Novel high-yield method for the synthesis of [Os3(CO) 10(μ-H)(μ-OSiR2R′)] species and X-ray characterization of [Os3(CO)10(μ-H)(μ-OSiPh 2OSiPh 2OH)] and [Os3(CO)10 (μ-H)(μ-OSiPh2OH)], first molecular models of osmium surface species anchored to vicinal

Article abstract of DOI:10.1021/om9908074

The easy replacement of the hydroxy ligand in [Os₃(CO)₁₀(μ-H)(μ-OH)] by siloxy ligands affords a high-yield procedure to model compounds of the type [Os₃(CO)₁₀(μ-H)(μ-OSiR₂R′)] (R = Et, Ph; R′ = Et, Ph, OH, OSiPh₂OH), a springboard to better understand the reactivity of osmium carbonyl species on the silica surface. The new clusters [Os₃(CO)₁₀(μ-H)(μ-OSiPh₂-OH)] and [Os₃(CO)₁₀(μ-H)(μ-OSiPh₂OSiPh ₂OH)], characterized by X-ray crystallography, represent the first molecular models of silica-anchored [Os₃(CO)₁₀(μ-H)(μ-OSi≡)] having a ligand that mimics surface geminal and vicinal silanols, respectively.

Full text of DOI:10.1021/om9908074

Organometallics 2000, 19, 1051-1059
1051
Novel High -Yield Meth od for th e Syn th esis of
[Os3(CO)10(µ-H)(µ-OSiR2R′)] Sp ecies a n d X-r a y
Ch a r a cter iza tion of [Os3(CO)10(µ-H)(µ-OSiP h 2OSiP h 2OH)]
a n d [Os3(CO)10(µ-H)(µ-OSiP h 2OH)], F ir st Molecu la r
Mod els of Osm iu m Su r fa ce Sp ecies An ch or ed to Vicin a l
a n d Gem in a l Su r fa ce Sila n ols
Elena Lucenti, Dominique Roberto,* Carmen Roveda, and Renato Ugo
Dipartimento di Chimica Inorganica, Metallorganica e Analitica and Centro CNR
“CSSSCMTBSO”, Universita` di Milano, Via G. Venezian, 21, 20133 Milano, Italy
Angelo Sironi
Dipartimento di Chimica Strutturale e Stereochimica Inorganica and Centro CNR
“CSSSCMTBSO”, Universita` di Milano, Via G. Venezian, 21, 20133 Milano, Italy
Received October 12, 1999
The easy replacement of the hydroxy ligand in [Os3(CO)10(µ-H)(µ-OH)] by siloxy ligands
affords a high-yield procedure to model compounds of the type [Os3(CO)10(µ-H)(µ-OSiR2R′)]
(R ) Et, Ph; R′ ) Et, Ph, OH, OSiPh2OH), a springboard to better understand the reactivity
of osmium carbonyl species on the silica surface. The new clusters [Os3(CO)10(µ-H)(µ-OSiPh2-
OH)] and [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)], characterized by X-ray crystallography,
represent the first molecular models of silica-anchored [Os3(CO)10(µ-H)(µ-OSit)] having a
ligand that mimics surface geminal and vicinal silanols, respectively.
In tr od u ction
and HCl(aq) to give [Os3(CO)10(µ-H)(µ-OH)]11 and
[Os3(CO)10(µ-H)(µ-Cl)],12 respectively, confirmed the
binding of the osmium carbonyl cluster to the silica
surface via only one silanol group and its hydridic
nature.
Since the pioneering work of Ugo and Basset on the
covalent binding of the “Os3(CO)10(µ-H)” fragment to the
silica surface via surface silanol groups,1-3 a lot of
spectroscopic work has been devoted to confirm the
nature of the species formed by interaction of [Os3-
(CO)12] with silica at ca. 150 °C because, based only on
infrared evidence, the triosmium core could bind a single
surface oxygen, to form [Os3(CO)10(µ-H)(µ-OSit)], or two
surface oxygens to form [Os3(CO)10(µ-OSit)2].1-3 Evi-
dence for the binding via a single surface silanol group
Such an organometallic carbonyl cluster species co-
valently bound to the silica surface not only is a catalyst
for olefin isomerization13 and hydrogenation14 but also
appears to be a possible intermediate in the high-yield
synthesis of various neutral and anionic osmium car-
bonyl clusters starting from OsCl3 or [Os(CO)3Cl2]2
supported on silica in the presence of alkali carbonates
and physisorbed water.15,16 However heterogeneous
systems are more difficult to characterize than homo-
geneous systems. Therefore, due to the usefulness of
specific organometallic molecules as models for getting
a better understanding of the nature of surface or-
ganometallic species,17-23 the synthesis of molecular
has been reached by infrared,1-6 Raman,4 EXAFS
1b,6,7
and 13C NMR8 spectroscopies. Computer modeling9,10
and the reaction of [Os3(CO)10(µ-H)(µ-OSit)] with HF(aq)
(1) (a) Ugo, R.; Psaro, R.; Zanderighi, G. M.; Basset, J . M.; Theolier,
A.; Smith, A. K. In Fundamental Research in Homogeneous Catalysis,
Vol 3; Tsutsui, M., Ed.; Plenum Press: New York, 1979. (b) Besson,
B.; Moraweck, B.; Smith, A. K.; Basset, J . M.; Psaro, R.; Fusi, A.; Ugo,
R. J . Chem. Soc., Chem. Commun. 1980, 569.
(2) (a) Smith, A. K.; Besson, B.; Basset, J . M.; Psaro, R.; Fusi, A.;
Ugo, R. J . Organomet. Chem. 1980, 192, C31. (b) Psaro, R.; Ugo, R.;
Zanderighi, G. M.; Besson, B.; Smith, A. K.; Basset, J . M. J . Organomet.
Chem. 1981, 213, 215.
(3) Basset, J . M.; Choplin, A. J . Mol. Catal. 1983, 21, 95.
(4) (a) Deeba, M.; Gates, B. J . Catal. 1981, 67, 303. (b) Deeba, M.;
Streusand, B. J .; Schrader, G. L.; Gates, B. C. J . Catal. 1981, 69, 218.
(5) Cook, S. L.; Evans, J .; Greaves, G. N. J . Chem. Soc., Chem.
Commun. 1983, 1287.
(6) Cook, S. L.; Evans, J .; McNulty, G. S.; Greaves, G. N. J . Chem.
Soc., Dalton Trans. 1986, 7.
(7) Drivenvoorden, F. B. M.; Konigsberger, Y. S.; Gates, B. C. J . Am.
Chem. Soc. 1986, 108, 6254.
(10) Shay, T. B.; Hsu, L. Y.; Basset, J . M.; Shore, S. G. J . Mol. Catal.
1994, 86, 479.
(11) Dossi, C.; Fusi, A.; Pizzotti, M.; Psaro, R. Organometallics 1990,
9, 1994.
(12) Roberto, D.; Pizzotti, M.; Ugo, R. Gazz. Chim. Ital. 1995, 125,
133.
(13) Barth, R.; Gates, B. C.; Zhao, Y.; Knozinger, J .; Hulse, J . J .
Catal. 1983, 82, 147.
(14) Choplin, A.; Besson, B.; D’Ornelas, L.; Sanchez-Delgado, R.;
Basset, J . M. J . Am. Chem. Soc. 1988, 110, 2783.
(15) Cariati, E.; Roberto, D.; Ugo, R. Gazz. Chim. Ital. 1996, 126,
339.
(16) Cariati, E.; Recanati, P.; Roberto, D.; Ugo, R. Organometallics
1998, 17, 1266.
(17) Evans, J .; Gracey, B. P. J . Chem. Soc., Dalton Trans. 1982,
1123.
(18) D’Ornelas, L.; Choplin, A.; Basset, J . M.; Hsu, L. Y.; Shore, S.
Nouv. J . Chim. 1985, 9, 155.
(8) Walter, T. H.; Frauenhoff, G. R.; Shapley, J . R.; Oldfield, E. Inorg.
Chem. 1988, 27, 2561.
(9) (a) Hsu, L. Y.; Shore, S. G.; D’Ornelas, L.; Choplin, A.; Basset,
J . M. Polyhedron 1988, 7, 2399. (b) Hsu, L. Y.; Shore, S. G.; D’Ornelas,
L.; Choplin, A.; Basset, J . M. J . Catal. 1994, 149, 159.
10.1021/om9908074 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 02/18/2000

1052 Organometallics, Vol. 19, No. 6, 2000
Lucenti et al.
models of the surface species [Os3(CO)10(µ-H)(µ-OSit)]
has attracted increasing attention.2,17,18,19d However,
models such as [Os3(CO)10(µ-H)(µ-OSiR3)] (R ) Et,
Ph)2,17,18 and [Os3(CO)10(µ-H)(µ-OSi7O10(c-C6H11)7)]19d
could be obtained only with rather low yields (ca.
8-28%) in one or more steps starting from [Os3(CO)12].
