Synthesis and Structural Characterization of a Series of Mn(III)-OR Complexes, Including a Water-Soluble Mn(III)-OH that Promotes Aerobic Hydrogen Atom Transfer

Abstract

Hydrogen atom transfer reactions (HAT) are a class of proton-coupled electron transfer (PCET) reactions used in biology to promote substrate oxidation. The driving force for such reactions depend on both the oxidation potential of the catalyst and the pK_a of the proton acceptor site. Both high-valent transition-metal oxo M(IV)=O (M= Fe, Mn) and lower-valent transition-metal hydroxo compounds M(III)–OH (M= Fe, Mn) have been shown to promote these reactions. Herein we describe the synthesis, structure and reactivity properties of a series of Mn(III)-OR compounds (R= ^pNO₂Ph(5), Ph(6), Me(7), H(8)), some of which abstract H-atoms. The Mn(III)-OH complex 8 is water-soluble and represents a rare example of a stable mononuclear Mn(III)-OH. In water, the redox potential of 8 was found to be pH-dependent and the Pourbaix (E_p,c vs pH) diagram has a slope (52 mV/pH) that is indicative of the transfer a single proton with each electron (ie, PCET). The two compounds with the lowest oxidation potential, hydroxide and methoxide-bound 7 and 8 are found to oxidize TEMPOH, whereas the compounds with the highest oxidation potential, phenol-ligated 5 and 6, are shown to be unreactive. Hydroxide-bound 8 reacts with TEMPOH an order of magnitude faster than methoxide-bound 7. Kinetic data (k_H/k_D= 3.1 (8), k_H/k_D= 2.1 (7)) are consistent with concerted H-atom abstraction. The reactive species 8 can be aerobically regenerated in H₂O, and at least 10 turnovers can be achieved without significant degradation of the “catalyst”. The linear correlation between redox potential and pH, obtained from the Pourbaix diagram, was used to calculate the BDFE= 74.0±0.5 kcal/mol for Mn(II)-OH₂ in water, and in MeCN its BDFE was estimated to be (70.1 kcal/mol). The reduced protonated derivative of 8, [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9), was estimated to have a pK_a of 21.2 in MeCN. The ability (7) and inability (5 and 6) of the other members of the series to abstract a H-atom from TEMPOH was used to estimate either an upper or lower limit to the Mn(II)-O(H)R pK_a based on their experimentally determined redox potentials. The trend in pK_a (21.2(R=H) > 16.2(R=Me) > 13.5(R=Ph) > 12.2(R=^pNO₂Ph)) is shown to oppose that of oxidation potential E_p,c (−220(R= ^pNO₂Ph) > −300(R= Ph) > −410(R= Me) > −600(R= H) mV vs Fc^+/0) for this particular series.

Introduction

Hydrogen atom (H•) transfer reactions (HAT) are a class of proton-coupled electron transfer (PCET) reactions used in biology to promote substrate oxidation. ^1,2 Typically these reactions are catalyzed by reactive intermediates formed upon the addition of dioxygen (O₂) to reduced transition metal ions (e.g., Fe, Cu, Mn). When strong C–H bond are involved, one would anticipate that the reactive catalyst would have to be highly oxidizing and therefore high-valent, e.g., M(IV)=O (M= Mn, Fe).^3-6 In some cases, however, a lower-valent M(III)-OH (M= Mn, Fe) has been shown to activate C-H bonds, an example of which is lipoxygenase (Figure 1). Lipoxygenases convert α-linoleic acid acid to hydroperoxy-octadecatreienoic acid via the hemolytic cleavage of an, albeit weaker (BDE~ 77 kcal/mol),⁷ allylic C–H bond. The thermodynamic driving force for oxidation reactions of this type depends on the relative strength of the substrate X-H (X= C, N, O) versus catalyst M(II)(HO–H) or M(III)(O–H) bond.^1,8-12 One can relate this driving force to both the redox potential (E_1/2(M(III)/M(II)) or E_1/2(M(IV)/M(III))), and basicity of the M(III)–OH or M(IV)=O oxygen.¹ As long as the coordinated oxygen is basic enough, then the oxidant can be rather mild (ie possess a low E_1/2) and still capable of promoting hydrogen atom abstraction. The increased basicity of a lower valent M(III)-OH oxygen vs higher valent M(IV)=O oxygen could conceivably offset the expected decrease in redox potential caused by the decrease in oxidation state, and this might explain the ability of M(III)-OH species to abstract H-atoms. A recent study addresses this point by comparing the thermodynamics and kinetics of H-atom abstraction involving structurally analogous synthetic LMn(IV)(O)₂, LMn(IV)(O)(OH)⁺, LMn(IV)(OH)(OH)²⁺, LMn(IV)(OH)(OH₂)³⁺ complexes.^13-16 This study shows that the H-atom abstracting capability of Mn(IV)=O and Mn(IV)-OH is approximately the same, due to the fact that a decrease in oxidation potential is offset by an increase in pK_a. This compensation effect is also observed upon sequential deprotonation of coordinated water molecules. The high-valent Mn(IV)=O was shown to react ~ 40 times faster than the hydroxo Mn(IV)-OH, however. Thiolate ligands have also been shown to facilitate biological H-atom transfer by creating a more basic oxo.^17-20 We are interested in developing these ideas further by exploring the H• atom transfer reactivity of thiolate-ligated small molecule systems. Herein, we describe the synthesis and properties of a series of thiolate- ligated Mn(III)-OR compounds. By altering the R-group, we are able to influence the pK_a of the coordinated ROH, and determine how this influences H-atom transfer capability. Included in this study is the reactivity of a rare example of a water-soluble thiolate-ligated Mn^III-OH that abstracts H• from organic substrates. A thermochemical cycle is used to obtain quantitative data, and determine the relative influence of redox potential versus pK_a in driving the observed HAT reaction in both water and acetonitrile.

Figure 1.

Reaction scheme showing the mechanism of lipoxygenase substrate oxidation which involves abstraction of an allylic H-atom from substrate by a catalytically active M(III)-OH (M= Mn, Fe)

Experimental

General Methods

All manipulations were performed using Schlenk line techniques or under a N₂ atmosphere in a glovebox. Reagents and solvents were purchased from commercial vendors were of highest available purity and were used without further purification unless otherwise noted. MeOH (Na) 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). Infrared spectra were recorded as nujol mulls on NaCl salt plates with a Perkin Elmer 1720 FT-IR spectrometer. 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. EPR spectra were recorded on a Varian CW-EPR spectrometer between 5-7 K equipped with an Oxford helium cryostat. 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. Solution magnetic moments were calculated by the Evans method,²¹ with temperature correction made in the manner described by Van Geet.²² Cyclic voltammograms were recorded in either MeCN (100 mM Buⁿ₄N(PF₆) supporting electrolyte) or H₂O (100 mM KClO₄ supporting electrolyte) on a PAR 263A potentiostat utilizing a glassy carbon working electrode, platinum auxillary electrode, and a Ag⁺/Ag reference electrode. pH measurements were made using Beckman Coulter 400 series handheld meter. X-ray crystallography data was recorded on either a Bruker APEX II single crystal X-ray diffractometer with Mo-radiation. Elemental analyses were performed by Atlantic Microlabs, Norcross, GA. TEMPOH²³ and [Mn^II(S^Me2N₄(tren))](PF₆) (1) ²⁴ were prepared according to literature procedures.

Synthesis of [Mn^III(S^Me2N₄(tren))₂-(μ-O)](PF₆)2•MeCN (2)

An anaerobic MeCN solution (5 mL) of [Mn^II(S^Me2N₄(tren))](PF₆) (1)²⁴ (500 mg, 1.12 mmol), containing the S^Me2N₄(tren) ligand, was prepared in a drybox. The solution was removed from the drybox and exposed to air for 5 minutes. Removal of all volatiles in vacuo afforded the title compound as a purple solid in nearly quantitative yield (499 mg, 0.55 mmol). Electronic absorption (MeCN; λ_max (nm) (ε (M⁻¹cm⁻¹))): 535 (510), 600 (380). Magnetic moment (299 K, solution state, CD₃CN): 2.86 B.M. per Mn. Elemental analysis for C₂₂H₅₀N₈OF₁₂P₂S₂Mn₂; Calculated: C, 29.15; H, 5.56; N, 12.36. Found: C, 28.19; H, 5.40; N, 12.28.

