A new class of chiral organogermanes derived from C2-symmetric dithiols: Synthesis, characterization and stereoselective free radical reactions

Article abstract of DOI:10.1021/jo026625s

A new class of dithiostannanes and dithiogermanes have been prepared from 1,1′-binaphthyl-2,2′-dithiol and 3,3′-bis(trimethylsilyl)-1,1′ -binaphtho-2,2′-dithiol. While reduction of 4-butyl-4-chloro-3,5-dithia-4-stanna-cyclohepta[2,1-a;3,4- a′]dinaphthalen

Full text of DOI:10.1021/jo026625s

VOLUME 68, NUMBER 13
J UNE 27, 2003
©
Copyright 2003 by the American Chemical Society
A New Cla ss of Ch ir a l Or ga n oger m a n es Der ived fr om
C
2
-Sym m etr ic Dith iols: Syn th esis, Ch a r a cter iza tion a n d
Ster eoselective F r ee Ra d ica l Rea ction s
,
‡
Giovanna Gualtieri,* Steven J . Geib,* and Dennis P. Curran*
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260
curran@pitt.edu
Received October 29, 2002
A new class of dithiostannanes and dithiogermanes have been prepared from 1,1′-binaphthyl-2,2′-
dithiol and 3,3′-bis(trimethylsilyl)-1,1′-binaphtho-2,2′-dithiol. While reduction of 4-butyl-4-chloro-
3
,5-dithia-4-stanna-cyclohepta[2,1-a;3,4-a′]dinaphthalene to the corresponding tin hydride was
unsuccessful, 4-tert-butyl-3,5-dithia-4-germa-cyclohepta[2,1-a;3,4-a′]dinaphthalene and 4-tert-butyl-
,6-bis(trimethylsilyl)-3,5-dithia-4-germa-cyclohepta[2,1-a;3,4-a′]dinaphthalene were obtained by
reduction of the parent germanium chlorides with NaBH and LiBH , respectively. Kinetic constants
2
4
4
for hydrogen transfer to a primary alkyl radical were measured for both germanium hydrides.
Reduction of R-halo carbonyl compounds by these germanium hydrides occurs with moderate ee
values (up to 42%), while hydrogermylation of methyl methacrylate occurs with low selectivity
(<3/1) for the former hydride but high selectivity (>10/1) for the latter.
In tr od u ction
center, in 1996 we introduced the first chiral tin hydride
4
derived from a C
2
-symmetric binaphthyl ligand. Since
Stereoselectivity in reactions of acyclic radicals is a
rapidly growing, yet still challenging field of radical
then, the groups of Metzger5 and Schiesser6 have syn-
2
thesized more chiral tin hydrides based either on a C -
1
chemistry. One way in which stereoselectivity can be
symmetric substituent or chiral ligands and proven that
they can indeed be used as stereoselective radical reduc-
ing agents (either alone or in conjunction with chiral
achieved is by use of chiral reagents that are analogues
of the popular tin hydride and, to a lesser extent, silicon
and germanium hydride reducing agents. Chiral, enan-
tiomerically pure, tin and germanium hydrides are
known,2 but to this date only a few of them have been
used for stereoselective free radical reactions.
7
6
Lewis acids) with high ee values (>90%).
We continued to elaborate on the C -symmetric derived
metal hydrides by exploring different patterns of substi-
tution in the C -symmetric ligand. We started by explor-
2
,3
2
To avoid possible racemization problems associated
with metal hydrides possessing an asymmetric metal
ing the replacement of metal-carbon bonds with metal-
sulfur bonds, and observing how this affects both the
stability and the reactivity of the resulting metal hy-
drides. This eventually led us to synthesize a new class
of chiral germanium hydrides as alternatives to the
traditional tin hydrides, and a preliminary report on
‡
Present address: Berlex Biosciences, 15049 San Pablo Avenue,
P.O. Box 4099, Richmond, CA 94804-4578.
1) (a) Sibi, M. P.; Porter, N. A. Acc. Chem. Res. 1999, 32, 163. (b)
(
Curran, D. P.; Porter, N. A.; Giese, B. Stereochemistry of Radical
Reactions; VCH: Weinheim, Germany, 1996. (c) Giese, B. Angew.
Chem., Int. Ed. Engl. 1989, 28, 969.
8
some of this work has appeared. We now report the
(
2) (a) Gielen, M. Top. Curr. Chem. 1982, 104, 57. (b) Vitale, C. A.;
(4) Nanni, D.; Curran, D. P. Tetrahedron: Asymmetry 1996, 7, 2417.
(5) (a) Schwarzkopf, K.; Blumenstein, M.; Hayen, A.; Metzger, J .
O. Eur. J . Org. Chem. 1998, 177. (b) Blumenstein, M.; Schwarzkopf,
K.; Metzger, J . O. Angew. Chem., Int. Ed. Engl. 1997, 36, 235. (c)
Schwarzkopf, K.; Metzger, J . O.; Saak, W.; Pohl, S. Chem. Ber./ Recl.
1997, 130, 1539.
