Direct Carboxylation of C(sp³)-H and C(sp²)-H Bonds with CO₂ by Transition-Metal-Catalyzed and Base-Mediated Reactions

Abstract

This review focuses on recent advances in the field of direct carboxylation reactions of C(sp³)-H and C(sp²)-H bonds using CO₂ encompassing both transition-metal-catalysis and base-mediated approach. The review is not intended to be comprehensive, but aims to analyze representative examples from the literature, including transition-metal catalyzed carboxylation of benzylic and allylic C(sp³)-H functionalities using CO₂ which is at a “nascent stage”. Examples of light-driven carboxylation reactions of unactivated C(sp³)-H bonds are also considered. Concerning C(sp³)-H and C(sp²)-H deprotonation reactions mediated by bases with subsequent carboxylation of the carbon nucleophile, few examples of catalytic processes are reported in the literature. In spite of this, several examples of base-promoted reactions integrating “base recycling” or “base regeneration (through electrosynthesis)” steps have been reported. Representative examples of synthetically efficient, base-promoted processes are included in the review.

1. Introduction

Viable and efficient conversion of CO₂ into chemicals and fuels is a challenging goal and, at the same time, a subject of scientific debate [1]. The high thermal and kinetic stability of the CO₂ molecule require for a suitable catalyst and energy input when employed in E-CO₂ bond formation (E = O, N, C) or reduction reactions. Current synthetic options for CO₂-fixation into useful chemicals encompasses chemical, electro-chemical, photo-chemical, photo-electrochemical and biological approach that have been overviewed in a number of excellent books and comprehensive reviews [2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17].

An analysis of the scientific literature suggests that a convenient energy-balance in CO₂-conversion processes requires improving energy efficiency through the use of renewable energy sources, advanced reactor technologies (as for example continuous-flow reactors) and efficient catalysts.

It is also widely accepted that present chemical CO₂-utilization does not contribute significantly to reduction of CO₂ emission (chemical industry is estimated to fix about 1.1 Mt of CO₂ per year in a variety of chemicals as urea, salicylic acid, cyclic carbonates and polycarbonates [18]).

Looking from a different perspective, a research topic receiving particular attention over the last decades concerns the possibility to substitute traditional protocols employed in the synthesis of carboxylic acids with more environmental friendly procedures based on the use of CO₂ as a cheap and non-toxic C1 source [19].

Indeed, carboxylic acids and derivatives are commercially valuable synthetic compounds and key intermediates in chemical and pharmaceutical industries [20,21,22,23].

In the context of thermal catalysis, an analysis of synthetic strategies pursued in the synthesis of carboxylic acids (and derivatives) from organic substrates and CO₂ encompasses:

(1)

transition-metal-catalysed carboxylation of preactivated substrates including allylstannanes [24,25,26,27], organoboronic esters [28,29,30,31,32], organozinc reagents [33,34] and aryl halides [35,36,37] as exemplified in Scheme 1a–d;

(2)

transition-metal-catalysed hydrocarboxylation of unsaturated compounds (olefins, allenes, alkynes) requiring AlEt₃ or ZnEt₂ co-reactants to generate a metal-hydride species undergoing insertion of the unsaturated substrate and subsequent carboxylation of the organometallic intermediate [38,39,40,41,42] as exemplified in Scheme 1e. It is worth citing very recent examples of hydrocarboxylation reactions involving substrate reduction by a “formal hydride donor”: (i) in Scheme 1f, [43] hydride is formally generated from H₂O and Mn; (ii) in Scheme 1g, [44] the substrate is reduced by photo-induced transfer of 2e⁻ and 2H⁺ from a sacrificial amine; (iii) in Scheme 1h, [45] the substrate is carboxylated and subsequently reduced with H₂ in a Poly-NHC/Ag/Pd mediated process;

(3)

use of CO₂ in conjunction with alkenes, alkynes, dienes, allenes, diynes, in oxidative cycloaddition reaction of low valent metal complexes to form five-membered metallacycles [46,47,48,49,50,51,52] as exemplified in Scheme 1i;

(4)

direct carboxylation of C-H bonds with CO₂ avoiding C-H pre-functionalization, hydrocarboxylation or metallacycle formation as exemplified in Scheme 1j [53]).

Among abovementioned synthetic strategies, C-H bond functionalisation with CO₂ as defined in point 4 seems to be particularly attractive as it can give straightforward access to value added products. In several reports, the term “direct carboxylation” is used to designate all synthetic strategies exemplified in Scheme 1a–j. Here we will use the term “direct carboxylation” with the meaning of point 4. In this review, we inspect recent advances concerning direct CO₂-based carboxylation reactions of C(sp³)-H and C(sp²)-H bonds encompassing both transition-metal-catalyzed and Brønsted-base mediated reactions. Also a few examples of carboxylation of CH₄ in the presence of heterogeneous catalysts as well as light-driven CO₂-based carboxylation of unactivated C(sp³)-H bonds will be considered.

It is worth noting that the research field of direct functionalisation with CO₂ of aromatic and heteroaromatic compounds has registered many developments in the last decade which have been object of excellent reviews [19,54,55]. Due to space limitations, Friedel-Crafts carboxylation of aromatics [12,55] will be not considered here.

Direct carboxylation of compounds containing activated C(sp³)-H bond has also received considerable attention. In contrast, the synthetically appealing direct carboxylation of unactivated C(sp³)-H bonds with CO₂ is a research field in its “nascent stage” (as stated by Sato, Cazin and Nolan [53,56]). We will include in the analysis recent examples of carboxylation of benzylic and allylic C(sp³)-H functionalities with CO₂ that promise to disclose new future research directions.

Base-mediated carboxylation reactions using CO₂ are initiated by abstraction of a C-H proton with production of a carbon anion able to nucleophilically attack the heterocumulene (Scheme 2). In a few cases the Brønsted base is playing a catalytic role (being deprotonated at the end of the cycle, Scheme 2a,b) while, in most cases, it reacts with the substrate in stoichiometric amount (occasionally a large base excess is required) (Scheme 2c). Stoichiometry applying throughout the course of the reaction for the abovementioned cases is shown in Scheme 2a–c.

For stoichiometric reactions falling under Scheme 2c, recent reports have dealt with the problem of Brønsted-base recycling [58,59] or Brønsted-base regeneration through electrodialysis [60] providing, on the overall, very efficient synthetic processes (see Section 4.1 and Section 6.3).

Interestingly, several reports [56,57,61] describe the employment of metal complexes bearing OH⁻ or ^tBuO⁻ ligands acting as proton acceptors during the catalytic cycle (an example is reported in Scheme 2d [57]) (see Section 5.2). These efficient synthetic methods can be “formally” considered as processes in which a base-promoted reaction takes place with the assistance of a metal catalyst.

We consider that it is worth including here selected examples reported in the literature concerning Brønsted-base promoted reactions (proceeding according to stoichiometry shown in Scheme 2c) as these processes provide access to efficient synthetic protocols to added value chemicals. Moreover, in the light of the abovementioned synthetic advances, base-promoted reactions may be integrated with a base-regeneration step or inspire future base-promoted reactions assisted by metal-catalysis.

2. Transition-Metal-Catalyzed Carboxylation of C(sp³)-H Bonds with CO₂

2.1. Carboxylation of Compounds Possessing Activated C(sp³)-H Bonds Catalyzed by Ag-Salts in Conjunction with Strong Bases

A recent report from the Yamada group [62] describes catalytic incorporation of CO₂ into specific substrates as ketones containing an alkyne unit (Scheme 3).

The reaction proceeds through a carboxylation/dig-cyclization pathway under mild conditions (P_CO2 = 1.0 MPa, 25 °C) affording γ-lactone derivatives (1).

This reaction was inspired by the efficient synthesis of cyclic α-methylene carbonates (2) from prop-2-ynyl alcohol derivatives and CO₂ shown in Scheme 4 [63,64,65,66,67].

α-methylene cyclic carbonates (2) can be obtained in good to excellent yields under transition-metal- or base-catalysis in a synthesis requiring high CO₂ pressure (P_CO2 = 0.5–5 MPa) and relatively high temperature (80–120 °C). In previous works dealing with carboxylative cyclization reaction of prop-2-ynyl alcohol derivatives under CO₂, Yamada [68,69] reported the combined use of silver salts and organic bases to enable the employment of milder reaction conditions (P_CO2 = 0.1 MPa, 80 °C). Among screened metal salts (Rh(acac)₃, PdCl₂, PtCl₂, CuCl, AgOAc, AuCl, Hg(OTf)₂) silver cation showed superior catalytic properties as π-Lewis acid in assisting the dig-cyclization process (Scheme 5).

Yamada applied the silver-based protocol to acetophenone derivatives possessing an alkynyl functionality to afford γ-lactone derivatives (1) (Scheme 3). The reaction was typically carried out using AgOBz (20 mol %) and MTBD (7-Methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene) (4 eq.) under P_CO2 = 1.0 MPa at 25 °C. By varying the substituent on the 1-phenyl-4-pentyne-1-one skeleton (16 substrates), γ-lactone derivatives 1 were obtained in yield ranging from 36 to 91% with 100% Z selectivity with respect to the double C-C bond. Dihydrofuran derivatives were detected in reaction products at lower amount (5% yield). The scope of the reaction could be extended to aliphatic ketone derivatives (Scheme 3, 1h–j) [62] possessing two different protons α and α′ to the carbonyl group. Also in this case γ-lactone derivatives were obtained with high selectivity (3 examples, 31–59% yields) without any control of the formation of enolate.

The combined use of silver salts (AgOAc, 10 mol %) and organic base [DBU, (1,8-diazabicyclo(5.4.0)-7-undecene), 2 eq.] catalysis under CO₂ atmosphere was extended by Yamada to o-alkynyl acetophenone derivatives (Scheme 6) leading selectively to 1(3H)-isobenzofuranylidene acetic acids derivatives (3) in moderate to excellent yield (12 examples, 47–99% yields) [70].

In situ esterification of the carboxylic acid functionality with MeI afforded the corresponding methyl esters. Dihydroisobenzofuran products bearing an aromatic substituent (R², Scheme 6) at the alkynyl-moiety were isolated in 100% Z selectivity with respect of the two C-C double bonds.

