An oxetane can stand in for a diaryl ketone
A 3,3-diaryloxetane is a four-membered oxygen ring bearing two aryl groups on the same carbon. The paper treats this motif as a bioisostere of a diaryl ketone: the strained ring's polarized C–O bonds reproduce part of the ketone's polarity without the planar carbonyl. The authors cite a dipole moment of 1.93 ± 0.01 D for oxetane against roughly 2.7–3.0 D typical of diaryl ketones, so the replacement preserves overall polarity while creating a three-dimensional, tetrahedral centre.
That electronic resemblance does not carry over to reactivity. A ketone's C=O π bond gives the carbonyl carbon a low-lying π* orbital and makes it electrophilic; an oxetane has no π system, and its polarity comes from σ-bond polarization spread around the ring. No oxetane carbon is comparably reactive. This is why the paper needs a special activating group rather than simply borrowing a carbonyl substitution reaction.
Why Friedel–Crafts routes restrict the aryl scope
The established route to 3,3-diaryloxetanes is a Lewis acid-promoted Friedel–Crafts reaction through oxetanyl cation intermediates, later extended with iron catalysis. The paper states that these reactions are largely restricted to electron-rich arenes, require strongly acidic conditions, and can suffer from ring-opening, rearrangement and competing side reactions. Decarboxylative coupling of oxetane-3-carboxylic acid derivatives is another option, but it relies on tertiary benzylic radicals whose stability and steric hindrance reduce the driving force for addition to a metal, and it needs multistep precursor synthesis.
The practical consequence is a scope problem, not just a yield problem. The new method's authors argue that existing strategies depend on polar bond disconnections and electronically biased substrates, which limits modularity. Their target is therefore a route in which the two aryl groups can be varied independently rather than dictated by how nucleophilic each arene happens to be.
A tetrafluoropyridyl ether makes the C–O bond reactive
The first step lithiates an aryl bromide, adds 3-oxetanone, then traps the resulting alkoxide with pentafluoropyridine after adding sodium hydride. This one-pot sequence gives a 3-aryl oxetan-3-yl tetrafluoropyridyl ether; for the 2-naphthyl example the yield was 89% on a 6.2 g (17.8 mmol) scale, and the reagent was stored at room temperature for at least six months without detectable decomposition. The sodium hydride addition raised the yield by about 46% relative to the lithium alkoxide alone.
The tetrafluoropyridyl group is the design choice that makes the rest work. It polarizes the exocyclic benzylic C–O bond and stabilizes the departing group through π delocalization, which the authors say facilitates oxidative addition. Oxidative addition is proposed to occur preferentially at that exocyclic bond, where developing benzylic character is stabilized by the adjacent aromatic ring, while the ring's endocyclic C–O bonds lack comparable stabilization and stay intact. The leaving group's low nucleophilicity and weak basicity are also proposed to suppress ring-opening. Alternative leaving groups, including sulfonates and carboxylates, gave substrates too unstable to isolate or no productive coupling.
Nickel, base and isobutanol are each required
The coupling of the naphthyl tetrafluoropyridyl ether with 4-methoxyphenylboronic ester reached 99% yield using Ni(cod)2 at 10 mol%, K3PO4 at 2.0 equivalents and iBuOH at 2.0 equivalents in THF at 30 °C. Control experiments show how tightly the conditions are defined: omitting Ni(cod)2 or K3PO4 gave no detectable product, removing iBuOH lowered the yield to 73%, and running under air gave below 3%. NiBr2·diglyme and Pd(PPh3)4 each gave below 3%, while the air-stable Ni(4-tBustb)3 complex gave 46%. Strongly coordinating ligands such as dtbbpy and dppf suppressed reactivity entirely.
The iBuOH effect is attributed to promoting transmetalation of the boronic ester, and the ligand results are read as evidence that strongly coordinating ligands inhibit formation of the reactive nickel species. These are the authors' interpretations of control experiments, not isolated intermediates. The practical point for a reader is that this is a nickel-specific, additive-dependent reaction rather than a general palladium-style coupling.
