Addition of allylindium reagents to aldehydes substituted at Cα or Cβ with heteroatomic functional groups. Analysis of the modulation in diastereoselectivity attainable in aqueous, organic, and mixed solvent systems

Article abstract of DOI:10.1021/ja953682c

The stereochemical course of indium-promoted allylations to α- and β-oxy aldehydes has been investigated in solvents ranging from anhydrous THF to pure H₂O. The free hydroxyl derivatives react with excellent diastereofacial control to give significantly heightened levels of syn-1,2-diols and anti-1,3-diols. Relative reactivities were determined in the α-series, and the hydroxy aldehyde proved to be the most reactive substrate. This reactivity ordering suggests that the selectivity stems from chelated intermediates. The rate acceleration observed in water can be heightened by initial acidification. Indeed, the indium-promoted allylation reaction mixtures become increasingly acidic on their own. Preliminary attention has been accorded to salt effects, and tetraethylammonium bromide was found to exhibit a positive synergistic effect on product distribution. Finally, mechanistic considerations are presented in order to allow for assessment of the status of these unprecedented developments at this stage of advancement of the field.

Full text of DOI:10.1021/ja953682c

J. Am. Chem. Soc. 1996, 118, 1931-1937
1931
Addition of Allylindium Reagents to Aldehydes Substituted at
CR or Câ with Heteroatomic Functional Groups. Analysis of
the Modulation in Diastereoselectivity Attainable in Aqueous,
Organic, and Mixed Solvent Systems
Leo A. Paquette* and Thomas M. Mitzel
Contribution from the EVans Chemical Laboratories, The Ohio State UniVersity,
Columbus, Ohio 43210
ReceiVed NoVember 1, 1996X
Abstract: The stereochemical course of indium-promoted allylations to R- and â-oxy aldehydes has been investigated
in solvents ranging from anhydrous THF to pure H2O. The free hydroxyl derivatives react with excellent diastereofacial
control to give significantly heightened levels of syn-1,2-diols and anti-1,3-diols. Relative reactivities were determined
in the R-series, and the hydroxy aldehyde proved to be the most reactive substrate. This reactivity ordering suggests
that the selectivity stems from chelated intermediates. The rate acceleration observed in water can be heightened by
initial acidification. Indeed, the indium-promoted allylation reaction mixtures become increasingly acidic on their
own. Preliminary attention has been accorded to salt effects, and tetraethylammonium bromide was found to exhibit
a positive synergistic effect on product distribution. Finally, mechanistic considerations are presented in order to
allow for assessment of the status of these unprecedented developments at this stage of advancement of the field.
The Cram rule,1 which was put forward more than four
decades ago as a predictor of the stereochemical course of
nucleophilic additions to acyclic aldehydes and ketones, initiated
organic chemists into thinking systematically about diastereo-
meric transition states. As a consequence of those complications
introduced by the dynamic conformational nature of substrates
bearing stereogenic centers proximal to the carbonyl group,
others have presented paradigms in which the interplay of steric
and stereoelectronic effects has been somewhat modified.2-4
Subtle intramolecular (in the form of torsional strain) and
intermolecular interactions (e.g., nonbonded compressions as-
sociated with the preferred trajectory for nucleophilic attack5)
have garnered serious attention. When R- and â-alkoxy
carbonyl compounds are involved and chelated intermediates
intervene, mechanistic analysis is simplified. In such instances,
addition often occurs from the sterically less hindered π-face
of the preorganized complex.6 Grignard reagents are reputed
to be particularly well suited to chelate control,7-9 while non-
chelate behavior has been reported for organolithium,10 alkyl-
titanium,6 and allylchromium reagents,11 as well as lower-order
cuprates.12
The sensitivity of the organometallic reagents mentioned
above to moisture requires that their addition reactions be
performed in anhydrous organic solvents. The metal indium
has recently been shown to offer intriguing advantages for
effecting C-C bond formation in an aqueous environment.12-14
The Cram and Felkin-Anh proposals have not been founded on
X Abstract published in AdVance ACS Abstracts, February 15, 1996.
(1) Cram, D. J.; Abd Elhafez, R. A. J. Am. Chem. Soc. 1952, 74, 5828.
(2) Karabatsos, G. J. J. Am. Chem. Soc. 1967, 89, 1367.
(3) Cherest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 2199.
(4) Ahn, N. T.; Eisenstein, O. NouV. J. Chim. 1977, 1, 61.
(5) (a) Bu¨rgi, H. B.; Dunitz, J. D.; Lehn, J. M. Tetrahedron 1974, 30,
1563. (b) Bu¨rgi, H. B.; Dunitz, J. D.; Shefter, E. J. J. Am. Chem. Soc. 1973,
95, 5065.
reactions carried out in water or under “wet conditions” of any
type. This change to a significantly more polar hydrogen
bonding medium could conceivably damp those factors control-
ling facial selectivity in the absence of water.13b,14e It need not,
however, if coordination to the indium ion overrides those
solvation forces that would break down the chelate.15 These
and many other related questions were viewed by us to warrant
a detailed comparative analysis of the indium-promoted ally-
lation of chiral R- and â-oxygenated aldehydes under conditions
where the reactive medium would range from anhydrous THF
to pure water. The observed stereoselectivities would, without
question, lift the limitations imposed by our current knowledge
base, which has been confined to observations made in the strict
absence of moisture. As others have emphasized,14-16 orga-
nometallic reactions conducted in aqueous media preclude any
need to make recourse to protection-deprotection tactics for a
number of functional groups. Perhaps still more advantageous
is the fact that water is the quintessentially benign solvent from
the environmental and flammability perspectives. Furthermore,
(10) (a) Rosen, T.; Taschner, M. J.; Heathcock, C. H. J. Org. Chem.
1984, 49, 3994. (b) Mahler, U.; Devant, R. M.; Braun, M. Chem. Ber. 1988,
121, 2035. (c) Montgomery, S H.; Pirrung, M. C.; Heathcock, C. H.
Carbohydr. Res. 1990, 202, 13. (d) Braun, M.; Mahler, H. Liebigs Ann.
1995, 29.
