Nitric oxide activation facilitated by cooperative multimetallic electron transfer within an iron-functionalized polyoxovanadate–alkoxide cluster

Cooperative multimetallic electron transfer to accommodate substrate binding.


Introduction
The chemical reactivity of nitric oxide (NO) has captivated the eld of bioinorganic chemistry, due to the participation of this small molecule in vasodilation, mammalian signalling, and immune defence processes. 1 Given the established signicance of this substrate in biological systems 2,3 and the biogeochemical nitrogen cycle, 4,5 the interaction of NO with metal centres, specically iron, has been an area of intense research. Indeed, the prevalence of heme-and non-heme-containing metalloenzymes in NO reductases has driven interest in understanding the electronic structure of the {FeNO} subunit in metalloproteins and model complexes. [6][7][8] A combination of spectroscopic, crystallographic, and theoretical methods have shown that substrate binding and reduction are key steps in NO activation. 2,3,[9][10][11][12][13] However, despite reports describing the behaviour of NO with ferric and ferrous heme complexes, the chemistry of this substrate with non-heme derivatives remains underdeveloped (Fig. 1). Toward a more complete understanding of the redox chemistry involved during NO activation, the synthesis, characterization, and reactivity of non-heme models, capable of supporting variable oxidation states of the {FeNO} subunit, are of interest.
The activation of NO requires the simultaneous transfer of multiple electrons and protons to the substrate. In nature, similar chemical transformations of gaseous substrates (e.g. H 2 , Fig. 1 Select examples of non-heme high-spin 5-and 6-coordinate {FeNO} complexes. N 2 , CO 2 ) are oen mediated by metalloenzymes that feature multiple, closely associated metal centres within the active site. [14][15][16][17][18][19][20] Cooperative, multimetallic reactivity has also been noted in heterogeneous systems, where reducible metal-oxide supports have been demonstrated to enhance the activity of metal-catalysts for the activation of small molecules. [21][22][23] In an effort to design homogeneous complexes that mimic the activity of these reactive multimetallic assemblies, our group 24,25 and others [26][27][28][29][30][31] have recently reported heterometallic complexes that possess varying degrees of multimetallic electronic communication. Extensive structural and spectroscopic characterization of these clusters has illustrated the participation of the assembly in molecular redox tuning and the storage of electron density. While these studies have expanded our understanding of the static electronic properties of well-dened multinuclear systems, comparatively less is known about the role of electron transfer across a multimetallic assembly during substrate activation.
Recently, our research group has reported the synthesis and characterization of a family of iron-functionalized polyoxovanadate-alkoxide (FePOV-alkoxide) complexes ( Fig. 1). 24,25,32 Using a suite of analytical techniques, we have demonstrated that the metal-oxide metalloligand functions as a redox reservoir for the ferric centre, storing reducing equivalents across the vanadyl ions in a delocalized cloud of electron density. The established spectroscopic handles, unique to this heterometallic framework, provide the opportunity for in situ analysis of the oxidation state of the metal-oxide scaffold. Therefore, our FePOV-alkoxide complexes are ideal for the investigation of electron transfer within heterometallic assemblies during substrate activation. Given the well-described electronic structures of the {FeNO} subunit, and the large degree of substrate sensitivity to the electronic environment of the bound iron centre, 33 we selected NO as an ideal substrate to investigate the reactivity of FePOValkoxide clusters. Herein, we report our results that showcase the ability of the POV-alkoxide scaffold to shuttle stored electron density to NO through the ferric centre. This work highlights the role of electron transfer within a multimetallic construct during chemical transformations and substrate activation.

Reactivity of NO(g) with an FePOV-alkoxide cluster
To investigate the reactivity of NO with high-spin ferric centres embedded within a multimetallic cluster, NO(g) was added to a solution of the neutral, FePOV-alkoxide cluster, [V IV 4 V V O 6 (OCH 3 ) 12 Fe III ] (1-V 5 Fe) (Scheme 1). A gradual colour change from dark-green to brown was observed. Analysis of the reaction mixture via 1 H NMR spectroscopy revealed formation of a new product, with three paramagnetically shied and broadened resonances located at 13.86, 20.12, and 127.44 ppm (Fig. S1 †). Formation of the desired NO adduct of the FePOV-alkoxide cluster, [V 5 O 6 (OCH 3 ) 12 {FeNO}] (2-V 5 FeNO), was conrmed by infrared spectroscopy. A new band corresponding to v(NO) was located at 1732 cm À1 (Fig. 2, Table 1), substantially shied from that of free NO (1900 cm À1 ). 34,35 Due to the coordinative saturation of the vanadyl ions of the FePOV-alkoxide cluster and the shi in v(NO) consistent with formation of an iron-nitrosyl subunit, 32,36-38 we concluded that NO most likely binds to the cluster through the vacant site of the 5-coordinate, square pyramidal Fe(III) centre of the multinuclear assembly (Scheme 1).
