COMMUNICATION "This is the peer reviewed version of the following article: Angew. Chem. Int. Ed. 2017, 56 (23), 6558-6562 , which has been published in final form at http://onlinelibrary.wiley.com/doi/10.1002/anie.201702857/abstractThis article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving." Switchable Site-Selective Catalytic Carboxylation of Allylic Alcohols with CO2 Manuel van Gemmeren, † Marino Börjesson, † Andreu Tortajada, † Shang-Zheng Sun, † Keisho Okura and Ruben Martin* Abstract: A switchable site-selective catalytic carboxylation of allylic alcohols has been developed in which CO2 is used with dual roles, both facilitating C–OH cleavage and as C1 source. This protocol is characterized by its mild conditions, absence of stoichiometric organometallic reagents, broad scope and exquisite regiodivergency that can be modulated by the ligand employed. C–O bond strength sulfonates esters & carbamates ethers OR O O O S R O alcohols OH R O availability Metal-catalyzed cross-coupling reactions of C–O electrophiles have recently gained considerable momentum due to the ready availability of alcohols and the design of orthogonal techniques in the presence of organic halides.[1] Unlike common C–C bondformations with activated organic sulfonates, esters, and ethers, the use of simple alcohols as counterparts, arguably the most accessible and simplest C–O derivative, has received much less attention. Despite advances realized in nucleophilic/electrophilic regimes (Scheme 1, bottom left),[2],[3] alcohols should ideally be utilized within the realm of cross-electrophile couplings,[4] thus avoiding the need for highly reactive nucleophilic entities. However, the remarkable activation barrier required for C–OH scission and the high polarizability of the O–H bond[5] left a reasonable doubt that such a counterintuitive scenario could ever be implemented (Scheme 1, bottom right). If successful, however, the realization of such a technique could unravel the inherent potential of simple alcohols in cross-electrophile events, constituting a unique opportunity to provide new tactics for building up complexity from simple precursors. [*] [†] M. van Gemmeren, M. Börjesson, A. Tortajada, S. –Z. Sun, K. Okura, Prof. R. Martin Institute of Chemical Research of Catalonia (ICIQ) The Barcelona Institute of Science and Technology Av. Països Catalans 16, 43007 Tarragona (Spain) E-mail: rmartinromo@iciq.es Prof. R. Martin ICREA, Passeig Lluïs Companys, 23, 08010 Barcelona Spain These authors contributed equally to this work Supporting information for this article is given via a link at the end of the document.((Please delete this text if not appropriate)) classical cross-coupling δ– nucleophile ample precedents path a δ+ δ+ OH electrophile “unconventional” cross-coupling electrophile path b unprecedented no R–[M] Scheme 1. Catalytic C–C Bond-Formation with Alcohols as Coupling Partners Prompted by the seminal stoichiometric work of Osakada,[6] we[7] and others[8] launched a program aimed at unraveling the potential of nickel-catalyzed reductive carboxylation of organic (pseudo)halides with carbon dioxide (CO2). At present, these endeavors unfortunately remain limited to organic halides or particularly activated C–O electrophiles.[9] As C–O electrophiles ultimately derive from the corresponding alcohol, we recently questioned whether we could establish a design principle for promoting a more atom- and step-economical technique via direct Ni-catalyzed carboxylation of simple unprotected alcohols in the absence of stoichiometric or air-sensitive organometallic reagents.