Flash Communication: Unexpected oxidative alkene cleavage behavior of a molecular molybdenum bis(dithiolene) catalyst

Abstract

Whereas homogeneous molybdenum catalysis typically promotes alkene epoxidation with peroxide oxidants, here a molybdenum bis(dithiolene) compound is shown to instead catalyze oxidative alkene cleavage, resulting in carbonyl products akin to ozonolysis reactions. A mechanism involving O₂ transfer from a Mo(O)(η²-O₂) intermediate to form 1,2-dioxetanes is proposed with support from DFT calculations.

Graphical Abstract

Molecular molybdenum catalysts are essential for industrial olefin epoxidation, with the Halcon process serving as a benchmark for commercial success.^1,2 The Halcon process employs molecular Mo(VI) catalysts and tert-butyl hydroperoxide (TBHP, ^tBuOOH) as the oxidant, efficiently converting propylene to propylene oxide at ~4 million metric tons per year scale. The terminal molybdenum-oxo (Mo=O) group has been identified as the active catalytic oxidant,^3,4 inspiring continued research and development to improve the activity and selectivity of such intermediates.^5–9

We recently disclosed that bio-inspired molybdenum bis(dithiolene) complexes show good activity for sulfide oxidation using hydrogen peroxide as the oxidant, giving exceptional sulfoxide selectivity.¹⁰ During subsequent studies, we uncovered the unexpected reactivity pattern reported here: instead of the anticipated epoxidation reaction when exposing aryl-substituted alkenes to ^tBuOOH and a molybdenum bis(dithiolene) catalyst, oxidative C=C cleavage was observed. This behavior deviates from the conventional reactivity of homogeneous Mo catalysis such as in the Halcon process. Oxidative alkene cleavage to form carbonyl compounds is a standard method in organic chemistry that converts abundant hydrocarbon feedstocks into value-added compounds or enables the installation of oxygen functionality in complex molecule settings.^11–13 The two canonical methods for oxidative alkene cleavage are ozonolysis¹³ and the Lemieux-Johnson reaction.¹⁴ Ozonolysis presents serious safety and technical concerns due to O₃ being a highly reactive, toxic, and explosive gas generated using specialized apparatus.^15–17 The Lemieux-Johnson method uses stoichiometric KMnO₄ or OsO₄, thus generating toxic heavy metal waste.^14,18 Therefore, the identification of new methodologies for oxidative alkene cleavage remains an active area of research.^19–21

We began our investigation by attempting the epoxidation of cis-cyclooctene with TBHP (1.1 eq.) in chloroform at 85 °C in the presence of [MoO₂(mnt)₂][NⁿBu₄]₂ ([Mo], mnt = maleonitriledithiolate) as a catalyst at 1.5 mol% loading. The reaction proceeded as expected, yielding the epoxidation product, cyclooctene oxide, selectively in 90% yield at 4 h (Figure 1a). However, to our surprise, changing the substrate from cis-cyclooctene to aryl-substituted alkene substrates yielded unexpected oxidative C=C cleavage rather than epoxidation.

Figure 1.

Alkene oxidation reactions catalysed by [MoO₂(mnt)₂][NⁿBu₄]₂ ([Mo], mnt = maleonitriledithiolate). For Figure 1b, product yields are reported relative to theoretical formation of 0.5 equiv. based on the limiting reagent (alkene).

The effects of reaction conditions were studied using trans-stilbene as a model substrate. After extensive investigation (see ESI), the optimal reaction conditions we identified for selective oxidative C=C bond cleavage of trans-stilbene used [Mo] catalyst (1.5 mol%), 2 equiv. of TBHP, in 1,2-dichlorobenzene solvent at 120 °C for 24 h (Figure 1b). Under these conditions, 77% conversion of trans-stilbene to a mixture of benzaldehyde (43%) and benzophenone (72%) was observed (relative to theoretical yields of 1 equiv. vs. trans-stilbene). No improvements to conversion were obtained beyond 24 h or with higher catalyst loading. Whereas benzaldehyde is the expected product of oxidative C=C cleavage, benzophenone may form from intramolecular rearrangement of the putative 1,2-dioxetane intermediate (vide infra). No conversion was observed in the absence of [Mo] catalyst. Alternative peroxy reagents or solvent choices provided lower conversions and/or lower C=C cleavage selectivity. We believe the catalysis to be homogeneous based on the lack of any obvious solid precipitates or significant color changes. Importantly, replacing the [Mo] catalyst with Na₂[MoO₄] resulted exclusively in epoxidation rather than oxidative cleavage, highlighting that the dithiolene ligands play an important role in controlling selectivity.

