Sol-Gel Ormosils Doped with TEMPO as Recyclable Catalysts for the Selective Oxidation of Alcohols

Article abstract of DOI:10.1002/1615-4169(200202)344:2<159::AID-ADSC159>3.0.CO;2-Q

Organically modified silicas doped with TEMPO prepared via the sol-gel method are highly recyclable catalysts of the selective Montanari-Anelli oxidation of 1-nonanol; They show a notable "positive feedback" effect of matrix alkylation on the catalyst activity which is typical of doped sol-gel materials and markedly differentiates the behaviour of these materials from that of analogous silica-supported TEMPO.

Full text of DOI:10.1002/1615-4169(200202)344:2<159::AID-ADSC159>3.0.CO;2-Q

COMMUNICATIONS
Sol-Gel Ormosils Doped with TEMPO as Recyclable Catalysts
for the Selective Oxidation of Alcohols
Rosaria Ciriminna,a Carsten Bolm,b Thomas Feyb, Mario Pagliaro*,a
a
Istituto per lo Studio del Materiali Nanostrutturati Sezione di Palermo Consiglio Nazionale delle Ricerche,
via Ugo La Malfa 153, 90146 Palermo, Italy
Fax: ( ‡ 39)-091-680-9399, e-mail: pagliaro@ictpn.pa.cnr.it
b
Institut f¸r Organische Chemie der RWTH Aachen, Professor Pirlet Stra˚e 1, 52056 Aachen, Germany
Received: December 17, 2001; Accepted: February 15, 2002
oxidation of glucose 1-phosphate in water;[5] oligo-
amine-supported TEMPO, in its turn, could neither be
Abstract: Organically modified silicas doped with
TEMPO prepared via the sol-gel method are highly
employed in chlorinated solvents (where it readily
recyclable catalysts of the selective Montanari-Anel-
dissolves) nor in the presence of metals such as Ru
li oxidation of 1-nonanol; They show a notable
that coordinate to the amine moieties of the polymer.[2b]
∫positive feedback∫ effect of matrix alkylation on
Finally, hydrophilicMCM-41-grafted TEMPO oculd
the catalyst activity which is typical of doped sol-gel
not be used above pH 8 when silica started to dissolve
materials and markedly differentiates the behaviour
of these materials from that of analogous silica-
supported TEMPO.
leaching the radicals in solution.[2a]
Now, we report that organically modified silica gels
(ormosils) doped with TEMPO prepared under mild
alkaline conditions are selective catalysts in the Mon-
tanari-Anelli oxidation of the aliphaticalcohol 1-non-
anol showing a remarkable stability in their selective
activity. Several different catalytic ormosils were pre-
Keywords: heterogeneous catalysis; ormosils; selec-
tive oxidation; sol-gel; TEMPO.
Recently, we have independently heterogenised the pared through reductive amination of 4-oxo-TEMPO
stable organicnitroxyl radical TEMPO (2,2,6,6-tetra-
with 3-aminopropyltrimethoxysilane in the presence of
methylpiperidine 1-oxyl radical) over different silica NaBH3CN in methanol (see Experimental Section),
matrices in order to obtain recyclable catalysts for followed by the sol-gel polycondensation process of
alcohol oxidations showing the well-known high selec- suitable silicon alkoxide precursors. The N2-BET pore
tivity of homogeneous TEMPO catalyses both in the volumes and surface areas (Table 1) of the resulting gels
Montanari Anelli oxidation of alcohols to carbonyl are typical of mesoporous sol-gel ormosils obtained
compounds[1a,1b] and in the NaOCl/Br oxidation of under alkaline conditions showing a broad pore size
carbohydrates to carboxylates in water.[1c]
With similar aims, TEMPO has also been anchored to
distribution (80 300 ä).
Based on our most recent results with silica-supported
the mesoporous channels of silica MCM-41[2a] and TEMPO,[1b] in order to investigate the catalytic activity
immobilised at the surface of the oligomeric triazine and stability of the new ormosils, we carried out the
Chimassorb 944 by oxidation of its amine moieties.[2b] oxidation of substrate 1-nonanol in several consecutive
‡
The cyclic nitrosonium ion TEMPO is, in fact, a highly runs (see Experimental Section) at 0 8C with a reaction
active and versatile selective catalytic oxidant which is time of only 15 minutes, thus expecting an incomplete
readily formed in situ by several primary oxidants under conversion and in case of absolute stability a constant
mild conditions.[3] However, in contrast to homogeneous yield in all subsequent runs. The results, however, were
TEMPO that is equally applicable to the conversion of a very surprising (Figure 1).
