GModel
CATTOD-8304; No. of Pages8
ARTICLE IN PRESS
2
D. Durán-Martín et al. / Catalysis Today xxx (2013) xxx–xxx
Table 1
ZrO(NO ) ·xH O (Sigma–Aldrich, 99.9% metal) were dissolved in
3
2
2
Chemical analysis of CuZr precursors and CuO and Cu crystallites sizes for calcined
and reduced samples respectively calculated by applying the Scherrer equation.
deionized water (0.3 M total metal content) to achieve nominal
Cu/Zr atomic ratios of 0.2, 0.4, 1, 2.5, and 6. The mixture was
stirred at room temperature and constant pH = 7 by adding an
Sample
Cu/Zr atomic ratio by TXRF
Crystallite size (nm)
aqueous Na CO solution (0.5 M). Once the precipitation was over,
2
3
CuO
Cu
the precipitate was aged in the mother liquor under these condi-
tions for 4 h, and the solid was subsequently recovered by filtration
ZrO2
–
n.d.
n.d.
n.d.
7.1
8.0
9.4
n.d.
n.d.
4.5
23.8
25.8
29.8
48.9
0.2CuZr
0.2
0.5
1.1
2.6
5.0
–
and washed with deionized H O. The catalytic precursors are
0.4CuZr
1CuZr
2
named here as xCuZr-p, where x denotes the Cu/Zr atomic ratio.
2
6
.5CuZr
CuZr
Monometallic samples (denoted as Cu-p and ZrO -p) were also
2
prepared following identical protocols.
Cu
11.7
The precursors were first treated in flowing synthetic air at 673 K
−
heating rate of 10 K min ) for 1 h and then in H /Ar (5 vol.% H ) at
2 2
1
n.d.: not determined.
(
−1
5
73 K (heating rate of 5 K min ) for 1 h. The conditions of the treat-
ment protocols were selected based on characterization results
shown later). The obtained bimetallic solids are denoted as xCuZr,
while monometallic samples are named as Cu and ZrO2.
with Al-K␣ (hꢂ = 1486.6 eV) and Mg-K␣ (hꢂ = 1253.6 eV) X-ray
sources. Binding energies were calibrated relative to the C 1s peak
from adventitious carbon of the samples at 284.6 eV to correct the
potential contact differences between the sample and the sample
holder of spectrometer. Both binding energy (BE) values and peak
areas were computed by fitting the experimental spectra to Gaus-
sian/Lorentzian lines after removing the S-shaped background.
Surface atomic ratios were calculated from the peak area ratios
normalized by atomic sensitivity factors.
Chemical analysis of liquid product after reaction was carried
out in Elan 6000 Perkin-Elmer Sciex inductively coupled plasma
mass spectrometry. Samples were digested in open glass by a
mixture of HNO3 and H2O2. Afterwards, samples were diluted in
1 v/v% HNO3. Total reflection X-ray fluorescence (TXRF) analysis
was employed to determine the composition of the solids before
and after reaction. Previously, samples were ground (particle size
<30 mm) and homogenized with high-purity water by ultrasonic
disaggregation to disperse the possible agglomeration of particles.
Two microliters of the suspension was placed on a flat carrier where
the water was evaporated off under a vacuum.
(
2.2. Glycerol hydrogenolysis measurements
Glycerol hydrogenolysis was performed in a stainless steel Parr
3
reactor (100 cm ). First, the activated catalyst (∼0.6 g) and an aque-
ous glycerol solution (25 g, 40 wt.% glycerol) were loaded into the
reactor, and the system was then purged three times consecutively
with N2 and H . Subsequently, the temperature was increased to
2
4
73 K and H2 pressure adjusted to 4.0 MPa. Finally, the mixture
was stirred at 500 rpm to start the reaction, which was conducted
for a period of 8 h. Reaction products were analyzed with a gas
chromatograph (GC) equipped with a capillary column (ZB-WAX,
3
0 m × 320 m × 0.5 m) connected to a flame ionization detector.