(11) (a) Jin, H.; Uenishi, J.; Christ, W. J.; Kishi, Y. J. Am. Chem. Soc.
1986, 108, 5644. (b) Mulzer, J.; Schulze, T.; Strecker, A.; Denzer, W. J.
Org. Chem. 1988, 53, 4098.
(12) Metz, P.; Schoop, A. Tetrahedron 1993, 46, 10597.
(13) (a) Li, C.-J. Chem. ReV. 1993, 93, 2023. (b) Chan, T.-H.; Li, C. J.;
Lee, M. C.; Wei, Z. Y. Can. J. Chem. 1994, 72, 1181. (c) Lubineau, A.;
Auge, J.; Queneau, Y. Synthesis 1994, 741.
(14) (a) Chan, T.-H.; Li, C.-J. Organometallics 1990, 9, 2649. (b) Li,
C.-J.; Chan, T.-H. Tetrahedron Lett. 1991, 32, 7017. (c) Chan, T.-H.; Li,
C.-J. J. Chem. Soc., Chem. Commun. 1992, 747. (d) Issac, M. B.; Chan,
T.-H. J. Chem. Soc., Chem. Commun. 1995, 1003. (e) Chan, T. H.; Li, C.
J. Can. J. Chem. 1992, 70, 2726.
(15) (a) Kim, E.; Gordon, D. M.; Schmid, W.; Whitesides, G. M. J. Org.
Chem. 1993, 58, 5500. (b) Gordon, D. M.; Whitesides, G. M. J. Org. Chem.
1993, 58, 7937. (c) Gao, J.; Ha¨rter, R.; Gordon, D. M.; Whitesides, G. M.
J. Org. Chem. 1994, 59, 3714.
(16) (a) Prenner, R. H.; Binder, W. H.; Schmid, W. Liebigs Ann. Chem.
1994, 73. (b) Binder, W. H.; Prenner, R. H.; Schmid, W. Tetrahedron 1994,
50, 749.
(6) (a) Reetz, M. T. Angew. Chem., Int. Ed. Engl. 1984, 23, 556. (b)
Reetz, M. T. Acc. Chem. Res. 1993, 26, 462.
(7) Cram, D. J.; Wilson, D. R. J. Am. Chem. Soc. 1963, 85, 1245.
(8) Nicolaou, K. C.; Claremon, D. A.; Barnette, W. E. J. Am. Chem.
Soc. 1980, 102, 6611.
(9) Still, W. C.; McDonald, J. H., III. Tetrahedron Lett. 1980, 21, 1031.
0002-7863/96/1518-1931$12.00/0 © 1996 American Chemical Society

1932 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Mitzel
Table 1. Indium-Mediated Allylations of R-Oxygenated Aldehydes
in Various Solvents (CH2dCHCH2Br, 25 °C)a
the need to remove water of hydration from compounds and
the requirement for a dry atmosphere would cease to exist.
Interest in the synthetic applications of indium, a previously
little explored metal, has been on the upswing in recent
years.13-19 At 5.785 eV, the first ionization potential of indium
falls well below that of comparable metals such as zinc,
magnesium, tin, and aluminum.14b Indium metal is roughly
comparable to silver in cost. However, regeneration of the
indium after completion of the reaction can be readily and
efficiently accomplished electrochemically.16a As a consequence
of the low solubility of certain classes of compounds in water,
we have also included experiments in which the medium consists
of equivolume amounts of THF and H2O.20
Results and Discussion
r-Oxygenated Aldehydes. A variety of aldehydes contain-
ing R-oxy substituents of widely differing basicities, viz. 1-8,
were examined for the purpose of elucidating whether Lewis
acid-base interactions play a role in indium-promoted reactions
and to what extent (Table 1). If chelate control is operational,
the expected allylation product is the syn diastereomer; other-
wise, conversion to the anti product results (Scheme 1). The
known aldehydes 1,21 2,21,22 and 521 were complemented by the
MOM derivative 4 and cyclic hemiacetal 3. The preparation
of the latter two substrates began with addition of vinylmag-
nesium bromide to cyclohexanecarboxaldehyde. Whereas hy-
droxyl protection in 9 followed by oxidative cleavage leads to
4, direct ozonolysis of 9 gives rise to 3 (Scheme 2). D-Arabinose
(6) and the highly oxygenated systems 723 and 824 round out
this subset.
The results obtained with 1 (entries 1-4) and 2 (entries 5-8)
provide important calibration points for non-chelate-controlled
behavior. In both instances, the anti product is favored.
Presumably because the basicity of the tert-butyldimethylsiloxy
substituent falls below that of benzyloxy, the anti percentages
reach a maximum for 1. The product distributions were
quantified by chemical conversion to the acetonide in advance
of high-field 1H NMR analysis21,25 (entries 1-4, 9-12, and 17),
by prior acetylation to gain solubility (entries 18-24),15a,16a or
most simply by direct spectroscopic integration (entries 5-8).
It is noteworthy that in every example allylations performed in
either H2O or H2O-THF (1:1) proceeded at appreciably more
rapid rates than in THF alone. For 1, the diastereoselectivity
realized is constant whether H2O is present or not. Interestingly,
the anti preference in the case of 2 decreases by a factor of
about 3 in pure H2O. Product yields were found to be
consistently high.
a All of the reactions were performed at least in duplicate at a
concentration of 0.1 M with vigorous stirring for the indicated time
span. The product distributions for 1-5 and 7 were determined by 1H
NMR integration at 300 MHz; for 6 and 8, the relative amounts of
products following chromatography are given. b The THF solution of
allyl bromide was heated to reflux with the indium prior to reaction.
The mixture was cooled to rt prior to introduction of the aldehyde,
and the coupling was performed as in a. c Performed at the reflux
temperature of the solvent. d Based on the chemical purity of the sample
of 5 employed. e Promoted by sonication.
Hemiacetal 3 must, of course, undergo ring opening prior to
condensation with the allylindium reagent. Beyond that, it is
not clear that C-C bond formation materializes prior to, or only
after, the loss of formaldehyde. The two options are ap-
proximated in MOM ether 4, and the R-hydroxy aldehyde 6
Scheme 1
(17) (a) Araki, S.; Ito, H.; Butsugan, Y. Synth. Commun. 1988, 18, 453.
