Communication pubs.acs.org/IC
Pyridinediimine Iron Complexes with Pendant Redox-Inactive Metals Located in the Secondary Coordination Sphere Mayra Delgado,† Joshua M. Ziegler,† Takele Seda,‡ Lev N. Zakharov,§ and John D. Gilbertson*,† †
Department of Chemistry and ‡Department of Physics, Western Washington University, Bellingham, Washington 98225, United States § Department of Chemistry, University of Oregon, Eugene, Oregon 97403, United States S Supporting Information *
ABSTRACT: A series of pyridinediimine (PDI) iron complexes that contain a pendant 15-crown-5 located in the secondary coordination sphere were synthesized and characterized. The complex Fe(15c5PDI)(CO)2 (2) was shown in both the solid state and solution to encapsulate redox-inactive metal ions. Modest shifts in the reduction potential of the metal−ligand scaffold were observed upon encapsulation of either Na+ or Li+.
temperature zero-field Mö s sbauer parameters [ΔE Q = 0.949(8); δ = 0.824(5) mm/s] confirm a five-coordinate highspin (S = 2) FeII center.14 1 was reduced under a CO atmosphere with NaHg amalgam in CH2Cl2 to form Fe(15c5PDI)(CO)2 (2). Slow evaporation of a saturated diethyl ether solution of 2 yielded a green, diamagnetic crystalline solid in 87% yield. The ATR-FTIR spectrum of 2 displays two νCO stretches at 1946 and 1882 cm−1 (1951 and 1884 cm−1 in solution). These are shifted from previously reported Fe(PDI)(CO)2 complexes, suggesting a more electronrich 15c5PDI ligand.15 An ORTEP view of 2 is shown in Figure 1.
R
edox-inactive metals are often used in combination with redox-active transition metals in synthetic and biological systems to invoke reactivity involving electron transfer.1,2 One particularly fascinating example is the role of the redox-inactive Ca2+ in the oxygen-evolving complex of photosystem II.3 Another is the Haber−Bosch process, which is carried out on potassium-promoted iron surfaces.4 The addition of redoxinactive Lewis acids shows enhanced electron-transfer rates, more positive reduction potentials,5 and enhanced rates of dioxygen activation6 in synthetic chemical systems. They also facilitate O−O,7 N−N,8 and H−H9 bond cleavage, as well as enable O- and H-atom transfer.10 Our group has been investigating the role of the secondary coordination sphere in complexes containing redox-active ligands of the pyridinediimine (PDI) scaffold. We have shown that pendant Bronsted bases/acids are capable of stabilizing rare intermediates11a and tuning the reduction potential of the ligandbased redox-active sites.11b,c We envisioned that positioning a Lewis acid binding site in the secondary coordination sphere would affect the reduction potentials in the PDI system (similar to redox-responsive pendant crown ferrocene systems).12 Herein we report a new family of PDI complexes that contain a pendant 15-crown-5 (15-c-5) capable of binding redox-inactive metals (Lewis acids) in the secondary coordination sphere. In order to synthesize a reduced Fe(PDI)(CO)2 complex with a pendant crown ether, the deep blue FeII(PDI) precursor Fe(15c5PDI)Cl2 (1) was first synthesized in high yields13 (∼90%) via the Schiff base condensation of the asymmetric ketone imine [(2,6-iPrC6H3NCMe)(OCMe)C5H3N] and 2-(aminomethyl)-15-crown-5 in the presence of FeCl2 (eq 1). Multiple attempts at achieving a publication-quality structure of 1 were unsuccessful because of severe disorder in the 15c5 (Figure S40). It is clear from the preliminary solution of 1 that the Fe center is five-coordinate, with the PDI N atoms and the Cl− ligands constituting the primary coordination sphere. The room © XXXX American Chemical Society
Figure 1. Chemdraw (left) and solid-state structure (right, 30% probability) of 2. The H atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Fe1−C34 1.784(4), Fe1−C33 1.787(5), Fe1−N1 1.964(3), Fe1−N2 1.857(3), Fe1−N3 1.965(3), C2−N1 1.333(5), C8−N3 1.323(5); C34−Fe1−C33 95.56(19), N2− Fe1−C33 153.18(17), N1−Fe1−N3 155.54(13).
