Emma N.
Cook
,
Ian M.
Courter
,
Diane A.
Dickie
and
Charles W.
Machan
*
Department of Chemistry University of Virginia, PO Box 400319 McCormick Rd, Charlottesville, VA 22904-4319, USA. E-mail: machan@virginia.edu
First published on 13th February 2024
The catalytic reduction of dioxygen (O2) is important in biological energy conversion and alternative energy applications. In comparison to Fe- and Co-based systems, examples of catalytic O2 reduction by homogeneous Mn-based systems is relatively sparse. Motivated by this lack of knowledge, two Mn-based catalysts for the oxygen reduction reaction (ORR) containing a bipyridine-based non-porphyrinic ligand framework have been developed to evaluate how pendent proton donor relays alter activity and selectivity for the ORR, where Mn(p-tbudhbpy)Cl (1) was used as a control complex and Mn(nPrdhbpy)Cl (2) contains a pendent –OMe group in the secondary coordination sphere. Using an ammonium-based proton source, N,N′-diisopropylethylammonium hexafluorophosphate, we analyzed catalytic activity for the ORR: 1 was found to be 64% selective for H2O2 and 2 is quantitative for H2O2, with O2 binding to the reduced Mn(II) center being the rate-determining step. Upon addition of the conjugate base, N,N′-diisopropylethylamine, the observed catalytic selectivity of both 1 and 2 shifted to H2O as the primary product. Interestingly, while the shift in selectivity suggests a change in mechanism for both 1 and 2, the catalytic activity of 2 is substantially enhanced in the presence of base and the rate-determining step becomes the bimetallic cleavage of the O–O bond in a Mn-hydroperoxo species. These data suggest that the introduction of pendent relay moieties can improve selectivity for H2O2 at the expense of diminished reaction rates from strong hydrogen bonding interactions. Further, although catalytic rate enhancements are observed with a change in product selectivity when base is added to buffer proton activity, the pendent relays stabilize dimer intermediates, limiting the maximum rate.
Interest in bioinspired and biomimetic systems has spurred the development of a significant number of Fe- and Co-based porphyrinic systems.3,9,12,13 By comparison, Mn has been less widely studied, despite its reactivity toward O2 and prevalence in dioxygen-dependent biological systems.1,2,9 Because of the relatively high spin-pairing energy of Mn ions, open-shell configurations with side-on coordination modes of O2 are often observed to be relatively stable on the potential energy surface of catalytic reactions, limiting activity.2,14–16 This issue arises, in part, as a result of the low basicity of O atoms in side-on Mn–O2 species, preventing effective protonation and cleavage of the O–O bond.2,14 This remains a challenge in the development of Mn-based catalysts for ORR, where destabilization of side-on Mn–O2 intermediates (and other stable reactive oxygen species, ROSs) is crucial to achieve enhanced rates of catalytic turnover (Fig. 1).2,14
There are a number of strategies that can be employed to avoid stable Mn ROS intermediates, including synthetically modifying the supporting ligand framework to tune the electronic structure at the metal center, the introduction of hydrogen-bonding interactions that alter the potential energy surface for Mn–O2 intermediates, or the use of steric hindrance at the active site.2 For instance, to explore how secondary-sphere ligand modification influenced ORR activity and selectivity, Nocera and coworkers introduced a xanthene-based pendent hangman moiety in a Mn tetraphenylporphyrin complex for ORR.17 They found that the hangman group promoted intramolecular proton transfer during catalysis and that hydrogen-bonding interactions between Mn-bound O2 and proton donors helped to favor end-on coordination modes.17 A firm fundamental understanding of how synthetic control over ORR at Mn-based active sites can avoid these potential thermodynamic limitations during catalysis remains an important question. Among the reports on Mn-based ORR catalysts, there are only a handful of non-porphyrinic based systems, despite the possibilities for alternative synthetic approaches for reaction control.2,18–23
Previous studies on the electrochemical ORR mediated by Co(N2O2) molecular complexes with 2,2′-bipyridine based ligand backbone24 found that the introduction of a –OMe pendent relay in the secondary coordination sphere resulted in a shift in selectivity from H2O to H2O2, as well as a shift in the observed rate-determining step (RDS) based on acid strength.25 It was hypothesized that directed protonation of the proximal oxygen of a Co–OOH intermediate was mediated by the pendent –OMe group.25 Reasoning that pendent relays could favor end-on O2 coordination and accelerate proton transfer in analogous Mn-based systems,22,23,26 two new Mn-based non-porphyrinic electrocatalysts for ORR have been prepared. Using an alkyl ammonium-based proton source, it was found that a secondary sphere –OMe pendent relay shifts selectivity to quantitative H2O2 formation under unbuffered conditions (proton donor is present without its conjugate base), although the rate of catalysis is suppressed. However, under buffered conditions (the proton donor and its conjugate base are present in equivalent amounts), hydrogen bonding between the ligand framework and the added acid is mitigated, resulting in a dramatic increase in rate and a shift in selectivity to H2O production. Interestingly, the shift in selectivity occurs in complexes with and without the pendent relay, with systematic variation of added base concentration implying that the change in mechanism is related to the deprotonation of a Mn(III)–OOH intermediate. Further, the complex with the pendent relay experiences a shift in RDS for catalysis under buffered conditions, which is proposed to originate from the pendent relay stabilizing Mn dimer formation relative to O2 binding and activation.
