Increasing reactivity by incorporating π-acceptor ligands into coordinatively unsaturated thiolate-ligated iron(II) complexes

Abstract

Reported herein is the structural, spectroscopic, redox, and reactivity properties of a series of iron complexes containing both a π-donating thiolate, and π-accepting N-heterocycles in the coordination sphere, in which we systematically vary the substituents on the N-heterocycle, the size of the N-heterocycle, and the linker between the imine nitrogen and tertiary amine nitrogen. In contrast to our primary amine/thiolate-ligated Fe(II) complex, [Fe^II(S^Me2N₄(tren))]⁺ (1), the Fe(II) complexes reported herein are intensely colored, allowing us to visually monitor reactivity. Ferrous complexes with R = H substituents in the 6-position of the pyridines, [Fe^II(S^Me2N₄(6-H-DPPN)]⁺ (6) and [Fe^II(S^Me2N₄(6-H-DPEN))(MeOH)]⁺ (8-MeOH) are shown to readily bind neutral ligands, and all of the Fe(II) complexes are shown to bind anionic ligands regardless of steric congestion. This reactivity is in contrast to 1 and is attributed to an increased metal ion Lewis acidity assessed via aniodic redox potentials, E_p,a, caused by the π-acid ligands. Thermodynamic parameters (ΔH, ΔS) for neutral ligand binding were obtained from T-dependent equilibrium constants. All but the most sterically congested complex, [Fe^II(S^Me2N₄(6-Me-DPPN)]⁺ (5), react with O₂. In contrast to our Mn(II)-analogues, dioxygen intermediates are not observed. Rates of formation of the final mono oxo-bridged products were assessed via kinetics and shown to be inversely dependent on redox potentials, E_p,a, consistent with a mechanism involving electron transfer.

1. Introduction

Thiolate ligands have been shown to lower the activation barrier to dioxygen binding, increase the ability of $\text{Fe}-{\text{O}}_2^{•-}$ superoxo compounds to abstract H-atoms from strong C–H bonds [1], promote peroxo O–O bond cleavage [2], and create potent high-valent metal-oxos capable of activating strong C–H bonds [1,3–5]. Iron thiolate (Fe–SR) bonds are highly covalent, stabilize low spin-states due to the nephelauxetic effect [6], and are capable of compensating for changes in electron density at the metal ion thereby maintaining a relatively constant redox potential [7]. This compensatory effect reflects not only the highly covalent Fe-S bonds, but also the redox non-innocence of thiolate ligands, which facilitates delocalization of oxidizing equivalents onto the sulfur [8]. This unique property is important to the function of both electron-transfer proteins, such as blue copper proteins [9,10], as well as non-heme iron enzymes [11], such as isopenicillin N-synthase (IPNS) [3,12,13], superoxide reductase (SOR) [14–19], cysteine dioxygenases (CDO) [20–26], and nitrile hydratases (NHases) [27–31], as well as the heme enzyme cytochrome P450 [32–34].

The primary coordination sphere of a transition-metal complex can have a profound influence on the electronic and geometric structure, as well as its magnetic, redox, and reactivity properties. Even subtle changes to the coordination sphere, involving the replacement of a single hydrogen-bond [35–38], or protonation of a coordinated ligand [7,39], can dramatically alter these properties. For example, if the imidazole of haemoglobin [38,39] is replaced with a cysteinate [5,40], then O₂ activation and reduction, and peroxo O–O bond cleavage, as opposed to O₂ transport, is favored, and the resulting iron-oxo becomes a potent oxidant capable of cleaving strong C–H bonds [5,40]. Our work has shown that thiolate ligands make low–spin iron accessible even in a non–heme environment [30,41,42], stabilize iron in the +3 oxidation state [30,41,42], and labilize sites trans to the thiolate thereby promoting reactivity by releasing product [15], even with low–spin Co(III) which is typically inert [43,44].

