A density functional study of chiral phosphoric acid-catalyzed direct arylation of trifluoromethyl ketone and diarylation of methyl ketone: Reaction mechanism and the important role of the CF3 group

Article abstract of DOI:10.1039/c3ob42157k

The detailed mechanism of the chiral phosphoric acid-catalyzed diarylation reaction between acetophenone and indole has been investigated by DFT methods and compared with that of the reaction between 2,2,2-trifluoroacetophenone and indole. The calculated results confirm our previous hypothesis that the CF ₃ group in the ketone plays a perfect double role in activating the substrate and stabilizing the single arylation product of tertiary alcohol. It is also demonstrated that the different ratio of the F-substitution in the CH₃ group of methyl ketone (CH_3-nF_n, n = 0, 1, 2, 3) affects the activation energy of the key dehydration step for the proposed diarylation process differently, and determines whether the subsequent re-arylation proceeds or is being suppressed. The computational prediction that the prohibitive barriers for CF₃ and CHF₂ ketones in the rate-determining dehydration step for the diarylation process could be overcome at higher reaction temperature has been validated by our additional experiments at 80 °C. Furthermore, the origin of the high enantioselectivity of the chiral phosphoric acid-catalyzed single arylation of trifluoromethyl ketone has been studied with the two-layer ONIOM method. The experimentally observed enantiomeric excess can be successfully rationalized.

Full text of DOI:10.1039/c3ob42157k

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
A density functional study of chiral phosphoric
acid-catalyzed direct arylation of trifluoromethyl
ketone and diarylation of methyl ketone: reaction
mechanism and the important role of the CF3 group†
Cite this: Org. Biomol. Chem., 2014,
12, 1908
Aiping Fu,*a Wei Meng,b Hongliang Li,a Jing Nieb and Jun-An Ma*b
The detailed mechanism of the chiral phosphoric acid-catalyzed diarylation reaction between acetophe-
none and indole has been investigated by DFT methods and compared with that of the reaction between
2,2,2-trifluoroacetophenone and indole. The calculated results confirm our previous hypothesis that the
CF3 group in the ketone plays a perfect double role in activating the substrate and stabilizing the single
arylation product of tertiary alcohol. It is also demonstrated that the different ratio of the F-substitution in
the CH3 group of methyl ketone (CH3−nFn, n = 0, 1, 2, 3) affects the activation energy of the key de-
hydration step for the proposed diarylation process differently, and determines whether the subsequent re-
arylation proceeds or is being suppressed. The computational prediction that the prohibitive barriers for
CF3 and CHF2 ketones in the rate-determining dehydration step for the diarylation process could be over-
come at higher reaction temperature has been validated by our additional experiments at 80 °C. Further-
more, the origin of the high enantioselectivity of the chiral phosphoric acid-catalyzed single arylation of
trifluoromethyl ketone has been studied with the two-layer ONIOM method. The experimentally observed
enantiomeric excess can be successfully rationalized.
Received 31st October 2013,
Accepted 21st January 2014
DOI: 10.1039/c3ob42157k
www.rsc.org/obc
great challenge owing to the easy diarylation of the ketones.1
Very recently, we have reported the first highly enantioselective
1. Introduction
The asymmetric Friedel–Crafts (FC) reaction represents one of chiral phosphoric acid-catalyzed direct arylation of trifluoro-
the most powerful methods for the synthesis of versatile build- methyl ketones.6 In these novel reactions, chiral phosphoric
ing blocks of biologically active aromatic compounds.1 acid 1 can catalyze the arylation reactions of indole 2 with the
Recently, chiral phosphoric acid derivatives have emerged as prochiral 2,2,2-trifluoroacetophenone 3a, affording the desired
powerful organocatalysts for the asymmetric Friedel–Crafts- α-trifluoromethyl tertiary alcohols 4a in high yields (99%) with
type reactions of indoles with various electrophilic partners, excellent enantioselectivities (92% ee) (eqn (1)). Moreover, dia-
such as imine,2 enamide,3 α,β-unsaturated carbonyls,4 and rylation was not observed under the given reaction conditions.
nitroolefins.5 In most of the reported reactions, the chiral
It is noteworthy that this direct arylation reaction can also
phosphoric acids have been found to exhibit high degrees of be extended to CHF2 and C2F5-ketones. As shown in eqn (2),
reactivity and enantioselectivity.1–5 Despite the impressive pro- the desired products were also achieved in high yields and
gress with chiral phosphoric acids, the direct catalytic asym- excellent enantioselectivities. In sharp contrast, under the same
metric arylation reactions between simple ketones and indole reaction conditions, when 2-fluoroacetophenone and non-
for the synthesis of valuable chiral tertiary alcohols remain a fluorinated acetophenone were used as the substrates, only the
diarylated products were observed (shown in eqn (3) and (4)).
