At the outset, the feasibility of the intraannular RCM
process was examined. The reaction of 1,2-diallylferrocene
(1a) proceeded smoothly in the presence of the Grubbs II
After the optimization studies, Mo/(R)-L1 or Mo/(R)-L3
was applied to the other substrates. Substrate rac-1c, which
has a ruthenocene core in place of the ferrocene in 1b, was
also resolved as described above, however the enantioselec-
tivity was somewhat lower (krel =11; Table 1, entry 4). The
position of R3 in 1 is crucial for a high enantioselectivity of
the ARCM reaction. Substrate 1d, which possesses a p-tolyl
group adjacent to the methallyl substituent, was resolved ef-
ficiently by using Mo/(R)-L1, and 1d with 91% ee was re-
covered in 30% yield (Table 1, entry 5). On the other hand,
the reaction of 1e, which has a p-tolyl substituent next to
the allyl group, under the identical conditions gave recov-
ered 1e with 4% ee and cyclized 2e with less than 1% ee
(Table 1, entry 6). Based on our previous studies,[9] it is
postulated that the initial metathesis between the Mo*–al-
catalyst[11] A (1 mol%) or the Schrock Mo catalyst[12]
B
(1 mol%) to give (h5-4,7-dihydroindenyl)FeCp (2a, Cp=cy-
clopentadienyl) in >99% yield [Eq. (2)].
Substrates 1b–1i are planar chiral because the two adja-
cent allylic substituents are non-identical, and their race-
mates were subjected to the kinetic resolution studies by
using a chiral Mo–alkylidene catalyst (Table 1).[13] Screening
of the chiral catalysts was examined on racemic 1-allyl-2-
methallylferrocene (1b) as a prototypical substrate. The re-
actions were carried out in benzene at 238C in the presence
of an appropriate chiral Mo catalyst (5 mol%), that was
generated in situ from the Mo precursor (pyrrolyl)2Mo(=
CHCMe2Ph)(=N-C6H3-2,6-iPr2) and an axially chiral biphe-
nol derivative.[14] Under these conditions, the Mo catalyst
generated with (R)-L1[15a] gave the RCM product 2b with
53% ee in 66% yield, and the unreacted 1b with 98% ee
was recovered in 24% yield (Table 1, entry 1). The krel value
([reaction rate of the fast-reacting enantiomer]/[reaction
rate of the slow-reacting enantiomer]; selectivity factor) for
this reaction is estimated to be 14.[16] The Mo/(R)-L2 spe-
cies[15b] was highly active; the reaction of rac-1b by using
this catalyst showed complete conversion to rac-2b within
4 h. Thus, the kinetic resolution study by using Mo/(R)-L2
was conducted at 08C in toluene for 1 h, but the enantiose-
lectivity was minimal (krel =2.0; Table 1, entry 2). The Mo
catalyst coordinated with (R)-L3[15c] showed excellent enan-
tioselectivity giving 1b with 96% ee in 37% yield and 2b
with 72% ee in 60% yield, respectively. The krel value for
the reaction is as high as 22 (Table 1, entry 3).
A
ACTHUNGERTNNUNGiAHCTUNGTRNENdUNG ene precatalyst and a substrate takes place at the un-
AHCTUNGTRENNUNG
Mo*–alkylidene, the steric factors between the two substitu-
ents adjacent to the [(R)-Mo*]-binding tether are fairly dif-
À
À
ferent (C H vs. C methallyl). As a result, [(R)-Mo*]/(R)-1d
forms preferentially over the other diastereomeric inter-
mediate (Scheme 2, left). On the other hand, the two sub-
stituents proximal to the allyl group are sterically compara-
ble in rac-1e. And thus, the (R)-Mo* species failed to distin-
guish the two enantiomers in 1e leading to the low enantio-
selectivity in its kinetic resolution (Scheme 2, right). Ferro-
cenylphosphine sulfide (1 f), in which the relative
arrangement of the three substituents is similar to that in
1d, was also a suitable substrate for the kinetic resolution,
and unreacted 1 f was recovered with 99% ee and 32%
yield with a krel value of 25 (Table 1, entry 7).
The enantioselectivity of the
kinetic resolution was sensitive
to the substituent at the allylic
2-position (R2 group in 1). In-
troduction of a bulkier substitu-
ent in place of the methyl
group in 1b improved the enan-
tioselectivity. The ARCM of
rac-1g, which has a 2-ethylallyl
tether, proceeded with excellent
selectivity (krel =30) to give the
Table 1. Molybdenum-catalyzed intraannular ARCM kinetic resolution of planar-chiral 1,2-diallylmetallo-
ACHTUNGTRENNUNG
cenes.[a]
[e]
1
Chiral L
T [8C],
t [h]
ee (yield) of
ee (yield) of
Config./optical
krel
recovered 1 [%][b,c,d]
cyclized 2 [%][b,c]
rotation of 2
recovered
substrate
with
99% ee in 37% yield (Table 1,
entry 8). With a benzyl group as
the R2 substituent in rac-1h, the
krel value was further improved
to 33 (Table 1, entry 9). Sub-
strate 1i, possessing a 2-(tri-
1
rac-1b
rac-1b
rac-1b
rac-1c
rac-1d
rac-1e
rac-1 f
rac-1g
rac-1h
rac-1i
(R)-L1
(R)-L2
(R)-L3
(R)-L3
(R)-L1
(R)-L1
(R)-L3
(R)-L3
(R)-L3
(R)-L2
23, 24
0, 1
98 (24)
5 (80)
96 (37)
96 (31)
91 (30)
4 (26)
99 (32)
99 (37)
96 (41)
97 (35)
53 (66)
31 (15)
72 (60)
52 (60)
31 (66)
<1 (74)
67 (67)
74 (59)
79 (53)
87 (57)
(+)
(+)
(+)
14
2.0
22
11
5.3
<1.1
25
30
2[f]
3
23, 12
23, 12
23, 0.5
23, 0.5
23, 12
23, 12
23, 12
23, 3
4
5
6
7
8
9
10
(+)
(À)
(À)
(R)-(+)
(+)
ACHUTNGRENmNUG ethACHTUNGTRENNyUGN lsilyl)allyl tether, was less
(+)
(À)
33
60
reactive probably due to the ex-
cessively bulky SiMe3 substitu-
ent. Both Mo/(R)-L1 and Mo/
(R)-L3 failed to afford the cy-
clized product for the reaction
with 1i. It was found that the
[a] The reaction was carried out in benzene in the presence of a metathesis catalyst (5 mol%) generated in
situ unless otherwise noted. [b] The enantiomeric excess was determined by chiral HPLC (see the Supporting
Information for details). [c] The yield of the product isolated by silica gel chromatography is given in paren-
theses. [d] The enantiomeric excess of recovered 1 was determined after conversion into 2 by a reaction with
the Grubbs II catalyst. [e] Calculated based on a first-order equation (Ref. [16]). [f] In toluene.
4152
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 4151 – 4154