Bifunctional mesoporous organosilica materials and their application in catalysis: Cooperative effects or not?

Article abstract of DOI:10.1021/cm903412e

The synthesis and characterization of a new periodically ordered mesoporous organosilica material (PMO) containing aniline as a bridging organic building block is presented. Furthermore the coordination of the vanadyl cation [VO] ²⁺ to the carb

Full text of DOI:10.1021/cm903412e

1472 Chem. Mater. 2010, 22, 1472–1482
DOI:10.1021/cm903412e
Bifunctional Mesoporous Organosilica Materials and Their Application
in Catalysis: Cooperative Effects or Not?
Andreas Kuschel, Malte Drescher, Thomas Kuschel, and Sebastian Polarz*
Department of Chemistry, University of Konstanz, 78457 Konstanz, Germany
Received August 30, 2009. Revised Manuscript Received January 18, 2010
The synthesis and characterization of a new periodically ordered mesoporous organosilica
material (PMO) containing aniline as a bridging organic building block is presented. Furthermore,
the coordination of the vanadyl cation [VO]2þ to the carboxy function of a second PMO material
containing benzoic acid is described. Bifunctional materials containing both mentioned functional
groups in different amounts (HOOC-PhSi2O3)1-x(H2N-PhSi2O3)x have been prepared. The investi-
gation of the dipolar coupling of [VO]2þ coordinated to the benzoic acid groups via EPR
spectroscopy has been used to derive a detailed picture about the surface structure of the pores on
the molecular scale. The catalytic activity of the bifunctional materials in a two-step chemical
transformation is investigated as a possible application. The vanadyl group catalyzes the cleavage of
an acetal followed by Knoevenagel condensation catalyzed by aniline. A cooperative effect of the two
groups has been observed which is, however, not caused by a special interaction between the groups
but is due to pure statistic reasons.
1. Introduction
1000 m2/g came into focus after the preparation of
mesoporous materials (meso-SiO2) like MCM-41 or
SBA-15 had been established.5-7 Several excellent review
articles focusing on the application of meso-SiO2 in
catalysis have already been published.4,8,9
Surface modification of meso-SiO2 with organic groups
can be achieved in three different ways.10 One method is
grafting respectively the covalent anchoring to silanol
groups present in mesoporous silica. A nice review ad-
dressing postfunctionalized meso-SiO2 and their applica-
tion in catalysis has been published by Ying et al.9 Two
sources, one of them organically modified, are used for
SiOx-network formation in the co-condensation method.
The most severe problem associated with the latter methods
is the inevitable dilution of the organosilica matrix with
pure silica. The composition of the materials has to be
Catalysis possesses without any doubt an actual and
future paramount importance. The recent progress re-
garding the synthesis of fine chemicals via molecular,
metal containing catalysts in the homogeneous phase is
impressive. It is the specific interplay between the metal
and the attached ligands which is responsible for the high
selectivity and activity in many molecular catalysts. On
the other hand, one of the advantages of heterogeneous
catalysis is the ease of separation of products and catalyst
from each other. Therefore, it is a relevant task to bridge
from catalytically active, molecular compounds to het-
erogeneous catalysis.1 Consequently, there has been a
tremendous and long lasting effort in the immobilization
of molecular catalysts on solid supports.2 Silica represents
the dominant support material because of the numerous
possibilities for shaping and for surface modification.3,4
For instance, silica with internal surface area of up to
described as (SiO2)1-x (RSiO1.5)x. x can be as high as
3
25%, but cases of only 5% modification have also been
reported frequently.5,11 It is also very difficult to guarantee
for a homogeneous distribution of the organic groups along
*Author to whom correspondence should be addressed. E-mail:
sebastian.polarz@uni-konstanz.de. Fax: (þ49)-7531 884406.
(1) Coperet, C.; Chabanas, M.; Saint-Arroman, R. P.; Basset, J. M.
Angew. Chem., Int. Ed. 2003, 42, 156. Cole-Hamilton, D. J. Science
2003, 299, 1702. Jas, G.; Kirschning, A. Chem.;Eur. J. 2003, 9, 5708.
(2) Kobayashi, S.; Akiyama, R. Chem. Commun. 2003, 449. Corma, A.;
Garcia, H. Adv. Synth. Catal. 2006, 348, 1391. Cozzi, F. Adv. Synth.
Catal. 2006, 348, 1367.
(3) Corma, A. Chem. Rev. 1997, 97, 2373. Sutra, P.; Brunel, D. Chem.
Commun. 1996, 2485. Johnson, B. F. G.; Raynor, S. A.; Shephard,
D. S.; Mashmeyer, T.; Thomas, J. M.; Sankar, G.; Bromley, S.; Oldroyd,
R.; Gladden, L.; Mantle, M. D. Chem. Commun. 1999, 1167.
(4) Brunel, D. Microporous Mesoporous Mater. 1999, 27, 329.
(5) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge,
C. T.; Schmitt, K. D.; Chu, C. T.; Olson, D. H.; Sheppard, E. W.;
McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc.
1992, 114, 10834.
(6) Kresge, C. T.; Leonowicz, M.; Roth, W. J.; Vartuli, J. C.; Beck, J. S.
Nature 1992, 359, 710. Yang, P. D.; Zhao, D. Y.; Chmelka, B. F.;
Stucky, G. D. Chem. Mater. 1998, 10, 2033. Polarz, S.; Smarsly, B.
J. Nanosci. Nanotechnol. 2003, 2, 581.
(7) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.;
Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548.
(8) Taguchi, A.; Schueth, F. Microporous Mesoporous Mater. 2005, 77,
1. Climent, M. J.; Corma, A.; Iborra, S.; Navarro, M. C.; Primo, J.
J. Catal. 1996, 161, 783. Moeller, K.; Bein, T. Chem. Mater. 1998, 10,
2950. Ciesla, U.; Schueth, F. Microporous Mesoporous Mater. 1999,
27, 131.
(9) Ying, J. Y.; Mehnert, C. P.; Wong, M. S. Angew. Chem., Int. Ed.
1999, 38, 56.
(10) Hoffmann, F.; Cornelius, M.; Morell, J.; Froeba, M. Angew.
Chem., Int. Ed. 2006, 45, 3216.
