The chemistry at stake is one of the most consequential in modern synthesis. Carbon(sp2)–heteroatom bonds — links between an aromatic ring and nitrogen, oxygen, sulfur or phosphorus — appear throughout pharmaceuticals, agrochemicals and functional materials. For decades, forming them reliably has meant palladium catalysis operating through a Pd(0)/Pd(II) cycle. That manifold demands exquisitely tuned ligands, because two steps pull in opposite directions: oxidative addition, in which the metal inserts into the aryl halide bond, favours electron-rich, sterically accessible metal centres, while reductive elimination, the bond-forming step, often benefits from the opposite electronic character. Because these demands are orthogonal, chemists routinely face extensive ligand screening — increasingly assisted by high-throughput experimentation or machine learning — for each new substrate class.

Nickel offers an elegant escape from this tug-of-war. In a Ni(I)/Ni(III) cycle, the bond-forming reductive elimination from Ni(III) is intrinsically favourable, proceeding with a low activation barrier and often without any elaborate ligand at all. The bottleneck shifts instead to the front end of the cycle: Ni(I) is a tempering, highly reactive metalloradical whose oxidative addition into aryl halides is often too sluggish, particularly for electron-rich aryl bromides, the very substrates that palladium struggles with too. Worse, Ni(I) species tend to die in unproductive ways — forming inactive dimers, resting states or insoluble nickel black — which forces chemists to compensate with high catalyst loadings, sometimes several mole percent, undermining the economic appeal of a non-precious metal.

The Pieber group had previously shown that so-called donor–acceptor ligands allow Ni(I) to be generated simply by irradiating a bench-stable Ni(II) precatalyst with blue light, through an intraligand charge-transfer mechanism. But those systems, like other contemporary Ni(I)/Ni(III) protocols, still could not generally couple electron-rich aryl bromides. An alternative strategy developed by Cornella and Ritter sidestepped oxidative addition altogether by using aryl thianthrenium salts that fragment under single-electron transfer — but that approach falters with electron-poor substrates and demands specialized starting materials. Sterically encumbered nucleophiles such as tertiary alcohols remained out of reach for everyone.

The Austrian team’s solution rests on two design decisions, each grounded in mechanism. The first concerns the additive. Earlier work had shown that the amidine base 1,1,3,3-tetramethylguanidine (TMG) helps weak nucleophiles in dual nickel/photoredox couplings, but TMG is itself a nucleophile and competes in the reaction, which likely explains poor results with secondary alcohols and the complete absence of tertiary alcohol couplings. The researchers reasoned that a truly non-nucleophilic additive was needed. They settled on 2-tert-butyl-1,1,3,3-tetramethylguanidine (BTMG), a TMG derivative whose tert-butyl group blocks its nucleophilic site. Experiments showed that BTMG coordinates reversibly to the standard Ni(II) precatalyst Ni(dtbbpy)Cl2, creating a new absorption band reaching into the visible spectrum. Electron paramagnetic resonance measurements then delivered the key observation: a paramagnetic signal consistent with Ni(I) appeared only when the nickel complex, BTMG and an alcohol were combined under irradiation — never when any component was missing. In other words, BTMG enables light absorption, while the nucleophile’s coordination opens the door to Ni(I) formation, all without any exogenous photoredox catalyst.

The second decision concerns the ligand itself. Building on work by Hadt and coworkers showing that oxidative addition rates of Ni(I) into aryl halides correlate with the energy of the metal’s 3d(z2) orbital, the team hypothesized that strongly electron-donating substituents on a 2,2′-bipyridine scaffold would raise that orbital energy and supercharge oxidative addition. Computational comparisons confirmed the trend across a series of bipyridines. But there was a catch: a hyperactive Ni(I) metalloradical is also a fragile one, prone to dimerization and decomposition. The ideal ligand must therefore balance reactivity with stability. The researchers’ answer was a rationally designed ligand they call dpabpy — 4,4′-bis(diphenylamino)-2,2′-bipyridine. Its diphenylamino substituents are powerful electron donors, boosting the nucleophilicity of the Ni(I) centre, while their extended conjugation creates a low-lying π* orbital that can host the unpaired electron as a ligand-centred radical on a square-planar Ni(II) species. Cyclic voltammetry showed that dpabpy’s reduction potential remains comparable to ordinary bipyridines, meaning the ligand can still be readily reduced to generate Ni(I) — the best of both electronic worlds.

