The effect of Lewis acids on the pinacol homocoupling reaction of aldehydes promoted by samarium diiodide

Article abstract of DOI:10.1002/(SICI)1099-0690(199912)1999:12<3369::AID-EJOC3369>3.0.CO;2-P

The effect of various Lewis acids on the samarium diiodide promoted pinacol homocoupling of aldehydes was investigated. The reaction of benzaldehyde proceeded with fair to good 1,2-anti-stereoselectivity, while in the case of other aromatic and aliphatic aldehydes syn-stereoselectivity was generally observed. Chiral α-alkylaldehydes allowed for an almost complete stereocontrol favoring syn-1,2-diols.

Full text of DOI:10.1002/(SICI)1099-0690(199912)1999:12<3369::AID-EJOC3369>3.0.CO;2-P

FULL PAPER
The Effect of Lewis Acids on the Pinacol Homocoupling Reaction of Aldehydes
Promoted by Samarium Diiodide
Rita Annunziata,[a] Maurizio Benaglia,[a] Mauro Cinquini,[a] and Laura Raimondi*[a]
Keywords: Pinacol reaction / Lewis acid / Samarium diiodide / 1,2-diols / Carbonyl compounds
The effect of various Lewis acids on the samarium diiodide
other aromatic and aliphatic aldehydes syn-stereoselectivity
promoted pinacol homocoupling of aldehydes was was generally observed. Chiral α-alkylaldehydes allowed for
investigated. The reaction of benzaldehyde proceeded with
an almost complete stereocontrol favoring syn-1,2-diols.
fair to good 1,2-anti-stereoselectivity, while in the case of
Introduction
Results and Discussion
SmI2-Promoted Homocoupling of Benzaldehyde in the
Presence of Lewis Acids
The pinacol coupling reaction is receiving renewed atten-
tion because of the recent availability of mild and selective
reducing agents.[1] Among these, samarium diiodide has
been successfully employed in the synthesis of several natu-
ral and nonnatural products.[1][2] While the intramolecular
coupling procedure is well established, less is known about
the intermolecular reaction: since the pioneering work of
Kagan et al.,[3] only a few studies have dealt with the stereo-
chemical outcome of the pinacol homocoupling of alde-
hydes promoted by SmII species.[1,2,4]
A great deal of effort has been devoted to make SmI2-
promoted reactions more convenient.[2] Interesting pro-
cedures have recently been proposed for the generation of
SmI2 from metallic samarium.[5] SmI2 can be regenerated
from SmIII species employing electrochemical procedures[6a]
or metallic species as coreductants,[6b,6c] and in the presence
of metallic magnesium a catalytic cycle was proposed.[6b,7]
Sm(Hg) was very recently shown to promote pinacol homo-
coupling of aromatic aldehydes.[8] Metal ketyl complexes
have been characterized by X-ray techniques,[9] while the
reducing power of SmI2 in THF in the presence of cosol-
vents such as HMPA has been measured.[10][11]
The SmI2-promoted homocoupling reaction of benzal-
dehyde was chosen to screen the behavior of different Lewis
acids. All reactions were run in THF under an argon atmos-
phere; two molar equivalents of a 0.1 THF solution of
SmI2 were added dropwise to a 0.1 THF solution of ben-
zaldehyde and the required additive (reaction conditions,
additives, chemical yields and diastereoisomeric ratios are
collected in Table 1). Assignment of the relative configura-
tion to the known diols 1a,b was based on the chemical
shift values of the CHOH proton that resonates at 4.65 and
4.80 ppm in the syn and anti isomer, respectively.[16]
Table 1. SmI2-promoted homocoupling of benzaldehyde
An increasing amount of studies also deals with the effect
of added Lewis acids on the intermolecular pinacol coup-
ling.[4,12Ϫ14] The stereoselection of the pinacol coupling is
often attributed to the capability of the metal to complex
both reactants in the transition state; the use of Lewis acids
more coordinating than the samarium(III) species is thus
likely to influence the syn/anti ratio. Very recently, we pub-
lished some preliminary results on the SmI2-promoted
homocoupling of imines in the presence of Yb(OTf)3,[15]
and wish now to report on our studies on the effect of
Lewis acids on the homocoupling reaction of aldehydes to
give syn and anti 1,2-diols.
