Characterization of Metastable Intermediates Formed in the Reaction Between a Mn(II) Complex and Dioxygen, Including a Crystallographic Structure of a Binuclear Mn(III)-Peroxo Species

Abstract

Transition-metal peroxos have been implicated as key intermediates in a variety of critical biological processes involving O₂. Due to their highly reactive nature, very few metal-peroxos have been characterized. The dioxygen chemistry of manganese remains largely unexplored despite the proposed involvement of a binuclear Mn-peroxo, either as a precursor to O₂, or derived from O₂, in both photosynthetic H₂O oxidation and DNA biosynthesis, arguably two of the most fundamental processes of life. Neither of these biological intermediates has been observed. Herein we describe the dioxygen chemistry of coordinatively unsaturated [Mn^II(S^Me2N₄(6-MeDPEN))] ⁺(1), and the characterization of intermediates formed en route to a binuclear mono-oxo bridged Mn(III) product {[Mn^III(S^Me2N₄(6-MeDPEN)]₂-(μ-O)}²⁺ (2), the oxo atom of which is derived from ¹⁸O₂. At low-temperatures, a dioxygen intermediate, [Mn(S^Me2N₄(6-MeDPEN))(O₂)]⁺ (4), is observed (by stopped-flow) to rapidly and irreversibly form in this reaction (k₁(−10 °C)= 3780±180M⁻¹s⁻¹, ΔH₁^‡ = 26.4±1.7 kJ mol⁻¹, ΔS₁^‡ = − 75.6±6.8 J mol⁻¹K⁻¹), and then convert more slowly (k₂(−10 °C)= 417±3.2 M⁻¹s⁻¹, ΔH₂^‡ = 47.1±1.4 kJ mol⁻¹, ΔS₂^‡ = − 15.0±5.7 J mol⁻¹K⁻¹) to a species 3 with isotopically sensitive stretches at ν_o-o (Δ¹⁸O) = 819(47) cm⁻¹, k_O–O= 3.02 mdyn/Å, and ν_Mn-O(Δ¹⁸O) = 611(25) cm⁻¹ consistent with a peroxo. Intermediate 3 releases approximately 0.5 equiv of H₂O₂ per Mn ion upon protonation, and the rate of conversion of 4 to 3 is dependent on [Mn(II)] concentration, consistent with the formation of a binuclear Mn-peroxo. This was verified by X-ray crystallography, where the peroxo of {[Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)}²⁺ (3) is shown to be bridging between two Mn(III) ions in an end-on trans-μ-1,2-fashion. This represents the first characterized example of a binuclear Mn(III)-peroxo, and a rare case in which more than one intermediate is observed en route to a binuclear μ–oxo bridged product derived from O₂. Vibrational and metrical parameters for binuclear Mn-peroxo 3 are compared with those of related binuclear Fe- and Cu-peroxo compounds.

Introduction

Transition-metal ions promote a wide variety of biochemical processes involving O₂.^1-6 Metal-peroxos are proposed as key intermediates in a majority of these processes.^2,3,7,8 For example, binuclear peroxo-bridged Mn intermediates (Mn–μ–(O₂²⁻)–Mn) are proposed to form during two of the most fundamental processes of life, that of photosynthetic H₂O splitting^9-15 and DNA biosynthesis.^16-18 Photosynthetic H₂O splitting^9-15 generates approximately 80% of the O₂ on our planet. However, neither the photosynthetic Mn-peroxo, nor the Mn-peroxo implicated in DNA biosynthesis has been observed. Under extremely high (30 bar) O₂ pressures a transient Mn species is observed during photosynthetic H₂O splitting,¹⁹ however, the identity of this species is unknown. Although similar peroxo-bridged intermediates are implicated in both of these fundamental processes, one forms via O–O bond cleavage en route to a more reactive high-valent metal-oxo intermediate,^16-18,20 while the other forms via O–O bond formation. Our understanding of the structural and electronic properties that favor O–O bond formation versus cleavage is extremely limited for Mn. Understanding how to efficiently oxidize H₂O to O₂ by examining the microscopic reverse process involving the reduction of dioxygen (O₂) to water (H₂O) is critical²¹ to the development of renewable solar fuels.²² Given that it can promote these reactions under mild ambient conditions, Nature has much to teach us. Peroxo O–O bond cleavage has been shown with other metals (Cu and Fe) to be facilitated either by binding the peroxo between two metal ions, ideally in a side-on μ–η²:η²–bridging mode,^23-26 or, by protonating one of the oxygens.^27-29 The activation barrier to O₂ binding and activation has been shown to be lowered by thiolate ligands (RS⁻).³⁰ Thiolate ligands also provide a convenient spectroscopic handle with which to probe reactivity.³¹ Despite their key role in biological catalysis, very few Mn-peroxo species have been characterized,^32-43 due, in part, to their instability.^40,44-46 In fact, structurally characterized examples of middle-to-late first-row transition-metal peroxos are, in general, scarce.^1,23,47,48 Of the seven structurally characterized metastable Mn(III)-peroxo compounds,^33,36-39,43 all but one⁴⁹ are mononuclear, and five contain the peroxo in a side-on (η²) binding mode. The only structural evidence for an end-on (η¹) peroxo binding mode with Mn(III) was reported in 1988,⁴⁹ and more recently by our group.³³ There are only two examples of peroxo-bridged Mn-clusters, one binuclear Mn(IV)₂ peroxo compound,⁴⁰ and one trinuclear Mn(III)₃ peroxo compound.⁴⁹

Recently we reported a series of coordinatively unsaturated thiolate-ligated Mn(II) complexes that react with dioxygen, via an observable intermediate, to afford rare examples of unsupported binuclear mono oxo bridged Mn(III) complexes.⁵⁰ Isotopic labeling studies showed that the oxo atom is derived from O₂.⁵⁰ Herein we describe the kinetics of this O₂ reaction, and provide evidence that more than one intermediate is involved, one of which is shown to be a binuclear peroxo-bridged Mn(III) complex. Crystallographic characterization of this dioxygen-derived intermediate provides the first structurally characterized example of a binuclear Mn(III)-peroxo. The significance of this structure is that it is analogous to key intermediates proposed to be involved in photosynthetic O₂ evolution and DNA biosynthesis.^9-18,20

