Tris(tert-butoxy)siloxy derivatives of boron, including the boronous acid HOB[OSi(OtBu)3]2 and the metal (siloxy)boryloxide complex Cp2Zr(Me)OB[OSi(OtBu)3]2: A remarkable crystal structure with 18 independent molecules in its asymmetric unit

Article abstract of DOI:10.1021/ic0205482

Silanolysis of B(O^tBu)₃ with 2 and 3 equiv of HOSi(O^tBu)₃ led to the formation of BuOB[OSi(O^tBu)₃]₂ (1) and B[OSi(O^tBu)₃]₃ (2), respectively. Compounds 1 and 2 are efficient single-source molecular precursors to B/Si/O materials via thermolytic routes in nonpolar media, as demonstrated by the generation of BO_1.5·2SiO₂ (BOSi2_xg and BO_1.5·3SiO₂ (BOSi3_xg) xerogels, respectively. Use of a block copolymer template provided B/Si/O materials (BOSi2_epe and BOSi3_epe) with a broad distribution of mesopores (by N₂ porosimetry) and smaller, more uniform particle sizes (by TEM) as compared to the nontemplated materials. Hydrolyses of 1 and 2 with excess H₂O resulted in formation of the expected amounts of ^tBuOH and HOSi(O^tBu)₃; however, reaction of 1 with 1 equiv of H₂O led to isolation of the new boronous acid HOB[OSi(O^tBu)₃]₂ (3). This ligand precursor is well suited for the synthesis of new metal (siloxy)boryloxide complexes via proton-transfer reactions involving the BOH group. The reaction of 3 with Cp₂ZrMe₂ resulted in formation of Cp₂Zr(Me)OB[OSi(O^tBu)₃]₂ (4) in high yield. This rare example of a transition metal boryloxide complex crystallizes in the triclinic space group P1 and exhibits a crystal structure with an unprecedented number of independent molecules in its asymmetric unit (i.e., Z = 18 and Z = 36). This unusual crystal structure presented an opportunity to perform statistical analyses of the metric parameters for the 18 crystallographically independent molecules. Complex 4 readily converts to Cp₂Zr[OSi(O^tBu)₃]₂ (5) upon thermolysis or upon dissolution in Et₂O at room temperature.

Full text of DOI:10.1021/ic0205482

Inorg. Chem. 2003, 42, 1140−1150
Tris(tert-butoxy)siloxy Derivatives of Boron, Including the Boronous
t
Acid HOB[OSi(O Bu)3]2 and the Metal (Siloxy)boryloxide Complex
t
Cp2Zr(Me)OB[OSi(O Bu)3]2: A Remarkable Crystal Structure with 18
Independent Molecules in Its Asymmetric Unit
,‡
,‡
,†
Kyle L. Fujdala,† Allen G. Oliver,* Frederick J. Hollander,* and T. Don Tilley*
Department of Chemistry, UniVersity of California, Berkeley, Berkeley, California 94720-1460,
and Chemical Sciences DiVision, Lawrence Berkeley National Laboratory, 1 Cyclotron Road,
Berkeley, California 94720-1460
Received September 10, 2002
t
t
t
t
Silanolysis of B(OBu)3 with 2 and 3 equiv of HOSi(O Bu)3 led to the formation of BuOB[OSi(OBu)3]2 (1) and
t
B[OSi(OBu)3]3 (2), respectively. Compounds 1 and 2 are efficient single-source molecular precursors to B/Si/O
materials via thermolytic routes in nonpolar media, as demonstrated by the generation of BO ‚2SiO (BOSi2 )
1.5
2
xg
and BO ‚3SiO (BOSi3 ) xerogels, respectively. Use of a block copolymer template provided B/Si/O materials
1.5
2
xg
(BOSi2epe and BOSi3epe) with a broad distribution of mesopores (by N porosimetry) and smaller, more uniform
2
particle sizes (by TEM) as compared to the nontemplated materials. Hydrolyses of 1 and 2 with excess H O
2
t
resulted in formation of the expected amounts of tBuOH and HOSi(OBu)3; however, reaction of 1 with 1 equiv of
t
H O led to isolation of the new boronous acid HOB[OSi(OBu)3]2 (3). This ligand precursor is well suited for the
2
synthesis of new metal (siloxy)boryloxide complexes via proton-transfer reactions involving the BOH group. The
t
reaction of 3 with Cp2ZrMe2 resulted in formation of Cp2Zr(Me)OB[OSi(OBu)3]2 (4) in high yield. This rare example
of a transition metal boryloxide complex crystallizes in the triclinic space group P1h and exhibits a crystal structure
with an unprecedented number of independent molecules in its asymmetric unit (i.e., Z′ ) 18 and Z ) 36). This
unusual crystal structure presented an opportunity to perform statistical analyses of the metric parameters for the
t
18 crystallographically independent molecules. Complex 4 readily converts to Cp2Zr[OSi(OBu)3]2 (5) upon thermolysis
or upon dissolution in Et2O at room temperature.
Introduction
process involves the hydrolysis and condensation of one or
more precursors, often alkoxides, in polar media (e.g., H2O
or ROH). Due to large differences in the hydrolysis rates of
individual precursors, undesirable heterogeneous (poorly
mixed) materials may form in sol-gel processes when
multicomponent oxides are desired.2 Although some degree
of control is possible through the careful modification of each
precursor, alternative methods that do not require added H2O
or polar media are useful for the preparation of highly
homogeneous oxide materials.3
The design and synthesis of advanced materials with
tailored properties often requires low-temperature methods
to permit the formation of metastable structures.1 A widely
used low-temperature approach for the synthesis of high
surface area oxide materials is the sol-gel method.2 This
* To whom corrrespondence should be addressed. E-mail: tdtilley@
socrates.berkeley.edu (T.D.T.).
† University of California and Lawrence Berkeley National Laboratory.
‡ University of California College of Chemistry X-ray Crystallographic
Facility (CHEXRAY).
(1) (a) Ultrastructure Processing of AdVanced Materials; Uhlmann, D.
R., Ulrich, D. R., Eds.; Wiley-Interscience: New York, 1992. (b)
Inorganic Materials; Bruce, D. W., O’Hare, D., Eds.; Wiley: New
York, 1992. (c) Ceramic Precursor Technology and Its Applications;
Narula, C. K., Ed.; Marcel Dekker Inc.: New York, 1995. (d)
Mehrotra, R. C. J. Non-Cryst. Solids 1988, 100, 1. (e) Stein, A.; Keller,
S. W.; Mallouk, T. E. Science 1993, 259, 1558. (f) Amabilino, D. B.;
Stoddart, J. F. Chem. ReV. 1995, 95, 2725. (g) Bowes, C. L.; Ozin, G.
A. AdV. Mater. 1996, 8, 13.
(2) (a) Brinker, C. J.; Scherer, G. W. Sol-Gel Science; Academic Press:
Boston, MA, 1990. (b) Sol-Gel Technology for Thin Films, Fibers,
Preforms, Electronics, and Specialty Shapes; Klein, L. C., Ed.;
Noyes: Park Ridge, NJ, 1988. (c) Brinker, C. J. J. Non-Cryst. Solids
1988, 100, 31. (d) Schmidt, H. J. Non-Cryst. Solids 1988, 100, 51.
(d) Schubert, U.; Hu¨sing, N.; Lorenz, A. Chem. Mater. 1995, 7, 2010.
(e) Corriu, R. J. P.; Leclercq, D. Angew. Chem., Int. Ed. Engl. 1996,
35, 1421. (f) Schubert, U. J. Chem. Soc., Dalton Trans. 1996, 3343.
1140 Inorganic Chemistry, Vol. 42, No. 4, 2003
10.1021/ic0205482 CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/24/2003

Tris(tert-butoxy)siloxy DeriWatiWes of Boron
In this context, we have been investigating the structure
and reactivity of complexes containing OSi(OtBu)3 and
O2P(OtBu)2 ligands of the general formula LnM[OSi-
a single ligand precursor containing two or more hetero-
elements.
The aim of the current study was to synthesize new
molecules containing boron and the -OSi(OtBu)3 ligand for
use as single-source molecular precursors to borosilicate
materials. In addition, the development of a general route
for the formation of metal (siloxy)boryloxide species (with
M-O-B-OSi(OtBu)3 linkages) was desired. In 1971 Abe
et al. reported that silanolysis of B(OnBu)3 with HOSi-
(OtBu)3 produced (nBuO)2BOSi(OtBu)3 and (nBuO)B[OSi-
(OtBu)3]2 as viscous oils.8 To provide clean conversions to
materials at low temperatures (via CH2CMe2 elimination),
we targeted species containing only tBuO groups. This work
4
5
(OtBu)3]m and LnM[O2P(OtBu)2]m (where Ln ) alkoxide,
amide, alkyl, etc.). These oxygen-rich molecules are excellent
single-source precursors6 to homogeneous mixed-element
M/Si/O and M/P/O oxide materials due to their clean thermal
decompositions (in organic solvent or in the solid state) at
low temperatures (<200 °C) via the elimination of isobutene
and H2O. In addition, these complexes serve as soluble
molecular models for heterogeneous catalysts based on oxide-
supported metal species.4m
With the thermolytic molecular precursor route firmly
established as an efficient method for the preparation of
homogeneous, bicomponent oxide materials, we have sought
to extend this approach to homogeneous tri- and tetra-
component oxide materials. In this context, we recently
reported the first complex containing both -OSi(OtBu)3 and
-O2P(OtBu)2 ligands, [(tBuO)3SiO]2Al[(µ-O)2P(OtBu)2]2Al-
(Me)OSi(OtBu)3, and demonstrated its use as an efficient
single-source molecular precursor to high surface area
silicoaluminophosphate materials.7 The development of
general routes to complexes with three or more hetero-
elements (e.g., Al, P, and Si) appropriate for use in the
thermolytic molecular precursor method should lead to new
homogeneous oxide materials with unique properties. One
approach for the synthesis of such complexes would involve
t
has provided two new crystalline compounds, BuOB[OSi-
(OtBu)3]2 (1) and B[OSi(OtBu)3]3 (2), that are efficient single-
source molecular precursors to borosilicate materials via
solution thermolyses.
