Theoretical view on a linear end-on manganese–dioxygen complex bearing a calix[4]arene ligand

Xiang-Cui Huanga, Hui-Fang Wang*a and Jian-Ping Lang*ab
aCollege of Chemistry, Chemical Engineering and Materials Science, Soochow University, Suzhou 215123, People's Republic of China. E-mail: wanghf@suda.edu.cn; jplang@suda.edu.cn
bState Key Laboratory of Organometallic Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai 200032, People's Republic of China. E-mail: jplang@suda.edu.cn

Received 30th April 2016 , Accepted 3rd August 2016

First published on 22nd August 2016


Abstract

By taking a Mn–dioxygen complex supported by a zwitterionic calix[4]arene ligand ([MnL(O2)(H2O)](PF6)2; H4L = [5,11,17,23-tetrakis(trimethylammonium)-25,26,27,28-tetrahydroxycalix[4]arene], CSD refcode: CAGVAJ) as a model system, we investigated the geometric and electronic structures of [MnL(O2)(H2O)]2+ through density functional theory (DFT) calculations without and with the presence of additional waters. It was found that the ground-state structure of [MnL(O2)(H2O)]2+ possesses a typical bent end-on binding mode (Mn–O–O angle at around 130° and dioxygen bond length at ∼1.32 Å) within a sextet state (S = 5/2) irrespective of calculated methodologies. The experimently proposed linear end-on structure (Mn–O–O angle at ∼180°) and the side-on structure (Mn–O–O angle ∼75°) are shown to be transition-state structures in the process of dioxygen flipping from one bent-end-on structure to another. By the inclusion of water solvation effect via both the explicit water-cluster model ([MnL(O2)(H2O)]2+nH2O, n = 1, 4, 6, 8) and implicit CPCM solvation model, a good correlation was concluded: the more additional waters and the more contracted Oe square benefit the larger Mn–O–O angle, indicating a water solvation effect and a ligand size effect. Inspired by the relationship of linear end-on mode with bent end-on mode and the consistence of the projected dioxygen bond length of [MnL(O2)(H2O)]2+–8H2O in the CPCM model along the Mn–O bond direction (dO1–O2(p) = 1.240 Å) with the experimental dioxygen bond length (1.249 Å), we concluded that the linear Mn–O–O arrangement fitted by experimentalists should be a result of the flipping motion of dioxygen rotating about the Mn–O bond in solid-state structures of the [MnL(O2)(H2O)](PF6)2 complex based on analyses of the crystallographic data. Natural bond orbital (NBO) analyses show the sextet [MnL(O2)(H2O)]2+nH2O (n = 0, 1, 4, 6, 8) have a Mn(III)-superoxo nature, consistent with experimental observations. The differences in the bonding structures of the three binding modes were discussed by the second-order perturbation energies ΔEij from the superoxo lone pair (LP) orbitals to Mn(III) 3d orbitals in β electron space. It can also give clues about the instability of the linear end-on binding mode and the increased Mn–O–O angle in the presence of water solvents.


Introduction

Due to the great interest in the mechanistic understanding of many enzymatic and biomimetic oxidation reactions,1 mononuclear metal–dioxygen (M–O2) complexes have been extensively synthesized and examined with various spectroscopic methods including X-ray crystallography.2 Most of the synthetic M–O2 complexes were reported as metal-superoxo, -peroxo and -hydroperoxo species, adopting either a η1 bent end-on binding mode or η2 side-on binding mode.3 In the η1 bent end-on binding mode, dioxygen coordinates to the central metal on one end with a M–O–O bond angle in a range of 121–146°;4–8 In the η2 side-on binding mode, dioxygen binds the central metal with two O atoms, featuring symmetric M–O bond lengths and M–O–O angles smaller than 80°.9–12 Different binding modes of dioxygen usually lead to different electronic structures of metal–O2 adducts. For example, the η1 end-on copper(II)-superoxo species are generally with a triplet ground-state while the η2 side-on copper(II)-superoxo are generally with a singlet ground-state.13 Two isomers of [Ni(13-TMC)O2]+ (TMC = macrocyclic N-tetramethylated cyclam) are demonstrated to have, respectively, a superoxo ligand bound to nickel(II) center in η1 end-on fashion ([NiII(13-TMC)O2]+) and a peroxo ligand bound to nickel(III) center in η2 side-on fashion ([NiIII(13-TMC)O2]+).14 More cases can be found elsewhere.15 Besides, the η1 end-on M–O2 complexes are frequently proposed as active intermediates in catalytic oxidative reactions, seemingly more reactive than their side-on analogs. For example, an end-on binding copper–dioxygen was identified as the active site of peptidyl α-hydroxylating monooxygenase (PHM) by X-ray crystallography,16 while side-on copper–dioxygen adducts with β-diketiminate and anilido–imine ligands were shown to have poor oxidizing power.17 Annaraj et al. demonstrated that binding of an axial ligand trans to the peroxo group of [MnIII(13-TMC)(O2)]+ complex may facilitate conversion of the side-on peroxo ligand into an end-on peroxo ligand, thus making it more reactive toward nucleophilic reactions.12 Also, side-on non-heme peroxoiron(III) centers were reviewed to be relatively inactive as compared to the end-on peroxoiron(III) complexes.18 Therefore, a more end-on geometric Mn–O2 complex was targeted by Geiger and coworkers and a complete conversion from a side-on ligated dioxygen to its end-on analog was expected through supporting ligands perturbations.19 However, fundamental studies on the transformation from side-on bound dioxygen ligands to end-on ones (or conversely) are still devoid because of the limited thermal stabilities of dioxygen complexes.20

Worth mentioning, a confusing binding mode of dioxygen was reported in a Mn–dioxygen complex [MnL(O2)(H2O)](PF6)2 (H4L = [5,11,17,23-tetrakis(trimethylammonium)-25,26,27,28-tetrahydroxycalix[4]arene]) by Liu et al. recently.21 It shows a linear Mn–O–O unit (Mn–O–O angle is 180°) enclosed by a NMe3+-substituted calix[4]arene ligand, and the distal O atom is surrounded by several water molecules distributing at top of the ‘bowl’ (Fig. 1). The CIF of this complex is enclosed as ESI and the view of its cationic structure [MnL(O2)(H2O)]2+ with 50% thermal ellipsoids is shown in Fig. S1 (ESI). Liu et al. assigned it as a Mn(III)-superoxo complex due to the typical signal for a superoxo radical observed in the electron paramagnetic resonance (EPR) spectrum of [MnL(O2)(H2O)](PF6)2 (a rhombic signal with g values of 2.0037 (293 K) and 2.0036 (110 K)). As far as we knew, such a linear coordination mode was seldom found in other structurally-defined dioxygen complexes except in some theoretical models.22,23 Most of the reported Mn–dioxygen complexes are established as side-on MnIII/IV-peroxo species24–26 and a few examples are found as η1 bent end-on hydroperoxo (Mn(III)–OOH27) and alkylperoxo (Mn(III)–OOR28) compounds. Few Mn-superoxo complex has been reported. Thirdly, M–O2 adducts are frequently supported by tetradentate nitrogenous ligands like porphyrin ligands,29 aminopyridyl ligands,30 and TMC ligands,3,31 but very few with calix[n]arene ligands. Thus, it intrigues us to investigate the geometric and electronic structures of [MnL(O2)(H2O)]2+ from density functional theory (DFT) calculations.


image file: c6ra11199h-f1.tif
Fig. 1 Experimentally determined structure of [MnL(O2)(H2O)]2+ from X-ray crystallography. Hydrogen atoms except those of waters are omitted for clarity. There are 8 water solvent molecules averaged in an asymmetric unit of the solid-state structure of [MnL(O2)(H2O)](PF6)2 and only four nearest waters are shown here. Distances for some key bonds and H-bonds are shown in Å. Color code here and after: magenta, Mn; red, O; blue, N; grey, C; white, H.