Given the propensity of NO to disrupt the molecular structure of self-assembled, cluster complexes, 39,40 retention of the polynuclear Lindqvist core following exposure of 1-V 5 Fe to NO was evaluated by IR and electronic absorption spectroscopies. Two characteristic bands of the Lindqvist core were observed in the IR spectrum of 2-V 5 FeNO, corresponding to v(V]O t ) and v(O b -CH 3 ) at 968 and 1022 cm À1 , respectively (O t ¼ terminal vanadyl oxygen atoms, O b ¼ bridging oxygen atoms of methoxide ligands) (Fig. 2, Table 1). Additionally, the electronic absorption spectrum of 2-V 5 FeNO reveals two intervalence charge-transfer (IVCT) bands, located at 382 nm (3 ¼ 5.06 Â 10 3 M À1 cm À1 ) and 982 nm (3 ¼ 5.80 Â 10 2 M À1 cm À1 ) (Fig. 3). These features correspond to d xy (V IV ) / d x 2 Ày 2(V V ) and d xy (V IV ) / d xy (V V ) electronic transitions, respectively. Previous reports have established that these absorptions are characteristic of mixed-valent, POV-alkoxide clusters, thus demonstrating retention of the Robin and Day Class II delocalization of electron density across the cluster core upon coordination of NO. 25,41 Collectively, IR and electronic absorption spectroscopy conrm retention of the multimetallic assembly following substrate activation. 24,25,41,42 The reactivity of complex 1-V 5 Fe and NO is surprising, as metalloproteins and model complexes containing high-spin ferric centres have been reported to have a lower affinity for this substrate, in comparison to their ferrous congeners. 43,44 This unusual binding of substrate at an Fe III ion prompted the investigation into the participation of the metal-oxide metalloligand in substrate activation. To establish the role of multimetallic, cooperative reactivity, a ve coordinate ferric complex, Fe III (salen)Cl (salen ¼ N,N 0 -bis(salicylidene)ethylenediamine) was reacted with NO(g). 45 This monometallic complex was selected as a suitable model for the FePOV-alkoxide clusters, due to their similar iron coordination environments (pseudo-square pyramidal) and the redox non-innocence of the respective organic and inorganic ligands. However, exposure of Fe III (salen)Cl to 20 equiv. of NO(g) yields no reaction. The disparity in the reactivity of the two, 5-coordinate, high-spin ferric complexes (Fe III (salen)Cl and 1-V 5 Fe) indicates that the POV-alkoxide scaffold is directly involved in NO activation, most likely through the efficient transfer of electron density to the iron centre (vide infra).

Analysis of reduced and oxidized derivatives of 2-V 5 FeNO
Elucidation of the oxidation state distribution of metal ions in complex 2-V 5 FeNO is necessary to dene the role of the vanadium-oxide metalloligand ligand during NO activation. Previously, our group 25 and others 41,42,46 have demonstrated that the isolation of the reduced and oxidized derivatives of multimetallic complexes aids in the characterization of the charge distribution of metal ions that compose the cluster core. To assess the accessibility of a redox series of the NOfunctionalized FePOV-alkoxide ({FeNO}POV-alkoxide) clusters, the electrochemical prole of 2-V 5 FeNO was explored via cyclic voltammetry (CV). A voltammogram possessing four redox events located at E 1/2 ¼ +0.03, À0.58, À1.16, and À2.68 V vs. Fc/Fc + was observed (Fig. 4). The reversibility of these events was established by varying the scan rate from 20 to 500 mV s À1 and tting the data to the Randles-Sevcik equation (Fig. S2 †). Linearity of each fully-reversible redox event was conrmed by plotting the current density (j p reported in A cm À2 ) versus the square root of the scan rate (v 1/2 reported in V s À1 ). 