Solvent-Mediated Allylation of Carbonyl Compounds with Allylic Stannanes

Article abstract of DOI:10.1021/jo962332l

Methanol promotes the addition of allyltrimethylstannane (1a) to isobutyraldehyde (2a, 30 °C) yielding the corresponding homoallylic alcohol (3a), without the necessity for added catalyst. The corresponding reaction of aldehydes 2a-e or activated ketone 2

Full text of DOI:10.1021/jo962332l

J . Org. Chem. 1997, 62, 1961-1964
1961
Solven t-Med ia ted Allyla tion of Ca r bon yl Com p ou n d s w ith
Allylic Sta n n a n es
Teresa M. Cokley,† Peta J . Harvey,† Raymond L. Marshall,† Adam McCluskey,‡ and
David J . Young*,†
Faculty of Science and Technology, Griffith University, Nathan 4111, Brisbane, Australia and
Department of Chemistry, University of Newcastle, Callaghan NSW 2308, Australia
Received December 12, 1996X
Methanol promotes the addition of allyltrimethylstannane (1a ) to isobutyraldehyde (2a , 30 °C)
yielding the corresponding homoallylic alcohol (3a ), without the necessity for added catalyst. The
corresponding reaction of aldehydes 2a -e or activated ketone 2f with tetraallyltin (1b, 0.25 equiv)
is substantially faster and proceeds in high yield (81-98%) and with easy separation of the product
from tin residues. Aliphatic ketones 2g and 2h also react, but require more forcing conditions.
Competitive experiments involving equimolar mixtures of selected aldehydes and ketones with 1b
indicates very high aldehyde chemoselectivity. The reaction of 1b with aldehydes proceeds slowly
at first, followed by a rapid acceleration which may be attributable to a build up of partially soluble
tin(IV) methoxide. The increased rate of carbonyl allylation by 1a and 1b in methanol, relative to
dimethyl sulfoxide, suggests that the primary activating influence of the solvent is via hydrogen
bonding to the carbonyl oxygen. There is no NMR spectroscopic evidence for a significant change
in the ground state structure of these allylic stannanes in methanol, relative to other solvents.
Ta ble 1. Allyla tion of Ald eh yd e 2a w ith
In tr od u ction
Allyltr im eth ylsta n n a n e 1a in Differ en t Solven tsa
The addition of allylic metal compounds to aldehydes
and ketones to yield homoallylic alcohols is a useful
transformation in organic synthesis and consequently has
received considerable attention in recent years.1 The
reaction is synthetically analogous to the aldol condensa-
tion but allows for the subsequent introduction of a
variety of alternative functional groups by manipulation
of the alkene moiety.2 Like the aldol reaction, addition
can be achieved with high levels of regio- and stereose-
lectively by judicious choice of substrates and reaction
conditions.3 Allylic stannanes offer an attractive com-
bination of configurational stability with relatively high
reactivity and have been extensively employed for the
allylation of aldehydes, in particular.4 Other than for
particularly reactive aldehydes (e.g. chloral), some form
of promotion is usually required such as heat,5 high
pressure6 or, more commonly, activation of the aldehyde
with a Lewis acid.7 Brønsted acids have also been
employed for this purpose, and of particular relevance
to this paper is a report of the allylation of carbonyl
compounds with tetraallyltin in acidic aqueous media.8
This procedure involved the use of 1 equiv of HCl (relative
to tetraallyltin) in aqueous THF and resulted in the
transfer of all four allyl groups with high chemoselectivity
carbonyl
compound
conversionb
(%)
entry
solvent
product
1
2
3
2a
2a
2a
CD2Cl2
CD3OD
(CD3)2SO
3a
3a
3a
3
73
5
a
Reactions conducted at a concentration of 1.0 M in each
reagent at 30 °C for 8 days. b Determined by 1H NMR spectroscopy.
toward aldehydes relative to ketones, esters, and acyl
chlorides and discrimination between different types of
ketone.
Also pertinent to the present discussion is that allylic,9
propargylic,10 and indenyl11 stannanes are configuration-
ally unstable in methanol and other polar solvents,
undergoing a facile allylic isomerization at ambient
temperatures. Speculation concerning the mechanism of
this process has invoked a solvent stabilized ion-pair
intermediate. It occurred to us that such an intermediate
might be trapped with an electrophile, and we have
recently communicated12 that aldehydes react with tet-
raallyltin in methanol and other polar solvents at ambi-
ent temperatures to provide the corresponding homoal-
lylic alcohols in high isolated yield. This reaction is
comparable to the corresponding aqueous HCl-promoted
process in that all four allyl groups are transferred
quantitatively but require no added activating agent. We
now report on aspects of the mechanism and chemose-
lectivity of this unusual effect of solvent on the addition
of allylic stannanes to carbonyl compounds.