Podest a´ , J . C. J . Chem. Soc., Perkin Trans. 1 1996, 2407 and references
therein. (c) Schumann, H.; Wassermann, B. C.; Pickardt, J . Organo-
metallics 1993, 12, 3051 and referenced cited therein. (d) Pratt, R. M.;
Thomas, E. J . J . Chem. Soc., Perkin Trans. 1 1998, 727 and referenced
cited therein.
(
3) (a) Taoufik, M.; Santini, C. C.; Basset, J .-M. J . Organomet. Chem.
(6) (a) For a recent review see: Schiesser C. H. Aust. J . Chem. 2001,
54, 89. (b) Schiesser C. H. Aust. J . Chem. 2001, 54, 199. (c) Dakternieks,
D.; Dunn, K.; Perchyonok, V. T.; Schiesser, C. H. Chem. Commun.
1999, 1665.
(7) For a review of Lewis acids in radical reactions see: Renaud,
P.; Gerster, M. Angew. Chem., Int. Ed. 1998, 37, 2563.
(8) Curran, D. P.; Gualtieri, G. Synlett 2001, SI, 1038.
1
999, 580, 128. (b) Colesdan, F.; Castel, A.; Rivi e` re, P.; Satg e´ , J .; Veith,
M.; Huch, V. J . Organomet. Chem. 1998, 17, 2222. (c) Tacke, R.; Kosub,
U.; Wagner, S. A.; Bertermann, R.; Schwarz, S.; Merget, S.; G u¨ nther,
K. Organometallics 1998, 17, 1687. (d) Eaborn, C.; Simpson, P.; Varma,
I. D. J . Chem. Soc. (A) 1966, 1133. (e) Brook, A. G.; Peddle, G. J . D. J .
Am. Chem. Soc. 1963, 85, 2338.
1
0.1021/jo026625s CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/05/2003
J . Org. Chem. 2003, 68, 5013-5019
5013

Gualtieri et al.
SCHEME 1. Syn th esis of Dith iosta n n a n es 2a -d
1
F IGURE 1. Representative H NMR resonances of 2a under
different conditions.
synthesis, detailed characterization, and reactivity data
for these novel germanium hydrides.
Resu lts a n d Discu ssion
Syn th esis a n d Ch a r a cter iza tion . The tin hydride
that we had previously synthesized (4,5-dihydro-4-meth-
yl-3H-dinaphtho[2,1-c:1′,2′-e]stannepin)4 suffered from
partial cleavage of the benzylic carbon-tin bond in
solution at room temperature. To strengthen the bond
F IGURE 2. Postulated dynamic chlorine exchange at the tin
center in 2a .
to tin, we decided to replace the benzylic -CH
2
- groups
clearly defined doublets at 6.89 (J ) 8.4 Hz) and 6.96
ppm (J ) 8.4 Hz) or as one broad doublet at 6.93 ppm
with sulfur atoms, by using 1,1′-binaphthyl-2,2′-dithiol
9
1
a
as the starting template. Racemic 1,1′-binaphthyl-
,2′-dithiol 1a reacted smoothly with 1.1 equiv of R SnCl
(R ) Bu or Ph) in the presence of triethylamine
to afford compounds 2a -d in excellent yields (Scheme
a). Alternatively, treatment of 1a with butyl(chloro)tin
diisopropoxide [Bu(Cl)Sn(Oi-Pr) ] or dibutyltin dimeth-
oxide [Bu Sn(OMe) ] in toluene at 130 °C with azeotropic
(J ) 8.2 Hz) (Figure 1, patterns A and C, respectively).
2
2
2
Other spectra showed intermediate patterns, where the
two doublets had coalesced into one very broad singlet
or RSnCl
3
(Figure 1, pattern B). It was found that 0.02 M samples
1
in CDCl at room temperature reproducibly showed the
3
2
aromatic pattern B. Maintaining this concentration, the
2
2
resonances further coalesced into pattern C at T > 40
removal of the alcohol also produced good yields of 2a
and 2c (Scheme 1a). All these new organotin compounds
were fully characterized and are stable both as solids and
in solution in the air, indicating that the tin-sulfur bond
is hydrolytically stable.10 Butyltin chloride 2a formed
colorless crystals suitable for X-ray crystal structure
analysis (see Supporting Information).
The appearance of the 1H NMR spectrum of 2a
depended on concentration and temperature of the
sample. Several spectra recorded from different samples
of 2a showed different aromatic patterns. Initially, in
some spectra the 12 aromatic protons were all different
from each other, as expected since the molecule is not
°
C and decoalesced into pattern A at T < 0 °C. Con-
versely, at room temperature, the spectrum varied from
pattern C to pattern A as the concentration of the sample
was decreased from 0.5 to 0.0008 M.
We postulated that this behavior could be due to small
amounts of free chloride ions in solution that promote a
dynamic exchange of the chlorine ligands on tin, probably
through a pentacoordinated symmetric tin species as
shown in Figure 2. It is known that pentacoordinated tin
1
1
species have a trigonal-bipyramid arrangement.
