J. Quesada et al.
Applied Catalysis A, General 559 (2018) 167–174
-TPD, being the main results summarized in Table 1. No
−
1
the reactor at 20 mL min
(STP). These conditions were chosen ac-
TPD and NH
3
cording to the optimization reported in our previous work and they
significant changes in surface area were observed, with only the ex-
pected slight decrease after the metal deposition on the parent mixed
oxide. In good agreement, pore volume and diameter also slightly de-
crease. Metals mainly are deposited on acid sites [29,30], being the
strongest ones the most affected by the metal deposition. A very similar
behavior, disappearing more than 85% of the initially present was
observed in both cases. As to the basicity, the decrease respect to the
Mg-Al is less marked, being only relevant in the case of the strongest
sites. This phenomenon is more evident in the case of Ni material,
catalyst that only keeps the weak basic sites.
XRD analyses (Fig. S1) corroborate that there are not significant
changes in the general structure of the bulk Mg-Al oxides, the same
peaks for all the metal-modified catalysts being observed. Periclase is
the main phase in all the cases, with similar diffraction patterns for all
the catalysts. No signals related to the added metal species were de-
tected, as expected considering the low metal content. HRTEM analyses
were carried out in order to determine the metal particle size and dis-
persion. Representative histograms of crystallite sizes distributions are
depicted in Fig. 1 (corresponding micrographs are included in the
Supplementary information, Fig. S2); whereas crystallite sizes and
metal dispersion data are summarized in Table 1. A very high disper-
sion is observed (> 75% in both the cases), with nanoparticles around
−1
correspond to a weight hourly space velocity (WHSV) of 7.9 h
[20].
The outlet gases were on-line analysed with a HP6890 Plus gas chro-
matograph with a flame ionization detector (GC-FID), using a TRB-5MS
capillary column. Additional GC-FID analyses were off-line performed
combining two columns (HP-Plot Q and HPPlot MoleSieve 5 A) in order
to distinguish and quantify ethylene and methane. Products identifi-
cation was performed using commercial standards and supported by
GC-MS (Shimadzu QP-2010) by the same methodology in the GC-FID.
Operation conditions were selected in order to ensure that the reported
experiments are performed under kinetic regime, being mass transfer
effect negligible.
Conversions (x) were calculated from the ethanol concentrations at
the reactor inlet and outlet. Carbon balances were calculated by con-
trasting the total quantity of carbon atoms at the reactor inlet and
outlet, taking into account only the identified products (compounds in
Scheme 1). Yield was calculated by the following equation:
moles of ethanol fed converted to the product i
⎛
⎝
⎞
⎠
ηi(%) =
·100
moles of ethanol fed
(1)
The productivity of the different compounds (P
i
) during the reaction
(
average formation rate) were determined as follows:
1.3 nm large. These values are in good agreement with those reported in
F·x·φi
the literature for similar catalysts prepared by this procedure [31].
The low crystallite sizes observed by HRTEM suggest a strong in-
teraction between metal and support. This hypothesis is congruent with
the high reduction temperatures observed in the TPR results (Fig. 1c).
According to the literature, the reduction of Ni nanoparticles takes
place at 640 K [32], whereas in our sample metal is mainly reduced at
−
(mmol·s 1·g ) =
−1
P
i
cat
W
(2)
F ≡ ethanol molar flow fed to the reactor (mmol s−1)
W ≡ catalyst mass (g)
i
ϕ ≡ Selectivity for product i (moles of ethanol fed converted to a
product i/ moles of converted ethanol).
3 4
698 K, with only a minor shoulder at 640 K. The reaction from Co O to
Diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy
experiments were performed using a Thermo Nicolet Nexus FTIR
equipped with a Smart Collector Accessory and a MCT/A detector. The
material (20 mg) was placed inside the catalytic chamber where the
temperature was controlled. The sample was pre-treated at 473 K for
Co is reported at T ≤ 623 K when is supported over alumina, and peaks
around 773 K or above are associated with the reduction of CoAl mixed
oxides [33]. Observing the TPR results obtained with the Co/MgAl
material, two peaks are highlighted at 565 and 750 K, which are related
to the former and the latter reductions, respectively.
−
1
1
h in He flow. Spectra were acquired in the 4000-650 cm
wave-
number range, after subtraction of the KBr standard background.
Spectra were recorded at same temperatures as in the reactor allowing
the comparison between both results, and working under inert (He) or
3.2. Reaction results under reducing atmosphere
2
reducing conditions (10 vol.% H /He), as needed. Signals were trans-
The role of Co and Ni was studied by introducing hydrogen in the
formed to Kubelka-Munk units to obtain semiquantitative results. This
method allows quantitatively analyze the amount of adsorbed species
on the surface, being the signal (for a same support and comparing
analogous conditions) proportional to the concentration of adsorbed
species [28].
2
helium stream (10 vol.% of H ). In a previous blank experiment
(
without catalyst), ethanol conversions were negligible at the tem-
perature range considered in this article, so reported data are directly
related to the catalytic activity. The reducing conditions are expected to
promote the hydrogenation steps of the reaction mechanism, enhancing
the 1-butanol yield, and to hinder oligomerization that could strongly
affect to the catalytic stability.
3. Results and discussion
The evolution of the conversion, carbon mass balance and se-
lectivity to the main compounds with the reaction temperature is
showed in Fig. 2 for the different tested materials (Mg-Al, Co/Mg-Al
and Ni/Mg-Al). Reactions with Mg-Al are also considered in order to
analyze any change in the basis mechanism because of the presence of
3
.1. Characterization of fresh catalysts
The morphological properties and surface chemistry of parent and
2 2
metal-modified materials have been analyzed by N physisorption, CO -
Table 1
Main results of the fresh catalysts characterization: morphological properties, density and distribution of the acid and basic sites, and HRTEM results. *Results taken
from a previous work [20].
1
1
Catalyst Morphological properties
Acid sites (μmol g ), [T (K)]
Basic sites (μmol g ), [T (K)]
HRTEM
S (m2 g1)
Mg-Al* 226
D
(Å)
V
p
(cm3 g ) weak
1
medium strong
weak
medium
strong
Metal dispersion (%) Crystallite diameter (nm)
p
135
59
0.7
0.4
0.4
11.3
12.5 41.8
49.7
71.7
[400]
10.6
238.6
[630, 670, 800]
–
–
–
[
345, 370] [450]
3.9 3.3
329, 365] [419]
2.9 3.1
325, 358] [408]
[630, 800] [340]
4.0
[517]
5.4
[508]
Ni
182
207
55.8
78.6
78.7
1.3
1.3
[
[328, 377] [430]
51.0
[355]
Co
60
70.2
53.8
[
[415, 491] [607, 738]
169