O2-Mediated Te(II)-Redox Catalysis for the Cross-Dehydrogenative Coupling of Indoles.

Very few elements in the periodic system can catalytically activate O2, such as in the context of cross-dehydrogenative couplings. The development of O2-activating catalysts is essential to enable new and sustainable reactivity concepts to emerge, because these catalysts also often feature specific activating interactions with the target substrates. In this context, the unprecedented Te(II)/Te(III) catalyzed dehydrogenative C3-C2 dimerization of indoles is described herein. The fact that O2 can be directly utilized as a terminal oxidant in this reaction, as well as the absence of any background reactivity without the redox-active Te catalyst, constitute very important milestones for the fields of cross-dehydrogenative couplings and tellurium catalysis.

According to Klaproth, tellurium is only the 20th metal to have been discovered,[1] yet its use as a metal catalyst in chemical reactions is still remarkably rare.[2−22] Several of these rely on chalcogen-bond[23,24] activation, usually of organic electrophiles containing heteroatom-derived polarizable functional groups. For example, Huber and collaborators described in 2019 a nitro-Michael addition reaction enabled by the tellurium dinuclear catalyst Te-cat1 (Scheme a).[14,15] In 2021, Pale, Mamane, and coauthors developed tellurium catalyst Te-cat2 for the activation of N-bromosuccinimide in a series of electrophilic bromination reactions (Scheme b).[16] Also in 2021, Gabbai and collaborators developed dinuclear catalyst Te-cat3, which was designed to activate a platinum complex by sequestrating chloride anions, notably enabling a cycloisomerization reaction (Scheme c).[17] However, tellurium can also readily accommodate many different oxidation states and is relatively versatile in its redox reactivity, as demonstrated by the seminal works of Detty, Ando, McCormick, and other research groups.[2−13,25] Tellurium therefore possesses key features for complex redox catalysis. On the basis of a recent serendipitous discovery that the phenotellurazine Te-cat4 could catalyze the O2-mediated cross-dehydrogenative coupling of phenols with phenothiazines (Scheme d),[26,27] we envisioned that this class of Te(II)/Te(III) redox-labile heterocycles could furnish the basis for a new generation of broadly applicable redox catalysts. In addition, the specific chalcogen-bonding ability[23,24] of tellurium might enable new classes of coupling reactions. The objective of the present work is therefore double: (1) to demonstrate that the Te(II)/Te(III) redox catalysis concept can be applied to more than just one type of oxidative coupling reaction, thus proving its generality, and (2) to apply this concept to more general and useful classes of reactions and substrates, such as unprotected indoles (Scheme e).[28−31]

Scheme 1

Recent Selected Te-Catalysis Applications

Cross-dehydrogenative couplings (CDCs) are a very important class of coupling reactions, because they bypass preactivation steps and are thus inherently step and atom efficient.[32−39] Unfortunately, the chemical oxidants that CDCs often require in order to become thermodynamically feasible are not always atom efficient. Thus, the use of molecular oxygen as a terminal oxidant for dehydrogenative coupling reactions is quite attractive—producing only water as the final reduction byproduct—although this is challenging because of the inherent stability of O2.[40] Therefore, the activation of O2 almost always requires a catalyst in the context of CDCs. Traditionally, Cu salts have often been utilized for this purpose because of their propensity to form activated oxo complexes.[41] Nevertheless, the use of O2 as the sole terminal oxidant in CDCs remains rare, and therefore new technologies that can activate it must be developed further. We thus commenced our study by associating our recently discovered Te(II)-phenotellurazine catalyst (Te-cat4, 10 mol %) with some indoles and phenols under an O2 atmosphere, initially under basic reaction conditions (Table ).[26,27] While our original idea was to intercept the indole’s oxidation with a phenol, we soon identified the dehydrogenative C3–C2-coupled product 2 in 73% isolated yield (entry 1). Product 2 can be formed in a dehydorgenative fashion by a number of previously known methods.[42−48] However, only a few utilized O2 as the terminal oxidant, and these sometimes required a photochemical setup.[49−52] It soon transpired that both the phenol additive and the Te catalyst are essential for this reaction (Table , entries 1–3). Further optimization experiments (entries 4–11) found that (1) the base, which we had used in a previous Te-catalyzed reaction,[26,27] is not required, (2) 4-Ph-phenol (add1) is the optimal phenol additive, and importantly (3) 2-methoxyphenotellurazine (Te-cat5) is the optimal Te(II) catalyst, in comparison to Te-cat4 (no functional group on the catalyst backbone) or Te-cat6 (electron-withdrawing functional group on the catalyst backbone: 2-trifluoromethylphenotellurazine). The optimal conditions are reported in Table , entry 11 (product 2a, 92% isolated yield).

