Keywords

9.1 Introduction

Selective oxidative functionalization of hydrocarbons with O2 as the terminal oxidant is an attractive goal. From an economic standpoint, atmospheric oxygen is one of the least expensive and abundant oxidizing agents. From an environmental perspective, the development of selective aerobic oxidation processes could minimize or even eliminate chemical waste. Organotransition metal catalysis is a viable approach to achieve these goals. In this Chapter, catalytic transformations of hydrocarbon C–H bonds will be considered that involve the use of O2 as terminal oxidant and soluble organometallic group 10 metal complexes as catalysts or catalytic intermediates. As it will be shown in this Chapter, the involvement of organometallic species may allow for diverse transformations of their metal–carbon bonds [1] leading toward various value-added hydrocarbon functionalization products. Along with an overview of representative examples of major reaction types, a discussion of the mechanisms of the reactions will be provided.

Among the transition metals, and the group 10 metals, in particular, palladium has played a prominent role in the development of selective organic oxidation reactions with O2 (“aerobic oxidation reactions”). A classic example of such processes is the oxidation of ethylene with O2 to acetaldehyde in the presence of aqueous [PdIICl4]2– and CuIICl2 cocatalysts (the Wacker process) developed in the 1950s (Eq. 9.1) [2,3,4]:

$$ 2{\text{CH}}_{2}{=}{\text{CH}}_{2}\,{{ +\, \text{O}}}_{2} \to 2{\text{CH}}_{3} {\text{CH}{=}\text{O}} $$
(9.1)

The Wacker process is an organometallic oxidation involving reactive alkylpalladium(II) species as key intermediates. The [PdIICl4]2–complex is responsible for ethylene oxidation to acetaldehyde with palladium(0) as another reaction product, which is a known stoichiometric reaction. In turn, the role of the copper cocatalyst is twofold. In its oxidized form, CuIICl2, it can convert palladium(0) back to palladium(II) producing copper(I) species as another product. In its reduced form, [CuICl2]–, it activates O2 with concomitant conversion of copper(I) species back to copper(II). Reaction (9.1) has been used on an industrial scale since the 1960s.

The use of soluble platinum complexes in aerobic oxidation of organic substrates has been known since the 1980s. An early example of such transformations utilizing O2 as the oxidant is the Shilov reaction, conversion of methane to CH3X products (X=OH, Cl) catalyzed by aqueous [PtIICl4]2– and a heteropolyacid redox cocatalyst (Eqs. 9.2, 9.3) [5]:

$$ 2 {\text{CH}}_{ 4} + {\text{ O}}_{ 2} \to 2 {\text{CH}}_{ 3} {\text{OH}} $$
(9.2)
$$ 2 {\text{CH}}_{ 4} + {\text{ O}}_{ 2} + {\text{ 2HCl}} \to 2 {\text{CH}}_{ 3} {\text{Cl }} + {\text{ 2H}}_{ 2} {\text{O}} $$
(9.3)

The original version of the reaction reported about 10 years earlier [6] utilized expensive H2PtCl6 as the oxidant. It was then discovered that H2PtCl6 could be used in a catalytic fashion with O2 as the terminal oxidant when a heteropolyacid redox cocatalyst is employed [5]. The Shilov reaction, similar to Wacker process, is also an organometallic oxidation process which involves methylplatinum(II) intermediates resulting from methane activation by platinum(II), as well as their methylplatinum(IV) derivatives resulting from oxidation of the former [7, 8]. The Shilov reaction’s main limitations are low catalyst turnover numbers (TON) resulting from gradual conversion of platinum(II) catalyst to inactive platinum(IV) and platinum(0) species, as well as poor (≤50%) selectivity in CH3X-type products due to “overoxidation” of methane leading to formaldehyde, formic acid, and CO2. Aerobic C–H functionalization reactions mediated by soluble platinum complexes have gained, so far, no practical applications although they remain in the focus of academic research [9, 10].

Finally, organometallic functionalization reactions of organic substrates mediated by nickel complexes are under development [11, 12] and the use of O2 as the terminal oxidant in such reaction has not yet been reported.

This quick introduction suggests that the most part of this chapter will be dedicated to organopalladium catalysis with much less attention paid to reactions of the other two group 10 metals.

9.2 Homogeneous Organometallic Palladium-Catalyzed Aerobic C–H Functionalization

Various types of organometallic palladium-catalyzed C–H oxidation (Eq. 9.4) and aerobic oxidative coupling of C–H (R–H) and X–H fragments (Eq. 9.5) leading to products with new C–X (R–X) bonds have been reported:

$$ 2{\text{R}}{-}{\text{H}} + {\text{O}}_{2} \to 2{\text{ROH}} $$
(9.4)
$$ 2{\text{R}}{-}{\text{H}} + 2{\text{X}}{-}{\text{H}} + {\text{O}}_{2} \to 2{\text{R}}{-}{\text{X}} + 2{\text{H}}_{2} {\text{O}} $$
(9.5)

