Cu(II)-Catalyzed Hydrofunctionalization of Unactivated Alkynes via π-Lewis Acid Activation

Abstract

A copper(II)-catalyzed stereodivergent hydrooxygenation of electronically unactivated alkynes with carboxylic acids is reported. Regioselectivity and kinetic reactivity are facilitated by a bidentate auxiliary, and syn- or anti-stereoselectivity is controlled through judicial tuning of reaction conditions. The method affords trisubstituted E- or Z-alkenes with enol ester functionality in a highly selective manner. Beyond carboxylic acids, a variety of other OH- and NH-nucleophiles react smoothly to furnish enol ether, enamine and alkylidene β-lactam building blocks. Mechanistic experiments and density functional theory (DFT) calculations shed light on the nature of π-Lewis acid activation with Cu(II) and support a catalytic cycle that features inner-sphere nucleocupration mechanism via 6-membered transition state for syn-addition and base-assisted outer-sphere nucleocupration mechanism for anti-addition.

Graphical Abstract

Transition-metal-catalyzed carbon–carbon π-bond functionalization reactions constitute a synthetically enabling and versatile toolkit to rapidly advance structural complexity of simple starting materials towards useful building blocks.¹ Carbophilic second- and third-row late transition-metal salts, such as Pd(II)², Au(I) and Pt(II),³ are known to catalyze various anti-selective Wacker-type nucleometalation processes through π-Lewis acid activation (Scheme 1A). ^4,5 Devising analogous strategies for π-activation with earth-abundant first-row metals would not only improve the sustainability, but expand catalytic capabilities of these reactions.⁶ In the past decades, copper salts have been shown to mimic some of the catalytic chemistry of palladium, for example, in C–H activation reactions.⁷ Generally speaking, Cu-catalyzed alkyne/alkene functionalization reactions typically proceed through one of two major manifolds: 1) syn-migratory insertion^8–9 of organocopper species (e.g., Cu–H, Cu–BR₂)^10–12, or 2) a radical addition initiated by single-electron transfer from Cu to a carbon–heteroatom or heteroatom–heteroatom bond (Scheme 1B).¹³ Although organometallic studies of copper as a π-Lewis acid have been reported,¹⁴ the scarce examples of Wacker-type nucleocupration catalysis in the literature are limited to intramolecular cyclizations¹⁵ or intermolecular couplings of electronically activated substrates¹⁶.

Scheme 1.

Background and project synopsis.

With the long-term goal of developing diverse carbon–carbon and carbon–heteroatom bond-forming reactions via nucleocupration, we chose intermolecular alkyne hydrooxygenation as a model reaction.¹⁷ Compared to other enol ester synthesis strategies, including O-acylation of ketones¹⁸, C–O cross-coupling¹⁹, enol ether/ester exchange²⁰, rearrangement of propargylic esters²¹ and isomerization of allylic esters²², addition of carboxylic acids to alkynes does not require pre-functionalized substrates or toxic reagents and is characterized by intrinsically high atom-economy and robust functional group tolerance. However, preparatively useful levels of regio- and stereoselectivity are typically only observed with terminal, electronically activated or symmetrical alkynes. For electronically unbiased, non-symmetrical internal alkynes, control of selectivity is particularly challenging. Our group has leveraged alkene and alkyne substrates containing polydentate Lewis basic auxiliaries as a platform to rapidly prototype new reactions.^2,9 Previous studies by Li and Corpas, Mauleón, and Arrayás have described directed alkyne anti-selective hydrooxygenation with Pd(II) and Ru(II) catalysts, respectively, using solvent-quantities of the carboxylic acid coupling partners.²³ Methodology that predictably grants access to either syn- or anti-addition products with stoichiometric quantities of the carboxylic acid coupling partner using an earth-abundant metal catalyst would dramatically expand the scope of accessible products and improve the sustainability of this approach. Herein, we report a Cu(II)-catalyzed directed regioselective and stereodivergent alkyne hydrooxygenation method.²⁴ Selectivity for Z- or E-enol ester products is controlled through careful tuning of the reaction conditions, particularly solvent polarity, to favor inner versus outer-sphere oxycupration, as revealed through density functional theory (DFT) calculation and mechanistic experiments (Scheme 1C).

