CH Carboxylation of Terminal Alkynes Catalyzed by Low Loadings of Silver(I)/DMSO at Ambient CO₂ Pressure

Terminally active: A reaction protocol that permits the CH carboxylation of alkynes at low temperatures, with mild bases, at ambient CO2 pressure, in the presence of ppm-quantities of an inexpensive silver catalyst. Carbon dioxide is arguably the most attractive C1 building block for organic synthesis. Dry air contains 0.039 % CO2, and every year, human activity releases 29 billion tons of this greenhouse gas into the atmosphere.1 It would be desirable to instead utilize at least a small part of the CO2 for chemical value creation.2 However, this is a great challenge, owing to its high thermodynamic stability and low reactivity.3 Traditionally, CO2 could thus be activated only with reactive organometallic species or strained ring systems.2 In the last decades, several catalytic processes for CO2 fixation have emerged that allow for example, the carboxylation of organometallic reagents, epoxides, the carboxylative cyclization of propargylic amines and alcohols, the reductive carboxylation of alkynes and alkenes, and the hydrosilylation of CO2 under mild conditions (Scheme 1).3, 4 A substantial advance was recently made in the development of direct CH carboxylations mediated by coinage metals.5–7 Overview of direct carboxylation processes. In 2010, Nolan and co-workers disclosed the first catalytic carboxylations of NH and CH bonds in the presence of hydroxide bases to form heterocyclic carboxylic acids.6 They used copper(I) or gold(I) hydroxides ligated with 1,3-bis(diisopropyl)phenylimidazol-2-ylidene as catalysts. The problem associated with catalytic carboxylations of terminal alkynes is that the products are thermally so unstable that with most catalysts, high yields can only be obtained when trapping the products by in situ esterification.8 Our group and that of Zhang disclosed the first catalysts that operate at such low temperatures that the free propiolic acids can be isolated.9 We employed (4,7-diphenyl-1,10-phenanthroline)bis[tri(4-fluorophenyl)phosphine]-copper(I) nitrate ([CuI]) as the catalyst,9a and Zhang described the use of N-heterocyclic carbene-copper complexes and copper(I) N,N,N′,N′-tetramethylethylenediamine complexes (Scheme 2).9b Lu et al. recently disclosed a silver(I) iodide-catalyzed protocol.10 In all three protocols, only mild bases and low reaction temperatures are required. CH carboxylation of terminal alkynes. Both CH acidic heterocycles11 and terminal alkynes12 can be carboxylated also in the absence of a catalyst, but only at substantially higher temperatures and increased CO2 pressures. As in the metal-catalyzed cases, alkali carbonates, usually Cs2CO3, were employed as the base. The propiolic acids produced in these transformations are valuable synthetic intermediates. They serve as substrates for cycloaddition, hydroarylation and decarboxylative cross-coupling reactions,13 giving access for example, to coumarins, flavones, aminoalkynes, alkynylarenes, and arylidene-oxindoles.14 The concept of a CH carboxylation of terminal alkynes compares favorably with traditional synthetic entries, for example, the addition of alkynes to formaldehyde and subsequent oxidation,15 the reaction of alkynylmetal species with CO2 or chloroformate,16 and the oxidative carbonylation of alkynes.17 A substantial improvement of catalyst activity, however, is needed before carboxylative CH activations of alkynes can compete with the established syntheses of arylpropiolic acids on laboratory, and especially on industrial scale. We herein disclose a reaction protocol that fulfills these requirements by permitting the CH carboxylation of alkynes at low temperatures, with mild bases, and at ambient CO2 pressure in the presence of ppm-quantities of an inexpensive silver catalyst. The carboxylation of 1-octyne (1 a) was used as a model reaction to systematically study the influence of various parameters on the reaction outcome (Table 1). At ambient CO2 pressure, the performance limit of the copper(I)- or silver(I)-based literature systems was reached when the catalyst loading was reduced to 2500 ppm (entries 1 and 2). The yield could not be improved by variation of the ligand, but we detected a surprisingly strong dependence on the reaction solvent (entries 3–8). Simple copper(I) or silver(I) salts gave high conversions specifically in DMSO as the solvent, or at least a co-solvent (entry 9). At lower DMSO concentrations, the conversion is reduced, indicating that its solvent properties contribute to its beneficial effect (entry 10). A gold(I) complex also showed detectable conversion (entry 15). Entry Catalyst [ppm] Solvent Conversion [%] Yield [%] 1 2500 [CuI] DMF 31 24 2 2500 AgI DMF 86 82 3 2500 [CuI] DMSO 54 46 4 2500 AgI DMSO 100 99 5 2500 AgI MeCN 51 10 6 2500 AgI PC 64 50 7 2500 AgI sulfolane 71 65 8 2500 AgI H2O 26 0 9[b] 2500 AgI DMF/DMSO 100 95 10[c] 2500 AgI DMF/DMSO 93 92 11 1250 Ag2O DMSO 100 94 12 2500 AgNO3 DMSO 100 98 13 2500 