Probably for this reason, the more easily prepared
[Os3(CO)10(µ-H)(µ-OPh)] was used as molecular model
in order to obtain mechanistic insight into the olefin
hydrogenation catalyzed by silica-anchored [Os3(CO)10-
(µ-H)(µ-OSit)].14
Surprisingly, only molecular models with a ligand
mimicking surface-isolated silanols had been pre-
pared,2,17,18,19d although on the silica surface there are
different silanol types that could exhibit different
reactivity and whose concentration is a function of the
temperature:24(i) the isolated groups (or free silanols,
where the surface silicon atom has three bonds into the
bulk structure and the fourth bond attached to a single
OH group), which correspond to 25% of the total surface
silanols at 200 °C, a temperature similar to that used
for the preparation of silica-anchored [Os3(CO)10(µ-H)-
(µ-OSit)];1-3 (ii) the geminal silanols (where two hy-
droxyl groups are attached to one silicon atom), which
correspond to 13% of the total surface silanols at 200
°C; (iii) the vicinal silanols (where two single silanol
groups, attached to different silicon atoms, are close
enough to hydrogen bond), which correspond to 62% of
the total surface silanols at 200 °C.
a high-yield synthesis of [Os3(CO)10(µ-H)(µ-OR)] (R )
Me, Bu, or Ph) by reaction of [Os3(CO)10(µ-H)(µ-OH)]
with ROH.25 Since the displacement of the hydroxy
ligand was favored by an increased acidity of the ROH
species (e.g., PhOH better than MeOH), it was expected
that molecular silanols should displace rather easily the
hydroxy ligand from [Os3(CO)10(µ-H)(µ-OH)] because
their acidity is higher than that of alcohols, although it
is lower than that of phenol.26 To mimic in an extensive
way the behavior of a silica surface, we choose three
types of molecular silanols: (i) R3SiOH (R ) Et, Ph) to
mimic surface-isolated silanols; (ii) Ph2Si(OH)2 to mimic
surface geminal silanols, and (iii) HOSiPh2OSiPh2OH
to mimic surface vicinal silanols.
In this work we present the results of this investiga-
tion which, by reproducing a reaction that occurs on the
silica surface, led to a new high-yield method for the
preparation, under mild conditions and in excellent
yields, of the known models [Os3(CO)10(µ-H)(µ-OSiR3)]
(R ) Et, Ph), obtained up to now in very low yields.2,17,18
We also report the synthesis and X-ray characterization
of the new species [Os3(CO)10(µ-H)(µ-OSiPh2OH)] and
[Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)], which are the
first molecular models of silica-anchored [Os3(CO)10(µ-
H)(µ-OSit)] having a ligand that mimics surface gemi-
nal and vicinal silanols, respectively.
Resu lts a n d Discu ssion
1. Syn th esis of [Os3(CO)10(µ-H)(µ-OSiR2R′)] (R )
Et, P h ; R′ ) Et, P h , OH, OSiP h 2OH). We have
evidence from our investigation on the reactivity of silica
physisorbed [Os3(CO)10(µ-H)(µ-OH)]15 that the displace-
ment of the hydroxy group by surface silanol groups is
an equilibrium reaction:
Computer modeling has suggested binding of the
“Os3(CO)10(µ-H)” moiety to vicinal or geminal surface
silanols;9,10 therefore it was an interesting challenge to
find an easy route for a simple, high-yield preparation
of molecular models of the silica-anchored [Os3(CO)10-
(µ-H)(µ-OSit)] species grafted not only to a surface-
isolated silanol but also to a surface vicinal or even
geminal silanol.
In the course of our investigation on the multistep
process of formation of various osmium carbonyl clusters
from silica physisorbed OsCl3 or [Os(CO)3Cl2]2,16 we
reached clear evidence for a facile equilibrium on the
silica surface between physisorbed [Os3(CO)10(µ-H)(µ-
OH)] and [Os3(CO)10(µ-H)(µ-OSit)].15 Such an observa-
tion prompted us to investigate the reaction of [Os3-
(CO)10(µ-H)(µ-OH)] with different molecular silanols in
order to prepare a significant series of molecular models.
Our choice was also supported by our recent finding of
[Os3(CO)10(µ-H)(µ-OH)] + tSiOH /
[Os3(CO)10(µ-H)(µ-OSit)] + H2O
Similar evidence has been reached by Bergman et al.
with a hydroxo iridium complex.22 Therefore when
extending this kind of reactivity to organic silanols, we
had to work with continuous water removal, for example
by use of a Markusson apparatus or by working in the
presence of a water scavenger such as P2O5. This
procedure must be carefully defined in order to avoid
self-condensation of silanols to siloxanes.27
1.1. Syn th esis of [Os3(CO)10(µ-H)(µ-OSiP h 3)]. The
reaction of Ph3SiOH with [Os3(CO)10(µ-H)(µ-OH)], pre-
pared by hydrolysis of silica-anchored [Os3(CO)10(µ-H)-
(µ-OSit)],25 was first investigated under nitrogen in
anhydrous m-xylene at reflux temperature (138 °C),
using a Markusson apparatus in order to shift the
equilibrium by removing water. The reaction was best
followed by 1H NMR spectroscopy, looking at the molar
ratio of the H-Os resonance of [Os3(CO)10(µ-H)(µ-OH)]
(δ in CDCl3 ) -12.64 ppm) and [Os3(CO)10(µ-H)(µ-
OSiPh3)] (δ in CDCl3 ) -11.36 ppm).
(19) (a) Feher, F. J . J . Am. Chem. Soc. 1986, 108, 3850. (b) Feher,
F. J .; Gonzales, S. L.; Ziller, J . W. Inorg. Chem. 1988, 27, 3440. (c)
Feher, F. J .; Walzer, J . F. Inorg. Chem. 1990, 29, 1604. (d) Liu, J . C.;
Wilson, S. R.; Shapley, J . R.; Feher, F. J . Inorg. Chem. 1990, 29, 5138.
(e) Feher, F. J .; Walzer, J . F. Inorg. Chem. 1991, 30, 1689.
(20) (a) Palyi, G.; Zucchi, C.; Ugo, R.; Psaro, R.; Sironi, A.; Vizi-
Orosz, A. J . Mol. Catal. 1992, 74, 51. (b) Vizi-Orosz, A.; Ugo, R.; Psaro,
R.; Sironi, A.; Moret, M.; Zucchi, C.; Ghelfi, F.; Palyi, G. Inorg. Chem.
1994, 33, 4600. (c) Zucchi, C.; Shchegolikhina, O. I.; Borsari, M.; Cornia,
A.; Gavioli, G.; Fabretti, A. C.; Rentschler, E.; Gatteschi, D.; Ugo, R.;
Psaro, R.; Pozdniakova, Y. A.; Lindeman, S. V.; Zhdanov, A. A.; Palyi,
G. J . Mol. Catal. A 1996, 107, 313.
(21) Roberto, D.; Cariati, E.; Pizzotti, M.; Psaro, R. J . Mol. Catal. A
1996, 111, 97.
(22) Meyer, T. Y.; Woerpel, K. A.; Novak, B. M.; Bergman, R. G. J .
Am. Chem. Soc. 1994, 116, 10290.
(23) Murugavel, R.; Voigt, A.; Walawalkar, M. G.; Roesky, H. W.
Chem. Rev. 1996, 96, 2205.
(24) Vansant, E. F.; Van Der Voort, P.; Vrancken, K. C. In
Characterization and Chemical Modification of the Silica Surface in
Studies in Surface Science and Catalysis, Vol. 93; Delmon, B., Yates,
J . T., Eds.; Elsevier: Amsterdam, 1995.
(25) Roberto, D.; Lucenti, E.; Roveda, C.; Ugo, R. Organometallics
1997, 16, 5974.
(26) West, R.; Baney, R. J . Am. Chem. Soc. 1959, 81, 6145.
