Enantioselective hydrogen transfer reactions from chiral binaphthyl variants of tin hydrides to prochiral radicals

Article abstract of DOI:10.1016/j.tetasy.2003.07.018

Enantioselective, reagent-controlled radical reductions of prochiral alkyl radicals, mediated by binaphthyl variants of tin hydrides can be carried out in a highly selective fashion: a maximum selectivity of 68% e.e. was reached. Temperature, solvent, Lewis acid and substituent effects are selectivity-controlling elements. The reactions can be conducted catalytically with 1 mol% of chiral information and excess NaCNBH₃.

Full text of DOI:10.1016/j.tetasy.2003.07.018

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 14 (2003) 3069–3077
Enantioselective hydrogen transfer reactions from chiral
binaphthyl variants of tin hydrides to prochiral radicals
Michael Blumenstein, Matthias Lemmler, Ahlke Hayen and Ju¨rgen O. Metzger*
Department of Chemistry, University of Oldenburg, PO Box 2503, 26111 Oldenburg, Germany
Received 16 June 2003; accepted 25 July 2003
Abstract—Enantioselective, reagent-controlled radical reductions of prochiral alkyl radicals, mediated by binaphthyl variants of
tin hydrides can be carried out in a highly selective fashion: a maximum selectivity of 68% e.e. was reached. Temperature, solvent,
Lewis acid and substituent effects are selectivity-controlling elements. The reactions can be conducted catalytically with 1 mol%
of chiral information and excess NaCNBH3.
© 2003 Elsevier Ltd. All rights reserved.
1. Introduction
finding suitable chiral reagents as well as selectivity
control seem to be more challenging. However, our
research has been dedicated to this approach for some
time now and we report herein about our progress
specifically involving binaphthyl variants of tin
hydrides as chiral hydrogen donors in enantioselective
hydrogen transfer reactions.
In the past decade, a growing number of contributions
concerning enantioselective radical transformations
have been published, proving the potential of this reac-
tion type after highly selective diastereoselective radical
reactions had already been receiving great attention and
gaining importance as synthetic tools for some time.1
p-Facial discrimination, which is required for the enan-
tioselective transformation of a prochiral radical or
radical trap, can be achieved via two different
approaches: In a complex-controlled reaction, an achi-
ral substrate (radical or radical trap) is coordinated by
or even covalently attached to a chiral reagent prior to
reaction, rendering the faces of the radical
diastereotopic and thus allowing the reaction to pro-
ceed through diastereomeric transition states with an
overall enantioselective reaction outcome after cleavage
of the auxiliary if necessary. A reagent-controlled reac-
tion on the other hand involves a prostereogenic radical
or radical trap and a chiral reagent. In this strategy, the
radical’s (or radical trap’s) faces become diastereotopic
by virtue of their association with the chiral reagent in
the transition state.
In the history of chiral tin hydride donors, several
contributions were based on terpene-derived reagents,
where the tin atom was bearing for example a menthyl-
group, and alkyl- or aryl-groups in addition to hydro-
gen.3–5 None of these compounds were able to reduce
the alkyl a-bromo phenylalkanoates, commonly used as
model substrate systems with significant enantioselectiv-
ity, unless they were used in combination with sterically
demanding chiral Lewis acidic auxiliaries.6 Research in
our group has been previously directed towards tin
hydrides, bearing chiral 2-[(1-dimethylaminoalkyl)-
phenyl] ligands. With these compounds, we demon-
strated that the commonly used a-bromo alkanoates
could be enantioselectively reduced with enantiomeric
excess of up to 25% in radical hydrogen transfer reac-
tions.7
Examples for complex-controlled enantioselective radi-
cal reactions have been reported since 1995,2 and excel-
lent enantioselectivities have been observed using this
strategy, while, in the latter type of transformations,
2. Results
C2-Symmetrical compounds have shown excellent per-
formance in various reactions and especially the
binaphthyl moiety has evolved as an important tool in
stereoselective synthesis. Thus, we decided to direct our
focus to the development of enantiomerically pure
* Corresponding author. Tel.: +49 (0) 441-798 3718; fax: +49 (0)
441-798 3329; e-mail: juergen.metzger@uni-oldenburg.de
0957-4166/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2003.07.018

3070
M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
Scheme 1.
binaphthyl tin hydrides and their application in
reagent-controlled enantioselective hydrogen transfer
reactions. A few tin organic compounds, containing the
binaphthyl unit had been described before,8–10 but tin
hydrides remained unknown.
pentyl-4,5-dihydro-3H-dinaphtho[2,1-c:1%,2%-e]stannepin
5a with lithium alanate lead to the desired tin hydride
1b. Yields for the synthesis of tin hydride 1a were
comparable.
Our efforts coincided with the development of a
binaphthyl-derived tin hydride by Curran et al., who
gained access to a similar methyl-substituted compound
by a different synthetic way.16
We developed a synthetic route to alkyl-substituted
binaphthyl tin hydrides, that basically allows the intro-
duction of any alkyl substituent to the tin atom, as long
as the corresponding dichloro alkylphenyl tin com-
pound is available.11 This method proved to be suitable
for the introduction of sterically demanding groups like
tert-butyl and neopentyl: both enantiomers of 4-tert-
butyl-4,5-dihydro-3H-dinaphtho[2,1-c:1%,2%-e]stannepin
1a as well as (R)-4-neopentyl-4,5-dihydro-3H-dinaph-
tho[2,1-c:1%2%-e]stannepin 1b were synthesized. The syn-
thetic route is exemplified for the new tin hydride 1b in
Scheme 1.
1H NMR-spectroscopic examination of the new tin
hydrides surprisingly revealed that the hydride proton
produces a doublet, so obviously coupling with at least
one of the methylene protons takes place. This was
confirmed by irradiation experiments.
To get a more detailed impression of the structure of
the new tin hydrides, molecular modelling, based on the
semiempirical PM3 method17 (RHF—restricted Hartree
Fock approach, structures were minimized with the
default options available within the MOPAC 93 pro-
gram package18) was carried out, and the result,
obtained for tin hydride 1b, is presented in Figure 1.
