Effect of the ZrO2 phase on the structure and behavior of supported Cu catalysts for ethanol conversion

Journal of Catalysis 307 (2013) 1–17

Contents lists available at SciVerse ScienceDirect

Journal of Catalysis

journal homepage: www.elsevier.com/locate/jcat

Effect of the ZrO phase on the structure and behavior of supported

2

Cu catalysts for ethanol conversion

a

b

a

c

b

a,

⇑

A.G. Sato , D.P. Volanti , D.M. Meira , S. Damyanova , E. Longo , J.M.C. Bueno

a

Laboratório de Catalise, Departamento de Engenharia Química, Universidade Federal de São Carlos, C.P. 676, 13565-905 São Carlos, SP, Brazil

Laboratório Interdisciplinar de Eletroquimica e Cerâmica, Departamento de Físico-Química, Instituto de Química Universidade Estadual Paulista,

4800-900 Araraquara, SP, Brazil

b

1

c

Institute of Catalysis, Bulgarian Academy of Sciences, 1113 Soﬁa, Bulgaria

a r t i c l e i n f o

a b s t r a c t

Article history:

The effect of amorphous (am-), monoclinic (m-), and tetragonal (t-) ZrO₂phase on the physicochemical

and catalytic properties of supported Cu catalysts for ethanol conversion was studied. The electronic

Received 23 April 2013

Revised 17 June 2013

Accepted 22 June 2013

parameters of Cu/ZrO

deﬁned by XPS and DRIFTS of CO-adsorbed. The results demonstrated that the kind of ZrO

a key role in the determination of structure and catalytic properties of Cu/ZrO catalysts predetermined

by the interface at Cu/ZrO . The electron transfer between support and Cu surface, caused by the oxygen

vacancies at m-ZrO and am-ZrO , is responsible for the active sites for acetaldehyde and ethyl acetate

formation. The highest selectivity to ethyl acetate for Cu/m-ZrO

2

were determined by in situ XAS, and the surface properties of Cu/ZrO

2

were

2

phase plays

2

Keywords:

Ethanol dehydrogenation

Cu/ZrO catalysts

2

Ethyl acetate

2

catalyst up to 513 K was caused by

the optimal ratio of Cu /Cu species and the high density of basic sites (O ) associated with the oxygen

0

+

2ꢀ

2

mobility from the bulk m-ZrO .

Characterization

1

. Introduction

The establishment targets for reducing greenhouse gas emis-

investigations have been focused on the effect of support material

on secondary product selectivity. In addition to simple dehydroge-

nation of ethanol, dehydrogenative coupling between the product

acetaldehyde and the ethanol could also occur to produce ethyl

acetate and hydrogen. Studies have shown that ethyl acetate can

be obtained directly from ethanol employing speciﬁc heteroge-

sions that contribute to the global warming should have a signiﬁ-

cant increase in the consumption of biofuels in coming years,

especially ethanol. This requires a substantial increase in ethanol

production and, in this sense, encourages research and develop-

ment of new raw materials for ethanol, such as lignocellulosic bio-

mass, and the construction of integrated bio-reﬁneries, a similar

concept to that of oil reﬁneries. Attention also turned to chemical

ethanol or ethanol chemistry, as the way to increase the chemical

production using a green chemistry route employing renewable re-

sources [1].

neous catalysts [3–8]. Iwasa and Takezawa [9] examined SiO

2

,

ZrO , Al , MgO, and ZnO as support for Cu-based catalysts and

2

2 3

O

found that the selectivity to acetaldehyde or ethyl acetate depends

strongly on the support kind. As the ethanol dehydrogenation is an

exothermic reaction, there is a formation of hot spots in the reactor

which leads to sintering the Cu particles and to decrease in the cat-

2

alyst activity [8]. ZrO has been described as preventive element

Acetaldehyde is a product of ethanol dehydrogenation, which

can be used as an intermediate in the production of acetic acid,

acetic anhydride, ethyl acetate, butyraldehyde, crotonaldehyde,

n-butanol, and many other chemicals. Ethyl acetate as one of the

main products of ethanol dehydrogenation is an environmentally

friendly organic solvent, which can be used for organic synthesis,

printing inks, adhesives, and cosmetics.

for sintering of Cu crystallites under the reaction conditions and

therefore was considered as a structural promoter and as an alter-

native support [10,11]. Recently, rice husk ash (RHA) composed of

amorphous silica has been used as a carrier for supported copper

catalysts prepared by ion-exchanged, which showed very high Cu

dispersion and stability in ethanol dehydrogenation to acetalde-

hyde at reaction temperature of 483–548 K [12].

Copper-based catalysts were found to be excellent catalysts for

ethanol dehydrogenation because of their ability to maintain the

C–C bond intact while dehydrogenating the CO–H bond [2]. Some

Zirconium oxide has received considerable attention as a cata-

lyst and support due to its unique properties; the preservation of

both acid and basic sites, on the one hand, and reducing and oxi-

dizing properties, on the other hand. As a result, the use of zirconia

rather than silica or alumina as a structural promoter or more fre-

quently as a support material has attracted considerable interest

for a number of catalytic applications in the recent years. In partic-

⇑

Corresponding author.

E-mail address: jmcb@ufscar.br (J.M.C. Bueno).

0

http://dx.doi.org/10.1016/j.jcat.2013.06.022

2

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

ular, it has been found that the activity of Cu/ZrO

strongly dependent on the phase structure of ZrO

context, it was reported [13] that Cu catalysts supported over m-

ZrO are more active in methanol synthesis than catalysts with

the same Cu surface density deposited on t-ZrO . The origin of

these differences was explained by the presence of higher concen-

tration of anionic defects on m-ZrO than on t-ZrO [14]. The differ-

ent spacing and symmetry of the Zr–O and –OH bonds at t-ZrO

and m-ZrO are considered to play key roles in determining the dis-

persion of the active metal component and the catalytic properties

of Cu/ZrO system [14].

The high copper–zirconia interface has been considered as a key

for obtaining highly active and selective copper catalysts with im-

proved stability [15]. The surface metallic copper area as well as

the kind of supported copper species is also expected to play a vital

role in the catalytic systems. A correlation between the surface

metallic Cu area and the catalyst activity was found [18]. New type

2

catalysts is

. Within this

2. Experimental

2

2.1. Catalyst preparation

2

Amorphous ZrO was prepared by sol–gel method described in

Ref. [24]. Zirconium n-propoxide (Zr(OCH CH CH ) (10.9 ml, 70%

2

3 4

2

in propanol) was dissolved in 233 ml anhydrous ethanol to obtain

a 0.1 M solution. Deionized water was added dropwise to the zirco-

nium n-propoxide solution to give water/alkoxide ratios of 32. A

cloudy solution was obtained on the initial addition of water,

which turned to a white suspension on complete addition. The sus-

pension was aged at room temperature for 24 h, followed by the

removal of the solvent by rotary evaporation. The obtained wet so-

lid was suspended in 350 ml deionized water with pH adjusted to

2

10 with ammonium hydroxide (NH OH, 30% in volume) and then

was reﬂuxed at 373 K in a 500 ml round-bottomed ﬂask. The pH

4

was maintained constant throughout the entire digestion that

was a period of 240 h. The solid was washed with deionized water.

The samples were dried at 373 K overnight and successively cal-

cined from room temperature to 673 K at a heating ramp of

10 K/min and a dwell time of 5 h under synthetic air ﬂow

(100 ml/min). Monoclinic and tetragonal ZrO₂commercial grades

(Saint-Gobain NorPro) were used as supports.

of nanostructured Cu/ZrO

nent dispersion has been obtained for methanol steam reforming

16].

There are several reports in the literature concerning the ethyl

2

materials featured with high compo-

[

acetate and acetaldehyde formation from ethanol on Cu-based cat-

alysts [9,17,18]. But these reports are not focused on the active cat-

alytic sites. To establish reliable reaction pathways models, the

ﬁrst step is the knowledge of the catalyst active sites, a point of vi-

tal importance to understand and predict the catalytic properties,

the most ambitious goal in the ﬁeld of catalysis.

ZrO -supported Cu catalysts have been used in many others

2

reactions such as methanol synthesis [14], steam reforming of

methanol [19,20], and water–gas shift reaction [21]. Based on

Cu/ZrO samples were prepared by incipient wetness impregna-

2

tion of the supports with a solution of (Cu(NO ) ꢁ2.5H O, 98%; Al-

3

2

drich) in methanol. The mixture was stirred in a round-bottomed

ﬂask under ultrasonic bath coupled with a rotary evaporation.

The solids were dried at 373 K overnight and calcined from room

temperature to 673 K at a heating ramp of 10 K/min and a dwell

time of 5 h under synthetic air ﬂow (100 ml/min). The theoretical

the XPS and XAS experiments, it was reported the presence of

amount of Cu was about 10 wt%. Cu supported on am-ZrO , m-

2

+

Cu species in the CuO pathway reduction in Cu/ZrO

2

catalysts

ZrO₂and t-ZrO₂here referred as Cu/am-ZrO , Cu/m-ZrO₂and Cu/

2

+

[

20,22,23]. Also, it was reported [19,20] that quite stable Cu as

intermediate Cu species in the CuO reduction were formed as ac-

tive species in the steam reforming of methanol over Cu/ZrO cat-

alysts, but there was no conclusion whether the partial copper

oxidation was due to the Cu/ZrO interaction or to the presence

t-ZrO , respectively.