† Griffith University. email: D.Young@sct.gu.edu.au.
‡ University of Newcastle.
X Abstract published in Advance ACS Abstracts, March 15, 1997.
(1) For a review see: Yamamoto, Y.; Asao, N. Chem. Rev. 1993, 93,
2207.
(2) For a recent example see: Marshall, J . A.; Hinkle, K. W. J . Org.
Chem. 1996, 61, 4247.
(3) For a recent example see: Carey, J . S.; Coulter, T. S.; Hallett,
D. J .; Maguire, R. J .; McNeill, A. H.; Stanway, S. J .; Teerawutgulrag,
A.; Thomas, E. J . Pure Appl. Chem. 1996, 68, 707.
(4) Yamamoto, Y. Acc. Chem. Res. 1987, 20, 243.
(5) McNeill, A. H.; Thomas, E. J . Tetrahedron Lett. 1990, 31, 6239
and references therein.
(6) (a) Yamamoto, Y.; Maruyama, K.; Matsumoto, K. J . Chem. Soc.,
Chem. Commun. 1983, 489. (b) Isaacs, N. S.; Marshall, R. L.; Young,
D. J . Tetrahedron Lett. 1992, 33, 3023.
(9) (a) Young, D.; Kitching, W. Silicon, Germanium, Tin Lead
Compd. 1986, 9, 67. (b) Verdone, J . A.; Margravite, J . A.; Scarpa, N.
M.; Kuivila, H. G. J . Am. Chem. Soc. 1975, 97, 843.
(10) Lequan, M. M.; Guillerm, G. C. R. Hebd. Seances Acad. Sci.
Ser. C 1969, 268, 858.
(11) Beletskaya, I. P.; Kashin, A. N.; Reutov, O. A. J . Organomet.
Chem. 1978, 155, 31.
(7) Marshall, J . A. Chem. Rev. 1996, 96, 31.
(8) Yanagisawa, A.; Inoue, H.; Morodome, M.; Yamamoto, H. J . Am.
Chem. Soc. 1993, 115, 10356.
(12) Cokley, T. M.; Marshall, R. L.; McCluskey, A.; Young, D. J .
Tetrahedron Lett. 1996, 37, 1905.
S0022-3263(96)02332-8 CCC: $14.00 © 1997 American Chemical Society

1962 J . Org. Chem., Vol. 62, No. 7, 1997
Ta ble 2. Allyla tion of Ald eh yd es 2 (R2 ) H) w ith Tetr a a llytin 1b in
Cokley et al.
Differ en t Solven tsa
Ta ble 3. Allyla tion of Keton es 2 w ith Tetr a a llyltin (1b)
in Meth a n ola
carbonyl
entry compound
3
R1
R2
product yield (%)b
1
2
3
4
5
2f
2g
2h
2i
CH3
CH3
CO2Et
C2H5
3fc
3g
3h
3i
95
74
76
<1
-(CH2)5-
-CHdCH(CH2)3-
CH3 Ph
h to reach completion, depending upon the aldehyde and
also the purity of 1b. Reactions involving 1b which had
developed a faint cloudiness on standing, proceeded faster
than reactions in which 1b had been distilled im-
mediately prior to use, although all reactions were
complete within 20 h and typically under 4 h. The
product could be isolated either by aqueous workup or
by evaporation of the solvent and washing the resulting
slurry of 3 and tin salts with dichloromethane followed
by kugelrohr distillation or filtration through silica gel.
This simple workup is in marked contrast to the tedious
separation of product from organotin residues often
associated with the use of allyltrialkylstannanes.
2j
3j
(ca 20)
a
Reactions conducted at reflux for 24 h unless otherwise stated.
b
Isolated yields, the value in parentheses is a percentage yield
determined by H NMR spectroscopy. c Conducted at 30 °C for 16
1
h. The product was hydrolyzed and isolated as the carboxylic acid.