(b) Araki, S.; Ito, H.; Butsugan, Y. J. Org. Chem. 1988, 53, 1831. (c) Araki,
S.; Katsumura, N.; Ito, H.; Butsugan, Y. Tetrahedron Lett. 1989, 30, 1581.
(d) Araki, S.; Butsugan, Y. J. Chem. Soc., Chem. Commun. 1989, 1286.
(18) Chao, L.-C.; Rieke, R. D. J. Org. Chem. 1975, 40, 2253.
(19) (a) Basile, T.; Bocoum, A.; Savoia, D.; Umani-Ronchi, A. J. Org.
Chem. 1994, 59, 7766. (b) Li, C.-J. Tetrahedron Lett. 1995, 36, 517. (c)
Li, C.-J.; Lu-Y.-Q. Tetrahedron Lett. 1995, 36, 2721.
(20) Preliminary communication: Paquette, L. A.; Mitzel, T. M.
Tetrahedron Lett. 1995, 36, 6863.
(21) Keck, G. E.; Boden, E. P. Tetrahedron Lett. 1984, 25, 265.
(22) Davis, A. P.; Jaspars, M. J. Chem. Soc., Perkin Trans. 1 1992, 2111.
(23) Schmid, C. R.; Bryant, J. D. Org. Synth. 1995, 72, 6.
(24) Roush, W. R.; Straub, J. A.; Van Nieuwenhze, M. S. J. Org. Chem.
1991, 56, 1636.
and, for this reason, these substrates were also included in the
study. Unmasking of the carbonyl group in 3 is without doubt
(25) Keck, G. E.; Murry, J. A. J. Org. Chem. 1991, 56, 6606.

Allylations of CR- or Câ-Substituted Aldehydes
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1933
Table 2. Effect of pH on Rate and Diastereoselectivity of
Allylindium Additions in Water at 25 °Ca
Scheme 2
product ratio
reaction
entry
aldehyde
pH
time, h
syn
anti
yield, %
1
33
34
5
35
36
9
1
b
3.5
5.5
4.0
3.0
12.5
4.0
24-30
48
0.5
1
1
1
1
1
3.9
3.0
3.0
1.2
1.4
1.5
1
90
85
80
92
84
86
90-95
80-87
85-88
7.0
4.0
b
7.0
4.0
b
2
3
1
2.3
2.0
10.0
37
38
7.0
4.0
1
1
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments. b pH not
controlled.
condensations does not seem to have been previously recognized
and was therefore explored more fully in order to elucidate the
accompanying advantages or disadvantages. As indicated in
Table 2, aldehydes 1-3 were closely scrutinized.
a relatively slow process in H2O, since reaction times of 20-
30 h were necessary to complete the consumption of starting
material (entries 9 and 10). For 5, allylation was complete in
much less time (entry 17). No reaction was observed with 3 in
THF unless the reaction mixture was heated. The product
distribution in both examples was now in favor of the syn
diastereomer, the crossover suggesting that chelate control may
now be operating. For 5, the syn/anti ratio (9.8) is appreciably
higher than that observed for 3 (2.3). Evidently, 3 is not
undergoing condensation as the free R-hydroxy aldehyde. The
results provided by the MOM ether 4 (entries 13-16) hold
interest because syn selectivity is operational. When comparison
is made with the O-benzyl ether 2, the added oxygen in 4 is
seen to exert a respectable directing effect as a consequence of
its chelating, or perhaps more accurately co-chelating, ability.
However, this capacity falls somewhat short of that exhibited
by the ring-opened tautomer of 3 where the side chain features
a terminal hydroxyl. The impressive role of OH substituents
becomes still more apparent when they reside directly adjacent
to the carbonyl as in 5 and 6.
Extensive studies were not performed on 5 because this
substance is difficult to obtain pure and is quite prone to
polymerization. On the other hand, D-arabinose (6) does not
suffer from these drawbacks, although its solubility in THF is
low. The responses exhibited by 5 and 6 in H2O (entries 17
and 18) include the most powerful examples of chelate-
controlled diastereoselectivity uncovered to date for reactions
promoted by indium.16b As concerns 6, dilution with THF (entry
19) appears to erode syn selectivity to a greater extent than does
dilution with ethanol (entries 19 and 23). Sonication accelerates
the latter reaction but has little effect on product ratio (entry
24).
When the pH was maintained at 7 by controlled infusion of
sodium hydroxide solution, an increase in reaction times became
necessary to achieve complete allylation (entries 33, 35, and
37). This phenomenon could be due in part to increased dilution
due to addition of the aqueous base and not constitute a
manifestation of the pH itself. Reactions allowed to proceed
without pH control were accompanied by a progressive devel-
opment in acidity to a point below pH 4. When the allylations
were initiated at a preset pH of 4 (entries 34, 36, and 38), the
transformations took place at notably accelerated rates. As
expected on structural grounds, these conditions are notably
effective in the case of 3 where the presence of acid serves to
facilitate hydrolysis of this acetal to 5.
Noteworthily, the benzyl- and silyl-protected substrates
exhibit the same diastereoselectivities at all ranges of pH tested.
These findings dispel any concerns that product distribution
might be dependent upon pH to the point where the syn/anti
ratios would vary as the reaction progressed. As concerns 1
and 2, the significant observable associated with the develop-
ment of acidic character in the reaction mixture is an increase
in rate.
This feature of indium catalysis in water requires that care
be exercised when acid-sensitive reactants are involved. When
working with 7, for example, the progress of allylation needs
to be carefully monitored. If workup is initiated as soon as 7
is completely consumed, good yields of the homoallylic alcohols
are obtained (entry 25). Failure to act promptly allows for
extensive deprotection of the acetal. Epoxide 8 also exhibits
degradation if left unattended, but to a lesser extent. Conse-
quently, these considerations must be taken into account when
designing alternative applications of this chemistry.