[a]
Centro CNR and Dipartimento di Chimica Organica e Industri-
[a]
[b]
`
ale, Universita di Milano,
Molar equivalents with respect to benzaldehyde. Ϫ Only ben-
zyl alcohol was isolated.
via Golgi 19, I-20133 Milano, Italy
Eur. J. Org. Chem. 1999, 3369Ϫ3374
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1434Ϫ193X/99/1212Ϫ3369 $ 17.50ϩ.50/0
3369

R. Annunziata, M. Benaglia, M. Cinquini, L. Raimondi
FULL PAPER
As can be seen from the literature,[1Ϫ3] the coupling reac- direct reduction by SmI2 of the Lewis acid-complexed ben-
tion also proceeds smoothly at Ϫ78°C in the absence of zaldehyde was possible, the reaction rate should be in-
additives with excellent chemical yields, but with no stereo- creased due to the more electrophilic character of the com-
selection (entry 1). The addition of proton donors (entries plexed CϭO bond. However this was not the case (Table 1,
2,3) did not affect the stereoselection, but lowered the entries 5Ϫ9).
chemical yield,[3] leading to the formation of benzyl alcohol
In Figure 1 we indicate two possible transition structures
as the only by-product. On passing from protic acids to to account for the observed stereoselectivity. In the presence
BF3·Et2O, the diols 1a,b were obtained in good yields, but of TiIV or NiII species, the SmIII-ketyl radical complex re-
the reaction was completely stereorandom (entry 4). More acts immediately with the strongly electrophilic Lewis acid-
interestingly, in the presence of titanium(IV) a predomi- complexed aldehyde before the chelation equilibrium takes
nance of the anti diastereoisomer was observed.[5a] Ad- place. In TS A, leading to the anti diol, steric and electro-
dition of Ti(OiPr)4 slowed the reaction and higher tempera- static interactions are minimized. The syn isomer, on the
tures were required for the coupling to proceed (entries other hand, may be derived from the chelated TS B. The
5Ϫ9), the best results being achieved at 20°C (entry 8). The intervention of different reducing agents, generated in situ
stereoselection was independent of the reaction tempera- from the Lewis acid by SmI2, can be ruled out. SmII is not
ture. More acidic TiIV species, such as Ti(OiPr)2Cl2, led to likely to reduce all the added metal species. Moreover, sev-
less satisfactory results (entry 10). The use of the more hin- eral low-valent titanium species are known to promote the
dered Ti(TADDOL)2 did not allow for significant improve- coupling reaction, but with significant syn selectivity.[1][19]
ments (entry 11); unfortunately, only racemic syn-1a was
observed.
[12]
The reaction performed in the presence of NiCl2
showed an analogous behavior to the one observed when
Ti(OiPr)4 was employed: higher temperatures (20°C) were
required for the reaction to proceed, and anti selectivity was
observed (entries 12Ϫ14).[17] All the other Lewis acids we
tested allowed for modest to good anti stereoselectivity, but
with no significant improvement (entries 15Ϫ20).
HMPA is indeed the most commonly used cosolvent in
SmI2-promoted reactions, including pinacol coupling,[1][2]
and influences both the chemical yield and the stereoselec-
tivity of the reaction.[11] Addition of HMPA to SmI2 solu-
tions is known to lead to the formation of species with an
increased reducing power. In fact, the oxidation potential
Figure 1. Proposed TS for the formation of 1a,b
of a 0.5 solution of SmI2 in THF (Ϫ1.33 V) rises to Ϫ1.46
V for SmI2(HMPA)2, and to Ϫ2.05 V for SmI2(HMPA)4.[10]
The combined use of 2 molar equivalents of SmI2(HMPA)4
(generated in situ from SmI2 and HMPA) together with one
equivalent of Ti(OiPr)4 allowed for the formation of only a
small quantity of diols (27%) with a modest anti stereoselec-
tivity (entry 21). As by-products, benzyl alcohol and vari-
SmI2-Promoted Coupling of Aldehydes in the
Presence of TiIV Species
ous deoxygenated and unsaturated compounds were ob-
The homocoupling reaction of aldehydes other than ben-
zaldehyde in the presence of SmI2 and Lewis acids was then
served.[18]
This disappointing result prompted us to investigate the investigated. These aldehydes, however, were either nonre-
effect of HMPA alone as an additive. In fact, the reaction active, or reacted with low selectivity, and only Ti(OiPr)4
performed with SmI2(HMPA)4 led to the formation of was an efficient additive in this reaction. As for the coupling
benzyl alcohol as the only product (entry 22). With 2 molar of benzaldehyde, the presence of the TiIV species slowed the
equivalents of HMPA per SmI2, 1a,b were recovered in reaction, so that higher temperatures and sometimes longer
good chemical yield, but the reaction was almost stereoran- reaction times were required for the reaction to proceed.
