PAPER
www.rsc.org/dalton | Dalton Transactions
Supramolecular ‘flat’ Mn9 grid complexes—towards functional molecular platforms†‡ Victoria A. Milway,a S. M. Tareque Abedin,a Virginie Niel,a Timothy L. Kelly,a Louise N. Dawe,a Subrata K. Dey,a David W. Thompson,a David O. Miller,a Mohammad Sahabul Alam,b Paul M¨ullerb and Laurence K. Thompson*a Received 8th November 2005, Accepted 27th January 2006 First published as an Advance Article on the web 12th May 2006 DOI: 10.1039/b515801j Flat, quantum dot like arrays of closely spaced, electron rich metal centres are seen as attractive subunits for device capability at the molecular level. Mn(II)9 grids, formed by self-assembly processes using ‘tritopic’ pyridine-2,6-dihydrazone ligands, provide easy and pre-programmable routes to such systems, and have been shown to exhibit a number of potentially useful physical properties, which could be utilized to generate bi-stable molecular based states. Their ability to form surface monolayers, which can be mapped by STM techniques, bodes well for their possible integration into nanometer scale electronic components of the future. This report highlights some new Mn(II)9 grids, with functionalized ligand sites, that may provide suitable anchor points to surfaces and also be potential donor sites capable of further grid elaboration. Structures, magnetic properties, electrochemical properties, surface studies on HOPG (highly ordered pyrolytic graphite), including the imaging of individual metal ion sites in the grid using CITS (current imaging tunneling spectroscopy) are discussed, in addition to an analysis of the photophysics of a stable mixed oxidation state [Mn(III)4 Mn(II)5 ] grid. The grid physical properties as a whole are assessed in the light of reasonable approaches to the use of such molecules as nanometer scale devices.
Introduction Molecular construction techniques for coordination complexes have evolved from the simple starting point of mixing solutions of a ligand and metal ion, and taking a chance on the outcome, to the point where some predictability can be built into the reaction by pre-programming of the ligand itself. This does not come about initially by design, but usually follows a sequence of events where rational observations on reaction outcomes reveal the reaction secrets. Self-assembly methods play a pivotal role in biological chemistry, where subtle intermolecular forces can lead to the organization of molecular precursors, in order to facilitate important chemical processes. It is in the spirit of this molecular pre-arrangement that self-assembly methods have found an important role in synthetic coordination chemistry, where rational approaches to the synthesis of desired products are required. The ligand can express the outcome of a reaction with a metal ion if it takes into account the coordination preferences a Department of Chemistry, Memorial University, St. John’s, Newfoundland, A1B 3X7, Canada. E-mail:
[email protected]; Fax: +1-709-737-3750 b Physikalisches Institut III, Universit¨at Erlangen-N¨urnberg, D-91058, Erlangen, Germany † Based on the presentation given at Dalton Discussion No. 9, 19–21st April 2006, Hulme Hall, Manchester, UK. ‡ Electronic supplementary information (ESI) available: Fig. S1 (structural representation of cation 4), Fig. S2 (magnetic data (lmol /T) for complex 5), Fig. S3 (magnetic data (vmol /T, lmol /T) for complex 7), Fig. S4 (M/H at 2 K for complex 7), Fig. S5 (p–p contacts in 10), Fig. S6 (extended p–p and NH2 –NH2 contacts in 10), Fig. S7 (UV data on 8, 9); Table S1 (Band I and II analysis for 9). See DOI: 10.1039/b515801j
This journal is © The Royal Society of Chemistry 2006
of the particular metal. Lehn et al. described this in terms of the ‘coordination algorithm’.1 In essence one builds compartments or pockets into the ligand with desired donor arrangements, and places these pockets in strategic locations along the ligand backbone. Substituted pyrimidines (see for example Chart 1a) have been particularly successful in this context,2,3 leading to the formation of ‘ditopic’ ligands, which self assemble to form tetra-nuclear [2 × 2] grid complexes with a variety of metal ions.3 Extended ligand systems in this class have also produced grids with much higher nuclearity, with up to 16 metals in [4 × 4] arrays in the case of lead.4 Property–structure relationships are critical in terms of the design of a ligand, and within a multi-metallic assembly remote positioning of paramagnetic metal ions does not encourage intramolecular spin communication or lead to novel electronic properties. Small bridging groups (e.g. oxygen), capable of propagating efficient spin exchange between metal ions are