This  document  is  the  Accepted  Manuscript  version  of  a  Published  Work  that  appeared  in  final  form  in   Journa  of  the  American  Chemical  Society,  copyright  ©  American  Chemical  Society  after  peer  review  and  technical  editing   by  the  publisher.  To  access  the  final  edited  and  published  work  see  http://pubs.acs.org/doi/abs/10.1021/ja509077a           Ligand-controlled Regiodivergent Ni-Catalyzed Reductive Carboxylation of Allyl Esters with CO2 Toni Moragas†‡, Josep Cornella†‡ and Ruben Martin*†§ † § Institute of Chemical Research of Catalonia (ICIQ), Av. Països Catalans 16, 43007, Tarragona, Spain Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluïs Companys, 23, 08010, Barcelona, Spain Supporting Information Placeholder ABSTRACT: A novel Ni-catalyzed regiodivergent re- ductive carboxylation of allyl esters with CO2 has been developed. This mild, user-friendly and operationallysimple method is characterized by an exquisite selectivity profile that is dictated by the ligand backbone. The ability to control the outcome of catalytic reactions by the fine-tuning of the catalyst structure is central in the cross-coupling arena.1 Despite the advances realized, the development of catalytic regiodivergent protocols from a common precursor in a rational and predictable manner remains a formidable challenge,2 thus offering a unique opportunity to improve our ever-growing chemical portfolio. Intriguingly, while allyl electrophiles have been successfully employed as coupling partners with nucleophilic counterparts,3 the utilization of these motifs in catalytic reductive protocols is not as commonly practiced as one might anticipate.4,5 This is probably due to the difficulty for discriminating at will both ends of the initially generated π-allyl metal complex,3 resulting in regioselectivity issues (Scheme 1, II vs III). Indeed, a catalyst-controlled regiodivergent reductive event for selectively obtaining II and III from a common allyl electrophile (I) remains an unexplored area of research. Scheme 1. Regiodivergency in Allyl Electrophiles R1 X I X = I,Br,Cl,OR catalyst M X R R1 Catalyst-controlled regiodivergency ? R R or 1 R II III R = Nucleophile: known R1 R = Electrophile: unknown Carbon dioxide (CO2) has emerged as a powerful synthon and renewable chemical feedstock for organic synthesis.6 The interest for designing new catalytic reactions using CO2 arises from its low cost, high abundance and lack of toxicity and flammability. Nonetheless, the design of catalytic processes based on carbon dioxide is particularly challenging since CO2 is kinetically inert and not particularly soluble in commonly employed organic solvents at atmospheric pressure, thus resulting in competitive side-reactions. In recent years, we7 and others8 launched a program to unravel the potential of cata- lytic reductive carboxylation events using aryl or alkyl electrophiles en route to carboxylic acids, privileged motifs in a wide variety of pharmaceuticals and agrochemicals.9 Although these reactions have reached remarkable levels of sophistication,7,8 a ligand-controlled selectivity in carboxylation events is unknown, leaving ample opportunities to improve upon existing carboxylation techniques. Herein, we summarize our investigations aiming at the development of an unprecedented regiodivergent catalytic reductive carboxylation strategy (Scheme 2).10 The protocol is inherently modular, allowing for the introduction of the carboxylic motif at any site of the allyl terminus depending on the ligand employed (paths a & b). To the best of our knowledge, this