Z. Császár, et al.
CatalysisCommunications146(2020)106128
reaction of the cyclic sulfate with the amine and the second substitution
reaction take place with complete inversion at the stereogenic centers. The
31P{1H} NMR spectra of the compounds L1-L6 exhibit one singlet line
indicating the formation of one single diastereomer. It is important to note
that the present synthetic methodology is of high modularity and does not
require tedious multiple reactions steps. Concerning the structural versa-
tility of the commercially available chiral (or non-chiral) cyclic sulfates
and diamines the structural fine tuning of the corresponding ligands can
easily be implemented without tedious workup procedures.
2.2. Catalytic studies
At first, we chose (E)-chalcone (S1) as a standard substrate for the
Ir-catalyzed hydrogenation to screen ligands L1-L6 (Scheme 2). Hy-
drogenation of S1 in methanol in the presence of the Ir catalyst, syn-
thesized in situ from [Ir(COD)Cl]2 and chiral P,N,N ligand (P,N,N/
Ir = 1), at a substrate catalyst molar ratio of 1000, 5 bar hydrogen
pressure and room temperature afforded hydrogenation products A1,
B1 and C1 in different ratio.
Fig. 1. Subclasses of chiral P,N,N ligands used in Ir-catalyzed asymmetric hy-
drogenation.
Generally, less common are instances of the manipulation of the
chelate ring size, despite the fact that such changes can be, in many
cases, readily implemented resulting in steric and also electronic al-
teration and producing similarly dramatic improvements in catalytic
activity and enantioselectivity [25–27]. Surprisingly, the variation of
the ring size of potentially tridentate ligands in catalysis is very rare
[28–31] and to the best of our knowledge there is no such example
concerning asymmetric catalytic transformations. However, as it was
underlined by Crabtree and Peris, “this little studied area, analogous to
bite angle effects in chelates, seems worth further efforts” [32].
In the present study we report on the development of a highly
modular synthetic approach leading to a novel class of chiral P,N,N
ligands based on two alkane-diyl backbones of different P,N and N,N
tether lengths (Fig. 1). In order to compare the catalytic behavior of the
new ligands, they were tested in the iridium catalyzed chemo- and
enantioselective hydrogenation of enones, a challenging substrate class,
with the intention to compare their activity, chemo- and enantioselec-
tivity. Our primary aim was to vary the chelate ring size formed by the
ligands and hence influence the bite angle and the conformational
flexibility of the catalysts. Additionally, the effect of the reaction con-
ditions and the substrate scope was carefully screened.
The reactions completed within 60 min with ligands L2 and L4
(Table 1, entries 2 and 4) and 87% conversion could be achieved with
L3 (entry 3). Catalysts with more rigid skeleton, i.e. ligands of shorter
backbone(s) (L1) or pyridyl containing side chain (L5 and L6), provided
low turnovers. For the sake of comparison, we tested the bidentate
analogue of ligand L3, with a simple N-ethyl substituent, under iden-
tical conditions in the hydrogenation reaction of S1 (entry 7). No
conversion could be achieved, indicating that the presence of the third
coordination site is necessary to obtain catalytic turnover. Furthermore,
it also suggests that the N atom of the side arm coordinates to the metal
during the catalytic reaction forming an N,N chelate in addition to the
P,N cycle. Zhou et al. reported that the introduction of a third co-
ordination site to a P,N/Ir system (SpiroAP/Ir) lead to increased sta-
bility and activity (SpiroPAP/Ir) [8]. In this case the pronounced dif-
ference between the P,N and P,N,N systems was attributed to the ability
of the former system to irreversibly form inactive dimers or trimers
under hydrogenation conditions [33].
As the bidentate system was totally inactive in our case (entry 7), it
was surmised that the introduction of the third coordination site not
only prevents the formation of inactive metal species but through co-
ordination to the metal may also change the conformation of the P,N
chelate, thus potentially changing the substrate-catalyst interaction in
the catalytic cycle. This is also in line with the data of Table 1, where
the N,N backbone. The highest conversions could be achieved with li-
gands L2 and L4 (Table 1, entries 2 and 4) having the most flexible
2. Results and discussion
2.1. Synthesis of the ligands
The novel ligands were synthesized in two simple steps. At first, en-
antiomerically pure (4S,5S)-4,5-dimethyl-1,3,2-dioxathiolane 2,2-dioxide
(1a) or (4R,6R)- or (4S,6S)-4,6-dimethyl-1,3,2-dioxathiane-2,2-dioxide
(1b) and the corresponding diamine were mixed in THF leading to ami-
noalkyl sulfates 2a-f (Scheme 1). A remarkable feature of this metho-
avoided. The addition of three equivalents of LiPPh2 in THF provided the
desired P,N,N (L1-L6) ligands in good yields. Both the ring opening
The chemoselectivity of the reaction drastically changes with the
length of the P,N backbone. Butane-2,3-diyl based systems L1 and L2 are
less selective toward allylic alcohol A1 (entries 1 and 2). In fact, L2
provided saturated ketone B1 with 94% chemoselectivity. Contrarily,
ligands with pentane-2,4-diyl backbone and sp3 N-atoms, L3 and L4,
proved to be rather selective to A1, giving 66 and 80% chemoselectivity,
respectively (entries 3 and 4).
Scheme 1. Synthesis of chiral ligands L1-L6
2