Ta ble 4. Com p etitive Allyla tion of Ca r bon yl Com p ou n d s
(2) w ith Tetr a a llyltin (1b) in Meth a n ola
entry
aldehyde
ketone
ratiob
1
2
3
2a
2c
2c
2g
2h
2j
>98:<2
95:5
96:4
a
Conducted with aldehyde (1.0 M), ketone (1.0 M), and 1b (0.25
Again, allylation in dichloromethane (Table 2, entry
2) or dimethyl sulfoxide (entry 7) was substantially
slower, although reaction in the latter solvent did proceed
to completion (1H NMR) and in 70% isolated yield over a
longer period (63 h). Only reactive aldehyde 2e under-
went complete allylation in dichloromethane and yielded
the expected Cram (erythro, de ) 70%) addition product13
predominantly in either dichloromethane (entry 9) or
methanol (entry 8). Ketones other than the reactive ethyl
pyruvate (2f) did not react at 30 °C, but alkyl ketones
(2g,h ) did proceed with complete conversion (1H NMR)
and in good isolated yield at reflux for 24 h while
allylation of acetophenone (2h ) did proceed to a limited
extent under these conditions (Table 3). Competitive
experiments involving equimolar mixtures of selected
M) at 30 °C for 16 h. Determined by 1H and
spectroscopy.
C NMR
b
13
Resu lts
We first examined the reaction of allyltrimethylstan-
nane (1a ) with isobutyraldehyde (2a , 1 equiv, 30 °C) in
different solvents (Table 1). The reactions were moni-
tored by 1H NMR spectroscopy for 8 days and while little
conversion was observed in dichloromethane or dimethyl
sulfoxide over this period, a significant amount of ho-
moallylic alcohol 3a was formed in methanol with tri-
methyltin methoxide as the only other product.
More useful from a synthetic perspective were the
reactions of tetraallyltin (1b, 0.25 equiv) with aldehydes
2 in polar solvents, and particularly in methanol. These
reactions proceeded at ambient temperature (30 °C) and
provided the corresponding homoallyl alcohols 3 in good
yield (Table 2). Reactions required approximately 2-20
(13) Roush, W. R.; Walts, A. E.; Hoong, L. K. J . Am. Chem. Soc.
1985, 107, 8186.

Solvent-Mediated Allylation of Carbonyl Compounds
J . Org. Chem., Vol. 62, No. 7, 1997 1963
Ta ble 5. Th e Effect of Solven t on Selected NMR P a r a m eter s for Or ga n osta n n a n esa
compound
NMR parameter
CD2Cl2
(CD3)2SO
CD3OD
HCONH2
1b
1b
1a
δ
119Sn
-49.4
264
331
-47.1
261
326
-51.1
267
329
-48.6
1J (119Sn-13CH2)
1J (119Sn-13CH3)
1J (119Sn-13CH3)
b
b
b
(CH3)4Sn
336
332
339
a
b
All spectra obtained at a concentration of 0.2 M. Could not be accurately determined in this non-deuterated solvent.
ence compound tetramethyltin in different solvents are
included in Table 5. These exhibited surprisingly little
variation with solvent.
Discu ssion
While it is interesting to note that methanol promotes
the addition of allyltrimethylstannane (1a ) to aldehydes,
this reaction is too slow to be of practical use. The
allylation of aldehydes with 1b, however, occurs at a
convenient rate, proceeds with clean and quantitative
transfer of all four allyl groups, and is very simple to
conduct and to monitor by NMR spectroscopy. Also
important from a synthetic perspective is that the result-
ing homoallylic alcohols can be easily separated from the
relatively insoluble and involatile polymeric tin salts.
Such mild reaction conditions should be compatable with
a wide variety of other functional groups, although it was
noted that substrate 2f did undergo some transesterifi-
cation to yield a small amount of the corresponding
methyl ester over the period of the reaction and conse-
F igu r e 1. Allylation of aldehyde 2a (1.0 M) with tetraallyl-
stannane (1b, 0.25 M) in CD3OD at 30 °C.
quently the product 3f was hydrolyzed and isolated as
the carboxylic acid.
