Metallaphotoredox-Catalyzed sp³–sp³ Cross-Coupling of Carboxylic Acids with Alkyl Halides

Abstract

Over the last half-century, transition metal-mediated cross-coupling reactions have changed the way in which complex organic molecules are synthesized. Indeed, the predictable and chemoselective nature of these transformations has led to their widespread adoption across a vast array of chemical research areas¹. However, the construction of sp³–sp³ bonds, a fundamental unit of organic chemistry, remains an important yet elusive objective for cross-coupling reaction engineering². In comparison to related procedures with sp²-hybridized species, the development of methods for sp³–sp³ bond formation via transition metal catalysis has been historically hampered by deleterious side-reactions, such as β-hydride elimination with Pd-catalysis, and the reluctance of alkyl halides to undergo oxidative addition^3,4. To address this issue, a number of research groups have demonstrated the feasibility of nickel-catalyzed cross-coupling processes to form sp³–sp³ bonds that utilize organometallic nucleophiles and alkyl electrophiles^5–7. In particular, the coupling of alkyl halides with pregenerated organozinc^8–10, Grignard^11,12, and organoborane¹³ species has been used to furnish diverse molecular structures. However, the poor step and atom economies along with the operational difficulties associated with making, carrying, and using these sensitive coupling partners has hindered their widespread adoption. The prospect of establishing a generically useful sp³–sp³ coupling technology that employs bench-stable, native organic functional groups, without the need for pre-functionalization or substrate derivatization, would therefore be a valuable addition to fields of research that rely on organic molecule construction. Here, we demonstrate that the synergistic merger of photoredox and nickel catalysis enables the direct formation of sp³–sp³ bonds using only simple carboxylic acids and alkyl halides as the nucleophilic and electrophilic coupling partners, respectively. The outlined protocol is suitable for a wide array of primary and secondary carboxylic acids and does not require the presence of radical stabilizing groups. The merit of this coupling strategy is illustrated by the expedient synthesis of the pharmaceutical tirofiban in four steps from commercially available starting materials.

Within the field of drug discovery, there is a demonstrated statistical correlation between clinical success and the molecular complexity of medicinal candidates with respect to the inherent ratio of sp²–sp³ to sp³–sp³ bond content¹⁴. Not surprisingly, these findings have created an emerging demand within medicinal chemistry for new reaction technologies that enable rapid access to drug-like molecules via the coupling of fragments that incorporate or build novel sp³–sp³ bonds. However, a major hurdle associated with achieving sp³–sp³ bond formation via transition metal catalysis is the limited availability of a diverse suite of nucleophilic coupling partners that are bench-stable, inexpensive, and easily procured. An attractive option would be to employ simple carboxylic acids, an abundant native functional group that is chemically robust yet can be readily exploited as a latent leaving group after multistep synthetic sequences (Fig. 1).

Figure 1. Carboxylic acids as coupling partners in a metallaphotoredox-mediated process to form sp³–sp³ bonds.

The majority of transition metal-catalyzed cross-couplings generally employ at least one sp^2-hydridized coupling partner. The direct utilization of robust, widely available, native functional groups such as carboxylic acids should encourage greater adoption of sp³–sp³ bond forming methods.

The emergence of visible light-mediated photoredox catalysis within the field of synthetic organic chemistry has enabled the discovery and invention of numerous unique and valuable transformations^15,16. Indeed, the electronic duality of photocatalyst excited states (which are simultaneously strong oxidants and reductants) has prompted the exploitation of these polypyridyl transition metal complexes in unconventional bond disconnections¹⁷ and facilitated the manipulation of oxidation states in organometallic complexes to enable previously elusive reactions^18–22. For example, the synergistic merger of photon-driven single-electron transfer (SET) processes with nickel-activated electrophiles has promoted the formation of valuable sp²–sp³ bonds whilst broadening the field of cross-coupling chemistry via the use of non-conventional reaction substrates^23,24. On this basis, we hypothesized that a straightforward and generic procedure might be developed to enable analogous sp³–sp³ bond formations via the application of ubiquitous carboxylic acids in a decarboxylative cross-coupling with alkyl halides using photoredox catalysis. Moreover, we hoped to introduce a paradigm for carbon–carbon bond construction that would (i) provide rapid access to complex molecular fragments via sp³–sp³ coupling, (ii) systematically streamline synthetic routes towards drug candidates, and (iii) enable alkyl–alkyl cross-coupling broadly in medicinal, process, and natural product chemistry programs.