(11) Sutra, P.; Brunel, D. Chem. Commun. 1996, 2485.
r
pubs.acs.org/cm
Published on Web 01/28/2010
2010 American Chemical Society

Article
Chem. Mater., Vol. 22, No. 4, 2010 1473
the pore surface. Nevertheless, there was some impressive
progress especially regarding the use of organically modified
mesoporous silica materials in catalysis. For direct relevance
of the work reported here are papers about materials
containing two different functional groups and their use in
catalysis.12 For instance, Lin et al. have described coopera-
tive effects in different condensation reactions (aldol reac-
tion, Henry reaction, and cyanosilylation) observed for the
catalytic activity of meso-SiO2 postmodified by acid and
base groups.13 Davis et al. have investigated the role of the
pKa value of the used acid in such reactions in further
detail.14 One further example is the work of Sanchez et al.
who have combined the Knoevenagel condensation with an
surface immobilized ruthenium hydrogenation catalyst.15
Much higher content of organic modification can be
reached when sol-gel precursors with a bridging organic
group, (R0O)3Si-R-Si(OR0)3, are used for the preparation
of the so-called periodically ordered mesoporous orga-
nosilica (PMO) materials reported by Ozin et al., Inagaki
et al., and Stein et al. independent of one another.16 It has
to be emphasized that, unlike the co-condensation case,
(R0O)3Si-R-Si(OR0)3 is used in an undiluted form leading
to mesoporous materials with the composition RSi2O3,
thus, with a degree of organic modification of 100%.
Consequently, the density of the organic groups is
very high.17 Until recently, the number of functional
groups which have been introduced to the pore walls via
the PMO technology was relatively restricted.10 PMO
materials with walls containing bridging ;CH2;,
;CH2CH2;, ;CHdCH;, and so forth have been
reported among others. By attaching two bridging units
at the silicon it could be shown that even higher densities
of organic groups are accessible.18 Much attention was
also given to PMO materials prepared from precursors
containing a bridging phenyl ring.19,20 Such PMOs are
very interesting because the π-π interaction between the
phenyl rings can induce partial crystallization of the pore
walls.21 A limited number of papers describe PMO ma-
terials possessing ligand functionality.22 For instance, the
derivatization of the phenyl ring in the corresponding
PMOs with amino or sulphonium groups has been re-
ported.23-25 The authors have introduced the functional
group by postfunctionalization starting from the un-
modified phenylene PMO. Consequently, only a fraction
of the phenylene groups has been functionalized. Matsuoka
et al. have reported a very interesting study about the
attachment of an organometallic Cr(CO)3 fragment to the
phenyl ring of the PMO.24 Our group reported about a
system which allows for the transformation of the bridging
organic entity into all different sorts of functional groups:
PMOs containing a bridging 1,3-bis-siloxy benzene deriva-
tive functionalized with additional substituents in the
5-position, the so-called UKON materials.26,27
There are only a few papers reporting about PMOs
containing two distinct functional groups.23,28 Ozin et al.
have prepared a mesoporous organosilica material by co-
condensation of a PMO precursor with bridging ethylene
function and an alkoxysilane modified by a vinyl group23
Alcohol groups could be introduced later by the selective
hydroboration of the vinyl functionality. Corriu et al.
published an interesting approach in 2006.29 A PMO
precursor containing a bis-sulfide bridge was co-con-
densed with an alkoxysilane modified by a terminal
amine. The cleavage and oxidation of the S-S bridge
resulted in sulfonic acid groups located in the pore walls
and amine groups at the surface of the pores. It should be
noted that the authors had to use a 10-fold excess of
Si(OEt)4 for the preparation of their materials.
Some papers about catalytic applications of PMOs
have also been published.22,30 For instance Corma et al.
have reported in a nice paper a material containing a
chiral, bridging vanadyl salen complex and its application
in catalysis.31 However, the authors had to dilute their
precursors with Si(OEt)4, and therefore the materials
might rather be categorized under the co-condensation
category rather than PMOs.
We aim at the synthesis of mesoporous organosilica
materials related to the PMO type with walls constructed
from two different functional building blocks assembled
with geometrical precision. First, we describe the forma-
tion of surface bound, catalytically active metal com-
plexes with the benzoic acid groups in UKON2a (A). The
second functional group is aniline, and the corresponding
PMO UKON2d is reported here for the first time (B).
Then, the mesoporous materials containing both groups
embedded in the pore walls are described (C). Finally, the
influence of the presence of the two groups in a two-step
catalytic transformation is reported (D).
(12) Margelefsky, E. L.; Zeidan, R. K.; Davis, M. E. Chem. Soc. Rev.
2008, 37, 1118.
(13) Huh, S.; Chen, H. T.; Wiench, J. W.; Pruski, M.; Lin, V. S. Y.
Angew. Chem., Int. Ed. 2005, 44, 1826.
(14) Zeidan, R. K.; Hwang, S. J.; Davis, M. E. Angew. Chem., Int. Ed.
2006, 45, 6332. Zeidan, R. K.; Davis, M. E. J. Catal. 2007, 247, 379.
(15) Goettmann, F.; Grosso, D.; Mercier, F.; Mathey, F.; Sanchez, C.
Chem. Commun. 2004, 1240.
(16) Asefa, T.; MacLachan, M. J.; Coombs, N.; Ozin, G. A. Nature
1999, 402, 867. Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.;
Terasaki, O. J. Am. Chem. Soc. 1999, 121, 9611. Melde, B. J.; Holland,
B. T.; Blanford, C. F.; Stein, A. Chem. Mater. 1999, 11, 3302.
(17) Hatton, B.; Landskron, K.; Whitnall, W.; Perovic, D.; Ozin, G. A.
Acc. Chem. Res. 2005, 38, 305.
(18) Landskron, K.; Hatton, B. D.; Perovic, D. D.; Ozin, G. A. Science
2003, 302, 266.
(19) Yoshina-Ishii, C.; Asefa, T.; Coombs, N.; MacLachlan, M. J.;
Ozin, G. A. Chem. Commun. 1999, 2539.
(20) Guan, S.; Inagaki, S.; Ohsuna, T.; Terasaki, O. J. Am. Chem. Soc.
2000, 122, 5660. Goto, Y.; Inagaki, S. Chem. Commun. 2002, 2410.
Yang, Q. H.; Kapoor, M. P.; Inagaki, S. J. Am. Chem. Soc. 2002, 124,
9694. Kapoor, M. P.; Yang, Q. H.; Inagaki, S. Chem. Mater. 2004, 16,
1209.
