Portal de Periódicos da CAPES

Mechanically Interlocked Molecules (MIMs)—Molecular Shuttles, Switches, and Machines (Nobel Lecture)

2017; Wiley; Volume: 56; Issue: 37 Linguagem: Inglês

10.1002/anie.201703216

1521-3773

J. Fraser Stoddart,

Molecular Sensors and Ion Detection

Chemistry welcomes a new bond: The mechanical bond has endowed molecules with component parts whose movements can be controlled and monitored. In his Nobel Lecture, J. F. Stoddart describes how being able to template the formation of mechanically interlocked molecules has led to the design and synthesis of shuttles, switches, and machines at the nanoscale. One of the most influential books to have been written in the field of chemistry is The Nature of the Chemical Bond by Linus Pauling,1 the first edition of which was published in 1939. In this classic work, Pauling distinguishes between electrostatic, covalent and metallic bonds, while recognizing that chemical bonds, between two atoms or groups of atoms, exist when the forces acting between them lead to the formation of aggregates we call molecules. Three decades later came the realization that there is another field of chemistry that exists beyond the molecule, which Donald Cram2, 3 referred to as host–guest chemistry, and Jean-Marie Lehn4, 5 as supramolecular chemistry. Chemistry beyond the molecule relates to organized entities that can be neutral molecules, or even cations, anions or radicals, which come together to form higher-order aggregates—call them adducts or complexes—under the influence of stabilizing intermolecular forces that are considerably weaker than are the covalent bonds which define the entities themselves. The process involving the coming together by organized entities is often referred6 to as self-assembly and the noncovalent bonding that accompanies it as molecular recognition. The noncovalent bonds include hydrogen and halogen bonds amidst a gamut of weak interactions which have been exploited to considerable effect during the past half-century. On this time-scale, another type of bonding in chemistry had been lurking in the background. That bond is the mechanical bond,7 which was the subject of an extensive treatise8 entitled, The Nature of the Mechanical Bond: From Molecules to Machines, published as recently as 2016. Just as the chemical bond is associated in our minds with attractive forces, such as those associated with the sharing of electrons between atoms or the electrostatic forces that exist between ions of opposite charges, the mechanical bond is first and foremost a physical bond which is governed in the final analysis by repulsive forces that prevent chemical bonds from intersecting. Whereas chemical bonds are shared between atoms or groups of atoms, mechanical bonds are shared between molecular entities called component parts. It follows that a mechanical bond can be defined as an entanglement in space between two or more component parts such that they cannot be separated without breaking or distorting the chemical bonds between atoms. We have referred to molecules that possess mechanical bonds as mechanically interlocked molecules or MIMs for short. The two archetypal examples (Figure 1) of MIMs are the catenanes and the rotaxanes which were the subject of a monograph9 published in 1971 entitled Catenanes, Rotaxanes, and Knots by Gottfried Schill. A catenane, whose name is derived from the Latin word catena, meaning chain, is a molecule with two or more topologically linked macrocyclic component parts, while a rotaxane, whose name is derived from the Latin words rota for wheel and axis for axle, is a molecule comprising at least one macrocyclic component part, i.e., ring(s) with at least one linear component part, i.e., axle(s), threaded through the ring(s) and terminated by bulky end-groups (stoppers) large enough to prevent dethreading of the dumbbell(s) resulting from the existence/formation of chemical (covalent and/or coordinative) bonds between the stoppers and the axle(s). Two points are worthy of mention at this juncture. Catenanes and rotaxanes are molecules: they are not supramolecular entities or supermolecules despite the fact that they most likely will harbor intramolecular noncovalent bonds. While catenanes assume the non-trivial topologies of links, rotaxanes are topologically trivial for the simple reason that their component parts may be separated by continuous deformation, e.g., expanding the diameter of a ring or shrinking the cross-section of a stopper, both of which have been demonstrated8 chemically. Related is the physical process of making rotaxanes known as slippage.10, 11 Graphical representations of a) a catenane and b) a rotaxane. A catenane is a mechanically interlocked molecule consisting of two or more interlocked rings. The word catenane is derived from the Latin catena meaning chain. A rotaxane is a mechanically interlocked molecule consisting of a dumbbell component threaded by one or more rings. The word rotaxane is derived from the Latin rota for