Pd ‐ Catalyzed Intermolecular Dehydrogenative Heck Reactions of Five ‐ Membered Heteroarenes

: The Pd ‐ mediated cross ‐ coupling of (hetero)arenes with alkenes may be an effective method for the formation of a C–C bond from two C–H bonds. Discovered by Fujiwara and co ‐ workers in 1967, this reaction led to a number of reports that we firstly highlighted in 2011 (review with references till June 2010) and for which, we retained the name “dehydrogenative Heck reaction”. The topic, especially the reactions of five ‐ membered heteroarenes, has been the subject of intensive research over the last ten years. The present review is limited to these dehydrogenative Heck reactions published since 2010, underlining the progress of the procedures.


Introduction
Palladium-catalyzed C-H functionalization has been at the forefront of organic synthesis over the last half-century [1][2][3][4]. Through the multitude of reactions, the coupling leading to a C-C bond from two C-H bonds is an atom-economical process because any prefunctionalization is required. In the late sixties, Fujiwara's team disclosed the Pd II -mediated synthesis of stilbene from the cross-coupling of styrene with benzene [5,6]. This synthesis contrasts with the usual Pd-catalyzed method independently discovered by the teams of Mizoroki [7], Julia [8], and Heck [9] (The Mizoroki publication, submitted on 20, October, 1970, was cited by Heck's team which submitted their report on 13, January, 1972. On 12, January, 1971, Julia's team deposited a "Pli cacheté", i.e., a sealed envelope, to the Société Chimique de France, which was open on 5, May, 1973.), which requires phenyl halide instead of benzene. Although the above Fujiwara cross-coupling initially occurred with a very low palladium turnover [6], such dehydrogenation reactions, which could be named "Dehydrogenative Heck Reactions" (DHRs), have been the subject of intensive research over the last half century leading to efficient catalytic procedures. We previously reviewed the corresponding literature with references until sep 2010 [10]. Through the DHRs, those of the five-membered heteroarenes have especially retained our attention, and are the subject of the present review which is limited to reports of the last ten years (For more general reviews containing some examples, see [11][12][13][14][15].) The reactions of fused bicyclic heteroarenes that do not involve the C-H bond of the small hetero ring (See examples in [16][17][18]) will be discarded. For convenience, the framework of the text depends on the nature of the aromatic substrate.
The simplified mechanism of DHRs usually admitted implicates the activation of a C-H bond of the arene by Pd II species leading to an arylpalladium intermediate which adds to the alkene to provide 1A or 1A' (Scheme 1, arene activation). Subsequent β-H elimination delivers the cross-coupling product and Pd 0 . The active Pd II species are regenerated with an oxidant. Another mechanism involving the coordination of the alkene to afford a η 2 -palladium complex susceptible to nucleophilic attack by the arene has been hypothesized (Scheme 1 alkene activation) [19]. Such a reaction would also afford intermediate 1A or 1A'. Both catalytic cycles describe dehydrogenative cross-coupling reactions, but a reaction via the alkene complex cannot be named DHR. Broggini's team proposed the term "alkene activation" for the corresponding catalytic cycle [19]. In fact, coordination of the alkene to palladium decreased the electron density of the double bond, leading to its activation towards nucleophilic attack, that is, a Friedel-Crafts reaction. Electrospray ionization mass spectrometry (ESI-MS) [20][21][22][23][24][25][26] and nuclear magnetic resosnance (NMR) [27] studies would favor the arene activation process for the majority of the dehydrogenative cross-coupling reactions.
The site selectivity of the heteroarene is governed by the "innate" or "guided" C-H bond activation [13]. The former relies on the electronic properties of the heterocycle induced by the heteroatom [28], whereas the latter is relevant to either a directing group or a specific additive. Scheme 1. Catalytic cycles of the dehydrogenative cross-coupling reactions.
A Catellani-type reaction [50][51][52] was related by Dong's team, leading to both 5-ethylenation and 4-arylation of 1-substituted thiophenes (Equation (18)) [53]. Sequential stepwise reactions were hypothesized by the authors. That led us to consider the steps of Scheme 3. Insertion of amide-based norbornene NB into the C-Pd bond of heteroarylpalladium intermediate 3A affords 3B. Use of the weak and π-acidic ligand AsPh3 would prevent formation of 3C by intramolecular coordination, and allows access to palladacycle 3D. Reaction of 3D with ArI provides 3E via a putative Pd IV intermediate [50,54]. β-C elimination liberates NB, giving 3F, this step being favored by the size of NB. 3F undergoes reaction with the alkene to deliver the difunctionalized compound.