This report appears synthetically interesting as the dihydroisobenzofuran structure is present in several bioactive compounds as Pestacin and Escitalopram. The literature documents the dihydroisobenzofuran skeleton can be obtained by a transition-metal-catalyzed cyclization reaction of o-alkynylbenzylalcohol or o-alkynylbenzaldehyde derivatives [71,72,73]. However, this synthetic approach allows only a limited number of substituents of the dihydroisobenzofuran skeleton. In contrast, the Yamada approach shown in Scheme 6 give access to dihydroisobenzofuran derivatives (12 examples) bearing a carboxyl-group substituent. The reaction is proposed to proceed via formation of an enolate that undergo carboxylation with CO₂ producing intermediate 4 (Scheme 7). The oxygen atom of the enol tautomer 4 reacts through an intramolecular 5-exo-dig regioselective cyclization pathway affording dihydroisobenzofuran derivatives 3.

A more detailed description of the reaction mechanism was reported Lu (see Scheme 9) [74] who further optimized the reaction conditions for the carboxylation/intramolecular cyclization reaction of o-alkynyl acetophenone with CO₂ by using MTBD (2 eq.) and AgBF₄ (2 mol %) in DMF under atmospheric CO₂ pressure (Scheme 8).

By varying the substituents at the aromatic ring (R¹, Scheme 8) and at the alkyne moiety (R², Scheme 8), 1(3H)-isobenzofuranylidene acetic acids (or ester derivatives 3) were obtained (15 examples) in 50–95% yields (Scheme 8). On the basis of deuterium-labelling experiments and DFT calculations Lu and co-workers proposed the reaction mechanism shown in Scheme 9.

The reaction proceeds through proton abstraction from the methyl group by DBU generating, thus, an enolate that further reacts with CO₂ affording the β-keto carboxylate anion 5. Subsequently, the negatively charged oxygen atom of the enolate tautomer nucleophilically attacks the Ag-activated alkyne moiety leading to the cyclic product. A deuterium-labelling experiment on 2-(phenylethynyl)phenylethanone has shown that the hydrogen α to the phenyl substituent (R = C₆H₅, Scheme 10) in 3 comes from the methyl group of the reagent, indicating that hydrogen shift is involved in the reaction.

To explain the high selectivity towards the 5-exo oxygen cyclization product a DFT study using the wB97XD function was implemented. The study revealed the 5-exo-oxygen cyclization pathway having the lowest energy barrier for the O atom attack to the alkynyl-moiety during cyclization.

2.2. Carboxylation of Benzylic and Allylic C(sp³)-H Bonds Catalyzed by Transition-Metal-Complexes

The group of Sato has recently reported direct functionalization with CO₂ of benzylic and allylic C(sp³)-H bonds catalyzed by transition-metal-complexes.

In 2012, Sato published a sequential protocol for the selective “formal” carboxylation of benzylic C(sp³)-H bonds applied to substrates possessing a nitrogen directing group (Scheme 10) including substituted 8-methylquinoline, 2-(o-tolyl) pyridine, 2-(o-tolyl) pyrimidine and 2-(o-tolyl) quinoline (11 substrates, 29–90% yields) [75].

At first substrates were silylated with Et₃SiH in toluene by Ru₃(CO)₁₂ catalyst according to the Kakiuchi protocol [76] to give benzylsilane derivatives 6 (Schemes 10a and 11a). Interestingly, the [Ir(cod)Cl]₂ catalyst was tested for the first time for C(sp³)-H silylation reactions (Schemes 10b and 11a). Both Ir and Ru catalysts afforded silylated products (6,8) in almost quantitative yield.

As shown in Scheme 11a, silylation of 2-(o-tolyl) pyridine exhibited a different regioselectivity toward C(sp²)-H and C(sp³)-H bonds functionalization in the presence of [Ir(cod)Cl]₂ and Ru₃(CO)₁₂ catalysts. While the Ir catalyst was found to afford selectively the C(sp²)-silylated product (8), the Ru catalyst afforded both C(sp²)- and C(sp³)-silylated derivative (6). The difference in regioselectivity was explained on the basis of the preference of Ir to participate in 5-membered metallacycles in contrast to Ru that is reported to form both 5- and 6-membered rings (Scheme 11b).

Benzyl-silane derivatives were subsequently carboxylated at the benzylic C(sp³)-Si bond using CsF and CO₂ after solvent exchange. Unfortunately, the carboxylation reaction was affected by undesired proto-desilylation, therefore carboxylated products (7) were obtained in moderate to good yields (29–90%).

A critical examination of the protocol reported by Sato evidentiates the process is a carboxylation of pre-activated silyl-derivatives in analogy with the well documented transition-metal-catalyzed carboxylation of pre-activated substrates shown in Scheme 1a–d. The originality of the process resides in the way the authors concatenate the pre-functionalisation and the carboxylation reactions that require different solvent and reaction conditions. After the silylation procedure, by simple evaporation of toluene and other volatile species (Et₃SiH and norbornene), subsequent addition of DMF and CsF allows to develop the original in situ protocol for the two reactions.

Very recently Mita and Sato have reported the catalytic carboxylation of allylic C(sp³)-H bond of terminal alkenes with CO₂ in the presence of a Co(acac)₂/Xantophos/AlMe₃/CsF system [53] (Scheme 12). Allylarenes (15 examples) were carboxylated to linear trans-styrylacetic acids (9) with moderate to good yields (41–84%) and 100% regioselectivity. Olefin isomers were detected in reaction products at lower amount (Scheme 12a). The scope of the reaction was extended to 1,4-diene derivatives giving access to hexa-3,5-dienoic acid derivatives in moderate to good yields (7 examples, 24–78% yields) (Scheme 12b).

The addition of CsF to the reaction mixture was shown to increase by approximately 40% the yield of the reaction. Sato tentatively attributed the CsF beneficial effect to the interaction of F⁻ with CO₂ which allows for dissolving more CO₂ into the reaction system.

The mechanism of the reaction has been proposed to involve a L_nCo(I)Me active species (Scheme 13) formed from Co(acac)₂ and AlMe₃ in the presence of the Xantphos ligand. The L_nCo(I)Me active species is proposed to form through L_nCo(II) alkylation by AlMe₃ to L_nCo(II)Me₂ and subsequent disproportionation of the latter to L_nCo(III)Me₃ and L_nCo(I)Me.

The L_nCo(I)Me complex coordinates, thus, the terminal olefin (11) with subsequent cleavage of the adjacent allylic C-H bond producing a η³-allyl-Co specie (12). The authors suggest a key role could be played at this stage by the oxygen donor atom of Xantphos ligand as its coordination to Co would assist the tautomerization of η³-allyl-Co to η¹-allyl-Co complex (Scheme 13b). Reductive elimination of CH₄ from η¹-allyl-Co(H)(CH₃) intermediate (13) leads to a low valent allyl-Co complex (14) able to undergo electrophilic attack from CO₂ at the γ-position. The last stage involves transmetallation between the carboxylated-Co complex (15) and AlMe₃ regenerating the L_nCo(I)Me catalyst and producing R-COOAlMe₂. Addition of HCl to the reaction mixture gives the styrylacetic acids.

For selected substrates the authors achieved further transformation of styrylacetic acids into γ-butyrolactones through classical synthetic procedures (Scheme 14).

(E)-4-(Benzo[d][1,3]dioxol-5-yl)but-3-enoic acid (16) was converted into anti-5-(Benzo[d][1,3]dioxol-5-yl)-4-hydroxydihydrofuran-2(3H)-one (17) by treatment with dimethyldioxirane (68% yield) (Scheme 14a). The ester methyl (E)-4-(benzo[d][1,3]dioxol-5-yl)but-3-enoate (18) was converted into the optically active syn-β-hydroxy-γ-butyrolactone (4R,5R)-5-(Benzo[d][1,3]dioxol-5-yl)-4-iododihydrofuran-2(3H)-one (19) (80 Yield, 99% ee) by Sharpless asymmetric dihydroxylation.

Very recently, the Hou group succeeded in achieving a “formal” Cu(I)-catalyzed carboxylation of C-H bonds of allyl phenyl ethers (20) (Scheme 15a) [77] by using an aluminium ate compound such as (ⁱBu)₃Al(TMP)Li (28, Scheme 16b) (TMP = 2,2,6,6-tetramethylpiperidide) [78] as base in proton abstraction form the substrate without side reactions.

As shown in Scheme 15a, the reaction consists in selective proton abstraction from phenyl allyl ethers (20) by (ⁱBu)₃Al(TMP)Li (2 eq.) (28) to obtain an η¹-allyl-anion coordinated to Al (21) that undergo subsequent transmetallation/carboxylation in the presence of the [(IMes)Cu(O^tBu)] catalyst (29) to give methyl 3-butenoate derivatives (22) in high yield and high stereoselectivity. Phenyl allyl ethers bearing slightly electron-donating alkyl substituents at the phenyl nucleus as well as electron rich and electron-withdrawing substituents were carboxylated and subsequently esterified to 2-aryloxy-3-butenoate methyl ester derivatives (17 examples, 65–90% yields) with high regio- and stereo-selectivity (Scheme 15a, pathway A). The authors noticed that, during the Cu(I)-catalysed step, by increasing temperature and reaction period significant amount of the (Z)-2-aryloxy-2-butenoate isomers (23) were formed. The group was able to obtain products 23 with high regio- and stereo-selectivity by addition of DBU in catalytic amounts after the Cu(I)-catalysed reaction step (Scheme 15a, pathway B) (16 examples, 52–91% yields). The scope of the reaction could be extended also to allyl benzenes (24) (Scheme 15b) that were converted into 2-aryl-3-butenoate derivatives (25) (84–85% yield) via pathway A and into (E)-2-aryl-3-butenoate derivatives via pathway B (obtained in high yield). Finally, carboxylation of p-methylbenzamide (26, Scheme 15c) at the C(sp³)-H benzylic functionality was obtained via pathway A (Scheme 15c) affording the corresponding carboxylic acid (27) in 52% yield and high regioselectivity.