(23) Asefa, T.; Kruk, M.; MacLachlan, M. J.; Coombs, N.; Grondey,
H.; Jaroniec, M.; Ozin, G. A. J. Am. Chem. Soc. 2001, 123, 8520.
(24) Kamegawa, T.; Sakai, T.; Matsuoka, M.; Anpo, M. J. Am. Chem.
Soc. 2005, 127, 16784.
(25) Ohashi, M.; Inagaki, S. Chem. Comm. 2008, 841.
(26) Kuschel, A.; Polarz, S. Adv. Funct. Mater. 2008, 18, 1272.
(27) Kuschel, A.; Polarz, S. Angew. Chem., Int. Ed. 2008, 49, 9513.
(28) Hunks, W. J.; Ozin, G. A. Adv. Funct. Mater. 2005, 15, 259. Cho,
E. B.; Kim, D.; Jaroniec, M. J. Phys. Chem. C 2008, 112, 4897.
(29) Alauzun, J.; Mehdi, A.; Reye, C.; Corriu, R. J. P. J. Am. Chem. Soc.
2006, 128, 8718.
(21) Inagaki, S.; Guan, S.; Ohsuna, T.; Terasaki, O. Nature 2002, 416,
304.
(22) Fujita, S.; Inagaki, S. Chem. Mater. 2008, 20, 891.
(30) Yang, Q.; Liu, J.; Zhang, L.; Li, C. J. Mater. Chem. 2009, 19, 1945.
(31) Baleizao, C.; Gigante, B.; Das, D.; Alvaro, M.; Garcia, H.; Corma,
A. Chem. Commun. 2003, 1860.

1474 Chem. Mater., Vol. 22, No. 4, 2010
Kuschel et al.
2. Experimental Section
ratios of the two sol-gel precursors 3,5-bis-(tri-iso-propoxysilyl)-
aniline and 3,5-bis-tri-iso-propoxysilylbenzoic acid were used. The
composition of the resulting mesoporous organosilica materials
can be described as (HOOC-PhSi2O3)1-x(H2N-PhSi2O3)x. The
treatment with a solution of the vanadyl cation is performed like
described for VO@UKON2a.
All reactions were performed with dried glassware under exclu-
sion of air and moisture if not indicated otherwise. All chemicals
were obtained from Aldrich and were dried and purified carefully.
1,3-Bis-tri-iso-propoxysilyl-5-bromobenzene, 3,5-bis-tri-iso-pro-
poxysilylbenzoic acid, and the UKON2a material were prepared
via procedures described in literature.26,27
PMOs Containing VO@COOPh and Ph-Br. The materials
were prepared similar to the method described for the synthesis of
VO@UKON2a. However, mixtures containing different ratios of
the two sol-gel precursors 1,3-bis-(tri-iso-propoxysilyl)-5-bromo-
benzene (1) and 3,5-bis-tri-iso-propoxysilylbenzoic acid were used.
The composition of the resulting, mesoporous organosilica mate-
rials can be described as (HOOC-PhSi2O3)1-x(Br-PhSi2O3)x. The
treatment with a solution of the vanadyl cation is performed like
described for VO@UKON2a.
Preparation of the Vanadyl-Containing PMO VO@UKO-
N2a. A total of 0.1 g of UKON2a is dried at 100 ꢀC under
vacuum for 15 h. A solution of 0.5 g of VO(acac)2 (1.9 mmol)
dissolved in 5 mL of chloroform is added. The mixture is
refluxed overnight, and the resulting green/yellow solid is
filtered off and is washed several times with 50 mL of chloro-
form. To remove the remaining VO(acac)2 from the pores the
material is stirred two times in 50 mL of CH2Cl2 overnight.
Washing is repeated with 20 mL of CH2Cl2/1 mL of pyridine.
After filtering and washing with small portions of CH2Cl2 the
material is dried under vacuum. FT-IR (ATR): 1049 (Si-O,
Si-C), 1445 (COO--symmetric valence vibration), 1523
(COO-asymmetric valence vibration), 1583 (aromatic C;C),
1706 (CdO), 2978 (C;H), and 3350 (O;H). EDX: V/Si/O ∼ 1/
16/40.
Catalytic Investigations Using the Bifunctional PMO Materials.
The material ([VO]0.5OOC-PhSi2O3)0.5(H2N-PhSi2O3)0.5 is given as
an example. A total of 35 mg of a bifunctional PMO material is dried
in vacuum. A mixture of freshly distilled benzyldioxolan (0.1 g, 0.69
mmol) and 0.75 g of ethyl cyano-acetate (0.69 mmol) is added. The
reaction mixture is heated to 62 ꢀC for 15 h. The material is filtered off,
and the products are analyzed. 1H NMR (400 MHz, CDCl3) δ/[ppm]:
1.24 (t, 3J = 7.2 Hz, H3C-CH2-O-ethyl cyano-acetate); 1.32 (t,
3J = 7.1 Hz, H3C-CH2-O-ethyl-2-cyano-3-phenylacrylate); 3.38
(s, bridging-CH2-ethyl cyano-acetate); 3.96 (“m”, O-CHH0-
CHH0-O-benzyldioxolan); 4.05 (“m”, O-CHH0-CHH0-O-
benzyldioxolan); 4.19 (q, 3J=7.2Hz, H3C-CH2-O-ethyl cyano-
acetate); 4.31 (q, 3J = 7.1 Hz, H3C-CH2-O-ethyl-2-cyano-
3-phenylacrylate); 5.73 (s, CH-benzyldioxolan); 7.3-8.1 (CH-arom
of all products and educts; detailed analysis not possible); 8.18 (s,
CH-ethyl-2-cyano-3-phenylacrylate); 9.94 (s, CHO-benzaldehyde).
The occurrence of the different reactants and products was monitored
Synthesis of 3,5-Bis-(tri-iso-propoxysilyl)-aniline (2). A total
of 0.18 mL of tBu3P (0.49 M solution in THF) is added to 0.05 g
of Pd(dba)2 in 5 mL of THF. The mixture is stirred for 30 min,
and 0.5 g of 1,3-bis-tri-iso-propoxysilyl-5-bromobenze (1) (0.88
mmol) in 20 mL of THF is added and then stirred for 30 min.
