Photocatalytic Aerobic Oxidation by a bis-Porphyrin-Ruthenium(IV) μ-Oxo Dimer: Observation of a Putative Porphyrin-Ruthenium(V)-Oxo Intermediate

Abstract

The title complexes catalyze the aerobic oxidations of hydrocarbons using visible light and atmospheric oxygen as oxygen source in sequences employing photo-disproportionation reactions. The putative oxidants, ruthenium(V)-oxo porphyrin species, can be detected and studied in real time via laser flash photolysis methods.

Selective oxidation is a key technology for the synthesis of high value chemicals in the pharmaceutical and petrochemical industries, but oxidations are among the most problematic processes to control.¹ Many stoichiometric oxidants with heavy metals are expensive and/or toxic, and thus impractical. The ideal green catalytic oxidation process would use molecular oxygen or hydrogen peroxide as the primary oxygen source, recyclable catalysts in nontoxic solvents, and an inexpensive energy source.² In this context, we have an interest in photochemical generation of highly reactive metal-oxo intermediates which, upon oxidization of substrates, give low-valent metal complexes that can be recycled for catalytic oxidations. One example of a catalytic aerobic oxidation driven by a photo-disproportionation reaction employed a diiron(III)-μ-oxo bisporphyrin complex,³ and photocatalytic oxidations of hydrocarbons using molecular oxygen as the oxygen source with no added reducing agent for the oxygen have been developed.⁴ The low reactivity of the formed iron(IV)-oxo transients and poor quantum efficiency are serious obstacles limiting the use of iron porphyrins as practical photocatalysts, but the catalytic efficiency of diiron(III)-μ-oxo bisporphyrin systems was improved by Nocera and coworkers who employed “Pacman” ligand designs with organic spacer-hinges serving to preorganize two iron centers in a co-facial arrangement.^5-8

Among the high-valent metal-oxo species, metal(V)-oxo complexes deserve special attention because they are highly reactive although rare and elusive. The known porphyrin manganese(V)-oxo complexes showed higher reactivity than well-studied iron(IV)-oxo porphyrin radical cation analogs.^{9, 10} A recent paper by Collins et al. reported spectroscopic evidence for an oxoiron(V) complex supported by a tetraanionic ligand that showed unprecedented reactivity.¹¹ Putative porphyrin- and corrole-iron(V)-oxo transients produced by laser flash photolysis methods displayed the appropriate high levels of reactivity expected for iron(V)-oxo species,^{12, 13} and we recently reported a photodisproportionation of a bis-corrole-iron(IV) μ-oxo dimer that apparently gave the same type of corrole-iron(V)-oxo transient in a system that has potential for light-driven oxidation catalysis.¹⁴

Porphyrin-ruthenium(V)-oxo transients^{15, 16} are attractive candidates for oxidations, and these species are proposed intermediates in very efficient catalytic processes,¹⁷ although not yet observed directly; computational studies suggest that they are stable with respect to ruthenium(IV)-oxo porphyrin radical cations.¹⁸ In this study, we explore the applicability of ruthenium porphyrins, which are known to display good oxidative robustness,¹⁹ in a light-driven catalytic process. We report that ruthenium(IV) μ-oxo bisporphyrin complexes catalyze the aerobic oxidation of hydrocarbons using visible light and atmospheric oxygen as oxygen source in sequences employing photo-disproportionation reactions to give putative ruthenium(V)-oxo species as shown in Scheme 1.

Scheme 1.

The diruthenium(IV)-μ-oxo-bis[5,10,15,20-tetra phenylporphyrin] precursor 1a was synthesized according to reported methods.²⁰ The hydroxyl axial ligands were readily replaced with chloride anions to give a μ-oxo complex formulated as [Ru^IV(TPP)Cl]₂O (Supporting Information, Figure S1). Two other dimer complexes, [Ru^IV(4-CF₃-TPP)OH]₂O (1b) and [(Ru^IV(4-MeOTPP) OH]₂O (1c), were prepared in a similar manner. The complexes were characterized by UV-visible, ¹H NMR, and IR spectra that matched those reported.²⁰ When dissolved in acetonitrile solution, complexes 1 were thermally stable to hydrocarbons, but irradiation with light (λ_max = 420 nm) of anaerobic solutions of complexes 1 in the presence of excess triphenylphosphine (50 mM) resulted in changes in the absorption spectra with isosbestic points (Figure 1). Similar spectral changes were previously observed upon the photolyses of diiron(III)-μ-oxo bisporphyrin complexes in the presence of Ph₃P.²¹ According to previous studies by Collamn and coworkers,²² the product formed with λ_max at 430 nm was assigned as Ru^II(TPP)(PPh₃)₂. The spectra signature of Ru^II(TPP)(PPh₃)₂ was further confirmed by production of the same species in the known experiment of the Ru^VI(TPP)O₂ with Ph₃P.²³

Figure 1.