The Power,9,10 Chisholm,11 Gibson,12 and Serwatowski13
groups have synthesized metal diarylboryloxide complexes
containing [OBAr2]- ligands (where Ar ) 2,4,6-Me3C6H3,9-13
2,4,6-iPr3C6H2,10 and 2,4,6-(CF3)3C6H211) using HOBAr2 or
LiOBAr2 reagents. Additionally, complexes of Zn and Cd
are known for the 9-BBN-9-O ligand (BBN ) borabicyclo-
[3.3.1]nonane).14 Thus far, these OBAr2 and 9-BBN-9-O
groups are the only boryloxide ligands that have been
observed to support transition metal boryloxide com-
plexes.9-12,14 Simple (RO)2BOH compounds do not appear
to be stable and have only been observed as components of
product mixtures derived from hydrolysis of orthoborates
B(OR)3.15 To obtain a (siloxy)boryloxide ligand pre-
cursor containing a BOH linkage, hydrolysis of 1 with 1
equiv of H2O was investigated, and this led to isolation of
HOB[OSi(OtBu)3]2 (3). This boronous acid is, to our
knowledge, the first isolable compound containing a
HOB(OSi)2 moiety. Its reaction with Cp2ZrMe2 dem-
onstrates the use of this compound in the synthesis of
the first boryloxide complex with a MOB(OR)2 core,
Cp2Zr(Me)OB[OSi(OtBu)3]2 (4).
(3) (a) Jansen, M.; Guenther, E. Chem. Mater. 1995, 7, 2110. (b) Vioux,
A. Chem. Mater. 1997, 9, 2292.
(4) (a) McMullen, A. K.; Tilley, T. D.; Rheingold, A. L.; Geib, S. J. Inorg.
Chem. 1989, 28, 3772. (b) McMullen, A. K.; Tilley, T. D.; Rheingold,
A. L.; Geib, S. J. Inorg. Chem. 1990, 29, 2228. (c) Terry, K. W.;
Tilley, T. D. Chem. Mater. 1991, 3, 1001. (d) Terry, K. W.; Gantzel,
P. K.; Tilley, T. D. Chem. Mater. 1992, 4, 1290. (e) Terry, K. W.;
Gantzel, P. K.; Tilley, T. D. Inorg. Chem. 1993, 32, 5402. (f) Terry,
K. W.; Lugmair, C. G.; Gantzel, P. K.; Tilley, T. D. Chem. Mater.
1996, 8, 274. (g) Su, K.; Tilley, T. D.; Sailor, M. J. J. Am. Chem.
Soc. 1996, 118, 3459. (h) Su, K.; Tilley, T. D. Chem. Mater. 1997, 9,
588. (i) Terry, K. W.; Lugmair, C. G.; Tilley, T. D. J. Am. Chem.
Soc. 1997, 119, 9745. (j) Rulkens, R.; Tilley, T. D. J. Am. Chem.
Soc. 1998, 120, 9959. (k) Lugmair, C. G.; Tilley, T. D. Inorg. Chem.
1998, 37, 764. (l) Terry, K. W.; Su, K.; Tilley, T. D.; Rheingold, A.
L. Polyhedron 1998, 17, 891. (m) Rulkens, R.; Male, J. L.; Terry, K.
W.; Olthof, B.; Khodakov, A.; Bell, A. T.; Iglesia, E.; Tilley, T. D.
Chem. Mater. 1999, 11, 2966. (n) Coles, M. P.; Lugmair, C. G.; Terry,
K. W.; Tilley, T. D. Chem. Mater. 2000, 12, 122. (o) Kriesel, J. W.;
Sander, M. S.; Tilley, T. D. Mater. Chem. 2001, 13, 3554. (p) Kriesel,
J. W.; Sander, M. S.; Tilley, T. D. AdV. Mater. 2001, 13, 331.
(q) Kriesel, J. W.; Tilley, T. D. J. Mater. Chem. 2001, 11, 1081.
(r) Fujdala, K. L.; Tilley, T. D. Chem. Mater. 2001, 13, 1817.
(s) Lugmair, C. G.; Fujdala, K. L.; Tilley, T. D. Chem. Mater.
2002, 14, 888. (t) Fujdala, K. L.; Tilley, T. D. Chem. Mater. 2002,
14, 1376.
(5) (a) Lugmair, C. G.; Tilley, T. D.; Rheingold, A. L. Chem. Mater. 1997,
9, 339. (b) Lugmair, C. G.; Tilley, T. D. Inorg. Chem. 1998, 37, 1821.
(c) Lugmair, C. G.; Tilley, T. D. Inorg. Chem. 1998, 37, 6304. (d)
Lugmair, C. G.; Tilley, T. D.; Rheingold, A. L. Chem. Mater. 1994,
6, 1615.
(6) Selected references on the single-source precursor approach: (a)
Hubert-Pfalzgraf, L. G. New J. Chem. 1987, 11, 663. (b) Cowley, A.
H.; Jones, R. A. Angew. Chem., Int. Ed. Engl. 1989, 28, 1208. (c)
Chaput, F.; Lecomte, A.; Dauger, A.; Boilot, J. P. Chem. Mater. 1989,
1, 167, 199. (d) Mehrotra, R. C. J. Non-Cryst. Solids 1990, 121, 1.
(e) Apblett, A. W.; Warren, A. C.; Barron, A. R. Chem. Mater. 1992,
4, 167. (f) Chandler, C. D.; Roger, C.; Hampden-Smith, M. J. Chem.
ReV. 1993, 93, 1205. (g) Bradley, D. C. Polyhedron 1994, 13, 1111.
(h) Narula, C. K.; Varshney, A.; Riaz, U. Chem. Vap. Deposition 1996,
2, 13. (i) Altherr, A.; Wolfganger, H.; Veith, M. Chem. Vap. Deposition
1999, 5, 87.
Results and Discussion
Synthesis of Boron Tris(tert-butoxy)siloxide Derivatives.
The reaction of 2 equiv of HOSi(OtBu)3 with B(OtBu)3 as a
t
neat mixture at 80 °C led to the formation of BuOB[OSi-
(OtBu)3]2 (1) (eq 1). Analytically pure colorless crystals of
(7) Fujdala, K. L.; Tilley, T. D. J. Am. Chem. Soc. 2001, 123, 10133.
(8) Kijima, I.; Yamamoto, T.; Abe, Y. Bull. Chem. Soc. Jpn. 1971, 44,
3193.
(9) Weese, K. J.; Bartlett, R. A.; Murray, B. D.; Olmstead, M. M.; Power,
P. P. Inorg. Chem. 1987, 26, 2409.
(10) Chen, H.; Power, P. P.; Shoner, S. C. Inorg. Chem. 1990, 30, 2884.
(11) Chisholm, M. H.; Folting, K.; Haubrich, S. T.; Martin, J. D. Inorg.
Chim. Acta 1993, 213, 17.
(12) (a) Gibson, V. C.; Redshaw, C.; Clegg, W.; Elsegood, M. R. J.
Polyhedron 1997, 16, 2637. (b) Gibson, V. C.; Mastroianni, S.; White,
A. J. P.; Williams, D. J. Inorg. Chem. 2001, 40, 826.
(13) Anulewicz-Ostrowska, R.; Lulinski, S.; Serwatowski, J.; Suwinska,
K. Inorg. Chem. 2000, 39, 5763.
(14) Lulinski, S.; Madura, I.; Serwatowski, J.; Zachara, J. Inorg. Chem.
1999, 38, 4937.
(15) Steinfeldt, K.-P.; Heller, G. Z. Naturforsch., B: Chem. Sci. 1981, 36B,
691.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1141

Fujdala et al.
