Characterization and chemical reactivity of room-temperature-stable MnIII–alkylperoxo complexes

While alkylperoxomanganese(iii) (MnIII–OOR) intermediates are proposed in the catalytic cycles of several manganese-dependent enzymes, their characterization has proven to be a challenge due to their inherent thermal instability. Fundamental understanding of the structural and electronic properties of these important intermediates is limited to a series of complexes with thiolate-containing N4S− ligands. These well-characterized complexes are metastable yet unreactive in the direct oxidation of organic substrates. Because the stability and reactivity of MnIII–OOR complexes are likely to be highly dependent on their local coordination environment, we have generated two new MnIII–OOR complexes using a new amide-containing N5− ligand. Using the 2-(bis((6-methylpyridin-2-yl)methyl)amino)-N-(quinolin-8-yl)acetamide (H6Medpaq) ligand, we generated the [MnIII(OOtBu)(6Medpaq)]OTf and [MnIII(OOCm)(6Medpaq)]OTf complexes through reaction of their MnII or MnIII precursors with tBuOOH and CmOOH, respectively. Both of the new MnIII–OOR complexes are stable at room-temperature (t1/2 = 5 and 8 days, respectively, at 298 K in CH3CN) and capable of reacting directly with phosphine substrates. The stability of these MnIII–OOR adducts render them amenable for detailed characterization, including by X-ray crystallography for [MnIII(OOCm)(6Medpaq)]OTf. Thermal decomposition studies support a decay pathway of the MnIII–OOR complexes by O–O bond homolysis. In contrast, direct reaction of [MnIII(OOCm)(6Medpaq)]+ with PPh3 provided evidence of heterolytic cleavage of the O–O bond. These studies reveal that both the stability and chemical reactivity of MnIII–OOR complexes can be tuned by the local coordination sphere.


Introduction
Metal-alkylperoxo adducts are essential species in industrial and biological oxidation reactions. [1][2][3] For example, Co III -alkylperoxo adducts are proposed as intermediates in the industrial oxidation of cyclohexane to adipic acid. 1,[4][5][6][7] In the oxidation mechanism, homolytic cleavage of the O-O bond of a Co IIIcyclohexylperoxo species leads to the production of cyclohexanol and cyclohexanone. 4,8 Further radical-induced oxidation of cyclohexanone by C-C bond cleavage yields adipic acid. 4 In biological systems, metal-alkylperoxo adducts are common intermediates in a variety of oxygenase enzymes, where they can be directly involved in substrate oxidation or precede the formation of high-valent metal-oxo species. 9,10 Given the importance of metal-alkylperoxo species in such reactions, there are now many examples of synthetic Fe-, 11,12 Co-, 1,4,5 and Cu-alkylperoxo 3,13,14 adducts, and these complexes are capable of oxidizing substrates such as 1,4-cyclohexadiene, 2-phenylpropionaldehyde, and triphenylphosphine.
While such studies of synthetic model complexes have probed the properties and reactivity of many types of metalalkylperoxo complexes, examples of Mn-alkylperoxo adducts are more limited, and there remain many open questions concerning the factors governing the decay and reactivity of these complexes. Kovacs and co-workers have performed pioneering investigations of Mn III -alkylperoxo adducts, including structural characterization of a family of complexes by X-ray crystallography. 15,16 These studies employed pentadentate, thiolate-containing N 4 S À ligands, which in the corresponding [Mn III (OOR)(N 4 S)] + complexes placed the thiolate donor cis to the alkylperoxo ligand, with bulky quinolinyl or 6methylpyridyl substituents trans to each other and cis to the alkylperoxo moiety (Fig. 1, le). 15,16 The crystallographically observed Mn-N distances for the quinolinyl and 6-methylpyridyl donors range from 2.35 to 2.52Å, which are quite long for Mn III -N interactions. Interestingly, these long Mn-N distances are correlated with the alkylperoxo O-O bond lengths, which vary from 1.43 to 1.47Å. 16 As shorter Mn-N distances gave longer O-O bonds, it was proposed that less Lewis acidic Mn III centers yielded more activated Mn IIIalkylperoxo adducts. By using variable-temperature kinetic studies, the O-O bond lengths for these Mn III -alkylperoxo complexes were in turn related to their thermal decay rates. Mn III -alkylperoxo adducts with longer O-O bonds decayed more rapidly, with lower DH ‡ values and DS ‡ values that were more negative. 16 Because of the correlation between the Mn-N distances and the O-O bond lengths, these results suggest that activation of Mn III -alkylperoxo complexes can be controlled by the donor strength of groups cis to the alkylperoxo unit. Thermal decomposition studies and analysis of the decay products of the Mn III -cumylperoxo adduct supported a decay by homolytic cleavage of the alkylperoxo O-O bond. 16 Very recently, Kovacs et al. reported a roomtemperature stable Mn III -alkylperoxo complex supported by an alkoxide analogue of the N 4 S À ligands. 17 DFT computations for the [Mn III (OO t Bu)(N 4 O)] + and [Mn III (OO t Bu)(N 4 S)] + pair support the notion that the enhanced stability of the alkoxideligated complex arises from greater Lewis acidity of the Mn III center.
