Nitrate as a Redox Co-Catalyst for the Aerobic Pd-Catalyzed Oxidation of Unactivated sp³-C–H Bonds

Abstract

This paper describes a new method for the catalytic aerobic oxygenation of unactivated sp³-C–H bonds. This transformation utilizes Pd(OAc)₂ as a catalyst in conjunction with NaNO₃ as a redox co-catalyst. Both oxime ether and pyridine derivatives are effective directing groups for these reactions. The oxygen incorporated into the product derives from the solvent (acetic acid). Preliminary results show that the addition of simple NaCl to the reaction mixture results in aerobic chlorination under analogous conditions.

Introduction

The development of metal-catalyzed methods for converting sp³-C–H bonds into C–O bonds using dioxygen as a terminal oxidant remains a grand challenge in organometallic chemistry.^1,2 Methods for the selective aerobic oxygenation of unactivated 1° C–H bonds in the presence of weaker benzylic, allylic, 2°, or 3° C–H bonds remain particularly elusive.^1,2 Over the past decade Pd-catalyzed ligand-directed C–H oxidation has emerged as a powerful approach to achieve sp³-C–H acetoxylation, alkoxylation, and hydroxylation.³ However, the oxidants used in these transformations are most typically reagents such as PhI(OAc)₂, IOAc, or K₂S₂O₈, which have the significant disadvantages of high cost, poor atom economy, the formation of stoichiometric byproducts, and/or moderate functional group tolerance.

Mechanistic studies suggest that these reactions proceed via three key steps: (a) ligand-directed C–H activation to form cyclometalated Pd^II complexes of general structure I, (b) 2e^– oxidation of I to form high valent Pd complex II, and (c) C–O bond-forming reductive elimination from II to release the oxygenated product (Figure 1).⁴ Importantly, at pH = 1, O₂ has a comparable oxidation potential to PhI(OAc)₂ ,^5,6 suggesting that it should be thermodynamically capable of effecting the 2e^– oxidation of palladacycle I to high valent Pd intermediate II. However, efforts to utilize dioxygen as a terminal oxidant in these transformations have generally been hampered by the slow kinetics of the aerobic oxidation of palladacycles.^7,8 Recent elegant studies by Vedernikov^2a,b and by Yu^2c have shown that it is possible to address this challenge through careful selection of the supporting ligand and/or the substrate. However, these successful examples of aerobic Pd-catalyzed ligand-directed C–H oxygenation are far from a general solution, because they exhibit a narrow substrate scope. Most notably, neither is effective for promoting the C–H oxygenation of unactivated sp³-C–H bonds, which are arguably the most challenging substrates for these transformations.³

Figure 1.

Proposed mechanism for Pd-catalyzed ligand-directed C–H oxidation.

We reasoned that more general aerobic C–H oxidation reactions might be possible by using a redox co-catalyst⁹ that has high kinetic reactivity towards both O₂ and palladacycles like I (Figure 2). This concept is inspired by the Pd^II/0-catalyzed Wacker oxidation, in which Cu(II)¹⁰ or polyoxometalates (POM's)¹¹ are utilized as co-catalysts in conjunction with O₂ as the terminal oxidant. A similar strategy has also been employed in some Pd-catalyzed benzene and methane oxidation reactions, most typically using POM or POM/NO₂ as redox co-catalysts.¹² The current manuscript uses this approach to achieve the Pd-catalyzed aerobic oxidation of unactivated sp³-C–H bonds in pyridine and oxime ether substrates using nitrate salts as redox co-catalysts. Both the scope and mechanism of these reactions are described, and their implications for expanding this method to diverse sp³-C–H functionalization reactions are discussed.

Figure 2.

Concept of redox co-catalysis for aerobic C–H oxidation

Results and Discussion

The Pd(OAc)₂-catalyzed C–H acetoxylation of substrate 1 was selected as a test reaction, since it has been reported to proceed efficiently with oxidants like PhI(OAc)₂ and K₂S₂O₈.¹³ With air as the oxidant, this transformation provided only traces of product 2 (1% yield, Table 1, entry 1). Redox co-catalysts that are commonly used in Pd^II/0 catalysis (e.g., H₅PMo₁₀V₂O₄₀ (POM-V₂) and Cu(OAc)₂) proved ineffective (entries 2 and 3). Thus, we next explored nitrate and nitrite salts as precursors to more kinetically reactive NO₂.¹⁴ Importantly, Campora has shown that NO₂ can oxidize Pd^II to Pd^IV under mild conditions.¹⁵ Additionally, NO₂ has been used as a redox co-catalyst in a wide variety of different transformations.^9,10b,16 Gratifyingly, the use of NaNO₂ or NaNO₃ as the co-catalyst resulted in moderate yield (48 and 53%, respectively) of 2 (entries 4 and 5). Interestingly, while these reagents have previously been used for the Pd-catalyzed nitration of aromatics,¹⁷ the corresponding nitrated products were not observed in the current transformations.^18,19 Under an N₂ atmosphere, the NaNO₃-mediated reaction proceeded in significantly lower yield (14%), corresponding to approximately 1 turnover of the redox co-catalyst (entry 6).

Table 1.

Optimization of redox co-catalyst

entry	co-cat	mol % co-cat	oxidant	yielda
1	none	0	Air	1%
2	POM-V2b	10	Air	0%
3	Cu(OAc)2	10	air	1%
4	NaNO2	10	air	48%
5	NaNO3	10	air	53%
6	NaNO3	10	nonec	14%
7	NaNO3	10	O2	70%
8	NaNO3	25	O2	91%
9d	NaNO3	25	O2	>95%

^aYields determined by ¹H NMR spectroscopic analysis and represent an average of 2 runs.

^bPOM-V₂ = H₅PMo₁₀V₂O₄₀.

