Transient axial-to-point chirality transfer: Pd/Sadphos spiroannulation of phenolic biaryls

From single chirality transfer to a sequential induction-and-transfer cascade

Transient axial-to-point chirality transfer has become a productive strategy for building spiro quaternary carbons from racemic or prochiral arenes. Since the 2015 reports from Luan and You, asymmetric formal spiroannulation has been used to construct chiral spirocycles by generating an axially chiral intermediate and then converting that axial element into a point stereocenter [1]. Most developed systems, however, remain limited to a single chirality transfer event and typically deliver products with only one stereocenter [1]. The anchor paper extends this paradigm by using a Pd/Sadphos catalyst to induce axial chirality in a racemic phenolic biaryl, then transfer it while a second stereochemical event is set, assembling three contiguous stereocenters in one step [1].

The mechanistic proposal is a DyKAT manifold: the two atropisomers of racemic biaryl 1 interconvert rapidly through reversible π-coordination to Pd(0), and the chiral catalyst selectively promotes atroposelective oxidative addition of the reactive (Sa)-atropisomer [1]. A stereospecific syn-migratory insertion of the 1,3-diene then gives a chiral σ-allylpalladium species, which isomerizes to a π-allyl complex; deprotonation of the pendant naphthol triggers dearomatization, and intramolecular attack at the less hindered π-allylic carbon followed by reductive elimination closes the spirocycle [1]. This is an interpretation supported by the observed sense of stereoinduction and by DFT-derived atropisomerization barriers, not a directly observed catalytic cycle.

What the substrate scope does and does not show

Under the standard conditions, more than 40 polycyclic spiro enones were obtained in moderate to high yields with excellent regio-, diastereo- and enantioselectivity, up to 99% ee [1]. Phenolic biaryls bearing electron-donating, electron-withdrawing and halide substituents delivered products in 81–98% ee, and substituted 1-arylcyclohexadienes with fluoro, chloro, methyl, phenyl, methoxyl, methylthio, trimethylsilyl, trifluoromethoxy and trifluoromethyl groups gave 84–99% ee [1]. The reaction also tolerated 2-naphthyl, dioxaphthyl, benzofuryl, benzothienyl, thienyl, benzyl, n-butyl and cyclohexyl-derived dienes, as well as cyclohexadiene itself, in 91–97% ee [1]. A gram-scale reaction furnished 1.2 g of 3aa in 63% yield with 96% ee, and the spiro enone could be reduced, rearranged or epoxidized with retention of enantioselectivity [1].

The boundary is equally informative. Acyclic 1,3-dienes with a terminal phenyl substituent were much less reactive and gave only 52% yield with 32% ee, which the authors explicitly flag as a signal that new ligands should be screened [1]. The method is therefore demonstrated for the tested phenolic biaryl and cyclic diene classes, not for all spirocyclic skeletons, and the acyclic diene result shows that the stereodetermining insertion step is sensitive to diene substitution. The reported ee values also come from HPLC analysis of isolated products under the optimized conditions, so they do not by themselves guarantee that scale-up or altered substrate combinations will preserve the same selectivity [1].

Why Sadphos, and how this compares with the precursor Heck/Tsuji–Trost chemistry

The ligand screen is the clearest point of comparison with earlier work. Common C2-symmetric bisphosphines, Trost ligands, bisoxazoline ligands and monodentate phosphoramidites gave poor yields and negligible enantiocontrol in this spiroannulation, whereas the non-C2-symmetric Sadphos family proved superior [1]. A new PC4 ligand with an electron-donating group on the xanthene backbone raised enantioselectivity from 92% to 96% ee, and the standard conditions ultimately delivered 77% yield with 96% ee for 3aa [1]. This mirrors the broader Sadphos design logic: these sulfinamide phosphines coordinate to Pd(0) and Pd(II) in multiple modes and have been tuned across Ming-Phos, Xu-Phos, Xiang-Phos, TY-Phos, PC-Phos and related families to solve different selectivity problems in palladium catalysis [5].

The closest precursor is the 2025 asymmetric domino Heck/Tsuji–Trost reaction of flexible vinylic halides with 1,3-dienes, which used Xu-Phos and Pd pivalate with Ag2SO4 in DMAc to make sp3-rich cyclic isoprenoids [3]. That work established that Heck insertion is the stereodetermining step and that Sadphos ligands are critical for reactivity and enantioselectivity, but its products are formed through a formal (4+2) Heck/etherification and do not involve a racemic atropisomeric biaryl or a DyKAT manifold [3]. The new paper changes the substrate logic: instead of a flexible vinylic halide, it uses a racemic phenolic biaryl whose axial chirality is transiently induced and then transferred, which is what enables three contiguous stereocenters rather than the simpler stereochemical outcome of the precursor [1,3].

Competing approaches and what remains open

Other asymmetric palladium methods show that Sadphos ligands are not limited to this cascade. Photoinduced palladium catalysis has been used for dynamic kinetic asymmetric allylic imidation, where minor changes in base and conditions switch between chiral 1,2- and 1,4-amino alcohols [2]. Sadphos has also enabled enantioselective Suzuki–Miyaura coupling to construct inherently chiral pillar[5]arenes, with PC-Phos and a bulkier TY-Phos analogue giving up to 96% ee [6]. Axially chiral alkylidenecyclobutanes have been made by a palladium/Sadphos carbene coupling that proceeds through enantiodetermined migratory insertion followed by central-to-axial chirality transfer in β-hydride elimination [7]. These examples confirm the general utility of Sadphos in asymmetric palladium catalysis but use different chirality-transfer directions and different substrate classes, so they do not directly test the new spiroannulation mechanism [2,7,8].

The unresolved questions are mechanistic and practical. The DyKAT interpretation rests on the measured atropisomerization barrier of 32.5 kcal/mol and on selectivity data, not on direct observation of the (Sa)-Pd and (Ra)-Pd intermediates [1]. The acyclic diene result, at 32% ee, shows that the current ligand does not enforce high selectivity across all diene classes [1]. And because the method is a cascade dearomative Heck/Tsuji–Trost reaction, it inherits the classic challenges of suppressing 1,4-palladium migration and β-hydride elimination; the paper argues that the Sadphos ligand is the key to overcoming them, but the evidence for that claim is the ligand screen and the product distribution rather than a direct kinetic partitioning of the off-cycle pathways [1]. Earlier work on palladium-catalyzed alkene difunctionalization also shows that dearomative and heterocycle-forming cascades can be sensitive to substrate and ligand bulk, with some cyclizations giving only 2:1 diastereomer ratios, which is a useful reminder that high selectivity in one cascade does not automatically generalize to related dearomative systems [4].

About These Sources

This research page is built on 7 peer-reviewed studies — published from 2018 to 2026, 6 from 2024 or later — selected as the most relevant from 12 studies that passed quality screening, drawn from 43 papers retrieved from a database of over 500 million.

Sources used in this answer