A detailed mechanism for the proposed decarboxylative sp³–sp³ coupling is delineated in Fig. 2. Initial visible-light excitation of the iridium(III) photocatalyst Ir[dF(CF₃)ppy]₂(dtbbpy)PF₆ [dF(CF₃)ppy = 2-(2,4-difluorophenyl)-5-(trifluoromethyl)pyridine, dtbbpy = 4,4′-di-tert-butyl-2,2′-bipyridine] (1) would generate the long-lived (τ = 2.3 μs)²⁵ excited-state *Ir^III complex 2. This species is a strong single-electron oxidant (E_1/2^red [*_Ir^III_/Ir^II] = +1.21 V vs. SCE in CH₃CN)²⁵ and should undergo reduction by a carboxylate anion derived from deprotonation of the acid 3. The resultant carboxyl radical is expected to rapidly extrude CO₂ to produce alkyl radical 4 and the reduced Ir^II catalyst 5_. Concurrently, the ligated nickel(0) complex 6 is generated in situ via two discrete SET reductions of (dtbbpy)Ni(II)Cl₂ by the iridium(II) state of the photocatalyst through sacrificial carboxylic acid consumption (E_1/2^red [Ir^III/Ir^II] =−1.37 V vs. SCE in CH₃CN, (E_1/2^red [Ni^II/Ni⁰] = −1.2 V vs. SCE in DMF)^25,26. The Ni⁰ complex 6 can intercept radical 4 to produce alkyl-Ni^I intermediate 7²⁷. Subsequent oxidative addition with alkyl halide 8 forms the putative organometallic Ni^III species 9, which after reductive elimination forges the desired sp³–sp³ bond to furnish the coupled product 10 and Ni^I adduct 11^28–30. At this stage the two catalytic cycles would converge by reduction of nickel(I) intermediate 11 by the reduced form of the iridium photocatalyst to reestablish both the Ir^III complex 1 and the Ni⁰ catalyst 6²⁵. Presently, we cannot rule the possibility of an alternative mechanism that involves Ni⁰-mediated oxidative addition and trapping of the alkyl radical 4 by a Ni^II-species^23,27.

Figure 2. Proposed mechanism for the metallaphotoredox-mediated cross-coupling of carboxylic acids to generate sp³–sp³ bonds.

The photoredox catalytic cycle commences with excitation of Ir^III 1 to give the excited state 2. Single electron oxidation of the carboxylate anion derived from acid 3 produces the alkyl radical 4 after CO₂-extrusion along with Ir^II 5. Ni⁰ catalyst 6 captures the alkyl radical 4 to form the Ni^I species 7. Ensuing oxidative addition with alkyl halide 8 leads to nickel(III) intermediate 9. Reductive elimination would then liberate the desired product 10 and Ni^I 11. Both catalytic cycles converge to complete a single turnover via a SET event that regenerates the photoredox and nickel catalysts

Predicated on our envisioned decarboxylative sp³–sp³ coupling mechanism, we began our primary investigations with consideration of the metallaphotoredox conditions previously developed within our group²³. In these studies we employed N-Boc proline and 1-bromo-3-phenylpropane as coupling partners. Unfortunately, owing to the basic reaction conditions in combination with the polar aprotic solvent dimethylformamide, exclusive ester formation was observed. Therefore, we recognized that judicious selection of solvent and base would be necessary to suppress this unwanted byproduct formation without obstructing the desired photocatalytic oxidation reaction pathway. To accomplish this goal a survey of solvents and inorganic bases was conducted in the presence of Ir[dF(CF₃)ppy]₂(dtbbpy)PF₆, NiCl₂×glyme, and dtbbpy with visible light irradiation from blue light-emitting diodes (LEDs). Pleasingly, this revealed that the combination of acetonitrile and K₂CO₃ greatly reduced the rate of carboxylate alkylation and furnished the desired product in 68% yield. Further optimization improved conversion to product through switching to the more electron-rich ligand 4,4′-dOMe-bpy (4,4′-dimethoxy-2,2′-bipyridine). Finally, the addition of water, which decelerated ester formation, provided an additional enhancement and an isolated yield of 85%. A series of control experiments omitting each individual reaction component highlighted the importance of photocatalyst, nickel, and light for promoting this decarboxylative sp³–sp³ bond-forming reaction (see Supplementary Information).