In an ideal case of a 100% selective transformation, reactions of the first type would produce no chemical waste. In the second case, ideally, the only by-product would be water. An extensive recent review covering aerobic functionalization of olefinic substrates is available [13]. Some representative examples of reactions of both types, (Eq. 9.4) and (Eq. 9.5), are listed in Table 9.1 and structures of the specific ligands 1–11 used in these reactions are given in Fig. 9.1. A literature analysis shows that oxidative aerobic transformations of olefinic substrates and arenes are explored better than those of alkanes. While compiling representative organometallic palladium-catalyzed aerobic C–H functionalization reactions, preference was given to more challenging processes involving arene, olefin, or alkane C–H activation (all entries except 9–12). The assignment of the type of a C–H bond involved in oxidative functionalization is purely formal for some reactions involving olefins serving either as hydrocarbon substrates (entries 9–12) or as coupling partners (entries 20–26). In these specific examples, the relevant organopalladium intermediates result from the addition of palladium(II) species across an olefin C=C bond (olefin insertion into PdII-ligand bond) and not from the olefin C–H bond activation. As an alternative to the olefin insertion, activation of olefins at a PdII center may involve direct allylic C–H bond cleavage leading to allylpalladium(II) intermediates (entry 5).

Table 9.1 Some representative examples of C–H oxidation and oxidative aerobic coupling of C–H bonds of organic substrates and X–H bonds of their coupling partners mediated by Pd complexes. The structures of specific ligands 1–11 used in these reactions are given in Fig. 9.1
Full size table
Fig. 9.1
figure 1

Ligands 1–11 used in reactions in Table 9.1

Full size image

The reactions in Table 9.1 are organized according to the type of the functional group introduced and new C–X bond formed, the type of the substrate C–H bonds involved, and the type of the X–H coupling partner, when applicable. Reactions leading to functionalization of substrate C(sp2)–H bonds include hydroxylation (entry 1), acetoxylation (entry 2), amidation/amination (entries 6–9), imidoylation (entries 13–15), homocoupling (entry 16), arylation (entries 17–19), alkenylation (entries 20–24), and alkylation (entry 26). Transformations of substrate C(sp3)–H bond include acetoxylation (entries 3–5), chlorination (entry 4), amidation/amination (entries 10–12), alkenylation (entry 25), and oxidative dehydrogenation of cyclohexane derivatives (entries 27–29). Besides X–H type coupling partners, boronic acid derivatives were also used in some cases (entry 19). Notably, the use of chiral supporting ligands may lead to a highly enantioselective product formation with the product enantiomeric excess up to 98% (entry 11).

The key to understanding and overcoming challenges associated with the development of aerobic C–H functionalization reactions and, in particular, most difficult aerobic functionalization of alkanes, lies in the understanding of their mechanisms. Notably, in the past two decades, the rapid development of Pd-catalyzed aerobic oxidative C–H functionalization has become possible thanks to close attention paid to mechanisms of palladium-mediated C–H bond activation [46] and O2 activation reactions [47]. Accordingly, in the subsequent discussion, some key mechanistic details of these reactions will be considered.

9.2.1 General Mechanisms of Palladium-Mediated Aerobic C–H Functionalization

Three plausible catalytic cycles showing major steps of palladium-catalyzed aerobic C–H functionalization are given in Scheme 9.1. In the “non-redox” PdII—only catalytic cycle A, the metal oxidation state remains the same during three major steps, the substrate activation step a, the O2 activation step b, and the product-releasing step c leading to the substrate functionalization product (hydrocarbyl-OH).

Scheme 9.1
scheme 1

Plausible simplified catalytic cycles for palladium-mediated aerobic C–H bond functionalization

Full size image

In mechanism B, a PdII/Pd0 redox couple is involved. The +2 metal oxidation state is not changed at the substrate activation step a. The step b leading to substrate functionalization product (hydrocarbyl-X) involves elimination of C–X bond from a palladium(II) center with concomitant reduction of PdII to Pd0, and the O2 activation step c leads to reoxidation of Pd0 to PdII.

Finally, the mechanism C is also palladium redox-based and involves PdII/PdIV redox couple. The PdII center does not change its oxidation state during the substrate activation step a and subsequent reaction with a coupling partner HX, step b. Two other steps, c and d, involve oxidation of PdII hydrocarbyls to their PdIV derivatives and C–X elimination of the product (hydrocarbyl-X) from the palladium(IV) center with concomitant reduction of PdIV to PdII, respectively.

9.2.1.1 Substrate Activation Step

All the basic mechanisms A–C in Scheme 9.1 imply that substrate activation leading to hydrocarbylpalladium(II) species (step a) occurs without change of the metal oxidation state +2.