Starting point for the general synthetic scheme is the
chiral ligand (R)-2,2%-bis-(chloromethyl)-1,1%-binaphthyl
2, which we synthesized according to published proce-
dures.9,12 Enantiomeric separation was achieved by
fractionated
crystallization
with
(−)-ephedrine.
Although this route might seem rather time consuming,
it avoids the use of expensive catalysts. For the follow-
ing Grignard-activation and reaction of binaphthyl
compound 3, the quality of the activated magnesium
The selective potential of all new tin hydrides in enan-
tioselective hydrogen transfer reactions was tested with
a system of methyl bromo phenylalkanoates as sub-
strates. Special focus was placed on temperature depen-
[magnesium-anthracene-complex,
9,10-dihydro-9,10-
anthracendiyl)-tris-(tetrahydrofuran)-magnesium]13 is
crucial and a subsequent successful alkylation of the
dichlorotin compound 7 can already be predicted by
characteristic colour changes during the reaction.
Dichloro neopentylphenyltin 7b as well as the tert-butyl
derivative 6a necessary for the synthesis of 1a, were
synthesized from triphenyltin chloride with neopentyl14
or tert-butyl lithium, respectively, leading to the alkyl
triphenyl tin derivatives 6a and 6b, which upon treat-
ment with hydrochloric acid then gave 7a and 7b.15
After this ring formation, in the next step the phenyl
group of stannepin 4 could be selectively exchanged for
bromo if the reaction with bromine was carried out in
diethyl ether in high dilution instead of methanol,
which is the commonly used solvent for comparable
brominations.4,11 Finally, reduction of 4-bromo-4-neo-
Figure 1. Preferred conformation of stannepin 1b: the neo-
pentyl group is pointing in the opposite direction relative to
the reactive hydrogen.

M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
3071
dence and steric substrate effects, but solvent effects
and influence of Lewis acids were examined as well.
Finally, the feasibility of a catalytic reaction procedure
was investigated. Scheme 2 summarizes the substrates 8
that were employed and the reduction products 9 that
resulted.3,4 The enantioselective reduction of substrate
8g had been reported for a complex-controlled reaction
before and was included here for comparison.19,20
initiators, where applicable. At 0°C and below, the
initiator system Et3B/O2 was added to accelerate the
reaction.
The enantioselectivity of the reduction of bromo ester
8a by tin hydride 1a was examined over a temperature
range of −110 to 110°C, and all results proved to be
reproducible. Neopentyl-substituted tin hydride 1b was
tested at three different temperatures (Table 1). All
reactions mediated by (S)-configured 1a showed an
(R)-selective preference in the product ratio, whereas
(R)-1a and (R)-1b both favoured the formation of
(S)-configured 9. The enantioselectivities observed in
the reaction series with 1a correlate well with the recip-
rocal temperature: starting with 28% at 110°C, the
selectivity increases by lowering the reaction tempera-
ture, and finally 64% e.e. are reached at −110°C. At two
different temperatures, the enantioselectivities obtained
by using (R)-1a or (S)-1a while applying the same
conditions, respectively, were compared (entries 3/4 and
10/11, respectively). As expected, the use of opposite
enantiomers of 1a resulted in a quantitative reversal of
the selectivity observed for reaction products (R)-9a
and (S)-9a. This outcome also supports the reliability
of the analytical determination of enantioselectivities.
The results, obtained by using neopentyl-substituted tin
hydride (R)-1b, indicate a somewhat lower inductive
potential of this compound relative to 1a (entries 5–7).
At room temperature, only 4% e.e. were observed,
whereas at −30°C, as well as at −60°C, 28% e.e. were
detected.
Scheme 2.
All reactions at and below room temperature were
carried out on an analytical scale in diethyl ether.
Enantioselectivity, product yields and degree of conver-
sion of bromo substrates were determined by gas chro-
matography using dodecane as internal standard.
Generally, excellent yields of reduction products were
obtained (>90%) and with the exception of substrates
8b and 8f only minimum amounts of byproducts were
detectable. To ensure a complete conversion of tin
hydride and for reasons of practicability, the ratio of 8
to 1a or 1b usually was varied between 1:1 and an
excess of 3:1. In additional experiments, it had been
confirmed that within this range no change in enan-
tioselectivity occurs.4 The reaction time, until complete
conversion of tin hydride had been reached, depended
on the reaction temperature and mode of addition of
Next, differently substituted bromo esters 8b–f and the
iodo lactone 8g (Scheme 1) were submitted to the same
reaction conditions, using tin hydride (S)-1a. As in the
case of tert-butyl compound 8a, for most of the sub-
strates examined here an increase of selectivity was
observed by decreasing the reaction temperature; how-
ever, selectivities were generally lower than with 8a.
Table 1. Enantioselectivity of the radical reduction of a-bromo ester 8a, mediated by tin hydrides 1a and 1b at different
temperatures
Entry
H-donor
8a:1a
T [°C]
tb [h]
[(S)-9a]:[(R)-9a]
Ee %
1
2
3
4
(S)-1a
(S)-1a
(S)-1a
(R)-1a
(R)-1b
(R)-1b
(R)-1b
(S)-1a
(S)-1a
(S)-1a
(R)-1a
(S)-1a
3:1
3:1
2:1
110c
69d
24
24
20
−30
−60
0
−15
−78
−78
−110
3
3
22
16
5
18
24
48
48
27
24
30
36:64
35:65
34:66
64:36
58:42
64:36
64:36
30:70
28:72
24:76
76:24
18:82
28
30
32
28
16
28
28
40
44
52
52
64
2:1
f
5
6e
7e
8e
9e
10e
11e
12e
f
f
1.1:1
1:1
1.8:1
1.8:1
2:1
a The conversion of 8a was quantitative in most cases (with reference to the ratio 8:1), and yields of 9a were usually ]90%, referenced to the
conversion of 8a.
b Time until complete conversion of tin hydride was reached.
c The reaction was carried out in toluene.
d Hexane was used as solvent.
e Addition of 10 mL of a solution of Et3B (15%) in hexane (entries 8–12) or diethyl ether (entries 6 and 7) at the beginning and every 4 h.
f 8a was used in excess, conversion and yield were not determined.