Overall, the proposed mechanism suggests that the 4H+/4e− product H2O is accessed by deprotonating an intermediate Mn(III) hydrogen peroxide complex species, implying that proton activity buffering at a pKa which favors an intermediate metal hydroperoxo is a potential point of reaction control over the selectivity of product formation, as well as over hydrogen-bonding interactions. These observations are representative of the complexity of ORR with Mn active sites: proton transfer is required to facilitate the reduction of superoxide to hydroperoxide (which we attribute to the competition between end-on and side-on coordination modes), but O–O bond scission occurs most efficiently if a second proton transfer to the hydroperoxide does not occur.
UV-vis, ESI-MS, and microanalyses are consistent with the proposed formulation of the Mn complexes shown in Fig. 1. Evans' method measurements exhibited μeff = 4.78 ± 0.11 and μeff = 4.59 ± 0.10 for 1 and 2, respectively; both values are consistent with high-spin d4 Mn(III) complexes (Tables S1 and S2†).27,28 Single crystals suitable for X-ray diffraction studies of complex 2 were grown by slow cooling of a saturated, boiling acetonitrile (MeCN) solution layered with diethyl ether. The solid-state structure of 2 is a dimeric species where a single O atom from each atom is coordinated in the axial position of a second equivalent of complex 2 to create a six-coordinate environment for each Mn center, analogous to our previous studies on a comparable Fe complex (Fig. 2C and D).29
![]() | ||
Fig. 2 (A) Structure of 1 Mn(p-tbudhbpy)Cl, (B) structure of 2 Mn(nPrdhbpy)Cl, (C) molecular structure of Mn(nPrdhbpy)Cl 2 obtained from single crystal X-ray diffraction studies by omitting one Mn complex from the (D) dimeric solid-state species. Purple = Mn, red = O, green = Cl, gray = C; thermal ellipsoids 50%, H atoms and disordered atoms omitted for clarity. CCDC 2255849.† |
N,N′-Diisopropylethylammonium hexafluorophosphate (DIPEAHPF6) was synthesized according to a previously reported procedure.30 Toluene was brought to reflux in the presence of ammonium hexafluorophosphate (PF6) and N,N′-diisopropylethylamine (DIPEA), resulting in the precipitation of a solid product after 48 hours. The suspension was cooled to 0 °C on an ice bath before being filtered and washed with dichloromethane (DCM). Rotary evaporation of the DCM solvent yielded a spectroscopically pure product: NMR spectroscopic characterization and microanalysis were consistent with a N,N′-diisopropylethylammonium salt with a PF6− counteranion. Single crystals suitable for X-ray diffraction studies were grown via slow evaporation from a concentrated DCM solution (Fig. S3†). 1H NMR spectroscopy was used to estimate a pKa of 18.7 for DIPEAHPF6 in MeCN via titration and competition experiments (see ESI†).31
Initial addition of 10 mM of the DIPEAHPF6 proton donor to both 1 and 2 under inert conditions resulted in the loss of reversibility of the Mn(III/II) redox couple as well as a shift to more positive potentials, Ep,a = −0.56 V vs. Fc+/Fc for 1 (Fig. S9†) and −0.53 V vs. Fc+/Fc for 2 (Fig. S16†). This is consistent with a contribution from an EC mechanism (reversible electron transfer followed by an irreversible chemical reaction), which is proposed to be the strong hydrogen-bonding interactions and protonation of an O atom in the inner-coordination sphere following formal reduction to Mn(II), consistent with our previous studies on acids with similar pKa values.22,32 Although titration of increasing amounts of DIPEAHPF6 beyond 10 mM did not cause a further potential shift for 1, the maximum positive potential shift for 2 was reached after the addition of 20 mM. Similarly, titration of increasing amounts of the DIPEAHPF6 proton donor in the presence of 10 mM DIPEA revealed that the Mn(III) reduction potential observed for 1 and 2 shifted to more positive potentials, however this shift was not dependent on the concentration of DIPEAHPF6 (Fig. S11 and S17†).
