C−H Carboxylation of Aromatic Compounds through CO2 Fixation

Abstract Carbon dioxide (CO2) represents the most abundant and accessible carbon source on Earth. Thus the ability to transform CO2 into valuable commodity chemicals through the construction of C−C bonds is an invaluable strategy. Carboxylic acids and derivatives, the main products obtained by carboxylation of carbon nucleophiles by reaction of CO2, have wide application in pharmaceuticals and advanced materials. Among the variety of carboxylation methods currently available, the direct carboxylation of C−H bonds with CO2 has attracted much attention owing to advantages from a step‐ and atom‐economical point of view. In particular, the prevalence of (hetero)aromatic carboxylic acids and derivatives among biologically active compounds has led to significant interest in the development of methods for their direct carboxylation from CO2. Herein, the latest achievements in the area of direct C−H carboxylation of (hetero)aromatic compounds with CO2 will be discussed.


Introduction
Carbon dioxide (CO 2 )r epresents the most abundantc arbon source in the Earth'sa tmosphere.I ti sg enerally considered a green carbon source as it is non-toxic, low-cost, and renewable. Furthermore,C O 2 is proposed to be the major cause of climate change, because of its greenhouse properties and levels of CO 2 in the atmosphere are steadily increasing. Thus, it is highly attractive to develop methodologies fort he transformation of CO 2 into valuable commodity chemicals. [1] Particularly, (hetero)aromatic carboxylic acids and derivatives are important motifs among natural productsa nd biologically active compounds. [2] The catalytic coupling of CO 2 with energy-rich substrates, such as epoxides, aziridines, anda mines has previously been widely explored for the construction of CÀOo rC ÀN bonds. [3] However,o wing to the thermodynamic stabilitya nd high oxidations tate of CO 2 ,i ts coupling with aromatic compounds its traditionally achieved by the use of highly reactive organolithium or Grignardr eagents (Scheme 1a and b). [4] These methodss uffer drawbacks of low functional-group compatibility thusl imiting their applications. To avoid this issue, many methodologies have been developedo ver the last two decades for the carboxylation of different organometallic reagents. [5][6][7] Pioneering work was reported by Nicholasa nd coworkers on the insertion of CO 2 into SnÀCbonds under Pd catalysis although high pressures of CO 2 were required (Scheme 1c). [5] More general methods involved the use of organoboron and organozinc compounds as nucleophiles (Scheme 1d and e). [6] Additionally,t he groups of Martin,D augulis, and Ts ujih ave demonstrated useful methodologies using readily availablea ryl halidesa ss tarting materials for transitionmetal-catalyzed carboxylation under mild conditions (Scheme 1f). [7] CÀHf unctionalization is one of the most encompassing transformations in organic chemistry as it allows efficient and economical access to myriad knownand unknown molecules. [8] In this regard, am ore sustainable synthesis of (hetero)aromatic carboxylic acids is the direct insertion of CO 2 into aC ÀHb ond of an organic substrate. Direct CÀHc arboxylationm ethods obviate the need for pre-functionalized substrates affording, at Carbon dioxide (CO 2 )r epresents the most abundant and accessible carbon source on Earth. Thus the ability to transform CO 2 into valuable commodity chemicals through the construction of CÀCb onds is an invaluable strategy.C arboxylica cids and derivatives, the main products obtained by carboxylation of carbon nucleophilesb yr eactiono fC O 2 ,h ave wide application in pharmaceuticals and advanced materials. Among the variety of carboxylationm ethods currently available, the direct carbox-ylation of CÀHb onds with CO 2 has attracted much attention owing to advantages from as tep-and atom-economical point of view.I np articular,t he prevalence of (hetero)aromatic carboxylic acids and derivativesa mong biologically active compounds has led to significant interest in the development of methods for their directc arboxylation from CO 2 .H erein, the latest achievements in the area of direct CÀHc arboxylation of (hetero)aromatic compounds with CO 2 will be discussed. least in theory,t he highest possible step-and atom-economy. Herein, we will review the current state-of-the-art of direct CÀHc arboxylation of (hetero)aromatic compounds with CO 2 . This Minireview is organized by the type of reagent used in four different sections:1 )base-mediated carboxylation; 2) Lewis-acid-mediated carboxylation; 3) transition-metal-catalyzed carboxylation;a nd 4) enzymatic carboxylation.W eh ave avoidedaclassification based on the underlying mechanism as often new experiments may revealaproposed mechanism to be incorrect. However,o ne can appreciate three main types of carboxylation mechanisms in the reactions discussed herein, depending on the mode of CÀHc leavage:1 )ane lectrophilic aromatic substitution;2 )a C ÀHd eprotonation by base;a nd 3) aC ÀHo xidativei nsertion. The proposedm echanisms for each transformation and their alignment within the three mechanistic classes will be discussed throughout the Minireview.