Experimental Section

General Methods

All manipulations were performed using Schlenk line techniques or under an N₂ atmosphere in a glovebox. Reagents and solvents were purchased from commercial vendors, were of the highest available purity, and were used without further purification unless otherwise noted. Propionitrile (CaH₂) and MeCN (CaH₂), were dried and distilled prior to use. Et₂O was rigorously degassed and purified using solvent purification columns housed in a custom stainless steel cabinet and dispensed by a stainless steel schlenk-line (GlassContour). UV/vis spectra were recorded on a Varian Cary 50 spectrophotometer equipped with a fiber optic cable connected to a “dip” ATR probe (C-technologies). A custom-built two-neck solution sample holder equipped with a threaded glass connector was sized specifically to fit the “dip” probe. Magnetic moments (solid state) were obtained with polycrystalline samples in gel-caps from 5 to 300 K by zero-field cooling experiments using a Quantum Design MPMS S5 SQUID magnetometer. Pascal's constants were used to correct for diamagnetic contributions to the experimental magnetic moment. X-ray crystallography data was recorded on either a Bruker APEX II single crystal X-ray diffractometer with Mo-radiation or a Bruker SMART Apex CCD diffractometer with Mo Kα radiation.

Synthesis of Peroxo-Bridged [Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2– O₂)(BPh₄)₂•2CH₃CH₂CN (3)

[Mn^II(S^Me2N₄(6-Me-DPEN)](BPh₄) (1) was synthesized as previously described for [Mn^II(S^Me2N₄(6-Me-DPEN)](PF₆), ⁵⁰ by replacing the NaPF₆ counterion with NaBPh₄. Peroxo–bridged 3 was prepared by cooling a propionitrile solution of 1 to −80°C, opening the flask to air, and then layering the cold solution with pre-cooled Et₂O.

Hydrogen Peroxide Detection Assay

A concentrated propionitrile solution of 1 (0.1 mL, 5.5 mM) was prepared under an inert atmosphere. After cooling the solution to −80 °C (with a dry ice/acetone bath), O₂ was gently bubbled into the solution, directly from a cylinder, for approximately two minutes. This procedure promoted a color change from light yellow to dark green, indicating the formation of peroxo-bridged 3. A minimal amount of concentrated H₂SO₄ (3-4 drops) was then added to the dark green solution, causing the solution to turn clear. The resulting reaction mixture was then passed through a very small silica plug, collected, and then added in a single aliquot to a stirring aqueous solution of KMnO₄ (3 mL, 6.15 × 10⁻⁵ M). The resulting changes in absorbance values were then recorded in one minute intervals until no further changes were observed. The amount of hydrogen peroxide present was calculated based on the absorbance at 550 nm (ε (M⁻¹cm⁻¹, 298 K) = 2455) caused by the following reaction:

$$2{\text{KMnO}}_4+5{\text{H}}_2{\text{O}}_2+6{\text{H}}^{+}\to 2{\text{Mn}}^{2+}+5{\text{O}}_2+8{\text{H}}_2\text{O}+2{\text{K}}^{+}$$

In a control experiment, the same procedure was performed upon oxo-bridged 2, which failed to illicit any considerable absorbance changes to aqueous solutions of KMnO₄ of known concentration.

Stopped Flow Kinetic Measurements

Acetonitrile and propionitrile solutions of the reagents were prepared in an MBraun glovebox filled with high purity argon and placed in Hamilton gastight syringes. Time-resolved spectra (400-800 nm) were acquired at low temperatures using a TgK Scientific (U.K.) SF-61DX2 Multi-Mixing CryoStopped-Flow Instrument, a J&M TIDASD-AQ diode array detector and MCS UV/NIR light source (Spectralytics, DenMark). The stopped-flow instrument was equipped with PEEK tubings fitted inside stainless steel plumbing, a 1.00 cm³ quarz mixing cell, and an anaerobic kit purged with an inert gas. The temperature in the mixing cell was maintained to 0.1 °C and the mixing time was 2-3 ms. All flow lines of the instrument were extensively washed with degassed, anhydrous acetonitrile or propionitrile before charging the driving syringes with reactant solutions. The reactions were studied by rapid scanning spectrophotometry under pseudo-first-order conditions with excess oxygen (see experimental details in Supplemental Material). Saturated solutions of O₂ were prepared by bubbling dry O₂ gas for 15 minutes at 25 °C into gastight syringes containing dry CH₃CN or CH₃CH₂CN; dilutions of the O₂ saturated solvent were performed anaerobically to obtain the desired [O₂]. The solubility of O₂ was taken as 8.1 mM in CH₃CN and 8.8 mM in CH₃CH₂CN at 25°C.⁵¹ All of the experiments were performed in a single-mixing mode of the instrument, with a 1:1 (v/v) mixing ratio. A series of three or four measurements gave an acceptable standard deviation (within 10%). Data analysis was performed with Kinetic Studio software from Hi-Tech Scientific and IGOR Pro software from Wavemetrics, Inc.

Resonance Raman Experiments

The Raman system consists of a Coherent I90C-K Kr⁺ laser, a SPEX model 1877 CP triple monochromator with a filter and blazed holographic gratings of 2400 grooves/mm and a Newton DU940N-BU 2048 × 512 pixel back-illuminated, thermoelectric-cooled CCD array. Crystalline samples of ¹⁶O and ¹⁸O-labelled peroxo-bridged 3 were prepared as described above. Spectra were taken at 77K with 5 - 40 mw laser power, and the sample was continuously rotated to minimize photo-decay.

X-Ray Crystallographic Structure Determination

A green needle of 3 measuring 0.20 × 0.15 × 0.05 mm³, was mounted on a glass capillary with oil. Data was collected at −173°C on a Bruker APEX II single crystal X-ray diffractometer, using Mo-radiation. The crystal-to-detector distance was set to 40 mm, and the exposure time was 30 s deg⁻¹ for all sets of exposure. The scan width was 0.5°. Data collection was 89.4% complete to 25.0° in θ. A total of 27,190 merged reflections were collected covering the indices, h = −15 to 15, k = −22 to 22, l = −22 to 22. 7,909 reflections were symmetry independent and the R_int = 0.0964 indicated that the data was of less than average (0.07) quality. Indexing and unit cell refinement indicated a monoclinic P lattice with the space group be P 2₁/c (No.14).