2

2.2. Methods

2

+

of water. Although there was not directly detection of Cu , it

The crystallographic phase of supports and Cu catalysts was

achieved by XRD. XRD patterns were collected by Rigaku DMax

+

was proposed [8] that the presence of a certain amount of Cu

species over Cu/ZrO

2

catalyst is very important for improving

2500PC, 40 kV e 150 mA, using CuKa (k = 1.5406 Å) radiation.

the selectivity to ethyl acetate.

The data were collected from 20° to 45° in 2h range with 0.5° diver-

gence slit, 0.3 mm receiving slit, in ﬁxed-time mode with 0.02° step

size.

Understanding the chemistry occurring on the surface of oxide-

supported metal clusters provides critical information for the

design and development of superior heterogeneous catalysts. At

the present, a few information exists in the literature about the

direct relation between the electronic properties of copper-based

catalysts supported on zirconia and the product selectivity in the

ethanol conversion. In the present work for ﬁrst time, it is at-

tempted to demonstrate that the catalytic behaviors of Cu catalysts

p

Speciﬁc surface area (SBET) and pore volume (V ) of the samples

were determined from the corresponding nitrogen adsorption–

desorption isotherms measured at 77 K on an ASAP2010 apparatus

(Micromeritics, Norcross, GA, USA).

TPR proﬁles of the samples were performed with a Micromeri-

2

tics Pulse Chemisorb 2705. A ﬂow of 10% H /He (30 ml/min) was

supported on different ZrO

2

phases, such as amorphous (am-),

passed through the sample, and the temperature was raised at a

rate of 10 K/min up to 523 K, while the thermal conductivity detec-

tor (TCD) signal was recorded.

monoclinic (m-), and tetragonal (t-), in the reaction of ethanol con-

version are directly related to the electronic properties of supported

copper species as well as to the speciﬁc properties of support. The

study reports in detail the physicochemical characterization of

XPS data were obtained with an SPECSLAB II (Phoibos-Hsa 3500

150, 9 channeltrons) SPECS spectrometer, with AlK

a source

Cu/ZrO

2

catalysts in relation to their structure and acid–base prop-

(E = 1486.6 eV) operating at 15 kV, Epass = 40 eV, 0.05 eV energy

step and acquisition time of 1 s per point. The samples were placed

on stainless steel sample-holders and were transferred to the XPS

pre-chamber under inert atmosphere and stayed there for 12 h in

a vacuum atmosphere. The residual pressure inside the analysis

chamber was smaller than 1 ꢂ 10 Torr. The binding energies

(BE) were referenced to the C 1s peak at 284.6 eV, providing accu-

racy within ±0.2 eV. All samples were reduced in a furnace from

erties as a function of the support kind. The inﬂuence of the ethanol

pressure, the contact time, and the reaction temperature on the rate

of ethyl acetate and acetaldehyde formation as well as on the prod-

uct selectivity is clariﬁed. To gain insight into the nature of sup-

ported copper species and supports, extensive characterization

ꢀ9

2

was carried out by: N adsorption, X-ray diffraction (XRD), X-ray

photoelectron spectroscopy (XPS), diffuse reﬂectance Fourier trans-

form infrared spectroscopy of CO adsorption (DRIFTS), tempera-

ture-programmed reduction (TPR), in situ X-ray absorption near

edge structure (XANES), and extended X-ray absorption ﬁne

structure (EXAFS) spectroscopy.

room temperature to desired temperature according to H

behavior at 10 K/min, staying at 503 K for 30 min under a 60 ml/

min ﬂow rate of 5% H /He, and then cooled to room temperature

to precede subsequent analysis.

2

-TPR

2

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

3

In the same TPR apparatus described above, N

used to determine the surface area of dispersed metallic copper

2

O titration was

where NCu–Cu is the coordination number of Cu cluster; L is the cop-

per cluster order; and DCu is the dispersion related to the number of

Cu atoms in a crust per total number of atoms in a cuboctahedral

cluster.

0

(

Cu ). The Cu/ZrO

that were cooled to 303 K and exposed to 1% N

for 10 min. It was expected that neither of the ZrO

its a signiﬁcant interaction with N O at that temperature [17]. A

second H -TPR by increasing the temperature up to 773 K at a rate

of 10 K/min under 5% H /He was acquired. The hydrogen consump-

tion was used to calculate the amount of oxygen deposited after

2

samples were reduced up to 523 K and after

O/He (30 ml/min)

supports exhib-

2

The particle size d

p

of copper cluster was calculated according

2

to the (Eq. (2)):

2

D

p

¼ LRCuꢀCu

ð4Þ

2

where RCu–Cu is the inter-atomic distance between copper–copper.

The diffuse reﬂectance FTIR spectra of adsorbed CO were re-

corded using a Thermo Nicolet 4700 Nexus FTIR spectrophotome-

ter with MCT detector and a (DRIFTS) reactor cell Spectra Tech with

19

2

N

2

O titration. A total of 1.46 ꢂ 10 Cu atoms/m and a stoichiom-

etry of 2Cu/H were used [25]. No signiﬁcant bulk oxidation was

2

0

observed for the samples. The Cu surface area was calculated on

the basis of the following relationship in:

CaF

2

windows (DRIFT HTHV cell). IR spectra were collected in 64

SACu ¼ 6:4955 ꢂ 10^ꢀꢂ C ꢂ D

where C is Cu content (%) and D is Cu dispersion (%).

2

ð1Þ

ꢀ1

scans at 4 cm resolution. The samples were dried at 473 K in a

He ﬂow during 1 h, followed by reduction at 503 K in a ﬂow of

0

3

2

4% H /He mixture for 2 h. The CO was introduced for 1 s and ad-

In situ XANES spectra at the Cu K-edge were carried out at the

D06A-DXAS beamline of the Brazilian Synchrotron Light Labora-

tory (LNLS) at Campinas-SP, Brazil. The D06A-DXAS beamline is

equipped with a focusing curved Si (111) monochromator, operat-

ing in Bragg mode, for the selection of the desired range of X-ray

wavelengths (8900–9400 eV). The samples were prepared as self-

supporting pellets containing ꢃ25 mg of catalyst and placed into

sorbed at 323 K at CO pressure of 20 Torr. Temperature-pro-

grammed desorption (TPD) of adsorbed CO was carried out in

ﬂowing He (30 ml/min) with a heating rate of 10 K/min.

2.3. Catalytic test

Activity and selectivity measurements for ethanol reaction were

a

tubular quartz furnace (di = 20 mm and X-ray path

carried out in a continuous-ﬂow, tubular ﬁxed-bed glass reactor

10 mm i.d.) over the temperature range of 473–548 K at atmo-

spheric pressure. Samples were reduced in situ by heating in pure

ﬂow from room temperature to 503 K at 10 K/min staying at

length = 440 mm) sealed with kapton refrigerated windows for

the transmission measurements. Temperature-resolved XANES

spectra at the Cu K-edge were acquired during TPR under a 5%

(

H

2

H /He ﬂow, heating from room temperature up to 800 K at

ꢀ1

this temperature for 1.5 h. Attempting to analyze the effect of eth-

anol partial pressure on the reaction, ﬁrst experiments were car-

ried out isothermally at 473 K with catalyst weight of 0.2 g at

high ethanol conversion at contact time of W/F = 38 min, where

W and F correspond to catalyst weight (g) and ﬂow rate of ethanol

1

0 K min (named XANES-TPR). The information about the pro-

portion of the CuO, Cu O and Cu metal evolution in the H -TPR

experiments to Cu/ZrO polymorphs was achieved by the linear

2

combination ﬁtting (LCF) method. In the LCF method, the X-ray

absorption spectrum is modeled by least-squares ﬁtting using a

linear combination (LCF) of known species (patterns) to ﬁt an un-

known spectrum (samples).

EXAFS spectra were carried out at the D04B-XAFS1 beamline of

LNLS. This bending magnetic beamline is equipped with a (111)

silicon channel cut monochromator and allowed us to perform EX-

AFS measurements. Spectra were acquired at room temperature

(

g/min), respectively. Ethanol (99.9%) in gas phase was fed to cat-

alytic reactor by passing He ﬂow through a saturator submerged

in a water bath with temperature at 296, 313, 328, and 338 K,

allowing a partial pressure of ethanol feeding of 7.0, 17.9, 37.4,

and 58.6 kPa, respectively. The helium gas ﬂow was adjusted keep-

ing a constant ﬂow of reagent dragged into the reactor for all par-

tial pressures of analyzed ethanol. The temperature-resolved

reactions were performed with catalyst of 0.04 and 0.2 g at low

and at high ethanol conversion at low and high contact time of

W/F = 1.2 min and 38.0 min, respectively. Reaction products were

collected periodically and analyzed by online gas chromatography

2

from samples reduced in H at 523 K. The EXAFS signal was ex-

tracted in a standard way using Athena/Artemis software packages.

