Aldehyde bis-arylation by metal triflates including bismuth triflate powder

Article abstract of DOI:10.1021/op0000733

Development of an efficient catalytic reaction promoted by a metal triflate has been emphasized in contemporary synthesis. To evaluate the potential of some metal trflates such as bismuth triflate in our process research, we devised here an expedient protocol for its preparation and characterization. Comparison of its catalytic activity with other well-known metal triflates such as scandium triflate and copper triflate was made in a glyoxylate arylation reaction, which is a useful transformation for the construction of a 1,1-diarylethane skeleton. Furthermore, NMR analysis of Bi(OTf)₃ in acetone-d₆ revealed the unique nature and reactivity of the triflate, which has not heretofore been disclosed.

Full text of DOI:10.1021/op0000733

Organic Process Research & Development 2001, 5, 84−88
Aldehyde Bis-arylation by Metal Triflates Including Bismuth Triflate Powder
Yasuhiro Torisawa,* Takao Nishi, and Jun-ichi Minamikawa
Process Research Laboratory, Second Tokushima Factory, Otsuka Pharmaceutical Co., Ltd.,
Kawauchi-cho, Tokushima 771-0182, Japan
Abstract:
On the basis of the Dubac protocol, we devised a simpler
preparative protocol as shown below. We selected toluene
(instead of the undesirable CH2Cl2) as the best solvent for
preparation and work-up. Protodemetalation of triphenylbis-
muth with triflic acid was complete within 3 h at room
temperature (as opposed to 10 h in CH2Cl2 as per Dubac)
and gave a homogeneous solution, from which a fine
powdered Bi(OTf)3 [BIT powder] was obtained after simple
evaporation of toluene as shown below. As an extension,
bismuth mesylate was obtained in the similar manner.
NMR Analysis. Dubac recently reported key spectro-
scopic information about Bi(OTf)3. Our BIT powder, as
prepared above, was identical in 1H-, 13C-, and F NMR with
those reported by Dubac in acetone-d6 solution. These
analyses precluded the major contamination by the bis-triflate
[PhBi(OTf)2] and toluene. Besides key distinctive peaks in
1H and 13C NMR, we observed new peaks gradually
appearing during an oVernight scan for 13C measurement
in acetone-d6. Analysis of these new peaks indicated the
formation of acetone aldol and mesityl-d oxide through a
catalytic action of Bi(OTf)3. While Dubac did not mention
this side reaction, the aldol reaction of deuterated acetone
was useful and effective in judging the activity of self-made
triflate prior to its use (Scheme 1).
As a preliminary survey, it was interesting to compare
the NMR time course of Bi(OTf)3, Sc(OTf)3, and Cu(OTf)2.
Just after sampling, only faint peaks of aldol products were
observed in the case of BIT solution in CD3COCD3. Within
12 h, the solution of BIT in CD3COCD3 turned to black,
and peaks of mesytil-d oxide became evident. On the other
hand, progress of the aldol reaction was very slow with
Sc(OTf)3, and it took 7 days to become the same mixture as
BIT mentioned above. In the case of Cu(OTf)2, no aldol
reaction was observed even after 1 month standing as CD3-
COCD3 solution. We thus observed the difference in Lewis
acid activity of some triflates by simple NMR analysis.
Glyoxylate Bis-arylation. After the spectroscopic analy-
sis, we conducted some synthetic reactions with metal
triflates to evaluate their synthetic potential. Especially, in
relation to our process research on some arylpiperazines
[OPC-14597 (Aripiprazole)],4 we investigated a glyoxylate
arylation as shown below (Scheme 2).
Development of an efficient catalytic reaction promoted by a
metal triflate has been emphasized in contemporary synthesis.
To evaluate the potential of some metal trflates such as bismuth
triflate in our process research, we devised here an expedient
protocol for its preparation and characterization. Comparison
of its catalytic activity with other well-known metal triflates
such as scandium triflate and copper triflate was made in a
glyoxylate arylation reaction, which is a useful transformation
for the construction of a 1,1-diarylethane skeleton. Furthermore,
NMR analysis of Bi(OTf)3 in acetone-d6 revealed the unique
nature and reactivity of the triflate, which has not heretofore
been disclosed.