wheel and axis for axle. In the early 1990s the community were somewhat skeptical about the existence and worth of MIMs. Did they really exist and, if they did, how easy were they to make? If they were easy to make, in what context would they become useful? These were not unreasonable questions and they had to be addressed. In the event, it might well take, as of today, a decade or more to start providing answers to the second question. As far as the first question was concerned, it could be addressed rather easily, given the ability of X-ray crystallographer and structural chemist David Williams at Imperial College London to provide a fast turn around on the solid-state structures of catenanes, and occasionally rotaxanes as well. In the case of donor–acceptor catenanes,12 we were intent on demonstrating that higher-order analogues could be produced using a template-directed protocol.13 In this manner, both [3]- and [4]catenanes were synthesized in good yields overall and fully characterized. The envelope was pushed by two postdoctoral researchers, Anatoly Reder and David Amabilino, who, one after the other, took up the formidable challenge of making a [5]catenane (olympiadane) whose constitution was confirmed14 in 1996 by a solid-state structure (Figure 2), courtesy of David Williams. Ju-Young Lee gilded the lily by synthesizing a branched [7]catenane15 in one step from olympiadane. Both the collection of the crystallographic data, which took a couple of weeks, and the solving of the solid-state structure of the branched [7]catenane after a period of five months constituted a tour de force at the time by David Williams. The presence of 20 disordered PF6− counterions plus disordered solvent molecules on top of a crystallographic symmetry highlights the complexity of the task from both an experimental point of view and a computational one. The solid-state structures of these two higher-order catenanes revealed a veritable array of noncovalent bonding interactions in the form of face-to-face [π⋅⋅⋅π] stacking interactions (cf. DNA) and [C−H⋅⋅⋅π] edge-to-face alignments of aromatic rings which are commonplace in proteins rich in aromatic amino acids, along with multiple [C−H⋅⋅⋅O] hydrogen bonds. It is the installing of these intramolecular (ultimately) interactions that aids and abets the templation that makes it possible to synthesize these higher-order donor–acceptor catenanes. The point one learns from all this information is that it is the same weak interactions that are present in naturally occurring compounds that show up time and and time again in exotic unnatural products. At the level of noncovalent bonding interactions, their presence is ubiquitous throughout biology, chemistry and materials science. a) The Olympics logo consisting of five interlocked rings picked out from left to right in blue, yellow, black, green and red. b) A space-filling representation of the solid-state structure of a [5]catenane called olympiadane in which the five mechanically interlocked rings are colored according to the Olympics logo. c) A ball-and-stick structure of olympiadane which reveals that two outer blue rings are the same while the blue ring in the middle is a larger homologue of the two terminal rings. The other two red rings enjoy the same constitution in which three 1,5-disubstituted naphthalene rings are linked in a circular fashion, by tetraethylene glycol links. d) A graphical representation of olympiadane where the (red) π-electron rich rings are mechanically interlocked with three (blue) π-electron poor rings. It was an absence of the appreciation by chemists, in the era before there was a realization of the importance of noncovalent bonds and a recognition regime beyond the molecule, that led to a lack of success in the making of catenanes and rotaxanes. Wasserman's synthesis (Figure 3) of a [2]catenane16 in no more than a 1 % yield in 1960 bears witness to the fact that, without any appreciable noncovalent bonding interactions between the precursors to the component parts, the possibility of achieving mechanical interlocking to afford a [2]catenane was largely down to chance, hence the use of the term statistical synthesis to describe much of the research carried out in the 1960s. The achievement of Ed Wasserman (Figure 4) was to serve notice on the chemical community that, although catenation which relies on a chance event is, most likely not going to be efficient, it can be demonstrated. While Wasserman was carrying out his research at Bell Laboratories in Murray Hill, New Jersey, Gottfried Schill and Arthur Lüttringhaus at the University of Freiburg in Germany were devising ways17 by which the component parts of a [2]catenane could be brought together using a covalent bond that could be cleaved in the final steps of the synthesis. The key compound shown at the top in Figure 5 was obtained crystalline after 14 initially linear steps towards the directed synthesis of the [2]catenane shown at the bottom in Figure 5. The authors described the