Scheme 16a shows formation of an (η¹-ally)-Al(ⁱBu)₃Li complex 21 by reaction of (ⁱBu)₃Al(TMP)Li (28, Scheme 16b) with the substrate through TMP proton abstraction. Subsequently [(IMes)Cu(O^tBu)] (29) (10 mol %) is added to the reaction mixture. Concerning the mechanism of the Cu(I)-catalyzed reaction, the author proposes that 21 undergo transmetallation with [(IMes)Cu(O^tBu)] affording (η¹-ally)-Cu(IMes) complex 30. Carboxylation of 30 occurs in extremely mild reaction conditions (P_CO2 = 0.1 MPa, 0 °C) affording 2-aryloxo-3-butenoate anion coordinated to “Cu(I)(IMes)” fragment (31). As final steps, transmetallation between 31 and 21 complexes produces (32) and regenerates the (η¹-ally)-Cu(IMes) species (30).

2.3. Carboxylation of CH₄ with CO₂ by Heterogeneous Catalysis

The reaction of CH₄ with CO₂ to form acetic acid was investigated by the Spivey group under heterogeneous catalysis (5% Pd/carbon and 5% Pt/alumina) at the temperature of 400 °C. (Scheme 17a) [79]. The reaction is highly thermodynamically unfavourable, therefore, to overcome the unfavorable equilibrium, CO₂ and CH₄ were reacted on the metal oxide to form acetate adsorbed on the catalyst surface. Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRITS) experiments evidenced formation of acetate adsorbed on the catalyst. Subsequently, a Temperature Programmed Reaction (TPR) experiment showed formation of gas phase acetic acid.

An alternative method to overcome the unfavorable equilibrium of reaction shown in Scheme 17a consisted of reacting an equimolecular gaseous mixture of CH₄, CO₂ and C₂H₂ over 5% Pt/Al₂O₃ and Zn(OAc)₂/carbon catalysts at 1 atm over a temperature range of 200 to 400 °C (Scheme 17b) [80]. The reaction of acetic acid with acetylene to form vinyl acetate shifts to the right the unfavourable formation of acetic acid from CH₄ and CO₂.

The reaction is proposed to proceed following a two steps pathway: first acetic acid is formed at the Pt-catalyst, then the acid reacts with acetylene affording vinyl acetate at the Zn-catalyst.

3. Light-Driven CO₂-Based Carboxylation Reactions of C(sp³)-H Bonds

Very recently, two examples of photo-catalysed carboxylation reactions of C(sp³)-H bonds have been reported by Jamison [81] and Macyck [82]. The Jamison process employs an organic photocatalyst (p-terphenyl) while the latter employs a ZnS photocatalyst. In addition, two examples of light-driven reactions exploiting the photochemical properties of ketones reported by Murakami are reviewed [83,84].

3.1. Photo-Catalysed Carboxylation of Acetylacetone with CO₂

Very recently, Macyk has reported the carboxylation of acetylacetone to 2-acetyl-3-oxobutanoic (34) and 3,5-dioxohexanoic (35) acids by using a photocatalytic system based on the wide band gap (3.6 eV) semiconductor zinc sulphide (E° = −2.0 V vs. SHE) with deposited ruthenium nanoparticles (Scheme 18) [82].

Temperature-programmed desorption (TPD) experiments showed that Ru-nanoparticles significantly increased CO₂ adsorption at the catalyst surface. Upon irradiation of ZnS nanoparticles suspended in CCl₄ acids 34 and 35 were detected in solution by application of IR and ¹³C NMR spectroscopies. Although the organic products could not be isolated, 34 and 35 were estimated to form in a 2:1 molar ratio. As shown in Scheme 18 the CO₂^•− radical is proposed to form at the Ru-particles surface while interaction of acetylacetone with photogenerated holes produces the organic radicals. At the last stage the reaction involves a radical-radical coupling to form acids 34 and 35. EPR spin-trapping experiments confirmed formation of CO₂^•− supporting, thus, the proposed mechanism.

3.2. Light-Driven Carboxylation of Benzylic and Allylic C(sp³)-H Bonds with CO₂

Very recently, Jamison reported the photoredox activation of carbon dioxide to promote carboxylation of N-benzyl substituted cyclic amines (36) in a continuous flow apparatus (Scheme 19) [81]. In typical reaction conditions, a mixture of substrate (N-benzyl substituted cyclic amines, 36), photoredox catalyst (para-terphenyl, 20 mol %), CF₃COOK (3 eq.) and CO₂ was irradiated in DMF under an ultraviolet lamp at the temperature of 30–35 °C during 10 min. CO₂ was introduced in the flow apparatus through a mass flow controller. In optimized conditions irradiation with a λ > 280 nm was used to avoid unproductive reactions promoted by short-wavelenght ultraviolet light.

Tertiary N-benzylpiperidines bearing electron-neutral or electron-rich aryl-substituents and various N-benzyl cyclic amines (16 examples) were carboxylated with high regioselectivity at the benzylic position in good to high yields (46–99% yields) (Scheme 19a). Other N-substituted cyclic amines as ticlopidine (38) and N-cyclohexylpiperidine (40) were carboxylated at α-position with respect to N showing that the scope of the reaction may expand to other amines with different regioselectivity (Scheme 19b,c).

Concerning the reaction mechanism, (Scheme 20) Jamison proposes the photochemical reaction is initiated by the photo-excited p-terphenyl catalyst (E° = −2.63 V vs. SCE in DMF) favoring single-electron transfer with the tertiary amine to provide the strongly reducing p-terphenyl radical anion (42) and the amine radical cation (43). Subsequent deprotonation of the amine radical cation by CF₃COO⁻ affords the neutral α-amino radical (44). The strong reducing p-terphenyl radical anion (42) is able to reduce carbon dioxide to CO₂^•− (45) which undergoes radical-radical coupling with the α-amino radical (44) leading to α-amino acids (37).

A report by Murakami group describes irradiation by UV or solar light of o-alkylphenyl ketones under CO₂ atmosphere (0.1 MPa) to produce o-acylphenylacetic acids (46) in good to high yields (14 examples, 61–95% yields) (Scheme 21) [83].

The reaction is proposed to proceed through endergonic isomerisation of o-alkylphenyl ketones to o-quinodimethanes (47) (Norrish Type II photoreaction) and subsequent [4 + 2] cycloaddition reaction of the (E)-o-quinodimethane with CO₂ (Scheme 22).

Interestingly, the reaction does not require any sacrificial reagent.

More recently Murakami reported a carboxylation reaction of C(sp³)-H allylic bonds of simple alkenes (10 examples) to form alkenylacetic acids (48) (β,γ-unsaturated carboxylic acids) in the presence of 3,6-diphenylxanthone (49), IPrCuCl complex (50) and ^tBuOK (Scheme 23) [84]. Although the carboxylation of cyclohexene afforded 2-cyclohexene-1-carboxylic acid in approximately 2% yield (based on the alkene reagent) the ketone and the Cu-complex showed to react catalytically with a TON of 9 and 45 respectively (Scheme 23).

The mechanism proposed (Scheme 24) involves light-driven excitation of ketone 49 followed by abstraction of an allylic proton from cyclohexene by the carbonyl oxygen affording a radical pair (51). Subsequent radical-radical coupling leads to a tertiary homoallyl alcohol (52) (high energy intermediate) which undergo subsequent deprotonation by [Cu]O^tBu (50) and complexation by [Cu]. The copper alkoxyde (53) undergoes β-carbon elimination leading to an allyl-Cu complex (54) regenerating the ketone (49). Insertion of CO₂ into the C-Cu bond of 54 leads to a carboxylate-Cu complex (55) that undergoes exchange with KO^tBu producing the carboxylate potassium salt (56).

4. Brønsted-Base-Mediated Carboxylation of C(sp³)-H Bonds with CO₂

A significant knowledge has been acquired over the past decades concerning Brønsted-base-promoted carboxylation of acidic C(sp³)-H bonds proceeding according to stoichiometry reported in Scheme 2c.

As reported in Section 1, acquired kwoledges in this research area might allow future developments through optimization of the base-recycling process or implementation of base/transition-metal co-catalyzed reactions. In light of this view, an analysis of recent advances in this field and relevant developments of this chemistry over the last two decades are reported in the following two sections.

4.1. Recent Advances in Brønsted-Base Mediated Carboxylation of Acidic C(sp³)-H Bond with CO₂

Thanks to reports by Jessop [58] and Beckman [59], significant advances in the synthesis of β-keto carboxylic acids from ketones and CO₂ and their further conversion into β-ketoesters and β-hydroxycarboxylic acids have been introduced. Very recently an interesting example of base-promoted carboxylative cyclization reaction of 1-propenyl ketones to α-pyrones has been reported by Lu [85].

Interestingly, in the Jessop process DBU can be easily recycled for subsequent reactions while in the Beckmann process DBU is shown to react catalytically. From a synthetic point of view, β-ketoesters are valuable building-block in the preparation of a wide variety of molecular systems as their structural unit is composed of two different electrophilic carbonyls and two nucleophilic carbons that can react selectively under suitable reaction conditions [86].

Moreover, chiral β-hydroxycarboxylic acids are important intermediates widely used as chiral precursors of anti-inflammatory products, [87] β-amino acids, [88] β-lactams, [89] β-lactones, [90] and pheromones [91].

The Jessop group developed an efficient two step procedure for the synthesis of chiral β-hydroxycarboxylic acids from ketones, CO₂ and H₂ [58]. First β-ketocarboxylic acids (57) are synthesized in a DBU-promoted reaction, subsequently the acids are stereoselectively reduced with H₂ to chiral β-hydroxycarboxylic acids (58) by using the RuCl₂[(S)-BINAP] catalyst (Scheme 25). It must be pointed out that, while asymmetric hydrogenation of β-keto carboxylic esters is a well known procedure, there was only one previous example of catalytic hydrogenation of β-keto acids reported by Genêt et al. [92].