Table 1. Crystallographic Data for 1 and 4
tBuOB[OSi(OtBu)3]2 Cp2Zr(Me)OB[OSi(OtBu)3]2
(1)
(4)
formula
cryst size (mm)
morphology
cryst system
space group
Z, Z′
C28H63BO9Si2
C35H67BO9Si2Zr
0.50 × 0.40 × 0.20
tablet
triclinic
P1h
36, 18
0.30 × 0.19 × 0.08
plate
triclinic
P1h
2, 1
cell consts
a (Å)
b (Å)
9.4639(3)
11.8965(2)
17.8225(7)
93.089(2)
101.872(1)
101.340(2)
1916.0(2)
1.059
33.0365(5)
34.9417(6)
40.9629(6)
103.134(1)
92.809(1)
118.206(1)
39900(1)
1.184
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
D
calc (g cm-3
)
F000
672
1.3
15 192
4.0
Mo KR (0.710 69)
46.7
ω (0.3°/frame)
-118
40 s/frame
185 691
112 301
7648/0
Figure 1. Thermal ellipsoid plot of 1 at the 50% probability level. The
methyl groups have been omitted for clarity.
µ(Mo KR) (cm-1
radiatn (λ, Å)
2θmax (deg)
)
Mo KR (0.710 69)
46.0
ω (0.3°/frame)
-114
30 s/frame
8698
5153
351/0
0.962/0.333
0.047
1 were isolated in 79% yield from a mixture of toluene and
acetonitrile (4:1) at -30 °C.
scan type
temp (°C)
exposure time
measd reflcns
unique reflcns
params/restr
80 °C
B(OtBu)3 + 2HOSi(OtBu)3
8
( BuO)B[OSi(O Bu)3]2
t
t
-2tBuOH
1
(1)
T
max/Tmin
0.862/0.603
0.061
Rint
A single-crystal X-ray crystallographic analysis revealed
that molecules of 1 have a monomeric structure in the solid
state (Figure 1). Compound 1 crystallizes in the triclinic space
group P1h with two molecules in the unit cell (Z ) 2) and
therefore one independent molecule in the asymmetric unit
(Z′ ) 1). A summary of the crystallographic data is provided
in Table 1, and selected bond distances and angles are listed
R
obs/Rall
0.083/0.136
0.212/0.248
1.127/0.983
0.44/-0.37
0.067/0.163
0.154/0.204
0.943/0.943
1.13/-0.79
wR2obs/wR2all
GOFobs/GOFall
max/min (e Å-3
)
linkages are quite flexible and reported values for such angles
range from ca. 129 to 161°.18 The O-Si-O-B-O-Si-O
linkage in 1 makes this compound an interesting molecular
model for borosilicate materials, given that it is the first
crystallographically characterized compound with such a
linkage.
Attempts to prepare 1 in the presence of organic solvents
(toluene, benzene, Et2O) were met with reduced yields and
the formation of significant amounts of a liquid product.
Isolation of this new product in pure form was not achieved;
t
in Table 2. Some of the methyl carbon atoms of the Bu
groups associated with the OSi(OtBu)3 ligands exhibited
rotational disorder, and these were modeled over two
positions with 50% occupancy and isotropic thermal param-
eters. All nondisordered C atoms and all O atoms were
modeled using anisotropic thermal parameters. The boron
atom is in a trigonal planar environment with the sum of
the angles about the boron atom being 360.0(6)°. The
B-O(C) (1.361(7) Å) and B-O(Si) (1.349(7) and 1.374(7)
Å) bond distances are not significantly different from normal
B-O distances (1.36 Å)16 and are similar to the average
B-O bond distances in B(OSnPh3)3 (1.368(6) Å),17
B(OSiPh3)3 (1.357(4) Å),18 PhB(OGePh3)2 (1.342(9) Å),18
and HOBMes2 (1.367(6) Å).9 For further comparison, the
structures of salen-supported borates containing phenyl-
siloxides have been recently reported by Atwood and co-
workers.19 These borates exhibit B-O(Si) distances (from
1.411(6) to 1.481(4) Å) that are significantly longer than
those observed for the 1. Presumably due to the steric strain
induced by the bulky OSi(OtBu)3 ligands in compound 1,
the Si-O-B angles (141.8(4) and 136.7(4)°) are larger than
the C-O-B angle (127.0(4)°). It is known that B-O-Si
1
however, H NMR spectra of reaction mixtures suggested
that it was (tBuO)2BOSi(OtBu)3. Reactions of 1 equiv of
HOSi(OtBu)3 with B(OtBu)3 as a neat mixture or in solution
also did not result in the clean isolation of this compound.
The reaction of 3 equiv of HOSi(OtBu)3with B(OtBu)3 in
toluene at 80 °C resulted in the formation of B[OSi(OtBu)3]3
(2) (eq 2). Analytically pure, colorless crystals were obtained
in 61% overall yield by crystallization from a mixture of
toluene and acetonitrile (4:1) at -30 °C. Despite many
crystallization attempts, crystals of 2 were consistently of
poor quality and not suitable for an X-ray crystallographic
analysis. Given that 1 and the B(OEPh3)3 compounds17,18 are
monomeric in the solid state it is likely that 2 is also
monomeric.
B(OtBu)3 + 3HOSi(OtBu)3 toluene, 80 °C8
(2)
(16) Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements; Perga-
mon: Oxford, England, 1984.
(17) Ferguson, G.; Spalding, T. R.; O′Dowd, A. T. Acta Crystallogr., Sect.
C 1995, 51, 67.
t
B[OSi(O Bu)3]3
-3tBuOH
2
(18) Murphy, D.; Sheehan, J. P.; Spalding, T. R.; Ferguson, G.; Lough, A.
J.; Gallagher, J. F. J. Mater. Chem. 1993, 3, 1275.
(19) Wei, P.; Keizer, T.; Atwood, D. A. Inorg. Chem. 1999, 38, 3914.
Hydrolyses of 1 and 2. The hydrolysis chemistry of 1
and 2 was investigated to characterize the utility of these
1142 Inorganic Chemistry, Vol. 42, No. 4, 2003

Tris(tert-butoxy)siloxy DeriWatiWes of Boron
Table 2. Selected Bond Distances (Å) and Angles (deg) for 1
Distances
B(1)-O(1)
B(1)-O(2)
B(1)-O(3)
Si(1)-O(1)
Si(1)-O(12)
Si(1)-O(13)
1.349(7)
1.374(7)
1.361(7)
1.610(4)
1.601(3)
1.590(4)
Si(1)-O(14)
Si(2)-O(2)
Si(2)-O(15)
Si(2)-O(16)
Si(2)-O(17)
O(3)-C(1)
1.592(4)
1.610(4)
1.605(4)
1.601(4)
1.594(4)
1.438(6)
O(12)-C(2)
O(13)-C(3)
O(14)-C(4)
O(15)-C(5)
O(16)-C(6)
O(17)-C(7)
1.457(6)
1.436(6)
1.413(6)
1.410(6)
1.383(6)
1.407(6)
Angles
O(1)-B(1)-O(3)
O(1)-B(1)-O(2)
O(3)-B(1)-O(2)
O(13)-Si(1)-O(14)
O(13)-Si(1)-O(12)
O(14)-Si(1)-O(12)
O(13)-Si(1)-O(1)
O(14)-Si(1)-O(1)
123.4(6)
O(12)-Si(1)-O(1)
O(17)-Si(2)-O(16)
O(17)-Si(2)-O(15)
O(16)-Si(2)-O(15)
O(17)-Si(2)-O(2)
O(16)-Si(2)-O(2)
O(15)-Si(2)-O(2)
B(1)-O(1)-Si(1)
111.6(2)
B(1)-O(2)-Si(2)
B(1)-O(3)-C(1)
Si(1)-O(12)-C(2)
Si(1)-O(13)-C(3)
Si(1)-O(14)-C(4)
Si(2)-O(15)-C(5)
Si(2)-O(16)-C(6)
Si(2)-O(17)-C(7)
136.7(4)
127.0(4)
134.3(3)
136.3(3)
139.8(3)
136.3(4)
139.9(4)
140.4(4)
120.0(5)
116.6(5)
109.3(2)
108.5(2)
110.4(2)
107.5(2)
109.5(2)
110.3(2)
108.9(2)
110.1(2)
110.4(2)
110.0(2)
107.1(2)
141.8(4)
compounds as potential sol-gel precursors to borosilicate
materials.20 The reaction of 3 equiv of H2O with 1 led to the
rapid formation of tBuOH and HOSi(OtBu)3 (1 and 2 equiv,
respectively), and the reaction of 3 equiv of H2O with 2 led
to the formation of 3 equiv of HOSi(OtBu)3. The formation
of only HOSi(OtBu)3 (i.e., no tBuOH) from hydrolysis of 2
and the formation of only 1 equiv of tBuOH from hydrolysis
of 1 (due to cleavage of the B-OtBu linkage) indicate that
the B-O-Si linkages are much more susceptible to hy-
drolysis than are the Si-OtBu linkages. This indicates that
the use of 1 or 2 in sol-gel procedures would lead to the
preferential formation of B2O3 networks, with little initial
incorporation of Si, thus leading to a heterogeneous material.
