Journal of the American Chemical Society
Article
form an FLP (or weak adduct) with both the Lewis acid
borane and the (di)aryl methylene cation (I3) is critical, as
other strong, less hindered Lewis acids such as BF3 do not
work in the reaction. Likewise, smaller phosphines or amines
tend to coordinate more strongly to the carbocation (I3). In
addition to being able to form an FLP, the base also functions
to trap reactive I8. Indeed, one could possibly conceive that
the reaction could proceed with the Lewis acid only, whereby
compared to the TS5cationic pathway energy barrier (range 8.1−
25.5 kcal/mol) when changing the substituent on the
phenylacetylene (Table 4, entries 6−9). As with ester 1a,
both esters disclosed a larger energy difference between the
two pathways when reacted with acetylenic compounds
bearing an electron-donating group (p-OMe). Likewise, a
much smaller energy difference was noted for both esters when
reacted with acetylenic compounds bearing an electron-
withdrawing group (p-CF3). Interestingly, DFT studies showed
that for the case of the reaction of ester 1j with 1-ethynyl-4-
(trifluoromethyl), the radical pathway is slightly energetically
more favorable compared with the cationic pathway (Table 4,
entry 8). These results suggest that for electron-withdrawing
arylacetylenes both paramagnetic and diamagnetic mechanisms
are potentially possible, whereas for electron-rich arylacety-
lenes a purely diamagnetic pathway is operative.
The key difference in the two mechanisms involves the
reaction of either a cationic intermediate or a radical
intermediate with the arylacetylene, generating a new cationic
or radical species. Whether this new intermediate is a cationic
or a neutral radical species can be probed using a Hammett
plot (cf. computational studies) based on substituted
arylacetylenes. To gain this insight into the reaction pathway
and substituent effects also from experimental evidence, we
examined competition reactions among the FLP, aryl ester 1a,
and arylacetylenes p-XC6H4CCH bearing electron-with-
drawing, electron-neutral, and electron-releasing groups. The
Hammett plot requires relative rate constants for the reaction
of different substituted alkynes that were obtained using a
series of competition experiments. Initial competition experi-
ments in the presence of 1.5 equiv of five arylacetylenes p-
XC6H4CCH (X = CF3, F, Cl, H, and OMe) were
unsuccessful. The excess arylacetylene present in the reaction
mixture destroyed the efficacy of the FLP system, producing a
complicated reaction mixture which was not suitable for in situ
NMR analysis. We therefore carried out three binary
competition experiments with two alkynes being present at
1.5 equiv each (Table 5).
−
R1CO2 accepts the proton in the last step (I8 → I10 →
product, Figure 2). However, DFT calculations showed that
this pathway was less favorable thermodynamically, and
experimentally this reaction showed only a 22% yield with
many side products formed.
We were curious to know whether both the paramagnetic
and diamagnetic pathways are operative in parallel or if a
diamagnetic mechanism is purely responsible for the product
formation for all alkynes and the radicals observed from EPR
studies were simply off-pathway intermediates. To establish
this, we undertook further DFT calculations to investigate the
effect of electron-withdrawing, electron-donating, and electron-
neutral substituents on the ester (1) and the alkyne on the
activation barrier for the reaction.
The energy barriers TS4radical pathway and TS5cationic pathway
varied the electronic properties of the alkyne using p-
XC6H4CCH (X = NO2, CF3, H, OMe, NMe2) with ester
1a. As evidence from the DFT calculations, changing the
substitution at the para position of the arylacetylene did not
make significant difference for TS4radical pathway in their
respective energy barrier (23.9−27.0 kcal/mol) (Table 4,
entries 1−5). However, the energy barrier for TS5cationic pathway
changed dramatically (8.2−23.2 kcal/mol). When electron-
withdrawing groups (p-NO2 and p-CF3) on the arylacetylene
were employed, the differences between TS4
and
radical pathway
TS5
are 0.7 and 6.1 kcal/mol (Table 4, entries 1
cationic pathway
and 2). Electronically neutral phenylacetylene exhibits a
TS4radical pathway → TS5cationic pathway difference of 7.7 kcal/mol
(Table 3, entry 3). Electron-donating groups such as methoxy
(TS4radical pathway − TS5cationic pathway = 12.1 kcal/mol) and N,N-
dimethylamine (TS4radical pathway − TS5cationic pathway = 18.8 kcal/
mol), on the other hand, showed a significant energy difference
(Table 4, entries 4 and 5). These observations can be seen in
Figure 3, which shows that for TS4 there is a negligible change
in the energy barrier when changing the electronic properties
of the acetylenic substrate. This is in agreement with little
charge formation on the reaction center in the radical
mechanism. Conversely, for TS5, there is a strong positive
correlation between the TS5 energy barrier and the substituent
σp constant. This is expected for the cationic pathway because
a developing positive charge adjacent to the substituted phenyl
ring will be stabilized by electron-donating groups (e.g., p-
NMe2 or p-OMe) on the acetylene. As can be seen in Figure 3,
TS5 is lower than TS4 for all substituents explored, although
the difference becomes small for strongly electron-withdrawing
groups.
We subsequently computed the energy barrier for the two
transition states by varying the electronic properties of the aryl
ester using electron-donating (p-OMe, 1d) and electron-
withdrawing (p-CF3, 1j) esters with electron-deficient (1-
ethynyl-4-(trifluoromethyl)benzene) and electron-rich (1-
ethynyl-4-methoxybenzene) acetylenes (Table 4, entries 6−
9). For both esters, a smaller change in the TS4radical pathway
energy barrier was observed (range 22.2−27.8 kcal/mol)
Using the optimized reaction conditions, three parallel
reactions were carried out in which equimolar mixtures of (a)
1-ethynyl-4-methoxybenzene and phenylacetylene, (b) 1-
ethynyl-4-(trifluoromethyl)benzene and phenylacetylene, and
(c) 1-ethynyl-4-methoxybenzene and 1-ethynyl-4-(trifluoro-
methyl)benzene were reacted with 1 equiv of ester 1a. The
3
ratios of Csp −Csp cross-coupled products (2d/2c, 2b/2c, and
2d/2b) were determined from the crude reaction mixture
For entries 1 and 2 in Table 5, the relative integrals for the
products were used to calculate the remaining equivalents for
the alkynes after reaction. Using the approach developed by
Ingold and Shaw30 and proposed for one-pot Hammett plots
by Harper and co-workers,31 we obtained relative rate
constants kx/kx′ for the reaction of differently substituted
arylacetylenes with reaction intermediate I3 or its equilibrium
species. Entry 3 confirms that in the competition between 1a
and 1-ethynyl-4-(trifluoromethyl)benzene/1-ethynyl-4-
methoxybenzene, product 2b is undetectable, in agreement
with the >200-fold difference in rate constants deduced from
entries 1 and 2. The relative rate constants in Table 5 are
normalized with respect to the unsubstituted alkyne and can
therefore be used directly to construct a Hammett plot (using
Hammett substituent constants from ref 32). The Hammett
plot (Figure 4) shows a clearly negative slope of −6.6 1.7,
4461
J. Am. Chem. Soc. 2021, 143, 4451−4464