(27) (a) Rutz, W.; Lange, D.; Kelling, H. Z. Anorg. Allg. Chem. 1985,
528, 98. (b) The Chemistry of Organic Silicon Compound; Patai, S.,
Rappoport, Z., Eds.; J ohn Wiley & Sons: New York, 1989. (c) Lickiss,
P. D. Adv. Inorg. Chem. 1995, 42, 147.

Synthesis of [Os3(CO)10(µ-H)(µ-OSiR2R′)] Species
Organometallics, Vol. 19, No. 6, 2000 1053
Ta ble 1. Syn th esis of [Os3(CO)10(µ-H)(µ-OSiR2R′)] (R ) Et, P h ; R′ ) Et, P h , OSiP h 2OH, OH) fr om
[Os3(CO)10(µ-H)(µ-OH)]a
molar ratio
[HOs3(CO)10OSiR2R′]
R2R′SiOH
[HOs3(CO)10OH]mol/L
[HOs3(CO)10OH]:R2R′SiOH
t (h)
yield (%)
Ph3SiOHb
6 × 10-4
1:30
1:300
5
5
24
10
5
7
24
2.5
4
11
3
25c
Ph3SiOHb,d
6 × 10-4
50c
100c
b
Ph3SiOH + HBF4
Ph3SiOHe
6 × 10-4
1 × 10-1
1 × 10-1
4 × 10-4
6 × 10-1
1 × 10-1
1 × 10-1
5 × 10-2
9 × 10-2
3 × 10-2
1:300
1:2
1:1
1:300
1:9
1:2
2:1
1:2
1:10
1:60
0c
100c (84)f
67c,g
Ph3SiOHe
Et3SiOHb
50c
Et3SiOHe,h
100c (81)f
100c (84)f
50c
e
O(SiPh2OH)2
O(SiPh2OH)2
Ph2Si(OH)2
Ph2Si(OH)2
e
e
40c,i
e
2.5
0.83
2.5
69c,j
e
Ph2Si(OH)2
100c (79)f
77c,k
Reactions carried out under N2 in m-xylene at 138 °C. Using the Markusson apparatus as water scavenger. c 1H NMR yield; obtained
a
b
from the molar ratio of the H-Os resonance of [Os3(CO)10(µ-H)(µ-OH)] and [Os3(CO)10(µ-H)(µ-OSiR2R′)]; the remainder is unreacted
d
[Os3(CO)10(µ-H)(µ-OH)]. Similar results are obtained using P2O5 as water scavenger. e Using P2O5 as water scavenger. f Isolated yields
g
h
of material after chromatography on silica gel. [H2Os3(CO)10] (4% yield) is also formed. Reaction carried out in pure Et3SiOH.
i [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] (20% yield) is also formed. j [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] is also formed (31% yield); similar
results are obtained with a concentration of [Os3(CO)10(µ-H)(µ-OH)] ) 3 × 10-2 M. [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] (23% yield) is
k
also formed.
When a dilute solution of [Os3(CO)10(µ-H)(µ-OH)] (6
× 10-4 M) and Ph3SiOH (molar ratio [Os3(CO)10(µ-H)-
(µ-OH)]:Ph3SiOH ) 1:30) in m-xylene is refluxed for 5
h, [Os3(CO)10(µ-H)(µ-OSiPh3)] is obtained in 25% yield
(1H NMR yield; Table 1), the remainder being unreacted
[Os3(CO)10(µ-H)(µ-OH)]. An increase of the reaction time
leads to thermal decomposition of the desired cluster,
affording a mixture of [H2Os3(CO)10], [H4Os4(CO)12], and
traces of a still unidentified hydridic carbonyl osmium
species (δ in CDCl3 ) -15.54 ppm). The thermal
decomposition of [Os3(CO)10(µ-H)(µ-OSiPh3)] to metallic
particles and unidentified carbonyl species was previ-
ously reported.18
However, in agreement with the increased reactivity
of [Os3(CO)10(µ-H)(µ-OH)] when working with excess
phenol to give [Os3(CO)10(µ-H)(µ-OPh)],25 the formation
of [Os3(CO)10(µ-H)(µ-OSiPh3)] is favored when the rela-
tive amount of Ph3SiOH is increased. By using a molar
ratio [Os3(CO)10(µ-H)(µ-OH)]:Ph3SiOH ) 1:300, [Os3-
(CO)10(µ-H)(µ-OSiPh3)] is obtained in 50%, 66%, and
100% yields after 5, 10, and 24 h, respectively. Similar
results are obtained by using a small reaction flask
equipped with a drying tube containing P2O5, instead
of the Markusson apparatus (Table 1).
Evaporation of the solvent, followed by chromatography,
affords the final product in 84% isolated yield.
1.2. Syn th esis of [Os3(CO)10(µ-H)(µ-OSiEt3)]. When
a dilute solution of [Os3(CO)10(µ-H)(µ-OH)] ((4-6) × 10-4
M) and Et3SiOH (molar ratio [Os3(CO)10(µ-H)(µ-OH)]:
Et3SiOH ) 1:300) in m-xylene is refluxed, under nitro-
gen using a Markusson apparatus, for 10 and 24 h,
[Os3(CO)10(µ-H)(µ-OSiEt3)] is obtained in 25% and 50%
yields only, respectively (1H NMR yield; Table 1), the
remainder being unreacted [Os3(CO)10(µ-H)(µ-OH)].
However, by working in pure Et3SiOH (molar ratio
[Os3(CO)10(µ-H)(µ-OH)]:Et3SiOH ) 1:9), at 138 °C under
nitrogen and in the presence of P2O5, [Os3(CO)10(µ-H)-
(µ-OH)] is rapidly (less than 3 h) converted to [Os3(CO)10-
(µ-H)(µ-OSiEt3)] in quantitative yield (Table 1). Evapo-
ration of the excess of Et3SiOH, followed by chromatog-
raphy, affords the final product in 81% isolated yield.
1.3. Syn t h esis of [Os3(CO)10(µ-H )(µ-OSiP h 2OSi-
P h 2OH)]. Treatment of a solution of [Os3(CO)10(µ-H)-
(µ-OH)] (1 × 10-1 M) and HOSiPh2OSiPh2OH (molar
ratio [Os3(CO)10(µ-H)(µ-OH)]:HOSiPh2OSiPh2OH ) 1:2)
in m-xylene, for 4 h at 138 °C under nitrogen and in
the presence of P2O5, affords [Os3(CO)10(µ-H)(µ-OSiPh2-
OSiPh2OH)] in quantitative yield, as shown by 1H NMR
spectroscopy (Table 1). Evaporation of the solvent,
followed by chromatography, affords the final product
in 84% isolated yield.
While the formation of [Os3(CO)10(µ-H)(µ-OR)] (R )
Me, Bu, Ph) from [Os3(CO)10(µ-H)(µ-OH)] and ROH is
acid-catalyzed,25 the addition of a few drops of HBF4‚
Et2O inhibits the formation of [Os3(CO)10(µ-H)(µ-
OSiPh3)] due to the known fast condensation of silanols
under acidic conditions.27
The structure of the new cluster [Os3(CO)10(µ-H)(µ-
1
OSiPh2OSiPh2OH)] was characterized by infrared, H
NMR, 29Si NMR, and mass spectra (see Experimental
Section) and confirmed by X-ray diffraction (Figure 1,
see later). This is the first molecular model of silica-
anchored [Os3(CO)10(µ-H)(µ-OSit)] having a ligand that
better mimics the silica surface. In fact at the temper-
ature (less than 200 °C) used for the formation of silica-
anchored [Os3(CO)10(µ-H)(µ-OSit)],1-3 most surface si-
lanols are vicinal (more than 62%);24 in addition computer
modeling showed that a good fit can be achieved by
placing the “Os3(CO)10(µ-H)” fragment on a vicinal
oxygen of a silica surface siloxane group.9,10
The use of excess silanol, which requires a difficult
purification, can be avoided by working with a concen-
trated solution of [Os3(CO)10(µ-H)(µ-OH)] (1 × 10-1
M
instead of 6 × 10-4 M). In this latter case, the formation
of [Os3(CO)10(µ-H)(µ-OSiPh3)] occurs readily, even with-
out an excess of Ph3SiOH, and is much faster than its
thermal decomposition. Treatment of a solution of
[Os3(CO)10(µ-H)(µ-OH)] (1 × 10-1 M) and Ph3SiOH
(molar ratio [Os3(CO)10(µ-H)(µ-OH)]:Ph3SiOH ) 1:2) in
m-xylene, at 138 °C under nitrogen and in the presence
of P2O5 for 5 h, affords [Os3(CO)10(µ-H)(µ-OSiPh3)] in
1.4. Syn th esis of [Os3(CO)10(µ-H)(µ-OSiP h 2OH)].
Computer modeling for the surface sites of hydroxylated
1
quantitative yield, as shown by H NMR spectroscopy.

1054 Organometallics, Vol. 19, No. 6, 2000
Lucenti et al.
F igu r e 1. ORTEP diagram of [Os3(CO)10(µ-H)(µ-OSiPh2-
OSiPh2OH)] showing 50% thermal ellipsoids. C atoms of
the CO groups bear the same numbering of the corre-
sponding O atoms. H atoms of the phenyl groups are
omitted for clarity.
F igu r e 2. ORTEP diagram of [Os3(CO)10(µ-H)(µ-OSiPh2-
OH)] showing 50% thermal ellipsoids. C atoms of the CO
groups bear the same numbering of the corresponding O
atoms. H atoms of the phenyl groups are omitted for clarity.