In the past years, DFT calculations have given reliable descriptions about geometric and electronic structures of synthesized metal–O2 complexes, making good supplementary for spectroscopic observations.19,32,33 Schatz et al.34 and Donoghue et al.33 detected copper(II)-superoxo complexes with various spectroscopic methods including UV/vis, EPR and nuclear magnetic resonance (NMR) spectroscopy, and determined the dioxygen bound in η1 bent end-on fashions based on DFT calculations. Zapata-Rivera et al.35,36 calculated the potential energy surfaces (PESs) involved in the complete intermolecular O2 transfer between TMC ligated Ni(II) and Mn(III) complexes. The calculated results supported the experimentally proposed two-step mechanism and the formation of intermediate was in agreement with kinetic studies. Therefore, in the present work, aimed at clarifying the most favourable binding mode of [MnL(O2)(H2O)]2+, different configurations of [MnL(O2)(H2O)]2+ have been investigated in combination of three different binding modes (bent end-on, linear end-on and side-on binding modes) with four different overall spin states (S = 7/2, 5/2, 3/2, 1/2) using DFT calculations. Water solvation effect was considered by combining an explicit water-cluster model and an implicit water solvation model. Influences of differently sized calix[4]arene ligands were explored and the Mn–O2 bond electronic structure was identified from NBO calculations. Through our calculations, a reasonable understanding on the experimentally fitted linear Mn–O–O unit was achieved and the water solvation effect and ligand size effect on Mn–O–O angle were clarified. It would shed light on controllable synthesis of more bent or more end-on Mn–dioxygen complexes through solvent polarity perturbations and ligand size adjustment.

Computational details

Based on the bowl-shaped framework built from crystallographic data of [MnL(O2)(H2O)](PF6)2, different binding modes of dioxygen were tried as starting geometries during our optimizations, including bent end-on mode (∠MnOO was initially set around 125°), linear end-on mode (∠MnOO was initially set at 180°) and side-on mode (featuring symmetric Mn–O bonds). Spin-unrestricted DFT calculations were performed by Gaussian 09.37 Seven functionals were assessed in the first part of this study, including a pure GGA functional, BP86, which is capable of producing very good geometries for the cobalt– and nickel–dioxygen complexes;14,32 three well-known hybrid functionals including B3LYP,38 M06-2X,39 TPSSh;40 two functionals including empirical dispersion corrections like B97-D41 and wB97X-D42 and an empirical dispersion corrected functional B3LYP-D3.43 Benchmarked by the experimental structural parameters, a reliable functional was chosen as our target functional. Applied with the chosen functional, at least 8 different basis set levels were then estimated. For example, B1 refers to a mixed basis set for whole system: polarized triple-ζ basis sets 6-311+G(d)44 for Mn, O atoms and polarized double-ζ basis sets 6-31G(d)45 for C, N, H atoms. B2 refers to 6-311+G(d) for Mn, O atoms and 6-31G(d,p) for other atoms, and B8 refers to 6-311+G(d) basis set for Mn, O atoms, 6-31G(d,p) basis set for H of the additional waters and 6-31G(d) basis set for the rest of atoms in the system. All the other basis set levels can be found in ESI. The calculated results are elaborated below.

Employing the most suitable functional and basis set level, we then compared energies and structures of [MnL(H2O)(O2)]2+ within different binding modes and different spin states. All possible spin states were considered for each binding mode, including octet state (S = 7/2), sextet (S = 5/2), quartet (S = 3/2) and doublet (S = 1/2). Optimization of octet state shows O2 flies away from Mn center with the nearest Mn–O distance more than 4.5 Å, implying hardly any ferromagnetically coupling interaction between the triplet dioxygen and sextet Mn(II). Based on the most stable configuration, we further studied the water solvation effect on the binding mode of [MnL(H2O)(O2)]2+ by combining an explicit water cluster model and an implicit conductor-like polarizable continuum model (CPCM) encoded in Gaussian software. The calculated results show that inclusion of the explicit water solvent molecules has significant influences on Mn–O–O bond angle. The details are elaborated in the following sections.

None structural or symmetry constraints were imposed during optimizations. Most structures were converged at default convergences, while the linear end-on configurations were located at tight convergences within ultrafine integration grids (grid = 99[thin space (1/6-em)]770). Analytical frequencies were calculated to verify local minima or transition states. Vibration frequencies reported here are scaled by 0.946.46 In some cases, different structures have been located for the same binding mode within the same spin state, e.g., the bent end-on mode within S = 5/2 has more than one equilibrium structures (like S1 and S2 shown in Fig. S2 in this work), only structures with the lowest electronic energies were discussed in this work. NBO calculations47 were done using NBO 3.0. NBO orbital contour plot were generated by Gview 5.0.

Results and discussions

Assessment of functionals and basis set levels

Performances of different functionals and basis set levels on geometry of sextet [MnL(O2)(H2O)]2+ were demonstrated in this part (the 1H NMR result show [MnL(O2)(H2O)](PF6)2 complex has a spin state of S = 5/2).21 All the calculated structural parameters under different methods are summarized in ESI (Tables S1 and S2). In Fig. 2a we show the optimized structure of sextet [MnL(O2)(H2O)]2+ under B3LYP/B1 method. We labelled the distal and middle O atoms of dioxygen ligand as O1 and O2, respectively. O atoms of the calix[4]arene ligand are labeled as Oe since they are on a mean equatorial plane. According to our calculations (Table S1), All tested functionals predict a similar bent end-on structure for sextet [MnL(H2O)(O2)]2+ irrespective of starting geometries. The converged bent end-on structure features Mn–O2–O1 angle less than 130° and dioxygen bond length (dO1–O2) around 1.30–1.32 Å. Besides, wB97X-D and B3LYP produce the largest Mn–O2–O1 angles; B3LYP-D3 gives better bond-length descriptions with the smallest root-mean-square deviation (RMSD) of distances as relative to the experimental parameters. Still, we chose B3LYP as our target functional because of its less computational cost than wB97X-D functional and the predicted larger bond angles than B3LYP-D3 functional. B3LYP has been also evaluated as a reliable functional to predict structural, electronic and magnetic properties of synthetic oxomanganese complexes48,49 and were frequently used in theoretical calculations on C–H bond activation mechanisms by non-heme MnIII/IV = O complexes.50,51 Calculations under different basis set levels show, similarly, that the sextet [MnL(H2O)(O2)]2+ converged to a typical bent end-on structure with Mn–O2–O1 angle around 130° (Table S2). Otherwise, we found that basis sets of Mn, using either Pople-type basis set 6-311+G(d) or LanL2DZ basis set with effective core potentials, give comparable geometrical parameters (differences within ∼0.02 Å and ∼0.3°) (see Table S2, B1 vs. B3; B6 vs. B7). Reducing O atom basis set from 6-311+G(d) to 6-31G(d) results in even smaller Mn–O2–O1 angles and increased RMSDs of distances (see Table S2, B1 vs. B7). Diffusions on the C, N atoms and polarizations on the H atoms have marginal effect on the optimized parameters (see Table S2, B1 vs. B2, B5). Reduction basis sets of Oe and Ow (O atoms of the axial water ligand) to a double-zeta 6-31G(d) results in a more deviated O2–Mn–Ow angle (see Table S2, B1 vs. B4). Therefore, we chose B1 as our favourable basis set level.
image file: c6ra11199h-f2.tif
Fig. 2 Optimized structures of [MnL(H2O)(O2)]2+ in (a) bent end-on, (b) linear end-on and (c) side-on configurations. Only structural parameters of S = 5/2 are shown here and those of the quartet and doublet can be found in Table 1.

