Complexes of an Extended Inherently Chiral Tris-Bipyridyl Cage

Manganese(II), iron(II), cobalt(II), and copper(II) complexes of an extended inherently chiral tris-bipyridyl cage

David F. Perkins*, Leonard F. Lindoy*†, Alexander McAuley*‡, George V. Meehan§, and Peter Turner*

*Centre for Heavy Metals Research, School of Chemistry, University of Sydney, Sydney NSW 2006, Australia; and §School of Pharmacy and Molecular Sciences, James Cook University, Queensland 4811, Australia

Edited by Jack Halpern, University of Chicago, Chicago, IL, and approved November 28, 2005 (received for review September 29, 2005)

Manganese(II), iron(II), cobalt(II), and copper(II) derivatives of two -[inherently chiral, Tris(bipyridyl) cages (L and L؅) of type [ML PF6)2(solvent)n and [FeL؅](ClO4)2 are reported, where L is the) hexa-tertiary butyl-substituted derivative of L؅. These products were obtained by using the free cage and metal template procedures; the latter involved the reductive amination of the respective Tris-dialdehyde precursor complexes of iron(II), cobalt(II), or nickel(II). Electrochemical, EPR, and NMR studies have been used to probe the nature of the individual complexes. X-ray structures of the manganese(II), iron(II), and copper(II) complexes of L and the .iron(II) complex of L؅ are presented; these are compared with the Scheme 1. Structure of cage 1 previously reported structures of the corresponding nickel(II) complex and metal-free cage (L). In each complex the metal cation occupies the cage’s central cavity and is coordinated to six nitro- communication we (22) reported the successful expansion of the gens from the three bipyridyl groups. The cations [MnL]2؉ and size of the cavity in 2 by replacement of each of the 2,6-pyridyl FeL]2؉ are isostructural but both exhibit a different arrangement groups with 5,5Ј-substituted 2,2Ј-bipyridyl moieties to yield the] of the bound cage to that observed in the corresponding nickel(II) extended cage 3 [R ϭ tertiary butyl (t-Bu)] (Scheme 3). Semiem- and copper(II) complexes. The latter have an exo-exo arrangement pirical (pm3) calculations indicated that 3 (R ϭ H) in its exo-exo of the bridgehead nitrogen lone pairs, with the metal inducing a configuration should readily accept a divalent octahedral metal triple helical twist that extends Ϸ22 Å along the axial length of ion such that the latter induces a triple helical twist about the each complex. In contrast, [MnL]2؉ and [FeL]2؉ have their terminal (long) cage axis; comparison of Corey–Pauling–Koltun models nitrogen lone pairs directed endo, causing a significant change in indicated that 3 (R ϭ t-Bu) will almost certainly exhibit similar the configuration of the bound ligand. In [FeL؅]2؉, the cage has behavior. In both cases metal-ion-induced helicity is expected to both bridgehead nitrogen lone pairs orientated exo. Semiempirical be aided by the presence of the six salicycl-derived fragments calculations indicate that the observed endo-endo and exo-exo adjacent to the nitrogen bridgeheads (see below). Indeed the arrangements are of comparable energy. x-ray structure of the nickel(II) complex of 3 (R ϭ t-Bu) confirmed that the presence of this central metal results in a encapsulation ͉ template ͉ transition metal ͉ triple helix triple-helical twist of the cage and that the twist extends Ϸ22 Å along its axial length. The cage is present in its exo-exo config- ince the first reports of the binding of metal ions in the uration, with the nickel bound to the six pyridyl nitrogens to yield Scavities of aza molecular cages (1, 2), the metal complexation an octahedral coordination geometry. The x-ray structure of chemistry of such systems has received continuing attention metal-free 3 (R ϭ t-Bu) was also obtained (22). In this case the (3–9). Such 3D cages will normally result in enhanced steric and cage adopts a ‘‘loose’’ helical arrangement in the solid state. electronic restraints on the bound metal ion relative to related Nevertheless, a solution NMR study in CDCl3 and semiempirical 2D macrocyclic systems (the ‘‘cryptate effect’’) (10, 11), with [pm3] calculations both indicate that it is conformationally unusual spectral, photoactive, and electrochemical properties flexible. The above results are in accord with the original also often resulting. postulate that 3 (R ϭ t-Bu) is expected to form metal complexes We have previously reported the synthesis of new cages that with octahedral metal ions that exhibit extended triple-helical include 1 (12, 13) (Scheme 1) and 2 (14, 15) (Scheme 2) by means structures. Whereas 3 (R ϭ t-Bu) was initially synthesized in very of stepwise procedures based on functional group interconver- low overall yield {Ͻ2% from 5,5-bis[(2-formyl-4Ј-tert- sions from corresponding dialdehyde precursors. A Corey– butylphenoxy)methyl]-2,2Ј-bipyridine (4;Rϭ t-Bu) (Scheme 4) Pauling–Koltun model of 2 in its exo-exo configuration indicated by a ‘‘multistep’’ procedure}, the yield was subsequently in- that it has a ‘‘slot-like’’ cavity. Recrystallization of this cage from benzene yielded the corresponding host–guest product in which a benzene molecule is symmetrically incorporated in the cage’s Conflict of interest statement: No conflicts declared. cavity, aligned normal to the long axis (14). An additional This paper was submitted directly (Track II) to the PNAS office. feature of these cage structures is that the position of each set of Abbreviations: t-Bu, tertiary butyl; dmb, 5,5Ј-dimethyl-2,2Ј-bipyridyl. three benzo rings results in there being a strong tendency for the Data deposition: The atomic coordinates have been deposited in the Cambridge Structural corresponding bridgehead caps to adopt a chiral ‘‘three-bladed Database, Cambridge Crystallographic Data Centre, Cambridge CB2 1EZ, United Kingdom propeller’’ arrangement about each long axis. (CCDC reference nos. 284933–284938). In previous studies 2,2-bipyridyl groups have been successfully †To whom correspondence should be addressed. E-mail: [email protected]. incorporated in the structures of new cage species; the compl- ‡Permanent address: Department of Chemistry, University of Victoria, Victoria, BC, Canada exation behavior of such species toward a variety of metal ions V8W 3P6. and other guest species has been investigated (16–21). In a recent © 2006 by The National Academy of Sciences of the USA