Synthesis of [Mn^III(S^Me2N₄(tren)(OPh-^pNO₂))](PF₆)•MeCN (5)

An anaerobic MeCN solution (2 mL) of 2 (100 mg, 0.110 mmol) was prepared in a drybox. An excess of para-NO₂-PhOH (765 mg, 5.50 mmol) was added to as a solid to this solution. After allowing the resulting solution to stir at room temperature for 10 minutes, all volatiles were removed in vacuo to afford a black solid. The resulting solid was recrystallized twice from MeCN/Et₂O (1/5 vol:vol) to afford the title compound as a dark red solid in 95 % yield (122 mg, 0.210 mmol). Electronic absorption (MeCN; λ_max (nm) (ε (M⁻¹cm⁻¹))): 395 (1180), 460 (820). Magnetic moment (299 K, solution state, CD₃CN): 4.80 B.M. Elemental analysis for C₁₇H₂₉N₅O₃F₆PSMn; Calculated: C, 35.00; H, 5.01. Found: C, 35.49; H, 5.21.

Synthesis of [Mn^III(S^Me2N₄(tren)(OPh))](PF₆)•Et₂O•MeCN (6)

An anaerobic MeCN solution (2 mL) of 2 (100 mg, 0.110 mmol) was prepared in a drybox. An excess of PhOH (1.04 g, 11.0 mmol) was added to as a solid to this solution. After allowing the resulting solution to stir at room temperature for 10 minutes, all volatiles were removed in vacuo to afford a dark red solid. The resulting solid was recrystallized twice from MeCN/Et₂O (1/5 vol:vol) to afford the title compound as a red solid in 89 % yield (106 mg, 0.200 mmol). Electronic absorption (MeCN; λ_max(nm)(ε (M⁻¹cm⁻¹))): 506(370), 820(50). Magnetic moment (299 K, solution state, CD₃CN): 4.88 B.M. Elemental analysis for C₁₇H₃₀N₄OF₆PSMn; Calculated: C, 37.92; H, 5.62; N, 10.41. Found: C, 38.00; H, 5.70; N, 10.49.

Synthesis of [Mn^III(S^Me2N₄(tren)(OMe))](PF₆) (7)

An anaerobic MeOH solution (1 mL) of 1 (100 mg, 0.224 mmol) was prepared in a drybox. The solution was removed from the drybox and exposed to air for 5 minutes. Removal of all volatiles in vacuo afforded the title compound as a red solid in nearly quantitative yield (106 mg, 0.222 mmol). Electronic absorption (MeCN; λ_max (nm) (ε (M⁻¹cm⁻¹))): 342(410), 439(330), 510(250), 768(85). Magnetic moment (299 K, solution state, CD₃CN): 4.99 B.M. Elemental analysis for C₁₂H₂₈N₄OF₆PSMn; Calculated: C, 30.26; H, 5.92; N, 11.76. Found: C, 30.29; H, 5.56; N, 11.61.

Synthesis of [Mn^III(S^Me2N₄(tren))(OH)](PF₆)•H₂O (8)

A 1 mL anaerobic aqueous solution of 1 (500 mg, 1.12 mmol) was prepared in a drybox. The solution was then removed from the drybox, opened to air, and allowed to stir at room temperature for 2-3 minutes. Removal of volatiles in vacuo afforded the title compound as a red solid in quantitative yield (518 mg, 1.12 mmol). Single crystals of 8 were grown from a 1/6 mixture of MeCN/Et₂O at 0 °C overnight. Solid 8 was stored at −80 °C to avoid small amounts of decomposition which occurs over a 24 hr period at room temperature. Electronic absorption (H₂O; λ_max (nm) (ε (M⁻¹cm⁻¹))): 287 (3720), 418(229), 489(364), 680(25). Electronic absorption (MeCN; λ_max (nm)(ε (M⁻¹cm⁻¹))): 299(2221), 411(260), 500(320), 803(40). IR (Nujol): ν_O-H = 3367 cm⁻¹, ν_O-D = 2457 cm⁻¹, ν_C=N = 1585 cm⁻¹. Magnetic moment (299 K, solution state, CD₃CN): 4.89 B.M. Elemental analysis for C₁₁H₂₆N₄OF₆PSMn; Calculated: C, 28.58; H, 5.67; N, 12.12. Found: C, 28.43; H, 5.43; N, 11.37.

Reaction Between [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) or [Mn^III(S^Me2N₄(tren)(OMe))](PF₆) (7) and TEMPOH

In a typical reaction, a 1 mM solution of 7 or 8 was prepared in 4 mL of either H₂O or MeCN under an inert atmosphere in a drybox. The solution was transferred via a gas-tight syringe to a custom-made two-neck vial equipped with a septum cap and threaded dip-probe feed-through adaptor that had previously been purged with argon and contained a stir bar. The anaerobic solution of 7 or 8 was continuously stirred while 0.2 equivalent aliquots of a TEMPOH solution (0.1 M stock solution) prepared in the appropriate solvent were added to the reaction mixture. Stirring was discontinued five minutes following the addition of an aliquot of TEMPOH in order to record a UV/Vis absorption spectrum. It was observed that all of the absorption bands characteristic of 7 or 8 decreased in a uniform fashion upon addition of TEMPOH to the reaction mixture, with 1.0 equivalents of TEMPOH resulting in complete disappearance of all visible region absorption features. The resulting spectrum following the addition of 1.0 equivalents of TEMPOH was compared to, and matched, the reported absorption spectrum of the reduced Mn complex, [Mn^II(S^Me2N₄(tren))]⁺.²⁴

Kinetics Measurements of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8)- and [Mn^III(S^Me2N₄(tren)(OMe))]⁺ (7)-Promoted Hydrogen Atom Abstraction from TEMPOH(D)

In a typical experiment, 3 mL of a 0.5 mM solution of 7 or 8 was prepared in a drybox and injected via a gas-tight syringe into a custom-made two-neck vial. The vial was equipped with a septum cap and threaded dip-probe feed-through adaptor that had previously been purged with argon and contained a stir bar. A known excess amount of TEMPOH(D) (0.1 mM stock solution in MeCN) was then injected into the solution containing 7 or 8. Reaction progress was independently monitored by changes in the absorption intensity at 511(8) and 520 nm (7), respectively, until no further change in the absorption intensity persisted for at least 60 seconds. First-order rate constants were calculated by plotting ln[(A_t-A_f)/A₀-A_f)] versus time. Experiments were repeated at least four times at each concentration of TEMPOH(D). Stirring was maintained at a relatively slow rate throughout the duration of each experiment in order to maximize the signal-to-noise ratio in the recorded spectra. Variable temperature measurements for an Eyring plot were performed as described with the exception that the reaction vessel was submerged in a cryogenic bath at a desired temperature. Temperatures were invariant throughout the duration of each experiment.