After the addition of 0.05 g of LiCl and 0.4 g of [(Me3Si)2N]2Zn
(1.06 mmol) the mixture is heated to 60 ꢀC for 72 h, filtered
through a short column of silica gel 60, and washed with
CH2Cl2. An orange gel is obtained after solvent removal. It is
dissolved in 20 mL of iPrOH, and 0.3 mL of Me3SiCl are added.
The mixtured was stirred overnight. The evolving hydrochloric
acid was trapped by pyridine. The product was obtained after
centrifugation with pentane. Column chromatography (silica
gel 60; CH2Cl2/AcOEt 30/1 f 20/1 f 10/1 f 0/1) was applied
for further purification. Finally 0.35 g (80%; 0.70 mmol) of a
slightly orange solid were obtained. 1H NMR (400 MHz,
1
by H NMR, and the qualitative integration of the characteristic
signals allowed the observation of the change in concentration of these
species. The total conversion in [%] is defined as the amount of
product III formed in the experiment in relation to the full transfor-
mation of the entire amount of reactant I into product III. Although
there are only small deviations regarding the specific surface area of
the different PMO materials prepared in the current work, possible
effects have been considered by dividing the value for the total
conversion through the internal surface area obtained from BET
evaluation of N2-physisorption data.
3
i
CDCl3): δ/[ppm]: 1.13 (d, 36H, J = 6.1 Hz, Pr-CH3); 4.18
(sept, 6H, 3J = 6.1 Hz, iPr-CH); 7.80 (s, 2H, o-arom. H); 7.94
(s, 1H, p-arom. H); 10.71 (s, 3H, NH3þ). 13C NMR (100.61
MHz, CDCl3): δ/[ppm]: 25.5 (iPr-CH3); 65.7 (iPr-CH); 129.5
(p-arom. C); 130.8 (Si-arom. C); 135.0 (o-arom. C); 141.7
(arom.-NH3þ). FT-IR (ATR): ν/[cm-1]: 1026 (Si-O); 1116
(Si-C); 1368-1577 (aliph. þ arom. C-C); 1621 (C-NH3þ);
2892-2980 (aliph þ arom. C-H); 3375, 3462 (NH3þ). EI-MS
(70 eV): m/z: 537 (Mþ) (very weak), 501 (-HCl) (main peak);
486 (-Me); 442 (-OiPr).
EPR Measurements and Analysis. All experiments were per-
formed at room temperature. A Bruker Elexsys E580 X-band
spectrometer equipped with a Super High-Q microwave-cavity
ER 4122SHQE was used. Performing modulated field sweeps
containing 1024 data points (total accumulated sweep time ca.
900 s) a modulation frequency of 100 kHz was used. Modulation
amplitude of 12.5 G has been chosen such that the signal was not
distorted. Spectrometer control and data postprocessing were
performed by the Bruker Xepr software. Since the line width of
the EPR spectra depends among other factors also on the
spin-spin interaction, the spectral second moment was used
to quantify this interaction. Therefore, sample x = 0.1 with a
spectrum SM(B) was compared to sample x = 1 featuring
spectrum SD(B). The difference of spectral second moments
ÆΔB2æ was calculated according to the following equation:32
Mesoporous Organosilica Material with Ph-NH2 as a Bridging
Organic Group (UKON2d). A total of 0.5 g of precursor (2) (0.93
mmol) and 0.295 g of Pluronic P123 are dissolved in 1.6 g of
EtOH. A total of 0.29 g of HCl (1 M) is added drop by drop. The
sols are aged in an open container for around one week. The
resulting monolithic pieces are dried in a vacuum at 100 ꢀC for
24 h. The template is removed by extraction first in 25 mL of
H2O and 25 mL of H2SO4 (concd) at 90 ꢀC and then in 25 mL
EtOH and 25 mL HCl (concd) at 60 ꢀC. Complete removal of the
Pluronic occurs within 2-4 days. The deprotonation of the Ph-
R
R
2
2
ðB -BFDÞ SDðBÞ dB
ðB -BFSÞ SMðBÞ dB
ÆΔΔB2æ ¼
-
R
R
þ
NH3 groups was achieved by treatment with triethylamine
followed by extensive washing.
SDðBÞ dB
SDðBÞ dB
ð1Þ
Bifunctional PMOs Containing VO@COOPh and Ph-NH2.
The materials were prepared similar to the method described for
the synthesis of UKON2d. However, mixtures containing different
(32) Steinhoff, H. J. Front. Biosci. 2002, 7, C97.

Article
Chem. Mater., Vol. 22, No. 4, 2010 1475
Chart 1. Modification of the PMO UKON2a with [V = O]2þ
where SD(B) is the absorption spectrum broadened due to
dipolar spin-spin interaction, SM(B) is the absorption spectrum
with negligible spin-spin interaction, and BFi and B are the first
spectral moments and the magnetic field, respectively. The
absorption EPR spectra were baseline corrected using Xepr
software (Bruker Biospin, Rheinstetten, Germany), and the first
and second moment were calculated numerically using MatLab
(MathWorks, MA, U.S.A.). To estimate the average nearest
neighbor distance all pair interactions to nearest neighbors have
been taken into account
X
where the second moments of each two-particle interaction are
X
1
dt6
ÆΔΔB2æ ¼
ÆΔΔBt2æ ¼ p
ð2Þ
t
t
added33 and where d is the spin-spin distance and p = 1.56 ꢀ
1060 T2 m-6 32
.
3. Results and Discussion
3.1. Modification of UKON2a with a Catalytically
Active Metal Center (A). The carboxylic group of
UKON2a should be able to coordinate to a range of
metal centers. For the current paper it is important that a
catalytically active species is used which coordinates
exclusively to the -COOH group. Other and undesirable
possibilities are either the amine function in UKON2d
later on (see paragraphs 3.2 and 3.3) or surface silanol
groups. The vanadyl cation [VIVO]2þ was selected be-
cause it can be utilized as a catalyst in various chemical
transformations.34 The successful binding of [VO]2þ can
be achieved if the pKa value of the corresponding acid of
the ligand attached to [VO]2þ is significantly higher than
that of benzoic acid (pKa = 4.22). After treatment of
UKON2a with an aqueous solution of [VO]SO4 (pKa =
-3), no uptake of [VO]2þ could be determined. The latter
result shows as well that the vanadyl cation exhibits no
tendency to bind to Si-OH functions located on the
surfaces of the pore walls. As opposed to this, the reaction
of UKON2a with [VO](acac)2 (pKa(acac) = 8.99) leads to
Figure 1. (a) EPR spectrum of [VdO]2þ@UKON2a. (b) The EPR line
broadening ΔBpp (black squares) and the average distance of two
carboxylic groups determined experimentally (gray filled circles) and in
comparison tothe values obtainedfromthe statistical model (graycrossed
circles) are shown as a function of the PMO composition (HOOC-
PhSi2O3)1-x(H2N-PhSi2O3)x.
the formation of the desired material with [VO]2þ at-
tached to the benzoic acid groups (Chart 1).
