Article pubs.acs.org/IC
Redox Pairs of Diiron and Iron−Cobalt Complexes with High-Spin Ground States Deanna L. Miller,† Randall B. Siedschlag,† Laura J. Clouston,† Victor G. Young, Jr.,† Yu-Sheng Chen,‡ Eckhard Bill,*,§ Laura Gagliardi,*,†,∥ and Connie C. Lu*,† †
Department of Chemistry, University of Minnesota, 207 Pleasant Street SE, Minneapolis, Minnesota 55455-0431, United States Supercomputing Institute and Chemical Theory Center, University of Minnesota, Minneapolis, Minnesota 55455, United States § Max Planck Institut für Chemische Energiekonversion, Stiftstraße 34−36, 45470 Mülheim an der Ruhr, Germany ‡ ChemMatCARS, University of Chicago, Argonne, Illinois 60439, United States ∥
S Supporting Information *
ABSTRACT: A series of iron and iron−cobalt bimetallic complexes were isolated: LFe 2 Cl (1), LFe 2 (2), Li(THF)3[LFe2Cl](Li(THF)3[2-Cl]), LFeCoCl (3), and LFeCo (4), where L is a trianionic tris(phosphineamido)amine ligand. As elucidated by single-crystal X-ray diffraction studies and quantum-chemical calculations, the Fe II Fe II and Fe IICo II complexes, 1 and 3, respectively, have weak metal−metal interactions (the metal−metal distances are 2.63 and 2.59 Å, respectively) with a partial bond order of 0.5. The formally mixed-valent complexes, FeIIFeI (3) and FeIICoI (4), have short metal−metal bonds (2.32 and 2.26 Å, respectively) with a formal bond order of 1.5. On the basis of magnetic susceptibility measurements, complexes 1−4 are all paramagnetic with highspin ground states, S = 3−4, which are proposed to arise from ferromagnetic coupling of the two metals’ spins through a direct exchange mechanism. Zero- and applied-field Mössbauer spectra corroborate the presence of distinct oxidation and spin states for the iron sites. The reduction potentials of 1 and 3 are −1.48 and −1.60 V (vs Fc+/Fc), respectively. Other characterization data are also reported for this series of complexes, electronic absorption spectra and anomalous X-ray scattering data.
■
INTRODUCTION Research in molecular magnetism is driven by the desire to understand, control, and/or elicit new magnetic behavior through the rational design of coordination complexes. For polynuclear metal clusters, the magnetic behavior is primarily dictated by the strength of the metal−metal interaction.1 Clusters with strong interactions are typically diamagnetic, or low-spin, while those with weaker interactions can give rise to either ferromagnetic or antiferromagnetic exchange.2,3 Past research efforts have generated fundamental magnetostructural relationships for polynuclear complexes with weakly coupled spins.1 Metal−metal-bonded complexes typically belong to the strong interaction regime and are low-spin.2,3 However, a small but growing number of metal−metal bonded complexes with an approximate single covalent bond between Fe and/or Co centers are high-spin.4−15 In an isostructural series of firstrow bimetallics in a tris(pyridylamido)amine scaffold, the spin ground-state switches, counterintuitively, from low to high spin as the metal−metal interaction changes from weak to covalent.16 The switch was explained by a competition between a superexchange mechanism (through a bridging ligand) versus direct exchange (via a metal−metal bonding interaction), where © XXXX American Chemical Society
