Formal C−H Carboxylation of Unactivated Arenes

Abstract A formal C−H carboxylation of unactivated arenes using CO2 in green solvents is described. The present strategy combines a sterically controlled Ir‐catalyzed C−H borylation followed by a Cu‐catalyzed carboxylation of the in situ generated organoboronates. The reaction is highly regioselective for the C−H carboxylation of 1,3‐disubstituted and 1,2,3‐trisubstituted benzenes, 1,2‐ or 1,4‐symmetrically substituted benzenes, fluorinated benzenes and different heterocycles. The developed methodology was applied to the late‐stage C−H carboxylation of commercial drugs and ligands.


Introduction
The last two decades have witnessed an exponential growth in the field of directc arbon-hydrogen (CÀH) bond functionalization. An umber of challenging carbon-carbon (CÀC) and carbon-heteroatom (CÀX) bond forming reactions can now be realized by directt ransitionm etal-catalyzed CÀHa ctivation. [1] Well-established CÀHa ctivations,w hich operate on unactivated systemsw ith good functionalg roup tolerance, can be applied to the late-stage substitutiono fv aluable and ratherc omplex systems, such as commercial drugs and natural products. [2] Ah oly grail in this field is the developmento fp rotocols that allow the direct carboxylation of CÀHb onds with CO 2 ;asustainable and fossil-free carbon source. [3] The resulting products, carboxylic acids and their derivatives, are widespread structural motifs in commercial drugs and natural products. [4] The use of CO 2 as ac arboxylating agent in CÀHf unctionalizations is also attractive for the late-stage isotopicl abeling of pharmaceuticals and other biologically activemolecules. [5] Despitec onsiderable progress in the field, knownp rotocols for CÀHf unctionalization with CO 2 still have pronouncedl imitations. Most of them are workings electively only on activated molecules. [3a] Nolan, [6a,b] Hou [6c] and co-workersh ave reported good regioselectivities for carboxylation of oxazolesa nd perfluorinated arenes (Scheme 1A); however,t his protocoli sl imited to activated aromatic systems with acidic CÀHb onds.F or unactivated systems, Iwasawa et al. found that Rh-catalyzed CÀHc arboxylationsc an provide good regioselectivities, but only in the presence of nitrogen-based directing groups (Scheme 1B). [7] Practical applications of directedC ÀHf unctionalizations are limited by the fact that directing groupsm ay not be removable or modifiable. [8] The aim of the work described here was to establisha methodf or regioselective CÀHc arboxylationo fu nactivated arenes with focus on substrates that are not reactive in currently knownp rotocols. [3a, 6, 7] We envisioned that the known reactivity and selectivity issues in the CÀHc arboxylation of unactivated arenes may be overcome by applyingasequential Ir/ Cu-catalyzed CÀHt ransformation strategy:R egioselective CÀH activation may be achieved throughasterically controlled Irmediated CÀHb orylation, [9,10] which we hypothesized could be followed by aC u-catalyzed carboxylation of the in situ generated organoboronates (Scheme 1C). If successful, our strategy would provide new opportunitiesf or formal CÀHc arboxylation of real-life systems, such as pharmaceuticals. Besides using CO 2 as as ustainable carboxylating agent, we furtherd ecidedt o ensure that our procedure would be applicable in green solvents, making our protocolr elevant in the context of green chemistry.
Carbonates were not suitable as solvents for the formal CÀH carboxylation ( Table 1, entry 1), even thought hey were the best solvents for the carboxylation of the intermediate organoboronate (Supporting Information, Ta ble S1-S4). Stepwise analysis of the reaction showedt hat the Ir-catalyzed CÀHb orylation of r2 is not working in carbonates (Scheme 2A). Ap ossible explanation may be that carbonates are being reduced by B 2 pin 2 /HBpin (bis(pinacolato)diboron/pinacolborane), thus consuming the reagent of CÀHb orylation. [14]  In contrast to Ir-catalyzed CÀHb orylation in carbonates as solvents, borylation of r2 in methylal leads to the corresponding organoboronate b2 in 65 %i solated yield (Scheme 2A). The followingC u-catalyzed carboxylation of b2 in methylalp rovided p2 in 90 %i solated yield (Scheme 2B). Application of at wosolvents ystem, applying methylali nt he CÀHb orylation step and exchangingt he solventt oc arbonates (DEC or DMC) for the carboxylation step,gave lower yields than the reaction performed using only methylal as solvent (Table 1, entry 3v ersus 4a nd 7). The yields for the two-solvents ystem can be slightly improved (72 %) by extending the initial CÀHb orylation in methylalto3 6h (Table 1, entry 5).
[a] The catalystwas generated in situ. [b] If not otherwise mentioned, the reaction was performedi nmethylal.
The influence of solvent on the outcomeo ft he carboxylation of arylboronic acid pinacole sters showed that the best yields are achieved in DEC, although in some cases the yield differences between the solvents were negligible (p1, p2, Scheme 5). Thiophene-2-boronica cid pinacol ester b25 and b30 were the single exception, providing best results in methylal (Scheme5,S upporting Information, Scheme S2). However, note that for the sequential Ir/Cu-catalyzed formal CÀHc arboxylation, the CO 2 -derived methylal proved to be the best, as evaluated fors everalr eactions (p2, p9, p23,S cheme 3a nd 4, Supporting Information, Table S4).
The exceptional substrate scopem ainly based on unactivated arenesand the excellentf unctional group tolerance allowed us to use the formal CÀHc arboxylationf or the late-stage functionalization of complexa nd practicallyv aluable systems, such as commerciald rugs and naturalp roducts (Scheme 6). [2,10,15] For example, we could carboxylate the natural product guaiazulene r39 (cosmetic ingredient) with 46 %y ield of p39 as a mixtureo fr egioisomers (10:~4). The commercial drugs praziquantel r40 (worm treatment) and clofibrate r41 (lipid-lowering agent)w ere carboxylated to provide, respectively, p40 (40 %) and p41 (34 %) in decent yields and with better regioselectivity( 10:~1). Observed regioselectivities are similart op reviously reported late-stage functionalizations of guaiazulene, praziquantel and clofibrate, which in all cases provided mixtures of regioisomers. [10e, 15e] The developed methodology was further evaluated for the late-stage carboxylation of organometallics and phosphine ligands (Scheme 6, p42 to p44), as the generation of carboxylated organometallics and phosphines could be highly relevant to the production of water-soluble homogeneous catalysts. [16] The formal CÀHc arboxylation of ferrocene r42 provided p42 in 37 %y ield, as as ingle product. All attempts to introduce a carboxyl group directly into unprotectedp hosphines failed, however,p hosphine oxidesw ere successfully carboxylated by our method.F or triphenylphosphine oxide r43 and diphenyl(cyclohexyl)phosphineo xide r44,t he reaction gave ar egioisomeric mixture of carboxylated phosphine oxides p43, p44 in 61 %a nd 67 %y ields, respectively.T ot he best of our knowledge, the late-stage CÀHf unctionalization of phosphine ligands hasnot been described before.