Even after diligent washing the signals that are typical
for the paramagnetic vanadyl cation implying the super-
position of hyperfine-structure lines of 51V, correspond-
ing to the perpendicular and parallel directions of the
g-tensor, can be seen in the electron paramagnetic reso-
nance (EPR) spectrum (Figure 1a).35 Furthermore, in
FT-IR spectra two additional bands at 1523 cm-1 and
(33) Abragam, A. The Principles of Nuclear Magnetism; Oxford
University Press: Oxford, 1961.
(34) Vedrine, J. C.; Millet, J. M. M.; Volta, J. C. Catal. Today 1996, 32,
115. Harvey, J. N.; Diefenbach, M.; Schroder, D.; Schwarz, H. Int. J.
Mass Spectrom. 1999, 182, 85. Baran, E. J. J. Inorg. Biochem. 2000,
80, 1.
(35) Gupta, S.; Khanijo, N.; Mansingh, A. J. Non-Cryst. Solids 1995,
181, 58. Carl, P. J.; Isley, S. L.; Larsen, S. C. J. Phys. Chem. A 2001,
105, 4563. Smith, T. S.; LoBrutto, R.; Pecoraro, V. L. Coord. Chem.
Rev. 2002, 228, 1. Zamotrin.Ea; Kulemin, N. P. Izvestiya Vysshikh
Uchebnykh Zavedenii Fizika 1969, 143.

1476 Chem. Mater., Vol. 22, No. 4, 2010
Kuschel et al.
Figure 2. Structural data of UKON2d: TEM migrograph (a), SAXS pattern (b), N2 physisorption isotherm (adsorption and desorption), and BJH pore-
size distribution function (c and d).
Chart 2. Preparation of the New PMO Material UKON2d Containing Aniline
1445 cm-1 appear which are typical for carboxylate
acting as a bidentate ligand (spectra shown in the Sup-
porting Information SI-1). No changes can be observed
for the intensity of the band at 3350 cm-1 associated with
the SiOH groups present in the material. It is also
important to note that there are no signals present
correlating to the acac-ligand which has been cleaved
quantitatively. It can be concluded that [VO]2þ coordi-
nates exclusively to the benzoic acid groups in UKON2a.
According to a quantitative evaluation of energy disper-
sive X-ray spectroscopy (EDX) data 25% of the benzoic
groups in UKON2a are involved as ligands whereas two
carboxylic groups bind to one vanadyl group (Chart 1).
Because we know from previous results that the amount
of functional groups which is accessible on the surface of
the pore wall of UKON2a is also of the order of 25%,27
the statement that the entire surface is functionalized with
[VO]2þ is reasonable.
of 3,5-bis(tri-iso-propoxy-siliyl)-bromobenzene with Zn-
(hmds)2 (see Chart 2). The resulting PMO was prepared
under acidic conditions using the true liquid crystal ap-
proach with Pluronic P-123 as a template (Figure 2a).7,36
Transmission electron microscopy (TEM) data show a
highly ordered material with cylindrical, hexagonally
aligned pores (Figure 2a). However, it should be noted that
TEM is only a local analytical technique. Therefore, small-
angle X-ray scattering (SAXS) was applied as an indepen-
dent method for the characterization of the structure of
UKON2d. Three clear maxima (s100 = 0.091 nm-1, s110
=
0.157 nm-1, s200 = 0.183 nm-1) are present in the SAXS
pattern (Figure 2b) indicating a pore to pore distance of
Dpore-pore ≈ 10 nm. Even a fourth, weak diffraction maxi-
mum can be identified at s = 0.248 nm-1. Such a high
number of high order reflections are seen if mesoporous
materials possess a well structured pore system. Therefore,
the SAXS data confirm the conclusions drawn from TEM
3.2. New PMO Constructed from Aniline (B). The
needed sol-gel precursor, 3,5-bis(tri-iso-propoxysilyl)-
aniline, can be obtained from the Pd-catalyzed coupling
(36) Polarz, S.; Antonietti, M. Chem. Commun. 2002, 2593. Goeltner,
C. G.; Antonietti, M. Adv. Mater. 1997, 9, 431. Goeltner, C. G.; Berton,
B.; Kraemer, E.; Antonietti, M. Adv. Mater. 1999, 11, 395.

Article
Chem. Mater., Vol. 22, No. 4, 2010 1477
Figure 3. Solid-state NMR data for the new PMO material UKON2d: 13C (a; black graph), 29Si (a; gray graph) and 15N spectra (b). In (b) a comparison to
the spectrum of the precursor 3,5-bis(tri-iso-propoxysilyl) aniline (gray line; measured in C6D6) is shown.
measurements. Next, N2 physisorption data was acquired as
shown in Figure 2c. The isotherm of UKON2d possesses a
shape which is typical for mesoporous solids with a capillary
condensation step at p/p0 ≈ 0.57.37 The material possesses
an internal surface area of 625 m2/g determined by the BET
method.38 The steady increase of adsorbed volume in the
pressure range below p/p0 < 0.4 indicates the presence of
large micropores (Dpore ≈ 2 nm). The pore-size distribution
function of UKON2d was determined from the BJH eval-
uation of the adsorption branch of the isotherm.39 A pore-
size maximum DP = 4.9 nm is found (Figure 2d). The pore
size determined from N2 physisorption is in good agreement
with TEM data. In combination with the SAXS data it can
be concluded that pore walls of UKON2d are relatively
thick (Dthickness ≈ Dpore-pore(SAXS) - Dpore(N2) ≈ 5 nm).
The latter finding can also be confirmed by TEM images. It
is seen that the pore walls have extensions similar to those of
the pores (see Figure 2a).