25,47 Comparison of the CV of complex 2-V 5 FeNO to that of the starting material, 1-V 5 Fe, illustrates the drastic changes in the redox properties of the cluster upon coordination of NO, suggesting substrate coordination has a direct effect on the electronic structure of the multimetallic assembly. Table 1 Comparison of spectroscopic parameters for 2-V 5 FeNO, 3-V 5 FeNO À , 4-V 5 FeNO 2À , 5-V 5 FeNO + and 6-V 5 FeNO 3À to selected {FeNO} complexes  The rst three redox events in the CV of complex 2-V 5 FeNO, observed at E 1/2 ¼ +0.03, À0.58, and À1.16 V, possess peak-to peak separations (DE 1/2 ) of approximately 0.60 V. Similar values of DE 1/2 have been observed in the case of FePOValkoxide clusters 25 and their homometallic, hexavanadate congeners, 41,42 and are associated with redox events localized to the vanadyl ions. In the case of complex 2-V 5 FeNO, the DE 1/2 values for these redox events are similar to those reported for the parent complex, 1-V 5 Fe, resulting in a similar range of comproportionation constants (K c ) between the two series of FePOV-alkoxide clusters (K c ¼ 8.8 Â 10 9 to 2.7 Â 10 10 ; 48 To unambiguously corroborate the hypothesis that the evenly spaced redox events correspond to vanadium-based electrochemical processes, the chemically reduced and oxidized derivatives of 2-V 5 FeNO were isolated. Exposure of 2-V 5 FeNO to stoichiometric equivalents of cobaltocene afforded access to the one-and two-electron reduced species, namely 3-V 5 FeNO À and 4-V 5 FeNO 2À (Scheme 2). Oxidation of 2-V 5 FeNO with AgClO 4 at À40 C led to the formation of 5-V 5 FeNO + . The NO adducts (3-V 5 FeNO À and 5-V 5 FeNO + ) could be independently synthesized by reaction of NO(g) with the appropriate FePOV-alkoxide precursor ([V 5 O 6 (OCH 3 ) 12 Fe] À and [V 5 O 6 (OCH 3 ) 12 Fe] + , respectively) (Scheme 2). In all cases, the products were analysed by 1 H NMR spectroscopy (Fig. S3-S5 †).

5-V 5 FeNO
Upon sequential reduction from 5-V 5 FeNO + to 4-V 5 FeNO 2À , the IR spectra reveal that v(V]O t ) decreases from 974 to 935 cm À1 , with corresponding increases in the values of v(O b -CH 3 ) (1003 to 1055 cm À1 ) (Fig. 2, Table 1). Similar changes in the v(V]O t ) and v(O b -CH 3 ) stretching frequencies have been reported for the redox series of homo-and heterometallic POValkoxide complexes. 25,41,42 Furthermore, with each electron added to the system, v(NO) shis $30 cm À1 toward lower energies (Fig. 2, Table 1). This small change in stretching frequency of the nitrosyl ligand is inconsistent with the direct reduction of the iron-nitrosyl moiety, as this change in oxidation state of the {FeNO} subunit is typically associated with changes in v(NO) of $100 cm À1 . 37,38,49 Instead, the decrease in stretching frequency of v(NO) compares favourably to changes observed in the {FeNO} 6/7 dithiolene systems. 50,51 The redox active dithiolene ligands of these complexes are able to store equivalents of electron density upon sequential reduction from the mono-cationic to the mono-anionic species, [Fe(NO)(S 2 C 2 -R 2 ) 2 ] n (R ¼ p-tolyl; n ¼ +1, 0, À1). As a result of ligand participation in the reduction events of the complex, retention of the {FeNO} 6 oxidation state was observed throughout the series. IR spectroscopy of the {FeNO} dithiolate complexes revealed small shis in v(NO) ($35 cm À1 /e À ), consistent with the changes in v(NO) observed in the spectra of the complexes 2-5. These small, yet distinct changes are a consequence of the dependence of the electronic structure of the {FeNO} moiety on the oxidation state of the redox-active ligand.
Likewise, electronic storage across a metal-oxide base has been observed previously, in the case of [LFe 3 (PhPz) 3 56,57 Agapie and coworkers demonstrate through spectroscopic and crystallographic investigations that changes in oxidation state of the tetra-iron cluster, [LFe 3 (PhPz) 3 OFeNO] 1+ , occur exclusively across the distal iron centres. Mössbauer parameters of complexes [LFe 3 (PhPz) 3 OFeNO] n+ (n ¼ 1, 2, 3) reveal little change in the isomer shi of the apical iron centre upon oxidation of the cluster core, indicating the retention of the {FeNO} 7 state for the iron-nitrosyl adduct across the redox series. Furthermore, nominal changes in v(NO) ($30 cm À1 /e À ) corroborate that no change in the oxidation state of the {FeNO} 7 subunit occurs during cluster oxidation. Collectively, these results establish the role of the tri-iron base as a redox-reservoir for the apical iron centre. The POV-alkoxide scaffold acts in analogy to the iron-oxide base of [LFe 3 (PhPz) 3 OFeNO] 1+ during changes in oxidation state of the cluster, as established via infrared analysis.