Therefore, our chiral phosphoric acid-promoted arylation
reactions for CF3- and CHF2-ketone have overcome the main
drawbacks of the formation of the bisindole products, and can
be a powerful strategy to construct the optically active trifluoro-
methyl or difluoromethyl-substituted tertiary alcohols, which
are valuable pharmaceutical intermediates and chiral building
blocks. Furthermore, the interesting result that the number of
the F atoms in the CH3−nFn (n = 0, 1, 2, 3) group of the sub-
strate ketones determines the final products was found in the
aInstitute for Computational Science and Engineering, Laboratory of New Fiber
Materials and Modern Textile, the Growing Base for State Key Laboratory, Qingdao
University, Qingdao 266071, China. E-mail: faplhl@eyou.com, apfu@qdu.edu.cn;
Fax: +86-532-85950768; Tel: +86-532-85950767
bDepartment of Chemistry, Key Laboratory of Systems Bioengineering (the Ministry of
Education), State Synergetic Innovation Center of Chemical Science and Engineering,
Tianjin University, Tianjin 300072, China. E-mail: majun_an68@tju.edu.cn
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c3ob42157k
1908 | Org. Biomol. Chem., 2014, 12, 1908–1918
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
ð1Þ
ð2Þ
ð3Þ
ð4Þ
present reactions. This intriguing fluorine-substitution effect bifunctional catalysts bearing both Lewis-basic site and
of the substrates on the reaction outcome and the important Brønsted-acidic site which activate both the electrophile and
role of the chiral phosphoric acid catalyst in the reaction nucleophile simultaneously, and this type of catalytic model
between carbonyl compounds and indole motivated us to was successfully used by Akiyama and Yamanaka, Goodman
study the reaction mechanism of the above mentioned aryla- and Simón, and Himo et al. in their computational insight
tion or diarylation process in detail. Actually, a number of into the origin of the enantioselectivity of various chiral
chiral phosphoric acid-catalyzed asymmetric Mannich,7 trans- BINOL-phosphoric acid derivatives-catalyzed asymmetric
fer hydrogenation,8 Strecker,9 and Friedel–Crafts reactions10 reactions.7–13 However, all these pioneering computational
have been studied by density functional theory methods. studies mainly focused on the reactions involving imine as the
These pioneering theoretical and experimental studies1–13 electrophile, and the reactions employing other electrophiles
have established that the chiral phosphoric acids are e.g. carbonyl compounds have been rarely seen.12,13 In fact,
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 1908–1918 | 1909

View Article Online
Paper
Organic & Biomolecular Chemistry
our chiral phosphoric acid-catalyzed arylation or diarylation trifluoromethyl ketone. Then the important role of the CF3
reactions6 represent one type of Friedel–Crafts reactions group of the substrate involved in the Friedel–Crafts reaction
employing different aromatic ketones as the electrophiles. As and the different ratio of the F-substitution effect on the final
one of the most important carbon–carbon bond forming reac- product of the fluorinated methyl ketones including mono-
tions, theoretical investigation into the detailed mechanism of fluoro, difluoro, and trifluoro groups are investigated, and
those kinds of reactions is crucial for further development of further experimental verification of our computational predic-
the bifunctional chiral Brønsted acid catalyst and for substrate tion is reported. Finally, we will explore the origin of high
scope widening. Furthermore, the insight into the remarkable enantioselectivity in the chiral phosphoric acid-catalyzed single
product dependence on the F-substitution in the substrates for arylation between trifluoromethyl aromatic ketone and indole.
the chiral phosphoric acid-promoted process has not been
3.1 Full reaction mechanism using the simplified catalyst
model
explored so far. Hence, to extend the general understanding of
the mechanism and stereoselectivity of the chiral phosphoric
acid-catalyzed organic reactions, the present theoretical study On the basis of our experimental results6 and the relative
was performed to address the following questions: (a) what is studies of others,7–13,18–19 the reaction mechanism of the
the mechanism of the chiral phosphoric acid-catalyzed aryla- chiral phosphoric acid-catalyzed diarylation reaction between
tion or diarylation processes of different ketones? (b) What is indole and methyl ketone is proposed and illustrated in
the intriguing role of the CF3 group and how does the different Scheme 1. Although the diarylation product for the CF3 ketone
ratio of the F-substitution in methyl ketone determine the was not observed experimentally, we also tried this process for
final products? (c) What is the origin of enantioselectivity for comparison purposes. As shown in Scheme 1, the reaction
the single arylation of trifluoromethyl ketones?
started with a condensation of the ketone and indole in the
presence of Brønsted acid, to give a hydrogen bond complex
IN1. Then IN1 underwent the enantioselective first C–C bond
forming reaction through TS1 to yield zwitterion intermediate
IN2 with the chiral phosphoric acid catalyst activating both the
ketone and the indole simultaneously. The subsequent C–H
deprotonation via TS2 recovered the aromaticity of the indole
ring and generated the hydrogen bond complex IN3, and then
the single Friedel–Crafts product and catalyst could be released.
As our experimental results reported, the arylation reaction of
CF3 ketone ceased here, while for the methyl ketone, the follow-
ing chiral phosphoric acid catalyzed-dehydration process
continued through TS3 to yield the carbocation IN4. After
removing the water and adding another indole molecule, the
second Friedel–Crafts reaction started via TS4 to generate IN6.
Similarly, the subsequent C–H deprotonation yielded the cata-
lyst and the double Friedel–Crafts product. Then the diarylation
reaction was completed. According to this full mechanism,
the whole diarylation reactions proceeded through three major
stages: the first Friedel–Crafts, dehydration and the second
Friedel–Crafts reactions. Based on our experimental obser-
vations and the envisaged mechanism, we can assume that the
first Friedel–Crafts reaction controls the stereochemistry of the
single arylation product and the dehydration process deter-
mines the re-arylation to proceed or not. These two steps would
be the key steps for us to investigate. Next, we have performed
the DFT calculations to confirm our hypothesis.