Base-mediated carboxylation
The first example of aC ÀHc arboxylation of aromatic compounds was developed by Kolbe and Schmitt during the 1860'sa nd it is stillw idely used in industry today.T he Kolbe-Schmitt reaction is most notable for the synthesis of Aspirin. It is one of the most important and well-known carboxylation reactions,p roviding direct access to salicylic acids through the ortho CÀHc arboxylation of phenoxides with CO 2 . [9] However, this process generallyr equires high CO 2 pressure (20-100 atm; 1atm = 0.101325 MPa) andt emperature (130-280 8C) to achieve good conversion. Additionally,t he preparation and isolation of completely dryp henoxides from the corresponding phenolsi sn ecessary,a st he presence of water inhibitst he Kolbe-Schmitt carboxylation. [10] These drawbacks can preclude use in somel aboratories. It is generally accepted that in the Kolbe-Schmitt reaction, CO 2 could be captured by am etal phenolate through weak coordination between the alkali metal and CO 2 . [11] Water molecules present in the reaction vesselc ould strongly chelate with the alkali metal phenoxides, consequently preventing the initial addition of CO 2 .T he hydrolysis of the alkali metal phenoxides to the phenols may also occur.
In 2016, the Larrosa group developed the first protocol for Kolbe-Schmitt-type carboxylations that occur efficiently under an atmospheric pressureo fC O 2 . [12] The use of NaH as the base avoided the undesired formation of H 2 O, and allowed the reaction to proceed in ao ne-pot process without isolation of the phenoxide precursor.M ore importantly,t he authors reported that using 2,4,6-trimethylphenol (TMP) as ar ecyclable additive significantly increased the carboxylation rate allowing the process to be carriedo ut under 1atm of CO 2 .T he reaction was compatible with both electron-donating and halogen substituents, and simple phenol (Scheme 2). However,s ubstrates containing strongly electron-withdrawing nitro groups were not reactive under these conditions (Scheme 2, 2l). It is noteworthy that this procedure is amenable to scalingu pa nd the additive 2,4,6-trimethylphenol (3)c an be easily recovered after the reaction( Scheme2,2n).
The electrophilica romatic substitution of phenoxide with CO 2 wast he most accepted mechanism for this typeo fr eaction. [11] It is proposed that sodium2 ,4,6-trimethylphenoxide 4, generated in situ from 3 and excess NaH, may play ar ole in aiding CO 2 fixation, although 4 is inert toward ortho-carboxylation itself (Scheme 3).
Besidet he CÀHc arboxylation of phenol substrates, much attention has been given to the base-mediated CÀHc arboxylation of heteroarenes. In this process, heteroarenes bearing sufficiently acidic CÀHb onds can undergo carboxylation. In 2010, the research group of Hu developedam ethodology for the CÀHc arboxylation of benzothioazole substrates under 1.4 atm CO 2 . [13] The group found that the carboxylation proceeds well when using LiOtBu and Cs 2 CO 3 as bases, whereas K 2 CO 3 , NaOMe, NaOH, and KOH were found to be ineffective. Further investigations revealed that some of the heteroaromatic carboxylic acids are unstable in solutiona nd slowly revert to the heteroarene upon loss of CO 2 .F or example, benzothiazole-2carboxylic acid 10 a shows 20 %d ecomposition over 5h.I n light of this, the carboxylated compounds were instead isolated as the corresponding methyl esters. The isolation of carboxylationp roducts as the corresponding esters, usually by reaction with an alkyl halide or trimethylsilyldiazomethane (TMSCHN 2 ), is acommon strategy employed in many of the ex- amplesd iscussed throughout this Minireview.T he carboxylation procedure developed by Hu and co-workersw as applicable to benzothiazoles bearing both electron-withdrawing groups and moderately electron-donatingg roups at C6 (Scheme 4, 10 a-10 c). Furthermore, benzoxazole, oxazole, and 1,3,4-oxadiazole derivativesw ere also found compatible, leading to the corresponding carboxylic acid derivatives in good yields (Scheme 4, 10 d-10 i).
The mechanismi ss omewhat different to the Kolbe-Schmitt reaction; Hu et al. proposed that aCs 2 CO 3 -mediated the deprotonation of CÀHb ond at the C2 position of benzothiazole 11, followed by reaction with CO 2 (Scheme 5). The detection of benzothiazole 11 and 2-benzothiazolecarboxylate 13 by NMR spectroscopy,b ut not the 2benzothiazolyl anion 12,s uggested that the C2-deprotonation is thermodynamically disfavored.