The ﬁts were done ﬁxing the same passive electron amplitude

reduction factor (S02) and photoelectron energy origin correction

(D

0

E ) found for the references compounds. The coordination num-

(

Varian GC-3400 CX) with a Chromosorb 102 column. Conversion

ber of the path (N ), the effective half-path length, R (which is

i

and selectivity were determined on the carbon basis products.

Turnover frequencies (TOFs) of acetaldehyde and ethyl acetate for-

mation using the Cu metal surface area per gram of catalyst based

on N O titration data (TOFs(N2O)) and EXAFS analysis (TOFs(EXAFS))

2

were calculated. The extent of ethanol conversion was maintained

below 15%.

equal to the inter-atomic distance for single-scattering paths),

2

and the Debye–Waller factor, r_i, were the parameters that were

adjusted. The local environment of the Cu atoms was determined

from the EXAFS using the phase shift and amplitude function for

Cu–Cu and Cu–O calculated including multiple scattering pro-

cesses (FEFF version 6.0). The oscillations were weighted with k

and Fourier transformed within the limit k = 3.5–12 Å . The pro-

2

ꢀ1

cedure to prepare the sample was the same used on DXAS beam-

3

. Results

line. The relatively narrow ﬁtting range

present catalyst systems, since the Hf L

HfO (present as an impurity in the ZrO

Dk, was used in the

3

-edge (9651 eV) due to

2

used) severely hampers

3.1. XRD

2

extraction of Cu K-edge oscillations at higher k values [26].

XRD patterns of pure ZrO

ples are shown in Fig. 1. Intensive diffraction patterns at 2h = 24.2°,

8.2°, 31.4°, and 34.3° of monoclinic ZrO

484) are predominant for Cu/m-ZrO . In addition to the character-

2 2

supports and calcined Cu/ZrO sam-

Based on EXAFS data and considering copper particles in cuboc-

0

tahedral model [27], the Cu surface area was calculated on the ba-

2

1

2

crystal phase (JCPDS 37-

sis of (Eq. (1)) combined with the following relationship:

2

istic peaks of zirconia, XR diffractions at 2h = 35.5° and 38.7° are

detected for Cu/m-ZrO , corresponding to crystalline CuO phase

with tenorite structure (JCPDS 48-1548). The absence of crystalline

CuO peaks in the XRD of Cu/am-ZrO and Cu/t-ZrO samples cannot

be ruled out as they might be less than 4 nm in size, which is be-

yond the detection capacity of the XRD technique. A broad X-ray

2

4Lð5L þ 3L þ 1Þ

N

D

Cu—Cu

¼

ð2Þ

ð3Þ

2

3

2

1

0L þ 15L þ 11L þ 3

2

3

0L þ 6

0L þ 15L þ 11L þ 3

Cu

¼

1

3

2

4

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

satellite at about 942 eV was not detected, which suggests the ab-

2

+

sence of Cu species [29]. It should be noted that the satellite

+

0

peaks are not seen in the case of Cu or Cu species because of com-

pletely ﬁlled 3d shells. The FWHM of Cu 2p3/2 is 2.0–2.4 eV, which

suggests the presence of at least two kinds of surface copper spe-

cies different in their chemical environments. Therefore, it cannot

+

be excluded traces of other species, like Cu , since the BE of Cu

0

generally overlaps with that of Cu in Cu 2p core level.

To distinguish better Cu⁰from Cu species, having similar BE,

the X-ray induced Auger electron spectra of reduced samples in

the kinetic region of 928–902 eV are presented in Fig. 2B. A clear

+

3 45

difference is observed in the shapes of Cu L M M45 Auger photo-

electron spectra of the samples, being in agreement with those

previously reported in the literature for Cu⁰and Cu species [29].

The reported changes have been attributed to the speciﬁc bonding

interactions appearing at the interface between the two oxide

phases species. Peaks at ca. 918.4 and ca.915.5 eV are detected,

+

2 2 2 2

Fig. 1. XRD of ZrO polymorphs and Cu/ZrO catalysts: m-ZrO (j); t-ZrO (N), and

CuO ( ).

_w_h_i_c_h_c_h_a_r_a_c_t_e_r_i_z_e_C_u0 and Cu

spectra, the position of the main peak for Cu/am-ZrO

about 915.6 eV is shifted to a higher kinetic energy, up to

18.4 eV for Cu/t-ZrO , which is usually assigned to the increase

2

O, respectively [29]. In the Auger

sample at

2

diffraction patterns are observed for am-ZrO

an amorphous or disordered at long range nature of zirconia. For

Cu/t-ZrO , diffraction peaks at 2h = 30.3° and 35.4° of t-ZrO crystal

phase (JCPDS 50-1089) are detected.

2

sample, pointing to

9

2

0

inf Cu amount. The latter is in agreement with the highest Cu sur-

face area detected over t-ZrO (Table 1). The modiﬁed Auger

parameter ( Cu) was calculated to determine more precisely the

chemical state of copper. This parameter is deﬁned by [29]:

2

a

3.2. Textural properties

aCu ¼ E

B

þ E

K

ð5Þ

Some textural characteristics of the supports and Cu/ZrO

2

sam-

where E

netic energy of the Cu LMM Auger electron. The

B

is the binding energy of Cu 2p core level and E

K

is the ki-

0

ples are listed in Table 1. The Cu surface area (SACu) and the cor-

a

Cu values at ca.

0

responding apparent Cu dispersion (DCu) for each sample obtained

0

1+

1

851.0 and 1848.1 eV (Table 2) correspond to Cu and Cu species,

by N

the Table 1. The SBET values of calcined Cu/ZrO

compared to those of the pure supports. Within the zirconia sup-

ports, series am-ZrO has the highest speciﬁc surface area.

The data N O titration and EXAFS analysis in Table 1 demon-

strate that the Cu/t-ZrO

sion. Similar results have been observed by Rhodes and Bell [14],

2

O titration method and from EXAFS analysis are also given in

respectively. The aCu value of 1851.0 eV obtained for reduced Cu/t-

ZrO sample means that the prevailing species over catalyst surface

2

samples are lower

0

+

is Cu , whereas the value of ca. 1848.1 eV suggests that Cu is a

dominant species for reduced Cu/am-ZrO (Table 2 and Fig. 2B).

The both features at 1848.0 and 1850.8 eV correlate with the pres-

2

0

2

exhibits the highest apparent Cu disper-

ence of Cu_a_n_d_C_u0 species, respectively, over m-ZrO

that the KE of Cu increases as a function of the amount of Cu , sim-

ilar to the observations in Ref. [29]. It is interesting to note that for

+

2

. It is clear

0

a

0

who assigned the lower Cu dispersion on m-ZrO

of zero charge of 8.5 and charge density of the monoclinic poly-

morph [28]. The net positive charge at the surface of m-ZrO leads

2

to a higher point

0

2

Cu and Cu O, the changes in aCu are larger than those of BE [22].

2

All samples exhibit a spin–orbit doublet of the Zr 3d core level

into 3d5/2 and 3d3/2 levels with an energy gap of 2.4 eV between

them and a relative intensity ratio (I 3d5/2/I 3d3/2) of 1.5 (Fig. 2C).

to a weaker interaction between the support and Cu species [28],

which contributes to a lower dispersion and formation of bulk

CuO as was detected by the XRD in Fig. 1.

2

This indicates the existence of ZrO -like species according to the

literature data [30]. Decomposition of the spectra produces peaks

attributed to the existence of two kinds of zirconium species, re-

3.3. XPS analysis

ferred as species I with low BE in the range 181.9–182.4 eV (Zr

and species II with higher BE in the range 183.2–183.9 eV (ZrII) (Ta-

ble 2). It should be noted that the fraction of Zr species for all sam-

ples is larger compared to that of species ZrII. The BEs of Zr are

similar to that of Zr ions in pure zirconia, but with slightly lower

values compared to that of stoichiometric ZrO (182.6 eV), espe-

cially for sample Cu/am-ZrO (181.9 eV), probably, due to some

I

)

XPS of Cu 2p and Auger peak, Zr 3d and O 1s core electron level

of reduced Cu/ZrO

2

samples are shown in Fig. 2A, B, C, and D,

I

respectively. The XPS parameters are summarized in Table 2. All

reduced samples exhibit symmetric Cu 2p3/2 and Cu 2p1/2 main

peaks (Fig. 2A) with BEs values at ca. 932.5 and 952.4 eV, respec-

tively, with a spin–orbit coupling energy of 20 eV. A shake-up

I

4

+

2

Table 1

Textural characteristics for supports and Cu/ZrO

2

catalysts, surface metallic area (SACu) and apparent metallic dispersion of Cu (DCu) based on N

2

O titration and EXAFS analysis.