Introduction
Continuing efforts have been made to attain efficiency
and safety (in both the operational and environmental sense)
for the conventional aromatic acylation and alkylation
processes (Friedel-Crafts reactions). Currently, a catalytic
use of metal trifluoromethanesulfonate derivatives in non-
halogenated media is often recommended in these respects.1
Of special interest is the rare earth triflates [RE(OTf)3],
because of their water-tolerance and recyclable nature.2 It is
as yet premature for us to judge whether these new promoters
(of large molecular weight) will supplant the classic Lewis
acid catalysts (of small molecular weight) in a more
demanding substrate. We thus started to experiment with such
triflates as Sc(OTf)3, Cu(OTf)2, Bi(OTf)3, and other triflates
in our process-oriented research. We disclose herein our
preliminary experimentation with such metal triflates, focus-
ing on some beneficial aspects of bismuth triflate.
Preparation and Characterization of Bi(OTf)3 Powder.
Soon after we started our study on catalytic acylation, Dubac
and co-workers announced a surprising catalytic activity of
Bi(OTf)3 in a Friedel-Crafts acylation.3 This prompted us
to investigate large-scale preparation of bismuth triflate and
its subsequent application in our process research.
* Author for correspondence. E-Mail: torisawa@fact.otsuka.co.jp.
(1) For very recent topics, see: (a) Kotsuki, H.; Ohishi, T.; Inoue, M.; Kojima,
T. Synthesis 1999, 603-606 and references therein. (b) Matsuo, J.;
Odashima, K.; Kobayashi, S. Synlett 2000, 403-405 and references therein.
(2) For reviews, see: (a) Kobayashi, S. Synlett 1994, 689-701. (b) Marshman,
R. W. Aldrichimica Acta 1995, 28, 77-84. (c) Kobayashi, S. J. Synth.
Org. Chem. Jpn. 1995, 53, 370-380.
(3) (a) Desmurs, J. R.; Labrouillere, M.; Roux, C. Le; Gaspard, H.; Laporterie,
A.; Dubac, J. Tetrahedron Lett. 1997, 38, 8871-8874 and references cited
threin. (b) Labrouillere, M.; Roux, C. Le; Gaspard, H.; Laporterie, A.;
Dubac, J. Tetrahedron Lett. 1999, 40, 285-286.
The reaction of dichloroanisole (DCA) with ethyl gly-
oxylate polymer (EGP) under the influence of a catalytic
metal triflate afforded para,para-dimer 1 in good yield after
(4) Oshiro, Y.; Sato, S.; Kurahashi, N.; Tanaka, T.; Kikuchi, T.; Tottori, K.;
Uwahodo, Y.; Nishi, T. J. Med. Chem. 1998, 41, 658-667.
84
•
Vol. 5, No. 1, 2001 / Organic Process Research & Development
10.1021/op0000733 CCC: $20.00 © 2001 American Chemical Society and The Royal Society of Chemistry
Published on Web 11/03/2000

Scheme 1. Preparation and NMR measurement of BIT
heating for 2-3 h. Both with Sc(OTf)3 (at 150-170 °C) and
Bi(OTf)3 (at 135 °C), the reaction worked well at 40-100 g
scale with 0.05-0.1 mol % triflate. These catalysts were
easily filtered off from the mixture after dilution with
solvents, and could be reused after azeotopic drying with
toluene. The symmetric structure of the product (1) was
easily assigned on the basis of the NMR analysis (COSY
and NOESY indicated para-substitution as shown). Thus,
the OMe group in the substrate was a key functionality for
the successful transformation, as well as reliable structural
determination. Attempted reactions with some aniline deriva-
tives (2,3-dichloroanilines and 2,3-dichlorophenylpiperazines)
gave no coupling products.
product. These results indicated Sc(OTf)3 catalyzed both
arylation by EGP and subsequent hydrolysis by H2O formed
during the reaction.