Even in the cases where palladium(0) complexes are used as pre-catalysts (examples in entries 13–15 in Table 9.1), the authors argue that the actual catalytically active species are palladium(II) complexes. The latter result from oxidation of palladium(0) species with O2 involving N–H acidic substrates, O-methyl hydroxamic acids, which serve as a source of anionic amido ligands for the resulting PdII center (see step b, mechanism B) [29], see, e.g. Eq. 9.6:

$$ {2}\left[ {{\text{Pd}}^{0} } \right] + {\text{O}}_{2} + 4{\text{H}{-}\text{N}}\left( {\text{OMe}} \right)\left( {\text{COR}} \right) \to {2 }\left[ {{\text{Pd}}^{\text{II}} \left( {{\text{N}}\left( {\text{OMe}} \right){\text{COR}}} \right)_{2} } \right] + {2{\text{H}}}_{2} {\text{O}} $$
(9.6)

Similar may be the role in Pd0 recycling of protected aminoacids 6 and 7 (Fig. 9.1) used as ligands in some aerobic C–H functionalization reactions (examples in entries 19–21, 24, Table 9.1).

Importantly, all the mechanisms of substrate activation by palladium(II) complexes discussed below require prior substrate coordination to the metal. Hence, the slow rates of ligand substitution in catalytically active metal species and strong coordination to palladium(II) center of a supporting ligand or substrate may decrease dramatically the overall catalyst turnover frequency. As a result, the judicious choice of supporting ligand for a catalyst may be very important, which is, in particular, a reason for the success of catalytic systems utilizing weakly coordinating 2-fluoropyridine 4 [31, 32] or bidentate 4,5-diazafluoren-9-one 2 [25] ligands (Fig. 9.1).

Holding these considerations in mind, it is very remarkable that the \( {\text{Pd}}_{2}^{{^{0} }} \left( {\text{dba}} \right)_{{_{ 3} }} \)-based catalytic system in examples in entries 13–15, Table 9.1 is very tolerant of heterocyclic donor groups present in substrates. These donor groups can strongly bind to palladium(II) center and severely inhibit catalysis of C–H functionalization by palladium(II) complexes. A possible explanation to this tolerance is that the palladium(II) center which is generated in reaction (9.6) above is coordinated to basic amido ligands and as such can be involved in substrate C–H bond activation/deprotonation (vide infra), experiencing minimal inhibiting effect of coordinating heterocyclic donor groups of the substrates [29].

9.2.1.1.1 Olefin Insertion into PdII-Ligand Bond

Following substrate coordination to a palladium(II) center, its activation by the metal leading to hydrocarbylpalladium(II) species may proceed either as the substrate C–H bond cleavage or, for unsaturated substrates, as an addition (insertion) reaction. For olefinic substrates, even with available relatively acidic allylic C–H bonds, step a may not involve the substrate C–H activation, as it is the case in oxidative amination of olefins in the examples in entries 9–12, Table 9.1. Instead, based on available mechanistic tests, authors of these catalytic systems propose that formation of alkylpalladium(II) intermediates in step a occurs as cis-aminopalladation of the olefinic C=C bonds (Scheme 9.2), and not as allylic C–H bond cleavage [27].

Scheme 9.2
scheme 2

Formation of alkylpalladium(II) intermediates as a result of cis-aminopalladation of olefinic substrates

Full size image
9.2.1.1.2 Deprotonation of PdII–Coordinated C–H Bonds
9.2.1.1.2.1 Allylic C–H Deprotonation of Olefins

In turn, in the absence of strong nucleophilic ligands such as amides in examples given in entries 9–12, Table 9.1, olefinic substrates with available allylic C–H bonds can undergo allylic C–H deprotonation by the action of a metal-coordinated carboxylate (Scheme 9.3) [19] or a similar basic ligand [29], or even a free carboxylate serving as a base, as it was found computationally [48]. Some representative reactions involving activation of olefinic substrates via allylic C–H bond cleavage are given in entries 5, 27–29, Table 9.1.

Scheme 9.3
scheme 3

Formation of allylpalladium(II) intermediates as a result of deprotonation of allylic C–H bonds of olefinic substrates involving coordinated carboxylate ligand

Full size image
9.2.1.1.2.2 Directed C–H Activation. Concerted Metallation–Deprotonation Mechanism

Considering non-olefinic substrates in Table 9.1, their quick inspection shows that many of them have metal-coordinating heteroatoms, e.g., carboxylate oxygens, examples in entries 1, 19–20, 24, a quinoline nitrogen, entry 3, an oxime or pyridine nitrogen in entries 4, 25, an anionic amide nitrogen resulting from N–H bond deprotonation in entries 6–8, 13–15, 17, 23, in a close proximity to C–H bonds involved in subsequent oxidative functionalization reaction. It was shown that functionalized hydrocarbon substrates containing suitable donor groups undergo C–H bond activation only after prior coordination of the donor group to the metal [49]. As a result, only those C–H bonds of the substrate that are accessible for the donor group-coordinated metal can be involved in subsequent transformations. Hence, the position of the donor groups relative to the substrate’s various C–H bonds determines the regioselectivity of the C–H activation step a. Formation of five-membered palladacyclic intermediates is usually kinetically favored over six-membered metallacycles. Notably, metallacyclic intermediates with both smaller and larger rings can form. As a result, for arene derivatives with donor groups (DG) attached to one of the arene carbon atoms, such as CO2–, CH2CO2–, CONR2, NHCOR, CH2NR2, 2-pyridyl, 2-oxazolyl, 2-imidazolyl, N=NAr, CH2SR, or CH2OH selective metallation and subsequent functionalization is most facile for the arene C–H bonds that are positioned ortho- to the donor group [49, 50] (examples in entries 1, 6–8, 13–15, 17–20, 23–24, Table 9.1). Similar rules apply for alkane C–H bond functionalization when dealing with functionalized alkane substrates bearing directing groups [49, 51] (examples in entries 3, 4, 25, Table 9.1). Importantly, by changing the length and configuration of a tether between an arene carbon to which the tether is attached and the donor group, one can achieve a rare selective meta-C–H bond functionalization of the arene, as it is the case in an example in entry 21 in Table 9.1 [37].

The considerations above also imply that for substrates having several types of chemically nonequivalent C–H bonds, selective functionalization of some of them may be a daunting problem in palladium catalysis. At the same time, this is one of the points of growth and development of this area [37, 47], where joint experimental and computational modeling efforts are especially promising [52].

The mechanism of C–H activation most common for substrates with donor groups and substrates not having directing groups such as non-functionalized arenes or alkanes (entries 2, 16, Table 9.1) is the concerted metallation–deprotonation (Scheme 9.4) [46], which was also studied computationally [53, 54]. In either case, before the deprotonation step can occur, a substrate C–H bond has to be coordinated to the metal center. Such coordination can enhance dramatically the C–H bond acidity and facilitate subsequent C–H deprotonation. Expectedly, C–H bonds of nonactivated alkane fragments are the least acidic and, as such, are most difficult to activate and functionalize.

Scheme 9.4
scheme 4

Formation of palladium(II) hydrocarbyls as a result of a concerted metallation-deprotonation (CMD)

Full size image

A substrate C–H bond coordination to the metal is greatly facilitated when the substrate has a donor group that can coordinate to the metal, thanks to the absence of the entropic penalty for the C–H bond coordination step (Scheme 9.4, top). Since hydrocarbon C–H bonds are very poor electron donors, the latter effect is of immense importance for metal-mediated C–H activation and functionalization. Interestingly, coordination of a substrate C–H bond to the metal center can be facilitated when a palladium(II) carboxylate is a strained chelate, such as palladium(II) pyridine-2,6-dicarboxylate (Scheme 9.5, top) [16, 17]. Dissociation of a carboxylate arm from the metal relives the chelate ring strain and is, therefore, facilitated, which accelerates coordination of the substrate C–H bond to be functionalized. Non-strained analogs with larger chelate size are less reactive (Scheme 9.5, bottom).

Scheme 9.5
scheme 5

C–H bond activation by palladium(II) pyridine-2,6-dicarboxylate complexes [16]

Full size image

9.2.1.2 C–X Bond Formation Step

This step is the most critical for achieving a desirable type of C–H bond functionalization. Since the key reaction intermediates produced at the step a are organopalladium(II) species, knowledge of their reactivity [1] is very important when designing new catalytic reactions. This step can be viewed as functionalization of transient organopalladium species.

In any of the mechanisms A–C (Scheme 9.1), the C–X bond formation may result already at the step a when the substrate is an olefin involved in an insertion reaction (see, e.g., Scheme 9.2).

If step a involves a substrate C–H bond cleavage which results in a hydrocarbylpalladium(II) intermediate, then the C–X bond formation occurs typically at the step b (mechanisms A, B) or d (mechanism C).

9.2.1.2.1 O2 Insertion into PdII–C Bond

According to mechanism A in Scheme 9.1, the PdII–C bond functionalization (step b) may be as “simple” as O2 insertion into PdII–C bond which does not involve the metal redox change, (see an example in entry 1 in Table 9.1 as well as a discussion in the next section, “O2 activation step”). The expected O2 insertion product is a hydrocarbylperoxo complex. A few such well-defined stoichiometric reactions are known. They involve methylpalladium(II) complexes [55,56,57,58,59]. One of the first reported O2 insertion reactions (Scheme 9.6, top) involves a dimethylpalladium(II) species and is a radical chain process [55], similar to a reported later analogous O2 insertion involving a neutral monomethylpalladium(II) compound [56]. Another reaction in Scheme 9.6 is photochemical and involves cationic monomethylpalladium(II) species (Scheme 9.6, bottom) [57,58,59]. All of these reactions occur in aprotic media and the resulting methylperoxo palladium(II) species are formed in high yields.

Scheme 9.6
scheme 6

Some reported examples of direct thermal [55] or photochemical [57,58,59] O2 insertion into PdII–CH3 bond of methylpalladium(II) species

Full size image

Protonolysis of the products of O2 insertion into PdII–CH3 bond in Scheme 9.6 can lead to free methylhydroperoxide, an unstable and explosive chemical. Hence, the practical value of such products may be low. In this regard, an in situ conversion of hydrocarbylperoxo metal species into palladium(II) alkoxo complexes or free alcohols would be more desirable. In fact, both MeO2H and MeOH form in a photocatalytic reaction of O2 with a water-soluble anionic methyl palladium complex [(dpms)PdIIMe(OH)]–, besides ethane which is a major reaction product (Scheme 9.7, top) [60]. Importantly, the formation of MeO2H can be fully suppressed by a slight modification of the reaction conditions with a concomitant increase of the MeOH yield up to 50% with the rest of the balance being ethane. It was shown that various hydroperoxides RO2H (R=H, Me, t-Bu) react cleanly and rapidly with the methylpalladium(II) reagent, [(dpms)PdIIMe(OH)]–, to form the corresponding ROH in high yields. A proposed reaction sequence shown in Scheme 9.7 (center and bottom) involves formation of a hypothesized highly electrophilic methylpalladium(IV) species responsible for the production of various Me–X products detected in the mixtures with various nucleophiles and resulting from their attack at the CH3–PdIV fragment of the proposed methylpalladium(IV) transient.