Salt Effects. Although an increase in acidity per se does
not impact on product stereoselectivity, it became of interest to
examine if the addition of salts to the reaction mixture would
affect product distribution. It is known that, for Diels-Alder
cycloadditions performed in water, the presence of salts
increases the amount of endo product due to an increase in the
internal pressure of the system.27 Were the reaction volumes
for formation of the syn and anti homoallylic alcohols to differ
comparably, the possibility exists that product ratios could be
conveniently manipulated to synthetic advantage in this manner.
The pair of representative aldehydes selected for study were
the MOM-protected derivative 4 and D-arabinose (6). The first
example exhibits only modest stereoselectivity in water (entry
pH Considerations. The rate acceleration noted above for
allylations promoted in water could, as for example with
D-arabinose, be attributed to the improved solubility of the
substrate in the aqueous medium. However, the phenomenon
persists when the solubilities of the reagents are lower in H2O
than in THF. This behavior could be explained by attributing
enhanced stability to the allylindium reagent in THF. Under
these circumstances, reactivity toward an incoming aldehyde
carbonyl would be reduced and condensation would proceed
more slowly. Indeed, an increase in reaction rate has been
observed by Araki et al.26 when progressing from benzene to
THF.
In the present study, it was noted that the pH of all allylations
performed in water or aqueous THF dropped significantly as
the reactions progressed. This aspect of indium-promoted
(26) Araki, S.; Jin, S.-J.; Idou, Y.; Butsugan, Y. Bull. Chem. Soc. Jpn.
1992, 65, 1736.
(27) Pindur, U.; Lutz, G.; Otto, G. Chem. ReV. 1993, 93, 741.

1934 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Mitzel
Table 3. Salt Effects on Indium-Promoted Allylations in Water at
25 °Ca
Table 4. Competitive Indium-Promoted Allylations in Water at 25
°Ca
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments.
a All experiments were conducted minimally in duplicate, and the
reported data represent the average of these experiments.
13), while 6 represents the most discriminating example under
standard aqueous conditions (entry 18). When arabinose was
admixed with 1 mol equiv of LiBr or MgCl2 and subjected to
conventional allylation in H2O, the syn/anti ratios fell somewhat
to 8.5:1 in both cases (Table 3, entries 39 and 40). On the
other hand, the use of tetraethylammonium bromide resulted in
a substantial increase to a record level of 13.5:1 (entry 41).
Increasing the relative amounts of this salt to 5 equiv did not
lead to a further increase in the level of syn product but did
promote a faster reaction presumably because of the ac-
companying increase in the ionic strength of the medium (entry
42). For reasons not yet understood, tetra-n-butylammonium
iodide was not comparably effective (entries 43 and 44). In
fact, the quaternary salt performed less well than the lithium
and magnesium halides.
Aldehyde 4 was unresponsive to the presence of LiBr and
MgCl2, giving rise to the same product distribution (syn/anti
2.1-2.2:1) as observed in H2O alone (entries 45 and 46).
However, tetraethylammonium bromide caused an almost 4-fold
increase in formation of the chelation-controlled product (entry
47)! As before, an increase in the relative proportion of this
salt above the 1 equiv level had no additional effect (entry 48).
Accordingly, a synergistic effect is observed when tetraethyl-
ammonium bromide is added at a modest level to the allylindium
reagent. Although the conditions of the selectivity enhancement
appear to be somewhat restrictive at this point in time, additional
studies need to be implemented before proper rationalization
of these observations can be offered.
significant differences in allylation rate between 5 and 8 (35-
fold) as well as 5 and 1 (29-fold) are very telling and indicate
that the adjacent epoxide and tert-butylsilyl groups are not at
all conducive to chelation. The somewhat more comparable
kinetic behavior of 4 agrees with the concept that a chelated
transition state competes favorably with a non-chelated alterna-
tive when a methoxymethyl protecting group is positioned R.
As pointed out by others,28,29 steric and electronic factors also
impact on relative rate. The extent to which these influences
are contributory under the present circumstances remains to be
evaluated. Notwithstanding, the results reported here show
conclusively that chelates are true intermediates in the allylation
in water of acyclic aldehydes carrying proper R-substituents.
â-Oxy-Substituted Aldehydes. As a class, â-alkoxy alde-
hydes respond with high diastereofacial selectivity to Lewis acid-
promoted condensations. These reactions include the TiCl4-
mediated addition of enolsilanes30 and allylsilanes,30,31 the use
of acidic titanium reagents,32 and recourse to other promoters
such as stannic chloride33 and boron trifluoride etherate.34 In
contrast, organometallic reagents of the RMgX, RLi, and R2-
CuLi type do not generally perform well despite the potential
for chelation control.35-39 In our view, a free hydroxyl
substituent â to a carbonyl group was considered to be
exploitable for 1,3-asymmetric induction during condensation
with allylindium reagents in water.
Competition Studies. Internal chelation, if operative, can
be expected to lend itself to more rapid conversion to product
if the chelated intermediate resides on the direct reaction
pathway.28 The increase in reaction rate is linked in turn to a
lowering in transition state energy associated with preformation
of the complex. The corollary to this analysis is that the more
selective substrates are also the more reactive.
In order to gauge the relative reactivity of selected R-oxy
aldehydes, the OMOM derivative 4 was allowed to vie
competitively with 1, 5, and 8 for a limited amount of the
allylindium reagent. The results are compiled in Table 4. It is
immediately obvious that 4 is somewhat more reactive than 1
and 8 (entries 49 and 50) but appreciably less reactive than 5
(entry 51). This ordering conforms very well with the stereo-
selectivity exhibited by these aldehydes. We have already
recognized that the presence of a free hydroxyl group as in 5 is
particularly conducive to high-level syn selectivity. The
Aldehydes 10-13 were selected because of their structural
simplicity, similarity to 1, 2, 4, and 5, and varied basicity at
the â-oxygen. If chelation were to gain importance and
nucleophilic attack were to occur from the less hindered
diastereotopic π-face of the aldehyde carbonyl, then anti adduct
(29) Das, G.; Thornton, E. R. J. Am. Chem. Soc. 1993, 115, 1302.
(30) (a) Reetz, M. T.; Kesseler, K.; Jung, A. Tetrahedron Lett. 1984,
25, 729. (b) Reetz, M. T.; Kesseler, K. J. Chem. Soc., Chem. Commun.