To reduce this general idea to practice, reaction conditions were tested with model substrate 1a, which consists of an electronically non-biased alkyne tethered to an 8-aminoquinoline (AQ)-derived amide directing auxiliary (Table 1). With Cu(OBz)₂•H₂O as catalyst and benzoic acid (BzOH) as coupling partner at 40 °C, we noticed stereoselectivity could be controlled through choice of appropriate solvent and Lewis basic additive. In absence of an additive, non-polar and non-coordinating solvents, such as DCE, ethers, and toluene, gave syn-addition product 2aa as the major product (entries 1–4). Among solvents tested, 1,4-dioxane afforded 2aa with highest yield (entry 3). Having optimized an effective protocol to access the E-configured product 2aa, we moved on to exploring variables that would invert stereoselectivity. Among alcoholic solvents tested, HFIP afforded ketone 4, the product of formal alkyne hydration, in 56% yield, presumably via alcoholysis of enol ester (entry 5). With methanol, both Z- and E- enol esters as well as ketone were observed (entry 6). When switching to polar and aprotic solvents, such as acetonitrile, DMSO, DMF and DMA, a mixture of syn- and anti-addition products were formed (entries 7–10). Based on this observation, we hypothesized that the formation of Z-enol ester was derived from stabilized carboxylate anion in polar solvents. Thus, we next examined some organic bases, which we envisioned could perturb Z/E-selectivity by deprotonating the carboxylic acid. Although adding pyridine or TMEDA (1 equiv) resulted in low yields (entries 11–12), to our delight, with tertiary amine additives, Z-enol ester 3aa was formed with excellent anti-selectivity and high yields (entries 13–14). The optimal combination proved to be diisopropylethylamine (DIPEA) as additive (1 equiv) and DMA as solvent at room temperature, giving the desired anti-addition product 3aa with 93% isolated yield (entry 15). While stoichiometric excess benzoic acid was beneficial for yield, under the optimized conditions, >70% yield and high selectivity could be maintained with benzoic acid loadings as low as 1.5 equiv (see Tables S3 and S4, Supporting Information).

Table 1.

Reaction optimization.^a

Entry	Additive	Solvent	Yield (2aa)	Yield (3aa)	Yield (4)
1	-	DCE	68%	0%	16%
2	-	THF	76%	0%	2%
3b	-	1,4-dioxane	96% (94%)	0%	0%
4	-	PhMe	83%	0%	7%
5	-	HFIP	0%	0%	56%
6	-	MeOH	38%	5%	10%
7	-	MeCN	63%	9%	5%
8	-	DMSO	17%	36%	1%
9	-	DMF	25%	44%	0%
10	-	DMA	24%	50%	0%
11	pyridine	DMA	2%	13%	0%
12	TMEDA	DMA	0%	0%	0%
13	Et3N	DMA	3%	70%	0%
14	DIPEA	DMA	1%	82%	0%
15b,c	DIPEA	DMA	1%	96% (93%)	0%

^aReactions performed on 0.05-mmol scale. Yield based on ¹H NMR analysis of the crude reaction mixture with benzyl 4-fluorobenzoate as internal standard.

^bReactions performed on 0.1-mmol scale. Isolated yield in parenthesis.

^cBzOH (3.0 equiv), room temperature, 12 h.

^d1,4-dioxane (0.1 M).

^eDMA (0.1 M).