Scheme 3. Plausible pathways of vicinal difunctionalizations.
Subsequent studies of the kinetic profile indicated "that the difunctionalization product was formed immediately at the beginning of the reaction and there was no accumulation of the C4-arylation intermediate during the course of the reaction" [53]. This "coupled" difunctionalization which also occurs for the Catellani diarylation of six-membered aromatic substrates [50,54], led us to suggest an alternative pathway. Exchange of ligand produces 3G from 3E, which undergoes β-C elimination leading to 3H. Subsequent intramolecular insertion of the alkene followed by reductive elimination affords the compound. According to this pathway, the coordination of the alkene prior to the β-C elimination precludes the release of the C4-arylation product.
2-Thienyl-N-tosylaniline undergoes regioselective DHR with ethyl acrylate leading to the C3 substituted thiophene and a fused heterocycle due, according to Youn and co-workers, to subsequent cyclization of the C3 substituted thiophene (Equation (19)) [55]. We will come back in Sub-chapter 4.1. to the mechanism and catalytic cycle of substrates with a directing group (Scheme 9).
Yu's team disclosed the promotion of DHR's by monoprotected amino acids [56,57]. Such an additive led however to modest yields for the cross-coupling of benzothiophenes with ethyl acrylate, as reported by Huang, Lin, and co-workers, using Pd(OAc)2 combined with 11molybdovanadophosphoric acid and N-acetylglycine under oxygen [58]. The regioselectivity depended on the substituents of the six-membered ring (Equation (24)). The procedure was nevertheless effective for benzofurans (see Equation (32)). (24) In contrast to the above example, the presence of N-acetylvaline allowed high yields of the Pd(OAc)2-catalyzed DHR of 1-(benzo[b]thiophen-2-yl)ethanone O-methyl oxime with acrylates (except butyl acrylate) using AgOCOCF3/air as oxidant (Equation (25)) [59]. Under these conditions, the crosscoupling with styrenes led to domino reactions giving annulation products, in improved yields with pyridine instead of N-acetylvaline (Equation (26)). According to Xia, Ji, and co-workers, the DHR product 4A undergoes reaction with the simultaneously formed Pd 0 , giving 4B via oxidative addition to N-O bond [60,61] (Scheme 4). Subsequent amino-Heck reaction provides 4C, which endures β-H elimination, affording the annulation product and MeOPdH. Under the reaction conditions, the latter leads to Pd II via, possibly, Pd 0 .

Thieno[3,2-b]thiophene
The best results of the Pd(OAc)2-catalyzed DHR of thienothiophene with styrenes were obtained with AgOAc in CF3CH2CO2H, while the coupling with acrylates and N,N-dimethylacrylamide was best carried out with AgOCOCF3 in EtCO2H (Equation (27)

Sub-Conclusion
The 2-position of thiophenes and benzothiophenes is the most reactive towards the DHR. Regiodivergent reaction of thiophenes occurs when the 2-and 5-positions are already substituted or in the presence of a directing substituent. Regioselectivity may also depend on the ligand.
In contrast to thiophenes and benzothiophenes, thienothiophene and thienofuran react in the 3position.