The electronic absorption spectra of complexes 2-5 provide additional insight into the electronic structure of the heterometallic Lindqvist cluster. Absorptions diagnostic of IVCT events between V(IV) and V(V) centres are observed in complexes 2-V 5 FeNO, 3-V 5 FeNO À , and 5-V 5 FeNO + , suggesting the presence of a mixed-valent POV-alkoxide core. 25,41,42 In previous work, similar features have been considered tell-tale signs of extensive delocalization of electron density across the POV-alkoxide assembly, 25,41,42,46,52 In contrast, the electronic absorption spectrum of 4-V 5 FeNO 2À has no IVCT absorptions. Instead, the spectrum contains a weak feature at 542 nm (3 ¼ 661 M À1 cm À1 ), which has been previously assigned to a forbidden d xy (V IV ) / d x 2 Ày 2 (V IV ) excitation of an isovalent POV-alkoxide cluster. 25 As such, the charge distribution for complex 4-V 5 FeNO 2À can be described as [V IV 5 O 6 (OCH 3 ) 12 {FeNO}] 2À . Taking into consideration that the sequential one-electron oxidations occur across the POV-alkoxide framework, the oxidation state distributions of the vanadium ions in the remaining {FeNO}POV-alkoxide clusters can be assigned 12 {FeNO}] + (5-V 5 FeNO + ) (Scheme 2). Thus, following exposure to NO, the POV-alkoxide core is oxidized by one electron, with respect to the parent cluster (Scheme 2).

Electronic structure of the {FeNO} subunit
Following identication of the oxidation states of the vanadium ions within the POV-alkoxide assembly, charge balance reveals, in all cases, an overall 2 + charge on the {FeNO} subunit for complexes 2-5. The required charge is satised by multiple resonance structures, namely [Fe II NOc] 2+ and [Fe III NO À ] 2+ . Due to the ambiguity in the explicit oxidation state of the iron centre, the electronic structures of the iron-nitrosyl subunits of the FePOV-alkoxide complexes are best described as {FeNO} 7 in the Enemark-Feltham notation. 53 The v(NO) observed in the IR spectra of clusters 2-5 are consistent with this assignment, with values that resemble previously reported 6-coordinate {FeNO} 7 complexes (Table 1). 11 To more rigorously dene the electronic structure of the iron-nitrosyl moiety, zero-eld Mössbauer spectroscopy was performed on solid samples of complexes 2-4 (collected at 80 K) (Fig. 5). The broad, slightly asymmetric, quadrupole doublets resemble values reported previously for monodisperse FePOValkoxide clusters, indicating a single electronic environment for the iron centre. 25 It is worth noting that the isomer shis obtained from the Mössbauer spectra of the {FeNO}POV-alkoxide clusters resemble previously reported parameters for {FeNO} 7 complexes ( Table 1). The observed decrease in quadrupole splitting of the {FeNO}POV-alkoxide species as compared to their monometallic, 6-coordinate, {FeNO} 7 congeners is likely reective of the unique structure of the multimetallic cluster complexes.
Electron paramagnetic resonance (EPR) spectroscopy was also performed on complex 3-V 5 FeNO À in order to obtain further insight into the spin state of the {FeNO} 7 unit of these   Table 1. complexes. In addition to a broad S ¼ 1/2 signal attributed to the POV-alkoxide scaffold, 54 signicant signal intensity in the g $ 4 region is present, consistent with the presence of a highspin S ¼ 3/2 {FeNO} 7 unit (Fig. S6 †).
The formation of the {FeNO} 7 subunit upon exposure of the FePOV-alkoxide clusters to NO(g) is remarkable, given addition of a nitrosyl ligand to a ferric centre would be expected to result in the formation of an {FeNO} 6 subunit. Generation of the {FeNO} 7 moiety is consistent with either a metal-centred or substrate-based reduction upon coordination of NO to the multimetallic species. The reduction of the iron nitrosyl moiety occurs simultaneously with the observed oxidation of the POValkoxide scaffold, thus we can conclude that the metal-oxide metalloligand transfers electron density to the iron centre to facilitate activation of the substrate. This observed electron transfer between metal-oxide scaffold and iron constitutes a rare example of cooperative multimetallic reactivity for substrate activation. Similar exibility in site-differentiated multimetallic clusters has recently been reported by Agapie and coworkers. 55 This report demonstrates the utility of electron transfer from a mixed-valent iron oxide base to a sitedifferentiated ferric ion upon coordination of CO, resulting in the formation of [LFe 3 O(PhIm)Fe(CO) n ] (L ¼ 1,3,5-triarylbenzene ligand motif; PhIm ¼ 1-phenyl imidazole) complexes.