To test this hypothesis, a small crystal of AgBF
4
was
added to a 0.02 M NMR sample of 2a (pattern B) to trap
the free chloride ions. The aromatic region of the spec-
2
C -symmetric. In other spectra, the six protons on one
naphthyl unit were magnetically equivalent to the cor-
responding protons on the other naphthyl unit, as if the
trum became non-C
small crystal of Bu NCl was added to that same sample,
the aromatic region simplified into the C -symmetric
2
-symmetric (pattern A). When a
4
2
molecule were C
2
-symmetric. In particular, the two most
pattern (pattern C). Apparently, at high dilution or low
temperature the process of chlorine exchange is slow
upfield diastereotopic protons appeared either as two
2
relative to the NMR time scale, therefore, the non-C -
(
9) Fabbri, D.; Delogu, G.; De Lucchi, O. J . Org. Chem. 1993, 58,
748.
10) (a) Chemistry of Tin, 2nd ed.; Smith, P. J ., Ed.; Blackie: London,
1
(
(11) (a) Shibata, I.; Suwa, T.; Baba, A. Main Group Met. Chem.
2001, 24, 637. (b) Pereyre, M.; Quintard, J .-M.; Rahm, A. Tin in
Organic Synthesis; Butterworth: London, UK, 1987.
UK, 1997. (b) Davies, A. G. Organotin Chemistry; VCH: Weinheim,
Germany, 1997.
5
014 J . Org. Chem., Vol. 68, No. 13, 2003

A New Class of Chiral Organogermanes
SCHEME 2. Syn th esis of Dith ioger m a n es 7a a n d
7
b a n d Ger m a n iu m Hyd r oxid e 6
symmetric pattern is observed. At high concentration or
temperature, the rate of chloride exchange increases and
the peaks of the diastereotopic aromatic protons coalesce.
A facile exchange of electronegative ligands such as Cl
and S between several organotin mercaptoesters has also
1
2
been observed by other groups. We viewed the mobility
of the chloro atom attached to tin in 2a as a promising
feature for further manipulation of this functionality.13
Unfortunately, all efforts to reduce (rac)-2a to the
corresponding tin hydride (rac)-3a failed. Mild reducing
agents (Bu
reacted starting material, while stronger reducing agents
LiAlH , DIBAL) induced cleavage of the sulfur-tin bond
3 3 3 4
SnH, Ph SnH, AlH , NaBH ) returned un-
(
4
to give back dithiol 1a . The chloride 2a was easily
converted into the iodide 4a by treatment with NaI in
acetone, but attempted reduction of the Sn-I bond was
once again unsuccessful (Scheme 1b).14
To strengthen the bond to sulfur we next turned our
attention to the nuclear substitution of tin by germanium.
Although trialkylgermanium hydrides are poorer hydro-
1
5
gen donors than their corresponding tin hydrides, there
is reason to expect that the thio substituents could have
F IGURE 3. X-ray crystal structures of germanium chlorides
a (a) and 5b (b).
5
an activating effect on the hydrogen donor ability of
germanium hydrides.16
ature in THF as shown in Scheme 2 (the scheme shows
syntheses of one enantiomer, but the other enantiomer
and racemic samples were also made by the same route).
Dilution of the reaction mixture with diethyl ether and
filtration consistently afforded 95-100% yield of clean
product 5a as a white solid. The germanium chloride 5a
was very stable both as a solid and in solution, being only
prone to hydrolysis on silica gel. Colorless crystals
suitable for X-ray analysis were obtained from chloroform
As with its tin counterpart, the reaction between 1a
and commercially available tert-butyl germanium trichlo-
ride proceeded very smoothly over 1 h at room temper-
(12) (a) Burley, J . W.; Hutton, R. E.; Dworkin, R. D.; Hutton, R. E.
J . Organomet. Chem. 1985, 284, 171. (b) Blunden, S. J .; Hobbs, L. A.;
Smith, P. J .; Davies, A. G.; Teo, S. B. J . Organomet. Chem. 1985, 282,
9
. (c) Burley, J . W.; Dworkin. R. D. J . Vinyl Technol. 1986, 8, 15.
13) The ability of 2a to reduce alkyl halides under catalytic
conditions (NaCNBH /t-BuOH/benzene/AIBN) was tested in the reduc-
(
3
(Figure 3a).
tion of bromoadamantane. Gas chromatographic analysis of the
reaction mixture after 1 h at 80 °C showed 99% yield of adamantane
To evaluate the effect of bulky substituents on the
(
using naphthalene as internal standard). Sodium cyanoborohydride
ortho positions of the binaphthyl unit, we also prepared
alone does not reduce bromoadamantane under these conditions,
therefore an organotin hydride formed in situ is probably the active
reducing agent.