variety of alcohols in different reaction media,[3] heter-
Hence, the unmodified silica-encapsulated TEMPO
ogeneous TEMPO often fails to show analogous versa- (SG-TEMPO-0) lost half of its initial high activity upon
tility. This is unfortunate since the low product volumes just 1 oxidative run, its colour turning to white as a clear
of the fine chemicals industry and the new combinatorial indication of material dissolution and leaching of the
applications of solid-phase syntheses[4] do indeed re- radical in the reaction mixture. However, it was
quire such a high degree of versatility. Hence, for sufficient to use the partly (25% of Si atoms substituted)
instance, while TEMPO anchored on commercial methylated ormosil SG-TEMPO-1 to observe a re-
aminopropyl-silica was found to be sufficiently recycla- markable stability in the activity along with an increment
ble in the Montanari Anelli oxidation of various in the intermediate yields upon the first 3 runs in which
alcohols,[1a] an analogously immobilised catalyst was the catalyst was tested that, remarkably, was also shown
unstable and practically inactive in the corresponding by the fully (100%) methylated gel SG-TEMPO-2
Adv. Synth. Catal. 2002, 344, No. 2
¹ VCH Verlagsgesellschaft mbH, D-69451 Weinheim, 2002 1615-4150/02/34402-159 163 $ 17.50+.50/0
159

Rosaria Ciriminna, et al.
COMMUNICATIONS
Table 1. Textural properties and composition of the catalytic ormosils doped with TEMPO.
Organosilane[a]
Si:MeOH
DF[b] [mmol/g]
SBET [m2/g]
Vp [mL/g]
[c]
Catalyst
SG-TEMPO-0
SG-TEMPO-1
SG-TEMPO-2
SG-TEMPO-E
SG-TEMPO-P
0% MTMS
25% MTMS
100% MTMS
75% ETMS
75% PTMS
1:3
1:3
1:2
1:3
1:3
0.27
0.28
0.30
0.29
0.25
217
213
267
160
75
1.24
1.19
1.25
0.74
0.38
[a]
Molar percentages relative to TMOS used in each gel preparation.
[b]
[c]
Degree of functionalisation based on the assumption of complete radical encapsulation during the sol-gel catalyst synthesis.
Silica-supported TEMPO used for comparison (SG-TMP-O) has BET surface area SBET ˆ 350 m2/g.[1b]
Figure 1. Yields in the Montanari Anelli oxidation of 1-nonanol to give nonanal in the presence of silica-supported TEMPO
(SG-TMP-O, front row), and of sol-gel ormosils doped with TEMPO [SG-TEMPO-1 is 25% and SG-TEMPO-2 is 100%
methylated (middle and back row, respectively)].
which after 6 runs was 8% more active than the fresh reaction run. Interestingly, the transparent microporous
catalyst and showed the same pronounced stability silica glass doped with TEMPO prepared under acidic
(Figure 1) of SG-TEMPO-1. A similar trend of incre- conditions[1c] employed in the same oxidation protocol
mental reaction rates was also shown by the partly of 1-nonanol was stable (no leaching or catalyst
(75%) propyl-modified silica SG-TEMPO-P which had modification) but only modestly active (8% yield after
an activity comparable to that of SG-TEMPO-2, while 15 min).
the analogous (75%) ethyl-modified gel SG-TEMPO-E
To investigate the effect of reaction on the catalyst
was less active than SG-TEMPO-1 but, again, recyclable structure, EPR spectra of SG-TEMPO-2 before and
and stable. In each catalytic run mediated by the doped after 7 consecutive oxidation runs (a and b in Figure 2,
ormosils no other products besides the carbonyl were respectively) were recorded.
detected in the reaction mixture and there was no
The spectra show that the distance between the outer
leaching of encapsulated radicals in solution as shown lines (related to the coupling constant of the interaction
also by the lack of further reactivity of the reaction of the unpaired electron with the spin of 14N nucleus)[6]
filtrate obtained upon separation of the catalyst from increases slightly, while the lesser increment (20%) of
the reaction mixture shortly (7 minutes) after the the line widths (related to the isotropicrotation
beginning of the oxidation.