The data for 3 was integrated and scaled using SAINT, and SADABS within the APEX2 software package by Bruker. Solution by direct methods (SHELXS, SIR97) produced a complete heavy atom phasing model consistent with the proposed structure. The structure was completed by difference Fourier synthesis with SHELXL97.^52,53 Scattering factors are from Waasmair and Kirfel.⁵⁴ Hydrogen atoms were placed in geometrically idealised positions and constrained to ride on their parent atoms with C---H distances in the range 0.95-1.00 Angstrom. Isotropic thermal parameters (U_eq) for the hydrogens were fixed such that they were 1.2U_eq of their parent atom Ueq for methylenes and CH's, and 1.5U_eq of their parent atom U_eq in case of methyl groups. All non-hydrogen atoms were refined anisotropically by full-matrix least-squares. Crystal data for 3 are summarized in Table 1, and metrical parameters are provided in Table 2.

Table 1.

Crystal data for [Mn^III(S^Me2N₄(6-Me-DPEN)]₂ (trans–μ–1,2–O)(BPh⁴)₂ •2CH₃CH₂CN (3).

	3
Formula	C108 H128B2Mn2N14O2S2
MW	1849.86
T, K	100(2) K
Unit cellA	Monoclinic
a, Å	14.2799(10)
b, Å	18.3957(14)
c, Å	18.5282(13)
α, deg	90
β, deg	94.541(2)
γ, deg	90
V, Å3	4851.9(6)
Z	2
d(calc), g/cm3	1.266
Space group	P 21/c
R	0.0613b
Rw	0.1614c
GOF	1.002

^aIn all cases: Mo Kα(λ = 0.71070 Å) radiation.

^bR = Σ ‖F_o| - |F_c‖/Σ |F_o|; R_w = [Σw(|F_o| = |F_c|)²/ΣwF_o²]^1/2, where w⁻¹ = [σ_count² + (0.05F²)²]/4F².

Table 2.

Comparison of Vibrational and Metrical Parameters for Selected Metal-Peroxo Complexes.

Complex	vO-O, cm-1	O–O, Å	Peroxo binding mode	Ref
[Mn2(SMe2N4(6-Me-DPEN))2(O2)]2+ (3)	819	1.452(5)	trans μ-1,2	This work
[Mn(3,5-iPr2pzH){(HB(3,5-iPr2pz)3}(O2)]	892	1.428(7)	side-on η2	(37)
[Mn(H3bupa)(O2)]−	885	N/A	side-on η2	(35)
[Mn(tmc)(O2)]+	N/A	1.403(4)	side-on η2	(43)
[Fe2(O2)(O2CCH2Ph) {(HB(pz′)3}2]	888	1.408(9)	trans μ-1,2	(7), (76)
[Fe2 (O)2(N-Et-hptb)(Ph3PO)2)2]3+	900	1.416(7)	cis μ-1,2	(63), (61)
[Fe2 (6Me2-BPP)2(OH)(O2)]+	908	1.396(5)	cis μ-1,2	(59)
[Cu2(tmpa)2(O2)]+	825	1.432(6)	trans μ-1,2	(26), (64), (63)
[Cu2(Me6tren)2(O2)]2+	814	1.368(9)	trans μ-1,2	(92)
[Cu2{(HB(3,5-iPr2pz)3}2(O2)]	741	1.412(12)	μ-η2: η2	(25), (56)
[Cu2(MeAN)2(O2)]2+	721	1.540(5)	μ-η2: η2	(57)

Two highly disordered propionitriles were found to co-crystallize with peroxo-bridged [Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)(BPh₄)₂•2CH₃CH₂CN (3), one modeled over two positions (83/17 ratio), and the second modeled over three positions (66/17/17 ratio). The occupancies of each disorder component were refined freely while restraining the total occupancy sum to one. Disorder in one of the phenyl rings of the tetraphenylborate anion was modeled over two positions (66/34 ratio) without any additional restraints.

Results and Discussion

Coordinatively unsaturated [Mn^II(S^Me2N₄(6-MeDPEN))](BF₄) (1) was synthesized and structurally characterized as previously described.⁵⁰ Dioxygen addition to 1 at ambient temperature causes an immediate color change, from colorless to purple, which is accompanied by the growth of a peak at 560(ε= 520 M⁻¹cm⁻¹) nm in the electronic absorption spectrum (Figure 1). Crystallization of the product and characterization by X–ray crystallography showed the final O₂–derived product to be a rare example of a stable unsupported mono-oxo-bridged binuclear Mn(III) compound, [Mn^III(S^Me2N₄(6-MeDPEN)]₂-(μ-O)(BF₄)₂(2) (Scheme 1).^{50 18}O-labelling studies⁵⁰ showed that the bridging oxo atom (blue, Scheme 1) of 2 is derived from dioxygen. When this reaction is monitored at low temperatures (−40 °C, in MeCN) via electronic absorption spectroscopy, a green metastable intermediate, 3, is detected, which displays an absorption band at 640(ε= 830 M⁻¹cm⁻¹) nm (Figure 1). The BF₄⁻ salt of this intermediate converts to oxo-bridged 2 within minutes at −40 °C. If the BF₄⁻ or PF₆⁻ counterion is replaced with BPh₄⁻, then the low-temperature (−40 °C) lifetime of 3 can be extended (to ∼45min) in MeCN. The conversion of 3 to oxo-bridged 2 is nearly quantitative (96-98%) with either counterion, based on the amount of Mn(II) starting material, implying that metastable 3 is well-behaved, and cleanly converts to 2. Low temperature addition of H₂SO₄ to in situ generated solutions of 3, reproducibly affords 0.43(4) equiv of H₂O₂ per equiv of Mn²⁺, as detected and quantified via a MnO₄⁻ titration using the method described in the experimental section (Figure S-1). The reproducible observation of hydrogen peroxide in a 2:1 (Mn: H₂O₂) stoichiometry suggested that metastable 3 is a bimetallic peroxo, Mn-(μ-O₂²⁻)-Mn.

