Reactivity of heterobimetallic alkoxysilyl and siloxy complexes in the catalytic dehydrogenative coupling of tin hydrides

Article abstract of DOI:10.1021/om970353f

Heterobimetallic complexes [(OC)₃(R₃Si)Fe(μ-dppm)Pd(η³-allyl)] 1 (R = OMe, Me, OSiMe₃, OSi(H)Me₂) and [(OC)₃Fe{μ-Si(OR)₂(OR)}(μ-dppm)PdCl] 6 (R = Me, SiMe₃) are effective catalyst in the dehydrogenative coupling of triorganotin hydrides HSnR′₃ (R′ = Ph, ⁿBu). Although the elementary transformations during catalysis appear to take place at palladium, the function of the iron fragment is to provide the palladium center with the appropriate coordination environment through metal-metal bonding and the Si-containing ligand. Indeed, complexes 1 and 6 revealed a higher catalytic activity than mononuclear Pd catalysts. Modifications of the substituents at silicon resulted in considerable variations of the TON (turnover number) and TOF (turnover frequency) values as well as in the lifetime of the catalysts. In the case of siloxy derivatives, TON and TOF values higher than in the case of the alkoxysilyl analogs have been obtained whereas the lifetime of the catalyst is much longer for the latter. Possible mechanisms which rationalize the role of the silicon ligand are discussed. Solvent effects have also been observed. One of the key features of these systems is the retention of the bimetallic nature of the catalyst. TON and TOF higher than 2 × 10⁵ and 3 × 10⁷ h^-1, respectively, have been obtained in the case of HSnⁿBu₃. The catalytic activity of 1 toward the dehydropolymerization of tin dihydrides has been tested.

Full text of DOI:10.1021/om970353f

5
40
Organometallics 1998, 17, 540-550
Rea ctivity of Heter obim eta llic Alk oxysilyl a n d Siloxy
Com p lexes in th e Ca ta lytic Deh yd r ogen a tive Cou p lin g of
Tin Hyd r id es
Pierre Braunstein* and Xavier Morise
Laboratoire de Chimie de Coordination, Associ e´ au CNRS (UMR 7513),
Universit e´ Louis Pasteur, 4 rue Blaise Pascal, F-67070 Strasbourg Cedex, France
Received April 25, 1997
3
Heterobimetallic complexes [(OC)
OSi(H)Me ) and [(OC) Fe{µ-Si(OR)
catalyst in the dehydrogenative coupling of triorganotin hydrides HSnR′
3
(R
3
Si)Fe(µ-dppm)Pd(η -allyl)] 1 (R ) OMe, Me, OSiMe
3
,
2
3
2
(OR)}(µ-dppm)PdCl] 6 (R ) Me, SiMe
3
) are effective
n
3
(R′ ) Ph, Bu).
Although the elementary transformations during catalysis appear to take place at palladium,
the function of the iron fragment is to provide the palladium center with the appropriate
coordination environment through metal-metal bonding and the Si-containing ligand.
Indeed, complexes 1 and 6 revealed a higher catalytic activity than mononuclear Pd catalysts.
Modifications of the substituents at silicon resulted in considerable variations of the TON
(turnover number) and TOF (turnover frequency) values as well as in the lifetime of the
catalysts. In the case of siloxy derivatives, TON and TOF values higher than in the case of
the alkoxysilyl analogs have been obtained whereas the lifetime of the catalyst is much
longer for the latter. Possible mechanisms which rationalize the role of the silicon ligand
are discussed. Solvent effects have also been observed. One of the key features of these
systems is the retention of the bimetallic nature of the catalyst. TON and TOF higher than
5
7
-1
n
2
× 10 and 3 × 10 h , respectively, have been obtained in the case of HSn Bu
3
. The
catalytic activity of 1 toward the dehydropolymerization of tin dihydrides has been tested.
In tr od u ction
During the last few years, we have been interested
in studying the properties of metal-silicon bonds in a
heterobimetallic environment.1 This led to the dis-
covery of a new class of complexes that display unusual
Fe-Si-O-M four-membered rings involving the silicon
backbone. The labile character of the dative oxygen-
,2
1
metal interaction, shown by VT H NMR studies, may
We have already reported that treatment of [(OC)3-
represent an interesting tool for catalytic purposes
where the storage of a “masked” coordination site plays
a key role.
3
{
(MeO)3Si}Fe(µ-dppm)Pd(η -allyl)] 1a (dppm ) Ph2-
PCH2PPh2) with excess HSnPh3 afforded, via propene
elimination, the trinuclear complex [(OC)3Fe{Si(OMe)2-
(OMe)}(µ-dppm)Pd(SnPh3)] (2a ), which is stabilized by
(1) (a) Braunstein, P.; Knorr, M.; Tiripicchio, A.; Tiripicchio Camel-
lini, M. Angew. Chem., Int. Ed. Engl. 1989, 28, 1361. (b) Braunstein,
P.; Knorr, M.; Piana, H.; Schubert, U. Organometallics 1991, 10, 828.
an intramolecular Fe-Si-O-Pd four-membered ring
eq 1). This reaction was also accompanied by evolu-
tion of H2 and formation of Sn2Ph6 (ca. 10%). Interest-
ingly when HSnPh3 was added to a solution of
(
c) Braunstein, P.; Knorr, M.; Schubert, U.; Lanfranchi, M.; Tiripicchio,
A. J . Chem. Soc., Dalton Trans. 1991, 1507. (d) Braunstein, P.; Knorr,
M.; Villarroya, E.; DeCian, A.; Fischer, J . Organometallics 1991, 10,
1b
(
3
6
4
714. (e) Knorr, M.; Braunstein, P. Bull. Soc. Chem. Fr. 1992, 129,
63. (f) Knorr, M.; Faure, T.; Braunstein, P. J . Organomet. Chem. 1993,
47, C4. (g) Braunstein, P.; Faure, T.; Knorr, M.; Balegroune, F.;
Grandjean, D. J . Organomet. Chem. 1993, 462, 271. (h) Braunstein,
P.; Knorr, M.; Strampfer, M.; Tiripicchio, A.; Ugozzoli, F. Organome-
tallics 1994, 13, 3038. (i) Knorr, M.; Braunstein, P.; Tiripicchio, A.;
Ugozzoli, F. Organometallics 1995, 14, 4910. (j) Braunstein, P.; Knorr,
M.; St a¨ hrfeldt, T. J . Chem. Soc., Dalton Trans. 1994, 1913 (k)
Braunstein, P.; Faure, T.; Knorr, M.; St a¨ hrfeldt, T.; DeCian, A.;
Fischer, J . Gazz. Chim. Ital. 1995, 125, 35. (l) Braunstein, P.; Knorr,
M. J . Organomet. Chem. 1995, 500, 21.
3
[
(OC)3{(Me3SiO)2MeSi}Fe(µ-dppm)Pd(η -allyl)] 1c, for-
mation of the Fe-Pd-Sn complex 2c was not evidenced.
Instead a vigorous gas evolution and a complete trans-
formation of HSnPh3 into Sn2Ph6 occurred (eq 1). These
observations clearly indicate that not only are complexes
1
a and 1c catalysts (precursors) for the dehydrogenative
(
2) Braunstein, P.; Morise, X.; Blin, J . J . Chem. Soc., Chem.
Commun. 1995, 1455.
tin-tin coupling of HSnPh3, but that the nature of the
S0276-7333(97)00353-1 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 01/24/1998

Heterobimetallic Alkoxysilyl and Siloxy Complexes
Organometallics, Vol. 17, No. 4, 1998 541
Sch em e 1
low catalytic activities were reported for the Pd and Yb
complexes, whereas in the case of the copper derivative
the tin-tin dehydrogenative coupling was observed as
a side reaction. It should also be mentioned that the
dehydropolymerization of H2SnR2 in the presence of
metallocene derivatives of Ti and Zr has been recently
silicon ligand strongly influences the course of the
reaction.
The wide range of industrial3-6 and biological4-7
applications of hexaorganoditin compounds and their
derived products has stimulated numerous studies
toward their synthesis. They are usually obtained by
various organic methods, such as condensation reactions
of HSnR3 with XSnR3 (X ) OH, NH2, halide,
1
8
described.