Scheme 9.7
scheme 7

Photochemical dioxygen activation by a water-soluble methylpalladium(II) complex and conversion of RO2H to ROH [60]

Full size image
9.2.1.2.1.1 Reductive Elimination of C–X Bond from a PdII or a PdIV Center

According to mechanisms B and C (Scheme 9.1), C–X bond formation occurs as a result of reductive elimination from PdII(X) hydrocarbyls (mechanism B, step b) or PdIV(X) hydrocarbyls (mechanism C, step d) species. The ligands X necessary for such reductive coupling are introduced into palladium(II) coordination sphere prior to the C–X bond elimination (mechanism B) and, typically, but not always, prior to the PdII to PdIV oxidation step c (mechanism C) as a result of a ligand exchange, olefin insertion into palladium(II)-ligand bond (examples in entries 20–26) or, rarely, C–H bond activation at the electrophilic PdIV center [61].

Mechanism C involving a PdII/PdIV redox couple is rare in aerobic C–H functionalization chemistry. In particular, the authors of the catalytic system in entry 2 [15], Table 9.1, proposed involvement of a PdII/PdIV redox couple with a C(sp2)–O reductive elimination from a PdIV hydrocarbyl resulting from the oxidation of its PdII precursor with HNO3. Similarly, possible involvement of the PdII/PdIV redox couple was discussed for reactions in entries 3 [16] and 4 [18] leading to C(sp3)–X (X=O, Cl) reductive elimination from PdIV species which, in fact, occurs as an SN2 process. In the latter system, the oxidant responsible for the generation of PdIV hydrocarbyls is HNO3, similar to the reaction in entry 2. In turn, for the catalytic system in entry 3, the PdIV hydrocarbyls were speculated to be produced aerobically from their PdII precursors [16]. It was shown computationally that the formation of PdIV transients is thermodynamically viable thanks to the ability of the tripod ligand 1 (Fig. 9.1) to adapt a facial coordination mode. At the same time, a mechanism involving an O2 insertion into PdII–C bond also could not be excluded [17]. Notably, there are precedents of reactions between O2 and dimethylpalladium(II) complexes supported by facially chelating ligands that lead to palladium(IV) derivatives [62, 63].

The involvement of PdII/Pd0 redox catalysis (mechanism B) is most commonly proposed in various aerobic C–H functionalization reactions. The relevant examples in Table 9.1 are the reactions leading to C(sp3)–O bond formation, such as in the catalytic system in entry 5, C(sp2)–N bond formation, such as in arene and alkene amination reactions in entries 6–12, and C–C bond formation, such as in the arene imidoylation reactions (entries 13–15), arene homocoupling (entry 16), arene arylation (entries 17–19), arene and alkane alkenylation (entries 20–25) and alkylation (entry 26).

Finally, a special case of the product forming step associated with the mechanism B which does not require the presence of an actual coupling partner is dehydrogenation of various cyclohexane derivatives in reactions in entries 27–29. In this case, the new C=C bonds result from β-hydrogen atom elimination of palladium(II) hydrocarbyl intermediates.

9.2.1.2.2 O2 Activation Step

Another step which is critical for any catalytic aerobic C–H functionalization process is O2 activation. Direct O2 insertion into PdII–C bond is one of the possibilities which has been already characterized in section “O2 Insertion into PdII–C Bond”.

9.2.1.2.2.1 O2 Activation by Redox Cocatalysts

O2 activation carried out by a redox cocatalyst is very common in aerobic Pd-catalyzed C–H functionalization, especially in its older versions. Some redox cocatalysts that were proven efficient are copper(II) complexes, heteropolyacids, and lower nitrogen oxides NOx (x = 1, 1.5, 2). These cocatalysts in their reduced form, e.g., copper(I) or NO, react rapidly with O2 to form species capable of oxidizing palladium center from lower to higher oxidation states, Pd0 to PdII (mechanism B, Scheme 9.1) or, in some cases, converting PdII hydrocarbyls to PdIV hydrocarbyls (mechanism C). Some of the catalytic systems in Table 9.1 utilize these cocatalysts, CuX2 (entries 7, 8, 23), NOx (entries 2, 4), and a heteropolyacid H4[PMo11VO40] (entry 25). While copper(II) cocatalysts are traditionally assumed to support recycling of Pd0 to PdII species (mechanism B), the systems utilizing NOx as redox mediators (entries 2, 4) are proposed to support reactions involving a PdII/PdIV redox couple (mechanism C, Scheme 9.1).

