80
P.A.A. Ignacio-de Leon et al. / Applied Catalysis A: General 511 (2016) 78–86
cyanopropyltriethoxysilane in a mixture of ethanol, water and
n-dodecylamine [59]. Carboxylic acids were afforded by subse-
quent nitrile hydrolysis in aqueous H2SO4 at reflux conditions
[60–63]. All samples were dried under dynamic vacuum prior
to further use. Actual organic loadings in the hybrid SiO2 mate-
rials were quantified by TGA, as detailed below. Carbon (C,
Calgon brand, granular, <1000 m) was treated with 2 M HCl
then rinsed with 18 Mꢀ water, ethanol and ether. Acid-treated
carbon (AC) was obtained by further treatment with 5 M HNO3,
followed by copious rinsing with 18 Mꢀ water [54]. See the sup-
porting information for full details on molecular and materials
syntheses.
3. Results and discussion
Commercially available mesoporous SiO2 gel particles are mod-
ified by grafting silane esters from a colloidal mixture in pyridine,
followed by hydrolysis (Scheme 2). The overall particle morphology
is maintained by SEM (Supporting information Fig. S1) after graft-
ing. Solid-state 13C CP/MAS NMR (Supporting information Table
S1) shows resonances at 6, 27, and 177 pm for the final grafted
acid, with loss of a resonance at 50 ppm from the intermediate
methyl ester, supporting the successful covalent anchoring of the
organosilanes, and IR spectra for unmodified and functionalized
carbonyl and alkyl stretches at ca. 1700 and 2900 cm−1, respectively
(Supporting information Fig. S2).
2.4. Epoxidation of p-substituted styrenes
N2 physisorption on the solid co-catalysts (Supporting infor-
mation Fig. S3) gives surface areas from the BET model and pore
dimensions from the BJH model (Table 1). The grafted materi-
als remain mesoporous but show small decreases in surface area
and pore diameter relative to the parent SiO2, as expected for
sub-monolayer coverage of the organosilanes. The co-condensed
material used is also mesoporous, while the carbon sample has
bimodal porosity. TGA quantifies the loss of organics during com-
bustion of the solid co-catalysts to estimate carboxylate molar
in color after heat treatment to 800 ◦C in O2, consistent with
complete combustion of the organics. Mass loss occurs between
350 and 550 ◦C. Loadings (mmol/g) and average surface densities
(groups/nm2) are given in Table 1 and are calculated assuming loss
of a surface C3H5O2 fragment for PA-SiO2, C7H5O2 for BA-SiO2, and
C4H7O2 for the sol–gel material. Overall, the loadings, pore struc-
tures, and functional group assignments for the supports are very
comparable to similar materials reported by some of us in previous
studies [54]. A detailed analysis of the Mn2-modified materials has
Styrene (≥99.0%, ReagentPlus, Sigma–Aldrich), styrene
oxide (97%, Aldrich), 4-methylstyrene (≥99.0%, 0.005% 4-tert-
butylcatechol as stabilizer, Aldrich), 4-vinylanisole (97%, Aldrich),
4-cyanostyrene (97%, stabilized with 0.05% 4-tert-butylcatechol,
Alfa Aesar), 4-trifluoromethylstyrene (98%, contains 0.1% 4-tert-
butylcatechol as inhibitor, Aldrich), divinylbenzene (DVB, 80%,
technical grade, 1000 ppm tert-butylcatechol as inhibitor, Aldrich),
anisole (anhydrous, 99.7%, Sigma–Aldrich), and benzonitrile
(anhydrous, >99%, Sigma–Aldrich), were passed through an acti-
vated alumina column in order to remove inhibitors and other
imputities and were stored at 2–6 ◦C until use. Acetonitrile (Spec-
trophotometric grade, 99.7+%, Alfa Aesar), 1,2-dichlorobenzene
(99%, Alfa Aesar), 3-chloroperoxybenzoic acid (77% max., Aldrich),
H2O2 (30 wt% in water, ACS reagent, Sigma–Aldrich) and Ag
(powder, <250 m, 99.99% trace metals basis, Aldrich) were used
as received. Standard catalytic tests were performed at 0 ◦C at
3000:1000:1:10 H2O2: C C: complex: carboxylate mole ratios.