While these studies provide structure-reactivity correlations with regards to the thermal decay pathway, the Mn III -alkylperoxo adducts of the N 4 S À ligands failed to show any direct reaction towards a range of substrates. 16 Product analysis of the reaction solutions provided evidence of the oxidation of a variety of substrates (i.e., PEt 3 , TEMPOH, and cyclohexane carboxaldehyde) following the decay of the Mn III -alkylperoxo adducts, implying that a product of the decay pathway is a capable oxidant. In contrast, Mn III -hydroperoxo complexes supported by neutral, macrocyclic N 4 ligands are known to react directly with aldehydes, suldes, and hydrocarbons possessing weak C-H bonds. 18,19 Given that metal-alkylperoxo adducts are oen taken as analogues of metal-hydroxoperoxo species, [18][19][20][21][22][23][24][25] the stark difference in reactivity between Mn III -alkylperoxo and Mn III -hydroperoxo adducts in substrate oxidation reactions is striking. The disparate reactivities of these complexes might reect the differences in the properties of the supporting ligands employed (N 4 S À for Mn III -alkylperoxo versus neutral N 4 for Mn III -hydroperoxo). There is a clear need to understand better the role of non-thiolate-containing supporting ligands in inuencing the properties and reactivity of Mn III -alkylperoxo complexes.
We previously generated a pair of Mn III -alkylperoxo complexes with ligands lacking thiolate ligation. 26 These Mn IIIalkylperoxo complexes were supported by the pentadentate dpaq and dpaq 2Me ligands, both of which feature strongly donating amide groups trans to the alkylperoxo moiety ( Fig. 1 26 Both [Mn III (OO t Bu)(dpaq)] + and [Mn III (OO t Bu)(dpaq 2Me )] + were unstable (t 1/2 ¼ 3200 and 3600 s, respectively, for 2 mM solution in CH 3 CN at À15 C) but were characterized by electronic absorption, Mn K-edge Xray absorption, and FT-IR spectroscopies. 26 The observation of tBuOOc in EPR spectra of the complexes following their thermal decay provided support for a decay pathway involving Mn-O bond homolysis. 26 While these data suggest differences in decay pathways for the thiolate-versus non-thiolate-ligated complexes, the large excess of t BuOOH (ca. 100 equivalents) required to form the [Mn III (OO t Bu)(dpaq)] + and [Mn III (OO t Bu)(dpaq 2Me )] + complexes made a complete analysis of their reactivity and decay pathways unfeasible. The large excess of t BuOOH also complicated any investigations of substrate oxidation.
Given the limitations of the Mn III -alkylperoxo complexes of the dpaq and dpaq 2Me ligands, we sought to develop a derivative of the dpaq ligand that would better stabilize the Mn III -alkylperoxo adduct. Herein, we report Mn III -alkylperoxo adducts supported by 6Me dpaq (Fig. 2). This new ligand incorporates steric bulk at the 6 position of the pyridyl substituents in the equatorial plane. This choice was inspired by the higher stability of the Mn III -alkylperoxo complexes supported by N 4 S À ligands with two bulky N-donor ligands cis to the alkylperoxo ligand ( Fig. 1, le) (Fig. S5 †), ESI-MS (Fig. S6 †), and Evans NMR (Fig. S7 †) provide evidence that the mononuclear structure observed in the solid state is retained in solution.
[Mn II (H 2 O)( 6Me dpaq)]OTf reacts very slowly with dioxygen. Electronic absorption data collected during the oxygenation of a 2.5 mM solution of [Mn II (H 2 O)( 6Me dpaq)]OTf in CH 3 CN shows the appearance of a single feature at 510 nm, but this new chromophore is still forming even aer 48 hours (Fig. S8, † le). This reactivity contrasts with that of [Mn II (dpaq)](OTf) and the majority of its derivatives, as these Mn II complexes reacted with dioxygen with full conversion to Mn III products over the course of several hours. [27][28][29] Presumably, the elongation of the Mn-N4 and Mn-N5 bonds in [Mn II (H 2 O)( 6Me dpaq)]OTf leads to a more electron-decient Mn II center with muted reactivity with dioxygen. In contrast, the reaction between [Mn II (H 2 -O)( 6Me dpaq)]OTf and 0.5 equiv. PhIO is rapid, resulting in the formation of a bronze colored solution with a single electronic absorption feature at ca. 510 nm (Fig. S8, † right), which can be attributed to [Mn III (OH)( 6Me dpaq)](OTf) (vide infra).
