Meso-N-methylation of a porphyrinoid complex: Activating the H-atom transfer capability of an inert Re^V(O) corrolazine

Abstract

The selective alkylation of a single meso-N atom of a corrolazine macrocycle is reported. Alkylation has a dramatic impact on the physicochemical properties of Re^V(O)(TBP₈Cz). New electron-transfer and hydrogen-atom-transfer reactivity is also seen for this complex, including one-electron reduction, which gives an air-stable 19π-electron aromatic radical complex.

Graphical Abstract

Meso-N-methylation of a corrolazine macrocycle results in changes in both the physicochemical properties and in the reactivity of a Re^V(O) corrolazine complex.

The development of novel porphyrinoid compounds is essential for applications in a range of fields that rely on porphyrinoid structures, including bioinorganic models of heme proteins,¹ electrochemical energy storage systems,² new dye-sensitized solar cells (DSSCs),^3a,b and molecules for photodynamic therapy (PDT).^3c The synthesis of ring-contracted porphyrins, and in particular, ring-contracted tetraazaporphyrins, or corrolazines (Cz),^{1a,b, 4} has been a focus of our research group. The corrolazine ligands are able to support a range of metal ions in various oxidation states, including high-valent metal-oxo and metal-imido species.⁵ However, synthetic modification of the corrolazine ring has thus far been limited to the peripheral substituents on the β-carbon positions.⁴

Porphyrins and corroles can be modified at the meso-carbon atoms, leading to a large array of porphyrinoid compounds with varying structural, electronic, and reactivity properties.⁶ The meso-N atoms in Cz, and in the related porphyrazines and phthalocyanines, appear to preclude facile alkyl-functionalization at the meso position.⁷ We envisioned that alkylation of the meso-N positions might be feasible given that protonation of these positions was occurring in the former complexes.⁸ If a generalized alkylation procedure could be devised, it would open a pathway to meso-N substituted porphyrinoid compounds with a range of new steric and electronic properties.⁹ Such a synthetic strategy would allow access to meso substitution in corrolazines and porphyrazines similar to what has been seen in porphyrin chemistry.

Herein we describe the synthesis of a cationic rhenium corrolazine [Re^V(O)(N-MeTBP₈Cz)]⁺[OTf]⁻ (1), which has been alkylated at one of the meso-N atoms of the Cz ring. To our knowledge, this synthesis is the first report of direct, selective alkylation at a meso-N position in a porphyrinoid compound.⁹ The methylated meso-N site was confirmed by single crystal X-ray diffraction (XRD). The selective alkylation of the starting Re^V(O)(TBP₈Cz) complex has a profound effect on the spectroscopic features, redox potentials and reactivity of this metallocorrolazine complex. The unalkylated Re^V(O) complex is unreactive toward electron- and hydrogen-atom donors, whereas alkylated 1 undergoes one-electron reduction to give an air-stable 19π-electron radical complex. The stabilization of organic radicals with extended porphyrinoid π systems has been the focus of much effort,^{9b, 10} and reduced 1 is a rare example of such a species. A PCET mechanism is indicated from kinetic and thermodynamic analyses of the reaction of 1 with H-atom donors.

The synthesis of 1 was accomplished by refluxing Re^V(O)(TBP₈Cz) with excess MeOTf in toluene (Scheme 1). The solution slowly changes from dark-green to brown upon the formation of complex 1. Purified complex 1 can be isolated as a dark brown solid in 38% yield. The meso-N methylation causes a red-shift and a decrease in the extinction coefficient of the Soret band (464 nm), and a red-shifting of the Q-band (732 nm) by 62 nm compared to Re^V(O)(TBP₈Cz) (Fig. 1a). The changes in the absorbance spectrum are similar to the changes seen for meso-N protonation of Re^V(O)(TBP₈Cz), as well as Mn corrolazines.⁸ The ¹H NMR spectrum for 1 shows that the two-fold symmetry of the starting Re^V(O) complex has been disrupted (Fig. 2). The NMR spectrum indicates that methylation occurs on one of the meso-N atoms that does not lie on the mirror plane bisecting the pyrrole-pyrrole linkage. The singlet at 5.05 ppm arises from the new meso-CH₃ group, which is shifted downfield by the aromatic ring current. Similar chemical shifts are seen for porphyrins methylated at the meso-carbon position.¹¹ Additional support of the assignment of 1 as the monoalkylated product is afforded by LDI-MS, where the parent ion was found at 1572.53 m/z (M⁺) (calc’d 1572.81 m/z).