3072
M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
in selectivity (entries 4 and 5). Tetrahydrofuran surpris-
ingly is the least suitable solvent for the examined
reaction, judging by the obtained product ratio of only
40:60 and although in methanol the best enantioselec-
tivity was found, the extended reaction time of 72 h,
which is probably due to the heterogeneity of the
reaction caused by insolubility of the tin hydride, dis-
qualifies this solvent (entry 7).
Also, the results clearly demonstrate a steric substituent
effect: when the sterical demand of alkyl substituents of
8 was reduced (8b: iso-propyl, Table 2, entries 1 and 7;
8c: ethyl, entries 2, 8 and 8c: methyl, entries 3 and 9),
enantioselectivity levels decreased. The same applied for
compounds 8e and 8f (entries 4, 5 and 10), where an
alkyl substituent R1 is combined with a small alkyl
group R2 instead of a phenyl group. For lactone 8g, the
best enantioselectivities in this series of different sub-
strates were found, showing a decent product ratio of
75:25 at −110°C as well as a temperature dependent
selectivity (entries 6, 11 and 12). While tin hydride
As Lewis acids play such an important role for
stereoselectivity in complex-controlled radical reactions
mediated by achiral reagents, we decided to test the
influence of some of these additives on the outcome of
our reactions. In Table 4, the results obtained by
adding MgI2, MgBr2, the chiral, enantiomerically pure
tin bromide (S)-5, or a combination of MgI2/TMEDA
to the usual reaction mixture, containing substrates 8a
or 8g, and tin hydride (S)-1a are compiled. Obviously,
in the case of substrate 8a, neither MgI2, nor MgBr2
exerted an influence on the selectivity (Table 4, entries 2
and 3), whereas an addition of the Lewis-acidic corre-
sponding tin bromide (S)-5a both significantly influ-
enced reaction time and enantioselectivity: At room
temperature, complete conversion was observed after
only 15 min, accompanied by a slightly increased selec-
(S)-1a lead to
a preferential formation of (R)-
configured products 9a–c with alkylphenyl bromo esters
8a–c, for dialkyl substrate 8f (S)-9 was favoured.
The results presented in Table 3 show, that the reaction
and the enantioselectivity can be significantly affected
by the solvent. Diethyl ether, methylene chloride and
benzene obviously are very comparable, equally suit-
able solvents for the examined transformation: reaction
times and product ratios were very similar (Table 3,
entries 1–3). In toluene and hexane, a marginal exten-
sion of necessary reaction time to reach complete con-
version of 1a was observed, as well as a small increase
Table 2. Enantioselectivity of the radical reduction of a-bromo esters 8b–g, mediated by tin hydride (S)-1a at different
temperatures
Entry
8
8:1a
T [°C]
tb [h]
[(S)-9]:[(R)-9]
E.e. %
1
2
3
4
5
6
7
8
b
c
d
e
f
g
b
c
d
f
1.7:1
2:1
1.1:1
2:1
2:1
1.4:1
1:1
1.5:1
1.1:1
1.4:1
2:1
21
21
21
21
21
24
24
24
22
24
7
48
48
48
45
19
31
39:61
47:53
50:50
50:50
51:49
54:46
33:67
45:55
52:48
57:43
69:31
75:25
22
6
0
0
2
21
8
−78c
−78c
−78c
−78c
−78c
−110c
34
10
4
14
38
50
9
10
11
12
g
g
2.5:1
a The conversion of 8 was quantitative in most cases (with reference to the ratio 8:1), and yields of 9 were usually ]90%, referenced to the
conversion of 8 [exceptions: entries 1 (55%), 5 (7%), 7 (47%)].
b Time until complete conversion of tin hydride was reached.
c Addition of 10 mL of a solution of Et3B (15%) in hexane during the first 4 h of reaction time.
Table 3. Enantioselectivity of the radical reduction of a-bromo ester 8a, mediated by tin hydride (S)-1a in different solvents
at room temperature
Entry
Solvent
8a:1aa
tb [h]
[(S)-9a]:[(R)-9a]
E.e. %
1
2
3
4
5
6
7
Diethyl ether
Methylene chloride
Benzene
Toluene
Hexane
2:1
2.5:1
1:1
22
24
26
28
28
72
24
34:66c
34:66
34:66
32:68
31:69
30:70
40:60
32
32
32
36
38
40
20
1:1
1.2:1
0.9:1
2.5:1
Methanol
Tetrahydrofurane
a The conversion of 8a was quantitative in most cases (with reference to the ratio of 8a:1), and yields of 9 were usually ]89%, referenced to the
conversion of 8a.
b Time until complete conversion of tin hydride was reached.
c conf. Table 1, entry 3.

M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
3073
Table 4. Influence of Lewis acidic additives on the enantioselective reduction of selected a-halo esters, mediated by tin
hydride (S)-1a in diethyl ether at room temperature or −78°C
Entry
Additivec
8
8:1aa
T [°C]
tb [h]
[(S)-9]:[(R)-9]
E.e. %
1d
2
3
–
a
a
a
a
a
a
g
g
g
2:1
1.7:1
2.2:1
1:1
1.8:1
1:1
2:1
1:1
3.6:1
24
21
21
22
48
48
34:66
34:66
32:68
30:70
24:76
16:84
69:31
66:34
51:49
32
32
36
40
52
68
38
32
2
MgI2
MgBr2
(S)-5a
–
(S)-5a
–
(S)-5a
MgI2, TMEDA
4
22
0.25e
27
5d
6
−78f
−78f
−78f
−78f
−78f
45
19
45
14
7d
8
9
a The conversion of 8a was quantitative in most cases (with reference to the ratio of 8a:1), and yields of 9 were usually ]90% [exception: entry
9 (81%)], referenced to the conversion of 8a.
b Time until complete conversion of tin hydride was reached.
c Molar ratio of additive to substrate app. 1:1.
d Conf. Tables 1 and 2.
e Complete conversion after 15 min, the enantiomeric ratio was observed for 20 h and did not change.
f Addition of 10 mL of a solution of Et3B (15%) in hexane during the first 3 h of reaction time.
tivity compared to the reaction without additive (entries
1 and 4) and at −78°C, the enantiomeric ratio increased
from 24:66 to 16:84 in the presence of tin bromide (S)-5a
(entries 5 and 6). However, substrate 8g was reduced with
lower enantioselectivity upon addition of Lewis acids
(entries 7–9). No effect of deracemication was observed
when racemic product 9a was mixed with 5a under
usual experimental conditions. Also, the reduction of
bromo ester 8a by achiral tributyltin hydride in the
presence of tin bromide 5a only lead to racemic products
9a.
