Effect of InCl3 on the addition of Grignard reagents to α,β-unsaturated carbonyl compounds

Article abstract of DOI:10.1016/S0040-4039(01)02263-8

Control of 1,2- versus 1,4-addition of organometallic reagents to enones remains a long-standing problem. There is still no satisfactory 1,2-directing agent comparable to the 1,4-directing effect of copper salts. We report that the presence of just 5 mol% indium(III) chloride can significantly alter the amount of 1,2-product formed in these reactions.

Full text of DOI:10.1016/S0040-4039(01)02263-8

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 887–890
Effect of InCl3 on the addition of Grignard reagents to
a,b-unsaturated carbonyl compounds
Brian G. Kelly and Declan G. Gilheany*
Chemistry Department and Conway Institute of Biomolecular and Biomedical Sciences, University College Dublin, Belfield,
Dublin 4, Ireland
Received 7 November 2001; accepted 28 November 2001
Abstract—Control of 1,2- versus 1,4-addition of organometallic reagents to enones remains a long-standing problem. There is still
no satisfactory 1,2-directing agent comparable to the 1,4-directing effect of copper salts. We report that the presence of just 5
mol% indium(III) chloride can significantly alter the amount of 1,2-product formed in these reactions. © 2002 Elsevier Science
Ltd. All rights reserved.
Carbonꢀcarbon bond formation is the fundamental
operation in the construction of organic molecules and
the addition of organometallic reagents to carbonyl
compounds is one of the most common CꢀC bond
forming reactions. However, when the carbonyl sub-
strate is a,b-unsaturated, a mixture of products may
arise (Scheme 1).
its infancy, a number of interesting and useful synthetic
aspects are already known.3 But only allylation has
been studied in much detail and indeed this is by far the
best known use for indium in organic synthesis. How-
ever, Trost and co-workers4 found that in palladium-
catalysed trimethylenemethane (TMM) cycloadditions
(which can be considered as a 1,4-addition followed by
cyclisation), the presence of 10 mol% indium(III) acetyl-
acetonate redirected the reaction from a conjugate
pathway to a 1,2 route with enones. It was this, allied
to the fact that the presence of indium has also been
reported to catalyse the reaction of allylic stannanes
with carbonyls,5 which led us to believe that our goal of
regioselectivity in the Grignard reaction of enones was
attainable.
Control of 1,2- versus 1,4-addition in such systems is
not satisfactory in that no 1,2-directing agent compara-
ble to the 1,4-directing effect of copper(I) salts1 has
been identified thus far—although CeCl3 has had a
significant effect in some cases.2 This is particularly true
when Grignard reagents must be used in cases where
the corresponding organolithium compounds (which
generally give the 1,2-adduct) are commercially unavail-
able or cause unwanted reactions by acting as bases.
With the foregoing in mind, we sought an agent which
could regioselectively direct such nucleophile/carbonyl
systems, thereby reducing or completely eliminating
unwanted side-products. Herein we report our prelimi-
nary results, which show that the presence of 55 mol%
InCl3 as an additive can significantly alter the amount
of 1,2-product formed. We chose to study indium com-
pounds because although organoindium chemistry is in
Following Trost’s success using In(III) compounds, we
chose InCl3 as an additive in the test reaction of
PhMgBr with benzylidene acetone. Gratifyingly, the
presence of just 5 mol% InCl3 did indeed alter the
outcome of the reaction, increasing the 1,2/1,4-product
ratio to 88:12 as compared to 76:24 in the absence of
any additive. In order to establish if InCl3 was altering
the rate of this reaction, we carried out a simple compe-
O
R3
O
HO
R3
R2
R3MgX
+
+
R1
R2
R1
1,4 product
R2
R1
1,2 product
Scheme 1.
* Corresponding author. Tel.: +0-353-1-7162308; fax: +0-353-1-7162127; e-mail: declan.gilheany@ucd.ie
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)02263-8