1984, 1079. (c) Reetz, M. T.; Jung, A. J. Am. Chem. Soc. 1983, 105, 4833.
(31) Keck, G. E.; Park, M.; Krishnamurthy, D. J. Org. Chem. 1993, 58,
3787.
(32) Reetz, M. T.; Jung, A. J. Am. Chem. Soc. 1983, 105, 4833.
(33) (a) Kiyooka, S.; Heathcock, C. H. Tetrahedron Lett. 1983, 24, 4765.
(b) Heathcock, C. H.; Kiyooka, S.; Blumenkopf, T. A. J. Org. Chem. 1984,
49, 4214.
(34) (a) Keck, G. E.; Abbott, D. E. Tetrahedron Lett. 1984, 25, 1883.
(b) Reetz, M. T.; Jung, A.; Bolm, C. Tetrahedron 1988, 44, 3889.
(35) Reetz, M. T.; Kesseler, K.; Schmidtberger, S.; Wenderoth, B.;
Steinbach, R. Angew. Chem., Int. Ed. Engl. 1983, 22, 989.
(36) Leitereg, T. J.; Cram, D. J. J. Am. Chem. Soc. 1968, 90, 4019.
(37) Still, W. C.; Schneider, J. A. Tetrahedron Lett. 1980, 21, 1035.
(38) Davis, A. P.; Jaspars, M. J. Chem. Soc., Perkin Trans. 1 1992, 2111.
(39) Braun, M.; Mahler, H. Liebigs Ann. 1995, 29.
(28) (a) Chen, X.; Hortelano, E. R.; Eliel, E. L.; Frye, S. V. J. Am. Chem.
Soc. 1990, 112, 6130; 1992, 114, 1778. (b) Frye, S. V.; Eliel, E. L.; Cloux,
R. J. Am. Chem. Soc. 1987, 109, 1862.

Allylations of CR- or Câ-Substituted Aldehydes
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1935
Table 5. Indium-Promoted C-Allylations of â-Oxygenated
Aldehydes in Various Solvents (CH2dCHCH2Br, 25 °C)a
Mechanistic Considerations. Additions of the allylindium
reagent to R- and â-hydroxy aldehydes in water have been
demonstrated to be highly stereoselective and synthetically
useful operations. The corresponding methoxy and MOM
derivatives exhibit comparable properties, although to a de-
monstrably lessened degree. The free hydroxyl derivatives
represent the more reactive substrates in either series, this
reactivity ordering conforming expectedly to chelation-controlled
addition. This ability of the indium cation to lock the carbonyl
substrate conformationally prior to nucleophilic attack is indica-
tive that coordination to the substrate can indeed overcome the
H2O solvation forces, especially when the neighboring func-
tionality is an unprotected hydroxyl substituent.
The sense of asymmetric induction in the R-series, viz. a
strong kinetic preference for formation of the syn diol, is
consistent with operation of the classic Cram model as in A.
Once complexation occurs, the allyl group is transferred to the
carbonyl carbon from the less hindered π-surface opposite to
that occupied from the R group. In B, the chelation pathway
a All of the reactions were performed at least in duplicate at a
concentration of 0.1 M with vigorous stirring for the indicated time
1
span. The product distributions in all cases were determined by H
NMR integration at 300 MHz. b The THF solution of allyl bromide
was heated to reflux with the indium prior to reaction. The mixture
was cooled prior to introduction of the aldehyde and the coupling was
performed as in a.
14 would result. Homoallylic alcohols 15 are the Felkin-Anh
(non-chelation-controlled) products.
The diastereomeric ratios of 14 and 15 resulting from
exposure of 10-13 to allyl bromide and indium in solvents
ranging from anhydrous THF to water are compiled in Table
5. The product distributions are seen to correlate closely with
is seen to be capable of adoption of a chair confomation which
concisely accommodates favored formation of the syn diol. The
reversal in stereoselectivity in going from 1 to 5 in the same
aqueous environment is the classical test for chelation.
For the â-chelate reactions, the factors which influence
product formation appear to be the same. When C forms,
intramolecular attack is guided to occur syn to the preexisting
hydroxyl. This reaction trajectory leads preferentially to the
anti diol, provided that a chairlike transition state approximating
D is followed.40 Importantly, it is one single allylindium that
the R-alkoxy series. Thus, a â-methoxyl substituent is capable
of modest levels of chelation control, irrespective of whether
the allylation is performed in an anhydrous or aqueous medium
(entries 62-65). Since a benzyloxy or a tert-butyldimethyl-
siloxy group results in production of totally stereorandom
mixtures of 14 and 15 (entries 55-60), there is no evidence
for transient structural rigidification prior to nucleophilic attack
in these examples. Aldehyde 12 actually produces syn isomer
15c preferentially to 14c when THF is the reaction medium
(entry 61). This crossover could reflect the operation of a steric
effect.
The unprotected hydroxyl derivative 10 exhibits the most
pronounced face selectivity as anticipated (entries 52 and 53).
In fact, the 8.5:1 ratio of 14a to 15a compares quite favorably
with the product distributions exhibited by the R-hydroxy
aldehydes 5 and 6. Clearly the free â-OH group is capable of
chelation control in water, finding it possible to coordinate to
the indium ion despite its preexisting solvation by water
molecules.
chelates and reacts. Although similar working models have been
advanced in explanation of the mode of addition of titanium32
or borane reagents,41 this behavior is distinct from other
chelation-controlled reactions where the reacting reagent is
different from the chelating agent. This may well be an
argument that the indium-mediated reaction takes place on the
metal surface.
The present studies have demonstrated a direct kinetic link
between stereoselectivity and the presence of a neighboring
hydroxyl group. While this relationship has been extensively
discussed,28,42 the support of this concept is not universal.
Several experimental and theoretical reports have appeared
supporting the notion that π-complexation is not a kinetically
(40) Mori, S.; Nakamura, M.; Nakamura, E.; Koga, N.; Morukuma, K.
J. Am. Chem. Soc. 1995, 117, 5055.
(41) (a) Narasaka, K.; Pai, H. C. Chem. Lett. 1980, 1415. (b) Narasaka,
K.; Pai, H. C. Tetrahedron 1984, 12, 2233. (c) Evans, D. A.; Chapman, K.