The presence of a redox-active cocatalyst in a catalytic system, such as those mentioned above, does not exclude the option that O2 activation will actually be carried out by palladium species. In particular, the authors of the reaction in entry 16 have observed only negligible effect of Cu(OTf)2 additive on Pd0 reoxidation. They have concluded that the major role of Cu(OTf)2 cocatalyst in the palladium-catalyzed oxidative homocoupling of o-xylene is not O2 activation but rather that of a Lewis acid enhancing reactivity of Pd(OAc)2 [31].

9.2.1.2.2.2 O2 Activation by Pd0 Species

The catalytic systems where O2 activation is carried out by Pd0 species are becoming increasingly important practically and are interesting mechanistically. These reactions convert Pd0 complexes to PdII peroxo species (Eq. 9.7), e.g., Pd(PPh3)4 is oxidized to Pd(κ2-O2)(PPh3)2 [64]. Notably, triphenylphosphine liberated in the latter reaction can reduce palladium peroxide and form corresponding phosphine oxide. This fact suggests that the practical value of phosphine ligands in aerobic palladium catalysis may be limited.

O2 activation by Pd0 species can be most efficient when suitable ligands are present. For instance, a PdL2 complex 12 with very bulky N-heterocyclic carbene (NHC) ligands L=N,N’-bis(2,2”,6,6”-tetramethyl-m-terphen-5’-yl)imidazole-2-ylidene) (Fig. 9.2) reacts with O2 at room temperature even in a solid state [65].

Fig. 9.2
figure 2

Some palladium complexes characterized in O2 activation

Full size image
$$ {\text{L}}_{\text{n}} {\text{Pd}}^{ 0} + {\text{O}}_{2} \to\,{\text{cis}}{\text{-L}}_{2} {\text{Pd}}^{\text{II}} \left( {{\text{O}}_{2} } \right) + \left( {{\text{n}} - 2} \right){\text{L}} $$
(9.7)

The resulting palladium(II) peroxo complexes are relatively basic and can react stepwise with acids to form first palladium(II) hydroperoxo complexes (Eq. 9.8) and, eventually, H2O2 (Eq. 9.9):

$$ \text{cis}{\text{-L}}_{2} {\text{Pd}}^{\text{II}} \left( {{\text{O}}_{2} } \right) + {\text{HX}} \to {\text{Pd}}^{\text{II}} {\text{X}}\left( {\text{OOH}} \right){\text{L}}_{2} $$
(9.8)
$$ {\text{Pd}}^{\text{II}} {\text{X}}\left( {\text{OOH}} \right){\text{L}}_{ 2} {\text{ + HX}} \to {\text{Pd}}^{\text{II}} {\text{X}}_{ 2} {\text{L}}_{ 2} {\text{ + H}}_{ 2} {\text{O}}_{ 2} $$
(9.9)

In turn, H2O2 released in the last reaction may act as an oxidant with respect to Pd0 and/or reactive hydrocarbyl PdII species (e.g., Scheme 9.7, bottom) [60] or decompose into O2 and H2O.

In the absence of suitable ligands, the rate of the oxidation reaction in Eq. 9.7 may be too slow and/or the stability of Pd0 complexes may be too low, so that a catalyst deactivation leading to the formation of catalytically inactive Pd black may become highly competitive with the reaction (9.7). Such catalyst deactivation is the major reason why many palladium-based catalytic systems involving PdII/Pd0 couple (mechanism B, Scheme 9.1) require high catalyst loading, 10–20% and even higher (see Table 9.1 for examples). Notably, at some intermediate stages leading to the formation of palladium black, palladium(I) species [25, 66] and, ultimately, soluble palladium clusters/nanoparticles may be produced which often are also catalytically active in aerobic oxidation reactions [67, 68].

Interestingly, until recently, the utilization of organic ligands in aerobic functionalization catalysis by palladium compounds was not practiced, although some polar solvents such as DMSO that can coordinate to palladium(II) center were used successfully in a number of aerobic palladium-catalyzed C–H functionalization reactions (examples in entries 6, 22, Table 9.1). This situation is, in part, a reflection of a formerly poor understanding of the underlying aerobic chemistry of Pd0 species [47]. One of the important reasons for this lag is related to the fact that rates of aerobic C–H functionalization by soluble palladium complexes are often zero order in [O2], and their turnover-limiting step is the substrate C–H activation, so making the characterization of the O2 activation step difficult in such systems.