8a that could be reached by using only 1 mol% of chiral
tin hydride (R)-1a in combination with excess of achiral
hydride NaCNBH3 (entry 2), was lower than in the case
where stoichiometric amounts of the chiral reagent had
been used (entry 1), while was comparable when a
catalytic amount of tin bromide 5a in connection with
NaCNBH3 was used as reduction system (entry 3). As
observed for the stoichiometric reactions listed in Table
1, opposite enantiomers of tin bromide 5a lead to a
reversal in selectivity in the reduction of 8a (entries 3 and
4). When the same reaction was conducted with only 0.01
mol% oftinbromide 5a, asignificantlyprolonged reaction
time was necessary until complete conversion of substrate
8a had occurred and also, the enantioselectivity only
reached 8% e.e. With iso-propyl-substituted bromo com-
pound 8b and with lactone 8g, the catalytic (1 mol% tin
bromide 5a) and the stoichiometric pathways lead to
comparable results, but in the case of 8b, again an
extended reaction time for complete conversion became
necessary (entries 6–9).
By using the achiral hydride sodium cyanoborohydride
we were able to carry out radical reductions enantioselec-
tively with only catalytic amounts of chiral information
(<1 mol% catalyst) for the first time (Table 5). During
the reaction, chiral tin hydride (R)-1a was generated or
regenerated in situ from catalytic amounts of tin bromide
(R)-5a or tin hydride (R)-1a with excess NaCNBH3.
Curiously, the enantioselectivity level of the reduction of
Table 5. Enantioselective radical reduction of substrates 8a, 8b, and 8g, mediated by catalytic amounts of tin hydrides
(R)-1a and (S)-1a in diethyl ether at room temperature
Entry
H-donor system
8a
tb [h]
[(S)-9]:[(R)-9]
E.e. %
1
2
3
4
5
6
7
8
9
(R)-1a
a
16
16
16
16
48
24
48
70
70
64:36
58:42
63:37
35:65
46:54
39:61
38:62
54:46
52:48
28
16
26
30
8
22
24
8
(R)-1a, NaCNBH3
(R)-5a, NaCNBH3
(S)-5a, NaCNBH3
(S)-5a, NaCNBH3
(R)-1a
(S)-5a, NaCNBH3
(R)-1a
(S)-5a, NaCNBH3
ac
ad
ad
ae
b
bd
g
gd
4
a The conversion of 8 was quantitative in most cases, and yields of 9 were usually ]93% [exception: entry 5 (74%)], referenced to the conversion
of 8.
b Time until complete conversion of tin hydride was reached.
c The ratio of 8 to 1a to NaCNBH3 was 100:1:300.
d The ratio of 8 to 5a to NaCNBH3 was 100:1:300.
e The ratio of 8 to 5a to NaCNBH3 was 10000:1:30000.

3074
M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
3. Discussion
cal center will most likely be oriented beneath the
binaphthyl substituent and the medium sized sub-
stituent M=Ph will probably point backwards, which
leaves the least hindered front space for the large
substituent L=tBu. In accordance with the experimen-
tal results, the situation depicted above will preferen-
tially yield (S)-configured product 9.
Our experimental results show a remarkable potential
for enantioselective, reagent-controlled radical reduc-
tions: the highest selectivity ever obtained in this kind
of transformation was observed in the reactions of tin
hydride 1a and methyl-2-bromo-3,3-dimethyl-2-phenyl-
butanoate 8a: 64% e.e. (Table 1, entry 12) and 68% e.e.
(with the addition of tin bromide 5a; Table 4, entry 6),
respectively.
Neopentyl-derived tin hydride 1b displays a poorer
performance than tert-butyl-substituted tin hydride 1a,
when reacted with compound 8a. Obviously, this also is
a matter of sterical factors: although the neopentyl
group of 1b is sterically more demanding overall, it also
possesses a certain flexibility, whereas the more com-
pact tert-butyl group of 1a is able and forced to exert a
greater influence towards the reaction center by intensi-
fying selective interactions with the substrate (Table 1).
Furthermore—and in contrast to the low yields
reported by Nanni and Curran16 for the enantioselec-
tive reduction of 2-bromo-1,2-diphenyl-propane-1-one
by a related binaphthyl tin hydride—we generally
obtained high yields of >90%, that are common for
reductions mediated by achiral tin hydrides. In the case
of a bromoketone as radical precursor, the formation
of byproducts resulting from carbonyl group reduction
seems likely to be responsible, and although we also
occasionally observed somewhat lower yields—primar-
ily for reactions involving bromo ester 8f—we do not
think that this is a matter of reduction of the ester
function, but rather of subsequent reactions.
For all compounds investigated, the enantioselectivity
increases with a decrease in reaction temperature,
although the extent of this dependence differs consider-
ably and is not correlated to the amount of selectivity.
Bromo ester 8f, for instance, exhibits a stronger temper-
ature dependence than the much more selectively
reduced substrate 8b (Table 2). The same is true for
bromo ester 8a, compared to iodo lactone 8g: although
8a is reduced with much higher selectivity, the tempera-
ture impact on the enantioselectivity is more distinct for
8g. A comparison of the binaphthyl tin hydride medi-
ated reduction of iodo lactone 8g with the previously
A more detailed look into our results reveals, that at
room temperature only tert-butyl- and iso-propyl-sub-
stituted bromo esters 8a and 8b, respectively, are
reduced with enantioselectivities higher than 10% e.e.,
whereas 8c–f obviously do not fulfill a certain sterical
minimum size requirement of substrate for a sufficient
selective interaction with the tin hydride (Tables 1 and
2).