OSiPh2OSiPh2OH)] (23% yield), due to further conden-
sation of [Os3(CO)10(µ-H)(µ-OSiPh2OH)] with excess
Ph2Si(OH)2.
silica, based on the structure of â-cristobalite, suggested
that the “Os3(CO)10(µ-H)” fragment cannot be accom-
modated on a geminal silanol of a fully hydroxylated
(100) surface. However, a satisfactory fit can be achieved
by placing the cluster on a geminal silanol of a defect
site created by the intersection of the (100) face with
the (001) face.9 To confirm that this kind of unusual and
more difficult binding to the silica surface could occur,
we choose Ph2Si(OH)2 as a simple model for geminal
silanols of the silica surface. Treatment of a solution of
[Os3(CO)10(µ-H)(µ-OH)] (5 × 10-2 M) and Ph2Si(OH)2
(molar ratio [Os3(CO)10(µ-H)(µ-OH)]: Ph2Si(OH)2 ) 1:2)
in m-xylene, for 3 h at 138 °C under nitrogen and in
the presence of P2O5, affords [Os3(CO)10(µ-H)(µ-OSiPh2-
OH)] (40% 1H NMR yield; Table 1) along with some
[Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] (20% 1H NMR
yield), the remainder being unreacted [Os3(CO)10(µ-H)-
(µ-OH)].
The reaction is easier in the presence of a higher
amount of Ph2Si(OH)2 (molar ratio [Os3(CO)10(µ-H)(µ-
OH)]:Ph2Si(OH)2 ) 1:10; concentration of [Os3(CO)10(µ-
H)(µ-OH)] ) (3-9) × 10-2 M). After 2.5 h, the conversion
of [Os3(CO)10(µ-H)(µ-OH)] is complete, but a mixture of
[Os3(CO)10(µ-H)(µ-OSiPh2OH)] (69% 1H NMR yield) and
[Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] (31% 1H NMR
yield) is obtained, which can be separated by column
chromatography.
The structure of the new cluster [Os3(CO)10(µ-H)(µ-
1
OSiPh2OH)] was characterized by infrared, H NMR,
29Si NMR, and mass spectra (see Experimental Section)
and confirmed by X-ray diffraction (Figure 2, see later).
This is the first molecular model of silica-anchored
[Os3(CO)10(µ-H)(µ-OSit)] having a ligand that mimics
surface geminal silanols.
1.5. An In vestiga tion of th e Rea ctivity of th e
F r ee Sila n ol Gr ou p s of [Os3(CO)10(µ-H)(µ-OSiP h 2-
OSiP h 2OH)] a n d [Os3(CO)10(µ-H)(µ-OSiP h 2OH)].
Both silanol groups of either the disilanol HOSiPh2-
OSiPh2OH or the silanediol Ph2Si(OH)2 can react with
various early transition metals and with main group
metals, affording cyclic metallasiloxanes.23 For example
reaction of HOSiPh2OSiPh2OH with TiCl4 affords the
spirocyclic complex [Ti(OSiPh2OSiPh2O)2],28 while reac-
tion with [Zr(CH2SiMe3)4] leads to the formation of
[(Me3SiCH2)2Zr(OSiPh2OSiPh2O)] with a six-membered
ZrO3Si2 ring.23,29 Recently it was reported that the same
disilanol reacts with chromium trioxide to give [Cr(dO)2-
(OSiPh2OSiPh2O)]2, where the two disilanolate ligands
bind to two chromium dioxo units, generating a unique
12-membered metallacyclic fragment.30 In addition,
Ph2Si(OH)2 shows a tendency to self-condense. For
example, it reacts with Ti(OBu)4 to yield the spirocyclic
complex [Ti(OSiPh2(OSiPh2)3O)2],31 although it reacts
with R2GeX2 (R ) Me, Ph) to give [Ph2SiO2GeR2]2, which
[Os3(CO)10(µ-H)(µ-OH)] can be selectively converted
to [Os3(CO)10(µ-H)(µ-OSiPh2OH)] only by working under
well-controlled conditions, with a huge amount of
Ph2Si(OH)2 (molar ratio [Os3(CO)10(µ-H)(µ-OH)]:Ph2Si-
(OH)2 ) 1:60). After 50 min, the desired cluster is
formed in 100% 1H NMR yield. Evaporation of the
solvent, followed by chromatography, affords the pure
cluster in 79% isolated yield. Interestingly, an increase
of the reaction time to 2.5 h decreases the selectivity,
leading to the formation of some [Os3(CO)10(µ-H)(µ-
(28) Andrianov, K. A.; Kurasheva, N. A.; Kuteinikova, L. I. Zh.
Obshch. Khim. 1976, 46, 1533; Chem. Abst. 1976, 85, 177526g.
(29) Hrncir, D. C. Mater. Res. Symp. Proc. 1988, 121, 127.
(30) Abbenhuis, H. C. L.; Vorstenbosch, M. L. W.; van Santen, R.
A.; Smeets, W. J . J .; Spek, A. Inorg. Chem. 1997, 36, 6431.
(31) (a) Zeitler, V. A.; Brown, C. A. J . Am. Chem. Soc. 1957, 79,
4618. (b) Hursthouse, M. B.; Hossain, M. A. Polyhedron 1984, 3, 95.

Synthesis of [Os3(CO)10(µ-H)(µ-OSiR2R′)] Species
Organometallics, Vol. 19, No. 6, 2000 1055
has a Ge2Si2O4 eight-membered ring core structure.32
Therefore it was interesting to find whether the remain-
ing free silanol groups of [Os3(CO)10(µ-H)(µ-OSiPh2-
OSiPh2OH)] and [Os3(CO)10(µ-H)(µ-OSiPh2OH)] could
react with an excess of [Os3(CO)10(µ-H)(µ-OH)].
However, various attempts to prepare bridged di-
nuclear species such as [Os3(CO)10(µ-H)(µ-OSiPh2-
OSiPh2O-µ)(µ-H)(CO)10Os3] or [Os3(CO)10(µ-H)(µ-OSi-
Ph2O-µ)(µ-H)(CO)10Os3] failed. Working under standard
reaction conditions used for the preparation of mono-
nuclear models starting from [Os3(CO)10(µ-H)(µ-OH)],
an excess of [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] does
not react with [Os3(CO)10(µ-H)(µ-OH)] (molar ratio
[Os3(CO)10(µ-H)(µ-OH)]:[Os3(CO)10(µ-H)(µ-OSiPh2OSi-
Ph2OH)] ) 1:4; see Experimental Section). The same
lack of reactivity was shown by [Os3(CO)10(µ-H)(µ-
OSiPh2OH)] working under similar conditions.
temperature, to give [Os3(CO)10(µ-H)(µ-CtCPh)].35 How-
ever both Ph3SiOLi and O(SiPh2OLi)2 do not react
with [Os3(CO)10(µ-H)(µ-Cl)] and [Os3(CO)10(µ-H)(µ-
O2CCF3], by working under nitrogen in benzene in a
range of temperatures between 25 and 75 °C.
36
2. X-r a y St r u ct u r es of [Os3(CO)10(µ-H )(µ-OSi-
P h 2OSiP h 2OH)] an d [Os3(CO)10(µ-H)(µ-OSiP h 2OH)].
The structures of [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)]
and [Os3(CO)10(µ-H)(µ-OSiPh2OH)] are of a certain
interest, being the first examples of a “Os3(CO)10(µ-H)”
cluster core carrying a free silanol group. The molecular
arrangements are shown with the atomic labeling in
Figures 1 and 2. Relevant bond distances and angles
are listed in Table 2. The two compounds share the
expected presence of a triangular metal atom framework
with two “long” (Os1-Os3 and Os2-Os3) and one
“short” Os-Os (Os1-Os2) edge. The short edge, in both
clusters, is bridged by the µ-OSiR2OX ligand and by the
µ-H (hydride) atom (which has been located in the
Fourier map and refined in both cases). As commonly
observed for M(µ-H)(µ-X)M moieties, the lengthening
that usually accompanies the formation of a three-center
two-electron bond upon H-bridging is counterbalanced
by an effect in the opposite direction due to the µ-OX
ligand; this eventually results in a shortening of the
Os-Os distance. The Os-C bond distances are strictly
comparable in the two clusters, and their pattern clearly
convey the picture that their relative length depends
on the nature of the trans-ligand: trans-to-O (C11 and
C21, av 1.873(2) Å) < trans-to-H (C12 and C22, av
1.907(5) Å) ≈ trans-to-Os (C33 and C34, av 1.913(4) Å)
< trans-to-Os (C13 and C23, av 1.934(9) Å) ≈ trans-to-
CO (C31 and C32, av 1.943(12) Å).
The “Os3(CO)10(µ-H)(µ-O-)” cores, as far as only the
bond lengths are taken into account, conform to an
idealized Cs symmetry (the pseudo mirror plane passing
through the Os3, H, and O atoms). This is no longer
true if we consider that the Os2-Os3-C33 angle
(average in the two clusters 105.0°) is much larger than
the Os1-Os3-C34 one (average in the two clusters
96.1°). However, it is not clear if this behavior is due to
the fact that the µ-OSiR2OX ligands do not conform to
Cs symmetry. The different steric requirements of the
µ-OSiR2OX and µ-H ligands are easily quantified by the
analysis of the distortion of their environment [the Os1-
Os2-C22 and Os2-Os1-C12 (which range from 133.0°
to 136.6°) are definitely larger than the Os1-Os2-C21
and Os2-Os1-C11 bond angles (which range from
117.4° to 120.4°)].