2. Computational details
To reduce the computational cost, for the study of the full reac-
tion mechanisms of catalyst 1-mediated diarylation processes
for acetophenone and 2,2,2-trifluoroacetophenone, the simpli-
fied model e.g. buta-1,3-diene-1,4-diol-phosphoric acid was
initially used instead of the real catalyst system employed in
the reaction. All ground state and transition state (TS) geome-
tries were located using full DFT calculations at the BH and
HLYP/6-31G (d) level of theory.14 To check the validity of the
results at the above computational level, we have reoptimized
several important key steps employing the 6-31+G (d, p) basis
set and other popular functionals, such as PBE1PBE, M06-2X
and B3LYP.14 Calculations on the real catalysts system were
performed using the two-layer ONIOM method,15 in which the
higher-level layer was treated using the BH and HLYP/6-31G(d)
level of theory (or the PBE1PBE/6-31G(d) and M06-2X/6-31G(d)
level for comparison purposes), while the HF/3-21G was
chosen for the lower-level region. Frequency calculations were
also carried out at the same level of theory. Bulk effects of the
solvent CH2Cl2 have been taken into account by means of a
dielectric continuum represented by the conductor polarizable
continuum model (CPCM),16 with united-atom Kohn–Sham
(UAKS) radii. The single-point calculations were done upon the
optimized gas phase geometries with a dielectric constant ε =
8.93 for CH2Cl2. The natural population analysis was per-
formed at the BH and HLYP/6-31G (d) level. Most of the calcu-
lations were carried out using the Gaussian 03 program,17 and
M06-2X computations were performed using Gaussian09.17
To reduce the computational cost, we carried out the study
of the possible reaction mechanism with a simplified catalyst
model, buta-1,3-diene-1,4-diol-phosphoric acid, which has
been successfully used by Simón and Goodman in their
theoretical investigations of the chiral BINOL-phosphoric acid-
catalyzed reactions.8a,9,10a,13a,b Because there may exist several
competing reaction channels due to the different approaching
modes of the two reaction partners of indole and ketone, we
3. Results and discussion
In this section, we will first discuss the full mechanism of the have performed a preliminary investigation of the first C–C
Brønsted acid-catalyzed diarylation of methyl ketone and bond formation process of indole with 3a and 3b since it is
1910 | Org. Biomol. Chem., 2014, 12, 1908–1918
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 1 Plausible reaction mechanism of the diarylation reactions for acetophenone and trifluoroacetophenone catalyzed by chiral phosphoric
acid.
considered to be the stereochemistry-determining step for the into the role of the Brønsted acid in the reaction of aromatic
chiral-phosphoric acid catalyzed Friedel–Crafts reaction and is C–H bonds with carbonyl compounds. The detailed discussion
crucial for the structure of the final product. Meanwhile, for is presented in the ESI.† Based on the preliminary exploration
comparison purposes, we have also considered the corres- of the major channel in the first C–C bond forming process
ponding uncatalyzed process for 2,2,2-trifluoro acetophenone (Fig. S1 and the related discussion in the ESI†), the intermedi-
(3a) and acetophenone (3b) in an attempt to obtain insight ates and the transition states for the whole mechanism of the
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 1908–1918 | 1911

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 1 BH and HLYP/6-31G (d) computed energy profiles for the simplified model of the chiral phosphoric acid-catalyzed diarylation reaction
between indole and 2,2,2-trifluoroacetophenone. The values in parentheses correspond to the single point CPCM calculation in CH2Cl2.
diarylation process were optimized at the BH and HLYP/6-31G(d) which proceeds via C–C bond formation followed by deproto-
level. The important structures and key geometric parameters nation. It is generally accepted that in the Friedel–Crafts reac-
for the reactions involving 3a and 3b are presented in Fig. S2 tions, the first nucleophilic attack of the aromatic ring to the
and S3,† respectively. The computed energy profiles of the electrophilic partner is the rate-determining step and the sub-
subject reactions are illustrated in Fig. 1 and 2.