dom (entry 23).
Some significant data are collected in Table 2.
A few comments can be added concerning the SmI2/
Both in the absence or in the presence of Ti(OiPr)4, ortho-
Lewis acid-promoted pinacol coupling of benzaldehyde substituted electron-rich aromatic aldehydes did not couple
(Table 1, entries 5Ϫ20). First of all, the significant decrease to give the corresponding diols (entries 1,2), with the sig-
of the reaction rate observed when TiIV (entries 5Ϫ11) or nificant exception of salicylaldehyde.[20] In the absence of a
NiII (entries 12Ϫ14) species were added, could be due to Lewis acid the coupling is stereorandom, while when
the presence of complexation equilibria in solution. Before Ti(OiPr)4 was added an 80:20 selectivity was observed, fav-
addition of SmI2, benzaldehyde is complexed with the oring the syn isomer.
Lewis acid; decomplexation and complexation with SmII is
probably needed for the reaction to take place.[2][19] If a in the presence of SmI2; upon addition of Ti(OiPr)4 only
2-Thienylcarboxaldehyde led to decomposition products
3370
Eur. J. Org. Chem. 1999, 3369Ϫ3374

Pinacol Homocoupling Reaction of Aldehydes Promoted by Samarium Diiodide
Table 2. SmI2-promoted homocoupling of aldehydes
FULL PAPER
Figure 2. Proposed TS for the formation of 2a,b
[a]
[b]
Molar equivalents with respect to benzaldehyde. Ϫ
Only 2-
SmI2-Promoted Coupling of Chiral Aldehydes
[c]
phenylethanol was isolated. Ϫ
Determined by 13C NMR spec-
troscopy, see ref.[22]
Despite its major drawbacks, we tested this procedure to
the pinacol coupling of chiral aldehydes. Reaction of op-
tically active α-alkoxy-aldehydes in the presence of SmI2,
with or without Ti(OiPr)4, led only to deoxygenation of the
substrates,[2] while the reaction performed on α-amino-alde-
hydes gave only the corresponding α-amino alcohols. Better
results were obtained with α-methylaldehydes such as rac-
2-phenylpropanal (Table 3). The coupling reaction per-
formed in the presence of 2 molar equivalents of SmI2 at
0°C led to the formation of only three out of the six (two
meso and four d,l couples) possible diastereoisomers of diol
8 (entry 1; cf. Figure 3), namely 3,4-syn-8a and 8b and 3,4-
anti-8c in a 56:22:22 ratio (see below for structural assign-
ment). Upon raising the reaction temperature to 20°C, both
chemical yield and reaction stereoselectivity improved, and
only 8a and 8b (in a 90:10 ratio) were formed in 55% chemi-
cal yield (entry 2). At this temperature and in the presence
of 1 molar equivalent of Ti(OiPr)4, only 8b was detected
and isolated in 77% yield (entry 3). Other reaction con-
ditions, i.e. higher reaction temperature (entry 4), addition
of HMPA (entries 5,6) and of Ti(OiPr)4 and HMPA (entry
7) were detrimental to yield but not to stereoselection.