constitutes the first time that the nature of the ligand dictates the outcome of carboxylation events.11 The transformation is mild and user-friendly, constituting an added value when compared with classical techniques based on well-defined allyl organometallic species,12,13 halide counterparts and/or high CO2 pressures. Scheme 2. Regiodivergent Catalytic Carboxylation R2 OAc R1 3 or R OAc 2 R R1 R3 R2 catalyst A / CO2 Reductant path b CO2H R1 catalyst B / CO2 R2 R1 Reductant path b R3 CO2H Mild reaction conditions Exquisite regiodivergency Operationally-simple R3 We started our investigations using 1a as the model substrate and the influence of all reaction components was systematically examined. As for other carboxylation reactions,7,8 we anticipated that the efficiency of the reaction would be strongly ligand dependent. As shown in Table 1, this was indeed the case. After some experimentation,14,15 we found that C2-substituted bipyridine L2 in DMF and Mn as reductant at atmospheric CO2 pressure was particularly suited for our purposes (entry 2). More importantly, such seemingly trivial modification at C2 was critical for improving the reactivity and selectivity pattern (entry 1 vs 2). Although L3 and L4 resulted in a decrease of selectivity (entry 3), a survey of additives revealed that both reactivity and 2a:3a ratio could be accentuated by adding MgCl2 with L3,16 afford- ing exclusively 2a in 77% isolated yield at 5 mol% catalyst loading (entry 4).17,18 Intriguingly, the use of MgCl2 did not have any influence for L2, thus showing the subtleties of our system. Strikingly, the use of commercially available quaterpyridine L5 resulted in a selectivity switch under identical reaction conditions, favouring the formation of 3a, albeit in lower yields (entry 6). These results tacitly suggest that the ligand backbone exclusively dictates the selectivity pattern. The fine-tuning of the Ni:L5 ratio, reductant, solvent and the inclusion of Na2CO3 as additive allowed for obtaining 3a in 72% isolated yield with an excellent 3a:2a ratio (entry 9).17,18 While similar selectivity was observed for L6 and L7 (entries 10 and 11), the best results were found with L5. As anticipated, control experiments revealed that all reaction components were crucial for success.14 Taking into consideration the lack of precedents when using L5 in the cross-coupling arena, we anticipate that L5 might open up perspectives in ligand design for effecting otherwise inaccessible coupling processes. Table 1. Optimization of the Reaction Conditions NiBr2·glyme (10 mol%) L (x mol%) C5H11 OAc C5H11 Reductant (2.40 equiv.) CO2 (1 atm), DMF, 40 ºC 1a R2 R1 R3 Entry Reductant   Yield  2+3  (%)   2 3 4 OAc 5 1e L1  (22)   Mn   8   L2  (22)   Mn   47   6 7 8 L3  (22)   Mn   58   1f, R = 3-CO2Et 1g, R = 3-Cl 1h, R = 2-SMe L3  (15)   c,d Mn   77   e O 9 5   L4  (22)   Mn   70   L5  (22)   Mn   10   L5  (22) f   Zn   28   5:95   f   Zn   41   f,g Zn   h 72   e h 3:97     2:98     4   1:99   8   L5  (15) 9   L5  (15)   f,g 10   L6  (15)   h L7  (15)     N Zn   f,g 11     Zn   11     N N R2   L5 (R = 2-py) L6 (R = Me) L7 (R = Ph) N R1 N 5:95     L1 (R1 = H) L2 (R1 = Me) N R1 R2 10 L3 (R2 = Me) L4 (R2 = Et) N R a 1a (0.25 mmol), NiBr2·glyme (10 mol%), L (x mol%), reductant (2.40 equiv.), DMF (0.17 M), CO2 (1 atm) at 40 ºC for 16 h. b Determined by GC using anisole as internal standard. c MgCl2 (2 equiv.) was added. d NiBr2·glyme (5 mol%). e Isolated yield. f DMA (0.17 M). g Na2CO3 (20 mol%) was added. h Zn (1.75 equiv.).   