In the present paper, we wish to report investigations
on the dehydrogenative coupling of stannanes catalyzed
by heterobimetallic Fe-Pd alkoxysilyl and siloxy com-
plexes. In addition to its intrinsic interest, we have used
this reaction as a tool in order to study the influence of
the heterobimetallic structure of the catalysts and of the
ligand environment around the palladium center on the
reaction process and provide an insight into homoge-
OSnR3...),4
-6,8,9
alkylations of tin halide derivatives,4-6,10
Wurtz-type reactions between XSnR3 (X ) halide) and
4
-6,11
metallic Li or Na,
or reduction of bis(trialkyltin)-
oxides.1
2,13
Surprisingly, the dehydrogenative coupling
of HSnR3 has been little studied. Only a few examples
where this reaction is catalyzed by organometallic
1
9
1
4
neous heterobimetallic catalysis.
Some preliminary
complexes have been described: [PdCl2(MeCN)2], [Pd-
2
1
4
15
16
results have been communicated.
(
Ph)I(PPh3)2], [Pd(PPh3)4], [Cp′Yb(µ-H)]2 (Cp′ )
1
7
C5H4Me), and [Ph3PCuH]6 have been used, however,
Resu lts
(
3) Evans, C. J . Chemistry of Tin; Harrison, P. G., Ed.; Blackie:
Glasgow, 1989; Chapter 13 and references cited therein.
4) Davies, A. G.; Smith, P. G. Comprehensive Organometallic
1
.
Syn th esis of [(OC)3{(Me2HSiO)Me2Si}F e(µ-
(
Chemistry; Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; Perga-
mon: Oxford, 1982; Vol. 2, p 519 and references cited therein.
3
d p p m )P d (η -a llyl)] 1d . We have previously described
(5) Wardell, J .; Spencer, G. M. Encyclopedia of Inorganic Chemistry;
the synthesis of heterobimetallic Fe-Pd silyl complexes
King, R. B., Ed.; Wiley: New York, 1993; Vol. 8, p 4172 and references
cited therein.
3
1
a -c by treatment of the dinuclear complex [Pd(η -
(6) Weis, R. W. Organometallic Compounds, 2nd ed.; Springer: New
C3H5)(µ-Cl)]2 with 2 equiv of 4a -c (obtained by depro-
York, 1973; Vol. II, first supplement, pp 826-832 and references cited
therein.
tonation of the corresponding hydrido silyl complexes
1
b,g
[
HFe(SiR3)(CO)3(dppm-P)] (3), Scheme 1).
The new
(
7) Selwyn, M. J . Chemistry of Tin; Harrison, P. G., Ed.; Blackie:
Glasgow, 1989; Chapter 11 and references cited therein.
8) Glocking, F. Chemistry of Tin; Harrison, P. G., Ed.; Blackie:
Glasgow, 1989; Chapter 8 and references cited therein.
9) (a) Neumann, W. P.; Pedain, J . Tetrahedron Lett. 1964, 2461.
b) Neumann, W. P.; Schneider, B. Angew. Chem., Int. Ed. Engl. 1964,
3
complex [(OC)3{(Me2HSiO)Me2Si}Fe(µ-dppm)Pd(η -al-
lyl)] 1d has been synthesized according to this proce-
dure. The hydrido-silyl complex mer-[HFe{SiMe2-
(OSiHMe2)}(CO)3(dppm-P)] (3d ) was prepared by
irradiation of a hexane solution of [Fe(CO)5] and (Me2-
HSi)2O (ratio 1:4) followed by substitution of the car-
bonyl group trans to Si with dppm. Formation of the
intermediate cis-[FeH{SiMe2(OSiHMe2)}(CO)4] was moni-
tored by IR spectrometry (see Experimental Section).
The three ν(CO) absorptions expected for a meridional
arrangement of the CO ligands around the iron center
(
(
(
3
, 751.
(10) Wittig, G.; Meyer, E. J .; Lange, G. Liebigs Ann. Chem. 1951,
5
71, 167.
(11) Gr u¨ ttner, G. Chem. Ber. 1917, 50, 1808.
12) (a) Poller, R. C. The Chemistry of Organotin Compounds;
(
Logos: London, 1970. (b) Neumann, W. P. The Organic Chemistry of
Tin; Wiley: London, 1970. (c) Davies, A. G.; Osei-Kissi, D. K. J .
Organomet. Chem. 1994, 474, C8. (d) McAlonan, H.; Stevenson, P. J .
Organometallics 1995, 14, 4021 and references cited therein.
(
13) Ulrich, S. (Witco GmbH) Ger. Offen. DE 4.231.083; Chem. Abstr.
994, 120, 245521v.
14) Bugamin, N. A.; Gulevitch, Y. V.; Beletskaya, I. P. Izv. Akad.
Nauk. SSR, Ser. Khim. 1982, 11, 2639.
15) (a) Mitchell, T. N.; Killing, H.; Rutschow, D. J . Organomet.
1
(
(17) Schubert, U.; Mayer, B.; Russ, C. Chem. Ber. 1994, 127, 2189.
(18) (a) Imori, T.; Tilley, T. D. J . Chem. Soc., Chem. Commun. 1993,
1607. (b) Imori, T.; Lu, V.; Cai, H.; Tilley, T. D. J . Am. Chem. Soc.
1995, 117, 9931.
(
Chem. 1986, 304, 257. (b) Zhang, H. X.; Guibe, F.; Balavoine, G. J .
Org. Chem. 1990, 55, 1857 and references cited therein.
(
16) Voskoboynikov, A. Z.; Beletskaya, I. P. New J . Chem. 1995, 19,
(19) Braunstein, P; Ros e´ , J . Comprehensive Organometallic Chem-
istry II; Elsevier Science Ltd.: New York, 1995; Vol. 10, pp 351-85.
7
23.

5
42 Organometallics, Vol. 17, No. 4, 1998
Braunstein and Morise
F igu r e 1. Evolution of the catalytic activity of 1c in the
formation of Sn Ph from HSnPh in dichloromethane;
inset represents an expansion of the plot in the first 30 s
of the reaction.
F igu r e 2. Comparison between the catalytic activities of
the siloxyl complexes 1b (×), 1c (b) and 1d (0) in the
2
6
3
2 6 3
formation of Sn Ph from HSnPh in dichloromethane.
Ta ble 1. Selected Ca ta lysis Da ta for 1b, 1c, a n d 1d
in 3d are observed at 2050 (w), 1980 (s), and 1965 (vs)
in the IR spectrum (toluene), whereas the H NMR
catalyst
TON maxa
plateau, hb
TOFc
1
5
5
5
1
b
1380
1630
1800
>1
1
0.5
1.15 × 10
1.98 × 10
2.16 × 10
spectrum contains a doublet at δ -9.1 ppm (J PH ) 26
Hz) and a signal at δ 5.0 ppm characteristic of the Fe-H
and Si-H groups, respectively. Compound 3d was
obtained as an air-sensitive sticky material that rapidly
decomposes into the chelate complex [Fe(CO)3(dppm-
P,P)] after reductive elimination of silane.20 It was,
therefore, used as soon as prepared for subsequent
reactions. Treatment of 3d with 1 equiv of KH in THF
1c
1
d
a
Total catalytic activity expressed in turnover numbers (aver-
b
age value of a minimum of five experiments). Time after which
the plateau is reached. c Maximum turnover frequency expressed
in turnovers per h.
31
1
P and H NMR spectroscopic methods, after Sn2Ph6
had been separated, showed the presence of a set of
followed by addition of the resulting mixture to a cold
31
1
signals ( P{ H} NMR δ 43.62 (P(Fe), J PP ) 44 Hz,
3
(
233 K) THF solution of [Pd(η -C3H5)(µ-Cl)]2 (0.5 mol
1
19
117
J P Sn ) 169 Hz, and J P Sn ) 137 Hz), -23.47 (J PP )
4 Hz)) that we assigned to complex 5c by comparison
equiv) yielded 1d as an air-stable yellow-orange powder
after purification by flash chromatography on alumina
4
of these data with those of an authentic sample pre-
pared by reaction of the anion 4c with ClSnPh3. The
Sn-P coupling observed in the downfield signal indi-
cates a cis arrangement of these two nuclei about the
Fe center, as represented below. However, this com-
(
eluent ) Et2O) (25% overall yield, based on dppm)
22
(Scheme 1). Spectroscopic data are given in the Ex-
perimental Section.
. Deh yd r ogen a tive Cou p lin g of HSn P h 3. Note
2
that all of the data refer to average values determined
for at least three catalytic runs. (i) F e-P d Allyl Siloxy
Ca ta lysts. Addition of a large excess (ca. 8000 mol
2
1
equiv) of HSnPh3 to a CH2Cl2 solution of 1c (293 K)
resulted in a vigorous gas evolution accompanied by an
immediate color change from yellow to orange-red.