Scheme 1.

Synthesis of [Re^V(O)(N-MeTBP₈Cz)]⁺[OTf]⁻ (1).

Fig. 1.

UV-vis spectra of (a) neutral Re^V(O)(TBP₈Cz) (green), meso-N-protonated [Re^V(O)(TBP₈Cz)(H)]⁺ (red), and meso-N-methylated 1 (blue) in CH₂Cl₂. (b) UV-vis spectral changes upon addition of excess HBArF to 1 (blue) in CH₂Cl₂ to form dicationic 2 (red).

Fig. 2.

Comparison of ¹H NMR spectra showing the (a) aromatic and (b) t-Bu region of 1 (green) and Re^V(O)(TBP₈Cz) (violet) in CD₂Cl₂ at 25 °C.

Crystallization of complex 1 was not successful, but addition of the proton donor [H(OEt₂)₂]⁺[B(C₆F₅)₄]⁻ (HBArF) to 1 in CH₂Cl₂ led to red shifts in both the Soret and Q-bands (Fig. 1b) indicative of meso-N protonation, and resulted in crystallization of the dicationic derivative [Re^V(O)(N-MeTBP₈Cz)(H)]²⁺[BArF]⁻₂ (2). The crystal structure is shown in Fig. 3, and confirms that methylation has occurred at a single meso-N atom. The proton on the opposing meso-N position was successfully located from difference Fourier maps (Fig. S9, ESI^†), and two BArF⁻ counterions were found in the crystal lattice, confirming the charge balance. The structure exhibits a short Re–O bond of 1.6643(17) Å, and an out-of-plane (N_pyrrole) distance for Re of 0.744 Å, both of which are similar to the neutral Re^V(O) complex. The phenyl groups adjacent to the N–CH₃ substituent are canted to accommodate the methyl group (C_α–C_β–C_ipso–C_ortho) (dihedral angle: 64.2°, 65.7°), as compared to the adjacent phenyl groups (dihedral angle: 44.0°, 29.2°). The angles around the N–CH₃ group (C(1)–N(7)–C(97) = 117.9(2)°; C(16)–N(7)–C(97) = 117.7(2)°; C(1)–N7–C16 = 124.3(2)°) are indicative of sp² hybridization, with the CH₃ group lying slightly above the mean plane of the 23-atom macrocyclic core (CH₃-plane = 0.43 Å). The ¹H NMR spectrum of 2 (Fig. S5, ESI^†), generated in situ by addition of HBArF to 1 in CD₂Cl₂, displays a broad resonance at 12.8 ppm consistent with meso-N protonation. Unfortunately, the dicationic species 2 is extremely moisture-sensitive and converts back slowly to 1 even in anhydrous CH₂Cl₂, making further reactivity studies difficult.

Fig. 3.

Displacement ellipsoid plots (40% probability level) at 110(2) K of (a) the dication of 2 and (b) side view with peripheral aryl groups omitted. All H-atoms except for the meso-H and disorder were removed for clarity.