More recently,F enner and Ackermann found that KOtBu enabled the efficient CÀHc arboxylation of heteroarenes with an ample substrate scope such as benzoxazole, benzothiazole, oxazole, and 1,3,4-oxadiazole derivatives (Scheme 6). [14] Compared with the Cs 2 CO 3 -mediated process (Scheme 4), the use of KOtBu allows for carboxylation at ar elativelyl ower reactiont emperature and, more importantly,a to nly 1atm CO 2 pressure. Therefore, this process can be carried out with as imple CO 2 balloon insteado f using specialist high-pressuree quipment. Similart o the Cs 2 CO 3 -mediated process, the authors suggested the reaction proceeded via initial reversible CÀHd eprotonation, followed by subsequent CO 2 insertion.
tert-Butoxide can not only mediate the CÀHc arboxylation of benzothiazole, benzoxazole, oxazole, and 1,3,4-oxadiazole derivatives, but also the direct carboxylation of indoles. In 2012, the Kobayashi group found that LiOtBu allowed the formation of indole-3-carboxylic acid derivatives under ambient CO 2 atmosphere. [15] Interestingly,t he authors found the use of al arge excess of LiOtBu (5 equiv) wase ssentialt op revent competing product decarboxylation and to improve yields and reproducibility.T he process is compatible with both electron-rich and electron-poor substituents (Scheme 7) at various positions, albeit substitution at C2 and C4 led to lower yields. The method was subsequently found to be suitable for pyrrole derivatives. [16] At entativem echanism proposed by the authors involves deprotonation at the most acidic NÀHp roton to anion 17.T his anion can be initially reversibly captured with CO 2 to form Ncentered carboxylate 18, [17] but this compound eventually reacts at the high temperatures of the reactiontof orm the C3carboxylated product 20 (Scheme 8). It is important to note that the free NÀHi se ssential for this carboxylation,a st he more nucleophilic N-methylindoled id not undergo base-mediated carboxylation under the reported conditions. This fact suggested that the reactionm ay proceedt hrough electrophilic aromatic substitution rather than the deprotonation of the CÀ Hb ond at the C3 positionofi ndoles.
ChemSusChem 2017, 10,3317 -3332 www.chemsuschem.org 2017 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim acid (23)c an be obtained by the carbonate promoted carboxylation of 2-furoic acid. This is ar emarkable transformation as it was previously believed that the deprotonation of 2-furoic acid could only be achieved with very strongb ases such as LDA or nBuLi. [19] Since 2-furoica cid can be readily made from lignocellulose, this strategy provides af ast and direct route to 23.T he reactionc an be conducted under af low of CO 2 to give good yield, or under 8atm of CO 2 for maximum efficiency (Table 1).
Encouraged by the resultso nt he carboxylationo f2 -furoic acid, Kanan and co-workers speculated whether benzoates (substantially weaker acids) would also be suitable substrates. Satisfyingly,w hen heating the cesium benzoate in the presence of 0.55 equivalents of Cs 2 CO 3 to 320 8Cu nder 8atm CO 2 pressure, ac ombined yield of 66 %f or am ixture of phthalates and tri-and tetracarboxylates was obtained (Scheme 9a). This contrasts with an earlier study that had reported that the carboxylation of cesium benzoate with Cs 2 CO 3 required much higherC O 2 pressure (ca. 400 atm) and reactiont emperature (380 8C). [20] The carboxylation of benzene is more challenging than benzoate, owing to the larger entropic penalty and the possible lower solubility of benzene in the molten salt. Kanan and co-workersr eported that carboxylation was possible, in the presence of cesium isobutyrate additive, at 380 8Cu nder 30 atm CO 2 pressure and 40 atm benzene pressure. It is important to note that the cesium isobutyrate molten component is criticalf or the successo ft he reaction, since Cs 2 CO 3 does not melt at 380 8Ci nt he absence of cesium isobutyrate (Scheme 9b). The carboxylation of isobutyrate to dimethyl malonate and decomposition of isobutyrate to formate and acetate also took place under these conditions. This work by Kanan and co-workerss hows that the CO 3 2À -promoted carboxylationo fb enzoate and benzene is possible albeit under extremelyf orcing conditions.
The same authors envisioned that aM 2 CO 3 reversibly deprotonates aC ÀHb ondt og enerateM HCO 3 and ac arbon-centered nucleophile M + C À that could attack CO 2 to form CÀCO 2 M. The decomposition of MHCO 3 resultsi nanet consumption of 0.5 equiv of M 2 CO 3 and CO 2 when per CÀCO 2 Mp roduced (Scheme 10 a). The CÀCO 2 Mc ould be protonated by the treatment of as trong acid such as HCl to form the carboxylic acid and by-product MCl. The metal salt MCl could then be processed by electrodialysis to regenerate M 2 CO 3 and the acid HCl (Scheme 10 b). [21] In this regard,t his whole cycle would effectively transform CÀHi nto CÀCO 2 Hu sing only CO 2 and no other stoichiometric reagents. This protocolw ould be classifiedw ithin the CÀHd eprotonation class by ab ase-type pathway.

