Stereoselectivity in Pinacol-Homocoupling Mediated by Samarium Diiodide

Article abstract of DOI:10.1007/s00706-003-0061-x

The stereoselectivity of the pinacol-coupling of various substituted benzaldehydes mediated by samarium diiodide was investigated. The dependence of product-ratios, yields, and stereoselectivities on the substrate, the substrate-reagent-ratio, and the solvent is described.

Full text of DOI:10.1007/s00706-003-0061-x

Monatshefte f€ur Chemie 134, 1607–1615 (2003)
DOI 10.1007/s00706-003-0061-x
Stereoselectivity in Pinacol-Homocoupling
Mediated by Samarium Diiodide
ꢀ
¨ ¨
¨
Peter Gartner , Max Knollmuller, and Joachim Broker
Institute of Applied Synthetic Chemistry, Vienna University of Technology,
A-1060 Vienna, Austria
Received June 30, 2003; accepted July 2, 2003
Published online August 28, 2003 # Springer-Verlag 2003
Summary. The stereoselectivity of the pinacol-coupling of various substituted benzaldehydes mediated
by samarium diiodide was investigated. The dependence of product-ratios, yields, and stereoselectivities
on the substrate, the substrate-reagent-ratio, and the solvent is described.
Keywords. Lanthanides; Aldehydes; Reductions; Hydrobenzoin; Diastereoselectivity.
Introduction
Being interested in a general approach to arylsubstituted hydrobenzoins and
realizing that not all of these compounds are accessible via benzoin-condensa-
tion and subsequent reduction in acceptable yields, the direct pinacol-coupling
seemed to be a promising alternative. To be even more precise, we expected
the application of SmI2 in the reductive coupling of 4-nitrobenzaldehyde to be a
promising way to synthesize 1,2-bis(4-nitrophenyl)ethane-1,2-diol [1]. Further-
more, we found that described stereoselectivities vary in a broad range which
seemed to be due to different reaction conditions [1, 2], but no general inves-
tigation on this matter has been performed so far. Thus, we decided to test the
pinacol-coupling of different nitro-, methoxy-, and unsubstituted derivatives
with regard to the dependence of stereoselectivity on the substrate-reagent-ratio
as well as other reaction conditions. The results of this study are presented
herein.
Results and Discussion
The pinacol-coupling was carried out in two different solvents (THF and
MeOH) using 5 different aromatic aldehydes as substrates (1a: Ar ¼ Ph, 1b:
Ar ¼ 2-MeOPh, 1c: Ar ¼ 4-MeOPh, 1d: Ar ¼ 2-NO2-Ph, 1e: Ar ¼ 4-NO2-Ph).
ꢀ
Corresponding author. E-mail: peter.gaertner@tuwien.ac.at

€
P. Gartner et al.
1608
Scheme 1
Table 1. Results of the reactions using benzaldehyde (1a) or 2- and 4-anisaldehyde (1b and 1c) as
substrates following the general procedures for the reactions in THF and MeOH described in the
experimental section
Entry
Ar
Solvent
Yield 2=3=%
Yield 4=%
dl=meso-ratio
(2:3)a
1
2
3
4
5
6
Ph
THF
52
65
35
54
49
70
18
34
29
14
27
10
63:37
68:32
60:40
62:38
58:42
52:48
Ph
MeOH
THF
2-MeO-Ph
2-MeO-Ph
4-MeO-Ph
4-MeO-Ph
MeOH
THF
MeOH
a
1
dl=meso-ratios were determined via H-NMR-spectra of the crude products
Furthermore, additional experiments with different substrate-reagent-ratios were
carried out in THF.
In the case of benzaldehyde (1a) and 2- and 4-methoxybenzaldehyde (1b and
1c) as substrates the reaction lead to a mixture of meso- and dl-diols (3a–3c and
2a–2c) as well as small amounts of benzyl alcohol (4a–4c) (Scheme 1 and Table
1). The yields and stereoselectivities obtained from the reactions in MeOH are so
far corresponding to Ref. [3].
On the contrary, the pinacol-coupling of aromatic aldehydes with 1 equivalent
of SmI2 in THF without addition of lewis-acids or complexing-agents is, according
to literature, proceeding with only very low stereoselectivity and when 2 equiva-
lents of the reagent are used with benzaldehyde at ꢁ78ꢂC no selectivity should be
observed [1, 2, 4–6], apart from the fact that 1b should not give any hydrobenzoin
derivative at all [2]. However, although 1b gave the lowest yield of the desired
product compared to aldehydes 1a and 1c, it was actually possible to carry out a
pinacol-coupling yielding 35% of the corresponding hydrobenzoin with a dl=meso-
ratio of 1.5:1 (Table 1, entry 3). Further on, the observed diastereoselectivities drop
proceeding from benzaldehyde (1a) to 2-methoxybenzaldehyde (1b) and the para-
substituted derivative 1c both in THF and MeOH. Whereas the drop in selectivity
from 1a to 1b may be due to an additional complexation of the 2-OMe-group such
an effect is impossible in 1c due to steric reasons. Yields of hydrobenzoins tend to
be higher when the reaction is carried out in MeOH but any further interpretation of
the obtained yields and dl=meso-ratios regarding the dependence from solvent and
substrate seems to be quite difficult, as no obvious trends have been observed.
To investigate the dependence of the achievable diastereoselectivity on the
amount of reagent used for the reaction in THF, additional experiments, using a