With optimal metallaphotoredox conditions in hand, we probed the generality of this process with respect to the aliphatic halide electrophile. As representative nucleophilic coupling partners, N-Boc- and N-Cbz proline were utilized interchangeably for this purpose (Fig. 3). Simple unfunctionalized primary alkyl halides such as 1-bromo-3-phenylpropane and 1-bromobutane were examined, and these underwent efficient cross-coupling (12 and 13, 85% and 76% yield, respectively). A substrate bearing a terminal olefin performed favorably under the reaction conditions to deliver the desired product (14, 84% yield). As this protocol is conducted at near ambient temperature, hydrolysis of an ethyl ester does not occur under the optimized basic reaction conditions, and the corresponding carbamate-protected amine was isolated in good yield (15, 64% yield). In addition, perfect chemoselective functionalization of an alkyl bromide in the presence of a primary chloroalkane was observed (16, 96% yield). A deprotection-cyclization sequence with this material would provide rapid access to the bicyclic tertiary amine core of the naturally occurring pyrrolizidine alkaloids³¹. The presence of free hydroxyl groups is fully compatible with the iridium photocatalyst and the nickel complex (17, 86% yield). Moreover, reactive Lewis basic functionalities, such as epoxides and aldehydes, are well tolerated in this cross-coupling procedure and provide numerous opportunities for further derivatization (18 and 19, 83% and 62% yield, respectively). The influence of substitution on the alkyl halide was also investigated to probe the steric limits of the electrophilic coupling partner. No detrimental effects to the efficiency of this process were observed when a β,β-disubstituted bromoalkane was employed, and pleasingly, neopentyl bromide coupled in modest yield (20 and 21, 75% and 52% yield, respectively). The higher propensity for activated electrophiles to promote esterification led to the utilization of benzyl chloride, as opposed to benzyl bromide, which generated a homobenzylic amine in good yield (22, 84% yield). Notably, bromomethane was a competent coupling partner in this protocol which formally affords the product of a fully reduced carboxylic acid moiety in a single step (23, 62% yield).

Figure 3. Carboxylic acid and alkyl halide scope in the dual nickel-photoredox catalyzed sp³–sp³ coupling reaction.

A broad array of alkyl halides and carboxylic acids are amenable coupling partners in this transformation. Top, generalized reaction; bottom, substrate scope. Primary and secondary electrophiles were coupled efficiently with proline derivatives. Alternative α-heteroatom substituted carboxylic acids could also be employed to form functionalized amines and ethers. Challenging substrates lacking apparent radical stabilization could also be employed successfully. Isolated yields are reported below each entry. See Supplementary Information for full experimental details. *Reaction run in flow (GC yield), see S.I. ^†Cyclopropylacetic acid was used.

Expanding the substrate scope to encompass secondary alkyl halides permitted us to forge sp³–sp³ bonds with adjacent tertiary carbon centers. For example, five- and six-membered cyclic bromoalkanes smoothly reacted to form the desired alkylated products in good to excellent yields (24–26, 57–91% yield). Smaller ring systems, including cyclopropane and oxetane, were also introduced via this metallaphotoredox procedure (27 and 28, 50% and 79% yield, respectively). These motifs have found application in drug discovery programs as chemically and metabolically stable bioisosteres³². Lastly, an acyclic secondary alkyl bromide was also successfully cross-coupled to generate the desired Cbz-protected amine (29, 69% yield).