Table 2. Ligand-controlled Regiodivergent Carboxylation CO2H O O Me Me 2i, 67%; 99:1 (2i:3i) Me CO2H Me Me Me 11 1k 3i, 35%; 9:91 (2i:3i)h CO2H 2j, 62%; 99:1 (2j:3j)i Me Me 1j OAc 8:92   7   O 1if 53:47   6   3f, 57%; 7:93 (2f:3f) 2f, 52%; 99:1 2g, 63%; 90:10 (2g:3g) 3g, 64%; 7:93 (2g:3g) 2h, 58%; 82:18 (2h:3h)g 3h, 52%; 5:95 (2h:3h) HO2C OAc 99:1   R (2f:3f)g OAc O O 75:25   4   CO2H R 93:7   3   R1 PMBO 3e, 47%; 7:93 (2e:3e) CO2H R 81:19   2   HO2C CO2H 2e, 70%; 99:1 (2e:3e) OAc 2a:3a   1   R1 Ph CO2H Ph R2 R2 R2 1b, R1=R2=H 2b, 84%; 97:3 (2b:3b)d 3b, 71%; 6:94 (2b:3b) 1c, R1=H; R2=Me 2c, 59%; 97:3 (2c:3c)e 3c, 60%; 1:99 (2c:3c) 1d, R1=Me; R2=H 2d, 57%; 97:3 (2d:3d)e,f 3d, 55%; 3:97 (2d:3d) CO2H OPMB OPMB b L  (x  mol%)   Me 3a,72%; 2:98 (2a:3a) 2a, 77%; 99:1 (2a:3a) R1 Ph 3a Entry   Me 1a AcO 3 (using L5)c CO2H CO2H Me 1 CO2H 2a or CO H 2 b 2 (using L3)a,b 1 OAc a C5H11 CO2H OAc R2 1 1 Ni/L3 or Ni/L5 or R or R R1 CO2H R2 Reductant OAc R2 3 R3 CO2 (1 atm) R3 2a-k 1a-k 3a-k R Me CO2H 3j, 57%; 7:93 (2j:3j) Me Me Me HO2C Me 2k, 64%; 99:1 (2k:3k)j Me Me 3k, 78%; 1:99 (2k:3k) a Using L3: 1 (0.25 mmol), NiBr2·glyme (5 mol%), L3 (15 mol%), Mn (0.60 mmol), MgCl2 (0.50 mmol) in DMF at 40 ºC. b 2a-2j were obtained in ≥9:1 E:Z ratio. c Using L5: 1 (0.25 mmol), NiBr2·glyme (10 mol%), L5 (15 mol%), Zn (0.44 mmol), Na2CO3 (20 mol%) in DMA at 40 ºC. d At 50 ºC. e NiBr2·glyme (10 mol%) at 60 ºC. f 1.5:1 (E:Z). g NiBr2·glyme (10 mol%) and L4 (30 mol%). h 1:1 syn:anti. i NiBr2·glyme (3 mol%). j 2.3:1 (E:Z). Encouraged by these precedents, we turned our attention to the preparative scope of our Ni-catalyzed regiodivergent carboxylation protocol (Table 1). As shown, a variety of allyl acetates were all carboxylated in good yields and excellent regioselectivities depending on the ligand utilized. As expected, the carboxylation strategy based on L3 resulted in the predominant formation of Econfigured isomers (2a-k).19 Remarkably, a high selectivity profile was obtained regardless of whether linear or α-branched allyl acetates were utilized. These results reinforce the notion that our regiodivergent protocol does not operate under substrate-control and that the ligand exclusively dictates the selectivity pattern. As shown for 1c-1d, the inclusion of substituents on the allyl motif did not have a deleterious effect on selectivity. The preparation of carboxylic acids bearing a quaternary center (3d and 3k) is particularly noteworthy since Ni-catalyzed reductive coupling reactions of tertiary alkyl electrophiles are virtually inexistent.20 The chemoselectivity profile of our method is further illustrated by the presence of ethers (1e), acetals (1i), esters (1f), thioethers (1h) or alkenes (1j and 1k). Strikingly, while the inclusion of thioether motifs in the side chain had a negative impact for 2h, no erosion in selectivity was found when operating under a L5 regime, hence suggesting that thioethers compete with substrate binding with L3. Interestingly, the selectivity towards 3i was not affected by substituents in the α position of the allyl acetate fragment (1i), albeit 3i was obtained in lower yield.21 The successful preparation of 2k and 3k from naturally ocurring farnesyl acetate 1k highlights the robustness of our protocol in the presence of multiple double bonds. Moreover, the carboxylation could be conducted without affecting the aryl chloride entity, providing an additional functional handle via cross-coupling techniques (2g and 3g). Importantly, we found that the carboxylation of 1j could be conducted without noticeable 5-exo-trig cyclization (2j and 3j).22 Overall, the data in Table 2 demonstrates the robustness and prospective impact of our