After ca. 30 s, the gas evolution progressively decreased
and finally ceased after ca. 1 h. During this time, the
color of the mixture faded and a pale-yellow solution
was obtained. The solvent was removed under vacuum,
and Sn2Ph6 was precipitated by addition of Et2O.
Monitoring the volume of H2 released allowed for a plot
TON (turnover numbers) vs time to be made, which is
shown in Figure 1. After a short induction period (see
inset), the reaction rate reaches a maximum after ca.
pound was not isolated in pure form from these residues
nor was any mononuclear Pd complex evidenced.
The siloxy complexes 1b and 1d are also efficient
catalysts for the dehydrogenative coupling of HSnPh3.
Under the reaction conditions reported above for 1c, the
volume of H2 released was monitored in both cases
1
0 s, which corresponds to a turnover frequency (TOF)
5
-1
of ca. 2 × 10
h
. This maximum reaction rate is
(TON vs time, Figure 2). Selected catalysis data are
observed for only 5 s, after which a regular decrease
leads, after ca. 50-60 min to a plateau. The latter
corresponds to a TON of 1630; the same value was
determined from the amount of Sn2Ph6 formed. Analy-
sis of the combined residues from 5-6 experiments by
reported in Table 1. The presence of complex 5b in the
combined residues of the different catalytic experiments
3
1
1
using 1b has been evidenced by P{ H} NMR. We have
independently tested pure complexes 5a (see below) or
5
c and observed no catalytic activity.
(
ii) F e-P d Allyl Alk oxysilyl Ca ta lysts. The above
(
20) Formation of [{Me
observed.
21) The determination of these reaction conditions has been
discussed in ref 2.
2 2 3
SiOMe Si}Fe(CO) (dppm-P)] was never
results prompted us to reinvestigate the reaction be-
(
(22) Braunstein, P.; Charles, C. Unpublished results.

Heterobimetallic Alkoxysilyl and Siloxy Complexes
Organometallics, Vol. 17, No. 4, 1998 543
Ta ble 2. Ca ta lytic Activity of Heter obim eta llic
F e-P d Alk oxysilyl a n d Siloxy Com p lexes a n d
Mon on u clea r P d Com p lexes in th e
Deh yd r ogen a tive Cou p lin g of HSn P h
Fe-Pd complexes mononuclear Pd complexes
catalyst
TONa
[Pd(η -allyl)(µ-Cl)]2 (13)
3
catalyst
TONa
3
1
1
1
1
2
6
6
a
b
c
d
a
a
b
700
1380
1630
1800
700
800
860
300
370
430
460
3
b
Pd(η -allyl)Cl(PPh3) (14)
PdCl2(PhCN)2 (15)
c
Pd(nbd)Cl2 (16)
d
PdMeCl(P∼O)2 (17)
700
1380
Pd(OAc)2 (18)
a
Total catalytic activity expressed in turnover numbers, deter-
mined from the mass of Sn2Ph6 recovered (average value of a
minimum of five experiments); T ) 293 K, ratio [HSnPh3]/[cat.]
)
8000, solvent CH2Cl2, reaction time ) 3 h, except for 1a , 2a ,
b
and 6a where reaction time ) 96 h. Formed in situ by addition
F igu r e 3. Comparison between the catalytic activities of
3
c
of 2 equiv of PPh3 to a solution of [Pd(η -allyl](µ-Cl)]2. nbd )
norbornadiene. d P∼O ) Ph2PCH2C(O)Ph.
1
a (O) and 1c (b) in the formation of Sn
2 6 3
Ph from HSnPh
in dichloromethane.
Ta ble 3. Selected Ca ta lysis Da ta for th e
Deh yd r ocou p lin g of HSn P h 3 Ca ta lyzed by 1c in
Differ en t Solven ts
tween HSnPh3 and the alkoxysilyl complex 1a 1b in order
to evaluate the factors responsible for the large differ-
ence of reactivity between the latter and the siloxy
complexes under similar reaction conditions. When a
large excess of HSnPh3 (ca. 8000 mol equiv) was added
to a CH2Cl2 solution of 1a , a very mild gas evolution
was observed, accompanied by an immediate color
change from yellow to red. The latter is characteristic
TON maxa
plateau, hb
TOFc
solvent
Et2O, Tol, CH2Cl2d
THF
acetone
1600e
600
6800
1
0.5
3
ca. 2 × 10
5
3
6.5 × 10
3
8.9 × 10
a
Total catalytic activity expressed in turnover numbers (aver-
age value of a minimum of five experiments). b Time after which
1
b
c
of the formation of the trimetallic complex 2a (eq 1),
which can be isolated (yield >80%) if the reaction is
quenched after a few minutes by the addition of 2-3
drops of water. Contrary to what occurred in the case
of complexes 1b-d , the evolution of H2 did not stop nor
decrease after 1 h and was still observed after 48 h while
a progressive fading of the color of the solution was
noticed. Monitoring the formation of Sn2Ph6 (see Ex-
perimental Section) as a function of time gave the plot
shown in Figure 3 (white dots (O)). Although the curve
is similar to that obtained with 1c (black dots b), the
scale is by far different: the plateau is only reached
after 48 h, which corresponds to a TON of 700. The TOF
corresponding to the maximum activity of the catalyst,
the plateau is reached. Maximum turnover frequency expressed
in turnovers per h. d In these solvents the catalytic activity was
almost not modified. e Average value, ∆ within (2%.
(
iv) Mon on u clea r P a lla d iu m Com p lexes. We
have determined the catalytic activity of a series of
2
3
mononuclear Pd complexes under the same experi-
mental conditions as those reported above for the Fe-
Pd heterobimetallic complexes. The TON determined
from the quantities of Sn2Ph6 obtained are listed in
Table 2. These values range from 300 for [PdCl2-
3
(
PhCN)2] to 860 for [Pd(η -allyl)Cl(PPh3)]. In all cases
the evolution of H2 was over after 2 h, whereas the
lifetime of the heterobimetallic catalysts varied from 0.5
to 72 h depending on the nature of the silyl group.
-
1
observed between 1 and 4.5 h, is lower than 100 h
.
(
v) Solven t Effects. The dehydrogenative coupling
The combined pale yellow residues from 5-6 experi-
of HSnPh3 catalyzed by 1c has been carried out in
various solvents with monitoring of the volumes of H2
released. Selected catalysis data are reported in Table
3. Almost identical data have been determined when
the reaction was carried out in Et2O, toluene, or CH2-
Cl2. However, it was much more sensitive to acetone
and THF, as illustrated in Figure 4. The role of THF
resembles that of a reaction inhibitor, whereas the
efficiency and lifetime of the catalyst were enhanced in
acetone. It is remarkable in this case that although a
ments, obtained after the solvent had been removed and
Sn2Ph6 separated, were analyzed by 3 P{ H} and
1
1
1
H
NMR spectroscopic methods. Signals assignable to
2
2
31
complex 5a analogous to 5b and 5c were observed ( P-
1
H} NMR δ 44.7 (P(Fe), J PP ) 45 Hz, J P119Sn ) 160 Hz,
{
1
17
and J P Sn ) 145 Hz), -23.1 (J PP ) 45 Hz)), whereas
no trace of 1a or 2a was evidenced. When the catalytic
experiments were carried out with 2a instead of 1a , the
same catalytic activity was observed.
(
iii) F e-P d -Cl Ca ta lysts. We have also tested the
-
1
relatively low TOF (6800 h ) was observed, when
compared to that obtained in solvents like Et2O, toluene,
reactivity of [(OC)3Fe{µ-Si(OR)2(OR)}(µ-dppm)PdCl] 6a
5
-1
R ) Me)1a and 6b (R ) SiMe3)1i toward the dehydro-
and CH2Cl2 (ca. 2 × 10 h ), the reaction proceeded at
this rate for almost 1 h, whereas the maximum reaction
rate was maintained for only a few seconds in the other
cases. Higher catalytic activity and enhancement of the
lifetime of the catalyst in acetone have also been
observed for 1b (Table 4).
(
genative coupling of HSnPh3. Interestingly, these two
complexes exhibited the same catalytic activity as their
Pd(allyl) analogs 1a and 1b, respectively. Furthermore
the Fe-Pd-Sn complex 2a was isolated in high yield
(>80%) when the reaction with 6a was quenched with
water after a few minutes. This observation suggest a
similar reaction mechanism for the Fe-Pd-Cl and Fe-
Pd(allyl) catalysts.
(
23) Fe complexes such as [Fe
dppm-P)] (R ) OMe, OSiMe ) have also been tested under similar
conditions; they do not catalyze the dehydrogenative Sn-Sn coupling.