Sub-Conclusion
The exchange of the S atom for the O atom has no effect on the DHR regioselectivity of the fivemembered heterocycle, the 2-position of furans and benzofurans being the most reactive in the absence of particular substituents.
The regioselectivity of the alkenylation of 4-aryl-1H-pyrrole-3-carboxylates is highly dependent on the solvent (Scheme 5) [69,70]. In benzene, assisted chelation by the carboxylate would favor the formation of 6A (Scheme 6). DMSO which is a strong coordinating solvent would add to Pd(OAc)2 to afford 6B [71]. That overrides the chelation with the carboxylate and promotes the palladation at C5 position giving 6C as intermediate. Subsequent reaction of 6A and 6C with the alkenylating agent affords the two DHR products (Scheme 1).
The DHR with Pd(OAc)2/Cu(OAc)2 in DMSO of free (NH)-pyrrole occurred in C2 and C5 positions, while N-tosylpyrrole afforded the mono-adduct (Equation (33)) [72]. In contrast, free (NH)-pyrrole with an electron-withdrawing C2 substituent led to alkenylation in C4 position (Equation (34)). The same regioselectivity was observed from N-methyl, N-benzyl, and N-tosylpyrroles, the latter reacting under different conditions (Equation (34)). Subjected the cross-coupling products to a second DHR led to C5 substitution with, in the case of the tosylpyrrole, a domino reaction leading to an indole (Scheme 7). In contrast to the above, the DHR of N-(p-NCC6H4CO)-2-acylpyrroles mainly occurred at the C5 position (Equation (35)) [73]. According to Sun and co-workers, the carbonyl of the protecting group could have a directing effect but "should not be dominant for this C5-alkenylation". We suspect a C5 selectivity favored by coordination of the nitrile moiety to Pd-intermediates.

Indoles
Our 2011 review [10] contains a great number of DHRs of indoles. The topic continued to retain the attention of researchers resulting into papers with different N-protecting groups and experimental conditions. Reviews not limited to palladium catalysis and DHR have been published [77][78][79]. As pointed in the introduction chapter, we will limit the examples to the DHRs of the heterocycle (For reactions at C4, C5, C6, and C7 positions of the benzenoid ring, see [18]).
Before presenting the results of the ten last years, it is necessary to remember the important communication of Gaunt and co-workers who, in 2005, disclosed the selective 2-or 3-alkenylation of free (NH) indoles, depending on the experimental conditions, especially the solvent (Scheme 10) [80]. The optimum conditions also allowed the selective 3-alkenylation of N-methyl indole but were ineffective for the C2. According to the authors, the reaction starts with palladation at C3. Acidic conditions would allow migration of the C3-Pd bond leading to C2-Pd bond, hence the observed results. This chapter is divided in three parts corresponding to reactions in C3, C2, or both C2 and C3 positions.
t-Butylhydroperoxide associated to catalytic amounts of a heterogeneous copper catalyst has been proposed as reoxidant for the Pd(OAc)2-catalyzed DHR of free (NH) and N-methylindoles in DMSO (Equation (49)) [86].
Bolm's team introduced the mechanochemical activation for solventless Rh-catalyzed DHRs [87]. Su and co-workers used this high-speed ball-milling process for the fast Pd-catalyzed C3 alkenylation of indoles with MnO2 as oxidant (Equation (50)) [25]. (50) On the basis of their previous studies [88,89], Sigman and co-workers used chiral pyridine oxazoline ligands to perform the enantioselective DHR of indoles with prochiral unsaturated alcohols [90]. Yields and enantioselectivities greatly depended on the substitution of the ligand. Ligands with a naphthyl group such as L1* and L2* were particularly effective providing aldehydes and ketones in good yields with e.e. up to 92% (Equation (51)).
The annulation depicted in Equation (61) [102] differs from those of Equations (56) and (57). Indeed, the reaction of the exocyclic C=C bond with the unsaturated reagent was one of the key intermediates. Plausible intermediates were proposed by Verma's team. According to Scheme 12, the cyclization step is Pd-catalyzed. Similar annulations were reported [85,95,103], but, differing from the DHRs, they will not be more commented. Increasing the indole/maleimide ratio provided indolopyrrolocarbazoles (Equation (62)) under experimental conditions previously used in Equation (47) [84]. The mechanism proposed by Zhao and co-workers involves the DHR leading to C3 alkenylation product 13A (Scheme 13). Activation of the C2-H bond of 13A affords 13B which reacts with indole to afford 13C. Subsequent thermal electrocyclization gives the product. In agreement with this proposal, the carbazole was also obtained from the reaction of C3 alkenylation product with maleimide. The cross-coupling of N-methyl and N-benzylindoles with methyl acrylate afforded a mixture of two isomeric carbazoles (Scheme 14, path a) [104]. The process became selective with others Nsubstituents, (path b) arising from a double DHR reaction followed, according Equation (63), thermal electrocyclization. The annulation was not observed from free (NH) indole (path c).

