Time-resolved XAS/MS/Raman monitoring of mutual copper self-reduction and ethanol dehydrogenation reactions

Article abstract of DOI:10.1039/c5ra27403f

Selective ethanol dehydrogenation using a Cu/Al₂O₃ catalyst was investigated by time-resolved XAS/MS/Raman techniques. On-line monitoring of the reaction products revealed that formation of H₂ and acetaldehyde occurs over intermediate Cu⁺ species self-reduced upon reaction and the selectivity to ethyl acetate results in coupling of acetyl and ethoxy species over mixed Cu°/Cu⁺ active sites.

Full text of DOI:10.1039/c5ra27403f

RSC Advances
View Article Online
View Journal | View Issue
COMMUNICATION
Time-resolved XAS/MS/Raman monitoring of
mutual copper self-reduction and ethanol
Cite this: RSC Adv., 2016, 6, 20453
dehydrogenation reactions†
a
Wellington H. Cassinelli, Leandro Martins,a Marina Magnani,a Sandra H. Pulcinelli,a
Received 22nd December 2015
Accepted 10th February 2016
*
Valerie Brioisb and Celso V. Santillia
´
DOI: 10.1039/c5ra27403f
www.rsc.org/advances
Selective ethanol dehydrogenation using a Cu/Al2O3 catalyst was information about the oxidation state of the active Cu species in
investigated by time-resolved XAS/MS/Raman techniques. On-line the EDR. Some studies have suggested that metallic copper is
monitoring of the reaction products revealed that formation of H2 the active species,7,8 while others have found that the presence
and acetaldehyde occurs over intermediate Cu+ species self-reduced of Cu+ species in association with Cuꢀ can optimize the activity
upon reaction and the selectivity to ethyl acetate results in coupling of and selectivity.5,9,10
acetyl and ethoxy species over mixed Cuꢀ/Cu+ active sites.
In the present work, it is shown that the activation process
can be integrated with the reductive atmosphere provided by
the EDR stream, which is important from the economic
perspective. Clear correlation between the oxidation state of the
active copper, the catalytic performance, and the selectivity of
the catalyst was revealed from monitoring of the copper species
by time-resolved XAS combined with determination of the EDR
products using on-stream mass spectrometry and Raman
spectroscopy measurements.
The copper-based catalyst was composed of well-dispersed
20 wt% of copper (20Cu/Al) supported on macro–mesoporous
alumina.10 The catalyst was synthesized by wetness impregna-
tion of calcined porous Al2O3 (SBET ¼ 577 m2 gꢁ1) using an
aqueous copper nitrate solution. Aer impregnation, the
sample was calcined in a conventional muffle furnace at 500 ꢀC
for 2 h. Transmission electron microscopy (TEM) characteriza-
tion of the as-prepared catalyst was carried out at the Brazilian
Nanotechnology National Laboratory (LNNano), using a JEOL
In recent years, the need to reduce dependency on fossil fuels
and decrease emissions of greenhouse gases (GHG) has moti-
vated the use of clean and renewable energy sources inducing
large increases in the production of biofuels such as biodiesel,
biogas, and bioethanol. The advances in research concerning
rst and second generation ethanol technologies have led to
a signicant increase in global ethanol production, reaching 86
billion liters in 2011.1 The use of ethanol in motor vehicles,
alone or in mixtures with gasoline, can decrease net GHG
emissions by 90%, compared to an equivalent amount of
gasoline.2 This renewable energy source can also be used in the
so-called ethanol dehydrogenation reaction (EDR) for the green
production of chemicals such as H2, acetaldehyde, ethyl acetate,
and n-butanol, amongst others. Since these compounds are
normally obtained from fossil fuel sources, their production
from bioethanol, for instance produced from sugarcane, offers
a way of overcoming the shortage of fossil fuel sources as well as
reducing emissions of GHG.
˚
JEM 3010 microscope operated at 300 kV (1.7 A resolution). The
sample was prepared by dropping an isopropanolic suspension
of the catalyst onto amorphous carbon lms supported on
nickel grids.