Comparisons of different binding modes within different spin states

Using B3LYP/B1 method, we located 8 configurations of [MnL(H2O)(O2)]2+ as summarized in Table 1, including the bent end-on, linear end-on and side-on modes in combination with the overall spin state of S = 5/2, 3/2 and 1/2 (the octet state are excluded here as described in Computational details). The bent end-on configurations were located by optimizing to local minima, while the linear end-on and side-on configurations were located by optimizing to transition states with Berny algorithm using GEDIIS in redundant internal algorithm.52 Frequency analyses show that bent end-on structures are true local minima with none imaginary frequencies while linear end-on and side-on structures feature single imaginary frequencies (vimg in Table 1). The calculated Gibbs free energies (ΔG) as relative to the sextet bent end-on configuration show the bent end-on structures are more stable than the linear end-on and side-on structures within the same spin state. This is different from the Mn(III)-peroxo complexes supported by N-donor ligands, which have been shown the peroxo ligand in side-on fashion is thermodynamically more stable than the bent end-on fashion.10,30 The optimized bent end-on, linear end-on and side-on structures can be found in Fig. 2. Seen from Table 1 and Fig. 2, the bent end-on structures possess asymmetric Mn–O1/Mn–O2 distances and Mn–O2–O1 angles are around 130°; the linear end-on structures feature almost linear Mn–O2–O1 angles (∠MnO2O1 ≈ 180°); the side-on structures possess nearly equal Mn–O1/Mn–O2 distances and Mn–O2–O1 angles are less than 80°. The calculated dO1–O2 are generally ∼1.32 Å regardless of the binding modes and spin states. Besides, ΔG in Table 1 indicate that within the bent end-on binding mode, the sextet state (S = 5/2) is the most stable with the lowest free energy, and the doublet state (S = 1/2) is the least stable. This result is consistent with the obtained magnetic moment of [MnL(O2)(H2O)](PF6)2 (5.8 μB) by the temperature-dependent 1H NMR spectroscopic method of Evans, which showed an overall spin state of S = 5/2.21 Moreover, the calculated stretching frequency of the dioxygen ligand (vO1–O2 in Table 1) for sextet [MnL(H2O)(O2)]2+ (1131.6 cm−1) is closer to the resonance Raman (rR) result (1124 cm−1)21 as compared to the quartet (S = 3/2, 1136.9 cm−1) and doublet states (1143.0 cm−1). The free energy difference of the fully optimized bent end-on structure in quartet and sextet states is, respectively, 1.80 kcal mol−1 under B3LYP/B1 method, 3.14 kcal mol−1 under M06-2X/B1 method, 22.71 kcal mol−1 under wB97X-D/B1 method and 2.98 kcal mol−1 under B3LYP-D3/B1 method, validating the conclusion that [MnL(H2O)(O2)]2+ has a sextet ground state. However, this free energy difference (quartet − sextet) is disputable within BP86/B1 method (−7.93 kcal mol−1), which was used to predict structures of peroxomanganese(III) complexes bearing amino-pyridine ligands.30
Table 1 Relative Gibbs free energies at 298.15 K (ΔG, kcal mol−1), imaginary frequencies (vimg), stretching vibration frequencies of O1–O2 bond (vO1–O2) and selected structural parameters of three binding configurations within different spin states. The experimentally structural parameters are listed in parentheses for comparison
  Bent end-on Linear end-ona Side-on
S = 5/2 S = 3/2 S = 1/2 S = 5/2 S = 3/2b S = 5/2 S = 3/2 S = 1/2
a Location of the doublet linear end-on configuration failed and the predicated relative energy should be higher than 16.6 kcal mol−1.b The structure is converged at default convergences.c dMn–Oe is the averaged distances of dMn–Oe1, dMn–Oe2, dMn–Oe3, dMn–Oe4.d dOe–Oe is the averaged distances of Oe1–Oe3 and Oe2-Oe4, which is considered as a size measurement of the Oe square at the lower rim of the calix[4]arene cone and will be elaborated in main text.
ΔG 0.00 1.80 15.6 4.2 4.7 9.5 14.2 40.4
vimg/cm−1 None None None 153.7i 153.7i, 146.4i 186.6i 263.3i 317.7i
vO1–O2/cm−1 1131.6(1124.0) 1136.9 1143.0 1134.5 1160.9 1133.1 1130.8 1085.2
∠MnO2O1 129.3°(180.0°) 131.7° 125.0° 178.9° 179.5° 75.0° 75.2° 79.7°
∠O2MnOw 170.6°(180.0°) 170.3° 172.7° 173.4° 173.5° 168.2° 169.5° 164.5°
dO1–O2 1.321(1.249) 1.322 1.307 1.320 1.318 1.321 1.325 1.337
dMn–O1 3.248(3.693) 3.201 2.875 3.492 3.426 2.467 2.572 2.208
dMn–O2 2.245(2.444) 2.165 1.919 2.173 2.108 2.454 2.569 2.013
dMn–Ow 2.335(2.210) 2.328 2.081 2.298 2.308 2.325 2.286 2.104
dMn–Oec 1.928(1.887) 1.933 1.935 1.934 1.939 1.937 1.926 1.941
dOe–Oed 3.858(3.773) 3.864 3.865 3.867 3.874 3.859 3.848 3.856


The vibration modes of the imaginary frequencies for the linear end-on and side-on configurations are indicated by the green arrows in Fig. 2b and c. They correspond to the motion of dioxygen ligand over Mn center, either flipping between right and left or swing up and down, respectively. It indicates that the linear end-on and side-on configurations should be transients during the dioxygen flipping from one equilibrium bent end-on configuration to another. To confirm this assumption, we performed optimizations on the linear end-on and side-on configurations by exerting very small displacements along the vibration mode of the imaginary frequency. It was found both the linear end-on configuration and side-on configuration thermodynamically converged into two bent end-on structures. The relevance of these three binding structures can be described by PES as shown in ESI (Fig. S2).

Water solvation effect and theoretical view on the experimentally fitted linear end-on unit

As a summary of above results, the equilibrium structure of [MnL(H2O)(O2)]2+ should possess a Mn–O2–O1 angle around 130° and dioxygen bond length of ∼1.32 Å irrespective of the calculation methodologies and the overall spin states of system, assigning a typical bent end-on binding mode. This result suggests the NMe3+ groups at the upper rim of calix[4]arene ligand have little effect on the binding mode of the dioxygen ligand, which was thought as one of the factors leading to the linear Mn–O–O arrangement in the previous experimental work.21 The possible reason can attribute to the long distances of N⋯O1 (5.26 Å in the X-ray structure and longer than 5.16 Å in the optimized structures). So, what are the exact factors leading to a more end-on Mn–O2–O1 angle and how can the experimentally fitted linear end-on mode be interpreted?