regiodivergent carboxylation protocol. Scheme 3. Convergent Synthesis of 2l and 3l from 1l-1n OAc Me Me Me Me Me 1l OAc Me 1m Me Me Me OAc 1n CO2H Me Ni / L5 Ni / L3 Zn Mn Me Me Me 3l Me CO2H 2l CO2 from 1l, 63%, 1:99 (2l:3l) from 1l, 80%, 99:1 (2l:3l), E:Z 2:1 from 1m, 73%, 1:99 (2l:3l) from 1m, 71%, 99:1 (2l:3l), E:Z 1:1.3 from 1n, 54%, 1:99 (2l:3l) from 1n, 52%, 99:1 (2l:3l). E:Z 1.5:1 Me Guided by the assumption that the reaction might not be substrate-controlled, we speculated that a different set of constitutional and configurational isomers could converge to a single carboxylic acid with a protocol based on L3 and L5. In line with our expectations, 1l-1n were exclusively converted into either 2l or 3l in good yields with variable E/Z ratios (Scheme 3).19 We believe these results suggest common reaction intermediates23 and increase the flexibility in synthetic design for preparing carboxylic acids from different precursors. Although a mechanistic study should await further investigations, we set out to explore the intermediacy of L3- and L5-Ni complexes. Following a procedure described by Nocera,24 we prepared air-sensitive 4 and 5 by reacting L3 or L5 with Ni(COD)2 in THF and their structures were univocally characterized by X-ray crystallography (Scheme 4).14,25 Intriguingly, while 2a could only be obtained in the presence of a reducing agent by using 4, 3a was cleanly produced with 5, even in the absence of reductant.26 These experiments confirm that the ligand backbone dictates the selectivity pattern and strongly suggest a different mechanistic pathway for L5 that differs from other reductive coupling events. At present, we believe that L5 might behave similarly to pincer-type ligands in related carboxylation events via η1-allyl intermediates27 and that the additional pyridine motif might be acting as a hemilabile ligand, thus tempering the catalytic activity on the Ni center and preventing decomposition pathways. Scheme 4. Stoichiometric Experiments 1a (1 equiv) CO2 (1 atm) MgCl2 (3 equiv) Me N Me N Ni N N Me Me N 4 N N Ni N 5 Mn (x equiv) DMF, 40 ºC 0% yield (x = 0) 46% yield (x = 2) 1a (1 equiv) CO2 (1 atm) Na2CO3 (20 mol%) Zn (x equiv) DMA, 40 ºC 54% yield (x = 0) 68% yield (x = 1.75) 2a 4 3a 5 In summary, we have described a novel, mild and userfriendly Ni-catalyzed regiodivergent carboxylation of allyl acetates with CO2. This protocol constitutes the first regiodivergent catalytic reductive coupling of allyl electrophiles and provides consistent evidence that the ligand dictates the selectivity pattern. We anticipate that this study will lead to new knowledge in catalyst design by using unconventional ligand backbones. Further investigations into the mechanism and the development of an asymmetric version are currently underway. ASSOCIATED CONTENT Supporting Information. Experimental procedures and spectral data. This material is available free of charge via the Internet at http://pubs.acs.org. AUTHOR INFORMATION Corresponding Author * rmartinromo@iciq.es Author contributions ‡ These authors contributed equally to this work Funding Sources No competing financial interests have been declared. ACKNOWLEDGMENT We thank ICIQ, the European Research Council (ERC277883) and MINECO (CTQ2012-34054 & Severo Ochoa Excellence Accreditation 2014-2018; SEV-2013-0319) for support. Johnson Matthey, Umicore and Nippon Chemical Industrial are acknowledged for a gift of metal & ligand sources. J.C thanks European Union (FP7-PEOPLE-2012- IEF-328381) for a fellowship. We thank Prof. V. Grushin & Prof. S. Ogoshi for useful discussions and Eduardo   Es-­‐ cudero   