Accessibility control on copper(II) complexes in mesostructured porous silica obtained by direct synthesis using bidentate organosilane ligands

Article abstract of DOI:10.1021/la101379p

The accessibility of metal(II) complexes in 2D hexagonal mesostructured porous silicas obtained by direct synthesis is controlled using an appropriate organosilane ligand. This is exemplified here using copper(II) as a transition metal probe and a neutral

Full text of DOI:10.1021/la101379p

pubs.acs.org/Langmuir
© 2010 American Chemical Society
Accessibility Control on Copper(II) Complexes in Mesostructured Porous
Silica Obtained by Direct Synthesis using Bidentate Organosilane Ligands
†
§
‡
,†
Wen-Juan Zhou,†,‡ Belen Albela, Pascal Perriat, Ming-Yuan He, and Laurent Bonneviot*
ꢀ
†Laboratoire de Chimie, Ecole Normale Supeꢀrieure de Lyon UMR-CNRS 5182, Universiteꢀ de Lyon,
‡
ꢀ
46 Allee d’Italie, 69364 Lyon Cedex 07, France, Shanghai Key Laboratory of Green Chemistry and Chemical
Process, East China Normal University, No. 3663, North Zhongshan Road, 200062, Shanghai, China, and
§Laboratoire MATEIS, INSA de Lyon, Ba^timent Blaise Pascal, 7 Avenue Jean Capelle, Lyon, France
Received April 7, 2010. Revised Manuscript Received June 18, 2010
The accessibility of metal(II) complexes in 2D hexagonal mesostructured porous silicas obtained by direct synthesis is
controlled using an appropriate organosilane ligand. This is exemplified here using copper(II) as a transition metal
probe and a neutral or negatively charged ligand: N-(2-aminoethyl)-3-aminopropyltrimethoxysilane, LA, and,
N-salicylaldimine-propylamine-trimethoxysilane, LB-, respectively. LA leads to inaccessible complexes located into
the pore wall and called “embedded” sites here where silanolate groups from the siliceous network block the access to
-
Cu(II) ions. By contrast, LB generates accessible complexes, named “showing-on” sites here. The copper-containing
silicas were synthesized with various metal molar ratios (M/SiO2 = 0.5-3%) in basic media, with cetyltrimethyl-
ammonium p-toluenesulfonate (CTATos) as template and with sodium silicate solution as silicon source. A soft
template extraction procedure has been developed to preserve the complex integrity of the showing-on copper sites
during the treatment. The embedded copper(II) and nickel(II) sites were compared. Materials containing embedded,
showing-on, and grafted sites were also compared with regard to pore size, surface polarity, and metal leaching. The
material containing showing-on sites was found to be catalytically active for the hydroxylation of phenol into catechol
and hydroquinone. Both textural and structural properties of the material and the copper sites were investigated using
XRD, TEM, N2 sorption isotherms, TGA, FT-IR, UV-visible, and EPR spectroscopies.
Introduction
directly introduced in the sol-gel synthesis of the mesostructured
silica according to direct synthesis and eventually synthesis
route.11-15 These latter approaches lead to species located in
the framework of the pore wall by contrast to surface species
obtained by grafting. Both techniques may be combined to
generate multifunctional mesostructured systems. For a direct
synthesis, precursors with theorganic moietylinked to at least two
condensable alkyltrialkoxysiloxane arms are required; then, per-
iodic mesostructured organosilicas materials (PMOs) may be
obtained following a sol-gel route in the presence of templating
surfactants.16-19 The latter technique is rather successful for small
organic moieties but seems up to now limited for bulky ones such
as transition metal complexes. The present study focuses on such
direct incorporation of transition metal ions adding an accessi-
bility control of the coordination to external molecules.
The developement of various and eventually “orthogonal”
strategies to synthesize functional solids are important particu-
larly to combine different types of active sites with various
properties. For instance, combining conductivity and magnetism
or superparamagnetism and fluorescence is crucial in diagnosis
and therapy.1-5 In catalysis, combining antagonist sites such as
acid and base functions can lead to synergic effects as in
enzymes.6-8 In that context, mesostructured porous silicas9 are
good candidates to develop tunable multifunctional materials.
Conveniently, their internal surface presents a high silanol density
useful for grafting various specific functions. For example, dual
grafting of molecular “stirrers” and “gates” has been performed
in a MCM-41 solid.10 Alternatively, organic functions can be
Indeed, because of their large size, the incorporation of metal
complexes in the pore wall of ordered mesoporous silicas using a
direct synthesis route remains challenging: ethylenediamine com-
plexes with Cu(II) in HSM silicas11 and cyclam complexes with
Cu(II) and Co(II) in SBA-1513,14 both in neutral synthetic
*Corresponding author. Laurent. Bonneviot, fax: þ33 472 72 88 60; tel:
þ33 472 72 83 91; e-mail: laurent.bonneviot@ens-lyon.fr.
(1) Tsai, C. P.; Hung, Y.; Chou, Y. H.; Huang, D. M.; Hsiao, J. K.; Chang, C.;
Chen, Y. C.; Mou, C. Y. Small 2008, 4, 186.
(2) Kim, J.; Kim, H. S.; Lee, N.; Kim, T.; Kim, H.; Yu, T.; Song, I. C.; Moon,
W. K.; Hyeon, T. Angew. Chem-Int. Edit. 2008, 47, 8438.
(3) Julian-Lopez, B.; Boissiere, C.; Chaneac, C.; Grosso, D.; Vasseur, S.;
Miraux, S.; Duguet, E.; Sanchez, C. J. Mater. Chem. 2007, 17, 1563.
(4) Cornia, A.; Costantino, A. F.; Zobbi, L.; Caneschi, A.; Gatteschi, D.;
Mannini, M.; Sessoli, R. Single-Mol. Magn. Relat. Phenom. 2006, 122, 133–161.
(12) Karakassides, M. A.; Bourlinos, A.; Petridis, D.; Coche-Guerente, L.;
Labbe, P. J. Mater. Chem. 2000, 10, 403.
€
ꢀ
(5) Lu, A. H.; Schuth, F. Adv. Mater. 2006, 18, 1793.
(13) Corriu, R. J. P.; Mehdi, A.; Reye, C.; Thieuleux, C. Chem. Commu. 2003,
1564.
(14) Corriu, R. J. P.; Mehdi, A.; Reye, C.; Thieuleux, C. Chem. Commu. 2002,
1382.
(15) Corriu, R. J. P.; Mehdi, A.; Reye, C.; Thieuleux, C. New J. Chem. 2003, 27,
905.
(6) Anan, A.; Sharma, K. K.; Asefa, T. J. Mol. Catal. A: Chem. 2008, 288, 1.
(7) Gelman, F.; Blum, J.; Schumann, H.; Avnir, D. J. Sol-Gel Sci. Technol. 2003,
26, 43.
(8) Zhang, F.; Liu, G. H.; He, W. H.; Yin, H.; Yang, X. S.; Li, H.; Zhu, J.; Li,
H. X.; Lu, Y. F. Adv. Funct. Mater. 2008, 18, 3590.
ꢀ
ꢀ
(9) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.;
Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.;
Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834.
(10) Zhu, Y. C.; Fujiwara, M. Angew. Chem., Int. Ed. 2007, 46, 2241.
(11) Karakassides, M. A.; Fournaris, K. G.; Travlos, A.; Petridis, D. Adv.
Mater. 1998, 10, 483.
(16) Inagaki, S. Nanotechnol. Mesostruct. Mater. 2003, 146, 1.
(17) Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.; Terasaki, O. J. Am.
Chem. Soc. 1999, 121, 9611.