correlation times)[6] indicates an analogous increase of
Silica-supported TEMPO (SG-TMP-O, 500 mmol/g the viscosity sensed by the spin label within the sol-gel
radical, obtained by impregnation of commercial prop- cage, likely to be due to water adsorption within the
yl-aminosilica with 4-oxo-TEMPO)[1a] was thus used for cages (resulting in more hydrogen bonding between
comparison in the same conversion of 1-nonanol to silanols and water) upon its prolonged use in the CH2Cl2
nonanal and, again,[1b] its activity measured by the -H2O reaction medium (see below for a consistent
intermediate yields continuously diminished upon every interpretation of the FTIR spectra). The reduction in
160
Adv. Synth. Catal. 2002, 344, 159 163

Sol-Gel Ormosils Doped with TEMPO as Recyclable Catalysts
COMMUNICATIONS
1
The overlapping bands at 2930 and 2871 cm in the
spectra in Figure 3 are assigned to the asymmetrical and
symmetrical vibration of CH2 groups in the propyl
chains[2a] carrying the bound nitroxyl radicals and clearly
indicate that these groups are unmodified prior to and
after reaction; the broad signals overlapping in the 1085
1
1100 cm region are, in turn, due to the Si-O-Si
[7]
asymmetricstrethcing
and show that the silica gel
polymeric backbone is unaffected by the prolonged use
in the reaction system. Interestingly, the only notable
difference in the FTIR spectra in Figure 3 is shown by
1
the enhancement of the signal at 1635 cm (which is due
to H-O-H bending)[7] and by the related enhancement of
1
the broad signal in the 3000 3600 cm region due to the
stretching vibrations of the H-bonded silanols with
water,[7] thus confirming the uptake of water molecules
in the sol-gel cages during the reaction (in accord with
the results of the EPR experiments given above).
Finally, the difference in the EPR spectra of silica-
encapsulated and silica-supported TEMPO shows how
different the sol-gel encapsulation of dopant molecules
is as compared to adsorption at an external surface: the
silica-supported TEMPO gives a broad low signal
(Figure 2, d) reflecting extensive intermolecular spin
interactions among the radicals anchored at the surface,
while the triplet (Figure 2, a) shown by the encapsulated
radicals reflects the isolation of the radicals within the
sol-gel cages. Incidentally, it may also be noted that the
100% methylated gel SG-TEMPO-2 yields a 3-times
more intense peak (Figure 2, a) compared to the same
amount of TEMPO encapsulated in unmodified SiO2
(Figure 2, c), thus indicating a pronounced protection of
the radicals against quenching that is probably related to
a lesser degree of free radical formation[8] during the sol-
gel synthesis of the ormosils as compared to the
synthesis of silica using TMOS alone.
These results indicate that in the heterogeneous
Montanari Anelli alcohol oxidation mediated by sol-
gel ormosils a cooperative mechanism may be in action
through which the more alcohol (and aldehyde) mole-
cules are adsorbed (and formed) within the inner cage
surface, the faster is the diffusion of the reactant and
product molecules through the porous network. Nota-
bly, an analogous explanation was recently proposed to
Figure 2. EPR spectra of sol-gel ormosil doped with SG-
TEMPO-2 prior to (a) and after (b) 7 consecutive reaction
runs in the Montanari-Anelli oxidation of 1-nonanol. The
spectra of sol-gel (c) and commercial (d) aminopropyl-silicas
doped with TEMPO are also shown.
intensity of the resonance peak in the used catalyst to explain the results of the adsorption properties of
about 1/2 of its original value might, in turn, be due to the similar sol-gel doped ormosils,[9] and was based on the
partial conversion of TEMPO radicals to nitrosonium observation that these materials are chemical sponges
‡
ions TEMPO that are rapidly formed upon addition of with chromatographic properties, i.e., they adsorb and
aqueous NaOCl (in 20% excess) to the reaction mixture concentrate the reagents at the cage surfaces where
containing the catalyst (when the slightly orangepowder reactions take place while the adsorbed molecules
suddenly changes its colour from pale to bright orange), promote further diffusion of incoming reactants.
and that after several consecutive reactions are left
attached as oxoammonium ions.
Indeed, besides the diffusional limitations imposed by
the narrow pore network, the chemical behaviour of sol-
Evidence for lack of degradation of the ormosil gel doped ormosils is affected by other fundamental
structure in the buffered alkaline reaction environment factors, and namely as in the case of analogous
(pH 9.1) was obtained by FTIR analysis of the catalyst ormosils-mediated esterifications[10a] and oxidations[10b]
prior and after 7 oxidative runs (Figure 3).
in the liquid-phase
by the hydrophobicity[11a] and
Adv. Synth. Catal. 2002, 344, 159 163
161

Rosaria Ciriminna, et al.