Interestingly, in the presence of HMPA the reaction tem-
perature seems to exert a major influence (entries 5,6), but
traces of diols 3a,b were recovered (entries 5,6).[2][21] 2-Fur-
ylcarboxaldehyde, on the other hand, underwent the pina-
col reaction in both cases affording diols 4a,b, but with no
significant stereoselectivity (entries 7,8).
The behavior of aliphatic aldehydes was less puzzling. 2-
Phenylacetaldehyde in the presence of SmI2 gave diols 5a,b
with good syn selectivity, while in the presence of Ti(OiPr)4
only the corresponding alcohol was isolated (entries 9,10).
Aliphatic aldehydes, such as cyclohexancarboxyaldehyde,
were known to give the corresponding pinacols in good
yields and with no selectivity by reaction with SmI2.[1][3]
Addition of Ti(OiPr)4 slowed the reaction, and even at
higher temperatures diols 6a,b were isolated in low yields
and with little syn selectivity (entry 11). When performed
in the presence of SmI2(HMPA)4 and Ti(OiPr)4 the reaction
was completely syn selective, although in low chemical
yield, and was independent of the reaction temperature (en-
tries 12,13). The stereoselectivity is due to SmI2(HMPA)4:
in the absence of added Lewis acid, pinacol 6a was formed
with complete stereoselection at 60°C, while at room tem-
perature the reaction is stereorandom (entries 14,15). Ad-
dition of Ti(OiPr)4 allows complete stereoselectivity also at
room temperature, but at the expense of longer reaction
times.[22][23]
Table 3. SmI2-promoted homocoupling of rac-2-phenylpropanal
From the data reported in Table 2, it can clearly be seen
that the reaction conditions need to be tuned to the particu-
lar substrate. However, we preferred to use standard reac-
tion conditions in order to compare the behavior of differ-
ent substrates. When the coupling was stereoselective (en-
tries 9,12,13,15), a preference for the syn isomer was ob-
served in all cases, while for benzaldehyde reactions anti
diol was favored. Complexation equilibria can take place
before the actual coupling reaction occurs; in these con-
ditions, TS B (Figure 1) is the most favored and well ac-
counts for the preferential syn stereoselectivity.[24] The reac-
tions of salicylaldehyde to give 2a,b probably proceed
through chelated TSЈs C and D, depicted in Figure 2.
[a]
rac-diol 8a possesses C2 symmetry; relative configuration at C-2
[b]
and C-3 was not assigned (see text). Ϫ
Only one enantiomer is
indicated for sake of simplicity. Ϫ [c] Molar equivalents with respect
[d]
to benzaldehyde. Ϫ Only 2-phenyl-1-propanol was isolated.
Eur. J. Org. Chem. 1999, 3369Ϫ3374
3371

R. Annunziata, M. Benaglia, M. Cinquini, L. Raimondi
FULL PAPER
3,4-syn configuration of 8a. Assignment of the relative con-
figuration at C-2 and C-5 (which can be either 2,3-syn/4,5-
syn or 2,3-anti/4,5-anti) was not possible.
Conversion of an 80:10 mixture of 8b and 8c into the
corresponding acetonides gave the asymmetric (C1) com-
pounds 9b and 9c (Figure 4). The coupling constant values
observed for the hydrogens at C-4 and C-5 of the dioxolanyl
moiety (J ϭ 7.5 and < 0.5 Hz for 9b and 9c, respectively)
supported the trans configuration for 9b and the cis con-
figuration for 9c. On the basis of these observations and of
the asymmetric (C1) structure of 9b and 9c the 2,3-syn-3,4-
syn-4,5-anti configuration was assigned to diol 8b, and the
2,3-anti-3,4-anti-4,5-syn configuration to diol 8c.
It is tempting to discuss the stereochemical outcome of
the homocoupling of 2-phenylpropanal; however, reliable
suggestions can be advanced only for the reaction per-
formed in the presence of Ti(OiPr)4 as additive (Table 3,
entry 3).[26] Only in this case could the stereoselectivity be
unambiguously assigned as 2,3-syn-3,4-syn-4,5-anti Ϫ only
8b was formed Ϫ and chemical yields were satisfactory
(77%). A rationale is tentatively depicted in Figure 5.