2 9 3 3
(CO) ] in THF and [HFe(SiR )(CO) -
(
3

5
44 Organometallics, Vol. 17, No. 4, 1998
Braunstein and Morise
F igu r e 5. Comparison between the catalytic activities of
F igu r e 4. Catalytic activity of 1c in the formation of Sn
Ph from HSnPh in Et O, CH Cl , or toluene (4), acetone
[) and THF (+).
2
-
n
n
1
a (.) and 1b (×) in the formation of Sn
2 6
Bu from HSn -
6
3
2
2
2
Bu in dichloromethane.
3
(
Ta ble 4. Tota l Ca ta lytic Activity a n d Rea ction
Tim e for th e Deh yd r ogen a tive Cou p lin g of HSn P h
Ca ta lyzed by 1b a n d 1c in Aceton e a n d CH Cl
presence of a radical scavenger (hydroquinone) and that
no solvent effect was noticed.
3
2
2
4
. Deh yd r ogen a tive Oligom er iza tion of Tin Di-
CH2Cl2
acetone
h yd r id es. We have also tested the reactivity of com-
plex 1b toward the dehydrogenative oligomerization of
tin dihydrides H2SnR2 (R ) Ph, Bu). Upon addition
TON maxa
RT, hb
TON maxa
RT, hb
catalyst
n
1
1
b
c
1300
1630
ca. 1.25
1
4380
6800
ca. 3.5
3
n
n
of 1b to an Et2O solution of H2Sn Bu2 (ratio H2Sn Bu2/
1
5
b ) ca. 1 × 10 ), a vigorous evolution of H2 was
a
Total catalytic activity expressed in turnover numbers (aver-
observed whose intensity rapidly decreased and ceased
after 2 h. Assuming that 1 mol equiv of H2 was released
per catalytic cycle, a TON of ca. 20 000 was determined.
In order to avoid the presence of a large excess of
age value of a minimum of five experiments). b RT ) Reaction time
corresponding to the end of gas evolution.
Ta ble 5. Deh yd r ogen a tive Cou p lin g of Tin
a
Hyd r id es Ca ta lyzed by 1a or /a n d 1b
n
unreacted H2Sn Bu2 and thus facilitate analysis of the
tin hydride
HSnPh3
catalyst
product (mn)b
TON maxc
polystannanes obtained, the reaction was repeated
n
using a H2Sn Bu2/1b ratio of ca. 25 000. When the
1a
Sn2Ph6 (60)
30
1
1a
1
1b
1b
b
Sn2Ph6 (60)
Sn2Bu6 (3)
Sn2Bu6 (3)
1300
reaction was over, the volatiles were removed under
vacuum, affording a brown waxy residue. Its UV-vis
spectrum (CH2Cl2 solvent) showed λmax values at 380,
350, and 314 nm, however, the two former bands were
of low intensity and a rapid photobleaching was ob-
served, resulting in their disappearance. In the 1 Sn
NMR spectrum (benzene-d6), several signals were ob-
served between δ -164 and -203. On the basis of
HSnBu3
ca. 114 000
ca. 202 000
1100
b
d
H2SnPh2
[SnPh2]n (60)
n
n
H2Sn Bu2
[Sn Bu2]n and
20 000
n
d
H[Sn Bu2]nH (120)
19
a
In CH2Cl2; T ) 20 °C. b Reaction time in minutes. c Turnover
numbers: average values of a minimum of three experiments/run.
Structure of oligomers not determined.
d
1
19
18b,24
recently reported
Sn NMR data,
the main reso-
3
. Deh yd r ogen a t ive Cou p lin g of H Sn Bu 3. We
nance at δ -190 (broad signal) was assigned to linear
have also evaluated the reactivity of 1a and 1b toward
the dehydrogenative coupling of HSnBu3. Vigorous
evolution of H2 was observed upon addition of the
n
H(Sn Bu2)nH chains whereas weaker resonances at δ
-
201 and -203 correspond to the cyclic species cyclo-
n
n
(
Sn Bu2)5 and cyclo-(Sn Bu2)6, respectively. Other sig-
catalysts to Et2O (or CH2Cl2) solutions of the tin hydride
6
(
ratio HSnBu3/cat. ca. 1 × 10 ). The volume of H2
(24) J ousseaume, B.; Noiret, N.; Pereyre, M.; Saux, A.; Franc e` s, J .-
M. Organometallics 1994, 13, 1034.
25) See, for example: (a) Canty, A. J . Acc. Chem. Res. 1992, 25,
3. (b) Canty, A. J . Platinum Met. Rev. 1993, 37, 2. (c) Canty, A. J .
released was monitored and converted into turnover
numbers. The plot of turnover numbers vs time is given
in Figure 5. In both cases a plateau is reached after
(
8
Comprehensive Organometallic Chemistry II; Abel, E. W., Stone, F.
G. A., Wilkinson, G., Eds.; Elsevier: New York, 1994; Chapter 5 and
references cited therein.
2
-3 min, whereas much longer reaction times were
observed in the case of HSnPh3. Note that further
addition of catalyst after this period produced a vigorous
gas evolution, thus showing that consumption of HSn-
Bu3 was not responsible for the loss of catalytic activity
of 1a ,b and the short catalyst lifetime. Interestingly,
very high TONs have been observed: 114 000 and
(26) (a) Woo, H. G.; Walzer, J . F.; Tilley, T. D. J . Am. Chem. Soc.
1
992, 114, 7047. (b) Tilley, T. D. Acc. Chem. Res. 1993, 26, 22. (c) Imori,
T.; Tilley, T. D. Polyhedron 1994, 10, 2231.
27) (a) J essop, P. G.; Morris, R. H. Coord. Chem. Rev. 1992, 121,
(
155. (b) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T.
H. Acc. Chem. Res. 1995, 28, 154 and references cited therein. (c)
Crabtree, R. H. Acc. Chem. Res. 1995, 95, 987. (d) Folga, E.; Woo, T.;
Ziegler, T. In Theoretical Aspects of Homogeneous Catalysis; van
Leeuwen, P. W. N. M., Morokuma, K., van Lenthe, J . H., Eds.;
Kluwer: Dordrecht, 1995; p 115 and references cited therein.
2
02 000 for 1a and 1b, respectively. These values are
more than 150 times those determined for the same
catalysts in the presence of HSnPh3. Furthermore, the
(28) Milet, A.; Dedieu, A.; Kapteijn, G.; van Koten, G. Inorg. Chem.
1997, 36, 3223.
7
-1
(29) (a) Roskamp, E. J .; Pedersen, S. F. J . Am. Chem. Soc. 1987,
09, 3152. (b) Carofiglio, T.; Floriani, C.; Roth, A.; Sgamellotti, A; Rosi,
M.; Chiesi-Villa, A.; Rizzoli, C. J . Organomet. Chem. 1995, 488, 141.
(30) We thank a reviewer for raising this question.
TOFs are also amazingly high: ca. 2.3 × 10 h for 1a
1
7
-1
and ca. 3 × 10 h for 1b. It should be mentioned that
these reactions were carried out in the dark and in the

Heterobimetallic Alkoxysilyl and Siloxy Complexes
Organometallics, Vol. 17, No. 4, 1998 545
nals (δ -164, -174, -179, and -184) were not assigned.
also noneffective, which is in agreement with a purely
homogeneous catalytic process.
1
The H NMR spectrum also indicateded the presence
n
n
of a mixture of linear H(Sn Bu2)nH and cyclic (Sn Bu2)n
polystannanes (see Experimental Section). Analysis of
a THF solution by GPC with a refractive index detector
indicated a large mass distribution with Mw/Mn )
After the initial formation of the Fe-Pd-Sn species,
we propose that the tin-tin dehydrogenative coupling
reactions occur either by successive oxidative additions
and reductive eliminations or by a σ-bond metathesis-
type mechanism (Scheme 2). In both cases the hemi-
5
5700/12800 for the linear oligomers.
2
2
When R ) Ph, a mild gas evolution was observed after
lability of the bridging µ -η -SiO ligand is involved while
addition of 1b (H2SnPh2/1b ratio ) ca. 10 000). It
ceased after 1 h. A pale yellow-brown precipitate was
formed, which was separated and washed with pentane
affording a pale yellow solid. A TON of 1100 was
determined on both the volume of H2 released and the
amount of polystannanes recovered, assuming that it
was constituted of [SnPh2]n species and that one Sn-
Sn bond was formed per catalytic cycle. Mass spectros-
copy only revealed the presence of SnnPhm fragments
with 1 e n e 6 and 1 e m e 7. A λmax value of 350 nm
was observed in the UV-vis spectrum, which corre-
sponds to oligomers with Mw/Mn ) 1140/1100 (GPC
measurements), probably cyclo-[SnPh2]n oligomers with
the heterobimetallic structure is preserved. In the case
of migratory insertions of small molecules (CO, CNR,
olefins) into the M-C bond (M ) Pd, Pt) of heterobi-
1
j
metallic alkoxysilyl Fe-M complexes, a mechanism
also based on successive Fe-Si-O-M ring openings
and closures resulting from the hemilability of the
3
bridging -Si(OMe) group has been described.