Previously our group showed that thiolate-ligated [Fe^II(S^Me2N₄(tren))]⁺ (1, Scheme 1) reacts with superoxide in a proton-dependent mechanism to afford a well-characterized metastable hydroperoxo complex, [Fe^III(S^Me2N₄(tren))(OOH)]⁺ (2), which then releases H₂O₂ upon protonation at the proximal oxygen [45], thus mimicking the function of SOR [46,47]. Two potential mechanisms for both the formation of 2, as well as the SOR ferric hydroperoxo intermediate, would involve the sequential addition of ${\text{O}}_2^{•-}$ and a proton, (a) + (b) to afford an $\text{Fe}(\text{II})—{\text{O}}_2^{-•}$, followed by Fe(III)–_OOH intermediate (Fig. 1), or the concerted addition of HO₂ to the metal ion (pathway (c) of Fig. 1). Primary amine-ligated 1, the precursor to hydroperoxo 2, is colorless and only displays bands in the UV, due to the high energy of the S → Fe(II) CT band [45,48]. This makes it difficult to detect ferrous Fe (II)-L (e.g., $\text{L}={\text{O}}_2^{-}$) intermediates. By incorporating π-acceptor N-heterocycles (N^Ar) into the ligand scaffold (Figs. 2 and 3), it should be easier to monitor reactivity via electronic absorption spectroscopy using the M → π-acceptor chromophore. We recently showed that steric encumbering groups (Me, OMe, or benzo) on the N^Ar ligand influence the kinetic barrier to O₂ binding to Mn(II), peroxo O–O bond cleavage, and hydrogen atom transfer (HAT) reactivity of the corresponding high-valent Mn-oxo intermediates [49,50]. We have also shown that the combination of redox active π-accepting N-heterocycle ligands and π-donor thiolate ligands can have a profound effect on the electronic structure. Sulfur K-edge X-ray absorption spectroscopic (XAS) data coupled with time-dependent density functional theory calculations (TD DFT) provided evidence for direct communication between the thiolate sulfur of ${\left[{\text{Fe}}^{\text{II}}\left(\text{tame}-{\text{N}}_2\text{S}{(\text{py})}_2\right)\right]}_2^{2+}$ and the roughly orthogonal (α-imino)-N-heterocycle ligand π*-orbitals [8]. Extensive delocalization of bonding electrons over the sulfur, iron, and (α-imino)-N-heterocycle was observed in the frontier orbitals, and the α-imino-pyridine C=N and C^ipso-C bonds were shown to elongate close to the length of a reduced α-imino-pyridine moiety, even in the absence of a strong reducing agent [8].

Scheme 1.

ChemDrawings of the Fe(II) complexes discussed herein. The 6th ligand of six-coordinate complexes is highlighted in blue.

Fig. 1.

Sequential, (a) + (b), versus concerted, (c), addition of HO₂ to Fe(II) to afford an Fe(III)–OOH.

Fig. 2.

ORTEP diagrams of coordinatively unsaturated thiolate-ligated 3 and 4. Hydrogen atoms, counterions, and cosolvents have been removed, and chemdrawings included for clarity.

Fig. 3.

ORTEP diagrams of coordinatively unsaturated thiolate-ligated 5 and 7. Hydrogen atoms, counterions, and cosolvents have been removed, and chemdrawings included for clarity.

Herein we report the synthesis, structure, redox properties, and reactivity of a new series of structurally related 5-coordinate thiolate-ligated Fe(II) complexes (Scheme 1) that incorporate both a π-donating thiolate ligand and π- accepting pyridine or quinoline ligands (Figs. 2 and 3). We show that by incorporating π-accepting N-heterocycles (N^Ar = 6-H-pyridine, 6-Me-pyridine, or quinoline) in place of the primary amines of [Fe^II(S^Me2N₄(tren))]⁺ (1) that (a) the ferrous compounds are intensely colored making it is easier to monitor reactions via electronic absorption spectroscopy, (b) the binding affinity for a sixth ligand increases, and (c) metal ion Lewis acidity increases [51].

2. Results and discussion

2.1. Syntheses and structure of coordinatively unsaturated Fe(II) complexes containing both a π-accepting N-heterocycle and a π-donating thiolate

The N-heterocycle ligands used in this study (Fig. 4) were synthesized as previously described for the corresponding Mn-derivatives [52]. The ligand scaffold was varied to include (a) pyridine or quinoline N-heterocycles, (b) H or Me substituents on the pyridine rings, or (c) ethyl versus propyl linkers between the tertiary amine and imine nitrogens (Fig. 4, highlighted in red) [52]. A metal templated Schiff-base reaction was used to assemble the corresponding Fe(II) complexes (Scheme 1, Figs. 2, 3, 5, and 10). Single crystals of each complex were obtained as described in the supplemental material. Crystallographic tables are provided in the supplemental material (Tables S5–S39). Selected metrical parameters are assembled in Table 1. As shown in the ORTEP diagrams of Figs. 2,3, 5, and 10, all of these complexes, with the exception of 6-MeOH (Fig. 5) and 8-MeCN (Fig. 10), were found to contain Fe in a five-coordinate, and thus coordinatively unsaturated environment, despite crystallization from coordinating solvents. With the exception of 6 (Fig. 5), those containing a Pr-linker (Fig. 3) were found to be approximately square pyramidal (τ-value = 0.01(5), 0.04(7), Table 1, highlighted in peach), and those containing an Et-linker were found to be approximately trigonal bipyramidal (τ-value = 0.60(3), 0.59(4); Table 1). Propyl-linked ([Fe^II(S-^Me2N₄(6-H-DPPN)]⁺ (6, Fig. 5), which contains less sterically encumbering 6-H substituents, is closer to being ideal trigonal bipyramidal (τ-value = 0.81) than any of the other structures (Table 1). The Fe-S bonds of the complexes containing a propyl linker (5 and 7) are shorter than those containing an ethyl linker (3 and 4). These trends make sense given that the more orthogonal L-Fe-L’ angles of a square pyramidal geometry provides better overlap with the metal ion d-orbitals. Square pyramidal [Fe^II(S^Me2N₄(6-Me-DPPN))]⁺ (5) possesses the shortest Fe-S bond (2.3005(9) Å, Table 1), and trigonal bipyramidal [Fe^II(S^Me2N₄(quinoEN))]⁺ (3) and [Fe^II(S^Me2N₄(6-Me-DPEN)]⁺ (4), which contain the more sterically encumbering 6-Me or benzo substituents, possess the longest Fe-S bonds (2.3357(9) Å and 2.326(1) Å, respectively). In addition to steric interactions, poorer overlap between the sulfur and the metal ion d-orbitals in a trigonal bipyramidal geometry, are likely responsible for the longer Fe-S bonds in the latter. Metal-nitrogen bonds were similarly elongated in the trigonal bipyramidal complexes 3 and 4. The average Fe-N bond length (Fe-N_avg = 2.12 Å) in primary amine-ligated [Fe^II(S^Me2N₄(tren))]⁺ (1) is shorter than that of N^Ar-ligated 3–6. This reflects the unfavorable steric interactions in the latter. Metal-nitrogen bonds were similarly elongated in the structurally analogous Mn(II) derivatives containing sterically encumbered N-heterocycles [52].