the former leads to antiferromagnetic exchange and the latter presumably results in ferromagnetic exchange.16,17 To further understand high-spin bimetallics with metal− metal bonding, we targeted metal−metal bonded complexes featuring iron and/or cobalt in a tris(phosphineamido)amine scaffold, which is referred to as L. Unlike the tris(pyridylamido)amine system, the PCH2N ligand arms of L are saturated, impeding superexchange interactions (via a bridging ligand) but not direct metal−metal exchange. A previously reported high-spin dicobalt complex, LCo2, showed moderate ferromagnetic coupling (2J = 120 cm−1) between high-spin S = 3/2 CoII and S = 1 CoI centers.18 Adding to this dicobalt complex, we present diiron and iron− cobalt analogues where the metals are both divalent, FeIIFeII and FeIICoII, or formally mixed-valent, FeIIFeI and FeIICoI. The diiron and iron−cobalt complexes in this series are all high-spin with strong ferromagnetic coupling (2J ≥ 400 cm−1). The compounds were investigated by single-crystal X-ray diffraction and X-ray anomalous scattering studies, cyclic voltammetry (CV), magnetometry, several spectroscopies [NMR, vis−nearReceived: June 21, 2016
A
DOI: 10.1021/acs.inorgchem.6b01487 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry Table 1. Crystallographic and Refinement Details for Complexes 1−4 chemical formula empirical formula fw cryst syst space group a (Å) b (Å) c (Å) α (deg) β (deg) γ (deg) V (Å3) Z Dcalcd (g cm−3) λ (Å), μ (mm−1) T (K) Θ range (deg) reflns collected unique reflns data/restraint/param R1, wR2 [I > 2σ(I)] R1, wR2 (all data)
1
1a
2
3
4
LFe2Cl C39H60N4P3Fe2Cl, C5H12 897.11 monoclinic P21/c 11.8877(6) 16.2378(8) 23.9950(12) 90 101.0640(10) 90 4545.7(4) 4 1.311 0.71073, 0.837 173(2) 1.52−27.46 52098 7801 10365/7/501 0.0465, 0.1121 0.0675, 0.1258
LFe2Br C39H60N4P3Fe2Br, C5H12 941.57 monoclinic P21/c 11.8696(5) 16.2190(7) 24.3507(11) 90 100.5650(10) 90 4608.4(3) 4 1.357 0.71073, 1.635 173(2) 1.52−26.26 48121 7067 9291/11/476 0.0435, 0.1052 0.0646, 0.1165
LFe2 C39H60N4P3Fe2 789.52 trigonal P321 15.8611(16) 15.8611(16) 11.7814(12) 90 90 120 2566.8(6) 2 1.022a 0.71073, 0.684 173(2) 1.48−26.32 9068 2285 3451/0/150 0.0383, 0.0662 0.0608, 0.0704
LFeCoCl C39H60N4P3FeCoCl, C5H12 900.19 monoclinic P21/c 11.8500(7) 16.2123(10) 23.9996(14) 90 101.4750(10) 90 4518.5(5) 4 1.323 0.71073, 0.889 173(2) 2.16−27.47 51322 7775 10289/5/495 0.0435, 0.1045 0.0627, 0.1153
LFeCo C39H60N4P3FeCo, C4H8O 864.70 monoclinic P21 11.837(2) 15.715(3) 11.986(2) 90 101.343(2) 90 2186.2(8) 2 1.314 0.71073, 0.858 173(2) 2.16−27.48 18111 7996 9688/5/491 0.0515, 0.1284 0.0660, 0.1406
a
This density does not consider the six disordered benzene molecules in the unit cell. The density with these disordered solvent molecules is 1.325 g cm−3. The formula weight of 2·3C6H6 is 1023.89. LFe2Br (1a). The synthetic procedure was identical with that used to prepare 1, except that FeBr2 was used as the metal reagent. Single crystals were obtained by layering pentane onto a concentrated toluene solution. 1H NMR (THF-d8, 500 MHz): δ 108.4, 98.9, 86.0, 49.7, 42.4, 20.8, 16.3, −66.6, −90.0. Anal. Calcd for C39H60N4P3Fe2Br: C, 53.88; H, 6.96; N, 6.44. Found: C, 53.81; H, 7.03; N, 6.35. LFe2 (2). To a THF solution of 1 (212 mg, 0.257 mmol) cooled in a dry ice/acetone cold well bath was added dropwise PhMgCl (2 M in THF, 1 equiv). The reaction was warmed to rt and stirred for 8 h. After filtering through Celite, the filtrate was dried in vacuo to give a residue that was dissolved in benzene and filtered through Celite. The filtrate was again dried in vacuo to give a crude, dark powder (126.8 mg, 60% crude yield). Single crystals were obtained via vapor diffusion of pentane into a concentrated toluene solution. 