Intermediates 2 and 8 are stabilized by an intramo-
lecular O f Pd interaction. Opening of this Fe-Si-
O-Pd four-membered ring releases a vacant coordina-
tion site at the Pd center, and this could be followed
either by oxidative addition of HSnPh3 and formation
IV
of Pd intermediates 7 and 9 or by formation of
1
8,31
n ) 4 and 5.
No discernible peaks assignable to
transient species 10 and 11 if the catalysis proceeds via
a concerted σ-bond metathesis mechanism (Scheme 2).
The former mechanism is supported by the fact that
formation, via oxidative additions at a Pd(II) centers,
of organopalladium(IV) intermediates including hydrido
Sn-H groups were detected in the 1H NMR and IR
spectra.
Discu ssion
2
5
complexes is now well-established. On the other hand,
a σ-bond metathesis mechanism has recently been
invoked for the dehydropolymerization of secondary
stannanes catalyzed by zirconocene and hafnocene
Mech a n istic Con sid er a tion s for th e Deh yd r oge-
n a tive Cou p lin g of HSn P h 3 Ca ta lyzed by Heter o-
bim eta llic F e-P d Silyl Com p lexes. The structure
of complexes 1 (18 electron iron center and allyl ligand
coordinated to the 16 electron palladium atom) along
with the fact that mononuclear palladium complexes
18b
derivatives and described for hydrosilane polymeriza-
26
tions catalyzed by the same complexes. Although this
mechanism has generally been applied to early transi-
1
4,15
catalyze the dehydrogenative coupling of stannanes
see Table 2) strongly suggest that the elementary
27
tion metals, lanthanides, and actinides, recent studies
(
28
indicate its possible extension to palladium chemistry.
transformations during the catalysis take place at the
palladium center, the iron fragment participating in its
coordination environment through the metal-metal
bonding, the bridging phosphine, and the silyl ligand.
We have found that 1a and the trinuclear Fe-Pd-
Sn complex 2a exhibited the same catalytic activity in
the dehydrogenative coupling of HSnPh3. We have
previously reported that reaction of 1a with HSnPh3 led
to the immediate formation of 2a , via propene
elimination.1 We, therefore, assume that this reaction
represents the initial step of the catalytic process
studied here and may be responsible for the induction
period observed in the plots shown in Figures 1 and 2.
Although no siloxy analog of 2a could be isolated,
probably owing to the high reactivity of these species
under the reaction conditions, nor prepared by an
alternative route, the rapid transformation of complexes
The combined observations that complexes 6a and 6b
exhibit the same catalytic activity as the corresponding
allyl derivatives 1a and 1b and that the trinuclear
complex 2a could be isolated when the reaction with 6a
was quenched after a few minutes strongly indicate that
the reactions follow the same pathway for all complexes.
The formation of 2a could either result from HCl
elimination (route a, Scheme 3) or from the reaction of
the Pd-Cl bond with the tin hydride to afford the
transient species 8a , which would then rapidly react
with a second equivalent of HSnPh3 to yield 2 and H2
b
(route b, Scheme 3). Although the former cannot be
ruled out, we favor the latter, which also corresponds
to part of the proposed catalytic cycle, on the basis of
the following observations: (i) the pH of the reaction
mixture and that of the gas evolving were not acidic,
even when the reactions were carried out on a large
scale (>100 mg of catalyst); (ii) tin hydrides, which are
known to reduce organic halides, have already been
used to prepare metal hydrides from M-Cl complexes,
this reaction being accompanied by elimination of R3-
1
b-d into Fe-Pd-Sn intermediates in the presence of
HSnPh3 is supported by the immediate color change of
the reaction mixture upon addition of the stannane from
yellow to red, the latter being the characteristic color
of 2a . It should also be mentioned that no modification
of the kinetic parameters or TON values for the different
catalysts has been noticed when the reactions were
carried out in the presence of radical scavengers or/and
in the dark, which seems to rule out a radical-chain
mechanism. Furthermore, addition of Hg metal was
1
5b,29
SnCl.
Since we did not observe any reactivity of
the Si-H bond of the siloxy ligand in 1d nor any Si-
H-Pd interaction, we assume that the above discussion
3
0
can also be applied in this case.
Isolation of 2a , even in the presence of a large excess
of HSnPh3, suggests that its subsequent reaction rep-
resents the rate-determining step in the catalytic cycle.
Identification, after completion of the catalysis, of
(31) Devylder, N.; Hill, M.; Molloy, K. C.; Price, G. J . J . Chem. Soc.,
Chem. Commun. 1996, 711.

5
46 Organometallics, Vol. 17, No. 4, 1998
Braunstein and Morise
Sch em e 2
complexes 5, which are inactive, suggests that one of
the factors responsible for the loss of catalytic activity
of the bimetallic complexes is the occurrence of Fe-Pd
bond cleavage. We could not evidence the fate of the
palladium center, which might display residual catalytic
activity (see below), and found no indication for the
formation of colloidal or metallic Pd.
the lower reactivity of the latter since the coordination
site on the palladium, which is required for the catalysis
to proceed, is more efficiently masked in 2a than in the
case of the siloxy complexes. This higher degree of
stabilization would also account for the enhancement
of the lifetime of the alkoxysilyl catalysts and for the
possible isolation of 2a in contrast to its siloxy analogs
2b-d .
Effect of th e Silyl Liga n d in th e Ca ta lytic Activ-
ity of Com p lexes 1. The most striking feature of the
dehydrogenative coupling of HSnPh3 catalyzed by com-
plexes 1 is the considerable difference of reactivity
between the alkoxysilyl complex 1a and the siloxy
complexes 1b-d , as shown in Figure 3. The latter
exhibit a catalytic activity that is about twice that of
the former, whereas the duration of the reaction does
not exceed 1 h, vs more than 48 h for 1a . As reported
above, the proposed mechanism involves the formation
Differences in reactivity are also observed between
the siloxy complexes 1b-d , as shown in Table 1 and
Figure 2. The trend is that the reactivity of a complex,
expressed by the TON and TOF values, is enhanced
when the number of OSiMe3 groups decreases. The
duration of the reaction, which can be related to the
lifetime of the catalyst, follows an opposite trend. As
an example, the TON and TOF are 30 and ca. 90%
higher, respectively, for 1d , which has only one OSiMe3
group, than in the case of 1b, which possesses three
OSiMe3 groups; however, the reaction is over after ca.
0.5 h with the former whereas it lasts over 1 h with the
2
of transient Fe-Pd species stabilized by µ2-η -SiO
bridges. Due to the reduced O-donor capacity of the
siloxy ligand compared to that of an alkoxysilyl ligand,
owing to the more electropositive character of the Si
atom in the former, the O f Pd interaction should be
less stabilizing in 2b-d than in 2a . This may explain
1
latter. Recent H NMR studies have evidenced a
dynamic behavior of the -SiO f Pd interactions in
related systems resulting in a rapid exchange between

Heterobimetallic Alkoxysilyl and Siloxy Complexes
Organometallics, Vol. 17, No. 4, 1998 547
Sch em e 3
the different OR groups of the silyl ligand which rotates
about the Fe-Si bond.1a,b Therefore, from a statistical
point of view, intermediates 2 and 9 should be more
stabilized by the presence of a tris(siloxy) ligand than
by that of mono- or bis(siloxy) one. This difference in
stabilization would then account for the observed dif-
ference of the reactivity between 1b, 1c, and 1d .