Fig. 4.

Ligands incorporated into the molecules discussed herein. Highlighted in red are the portions of the ligand scaffold that were varied.

Fig. 5.

ORTEP diagrams of coordinatively unsaturated thiolate-ligated 6 and its MeOH bound derivative 6-MeOH. Hydrogen atoms, counterions, and cosolvents have been removed, and chemdrawings included for clarity.

Fig. 10.

ORTEP diagram of thiolate-ligated 8. Hydrogen atoms, counterions, and cosolvents have been removed, and a chemdrawing was included for clarity.

Table 1.

Selected Bond Distances (Å) and Bond Angles (deg) for Five Coordinate [Fe^II(S^Me2N₄(quinoEN))](PF₆) Et₂O (3), [Fe^II(S^Me2N₄(6-Me-DPEN)](PF₆) (4), [Fe^II(S^Me2N₄(6-Me-DPPN)](PF₆) (5), [Fe^II(S^Me2N₄(6-H-DPPN))](PF₆) (6), and [Fe^II(S^Me2N₄(quinoPN))](BPh₄)⋅MeCN (7).

	3	4	5	6	7
Fe-S(1)	2.3357(9)	2.326(1)	2.3005(9)a	2.319(1)	2.3057(5)
Fe-N(1)	2.111(2)	2.132(4)	2.155(3)	2.077(3)	2.148(1)
Fe-N(2)	2.226(2)	2.253(4)	2.206(2)	2.299(3)	2.169(1)
Fe-N(3)	2.152(2)	2.165(4)	2.210(3)	2.092(3)	2.226(1)
Fe-N(4)	2.159(2)	2.191(4)	2.133(2)	2.094(3)	2.147 (1)
S(1)-Fe-N(1)	82.91(7)	81.5(1)	83.78(7)	84.00(8)	83.69(3)
S(1)-Fe-N(2)	157.80(6)	158.7(1)	149.70(7)	174.86(8)	148.52(3)
S(1)-Fe-N(3)	106.98(7)	116.1(1)	94.04(7)	106.18(9)	94.28(3)
S(1)-Fe-N(4)	121.74(7)	115.0(1)	129.45(7)	105.73(8)	131.85(3)
N(1)-Fe-N(3)	120.28(9)	123.1(1)	150.42(9)	114.3(1)	154.97(4)
N(1)-Fe-N(4)	119.21(9)	121.0(1)	105.3(1)	126.2(1)	100.55(4)
N(3)-Fe-N(4)	104.91(9)	100.5(1)	98.52	113.1(1)	99.29(4)
τ-value	0.60	0.59	0.01	0.81	0.11

2.2. π-Accepting N^Ar ligands shift the electronic absorption bands into the visible region