Supported copper is the most widely used catalyst for alcohol
dehydrogenation, and its selectivity towards different products
changes according to catalytic features such as the acid–basic
characteristics of the support,3,4 Cu loading, metal dispersion,
and residence time.5,6 However, there is a lack of conclusive
Fig. 1(a) shows a TEM image of the calcined 20Cu/Al sample,
revealing that the copper species were well dispersed on the
Al2O3 surface. The pore size distributions determined by means
of N2 adsorption–desorption isotherms (Fig. S1, ESI†) showed
that the porous Al2O3 used as support had a mesopore size
distribution centered at 7.8 nm that could limit the size of the
unreduced copper, which was mainly present as CuO particles
in the as-synthesized samples. The bright-eld and dark-eld
images (Fig. 1(a) and (b)) revealed an average particle size of
2.4 ꢂ 1.1 nm, which was smaller than the average mesopore size
of the Al2O3 support. The electron diffraction pattern measured
a
´
Instituto de Quımica, Universidade Estadual Paulista – UNESP, Rua Professor
Francisco Degni, 55, 14800-060, Araraquara, SP, Brazil. E-mail: cassinelli@iq.
unesp.br; Fax: +55-16-33222308; Tel: +55-16-33019758
bSynchrotron SOLEIL, L'Orme des Merisiers, BP48, Saint Aubin, 91192 Gif-sur Yvette,
France
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c5ra27403f
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 20453–20457 | 20453

View Article Online
RSC Advances
Communication
from the standpoint of catalytic performance, because a greater
quantity of well-dispersed copper species increases the number
of available active sites.
In order to demonstrate the feasibility of self-activation of
the as-prepared oxidic copper catalyst, time-resolved operando
monitoring of the effluents using XAS, mass spectrometry, and
Raman analysis was performed during the EDR, using a dedi-
cated cell connected to the gas feeding system installed at the
SAMBA beamline of the SOLEIL synchrotron.12 Time-resolved
spectra were collected in situ using a Quick-XAS mono-
chromator equipped with a Si(111) channel-cut crystal whose
oscillatory movement was set to acquire one spectrum every 0.5
s.13 The effluent gas at the reactor outlet was monitored on-line
using mass spectrometry (Cirrus system, MKS) and Raman
spectroscopy (RXN1 system equipped with a sensitive Air-
Head™ probe, Kaiser Optical Systems, Inc.). Table 1 (ESI†) gives
the mass fragments used for product monitoring, together with
the Raman band positions of gaseous products of interest.
Evolution of Raman bands regarding ethanol consumption
(2865 cmꢁ1) and H2 formation (585 cmꢁ1) are also illustrated in
the Fig. S2.† The reaction was performed using isothermal steps
ꢀ
ꢀ
of 50 C in the range from 200 to 400 C, with 45 min at each
temperature, using 15.2 mg of 20 wt% copper catalyst diluted in
boron nitride (mass dilution of about 45%). The heating rate for
changing the isothermal reaction temperature was set to 10 ꢀC
minꢁ1. Ethanol was introduced to the catalytic reactor by
passing a ow of helium (at 40 mL minꢁ1) through a saturator
containing liquid ethanol immersed in
a temperature-
controlled bath at 43 ꢀC. The XANES data analysis was per-
formed using Athena14 soware and copper speciation was ob-
tained by linear combination ts (LCF). The XANES spectra of
the pure copper reference species (in the Cu2+, Cu+, and Cuꢀ
oxidation states) employed in the LCF were processed using
a multivariate curve resolution with alternating least squares
procedure (MCR-ALS). The chemometric method was described
fully in a previous paper that focused on the copper speciation
resulting from temperature programmed reduction (TPR) of the
same Cu-based catalysts.15
Fig. 2(a) shows the Cu K-edge normalized XANES and Fourier
transforms (FT) of the EXAFS spectra (not corrected for phase
shi) obtained for the calcined 20Cu/Al catalyst during the EDR.
Each spectrum in Fig. 2(a) was collected aer 45 min of reaction
at a given isothermal temperature. For comparison, the data for
Fig. 1 TEM images of the 20Cu/Al catalyst. (a) Bright-field and particle
size distribution, and (b) dark-field and electron diffraction.