UV-visible spectral change of 1a (6 × 10^-6 M) in the presence of excess Ph₃P (50 mM) upon irradiation with a 300-W visible lamp (λ_max = 420 nm) in anaerobic CH₃CN solution. The Spectra were recorded at t = 0, 2, 4, 6 and 10 min.

A series of ruthenium(IV)-μ-oxo bisporphyrins was evaluated in the aerobic oxidation of cis-cyclooctene (Table 1). The reactions were carried out with 0.5 μmol of catalyst in 5 mL of oxygen-saturated solution containing 4 mmol of cis-cyclooctene. After 24 hours of photolysis with visible light (λ_max = 420 nm), cis-cyclooctene oxide was obtained as the only identifiable oxidation product (> 95% by GC) with ca. 220 turnovers (abbreviated as TON representing moles of product/moles of catalyst) of catalyst 1a (entry 1). The trend in the TONs roughly paralleled irradiation times. Control experiments showed that no epoxide was formed in the absence of either the catalyst or light. The results cannot be ascribed to the chemistry of singlet oxygen, which is characterized by efficient “ene” reactions of alkenes.²⁴ The use of other solvents instead of CH₃CN, or air as oxygen source (data not shown), resulted in reduced TONs (entries 2-4). Catalyst degradation was a problem with higher-energy light, but the use of UV irradiation increased catalytic activity (entry 5). It is interesting to note that the catalytic activity was enhanced by adding small amounts of anthracene (entries 6 and 9). Quite surprisingly, the axial ligand on the metal had a significant effect, and the [Ru^IV(TPP)Cl]₂O (entry 7) gave reduced activity compared to [Ru^IV(TPP)OH]₂O. The substituent in the porphyrin ligand gave a noticeable effect on the catalytic activity (entries 1, 8 and 10), with the most electron demanding system, namely [Ru^IV(4-CF₃-TPP)OH]₂O, being the most efficient catalyst.

Table 1.

Aerobic Photocatalytic Oxidation of cis-Cyclooctene with Diruthenium(IV) μ-Oxo Porphyrins^a

entry	catalyst	solvent	t/day	TONb,c
1	[RuIV(TPP)OH]2O	CH3CN	1	220±16
	1a		2	460±40
			3	640±65
2		CHCl3	1	110±18
3		C6H6	1	140±6
4		THF	1	190±10
5d		CH3CN	1	340±8
6e		CH3CN	1	300±35
7	[RuIV(TPP)Cl]2O	CH3CN	1	70±5
8	[RuIV(4-CF3-TPP)OH]2O	CH3CN	1	250±21
9e	1b	CH3CN	1	340±9
10	[RuIV(4-MeOTPP)OH]2O 1c	CH3CN	1	190±23

^aThe reaction was carried out in a Rayonet reactor, with 0.5 μmol of catalyst in 5 mL of solvent containing 4 mmol of cis-cyclooctene. Oxygen-saturated solutions were irradiated with visible light (λ_max= 420 nm) or otherwise noted. Products were analyzed on an HP 5890 GC with a DB-5 capillary column employing an internal standard

^bTON represents the total number of moles of product produced per mole of catalyst. All reactions were run three times, and the data reported are the averages with standard deviation (1σ).

^cThe major product was cis-cyclooctene oxide, detected in > 95% yield.

^dUV-visible light (λ_max = 350 nm).

^e5 mg of anthracene was added.

In a fashion similar to that described for cis-cyclooctene epoxidation, the photocatalytic oxidations of a variety of organic substrates were examined. Table 2 lists the oxidized products and corresponding TONs using the [Ru^IV(4-CF₃-TPP)OH]₂O (1b) as the photocatalyst. The trend in the TONs roughly parallels the substrate reactivity and significant activity was observed with up to 3900 TON. After 24 h photolysis, norborbene was oxidized to norbornene oxide (exo mainly) with 200 TON (entry 1). Cyclohexene, in contrast, is susceptible to the allylic oxidation that gave primarily 2-cyclohexenone and 2-cyclohexenol along with minor expoxide (entry 2). This product distribution is no different than those typically obtained from cofacial iron porphyrin photocatalysts.⁷ Activated hydrocarbons including triphenylmethane, diphenylmethane, ethylbenzene and xanthenes were oxidized to the corresponding alcohols and/or ketones from over-oxidation with total TONs ranging from 560 to 2900 (entries 3-6). Noticeably, the oxidation of secondary benzylic alcohols gave the highest catalytic activities (entries 7-8). Competitive catalytic oxidation of ethylbenzene and ethlybenzene-d₁₀ revealed a kinetic isotope effect (KIE) of k_H/k_D = 4.8 ± 0.2 at 298 K, similar to the KIE reported for the reaction of ethylbenzene with an electron-deficient iron(IV)-oxo porphyrin radical cation species.²⁵ The observed KIE is larger than those observed in autooxidation processes,²⁶ suggesting a non-radical mechanism that involves the intermediacy of ruthenium(V)-oxo species as postulated in Scheme 1.

Table 2.