[10] To such end, we hypothesized that CO2 could be used with dual roles, both as C1 source and as an activating group for C–O bond-cleavage (Scheme 2). Our hypothesis is driven by the known ability of CO2 to react reversibly with alcohols en route to carbonic acids I,[11] thus lowering the activation energy for promoting C–O scission while accelerating the rate of oxidative addition to Ni(0)Ln (III vs II) prior CO2 insertion.[12] Herein, we describe our results towards this goal by designing a mild, and user-friendly Ni-catalyzed carboxylation of allylic alcohols under atmospheric pressure of CO2. We demonstrate that the site-selectivity at the allyl terminus can be discriminated and modulated at will by the ligand employed, leading to either linear or a-branched carboxylic acids.[13] COMMUNICATION R OH L Ni catalyst slow strong C–O bond R Ni OH C5H11 II CO2 OH CO2H R α-branched acid entry Deviation from standard conditions[a] CO2H 2a or CO2H Zn, MgCl2 (1.20 equiv) CO2 (1 atm), DMF 40 ºC (16 h) 1 CO2 facilitates C–O cleavage C5H11 NiBr2·glyme (10 mol%) L3 (26 mol%) C5H11 2b yield 2a + 2b (%)[b] 2a:2b[b] O OH I Ni catalyst fast weak C–O bond L R Ni III OR CO2 switchable selectivity mild & broad scope CO2 with dual roles no R–[M] reagents R CO2H linear acid Scheme 2. Design Principle for a Carboxylation of Allylic Alcohols with CO2 Our study began by evaluating the carboxylation of 1 at atmospheric pressure of CO2. Interestingly, neither 2a nor 2b were observed under conditions previously developed for other carboxylation events.[7,8] After some experimentation, a combination of NiBr2·glyme, commercially available L3, Zn, MgCl2 and DMF at 40 ºC afforded 2a in 70% isolated yield with exquisite site-selectivity (2a:2b = 99:1).[14] Inferior results were consistently observed with Ni sources, additives and solvents other than NiBr2·glyme, MgCl2, and DMF (entries 2 and 3). Similarly, lower yields were found when MgBr2 or ammonium salts were used in lieu of MgCl2 (entries 4 and 5).[15] Traces of 2a, if any, were observed when using Mn or silanes as reducing agents (entry 6). The former finding is particularly noteworthy given the notorious success of Mn as reductant in crosselectrophile couplings,[4] including carboxylation protocols.[7,8] As expected, the nature of the ligand exerted a profound influence in both reactivity and site-selectivity. While excellent regioselectivities towards 2a were found with 2,2’-bipyridines and 1,10-phenanthrolines possessing substituents adjacent to the nitrogen atom, better yields were found for the latter (entries 7-11). At present, we hypothesize that these results account for the higher rigidity of 1,10-phenanthrolines, likely increasing the stability of the putative reaction intermediates within the catalytic cycle. Interestingly, bench-stable C6-substituted terpyridines L7L8, resulted in a selectivity switch en route to 2b under identical reaction conditions to that shown for L3, albeit in lower yields (entry 12 and 13).[16] It is worth noting that related quaterpyridine analogue such as L6 failed to provide 2b.[17] These results tacitly suggest that the coordination geometry of the ligand dictates the selectivity pattern, with tridentate ligands being particularly suited for 2b. Careful fine-tuning of the reaction conditions afforded 2b in 81% isolated yield with an excellent 3:97 2a:2b ratio when using Ni(COD)2 as catalyst in DMA at rt, and employing both CaCl2 and NEt3 as additives (entry 14).