(18) Hunks, W. J.; Ozin, G. A. J. Mater. Chem. 2005, 15, 3716.
(19) Asefa, T.; MacLachlan, M. J.; Coombs, N.; Ozin, G. A. Nature 1999, 402,
867.
Langmuir 2010, 26(16), 13493–13501
Published on Web 07/14/2010
DOI: 10.1021/la101379p 13493

Article
Zhou et al.
Scheme 1. Organosilane Ligands LA and HLB Used for Incorpora-
tion of Copper or Nickel in the Mesoporous Materials
structure sensitive electron paramagnetic resonance (EPR)
response.27 For comparison, mesoporous materials possessing
framework Cu(II) and Ni(II) species using LA were prepared
using a similar method as before except for the extraction of the
surfactant cetyltrimethylammonium, CTAþ. Previously, the dis-
placement of CTAþ and the capping of the silanol by trimethyl-
silyl using a silylation agent was performed to avoid the collapse
of the silica mesostructure (reminding that it is synthesized at
low temperature ∼60 ꢀC).21,22 Here, the new extraction method
consists merely of a soft acidic treatment using 1 mol equiv of HCl
vs the surfactant that allows better control on the complex built
-
from LB . In complement, a metal free and MCM-41 (LUS) where
[Ni(LA)2]2þ complexes have been postgrafted were used also for
comparison. Note that the new extraction procedure has been
also applied for the nickel-containing material generating here an
original set of samples that are not silylated (no silanol capping in
the channel) in comparison to our previous study.22 The structure
and the texture of the materials have been investigated using
XRD, N2 sorption isotherms, and TEM techniques. The nature
and the quantity of the loaded species have been determined from
FT-IR spectroscopy, elemental analysis (EA), ICP-MS, and TGA
measurements. Furthermore, the metal displacement upon acidic
treatment has been tested in both types of materials and mon-
itored using EPR spectroscopy. Finally, the catalytic activity on
the hydroxylation of phenol into diphenols (catechol and hydro-
quinone) has been also studied.
conditions, again ethylenediamine copper(II) complexes and also
vanadyl Schiff basecomplexes in MCM-41 using a basic route12,20
and, also, phosphinorhodium(I) complexes in SBA-3 using an
acidic route.21 In all the cases, the metal ion is complexed by a
single tetradentate macrocycle with four polycondensable groups,
by a single tetradentate ligand with two polycondensable groups,
or by three monodentate organosilane ligands, respectively.
Bidentate ligands with one organosilane polycondensable arm
can be used as long as there are at least two such ligands per metal
ion.11,12,22 The metal ion concentration has to remain relatively
low (<5%) mainly to preserve the mesostructure; therefore, they
may not be regarded as PMO materials, despite the high C/Si
molar ratio that can be reached (ca. 20%). Furthermore, the pH,
the temperature, and the choice of ligand are key parameters to
maintain the complex integrity and to obtain a control on their
location in the final material.12,21,22 However, it is not clear yet
what are the key parameters that control the metal accessibility
required for eventual adsorption or catalytic properties. To our
knowledge, there is no other work reporting specifically on that
matter.
Using the basic synthesis route, we have previously reported
that nickel(II) complexed with the bidentated neutral ligand, LA
(see abstract and Scheme 1), leads to framework complexes that
are not accessible to external molecules and are resistant to acid
leaching despite the very small pore wall thickness (∼1 nm) of the
mesostructured host matrix.22 The coordination of nickel(II) is
characterized by two bidentate LA ligands and, in addition, two
silanolate groups that are locking the metal accessibility. This
observation prompted us to find a way to avoid such a coordina-
tion blockage by moving from a neutral bidentate ligand to a
negatively charged ligand that might compete favorably with the
blocking effect of the silanolate ligands. From the demonstration,
this ligand had to be structurally related to LA.
Experimental Section
1. Materials. Copper(II) acetate trihydrate (99%, Acros),
copper(II) nitrate pentahydrate (99%, Acros), nickel(II) nitrate
hexahydrate (99%, Acros), sodium hydroxide (Acros), 3-amino-
propyl-trimethoxysilane (95%, Acros), salicylaldehyde (98%,
Avocado), N-(2-aminoethyl)-3-aminopropyltriymethoxysilane
(LA; 97%, Acros), Ludox HS-40 (40% SiO2; Aldrich), hexadecyltri-
methylammonium-p-toluene-sulfonate, (CTATos; > 99% Merck),
HCl (1 mol L-1, Acros), ethanol (96%, Elvetec; anhydrous, Carlo
3
Erba, HPLC quality), potassium phosphate monobasic (99%,
Acros), disodium hydrogen phosphate (anhydrous, >99%, Merck),
phenol (g99%, Sigma-Aldrich), catechol (99%, Acros), and hydro-
quinone (g99%, Fluka) were used as received. The sodium silicate
solution was prepared as follows: Ludox HS- 40 (187 mL) was added
to sodium hydroxide (32 g) in deionized water (800 mL) and stirred
at 40 ꢀC until clear.
2. Synthesis. 2.1. Direct Synthesis of Materials DS-
NiLA and DS-CuLA. A sodium silicate solution (40 mL) was
stirred at 60 ꢀC for 1 h. A second solution of hexadecyltrimethyl-
ammonium p-tolunensulfonate (CTATos, 1.8 g) in deionized
water (50 mL) was stirred during 1 h at 60 ꢀC. A solution
containing 0.28 mmol of [Ni(LA)2](NO3)2 (M/SiO2 = 0.3%,
LA = N-(2-aminoethyl)-3-aminopropyltrimethoxysilane) or 0.84
mmol of [Cu(LA)2](NO3)2 (M/SiO2 = 1%) was added to the first
solution and stirred until homogeneous. The solutions remain clear
up to this point. By contrast, a rapid condensation of the silica takes
place when this solution is added dropwise during 20 min at 60 ꢀC to
the second solution containing the templating ammonium. The
colloidal solutions formed were further stirred at 60 ꢀC for 24 h.
After filtration and washing with deionized water (200-300 mL), the
as-made solids were dried in the oven at 60 ꢀC for 2 days, obtaining
2.1 g of DS-NiLA and 3.0 g of DS-CuLA, respectively. EA for DS-
NiLA: C (30.41%), N (2.23%), Ni (0.61%), weight loss at 1000 ꢀC
(46.59%), and DS-CuLA: C (31.16%), N (3.22%), Cu (1.67%),
weight loss at 1000 ꢀC (51.07%).
The organosilane ligand of choice is the hemisalen
(salicylaldimine), HLB (see Abstract and Scheme 1), that pro-
duces well-characterized neutral M(LB)2 complexes with M =
Co2þ, Cu2þ, Ni2þ 23-27
Besides, such metal complexes are very
.
interesting per se, since they are known to afford selective catalytic
oxidation reactions.26 Here, Cu(II) is preferred owing to its highly
(20) Baleizao, C.; Gigante, B.; Garcia, H.; Corma, A. J. Catal. 2003, 215, 199.
(21) Dufaud, V.; Beauchesne, F.; Bonneviot, L. Angew. Chem., Int. Ed. 2005, 44,
3475.
(22) Zhou, W. J.; Albela, B.; Ou, M.; Perriat, P.; He, M. Y.; Bonneviot, L. J.
Mater. Chem. 2009, 19, 7308.
(23) Chandran, D.; Kwak, C. H.; Ha, C. S.; Kim, I. Catal. Today 2008, 131, 505.
(24) Chakraborty, J.; Nandi, M.; Mayer-Figge, H.; Sheldrick, W. S.; Sorace, L.;
Bhaumik, A.; Banerjee, P. Eur. J. Inorg. Chem. 2007, 5033.
