Molecular Shuttles Entropy-Driven Translational Isomerism: A Tristable Molecular Shuttle** Giovanni Bottari, Francois Dehez, David A. Leigh,* Phillip J. Nash, Emilio M. PØrez, Jenny K. Y. Wong, and Francesco Zerbetto* Stimuli-responsive molecular shuttles translocate a macro- cycle between different sites (“stations”) on a rotaxane thread under the influence of an external trigger. [1] In bistable shuttles the relative macrocycle binding affinities of the stations are reversed by the stimulus, generally through it bringing about a chemical change in the molecule that targets the enthalpy of binding of the macrocycle to one or both stations. [2] Immediately following the chemical transforma- tion the molecule is no longer in the most energetically favored co-conformation and the macrocycle moves along the thread to its newly preferred position through biased Brown- ian motion as the system relaxes to the global minimum. [3] Although many external stimuli can be used to induce shuttling in this way, for example pH change, [4] light, [5] and electrochemistry [4a,6] , a simple temperature change is not generally considered one of them. [7] The Boltzmann distribu- tion of the macrocycle between the different binding sites within a shuttle ensures that heating or cooling changes the degree of discrimination the macrocycle expresses for the various stations, but not the actual station preference of the macrocycle. However, a change of relative-station binding affinity with temperature is possible in principle, since DG binding = DH binding ÀTDS binding . If the entropy terms are sufficiently different then the relative binding affinity of the macrocycle for the two stations can be reversed by increasing or lowering the temperature. Here we describe an example of this phenomenon. [8] The [2]rotaxane 1 is, in fact, a tristable molecular shuttle; the first rotaxane in which a ring can be switched between three different positions on a thread (Figure 1). [9] Rotaxane E-1 was prepared in 32 % yield from thread E-2 (Scheme 1). E-2 has previously [5g] been utilized as the thread for a light- and heat-switchable bistable molecular shuttle 3, and contains two sites designed to hydrogen bond to a benzylic amide macrocycle, namely a fumaramide group (shown in green) and a succinic amide ester unit (orange), separated by a dodecane chain (purple). Shuttle 1 differs from 3 only in that the macrocycle contains endo-pyridine units instead of isophthalamide groups. Photoisomerization of E-1 at 254 nm afforded the cis-rotaxane Z-1 in 54 % yield. Since the xylylene units of the macrocycle shield the encapsulated regions of the thread, the position of the ring in E- and Z-1 could be determined by comparing the chemical shift of the protons in the [2]rotaxanes with those of the corresponding threads (Figure 2). The 1 H NMR spectra (400 MHz, 298 K; Figure 2 a and b, see page 5888) confirm the position of the macrocycle over the fumaramide station of E-1 in CDCl 3 . The olefin protons H i and H j are shielded by more than 1.5 ppm in the rotaxane relative to the thread, while the chemical shifts of the succinic amide ester protons H c and H d are unchanged. Lowering the temperature had no effect on the chemical-shift values, the only significant change in the spectra being that the macro- cycle H E protons sharpen as the ring pirouetting about the thread becomes slow on the NMR timescale. In Z-1, the strong binding fumaramide station is replaced with a group of much poorer macrocycle-binding affinity (maleamide) and we expected the macrocycle to be displaced to the succinic amide ester site on the thread (that is, co- conformer succ-Z-1), as occurs with Z-3. [5g] Whilst the chemical-shift differences (> 1.2 ppm, COSY) of the H c and H d protons confirm that this is largely the case [10] at room temperature and above (e.g., at 308 K; Figure 2d), to our surprise the 1 H NMR spectrum of Z-1 proved highly temper- ature dependent. Indeed, at 258 K (Figure 2e) the major signals for H c and H d of Z-1 appear at the same chemical shifts as they do in the thread (Z-2). In addition the olefin protons H i’ and H j’ are also unchanged indicating that the macrocycle is not primarily located over either of the designed stations! In fact, it is the alkyl protons of the C 12 chain that experience significant upfield shifts (up to 1 ppm at 258 K), which indicates that the pyridine macrocycle is actually positioned Figure 1. Macrocycle translation in a tristable molecular shuttle. Mac- rocycle movement between 1) and 2) is an entropy-driven process involving no change to the covalent structure of the molecule. [*] Prof. D. A. Leigh, Dr. G. Bottari, P. J. Nash, E. M. PØrez, Dr.J.K.Y.Wong SchoolofChemistry,UniversityofEdinburgh,TheKing'sBuildings, West Mains Road, Edinburgh EH9 3JJ (UK) Fax:(+ 44)131-667-9085 E-mail: David.Leigh@ed.ac.uk Prof. F. Zerbetto, Dr. F. Dehez Dipartimento di Chimica “G. Ciamician”, Università degli Studi di Bologna, via F. Selmi 2, 40126 Bologna (Italy) Fax (+ 39)051-2099456 E-mail: gatto@ciam.unibo.it [**] This work was supported by the European Union Future and Emerging Technology Program MechMol, the EPSRC, and the MURSTproject “Dispositivi Supramolecolari”. D.A.L. is an EPSRC Advanced Research Fellow (AF/982324). Supporting information for this article is available on the WWW under http://www.angewandte.org or from the author. Communications 5886 # 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim DOI: 10.1002/anie.200352176 Angew. Chem. Int. Ed. 2003, 42, 5886–5889