Solutions of N^Ar/thiolate complexes 3 – 6 are all intensely colored and contain intense charge transfer bands in the visible region (λ_max = 387–502 nm) of the electronic absorption spectra (Figs. 6 and 7, and Figs. S4–S9), making it relatively easy to monitor reactivity (vide infra). This is in contrast to the primary amine complex, [Fe^II(S^Me2N₄(tren))]⁺ (1), which is colorless, and only contains absorption bands in the UV region of the spectrum (λ_max = 290 nm) [48]. The energy of the lowest-energy band is sensitive to ring size, and red shifts when pyridine is replaced with a larger quinoline N-heterocycle, as was shown with previously reported π-donating/π-accepting thiolate/(α-imino)-N-heterocycle-ligated ${\left[{\text{Fe}}^{\text{II}}\left(\text{tame}-{\text{N}}_2\text{S}{(\text{Me}-\text{Im})}_2\right)\right]}_2^{2+}$, ${\left[{\text{Fe}}^{\text{II}}\left(\text{tame}-{\text{N}}_2\text{S}{(\text{py})}_2\right)\right]}_2^{2+}$ versus ${\left[{\text{Fe}}^{\text{II}}\left(\text{tame}-{\text{N}}_2\text{S}{\left(\text{quino}\right)}_2\right)\right]}_2^{2+}$ [8]. Although the λ_max of N^Ar/thiolate-ligated 3 (Fig. S4), 4 (Fig. S5), and 5 (Fig. S6) lie in the UV region of the spectrum, the intense (ε ≥ 1300 cm⁻¹ M⁻¹) absorption bands are relatively broad and tail into the visible region to ~500 nm, resulting in an intense red or orange color.

Fig. 6.

Quantitative electronic absorption spectra of 3 (green) and 4 (blue) in THF, and 8 (red) in CH₂Cl₂ at 25 °C.

Fig. 7.

Solvent-dependent electronic absorption spectrum of 6 at 298 K.

2.3. Reversible solvent binding to Fe(II)-6

The electronic absorption spectra of the more sterically encumbered complexes 3–5, and 7 (Scheme 1), are independent of solvent (Figs. S4–S8, Table S1), indicating that coordinating solvents do not bind to the open coordination site. The spectrum of the least sterically encumbered member of the series, 6, on the other hand, is both solvent-dependent (Fig. 7) and temperature-dependent (Fig. 8, and Figs. S9 and S11, Tables S2–S4). This indicates that coordinating solvents (MeCN, MeOH, and THF) bind reversibly to the vacant site of 6. Consistent with this, five-coordinate [Fe^II(S^Me2N₄(6-H-DPPN))]⁺ (6, Fig. 5), could only be isolated from the non-coordinating solvent, CH₂Cl₂. Crystallization from MeOH affords a MeOH-bound derivative, [Fe^II(S^Me2N₄(6-H-DPPN)) (MeOH)]⁺ (6-MeOH, Fig. 5). The red-shift in λ_max (Figs. 7 and 8, Figs. S9, and S11) is consistent with the conversion of a 5- to 6-coordinate structure [6]. The Fe-O(1) bond length (Fig. 5, 2.210(2) Å), which is long relative to a typical methoxide Fe(II)-O bond (range: 1.864 Å–1.927 Å) [53,54], the H-bonding network in the crystal structure packing diagram, and the number of counterions, support the presence of a proton on O(1).

Fig. 8.

Temperature-dependent equilibrium between five-coordinate 6 (0.68 mM) and MeOH-bound 6-MeOH, formed via the addition of 243 equiv of MeOH to 6 in CH₂Cl₂.

Temperature-dependent equilibrium constants (Tables S2–S4) for solvent binding to 6 were obtained via the quantitative addition of MeCN (Fig. S11), MeOH (Fig. 8), or THF (Fig. S9) to five-coordinate 6 in CH₂Cl₂ over the temperature range 0 °C to −70 °C, and thermodynamic parameters for solvent binding (ΔH, ΔS) were obtained from a Van’t Hoff plot (Fig. 9, Table 2, Figs. S10 and S12). Of the three solvents investigated, 6 has the highest affinity for MeOH ${\text{K}}_{\text{eq}}^{213\text{K}}\left(\text{MeOH}\right)=26.7{\underset{\_}{\text{M}}}^{-1}\text{ versus }{\text{K}}_{\text{eq}}^{213\text{K}}\left(\text{MeCN}\right)=0.70{\underset{\_}{\text{M}}}^{-1}$, and ${\text{K}}_{\text{eq}}^{213\text{K}}\left(\text{THF}\right)=1.33{\underset{\_}{\text{M}}}^{-1}$), with an associated enthalpy (ΔH) that is approximately twice that of MeCN and THF (Table 2). The negative entropy parameters (ΔS) are consistent with an associative process.

Fig. 9.

Van’t Hoff plot for MeOH binding to 5-coordinate 6 in CH₂Cl₂.

Table 2.

Thermodynamic Parameters Associated with Solvent Binding to Five-Coordinate [Fe^II(S^Me2N₄(6-H-DPPN)](PF₆) (6) in CH₂Cl_2.