X-ray diffraction studies of crystals obtained from this dark orange solution establish the oxidation product as [Mn III (OH)( 6Me dpaq)](OTf) (Fig. 3, right). In this complex, a sixcoordinate Mn III center is in a distorted octahedral geometry with the hydroxo ligand trans to the amide function. This coordination mode is identical to that observed in Mn IIIhydroxo complexes of dpaq and its derivatives. [27][28][29] The Mn-OH distance of 1.806(6)Å is within error of that observed for [Mn III (OH)(dpaq)](OTf) (1.806(13); see Table 1) 27 and is on the low end of the range of Mn-OH bond lengths reported for other Mn III -hydroxo complexes (1.81-1.86Å). 28,[30][31][32][33][34][35][36][37][38] Further comparisons of the X-ray structure of [Mn III (OH)( 6Me dpaq)](OTf) with those of related Mn III -hydroxo species reveals that the 6-Mepyridyl groups give rise to substantial bond elongations. Specically, the Mn-N4 and Mn-N5 distances of [Mn III (OH)( 6-Me dpaq)](OTf) are 2.322(6) and 2.381(7)Å, while the corresponding distances in [Mn III (OH)(dpaq)](OTf) are 2.260(14) and 2.216(15)Å (Table 1). 27 Mn III complexes of the N 4 S À class of ligands ( Fig. 1) with 6-Me-pyridyl or quinolinyl substituents in the equatorial eld had Mn-N distances ranging from 2.352-    Previous investigations of [Mn III (OH)(dpaq)](OTf) and a subset of its derivatives have shown that dissolution of the salts of these Mn III -hydroxo complexes in dried CH 3 CN leads to the formation of an equilibrium mixture of Mn III -hydroxo and (m-oxo)dimanganese(III,III) complexes that can both be detected by 1 H NMR spectroscopy. 29,40 The 1 H NMR spectrum of [Mn III (OH)( 6Me dpaq)](OTf) in CD 3 CN at 298 K exhibits seven hyperne-shied peaks that lie well outside 0-20 ppm (the diamagnetic region), as well as two well-resolved peaks in the 0-10 ppm region (Fig. 4, red trace). The lack of a large number of peaks in the diamagnetic region suggests that dissolution of [Mn III (OH)( 6Me dpaq)](OTf) in CD 3 CN does not result in the formation of (m-oxo)dimanganese(III,III) species. The chemical shis for the 1 H NMR signals of [Mn III (OH)( 6Me dpaq)] + are quite similar to those of [Mn III (OH)(dpaq)] + ( Fig. 4 and Table 2). 29,40 The four most upeld-shied peaks in the 1 H NMR spectrum of [Mn III (OH)(dpaq)] + were assigned to protons from the quinolinyl moiety. 40 The upeld region of the 1 H NMR spectrum of [Mn III (OH)( 6Me dpaq)] + shows three sharp peaks at À19.3, À45.0, and À61.6 ppm that resemble the peaks of [Mn III (OH)(dpaq)] + at À15.5, À33.7, À53.8 ppm (Fig. 4). The 1 H NMR spectrum of [Mn III (OH)( 6Me dpaq)] + lacks a broad, highly upeld-shied peak analogous to the weak, broad À63.4 ppm signal of [Mn III (OH)(dpaq)] + , but the breadth of this signal renders it difficult to resolve. The overall similarities between the upeld regions of the 1 H NMR spectra of [Mn III (OH)( 6Me dpaq)] + and [Mn III (OH)(dpaq)] + are expected given the lack of changes to the quinolinyl group in the former complex. The downeld 1 H NMR signals of [Mn III (OH)(dpaq)] + and the resonance at À4.6 ppm were attributed to pyridyl protons. 40 Consequently, the larger relative perturbations in the downeld regions of the 1 H NMR spectra of [Mn III (OH)(dpaq)] + and [Mn III (OH)( 6Me dpaq)] + can be rationalized by changes in chemical shis of pyridyl protons. The peak at 130.5 ppm for [Mn III (OH)(dpaq)] + was assigned to  the a-H of the pyridine substituent. 40 The lack of a corresponding peak in the 1 H NMR spectrum of [Mn III (OH)( 6Me dpaq)] + is consistent with the functionalization of the pyridyl functions in the 6-position.  + shows an initial rise in absorbance intensity at 510 nm that maximizes near ca. 60 minutes and then drops and levels by 120 minutes. In contrast, the absorbance intensity at 650 nm shows a steep rise from 0 to 40 minutes, grows more slowly from 40-100 minutes, and then rises quickly and levels by 120 minutes (Fig. 5 While all attempts at obtaining crystalline material for [Mn III (OO t Bu)( 6Me dpaq)](OTf) were unsuccessful, we were able to obtain diffraction-quality crystals for [Mn III (OOCm)( 6Me dpaq)](OTf), which conrmed the formulation of this complex (Fig. 6). The cumylperoxo ligand of [Mn III (OOCm)( 6Me dpaq)](OTf) is bound trans to the amide nitrogen (N2-Mn-O1 angle of 176.6 ), occupying the position of the hydroxo ligand in [Mn III (OH)( 6Me dpaq)](OTf). The Mn-O1 bond length for [Mn III (OOCm)( 6Me dpaq)](OTf) is longer than the Mn-O1 bond in the Mn III -hydroxo analogue (1.849(3) and 1.806(6)Å, respectively), but similar to crystallographic Mn-OOCm distances for the [Mn III (OOCm)(N 4 S)] + complexes (1.848(4) and 1.84(1)Å). 15,16 The O1-O2 bond is oriented such that the projection of this bond onto the equatorial plane bisects the N1-Mn-N4 bond angle. The aryl ring of the cumyl moiety is parallel to the plane of the pyridines coordinated to the Mn III center, as opposed to the out-of-plane orientation observed in the [Mn III (S Me2 N 4 (6-Me-DPEN))(OOCm)](BPh 4 ) complex. 16 This orientation of the cumyl moiety is stabilized by p-CH interactions between the aryl ring and a methylene group of the 6Me dpaq ligand, as evidenced by short H/C contacts of ca. 2.8  Although we were unable to obtain crystallographic information concerning [Mn III (OO t Bu)( 6Me dpaq)] + , the 1 H NMR spectrum of this species in CD 3 CN is essentially identical to that of [Mn III (OOCm)( 6Me dpaq)] + ( Fig. 4 and Table 2). Each spectrum shows six peaks outside the diamagnetic region, three downeld and three upeld. The only notable difference between the spectra of [Mn III (OO t Bu)( 6Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + are the number of peaks from 10-5 ppm (Table 2). While [Mn III (OO t-Bu)( 6Me dpaq)] + shows a single prominent peak at 9.2 ppm, [Mn III (OOCm)( 6Me dpaq)] + shows ve peaks in this region. We tentatively attributed these peaks to protons from the t-butyl and cumyl moieties, respectively. The 1 H NMR spectra of [Mn III (OOCm)( 6Me dpaq)] + and [Mn III (OO t Bu)( 6Me dpaq)] + are also very similar to that of [Mn III (OH)( 6Me dpaq)] + (Fig. 4 and Table 2), which is consistent with the same binding mode of the 6Me dpaq ligand in each of these complexes. In particular, the three upeld peaks of the Mn III -alkylperoxo adducts at ca. À22, À47, and À60 ppm have chemical shis very similar to the upeld resonances observed for [Mn III (OH)( 6Me dpaq)] + (À19.3, À45, and À61.6 ppm; see Table 2). Similarly, the downeld peaks at ca. 67, 46, and 42 ppm in the 1 H NMR spectrum of the Mn III -alkylperoxo adducts show only slight deviations from the corresponding resonances of [Mn III (OH)( 6Me dpaq)] + (66.0, 51.4, and 44.8 ppm; see Table 2). The solution FT-IR spectra of [Mn III (OO t Bu)( 6Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + show features at 875 cm À1 and 861 cm À1 , respectively, that are absent in the FT-IR spectrum of [Mn III (OH)( 6Me dpaq)] + (Fig. 7) (Fig. 7). EPR analysis of frozen 5 mM CH 3 CN solutions of [Mn III (OO t-Bu)( 6Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + collected at 10 K showed no signals in either perpendicular-or parallel-mode (Fig. S19 †). This observation is consistent with the lack of X-band EPR signals for many Mn III complexes due to moderate to large zero-eld splitting relative to the microwave energy. [45][46][47][48][49] The magnetic moments of these complexes (4.8m B ) are consistent with S ¼ 2 Mn III centers ( Fig. S15 and S16 †). Additional ESI-MS data and solution-phase magnetic moments further support the formulations for these Mn III -alkylperoxo complexes (Fig. S14 †).   