It appears that, under these conditions, the free
silanol groups of [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)]
and [Os3(CO)10(µ-H)(µ-OSiPh2OH)] do not displace eas-
ily the hydroxy ligand of [Os3(CO)10(µ-H)(µ-OH)]. The
reason for this much lower reactivity is not clear yet. It
cannot be attributed to a much lower acidity of the
remaining free silanol groups since the 29Si NMR
chemical shift of HOSiPh2OSiPh2OH (δ ) -36.37 ppm
in CDCl3) is very similar to that of the free silanol of
the complex [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] (δ )
-36.41 ppm in CDCl3), suggesting a similar electron
density at silicon in both compounds.33 The approach
of another [Os3(CO)10(µ-H)(µ-OH)] cluster to the free
silanol group of a geminal structure could be difficult
due to the proximity of the fragment “Os3(CO)10(µ-H)-
(µ-OSiPh2)” (see Figure 2). However this approach
cannot be so hindered in the case of [Os3(CO)10(µ-H)(µ-
OSiPh2OSiPh2OH)], where the silanol group should be
pointing, as in the crystal structure (see Figure 1), to
the other side with respect to the “Os3(CO)10(µ-H)”
fragment.
Although we cannot give at the moment a satisfactory
explanation, these results are in agreement with the
observation that silica-anchored [Os3(CO)10(µ-H)(µ-
OSit)] can be prepared only with loadings less than 4
wt % Os/SiO2,8,25 which corresponds to a grafting on less
than 5% of the total surface silanol groups when
working with Degussa Aerosil 200 (number of surface
silanols equals ca. 4.9 OH/nm2, corresponding to 1.5
mmol OH/g SiO2).24 The failure to obtain silica-anchored
[Os3(CO)10(µ-H)(µ-OSit)], with high loadings, could be
due to the low reactivity of silanol groups located near
surface silanolate groups binding a “Os3(CO)10(µ-H)”
fragment.
These molecular characteristics parallel those of the
related [Os3(CO)10(µ-H)(µ-Y)] models (Y ) OSiEt3,18
OSi7O10(c-C6H11)7,19d OR37), confirming that the other
substituents attached to the silicon atom linked to the
Os3 core do not have a significant influence on the
dimensions of the “Os3(CO)10(µ-H)” moiety and in par-
ticular on the “Os2(µ-H)(µ-OSit)” bond distances.
Finally in both crystal structures the free silanol
groups do not interact via intramolecular or intermo-
lecular hydrogen bonding either with the carbonyl
groups of a “Os3(CO)10(µ-H)” moiety or with an O atom
We investigated also another route to [Os3(CO)10(µ-
H)(µ-OSiPh2OSiPh2O-µ)(µ-H)(CO)10Os3] since it is known
that reaction of the dilithium salt O(SiPh2OLi)2 with
various metal chlorides affords silanolate derivatives
(e.g., CoCl2 can be converted to [Co(OSiPh2OSiPh2O)2]-
[Li(THF)2]2);34 in addition, [Os3(CO)10(µ-H)(µ-Cl)] was
reported to react with lithium phenylacetylide, at room
(32) Puff, H.; Bockmann, M. P.; Kok, T. R.; Schuh, W. J . Organomet.
Chem. 1984, 268, 197.
(33) Williams, E. A.; Cargioli, J . D.; Larochelle, R. W. J . Organomet.
Chem. 1976, 108, 153.
(34) Hursthouse, M. B.; Mazid, M. A.; Motevalli, M.; Sanganee, M.;
Sullivan, A. C. J . Organomet. Chem. 1990, 381, C43.
(35) Koridze, A. A.; Kizas, O. A.; Petrovskii, P. V.; Kolobova, N. E.;
Struchlov, Y. T.; Yanovsky, A. I. J . Organomet. Chem. 1988, 338, 81.
(36) Bostick, E. E.; Zdaniewski, J . J . Ger. Offen. 2,048,914; Chem.
Abstr. 1971, 75, 21496e.
(37) Churchill, M. R.; Wasserman, H. J . Inorg. Chem. 1980, 19, 2391.

1056 Organometallics, Vol. 19, No. 6, 2000
Lucenti et al.
Ta ble 2. Bon d Len gth s (Å) a n d An gles (d eg) for [Os3(CO)10(µ-H)(µ-OSiP h 2OSiP h 2OH)] a n d
[Os3(CO)10(µ-H)(µ-OSiP h 2OH)]
Os(1)-C(11)
1.875(5)
1.914(5)
1.939(4)
2.127(2)
1.72(3)
2.7999(2)
2.8254(2)
1.871(4)
1.905(5)
1.921(5)
2.133(3)
1.83(3)
1.873(5)
1.901(6)
1.934(5)
2.128(3)
1.65(4)
2.8090(3
2.8233(3
1.872(5)
1.909(6)
1.941(5)
2.126(3)
1.72(4)
2.8280(3
1.913(6)
1.917(6)
1.924(6)
1.953(6)
1.143(5)
1.141(6)
1.149(6)
1.149(5)
1.138(6)
1.135(5)
1.139(6)
1.149(6)
1.132(6)
1.124(6)
1.654(3)
1.645(3)
1.852(5)
1.857(5)
C(22)-Os(2)-Os(1)
C(23)-Os(2)-Os(1)
O(1)-Os(2)-Os(1)
C(21)-Os(2)-Os(3)
C(22)-Os(2)-Os(3)
C(23)-Os(2)-Os(3)
O(1)-Os(2)-Os(3)
Os(1)-Os(2)-Os(3)
C(33)-Os(3)-C(34)
C(33)-Os(3)-C(32)
C(34)-Os(3)-C(32)
C(33)-Os(3)-C(31)
C(34)-Os(3)-C(31)
C(32)-Os(3)-C(31)
C(33)-Os(3)-Os(1)
C(34)-Os(3)-Os(1)
C(32)-Os(3)-Os(1)
C(31)-Os(3)-Os(1)
C(33)-Os(3)-Os(2)
C(34)-Os(3)-Os(2)
C(32)-Os(3)-Os(2)
C(31)-Os(3)-Os(2)
Os(1)-Os(3)-Os(2)
O(2)-Si(1)-O(1)
132.98(15)
117.04(15)
48.81(7)
91.40(15)
86.51(16)
176.69(15)
82.44(7)
60.123(6)
100.0(2)
90.9(2)
136.46(14)
111.73(15)
48.70(7)
90.59(14)
90.86(16)
171.59(14)
81.60(7)
60.110(7)
98.9(3)
92.8(2)
92.4(2)
94.0(2)
95.7(2)
168.6(2)
164.6(2)
96.41(19)
84.23(15)
86.77(15)
105.1(2)
155.98(19)
84.36(14)
84.99(14)
59.612(7)
108.95(17)
109.1(2)
106.72(19)
109.8(2)
108.90(19)
113.2(2)
Os(1)-C(12)
Os(1)-C(13)
Os(1)-O(1)
Os(1)-H
Os(1)-Os(2)
Os(1)-Os(3)
Os(2)-C(21)
Os(2)-C(22)
Os(2)-C(23)
Os(2)-O(1)
93.9(2)
93.3(2)
94.2(2)
170.2(2)
Os(2)-H
Os(2)-Os(3)
2.8400(3)
1.908(5)
1.910(6)
1.948(5)
1.948(5)
1.135(5)
1.128(5)
1.126(5)
1.148(5)
1.140(5)
1.146(6)
1.135(6)
1.132(6)
1.133(6)
1.147(6)
1.631(3)
1.638(3)
1.858(4)
1.862(4)
1.631(3)
1.632(3)
1.853(4)
1.862(4)
89.1(2)
92.61(19)
95.83(18)
166.31(15)
100.39(16)
96.14(15)
117.42(14)
134.64(13)
117.16(14)
49.01(7)
87.94(14)
86.44(13)
177.68(14)
82.91(7)
60.642(7)
91.06(19)
88.7(2)
Os(3)-C(33)
Os(3)-C(34)
164.14(16)
95.73(15)
85.69(14)
87.86(15)
105.05(16)
154.96(15)
84.55(15)
85.77(15)
59.235(6)
105.77(14)
109.96(17)
109.33(16)
112.40(18)
111.01(17)
108.33(19)
108.72(16)
110.97(17)
106.54(19)
108.38(17)
111.09(17)
111.13(19)
178.5(4)
Os(3)-C(32)
Os(3)-C(31)
C(11)-O(11)
C(12)-O(12)
C(13)-O(13)
C(21)-O(21)
C(22)-O(22)
C(23)-O(23)
C(31)-O(31)
C(32)-O(32)
O(2)-Si(1)-C(111)
O(1)-Si(1)-C(111)
O(2)-Si(1)-C(121)
O(1)-Si(1)-C(121)
C(111)-Si(1)-C(121)
O(2)-Si(2)-O(3)
C(33)-O(33)
C(34)-O(34)
Si(1)-O(2)
Si(1)-O(1)
Si(1)-C(111)