The observation of these shifts in the presence of the proton donor and its conjugate base, as well as the absence of a Nernstian relationship suggested that the axial chloride ligand was participating in related chemical steps during the EC mechanism. To better understand the role of DIPEA and DIPEAHPF6 in axial Cl− ligand solvation, CVs of 1 (Fig. S13 and S14,† respectively) were taken in the presence of 0.1 M TBACl. The addition of increasing amounts of both DIPEA by itself and DIPEAHPF6 with a fixed concentration of DIPEA (buffered) in the presence of excess Cl− under Ar saturation conditions suppressed positive potential shifts of the Mn(III/II) redox feature in 1. The suppression of the potential shift in the presence of excess Cl− suggests that the shift observed originally is due to axial ligand solvation. Since it was observed that increasing concentrations of DIPEAHPF6 led to a loss of reversibility for the Mn(III/II) feature, these data taken in aggregate imply that axial Cl− ligand loss enables hydrogen-bonding interactions and protonation of the O atom in the inner coordination sphere upon reduction.
Similar studies were conducted on 2, where the addition of 0.1 M TBACl to a solution of 2 resulted in a negative potential shift of a quasireversible Mn(III/II) feature to −0.76 V vs. Fc+/Fc (Fig. S22†). Addition of 10 mM DIPEA caused a positive potential shift of this Mn(III/II) redox feature back to −0.56 V vs. Fc+/Fc (Fig. S22†) and the titration of increasing amounts of DIPEAHPF6 under buffered conditions suppressed the observed positive potential shift (Fig. S23†). As was the case with complex 1, it is hypothesized that the presence of both DIPEA and DIPEAHPF6 contribute to axial Cl− ligand loss for 2, with irreversibility again suggesting strong hydrogen-bonding interactions and protonation involving a ligand O atom bound to Mn. Consistent with this, UV-vis spectroscopic analysis of 1 and 2 in the presence of DIPEA and DIPEAHPF6 showed no evidence of interaction with Mn(III) (Fig. S58–S61†).
Under O2 saturation in the presence of 10 mM DIPEAHPF6, there is an increase in current at the Mn(III/II) reduction feature for both 1 (Fig. S9†) and 2 (Fig. S16†), indicative of electrocatalytic activity for the ORR. Likewise, under O2 saturation in the presence of buffered DIPEAHPF6 (1:
1 ratio of ammonium to its conjugate base), there is an increase in current density at the Mn(III/II) redox couple for both 1 and 2, consistent with catalytic O2 reduction (Fig. 3). Effective overpotentials (η) for the ORR by 1 and 2 calculated under buffered conditions were determined to be 0.58 and 0.56 V, respectively (see ESI†). Notably, the catalytic current density observed under buffered conditions with complex 2 is much higher than unbuffered conditions (Fig. S19†), suggesting that the presence of conjugate base enhances electrocatalytic ORR by 2, vide infra. Oxidative current due to amine oxidation precludes the use of rotating ring-disk methods for H2O2 detection under these conditions for both 1 and 2 (Fig. S24†).
Variable concentration studies with 1 and DIPEAHPF6 under unbuffered conditions revealed a rate law for the ORR with first-order dependencies on [1] and [O2] and zero-order dependencies on [DIPEAHPF6] and [Cp*2Fe] (Fig. S36–S39†), corresponding to the rate law shown in eqn (1).