Optimal reaction conditions for the carboxylation of acetophenone to benzoylacetic acid (91% yield) were found by working in dry conditions and in the absence of solvents (neat DBU/acetophenone mixture, 2 eq. of DBU, P_CO2 = 6.0 MPa, 0 °C). Optimized reaction conditions were applied to other ketones as acetone, cyclohexanone and aryl-substituted acetophenones affording the corresponding β-keto acids in yield ranging from 11 to 78% and high regioselectivity. Several substrates such as isobutyrophenone and (1R)-camphor could not be carboxylated by using DBU. Instead NEt₃/MgI₂ and LDA were suitable bases.

The author underscores that, due to mild conditions adopted and high selectivity of the reaction, unconverted ketones and DBU can be recovered from the reaction mixture and easily recycled. The developed method meets, thus, most of requirement for an industrial application.

In the second reaction step (Scheme 25) β-ketocarboxylic acids are asymmetrically hydrogenated in various polar and nonpolar solvents by using RuCl₂[(S)-BINAP] catalyst (1 mol %, P_H2 ranging from 0.5 to 8.0 MPa, 25 °C).

As β-ketoacids easily undergo a decarboxylation reaction, a thorough investigation of solvent effect and other factors affecting the β-hydroxycarboxylic acids yield was required to set conditions favouring slow decarboxylation and fast hydrogenation of ketoacids. Methanol and CH₂Cl₂ were found as optimal solvents allowing to obtain high yield of β-hydroxyphenylacetic acid (85% and 93% yields were respectively obtained in the two solvents) and high ee (99% for the R enantiomer).

Beckman has developed a “Reversible CO₂-Carrier” (RCC) based on DBU covalently bound to siloxane (methylhydrosiloxane dimethyl siloxane copolymer, HMS) support (Scheme 26b) [59].

Interestingly, HMS-DBU suspended in methanol absorbed CO₂ up to 100% capacity at room temperature under low CO₂-pressure to form HMS-DBU-CO₂. Moreover, HMS-DBU-CO₂ recovered by filtration was shown to release CO₂ at 120 °C in thermogravimetric analysis (TGA) measures. RCC showed, thus, to possess highly desirable properties and to enable high yield conversion of ketones to the corresponding β-ketoacids (59) under ambient CO₂ pressure and temperature (Scheme 26a). Reacting in situ β-keto acids with CH₃I allowed easy RCC recovery (by simple filtration) and recycle up to five times without loss of activity.

Very interestingly, by careful determination of the DBU-active sites in the solid carrier it was possible to show that the covalently-bound-DBU reacts catalytically. In addition, the author notes the ester yields strongly correlates with the enolic content of the reacting ketone. Ketones with higher equilibrium constant for keto ↔ enol equilibrium as 2,4-pentanedione, ethyl acetoacetate, 2-indanone, cyclopentanone and cyclohexanone afforded higher carboxylic ester yield (96–99%) with TON of 200 (over 4 h). Reactants as acetophenone and para-substituted acetophenones were carboxylated/esterified with low yields (9–29%, 21–61 TON over 4 h).

The research group of Lu reported in 2016 the carboxylative cyclization of substituted 1-propenyl ketones (21 examples) with CO₂ for the synthesis of α-pyrones (61) in moderate to good yield (22–83%) (Scheme 27) [85].

The reaction is proposed to proceed through base-promoted enolization of the ketone followed by γ-carboxylation of the enolate (62) leading to a β,γ-unsaturated carboxylate intermediate (63) (Scheme 28). Proton abstraction from intermediate (63) promotes subsequent cyclization leading to α-pyrones (61).

Optimized reaction conditions were found using CsF (4 eq.), P_CO2 = 1.0 MPa, at 100 °C. Suitable bases for carboxylative cyclization of propenyl ketones were Cs₂CO₃, NaH, and CsF.

A mechanistic investigation using β-methylchalcone (64) and C¹⁸O₂ (Scheme 29) afforded the singly isotopically ¹⁸O-labeled 4,6-diphenyl-2H-pyran-2-one (65) in 45% yield while the doubly labeled isotopologue 66 was not detected.

This result indicates that the cyclization proceeds via intramolcular attack of enolate oxygen on the carbonyl group of the carboxylate intermediate 63.

4.2. Carboxylation of C(sp³)-H Active Bonds by Using CO₂ Carriers

It is known that Grignard and organolithium reagents readily react with CO₂ affording carboxylates. However, these reactions do not tolerate functionalities as ketone, ester and nitrile. Alternatively, a wide range of organic bases [93,94,95,96,97,98] have been used to deprotonate an activated C(sp³)-H bond providing a nucleophilic substrate able to react with CO₂ [99,100,101,102,103,104,105,106].

Table 1. Organic bases frequently used in Brønsted-base-promoted carboxylation of active hydrogen compounds with CO₂.

Entry	Acid-Base Equilibrium	pKa Value (of the Conjugated Acid)	Reference
1		18.0 in DMSO [93]	[99]
2	(in conjuction with MgI2)	10.70 in H2O [94]	[100]
3		11.82 in MeCN [95]	[101]
4		28.5 in DMSO [96]	[102]
5		29.0 in DMSO [93] considering CH3OH	[103]
6		17.97 in H2O [97] (calculated value)	[104]
7		21.5 in DMSO [96]	[105]
8		21.1 in DMSO [98] 32.4 in CH3CN (calculated value)	[106]

lists organic bases frequently used in Brønsted-base-promoted carboxylation of active hydrogen compounds (pKa value refers to the conjugated acids) using CO₂.

An analysis of the scientific literature available on the subject shows that base-promoted carboxylation reactions may proceeds following a “one-step” or a “two-steps” pathway as shown in Scheme 30. In the one-step procedure, the substrate is reacted with the base under CO₂ atmosphere while, in the two-step procedure at first a CO₂-carrier is formed which then transfers CO₂ to the substrate mimicking biotin-dependent enzymatic carboxylation reactions. In both cases, even apparently simple reactions exhibit relatively complex mechanisms.

Particular interest has been raised recently for “CO₂-carriers” as carboxylates (67) and carbamates (68) (Scheme 31) which have been effectively used as “CO₂-bent surrogates” in electrochemical and photochemical production of chemicals from CO₂ [10,107,108,109].

We list here on several CO₂-carriers effectively used in carboxylation reactions of activated C(sp³)-H bond according the “two step pathway” shown in Scheme 30b.

Inspired by biotin-dependent enzymatic carboxylation reactions Saegusa and co-workers synthesized in 1979 a 2-oxoimidazolidine-1-carboxylate complex 70 obtained by reacting 2-oxoimidazolidine (69) with M(OCH₃)₂ (M = Mg²⁺, Mn²⁺) and CO₂ in DMF at 50 °C (Scheme 32a) [104]. Both the Mg²⁺- and Mn²⁺-ureido-carboxylate-complexes were able to transfer CO₂ to cyclohexanone at 110 °C in DMF to form 2-oxocyclohexanecarboxylic acid that was converted into ester by reaction with MeI (Scheme 32b). By using the Mg²⁺-ureide-carboxylate complex 70 in a 10:1 molar ratio with respect to cyclohexanone the 2-oxocyclohexanecarboxylic acid methyl ester was obtained in 43% yield.

In 1983, Matsumura reported the easy carboxylation of cyclic thioureas 71 as ethylenethiourea and 1,3-propylenethiourea with CO₂ in the presence of ethylmagnesium bromide at room temperature (Scheme 33a) [110]. The carboxylato-complexes (73) were isolated and used in a transcarboxylation reaction of ketones in DMF at 15 °C (Scheme 33b). After acidification, the corresponding β-keto carboxylic acids were isolated in 49–80% yields.

Interestingly, the thioureide-CO₂ complexes 73 were able to act as CO₂-carriers at temperature significantly lower (15 °C) with respect to analogous 2-oxoimidazolidine-1-carboxylato complexes (Scheme 32b).

Scheme 32 and Scheme 33 show the use of molecules containing the carbamate functionality (N-CO₂⁻) as CO₂-carriers in the carboxylation of C-H active substrates. More recently 1,3-dialkyl- (76–77) or 1,3-diaryl-imidazolium-2-carboxylates (78) have been used as CO₂-transfer agents for the carboxylation of ketones and nitriles [66,106,111,112].

Reactions in which a “C-COO⁻” moiety behave as CO₂-transfer agent towards an organic compound are not common. Previous report is the Henkel reaction converting potassium benzoate into benzene and terephthalic acid, a valuable component for polyesters derivatives [113].

We have reported direct carboxylation of 1,3-dimethyl- (74) and 1-butyl,3-methyl-imidazolium chloride (75) with Na₂CO₃/CO₂ at high temperature (110 °C) and P_CO2 = 5 MPa (Scheme 34a) [111] to afford 1,3-dialkylimidazolium-2-carboxylates (76–77) in 92% yield and 91% selectivity. Subsequently 76–77 have been used as CO₂-carriers able to transfer CO₂ to acetone, acetophenone, cyclohexanone and benzyl cyanide (Scheme 34b) [66,106]. Interestingly, the reaction proceeds at room temperature requiring the presence of a metal salt as NaBF₄, NaBPh₄ or KPF₆ in stoichiometric amount. Carboxylated products are isolated in the form of metal salts in yield ranging from 60 to 81%.

More recently Louie has undertaken a detailed mechanistic study on the abovementioned trans-carboxylation reaction by using IPr-CO₂ (78), IMes-CO₂ and I^tBut-CO₂ as CO₂-carriers [112]. Investigating the role played by the alkali metal, Louie succeeded in isolating NHC-CO₂·M (M = Li⁺, Na⁺) adducts (79,80) that were characterized by single crystal X-ray diffraction analysis.

The X-ray structure showed Li⁺ and Na⁺ cations affording respectively monomeric (79) and dimeric (80) complexes (Scheme 35a). Most important, it was shown the interaction of the metal ion with the CO₂-moiety decreasing the CO₂-torsion angle with respect to the plane of the ring (Scheme 35b) suggesting, thus, a stabilisation of the carboxylate [114].