(25) Luo, Y.; Lin, J. Microporous Mesoporous Mater. 2005, 86, 23.
(26) Murphy, E. F.; Schmid, L.; Burgi, T.; Maciejewski, M.; Baiker, A.;
Gunther, D.; Schneider, M. Chem. Mater. 2001, 13, 1296.
(27) Murphy, E. F.; Ferri, D.; Baiker, M.; Van Doorslaer, S.; Schweiger, A.
Inorg. Chem. 2003, 42, 2559.
2.2. Synthesis of HLB Ligand. A two-neck round-bottom
flask with a condenser was first vacuumed at room tem-
perature and then filled with argon, and the same procedure
was repeated three times. Salicylaldehyde (0.57 mL, 5 mmol) and
anhydrous ethanol (22 mL) were introduced into the flask.
13494 DOI: 10.1021/la101379p
Langmuir 2010, 26(16), 13493–13501

Zhou et al.
Article
3-Aminopropyltrimethoxysilane (0.90 mL, 5 mmol) was then
added dropwise. The resulting mixture was stirred under reflux
in the argon atmosphere for 6 h. The obtained yellow product was
analyzed by 1H NMR (300 MHz, CDCl3): δ 0.76 (t, 2H, J = 105
8.30Hz), 1.89(tt, 2H, J = 8.20, 6.75 Hz), 3.63(t, 2H), 3.64(s, 9H),
6.92(t, 1H, J = 7.49 Hz), 7.01 (d, 1H, J = 8.20), 7.30 (dd, 1H, J =
7.50 Hz), 7.37 (d, 1H, J = 7.46 Hz), 8.40 (s, 1H). It was found
impossible to crystallize this compound. Since the reaction was
quantitative, there were negligible quantities of reactant detected
onthe NMR signal and the reaction mixture was used directly as a
source for the synthesis of the solid.
complexes, using trimethylsilyl functions (TMS) to separate each
complex one from another.22,31,32
2.6.1. Synthesis of Standard Metal Free 2D Hexagonal
Mesoporous Silica PS. The synthesis was processed as for
material DS replacing the 24 h stirring at 60 ꢀC by autoclaving
at 130 ꢀC for 20 h. Six grams of LUS was obtained and designated
hereas PS to simplify the notation.29 EA: C (32.7%), H (6.6%), N
(2.0%), weight loss at 1000 ꢀC (48.4%). PS-E was obtained after
extraction of the template following the same protocol as for
DS-E (see above).
2.6.2. Partial Template Extraction of PS. PS (10 g) was
placed in a round-bottom flask, and then ethanol (400 mL, 96%)
2.3. Direct Synthesis of DS-CuLB Materials. The above
mixture was evaporated at 40 ꢀC under vacuum, and the volume
was reduced to 10 mL. Then, an ethanol solution (40 mL) of
copper acetate monohydrate (0.479 g, 2.4 mmol) was added
dropwise. The mixture was stirred under reflux in argon atmo-
sphere for 2 h obtaining a solution (48.5 mL) containing the CuLB
complexes (no hydrolysis was observed despite the presence of
water in the solid copper salt). Then, 1.7, 3.4, or 8.5 mL of this
solution (nCu = 0.084, 0.17, 0.42 mmol, M/SiO2 = 0.1%, 0.2%,
0.5%, respectively) were used to prepare materials DS-CuLB,
series using the same experimental procedure as for materials DS-
NiLA and DS-CuLA. After washing with water and drying at 60
ꢀC, about 2.8 g of solid was obtained in all cases. EA: C (33.3%),
N (1.97%), Cu (0.24%), weight loss at 1000 ꢀC (51.58%); C
(32.09%), N (1.79%), Cu (0.35%), weight loss at 1000 ꢀC
(53.27%); C (31.94%), N (2.17%), Cu (0.94%), weight loss at
1000 ꢀC (46.59%), respectively.
and hydrochloric acid 1 mol L-1 (6.8 mL, 1.1 equiv) were added.
3
The mixture was stirred at 40 ꢀC for 1 h. After filtration and
washing with ethanol (100 mL ꢀ 2) and acetone (50 mL ꢀ 2), the
solid was dried at 80 ꢀC for 20 h. Material PS-PE (7.1 g) was
obtained.
2.6.3. Partial Silylation of PS-PE. PS-PE (6.8 g) was added
into a round-bottom three-neck flask, and then dried at 130 ꢀC for
1 h under argon flow and during 2 h under vacuum. Cyclohexane
(170 mL) and HMDSA (30 mL) were added under argon. The
mixture was refluxed for 18 h. The obtained solid was finally
washed with cyclohexane (2 ꢀ 30 mL), ethanol (2 ꢀ 60 mL), and
acetone (2 ꢀ 60 mL) and then dried at 80 ꢀC for 18 h. This
sequence of steps was repeated twice. Partially silylated material
PS-PES (7.1 g) was obtained. EA: C (19.7%), H (4.37%), N
(0.7%), weight loss at 1000 ꢀC (22.6%).
2.6.4. Extraction of Remaining Surfactant of PS-PES. PS-
PES (2 g) was placed in a round flask, and then ethanol (86 mL,
2.4. Direct Synthesis of Metal-Free Reference DS Material.
A metal-free silica DS was prepared with similar conditions than
materials DS-NiLA and DS-CuLA. EA: C (35.0%), H (7.4%), N
(2.0%), weight loss at 1000 ꢀC (58.1%). It is basically the low-
temperature synthesis version of LUS usually autoclaved at
130 ꢀC (vide infra).28,29
96%) and hydrochloric acid (1.28 mL, 1 equiv, 1 mol L-1) were
3
added. The mixture was stirred at 40 ꢀC for 1 h. After filtration
and washing with ethanol (50 mL) and acetone (50 mL), the
obtained solid was dried at 80 ꢀC for 20 h. The same procedure
was repeated. Finally, material PS-PESE (1.4 g) was obtained.
EA: C (7.8%), H (2.4%), N (<0.1%), weight loss at 1000 ꢀC
(10.7%).
2.5. Template Extraction. The solid (0.4 g, DS-NiLA, DS-
CuLA or DS-CuLB, and DS) was added into a two-neck round-
bottom flask with a condenser. Then, technical ethanol (96%,
2.6.5. Synthesis of Material PS-NiLA. PS-PESE (0.5 g) was
dried at 130 ꢀC under argon flow during 1 h, and then evacuated
under vacuum at 130 ꢀC during 2 h. A solution of Ni(II) nitrate
(0.28 mmol) and LA ligand (nLA/nNi = 2) in 2 mL anhydrous
ethanol was added together with 40 mL of toluene. The resulting
mixture was stirred at 60 ꢀC for 18 h under argon. After filtration,
the obtained solid was washed with toluene (50 mL) and ethanol
(25 mL), and then dried at 60 ꢀC, obtaining 0.56 g of material PS-
NiLA. EA: C (12.1%), H (3.0%), N (3.5%), Ni (2.5%), weightloss
at 1000 ꢀC (19.3%).