To understand the importance of the directing group identity, we evaluated the reactivity of various mono- and bidentate alternatives. Mono-dentate acid and amide (DG2 and DG3) and alternative bidentate N,N- and N,S-directing groups were ineffective (DG4–DG7).²⁵ Notably, oxazoline-derived directing group (DG7), which is widely used in Cu(II)-mediated C–H activation reactions, ^{7c–7e, 26} failed to afford the desired product. Electronically modified AQ directing auxiliaries were also examined. Using electron-rich ^5-OMeAQ as directing group hardly influenced reactivity and selectivity in both syn and anti-conditions. In contrast, with electron-deficient ^5-NO2AQ, syn-selective addition proceeded well, whereas anti-addition was low-yielding.

The possible involvement of trace metal contaminants in promoting hydrooxygenation was considered. In a series of control experiments, we tested alternative batches of Cu(OAc)₂ with varying levels of copper purity as well as other transition-metal catalysts previously employed in directed alkyne hydrooxygenation reactions²³ under standard and literature conditions. These experiments revealed that the quality of Cu(OAc)₂ did not influence the reaction and other transition-metal catalysts were unreactive, confirming that copper is the reactive catalyst in the reaction (Table S2, Supporting Information).

With the optimized conditions in hand, we first explored the scope of carboxylic acids (Table 2). Most aromatic acids that were examined gave the desired syn- and anti-addition products with moderate to good yields and excellent regio- and stereoselectivity. Benzoic acid derivatives with electron-withdrawing (–SO₂Me, –SO₂F, –CHO, –CF₃, –CN, –F, –NO₂, –CO₂Me) and electron-donating groups (–SMe, –OAc, –OMe, –NHAc) at the para (2ab–2an, 3ab–3an), ortho (2ao–2ap, 3ao–3ap), or meta (2aq–2at, 3aq–3at) positions did not significantly affect the yield or selectivity. Some potentially reactive functional groups, such as sulfonyl fluoride (2ac, 3ac), aldehyde (2ad, 3ad), halides (2al–2am, 3al–3am), and vinyl groups (2an, 3an) were well tolerated under mild conditions. For pentafluorobenzoic acid (2au, 3au), despite its low pK_a (1.48), the reactions proceeded successfully under both syn- and anti-conditions in >99% and 44% yields, respectively. Sterically hindered 2,4,6-trimethylbenzoic acid (2av, 3av) delivered corresponding syn- and anti-addition products in decent yields. Additionally, ferrocene and heteroaromatic rings such as pyrrole, furan and thiophene are also compatible (2aw–2az, 3aw–3az). However, picolinic acid is inactive (Table S8, Supporting Information), probably because the strongly coordinating pyridine-type nitrogen inhibits the copper catalyst.

Table 2.

Carboxylic acid scope.^a

^aReactions performed on 0.10-mmol scale. Percentages represent isolated yields. Stereoisomeric ratios were determined by ¹H NMR analysis of the crude reaction mixtures and isolated products, and in both cases, values were ≥95:5 for all examples tested unless otherwise noted.

^b80 °C.

^cZ:E = 93:7 in the crude reaction mixture and ≥95:5 in the isolated product.

^dZ:E = 78:22 in the crude reaction mixture and ≥95:5 in the isolated product.

^e30 °C.

^fWith PiV₂O (2.5 equiv) as additive.

^gZ:E = 85:15 in the crude reaction mixture and ≥95:5 in the isolated product.

^h60 °C.

ⁱCu(OAc)₂ (20 mol%), PhC(O)CO₂H (5.0 equiv).