This allylation procedure is also highly chemoselective
as indicated by the varying reaction times and temper-
atures required for different types of carbonyl compounds
(Tables 2 and 3) and the high level of discrimination
between aldehydes and ketones (Table 4). This degree
of chemoselectivity is not possible with other more
commonly used allylating reagents such as allylmagne-
sium bromide, allyllithium, or allyltributyltin/BF3‚OEt2.8
The diastereoselectivity observed on reaction with chiral
aldehyde 2e is better than that obtained with some
achiral allylmagnesium, titanium, and -chromium re-
agents,15 but less than that obtained with chiral allyl-
stannane16 and -boron13 reagents. The diastereomers of
3e are, however, easily separated by derivatization and
chromatography17 and, in our experience, the high chemi-
cal yield and convenience of the present procedure
compensates for the minimal effort of the subsequent
separation.
Our NMR spectroscopic study of the ground state
structure of 1a and 1b in different solvents provided no
evidence for the solvent-stabilized ion-pair species pro-
posed as an intermediate in the isomerization of allylic
stannanes.9-11 If such a species is in equilibrium with
the tetrahedral allylic stannanes in polar solvents, it can
only be present as a minor component undergoing
exchange at a rate faster than the NMR timescale. The
observation that both reaction of 1a and of 1b are
accelerated in methanol to a greater extent than in the
more polar, aprotic dimethyl sulfoxide also suggests that
the primary influence of the solvent is not to activate the
aldehydes and ketones with 1b (0.25 equiv) at 30 °C in
methanol (Table 4) indicated very high aldehyde selectiv-
ity.
The allylation of carbonyl compounds 2a -f with 1b
were performed in CD3OD at 30 °C and monitored by 1H
NMR spectroscopy. The allylation of 2a was also moni-
tored by 119Sn NMR spectroscopy which revealed a
decrease in the signal for 1b (δ -51.15) and the concomi-
tant appearance of a series of signals (some broad)
beween approximately -600 and -640 ppm which is
consistent with the formation of polymeric tin(IV) meth-
oxide species.14 No intermediate allyl tin species
(allyl)4-nSn(OR)n (n ) 1-3) were observed in any of these
spectroscopic studies, indicating that transfer of the first
allyl group is slower than subsequent transfers. Each
reaction proceeded slowly before undergoing a rapid
acceleration (Figure 1). This kinetic profile was quite
reproducible, only varying in the length of the lag period
which was dependent on both the aldehyde and the
purity of 1b. The aliphatic aldehydes 2a and 2b exist
primarily (ca. 85-90%) as the hemiacetals in CD3OD
under these conditions (1.0 M, 30 °C).
In an attempt to elucidate the enhanced reactivity of
allylic stannanes toward carbonyl compounds in polar
solvents, and in methanol in particular, we carefully
examined the ground state structure of 1a and 1b in
different solvents by NMR spectroscopy. Parameters
such as 119Sn chemical shift and 119Sn-13C coupling
constants are very sensitive to the substituents and
symmetry about the tin atom and are, therefore, a
convenient probe for small structural changes including
weak donor-acceptor interactions in solution.14 Selected
NMR parameters for these substrates and for the refer-
(15) Mulzer, J .; Angermann, A. Tetrahedron Lett. 1983, 24, 2843.
(16) Boldrini, G. P.; Lodi, L.; Tagliavini, E.; Tarasco, C.; Trombini,
C.; Umani-Ronchi, A. J . Org. Chem. 1987, 52, 5447.
(17) We have found that the diasteromers of 3e can be easily
separated as the corresponding SEM ethers by chromatography on
silica (10% ethyl acetate in hexane). They can also be separated after
benzylation using Mulzer’s procedure (reference 15).
(14) Wrackmeyer, B. In Annual Reports on NMR Spectroscopy;
Webb, G. A., Ed.; Academic Press: London, 1985; Vol. 18, p 73.

1964 J . Org. Chem., Vol. 62, No. 7, 1997
Cokley et al.
F igu r e 2. Possible transition state for the methanol-mediated
addition of allylic stannanes to carbonyl compounds.
lidene-D-mannitol as previously described.20 All other carbonyl
compounds were purchased from Aldrich and distilled im-
mediately before use.
allylic stannane via formation of an ionized intermediate.
Rather, we suggest that hydrogen bonding between the
carbonyl compound and methanol is the predominant
activating effect, and an eight-membered transition state
involving methanol coordination to tin and hydrogen
bonding to the carbonyl compound can be invoked (Figure
2). A similar mechanism can be used to account for the
corresponding HCl-mediated reaction of 1b with carbonyl
compounds.