In both examples, electron transfer claries that the metalloligand plays an important part in substrate activation, providing electron density to a ferric centre to facilitate substrate coordination and reduction.

Attempts to access the fully reduced {FeNO}POV-alkoxide
Our attention turned toward identication of the fully reduced species, generated by the one-electron reduction of complex 4-V 5 FeNO 2À . Interest in characterizing the product of this reaction stems from the surprising peak-to-peak separation associated with the redox event centred at E 1/2 ¼ À2.68 V vs. Fc/Fc + (DE 1/2 ¼ 1.52 V). Similar separations of redox events in heterometal-functionalized polyoxometalates have been attributed to changes in oxidation state of the distinct metal centre incorporated within the cluster. 58,59 Consistent with these observations, the reducing nature of this electrochemical process suggests the reduction event is localized to the sitedifferentiated, iron-nitrosyl moiety. Accordingly, we hypothesized that reduction of the {FeNO} 7 subunit would give rise to the formation of a rare, non-heme {FeNO} 8 species. [36][37][38][60][61][62][63] Reduction of 2-V 5 FeNO in THF was attempted at lowtemperatures with a freshly prepared solution of sodium naphthalenide. Stirring the solution for 1 hour results in apparent formation of a single product, as observed by 1 H NMR spectroscopy (Fig. S7, † d ¼ 90. 33, 34.48, 16.03 ppm in CD 3 CN). This three-resonance pattern in the product is similar to complexes 2-5, suggesting clean formation of the desired fullyreduced {FeNO}POV-alkoxide (Scheme 3). However, multiple attempts to characterize the reduced cluster via Mössbauer spectroscopy resulted in identication of three distinct electronic environments for iron within the sample (Fig. S9, Table S1 †). Despite the chemical reversibility observed for the most-reducing event in the CV of 2-V 5 FeNO, and the seemingly straightforward 1 H NMR spectrum of the product, the complicated Mössbauer spectrum suggests that substantial decomposition occurs upon attempts to chemically generate [V 5 O 6 (OCH 3 ) 12 {FeNO}] 3À (6-V 5 FeNO 3À ), to paramagnetic, 1 H NMR-silent byproducts.
Given the apparent thermal instability of the fully reduced cluster, in situ electroanalytical techniques were employed for characterization of the {FeNO}POV-alkoxide clusters (Fig. 6, Table 2). Infrared-spectroelectrochemistry (IR-SEC) allows IR active modes in a compound of interest to be monitored with respect to changes in electrochemical potential as a function of time. 64,65 To correlate the electrochemical response directly to the independently synthesized compounds, we started with the electrochemical generation of complexes 2-5. Indeed, a 3 mM solution of 2-V 5 FeNO in 0.4 M TBAPF 6 /THF supporting electrolyte at À0.27 V vs. Fc/Fc + shows an IR absorbance at 1735 cm À1 . This value compares well with the IR data from the independently synthesized sample (1732 cm À1 , Fig. 3, Table 1). When the applied cell potential was brought to 0.08 V vs. Fc/Fc + , an absorption band at 1785 cm À1 appears and grows in intensity with the concomitant loss of the band at 1735 cm À1 , Scheme 3 Attempted synthesis of complex 6-V 5 FeNO 3À via the triple reduction of 2-V 5 FeNO. consistent with the electrochemical generation of the oneelectron oxidized species 5-V 5 FeNO 1+ (v(NO) (ATR) ¼ 1780 cm À1 ). Upon returning to À0.27 V, the n(NO) of the starting material 2-V 5 FeNO at 1735 cm À1 was regenerated with the loss of the band at 1785 cm À1 (Fig. 6), indicating good reversibility of this electrochemical process. At more negative potentials, À0.68 V vs. Fc/Fc + , an absorption at 1699 cm À1 appears with the disappearance of the band at 1735 cm À1 , consistent with the electrochemical generation of the one-electron reduced species 3-V 5 FeNO 1À (v(NO) (ATR) ¼ 1709 cm À1 ). At À2.0 V vs. Fc/Fc + , a new IR absorbance band at 1664 cm À1 is observed to grow in intensity with the loss of the band at 1699 cm À1 , indicating formation of 4- To probe the formation of the fully reduced cluster, [V 5 O 6 (-OCH 3 ) 12 FeNO] 3À (6-V 5 FeNO 3À ), via IR-SEC, the cell potential was held at À2.8 V vs. Fc/Fc + . An IR absorption band appears at 1640 cm À1 and increases in intensity with loss of the band at 1664 cm À1 . Aer holding a potential of À2.8 V for 25 min, the band noted at 1640 cm À1 corresponding to tri-reduced complex, 6-V 5 FeNO 3À , was nearly consumed, with concomitant formation of a new complex with a v(NO) located at 1604 cm À1 . This new species is indicative of a subsequent chemical reorganization of the {FeNO}POV-alkoxide cluster, as attempts to return the potential toward oxidizing regimes did not result in regeneration of 2-V 5 FeNO.