19
reported tributyl tin hydride mediated variant involv-
ing a C2-symmetrical diamine as chiral auxiliary at
−78°C reveals, that the complex-controlled transforma-
tion is superior to the reagent-controlled reaction con-
cerning selectivity under these conditions: a ratio of
81:19 (R)-9g:(S)-9g was obtained using complex con-
trol, while 1a as reducing agent lead to a ratio of 69:31
(S)-9g:(R)-9g.
Lactone 8g cannot be directly compared due to its
cyclic character, however, for radical precursors 8a–f it
can be deducted from the results, that for a sufficient
stereodifferentiation certainly all substituents at the
radical center need to be as different as possible con-
cerning their sterical demand. Substrate 8f, carrying a
tert-butyl, a methyl and an ester group, for example, is
reduced almost unselectively, while compound 8a, that
only differs in having a phenyl instead of a methyl
group, is the most selectively reduced precursor.
An examination of the solvent nature with respect to
the reaction outcome revealed a strong influence, as
shown in Table 3 for the reduction of substrate 8a by
tin hydride (S)-1a. The highest enantioselectivity was
observed in methanol and the lowest in tetra-
hydrofuran, while other solvents were marginally worse
than methanol. The results show, that with methanol
and THF on both ends of the enantioselectivity range
and hexane in the middle, obviously polar effects can-
not be held responsible for these findings. It can rather
be assumed, that the complexing properties of the
solvent play an important role: Tetrahydrofuran is an
excellent complexing solvent for Lewis acids, while
diethyl ether is less complexing. Methylene chloride
possesses rather weak complexing properties and hex-
ane none. The same goes for methanol, which in this
case does not even solubilize tin hydride 1. With ben-
zene, p-stacking effects allow a small amount of com-
plexation, that is—due to less symmetry—even weaker
in toluene. The conclusion must be, that solvents with
strong complexing ability for Lewis acids lead to a
decrease in enantioselectivity by probably unfavourably
interacting with the hydrogen donor.
In Figure 2, a likely transition state arrangement for
hydrogen transfer from an (R)-configured tin hydride 1
to a radical is depicted: the linear alignment of donor,
hydrogen and acceptor atoms Sn, H and C should
enable a stereodifferentiation between the enantiotopic
faces of the radical caused by steric interactions of all
substituents. The small group S=COOMe of the radi-
Figure 2.

M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
3075
4. Conclusion
In Table 4, the results concerning effects of additional
Lewis acids on the enantioselectivity of the reduction
of several substrates 8 are compiled. Especially the
addition of tin bromide 5a before reaction had a selec-
tivity-enhancing effect, which is particularly pro-
nounced in the reduction of bromo ester 8a at −78°C
(Table 4, entry 6), but can also be observed at rt
(entry 4). In contrast, in the reduction of 8g the enan-
tioselectivity is decreased, compared to the reaction
without additive (entries 7 and 8). A subsequent alter-
ation of the reduction product composition by tin
bromide 5a can be ruled out, because no effect was
observed when racemic product 9a was mixed with 5a
under usual experimental conditions. Also, the reduc-
tion of bromo ester 8a by achiral tributyltin hydride in
the presence of tin bromide 5a only lead to racemic
products 9a, so obviously a chiral tin hydride like 1 is
mainly responsible for selectivity control, and an addi-
tive like tin bromide 5a merely causes an amplification
of enantioselectivity in conjunction with this chiral
reducing agent.
We have developed a synthetic route to enantiomeri-
cally pure alkyl-substituted binaphthyl tin hydrides
that is suited for the introduction of any kind of alkyl
group. With the high yielding syntheses of both enan-
tiomers of a tert-butyl variant and one enantiomer of
the neopentyl-substituted compound the accessibility
of this strategy was demonstrated. These new, chiral
tin hydrides were used as hydrogen donors in enan-
tioselective radical reductions, and a variety of inter-
esting results was found. Both the encouragingly high
enantioselectivities we obtained as well as the feasibil-
ity of conducting the reactions catalytically without
loss of selectivity are promising features in the field of
reagent controlled enantioselective radical reactions.
5. Experimental
5.1. General procedures
All reactions involving metalloorganic compounds
were carried out in standard Schlenk technique under
an atmosphere of dry, oxygen free argon. All commer-
cially available reagents were employed as supplied,
solvents were dried and distilled according to standard
procedures. Analytical GC was performed on a Carlo
Erba Vega Series GC 6000 with FID detector and 20
m DB1 capillary column or a Carlo Erba HRGC with
FID detector and chiral capillary column [heptakis-
Tin bromides like compound 5 possess a weak, Lewis
acidic character, and Lewis acids are known for their
ability to form complexes with carbonyl groups.
Therefore, a coordinating interaction of 5a with the
ester function of substrates 8 can be assumed to take
place in our case. As Sibi et al.21 have found before,
organo tin halides can act as weak Lewis acids and
can enhance both the chemo- and the diastereoselec-
tivity of a radical reaction. The present work shows
for the first time that this probably also applies to
enantioselective radical reactions.
(2,6-di-O-methyl-3-O-pentyl)-b-cyclodextrine
phase
diluted in 50% of OV-1701, 25 m or 25% 2,3-dimethyl-
6-tert-butyl-dimethylsilyl-b-cyclodextrin phase, 30 m].
1H, 13C and 119Sn NMR spectra were recorded on a
Bruker AM 300 (1H 300.1 MHz, 13C 75.5 MHz and
119Sn 122 MHz) or a Bruker AM 500 (1H 500.1 MHz
and 13C 125.8 MHz) spectrometer at 20°C using TMS
(1H NMR), SnMe4, or solvent signals as internal or
external standard, respectively. Mass spectra were
recorded on a Finnigan MAT 212 or a MAT 95
(HRMS). Elemental analysis was conducted by Ana-
lytische Laboratorien 51789 Lindlar. Specific rotations
were measured on a Perkin Elmer polarimeter 343.
Melting points were determined on a Laboratory
Devices Mel-Temp apparatus and are uncorrected.
Products were identified by comparison with GC
retention times of independently synthesized reference
compounds.