In the cases of non-aromatic acids, performing syn-selective reactions at lower temperature (30 °C) suppressed the competitive formation of ketone byproduct (2ba, 2be, 2bg, 2bh). Alternatively, in the case of pivalic acid, addition of pivalate anhydride can also prevent ketone formation (2bb). Under anti-selective conditions, the ketone byproduct arising from formal hydration was not observed; however, pivalic acid (3bb) and cyclopropanecarboxylic acid (3bc) gave the corresponding products with diminished stereoselectivity. We also tested the reactivity of α-functionalized acetic acids. 2-Phthalimidoacetic acid coupled smoothly under both syn and anti-conditions (2bd, 3bd). With 2-chloroacetic acid, syn-selective addition proceeds in 61% yield (2be). However, under anti-conditions, after initial addition of the first equivalent of acid to the alkyne, the highly electrophilic α-chloride is substituted rapidly by a second equivalent of acid (Scheme S14, Supporting Information). Phenylglyoxylic acid afforded the corresponding products in 52% and 12% yields at 60 °C (2bf, 3bf). The diminished reactivity presumably arises from enhanced stability and attenuated nucleophilicity of the carboxylate conjugate base (pK_a = 1.21 in H₂O). Conjugated alkenyl and alkynyl acids are also good substrates, giving satisfactory yields and selectivity (2bg–2bh, 3bg–3bh).

Regarding the alkyne scope (Table 3), an internal methyl-substituted alkyne led to good yields and stereoselectivity (2ca, 3ca). Increasing the steric hindrance at the propargylic position distal from the AQ group had little influence on syn-selective reactivity (2da) but resulted in poor reactivity in anti-addition. In the case of cyclohexyl alkyne, increasing the catalyst loading to 50 mol% afforded Z-enol ester in 49% yield (3da) (for optimization details, see Table S5, Supporting Information). Substrates with a pendant ether or chloride group worked smoothly, giving good yields for syn- and anti-addition, respectively (2ea, 2fa, 3ea, 3fa). The α-gem-dimethyl substrate also underwent both syn- and anti-hydrooxygenation, giving excellent yields and selectivity (2ga, 3ga). Under syn-conditions, terminal alkyne substrate afforded the desired E-enol ester in 71% yield (2ha) along with some Glaser-type alkyne homo-dimerized side-product.²⁷ Under anti-conditions, however, the desired hydrooxygenation product was formed in low yield (3ha), presumably because the presence of basic DIPEA facilitates copper–acetylide formation, thereby favoring Glaser homo-dimerization.²⁸ While the formation of these Glaser side products suggests that Cu(I) species may be formed under the reaction conditions, for example, by amine-reduction or disproportionation, mechanistic data point to Cu(II) as the catalytically active species for hydrooxygenation (vide infra). Aryl alkynes exhibited diminished reactivity compared to alkyl alkynes. Syn-selective addition gave decent yields (2ia, 2ja) with higher temperature and copper loading, while there is no desired product detected under anti-condition because of rapid alkyne-to-allene isomerization. To prevent the undesired isomerization pathway, we next examined an aryl alkyne substrate with α-gem-dimethyl substitution. To our delight, anti-selective addition proceeded smoothly with a higher temperature and catalyst loading (3ka), while a low yield was obtained under syn-selective conditions (2ka). Introduction of an additional methylene spacer between the directing group and alkyne was tolerated but required much higher temperatures (2la, 3la). Alkynyl amine substrate containing a picolinyl directing auxiliary gave syn-addition product in 91% yield under 120 °C (2ma), but only unreacted starting material was observed in the anti-selective conditions for reasons that remain unclear.

Table 3.

Alkyne scope.^a

^aReactions performed on 0.10-mmol scale. Percentages represent isolated yields. Stereoisomeric ratios were initially determined by ¹H NMR analysis of the crude reaction mixtures and isolated products, and were ≥95:5 in all examples tested unless otherwise noted.

^bCu(OBz)₂•H₂O (50 mol%).

^cZ:E = 93:7 in product.

^dCu(OBz)₂•H₂O (20 mol%), 50 °C.

^e1,4-dioxane (0.1 M), 50 °C.

^fCu(OBz)₂•H₂O (20 mol%), BzOH (5.0 equiv), 80 °C, Ar.

^gCu(OBz)₂•H₂O (20 mol%), DIPEA (2.0 equiv), BzOH (5.0 equiv), 80 °C.

^h120 °C.