Note that when the Fe-Pt complex [(OC)3{(Me3-
Sch em e 4
3
1i
SiO)3Si}Fe(µ-dppm)Pt(η -allyl)] (12) was treated with
HSnPh3, no gas evolution nor formation of Sn2Ph6 was
observed, even after several hours or upon heating (CH2-
Cl2 reflux). However, this reaction yielded the hydrido
complex [HFe{Si(OSiMe3)3}(CO)3(dppm-P)], which has
been characterized by 31P and 1H NMR spectroscopy
(
presence of a doublet due to the hydride at δ ) -9.1
1
a,c
ppm, J PH ) 27 Hz),
plexes, which were not evidenced.
and presumably Pt-Sn com-
Com p a r ison b et w een H et er ob im et a llic a n d
Mon on u clea r Com p lexes. Previous work by Be-
catalysts [Pd(π-allyl)(µ-Cl)]2 (13) and [Pd(π-allyl)Cl-
(PPh3)] (14) are slightly more active than the alkoxysilyl
heterobimetallic derivatives 1a , 2a , and 6a . The initial
step of the reaction between tin hydrides and complexes
13 and 14 should follow either of the four pathways
presented in Scheme 4. Elimination of HCl (pathway
a) does not appear likely, since the pH of the reaction
mixture was not acidic. From the remaining pathways,
we favor b, which involves formation of a stable Sn-Cl
bond and maintains the allyl ligand in the coordination
sphere of the palladium center. Propene elimination in
the cases of complexes 1 resulted from both the allyl
ligand being present and the stability of the trinuclear
products. One can speculate that the allyl ligand,
through an η h η equilibrium, may generate a vacant
site on the palladium and thus play a role similar to
that of the silyl ligand in 1. The environment of the
palladium(II) center would, therefore, mimic that of the
Fe-Pd heterobimetallic complexes, a feature that may
account for the higher reactivity of 13 and 14 than the
other mononuclear species. The lower TONs for com-
plexes 1a , 2a , and 6a can be attributed to the fact that
the Si(OMe)3 ligand is more stabilizing than the allyl
ligand, which in turn may be responsible for their longer
lifetime.
1
4
letskaya et al. has shown that tin hydrides HSnR3
were dehydrogenated and converted to the correspond-
ing distannanes Sn2R6 in the presence of 1 mol % of
mononuclear Pd(II) complexes. Although good conver-
sions (ca. 80%) were reported, the catalytic activity
appears relatively low (TON < 45) in comparison with
those observed in the case of the heterobimetallic
complexes 1. However, the authors did not attempt
to optimize these reactions, and the experimental con-
ditions are somewhat different from those we have
used in the present study. We were, therefore, inter-
ested in determining the catalytic activity of a series
of mononuclear Pd(II) complexes toward the dehydro-
genative coupling of HSnPh3 in the reaction condi-
tions reported above for complexes 1. The TONs
obtained are given in Table 2. In general, mononuclear
Pd(II) complexes appear to be less efficient catalysts
than their heterobimetallic counterparts. This observa-
tion emphasizes the role of the heterobimetallic struc-
ture and more precisely that of the iron fragment, which
seems to provide an appropriate environment for the
Pd center through metal-metal bonding and the bridg-
ing ligands.
1
3
Note, however, that the most active mononuclear

5
48 Organometallics, Vol. 17, No. 4, 1998
Braunstein and Morise
Solven t Effects. The catalytic activity of complexes
in the dehydrogenative coupling of HSnPh3 was
H2SnR2 (R ) Bu, Ph). Surprisingly, in these cases the
catalytic activity of 1b was less than that observed in
the presence the corresponding HSnR3. When R ) Ph,
cyclic oligomers [SnPh2]n (n ) 5 and 6) were formed,
whereas higher molecular weight polystannanes have
not been detected. However when R ) Bu, both cyclic
1
comparable in Et2O, toluene, and CH2Cl2. However, in
THF, the reaction was less efficient: the TON and the
TOF values were 2.5 and 30 times smaller, respectively,
than in other solvents, whereas the reaction was already
over after 0.5 h (Table 3). It appears reasonable to
associate this reduction of catalytic activity with the
coordinating properties of THF, which may either
coordinate at the palladium center and/or form adducts
with the tin hydride, thus rendering the activation of
the Sn-H bond less efficient. Surprisingly high TONs
were obtained in acetone and were associated with a
significant increase of the catalyst lifetime (Table 3).
Although the TOF was smaller in acetone than in CH2-
Cl2 or related solvents, the reaction proceeded at a
maximum rate for more than 40 min, in contrast in the
other cases where it was only maintained for a few
seconds. Although we have observed that a small
fraction of acetone was reduced by HSnPh3 with forma-
tion of Sn2Ph6 and isopropyl alcohol, this side reaction
remains a minor process that does not release H2 and
thus did not affect our measurements. Furthermore,
we have studied the behavior and stability of 1 and 2a
in acetone and did not find any significant interactions
between these complexes and the solvent.
and linear chains polystannanes were obtained.
A
relationship between λmax and chain length for (R2Sn)n
3
1
polymers has recently been presented. Accordingly,
the λmax value we observed at 380 nm is consistent with
an unspecified degree of polymerization of greater than
80 Sn atoms. The presence of linear polystannanes is
1
19
also confirmed by the
Sn NMR resonance (C6D6) at
-190 ppm. The other observed λmax values of 350 and
314 nm should correspond to cyclic or linear polystan-
nanes with degrees of oligomerization lower than 15.
Con clu d in g Rem a r k s
In studying the dehydrogenative coupling of tin
hydrides catalyzed by heterobimetallic Fe-Pd silyl
complexes, we have witnessed that the silyl group
played a major role in the catalytic process and that
substituent modifications resulted in important varia-
tions of catalytic activity and reaction kinetics. The
influence of the metal-metal bonding has been evi-
denced by comparison of the catalytic activity of mono-
nuclear Pd species with that of the title complexes.
Owing to different observations made throughout this
study, we propose a reaction mechanism in which the
bimetallic structure of the catalyst is retained. This
work provides an insight into homogenous bimetallic
Deh yd r ogen a tive Cou p lin g of Oth er Tin Hy-
d r id es. The efficiency of catalysts 1a ,b toward the
dehydrogenative coupling of HSnBu3 was largely en-
hanced compared to that observed in the case of
HSnPh3. TON and TOF values were more than 100
times those obtained in the latter case. The difference
of reactivity between the alkoxysilyl derivative 1a and
its siloxy analog 1b, already observed in the presence
of HSnPh3, still occurred with HSnBu3, which indicates
the involvement of the silyl unit in the catalytic process
as discussed above. The very high reaction rate and
the fact that tributyltin units are often involved in
radical-mediated reactions may suggest that the cataly-
sis proceeds via a radical chain mechanism. However,
no modification of the catalytic activity of 1a ,b was
noticed when carrying out the experiments in the dark
and in the presence of a radical scavenger. Further-
more, we have been able to detect by 31P{ H} NMR the
formation of transient Fe-Pd-Sn trimetallic complex
1
9
catalysis and illustrates the “ligand effect” that a
metal center (i.e., Fe) may have on the adjacent metal
(i.e., Pd).
Exp er im en ta l Section
All of the reactions and manipulations were carried out
under an inert atmosphere of purified nitrogen using standard
Schlenk-tube techniques. Solvents were dried and distilled
under nitrogen before use: hexane and toluene over sodium,
tetrahydrofuran and diethyl ether over sodium benzophenone,
acetone over calcium chloride, dichloromethane over calcium
hydride or phosphorus pentoxide. Nitrogen (Air liquide,
R-grade) was passed through BASF R3-11 catalyst and mo-
lecular sieves columns to remove residual oxygen and water.
Elemental C, H, and N analyses were performed by the Service
de microanalyses (Universit e´ Louis Pasteur, Strasbourg).
Infrared spectra were recorded on a IFS 66 Bruker FT-IR
spectrometer. UV-vis spectra were recorded on a Kontron
1
[
(CO)3Fe{Si(OMe)2(OMe)}(µ-dppm)Pd(SnBu3)] (15) upon
treatment of 1a with excess HSnBu3.1b We assume,
therefore, that the reaction proceeds according to a
similar catalytic cycle as that presented above for
HSnPh3 and that the enhancement of the TON and TOF
values associated with shorter reaction times, compared
to experiments carried out with HSnPh3, are related to
the higher reactivity of intermediates 15 than 2 toward
the tin hydride substrate present. The high efficiency
of catalysts 1 in the dehydrogenative coupling of HSn-
Bu3 renders this reaction a competitive approach toward
the synthesis of hexa-n-butylditin compared to other
procedures that usually require harsher conditions or
long reaction times or are expensive.
1
31
1
13
1
Uvikon 860 spectrometer. The H, P{ H}, and C{ H} NMR
spectra were recorded at 300.1, 121.5, and 75.5 MHz, respec-
tively, on a Bruker AM300 instrument. Phosphorus chemical
3 4 2
shifts were externally referenced to 85% H PO in H O with
downfield chemical shifts reported as positive. Mass spectra
were measured on a Fisons ZAB-HF spectrometer (Universite´
Louis Pasteur, R. Hueber). The GPC measurements were
recorded on a Chromatograph (vs polystyrene standards)
equipped with a refractive index detector and 50, 100, 500,
3
and 10 Å PLgel columns in series (tetrahydrofuran solvent).