conditions was monitored by electronic absorption spectroscopy. Each decay reaction progressed with the disappearance of the 650 nm feature associated with the Mn III -alkylperoxo adduct and the appearance of a feature at 510 nm (Fig. 8). The feature at 510 nm, along with ESI-MS analyses of the product solutions ( Fig. S20 †), marks [Mn III (OH)( 6Me dpaq)] + as the major product of these decay reactions (99 and 92% formation from [Mn III (OOCm)( 6Me dpaq)] + and [Mn III (OO t Bu)( 6Me dpaq)] + , respectively, on the basis of the extinction coefficient of the Mn III -hydroxo complex). For [Mn III (OOCm)( 6Me dpaq)] + , the decay proceeds with isosbestic behavior, and the respective decay and formation rates of the 650 and 510 nm electronic absorption signals are identical (Fig. 8, right; k obs ¼ 0.016 min À1 for both processes). In contrast, the thermal decay for [Mn III (OO t Bu)( 6Me dpaq)] + is not isosbestic, and the rate of formation of the 510 nm chromophore lags behind the decay of the 650 nm band (Fig. 8, le). We tentatively attribute this difference to the higher purity of [Mn III (OOCm)( 6Me dpaq)] + used in these experiments, as this complex was obtained as a recrystallized solid. In support, thermal decay studies of crude [Mn III (OOCm)( 6Me dpaq)] + also failed to show isosbestic behavior (Fig. S22 †). Additional product analysis following the thermal decay of 18 mM [Mn III (OOCm)( 6Me dpaq)] + in CH 3 CN at 323 K revealed 61.3 AE 0.1% 2-phenyl-2-propanol and 25.7 AE 0.1% acetophenone formed relative to the initial [Mn III (OOCm)( 6Me dpaq)] + concentration. (The organic products from the thermal decay of [Mn III (OO t Bu)( 6Me dpaq)] + were not quantied, because the volatility of acetone, a potential product, renders quantication unreliable.) When the decay of 6 mM [Mn III (OOCm)( 6Me dpaq)] + was performed in CD 3 CN, we observed 50 AE 0.3% 2-phenyl-2propanol and 40 AE 0.3% acetophenone relative to the initial [Mn III (OOCm)( 6Me dpaq)] + , a marked change in the product distribution. In addition, when [Mn III (OOCm)( 6Me dpaq)] + was allowed to decay in benzonitrile (PhCN; see Fig. S23 †), we observed 30 AE 1.1% 2-phenyl-2-propanol and 70 AE 1.1% acetophenone. The implications of these results with respect to the decay mechanism are explored in the Discussion section (vide infra).
To further probe the decay reactions, we monitored decay kinetics for 1.25 mM CH 3 CN solutions of each complex from 303-348 K in CH 3 CN under anaerobic conditions. At each temperature, the decay could be t to a pseudo-rst-order process, and the k obs values at different temperatures were t to the Eyring equation to obtain activation parameters (Fig. 9). This analysis yielded DH ‡ ¼ 21.4 AE 1.5 kcal mol À1 , DS ‡ ¼ À9. 5  Substrate oxidation by Mn III -alkylperoxo adducts 1 Direct oxidation of triphenylphosphine. The addition of 100 equiv. PPh 3 to an anaerobic solution of [Mn III (OO t Bu)( 6-Me dpaq)] + (2.0 mM in CH 3 CN) at 298 K resulted in the loss of intensity at 650 nm over the course of two hours, resulting in an electronic absorption spectrum consistent with the generation of Mn II products (Fig. 10, le). The decay of the 650 nm absorption signal could be well-t to a rst-order model, yielding a pseudo-rst-order rate constant k obs (Fig. 10, le inset). A 31 P NMR analysis of the organic products revealed the formation of Ph 3 PO in 70% yield relative to the Mn III -alkylperoxo concentration (Fig. S24 †). EPR analysis of the reaction mixture shows a signal centered at g ¼ 2.03 that is similar in appearance to that of the [Mn II (H 2 O)( 6Me dpaq)]OTf starting material, albeit with the lack of apparent hyperne splitting (Fig. S25 †). 28,39,50 This evidence is in accordance with the featureless UV-vis spectrum of the nal reaction mixture, which is characteristic of a Mn II product (Fig. 10, le). The rate of decay of [Mn III (OO t Bu)( 6Me dpaq)] + increased linearly with increasing concentrations of PPh 3 (Fig. 10, right). A linear t of k obs versus PPh 3 concentration yields a second-order rate constant for PPh 3 oxidation by [Mn III (OO t Bu)( 6Me dpaq)] + of  Further insight into the reaction of PPh 3 with [Mn III (OO t-Bu)( 6Me dpaq)] + was obtained through an Eyring analysis of variable-temperature kinetic experiments (Fig. 11). These experiments yielded DH ‡ ¼ 17.6 AE 1.4 kcal mol À1 , DS ‡ ¼ À12.6 AE 4.6 cal mol À1 K À1 , and DG ‡ ¼ 21.3 AE 2.8 kcal mol À1 at 298 K. The high activation enthalpy and Gibbs free energy of activation account for the sluggishness of this reaction, and the negative entropy of activation is consistent with a bimolecular reaction.