Surprisingly, an examination of a mixture of tin bro-
mide 5a and bromo ester 8a by NMR neither revealed
a complexation of the carbonyl group nor a complexa-
tion via p-stacking, so exactly what kind of complexa-
tion or interaction actually takes place remains
unknown. Based on the missing signal shifts in the
NMR spectra, it can be assumed though, that a rather
dynamic contact between tin hydride and tin bromide
dominates over carbonyl complexation. However, this
influence of tin bromide 5 on the tin hydride probably
is rather weak, otherwise enantioselectivity would be
evident in the reduction of bromo ester 8a by achiral
tributyl tin hydride in conjunction with tin bromide
5a.
5.2. Neopentyltriphenyl tin 6b
Another important finding of our research was the
feasibility of enantioselective radical reductions follow-
ing catalytic methodology (Table 5). Even if just 1
mol% of chiral ‘catalyst’ was employed, in most exam-
ples no or only a small decrease in enantioselectivity
was observed compared to the experiments carried out
with stoichiometric amounts of chiral tin hydride and
it was even possible to convert preparative amounts of
substrate (2.8 g) using the catalytic reaction variant.
However, if the quantity of chiral information was
further lowered to 0.01 mol% relative to the substrate,
a pronounced loss of selectivity was noted.
Chlorotriphenyltin (10 g, 25.94 mmol) was suspended
in 200 mL benzene. Over a time period of 6 h, 40 mL
(28.6 mmol) of a solution of neopentyllithium in ben-
zene (0.715 mol L−1) were added dropwise. The mixture
was stirred at room temperature overnight, then
refluxed for 1 h and allowed to cool. After hydrolysis
with 20 mL H2O and phase separation, the aqueous
layer was extracted with benzene and the pooled organ-
ics were dried over MgSO4. Evaporation of the solvent
yielded 7.65 g (70%) of 6b as a yellow solid. Mp 52°C;
1H NMR (500 MHz, CDCl3) l 1.04 (9H, s), 1.75 (2H,
3
3
s, JH,Sn(119)=29.09 Hz, JH,Sn(117)=28.54 Hz), 7.33 (9H,

3076
M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
m), 7.56 (6H, m); 13C NMR (125.7 MHz, C6D6) l
29.71, 31.00, 33.50 (3JC(13),Sn(117/119) 18.69/18.68 Hz),
128.40 (2JC(13),Sn(117/119) 23.87/25.65 Hz), 128.62, 137.07
(3JC(13),Sn(117/119) 17.64/17.65 Hz), 140.18; 119Sn NMR
(186.501 MHz, CDCl3) l 108.51.
[h]D20=+245.2 (c 2.31, CCl4); (S)-4a [h]2D0=−240.5 (c
1.12, CCl4); [h]D20=−226.9 (c 1.48, Et2O); [h]2D0=−116.1
1
(c 1.45, C6H6); H NMR (300 MHz, CDCl3) l 6.9–7.9
(17H, m), 2.21 and 2.44 (2H, AB, JAB=11.0 Hz), 2.19
and 2.56 (2H, AB, JAB=11.2 Hz), 1.18 (9H, s, 3J
(
H,Sn(117/119)=71.1 Hz/74.3 Hz); 1H NMR (300 MHz,
5.3. Dichloro tert-butylphenyl tin 7a
C6D6) l 6.9–7.8 (17H, m), 2.38 and 2.55 (2H, AB,
AB=11.0 Hz), 2.26 and 2.39 (2H, AB, JAB=11.2 Hz),
J
tert-Butyltriphenyl tin15 (20 g, 0.05 mol) was heated
with concentrated hydrochloric acid to 90°C for 7 min.
and the mixture was then cooled immediately. Extrac-
tion with dichloro methane was followed by drying
over CaCl2. After evaporation of the solvent, a colour-
less liquid was obtained by distillation in vacuo. Yield:
1.06 (9H, s, JH,Sn(117/119)=70.4 Hz/73.5 Hz]; 13C NMR
(75.5 MHz, CDCl3) l 139.91, 138.35, 138.03, 136.65,
133.18, 133.10, 131.56, 131.40, 131.28, 130.91, 128.62,
128.24, 127.88, 127.73, 127.64, 127.52, 125.90, 125.83,
125.79, 125.70, 124.08, 123.91, 31.10, 30.06, 17.84,
17.07; 119Sn NMR (122 MHz, C6D6) l 31.5; MS (70 eV,
EI) m/z (%) 534 (4) [M+], 477 (100) [M+−tBu], 400 (4),
280 (40), 197 (58), 195 (42), 193 (26), 179 (18), 178 (8);
HRMS (isobutane, CI): calcd. for C32H31Sn [MH+]
535.1447, found 535.1440; elemental anal. calcd for
C32H30Sn (533.3) C, 72.07; H, 5.67; Sn, 22.26; found C,
72.22; H, 5.61; Sn, 22.05.
3
1
12 g (74%); bp (0.03 mbar) 97°C; H NMR (300 MHz,
CDCl3) l 7.43–7.46 (2H, m), 7.10–7.13 (3H, m), 1.18
3
(9H, s, JH,Sn(117/119)=132 Hz/138 Hz); 13C NMR (75.5
MHz, CDCl3) l 138.6, 135.4, 131.5, 129.8, 43.7 [1J (13C
117/119Sn)=545 Hz/571 Hz], 28.4; 119Sn NMR (122
MHz, [D8]toluene) l 16.5; HRMS (isobutane, CI) calcd
for C10H15Cl2Sn [MH+] 324.9572, found 324.9567.
5.6. (R)-4-Neopentyl-4,5-dihydro-4-phenyl-3H-dinaph-
tho[2,1-c:1%,2%-e]stannepin 4b
5.4. Dichloro neopentylphenyl tin 7b
Same procedure as for 7a. Yield 70%; bp (0.05 mbar)
Same procedure as described for the synthesis of 4a.
