146
M. Rocha et al. / Journal of Catalysis 361 (2018) 143–155
experiment was carried out in a 50 mL round bottom flask. In this
case, the electronic spectra of the reaction mixtures were acquired
every minute by withdrawing 3 mL aliquots from the reaction
medium; the reactions were stirred at controlled temperature
(25 °C). After each catalytic cycle, the catalysts were filtered and
then reused in a new cycle. This procedure was repeated 9 times
and at the end of the 10 reuse cycles, the materials were dried
and characterized as described in the results and discussion section
adsorbed water), respectively [34,35], a strong and broad band at
1045 cmÀ1 with a shoulder at ꢁ1093 cmÀ1 corresponding to
SiAOASi asymmetric stretching vibrations and a shoulder at 917
cmÀ1 due to AlAOHAAl bending vibrations [34–36]. The band near
800 cmÀ1 could be assigned to several vibrations namely to the
presence of quartz mixture in the sample [35], to SiAOASi vibra-
tions [37] or to AlAOASi vibrations [38]. The band at 525 cmÀ1 cor-
responds to SiAOAAl (octahedral Al) vibration and the band at 468
cmÀ1 is assigned to SiAOASi bending vibrations [36,39].
(Section 3.3.2). Control experiments were performed with the L-
serine-functionalized clays (K10_A and K10_B) using the same cat-
alytic reaction conditions. Adsorption of 4-NP substrate into the
catalysts was also evaluated using similar catalytic reaction condi-
tions, but in the absence of NaBH4. The catalytic performance of
gold immobilized directly onto K10 was also assessed on the 4-
NP reduction.
Upon silylation, the FTIR spectra of the two materials clearly
show characteristic bands of the grafted organosilanes. In the case
of K10_A (Fig. 1a), the bands at 2915 and 2935 cmÀ1 are due to
CAH stretching vibration modes, the band at 1687 cmÀ1 corre-
sponds to the carbonyl group of the ester moiety (C@O stretching).
The bands at 1630 and 1570 cmÀ1 are assigned to the carbonyl
group stretching vibration (superimposed with HAOAH bending
vibrations from adsorbed water) and to the NAH bending of carba-
mate, respectively. The band at 1440 cmÀ1 is assigned to CH2 bend-
ing deformations [32]. In the case of K10_B (Fig. 1b), the peak at
3040 cmÀ1 is due to stretching vibration modes of aromatic CAH
bonds, the peaks at 2925 and 2870 cmÀ1 correspond to CAH
stretching vibrations, the band at 1690 cmÀ1 is related to vibration
modes of superimposed carbonyl groups of carbamate and amide
moieties, the band at 1555 cmÀ1 is assigned to NAH bending of
carbamate moieties and the band at 1455 cmÀ1 is assigned to
CH2 bending vibrations [32].
3. Results and discussion
3.1. Characterization of
L-serine-functionalized K10 supports
Two -serine-functionalized K10 montmorillonite clay materials –
L
K10_A and K10_B – were prepared as shown in Scheme 1 and used as
solid supports for the immobilization of Au NPs. 1H and 13C NMR and
ESI-MS were used to monitor the formation of the intermediate
serine derivatives; elemental analyses, FTIR and solid-state NMR
(
29Si and 13C) were performed in order to characterize the prepared
Solid-state 13C CPMAS NMR gave vital information about the
solid supports K10_A and K10_B.
surface silicon environments and attached L-serine silanes, specif-
The 1H NMR data of the
L-serines 2 and 3 is provided in Figs. S2
ically of their grafting mechanisms. The solid-state 13C CPMAS
NMR spectra of K10_A and K10_B are presented in Fig. S4 in the
Supporting Information. In both spectra, carbonyl signals are iden-
tified near d = 160 and 173 ppm and the signals corresponding to
CH, CH2 and CH3 are in the range of d = 5–60 ppm, as expected [32].