1H NMR (C6D6, 500 MHz): δ 51.8, 28.1, −62.8. UV−vis−NIR [C6H6; λmax, nm (ε, M−1 cm−1)]: 430 sh (4300), 490 sh (3000), 550 sh (2040), 660 sh (1170), 1040 sh (340). Anal. Calcd for C39H60N4P3Fe2: C, 59.33; H, 7.66; N, 7.10. Found: C, 59.30; H, 7.54; N, 6.97. Li(THF)3[LFe2Cl](Li(THF)3[2-Cl]). To a THF solution of LH3 (538 mg, 0.792 mmol) cooled in a dry ice/acetone cold well bath was added dropwise excess n-BuLi (2.5 M in hexane, 980 uL, 2.45 mmol). The reaction was warmed to rt slowly and stirred for 1 h. The reaction was then frozen in a LN2 cold well bath and, upon thawing, was added to the top of a frozen THF (3 mL) solution of FeCl2(THF)1.5 (370 mg, 1.58 mmol). The reaction was warmed to rt and stirred for 12 h, giving a red solution. The reaction solution was dried in vacuo, and the resulting residue was dissolved in toluene. The solution was filtered through Celite, and the filtrate dried in vacuo to give a red powder (472 mg, 70% yield). Single crystals were obtained via vapor diffusion of pentane into a concentrated THF solution. 1H NMR (THF-d8, 500 MHz): δ 58.3, 28.3, 9.8, −14.9, −61.7. Anal. Calcd for C39H60N4P3Fe2(LiCl(OC4H8)3): C, 58.44; H, 8.08; N, 5.34. Found: C, 58.34; H, 7.98; N, 5.16. LFeCoCl (3). To a THF solution of 1 (444 mg, 0.54 mmol) was added portionwise a THF solution of CoCl2(THF)1.5 (105.5 mg, 0.54 mmol). The reaction color immediately changed to a deep purple. The reaction was stirred at rt for 3 h. After filtering through Celite, the filtrate was dried in vacuo, and the resulting residue was dissolved in toluene and filtered again through Celite. The filtrate was collected,
IR (NIR), and Mössbauer], and theoretical calculations [density functional theory (DFT) and CASSCF/CASPT2], so as to better understand the relationships between the bonding, electronic structure, and properties in this intriguing class of high-spin metal−metal bonded complexes.
■
EXPERIMENTAL SECTION
General Considerations. Unless otherwise stated, all manipulations were performed under a dinitrogen atmosphere in a VAC Atmosphere glovebox. Standard solvents were deoxygenated by sparging with dinitrogen and dried by passing through activated alumina columns of a SG Water solvent purification system. Deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc., degassed via freeze−pump−thaw cycles, dried over activated alumina, and stored over activated 4 Å molecular sieves. All other reagents were purchased from Aldrich or Strem and used without further purification. Elemental analyses were performed by Complete Analysis Laboratories, Inc. (Parsippany, NJ). The ligand 2,2′,2″-tris[diisopropyl(phosphinomethyl)amino]triphenylamine (abbreviated as LH3) was synthesized according to the literature.19 Syntheses. LFe2Cl (1). To a frozen diethyl ether (Et2O) solution of LH3 (538.4 mg, 0.792 mmol) in a LN2 cold well was added slowly nBuLi (2.5 M in hexane, 982 μL, 2.45 mmol). The reaction was warmed to room temperature (rt) and stirred for 30 min to give a pale-yellow solution, which was then frozen in a LN2 cold well. After thawing, the ligand solution was layered on top of a frozen tetrahydrofuran (THF) solution of FeCl2(THF)1.5 (370.5 mg, 1.58 mmol), and then allowed to warm to rt. The reaction was stirred for 4 h, affording a deep-blue color. After removal of the solvent in vacuo, the resulting residue was dissolved in toluene and filtered through a Celite plug. The filtrate was collected and dried in vacuo to provide a deep-blue crystalline powder (471.8 mg, 70% yield). Single crystals were obtained by layering pentane onto a concentrated toluene solution. 1H NMR (C6D6, 500 MHz): δ 118.9, 97.0, 86.1, 56.7, 34.7, 19.7, 10.3, 4.4, −66.8, −86.8. UV−vis−NIR [THF; λmax, nm (ε, M−1 cm−1)]: 520 sh (5200), 635 (6800), 1075 (500). Anal. Calcd for C39H60N4P3Fe2Cl: C, 56.78; H, 7.33; N, 6.79. Found: C, 56.68; H, 7.26; N, 6.72. B