X-Ray Crystallography

A purple 0.48 × 0.36 × 0.07 mm crystal plate of 2 was mounted on a glass capillary with oil. Colorless plates of 5 (0.15 × 0.10 × 0.10 mm³), 6 (0.25 × 0.10 × 0.05 mm³), and 8 (0.50 × 0.40 × 0.15 mm³) were each mounted on a glass capillary with oil. A red needle of 7, measuring 0.15 × 0.05 × 0.05 mm³ was mounted on a glass capillary with oil. Data was collected for all five compounds on a Bruker APEX II single crystal X-ray diffractometer at −143 °C for 2, and −173 °C for 5-8. For 2, the crystal-to-detector distance was 43.5 mm and exposure time was 60 seconds per degree for all sets. The crystal-to-detector distance for 8 was 40 mm and exposure time was 20 seconds per degree for all sets. The crystal-to-detector distance for 5 and 7 was 40 mm and exposure time was 10 seconds per degree for all sets. The scan width for 2 was 1.0°. The scan width for 5-8 was 0.5°. Data collection for 2 was 99.0% complete to 25.68^o and 99.0% complete to 25^o in ϑ. Data collection for 5 was 99.1% complete to 25.0° in ϑ. Data collection for 6 was 100% complete to 25.0° in ϑ. Data collection for 7 was 96.6% complete to 25.35° in ϑ. Data collection for 8 was 99.5% complete to 28.48° in ϑ. A total of 74378 partial and complete reflections for 2 were collected covering the indices, h = −15 to 15, k = −20 to 20, l = −44 to 44. 7279 reflections were symmetry independent and the R_int = 0.0971 indicated that the data was of less than average quality (0.07). Indexing and unit cell refinement indicated an orthorhombic P lattice. The space group for 2 was found to be P b c a (No.61). For 5, a total of 85,796 partial and complete reflections were collected covering the indices h = −41 to 41, k = −18 to 18, l = −16 to 16. 6,622 reflections were symmetry independent and the R_int = 0.0361 indicated that the data for 5 was good (0.07 average quality). Indexing and unit cell refinement indicated a monoclinic C lattice with the space group C 2/c (No.15). One PF6- molecule was found disordered over two positions, each with 50% occupancy. For 6, 4905 reflections were symmetry independent and the R_int = 0.0195 indicated that the data was good (average quality 0.07). Indexing and unit cell refinement indicated a monoclinic P lattice. The space group for 6 was found to be P 2₁/c (No. 14). A total of 106,880 partial and complete reflections were collected covering the indices h = −11 to 11, k = −22 to 22, l = −29 to 29. 7,668 reflections were symmetry independent and the R_int = 0.0313 indicated that the data for 6 was good (0.07 average quality). Indexing and unit cell refinement indicated a monoclinic P lattice with the space group P 21/n (No.14). The PF6- and MeCN molecules are each slightly disordered in structure 6. For 7, a total of 26,870 partial and complete reflections were collected covering the indices h = −16 to 16, k = −9 to 10, l = −19 to 19. 1,868 reflections were symmetry independent and the R_int = 0.0792 indicated that the data for 7 was of average quality (0.07 average quality). Indexing and unit cell refinement indicated an orthorhombic P lattice with the space group P n m a (No.62). The structure of 7 exhibits disorder across a mirror symmetry plane through the Mn ion, sulfur atom, and imine nitrogen. The PF6- molecule is heavily disordered with two superimposed geometries at 0.66:0.34 occupancies. A total of 67058 (merged) reflections for 8 were collected covering the indices, h = −11 to 11, k = −22 to 22, l = −18 to 18. 4905 reflections were symmetry independent and the R_int = 0.0195 indicated that the data was good (average quality 0.07). Indexing and unit cell refinement indicated a monoclinic P lattice. The space group for 8 was found to be P 2₁/c (No. 14).

Data for 2 was integrated and scaled using hkl-SCALEPACK. Solution by direct methods (SIR97) produced a complete heavy atom phasing model consistent with the proposed structure. All non-hydrogen atoms were refined anisotropically by full-matrix least squares. All hydrogen atoms were located using a riding model. Data for 5 – 8 was integrated and scaled using SAINT, SADABS within the Bruker APEX2 software package by Bruker. Solution by direct methods (SHELXS, SIR97)²⁵ produced a complete heavy atom phasing model for each consistent with the proposed structure. The structures were each completed by difference Fourier synthesis with SHELXL97.^26,27 Scattering factors are from Waasmair and Kirfel.²⁸ 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 were fixed such that they were 1.2U_eq of their parent atom Ueq for 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. For 7, the structure exhibited disorder across a crystallographic mirror symmetry plane containing the manganese, sulfur and nitrogen atom N1. For 8, a co-crystallized water molecule was found to be disordered over two different hydrogen geometries, one in which there is hydrogen-bonding to the fluoride atom F(2) of the PF₆⁻, and a second in which there is hydrogen-bonding to a thiolate sulfur of an adjacent molecule as well as another fluoride of the PF₆⁻.

Results and Discussion

Synthesis and Structure of Mono Oxo Bridged [Mn^III(S^Me2N₄(tren)]₂-(μ-O)(PF₆)₂ (2)

Addition of dioxygen (O₂) to coordinatively unsaturated [Mn^II(S^Me2N₄(tren))](PF₆) (1)²⁴ in MeCN affords a rare example of an unsupported mono-oxo bridged Mn(III) dimer, [Mn^III(S^Me2N₄(tren))]₂-(μ-O)(PF₆)₂ (2). Single crystals of 2 suitable for X-ray diffraction studies were obtained via crystallization from MeCN/Et₂O. As shown in the ORTEP of Figure 2, each Mn center of 2 is six-coordinate and ligated by an oxo trans to an imine and cis to a thiolate sulfur. Although there are numerous examples of μ-carboxylate-μ-oxo and bis-oxo Mn(III) dimers,^29-33 there are significantly fewer examples containing a single, unsupported oxo bridge.^34-37 Thiolate-ligated mono oxo-bridged dimanganese complexes are even more rare, and include pyridine and quinoline derivatives of 2, [Mn^III(S^Me2N₄(6-Me-DPEN))]₂(μ-O)(PF₆)₂•2MeCN (3),³⁵ and [Mn^III(S^Me2N₄(2-QuinoEN))]₂(μ-O)(PF₆)₂ (4).³⁵ The Mn–S(1) bond length of oxo–bridged 2 is 0.13 Å shorter than that of reduced [Mn^II(S^Me2N₄(tren))](PF₆) (1)²⁴ (Table 2), and comparable to those of N-heterocyclic amine ligated oxo-bridged {[Mn^III(S^Me2N₄(6-Me-DPEN)]₂-(μ-O)}²⁺ (3) and [Mn^III(S^Me2N₄(QuinoEN)]₂-(μ-O)}²⁺ (4),³⁵ consistent with an increase in oxidation state from Mn(II) to Mn(III). The trans Mn-N(3) and Mn-N(4) bonds in 2 are elongated somewhat relative to Mn-N(2) reflecting a Jahn-Teller-like distortion, however this distortion is significantly less than that seen with the more sterically encumbered complexes 3 and 4.³⁵ No intermediates are observed in the reaction between 1 and O₂, even at temperatures as low as −80 °C. This is in contrast to the more sterically encumbered, π-accepting N-heterocyclic amine compound [Mn^II(S^Me2N₄(6-Me-DPEN)]⁺ which forms observable Mn-O₂ and peroxo intermediates in its low-temperature reaction with O₂.³⁸ The Mn–O(1) bond to the bridging oxo is slightly longer in 2 (1.791 Å; average) relative to 3 (1.7602(4)Å) and 4 (1.7599(6) Å) indicating that it has slightly less double bond character. Consistent with this, the Mn-O-Mn bridging angle of [Mn^III(S^Me2N₄(tren))]₂-(μ-O)(PF₆)₂ (2) is fairly bent (158.75°), and the S(1)N(1)N(2)O(1)Mn(1) and S(2)N(5)N(6)O(1)Mn(2) planes (Figure S-1) are close to orthogonal (dihedral angle= 69.1°). The bent bridging Mn-O-Mn angle is in contrast to the majority of reported mono oxo bridged dimanganese complexes,³⁶ including 3(180°) and 4 (172.3(2)°),³⁵ which are approximately linear. The only example of a more acute Mn-O-Mn angle (146.15(11)°) is in the unstable oxo-bridged {[Mn^III(S^Me2N₄(6-Me-DPPN))]₂-(μ-O)}²⁺.³⁵ A possible explanation for the acute angle is that, in the absence of steric repulsion between the gem-dimethyls adjacent to the sulfur and the amine ligand, the highly covalent π-donor thiolate ligand would compete more effectively than the oxo for π-overlap with the Mn d-orbitals, resulting in less Mn=O–Mn ↔ Mn–O=Mn double bond character. This would free up a lone pair on the oxo, which would cause the Mn-O-Mn angle to bend. An acute bridging angle would preclude the formation of strong, stabilizing π bonds between the oxo and metal ion, resulting in a more basic oxo. The thiolate ligand would also contribute to an increased basicity. The iron analogue of 2, {[Fe^III(S^Me2N₄(tren))]₂-(μ-O)}²⁺,³⁹ has been shown to readily react with proton donors HA, and has a Fe-O-Fe angle (155.3(5)°) similar to that of 2.

Figure 2.

ORTEP of mono oxo-bridged [Mn^III(S^Me2N₄(tren))]₂-(μ-O)(PF₆)₂ (2) with hydrogen atoms, and counter anions omitted for clarity.

Table 2.