To answer these questions, we investigated water solvation effect on the equilibrium structure of [MnL(O2)(H2O)]2+. Based on the optimized bent end-on [MnL(O2)(H2O)]2+ complex of S = 5/2, we progressively added additional waters to solvate the dioxygen moiety in the calix[4]arene ‘bowl’, building a set of solvated complexes [MnL(O2)(H2O)]2+nH2O (n = 1, 4, 6, 8). As starting structures for each [MnL(O2)(H2O)]2+nH2O complex, the additional waters were intentionally arranged in different patterns with the most hydrogen bonds being formed. Then they were optimized to local minima. Only structure with the lowest electronic energy was chosen as our target structure, which was found to possess the shortest hydrogen bonds between dioxygen moiety and additional waters. The optimized [MnL(O2)(H2O)]2+nH2O structures are listed in Fig. 3, as well as some key structural parameters and H-bonding distances. The other structural parameters are given in ESI (Table S3). It is seen that the distal O1 atom is hydrogen-bonded to additional waters with H-bonding distances in a range of 1.61–2.07 Å. In MnL(O2)(H2O)]2+–8H2O complex (Fig. 3d), the explicit additional waters tend to form a water-cubic cluster by H-bonding to each other. With the number of additional waters increasing from 1 to 8, the Mn–O2–O1 angle (∠MnO2O1) increases gradually from 132.1° to 162.6°, getting closer to a linear angle; the Mn–O2 distance elongates from 2.366 Å to 2.552 Å, becoming less deviated from the experimental distance (2.444 Å) than that of n = 0 (2.245 Å). The dOe–Oe decreases from 3.853 Å to 3.826 Å, getting nearer to the X-ray distance (3.773 Å). All these improved bond distances finally lead to reduced RMSDs of [MnL(O2)(H2O)]2+–8H2O (0.066 in Table S3) as compared to that of [MnL(O2)(H2O)]2+ (0.118 in Table S1). It implies that additional waters are necessary in describing the structure of [MnL(O2)(H2O)]2+ from a theoretical model. Besides, the shortest H-bonding distance in each [MnL(O2)(H2O)]2+nH2O complex is generally located at the back face of the O1 atom (Fig. 3). It should be the right reason for the enlarged Mn–O2–O1 angle. Attempts with even more additional waters (e.g., n = 10 or 12) have been done, but none equilibrium structures with Mn–O2–O1 angle larger than 162.6° were located. It indicates that a linear Mn–O2 unit could not be anticipated even if more additional waters are included. Otherwise, the optimized structure of [MnL(O2)(H2O)]2+–8H2O at B3LYP-D3/B1 level (Table S3) show that the empirical dispersion correction benefit good bond length description (RMSD = 0.045) as expected,43 while predicts more bent bond angles than B3LYP functional (e.g. 156.4° vs. 162.2° for Mn–O2–O1 angle, and 172.9° vs. 174.9° for O2–Mn–Ow angle).


image file: c6ra11199h-f3.tif
Fig. 3 Optimized structures of sextet [MnL(O2)(H2O)]2+nH2O (n = 1, 4, 6, 8) complexes under B3LYP/B1 method. The H-bonds are shown in light-blue dot lines. The key structural parameters are labeled in bigger-size numbers and main H-bond distances (in Å) in smaller-size numbers. Data in parentheses correspond to results at B3LYP/B2 level in CPCM model (see B3LYP/B2* in Table S4). dO1–O2(p) is the projected distances of dO1–O2 along the Mn–O2 bond direction.

Subsequently, we investigated the basis set effect and the bulk water solvation effect on geometry of [MnL(O2)(H2O)]2+–8H2O. We re-optimized structure of [MnL(O2)(H2O)]2+–8H2O under different basis set levels and in an implicit water solvation model, CPCM model, respectively. The calculated results are shown in ESI (Table S4). It shows that polarizations on H atoms, especially on those of the additional waters, benefit a larger Mn–O2–O1 angle (e.g. the Mn–O2–O1 angle of [MnL(O2)(H2O)]2+–8H2O increases to 164.6° under B3LYP/B2 method and 168.6° under B3LYP/B8 method as in Table S4). It suggests that the polarization effect of additional waters is important on the bonding angle of Mn–O2 unit. While, the bulk solvation effect via CPCM model, makes the Mn–O2–O1 angle slightly reduced, e.g., the Mn–O2–O1 angle of [MnL(O2)(H2O)]2+–8H2O decreases from 162.6° in vacuum (B3LYP/B1) to 159.6° in CPCM model (B3LYP/B2*) (Fig. 3d and Table S4). Whereas, the Mn–O2 distance in CPCM model (2.467 Å under B3LYP/B2* level in Table S4) gets more closed to the experimental distance (2.444 Å), resulting in a smaller RMSD (0.073 in Table S4). However, the calculated Mn–O2–O1 angle (∼160°) and dO1–O2 (∼1.32 Å) are still the two most deviated parameters from the X-ray results (180.0° and 1.249 Å, respectively). While, such deviations can be easily reconciled if we take into consideration the projected distances of dO1–O2 (dO1–O2(p) in Fig. 3) along the Mn–O2 bond directions (the Mn–O2 bond is nearly vertical to the mean equatorial plane of Oe square). We noticed that within CPCM model (B3LYP/B2*), the dO1–O2(p) of [MnL(O2)(H2O)]2+–8H2O is 1.240 Å, approximately equal to the X-ray distance of dioxygen bond (1.249 Å). The other key bond lengths like dMn–O2 and dMn–Oe under B3LYP/B2* method also coincide well with the crystallographic results with deviation less than 0.05 Å (see data in parentheses in Fig. 3 and Table S4). Such good consistencies suggest two possibilities: (i) the proposed linear Mn–O2 unit in [MnL(O2)(H2O)](PF6)2 complex would result from superposition of disordered Mn–O2 moieties rotating about the Mn–O2 bond in equilibrium bent end-on configuration, similar to the observed dioxygen disorder in solid-state structures of cobalt–dioxygen and iron–dioxygen complexes ligated by porphrins.53–55 Li and co-workers have treated such disorder as the dynamics of dioxygen rotation around the M–O bond based on detailed analyses about the solid-state structures of oxycobalt and oxyiron picket fence porphyrins under different temperatures.5,55 (ii) When the diffraction data of [MnL(O2)(H2O)](PF6)2 were collected, the dioxygen ligand would be in a constant flipping motion around the Mn–O bond as indicated by the vibration mode of the imaginary frequency of the linear end-on structure shown in Fig. 2b. The resulting averaged electron density of O atoms would then misguide experimentalists with a linear Mn–O2–O1 arrangement as Liu et al. made.21 By examination on the crystallographic data of [MnL(O2)(H2O)](PF6)2, we inclined to the latter possibility. The proofs are: (1) the O1, O2, Mn and Ow atoms are ordinary thermal ellipsoids, with little sign of normal disorder (Fig. S1, ESI), even if the crystallographic data were collected at −50 °C; (2) the thermal factors of O1, O2, Mn and Ow atoms are, respectively, 0.052(2) e/Å2, 0.045(2) e/Å2, 0.0433(5) e/Å2 and 0.043(2) e/Å2. They are generally larger than those of Oe atoms (0.0296(8) e/Å2) and C atoms of phenyl rings (0.016–0.026 e/Å2) (see CIF, ESI), implying an amplitude librational motion of these atoms around their equilibrium positions. Also, the larger thermal factor of O1 atom than that of O2 atom indicates a larger amplitude librational motion of O1 than O2. (3) The calculated flipping barrier between the bent end-on structure and the liner end-on structure would be not more than 4.20 kcal mol−1 as indicated by the PES in Fig. S2 (ESI) and thus can be easily conquered at −50 °C where the crystallographic data were collected.