Conclusions
We have developed ar obusta nd versatile strategy for af ormal CÀHc arboxylation of unactivated arenes.T he present strategy consists of Ir-catalyzed CÀHb orylation and subsequent Cu-catalyzed carboxylation of in situ generated organoboronates. The protocol does not require any workup or purification duringt he two steps. The formal CÀHc arboxylation reaction proceeds with remarkable regioselectivity for 1,3-disubstituted and 1,2,3-trisubstituted benzenes, 1,2-and 1,4-symmetrically substituted benzenes, fluorinated benzenes and several heterocycles. The developed methodology shows excellent functional group tolerance and can be appliedf or the late-stage CÀHf unctionalization of commercial drugs and ligands. Thus, the present protocol has capacityf or creatings tructurally diverse molecularl ibraries for modern medicinal chemistrya nd drug discovery, avoiding parallel de novo synthesis.
Evaluation of ar ange of green solvents showed that the formal CÀHc arboxylation can be conducted in CO 2 -derived solvents, which perform better than commono rganic solvents for these reactions. We believe that the present methodology will open anew chapter for the application of CO 2 as asustainable carboxylating agent in medicinal chemistry,materialsciences and catalysis.
Other renewable solvents like 2MeTHF,d iethoxymethane or methylal can replace Et 2 Ow ithout any noticeable difference (the difference was in the range AE 3%). Similarly,s aturated solution of NaHCO 3 can be replaced by 2 m solution of KOH.