The chemical nature of UKON2d was investigated via
solid-state MAS NMR spectroscopy. The 29Si NMR
spectrum of UKON2f (Figure 3a) contains the character-
istic three signals at δ = -70 ppm for (HO)2RSi(OSi) =
T1, -78 ppm for (HO)RSi(OSi)2 = T2, and -86 ppm for
RSi(OSi)3 = T3.19 Signals indicating the presence of pure
silica (SiO2) are not seen demonstrating that the precursor
has been used in an undiluted form and that the Si-C
bonds in UKON2d are stable. In the 13C NMR spectrum
only the signals for Carom-NH2 (δ = 147 ppm) and the
remaining Carom (δ = 135 ppm, 139 ppm) are observed
(Figure 3a). In addition 15N NMR data were recorded
(Figure 3b). The chemical shift of the nitrogen atom (δ =
55.8 ppm) in the precursor (2) fits very well to the value for
pure aniline (δ = 56.1 ppm). Comparison to the solid-
state NMR spectrum of UKON2d indicates clearly that
its composition can be described as NH2C6H3Si2O3 (see
Chart 2).
3.3. PMOs Containing Two Different Functional
Groups (C). An even higher level of complexity can be
achieved if a PMO material is prepared containing two
very different functional groups like in UKON2a and
UKON2d organized on the nanoscale. The desired bi-
functional PMOs have been synthesized from mixtures of
the precursors 3,5-bis(tri-isopropoxysilyl)benzoic acid
and 3,5-bis(tri-isopropoxysilyl) aniline in different ratios.
Materials with the formal composition of a copolymer
(HOOC-PhSi2O3)1-x(H2N-PhSi2O3)x with (x = 0.0, 0.1,
..., 1.0) can be obtained. This can be shown by FT-IR
spectroscopy and solid-state NMR spectroscopy. The
corresponding data are shown in Figure 4. The main
feature in the fingerprint region of the FT-IR spectrum of
UKON2a is the band at 1700 cm-1 correlating to the
carboxylic group. The characteristic vibrations for the
-NH2 function in UKON2d are found at 1623 cm-1 and
1387 cm-1 (see also Supporting Information SI-2). The
band at 1700 cm-1 becomes gradually weaker and the
bands correlating to aniline stronger with increasing x.
13C-solid-state NMR spectroscopy was applied as an
independent technique to support the conclusions made
from FT-IR. Also in 13C NMR it is observed that the
signal at δ = 173 ppm for the carboxy group becomes
smaller with increasing x.
However, because the FT-IR and solid state NMR
methods average over the entire sample there is not yet
any information about the distribution of the two differ-
ent groups on the nanoscale. Four different scenarios are
possible which are depicted in Scheme 1. Obviously, the
macrophase separation into a physical mixture of two
different mesoporous organosilica materials represents
an undesired event (see Scheme 1a). It could occur if the
two precursors are associated with very different hydro-
lysis rates or very different solubility. If the latter two
properties are only slightly different from each other it
could come to microphase separation (Scheme 1b). It can
be seen that the majority of the pore surfaces would be
(37) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity,
4th ed.; Academic Press: New York, 1982; Vol. 2. Rouqerol, J.; Avnir,
D.; Fairbridge, C. W.; Everett, D. H.; Haynes, J. H.; Pernicone, N.;
Ramsay, J. D.; Sing, K. S. W.; Unger, K. K. Pure Appl. Chem. 1994, 66,
1739.
(38) Brunauer, S.; Emmet, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60,
309.
(39) Barret, E. P.; Joyner, L. G.; Halenda, P. H. J. Am. Chem. Soc. 1951,
73, 373.

1478 Chem. Mater., Vol. 22, No. 4, 2010
Kuschel et al.
Figure 4. FT-IR spectra (a) and 13C-solid state NMR spectra (b) of the bifunctional PMOs (HOOC-PhSi2O3)1-x(H2N-PhSi2O3)x.
Scheme 1. Different Scenarios for the Organization of the Two Functional Groups at the Nanoscalea
a (a) Macrophase separation, (b) microphase separation, (c) pairing of amine and carboxylic acid, and (d) random mixture.
either covered with carboxylic acid or amine groups. As a
result the pore surfaces would not exhibit bifunctional
character. Only if phase separation does not occur, one
can expect that the surfaces will be characterized by the
presence of both functionalities. However, even under
these circumstances two further possibilities need to be
considered. Either the distribution of the two building
blocks is entirely random, or it could be that an attractive
interaction between the benzoic acid and aniline building
blocks occurs. It should be noted that the event of
attraction is less likely because at the pH used for the
preparation of the materials (pH = 0; see Experimental
Section) benzoic acid and aniline will both be protonated
(-COOH and -NH3þ). Nevertheless, both possibilities
will be discussed (Scheme 1 c,d). A cross section of one
mesopore and the surrounding pore wall is considered as
a model. The pore wall has been divided into equal grids
representing the individual building blocks. Then, by
using a random generator either an aniline (red) or a
benzoic acid (blue) building block is assigned to each grid.
The latter method leads to the purely statistical situa-
tion, and the result of this scenario is shown in Scheme 1d.

Article
Chem. Mater., Vol. 22, No. 4, 2010 1479
The following method has been used for the simulation of
the interaction between the aniline and the benzoic acid
building blocks. In the first step one grid has been
assigned randomly. Then, exactly one grid in direct
contact to it is set to the second character. The procedure
is repeated until all grids are assigned. The result of the
latter scenario is shown in Scheme 1c. It is seen that the
pores possess bifunctional character for both scenarios
but that the distribution of groups in the pore walls is
different. The statistic clustering of equal functional
groups is much less likely for an interaction between
aniline and benzoic acid. It can be concluded that each
of the four shown scenarios is connected to a certain
composition dependency of the distance distribution
functions of the functional groups in the material. Find-
ing the experimental composition dependency of the
distance functions should enable us to distinguish bet-
ween the different scenarios.