in the same region shown in Fig. 1(a) and (b) indicated a low
degree of crystallization, reecting a distribution of Cu species
with no long-range order on the alumina support. Despite the
high loading of copper used for catalyst preparation, the
absence of large CuO agglomerates in the TEM images clearly
evidences the capacity of the porous Al2O3 with high surface
area to disperse a high amount of oxidic Cu2+ species, with
monolayer or sub-monolayer coverage. Previous work11 found
that surface saturation of Al2O3 by copper was achieved with
approximately 4–5 wt% of Cu per 100 m2 gꢁ1 of g-Al2O3. Here,
the high surface area of the support (SBET ¼ 577 m2 gꢁ1) was
responsible for the formation of an incomplete monolayer of
copper on the surface of the alumina, even with such a high Cu
loading (20 wt%). This feature is a fundamental prerequisite
ꢀ
the sample collected before the reaction at 200 C are also dis-
played. The XANES spectrum presented an intense white line
characteristic of divalent oxidic Cu2+ species. It can be seen
from Fig. 2(a) that the intensity of the white line did not change
ꢀ
during the rst isothermal step at 200 C. However, there were
clear decreases of the white line during the second and third
isothermal temperature steps at 250 and 300 ꢀC, respectively
(Fig. 2(b)). At the end of these isothermal steps, there was
a decrease in the intensity of the rst FT peak located between
˚
1.0 and 2.0 A, corresponding to the Cu–O contribution. There
was also an increase in the intensity of the peak located between
ꢀ
˚
2.2 and 2.7 A, corresponding to the Cu–Cu contribution of Cu
species. There was a progressive decrease of the white line and
changes in the intensity of the FT peaks during the 45 min of
20454 | RSC Adv., 2016, 6, 20453–20457
This journal is © The Royal Society of Chemistry 2016

View Article Online
Communication
RSC Advances
Fig. 2 (a) Cu K-edge XANES and Fourier transformed EXAFS spectra of
the 20Cu/Al catalyst collected during the ethanol dehydrogenation
reaction from 200 to 400 ꢀC. (b) Quick-XANES spectra recorded
during the ethanol dehydrogenation reaction from 200 to 300 ꢀC.
Fig. 3 (a) Copper speciation during the ethanol dehydrogenation
reaction obtained from time-resolved Cu K-edge XANES measure-
ments; (b) and (c) distribution of products with ethanol on-stream
obtained using mass spectrometry and Raman spectroscopy,
respectively.
isothermal treatment at 250 ꢀC with ethanol on-stream. During
the third isothermal step at 300 ꢀC, copper was almost
ꢀ
completely reduced; full reduction was achieved at 350 C, as
evidenced by the XANES ngerprints shown in Fig. 2(a). The
decrease in intensity of the Cu–Cu contribution to the FT during
heating from 300 to 400 ꢀC was mainly related to the increase of
Debye–Waller factors, due to the increase of the thermal
vibrations.
Fig. 3(a)–(c) show the Cu speciation obtained by LCF analysis
of the Cu K-edge XANES and reaction effluent monitoring data
acquired during the ethanol dehydrogenation. The speciation
indicated that at the time that mainly Cu2+ species were present
in the reaction at 200 ꢀC, there was no substantial formation of
products with ethanol on-stream. The increase in the temper-
ature from 200 to 250 ꢀC promoted self-activation of the copper.
At this stage, there was an abrupt consumption of ethanol
(Fig. 3(c)), and the formation of H2 and acetaldehyde (Fig. 3(b)
and (c)) was identied when Cu+ species started to be formed by
partial reduction of Cu2+ (Cu2+ / Cu+). Similar abrupt
consumption of ethanol was not observed for reduced 20Cu/Al
catalyst in the ethanol dehydrogenation reaction (Fig. S3†).
Aer reaching a maximum value of about 60%, a decrease in the
amount of Cu+, together with a concomitant increase of the Cuꢀ
species. As there was no substantial change in the Cu2+
concentration during this stage, the formation of Cuꢀ was
related to a second reduction process involving monovalent
copper (Cu+ / Cuꢀ). This second reduction stage was accom-
panied by a two-fold decrease in the conversion of ethanol and
production of acetaldehyde. The selectivity towards H2
continued to increase, despite a slight slowdown of the forma-
tion rate, while the ethyl acetate concentration started to
increase due to the dehydrogenation of acetaldehyde over the
mixture of Cuꢀ/Cu+ active sites. It is noteworthy that there was
consumption of H2 during the second reduction stage, so the
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 20453–20457 | 20455

View Article Online
RSC Advances
Communication
hydrogen concentration observed did not reect the full quan- on the porous alumina was proposed to be related to the iso-
tity of H2 formed during this stage. The increase of the reactor lated Cu ions bound to the vacant Al3+ sites, as well as to the
temperature to 300 C promoted a new abrupt increase in the large diffusion distances between Cu+ ions adsorbed in
ꢀ
Cu+ concentration, resulting from the fast reduction of Cu2+
.