The benchmarking experiments were striking. In carbon–nitrogen couplings across a series of para-substituted aryl bromides, the standard ligand dtbbpy showed yields that fell steadily as the aryl bromide became more electron-rich. Methoxy- and dimethylamino-substituted bipyridines improved matters but still struggled with 4-bromoanisole. Dpabpy outperformed every competitor across the board. More remarkably, at catalyst loadings of 0.1 mole percent — and in one case even 0.01 mole percent — dpabpy delivered near-quantitative product formation with challenging aryl bromides, and it retained its advantage when the reaction time was slashed to a single hour. A light on–off experiment revealed that the reaction halts in the dark, indicating that thermal pathways regenerating Ni(II) resting states are fast, and that the ligand’s steric protection is what keeps Ni(I) from sliding into irreversible dimerization and nickel black. Control studies confirmed that BTMG, dpabpy, nickel and light are each essential.

The substrate scope reads like a medicinal chemist’s wish list. With 1 mole percent of Ni(dpabpy)Cl2 and BTMG at 45 degrees Celsius, aryl bromides bearing nitriles, amines, cyclopropyl groups, sulfones, boronic esters and amides all coupled efficiently with benzyl alcohol or benzyl amine. Six-membered heteroaryl bromides performed well, as did a wide range of five-membered heterocycles including thiophenes, indazoles, benzothiazoles, pyrazoles and triazoles — though some electron-poor pyrroles, oxazoles and imidazoles succumbed to protodehalogenation instead. Primary, secondary, benzylic, allylic and aliphatic alcohols gave yields from 51 to 99 percent, and even water could serve as the nucleophile to make phenols. Most strikingly, tertiary alcohols — long considered unreactive in Ni(I)/Ni(III) couplings — participated successfully, as did sterically demanding secondary amines and α,α,α-trisubstituted primary amines. Beyond oxygen and nitrogen, the platform handled phosphine oxides, amides, carbamates, thiols, sulfinate salts, carboxylic acids, chloride, sulfoximines and sulfonamides.

The practical demonstrations are where the chemistry becomes genuinely viral material for the synthesis community. Late-stage functionalizations succeeded on amoxapine, perphenazine, leelamine, linagliptin and cholesterol as nucleophiles, and on derivatives of tocopherol, celecoxib, D-glucose and flumazenil as electrophiles — including a cross-coupling that stitched a tocopherol derivative directly to linagliptin. On scale, two consecutive coupling steps delivered the antipruritic drug pramocaine in a telescoped one-pot sequence with an overall yield of 47 percent, exploiting the selective mono-etherification of 1,4-dibromobenzene at just 0.1 mole percent catalyst. Even more eye-catching, the N-Boc-protected antidepressant fluoxetine was prepared on a 3.65-gram scale in 89 percent isolated yield using only 100 parts per million of nickel — roughly a thousandth of the metal loading typical of many nickel-catalysed couplings.

What makes this work resonate beyond its impressive numbers is the conceptual lesson: in Ni(I)/Ni(III) catalysis, the ligand’s job is not to force reductive elimination, as in palladium chemistry, but to keep a reactive radical in check while feeding it substrates it can actually activate. By tuning donor strength and redox activity in a single bipyridine scaffold, and pairing it with a non-nucleophilic base that turns visible light into a Ni(I) source from a bench-stable precatalyst, the ISTA team has shown that enhancing oxidative addition is the key that unlocks a general, low-loading, photocatalyst-free platform for carbon–heteroatom bond formation. With a patent application filed and gram-scale drug synthesis already demonstrated, the method seems poised to move quickly from journal pages to medicinal chemistry labs — a reminder that sometimes the biggest leaps in catalysis come not from new metals, but from rethinking the molecules that surround them.

Subject of Research: Mechanistically guided ligand design for low-loading Ni(I)/Ni(III)-catalysed C(sp2)–heteroatom cross-coupling of aryl bromides

Article Title: Ligand design broadens NiI-catalysed C(sp2)–heteroatom couplings of aryl bromides at low catalyst loadings

Article References: Bena, A. R., Banik, T., Giannoudis, C., Ortis, F., Bím, D., Baunis, H., Palissery, G. H., & Pieber, B. (2026). Ligand design broadens NiI-catalysed C(sp2)–heteroatom couplings of aryl bromides at low catalyst loadings. Nature Catalysis. https://doi.org/10.1038/s41929-026-01616-6

Image Credits: AI Generated

DOI: 10.1038/s41929-026-01616-6

Keywords: nickel catalysis, cross-coupling, ligand design, bipyridine, photocatalysis, C–N coupling, C–O coupling, aryl bromides, oxidative addition, late-stage functionalization, pharmaceutical synthesis, Nature Catalysis