Information obtained from X-ray structure was contradicted by results of TGA measures on the NHC-CO₂·M adducts. As matter of fact, the metal adducts were shown to lose CO₂ at lower temperature with respect to their parent carboxylate. In addition, kinetic information inherent the carboxylation of acetophenone to benzoylacetic acid showed the reaction was first order in IPrCO₂ but independent of NaBPh₄ concentration.

Concerning the reaction mechanism, Louie proposes the trans-carboxylation reaction is initiated by spontaneous decarboxylation of the NHC-CO₂ complex (78) with generation of the free NHC carbene (81) able to promote hydrogen abstraction from the substrate producing, thus, the corresponding enolate anion (Scheme 36). Subsequent carboxylation of the enolate anion with CO₂ provides benzoylacetate.

On the basis of abovementioned experimental evidence it was concluded that the metal salts increase the solubility of poorly soluble imidazolium-2-carboxylates and is involved in product stabilization through ion pairing.

5. Transition-Metal-Catalyzed Carboxylation of C(sp²)-H Bonds with CO₂

5.1 Carboxylation of Alkenyl-C-H Bonds by Pd-Catalysis

As reported in the introductory section, direct carboxylation of aromatic and/or heteroaromatic substrates with CO₂ has received particular attention in the last decade. Excellent reviews on this topic have appeared recently [19,54,55,115]. Direct carboxylation of alkenyl-C-H bonds is more difficult to achieve, therefore, to the best of our knowledge, only one report from Iwasawa [116] describes direct carboxylation with CO₂ of 2-hydroxystyrenes in the presence of the Pd(acac)₂/Cs₂CO₃ catalytic system (Scheme 37). In optimized conditions, α-phenyl-2-hydroxystyrene (82) is carboxylated to 4-phenylcoumarin (83) (86% yield) in the presence of Pd(OAc)₂ (5 mol %) and Cs₂CO₃ under atmospheric CO₂ pressure at 100 °C. A wide range of 2-hydroxystyrene derivatives (16 examples) are carboxylated to the corresponding coumarin derivatives in yield ranging from 73 to 90%.

As possible reaction mechanism Iwasawa proposes the hydroxyl functionality of 82 undergoing deprotonation and subsequent coordination to the metal (Scheme 38). Then, the alkenyl C-H bond of α-phenyl-2-hydroxystyrene is cleaved (by chelation-assisted alkenyl C-H bond cleavage) producing a six-membered alkenyl palladium intermediate (84) bearing an additional 2-hydroxystyrene molecule (as cesium salt) as ligand. The organometallic intermediate 84 undergoes, thus, reversible carboxylation at the alkenyl functionality (affording 85) and subsequent exchange with an additional substrate molecule (in the presence of Cs₂CO₃) to eliminate the coumarin 83. At the same time the six-membered alkenyl palladium catalyst 84 is regenerated.

Iwasawa notes the carboxylation step is reversible and the cyclization reaction (lactonization) helps to shift to the right side the carboxylation-decarboxylation equilibrium.

5.2. Carboxylation of Aromatic and Heteroaromatic Compounds

The ability of dicoordinate, linear [Au(NHC)OH] complexes (NHC = 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene or 1,3-di-tert-butylimidazol-2-ylidene) to catalyze C-H activation/carboxylation reactions of carbo- and hetero-cycles under mild reaction conditions (P_CO2 = 0.15 MPa, ambient temperature) was reported by Nolan and coworkers in 2010 (Scheme 39) [57]. A wide range of substrates (including oxazoles, isoxazoles, thiazoles, isothiazoles, imidazoles, pyridazine and some aromatic heterocycles containing three or four heteroatoms) were carboxylated, in the presence of a stoichiometric amount of KOH, in yield ranging from 76 to 92% at the most acidic C-H bond. Furthermore, the synthetic method could be applied to polyfluoro- and polychloro-substituted benzene derivatives (Scheme 39).

There is also an almost contemporary report by Larrosa concerning C-H bond activation of electron deficient (hetero)aromatics by L_nAu(I)-complex to provide the corresponding aryl-Au(I) species [117].

The reaction described in Scheme 39 is considered to proceed via C-H deprotonation assisted by the Au-complex (Scheme 40). The basicity of the [Au(NHC)OH] complexes was measured by acid-base potentiometric titrations giving a pKa_DMSO = 30.3(4) value for [(IPr)AuOH] and a pKa_DMSO = 32.4(2) value for the [(I^tBu)AuOH] complexes. As a consequence, assuming simple Brøensted acid/base theory could be used to predict the feasibility of C-H bond functionalisation, substrates possessing C-H bond with a pKa below 32.1 could be carboxylated [118]. Nolan noted that mechanistic features of the Au(I)-catalyzed carboxylation reaction (shown in Scheme 40) are similar to the generally accepted mechanism of the copper(I)-catalyzed carboxylation of organoboronic esters [30,119]. Indeed, the [(IPr)AuOH] complex (90) is proposed to abstract the most acidic C-H proton from oxazole affording a [(IPr)Au(oxazolyl)] complex (91) able to insert CO₂ into Au-C bond. The author was able to isolate [2-oxazolyl-Au(NHC)] (91) and [2-oxazolyl-COOAu(NHC)] (92) intermediates by reacting [(IPr)AuOH] with oxazole in stoichiometric amounts (according to reaction conditions shown in Scheme 40) to obtain 91 and 92 in 93% and 86% yield respectively. Experimental evidence supports, thus, the proposed mechanism. The last stage of the cycle involves metathesis of the carboxylate complex 92 with KOH releasing oxazolyl-COOK and regenerating the [(IPr)AuOH] complex 90.

Further development of the abovementioned study led to the carboxylation of oxazole, benzoxazole, benzothiazole, and polyfluorobenzenes with CO₂ by employing the cheaper [(IPr)CuOH] catalyst [56] in the presence of CsOH (2.2 eq.) under low CO₂ pressure (0.15 MPa) at the temperature of 65 °C (Scheme 41).

Acid-base potentiometric titration of [(IPr)CuOH] complex gave a pKa_DMSO = 27.7(2) value. The reaction is proposed to occur through a mechanism similar to that described in Scheme 40 for the Au(I)-catalyst. The lower pKa value for [(IPr)CuOH] complex with respect to [(IPr)AuOH] complex allows to predict a more limited substrate scope. Substrates 93 and 95 were carboxylated with high selectivity at the most acidic C-H position in high yields (77–93%).

Cu(I)-NHC-based catalysts were also used by the Hou [54] group. The Cu(I)-chloride pre-catalyst [(IPr)CuCl] (5 mol %) in the presence of KO^tBu (1.1 eq.) under atmospheric CO₂ pressure at 80 °C was employed in carboxylation of various benzoxazole derivatives bearing both electron-donating and electron-withdrawing substituents at the aryl moiety. Carboxy-derivatives were esterified in situ with alkyl iodides providing the corresponding ester derivatives (11 examples, 50–89% esters yields) (Scheme 42).

1-methyl-benzimidazole and 5-phenyl-1,3,4-oxadiazole were also carboxylated selectively at C2 position in 14% and 36% yield respectively.

Concerning the reaction mechanism, the [(IPr)Cu(O^tBu)] (99) active species is proposed to form by metathesis with ^tBuOK of the [(IPr)CuCl] pre-catalyst. Isolation and fully characterization of [2-benzoxazolyl-Cu(IPr)] (100) and [2-benzoxazolyl-COO-Cu(IPr)] (101) complexes by the Hou group allowed to propose a catalytic cycle analogous to that proposed by Nolan for the Au-catalyzed reaction (Scheme 43a). Therefore, also Cu(I)-catalyzed carboxylation reactions are considered to proceed via C-H deprotonation.

Interestingly, X-ray crystallographic analysis of [2-benzoxazolyl-COO-Cu(IPr)] complex (101) showed the benzoxazolylcarboxylate ligand is bound in a chelating fashion to the Cu-atom (as shown in Scheme 43a).

Carboxylation of benzoxazoles derivatives (8 examples, 41–88% yields) and benzothiazole derivatives (7 examples, 33–79% yield) was reported by Hou employing the 1,4-di(2,6-diisopropylphenyl)-3-methyl-1,2,3-triazol-5-ylidene ligand (TPr) coordinated to Cu(I) (102) (Scheme 43b) [120] under the same reaction conditions adopted for the [(IPr)Cu(OH)] analogue.

The aluminium ate compound (ⁱBu)₃Al(TMP)Li (28) employed by Hou in the Cu(I)-catalyzed carboxylation of allyl aryl ethers (see Section 2.2) was employed also in carboxylation of arylic-C-H bonds (Scheme 44) [121]. Benzene derivatives as N,N-diisopropylbenzamide, benzonitrile and anisole (bearing both electron-withdrawing and electron-donating substituents) were carboxylated in high yield (16 examples, 50–95% yields) and high selectivity. In addition, heteroarenes such as benzofuran, benzothiophene, and indole derivatives (7 examples) were effectively carboxylated at the C(2)-position (72–94% yields).

The aluminate base (28, Scheme 45b) (TMP = 2,2,6,6-tetramethylpiperidide) [78] was used in proton abstraction from the substrate to give a Ar-Al(ⁱBu)₃Li complex (107) (Scheme 45a). Subsequent transmetallation in the presence of [(IPr)CuO^tBu] (5 mol %) and KO^tBu (5 mol %) under atmospheric CO₂ pressure afforded carboxylic acids in extremely mild reaction conditions (room temperature).

Having isolated and fully characterizated several key intermediates as copper aryl- and isobutyl-complexes and their carboxylated products, Hou proposed a reaction mechanism implying transmetallation of the (Ar)Al(ⁱBu)₃Li complex 107 with (IPr)Cu(O^tBu) to obtain the (Ar)Cu(NHC) species 108 (Scheme 45a). The (Ar)Cu(NHC) complex 108 undergoes, thus, CO₂-insertion into Cu-C bond affording the (ArCOO)Cu(NHC) carboxylate complex 109. As final step, transmetallation between (ArCOO)Cu(NHC) and (Ar)Al(ⁱBu)₃Li complex produces (ArCOO)Al(ⁱBu)₃Li (110) and regenerates the (NHC)CuAr species 108.