Turnover Numbers for Alkenes and Benzylic C-H Oxidations Using 1b as the photocatalyst^a

entry	substrate	product	TONb
1	norbornene	norbornene oxidec	200±40
2	cyclohexene	2-cyclohexenol	160±14
		2-cyclohexenone	350±31
		cyclohexe oxide	30±3
3	triphenylmethane	triphenylmethanol	1120±48
4	diphenylmethane	diphenylmethanol	820±102
		benzophenone	140±18
5	ethylbenzene	1-phenylethanol	380±41
		acetophenone	180±19
6d	xanthene	9-xanthone	2900±140
7	1-phenylethanol	acetophenone	3300±240
8e	9-xanthenol	9-xanthone	3900±280

^aTypically with 0.25 μmol of 1b in 5 mL of CH₃CN containing 2~4 mmol of substrate and 5 mg anthracene.

^bDetermined for a 24 h photolysis (λ_max= 420 nm). The values reported are the averages of 2-3 runs (1σ deviation).

^c> 90% exo isomer.

^dOne minor product was detected by GC but not identified.

^e48% product yield.

To probe the nature of the transient oxidizing species, we conducted laser flash photolysis (LFP) studies similar to those previously described for iron and manganese complexes.²⁷ As shown in Figure 2A, the photo-disproportionation of the bisporphyrin diruthenium (IV)-μ-oxo complexes exhibited a fast decay followed by a slow grow that were observed in different time scales (Figure 2A). Figure 2B shows a time-resolved difference spectrum for a 50 ms time scale, where the only observable transients are those that increase or decrease in concentration on the short time-scale, 2a and 3a in this case. Irradiation of the complex [Ru^IV(TPP)OH]₂O 1a with 355 nm laser light at ambient temperature in CH₃CN solution instantly produced a highly reactive transient 2a displaying a strong Soret band at 390 nm, which rapidly decayed to form a compound (3a) with Soret band at 410 nm and Q band at 530 nm (Figure 2B). The spectrum of 3a was essentially identical to that of Ru^III(OMe)TPP, which was independently prepared from a reported method.¹⁵ Accordingly, we assigned the structure of 2a as [Ru^V(O)(TPP)(OH)] and that of 3a as Ru^III(TPP)OH. It is noteworthy that the photolysis efficiency could be enhanced by adding benzophenone or anthracene, which presumably acts as a photosensitizer. In the presence of anthracene (10 mM), the photolysis of 1a had a quantum yield of 1.1× 10^-3 (see Supporting Information for details), which is 10 times greater than that reported for the photolysis of a bis-porphyrin diiron(III) μ-oxo complex.³ In a slower subsequent phase of the reaction (Figure S2 in Supporting Information), exposing photoproduct 3a to oxygen led to regeneration of the dimetal(IV)-μ-oxo complex 1a through an autooxidation pathway.²²

Figure 2.

(A) Kinetic trace at λ_max (390 nm) showing a rapid decay followed by a slow growing process after laser pulse; (B) Time-resolved difference spectrum at 0.2, 1, 2, 5, 8 and 10 ms, following 355 nm irradiation of [Ru^IV(TPP)OH]₂O (1a) in the presence of benzophenone (10 mM) in CH₃CN at 25 °C; difference spectrum = spectrum(t) - spectrum (final= 50 ms). In this representation, decaying peaks have positive absorbances, whereas growing peaks have negative absorbances. (C) Observed pseudo-first-order rate constants for reactions of 2a with cyclohexene (circles) and cis-cyclooctene (triangles) in CH₃CN.

Kinetic studies were accomplished by generating transient 2a in the presence of organic substrates at varying high concentrations under pseudo-first-order conditions. The kinetics are described by Eq. 1, where k_obs is the observed pseudo-first-order rate constant, k₀ is the background rate constant for decay in the absence of substrate, k_ox is the second-order rate constant, and [Substrate] is the concentration of substrate.

$$k_{\text{obs}}=k_0+k_{\mathrm{ox}}\left[\text{Substrate}\right]$$ (eq 1)

The results from reactions of 2a formed by photolysis of 1a are shown graphically in Figure 2C, where plots of k_obs versus the concentrations of cyclohexene and cis-cyclooctene were linear, and the rate constants for their reactions with 2a are k = 1.8 × 10³ and 1.3 × 10³ M^-1 s^-1, respectively. Second-order rate constants with other substrates are collected in Table S1 (Supporting Information). Remarkably, the observed rate constants for reactions of the putative ruthenium(V)-oxo species 2a with alkenes are 5-6 orders of magnitude greater than those for similar reactions of the well characterized trans-dioxoruthenium(VI) complexes.²³

In summary, we have demonstrated that the ruthenium(IV)-μ-oxo bisporphyrins catalyzed efficient aerobic oxidation of alkenes and activated hydrocarbons using visible light and atmospheric oxygen. The observed photocatalytic oxidation is ascribed to a photo-disproportionation mechanism to afford a putative porphyrin-ruthenium(V)-oxo species that can be directly observed and kinetically studied by laser flash photolysis methods. Further studies to characterize the observed transients more fully and to define synthetic applications are ongoing in our laboratory.