Figure 1.

Electronic absorption spectrum of thiolate-ligated [Mn^II(S^Me2N₄(6-MeDPEN))] ⁺(1), and the reaction product 2, and intermediate 3, formed in its low-temperature reaction with dioxygen (O₂) in MeCN.

Scheme 1.

Although there is only one reported example of a resonance Raman–characterized Mn-peroxo species,^34,49 we were able to successfully perform these experiments using crystalline samples of 3. Resonance Raman spectra were collected using a series of excitation wavelengths, and resonance enhanced ν_O-O and ν_Mn-O vibrations are only observed when an excitation frequency of 413 nm near the high energy absorption band of 3 (at 385 nm) is used. Excitation wavelengths near the lower energy visible band (at 640 nm; Figure S–2) did not, on the other hand, result in resonance enhanced peroxo vibrations, indicating that the higher energy, as opposed to lower energy band, corresponds to a peroxo-to-metal charge transfer band. This is in contrast to iron-peroxo compounds, which display a peroxo-to-metal charge transfer band in the visible region (at ∼700 nm),^20,55 but is in agreement with the TD-DFT calculated peroxo-to-metal charge transfer transition (π_op*(O-O) → Mn(d_yz)) energies (∼450 nm) previously reported for Mn(III)–peroxo complexes [Mn^III(O₂)(L₇py₂^R)]²⁺.³⁴ As shown in Figure 2, isotopically sensitive resonance-enhanced vibrations for 3, not seen with 1 and 2, are observed at ν_O-O(Δ¹⁸O) = 819(47) cm⁻¹ and ν_Mn-O(Δ¹⁸O) = 611(25) cm⁻¹, consistent with a metal– peroxo species. This indicates that two electron reduction of O₂ has occurred, and the dioxygen O-O bond has been activated. The ν_O-O stretching frequency of 3 is significantly lower than the few reported ν_O-O frequencies (IR) for monomeric side-on Mn(III)-peroxo compounds such as [Mn(3,5-iPr₂pzH){(HB(3,5-iPr₂pz)₃}(O₂)], and [Mn(H₃bupa)(O₂)]⁻ (Table 2).^35,37 A comparison of force constants (vide infra) and an analysis of the coupling between O–O and Mn–O stretches would be necessary in order to comment on the relative extent of O–O bond activation in 3 versus [Mn(3,5-iPr₂pzH){(HB(3,5-iPr₂pz)₃}(O₂)], and [Mn(H₃bupa)(O₂)]⁻, however this data is not available for the side-on Mn(III)-peroxo compounds. Given the known mechanisms for O–O bond activation,^27-29 this would suggest that the peroxo of 3 is either protonated or coordinated in a bridging mode. The reproducibly observed 2:1 ratio of H₂O₂: Mn would be more consistent with the latter, and the one example of a vibrationally characterized multinuclear Mn-peroxo compound,⁴⁹ [Mn^III₃(dien)₃(μ-OAc)₂(μ₃-O)(cis μ-1,2-O₂)]³⁺, also displays a low ν_O-O frequency (814 cm⁻¹) consistent with this. Although they are very different in terms of electronic structure, side-on binuclear μ-η²: η² Cu peroxo compounds display ν_O-O frequencies noticeably lower than that of 3 (Table 2),^25,56,57 whereas end-on binuclear Cu peroxo complexes display ν_O-O frequencies comparable to that of 3 (Table 2). Binuclear end-on Fe(III)-peroxos species,^55,58-60 including intermediates formed in RNR and MMO,²⁰ however, show significantly higher ν_O-O stretching frequencies (843 – 908 cm⁻¹) relative to 3 (Table 2). This has been attributed to the large Fe-O-O angles associated with these binuclear Fe(III)-peroxo compounds which induces Fe-O mixing into the O-O bond.⁵⁵ The ν_O-O stretching frequency for 3 (819 cm⁻¹) does fall in the range (781-844 cm⁻¹) of mononuclear Fe(III) hydroperoxo compounds,²⁹ where there is no Fe-O mixing into the O-O vibration due to the lack of mechanical coupling.^29,55 As far as vibrations involving the Mn-peroxo bond are concerned, experimentally determined ν_Mn-O data is not available for comparison.

Figure 2.

Resonance Raman spectrum of the metastable intermediate 3, formed in the low-temperature reaction between coordinatively unsaturated [Mn^II(S^Me2N₄(6-MeDPEN))](BPh₄) (1), and ¹⁶O- (green) and ¹⁸O-labled (red) dioxygen (O₂). Counterion (BPh₄⁻) stretch seen at 996 cm^-1.