888
B. G. Kelly, D. G. Gilheany / Tetrahedron Letters 43 (2002) 887–890
Before we can discuss the role of InCl3, some discussion
of the 1,2/1,4 ratio in the absence of additive is merited.
Clearly there is a large variation between each of the
substrate/Grignard combinations indicating that the
ratio is both substrate and reagent dependent. It is a
phenomenon that has been discussed elsewhere in the
literature6 but a universal theory which adequately
explains it has yet to be generally accepted. Holm7 has
carried out a series of kinetic experiments in order to
shed some light on the situation. He asserted that only
substrates which have or may assume a cisoid confor-
mation may react in a 1,4 fashion via a polar mecha-
nism. By invoking the concept of a homolytic reaction
mechanism he rationalised why certain transoid sub-
strates such as 2-cyclohexenone also underwent 1,4-
addition with certain Grignard reagents. However, he
did note anomalies within this theory whereby certain
Grignard reagents which are unlikely candidates for
homolysis were still observed to undergo 1,4-addition
with transoid substrates.
tition experiment. Two reactions were performed simul-
taneously under identical conditions—one in the
absence of additive and the other in the presence of 5
mol% InCl3. Both reactions were quenched after 60 min
1
and after workup, analysis by H NMR revealed a 98%
conversion of enone in the case where InCl3 was
present, and 88% conversion when it was not. Thus it
was evident that (at least in this case) InCl3 was capable
of inducing an increase in the rate of reaction. Conse-
quently we decided to screen a number of Grignard
reactions with various enones in the presence of InCl3
(Table 1). It can be seen from Table 1 that in many
cases there was a significant alteration in the product
distribution when just 5 mol% of InCl3 was added to
the reaction.
We can make the following generalisations about the
results we have obtained. In the cases of MeMgCl and
strikingly with EtMgBr when there is not already exclu-
sive 1,2-product formation in the absence of any addi-
tive, the addition of InCl3 causes an increase in the
proportion of 1,2-product formed. The opposite is the
case (i.e. there is a proportional decrease in the amount
of 1,2-product) with allylmagnesium bromide and tert-
butylmagnesium chloride as the Grignard reagents.
However, the situation with phenylmagnesium halide is
less clear cut.
Initially we had envisaged the role of InCl3 as that of a
Lewis acid, whereby it would coordinate to the oxygen
of the carbonyl moiety of the enone, thereby rendering
it more electrophilic and thus promote attack by the
incoming nucleophile at the carbonyl position (Scenario
1). Our initial postulate therefore, was that if we added
Table 1. Effect of added InCl3 on the 1,2/1,4-product ratiosa in the reactionb of Grignard reagents with various enones
a
Determined by 1H NMR; no other products formed unless otherwise stated. Figures in parentheses refer to the percentage conversion of enone
as determined by 1H NMR.
b Typical procedure was as follows: To a solution of enone (1 equiv.) in dry THF (ca. 1.5 mmol/30 mL) was added InCl3 (0.05 equiv.). The
resulting suspension was stirred rapidly and cooled to 0°C. Grignard reagent (1.1 equiv.) was then added all at once via syringe, at which point
dissolution of the suspension occurred. The reaction mixture was allowed to warm to ambient temperature over 16 h, after which it was
quenched with a saturated ammonium chloride solution (ca. 30 mL), extracted with ether (3×20 mL), dried over sodium sulfate and concentrated
in vacuo to give the crude product(s).
c PhMgCl was used.
d Reduction and other side products formed.