ⁱ100 °C, Z:E = 83:17 in crude reaction mixtures and ≥95:5 in isolated products.

In preliminary experiments, we have found that this Cu(II)-catalyzed hydrooxygenation method can be extended to an alkyne substrate containing only a simple, mono-dentate directing group (Scheme 2). Following optimization of reaction conditions for this substrate (Table S6, Supporting Information), the reaction was found to proceed at elevated temperature. The corresponding anti-addition product (2na) formed via the 5-exo-type metallacycle was isolated in 55% yield, accompanied by formation of regioisomer (2na’) in 27% yield via 6-endo-type metallacycle. In this system, competitive formation of two regioisomers likely stems from the higher reaction temperature and more flexible chelate structure from the mono-dentate directing group.

Scheme 2.

Preliminary results of alkyne substrate with a mono-dentate directing group.

To establish the generality of Cu(II) as a catalytic π-Lewis acid for alkyne activation, other OH- and NH-nucleophile classes were next evaluated (Table 4). Under standard or slightly modified anti-selective conditions, acidic OH-nucleophiles, such as pentafluorophenol, yielded corresponding enol ether with good yields and excellent anti-selectivity (4a). Although competitive C-²⁹ versus O-functionalization³⁰ of the enol/enolate-type pronucleophiles 4-hydroxycoumarin and dimedone in related previous work, in this case anti-selective O-functionalization products were formed exclusively in both cases (4b, 4c). Hydroamination of alkynes is a powerful and efficient tool for constructing C–N bonds.³¹ Previously reported Cu-catalyzed alkyne hydroamination methods relied on either activated alkyne substrates³² or an umpolung strategy³³ with amine electrophiles and silanes as H-atom source. In presence of Cu(II) catalyst, DIPEA and polar solvent (EtCN), sulfonamide nucleophiles underwent anti-addition to the standard alkyne substrate to afford the corresponding enamines in moderate to good yields (4d–4h). Additionally, with alkynyl bromide as substrate, intramolecular syn-selective hydroamination was achieved to give alkylidene β-lactam³⁴ building block, which exhibits excellent anti-bacterial activity³⁵ and is utilized as an important linchpin in organic synthesis³⁶, via 4-exo-dig cyclization (4i).

Table 4.

OH- and NH-nucleophile scope.^a

^aReactions performed on 0.10-mmol scale. Percentages represent isolated yields. Stereoisomeric ratios were initially determined by 1H NMR analysis of the crude reaction mixtures and isolated products and were ≥95/5 in all examples tested unless otherwise noted.

^bCu(OAc)₂ (10 mol%), 1a (1.0 equiv), [Nu] (3.0 equiv), DIPEA (1.0 equiv).

^cDMA (0.25 M), r.t.

^dEtCN (0.25 M), r.t.

^eEtCN (0.25 M), 40 °C.

^fCu(OAc)₂ (20 mol%), 1a (1.0 equiv), [Nu] (5.0 equiv), DIPEA (1.0 equiv), EtCN (0.25 M), 40 °C.

^gCu(OAc)₂ (20 mol%), 1a (1.0 equiv), saccharin (5.0 equiv), DIPEA (2.0 equiv), EtCN (0.25 M), 70 °C.

To underscore the preparative utility of this method, we performed reactions with a representative substrate on larger scale. Both the syn- and anti-selective addition can be performed on gram-scale to furnish the desired products in 89% and 85% yield with excellent selectivity. Regio- and stereoselectivity were confirmed by single-crystal X-ray diffraction experiments (Scheme 3A). Additionally, several product transformations were carried out. The 8-aminoquinoline (AQ) directing auxiliary can be removed via a transamidation procedure shown in Scheme 3B.³⁷ After activating the amide through installation of an N-Boc group, the resulting twisted amides can react with a free amine to form other valuable amides (2aa-2, 3aa-2). To demonstrate the potential utility of stereodivergent access to enol esters, we conducted stereospecific alkene epoxidation reactions. A pair of epoxide diastereomers (2aa-3, 3aa-3) could thus be accessed in good yields and diastereoselectivity (Scheme 3C).