17
3
,3′-bis(trimethylsilyl)-1,1′-binaphtho-2,2′-dithiol 1b. This
was reacted with tert-butyl germanium chloride and
triethylamine in THF at room temperature to afford
crude 5b in quantitative yield. Due to the steric demand
of the ortho TMS groups, this reaction was slower than
that for the nonsilylated dithiol 1a , and was completed
in about 2 h (Scheme 2). Germanium chloride 5b was
(
14) Attempts to synthesize 3a directly by reaction of dithiol 1a
with BuSnH in benzene at 80 °C also failed, although the same
reaction between 1a and Bu SnH successfully afforded clean 2c.
15) (a) Chatgilialoglu, C.; Newcomb, M. Adv. Organomet. Chem.
999, 44, 67. (b) Chatgilialoglu, C.; Ballestri, M.; Escudi e´ , J .; Phailhous,
3
2
2
(
1
I. Organometallics 1999, 18, 2395. (c) Chatgilialoglu, C. Acc. Chem.
Res. 1992, 25, 188.
(
16) Chatgilialoglu, C.; Guerra, M.; Guerrini, A.; Seconi, G.; Clark,
K. B.; Griller, D.; Kanabus-Kaminska, J .; Martinho-Sim o˜ es, J . A. J .
Org. Chem. 1992, 57, 2427.
(17) Cossu, S.; De Lucchi, O.; Fabbri, D.; Valle, G.; Painter, G. F.;
Smith R. A. J . Tetrahedron 1997, 53, 6073.
J . Org. Chem, Vol. 68, No. 13, 2003 5015

Gualtieri et al.
obtained as a white solid, very stable to air and moisture.
Recrystallization from hexanes gave colorless crystals
suitable for X-ray crystallography (Figure 3b).
Analysis of the crystal structures of 5a and 5b shows
that the germanium atom has a distorted tetrahedral
geometry with angles between the tert-butyl group and
S(2) significantly larger than 109° (117° and 118°,
respectively) to accommodate the steric bulk of the tert-
butyl group, at the expense of the angles between the
chlorine and S(2) which are significantly smaller (101°
and 99°, respectively). As a result of the steric strain
imposed by the seven-membered ring, angles between the
two aromatic carbons at the binaphthyl junction are
slightly larger than 120°. The dihedral angle between the
four aromatic carbons of the ring increases from 76° in
5
a to 86° in 5b, to relieve the steric repulsion between
the tert-butyl group and the ortho TMS groups. Further-
more, in 5b, angles at the ortho carbons (bonded to the
TMS groups) are considerably larger than 120° as a result
of this steric repulsion (125° and 126°). The seven-
membered ring also has a distorted conformation, with
bond angles about S(1) and S(2) being different, and the
difference being larger in 5b than in 5a (95° and 104° vs
F IGURE 4. X-ray crystal structure of germanium hydride 7b.
1
judged by H NMR analysis), and was normally used as
such. If higher purity was needed, a small sample of 7a
was first filtered through a silica gel plug, then dissolved
in diethyl ether and washed with aqueous base. Alter-
natively, the crude hydride was purified by semiprepara-
tive chiral HPLC and obtained in g98% purity.
On the other hand, treatment of 5b with sodium
borohydride in THF at room temperature or at reflux did
produce some of the desired germanium hydride 7b, but
the reaction was not clean, as reduction of the Ge-S
bonds competed with reduction of the Ge-Cl bond. The
smaller and more powerful lithium borohydride proved
to be much more effective at room temperature or below
for the reduction of 5b, although some formation of dithiol
1
00° and 97°, respectively). Bond lengths of 2.18 and
2
.19-2.21 Å for the Ge-Cl and Ge-S bonds are within
1
8
normal ranges.
Although stable in the solid state, 5b was found to be
prone to partial hydrolysis in solution or on silica gel to
give the germanium hydroxide 6. The latter was also
independently prepared by treatment of 5b with an
aqueous solution of ammonium hydroxide in ether
(Scheme 2). The white solid obtained in quantitative yield
was recrystallized from hexanes and fully characterized,
including by X-ray crystallography (see Supporting
Information).
1
b was still observed. The hydroxide 6 was also reduced
For the reduction of chloride 5a to hydride 7a , we
effectively by lithium borohydride but provided a lower
hydride-to-dithiol ratio.19 Thus, we focused on optimizing
the reduction of the chloride 5b (Scheme 2).
screened some mild reducing agents (AlH
3 3 3
, Bu SnH, Ph -
SnH, NaCNBH , NaBH ) and found that sodium boro-
3
4
hydride was the best choice, although the reaction
conditions had to be carefully controlled. Thus, 5a was
treated with sodium borohydride in THF at room tem-
perature, and the desired germanium hydride 7a was
obtained as the major product. The yield of 7a was
dependent on the reaction time, and dithiol 1a formed
upon prolonged stirring so it was important to stop the
reaction as soon as the starting material had reacted
completely (typically about 1 h and 40 min). In reaction
scales larger than 1 mmol, 2-10% of dithiol 1a was
usually found as a byproduct, together with variable
amounts of another unidentified byproduct (2-10%).
Short column chromatography was effective in removing
most of the impurities of the crude product, but at the
expense of 30-40% material loss.