for   all   X-­‐Ray   crystallographic   data. This paper is dedicated to the memory of Prof. Gregory L. Hillhouse. (13) REFERENCES (1) Diederich, F.; Meijere, A., Eds. Metal-Catalyzed CrossCoupling Reactions; Wiley-VCH: Weinheim, 2004. (2) For selected reviews: (a) Mahatthananchai, J.; Dumas, A. M.; Bode, J. W. Angew. Chem., Int. Ed. 2012, 51, 10954. (b) Afagh, N. A.; Yudin, A. K. Angew. Chem., Int. Ed. 2010, 49, 262. (c) Trost, B. M. Science 1983, 219, 245. (3) For selected reviews on metal-catalyzed allylic substitution: (a) Arnold, J. S.; Zhang, Q.; Nguyen, H. M. Eur. J. Org. Chem. 2014, 23, 4925. (b) Sundararaju, B.; Achard, M.; Bruneau, C. Chem. Soc. Rev. 2012, 41, 4467. (c) Helmchen, G.; Dahnz, A.; Dübon, P.; Schelwies, M.; Weihofen, R. Chem. Commun. 2007, 675. (d) Trost, B. M.; Crawley, M. L. Chem. Rev. 2003, 103, 2921. (4) For reviews on metal-catalyzed reductive couplings: (a) Moragas, T.; Correa, A.; Martin, R. Chem. Eur. –J. 2014, 20, 8242. (b) Tasker, S. Z.; Standley, E. A.; Jamison, T. F. Nature 2014, 509, 299. (c) Everson, D. A.; Weix, D. J. J. Org. Chem. 2014, 79, 4793. (d) Knappe, C. E. I.; Grupe, S.; Gärtner, D.; Corpet, M.; Gosmini, C.; Jacobi von Wangelin, A. Chem. Eur. –J. 2014, 20, 6828. (5) For selected regioselective (not regiodivergent) reductive couplings of allyl electrophiles: (a) Tan, Z.; Wan, X.; Zang, Z.; Qian, Q.; Deng, W.; Gong, H. Chem. Commun. 2014, 50, 3827. (b) Anka-Lufford, L. L.; Prinsel, M. R.; Weix, D. J. J. Org. Chem. 2012, 77, 9989. (c) Wang, S.; Qian, Q.; Gong, H. Org. Lett. 2012, 14, 3352. (d) Durandetti, M.; Nedelec, J. –Y.; Perichon, J. J. Org. Chem. 1996, 61, 1748. (6) For selected reviews: (a) Zhang, L.; Hou, Z. Chem. Sci. 2013, 4, 3395. (b) Tsuji, Y.; Fujihara, T. Chem. Commun. 2012, 48, 9956. (c) Cokoja, M.; Bruckmeier, C.; Rieger, B.; Herrmann, W. A.; Kuhn, F. E. Angew. Chem., Int. Ed. 2011, 50, 8510. (d) Huang, K.; Sun, C. –L.; Shi, Z. –J. Chem. Soc. Rev. 2011, 40, 2435. (e) Martin, R.; Kleij, A. W. ChemSusChem. 2011, 4, 1259. (7) (a) Liu, Y.; Cornella, J.; Martin, R. J. Am. Chem. Soc. 2014, 136, 11212. (b) Correa, A.; Leon, T.; Martin, R. J. Am. Chem. Soc. 2014, 136, 1062. (c) Leon, T.; Correa, A.; Martin, R. J. Am. Chem. Soc. 2013, 135, 1221. (d) Correa, A.; Martin, R. J. Am. Chem. Soc. 2009, 131, 15974. (8) (a) Nogi, K.; Fujihara, T.; Terao, J.; Tsuji, Y. Chem. Commun. 2014, 50, 13052. (b) Tran-Vu, H.; Daugulis, O. ACS Catal. 2013, 3, 2417. (c) Fujihara, T.; Nogi, K.; Xu, T.; Terao, J.; Tsuji, Y. J. Am. Chem. Soc. 2012, 134, 9106. (9) Maag, H. Prodrugs of Carboxylic Acids; Springer: New York, 2007. (10) For selected metal-catalyzed regiodivergent reductive coupling processes not involving CO2 or allyl electrophiles: (a) Zhao, Y.; Weix, D. J. J. Am. Chem. Soc. 2014, 136, 48. (b) Köpfer, A.; Sam, B.; Breit, B.; Krische, M. J. Chem. Sci., 2013, 4, 1876. (c) Shareef, A. –R.; Sherman, D. H.; Montgomery, J. Chem. Sci. 2012, 3, 892. (d) Miller, K. M.; Jamison, T. F. J. Am. Chem. Soc. 2004, 126, 15342. (11) This work was presented at the 19th International Symposium on Homogeneous Catalysis, Ottawa (July 6-11, 2014). At the conference, the Tsuji group described a nonreductive Cu-catalyzed regiodivergent carboxylation of allenes using silylboranes as coupling partners. (12) For the carboxylation of allyl boronates, see: (a) Duong, H. A.; Huleatt, P. B.; Tan, Q. –W.; Shuying, E. L. Org. Lett. (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) 2013, 15, 4034. For the carboxylation of allyl stannanes: (b) Hruszkewycz, D. P.; Wu, J.; Hazari, N.; Incarvito, C. D. J. Am. Chem. Soc. 2011, 133, 3280. (c) Shi, M.; Nicholas, K. M. J. Am. Chem. Soc. 1997, 119, 5057. For selected carboxylation of well-defined organometallics derived from allyl halides: (a) Wu, J.; Green, J. C.; Hazari, N.; Hruszkewycz, D. P. Incarvito, C. D.; Schmeier, T. J. Organometallics 2010, 29, 6369. (b) Miao, B.; Ma, S. Chem. Commun. 