Note that the catalyst was added based on moles of the complex
and alkene based on moles of the double bond; thus, reactions
with Mn2 contain twice as many Mn atoms as for reactions using
Mn1 and reactions with DVB contain half as many moles of the
substituted styrene as for the other reactants. 1,2-Dichlorobenzene
was used as an internal standard.
In a standard reaction, a measured amount of the carboxylic
acid-modified SiO2 was added to a vial equipped with a stir bar,
followed by 2.5 mL of 0.4 mM Mn1 or Mn2 in acetonitrile. The mix-
ture was allowed to equilibrate for 40 min at 0 ◦C with stirring
on a cold plate, after which 7.5 mL of a 0.15 M alkene solution
in acetonitrile was added. Timing commenced upon addition of
0.7 mL of a solution of dry H2O2 in acetonitrile (prepared by dilut-
ing 10 mL 30% aqueous H2O2 with 20 mL acetonitrile containing
7 g anhydrous MgSO4 followed by decanting the supernatant).
Aliquots were withdrawn through a Whatman GF/F syringe fil-
ter at certain time points over a 24 h interval and delivered
to GC vials containing approximately 10 mg Ag powder to con-
sume any unreacted H2O2 and stabilize the oxygenates prior
to GC–MS analysis. Analyte concentrations were quantified with
a Shimadzu GCMS-QP2010 SE on a Phenomenex ZB-624 capil-
lary column (30m × 0.25 mm diameter × 1.4 m thickness) based
from prepared calibration standards of 0–5 mM analyte in ace-
tonitrile. p-X-styrenes, styrene oxide, 1-phenyl-1,2-ethanediol,
benzaldehyde, phenylacetaldehyde, acetophenone and divinylben-
zene were calibrated against commercially available standards.
Ethylvinylbenzene monoxide (EVBMO), divinylbenzene monoxide
(DVBMO) and divinylbenzene dioxide (DVBDO) were calibrated
against authentic samples obtained as a gift. Calibration standards
for p-X-styrene oxides were prepared from epoxidation with excess
m-chloroperbenzoic acid.
3.2. Catalytic epoxidation of styrene
Fig. 1 shows typical kinetic profiles for the first 2 h of the
epoxidation of styrene with H2O2 using Mn1 and Mn2 supported
ings of BA-SiO2 and PA-SiO2 were used, 0.87 and 0.42 mmol/g.
Styrenics are well-studied benchmark reactants, [64]. which also
these catalysts exhibit a ∼15 min induction period following H2O2
addition. We and others have documented this induction period
and attribute it to assembly of the catalyst–carboxylate com-
plex [53,67]. Major products styrene oxide and styrene diol, and
rearrangement/overoxidation products phenylacetaldehyde and
benzaldehyde (presumably from oxidative cleavage, Scheme 3)
increase monotonically in concentration for 300–450 min, after
which product concentrations generally plateau. See the support-
ing information for full kinetic traces out to 24 h (Supporting
information Fig. S5). Interestingly, Mn1/BA-SiO2 shows has an ini-
tial epoxidation rate that is <10% that of Mn1/PA-SiO2, but its
productivity remains steady up to and past 450 min, ultimately
ity to epoxide exceeded 90% in all cases. Mass balance, expressed as
the total mmol of all observed products relative to mmol consumed
alkene, is >95%, demonstrating that all key products are accounted
for by this analysis.
Table 2 gives the initial rates and total turnover numbers (TON,
mole product per mole of Mn1 or Mn2, not per Mn atom) and prod-
uct selectivity at 24 h. Initial rates are calculated from sustained