9.2.1.2.2.3 O2 Activation by PdII Hydrides Versus Pd0 Species

Catalytically competent palladium(0) species are expected to result from C–X reductive elimination of PdII(hydrocarbyl)X complexes in mechanism B, Scheme 9.1. Alternatively, palladium(0) species may be produced as a result of H–Y elimination of palladium(II) hydrides (Eq. 9.10) which, in turn, are formed as a result of β-hydrogen atom elimination of suitable palladium(II) alkyl, alkoxo, or similar species.

$$ {\text{L}}_{\text{n}} {\text{Pd}}^{\text{II}} \left( {\text{H}} \right){\text{Y}} \to {\text{L}}_{\text{n}} {\text{Pd}}^{ 0} {\text{ + HY}} $$
(9.10)

Importantly, palladium(II) hydrides are also able to react with O2. The reaction proceeds via O2 hydrogen atom abstraction/radical recombination pathway leading to O2 insertion into Pd–H bond and formation of palladium(II) hydroperoxides [69] (Eq. 9.11), so allowing to return PdII back to the catalytic cycle.

$$ {\text{L}}_{\text{n}} {\text{Pd}}^{\text{II}} \left( {\text{H}} \right){\text{Y + O}}_{ 2} \to {\text{L}}_{\text{n}} {\text{Pd}}^{\text{II}} \left( {\text{OOH}} \right){\text{Y}} $$
(9.11)

The reaction between O2 and a (PCP)PdII(H) complex 13 (Fig. 9.2) was characterized kinetically to reveal a first-order dependence of its rate on pO2 and a large deuterium kinetic isotope effect, kPdH/kPdD = 5.8, all consistent with an H-atom abstraction mechanism. The mechanism was also analyzed computationally [70].

More extensive studies of reactions between various palladium(II) hydride complexes and O2 (Eq. 9.11) have led to a conclusion that an alternative reaction sequence (9.10)–(9.7)–(9.8), that is HY reductive elimination—oxidation, leading to palladium(II) hydroperoxo complexes can be faster than the direct route (9.11) [71], although, in general, both pathways may be very competitive kinetically [72, 73]. In some cases, just a minor variation in the electronic properties of the anionic ligand Y, e.g., a p-substituted benzoate in bis-NHC palladium(II) hydride complexes trans-L2PdII(H)(O2CC6H4-p-X) 14 (Fig. 9.2), can lead to a change in the reaction mechanism from the direct O2 insertion (Eq. 9.11), with a large deuterium kinetic isotope effect, kPdH/kPdD = 3.1 for X = OMe [73], to a stepwise transformation (9.10)–(9.7)–(9.8), with a very small kPdH/kPdD = 1.3 for X = H [71]. Interestingly, benzoquinone additives which are often present as a cocatalyst in palladium-catalyzed aerobic C–H functionalization reactions (see, e.g., an example in entry 19 in Table 9.1) were found to accelerate the reaction sequence (9.10)–(9.7)–(9.8) [74].

Notably, the HY reductive elimination—oxidation reaction sequence (9.10)–(9.7)–(9.8) and, in particular, its first step (9.10), is strongly favored in palladium(II) complexes bearing labile monodentate L-type ligands since three-coordinate LPdII(H)Y species resulting from a ligand L dissociation eliminate H–Y at faster rates (Eq. 9.10, n = 1 vs. n = 2). The use of bidentate ligands appears to also favor the HY reductive elimination—oxidation reaction sequence, as compared to the direct pathway (Eq. 9.11), when one of the ligand’s donor atoms is basic enough to deprotonate the PdII–H bond. That is usually the case for N-donor ligands. The deprotonation can occur upon this donor atom dissociation from the metal. As a result, the authors of [47] conclude that the O2 activation in most aerobic catalytic systems used till date is carried out, most typically, by Pd0 species and not by palladium(II) hydrides.

9.3 Homogeneous Organometallic Platinum—Catalyzed Aerobic CH Oxidation

As it was mentioned in the introduction, the first-ever developed platinum-based catalytic system for aerobic C–H functionalization allowed to carry out an overall very challenging transformation, the conversion of gaseous methane to CH3X products (Eqs. 9.2, 9.3), albeit with low [PtCl4]2– catalyst turnover (≤6) and poor selectivity in CH3X products (≤50%) [5]. The reaction mechanism [7, 8] is similar to mechanism C shown in Scheme 9.1 for aerobic palladium catalysis. Notably, the heteropolyacid used in these experiments as a redox mediator was also shown by the authors to oxidize methanol, so contributing to the overall low reaction selectivity in CH3X products. Subsequent attempts to develop more efficient variants of the reaction were made. In 2001, some modifications to the aerobic system were undertaken by introducing aqueous CuCl2 as a redox mediator instead of a heteropolyacid and using water-soluble alkanesulfonic acids as substrates which are much easier to handle than gaseous methane. Water was used as the reaction medium [75]. These changes allowed to achieve the catalyst turnover numbers up to 43–52 after 4 h of reaction at 160 °C for ethanesulfonic acid as a substrate (Eq. 9.12):

$$ 2 {\text{C}}_{ 2} {\text{H}}_{ 5} {\text{SO}}_{ 3} {\text{Na + O}}_{ 2} \to 2 {\text{HOCH}}_{ 2} {\text{CH}}_{ 2} {\text{SO}}_{ 3} {\text{Na}} $$
(9.12)

The reaction was ~50–76% selective with respect to the methyl group oxidation product, 2-hydroxyethanesulfonic acid shown in Eq. 9.12, with the rest of the balance being mostly the corresponding aldehyde and carboxylic acid.