^cN₂ atmosphere.

^d110 °C.

Further optimization showed that moving from 1 atm of air to 1 atm of O₂ in the Pd(OAc)₂/NaNO₃-catalyzed reaction resulted in a significant boost in yield to 70% of 2 (entry 7). Increasing the loading of NaNO₃ to 25 mol % and the temperature to 110 °C led to further increases in yield. Under the final conditions (5 mol % of Pd(OAc)₂, 25 mol % of NaNO₃, and 1 atm O₂ in AcOH/Ac₂O at 110 °C for 18 h), 2 was formed in quantitative yield (entry 9).

We next evaluated the substrate scope of this transformation. As shown in Table 2, a variety of substrates containing both oxime ether and pyridine directing groups underwent aerobic Pd(OAc)₂/NaNO₃-catalyzed sp³-C–H acetoxylation.²⁰ Substrates containing 1° sp³-C–H sites β to the directing group generally underwent smooth and high yielding C–H oxygenation. In contrast, the functionalization of a 2° C–H bond to form 6 proceeded in only modest 41% isolated yield, even with 1 equiv of NaNO₃.²¹ In substrates containing tert-butyl substituents proximal to the directing group, triacetoxylated products predominated (for example, entries 2 and 6-9). Importantly, ligand-directed aerobic oxidation at unactivated 1° sp³-C–H sites β to the directing group proceeded selectively, even in the presence of much weaker (but not proximal) benzylic C–H bonds (entry 9). In some instances, increased loading of NaNO₃ (up to 1 equiv per C–H bond being functionalized) was necessary to achieve high yields (e.g., entries 5 and 10); however, the low cost of this oxidant relative to something like PhI(OAc)₂ still makes this a synthetically useful procedure.²²

Table 2.

Substrate scope for C–H oxygenation

entry	substrate	product	isolated yield
1			80%
2			83%
3			60%
4			51%
5			41%
6a			79%
7a			78%
8a			77%
9			68%
10			80%

^a4 Å molecular sieves added.

We hypothesize that these transformations proceed via the series of catalytic cycles shown in Figure 3. Step i involves the decomposition of nitrate to NO₂. This process is well-precedented,¹⁴ and has been exploited to access NO_X redox co-catalysts in a number of other aerobic oxidation reactions.¹⁶ Next the NO₂ and 2 equiv of AcOH react with cyclopalladated intermediate I to generate high oxidation state Pd compound II along with NO and H₂O (step ii).¹⁵ Carbon-oxygen bond-forming reductive elimination (step iii) then releases the C–H oxidation product, while the NO is oxidized by O₂ to regenerate NO₂ (step iv).¹⁶

Figure 3.

Proposed catalytic cycle. The complete supporting ligand set on Pd is unclear at this time, and nitrate and/or nitrite may serve as ligands during catalysis.

To directly probe for the intermediacy of NO, we conducted the catalytic reactions in the presence of 2,6-di-tert-butyl-4-methylphenol (BHT). NO is known to react rapidly with BHT to afford 2,6-di-tert-butyl-4-methyl-4-nitrosocyclohexa-2,5-dienone (TBMND),²³ and the consumption of NO in this manner would be expected to have a detrimental effect on the C–H acetoxylation reaction. As anticipated, the addition of 25 mol % of BHT to the catalytic oxidation of 1 resulted in a significant reduction in the yield of 2, from >95% to 12% under otherwise analogous conditions (eq. 1).²⁴ When the NaNO₃/O₂ oxidation reaction of 1 was assayed after 30 min, ¹H NMR spectroscopic analysis showed resonances consistent with the formation of TBMND (in 26% yield based on BHT).²⁵ Together, these two experiments provide preliminary support for the generation of NO as a key intermediate in this transformation.

(1)

Another important feature of the mechanism proposed in Figure 3 is that the oxygen atom in the product derives from acetic acid and not from O₂. To test this, we performed the reaction of substrate 1 in the presence of 1 atm of ¹⁸O₂. Consistent with the proposed mechanism, HRMS analysis of the isolated product showed that <5% of the ¹⁸O label was incorporated into 2 under these conditions (eq. 2).²⁶

(2)

This observation suggests that the use of different nucleophilic solvents and/or additives should enable the incorporation of other functional groups under aerobic oxidation conditions. To test this possibility, we examined the Pd(OAc)₂/NaNO₃-catalyzed aerobic oxidation of 2-tert-butylpyridine in propionic acid. Gratifyingly, this transformation provided 80% yield of the corresponding oxygenated product 12 (eq. 3).

(3)

A second preliminary experiment showed that the addition of 10 equiv of NaCl to the Pd(OAc)₂/NaNO₃-catalyzed aerobic oxidation of 2-tert-butylpyridine produces a mixture of chlorinated products 13, 14, and 15 (eq. 4). While the latter reaction has thus far proven challenging to optimize further, this result provides extremely promising precedent that redox cocatalysis can be utilized to achieve diverse aerobic C–H functionalization reactions.

(4)

Conclusions

In summary, this paper demonstrates the use of a combination of Pd(OAc)₂ and NaNO₃ or NaNO₂ to catalyze the aerobic C–H oxygenation of unactivated sp³-C–H bonds. Preliminary mechanistic and synthetic studies suggest that this approach could enable more general C–H functionalization reactions with dioxygen as the terminal oxidant. Investigations aimed at understanding the mechanism, decreasing catalyst loadings, and further increasing the scope of these reactions are currently underway in our laboratory. More generally, since commercial Pd(OAc)₂ is frequently contaminated with nitrite salts,¹⁹ we suggest that redox co-catalysis by NO_x may play a previously unappreciated role in a variety of other ‘Pd(OAc)₂’-catalyzed aerobic oxidation reactions.