The instability of 6-V 5 FeNO 3À , as conrmed by IR-SEC, is consistent with our experimental data. In the chemical reduction of 2-V 5 FeNO with excess sodium naphthalenide, the reaction mixture is stirred for one hour, well beyond the lifetime of the reduced cluster as suggested by IR SEC ($25 min). Attempts to chemically re-oxidize the product mixture isolated from chemical reduction were unsuccessful, indicating that over the course of an hour, complex 6-V 5 FeNO 3À disproportionates to new products with vastly different electronic properties.
The small change in v(NO) ($25 cm À1 ) observed in the IR-SEC data of the transiently generated tri-reduced cluster (6-V 5 FeNO 3À ) suggests that despite the large DE 1/2 of the most reducing event, the nal reduction likely is localized to the POV-alkoxide metalloligand, as opposed to the {FeNO} subunit. An additional reduction across the POV-alkoxide scaffold requires formation of an FePOV-alkoxide cluster with a single V(III) ion. Accessing a V(III) ion within an iron-functionalized POV-alkoxide scaffold has been reported previously by our laboratory, 25 however, the signicant peak-to-peak separation observed in the CV of complex 2-V 5 FeNO was not observed in the case of the parent FePOV-alkoxide cluster, 1-V 5 Fe. In the case of the nitrosyl functionalized cluster, accessing the V(III) ion within the POV-alkoxide scaffold only occurs at extremely reducing potentials, and immediately affords disproportionation. The seemingly straight-forward 1 H NMR spectrum of the product of decomposition and the reduced v(NO) observed by IR-SEC suggests that degradation of 6-V 5 FeNO 3À results in formation of an FePOV-alkoxide cluster with a chemically activated nitrosyl unit. Current investigations into the molecular composition of this highly reactive by-product are underway.

Conclusions
In summary, a series of NO-bound FePOV-alkoxide clusters, 5-V 5 FeNO + , 2-V 5 FeNO, 3-V 5 FeNO À , and 4-V 5 FeNO 2À , varying in oxidation state by a single electron have been synthesized and systematically characterized using a variety of spectroscopic techniques. The charge distributions of complexes 2-5 were determined, revealing electron transfer from the POV-alkoxide scaffold to the iron-nitrosyl moiety upon coordination of NO. These results demonstrate that the metal-oxide metalloligand is capable of providing electron density to the iron centre for substrate coordination and activation.
The formation of {FeNO} 7 motifs in FePOV-alkoxide clusters illustrates our initial foray into utilizing metal-oxide scaffolds as redox reservoirs for rst-row transition metal complexes. The diffuse electron density of the POV-alkoxide scaffold, coupled with a suitable vacant site on the heterometal, enables coordination and the reduction of NO at a ferric centre. Future studies will expand upon this seminal work-extending it to other small molecule substrates and uncovering new approaches to mediating multielectron transformations using earth abundant elements.

General considerations
All manipulations were carried out in the absence of water and oxygen in a UniLab MBraun inert atmosphere glovebox under a dinitrogen atmosphere. Glassware was oven dried for a minimum of 4 h and cooled in an evacuated antechamber prior to use. Unless otherwise noted, solvents were dried and deoxygenated on a Glass Contour System (Pure Process Technology, LLC) and stored over activated 3Å molecular sieves (Fisher Scientic). Celite 545 (J. T. Baker) was dried in a Schlenk ask for at least 14 h at 150 C under vacuum prior to use. Sodium (Na, 99%+), Silver perchlorate (AgClO 4 , anhydrous, 97%) and bis(cyclopentadienyl)cobalt(II) (CoCp 2 , 98%) were purchased from Sigma-Aldrich and used as received. Nitric Oxide (NO, Matheson Gas Products, Inc. 99%) was puried according to literature precedent. 66 [V 5 O 6 (OCH 3 ) 12 Fe] + , We were unable to isolate complex 6-V 5 FeNO 3À chemically, therefore the IR spectrum of this compound was not obtained. b v(NO) absorbance corresponds to an irreversible chemical transformation that occurs following formation of 6-V 5 FeNO 3À .
