Some uses of mischmetall in organic synthesis

Article abstract of DOI:10.1016/j.tet.2003.07.017

Mischmetall, an alloy of the light lanthanides, has been used in a variety of organic reactions, either as a coreductant in samarium(II)-mediated reactions (Barbier and Grignard-type reactions, pinacolic coupling reactions) or as the promoter of Reformatsky-type reactions. It has been also employed as the starting material for easy syntheses of lanthanide trihalides, the reactivity of which has been explored in Imamoto and Luche-Fukuzawa reactions and in Mukaiyama aldol reactions.

Full text of DOI:10.1016/j.tet.2003.07.017

Tetrahedron 59 (2003) 10551–10565
Some uses of mischmetall in organic synthesis
*
´
Marie-Isabelle Lannou, Florence Helion and Jean-Louis Namy
´ ´ ´
Laboratoire de Catalyse Moleculaire, associe au CNRS, ICMO, Bat 420, Universite Paris-Sud, 91405, Orsay, France.
Received 21 May 2003; accepted 16 July 2003
Abstract—Mischmetall, an alloy of the light lanthanides, has been used in a variety of organic reactions, either as a coreductant in
samarium(II)-mediated reactions (Barbier and Grignard-type reactions, pinacolic coupling reactions) or as the promoter of Reformatsky-type
reactions. It has been also employed as the starting material for easy syntheses of lanthanide trihalides, the reactivity of which has been
explored in Imamoto and Luche–Fukuzawa reactions and in Mukaiyama aldol reactions.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
which is rapidly covered with an oxide layer. Here we
report some interesting uses of mischmetall in organic
synthesis.
The use of rare earth compounds in organic chemistry has
grown considerably during the past 20 years. Mainly
samarium, cerium, lanthanum, ytterbium, neodymium,
dysprosium, lutetium, scandium and yttrium metals and
derivatives have been studied. These elements clearly differ
in terms of reactivity. For instance, samarium(II) com-
pounds can be used as powerful reductants, whereas,
cerium(IV) ones are strong oxidants, some scandium and
ytterbium(III) derivatives are very efficient catalysts in a
variety of reactions and chemistry of lanthanum metal
shows special features. However, we thought that in many
reactions, a mixture of lanthanide could be used instead of
individual element without significant change in results.
This should reduce the cost of reactions since, separation of
the elements remains an expensive process. Mischmetall is
one of the main sorts of mixed rare earths. It is a well known
alloy produced by fused chloride electrolysis of the light
lanthanide elements. It comes from a lanthanide ore called
bastnaesite (a fluorocarbonate). It constitutes over 90% of
the rare earth metal consumed for steelmaking. The relative
concentration of the four light lanthanide elements and of
the ‘heavies’ can vary depending on the origin of the ores.
The usual mischmetall contains 48–50% Ce, 32–34% La,
13–14% Nd, 4–5% Pr and 1.5% other rare earth metals. It
is at least 10 times cheaper than cerium. Mischmetall is
commercially available as ingots which are easily
powered with a rasp (caution! see Section 4). After scraping,
it can be stored in a Schlenck tube under argon for several
weeks. This is impossible with most of the pure metal
powders (especially cerium and lanthanum), the surface of
2. Results and discussion
2.1. Mischmetall as a coreductant in samarium(II)
catalysed reactions
Since its introduction in synthetic organic chemistry in
1977,1 samarium diiodide has become one of the most
popular reagents. However, its cost could be considered as a
major drawback, potentially limiting the development of its
chemistry. For this reason several attempts have been
reported to run reactions with catalytic quantities of SmI2
and an in situ regeneration system for the Sm(II) species.2,3
Ideally, a system for in situ regeneration of SmI2 should be:
(1) cheaper than samarium metal, (2) able to quickly reduce
samarium(III) species into samarium(II) ones, which
implies the ability to cleave Sm(III)–O bonds, (3)
unreactive towards organic substrates and (4) simple and
usable in a variety of the organic reactions mediated by
SmI2. Due to their thermochemical properties (redox
properties and strength of the Ln–O bonds) and the low
reactivity of lanthanide metals (if they are not activated with
iodine or mercury salts),4 cerium, lanthanum, neodymium
or praseodymium fit all the above requirements, except for
the first one. Fortunately, the alloy of the light lanthanides
known as mischmetall (or cerium mixed metal) is available
at a low price. Here we refer to some previously published
results and we report on new ones concerning the use of
mischmetall as a coreductant in Sm(II) mediated reactions.
Considering the percentages of each element, the molecular
weight of mischmetall is taken as 140 g mol21. It is denoted
below as Ln.
Keywords: mischmetall; samarium; lanthanide; catalysis; organic reactions.
*
Corresponding author. Fax: þ33-1-69-15-4690;
e-mail: jelonamy@icmo.u-psud.fr
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.07.017

10552
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
2.1.1. Catalytic Barbier and Grignard-type allylation
and benzylation of ketones and aldehydes.5 Catalytic
reactions of allylic halides were tested with ketones and
aldehydes, either according to a Barbier-type Procedure
(Catalytic-BP) or a Grignard-type one (Catalytic-GP),
(Scheme 1).
the reaction should occur through formation of organome-
tallic species.
In the catalytic scheme I, an organosamarium compound,
RSmIX (or RSmX2 or RSmI2) is formed first. Its addition to
a carbonyl compound gives a samarium alkoxide. This in
turn, through a reduction–transmetallation process with Ln,
provides another lanthanide alkoxide, with the regeneration
of SmI2 (or SmX2). In the catalytic scheme II, the initially
formed RSmIX reacts with Ln to give another organo-
metallic compound (LnR3; Ln–Sm) through a reduction–
transmetallation process, with the regeneration of SmI2.
This organolanthanoid compound then adds to the carbonyl
compound to furnish an alkoxide. Alternatively, reduction
of the initially produced radicals (from a reaction
between organic halide and SmI2) with Ln instead of
Sm(II) could give an organometallic compound (R3Ln)
directly, SmI2X being concomitantly reduced with Ln into
SmI2. This catalytic scheme III (Fig. 1) should not require
formation of an organosamarium compound. However, it
can be seen as a variant of catalytic scheme II since it also
implies a reaction between R3Ln (Ln–Sm) and a carbonyl
compound.
Scheme 1.
In catalytic-BP, yields of tertiary alcohols were close to
those obtained under stoichiometric conditions.6 In
catalytic-GP, tertiary alcohols were obtained in moderate
(X¼I) to good yield (X¼Br). In contrast, the sequential
procedure cannot be used with samarium diiodide in
stoichiometric amount, since the dienes arising from a
Wurtz coupling are the only detected products.7
In catalytic-BP, all three mechanisms are possible, whereas,
in catalytic-GP the catalytic scheme I is not possible, as it is
known that organosamarium compounds are unstable. On
the other hand, the success of this procedure would agree
well with the catalytic schemes II or III which imply
formation of R3Ln, R2LnX or RLnX2 (Ln other than Sm).
We proposed three catalytic schemes (Fig. 1). In all cases
As the main components of mischmetall are cerium,
lanthanum and neodymium they were tested in both
procedures in place of mischmetall under similar experi-
mental conditions. It was found that cerium and lanthanum
were both efficient coreductants in these reactions, as well in
Catalytic-BP and Catalytic-GP. The results, obtained either
with Ce, La or Nd were close to those described with
mischmetall. However, the alloy is much cheaper than
cerium, lanthanum or neodymium, and moreover is less
prone to oxidation. All this is in favour of the use of
mischmetall instead of separated lanthanide metals.
Interestingly, a simpler experimental procedure starting
from a mixture of samarium metal and mischmetall could be
used as depicted below (Scheme 2), while, a sequential
procedure also gave excellent results. Thus, it had been
demonstrated that a stable organometallic compound was
also formed under these experimental conditions
(Scheme 3). In both cases the product was obtained in
excellent yield.
Presumably, in situ formation of samarium diiodide occurs
Figure 1.
Scheme 2.