important in this context, and bring the metal ions into close proximity when strategically located in a ligand framework. Substituted amidrazone ligands like poap (ditopic) and 2poap (tritopic) and their variants (Chart 1b) are ideally suited for this purpose, and self-assemble in high yield in the presence of a variety of metal ions (e.g. Mn(II), Fe(II), Fe(III), Co(II), Ni(II), Cu(II), Zn(II)) to produce tetra-nuclear [2 × 2]5,6 and nona-nuclear [3 × 3] grids (Scheme 1) respectively.7–17 The metal ions are held in close proximity within the grid framework by deprotonated hydrazone oxygen bridging atoms, with metal–metal separations ˚ . This leads to intramolecular exchange coupling with of ∼4 A all copper(II) examples exhibiting ferromagnetic exchange and all others antiferromagnetic exchange. Dalton Trans., 2006, 2835–2851 | 2835
mixed oxidation state manganese (Mn(III)/Mn(IV)) clusters (e.g. [Mn4 ]2 , Mn12 ),18,19 which display single molecule magnetization at low temperature, quantum tunneling of magnetization and single molecule hysteresis. Much larger systems (e.g. Mn84 , diameter 4.2 nm) can be synthesized, and also display single molecule magnetic behaviour.20 While magnetic systems in this class have the potential for storing binary code based on their single molecule magnetic properties, critical blocking temperatures for such systems are low (90%). Thus far the crystals obtained by this route have not been suitable for X-ray diffraction. Vis/NIR (k/nm) (methanol–acetonitrile (4 : 1)): 700 (e = 760 M−1 cm−1 ), 1000 (e = 800 M−1 cm−1 ); (mull) 650 (sh), 950 (br). Elemental analysis: Found (%): C, 34.18; H, 2.76; N, 18.53; Cl, 9.48. Calc (%) for 2838 | Dalton Trans., 2006, 2835–2851
(C19 H15 N9 O2 )6 Mn9 (ClO4 )6 Cl5 (H2 O)19 (CH3 OH)2 : C, 34.15; H, 3.33; N, 18.55; Cl, 9.45. Crystallography The diffraction intensities of a red-orange prismatic crystal of 1 of dimensions 0.52 × 0.30 × 0.26 mm, were collected with graphite-monochromatized Mo-Ka X-radiation using a Bruker P4/CCD diffractometer at 193(1) K to a maximum 2h value of 52.9◦ . The data were corrected for Lorentz and polarization effects. The structure was solved by direct methods.24,25 All atoms except hydrogens, and constituent atoms in some lattice acetonitrile molecules and some PF6 − anions were refined anisotropically. Hydrogen atoms were placed in calculated positions with isotropic thermal parameters set to twenty percent greater than their bonded partners, and were not refined. Neutral atom scattering factors26 and anomalous-dispersion terms27,28 were taken from the usual sources. All other calculations were performed with the teXsan29 crystallographic software package. The maximum and minimum peaks on the final difference Fourier map corresponded to 1.23 ˚ −3 , respectively. Abbreviated crystal data for 1 and −0.61 e A are given in Table 1. Diffraction intensities for 4, 5 and 6 were collected similarly, and structural solutions achieved using the same methods (see Table 1). The refinement of the putative Mn8 Co grid (6) was carried out assuming that all metal atoms were Mn(II), because of the difficulty of distinguishing the different metals. CCDC reference numbers 288821–288824. For crystallographic data in CIF or other electronic format see DOI: 10.1039/b515801j
Results and discussion Structures [Mn9 (2pomp)6 ](PF6 )6 ·3H2 O·4CH3 OH·2CH3 CN (1). A structural representation of the core grid cationic fragment in 1 is shown in Fig. 2. and important bond distances and angles are given in Table 2. The grid itself is similar to that reported for
C ) of the grid core structure in 1. Fig. 2 Structural representation (Povray
This journal is © The Royal Society of Chemistry 2006
Table 1 Crystal data and structure refinement for compounds 1, 4, 5 and 6
Chemical formula
1
4
5
6
C138 H120 O12 F36 N48 P6 Mn9
C140.5 H156 O24 N60 S12 Mn9
C135 H138 O39 F18 N54 S12 Mn9
C111 H114 Cl6 O47 N70 Mn9
4362.05 193(1) Triclinic P1¯ 19.171(2) 20.758(2) 24.905(2) 81.207(2) 75.954(2) 83.767(2) 9475(1) 2 1.529 8.11 73525 38653 (0.035) 0.093 0.331
3887.72 193(1) Triclinic P1¯ 18.119(2) 18.283(2) 26.240(2) 108.471(1) 93.504(2) 93.830(2) 8195(1) 2 1.575 8.66 55006 33117 (0.03) 0.097 0.330
M 4007.01 3948.33 T/K 193(1) 193(1) Crystal system Monoclinic Monoclinic Space group Cc P21 /c ˚ a/A 26.071(2) 28.447(3) ˚ b/A 26.150(2) 21.338(2) ˚ c/A 26.592(2) 33.290(4) a/◦ ◦ b/ 102.467(2) 111.525(2) c /◦ ˚3 V /A 17702(2) 18798(3) Z 4 4 Dc /g cm−3 1.503 1.395 l(Mo-Ka)/cm−1 7.8 7.92 Reflections total 70067 141230 Reflections unique (Rint ) 36112 (0.053) 38393 (0.082) a R1[I > 2r(I)] 0.088 0.104 wR2a 0.259 0.369 a R1 = [|F o | − |F c |]/ |F o |, wR2 = [ [w(|F o |2 − |F c |2 )2 ]/ [w(|F o |2 )2 ]]1/2 .