Rateunbuffered = kcat[Mnp-tbu]1[O2]1 | (1) |
A Ti(O)SO4 colorimetric assay was used to determine the selectivity of the ORR as previously described.23,33 Aliquots were taken after the reaction was allowed to reach completion and the amount of H2O2 produced was quantified. Selectivity testing under unbuffered conditions revealed that 1 is 64.2 ± 6.9% selective for H2O2 corresponding to an ncat of 2.72 (Fig. S26† and Table 1). Under these unbuffered conditions (only proton donor present), control studies showed no degradation of H2O2via disproportionation, with quantitative recovery of H2O2 (Fig. S28 and Table S6†).
Unbuffered | Buffered (15 s) | |||
---|---|---|---|---|
% H2O2 | % H2O | % H2O2 | % H2O | |
1 | 64.2 ± 6.9 | 35.8 ± 6.9 | 18.1 ± 4.1 | 81.9 ± 4.1 |
2 | 96.2 ± 4.1 | 3.8 ± 4.1 | 37.9 ± 6.7 | 62.1 ± 6.7 |
Similarly, variable concentration studies of ORR by 2 with unbuffered DIPEAHPF6 showed first-order dependencies on [2] and [O2] (Fig. S45–S49†). Selectivity studies showed 96.2 ± 4.1% H2O2 selectivity corresponding to an ncat = 2.08 (Fig. S27 and Table S5†), with control studies showing that 93.6 ± 4.6% H2O2 was recovered after 20 minutes (Fig. S29 and Table S7†), consistent with very slight activity for H2O2 disproportionation. Notably, the observation of an increased amount of the H2O2 product for complex 2 – which contains pendent proton donor relays – mirrors similar observations made with Co-based analogues of these compounds previously.25 Thus, eqn (2) is proposed as the rate law for ORR mediated by 2 under unbuffered conditions.
Rateunbuffered = kcat[MnnPr]1[O2]1 | (2) |
Subsequent re-examination of 1 under buffered conditions (equal amounts of ammonium proton donor and its conjugate base) showed first-order dependencies on [1] and [O2] and zero-order dependencies on [1:
1 DIPEAHPF6
:
DIPEA] and [Cp*2Fe] (Fig. S40–S44†). Systematically varying the concentration of either [DIPEAHPF6] or [DIPEA] against a fixed concentration of the other also exhibited no concentration dependence on the catalytic reaction (Fig. S42†). Selectivity studies revealed that after 15 s, 1 showed 81.9 ± 4.1% selectivity for H2O as the product (Fig. S30† and Table 1). H2O2 disproportionation was observed under catalytic conditions: only 10.8 ± 6.6% H2O2 was recovered after 105 s with 1
:
1 DIPEAHPF6
:
DIPEA present (Fig. S32 and Table S10†). Interestingly, control studies under buffered catalytic conditions without the presence of O2 revealed minimal degradation of H2O2 by the Mn(II) form of the complex (Fig. S34 and Table S12†), suggesting that the observed reaction selectivity difference is not due to disproportionation alone. These results are summarized in the following rate law, eqn (3), for ORR mediated by 1 under buffered conditions:
Ratebuffered = kcat[Mnp-tbu]1[O2]1 | (3) |
Conversely, variable concentration studies of ORR catalyzed by 2 under buffered conditions revealed mechanistic differences in comparison to the data obtained for 1 and 2 under unbuffered conditions, as well as a change in selectivity. The ORR mediated by 2 showed a second-order dependence on [2], a first-order dependence on [1:
1 DIPEAHPF6
:
DIPEA], an inverse first-order dependence on [Cp*2Fe] and a zero-order dependence on [O2] (Fig. S49–S53†). Unlike 1, when [DIPEA] was varied against a fixed concentration of [DIPEAHPF6] with 2, a first-order dependence on rate was observed; experiments where [DIPEAHPF6] was varied against a fixed [DIPEA] showed a zero-order dependence. These results lead to the proposed rate law shown in eqn (4). The observed first-order dependence on [DIPEA] implies that the dependence observed on [1
:
1 DIPEAHPF6
:
DIPEA] is a result of varying DIPEA concentration. This interpretation is validated by the zero-order dependence on [DIPEAHPF6] with fixed [DIPEA], ruling out proton activity or proton donor concentration as influencing the observed reaction rate.