Before GC analysis, samples were prepared by adding acetone (10 g,
solvent) and ethyl valerate (0.2 g, internal standard) to the reaction
aliquot (0.4 g). Glycerol conversion is defined as the ratio between
moles of glycerol consumed in the reaction to the total moles of gly-
cerol initially present. Yield to 1,2-PDO is defined as the ratio
of moles of product formed to the total moles of glycerol ini-
tially present, being the selectivity defined as the result of
dividing the yield value by glycerol conversion calculated pre-
3. Results and discussion
3.1. Characterization of catalytic precursors
viously. H O content in the reaction products was determined by
2
Karl Fisher titration of an aliquot of reaction highly diluted with
pure methanol.
Table 1 shows in the first column the Cu/Zr atomic ratios for the
different precursors as measured by TXRF. In general, the experi-
mental Cu/Zr atomic ratio is in all cases quite close to nominal ones
(within a 15% deviation) that indicate that precipitation of both Cu
and Zn nitrates has been quantitative.
2.3. Characterization techniques
Powder X-ray diffraction (XRD) patterns were recorded in the
0–80 2ꢀ range in scan mode (0.02 , 1 s) using a X Pert Pro PANa-
Fig. 1 shows the XRD patterns recorded with the copreci-
pitated CuZr samples. The Cu-free and low Cu content solids
◦
◦
ꢀ
1
lytical diffractometer with Cu K␣1 (ꢁ = 0.154046 nm) radiations.
The crystallite size (D) was calculated by Scherrer equation, D = 0.90
(ZrO -p, 0.2CuZr-p, and 0.4CuZr-p) exhibit a broad diffraction peak
(2ꢀ = 20–40 ) attributed to low-crystalline zirconium hydroxycar-
2
◦
ꢁ
/ˇ cos ꢀ, where ꢀ is the diffraction angle and ˇ is the full width at
bonate (1, Zr(CO )(OH) ) [11,12], however its intensity is very
3
2
half-maximum (FWHM).
small, see multiplicity factor for these diffractograms. As the Cu
concentration increases (1CuZr-p, 2.5CuZr-p, 6CuZr-p, and Cu-p),
diffraction lines assigned to Cu (CO )(OH) (2, malachite; JCPDS
Evolved gas analysis by mass spectrometry (EGA-MS) was per-
formed by loading the samples (ca. 0.1 g) in a U-shaped quartz
reactor connected to a Balzer PrismaTM quadrupole mass spec-
trometer (QMS 200). The analysis was conducted while flowing
2
3
2
010-0399) are clearly observed, although the presence of CuCO3
(3, JCPDS 046-0858) cannot be discarded because these two phases
show similar XRD patterns. Additionally, Cu (NO )(OH) (4, ger-
3
−1
an O /Ar mixture (50 cm min , 20 vol.% O ) from room tempe-
2
2
2
3
3
−
1
rature to 1000 K at a heating rate of 10 K min . The fragments
hardtite; JCPDS 15-0014) is also detected in the Cu-p sample.
The thermal decomposition of these catalytic precursors (Cu and
Zr carbonates and hydroxycarbonates) in synthetic air has been
explored by evolved gas analysis with mass spectrometry (EGA-
MS) with the aim of determining the temperature required to form
the corresponding oxides. Fig. 2 depicts the evolution of fragments
with m/z = 18 and 44 (associated to H O and CO formation, respec-
m/z = 18 (H O ) and 44 (CO2+) were continuously monitored with
+
2
the mass spectrometer. Gas lines from the reactor outlet to the MS
inlet were heated to 393 K to avoid H O condensation phenomena.
2
Temperature-programmed reduction (TPR) experiments were also
performed with the U-shaped reactor connected to the mass spec-
trometer. A H /Ar mixture (5 vol. % H /Ar) was contacted with the
2
2
2
2
sample while heating from room temperature to 1000 K at a rate of
tively). The Cu-free and low Cu content solids (ZrO -p, 0.2CuZr-p,
2
−
1
5
K min
recorded with a VG Escalab 200 R Fisons spectrometer equipped
.
2
Cu content increases (Fig. 2a). Minor CO2 formation (Fig. 2b) is
Please cite this article in press as: D. Durán-Martín, et al., Stability and regeneration of Cu–ZrO2 catalysts used in glycerol hydrogenolysis to
1