Selected Bond Distances (Å) and Bond Angles (deg) for [Mn^II(S^Me2N₄(tren))](PF₆) (1),²⁴ [Mn^III(S^Me2N₄(tren))]₂-(μ-O)(PF₆)₂•MeCN (2), [Mn^III(S^Me2N₄(6-Me-DPEN))]₂(μ-O)(PF₆)₂•2MeCN (3),³⁵ [Mn^III(S^Me2N₄(2-QuinoEN))]₂(μ-O)(PF₆)₂ (4),³⁵ [Mn^III(S^Me2N₄(tren))(OPh-^pNO₂)](PF₆)•MeCN (5), [Mn^III(S^Me2N₄(tren))(OPh)](PF₆)•MeCN•Et₂O (6), [Mn^III(S^Me2N₄(tren))(OMe)](PF₆) (7), and [Mn^III(S^Me2N₄(tren))(OH)](PF₆)•H₂O (8).

	1	2	3	4	5	6	7	8
Mn-S(1)	2.412(3)	2.286(2)	2.2767(7)	2.292(1)	2.2578(8)	2.2675(3)	2.2937(14)	2.2840(4)
Mn-N(1)	2.166(8)	2.017(6)	1.999(3)	2.010(3)	1.999(2)	2.0043(10)	2.018(4)	2.0133(9)
Mn-N(2)	2.334(8)	2.172(7)	2.151(2)	2.130(3)	2.148(2)	2.1474(10)	2.176(5)	2.1634(9)
Mn-N(3)	2.201(8)	2.351(8)	2.581(2)	2.543(3)	2.288(2)	2.3152(11)	2.295(4)	2.275(1)
Mn-N(4)	2.198(8)	2.280(6)	2.501(2)	2.370(3)	2.300(2)	2.3173(10)	2.295(4)%	2.3322(9)
Mn(1)–O(1)	N/A	1.783(5)	1.7602(4)	1.7599(6)	1.901(2)@	1.8678(8)**	1.836(5)&	1.8540(8)#
Mn(2)–O(1)	N/A	1.799(5)	1.7602(4) $	1.7599(6)	N/A	N/A	N/A	N/A
Mn(1)•••.Mn(2)	N/A	3.521	3.520	3.512	N/A	N/A	N/A	N/A
S(1)-Mn-N(1)	81.9(3)	83.1(2)	82.08(8)	82.60 (8)	83.94(7)	83.26(3)	83.53(12)	83.48(3)
S(1)-Mn-N(2)	158.8(2)	164.6(2)	164.17(7)	163.90(9)	166.11(7)	165.65(3)	166.05(13)	165.40(3)
S(1)-Mn-N(3)	115.0(3)	104.6(2)	106.19(6)	101.40(8)	104.36(7)	98.47(3)	103.42(9)	98.85(3)
S(1)-Mn-N(4)	109.1(2)	102.3(2)	106.84(6)	109.23(8)	103.45(8)	107.23(3)	103.42(9)	106.63(3)
N(1)-Mn-N(3)	125.0(3)	87.1(3)	86.85(6)	94.92(11)	91.19(9)	95.36(4)	90.92(9)	96.12(4)
N(1)-Mn-N(4)	109.5(3)	89.7(3)	92.68(13)	84.75(11)	97.35(9)	87.89(4)	90.92(9)	89.19(3)
N(3)-Mn-N(4)	112.2(3)	152.3(3)	146.57(6)	149.03(10)	151.61(10)	154.30(4)	153.1(2) %	154.41(3)
S(1)–Mn–O(1)	N/A	96.6 (2)	95.37(2)	99.20(3)	98.38(7)@	99.20(3) **	98.78(15) &	96.88(3) #
N(1)–Mn–O(1)	N/A	177.7(3)	177.21(9)	177.02(12)	173.15(9)@	171.55(4) **	173.2(7) &	177.30(4) #
Mn–O(1)–Mn	N/A	158.8(3)	180.0	172.3(2)	N/A	N/A	N/A	N/A
Mn–O(1)–C	N/A	N/A	N/A	N/A	136.8(2)@	134.25(7) **	124.0(9) &	N/A

^#In this case, O(1)= hydroxide oxygen.

^&In this case, O(1)= methoxide oxygen.

^%For this structure N(3) and N(4) are equivalent and related by a crystallographically imposed mirror plane.

^**In this case, O(1)= phenoxide oxygen.

^@In this case, O(1)= ^pNO₂-phenoxide oxygen.

^$For this structure the two halves of the “dimer” are related by crystallographic symmetry so that Mn(1)= Mn(2).

Proton-Induced Cleavage of the μ-oxo Dimer in MeCN

The μ-oxo bridge of 2 is readily cleaved with proton donors ROH (PhOH, p-NO₂-PhOH, MeOH, and H₂O) in MeCN to afford a series of new Mn(III)-OR complexes, [Mn^III(S^Me2N₄(tren))(OR)]⁺ (R= ^pNO₂Ph (5), Ph (6), Me (7), or OH)]⁺ (8)) including a rare example^40,41 of a monomeric Mn(III)-OH compound (Figure 3). Compound 8 is unusual in that it is monomeric (vide infra) despite the lack of steric bulk. The only other reported examples of monomeric Mn(III)-OH compounds contain bulky substituents designed to protect the hydroxide.^40,41 The monomeric structure of each RO-bound derivative described herein, including hydroxide-bound 8, was verified by X-ray crystallography (vide infra). Preliminary evidence to suggest that ROH reacts with {[Mn^III(S^Me2N₄(tren))]₂-(μ-O)}²⁺ (2) involved monitoring the reaction by electronic absorption spectroscopy. A color change, from purple to red, was observed, and peaks grew in at 500 and 411 nm (H₂O; Figure 4; Figure S-2), 439 nm (MeOH; Figure S-3, Figure S-4), 505 nm (PhOH; Figure S-5), or 460 nm (p-NO₂-PhOH; Figure S-6, Figure S-7). These spectral changes involved either a blue-shift, or complete disappearance, of the 520 nm band associated with binuclear 2 (Figure S-8). The number of equivalents of proton donor required for complete conversion was found to vary with the acidity of the proton donor: p-NO₂-PhOH (9 equiv; pK_a= 22 in MeCN)^42,43 and PhOH (15 equiv; (pK_a= 27.2 in MeCN)^42,43, versus H₂O (40 equiv) and MeOH (75 equiv; pK_a>> 33 in MeCN).^42,43 t-Butanol does not react with 2, setting upper and lower limits to the bridging oxo pK_a in MeCN: (pK_a(MeOH) < pK_a(Mn-oxo) < pK_a(tBuOH)). No reaction is observed between 2 and NaOPh-^pNO₂, NaOPh, and NaOMe indicating that these reactions (Figure 3) are proton-dependent. This would be consistent with a mechanism involving initial protonation of the bridging oxo of 2 to form binuclear μ-hydroxo bridged, [Mn^III(S^Me2N₄(tren)]₂-(μOH)(PF₆)₂ (2-H⁺). A μ–OH bridge would be weakened relative to an oxo bridge.⁴⁴ Alkoxide/phenoxide (RO⁻)-induced cleavage of the hydroxo-bridge would then afford one equiv of [Mn^III(S^Me2N₄(tren))(OR)]⁺ (R=^pNO₂Ph (5), Ph (6), or Me (7)) and one equiv of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8). The hydroxo-bridged intermediate, 2-H⁺, is not observed in any of these reactions (Figure 3) indicating that RO⁻-induced cleavage of the μ–OH bridge occurs more rapidly than the initial protonation step. Rate limiting μ-oxo protonation would be consistent with Norton’s previously reported observations.⁴⁵ It’s also possible that the weakened μ–OH bridge cleaves prior to RO⁻ attack, to afford 8 and a solvent-bound Mn(III)-NCMe species, which then reacts with the first equivalent of RO⁻ released during the bridge protonation step to afford 5, 6, or 7. Addition of a second equiv of ROH to the equiv of 8 formed following rupture of the μ–OH bridge would afford a second equiv of 5-7 and H₂ Consistent with this, phenoxy and methoxy derivatives 5, 6, and 7 can also be generated via the addition of ROH to hydroxo-bound 8 (Scheme S-1; Figures S-9 - S-11). Two equiv of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) are formed directly via the addition of H₂O to 2 in MeCN. Again, the number of equivalents of ROH required for complete conversion 8 → 5 (8 eq), 6 (15 eq), or, 7 (120 eq) was found to be dependent on the relative acidity of Mn-OH₂ vs ROH. The ability of proton donors as basic as MeOH to cleave the oxo bridge of 2 contrasts with the lack of reactivity between pyridine- and quinoline-ligated, {[Mn^III(S^Me2N₄(6-Me-DPEN)]₂-(μ-O)}²⁺ (3), and {[Mn^III(S^Me2N₄(QuinoEN)]₂-(μ-O)}²⁺ (4)³⁵ and H₂O or MeOH. This likely reflects the reduced basicity of the oxo atom of 3 and 4 relative to that of primary amine-ligated 2 due to the π-acceptor properties of the pyridine and quinoline ligands and elongated Mn•••N(3,4) bonds. Both of these would create a more Lewis acidic metal center. Although it is ligated by σ-donating primary amines, the μ-oxo bridge of {[Fe^III(S^Me2N₄(tren)]-(μ-O)-[Fe^III(S^Me2N₄(tren)]}²⁺ is only cleaved with more acidic proton donors (HCl, LutNH⁺, TfOH, HOAc). This reflects the increased metal ion Lewis acidity of Fe³⁺ relative to Mn³⁺, which decreases the basicity of the oxo atom. The bridging oxo of pyridine-, carboxamido-ligated [(Mn^III(PaPy₃))2(μ-O)]²⁺ reported by Mascharak and coworkers is also readily protonated by proton donors as basic as MeOH to afford the corresponding monomeric complexes [Mn^III(PaPy₃)(L)]⁺ (L= MeO⁻, PhO⁻, AcO⁻, BzO⁻).⁴⁶ An L= HO⁻ derivative was not reported for this series. Hydroxo-bound [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) is most readily isolated via aerobic oxidation of 1 in aqueous solutions as described below.