25 mL) and HCl (1 mol L-1, 0.54 mL, nHCl:nCTAþ = 1:1) were
3
added. The mixturewas stirred at40ꢀC for 1 h. The obtained solid
was finally washed with technical ethanol (50 mL ꢀ 4) and
acetone (50 mL ꢀ 2). The solids, DS-NiLA-E (0.24 g), DS-
CuLA-E (0.25 g), DS-CuLB-E (ca. 0.22 g), and DS-E (0.25 g) were
respectively obtained. EA for DS-NiLA-E: C (2.43%), N (1.01%),
Ni (0.88%), weight loss at 1000 ꢀC (13.37%); DS-CuLA-E: C
(5.97%), N (2.60%), Cu (2.50%), weight loss at 1000 ꢀC
(16.30%); DS-CuLB-E with the smallest Cu/Si ratio: C (3.20%),
N (0.21%), Cu (0.31%), weight loss at 1000 ꢀC (89.62%); with the
intermediate Cu/Si ratio: C (7.95%), N (1.74%), Cu (0.52%),
weight loss at 1000 ꢀC (83.49%); with the higest Cu/Si ratio: C
(4.34%), N (0.56%), Cl (0.04%), Cu (1.06%), weight loss at
1000 ꢀC (19.56%); DS-E: C (12.4%), H (3.6%), N (0.1%), weight
loss at 1000 ꢀC (12.8%).
2.7. Metal Extraction. The solid DS-NiLA-E, DS-CuLA-E,
or DS-CuLB-E (0.4 g) was added into a flask with a condenser.
Then, ethanol (25 mol) and HCl (1 mol L-1, 0.5 mL, nHCl:nmetal
=
3
8, 3, and 8, respectively) were added. The mixture was stirred at
40 ꢀC for 1 h. After filtration and washing with technical ethanol
(50 mL ꢀ 4) and acetone (50 mL ꢀ 2), the solids were dried at 60 ꢀC
for one day. Materials DS-NiLA-H, DS-CuLA-H, and DS-CuLB-H
(ca. 0.24 g) were obtained. For PS-NiLA, the same process was
2.6. Postsynthesis of Material PS-NiLA. Material PS-NiLA
was synthesized starting from a 2D hexagonal silica prepared at
130 ꢀC called LUS.29,30 The “molecular stencil patterning” (MSP)
technique was used to homogeneously distribute the grafted
employed (HCl 1 mol L-1, nHCl:nmetal = 9) obtaining material PS-
3
NiLA-H.
2.8. Catalytic Test. Phenol hydroxylation experiments were
performed at 80 ꢀC for 2 h using a 50 mL two-neck round-bottom
flask with a condenser. In a standard run, phenol (0.1 g, 1.1
mmol), catalyst (molar ratio Cu:phenol = 1:130), buffer solution
pH=6 (7 mL) and H2O2 (50% aqueous) were added in this order.
The phenol and the products, catechol (CAT) and hydroquinone
(HQ), were analyzed by HPLC.
(28) Bonneviot, L.; Morin, M.; Badiei, A. Patent WO 01/55031 A 1 2001. this
patent relates the use of CTATos as a surfactant where the presence of tosylate instead of
inorganic anions reported in the other patent literature allows to produce an hexagonal
mesostructured silica of high stability with much lower surfactant concentration. Then,
the yield in surfactant is nearly 100% (no foam during washing). Incidentally, the CTA
Tosylate salt is a cheaper surfactant source than the chloride or bromide analogue.
(29) Reinert, P.; Graillat, C.; Spitz, R.; Bonneviot, L. Proc. 3rd Int. Meso-
structured Mat. Symp., Ryoo, R., Park, S., Eds. Stud. Surf. Sci. Catal., 2003, 146,
439.
3. Analytical Techniques. XRD: low-angle X-ray powder
diffraction experiments have been carried out using a Bruker
(Siemens) D5005 diffractometer using Cu K monochromatic
radiation. IR: infrared spectra were recorded from KBr pellets
using a Mattson 3000 IR-TF spectrometer. UV-visible: liquid
UV-visible spectra were recorded using a Vector 550 Bruker
(30) Bonneviot, L.; Badiei, A.; Crowther, N. Patent WO 02216267 2002.
ꢀ
(31) Calmettes, S.; Albela, B.; Hamelin, O.; Menage, S.; Miomandre, F.;
Bonneviot, L. New J. Chem. 2008, 32, 727.
(32) Abry, S.; Thibon, A.; Albela, B.; Delichere, P.; Banse, F.; Bonneviot, L.
New J. Chem. 2009, 33, 484.
Langmuir 2010, 26(16), 13493–13501
DOI: 10.1021/la101379p 13495

Article
Zhou et al.
spectrometer; solid diffuse reflectance UV-visible spectra were
recorded from aluminum cells with Suprasil 300 quartz windows,
using a PerkinElmer Lambda 950 and PE Winlab software.
Nitrogen sorption isotherms at 77 K were determined with a
volume device Micromeritics ASAP 2010 M on solids that were
dried at 80 ꢀC under vacuum overnight. TGA measurements were
collected from Al2O3 crucibles on a DTA-TG Netzsch STA 409
PC/PG instrument, under air (30 mL/min), with a 25-1000 ꢀC
(10 ꢀC/min) temperature increase. The silicon content was system-
atically measuredfrom the residual mass at 1000 ꢀC assuming that
C, H, and N were all removed. This was cross-checked by
elemental analysis on several samples with or without metal. In
the presence of metal, a correction was performed assuming the
formation of Cu2O remaining at 1000 ꢀC. Finally, the silicon
coming from the organosilane was neglected when, as in most of
the cases, it accounted for less than 4%.21,22,31,32,36,48 pH-meter
(Meterlab pHM210) was calibrated in aqueous solution, so that
the pH measured for the EtOH-H2O is only indicative. EPR
spectra were recorded using a Bruker Elexsys e500 X-band (9.4
GHz) spectrometer with a standard cavity. Quantification of
copper free reference, DS, whereas the yield of copper was nearly
total (Table S2). In the case of the Ni-containing material, DS-
1NiLA, both silicon and metal yields were lower, i.e., ∼ 20% and
∼80%, respectively. This is consistent with a higher rate of
hydrolysis for the metalated alkoxysiloxane than that of the
silicate oligomers.22
The metal-free material, DS-, was prepared in the very same
conditions as the above-described metal-containing materials.
The postgrafted material, PS-NiLA, with the metal complexes
located inside the pores, was prepared also with an L/M molar
ratio equal to 2. The latter was synthesized using the so-called
molecular stencil patterning (MSP) approach, which allows
homogeneous distribution of a function on the internal surface
of the pore of a 2D hexagonal mesostructured porous silica in
basic conditions.22,31,32 Accordingly, starting form an as-made
silica, PS, a controlled amount of the templating agent CTAþ is
removed to lead to the partially extracted material PS-PE. Then,
a silylation step with hexamethyldisilazane (HMDSA) is per-
formed without displacement of the remaining CTAþ molecules
to afford the partially extracted and silylated material PS-PES.
After removal of the remaining CTAþ moieties, which play the
role of stencil, the material PS-PESE is obtained. The latter
undergoes the grafting of the metal complexes [M(LA)2]2þ, yield-
ing the postgrafted material PS-NiLA.
The extraction of the template is a key point, since it should
neither affect the coordination state of the metal nor provoke any
metal leaching. A mere extraction of the templating agent may be
obtained from either an acidic or nearly neutral treatment using
HCl or NH4AcO, respectively. Other extraction routes were
investigated allowing concomitantly the capping of some of the
silanol groups of the channel surface:22 direct silylating treatment
using (i) chlorotrimethylsiliane (TMSCl) with HCl evolution or
(ii) a mixture of TMSCl and hexamethyldisilazane (HMDSA).