532–537 ͉ PNAS ͉ January 17, 2006 ͉ vol. 103 ͉ no. 3 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0508539103 Downloaded by guest on September 25, 2021 Scheme 4. Structure of dialdehyde precursor 4.

copper(II) complexes each were obtained by the direct interaction of 3 (R ϭ t-Bu) with the corresponding metal nitrate in acetonitrile. The crude product was in each case then converted Scheme 2. Structure of cage 2. to its hexafluorophosphate salt and recrystallized from an acetonitrile͞ether mixture. In contrast to the direct synthetic procedures from the pre- creased to 11% by using a ‘‘one pot’’ reductive amination formed cage used for the above complexes, [CoL](PF6)2⅐2- procedure in which three equivalents of the dialdehyde were CH3CN⅐5H2O was obtained in 72% yield from an in situ (metal reacted with excess ammonium acetate and sodium cyanoboro- template) reductive amination procedure of the type mentioned hydride. The ammonium ion ultimately forms the bridgehead above, with the procedure being similar to that used previously nitrogens of 3 (R ϭ t-Bu) via three (sequential) imine conden- for obtaining [NiL](PF ) (22). In the cobalt complex case, ͞ 6 2 sation reduction steps (22). An additional modification, in cobalt(II) nitrate tetrahydrate was reacted with three equiva- which nickel ion was added to the above reaction mixture to act lents of dialdehyde 4 (R ϭ t-Bu) in hot acetonitrile, and the as a metal template, yielded the corresponding nickel(II) com- resulting solution was added dropwise to a warm acetonitrile ϭ plex of 3 (R t-Bu) in 75% yield. solution containing excess ammonium acetate and sodium cya- We now report the results of a comparative investigation of the noborohydride. The crude product was converted to its hexaflu- interaction of manganese(II), iron(II), cobalt(II), and copper(II) orophosphate salt before recrystallization from acetonitrile͞ ϭ ϭ with 3 (R t-Bu) and iron(II) with 3 (R H); the results are ether. also compared with those obtained in our previous study involv- A related in situ procedure was also used to obtain [FeL](PF6)2 ing the interaction of nickel(II) with these ligands (22). ⅐3H2O(Lϭ 3,Rϭ t-Bu) but in this case two distinct steps were involved. In the first the deep red Tris-ligand complex of Results and Discussion ϭ ϭ dialdehyde 4 (R t-Bu) was isolated. This solid product was Cage Synthesis. The free cage 3 (R Hort-Bu) was synthesized dissolved in warm acetonitrile, and the solution was then added by using two methods. The first involved the reductive amination slowly to excess ammonium acetate͞sodium cyanoborohydride of dialdehyde 4 (R ϭ Hort-Bu) in a one-pot procedure of the solution in a similar manner to that used for the cobalt system. type mentioned earlier; this procedure involved the dropwise