sequent proton transfer is a fast process. As confirmed by the
As has been proposed, in the initial step of the Friedel– potential energy surfaces for the phosphoric acid-catalyzed dia-
Crafts reaction, the Brønsted acid, ketone, and indole first rylation processes of two ketones illustrated in Fig. 1 and 2,
form a termolecular complex IN1, which is stabilized by the the activation barrier for the nucleophilic attack is much
favorable hydrogen bond interactions of 19.4 kcal mol−1 for higher than that of the following deprotonation process
IN1a and 21.5 kcal mol−1 for IN1b. Calculations indicate that whether for the CF3 or CH3 ketone (19.0 vs.16.0 kcal mol−1 for
phosphoric acid tends to donate the acidic proton to the car- CF3 ketone, 22.7 vs. 15.5 kcal mol−1 for CH3 ketone). Further-
bonyl oxygen of ketone with the O⋯H distance of 1.76 Å in more, in the key C–C bond forming step, the barrier for TS1
IN1a and 1.66 Å in IN1b. Simultaneously, the basic site of involving CF3 ketone is much lower than that involving non-
phosphoryl oxygen can accept a hydrogen from the NH moiety fluorinated CH3 ketone (19.0 vs. 22.7 kcal mol−1), and the
of indole with the O⋯H distance of 1.88 Å in IN1a and 1.90 Å larger positive reaction enthalpy also suggests that the reaction
in IN1b. These strong hydrogen-bonding interactions can involving methyl ketone is more endothermic than that invol-
result in an increase of the electrophilic character of the carbo- ving CF3 ketone. Hence, the calculated data confirm the
nyl compound and the nucleophilicity of indole. Furthermore, important role of the electron-withdrawing CF3 group in acti-
both the H⋯O distances and the stabilization energies demon- vating the substrate in the Friedel–Crafts reaction, which corre-
strate that the hydrogen-bond interaction between the catalyst sponds well with the initial FMO analysis (ESI†) and the
and the non-fluorinated methyl ketone 3b is stronger than experimental observation of the higher reactivity of CF3
that of the CF3 ketone 3a. This can be attributed to the strong ketone. Then after the fast proton transfer process through
electron-withdrawing nature of the CF3 group, which lowers TS2, the phosphoric acid bonded single arylation product IN3
the electron density of the oxygen in the carbonyl group (−0.60 is generated. Fig. 1 and 2 clearly show that the single FC
in IN1a vs. −0.65 in IN1b) and makes it a weaker hydrogen product IN3a is 8.2 kcal mol−1 more stable than that of its
bond acceptor than methyl ketone. Then the pre-activated counterpart IN3b. Hence, in accord with the experimental
complex IN1 enters the following Friedel–Crafts reaction, observations, the presence of the trifluoromethyl group in the
1912 | Org. Biomol. Chem., 2014, 12, 1908–1918
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 2 BH and HLYP/6-31G (d) computed energy profiles for the simplified model of the chiral phosphoric acid-catalyzed diarylation reaction
between indole and acetophenone. The values in parentheses correspond to the single point CPCM calculation in CH2Cl2.
substrate has resulted in a reduction of the activation barrier activation barrier for 3a (30.7 kcal mol−1) is substantially
and an enhancement of the product stability in the single higher than its 3b counterpart (23.0 kcal mol−1) by 7.7 kcal
arylation process, and enables the FC reaction involving 3a to mol−1. Furthermore, the formation of the carbocation for 3a is
be more favorable than its 3b counterpart.
much more endergonic than 3b (12.5 vs. 5.6 kcal mol−1). In
Following the first Friedel–Crafts reaction, in most cases, fact, the difficulty in formation of the carbocation for 3a is
the resulting product of tertiary alcohol IN3 would undergo a consistent with our chemical intuition: (a) the apparent short-
dehydration process because of the good leaving-group charac- ening of the C–O bond when comparing IN3a to its non-fluori-
ter of the hydroxyl moiety under the employed acidic con- nated analogue IN3b (1.41 vs. 1.42 Å) indicates that cleavage of
ditions.1 In fact, this is the main drawback of many analogous the C–O bond is difficult for the trifluoromethyl-substituted
reactions between carbonyl compounds and indole promoted tertiary alcohol; (b) compared with the CH3 group, the strong
by Brønsted acid, in which the indolemethanol derivatives electron-withdrawing nature of the CF3 moiety could highly
formed in the first stage are apt to dehydrate and undergo destabilize the positive charge in carbocation IN4a. Hence,
further indole addition with the formation of bisindole pro- compared with that of 3b substrate, all the above discussions
ducts. More intriguingly, our experiments have demonstrated suggest that the dehydration process for 3a can be effectively
that the incorporation of the CF3 group into the substrate can prohibited at room temperature, and the reaction will stop at
substantially change this inclination of dehydration and the single Friedel–Crafts reaction stage.
makes the reaction cease at the single F–C stage. Then theore-
tical insight into the product switch arising from the trifluoro- another indole molecule triggered the second Friedel–Crafts
methyl group is continually presented in the following. process. Considering the overall mechanism of the diarylation
Afterwards, removal of the water molecule and addition of
As shown in Fig. 1 and 2, the dehydration of IN3 through process shown in Fig. 1 and 2, as we have proposed, the first
TS3 to form the carbocation IN4 with the cleavage of the C–O C–C bond formation and the dehydration process would be
bond also takes place with the assistance of the phosphoric the most important steps which determine the enantio-
acid catalyst which donates the proton to the single arylation selectivity and the final product, respectively. Therefore, to
product of tertiary alcohol. Although TS3a and TS3b feature verify our results at the BH and HLYP/6-31G(d) level, the inter-
nearly the same C–O bond lengths (1.88 Å), the predicted mediates and the transition states for these two key steps have
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 1908–1918 | 1913

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 1 Computed activation enthalpies (kcal mol−1) of the first C–C bond forming and dehydration steps for the simplified model of chiral phos-
phoric acid-catalyzed diarylation process between acetophenone and indole or different ratio of F-substituted acetophenone and indole at various
levels of theory
First C–C bond formation
Dehydration
CF3
Method and basis sets
CF3
CHF2
CH2F
CH3
CHF2
CH2F
CH3
BHandHLYP/6-31G(d)
BHandHLYP/6-31+G(d, p)
PBE1PBE/6-31G(d)
PBE1PBE/6-31+G(d, p)
M06-2X/6-31+G(d, p)
B3LYP/6-31+G(d, p)
19.0
19.6
10.9
10.4
11.6
16.9
19.1
19.6
11.3
9.5
12.6
15.7
19.8
20.7
11.3
12.3
13.1
18.0
22.7
23.3
15.9
15.9
16.4
21.2
30.7
30.0
27.2
27.2
28.7
21.0
30.3
26.2
26.3
23.3
26.7
17.7
24.4
21.6
20.9
18.3
21.5
15.1
23.0
20.7
20.3
18.0
20.8
14.8
been reoptimized employing three other popular DFT CF3, CHF2, CH2F, and CH3 ketones has also been carried out
methods such as PBE1PBE, M06-2X, and B3LYP at the 6-31+G- by the semiempirical FMO theory developed by Yu et al.20 and
(d,p) basis set level. The calculated activation barriers for CF3 the details have been shown in Table S1.† Considering the
and CH3 ketone have been presented in Table 1.