Si(1)-C(121)
O(2)-Si(2)-C(221)
O(3)-Si(2)-C(221)
O(2)-Si(2)-C(211)
O(3)-Si(2)-C(211)
C(221)-Si(2)-C(211)
O(11)-C(11)-Os(1)
O(12)-C(12)-Os(1)
O(13)-C(13)-Os(1)
O(21)-C(21)-Os(2)
O(22)-C(22)-Os(2)
O(23)-C(23)-Os(2)
O(31)-C(31)-Os(3)
O(32)-C(32)-Os(3)
O(33)-C(33)-Os(3)
O(34)-C(34)-Os(3)
Si(1)-O(1)-Os(1)
Si(1)-O(1)-Os(2)
Os(1)-O(1)-Os(2)
Si(2)-O(2)-Si(1)
C(112)-C(111)-Si(1)
C(116)-C(111)-Si(1)
C(126)-C(121)-Si(1)
C(122)-C(121)-Si(1)
C(216)-C(211)-Si(2)
C(212)-C(211)-Si(2)
C(226)-C(221)-Si(2)
C(222)-C(221)-Si(2)
Si(2)-O(2)
Si(2)-O(3)
Si(2)-C(221)
Si(2)-C(211)
C(11)-Os(1)-C(12)
C(11)-Os(1)-C(13)
C(12)-Os(1)-C(13)
C(11)-Os(1)-O(1)
C(12)-Os(1)-O(1)
C(13)-Os(1)-O(1)
C(11)-Os(1)-Os(2)
C(12)-Os(1)-Os(2)
C(13)-Os(1)-Os(2)
O(1)-Os(1)-Os(2)
C(11)-Os(1)-Os(3)
C(12)-Os(1)-Os(3)
C(13)-Os(1)-Os(3)
O(1)-Os(1)-Os(3)
Os(2)-Os(1)-Os(3)
C(21)-Os(2)-C(22)
C(21)-Os(2)-C(23)
C(22)-Os(2)-C(23)
C(21)-Os(2)-O(1)
C(22)-Os(2)-O(1)
C(23)-Os(2)-O(1)
C(21)-Os(2)-Os(1)
89.5(2)
91.6(2)
96.2(2)
179.7(5)
179.9(6)
172.9(5)
178.8(4)
177.4(5)
176.3(5)
178.5(5)
177.6(5)
176.5(6)
176.4(6)
125.87(16)
130.60(16)
82.65(10)
177.6(5)
172.8(4)
178.1(4)
178.8(5)
173.8(5)
179.9(6)
177.3(4)
177.4(6)
168.29(18)
98.30(16)
96.18(16)
120.07(17)
136.58(15)
112.47(15)
48.65(7)
89.38(17)
91.94(17)
171.82(15)
81.69(7)
176.9(5)
137.44(15)
135.86(16)
82.18(9)
143.33(19)
123.5(3)
120.1(3)
123.9(3)
119.8(3)
121.4(3)
60.278(7)
88.8(2)
92.4(2)
123.1(4)
120.6(4)
124.2(4)
120.1(4)
96.8(2)
97.1(2)
168.32(16)
98.43(15)
96.86(16)
119.51(14)
168.94(17)
99.08(16)
94.33(16)
120.43(15)
121.3(3)
121.0(4)
121.7(3)
of the organic silicon skeleton (either Os-O-Si or Si-
O-Si or SiOH bonds). Therefore, from the X-ray evi-
dence, it appears that intramolecular interactions of the
free silanol group with the “Os3(CO)10(µ-H)” core are not
relevant, at least in these new molecular models.
instance it seems possible by this new synthetic meth-
odology to achieve the bonding of more than one osmium
carbonyl cluster to more sophisticated molecular models
of the silica surface carrying many vicinal or geminal
silanol groups.19,20c
From the chemical behavior and from the X-ray
structure of two new molecular models of the silica-
anchored [Os3(CO)10(µ-H)(µ-OSit)] species, with ligands
mimicking geminal and vicinal surface silanols, we have
reached some evidence that the free silanol groups are
rather unreactive toward another [Os3(CO)10(µ-H)(µ-
OH)] moiety, despite that there is no evidence of
perturbation (for instance by intramolecular hydrogen
bonding) of the free silanol groups by the already bonded
“Os3(CO)10(µ-H)” core. This is a new observation, which
Con clu sion
The high-yield synthetic method, described in this
paper for making Os-OSi bonds in [Os3(CO)10(µ-H)(µ-
OSiR2R′)] clusters, not only suggests a new way of
strong interaction of slightly oxidized metal particles
with the silica surface but also may open an easy route
for the preparation of new molecular models of osmium
carbonyl clusters grafted to surface silanol groups. For

Synthesis of [Os3(CO)10(µ-H)(µ-OSiR2R′)] Species
Organometallics, Vol. 19, No. 6, 2000 1057
syntheses for the various [Os3(CO)10(µ-H)(µ-OSiR2R′)] clusters
are described below.
can be of relevance to the understanding of the process
of dispersion of small particles or metal clusters on the
silica surface, although it should be confirmed by more
sophisticated models of the silica surface carrying many
vicinal or geminal silanol groups.19,20c
[Os3(CO)10(µ-H)(µ-OSiP h 3)]. A solution of [Os3(CO)10(µ-H)-
(µ-OH)]25 (101 mg; 0.116 mmol) and Ph3SiOH (66.0 mg; 0.238
mmol; molar ratio Ph3SiOH:[Os3(CO)10(µ-H)(µ-OH)] ) 2:1) in
anhydrous m-xylene (1 mL; [Os3(CO)10(µ-H)(µ-OH)] ) 1 × 10-1
M) was heated (138 °C) under N2 in a small flask (5 mL)
equipped with a closed Pyrex filter funnel (diameter ) ca. 30
mm; height ) ca. 250 mm) with a fritted disk containing P2O5
(2 g; the distance between the reaction solution and the fritted
disk is ca. 120 mm). After 5 h, the formation of [Os3(CO)10(µ-
H)(µ-OSiPh3)] was quantitative, as shown by 1H NMR spec-
troscopy. The solution was evaporated to dryness, pentane was
added to dissolve [Os3(CO)10(µ-H)(µ-OSiPh3)], and the resulting
suspension was filtered in order to remove the excess Ph3SiOH.
The filtrate was evaporated to dryness, and the residue was
chromatographed on a silica column using as eluant pentane/
CH2Cl2 (10:1), to obtain pure [Os3(CO)10(µ-H)(µ-OSiPh3)] (111
mg; 0.098 mmol; 84% yield). The product was characterized
by elemental analysis (calcd C, 29.82; H, 1.42; found C, 30.16;
H, 1.42), mass spectrometry (in the EI mass spectrum, there
is the molecular ion peak at m/e ) 1128 [M]+, followed by peaks
corresponding to the successive loss of 10 CO’s), infrared
spectroscopy (in CH2Cl2: ν(CO) 2109(w), 2070(vs), 2057(s),
Finally the ready availability of sizable amounts of
these molecular models allows now a detailed investiga-
tion of their reactivity.38 Therefore, since homogeneous
models can be studied more deeply and more easily than
their related surface species, it is now possible to make
use of this kind of molecular model not only for the
elucidation of the structure of surface species, as was
done up to now,2,17,18,19d but also for confirming or better
defining their rather complex surface behavior, which,
up to now, has been suggested from spectroscopic
evidence1-5 or indirectly deduced from mass balance of
gaseous products produced by chemical reaction with
the surface1,2 or by oxidation of surface species39 or even
by analogies with the chemical behavior of known
related metal carbonyl species.1
Exp er im en ta l Section
1
2021(vs), 1999(m), and 1980(w) cm-1), H NMR (in CDCl3: δ
) 7.44 (m, 15 H, 3 Ph) and -11.36 (s, 1 H, HOs) ppm), and
29Si NMR (in CDCl3: δ ) 0.26 ppm; for comparison the 29Si
NMR signal of HOSiPh3 in CDCl3 is at δ ) -12.47 ppm).