[18] As anticipated, rigorous control experiments revealed that all of the reaction parameters were critical for success.[14] 99:1 Using NiCl2·glyme (Ni(COD)2) 58 (69) 99:1 3 DMA (NMP) as solvent 51 (52) 99:1 4 MgBr2 instead of MgCl2 52 99:1 5 nBu4NBr (nBu4NOAc) instead of MgCl2 0 (48) 99:1 6 Mn (Et3SiH) instead of Zn 3 (0) 99:1 7 L1 instead of L3 3 66:34 L2 instead of L3 42 99:1 9 L4 instead of L3 54 99:1 10 L5 instead of L3 23 99:1 11 O 70,[c] 26[d] 8 R L6 instead of L3 0 - 12 L7 instead of L3 18 23:77 13 L8 instead of L3 14 L8 instead of L3, with CaCl2 and NEt3 1 none 2 R2 30 R2 R1 3:97 R2 N N N R1 4:96 55, 81[c,e] N R R2 N N R N R1 R1 = H; R2 = Ph, L1 R1 = Me; R2 = H, L2 R1 = Me; R2 = Ph, L3 R1 = n-Bu; R2 = Ph, L4 R = n-Bu, L5 R = 2-py, L6 R1 R1 = Me; R2 = H, L7 R1 = R2 = Me, L8 Scheme 3. Optimization of the Reaction Conditions. 1 (0.20 mmol), NiBr2·glyme (10 mol%), L3 (26 mol%), Zn (4 equiv), MgCl2 (1.20 equiv), DMF (0.10 M), CO2 (1 atm) at 40 ºC for 16 h (yields in parentheses refer to yields of the reagents used in parentheses). standard. [c] Isolated yield. [d] [b] No MgCl2. GC yields using anisole as internal [e] Ni(COD)2 (10 mol%), L8 (10 mol%), CaCl2 (4 equiv), NEt3 (3 equiv), Zn (1.50 equiv), DMA (0.20 M) at rt. COMMUNICATION R2 linear acid R1 R2 OH R3 Ni/L3 R1 CO2H R3 OH CO2 R2 R1 2a-14a α-branched acid CO2H R2 1 R R3 2b-14b Ni/L8 or CO2 R3 conditions A (Ni/L3)[a] Me CO2H CO2H CO2H 2a, 70%,[b] 63%[c] 4a, 64%[d] 3a, 68% Me Ph CO2H Ph CO2H CO2H 7a, 66%[d] 6a, 66% 5a, 63% R O R CO2H O CO2H R = OMe, 8a, 63%[d] R = OPiv, 9a, 71%[d] Cl CO2H Me R = CF3, 11a, 64%[d] R = CO2Et, 12a, 57%[d] Me 10a, 60%[d] CO2H Me Ph CO2H Me CO2H Me 14a, 66%[e] 13a, 65%[d] conditions B (Ni/L8)[i] Me 15a, 76%,[f] 60%,[g] 68%[h] CO2H CO2H CO2H 2b, 81%,[b] 81%[c] 3b, 67%[j] Ph 4a, 71% CO2H CO2H Ph CO2H 5b, 76% Me 6b, 82% 7b, 70% R CO2H R O Me O CO2H 10b, R = OMe, 8b, 64% R = OPiv, 9b, 66% Cl CO2H CO2H R = CF3, 11b, 73% R = CO2Et, 12b, 73% 80%[k] Me CO2H Ph Me Me CO2H Me O C5H11 70% 15b, 72%,[l] 65%,[m] 65%[n] 14b, 92% 13b, 65% CO2 by using a Ni/L3 or Ni/L8 couple (Scheme 4). In all cases analyzed, an exquisite site-selectivity was observed regardless of whether linear or a-branched allylic alcohols were utilized, thus indicating that a substrate-controlled site-selectivity does not come into play. Notably, the substitution pattern on the allyl terminus was largely inconsequential for both reactivity and siteselectivity, obtaining predominantly the corresponding linear or a-branched carboxylic acids. The presence of esters (12a, 12b), acetals (10a, 10b), alkenes (15a, 15b) or even aryl chlorides (13a, 13b) do not interfere, providing ample room for further functionalization. Likewise, the presence of C–O electrophilic sites susceptible to Ni-catalyzed cross-coupling reactions such as aryl methyl ethers (8a, 8b) or aryl pivalates (9a, 9b) did not compete with the efficacy of the carboxylation event. Particularly noteworthy was the ability to access quaternary centers (14b and 15b), as cross-electrophile endeavors have found little echo with tertiary alkyl electrophiles.[19] While a Ni/L3 regime provided predominantly (E)-configured isomers, we observed an intriguing dichotomy exerted by the geometry of the starting precursor. Specifically, naturally-occurring geraniol gave rise to 15a with a higher E:Z ratio than structurally isomeric nerol and linalool. While tentative, a subtle h1 to h3 hapticity of the intermediate allyl-Ni complex or the known ability of tethered alkenes on the side-chain to act as intramolecular directing groups might account for these results.