This is not surprising given the well-known fact that B-O-
Si linkages are unstable under typical sol-gel conditions,
giving rise to heterogeneous B2O3/SiO2 materials.20
Thermal Stability of 1 and 2. Thermogravimetric analysis
(TGA) and differential scanning calorimetry (DSC) studies
of 1 and 2 revealed that these species exhibit a melting
transition (at ca. 70 and 160 °C for 1 and 2, respectively)
and subsequently boil (at ca. 170 and 255 °C for 1 and 2,
respectively) without thermal decomposition. Although boil-
ing behavior has not been observed for metal complexes
containing the OSi(OtBu)3 ligand,4 HOSi(OtBu)3 displays
similar melting and boiling behavior without thermal de-
composition. The TGA and DSC results suggest that 1 and
2 will require a catalyst to facilitate their thermal conversions
to borosilicate materials. For example, we have found that
the thermal decomposition of M[OSi(OtBu)3]n complexes is
acid catalyzed (e.g., HCl, AlCl3)4n,p and that aluminosilicate
materials catalyze the thermal decomposition of HOSi-
(OtBu)3.4s
molecular precursors to B/Si/O materials. Thermolyses of
sealed tubes containing 1 or 2 dissolved in toluene at 180
°C in the presence of a catalytic amount of AlCl3 (1 mol %)
resulted in the formation of transparent monolithic gels that
formed xerogels of composition BO1.5‚2SiO2 (BOSi2xg) or
BO1.5‚3SiO2 (BOSi3xg) upon drying, respectively. The ce-
ramic yields for BOSi2xg (23.4%) and BOSi3xg (25.8%)
correspond closely to the expected yields (25.4% and 26.8%,
respectively). Additional solution (toluene) thermolyses of
1 or 2 in the presence of the structure-directing block
copolymer EO20PO70EO20 (where EO ) ethylene oxide and
PO ) propylene oxide) led to the mesoporous xerogels
BOSi2epe and BOSi3epe, respectively, upon drying. It has
previously been reported that use of this and related block
copolymers have provided mesostructured oxide materials.7,21
To our knowledge, these are the first examples of borosilicate
materials obtained from single-source molecular precursors
in nonpolar media.20
To characterize the chemistry involved in materials forma-
tion, the solution thermolyses of 1 and 2 were monitored
using solution NMR spectroscopy by heating a toluene-d8
solution of the precursor, AlCl3 (∼1 mol %), and Cp2Fe (as
a standard) to 180 °C for 1 h in a sealed NMR tube. The
major organic product detected in each case was CH2CMe2
t
with generation of only small amounts of BuOH and trace
amounts of HOSi(OtBu)3. This indicates that the thermolytic
conversions of 1 and 2 proceed efficiently.4,5,7
Nitrogen porosimetry was used to characterize the surface
areas and pore structures of the B/Si/O materials generated
from 1 and 2 (Figure 2). The Brunnaur-Emmett-Teller
(BET) method22 was used to calculate the surface areas, and
the Barrett-Joyner-Halenda (BJH) method23 was used to
obtain the pore size distributions. The surface area and pore
Borosilicate Materials from 1 and 2. In light of the above
observations, we elected to perform thermal decompositions
of 1 and 2 in the presence of a small amount of AlCl3 to
determine the efficiency of these compounds as single-source
(21) For example: (a) Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. J. Science
1995, 269, 1242. (b) Attard, G. S.; Glyde, J. C.; Go¨ltner, C. G. Nature
1995, 378, 366. (c) Go¨ltner, C. G.; Antonietti, M. AdV. Mater. 1997,
9, 431. (d) Prouzet, E.; Pinnavaia, T. J. Angew. Chem., Int. Ed. Engl.
1997, 36, 516. (e) Zhao, D. Y.; Feng, J. L.; Huo, Q. S.; Melosh, N.;
Fredrickson, G. H.; Chmelka, B. F.; Stucky, G. D. Science 1998, 279,
548. (f) Kriesel, J. W.; Sander, M. S.; Tilley, T. D. AdV. Mater. 2001,
13, 331.
(22) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60,
309.
(23) Barret, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951,
73, 373.
(20) (a) Irwin, A. D.; Holmgren, J. S.; Zerda, T. W.; Jonas, J. J. Non-
Cryst. Solids 1987, 89, 191. (b) Irwin, A. D.; Holmgren, J. S.; Jonas,
J. J. Non-Cryst. Solids 1988, 101, 249. (c) Zha, C.; Atkins, G. R.;
Masters, A. F. J. Non-Cryst. Solids 1998, 242, 63. (d) Kasgo¨z, A.;
Misono, T.; Abe, Y. J. Non-Cryst. Solids 1999, 243, 168. (d) Soraru`,
G. D.; Babonneau, F.; Gervais, C.; Dallabona, N. J. Sol-Gel Sci.
Technol. 2000, 18, 11. (e) Beckett, M. A.; Rugen-Hankey, M. P.;
Varma, K. S. Chem. Commun. 2000, 1499.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1143

Fujdala et al.
Figure 3. TEM micrographs of (a) BOSi2xg and (b) BOSi3xg after
calcination at 500 °C.
Figure 2. N2 adsorption-desorption isotherms and pore size distributions
(inset) after calcination at 500 °C for (a) BOSi2xg and BOSi2epe and (b)
BOSi3xg and BOSi3epe
.
volume data for BOSi2xg (48 m2 g-1 and 0.17 cm3 g-1) after
calcination at 500 °C reveal a 1 order of magnitude (smaller)
difference when compared to those of calcined BOSi3xg (590
m2 g-1 and 1.21 cm3 g-1). In contrast, the surface areas and
pore volumes of BOSi2epe (168 m2 g-1 and 0.70 cm3 g-1)
and BOSi3epe (151 m2 g-1 and 0.41 cm3 g-1) were quite
similar. Thus, the presence of the copolymer appears to have
a moderating effect on the pore structure and surface area.
The pore size distributions as calculated from the adsorption
branch of the adsorption-desorption isotherms revealed that
BOSi2xg and BOSi3xg possess mostly pores with radii less
than 10 Å, although larger pores were evident (Figure 2
insets). Analysis of the adsorption pore size distributions for
BOSi2epe and BOSi3epe revealed a wide range of pore sizes
in the mesoporous region (Figure 2 insets). The average pore
radii as calculated using the BJH method and the adsorption
(desorption) data for BOSi2epe and BOSi3epe were 118 Å
(78 Å) and 59 Å (48 Å), respectively. Although the pore
size distributions for BOSi2epe and BOSi3epe are not as
narrow as those for other materials generated thermolytically
in the presence of EO20PO70EO20,7,21f their mesoporous nature
is evident.
Transmission electron microscopy (TEM) studies revealed
that the nontemplated xerogels (BOSi2xg and BOSi3xg)
consisted of relatively large spherical particles of varying
size (Figure 3). In contrast to this, the mesoporous boro-
silicates (BOSi2epe and BOSi3epe) exhibited significantly
smaller particles with a more uniform size distribution
(Figure 4). The presence of a wormhole-type motif is evident
for both of these samples. Similar (although more ordered)
wormhole structures have been previously observed by TEM
for materials generated thermolytically from molecular
Figure 4. TEM micrographs of (a) BOSi2epe and (b) BOSi3epe after
calcination at 550 °C.
precursors in the presence of block copolymers.7,21f The
highly disordered nature of the pore structures for BOSi2epe
and BOSi3epe likely arises from the structure-directing
copolymer acting merely as an anisotropic space filler with
little influence over the long-range ordering.7,21
Low-angle powder X-ray diffraction (PXRD) is often used
to indicate the presence of mesostructural ordering with a
single (well-defined) peak indicating a disordered structure
with little long-range order and multiple peaks suggesting
significant long-range order.21 Such studies of the meso-
porous B/Si/O materials (BOSi2epe and BOSi3epe) from 2θ
) 0.5 to 3° revealed only a single broad feature originating
at 2θ ) 0.9 and 1.1° for BOSi2epe and BOSi3epe, respec-
tively, and increasing in intensity to 2θ ) 0.5° (the start of
the measurement). The ill-defined nature and broadness of
these peaks suggest that there is no long-range meso-
structural order and that the mesopores cover a wide size
distribution.
Synthesis of the New Boronous Acid HOB[OSi-
(OtBu)3]2. Hydrolysis of 1 with only 1 equiv of H2O in
benzene-d6 led to the appearance of two new resonances at
5.71 and 1.41 ppm in the 1H NMR spectrum. Integration of
these resonances suggested that this new species could be
HOB[OSi(OtBu)3]2 (3, eq 3). Nearly quantitative amounts
1
of 3 (by H NMR spectroscopy and integration against a
Cp2Fe standard) were observed initially; however, slow
1144 Inorganic Chemistry, Vol. 42, No. 4, 2003

Tris(tert-butoxy)siloxy DeriWatiWes of Boron
decomposition to insoluble material occurred over the course
of a few hours.
+ H2O benzene, 25 °C8
(tBuO)B[OSi(OtBu)3]2
-tBuOH
1
HOB[OSi(OtBu)3]2
(3)
3
Isolation of 3 was ultimately achieved by performing the
reaction of 1 with 1 equiv of H2O in benzene at 25 °C for 1
h. Subsequent freezing of the reaction mixture (at 0 °C) and
removal of the solvent in vacuo (via lyophilization) allowed
the isolation of a mixture of 3 and small amounts of 2 (ca.