	MeOH	MeCN	THF
ΔH (kcal/mol)	−6.9(5)	−3.2(1)	−3.27(3)
ΔS (cal/mol⋅K)	−25(2)	−16(2)	−14.5(1)

Aside from N-heterocycle-ligated 6, [Fe^II(S^Me2N₄(6-H-DPEN)) (μ-S^Me2N₄(6-H-DPEN)Fe^II (MeCN)](PF₆)₂ (8) is the only other thiolate-ligated ferrous complex synthesized by our group that readily binds neutral ligands (i.e., solvents, Fig. S13). Complex 8 contains an Et-linker in place of the Pr-linker of 6, and an unencumbering 6-H substituent, and was crystallized from MeCN. As shown in the ORTEP diagram of Fig. 10, both Fe(II) ions are six-coordinate: Fe(2) contains a MeCN coordinated to the sixth coordination site, and Fe(1) contains a thiolate provided by the adjacent Fe(2) in the sixth coordination site. This implies that the Fe(II) ion of both 8 and 6 is either more accessible (a steric effect), or more Lewis acidic (an electronic effect) relative to the other complexes described herein (vide infra). Previously we showed that electron-rich π-donating thiolate ligands decrease metal ion Lewis acidity in primary amine-ligated Fe(II) complexes favoring coordinatively unsaturated structures [27,30,55]. The greater affinity that 6 and 8 have for additional ligands can therefore be attributed to the π-accepting properties of the pyridine partially off-setting the π-donor properties of the thiolate, as well as the 6-H substituents creating more space for incoming ligands. Comparison of the space-filling diagrams of 3, 4, 5, and 6 in Fig. 11 shows that metal ion accessibility not only depends on steric bulk (e.g., 3 and 4) but also the geometry of the complex. The metal ion is most visible, and thus most accessible, in trigonal bipyramidal 6, and least accessible in square pyramidal 5, and this influences reactivity with substrates such as O₂. Although 8 (Fig. 10) is dimeric in the solid state, magnetic susceptibility data (Fig. S14) indicates that the dimer cleaves to form monomers in coordinating solvents. In MeOH, the magnetic moment (μ_eff = 4.92 BM at 298 K), determined using thse Evan’s method, indicates that 8-MeOH is monomeric and high-spin, S = 2. Confirmation of this was obtained using high resolution linear trap quadropole (LTQ) Orbitrap mass spectroscopy, which shows isotopomers differing by single, as opposed to half mass, units (Fig. S15). The detected peak at 428.13 m/z (M + 31) matches the calculated mass for a six-coordinate methanol-bound complex, [Fe^II(S^Me2N₄(6-H-DPEN))(MeOH)]⁺ (8-MeOH). In MeCN, the high resolution LTQ ion trap mass spectrum of 8 shows isotopomers that differ both by half mass units, as well as single mass units (Fig. S16), indicating that an equilibrium mixture (Fig. S17) of dicationic dimer and monocationic monomers exists in this solvent. The ratio of ${\mu }_{\text{eff}}^2\left(\text{expt}\right)/{\mu }_{\text{eff}}^2\left(\text{theoretical}\right)$ provides the percentage of monomeric species 8-MeCN in solution at a given temperature (52.3% at 298 K, Fig. S17) [8].

Fig. 11.

Space-filling diagrams comparing metal ion accessibility of sterically encumbered trigonal bipyramidal 3 and 4, versus sterically encumbered square pyramidal 5, and unencumbered (6-H substituent) square pyramidal 6.

2.4. Anion binding

Although neutral solvent molecules were found only to bind to the least sterically hindered members of the series, 6 and 8, anionic ligands, such as OAc⁻ and ${\text{N}}_3^{-}$, on the other hand, were shown to bind even to some of the more sterically hindered members of the series. For example, Bu₄N(N₃) (Figs. S18 and S19) and acetate (Fig. 12 and Fig. S20) were shown to bind to 3 and 4. Azide was also shown to bind to 8 (Fig. S21). The red-shift in λ_max (from e.g., 474 nm to 554 nm for ${\text{N}}_3^{-}$ binding to 8) caused by anion binding would suggest that the bands involve a Fe(d) → π*(N-heterocycle) charge transfer transition, analogous to what we previously observed [8]. The ability to detect anion binding to Fe(II) contrasts with primary amine-ligated [Fe(II) (trenS)]⁺ (1) and bodes well for our goal of probing superoxide $\left({\text{O}}_2^{•-}\right)$ binding to Fe(II). Reactivity with the ${\text{O}}_2^{•-}$ will be the subject of a separate manuscript.

Fig. 12.

Monitoring the addition of Bu₄NOAc (in 0.2 equiv aliquots) to 3 in MeCN at 25 °C by electronic absorption spectroscopy.