Pinacol-Homocoupling Mediated by Samarium Diiodide
1609
Table 2. Results of experiments using different amounts of 0.1 M SmI2-solution or SmI2 generated
in situ
Entry
Equiv. SmI2
Yield 2=3a=%
dl:mesoa
Yield 4a=%
Yield 5a=%
1
2
3
4
5
6
7
8
9e
1.5b
52
46c
39
55
95
93
14
82
61c
63:37
43:57
37:63
43:57
31:69
32:68
71:29
56:44
54:46
18
–
–
2.0b
–
3.0b
–
–
1.1d
14
–
30
–
2.0d
3.0d
–
–
0.5d
25
18
10
31
–
0.8d þ 1.2 Sm(0)
0.1d þ 3 Sm(0)
–
a
dl=meso-ratios were determined via 1H-NMR-spectra of the crude products; b SmI2 was generated
c
in situ from Sm(0) and 1.5 equiv. of 1,2-diiodoethane; approx. 20% of aldehyde were recovered;
d
a commercially available 0.1 M solution of SmI2 in THF was used; e the experiment was carried out
at ꢁ78ꢂC
Scheme 2
commercially available SmI2-solution, were carried out. This way an accurate
amount of reagent could be provided in the reaction mixture. The results of these
experiments, using the 0.1 M SmI2-solution as well as SmI2 generated in situ, are
shown in Table 2.
Using a large excess of reagent, the reaction gave a mixture of only meso- and
dl-hydrobenzoin, but equal amounts of reagent and substrate or an excess of sub-
strate lead to mixtures of meso- and dl-hydrobenzoin, benzyl alcohol and benzoin
as the main products (Scheme 2). Furthermore, the dl=meso-ratio seems to be
highly dependent on the substrate-reagent-ratio. When an excess of reagent is used,
the reaction is moderately meso-selective (Table 2, entries 5 and 6). With about one
equivalent of reagent or less the yield of hydrobenzoin drops and when less than
one equivalent of reagent is used, mainly the dl-diastereomer is obtained (Table 2,
entries 4 and 7). This trend is also obvious in the experiments where SmI2 was
formed in situ (entries 1–3). The shift towards dl-selectivity (Fig. 1) compared to
the experiments with the SmI2-solution can be explained by a lower amount of
reagent actually formed in the reaction mixture. Another explanation could be that
only a certain amount of SmI2 is present in the mixture when the aldehyde is added
and it is formed as the reaction proceeds due to shift of equilibrium, providing a
lower concentration over the whole reaction time.
To find conditions for a dl-selective pinacol-coupling leading to a satisfying
yield of hydrobenzoin, two experiments with a combination of less than 1 equiva-
lent of SmI2-solution and an excess of Sm(0) were carried out (Table 2, entries 8