T.; Carreira, E. M. J. Am. Chem. Soc. 1988, 110, 3560.
(42) Corcoran, R. C.; Ma, J. J. Am. Chem. Soc. 1992, 114, 4536.

1936 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
Paquette and Mitzel
important event.40,43 Clearly, additional studies of this entire
question would be welcomed.
(75 MHz, CDCl3) δ 97.2, 93.7, 87.4, 39.3, 28.9, 28.7, 26.3, 25.7. 25.6;
MS m/z (M+) calcd 172.1099, obsd 172.1093.
r-(Methoxymethoxy)cyclohexaneacetaldehyde (4). A magneti-
cally stirred solution of 9 (900 mg, 6.42 mmol) in a 1:1 mixture of
THF and DMF (30 mL) was treated with sodium hydride (277 mg,
11.56 mmol) and chloromethyl methyl ether (616 mg, 7.70 mmol),
and the reaction was allowed to proceed at 20 °C for 48 h. Water (25
mL) and CH2Cl2 (25 mL) were introduced, and the separated aqueous
layer was extracted with CH2Cl2 (2 × 30 mL). The combined organic
phases were dried and concentrated, and the residue was purified by
chromatography on silica gel (elution with 10:1 hexanes-ethyl acetate)
to give the protected alcohol as a colorless oil (945 mg, 80%): 1H
NMR (300 MHz, CDCl3) δ 5.70-5.58 (m, 1 H), 5.25-5.11 (m, 2 H),
4.69 (d, J ) 6.7 Hz, 1 H), 4.50 (d, J ) 6.7 Hz, 1 H), 3.70 (t, J ) 7.4
Hz, 1 H), 3.36 (s, 3 H), 1.94-1.00 (m, 11 H); 13C NMR (75 MHz,
CDCl3) δ 136.9, 118.0, 93.7, 82.5, 55.4, 42.2, 29.0, 26.6, 26.1, 26.0.
Despite the considerable success of the present investigation,
the precise mechanism of indium-promoted reactions remains
unclear. In the mid-1980s, the involvement of radical pairs was
advanced in explanation of tin-promoted allylations.44 Subse-
quent recourse to radical clock experiments demonstrated
unambiguously that radicals could not be involved.45 We have
come to favor a single electron transfer process similar to that
advanced by Chan.13b According to this reaction profile, the
allyl bromide approaches the surface of the indium metal where
the SET process generates the reactive radical anion/indium
radical cation pair E. These conditions operate, of course, only
A 630 mg (3.38 mmol) sample of this ether was ozonolyzed in the
predescribed manner and purified by flash chromatography on silica
gel to give 4 as an unstable colorless oil (472 mg, 75%): 1H NMR
(300 MHz, CDCl3) δ 9.62 (d, J ) 2.5 Hz, 1 H), 4.71 (d, J ) 6.8 Hz,
1 H), 4.66 (d, J ) 6.8 Hz, 1 H), 3.65 (dd, J ) 5.5, 3.1 Hz, 1 H), 3.39
(s, 3 H), 1.81-1.64 (m, 6 H), 1.28-1.17 (m, 5 H); 13C NMR (75 MHz,
CDCl3) δ 203.6, 97.0, 86.5, 55.9, 39.5, 28.9, 27.8, 26.1, 26.0, 25.9.
Prototypical Allylation Reactions. A. In H2O. A magnetically
stirred solution of 4 (150 mg, 0.806 mmol) in water (8.9 mL) was
treated with indium powder (101 mg, 0.887 mmol) and allyl bromide
(145 mg, 1.21 mmol). The reaction was allowed to proceed until no
4 remained (TLC analysis). Ethyl acetate was added, stirring was
maintained for 60 min, and the separated aqueous phase was extracted
with ethyl acetate (2 × 20 mL). The combined organic layers were
dried and evaporated. The residue was taken up in anhydrous methanol
(15 mL) containing a few milligrams of p-toluenesulfonic acid and
refluxed for 12 h to provide the diol. The methanol was removed in
vacuo and replaced by acetone (15 mL). The resulting solution was
stirred for 6 h and concentrated to leave an oil, purification of which
was accomplished by flash chromatography on silica gel (elution with
50:1 hexanes-ethyl acetate) to give the acetonide in 80-95% yield
(Table 1).
The determination of diastereomer composition was performed as
described by Keck.25 The key identifying NMR signals are as
follows: syn, dd at δ 3.46 and 13C absorptions at 107.9, 84.5, and 78.0
ppm. For the anti isomer: dd at δ 3.73 and 13C peaks at 107.2, 82.3,
and 76.9 ppm.
Comparable processing of 7 (200 mg, 1.54 mmol) afforded 207 mg
(83%) of a 1:3.2 mixture of syn and anti alcohols.50 Syn isomer: 1H
NMR (250 MHz, CDCl3) δ 5.93-5.76 (m, 1 H), 5.12 (d, J ) 17 Hz,
1 H), 5.11 (d, J ) 10.5 Hz, 1 H), 4.07-3.95 (m, 2 H), 3.78-3.67 (m,
1 H), 3.58 (quintet, J ) 6.3 Hz, 1 H), 3.27-2.19 (series of m, 3 H),
1.43 (s, 3 H), 1.36 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 134.2, 117.8,
109.4, 78.5, 71.5, 66.0, 38.2, 27.0, 25.3. Anti isomer: 1H NMR (250
MHz, CDCl3) δ 5.92-5.75 (m, 1 H), 5.25 (d, J ) 16 Hz, 1 H), 5.13
(d, J ) 10.5 Hz, 1 H), 4.05-3.87 (m, 3 H) 3.77 (dq, J ) 8.8, 4.4 Hz,
1 H), 2.43-2.10 (m, 2 H), 2.00 (d, J ) 3.4 Hz, 1 H), 1.43 (s, 3 H),
1.37 (s, 3); 13C NMR (75 MHz, CDCl3) δ 133.9, 118.2, 109.0, 78.1,
70.4, 65.2, 37.6, 26.5, 25.2.
when indium metal is present as a reactant. Acyclic diastereo-
facial control is presently recognized to occur in a wide range
of reactions.46,47 Suffice it to indicate at this point that the
preformation of allylindium reagents may well bypass the
involvement of E, suggesting an alternative pathway involving
the more conventional species F can also operate.17a,48 Proper
selection of reaction conditions could alter the precise pathway
at work.