We also discovered that Bi(OTf)3 prepared in situ in the
reaction flask was a suitable reagent for related reaction with
m-chlorophenol (MCP) and acetal. MCP was more reactive
in this bis-arylation reaction, and some polymerization
products or other side-products were formed as a over
reaction at phenol OH. Addition of SiO2 and cosolvents such
as CH2Cl2 was particularly useful to prevent polymerization
and to form a homogeneous mixture. For the structural
assignment, the product 2 obtained from the reaction with
MCA was converted to the product 3 obtained from MCP
through the sequence shown below (Scheme 2).
5
Among the other triflates examined, Hg(OTf)2 showed
exceptional catalytic activity and led to a clean formation
of the dimer 1. The mild reagent Cu(OTf)2 was an alternative
safe choice for this conversion, albeit at low yield (∼50%).
It should be noted here that Hg(OTf)2 formed a clear reaction
mixture, while the other triflates led to an insoluble mixture
(suspension), from which catalysts were removed without
difficulty. The bismuth mesylate [Bi(OMs)3] prepared above
was less effective as Cu(OTf)2. TfOH itself afforded a rather
messy mixture under comparable reaction conditions (pre-
sumably due to demethylation at high temperatures).
Further experimentation revealed that a similar reaction
of m-chloroanisole (MCA) gave a more complex mixture,
in which a substantial amount of the isomer (o,p-dimer) and
their hydrolysis product (carboxylic acid) were formed along
with the desired product ( ∼30-40% yield after chroma-
tography) as indicated in the scheme. For such reaction, we
utilized a convenient catalyst system: Sc(OTf)3-MgSO4-
SiO2. We observed that less hydrolysis product was formed
in the mixture, which resulted in an easy separation of the
Further applications of Bi(OTf)3 in catalytic acylation
(chloroacetylation and trichloroacetylation) have also been
carried out with substrates related to our process research.6
Details of this progress will be reported in due course.
In summary, we disclose herein a user-friendly preparation
of the Bi(OTf)3 (BIT) powder which was stable and active
after ∼2-year storage in the usual flask container. It usually
contains about 7% H2O after long-term storage at room
temperature, which was determined by the analysis (Karl-
Fisher apparatus) and monitored by NMR in acetone-d6. In
the glyoxylate arylation reaction shown above, BIT powder
showed comparable activity with Hg(OTf)2 and Sc(OTf)3.
Easy preparation in a large quantity (from triphenylbismuth)
as well as a stable and recyclable nature demonstrated its
superiority to other triflates for contemporary synthesis.
Exceptionally low toxicity is another benefit of bismuth salts,
which will expand its catalytic and stoichiometric use in
medicinal chemistry.
In our NMR analysis of the metal triflates in acetone-d6,
we observed an aldol reaction of acetone promoted by the
(5) (a) Nishizawa, M.; Morikuni, E.; Asoh, K.; Kan, Y.; Uenoyama, K.;
Imagawa, H. Synlett 1995, 169-170. (b) We also prepared a powder of
Hg(OTf)2 from the reaction of HgPh2 and triflic acid as described in Bi-
(OTf)3.
(6) Torisawa, Y.; Nishi, T.; Minamikawa, J. Abstracts of the 120th meeting
of Pharmaceutical Society of Japan, Gifu, 2000, p 111.
