52
J.W. Brown et al. / Catalysis Communications 59 (2015) 50–54
the exception that in our heterogeneous system the MOF catalyst could
easily be separated from the reaction solution at the end of the catalytic
run using a simple filtration step. The epoxidation reactions were per-
formed by bubbling O2 (30 sccm) through a mixture of methylene chlo-
ride (10 mL), isobutyraldehyde (0.913 mL, 10 mmol), alkene (2 mmol),
MOF (5 mg) and heptane (1.0 mmol, inert internal standard) in a reac-
tion vessel for an extended period of time. The reaction mixture was
monitored using gas chromatography–mass spectrometry (GC–MS).
The results for catalysis of different alkenes by both MOF-525-Mn and
the free TCPP-Mn linker are summarized in Table 1, which lists the
main reaction products for each alkene reactant and corresponding re-
action yields and conversion rates. Note that the sizes of all alkenes
and sacrificial aldehyde are small enough to fit inside the MOF pore
(19 Å in diameter), allowing full access to all active sites in the frame-
work. Some yields and conversion rates for the smaller alkenes are
close to maximal for MOF-525-Mn. Ethylene and propylene were not in-
vestigated in this study due to the gas-phase nature of these reactions,
requiring an entirely different reaction setup than the one used here.
In the case of styrene epoxidation, GC–MS analysis revealed a series
of reaction products, with the main reaction product being styrene
oxide (see SI, Fig. S6). Due to the large number of side products, we op-
timized the styrene reaction yield by varying reactor temperature.
While this optimization technique could be applied to the other alkene
reactions, we decided to focus on styrene and therefore analyzed the
reaction kinetics to determine the main parameters of the epoxidation
reaction pathway (Fig. 1). Industrial applications of styrene epoxidation
include epoxy resins and production of chemicals such as phenethyl
alcohol and styrene glycol. [21]
where Ea,i is the activation energy for the i-th pathway, R is the ideal gas
constant, T is the temperature and Ai(T) is the pre-exponential prefactor.
The pseudo-rate constant for styrene epoxidation, k'1, is related to the
inherent rate constant, k1, according to k'1 = k1[O2], where [O2] is the
concentration of oxygen dissolved in the solvent.
The apparent rate constants and reaction orders were computed
over the temperature range 293–329 K by fitting the decline in styrene
concentration over time (see SI section for details of the procedure). The
pseudo-rate constants at different temperatures were used to derive the
activation energy of the reaction from a fit to the Arrhenius equation.
The temperature dependence of the reaction pseudo-rate constants
for MOF-525-Mn was calculated from experimental data. A linearized
Arrhenius plot showing the fit over the temperature range 293–329 K
is found in Fig. 2. The styrene epoxidation reaction was found to be of
the first order with respect to styrene, and the activation energy for
the epoxidation of styrene was found to be 42.6 5.7 kJ/mol. This acti-
vation energy agrees with calculated theoretical values within experi-
mental error [22]. (We are unaware of any experimentally measured
values for Ea reported in the literature at the time of writing.) The
reaction order for this heterogeneous catalyst is in agreement with
published results from homogeneous manganese porphyrin systems
[23]. This result suggests that the manganese porphyrin heterogeneous
catalyst likely operates according to a mechanism similar to its homoge-
neous counterpart.
The heterogeneous nature of the epoxidation reaction was confirmed
by filtration. After the initial reaction of styrene, the MOF was filtered
and fresh isobutyraldehyde and styrene were added to the filtrate.
Using the filtrate, the reaction was run again under analogous conditions
to the initial experiment and no catalytic conversion was observed. In
addition, no catalytic conversion was observed if isobutyraldehyde,
MOF-525-Mn, or metalated porphyrin was not present in the initial re-
action mixture (see SI).
MOF-525-Mn showed minimal deactivation and maintained its
structural stability and crystallinity at the end of the catalytic cycle, as
evidenced by structural analysis with PXRD (Fig. 3). Leeching of manga-
nese and porphyrin from MOF-525-Mn was investigated by performing
ICP–AES and elemental analysis on the product solution after filtration.
We found no evidence of leeching within experimental error of these
measurement techniques (see SI). X-ray photoelectron spectroscopy
(XPS) measurements indicated an oxidation state of Mn(III) within
the framework and no change in oxidation state was observed after ca-
talysis (Fig. 3). To test catalytic activity after multiple runs, several cycles
of epoxidations were run back-to-back using the same MOF sample, fil-
tering it between each run while monitoring the reaction via GC–MS.
The reactor was modeled as a constant-volume batch reactor. Rapid
stirring of the reaction mixture served to ensure homogenous mixing of
all components in the reactor, and the concentration of dissolved oxy-
gen in the reaction mixture was assumed constant due to the large
and continuous flux of oxygen gas into the reaction vessel. The reaction
pathway is assigned the pseudo-rate constants k'1, k2' , and k3' . The latter
are assumed to follow the Arrhenius behavior:
0
a;i=RT
kiðTÞ ¼ AiðTÞe−E
ð1Þ
Fig. 1. Reaction pathway for the epoxidation of styrene to styrene oxide including side
products. Relative yields of each species measured at equilibrium are shown as
percentages.
Fig. 2. Arrhenius plot of the rate constants k'1 over the temperature range 293–329 K.