J. Segato et al.
Inorganica Chimica Acta 522 (2021) 120372
complexes RCBF4, RCBF4/H2O and RCOTf, where N-(prop-2-ynyl)benza-
mide is coordinated to gold and the anion is in the second coordination
sphere weakly interacting with the NH moiety of the substrate through a
hydrogen bond. The OTsꢀ replacement with the substrate is instead a
thermoneutral process, yielding a RCOTs (0.1 kcal/mol) through a
calculated energy barrier of 7.8 kcal/mol and a tri-coordinated transi-
tion state where both N-(prop-2-ynyl)benzamide and OTsꢀ are bound to
gold (see Fig. S3 in the SI). The optimized geometries of all species are
shown in Figs. S4–S7. In all the RC complexes the counterion Xꢀ is
placed above gold, weakly interacting with the metal center with at least
one basic atom and forming a hydrogen bond with substrate NH group.
Starting from these RC adducts, the oxygen atom of the substrate CO
group intramolecular nucleophilic attack to one carbon atom of the
activated triple bond can easily occur, with relativity low energy bar-
riers (12.2, 10.1, 13.9 and 12.4 kcal/mol for BFꢀ4 , RCBF4/H2O, OTfꢀ and
OTsꢀ , respectively) (Fig. 2).
As it can be seen, this step requires a higher energy barrier than that
of the nucleophilic attack step for all the three anions, in full agreement
with the experimental observations. The geometries of TSII are
compared in Fig. 4, whereas PC structures are reported in Fig. S8 in the
SI.
The protodeauration energy barrier trend exactly matches both that
of the catalytic performances, and that of the anion basicity and
hydrogen-bond acceptor power (BFꢀ4 < OTfꢀ < OTsꢀ ), thus suggesting
an impact of the anion in controlling the reaction rate through
abstraction of the proton from the NH moiety and its releasing to the C2
atom bond to gold. Interestingly, inspection of the TSII structures in
Fig. 4 reveals that both the coordinating ability trend (Au… O = 3.523 Å
for OTsꢀ , Au… O = 3.611 Å for OTfꢀ , and Au… F = 3.914 Å for BFꢀ4 /
H2O) and the hydrogen-bond acceptor power (H… C2 = 3.079 Å for
OTsꢀ , H… C = 2.984 Å for OTfꢀ , and H… C = 2.288 Å for BF4ꢀ /H2O) are
able to account for the energy barrier trend. In addition, hydrogen
abstraction from substrate NH appears to be hindered by the coordi-
nating capability of the anion (H… N = 2.328 Å for OTsꢀ , H… N = 2.498
Å for OTfꢀ , and H… N = 4.444 Å for BF4ꢀ /H2O). Thus, the anion basicity
seems to be mainly responsible for the energy activation barrier value.
Overall, these results suggest that too basic anions, with higher
hydrogen-bond acceptor power (OTsꢀ ), do not easily abstract and
transfer the proton, thus slowing the reaction rate, consistently with
experimental findings, where the anion “proton shuttle ability” is found
to be crucial for the catalytic activity. Finally, a calculation has been
performed on the overall reaction path for NHC-Au-OTf, using the full
ligand (NHC = 1,3-bis(2,6-di-isopropylphenyl)-imidazol-2-ylidene), to
test the reliability of the results previously obtained with the model
NHC. The optimized structures of IC, RC, TSI, I, TSII and PC and their
relative energies are reported in the SI (Fig. S9). The results show that
the interactions of the anion with the bulky NHC ligand do not affect the
energetics of the whole path, particularly at the transition states, which
show comparable energies with respect to the corresponding IC (for
model NHC E (TSIOTf) = 8.2 kacl/mol, E(TSIIOTf) = 21.9 kcal/mol vs. full
NHC E(TSIOTf) = 6.3 kcal/mol, E(TSIIOTf) = 21.6 kcal/mol).
Formation of the vinyl gold complex (intermediate I) from this step is
an exergonic process for all the anions (Fig. 2). In the intermediate I, the
anion, still forming a hydrogen bond with the substrate NH group,
moves from the gold center, localizing in the second coordination
sphere. The acidic NH hydrogen should then be transferred to the C2
carbon atom coordinated to Au by the anion, acting as proton shuttle. To
calculate the overall path starting from NH proton detachment, due to
the very large number of different conformations involved, molecular
dynamics simulations would be needed, which are beyond the scope of
this work.
However, starting our analysis of the protodeauration step from in-
termediate I, a transition state for the final proton transfer from the
anion to C2 (TSII) has been located, which allows to completely ratio-
nalize the experimental findings.
Energy profiles for the protodeauration step are shown in Fig. 3.
The energy barrier calculated from I follows the trend BFꢀ4 /H2O <
OTfꢀ < OTsꢀ (21.8, 25.0 and 27.4 kcal/mol, respectively).
3. Conclusion
In this work, we extended our previous works studying the cyclo-
isomerization of N-propargylcarboxamides promoted by NHC-Au-X
complexes in green solvents from an experimental and theoretical
point of view. At first, we investigated the effect of the solvent
employed. As a primary and important goal, we found that the reaction
proceeds very well in most of the alternative solvents employed, such as
cyclohexanone, isopropyl acetate, MIBK, ethyl lactate, furfuryl alcohol,
γ-valerolactone and propylene carbonate. TOFs are comparable or even
better with respect to those obtained using VOS. On the other hand,
when the reaction is performed in DMSO or propionitrile the TOFs are
much lower, owing to the presence of a coordinating functional group
–
–
– –
C N, respectively) that upon coordination to the metal
–
(>S O and
–
centre affords inactive NHC-Au(solvent)OTf species. We also observed
that the use of propionic acid resulted in a speed up of the reaction,
according to the view that protodeauration is the RDS.
Secondarily, we found, according to our previous work, that the
activity of the catalyst in propionic acid (in terms of TOF) seems to be
inversely correlated to the coordinating ability and hydrogen-bond
acceptor power of Xꢀ (basic and coordinating strength: BFꢀ4 < OTfꢀ
<
OTsꢀ < TFAꢀ ), with TFAꢀ being by far the worst counterion. This
behaviour is consistent with the active role of the counterion in all the
three steps of the reaction pathway (pre-equilibrium, nucleophilic attack
and protodeauration).
DFT calculations corroborate these peculiar experimental kinetic
findings showing an excellent agreement, and the proposed mechanism
fully rationalizes the experimental reactivity, which is highly dependent
on both anion and solvent effects.
Fig. 3. Energy profiles for the protodeauration step of the cycloisomerization of
N-propargylcarboxamides reaction mechanism catalyzed by NHC-Au-X (Xꢀ
=
BFꢀ4 /H2O, OTfꢀ , and OTsꢀ ). Energy values (kcal/mol) refer to corresponding
ICX (Fig. 2) taken as zero.
These results are indeed remarkable and suggest that a sustainable
4