15. P. L. Anelli, N. Spencer, J. F. Stoddart, J. Am. Chem. Soc. 113, 5131 (1991). 16. J.-P. Sauvage, C. Dietrich-Buchecker, Eds., Molecular Catenanes, Rotaxanes and Knots (VCH-Wiley, Wein- heim, Germany, 1999). 17. Y. Luo et al., ChemPhysChem 3, 519 (2002). 18. M.-V. Martı ´nez-Dı ´az, N. Spencer, J. F. Stoddart, An- gew. Chem. Int. Ed. 36, 1904 (1997). 19. S. J. Cantrill, A. R. Pease, J. F. Stoddart, J. Chem. Soc. Dalton Trans. 21, 3715 (2000). 20. M. Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. Int. Ed. Engl. 37, 2754 (1998). 21. V. Balzani et al., Chem. Eur. J. 9, 5348 (2003). 22. S. Anderson, H. L. Anderson, J. K. M. Sanders, Acc. Chem. Res. 26, 469 (1993). 23. P. M. S. Monk, The Viologens. Physicochemical Prop- erties, Synthesis and Application of the Salts of 4,4'- Bipyridine (Wiley, Chichester, UK, 1998). 24. Although multivalency is a concept that is well- established with regard to numerous biological systems, its investigation has been restricted large- ly to the acquisition of thermodynamic parame- ters, e.g., association or dissociation constants and free energies, enthalpies, and entropies of compl- exation. In the present investigation, the opportu- nity arose to study within a molecularly con- strained environment how three equivalent recog- nition sites respond to being addressed chemically. The fact that they switch in series rather than in parallel is a feature worthy of note. Although many might argue intuitively that multivalency from a mechanistic standpoint has to be a stepwise pro- cess, the results presented in this article provide indirect yet convincing experimental evidence for the one-at-a-time mechanism; i.e., by taking single steps rather than moving in concert, the molecular elevator is more reminiscent of a legged animal than it is of the passenger elevator. 25. This research was supported by NSF (CHE- 0317170, CHE-9974928, and CHE-0116853) in the United States and the University of Bologna (Funds for selected topics), Ministero dell’Istruzione, dell’Universita ` e della Ricerca (Supramolecular de- vices project), Fondo per gli Investimenti della Ricerca di Base (RBNE019H9K) and the European Union (Molecular-Level Devices and Machines Net- work HPRN-CT-2000-00029) in Italy. We thank M. Venturi and P. Ceroni for useful discussions. Supporting Online Material www.sciencemag.org/cgi/content/full/303/5665/1845/ DC1 Materials and Methods SOM Text Scheme S1 Figs. S1 and S2 References 17 December 2003; accepted 12 February 2004 Electrical or Photocontrol of the Rotary Motion of a Metallacarborane M. Frederick Hawthorne, 1 * Jeffrey I. Zink, 1 Johnny M. Skelton, 1 Michael J. Bayer, 1 Chris Liu, 1 Ester Livshits, 2 Roi Baer, 2 Daniel Neuhauser 1 Rotary motion around a molecular axis has been controlled by simple electron transfer processes and by photoexcitation. The basis of the motion is intramo- lecular rotation of a carborane cage ligand (7,8-dicarbollide) around a nickel axle. The Ni(III) metallacarborane structure is a transoid sandwich with two pairs of carbon vertices reflected through a center of symmetry, but that of the Ni(IV) species is cisoid. The interconversion of the two provides the basis for con- trolled, rotational, oscillatory motion. The energies of the Ni(III) and Ni(IV) species are calculated as a function of the rotation angle. The smallest nanomachines made to date are based on changes in molecular bonding. In bio- logical motors, such as those that power the linear motion of myosin head in muscle contraction ( 1, 2) or the rotary motion of F 1 -adenosine triphos- phatase ( 3–5) in bacterial flagella, the hydrolysis energy of the adenosine triphosphate (ATP) to diphosphate reaction creates new bonds that can exert forces that change the shape of the larger molecule and perform work. Given the existence of biological motors, the interest of chemists in designing molecular motors stems from the challenge not only of making even smaller nanomachines that perform controllable motion ( 6), but also of creating systems that can be powered with light or electrical energy, rather than depending on the delivery of ATP. Linear motors have been developed such as the rotaxane systems ( 7–9), in which a shuttle ring component slides from one physically discrete zone of a rigid rod to another. Assessing a potential molecular motor based on changes in bonding must include determining whether the system has well- defined reactant and product states that are separated by a considerable energy barrier. The reasons are twofold. First, there has to be some way of putting into the molecule energy that will be recovered later as work. For photon-driven systems, this would likely be an excited state; for electrically driven sys- tems, this could be the formation of an inter- mediate with a different redox state. Second, in either case, the product formed must have a barrier against back reaction, or it is unlike- ly that the system would exert any force; it would more likely relax back unproductively to the reactant configuration. With these configurations in mind, we reex- plored a small metal complex with potential as a two-state rotary machine that could be used, for example, to open or close a valve or switch. Examples to date of artificial rotors have been few. Rotary motion in synthetic arrays is rep- resented by molecular windmills and turnstiles in which one component of the molecular de- vice is capable of freewheeling rotation on a molecular axis (10), but such barrier-free unre- stricted rotation cannot be powered or con- trolled. The rotation of a dipolar rotor in a rotating electric field has been modeled (11). Examples in which control is possible are provid- ed by the interlocked rings of catenanes that undergo circumrotation by passing through each other ( 12, 13). However, machines with rotary motion about a rigid molecular axis, controlled by simple electron transfer processes or by photoex- citation, have not been previously reported. The basis of the molecular device reported here has been known since the discovery (14 ) of the d 7 Ni(III) and d 6 Ni(IV) commo-bis- 7,8-dicarbollyl metallacarboranes, denoted as 1T (15 ) and 1C (16 ) (Fig. 1), and their palladium analogs. These complexes are pro- duced by the coordination of two dicarbollide ions (17 ) (Fig. 1, 2) with a Ni(II) ion, fol- lowed by subsequent oxidation. The “sand- wich” species may undergo rotary motion of the ligands with respect to each other. This is analogous to the well-known metallocenes (18), but those have rotational barriers only of 2 kcal/mol (19.) The present metallacarbo- ranes have barriers about three times as large, leading to temperature invariance in their nu- clear magnetic resonance (NMR) spectra. The presence of two adjacent CH vertices in the bonding face of 2 introduces localized re- gions of reduced negative charge (20) and an antipodal concentration of negative charge. The resulting interligand interaction leads to a trans configuration (15, 17, 21) (such as in 1T) in most examples of commo-bis-7,8-dicarbollyl metallacarboranes. One of the few exceptions is the Ni(IV ) species (1C), which is cisoid, with its pairs of carbon vertices on the same side of the molecule (14, 15 ). Nickel rarely appears in the formal +4 oxidation state in inorganic structures (22) and never in organometallic compounds, thereby making 1C quite unusual. The interconversion of the 1T and 1C geometries when the nickel oxidation state is changed provides the basis for the controlled, rotational, oscillatory motion and can be achieved electrochemically, by redox reac- tions, or photochemically. Thus, an oscillato- ry molecular rotor, providing useful work on the molecular scale, is plausible, based on the 1 Department of Chemistry and Biochemistry, Univer- sity of California, Los Angeles, CA, 90095–1569, USA. 2 Department of Chemistry, Hebrew University of Jerusalem, Givat Ram, Jerusalem, Israel. *To whom correspondence should be addressed. E- mail: mfh@chem.ucla.edu R EPORTS www.sciencemag.org SCIENCE VOL 303 19 MARCH 2004 1849