Information about the distances between the carboxy-
late groups as a function of x can be acquired by studying
the magnetic interaction of the paramagnetic vanadyl
centers via EPR spectroscopy. First, the organosilica
material containing only aniline (x = 0; UKON2d) was
treated with a [VO](acac)2 solution. After washing (see
Experimental Section) there were no signals found in the
EPR spectrum which are characteristic for [VO]2þ. It is
proven that [VO]2þ does not coordinate to the amine
function in UKON2d. Therefore, if the same procedure is
nates the possibility of a microphase separation since here
the average distance between neighboring carboxylate
groups would be more or less independent from composi-
tion (see Scheme 1b). Finally, a quantitative correlation
between DC-C and the composition parameter x has been
developed from a statistical evaluation of the grid model
mentioned above. An extract of the surface is represented
by a 10 ꢀ 10 grid in which each cell represents one single
PMO building block. The carboxy function is randomly
assigned to the cells accounting for the predefined com-
position x as a constraint (see also Supporting Informa-
tion SI-4), and the distances of the cells containing the
-COOH groups are determined and averaged. The pro-
cedure is repeated 105 times for each x leading to an
averaged value for DC-C. The resulting dependency of
D
C-C as a function of x is also shown in Figure 1b. The
experimental values for DC-C(x = 0) and DC-C(x = 0.9)
are reproduced with high precision. Because there is a
close agreement between the model and the experimental
data, it can be assumed that the building blocks are
statistically distributed in the pore walls. The scenario
involving the interaction between aniline and benzoic
acid building blocks (Scheme 1c) is difficult to exclude
for a certainty, but the presented arguments are in favor
of the purely statistic assembly (Scheme 1d). In addition,
a pair formation of aniline and benzoic acid should be
seen as a change in the vibration of the carboxy group in
FT-IR spectroscopy. There are no changes of the band at
1700 cm-1 seen for the bifunctional materials in compar-
ison to UKON2a (see Figure 4a).
applied to the bifunctional materials (HOOC-PhSi2O3)1-x
-
(H2N-PhSi2O3)x with x > 0 one can be sure that [VO]2þ is
located exclusively at the benzoic acid entities. The ma-
terial with x = 0.5 was investigated as a representative
case with scanning electron microscopy (SEM) in back-
scattering modus (data shown in Supporting Information
SI-3). It is seen that the distribution of the elements, in
particular the vanadium centers, is absolutely homo-
geneous. The latter observation rules out the macrophase
separation scenario (Scheme 1a). Because the benzoic
acid groups become more and more diluted with increas-
ing x, the density of [VO]2þ is expected to decrease as well.
Interestingly, not only a decrease of the EPR signal
intensity is found, but in addition a monotonic depen-
dency of the line-broadening ΔBpp of the EPR signals
caused by spin-spin coupling as function of x has been
observed (dependency shown Figure 1b). The spin-spin
coupling is reduced because the distance between the
[VO]2þ centers becomes larger. The average distance of
[VO]2þ ions present at the pore-surface DC-C can be
obtained from evaluating the second momentum of the
EPR data (see Experimental Section). DC-C(x = 0) =
0.8 nm and DC-C(x = 0.9) = 1.6 nm have been deter-
mined. The experimental value for the distance between
two vanadyl groups, respectively, two carboxylate groups
in UKON2a (x = 0) fits well to the expected value
obtained from force-field calculations DC-C(th.) = 0.8 nm
(see Supporting Information SI-4). Therefore, EPR
spectroscopy seems to be a relaible technique for the
experimental determination of the required distances.
Furthermore, the distinct dependency of DC-C on x elimi-
To the best of our knowledge, this is the first experimen-
tally underpinned picture for the structure of the surfaces of
amorphous mesoporous organosilica materials with mole-
cular scale precision. Furthermore, the results suggest that
the properties of the surfaces can be tuned by adjusting the
composition parameter x. One possible application for such
materials with adjustable bifunctional surfaces could be in
catalysis which is described in the next paragraph.
3.4. Activity in Consecutive Chemical Transformations
(D). It is an interesting question if the presence of both
different functional groups at the surface of the meso-
pores might lead to cooperative effects regarding che-
mical transformations proceeding inside the pore
system. Others have also addressed a similar task,
however, not with PMO materials.12-15 The coordina-
tion compounds of [VO]2þ are valuable Lewis acids,
and they can, for instance, catalyze the cleavage of an
acetal-like benzyldioxolan (see Figure 5; steps I f II).40
Aniline on the other hand due to its Lewis base
characteristics is known to catalyze the Knoevenagel
condensation, for instance, of benzaldehyde with
cyan-acetic ethyl ester (see Figure 4; steps II f III).41
(40) Kantam, M. L.; Neeraja, V.; Sreekanth, P. Catal. Commun. 2001, 2,
301.
(41) Fiuza, S. M.; VanBesien, E.; Milhazes, N.; Borges, F.; Marques,
M. P. M. J. Mol. Struct. 2004, 693, 103. Esteves, M.; Siquet, C.;
Gaspar, A.; Rio, V.; Sousa, J. B.; Reis, S.; Marques, M. P. M.; Borges, F.
Arch. Pharm. Chem. Life Sci. 2008, 341, 164. Gaspar, A.; Garrido,
E. M.; Esteves, M.; Quezada, E.; Milhazes, N.; Garrido, J.; Borges, F.
Eur. J. Med. Chem. 2009, 44, 2092.

1480 Chem. Mater., Vol. 22, No. 4, 2010
Kuschel et al.
Figure 5. Schematical representation of a two-step chemical transformation (acetal cleavage I f II; Knoevenagel condensation II f III) performed in the
bifunctional PMO. The surfaces of the pores have been systematically varied from aniline rich to aniline poor (and benzoic acid rich) architecture.
The mentioned reactions are also suitable because the
individual reaction rate constants are comparable to
each other. If the reaction rate constants for one step
would be significantly higher, the slower step and the
associated catalytic group would dominate the beha-
vior of the entire system.
It is interesting to study the mentioned consecutive
reaction sequence I f II f III in dependence of the
architecture of the surfaces containing both groups
VO2þ@OOCPh and Ph-NH2. All samples have been
treated under identical conditions (reaction temperature,
conversion time, etc.). For details see the Experimental
Section. It is important to note that despite the fact that
there are only small deviations regarding the specific
surface area of the different PMO materials prepared in
the current work, effects caused by such minor differences
Figure 6. Total conversion normalized against the specific surface area
(Conv.(I-III)BET) of the respective materials for the mesoporous organo-
silicas as a catalyst containing either aniline and the vanadyl cation
have been excluded by dividing the value for the total
conversion through the internal surface area obtained
attached to benzoic acid (blue graph) or the latter and bromobenzene
(red graph). The probability P accessing two different functional groups
(dashed gray line) as a function of the materials composition x is also
shown.
from BET evaluation of N2 physisorption data. First, the
reference experiments with the monofunctional PMO
materials (x = 0; x = 1) have been performed. No
noteworthy conversion was observed when the either
the pure material UKON2a or pure UKON2d (see
Figure 6; x = 1) was used. When VO@UKON2a was
present as a catalyst product II could be separated from
the reaction mixture in low yield (see Figure 6; x = 0).