The Fe center is five-coordinate with square-pyramidal geometry (τ = 0.04).16 The Cimine−Nimine bond lengths in 2 are 1.333(5) and 1.323(5) Å, and the Cimine−Cipso bond lengths are 1.431(5) and 1.444(5) Å. These data, taken in conjunction with the room temperature zero-field Mössbauer parameters [ΔEQ = 1.198(7); δ = −0.080(2) mm/s], suggest that the complex is best described as a S = 0 FeII center with a doubly reduced 15c5PDI ligand.14c,17 However, because of the known ambiguity17 in the oxidation Received: November 4, 2015
A
DOI: 10.1021/acs.inorgchem.5b02544 Inorg. Chem. XXXX, XXX, XXX−XXX
Communication
Inorganic Chemistry
The pendant Li+ ion in 4 (Figure S39) also lies to one side of the 15c5 ring at 0.3789(15) Å above the least-squares plane. The Li+ ion in 4 is six-coordinate, with five Li−O and one Li−F contacts (from PF6−). The Li−O contacts are shorter in 4 than Na−O in 3 because of the smaller alkali ion: Li1−O6 2.104(12) Å, Li1−O4 2.318(12) Å, Li1−O3 2.147(11) Å, Li1−O7 2.277(11) Å, and Li1−O5 2.153(11) Å. The Li1−F1 contact is 1.860(10) Å. Numerous structures containing Li+ and 15c5 from the CSD19 revealed an arrangement equivalent to 4. In order to probe the solution interaction of the pendant 15c5 in 2 with alkali metals, the C−H resonances of the 15c5 ring were interrogated via a series of 1H NMR titrations of varying amounts of either Na+ or Li+ in acetonitrile-d3 (MeCN-d3). As can be seen in Figure 3, the addition of increasing amounts of NaPF6 to a
state(s) in these systems, we are representing 2 in Figure 1 as a delocalized ligand system with mirrored symmetry.18 Compound 2 is able to chelate the alkali-metal ions Na+ and Li+, as illustrated by the ORTEP views of [Fe(15c5PDI)(CO)2Na][PF6] (3; Figure 2) and [Fe(15c5PDI)(CO)2Li][PF6]
Figure 2. Solid-state structure (30% probability) of 3. The H atoms have been omitted for clarity. Selected bond lengths (Å) and angles (deg): Fe1−C34 1.795(6), Fe1−C33 1.792(6), Fe1−N1 1.980(4), Fe1−N2 1.863(4), Fe1−N3 1.972(4), C2−N1 1.331(6), C8−N3 1.326(6), Na1−F1 2.438(4), Na1−F2 2.401(4), Na1−O(avg) 2.414; C34−Fe1− C33 97.5(2), N2−Fe1−C33 155.7(2), N1−Fe1−N3 154.79(16).
(4; Figure S39). Compounds 3 and 4 were prepared by stirring a solution of 2 in CH3CN with a slight excess of either NaPF6 or LiPF6, respectively. After workup, green diamagnetic X-rayquality crystals were obtained. As in 2, the Fe center in both 3 (τ = 0.01) and 4 (τ = 0.36) is five-coordinate with square-pyramidal geometry, with the Fe center in 4 being distorted. The oxidation state of the ligand is not changed upon alkali-metal chelation, as indicated by the Cimine−Nimine [1.326(6) and 1.331(6) Å in 3 and 1.312(7) and 1.325(6) Å in 4] and the Cimine−Cipso [1.441(7) and 1.434(7) Å in 3 and 1.424(8) and 1.428(7) Å in 4] bond lengths. The Mössbauer parameters for 3 and 4 (Table 1)
Figure 3. Plot of the chemical shift of the 15c5 C−H resonances in 2 versus added equivalents of NaPF6 (left) and corresponding stacked 1H NMR spectra of the C−H region of 15c5 in 2 (right). The spectra plotted correspond to 0 equiv of NaPF6 (bottom) up to 3 equiv (top).
solution of 2 causes a downfield shift in the resonances associated with the pendant 15c5 up to 1 equiv of NaPF6, suggesting formation of the 1:1 Na+ complex (3) in solution. Each spectrum was internally referenced to TMS, and the shift in the 15-c-5 resonances was significantly larger than any other resonance.20 The same behavior is observed when LiPF6 is used in place of NaPF6 (Figures S24 and S25), suggesting that the 1:1 Li+ complex (4) also forms in solution. The resonance corresponding to the tertiary C−H of the 15c5 ring at 4.04 ppm has a larger downfield shift in the LiPF6 titration (0.21 ppm) than NaPF6 (0.14 ppm). The resonance also retains its fine structure, whereas it is significantly broadened with NaPF6, perhaps suggesting more fluxional binding in the case of NaPF6. This hypothesis is further supported by an increased charge−radius ratio of Li+ and those data reported above on the positions of the alkali ions in the plane of O atoms (Li+ lies ∼0.30 Å deeper into the 15-c-5). To investigate the effect of the pendant alkali ion on the reduction potentials of 2−4, we examined the cyclic voltammetry (CV) in various solvents. Previous work11b,15,21 has shown that Fe(PDI)(CO)2 complexes are capable of undergoing a reversible (or quasi-reversible) “ligand-based” one-electron oxidation to form [Fe(PDI)(CO)2]+ and a reversible (or quasi-reversible) one-electron “metal-based” reduction to form [Fe(PDI)(CO)2]− (eq 2).22 The cyclic voltammograms of 2 in
Table 1. Selected Spectroscopic and Electrochemicala,b Data νCO (cm−1)c νCO (cm−1)d δ (mm/s) ΔEQ (mm/s) Eox(DCM) (V)e Ered(THF) (V)f Ered(MeCN) (V)g
2
3
4
1946, 1882 1951, 1884 −0.080(2) 1.198(7) −0.602 −2.591 −2.303
1936, 1868 1951, 1886 −0.087(5) 1.418(9) −0.544 −2.543 −2.293
1931, 1866 1951, 1886 −0.080(5) 1.75(1) −0.549 −2.591 −2.301
a
1 mM analyte, 0.1 M TBAPF6, glassy carbon working electrode, platinum counter electrode, and Ag/Ag+ in CH3CN reference electrode. bInternally referenced to Fc+/0. cATR-FTIR. dCH3CN FTIR. eOxidation (