Ligand size effect and implications on Mn–O2–O1 angle modulation

Based on the optimized structures of the sextet bent end-on [MnL(O2)(H2O)]2+nH2O (n = 0, 1, 4, 6, 8) complexes as shown in Table 1 and Fig. 3, we found two linear dependences on the number of additional waters n, namely, ∠MnO2O1 ∼ n linear dependence and dOe–Oen. linear dependence. Two linear relationships are exhibited in Fig. 4. It shows that a larger n corresponds to a larger Mn–O2–O1 angle (red square line in Fig. 4, R2 = 0.968) and a shorter dOe–Oe (black triangle line in Fig. 4, R2 = 0.967), suggesting a correlation between the shrinking Oe square and the increasing Mn–O2–O1 angle. We also plotted the Mn–O2–O1 angle with the diagonal distance of the CN square (CN denote carbon atoms bonding with NMe3+ groups) at the upper rim of calix[4]arene cone. It was found that the larger CN square corresponds to the larger Mn–O2–O1 (Fig. S3). These linear dependences were not found in doublet and quartet [MnL(O2)(H2O)]2+nH2O complexes.
image file: c6ra11199h-f4.tif
Fig. 4 Linear relationships of Mn–O2–O1 bonding angles (∠MnO2O1) (squares, red line) and dOe–Oe (triangles, black line) of [MnL(O2)(H2O)]2+nH2O (n = 0, 1, 4, 6, 8) complexes with the number of additional waters (n). Insets shown are the structural variation of [MnO2]2+ cores depending on n (distances in Å), atoms of the equatorial ligand and additional water molecules are omitted for clarity.

Presumably, the hydrophobic repulsion between the additional waters and the phenyl rings would expand the upper rim of the calix[4]arene (CN square), thus result in a contracted Oe square at the lower rim due to the rigidity of phenyl rings. Such a contracted Oe square would finally lead to a increased Mn–O2–O1 angle because of the constriction effect of the ‘bowl’ bottom on the Mn–O2 unit. Therefore, the more additional waters would induce the stronger hydrophobic repulsion, thus the more expanded CN squares and more contracted Oe squares, finally leading to the larger Mn–O2–O1 angles. To certify this supposition, we modified the calix[4]arene ligand of [MnL(O2)(H2O)]2+ by substituting the NMe3+ groups with four more bulky groups like NEt3+ (Et = ethyl) and N(t-Bu)3+ (t-Bu = tert-butyl), producing two complexes [MnLEt(O2)(H2O)]2+ (NEt3+-appended complex) and [MnLt-Bu(O2)(H2O)]2+ (N(t-Bu)3+-appended complex). Evidenced from the enlarged dCN–CN and reduced dOe–Oe in [MnLEt(O2)(H2O)]2+ and [MnLt-Bu(O2)(H2O)]2+ complexes as tabulated in Table 2, we see the more bulky groups at the upper-rim of the cone directly lead to more expanded upper rims and thus more contracted lower rims. As expected, the larger Mn–O2–O1 angles are obtained in [MnLEt(O2)(H2O)]2+ and [MnLt–Bu(O2)(H2O)]2+ as shown in Table 2 (131.1° and 133.4°, respectively). Additionally, we directly enlarged the Oe square (dOe–Oe) of the NMe3+-substituted calix[4]arene (i.e., [MnL(O2)(H2O)]2+) by replacing the four connecting methylene (–CH2–) groups with four –S– groups ([MnL−S–(O2)(H2O)]2+ in Table 2 and Fig. S4). Then we got a decreased Mn–O2–O1 angle at 120.5° expectedly. Therefore, we can conclude that calix[4]arene ligand with a shrinked bottom benefits a more end-on Mn–O2 unit and that with an expanded bottom favors a more bent Mn–O2 unit, indicating a ligand size effect. The supporting ligand ring-size effect has been also reported by Nam and coworkers on the metal dioxygen complexes supported by TMC ligands.3,14 Whereas, such a ligand size effect is less important than the water solvation effect on modulating Mn–O2–O1 angle. The evident proof can be found in Table 2. From [MnL−S–(O2)(H2O)]2+ complex to [MnLt-Bu(O2)(H2O)]2+ complex, the dOe–Oe decreases by ∼0.13 Å (from 3.975 Å to 3.848 Å) and the Mn–O2–O1 angle increases by ∼13° (from 120.5° to 133.4°). By contrast, from [MnL(O2)(H2O)]2+ to [MnL(O2)(H2O)]2+–8H2O, the dOe–Oe decreases by only 0.05 Å (from ∼3.87 Å to ∼3.82 Å, see Fig. 2 and 3), the Mn–O2–O1 angle increases by about 33.3° (129.3° vs. 162.6°).

Table 2 Calculated dCN–CN, dOe–Oe and Mn–O2–O1 angle (∠MnO2O1) of modified [MnLR(O2)(H2O)]2+ complexes with the NMe3+-appended, NEt3+-appended, N(t-Bu)3+-appended and –S– connecting calix[4]arene ligands
  NMe3+-appended a NEt3+-appended N(t-Bu)3+-appended –S– connecting
a Calculated structural parameters of [MnL(O2)(H2O)]2+ at B3LYP/B1 level.b One of the dCN–CN in NEt3+-substituted calix[4]arene is longer than NMe3+-substituted calix[4]arene.
dCN–CN 8.922 8.916b 9.164 9.691
dOe–Oe 3.858 3.854 3.848 3.975
∠MnO2O1 129.3° 131.1° 133.4° 120.5°


Would any other polar solvents benefit a larger Mn–O2–O1 angle? To answer this question, we further optimized the structures of [MnL(O2)(H2O)]2+ and [MnL(O2)(H2O)]2+–8H2O in CPCM model with different solvents. The chosen solvents include water (ε = 78.36), methanol (ε = 32.61), chloroform (ε = 4.71) and benzene (ε = 2.27). The calculated Mn–O2–O1 angles and dOe–Oe are exhibited in Table 3 and the Cartesian coordinates of the optimized structures in different solvents are enclosed in ESI. From Table 3, we can see that when the polarity of the solvent (indicated by the dielectric constant ε) increases in the order: vacuum < benzene < chloroform < methanol < water, the Mn–O2–O1 angle as well as the dOe–Oe of [MnL(O2)(H2O)]2+ in solvents increase, suggesting the more polarized solvent favors larger Mn–O2–O1 angle. It is consistent with the suggestion that the bulk polarization of the solvent can affect the binding modes of metal–O2 complexes.56 On the contrary, the Mn–O2–O1 angle of [MnL(O2)(H2O)]2+–8H2O complex increases as the solvent polarity decreases (Table 3). Such a result can be explained by the deeper insertion of the water-cubic cluster into the calix[4]arene bowl in a less polarized solvent (e.g. chloroform and benzene) due to the strengthened water-phobic interaction between the additional water-cluster and the implicit water-insoluble solvents. The evidence is that the vertical distances from the Mn center to the water-cubic cluster decrease in the following order: water > methanol > chloroform > benzene > vacuum (see the Cartesian coordinates listed in ESI). The deeper insertion of the water-cubic cluster in a less polarized solvent is assumed to make a larger steric effect with the calix[4]arene cone, thus the reduced dOe–Oe and increased Mn–O2–O1 angle (Table 3). It confirms that the ligand size effect induced by the explicit additional water molecules is maintained in non-water solvents. Moreover, the deep insertion of the water-cubic cluster also leads to enhanced hydrogen bonding interaction between the dioxygen ligand with additional water molecules in [MnL(O2)(H2O)]2+–8H2O complex in non-water solvents evidenced from the shortened H-bond distances (see Cartesian coordinates in ESI). All above results indicate that the bending degree of the M–O2 unit in metal–dioxygen complexes can be adjusted through solvent polarity perturbations and ligand size modulations.