Scheme 3.

Gram-scale experiments and product transformations.

The stereodivergent nature of the method and the ability of copper to promote this transformation prompted us to probe the underlying reaction mechanism. Firstly, we designed a series of control experiments to probe the viability of different mechanisms that could theoretically be operative (Scheme 4A). Given that propargylic carboxamides are prone to isomerization to form the corresponding conjugated allenes,^38,39 we first wanted to establish whether the putative allene was a competent intermediate. To this end, we independently prepared allene 5 and then subjected it to the syn-and anti-selective conditions.⁴⁰ From the two experiments, allylic ester 6 was obtained as the major product. In product 6, the intended C–O bond had formed at the γ-position but the alkene remained in conjugation with the amide at α,β-position, and subsequent alkene isomerization did not take place. Second, we sought to ascertain the oxidation state of copper responsible for the observed reactivity. We recognized the potential for the formation of Cu(I) via Cu(II) precatalyst disproportionation⁴¹ or reduction by the amine additive⁴². which could then be the reactive catalytic species in the reaction. To probe this possibility, the reaction was attempted with Cu(OAc) as the catalyst under inert atmosphere, but in these experiments the desired syn- or anti-addition products (2aa and 3aa) were observed in very low yields, suggesting that Cu(I) was far less reactive than Cu(II). Third, we clarified whether the selectivity for each of the two stereoisomers was a result of thermodynamic versus kinetic control enabled by Z/E alkene isomerization. To this end, independently prepared syn (2aa) and anti (3aa) products were subjected to anti or syn conditions, respectively. However, no evidence of alkene isomerization was observed, precluding the possibility that secondary product isomerization contributed to selectivity control.

Scheme 4.

Mechanistic investigation and proposed mechanisms.

Next, to gain insight into the product-determining step of the catalytic cycle and establish what structural features of the nucleophile are important for high kinetic reactivity, competition experiments were performed with a mixture of electron-rich 4-methoxybenzoic acid (pK_a = 4.47) and electron-deficient 4-cyanobenzoic acid (pK_a = 3.55) (Scheme 4B). In both syn- and anti-selective conditions, the electron-deficient acid addition products (2af or 3af) were found to be the major products. This product distribution presumably arises from the fact that 4-cyanobenzoic acid is more acidic, which gives rise to higher concentration of the reactive benzoate conjugate base in solution, consistent with alkyne oxycupration as the product-determining step. In line with this hypothesis, under anti-conditions, when the equivalents of DIPEA are increased, the ratio of electron-rich to electron-poor acid addition product (3ai:3af) increases. Here, higher concentration of the Brønsted base leads to near complete deprotonation of both benzoic acids, and in this scenario the more nucleophilic 4-methoxybenzoate reacts faster than its electron-deficient counterpart as expected.

To gain insight into the proton source and potential reversibility of C–H bond formation, deuterium incorporation experiments were performed (Scheme 4C). To minimize the concentration of H⁺ in reaction system, benzoic acid-d and deuterium oxide were employed. Under syn-selective conditions, 40% deuterium was observed exclusively at the vinylic position, with the remaining non-deuterated product presumably arising from H⁺ in AQ group. In anti-selective conditions, deuterium incorporation was observed in both vinylic and allylic positions. The allylic deuteration indicated facile deprotonation of propargylic position with DIPEA, resulting in allene formation in absence of coupling partner. Additionally, when adding the hydrooxygenation products 2aa and 3aa to corresponding syn- and anti-conditions in presence of deuterium sources, no H/D exchange took place at vinylic positions, suggesting the protodecupration step is irreversible under standard conditions.