5%) as a colorless powder. Sublimation of the powder at 50
°C in vacuo resulted in the isolation of analytically pure 3
in 86% yield. Attempts to isolate 3 from other solvents
(toluene, Et2O, THF, and pentane), either upon cooling (for
crystallization) or upon concentration in vacuo, resulted in
the decomposition of 3. However, 3 is stable for several
months in the solid state at room temperature. Although we
were unable to crystallographically characterize 3, its identity
was confirmed by NMR (1H, 13C, and 11B) and IR spec-
troscopies and combustion analysis (C and H). Attempts to
prepare HOB(OtBu)2 via hydrolysis of B(OtBu)3 were
hindered by rapid hydrolysis and condensation processes to
form B2O3. The exclusive formation of 3 during the initial
stages of the hydrolysis of 1 indicates that the B-OtBu
linkages are significantly more susceptible to hydrolytic
cleavage than are the B-OSi(OtBu)3 fragments. Hence, it
follows that the order of hydrolytic stability is B-OtBu <
B-OSi , Si-OtBu.
Figure 5. Thermal ellipsoid plot showing one unique molecule of 4 from
the asymmetric unit at the 50% probability level. Si1 represents the Si atom
farthest from the Zr-Me group. Hydrogen atoms have been omitted for
clarity.
involved in identifying the structures of solid-supported
organometallic fragments make the synthesis of well-defined
molecular model complexes that mimic potential structures
of grafted fragments (containing E-O-M linkages) highly
desirable. In a recent investigation by Basset and co-workers
on the nature of surface-bound species resulting from the
grafting of CpZrMe3 and Cp2ZrMe2 onto various Lewis
acidic oxide materials (e.g., Al2O3 and Al2O3/SiO2), structures
with Zr-O surface linkages that closely resemble that of 4
were proposed.25
Structural Characterization of 4. Colorless crystals
(tablets) of 4 that were of sufficient quality for an X-ray
crystallographic analysis were grown from a concentrated
pentane solution after storage at -30 °C for several weeks.
Complex 4 crystallized in the triclinic space group P1h with
36 molecules in the unit cell (Z ) 36). This corresponds to
18 unique molecules in the asymmetric unit (Z′ ) 18), which
we believe is the largest number of molecules ever observed
in the asymmetric unit of a fully documented crystal
structure. A thermal ellipsoid plot of one of the molecules
from the asymmetric unit of 4 is shown in Figure 5, and a
space-filling diagram of the entire asymmetric unit is shown
in Figure 6.
Synthesis of Cp2Zr(Me)OB[OSi(OtBu)3]2 (4). With the
potential ligand precursor 3 in hand, we sought to explore
its use as a reagent for synthesis of new boryloxide
complexes of the type LxM{OB[OSi(OtBu)3]2}y. The reaction
of 3 with Cp2ZrMe2 in pentane at 25 °C resulted in the
formation of Cp2Zr(Me)OB[OSi(OtBu)3]2 (4) (quantitative
by 1H NMR spectroscopy, eq 4). Colorless crystals of
analytically pure 4 were isolated in 90% yield from a
concentrated pentane solution at -78 °C. The crystals of 4
proved to be extremely moisture sensitive and exhibited a
melting/decomposition transition at just above room tem-
perature (ca. 30 °C). Complex 4 is stable in the presence of
1 equiv of 3 in benzene-d6 solution for 2 h at 25 °C, after
which time decomposition was evident.
A search of the Cambridge Structural Database (CSD,
version 5.22, October 2001) for structures with Z′ g 16
revealed only 6 compounds.27 Of these, two structures (a
diterpenoid27d and HOSnMe327a) had Z′ g 18; however, no
3D coordinates were listed in the CSD so it was not possible
to analyze the data independently. The information given in
the original paper for the diterpenoid structure (Z′ ) 20.5)
(26) Duchateau, R.; Cremer, U.; Harmsen, R. J.; Mohamud, S. I.;
Abbenhuis, H. C. L.; van Santen, R. A.; Meetsma, A.; Thiele, S. K.-
H.; van Tol, M. F. H.; Kranenburg, M. Organometallics 1999, 18,
5447.
(27) (a) Kasai, N.; Yasuda, K.; Okawara, R. J. Organomet. Chem. 1965,
3, 172. (b) Bruckner, S. Acta Crystallogr., Sect. B 1982, 38, 2405. (c)
Hsu, L.-Y.; Nordman, C. E. Science 1983, 220, 604. (d) Hashimoto,
T.; Toyota, M.; Koyama, H.; Kikkawa, A.; Yoshida, M.; Tanaka, M.;
Takaoka, S.; Asakawa, Y. Tetrahedron Lett. 1998, 39, 579. (e)
Hassaballa, H.; Steed, J. W.; Junk, P. C.; Elsegood, M. R. J. Inorg.
Chem. 1998, 37, 4666. (f) Chou, S.-S. P.; Liu, S.-H. J. Organomet.
Chem. 1998, 555, 227.
This new boryloxide complex, containing both Zr-O-B
and Zr-Me linkages, could serve as a molecular model for
oxide-supported metallocene complexes.24-26 The difficulties
(24) (a) Hlatky, G. G. Coord. Chem. ReV. 1999, 181, 243. (b) Hlatky, G.
G. Chem. ReV. 2000, 100, 1347.
(25) Jezequel, M.; Dufaud, V.; Ruiz-Garcia, M. J.; Carrillo-Hermosilla,
F.; Neugebauer, U.; Niccolai, G. P.; Lefebvre, F.; Bayard, F.; Corker,
J.; Fiddy, S.; Evans, J.; Broyer, J.-P.; Malinge, J.; Basset, J.-M. J.
Am. Chem. Soc. 2001, 123, 3520.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1145

Fujdala et al.
amined for unusual systematic absences, which might
indicate a twinned crystal or a falsely determined cell. The
data were also analyzed for twinning using the program
“Gemini”28 which again indicated no sign of twinning.
Further, the cell determined in this analysis was the same as
that found originally without the original being provided at
the outset of the analysis.29
That this was not a unique crystal sample (an artifact of
the “selection effect”) is demonstrated by our observation
of the same unit cell determined from four different crystals
of varying quality selected from two different recrystalliza-
tions of complex 4 over the period of 1 month. The
observation of the same unit cell for multiple crystals also
suggests that crystal cracking is not responsible for the
unusual nature of the structure. The cell and the structure
were additionally subjected to the routines of MISSYM and
ADDSYM in the PLATON suite of programs.30 No ad-
ditional symmetry or higher symmetry space group was
found. Careful examination of each molecule in the asym-
metric unit shows that each is in some way unique either in
conformation or (equally important in a crystal structure)
orientation.
Details of the data collection, structure solution, and
refinement are provided in the Experimental Section and in
Table 1. Seven of the molecules in the asymmetric unit
t
exhibited rotational disorder in a single Bu group, and this
t
commonly observed Bu disorder was modeled using two
carbon atoms at 50% occupancy for each disordered Me
group. Some of the OtBu oxygen atoms (18 out of 108)
exhibited disorder, and this was modeled over two positions
at 50% occupancy. The Cp rings of one of the molecules
had positional disorder that was easily modeled over two
sites with appropriate atoms at 50% occupancy. All other
Cp ligands were normal. Remarkably, all 18 MeZrOB(OSi)2
moieties were well behaved and contained no disordered
atoms. The Zr, Si, and nondisordered C and O atoms were
modeled with anisotropic thermal parameters. There are no
significantly close intermolecular contacts (i.e., no “hydrogen
bumping”).
Figure 6. Space-filling diagram of the asymmetric unit for the crystal
structure of 4 as generated from the crystallographic data.
is inconsistent and strongly suggests a typographical error
in the cell parameter a (a ) 14.82 Å rather than a ) 114.82
Å as reported) and in the Z value (Z ) 8 rather than Z ) 82
reported).27d If these changes are made, the cell volume
agrees with the reported volume (4357(1) Å3) and a more
reasonable Z′ value of 2 is obtained. The structure of
HOSnMe3 (Z′ ) 32) was determined from a poor quality
data set (232 observations), and the reported unit cell was
an extension of a smaller subcell (a ) 2a′ and c ) 2c′).27a
The extension to the larger unit cell was based upon 10 “very
weak” reflections. The actual unit cell is likely one of the
two given subcells (Pn, Z ) 16, or P21nm, Z ) 2).27a The
only other metal-containing structures in the search results,
[UO2(H2O)3Cl2]‚15-crown-527e and tricarbonyl[(1-4-η4)-6-
exo-cyano-2-(phenylsulfonyl)-1,3-cyclohexadiene]iron,27f have
Z′ ) 16.