Scheme 14.
Reactivity depending on the N substituent.
Using PdCl2/K2CrO4/H2O/O2 in DMF, C1-substituted indolizines underwent cross-coupling with α,β-unsaturated carboxylic acids to produce C3 acylated compounds (Equation (69)) [108]. Traces of annulation products were identified; they were selectively formed under modified experimental conditions (Equation (70)). According to Zhang and co-workers, the first steps of the acylation are those of the DHR leading to 15A (Scheme 15). Oxidation of the activated tertiary hydrogen bond of 15A with K2CrO4 provides 15B. Subsequent hydrolysis of the C-Pd bond affords 15C that is oxidized with K2CrO4 giving the acylated compound. The annulation product would arise from 15A via β-H elimination leading to 15D. Subsequent oxidation with BQ/K2CrO4 affords 15E. Activation of C-H bond in 5 position gives the palladacycle 15F, which delivers the product via reductive elimination and decarboxylation.

Sub-Conclusion
The regioselectivity of the DHR of pyrroles depends on solvent, N protective group, and, if present, electron-withdrawing substituents. Indoles react in 3 position in the absence of directing substituent.
Under DHR conditions, the C3 position of 7-azaindoles and indolizines is the more reactive. It seems however necessary to note that the C3 position relative to the N atom is β for the former and α for the latter.
Maiti's team studied the olefination of 4-methylthiazole [112]. With Pd(OAc)2/AgOAc in trifluoroethanol containing catalytic amounts of N-acetylglycine, the reaction with methyl acrylate led to a complex mixture (Equation (76)). The authors examined the DHR in the presence of tricoordinating templates to direct the metal catalyst to the desired C5−H bond. They founded an effective recyclable template for the purpose (Equation (76)). Interestingly, selective C5-olefination also occurred from 4chlorothiazole, 2-methylthiazole, and unsubstituted thiazole (Equation (77)).
The key intermediates proposed by Maiti's team for the template reactions are depicted in Scheme 16 [112]. Exchange between acetonitrile and thiazole provides 16A. Activation of C5-H bond due to interaction with the aromatic nitrile allowed regioselective palladation leading to 16B, while the C2-H bond is preserved. Subsequent reaction with the olefin gives 16C, which liberates the product and regenerates the template. Scheme 16. Template-mediated regioselectivity.
An efficient and highly regioselective Pd II -catalyzed C-3 alkenylation of imidazo[1,2-a]pyridines with acrylates and styrenes was developed by Cao's team using oxygen as the main oxidant [131]. Catalytic amounts of Ag2CO3 and stoichiometric quantities of both AcOH and Ac2O were required for optimum results. Acrylates and acrylonitrile tended to form β-products, while styrenes tended to form α-products (Scheme 19). The authors proposed two different catalytic cycles like those of Scheme 1, "arene activation" for the β-products and "alkene activation" for the α-products. Using oxygen as the sole oxidant, Pd(OAc)2-catalyzed 2-phenylimidazo[1,2-a]pyridine with styrene in DMSO provided the C3 branched adduct in low yield [132]. Changing the solvent for DMA with, especially Bu4NBr as additive, led to a fair yield (Equation (100)). Other substituted substrates and styrenes were reacted under the same conditions. The branched product was also obtained with 1octene. Hajra and co-workers retained the mechanism of Scheme 1, alkene activation.  (102) and (103)) under experimental conditions used for DHRs of indolizines (Equation (67)), leading to fluorescent molecules with potential role in biological imaging [106].