key compound as “one which is linked intra-annularly and in which the chains of the double ansa-system are situated on opposite sides of the benzene ring.” They also drew attention to the fact that the chain in the precursor to the key compound, which is attached by means of a cyclic acetal to the benzene ring, is held at right angles to the plane of the ring on account of the tetrahedral configuration about the aliphatic (acetal) carbon atom, thus ruling out the formation of the isomer with the extra-annular attachment of the ring. The final four steps of the 18-step synthesis all went in nearly quantitative yields to afford the [2]catenane. Aside from it being a long and difficult synthesis, the fact that once all the covalent bonds holding the component parts together have been cleaved, the two mechanically interlocked rings are essentially devoid of any ‘cross-talk’ between them. Nonetheless, Schill continued to be active in the field of directed synthesis of MIMs, including rotaxanes, until the early ’90s. I have commented elsewhere18 that Gottfried Schill (Figure 6) can be looked up as the father of the mechanical bond: he is a chemist who was decades ahead of his time! Wasserman's statistical synthesis reported in 1960 of arguably the first wholly synthetic [2]catenane employing an acyloin condensation in order to clip an acyclic diester around a deuterated cyclohexane derivative. The ring sizes, expressed as the number of carbon atoms in the rings, are denoted in square brackets. Deuterium labeling was employed as evidence in support of catenation following cleavage of the α-hydroxyketone functions with alkaline hydrogen peroxide. Ed Wasserman. The telescoped final four steps in a 22-step covalent-directed synthesis of a [2]catenane reported by Schill and Lüttringhaus in 1964. The ring sizes, given in terms of the number of carbon atoms present in the rings, are denoted in square brackets. Gottfried Schill. A seminal publication19 by Jean-Pierre Sauvage (Figure 7) in 1983 describing (Figure 8) the use of a copper(I) ion to template the formation of catenates—from whence, on demetallation, catenanes can be obtained—was a game-changer. His introduction of transition metal templation into the field of MIMs was transformative: it demonstrated that catenanes and rotaxanes were readily accessible and set the stage for the subsequent emergence and rapid growth of the field. Jean-Pierre Sauvage. Sauvage's entwining (Strategy A) and threading (Strategy B) for the copper(I)-templated synthesis of the first catenate in 1983. In the bottom right-hand corner is a tubular representation of the solid-state structure of the catenate. During a three-year secondment to the ICI Corporate Laboratory between 1978 and 1981, I joined forces with Howard Colquhoun in the investigation of the second-sphere coordination of transition metal ammines by crown ethers.20 This program of research was sustained in its speed and efficiency by our striking up a highly fruitful collaboration with David Williams. One of the solid-state superstructures he obtained in 1981 was to set us on a road to discovery and invention. The superstructure (Figure 9a,b) in question was that of a 1:1 crystalline adduct that is formed21 when a dicationic platinum ligand in addition to two cis-diammine ligands is crystallized (CH2Cl2/Et2O) in the presence of a molar equivalent of dibenzo[30]crown-10 (DB30C10). While the two ammine ligands from [N−H⋅⋅⋅O] hydrogen bonds with the two polyether loops of DB30C10, the bipyridyl ligand finds itself slotted (Figure 9 c) in between the two catechol rings of the crown ether. The structural similarity between the platinum complex and diquat—a compound marketed by ICI at the time in admixture with paraquat as a ‘wipe-out’ weedkiller—led us to show that it forms22 a 1:1 crystalline complex with DB30C10 as well. Its solid-state superstructure (Figure 9 d,e) mirrors that of the 1:1 adduct DB30C10 forms with the platinum complex. The diquat dication finds itself enjoying (Figure 9 f) charge transfer and π–π stacking interactions with the two catechol rings in the crown ether, aided and abetted by [C−H⋅⋅⋅O] hydrogen bonds. a) Space-filling and b) ball-and-stick representations of the 1:1 adduct formed between [Pt(bipy)(NH3)2]2+ and DB30C10 in the solid state. c) A graphical representation of the 1:1 adduct showing the [π⋅⋅⋅π] stacking interactions (sets of vertical lines) and the [N−H⋅⋅⋅O] hydrogen bonds and pole–dipole interactions between the dicationic transition metal diammine and some of the oxygen atoms in the polyether loops of DB30C10. d) Space-filling and e) ball-and-stick representations of the 1:1 complex formed between the diquat dication and DB30C10. f) A graphical representation showing the [π⋅⋅⋅π] stacking interactions (sets of vertical lines) and the [C−H⋅⋅⋅O] hydrogen bonds between the bismethylene bridge in the diquat cation and some of the oxygen atoms in the polyether loops of DB30C10. The next step, on my return to Sheffield in 1981, was to uncover a good crown ether receptor for paraquat. We had observed that DB30C10 was able to form a weak 1:1 complex with diquat and so we argued that a small change to the constitution of the crown ether might be all that is required in order to pinpoint a good crown ether receptor for paraquat. This strategy proved to be successful. We found that a constitutional isomer of DB30C10, namely bis-para-phenylene[34]crown-10 (BPP34C10), forms23 a strong 1:1 complex with paraquat. Moreover, the solid-state superstructure (Figure 10 a) of the crystalline 1:1 complex, when compared with the solid-state structure of BPP34C10, carried two very strong messages. One was the fact that the crown ether, at least as portrayed in the solid state, is preorganized to complex with the paraquat (PQT2+) dication. The other was that the manner in which the PQT2+ dication threads through the BPP34C10 ring is highly suggestive of it being a MIM precursor. In the context of molecular recognition, we had established how to bind a π-electron rich host—or using our color scheme, we could put blue inside red. The next question was—could we reverse the recognition motif and put red inside blue? Or expressed another way, could we bind a π-electron rich guest, e.g., 1,4-dimethoxybenzene (1/4DMB), inside a π-electron deficient host? With his synthesis24 of cyclobis(paraquat-p-phenylene) (CBPQT4+), Mark Reddington was able to show that 1/4DMB does indeed bind, albeit somewhat weakly, with CBPQT4+. Once again the solid-state superstructure (Figure 10 b) of the crystalline 1:1 complex, when compared25 with the solid-state structure of CBPQT4+, reveals a highly preorganized match between the guest and the rigid host. Comparison of the space-filling representations of the solid-state (super)structures of the two receptors a) BPP34C10 and b) CBPQT4+ and their 1:1 complexes obtained, respectively, with paraquat (PQT2+) and 1,4-dimethoxybenzene (1/4DMB). The box at the bottom relates the solid-state (super)structures to the structural formulas for the two receptors, BPP34C10 and CBPQT4+ and their 1:1 complexes with PQT2+ and 1/4DMB, respectively. We were now poised to perform the all-important experiment which would answer the question—what would happen if we carried out the synthesis of CBPQT4+ in the presence of BPP34C10? Cristina Vicent and Neil Spencer provided the answer by carrying out the reaction (Figure 11) in acetonitrile at room temperature in the presence of BPP34C10 as a template. The outcome from the first reaction exceeded our wildest dreams: we were able to isolate our first donor–acceptor [2]catenane in a remarkable 70 % yield. Our excitement did not stop there: a space-filling representation of the solid-state structure (Figure 12), which graced the front cover of the October 1989 issue of Angewandte Chemie,26 was a sight to behold. Moreover, both 1H NMR spectroscopy and electrochemical experiments, performed at the University of Miami by Angel Kaifer, conveyed a very strong message. The weak noncovalent bonding interactions (Figure 13 a) that are accrued during the template-directed synthesis ‘live on’ in the [2]catenane afterwards. The implications of this observation were to be profound when it came to designing and synthesizing, first of all molecular switches, and then ultimately molecular machines. The template-directed synthesis of the first donor–acceptor [2]catenane under kinetic control at room temperature in acetonitrile. BPP34C10, which was present in 3 molar equiv, acts as a template in the stepwise formation of the CBPQT4+ ring in a process where covalent bond formation leads to the production of a bipyridinium unit which is recognized noncovalently by the BPP34C10 when threading occurs: thereafter a second covalent bond is formed resulting in the formation of the [2]catenane where the noncovalent bonding interactions ‘live on’ inside the MIM. They are indicated by the parallel vertical lines. A space-filling representation of the solid-state structure of the first donor–acceptor [2]catenane adorning of the front cover of the October issue of Angewandte Chemie in 1989. The solid-state structures, depicted in ball-and-state representations, of a) the first donor–acceptor [2]catenane and b) a switchable donor–acceptor [2]catenane based on a stronger tetrathiafulvalene (green) and a weaker 1,5-dioxynaphthalene donor (pink). Note the presence in both [2]catenanes of π–π stacking interactions (3.5 Å) between the aromatic donors and acceptors. In the degenerate [2]catenane, note the presence of strong [C−H⋅⋅⋅O] interactions. As we took our leave of Sheffield in 1990 to move to Birmingham, Pier Lucio Anelli had completed the template-directed synthesis (Figure 14) of a degenerate [2]rotaxane27 modeled on that of the degenerate [2]catenane. The lower yield of 32 % reflects the fact that the dumbbell template is much less preorganized than BPP34C10. Dynamic 1H NMR spectroscopy, performed on the rotaxane in hexadeutero acetone, revealed that the CBPQT4+ ring shuttles (Figure 15) back and forth between the two hydroquinone units (‘stations’) around 1000 times a second. For obvious reasons, I found myself referring to the compound as a molecular shuttle and was prompted to draw the following conclusion in the communication27 published with this title in the Journal of the American Chemical Society in 1991: The template-directed synthesis of a degenerate donor–acceptor [2]rotaxane, also known as a ‘molecular shuttle’, under kinetic control at room temperature in acetonitrile. The dumbbell acts as template for the formation of the CBPQT4+ ring in a stepwise manner. The noncovalent bonding interactions which ‘live on’ inside the [2]rotaxane are indicated by parallel vertical lines. A degenerate donor–acceptor [2]rotaxane for which the phrase ‘molecular shuttle’ was introduced in 1991. The CBPQT4+ ring darts back and forth between the two hydroquinone recognition sites about 1000 times a second in acetone at room temperature. The molecular shuttle was seen as a prototype for the construction of molecular switches and machines based on molecules containing mechanical bonds. “The opportunity now exists to desymmetrize the molecular shuttle by inserting nonidentical ‘stations’ along the polyether ‘thread’ in such a manner that these different ‘stations’ can be addressed selectively by chemical, electrochemical, or photochemical means and so provide a mechanism to drive the ‘bead’ to and fro between ‘stations’ along the ‘thread’. Insofar as it becomes possible to control the movement of one molecular component with respect to the other in a [2]rotaxane, the technology for building ‘molecular machines’ will emerge. The molecular shuttle described in this communication is the prototype for the construction of more intricate molecular assemblies where the components will be designed to receive, store, transfer and transmit information in a highly controllable manner, following their spontaneous self-assembly at the supramolecular level. Increasingly, we can look forward to a ‘bottom up’ approach to nanotechnology, which is targeted toward the development of molecular-scale information processing systems.” The next challenge we faced was to desymmetrize the degenerate [2]rotaxane in an attempt to obtain a rotaxane with two recognition units on the dumbbell component, one of which is more attractive to being encircled by the CBPQT4+ ring than the other. After much initial experimentation we settled on a design where one of the hydroquinone units in the molecular shuttle is replaced by a benzidine unit and the other by a biphenol residue. Since we were prevented from using benzidine in the United Kingdom, postdoctoral scholar Richard Bissell made the journey across the Pond in order to complete the template-directed synthesis of the non-degenerate [2]rotaxane in collaboration with Angel Kaifer. On this occasion, dynamic 1H NMR spectroscopy revealed28 that the CBPQT4+ ring spends 84 % of its time on the benzidine unit and 16 % on the biphenol unit at equilibrium at room temperature in CD3CN solution. The unequal distribution of the ring between the two recognition units was sufficient to allow us to intervene in the equilibrium between the two translational isomers. It was possible, by adding acid to protonate the nitrogen atoms on the benzidine unit, to induce (Figure 16), as a result of Coulombic forces, the CBPQT4+ ring to move to the neutral biphenol unit rather than reside on the diprotonated benzidene unit carrying its two positive charges. While this bistable [2]rotaxane was to rank as our first switchable donor–acceptor MIM, a bistable [2]catenane,29 in which one of the two hydroquinone rings in the BPP34C10 component of the original [2]catenane (Figure 13 a) was replaced by a tetrathiafulvalene (TTF) unit and the other with a 1,5-dioxynaphthalene (DNP) unit was designed and synthesized by Gunter Mattersteig. The solid-state structure (Figure 13 b) of this bistable [2]catenane reveals that the TTF unit resides inside the cavity of the CBPQT4+ ring, while the DNP unit is located outside and alongside one of the two bipyridinium (BIPY2+) units in the CBPQT4+ ring. The first switchable donor–acceptor [2]rotaxane, where switching can be implemented by adding a drop of acid to protonate the benzidine recognition unit (the more preferred site for the location of the ring) whereupon the CBPQT4+ ring migrates to the biphenol unit, the less preferred site for the location of the ring on the neutral dumbbell component. The switch can be reset by adding base. This bistable [2]catenane was used30 to construct a solid-state electronically reconfigurable switch in a simple crossbar device by Jim Heath (Figure 17) when we entered into a highly fruitful collaboration, following my move from Birmingham to the University of California Los Angeles (UCLA). Prior to introducing this bistable [2]catenane into the crossbar device we had ascertained31 how to effect its transfer onto the device using the Langmuir–Blodgett (LB) technique from a