The nature of the crystal structure afforded a rare op-
portunity to perform detailed statistical analyses of the metric
parameters with a large pool of independent molecules to
analyze. The usual metric parameters (bond lengths and
angles) show regular, normal distributions as would be
expected if the data were correct and accurate. The disorder
discussed above gives rise to a few outliers, which can be
ignored for statistical purposes. Figure 7 presents the
distribution of C-C bond distances and clearly shows a
t
normal distribution for both the Cp and Bu C-C bond
lengths. The observed mean C-C bond distances are
Whenever an “unusual” or “unique” structure is deter-
mined, it is proper to consider whether the “uniqueness” is
due to nature or is due to the failure of the experimenters to
recognize other explanations for the results. While no such
examination is foolproof, we did apply stringent tests to the
crystallographic data obtained for 4. The raw data (both CCD
frames and derived structure factors) were repeatedly ex-
(28) Sparks, R. A. GEMINI; Bruker-AXS, Inc.: Madison WI, 1999 (part
of the SMART software suite).
(29) We thank Dr. Victor G. Young, Jr., Director, X-ray Crystallographic
Facility, Department of Chemistry, University of Minnesota, for these
analyses.
(30) Spek, A. L. PLATON, a multipurpose crystallographic tool; Utrecht
University: Utrecht, The Netherlands, 2001. We thank Prof. Anthony
L. Spek, Utrecht University, Utrecht, The Netherlands, for these
analyses.
1146 Inorganic Chemistry, Vol. 42, No. 4, 2003

Tris(tert-butoxy)siloxy DeriWatiWes of Boron
Figure 7. Plot of the distribution of C-C bond lengths for 4. The space between x-axis tick marks corresponds to 0.02 Å. The outliers are due to disorder
in the crystal structure.
Table 3. Bond Distance (Å) and Angle (deg) Statistics for 4
min
max.
mean
N
valuea
valuea
valueb
Distances
Zr-O
18
18
180
18
36
1.921
2.250
2.474 (2.455)
1.296
1.338
1.999
2.297
2.566
1.345
1.394
1.637
1.629 (1.813)
1.490 (1.716)
1.586 (1.990)
1.450 (1.760)
1.974(4)
2.281(3)
2.520(2)
1.329(3)
1.370(3)
1.604(3)
1.610(4)
1.460(5)
1.509(3)
1.375(6)
Zr-C(CH3)
Zr-C(Cp)
B-O(Zr)
B-O(Si)
Si-O(B)
Si-O(C)
C-O
36
1.534
126
126
362
203
1.492 (1.398)
1.398 (1.269)
1.402 (1.250)
1.292 (0.910)
C-C(tBu)
C-C(Cp)
Angles
96.9
146.5
111.3
133.2
102.7 (84.6)
129.6 (118.7)
104.2 (78.5)
C(CH3)-Zr-O
Zr-O-B
O-B-O
B-O-Si
O-Si-O
Si-O-C
C-C-C
18
18
54
36
251
126
549
100.3
177.5
124.7
164.0
116.2 (138.8)
141.0 (151.7)
113.2 (148.5)
98.4(2)
160(2)
120.0(4)
144(1)
109.3(4)
132.0(4)
109.9(2)
Figure 8. Plot of the distribution of B-O(Zr) and B-O(Si) bond lengths
for 4. The space between x-axis tick marks corresponds to 0.01 Å.
t
1.375(6) and 1.509(3) Å for the Cp and Bu groups,
respectively. Figure 8 shows the distribution of B-O bond
distances and illustrates that the B-O(Zr) distances are
generally shorter than the B-O(Si) distances. The mean
B-O(Zr) distance is 1.329(3) Å, and the mean B-O(Si)
distance is 1.370(3) Å with no significant difference between
the B-O(Si1) and B-O(Si2) mean distances (1.364(3) and
1.377(3) Å), where Si1 corresponds to the siloxide group
farthest from the Zr-CH3 linkage in each molecule. The
mean Si-O distance in 4 is 1.609(4) Å with no significant
difference between Si(1)-O and Si(2)-O distances or
between Si-O(C) and Si-O(B) distances. Table 3 provides
a summary of the statistical analyses for selected bond
distances and angles.
The mean O-B-O angle is 120.0(4)°, as would be
expected for a trigonal planar boron center. The B-O-Si
angles exhibit a wide range of values (from 133.2 to 164.0°);
however, these approximate a (wide, tailed) normal distribu-
tion as shown in Figure 9. Interestingly, the Zr-O-B angles
show an apparent trimodal distribution over a wide range of
a The value in parentheses indicates a distance or angle involving a
disordered atom. b Calculated using “N” parameters, including disorder
where appropriate. Esd in parentheses.
values from 146.5 to 177.5° with a mean value of 160(2)°
(Figure 9). Of the 36 Zr-O-B-O torsion angles that ranged
from -180 to 180°, 9 were between -90 and -60°, another
9 were between 90 and 120°, and the rest were spread evenly
within given 30° intervals. The net effect of the OB[OSi-
(OtBu)3]2 ligand appears to be a steric shrouding of the metal
center in an umbrella-like fashion with the rather long Zr-
O-B-O linkage as the handle. The Zr-Me group, however,
is relatively unprotected in all observed conformations. The
oxygen atom of the Zr-O-B linkage behaves as if it were
a universal joint with a soft bending potential, allowing the
tris(tert-butoxy)siloxide groups to assume many not-so-subtly
different orientations relative to the Zr atom and the Cp
groups. This appears to be a major factor in the crystallization
of 18 independent molecules in the asymmetric unit (but does
Inorganic Chemistry, Vol. 42, No. 4, 2003 1147

Fujdala et al.
spectra to those of an authentic sample synthesized by the
addition of 2 equiv of HOSi(OtBu)3 to a pentane solution of
Cp2ZrMe2.33 The thermolysis of 4 in benzene-d6 at 80 °C
(in a sealed J-Young NMR tube) also led to the formation
1
of 5 (90% yield by H NMR spectroscopy).
The solvent-induced conversion of 4 to 5 suggests the
participation of a polar intermediate. This, coupled with the
known ability of boryloxides to abstract alkyl groups from
Lewis acidic centers,32,34 leads us to propose an intra-
molecular methide transfer from Zr to B, with subsequent
transfer of a tris(tert-butoxy)siloxide ligand from the resulting
{OBMe[OSi(OtBu)3]2}- group to the now cationlike Zr
center. Elimination of “BMeO” from a Zr-OB(Me)[OSi-
(OtBu)3] species would then form the observed product (5).
Attempts to observe intermediates by performing thermolyses
in the presence of a base (pyridine, MeCN) or upon
dissolution of 4 in Et2O/base mixtures led to the exclusive
formation of 5. Unfortunately, no intermediate species could
Figure 9. Plot of the distribution of Zr-O-B and Si-O-B angles for 4.
The space between x-axis tick marks corresponds to 2°.
Scheme 1
1
be identified in the above reaction mixtures (by H NMR
spectroscopy) and no resonances were detected in their 11B
NMR spectra. On the basis of mass balance considerations,
the boroxine (MeBO)3 is a reasonable product that could
account for the eliminated “BMeO” fragments from 4.
However, we were unable to detect this (or any other) boron-
1
containing species by H and 11B NMR spectroscopy or by
GC/MS analysis of reaction mixtures; hence, the fate of the
boron upon conversion of 4 to 5 is unknown at this time.
The formation of nonvolatile oligomeric or polymeric species
is possible and would lead to severely broadened NMR
resonances that could account for the difficulties encountered
in locating the boron.
not explain why there are 18 molecules, nor does it describe
the nature of the long-range ordering forces).
Complex 4 is a rare example of a crystallographically
characterized transition metal boryloxide and is, to our
knowledge, the first transition metal boryloxide with a
M-OB(OR)2 moiety. The other structurally characterized
transition metal boryloxides are [Co(µ-OBMes2)2Li(THF)2Cl2-
Li(THF)2,9 [Mn(OBTrip2)(µ-OBTrip2)]2,10 [Fe(OBMes2)(µ-
OBMes2)]2,10 [(NMe2)2WOBMes2]2,11 [(tBuO)2WOBMes2]2,11
[(µ3-9-BBN-9-O)ZnEt]4,14 and [(µ3-9-BBN-9-O)CdMe]4.14
Several main group element boryloxides have also been
structurally characterized including those containing -OBAr2
(Li9,12a and Al12b,13), -OB[CH(SiMe3)2]2 (Li),31 and 9-BBN-
9-O ligands (Al, In, Ga),32 as well as B(OEPh3)3 (E ) Sn,17
Si18) and PhB(OGePh3)2.18
Conclusions
Syntheses of the new boron-based tris(tert-butoxy)siloxide
t
species BuOB[OSi(OtBu)3]2 (1), B[OSi(OtBu)3]3 (2), and
HOB[OSi(OtBu)3]2 (3) have been achieved. Compounds 1
and 2 are efficient single-source molecular precursors that
have been used to produce borosilicate xerogels (BOSi2xg
and BOSi3xg, respectively) from nonpolar media utilizing
the thermolytic molecular precursor method. BOSi3xg is a
high surface area borosilicate xerogel (590 m2 g-1) with
formula BO1.5‚3SiO2, while the xerogel BOSi2xg (BO1.5‚2SiO2)
has a smaller area (48 m2 g-1). Use of EO20PO70EO20 during
the solution thermolyses of 1 and 2 resulted in formation of
new xerogels (BOSi2epe and BOSi3epe, respectively) that did
not have ordered mesostructures; however, these materials
were mesoporous in nature and had moderately high surface
areas (ca. 160 m2 g-1).