[20] In contrast, 15b was exclusively obtained from either geraniol, nerol or linalool with a Ni/L8 couple. It is worth mentioning that substituents in the a-position of the allyl motif did not erode the site-selectivity, giving rise to 10a and 10b in good yields, respectively. The synthetic applicability of our Ni-catalyzed carboxylation in the context of natural product synthesis is further illustrated in Scheme 5. Specifically, a simple three-step sequence consisting of a direct carboxylation of 1a followed by esterification with TMSCHN2 and Sharpless asymmetric dihydroxylation afforded 16 in good overall yield, constituting a formal total synthesis of (+)-transcognac lactone[21] 17 and other related natural products.[22] Taken together, the results compiled in both Schemes 4 and 5 shows the prospective potential of our methodology. X O (a) TMSCHN2 (b) AD-mix α 58% yield X = OH (1) X = CO2H (2a) C5H11 16 O O ref. 20 OH er = 97.5:2.5 C5H11 17 Me (+)-trans-cognac lactone Scheme 5. Formal Total Synthesis of (+)-trans-Cognac Lactone. Scheme 4. Scope of the Regiodivergent Carboxylation of Allylic Alcohols with CO2. [a] Conditions A: as scheme 3, entry 1, at 0.25 mmol scale; Yields of isolated products, average of at least two independent runs. 2a-15a were all obtained in 99:1 (linear:branched) ratio. 2a-13a were obtained in ≥92:8 E:Z ratio. [b] from linear alcohol. DMF (0.70 M). 16:84. [h] [e] [c] E:Z = 45:55. from a-branched alcohol. [f] [d] from geraniol, E:Z = 94:6. from linalool, E:Z = 42:58. [i] Zn (2.50 equiv) in [g] from nerol, E:Z = Conditions B: as scheme 3, entry 14, at 0.25 mmol scale; Yields of isolated products, average of at least two independent runs. With the exception of 3b and 10b, all products were obtained in ≥92:8 (branched: linear) ratio. and 1:1 diastereomeric ratio. [l] [j] 82:18 (3b:3a). from geraniol. [m] from nerol. [k] [n] 85:15 (10b:10a) from linalool. With a reliable access to either 2a or 2b, we next turned our attention to evaluate the generality of our Ni-catalyzed switchable site-selective carboxylation of allylic alcohols with While an unambiguous understanding on how these reactions work at the molecular level should await further investigations, we turned our attention to gather empirical evidence about the intriguing role of the ligand by studying the reactivity of the putative Ni(0)Ln species (Scheme 6). Particularly noteworthy was the absence of reactivity when simply exposing 18 to stoichiometric amounts of either Ni(0)(L3)2[23] or Ni(COD)2/L8[24] in the absence of CO2, thus reinforcing the notion that a direct oxidative addition of an allylic alcohol to Ni(0)Ln does not come into play.[25] More importantly, while a Ni(0)(L3)2 regime required the presence of Zn en route to 19a,[23] the corresponding Ni(COD)2/L8 couple cleanly produced 19b in the absence of reductant, thus evidencing the unique role exerted by the ligand COMMUNICATION backbone.[26] Although speculative, these results might suggest the intermediacy of h1-Ni(I) intermediates with bidentate and rigid L3, in which CO2 insertion takes place at the a-carbon (19a),[27-30] whereas the presence of tridentate L8 might result in h1-Ni(II) intermediates with C–C bond-formation occurring at the g-carbon (19b).