[V 5 O 6 (OCH 3 ) 12 Fe] (1-V 5 Fe), [V 5 O 6 (OCH 3 ) 12 Fe] À , and [V 5 O 6 (-OCH 3 ) 12 Fe] 2À were synthesized according to our previous report. 24,25 1 H NMR spectra were recorded on a Bruker DPX-500 MHz spectrometer locked on the signal of deuterated solvents. All chemical shis were reported relative to the peak of a residual 1 H signal in deuterated solvents. CD 3 CN and CDCl 3 were purchased from Cambridge Isotope Laboratories, degassed by three freeze-pump-thaw cycles, and stored over activated 3Å molecular sieves. Infrared (FT-IR, ATR) spectra of complexes were recorded on a Shimadzu IR Affinity-1 Fourier Transform infrared spectrophotometer and are reported in wavenumbers (cm À1 ). Electronic absorption measurements were recorded at room temperature in anhydrous acetonitrile in a sealed 1 cm quartz cuvette with an Agilent Cary 60 UV-Vis spectrophotometer. Elemental analyses were performed on a PerkinElmer 2400 Series II Analyzer at the CENTC Elemental Analysis Facility (University of Rochester).
Cyclic Voltammetry experiments were recorded with a CH Instruments Inc. 410c time-resolved electrochemical quartz crystal microbalance. All measurements were performed in a three-electrode system cell conguration that consisted of a glassy-carbon (ø ¼ 3.0 mm) working electrode, a Pt wire counter electrode, and an Ag/AgCl wire reference electrode. All electrochemical measurements were performed at room temperature in a N 2 -lled glovebox. A 0.4 M n Bu 4 NPF 6 solution (anhydrous THF) was used as the electrolyte solution. All redox events were referenced against the ferrocene/ferrocenium (Fc/ Fc + ) redox couple.
All samples for 57 Fe Mössbauer spectroscopy were run as isolated solid samples made from natural abundance iron. All samples were prepared in an inert-atmosphere glovebox equipped with a liquid-nitrogen ll port. This enables freezing of the samples to 77 K within the glovebox. Samples were loaded into a Delrin Mössbauer cup for measurements and loaded under liquid nitrogen. 57 Fe Mössbauer measurements were performed using a SEE Co. MS4 Mössbauer spectrometer integrated with a Janis SVT-400T He/N 2 cryostat for measurements at 80 K. Isomer shis were determined relative to a-iron at 298 K. All Mössbauer spectra were t using the program WMoss (SEE Co.).
Samples for Electron Paramagnetic Resonance analysis were prepared in an inert atmosphere glovebox (N 2 ). EPR samples were prepared as solutions in acetonitrile and loaded into 4 mm OD Suprasil quartz EPR tubes from Wilmad Labglass. X-band EPR spectra were collected at 10 K on a Bruker EMXplus spectrometer equipped with a 4119HS cavity and an Oxford ESR-900 helium ow cryostat. Instrumental parameters employed for all samples were as follows: 1 mW power; 80 ms time constant; 8 G modulation amplitude; 9.38 GHz frequency; and 100 kHz modulation frequency.
IR Spectroelectrochemistry (IR-SEC) experiments were conducted using a custom cell based on a previously published design. [67][68][69] The three-electrode set-up consists of an inner glassy carbon working electrode disc (10 mm diameter), a central circular silver bare metal pseudoreference electrode, and an outer circular glassy carbon counter electrode embedded within a PEEK block. All data were referenced to an internal ferrocene standard (ferricenium/ferrocene reduction potential under stated conditions; obtained by taking a CV with the cell prior to injecting analyte for IR-SEC experiments) unless otherwise specied. IR-SEC solutions were prepared inside a glovebox under N 2 atmosphere and injected in IR-SEC cell using an SGE airtight syringe. IR spectra were collected using a Bruker Vertex 80 spectrometer with a liquid nitrogen-cooled detector.