Figure 3. All possible diastereomeric structures for diols 8
the extremely low chemical yields donЈt allow for further
speculation.
The assignment of the relative configuration at C-3 and
C-4 of compounds 8aϪc was carried out on converting 8 to
acetonides 9 and then using a combination of spectroscopic
evidence and symmetry considerations.[25] The reduced
1
number of signals observed in both the H and 13C spectra
of 8a suggested a C2 or CS symmetry for this compound.
Acetonide 9a, readily obtained from 8a (2,2-DMP,
1
BF3·Et2O) also showed H and 13C NMR spectra compat-
ible with these symmetries (Figure 4). The observed isoch-
ronicity of the two geminal methyls (homotopic in the C2-
symmetric structure, diastereotopic in the CS one) strongly
supported the 4,5-trans configuration of 9a, and hence the
Figure 5. Proposed TS for the formation of 8b
The complete 3,4-syn stereoselectivity arises from coordi-
nation of both reagents to the TiIV species, as shown in
Figure 1 (TS B). The well-known tendency of 2-phenylpro-
panal to react with nucleophiles in a Felkin-Ahn mode can
account for the 2,3-syn relative stereochemistry.[27] In the
corresponding radical-anion species, however, the increased
steric requirements of the alkoxide changes its reactivity to
an anti-Felkin mode,[28] thus a 4,5-anti relative stereochem-
istry is favored. In the proposed TS, moreover, all steric
interactions are minimized.[24]
Conclusion
The major pitfall of the SmI2-based pinacolic homocoup-
ling lies in the great versatility of this reagent: the reaction
conditions need to be tuned carefully for each substrate in
Figure 4. All possible diastereomers for acetonides 9
3372
Eur. J. Org. Chem. 1999, 3369Ϫ3374

Pinacol Homocoupling Reaction of Aldehydes Promoted by Samarium Diiodide
FULL PAPER
mmols) and BF3·Et2O (0.1 mmols) at 0°C, and the reaction kept at
that temperature for 2 h. The reaction was quenched with saturated
aqueous NH4Cl, and extracted twice with Et2O. The crude product
was purified by flash chromatography (Et2O/hexanes 90:10). Com-
pound 9a was obtained from 8a in 81% yield; 9b,c were obtained
from 8b,c in 95% yield. C21H26O2 (310.437): calcd. C 81.25, H 8.44;
found C 81.31, H 8.46. Significant 1H and 13C NMR resonances
are reported in Table 4.
order to achieve the best results each time. The intermolecu-
lar homocoupling reaction of benzaldehyde upon addition
of Lewis acids stronger than SmIII allows for synthesis of
the anti diol, while syn diastereoisomers are always favored
with different, achiral aldehydes.
The presence of chiral substituents on the aldehyde has a
major influence on the reaction selectivity. The intermolec-
ular coupling of chiral α-methylaldehydes is extremely ster-
eoselective, and syn stereoselectivity is often complete.
Further investigation is needed in order to extend this meth-
odology to substrates of greater synthetic interest.
1
Table 4. Significant H and 13C NMR spectroscopic data for com-
pounds 8a؊c and 9a؊c
Experimental Section
CHN Analyses: PerkinϪElmer 240 instrument. 1H and 13C NMR:
Bruker AM300 at 300.133 and 75.47 MHz, respectively; CDCl3 as
1
solvent; H chemical shifts are reported in δ relative to TMS; 13C
chemical shifts and JHϪH coupling constants are reported in Hz.
Silica gel (230Ϫ400 mesh) was used for flash chromatography. Or-
ganic extracts were dried over Na2SO4 and filtered before removal
of the solvent. Dry solvents were distilled as follows: THF from
Na and benzophenone (twice), CH2Cl2 from CaH2; HMPA from
CaH2 (in vacuo). Distilled HMPA was stored under a nitrogen at-
mosphere over 4A molecular sieves. All reactions employing dry
solvents were performed under an argon atmosphere. Ti(OiPr)4 and
all aldehydes were distilled before use. ZnCl2 was dried by sub-
sequent fusions in vacuo. TiCl2(OiPr)2 was prepared from TiCl4
(1 mol. equiv.) and Ti(OiPr)4 (1 mol. equiv.); Ti(TADDOL)2 from
Ti(OiPr)4 (1 mol. equiv.) and TADDOL (2 mol. equiv.). Commer-
cially available (Aldrich) 0.1 solution of SmI2 in dry THF was
used in the coupling procedure.