DOI: 10.1021/acs.inorgchem.6b01487 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry concentrated in vacuo, and then layered with pentane to provide purple crystalline blocks (398 mg, 90% yield). 1H NMR (C6D6, 500 MHz): δ 150.7, 88.3, 48.3, 32.8, 15.3, 11.9, −70.7, −86.3. UV−vis− NIR [THF; λmax, nm (ε, M−1 cm−1)]: 510 (5300), 600 sh (4600), 1000 (430), 1500 (350). Anal. Calcd for C39H60N4P3FeCoCl: C, 56.57; H, 7.30; N, 6.77. Found: C, 56.48; H, 7.38; N, 6.63. LFeCo (4). To a THF solution of 3 (110 mg, 0.13 mmol) was added dropwise MeMgCl (3 M in THF, 1 equiv). The reaction color immediately changed to dark brown-orange. The reaction was stirred at rt for 2 h. After filtering through Celite, the solution was dried in vacuo, and the resulting residue was dissolved in benzene and filtered again through Celite. The filtrate was again dried in vacuo to afford a dark powder (98 mg, 90% crude yield). Single crystals were obtained via vapor diffusion of Et2O into a concentrated THF solution. 1H NMR (C6D6, 500 MHz): δ 135.8, 53.3, 26.8, 22.4, 17.0, −4.0, −34.0, −57.7. UV−vis−NIR [THF; λmax, nm (ε, M−1 cm−1)]: 420 (7800), 510 sh (3300), 580 sh (1500), 760 (700), 1520 (100). Anal. Calcd for C39H60N4P3FeCo: C, 59.10; H, 7.63; N, 7.07. Found: C, 59.19; H, 7.68; N, 6.98. X-ray Crystallography and Structure Refinement Details. A blue block (0.30 × 0.25 × 0.20 mm3) of 1, a blue block (0.40 × 0.30 × 0.20 mm3) of 1a, a yellow plate (0.15 × 0.05 × 0.05 mm3) of 2, a purple block (0.30 × 0.25 × 0.20 mm3) of 3, and a brown block (0.25 × 0.20 × 0.20 mm3) of 4 were each placed on the tip of a 0.1-mmdiameter glass capillary and mounted on a Bruker APEX II CCD diffractometer for data collection at 173(2) K. The data collection was carried out using Mo Kα radiation (graphite monochromator).20 The data intensity was corrected for absorption and decay (SADABS). Final cell constants were obtained from least-squares fits of all measured reflections, and the structure was solved using SHELXS-08 and refined using SHELXL-2013.21 A direct-methods solution was calculated and provided most non-H atoms from the E map. Full-matrix leastsquares/difference Fourier cycles were performed to locate the remaining non-H atoms, and all non-H atoms were refined with anisotropic displacement parameters, except when noted. H atoms were placed in ideal positions and refined as riding atoms with relative isotropic displacement parameters. Disordered pentane molecules in 1, 1a, and 3 were refined using SAME and DFIX restraints, with C atoms in the pentane molecule in 1a needing to be refined isotropically. Compound 2 was refined as an inversion twin with a minor component of 6%, and a PLATON SQUEEZE was used to remove the electron density in a channel consistent with six benzene molecules in the unit cell.22 Compound 4 was refined as an inversion twin, resulting in a 46% minor component, and a disordered molecule of THF was refined using SAME restraints. Crystallographic details are shown in Table 1. Anomalous Diffraction Data Collection and Refinement of Fe and Co Occupancies. Single crystals of 3 and 4 were mounted on a glass fiber and cooled to 100 or 230 K, respectively, using an Oxford Instruments Cryojet cryostat. The Bruker D8 diffractometer, integrated with an APEX-II CCD detector, was modified for synchrotron use at the ChemMatCARS 15-ID-B beamline at the Advanced Photon Source (Argonne National Laboratory). Diffraction data were collected at three different energies each with 0.3 s frames with manual attenuation of the beam to minimize overages of individual pixels. The scan at 30.0 keV (λ = 0.41328 Å), which is energetically well above the atomic absorption energies, gave a leastsquares refinement