Vol. 5, No. 1, 2001 / Organic Process Research & Development
•
85

Scheme 2. Glyoxylate bis-arylation
triflate to produce mesityl-d oxide. We are now extending
this NMR assay to other metal triflates to monitor the
reactivity of metal triflates towards solvent acetone. This
convenient assay reflects the nature of the reagent (associa-
tion with water) and its catalytic activity towards simple
carbonyls (formation of mesityl oxide).
and attached H2O, and elemental analysis of this sample did
not give correct values (Anal. Calcd for BiC3F9S3: C, 5.49;
H, 0.00. Found: C, 6.32; H, 1.95). Karl-Fisher analysis
indicated constant water content at 6.98-7.00%. This BIT
powder prepared in our hands was identical with solvent-
free material reported by Dubac in 13C-and F NMR.3 After
13C NMR measurement (one overnight, ∼12 h), there
appeared the clear peaks of mesityl-d oxide: 196.6, 152.7,
122.4t, plus satellites at 67.7, 52.3, 26.70, 17.71. The long-
stored material could be further azeotropically dried by
toluene under reduced pressure to furnish the sample of
constant water content (∼7%) as a nonhygroscopic powder.
Reaction of DCA with EGP. (A) By Bi(OTf)3. EGP
solution (50% in toluene, TCI, 40.0 mL) was first concen-
trated under reduced pressure to obtain neat EGP (20 g), to
which was added DCA (Aldrich, 45 g, 257 mmol), and the
whole was heated at around 80 °C to form a clear
homogeneous solution. To this was next added Bi(OTf)3
powder (1.0 g) in one portion. The resulting mixture was
gradually heated to 110°C ( ∼1 h). After monitoring the
reaction by TLC, the mixture was further heated at ∼135
°C for 1 h. The insoluble reagent precipitated after dilution
with CH2Cl2 at room temperature was recovered through
filtration and washing with CH2Cl2. The filtrate was washed
with saturated aqueous NaHCO3 and concentrated to give
nearly pure product as orange oil (1) (60 g). 1: 1H NMR
(CDCl3) δ 1.28 (3H, t, J ) 4 Hz), 3.90 (6H, s), 4.25 (2H, q,
J ) 4 Hz), 5.70 (1H, s), 6.82 (2H, d, J ) 8.7 Hz), 6.94 (2H,
d, J ) 8.7 Hz). 13C NMR (CDCl3) δ 171.2; 155.4; 134.0;
128.6; 127.2; 122.3; 109.5; 61.5; 56.2; 52.0; 13.9. MS (EI)
m/z 436 (M+), 438, 440.
Experimental Section
Materials and Instrumentation. Chemicals were pur-
chased from commercial firms indicated and used without
further purification. TLC analysis was carried out using Merk
silica gel 60 F254 plate (Art 5715). 1H- and 13C NMR spectra
were measured at 300 and 75 MHz, respectively, with
tetramethylsilane as the internal standard. NMR solvents
(CDCl3, CD3COCD3) were purchased from Cambridge
Isotope Laboratory (CIL). Infrared (IR) spectra were recorded
with a Perkin-Elmer 1600 series FTIR apparatus. Mass
spectra were recorded with a Shimadzu GCMS-QP1000
spectrometer at 70 eV.
Preparation of BIT Powder. A neat solution of CF3-
SO3H (TCI, 25.0 g) was added at room temperature to a
stirred solution of Ph3Bi (Strem, 22.5 g) in toluene (250 mL).
After an exothermic process had ceased, the colored mixture
was kept stirring for 3 h at room temperature before being
concentrated under reduced pressure to form a free-flowing
solid mass of Bi(OTf)3 (quant).
Solvent System. As indicated above, toluene was found
to be the most convenient solvent to obtain a good powder
from the reaction mixture by simple evaporation with proper
heating. In a reaction using CH2Cl2, acetonitrile, and DME,
the resulting solid mass was gradually changed in color to
orange-brown during storage at room temperature.
NMR Measurement. 1H NMR measurement in acetone-
d6 (1% TMS, CIL) indicated the presence of some toluene
(B) By Sc(OTf)3. To a homogeneous mixture of the neat
EGP (40.0 g) and DCA (100 g) preheated at around 70 °C
was added Sc(OTf)3 (Aldrich, 1.0 g), and the resulting
86
•
Vol. 5, No. 1, 2001 / Organic Process Research & Development

mixture was gradually heated to 150-170 °C. Heating was
continued for 1 h to form a dark brown mixture. TLC
analysis indicated the complete disappearance of starting
material with concomitant formation of the more polar
products. The mixture was cooled to room temperature and
diluted with CH2Cl2. The insoluble precipitate was then
recovered through filtration and washing with CH2Cl2. The
filtrate was washed with saturated aqueous NaHCO3 and
concentrated to give crude product as a brown oil (1)
(130 g).