[a] 13C NMR spectrum not measured. Ϫ
Resonances cannot be
[c]
[b]
assigned unambiguously to C(H)-1 or C(H)-6. Ϫ
Resonances
cannot be assigned unambiguously to C-2 or C-5. Ϫ [d] Resonances
cannot be assigned unambiguously to C-3 or C-4.
General Procedure for the Homocoupling Reaction of Aldehydes:
Synthesis of Diols 1, 6a؊b and 8a؊c: To a stirred 0.1 solution of
the required aldehyde in dry THF, kept at the desired temperature
(see Tables 1Ϫ3), were added the appropriate additive (see Tables
1Ϫ3) and SmI2 (0.1 solution in THF, 2 equiv.). The reaction was
monitored by TLC, and quenched with 10% aqueous HCl. The two
phases were separated, and the water layers were extracted twice
with Et2O. The crude product was purified by flash chromatogra-
phy (Et2O/ hexanes 50:50 Ǟ 80:20). Reaction conditions, chemical
yields and diastereoisomeric ratios are reported in Tables 1Ϫ3.
Acknowledgments
Special thanks are due to Prof. Franco Cozzi for helpful stereo-
chemical discussion, to Marco Covini and Valerio Chiroli for their
collaboration. Partial financial support by MURST (60%) and
CNR is gratefully acknowledged.
[1]
[1a] G.M. Robertson, “Pinacol Coupling Reactions”, in Compre-
Diols 1, 3؊6a,b are known compounds.[16,22,23]
hensive Organic Synthesis Vol. 3 (Eds.: B. M. Trost, I. Fleming),
[1b]
Pergamon Press, Oxford, 1991, pp. 563Ϫ611. Ϫ
M. Benag-
2a,b: C14H14O4 (246.262): calcd. C 68.28, H 5.73; found C 68.35,
H 5.70. Ϫ 1H NMR: δ ϭ 7.13 (dt, J ϭ 7.6, 1.7 Hz, 2 H), 6.86 (dd,
J ϭ 7.1, 1.1 Hz, 2 H), 6.60 (dt, J ϭ 7.1, 1.1 Hz, 2 H), 6.45 (dd,
J ϭ 7.6, 1.7 Hz, 2 H), 5.05 (s, 2 H, 2b), 5.02 (s, 2 H, 2a).
lia, L. Raimondi, “Stereoselection in Pinacol Coupling Reac-
tions of CϭO and CϭN Double Bonds”, in Seminars in Or-
`
ganic Synthesis (Ed.: F. Cozzi), Societa Chimica Italiana, 1998,
pp. 225Ϫ250.
[2]
[3]
G. A Molander, C. R. Harris, Chem. Rev. 1996, 96, 307Ϫ338.
[3a] J. L Namy, J. Souppe, H. B. Kagan, Tetrahedron Lett. 1983,
8aϪc: C18H22O2 (270.372): calcd. C 79.96, H 8.20; found C 79.91,
H 8.24. For significant 1H and 13C NMR resonances, see Table
4.[25]
[3b]
24, 765Ϫ766. Ϫ
J. Souppe, L. Danon, J. L. Namy, H. B.
Kagan, J. Organomet. Chem. 1983, 250, 227Ϫ236.
H. L. Pedersen, T. B. Christensen, J. Enemærke, K. Daasbjerg,
T. Skrydstrup, Eur. J. Org. Chem. 1999, 565Ϫ572.
[4]
Synthesis of 7a,b From 2a,b: To a stirred solution of 2a,b (1 mmol)
in dry THF (10 mL) was added K2CO3 (5 mmols) at room tem-
perature. After 45 min., methyl iodide (3 mmols) was added drop-
wise; the mixture was stirred overnight, then diluted with water
(10 mL) and Et2O (10 mL). The two phases were separated, and
the aqueous layers extracted twice with Et2O. The crude product
was purified by flash chromatography (Et2O/hexanes 90:10), af-
fording known 7a,b[23] in 70% yield.