reaction occurring at room temperature, the predicted values
From the table we can see that all the DFT methods exhibit can also satisfactorily explain why CHF2 only affords the single
a similar trend of a lower activation barrier in the first C–C F–C product while CH2F yields the double F–C product.
bond-forming step and a much higher barrier in the dehy-
It is also worth noting that although the higher energetic
dration process for CF3 ketone 3a in comparison with the non- barriers of the rate-determining dehydration process for CF3
fluorinated ketone 3b. In the reaction pathway of 3b, the rela- and CHF2 methyl ketones suggest that the formation of the
tive energies of the two transition states for the nucleophilic diarylation product is prohibited, the predicted barriers for
attack and the dehydration step are comparative and the rate- them are in the range of 24 kcal mol−1 to 31 kcal mol−1, which
determining step depends on the basis sets and methods. is not indeed too high to be overcome. Considering that our
However, all the methods give a relatively low barrier which reported reactions took place at room temperature, we can
can be easily overcome at room temperature, then the experi- expect that raising the reaction temperature could facilitate the
mentally observed diarylation can proceed smoothly. Instead, rate-determining dehydration process for the subsequent re-
for the reaction pathway of CF3 ketone 3a, the dehydration arylation reactions. Therefore, further experimental investi-
step obviously became rate-determining with the significant gation was carried out to verify the prediction based on the
barrier that could not be surmounted at the room temperature. above discussions.
It is worth noting that except the B3LYP method which pre-
dicts the much lower activation barrier for the dehydration
3.2 Experimental validation
step, the other three DFT methods could satisfactorily repro-
duce the perfect double role of the CF3 group in activating sub-
strates and stabilizing single arylation adduct of tertiary
alcohol. In summary, given that the reactions were performed
at room temperature and the dehydration barrier was enlarged
to a great extent, the diarylation process is unlikely to proceed
for CF3 ketone, which is consistent with the experimental
observation that indole and 3a only yield the single Friedel–
Crafts product at room temperature, while indole and CH3
ketone afford the diarylated product under the same
conditions.
Moreover, we have also considered the two key steps of dia-
rylation for ketones involving CHF2 and CH2F moieties. The
activation enthalpies for those two substrates at different levels
are also shown in Table 1. Combined with their CF3 and CH3
counterparts, the remarkable fluorine effect on the final
product control is also identified by the calculated barriers. As
expected, the different ratios of F-substitution in CH3 of
methyl ketone can activate the carbonyl group in the C–C
bond forming step, and stabilize the FC product of tertiary
alcohol to a different extent. The analysis of the relationship
between the LUMO energies of the electrophiles and the acti-
vation energies in the C–C bond formation steps for the set of
To extend the hypothesis that the higher temperature could
effectively activate the CF3 and CHF2 stabilized arylation
product of tertiary alcohol, we next tried the reaction of indole
with the single arylation product of trifluoromethyl ketone e.g.
2,2,2-trifluoro-1-(1H-indole-3-yl)-1-phenylethanol (product 4a
in the previous reactions) in the presence of racemic phospho-
ric acid 1 at the reaction conditions with higher temperature.
To our delight, the desired double arylation product was
obtained in 79% yield at 40 °C (Table 2, entry 2) while the
control experiment demonstrated that a higher temperature
was indispensable for the success of the target reaction
(Table 2, entry 2). Different polar aprotic solvents were also
screened (Table 2, entries 3–5), and sharp slide of yields were
observed when THF and CH3CN were employed. In contrast,
an apolar solvent toluene gave a good yield in 60 h. After per-
forming the reaction at a higher temperature of 80 °C, a quan-
titative conversion of the re-arylation product was observed in
24 h (Table 2, entry 7). Otherwise, a blank experiment was con-
ducted, showing that a phosphoric acid was vital to this trans-
formation even at a higher temperature (Table 2, entry 10). Not
surprisingly, the single Friedel–Crafts product of CHF2 ketone
(2,2-difluoro-1-(1H-indole-3-yl)-1-phenylethanol) also has
a
strong tendency to generate the re-arylation product at the
1914 | Org. Biomol. Chem., 2014, 12, 1908–1918
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 2 The Brønsted acid-catalyzed rearylation reaction of 2,2,2-trifluoro-1-(1H-indole-3-yl)-1-phenylethanol and 2,2-difluoro-1-(1H-indole-3-
yl)-1-phenylethanol
Entry
R
Catalyst (mol%)
Solvent
Temp. (°C)
Time (h)
Yielda (%)
1
2
3
4
5
6
7
8
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CHF2
5
5
5
5
5
5
5
1
CH2Cl2
CH2Cl2
CHCl3
25
40
40
40
40
40
80
80
80
80
80
72
60
60
60
60
60
24
120
16
120
12
—
79
75
23
40
82
98
73
91
—
95
THF
CH3CN
Toluene
Toluene
Toluene
Toluene
Toluene
Toluene
9
10
11
10
—
5
a Isolated yield.