Figure 3.

Reaction scheme outlining the observed conversion of oxo-bridged [Mn^III(S^Me2N₄(tren)]₂-(μ-O)(PF₆)₂•MeCN (2) to monomeric para-nitro-phenoxide-, phenoxide-, methoxide-, and hydroxide-ligated 5-8 via the addition of the corresponding ROH in MeCN..

Figure 4.

Conversion of oxo-bridged [Mn^III(S^Me2N₄(tren)]₂-(μ-O)(PF₆)₂•MeCN (2) to monomeric hydroxide-ligated [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) via the addition of 0-40 equiv H₂O in MeCN, as monitored via electronic absorption spectroscopy.

Synthesis of Hydroxo-Bound [Mn^III(S^Me2N₄(tren))(OH)](PF₆)•H₂O (8) in Water

Aerobic oxidation of 1 in H₂O cleanly affords [Mn^III(S^Me2N₄(tren))(OH)](PF₆)•H₂O (8) in high yields (99 %). Hydroxo-bound 8 can also generated by dissolving pre-isolated crystalline samples of μ-oxo 2 in H₂O, indicating that oxo-bridged 2 forms en route to hydroxo-bound 8 (Scheme 1). By varying the aqueous pH and monitoring changes to the electronic absorption spectrum we determined that the Mn(III)-O-Mn(III) bridge of 2 is readily cleaved in solutions as basic as pH=10. This observation implies that the bridging oxo atom of 2 is fairly basic (pK_a ≥ 10). Our inability to observe hydroxo-bridged {[Mn^III(S^Me2N₄(tren))]₂-(μ-OH)}²⁺ (2-H⁺), which is presumed to be an intermediate in the conversion of 2 to 8 did not allow us to more quantitatively determine the pK_a of the oxo atom of 2. Another factor is the limited aqueous solution stability of 8 at extremely basic pHs. Optimum stability (on the order of hours) is observed in the pH range 6 ≤ pH ≤ 8.5. At pH > 10.5, 8 is short-lived, and converts to a black precipitate, presumably Mn₂O₃, or possibly MnO₂ (if disproportionation is involved), within minutes.

Scheme 1.

Comparison of the Crystallographic Structures of Thiolate-Ligated Mn(III)-OR Compounds

The structures of the four Mn(III)-OR compounds described in this study, [Mn^III(S^Me2N₄(tren)(OPh-^pNO₂))](PF₆)•MeCN (5), [Mn^III(S^Me2N₄(tren)(OPh))](PF₆)•MeCN•Et₂O (6), [Mn^III(S^Me2N₄(tren)(OMe)](PF₆) (7) and [Mn^III(S^Me2N₄(tren))(OH)](PF₆)•H₂O (8), were determined by X-ray crystallography. ORTEP diagrams of the more stable complexes are shown in Figure 5, and the slightly less stable, more reactive complexes in Figure 6. All four compounds are mononuclear. In particular, mononuclear Mn(III)–OH hydroxide complexes are extremely rare.^{40, 47-50} The Mn^III ion in all four compounds is pseudo-octahedral, and contains either an ^pNO₂-PhO⁻ (5),PhO⁻ (6), MeO⁻ (7)or OH⁻ (8) trans to an imine nitrogen (N(1)), and cis to a thiolate sulfur. Metrical parameters are compared in Table 2. Comparison of the Mn-S(1) bond lengths in 5-8 with that of reduced [Mn^II(S^Me2N₄(tren))]⁺ (1)²⁴ and oxo bridged 2-4 indicates that the Mn ions are clearly oxidized in 5-8 relative to 1. Although much less exaggerated than with sterically encumbered oxo-bridged 3 and 4, elongation of the trans Mn-N(3) and Mn-N(4) bonds is observed in 5-8 relative to reduced 1. This would also be consistent with each compound containing a Mn⁺³ ion containing an odd antibonding electron. With the exception of hydroxide-bound 8 (vide infra), the Mn-O(1) distance decreases (from 1.901(2) Å in 5 to 1.836(5) Å in 7) as the electron withdrawing ability (via resonance) of the R-group decreases. Resonance between the oxygen lone pairs and the phenyl rings of 5 and 6 removes negative charge from the oxygen, and ties up one of the lone pair that would otherwise be available to interact with protons. This would be expected to influence reactions involving proton addition to the coordinated oxygen (vide infra). The short Mn-O(1) bond distances and Mn-O(1)-C bond angles near 120° (Table 2) indicate that the oxygen O(1) in these structures is sp² hybridized and involved in π-bonding to the metal, consistent with some Mn=OR double bond character. The Mn-O(1)-C angle increases in the order 7 < 6 < 5 indicating that an oxygen lone pair is partially involved in resonance with the aromatic ring in structures 5 and 6 as well resulting in partial O=C^phenyl character. The short aromatic C-O(1) bond distance (1.321(3) Å in 5, and 1.3404(15) Å in 6 vs the typical C^aromatic-O bond distance of 1.362 Å in free PhOH) would be consistent with this. In contrast, the bond methoxide C–O bond length in 7 (1.413(13) Å) matches that of free MeOH (1.413 Å). The Mn-O(1) hydroxide distance (1.8540(8)) in 8 is longer than expected, based on the pK_a of H₂O relative to the other members of the series. This is because the hydroxide ion is hydrogen bonded to a co-crystallized H₂O molecule (O(1)•••H(2G)O(2)H= 1.890 Å; O(1)•••O(2)= 2.660 Å) in the solid state. There is also a weak intramolecular hydrogen bond between one of the primary amine N(3,4)H₂ protons and the bound hydroxide oxygen O(1) (N(3)-H•••O(1)= 2.730 Å) in 8. This effect is modified somewhat by an intermolecular H-bond of the opposite nature (ie, one which pushes electron density back onto the hydroxide oxygen) between the hydroxide proton and the thiolate sulfur of a neighboring molecule (O(1)-H•••S(1)= 2.466 Å; Figure S-17). Weak intramolecular H-bonds are also observed between the alkoxide (N(3)-H•••O(1)= 2.632 Å in 7), and phenoxide (N(4)-H•••O(1)= 2.796 Å in 5, N(4)-H•••O(1)= 2.682 Å in 6) oxygens and primary amine N–H protons. Again with the exception of hydroxide-bound 8, these intramolecular H-bond distances roughly correlate with ROH pK_a. There is an inverse correlation between the Mn(III)-SR and Mn-OR bond lengths (Figure S-18): the Mn-S(1) bond length increases as the Mn-O(1) bond length decreases (Table 2). This reflects the flexible nature of the Mn-SR bond and its ability to compensate for changes in electron density at the metal ion.⁵¹ It also reflects the fact that both the thiolate and alkoxide are competing for overlap with the same π-d orbital. An increased Mn(III)-SR bond covalency might be responsible in part for the increased stability of phenoxide ligated 5 and 6 versus methoxide- and hydroxide-ligated 7 and 8.

Figure 5.