Sydnones
At room temperature, using AgOAc as the oxidant, 3-arylsydnones reacted with acrylates, acrylamide, allyl acetate, and styrenes leading to cross-coupling products in fair to good yields, while the 4-vinyl adduct arose with vinyl acetate (Scheme 20) [134]. Yang and Kuang did no comment on the cleavage mechanism of the C-OAc bond. We propose two plausible pathways based on Scheme 1 (Scheme 21). The DHR mechanism provides 21B and/or 21C through heteroarylpalladium complex 21A. β-H elimination from 21B would afford the cross-coupling adduct while β-HOAc elimination from 21C would lead to the vinyl adduct 4 . Another possibility is the formation of alkene complex 21D which underwent nucleophilic attack giving 21B and/or 21C. The addition of Heck reagent to vinyl acetate may add to α and β positions, with slight preference for the β [135], leading to mixture of products in a ratio depending on the experimental conditions [36,116,136], while nucleophilic attack on the alkene complex could prefer the α position [119]. As the 3-arylsydnone/vinyl acetate cross-coupling product was not observed (Scheme 20), the 4-vinyl adduct would be rather formed via alkene activation.

1,2,3-Triazole N-Oxides
The Pd(OAc)2/Ag2CO3/pyridine system in t-BuOH/dioxane led to efficient DHRs of 2-aryl-2H-1,2,3triazole 1-oxides with acrylates and styrene (Scheme 22) [138]. The presence of pyridine and the 1:4 ratio of t-BuOH/dioxane were essential for yields. The reaction with vinyl acetate provided a mixture of the cross-coupling and corresponding deacetoxy products, while that of 1-octene led to three isomers. Kuang and co-workers proposed that the three isomers were issued from the DHR (Scheme 1, arene activation). The formation of the branched product led us to suspect the possibility of a nucleophilic attack on the octene-coordinated palladium complex (Scheme 1, alkene activation). The mixture obtained from vinyl acetate does not allow the prioritization of a catalytic cycle (see Scheme 21 and discussion of Sub-chapter 9.2).

Scheme 22.
Products depending on the olefinating agents.

Conclusion and Remarks
Undeniably, the development of DHRs of five-membered heteroarenes has greatly progressed in the last ten years. Their exclusive properties have attracted the attention of scientists all over the world, leading to processes that would be valuable tools for academic and industrial applications.
A number of Pd(OAc)2-catalyzed methods seem almost identical. In fact, the oxidant, required for the turnover, and the solvent often varied with the substrate. Additives, which may also be ligands, have, in most cases, a decisive role on the efficiency. As the range of oxidants, solvent mixtures, and additives is large, it is not obvious to choose the right combination for a new reaction. Complexity increases with N-heterocycles because their reactivity may depend on the N-substitution. Even if the wide published procedures give valuable indications, the chemist should try and see to reach fine results from a new DHR.
Compared to traditional Heck reactions, the DHR's have a "green" character, because prefunctionalization of the starting aromatic partner is avoided, improving step economy, and also the high atom efficiency gained by formal removal of only the molecular hydrogen. Nevertheless, the DHR's are usually performed with an excess of heteroarene or alkenating reagent, and only a few methods use solely oxygen to regenerate the Pd active species. Consequently, expensive silver salts are often required as oxidants. Leading to wastes, these conditions decrease the "green" aspect of the procedures and their compatibility with the atom economy principle [139].
In our opinion, five publications are particularly important [49,69,70,112,126] and complete the older Gaunt report [80]. One of them revealed a regioselectivity depending on the metal catalyst [49], while three others related regioselectivity mainly depending on solvent [69,70] or ligand [126]. The fifth publication disclosed a template directing the regioselectivity [112]. Another important report related regioselectivity depending on the substitution [72]. Although these reports concern DHRs of only a few substrates, they represent breakthroughs for researchers involved in the selective synthesis of regioisomers from the same starting material. Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest