Struct Chem
(CD3)2CO): δ (ppm) 137.02, 130.01, 125.95, 124.3, 123.03,
121.72, 121.27, 111.34, 104.56, 77.03 (q, J = 36.3 Hz); 19F
NMR (376 MHz, (CD3)2CO): δ (ppm) − 81.05 (d, J = 5.2 Hz)
2,2,2-trifluoro-1-(indol-3-yl)ethanol (B): MS (EI) m/z: 118
states, or products. The reactant-solvent interaction is further de-
pendent upon the type of solvent used and its polarity index [39]
which can be categorized as non-polar, polar aprotic, or polar
protic. Polar solvents can interact with electron-rich substrates
and hence assist in specific bond cleavages [40] via solvation
effects. Polar protic solvents due to their ability to participate in
hydrogen bonding can help further activate and stabilize the in-
termediate. Non-polar solvents, however, only aid in solubiliza-
tion of reactants and hence are not able to show any predominant
solvation effects [28]. In the present work in addition to the
experimental investigations, density functional theory (DFT)
and NMR studies have been performed to better understand the
driving forces of the solvent effect. Based on our previous works
[8–13], the natural reactivity of indoles towards
trifluoroalkylation lies at the C3 position (thermodynamically
favored) under acidic conditions with microwave irradiation.
The synthesis of N-substituted indole (kinetic control) is tradi-
tionally favored using strong bases or with metal-mediated reac-
tions. However, as pointed out in our earlier work [8–13], the use
of cinchona alkaloids, considered as strong organic bases, still
resulted in the formation of the C3 product; thus, the base itself
cannot ensure the selective formation of the N1-product. In con-
trast, as shown in our recent studies [15, 30], using polar aprotic
solvents resulted in the exclusive formation of N-
hydroxyalkylated indoles. The goal of the current study is to
understand the fundamental basis of this exquisite solvent control
on the regioselective outcome of the reaction by using different
solvents ranging from non-polar to polar aprotic solvents in both
experimental and theoretical investigations.
1
(100%), 146 (92%), 215 (48%, M+), 197 (11%); H NMR
(400 MHz, CDCl3): δ (ppm) 8.3 (br s, 1H), 7.75 (d, J = 8.4 Hz,
1H), 7.17–7.45 (m, 4H), 5.38 (q, J = 6.8 Hz, 1H), 2.43 (br,
1H); 13C NMR (101 MHz, CDCl3): δ (ppm) 136.01, 126.2,
125.68, 123.65, 123.63, 122.93, 120.59, 119.32, 111.40,
67.56 (q, J = 33.3 Hz); 19F NMR (376 MHz, CDCl3): δ
(ppm) − 77.91 (d, J = 6.8 Hz)
Computational methods Geometry optimization of the reac-
tant, product, and transition state structures have been carried
out at B3LYP [31, 32]/6-31G(d,p) level of theory using
Gaussian 09 suite [33]. Frequency calculations of all opti-
mized structures were performed to confirm the minima and
transition state structure with one imaginary frequency. The
effect of solvent on overall reaction has been investigated
implicitly as well as explicitly. For that, the implicit integral
equation formalism polarizable continuum model (IEF-PCM)
of solvation has been used and those solvents that are utilized
in experiments were considered [34]. In addition, the temper-
ature factor (80 °C) was incorporated so as to mimic the ex-
periment. Intrinsic reaction coordinate (IRC) calculations
were performed for transition state structures to confirm that
it connects well to product and reactant on the potential energy
surface. Zero-point-corrected activation energy, free energy of
activation, and other thermodynamic parameters are estimated
from implicit calculations. The explicit role of solvent in the
reaction has further been studied, and for that, possible stable
complexes of indole with the solvent (DMSO, Et2O, acetone,
and benzene) and catalyst TEA were formed. Relative stabili-
zation energies and binding energies of the complexes were
The trifluoroacetaldehyde alkyl (ethyl/methyl) hemiacetals
and their hydrates are commercially available precursors for
the in situ generation of trifluoroacetaldehyde (fluoral) via
conventional or microwave heating [41] and consistent results
(conversion or enantioselectivity) are observed using any al-
kyl hemiacetals [8–13, 42].
1
estimated. H NMR chemical shifts were obtained for these
complexes in order to compare with experimental data [35].
Moreover, electrostatic potential (ESP) fit charges were calcu-
lated to identify the potential nucleophilic center on indole.
CYLview software was used to generate the images from
optimized geometries [36].
In order to observe the effect of solvents on the regioselective
outcome of the reaction, all other reaction parameters were first
optimized. The goal was to test the model reaction with sufficient
conversion and high regioselectivity. The conditions were stan-
dardized with dimethyl sulfoxide (DMSO) as solvent.
The initial screening was carried out at 120 °C to determine
the best molar ratio of the reactants and the amount of the base
catalyst (triethylamine [TEA]) used, which was proposed to
link to and partially deprotonate the slightly acidic N-H of
indole to increase the electron density of the heterocyclic ring,
thus increasing its reactivity. The results, tabulated in Table 1,
show three different indole to trifluoroacetaldehyde
ethylhemiacetal (TFAE) molar ratios (1:5, 1:7, 1:9) that were
studied respectively (Table 1, entries 1, 3, and 5). Although
use of nine equivalents of TFAE gave better percent conver-
sion (78% in 15 min, Table 1, entry 5), the ratio of the N-
substituted product was higher when 7 equivalent of TFAE
was used (69% in 15 min, Table 1, entry 3) with no by-product
Results and discussion
Experimental solvent studies
The effect of solvents on controlling the regioselectivity during
the substitution of indoles has previously been studied but only in
the presence of metal counter ions [21]. There are, however,
many examples in the literature on the exquisite role of solvents
in determining reaction rates [37] and selectivity [38]. The ability
of solvents to control the outcome of a reaction depends on their
capacity to coordinate with the reactants, intermediates, transition