[30-32] Ni(COD)2 (1 equiv) L8 (1 equiv) CO2, CaCl2, NEt3 Zn (x equiv) DMA, rt CO2H OH Ph 18 Ph N Me Me Ni Ph 19b 97:3 (19b:19a) x = 0, 73% x = 1.5, 81% N Me Ph N N Me Ni(0)(L3)2 [3] Ni(0)(L3)2 (1 equiv) L3 (0.60 equiv) CO2, MgCl2 Zn (x equiv) DMF, 40 ºC Ph CO2H Ph Ph 19a 99:1 (19a:19b) x = 0, 0% x = 4, 58% Scheme 6. Preliminary Mechanistic Experiments. In summary, we have documented a Ni-catalyzed switchable site-selective reductive carboxylation of allylic alcohols, in which regiodivergency is exclusively dictated by the geometry of the ligand backbone. The salient features of this user-friendly method are the broad scope, mild conditions, exquisite siteselectivity profile and the ability of CO2 to act with dual roles, both as C1 source and as a mediator for facilitating C–O scission. We anticipate that this method will foster systematic investigations for a more prolific use of simple alcohols in crosselectrophile reactions. Further work along these lines is being conducted in our laboratories. Acknowledgements We thank ICIQ, the European Research Council (ERC-277883), MINECO (Severo Ochoa Excellence Accreditation 2014-2018, SEV-2013-0319 & CTQ2015-65496-R, MINECO-FEDER-UE) and Cellex Foundation for support. Johnson Matthey, Umicore and Nippon Chemical Industrial are acknowledged for a gift of metal & ligand sources. MvG, M.B and A.T thank Alexander von Humboldt-Foundation and MINECO for a postdoctoral and predoctoral scholarships. [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] Keywords: cross-coupling • nickel • carbon dioxide • C–O cleavage [14] [1] [15] [2] (a) C. Zarate, M. van Gemmeren, R. J. Somerville, R. Martin, Adv. Organomet.Chem. 2016, 66,143; (b) M. Tobisu, N. Chatani, Acc. Chem. Res. 2015, 48,1717; (c) E. J. Tollefson, L. E. Hanna, E. R. Jarvo, Acc. Chem. Res. 2015, 48,2344; (d) B. Su, Z. –C. Cao, Z. –J. Shi, Acc.Chem. Res. 2015, 48, 886; (e) J. Cornella, C. Zarate, R. Martin, Chem. Soc.Rev. 2014, 43, 8081; (f) J. Yamaguchi, K. Muto, K. Itami, Eur.J.Org. Chem. 2013,19 (g) B. M. Rosen, K. W. Quasdorf, D.A. Wilson, N. Zhang, A. –M. Resmerita, N. K. Garg, V. Percec, Chem. Rev. 2011, 111,1346 While activation of allylic alcohols with CO2 has been explored using Pd, Ir, Rh as catalysts, analogous activation technologies using nickel catalysts are rare. For excellent reviews dealing with cross-coupling reactions of allylic alcohols with nucleophilic entities via the intermediacy of p-allylic species: (a) N. A. Butt, W. Zhang, Chem. Soc. [16] [17] [18] Rev. 2015, 44, 7929; (b) B. Sundararaju, M. Achard, C. Bruneau, Chem. Soc. Rev. 2012, 41, 4467; (c) Y. Tamaru, Eur. J. Org. Chem. 2005, 2647. For selected cross-coupling reactions of alcohols with nucleophilic species, see: (a) Y. Kita, R. D. Kavthe, H. Oda, K. Mashima, Angew. Chem., Int. Ed. 2016, 55, 1098; Angew.Chem. 2016, 128,1110; (b) Z. – C. Cao, D. –G. Yu, R. –Y. Zhu, J. –B. Wei, Z. –J. Shi, Chem. Commun. 2015, 51, 2683; (c) X. Huo, G. Yang, D. Liu, Y. Liu, I. D. Gridnev, W. Zhang, Angew. Chem., Int. Ed. 2014, 53, 6776; Angew. Chem. 2014, 126, 6894; (d) J. Y. Hamilton, S. Krautwald, D. Sarlah, M. Schafroth, E. M. Carreira, Science 2013, 340, 1065; (e) J. Y. Hamilton, D. Sarlah, E. M. Carreira, J. Am. Chem. Soc., 2013, 135, 994; (f) G. -X. Jiang, B. List, Angew. Chem., Int. Ed. 2011, 50, 9471; Angew. Chem. 2011, 123, 9643; (g) D. –G. Yu, B. –J, Li, S. –F. Zheng, B. –T. Guan, B. –Q. Wang, Z. –J. Shi, Angew. Chem. Int. Ed. 2010, 49, 4566; Angew. Chem. 2010, 122, 4670; (h) T. Yamamoto, J. Ishizu, A. Yamamoto, J. Am. Chem. Soc., 1981, 103, 6863, and references therein. For selected reviews: a) J. Gu, X.Wang, W. Xue, H. Gong, Org. Chem. Front. 