[Os3(CO)10(µ-H)(µ-OSiEt3)]. [Os3(CO)10(µ-H)(µ-OH)] (153
mg; 0.176 mmol) was dissolved in pure Et3SiOH (0.25 mL; 1.62
mmol; molar ratio Et3SiOH:[Os3(CO)10(µ-H)(µ-OH)] ) 9:1;
[Os3(CO)10(µ-H)(µ-OH)] ) 6 × 10-1 M) and heated at 138 °C
under N2 in a small flask (5 mL) equipped with a closed Pyrex
filter funnel (diameter ) ca. 30 mm; height ) ca. 250 mm)
with a fritted disk containing P2O5 (2 g; the distance between
the reaction solution and the fritted disk is 120 mm). The
reaction was complete after 2.5 h, as shown by 1H NMR
spectroscopy. The solution was evaporated to dryness, and the
residue was chromatographed on a silica column using as
eluant hexane, to obtain pure [Os3(CO)10(µ-H)(µ-OSiEt3)] (140
mg; 0.143 mmol; 81% yield). The product was characterized
by elemental analysis (calcd C, 19.53; H, 1.63; found C, 19.84;
H, 1.88), mass spectrometry (in the EI mass spectrum, there
is the molecular ion peak at m/e ) 983 [M]+, followed by peaks
corresponding to the successive loss of 10 CO’s), infrared
spectroscopy (in CH2Cl2: ν(CO) 2108(w), 2069(vs), 2056(s),
Gen er a l Com m en ts. [Os3(CO)10(µ-H)(µ-OH)] was prepared
according to the literature.25 Ph3SiOH and Ph2Si(OH)2, pur-
chased from Sigma-Aldrich, were recrystallized (CH2Cl2) before
use. HOSiPh2OSiPh2OH was prepared according to the lit-
erature40 by hydrolysis of ClSiPh2OSiPh2Cl (purchased from
ABCR GmbH & Co), whereas Et3SiOH was purchased from
Sigma-Aldrich and used without further purification. The
organic solvents were dried over molecular sieves (3 Å),
whereas nitrogen was dried by flowing over drierite. Products
were purified by recrystallization or by column chromatogra-
phy using silica gel 60 (purchased from Fluka). Spectra were
obtained by use of the following spectrometers: Bruker-Vector
22 or J asco FT-IR 420 (IR), Bruker AC-200 or Bruker DRX-
300 (1H NMR and 29Si NMR), Varian VG9090 (MS). Elemental
analyses were carried out in the analytical laboratory of our
department.
1. Syn th eses. To reach high yields, we used two strategies
to remove water: (i) use of a Markusson apparatus for
continuous distillation of water, when working with large (50
mL) amounts of solvent and dilute solutions, and (ii) use of a
drying tube containing P2O5, particularly when working with
small (few mL) amounts of solvent. The second method turned
out to be the most convenient also because the best yields were
reached by working with concentrated solutions of [Os3(CO)10-
(µ-H)(µ-OH)] (see Table 1).
1
2019(vs), 1996(m) and 1981(w) cm-1), H NMR (in CDCl3: δ
) 0.99 (t, 6 H, 3 CH2), 0.62 (q, 9 H, 3 CH3), -12.36 (s, 1 H,
HOs) ppm), and 29Si NMR (in CDCl3: δ ) 38.36 ppm; for
comparison, the 29Si NMR signal of HOSiEt3 in CDCl3 is at δ
) 19 ppm).
Because the infrared spectra in the carbonyl region of
[Os3(CO)10(µ-H)(µ-OH)]25 and of [Os3(CO)10(µ-H)(µ-OSiR2R′)]
species (R ) Et, Ph; R′ ) Et, Ph, OH, OSiPh2OH) are very
similar, reactions were best followed by 1H NMR spectros-
copy: samples of the reaction mixtures were either analyzed
[Os3(CO)10(µ-H )(µ-OSiP h 2OSiP h 2OH )]. A solution of
[Os3(CO)10(µ-H)(µ-OH)] (165 mg; 0.190 mmol) and HOSiPh2-
OSiPh2OH (159 mg; 0.384 mmol; molar ratio HOSiPh2OSiPh2-
OH:[Os3(CO)10(µ-H)(µ-OH)] ) 2:1) in anhydrous m-xylene (1.5
mL; [Os3(CO)10(µ-H)(µ-OH)] ) 1.2 × 10-1 M) was heated under
N2 in a small flask (5 mL) equipped with a closed Pyrex filter
funnel (diameter ) ca. 30 mm; height ) ca. 250 mm) with a
fritted disk containing P2O5 (2 g; the distance between the
reaction solution and the fritted disk is ca. 120 mm). The
reaction was complete after 4 h, as shown by 1H NMR
spectroscopy. The solution was evaporated to dryness, and the
residue was chromatographed on a silica column using as
eluant CH2Cl2/pentane (10:3), to obtain pure [Os3(CO)10(µ-H)-
(µ-OSiPh2OSiPh2OH)] (203 mg; 0.160 mmol; 84% yield). Crys-
tals, suitable for X-ray diffraction, of this new cluster were
obtained by slow addition at room temperature of pentane to
its solution in CH2Cl2. [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)]
was also characterized by elemental analysis (calcd C, 32.25;
H, 1.73; found C, 32.51; H, 1.79), mass spectrometry (in the
EI mass spectrum, there is the molecular ion peak at m/e )
1
as such or evaporated to dryness and dissolved in CDCl3. H
NMR yields were obtained by determining the molar ratio from
the H-Os resonance of [Os3(CO)10(µ-H)(µ-OH)] (δ in CDCl3 )
-12.64 ppm)25 and of the reaction product [Os3(CO)10(µ-H)(µ-
OSiR2R′)]. In some cases, for example by working with dilute
solutions of [Os3(CO)10(µ-H)(µ-OH)] (6 × 10-4 M) under not
optimized reaction conditions, [H2Os3(CO)10] (δ in CDCl3
)
-11.77 ppm), [H4Os4(CO)12] (δ in CDCl3 ) -20.45 ppm), and/
or a still unidentified hydride species (δ in CDCl3 ) -15.54
ppm) were also formed (see Results and Discussion). The best
(38) Garegnani, M.; Lucenti, E.; Roberto, D.; Roveda, C.; Ugo, R.
Organometallics, to be submitted.
(39) Dossi, C.; Fusi, A.; Psaro, R.; Zanderighi, G. M. Appl. Catal.
1989, 46, 145.
(40) Burkhard, C. A. J . Am. Chem. Soc. 1945, 67, 2173.

1058 Organometallics, Vol. 19, No. 6, 2000
Lucenti et al.
Ta ble 3. Su m m a r y of Cr ysta l Da ta a n d Str u ctu r e Refin em en t P a r a m eter s for
[Os3(CO)10(µ-H)(µ-OSiP h 2OSiP h 2OH)] a n d [Os3(CO)10(µ-H)(µ-OSiP h 2OH)]
formula
fw
cryst syst
space group
a, Å
C34H21O13Os3Si2
1264.29
monoclinic
P21/a
19.555(1)
9.326(1)
22.079(1)
90.00(1)
111.97(1)
90.00(1)
3734.0(3)
4
C22H11O12Os3Si
1066.00
triclinic
P1h
9.974(1)
10.413(1)
14.305(1)
100.59(1)
104.50(1)
101.51(1)
1366.01(12)
2
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å3
Z
F(000)
2340
2.249
293(2)
962
2.592
293(2)
D(calc), g cm-3
temp, K
diffractometer
radiation (graph monochr), Å
abs coeff, mm-1
cryst size, mm
scan method
frame width, deg
max time per frame, s
θ range, deg
index ranges
SMART
0.71073
10.310
SMART
0.71073
14.020
0.20 × 0.10 × 0.10
0.15 × 0.15 × 0.10
ω
ω
0.3
0.3
20
20
0.99-29.34
1.52-29.23
-25 e h e 25, -12 e k e 12,
-29 e l e 29
56 804
-13 e h e 13, -13 e k e 14,
-19 e l e 19
16 756
reflns collected
ind reflns
cryst decay, %
refinement method
no. of data/restraints/params
goodness-of-fit on Fo
9414 [R(int) ) 0.0283]
6636 [R(int) ) 0.0277]
no
no
2
2
least-squares on Fo
least-squares on Fo
9414/0/473
1.104
1.104
6636/0/347
0.816
0.816
2
2
rest goodness-of-fit on Fo
R indices [Fo > 4σ(Fo)]
R1 0.0242, wR2 0.0569
R1 0.0349, wR2 0.0591
0.896 and -1.553
calc
R1 0.0215, wR2 0.0421
R1 0.0320, wR2 0.0436
0.674 and -1.064
calc
R indices (all data)
largest diff peak and hole, e Å-3
weighting scheme
1266 [M]+, followed by peaks corresponding to the successive
loss of 10 CO’s), infrared spectroscopy (in CH2Cl2: ν(CO)
2110(w), 2072(vs), 2058(s), 2021(vs), 2001(m), and 1981(w)
successive loss of 10 CO’s), infrared spectroscopy (in CH2Cl2:
ν(CO) 2109(w), 2072(vs), 2058(s), 2019(vs), 2002(m), and
1
1979(w) cm-1), H NMR (in CDCl3: δ ) 7.55 (m, 10 H, 2 Ph),
cm-1), H NMR (in CDCl3: δ ) 7.45 (m, 20 H, 4 Ph), 2.78 (s,
2.95 (s, 1 H, OH), -12.50 (s, 1 H, HOs) ppm), and 29Si NMR
(in CDCl3: δ ) -20.72 ppm; for comparison, the 29Si NMR
signal of Ph2Si(OH)2 in CD3COCD3 is at δ ) -33.01 ppm).