2

3

EXAFS^a

SACu (m /gcat

Samples

S

BET (m /gcat

)

V

p

(cm /gcat

)

N

2

O

Data

2

b

SACu (m /gcat

)

D

Cu (%)

)

D

Cu (%)

am-ZrO

m-ZrO

t-ZrO

Cu/am-ZrO

Cu/m-ZrO

Cu/t-ZrO

2

376

92

152

265

76

0.59

0.28

0.19

0.43

0.22

0.13

–

14

13

22

–

22

20

34

–

52.4

37.6

58.6

–

80.7

57.8

90.2

2

115

a

Assuming cuboctahedral model;

Number of atoms in a crust per total number of atoms.

b

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

5

2

Fig. 2. XPS of Cu 2p (A), Cu Auger (B), Zr 3d (C), and O 1s core level (D) for reduced Cu/ZrO samples.

Table 2

XPS parameters of reduced Cu/ZrO

2

samples.

Cu Auger (eV)

Samples

Cu 2p3/2 (eV)

a

Cu (eV)

Zr 3d5/2 (eV)

Zr species

O 1s (eV)

species

Cu/Zr

I

ZrII species

O

I

OII species

Cu/am-ZrO

2

932.5 (2.4)^a

932.5 (2.3)

932.6 (2.0)

915.6

915.5/918.3

918.4

1848.1

1848.0/1850.8

1851.0

181.9 (83)^b

182.1 (83)

182.4 (88)

183.2 (17)^b

183.4 (17)

183.9 (12)

529.8 (44)^b

530.0 (49)

530.3 (53)

530.9 (56)²

531.6 (51)

532.2 (47)

0.40

0.49

0.64

Cu/m-ZrO

Cu/t-ZrO

2

a

Full width at half maximum (FWHM).

Percent of Zr and O species.

b

6

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

oxygen deﬁciency. The position shift toward the lower binding en-

ergy might be associated with the holes created by oxygen vacan-

oxide particles, and (ii) a higher temperature region (from 450 to

530 K) assigned to highly dispersed copper species in a strong

interaction with the support as well as to well crystallized CuO

2

cies in the ZrO lattice [31]. The BEs of the higher energy

component (183.2–183.9 eV) correspond to the formation of a Zr

species bound to a more electron attractive species and formation

of partially reduced Zr sites [30].

in weak interaction with ZrO

an intimate contact between Cu and Zr for Cu/m-ZrO

2

. It can be suggested that there is

sample,

2

d+

which facilitates the reduction in supported copper oxide species

at lower temperatures.

The TPR ﬁndings indicate the presence of Cu–Zr interaction in

the samples in different levels as well as distinct active sites behav-

ior where highly dispersed copper oxide species, crystallized cop-

The value of XPS atomic Cu/Zr ratio for reduced Cu/ZrO

increases in the following order: Cu/am-ZrO > Cu/m-ZrO

Table 2). The highest Cu/Zr ratio value of 0.64 for Cu/t-ZrO

2

samples

> t-ZrO

means

2

surface. As is shown in

2

(

2

a pronounced Cu enrichment on the t-ZrO

Table 1, the highest apparent Cu⁰density and Cu surface area

0

per oxide strongly or weakly interacted with the ZrO

coexist.

2

surface

measured by N

2 2

O titration are found for Cu/t-ZrO sample.

The changes in the XPS of O 1s core electron of reduced Cu/ZrO

2

samples can be explained by the superposition of contributions

from the substrate and deposited overlayers, as shown in Fig. 2D.

The O 1s broad peaks can be decomposed in two peaks at the cor-

responding position using XPS peak splitting program (XPS Casa

Software), whose relative contents are shown in Table 2. According

to the literature [31,32], there might be two types of oxygen spe-

3.5. In situ XANES analysis

The change in the chemical state and local symmetry of copper

is clariﬁed by analyzing the normalized (Fig. 4A) and ﬁrst deriva-

tive (Fig. 4B) of Cu K-edge XANES spectra of calcined Cu/ZrO

ples, including the patterns of CuO and Cu(OH) references.

Normalized spectra of reduced Cu/ZrO samples at the end of

reduction process (800 K) as well as of Cu O and Cu metal foil

2

sam-

2

cies in the Cu/ZrO

ZrO and/or Cu O (O

2

system after reduction: oxygen species of

) and oxygen species of Zr–OH (OII), whose

2

I

2

binding energy is in the range of 529.8–530.3 eV and 530.9–

32.2 eV, respectively (Table 2). It was shown that the oxygen spe-

cies with BE P531.0 eV are attributed to OH, chemisorbed oxygen

are depicted in Fig 4C. There is a relationship between the position

5

of the pre-edge peak (due to the dipole forbidden 1s ? 3d) and the

2

+

coordination geometry of Cu compounds patterns. The compari-

son between the positions of the pre-edge peak of calcined samples

or carbonates groups [33].

2

+

and of Cu compounds patterns (Fig. 5A) suggests the following:

i) a square planar geometric coordination for Cu/m-ZrO, analogous

to CuO and (ii) an octahedral symmetry for Cu/am-ZrO and Cu/t-

ZrO , analogous to Cu(OH) [36,37]. The energy position of the

maximum in the pre-edge peak increases from octahedral to

square planar [37]. The peak assigned to the 1s ? (4p + L) shake-

down transition involving a 1s ? 4p transition with simultaneous

ligand-to-metal charge transfer (LMCT) is also known as the edge

transition. The calcined Cu/m-ZrO sample shows this transition

at ca. 8984 eV typically for Cu compounds. The edge transition

for Cu/am-ZrO and Cu/t-ZrO samples shows a small shift to

986 and 8985 eV, respectively (Fig. 4A) [38,39].

A strong intensive white line (WL) at ca. 8997 eV assigned to di-

pole 1s ? 4pxy electron transition [40] is observed for CuO,

Cu(OH) , and calcined Cu/ZrO samples (Fig. 4A). The electron tran-

sition for Cu O appears at 8996 eV. The observed shoulder at

990 eV for CuO reference in Fig. 4B could be assigned to the cova-

(

3

.4. TPR

2

Fig. 3 shows that the TPR proﬁles of calcined Cu/ZrO

2

samples

2

shows a

depend on the kind of ZrO phase. TPR of Cu/am-ZrO

broad intensive heterogeneous peak ranging from 415 to 480 K

with a maximum at 454 K and a shoulder at ca. 440 K. Similarly,

2

z

Cu/t-ZrO

ature region; there is a peak at 440 K with a shoulder at 418 K. In

addition, a small peak at 477 K is appeared in the TPR of Cu/t-ZrO

Similar reduction peaks for Cu/ZrO samples have been previ-

ously observed. It has been found [34,35] that the presence of Cu

oxide species differently interacted with the ZrO matrix, depend-

ing on the CuO dispersion and crystallinity as well as on the ZrO

phase. Usually, dispersed CuO is reduced more easily than larger

CuO particles. Comparing the TPR proﬁles of Cu/am-ZrO and Cu/

t-ZrO , it can be concluded that supported Cu oxide species on t-

ZrO are more easily reduced than that over am-ZrO

In contrast to the TPR proﬁles of the above-mentioned samples,

the TPR of Cu/m-ZrO exhibits more heterogeneous distribution of

Cu oxide species (Fig. 3). Reduction peaks for Cu/m-ZrO could be

divided in two well deﬁned regions: (i) a lower temperature region

from 400 to 425 K), attributed to the reduction in small copper

2

exhibits also an intensive peak but in the lower temper-

2

+

2

.

2

8

2

8

2

lence degree of the interaction and/or to a disorder degree of the

local structure.

2

.

The absorption edge for Cu

s ? (4p + L) appears at 8981.5 eV, while Cu metal foil exhibits

absorption edge at 8979 eV and a well-resolved doublet in the

post-edge region (Fig. 4C). The values for CuO, Cu foil, and Cu

2

O due to electronic transition

2

1

z

2

O

(

reference are in a good agreement with those reported in the liter-

ature [23,29]. The XANES spectra collected at the highest temper-

ature of reduction (800 K) suggest that the absence of WL and the

1

s ? (4p

and that Cu is the dominant species in reduced Cu/ZrO

Fig. 5(A-C) shows in situ Cu K-edge XANES spectra evaluated

during the H -TPR of Cu/am-ZrO (A), Cu/m-ZrO (B), and Cu/t-

ZrO (C) samples in the photon energy region of 8940–9040 eV.

z

+ L) electron transition at ca. 8979 eV is an indication

0

2

samples.

2

The respective semiquantitative change in the percentage of differ-

ent Cu species with increasing the reduction temperature within

3

00–800 K is shown in Fig. 5D–F. The corresponding ﬁts, using a lin-

ear combination of CuO, Cu O, and Cu foil standards, provided a

2

form to monitor the electronic state of copper species evolution.

It is observed that the reduction in copper oxide species on Cu/t-

ZrO

to Cu/am-ZrO

correlates with the highest apparent metal dispersion detected for

Cu/t-ZrO sample (Table 1). On the other hand, the gradual reduc-

2

starts at signiﬁcantly lower temperature (at 398 K) compared

2

and Cu/m-ZrO (at 448 and 443 K, respectively). This

2

Fig. 3. TPR proﬁles of calcined Cu/ZrO

2

samples.