Table 1

Screening of Reaction Conditionsa

entry	catalyst	additive	base	yield of 2a (%)b
1	Te-cat4	add1	K2HPO4	73
2		add1	K2HPO4	nd
3	Te-cat4		K2HPO4	52
4	Te-cat5	add1	K2HPO4	84
5	Te-cat6	add1	K2HPO4	63
6	Te-cat5	add2	K2HPO4	51
7	Te-cat5	add3	K2HPO4	75
8	Te-cat5	add4	K2HPO4	71
9	Te-cat5	add1	NaHCO3	81
10	Te-cat5	add1	K3PO4	15
11	Te-cat5	add1		92
12c	Te-cat5	add1		91
13	Te-cat5	add2		60
14	Te-cat5	add3		81
15	Te-cat5	add4		60
16	Te-cat5	add5		nd
17	Te-cat5	add6		nd
18	Te-cat5		PhCO2Na	36
19	Te-cat5			32
20d	Te-cat5	add7		nd
21d	Te-cat5	add8		nd
22e	Te-cat5	add1		81

Reaction conditions unless specified otherwise: 1a (0.5 mmol), catalyst (10 mol %), additive (1 equiv), and base (1 equiv) in o-dichlorobenzene (ODCB, 1.5 mL). nd = product not detected.

Isolated yields.

Reaction performed in the dark.

Some decomposition noted.

With 1 mol % catalyst and 10 mol % additive after 8 h.

Performing the reaction in the dark yielded a similar result (2a, 91%; Table , entry 12), indicating that the involvement of singlet O2 can be reasonably excluded.[4] Importantly, none of the tested additives (add2–add8) performed any better than the 4-Ph-phenol add1, whether in the presence or the absence of a base (entries 13–21). The role of the 4-Ph-phenol additive (add1) must therefore be specific to its structure and associated properties, whether in terms of radical persistency and scavenging or in terms of pKa.

Impressively, decreasing the Te-cat5/4-Ph-phenol catalytic loading by 1 order of magnitude (1 mol % of Te catalyst, 10 mol % of 4-Ph-phenol; Table , entry 22) still afforded coupling product 2a in 81% isolated yield, albeit with a longer reaction time of 8 h. On consideration that there is no background reaction in the absence of a Te catalyst, this constitutes an important proof of concept for the further development of Te-catalyzed cross-dehydrogenative coupling reactions. In addition, the finding that the 2-methoxy-phenotellurazine Te-cat5 is a more potent catalyst in comparison to its more electron-neutral and -poor congeners (respectively Te-cat4 and Te-cat6) will probably guide the design of future Te catalysts.

With the optimized conditions in hand, we then explored the scope of this reaction (Scheme ). Several common functional groups (OMe, CF3, Me) at either the indole scaffold or the aryl part have been successfully utilized with moderate to good yields. Halogens (F, Cl, Br) are also well tolerated at various positions. The highest yield was obtained with a 2-naphthyl unit as the aryl moiety (2l, 95%). Moreover, the method was found to be easily scalable. Indeed, product 2a could be isolated in 90% isolated yield on a 2 mmol scale (92% for 0.5 mmol, Scheme ; see also the Supporting Information). However, indoles bearing strong electron-withdrawing groups, such as nitro or cyano derivatives, deliver only very low yields.

Scheme 2

Reaction Scope and Isolated Yields

Standard conditions: indole (0.5 mmol), Te-cat5 (10 mol %), and add1 (1 equiv) in ODCB (1.5 mL) under an O2 atmosphere (1 atm), for 3 h.

Neither 2-methylindole 3 nor ethyl 1H-indole-2-carboxylate 4 delivered any C3–C2 coupling product. Moreover, 2-phenylindol-3-one 5, a species that is sometimes detected as a trace byproduct (confirmed by NMR), does not deliver any of the corresponding C3–C2 product either, whether alone or in combination with another indole substrate as the nucleophile, such as 1q (Scheme , eq 1). This clearly demonstrates that 2-phenylindol-3-one 5 cannot be a productive reaction intermediate. In addition, both BHT and TEMPO poison the reaction (Scheme , eq 2). No trapped adducts could be identified in the reaction mixtures, however.

