrrole 2,2′-dicarboxylic acid 18,19 (1a) is a catalysis-driven 20,21 motor that can continuously transduce energy from a chemical fuel 9,20-27 to induce repetitive 360 directional rotation of the two aromatic rings around the covalent N-C bond that connects them. Upon treatment of 1a with a carbodiimide, 21,25-27 intramolecular anhydride formation between the rings, and the anhydride's hydrolysis, both occur incessantly. Both reactions are kinetically gated 28-30 causing directional bias. Accordingly, catalysis of carbodiimide hydration by the motor-molecule continuously drives net directional rotation around the N-C bond. The directionality is determined by the handedness of both an additive 31 that accelerates anhydride hydrolysis and that of the fuel, and is easily reversed. >97% of fuel molecules are consumed through the chemical engine cycle 24 with a directional bias of up to 71:29 with a chiralitymatched fuel and additive. In other words, the motor makes a 'mistake' in direction every 3-4 turns. The 26-atom motor-molecule's simplicity augurs well for its structural optimisation and the development of derivatives that can be interfaced with other components for the performance of work and tasks. 32-36In 1999, Feringa reported 13 the first synthetic molecular rotary motor, based on light-driven isomerisation around a double bond. 14 Contemporaneously, 5 Kelly explored a chemically fuelled triptycene derivative, 15 conceptually related to Feynman's ratchet, 16,17 that directionally rotated 120 about a single bond. Our group have described catenanes that undergo pulsed 11 or autonomous 9 chemically fuelled rotation of one macrocycle around another and have recently used 21 carbodiimide chemical fuels to drive a rotaxane-based information 37-41 ratchet. A feature of the latter system's fuelling cycle (Supplementary Information, Section S2.5) is the formation of a transient activated ester. 23 Similar activated ester and anhydride intermediates have been exploited in dissipative assembly [25][26][27] and Hartley has shown that tethering diphenic acids as anhydrides produces a dynamic change in the angle between the biphenyl rings. 42,43 As both carbodiimide-fuelled formation of an anhydride and its subsequent hydrolysis can occur under the same set of conditions, 25,26,42,43 we reasoned that asymmetric catalysis of the formation and/or hydrolysis of the anhydride could potentially directionally bias the rotation of rings joined by

Biological systems exhibit a range of complex functions at the micro-and nanoscale under non-equilibrium conditions (e.g. transportation and motility, temporal control, information processing, etc.). Synthetic chemists also use out-of-equilibrium systems, for example in kinetic selection during catalysis, self-replication, dissipative self-assembly, synthetic molecular machines, and in the form of chemical oscillators. Key to non-equilibrium behavior are the mechanisms through which systems are able to extract energy from the fuel. In this Perspective we consider different examples using a common conceptual framework. We discuss how reaction cycles can be coupled to other dynamic processes through positive (acceleration) or negative (inhibition) catalysis to provide the thermodynamic impetus for diverse non-equilibrium behavior, in effect acting as a chemical engine. We explore the way that the energy released from reaction cycles is harnessed through kinetic selection in a series of what may have previously been considered somewhat disparate fields (systems chemistry, molecular machinery, supramolecular assembly and chemical oscillators), highlight common mechanistic principles, introduce concepts for the synchronization of chemical reaction cycles, and identify future challenges for the invention and application of non-equilibrium systems. MainBiological systems exhibit a broad range of complex functions, from transportation and motility 1-4 to temporal control 5,6 and information processing 7,8 , under non-equilibrium conditions realized by dissipating the chemical potential of high-energy species (typically the hydrolysis of adenosine triphosphate (ATP) or guanosine triphosphate (GTP) [3][4][5]7 ). Synthetic chemists are creating their own non-equilibrium systems, resulting in kinetic selection in catalysis, self-replicating systems 9,10 , dissipative assembly [11][12][13] , synthetic molecular machines 14-17 , and

We report a rotaxane-based information ratchet in which the macrocycle distribution is pumped away from equilibrium using a carbodiimide fuel. A carboxylate group on the axle, nonequidistant between two macrocycle binding sites, efficiently catalyzes the hydration of a carbodiimide fuel to the corresponding urea waste, with >80% of the fuel molecules reacting through the machine-catalyzed pathway. The energy of the reaction is harnessed by kinetic differentiation of the mechanical states of the machine driving the macrocycle to the binding site distal to the catalyst. Steric hindrance between the macrocycle and the fuel slows the reaction of the carboxylate group (to form a barrier to macrocycle movement) in the proximal co-conformer, whereas hydrogen bonding between the macrocycle and the barrier accelerates hydrolysis of the activated ester proximal isomer. The two directionally biased processes reinforce each other’s effect, resulting in a doubly kinetically gated ratchet that achieves 1:18 directionality, an exceptional degree of selectivity for a synthetic chemically fueled molecular motor.

Metallophilic interactions are increasingly recognized as playing an important role in molecular assembly, catalysis, and bio‐imaging. However, present knowledge of these interactions is largely derived from solid‐state structures and gas‐phase computational studies rather than quantitative experimental measurements. Here, we have experimentally quantified the role of aurophilic (AuI⋅⋅⋅AuI), platinophilic (PtII⋅⋅⋅PtII), palladophilic (PdII⋅⋅⋅PdII), and nickelophilic (NiII⋅⋅⋅NiII) interactions in self‐association and ligand‐exchange processes. All of these metallophilic interactions were found to be too weak to be well‐expressed in several solvents. Computational energy decomposition analyses supported the experimental finding that metallophilic interactions are overall weak, meaning that favorable dispersion and orbital hybridization contributions from M⋅⋅⋅M binding are largely outcompeted by electrostatic or dispersion interactions involving ligand or solvent molecules. This combined experimental and computational study provides a general understanding of metallophilic interactions and indicates that great care must be taken to avoid over‐attributing the energetic significance of metallophilic interactions.