Table 3 Calculated Mn–O2–O1 angles and dOe–Oe of [MnL(O2)(H2O)]2+ and [MnL(O2)(H2O)]2+–8H2O in different solvents within CPCM scheme
Solvent (ε) [MnL(O2)(H2O)]2+ [MnL(O2)(H2O)]2+–8H2O
∠MnO2O1 dOe–Oe ∠MnO2O1 dOe–Oe
Water (ε = 78.36) 133.3° 3.874 153.1° 3.843
Methanol (ε = 32.61) 133.1° 3.872 153.5° 3.843
Chloroform (ε = 4.71) 132.7° 3.865 155.1° 3.840
Benzene (ε = 2.27) 131.8° 3.859 158.6° 3.834
Vacuum (ε = 1) 129.3° 3.858 162.6° 3.826


Electronic structures of [MnL(O2)(H2O)]2+nH2O

Here we are at a position to examine the electronic structure of Mn–dioxygen bond in [MnL(O2)(H2O)]2+ with and without the presence of additional waters. A preliminary understanding can be obtained by NBOs analyses, although a detailed analysis would require the use of multiconfigurational methods.15,36,57 The natural populations (Q/e) on Mn–O2 unit are listed in Table 4 for all [MnL(O2)(H2O)]2+nH2O (n = 0, 1, 4, 6, 8) complexes. It shows there are about 0.69–076e negative charge on dioxygen ligand (QO1 + QO2) and 1.15–1.24e positive charge on Mn(II) center, suggesting a net charge transfer from Mn(II) to dioxygen and leading to a Mn(III)-superoxo nature. This result coincides with the observed EPR signal as we referred before. Such a Mn(III)-superoxo nature can be evidenced by the occupations of NBOs involving Mn 3d orbitals and O2 lone pair (LP) orbitals as given in ESI (Table S5): four of Mn 3d orbitals, i.e., 3dxz, 3dyz, 3dxy, 3dz2 have nearly 1e occupations in α space while with very small occupations (less than 0.20e) in their β counterparts (3dx2−y2 orbital is nearly empty in both α and β spaces), corroborating a Mn(III) nature; the occupation differences of O2 LP orbitals (including orbitals of both O1 and O2 atoms) between α and β counterparts (α–β) are generally 1.0–1.1e, corroborating a superoxo nature. Besides, a Mn(III)-superoxo character is also reflected in the Mulliken spin populations on Mn–O2 unit (SMn = 3.9e–4.0e, SO1+O2 ≈ 1.0e) and Wiberg bond index of O1–O2 (BO1–O2 ≈ 1.5) as shown in ESI (Table S6). The presence of additional waters does not affect the Mn(III)-superoxo electronic structure, but induce more negative charges on distal O1 atoms than those of the middle O2 atoms (Table 4). It should be the reason why a more end-on Mn–O2 unit was more reactive toward nucleophilic reactions as experimentally proposed.12 Otherwise, the fractional occupations on Mn 3d orbitals and dioxygen LP orbitals in β space are responsible for the non-integer natural populations on Mn and O atoms (see Fig. S5, ESI). It implies back charge transfer from superoxo to Mn(III) in β electron space. Taking the bent end-on [MnL(O2)(H2O)]2+ for instance (Table S5, ESI), the pip orbital (the subscripts ip here and after denote orbitals in the plane of MnO2O1) of O2 atom has a β-occupation about 0.8e and the Mn 3dz2 orbital has a β-occupation about 0.2e (both of their α-counterparts are singly occupied), hinting a back charge transfer from pip(O2) to Mn 3dz2 orbital and thus resulting in the less than 1e population on dioxygen ligand (Table 4).
Table 4 Natural populations (Q/e) on Mn–O2 unit in sextet [MnL(O2)(H2O)]2+nH2O complexes. The results of [MnL(O2)(H2O)]2+–8H2O in CPCM model is at B3LYP/B2 level
  n = 0 n = 1 n = 4 n = 6 n = 8 n = 8 (CPCM)
QMn/e 1.15 1.18 1.21 1.21 1.24 1.22
QO1/e −0.32 −0.37 −0.44 −0.44 −0.43 −0.41
QO2/e −0.37 −0.34 −0.32 −0.31 −0.31 −0.30


To estimate the stabilizing effect of charge delocalization from superoxo to Mn(III) in β space (similar charge delocalization in α space is negligible since the referred 3d orbitals is occupied), we listed in Fig. 5 the dominant second-order perturbation energies ΔEij from superoxo LP orbitals to Mn(III) unoccupied 3d orbitals in β space for bent end-on [MnL(O2)(H2O)]2+nH2O complexes. The larger ΔEij indicates the more tendency for electrons delocalization from superoxo to Mn(III), thus the more intensive stabilizing interaction. It was found that there are two important stabilization energies for bent end-on structures (Fig. 5 and S5): one corresponds to charge delocalization from the pip(O2) orbital to Mn(III) 3dz2 orbital (pip(O2) → 3dz2); the other one corresponds to charge delocalization from pip(O1) orbitals to the σ anti-bonding orbitals of the nearest additional water (image file: c6ra11199h-t1.tif) (none for n = 0). pip(O2) → 3dz2 charge delocalization should be a σ-donation. Seen from Fig. 5, ΔEij (pip(O2) → 3dz2) is quite large for [MnL(O2)(H2O)]2+ (i.e., n = 0, 41.47 kcal mol−1), corresponding to a σ-bonding interaction of superoxo image file: c6ra11199h-t2.tif with Mn 3dz2 as described in other work.15,36,58 The corresponding molecular orbital (image file: c6ra11199h-t3.tif) is depicted in ESI (Fig. S6). While, this interaction gets greatly reduced as number of additional waters increases, e.g., from 41.47 kcal mol−1 with n = 0 to 1.08 kcal mol−1 with n = 8. It is possibly because of the elongated Mn–O2 distances in the presence of water as shown in Fig. 3 and the enlarged Mn–O2–O1 angle that would make pip(O2) orbital more horizontal, thus unfavourable charge delocalization to Mn 3dz2 orbital. The image file: c6ra11199h-t4.tif stabilization energy in the presence of water suggests strong H-bonding interactions between O1 atom with additional waters. Moreover, the dominant image file: c6ra11199h-t5.tif donation are always trans to the bending direction of dioxygen (Fig. 5 and S5). It should be the exact reason for the enlarged Mn–O2–O1 angles in the presence of water.


image file: c6ra11199h-f5.tif
Fig. 5 NBOs of superoxo donors (left) and Mn(III) acceptors (right) for sextet [MnL(O2)(H2O)]2+nH2O (n = 1) complex, as well as second-order perturbation energies (ΔEij in kcal mol−1) for n = 0, 1, 4, 6, 8 in β space. All hydrogen atoms are omitted for clarity. NBOs for [MnL(O2)(H2O)]2+nH2O (n = 0, 4, 6, 8) are similar to n = 1 and can be found in ESI.