Based on mechanistic studies and literature,^{2d,16e,23b,43} we proposed possible reaction mechanisms for syn- and anti-selective reactions (Scheme 4D). In syn-condition, the carboxylate ligand on Cu(II) center can be added to alkyne from conventional migratory insertion via a 4-membered transition state or inner-sphere nucleometallation via a 6-membered transition state. In anti-condition, the coordinating basic additive or solvent is likely to block the coordination site for carboxylate ligand, forcing it to proceed with outer-sphere nucleocupration and lead to anti-addition product. Besides, in presence of a carboxylate ligand, hydrogen-bonding-assisted anti-selective attack is also plausible. ^{2d,16e,23b,43}

Finally, to distinguish possible pathways in syn- and anti-selective conditions, we turned to density functional theory (DFT) calculations. The DFT-optimized structure of the π-alkyne complex of substrate 1c with the Cu(II) catalyst (7) indicates relative strong binding of the alkyne to the square planar^44,45 Cu(II) center and a significant amount of charge transfer from the alkyne to the Cu due to the π-Lewis acid activation (Figure S4). Several nucleocupration–protodecupration pathways from complex (7) were computed in both 1,4-dioxane and DMA solvents to mimic the experimental conditions leading to the syn- and anti-selective products 2ca and 3ca, respectively (see Table 3). In 1,4-dioxane solvent (Figure 1, left), the most favorable pathway is the inner-sphere syn-nucleocupration of the benzoate-supported Cu(II) complex 7 via a six-membered cyclic transition state TS1a. The calculated free energy barrier and the transition state geometry is comparable with the previously reported syn-nucleopalladation reactions.^46–48 The inner-sphere syn-nucleocupration via four-membered cyclic transition state TS1b is 9.7 kcal/mol less favorable. Next, we considered the outer-sphere nucleocupration pathway that involves the replacement of the benzoate ligand in 7 with a dioxane solvent molecule followed by the anti-nucleocupration of the dissociated BzO⁻. This outer-sphere pathway via TS1c requires a much higher barrier (ΔG^‡ = 34.1 kcal/mol with respect to 7), due to the insufficient stabilization of charge separation between the BzO⁻ nucleophile and the cationic Cu(II) center in the non-polar 1,4-dioxane solvent. The hydrogen-bonding-assisted nucleocupration pathway (Figure S2) was also found to be less favorable than the syn-nucleocupration via TS1a. Although the nucleocupration of 7 to 8 is slightly endergonic by 4.7 kcal/mol, subsequent coordination of benzoic acid to the Cu(II) center to form 9 and the protodecupration via TS2 are highly facile, leading to Cu(II)–product complex 10, which is 16.6 kcal/mol more stable than 7. The facile protodecupration process indicates that the product stereoselectivity is determined in the nucleocupration step.

Figure 1.

Computed reaction energy profiles for the syn- and anti-nucleocupration pathways in 1,4-dioxane and DMA solvents (i.e., “syn-” and “anti-conditions”, respectively). All calculations were carried out at the M06-D3/SDD(Cu)–6–311+G(d,p)/SMD//M06-D3/SDD(Cu)–6–31G(d)/SMD level of theory.

The stability of the 6-membered transition state for syn-nucleocupration compared to alternative pathways explains why other nucleophiles in Table 4 failed in intermolecular syn-selective addition. For non-carboxylate oxygen nucleophiles, an analogous six-membered transition state is inaccessible due to the presence of only one nucleophilic/coordinating oxygen atom. In the case of nitrogen nucleophiles, we hypothesize that since nitrogen is more strongly coordinating than oxygen, formation of the desired six-membered transition state—with the oxygen coordinating to copper and the nitrogen acting as the nucleophile—is less favorable.