Lewis-acid-mediated carboxylation
Friedela nd Crafts found that am inor amount of benzoica cid was obtained when CO 2 was bubbled through am ixture of aluminumc hloride and benzene when heatedt oi ts boiling point along with the generation of as mall amount of hydrogen chloride. [22] They suggested this process may involveani nitial complex between benzene and Al 2 Cl 6 ,w ith subsequent formation of phenyl aluminum dichloride intermediates (PhAl 2 Cl 5 ). The organoaluminum speciest hen reacted with CO 2 to obtain the benzoic acid after aqueous workup (Scheme 11). However,o wing to the low electrophilicity of CO 2 and the side-reactions attributed to the strong Lewis-acidity of aluminum-based compounds, the carboxylic acids are generally obtained in poor yields when using this procedure. [23] Furthermore, secondary reactionp roducts such as benzophenones and diphenylmethanes are formed in significant amounts.
Scheme6.Selectedexamples of KOtBu-mediatedc arboxylation. More recently,O lah and Prakash reportedt hat the use of AlCl 3 /Al can dramatically improve the arene carboxylation yields. [24] In this procedure, the addition of aluminumm etal powder is key,w hich presumably scavenges the HCl liberated in the process, thus shifting the equilibria toward product formation.M oreover,t he AlCl 3 generated in situ upon reaction of Al with HCl furtherp romotes carboxylation.T his direct carboxylation could be carried out at moderate temperatures in good-to-excellent yields. The more active substrates such as mesitylene underwent the carboxylation to the corresponding acid with 80 %y ield even at 20 8C. Conversely, deactivated aromatics such as benzene halides gave relatively low yields and nitrobenzene substrates were not carboxylated (Table 2).
Unfortunately,a mong alkylbenzenes only methylbenzenes are suited toward selective carboxylation as other homologuest end to disproportionate when treated with anhydrous AlCl 3 /Al, decreasing the yield of the desired carboxylic acids( Scheme 12). It is important to note that, unlike the previous report that led to the formation of as ignificant amount of diaryl ketone, [25] this procedure only produced as mall amount of those side products.
Scheme8.Proposed mechanism for carboxylation of indole. To investigate the reaction mechanism, density functional theory (DFT) studies on the activation process werec arried out. The calculations favor ap athway involving the formation of aC O 2 -AlCl 3 complexo ver the previously proposed formation of PhAlCl 2 .Onthe basis of these calculations, anew mech-anistic pathway wasp roposed. Initial activation of CO 2 by two molecules of aluminum chloride afford six-memberedr ing CO 2 -(AlCl 3 ) 2 complex 31.I nt he presence of af urther AlCl 3 molecule, 31 would react with benzene via transition state 33 to give intermediate 34.F inally,r e-aromatization with loss of HCl affords aluminumc arboxylate 35 (Scheme 13). More recently, Munshi and co-workers found that the carboxylation of toluene proceeded more efficiently by holdingA lCl 3 under pressured CO 2 (ca. 70 atm) for 1h prior to the addition of toluene. This manipulation enables the carboxylation to occurw ith weaker Lewis acids, such as SnCl 4 ,M oCl 5 ,a nd TiCl 4 withouta ppreciable loss of the product yield. [26] These observations further support Olah'sm echanism of Lewis-acid activation of CO 2 rathert han the activation of aromatic ring. This mechanism is similart ot hat of the Kolbe-Schmitt reaction proceeding via an electrophilic aromatic substitution pathway.
The research group of Hattori, on the other hand, found that the carboxylation of aromatic compounds with CO 2 can be significantly promoted by the addition of al arge excesso f chlorotrimethylsilane( TMSCl), giving the corresponding carboxylic acids in good-to-excellent yields. It is noteworthy that these reactionc onditions are mild, proceeding under room temperature at 30 atm of CO 2 pressure (Scheme 14). [27] The authors suggested that TMSX (X = Cl or Br) could be activated in the presence of the AlX 3 to form TMSAlX 4 (37). These speciesw ould then silylate the aromatic substrate through an electrophilic aromatic substitution, affording arylsilane 38 and Scheme10. Possible pathways for carbonate-promoted carboxylationo faromatics.