The stabilizing interactions between superoxo and Mn(III) center in linear end-on and side-on [MnL(O2)(H2O)]2+ structures can also be interpreted based on ΔEij (Fig. S5, ESI), although a general understanding on the bonding interaction of different configurated M–O2 complexes have been achieved in many other work.13,15,57 We found that the pip(O2) → 3dz2 charge delocalization is largely decreased in the linear end-on (6.94 kcal mol−1) and side-on configurations (2.39 kcal mol−1) as compared to the bent end-on structure, because of the unmatched orbital symmetry of pip(O2) with 3dz2 in these two binding modes. Instead, a π-type delocalization between pip(O2) and 3dxz orbital (pip(O2) → 3dxz), becomes dominant in the linear end-on structure (Fig. 5b) (about 8.63 kcal mol−1). It must cause the linear end-on structure less stable than the bent end-on structure; for side-on structure,ΔEij shows comparable stabilizing interactions between pip orbitals of O1, O2 atoms with Mn 3dxz orbital. It should correspond to the σ bonding interaction of the π* orbital of dioxygen with metal 3d orbital (e.g., 3dxz or else) as described elsewhere13,15,36

Conclusions

By taking an experimentally isolated [MnL(O2)(H2O)]2+ complex as model system, this work provided explorations on the most favourable binding mode of dioxygen in [MnL(O2)(H2O)]2+ complex based on DFT calculations. Influences of solvent and sizes of supporting ligands on Mn–O–O bond angle were investigated. From our calculations, we found that (i) functionals including B3LYP, wB97X-D, M06-2X and B3LYP-D3 predict [MnL(O2)(H2O)]2+ complex has a sextet ground-state, consistent with experimental results. While, BP86 predicts a quartet ground-state; (ii) [MnL(O2)(H2O)]2+ has an equilibrium structure with a bent end-on binding mode, featuring Mn–O–O angle at around 130° and dioxygen bond length at ∼1.32 Å. The experimentally proposed linear end-on structure (Mn–O–O angle ∼180°) and the side-on structure (Mn–O–O angle ∼75°) are shown to be transients in the process of dioxygen flipping from one bent-end-on structure to another; (iii) inclusion of water solvation effect via both explicit water-cluster model ([MnL(O2)(H2O)]2+nH2O, n = 1, 4, 6, 8) and implicit CPCM model benefits a larger Mn–O–O angle, e.g., Mn–O–O angle of [MnL(O2)(H2O)]2+–8H2O is 162.6° in vacuum and 159.6° in CPCM model. The key bond lengths of [MnL(O2)(H2O)]2+–8H2O, except dO1–O2, also get much closer to the experimental parameters. It indicates that inclusion of water solvation effect from explicit water solvent molecules is necessary in describing the geometric structure of [MnL(O2)(H2O)](PF6)2 complex from a theoretical model.

The deviation of DFT calculated dO1–O2 (1.321 Å) from the X-ray distance (1.249 Å) was reconciled by the projected distance of dO1–O2 on the Mn–O bond direction of [MnL(O2) (H2O)]2+–8H2O within CPCM model (1.240 Å). Together with analyses on thermal ellipsoids of O atoms and the transformation energies from a bent end-on structure to a linear end-on structure (PES in Fig. S2, ESI), this reconciliation gave clues about the dioxygen flipping motion around Mn–O bond in the solid state structure of [MnL(O2)(H2O)](PF6)2 complex when the diffraction data were collected. It thus misguided the experimentalists with a linear Mn–O–O arrangement. Otherwise, a good correlation is concluded that the more additional waters and the smaller Oe squares would benefit a larger Mn–O–O angle, indicating a water solvation effect and a ligand size effect. These results shed light on controllable synthesis of more end-on or more bent M–O2 complexes through solvent polarity perturbations and supporting ligand size modifications. Whereas, the ligand size effect is shown to be less important than the water solvation effect. The physical origin of the water solvation effect can attribute to the strongest H-bonding interaction locating at the back face of the distal O atom in each [MnL(O2)(H2O)]2+nH2O (n = 1, 4, 6, 8) complex (see Fig. 3 and 5). NBO analyses show [MnL(O2)(H2O)]2+ has a Mn(III)-superoxo nature, consistent with experimental results. The presence of additional waters would not affect the Mn(III)-superoxo nature. The second-order perturbation energies ΔEij from superoxo LP orbitals to Mn(III) unoccupied 3d orbitals in β space gave explanations about the instability of linear end-on mode as compared to the bent end-on mode and why the presence of water benefits a more end-on Mn–O–O angle.

Acknowledgements

The authors acknowledge the National Natural Science Foundation of China (No. 21101109, 21171124) and Specialized Research Fund for the Doctoral Program of Higher Education (No. 20113201120005). H. F. Wang appreciates the Priority Academic Program Development of Jiangsu Higher Education Institutions and suggestions of Prof. Z. P. Liu and Prof. Z. H. Li at Fudan University.