Verification of the bimetallic nature of peroxo-bound 3 was obtained by X-ray crystallography. Single crystals of the metastable intermediate were obtained by cooling a propionitrile (CH₃CH₂CN) solution of the tetraphenylborate salt of 1 to −80 °C, introducing O₂, and then layering it with Et₂O. As shown by the R-factor (6.1%; Table 1), and estimated standard deviations (Table 3), the structure of 3 is of high-quality. As shown in the ORTEP diagram of Figure 3, [Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)(BPh₄)₂•2CH₃CH₂CN (3) contains an O₂ ligand bridging two manganese ions in an end-on trans configuration (Figure S-3). Each Mn ion is ligated by a thiolate sulfur (S(1)), an O₂-derived oxygen (O(1)), an imine (N(1)) and tertiary amine (N(2)), and two N-heterocyclic amine nitrogens (N(3, 4)). The distance separating the pyridine nitrogens N(4) and N(4′) from the Mn ions (2.492(3) Å) is significantly greater than the sum of their covalent radii (2.105 Å).⁶¹ Most likely this is due to Jahn-Teller-like distortion which allows for the stabilization of an antibonding electron. Steric constraints involving the 6-Me substituent and the gem-dimethyls adjacent to the sulfur likely contribute as well.⁵⁰ The O–O bond length in 3 (1.452(5) Å, Table 3) is significantly longer than that (1.21 Å) of dioxygen (O=O), and close to that of reduced O–O²⁻ (1.49 Å),⁴⁸ consistent with reductive activation of the O–O bond. Dioxygen reduction is accompanied by oxidation of both Mn ions in 3 (from the +2 to the +3 oxidation state), as shown by comparison of the Mn–X (X= N, S) bond lengths with those of the reduced starting material 1 and final oxidized product 2 (Table 3).⁵⁰ The Mn(III)-S bond (2.2747(12) Å) of 3 is 0.096 Å shorter than that of reduced 1, but comparable to that of oxidized mono-oxo bridged {[Mn^III(S^Me2N₄(6-MeDPEN)]₂-(μ-O)}²⁺ (2), {[Mn^III(S^Me2N₄(2-QuinoEN)]₂(μ-O)}²⁺ (2.292(1) Å), [Mn^III(S^Me2N₄(6-Me-DPPN))]₂(μ-O)(BPh₄)₂ (2.255(2) Å), and {[Mn^III(S^Me2N₄(6-H-DPEN)]₂(μ-O)}²⁺ (2.2968(15) Å),⁵⁰ as well as alkylperoxo-ligated [Mn^III(S^Me2N₄(2-QuinoEN)(OOtBu)]⁺ (2.270(3) Å).³³ The peroxo in 3 is best described as binding in an end-on trans–μ–1,2–O₂ fashion (Figure S-3), as opposed to an asymmetric side-on μ-η²: η² fashion as shown by the significantly shorter Mn(1)–O(1)(1.832(3) Å) vs Mn(1)•••O(1′) (2.48 Å) distance. Peroxo-bridged 3 represents the first example of a binuclear Mn(III)-peroxo complex, structurally, or otherwise, characterized. Of the six structurally characterized metastable Mn(III)-peroxo compounds,^{33,36,37,39,40} all but one⁴⁹ are mononuclear, and five contain the peroxo in a side-on (η²) binding mode. The only other multinuclear Mn-peroxo compound, [Mn^III₃(dien)₃(μ-OAc)₂(μ₃-O)(cis μ-1,2-O₂)]³⁺,⁴⁹ contains a bridging cis μ-1,2-peroxo. Aside from this cis μ-1,2-peroxo compound, the only other crystallographically characterized end-on manganese-peroxo compound contains an alkyl-peroxo ligand.³³ The Mn–O(1) peroxo bond of 3 is noticebly shorter than all other reported structurally characterized Mn(III)-peroxo compounds (reported range: Mn–O= 1.850(6)–1.901(4) Å).^{33,36,37,39,40} The acute Mn-O(1)-O(2) angle (97.5(6)˚) indicates that the Mn–O(1) peroxo bond of 3 does not have significant double bond character, however. The peroxo O–O bond of 3 (1.452(5) Å), is significantly longer (Table 2) than all of the structurally characterized side-on Mn(III)-η²-O₂²⁻ (reported range: O–O= 1.410(4)–1.428(7) Å),^36,37,39 but similar in length to the only structurally characterized end-on peroxo compounds, [Mn^III(S^Me2N₄(QuinoEN))(OO^tBu)]⁺ (1.457(7) Å),³³ and within the error limits of [Mn^III₃(dien)₃(μ-OAc)₂(μ₃-O)(cis μ-1,2-O₂)]³⁺ (1.6(1) Å).⁴⁹ The long peroxo bond length of 3 implies that the O–O bond should readily cleave, consistent with its observed conversion to binuclear mono-oxo bridged 2. Comparison of the peroxo O–O bond length in binuclear Mn-3 with selected examples of binuclear Fe and Cu peroxo complexes (Table 2) illustrates how the O–O bond length in 3 falls at the high end of that observed in the majority of reactive binuclear metal-peroxo complexes regardless of the peroxo binding mode. This suggests that the peroxo in 3 is highly activated. As shown by a recent analysis of the factors which influence the extremely long (1.540(5) Å) peroxo bond in [Cu₂(MeAN)₂(O₂)]²⁺,⁵⁷ a variety of factors (i.e. extent of overlap with the peroxo σ* vs π*) influence the peroxo O–O bond length which do not necessarily reflect the extent of O-O bond activation. Another physical parameter which is predictive of the level of O–O bond activation is the vibrational force constant. The force constant associated with the ν_O-O of 3 (k= 3.02 mdyn/Å), calculated using normal coordinate analysis, is comparable to those of reactive binuclear Fe– and Cu–peroxo compounds, [Fe₂(cis-μ,-1,2-O₂)(OBz)₂{HB(pz′)₃}₂] (k= 3.1 mdyn/Å) and [{Cu(TMPA)}₂(trans-μ-1,2-O₂)]²⁺ (k= 3.103 mdyn/Å),^55,62 suggesting that 3 should be reactive as well. The experimentally observed ν_O-O stretching frequencies associated with the above Fe- and Cu-peroxo compounds (832 and 876 cm⁻¹),^55,62 on the other hand, are significantly higher than that of 3 (due to the mechanical coupling between the M-O and O-O vibrations) showing that ν_O-O stretching frequency can not be used alone to predict levels of O–O bond activation.⁵⁷ The low-temperature (−40 °C) half-life of 3 both in the presence (τ_1/2= 11(2) sec; Figure S-4) and absence (τ_1/2= 14(2) sec) of cyclohexane carboxaldehyde indicates 3 is unreactive towards electrophiles.^35,43 This is in contrast to side-on mononuclear peroxo complexes, [Mn^III(tmc)(O₂)],⁴³ and [Mn^III(H₃bupa)(O₂)]¹⁻.³⁵ This lack of reactivity is perhaps not surprising, however, given the bridging, and therefore inaccessible and less nucleophilic, nature of the peroxo in 3. As shown by the space–filling diagram of Figure 4, the aromatic pyridine rings provide a protective cavity for the peroxo, thus explaining the stability (albeit limited) of what appears to be a highly activated O–O bond. The aromatic rings are roughly coplanar (deviation= 7.95°) with an average C•••C separation of 4.16 Å, suggesting that π-stacking interactions might help to stabilize the structure. The crystallographic data show that the Mn³⁺ ions in 3 are significantly more separated (4.113 Å) than would be expected for a bridging side-on μ–η²:η²–peroxo (∼3.5 Å),^23,25,26,47 which is again consistent with an end-on trans–μ–1,2–peroxo description. This is a much a less commonly observed peroxo binding mode.⁶³ Fits to the magnetic susceptibility vs temperature (J= 0 cm⁻¹, D= −1.6 cm⁻¹; E/D= 0) curves (Figure 5A), and inverse magnetic susceptibility vs temperature data (μ_eff= 9.37 BM; Figure 5B), indicate that the Mn³⁺ ions of 3 are magnetically uncoupled and high–spin (S= 2). Peroxo ligands often mediate strong exchange coupling,^24,55,64,65 however DFT single point calculations on the crystal structure of 3 reproduce the low exchange coupling, and show that this mainly reflects the fact that the Mn d orbital forming a strong σ-bond with the peroxide π* orbital is unoccupied. Also, the π-interaction between peroxide and Mn d orbitals is very weak and therefore does not contribute to coupling between the two metal centers.