2

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

7

K-edge are shown in Fig. S1 (see Supplemental material). The num-

ber of neighboring of Cu atoms (N), the distance between them (R),

2

and the statistical order (Debye–Waller factor

r

) as well as the

particles size calculated from the EXAFS analysis are listed in Table

. The particle size was determined from the average coordination

3

number of the nearest metal neighbor atoms assuming a cubocta-

hedral geometry [41]. It should be noted that the values of the EX-

0

AFS parameters for references CuO, Cu

2

O, and Cu are in agreement

with those reported in the literature [42]. The values of inter-atom-

ic distances RCu–Cu for the samples (2.53 ± 0.001 and 2.54 ± 0.002 Å)

0

related to the Cu–Cu distance in bulk Cu indicate that the local

structure of supported Cu species is similar to that of Cu⁰

(

2.54 Å). However, the lower Cu–Cu coordination numbers for sup-

ported Cu samples in the range 6.0–9.3 compared to the those of

bulk CuO (10) and Cu O (12) as well as the different Cu–O bond

distances (1.89–1.87 Å) relative to those of bulk CuO (1.94 Å) and

Cu O (1.84 Å) indicate that the local structure around supported

Cu species differs from that of the bulk CuO and Cu O. It was

shown that the kinetic reductions in Cu/ZrO polymorphs are quite

2

different in relation to unsupported CuO; a direct CuO ? Cu trans-

formation occurs on unsupported CuO. Yang et al. [43] showed that

both the TPR and isothermal reduction detected easiness in the

reduction property of CuO than of Cu

2

O under the same reaction

conditions. The apparent activation energy for the reduction in

CuO is about 14.5 kcal/mol, while for Cu

2

O, it is 27.4 kcal/mol

[

(

44]. The coordination number of copper particles for Cu/m-ZrO

2

9.3) is the highest in comparison with those for Cu/am-ZrO

2

7.4) and Cu/t-ZrO (6), indicating an agglomeration of the copper

particles over monoclinic zirconia. The Cu–O bond distance for Cu/

m-ZrO was not found after reduction, since the EXAFS technique is

2

limited to large particles. It could be supposed that a part of the

oxygen species is located on the surface of copper particles accord-

ing to the observations by Knapp et al. [45]. The authors have

showed that the oxygen atoms are located in the tetrahedral or

octahedral vacancies of the copper surfaces, and about of one third

of the copper surface atoms are blocked by oxygen [45].

3.7. FTIR of CO adsorption

2

The FTIR spectra of CO adsorption on reduced Cu/ZrO samples

ꢀ1

in the high frequency (HF) region of 2175–2025 cm and in low

ꢀ

1

frequency (LF) region of 1850–1150 cm are shown in Figs. 6A–

C and D–F, respectively. It is well known [46] that the CO adsorp-

ꢀ1

tion on copper surfaces exhibits IR bands at 2250–2000 cm spec-

tral region associated with the linear or bridge-bonded CO species

0

interacting with CuO, Cu

2

O, or Cu sites. The bands at 2116–

ꢀ1

+

2

111 cm have been related to Cu -CO carbonyls, while the bands

at 62100 cm have been assigned to CO-adsorbed on Cu [46].

In the copper spectral region, the IR spectra of reduced Cu/am-

Fig. 4. Normalized XANES spectra at Cu K-edge (A) and respective derivative of

XANES spectra of calcined (B) and reduced Cu/ZrO samples (C).

ꢀ1

0

2

tion in oxide copper species for Cu/t-ZrO

er temperature (at 448 K, Fig. 5C) compared to those of the other

ones (at 573 and 623 K for Cu/am-ZrO and Cu/m-ZrO , respectively

Fig. 5A and B). It should be noted that all Cu/ZrO samples showed

a transition region in which a mixture of three copper species, Cu

2

sample is ﬁnished at low-

2 2

ZrO2, Cu/m-ZrO and Cu/t-ZrO exhibit strong bands with maxi-

ꢀ1

mum at 2113, 2102, and 2100 cm , respectively (Fig. 6A–C). The

observed shift in the band maximum of the samples may be caused

by changes in the structure and electronic properties of Cu species.

2

(

2

+

0

,

According to the literature data [46], the Cu –CO carbonyls can be

+

0

+

Cu , and Cu is detected, i.e., CuO is reduced to metallic Cu through

the intermediate Cu O. Similar results have been observed in Ref.

19]. It is interesting to note that a small percentage of Cu species

at the end of reduction (up to 800 K) still exists on Cu/m-ZrO

Fig. 4E). It means that a more difﬁcult reduction in the Cu oxide

species is occurred for Cu/m-ZrO sample.

adsorbed at the same frequency, like that of Cu –CO ones, when

the copper is highly dispersed on the surface. The both species

can be better distinguished by their thermal stability criterion:

2

+

[

0

the surface Cu –CO carbonyls are easily destroyed with increasing

2

+

(

the temperature, while Cu cations form more stable surface car-

bonyls, being seen in Fig. 6A–C. It should be noted that the high

2

+

stability of Cu –CO species is a result of the stabilization of stron-

ger

supported on am- and t-ZrO

ture (at 373 K) in comparison with that supported on m-ZrO

r

-bond [47]. The adsorbed carbonyl species for Cu samples

are decomposed at higher tempera-

(at

348 K). Thus, the IR band maximum at 2113 cm of Cu/am-ZrO

3

.6. EXAFS analysis

2

ꢀ1

Fourier transform magnitudes of the experimental EXAFS

spectra and respective ﬁts for Cu metal and oxides patterns at Cu

2

+

sample would be associated with the presence Cu –CO carbonyls

8

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

2

Fig. 5. Temperature-resolved XANES spectra at Cu K-edge of Cu/ZrO samples under hydrogen atmosphere: normalized spectra (A–C) and corresponding percentage of Cu

species (D–F).

Table 3

EXAFS parameters of bulk CuO, Cu

2

On and Cu⁰references and of CuZrO

2

catalysts reduced at 503 K in ﬂow of 5%H

Samples

2

/He.

References

CuO

Cu

2

O

Cu⁰

Cu/am-ZrO

2

Cu/m-ZrO

2

Cu/t-ZrO

2

N

Cu–O

4

10

2

12

0.6

7.4 ± 0.1

1.89 ± 0.01

2.53 ± 0.001

2.67 ± 0.002

0.008 ± 0.0002

4.4

0.7

6.0 ± 0.1

Cu–Cu

12

9.3 ± 0.1

RCu–O (Å)

1.94 + 0.006

1.84 + 0.007

1.87 ± 0.01

2.54 + 0.001

2.70 + 0.002

0.008 ± 0.0003

2.9

R

Cu–Cu (Å)

2.54 + 0.002

0.009 ± 0.0005

2.54 ± 0.001

a

Cu–Cu (Å)

3.03 + 0.004

0.004 ± 0.001

2

r

0.002 ± 0.001

0.009 ± 0.0002

8.7

b

d

P

(nm)

a

Contribution of Cu–Cu bond length for Cu

d : particle size in cuboctahedral model.

P

2

O.

b

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

9

Fig. 6. FTIR spectra of CO adsorption on reduced Cu/ZrO

2

samples in the HF region of 2175–2025 cm ¹(A–C) and in LF region of 1850–1150 cm (D–F).

ꢀ

ꢀ1

as a dominant species. The position of the IR band at ca. 2100 cm

for Cu/m-ZrO and Cu/t-ZrO can be related to the adsorbed car-

bonyls of metallic copper as a prevailing species. However, the car-

bonyl species on Cu/t-ZrO has a slightly higher thermal stability

compared to that on Cu/m-ZrO , which means the presence of cop-

1220 cm

are assigned to bridged bidentate carbonates

ꢀ

ꢀ1

2

(b-HCO ), whereas the peaks at ca. 1547 and 1375 cm and at

3

ꢀ1

2ꢀ

ca. 1343 cm are related to monodentate ðm-CO Þ and bidentate

3

2

3

ꢀ

2

ðb-CO Þ carbonate species, respectively [49–51]. The no well-re-

ꢀ1

2

solved band at ca. 1450 cm can be connected with the formation

2+

3ꢀ

per species with a lower electron density. No Cu –CO species were

detected in all IR spectra since they usually show bands

of ionic carbonate species (i-HCO ).

Usually, the bicarbonate species are formed by adsorption of CO

on basic OH groups. The formation of bicarbonates on m-ZrO is

2

ꢀ1

>2150 cm [46].

