Reactivity of bismuth(III) halides towards alcohols. A tentative to mechanistic investigation

Article abstract of DOI:10.1016/S0040-4020(01)00013-8

The reactivity of bismuth(III) halides (BiX₃; X=Cl, Br and I) towards a series of alcohols has been investigated. Three different reactions have been studied, namely: halogenation, dehydration and etherification. The behaviour of these bismuth derivatives was found to depend on the nature of the halide bonded to the bismuth atom. Their reactivities can be interpreted on the basis of the Hard and Soft Acids and Bases (HSAB) principle. A mechanism is proposed which involves the formation of a complex of the alcohol with Bi(III).

Full text of DOI:10.1016/S0040-4020(01)00013-8

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 1909±1916
Reactivity of bismuth(III) halides towards alcohols. A tentative to
mechanistic investigation
El Mehdi Keramane, Bernard Boyerp and Jean-Pierre Roque
Â
Laboratoire de Chimie Organique Physique, Universite Montpellier II, Place E. Bataillon, 34095 Montpellier Cedex 5, France
Received 3 July 2000; revised 20 December 2000; accepted 21 December 2000
AbstractÐThe reactivity of bismuth(III) halides (BiX3; XˆCl, Br and I) towards a series of alcohols has been investigated. Three different
reactions have been studied, namely: halogenation, dehydration and etheri®cation. The behaviour of these bismuth derivatives was found to
depend on the nature of the halide bonded to the bismuth atom. Their reactivities can be interpreted on the basis of the Hard and Soft Acids
and Bases (HSAB) principle. A mechanism is proposed which involves the formation of a complex of the alcohol with Bi(III). q 2001
Elsevier Science Ltd. All rights reserved.
1. Introduction
In order to understand better the role played by BiCl3 in this
reaction and to suggest a mechanism, we have completed
our investigations of this study by extending it towards a
large series of aliphatic, cyclic, benzylic and allylic
alcohols.4 We have used CCl4 as solvent to facilitate the
1H NMR analysis of the reaction medium. Reactions were
GC monitored and products were identi®ed by comparison
with authentic samples.
Over the past few years an increasing number of papers have
described the use of bismuth(III) salts as catalysts.1 A few
recent reports deal with the catalytic activity of Lewis acids
derived from Bi(III).1 This new interest to bismuth salts is
understandable. Bismuth and its derivatives are regarded as
the least toxic of the heavy elements2 and they have been
widely used in traditional medicine.3
The present paper describes a mechanistic study of this
chlorination reaction.
In our continuing effort to develop bismuth salts as reagent
for organic reactions, our laboratory recently reported the
use of bismuth(III) halides in reactions such as the chlori-
nation of alcohols4 with BiCl3, a convenient halogen
exchange reaction5 and an ef®cient benzylation reaction of
alcohols.6
The results concerning primary, secondary and tertiary
alcohols are presented in Table 1 and those for benzylic
and allylic alcohols in Table 2.
Examination of Tables 1 and 2 showed that:
We present here a study of the activity of bismuth(III)
halides for the halogenation of alcohols.
1. Primary alcohols did not react with BiCl3. They were
recovered unchanged after three days at re¯ux. This
result, that we had already noted earlier,4 agrees with
that previously reported in the chlorination of alcohols
by the Me3SiCl/BiCl3 system.7
2. Secondary alcohols gave alkenes (30%) together with the
chlorides (70%) as proved by a constant alkene/chloride
ratio during the reaction. At room temperature the reac-
tion was slower but the alkene/chloride ratio remained
unchanged.
3. Roughly the same ratios chloride and alkene were
obtained from pentan-3-ol, hexan-2-ol, octan-2-ol and
octan-3-ol. However, in these cases, several isomeric
chlorides and alkenes were formed which would suggest
a reaction through a carbocation pathway.
2. Results and discussion
2.1. Reactivity of BiCl3
Initially, we started this study by comparing the reactivity of
BiCl3 in the chlorination of various primary, secondary,
tertiary and benzylic alcohols. In the presence of BiCl3,
aliphatic alcohols yielded a mixture of the corresponding
chlorides (70%) and alkenes (30%) while benzylic alcohols
gave a mixture of chlorides (30%) and symmetrical ethers
(70%).4
4. Tertiary alcohols appear to be the most reactive
substrates towards BiCl3 yielding exclusively the corre-
sponding chlorides.