a diluted sub-monolayer, which increased as the Cu loading
This speciation change was accompanied by an additional decreased.
increase in acetaldehyde production. Aer total reduction of the
The development of Cu+ / Cuꢀ reduction from 250 ꢀC
Cu species with further increase of temperature (T > 300 ꢀC), the resulted in an increase in the concentration of Cuꢀ. During this
increment in ethanol conversion achieved at each temperature stage, the ethyl acetate formation rate started to increase due to
was smaller than the values observed at lower temperatures. interaction between the adsorbed ethoxy and acetyl species,
Above 350 ꢀC, the selectivity towards ethyl acetate decreased, leading to a decrease in the acetaldehyde formation rate
and there were increases in the concentrations of products such (Fig. S4†). The increase in the ethyl acetate yield was in agree-
as ethene, diethyl ether, and water, resulting from ethanol ment with previous ndings,9 where the presence of Cuꢀ
dehydration reactions (Fig. 3(b)).
together with isolated partially oxidized Cu+ was able to activate
It has been proposed that Cuꢀ is able to activate the O–H the C–H bond in the adsorbed CH3CH2O–Cu+ and promote the
bond of ethanol adsorbed on catalytic surfaces.3,16 However, formation of ethyl acetate by coupling with the acetyl (CH3CO)
recent studies have shown that an optimal Cuꢀ/Cu+ ratio can fragment.18 The present results (Fig. 3) showed that there was
also promote faster ethanol dehydrogenation on Cu catalysts no further increase in the concentration of ethyl acetate aer
supported on Al2O3 and ZrO2.5,10,17 In the present case, the the full reduction of copper species achieved by increasing the
operando quick-XAS/MS/Raman results revealed no signicant reaction temperature to 350 ꢀC. This conrms the key role of the
ethanol activation in the presence of supported Cu2+ species at Cu+/Cuꢀ balance in the production of ethyl acetate from ethanol
200 ꢀC. The onset of acetaldehyde formation was evident at dehydrogenation. Similar results were reported by Inui et al.,19
higher temperatures, when the concentration of Cu+ increased with the coupling of ethanol with acetaldehyde for ethyl acetate
considerably as a result of the reduction of Cu2+ species, rst formation occurring over a mixed metal-oxides surface (Cu+ and
aer heating from 200 to 250 ꢀC, and then from 250 to 300 ꢀC. Cuꢀ), but not over a Cuꢀ surface.
The maximum rate of acetaldehyde formation, as shown by the
In summary, we have demonstrated a correlation between
derivative curve (Fig. S4†), occurred without any substantial catalytic performance in the ethanol dehydrogenation reaction
presence of reduced Cuꢀ species, while the maximum rate of H2 and the most active oxidation state of copper species supported
production was observed when the quantity of intermediate Cu+ on porous alumina. The combination of time-resolved XAS/MS/
species reached a maximum. These results, together with the Raman techniques revealed an abrupt consumption of ethanol
decrease in the acetaldehyde formation rate with further and a pronounced increase of formation of H2 and acetaldehyde
reduction of Cu+ to Cuꢀ at the end of the isothermal treatment as Cu+ started to be formed (Cu2+ / Cu+) and reached its
at 250 ꢀC caused by the increase in the ethyl acetate formation maximum concentration. During the development of the
rate, provided strong evidence that Cu+ was more active than second reduction stage (Cu+ / Cuꢀ), the consumption of
Cuꢀ in promoting the rst ethanol dehydrogenation towards ethanol, and the formation of H2 and acetaldehyde, were not
acetaldehyde and H2. During the copper activation, it was maintained at elevated rates, because ethyl acetate production
possible to identify the most active species for acetaldehyde started. From these results, it can be concluded that Cu+ is the
formation, because formation of the Cu+ intermediate was most active species for promotion of the rst ethanol dehy-
associated with the self-production of H2 caused by the dehy- drogenation. On the other hand, the combined presence of Cu+/
drogenation of ethanol to acetaldehyde. The amount of H2 Cuꢀ species results in an increase in the formation of ethyl
formed in this process was at a concentration suitable for acetate by the coupling of acetyl (CH3CO) and ethoxy
sustaining the Cu+ intermediate throughout the treatment of (CH3CH2O) fragments. It should be noted that the methodology
the 20 wt% Cu catalyst at 250 ꢀC. In previous work investigating used in this work, based on the combination of time-resolved
the activation of Cu under an H2/He atmosphere (heating from XAS/MS/Raman techniques, is not only suitable for studying
50 to 250 ꢀC, with isothermal treatment at 250 ꢀC for 0.5 h), the ethanol dehydrogenation process, but also offers a valuable
using Cu-based catalysts prepared from the same porous means of understanding active species in many other indus-
alumina support, it was found that the lifetime of the Cu+ trially relevant processes involving the use of supported
species at 250 ꢀC was shorter than observed here.10 Although catalysts.