(C) By Hg(OTf)2. To a homogeneous mixture of the neat
EGP (2.5 g) and DCA (5.0 g) preheated at around 70 °C
was added Hg(OTf)2 (70 mg), and the resulting mixture was
gradually heated to 135 °C. Heating was continued for ∼1
h until TLC analysis indicated the complete disappearance
of starting material with concomitant formation of the more
polar products. The mixture was cooled to room temperature
and diluted with CH2Cl2. The insoluble precipitate formed
after standing was separated by decantation followed by
washing with CH2Cl2. The filtrate was washed with saturated
aqueous NaHCO3 and concentrated to give nearly pure
product as orange oil (1) (5.5 g).
the mixture was kept stirring at room temperature for 3 days.
After the completion of the reaction, the mixture was diluted
with CH2Cl2, and the insoluble solids were filtered . Filtrates
were concentrated to give crude product as black oil, which
was further purified by repeated SiO2 column chromatog-
raphy to give 3 (10.6 g, 38.5%). 3: 1H NMR (CDCl3) δ 1.50
(3H, d, J ) 7.1 Hz), 4.74 (1H, q, J ) 7.1 Hz), 4.83 (2H, bs,
OH), 6.68 (2H, dd, J ) 8.4, 2.7 Hz), 6.87 (2H, d, J ) 2.7
Hz), 6.97 (2H, d, J ) 8.4). 13C NMR (CDCl3) δ 154.2; 134.6;
134.3; 128.6; 116.4; 113.8; 36.8; 20.9; 13.9. MS (EI) m/z
282 (M+), 284.
Conversion of 2 into the 1,1-Diarylethane (3). To a
stirred solution of the ester 2 (22.8 g, 61 mM) in DME (250
mL) was added ZrCl4 powder (30.0 mM), and to the resulting
mixture was added NaBH4 (145.5 mM) at room temperature
slowly and carefully. The mixture was stirred at room
temperature for ∼1 h, followed by gradually heated to 80
°C for 0.5 h (only gentle reflux). TLC (AcOEt/n-hexane )
1/5) now indicated almost all the starting material consumed.
After cooling to room temperature, the mixture was carefully
quenched by the slow addition of H2O. The precipitated
mixture was diluted with CH2Cl2. After insoluble solid mass
was removed by filtration, resulting organic layer was washed
with brine. Evaporation of the dried solvent (MgSO4) gave
crude product which was further subjected to azeotoropic
drying by toluene to give a crude alcohol (19.1 g) suitable
for next bromination. Caution: do not heat the reduction
mixture immediately after the addition of the reagent
(NaBH4).
To a stirred solution of the above alcohol (18.0 g, 55 mM)
in CH2Cl2 (200 mL) was added Ph3P (64 mM) and CBr4
(64 mM) at room temperature. An instantaneous reaction
took place, and the resulting mixture was kept stirring for 3
h. TLC indicated no starting material remained. The mixture
was concentrated under reduced pressure to give crude
product, which was further purified by SiO2 column to give
nearly pure bromide (14.0 g).
To a preheated solution of the bromide obtained above
(14.0 g, 35.8 mM) in phenylsilane (14.0 g) was added AIBN
(150 mg), and the resulting mixture was heated at 110-120
°C for 0.5 h. TLC (AcOEt/n-hexane ) 1/6) indicated that
one-half of the starting material remained unchanged. Further
amount of AIBN (150 mg ×2) was added, and the mixture
was heated at 130-140 °C for 0.5 h interval. TLC showed
almost no starting material. The mixture was subjected to
evaporation to give yellow residue, which was further
purified by SiO2 column (AcOEt/n-hexane ) 1/7) to give
pure material (4) (11.3 g).