The research group of Gooßen developed a, (4,7-diphenyl-1,10-phenanthroline)bis(trisphenylphosphine) copper(I) nitrate (111) catalyst for the carboxylation of benzoxazole under atmospheric CO₂ pressure (Scheme 46) [122].

It is worth citing, catalyst (111) and (4,7-diphenyl-1,10-phenanthroline)bis-[tris(4-fluorophenyl)-phosphine] copper(I) nitrate catalysts effectively promotes the insertion of CO₂ into the C-H bond of terminal alkynes under very mild conditions.

Above reported examples of Au- and Cu-catalyzed direct C(sp²)-H carboxylation are proposed to proceedes via a C-H deprotonation type mechanism. Iwasawa and coworkers have recently reported the Rh(I)-catalyzed direct C-H carboxylation of 2-phenylpyridines and 1-phenylpyrazoles derivatives under ambient CO₂ pressure proceeding via a reaction mechanism involving chelation-assisted C-H bond activation (C-H oxidative addition) [123]. Iwasawa noted the new approach can be extended to substrates featuring C-H bond with pKa_DMSO above 30.

Iwasawa envisioned to generate a L_nRh(I)-(Me)-catalyst (114, Scheme 48) by reacting [RhCl(coe)₂]₂ in the presence of PCy₃ and a methylmetallic reagent. Among methylmetallic reagents AlMe₂(OMe) was shown as the most efficient in promoting carboxylation of target substrates. 2-Phenylpyridine derivatives (112) were shown as excellent substrates undergoing mono-carboxylation regioselectively at the ortho-position of the aryl-substituent (Scheme 47). The synthetic method included esterification in situ with TMSCHN₂ of the Ar-carboxylated products, therefore methyl-ester derivatives 113 were isolated in good to high yields (11 examples, 51–88% yield). Interestingly, 2-phenylpyridine derivatives bearing both electron-donating and electron withdrawing substituents in para-position of the benzene ring were suitable substrates.

When using 1-phenylpyrazole as substrate both mono- and di-carboxylated products at ortho-positions were obtained in 80% combined yield.

The reaction mechanism proposed by Iwasawa involves coordination of the substrate to the LnR(I)-Me active specie (114) with subsequent chelation assisted C-H bond activation (oxidative addition) at the C(2)-benzene position to obtain complex 115. Reductive elimination of CH₄ from the six-coordinated intermediate complex leads to a ArRh(I) cyclometallated specie (116) undergoing CO₂ insertion into C-Rh bond to give the mono-carboxylated product ArCOORh(I)L_n (117). Subsequent transmetallation of ArCOORh(I)L_n with methylalluminium reagent produces ArCOOAlX₂ (118) and regenerates the LnRh(I)Me complex 114 (Scheme 48).

More recently, Iwasawa and co-workers succeeded in CO₂-promoted carboxylation of simple arenes in the absence of directing groups [124].

A Rh(I)-complex coordinated by 1,2-bis(dialkylphosphino)ethane ligands (preferentially 1,2-bis(dicyclohexylphosphino)ethane) (119, Scheme 49) was shown to be an efficient carboxylation catalyst of simple arenes (Scheme 49) (under P_CO2 = 0.1 MPa, closed reactor) in the presence of AlMe_1.5(OEt)_1.5, 1,1,3,3-tetramethylurea (TMU) and DMA solvent. However, quite high temperature was required, ranging from 85 to 145 °C. TMU was believed to stabilize coordinatively unsaturated Rh(I)-species (Scheme 50). In optimized conditions, various arenes were carboxylated to afford the corresponding carboxylic acids in moderated TON. Electron-donating substituents such as methyl- or methoxy-group attached to the benzene-nucleus afforded mono-carboxylated products with TON ranging from 15 to 46.

The carboxylation of several heteroaromatic compounds as benzofuran and N-methylindole proceeded selectively at the 2-position with TON of 12 and 21 respectively.

A detailed mechanistic investigation was conducted by the Iwasawa group by application of an accurate kinetic study [125]. The reaction was shown to proceeds via a mechanism very similar to the carboxylation of 2-phenylpyridine derivatives (Scheme 48). Key steps were the formation of a L_nRh(I)-Me species (120) interacting with the Ar-H substrate which undergoes C-H oxidative addition to produce a six-coordinated L_nRh(III)(Ar)(H)(Me) complex (121). Further reductive elimination produces CH₄ and L_nRh(I)-Ar complex (122) that undergoes CO₂ insertion into the C-Rh bond (Scheme 50).

Moreover, the mechanistic study [125] evidenced that transmetallation of the ArCOORhL_n complex (123) with methylaluminium may involve a quite complex reaction pathway (Scheme 51). Dedicated experiments have shown that transmetallation of ArCOORh(I)L_n (123) and chloroaluminate species or AlMeX₂ may convert the catalyst back to [RhCl(dcype)]₂ (119) and to [RhMe(dcype)] (120) respectively. In turn, [RhCl(dcype)]₂ may reacts with AlMeX₂ affording the L_nRh(I)-Me species (120). The two Rh(I)-complexes ([RhCl(dcype)]₂ (119) and L_nRh(I)-Me (120) are, thus, in equilibrium. Interestingly, the equilibrium seems to limit the concentration of the active specie L_nRh(I)-Me and to suppress its decomposition.

More recently, the chlorobis(cyclooctadiene)rhodium(I) dimer ([Rh(ceo)₂Cl]₂) catalyst in the presence of methylaluminium was used to mono-carboxylate 2-phenylpyridine-5-4′-dicarboxylic acid (dcppy) (125) selectively at the ortho-position of the aryl-substituent to give the 126 derivative (Scheme 52) [126].

The reaction was used in post-modification of UiO-67(dcppy) MOF affording UiO-67(dcppy)-COOH. The resultant carboxylated MOF was used as a solid-state Brønsted acid catalyst in ring-opening reactions of epoxides with methanol.

6. Brønsted Base-Mediated Carboxylation Reactions of C(sp²)-H Bonds with CO₂

As reported in Section 1, particular attention has been devoted over the last decade to base-promoted carboxylation of C(sp²)-H bonds due to the relevance of aryl-carboxylates as bioactive substances [23]. Studied reactions often proceed according to stoichiometry reported in Scheme 2c. However, Kanan [60] has recently reported the carboxylation of 2-carboxy pyrrole to furan-2,5-dicarboxylic acid(a highly desirable biobased feedstock [127]) with Cs₂CO₃/CO₂ adding to the process a base-regeneration step (by electrodialysis). The CO₂-foot print of the overall synthetic method depends, thus, on intelligent use of renewable energy sources.

In the following Sections an example of base-catalyzed carboxylation of 2-alkynyl indoles with CO₂ is reported followed by an overview of base-promoted carboxylation reactions of C(sp²)-H bonds. The synthetic process reported by Kanan will be analyzed in Section 6.3.

6.1. Base-Catalyzed Carboxylation of 2-Alkynyl Indoles with CO₂

A TBD base was used by Skyrdrup to catalyze a carboxylation/cyclization reaction applied to 2-alkynyl indoles (127) with aromatic or aliphatic substituents at the alkyne unit (Scheme 53) to produce selectively the new heterocyclic structures 128 [128]. Aromatic 2-alkynyl indoles (R² = aryl, Scheme 53) (0.20 mmol) were reacted in the presence of 0.3 eq. of TBD at the temperature of 100 °C after injection of a definite amount of CO₂ (0.61 mmol) into a 10 mL reaction vial. Substrates underwent the carboxylation/cyclization reaction in high yield (10 examples, 71–90% yields). Using aliphatic 2-alkynyl indoles (R² = alkyl, Scheme 53) the reaction proceeded under the same reaction conditions but the substrates proved less reactive requiring 1 eq. of TBD for optimal conversion (9 examples, 53–86% yields).

A plausible reaction mechanism was proposed by Skrydstrup (Scheme 54) encompassing the formation of the well characterized TBD-CO₂ zwitterionic adduct (129). Interaction of the 2-alkynyl indole with the TBD-CO₂ adduct promotes deprotonation of the NH-functionality and subsequent electrophilic attack by CO₂ at the 3 position of the indole ring (130). Re-aromatization of the indole ring by 3,1-hydrogen shift produces the carboxylate (131) which is prone to attack the alkynyl-moiety via a 6-endo-dig-cyclization mechanism. Recently, the reaction mechanism proposed by Skrydstrup has been reanalyzed by Wong and co-workers via a DFT study [129] using the energetic span model developed by Shaik and Kozuch [130] to calculate turnover frequencies (TOFs). The DFT analysis suggests that the more favored activation pathway encompasses a bifunctional mechanism in which TBD is able to deprotonate 127 at the nitrogen atom (proton transfer 1, Scheme 54b) engaging, at the same time, hydrogen bond with CO₂ (132). Next, C-C bond formation affords the 133 species which requires a second TBD molecule for C-H proton abstraction (proton transfer 2) to give 134. Strong hydrogen bonding between guanidinium TBDH⁺ and the anionic CO₂ moiety stabilizes intermediate 134. Additional transfers of hydrogen ion between TBDH⁺ and nitrogen of indole-CO₂ regerates TBD and gives 135 which can undergo a ring closure reaction to provide 136. Proton transfer 4 from TBDH⁺ to 136 gives finally TBD and product 128. Calculated TOF for the bifunctional activation pathways was 1.2 × 10⁻¹³ h⁻¹, which was the higher value as compared to alternative pathways.