[5] [5a]
R. Yanada, N. Negoro, K. Yanada, T. Fujita, Tetrahedron
[5b]
Lett. 1997, 38, 3271Ϫ3274. Ϫ
Y. Nishiyama, E. Shinomiya,
S. Kimura, K. Itoh, N. Sonoda, Tetrahedron Lett. 1998, 39,
3705Ϫ3708.
[6] [6a]
´
´
E. Leonard, E. Dun˜ach, J. Perichon, J. Chem. Soc., Chem.
[6b]
Commun. 1989, 276Ϫ277. Ϫ
R. Nomura, T. Matsuno, T.
[6c]
Endo, J. Am. Chem. Soc. 1996, 118, 11666Ϫ11667. Ϫ
F.
´
Helion, J.-L. Namy, J. Org. Chem. 1999, 64, 2944Ϫ2946.
[7]
In our hands, however, metallic Mg promoted the pinacol
homocoupling of benzaldehyde also in the absence of SmII spec-
ies. See also: W.-C. Zhang, C.-J. Li, J. Org. Chem. 1999, 64,
3230Ϫ3236.
Synthesis of 9a؊c From 8a؊c: To a stirred solution of 8 (1 mmol)
in dry CH2Cl2 (10 mL) were added 2,2-dimethoxypropane (3
Eur. J. Org. Chem. 1999, 3369Ϫ3374
3373

R. Annunziata, M. Benaglia, M. Cinquini, L. Raimondi
FULL PAPER
[8]
L. Wang, Y. Shang, Synth. Commun. 1998, 28, 3991Ϫ3997.
Addition of Ti(OiPr)4 completely inhibited the reaction, in
agreement with our results (ref.[5a]).
[9] [9a]
Z. Hou, T. Miyano, H. Yamazaki, Y. Wakatsuki, J. Am.
[21]
[9b]
Desulfurization was the major (or only) reaction path allowed
in this case. While working on a different research project, we
tested the reactivity of aromatic aldehydes ortho-substituted
with sulfur-bearing residues in the presence of SmI2: we never
observed diol formation, but only desulfurization products to-
gether with substitution reaction at the aromatic ring in some
particular substrates.
Chem. Soc. 1995, 117, 4421Ϫ4422. Ϫ
Z. Hou, A. Fujita, Y.
Zhang, T. Miyano, H. Yamazaki, Y. Wakatsuki, J. Am. Chem.
Soc. 1998, 120, 754Ϫ766.
[10] [10a] M. Shabangi, R. A. Flowers II, Tetrahedron Lett. 1997, 38,
[10b]
1137Ϫ1140. Ϫ
M. Shabangi, J. M. Sealy, J. R. Fuchs, R.
A. Flowers II, Tetrahedron Lett. 1998, 39, 4429Ϫ4432. Ϫ [10c] J.
B. Shotwell, J. M. Sealy, R. A. Flowers II, J. Org. Chem. 1999,
Diols 3a, 4a, 5a,b and 6a are known compounds.[23] Diols 2a,b
were chemically correlated to the known o-methoxyderivatives
7a,b.[23] Stereochemical assignment to diols 2؊6a,b was per-
formed on the basis of both 1H and 13C NMR spectra and
chemical correlation. The CH(OH) resonance of the syn prod-
ucts often overlapped with the CH2OH singlet of the corre-
sponding alcohol (the most common by-product of these reac-
tions).
[22]
[10d]
64, 5251Ϫ5255. Ϫ
For X-ray structure of SmI2 complexes
with HMPA, see: Z. Hou, Y. Wakatsuki, J. Chem. Soc., Chem.
Commun. 1994, 1205Ϫ1206.