2015, 3, 1411-1421; b) J. D. Weix, Acc. Chem. Res. 2015, 48, 1767-1775; c) T. Moragas, A. Correa, R. Martin, Chem. Eur. J. 2014, 20, 8242-8258; d) C. E. I. Knappke, S. Grupe, D. Gärtner, M. Corpet, C. Gosmini, A. Jacobi von Wangelin, Chem. Eur. J. 2014, 20, 6828-6842 S. J. Blanksby, G. B. Ellsion, Acc. Chem. Res. 2003, 36, 255 K. Osakada, R. Sato, T. Yamamoto, Organometallics 1994, 13, 4645 For selected references: (a) T. Moragas, M. Gaydou, R. Martin, Angew. Chem., Int. Ed. 2016, 55, 5053; Angew. Chem. 2016, 128, 5137; (b) X. Wang, M. Nakajima, R. Martin, J. Am. Chem. Soc. 2015, 137, 8924; (c) X. Wang, Y. Liu, R. Martin, J. Am. Chem. Soc. 2015, 137, 6476; (d) Y. Liu, J. Cornella, Martin, R. J. Am. Chem. Soc. 2014, 136, 11212; (e) A. Correa, T. León, R. Martin, J. Am. Chem. Soc. 2014, 136, 1062. For selected references: (a) F. Rebih, M. Andreini, A. Moncomble, A. Harrison-Marchand, J. Maddaluno, M. Durandetti, Chem. - Eur. J. 2016, 22, 3758; (b) K. Nogi, T. Fujihara, J. Terao, Y. Tsuji, J. Org. Chem. 2015, 80, 11618; (c) K. Nogi, T. Fujihara, J. Terao, Y. Tsuji, Chem. Commun. 2014, 50, 13052; (d) H. Tran-Vu, O. Daugulis, ACS Catal. 2013, 3, 2417; (e) T. Fujihara, K. Nogi, T. Xu, J. Terao, Y. Tsuji, J. Am. Chem. Soc. 2012,134, 9106. M. Börjesson, T. Moragas, D. Gallego, R. Martin, ACS Catal. 2016, 6, 6739. For an elegant regioselective (not regiodivergent) carboxylation with stoichiometric, air-sensitive and pyrophoric Et2Zn as reducing agent: T. Mita, Y. Higuchi, Y. Sato, Chem. - Eur. J. 2015, 21, 16391 (a) R. R. Weikel, J. P.; Hallett, C. L. Liotta, C. A. Eckert, Ind. Eng. Chem. Res. 2007, 46, 5252; (b) M. Sakamoto, I. Shimizu, A. Yamamoto, Bull. Chem. Soc. Jpn. 1996, 69, 1065; For an elegant Pd-catalyzed cross-coupling reaction with nucleophilic entities promoted by CO2, see: S. B. Lang, T. M. Locascio, J. A. Tunge, Org. Lett. 2014, 16, 4308. For a selection of elegant linear-selective reductive-couplings of other allyl electrophiles, see: (a) L. L. Anka-Lufford, M.R. Prinsell, D.J. Weix, J. Org. Chem. 2012, 77, 9989; (b) X. Qian, A. Auffrant, A. Felouat C. Gosmini, Angew. Chem. Int. Ed. 2011, 50, 10402; Angew. Chem. 2011, 123, 10586; (c) Y. Dai, F. Wu, Z. Zang, H. You, H. Gong, Chem. Eur. J. 2012, 18, 808. See Supporting information for details. The mass balance accounts for unreacted starting material and reduction events. For the beneficial role of MgCl2 in cross-electrophile events, see: (a) C. Zhao, X. Jia, X. Wang, H. Gong J. Am. Chem. Soc. 2014, 136, 17645; (b) T. Leon, A. Correa, R. Martin, J. Am. Chem. Soc. 2013, 135, 1221; c) F. Wu, W. Lu, Q. Qian, Q. Ren, H. Gong, Org. Lett. 2012, 14, 3044. While L7 is commercially available, L8 could easily be prepared from simple precursors. See ref. 13. These results are in sharp contrast with the ability of quaterpyridine ligands to promote carboxylation processes of allyl ester derivatives: T. Moragas, J. Cornella, R. Martin, J. Am. Chem. Soc. 2014, 136, 17702. Although tentative, Lewis acids (MgCl2 or CaCl2) might react with the corresponding carbonic acid generated in situ upon exposure of 1 with CO2, thus accelerating C–O bond-cleavage. Importantly, an identical regioselectivity was found regardless of whether CaCl2 or MgCl2 were utilized, thus reinforcing the notion that the site-selectivity was dictated by the ligand employed. As for the role of NEt3, it could potentially act COMMUNICATION [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] as a proton scavenger to prevent the formation of reduced products. See ref. 13. (a) X. Wang, S. Wang, W. Xue, H. Gong, J. Am. Chem. Soc. 2015, 137, 11562; (b) C. Zhao, X. Jia, X. Wang, H. Gong, J. Am. Chem. Soc. 2014, 136, 17645; (c) M. Börjesson, T. Moragas, R. Martin, J. Am. Chem. Soc. 