Keywords: bismuth; alcohols; chlorination; dehydration; etheri®cation.
Corresponding author. Tel.: 133-4-67-14-38-38; fax: 133-4-67-14-38-
88; e-mail: bboyer@univ-montp2.fr
p
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)00013-8

1910
E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
Table 1. Chlorination of primary, secondary and tertiary alcohols by BiCl3 (1 equiv.) in CCl4
Alcohols (1 equiv.)
T (8C)
Reaction
times
Conversion
(%)
Products
Hexan-1-ol
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
Re¯ux
3 days
3 days
3 days
3 days
3 days
1 h
1 h 30
45 min
2 h
0
0
0
0
0
100
100
100
100
100
100
±
±
±
±
±
2-Ethylhexan-1-ol
2-Methylhexan-1-ol
Octan-1-ol
2-Methylpropan-1-ol
Cyclohexanol
Cyclopentanol
Butan-2-ol
Pentan-3-ol
Hexan-2-ol
Octan-2-ol
70% Chlorocyclohexane130% cyclohexene
70% Chlorocyclopentane130% cyclopentene
70% 2-Chlorobutane130% but-2-ene
41% 3-Chloropentane129% 2-chloropentane130% pent-2-ene
46.5% 2-Chlorohexane126% 3-chlorohexane127.5% hex-2-ene
41% 2-Chlorooctane119% 3-chlorooctane15.5% 4-chlorooctane112.5% oct-2-
ene112.5% oct-3-ene16% oct-4-ene13.5% oct-1-ene
14.5% 2-Chlorooctane128% 3-chlorooctane19.5% 4-chlorooctane113% 2-
octene115% 3-octene113% 4-octene17% 1-octene
100% 2-Chloro-2-methylpropane
1 h 30 min
12 h
Octan-3-ol
Re¯ux
12 h
100
2-Methylpropan-2-ol
2-Methylbutan-2-ol
1-Methylcyclohexanol
258C
258C
258C
5 min
5 min
5 min
100
100
100
100% 2-Chloro-2-methylbutane
100% 1-Chloro-1-methylcyclohexane
Table 2. Chlorination of benzylic and allylic alcohols by BiCl3 (1 equiv.) in CCl4
Alcohols (1 equiv.)
T (8C)
Reaction
times
Conversion
(%)
Products
Benzyl alcohol
1-Phenylethanol
2-Phenylpropany2-ol
Prop-2-en-1-ol
But-2-en-1-ol
258C
Re¯ux
258C
Re¯ux
258C
Re¯ux
258C
Re¯ux
258C
30 min
5 min
30 min
5 min
30 min
5 min
7 h
100
100
100
100
100
30% Benzyl chloride170% benzyl ether
100% Benzyl chloride
30% 1-Chloro-1-phenylethane170% bis(1-phenylethyl)ether
100% 1-Chloro-1-phenylethane
30% 2-Chloro-2-phenylpropane170% bis(2-phenylpropyl)ether
100% 2-Chloro-2-phenylpropane
70% 1-Chloroprop-2-ene130% bis(prop-2-enyl)ether
100% 1-Chloroprop-2-ene
60% 1-Chlorobut-2-ene
15 min
1 h
25% 2-Chlorobut-3-ene
15% (But-2-enyl)(1-methylprop-2-enyl)ether
70% 1-Chlorobut-2-ene130% 2-chlorobut-3-ene
60% 1-chlorobut-2-ene120% 2-chlorobut-3-ene
20% (But-2-enyl)(1-methylprop-2-enyl)ether
75% 1-Chlorobut-2-ene125% 2-chlorobut-3-ene
Re¯ux
258C
5 min
1 h 30
100
100
But-3-en-2-ol
Re¯ux
5 min
100
5. With benzylic alcohols, and in agreement with previous
work,4 at room temperature, the corresponding chlorides
were formed in 30% yield only; the main product was the
symmetrical ether. We have observed that ethers and
chlorides were obtained concurrently. However, when
the reaction is conducted in re¯uxing CCl4, the ether
was not formed and the chloride was obtained with
excellent yields (97%).
(but-2-enol) predominantly led to crotyl chloride (1-
chloro-but-2-ene) together with the same unsymmetrical
ether containing both crotylic and angelic groups.
We then have studied the chlorination of a series of cyclic
alcohols possessing a well-de®ned con®guration. Results
are reported in Table 3.