The influence of remote meso-N-methylation on the Re–O bond was probed by ATR-IR spectroscopy. The IR spectra for 1 with natural abundance ¹⁶O (1-¹⁶O) and isotopically enriched ¹⁸O (1-¹⁸O) (>99%, incorporated from H₂¹⁸O)^8c in the terminal oxo position is shown in Fig. S3 (ESI^†). Comparison of the spectrum for 1-¹⁶O with Re^V(O)(TBP₈Cz) does not reveal any obvious new peaks. However, the IR spectrum for 1-¹⁸O reveals a new peak at 953 cm⁻¹ in comparison with 1-¹⁶O, which is consistent with a Re–O stretch (Fig. S3, ESI^†). A value of ν(Re–¹⁶O) = 1011 cm⁻¹ was calculated by using a simple diatomic oscillator model, which falls under an intense corrolazine vibrational mode centered at 1001 cm⁻¹, also seen in other corrolazine compounds.¹² The Re–O stretch in 1-¹⁸O is shifted to higher energy by 8 cm⁻¹ as compared to ν(Re–¹⁸O) = 945 cm⁻¹ for unmethylated Re^V(¹⁸O)(TBP₈Cz).^8c This shift is quite similar to the 11 cm⁻¹ upshift observed for ν(Re–O) upon meso-N protonation of Re^V(O)(TBP₈Cz).^8c Thus both remote meso-N methylation and protonation appear to cause a slight strengthening of the Re–O bond in these complexes (an influence from a change in symmetry on ν(Re–O) cannot be ruled out). To support these assignments, frequency calculations were performed via density functional theory (DFT) computations. Geometry-optimized structures for truncated models of 1 and Re^V(O)(TBP₈Cz) were obtained at the PBE0/LANL2TZ/6-31G** level of theory, (Fig. S15, ESI^†). These structures led to calculated values of ν(Re–O) = 1054 cm⁻¹ for the parent compound and 1063 cm⁻¹ for 1. Although these vibrational frequencies are larger than the experimental data,¹³ the difference between the calculated ν(Re–O) values for 1 and the parent complex is Δν(Re–O)) = 9 cm⁻¹, in excellent agreement with experiment.

Further insights regarding the electronic differences between 1 and Re^V(O)(TBP₈Cz) were obtained through cyclic voltammetric measurements (Fig. S4, ESI^†). Complex 1 exhibits two reversible waves at −561 mV and −906 mV versus Fc⁺/Fc in CH₂Cl₂. This electrochemical profile is different from Re^V(O)(TBP₈Cz), which shows only one reversible wave at +565 mV, assigned to a Cz ring oxidation.^8c The addition of the methyl group makes Cz ring oxidation inaccessible over the solvent window, and the redox couples at −561 mV and −906 mV for 1 can be assigned to consecutive Cz ring reductions. We have shown previously that monoprotonation of Re^V(O)(TBP₈Cz) to form [Re^V(O)(TBP₈Cz)(H)]⁺ led to a similar disappearance of the ring oxidation wave at +565 mV and the appearance of an irreversible Cz ring reduction wave at −645 mV. Attempted reaction of this complex with H-atom, O-atom and electron transfer substrates led either to no reaction or deprotonation. These results contrast those for 1, where a reversible ring reduction is observed. The electrochemical data show that complex 1 is significantly more difficult to oxidize than Re^V(O)(TBP₈Cz), but should be susceptible to reduction by an appropriate one-electron reducing agent.

Addition of the one-electron reducing agent Cr(C₆H₆)₂ to 1 in CH₂Cl₂ led to the changes in the UV-vis spectrum seen in Fig. 4a. Isosbestic conversion is observed, giving a new species with λ_max = 440, 489 nm, 920 nm. The decrease in the extinction coefficient of the Soret band and the appearance of a long-wavelength band at 920 nm is consistent with a radical delocalized on the corrolazine π-system. The 920 nm band is significantly more red-shifted than the analogous red-shifted peaks in corrolazine π-radical-cations.^[8c,14] However, a similarly red-shifted band at 915 nm was recently reported for a P^V porphyrazine π-radical complex.^[10b] The new 920 nm species can be quantitatively converted back to 1 by titration with the one-electron oxidant Cp₂Fe⁺PF₆⁻ (Fig. S11, ESI^†). Taken together, the data indicate that cationic 1 undergoes a reversible, one-electron reduction to give a neutral π-radical complex, [Re^V(O)(N-MeTBP₈Cz)]^• (3) (Scheme 2). This complex could be isolated as a red-brown solid by removal of solvent in vacuo, which then slowly decays with a half-life of 23 h under aerobic conditions at 23 °C (Fig. S12, ESI^†). Thus 3 is a rare example of an isolable, air-stable porphyrinoid π-radical complex, and can be described as a 19π electron aromatic system. Very few other 19π electron porphyrinoid species are known.¹⁰

Fig. 4.