The product was obtained in 70% yield. An identical procedure, CHEMISTRY addition of the dialdehyde into a solution of ammonium acetate involving the initial isolation of [FeL ](PF ) (L ϭ 4,Rϭ H), and sodium cyanoborohydride. Alternatively 3 (R ϭ t-Bu) was 3 6 2 yielded the corresponding non-tert-butylated cage derivative, obtained by demetallation of the corresponding iron(II), cobal- [FeL](ClO ) (L ϭ 3,Rϭ H), in virtually quantitative yield. t(II), and nickel(II) complexes of 3 (R ϭ t-Bu) prepared by the 4 2 The 1H NMR spectrum of 3 (R ϭ t-Bu) at 298 and 318 K shows method outlined below. In this latter procedure, metal removal two singlets in the aliphatic region, one attributable to the involved refluxing the respective complex in methanol with -CH N- group at ␦ 3.80, the other caused by the -CH O- group sodium sulfide for 3 h. 2 2 at ␦ 5.02 in accord with the individual protons of each pair being 3 ϭ t Synthesis of Cage Complexes. The synthesis of selected metal equivalent, presumably reflecting the ability of (R -Bu) to complexes of 3 (R ϭ Hort-Bu) was performed by using both interchange between conformations via twisting about its N-N axis. In contrast, the 1H NMR spectrum of the low-spin dialde- direct synthesis from the preformed cage and an in situ metal-ion ϭ ϭ template directed route. The following solid complexes were hyde precusor complex [FeL3](PF6)2 (L 4,R t-Bu) at 298 K in CDCl3 shows nonequivalence of the methylene protons isolated and characterized: [MnL](PF6)2⅐2H2O, adjacent to the bipyridyl groups, giving rise to an AB system [FeL](PF6)2⅐3H2O, [CoL](PF6)2⅐2CH3CN⅐5H2O, and [CuL](PF ) ⅐2H O(Lϭ 3, R ϭ t-Bu) and [FeL](ClO ) (L ϭ 3, consistent with the presence of restricted rotation about the 6 2 2 4 2 1 R ϭ H). With the exception of the cobalt(II) complex, solid-state Ar-CH2-O- bonds. Also contrasting with the H spectrum of free ϭ ϭ ϭ x-ray diffraction structures of each of these complexes were 3 (R t-Bu), that of low-spin [FeL](PF6)2 (L 3, R t-Bu) in successfully determined (see below). The manganese(II) and acetone-d6 at both 300 and 320 K shows that the individual protons of both methylene groups (CH2O and CH2N) are nonequivalent; again each group appears as an AB system. Clearly the conformation of the bound ligand in this complex cation remains fixed in acetone-d6 over the 20°C temperature range.

EPR and Electrochemical Studies. The EPR spectrum of [MnL](PF6)2 (L ϭ 3, R ϭ t-Bu) at 4 K (see Fig. 5a, which is published as supporting information on the PNAS web site) is in accord with the manganese(II) ion being in a symmetrical octahedral environment; namely, the spectrum is representa- tive of an isotropic manganese(II) ion. There are several possible transitions for a d5 arrangement but the one least dependent on angle is the Ϫ1͞2toϩ1͞2 transition. This Scheme 3. Structure of cage 3. transition is split 6-fold because of the coupling of the 55Mn