Table 3.

Selected Bond Distances (Å) and Bond Angles (deg) for Five Coordinate [Mn^II(S^Me2N₄(6-Me-DPEN)](PF₆) (1), oxo bridged [Mn^III(S^Me2N₄(6-MeDPEN)]₂-(μ-O)(BF₄)₂•2MeOH (2), Peroxo-Bound [Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)(BPh₄)₂ •2CH₃CH₂CN (3)

	1	2	3
Mn(1)–S(1)	2.3710(6)	2.2767(7)	2.2747(12)
Mn(1)–N(1)	2.186(2)	1.999(3)	2.040(3)
Mn(1)–N(2)	2.297(2)	2.151(2)	2.203(3)
Mn(1)–N(3)	2.222(2)	2.581(2)	2.410(3)
Mn(1)–N(4)	2.239(2)	2.501(2)	2.492(3)
Mn(1)–O(1)	N/A	1.7602(4)	1.832(3)
O(1)–O(1′)	N/A	N/A	1.452(5)
Mn(1)•••O(1′)	N/A	N/A	2.481
Mn(1)•••Mn(1′)	N/A	3.520	4.113
Mn(1)-O(1)-O(1′)	N/A	N/A	97.5(6)
O(1)-Mn(1)-S(1)	N/A	95.37	85.12(9)
O(1)-Mn(1)-N(3)	N/A	92.79	92.4(1)
O(1)-Mn(1)-N(1)	N/A	177.21(9)	165.5(1)
O(1)-Mn(1)-N(2)	N/A	100.44(6)	111.1(1)
S(1)-Mn-N(2)	156.98(5)	164.17(7)	163.3(1)
S(1)-Mn-N(3)	117.36(5)	106.2 (1)	109.0(1)
S(1)-Mn-N(4)	117.38(5)	106.84(6)	105.6(1)
N(1)-Mn-N(4)	120.58(7)	92.68(1)	101.7(1)
Mn(1)–O–Mn(1′)	N/A	180.0	N/A

Figure 3.

ORTEP of peroxo-bridged {[Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)}²⁺(3) showing 50% probability ellipsoids and the atom labeling scheme. Hydrogens have been omitted for clarity.

Figure 4.

Space filling diagram of peroxo-bridged {[Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)}²⁺ showing that the peroxo (red) is nestled in a protective cavity provided by the pyridine rings.

Figure 5.

Magnetic susceptibility vs temperature data (A, blue diamonds) obtained for a solid sample of peroxo-bridged [Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2-O₂)(BPh₄)₂•2CH₃CH₂CN (3). Fits to the chi vs T data (A; pink line; J= 0 cm^-1, D= −1.6 cm^-1, and E/D= 0)⁶⁶, as well as the inverse magnetic susceptibility vs T data (B) are consistent with high–spin (S= 2) magnetically uncoupled (down to 5 K) Mn³⁺ ions in an axial environment.

Kinetics of the Reaction Between [Mn^II(S^Me2N₄(6-MeDPEN))](BPh₄) (1) and O₂

In order to determine the mechanism by which peroxo-bridged [Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)²⁺ (3) forms, we examined the kinetics of low-temperature reactions between [Mn^II(S^Me2N₄(6-MeDPEN))](BPh₄) (1) and dioxygen. Given the widely accepted proposed mechanisms for binuclear metal-peroxo and μ-oxo formation,⁶⁷ the likely mechanism was hypothesized to involve a mononuclear Mn-O₂ intermediate. An intermediate (with λ_max = 515 nm) is in fact observed to form en route to peroxo 3 on the msec time-scale using a stopped-flow instrument. This intermediate, [Mn(S^Me2N₄(6- MeDPEN))(O₂)]⁺ (4; Scheme 2), which is presumed to be an η¹- superoxo, accumulates rapidly (within fractions of a second) in MeCN at low temperatures, and then more slowly converts to μ-peroxo-bridged 3 (λ_max = 640 nm). Kinetics for both of these steps (shown in Scheme 2) were measured by the stopped-flow technique over the temperature range −40 °C to − 10 °C. A representative spectral overlay is shown in Figure 6, and the general kinetic scheme is shown in equations 1 and 2:

Scheme 2.

Figure 6.

Time resolved spectral changes obtained upon mixing acetonitrile solutions of Mn(II) complex 1 (1.5 mM) and O₂ (4.05 mM) at −10 °C. Inset: kinetic trace (λ = 515 nm) showing formation and subsequent decay of the Mn-O₂ intermediate 4. All reported concentrations are after mixing in the stopped-flow cell.

$$[{\text{Mn}}^{\text{II}}]+{\text{O}}_2\stackrel{k_1}{\to }[{\text{Mn}}^{\text{III}}-{\text{O}}_2]$$ (1)

$$[{\text{Mn}}^{\text{III}}-{\text{O}}_2]+[{\text{Mn}}^{\text{II}}]\stackrel{k_2}{\to }[{\text{Mn}}^{\text{III}}-{\text{O}}_2-{\text{Mn}}^{\text{III}}]$$ (2)

$$\mathit{\text{rate}}={\mathit{\text{ae}}}^{-k_{1\mathit{\text{obs}}}t}+{\mathit{\text{be}}}^{-k_{2\mathit{\text{obs}}}t}$$ (3)