6. As benzylic alcohols, allylic alcohols have led at room
temperature to a mixture of chloride and ether, the
chloride being predominant in that case. At re¯ux, only
the chloride was obtained. Also we have observed that
allylic alcohols are subject to an allylic rearrangement
induced by BiCl3, to give mixtures of isomers. Thus
both angelic alcohol (but-3-en-2-ol) and crotyl alcohol
With secondary alcohols (Tables 1 and 3), the chlorination
reaction seemed to be superseded by the formation of
alkenes. As seen with other secondary alcohols, the
temperature has no in¯uence on the nature and the propor-
tion of the products.
To evaluate the in¯uence of the BiCl3 concentration, some
Table 3. Chlorination of cyclic secondary alcohols by BiCl3 (1 equiv.) in CCl4 at re¯ux
Alcohols (1 equiv.)
Reaction
times
Conversion
(%)
Products
cis-2-methylcyclohexanol
trans-2-methylcyclohexanol
15 min
1 h
100
100
32% 1-Chloro-1-methylcyclohexane168% 1-methylcyclohex-1-ene
30.3% Trans-1-chloro-2-methylcyclohexane16% cis-1-chloro-2-
methylcyclohexane18.9% 1-chloro-1-methylcyclohexane142.4% 1-
methylcyclohex-1-ene16.3% 1-methylcyclohex-2-ene16.2% cyclohexylidene
35% Menthyl chloride165% menthenes
Menthol
4 h
100

E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
Table 4. Chlorination of alcohols by BiCl3 (1/3 equiv.) in CCl4.
1911
Alcohols (1 equiv.)
Octan-2-ol
T (8C)
Reaction
times
Conversion
(%)
Products
Re¯ux
45 h
100
40% 2-Chlorooctane119% 3-chlorooctane16% 4-
chlorooctane113% oct-2-ene112.5% oct-3-ene16% oct-4-
ene13.5% oct-1-ene
cis-2-Methylcyclohexanol
2-Methylpropan-2-ol
1-Phenylethanol
Re¯ux
258C
258C
1 h
2 h 30
1 h
100
100
100
33% 1-Chloro-1-methylcyclohexane167% 1-methylcyclohexene
100% 2-Chloro-2-methylpropane
30% 1-Chloro-1-phenylethane170% bis (1-diphenylethyl)ether
Scheme 1.
Scheme 2.
alcohols were treated with a third of one equivalent of BiCl3
to give the results reported in Table 4.
If such a mechanism occurs, it would imply: (a) the forma-
tion of a carbocation, (b) the liberation of HCl, (c) the
formation of oxychloride.
The use of smaller quantities of BiCl3 did affect neither the
nature nor the ratio of the products but increases the reaction
times. However, 2-methylpropan-2-ol was completely
converted into 2-chloro-2-methylpropane showing that the
three chlorine atoms of BiCl3 are usable in the chlorination
reaction.
Also the formation of ethers or alkenes would result from a
reaction involving BiCl3.
We have successively examined each of these points.
(i) To verify the existence of a carbocation intermediate, we
have compared reaction times of the benzyl alcohol and the
4-methylbenzyl alcohol. If a carbocation was formed, the
presence of the methyl group in para position should
enhance the reaction rate. In fact, when the reaction was
performed in the standard conditions (1 equiv. of BiCl3 in
CCl4 at 258C), the benzyl alcohol disappeared faster than the
4-methyl derivative, reaction times being 30 and 55 min,
respectively. Those results are inconsistent with the forma-
tion of a carbocation.
From all these results, we have considered two possible
mechanisms.
The ®rst mechanism involves the formation of alkoxy-
bismuthane intermediates and proceeds via a carbocation
derived from the alcohol.
The second mechanism assumes the alcohol activation by
means of complexes between the oxygen of the alcohol and
the bismuth atom.
Moreover, if there was a carbocation, cis- and trans-2-
methylcyclohexanol (Table 3) would lead to the same
carbocation and, consequently should then give the same
product mixtures which is not the case.
2.1.1. Mechanism A. The mechanism (Scheme 1) was
inspired from that proposed in the alcohols chlorination
by boron trichloride.8
The action of BiCl3 on the alcohol would lead to formation
of the dichloroalkoxybismuthane with liberation of HCl.
This intermediate would then give bismuth oxychloride
and an ion pair. The isomeric chlorides obtained in the
reactions would arise from a carbocation rearrangement.