Experimental Section49
5-Cyclohexyl-1,3-dioxolan-4-ol (3). A solution of vinylmagnesium
bromide in THF [from 11.36 g (107.2 mmol) of vinyl bromide] was
treated dropwise with cyclohexanecarboxaldehyde (3.00 g, 26.8 mmol).
The reaction mixture was refluxed for 5 h, cooled to 20 °C, treated
with 1 N HCl, and extracted with ether (4 × 50 mL). The combined
organic layers were washed sequentially with 1 N HCl, water, and brine,
then dried and evaporated. The residue was purified by chromatography
on silica gel (elution with 4:1 hexanes-ethyl acetate) to give 9 as a
colorless oil (3.60 g, 94%): 1H NMR (300 MHz, CDCl3) δ 5.85 (m,
1 H), 5.13 (m, 1 H), 3.83 (t, J ) 6.4 Hz, 1 H), 1.86-0.95 (series of m,
11 H); 13C NMR (75 MHz, CDCl3) δ 139.8, 115.4, 76.6, 43.5, 29.0,
28.3, 26.5, 26.1, 26.0.
A solution of 9 (1.00 g, 7.14 mmol) in CH2Cl2 (65 mL) was cooled
to -78 °C, ozonolyzed for 15 min, purged with oxygen, and treated
with dimethyl sulfide (12 mL). After 30 min, the cooling bath was
removed and the reaction mixture was stirred at 20 °C for 15 h prior
to solvent evaporation. Flash chromatographic purification (silica gel,
elution with 5:1 hexanes-ethyl acetate) gave 3 as a colorless oily
diasteromeric mixture (858 mg, 70%): 1H NMR (300 MHz, CDCl3) δ
5.25 (d, J ) 2.9 Hz, 1 H), 5.08 (d, J ) 9.8 Hz, 2 H), 3.50 (d, J ) 2.9
Hz, 1 H), 3.34 (br s, 1 H), 1.83-1.01 (series of m, 11 H); 13C NMR
Analogous treatment of 8 (100 mg, 0.463 mmol) gave rise to 94 mg
(78%) of a 1:2 mixture of syn and anti alcohols.24 Syn isomer: 1H
NMR (300 MHz, CDCl3) δ 5.82 (m, 1 H), 5.15 (m, 2 H), 3.78 (m, 2
H), 3.59 (m, 1 H), 3.19 (m, 1 H), 2.98 (dd, J ) 4.4, 7.5 Hz, 1 H), 2.37
(t, J ) 6.7 Hz, 2 H), 2.21 (br s, 1 H), 0.91 (s, 9 H), 0.10 (s, 3 H), 0.09
(s, 3 H); 13C NMR (75 MHz, CDCl3) δ 133.2, 118.3, 68.9, 61.6, 59.8,
57.6, 38.7, 25.8 (3 C), 18.2, -5.3, -5.4. Anti isomer: 1H NMR (300
MHz, CDCl3) δ 5.89 (m, 1 H), 5.25 (m, 2 H), 4.01 (dd, J ) 5.6, 11.6
Hz, 1 H), 3.71 (dd, J ) 6.5, 11.6 Hz, 1 H), 3.57 (m, 1 H), 3.14 (m, 1
H), 2.94 (dd, J ) 4.3, 7.9 Hz, 1 H), 2.69 (d, J ) 1.7 Hz, 1 H), 2.42 (m,
2 H), 0.89 (s, 9 H), 0.09 (s, 3 H), 0.08 (s, 3 H); 13C NMR (75 MHz,
CDCl3) δ 133.5, 118.1, 69.2, 62.0, 58.3, 55.4, 39.4, 25.8 (3 C), 18.2,
-5.3, -5.5.
(43) (a) Poll, T.; Metter, J. O.; Helmchen, G. Angew. Chem., Int. Ed.
Engl. 1985, 24, 112. (b) Buhro, W. E.; Georgiou, S.; Ferna´ndez, J. M.;
Patton, A. T.; Strouse, C. E.; Gladysz, J. A. Organometallics 1986, 5, 956.
(c) Arai, M.; Kawasuji, T.; Nakamura, E. J. Org. Chem. 1993, 58, 5121.
(44) (a) Petrier, C.; Luche, J. L. J. Org. Chem. 1985, 50, 910. (b) Einhorn,
C.; Luche, J. L. J. Organomet. Chem. 1987, 322, 177.
(45) Wilson, S. R.; Guazzaroni, M. E. J. Org. Chem. 1989, 54, 3087.
(46) Porter, N. A.; Giese, B.; Curran, D. P. Acc. Chem. Res. 1991, 24,
296.
(47) Smadja, W. Synlett 1994, 1.
(48) Marshall, J. A.; Hinkle, K. W. J. Org. Chem. 1995, 60, 1920.
(49) For generic experimental details, see: Paquette, L. A.; Lobben, P.
C. J. Am. Chem. Soc. 1996, 118, 1917.
(50) (a) Hoffmann, R. W.; Endesfelder, A.; Zeiss, H.-J. Carbohydrate
Res. 1983, 23, 320. (b) Roush, W. R.; Walts, A. E.; Hoong, L. K. J. Am.
Chem. Soc. 1985, 107, 8186.

Allylations of CR- or Câ-Substituted Aldehydes
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1937
B. In THF. A magnetically stirred slurry of D-arabinose (150 mg,
1.0 mmol) in THF (11 mL) was treated with indium powder (126 mg,
1.1 mmol) and allyl bromide (180 mg, 1.5 mmol). Reaction was
allowed to proceed until TLC analysis showed no residual starting
material to be present. After the evaporation of solvent, the residue
was taken up in pyridine (4 mL) and acetic anhydride (4 mL), and the
mixture was allowed to stir for 12 h prior to concentration in vacuo.