2.5. Oxidation potentials as a probe of metal ion Lewis acidity

In order to promote superoxide reduction and/or oxidation, a transition-metal complex must be stable in multiple oxidation states. Redox potentials not only provide information about the relative stability of two oxidation states, but also provide a quantitative measure of metal ion Lewis acidity. As shown by the redox potentials in Table 3, and the cyclic voltammograms of Fig. 13 and Figs. S22–S26, π-accepting N-heterocycles increase metal ion Lewis acidity relative to the primary amines of [Fe^II(S^Me2N₄(tren))]⁺ (1, −150 mV vs SCE), as we had anticipated. This is shown by the more positive potential at which Fe(II) converts to Fe(III) (E_p,a) for complexes 3–7, and 8, relative to that of 1 (Table 3), which shifts by as much as 585 mV. Given that redox potentials reflect the relative stability of two different oxidations states, the higher potential for complexes 3–7 and 8 indicate that π-accepting N-heterocycles make the complexes are more difficult to oxidize, and cause the highest occupied molecular orbitals (HOMOs) to lie lower in energy for the N-heterocycle amine-ligated complexes. The potentials associated with 6-coordinate 6-MeOH and 8-MeCN shift less than the others (by as little as +191 mV), relative to 1 reflecting an increase in electron density at the metal caused by the sixth ligand. The lower potential for these complexes also reflects the increased stability of Fe(III) in its preferred six-coordinate environment. The chemical irreversibility of waves associated with the Fe(II)/Fe(III) couple of complexes 3–7 (Figs. S22–S24, and S26–S27) indicates that the five-coordinate structures convert to six-coordinate structures upon oxidation, as observed previously by our group [56]. The reversible wave associated with the 8-MeCN/8^ox-MeCN couple (Fig. 13), as well as the 6-MeOH/6^ox-MeOH (Fig. S25) reflects the fact that, unlike the other complexes, 8-MeCN and 6-MeOH are six-coordinate to begin with.

Table 3.

Comparison of Redox Potentials for Structurally Analogous Thiolate–Ligated Fe (II) Complexes.

	NAr	Linker	E1/2 (mV vs SCE)
[FeII(SMe2N4(tren))]+ (1)	N/A	Et	−150
[FeII(SMe2N4(6-H-DPEN))(MeCN)]+ (8-MeCN)	6-H-DPEN	Et	−90
[FeII(SMe2N4(6-H-DPPN)(MeOH]+ (6-MeOH)	6-H-DPPN	Pr	+66
[FeII(SMe2N4(2-QuinoEN))]+ (3)	QuinoEN	Et	+401
[FeII(SMe2N4(6-Me-DPEN)]+ (4)	6-Me-DPEN	Et	+408
[FeII(SMe2N4(2-QuinoPN))]+ (7)	QuinoPN	Pr	+415
[FeII(SMe2N4(6-Me-DPPN)]+ (5)	6-Me-DPPN	Pr	+433

Fig. 13.

Cyclic voltammogram of [Fe(S^Me2N₄(6-H-DPEN))(MeCN)]⁺ (8-MeCN) in MeCN with 0.1 M Bu₄NPF₆ supporting electrolyte, and a scan rate of 100 mVs⁻¹.

2.6. Outer sphere oxidation

Although too unstable to isolate, in situ formation of oxidized Fe(III) derivatives via the addition of either (p-tolyliminium)PF₆ or (Cp₂Fe)PF₆, depending on the redox potential, could be monitored by electronic absorption spectroscopy (Fig. 14, and Figs. S27–S29). As illustrated in Fig. 14 for 8-MeCN, the absorption bands associated with the Fe(II) derivative disappear upon oxidation and are replaced with new absorption bands in the ~600 nm region of the spectrum. The observed red-shift would be more consistent with a LMCT transition, than a MLCT transition, being responsible for the new absorption features, given that the metal ion orbitals would drop in energy. The rhombic ⊥-mode EPR signals at g = 2.180, 2.145, 2.008 (Fig. 15), and g = 2.25, 2.17, 1.98 (Fig. S30), for 8^ox-MeCN and 6^ox-MeOH, respectively, indicate that the oxidized ferric products are monomeric and low-spin (S = 1/2).

Fig. 14.

Oxidation of 6-H-DPEN 8-MeCN with (p-tolyliminium)PF₆ in 0.1 eq aliquots in MeCN at 25 °C.

Fig. 15.

⊥-mode, continuous-wave 9.30 GHz EPR spectrum of 8^ox-MeCN in MeCN/toluene glass at 77 K.