6.2. Base-Promoted Carboxylation of Aromatic and Heteroaromatic Compounds with CO₂

One impressive example of improved energy efficiency in CO₂ reduction to CO comes from an electrochemical reduction process (96% selectivity) reported by Masel and co-workers by application of a cell overpotential of only 0.17 V on a 1-ethyl-3-methylimidazolium tetrafluoroborate (EMIM BF₄)-coated silver catalyst. Masel proposed CO₂ to react with EMIM to form a EMIM CO₂[BF₄] complex [131]. Analogous results were reported by Brennecke and co-workers [132] by using Pb electrodes and (EMIM)(Tf₂N) electrolyte. By spectroscopic characterization of the precipitate formed during electrolysis, Brennecke reported the formation of the NHC-CO₂ adduct shown in Scheme 55a. NHC-CO₂ adducts seems, thus, to play a significant role in electrochemical CO₂ reduction to CO at low overpotential. NHC-CO₂ adducts are well characterized species which chemistry has been developed after isolation of stable N-heterocyclic carbenes (NHC) (137) started by Arduengo in 1991 [133]. In Section 4.2, we have mentioned the carboxylation of 1,3-dialkyimidazolium chlorides with CO₂/Na₂CO₃ and the use of 1,3-dialkyimidazolium-2-carboxylates as CO₂-carriers in the carboxylation of C(sp³)-H acidic compounds. NHC-CO₂ compounds can be also synthesized by carboxylation of imidazol-2-yilidene precursors as reported by Kuhn in 1999 (Scheme 55b) [134] or by carboxylation of imidazol-2-ylidenes generated in situ by deprotonation of imidazolium salts (139) with potassium tert-butoxide or KN(SiMe₃)₂ as reported by Louie [135] and Delaude [136] (Scheme 55c).

Although the reaction is not involving CO₂, it is worth citing that, unexpected synthesis of 1,3-dimethylimidazolium-2-carboxylate was reported by Rogers et al. in 2003 [137] by reaction of 1-methylimidazole with dimethylcarbonate. Carrying the reaction at temperature lower than 95 °C affords selective formation of the imidazolium-2-carboxylate regioisomer (76) in 76% yield. At higher temperature (110 °C) the imidazolium-4-carboxylate regioisomer (141) forms in approximately 15% yield (Scheme 56).

The mechanism of this unexpected reaction was modelled via DFT by Crabtree in 2007 (Scheme 57) [138]. According to DFT analysis, N-methylation of 1-methyimidazole by dimethylcarbonate affords an imidazolium/CH₃OC(O)O⁻ ion pair (142). Proton abstraction from the imidazolium cation by CH₃OC(O)O^- generates the 1,3-dimethylimidazol-2-yilidene (71) that promptly reacts with CO₂. NHC-CO₂ adduct forms, thus, via a reaction pathway similar to the carboxylation reported by Louie and Delaude (Scheme 55c). Methanol formed during the reaction may be engaged into H-bonding stabilization of the carboxylate product (76).

Concerning the base-promoted carboxylation of heteroarenes, in 2010 Hu and coworkers reported the carboxylation of benzothioazole and benzoxazole derivatives under CO₂ (0.14 MPa) using Cs₂CO₃ (1.2 eq.) at the temperature of 125 °C in DMF solution (Scheme 58) [139]. In order to avoid decomposition of unstable carboxylic acids the products were converted in situ into stable esters with MeI or trimethylsilyldiazomethane (TMSCHN₂). Benzothiazole and benzoxazole derivatives bearing both electron-withdrawing and moderately electron-donating substituents at the phenyl ring (8 examples) gave the corresponding 2-carboxy methyl esters (143) in 68–97% yields. Excellent regioselectivity for carboxylation at the C(2)-position was registered. 5-Phenyloxyoxazole and 2-aryl-1,3,4-oxadiazole derivatives were also carboxylated (144) in good to high yield (8 examples, 55–96% yields).

Hu proposes the carboxylation to proceed via a mechanism different from the Kolbe-Schmitt reaction encompassing deprotonation of benzothiazole at the C(2) position with subsequent carboxylation of the 2-benzothiazolyl anion (145) (Scheme 59).

However, spectroscopic analysis of the reaction mixture did not provide evidence of 2-benzothiazolyl anion (145) formation. Therefore, it was concluded C-H cleavage to proceed via an uphill reaction step.

LiO^tBu was reported as active base in direct carboxylation of unprotected indole derivatives. (Scheme 60) [140]. Under optimized conditions (ambient CO₂ pressure, 100 °C) the use of a large excess of LiO^tBu was found (5 eq.) to suppress products decarboxylation. Indole derivatives with both electron-rich and -poor substituents at the C(5)-, C(4)- and C(7)-positions afforded selectively indole-3-carboxylic acid derivatives (146) with low to excellent yield (11 examples, 19–96% yields). 2-methyl-substituted indole afforded the carboxylic acid derivative in good yield (80%) while 2-phenyl-substituted indole was carboxylated in only 31% yield.

Later, Kobayashi published the extension of the carboxylation reaction to pyrrole derivatives [141].

Concerning the reaction mechanism, Kobayashi proposes substrate deprotonation to occur at the N-H proton affording indolyl anion (147) able to undergo subsequent carboxylation with CO₂ to form the 1-indolyl carbamate salt (148) (Scheme 61). As stated before the free N-H functionality was a required feature as N-CH₃ substituted indoles were inert towards carboxylation under reported conditions. The possibility for having a resonance structure with the negative charge delocalized at C3 position explains formation of the observed indole-3-carboxylic acid derivatives (146).

More recently, Ackermann reported efficient carboxylation of heteroarenes by using a strong base as KO^tBu (1.2 eq.) in DMF at 100 °C under atmospheric CO₂ pressure (Scheme 62) [142]. Benzoxazole, benzothiazole, oxazole and 1,3,4-oxadiazole derivatives were carboxylated selectively at the C(2) position (12 examples, 43–80% yields).

In analogy with the reaction mechanism proposed by Hu (Scheme 59), the reaction may proceed via initial reversible C-H deprotonation, followed by subsequent CO₂ carboxylation.

In 2016, Larrosa reported an advanced protocol for a Kolbe-Schmitt-type carboxylation reaction based on reacting PhOH (and phenol derivatives) with NaH under an atmospheric CO₂ pressure at the temperature of 185 °C [143] in the absence of organic solvent. The procedure avoided the isolation of the PhONa precursor and the undesired formation of H₂O.

Interestingly, the addition of 2,4,6-trimethylphenol (TMP, 153) as a recyclable additive had a beneficial effect on the reaction by increasing both the initial reaction rate and final yield (Scheme 63). Phenol derivatives possessing both electron-donating and halogen substituents were carboxylated with excellent regioselectivity affording ortho-carboxylated products (18 examples, 55–91% yields). Substrates as 3-methylphenol, 3-chlorophenol and 3-bromophenol afforded both C2- and C6-carboxylated regioisomers. However selectivity was significantly higher for the C6-carboxylated product.

The author proposes that sodium 2,4,6-trimethylphenoxide (153), formed in situ from 2,4,6-trimethylphenol and NaH, plays a role as CO₂-carrier. A large body of evidence suggests that only “fixed CO₂” (absorbed or adsorbed) by the solid reagent is involved in the carboxylation reaction, therefore 2,4,6-trimethylphenoxide may increase the amount of CO₂ able to react being inert towards ortho-carboxylation itself (Scheme 64).

Very recently Zhi reported a transition-metal-free carboxylation of C(sp²)-H bond with CO₂ followed by a lactonization reaction applied to 2-(imidazo[1,2-a]pyridine-2-yl)phenol (154, Scheme 65) and 2-(1-arylvinyl)phenol derivatives (82, Scheme 66) [144].

For reaction shown in Scheme 65 substrates bearing electron-donating and electro-withdrawing group substituents both at the phenoxy-(R¹) and pyridine-(R²)-moieties were functionalized in good to excellent yield (17 examples, 56–96% yields).

Concerning 2-alkenylphenol derivatives (82), mono-substituted substrates with electron-donor groups (R³) afforded the corresponding lactone (83) in moderate to good yield (4 examples, 38–71% yields).

A careful mechanistic investigation brought to propose that, under basic conditions, a carbonate (156) or a dicarbonate (157) potassium salt may form, that subsequently undergo cyclization to the lactone derivative (155) (Scheme 67).

Yu and coworkers succeded in obtaining the first example of lactamization of 2-alkenylanilines (158) with CO₂ to form 2-quinolinone derivatives (159) in high yields (26 examples, 42–95% yields) under atmospheric pressure (Scheme 68a) [145]. The 2-quinolinone motif is widely exploited in synthetic medicinal chemistry and photoelectric materials [146,147,148] and was previously accessed by the groups of Alper [149], Zhu [150] and Chuang [151] via lactamization of alkenyl and aryl C-H bonds with CO. Yu implemented, thus, a lactonization reaction using CO₂ founding that 2-alkenylanilines with R² = CH₃ (Scheme 68a) are more reactive than substrates with R² = H, Et. In contrast, both electron-donating and electron withdrawing substituents on the aniline nucleus do not affect significantly the substrate reactivity. Moreover, Yu has shown that also anilines with ortho-heteroarene substituents (Scheme 68b) as imidazo[1,2-a]pyridine (161a), pyrazole (161b) and pyrrole (161c) are suitable substrates for the reaction.

As aniline derivative 162 and urea 163 were isolated as reaction intermediates, Yu proposes the reaction mechanism shown in Scheme 69, encompassing formation of isocyanate 164 that undergo irreversible cyclization leading to 159.

6.3. Base-Promoted Carboxylation of Furan-2-Carboxylic Acid to Furan-2,5-Dicarboxylate

A capstone work reported by Kanan describes the use of molten Cs₂CO₃ in the carboxylation of furan-2-carboxylic (165) acid to furan-2,5-dicarboxylate (FDCA²⁻) (166) under CO₂ (pressure range from 0.1 to 0.8 MPa) at temperature ranging from 195 to 260 °C (Scheme 70) [60]. As shown in Scheme 70, at first furan-2-carboxylic acid (165) is converted into furan-2-carboxylate salt with Cs₂CO₃ (1.05 eq.) at the temperature of 150 °C. Then furan-2-carboxylate is converted into furan-2,5-dicarboxylate (166) with additional Cs₂CO₃ (0.55 eq.). Reacting 1 mmol of furan-2-carboxylate at 200 °C under P_CO2 = 0.8 MPa during 5 h afforded furan-2,5-dicarboxylate in 89% yield and 100% regioselectivity (detected decomposition products represented only 5%). Scale up of the reaction by using 10 or 100 mmol of the starting compound required a fine adjustement of the reaction parameters to give the dicarboxylate in 81% and 71% yield respectively.