COMMUNICATIONS
Figure 3. FTIR spectra of the solgel ormosil doped with SG-TEMPO-2 prior to (solid line) and after (broken) 7 consecutive
reaction runs in the Montanari Anelli oxidation of 1-nonanol.
flexibility[11b] of the sol-gel cages that are both known to high yields of 1-nonanal with the known selectivity of
be crucially affected by the employment of substituted homogeneous TEMPO and no loss of chemical activity
silanes R×Si(OR)3 in the sol-gel polycondensation. It is upon consecutive reactions. Considering the versatility
the enhanced hydrophobicity of the silica sol-gel cages of both technologies,[3,12] the method is thought to be of
doped with TEMPO and organically modified by the utility for the selective oxidation of other alcoholic
presence of the alkyl groups that promotes the diffusion substrates, while the results of the present work are
of the hydrophobic alcohol and aldehyde molecules intended to provide further guidelines for the develop-
within the porous network, and also largely diminishes ment of versatile heterogeneous catalysts for selective
the amount of hindering intra-cage hydrogen bonds oxidations.
between Si-OH groups and silanols and water (con-
firmed one EPR by the lack of spurious peaks in
Figure 2, a compared to c), which, in turn, largely
enhances the freedom of the encapsulated TEMPO by
Experimental Section
providing the needed flexibility.
Interestingly, the continuous decrease of the activity
of TEMPO anchored to commercial aminopropyl-silica
was recently proposed[1b] to be due to intramolecular
quenching of the radicals anchored in proximity at the
material surface. In fact, surface derivatisation of a
material×s surface requires the formation of a new
covalent bond through a slow heterogeneous reaction
Preparation of SG-TEMPO Catalysts
All the chemicals employed were obtained from Sigma Aldrich
and used without further purification. The radical species 4-
oxo-TEMPO (1 g, 7 mmol) was added to a solution of 3-
aminopropyltrimethoxysilane (5 mL, 28 mmol) containing
NaBH3CN (0.220 g, 4 mmol) dissolved in methanol (18 mL).
The reaction solution was stirred for 40 h and then the
that leaves the anchored molecules unprotected at the unreacted cyanoborohydride destroyed by the addition of
7 M HCl (1 mL).
material×s pore surface, while the homogeneous sol-gel
encapsulation process disperses and protects the dopant
species within nano- or mesoporous modified silica
cages, allowing the use of doped gels with higher
catalytic loads without loss in material activity.
In conclusion, we have found that the coupling of sol-
gel and TEMPO-mediated oxidation processes affords
the synthesis of doped materials that are recyclable
heterogeneous catalysts for the oxidation of the ali-
The precursor mixture thereby obtained was equally divided
into 7 portions and each vessel charged with different amounts
(in subsequent order) of MeOH, alkyltrimethoxysilane,
TMOS and H2O to yield gels with the chosen total molar
ratios Si:H2O:MeOH ˆ 1:8:2 (or 3, as specified in Table 1).
Sodium fluoride was added as condensation catalyst in the
ratio Si:F ˆ 1:0.017 for the preparation of the gel obtained from
methyltrimethoxysilane (MTMS) alone (ETMS and PTMS are
ethyl- and propyltrimethhoxysilane, respectively). Hence, for
phatic alcohol 1-nonanol with bleach, rapidly affording example, the gel SG-TEMPO-2 with Si:H2O ˆ 1:2 was ob-
162
Adv. Synth. Catal. 2002, 344, 159 163

Sol-Gel Ormosils Doped with TEMPO as Recyclable Catalysts
COMMUNICATIONS
tained by charging one vessel of the precursor mixture
mentioned above with MeOH (1 mL), MTMS (5.82 mL) and
H2O (5.35 mL). The resulting sol was stirred using a Vortex
shaker prior to the addition of aqueous 1 M NaF (0.750 mL)
under fast stirring that resulted in fast gelation. Likewise, all
the other sols gelled rapidly yielding elasticalcogels that were
left sealed to age for 20 h at room temperature prior to be
opened and densified at 60 8C for 3 days during which the gels
further shrank to about half of the initial volume reaching
constant weight. The resulting xerogels isolated as pale orange-
coloured powders, were washed with boiling CH2Cl2 (3 Â
20 mL, 50 8C) under reflux, dried and used as such in the
oxidations.