2.7. Dioxygen reactivity

The addition of dry O₂ to complexes 3, 4, 6, 7, and 8 at ambient temperatures in CH₂Cl₂ causes the orange/red solutions to become intensely purple in color. This color change is accompanied by the growth of a band in the electronic absorption spectrum in the range 494 – 587 nm (Fig. 16 and Figs. S31–S33). The λ_max is dependent on both the N-heterocycle and its substituent. These bands are red-shifted relative to that of primary amine-ligated [Fe^III(S^Me2N₄(tren))]₂(μ-oxo)⁺ (9) (λ_max = 485 nm), reflecting the greater Lewis acidity of the metal ions of 3, 4, 6, 7, and 8, due to the incorporation of π-accepting N-heterocycle ligands. The rate at which purple dioxygen products form is dependent on the redox potential of the Fe(II) precursor, and thus the metal ion Lewis acidity. For example, the rate at which the dioxygen product of 8-MeOH (E_1/2 = −90 mV,E_p,a = −60 mV vs SCE, Fig. 13) forms ($k_{\mathit{\text{obs}}}^{\mathit{200}K}=2.4\times {10}^{-2}{\text{s}}^{-1}$, t_1/2 = 29 sec, Fig. S34) is significantly faster than the rate at which the dioxygen product of both 3 (E_p,_a = +439 mV vs SCE, Fig. S22) and 4 (E_p,_a = +449 mV vs SCE, Fig. S23) forms ($k_{\mathit{\text{obs}}}^{\mathit{298}K}(3)=4.9\times {10}^{-3}{\text{ s}}^{-1}$, t_1/2(3) = 141 sec, Fig. 17); $k_{\mathit{\text{obs}}}^{\mathit{298}K}(4)=2.3\times {10}^{-3}{\text{ s}}^{-1}$, t_1/2(4) = 301 sec, Fig. S35), and the latter two are at a higher temperature. In contrast to all of the other Fe(II) complexes discussed herein, complex 5 (E_p,_a = +486 mV vs SCE, Fig. S24) does not react with O₂ at all, even with prolonged reaction times (48 hr). This likely reflects the more sterically congested, less accessible Fe site of 5 (Fig. 11). As shown in the ORTEP diagrams of Fig. 18, and Figs. S36–S38, the structure of the purple product, 3^μ-oxo, 4^μ-oxo, 6^μ-oxo, and 8^μ-oxo, is analogous to primary amine-ligated 9 [57], and contains a single unsupported oxo atom bridging between two ferric ions. There is a direct correlation between the λ_max of the absorption band associated with the mono oxo-bridged products, and the Fe⋯N^Ar distance (Fig. 19), and an inverse correlation between oxidation potential, E_p,a, and λ_max (Fig. 20). These trends can be explained based on the expected drop in the energy of metal ion d-orbitals expected with an increase in metal ion Lewis acidity. The less sterically encumbered complexes, 6^μ-oxo and 8^μ-oxo, with R = H substituents possess the shorter Fe-N^Ar bonds, and this increases electron density at the metal ion, thus decreasing metal ion Lewis acidity. If the absorption bands involve LMCT transitions, then for the reasons outlined above, one would expect the band to shift to lower energies, or higher wavelength, as metal ion Lewis acidity increases, as seen in Fig. 19. The absorption band, λ_max, of the least sterically hindered (R = H) oxo, 8^μ-oxo, with the shortest Fe-N^Ar bonds is closest in energy to that of primary amine-ligated 1. In contrast to the dioxygen chemistry of the corresponding Mn(II) derivatives, where as many as four metastable intermediates are observed [49,50,58], no intermediates are observed en route to our mono oxo-bridged Fe products, 3^μ-oxo, 4^μ-oxo, 6^μ-oxo, and 8^μ-oxo. This implies that with Fe(II), either O₂ binding is rate-determining, or the iron-dioxygen intermediates are less stable than our manganese-dioxygen intermediates. Both could be due to the fact that Fe(II) contains an additional destabilizing antibonding electron. Selected metrical parameters of mono oxo bridged 3^μ-oxo, 4^μ-oxo, 6^μ-oxo, and 8^μ-oxo are shown in Table 4. As shown previously with the mono oxo-bridged Mn(III)-analogues, Fe⋯N^Ar distances (Fe-N(3,4)_avg = 2.34 Å for 3^μ-oxo, and Fe-N(3,4)_avg = 2.41 Å for 4^μ-oxo) of the complexes containing encumbering substituents on the N-heterocycle, 3^μ-oxo and 4^μ-oxo, are significantly longer than the sum of their covalent radii (2.11 Å) [59]. This supports our earlier conclusions [52], suggesting that ligand constraints, as opposed to Jahn-Teller-like effects expected for high-spin d⁴-transition-metals [49], is responsible for the longer bonds in our mono oxo-bridged M(III) (M = Mn, Fe) complexes.

Fig. 16.

Monitoring the addition of O₂ to 4 (black trace) in CH₂Cl₂ at ambient temperature, formation of 4^μ-oxo (purple trace) was observed over 36 min. Scans recorded at 166 s intervals. [4] = 0.75 mM.

Fig. 17.

First order kinetics plot for the reaction between 3 and O₂ under pseudo 1st order conditions with excess O₂ in CH₂Cl₂ at 25 °C.

Fig. 18.

ORTEP diagram of oxo bridged 8^μ-oxo Hydrogen atoms, counterions, and cosolvents have been removed for clarity.

Fig. 19.

Correlation between Fe⋯N^Ar distance and λ_max for mono oxo-bridged dimers.

Fig. 20.

Correlation between oxidation potential E_p,a distance and λ_max for mono oxo-bridged dimers.

Table 4.