To minimize formation of byproducts it was best to
carry out the reaction at low temperature. Thus, a
solution of 5b and an excess of lithium borohydride (20
equiv) in THF was kept at -8 °C for 5 days, after which
the reaction mixture was worked up with a dilute
solution of ammonium chloride. The residue usually
contained less than 15% of combined dithiol and germa-
nium dihydride byproduct.20 After recrystallization from
hexanes, germanium hydride 7b was obtained with a
purity of g95% in 65-80% yield. The crystals obtained
were suitable for X-ray crystal structure analysis (Figure
4
). As observed in 5b, angles at S(1) and S(2) are
significantly different (93° vs 105°), the aromatic carbons
at the binaphthyl junction have angles larger than 120°
and the TMS groups are pushed back from the tert-butyl
group, with angles at the ortho carbons of 126° and 127°.
The dihedral angle between the four aromatic carbons
of the seven-membered ring is also large (84.6°).
The most reliable method to obtain 7a in purity g90%
and good yield was to run the reaction on a small scale
(0.3-0.5 mmol). After completion and standard extractive
workup, the dithiol byproduct was removed by washing
the organic layer with an aqueous solution of sodium
carbonate (pH 11.5-12). The germanium hydride 7a was
obtained from the organic layer with a dithiol content of
less than 2% molar and 2-5% of other impurities (as
(
(
19) Gualtieri, G. Ph.D. Thesis, University of Pittsburgh, 2000.
20) The byproduct found in the crude reaction mixture is believed
to be the C -symmetric dithiogermanium dihydride, as judged by the
2
1
singlet at 5.7 ppm (2H) in the H NMR spectrum, which correlate with
a singlet at 0.57 ppm (18H). This byproduct, probably the result of
â-elimination at the germanium center with loss of isobutene, was
never isolated and fully characterized, but it was found in small
amounts in almost every reaction condition employed.
(
18) Baines, K. M.; Stibbs, G. W. Coord. Chem. Rev. 1995, 145, 157-
2
00.
5
016 J . Org. Chem., Vol. 68, No. 13, 2003

A New Class of Chiral Organogermanes
SCHEME 3. Ra d ica l Ch a in Red u ction w ith 7a
TABLE 1. Deter m in a tion of th e Kin etic Con sta n t for
Hyd r ogen Tr a n sfer fr om 7a a n d 7b to a P r im a r y Alk yl
Ra d ica l
7
(
rac)-1a
T
10 kH(T)
-
1
-1
run hydride (mol %)
[7]
[9]
11/12 (°C) (M
s )
1
2
3
4
5
6
7
8
(rac)-7a
(rac)-7a
(rac)-7a
(rac)-7a
(rac)-7a
(rac)-7b
(rac)-7b
(rac)-7b
1.5
1.5
0.15 0.05
0.10 0.05
0.10 0.05
0.10 0.05
0.10 0.05
0.08 0.04
0.16 0.08
0.21 0.11
1.38
0.73
0.85
0.87
1.06
0.37
0.61
0.89
83
83
83
83
83
80
80
80
2.8
2.5
2.9
3.0
3.7
1.4
1.2
1.4
a
3
6
6
a
a
0
0
0
SCHEME 4. Com p etition Kin etic Exp er im en ts
w ith Ra d ica l Clock 9
a
Extra dithiol 1a added by doping.
Since the sample of 7a used for these reactions was
contaminated by 1.5% molar of dithiol 1a , we were
concerned that the dithiol could act as a polarity-reversal
catalyst, thus competing with the germanium hydride for
hydrogen transfer to the intermediate carbon radical.
This well-known phenomenon has been exploited in the
2
4
radical hydrosilylation of alkenes with silicon hydrides.
This hydride proved to be remarkably stable as a solid
and could be stored for long periods of time in closed
containers without any precautions. In solution, it is
somewhat sensitive to air and moisture, but oxidation
or hydrolysis are quite slow, so handling of this product
is very simple and practical.
Ra te Con sta n t Mea su r em en ts. The reactivity of 7a
as a radical chain reducing agent was first tested for two
simple substrates, bromoadamantane and carbon tetra-
chloride, at 80 °C and with AIBN as radical initiator
To check if thiol catalysis existed, we ran some of the
kinetic experiments doping the reaction mixture with
additional amounts of dithiol 1a (3% and 6% molar,
entries 3-5). We anticipated observing a difference in
the calculated rate constants if the thiol was acting as a
catalyst.
As can be seen from Table 1, the range of values
obtained for the kinetic constant of 7a at increasing
amounts of added 1a fluctuates only within experimental
error (entries 1-5). Therefore, it appears that the pres-
ence of up to 12% molar of -SH groups does not
significantly interfere with the rate of hydrogen abstrac-
tion by the primary radical 10 from the germanium
hydride. The average kinetic constant obtained for 7a
(Scheme 3). In both cases the reaction proceeded smoothly
to give high yields of the reduced products. In the
reduction of bromoadamantane, 88% yield of adamantane
was calculated by GC with dodecane as internal stan-
dard. With carbon tetrachloride, proof of reduction was
obtained by comparing the proton NMR spectrum of the
resulting germanium chloride with that of an authentic
sample of 5a . A quantitative yield of 5a was obtained by
7
-1 -1
(
k
H(83) ) 3.0 × 10 M
s
) shows that this hydride is an
SnH, for which kH(83)
even better hydrogen donor than Bu
3
6
-1 -1 21
)
6.8 × 10 M
s .