Scheme 3

Mechanistic Experiments

We were next intrigued by the role of the 4-Ph-phenol additive. One reasonable hypothesis was that the Te-cat5 (pre)catalyst might first click[53] on the 4-Ph-phenol additive (add1) to produce the dehydrogenative phenochalcogenazination product Te-cat7, which we speculated might be the true active catalyst. We therefore prepared and isolated Te-cat7 (see the Supporting Information) and then engaged it as a control catalyst (10 mol %, Scheme , eq 3), in the absence of any free 4-Ph-phenol add1. This did indeed afford product 2a, but only in 36% yield (versus 92% for standard conditions), thus proving the Te-cat7 hypothesis to be wrong. Another concurring indication is that we had to resort to Pd catalysis to actually make Te-cat7, as the direct O2-mediated dehydrogenative route from Te-cat5 and add1 proved unsuccessful (see the Supporting Information). It is nevertheless interesting to note that the N–H functional group within the phenotellurazine scaffold is not a strict requirement for catalytic activity. This observation was further confirmed when we engaged control catalyst Te-cat8, produced from the dehydrogenative phenochalcogenazination of 3,4,5-trimethylphenol with Te-cat4,[53] which showed much higher catalytic activity (2a, 75%). Impressively, simple N-phenylphenotellurazine Te-cat9 afforded even higher catalytic activity (2a, 85%). In this case, however, the addition of 1 equiv of 4-Ph-phenol add1 decreased the yield of 2a to 45%. Next, we verified that some trace transition metals arising from the synthesis of the successful catalyst Te-cat5 were not responsible for the observed catalytic activity, by engaging its precursor I-cat10 (Scheme ). While the synthesis of Te-cat5 from I-cat10 is a transition-metal-free process, the synthesis of the precursor, I-cat10, involves both copper and palladium salts in the two previous steps.[53] As expected, I-cat10 did not show any conversion (Scheme ), highlighting the crucial importance of the tellurium atom for catalytic activity. Finally, delayed substrate addition experiments showed a severe decrease in product yield when Te-cat5 was prestirred beforehand under a O2 atmosphere at 130 °C (eq 4, Scheme ). A 3 h pretreatment reduced the yield of product 2a from 92% under the standard conditions to a 56% isolated yield, even though the O2 atmosphere was renewed at the moment of substrate addition. A 6 h pretreatment further reduced it to 46% isolated yield. Therefore, it can be concluded that the true active catalyst is closer in structure to Te-cat5 than to any of its potential decomposition byproducts, even if some of these may also possess some background catalytic activity, as indicated by the persisting yield of 2a after prolonged pretreatment (18 h, 45% isolated).

On the basis of these elements, we propose the mechanism shown in Scheme . We already showed in a previous study that molecular oxygen can directly oxidize phenotellurazine to the persistent Te(III) radical cation, which can readily be characterized by EPR spectroscopy.[26] This process might be facilitated by the coordinative affinity of Te(II) for electronegative heteroatoms, a well-documented process in the literature.[2−24] The resulting Te(III) radical cation intermediate Te-cat5 would then oxidize the indole substrate 1 through a possible Te–N interaction, also a well-documented type of interaction,[54] inverting its philicity from a nucleophile to an electrophile (radical cation intermediate I). This would in turn react with another not yet oxidized 1 equiv of nucleophilic indole 1 to form an intermolecular C3–C2 bond (intermediate II), presumably by a classical homolytic aromatic substitution pathway (HAS).

Scheme 4

Proposed Mechanism

Because 2-phenylindol-3-one 5 was not found to be a productive intermediate (Schemes and 3), it is therefore reasonable to assume that the carbonyl group is obtained upon further oxidation after C3–C2 bond formation (Scheme ), from intermediate III. As to the role of the phenol additive (add1), it could be that of a Brønsted acid–base relay between O2 reduction intermediates, which are relatively basic, and proton scavenging, for example from intermediate II toward intermediate III. The fact that the structure and acidity of the phenol additive (Table ) are important for the reaction outcome is in line with this Brønsted relay hypothesis. Also in support of this Brønsted relay hypothesis, the second best phenol additive (add3, Table , entry 14) has a similar overall acidity in comparison to add1 in spite of a very different structure (p-CF3 versus p-phenyl, respectively). The phenol additive may have other roles as well, such as that of a secondary redox relay catalyst. Indeed, the phenoxyl radical arising from phenolate oxidation would have some persistency due to spin delocalization. Finally, it may also have a redox-stabilizing role on the Te catalyst, thus maintaining its integrity during the reaction. Phenols are indeed known for their antioxidant properties,[55,56] and Te-cat5 proved to be notably more robust in the presence of 4-Ph-phenol (add1) than in its absence (Scheme , eq 5).

In conclusion, we have reported here the unprecedented Te-catalyzed dehydrogenative C3–C2 dimerization of indoles. This represents a rare[26,27] redox-active phenotellurazine catalyst enabled method, and the first that does not involve a phenothiazine substrate. The most important features of this reaction are its O2-mediated character and the absence of a background reaction when the Te(II) catalyst is omitted. Furthermore, impressively, the reaction is still operational with only 1 mol % of Te catalyst loading. Finally, we have shown that electron-donating substituents (R = OMe) significantly improve the Te(II) catalytic activity in the context of a dehydrogenative coupling reaction. These findings will therefore have a considerable impact on the fields of cross-dehydrogenative couplings and tellurium redox catalysis, as well as on related pnictogen redox[57,58] and other unusual element redox catalyzed systems.[59,60]