Both aryl groups can be varied independently
The first step converted a wide range of aryl bromides into tetrafluoropyridyl ethers in 80–95% yields on a 5 mmol scale, covering electron-rich and electron-deficient arenes. Heteroaryl groups were also accessible, including indole, benzothiophene, benzofuran, furan and quinoline, with 28–81% yields. Pyridine and pyrimidine derivatives failed, giving target products in less than 10% yield.
The second step tolerated electron-rich and electron-poor arylboronic esters, as well as heteroaryl and alkenyl boronic esters. The authors highlight that electron-poor arenes, heteroarenes and alkenyl substituents couple efficiently, in contrast to Friedel–Crafts reactions where such substrates are typically unreactive. That contrast is the paper's central scope claim: reactivity is governed by the transition-metal cycle rather than by the intrinsic nucleophilicity of the arene. A methyl-ester diaryloxetane was made in 95% yield on a 1.2 g (3.8 mmol) scale, while the benzothiophene-containing substrate reached only 26%, possibly because sulfur coordinates to nickel.
Evidence points to a polar Ni(0)/Ni(II) cycle
Both catalytic and stoichiometric Ni(cod)2 promoted the reaction above 95% yield. Adding 10 mol% of the one-electron oxidant FeCp2BArF to a stoichiometric nickel reaction still gave 93% product, but adding a full equivalent suppressed reactivity, which the authors read as Ni(I) being insufficient to sustain turnover and Ni(0) being the catalytically relevant species. The radical scavenger BHT at 2.0 equivalents had little effect, whereas TEMPO completely inhibited the reaction with quantitative recovery of the starting ether. The authors caution that TEMPO also reacts readily with low-valent nickel, so inhibition could reflect oxidation rather than radical trapping.
A Hammett analysis of para-substituted aryl ethers gave a shallow negative reaction constant (ρ = –0.49), indicating only weak dependence on aryl electronics. The proposed oxidative addition complex was not isolated or detected by NMR, possibly because it is intrinsically unstable. The working mechanism is therefore a Ni(0)/Ni(II) polar sequence of oxidative addition, transmetalation and reductive elimination, distinct from the Ni(I)/Ni(II)/Ni(III) manifolds often invoked for radical cross-couplings. The authors present this as supported by the observations, not as a directly observed intermediate.
Two-step routes beat earlier sequences on yield
The method was applied to oxetane analogues of marketed compounds, including derivatives related to ketoprofen, oxybenzone, fenofibrate and a KIT inhibitor, and was extended to 3,3-diaryl azetidines. Attempts to include 3,3-diaryl cyclobutanes and other diaryl quaternary architectures failed because the corresponding tetrafluoropyridyl ethers were unstable. Late-stage variation of the boronic ester component was tolerated with fenbufen, a benzyloxazolidinone, estrone, tocopherol, flurbiprofen and pyriproxyfen.
The most decision-relevant comparison is step count and overall yield. A KIT inhibitor analogue was made in 20% overall yield over two steps from 3-oxetanone, against 0.1% over four steps in a prior approach. A diaryloxetane intermediate was obtained in 58% over two steps versus 22% over four steps previously, and an adiporon analogue was reached in 65% yield over two steps after deprotection and amide formation. These are specific matched comparisons from the paper, not a general claim that the route is superior for every target. The priority the paper establishes is modular installation of two independently chosen aryl groups at a fully substituted oxetane carbon, with the clearest advantage where electron-poor or heteroaryl groups and late-stage diversification are needed.
Sources used in this answer
Modular synthesis of 3,3-diaryloxetanes
A two-step route from 3-oxetanone via an oxetanyl tetrafluoropyridyl ether enables nickel-catalysed cross-coupling of two different aryl groups, giving 3,3-diaryloxetanes in up to 99% coupling yield and improving matched step-count and overall-yield comparisons against prior routes.