˚ ) and angles (◦ ) for Mn9 (2pomp)6 ](PF6 )6 ·3H2 O·4CH3 OH ·2CH3 CN (1) Table 2 Bond distances (A Mn1–O7 Mn1–N23 Mn1–O1 Mn1–N2 Mn1–N22 Mn1–N1 Mn2–O9 Mn2–N4 Mn2–N30 Mn2–O2 Mn2–O1 Mn2–N29 Mn3–N37 Mn3–O11
2.153(6) 2.169(8) 2.182(6) 2.189(7) 2.261(8) 2.268(7) 2.138(6) 2.159(8) 2.204(7) 2.211(6) 2.224(6) 2.279(8) 2.177(7) 2.177(6)
Mn4–N9 Mn4–O7 Mn4–O8 Mn4–N8 Mn5–N11 Mn5–N32 Mn5–O4 Mn5–O3 Mn5–O9 Mn5–O10 Mn6–O4 Mn6–N39 Mn6–O12 Mn6–N13
2.178(7) 2.208(6) 2.241(6) 2.293(7) 2.162(7) 2.170(6) 2.209(6) 2.211(5) 2.217(6) 2.222(6) 2.188(6) 2.197(7) 2.199(6) 2.200(7)
Mn3–O2 Mn3–N6 Mn3–N7 Mn3–N36 Mn4–O3 Mn4–N25 Mn7–N15 Mn7–N28 Mn8–N18 Mn8–O10 Mn8–N34 Mn8–O5 Mn8–O6
2.182(6) 2.182(8) 2.229(7) 2.283(9) 2.164(5) 2.164(7) 2.283(8) 2.300(8) 2.173(8) 2.180(6) 2.187(8) 2.224(6) 2.247(6)
Mn6–O11 Mn6–N14 Mn7–O8 Mn7–N16 Mn7–O5 Mn7–N27 Mn8–N35 Mn9–O6 Mn9–N41 Mn9–O12 Mn9–N20 Mn9–N42 Mn9–N21
2.239(6) 2.307(8) 2.175(6) 2.182(8) 2.200(6) 2.203(8) 2.299(8) 2.138(6) 2.153(9) 2.178(6) 2.186(9) 2.240(9) 2.301(7)
Mn1–O1–Mn2 Mn3–O2–Mn2 Mn4–O3–Mn5
128.7(3) 128.6(3) 128.7(3)
Mn1–O7–Mn4 Mn7–O8–Mn4 Mn2–O9–Mn5
128.6(2) 126.8(3) 128.5(3)
Mn6–O4–Mn5 Mn7–O5–Mn8 Mn9–O6–Mn8
128.4(3) 128.0(3) 128.1(3)
Mn8–O10–Mn5 Mn3–O11–Mn6 Mn9–O12–Mn6
127.8(3) 127.7(3) 129.8(3)
the complex [Mn9 (2pomp)6 ](MnCl4 )2 Cl2 ·2CH3 OH·7H2 O,16 but is unencumbered by the presence of competing paramagnetic subunits, which complicate the assessment of the grid magnetic ˚, exchange. The Mn–Mn distances fall in the range 3.92–3.98 A with Mn–O–Mn angles in the range 126.8–129.8◦ . Averaged Mn– ˚ , typical of Mn(II). L distances fall in the range 2.199–2.224 A This is typical for Mn grids in this class, and the overall grid dimensions are very similar to those of the parent complex [Mn9 (2poap)6 ](ClO4 )6 ·18H2 O (8) (Chart 1b; R = H, R = NH2 , X = CH).7 The substitution of the NH2 group in 8 for a methyl group has clearly had little impact on the way in which the ligands associate in the grid structure, and it is clear that the driving force for the ‘parallel’ arrangement of the two groups of ligands above and below the metallic plane is the alignment of the aromatic ˚ ). pyridine rings and close p contacts (ring–ring distances