Scheme11. Possible pathwaysfor the carboxylation of benzene promoted by Lewis-acids.  . The authors propose that the silylations tep is an equilibrium process, thus justifying the requiremento falarge excesso fT MSX reagent to ensure the reaction proceeds in high yields. However,t his explanation for the need of al arge excess of TMSX seems unlikely if, as proposed,t he step forming 37 is irreversible as the quantity of 37 would be determined by the concentration of limiting reagent AlX 3 . Subsequently,H attori'sg roup furthere xtended this work showingt hat av ariety of trialkyl-or triaryl-silyl chlorides efficiently promote the AlBr 3 -mediated carboxylationo fa romatic compounds. [28] Triphenyl silyl chloride performed best among the silyl chlorides tested and, especially when polycyclica renes were used, almost quantitative yield was obtained with high regioselectivity.O nt he basis of the recovery of Ph 3 SiCl at the end of the reaction, the authorss uggested that the silyl halides may act as catalysts in the carboxylation, although they are required in excess. In ar evision to their previous proposal, the authors suggested that silyl chlorides promote carboxyla-tion by reactingw ith CO 2 in cooperationw ith AlX 3 to give haloformate-like active species 42,w hich could react with the arene to give as ilyl ester 43 and the superacid HAlX 4 .T he decomposition of 43 in the presence of HAlX 4 leads to the carboxylic acid and regenerates the trialkylsilyl chloride R 3 SiX. Finally,t he carboxylic acid reacts with AlX 3 to afford an aluminum carboxylate 40 at the expense of an equimolar amount of AlX 3 (Scheme 16). This mechanistic proposal would fall within the electrophilicaromatic substitution-type class.
Scheme16. Revised mechanism for R 3 SiX-promoted carboxylation. Hattori et al. suggested that zwitterionic species 48 was formed initially through electrophilica ddition of Me 2 AlCl to 1-methylindole, followed by deprotonation to ate complex 50, and carboxylation( Scheme 18). The equilibrium between 44, 48,a nd 50 is proposed to lie significantly towardt he starting materials, the equilibrium being displaced by the carboxylation under high pressureo fC O 2 .N otably,H Cl may be eliminated from zwitterionic species 48 to form indolylaluminum species 51 instead of ate complex 50. [30] Inglesona nd co-workers reportedt hat treatment of 1-methylindole 44 with AlCl 3 at 80 8Ca fforded zwitterionic species similar to adduct 48,a se videnced by X-ray analysis. [31] The same treatment of thiophenes and benzothiophenes with EtAlCl 2 under pressurized CO 2 gave the corresponding carboxylic acids in up to 90 %yield. [32] Also, fused-ringaromatic compounds, such as naphthalene, anthracene, and phenanthrene, could undergo Lewis-acid-mediated carboxylation under CO 2 pressure (ca. 30 atm), giving 1-naphthoica cid, 9-anthracenecarboxylic acid, and 9-phenanthrenecarboxylic acid, respectively. [33] However,p lacing an electron-withdrawing group on the fused-ringa romatic compounds completely hin-dered carboxylation. Ap lausible mechanism may involve the attack of aL ewis-acid-activated CO 2 molecule on the aromatic ring to form an arenium ion, followed by re-aromatization to give ac arboxylic acid after aqueous workup.
Salicylic acid derivatives have long been synthesized using the Kolbe-Schmitt reaction mediatedb yabase (see Section2). However, Iijima and Yamaguchi found that the carboxylation of phenol with CO 2 (ca. 80 atm) could also take place in the presence of aL ewis acid at moderate temperatures. [34] Among the Lewis acids investigated, AlBr 3 was found to be the most efficient, leading to the salicylic acids in approximately 70 % yields. The proposed reaction mechanismi nvolved the formation of phenoxyaluminum dibromide 54 from phenol in the presence of AlBr 3 .T he intermediate 54 could react with CO 2 to produce the CO 2 complex, which after ortho-carboxylation and re-aromatization afforded the aluminums alt of salicylate 57. Simple workup of 57 furnished the salicylic acid 58 (Scheme 19). This proposed mechanism is similart ot he Kolbe-Schmitt reaction, proceeding via an electrophilic aromatic substitution.
It may be expected that bases would be incompatiblew ith Lewis-acid-mediatedc arboxylations. However,T anaka, Hattori, and co-workers developed am ethod for CÀHc arboxylation with CO 2 mediated by ac ombination of EtAlCl 2 and 2,6-dibromopyridinea saweak base. [35] Thism ethod allows indole, thiophene, and furan derivatives to be carboxylated to the correspondingc arboxylic acids (Scheme 20). It is important to note that the protocol also enables av ariety of a-arylalkenes and trialkyl-substitutedalkenes to undergo carboxylation Scheme17. Lewis-acid-mediated carboxylation of indoles and pyrroles.
Scheme18. Proposed mechanism for the carboxylation of indoles.
Scheme19. Proposed mechanism for the carboxylation of phenol, mediated by aL ewis acid.