Rate = kcat[MnnPr]2[DIPEA]1[Cp*2Fe]−1 | (4) |
Selectivity studies of ORR under buffered conditions showed that after 15 s, 2 is 62.1 ± 6.7% selective for H2O (Fig. S33† and Table 1). Control studies showed H2O2 degradation consistent with disproportionation: after 150 s only 3.81 ± 0.83% of H2O2 was recovered under buffered conditions with complex 2. Again, H2O2 reduction was not observed under the buffered catalytic conditions when placed under an inert N2 atmosphere, suggesting that the change in selectivity is not exclusively due to disproportionation (Fig. S35 and Table S13†).
The most probable thermodynamic step from the starting complex [Mn(nPrdhbpy)(Cl)]0 under protic conditions is reduction followed by a hydrogen–bonding interaction between the ligand framework and an equivalent of [DIPEAH]+ to generate [Mn(nPrdhbpy[AH])(Cl)]0 (Fig. 5). Note that the [AH] notation indicates the hydrogen-bonding interaction involving a Mn-bound O atom from the ligand framework. Although formal proton transfer to the ligand framework with loss of DIPEA is exergonic by −7.2 kcal mol−1, the binding of O2 with accompanying Cl− release to generate the cationic species [Mn(nPrdhbpy[H])(η2-O2)]+ is uphill by 34.2 kcal mol−1, precluding its involvement in the catalytic cycle. Instead, O2 binding with loss of [DIPEAH][Cl] is endergonic by 20.2 kcal mol−1, which is consistent with experimental observations, generating [Mn(nPrdhbpy)(η2-O2)]0. In this structure, the bound O2 is in a side-on coordination mode with a bond length of 1.309 Å, consistent with reduction to superoxide, O2˙−. Reduction and protonation of this species had a favorable CEPT pathway (−0.01 V vs. Fc+/0), where proton transfer has occurred from [DIPEAH]+ to generate [Mn(nPrdhbpy)(η1-O2H)]0. The shift in coordination mode of the O2 fragment from side-on to end-on reflects additional reduction: the bond lengthens to 1.451 Å, consistent with a peroxide. Protonation of the proximal O atom to generate [Mn(nPrdhbpy)(η1-O2H2)]+ by an equivalent of [DIPEAH]+ is endergonic by 9.9 kcal mol−1 and displacement by Cl− to facilitate H2O2 release is favorable by −23.6 kcal mol−1. These reaction steps align with experimental observations on the catalytic cycle which produces H2O2 under unbuffered conditions.
Next the favorability of dimerization from the hydroperoxide was assessed to explore the implied dimeric pathway to H2O production (Fig. S77†). The experimental studies described above established the viability of H2O2 as an intermediate species, therefore the [Mn(nPrdhbpy)(η1-O2H2)]+ adduct was considered as the starting point. Deprotonation of Mn-bound H2O2 by DIPEA to generate [Mn(nPrdhbpy)(η1-O2H)]0 is favorable by 9.9 kcal mol−1. Subsequent dimerization of [Mn(nPrdhbpy)(η2-O2H)]0 with an equivalent of [Mn(nPrdhbpy)(Cl)]0 is endergonic by 19.7 kcal mol−1 and generates [Mn(nPrdhbpy)(O)]0 (S = 3/2), [Mn(nPrdhbpy)(OH)]+ (S = 3/2), and Cl− as the products. Attempts to examine a stabilized bridging hydroperoxo dimer were unsuccessful, homolytic O–O bond scission occurred spontaneously in all cases. From the terminal manganese oxo [Mn(nPrdhbpy)(O)]0, a calculated CEPT potential of +1.00 V vs. Fc+/Fc was obtained for the production of a Mn(III) hydroxide [Mn(nPrdhbpy)(OH)]0, while the convergent pathway via the reduction of [Mn(nPrdhbpy)(OH)]+ was estimated to be +0.62 V; both processes are expected to be facile at the considered operating potential of −0.55 V vs. Fc+/Fc. Protonation of the neutral Mn–OH species to make the corresponding aquo species is uphill by 1.2 kcal mol−1 and subsequent H2O loss with Cl− coordination to Mn to regenerate the starting species and close the H2O2 reduction cycle is downhill by −20.2 kcal mol−1.