Since 2-furoic acid can be readily made from lignocelluloses (Scheme 71), the reaction give access to furan-2,5-dicarboxylic acid (166) which has been shown an excellent substitute for terephthalic acid.

Assuming the pKa for the C-H at the 5 position of furan-2-carboxylic acid being similar to the pKa of unsubstituted furan a pKa value of ≈35 can be considered.

Considering the reaction shown in Scheme 70, the stoichiometry of the overall process can be written as shown in Scheme 72. By acidification of the obtained furan-2-carboxylate salt with HCl, furan-2,5-dicarboxylic acid and CsCl were obtained. By using bipolar membrane electrodialysis CsCl was converted into HCl and CsOH. At this stage, HCl could be recycled in the acidification step and CsOH could be reacted with 2-furanoic acid to regenerate furan-2-carboxylate. At the same time reaction of 2 CsOH with CO₂ gave Cs₂CO₃ and H₂O.

The stoichiometry of the “protonation with HCl and CO₃²⁻ regeneration steps” can be written as shown in Scheme 73.

Considering the overall process encompassing:

(i)

carboxylation of furan-2-carboxylic to furan-2,5-carboxylate;

(ii)

protonation of furan-2,5-carboxylate with HCl (to give furan-2,5-carboxylic acid);

(iii)

regeneration of HCl and Cs₂CO₃ from CsCl and CO₂.

The stoichiometry of the reaction can be written as shown in Scheme 74.

Moreover, a method for esterification of FDCA²⁻ to dimethyl furan-2,5-carboxylate (DMFD) by heating 1 mmol of caesium FDCA²⁻ to 200 °C in 100 mL anhydrous methanol under P_CO2 = 0.45 MPa for 30 min was also developed. The dimethyl furan-2,5-carboxylate was obtained in 50% yield.

The scope of the carboxylation reaction was extended to benzoate salt and benzene. The carboxylation of benzoate salt afforded phthalates (53% combined yields) besides tri- and tetra-carboxylates (13% combined yields) (Scheme 75).

As expected, benzene carboxylation required more drastic reaction conditions. 1.5 mmol of Cs₂CO₃ and 1 mmol of caesium isobutyrate (used as additive to obtain a molten mixture) were heated to 350–360 °C, under P_CO2 = 3.1 MPa and P_benzene = 4.2 MPa (Scheme 76). At the temperature of 350 °C benzoate, phthalates and benzene tricarboxylates were obtained in 12% combined yield, while at 360 °C the yield increased to 19% (yield was based on Cs₂CO₃).

Very interestingly Kanan notes that accessing carboxylates derivatives of substrates possessing unactivated C-H bonds open to the possibility to synthesize alcohols and hydrocarbons derivatives by combining carboxylation and hydrogenation reactions with use of renewable H₂.

7. C-H Carboxylation with CO₂ via Heterogeneous Catalysis

The role of homogeneous catalysis in carboxylation reactions of C(sp³)-H and C(sp²)-H bonds with CO₂ has been touched upon in previous Sections. With a view to developing at industrial scale C-H carboxylation reactions the use of active heterogeneous catalysts is highly desirable to allow easy catalyst recovery from the reaction mixture and subsequent recycle. However supported metal catalysts are often affected by unsatisfactory activity [152]. Recent advances in the field have been reported by Huang and coworkers by using a Shiff base-modified silver catalyst (169, Scheme 77b) employed in conjuction with Cs₂CO₃ to carboxylate terminal alkynes (167) to alkynyl carboxylic acids (168, Scheme 77a) [152] in high yields (86–98%) at ambient pressure.

The robust and green catalyst 169 was obtained by deposition of silver nanoparticles (NPs) on the solid support by direct reduction of AgNO₃ with a metastable aminal under ultrasonic conditions. The catalyst was fully characterized by powder X-ray diffraction (PXRD), X-ray photoelectron spectroscopy (XPS), and inductively coupled plasma (ICP) analysis showing excellent recyclability under reaction conditions in five runs with a total TON of 3430.

The reaction mechanism proposed by Huang (Scheme 78) involves coordination of the alkyne to Ag-NPs which is thought to increase C(sp)-H acidity allowing deprotonation by Cs₂CO₃ and formation of a silver acetylide intermediate (170). CO₂ insertion into C-Ag bond leads to the silver carboxylate (172) that in the presence of a large amount of Cs₂CO₃ and alkyne undergo transmetallation producing cesium carboxylate (173) and regenerating 170. Propiolic acid derivatives (168) are finally produced by acidification with concentrated HCl.

Furthermore, very recently porous MOF have been used as efficient heterogeneous catalysts in carboxylation of terminal alkynes with CO₂. Considering that MOF have emerged as promising porous materials with a wide range of applications encompassing CO₂ capture, separation, storage and regeneration as well as CO₂ photo- and electro-catalytic reduction, future directions in MOF chemistry point at their use in CO₂-epoxides co-polymerization (to make commodity polymers) and in CO₂-olefins co-polymerization [153].

Cheng and coworkers [154] have reported the synthesis of an efficient nanoscale heterogeneous catalyst obtained by introducing Ag NPs into the pores of MIL-101(Cr) (174) via a simple impregnation-reduction method to obtain Ag@MIL-101 (4.16 wt % Ag) (Scheme 79a). In optimized conditions, terminal alkynes are carboxylated at C(sp)-H bond by reacting 1 mmol of the substrate with 70 mg of Ag@MIL-101 (2.7 mol % Ag), in the presence of Cs₂CO₃ (1.5 mmol) under 0.1 MPa of CO₂ (at 50 °C in DMF) (Scheme 79b). Subsequent acidification with HCl affords propiolic acids derivatives. Interestingly, Cheng reports that, after catalysis, Ag@MIL-101 can be easily separated by centrifugation and reused at least five times without decrease of the catalytic activity.

The success of MOF as catalysts is thought to be linked to their ability to act as a microreactor with the terminal alkynes entering the channels and coordinating to Ag NPs. The mechanism proposed by Cheng and coworkers for carboxylation of alkynes is analogous to pathway shown in Scheme 78.

Bhaumick has also reported the synthesis of Ag NPs/Co-MOF (175) containing Ag in 4.40 wt % (Scheme 80b) [155]. The catalyst has been fully characterized by powder X-ray diffraction, thermogravimetric analysis, energy dispersive X-ray spectrometry and high-resolution transmission electron microscopy. By reacting 1 mmol of terminal alkynes with 50 mg of Ag NPs/Co-MOF (2 mol % Ag) in the presence of Cs₂CO₃ (1.5 mmol) under 0.1 MPa of CO₂ at 80 °C in DMF the corresponding propiolic acid derivatives were obtained in 84–98% yield (7 examples) (Scheme 80a). Analogously to Ag@MIL-101, Ag NPs/Co-MOF can be easily recovered by filtration from reaction solution and reused in up to six reaction cycles with slight loss of activity. ICP-MS analysis of the reused catalyst confirmed that Ag NPs are strongly bound to the framework and insignificant amount of Ag is lost in the filtrate.

The reaction mechanism proposed by Bhaumick for carboxylation of alkynes is analogous to previously reported pathways (Scheme 78).

8. Thermodynamics of C-H Carboxylation Reactions with CO₂

Due to high thermodynamic stability of the CO₂ molecule (ΔG°_f = −394.6 kJ/mol) C-H carboxylation reactions are endoergonic processes. ΔG°_react values for several biologically relevant C-H carboxylation reactions have been calculated by Faber and co-worker using the IEFPCM (H₂O)-G3MP2B3 method [156]. Selected values of ΔG°_react relevant to the topic of this review are reported in Scheme 81.

For thermal carboxylation reactions considered in above Sections thermodynamic constraints are overcome: (i) by using high energy co-reactants in stoichiometric amount (or even in excess) with respect to the organic substrate (overall stoichiometry of selected reactions is shown in Scheme 82a–d); (ii) by promoting C-H carboxylation reactions followed by a lactonization or lactamization step (overall stoichiometry of selected reactions is shown in Scheme 82e–f).

With reference to light-driven carboxylation reactions discussed in Section 3.1 and Section 3.2, the proposed one electron reduction of CO₂ to CO₂^•− is affected by an extremely negative E°′ redox potential of −1.90 V versus NHE (under standard conditions in aqueous solution at pH = 7, 25 °C, 0.1 MPa of gases and 1M solutes) as can be seen in Figure 1 (red points, Equation (1)) [157]. This redox potential is even more negative (−2.21 V vs. SCE) in dry dimethylformamide [158]. Multi-electron and multi-proton reduction of CO₂ in acqueous solutions (Equations (2)–(6), Figure 1) are affected by a lower thermodynamic barrier.

Jamison [81] noted that para-terphenyl (E° = −2.63 V vs. SCE in DMF) could overcome this high energy barrier as previous work by Yanagida [159] report use of this photoredox catalyst for the reduction of CO₂ to formic acid under UV irradiation. Macyk also declares use of ZnS as a wide band gap semiconductor [82] with a E° in the range of −1.8 to −2.0 V vs. NHE previously used in photocatalytic water splitting and CO₂ reduction [160,161].

In summary, C-H carboxylation reactions using CO₂ are endoergonic processes requiring an energy input that can be supplied in the form of: (i) high energy co-reactants; (ii) electrons or irradiation; (iii) energy employed in electrodialysis to regenerate bases. As stated by Aresta [1] intelligent use of renewable energy sources (solar, wind and other renewable) and the analysis of the complete CO₂-balance will allow selecting environmentally sustainable processes that reduce (or avoid) the use of fossil fuels.

9. Conclusions

The above reported literature analysis shows that catalytic and stoichiometric reactions giving straightforward access to carboxylation of C-H bonds with CO₂ are a high appealing research field addressed by the research community. Transition-metal catalyzed carboxylation reaction of unactivated C(sp³)-H bond reported by Mita, Sato and Hou as well as photo-catalyzed carboxylation reported by Jamison provide new and promising routes in catalysis. Base-mediated carboxylation reactions reported by Jessop, Beckmann and Kanan represent also a significant advance in the field of metal-free conversion processes encompassing base-recycling or base regeneration.