Next, we computed the reaction pathways of the same π-alkyne–Cu(II) complex 7 in DMA solvent (Figure 1, right). The barriers to the two inner-sphere nucleocupration transition states (TS1a and TS1b) are similar to those computed in 1,4-dioxane solvent. However, the computed barrier to the outer-sphere anti-nucleocupration (TS1c’) is much lower (ΔG^‡ = 18.7 kcal/mol with respect to 7) than that in 1,4-dioxane (ΔG^‡ = 34.1 kcal/mol). The stabilization of TS1c’ is due to the combined effects of strong solvent coordination to the Cu(II) center and the electrostatic stabilization of the more polar transition state structure in the polar solvent (ε = 37.8 and 2.2 for DMA and 1,4-dioxane, respectively). The Gibbs free energy of the anti-nucleocupration transition state TS1c’ is 0.8 kcal/mol lower than that of TS1a, which is consistent with the moderate anti-selectivity observed in DMA in the absence of additives. Because the DIPEA additive further enhanced the anti-selectivity in DMA solvent (Table 1), we computationally explored two potential roles of the DIPEA additive. Explicit coordination of DIPEA to the Cu(II) center in place of the DMA solvent molecule was found to be unfavorable (Figure S1B), due to the steric bulk of the tertiary amine. Instead, the protonated diisopropylethylammonium cation (DIPEA•H⁺) could coordinate with the carbonyl oxygen of the benzoate anion nucleophile in the outer-sphere anti-nucleocupration (TS1d). The DIPEA•H⁺ coordination promotes the benzoate anion dissociation from 7 (Figure S1B) while also lowers the overall barrier to the outer-sphere anti-nucleocupration by stabilizing the charge-separated transition state. The DIPEA-assisted anti-nucleocupration (TS1d) requires a low barrier of only 11.2 kcal/mol, which is consistent with the high levels of anti-selectivity under these experimental conditions. Similar to the reaction in 1,4-dioxane, BzOH-promoted protodecupration of the alkenyl Cu(II) intermediate 11 is highly facile (via TS3), which confirms that the trans-selective nucleocupration is the product stereoselectivity-determining step.

To further explore the role of the ammonium carboxylate in anti-addition implicated by the DFT data, we performed an experiment in which DIPEA and benzoic acid were replaced with tetrabutylammonium benzoate (1 equiv) and coupled with 1a under anti-selective conditions (entry 15, Table 1). This modification led to comparable yield and excellent Z-selectivity of 3aa. This data is consistent with the proposed role of the ammonium cation in charge stabilization and reveals that the hydrogen-bonding interaction between DIPEA•H⁺ and the carboxylate anion is not a strict requirement for the reactivity and selectivity.

Effective reaction partners in the anti-addition system must be sufficiently acidic to be deprotonated with the amine base. In particular, competent nucleophiles possess a pK_a value in the 1.0–6.0 range in H₂O. Reactive carboxylic acids span from phenylglyoxylic acid (pK_a = 1.21) to pivalic acid (pK_a = 5.03). Likewise, electron-poor pentafluorophenol (pK_a = 5.5) reacts, whereas unsubstituted phenol (pK_a = 10) does not. Reactive nitrogen nucleophiles fall within a similar pK_a window.

In conclusion, we develop a versatile platform employing Cu(II) as π-Lewis acid to achieve of electronically unactivated alkynes with diverse OH- and NH-nucleophiles. Regioselectivity and kinetic reactivity are facilitated by a bidentate auxiliary, and syn- or anti-stereoselectivity of benzoic acid coupling partners is controlled through judicious tuning of reaction conditions. An alkyne substrate bearing a mono-dentate amide also exhibits reactivity under a more forcing condition. Mechanistic experiments and DFT calculations support an inner-sphere nucleometallation mechanism via a 6-membered transition state in syn-addition and highlight the essential roles of DIPEA and DMA in the anti-addition pathway. We anticipate that this Cu(II)-catalyzed π-activation strategy can be generalized to other π-bond containing substrates, providing a sustainable alternative to precious metal catalysts.

Supplementary Material

Supporting Information. Experimental details and details of computational studies. This material is available free of charge via the Internet at http://pubs.acs.org.

Scheme 5.

Effect of ammonium salt in anti-condition