![]() | ||
Scheme 1 Proposed catalytic cycle for ORR, with the key equilibrium for switching between the two reaction pathways indicated in gray. |
Following this activation of the catalyst, rate-limiting O2 binding to Mn(II) to form a Mn(III)-superoxide species, iii, occurs with loss of [DIPEAH][Cl]. Based on this assignment, species ii is considered to be the resting state of the catalytic cycle under these conditions. Reduction and protonation by a CEPT pathway results in the formation of a Mn-hydroperoxo species, iv, whose protonation leads to the formation of a Mn–H2O2 intermediate, v. This is the primary reaction pathway for 1 (64.2 ± 6.9% selectivity) and 2 (96.2 ± 4.1% selectivity) under unbuffered conditions, as suggested by selectivity studies and the relative absence of activity for H2O2 disproportionation in separate testing. The observed selectivity enhancement for H2O2 during ORR mediated by 2 is also aligned with the participation of the pendent methoxy group in hydrogen bonding interactions with added acid, as we have observed previously.25 This mechanistic proposal is consistent with the general thermodynamic landscape obtained by DFT methods (Fig. 5). It is worth noting that H2O2 disproportionation is observed by the parent Mn(III) species of both complexes in the presence of DIPEA. Further, control studies with Mn(II) revealed minimal interaction with H2O2 (Fig. S34†), suggesting that disproportionation is reliant on the availability of formally Mn(III) complexes.
Under buffered conditions, complex 2 demonstrated a substantial increase in the observed catalytic rate in comparison to unbuffered conditions with an accompanying mechanistic divergence involving dimerization. While we do not observe a change in the rate-determining step or infer a change in the resting state for 1, it is proposed that the shift in selectivity from H2O2 under unbuffered conditions to H2O under buffered conditions is similarly due to the accessibility of a dimerization pathway in the presence of added base. Based on the experimental and computational data presented, it is likely that the observed mechanistic divergence arises from differences in equilibrium control over the speciation of complexes iv and v, as depicted in eqn (5).
Mn(O2H) + BH+ ![]() | (5) |
Therefore, under buffered conditions, an alternative reaction pathway dominates the observed selectivity. With suitable concentrations of base present, eqn (5) shifts to the left, favoring the hydroperoxo species, iv, which becomes the resting state for 2 under buffered conditions. Species iv dimerizes with an equiv of the Mn(III) complex i leading to the formation of species vi and vii from spontaneous O–O bond cleavage, with accompanying chloride loss. Following one-electron reduction or a one-electron CPET process, these products converge at a Mn(III) hydroxide species viii. Following this, formal protonation and displacement of H2O by Cl− complete the catalytic cycle. For complex 2, spontaneous O–O bond cleavage from dimerization becomes the RDS of the reaction, which is attributed to the pendent relay groups slowing dimer formation through intermolecular coordination and hydrogen-bonding. This proposal is consistent with the second-order rate dependence on [Mn] and first-order rate dependence on [DIPEA] under buffered conditions. Further, the inverse first-order dependence observed for [Cp*2Fe] implies that Mn(III) is required for dimerization, rather than Mn(II). This reaction pathway is in good agreement with that computed using DFT methods (Fig. S77†).
