Angewandte
Research Articles
Chemie
kenes) were obtained predominantly (Scheme 1c, left) via the
p-coordination.
As is well-known, rationally manipulating reaction selec-
tivity for regio-divergent synthesis by utilizing and/or over-
coming the intrinsic electronic and steric bias is a perpetual
subject for chemists.[11] In view of the privileged significance
of a-fluoroalkene motifs in drugs, agrochemicals and materi-
als, we wondered if we could overcome the innate reactivity to
gain high level of branch-enriched fluorinated allylic skele-
tons also from the same gem-difluorocyclopropane precursors
via the s-coordination.
Inspired by Echavarren[12] and Morkenꢀs elegant works[13]
on the regio-controlled allyl-allyl cross-couplings via 3,3’-
reductive elimination pathway (Scheme 1b), and combined
with our expertise on hydrazones acting as alkyl carboanion
equivalents in the nucleophilic additions[14] and cross-coupling
reactions,[15] we hypothesized that gem-difluorocyclopropane
could be used as allyl electrophile with simple hydrazone as
diazaallyl partner.[16] The readily formed putative allyl,
diazaallyl-palladium intermediate would allow for the anal-
ogous inner-sphere 3,3’-reductive elimination, thus ensuring
the nucleophiles to attack at the sterically more hindered C3
atom (Scheme 1c, right). Besides, based on the literature[17]
and our previous work on Pd-catalyzed hydroalkylation of
methylenecyclopropanes with hydrazones as pronucleo-
philes,[15e] we also reasoned that a subsequent denitrogenation
activity might take place, which in turn would accelerate the
3,3’-reductive elimination process cooperatively. Assuming
that our proposal is viable, there comes with an affiliated
benefit that alkyl groups could be incorporated into the
monofluoroalkene scaffolds, which is potentially complemen-
tary to the Au-catalyzed hydrofluorination reaction of
terminal alkyne.[4a] With this design in mind, herein we wish
to report that a switched regioselectivity of Pd-catalyzed
defluorination alkylation of gem-difluorocyclopropanes was
enabled by simple hydrazones, and the branch-enriched
monofluorinated alkene products were obtained selectively
(Scheme 1d).
Scheme 2. Possible intermediates encountered in the allyl-(diaza)allyl
cross-couplings.
cross-coupling, the ambident nucleophilic sites (C-terminal
and N-terminal)[20] in hydrazone part may result in additional
bis(h1-allyl) bonding modes (V, VI and VII), which are
detrimental to the designed regioselectivity. Last but not the
least, different from allyl-allyl cross-coupling, an appended
denitrogenative event is expected to complete the ring
opening-alkylation process.
With these concerns in mind, a palladium/bidentate
phosphine system was initially adopted to start our study
with the model reaction of 2-(2,2-difluorocyclopropyl)naph-
thalene (1a) and phenyl hydrazone (2a) (Table 1). However,
no expected product 3a was obtained with small bite-angle
bidentate phosphine ligand, which indicated that the allyl-
(diaza)allyl cross-coupling might be quite different from the
previous allyl-allyl one. We then reanalyzed the possible
intermediates (V–VIII) in Scheme 2 and envisioned that in
order to make the least hindered bis(h1-allyl) species VIII
favorable, the steric-embedded coordination ligand might be
beneficial due to its repulsion to the substituents (R or R1).
Therefore, a series of bulky ligands, such as PCy3, PtBu3,
RuPhos, Xphos and NHCs (N-heterocyclic carbenes) were
examined accordingly (entries 1–8). We were pleased to find
that the anticipated branched fluoroalkene product 3a could
be selectively obtained in 97% yield when the carbene ligand
precursor SIPr·HCl [1,3-bis(2,6-diisopropylphenyl)imidazoli-
um chloride] was used in presence of KOH in THFat 458C for
12 h (entry 8). In order to simplify the experimental proce-
dure, the air-stable and commercially-available Pd-PEPPSI-
SIPr catalyst was employed and 3a was afforded in compa-
rable yield and regioselectivity (entry 9). Base was utilized to
mediate the hydrazone deprotonation, and its choice was also
crucial to facilitate the final denitrogenation. No desired
product was observed when strong bases such as KOtBu and
LiOtBu were tested, owing to the dominant Wolff–Kishner
reduction of hydrazone (entries 10 and 11). Organic base
DBU (1,8-diazabicyclo[5.4.0]undec-7-ene) did not work for
the defluorination alkylation reaction (entry 12). Relatively
Results and Discussion
To realize this hypothesis, several challenges must be
taken into consideration (Scheme 2). Firstly, to make the
reaction operate via a 3,3’-reductive elimination, the coordi-
nation ligand must play a key role, which should render both
(aza)allyl groups to adopt h1-bonding mode (II or III) other
than the more common h3-bonding mode (I). Secondly, two
kinds of bis(h1-allyl) bonding modes are available when the
substitution (R) is present, and the sterically less hindered
bis(h1-allyl) species (III) is supposed be predominant to favor
3,3’-reductive elimination (the direct elimination from inter-
mediate II is less likely[18]) relative to the 1,1’-path. As
revealed by Morkenꢀs work,[19] small-bite-angle bidentate
phosphine ligands could effectively increase both the C1-C1’
separation and promote the formation of desired bis(h1-allyl)
bonding mode. Whether it also works or not in our system, the
general principle to choose appropriate ligands is well
applicable. Thirdly, considering the analogous allyl-diazaallyl
Angew. Chem. Int. Ed. 2021, 60, 2 – 9
ꢀ 2021 Wiley-VCH GmbH
&&&&
These are not the final page numbers!