of all model positional and displacement parameters to 0.5 Å resolution. To determine the compositions of Fe/Co at the two independent metal sites, two additional sets of data below the K-edge of Fe and Co were collected. For 3, the data were collected 50 eV below the K-edge (keV [wavelength (Å)]) of Fe (7.062 [1.756]) and Co (7.659 [1.619]), and for 4, the data were collected 25 eV below the K-edge of Fe (7.087 [1.749]) and Co (7.684 [1.614]). The anomalous diffraction can readily distinguish Fe/Co compositions at the two metal sites because of the expected differences in the Δf ′ and Δf ″ values for these two elements, as shown in Figure 1. The Δf ′ and Δf ″ values of an element change dramatically near the element’s absorption edge, but for other element(s), they remain
Figure 1. Anomalous dispersion corrections to normal scattering factors, including the real (Δf ′) and imaginary (Δf ″) components, for Fe (red) and Co (blue) as a function of the wavelength (Å). The dotted lines represent the experimental wavelengths for the anomalous experiments for 3 (gray) and 4 (black), which were selected to be slightly lower than the Fe and Co absorption edge energies. relatively constant. Each of these two anomalous diffraction data sets thus provides a different view of the electrons present at both sites. Two anomalous data sets (λ > λedge) were used to solve for the metal occupancies. (Note: For λ = λedge, the data are less reliable because of inaccuracies in the metal K-edge energies, which shift in coordination compounds. For λ < λedge, potential issues due to adsorption and/or fluorescence make the data less reliable.23) The two anomalous data sets were simultaneously used in a least-squares refinement to determine the Fe/Co occupancies at the two metal sites. GSAS-II was employed for these least-squares refinements because it allows multiple diffraction data sets as an input with subsequent refinement using a common crystallographic model.24 Compound 4 was collected at a higher temperature (230 K) because of its potential phase transition at lower temperatures and was also refined with the twin law [−1, 0, 0; 0, −1, 0; 0, 0, −1]. Physical Measurements. NMR spectra were collected on a Varian Inova 500 MHz spectrophotometer. UV−vis−NIR absorption data were collected on a Cary-14 spectrophotometer. CV was conducted using a CH Instruments 600 electrochemical analyzer. The one-cell setup utilized a platinum working electrode, a platinum wire counter electrode, and a Ag/AgNO3 reference electrode in CH3CN. Analyte solutions were prepared in a 0.4 M NBu4PF6/THF solution and referenced internally to the FeCp2/FeCp2+ redox couple. Mössbauer data were recorded on an alternating constant-acceleration spectrometer. The minimum experimental line width was 0.24 mm s−1 (full width at half-height). The 57Co/Rh source (1.8 GBq) was positioned at rt inside the gap of the magnet system at a zero-field position. Isomer shifts are quoted relative to Fe metal at 300 K. Magnetic Mössbauer spectra were simulated by using a spinHamiltonian description of the electronic ground state of the spincoupled dimers with total spin S: 2 2 2 ̂ Ĥ ZFS = gμB⇀ S ·⇀ B + D[Sẑ − 1/3S(S + 1) + E /D(Sx̂ − Sŷ )]
(1) where D and E/D are the axial and rhombic zero-field parameters of the ground-state manifold. The hyperfine interactions for 57Fe were calculated by using the usual nuclear Hamiltonian. Magnetic susceptibility data were measured from powder samples of solid material in the temperature range 2−300 K by using a SQUID susceptometer with a field of 1.0 T (MPMS-7, Quantum Design, calibrated with a standard palladium reference sample, error