(a) UV-vis spectral titrations of 1 + Cr(C₆H₆)₂ (0 – 1 equiv) in CH₂Cl₂. (b) EPR spectrum of 1 (0.84 mM) and PhNHNH₂ (1.2 M) in CH₂Cl₂ at 295 K.

Scheme 2.

Electron-transfer and hydrogen-atom-transfer reactions with meso-N-methylated 1 and the 19π-electron radical complex 3.

The neutral complex Re^V(O)(TBP₈Cz) is unreactive toward H-atom donors. However, given that 1 could be reduced by one electron-transfer agents, we examined this complex for its ability to react with H-atom donors. The reaction of 1 with TEMPOH (BDE(O–H)_DMSO = 72 kcal/mol)¹⁵ in CH₂Cl₂ at 23 °C led to conversion to 3 (Fig. S13, ESI^†). For TEMPOH, E_1/2 = 0.71 V (vs Fc⁺/Fc) makes outer-sphere ET to 1 prohibitively disfavored (ΔG_ET = +1.27 eV). A second H-atom donor, phenylhydrazine (PhNHNH₂ BDE (N–H)_DMSO = 75 kcal/mol) also reacts with 1 to give the same UV-vis spectral changes as TEMPOH (Scheme 2). These reactions indicate that 1 is capable of abstracting H atoms from weak O–H and N–H bonds.

Reaction of 1 with excess PhNHNH₂ led to the fluid solution EPR spectrum for 3 shown in Figure 4b. A narrow, six-line signal centered near g ≈ 2 is seen. This spectrum can be assigned to an S = ½ radical delocalized over the Cz π system and split by ¹⁴N hyperfine coupling. DFT calculations for the doublet state for 3 (PBE0/LANL2TZ/6-31G** level of theory) yield a spin density plot (Figure S16, ESI^†) that shows the unpaired electron is delocalized in a Cz π orbital that includes the three meso-N atoms, consistent with the observed ¹⁴N hyperfine coupling. In comparison, the π-cation radical complex [Re^V(O)(TBP₈Cz]^•+ exhibits a sharp singlet at g = 2.00 with no ¹⁴N hyperfine splitting,^8c and the spin density calculated for this complex (Fig. S16, ESI^†) is concentrated on the pyrrole carbon atoms.

Mechanistic insights regarding the HAT reactivity of 1 were obtained from kinetic measurements. Reaction of 1 with excess TEMPOH led to pseudo-first-order decay of 1 and production of 3 as seen by UV-vis, and varying the TEMPOH concentration led to a linear second-order plot with k₂ = 0.76(2) M⁻¹ s⁻¹. A kinetic isotope effect = 1.4 was found with TEMPOD. Concerted HAT reactions involving metal-oxo complexes often exhibit larger KIEs,¹⁶ although with some exceptions.¹⁷ The relatively small KIE seen for 1 may suggest a proton-coupled electron-transfer (PCET) process.¹⁸ It is reasonable to speculate that a weakly basic meso-N atom of the Cz ring is the initial proton acceptor, which then is likely rapidly deprotonated by the OTf⁻ counterion.^8b However, the exact fate of the proton in the PCET reactions with 1 is not known at this time.

In summary, a facile method for alkylation of meso-N-substituted porphyrinoid compounds has been reported. The post-cyclization treatment of other porphyrazines and phthalocyanines with appropriate alkyl electrophiles should be straightforward, opening the door to a wide range of new porphyrinoid compounds. The attachment of the methyl group to the meso-N atom in 1 has a profound influence on spectral properties and redox potentials, and reduction of 1 leads to a rare, air-stable 19π-electron radical species. Complex 1 is also capable of abstracting hydrogen atoms from weak O–H and N–H bonds, suggesting that porphyrinoid π-radical complexes in synthetic systems or heme proteins may have as yet undiscovered potential for oxidative reactivity toward substrates. The strategy of meso-N alkylation may also allow for the future installation of pendant groups orthogonal to the tetrapyrrolic plane of the Cz ligand, which may help in activating or stabilizing metal-oxygen intermediates.