Reactions of Complex 4. Reaction of 3 with Cp2ZrMe2
in Et2O gave a new product, Cp2Zr[OSi(OtBu)3]2 (5), that
results (formally) from the elimination of “BMeO”. Complex
1
5 was formed quantitatively as indicated by a H NMR
spectrum of the crude product mixture. Isolation of complex
5 was achieved by crystallization from the reaction mixture
(Et2O) at -78 °C in 84% yield. Addition of Et2O to neat 4
at 25 °C also led to the quantitative formation of 5,
suggesting that 4 may be formed in the reaction of Cp2ZrMe2
with 3 in Et2O but that it readily converts to 5 under these
conditions (Scheme 1). The identity of complex 5 was
confirmed by comparison of its 1H NMR, 13C NMR, and IR
(33) Original synthesis of 5: reaction of 2 equiv of HOSi(OtBu)3 with
Cp2ZrMe2 in pentane at 25 °C and subsequent crystallization from
pentane at -78 °C (66%). Terry, K. W. Ph.D. Thesis, University of
California at San Diego, 1993. Data for 5 are as follows. Anal. Calcd
for C34H64O8Si2Zr: C, 54.58; H, 8.62. Found: C 54.79; H, 8.33. IR
(cm-1): 3100 w br, 1386 m, 1362 s, 1240 m, 1215 m, 1192 s, 1050
vs br, 1022 s sh, 1000 s sh, 946 s, 910 w sh, 845 vw sh, 822 m sh,
802 s, 720 vw sh, 696 s, 645 w, 614 vw sh, 525 vw sh, 504 vw sh,
481 w sh, 454 w, 424 w sh, 386 w, 350 w, 326 w. 1H NMR (benzene-
d6, 300 MHz): δ 1.46 (s, 54H, Si(OtBu)3), 6.43 (s, 10H, Cp2Zr).
13C{1H} NMR (benzene-d6, 75.5 MHz): δ 113.25 (s, Cp2Zr), 72.08
(s, SiOC(CH3)3), 31.98 (s, SiOC(CH3)3).
(31) Beck, G.; Hitchcock, P. B.; Lappert, M. F.; MacKinnon, I. A. Chem.
Commun. 1989, 1312.
(32) Anulewicz-Ostrowska, R.; Lulinski, S.; Serwatowski, J. Inorg. Chem.
1999, 38, 3796.
(34) Synoradzki, L.; Boleslawski, M.; Lewinski, J. J. Organomet. Chem.
1986, 284, 1.
1148 Inorganic Chemistry, Vol. 42, No. 4, 2003

Tris(tert-butoxy)siloxy DeriWatiWes of Boron
JEOL 200cx electron microscope operating at an accelerating
voltage of 200 kV. Samples for TEM analyses were prepared via
deposition using a pentane suspension of finely ground material
on a “lacey carbon” copper grid obtained from Ted Pella Inc. X-ray
powder diffraction data were obtained using a Siemens D5000
diffractometer operating with θ-2θ geometry at room temperature.
The PXRD data were collected using Cu KR radiation (λ ) 1.5406
Å). Collection times of 1.5 or 5.0 s and step sizes of 0.02 or 0.01°
were used for high- or low-angle measurements, respectively.
Elemental analyses were performed by the College of Chemistry
Micro-Mass Facility at the University of California, Berkeley, CA.
B(OtBu)3 was purchased from Gelest and used as received, the block
copolymer EO20PO70EO20 (Pluronic 123, Mav ) 5800) was a
generous gift from BASF (Mt. Olive, NJ), and HOSi(OtBu)335 and
Compound 3 appears to be the first boronous acid to
contain a siloxide group, and it is a rare example of an
isolable boronous acid.9,15 Apparently, the steric bulk of the
tris(tert-butoxy)siloxide groups suppresses condensation
reactions and permits the isolation of 3 in high yield. This
species should prove to be a valuable ligand precursor for
the synthesis of new (siloxy)boryloxide complexes. The use
of such compounds as single-source molecular precursors
to complex oxide materials, and as molecular models for
oxide-supported metal species, will be reported soon. Initial
reactivity studies of 3 with Cp2ZrMe2 demonstrate the clean
formation of a (siloxy)boryloxide complex, Cp2Zr(Me)OB-
[OSi(OtBu)3]2 (4), with a unique crystal structure containing
18 independent molecules in its asymmetric unit.
36
Cp2ZrMe2 were prepared by literature procedures.
tBuOB[OSi(OtBu)3]2 (1). A neat mixture of B(OtBu)3 (2.436 g,
10.58 mmol) and HOSi(OtBu)3 (5.598 g, 21.17 mmol) was heated
at 80 °C for 18 h in a Schlenk tube. The Schlenk tube was cooled
to room temperature, and the volatile materials were removed in
vacuo leaving a colorless solid. The solid was dissolved in toluene
(20 mL), and acetonitrile (5 mL) was subsequently added. The
resulting solution was stored at -30 °C for 24 h, yielding colorless
platelike crystals. The crystals were washed with acetonitrile to
yield 5.074 g (8.31 mmol) of analytically pure 1 (79%). Anal. Calcd
for C28H63BO9Si2: C, 55.06; H, 10.40. Found: C, 55.10; H, 10.42.
IR (cm-1): 2977 vs, 2933 m, 2902 w, 2876 w, 1474 m, 1414 m,
1389 s, 1366 s, 1341 s, 1244 m, 1194 s, 1069 vs, 1027 m, 879 m,
The statistical analyses of the crystallographic data for 4
show a normal distribution for the metric parameters allowing
a high degree of confidence, even from a not overly
determined structure (i.e., low symmetry and low data
redundancy). Often crystals are deemed to be of insufficient
quality when large unit cells are found for low-symmetry
space groups during crystallographic investigations. How-
ever, the structure of 4 clearly illustrates that such large unit
cells can be real and provide accurate structures, although
such cases are obviously rare. The extreme flexibility of
B-O-X bond angles is illustrated here by the fact that, for
the 18 unique molecules in the asymmetric unit of 4,
chemically identical B-O-Zr and B-O-Si angles each
exhibit ranges covering more than 30°.
The chemistry of transition metal boryloxide species
appears to be relatively unexplored and, given the recently
reported interesting catalytic activity of aluminum boryl-
oxides,12b,13 may provide access to new reactive and poten-
tially catalytic species. Initial reactivity studies of 4 suggest
that the Zr-CH3 and Zr-O-B linkages are highly reactive
and easily cleaved, even upon simple dissolution in Et2O.
Further investigations of the reactivity of HOB[OSi(OtBu)3]2
with various transition and main group metal species are
currently underway.
1
832 m, 704 m, 675 w, 622 w, 514 m, 483 m, 425 w. H NMR
(300 MHz): δ 1.48 (s, 9 H, BOtBu), 1.47 (s, 54 H, Si(OtBu)3).
13C{1H} NMR (100.6 MHz): δ 72.92 (s, Si(OC(CH3)3), 72.62 (s,
BOC(CH3)3), 31.75 (s, Si(OC(CH3)3), 30.39 (s, BOC(CH3)3).
11B{1H} NMR (160.4 MHz): δ 12.9 (s, br). Mp: ∼72 °C. Bp:
∼170 °C.
B[OSi(OtBu)3]3 (2). To a mixture of HOSi(OtBu)3 (4.197 g, 15.9
mmol) and B(OtBu)3 (1.218 g, 5.29 mmol) in a Schlenk tube was
added toluene (5 mL). The tube was heated at 80 °C, and the
reaction mixture was magnetically stirred for 24 h. Subsequent
cooling to room temperature and removal of the volatile materials
in vacuo left a colorless solid. The resulting solid was then dissolved
in toluene (10 mL), and acetonitrile (2.5 mL) was added. Analyti-
cally pure colorless needlelike crystals were obtained after storage
at -30 °C for 24 h and subsequent washing of the isolated crystals
with acetonitrile (61% yield; 2.582 g, 3.22 mmol). Anal. Calcd for
C36H81BO12Si3: C, 53.98; H, 10.19. Found: C, 53.79; H, 10.15.
IR (cm-1): 2976 vs, 2932 m, 2907 w, 2875 vw, 1474 vw, 1388 m,
1366 s, 1244 m, 1194 s, 1075 vs, 1028 m, 881 m, 832 m, 704 m,
Experimental Section
General Methods. All synthetic manipulations were performed
under an atmosphere of nitrogen using standard Schlenk techniques
and/or a Vacuum Atmospheres drybox, unless noted otherwise. All
solvents used were distilled from sodium/ benzophenone, potassium/
benzophenone, sodium, or calcium hydride as appropriate. The
solution NMR spectra were recorded at ambient temperature using
a Bruker AMX300 spectrometer at 300 (1H) MHz, a Bruker AM400
spectrometer at 400 (1H) or 100.6 (13C) MHz, or a Bruker DRX500
spectrometer at 160.4 (11B) MHz. Chemical shifts are reported in
ppm relative to tetramethylsilane (1H and 13C) and BF3‚Et2O (11B).