Kinetic traces (λ = 515 nm) obtained under pseudo first-order conditions with excess O₂ (Figure 6) were fit to the general two-term rate law shown in equation (3), which represents a general solution to the mechanism outlined in equations (1) and (2). Pseudo first-order rate constants, k_1obs and k_2obs, were obtained (see Supplemental material) by fitting the kinetic traces to equation (3). The observed rate constant associated with formation of intermediate 4, k_1obs, was found to increase linearly with increasing O₂ concentrations, allowing us to determine the second order rate constant (k₁) from the slope of k_1obs vs [O₂] plots. Rate constants were obtained in this manner at four different temperatures (Table 4). The intercepts of the k_1obs versus O₂ plots were found to be close to zero (Figure 7) indicating that reaction (1) in Scheme 2 is essentially irreversible. The observed rate constant, k_2obs associated with the conversion of superoxo-4 to peroxo-3 (rxn (2) in Scheme 2; eqn (2)), on the other hand, was found to be essentially independent of O₂ concentration over the entire temperature range studied (Figure S-5). This indicates that the rate of formation of the μ-peroxo 3 is zero order in [O₂]

Table 4.

Experimentally Measured Temperature-Dependent Rate Constants for the Formation of η¹-Superoxo 4 (k₁) and Binculear Peroxo-Bridged 3 (k₂) in the Reaction Between [Mn^II(S^Me2N₄(6-MeDPEN))](BPh₄) (1) and O₂ in MeCN.

Temperature (°C)	k1 (M−1 s−1)	k2 (M−1 s−1)
− 40 °C	717 ± 45	22. 7 ± 0. 2
− 30 °C	1151 ± 24	60. 7 ± 0. 3
− 20 °C	2106 ± 259	153. 6 ± 1. 5
− 10 °C	3779 ± 180	416.9 ± 3.2

Figure 7.

Plots of the observed rate constants (k_1obs) for the formation of superoxo intermediate 4 versus O₂ concentration over the temperature range − 40 °C to − 10 °C in the reaction between Mn(II) complex 1 (1.5 mM) and O₂ in CH₃CN. The slope of each line represents the value of the second-order rate constant for superoxo 4 formation (k₁) at each temperature.

In order to determine the reaction order with respect to Mn(II) for steps (1) and (2) of Scheme 2 at −30 °C, the [Mn^II(S^Me2N₄(6-MeDPEN))]⁺(1) concentration was varied (from 0.8, 1.1, 1.5 to 2.0 mM after mixing) while maintaining a constant amount of excess oxygen (4.1 mM after mixing). The rates of formation of η¹-superoxo 4 and binuclear peroxo-bridged 3 were determined as described above, by following the build-up and decay of 4 at λ = 515 nm. The observed rate constants (k′_1obs and k′_2obs) were obtained by fitting the kinetic traces to equation (3). The observed pseudo-first-order rate constant for η¹-superoxo intermediate 4 formation (k′_1obs) was found to be independent of [Mn(II)] in the presence of excess oxygen, consistent with first-order behavior with respect to [Mn(II)] for step (1) (Figure S-6). Observed rate constants for the formation of μ-peroxo 3 (k′_2obs) showed a linear dependence on [Mn(II)] in excess oxygen, confirming second-order behavior with respect to [Mn(II)] overall (Figure 8). This supports the reaction mechanism outlined in equations (1) and (2) and Scheme 2. Taken together, the [O₂]- and [Mn(II)]-dependency

Figure 8.

Plot of the observed rate constants (k′_2obs) for formation of the binuclear peroxo complex 3 versus [Mn(II)L](BPh₄) at − 30 °C in CH₃CN. [O₂] after mixing = 4.1 mM.

$${\mathit{\text{rate}}}_1=k_1\left[\mathit{\text{Mn}}\right(\mathit{\text{II}}\left)\right]\left[O_2\right]$$ (4)

$${\mathit{\text{rate}}}_2=k_2\left[\mathit{\text{Mn}}\right(\mathit{\text{III}}\left)O_2\right]\left[\mathit{\text{Mn}}\right(\mathit{\text{II}}\left)\right]$$ (5)

experiments yield the overall rate law for the formation of superoxo-4 (eqn (4)) and peroxo-3 (eqn (5)), respectively. Based on the effective second-order dependence on [Mn(II)] for μ-peroxo 3 formation (eqn 5), the decay event in kinetic traces at 515 nm was refit to a second-order equation (eqn (6)) from which k₂

$${\mathit{\text{rate}}}_2={\mathit{\text{Abs}}}_{\mathit{\text{final}}}-\left[\right({\mathit{\text{Abs}}}_{\mathit{\text{final}}}-{\mathit{\text{Abs}}}_0)/(1+k_2\ast {\left[\mathit{\text{Mn}}\right(\mathit{\text{II}}\left)\right]}_0\ast t\left)\right]$$ (6)

was obtained, and used for the calculation of activation parameters. The second-order rate constants k₁ and k₂ are summarized over the temperature range − 40 °C to − 10 °C in Table 4. Activation parameters for the formation of superoxo 4 (Δ H₁^‡ = 26.4±1.7 kJ mol⁻¹, ΔS₁^‡ = −75.6±6.8 mol⁻¹ K⁻¹) and μ-peroxo 3(ΔH₂^‡ = 47.1±1.4 kJ mol⁻¹, ΔS₂^‡ = − 15.0±5.7 J mol ⁻¹ K ⁻¹) were obtained from the Eyring plots shown in Figure S-7 The negative entropy of activation would be consistent with an associative process in both steps. The relatively small absolute value of the activation entropy, especially for the second step, may be attributed, at least in part, to desolvation of one (or both) reactants.