The authors suggested that zwitterion 62 could be generated via electrophilica ddition of startingm aterial 61 to EtAlCl 2 . Then, 2,6-dibromopyridine would abstract ap roton to afford ate complex 63.C arboxylation of the ate complex 63 affords the aluminumc arboxylate 64.O wing to the far greater acidity of the conjugate acid of the pyridine base (pK a %À2) compared with the carboxylic acid (pK a % 5), the carboxylic acid 65 could be liberated from the aluminum carboxylate 64 (Scheme 21). Apparently,t he use of 2,6-dibromopyridine base favors the dissociation of the aluminum-pyridine salt owing to its sterically bulky naturea nd lower basicity.

Au-catalyzed CÀHc arboxylation
During recent decades, transition-metal catalysis has set the stage for efficient direct CÀCb ond-forming reactions to proceed under exceedingly mild reaction conditions. They have been extended to include site-selective directc arboxylation of CÀHb onds. Our group has shown that the gold complex LAuCl was able to cleave the most electron-deficient CÀH bond of electron deficient (hetero)aromatics to provide the correspondingaryl-Au I species. This allowed for the first general protocol for direct CÀHa ctivation of av ariety of relatively acidic (hetero)arenes with Au I complexes via ac oncertedm etalationd eprotonation. [36] Av ariety of ligands (L) such as alkyl and aryl phosphines, phosphites, and N-heterocyclic carbenes (NHCs) were found to be compatible with this procedure. Simultaneously,t he research group of Nolan showedt hat a monomeric, dicoordinate, linear Au I /NHC hydroxide complex was capable of inducing CÀHa ctivation on (hetero)arenes. [37] Carrying on from this work, Nolan and co-workers reported that the N-heterocyclic carbene gold complex enabled CÀH carboxylation of (hetero)arenes with CO 2 at ambient temperature in the presence of as toichiometric amount of KOH base and only 1.5 atm of CO 2 pressure. [38] The IPr (1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene) gold complex (IPr)AuOH was found to be the most efficient of all examined, leadingt ot he carboxylated products in good-to-high yields. This methodology allowed directC ÀHc arboxylationo fc arbocyclesa nd heterocycles including oxazoles, thiazoles, isoxazoles,a nd some aromatic heterocycles containing multiple heteroatoms (Scheme 22, 66 a-e). The CÀHactivation is highly regioselective at the most acidic CÀHb ond. Additionally,t his transformation is efficient for polyfluoro-and polychloro-substituted benzene derivatives (Scheme 22, 68 a-c).
Stoichiometric reactions were carriedo ut to explore the mechanism of this transformation by the Nolan group. By simply mixing gold complex (IPr)AuOH 69 with oxazole (70), 93 %o fg old-oxazolespecies 71 waso btained. The insertion of CO 2 (5 atm, À78 8C) into the CÀAu bond by nucleophilic addition of oxazole to the electrophilic CO 2 carbon atom produced the carboxylate complex 72 in 86 %y ield. Finally,t he metathesis of carboxylate complex 72 with KOH regenerated the (IPr)AuOH catalyst and simultaneously released the potassium carboxylate 73 (Scheme 23). This methodi sr elated to the Cs 2 CO 3 -mediated carboxylations, proceeding by CÀHd eprotonation. Contrarily to this proposal,A hlquist, Wendt, and coworkers showedt hat the reactivity of (NHC)Au I Ar complexes with strongly electrophilic substrates such as methyl triflate or Scheme21. Proposed mechanism for EtAlCl 2 /2,6-dibromopyridine-promoted carboxylation.
Scheme22. Selectede xamples of (NHC)AuOH-catalyzedd irect CÀHc arboxylation. methyl iodide to form toluene, biphenyl, and ethane were most likely to occur througha no xidative mechanism. Based on experimentala nd computational results, Ahlquist, Wendt, et al. argued that the AuÀC s bond of (NHC)Au I Ar complexes exhibit poor nucleophilicity and suggested that they would be unlikely to react with CO 2 . [39] 4.2 Cu-catalyzedC ÀHc arboxylation Nolan'sg roup later improvedu pon this methodology by the use of ac heaper copper-based (IPr)CuOH catalysti nstead of ag old complex. [40] This methodology allows heterocycles including oxazoles, thiazoles, and polyfluorobenzenes to undergo CÀHc arboxylation with CO 2 at the most acidic position (Scheme 24). This copper-baseds ystem permits a significant range of NÀHc arboxylation (pK a < 27.7) such as imidazole, indole, and pyrazoled erivatives, which were transformedc leanly and quantitatively to the correspondingm ethyl esters. Preliminary mechanistic studies suggest am echanism analogous to that proposed when using (IPr)AuOHc atalyst (Scheme 23).