Discussion
Ligand-sphere inuence on the structure-property correlations of Mn III -alkylperoxo complexes The generation of the room-temperature stable [Mn III (OO t Bu)( 6-Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + complexes relied upon previous observations that Mn III centers with higher Lewis acidity give rise to corresponding Mn III -alkylperoxo adducts with shorter and more stable O-O bonds. 16 The basis of this correlation rests on the p-donating properties of the alkylperoxo ligand. According to this model, which was originally proposed by Kovacs, DeBeer, and co-workers, 16 a more Lewis acidic Mn III center fosters greater pdonation from the alkylperoxo p* MO, which strengths the O-O bond. The 6-Me-pyridyl groups of the structurally characterized [Mn III (OOCm)( 6Me dpaq)] + complex give two elongated Mn-N distances of 2.284(4) and 2.394(4)Å. These weak metal-ligand interactions increase the Lewis acidity of the Mn III center, stabilizing the Mn III -alkylperoxo unit. DFT computations lend credence to this model, showing a reduction in the Mulliken charge of [Mn III (OO t Bu)( 6Me dpaq)] + relative to [Mn III (OO t Bu)(dpaq)] + (0.48 and 0.52, respectively). The computations also predict a greater admixture of alkylperoxo character in the Mn-OO t Bu p-antibonding MO for [Mn III (OO t Bu)( 6Me dpaq)] + (Fig. S33 †), which supports stronger p-donation in this complex.
The crystal structure of [Mn III (OOCm)( 6Me dpaq)] + allows us to determine how closely this complex follows previously   observed, making this complex a clear outlier (Fig. 12). It is not completely surprising that the markedly different coordination spheres of [Mn III (OOCm)( 6Me dpaq)] + and the [Mn III (OOR)(N 4 S)] + series would cause such a deviation, as the O-O distance should be a reporter of the entire coordination sphere. It is additionally possible that there is a limit to the extent to which the O-O bond can be elongated in Mn III -OOR complexes, and that the limit is near 1.47Å.
The [Mn III (OOCm)( 6Me dpaq)] + complex also breaks the previously observed correlation that Mn III -alkylperoxo adducts with longer O-O bonds are less stable than those with shorter O-O bonds. 15 [Mn III (OOCm)( 6Me dpaq)] + has an O-O distance at the long end of the [Mn III (OOR)(N 4 S)] + series but has a roomtemperature half-life of ca. 5 days. In contrast, the most stable [Mn III (OOR)(N 4 S)] + complex has a half-life of ca. 5 minutes at 293 K. 16 One caveat that must be noted in comparing the thermal stability of [Mn III (OOCm)( 6Me dpaq)] + with the [Mn III (OOR)(N 4 S)] + series is the difference in solvents (CH 3 CN and CH 2 Cl 2 , respectively). 16 To address this complication, we determined the half-life of [Mn III (OOCm)( 6Me dpaq)] + in CH 2 Cl 2 at 298 K. The [Mn III (OOCm)( 6Me dpaq)] + complex did decay more rapidly in CH 2 Cl 2 than in CH 3 CN (half-life of 3 vs. 8 days, respectively). While solvent does have some effect on the stability of the Mn III -alkylperoxo complex, the solvent change alone cannot account for the dramatic increase in stability of the [Mn III (OOR)( 6Me dpaq)] + complexes relative to the [Mn III (OOR)(N 4 S)] + series. On this basis, while also noting the limited sample size, it is tempting to speculate that the presence of thiolate ligands in the [Mn III (OOR)(N 4 S)] 2+ series severely reduces the stability of the Mn III -alkylperoxo adducts. This conclusion makes it all the more remarkable that the rst isolable Mn III -alkylperoxo adducts contained thiolate ligands.