Perkins et al. PNAS ͉ January 17, 2006 ͉ vol. 103 ͉ no. 3 ͉ 533 Downloaded by guest on September 25, 2021 nucleus (I ϭ 5͞2). A giso of 2.035 and an Aiso hyperfine value of 64G were determined. The EPR spectrum for [CoL](PF6)2 (L ϭ 3, R ϭ t-Bu) (see Fig. 5b) suggests that at4Kanequilibrium exists between high-spin (quartet) and low-spin (doublet) states. However, the room temperature magnetic moment (␮) of the solid complex at 27°C is 4.62(12) ␮B, indicating the presence of high-spin cobalt(II) at ambient temperature. The EPR spectrum of [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) at 4 K (see Fig. 5c) at 4 K shows a typical (23–25) Jahn-Teller distorted (tetragonally elongated) axial d9 spectrum with gʈ ϭ 2.277, gЌ ϭ 2.069, and hyperfine splitting Aʈ of 156 G, consistent with a ground state in which the unpaired electron resides in the dx2-y2 orbital. Hyperfine nitrogen splitting was not observed for each of the above complexes, a not unexpected result for systems of this nature. Cyclic voltammograms at several scan rates for [FeL]2ϩ (L ϭ 3, R ϭ H) in acetonitrile were measured (see Fig. 6, which is published as supporting information on the PNAS web site). A quasi-reversible [FeL]3ϩ͞[FeL]2ϩ (L ϭ 3, R ϭ H) oxidation wave is present at Ϸ1,105 mV (at 50 mV⅐sϪ1) and 1,140 mV (at 700 mV⅐sϪ1), and a second peak is also present at Ϸ100 mV less positive. The shape of the latter peak suggests it reflects an adsorption process at the electrode; this peak disappears at scan rates Ͼ850 mV⅐sϪ1. The cyclic voltammetric study of the analogous t-Bu-substituted complex [FeL]2ϩ (L ϭ 3, R ϭ t-Bu) at a scan rate of 100 mV⅐sϪ1 (and faster) showed a quasi-reversible [FeL]3ϩ͞[FeL]2ϩ (L ϭ 3, R ϭ t-Bu) couple at Ϸ1,190 mV (see Fig. 7, which is published as supporting information on the PNAS web site). For this system a second wave at a similar position to that observed for [FeL]2ϩ (L ϭ 3, R ϭ H) was observed only when the scan rate was reduced to 10 mV⅐sϪ1. Cyclic voltammetry of the related noncage system 2ϩ [Fe(dmb3)] (where dmb ϭ 5,5Ј-dimethyl-2,2Ј-bipyridyl) was investigated for comparison with the behavior of [FeL]2ϩ (L ϭ 3ϩ 2ϩ 3, R ϭ H and t-Bu). The [Fe(dmb3)] ͞[Fe(dmb3)] couple is highly reversible and appears Ϸ140 mV less positive than that for the [FeL]3ϩ͞[FeL]2ϩ (L ϭ 3, R ϭ t-Bu) couple and is independent of scan rate (see Fig. 8, which is published as supporting information on the PNAS web site). The secondary wave observed in the cyclic voltammograms of [FeL]2ϩ (L ϭ 3, R ϭ H) and [FeL]2ϩ (L ϭ 3, R ϭ t-Bu) at slower scan rates was not observed for this system. The reversibility of the oxidation͞ 2ϩ reduction curve for the noncage complex, [Fe(dmb3)] , compared with that of the cage complexes probably reflects a ϭ ϭ reduced facility for the latter to readily accommodate the Fig. 1. ORTEP (26, 27) depiction of [FeL](ClO4)2 (L 3, R t-Bu) with 20% changed radius of the metal ion in its new oxidation state. In this displacement ellipsoids. Hydrogen atoms, anions, and solvent molecules have context, the redox process in the latter case might, for example, been omitted. be associated with a slow ligand conformational change coupled to a spin state change. oriented either exo-exo or endo-endo, with little difference in the 2ϩ ϭ ϭ The cyclic voltammogram of [NiL] (L 3, R t-Bu) gave heats of formation between these two configurations (22). In no evidence for the presence of a [NiL]3ϩ͞[NiL]2ϩ (L ϭ 3, R ϭ view of this similarity, it was of interest to investigate which of t-Bu) oxidation wave. During a bulk electrolysis experiment in ϭ ϭ these arrangements occurs in the solid-state structures of the which a small amount of [NiL](PF6)2 (L 3, R t-Bu) was ϭ ⅐ Ϫ3 respective metal complexes of 3 (R Hort-Bu). In our previous dissolved in a 0.1 mol dm solution of Bu4NClO4 in dry ϭ acetonitrile and the mixture was subjected to a potential of 2.10 study the exo-exo arrangement was found for metal-free 3 (R V for 1 h, no change in the UV-visible spectrum of the solution t-Bu) and, as mentioned already, it is also present in the occurred. Furthermore, an EPR spectrum of a sample solution corresponding nickel(II) complex of this ligand. ϭ ϭ from the electrolysis cell at 4 K indicated that no nickel(III) was X-ray crystal structures of [CuL](PF6)2 (L 3, R t-Bu), ϭ ϭ ϭ ϭ present. However, during the electrolysis a vivid purple color [FeL](ClO4)2 (L 3, R H), [MnL](PF6)2 (L 3, R t-Bu), and ϭ ϭ appeared in the solution at the counter electrode. A subsequent [FeL](PF6)2 (L 3, R t-Bu) together with those for the two EPR spectrum of this solution indicated the existence of a ‘‘reference’’ complexes, [Fe(dmb)3](PF6)2 and [Cu(dmb)3](PF6)2, ligand-centered radical species. These observations are in accord have been determined. 3ϩ ϭ ϭ with [NiL] (L ϭ 3, R ϭ t-Bu) being a strong oxidant. ORTEP (26, 27) depictions of the [FeL](ClO4)2 (L 3, R H) and [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) complexes are provided in X-Ray and Computational Studies. An initial semiempirical (pm3) Figs. 1 and 2 (see also Table 1, which is published as supporting computational study of 3 (R ϭ H) for the gas phase indicated information on the PNAS web site), together with details for the that the bridgehead nitrogen lone pairs in this cage might be previously reported (22) and isomorphous [NiL](PF6)2 (L ϭ 3,

534 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0508539103 Perkins et al. Downloaded by guest on September 25, 2021 Fig. 3. Space-ﬁlling depiction of the x-ray structure of [CuL]2ϩ (L ϭ 3, R ϭ t-Bu) showing the helical twist along its axial length.