Notes and references

  1. K. T. Horak and T. Agapie, J. Am. Chem. Soc., 2016, 138, 3443–3452 CrossRef CAS PubMed.
  2. K. Ray, F. F. Pfaff, B. Wang and W. Nam, J. Am. Chem. Soc., 2014, 136, 13942–13958 CrossRef CAS PubMed.
  3. J. Cho, R. Sarangi and W. Nam, Acc. Chem. Res., 2012, 45, 1321–1330 CrossRef CAS PubMed.
  4. J. Cho, J. Woo and W. Nam, J. Am. Chem. Soc., 2010, 132, 5958–5959 CrossRef CAS PubMed.
  5. J. Li, B. C. Noll, A. G. Oliver and W. R. Scheidt, J. Am. Chem. Soc., 2012, 134, 10595–10606 CrossRef CAS PubMed.
  6. C. Wurtele, E. Gaoutchenova, K. Harms, M. C. Holthausen, J. Sundermeyer and S. Schindler, Angew. Chem., Int. Ed., 2006, 45, 3867–3869 CrossRef PubMed.
  7. H. Hori, M. Ikedasaito and T. Yonetani, Nature, 1980, 288, 501–502 CrossRef CAS PubMed.
  8. E. Melamud, B. L. Silver and Z. Dori, J. Am. Chem. Soc., 1974, 96, 4689–4690 CrossRef CAS.
  9. P. Chen, D. E. Root, C. Campochiaro, K. Fujisawa and E. I. Solomon, J. Am. Chem. Soc., 2003, 125, 466–474 CrossRef CAS PubMed.
  10. H. Kang, J. Cho, K. B. Cho, T. Nomura, T. Ogura and W. Nam, Chem.–Eur. J., 2013, 19, 14119–14125 CrossRef CAS PubMed.
  11. K. Fujisawa, M. Tanaka, Y. Morooka and N. Kitajima, J. Am. Chem. Soc., 1994, 116, 12079–12080 CrossRef CAS.
  12. J. Annaraj, J. Cho, Y. M. Lee, S. Y. Kim, R. Latifi, S. P. de Visser and W. Nam, Angew. Chem., Int. Ed., 2009, 48, 4150–4153 CrossRef CAS PubMed.
  13. J. W. Ginsbach, R. L. Peterson, R. E. Cowley, K. D. Karlin and E. I. Solomon, Inorg. Chem., 2013, 52, 12872–12874 CrossRef CAS PubMed.
  14. J. Cho, H. Y. Kang, L. V. Liu, R. Sarangi, E. I. Solomon and W. Nam, Chem. Sci., 2013, 4, 1502–1508 RSC.
  15. H. Chen, K. B. Cho, W. Z. Lai, W. Nam and S. Shaik, J. Chem. Theory Comput., 2012, 8, 915–926 CrossRef CAS PubMed.
  16. S. T. Prigge, B. A. Eipper, R. E. Mains and L. M. Amzel, Science, 2004, 304, 864–867 CrossRef CAS PubMed.
  17. C. J. Cramer and W. B. Tolman, Acc. Chem. Res., 2007, 40, 601–608 CrossRef CAS PubMed.
  18. M. Costas, M. P. Mehn, M. P. Jensen and L. Que, Chem. Rev., 2004, 104, 939–986 CrossRef CAS PubMed.
  19. R. A. Geiger, S. Chattopadhyay, V. W. Day and T. A. Jackson, J. Am. Chem. Soc., 2010, 132, 2821–2831 CrossRef CAS PubMed.
  20. E. A. Lewis and W. B. Tolman, Chem. Rev., 2004, 104, 1047–1076 CrossRef CAS PubMed.
  21. L. L. Liu, H. X. Li, L. M. Wan, Z. G. Ren, H. F. Wang and J. P. Lang, Chem. Commun., 2011, 47, 11146–11148 RSC.
  22. K. B. Cho, H. Chen, D. Janardanan, S. P. de Visser, S. Shaik and W. Nam, Chem. Commun., 2012, 48, 2189–2191 RSC.
  23. A. Dedieu, M. M. Rohmer and A. Veillard, J. Am. Chem. Soc., 1976, 98, 5789–5800 CrossRef CAS.
  24. A. B. Wolk, C. M. Leavitt, J. A. Fournier, M. Z. Kamrath, G. B. Wijeratne, T. A. Jackson and M. A. Johnson, Int. J. Mass Spectrom., 2013, 354, 33–38 CrossRef.
  25. R. L. Shook, W. A. Gunderson, J. Greaves, J. W. Ziller, M. P. Hendrich and A. S. Borovik, J. Am. Chem. Soc., 2008, 130, 8888–8889 CrossRef CAS PubMed.
  26. C.-M. Lee, C.-H. Chuo, C.-H. Chen, C.-C. Hu, M.-H. Chiang, Y.-J. Tseng, C.-H. Hu and G.-H. Lee, Angew. Chem., Int. Ed., 2012, 51, 5427–5430 CrossRef CAS PubMed.
  27. H. So, Y. J. Park, K. B. Cho, Y. M. Lee, M. S. Seo, J. Cho, R. Sarangi and W. Nam, J. Am. Chem. Soc., 2014, 136, 12229–12232 CrossRef CAS PubMed.
  28. M. K. Coggins, V. Martin-Diaconescu, S. DeBeer and J. A. Kovacs, J. Am. Chem. Soc., 2013, 135, 4260–4272 CrossRef CAS PubMed.
  29. C. J. Weschler, B. M. Hoffman and F. Basolo, J. Am. Chem. Soc., 1975, 97, 5278–5280 CrossRef CAS PubMed.
  30. R. A. Geiger, D. F. Leto, S. Chattopadhyay, P. Dorlet, E. Anxolabehere-Mallart and T. A. Jackson, Inorg. Chem., 2011, 50, 10190–10203 CrossRef CAS PubMed.
  31. M. S. Seo, J. Y. Kim, J. Annaraj, Y. Kim, Y. M. Lee, S. J. Kim, J. Kim and W. Nam, Angew. Chem., Int. Ed., 2007, 46, 377–380 CrossRef CAS PubMed.
  32. R. Sarangi, J. Cho, W. Nam and E. I. Solomon, Inorg. Chem., 2011, 50, 614–620 CrossRef CAS PubMed.
  33. P. J. Donoghue, A. K. Gupta, D. W. Boyce, C. J. Cramer and W. B. Tolman, J. Am. Chem. Soc., 2010, 132, 15869–15871 CrossRef CAS PubMed.
  34. M. Schatz, V. Raab, S. P. Foxon, G. Brehm, S. Schneider, M. Reiher, M. C. Holthausen, J. Sundermeyer and S. Schindler, Angew. Chem., Int. Ed., 2004, 43, 4360–4363 CrossRef CAS PubMed.
  35. J. Zapata-Rivera, R. Caballol, C. J. Calzado, D. G. Liakos and F. Neese, Chem.–Eur. J., 2014, 20, 13296–13304 CrossRef CAS PubMed.
  36. J. Zapata-Rivera, R. Caballol and C. J. Calzado, Phys. Chem. Chem. Phys., 2015, 17, 2814–2822 RSC.
  37. M. J. T. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision C.01, Gaussian, Inc., Wallingford CT, 2010 Search PubMed.
  38. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652 CrossRef CAS.
  39. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241 CrossRef CAS.
  40. V. N. Staroverov, G. E. Scuseria, J. M. Tao and J. P. Perdew, J. Chem. Phys., 2003, 119, 12129–12137 CrossRef CAS.
  41. S. Grimme, J. Comput. Chem., 2006, 27, 1787–1799 CrossRef CAS PubMed.
  42. J.-D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008, 10, 6615–6620 RSC.
  43. K. S. Thanthiriwatte, E. G. Hohenstein, L. A. Burns and C. D. Sherrill, J. Chem. Theory Comput., 2011, 7, 88–96 CrossRef CAS PubMed.
  44. R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem. Phys., 1980, 72, 650–654 CrossRef CAS.
  45. P. Harihara and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213–222 CrossRef.
  46. NIST Computational Chemistry Comparison and Benchmark Database, NIST Standard Reference Database Number 101 Release 16a, ed. R. J. Johnson, National Institute of Standards and Technology, Gaithersbug, MD, August 2013, http://cccbdb.nist.gov/ Search PubMed.
  47. E. D. Glendening, C. R. Landis and F. Weinhold, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2012, 2, 1–42 CrossRef CAS.
  48. T. Wang, G. W. Brudvig and V. S. Batista, J. Chem. Theory Comput., 2010, 6, 2395–2401 CrossRef CAS PubMed.
  49. E. M. Sproviero, J. A. Gascon, J. P. McEvoy, G. W. Brudvig and V. S. Batista, J. Inorg. Biochem., 2006, 100, 786–800 CrossRef CAS PubMed.
  50. M. Jaccob, A. Ansari, B. Pandey and G. Rajaraman, Dalton Trans., 2013, 42, 16518–16526 RSC.
  51. K. B. Cho, S. Shaik and W. Nam, J. Phys. Chem. Lett., 2012, 3, 2851–2856 CrossRef CAS.
  52. X. S. Li and M. J. Frisch, J. Chem. Theory Comput., 2006, 2, 835–839 CrossRef CAS PubMed.
  53. J. H. Bowen, N. V. Shokhirev, A. M. Raitsimring, D. H. Buttlaire and F. A. Walker, J. Phys. Chem. B, 1997, 101, 8683–8691 CrossRef CAS.
  54. F. A. Walker and J. Bowen, J. Am. Chem. Soc., 1985, 107, 7632–7635 CrossRef CAS.
  55. J. F. Li, B. C. Noll, A. G. Oliver, C. E. Schulz and W. R. Scheidt, J. Am. Chem. Soc., 2013, 135, 15627–15641 CrossRef CAS PubMed.
  56. H. Chen, M. Ikeda-Saito and S. Shaik, J. Am. Chem. Soc., 2008, 130, 14778–14790 CrossRef CAS PubMed.
  57. Y. Chen and S. Sakaki, Inorg. Chem., 2013, 52, 13146–13159 CrossRef CAS PubMed.
  58. I. Bytheway and M. B. Hall, Chem. Rev., 1994, 94, 639–658 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available: Test calculations using different functional and different basis set levels, optimized structural parameters of [MnL(O2)(H2O)]2+nH2O (n = 1, 4, 6, 8) under different basis sets, PES of three binding configurations, NBO orbital interactions of [MnL(O2)(H2O)]2+nH2O (n = 1, 4, 6, 8) and Catesian coordinates for all structures described in this work. See DOI: 10.1039/c6ra11199h

This journal is © The Royal Society of Chemistry 2016
Click here to see how this site uses Cookies. View our privacy policy here.