Yield 69%; mp 160°C (dec.); [h]2D0=+170.6 (c 0.55,
CCl4); 1H NMR (300 MHz, CDCl3) l 1.26 (9H, s), 1.36
1
112°C; H NMR (300 MHz, CDCl3) l 0.79 (9H, s),
3
3
1.70 (2H, s, JH,Sn(119)=36.10 Hz, JH,Sn(117)=30.09 Hz),
3
3
7.03 (3H, m), 7.52 (2H, d, JH,H=5.83 Hz, JH,Sn(119)
=
2
2
(2H, d, JH,H=11.31 Hz, JH,Sn(117/119) 18.47/18.83 Hz),
2.15–2.47 (4H, m), 7.01–7.87 (17H, m); 13C NMR (75.5
MHz, CDCl3) l 19.31, 20.57, 29.71, 32.06, 33.48,
124.04–138.01 (20 C); 119Sn NMR (186.501 MHz,
CDCl3) l 18.1; MS (isobutane, CI) m/z (%) 604 (100)
[M+iBu]+; 470 (28) [M−Ph]+; MS (70 eV, EI) m/z (%)
548 (32) [M+] 477 (100) [M−neopentyl]+; HRMS (EI)
calcd for C33H32Sn [M+] 548.1526, found 548.1531.
39.02 Hz, JH,Sn(117)=36.09 Hz); 13C NMR (125 MHz,
C6D6) l 23.06, 32.64, 45.11, 130.72 (2JC(13),Sn(117/119)
38.04/38.04 Hz), 131.28, 134.69 (3JC(13),Sn(117/119) 31.48/
31.62 Hz), 141.06; 119Sn NMR (186.501 MHz, CDCl3)
l 38.5; HRMS (isobutane, CI) calcd. for C11H17Cl2Sn
[MH+] 338.9729, found 338.9722.
3
5.5. 4-tert-Butyl-4,5-dihydro-4-phenyl-3H-dinaphto-[2,1-
c:1%,2%-e]stannepin 4a
5.7. 4-Bromo-4-tert-butyl-4,5-dihydro-3H-dinaphtho[2,1-
c:1%,2%-e]stannepin 5a
2,2%-Bis-(chloromethyl)-1,1%-binaphtyl9 (2, 4.15 g, 11.87
mmol) was dissolved in 80 mL THF and subsequently
added to 10 g magnesium-anthracene-THF complex13
in 40 mL THF at rt over a 1 h period. Stirring was
continued for 2 h. In the course of the reaction, the
colour of the suspension changed from red-orange to
deep green. After complete addition, a yellow suspen-
sion was obtained, which was concentrated in vacuo to
a volume of 20 mL. 180 mL heptane were added, the
remainder of THF was removed in vacuo, and 300 mL
benzene were added. The resulting suspension was
stirred for 12 h, filtered through a glass frit (P4), and
the residue was washed with 100 mL heptane. The
yellow solid thus obtained was dissolved in 400 mL
THF and dichloride 7a (3.85 g, 9.5 mmol) was added in
one portion. Stirring was continued for 12 h, followed
by addition of 500 mL diethyl ether and hydrolysis with
100 mL saturated aqueous NH4Cl solution. After sepa-
ration, the organic layer was washed twice with 200 mL
H2O, with 100 mL saturated aqueous Na2CO3 solution,
with 100 mL brine, and concentrated. Drying over
MgSO4 was followed by further concentrating in vacuo,
and finally the residue was purified by column chro-
matography (alumina 90, neutral) with diethyl ether as
eluent. Yield: 4.8 g (74%); mp 165–168°C (dec.); (R)-4a
Stannepin 4a (1.4 g, 2.6 mmol) was dissolved in 50 mL
diethyl ether. The solution was cooled to −78°C and
under exclusion of light, bromine (400 mg, 2.5 mmol),
diluted in 25 mL diethyl ether was added over a 15 min.
period. Stirring at −78°C was continued for 3 h, the
reaction mixture was allowed to warm to rt and the
solvent was removed in vacuo. Yield: 1.28 g (91%);
(S)-5a: mp 180–185°C (brown), 240–250°C (dec.); (R)-
5a [h]2D0=+74.1 (c 0.27, C6H6); (S)-5 [h]2D0=−75.3 (c 1.2,
1
C6H6); H NMR (500 MHz, C6D6) l 6.9–7.8 (12H, m),
2.26 and 2.39 (2H, AB, JAB=11.4 Hz), 2.55 and 2.38
3
(2H, AB, JAB=10.8 Hz), 1.02 (9H, s, JH,Sn(117/119)
=
94.1 Hz/97.9 Hz]; 1H NMR (300 MHz, CDCl3) l
6.9–7.9 (12H, m), 2.89 and 2.38 (2H, AB, JAB=11.3
Hz), 2.62 a. 2.41 (2H, AB, JAB=11.4 Hz), 1.28 (9H, s,
3J(H,Sn(117/119)=95.1 Hz/99.4 Hz); 13C NMR (125.7
MHz, C6D6) l 125–136 (20C, m), 29.8, 28.1, 23.6, 21.3;
13C NMR (75.5 MHz, CDCl3) l 125 – 136 (20C, m),
30.0, 28.0, 23.6, 21.2; 119Sn NMR (122 MHz, C6D6) l
120.0; HRMS (isobutane, CI) calcd for C26H26BrSn
[MH+] 537.0240, found 537.0221; elemental anal. calcd
for C26H25BrSn (535.9) C, 58.28; H, 4.70; Sn, 22.15;
found C, 58.06; H, 4.64; Sn, 21.90.

M. Blumenstein et al. / Tetrahedron: Asymmetry 14 (2003) 3069–3077
3077
5.8. (R)-4-Bromo-4-neopentyl-4,5-dihydro-3H-dinaph-
tho[2,1-c:1%,2%-e]stannepin 5b
279 (100) [M+−tBuSnH]; HRMS (EI) calcd for
C27H27Sn [M+−H] 471.1134, found 471.1121; elemental
anal. calcd for C27H28Sn (471.22) C, 68.82; H, 5.99;
found C, 69.75; H, 5.85.
Same procedure as described for the synthesis of 5a.