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
10553
Scheme 3.
under these conditions. In addition, diiodoethane might
activate mischmetall, thus facilitating the catalytic process
(reduction–transmetallation of RSmIX—see catalytic
scheme II, or reduction of radicals—see catalytic scheme
III, Figure 1).
involving lactones or esters and alkyl halides yield the
expected diols in the presence of small amounts of nickel
diiodide (1 mol% with respect to SmI2).9 This experimental
process can be termed as NiI2-Catalysed Barbier
Procedure (BP/NiI2(cat.)). Therefore, the mischmetall/
0
[SmI2/NiI2(cat .) (cat.) system was used in a Catalytic Barbier
]
2.1.2. Catalytic Barbier and Grignard-type reactions of
esters and lactones. We first studied reactions of lactones
with activated halides (allyl iodide, allyl bromide and
benzyl bromide) mediated by the mischmetall/SmI2(cat.)
system using a Catalytic-BP or a Catalytic-GP and readily
observed the formation of diols. Some results are presented
in Table 1.8 Yields in diols were good or excellent.
Nevertheless, the yields were slightly better with the
Catalytic-BP than with the Catalytic-GP. In the absence of
samarium diiodide the starting materials were recovered
unchanged.
Procedure catalysed with NiI2 (Catalytic-BP/NiI2(cat.)) and
compared to the mischmetall/SmI2(cat.) system in a catalytic
Barbier procedure (Catalytic-BP). Some results are
presented in Table 2.8
We checked that both mischmetall alone and mischmetall in
the presence of nickel diiodide (without SmI2), were
unreactive in the reactions indicated above. The results
obtained with the five or six-membered-ring lactones
(entries 1, 2) were in good agreement with the expectations
since good yields in diols were obtained with nickel
diiodide, whereas, in the absence of this catalyst,
starting materials were recovered unchanged. Thus, this
‘double-stage catalysis’ is efficient in those cases. More
Next, we examined the reactions of less active alkyl halides.
In stoichiometric conditions, Barbier-type reactions
Table 1. Catalytic reactions of activated halides mediated by the mischmetall/SmI2 system
Entry
R1Br
Ester or lactone
Isolated yields (%)
Catalytic-BPa
Catalytic-GPa
1
2
3
4
5
6
7
8
All-I
Ethyl benzoate
Ethyl benzoate
Ethyl benzoate
Ethyl i-butyrate
Ethyl i-butyrate
Ethyl i-butyrate
b-Propiolactone
g-Butyrolactone
74
68
71
60
64
84
56
77
54
73
55
52
50
73
71
50
All-Br
Bn-Br
All-I
All-Br
Bn-Br
Bn-Br
Bn-Br
a
For experimental details, see Ref. 8.

10554
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Table 2. Catalytic Barbier-type reactions of lactones or esters with iodoheptane
Entry
Ester or lactone
Isolated yields (%)
a
Catalytic-BPa
Catalytic-BP/NiI2(cat.)
1
2
3
4
5
6
g-Butyrolactone
d-Valerolactone
1-Caprolactone
Ethylbenzoate
t-Butylacetate
Ethyl-i-butyrate
0
0
91
74
70
76
83
66
68
78
72
81
a
For experimental details, see Ref. 8.
Scheme 4.
unexpectedly, the diol was obtained with 1-caprolactone in
excellent yield in the absence of nickel diiodide (entry 3). It
was thus anticipated that esters should behave like this
unstrained lactone. This was confirmed: good yields in
tertiary alcohols were obtained in both procedures (entries
4–6). Thus, with 1-caprolactone and esters the catalytic
Barbier procedure does not require the presence of nickel
diiodide while the stoichiometric Barbier procedure
demands it. This shows that the mischmetall/SmI2(cat.)
system can be more reactive than stoichiometric samarium
diiodide. Some hypotheses are proposed in Ref. 8 to explain
these results.
coupling reactions using this reagent and mischmetall as a
coreductant.
Before studying catalytic pinacol coupling, we first tried to
find an easy and reliable method of preparation of samarium
dibromide. This compound can be prepared directly from
samarium metal and 1,2-dibromoethane in THF.12 How-
ever, the latter reaction usually needs a long induction time
and small amounts of diiodoethane have to be added to
ensure reproducible results. To overcome this drawback, we
attempted to use 1,1,2,2-tetrabromoethane instead of 1,2-
dibromoethane. Mixing samarium with 1 equiv. of C2H2Br4
in THF resulted in a fast exothermic reaction, evolution of a
gas, and disappearance of samarium metal. A cream-
coloured solid was formed which failed for the coupling
of 5-nonanone into diol. It was identified as being samarium
tribromide.13 With half an equivalent of 1,1,2,2-tetra-
bromoethane, a gas (acetylene) was also produced and a
dark blue solid, which was slightly soluble in THF, was
obtained. Its reaction with 5-nonanone furnished the expected
diol in 87% isolated yield. These properties are in agreement
with those of SmBr2. Scheme 5 sums up the results.
2.1.3. Pinacol coupling reactions. The pinacol coupling of
acetophenone can be readily performed using SmI2 as a
catalyst and mischmetall as a coreductant (Scheme 4). In
contrast to the previously reported procedure, which uses
magnesium as a coreductant, there is no need to add
trimethylsilyl chloride to cleave the Sm(III)–O bonds.3 The
pinacol coupling product is obtained in good isolated yield.
The dl/meso ratio (80:20) is the same as that measured when
the reaction is performed under stoichiometric conditions
(79:21).10 This result is in good agreement with a
pinacolisation mediated by SmI2 leading to a samarium(III)
pinacolate, the reduction of which by mischmetall
regenerates SmI2. Without samarium diiodide, no reaction
occurs.
This first result was interesting. However, the method is
limited to aromatic ketones.11 Samarium dibromide is a
more powerful reagent in pinacol coupling.12 This led us to
examine the possibility to perform catalytic pinacol
Scheme 5.

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Table 3. Catalytic pinacolizations using SmBr2 and mischmetall
10555
Entry
Carbonyl compounds
x/y/z
Conditions
Isolated yields (%)a
dl/meso
1
2
3
4
5
6
7
8
9
10
Cyclobutanone
Cyclohexanone
Cyclooctanone
Benzaldehyde
Pivalaldehyde
2-Octanone
3-Octanone
5-Nonanone
Acetophenone
1,3-Dibenzoylpropaneb
5:1.4:0.7
5:1.4:0.7
5:1.4:0.7
5:1.4:0.7
5:1.4:0.7
5:1.4:0.7
5:0.7:1.4
5:0.7:1.4
5:0.7:1.4
5:0.7:1.4
Slow
Slow
Slow
Slow
Slow
Slow
Fast
Fast
Fast
Fast
1: 66
55
69
62
57:43
100:0
50:50
50:50
2: 63
72
54
68
72
80:20
3: 70c
a
For experimental details, see Section 4.
Carbonyl compound: 2 mmol.
cis 1,2-Diphenylcyclopentan-1,2-diol.
b
c
In SmI2 catalytic Barbier-type reaction, we had observed
that a simple experimental procedure, which starts from a
mixture of samarium metal and mischmetall treated by 1,2-
diiodoethane, gave excellent results (see above). Here we
applied this methodology, changing the diiodide for the
tetrabromide. Some results are shown in Table 3.
2.1.4. Reformatsky-type reactions. Since the pioneering
works of Kagan et al. in the field of samarium diiodide
mediated reactions,7 many reports have demonstrated the
crucial role of this reagent in Reformatsky-type reactions, as
well in intermolecular reactions using chiral auxiliaries14
than in intramolecular ring-closure reactions.15,16 Never-
theless, no reaction using samarium diiodide in catalytic
amounts has been reported so far.
This procedure involving a mixture of Sm and mischmetall
has allowed us to isolate pinacols in moderate to good
yields. In some cases a slow addition of the carbonyl
compound was necessary in order to allow regeneration of
Sm(II) species thus preventing formation of by-products.
The quantities of each reagent (mischmetall, samarium,
tetrabromoethane) have been optimised according to the
procedure (fast or slow addition of the carbonyl com-
pounds). We observed in some cases that an excess of
tetrabromoethane, with respect to samarium, improved the
yields in pinacol (entries 7–10), probably in facilitating the
regeneration of samarium dibromide (more bromide ions
are available). In the absence of samarium, the diols were
not obtained, either the carbonyl compound was recovered
or a complex mixture of products was formed. Without
tetrabromoethane, no reaction took place.
Since our previous work using catalytic amounts of
samarium diiodide in Barbier and Grignard-type reactions
were very promising,5 Reformatsky-type reactions have
been performed using the mischmetal/SmI2 catalytic-
system. Reactants were added simultaneously following a
Barbier-type procedure. As expected, our system is an
efficient promoter for Reformatsky-type reactions using a
Barbier-type procedure. The expected b-hydroxyesters
were obtained in medium to good yields, the results are
presented in Table 4.
The electrophile/a-bromoester ratio had to be established
first. Thus, when a 1:1 ratio was used, yields were very low
due to side-reactions, mainly debromination and auto-
condensation reactions. With ethyl 2-bromobutyrate and
tert-butyl 2-bromoacetate, it was found that a 1:2 ratio
allows the formation of the b-hydroxyesters with satis-
factory yields. Besides, 3 equiv. of methyl 2-bromoacetate
(with respect to the ketone) were required for the reaction
to proceed with acceptable yields since it is more
susceptible to develop side-reactions than other
a-bromoesters.17 It is known that a-bromoesters are
subject to side-reactions such as reduction and auto-
condensation. They can react with tetrahydrofuran
either by a radical pathway or via an anionic pathway
(Scheme 7) to give the reduced product A, or the
autocondensation product B by reaction with another
a-bromoester (Scheme 8).18 Methyl 2-bromoacetate and
tert-butyl 2-bromoacetate are subject to debromination
reaction (Scheme 7) rather than autocondensation, whereas,
for ethyl 2-bromobutyrate, the autocondensation product
The following catalytic scheme can be proposed. In contrast
to the previously reported SmI2-catalytic pinacolization3
there is no need to use TMSCl together with a coreductant
since cerium and lanthanum (the main components of
mischmetall) are both able to cleave the Sm–O bond and to
reduce Sm(III) species into Sm(II) ones, (Scheme 6).
1
has been observed and fully characterised (IR, H and 13C
NMR).
Scheme 6.

10556
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Table 4. Reformatsky-type reactions mediated by the mischmetall/SmI2(cat.) system
Entry
Electrophile
a-Bromoester
Ratioa
Isolated yields (%)b
1
2
3
4
5
6
7
8
4-Phenyl-butan-2-one
4-Phenyl-butan-2-one
4-tert-Butyl-cyclohexanone
4-tert-Butyl-cyclohexanone
Acetophenone
Acetophenone
Cyclohexanone
Cyclohexanone
2-Indanone
Methyl 2-bromoacetate
tert-Butyl 2-bromoacetate
Methyl 2-bromoacetate
tert-Butyl 2-bromoacetate
Methyl 2-bromoacetate
tert-Butyl 2-bromoacetate
Ethyl 2-bromobutyrate
Ethyl 2-bromobutyrate
Ethyl 2-bromobutyrate
Methyl 2-bromoacetate
Methyl 2-bromoacetate
1:2.9
1:2
1:2.9
1:2
1:2.9
1:2
1:2
1:2
1:2
74
77
74c
51d
54
82
92
53e
4: 54
0
9
10
11
Benzaldehyde
Octanal
1:2.9
1:2.9
0
a
Electrophile/a-bromoester.
For experimental details, see Section 4.
b
c
d
e
cis/trans¼60:40.
cis/trans¼80:20.
Fast addition of the reactants.
Scheme 7.
Scheme 8.
When using aldehydes as electrophiles (entries 10 and 11),
complex mixtures are obtained. Thus, only trace amounts of
1
desired products are detected by H NMR, cyclotrimerisa-
tion of aldehydes being faster than the formation of the
b-hydroxyester.
Scheme 9.
Reformatsky-type reactions also seemed to be an interesting
target for our system since it has been previously observed
that stable organolanthanides species have been obtained
with allylic and benzylic bromides (see above). In a similar
way, we tried to obtain stable enolates. The reaction was
carried out in two stages: addition of the a-bromoester to the
mischmetall/SmI2(cat.) system, followed by quenching with
an electrophile. Many reactions have been performed
between ethyl 2-bromobutyrate and cyclohexanone
(Scheme 9), unfortunately, always leading either to the
starting material (temperatures lower than 2508C) or to
b-ketoester at higher temperatures since the autocondensa-
tion reaction is strongly predominant.
The formation of the autocondensation product has already
been studied in the reaction of ethyl 2-bromoacetate with
stoichiometric samarium diiodide by Utimoto et al.,18 who
have proved that an unreactive intermediate is produced at a
temperature above 2508C.
2.2. Reformatsky-type reaction mediated by activated
mischmetall
More than one hundred years after its discovery, the
Reformatsky-type reaction is still widely studied. In the last
decades it has been described to be promoted by several

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Table 5. Some additives of mischmetall in Reformatsky-type reactions
10557
Entry
Additive
None
Diiodoethane
Diiodoethane
Diiodoethane
Dibromoethane
Iodine
Isolated yield (%)a
1
2
3
4
5
6
0
65b
40c
65
53
79
a
For experimental details, see Section 4.
Reactants were added dropwise in 3 h; mischmetall/diiodoethane¼1:1.
b
c
Idem; ketone/a-bromoester¼1:1.
activated or non-activated metals among these: zinc,19
indium,20 germanium21 or cerium.22
enolates cannot be obtained and the reaction does not work
with aldehydes as electrophiles due to their ability to form
cyclotrimers in the presence of mischmetall. On the other
hand, activated mischmetall is also a good promoter for
Reformatsky-type reactions. Moreover, activation by
organic halides is a very simple procedure. The reaction is
less sensitive to air than with samarium diiodide and yields
are good in most of the cases even if slightly lower than with
samarium diiodide. However, dropwise addition of the
reactants is not necessary in that case. As described above,
the reaction does not work when aldehydes are used as
electrophiles.
Pioneering works of Imamoto et al. in this field22 have
revealed the role of the activation of cerium metal in the
Reformatsky-type reaction. Cerium was activated either by
amalgamation or by catalytic amounts of mercury(II)
chloride. Some attempts have been previously reported
concerning the use of mischmetall in Grignard and in
Reformatsky reactions.23 We then decided to investigate
further the reactivity of mischmetall in the Reformatsky-
type reaction. At first, activated mischmetall was tested in
the reaction between ethyl 2-bromo-butyrate and cyclo-
hexanone. The results are presented in Table 5.
To conclude, it seems obvious that there is no need to work
with cerium metal activated by mercury22 since mischmetall
seems to show a high reactivity in Reformatsky-type
reactions and the use of iodine or diiodoethane is much
more convenient than amalgamation.
Mischmetall alone is not capable of promoting the
Reformatsky-type reaction (reactants are fully recovered),
but the addition of catalytic amounts of iodine, dibromo-
ethane or diiodoethane to the alloy allows the formation of
the desired b-hydroxyesters. The best result was obtained
with iodine, but results were more reproducible with
2.3. Preparations and uses of anhydrous mischmetall
halides
diiodoethane,
for
that
reason,
the
system
mischmetall/diiodoethane(cat.) has been further investigated
and the results are listed in Table 6.
2.3.1. Preparations of anhydrous lanthanide trihalides.
Classical methods for the preparation of anhydrous rare
earth trihalides (LnX3, X¼Cl or Br) involve reaction of
lanthanide oxides with hydroacids in water, followed by
drying with NH4Cl, or trimethyl orthoformate or thionyl
chloride. Some years ago, Deacon et al.24 reported the
preparation of anhydrous lanthanide trichlorides from metal
As a conclusion, mischmetall is an efficient coreductant for
samarium diiodide catalysed Reformatsky-type reactions.
Nevertheless, the process requires a dropwise addition of the
reactants. The limitations of this method are that stable
Table 6. Use of mischmetall in Reformatsky-type reactions
Entry
Electrophile
a-Bromoester
Isolated yields (%)a
1
2
3
4
5
6
4-Phenyl-butan-2-one
4-Phenyl-butan-2-one
4-tert-Butylcyclohexanone
4-tert-Butylcyclohexanone
Acetophenone
Methyl 2-bromoacetate
tert-Butyl 2-bromoacetate
Methyl 2-bromoacetate
tert-Butyl 2-bromoacetate
Methyl 2-bromoacetate
tert-Butyl 2-bromoacetate
68
72
67b
61c
36
Acetophenone
46
a
For experimental details, see Section 4.
cis/trans¼35:65.
b
c
cis/trans¼65:35.

10558
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Scheme 10.
Table 7. Preparation of lanthanide trichlorides
Entry
Ln
x
(Coordinated thf)
Isolated yields
(%)
Colour
1
2
3
4
5
Nd
Pr
Sm
Gd
2.4
2.6
2.5
2.6
1.8
83
82
74
67
95
Pale blue
Dark beige
Cream-coloured
White
Mischmetall
White
and hexachloroethane in THF under ultrasonic activation
(Scheme 10).
tetrabromoethane instead of 1,2-dibromoethane, the results
are presented in Table 8.
We have performed a similar reaction with mischmetall
powder. In this case, ultrasonic activation was not required.
Considering this result we wondered whether pure metals
could furnish anhydrous lanthanide chlorides in a similar
fashion provided powder was freshly prepared by scraping
of ingots. We tested samarium, neodymium, praseodymium
and gadolinium metals and were pleased to observe the
expected reaction with hexachloroethane. The results are
presented in Table 7.
Yields in trihalides are good and this method gives a new
and simple access to anhydrous lanthanide tribromides. An
excess of tetrabromoethane was employed and we observed
both the formation of a mixture of (Z) and (E)-1,2-
dibromoethene and evolution of a gas which was identified
as being acetylene by trapping with aqueous Cu(NH3)2Cl
(brick red cuprous acetylide was formed). Though heating is
not absolutely necessary, it was used to make sure that all
the metal powder is consumed. It is worth noting that the
alloy is a good starting material for the preparation of
tribromides since the yield is good and the starting product
is cheap. The numbers of thf molecules coordinated to the
metals are larger with lanthanide tribromides than with
lanthanide trichlorides.
Titrations were performed as previously described.13 Yields
in complexes were good to excellent, and the number of thf
molecules coordinated to the metal was in accordance with
the results found by Deacon et al.
Lanthanoid tribromides are generally synthesized from
lanthanide oxides and hydrobromic acid in water, then
subsequently dried with NH4Cl. The main drawback is their
high tendency to undergo hydrolysis. Previous work showed
that anhydrous samarium dibromide can be prepared from
dibromoethane and samarium metal in THF.12 Thus, it has
been assumed that trivalent lanthanide bromides could be
prepared in the same manner. However, the reaction usually
needs a long induction time and small amounts of
diiodoethane must be added to ensure reproducible results.
To overcome this drawback, we attempted to use 1,1,2,2-
A few preparations of lanthanide triiodides have been
reported either requiring large excesses of reagents or
drastic conditions.25,26 Only samarium and ytterbium
triiodides have been easily prepared in two steps from the
metal.13 This procedure is possible due to their stability in
the divalent state. We found a very simple way to prepare
trivalent lanthanide iodides starting from freshly scraped
metals and diiodoethane in one step. This procedure seems
applicable to all lanthanide metals. It is similar to the one
reported for samarium diiodide. The use of diiodoethane is a
good choice since the formation of ethylene gas displaces
the reaction towards the complex formation. The results are
presented in Table 9.
Table 8. Preparation of lanthanide tribromides
The reaction is very vigorous. As a consequence, there is no
need for heating the reaction mixture and yields of
lanthanide triiodides are good to excellent. Lanthanum
triiodide, that is normally prepared in 1,2-dimethoxyethane
from lanthanum metal and iodine at reflux for 20 h,26b can
be prepared at room temperature in THF (entry 1). It comes
out that the number of thf molecules coordinated on the
metal increases along with the size of the anion.
Entry
Ln
x (Coordinated thf) Isolated yields
(%)
Colour
1
2
3
4
5
Nd
Pr
Sm
Gd
4
3.5
4
4
2.6
83
85
77
86
89
Beige
Beige
Pale green
Grey
Brown
As a conclusion, anhydrous lanthanide trihalides have been
prepared using a simple and efficient method in one step
Mischmetall

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
10559
Table 9. Preparation of lanthanide triiodides
stoichiometry. It is then necessary to use a 1:1 stoichiometry
for LiAlH4/LnCl3. Moreover, it was reported that on
addition of LiAlH4, a vigorous gas evolution is observed.
In our case, no gas evolution occurs. It seems that
commercially available lanthanide trichlorides hepta-
hydrates cannot be totally dehydrated even after drying
under drastic conditions. 1 or 2 mol equiv. of LiAlH4 are
consumed by the remaining water molecules coordinated to
the metal. Evans et al. have previously observed the
presence of water in cerium trichloride heptahydrate dried
by standard procedures such as high vacuum
(1023 mm Hg): at least one or two molecules of water
remain coordinated to the metal.28
Entry
Ln
x
(Coordinated thf)
Isolated yields
(%)
Colour
1
2
3
4
5
6
La
Nd
Pr
Sm
Gd
4.4
4
4.3
4.2
4
66
88
91
84
89
79
Beige
Yellow
Pale beige
Yellow–green
Beige
Mischmetall
4.3
Beige
from several lanthanide metals. It is worth noting that the
preparations of trihalides proceed in a satisfactory way with
mischmetall.
Anhydrous lanthanide trichlorides are very efficient
promoters for extended Luche-type reaction. Expected
allylic alcohols are obtained in good to excellent yields,
with a complete 1,2-selectivity and the method using
anhydrous lanthanide trichlorides and LiAlH4 is less
restrictive than typical Luche reactions since a,b-unsatu-
rated carbonyl compounds such as aldehydes, esters and
carboxylic acids can be reduced. Moreover, the anhydrous
lanthanide trichloride prepared from mischmetall shows a
very good reactivity and selectivity in this reaction.
Then our aim was to investigate the efficiency of trihalides
prepared from the alloy in order to compare their reactivity
with those of complexes prepared from pure metals. In the
first part, reactions involving stoichiometric amounts of
lanthanide trihalides such as Luche–Fukuzawa and
Imamoto-type reactions were performed and in the second
part, reactions involving catalytic amounts of anhydrous
lanthanide trihalides in Mukaiyama aldol-type reactions
were investigated.
As a conclusion, there is no need to use a trichloride
synthesized from a pure lanthanide metal since the
efficiency of complexes prepared from the alloy is very
good. Moreover, the reactivity of commercial lanthanides
trichlorides is different from that one of our complexes,
since lanthanide trichlorides prepared from lanthanide
oxides cannot be anhydrous even after drying under
drastic conditions. In contrast, our complexes are
anhydrous, cheap and their preparation can be carried out
easily in one step.
2.3.2. Anhydrous mischmetall trichloride in Luche–
Fukuzawa-type reductions of a,b-unsaturated carbonyl
compounds. Fukuzawa et al. have described an extended
version of the Luche reaction where the use of ‘anhydrous’
cerium trichloride along with LiAlH4 allows the selective
reduction of a,b-unsaturated carbonyl compounds in the
1,2-positions leading to allylic alcohols.27 We have
performed the same reactions using anhydrous lanthanide
trichlorides prepared from mischmetall. The results are
given in Table 10.
2.3.3. Anhydrous mischmetall trihalides in Imamoto-
type reactions. The chemistry of organocerium(III)
reagents has been previously reported by Imamoto et al.22,29
Their reactivity towards a,b-unsaturated aldehydes or
ketones shows a strong carbonylophilic character, leading
to the selective 1,2-addition on the carbonyl group.
Organocerium(III) reagents are formed by reaction of
alkyl lithium reagents with lanthanides trihalides.
As expected, a,b-unsaturated carbonyl compounds are
easily reduced, but many differences have been observed
between the reactivity of the anhydrous LnCl3(thf)1.8
(Ln¼mischmetall) and lanthanide trichlorides used by
Fukuzawa. Indeed, a 2:1 stoichiometry (LiAlH4/CeCl3)
(with respect to the carbonyl compound) is used by
Fukuzawa to reduce selectively an unsaturated ketone in
the 1,2 positions. In our case, fully reduced product was
observed along with the desired one when using Fukuzawa
Some reactions have been performed in order to compare
the selectivity of anhydrous mischmetall trihalide-promoted
addition of alkyl lithium reagents to a,b-unsaturated
carbonyl compounds with one of the commercially
available cerium trichloride. The results are presented in
Table 11.
Table 10. Mischmetall trichloride in some Luche–Fukuzawa-type
reductions
Anhydrous mischmetall trihalides (X¼Cl, Br, I) exhibit a
good reactivity and an excellent diastereoselectivity in 1,2-
addition of alkyl lithium reagents to a,b-unsaturated
carbonyl compounds. Yields of expected allylic alcohols
are excellent except for the reaction of b-ionone with
n-BuLi in the presence of LnI3 where a complex mixture is
formed (entry 10). In contrast to CeCl3 prepared according
to the classical procedure which remain hydrated, with the
anhydrous mischmetall trichloride, the 1:1 stoichiometry
(LnCl3/n-BuLi) is convenient to allow the formation of the
expected allylic alcohol since n-butyl lithium is directly
Entry
R1
Ph
Ph
Ph
Ph
R2
R3
Me
R4
Isolated yields (%)
1
2
3
4
3
H
H
H
H
Me
Me
H
H
H
H
94
94
68
71
84
H
OH
OC2H5
H

10560
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
widely developed. Many Lewis acids have been employed
either in stoichiometric or in catalytic amounts to carry out
these reactions. Especially, Kobayashi et al. have exten-
sively studied the use of rare earth triflates as catalysts.30
The first Mukaiyama aldol-type reaction catalysed by
lanthanides derivatives was described by Kagan et al.31
with the presence of 5 mol% of several lanthanide
trichlorides. The reactions were rather slow (12 h to several
days) and yielded either O-silylated or desilylated aldols. A
few years later, Collin et al. showed the efficiency of
lanthanide iodides as catalysts in Mukaiyama aldol-type
reaction.32 The reactions are very fast (5 min at 2788C) and
yields are excellent.
Table 11. Mischmetall trihalides in some Imamoto-type reactions
Entry
X
Carbonyl compound
Acetophenone
t (8C)
Isolated yields (%)
1
2
Cl
Cl
278
278
90
93
3
4
Br
I
278
265
97
98
We have studied the reactivity of anhydrous lanthanide
tribromides. Their reactivity was assumed to be inter-
mediate between that of lanthanide trichlorides and that of
lanthanide triiodides. Some reactions have been carried out
between aromatic aldehydes bearing either an electron
withdrawing-group or an electron-donating group in various
positions and ketene trimethylsilyl methyl acetal (KSA)
using 10 mol% (with respect to the aldehyde) of anhydrous
lanthanide tribromides prepared from mischmetall. The
results are presented in Table 12.
5
6
7
8
9
Cl
Br
I
278
278
265
278
278
5: 92
96
89
The reactions proceed smoothly at room temperature, and
reaction times differ according to the group which is on the
aromatic ring. For electron-withdrawing groups, the reac-
tion can require up to 20 h, whereas, for electron-donating
groups, the reaction is very fast. The expected products were
obtained in good yields. In the case of m-bromobenzalde-
hyde, a side-product coming from an Oppenauer-type
oxidation of the aldol has been observed. It is noteworthy
that reactions performed with anhydrous tribromides
prepared from individual metals (samarium, gadolinium,
praseodymium and neodymium), led to b-hydroxy esters in
slightly lower yields than with the catalyst prepared from
mischmetall. The results are presented in Table 13.
Cl
Br
6: 95
85
10
11
12
I
265
278
278
–
Br
Br
Benzaldehyde
98
93
Mischmetall and neodymium tribromides have been also
tested as catalysts in Mukaiyama aldol-type reactions
involving an silyl enol ether and aromatic aldehydes. The
results are presented in Table 14.
reacting to form the organocerium(III) reagent while with
CeCl3·H2O at least 1 equiv. of RLi is consumed by water
(Scheme 11).
The reactions with o-methoxybenzaldehyde and b-naphtal-
dehyde are efficiently catalyzed by the tribromides, but with
p-nitobenzaldehyde the reactions are slow and the silyl enol
ether is gradually changed into acetophenone. Finally, we
have tested mischmetall triiodide under the conditions used
by Collin et al. (2788C, 5 min) for samarium iodides.32 The
results are presented in Table 15.
Scheme 11.
As a conclusion, all anhydrous lanthanide trihalides
prepared are very efficient promoters for the 1,2-addition
of organolanthanide(III) reagents to a,b-unsaturated
carbonyl compounds. Moreover, there is no need to use
lanthanide trihalides prepared from pure metals since the
alloy leads to the formation of very active complexes.
Under these conditions, moderate (entries 3, 4) to excellent
(entries 1, 2) yields were obtained.
As a conclusion, this is the first study concerning the
reactivity of anhydrous lanthanide tribromides as
catalysts. Their reactivity in Mukaiyama aldol-type reaction
has been shown to be very good. Moreover, catalysts
prepared from mischmetall were found to be more efficient
than catalysts prepared from pure metals. This feature is
very interesting since the alloy gives access to low-cost
catalysts.
2.3.4. Mukaiyama aldol-type reactions catalysed by
anhydrous lanthanide tribromides and triiodides.
Mukaiyama aldol-type reactions as well as Michael
reactions are fundamental tools for the formation of
carbon–carbon bonds in organic chemistry and have been

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Table 12. Mischmetall tribromide in some Mukaiyama aldol-type reactions
10561
Entry
R
KSA
(equiv.)
Reaction time
(h)
Y
Isolated yields
(%)a
1
2
3
4
5
6
7
8
9
10
p-OMe
p-OMe
p-OMe
o-OMe
H
p-NO2
p-Cl
p-Br
1.1
1.4
1.4
1.1
1.4
1.1
1.1
1.4
1.1
1.4
2
2
2
1
6.5
4
20
3
18
3
TMS
TMS
TMS
TMS
H
H
H
H
H
78
77
86b
76
54
83
60
34c
68
82
p-CF3
a-Naphthaldehyde
H
a
In some cases (entries 5–10) it was easier to purify desilylated products (desilylation with HCl 1 M in THF).
With 20 mol% of LnBr3.
A by-product (19%) arising from an Oppenauer oxidation of the aldol was also obtained.
b
c
Table 13. Use of lanthanide tribromides in Mukaiyama aldol-type reactions
Entry
R
Ln
Reaction time (h)
Isolated yield (%)
1
2
3
4
5
6
7
8
p-OMe
p-OMe
p-OMe
p-OMe
p-OMe
p-CF3
p-CF3
p-CF3
p-CF3
p-CF3
p-NO2
p-NO2
p-NO2
p-NO2
p-NO2
Nd
Pr
Sm
Gd
Mischmetall
Nd
Pr
Sm
Gd
Mischmetall
Nd
Pr
Sm
1
1
1
1
1
24
24
24
24
18
24
24
24
24
4
67
63
63
63
76
55
40
40
33
68
67
42
54
50
83
9
10
11
12
13
14
15
Gd
Mischmetall
Table 14. Use of lanthanide tribromides (mischmetall and neodymium) in Mukaiyama aldol-type reactions
Entry
Aldehyde
Ln
Reaction time (h)
Isolated yields (%)
1
2
3
4
5
o-Methoxybenzaldehyde
o-Methoxybenzaldehyde
p-Nitrobenzaldehyde
p-Nitrobenzaldehyde
b-Naphtaldehyde
Mischmetall
Nd
Mischmetall
Nd
Mischmetall
3
3
48
48
24
94
82
44
20
69
3. Conclusion
it is the most widely applicable coreductant known so
far. Moreover, the system mischmetall/SmI2(cat.) can
sometimes be more efficient than stoichiometric
samarium diiodide. The formation of stable organo-
metallic species have been observed with allylic and
benzylic halides at room temperature without any
additives.
Mischmetall is a very interesting starting material in the
field of lanthanide chemistry. Many of its advantages have
been highlighted.
(1) Mischmetall has been used as a coreductant for
samarium diiodide and, to the best of our knowledge,
(2) In Reformatsky-type reaction, mischmetall can be used

10562
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
Table 15. Use of mischmetall triiodide in Mukaiyama aldol-type reactions
Entry
R
Isolated yields (%)
1
2
3
4
p-OMe
o-OMe
p-NO2
H
86
90a
64
54
a
With 10 mol% of LnI3.
either as the coreductant of samarium diiodide or as the
promoter of the reaction with catalytic amounts of
diiodoethane. The system is then more convenient than
the Reformatsky-type reaction described by Imamoto
et al. mediated by cerium with catalytic amounts of
mercury(II) chloride.
performed at 70 eV. High Resolution Mass Spectra
(HRMS) and Electrospray Mass Spectra were performed
on a GC/MS Finningan-MAT-95-S. Flash chromatography
was performed on silica gel (Merck 230–240 mesh;
0.0040–0.0630 mm). All commercially available organic
compounds were distilled before use. Lanthanide metals
(ingots) were purchased from the Acros Company and
mischmetall (cerium mixed metal) from Fluka (about $60
per 500 g). Mischmetall ingots (about 5 g) were easily
powered in air with a rasp (caution! mischmetall powder is
pyrophoric, sparks can come out from ingot when it is
scraped, powder must be collected in a flask filled with
argon). The powder was kept under argon in a Schlenck
tube, (average molecular weight of mischmetall is 140).
Tetrahydrofuran (THF) was distilled under argon from
sodium benzophenone ketyl. Samarium diiodide was
prepared as previously described.7 All reactions were
carried out under argon in Schlenck tubes by standard
vacuum line techniques. Lanthanide trihalides were stored
and handle in a glove box under argon.
(3) A very simple method of metal activation allowed the
preparation of anhydrous trichlorides, tribromides and
triiodides from mischmetall and also from several pure
lanthanide metals. The reactivity of complexes pre-
pared from mischmetall has been further investigated
in reactions involving lanthanide halides in stoichio-
metric amounts (Luche and Imamoto-type reactions)
or in catalytic conditions (Mukaiyama aldol-type
reaction).
(4) In Luche–Fukuzawa and Imamoto-type reactions,
complexes prepared from mischmetall showed an
efficiency and selectivity for 1,2-addition to the
carbonyl group similar to that of complexes prepared
from pure lanthanide metals. Moreover, as they were
synthesized in anhydrous conditions, a difference has
been observed between their reactivity and those of
commercial complexes dried by classical methods.
Finally, the catalytic activity of anhydrous lanthanide
tribromides has been investigated, showing that
catalysts prepared from mischmetall are often more
efficient than those made from pure metals.
4.2. Catalytic pinacol coupling procedure
Fast addition. Mischmetall powder (0.7 g, 5 mmol) was
suspended in THF (2 mL), with Sm (0.105 g, 0.7 mmol) and
Br2CHCHBr2 (1.4 mmol) in a Schlenck tube under argon at
room temperature for 2 h. Ketone or aldehyde (4 mmol) in
THF (12 mL) was then added to the THF/SmBr2/
mischmetall suspension. The mixture was then stirred for
an additional period of 16 h, diluted with ether, quenched
with HCl (1 M) and stirred for 15 min to obtain a clear
solution, which was extracted with ether. The combined
extracts were washed with brine, sodium thiosulfate and
brine again. The organic layer was dried with MgSO4, and
the solvents were removed under reduced pressure. The
crude material was purified by flash chromatography on
silica gel.
It has been demonstrated that mischmetall can be used in
different ways in the field of organic chemistry but much
remains to be explored. Indeed, for a number of years, the
chemistry of lanthanide metals has been in constant
expansion. Since mischmetall powder can be easily
obtained and stored and is much more reactive and cheaper
than commercially available lanthanide metals powders, it
should become a very useful reagent in organic synthesis.
Slow addition. Mischmetall powder (0.7 g, 5 mmol) was
suspended in THF (2 mL), with Sm (0.21 g, 1.4 mmol) and
Br2CHCHBr2 (0.7 mmol) in a Schlenck tube under argon at
room temperature for 2 h. THF (5 mL) was then added.
Ketone or aldehyde (4 mmol) in THF (7 mL) was slowly
added to the THF/SmBr2/mischmetall suspension over 14 h.
The mixture was then stirred for an additional period of 2 h,
diluted with ether, quenched with HCl (1 M) and stirred for
15 min to obtain a clear solution, which was extracted with
ether. The combined extracts were washed with brine,
sodium thiosulfate and brine again. The organic layer was
dried over MgSO4, and the solvents were removed under
4. Experimental
4.1. General
1H and 13C NMR spectra were recorded at 250 and 63 MHz
respectively, with a Bruker AC 250 instrument (unless
otherwise stated). Chemical shifts are reported in part per
million (d) downfield from TMS. Infrared (IR) spectra were
recorded neat on a FTIR IFS 66 Bruker and are reported in
cm21. Mass spectra (MS) were determined on a GC/MS
Ribermag R10-10 instrument. Electronic impact was

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
10563
reduced pressure. The crude material was purified by flash
chromatography on silica gel.
Titrations were performed on the solids (LnX3,(thf)x) as
described in Ref. 13. Numbers of coordinated thf were
established on weighed aliquots from determinations of
halide and lanthanide contents.
4.3. Reformatsky-type reactions mediated by the
mischmetall/SmI2(cat.) system
4.8. Luche–Fukuzawa-type reductions of a,b-
unsaturated carbonyl compounds
Mischmetall powder (0.7 g, 5 mmol) was suspended in THF
(7 mL), with SmI2 (0.7 mmol) in a Schlenck tube under
argon at room temperature. A solution of an a-bromoester
(4 mmol) and ketone (2 mmol) in THF (7 mL) was slowly
added to the THF/SmI2/mischmetall suspension over 3 h.
The mixture was then stirred for an additional period of
0.5 h, diluted with ether, quenched with HCl (1 M) and
stirred for 15 min to obtain a clear solution, which was
extracted with ether. The combined extracts were washed
with brine, sodium thiosulfate and brine. The organic layer
was dried with MgSO4, and the solvents were removed
under reduced pressure. The crude material was purified by
flash chromatography on silica gel.
Anhydrous LnCl3,(thf)1.8 (Ln¼mischmetall) powder
(0.376 g, 1 mmol) and a carbonyl compound (1 mmol) in
THF (3 mL) were stirred in a Schlenck tube under argon at
room temperature for 10 min. A solution of LiAlH4
(0.5 mmol for ketones and aldehydes, 1 mmol for acids
and esters) in THF (5 mL) was then added. The mixture was
stirred for 15 min, quenched with HCl (0.1 M, until pH¼7)
and extracted with ether. The combined extracts were
washed with brine, dried with MgSO4 and the solvents were
removed under reduced pressure. The crude material was
purified by flash chromatography on silica gel.
4.4. Reformatsky-type reaction mediated by mischmetall
in the presence of diiodoethane
4.9. Imamoto-type reactions
Anhydrous mischmetall trihalide powder (1.3 mmol) in
THF (2 mL) was stirred in a Schlenck tube under argon at
room temperature for 1 h, then cooled at 2788C. A solution
of n-butylithium (1.3 mmol) in hexane (2 mL) was added
and stirring was maintained for an additional period of 0.5 h.
A carbonyl compound (1 mmol) in THF (5 mL) was then
added. The mixture was stirred for 3 h at 2788C, then
hydrolysed with saturated ammonium chloride and
extracted with CH2Cl2. The combined extracts were dried
with MgSO4 and the solvents were removed under reduced
pressure. The crude material was purified by flash
chromatography on silica gel.
A mixture of mischmetall powder (0.7 g, 5 mmol) and
diiodoethane (0.197 g, 0.7 mmol) was vigorously stirred in
a Schlenck tube under argon at room temperature for
30 min. THF (7 mL) was then added. After a short induction
period an exothermic reaction occurred. After cooling to
room temperature a solution of an a-bromoester (4 mmol)
and a ketone (2 mmol) in THF (7 mmol) was added. An
exothermic reaction occurred again. The mixture was then
stirred for an additional period of 15 min and treated as
described above. Similar procedures were used with I2 and
1,2-dibromoethane.
4.5. Preparation of anhydrous lanthanide trichlorides
(Ln5mischmetall, Nd, Pr, Sm, Gd)
4.10. Mukaiyama aldol reactions
Catalysed by anhydrous lanthanide tribromides. Anhydrous
lanthanide tribromide powder (0.1 mmol) in CH2Cl2
(10 mL) was stirred in a Schlenck tube under argon at
room temperature for 10 min. Reagents were then added,
first an aldehyde (1 mmol) then KSA (1.1 or 1.4 mmol).
Stirring was continued for 1–24 h, with GC monitoring.
The mixture was treated with hexane (10 mL). Lanthanide
salts precipitate was then removed by filtration. Solvents
were removed under reduced pressure. The crude material
(silylated products) was purified by flash chromatography
on silica gel.
Metal powder (10 mmol) was vigorously stirred in a
Schlenck tube under argon at room temperature for
30 min. A solution of hexachloroethane (4.26 g, 18 mmol)
in THF was slowly added over 30 min. An exothermic
reaction occurred. The mixture was stirred for an additional
period of 4 h at room temperature (to make sure that all the
metal powder is consumed, it can be refluxed for 12 h).
Hexane (10 mL) was added to the cooled mixture, which
was allowed to settle. The supernatant solution was
then removed. Precipitate (LnCl3,(thf)x) was washed three
times with hexane (3£25 mL). Remaining solvents were
removed under reduced pressure (1021 Torr) and
heating (608C). Products were kept and weighed in a
glove box.
Desilylation was performed with HCl 1 M (3 mL) in THF
(3 mL). The mixture was stirred for 2 h, then extracted with
dichloromethane. The combined extracts were dried with
MgSO4 and the solvents were removed under reduced
pressure. The crude material was purified by flash
chromatography on silica gel.
4.6. Preparation of anhydrous lanthanide tribromides
(Ln5mischmetall, Nd, Pr, Sm, Gd)
The same procedure was used.
Catalysed by anhydrous lanthanide triiodides. The same
procedure was employed except that 0.05 mmol of lantha-
nide salt was used and the reactions were performed at
2788C within 5 min.
4.7. Preparation of anhydrous lanthanide triiodides
(Ln5mischmetall, La, Nd, Pr, Sm, Gd)
The same procedure was used, except that the mixture was
simply stirred for 12 h at room temperature.
All the compounds have been previously prepared and
characterised. Some selected data are indicated below.

10564
M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
4.10.1. Bicyclobutyl-1,10-diol (1). Purification. Column
chromatography on silica gel (pentane/ethyl acetate 60:40)
white solid, (66% yield). 1H NMR (CDCl3) d (ppm): 2.14 (s,
2H, 2OH); 2.00 (m, 6H); 1.63 (m, 2H). 13C NMR (CDCl3) d
(ppm): 78.4; 30.5; 12.8. HRMS (m/z): calcd for C8H12O
(Mþ2H2O), 124.0888; found, 124.0885. Anal. calcd for
C8H14O2: C, 67.57; H, 9.92; O, 22.50. Found: C, 67.31; H,
9.91. FTIR (CaF2/CCl4) nmax: 3625; 3573; 2989; 2948;
2870; 2841; 1458; 1417; 1355; 1301; 1251; 1151; 1125;
1041.
m, C(7)–CH3). 13C NMR (CDCl3) d (ppm): 137.9; 133.3
(C(20)); 127.3; 119.2; 72.2 (C(3)); 53.2 (C(10)); 41.7 (C(40));
31.0 (C(600)); 30.7; 27.4; 26.7; 26.1; 25.5; 22.2 (C(5), C(6));
21.9 (C(2 )–CH3); 13.2 (C(7)). GC/MS m/z (% base peak):
232 (9); 217 (2); 194 (14); 193 (23); 176 (25); 161 (8); 151
(10); 147 (10); 137 (22); 133 (26); 123 (27); 119 (41); 109
(22); 105 (29); 93 (28); 91 (36); 79 (20); 71 (19); 55 (50); 43
(100); 41 (82). FTIR (NaCl/liquid film) nmax: 3401; 3025;
2961; 2929; 2853; 1732; 1456; 1377; 1363; 1298; 1255;
1129; 976; 936; 826; 794.
4.10.2. 2,2,5,5-Tetramethyl-hexane-3,4-diol (2). Purifi-
cation. Column chromatography on silica gel (pentane/ethyl
acetate 90:10) white solid, (63% yield). dl/meso 100:0. H
NMR (CDCl3) d (ppm): 3.33 (d, J¼6.7 Hz, 2H, C(3)H2,
C(4)H2); 2.34 (d, J¼6.7 Hz, 2H, 2OH); 0.90 (s, 18H, 9CH3).
13C NMR (CDCl3) d (ppm): 74.8 (C(3), C(4)); 35.2 (C(2),
C(5)); 25.8 (CH3). HRMS (m/z): calcd for C10H22O2 (Mþ),
174.1620; found, 174.1620. FTIR (CaF2/CCl4) nmax: 3642;
3543; 2961; 2909; 2870; 1477; 1465; 1394; 1365; 1240;
1181; 1088; 1060; 1009.
4.10.6. (E)-3-Methyl-1-(2,6,6-trimethylcyclohex-1-enyl)-
hept-1-en-3-ol (6). Purification. Column chromatography
on silica gel (heptane/ethyl acetate 90:10), colourless oil,
(95% yield). 1H NMR (CDCl3) d (ppm): 5.95 (1H, d, C(1)H,
J1–2¼16.1 Hz); 5.40 (1H, d, C(2)H, J2–1¼16.1 Hz); 1.90
1
(2H, t, C(30)H2, J3 –4 ¼6.6 Hz); 1.61–1.16 (11H, m,
C(3)OH, C(4)H2, C(5)H2, C(6)H2, C(40)H2, C(50)H2); 1.58
(3H, s, C(20)–CH3); 1.24 (3H, s, C(3)–CH3); 0.91 (6H, s,
C(60)–(CH3)2); 0.83 (3H, t, C(7)H3, J7–6¼7.1 Hz). 13C
NMR (CDCl3) d (ppm): 140.8 (C(2)); 137.2 (C(10)); 128.1
(C(20)); 124.8 (C(1)); 73.5 (C(3)); 42.7 (C(30)); 39.5 (C(4));
34.1 (C(60)); 32.8; 28.9 (C(60)(CH3)2); 28.5 (C(3)–CH3);
26.6; 23.3 (C(5)); 21.5 (C(20)–CH3); 19.4 (C(6)); 14.2
(C(7)). GC/MS m/z (% base peak): 233 (18); 232 (93); 217
(40); 203 (5); 189 (22); 175 (30); 161 (27); 147 (55); 133
(36); 121 (30); 119 (100); 107 (26); 105 (65); 95 (19); 93
(18); 91 (34); 81 (14); 77 (14); 69 (18); 55 (33); 43 (23); 41
(51). FTIR (NaCl/liquid film) nmax: 3401; 2953; 2929; 2860;
1714; 1653; 1458; 1377; 1360; 1260; 1124; 1041; 1030;
976; 908; 804.
0
0
4.10.3. cis-1,2-Diphenylcyclopentane-1,2-diol (3). Purifi-
cation. Column chromatography on silica gel (pentane/ethyl
acetate 80:20), white solid, (70% yield). dl/meso 0:100. 1H
NMR (CDCl3) d (ppm): 7.05 (m, 10H, 2Ph); 3.00 (s, 2H,
2OH); 2.58 (m, 2H); 2.45–2.00 (m, 4H). 13C NMR (CDCl3)
d (ppm): 142.4; 127.2; 126.9; 126.2; 85.7 (C(1), C(2)); 36.9;
20.0. GC/Ms: m/z (%)¼55 (19); 77 (27); 104 (50); 105
(100); 120 (19); 133 (13); 218 (6); 236 (2); 254 (4). HRMS
(m/z): calcd for C17H18O2 (Mþ), 254.1307; found,
254.1310. FTIR (CaF2/CCl4) nmax: 3614; 3567; 3093;
3064; 3041; 3029; 2953; 2879; 1497; 1369; 1195; 1124;
1065; 1035.
Acknowledgements
4.10.4. Ethyl 2-(2-hydroxyindan-2-yl)-butanoate (4).
Purification. Column chromatography on silica gel
We thank the University of Paris-Sud and the CNRS for
their financial support.
1
(heptane/ethyl acetate 80:20), white solid, (54% yield). H
NMR (CDCl3) d (ppm): 7.17 (4H, m, Har); 4.24 (2H, q,
O–CH2–CH3, JCH –CH ¼7.3 Hz); 3.27 (1H, s broad,
2
3
C(20)OH); 3.00 (4H, m, C(10)H2, C(30)H2); 2.55 (1H, dd,
References
C(2)H, J2-Ha¼10.8 Hz, J2-Hb¼3.9 Hz); 1.86 (2H, m,
C(3)HaHb); 1.32 (3H, t, O–CH2–CH3, JCH –CH2¼7.3 Hz);
1. Namy, J. L.; Girard, P.; Kagan, H. B. New J. Chem. 1977, 1,
5–7.
3
0.98 (3H, t, C(4)H3, J4–3¼7.6 Hz). 13C NMR (CDCl3) d
(ppm): 175.8 (C(1)); 140.7 a0nd 140.5 (C(30a), C(70a)); 126.4
(C(40), C(70)); 124.6 (C(5 ), C(60)); 82.0 (C(20)); 60.4
(O–CH2–CH3); 55.5 (C(2)); 46.5 and 44.6 (C(10), C(30));
21.4 (C(3)); 14.1 (O–CH2–CH3); 12.2 (C(4)). GC/MS m/z (%
base peak): 248 (1); 230 (21); 157 (21); 156 (8); 143 (8); 133
(10);132(31);129(10);117(11);116(22);115(27);105(57);
104 (100); 103 (47); 101 (22); 91 (18); 88 (12); 79 (14); 78
(24); 77 (28); 73 (35); 71 (13); 70 (12); 55 (17); 43 (24); 41
(14); 39 (11). FTIR (CaF2/nujol) nmax: 3520; 1715; 1348;
1318; 1291; 1262; 1226; 1213; 1180; 1154; 1129; 1099; 1080.
2. Corey, E. J.; Zheng, G. Z. Tetrahedron Lett. 1997, 38,
2045–2048.
3. Nomura, R.; Matsumo, T.; Endo, T. J. Am. Chem. Soc. 1996,
118, 11666–11667.
4. (a) Fukuzawa, S.; Sumimoto, N.; Fujinami, T.; Sakai, S. J. Org.
Chem. 1990, 55, 1628–1631. (b) Fukuzawa, S.; Fujinami, T.;
Sakai, S. J. Chem. Soc. Chem. Commun. 1986, 475–476.
´
5. Di Scala, A.; Garbacia, S.; Helion, F.; Lannou, M. I.; Namy,
J. L. Eur. J. Org. Chem. 2002, 2989–2995.
6. Souppe, J.; Namy, J. L.; Kagan, H. B. Tetrahedron Lett. 1982,
23, 3497–3500.
4.10.5. (E)-3-Methyl-1-(2,6,6-trimethylcyclohex-2-enyl)-
hept-1-en-3-ol (5). Purification. Column chromatography
on silica gel (heptane/ethyl acetate 95:5), colourless oil,
(96% yield). 1H NMR (CDCl3) d (ppm): 5.37 (3H, m,
7. Girard, P.; Namy, J. L.; Kagan, H. B. J. Am. Chem. Soc. 1980,
102, 2693–2698.
´
8. Lannou, M. I.; Helion, F.; Namy, J. L. Tetrahedron Lett. 2002,
43, 8007–8010.
9. Machrouhi, F.; Hamann, B.; Namy, J. L.; Kagan, H. B. Synlett
1996, 633–634.
C(1)H, C(2)H, C(30)H); 2.02 (1H, d, C(10)H, J1 –1¼8.8 Hz);
0
1.90 (2H, m, C(40)H2); 1.63–1.04 (9H, m, C(3)OH, C(4)H2,
C(5)H2, C(6)H2, C(50)H2); 1.51 (3H, s, C(20)–CH3); 1.20
(3H, s, C(3)–CH3); 0.81 (6H, s, C(60)–(CH3)2); 0.74 (3H,
10. Namy, J. L.; Souppe, J.; Kagan, H. B. Tetrahedron Lett. 1983,
24, 765–766.

M.-I. Lannou et al. / Tetrahedron 59 (2003) 10551–10565
10565
´
11. Helion, F.; Namy, J. L. J. Org. Chem. 1999, 64, 2944–2946.
12. Lebrun, A.; Rantze, E.; Namy, J. L.; Kagan, H. B. New
J. Chem. 1995, 19, 699–705.
24. Deacon, G. B.; Feng, T.; Nickel, S.; Skelton, B. W.; White,
A. H. J. Chem. Soc., Chem. Commun. 1993, 1328–1329.
25. Hitchcock, P. B.; Lappert, M. F.; Prashar, S. J. Organomet.
Chem. 2000, 613, 105–110.
13. Namy, J. L.; Girard, P.; Kagan, H. B.; Caro, P. E. New J. Chem.
1981, 5, 479–484.
26. (a) Hazin, P. N.; Human, J. C.; Bruno, J. V. Organometallics
1987, 6, 23–27. (b) Giuseppone, N. PhD Thesis, University of
Paris-Sud, 2000..
14. Fukuzawa, S.; Matsuzawa, H.; Yoshimitsu, S. J. Org. Chem.
2000, 65, 1702–1706.
15. (a) Molander, G. A.; Etter, J. B.; Harring, L. S.; Thorel, J. P.
J. Am. Chem. Soc. 1991, 113, 8036–8044. (b) Molander, G. A.;
Brown, G. A.; Storch de Garcia, I. J. Org. Chem. 2002, 67,
3453–3463.
27. Fukuzawa, S. I.; Fujinami, T.; Yamaguchi, S.; Sakai, S.
J. Chem. Soc., Perkin Trans. 1 1986, 1929–1932.
28. Evans, W. J.; Feldman, J. D.; Ziller, J. W. J. Am. Chem. Soc.
1996, 118, 4581–4584.
16. Inanaga, J.; Yokoyama, Y.; Handa, Y.; Yamaguchi, M.
Tetrahedron Lett. 1991, 32, 6371–6374.
29. (a) Imamoto, T.; Kusumoto, T.; Yokoyama, M. J. Chem. Soc.,
Chem. Commun. 1982, 1042–1044. (b) Imamoto, T.;
Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya, Y.
J. Am. Chem. Soc. 1989, 111, 4392–4398.
17. Newman, M. S.; Evans, F. J., Jr. J. Am. Chem. Soc. 1955, 77,
946–947.
18. Utimoto, K.; Takai, T.; Matsubara, S. Bull. Soc. Chim. Fr.
1997, 134, 365–369.
30. (a) Kobayashi, S. Chem. Lett. 1991, 2087–2089.
(b) Kobayashi, S.; Hachiya, I. Tetrahedron Lett. 1992, 33,
1625–1628. (c) Kobayashi, S.; Sugiura, M.; Kitagawa, H.;
Lam, W. W. L. Chem. Rev. 2002, 102, 2227–2302.
31. Vougioukas, A. E.; Kagan, H. B. Tetrahedron Lett. 1987, 28,
5513–5516.
19. Fu¨rstner, A. Synthesis 1989, 571–590.
20. Chao, L. C.; Rieke, R. D. J. Org. Chem. 1975, 40, 2253–2255.
21. Kagoshima, H.; Hashimoto, Y.; Oguro, D.; Saigo, K. J. Org.
Chem. 1998, 63, 691–697.
22. Imamoto, T.; Kusumoto, T.; Tawarayama, Y.; Sugiura, Y.;
Mita, T.; Hatanaka, Y.; Yokoyama, M. J. Org. Chem. 1984,
49, 3904–3912.
32. (a) Van de Weghe, P.; Collin, J. Tetrahedron Lett. 1993, 35,
3881–3884. (b) Giuseppone, N.; Van de Weghe, P.; Mellah,
M.; Collin, J. Tetrahedron 1998, 54, 13129–13148.
23. Nakanishi, A.; Fukuzawa, S. I.; Fujinami, T.; Sakai, S. Chem.
Express 1986, 1, 275–278.