bond lengths in the crystal structure of the reference complex [Cu(dmb)3](PF6)2 are also essentially identical (see Table 2, which is published as supporting information on the PNAS web site). The apparent absence of Jahn-Teller bond length distor- tions in other copper(II) complexes has been reported (ref. 28 Fig. 2. ORTEP (26, 27) depiction of [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) with 20% displacement ellipsoids. The asymmetric unit contains two independent com- and references therein). Thus, based on the results of x-ray (29–33) or neutron (34) diffraction studies, other copper(II) plexes, each centered on a 3-fold rotation axis. Anions and solvent molecules CHEMISTRY have been omitted. systems with 3-fold symmetry axes have appeared to have apparently equivalent (or partially averaged) (24) coordination bond lengths. However, in most cases this equivalence has been R ϭ t-Bu) structure. The structures of these complexes are in attributed to underlying static disorder about the 3-fold axis accord with the extended triple-helical arrangement predicted masking the existence of Jahn-Teller-driven tetragonal distor- from the modeling studies (22); a space-filling depiction of the tion or alternatively to result from a dynamic Jahn-Teller x-ray structure of the copper complex illustrating the helical twist situation (35) where there is rapid interchange between the long is shown in Fig. 3. Notably, in each of these structures, an exo-exo and short coordination bonds in the tetragonally distorted terminal cap configuration is adopted. As mentioned previously, complex. In any case, as occurs in the present study, the above the analogous hexabenzo cages of types 1 and 2 (13) also appear apparent absence of a Jahn-Teller effect is contradicted by ESR to show a strong tendency for adopting an exo-exo disposition of evidence in a number of instances. their terminal bridgehead nitrogens. The displacement ellipsoids of the metal ion and donor atoms Both the nickel and copper structures are associated with two in [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) are unremarkable, and hence crystallographically independent complexes, each located on a give little evidence for underlying disorder. Intriguingly, refine- 3-fold axis and having opposing helical twists. In contrast, there ment of the [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) structure in the is only one complex in the asymmetric unit of the [FeL](ClO4)2 symmetry free space group P1 resulted in statistically signifi- (L ϭ 3,Rϭ H) structure, and it is not located on a symmetry cantly different coordination bonds lengths; however, this find- site. The two independent complex molecules of opposite hand- ing may be no more than an artifact of parameter correlation in edness observed for each of the nickel and copper structures may the least-squares refinement. Perhaps there is dynamic disorder be regarded as defining a dimeric unit, with the bipyridyl strand driven by the Jahn-Teller distortion in the present system, but of one molecule nestling into the ‘‘groove’’ formed by two there is no indication of its occurrence from the x-ray study. strands of its neighbor. The metal ion coordination is pseudooc- The helical pitch of the ligand strands in [CuL](PF6)2 (L ϭ 3, R ϭ tahedral in all three metal complexes. The two crystallographi- t-Bu), [FeL](ClO4)2 (L ϭ 3, R ϭ H), and [NiL](PF6)2 (L ϭ 3, R ϭ cally unique copper-to-nitrogen (coordinating) distances of one t-Bu) may be gauged from the independent angles subtended by the complex of type [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) are statistically principal cage axis and the principal bipyridyl axis (Table 1); they identical at 2.137(3) and 2.138(3) Å and yet differ in the second are much the same (38.2–40.1°) for these three complexes as complex being 2.152(2) and 2.142(2) Å, respectively. observed in the reference [Fe(dmb)3](PF6)2 and [Cu(dmb)3](PF6)2 The crystal structure of the above copper complex provides no structures (35.1–39.7°). Clearly the exo-exo configuration does not evidence for the anticipated tetragonal distortion, whereas, as constrain the coordination of the bipyridyl moieties in these cage mentioned earlier, the 4-K EPR spectrum of this complex shows complexes; thus it appears that whether it be nickel(II) or coppe- a classic d9 Jahn-Teller distorted patten, with anisotropic g r(II) the nature of the bound metal exerts little influence on the values, which are similar to those reported for a range of cage helicity or the disposition of the respective apical bridgeheads tetragonally distorted complexes. Similarly, the coordination in this set of exo-exo complexes.