Yield 90%; mp 151°C (dec.); [h]2D0=+38.1 (c 0.08, ben-
1
zene); H NMR (300 MHz, CDCl3) l 1.29 (9H, s), 1.41
2
(2H, d, JH,H=10.82 Hz), 2.32–2.67 (4H, m), 7.0–7.9
References
(12H, m); 13C NMR (75.5 MHz, CDCl3) l 23.06, 32.64,
45.11, 130.72 (2JC(13),Sn(117/119) 38.04/38.04 Hz), 131.28,
134.69 (3JC(13),Sn(117/119) 31.48/31.62 Hz), 141.06; 13C
NMR (75.5 MHz, CDCl3) l 22.42, 23.33, 29.80, 30, 32,
124.92–136.04 (20 C); 119Sn NMR (186.501 MHz,
C6D6) l=56.9 (s); MS (isobutane, CI) m/z (%)=606
(100) [M+iBu]+; 470 (100) [M−Br]+; MS (70 eV, EI) m/z
(%): 550 (30) [M+], 279 (100) [M+-neopentylSnH], 57
(85); HRMS (EI) calcd. for C27H27BrSn [M+] 550.0318,
found 550.0318.
1. Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of
Radical Reactions; VCH: Weinheim, 1996.
2. Sibi, M. P.; Porter, N. A. Acc. Chem. Res. 1999, 32,
163–171.
3. Schwarzkopf, K. Dissertation, Universita¨t Oldenburg,
1996.
4. Blumenstein, M. Stereoselektive Reaktionen in der
Radikalchemie; Shaker-Verlag: Aachen, 1997.
5. Dakternieks, D.; Dunn, K.; Henry, D. J.; Schiesser, C.
H.; Tiekink, E. R. T. Organometallics 1999, 18, 3342–
3347.
6. Dakternieks, D.; Dunn, K.; Perchyonok, V. T.; Schiesser,
C. H. J. Chem. Soc., Chem. Commun. 1999, 1665–1666.
7. Schwarzkopf, K.; Blumenstein, M.; Hayen, A.; Metzger,
J. O. Eur. J. Org. Chem. 1998, 177–181.
5.9. 4-tert-Butyl-4,5-dihydro-3H-dinaphtho[2,1-c:1%,2%-
e]stannepin 1a
A solution of tin bromide 5a (0.76 g, 1.4 mmol) in 20
mL diethyl ether was added to a suspension of LiAlH4
(57 mg, 1.5 mmol) in 10 mL diethyl ether at −10°C and
the mixture was stirred subsequently for 6 h at rt.
Workup included addition of 0.03 mL H2O and drying
over MgSO4. Yield 602 mg (93%); mp 210–220°C
(brown), 290–300°C (dec.); (R)-1a [h]2D0=−40 (c 0.1,
C6H6); (S)-1a [h]D20=+31 (c 0.88, C6H6); IR (KBr)
8. Gross, U.-M.; Bartels, M.; Kaufmann, D. J. Organomet.
Chem. 1988, 344, 277–283.
9. Chong, J. M.; Mac Donald, G. K.; Park, S. B.; Wilkin-
son, S. H. J. Org. Chem. 1993, 58, 1266–1268.
10. Noyori, R.; Kitamura, M.; Takemoto, K. Jpn. Kokai
Tokkyo JP 0491, 093, 1992.
11. Blumenstein, M.; Schwarzkopf, K.; Metzger, J. O.
Angew. Chem. 1997, 109, 245–247; Angew. Chem., Int.
Ed. Engl. 1997, 36, 235–236.
1
w˜=1790 cm−1; H NMR (300 MHz, C6D6) l 6.9–7.8
(12H, m), 6.0 (1H, d, 3J=5.1 Hz), 2.0–2.3 (4H, m), 1.01
(9H, s, 3J(H,Sn(117/119)=75.9 Hz/78.1 Hz); 13C NMR
(75.5 MHz, C6D6) l 125–134 (20C), 31.3, 28.3, 16.6,
16.1; MS (CI, benzene) m/z (%)=536.2810 (36) [M++
C6H6], 459.4769 (2) [MH++H]; HRMS (70 eV, EI)
calcd for C26H25Sn [M+−H] 457.0985, found 457.0981;
elemental anal. calcd for C26H26Sn (457.2) C, 68.31; H,
5.73; Sn, 25.96; found C, 68.04; H, 5.60; Sn, 25.65.
12. Maigrot, N.; Mazaleyrat, J.-P. Synthesis 1985, 3, 317–
320.
13. Bogdanovic, B.; Liao, S.; Schlichte, K.; Westeppe, U. In
Organic Syntheses; King, R. B.; Eisch, J. J., Eds.; Elsevier
publishing Comp: Amsterdam, 1988; Vol. 4, pp. 410–413.
14. Beringer, F. M.; Nathan, R. A. J. Org. Chem. 1970, 35,
2095–2096.
15. Kandil, S. A.; Allred, A. L. J. Chem. Soc. A 1970,
2987–2992.
16. Nanni, D.; Curran, D. P. Tetrahedron: Asymmetry 1996,
7, 2417–2422.
17. Stewart, J. J. P. J. Comp. Chem. 1991, 12, 320–341.
18. Stewart, J. J. P. QCPE 1993, Program 455 (version 93).
19. Murakata, M.; Tsutsui, H.; Hoshino, O. J. Chem. Soc.,
Chem. Commun. 1995, 481–482.
5.10. (R)-4-Neopentyl-4,5-dihydro-3H-dinaphtho[2,1-
c:1%,2%-e]stannepin 1b
Same procedure as described for the synthesis of 1a.
Yield 91%; mp 200°C (brown), 250°C (dec.); [h]2D0=
−34.5 (c 0.15, benzene); IR (KBr) w˜=1740 cm−1 (SnH);
1H MR (500 MHz, C6D6) l 1.92 (2H, m), 2.06 (9H, s),
2.45–2.51 (2H, m), 2.66–2.74 (2H, m), 5.94 (1H, d,
3J=4.4 Hz), 6.79–7.79 (12H, m); 13C NMR (125.7
MHz, C6D6) l 23.31, 24.14, 31.17, 29.27, 30.16, 125.32–
138.46 (20 C); MS (isobutane, CI) m/z (%)=470 (30)
[M−H2]+; MS (70 eV, EI) m/z (%): 471 (98) [M+−H],
20. Murakata, M.; Tsutsui, H.; Hoshino, O. Heterocycles
1997, 46, 517–522.
21. Sibi, M. P.; Ji, J. J. Am. Chem. Soc. 1996, 118, 3063–
3064.