A more recent reinvestigation of the Shilov reaction was undertaken in 2010 [9]. The authors used microfluidics technique and screened a number of redox mediators for the reaction of CH4 with O2 in water at 180 °C. They observed up to 49 catalyst turnovers after 6 h with the selectivity in CH3OH ~50% using either Fe2(SO4)3 or a heteropolyacid as a redox mediator. Formic acid accounted for the rest of the balance.

Although the catalyst turnover numbers in both cases are much better than in the original Shilov publication [5] and, in fact, in many palladium-based systems listed in Table 9.1, the resulting oxidation products, 2-hydroxyethanesulfonic acid and methanol, may, most likely, be readily available at a lower cost using traditional methods of their preparation. Further reaction developments are in order.

Notably, learning from recent progress in aerobic catalytic C–H functionalization by palladium complexes, a possible direction for future research in catalytic platinum chemistry may target a better understanding of the underlying C–H activation, aerobic oxidation, and C–X reductive elimination chemistry of platinum species involved and rational design of ligands for these transformations [76]. As an example of such efforts, platinum(II) complex 15 supported by a newly designed sulfonated pincer ligand in Scheme 9.8 supports facile aerobic stoichiometric C–H functionalization of a series of arenes leading to derived arylplatinum(IV) complexes 17 [10]. The reaction is more efficient for electron-rich arenes. The minor products of the arenes functionalization are oxidatively C–C coupled complexes 18. The fraction of the undesirable product 16 forming along with 18 can be significantly reduced in the presence of p-hydroquinone.

Scheme 9.8
scheme 8

Aerobic stoichiometric C–H functionalization of arenes mediated by platinum(II) complex 15 [10]

Full size image

9.4 Development of Homogeneous Organometallic Nickel-Catalyzed Aerobic CH Oxidation

Nickel-based catalytic systems for organometallic aerobic C–H functionalization are still at an early stage of development. In fact, no such systems have been reported so far. At the same time, some basic step of a plausible catalytic cycle incorporating nickel complexes can be envisioned. A recent publication [77] discloses a possible approach toward donor group/auxiliary 19-directed C(sp3)–H activation at a nickel(II) center to produce metallacyclic nickel(II) alkyls 20 (Scheme 9.9). The behavior of the reported system resembles that of similar palladium(II) systems, although at somewhat higher temperatures for the nickel-based one. The kinetics of the reaction in Scheme 9.9 has been characterized in detail, including a large deuterium kinetic isotope effect, kH/kD ~ 7, and the reaction was proposed to operate a concerted metallation—deprotonation mechanism also common in palladium chemistry.

Scheme 9.9
scheme 9

Auxiliary-directed C(sp3)–H activation at a nickel(II) center (top) [77] and aerobic oxidative functionalization of NiII-C(sp3) bonds (bottom) [78, 79]

Full size image

Notably, although not shown to involve organonickel intermediates, hydrogen atom transfer (HAT) chemistry involving C–H bonds of a series of alkylarenes such as 9,10-dihydroanthracene, toluene, and ethylbenzene has also been demonstrated in their reactions with a few isolated nickel(III) complexes [78].

In turn, stoichiometric aerobic Ni–C bond functionalization reactions have been known for a long time. Examples of aerobic oxidative C(sp2)–O [79] and C(sp3)–N [80] coupling reactions are given in Scheme 9.9, bottom.

Finally, there is a substantial body of recent work detailing the intimate chemistry of dioxygen activation at a nickel(I) center supported by various chelating ligands which can lead to nickel(II) superoxo or peroxo complexes and even nickel(III) peroxo species [81].

Overall, the fact that the key steps of a potential catalytic cycle of aerobic C–H functionalization by nickel are established, may be viewed as a promise of possible future development of the related catalytic chemistry.

9.5 Conclusions

Aerobic organometallic C–H functionalization catalysis by palladium complexes has shown a significant development over the past two decades. The major driving force behind this success was an increased attention to and an improved understanding of the reaction mechanisms and the role that the ligand environment at the metal plays in such reactions. Numerous challenges remain on the way toward making hydrocarbon C–H bonds a “functional group” that can be readily and selectively transformed to another “classic” functional group by using the right catalytic system. Understanding the mechanism of C–H activation and the factors that control its selectivity in substrates with chemically non-equivalent C–H bonds can drive the progress in this area. A collaboration of experimentalists and computational chemists in such a challenging area of research may become fruitful. While organometallic palladium aerobic oxidation catalysis has already been an established and has a solid reputation among synthetic chemists, analogous platinum-based systems for aerobic C–H functionalization are scarce. Slower reaction rates of reactions involving platinum species, as compared to analogous palladium chemistry, maybe a reason behind such poor performance in aerobic platinum catalysis. But as in the case of the recent development in aerobic palladium catalysis, greater attention to reaction mechanisms and ligand design might help improve the situation. Finally, organometallic nickel aerobic oxidation catalysis is not established yet but there are some promising reports that suggest possible future development of this field. Overall, selective catalytic aerobic C–H functionalization reactions are poised to grow in their importance in the coming years and the group 10 metals can continue contributing to this development.