Perkins et al. PNAS ͉ January 17, 2006 ͉ vol. 103 ͉ no. 3 ͉ 535 Downloaded by guest on September 25, 2021 respective sets of complexes is that the terminal t-Bu groups are further apart in the endo-endo structures than in the exo-exo structures. Perhaps surprisingly, the distance between the terminal bridgehead nitrogens in the endo-endo structures is longer than the corresponding distance in the exo-exo structures (Table 2). Likewise, the metal ions in the endo-endo species are further from the apical nitrogens than occurs in the exo-exo complexes. Such ‘‘stretching’’ of the complexes is reflected in the angles between the principal axis of each cage complex and the principal axis of its bipyridyl moieties, which are markedly smaller in the endo-endo structures than in the exo-exo species (Table 1) and also smaller than in the [Fe(dmb)3](PF6)2 and [Cu(dmb)3](PF6)2 structures (Table 2 and Figs. 9 and 10, which are published as supporting information on the PNAS web site). That is, the orientation of the terminal bridgehead nitrogens markedly affects the helical pitch of the ligand strands in the cage complexes. Nevertheless, as mentioned already, the two configurations are essentially energetically equivalent and it may be crystal-packing influences, and hence the crystal growth process, that determines which configuration is isolated in the solid state. In the above context it is noted that the synthetic route used does not appear to determine the orientation of the terminal bridgehead nitrogens. The [NiL](PF6)2 (L ϭ 3, R ϭ t-Bu) (exo-exo), [FeL](PF6)2 (L ϭ 3, R ϭ t-Bu) (endo-endo), and [FeL](ClO4)2 (L ϭ 3, R ϭ H) (exo-exo) were obtained by using the in situ procedure, whereas [CuL](PF6)2 (L ϭ 3, R ϭ t-Bu) (exo-exo) and [MnL](PF6)2 (L ϭ 3, R ϭ t-Bu) (endo-endo) were prepared by the reaction of the appropriate metal salt with the free cage. Full details of data collection and structure refinement for all complexes are included in Supporting Text, which is published as supporting information on the PNAS web site. Concluding Remarks In its exo-exo configuration, 3 (R ϭ t-Bu or H) is predisposed to form an extended triple helix on coordination to an octahedral metal ion. Clearly it is the tendency of the capping groups to twist into a propeller-like configuration (left- or right-handed) cou- ϭ ϭ Fig. 4. ORTEP (26, 27) depiction of [FeL](PF6)2 (L 3, R t-Bu) with 20% pled with the chiral octahedral geometry adopted by the central displacement ellipsoids. The asymmetric unit contains 1͞6 of a complex molecule, with the metal residing on a 3-fold roto-inversion site. Anions and solvent metal on coordination to the three bipyridyl residues that promotes the helical twist along the length of individual complex molecules have been omitted. [MnL](PF6)2 (L ϭ 3, R ϭ t-Bu) is isostructural. cations. Thus the x-ray structures of [NiL]2ϩ and [CuL]2ϩ (L ϭ 3,Rϭ t-Bu) confirmed that the central metal ion promotes a The distance between the metal ion and the apical nitrogen triple-helical twist that extends Ϸ22 Å along the axial length of atoms for each of the above complexes of 3 (R ϭ t-Bu) is each complex. This behavior is unusual relative to other ex- asymmetrical, presumably a manifestation of local packing in- tended helical metallo-structures reported previously (36, 37) in teractions. Further, the separation between the t-butyl residues which multiple metal-ion coordination is required to induce is greater at one end of each complex than at the other end helicity along the length of the respective systems. (Table 1). One complex is ‘‘longer’’ than its partner, with the Materials and Methods difference in the distance between the terminal bridgehead nitrogens of the two crystallographically independent molecules Details of the physical measurements, copies of EPR and cyclic being 0.88 Å for the nickel(II) species and 0.84 Å for the voltammograms, additional x-ray structural data {including copper(II) species. The separation between the respective met- structural figures for [Fe(dmb)3](PF6)2 and [Cu(dmb)3](PF6)2}, and the synthesis and characterization of the organic precursors als and the corresponding ‘‘capping’’ nitrogens precludes any 4 (R ϭ t-Bu, H) together with the reference complexes significant interaction between these centers in each complex. [Fe(dmb) ](PF ) and [Cu(dmb) ](PF ) are presented in Sup- The crystal structures of [FeL](PF ) (L ϭ 3, R ϭ t-Bu) and 3 6 2 3 6 2 6 2 porting Text. [MnL](PF6)2 (L ϭ 3, R ϭ t-Bu) (Fig. 4 and Table 2) are isostructural and, unlike the nickel and copper structures just t-Bu). Iron(II) chloride tetrahydrate ؍ R ;4 ؍Fe(L)3](PF6)2⅐3H2O(L] discussed, exhibit endo-endo terminal nitrogen configurations. (0.25 g, 1.26 mmol) was dissolved in acetonitrile (100 cm3). Further, there are no intermolecular groove interactions Dialdehyde 4 (2.03 g, 3.78 mmol) was added, and the red mixture present, and the ligand molecule is symmetrically disposed about was stirred at reflux for 20 min. The solvent was removed, and the central metal ion in each complex. In each case, the metal the deep-red solid was partitioned between aqueous potassium resides at the perpendicular intersection of a 3-fold axis, defining hexafluorophosphate (saturated, 50 cm3) and dichloromethane the molecule’s principal axis, and a 2-fold axis. All six MON (100 cm3). The organic fraction was washed twice with water (50 bonds are thus identical, and the apical nitrogen atoms are cm3), dried over anhydrous sodium sulfate, and evaporated to equi-distant from the respective metal centers. dryness on a rotary evaporator to give the crude product. This A consequence of the conformational differences between the material was recrystallized from dichloromethane͞ether to af- endo-endo and the exo-exo apical bridgehead dispositions in the ford [Fe(L)3](PF6)2 (L ϭ 4; R ϭ t-Bu) as a deep red powder (2.12