The spectrum of K10_B also shows peaks at d = 127 and 144
ppm corresponding to the phenyl ring. The intensity of the peaks
corresponding to SiOCH2CH3 and SiOCH2CH3 carbon signals
(expected near d = 20 and 60 ppm, respectively) is in accordance
with a successful bidentate grafting and a possible partial hydrol-
ysis of the remaining ethoxy group to afford a silanol [R–Si(OH)
(O–K10)2] or a lateral condensation of silanes [R–Si(O–Si(O–K10)
and S3 in the Supporting Information. 13C NMR spectroscopy was
also performed for 2 (all these spectra are provided in the Support-
ing Information). For compound 2 it is possible to clearly identify
the methyl ester group by the presence of signals corresponding
to OCH3 and C@O groups in the 1H and 13C NMR spectra, respec-
tively. The structure of compound 3 is also easily confirmed by
the presence of new signals corresponding to the NH amide bonds
and benzyl moieties.
The silylation of L-serines 2 and 3 was performed by following
our previously described protocol [32], and the formation of the
desired compounds 4/5 was confirmed by TLC and by ESI-MS spec-
trometry. The silylated
L
-serines 4 and 5 were not isolated because
2–R)(O–K10)2] (R = L-serine organosilane) [32].
the silylation reactions were not complete and the purification of
these compounds through chromatographic column results in sig-
nificant losses due to hydrolysis of the triethoxysilane group.
The grafting of the silylated serines 4 and 5 onto K10 was also
performed following our previously described methodology, in
which the mechanism was also studied [32]. The K10_4 and
K10_5 materials were subsequently treated with piperidine in
order to deprotect the amine group, thus obtaining the desired
materials K10_A and K10_B, respectively. As previously men-
tioned, the success of this step was confirmed by the ninhydrin
test, which gave negative results for K10_4 and K10_5 and positive
results for K10_A and K10_B. These tests clearly confirmed the
The 29Si MAS NMR spectra of K10_A and K10_B are presented in
Fig. S5 in the Supporting Information, in which the peaks are iden-
tified according to the commonly used Qn and Tn notation and the
chemical shifts are summarized in Table 1.
The spectra present an intense peak corresponding to Q4 reso-
nances (d = À111 ppm) which is attributed to fully condensed silica
units that do not react with the organosilane molecules and two
smaller peaks at d = À93 and À102 ppm corresponding to Q2 and
Q3 resonances, respectively, and are associated to the germinal
silanol groups and one terminal silanol group [32,40–42]. The
spectra also exhibit additional peaks at d = À57 ppm and d = À58
ppm for K10_A and K10_B, respectively, and a peak at d = À67
ppm in both materials associated to T2 and T3 silicon environ-
grafting of the L-serine organosilanes onto K10 and presence of free
amine groups on the resulting materials surface (see Fig. S1 in the
Supporting Information).
ments, revealing the anchorage of organosilated L-serines by
bidentate and tridentate grafting on the K10 surface [32,43,44].
In both K10_A and K10_B the T2 and T3 sites are observed with
similar intensity. However, the possibility of lateral condensation
of grafted organosilanes – which results in T3 signals – also needs
to be considered. These observations are in agreement with previ-
ously reported results for related materials [32,42].
The functionalization of K10 with L-serine derivatives was
assessed by FTIR, solid-state NMR (29Si and 13C) and elemental
analyses. The FTIR spectra of the resulting materials K10_A and
K10_B are presented in Fig. 1.
For all samples, characteristic vibrational bands of montmoril-
lonite clay are present in the corresponding FTIR spectra: a band
around 3620 cmÀ1 corresponding to stretching vibrations of OAH
groups coordinated to octahedral cations (mainly Al, but also Mg
and Fe) [34], bands at 3394 and 1640 cmÀ1 assigned to OAH
stretching and bending vibrations of interlayer water (physically
The relative concentrations of the Qn and Tn sites ascertained by
deconvolution of the 29Si MAS NMR spectra are presented in
Table 1. They allow estimating the amount of grafted organosilane
in K10 as 10.7 and 9.0% for K10_A and K10_B, respectively (based
on (T3 + T2)/(T2 + T3 + Q2 + Q3 + Q4)).