ORTEP of [Mn^III(S^Me2N₄(tren))(OPh)](PF₆)•MeCN•Et₂O (6) and [Mn^III(S^Me2N₄(tren))(OPh-^pNO₂)](PF₆)•MeCN (5), with hydrogen atoms, counter anions, and solvents of crystallization omitted for clarity.

Figure 6.

ORTEP of [Mn^III(S^Me2N₄(tren))(OMe)]⁺ (7) and [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) with hydrogen atoms omitted for clarity.

Solution Properties of [Mn^III(S^Me2N₄(tren))(OR)]⁺ (R= ^pNO₂Ph (5), Ph (6), Me (7), and H (8)) Relevant to Reactivity

The most stable thiolate-ligated Mn(III)-OR derivatives, [Mn^III(S^Me2N₄(tren))(OPh-^pNO₂)]⁺ (5) and [Mn^III(S^Me2N₄(tren))(OPh)]⁺ (6), are derived from the more acidic alcohols, PhOH (pK_a = 27.2 in MeCN),^42,43 and p-NO₂-PhOH (pK_a = 22 in MeCN).^42,43 The more reactive (vide infra), and consequently somewhat less stable derivatives, [Mn^III(S^Me2N₄(tren)(OMe))]⁺ (7) and [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8), are derived from the less acidic proton-donors MeOH and H₂O (pKa >> 32 in MeCN).^42,43 Stability was assessed based on changes to the electronic absorption spectrum. Ambient temperature solutions of 7 (τ_1/2(MeCN, 298 K)= 1.9 hrs) and 8 (τ_1/2(H₂O, 298 K)= 3.8 hrs) were found to decay in approximately half a day, whereas those of 5 and 6 showed no signs of decay even after a week. In the solid state, compounds 7 and 8 were found to be more robust, and were stored in a −80°C freezer. For solution studies (i.e., quantitative electronic absorption spectroscopy (vide supra), kinetics (vide infra), and electrochemistry), fresh samples of 7 and 8 were prepared in every case. Cyclic voltammograms of 5-8 (Figures S-13- S-17) display irreversible waves indicating that, like our previously reported compounds [M^III(S^Me2N₄(tren))(L)]⁺ compounds (M= Co, Fe),^24,52 the six-coordinate structure converts to a five-coordinate structure upon reduction. This has been attributed to the π-donating properties of the thiolate which decrease the Lewis acidity of the metal center, expecially when it is a lower oxidation state.⁵³ The cathodic peak potentials for 5-8 (E_pc, Table 3) were found to be reproducible to within ±10 mV, and become increasingly negative as the electron withdrawing ability of the R-group (via resonance) decreases. Given their potentials, none of these compounds would be expected to behave as a strong oxidant. In fact, hydroxide-, and methoxide-bound 8 and 7 have the lowest (ie, most negative) Mn^III/II potential within the series, and should therefore be the weakest oxidants, and yet they turn out to be the most reactive and capable of abstracting H-atoms (vide infra), suggesting that something other than redox potential is governing their reactivity. Phenoxide-ligated 5 and 6 were found to be unreactive. The reactivity of 8 was explored in both H₂O and MeCN, and that of 7 was explored only in MeCN.

Table 3.

Redox Properties, Proton Affinity and O–H BDFE of Mn(III/II)-O(H)R compounds in MeCN or H₂O

	Ep,c (mV vs Fc+/0) in MeCN	Estimated pKa of Mn(II)-O(H)R in MeCN	BDFE in MeCN (kcal/mol)	BDFE in H2O (kcal/mol)
Mn-OPh-pNO2 (5)	−220	< 12.2	N/A	N/A
Mn-OPh (6)	−300	< 13.5	N/A	N/A
Mn-OMe (7)	−450	≥ 16.2	N/A	N/A
Mn-OH (8)	−600	21.2	70.1	74.0(5)

Proton-Coupled Electron Transfer to [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) in Water

Of the four Mn(III)-OR complexes 5-8 only 8 is stable in water. This water solubility provides a wealth of thermodynamic information (vide infra), in addition to facilitating aerobic oxidation catalysis under environmentally friendly conditions (vide infra). The redox potential of 8 in H₂O is pH-dependent (Figure 7; Table S-0) indicating that proton-coupled electron transfer is involved. The slope of the aqueous Pourbaix diagram of Figure 7 (52 mV/pH) is close to that expected (59 mV/pH) for ideal Nernstian behavior, and is indicative of the transfer of single proton with each electron. The significance of this is that the reverse process, i.e., Mn(II)-H₂O → Mn(III)-OH + H⁺ + e⁻, is involved in photosynthetic water splitting.^54-56 The pH range examined (Figure 7) reflects the stability range of the Mn(III)-hydroxo compound, [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8). The low aqueous redox potentials (380 mV ≥ E_p,c (8) ≥ 150 mV vs NHE) over the pH range examined (6.2 ≤ pH ≤ 10.5; Figure 7) indicate that 8 is a very mild oxidant in water. Using the aqueous Pourbaix diagram for 8 (Figure 7) one can directly calculate the aqueous O–H bond dissociation free energy (BDFE) for the product of proton-coupled electron transfer, [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9; Scheme 2). using equation (1), where E_pH is the redox potential of 8 at a given pH, and the constant 57.6 kcal/mol is the solvent-dependent energy of formation and solvation (C_G,sol) of H^• in

$$\mathit{BDFE}\left({\mathit{Mn}}^{\mathit{II}}\right(\mathit{HO}-H\left)\right)=23.06\left[E_{\mathit{pH}}\right]+1.37\left(\mathit{pH}\right)+57.6\mathit{kcal}∕\mathit{mol}$$ (1)

H₂O.¹ This equation represents a modified form¹ of the Bordwell equation,^8,9 for use specifically in aqueous solution reactions. The use of cathodic peak potentials, E_p,c, in place of E_1/2 , has been previously shown to introduce essentially no error in BDFE calculations for species involved in irreversible redox reactions.^1,9 The redox potential used in equation (1) was referenced to NHE according to the convention for aqueous bond dissociation free energies calculations.¹ The relationship between the cathodic peak potential and pH (E_pH= E⁰_pH=0 – (0.052*pH; E⁰_pH=0 = +0.708V), obtained from the least-squares fit to the data of Figure 7 provided E_pH values for this calculation. Based on this calculation, the O–H bond dissociation free energy for 9 was determined to be 74.0±0.5 kcal/mol (Figure 8; Table 3). This O–H bond strength is significantly lower than that of free H₂O (119 kcal/mol), demonstrating that water coordination to Mn(II) significantly activates the O–H bond--- a property that is relevant to water splitting in photosynthesis. Using the Bordwell equation (2)^8,9 below (with C_G,H2O = 57.6 kcal/mol for H₂O), the experimentally determined pK_a= 5.3 of 9, and the BDFE obtained from the Pourbaix data (74.0 kcal/mol), the formal potential for the [Mn(III)-OH₂]²⁺ (10)/[Mn(II)-OH₂]⁺ (9) couple (+0.40 V vs NHE; Figure 8) was

$$\mathit{BDFE}\left(\mathrm{M}\right(\mathrm{II}\left)\right(\mathrm{HO}-\mathrm{H}\left)\right)=23.06E^0\left(\mathrm{M}\right(\mathrm{III})∕\mathrm{M}(\mathrm{II}\left)\right)+1.37\left({\mathit{pK}}_a\right)+C_{G,\mathit{sol}}$$ (2)

calculated (Scheme 2). The proton dissociation constant for 9 was determined to be pK_a= 5.3 via HOTf + 8 titration experiments (Figure S-20). Due to their instability in H₂O, Pourbaix diagrams were not obtained for the methoxide and phenoxide derivatives Mn(III)-OMe (7), Mn(III)-OPh (6), and Mn(III)-OPh-p-NO₂ (5), making it impossible to calculate the O–H Bw 8 suggested that it would be capable of promoting PCET-type reactions involving substrates with weaker X-H bonds.

Figure 7.

Pourbaix diagram for [Mn^III(S^Me2N₄(tren))(OH)]⁺ (2) in aqueous solution. pH was adjusted using dilute aqueous solutions of either HBF₄ or NH₄OH and cathodic peak potentials were measured at 298 K in H₂O with 0.1 M KClO₄ of supporting electrolyte and a glassy carbon electrode..

Scheme 2.

Figure 8.