The CO adsorption capacity on ZrO

2

support as well as the posi-

connected with the higher fraction of bridged-bonded hydroxyl

groups on its surface [51] The above-mentioned both bands corre-

sponding to bidentate carbonates mean different geometrical con-

ﬁgurations and/or different crystal planes. The surface bidentate

carbonate surface complexes involves acid–base pair sites (cus

ꢀ1

tion and intensities of the IR bands in the range of 1850–1150 cm

depend on the kind of ZrO phase (Fig. 6D–F). The different surface

2

species formed upon CO adsorption can yield information on the

existence of surface basic sites (basic OH groups and coordinatively

2ꢀ

4+

2ꢀ

unsaturated sites, cus O centers) or acid–base pair sites (cus Zr

–

Zr –O centers), while the monodentate carbonates involve cus

2

ꢀ

2ꢀ

O

centers) [48]. Fig. 6D–F reveals that the band distribution for

the three samples is quite different. Following the CO adsorption

on Cu/m-ZrO sample (Fig. 6E), the following relatively well-re-

solved bands are observed at 1635, 1547, 1450, 1375, 1343, and

O

centers of high basicity [50].

2

The comparison of the IR spectra of Cu/m-ZrO with those of Cu

2

samples supported on am- and t-ZrO

that for Cu/am-ZrO , there is a broad band in the higher wavenum-

ber range of 1700–1400 cm assigned to b-HCO ; b ꢀ CO and

2

phases (Fig. 6D–F) indicates

2

ꢀ

1

ꢀ1

ꢀ

2ꢀ

1

220 cm , which are related to different CO bonded species

3

ꢀ

caused by the presence of acid–base sites. The peaks at 1635 and

i-HCO₂species. Two kinds of carbonate species are observed for

1

0

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

Cu/t-ZrO

2

; (i) bidentate carbonate species revealed by the broad

ethanol for Cu catalysts supported on am- and m-ZrO

practically, do not change with the increase in PEtOH, while for

Cu/t-ZrO , there is a decrease in the conversion. It should be noted

that for all catalysts, the selectivity to ethyl acetate increases with

the partial pressure at the expense of the acetaldehyde consump-

2

phases,

ꢀ

1

ꢀ1

band at ca. 1064 cm and by the small one at ca. 1348 cm and

2

(

ii) polydentate carbonate species ðp-CO Þ [52] detected by the

2

3

ꢀ

1

2ꢀ

strong bands at ca. 1426 and 1411 cm . The p-CO₃species is

caused by the presence of strong Lewis acid sites due to the closely

4

+

spaced cus Zr sites at the t-ZrO

2

surface [52].

2

tion, as the highest values are observed for Cu/m-ZrO (Table 4

The effect of ZrO phase on the thermal stability of adsorbed CO

is illustrated in Fig. 6D–F, showing TPD of adsorbed CO. The resis-

tance to thermal desorption of CO is a way to qualify the strength

2

and Fig. S2). The selectivity to methyl ethyl ketone and crotonalde-

hyde also decreases with the increase in ethanol partial pressure;

2

the highest selectivity is observed for Cu/t-ZrO catalyst (Table

2

and nature of the acid–base sites over Cu/ZrO . A band intensity

4). Both the Cu/am-ZrO and Cu/m-ZrO catalysts show low depen-

2

decrease and some displacement in the IR spectra of all samples

with increasing the temperature of CO desorption are observed.

For CO-adsorbed at 323 K, the CO adsorption capacity on Cu/m-

dence of PEtOH in the crotonaldehyde formation.

It is interesting to note that the total amount of the products

(hereafter designed as PCD) formed via acetaldol by aldol conden-

sation (including MEK/BUT and CROT) and the products of ethanol

dehydration (including DEE and ETE) [5] are in parallel with the in-

crease in the ethanol partial pressure for Cu catalysts supported on

ZrO

73 K) than that on Cu/t-ZrO

to 398 K). It means that CO is bound more strongly on the m-

2

is retained to a signiﬁcantly higher temperature (up to

4

2

(up to 398 K) and Cu/am-ZrO (up

2

ꢀ

1

ZrO

1

2

surface revealed by the presence of bands at 1547 cm

,

m- and t-ZrO

tween the formation of PCD and the water released during the

reaction. However, the highest amount of PCD for Cu/t-ZrO cata-

lyst does not correspond to a relatively amount of released water

like that for Cu/am-ZrO and Cu/m-ZrO (Table 4). The ratio of

water/PCD is different for each catalyst.

2

(Table 4). All catalysts keep a speciﬁc relation be-

ꢀ1

2ꢀ

343 cm , and at 1375 cm assigned to b-CO₃and m-CO₃spe-

cies, respectively. It should be noted that the thermal stability of

hydrogen carbonates is very limited being eliminated at 373 K.

Most likely, the differences in the CO adsorption capacity and

2

strength on the different ZrO

the local environment of the OH groups. As was shown in Ref.

14], the difference in the adsorption capacity of CO on Cu/m-

ZrO and Cu/t-ZrO is expected to be a function of the relative con-

centration of anionic vacancies present on the surface of the two

polymorphs of ZrO . The higher CO adsorption capacity on m-

ZrO compared to that on t-ZrO might be related to a higher con-

2

phases are due to differences in

Fig. 7 shows the changes in ethanol conversion (A), ethyl ace-

tate selectivity (B), and acetaldehyde selectivity (C) as a function

of the reaction temperatures at different contact times at high

(W/F = 38 min) and at short contact time (W/F = 1.2 min) and at

constant PEtOH = 37.4 kPa. The activity of all catalysts increases

with increasing the reaction temperature. The behaviors of Cu/

[

2

centration of surface anionic vacancies at the surface of zirconia.

Such vacancies expose more cus zirconium cations and enhanced

the Brönsted acidity of adjacent Zr–OH groups, which are only

weak acidic. Thus, the creation of accessible Lewis acid sites in

combination with moderately acidic Brönsted acid sites facilitates

the adsorption of CO, which is a basic molecule.

2

ZrO catalysts at temperatures ranging from 473 to 548 K are

equivalent in terms of ethanol conversion for both the short and

high contact time (Fig. 7A). However, concerning the selectivity

to acetaldehyde and ethyl acetate a great difference is observed

for zirconia-supported Cu catalysts at different contact times. The

reaction prioritizes the ethyl acetate formation at high contact

time (Fig. 7B), whereas acetaldehyde is readily formed at short

contact time (Fig. 7C). It should be noted that the effect of the kind

3.8. Performance of Cu/ZrO

2

catalysts in the reaction of ethanol

of ZrO

selectivity to ethyl acetate of Cu/m-ZrO

higher compared to those of Cu supported on am- and t-ZrO

high contact time. However, there is a marked decrease in the

selectivity to ethyl acetate for Cu/m-ZrO catalyst at temperature

2

phase has a strong inﬂuence on the product selectivity. The

catalyst at 473–523 K is

at

conversion

2

The effect of the partial pressure of ethanol (PEtOH) on the etha-

nol conversion and products selectivities are listed in Table 4. The

main reaction products are acetaldehyde (AcH) and ethyl acetate

2

above 548 K (from 72.5–73.1% to 57.3%, Fig. 7B). The highest selec-

tivity to acetaldehyde at short contact time is observed for Cu/am-

(

EtOAc). By-products of ethanol reaction are also formed, which

are methyl ethyl ketone (MEK), butanol (BUT), crotonaldehyde

ZrO

Cu/t-ZrO

2

catalyst at reaction temperature of 473 K (80.1%), followed by

and Cu/m-ZrO (Fig. 7C).

(

CROT), diethyl ether (DEE), propanone (PROP) or acetone, ethene

ETE), and others like, CO and CO . The conversion values of

2

Table 4

Effect of the partial pressure of etanol (PEtOH) on conversion and selectivity of Cu/ZrO

2

catalysts. Catalytic test at Treac = 573 K and W/F = 38 min.

Catalysts

P

EtOH (kPa)

X (%)

Selectivity (%)

Quantity (mol)

Total PCD

Ratio

O/PCD

AcH

EtOAc

MEC^a

CROT

DEE

PROP

ETE

Others^b

H

2

O

H

2

Cu/am-ZrO

2

7.0

44.9

50.1

52.0

48.0

43.9

31.5

23.9

20.7

49.2

61.7

69.6

72.6

6.6

6.2

4.8

5.1

–

0.7

0.2

–

0.4

0.5

0.2

0.3

0.2

0.4

–

0.1

0.2

0.3

60.3

130.1

100.9

55.9

271.0

409.9

339.7

227.1

4.5

3.2

3.4

4.1

1

3

5

7.9

7.4

8.6

0.1

0.2

Cu/m-ZrO

2

7.0

7.9

7.4

8.6

46.2

46.8

47.7

48.9

40.7

29.1

21.3

15.6

54.8

66.8

75.0

80.5

3.2

2.6

1.9

1.6

0.3

0.6

0.7

1.2

–

0,9

–

0.4

0.5

–

0.1

0.3

40.6

81.6

108.5

160.4

126.6

274.8

499.5

741.3

3.1

3.4

4.6

1

3

5

0.6

0.4

0.3

Cu/t-ZrO

2

7.0

7.9

7.4

8.6

56.9

48.1

45.1

43.0

54.1

44.3

35.5

29.7

27.1

33.1

41.0

46.9

13.8

13.2

12.2

11.1

3.7

8.2

10.1

11.1

0.7

0.5

0.3

0.2

0.1

0.2

0.4

200.6

520.8

913.2

541.4

1148.7

2017.8

2688.5

2.7

2.2

2.3

1

3

5

1188.3

a

And/or butanol (BUT).

b

2

Others: CO and CO . Abbreviations: ethanol conversion (X); acetaldehyde (AcH); ethyl acetate (AcOEt); methyl ethyl ketone (MEK); crotonaldehyde (CROT); diethyl ether

(

DEE); propanone (PROP); ethene (ETE); and products of condensation and dehydration (PCD).