raising temperature, and the bisindole product was obtained excellent enantiomeric excess (eqn (1)). Based on the above
in 95% yield at 80 °C (Table 2, entry 11). Meanwhile, some study of the full mechanism involving a simplified model of
“one-pot” double-Friedel–Crafts reactions were carried out phosphoric acid, in the following we will focus on the origin of
smoothly under the optimal reaction conditions (eqn (5)). And enantioselectivity for the chiral phosphoric acid-catalyzed
it gave 96% (CF3 ketone) and 67% (CHF2 ketone) yield, respect- Friedel–Crafts reaction involving CF3 ketone. As has been dis-
ively. Those additional experimental results are in good accord cussed, the Friedel–Crafts reaction proceeds via two steps e.g.
with our computational predictions. The experimental details the C–C bond formation and the subsequent C–H deprotona-
are provided in the ESI.†
tion process. Since the deprotonation step is neither a rate-
In conclusion, both theoretical and experimental studies determining nor a stereochemistry-determining one, in this
demonstrate that it is possible to tune the single arylation and section we have only focused on the enantioselective C–C bond
diarylation process by introducing the CF3 (CHF2) group into forming process e.g. the aromatic indole addition to the
the substrate and altering the reaction temperatures.
ketone. Combined with the preliminary study using the simpli-
fied model (Fig. SI† and the related discussion), four reactive
channels corresponding to two stereoisomers have been con-
3.3 Enantioselectivity using the real catalyst model
Our experiments6 for the reaction of indole with the aromatic sidered. These include two possible diastereomeric transition
methyl ketones promoted by chiral phosphoric acid have states of indole attacking the re and si faces of CF3 ketone, and
shown that trifluoromethyl ketone can afford the resulting tri- two possible relative orientations of the unsymmetrical indole
fluoromethyl-substituted tertiary alcohol in high yield and to the electrophile ketone. Simón and Goodman have
ð5Þ
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 1908–1918 | 1915

View Article Online
Paper
Organic & Biomolecular Chemistry
suggested that the ONIOM method8a,9,10a,13a could be success- orientation suggests that the following two factors might be
fully used to treat large catalyst systems like BINOL-based responsible for the observed enantioselectivity. First, TS1-re-
phosphoric acid in the investigation of the reasonable mech- exo has the shortest hydrogen-bond distances between the
anism and the explanation of enantioselectivity in various phosphoric acid catalyst and the two reaction partners (TS1-re-
asymmetric reactions. Inspired by these successes, we have exo: 1.78 and 1.33 Å, TS1-si-exo: 2.07 and 1.39 Å). Secondly,
performed the DFT calculation of the four required TSs by a TS1-re-exo experiences less steric repulsion when the reaction
complete model for the phosphoric acid catalyst using the partners enter into the binding pocket of the catalyst, while in
ONIOM method, in which all atoms in the catalyst were TS1-si-exo, the aromatic ring of indole is close to the bulky
included in the lower-level layer except the phosphoric acid 2,4,6-triisopropylphenyl group of the catalyst and suffer from
moiety, which was included in the higher-level layer, together more steric hindrance. As suggested in Fig. S6,† this unfavor-
with the ketone and indole involved in the reaction. The DFT able interaction can be reflected by the large deviation from
functional of BH and HLYP combined with the 6-31G (d) basis the eclipsing conformation of the isopropyl Hs at 4-positions
set was used in the higher-level layer, and the HF/3-21G was on the 2,4,6-triisopropylphenyl substituents to the phenyl
used in the lower-level layer. Because the different confor- double bond which is required to diminish the steric repulsion
mation of the steric 2,4,6-triisopropylphenyl substituents on between the catalyst and the substrates when the catalyst tries
the catalyst can directly affect the stability of the different TSs to accommodate the substrates into its chiral pockets. Thus,
and subsequently the enantioselectivity, we have carried out the above two factors direct the incoming indole approaching
the conformational analysis of the catalyst and the related TSs the re-face of ketone and, as expected, the presence of the
first and the details are presented in the ESI (Fig. S5 and S6†). large substituent at the 3,3′-positions is a crucial element to
The most stable conformers of the TSs in the four reaction induce asymmetry in the chiral phosphoric acid catalyzed
channels have been used in the following discussion of the reactions.
reaction enantioselectivity.