Synthesis of {(VO) 5 O(OCH 3 ) 12 FeNO} (2-V 5 FeNO)
In the glovebox, a 250 mL Schlenk ask was charged with 1-V 5 Fe (0.119 g, 0.152 mmol) and 40 mL CH 2 Cl 2 . The nitrogen atmosphere was removed with two freeze-pump-thaw cycles, and one atmosphere of puried NO gas (normal temperature and pressure) was added via to the evacuated ask. The solution was vigorously stirred for three days, resulting in a colour change from dark-green to brown. Volatiles were removed under reduced pressure, and the ask was taken back into the glovebox. The product was extracted with diethyl ether (Et 2 O, 10 mL Â 3), followed by extraction with n-pentane (10 mL Â 3). Removal of solvent under reduced pressure results in isolation of the product, 2-V 5 FeNO (0.074 g, 0.091 mmol, 60%). 1  Method B. In the glovebox, a 250 mL round bottom Schlenk ask was charged with K(VO) 5 O(OCH 3 ) 12 Fe (0.130 g, 0.159 mmol) and 40 mL CH 2 Cl 2 /THF (1 : 1) mixture. The ask was capped with a septum, carefully taped, and removed from the glovebox. Puried NO gas (ca. 74 mL at normal temperature and pressure conditions) was added via gas-tight syringe. The solution was vigorously stirred for four days and afforded a colour change from dark-green to brown. The solvent was removed under reduced pressure, and the ask was taken back into the glovebox. Copious amount of CH 2 Cl 2 was added to extract the solid residue until the extraction turned nearly colourless, followed by removal of the volatiles to yield 3-V 5 FeNO À as a brown powder (0.050 g, 0.059 mmol, 37%). 1 H NMR and FT-IR spectra collected for the product are identical to that observed for method A.
Synthesis of (CoCp 2 ) 2 {(VO) 5 O(OCH 3 ) 12 FeNO} (4-V 5 FeNO 2À ) A 20 mL scintillation vial was charged with 2-V 5 FeNO (0.124 g, 0.153 mmol) and 16 mL THF. In a separate vial, cobaltocene (CoCp 2 , 59 mg, 0.311 mmol) was dissolved in 4 mL THF and added dropwise to the cluster solution while stirring. A dark grey-brown precipitate started to form within 5 minutes. The mixture was stirred vigorously for an additional 2 h. The precipitate was collected and washed with THF until the ltrate was nearly colourless. The grey-brown solid was dried under vacuum to yield complex 4-V 5 FeNO 2À in good yield (0.148 g, 0.125 mmol, 81%). 1  Method A. A 20 mL scintillation vial was charged with 2-V 5 FeNO (0.042 g, 0.052 mmol) and 6 mL THF. The vial was placed in a cold well cooled to À40 C. In a separate vial, silver perchlorate, (AgClO 4 , 0.011 g, 0.053 mmol) was dissolved in 2 mL THF, placed in the cold well for 10 minutes, and added dropwise to the cold cluster solution with vigorous stirring. The mixture was warmed to room temperature, during which time a dark-grey precipitate started to form. The reaction was subsequently stirred for 2 h to ensure completion. The solution was ltered over celite (1 cm), and the solvent was removed under reduced pressure, followed by extraction of the solid residue with diethyl ether. Volatiles were removed under vacuum to yield 5-V 5 FeNO + as a brown powder in good yield (0.037 g, 0.041 mmol, 78%). 1  . Elemental analysis of complex 5-V 5 FeNO + was not obtained because the compound is not sufficiently stable at room temperature.
Method B. In the glovebox, a 250 mL round bottom Schlenk ask was charged with (VO) 5 O(OCH 3 ) 12 FeClO 4 (0.132 g, 0.150 mmol) and 40 mL CH 2 Cl 2 . The ask was capped with a septum, carefully taped, and removed from the glovebox. Puried NO gas (ca. 74 mL at normal temperature and pressure conditions) was added via gas-tight syringe. The solution was vigorously stirred overnight and afforded a colour change from dark green to dark brown. The solvent was removed under reduced pressure, and the ask was taken back into the glovebox. Toluene (10 mL Â 3) was added to extract the solid residue, followed by removal of the volatiles to yield 5-V 5 FeNOClO 4 as a brown powder (0.088 g, 0.109 mmol, 71%). 1 H NMR and FT-IR spectra collected for the product are identical to that observed for method A.
Chemical reduction of 2-V 5 FeNO with Na(Nap) A 20 mL Scintillation vial was charged with 0.028 g naphthalene (0.218 mmol) and 6 mL THF, to which pieces of freshly cut sodium metal (0.073 g, 3.167 mmol) were added. The mixture was stirred at room temperature for 3 hours, affording a dark green solution of sodium naphthanelide. This solution was decanted into a fresh 20 mL Scintillation vial and frozen in an 80 K cold well. In a separate vial, 2-V 5 FeNO (0.054 g, 0.067 mmol) was dissolved in 2 mL of THF, and the solution was frozen at 80 K. The 2-V 5 FeNO solution was allowed to thaw and was added to the frozen sodium naphthanelide solution along with an additional 2 mL thawed THF to complete the transfer. The reaction was stirred vigorously and allowed to warm to room temperature for 1 hour. The resulting brown solution was ltered over a bed of celite (0.5 cm), the volatiles were removed, and the brown solid was washed with toluene (5 mL Â 2).

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