COMMUNICATIONS
completely shut down the production of 5, with only Gratifyingly, the expected deuterated product 11
the normal PKR product 4 observed (Entry 13); Mean- (Entry 1) was obtained with the addition of CD3OD in
while, H2 atmosphere also suppressed the production the reaction mixture. Then two additional deuterium
of 5; Reactions under either pure H2 atmosphere or H2/ labelling conditions (Entries 2–3) were devised to
CO atmosphere only produced the normal PKR determine whether the deuterium in the product 11 was
product 4 (Table 1, Entries 14–16).
stemmed from the methyl group or the hydroxyl group
In the search of the best additives for the reductive of the methanol. The results clearly indicated that
cleavage of the CÀ O bond, we found that water or hydrogen originated from the alkanol proton of the
alcoholic solvents (Entries 17–20) were able to methanol. Meanwhile, considering that the association
promote the CÀ O bond cleavage to generate 5. It is of the deuterium in 11 could either occur during the
interesting to note that both chemoselectivity and yield reaction or after it was quenched by proton sources, we
of 5 were significantly improved when isopropanol also added CH3OD after the disappearance of cobalt-
was used as an additive (Table 1, Entry 20). The yield alkyne complex 9 (Entry 4). No association of deute-
of the desired product 5 was further increased when rium in the product indicated that the reductive
more i-PrOH (MeCN: i-PrOH=5:1) was used in the cleavage of CÀ O bond took place during the reaction.
reaction (Entry 21). Finally, the yield of 5 reached 56%
Based on the widely recognized Pauson-Khand
when the reaction was performed at both a lower reaction mechanism[8] and the deuterium labelling
concentration (0.05 M) and a lower temperature experiment, a plausible reaction mechanism for the
°
(50 C).
CÀ O bond cleaved Pauson-Khand reaction was pro-
With the optimized reaction conditions in hand, we posed (Scheme 3).
further investigated the substrate scope of the CÀ O
Taking substrate 3 as an example, we assumed that
bond cleaved PKR (Table 2). It turned out that the cobalt-alkyne complex 9 went through a well-accepted
reaction was highly substrate-dependent. The substitu- mechanism to convert to intermediate B via A. Given
tion pattern and stereochemistry had a vital influence that Lewis base can accelerate the alcoholysis of
on the outcome of the reactions. Substrates 6a–c carbonyl cobalt,[9] we speculated that MeCN might
afforded the CÀ O bond cleaved PKR products 7a–c in serve as a Lewis base to promote the alcoholysis of
38–58% yields; while substrate 6d and 6e only intermediate B, which breaks the CoÀ Co bond to form
produced the normal PKR product 7d and 7e. the cobalt-hydrogen species C. Subsequently, reductive
Substrate 6f, the diastereoisomer of our primary elimination of carbonyl cobalt complex in C could
substrate 3, failed to undergo PKR and yielded a furnish an intermediate D (Path a). Alternatively, the
depropargylated product epi-8 instead. This result intermediate B could go through the same path in a
illustrated the crucial contribution of the substrate regular Pauson-Khand reaction to reach intermediate
configuration to the reaction. Besides syntheses of the C’. C’ was then converted to the intermediate D
above 5,6-carbocycles, 5,7-carbocycle 7g could also through alcoholysis (Path b). We speculated that
be generated from substrate 6g in 37% yield using this electrons on Co in the intermediate D could fill in the
chemistry. Although the yield was not ideal, this antibonding orbital of CÀ O bond to promote the
reaction holds a great advantage for preparing such a heterolysis of CÀ O bond. Finally, the resulting inter-
synthetic challenging compound from a simple sub- mediate E produces the CÀ O bond cleaved PKR
strate 6g. To our further delight, the CÀ O bond cleaved
PKR was generally successful with cyclopentene
substrates. Substrates 6h–m, which contained various
substitution patterns, all produced the corresponding
5,5-carbocycles in modest to good yields. Notably, the
reaction of 6h, which was carried out on a 5 g scale,
achieved 73% yield. Again, the influence of stereo-
chemistry of the substrate was observed. Although
very similar to 6h structurally, substrate 6n only
afforded depropargylated product 10 under our opti-
mized reaction conditions.
To acquire more insights into the mechanism of this
interesting reaction, we first tried to clarify the hydro-
gen source in this reductive process. Given the best
performance of alcoholic additives in our optimization
experiment, they were speculated to be the hydrogen
source. To verify this assumption, we conducted a
deuterium labelling experiment using various commer-
cially available d-methanol as an additive (Table 3).
Scheme 3. A Proposed Reaction Mechanism.
Adv. Synth. Catal. 2021, 363, 1887–1891
1889
© 2021 Wiley-VCH GmbH