Ether formation would result from the reaction between
the carbocation and unreacted alcohol molecules
(Scheme 2).
Additionally, Table 2 showed a rearrangement of allylic
alcohols. Thus at re¯ux, both but-2-en-1-ol and but-3-en-
2-ol led roughly to the same chlorobutene mixture (70/30
Finally, the deprotonation of the carbocation would produce
alkenes and hydrogen chloride (Scheme 3).
Scheme 3.

1912
E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
and 75/25, respectively). This result could suggest the
passage through two carbocations in a fast equilibrium.
However, if this was the case, the secondary chloride
corresponding to the more stable carbocation should be
expected. The experiment showed the opposite: 1-chloro-
but-2-ene was roughly three times more abundant than
2-chlorobut-3-ene. On the contrary, a slow equilibrium
between both carbocations should not give the same mixture
of chlorides.
According to mechanism A, this solid should not be able to
chlorinate an alcohol. This solid was recovered, washed
with CCl4 and reacted with another equivalent of 1-phenyl-
ethanol. This reaction led to the corresponding chloride and
ether in the same proportion as shown previously. This was
not compatible with mechanism A.
In conclusion, mechanism A had to be discarded.
2.1.2. Mechanism B. This mechanism suggests the forma-
tion of a complex between the alcohol and BiCl3 followed
by different routes leading either to the chloride, the ether or
the alkene.
All those experimental observations should preclude the
formation of a carbocation as reaction intermediate.
(ii) This mechanism involves the formation of HCl. We
were never able to detect HCl. Thus, 2-methylpropan-2-ol
(tBuOH), was reacted at 258C in CCl4 with 1 equiv. of BiCl3
and 1 equiv. of 2,3-dimethylbut-2-ene to trap HCl. The reac-
tion rate was strongly reduced (30 min instead of 5 min) but
no 2-chloro-2, 3-dimethylbutane traces could be detected.
At this stage, we have found no evidence for the production
of HCl.
2.1.2.1. Etheri®cation. Proposals described above (see
iv) indicate that this reaction does not consume chlorine
of BiCl3. Then, this last would play only a catalytic role.
To con®rm this catalytic role played by BiCl3 in the
benzylic alcohols etheri®cation, 1-phenylethanol was
reacted with a catalytic amount of BiCl3 (5% mol equiv.).
The reaction proceeds to completion in 10 days leading to
1,1-diphenylethylether (87%) and 1-chloro-1-phenylethane
(13%) (Scheme 4).
(iii) According to this mechanism, bismuth oxychloride
(OvBi±Cl) should be formed at the end of the reaction.
When reactions were performed in the presence of 1 equiv.
of BiOCl in CCl4 at re¯ux, the alcohols were recovered
unchanged. Thus BiOCl is totally inactive towards alcohols.
To verify the formation of BiOCl, the chlorination of
tBuOH by BiCl3 (1 mol equiv.) was performed in CCl4 at
room temperature. Upon completion of the reaction, a solid
was recovered, washed with CCl4 and reacted with another
equivalent of tBuOH. This gave the corresponding chloride
proving thus that the recovered solid was not BiOCl.
The alcohol was completely consumed showing that BiCl3
acts as catalyst in the ether formation. Its reaction
mechanism is shown in Scheme 5.
The studies of Lewis acidity of Bi(III) compounds,
especially if the bismuth bears electronegative atoms, has
led to the identi®cation of numerous complexes9 formed
with amines, alcohols, amides, esters, carbonyl group,¼
Such structures reported in those complexes lend additional
support to our mechanistic proposal.
Also we have observed (Table 4) that tBuOH (1 equiv.) was
totally converted into tBuCl by 1/3 equiv. of BiCl3. Thus
the stoichiometry 3ROH/1BiCl3 was observed which is
inconsistent with mechanism A.
The interaction between the lone electron pair of the alcohol
oxygen and the electronic de®ciency of bismuth involve the
elongation and the weakening of the C±O bond turning the
hydroxy-bearing carbon into an electrophilic centre. The
oxygen atom of a second alcohol molecule acting as nucleo-
phile attacks this positively carbon atom affording thus to
the ether formation with loss of water and that, without
direct BiCl3 consumption.