The latter treatment was found to be the best to maintain the
[Ni(LA)2] species in the framework position and preserve the
hexagonal structure of the silica due to the more neutral media
generated during the process.22 Nonetheless, these conditions
were soft enough to retain copper in the DS-CuLB series. The
capping treatment was abandoned. Finally, it was found that a
mere acid extraction using of only 1 equiv of HCl per surfactant
was sufficient to neutralize the silanolate groups and to produce a
nearly neutral solution that fully preserved the structure. This has
been applied to generate the series of extracted solids, DS-NiLA-
E, DS-CuLA-E, and DS-CuLB-E. The latter present M/Si molar
ratios of 0.010, 0.029, and 0.013 to be compared to 0.012, 0.032,
and 0.016 for the mother materials DS-NiLA, DS-CuLA, and DS-
CuLB, respectively (Supporting Information Table S1). These
figures indicate that the metal content remained nearly the same
after such a soft treatment. For the grafted material PS-NiLA, the
M/Si molar ratio was 0.038.
Cu(II) species was performed using crystals of CuSO4 5H2O as
3
calibration reference. TEM: imaging and analytical studies were
performed on a JEOL 2010F microscope equipped with a field
emission gun and operating at 200 kV. EDX (energy dispersive
X-ray analysis) was performed with the help of a dispersive X-ray
system (INCA-Oxford equipment) with a detector having a
polymer ultrathin window. HPLC analyses were performed on
a LC/MS Agilent 1100SLT. The working conditions are the
following: column, ZORBAX-SB-C18 of dimension 4.6 mm ꢀ
250mm;protect, 5 μm; loopsize, 8μL;temperature ofthe column,
room temperature; UV detector, 245 and 280 nm; mobile phase,
water (10 mmol/L formic acid, 60%) and acetonitrile (40%); flow
rate, 0.2 mL/min.
Results and Discussion
1. Synthesis of the Materials. Five materials have been
synthesized at 60 ꢀC using direct synthesis (DS): DS-NiLA and
DS-CuLA, with LA ligand and starting from Ni(NO3)2 or Cu-
(NO3)2, respectively, and three DS-CuLB materials with HLB and
various copper loading starting from Cu(AcO)2. In all cases, the
complex precursor was polycondensed in basic conditions with
sodium silicate in the presence of hexadecyltrimethylammonium
p-tolunensulfonate (CTATos) in water. For complexation with
HLB, the acetate counterion was preferred over nitrate to avoid
acidity and hydrolysis of the imine bond of the ligand (vide
infra).34,35 In all the materials, the initial molar L/M ratio was
fixed at 2, according to previous studies with the Ni-LA system,
showing that these conditions yield mainly [Ni(LA)2]2þ species,
whereas other ratios gave a mixture of complexes or/and free
ligand.22
The M/Si molar ratio in the mother gel was 0.3% and 1.0% for
DS-NiLA and DS-CuLA and 0.1%, 0.2%, and 0.5% in the DS-
3CuLB series. The actual metal content was 1.2% and 3.2% for
the MLA series and 0.5%, 0.7%, and 1.6% for the CuLB series,
respectively (Supporting Information Table S1 and Experimental
section). The DS-CuLB series was designed to test the effect of the
metal loading with ligand LB. For the sake of simplification, DS-
CuLB refers to as the highest copper loaded material of the series
otherwise stated. Note that the M/Si molar ratio in the solid is
much higher than in the mother gel. Indeed, the yield of silicon
was low, only ∼30% for DS-CuLA and DS-CuLB, as for the
The elemental analyses of the template extracted materials
provide a N/M molar ratio of 4.8 for both DS-NiLA and DS-
CuLA, indicating the presence of at least two LA ligands per metal
ion. However, in the postsynthesis material PS-NiLA, the N/M
value of 6.0 suggested the presence of two nitrate ions as
counterions in addition of the two LA.22 The presence of the
latter ions was confirmed by IR spectroscopy. Finally, for DS-
CuLB-E a N/M molar ratio of 2.4 is consistent with the presence
of at least two LB ligands per metal ion in the material and the
presence of non-coordinated LB- ligands due to a slight excess of
ligand in the preparation.
(33) Abry, S.; Albela, B.; Bonneviot, L. C. R. Chim. 2005, 8, 741.
(34) Fenton, D. E.; Westwood, G. P.; Bashall, A.; McPartlin, M.; Scowen, I. J.
J. Chem. Soc., Dalton Trans. 1994, 2213.
(35) Elias, H.; Hasserodttaliaferro, C.; Hellriegel, L.; Schonherr, W.; Wannowius,
K. J. Inorg. Chem. 1985, 24, 3192.
2. Structure and Morphology of the Materials. All the
materials are characterized by three XRD peaks assigned to
Æ100æ, Æ200æ, and Æ210æ reflections, consistent with a well-ordered
13496 DOI: 10.1021/la101379p
Langmuir 2010, 26(16), 13493–13501

Zhou et al.
Article
Figure 1. XRD patterns of materials DS-NiLA-E, DS-NiLA,
DS-CuLB-E, and DS-CuLB.
Figure 3. N2 sorption isotherms at 77K of DS-E, DS-1NiLA-E,
DS-CuLA-E, and DS-CuLB-E.
the pore diameter, φBdB, is the same within 0.1 nm, for materials
DS-NiLA-E and DS-CuLA-E and for the metal-free reference DS-
E, while it is smaller for DS-CuLB-E (Table 1). The introduction
of the metal complex in the DS series produces only a slight
decrease of the specific surface area, SBET, and the total porous
volume, Vp. Nonetheless, the effect is more pronounced in a
higher metal-loaded DS-CuLA-E material than in DS-NiLA-E.
The horizontal hysteresis seen on the N2 sorption isotherm for
P/P0 > 0.4 indicates the presence of defects found in a higher
concentration for higher metal loadings.39 Interestingly, the
incorporation of the metal complex in the sol-gel synthesis does
not affect the resulting pore wall thickness, Wt (Table 1). By
contrast, the grafting of complexes such as Ni(LA)2 or ami-
no-copper complexes in LUS, i.e., in both PS-NiLA and PS-
CuLC,32 respectively, leads not only to a drastic decrease of pore
volume, but also to a much thicker apparent wall thickness. This
is consistent with an expected pore filling and a location of the
complexes at the surface of the wall leading to an apparent
thickening of ca. 0.5 nm.
These data strongly suggest that the metal complexes are
located in the pore wall in the materials DSNiLA-E, DS-CuLA-E,
and DS-3CuLB-E. By contrast, the presence of grafted species in PS-
NiLA obviously leads to an increase of the apparent pore thickness
and to a decrease of the pore volume. Furthermore, the C parameter,
which is calculated from the BET equation, is related to the surface
polarity and provides complementary information on the surface
modification. Indeed, the materials prepared with the metalated
Ligand LA are characterized by relatively high C values (∼100) close
to that of the metal-free material DS and consistent with little surface
modification. By contrast, the organic grafted complexes in PS-
NiLA clearly decrease the surface polarity (C = 34). The material
DS-CuLB-E exhibits an intermediate value (C = 83) suggesting an
intermediate situation in the materials using LB.
Figure 2. TEM image of DS-CuLB-H (for Cu/Si = 0.5 mol %).
hexagonal array of channels like in MCM-41-type materials
(Figure 1).9,32,36 Transmission electron microscopy (TEM) in-
vestigations also confirmed the highly ordered nature of the
materials studied here as depicted in Figure 2 for DS-CuLB-E.
Note that the XRD pattern is about three times more intense for
the template-extracted materials than for the as-made materials;
indeed, this is expected when the channels are emptied without the
loss of structure. In addition, template removal leads to structure
shrinkage of about 5% according to a decrease of the a0 para-
meter from 4.9, 5.1, and 4.6 in DS-NiLA, DS-CuLA, and DS-CuLB
down to 4.7, 4.8, and 4.4 nm in DS-NiLA-E, DS-CuLA-E, and DS-
CuLB-E, respectively. This is likely due to an increase of surface
tension in the channel upon surfactant removal.