536 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0508539103 Perkins et al. Downloaded by guest on September 25, 2021 g, 86%), mp Ͼ250°C. Found: C, 60.78; H, 5.68; N, 3.97. cm3) was added dropwise over3htoavigorously stirred solution C102H108F12FeN6O12P2⅐3H2O requires C, 60.95; H, 5.72; N, of ammonium acetate (0.18 g, 2.5 mmol) and sodium cyanoboro- 4.18%. hydride (0.15 g, 2.5 mmol) in acetonitrile (500 cm3) at 50°C. The reaction mixture was maintained at 50°C for another hour, and H). In a similar manner to that described the solvent was removed on a rotary evaporator. The deep-red ؍ R ;4 ؍ FeL3](PF6)2 (L] above for [FeL3](PF6)2 (L ϭ 4;Rϭ t-Bu), dialdehyde 4 (R ϭ H) solid obtained was dissolved in dichloromethane and washed yielded a solid that was recrystallized from methanol to yield a with aqueous saturated potassium hexafluorophosphate. The deep-red microcrystalline solid, mp Ͼ250°C. Found: C, 57.66; H, organic phase was washed twice with water and dried over 3.91; N, 5.12. C78H60 F12 FeN6O12P2 requires C, 57.86; H, 3.74; anhydrous sodium sulfate, and the solvent was removed to yield N, 5.19%. a deep-red solid. This material was dissolved in a minimum of chloroform and chromatographed on a silica gel column. Frac- tions corresponding to the main band (second band) were -t-Bu). Manganese(II) nitrate hexahy ؍ R ,3 ؍MnL](PF6)2⅐2H2O(L] drate (0.0186 g, 0.0648 mmol) and 3 (R ϭ t-Bu) (0.1001 g, 0.0647 combined, and the solvent was removed. The resulting solid was 3 purified by using the following procedure: the compound was mmol) were dissolved in acetonitrile (100 cm ), and the solution 3 3 was stirred at reflux for 3 h. The solvent was removed on a rotary dissolved in acetonitrile (5 cm ) and chloroform (10 cm ), and evaporator, and the resulting pink solid was partitioned between diethyl ether was then added dropwise until the stirred solution dichloromethane and saturated KPF solution. The organic remained slightly cloudy. This mixture was left to stand at 4°C for 6 3 days. Deep-red microcrystalline [FeL](PF ) ⅐3H O(Lϭ 3, phase was washed three times with water, dried over anhydrous 6 2 2 R ϭ t-Bu) (0.218 g, 45%) was filtered off, mp Ͼ250°. Found: C, sodium sulfate, and evaporated to dryness on a rotary evaporator 63.11; H, 6.25; N, 5.60. C H F FeN P O ⅐3H O requires C, to give a pink solid. This solid was recrystallized from acetoni- 102 114 12 8 2 6 2 62.89; H, 6.21; N, 5.75%. trile͞ether to yield [MnL](PF6)2⅐2H2O(Lϭ 3,Rϭ t-Bu) as a reddish-pink microcrystalline solid that effloresced to become a H). In a similar manner to that described ؍ R ,3 ؍ FeL](PF ) (L] pink powder on filtration, mp Ͼ250°C. Found: C, 63.60; H, 6.31; 6 2 above for [FeL](PF6)2 (L ϭ 3,Rϭ t-Bu), [FeL3](PF6)2 (L ϭ 4, N, 5.91. C102H114F12MnN8P2O6⅐2H2O requires C, 63.51; H, 6.17; R ϭ H) (1.10 g, 0.668 mmol) yielded [FeL](PF6)2 (L ϭ 3,Rϭ N, 5.81%. H) as a deep-red solid (1.03 g, 99%), mp Ͼ250°C after recrystallization from acetonitrile͞ether. Found: C, 59.89; H, 4.50; N, t-Bu). In a similar manner to that ؍ R ,3 ؍CuL](PF6)2⅐3H2O(L] 6.97. C78H66F12FeN8O6P2 requires C, 60.16; H, 4.27; N, 7.20%. described above, copper(II) nitrate trihydrate (0.0200 g, 0.0828 mmol) and 3 (R ϭ t-Bu) (0.1281 g, 0.0828 mmol) in acetonitrile -t-Bu). This complex was pre ؍ R ,3 ؍CoL](PF6)2⅐2CH3CN⅐5H2O(L] 3 (100 cm ) gave a pale blue solid that was recrystallized from pared by an analogous procedure [from cobalt(II) nitrate tet- ͞ ⅐ ϭ ϭ acetonitrile ether, yielding [CuL](PF6)2 3H2O(L 3,R t-Bu) rahydrate] to that reported (22) for [NiL](PF6)2. The orange- as a pale greenish-blue microcrystalline solid that effloresced to yellow solid obtained was recrystallized from acetonitrile-ether become a pale blue powder on filtration, mp Ͼ250°C. Found: C, to yield [CoL](PF ) ⅐2CH CN⅐5H O(Lϭ 3,Rϭ t-Bu) as a ⅐ 6 2 3 2 62.88; H, 6.35; N, 5.81. C102H114F12CuN8P2O6 3H2O requires C, microcrystalline solid (72%), mp Ͼ250°C. Found (air dried): C, ⅐ ⅐ 62.65; H, 6.18; N, 5.73%. 61.35; H, 6.18; N, 6.87. C102H114CoF12N8O6P2 2CH3CN 5H2O CHEMISTRY requires C, 61.53; H, 6.33; N, 6.77%. t-Bu). A solution of [Fe(L)3](PF6)2 (L ϭ ؍ R ,3 ؍FeL](PF6)2⅐3H2O(L] 4;Rϭ t-Bu) (0.50 g, 0.25 mmol) dissolved in acetonitrile (400 We thank the Australian Research Council for support.

