Full Paper
idyl catalysts were found to be effective for the cyclisation of
nonterminal propargylic amides, but were ineffective for reac-
tions with substrates containing electron-withdrawing amide
substituents (R1). With this in mind, the cyclisation of a number
of propargylic amides was examined using 5 mol% of complex
4d, and the results were compared to that previously achieved
using 10 mol% [Ag(MeO-Py)2]PF6 under identical conditions.
Initial tests were performed with substrates 7a–g with differ-
ent amide substituents (R1). In all but one case (7d to 8d),
higher and faster turnovers were achieved using the NHC com-
plex. Particularly gratifying are the substantial improvements
achieved with substrates 7e–g, in which the Ag–pyridyl
system was ineffectual. The cyclisation of substrate 7g to 8g is
particularly notable. This reaction was only reported once
before, using 5 mol% of [Au(PPh3)]OTs to proceed to a 87%
conversion in 48 h.[49] In comparison, a comparable 92% yield
can be obtained using catalyst 4d in just 18 h under essential-
ly the same reaction conditions. On the other hand, double
cyclisation of di-propargylic amide substrates 7h and 7i pro-
ceeded in 65% and essentially quantitative conversions to 8h
and 8i, respectively. The lower productivity of the former was
attributed to catalyst inhibition by chelation of the bis-oxazoli-
dine 8h to the silver catalyst. Conversely, the cyclisation of
substrate 7j to 8j was achieved in 20 min at 1 mol% catalyst
loading, facilitated by the presence of gem-dimethyl groups at
the propargylic position (Thorpe–Ingold effect). This also al-
lowed the cyclisation of the nonterminal alkyne substrate 7k
to be achieved in a good yield. In contrast, the silver(I)–pyridyl
catalyst proved to be superior for the cyclisation of 7l and
7m. These results demonstrate that the silver–NHC catalyst is
rather sensitive to bulky substituents at the propargylic posi-
tion. It is conceivable that the catalytic activity may be reestab-
lished by exploring the synergistic effects between the neutral
and anionic ligands, but this is beyond the scope of this study.
systems, particularly towards electron-deficient amide substitu-
ents. On the other hand, [Ag(4-MeO-Py)2][PF6] is a better cata-
lyst for substrates containing substituents at the propargylic
and terminal substituents; the two silver catalytic systems are
highly complementary in their reaction scope. The most signifi-
cant result of this study is the observation of synergistic effects
between the NHC and carboxylate ligands. This is interesting,
as it can potentially offer independently tunable sites to opti-
mize the catalyst stability and reactivity towards a certain sub-
strate. It is conceivable that a better catalytic performance may
be achieved if saturated series of ITent ligands, for example
SITent, are available. The synthesis of these new [Ag(NHC)(car-
boxylate)] complexes, and their applications in other catalytic
reactions, will be explored in future work.
Experimental Section
General procedure for the synthesis of [(Ag(NHC)(O2CR)] com-
plexes: A mixture of NHC·HBF4 (0.2 mmol, 1 equiv) and Ag(O2CR)
(0.24 mmol, 1.2 equiv) in CH2Cl2 (5 or 10 mL) was stirred for 15 min
and K2CO3 (4.0 mmol, 20 equiv) was added. After a prescribed
period of time (x h, see the Supporting Information), the reaction
mixture was filtered through Celite and the solvent was removed
in vacuo. The residue was washed with 5 mL Et2O to obtain the
crude product as a white solid. Recrystallisation from CH2Cl2/
hexane at 08C or ꢀ208C afforded the desired product as colourless
crystals.
Single-crystal X-ray diffraction data for all 15 complexes have been
deposited with the Cambridge Cystallographic Data Centre
(CCDC 1449329–1449343).[36a]
Computational details: All calculations were performed using the
density functional theory (DFT) with the gradient corrections for
exchange and correlation proposed by Becke[51] and Perdew[52]
(BP86) implemented in the ADF package,[53–55] in combination with
a fine integration parameter (with a numerical integration parame-
ter set to 5). A triple-x basis set with two polarization functions on
all atoms (TZ2P) was used. Electrons of the core shells have been
treated within the frozen core approximation.[56] Scalar relativistic
corrections and spin orbit were included with the zeroth-order reg-
ular approximation (ZORA).[57] All geometries were optimized with-
out any symmetry constraint. These geometries were used to
obtain the carbene NMR chemical shielding properties (including
spin orbit corrections) in the studied NHC–AgI complexes. We in-
cluded the spin orbit term in the NMR calculations to understand
the spin-orbit coupling effects and to estimate the number of elec-
tron transitions from the filled NHC–Ag (s) MO to the empty NHC–
Ag (p*) MO and their contribution to the overall calculated para-
magnetic shielding.
Conclusion
A series of fifteen new [Ag(NHC)(carboxylate)] complexes has
been prepared, containing saturated and unsaturated NHC (in-
cluding ITent), as well as acetate and benzoate ligands. The
steric and electronic properties of these complexes have been
comprehensively and systematically explored by spectroscopic,
crystallographic, and DFT methods. These studies suggest that
the coordination chemistry and stability of these complexes
are dominated largely by steric, rather than electronic effects
imposed by the NHC ligand. The steric environment surround-
ing the Ag metal center is responsive to ligand saturation,
wingtip substituents, and, in the case of IPent, the nature of
the carboxylate ligand.
General procedure for the AgI–NHC catalyzed conversion of 7 to
8: The substrate (0.2 mmol) and the 1,3,5-trimethoxybenzene inter-
nal standard (ca. 6.8 mg) were dissolved in CD2Cl2 (0.5 mL) at room
temperature, followed by the catalyst addition (0.01 or 0.05 equiv).
The solution was transferred to an NMR tube, and the reaction
The catalytic activity of the NHC–Ag–carboxylate complexes
was assessed in the cyclisation of propargylic amides to give
oxazolines. In contrast to Au–NHC complexes, a ligand acceler-
ation effect is clearly observed using these Ag catalysts. Com-
parisons were also made with previously reported pyridyl–Ag
complexes. In all cases, the reaction proceeded under mild
conditions to afford 5-exo-dig oxazoline products. The use of
[Ag(IPent)(OBz-Cl)] 4d overcomes limitations of earlier catalytic
1
progress was monitored at ambient temperature by H NMR spec-
troscopy. The yields and conversions were calculated by compari-
son of the integrals of the product and substrate resonances to
that of the internal standard.
Chem. Eur. J. 2016, 22, 1 – 9
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