The nitrogen sorption isotherms performed on the extracted
materials (Figure 3) are all of type IV according to the IUPAC
classification.37 The steep capillary condensation step at about
0.38 P/P0 evidences a narrow pore size distribution.9,38 Strikingly,
In fact, the metal leaching test using an ethanolic solution of
HCl in excess provides the most striking difference between
(36) Abry, S.; Lux, F.; Albela, B.; Artigas-Miquel, A.; Nicolas, S.; Jarry, B.;
Perriat, P.; Lemercier, G.; Bonneviot, L. Chem. Mater. 2009, 21, 2349.
(37) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.;
Rouquerol, J.; Siemieniewska, T. Pure Appl. Chem. 1985, 57, 603.
(38) Kresge, C. T.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E. Mesoporous
Cryst. Relat. Nano-Struct. Mater. 2004, 148, 53.
(39) Lin, H. P.; Wong, S. T.; Mou, C. Y.; Tang, C. Y. J. Phys. Chem. B 2000, 104,
8967.
Langmuir 2010, 26(16), 13493–13501
DOI: 10.1021/la101379p 13497

Article
Zhou et al.
Table 1. Textural Properties of Mesoporous Hybrid Materials from
N2 Sorption Isotherms at 77 K
Vpa (cm3g-1
( 0.02
)
φBdB
( 0.1 (nm) ( 0.1 ( 0.1 C e
a0
Wtd (nm)
b
c
SBET
material (m2g-1) ( 50
DS-E
1050
910
1.02
0.85
0.76
0.83
0.46
0.45
3.8
3.8
3.9
3.5
3.4
3.3
4.7
4.7
4.8
4.4
4.8
4.7
0.9
0.9
0.9
0.9
1.4
1.4
108
106
99
83
34
-
DS-NiLA-E
DS-CuLA-E
DS-CuLB-E
PS-NiLA
650
1010
860
f
PS-NiLC
590
b
a Total porous volume determined at P/P0 = 0.98. φBJH can be
√
deduced by subtraction of 0.8 to the φBdB value. c a0 = 2d100
/
3. d Wt,
apparent wall thickness deduced from the BdB pore diameter and the
d100 value in the XRD pattern e C parameter from BET equation.
f Copper-amino complex grafted on a similar LUS support containing
1% wt Cu.31
materials. A neat discoloration was found for both DS-CuLB-E
and PS-NiLA, leading to materials DS-3CuLB-H and PS-NiLA-
H, respectively. By contrast, the color was maintained in DS-
NiLA-E and DSCuLA-E. This is fully consistent with the chemical
analysis showing a quantitative metal removal in the former cases
(M/Si molar ratio of 0.004 and 0.003 for DS-CuLB-H and PS-
NiLA-H, respectively, compared to 0.013 and 0.038 in the parent
materials DS-CuLB-E and PS-NiLA, respectively, Supporting
Information Table S1). The latter case (M/Si molar ratio of
0.007 and 0.026 for DS-1NiLA-H and DS-CuLA-H, respectively,
compared to 0.010 and 0.029 in DS-NiLA-E and DS-CuLA-E,
respectively, Supporting Information Table S1) indicates that the
metal retention is about 70% and 90% with LA, while a neat
leaching of ca. 90% takes place with LB.
3. Coordination of the Metal in the Materials. The coordi-
nation state of M-LA systems was characterized by UV-visible
and EPR techniques. The diffuse reflectance UV-visible spec-
trum of the solid DS-NiLA presents two d-d bands at ∼370 and
∼600 nm, corresponding to the ν3 and ν2 transitions, respectively
(Supporting Information Table S3).22
The position of these two bands indicates the presence of two
LA ligands per Ni(II) ion, as in the initial aqueous solution.22
These bands were unchangedafter extraction ofthe template (DS-
NiLA-E) and also after the metal leaching test (DS-NiLA-H). This
shows that the complex was not affected during these two
different treatments. By contrast, the disappearance of the d-d
bands in the visible region for the postsynthesis material, PS-
NiLA, after the metal leaching test is consistent with the quanti-
tative Ni removal characterized above from chemical analyses
(Supporting Information Table S1).
Figure 4. EPR of DS-CuLA, DS-CuLB, and DS-CuLB-E. Experi-
mentalconditions: frequency=9.35GHz;power= 10mWfor DS-
CuLA and DS-CuLB, 40 mW for DS-CuLB-E; modulation ampli-
tude=1 G for DS-CuLA 6 G for DS-CuLB, 9 G for DS-CuLB-E,
and modulation frequency = 100 kHz.
(=2.200) and a slightly smaller hyperfine value A =18.8 mT con-
sistent with a weaker crystal field in the equatorial plan (xy) than
the solid inserted species. This indicates that the environment
shifts either toward a square pyramidal symmetry expected with a
strong ligand in the z axis or toward an elongated octahedral
environment obtained with a pair of weak ligands in axial posi-
tion. Since the siliceous matrix may provide either silanol or
silanolate groups and since the former is close to water in the
spectrochemical series, the most probable ligands here are the
silanolate groups. The latter are σ and π donating group like OH-
and are weaker than H2O.42 The g^, less sensitive to any crystal
field effect, is not commented here and is indeed nearly the same
(∼2.050) in solution and in the solid. These conclusions are
consistent with an overall d-d bathochromic band shifting from
18000 cm-1 down to ca. 17000 cm-1 for complexes in solution to
material inserted species.
The integrity of ligand LB in the materials was monitored from
the IR spectrum, looking at the band at 1623 cm-1 assigned to the
stretching vibration, ν(CdN), of the imino group (Scheme 1).
Note that the ligand and metal-free DS-E material exhibits a small
band at 1630 cm-1 (deformation, δHOH) reminiscent of H2O
traces. A slight shoulder at 1623 cm-1 could hardly be seen and
precludes any definitive conclusion on the presence of the imine
group in metal-free materials. Nonetheless, note that the ν(CdN)
band is expected to shift upward by about 20 cm-1 upon complex-
ation as reported earlier.36 The progressive increase in intensity
and the shift of the band from 1630 up to 1650 cm-1 with higher
metal loading in the DS-CuLB series (M/Si = 0.5, 0.7, and 1.6 mol
%) is consistent with the imino group being engaged in the copper
complex (Supporting Information Figure S1). This band moves
back down to 1630 cm-1 in material DS-CuLB-H consistent with
metal decomplexation and leaching after the acid treatment.
The UV-visible investigation reveals that the spectrum of DS-
CuLB, DS-CuLB-E, and DS-CuLB-H materials exhibit in all cases
three main absorption bands below 300 nm, which can be attri-
buted to π-π* transitions of the LB ligand either free or com-
plexed to the copper ion (Figure 5 and Supporting Information
Material DS-CuLA exhibits a single axial EPR signal for d9
Cu(II) cations with g =2.19 and g^=2.050 characteristic of a
dx2-y2 ground-state matching either with a square planar, tetra-
gonal distorted octahedral, or square pyramidal environment.
The hyperfine coupling constant of the parallel component, A ,
was 18.8 mT (Figure 4) compares well with the characteristics of
[Cu(en)2(H2O)2]2þ elongated octahedral species (Supporting In-
formation Table S4).40,41 This is consistent with two LA ligands
per Cu(II) ion in the solid material. Note that a material prepared
with a LA/Cu molar ratio of 1 instead of 2 exhibits two EPR
signals consistent with a mixture of [Cu(LA)]2þ and [Cu(LA)2]2þ
inserted complexes. Though it appears that material DS-CuLA
contains a single copper(II) species, there are some slight differ-
ences with the genuine [Cu(LA)2]2þ complex. The latter like [Cu-
(en)2]2þ are characterized in solution by a slightly larger g
(40) Lewis, W. B.; Alei, M.; Morgan, L. O. J. Chem. Phys. 1966, 45, 4003.