B. G. Kelly, D. G. Gilheany / Tetrahedron Letters 43 (2002) 887–890
889
enough InCl3 to each reaction we would be able to
induce 1,2-addition in all cases.
than that of ‘ordinary’ Grignard reagent, but with a
sluggish reactivity in comparison (Scenario 2). How-
ever, the situation is not quite as simple as that because
there is an alternative scenario (Scenario 3), where InR3
acts as a Lewis acid and the 4th equivalent of Grignard
derived nucleophile is directed to the carbonyl (again,
similar to Scenario 1).
However, when we added 100 mol% InCl3 to the reac-
tion of PhMgBr with benzylidene acetone (Table 2,
entry 3), we found that no addition to the enone
occurred. This was quite puzzling. However, when we
closely examined the set of reactions we had originally
screened (Table 1), we found that in nearly every case
the addition of InCl3 caused a decrease in the conver-
sion of starting material and hence a decrease in the
yield of addition products. Thus it was apparent that
the InCl3 was competing with the enone for attack by
the Grignard nucleophile.
It is likely that Scenario 2 operates, as the low yield of
addition product would indicate that a species less
reactive than RMgX is the reactant. That said, how-
ever, to the best of our knowledge the only reported
indium ‘ate’ complexes have been made using organo-
lithium reagents, and not organomagnesiums. We also
cannot rule out the possibility of both scenarios (2 and
3) occurring in parallel.
Therefore we had to re-evaluate our initial assumptions
about the role of InCl3 and we considered a number of
possibilities, two of which are outlined below:
The above arguments may be used to explain the effect
of added InCl3 on the methyl and ethyl and to a certain
extent phenyl Grignard reactions. However, they can-
not be used for the allyl and tert-butyl reagents. Both
of these reagents are potential candidates for reaction
via a homolytic mechanism, as they can form radicals
which are stabilised by resonance and hyperconjuga-
tion, respectively. Therefore, it may be that the presence
of InCl3 stabilises the radicals formed thus inhibiting
the ‘usual’ reactions in the absence of any additive.
That said, there was no obvious ketyl coloration
observed in any of the reactions with or without an
indium additive. Alternatively, the putative InR3 Lewis
acid in operation in the cases of the reactions with
EtMgBr and MeMgCl has never been reported to form
from allyl or tert-butyl Grignard reagents. Thus, what-
ever species does form (presumably InR2Cl or InRCl2)
in these cases may in fact weakly direct 1,4-addition.
(a) The Grignard reagent reacts with InCl3 to form an
InR3 species which then delivers the R group to the
enone. The reported preparation of many trialkyl- and
triaryl-indiums is just this reaction,8 so this is a likely
process in our system. When we carried out a control
reaction by adding 3 equivalents of PhMgBr along with
100 mol% of InCl3 (Table 2, entry 4), again we found
there was no reaction with the enone. This rules out the
possibility of the InR3 species supplying the nucleo-
phile, at least in the phenyl case. Another possibility
(where 5 mol% InCl3 is present) is that the InR3 species
formed could activate the carbonyl position on the
enone for further attack by the remaining Grignard
reagent (similar to Scenario 1, above). However, this
can only be the case if the InR3 species does not further
react with the Grignard reagent to form an ‘ate’ com-
plex (see below).
The situation with phenylmagnesium halide is less
clear-cut. Although with PhMgBr there is an increase in
the proportion of 1,2-product formation in some cases,
when PhMgCl was used with 2-cyclohexenone (Table
1), the addition of InCl3 caused a marked decrease in
the proportion of 1,2-product. It is too early at this
stage to speculate as to why this should be so, although
major differences have been seen before in the reaction
of acyclic and cyclic a,b-unsaturated ketones.10
(b) An indium ‘ate’ species (InR4−+MgX) could be formed
(which is more reactive than the corresponding InR3
entity9) which could then act as the aryl-/alkylating
agent. When we added 4 equivalents of PhMgBr along
with 100 mol% InCl3, we obtained the 1,2-product
exclusively, but there was a low conversion of starting
material. This would appear to confirm the intermedi-
acy of an ‘ate’ species, whose regioselectivity is greater
Table 2. Variation of amount of InCl3 and Grignard reagenta
HO
O
Ph
O
InCl3
THF, 0oC, 16 h
Ph
Me
1,2 product
+
+
PhMgBr
Ph
Ph
Me
Ph
Me
1,4 product
b
Entry
Equivalents of PhMgBrb
Mol% of InCl3
1,2/1,4 Ratioc
1
2
3
4
5
1.1
1.1
1.1
3
–
76:24 (100)
88:12 (100)
– (0)
– (0)
100:0 (27)
5
100
100
100
4
a Figures in parentheses refer to percentage conversion of enone as determined by 1H NMR.
b W.r.t. benzylidene acetone.
c Determined by 1H NMR.