Next, we conducted a preliminary substrate scope study (Table 1). Changing the alkene substrate from trans-stilbene (entry 1) to 1,1-diphenylethylene (entry 2) did not have a significant impact on reaction efficiency and yielded a single product, benzophenone. (Presumably, formaldehyde was also produced.) However, further steric hindrance did influence reaction efficiency. The use of triphenylethylene (entry 3) resulted in only 57% conversion of starting material, and no conversion was observed for tetraphenylethylene (entry 4). Effects of electron-withdrawing and -donating groups on the catalysis are reported in Table S2.

Table 1.

Substrate scope of oxidative C=C cleavage.^[a]

Entry	R1	R2	R3	Conversion (%)[b]	Benzophenone (%)[c]	Benzaldehyde (%)[c]
1	H	Ph	H	77	72	43
2	Ph	H	H	74	74	0
3	Ph	Ph	H	57	39	40
4	Ph	Ph	Ph	0	0	0

^[a]Reaction conditions: alkene (2.0 mmol), TBHP (4.2 mmol), [Mo] (0.030 mmol), 1,2-Cl₂C₆H₄ (5 mL), 120°C, 24 h.

^[b]Determined by GC using mesitylene as an internal standard.

^[c]Relative to theoretical formation of 0.5 equiv. based on the limiting reagent (alkene).

Our hypothetical mechanism is informed by an alternative method for oxidative alkene cleavage that uses photosensitizers to generate singlet oxygen, which then reacts with alkenes to form 1,2-dioxetane intermediates that decompose to give carbonyl compounds.^22–24 Our proposed reaction mechanism using 1,1-diphenylethylene as a representative alkene is shown in Figure 2a. As in our previous work with this catalyst,¹⁰ we propose that the active oxidant is [(mnt)₂Mo(O)(η²-O₂)]^2-. Here, we hypothesize that this intermediate transfers O₂ to the alkene to generate [(mnt)₂MoO]^2- and 1,1-diphenyl-1,2-dioxetane. The dioxetane intermediate is known to readily decompose to benzophenone and formaldehyde at these temperatures.^22–24 Meanwhile, reoxidation of [(mnt)₂MoO]^2- to [(mnt)₂Mo(O)(η²-O₂)]^2- by 2 equiv. of TBHP would occur in a manner akin to the pathway we modeled computationally in our previous study.¹⁰

Figure 2.

(a) Hypothetical mechanism with (b) computational support. DFT calculations were conducted at the M062X-D3(SMD=CH₂Cl₂)//def2TZVPP/6–31+G** level.

Although molybdenum oxo/peroxo intermediates are well-known in the literature,^25–32 their ability to form dioxetanes from alkenes is unexplored to our knowledge. Therefore, we conducted DFT calculations to explore the energetic feasibility of this key step of our proposed mechanism (Figure 2b). The reaction between [(mnt)₂Mo(O)(η²-O₂)]^2- and 1,1-diphenyletheylene to produce [(mnt)₂MoO]^2- and 1,1-diphenyl-1,2-dioxetane was calculated to be thermodynamically favourable by ΔG = −3.3 kcal mol⁻¹. The unexpected formation of benzophenones from stilbenes likely involves rearrangement of the dioxetane intermediate by an unknown mechanism promoted at the high reaction temperature. A speculative mechanism is presented in Scheme S1. We presume that the same oxo/peroxo intermediate is responsible for epoxidation of cis-cyclooctene, and further work will be needed to determine the factors causing divergent epoxidation vs. oxidative cleavage behavior.

In conclusion, a molecular Mo catalyst with bioinspired bis(dithiolene) ligation was found to be active for oxidative cleavage of aryl-substituted C=C double bonds using TBHP as the oxidant. This behavior stands in contrast to the typical reactivity of molecular Mo complexes that are established epoxidation catalysts. The proposed mechanism involves O₂ transfer to the alkene from a key Mo(O)(η²-O₂) catalytic intermediate.

Supplementary Material

Supporting Information

The Supporting Information is available free of charge on the ACS Publications website.

Experimental & computational methods, supporting data (PDF)

Computational output (XYZ)