1. Boston D. R. & Rose, N. J. (1968) J. Am. Chem. Soc. 90, 6859–6860. 19. Seel, C. & Vo¨gtle, F. (1992) in Crown Compounds: Toward Future Applications, 2. Jackels S. C. & Rose, N. J. (1973) Inorg. Chem. 12, 1232–1237. ed. Cooper, S. R. (VCH, New York), p. 191. 3. Nelson, J., McKee, V. & Morgan, G. (1998) Prog. Inorg. Chem. 47, 167–316. 20. Echegoyen, L., Perezcordero, E., Regnouf de Vains, J. B., Roth, C. & Lehn, 4. Apperley, D. C., Clegg, W., Coles, S., Coyle, J. L., Martin, N., Maubert, B., J.-M. (1993) Inorg. Chem. 32, 572–577. McKee, V. & Nelson, J. (1999) J. Chem. Soc. Dalton Trans., 229–236. 21. Manchanda, V. K. & Mohapatra, P. K. (1993) J. Incl. Phenom. 15, 121–130. 5. Drew, M. G. B., Farrell, D., Morgan, G. G., McKee, V. & Nelson, J. (2000) 22. Perkins, D. F., Lindoy, L. F., Meehan, G. V. & Turner, P. (2004) Chem. J. Chem. Soc. Dalton Trans., 1513–1519. Commun., 152–153. 6. Arnaud-Neu, F., Fuangswasdi, S., Maubert, B., Nelson, J. & McKee, V. (2000) 23. Beveridge, K. A., McAuley, A. & Xu, C. (1991) Inorg. Chem. 30, 2074–2082. Inorg. Chem. 39, 573–579. 24. Bernhardt, P. V., Bramley, R., Engelhardt, L. M., Harrowfield, J. M., Hockless, 7. Gloe, K., Goretzki, G., McKee, V., Nelson, J., Nieuwenhuyzen, M., Pal, I., D. C. R., Korybut-Daszkiewicz, B. R., Krausz, E. R., Morgan, T., Sargeson, Stephan, H., Town, R. M. & Wichmann, K. (2003) J. Chem. Soc. Dalton Trans., A. M., Skelton, B. W. & White, A. H. (1995) Inorg. Chem. 34, 3589–3599. 1961–1968. 25. Kurosaki, H., Bucher, C., Espinosa, E., Barbe J.-M. & Guilard, R. (2001) Inorg. 8. Geoffrey, G., Jeffery, J. C. & Brooker, S. (2004) Eur. J. Inorg. Chem., 2570–2584. Chim. Acta 332, 145–149. 26. Johnson, C. K. (1976) ORTEPII: Report ORNL-5138 (Oak Ridge National 9. Escuer, A., McKee, V., Nelson, J., Ruiz, E., Sanz, N. & Vicente, R. (2005) Laboratory, Oak Ridge, TN). Chem. Eur. J. 11, 398–405. 27. Hall, S. R., du Boulay, D. J. & Olthof-Hazekamp, R. J., eds. (1999) Xtal3.6 10. Kauffmann, E., Lehn, J.-M. & Sauvage, J.-P. (1976) Helv. Chim. Acta 59, System (University of Western Australia, Perth). 1099–1111. 28. Persson, I., Persson, P., Sandstroem, M. & Ullstroem, A.-S. (2002) J. Chem. Soc. 11. Buschmann, H. J. (1987) Inorg. Chim. Acta 134, 225–228. Dalton Trans., 1256–1265. 12. Atkinson, I. M., Lindoy, L. F., Matthews, O. A., Meehan, G. V., Sobolev, A. N. 29. Joesten, M. D., Hussain, M. S. & Lenhert, P. G. (1970) Inorg. Chem. 9, 151–161. 47, & White, A. H. (1994) Aust. J. Chem. 1155–1162. 30. Cullen, D. L. & Lingafelder, E. C. (1971) Inorg. Chem. 10, 1264–1268. 13. Adam, K. R., Atkinson, I. M., Kim, J., Lindoy, L. F., Matthews, O. A., Meehan, 31. Ammeter, J. H., Burgi, H. B., Gamp, E., Sandrin-Meyer, V. & Jensen, W. P. G. V., Raciti, F., Skelton, B. W., Svenstrup, N. & White, A. H. (2001) J. Chem. (1979) Inorg. Chem. 18, 733–750. Soc. Dalton Trans., 2388–2397. 32. Astley, T., Ellis, P. J., Freeman, H. C., Hitchman, M. A., Keene, F. R. & 14. Janssen, R. J. A., Lindoy, L. F., Matthews, O. A., Meehan, G. V., Sobolev, A. N. Tiekink, R. T. (1995) J. Chem. Soc. Dalton Trans., 595–601. & White, A. H. (1995) J. Chem. Soc. Chem. Commun., 735–736. 33. Hitchman, M. A., Maaskant, W., Van der Plas, J., Simmons, C. J. & Strate- 15. Atkinson, I. M., Carroll, A. R., Janssen, R. J. A., Lindoy, L. F., Matthews, O. A. meier, H. (1999) J. Am. Chem. Soc. 121, 1488–1501. & Meehan, G. V. (1997) J. Chem. Soc. Perkin Trans. 1, 295–302. 34. Isaacs, N. W. & Kennard, C. H. L. (1969) J. Chem. Soc. A, 386–389. 16. de Mendoza, J., Mesa, E., Rodriguez-Ubis, J.-J., Vazquez, P., Vo¨gtle, F., 35. Murphy, B. & Hathaway, B. J. (2003) J. Coord. Chem. 243, 237–262. Windscheif, P.-M., Rissanen, K., Lehn, J.-M., Lilenbaum, D. & Ziessel, R. 36. Lindoy, L. F. & Atkinson, I. M. (2000) Self-Assembly in Supramolecular (1991) Angew. Chem. Int. Ed. Eng. 30, 1331–1333. Chemistry (Royal Society of Chemistry, Cambridge, UK). 17. Ebmeyer, F. & Vo¨gtle, F. (1989) Angew. Chem. Int. Ed. Eng. 28, 79–81. 37. Piguet, C., Borkovec, M., Hamacek, J. & Zeckert, K. (2005) Coord. Chem. Rev. 18. Seel, C. & Vo¨gtle, F. (1992) Angew. Chem. Int. Ed. Engl. 31, 528–549. 249, 705–726.

Perkins et al. PNAS ͉ January 17, 2006 ͉ vol. 103 ͉ no. 3 ͉ 537 Downloaded by guest on September 25, 2021