1
For germanium hydride 7b the kinetic constant was
H NMR with 1,3,5-trimethoxybenzene as internal stan-
calculated as the average of three runs (entries 6-8, kH(80)
dard.
7
-1 -1
)
1.4 × 10 M
s ). The value obtained shows that 7b
To learn more about the intrinsic reactivity of 7a and
is a roughly 2-fold slower hydrogen donor than 7a ,
probably due to the bulky trimethylsilyl substituents on
the binaphthyl unit. Considering the beneficial effect that
sulfur ligands have on the hydrogen donor ability of
silicon hydrides,16 it is not surprising that a similar trend
is observed for germanium hydrides.
7
b, we measured the kinetic constants for hydrogen
transfer from both hydrides to a primary alkyl radical
2
1
using a calibrated radical clock (Scheme 4). From the
ratio of cyclized/reduced products (11/12) and knowing
2
2
k
c
,
the kinetic constant for hydrogen transfer can be
calculated from the appropriate equations (see Support-
ing Information).23
Ster eoselectivity of F r ee-Ra d ica l Tr a n sfor m a -
tion s. Given the good stability of these novel germanium
hydrides and their favorable kinetic constants as radical
chain propagating agents, we next tested them as enan-
tioselective radical reducing agents for the reduction of
racemic 3-iodo-3-methoxymethyl-2-chromanone 13 and
methyl 2-bromo-2-phenyl-3,3-dimethylbutanoate 15
(Scheme 5 and Table 2). Both enantiomers of 1a were
prepared by resolution of (rac)-1a according to the
The kinetic measurements were carried out by mixing
1
equiv of the 7-bromo-1,1-diphenyl-1-heptene 9 with 2-3
equiv of 7a or 7b and a catalytic amount of AIBN in
deuterated benzene in a sealed tube. The resulting
mixture was then heated at the appropriate temperature
(
80-83 °C) for 4-11 h. After removal of the solvent, an
internal standard was added and the yields and ratios
of cyclized/reduced products (11/12) were obtained from
1
25
H NMR analysis. The data for these experiments are
procedure reported by De Lucchi and co-workers,
thereby granting access to enantiopure 7a and 7b.
summarized in Table 1.
(
(
21) Newcomb, M. Tetrahedron 1993, 49, 1151.
22) Newcomb, M.; Horner, J . H.; Filipkowski, M. A.; Ha, C.; Park,
(24) (a) For a review of polarity-reversal catalysis of hydrogen
abstraction reactions see: Roberts, B. P. Chem. Soc. Rev. 1999, 28,
25. (b) Haque, M. B.; Roberts, B. P.; Tocher, D. A. J . Chem. Soc., Perkin
Trans. 1 1998, 2881. (c) Haque, M. B.; Roberts, B. P. Tetrahedron Lett.
1996, 37, 9123.
S.-U. J . Am. Chem. Soc. 1995, 117, 3674.
23) Ha, C.; Horner, J . H.; Newcomb, M.; Varick, T. R.; Arnold, B.
R.; Lusztyk, J . J . Org. Chem. 1993, 58, 1194.
(
J . Org. Chem, Vol. 68, No. 13, 2003 5017

Gualtieri et al.
SCHEME 6. Hyd r oger m yla tion of Meth yl
SCHEME 5. Red u ction of r-Ha lo Ca r bon yl
Com p ou n d s 13 a n d 15 w ith 7a a n d 7b
Meth a cr yla te w ith (r a c)- a n d (R)-7a ,b
TABLE 2. Red u ction s of Ra cem ic Iod id e 13 a n d
Br om id e 15 w ith Ger m a n iu m Hyd r id es 7a a n d 7b
T
yield
ee
(%)
a
entry substrate hydride (°C) initiator product (%)
1
2
3
4
5
6
7
8
9
13
13
13
13
15
15
15
15
15
(R)-7a -78 Et
3
B/Ar
14
14
14
14
16
16
16
16
16
97
96
70
41
79
94
80
54
34
20 (R)
26 (R)
41 (S)
40 (R)
(R)-7a -78 AIBN/hν
b
b
(R)-7b -78 Et
(S)-7b -78 Et
(R)-7a -78 Et
(R)-7a -78 Et
(R)-7a -78 AIBN/hν
(R)-7b -25 Et
(S)-7b -60 Et
3
3
3
3
B/Ar
B/O
B/Ar
B/O
2
c
22 (S)
27 (S)c
2
c,d
29 (S)
32 (R)
42 (S)
b
e
products were obtained in high yields but low ee values
(20-29%, Table 2, entries 1,2 and 5-7).26 With 7b, the
ee of 14 increased to 41% (Table 2, entries 3 and 4),
although initiation with AIBN was not effective. The
reduction of 15 with 7b at -25 °C gave slightly better ee
than was obtained with 7a at -78 °C for the same
substrate (Table 2, compare entry 8 with entries 5-7).