A physical mixture (compare to scenario Scheme 1a) of
VO@UKON2a and UKON2d was also tested. How-
ever, only a vanishing amount of product III has been
formed. The latter results demonstrate that the spatial
proximity of the two catalytic functions can be a deciding
factor.

Article
Next we wanted to make sure that any observed
Chem. Mater., Vol. 22, No. 4, 2010 1481
distribution function is given by the binomial distribution
function
effects for the bifunctional PMO materials are related
to the differing organization of the functional groups
on the nanoscale and not any structural differences.
Therefore, all materials were characterized by SAXS
and N2-physisorption analysis (see Supporting Infor-
mation SI-3). The main reflection is found at a scatter-
ing vector q ≈ 0.70 nm-1 for all samples. Only one reflex
can be observed in SAXS which indicates that the
mesoporous organosilica materials possess a wormhole
pore system. There is a small variation in pore size of
Dp = 4.5 ( 0.3 nm. The resulting differences in internal
surface area have been considered by normalizing the
total conversion to the specific surface area of the
respective material. The results are shown in Figure 5.
A maximum in total conversion is seen for a material
which is composed of benzoic acid and aniline in equal
shares (x = 0.5). It is important to note that in principle
both steps (I f II and II f III) can be catalyzed by a
Lewis acid, but it is expected that the rate for the
Knoevenagel condensation will be relatively small.
Therefore, an analogous series of benzoic acid contain-
ing materials has been prepared in which aniline is not
the second building block but the sol-gel precursor
containing a bridging bromo-benzene has been used
(see Experimental Section) resulting in materials with
the composition (VO@OOC-PhSi2O3)1-x(Br-PhSi2O3)x.
The same catalytic tests were performed. It can be seen
that also the mesoporous organosilica materials contain-
ing only VO@OOCPh as the catalytically active group
lead to a minor conversion I f II f III, more or less
independent of x (Figure 6). Furthermore, it can be
deduced from NMR spectroscopy of the reaction mix-
tures that the conversion I f II drops almost linear with x
(see Supporting Information SI-5). As a result of the same
arguments, it can be expected that there would be a
monotonic increase of the conversion for step II f III
with x for materials containing aniline and bromoben-
zene (Br-PhSi2O3)1-x(NH2-PhSi2O3)x. Thus, it is ob-
vious that the conversion for the consecutive steps I f
II f III catalyzed by the bifunctional materials contain-
ing benzoic acid and aniline shown in Figure 6 is not a
simple superposition of step I f II catalyzed by
VO@OOCPh and step II f III catalyzed by NH2Ph.
The latter findings raise a new question. Is there some
special kind of interaction between the two functional
groups on the surface of the mesoporous materials,
respectively a cooperative effect, or is the variation in
diffusion time from the first catalytic center ([VO]2þ) to
the second depending on x between them the deciding
factor?
ꢀ ꢁ
n
P ¼
pkð1 - pÞn -k w Pꢁp ꢁ f ðxÞ
k ¼ 1; n.k
k
which becomes proportional to p (= the probability for
finding a benzoic building block; determined by the
composition x) when k = 1 meaning that one contact
to the aniline building block is sufficient for the conver-
sion (steps I f II f III) and n . k. The probability P as a
function of x is also plotted in Figure 6. It possesses a
maximum at x = 0.5. Although the mentioned statistical
model is relatively rough it can be seen that it explains the
maximum in total yield at x = 0.5. Furthermore, there is a
relatively good qualitative compliance regarding the depen-
dence on x especially for x > 0.5. It can be concluded that
there is no special chemical interaction between aniline and
the vanadyl cation. The cooperativity is caused by maximiz-
ing the probability of accessing both functional groups
during diffusion through the pore system.
4. Conclusion
UKON2a is a mesoporous organosilica material con-
structed from a building block containing a bridging
benzoic acid group.26,27 The coordination of the vandyl
cation [VO]2þ to the carboxylic groups in UKON2a is
presented.
A new PMO material constructed from aniline has been
reported (UKON2d). It was prepared from a sol-gel
precursor containing a bridging aniline unit. The result-
ing material possessed a well structured pore system of
cylindrical channels aligned in a hexagonal packing.
Because aniline is one of the most important entities
known in organic chemistry we believe that UKON2d
could be a bifurcation point for more advanced materials
in the future.
Next, bifunctional materials containing both building
blocks (aniline and benzoic acid) have been prepared. It
could be shown that the distribution of the two groups in
the pore walls and on the surfaces is purely statistic. We
obtained a refined picture of the molecular structure of
the surfaces of such mesoporous organosilica materials.
The results indicate that one can achieve a fine-tuning of
the chemical character of the pore surfaces by adjusting
the composition of the precursor mixture used for the
preparation of the materials.
One possible application of such bifunctional meso-
porous organosilica materials was probed. The cooperati-
vity of the two groups regarding a two-step chemical
process was tested. It could be shown that a maximum in
activity measured as yield per time was realized for a
material composed of 50% benzoic acid groups and 50%
aniline groups. However, the observed effects were not
caused by any special interaction between the two func-
tional groups. The results could be explained very well by
a simple statistical model which considers the probability
Because it could be proven that the distribution of the
benzoic acid groups is purely stochastic it can also be
concluded that this is true for the aniline building blocks
as well. Then, because diffusion through the pore system
is a stochastic element the mentioned reaction becomes a
Bernoulli experiment reflecting the probability distribu-
tion function that a species moving inside the pore system
accesses both functional groups. The latter probability

1482 Chem. Mater., Vol. 22, No. 4, 2010
Kuschel et al.
for accessing both functional groups during the course of
diffusion through the pore system.
Supporting Information Available: SI-1, Additional analytical
data for [VO]2þ@UKON2a. SI-2, Additional analytical data
for UKON2d. SI-3, Additional analytical data for bifunctional
PMO materials. SI-4, Statistical evaluation of the bifunctional
materials. SI-5, Investigation of steps I f III catalyzed by
(VO@OOC-PhSi2O3)1-x(Br-PhSi2O3)x (PDF).This material is
available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. We thank the Zukunftskolleg University
of Konstanz and the Deutsche Forschungs-gemeinschaft
(SP 780/6-1, DR743/2-1) for funding. We thank M. Spitzbarth
and K. Guth for the EPR measurements.