At t em p t ed Syn t h esis of [Os3(CO)10(µ-H )(µ-OSiP h 2O-
SiP h 2O-µ)(µ-H)(CO)10Os3]. (i) A solution of [Os3(CO)10(µ-H)-
(µ-OH)] (77.0 mg; 0.089 mmol) and HOSiPh2OSiPh2OH (18.7
mg; 0.045 mmol; molar ratio HOSiPh2OSiPh2OH:[Os3(CO)10-
(µ-H)(µ-OH)] ) 1:2) in anhydrous m-xylene (0.9 mL; [Os3(CO)10-
(µ-H)(µ-OH)] ) 0.99 × 10-1 M) was heated (138 °C) under N2
in a small flask (5 mL) equipped with a closed Pyrex filter
funnel (diameter ) ca. 30 mm; height ) ca. 250 mm) with a
fritted disk containing P2O5 (2 g; the distance between the
reaction solution and the fritted disk is ca. 120 mm). The
reaction was monitored by 1H NMR spectroscopy: after 3 and
11 h, [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] was formed in 25%
and 50% yields, respectively, the remainder being unreacted
[Os3(CO)10(µ-H)(µ-OH)]. (ii) A solution of [Os3(CO)10(µ-H)(µ-
OH)] (27.7 mg; 0.032 mmol) and [Os3(CO)10(µ-H)(µ-OSiPh2-
OSiPh2OH)] (154 mg; 0.122 mmol; molar ratio [Os3(CO)10(µ-
H)(µ-OSiPh2OSiPh2OH)]:[Os3(CO)10(µ-H)(µ-OH)] ) 4:1) in
anhydrous m-xylene (0.3 mL; [Os3(CO)10(µ-H)(µ-OH)] ) 1 ×
10-1 M) was heated (138 °C) for 3 h under N2 in a small flask
(5 mL) equipped with a closed Pyrex filter funnel (diameter )
ca. 30 mm; height ) ca. 250 mm) with a fritted disk containing
P2O5 (2 g; the distance between the reaction solution and the
fritted disk is ca. 120 mm). No reaction occurred, as evidenced
by 1H NMR spectroscopy of the reaction mixture. (iii) A
solution of O(SiPh2OLi)2 in anhydrous benzene (prepared by
reaction of HOSiPh2OSiPh2OH with BuLi;36 [O(SiPh2OLi)2] )
3.7 × 10-2 M; 0.7 mL corresponding to 0.026 mmol of O(SiPh2-
OLi)2) was added to a solution of [Os3(CO)10(µ-H)(µ-O2CCF3)]25
(50.4 mg; 0.052 mmol; molar ratio [Os3(CO)10(µ-H)(µ-O2CCF3)]:
O(SiPh2OLi)2 ) 2:1) in anhydrous benzene (20 mL) and stirred
1
1 H, OH), -12.02 (s, 1 H, HOs) ppm), and 29Si NMR (in
CDCl3: δ ) -28.72 (OsOSiPh2O), -36.41 (OSiPh2OH) ppm;
for comparison, the 29Si NMR signal of HOSiPh2OSiPh2OH in
CDCl3 is at δ ) -36.37 ppm and in CD3COCD3 is at δ )
-39.66 ppm).
[Os3(CO)10(µ-H)(µ-OSiP h 2OH)]. Ph2Si(OH)2 (1.738 g; 8.04
mmol) was added to a solution of [Os3(CO)10(µ-H)(µ-OH)] (115
mg; 0.133 mmol; molar ratio Ph2Si(OH)2:[Os3(CO)10(µ-H)(µ-
OH)] ) 60:1) in anhydrous m-xylene (4.3 mL; [Os3(CO)10(µ-
H)(µ-OH)] ) 3.1 × 10-2 M) in a small flask (5 mL) equipped
with a closed Pyrex filter funnel (diameter ) ca. 30 mm; height
) ca. 250 mm) with a fritted disk containing P2O5 (2 g; the
distance between the reaction mixture and the fritted disk is
ca. 120 mm), and the resulting slurry (due to the huge excess
of Ph2Si(OH)2) was heated under N2 at 138 °C. After about 15
1
min a homogeneous solution was obtained. H NMR spectros-
copy evidenced the formation of [Os3(CO)10(µ-H)(µ-OSiPh2OH)]
in 67% yield. After 50 min, the formation of this new cluster
was quantitative. The solution was evaporated to dryness, and
the residue was chromatographed on a silica column using as
eluant CH2Cl2/hexane (3:2), to obtain pure [Os3(CO)10(µ-H)(µ-
OSiPh2OH)] (112 mg; 0.105 mmol; 79% yield). Interestingly,
an increase of the reaction time to 2.5 h leads to the formation
of some [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] (23% yield) due
to the condensation of [Os3(CO)10(µ-H)(µ-OSiPh2OH)] with
excess Ph2Si(OH)2. Crystals, suitable for X-ray diffraction, of
[Os3(CO)10(µ-H)(µ-OSiPh2OH)] were obtained by slow evapora-
tion at room temperature of its solution in CH2Cl2. This new
cluster was also characterized by elemental analysis (calcd C,
24.75; H, 1.12; found C, 25.05; H, 1.41), mass spectrometry
(in the EI mass spectrum, there is the molecular ion peak at
m/e ) 1068 [M]+, followed by peaks corresponding to the

Synthesis of [Os3(CO)10(µ-H)(µ-OSiR2R′)] Species
Organometallics, Vol. 19, No. 6, 2000 1059
under N2. After 12 h at room temperature or after 6 h under
reflux (75 °C), no reaction occurred, as shown by infrared
spectroscopy in the carbonyl region where are present only the
bands due to unreacted [Os3(CO)10(µ-H)(µ-O2CCF3)].
Rσ ) 0.0431).42 The structures were solved by direct methods
(SIR97)43 and refined with full-matrix-block least-squares
(SHELX93);44 anisotropic temperature factors were assigned
to all non-hydrogen atoms. All phenyl hydrogens were riding
on their carbon atoms with individual isotropic displacement
parameters 1.2 times that of the pertinent carbon atom. The
hydridic ligand has been located on the Fourier map and
refined in both structures.
2. X-r a y Str u ctu r e Deter m in a tion of [Os3(CO)10(µ-H)-
(µ-OSiP h 2OSiP h 2OH)] an d [Os3(CO)10(µ-H)(µ-OSiP h 2OH)].
A suitable crystal of each compound was chosen and mounted,
in air, on a glass fiber onto a goniometer head and collected
at room temperature on a Siemens SMART CCD area-detector
diffractometer. Crystal data are reported in Table 3. Graphite-
monochromatized Mo KR (λ ) 0.71073 Å) radiation was used
with the generator working at 45 kV and 40 mA. Cell
parameters and orientation matrix were obtained from least-
squares refinement on reflections measured in three different
sets of 15 frames each, in the range 0° < θ < 23°. The data
collection was performed by measuring 2100 frames (20 s per
frame; ω scan method, ∆ω ) 0.3°; sample-detector distance
fixed at 5 cm), which, upon data reduction, afforded almost
all reflections belonging to the sphere with 2θ < 58°. The first
100 frames were recollected at the end to monitor crystal
decay, which was not observed; an absorption correction was
applied (SADABS).41
Ack n ow led gm en t . We deeply thank Pasquale
Illiano and Americo Costantino, for running 1H, 13C, and
29Si NMR spectra, and Prof. Tiziana Beringhelli (Uni-
versita` di Milano) for fruitful discussions. This work was
supported by the Ministero dell’Universita` e della
Ricerca Scientifica e Tecnologica and by the Consiglio
Nazionale delle Ricerche (CNR, Roma).
Su p p or tin g In for m a tion Ava ila ble: Refined atoms co-
ordinates (Tables S1 and S4), anisotropic thermal parameters
(Tables S2 and S5), idealized hydrogen atom coordinates
(Tables S3 and S6). This material is available free of charge
via the Internet at http://pubs.acs.org.
OM9908074
For [Os3(CO)10(µ-H)(µ-OSiPh2OSiPh2OH)] a total of 56 804
reflections were collected (9414 unique, Rint ) 0.0283; Rσ
)
2
2
(42) Rint ) ∑|Fo - Fmean |/∑Fo2; Rσ ) ∑σ(Fo2)/∑Fo2; R1 ) ∑||Fo| -
0.0236),42 while for [Os3(CO)10(µ-H)(µ-OSiPh2OH)] a total of
16 756 reflections were collected (6636 unique, Rint ) 0.0277;
|Fc||/∑|Fo|; wR2 ) (∑(Fo - Fc2)2/∑wFo4).1/2
2
(43) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.;
Burla, M. C.; Polidori, G.; Camalli, M. J . Appl. Crystallogr. 1994, 24,
435.
(41) Sheldrick, G. M. SADABS, University of Go¨ttingen: Germany,
1996.
(44) Sheldrick, G. M. SHELX93-Program for the refinement of crystal
structure; University of Go¨ttingen: Germany, 1993.