2014, 50, 3285. (c) Hung, T.; Jolly, P. W.; Wilke, G. J. Organomet. Chem. 1980, 190, C5. (d) Courtois, G.; Migniac, L. J. Organomet. Chem. 1974, 69, 1. (e) Friederich, L. E.; Cormier, R. A. J. Org. Chem. 1971, 36, 3011. See Supporting information for details. No products were detected under previously reported Nicatalyzed reductive carboxylation reactions (ref. 7b, 7c, 8c). For the beneficial role of MgCl2 in carboxylation reactions or reductive events, see: (a) Wu, F.; Lu, W.; Qian, Q.; Ren, Q.; Gong, H. Org. Lett. 2012, 14, 3044. (b) Metzger, A.; Bernhardt, S.; Manolikakes, G.; Knochel, P. Angew. Chem., Int. Ed. 2010, 49, 4665. (c) Ref. 7c. Dimerization and reduction account for the mass balance. The inclusion of H2O (10–100 mol%) shut down the reactivity. We found identical results when scaling up the reaction of 1a (1 mmol) under a L3 or L5 regime. Lower E/Z ratios were found for 2k-n, an observation that is in line with the directing effect of tethered alkenes in Nicatalyzed coupling reactions. See for example: ref. 10d. For remarkable exceptions: (a) Luo, L.; Zhang, J. –J.; Ling, W. –J.; Shao, Y. –L.; Wang, Y. –W.; Peng, Y. Synthesis 2014, 46, 1908. (b) Yang, D.; Belardi, J. K.; Micalizio, G. C. Tetrahedron Lett. 2011, 52, 2144. (c) Refs. 7c and 8a No reaction took place when exposing 4,4-dimethylpent-1en-3-yl acetate possessing a quaternary carbon in α-position under the conditions based upon L5. Likewise, no reaction was observed with cyclohex-2-en-1-yl acetate. At higher Ni/L3 loadings we observed 2j and 5-exo-trig cyclization in a linear relationship, suggesting that a radicalescape-rebound mechanism could be operating. Indeed, the reaction of 1j with Ni/L3 was inhibited by addition of radical scavengers such as TEMPO or galvinoxyl. Intriguingly, 3j was the only observable product with Ni/L5, reinforcing the notion that a different interplay operates for L5. In line with this notion, we found that racemic 3b was obtained from (R)-1b (65% ee) under a Ni/L5 regime. Powers, D. C.; Anderson, B. L.; Nocera, D. G. J. Am. Chem. Soc. 2013, 135, 18876. Taking into consideration the tetrahedral geometry for Ni(0) complexes, the square-planar environment found for 5 might suggest that this complex would be best described as a Ni(II) complex of a reduced quaterpyridine ligand dianion rather than a Ni(0) complex. For an excellent review dealing with redox-active ligands: Hu, X. Chem. Sci. 2011, 2, 1867. The use of stoichiometric Ni(COD)2/L3 or Ni(COD)2/L5 in provided otherwise identical reactivity to 4 and 5. Tridentate, pincer-type ligands have shown to promote a related α-branched carboxylation of allenes. The mechanism is believed to proceed via the formation of a new C-C bond between CO2 and the γ-carbon of an in situ generated η1-allyl metal complexes, see: (a) Suh, H.–W.; Guard, L. M.; Hazari, N. Chem. Sci. 2014, 5, 3859. (b) Takaya, J.; Iwasawa, N. J. Am. Chem. Soc. 2008, 130, 15254. Ligand-controlled catalytic regiodiodivergent reductive carboxylation N N R1 CO2H R2 R3 14 examples up to 78% yield up to 99:1 selectivity L6 N N R2 Ni catalyst Reductant CO2 (1 atm) R1 R3 OAc 1 or R OAc R2 R3 Me User-friendly & mild conditions Exquisite regiodivergency N N L4 Me R2 Ni catalyst Reductant CO2 (1 atm) R1 CO2H R3 14 examples up to 84% yield up to 99:1 selectivity 5