We subsequently examined the scope of the nucleophilic component and we were pleased to discover that an assortment of readily available carboxylic acids were viable for this transformation. For instance, Boc-protected pipecolic acid and an azetidine derivative both underwent decarboxylative coupling to furnish alkylated products in good yields (30 and 31, 61% and 70% yield, respectively). Naturally occurring amino acids, which are inexpensive and obtainable from ample biomass feedstocks, can also be exploited to form α-functionalized amines with excellent efficiency (32 and 33, 72% and 71% yield, respectively). Similarly, O-methylated glycolic acid functions well to provide access to linear ethers in a straightforward manner (34, 61% yield). The cyclic substrate tetrahydrofuran-2-carboxylic acid was also coupled with 1-bromo-3-phenylpropane under these dual nickel-photoredox conditions to afford the ethereal product in very good yield (35, 74% yield). Although beneficial, the inclusion of an α-heteroatom on the acid fragment is not a prerequisite for this sp³–sp³ bond forming process. For example, Cbz-protected isonipecotic acid and a tetrahydro-2H-pyran derivative were cleanly converted to the corresponding coupled products in an effective fashion (36 and 37, 62% and 66% yield, respectively). Simple alkyl precursors lacking heteroatoms can also be utilized, with cyclohexanecarboxylic acid reacting in reasonable yield (38, 52% yield). An acyclic β-amino acid that would generate a secondary radical upon decarboxylation exhibited respectable efficiency and offers the opportunity to synthesize β-functionalized amines with ease (39, 58% yield).

Finally, two primary substrates were evaluated in this system to fully exemplify the power of this new sp³–sp³ coupling paradigm. The rearrangement of a cyclopropyl system under the reaction conditions presents the opportunity to produce alkylated homoallylic products in a single step (40, 43% yield). Moreover, the monomethyl ester of glutaric acid was subjected to the optimized metallaphotoredox procedure and provided an encouraging quantity of the desired product (41, 40% yield). This substrate highlights the potential for downstream modification of latent carboxylates since hydrolysis of the methyl ester would unlock the potential for further sp³–sp³ bond formation. Thus, molecules that contain multiple carboxylic acids can function as linchpin reagents for the rapid assembly of complex molecular architectures. It should be noted that when tertiary acids were applied, the coupled products could only be obtained limited efficiencies (~5–10%). Attempts to improve the yields to synthetically useful levels are currently ongoing.

To further demonstrate the synthetic utility of this decarboxylative coupling protocol, we applied it to the synthesis of the antiplatelet drug tirofiban³³. As shown in Fig. 4, Boc-isonipecotic acid 42 and alkyl bromide 43 (protected to avoid cyclization to THF) were exposed to the optimized metallaphotoredox conditions to afford alcohol 44 in good yield, following deprotection of the silyl ether. Thereafter, utilization of the previously established dual photoredox-nickel catalytic etherification reaction enabled direct formation of the desired C–O bond¹⁸. Following acidic deprotection, tirofiban 45 was synthesized in 59% yield over the final two steps.

Figure 4. Application of two metallaphotoredox strategies to the synthesis of tirofiban.

The cross-coupling of acid 42 and alkyl halide 43 generates a new sp³–sp³ bond and subsequent TBAF deprotection provides alcohol 44. Tirofiban 45 is then synthesized in two further steps in good yield via a Ni-photoredox-mediated etherification reaction and acidic deprotection. ^†34% of bromide starting material recovered.

In summary, we have established a robust strategy for the direct formation of sp³–sp³ bonds from abundant carboxylic acids and alkyl halides. This new platform for carbon–carbon bond construction is enabled by the catalytic activation of both coupling partners through the synergistic merger of photoredox and nickel catalysis. The benign nature of the reaction conditions has been exemplified by the breadth of functional groups tolerated in this transformation. We believe that the generality of this methodology and the ready availability of the starting materials used will aid the uptake of sp³–sp³ cross-coupling across several fields of synthetic organic chemistry.