For the purpose of clarifying the steric effect of the substitu-
ent at the 3,3′-position in the chiral phosphoric acid catalyst,
we have also considered four similar TSs involving the simpli-
4. Conclusions
fied model (buta-1,3-diene-1,4-diol-phosphoric acid and biphe- In this work, the chiral phosphoric acid-catalyzed arylation
nyl-phosphoric acid), and the optimized geometrical and diarylation reactions between indole and CH3−nFn-ketones
structures and the relative enthalpies are illustrated in (n = 0, 1, 2, 3) have been investigated by DFT methods. The full
Fig. S4.† The TS structures and relative enthalpies involving mechanism of the diarylation process for 2,2,2-trifluoromethyl
the complete real catalyst are shown in Fig. 3. As depicted in ketone and methyl ketone has been considered employing the
these figures, different orientations of the indole ring affect simplified catalyst model e.g. buta-1,3-diene-1,4-diol-phospho-
the energies of the resulting transition structures greatly. “TS1- ric acid using BH and the HLYP/6-31G(d) method. The two
endos” with the two simplified catalyst models are about crucial steps, including the first C–C bond formation and the
5–7 kcal mol−1 higher than their “exo” counterparts, while for dehydration during the diarylation process, have been dis-
the real catalyst with the bulky substituent groups at 3,3′-posi- cussed employing different DFT methods. All the calculations
tion, the energy difference increases to about 7–10 kcal mol−1
.
confirmed that the dehydration process required a lower acti-
Therefore, in the following, we only focus on the two transition vation barrier in the case of the methyl ketone or CH2F ketone,
states of “exo” orientation, e.g. the indole aromatic ring is in and the diarylation process can proceed smoothly. Moreover,
the same direction with the phenyl group of ketone. It is the perfect double role of the CF3 (CHF2) group in activating
observed that the energy difference between the transition the substrate for the first C–C bond forming reaction and sta-
states involving re- and si-facial attacking of indole to ketone bilizing the single Friedel–Crafts product was identified satis-
adopting the “exo” arrangement is nearly negligible since the factorily. In addition, the predicted activation barriers in the
simplified model of the catalyst lacks the bulky groups. When dehydration process for CF3 (CHF2) ketone range from 24 to
the full structure of the catalyst is included, as shown in Fig. 3, 31 kcal mol−1, which suggests that the generation of the diaryl-
the most stable transition structure corresponds to the one in ation product is not really prevented at the higher tempera-
which indole attacks on the re face of the ketone, lying 2.5 kcal ture. The additional experiments at 80 °C sufficiently validated
mol−1 lower in enthalpy than the si-facial attack. When the this theoretical prediction. Hence, our theoretical and experi-
solvent effect is taken into account, the energy difference mental results suggest that the arylation and the diarylation
changes to 1.8 kcal mol−1. Furthermore, the combination of products can be switched by tuning the temperature and the
PBE1PBE/6-31G (d) or M06-2X/6-31G (d) with HF/3-21G F-substituent ratio. Finally, the issues of the re-facial selectivity
methods has also been used to reoptimize the four TSs of the in the BINOL-phosphoric acid catalyzed arylation reaction of
complete catalyst model and the differences in enthalpy are CF3 ketone are also addressed. The ONIOM calculations (BH
also shown in Fig. 3. All the methods reproduce the experi- and HLYP/6-31G (d): HF/3-21G) on the transition states of the
mentally observed configuration of the major product and the stereochemistry-determining C–C bond forming step showed
enantioselectivity (92% ee) satisfactorily. Inspection of the geo- that the ketone and indole were both activated by the bifunc-
metric parameters for the two lower-energy TSs with “exo” tional chiral phosphoric acid through the formation of
1916 | Org. Biomol. Chem., 2014, 12, 1908–1918
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 3 DFT computed transition states and relative enthalpies for the C–C bond formation step of the chiral phosphoric acid-catalyzed Friedel–
Crafts reaction (real system) between indole and 2,2,2-trifluoroacetophenone. CPCM values in CH2Cl2 are shown in parentheses. Atoms treated with
the high level of theory are shown as a ball and stick model, while atoms in the low-level layer are shown as a tube model. (a) ONIOM (BH and HLYP/
6-31G(d):HF/3-21G); (b) ONIOM (PBE1PBE//6-31G(d):HF/3-21G); (c) ONIOM (M062X//6-31G(d):HF/3-21G).
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 1908–1918 | 1917

View Article Online
Paper
Organic & Biomolecular Chemistry
hydrogen bonds. The stronger hydrogen-bond interaction
and the less steric repulsion between the bulky group at the
9 L. Simón and J. M. Goodman, J. Am. Chem. Soc., 2009, 131,
4070–4077.
3,3′-position of the catalyst and the substrates existing in the 10 (a) L. Simón and J. M. Goodman, J. Org. Chem., 2010, 75,
re-facial attacking TS cause the (R)-enantiomer to be preferen-
tially produced. Our calculations also successfully reproduce
the observed enantioselectivity.
589–597; (b) C. Zheng, Y.-F. Sheng, Y.-X. Li and S.-L. You,
Tetrahedron, 2010, 66, 2875–2880; (c) T. Hirata and
M. Yamanaka, Chem.–Asian J., 2011, 6, 510–516;
(d) L. Simón and J. M. Goodman, J. Org. Chem., 2011, 76,
1775–1778.
Acknowledgements
11 (a) M. Yamanaka and T. Hirata, J. Org. Chem., 2009, 74,
3266–3271; (b) T. Akiyama, H. Morita, P. Bachu, K. Mori,
M. Yamanaka and T. Hirata, Tetrahedron, 2009, 65, 4950–
4956; (c) F.-Q. Shi and B.-A. Song, Org. Biomol. Chem., 2009,
7, 1292–1298.
12 (a) N. Li, X.-H. Chen, J. Song, S.-W. Luo, W. Fan and
L.-Z. Gong, J. Am. Chem. Soc., 2009, 131, 15301–15310;
(b) S. Xu, Z. Wang, Y. Li, X. Zhang, H. Wang and
K. Ding, Chem.–Eur. J., 2010, 16, 3021–3035; (c) Q. Cai,
C. Zheng and S.-L. You, Angew. Chem., Int. Ed., 2010,
122, 8848–8851.