The residue was partitioned between ethyl acetate (100 mL) and water
(100 mL), and the separated aqueous phase was extracted with ethyl
acetate (2 × 50 mL). The combined organic solutions were dried and
concentrated. Chromatography of the residue on silica gel (elution with
5:1 ethyl acetate-hexanes) afforded the two diastereomeric penta-
acetates.
methyl doublets at δ 1.25 and 1.27 (in CDCl3). The 13C NMR spectra
also agreed fully with the literature data.25
From allyl bromide (260 mg, 1.66 mmol), indium metal (140 mg,
1.22 mmol), and aldehyde 13 (116 mg, 1.11 mmol), there was isolated
143 mg (91%) of unpurified homoallylic alcohols. This mixture was
dissolved in CHCl3 (50 mL), treated with iodotrimethylsilane (640 mg,
3.1 mmol) at room temperature (rt), and stirred for 12 h before water
(50 mL) was introduced. After 2 h, the aqueous layer was extracted
with CH2Cl2 (2 × 20 mL) and the combined organic layers were washed
with water and brine, dried, and evaporated. The residue was purified
by flash chromatography on silica gel (elution with 1:3 hexanes-ethyl
acetate) to give 14a and 14b in a combined yield of 120 mg (93%).
Anti isomer 14a: 1H NMR (300 MHz, CDCl3) δ 5.80 (m, 1 H), 5.14
(m, 2 H), 4.14 (m, 1 H), 3.99 (m, 1 H), 2.67 (br s, 2 H), 2.27 (m, 2 H),
1.61 (t, J ) 5.7 Hz, 2 H), 1.23 (d, J ) 6.3 Hz, 3 H); 13C NMR (75
MHz, CDCl3) δ 134.7, 118.0, 68.1, 65.3, 43.6, 41.9, 23.4. The key 1H
NMR signal for syn isomer 15a is the methyl doublet (J ) 6.3 Hz) at
δ 1.20.
Syn (threo) isomer: 1H NMR (300 MHz, CDCl3) δ 5.70 (m, 1 H),
5.41 (dd, J ) 11.1, 4.2 Hz, 1 H), 5.29 (dd, J ) 10.8, 4.2 Hz, 1 H),
5.09 (m, 2 H), 5.05 (m, 1 H), 5.03 (m, 1 H), 4.23 (dd, J ) 12.4, 3.4
Hz, 1 H), 4.10 (dd, J ) 12.4, 5.4 Hz, 1 H), 2.39-2.26 (m, 2 H), 2.11
(s, 3 H), 2.06 (s, 3 H), 2.052 (s, 3 H), 2.048 (s, 3 H), 2.04 (s, 3 H); 13
C
NMR (75 MHz, CDCl3) δ 170.4, 170.1, 169.8 (2 C), 169.7, 132.0,
118.9, 70.7, 70.4, 68.9, 68.7, 61.5, 35.1, 20.8, 20.7, 20.6, 20.5 (2 C).
Anti (erythro) isomer: 1H NMR (300 MHz, CDCl3) δ 5.71 (m, 1
H), 5.43 (dd, J ) 11.3, 2.4 Hz, 1 H), 5.29 (dd, J ) 10.7, 2.4 Hz, 1 H),
5.07 (m, 2 H), 5.03 (m, 1 H), 5.02 (m, 1 H), 4.21 (dd, J ) 12.5, 2.8
Hz, 1 H), 4.06 (dd, J ) 12.5, 5.4 Hz, 1 H), 2.39-2.28 (m, 2 H), 2.082
In those examples where the silylated aldehyde 12 was studied, the
homoallylic alcohol mixture was deprotected by stirring with a catalytic
quantity of p-toluenesulfonic acid in methanol51 at rt for 90 min. The
solvent was removed in vacuo, and the resulting diols were separated
by flash chromatography on silica gel (elution with 1:2 hexanes-ethyl
acetate). The syn and anti isomers were identified by comparison of
(s, 3 H), 2.078 (s, 3 H), 2.05 (s, 3 H), 2.04 (s, 3 H), 2.01 (s, 3 H); 13
C
1
their H NMR spectra with literature data.34b
NMR (75 MHz, CDCl3) δ 170.3, 170.0, 169.7, 169.62, 169.59, 131.9,
118.8, 70.6, 70.2, 68.8, 68.6, 61.3, 35.0, 20.7, 20.6, 20.50 (2 C), 20.4.
Reactions performed in 1:1 H2O-THF were carried out in analogous
fashion.
Competition Experiments. To a mixture of 1 mmol of each of
two aldehydes in water (11 mL) was added 1.1 mmol of powdered
indium metal and 1.5 mmol of allyl bromide. The reaction mixture
was stirred at rt for approximately 5 h. Ethyl acetate (50 mL) was
introduced, and after a period of vigorous mixing, the aqueous layer
was separated and extracted with additional ethyl acetate (3 × 15 mL).
The combined organic layers were dried and concentrated, and the
products were separated by flash chromatography on silica gel as
described above. The product ratios given in Table 4 are based upon
the weights of isolated homoallylic alcohols.
C. With Preformation of the Allylindium Reagent in THF. To
a magnetically stirred solution of allyl bromide (82 mg, 0.68 mmol) in
dry THF (5 mL) was added powdered indium metal (58 mg, 0.50
mmol), and the mixture was treated at reflux for 1 h and allowed to
cool to rt. An 80 mg (0.45 mmol) sample of 11 was introduced, and
the progress of reaction was monitored by TLC, which indicated the
reaction to be complete after 8 h. After water had been added, stirring
was prolonged for 5 min prior to extraction of the product into ethyl
acetate (3 × 15 mL). The combined organic layers were dried and
evaporated, and the resulting residue was purified by flash chroma-
tography on silica gel (elution with 5:1 hexanes-ethyl acetate) to give
the mixture of homoallylic alcohols as a colorless oil (82 mg, 82%).
Diastereomers 14b and 15b were distinguished by 1H NMR spectros-
copy, and the isomer distribution was quantified by integration of the
Acknowledgment. Funds in support of this research were
generously provided by the Emissions Reduction Research
Center at the New Jersey Institute of Technology.
JA953682C
(51) Corey, E. J.; Venkateswarlu, A. J. Am. Chem. Soc. 1972, 94, 6190.