€
P. Gartner et al.
1610
Fig. 1. Dependence of stereoselectivity on amount of SmI2 (
formed in situ)
: SmI2-solution;
: SmI2
and 9). The use of 0.8 equivalents of SmI2 together with 1.2 equivalents of Sm(0)
lead to much higher yields of hydrobenzoin but a significant loss of selectivity. The
use of catalytic amounts of SmI2 at lower temperature did not improve this result as
well.
It has to be noted that the pinacol-coupling of benzaldehyde using Sm(0) in
THF=HCl is already known, hence a direct participation of the samarium-metal
present in the reaction mixture is possible [7]. However, considering the poor
yields of hydrobenzoin and benzyl alcohol using Sm(0) without addition of
any activating reagent as I2 or HCl, this influence seems to be negligible [7]. As
the reaction-times using Sm(0) in THF=HCl are at least 24 h, the reaction of
remaining aldehyde with excessive samarium-metal during the hydrolysis of the
reaction mixture (see experimental section) does not seem to be very likely as
well.
The observed dependence of the obtained diastereoselectivities can be ex-
plained by two different and competing reaction-mechanisms following the possi-
ble mechanisms proposed in literature (Scheme 3) [2, 5]: If a large excess of
reagent is used for the reaction and the substrate is slowly added to the reaction
mixture, the aldehyde is quantitatively transformed to the ketyl-radical and the
main mechanism is a combination of two ketyl-radicals. In this case, the transition
state B seems to be sterically favoured, leading to the meso-product. If less than
one equivalent of reagent is used and the reagent is added to the substrate, the
ketyl-radical reacts mainly with excessive aldehyde. In this case the transition state
A is favoured due to the possible coordination of the aldehyde-carbonyl-O with the
Sm(III) of the ketyl-radical, leading to the dl-form as the main product.
On the other hand, according to Ref. [1], 4-nitrobenzaldehyde (1e) should
give the corresponding hydrobenzoin in 0.5 minutes and 95% yield, but the
reaction of 2- and 4-nitrobenzaldehyde (1d and 1e) with SmI2 in THF and
MeOH, respectively, yielded only the corresponding benzyl alcohol (4d and
4e) (Table 3). When the usual reaction conditions in THF (see experimental

Pinacol-Homocoupling Mediated by Samarium Diiodide
1611
Scheme 3
Scheme 4
Table 3. Results of the reactions using nitro-derivatives 1d and 1e as substrates
Entry
Solvent
Ar
Yield=%
1
THF
2-NO2-Ph
2-NO2-Ph
4-NO2-Ph
4-NO2-Ph
51
57
24
42
2
3a
MeOH
THF
4
MeOH
a
SmI2 (dissolved in THF) was added to a solution of the aldehyde in THF
section) were applied to 4-nitrobenzaldehyde (1e) the nitro-substituent was
reduced as well leading to 4-aminobenzyl alcohol (4f). Although the possibility
to reduce aromatic nitro-compounds with an excess of SmI2 in THF=H2O or
MeOH is already known in literature, this reaction was only observed in case of
the p-NO2-derivative in THF [8–10]. By inverting the sequence of reagent addi-
tion this side-reaction could be prevented and 4-nitrobenzyl alcohol (4e) was
isolated (Table 3, entry 3).
An additional experiment using benzoin as substrate was carried out to inves-
tigate if SmI2 and the conditions used for the pinacol-couplings are applicable for
the conversion of benzoin to meso- and dl-hydrobenzoin. It has to be noted that,
according to literature, the reduction of benzoin with a SmI2-LiNH2-system in

€
P. Gartner et al.
1612
Scheme 5
THF=MeOH should give a mixture of hydrobenzoin and 1,2-diphenylethanol as the
main products together with traces of desoxybenzoin and benzyl alcohol [11].
However, this was not the case when the reaction conditions described herein were
applied on benzoin as the substrate (Scheme 5 and experimental section).
1,2,3,4-Tetraphenyl-2,3-butanediol (9) was most likely formed by the
pinacol-coupling of desoxybenzoin 6. This coupling has been described earlier to
take place during the electrolytic reduction of desoxybenzoin in the presence
of Fe2þ or Cr3þ beside the formation of the corresponding alcohol [12]. The
NMR-spectrum of this compound corresponds with the data given in literature
for the dl-diol [13]. We also observed the formation of 1,2,3,4-tetraphenyl-
1,2,3,4-butanetetrole (10) which is the actual coupling product of 5. The for-
mation of 10 during the cathodic pinacolisation of benzoin was described
already earlier [14]. However, no configuration assignment was performed so
far. Although the determination of configuration was beyond the scope of this
investigation, a few things can be estimated from NMR results with regard to
stereoselectivity. The reduced set of signals gives evidence for a molecule with
C2- or CS-symmetry. Thus, relative configuration should be (Rꢀ,Rꢀ,Sꢀ,Sꢀ) or
(Rꢀ,Sꢀ,Rꢀ,Sꢀ) in case of CS-symmetry and (Rꢀ,Rꢀ,Rꢀ,Rꢀ) or (Rꢀ,Sꢀ,Sꢀ,Rꢀ) in case
of C2-symmetry. As shown in Scheme 6 for the transition-state of (R)-benzoin
the relative configuration of C1 and C2 should be (Rꢀ,Sꢀ) due to the possible
chelation in the transition-state and the following attack of either a second
ketyl-radical or a benzoin-carbonyl from the sterically less hindered side. Con-
sidering the dl-selectivity of the pinacol-coupling of benzaldehyde under the
same reaction-conditions (Table 2, entry 1) as well as the predominant config-
uration of 9 formed in the same reaction-mixture, the relative configuration of
C2 and C3 should be dl (in this case (Sꢀ,Sꢀ)) as well. Thus, regarding the 4
possibilities mentioned before, (Rꢀ,Sꢀ,Sꢀ,Rꢀ)-configuration seems to be the most
likely one.
Scheme 6