The differences in ORR catalyzed by 2 under unbuffered and buffered conditions in comparison to 1 suggest an important role for the pendent –OMe moiety in the secondary coordination sphere, as well as for the added base. Under unbuffered conditions, catalytic activity of 2 is suppressed in comparison to 1, which is observed under both electrochemical and spectrochemical conditions. This is attributed to strong hydrogen-bonding interactions between the –OMe groups and DIPEAHPF6 kinetically inhibiting O2 binding, the rate-determining step of the catalytic cycle. In addition to catalytic suppression, we observe a shift in selectivity from approximately 64% for H2O2 by 1 to 96% for H2O2 by 2 under unbuffered catalytic conditions. As introduced above, we propose that the observed shift in product selectivity is due to the hydrogen-bonding interaction of the –OMe group assisting in proton transfer to the proximal Mn–OOH oxygen in species iv, as we have previously suggested in similar Co systems.25
Interestingly, upon the addition of the DIPEA conjugate base to catalytic conditions with 2, catalysis is significantly enhanced under electro- and spectrochemical conditions. There is an accompanying change in mechanism, where the equilibrium responsible for H2O2 formation becomes unfavorable, allowing dimerization with accompanying O–O bond scission to become rate-determining. It is also likely that DIPEA mitigates the effects of the strong H-bonding interaction between DIPEAHPF6 and the –OMe moiety, allowing an increased rate of hydroperoxo intermediate generation. Indeed, in control studies with 2 chemically reduced to Mn(II) by CoCp2in situ, the addition of DIPEAHPF6 inhibited reactivity with O2: under unbuffered conditions the reaction between 2 and O2 took approximately 100 min to go to completion (Fig. S72†), compared with approximately 25 m for complex 1 (Fig. S65†). It is also worth noting the difference in selectivity under buffered conditions between 1 and 2, where 1 is 81.9% selective for H2O (Table 1) and 2 is only 62.7% selective for H2O. We attribute this difference to the H-bonding ability of the –OMe groups of 2 promoting the formation of species v during catalysis by directing the proton donor. However, the dimer pathway to water formation is the primary pathway under buffered conditions, as evidenced by the shift in observed rate law and the shift in product selectivity to H2O.
The observed rate constants (Rfit/ncat values) under identical buffered conditions for 1 and 2 were kobs = 1.23 ± 0.17 × 10−1 s−1 and kobs = 0.706 ± 0.25 × 10−1 s−1, respectively. While addition of one equivalent of conjugate base for every equivalent of proton donor mitigates some of the hydrogen-bonding induced suppression observed for 2 under unbuffered conditions, the accompanying inhibition of dimerization results in a slight decrease in the rate of catalysis relative to 1. Therefore, it has been demonstrated that the introduction of a pendent relay into homogeneous molecular Mn-based electrocatalysts plays an essential role in ORR through the hydrogen bond-assisted stabilization of key intermediates and by favoring the direct protonation of the proximal oxygen of the Mn–OOH intermediate. The relative hydrogen-bond donor ability of DIPEAHPF6 is such that inhibition of catalysis can also occur, which is mitigated through the introduction of the conjugate base. However, the mitigation of this effect results in a change in mechanism, where O–O bond scission occurs spontaenously upon dimerization. Thus, the conjugate base plays an important role during catalysis, allowing for an on-cycle dimer pathway that shifts the reaction pathway towards the formation of water.
It is worth briefly discussing the comparison between the data reported here for 1 and those reported previously with phenolic proton donors for a related complex with additional tert-butyl substituents.22,23 With buffered phenolic proton donors, the observed rate law for the ORR mediated by this more sterically hindered complex showed first-order dependencies on [catalyst] and [O2], but no dependence on proton donor activity. Similarly, the results obtained in the presence of a buffered ammonium proton donor with higher activity (pKa(MeCN) = 18.7) described here again demonstrate that proton activity is not relevant to the observed rate laws. The observed relevance of an off-cycle EC reaction involving the protonation of the ligand framework also aligns with the results of a potential-pKa diagram obtained during the previous study where acids with pKa < 20.11 were found to protonate the ligand framework of the tetra tert-butylated complex. The systems reported here show greater stability with respect to any H2O2 generated, which we had previously observed not to be the case in the presence of phenol and phenolate derivatives, consistent with the ease with which they can be oxidized as well as their competency as ligands relative to the sterically hindered ammonium/amine pairs used here.49–51 Indeed, control testing shows that the DIPEA does not react under experimental conditions (Fig. S58 and S59†) and is too sterically hindered to coordinate to Mn.
Footnote |
† Electronic supplementary information (ESI) available: Computational coordinates, experimental and computational details, as well as supplementary data from additional cyclic voltammetry, spectrochemical, spectroscopic experiments are available. Deposition numbers 2255849 and 2255850 contain the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre. CCDC 2255849 and 2255850. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3sc02611f |
This journal is © The Royal Society of Chemistry 2024 |