The synthesis of the bimetallic complexes was generally
5
monitored by IR in the ν(CO) region. Fe(CO) was obtained
New advances in the synthesis of high molecular
3
32a
3
1
from Aldrich and used as is. Complexes [Pd(η -allyl)(µ-Cl)]
2
,
polystannanes by either Wurtz coupling or dehydro-
polymerization of secondary stannanes18 have been
recently reported. We found, therefore, it of interest to
test complex 1b toward the dehydropolymerization of
(32) (a) Sakakibara, M.; Takahashi, Y.; Sakai, S.; Ishii, Y. J . Chem.
Soc., Chem. Commun. 1969, 396. (b) Hartley, F. R. Organomet. Chem.
Rev., Sect. A 1970, 6, 119.

Heterobimetallic Alkoxysilyl and Siloxy Complexes
,32b [Pd(nbd)Cl ],33 [PdMeCl(Ph
Organometallics, Vol. 17, No. 4, 1998 549
PdCl
2
(PhCN)
2
2
2
PCH
2
C(O)Ph)
2
],34
charged with 1.4 g of HSnPh
3
(4 × 10 mol) in 15 mL of
2 2
-3
1
b
1g
1a
1e
1
a , 1b,c, 6a , and 6b were all prepared using published
procedures. HSnPh , H SnPh , and H SnBu were obtained
by reduction of the corresponding tin chloride derivatives with
solvent and was placed in a water bath at 293 K. CH Cl was
3
2
2
2
2
used as the solvent, except for in the solvent effect studies.
H Mon itor in g. When the volume of H released (catalyst
2
2
1
8b,37
LiAlH
4
in diethyl ether.
HSnBu
3
was a commercial
1a -d and solvent effect studies) was monitored, the procedure
sample (Lancaster) and was used as received. The experi-
mental error on the TON and TOF values has been estimated
to be 3-4%, values given have been rounded off.
was as follows: The Schlenk flask was fitted onto a gas
-
4
burette, and 1 mL of a 5 × 10 M solution of catalyst (5 ×
10 mol) was rapidly added to the reaction mixture via syringe
through the serum cap. The volumes of H released were
-
7
2
Syn th esis of [(OC)
3
{(Me
η -a llyl)] (1d ). A magnetically stirred solution of [Fe(CO)
0.4 mL, 3 mmol) and (Me HSi) O (1.5 mL, 8.5 mmol) in hexane
100 mL) was irradiated with isopropyl alcohol cooling at 283
2 2
HSiO)Me Si}F e(µ-d p p m )P d -
directly read on the graduated burette, and the contents of
the syringe was taken into account, therefore VH2 ) Vread - 1
cm . The turnover numbers were calculated by the following
3
(
(
(
5
]
2
2
3
2 2
equation: TON ) nH /ncat., with nH being determined by
K for ca. 5 h. Irradiation was stopped when the ν(CO)
applying the gas equation PV ) nRT. Each experiment was
repeated at least 3 times, and the mean values were used for
the different plots. When the reaction was over, the volatiles
-
1
absorptions of [Fe(CO)
5
] (2001 and 2023 cm ) had almost
completely disappeared. The resulting complex cis-[HFe-
{
2
SiMe
2 2 4
(OSiHMe )}(CO) ] (ν(CO) (hexane) 2095 m, 2029 sh,
2
were removed in vacuo and the residue washed with Et O,
-1
021 vs, and 2009 vs cm ) was not isolated, and the resulting
3
8
1
affording Sn
NMR spectrum only showed signals in the aromatic region).
The TON determined from the quantity of Sn Ph recovered
was in almost all cases in good accordance with that deter-
mined from the volumes of H . The ether fraction was dried
2 6
Ph as a white solid (mp ) 225-235 °C, the H
pale-yellow mixture was added immediately to a toluene
solution (50 mL) of dppm (1.05 g, 2.73 mmol) in two portions.
After each addition, the CO evolved was removed under
reduced pressure for 1 min. The solution was stirred for 30
min at room temperature, and the volatiles were removed
2
6
2
under reduced pressure. Extraction with pentane or hexane
afforded a pale yellow solution. The extracts of different
experiments with a same catalyst were combined, and the
solvent was removed in vacuo. NMR spectroscopic analysis
of the residues of the reactions carried out with the heterobi-
metallic Fe-Pd catalysts 1a -c showed the presence of com-
plexes 5. Authentic samples of complexes 5a and 5c have been
under reduced pressure, affording mer-[HFe(CO)
3 2
{SiMe -
(
OSiHMe )}(dppm-P)] (3d ) as a brown solid, which was
2
3
1
1
contaminated by ca. 20% of [Fe(CO)
NMR δ 15 ppm).
3
(dppm-P,P)] ( P{ H}
The hydrido complex 3d was dissolved in THF and added
to a suspension of KH in excess in THF. An immediate gas
evolution was noticed (H
formation of transient K[Fe(CO)
was monitored by IR spectrometry in the ν(CO) region (1928
2
), whereas the color darkened. The
prepapred independently by reaction of the anions 4a and 4c
3
{SiMe (OSiHMe )}(dppm-P)]
2
2
22
with ClSnPh
P h
98 K) δ 3.76 (s, 9H, SiOMe), 4.25 (br, 2H, PCH
3
.
[
3
Sn Fe(CO)
3
{Si(OMe)
3
}(dppm -P )] (5a). 1H NMR (C
P), 6.5-7.7
m, 35H, aromatics); 31P{ H} NMR: see text. Anal. Calcd for
SiSn: C, 59.12; H, 4.66. Found: C, 58.95; H,
6 6
D ,
-
1
w, 1845 vs, 1824 vs cm ). When the reaction was complete
2
(
2
(
after ca. 30 min), the solution was filtered over a 1 cm Celite
1
3
pad and slowly added to a THF solution of [Pd(η -allyl)(µ-Cl)]
0.5 mol equiv) at 233 K. The reaction mixture was allowed
2
49 6 2
C H46FeO P
.77.
(
4
to warm to room temperature and stirred for 30 min before it
was filtered. The volatiles were removed under reduced
pressure, and the residue was chromatographed on alumina.
[
6
P h
3
Sn F e(CO)
, 298 K) δ 0.33 (s, 27H, OSiMe
3
{Si(OSiMe
3
)
3
}(d p p m -P )] (5b). 1H NMR
2
(C
D
6
3
), 3.28 (d, 2H, J (P-H) )
3
1
8
(
1
5
Hz, PCH
2
P), 6.7-7.7 (m, 35H, aromatics); P NMR δ 44.2
The front band (R ) 0.85) of yellow color containing 1d was
F
2
+3
2
119
P(Fe), d with Sn satellites,
J (P-P) ) 52 Hz, J (P- Sn)
collected and dried in vacuo. Complex 1d was obtained as a
yellow-orange air stable powder (yield ) 38%; overall yield )
2
117
2+3
55 Hz, J (P- Sn) ) 130 Hz), -23.1 ppm (d,
J (P-P) )
2 Hz).
[P h
-1
2
5% based on dppm). IR (THF) 1948 m, 1880 s, 1861 vs cm ;
}(dppm -P )] (5c). 1H NMR
), 4.27
P), 6.5-7.9 (m, 35H, aromatics);
P NMR see text. Anal. Calcd for C53 Si Sn: C,
58.09; H, 5.34. Found: C, 58.64; H, 5.52.
Sn P h Mon itor in g. When the formation of Sn
1
3
3 3 3 2
Sn Fe(CO) {SiMe(OSiMe )
H NMR (C
H)Me ), 0.48 (s, 6H, Fe-SiMe
Hz, allyl), 3.48 (br, 1H, allyl), 3.72 (t, 2H, J (P-H) ) 11.5 Hz,
PCH P), 5.16 (br, overlapping signals, 2H, allyl, SiH), 5.36 (t,
6 6
D , 298 K) δ 0.22 (d, J (H-H) ) 10 Hz, 6H, OSi-
3
(C
6
D
6
, 298 K) δ 0.22 (s, 3H, SiMe), 0.33 (s, 18H, OSiMe
3
(
2
2
), 2.61 (d, 2H, J (H-H) ) 8.5
2
2
(d, 2H, J (P-H) ) 11 Hz, PCH
2
3
1
H
58FeO
5
P
2
3
2
3
3
1
H, J (H-H) ≈ J (P-H) ≈ 7 Hz, allyl), 6.65-7.72 (m, 20H,
31 1 2+3
6 6
2
6
2 6
Ph was
aromatics); P{ H} NMR (C
1
for C35
H, 4.70.