Pinacol-Homocoupling Mediated by Samarium Diiodide
1613
Conclusion
Most of the desired hydrobenzoin-derivatives could be synthesized via the pinacol-
coupling induced by samarium diiodide with satisfactory yields, however, the
nitro-substituted compounds gave only the corresponding benzyl alcohols. The
stereoselectivity of the reaction could actually be controlled via variation of
the substrate-reagent-ratio. With an excess of SmI2 meso-hydrobenzoin could be
obtained in satisfying yield and moderate diastereoselectivity, whereas subequimolar
amounts of SmI2 gave preferably dl-hydrobenzoin. However, yields are lower than
for the reaction with an excess of SmI2 due to obvious reasons. The applicability of
SmI2 as a reducing agent for carbonyl-compounds, like in this case for benzoin,
seems to be limited due to the fact that both reduction and reductive coupling of
these compounds take place simultaneously, resulting in complex product-mixtures.
Experimental
Diethylether, petrolether and CH2Cl2 were distilled before use. Tetrahydrofuran (THF) and MeOH were
degassed under Ar and distilled over Na=benzophenone and Mg, under Ar directly before use. 1,2-
Diiodoethane was dissolved in ether and the solution was extracted with a Na2S2O3-solution, washed
with brine, dried with anhydrous Na2SO4, and evaporated under slightly reduced pressure. Aldehydes
and benzoin were purified by vacuum-distillation directly before use. Commercially available I2 and Sm
(Aldrich, 99.9%, ꢁ 40 mesh) were used without further purification. SmI2-solution (0.1M in THF) was
purchased from Aldrich and stored under Ar. All reactions were carried out under Ar atmosphere.
1
Meso=dl-ratios were determined from the crude products by integration of the H NMR spectra.
NMR spectra were recorded on a Bruker AC 200 spectrometer operating at 200MHz (1H) or
50MHz (13C) using the solvent peak as reference. The measurements were carried out in CDCl3 or
DMSO-d6 at 300 K.
Analytical thin-layer chromatography (TLC) was performed on Merck silica gel 60F254 coated
aluminum sheets and compounds were visualised with UV-light or staining with 5% phosphomolybdic
acid hydrate in ethanol and heating.
Elemental analyses (C, H, N) were conducted at the Institute of Physical Chemistry at the Uni-
versity Vienna, Mikroanalytical Laboratory, their results were found to be in good agreement (ꢃ0.3%)
with the calculated values.
General Procedure for the Pinacol-Coupling Carried Out in THF
A solution of 141 mg of 1,2-diiodoethane (0.50 mmol) in 2 cm3 of THF was added dropwise to a
suspension of 100 mg of samarium (0.67 mmol) in 1 cm3 of THF. (After the addition of 10% of the
diiodoethane solution the mixture was stirred until the blue colour of SmI2 in THF was apparent and
the remaining reagent was added afterwards.) After 30 minutes a solution of the aldehyde (0.33 mmol)
in THF was added and the reaction was stirred until the blue colour of SmI2 disappeared. The reaction
mixture was quenched with 0.1 N HCl and extracted with ether. The combined organic phases were
washed with 10% Na2S2O3-solution and brine, dried with anhydrous Na2SO4, filtered, and evaporated
to dryness. The crude product was purified by flash chromatography (petrolether:ether 20:1! ether).
Experiments with Different Substrate-Reagent-Ratio Using Sm(0) and 1,2-Diiodoethane
The procedure described above was only slightly modified: 1.34 mmol of samarium and 1.0 mmol of
1,2-diiodoethane were used for the in-situ-synthesis of the reagent. 0.5 and 0.33 mmol of benzaldehyde
were added to the mixture.