Reaction of MCA with EGP. (A) By Sc(OTf)3. Solids
of Sc(OTf)3 (500 mg) and SiO2-MgSO4 (2:1; 5 g) were
heated at 130-140 °C for 1 h. To this fine powder was added
MCA (TCI, 25.0 g, 175.3 mM) and neat EGP (10.0 g) at
around 80 °C. The mixture was then heated 110-120 °C
for ∼3 h. TLC then indicated the formation of desired
product as a less polar spot. Heating was then ceased, and
the mixture was diluted with CH2Cl2 before the addition of
SiO2 (10 g). The solid mass was filtered off and washed
with CH2Cl2. The filtrate was washed with saturated aqueous
NaHCO3 and concentrated to give crude products. The
mixture was then purified by SiO2 column (AcOEt/n-hexane
) 1/7) to give two less polar fractions, from which the
desired dimer 2 was isolated as a colorless oil (15.2 g,
1
47.0%). 2: H NMR (CDCl3) δ 1.25 (3H, t, J ) 7 Hz), 3.77
(6H, s), 4.22 (2H, q, J ) 7 Hz), 5.64 (1H, s), 6.76 (2H, dd,
J ) 8.7, 2.7 Hz), 6.97 (4H, m). 13C NMR (CDCl3) δ 171.9;
159.3; 135.0; 130.1; 127.5; 115.1; 112.8; 61.4; 55.4; 50.1;
14.1. MS (EI) m/z 368 (M+), 370.
(B) By Bi(OTf)3. A solid mass of Bi(OTf)3 (500 mg) with
SiO2-MgSO4 (1:1; 3 g) was heated at 130 °C for 2 h. To
this fine powder was added MCA (25.0 g) and neat EGP
(10.0 g) at around 80 °C. The mixture was then heated 130
°C for ∼2 h. Heating was ceased, and the mixture was diluted
with CH2Cl2. The solid mass was filtered off and washed
with CH2Cl2. The filtrate was concentrated to give crude
yellow products. The mixture was purified by SiO2 column
to give two less polar fractions, among which the desired
dimer 2 was obtained as a colorless oil (14.5 g, 44.8%).
Reaction of MCP with Acetal. To a stirred solution of
triphenylbismuth (8.0 g) in CH2Cl2 (40 mL) was added at
room temperature triflic acid (5.0 mL) dropwise under
cooling. The resulting mixture was kept stirring at room
temperature for 0.5 h, before the addition of SiO2 (50 g)
and CH2Cl2 (40 mL). To this suspension was added MCP
(25.0 g) and acetal (25.0 mL) under adequate stirring, and
To a stirred solution of 4 (3.44 g) in CH2Cl2 (30 mL)
was added at room temperature a CH2Cl2 solution of BBr3
(30 mL, 1 M solution). The resulting mixture was kept
stirring at room temperature overnight until TLC indicated
no starting material remained. The mixture was carefully
quenched by the addition of H2O and extracted with CH2-
Cl2. The organic layer was washed with H2O and brine and
concentrated after drying. Evaporation of the solvent gave
nearly pure material (3) as a colorless oil (2.77 g) identical
Vol. 5, No. 1, 2001 / Organic Process Research & Development
•
87

with the product obtained from the reaction on MCP and
acetal.
us useful information on bismuth chemistry. Thanks are also
extended to Mr. K. Tsuji in our laboratory for his technical
assistance. Finally, we thank Mr. K. Tada for his skillful
MS measurement.
Acknowledgment
We thank Professors M. Nishizawa and T. Sugihara of
Tokushima Bunri University for their generous gift of Hg-
(OTf)2 and helpful commentary. We are also grateful to
Professor M. Wada of Tokushima University for providing
Received for review June 27, 2000.
OP0000733
88
•
Vol. 5, No. 1, 2001 / Organic Process Research & Development