Thermochemical cycle for [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8)-promoted PCET reactions in aqueous solution. Potentials are referenced to NHE.

Aqueous Reactions Between [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) and TEMPOH

Reactions between hydroxide-bound 8 and 2,2′,6,6′-tetramethylpiperidine-1-ol (TEMPOH) in H₂O were monitored by electronic absorption spectroscopy (Figure 9), and shown to involve the disappearance of visible bands at 418 and 489 nm to afford a featureless UV-vis spectrum characteristic of [Mn^II(S^Me2N₄(tren))]⁺ (1)²⁴ (Scheme 2). Titration experiments show that a clean 1:1 (8:TEMPOH) stoichiometric reaction is involved. Kinetic data (vide infra) suggests that the reaction involves an unobserved reduced six-coordinate intermediate, [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9), which was shown by cyclic voltammetry (vide supra, Figure S-17) to be unstable on the time-scale of these experiments. This reactivity would be consistent with the relative O–H bond strengths of 9 (BDFE= 74 kcal mol⁻¹) versus that of TEMPOH (BDFE= 71 kcal mol⁻¹),¹ in H₂O. Quantitative conversion back to 8 (Figure S-21) is observed upon the addition of O₂ to the TEMPOH-reduced product (Scheme 2). Given that oxo bridged 2 forms upon O₂ addition to 1, and converts to hydroxo-bound 8 in H₂O, it is likely that oxo-bridged 2 forms as an intermediate en route to 8 (Scheme 2). At least 10 turnovers (likely more) can be achieved without significant degradation of the “catalyst” 8 (Figure S-22). The net reaction promoted by 8 (Scheme 2), ie, O₂ promoted TEMPOH oxidation in H₂O at ambient temperatures, is thus done under environmentally-friendly (“green”) conditions. No reaction is observed between hydroxide-bound 8 and 1,4-cyclohexadiene, 9,10-dihydroanthrocene, or toluene. Based on their self-exchange rate constants¹⁰ these substrates would be expected to react at a rate that is slower than the decomposition rate of 8 (τ_1/2(H₂O, 298 K)= 3.8 hrs).

Figure 9.

Anaerobic reaction between [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) and 0.2 equiv aliquots of TEMPOH in H₂O at 298 K, showing that 1.0 equiv of TEMPOH is required to convert 8 to five-coordinate 1, presumably via an unobserved reduced water-bound intermediate [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9).

Estimated Thermodynamic Ability of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) to Abstract Hydrogen Atoms in MeCN

Complex 8 is the only known Mn(III)-OH complex for which thermodynamic data, predictive of H-atom abstraction capability, is reported in water. This unfortunately precludes a comparison between the O–H BDFE of the corresponding reduced complex [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9) and other similar systems for which data is reported in MeCN⁴⁸ or acetone/H₂O (4:1).¹⁶ By determining thermodynamic parameters for 9 in organic solvents such as MeCN, comparisons could then be made. Although a cathodic peak potential could be obtained for 8 in MeCN (Table 3), the “pK_a“ of protonated 10 (Figure 8) in MeCN could not be experimentally determined, however, because the MeCN displaces the H₂O ligand. This ultimately prevented an experimental determination of the O–H BDFE for 9 (Scheme 2) in MeCN. A crude approximation of this O–H bond dissociation free energy (BDFE(9)) can be made, however, if we assume that the solvation energy of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) and [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9) is the same. If this is assumed, then the difference between the O–H BDFE for 9 in MeCN vs H₂O is equal to the difference between the free energy of solvation for H• in MeCN (5.12 kcal/mol) and H₂O (8.98 kcal/mol), affording a BDFE for [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9) in MeCN of 70.1 kcal/mol (Table 3). This estimated BDFE would suggest that 8 should be capable of abstracting a H-atom from TEMPOH (O–H BDFE= 66.5 kcal mol⁻¹ in MeCN)¹ in MeCN, consistent with what is observed.

Hydrogen Atom Transfer Reactivity of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) and [Mn^III(S^Me2N₄(tren)(OMe))]⁺ (7) in MeCN

As anticipated, based on the thermodynamic discussion above, hydroxide-bound 8 reacts quantitatively with TEMPOH in MeCN at 298 K. This was determined by monitoring the reaction between 8 and TEMPOH (0.2 eq aliquots) in MeCN using electronic absorption spectroscopy (Figure S-23). Methoxide-bound 7 was also found to react quantitatively with TEMPOH in MeCN (Figure S-24). One equivalent of TEMPOH was required for complete disappearance of 7 (Figure S-24) or 8 (Figure S-23) indicating that a clean 1:1 stoichiometric reaction is involved. The initial products expected in these reactions [Mn^II(S^Me2N₄(tren))(ROH)]⁺ (R= H, Me) are not observed, consistent with the electrochemical data (vide supra; Figures S-13 - S-14) which demonstrates that these six-coordinate reduced species are unstable^52,57 on the time-scale of these experiments. Phenoxide-bound 5 and 6 show no reaction with TEMPOH, even when excess substrate is added and the reaction mixture is monitored for prolonged periods under anaerobic conditions (~24 hrs). This would imply that the O–H BDFE of [Mn^II(S^Me2N₄(tren))(ROH)]⁺ (R= Ph, Ph-^pNO₂) is less than 66.5 kcal/mol, the O–H bond BDFE of TEMPOH.

Kinetics of the Reaction Between [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) and [Mn^III(S^Me2N₄(tren)(OMe))]⁺ (7) and TEMPOH in MeCN

Oxidation of TEMPOH by 8 is reasonably fast (Figure S-25) in MeCN, with a second order rate constant of k(298 K)= 2.1 × 10³ M⁻¹s⁻¹ (Figure S–26), a k_H/k_D= 3.1 (Figure 10), and activation parameters of ΔH^‡= 8.2 kcal/mol, and ΔS^‡= −25.5 eu (Figure S-27). TEMPOH oxidation by 7 is roughly an order of magnitude slower than 8, with a second order rate constant of k(298 K)= 3.6 × 10² M⁻¹s⁻¹ (Figure S–28), an ambient temperature k_H/k_D= 2.1 (Figure S-28), and activation parameters of ΔH^‡= 8.3 kcal/mol, and ΔS^‡= −29 eu (Figure S-29). The observed deuterium isotope effects for both the reaction between 7 and 8 and TEMPOH in MeCN, coupled with the high oxidation potential (E= +710 mV vs FeCp₂/FeCp₂⁺) and pK_a (pK_a= 41) of TEMPOH,¹ would be consistent with a mechanism involving concerted H-atom transfer, as opposed to sequential H⁺ followed by e^– transfer or vice versa.

Figure 10.

Kinetic isotope effect (k_H/k_D) for the reaction between [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) (0.5 mM) and TEMPOH(D) in MeCN at 298 K.

Factors Responsible for the Ability of [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8) and [Mn^III(S^Me2N₄(tren)(OMe))]⁺ (7) to Abstract H-Atoms