The kinetic scheme depicted in eq 2 is consistent with
the results from our spectroscopic monitoring of the
reactions in methanol (Figure 1) and from what is known
of the reactivity of allylic stannanes bearing electron-
withdrawing groups on tin. Each substitution of alkoxide
for an allyl group would be expected to yield a more
reactive allylating species.18 The kinetic profile of this
reaction, i.e. rapid rate acceleration after a lag period, is
difficult to explain but may result from autocatalysis
provided by build up in the concentration of partially
soluble tin(IV) methoxide oligomers. This autocatalysis
does not occur in the corresponding reaction of 1a for
which the trimethyltin methoxide produced is much less
Lewis acidic. The presence of a minute amount of a
Lewis acidic impurity would account for the decreased
lag time in reactions of older, slightly cloudy samples of
1b.
Gen er a l P r oced u r e for Allyla tion s w ith 1b. Carbonyl
compounds (2, 10 mmol) and tetraallyltin (1b, 2.5 mmol) were
dissolved in methanol (10 mL) and stirred at either 30 °C or
at reflux for up to 24 h. The methanol solution was then
poured into water (50 mL) and extracted with dichloromethane
(3 × 25 mL). The combined organic extracts were dried
(MgSO4), and the solvent was removed in vacuo or by fractional
distillation for particularly volatile products (e.g. 3a ). The
crude product 3 was purified by kugelrohr distillation or
filtration through a short bed of silica gel (dichloromethane).
Com p etition Exp er im en ts. The competitive allylation
experiments were conducted according to the standard proce-
dure above with aldehyde (1.0 M), ketone (1.0 M), and 1b (0.25
M) at 30 °C for 16 h. The proportion of each allylated product
3 in the mixture was determined by 1H and 13C NMR
spectroscopy.
Allyla tion P r od u cts. Homoallylic alcohols 2 obtained by
1
allylation of aldehydes and ketones were characterized by H
and 13C NMR spectroscopy and comparison with literature
values: 2-methyl-5-hexen-3-ol (3a ),21 1-cyclohexyl-3-butenol
(3b),21 1-phenyl-3-butenol (3c),21 2-phenyl-4-penten-2-ol (3d ),22
1,2-O -isopropylidene-5-hexene-1,2,3-triol (3e),13 ethyl 2-hy-
droxy-2-methyl-4-pentenoate (3f),23 3-methyl-5-hexen-3-ol (3g),24
1-allylcyclohexanol (3h ),21 2-phenyl-4-penten-2-ol (3j).25
Ack n ow led gm en t. We gratefully acknowledge As-
sociate Professor Ian J enkins for helpful discussion and
the Australian Research Council for a Research Fellow-
ship to D.Y.
J O962332L
Exp er im en ta l Section
Ma ter ia ls. Allyltrimethylstannane (1a ) was prepared from
allylmagnesium bromide and trimethyltin chloride as previ-
ously described.19 Tetraallyltin 1b was purchased from Ald-
rich and distilled immediately before use. Carbonyl compound
2e was prepared by periodate oxidation of 1,2:5,6-diisopropy-
(20) Schmid, C. R; Bryant, J . D.; Dowlatzedah, M.; Phillips, J . L.;
Prather, D. E.; Schantz, R. D.; Sear, N. L.; Vianco, C. S. J . Org. Chem.
1991, 56, 4056.
(21) Davis, A. P.; J aspars, M. J . Chem. Soc., Perkin Trans. 1 1992,
2111.
(22) Deleris, G.; Dunogues, J .; Calas, R. Tetrahedron Lett. 1976, 28,
2449.
(18) (a) Gambaro, A.; Ganis, P.; Marton, D.; Peruzzo, V.; Tagliavini,
G. J . Organomet. Chem. 1982, 231, 307. (b) Yano, K.; Baba, A.;
Matsuda, H. Bull. Chem. Soc. J pn. 1992, 65, 66.
(19) Harston, P.; Wardell, J . L.; Marton, D.; Tagliavini, G.; Smith,
P. J . Inorg. Chim. Acta 1989, 162, 245.
(23) Moorlag, H.; Kellogg, R. M.; Kloosterman, M.; Kaptein, B.;
Kamphuis, J .; Schoemaker, H. E. J . Org. Chem. 1990, 55, 5878.
(24) Zhou, W.; Liang, S.; Yu, S.; Luo, W. J . Organomet. Chem. 1993,
452, 13.
(25) Pilcher, A. S.; DeShong, P. J . Org. Chem. 1996, 61, 6901.