[11]
Some extremely interesting papers have recently been published
regarding the stereoselectivity in the homocoupling of benzal-
dehyde tricarbonylchromium complexes in THF, and its depen-
[11a]
dence on the presence of HMPA as cosolvent:
N. Tanigu-
chi, N. Kaneta, M. Uemura, J. Org. Chem. 1996, 61,
[23] [23a]
T. W. Wallace, I. Wardell, K.-D. Li, P. Leeming, A.
[11b]
6088Ϫ6089. Ϫ
N. Taniguchi, M. Uemura, Tetrahedron
Redhouse, S. R. Challand, J. Chem. Soc., Perkin Trans. I 1995,
1998, 54, 12755Ϫ12788.
2293Ϫ2308. Ϫ [23b] K. R. K. Prasad, N. N. Joshi, J. Org. Chem.
[12]
[13]
[14]
F. Machrouhi, B. Hamann, J.-L. Namy, H. B. Kagan, Synlett
1996, 633Ϫ634.
J. R. Fuchs, M. L. Mitchell, M. Shabangi, R. A. Flowers II,
Tetrahedron Lett. 1997, 38, 8157Ϫ8158.
For a review on the use of Lewis acids in free radical reactions,
see: P. Renaud, M. Gerster, Angew. Chem. Int. Ed. 1998, 37,
2562Ϫ2579.
R. Annunziata, M. Benaglia, M. Cinquini, F. Cozzi, L. Rai-
mondi, Tetrahedron Lett. 1998, 39, 3333Ϫ3336.
[23c]
1996, 61, 3888Ϫ3889. Ϫ
J. B. Lambert, H. W. Mark, E.
Stedman Magyar, J. Am. Chem. Soc. 1977, 99, 3059Ϫ3067. Ϫ
[23d] R. W. Hoffmann, K. Ditrich, G. Köster, R. Stürmer, Chem.
Ber. 1989, 122, 1783Ϫ1789.
[24]
We have no evidences about the reaction mechanism operating
in the different cases. Figure 1 shows a nucleophilic addition of
the ketyl radical to a carbonyl, but direct coupling of two ketyls
or nucleophilic addition of a C,O-dianion to the CϭO cannot
be ruled out. Complexation and/or chelation by the different
metal species complicates the picture. For details on the differ-
ent mechanisms proposed for the coupling reaction and a stere-
ochemical discussion, see ref.[1][4]
[15]
[16]
[17]
A. Clerici, O. Porta J. Org. Chem. 1985, 50, 76Ϫ81.
In the presence of a Ni0 species [Ni(PPh3)4], 1a,b was obtained
in 15% yield and a 35/65 syn/anti ratio.
[25]
1
Comparison of H NMR spectra of diols 8aϪc with those re-
[18]
[19]
In the presence of NiCl2 and HMPA, only unsaturated products
were detected by NMR spectroscopy.
ported in the literature was ambiguous; see: T. Hirao, M. Asah-
ara, Y. Muguruma, A. Ogawa, J. Org. Chem. 1998, 63,
2812Ϫ2813.
Precomplexation of the reducing species before formation of the
ketyl radical was demonstrated for some low-valent titanium-
promoted couplings: A. Fürstner, B. Bogdanovic, Angew. Chem.
Int. Ed. Engl. 1996, 35, 3442Ϫ3469.
The outstanding importance of a free hydroxyl to coordinate
the samarium species, although in reactions other than pinacol
coupling, is well-known in the literature.[2] Surprisingly enough,
Yanada et al. observed some pinacol coupling in the reaction
of the o-methoxybenzaldehyde (SmI2 was generated from Sm
and I2 in methanol), with good yields but low syn selectivity.
[26]
[27]
Interestingly enough, 8b arises from the coupling of two differ-
ent enantiomers (S* ϩ R*) of rac-2-phenylpropanal.
E. L. Eliel, S. H. Wilen, L. N. Mander, Stereochemistry of Or-
ganic Compounds, Wiley, New York, 1994.
[20]
[28] [28a]
Y. Yamamoto, K. Matsuoka, H. Nemoto, J. Am. Chem.
[28b]
Soc. 1988, 110, 4475Ϫ4476. Ϫ
Y.-D. Wu, K. N. Houk, J.
Am. Chem. Soc. 1992, 114, 1656Ϫ1661.
Received July 4, 1999
[O99407]
3374
Eur. J. Org. Chem. 1999, 3369Ϫ3374