(iv) Thus, in this mechanism the ®rst steps are common for
the formation of chlorides, ethers or alkenes. Whatever the
product, BiCl3 would be consumed. The reaction of 3 equiv.
of 1-phenylethanol with 1 equiv. of BiCl3 gave chloride
(30%), ether (70%) and a solid.
Scheme 4.
Scheme 5.

E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
1913
Scheme 6.
Scheme 7.
2.1.2.2. Elimination. When the elimination occurs, for
butan-2-ol mainly in a SNi mechanism (retention of con®g-
uration), which agrees with the four-centre reaction
mechanism proposed above and, therefore, with the
mechanism B.
example with the cyclohexanol, it would proceed through
a similar catalytic process to that postulated for the
etheri®cation reaction according to Scheme 6.
2.1.2.3. Chlorination. The alcohol/BiCl3 complex may
evolve according to a third way. The carbon atom bearing
the hydroxyl group can be attacked by one of chlorine atom
of BiCl3. In the bismuth atom an exchange between chlorine
atom and the hydroxyl group occurs involving a four-centre
transition state already previously mentioned in the halogen
exchange reactions.5 The expected chloride was formed
with a new species of bismuth, `BiCl2OH'. This reaction
would occur according to an SNi mechanism (Scheme 7).
We have now to explain the formation of different isomer
chlorides during the chlorination reactions which results
were already reported in preceding tables.
The formation of these different chlorides could be
explained in considering the geometry of the BiCl3±ROH
complex in which several parameters should be taken in
consideration, namely:
the location of positively charges-bearing carbons;
the position of chlorine atoms of BiCl3 compared to these
electrophilic centres;
steric hindrance susceptible to interfere with the chlorine
approach;
the size of bismuth atom.
Since one mole of BiCl3 can convert 3 mol of tBuOH into
the corresponding chloride, we should therefore admit that
the bismuth could exchange successively its chlorine atoms
according to an analog process.
In summary, BiCl3 plays a catalytic role in etheri®cation and
elimination reactions and acts as a reagent in the chlorina-
tion reaction.
In considering these different parameters, we can partly
explain the regioselectivity of the chlorination and thus
the formation of several positional isomers. For example,
the hexan-2-ol leads to the 2- and 3-chlorohexane in the
ratio 2:1 roughly (see Table 1). Let us consider the hexan-
2-ol/BiCl3 complex (Scheme 8).
Finally, another proof of the validity of mechanism B has
been given by the stereochemical study of the chlorinating
reaction of optically alcohols [(S)-(2)-1-phenylethanol and
(1)-butan-2-ol] in the presence of BiCl3 in CCl4 at re¯ux;
chlorides were isolated and their optical rotations measured
(Table 5).
We have attempted to explain our experimental results from
those criteria.
These ®ndings will suggest that in the chlorination
reactions, BiCl3 attacked the 1-phenylethanol and the
The Bi±O interaction creates a positive charge on the C2.
This charge can be spread partially toward the C3. Since the
Table 5. Chloriantion of optically alcohols by BiCl3 (1 equiv.) in CCl4 at re¯ux
b
10,11
[a]D
ROH
[a]D
RCla
[a]D
(S)-(1)-butan-2-ol
(S)-(2)-1-phenylethanol
1128 cˆ5; MeOH
2458 cˆ10; MeOH
(S)-(1)-2-chlorobutane
(S)-(2)-1-chloro-1-phenylethane
124.98 cˆ15.6; MeCN
2458 cˆ5.4; EtOH
124.78 cˆ15.6; MeCN
2458 cˆ5.4; EtOH
a
Chlorides isolated from the reaction mixture.
Optical rotations were measured at 208C.
b

1914
E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
It is worthwhile to mention that in the case of aliphatic
secondary alcohols, the chlorination never concerns the
primary carbon atom (see Table 1) whatever the position
of the hydroxyl group.
Finally, the mechanism B would explain part of experimen-
tal results namely:
the 3 chlorine atoms of BiCl3 are available and exchange-
able to achieve the chlorination reaction.
This reaction occurs according to a SNi mechanism and
depends of the steric hindrance of the bismuth±alcohol
complex.
Scheme 8.
Following these results, we supplemented our investigations
in studying the reactivity of two other bismuth(III) halides
(BiBr3 and BiI3) towards alcohols.