At -60 °C, the reaction became very slow, therefore it
was stopped after 2 days at incomplete conversion. In
this case, the ee increased further to 42% (Table 2, entry
3
B/Ar
B/Ar
f
3
a
By proton NMR analysis unless otherwise indicated. b Isolated
c
d
T
h
i
s
y
i
e
l
d
.
E
n
a
n
t
i
o
m
e
r
i
c
e
x
c
e
s
s
c
a
l
c
u
l
a
t
e
d
f
o
r
t
h
e
a
l
c
o
h
o
l
1
7
.
reaction was repeated in identical conditions and gave 87% yield
and 20% ee. Conversion of starting material incomplete. f This
reaction was repeated in identical conditions and gave 37% yield
and 41% ee.
e
3
In a typical run, the radical initiator (AIBN or Et B)
was added to a solution of the iodide 13 or bromide 15 (1
equiv) and the germanium hydride (2 equiv) in toluene
at the appropriate temperature. If needed, additional
triethylborane was added in several portions, until the
starting material had reacted completely. When TLC
analysis showed complete consumption of starting mate-
rial, the reaction mixture was concentrated, and p-
dimethoxybenzene was added as internal standard to
determine the yield of product by proton NMR spectros-
copy. The crude product was then purified by column
chromatography and analyzed by chiral HPLC to obtain
the enantiomeric excess.
9
). As expected, opposite enantiomers of 7b afforded
opposite enantiomers of reduced products (Table 2,
entries 3,4 and 8,9). The absolute configurations of the
reduced products were assigned by comparing their R
to the literature values.
Introduction of the ortho TMS groups on the binaph-
thyl ring of the hydride resulted not only in higher
enantiomeric excesses of the reduced substrates 14 and
D
1
6 but also in a reversal of the selectivity. The same
enantiomers of 7a and 7b afforded preferentially opposite
enantiomers of the reduced products (Table 2, entries 2,3
and 7,8).
In the particular case of the reduction of bromide 15
by 7a , isolation of pure 16 from the reaction mixture
proved rather cumbersome due to partial hydrolysis on
silica gel of the excess germanium hydride and the
byproduct germanium bromide. We found that reduction
of the crude reaction mixture with LAH produced the
alcohol 17 and dithiol 1a , which were easily separated
through a silica gel plug. Therefore, the enantiomeric
excesses for these reductions were obtained for the
alcohol 17, rather than the ester 16.
To investigate diastereoselective transformations with
both racemic and enantiopure 7a and 7b, we next studied
the hydrogermylation reactions of methyl methacrylate,
which give diastereomeric products 18a /19a and 18b/19b
(Scheme 6).
Methyl methacrylate (1.2-2 equiv), AIBN, and the
germanium hydride (1 equiv) dissolved in toluene (ben-
(
26) To verify whether any of the observed ee values could be
ascribable to traces of (R)-1a present in the reaction mixtures, we ran
a series of control experiments in which 15 was reduced with Bu SnH
Reduction of both halides with (R)-7a proceeded well
with either AIBN or Et B as initiator and the reduced
3
3
in the presence of variable amounts of (R)-1a (from 0.1 up to 1 equiv).
In all cases, the reduced product 16 was obtained as racemic, indicating
that if 1a takes part in the hydrogen abstraction step it does so without
any stereoselectivity. This is expected since the structure of 1a is not
as rigid and sterically demanding as that of 7a .
(
25) De Lucchi, O.; Fabbri, D.; Delogu, G. J . Org. Chem. 1995, 60,
6
599.
5
018 J . Org. Chem., Vol. 68, No. 13, 2003

A New Class of Chiral Organogermanes
TABLE 3. Hyd r oger m yla tion of Meth yl Meth a cr yla te
w ith Ger m a n iu m Hyd r id es 7a a n d 7b
the minor isomer was characterized by NMR spectroscopy
in CD
Cl or benzene-d solutions.
2
2
6
entry
hydride
T (°C)
products
yield (%)a
ratio
We were unable to obtain crystals of either diastere-
omer 18b or 19b; however, reduction of the major adduct
18b with DIBAL at -60 °C afforded the corresponding
alcohol 20, which was recrystallized from hexanes and
found to have the (R,R) relative configuration by X-ray
crystallography (see Supporting Information).