a similar maximum value of about 60% of intermediate Cu+
species was achieved when the temperature reached 250 ꢀC, the
second reduction stage leading to Cuꢀ occurred during the same
period. In our recent work,10 it was found that for Cu loadings in
Acknowledgements
the range from 5 to 20 wt% in catalysts prepared on the Al2O3 This work was supported by the Brazilian agencies: CAPES,
support, the intermediate Cu+ species formed by classical TPR CNPq, and FAPESP (project numbers 07/53073-4, 2010/01449-3,
treatment became stabilized as the loading was decreased. In 2013/50023-7, and 2013/05346-2) and by a FAPESP/CNRS bilat-
this case, the onset temperature of the Cu2+ / Cu+ reduction eral cooperation program. The authors are grateful to LNNano
was signicantly lower than the temperature leading to the (CNPEN) for the TEM images, and to SOLEIL Synchrotron for
formation of Cuꢀ species. The stability of the Cu+ intermediate providing beam time at the SAMBA beamline.
20456 | RSC Adv., 2016, 6, 20453–20457
This journal is © The Royal Society of Chemistry 2016

View Article Online
Communication
RSC Advances
11 R. M. Friedman, J. J. Freeman and F. W. Lytle, J. Catal., 1978,
55, 10–28.
Notes and references
12 C. La Fontaine, L. Barthe, A. Rochet and V. Briois, Catal.
Today, 2013, 205, 148–158.
1 S. Kumar, P. Shrestha and P. Abdul Salam, Renewable
Sustainable Energy Rev., 2013, 26, 822–836.
13 E. Fonda, A. Rochet, M. Ribbens, L. Barthe, S. Belin and
V. Briois, J. Synchrotron Radiat., 2012, 19, 417–424.
14 B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12,
537–541.
15 W. H. Cassinelli, L. Martins, A. R. Passos, S. H. Pulcinelli,
C. V. Santilli, A. Rochet and V. Briois, Catal. Today, 2014,
229, 114–122.
2 G. R. Timilsina and A. Shrestha, Energy, 2011, 36, 2055–2069.
3 P. C. Zonetti, J. Celnik, S. Letichevsky, A. B. Gaspar and
L. G. Appel, J. Mol. Catal. A: Chem., 2011, 334, 29–34.
4 J. M. Vohs, Chem. Rev., 2013, 113, 4136–4163.
5 I. C. Freitas, S. Damyanova, D. C. Oliveira, C. M. P. Marques
and J. M. C. Bueno, J. Mol. Catal. A: Chem., 2014, 381, 26–37.
6 K. Inui, T. Kurabayashi and S. Sato, J. Catal., 2002, 212, 207–
215.
7 N. Iwasa and N. Takezawa, Bull. Chem. Soc. Jpn., 1991, 64,
2619–2623.
8 S. W. Colley, J. Tabatabaei, K. C. Waugh and M. A. Wood, J.
Catal., 2005, 236, 21–33.
9 A. G. Sato, D. P. Volanti, I. C. de Freitas, E. Longo and
J. M. C. Bueno, Catal. Commun., 2012, 26, 122–126.
10 W. H. Cassinelli, L. Martins, A. R. Passos, S. H. Pulcinelli,
A. Rochet, V. Briois and C. V. Santilli, ChemCatChem, 2015,
7, 1668–1677.
16 D. Gao, Y. Feng, H. Yin, A. Wang and T. Jiang, Chem. Eng. J.,
2013, 233, 349–359.
17 A. G. Sato, D. P. Volanti, D. M. Meira, S. Damyanova,
E. Longo and J. M. C. Bueno, J. Catal., 2013, 307, 1–17.
18 J. C. Kenvin and M. G. White, J. Catal., 1992, 135, 81–91.
19 K. Inui, T. Kurabayashi, S. Sato and N. Ichikawa, J. Mol.
Catal. A: Chem., 2004, 216, 147–156.
This journal is © The Royal Society of Chemistry 2016
RSC Adv., 2016, 6, 20453–20457 | 20457