2.2. Reactivity of BiBr3 and BiI3
1-phenylethanol was reacted with 1 equiv. of BiBr3 in CCl4
at 258C. Within 15 min the alcohol was totally consumed,
giving bis-(1-phenylethyl)ether in 100% yield. BiBr3 does
not seem to be a brominating reagent. To verify this point,
this work has been extended to a series of alcohols (primary,
secondary, tertiary, benzylic and allylic). Results are
reported in Table 6.
Scheme 9.
BiCl3 size and its position with respect to the alcohol,
chlorine atoms of BiCl3 are close to these two attack centres
(C2 and C3) and chloride mixture was ensued via a four- or
a ®ve-centre transition state, respectively. The attack on the
C2, more positively polarized, is predominant. On the other
hand, menthol afforded only one chloride which could be
explained with the help of the menthol/BiCl3 complex
(Scheme 9) involving a four-centre intermediate.
From these results, we can make the following remarks:
Like with BiCl3, primary alcohols did not react in the
presence of BiBr3.
Secondary and tertiary alcohols afforded exclusively the
corresponding alkenes. Even tertiary alcohols which are
converted into chlorides in 100% yield by BiCl3, react
with BiBr3 to lead quantitatively the corresponding
alkenes.
Benzylic and allylic alcohols led to the corresponding
ethers without bromide formation, with reaction times
shorter than those obtained with BiCl3. As previously
noted, but-2-en-1-ol and but-3-en-2-ol gave mixture of
symmetrical and unsymmetrical ethers.
From the menthol structure, BiCl3 is relatively bulky and
there is a strong steric hindrance with the isopropyl group.
Thus BiCl3 can only adopt only one well-precise position
with respect to the ring. So, BiCl3 is located under the ring
and far from it. Thus, the chlorine approach could happen
only on the C1, which should be the less-hindered side.
Since BiCl3 is situated far from the ring, the chlorination
reaction was disfavoured leading thus to a higher alkene
formation.
BiBr3 is not a reagent for the bromination of alcohols and it
does not give its bromine atoms.
Also the observed transposition with allylic alcohols (Table
2) can result from the chlorine atom transfer involving a six-
centre (Scheme 10) transition state. Thus, it appears that the
size of the bismuth atom can permit the passage through a
four-, ®ve- or six-centre transition state.
Those results can be explained on the base of the Hard and
Soft Acids and Bases concept12 (HSAB principle): soft acids
prefer to bond to soft bases. The Bi31 ion is considered as a
Scheme 10.

E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
Table 6. Reactivity of BiBr3 (1 equiv.) towards alcohols in CCl4
1915
Alcohols (1 equiv.)
Reaction
times
Conversion
(%)
Products
Hexan-1-ola
7 days
7 days
1 h
1 h 30
2 h
5 min
5 min
5 min
30 min
20 min
15 min
4 h
0
0
Octan-1-ola
Cyclohexanola
100
100
100
100
100
100
100
100
100
100
100
100% Cyclohexene
100% 1-Methylcyclohex-1-ene
100% Pent-2-ene
100% 2-Methylprop-1-ene
100% 2-Methylbut-2-ene
100% 1-Methylcyclohex-1-ene
100% Dibenzylether
cis-2-Methylcyclohexan-1-ola
Pentan-3-ola
2-Methylpropan-2-olb
2-Methylbutan-2-olb
1-Methylcyclohexan-1-olb
Benzyl alcoholb
1-Phenylethanolb
2-Phenylpropan-2-ola
Prop-2-en-1-ola
100% Bis-(1-phenylethyl)ether
100% Bis-(2-phenylprop-2-yl)ether
100% Bis-(prop-2-enyl)ether
58.4% (But-2-enyl)(1-methylprop-2-enyl)ether 111.6% bis-(1-methylprop-2-
enyl)ether 130% bis-(but-2-enyl)ether
63.4% (But-2-enyl)(1-methylprop-2-enyl)ether 118.3% bis-(1-methylprop-2-
enyl)ether 118.3% bis-(but-2-enyl)ether
But-2-en-1-ola
30 min
But-3-en-2-olb
30 min
100
a
Reaction performed at 608C.
Reaction achieved at 258C.
b
soft (or eventually borderline) acid and should prefer to
bond to soft bases like I2 or Br2 rather than with a hard
base. It is worthwhile pointing out that Cl2 and OH2 are two
hard bases approximatively equivalent, which would
explain why BiCl3 can behave as chlorinating agent. In
other words, BiCl3 is able to exchange chlorine atom against
hydroxyl group, while BiBr3 does not exchange bromine
with hydroxyl group.