1.8:1b
1.7:1
3:1
1
2
3
4
5
6
7
8
(rac)-7a
(rac)-7a
(R)-7a
(rac)-7b
(R)-7b
(rac)-7b
(R)-7b
(rac)-7b
25
-78
-78
25
(rac)-18a /19a
(rac)-18a /19a
(R)-18a /19a
(rac)-18b/19b
(R)-18b/19b
(rac)-18b/19b
(R)-18b/19b
(rac)-18b/19b
57
77
61
61
56
63
b
b
c
11:1
d
15:1c
14:1c
25
-20
-20
-60
d
c
The stereoselectivity of the hydrogermylation reactions
with 7a and 7b follows the same course, yielding the
55
54
25:1
18:1c
(
R,R)-diastereomer as the major product in both cases.
a
b
Isolated yield of mixture unless otherwise indicated. By
c
d
Since in the reduction of halides 13 and 15 the same
enantiomers of 7a and 7b afforded opposite enantio-
enriched products, this result suggests that in the hy-
drogermylation it is the Ge atom â to the radical center
that dictates the stereochemical outcome of the hydrogen
abstraction step. The approaching germanium hydride
has a secondary effect, and enhances the selectivity if it
is of the same (R) configuration (matched pair). Com-
parison of the results with racemic and enantiopure
hydrides also supports this hypothesis and indicates that
the difference in energy between the matched and the
mismatched pairs is not very large.
proton NMR analysis. By HPLC analysis. Isolated yield of 18b.
zene for reactions at room temperature) were placed in
a sealed tube purged with argon. The mixture was
irradiated with a 450-W medium-pressure mercury lamp
and stirred at the appropriate temperature until TLC
analysis showed complete consumption of the germanium
3
hydride. In the reactions with Et B, 1.5 equiv of methyl
methacrylate was used, and triethylborane was added
to a solution of the reactants at the appropriate temper-
ature. At -60 °C, the reaction with 7b became sluggish
and addition of more triethylborane was necessary to
drive it to completion. The crude mixture was concen-
trated under reduced pressure and analyzed either by
proton NMR spectroscopy or by analytical HPLC to
obtain the ratio of products.27 Then, the product was
purified by column chromatography or preparative HPLC
to afford the adducts 18a /19a and 18b/19b. The yields
and diastereomeric ratios obtained from this series of
experiments are summarized in Table 3.
Con clu sion s
We have synthesized the first chiral enantiomerically
2
pure germanium hydrides 7a ,b containing a C -sym-
metric binaphthyl substituent and two germanium-
sulfur bonds. Although care is required in the last part
of the synthesis, the hydrides are very stable as solids
and can be conveniently handled without any special
precautions.
In the reactions run with racemic 7a , isolated yields
of combined 18a /19a were moderate (57-77%) but the
diastereomer ratio was low (1.7:1) and independent of the
reaction temperature (virtually no change from -78 °C
to room temperature, Table 3, entries 1 and 2). When
enantiomerically pure germanium hydride was used, the
diastereomer ratio increased to 3:1 (Table 3, entry 3). The
minor isomer 19a was isolated by column chromatogra-
phy and the X-ray crystal structure was solved to assign
the relative configuration, which was found to be (R,S)
The kinetic constants for hydrogen transfer from 7a ,b
to primary alkyl radicals are 2-4 times higher than that
of tributyltin hydride, indicating that the two Ge-S
bonds significantly enhance the hydrogen donor ability
of the germanium hydrides. These new germanium
hydrides are effective free-radical chain hydrogen trans-
fer agents.
As anticipated, the introduction of bulky TMS groups
on the ortho positions of the binaphthyl units resulted
in enhanced selectivity and decreased reactivity of the
germanium hydride. High diastereoselectivity was ob-
tained in the hydrogermylation of methyl methacrylate
with (R)-7b.
(see Supporting Information).
The diastereoselectivity of the hydrogermylation was
considerably higher when the more sterically biased
hydride 7b was used as a racemate (Table 3, 11:1, entry
To the best of our knowledge this is the first application
of chiral germanium hydrides in stereoselective radical
chain reactions. The synthesis of 7a and 7b should serve
as a starting point for the design and synthesis of
improved tin, germanium, or silicon hydrides of this type.
4
), and increased further at lower temperature to 18:1
(Table 3, entries 6 and 8). Also in this case, there was a
slight increase in selectivity in reactions run with enan-
tiomerically pure (R)-7b as opposed to racemic 7b (Table
3
, compare entries 4 and 6 with entries 5 and 7). The
major and minor products 18b/19b were isolated either
as a mixture by column chromatography or separately
by semipreparative HPLC. Isolated yields of combined
Ack n ow led gm en t. We thank Paul Badger for help
with the X-ray crystallographic data. We thank the
National Science Foundation for support of this work.
We thank Samantha Haedrich for preparing the cover
art.
1
8b and 19b were generally moderate (54-63%), al-
though they were not optimized. While the major product
8b was very stable, the minor isomer 19b was found to
be unstable in CDCl solution, decomposing rather
quickly into an unidentified yellow byproduct. However,
1
Su p p or tin g In for m a tion Ava ila ble: Complete experi-
mental details, copies of the 1H and C NMR spectra of all
the compounds, and crystallographic details. This material is
available free of charge via the Internet at http://pubs.acs.org.
3
13
(27) The 18a to 19a ratios were determined by proton NMR
spectroscopy, by integration of the -OCH resonances.
3
J O026625S
J . Org. Chem, Vol. 68, No. 13, 2003 5019