€
P. Gartner et al.
1614
Experiments with Different Substrate-Reagent-Ratio Using 0.1 M SmI2-Solution
When an excess of reagent was used for the reaction, the SmI2-solution (2–9 cm3, depending on the
substrate-reagent-ratio) was placed in the reaction-vessel without further dilution and the correspond-
ing amount of benzaldehyde dissolved in THF was added. In the case of 0.5 equivalents of SmI2 the
reagent was added slowly to the aldehyde dissolved in THF. For the experiments with additional Sm(0)
the aldehyde was added to the stirred suspension of samarium in the SmI2-solution. The initially blue
reaction mixture immediately turned yellow, but the blue colour returned during the reaction time of
12h. In all cases the workup was carried out as described above.
General Procedure for the Pinacol-Coupling Carried Out in MeOH
A solution of 99 mg of iodine (0.39 mmol) in 0.5 cm3 of MeOH was added dropwise to a mixture of
100 mg of samarium (0.67 mmol) and the aldehyde (0.39mmol) in 2 cm3 of MeOH. During the
exothermic reaction the colour of the reaction mixture turned from brown to light orange and the
reaction mixture was stirred until the TLC showed no further change. The reaction mixture was
quenched with 0.1 N HCl, MeOH was evaporated, and the residue was extracted with ether. The
combined organic phases were washed with 10% Na2S2O3-solution and brine, dried with anhydrous
Na2SO4, filtered, and evaporated to dryness. The crude product was purified by flash chromatography
(petrolether:ether 20:1! ether).
Reaction of Benzoin with SmI2 in THF
A solution of 141 mg of 1,2-diiodoethane (0.50 mmol) in 2 cm3 of THF was added dropwise to a
suspension of 100 mg of samarium (0.67 mmol) in 1 cm3 of THF. (After the addition of 10% of the
diiodoethane solution the mixture was stirred until the blue colour of SmI2 in THF was apparent and
the remaining reagent was added afterwards.) After 30 minutes a solution of 71mg of benzoin
(0.33 mmol) in THF was added and the reaction was stirred until the blue colour of SmI2 disappeared.
The reaction mixture was quenched with 0.1 N HCl and extracted with ether. The combined organic
phases were washed with brine, dried with anhydrous Na2SO4, filtered, and evaporated to dryness. The
crude product (82 mg) was purified by flash chromatography (petrolether:ether 20:1! ether) yielding
13mg of the substrate (5), 5 mg (9%) of desoxybenzoin (6), 5 mg (9%) of benzil (7), 8 mg (13%) of
1,2-diphenyl-1,2-ethandiol (8), 10mg (9%) of dl-diol 9 [13], and 12mg (10%) of 1,2,3,4-tetraphenyl-
1,2,3,4-butanetetrole (10) as a white solid.
1,2,3,4-Tetraphenyl-1,2,3,4-butanetetrole (10; C28H26O4)
1
Mp.: 145–151ꢂC; H NMR (200 MHz, CDCl3, 25ꢂC): ꢀ ¼ 2.26, 4.59 (2bs, 4OH), 5.88 (s, 2CHOH),
13
7.01–7.32 (m, 20aromatic-H); C NMR (50 MHz, CDCl3, 25ꢂC): ꢀ ¼ 77.09 (s, Ph-COH), 82.60 (d,
Ph-CHOH), 127.02, 127.37, 127.59, 128.83 (4d, Ph-C), 140.53, 140.31 (2s, Ph-C-1); M¼ 436.52.
References
[1] Namy JL, Souppe J, Kagan HB (1983) Tetrahedron Lett 24: 765
[2] Annunziata R, Tenaglia M, Cinquini M, Raimondi L (1999) Eur J Org Chem: 3369
[3] Yanada R, Negro N, Yanada K, Fujita T (1997) Tetrahedron Lett 38: 3271
[4] Wallace TW, Wardell I, Li KD, Leeming P, Redhouse AD, Challand SR (1995) J Chem Soc,
Perkin Trans 1: 2293
[5] Pedersen LH, Christensen TB, Enemærke RJ, Daasbjerg K, Skrydstrup T (1999) Eur J Org
Chem: 565
[6] Christensen TB, Riber D, Daasbjerg K, Skrydstrup T (1999) Chem Comm: 2051

Pinacol-Homocoupling Mediated by Samarium Diiodide
1615
[7] Talukdar S, Fang JM (2001) J Org Chem 66: 330
[8] Wang L, Zhou LH, Zhang YM (1999) Synlett 7: 1065
[9] Banik BK, Mukhopadhyay C, Venkataraman MS, Becker FF (1998) Tetrahedron Lett 39: 7243
[10] Kamochi Y, Kudo T (1993) Chem Lett 9: 1495
[11] Kamochi Y, Kudo T (1991) Tetrahedron Lett 32: 3511
[12] Fournier F, Fournier M (1986) Can J Chem 64: 881
[13] Clerici A, Porta O (1985) J Org Chem 50: 76
[14] Juday RE, Waters WL (1975) Tetrahedron Lett 16: 4321