Langmuir monolayer where the counterions are dimyristoylphosphatidyl anions. The ON/OFF ratios for the molecular switch tunnel junctions (MSTJs) in crossbar devices incorporating the bistable [2]catenane were only around 2. The desire to increase this ratio took us inexorably in the direction of bistable [2]rotaxanes (Figure 18) which, it transpired, can be switched with an order of magnitude higher ON/OFF ratios. Reversible, electronically driven switching was not only observed in MSTJs incorporating a molecular monolayer of this bistable [2]rotaxane sandwiched between a bottom polysilicon electrode and a top titanium/aluminum electrode, but a 160,000-bit molecular electronic memory circuit has also been fabricated (Figure 19 a) at a density of 100,000,000,000 bits per square centimeter.31,32 The entire 160-kbit crossbar assembly was smaller than the cross-section of a white blood cell and could be switched (Figure 19 b) up to around 100 times. The search for a more robust setting34 in metal–organic frameworks for these bistable MIMs continues.35 Jim Heath. Top left: Scanning electron microscope image of the nanowire crossbar memory. The array of 400 Si bottom nanowires is portrayed as the light grey rectangular patch extending diagonally up from the left. Bottom left: Scanning electron microscope image showing the cross-point of the top (red) and bottom (yellow) nanowire electrodes. Each cross-point corresponds to an ebit in memory testing. Top right: High resolution scanning electron microscope image of approximately 2500 junctions out of a 160,000-junction nanowire crossbar circuit. The red square highlights an area of the memory that is equivalent to the number of bits tested. Positioned in the center of the illustration is the structural formula of the bistable [2]rotaxane used in the memory. a) A 160-kilobit crossbar molecular memory device (orange in the middle of the illustration) is smaller than the cross-sectional area of a white blood cell (purple). b) A demonstration of point-addressability within the crossbar. Good ebits were selected from the defect mapping of the tested portion of the crossbar. A string of 0s and 1s corresponding to the ASCII characters for ‘CIT’ (California Institute of Technology) were stored and read out sequentially. The dotted line indicates the separation between the 0 and 1 states of the individual ebits. The black trace is raw data showing 10 sequential readings of each bit while the red bars represent the average of these 10 readings. In a collaboration with Jeff Zink (Figure 20) at UCLA, we have functionalized the surfaces of mesoporous silica nanoparticles (MSNPs) with both rotaxane-based nanovalves and snap-tops (Figure 21) for the controlled release of drugs. The nanovalves which adorn the surfaces of the MSNPs can be bistable donor–acceptor [2]rotaxanes36 where redox chemistry can be used to open and close the nanovalves. This type of integrated device holds considerable promise for drug delivery systems. Jeff Zink. Examples of mechanized mesoporous silica nanoparticles (MSNPs). Illustrations of the mode of operation of rotaxane-based nanovalves and snap-tops for the controlled release of drugs. Scanning electron microscopic (bottom left) image of MSNPs. Transmission electron microscopic (bottom right) image of MSNPs. A fundamental property of biological molecular machines, e.g., the motor proteins,37-44 is that they consume energy and drive systems away from equilibrium by controlling kinetic barriers. The emergence of the science of artificial molecular machines (AMMs) in the hands of chemists, for the most part, presents us with a steep learning curve to negotiate. Nonetheless, it is a challenge that the chemistry community has embraced increasingly during the past two decades with much of the teaching of the fundamental theory coming from the physics community, and, in particular, one very prominent physicist from the University of Maine, namely Dean Astumian45-48 with whom we have collaborated as well as published several reviews49-51 that stand alongside a crop52-69 that can be traced back for at least two decades from the present day. In my Nobel Lecture, I chose to highlight the progress50, 51 we have made in recent times on the design and synthesis of artificial molecular pumps, based on the building blocks that found their origins almost 40 years ago in the ICI Corporate Laboratory and were used by us21-27 to put the mechanical bond on a firm footing before introducing28 them into molecular switches in the first instance. At the outset, we designed and synthesized a constitutionally unsymmetrical dumbbell34 and demonstrated (Figure 22) that it transports the CBPQT4+ ring in a unidirectional manner. In its midriff, the dumbbell houses an electron rich 1,5-dioxynaphthalene (DNP) unit which is capable of recognizing the electron poor CBPQT4+ ring. Kinetic control of the threading and dethreading of the dumbbell by the ring is managed by arranging to have a neutral isopropylphenyl (IPP) unit at one