Related Technologies), Dr. Cristina D×Oca (Department of
Pharmacochemistry) for the EPR measurements, and Dr.
Filippo Saiano (Laboratory of Agriculture Chemistry), all
from the University of Palermo, for the FTIR analyses. M.P. and
C.B. thank the CNR×s Department of International Relation-
ships for financial support from the Short-Term Mobility
Program and the Fonds der Chemischen Industrie for general
funding, respectively.
References
[1] a) C. Bolm, T. Fey, Chem. Commun. 1999, 1795; b) T.
Fey, H. Fischer, S. Bachmann, K. Albert, C. Bolm, J. Org.
Chem. 2001, 66, 8154; c) R. Ciriminna, D. Avnir, J. Blum,
M. Pagliaro, Chem. Commun. 2000, 1441.
[2] a) D. Brunel, F. Fajula, B. Nagy, B. Deroide, L. Veum,
J. A. Peters, H. van Bekkum, Appl. Catal. A: General
2001, 213, 73; b) A. Dijksman, I. W. C. E. Arends, R. A.
Sheldon, Chem. Commun. 2000, 271.
[3] A. E. J. de Nooy, A. C. Besemer, H. van Bekkum, Syn-
thesis 1996, 1153.
[4] See, for instance: S. V. Ley, I. R. Baxendale, R. N. Bream,
P. S. Jackson, A. G. Leach, D. A. Longbottom, M. Nesi,
J. S. Scott, R. I. Storer, S. J. Taylor, J. Chem. Soc. Perkin
Trans. 1 2000, 3815.
General Oxidation Procedure with SG-TEMPO-2 as
Catalyst
The oxidations were performed in a glass tube with a cooling
mantle equipped at the bottom with a ceramic filter plate that
allowed easy separation of the catalyst upon reaction. In a
typical experiment the catalyst (30.0 mg, 8 mmol radical) was
added to the tube followed bya CH2Cl2 solution (2.0 mL) of the
alcohol (0.4 M) with the internal standard dodecane (0.12 M),
and 0.16 mL of aqueous KBr (0.5 M). After cooling the
mixture to 0 8C, 2.7 mL of aqueous NaOCl diluted to a
concentration of 0.37 M and buffered by NaHCO3 to pH 9.1
were added and the resulting mixture shaken vigorously for
15 min. After filtration, the organicphase was separated, dried
over MgSO4 and the product analysed by GC (ultra-2-column,
HP). The filtered sol-gel catalyst was washed 5 times with H2O,
MeOH and CH2Cl2 (2 mL each), air-dried and reused as such in
7 consecutive oxidation runs after which the catalyst was used
for the EPR measurements carried out at 296 K on a Bruker
spectrometer working at medium frequency (50 kHz) and low
amplitude (0.101 G) modulation. The BET measurements
were performed on a Carlo Erba Sorptomatic1900, and a
Bruker Vector 22 spectrometer was used for the FTIR analyses
obtained from samples dispersed in KBr.
[5] A. Heeres, H. van Doren, K. F. Gotlieb, I. P. Bleeker,
Carbohydr. Res. 1997, 299, 221.
[6] A. Shames, O. Lev, B. Iosefzon-Kuyavskaia, J. Non Cryst.
Solids 1994, 175, 14.
[7] M. Epifani, G. De, A. Licciulli, L. Vasanelli, J. Mater.
Chem. 2001, 11, 3326.
[8] See reference 15 cited in Ref.[6]
[9] S. Fireman-Shoresh, N. H¸sing, D. Avnir, Langmuir
2001, 17, 5958.
[10] a) M. T. Reetz, Adv. Mater. 1997, 9, 943; b) M. Pagliaro,
R. Ciriminna, Tetrahedron Lett. 2001, 42, 4514.
[11] a) B. M. De Witte, D. Commers, J. B. Uytterhoeven, J.
Non Cryst. Solids 1996, 202, 35; b) D. Avnir, Acc. Chem.
Res. 1995, 28, 328.
[12] J. D. Wright, N. A. J. M. Sommerdijk, Sol-Gel Materials:
Chemistry and Applications, Gordon and Breach Science
Publishers, Amsterdam, 2001.
Acknowledgements
Thanks to Dr. Ingo Schiffers (RWTH Aachen) for the GC
analyses, to Prof. Mariella Brai (Department of Physics and
Adv. Synth. Catal. 2002, 344, 159 163
163