2016, 138, 7504; (d) Ref. 16 K. M. Miller, T. F. Jamison, J. Am. Chem. Soc. 2004, 126, 15342 H. Takahata, Y. uchida, T. Momose, J. Org. Chem. 1995, 60, 5628 (a) T. Kapferer, R. Brückner, A. Herzig, W. A. König, Chem. Eur. J. 2005, 11, 2154; (b) C. García, M. Soler, V. S. Martín, Tetrahedron Lett. 2000, 41, 4127. For the synthesis of Ni(0)(L3)2 see ref. 7c. When Ni(0)(L3)2 was used in the absence of additional L3, 19a was obtained with a 22% isolated yield. An otherwise similar yield was obtained upon treatment of Ni(COD)2 with L3. At present, we do not have a rationale for the required 1:2.6 ratio of Ni:L3 ratio. Despite considerable synthetic efforts, we were unable to prepare Ni(0)(L8) in analytically pure form. Although one might argue that allyl chloride intermediates could be generated upon exposure of allyl alcohols to either MgCl2 or CaCl2, control experiments in the absence of CO2 ruled out this possibility. Additionally, we also demonstrate that allyl chlorides are not competent as reaction intermediates en route to either 19a or 19b. See ref. 13. Interestingly, 48% of 19b could be obtained when reacting 18 with Ni(COD)2/L8 in the absence of Zn, NEt3 or CaCl2. For the in situ generation of Ni(I) from Ni(II): (a) C. A. Laskowski, D. J. Bungum, S. M. Baldwin, S. A. Del Ciello, V. M. Iluc, G. L. Hillhouse, J. Am. Chem. Soc. 2013, 135, 18272; (b) J. Breitenfeld, J. Ruiz, M. D. Wodrich, X. Hu, J. Am. Chem. Soc. 2013, 135, 12004; (c) S. Biswas, D. J. Weix, J. Am. Chem. Soc. 2013, 135, 16192; (d) T. Fujihara, K. Nogi, T. Xu, J. Terao, Y. Tsuji J. Am. Chem. Soc. 2012, 134, 9106. The available experimental data do not allow us to rule out the formation of Ni(I) intermediates by comproportionation, see for example: (e) J. Cornella, E. Gomez-Bengoa, R. Martin, J. Am. Chem. Soc. 2013, 135,1997; (f) A. Velian, S. Lin, A. J. M. Miller, M. W. Day, T. Agapie, J. Am. Chem. Soc. 2010,132, 6296. Recently, Ni(I) species have shown to rapidly react with CO2: F. S. Menges, S. M. Craig, N. Tötsch, A. Bloomfield, S. Ghosh, H.-J. Krüger, M. A. Johnson Angew. Chem., Int. Ed. 2016, 55, 1282; Angew. Chem. 2016, 128, 1304. For a mechanistic hypothesis, see ref. 13 This mechanistic interpretation is somewhat reminiscent of the elegant 1 Pd-catalyzed carboxylation of allenes via h -allyl-Pd(II) intermediates possessing otherwise related tridentate pincer-type ligands in which the CO2 insertion occurred at the γ-carbon: (a) H.-W. Suh, L. M. Guard, N. Hazari, Chem. Sci. 2014, 5, 3859; (b) J. Takaya, N. Iwasawa, J. Am. Chem. Soc. 2008, 130, 15254. We cannot rule out that L8 might be redox non-innocent in our carboxylation event. For excellent reviews about redox-active ligands, see: (a) P. Chirik, Angew. Chem. Int. Ed. 2017, DOI: 10.1002/anie.201611959; (b) X. Hu Chem. Sci., 2011, 2, 1867. COMMUNICATION COMMUNICATION Manuel van Gemmeren, Marino Börjesson, Andreu Tortajada, ShangZheng Sun, Keisho Okura and Ruben Martin* Page No. – Page No. A switchable site-selective Ni-catalyzed carboxylation of allylic alcohols has been developed, in which CO2 is acting with dual roles, both facilitating the a priori uphill C–OH cleavage and as a C1 source This method constitutes the first example of a cross-electrophile coupling of unprotected alcohols in the absence of stoichiometric organometallic species. The protocol is distinguished by a broad scope and an exquisite regiodivergency that can be modulated by the ligand employed. Title Ni-Catalyzed switchable site-selective carboxylation of allyl alcohols with CO2