Dioxygen binding to iron(II), cobalt(II), and copper(II) is particularly relevant to bioinorganic chemistry, and synthetic model complexes provide important mechanistic insights.^51,68,79 Manganese dioxygen chemistry is, on the other hand, less well developed,^32,35 and no kinetic information is available for comparison to our results. The overall reaction pathway of oxygenation of [Mn^II(S^Me2N₄(6-MeDPEN))]⁺ (1) is similar to well-established mechanisms of dioxygen reactivity with other biologically-relevant first-row transition-metal complexes.^{51,68-74,80-82} Initial coordination of dioxygen to a mononuclear metal complex (Table 5) is often followed by the formation of a dinuclear peroxo species, which in many cases undergoes O-O bond cleavage, to afford either a high-valent mononuclear metal-oxo (M=O) species (Scheme 2) or a high-valent bimetallic bis μ-oxo (M(μ-O)₂M) species.^{8,26,67,75,83} Steric bulk may, in some cases, stabilize the mononuclear dioxygen adducts and prevent the formation of a dinuclear peroxo complex.^47,71,80,84 The relative rates of individual steps tend to vary as a function of the metal ion and ligand. Kinetic data for O₂ binding to selected examples of aminopyridine or polyamine ligated copper(I), cobalt(II), or iron(II) complexes are listed in Table 5. Although not directly comparable to Mn(II)-containing 1 due to differences in ligand structure and solvent, the activation barrier to O₂ binding to five-coordinate 1 is notably higher than that of coordinatively unsaturated complexes (Table 5). It is comparable, on the other hand, to that of six-coordinate diiron(II) complexes, such as [Fe₂(HPTP)(O₂CPh)₂]⁺, [Fe₂(OH)₂(TPA)₂]²⁺, or [Fe₂(OH)₂(BQPA)₂]²⁺.^51,86 Oxygen binding to six-coordinate iron(II) further decrease upon increasing the steric bulk of the ligand (e.g., [Fe₂(OH)₂(6-Me₃TPA)₂]²⁺).^75,86 Oxygen binding to coordinatively unsaturated Co(II) or Cu(I) is even more rapid (Table 5),^76,87-89 but it can be also retarded by steric hindrance (e.g. [Cu(HIPT₃)tren]⁺).⁹⁰

Table 5.

Comparison of Kinetic Parameters for Dioxygen Binding to Mn(II), Fe(II), Cu(I) and Co(II) Complexes.

Complex	k1, M−1 s−1	ΔH‡, kJ/mol	ΔS‡, J/K mol	Ref
[MnII(SMe2N4(6-Me-DPEN)]+ (1)	3.78(18) × 103 (233 K)a	26(2)	−76(7)	This work
[FeII2(HPTP)(O2CPh)]2 +	9.6(3) × 102 (233 K)c	16.5(4)3	−114(2)	(85)
[FeII (OH)2 (TPA)2]2+	1.21(6) × 101 (233 K)b	30(4)4	−94(10)	(60)
[FeII2(OH)2(6-Me3TPA)2]2+	1.94(11) (233 K)a	16(2)	−167(10)	(60), (86)
[FeII2(OH)2(BQPA)2]2+	3.2(2) (233 K)b	36(4)	−80(10)	(60)
[FeII2(OH)2(BnBQA)2]2	2.67(10) × 103 (233 K)a	16(2)	−108(10)	(60)
[CuI(Me6tren)(EtCN)]+	9.5(4) × 104 (183 K)c2	17.1(6)	−52(3)	(93)
[CuI(TPA)]+	1.50(2) × 108 (193 K)d1	7.62	−45	(76), (87), (88)
[CuI(TPA)(EtCN)]+	5.0(3) × 105 (223 K)c	31.6(5)	10.0(3)	(76)
[CuI(HIPT)3tren]+	2.3(2)e	N/A	N/A	(90)
[CoII(SalMeDPT)	8.8 × 104 (203 K)e	12.7	−80	(89)

^aMeCN.

^bCH₂Cl₂.

^cCH₃CH₂CN.

^dTHF.

^eAcetone.

Conclusions

In conclusion, we describe a rare instance in which two metastable intermediates are observed in the low-temperature dioxygen chemistry of a coordinatively unsaturated manganese complex, [Mn^II(S^Me2N₄(6-MeDPEN))] ⁺ (1). Although oxygen binding to biologically relevant Fe(II), Cu(I), and Co(II) has been extensively studied,^62,91,92 the observation of more than one intermediate in these processes is relatively rare.^76,87,88,93 Manganese dioxygen chemistry, on the other hand, remains largely unexplored.^32,35,40 Low-temperature stopped-flow experiments reported herein show that a dioxygen intermediate, [Mn^III(S^Me2N₄(6-MeDPEN))(O₂)]⁺ (4), rapidly forms upon the addition of O₂ to coordinatively unsaturated 1 consistent with O₂ binding to an open coordination site. The thiolate ligand helps to maintain an open coordination site.⁹⁴ O₂ binding to 1 is found to be relatively slow compared to related biologically relevant Fe and Cu complexes. In the absence of any other kinetic data for O₂ binding to Mn(II), it would be difficult to draw any conclusions regarding the role of the metal ion in this reaction. Dioxygen intermediate 4 was shown to convert more slowly to a species, 3, with isotopically sensitive stretches in the resonance Raman spectrum (ν_O-O(Δ¹⁸O) = 819(47) cm⁻¹, k_O–O= 3.02 mdyn/Å, ν_Mn-O(Δ¹⁸O) = 611(25) cm⁻¹) consistent with a peroxo. The frequency of this ν_O-O stretch is compared with that of related binuclear Cu- and Fe-peroxo compounds. Roughly ½ equiv H₂O₂ was shown to be released per Mn ion of 3, and the rate of conversion of 4 to 3 was shown to be dependent on [Mn(II)] concentration, consistent with its formulation as a binuclear Mn-peroxo. This was verified by X-ray crystallography, where the peroxo of {[Mn^III(S^Me2N₄(6-Me-DPEN)]₂(trans–μ–1,2–O₂)}²⁺ (3) is shown to be bridging between two Mn(III) ions in an end-on trans-μ-1,2-fashion. This represents the first characterized example of a binuclear Mn(III)-peroxo similar to intermediates proposed to be involved in both H₂O oxidation and DNA biosynthesis, arguably two of the most fundamental processes of life. Neither of these biological intermediates has been observed. The peroxo O–O bond of 3 is shown to be is considerably longer than that of all crystallographically characterized mononuclear side-on peroxo compounds, and fall at the high end of reported bond lengths for binuclear Cu- and Fe-peroxo compounds, regardless of the peroxo binding mode.