(41) Ganesan, R.; Viswanathan, B. J. Phys. Chem. B 2004, 108, 7102.
(42) Bonneviot, L.; Legendre, O.; Kermarec, M.; Olivier, D.; Che, M. J. Colloid
Interface Sci. 1990, 34, 534–547.
13498 DOI: 10.1021/la101379p
Langmuir 2010, 26(16), 13493–13501

Zhou et al.
Article
further acid treatment that yields metal leaching (DS-CuLB-H),
the low-energy band shifts from 370 nm down to 390 nm is
-
consistent with a decomplexation of LB . It is worth noting that
the presence of the peak at 310 nm in the spectrum of the starting
material DS-CuLB though rather weak is consistent with some
noncomplexed HLB ligands located in a nonpolar environment,
i.e., more likely entrapped in the pore wall. Furthermore, an
additional shoulder at ca. 470 nm was observed, which can be
tentatively assigned to a ligand-to-metal charge transfer (LMCT)
band.44-46 This band indeed disappears after metal extraction
(Figure 5).
Finally, there is a weak and broad band centered at ca. 635 nm
due to d-d transitions of copper(II) ion. This band is lower in
energy than the corresponding band of the [Cu(LB)2] molecular
complex in ethanol (622 nm, Supporting Information Table S5),
while the position is about the same in the spectrum of the
-
aqueous solution of the complex (2 equiv LB versus Cu). This
is consistent with a mixture of mono(salicylaldimine) copper(II)
and bis(salicylaldimine) copper(II) complexes both in water and
in the solid hybrid materials as already mentioned from the
discussion of the n-π* band position.
Figure 5. Solid-state UV-visible spectra of copper containing
materials as synthesized, after template extraction, and after acid
leachingtestin materials DS-CuLB, DS-CuLB-E, and DS-CuLB-H,
respectively: the horizontal bars indicates the wavelength domains
for the different electronic transitions in the inserted complex.
The EPR spectrum of material DS-CuLB (Figure 4) confirms
the presence of two species: one with a g tensor mostly axial
characterized by g = 2.31, A = 13.9 mT, and the other by g =
2.22, A =17.2 mT. Both are consistent with a dx2-y2 ground state
and with an elongated octahedral environment, as above. The
former is assigned to 1N3O environment in the equatorial plan
around copper(II) ion and the latter with the 2N2O environment,
Table S3). These bands are also observed in the molecular
analogues in solution (Supporting Information Figure S2). The
complexation to copper affects only their relative intensity and is
hardly exploitable per se. In the range 300-450 nm, two more
bands are observed, also in the free ligand and in the complex. By
contrast with the previous bands, their position depends on the
solvent polarity and on the complexation to a metal ion. They are
assigned to n-π* transitions of the imino group. The phenolic
proton of the free neutral ligand in nonpolar solvents is located on
the phenolic oxygen, giving rise to a band at ca. 330 nm. This band
position is associated with the imino group with a free electron
pair nonengaged in any sort of bonding. In polar solvents, the
proton is transferred onto the nitrogen of the imino group, and
the band undergoes a bathochromic shift down to ca. 400 nm
reminiscent of the engagement of the electron pair in a bonding
with the displaced proton.43 In fact, both bands are often
observed at the same time as a result of both proton locations
being in equilibrium one with another. HLB in ethanol exhibits
two bands at 310 nm (strong) and 400 nm (weak) revealing the
existence of an equilibrium in favor of the oxygen phenolic
position. Inversely, the imino-nitrogen position, which generates
a polar species, is the most favored situation in water, with a single
band at 400 nm (Supporting Information Figure S2-A). In the
presence of Cu(II) ions and in ethanol, the band position shifts to
the intermediate position of 360 nm (Supporting Information
Figure S2-B). In water, the bis hemisalen, Cu(LB)2, is in equilib-
rium with the monohemisalen complex, Cu(LB)(H2O)2, as re-
vealed by EPR (vide infra). As a consequence, some of the ligands
HLB are free in the water solution. The resulting band is broad
and is consistent with a mixture of two bands at about 360 and
380 nm. Decreasing the pH from 11 down to 6 favors the shift of
the band up to 380 nm, consistent with a complete leaching of the
copper ion (Cu-LB-W11 to W-6, Supporting Information Figure
S2-B). By comparison, in the absence of copper, the n-π* band of
the free ligand HLB does not shift in the pH range 6-11. For the
solid DS-3CuLB, there are also two bands observed at ca. 310 and
370 nm (Figure 5).
-
consistent with one and two LB ligands per Cu(II) ion, respec-
tively. Similar values have been reported by Murphy et al. for the
same complexes grafted on nonmesostructured porous silica.26
Indeed, such distribution in the solids is reminiscent of the
equilibrium between mono- and bis-LB complexes in water
solution.47,48 This mixture of copper species in the DS-3CuLB
materials is in full agreement with the conclusion drawn from the
UV-visible spectra and with the presence of some noncomplexed
ligand HLB entrapped in the pore wall.
4. Location of the Metal Complexes in the Solid Materi-
als. All the above results allow us to confirm that the direct
synthesis using LA leads to inaccessible [Ni(LA)2]2þ and [Cu-
(LA)2]2þ complexes that are embedded in the pore walls and called
as such in Scheme 2. The new spectroscopic characteristics of the
incorporated complexes and the resulting inaccessibility are
related to a direct coordination of the metal ions with silanolate
moieties of the siliceous network. Note that the metal coordina-
tion is well-defined and can be exploited to tune the redox
potential and/or the optical properties of the inorganic porous
structure for bioinspired designed of catalysts.31-33 At the ex-
pense of the complex integrity, these complexes might be rendered
accessible using calcination at high temperature as shown by
Karakassides et al. for direct synthesis HSM or MCM materials
containing copper(II) using also LA ligand.11,12
In material DS-CuLB-E, two complexes are present, i.e.,[Cu-
(LB)2X2] and [Cu(LB)XL] complexes (X = SiO- or OH- and
L = H2O or silanol ligands). Both are likely located in the wall
according to the pore wall thickness and the high value of the pore
volume. However, their weak resistance to the acid leaching test is
(44) Salavati-Niasari, M.; Mirsattari, S. N.; Bazarganipour, M. Polyhedron
2008, 27, 3653.
(45) Kurzak, K.; Kuzniarska-Biernacka, I.; Zurowska, B. J. Sol. Chem. 1999,
Upon careful extraction of the template, there is no change in
the UV spectra (compare DS-CuLB and DS-CuLB-E). Upon
28, 133.
(46) Kuzniarska-Biernacka, I.; Kurzak, K.; Kurzak, B.; Jezierska, J. J. Sol.
Chem. 2003, 32, 719.
(47) Baleizao, C.; Garcia, H. Chem. Rev. 2006, 106(9), 3987.
(43) Rospenk, M.; Krol-Starzomska, I.; Filarowski, A.; Koll, A. Chem. Phys.
2003, 287, 113.
Langmuir 2010, 26(16), 13493–13501
DOI: 10.1021/la101379p 13499

Article
Zhou et al.