890
B. G. Kelly, D. G. Gilheany / Tetrahedron Letters 43 (2002) 887–890
Having established that InCl3 did indeed alter the
product distribution in the reaction of Grignard
reagents with enones, but with a decrease in conversion
in most cases, we wanted to establish if less than 5
mol% InCl3 could still induce a similar redistribution of
products, but with less of a drop in conversion. We
investigated the reaction of EtMgBr with benzylidene
acetone because this combination of Grignard and sub-
strate had shown the most pronounced effect when we
had carried out our initial screening. The results of this
investigation are shown in Table 3.
assert that the effect is not just limited to allylic nucleo-
philes. Efforts are ongoing to investigate further the
scope of indium compounds in these nucleophile/car-
bonyl systems.
Acknowledgements
We gratefully acknowledge that this work was sup-
ported by Enterprise Ireland under Grant No. ST/2000/
040. We thank Dr. Nick Greeves for helpful
discussions.
As can be seen from these results, there is a correlation
between the amount of InCl3 and the decrease in con-
version of enone. However, as the InCl3 loading
increases from 1 to 20 mol%, there is very little change
in the 1,2/1,4-product ratio. Thus, in this limited case it
would seem that a reduction in the loading of InCl3 is
advantageous with regard to the conversion without a
significant diminution of the 1,2 selectivity.
References
1. (a) House, H. O.; Respess, W. L.; Whitesides, G. M. J.
Org. Chem. 1966, 31, 3128; (b) Posner, G. H. Org. React.
1972, 19, 1–113; (c) Posner, G. H. An Introduction to
Synthesis Using Organocopper Reagents; Wiley-Inter-
science: New York, 1980; (d) Lipshutz, B. H. Synthesis
1987, 325; (e) Taylor, R. J. K. Organometallic Reagents;
Oxford University Press: Oxford, 1994.
2. Molander, G. Chem. Rev. 1992, 92, 29–68.
3. (a) Cintas, P. Synlett 1995, 1087; (b) Li, C. J. Tetrahedron
1996, 52, 5643; (c) Li, C. J.; Chan, T. H. Tetrahedron
1999, 55, 11149; (d) Babu, G.; Perumal, P. T. Aldrichim.
Acta 2000, 33, 16; (e) Chauhan, K. K.; Frost, C. G. J.
Chem. Soc., Perkin Trans. 1 2000, 3015.
In conclusion, we have shown that the presence of
InCl3 can have a marked effect on the regioselective
outcome of the reaction of Grignard reagents with
enones. However, although it is too early at this stage
to comment definitively as to why this is so, we can
Table 3. Influence of InCl3 on the reaction of 1.1 equiv.
of EtMgBr with benzylidene acetone
4. Trost, B. M.; Sharma, S.; Schmidt, T. J. Am. Chem. Soc.
HO
Et
O
O
1992, 114, 7093.
5. Loh, T.-P.; Li, X.-R. Angew. Chem., Int. Ed. 1997, 36,
980.
InCl3
THF, 0oC, 16h
+
EtMgBr
+
Et
Ph
Ph
Ph
6. March, J. Advanced Organic Chemistry; Wiley & Sons,
1985.
Mol% InCl3
1,2/1,4 Ratioa
% Conversionb
7. Holm, T. Acta Chem. Scand. 1991, 45, 925.
8. Tuck, D.; Leman, J. T.; Barron, A. R. Encyclopedia of
Inorganic Chemistry; Wiley & Sons, 1994.
–
1
5
10
20
15:85
41:59
45:55
43:57
47:53
99
89
86
80
50
9. Araki, S.; Jin, S. J.; Butsugan, Y. J. Chem. Soc., Perkin
Trans. 1 1995, 549.
10. (a) House, H. O.; Thompson, H. W. J. J. Org. Chem.
1963, 28, 360; (b) Alexender, E. R.; Coraor, G. R. J. Am.
Chem. Soc. 1951, 73, 2721.
a Determined by 1H NMR.
b Refers to percentage conversion of enone.