13 (a) M. N. Grayson, S. C. Pellegrinet and J. M. Goodman,
J. Am. Chem. Soc., 2012, 134, 2716–2722; (b) M. N. Grayson
and J. M. Goodman, J. Am. Chem. Soc., 2013, 135, 6142–
6148; (c) H. Wang, P. Jain, J. C. Antilla and K. N. Houk,
J. Org. Chem., 2013, 78, 1208–1215.
This work was supported by the National Natural Science
Foundation of China (no. 21103096, 21172170 and 21225208),
the Natural Science Foundation of Shandong Province
(ZR2010BM024), and the Project of Shandong Province
Higher Educational Science and Technology Program, China
(J10LB06). We also thank the National Key Project on Basic
Research (grant no. 2012CB722705, 2014CB745100).
References
1 For reviews, see: (a) S.-L. You, Q. Cai and M. Zeng, Chem.
Soc. Rev., 2009, 38, 2190–2201; (b) A. Zamfir, S. Schenker,
M. Freund and S. Tsogoeva, Org. Biomol. Chem., 2010, 8,
5262–5276; (c) V. Terrasson, R. M. de Figueiredo and
J. M. Campagne, Eur. J. Org. Chem., 2010, 2635–2655; 14 (a) A. D. Becke, J. Chem. Phys., 1993, 98, 1372–1377;
(d) M. Rueping, A. Kuenkel and I. Atodiresei, Chem. Soc.
Rev., 2011, 40, 4539–4549.
2 (a) Q. Kang, Z.-A. Zhao and S.-L. You, J. Am. Chem. Soc.,
2007, 129, 1484–1485; (b) M. Terada, S. Yokoyama,
K. Sorimachi and D. Uraguchi, Adv. Synth. Catal., 2007,
(b) A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652;
(c) C. Lee, W. Yang and R. G. Parr, Phys. Rev. B:
Condens. Matter, 1988, 37, 785–789. For M06-2X, see:
(d) Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120,
215–241.
349, 1863–1867; (c) G.-W. Zhang, L. Wang, J. Nie and 15 (a) M. Svensson, S. Humbel and K. Morokuma, J. Chem.
J.-A. Ma, Adv. Synth. Catal., 2008, 350, 1457–1463.
3 (a) M. Terada and K. Sorimachi, J. Am. Chem. Soc., 2007,
129, 292–293; (b) Y.-X. Jia, J. Zhong, S.-F. Zhu, C.-M. Zhang
and Q.-L. Zhou, Angew. Chem., Int. Ed., 2007, 46, 5565–
5567.
Phys., 1996, 105, 3654–3661; (b) T. Vreven and
K. Morokuma, J. Comput. Chem., 2000, 21, 1419–1432;
(c) S. Dapprich, I. Komaromi, K. S. Byun, K. Morokuma and
M. J. Frisch, THEOCHEM, 1999, 461–462.
16 (a) V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102,
1995–2001; (b) B. Barone, M. Cossi and J. Tomasi,
J. Comput. Chem., 1998, 19, 404–417.
4 Q. Cai, Z.-A. Zhao and S.-L. You, Angew. Chem., Int. Ed.,
2009, 48, 7428–7431.
5 J. Itoh, K. Fuchibe and T. Akiyama, Angew. Chem., Int. Ed., 17 (a) M. J. Frisch,et al., GAUSSIAN03 (Revision D.01),
2008, 47, 4016–4018.
6 J. Nie, G.-W. Zhang, L. Wang, A. P. Fu, Y. Zheng and
J.-A. Ma, Chem. Commun., 2009, 2356–2358.
Gaussian, Inc., Wallingford, CT, 2004; (b) M. J. Frisch, et al.,
GAUSSIAN 09 (Revision C.01), Gaussian, Inc., Wallingford,
CT, 2010.
7 M. Yamanaka, J. Itoh, K. Fuchibe and T. Akiyama, J. Am. 18 A. S. Cothelf, T. Hansen and K. A. Jørgensen, J. Chem. Soc.,
Chem. Soc., 2007, 129, 6756–6764. Perkin Trans. 1, 2001, 854–860.
8 (a) L. Simón and J. M. Goodman, J. Am. Chem. Soc., 2008, 19 (a) F.-L. Sun, M. Zeng, Q. Gu and S.-L. You, Chem.–Eur. J.,
130, 8741–8747; (b) T. Marcelli, P. Hammar and F. Himo,
Chem.–Eur. J., 2008, 14, 8562–8571; (c) T. Marcelli,
P. Hammar and F. Himo, Adv. Synth. Catal., 2009, 351, 525–
2009, 15, 8709–8712; (b) S.-G. Wang, L. Han, M. Zeng,
F.-L. Sun, W. Zhang and S.-L. You, Org. Biomol. Chem.,
2012, 10, 3202–3209.
529; (d) Y. Shibata and M. Yamanaka, J. Org. Chem., 2013, 20 L.-G. Zhuo, W. Liao and Z.-X. Yu, Asian J. Org. Chem., 2012,
78, 3731–3736.
1, 336–345.
1918 | Org. Biomol. Chem., 2014, 12, 1908–1918
This journal is © The Royal Society of Chemistry 2014