We tried to extend to BiI3 experiments achieved with BiBr3.
However whatever the considered alcohol, BiI3 was not able
to convert any of them. Alcohols were recovered unchanged
after reaction in the presence of BiI3. Two factors can
explain this result:
According to the HSAB principle. I2 being a soft base
and OH2 a hard base, BiI3 is not able to play the role of
reagent (iodination reaction), which is in agreement with
BiBr3 behaviour.
Furthermore, the bulky iodine atoms around the bismuth
metal lead to a compact structure,13 which prevent any
alcohol molecule approach and the Bi±OH interaction
creation. Thus, BiI3 can act neither as reactant nor as
catalyst.
Thus, from this HSAB principle, BiBr3 would be unchanged
at the end of the reaction with alcohols and would act only
as an ef®cient catalyst for the elimination and etheri®cation
reactions.
To verify that, 1-phenylethanol was reacted with 0.05 equiv.
of BiBr3 in CCl4 at 258C. The alcohol was totally converted
to ether after 8 h. At the end of this reaction, the solid was
recovered, washed with CCl4 and reacted with one another
equivalent of 1-phenylethanol. The corresponding ether was
again obtained after 8 h. Therefore BiBr3 is really a catalyst
for the etheri®cation reaction of allylic and benzylic
alcohols.
3. Conclusions
In conclusion, this study concerning the bismuth(III)
halides reactivity towards alcohols has shown the following
points:
Experimental results are better explained by the
postulated mechanism B which implies the preliminary
formation of a complex between the alcohol and BiX3.
BiCl3 has two behaviour patterns. It plays the role of
reagent for the chlorination of alcohols and intervenes
as catalyst for both the etheri®cation and the elimination
reactions.
The catalytic role played by BiBr3 in the elimination
reaction was also con®rmed by examining the reaction of
cyclohexanol with 0.05 equiv. of BiBr3 in CCl4 at re¯ux.
The alcohol was completely converted to cyclohexene
within 24 h. Therefore BiBr3 is also a catalyst for the elimi-
nation reaction.
Otherwise we have observed that under indentical experi-
mental conditions in the presence of BiBr3, benzyl alcohol
give the corresponding ether faster than the 4-methylbenzyl
alcohol (40 and 65 min, respectively). Such a behaviour is
not consistent with mechanism involving a carbocation as
intermediate.
The chlorination by BiCl3 would involve
a
SNi
process, which agrees with the already proposed5 four-
centre reaction mechanism. Due to the bismuth atom large
size and the steric hindrance of the BiCl3±alcohol complex,
chlorination can also involve a ®ve- or six-centre transition
state.
From these results it appears that BiBr3 proceeds through
similar catalytic processes to that postulated for the same
reactions described above for the etheri®cation and elimina-
tion reactions catalysed by BiCl3.
The steric hindrance affects strongly the reaction: a large
steric interference leads to a low chloride formation and
consequently to a large alkene or ether amount.

1916
E. M. Keramane et al. / Tetrahedron 57 (2001) 1909±1916
London, 1987; Vol. 3, pp. 292. (b) Manhart, M. D. Rev. Infect.
BiBr3 behaves only as an ef®cient catalyst for both the
etheri®cation and the elimination reactions.
BiI3 is totally inactive towards alcohols.
Dis. 1990, 12, 311. (c) Kozak, R. W.; Waldmann, T. A.;
Atcher, R. W.; Gansow, O. A. Trends Biotechnol. 1985, 4,
259. (d) Baxter, G. F. Chem. Ber. 1992, 28, 445.
Â
4. Boyer, B.; Keramane, E. M.; Montero, J. L.; Roque, J. P.
Synth. Commun. 1998, 28, 1737.
5. Boyer, B.; Keramane, E. M.; Arpin, S.; Montero, J. L.; Roque,
Furthermore the difference in the halogenation ability
between BiCl3 and BiBr3 or BiI3 is in agreement with
HSAB principle.
Â
J. P. Tetrahedron 1999, 55, 1971.
6. Boyer, B.; Keramane, E. M.; Roque, J. P.; Pavia, A. A.
Tetrahedron Lett. 2000, 41, 2891.
Further studies on the application of Bi(III) derivatives to
others reactions are in progress.