Scheme 2. “Showing-On” and “Embedded” Framework Species in
Materials DS-CuLB-E and DS-NiLB-E or DS-CuLA-E, respectivelya
used as catalyst. The characterization of this catalyst is poorly
reported and does not allow clear comparison.49 Other mesopor-
ous materials with Cu(II) ions, such as Cu-SBA-15,50 Cu-HMS,51
and Cu-exchanged Y zeolites,52 favor the formation of CAT,
while TS-1 and Cu-CMM favor the formation of HQ. Despite the
different local structure around copper, the catalytic activity of
the showing-on sites does not differ much with previous zeolite or
mesoporous silica-supported copper systems cited above.
After the first run, DS-CuLB-E was recovered by filtration and
then washed with ethanol and acetone and dried at 60 ꢀC for 18 h.
This so-obtained material was characterized by EPR spectrosco-
py. The copper EPR signal decreases by about 20%. This result
suggested that the metal had partially leached out. In the second
run, the reaction was run in the same conditions than for the
first run, therefore with more catalysts to keep the phenol to
copper ratio at 130. The rather identical activity and selectivity
profiles shows that copper(II) is necessary to observe such
performance. However, the loss of copper precludes any further
conclusions on the real catalytic sites that could be either the
inserted complex or the leached copper (ii) species. The equilib-
rium observed in solution for the molecular analogues was
already an indication that LB is a much weaker ligand than LA
and might be not the best candidate for catalytic applications.
This may be the reason for the nonspecific difference between the
present catalysts and the above cited systems where slight leaching
may also occur.
a L is any neutral ligand such as water or silanol; for the sake of
simplicity, the axial ligand are not represented for the “showing-on”
species.
Table 2. Catalytic activity on phenol hydroxylation with H2O2 over
various catalystsa
phenol: phenol conv.
H2O2 (mol %) selectivity (mol %) HQ
diphenols
CAT/
material
DS-E and DS-CuLA-E 1:1
9
17
15
0
85
84
0
1.8
1.9
DS-CuLB-E, 1st run
DS-CuLB-E, 2nd run
3:1
3:1
a Reaction conditions: solvent, buffer solution pH=6; phenol weight,
0.0989 g; phenol:Cu = 130:1; t = 2 h; T = 80 ꢀC; [phenol] = 0.15 mol/L.
After the first run, 80% of Cu(II) species remained in the catalysts
(see text), and the catalyst mass was therefore corrected to afford the
same phenol/Cu molar ratio.
Conclusions
Hybrid mesostructured porous materials containing Ni2þ or
Cu2þ complexes with bidentate organosilane amino ligands in the
framework have been synthesized. A direct synthesis has been
optimized in order to condense the complex precursors with
sodium silicate in aqueous solution, maintaining the structure
integrity of both the porous material and the metal complex. The
location of the complex in the direct synthesis material is con-
trolled using two condensable organosilane ligands (Scheme 1):
N-(2-aminoethyl)-3-aminopropyltrimethoxysilane, LA, and, N-
salicylaldimine-propylamine-trimethoxysilane, HLB. The former
ligand leads to inaccessible complexes with a coordination locked
by two silanolate groups as exemplified in both DS-NiLA and DS-
CuLA materials; they are named “embedded” sites. By contrast,
reminiscent of their sensitivity to proton attack and their pro-
pensity to leach out Cu(II) ions. Then, it is proposed that both
sites in DS-CuLB-E are showing on the pore wall surface like
outcrops and are called as such in Scheme 2. The rationale for
-
such difference between LA and LB resides both in the more
hydrophobic nature and in the neutrality of the complexes formed
by the latter in comparison with the former. Both parameters are
indeed strong driving forces to favor the location of the complexes
in the organic side of the electrical palisade between the micelle
and the aqueous solution during the initial steps of the silicate and
organosilane moieties condensation. On the contrary, the ligand
LA favors direct binding of the metal to framework SiO- anions
leading to chemically inaccessible sites inside the pore wall.
5. Catalytic Activity of “Showing-On” Sites. To further
probe the accessibility of the showing-on sites, the DS-CuLB-E
material was tested as heterogeneous catalyst for phenol hydro-
xylation using H2O2 as oxidant (Table 2). The reaction was
performed at 80 ꢀC and at pH=6 (buffer solution) using a phenol/
Cu and a phenol/H2O2 molar ratio of 130:1 and 3:1, respectively.
Note that no reaction takes place in these conditions without
H2O2. Both metal-free DS-E and DS-CuLA-E materials exhibited
a conversion of ca. 9%, slightly higher than in the absence of any
of both materials (∼6-8%), but no selectivity toward catechol
and hydroquinone was observed in all these cases. By contrast,
DS-CuLB-E was found active (17% conversion) in phenol hydro-
xylation, 85% of which was catechol (CAT) and hydroquinone
(HQ) with a CAT/HQ ratio of 1.8. The remaining 15% of
products corresponded to byproducts. By contrast, Ray et al.
reported that the predominant product is the hydroquinone,
while no benzoquinone was formed when bis(salicylaldimine)
copper(II) complexes immobilized in the channel of MCM-41 are
-
LB , which prevents direct bonding between the silanolate groups
and the metal ion, yields accessible and catalytically active sites as
shown in the DS-CuLB material. These framework sites are
named here “showing-on” (Scheme 2). By contrast with the post-
grafted complexes, both embedded and showing-on complexes do
not occupy the pore volume and do not affect the apparent pore
wall thickness according to the nitrogen sorption investigations.
Tuning independently both location and accessibility of the metal
sites presents new opportunities to design multifunctional materi-
als. For instance, inaccessible embedded sites may intervene as
mere redox regulation centers, optical traps, or fluorescence
centers while showing-on sites may be chosen for a specific
catalytic activity and grafting of sites (metallic or organic) may
be added to tune specific adsorption, acid-basic, or steric pro-
perties of the surface.
(49) Ray, S.; Mapolie, S. F.; Darkwa, J. J. Mol. Catal. A: Chem. 2007, 267, 143.
(50) Wang, L. P.; Kong, A. G.; Chen, B.; Ding, H. M.; Shan, Y. K.; He, M. Y. J.
Mol. Catal. A: Chem. 2005, 230, 143.
(51) Fu, Z. H.; Chen, J. H.; Yin, D. L.; Yin, D. H.; Zhang, L. X.; Zhang, Y. Y.
Catal. Lett. 2000, 66, 105.
(52) Wang, J.; Park, J. N.; Jeong, H. C.; Choi, K. S.; Wei, X. Y.; Hong, S. I.; Lee,
C. W. Energy Fuels 2004, 18, 470.
(48) Zhou, W. J. Ph.D. Thesis, University of Lyon, France and East China
Normal University, China, 2009.
13500 DOI: 10.1021/la101379p
Langmuir 2010, 26(16), 13493–13501

Zhou et al.
Article
Acknowledgment. Wen-Juan Zhou thanks the cooperative
and Cu-LB complexes in solution), Table S1 (metal content
in mol % and molar ligand to metal ratio), Table S2 (yields
on metal, ligand and silicon for the materials), Table S3 (IR
and UV-visible data of the ligand and copper complexes),
Table S4 (EPR characteristics of complexes in solution and
in the materials), and Table S5 (IR and UV-visible data for
the molecular ligand and metal complex analogues).This
material is available free of charge via the Internet at
http://pubs.acs.org.
ꢀ
program between Ecole Normale Superieure (ENS) de Lyon and
East China Normal University (ECNU) in Shanghai, and thanks
both French Government for the Eiffel Doctoral Scholarship and
Chinese Government for Doctoral Scholarship to support her
long stay in France.
Supporting Information Available:
Figure S1 (FT-IR
spectra of materials), Figure S2 (UV-visible spectra of LB
Langmuir 2010, 26(16), 13493–13501
DOI: 10.1021/la101379p 13501