Spectroscopic data of 3d : IR(toluene) (also see text) 2050
D
, 298 K) 26.4 (d,
J (P-P) )
09 Hz, P), 66.1 (d, 2 J (P-P) ) 109 Hz, P(Fe)). Anal. Calcd
+3
monitored (catalysts 1a , 2a , and 6a ), the procedure was as
follows: a battery of six Schlenk flasks was prepared as
described above, except that they were not connected to gas
burettes. The catalyst was then added, and the reaction
mixtures were successively quenched at t ) 1, 4.5, 8.5, 25, 48,
and 72 h by addition of water. The volatiles were removed
under reduced pressure, and the residue was washed with
4 2 2
H40FeO P PdSi : C, 52.22; H, 5.01. Found: C, 52.45;
-
1 1
w, 1980 s, 1965 vs cm ; H NMR (C
2
6
D
6
, 298 K) δ -9.1 (d, J )
6 Hz, 1H, Fe-H), 0.21 (d, J ) 12 Hz, 6H, OSi(H)Me ), 0.89
), 3.18 (dd, J (P-H) ) 8 and 1 Hz, 2H,
2
2
(
s, 6H, Fe-SiMe
2
3
1
Et
weighed. The turnover numbers were determined by the
following equation: TON ) nSn Ph /ncat.
2 2 6
O, affording pure Sn Ph as a white solid, which was
PCH
{
2
P), 5.21 (br, 1H, SiH), 6.7-7.7 (m, 20H, aromatics); P-
H} NMR (C D , 298 K) -25.34 (d, J ) 105 Hz, P), 52.18 (d,
1
6 6
2
6
J ) 105 Hz, P(Fe)).
In the case of mononuclear palladium catalysts 13-18, a
battery of three Schlenk flasks was set up as described above.
After addition of the catalyst, the reaction mixtures were
3
Deh yd r ogen a tive Cou p lin g of HSn P h Ca ta lyzed by
Silylated Heter obim etallic Fe-P d Com plexes an d Mon o-
n u clea r P d Com p lexes. Gen er a l P r oced u r e. A Schlenk
flask equipped with a stirring bar and a serum cap was
stirred until the gas evolution had ceased. Sn
separated using the same work up as mentioned above.
Deh yd r ogen a tive Cou p lin g of HSn Bu . The approxi-
2 6
Ph was
3
(
33) Drew, D.; Doyle, J . R. Inorg. Synth. 1972, 13, 52.
(34) This complex was prepared according to the procedure described
mate catalytic activity of 1a ,b was first estimated as follows:
3
5
36
-4
for [PdCl
2
{Ph
2
PCH
2
C(O)Ph}
2
]
from [PdClMe(cod)] and Ph
2
PCH
2
C-
1 mL of a 5 × 10 M CH
Cl
solution of 1a ,b was added to a
Cl The flask was
2
2
(
O)Ph.35
Schlenk flask containing 15 mL of CH
2
2
.
(35) Bouaoud, S. E.; Braunstein, P.; Grandjean, D.; Matt, D.; Nobel,
D. Inorg. Chem. 1986, 25, 3765.
(
(
36) Lapido, F. T.; Anderson, G. K. Organometallics 1994, 13, 303.
37) Kuivila, H. G.; Sawyer, A. K.; Armour, A. G. J . Org. Chem. 1961,
(38) Literature values: (a) 226-228 °C, Gilman, H.; Rosenberg, S.
D. J . Org. Chem. 1953, 18, 1554. (b) 230-234 °C, Tamborski, C.; Ford,
F. E.; Soloski, E. J . J . Org. Chem. 1963, 28, 237.
2
6, 1426.

5
50 Organometallics, Vol. 17, No. 4, 1998
Braunstein and Morise
placed in a 293 K water bath, and ca. 2 mL of pure HSnBu
was added. An immediate vigorous gas evolution (H ) was
observed, whereas the color of the solution turned red. After
the evolution of H had ceased (ca. 1 min), another portion of
HSnBu was added. This procedure was repeated until the
addition of the tin hydride yielded no more H . The volatiles
were then removed, and the residue was analyzed by H NMR
spectroscopy. The ratio between Sn Bu and residual HSnBu
was determined by integration of the signals in the CH and
CH regions, and it was then used to determine the TON,
which was found to be >100 000 for 1a and >200 000 for 1b.
3
1H NMR (300 MHz, benzene-d
6
) δ 1.03 (br, CH
, linear), 1.39-1.56 (m, br, CH , cyclic), 1.58-1.66 (m,
, linear), 1.74-1.98 (m, br, CH , cyclic and linear).
3
, cyclic), 1.15
2
(br, CH
br, CH
3
2
2
2
1
19
2
Assignment made according to ref 18b. UV-vis and
Sn-
1
3
{ H} NMR: see text.
2
Deh yd r ogen a tive P olym er iza tion of H
polymerization of H SnPh followed procedures similar to that
described above for H
2 2
Sn P h .
The
1
2
2
n
2
6
3
2
Sn Bu
2
, except that 1 g of H
2
SnPh
2
was
3
used per run. A mild gas evolution was noticed, which ceased
after 1 h. A pale yellow precipitate was formed. It was
2
separated and washed with pentane. The average V 2
mL, and the average mass of [Ph Sn] ) 100 mg. Assuming
that 1 mol equiv of H was released and that one Sn-Sn bond
H
) 9.5
1
Bu
.33 (sext, CH2γ), 1.50 (quint, CH2â).
Mon itor in g. The procedure described for the dehydro-
genative coupling of HSnPh was used. HSnBu (8 g) was
placed in a Schlenk flask with 15 mL of CH Cl
6
Sn
2
:
H NMR (CDCl
3
, 298 K) δ 0.91 (t, CH
3
), 0.98 (t, CH2R),
2
n
1
2
H
2
was formed per catalytic cycle, a TON of ca. 1100 was
determined. IR (KBR) 3061 m, 3045 m, 1641 m, 1572 m, 1476
m, 1425 s, 725 vs, 693 vs, 442 s; H NMR (300 MHz, CDCl
3
3
1
2
2
before 0.25
3
) δ
-
4
-5
mL of a 2.03 × 10 M solution of 1a or 0.2 mL of a 9.3 × 10
7.1-7.8 (m, aromatics); MS (FAB, positive ions; m/ z) 1251
+
+
+
+
+
+
+
+
M solution of 1b (CH Cl ) was added. A vigorous H
2
2
2
evolution
(Sn
(Sn
(SnPh
6
Ph
Ph
7
), 895 (Sn
), 623 (Sn
3
Ph
7
), 861 (Sn
), 546 (Sn
4
Ph
Ph
5
), 741 (Sn
), 469 (Sn
3
Ph
Ph
5
), 700
), 351
was immediately observed in both cases.
2
6
2
Ph
5
2
4
2
3
n
+
+
Deh yd r ogen a tive P olym er iza tion of H
2
Sn Bu
2
. Addi-
3
), 197 (SnPh ). Anal. Calcd for C12
H
10Sn: C, 52.8;
-
4
tion of 1 mL of a 3.36 × 10 M CH
2
Cl
2
solution of 1b to 8 g of
H, 3.69. Found: C, 52.3; H, 3.55.
n
H
2
Sn Bu
2
in 30 mL of Et
2
O resulted in a vigorous gas evolution
(
the whole vessel was wrapped in an aluminium foil to protect
Ack n ow led gm en t. We are grateful to Dr. E. Franta,
Mr. R. Meens, and Mr. A. Rameau (Institut Charles
Sadron, Strasbourg) for the GPC analyses and to Rh oˆ ne-
Poulenc Chimie and Wacker-Chemie GmbH for the gifts
of chemicals. Financial support from the Centre Na-
tional de la Recherche Scientifique (Paris) and the
Minist e` re de la Recherche et de la Technologie and the
Commission of the European Communities (Contract
No. CHRX-CT93-0277) are gratefully acknowledged.
the mixture from room light). Hydrogen evolution slowed
down after ca. 1 min but continued over the next 2 h. The
volume of H
TON of ca. 20 000, assuming that 1 mol equiv of H
released per catalytic cycle. The experiment was repeated 3
2
released was 169 mL, which corresponded to a
2
was
n
times with 2.4 g of H
2
Sn Bu
and thus facilitate the analysis of the
product. The average volume of H was 165 mL, and the
2
in order to avoid the presence of
n
2 2
unreacted H Sn Bu
2
average TON ca. 20 000. After the volatiles were removed
under vacuum, the pale brown waxy residue was analyzed.
OM970353F