M.P. House et al. / Journal of Catalysis 252 (2007) 88–96
89
2
. Experimental
He, except in the case of extended reduction, in which 3-µL
injections were made every 2 min. The TPD experiments were
made by saturating the surface with methanol injections at room
A mixed-phase catalyst of iron molybdate with an excess
−
1
of molybdenite (25.9% by weight) was prepared by acidifica-
tion of a solution of ammonium heptamolybdate (BDH, ꢀ99%)
to pH 2 using nitric acid (Fisher, laboratory grade), before the
dropwise addition of iron nitrate (BDH, ꢀ98%) with stirring
temperature in a flow of 30 mL min He before the catalyst
◦
−1
.
temperature was ramped at a rate of 12 C min
In the TPD data displayed here, the contributions to the
30-amu formaldehyde signal by methanol was corrected for
methanol cracking and reaction on the mass spectrometer fil-
ament. Similarly, the 44-amu signal for carbon dioxide was
corrected for the combustion of methanol and formaldehyde on
the filament, and the 28-amu signal of carbon monoxide was
corrected for methanol and formaldehyde cracking and reac-
tion, as well as CO2 cracking. The levels of these contributions
were estimated by a series of calibrations in which pure gas, or
vaporized liquid in the case of formaldehyde, was passed over
the catalyst bypass. A similar procedure was followed for de-
termining the conversion–selectivity plots.
◦
at 60 C. This led to the formation of a canary yellow pre-
◦
cipitate, which was then heated to near dryness at 90 C. The
resulting solid was dried at 120 C overnight before being cal-
cined in air at 500 C for 48 h. As shown by XRD and reported
in more detail below, the catalyst comprised two phases, ferric
molybdate and molybdenite, as could be expected from the sto-
ichiometry of the catalyst, which was Mo:Fe 2.2:1, compared
with ferric molybdate at 1.5:1.
The pulsed-flow microreactor allows for the examination of
industrially important catalytic reactions on a small scale and
can provide kinetic and mechanistic data, and has been de-
scribed in detail elsewhere [6,7]. The reactor consists of a stain-
less steel U-tube (6 mm o.d., 4 mm i.d.) mounted vertically in a
GC oven that can be held at constant temperature or ramped
between two temperatures at a constant rate before product
monitoring on a mass spectrometer. First, 0.5 g of catalyst is
loaded into the tube. Several gases may be passed over the cat-
alyst bed; for example, two gases, a diluent gas (helium) and
a dosing gas, can be flowed continuously, and a third gas can be
introduced by means of a pulsing valve. Further gases or liquids
◦
◦
The oxygen removed from the catalyst was calculated from
the total yield of all products, making use of the number of oxy-
gen atoms removed for each product molecule as follows: H2, 0;
H2O, 1; CO, 0; H2CO, 0; CO2, 1; and CH3OCH3, −1. Oxygen
removal was calculated by multiplying the amount of a prod-
uct formed by the amount of oxygen required to form it and
totalling these values. The resulting value was converted into a
percentage of the total oxygen by assuming that all of the iron in
the catalyst formed Fe2(MoO4)3, whereas the excess molybde-
num formed MoO3. The monolayers of oxygen removed were
19
−2
(
in this case methanol) can be injected through a septum assem-
calculated by assuming each monolayer contained 10
m of
bly, and continuous flow conditions can be introduced here by
means of a syringe pump.
oxygen atoms, which is a typical value for surface adatom con-
centrations.
All gases were supplied by BOC Ltd with a purity of
99.5% and were passed through Puritubes (Phase Separation
Ltd.) filled with 5 Å molecular sieves to remove carbon dioxide
and water. Gas flows were controlled by mass flow controllers
The XRD spectra were obtained using a Enraf Nonus
FR590 diffractometer fitted with a hemispherical analyzer, us-
ing CuKα radiation (λ = 1.540598 Å), with a voltage of 40 kV
and a current of 30 mA. The XPS spectra were obtained us-
ing an ESCALAB 220 spectrometer equipped with AlKα and
MgKα sources and fitted with a fast entry lock for easy sample
loading. For this study, AlKα (1486.6 eV) irradiation was used,
so that the Fe Auger peaks would not overlap with the Fe 2p1/2
and Fe 2p3/2 peaks.
ꢀ
(
Brookes 5850TR series), allowing flow rates of the gases to
−
1
be controlled within 0.1 mL min , and were calibrated us-
ing a bubble flow meter. After flowing over the catalyst bed,
the reactants and products flowed down a heated capillary line.
The flow of gas to the capillary line was controlled by a nee-
dle valve, with most of the gas then vented by a rotary pump,
allowing a small fraction to be bled into the UHV chamber
containing the mass spectrometer (Hiden Analytical quadru-
pole Hal 201), which was computer-controlled and displayed
results in real time. To account for sensitivity drift within the
mass spectrometer, pulses of methanol were passed through the
bypass before each run. Reagent conversions and product se-
lectivities were calculated from the product distributions. To
accurately measure the temperature, a thermocouple was in-
serted into the catalyst bed.
The surface area of the catalyst was measured using a
Micromeritics Gemini 2360 instrument and was found to be
2
−1
.
6.7 (± 1) m g
3. Results and discussion
3.1. Selective oxidative dehydrogenation of methanol
Fig. 1 shows a temperature-programmed pulsed-flow reac-
tion profile for the iron molybdate catalyst in 10% O2/He;
a wider temperature range is shown in supplementary data
Fig. S1. Such raw data were obtained for other experiments in
this paper but usually are not given for the sake of brevity and
clarity. Clearly, formaldehyde selectivity was always high; for
Catalysts were subjected to pulsed reduction at various tem-
peratures while the activity and selectivity were monitored. To
establish the aerobic ability of the catalyst, 1-µL injections of
liquid methanol were passed over the catalyst every 2 min, in
−
1
◦
a flow of 30 mL min 10% O2/He, while the temperature was
instance, as shown in Fig. 1, after 48 min at 250 C, methanol
◦
◦
−1
ramped to ∼400 C at a rate of 8 C min . For the reduction
conversion was close to 100%; note the near-zero 31-amu sig-
nal, compared with the high (30-amu) formaldehyde signal.
experiments, the catalysts were held isothermally with 1-µL liq-
−
1
◦
uid methanol injections every 2 min in a flow of 30 mL min
The reaction started below 150 C, with methanol conversion