AgOAc DMSO 100 96 14 2500 AgBF4 DMSO 100 98 15 5000 [AuI] DMSO 9 6 16 500 AgBF4 DMSO 100 99 17 250 AgBF4 DMSO 80 74 18[d] 250 Ag2O DMSO 100 98 19 – DMSO 0 0 20[e] 500 AgBF4 DMSO 100 93 21[f] 500 AgBF4 DMSO 90 83 The nature of the silver counter-ion was found to have only a limited effect (entries 11–14). The high solubility of AgBF4 allows the preparation of stock solutions useful particularly on small scale. The inexpensive but less soluble silver(I) oxide is advantageous for preparative-scale reactions (entry 18). At 50 °C or above, 500 ppm of silver(I) are sufficient to achieve full conversion in overnight reactions (entries 16, 20, 21), and when reducing the amount to 250 ppm, 80 % conversion were still achieved which corresponds to a turnover number of ≈3000 (entry 17). A control experiment revealed that under such mild conditions, CH carboxylation does not proceed without a metal catalyst (entry 19). Having thus identified a catalyst system with a new level of efficiency for the carboxylation of alkyl- and aryl-substituted alkynes, we tested the generality of the new protocol by applying it to various terminal alkynes (Table 2). In many cases, full conversion was achieved using 500 ppm of the catalyst within 16 h. For some substrates, especially aryl alkynes, the conversion was incomplete after this time. However, increasing the catalyst loading to 0.25 mol % led to near full conversion also for these substrates. The propiolic acids were isolated in high yields and purity after a simple acidic workup. If desired, alkyl esters can directly be accessed by in situ alkylation. Thus, in method B, 1-bromohexane was added at the end of the reaction. This also allowed isolating ester derivatives of particularly unstable propiolic acids (3 d, 3 g, and 3 r). Product Yield [%] Product Yield [%] Product Yield [%] 3 a 99(94) 95;[c] 88[e] 3 b 95(99) 3 c 99(97) 85,[d] 90[f] 3 d 21(52)[b] 3 e 93(83)[b] 3 f 95(97) 74[e] 3 g 70(99) 96[e] 3 h 98(93) 3 i 99(97) 85;[c] 74[e] 3 j 97(99) 73;[c] 90[e] 3 k 99(91)[b] 62[c] 3 l 92(99)[b] 87[c] 3 m 97(96)[b] 99;[c] 96[f] 3 n 98(91)[b] 98;[c] 95[f] 3 o 99(96)[b] 89;[d] 90[f] 3 p 99[b] 73;[c] 81[d] 3 q 98(99)[b] 99;[c] 83[e] 3 r 40(92)[b] 0(75)[c] 3 s 96(95)[b] 3 t 97(99)[b] 86;[c] 86[d] 3 u 99(96)[b] 86;[d] 92[f] Most functional groups, including halogens, ethers, tertiary amines and trifluoromethyl groups, were tolerated. The isolated yields compare favorably with those previously reported with catalyst loadings of 1–7 mol %. The unique catalyst performance observed in DMSO, as well as the surprisingly small influence of the silver counter-ion, are best explained by a mechanism in which the catalytically active species is a DMSO-ligated silver(I) carbonate formed in situ from the silver salt, DMSO, and caesium carbonate (Scheme 3). [Ag(DMSO)1−2]+X− structures are known to form upon dissolving silver salts in DMSO.18 DMSO is known to act as an oxygen donor to AgI.19 The unique efficiency of DMSO as a reaction solvent may partially be explained by the higher solubility of CO2 in DMSO than in other solvents.20 There are also several reports that silver nanoparticles form when heating silver salts in DMSO.21 Although we never observed precipitation of the silver, it cannot be excluded that such silver nanoparticles are involved in the catalytic cycle. Proposed reaction mechanism. The carboxylation is likely to follow the catalytic cycle proposed by Lu.10 The reaction of silver salts or nanoparticles with a terminal alkyne is likely to afford a silver acetylide along with caesium bicarbonate.22 After insertion of CO2 into the CAg bond,23 caesium propiolate would then be released via salt metathesis with caesium carbonate, regenerating the initial silver species. Further mechanistic studies of this convenient and high-yielding propiolic acid synthesis are underway. A 10 mL vessel was charged with caesium carbonate (399 mg, 1.20 mmol). Under an atmosphere of carbon dioxide, a solution of silver tetrafluoroborate (0.01 mg, 0.50 μmol) in degassed DMSO (3.00 mL) was added, and the mixture was stirred at room temperature for 5 min. After purging the reaction vessel with CO2, the alkyne 1 a–u (1.00 mmol) was added via syringe. The resulting mixture was stirred for 16 h at 50 °C at ambient CO2 pressure. At the end of the reaction time, the mixture was cooled down to room temperature and the solvent was removed by lyophilization. The resulting salt was dissolved in H2O (10.0 mL) and extracted with n-hexane (3×20.0 mL). Then the aqueous layer was acidified with aqueous HCl (1N, 2.00 mL) and extracted with ethyl acetate (3×20.0 mL). The combined organic layers were washed with brine (10.0 mL), dried over MgSO4, filtered and the volatiles were removed under vacuum to afford the corresponding acids 3 a–u. For full experimental procedures, see the Supporting Information. We thank BASF SE for funding and S. Schäfer, A. W. Jones, C. Matheis, M. Weiland, and S. Brühl for technical support. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.