Selected Bond Distances (Å) and Angles (deg) for oxo bridged binuclear [Fe^III(S^Me2N₄(2-quinoEN))]₂(μ-O)(BF₄)₂·2MeCN (3^μ-oxo), [Fe^III(S^Me2N₄(6-Me-DPEN)]₂(μ-O)(PF₆)₂·(4^μ-oxo), [Fe^III(S^Me2N₄(6-H-DPEN)]₂(μ-O)(PF₆)₂⋅MeCN (8^μ-oxo), [Fe^III(S^Me2N₄(6-H-DPPN))]₂(μ-O)(PF₆)₂ (6^μ-oxo).

	3μ-oxo	4μ-oxo	8μ-oxo	6μ-oxo
Fe-S(1)	2.305(1)	2.303(1)	2.306(1)	2.271(1)
Fe-N(1)	2.174(3)	2.163(4)	2.155(3)	2.302(4)
Fe-N(2)	2.175(3)	2.208(4)	2.222(3)	2.240(4)
Fe-N(3)	2.303(3)	2.390(4)	2.160(3)	2.167(4)
Fe-N(4)	2.371(3)	2.429(4)	2.168(4)	2.132(4)
Fe–O	1.778(3)	1.7816(7)	1.787(3)	1.7985(6)
S(1)-Fe-N(1)	80.54(9)	80.3(1)	81.75(9)	79.17(11)
S(1)-Fe-N(2)	158.48(9)	158.8(1)	159.2(1)	168.12(11)
S(1)-Fe-N(3)	104.00(9)	103.63(1)	104.8(1)	99.95(12)
S(1)-Fe-N(4)	104.64(8)	106.9(1)	101.4(1)	106.90(12)
N(1)-Fe-N(3)	83.02(11)	98.1(1)	82.7(1)	91.95(15)
N(1)-Fe-N(4)	95.10(11)	80.5(1)	93.5(1)	87.77(15)
N(3)-Fe-N(4)	150.57(12)	148.7(1)	152.7(1)	152.58(17)
S(1)–Fe–O	104.15(9)	101.84(5)	104.6(1)	97.66(4)
N(1)–Fe–O	173.91(12)	168.7(1)	172.6(1)	171.62(11)
Fe–O–Fe	174.02(14)	180.00	175.9(2)	180.00

Mono oxo-bridged μ-oxo Fe dimers are typically unreactive and quite stable under most conditions [60]. However, we previously showed that proton-induced cleavage of the oxo-bridge of [Fe^III(S^Me2N₄(tren))]₂(μ-O)²⁺ (9) occurs when an electron-donating thiolate is included in the coordination sphere [57]. We hypothesized that this was because the thiolate creates a more basic oxo was, which is more susceptible to protonation. In the current study, we find that even when π-accepting ligands are incorporated into the coordination sphere, the oxo is basic enough to undergo proton-induced cleavage. For example, HOAc reacts with 4^μ-oxo in MeCN to afford a species with a red-shifted λ_max at 630 nm (Fig. 21). This spectrum is different from the (p-tolyliminium)PF₆ oxidized product derived from 4 (Figure S28). The former (Fig. 19) presumably contains an acetate coordinated to the sixth coordination site, analogous to the crystallographically characterized product derived from 9 [57,61], whereas (p-tolyliminium)PF₆ oxidized 4^ox presumably has a coordinated MeCN in the coordination sphere, given the preference that Fe(III) has for being six-coordinate. All of the ferric derivatives examined in the current study were too unstable to isolate, with the exception of the mono oxo-bridged dimers. The instability of ferric complexes containing π-accepting N-heterocycle ligands (Fig. 4) reflect their higher redox potentials (Figs. S22–S26), and thus greater metal ion Lewis acidity, of the complexes discussed herein.

Fig. 21.

Addition of HOAc to 4^μ-oxo in MeCN at 298 K.

3. Summary and conclusions

This work shows that metal ion Lewis acidity can be modulated by including π-accepting ligands in the coordination sphere, or, by increasing steric bulk and constraining both ligand geometry as well as Fe⋯N^Ar distances, especially in oxidized Fe(III)-derivatives. Ferrous, Fe (II) complexes containing more Lewis acidic metal ions have charge transfer bands in the visible region of the electronic absorption spectrum, and a higher affinity for a sixth ligand. This is in contrast to our previously reported primary amine-ligated complex, 1, which mimics superoxide reductase (SOR) function. The higher energy charge transfer transition associated with 1 made it difficult to visualize substrate-bound Fe(II) derivatives, and thus we were unable to distinguish between a concerted versus step-wise mechanism in the oxidative addition of HO₂ (Fig. 1). Neutral ligands were shown to bind reversibly to the least sterically encumbered complexes, and anionic ligands were shown to bind to all of the complexes reported herein. Thermodynamic parameters associated with the former were obtained from temperature-dependent equilibrium constants. With the exception of the most sterically encumbered complex (Fig. 11), all of the complexes described herein were shown to react with O₂, with rates dependent on oxidation potential, E_p,a indicating that an electron transfer event takes place. In contrast to our previously reported Mn(II) derivatives [49,52], dioxygen intermediates are not observed with the Fe(II) complexes reported herein.