Benzene-d6, vacuum transferred from a Na/K alloy, was used as
the solvent for all NMR studies. Infrared spectra were recorded
using a Mattson Infinity Series FTIR spectrometer with all samples
being pressed into KBr disks. Thermal analyses were performed
using a TA Instruments SDT 2960 integrated thermogravimetric/
differential scanning calorimetric analyzer (TGA/DSC). N2 poros-
imetry was performed using a Quantochrome Autosorb 1 surface
area analyzer with all samples heated at 120 °C for a minimum of
12 h prior to measurement. TEM images were acquired using a
1
616 w, 518 m, 493 w. H NMR (300 MHz): δ 1.50 (s, 81 H,
Si(OtBu)3). 13C{1H} NMR (100.6 MHz): δ 73.06 (s, Si(OC(CH3)3),
31.96 (s, Si(OC(CH3)3). 11B{1H} NMR (160.4 MHz): δ 12.9 (s,
br). Mp: ∼162 °C. Bp: ∼257 °C.
BOSi2xg and BOSi3xg. Representative solution thermolyses to
form BOSi2xg and BOSi3xg were carried out as follows: A thick-
walled pyrolysis tube was charged with a toluene (8.0 mL) solution
of 1 (1.000 g, 1.637 mmol) or 2 (1.311 g, 1.637 mmol) in a drybox.
To this solution was added AlCl3 (0.002 g, 0.016 mmol, 1 mol %)
in solid form. The tube was flame-sealed after three freeze-pump-
thaw cycles and placed in a preheated oven at 180 °C. A clear
monolithic gel was evident after ∼45 min. The tube was heated
for an additional 36 h to ensure complete conversion. After this
heating period the tube containing the gel was cooled to room
(35) Abe, Y.; Kijima, I. Bull. Chem. Soc. Jpn. 1969, 42, 1118.
(36) Hunter, W. E.; Hrincr, D. C.; Bynum, R. V.; Pentilla, R. A.; Atwood,
J. L. Organometallics 1983, 2, 750.
Inorganic Chemistry, Vol. 42, No. 4, 2003 1149

Fujdala et al.
temperature and opened. The gel was transferred to a Schlenk tube
under N2 and dried under flowing N2 for 5 days at 25 °C giving
the xerogel BOSi2xg or BOSi3xg. The xerogels were further dried
in vacuo at 25 °C for 24 h and finally at 120 °C in vacuo for an
additional 24 h. The mass of BOSi2xg obtained was 0.234 g (23.4%
ceramic yield), and that of BOSi3xg was 0.339 g (25.8%). The
xerogel pieces were then ground to a fine powder and calcined at
500 °C under O2.
BOSi2epe and BOSi3epe. Representative solution thermolyses in
the presence of EO20PO70EO20 were carried out as follows: A thick-
walled pyrolysis tube was charged with a toluene solution (6.0 mL)
of 1 (0.400 g, 0.655 mmol) or 2 (0.500 g, 0.624 mmol) and
EO20PO70EO20 (0.300 g) in a drybox. To this solution was added
AlCl3 (1 mg, 1 mol %) in solid form. The tube was flame-sealed
after three freeze-pump-thaw cycles and treated as described
previously except that the as-synthesized gel was washed several
times with hexanes (to extract any free polymer) prior to the drying
process. After calcination at 550 °C for 8 h (to remove any
remaining block copolymer) the xerogels BOSi2epe (0.083 g, 21%)
or BOSi3epe (0.109 g, 22%) were obtained.
w, 517 vw, 491 w, 459 w, 428 w. 1H NMR (400 MHz): δ 6.07 (s,
10 H, Cp2Zr), 1.48 (s, 54 H, Si(OtBu)3), 0.49 (s, 3 H, ZrCH3).
13C{1H} NMR (100.6 MHz): δ 111.63 (s, Cp2Zr), 72.80 (s,
SiOC(CH3)3), 31.86 (s, Si(OC(CH3)3), 24.18 (s, ZrCH3). 11B{1H}
NMR (160.4 MHz): δ 13.0 (s, br). Mp: ∼30 °C (d).
X-ray Structure Determinations. Each crystal was mounted
on a quartz fiber using Paratone N hydrocarbon oil. In the case of
4, the crystals were kept cold at all times (-30 °C) to minimize
reaction with/dissolution in the Paratone oil. All measurements were
made on a SMART CCD area detector with graphite-mono-
chromated Mo KR radiation and a crystal to detector distance of
60.0 cm. Data for 1 and 4 were integrated using the program SAINT
(SAX Area-Detector Integration Program, V5.04; Bruker, 1995)
and were corrected for Lorentz and polarization effects. The data
were analyzed for agreement and possible absorption using XPREP
(part of the SHELXTL Crystal Structure Determination suite of
programs; Bruker, 1995). An empirical absorption correction based
upon comparison of redundant and equivalent reflections was
applied using SADABS (Siemens Area Detector Absorption Cor-
rections; Sheldrick, G. M., 1996). The structure of 1 was solved
by direct methods and refined using Fourier techniques (based upon
F2) with all calculations performed using the SHELXTL crystal-
lographic software package (Sheldrick, G. M. SHELXTL, V5.03;
Bruker, 1994). The structure of 4 was also solved by direct methods
and refined routinely (F2); however, a blocking strategy was applied
due to the large number of refined parameters (7648). The first
block involved refinement of all 18 Zr atoms, and each subsequent
block consisted of one of the independent molecules (including their
respective Zr atoms). During the solution process, three potential
pseudo-inversion centers were located. The true center was selected
by inspection of the Zr and Si positions and confirmed by
refinement statistics and the complete solution and refinement of
the structure. The molecules in the asymmetric unit of 4 lie in three
chains of six molecules that run parallel to the [201h] plane within
the lattice. All non-hydrogen and non-boron atoms that did not
exhibit disorder were refined with anisotropic thermal parameters
for both 1 and 4, and the hydrogen atoms were included in
calculated idealized positions but not refined.
HOB[OSi(OtBu)3]2 (3). To a magnetically stirring solution of 1
(0.679 g, 1.112 mmol) in benzene (10 mL) was added dropwise
via syringe H2O (20 µL, 1.112 mmol) in a Schlenk tube. The
reaction mixture was stirred at 25 °C for 1 h and was subsequently
cooled to 0 °C using an ice bath. The reaction mixture solidified
due to freezing of the benzene. The tube was evacuated while cold
resulting in the removal of the benzene (via lyophilization) leaving
an opaque powder consisting of 3 and a small amount of 2 as an
impurity (ca. 5%). Subsequent sublimation of this powder at 50
°C in vacuo (ca. 0.01 mmHg) resulted in isolation of analytically
pure 3 in 86% yield (0.528 g, 0.952 mmol). Anal. Calcd for
C24H55BO9Si2: C, 51.97; H, 9.99. Found: C, 51.93; H, 9.96. IR
(cm-1): 3223 br, 2977 vs, 2934 m, 2910 w, 2876 w, 1474 m, 1423
m, 1390 m, 1367 s, 1339 m, 1245 m, 1194 s, 1068 vs, 1025 m,
877 m, 832 m, 705 s, 618 w, 516 m, 489 m. 423 w. 1H NMR (300
MHz): δ 5.71 (s, 1 H, BOH), 1.41 (s, 54 H, Si(OtBu)3). 13C{1H}
NMR (100.6 MHz): δ 73.39 (s, Si(OC(CH3)3), 31.48 (s, Si(OC-
(CH3)3). 11B{1H} NMR (160.4 MHz): δ 15.4 (s, br). Mp: 57 °C.
Bp: ∼240 °C.
Cp2Zr(Me)OB[OSi(OtBu)3]2 (4). To a solid mixture of Cp2ZrMe2
(0.046 g, 0.184 mmol) and HOB[OSi(OtBu)3]2 (0.102 g, 0.184
mmol) in a Schlenk tube was added pentane (2 mL) with magnetic
stirring at 25 °C. The evolution of a gas was observed. After 30
min, the reaction mixture was concentrated in vacuo to 1 mL and
stored at -78 °C for 24 h. Analytically pure colorless crystals of
4 were isolated in 90% yield (0.127 g, 0.166 mmol). Anal. Calcd
for C35H67BO9Si2Zr: C, 53.21; H, 8.55. Found: C, 53.29; H, 8.67.
IR (cm-1): 2975 vs, 2931 m, 2900 w sh, 2874 w, 1474 w, 1458
w, 1419 w, 1389 m, 1366 s, 1339 m, 1243 m, 1193 s, 1066 vs,
1025 s, 976 w, 861 vw, 831 w, 801 m, 748 w, 704 m, 675 w, 616
Acknowledgment. This work was supported by the
Director, Office of Energy Research, Office of Basic Energy
Sciences, Chemical Sciences Division, of the U.S. Depart-
ment of Energy under Contract No. DE-AC03-76SF00098.
We thank A. Stacy for the use of instrumentation (PXRD).
Supporting Information Available: Crystallographic informa-
tion files (CIF) for 1 and 4. This material is available free of charge
via the Internet at http://pubs.acs.org.
IC0205482
1150 Inorganic Chemistry, Vol. 42, No. 4, 2003