Á
7. Labrouillere, M.; Le Roux, C.; Oussard, A.; Gaspard-
Iloughmane, H.; Dubac, J. Bull. Soc. Chim. Fr. 1995, 132, 522.
8. (a) Hudson, H. R.; Kinghorn, R. R. F.; Murphy, W. S. J. Chem.
Soc. 1971, 3593. (b) Gerrard, W.; Hudson, H. R.; Murphy,
W. S. J. Chem. Soc. 1964, 2314. (c) Gerrard, W.; Lappert,
M. F. J. Chem. Soc. 1951, 1020. (d) Gerrard, W.; Lappert,
M. F. J. Chem. Soc. 1955, 3084.
4. Experimental
1H NMR was recorded in a BRUCKER ACE-250 instru-
ment at 250 MHz in CDCL3 solutions. GC analysis was
performed with a DELSI 330 apparatus. Optical rotations
were measured using a PERKIN ELMER 241 polarimeter
with speci®c rotations determined at 208C. All starting
materials, alcohols, optically active alcohols, BiCl3, BiBr3,
BiI3, BiOCl, were commercially available (Aldrich).
9. (a) Lange, K. C. H; Klapotke, T. M. In The Chemistry of
È
Arsenic, Antimony and Bismuth Compounds; Pataõ, S., Ed.;
Wiley: New York, 1994; pp 315. (b) Majafuji, T.; Mutoh,
T.; Satoh, K.; Tsunenari, K. Organometallics 1995, 14,
3848. (c) Carmalt, C. J.; Clegg, W.; Elsegood, M. R. J.;
Errington, R. J.; Havelock, J.; Lightfoot, P.; Norman, N. C.;
Scott, A. J. Inorg. Chem. 1996, 35, 3709. (d) Di Vaira, M.;
Mani, F.; Stoppioni, P. Eur. J. Inorg. Chem. 1999, 833.
(e) Kallal, K.; Coin, C.; Dunach, E.; Postel, M. New J.
Chem. 1997, 21, 495 (and references cited therein).
(f) Genge, A. R. J.; Levason, W.; Reid, G. Chem. Commun.
1998, 2159. (g) Murray, B.; Hvoslef, J.; Hope, H.; Power, P. P.
Inorg. Chem. 1983, 22, 3421. (h) Althaus, H.; Breunig, H. J.;
Rosler, R.; Lork, E. Organometallics 1999, 18, 328.
Whatever the bismuth derivative (BiCl3, BiBr3, BiI3,
BiOCl), halogenation reactions were performed following
the general procedure previously described.4 The same
method was used for the reaction achieved in the presence
of the 2,3-dimethylbut-2-ene.
References
10. Holcomb, H. L.; Nakanishi, S.; Flood, T. C. Organometallics
1996, 15, 4228.
1. (a) Suzuki, H.; Igekami, T.; Matano, Y. Synthesis 1997, 249.
Á
(b) Desmurs, J. R.; Labrouillere, M.; Le Roux, C.; Gaspard,
11. Moss, R. A.; Baicerzak, P. J. Am. Chem. Soc. 1992, 114, 9386.
12. (a) Pearson, R. G. J. Am. Chem. Soc. 1963, 85, 3533.
(b) Pearson, R. G. Science 1966, 151, 172. (c) Parr, R. G.;
Pearson, R. G. J. Am. Chem. Soc. 1983, 105, 7512.
(d) Pearson, R. G. J. Chem. Educ. 1987, 64, 561.
H.; Laporterie, A.; Dubac, J. Tetrahedron Lett. 1997, 38, 8871.
(c) Repichet, S.; Le Roux, C.; Dubac, J.; Desmurs, J. R. Eur. J.
Org. Chem. 1998, 2743. (d) Repichet, S.; Roux, C.;
Hernandez, P.; Dubac, J. J. Org. Chem. 1999, 64, 6479.
2. Irwing-Sax, N.; Bewis, R. J. In Dangerous Properties of
Industrial Materials; Van Nostrand Reinhold: New York,
1989; pp 283 (also pages 284, 522 and 523).
13. Fisher, G. A.; Norman, N. C. The Structures of the Group 15
Element (III) Halides and Halogeanions. In Advances in
inorganic chemistry; Academic Press: San Francisco, 1994;
Vol. 14, pp 336 (and references cited therein).
3. (a) Wilkinson, G., Gillard, R. D., McCleverty, J. A., Eds.;
Comprehensive Coordination Chemistry, Pergamon Press: