Allylation of aldehydes promoted by the cerium(III) chloride heptahydrate/sodium iodide system: The dependence of regio- and stereocontrol on the reaction conditions

Article abstract of DOI:10.1002/adsc.200505184

The cerium(III) chloride heptahydrate/sodium iodide complex (CeCl ₃·7 H₂O/NaI) acts as a useful promoter in the carbon-carbon bond forming reaction by addition of allyltributylstannanes to aldehydes. The reaction of 2-butenyltributyl

Full text of DOI:10.1002/adsc.200505184

FULL PAPERS
DOI: 10.1002/adsc.200505184
Allylation of Aldehydes Promoted by the Cerium(III) Chloride
Heptahydrate/Sodium Iodide System: the Dependence of Regio-
and Stereocontrol on the Reaction Conditions
G. Bartoli,a,* A. Giuliani,b E. Marcantoni,b M. Massaccesi,b P. Melchiorre,a
S. Lanari,b L. Sambria
a
Dipartimento di Chimica Organica “A. Mangini”, V. le Risorgimento 4, 40136 Bologna, Italy
Fax: (þ39)-(0)51-209-3654, e-mail: giuseppe.bartoli@unibo.it
b
Dipartimento di Scienze Chimiche, via S. Agostino 1, I-62032 Camerino (MC), Italy
Received: May 3, 2005; Accepted: July 23, 2005
Abstract: The cerium(III) chloride heptahydrate/so- is observed in solvent-free conditions. Conversely,
dium iodide complex (CeCl3 · 7 H2O/NaI) acts as a when the reaction is carried out in acetonitrile as
useful promoter in the carbon-carbon bond forming the solvent, the a-adduct largely prevails. In the last
reaction by addition of allyltributylstannanes to alde- case, a complete stereocontrol is observed, the less
hydes. The reaction of 2-butenyltributylstannane stable (Z)-isomer being obtained in high geometrical
shows that the regio- and the stereochemical out- purity.
comes depend on the reaction conditions. When the
promoter is adsorbed on a solid support (aluminum Keywords: allylation; cerium; diastereoselectivity;
oxide), a highly prevalent formation of the g-adduct Lewis acids; stannanes; synthetic methods
Introduction
presence of appropriate additives have been developed.
For example, cobalt(II) chloride[8] and tin(II) halides[9]
The addition of allylic metal compounds to aldehydes to promote the reaction of allylstannanes with aldehydes
yield homoallylic alcohols[1] is a useful transformation in to give the prevalent formation of a-adducts. These
organic synthesis and consequently has received consid- methodologies allow the preferential insertion of the
erable attention in recent years.[2] For this purpose, allyl- 2-butenyl group in the more stable (E)-structure. This
stannanes have been extensively employed since these is due to the combination of two factors: a) the 2-bute-
reagents offer an attractive combination of stability nyltributylstannane is easily available in the (E)-config-
and high reactivity.[3] In widespread available methodol- uration, or in a 75:25 mixture of (E)/(Z)-isomers respec-
ogies, the reaction with g-substituted allylstannanes gen- tively, owing to the high tendency of (Z)-compounds to
erally proceeds with g-regioselectivity.[4] Conversely, isomerize to the more stable (E)-isomer; b) the transfer
only few protocols for the regioselective synthesis of of an alkenyl moiety generally proceeds with retention
a-adducts have been reported, because almost all allylic of configuration at the D2 position.[10] In conclusion,
metal derivatives react with aldehydes to give the g-ad- the synthesis of a (Z)-allyl alcohol is still an unsolved
ducts exclusively. Therefore, much effort has been re- problem.[11]
cently devoted to solve this problem. However, a com-
During our program to develop new synthetic uses of
plete a-regioselectivity can hardly be achieved since it the CeCl3 ·7 H2O/NaI[12] combination in promoting C C
varies depending on the nature not only of the allylic bond formation,[13] we found that this system strongly fa-
metal reagents, but also of aldehydes, promoters, and re- cilitates the addition of allyltributylstannane to alde-
action conditions.[5] Appreciably high a-selective allyla- hydes in CH3CN.[14] We report now that by this protocol
tion is observed in the reaction of aldehydes with allylic a highly regio- and stereoselective addition of the 2-al-
barium,[6] and allylic cerium reagents.[7] However, these kenyltributyltin derivative is accomplished surprisingly
metal-mediated allylations are very difficult to handle leading to the prevalent formation of the a-adduct in
because the reactions have to be performed under strict- the less stable (Z)-configuration.
À
ly anhydrous, oxygen-free, and low temperature condi-
Furthermore, we report our studies on the same reac-
tions. In order to set up more practical procedures, effi- tion carried out in solvent-free conditions in the pres-
cient protocols based on the use of allylstannanes in the ence of the promoter supported on neutral alumina.
Adv. Synth. Catal. 2005, 347, 1673 – 1680
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1673

G. Bartoli et al.
FULL PAPERS
The adoption of these different conditions gave com- dure proved to be general and could be applied to a
pletely different results, the almost exclusive formation broad range of aldehydes (see Table 1). Good results
of the g-adduct having been observed.
were in fact obtained with aromatic, aliphatic and heter-
ocyclic aldehydes.[21]
Also an acid-sensitive aldehyde such as furfural (1i) is
converted into the corresponding homoallylic alcohol in
good yield using this procedure (Table 1, entry 9). It is
Results and Discussion
Before analyzing the regio- and stereocontrol observed noteworthy that cinnamaldehyde (1h) undergoes regio-
in the addition of crotyltributylstannane to aldehydes in selective 1,2-addition exclusively.
various reaction conditions, we wish to report our inves-
As expected, the reactivity of aryl aldehydes is strong-
tigations on the ability of the CeCl3 · 7 H2O/NaI system ly dependent on the nature of the substituents on the ar-
adsorbed on a solid support to promote the allylation of omatic ring. Electron-withdrawing groups (Table 1, en-
aldehydes by examining at first the reaction of alde- tries 4 and 5) accelerate the addition process, while elec-
hydes 1 with the simple allyltributylstannane 2 tron-donating ones show a strong deactivating effect
(Scheme 1).
(Table 1, entries 2 and 3). For these reasons, the allyla-
Since several studies on the use of rare earth com- tion of 4-methoxybenzaldehyde (1c) is effected in very
pounds supported on silica gel have been reported in low yields since in this case side processes become com-
the literature,[15] we tested our allylation reaction on a petitive.
silica gel surface in solvent-free conditions.[16] Unfortu-
The presence of NaI and the use of the promoter sup-
nately, the reaction of benzaldehyde (1a) with allyltribu- ported on Al2O3 are essential for the efficiency of the
tylstannane 2 in the presence of CeCl3 · 7 H2O/NaI-SiO2 process. In fact, in the absence of NaI or in the presence
does not work, only the unaffected 1a being recovered of unsupported CeCl3 ·7 H2O/NaI system in solvent-free
even after prolonged reaction times. This is consistent conditions the process becomes very sluggish and side
with the well-documented[17] instability of allylstan- processes largely prevail. Very likely, Al2O3 acts as a car-
nanes under acidic conditions, owing to their high ten- rier to increase the surface area available for the heter-
dency to undergo protodestannylation. We thought ogeneous reaction, and cerium salt interacts with oxy-
to solve this problem by changing the inorganic material gen atoms at the surface of the support, forming new ac-
support. It is known, in fact, that alumina is a particularly tive sites on the alumina local structure. The water also
interesting metal oxide widely used to carry out plays an important role in this allylation reaction. In
surface organic chemistry.[18] With this in mind, the fact, the reaction works only in the presence of hydrated
CeCl3[1·97] H2O/NaI system was immobilized on neutral CeCl3, while with dry CeCl3 no allylation product is de-
Al2O3 (Fluka, neutral, Brockman activity, grade 1, tected. However, it is still not clear how water molecules
150 mesh) and the activity of the resulting supported participate in the reaction.[22]
promoter was tested in the reaction of allyltributylstan-
nane 2 with benzaldehyde (1a; Scheme 1).
We wish to outline that the present methodology rep-
resents a very useful improvement with respect to the
process carried out in CH3CN, because it provides a
more practical work-up procedure[20] (see Experimental
Section). Moreover, we found that the activity of the cat-
alyst supported on Al2O3 is not weakened by absorption
of moisture from the air, and the catalyst can be stored
for long periods without any appreciable loss of activity.
Scheme 1.
Regio- and Stereoselective Control in the Reaction
We examined the influence of the reactants ratio and with Crotyltributylstannane (4)
of the relative combination of the two promoter compo-
nents on the reaction. We found that the best choice was The reaction of benzaldehyde (1a) with 1 equiv. of a
a 1:1 ratio between 1a and 2 in the presence of 1 equiv. of 85:15 mixture of (E)- and (Z)-2-butenyltributylstan-
CeCl3 · 7 H2O and 0.1 equiv. of NaI on Al2O3 (1 g/mmol nane (4)[23] in the presence of CeCl3 · 7 H2O/NaI combi-
aldehyde).
nation supported on Al2O3 at 508C (Method A) leads to
The reaction temperature is also important. In fact, the exclusive formation of the g-adduct 6a, (Table 2, en-
the process is sluggish at room temperature, with low try 1). The high preference towards the g-adduct is con-
conversion yields (40%) being obtained after 3 days. It firmed by the reaction of 4 with other aldehydes, (Ta-
was necessary to increase the temperature to 508C to ble 2, entries 3, 5, 8, 12 and 13). However, the regiocon-
have the reaction completed in acceptable times trol was strongly affected by the reactivity of the sub-
(24 h). A further increase in the temperature produced strate. In fact, with highly reactive aldehydes, such as
an increase of undesired side products.[20] This proce- 1a, 1e and 1h (Table 2, entries 1, 8 and 12), only the for-
1674
asc.wiley-vch.de
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2005, 347, 1673 – 1680

Allylation of Aldehydes Promoted by Cerium(III) Chloride Heptahydrate/Sodium Iodide
FULL PAPERS
Table 1. Allylation reaction of aldehydes promoted by CeCl3 · 7 H2O/NaI system
supported on Al2O3 at 508C.[a]
hyde 1h exclusively underwent
1,2-addition and 4-methoxybenzal-
dehyde (1c) did not react.
We observed that the a/g ratio is
not influenced by the double bond
configuration of the allyl moiety,
because similar results are obtained
by using 2-butenyltributylstannane
in complete (E)-configuration.[24]
Although it is known that both
(E)- and (Z)-2-butenylmetal re-
agents predominantly give syn-g-
adducts in Lewis acid-promoted re-
actions, the present reaction suffers
from poor stereoselectivity. In most
cases, in fact, approximately equal
yields for syn and anti homoallylic
alcohols are obtained.
Conversely, opposite results were
obtained when the reaction was car-
ried out in a solvent. For example,
the addition of 4 to benzaldehyde
(1a) in CH3CN in the presence of
CeCl3 ·7 H2O/NaI complex at
room temperature[14] (Method B)
yielded a 90:10 mixture of a- and
g-adducts, respectively, in 70%
global yields. The high preference
toward a-attack is confirmed by re-
sults obtained with other aldehydes
(see Table 2, entries 4, 6, 7, 9–11
and 14).
As already observed in the case
of Method A, even under the exper-
imental conditions adopted in
Method B, the formation of the mi-
nor g-adduct is not a stereocontrol-
led process, an almost 1:1 syn/anti-
mixture being obtained in all cases.
On the contrary, the process of
the a-adductformation surprisingly
proceeds with very high stereocon-
trol giving in all cases the (Z)-alco-
hol in very high purity, as shown by
accurate GC-MS and NMR analy-
sis.[25] These results are very re-
markable, since the formation of
the (E)-stereoisomer generally pre-
vails in this kind of reactions.
[a]
Reactions performed at 508C in the presence of 1.0 equiv. of CeCl3 ·7 H2O and 0.1
equiv. of NaI supported on neutral alumina (1.00 g/mmol aldehyde 1).
All starting aldehydes were commercially available.
All products were identified by their IR, NMR, and GC-MS.
Yields of products isolated by column chromatography over silica gel.
Yield by GC-MS analysis.
[b]
[c]
[d]
[e]
We wish to outline that, besides
the great relevance originated by
mation of the g-adduct 6 was observed, while with less the novelty of the unexpected stereochemical outcome,
reactive substrates, such as 1b and 1i (Table 2, entries 3 the present reaction appears very interesting from a syn-
and 13), an appreciable amount of a-adduct 5 was also thetic point of view. The a- and g-regioisomers in fact
detected.
can be easily separated by column chromatography,
In agreement with what was observed in the reaction and then this method provides a useful access to diffi-
with allyltributylstannane (2), a,b-unsaturated alde- cultly available (Z)-alcohols 5.
Adv. Synth. Catal. 2005, 347, 1673 – 1680
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1675

G. Bartoli et al.
FULL PAPERS
Table 2. Regioselective allylation of aldehydes promoted by CeCl3 · 7 H2O-NaI system.[a]
Entry
Aldehyde
Method[b]
Time [h]
Product
Regioisomeric Ratio[c] a/g
syn:anti[d]
Yield [%][e]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1a
1a
1b
1b
1c
1c
1d
1e
1e
1f
A
B
A
B
A
B
B
A
B
B
B
A
A
B
24
38
30
48
24
50
26
24
19
36
24
18
21
22
6a
0: 100
90: 10
42: 58
89: 11
50: 50
50: 50
46: 54
50: 50
85
70
77
81
0
81
93
91
99
79
71
74
80
85
5aþ6a
5bþ6b
5bþ6b
5cþ6c
5dþ6d
6e
5eþ6e
5fþ6f
5gþ6g
6h
81: 19
85: 15
0: 100
85: 15
81: 19
86: 14
0: 100
28: 72
100: 0
50: 50
50: 50
50: 50
50: 50
50: 50
50: 50
50: 50
36: 64
1g
1h
1i
5iþ6i
1i
5i
[a]
The allylation of aldehydes (1 mmol) by (E)-2-butenyltributylstannane 4 (1 mmol) was carried out with CeCl3 · 7H2O
(1 mmol) and NaI (0.1 mmol).
[b]
A: The CeCl3 · 7 H2O-NaI combination supported on neutral alumina (1.00 g/mmol aldehydes 1). B: A suspension of
CeCl3 · 7 H2O-NaI in acetonitrile.
[c]
[d]
[e]
1
The regioisomeric ratio was determined by GLC and H NMR spectroscopy.
1
The stereochemistry was determined by H and 13C NMR spectroscopy.
Total yield of products isolated by column chromatography over silica gel.
Mechanistic Considerations
more reasonably attributed to a weak stereocontrol in
the open chain transition state A.
Owing to the complexity of the obtained results, it is very
It is more difficult to explain the formation of the (Z)-
difficult to completely rationalize the observed stereo- a-adduct 5, when the reaction is carried out according to
chemical outcome, especially in the case of the reactions Method B. The simplest explanation is consistent with
carried out in CH3CN (Method B), which lead to the the hypothesis that in the presence of the solvent the in-
prevalent unexpected formation of (Z)-a-adducts. Our teraction between 4 and the aldehyde proceeds via an
difficulties arise from the fact that the various mecha- SE2 pathway according to a cyclic transition state B to
nisms proposed for other similar allylations of aldehydes give the linear alcoholate 8. According to this mecha-
do not fit our results (see below). However, a reasonable nism, the reaction must produce also Bu3SnI (9). In
explanation of our findings is tentatively proposed.
fact, the GC-MS analysis of the reaction mixture re-
It is well known that a carbonylic function can under- vealed the presence of small amounts of Bu3SnI and of
go attack from C1 or C3 carbon of 4.[26] This process is fa- Bu3SnOH, but not of Bu3SnCl. Bu3SnOH is very likely
cilitated by the presence of Lewis acids, like Ce(III) de- formed by reaction of Bu3SnI with the water present in
rivatives, able to coordinate the oxygen atom of the car- the reaction mixture. This hydrolysis process releases
bonyl group.[27]
the iodide ion and this accounts for the fact that the
In this context, the exclusive formation of the g-ad- use of a catalytic amount of sodium iodide is sufficient
duct 6 in the process carried out with supported catalyst to promote the reaction.[14]
(Method A) can be rationalized on the basis of an initial
Moreover, this mechanistic assumption is strength-
formation of a complex between the aldehyde 1 and the ened by the fact that during the drafts of this work, a pa-
Lewis acid species adsorbed on Al2O3, which undergoes per appeared in the literature, in which the authors pro-
a nucleophilic attack by the C3 carbon of 4 to give alco- posed a mechanism similar to that depicted in Scheme 2
holate 7 via an S’E2 pathway.[28] The lack of stereocontrol to explain the product distribution observed in the ally-
(see Table 2) observed could be attributed to the rever- lation of aldehydes catalyzed by carboxylic acids.[30]
sibility of the addition of 4 to 1, analogously to the hy-
As above mentioned, it is very difficult to rationalize
drated cerium(III) salt-promoted addition of g-substi- the formation of the (Z)-isomer 5. We can assume that
tuted allylmetal compounds to imine derivatives.[29] the addition of 4 to 1 through the cyclic transition state
However, we think that the poor selectivity can be B produces at first the alcoholate (E)-8, which under-
1676
asc.wiley-vch.de
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2005, 347, 1673 – 1680

Allylation of Aldehydes Promoted by Cerium(III) Chloride Heptahydrate/Sodium Iodide
FULL PAPERS
Scheme 2.
goes an isomerization promoted by an interaction be- does not occur under the adopted experimental condi-
tween Ce(III) and the carbon-carbon double bond. tions. Otherwise, it is not conceivable to assume the for-
This hypothesis is based on a recent study[31] which firm- mation of an allylcerium species since it is well known
ly established the ability of rare earth derivatives to co- that organocerium compounds are rapidly hydrolyzed
ordinate alkenes. On the other hand, the alternative by water. On the other hand, it has been proved that
hypothesis that alcoholate (Z)-8 is directly formed the presence ofwater is essential to promote the reaction.
from an interaction of 1 with 4 seems less plausible, be-
Another possible explanation, which has been recent-
cause it should imply a too severe stereocontrol in the ly proposed to account for the formation of linear a-ad-
cyclic transition state B. However, we do not have any ducts is based on the hypothesis that the reaction pro-
conclusive evidence supporting this mechanistic inter- ceeds through two separate steps. The first one implies
pretation. Infact, we were unabletoisolateor synthesize that the Ce(III) complex-promoted interaction between
compound (E)-8 in order to verify if, under our reaction the aldehyde 1 and 4 gives the g-adduct 6. In the second
conditions, it can undergo a facile isomerization to (Z)- step, 6 is converted to a-adduct 5 through an initial for-
8. Furthermore, it is impossible to follow the course of mation of a hemiacetal by addition to unreacted alde-
the reaction via NMR spectroscopy, owing to the pres- hyde 1 followed by a Lewis acid-promoted retroene re-
ence of paramagnetic Ce(III) species.[32]
action,[34] and then by a 2-oxonia [3,3]-sigmatropic rear-
Various other mechanistic hypotheses have been pro- rangement.[35]
posed to explain the formation of the a-adducts in other
However, this mechanism can be completely excluded
allylationreactionsystems. However, alltheseinterpreta- by the following findings. The treatment of g-adduct 6e,
tions are ruled out in our system by some unequivocal ex- (prepared with Method A), with a small or stoichiomet-
perimental evidence. In fact, the alternative hypothesis ric amount of the parent aldehyde 1e in the presence of
that the reaction in CH3CN proceeds through a prelimi- CeCl3 ·7 H2O/NaI system in CH3CN at room tempera-
nary transmetallation process[33] between stannane 4 ture, did not isomerize to the corresponding a-adduct
and the Ce(III) complex is excluded by our previous find- 5e. The treatment of 6e with a more reactive aldehyde,
ings[14] in which we demonstrated, through a direct spec- such as 3-phenylpropanal, gave analogously unsuccess-
troscopic analysis, that such a transmetallation process ful results.
Adv. Synth. Catal. 2005, 347, 1673 – 1680
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1677

G. Bartoli et al.
FULL PAPERS
trated at reduced pressure. The crude material was purified by
flash chromatography on a silica gel column (eluent, hexanes-
ethyl acetate, 80:20) to give a mixture of syn- and anti-2-meth-
yl-1-[4-(trifluoromethyl)phenyl]-3-buten-1-ol (6e); yield:
0.187 g (91%).
Finally we wish to outline that the opposite regio-
chemistry shown by Method A with respect to Method
B can be attributed to the fact that a highly organized cy-
clic transition state of type B is unlike to be arranged on
the Al2O3 surface.
Method B: To a suspension of CeCl3 · 7 H2O (0.373 g,
1.0 mmol) and NaI (0.015 g, 0.1 mmol) in acetonitrile
(12 mL) was added successively 4-(trifluoromethyl)benzalde-
hyde (1e; 1.0 mmol, 0.174 g, 0.14 mL) and 2-butenyltributyl-
stannane (4; 1.0 mmol, 0.313 g) at room temperature. The re-
sulting reaction mixture was stirred at that temperature until
the disappearance of the starting material (19 hours, checked
by TLC and GC analyses), then was quenched with 0.1 N
HCl solution. The aqueous layer was extracted with diethyl
ether and the combined organic layers were stirred for 1 h
with 10% KFin H2O (10 mL). Theorganic phase was separated
and washed with brine, dried with Na2SO4 and concentrated at
reduced pressure. The mixture of two diastereomers was sepa-
rated by silica gel chromatography (eluent, hexanes: ethyl ace-
tate 80/20) to furnish 0.173 (84%) g of (Z)-1-[4-(trifluorome-
thyl)phenyl]-3-penten-1-ol (5e) and 0.031 g (15%) of a mixture
of syn- and anti-2-methyl-1-[4-(trifluoromethyl)phenyl]-3-but-
en-1-ol (6e).
All obtained products, except 5c, are known compounds and
were identified by comparison of 1H and 13C NMR spectra with
literature data. Spectroscopic data of 5c are given below.
(Z)-1-(4-Methoxyphenyl)-3-penten-1-ol (5c): IR (neat): n¼
3400, 3017, 1648, 1541 cmÀ1; 1H NMR (300 MHz, CDCl3): d¼
1.61 (d, 3H, JHH ¼6.6 Hz), 2.21 (bs, 1H, OH), 2.40–2.62 (m,
2H), 3.98 (s, 3H), 4.78 (t, 1H, JHH ¼6.4 Hz), 5.38–5.47 (m,
1H), 5.58–5.68 (m, 1H), 6.89 (d, 2H, JHH ¼8.4 Hz), 7.30 (d,
2H, JHH ¼8.8 Hz); 13C NMR (75 MHz, CDCl3): d¼13.0, 31.6,
56.0, 73.4, 114.3, 122.6, 129.5, 130.6, 131.7, 160.6; MS (70 eV,
EI): m/z¼192 (Mþ), 137 (100), 109, 107, 77 , 51, 41, 39; anal.
calcd. (%) for C12H16O2 (192.115): C 74.97, H 8.39; found: C
74.95, H 8.27.
Conclusion
In conclusion, regardless of the mechanistic details, the
experimental simplicity of our CeCl3 · 7 H2O/NaI pro-
moted reaction, the low cost and ease of access to the re-
quired reagents, and the high regioselectivity observed
provide a convenient and practical method of allylation
of aldehydes. Furthermore, the combination of these ad-
vantages with the possibility to obtain (Z)-a-adducts
without isomerization to the more stable (E)-configura-
tion, makes our procedure a very efficient method for
the preparation of this important class of compounds.
The role of CeCl3 ·7 H2O and NaI is intriguing and
complex because the exact nature of the intermediate
obtained by the interaction of the reagents with the
CeCl3 · 7 H2O/NaI system is not yet known.[12a] Further
studies are in progress in our laboratories to analyze in
more detail the mechanistic outcome of the reaction.
Unfortunately, first attempts to extend this reaction to
other g-substituted allylstannanes such as cinnamyltri-
butylstannane, met with failure. We hope that under op-
timized conditions most undesired side reactions (poly-
merization and rapid protodestannylation) will be elim-
inated. The development of this new protocol also
prompted us to investigate further applications of our
reagent system in new schemes of synthesis, and at pres-
ent experiments for the preparation of biologically im-
portant substances are in progress in our laboratories.
Acknowledgements
Work carried out in the framework of the National Project “Ster-
eoselezione in Sintesi Organica. Metodologie e Applicazioni”,
and by National Project “Studio degli Aspetti Teorici ed Appli-
cativi degli Aggregati di Molecole Target su Siti Catalitici Ster-
eoselettivi” supported by MIUR, Rome, by the University of Bo-
logna and by University of Camerino. The authors are also
grateful to Drs. Massimo Bartolacci and Melissa Paoletti for
their assistance in various stages of this effort.
Experimental Section
Typical Experimental Procedure for Allylation
Method A: Neutral alumina (1 g) was added to a mixture of
CeCl3 · 7 H2O (0.373 g, 1.0 mmol) and NaI (0.015 g,
0.1 mmol) in acetonitrile (10 mL), and the mixture was stirred
overnight at room temperature. The acetonitrile was removed
by rotary evaporation and the resulting mixture stored in a bot-
tle at room temperature.
To the CeCl3 ·7 H2O-NaI combination supported on neutral
alumina (1.388 g) prepared as above was added successively 4-
(trifluoromethyl)benzaldehyde (1e; 1.0 mmol, 0.174 g,
0.14 mL) and 2-butenyltributylstannane (4; 1.0 mmol,
0.313 g) at room temperature. The resulting reaction mixture
was then stirred at that temperature until the disappearance
of the starting material (24 hours, checked by TLC and GC
analyses). After addition of Et2O the mixture was passed
through a short pad of Celite and the filtrate was stirred for
1 h with 10% KF in H2O (10 mL). The organic phase was sep-
arated and washed with brine, dried with Na2SO4 and concen-
References and Notes
´
[1] a) N. Solin, J. Kjellgren, K. J. Szabo, J. Am. Chem. Soc.
2004, 126, 7026–7033; b) J. W. Bode, D. R. Gauthier,
E. M. Carreira, Chem. Commun. 2001, 2560–2560;
c) A. Yanagisawa, in: Comprehensive Asymmetric Catal-
ysis, (Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Vol.
2, Springer-Verlag, Berlin, 1999, p. 965; d) H. Nakamura,
H. Iwama, Y. Yamamoto, J. Am. Chem. Soc. 1996, 118,
6641–6647; e) A. H. Hoveyda, J. P. Morken, Angew.
Chem. Int. Ed. 1996, 35, 494–498; f) W. R. Roush, in:
1678
asc.wiley-vch.de
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2005, 347, 1673 – 1680

Allylation of Aldehydes Promoted by Cerium(III) Chloride Heptahydrate/Sodium Iodide
FULL PAPERS
Comprehensive Organic Synthesis, (Eds.: B. M. Trost, I.
Fleming, C. H. Heathcock), Vol. 2, Pergamon Press, Ox-
ford, 1991, pp. 1–53.
in acetonitrile followed by complete removal of the sol-
vent. Thus, in this our methodology of Lewis acid pro-
moter solvent-free reaction the term ꢁsolvent–freeꢂ re-
fers solely to the reaction itself. On the other hand, prep-
aration of initial adsorbate and purification of products
invariably involve the use of solvent.
[2] Y. Yamamoto, N. Asao, Chem. Rev. 1993, 93, 2207–2293
and references cited therein.
[3] Y. Yamamoto, Acc. Chem. Res. 1987, 20, 243–249.
[4] a) I. Shibata, N. Yoshimura, M. Yabu, A. Baba, Eur. J. [17] Y. Naruta, Y. Nishigaichi, K. Maruyama, Chem. Lett.
Org. Chem. 2001, 3207–3211; b) G. E. Keck, D. E. Ab-
bott, E. P. Boden, E. J. Enholm, Tetrahedron Lett. 1984,
25, 3927–3930; c) Y. Yamamoto, H. Yatagai, Y. Naruta,
K. Maruyama, J. Am. Chem. Soc. 1980, 102, 7107–
7109; d) R. W. Hoffman, H.-J. Zeiss, Angew. Chem. Int.
Ed. 1979, 18, 306–307.
1986, 1857–1860.
[18] R. Maggi, R. Ballini, G. Sartori, R. Sartorio, Tetrahedron
Lett. 2004, 45, 2297–2299.
[19] CeCl3 ·7 H2O, NaI and Al2O3 were suspended in CH3CN
at room temperature. The removal of the solvent by ro-
tary evaporation at 358C affords the heterogeneous
Lewis acid promoter.
[20] The removal of any excess of allyltributylstannane and
other tin by-products has been achieved by treatment
of the filtrate with 10 mol % KF and rapid filtration
through silica gel (A. G. Davies, Organotin Chemistry,
Wiley-VCH, Weinheim, 1997).
[21] The procedure exhibits high chemoselectivity towards al-
dehydes in the presence of ketones. In fact, when a mix-
ture of 1a and acetophenone is allowed to react with 2,
only the allylic alcohol 3a and unaffected acetophenone
were found.
[5] a) K.-T. Tan, S.-S. Chang, H.-S. Cheng, T.-P. Loh, J. Am.
Chem. Soc. 2003, 125, 1007–1010; b) T.-P. Loh, K.-T. Tan,
Q.-Y. Hu, Tetrahedron Lett. 2001, 42, 8705–8708; c) A.
Takuwa, J. Shiigi, Y. Nishigaichi, Tetrahedron Lett.
1993, 34, 3457–3460; d) K. Kanagawa, Y. Nishigaichi, J.
Org. Chem. 1992, 57, 6988–6991; e) Y. Yamamoto, N.
Maeda, K. Maruyama, J. Chem. Soc. Chem. Commun.
1983, 742–743; f) A. Gambaro, P. Gains, D. Marton, V.
Peruzzo, G. Tagliavini, J. Organomet. Chem. 1982, 231,
307–313.
[6] A. Yanagisawa, S. Habane, K. Yasne, H. Yamamoto, J.
Am. Chem. Soc. 1994, 116, 6130–6141.
[7] a) S. Matsukawa, Y. Funabashi, T. Imamoto, Tetrahedron
Lett. 2003, 44, 1007–1010; b) T. Cohen, J. R. Matz, J. Am.
Chem. Soc. 1980, 102, 6900–6902.
[22] The role of water was investigated in previous findings
(see ref.[14]), and, without conclusive data, we postulate
that cerium center may require legation by 1 or more
equiv. of water for generating fully active species.
[23] Starting from commercial crotyl bromide (85% E config-
uration) the configuration of the allyl moiety is essential-
ly preserved, see: a) E. Winter, R. Brꢃckner, Synlett
1994, 1049–1053; b) A. Takuwa, Y. Naruta, O. Soga, K.
Maruyama, J. Org. Chem. 1984, 49, 1857–1862; c) Y.
Naruta, J. Am. Chem. Soc. 1980, 102, 3774–3775.
[24] S. Weigand, R. Brꢃckner, Synthesis 1996, 475–482.
[25] For sake of accuracy, typical NMR experiments were car-
ried out to confirm the (Z)-configuration of the double
bond in product 5. ROESY analysis of 5e showed re-
markable cross-peaks between vinylic protons, and be-
tween protons in C2 and protons in C5. Moreover, the
observed intramolecular NOE effect, the coupling con-
[8] J. Iqbal, S. P. Joseph, Tetrahedron Lett. 1989, 30, 2421–
2422.
[9] A. Ito, M. Kishida, Y. Kurusu, Y. Masuyama, J. Org.
Chem. 2000, 65, 494–498.
[10] a) F. von Gyldenfeldt, D. Marton, G. Tagliavini, Organo-
met. 1994, 13, 906–913; b) H. Yatagai, Y. Yamamoto, K.
Maruyama, J. Am. Chem. Soc. 1980, 102, 4548–4550.
[11] A. Takuwa, O. Soga, T. Mishima, J. Org. Chem. 1987, 52,
1261–1265.
[12] a) G. Bartoli, E. Marcantoni, L. Sambri, Synlett 2003,
2101–2116. b) a) G. Sabitha, R. Satheesh Babu, M. Raj-
kumar, R. Srividya, J. S. Yadav, Org. Lett. 2001, 3, 1149–
1151; c) S. Fukuzawa, T. Tsuruto, T. Fujinami, S. Sakai, J.
Chem. Soc. Perkin Trans. 1 1987, 1473–1472.
[13] a) G. Bartoli, M. Bosco, S. Giuli, A. Giuliani, L. Lucarel-
li, E. Marcantoni, L. Sambri, E. Torregiani, J. Org. Chem.
2005, 70, 1941–1944; b) G. Bartoli, M. Bosco, G. Foglia,
A. Giuliani, E. Marcantoni, L. Sambri, Synthesis 2004,
895–900; c) G. Bartoli, M. Bartolacci, M. Bosco, G. Fo-
glia, A. Giuliani, E. Marcantoni, L. Sambri, E. Torregia-
ni, J. Org. Chem. 2003, 68, 4594–4597, d) G. Bartoli, M.
Bosco, M. C. Bellucci, E. Marcantoni, L. Sambri, E. Tor-
regiani, Eur. J. Org. Chem. 1999, 617–620.
1
stant in H NMR of the vinylic protons [3J (Z) 11 Hz],
and the chemical shifts in 13C NMR of the terminal meth-
yl group [dCH3 (Z) dCH3 (E)] are fully consistent with this
assignment, demonstrating that 5e is obtained in high di-
astereomeric purity and its (E)-isomer is present only in
traces.
[26] J. A. Marshall, K. W. Hinkle, J. Org. Chem. 1996, 61,
105–108.
[27] G. A. Molander, C. R. Harris, Chem. Rev. 1996, 96, 307–
338.
[14] G. Bartoli, M. Bosco, A. Giuliani, E. Marcantoni, A. Pal-
mieri, M. Petrini, L. Sambri, J. Org. Chem. 2004, 69,
1290–1297.
[15] G. Bartoli, M. Bartolacci, A. Giuliani, E. Marcantoni, M.
Massaccesi, E. Torregiani, J. Org. Chem. 2005, 70, 169–
174, and references cited therein.
[28] S. E. Denmark, H. Shinzo, J. Org. Chem. 1994, 59, 5133–
5135.
[29] T. Basile, A. Bocoum, D. Savoia, A. Umani-Ronchi, J.
Org. Chem. 1994, 59, 7766–7773.
[30] G.-L. Li, G. Zhao, J. Org. Chem. 2005, 70, 4272–4278.
This procedure shows a general outcome similar to our
reaction in CH3CN (Method B). However, the reported
[16] The CeCl3 · 7 H2O-NaI system dispersed on chromatog-
raphy silica gel prepared by simple mixing both reagents
Adv. Synth. Catal. 2005, 347, 1673 – 1680
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1679

G. Bartoli et al.
FULL PAPERS
results show a less efficient regio- and stereocontrol with
respect to our protocol.
[31] S. Kobayashi, Lanthanides: Chemistry and Uses in Or-
ganic Synthesis, Springer, Berlin, 1999.
D. Schinzer), Kluwer Academic Publishers: Dodrecht,
The Netherlands, 1989, pp 73–105; c) S. E. Denmark,
E. J. Weber, T. M. Wilson, Tetrahedron 1989, 45, 1053–
1065.
´
[32] a) L. G. Hubert-Pfalzgraf, L. Machado, Polyhedron 1996,
15, 545–549; b) H. A. Stecher, A. Seu, A. L. Rheingold,
Inorg. Chem. 1989, 28, 3280–3282.
[33] a) S. E. Denmark, N. G. Almstead Tetrahedron 1992, 48,
5565–5578; b) G. E. Keck, S. Castellino, M. B. Andrus,
in: Selectivities in Lewis Acid Promoted Reactions, (Ed.:
[34] J. L. Ripoll, Y. Vallee, Synthesis 1993, 659–677.
[35] a) J. Nokami, M. Ohga, H. Nakamoto, T. Matsubara, I.
Hussain, K. Kataoka, J. Am. Chem. Soc. 2001, 123,
9168–9169; b) S. Sumida, M. Ohga, J. Mitani, J. Nokami,
J. Am. Chem. Soc. 2000, 122, 1310–1313; c) J. Nokami, L.
Anthony, S. Sumida, Chem. Eur. J. 2000, 6, 2909–2913.
1680
asc.wiley-vch.de
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2005, 347, 1673 – 1680