Thermal decay mechanism of [Mn III (OOCm)( 6Me dpaq)] + The organic products observed upon the decay of cumylperoxometal complexes are oen used to infer the nature of the decay pathway. 51 Homolytic cleavage of the O-O bond produces cumyloxyl radical that can rearrange by b-scission to produce acetophenone and cCH 3 (Scheme 2, path b). Alternatively, heterolytic cleavage of the O-O bond produces cumyl oxyanion that can deprotonate solvent to produce 2-phenyl-2-propanol (Scheme 2, path a). [51][52][53] Previous studies of Mn III -alkylperoxo 16 and some Fe III -alkylperoxo 12  There are recent reports of Mn IV -oxo adducts of the closely related dpaq ligand that react with C-H bonds to yield a Mn III -hydroxo product. 56,57 To clarify the decay pathway of [Mn III (OOCm)( 6Me dpaq)] + , we examined the products formed when the complex decayed in CD 3 CN. In this case, we observed increased formation of acetophenone and decreased formation of 2-phenyl-2-propanol (40 : 50%) compared to the decay in CH 3 CN (26 : 61%). A change in product distribution in deuterated solvent was also observed by Cho et al. in their investigations of a Cu II -alkylperoxo complex. 13 Itoh 2,3 and others 51,52 have rationalized a change in the acetophenone: 2-phenyl-2-propanol distribution in terms of solvent involvement in the decay pathway. In CH 3 CN, the cumyloxyl radical decays by competing reactions: (1) b-scission to yield acetophenone and methyl radical, and (2) hydrogen-atom abstraction from solvent to yield 2-phenyl-2- propanol (Scheme 2, paths c and d). 2,3,51,54,55,58 In CD 3 CN, the rate of the hydrogen-atom abstraction reaction from the solvent is decreased, yielding a marked increase in acetophenone formation by b-scission. When [Mn III (OOCm)( 6Me dpaq)] + decays in PhCN, we observe an even greater increase in acetophenone formation (70%), with only 30% formation of 2-phenyl-2propanol. We attribute this change to the strong C-H bonds in PhCN that suppress the reaction of solvent with the cumyloxyl radical (Scheme 2, path d). Taken together, the change in the distribution of acetophenone: 2-phenyl-2-propanol in CD 3 (Fig. 8). In contrast, the decay rate of [Mn III (OOCm)( 6Me dpaq)] + increases in CD 3 CN by about eight-fold relative to that in CH 3 CN (Fig. S34 †). In addition, the decay rate of [Mn III (OOCm)( 6Me dpaq)] + in CD 3 CN is ve-fold faster than the rate of formation of [Mn III (OH)( 6Me dpaq)] + . These observations are consistent with our proposal that a fraction of the [Mn III (OOCm)( 6Me dpaq)] + complex decays by homolytic O-O cleavage to give a Mn IV -oxo adduct and cumyloxyl radical. In CH 3 CN, the cumyloxyl radical and Mn IV -oxo intermediates react rapidly and preferentially with solvent to give the observed [Mn III (OH)( 6Me dpaq)] + and 2phenyl-2-propanol products. Under these conditions, a relatively small amount of cumyloxyl radical undergoes b-scission to yield acetophenone. In CD 3 CN, the Mn IV -oxo adduct and cumyloxyl radical decay products have slower rates of reaction with solvent, allowing for reaction with [Mn III (OOCm)( 6Me dpaq)] + , which hastens its decay.

Reaction mechanism of [Mn III (OOCm)( 6Me dpaq)] + with PPh 3
To the best of our knowledge, the reactions of [Mn III (OO t Bu)( 6-Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + with PPh 3 at 298 K are the rst observations of direct substrate oxidation by Mn IIIalkylperoxo complexes. The reaction of [Mn III (OOCm)( 6Me dpaq)] + with PPh 3 showed the near exclusive formation of 2phenyl-2-propanol, with only a trace amount of acetophenone (Fig. S35 †). This distribution suggests a change to O-O heterolysis under these conditions. An Eyring analysis for the reaction of [Mn III (OO t Bu)( 6Me dpaq)] + with PPh 3 gives DS ‡ ¼ À12.6 AE 4.6 cal mol À1 K À1 , which is consistent with a bimolecular reaction involving the association of [Mn III (OO t Bu)( 6Me dpaq)] + with PPh 3 to form the activated complex. The change in reaction rate as a function of PPh 3 concentration is further evidence of a direct reaction between the [Mn III (OOR)( 6Me dpaq)] + complexes and PPh 3 . In addition, an ESI-MS analysis of the products of the reaction of [Mn III (OO t Bu)( 6Me dpaq)] + with PPh 3 revealed a peak for [Mn(OPPh 3 )( 6Me dpaq)] + that shis by +2 mass units when the Mn III -alkylperoxo adduct is prepared using t Bu 18 O 18 OH (Fig. S27 †). Thus, the oxygen in the OPPh 3 product derives from the Mn III -alkylperoxo unit.
We propose a reaction mechanism where [Mn III (OOR)( 6-Me dpaq)] + and PPh 3 form an activated complex, with PPh 3 interacting with the proximal oxygen of the alkylperoxo ligand (Scheme 3). Recent reports show that Brønsted and Lewis acids, or the introduction of secondary coordination interaction through pendant amines which act as hydrogen-bond acceptor in an Fe III -OOR (R ¼ H, acyl) adduct could direct heterolytic cleavage. 11,[59][60][61] This interaction between [Mn III (OOR)( 6Me dpaq)] + and PPh 3 may also be able to instigate heterolytic cleavage of the Mn III -alkylperoxo O-O bond. For the [Mn III (OOCm)( 6Me dpaq)] + complex, this decay will lead to the formation of cumyloxy anion, which gives 2-phenyl-2-propanol aer protonation 2,54,55 and a Mn III -species that is reduced to the Mn II product observed by UV-vis and EPR spectroscopy (Scheme 3). The identity of the reductant for the Mn III center is unclear.

Reaction mechanism of [Mn III (OOCm)( 6Me dpaq)](OTf) with DHA
In contrast to the direct oxidation of PPh 3 , the reaction of [Mn III (OO t Bu)( 6Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + with DHA is an indirect oxidation caused by products of the Mn IIIalkylperoxo decay process (Scheme 4). Neither [Mn III (OOR)( 6-Me dpaq)] + complex shows any change in decay rate in the presence of an excess amount of DHA (Fig. S29 †), although the decay solutions reveal the formation of 1.4 equiv. anthracene relative to the initial Mn III -OOR concentration. The notion that the reaction is indirect is further supported by the lack of any change in decay rate when d 4 -DHA is used as substrate.
The formation of 1.4 equiv. anthracene is consistent with the thermal decay of the [Mn III (OOR)( 6Me dpaq)] + complexes by O-O homolysis (Scheme 4). The Mn IV -oxo decay product should be capable of DHA oxidation, as observed for several oxo-manganese complexes. 56,57 This reaction will result in a Mn II -aqua complex, consistent with the observation of a Mn II signal in the EPR spectrum of the nal reaction mixture (Fig. S30 and S31 †). The cumyloxyl radical also generated by O-O homolysis could be responsible for the remaining 0.4 equiv. anthracene.

Conclusions
Inspired by previously developed structure-reactivity correlations, we developed a new ligand derivative ( 6Me dpaq) that provides remarkable stability to Mn III -alkylperoxo complexes. A simple change of two pyridyl groups to 6-Me-pyridyl groups results in new Mn III -alkylperoxo complexes that (i) can be generated using stoichiometric amounts of oxidant rather than large excesses, and (ii) are stable at room temperature. This enhanced stability allowed us to structurally characterize a Mn III -cumylperoxo adduct by X-ray diffraction. In spite of the unusual stability of these Mn III -alkylperoxo adducts, these complexes are the rst members of their class to show direct reactivity with a substrate (triphenylphosphine). This result demonstrates that the ligand-sphere of Mn III -alkylperoxo adducts has great control over reactivity. Examination of the thermal decay of these N 5 À -ligated Mn III -alkylperoxo adducts provides evidence from both homolytic and heterolytic O-O bond cleavage, which is distinct from that observed for Mn IIIalkylperoxo adducts bound by thiolate-containing N 4 S À ligands. While the basis of the enhanced stability of [Mn III (OO t-Bu)( 6Me dpaq)] + and [Mn III (OOCm)( 6Me dpaq)] + will be the subject of future investigations, it is tempting to speculate that the thiolate ligands in the [Mn III (OOR)(N 4 S)] + complexes serve to lower the activation energy for decay. In addition, while the new Mn III -alkylperoxo adducts generally follow a previously identi-ed structural correlation between Mn-N and O-O distances, the observed O-O distance for the Mn III -cumylperoxo adduct is far shorter than expected on the basis of the Mn-N distances. Thus, perturbations to the primary coordination sphere not only affect reactivity but also render this complex an outlier compared to previous compounds. Future work will be aimed at understanding the basis for this outlier status in terms of both structural correlations and chemical reactivity.

Data availability
Crystallographic data are available through the Cambridge Crystallographic Data Centre (CCDC) at https:// www.ccdc.cam.ac.uk/ with structure codes 2048663, 2049911, and 2048664.

Author contributions
A. A. O., J. D. P. and T. A. J. conceived and planned the experiments. A. A. O. and J. D. P. performed all experiments and computations, except those involving X-ray crystallography, which were performed by V. W. D. All authors contributed to data analysis and provided contributions to writing of the nal manuscript.

Conflicts of interest
There are no conicts to declare.