According to the Bordwell equation (eqn (2)),^1,8-12 in relation to thermodynamic cycles such as that of Figure 8, the two factors which determine the driving force for H-atom transfer reactions are the redox potential of the oxidant (e.g., Mn(III)-OR) and the pK_a of the protonated reduced product (e.g., Mn(II)-O(R)H). If one compares the cathodic peak potentials (E_p,c) of 5-8 in MeCN (Table 3), one would not expect any of these complexes to be capable of oxidizing TEMPOH, and, based on their E_p,c values, one would expect the phenoxide complexes 5 and 6 to be better oxidants than either the methoxide or hydroxide complexes 7 and 8. This is in contrast to the observed reactivity, however, suggesting that the trend in pK_a values opposes that of E_p,c for these compounds. The pK_a values required in order to abstract a H-atom from TEMPOH in MeCN at a given potential are illustrated in Figure 11. These were calculated using the Bordwell equation (2), with potentials referenced relative to Cp₂Fc⁺/Cp₂Fc according to the convention.¹ From this, one can see that even with a cathodic peak potential as low as −600 mV (vs Cp₂Fc⁺/Cp₂Fc) a compound can abstract a H-atom from TEMPOH, as long as the pK_a for the reduced, protonated derivatives, [Mn^II(S^Me2N₄(tren))(ROH)]⁺ (R= H, Me, Ph, Ph-^pNO₂), is greater than 18.5. The anionic thiolate of compounds 5-8 should increase the basicity of the coordinated oxygen in all four complexes relative to, for example, a nitrogen-ligated Mn(III) complex such as [Mn^III(PY5)(OH)]²⁺ (11; PY5 = 2,6-bis(bis(2-pyridyl)methoxymethane)-pyridine).⁴⁸ Similarly, the thiolate ligand of choloroperoxidase and P450 has been shown to create a highly basic oxo that is believed to be responsible for the highly reactive nature of their mildly oxidizing active sites.^18-20 Using the estimated O–H BDFE (70.1 kcal/mol) obtained as described above for 9 in MeCN, and the cathodic peak potential for 8 in MeCN (−600 mV; Table 3), we can then calculate an estimated pK_a = 21.2 for 9 using the Bordwell equation (2). This indicates that the hydroxide of thiolate-ligated 8 is approximately eight-orders of magnitude more basic than that of nitrogen-ligated 11, the protonated reduced derivative of which has a pK_a= 13.0±0.5 in the same solvent.⁴⁸ These pK_a differences, in part, reflect differences in molecular charge (1+(8) vs 2+(11)), as well as the electron donating thiolate ligand present in 8, but not in 11.^18-20 The basic hydroxy oxygen therefore allows 8, which is nearly a volt (780 mV) less oxidizing than 11 (+180 mV vs. Fc^+/0 in MeCN), to abstract H-atoms from substrates, such as TEMPOH, with weak X-H bonds. The stronger O–H bond (BDFE= 82±2 kcal/mol in MeCN) ⁴⁸ of the reduced and protonated derivative of 11 [Mn^II(PY5)(OH₂)]²⁺ provides ~12 kcal/mol more driving force relative to 8 for the abstraction of H-atoms, consistent with the fact that this dicationic, N-ligated derivative abstracts H-atoms from stronger X-H bonds.⁴⁸

Figure 11.

The pKa required to abstract a H-atom from TEMPOH as a function of redox potential in MeCN. The data points shown provide the pK_a that would be required specifically for 5-8 to react with TEMPOH given their experimentally determined redox potentials.

Although the quantitative data obtainable for 8, due to its PCET properties in water (vide supra), is not available for methoxide-ligated 7 and phenoxide-ligated 5 and 6, we can use the fact that the former abstract H-atoms from TEMPOH, whereas the latter two do not, to place a lower and upper limit on the pK_a of the protonated reduced derivatives Mn(II)-O(H)R (R= ^pNO₂Ph, Ph, or Me (7)). According to the E_p,c vs pK_a graph of Figure 11, and the equation relating these two parameters ( E_p,c= −0.0594*pK_a + 0.503), we can estimate that, given their cathodic peak potentials (Table 3), the protonated reduced derivatives of phenoxide-ligated 5, Mn(II)-O(H)R (R= ^pNO₂Ph), and 6, Mn(II)-O(H)R (R= Ph), would have to have a pK_a of 12.2 and 13.5, respectively in order to abstract H-atoms from TEMPOH. The fact that they do not react implies that their pK_a-values are less than these values (Table 3). These estimated upper limits to these pK_a-values make sense given that coordination to a Mn³⁺ ion would be expected to substantially drop the pK_a of p-NO₂-PhOH and PhOH relative to that of the free phenols (pK_a(PhOH )= 27.2; pK_a(p-NO₂-PhOH )= 22 in MeCN)^42,43 in MeCN. Based on its cathodic peak potential (Table 3) methoxide-ligated 7 would have to have a pK_a of 16.2 in order to abstract H-atoms from TEMPOH. The fact this 7 does, in fact, react implies that its pK_a-value is greater than this.

Conclusions

In summary, we have described the synthesis, structural characterization, electrochemical and reactivity properties of a series of thiolate-ligated [Mn^III(S^Me2N₄(tren))(OR)]⁺ complexes (R= ^pNO₂Ph (5), Ph (6), Me (7), and H (8)). Each of these complexes was synthesized via the proton-induced cleavage of a structurally characterized oxo-bridged dimer containing an unusually bent bridging Mn-O-Mn angle (158.75°) with a fairly basic oxo. An alternative route to hydroxo-bound 8 involved the aerobic oxidation of the five-coordinate Mn(II) precursor in aqueous solution. In H₂O, 8 was shown to display a pH-dependent redox potential with a Pourbaix diagram slope (52 mV/pH) that is indicative of the transfer a single proton with each electron. The least stable, and least oxidizing members of this series, [Mn^III(S^Me2N₄(tren)(OMe))]⁺ (7) and [Mn^III(S^Me2N₄(tren))(OH)]⁺ (8), were shown to abstract H-atoms from TEMPOH, whereas the more oxidizing members (based on redox potential), [Mn^III(S^Me2N₄(tren))(OPh-^pNO₂)]⁺ (5) and [Mn^III(S^Me2N₄(tren))(OPh)]⁺ (6), were found to be unreactive. Kinetic parameters for Mn(III)-OR (R= Me(7), H(8))–promoted TEMPOH oxidation (k_H/k_D= 3.1 (8), k_H/k_D= 2.1 (7)) were shown to be consistent with concerted H-atom abstraction, and 8 was shown to react an order of magnitude faster than 7. The aqueous O–H BDFE for water-bound Mn(II)-OH₂ (74.0±0.5 kcal/mol; Table 3) was directly calculated using the best-fit line to the Pourbaix diagram in conjunction with the Bordwell equation, and converted to an estimated O–H BDFE in MeCN (70.1 kcal/mol; Table 3) so that comparisons could be made with other systems. Hydroxo-bound 8 is the only water soluble Mn(III)-OH compound shown to promote H-atom transfer. It was determined that the trend in pK_a values opposes that of E_p,c for the compounds reported herein. A pK_a-value of 21.2 for the reduced protonated derivative of 8, [Mn^II(S^Me2N₄(tren))(H₂O)]⁺ (9), was calculated using the Bordwell equation and its estimated O–H BDFE in MeCN (based on its experimental value in H₂O). It was suggested that an increased oxygen basicity due to the thiolate sulfur, which offsets its low cathodic peak potential (−300 mV vs Fc⁺/Fc⁰) is responsible for the observed HAT reactivity of 8. Reactivity or lack of reactivity with TEMPOH was used to place either upper or lower limits on the pK_a (Table 3) of the reduced protonated derivatives of methoxide-ligated 7 and phenoxide-ligated 5 and 6.

Table 1.

Crystal data for [Mn^III(S^Me2N₄(tren)]₂-(μ-O)(PF₆)₂•MeCN (2), [Mn^III(S^Me2N₄(tren)(OPh-^pNO₂))](PF₆)•MeCN (5), [Mn^III(S^Me2N₄(tren)(OPh))](PF₆)•MeCN•Et₂O (6), [Mn^III(S^Me2N₄(tren)(OMe))](PF₆) (7), and [Mn^III(S^Me2N₄(tren))(OH)](PF₆)•H₂O (8).

	2	5	6	7	8
Formula	C24H53F12Mn2 N9OP2S2	C19H32F6Mn N6O3PS	C23H43F6Mn N5O2PS	C12H28F6MnN4 OPS	C11H28F6MnN4 O2PS
MW	947.69	624.48	653.59	476.35	480.35
T, K	130(2)	100(2)	110(2)	100(2)	100(2)
Unit cellA	Orthorhombic	Monoclinic	Monoclinic	Orthorhombic	Monoclinic
a, Å	12.5660(2)	31.099(2)	8.4834(3)	13.5710(19)	8.5318(6)
b, Å	17.0110(4)	14.0709(8)	16.4970(7)	8.8115(12)	16.6962(13)
c, Å	36.2170(6)	12.1715(7)	22.0354(9)	16.452(2)	13.7039(10)
α, deg	90	90	90	90	90
β, deg	90	91.213(4)	97.872(2)	90	94.552(3)
γ, deg	90	90	90	90	90
V, Å3	7741.8(3)	5324.9(6)	3054.8(2)	1967.4(5)	1945.9(2)
Z	8	8	4	4	4
d(calc), g/cm3	1.626	1.558	1.421	1.608	1.640
Space group	P bca	C 2/c	P 21/c	P nma	P 21/c
R	0.0875b	0.0506b	0.0265b	0.0494b	0.0211b
Rw	0.2476 c	0.1099c	0.0668c	0.1339c	0.0559c
GOF	1.014	1.169	1.041	1.077	1.042

^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.