Scheme25. Selectede xamples of (IPr)CuCl-catalyzed direct CÀHc arboxylation. (IPr)CuCl-catalyzed CÀHc arboxylation. [41] The authors showed that complex 78 could be obtainedf rom 77 a by exposing it to 1atm of CO 2 pressure at room temperature. Ap roposed catalytic cycle is presented in Scheme 26. Initially,l igand exchange between (IPr)CuCl (75)a nd KOtBu affords (IPr)Cu(OtBu) (76), which deprotonates the heterocycle to form 77 a.CO 2 insertion of 77 a takes place to afford 78,w hich subsequently reacts with KOtBu to regenerate catalyst( IPr)Cu(OtBu)(76), and pro-duces the potassium carboxylate 79.A riafard, On the basis of DFT studies Yates and co-workerss uggested an alternative mechanism in which carbene intermediate 77 b was generated in the catalytic cycle, insteado f77 a. [44] The authors indicated that 77 b is more reactive towardC O 2 insertion reaction through nucleophilic attack. Similarly to the Au-catalyzed direct carboxylation, this method could also be regarded as a CÀHd eprotonation process.
More recently,H ou and co-workersd eveloped am ethod for the CÀHc arboxylation of various aromatic compounds with CO 2 . [46] The carboxylation products were afforded by deprotonative alumination using an iBu 3 Al(TMP)Li aluminate base [45] followed by the carboxylation of the resulting arylaluminum speciesw ith as tandard atmospheric CO 2 pressure in the presence of ac atalytic amount of an NHC-copper complex and KOtBu. Ar elativelyb road scope of benzene derivatives, such as N,N-diisopropylbenzamide,benzonitrile, and anisole, bearing both electron-withdrawing and electron-donatingg roups,u ndergo the alumination/carboxylation sequence to afford the corresponding carboxylic acids in high yield and high selectivity (Scheme 27, 81 a-i). Heteroarenes such as benzofuran, benzothiophene, and indole derivativesa re also compatible with this method (Scheme 27, 83 a-c).
To understand the mechanism of the reaction, some key intermediates such as the coppera ryl and copper isobutyl complexes (and their carboxylation products)w erei solated and structurally characterized by X-ray crystallography.Aplausible mechanism is shown in Scheme 28. Initially, the arylaluminum species 85 would be generated by the deprotonation of an ar-Scheme26. Proposed mechanism for (IPr)CuCl-catalyzed CÀHc arboxylation.

Rh-catalyzed CÀHc arboxylation
Besides gold andc opper catalyzed CÀHc arboxylation,t he research group of Iwasawa developed am ethodf or Rh I -catalyzed direct CÀHc arboxylation of aromatic compounds under ambient CO 2 pressure through chelation-assisted CÀHa ctivation. [47] This reactionshowedn ol imitation concerning the acidity of the CÀHb ond, as substrates bearing both electron-donating and electron-withdrawingg roups are tolerated (Scheme 29). Pyridyla nd pyrazolyl were used and found to also be suitable as directingg roups, leading to the ortho-carboxylatedp roducts in good yields. The use of as toichiometric methylaluminum alkoxide AlMe 2 (OMe), prepared upon the mixing of Me 3 Al and MeOH, is important for efficient carboxylation. It is noteworthy that the formation of dicarboxylated product was not observed using this method.
More recently,I wasawa and co-workersr eported am ethod for Rh-catalyzedC ÀHc arboxylationo fs imple arenes under at-mosphericC O 2 pressure without the assistance of ad irecting group. [48] The authors found that the bidentate ligand was essentialf or this reaction, since bulky monodentate phosphine complexes showedn or eactivity.T he use of AlMe 1.5 (OEt) 1.5 , which was prepared by mixing AlMe 3 and EtOH, dramatically improved the efficiency of the reaction. Adding1 ,1,3,3-tetramethylurea (TMU) can also promote the carboxylation, probably owing to the stabilization of coordinatively unsaturated Rh I species.I ti si mportant to note that polar solvents play important roles,s ince the reaction resulted in poor turnover number (TON) in the absence of dimethylacetamide( DMA). This method allows various arenes to undergo directed carboxylation to afford the correspondingc arboxylic acids with moderate TON. Impressively,e ven simple benzoica cid could be obtained by treatment of benzene under the Rh-catalyzedc onditions (Table 3; 91 a). Electron-donating groupss uch as methyl or methoxy led to mixed products whereby the meta-carboxylated product is the major product (Table 3; 91 b and 91 c). Electron-withdrawing-group-substituted arenes such as halobenzenes mainly afforded ortho-carboxylated products (Table 3; 91 d). Furthermore,b enzofuran, N-Boc indole, and fer-Scheme28. Proposed mechanism for Cu-catalyzed formal CÀHc arboxylation.
Competitive reactions of benzene and deuterated benzene were carried out by Iwasawa et al. finding ak inetic isotope effect (KIE) value of 5.5 after 1hreactiont ime. This result suggests that the rate-determining step is CÀHb ond activation. On the basis of these results andl iterature precedent, [6b, 49] a tentatively reaction mechanism was proposed by the Iwasawa et al. (Scheme 31). Initially,t he methylrhodium(I) complex 101 was generated by the mixture of Rh catalyst 100 and AlMe 1.5 (OEt) 1. 5 .O xidative addition of aC ÀHb ond of the arene affords an aryl(hydride)(methyl)rhodium(III) intermediate (102). Reductivee liminationo fi ntermediate 102 providesa rylrhodium(I) complex 103 and methane, which was detected by GC analysis.Nucleophilic addition of the highly reactive arylrhodium(I) complex 103 to CO 2 gives rhodium(I) benzoate complex 104,w hich then further reacts with AlMe 1.5 (OEt) 1.5 to furnish the aluminum carboxylate 105 and regenerating the active Rh catalyst 101.T he final product would be obtainedb y simple acidic workup.
Another detailed mechanistics tudy by Iwasawa andc oworkers confirmed their initially postulated mechanism above to be fundamentally correct. [50] However,i nsteado ft he reaction between 104 and methylaluminum speciest og ive 101, the transmetallation of 104 with chloroaluminum species could also occur,c onverting the catalystb ack to 100.T he carboxylation of catalyst 101 could also take place in the same manner as the carboxylation of 103,c onsequently affording the acetic acid by-product.
On the basis of the crystal structure of 2,6-dihydroxybenzoate decarboxylase (Strain MTP-1005;PDB code 2DVU) reported by Goto and co-workers, [58] it was suggested that proton abstraction of the phenolic hydroxy group occurs at the Asp287 residue, activated in at riad by His218 and Glu221 (Scheme 33). This in turn enhancest he nucleophilicity of the ortho carbon and enables an ucleophilic attack ontoc oordinated bicarbonate bound by at ightly coordinated Zn 2 + center bound by Asp287, Glu8, His10,and His164residues. The oxyanion intermediate 107 is formed initially through nucleophilic addition of 1,5-dihydroxybenzene (109)t ot he active site of 2,6-dihydroxybenzoate decarboxylase (106). The carboxylated product 2,6-dihydroxybenzoic acid docked into the 108 active site is then obtained by the re-aromatization of 107 aided by a network of three catalytically important and structurally conservedw ater molecules (W1, W2, and W3), which is triggered by Asn128t hrough ap rotonation-deprotonation sequence. The final carboxylatedp roduct 110 would be released from the active site by replacing another substrate 109 with 108,s imultaneously regeneratingt he decarboxylase with an undocked active site for as ubsequentc atalytic cycle. This enzymatic carboxylation relies on the electrophilica romatic substitution process.

Conclusions
The direct carboxylation of hydrocarbons by way of fixing CO 2 as aC 1s ource holds potentiali no rganic synthesis as an ideal strategy to construct the CÀCb ond in as tep-and atomeconomical fashion. Recent advances in CÀHc arboxylation using CO 2 were summarized;i np articular the carboxylation of aromatic compounds that are important motifs among natural products andbiologically active compounds. This progress represents the development of green chemistry methodologies to meet economic and environmental requirements, and provides an alternative to traditional CO 2 couplingr eactions that require organometallic reagents.
The classic strategies for CÀHc arboxylation with CO 2 require the use of base or Lewis-acid mediators. In this regard,t he base-mediated CÀHc arboxylation usually requires as trong base to deprotonatet he most acidic protont of orm as trong Scheme32. Selectede xamples of (de)carboxylase-catalyzed carboxylation of phenols.
ChemSusChem 2017, 10,3317 -3332 www.chemsuschem.org 2017 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim nucleophilic carbon atom, thus enabling attack on the weakly electrophilic CO 2 .H owever,t his base-mediated CÀHc arboxylation generally requires relatively high reactiont emperatures. On the other hand the strategy of Lewis-acid-mediated CÀH carboxylation relies on the activation of CO 2 through coordination with the Lewis acid. The aromatic compounds can therefore react with the activated CO 2 .T his strategy allows for CÀH carboxylation at relativelyl ow reaction temperatures. However, it usually requires high CO 2 pressure and the regioselectivity is often poor.
The development of methods operating under milder conditions, such as room temperature or low pressure of CO 2 ,s till remainsagreat challenge for synthetic chemists. Many catalytic transformationso fC O 2 have been accomplished in recent decades, and this progress allows for CÀHc arboxylationt o proceeda tr elativelyl ow reaction temperatures and CO 2 pressures. Despite advances in this area the substrate scope is often limited. Thus, it is still highly desirable to develop new methodologies for carboxylation to proceed in simple, mild, and generally applicable conditions using ab road scope of readily available startingm aterials.