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

11

catalysts showed active sites for methyl ethyl ketone/butanol and

crotonaldehyde formation. An increase in the selectivity to prod-

ucts of acetaldehyde condensation with increasing the reaction

temperature at short contact time is observed (Fig. S3). In opposite

2

to that, these by-products for Cu/t-ZrO catalyst decay pro-

nouncedly with increasing the reaction temperature at high con-

tact time (Fig. 8A and B).

Fig. 9 shows the rate of the formation of acetaldehyde and ethyl

acetate at short contact time (over a range of levels at steady state)

as a function of the reaction temperature. At all reaction tempera-

2

tures, the Cu/t-ZrO catalyst demonstrates the highest rates in the

dehydrogenation of ethanol to acetaldehyde. On the other hand,

the same catalyst shows the lowest rate of the ethyl acetate forma-

tion. The highest rate of ethyl acetate formation is observed for Cu/

m-ZrO

catalyst supported on am-ZrO

rate of ethyl acetate formation.

The apparent activation energy (E

acetate formation (Table 5) was calculated from the slop of Arrhe-

nius plots (Fig. S4). A major E for ethyl acetate formation is ob-

tained for Cu/am-ZrO catalyst, which means that the reaction

rate is very sensible to the temperature, being revealed by the

highest inclination of the Arrhenius curve (Fig. S4, B). Cu/t-ZrO

displays the lowest E for both the acetaldehyde and ethyl acetate

2

up to 513 K. However, at temperature above 513 K, the Cu

2

exhibits a strong increase in the

a

) of acetaldehyde and ethyl

a

2

a

formation, which means a low dependence of the reaction rate on

the temperature. The latter is supported by the lowest inclination

of the curve (Fig. S4, A).

The turnover frequencies (TOFs) of acetaldehyde and ethyl ace-

tate formation are listed in Table 5. The high activity of Cu/ZrO

2

catalysts in relation to ethyl acetate formation is supported by

0

the higher values of the apparent TOFs(N2O), based on the Cu sur-

face area per gram of catalyst, compared to those found for Cu/SiO

catalyst [53]. The latter is more evident for Cu/m-ZrO catalyst (Ta-

2

ble 5). TOFs(EXAFS) show similar activities of acetaldehyde forma-

tion over all catalysts and the highest activity of ethyl acetate

formation over Cu/m-ZrO2.

Fig. 10 exhibits the evolution of the selectivity to propanone

and ethyl acetate as a function of the reaction temperature at high

contact time. It is observed that the selectivity to propanone for

2 2

Cu/am-ZrO and Cu/m-ZrO catalysts increases at the expense of

the decrease in the selectivity to ethyl acetate, which is much more

visible for the second one at the highest reaction temperature.

The change in the amount of PCD and the water/PCD ratio for

each catalyst at different reaction temperatures is shown in

Fig. 11. The quantity of water increases in the following order:

Cu/am-ZrO

highest amount of water and PCD as reaction efﬂuent is observed

for Cu/t-ZrO , the lowest water/PCD ratio is calculated. For Cu/

am-ZrO and Cu/t-ZrO catalysts, practically, there is no signiﬁcant

2 2 2

< Cu/m-ZrO < Cu/t-ZrO (Table 4). Even though the

2

change in the water/PCD ratio with the increase in reaction tem-

perature (Fig. 11). It is interesting that the water/PCD ratio for

Cu/m-ZrO increases from 3.2 up to 5 at 498–523 K, and after that,

2

it is decreased to 3.4 at the highest reaction temperature (548 K).

The highest value of the ratio should be caused by the resulting

amount of PCD formed upon a temperature increase, but the

amounts of water do not follow the same trend (Fig. 11B). This

phenomenon will be discussed in Section 4.2.

Fig. 7. Effect of the contact time on the ethanol conversion (A), product selectivity

at high contact time (F/W = 38 min) (B) and at short contact time (F/W = 1.2 min)

(

C) as a function of the reaction temperature for Cu/ZrO

2 2

catalysts: Cu/am-ZrO ( )

Cu/m-ZrO (j) and Cu/t-ZrO ). (PEtOH = 37.4 kPa).

2

(

4. Discussion

2

4.1. Nature of the species in Cu/ZrO system

There is a similarity in the behavior of Cu/ZrO

tion to methyl ethyl ketone/butanol and crotonaldehyde selectivity

Table 4). Even though these by-products of acetaldehyde conden-

sation should not be formed at short contact time, the Cu/ZrO

2

catalysts in rela-

The present results show that the zirconia-supported copper

species in reduced catalysts are in different oxidation states, in the

form of Cu and Cu , which distribution depends on the nature of

(

+

0

2

1

2

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

Fig. 8. Evolution of the selectivity to methyl ethyl ketone/buthanol (A) and crotonaldehyde (B) as a function of the reaction temperature for Cu/ZrO

); Cu/m-ZrO (j) and Cu/t-ZrO ).(W/F = 38 min, PEtOH = 37.4 kPa).

2 2

catalysts: Cu/am-ZrO

(

2

(

Fig. 10. Evolution of selectivity to ethyl acetate and propanone as a function of the

reaction temperature for Cu/ZrO catalysts: Cu/am-ZrO ); Cu/m-ZrO (j) and

Cu/t-ZrO ). (W/F = 38 min, PEtOH = 37.4 kPa).

2

(

2

(

surface-to-volume ratio; (ii) increase in the electron density of sup-

ported particles in respect to bulk metal; (iii) bonding interactions

at the interface and its nature; and (iv) differences in the polariza-

tion properties of the oxide substrates put in contact. Therefore,

the changes in the electronic properties of copper have to be associ-

ated with one or more of these factors, which can be a function of the

Fig. 9. Rate of the formation of acetaldehyde and ethyl acetate over Cu/ZrO

catalysts: Cu/am-ZrO ); Cu/m-ZrO (j) and Cu/t-ZrO ). (W/F = 1.2 min,

2

(

2

(

PEtOH = 37.4 kPa).

+

0

ZrO

2

support phase. According to the literature data [54,55], the

ZrO

duced Cu/m-ZrO

part the different behaviors of this catalyst in the reaction of ethanol

2

kind. The observed equilibrium between Cu and Cu for the re-

0

+

change in the proportion Cu /Cu for copper samples supported

on different zirconia phases can be attributed to several reasons that

involve the following: (i) changes in the average Cu–Cu

coordination number of Cu nanoparticle atoms owing to a high

2

is a speciﬁc characteristic, which can explain in

conversion compared to Cu catalysts supported on am- and t-ZrO

2

phases.

Table 5

Apparent activation energy and pre-exponential factor of kinetics constant, turnover frequencies (TOF) for acetaldehyde and ethyl acetate formation. Reaction conditions for

kinetic constant calculations: W/F = 1.2 min and PEtOH = 37.4 kPa. Reaction conditions for TOF calculations: W/F = 1.2 min, PEtOH = 37.4 kPa, Treac = 473 K.

TOF(N2O) (s^ꢀ¹)

TOF(EXAFS) (s

ꢀ1

)

Catalysts

Apparent E

a

(kJ/mol)

4

AcH

EtOAc

AcH (ꢂ10 )

EtOAc (ꢂ10 )

AcH (ꢂ10 )

EtOAc (ꢂ10 )

Cu/am-ZrO

2

33.9 (±0.3)

38.5 (±1.1)

36.3 (±4.6)

37.5

54.9 (±0.2)

40.3 (±1.4)

30.8 (±3.3)

–

810

640

580

410 [59]

1930 [59]

78

114

55

3.9 [59]

26.0 [59]

217

222

217

–

21

40

21

–

Cu/m-ZrO

Cu/t-ZrO

Cu/SiO -Aq

CuSiO -IE

2

42.5

–

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

13

on the oxide surface of ZrO

2

revealed by the presence of oxygen an-

ions and/or oxygen vacancies might provide sites for nucleation

and growth of metal copper particles. For example, it was found

by measuring the number density of oxygen vacancies and gold

2

particles that the density of oxygen vacancies on TiO (110) de-

creases during gold particle growth [56]. The theoretical results

showed that the adsorption energy of a single gold atom on an oxy-

gen vacancy site is more stable by 0.45 eV than on the stoichiom-

etric surface [56].

0

+

2

The formation of different Cu /Cu ratios for Cu/ZrO samples

might be related to a strong interaction of the copper particles with

the matrix. According to some authors [22], the presence of Cu⁺

species on the surface of supported Cu catalysts is associated with

the dispersion degree of supported phase. They showed that the in-

crease in copper dispersion degree leads to an increase in the en-

ergy of Cu interaction as well as to a decrease in the value of the

Auger parameter,

, the change in

ZrO > Cu/am-ZrO

copper species supported on am-ZrO

a

Cu. Concerning the XPS results listed in Table

Cu is in the following order: Cu/t-ZrO > Cu/m-

. Therefore, the lowest Cu value is detected for

(1848.2 eV), to which corre-

2

a

2

sponds the highest BE value of Cu 2p3/2 (932.9 eV). The shift of the

valence band to a higher binding energy is a result of the reduced

particle size [57]. The small particles or isolated Cu atoms at the

interface between metal and oxide support become electropositive

charged, caused by the electron charge transfer from the metal to

support [57]. The IR of CO adsorption data (Fig. 6) show that the

+

Cu species are dominant over am-ZrO

2

surface. The presence of

+

Cu as dominant species is also supported by the low value of

the metallic copper surface area over am-ZrO , revealed by the

O titration (Table 1). It should be noted that the well-dispersed

copper oxide species is not easily reduced to metal due to the pres-

ence of oxygen vacancy in ZrO

, which protects the interfacial Cu⁺

2

N

2

from further reduction during catalytic reaction [58]. It can be con-

cluded that well-dispersed copper species present in reduced Cu/

2

am-ZrO . The lowest value of XPS atomic Cu/Zr ratio for Cu/am-

ZrO (Table 2) should be related to a high amount of exposed Zr

2

atoms under XPS measurements, due to the highest speciﬁc surface

area of support.

It was concluded [22] that ZrO

2

support strongly stabilizes the

Cu state because these species can be obtained just by annealing

under ultrahigh vacuum, while a reduction treatment (H at

23 K) is necessary for the case of SiO -supported samples, due

+

2

5

2

to the difference in the overall polarization property of the both

oxides; the covalence of Cu–O–Me bond was higher when copper

oxide is deposited over silica. Interactions with an oxide support

can accelerate or decrease the rate of CuO reduction [59]. Quanti-

tative analysis of the variations between the well-dispersed phases

and bulk materials has been done by comparing the changes with

quantum mechanical calculations of the Cu–O–Me bond structure

at the interface in terms of covalence, density of charge distribu-

tion, etc. [60]. The interface between Cu and Cu

by means of density functional theory (DFT), and it was shown that

a continuous chemical bonding across the Cu–Cu O system exists

61].

It is evident from the EXAFS analysis (Table 3) that the zirconia-

2

O was investigated

2

[

supported copper nanoparticles behave very differently from their

bulk forms, due to the electronic and structural effects of materials.

On the surface of copper, it is difﬁcult to avoid the perturbation of

2

oxygen from the OH groups as well as from the bulk ZrO , even if

Fig. 11. Effect of the reaction temperature on the change in total PCD and ratio of

water/PCD for (A) Cu/am-ZrO (B) Cu/m-ZrO (C) Cu/t-ZrO (W/F = 38 min,

EtOH = 37.4 kPa).

the catalyst was reduced at high temperature. Copper has a com-

pletely ﬁlled 3d band at the Fermi level; however, the smaller Cu

particles in a strong interaction with the support surface might

have a partially ﬁlled 3d band with a high density of states at Fermi

level, thereby resulting in a lower electron density. The variation in

the inter-atomic distances, R_C_u_–_C_uand R_C_u_–_O, indicates different

interactions between the copper atoms and between copper and

2

,

2

,

2

.

P

For zirconia-supported Cu samples, the interfaces between Cu

and ZrO could be different in respect to copper particle size and

copper oxide support contact structure. The presence of defects

2

1

4

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

oxygen atoms, respectively (Table 3). This is well visible for Cu spe-

cies supported on am- and t-zirconia, which have smaller nano-

Cu surface may occur only when a signiﬁcant fraction of the titania

surface in Cu/TiO system is still exposed. This suggests that the

role of m-ZrO support in the Cu/m-ZrO system is to provide an

extra charge to copper particles or oxygen to proximity of Cu. This

interpretation implies that relatively large metal Cu particles in Cu/

m-ZrO system distinguish themselves from smaller ones by the

2

presence of surface adatoms, which are quite stable even upon

hydrogen reduction up to 800 K [63].

2

cluster size (4.4 and 2.9 nm for Cu/am-ZrO

respectively, Table 3) compared to that for Cu/m-ZrO

In addition, the presence of oxygen causes a stretching in the

2

and Cu/t-ZrO

2

,

2

(8.7 nm).

0

Cu–Cu distance from 2.54 to 3.03 Å for bulk Cu and Cu

tively. The values of RCu–Cu for Cu/am-ZrO and Cu/t-ZrO

the Cu

shorten compared to that of the bulk Cu

geometrically, it is inferred that the oxygen species linked to cop-

per nanoclusters in Cu/am-ZrO and Cu/t-ZrO is more exposed,

2

O, respec-

2

related to

2

O contribution are 2.67 and 2.70 ± 0.002 Å, values which are

O (3.03 ± 0.004 Å). Thus,

2

The decreased value of the coordination number of Cu–Cu

atoms (Table 3) and the low temperature reduction in copper spe-

cies on t-ZrO (Figs. 3 and 5) consist with the increase in electron

2

being seen in Fig. S5. This kind of oxygen will be called by us ‘‘la-

bel’’ oxygen.

density of the Cu atoms in very small metal particles, which are

dominant on the support surface (Table 2). In general, it is ex-

pected that the fraction of oxygen over small metallic Cu particles

The ‘‘label’’ oxygen weakly linked to the copper nanoparticles

would be readily decomposed on the surface of Cu/t-ZrO

2

system

is higher compared to that on bigger ones [45]. For Cu/am-ZrO

2

ꢀ10

+

under high vacuum treatment (10

Torr), since during the XPS

sample, where the Cu species is dominant, it can be concluded

that a great part of the copper species is in interaction with the

oxygen sites originated from the oxygen-rich interface at Cu/am-

measurements, metallic copper was detected only (Table 2). This

could be caused by some photoreduction in copper oxide species

irradiated by X-ray and under vacuum. On the other hand, con-

cerning the EXAFS analysis, the electron transition of

ZrO

In opposite to that, a small fraction of the copper species should

be covered by oxygen in Cu/t-ZrO , i.e., the positive charged copper

2 2

by oxygen diffusion, similar to that observed for Cu/m-ZrO .

1

s ? (4pz + L) at K Cu-edge typical for bulk Cu

2

O reference is not

2

detected for Cu/t-ZrO ; copper oxide species is totally reduced to

metallic copper (Fig. S1). It should be noted that the photon energy

over the samples during XAS experiments was much higher

2

species is mainly caused by the direct connection with the support

surface. The higher stability of CO desorbed up to a higher temper-

ature on Cu/t-ZrO compared to that on Cu/m-ZrO would be re-

2 2

(

1

9000 eV) than that during the XPS measurements (about

000 eV). Probably, there is no real ionic interaction between the

oxygen and copper, due to the weakly bonded oxygen to copper

surface. The same consideration is observed for Cu/am-ZrO and

, but the copper oxide species (Cu ) is much more stable

on am-ZrO and m-ZrO compared to that on t-ZrO

lated to a covalence character of the Cu–O–Zr interaction due to

the different polarities of both zirconia phases.

Considering that the oxygen is highly negatively charged, it is

reasonable from electrostatic interaction that the higher the cat-

ionic charge on Cu is, the stronger the bonding ability to the O

atom will be. However, theoretical studies have shown [64] that

2

+

Cu/m-ZrO

2

.

Even though many studies, especially theoretical studies, have

been suggested that the copper oxide support interface plays a sig-

niﬁcant role for the creation of oxygen sites, one question arises

how the limited surface area of the copper catalytic system, like

2

the adsorption energy of oxygen for Cu/ZrO (2.25 eV) is higher

compared to those for Cu(111) (1.44 eV) or cus Cu (0.87 eV). This

implies that the location of adsorbed O prefers a high coordination

environment on neutral Cu system where the surface electron den-

sity is high, in line with the strong ionic bonding character of O and

2

Cu/m-ZrO catalyst with the highest particle size (8.7 nm), can be

responsible for the creation of oxygen species, being revealed by

the presence of stable Cu species under reduction up to 800 K.

Cu. Studying the interface between Cu and Cu

shown that a continuous chemical bonding exists across the Cu–

Cu O system, i.e., the copper atoms are simultaneously connected

by a metallic bonding with the copper atoms in both the metal

and the oxide phase, and besides, they form ionic-covalent connec-

tions with the oxygen.

2

O by DFT [61], it was

+

Caballero et al. [23] also reported evidence of high stability of

2

+

Cu oxide phase on ZrO

2

support based on in situ Cu K-edge XANES

measurements under reaction conditions. Sloczynski et al. [34]

found that the reduction in CuO is an autocatalytic consecutive

reaction, and the intermediate product is amorphous Cu

2

O.

The slowly transition of CuO to Cu and the high stability of Cu

species for Cu/m-ZrO during the XANES-TPR analysis (Fig. 5)

0

+

4.2. Relation between the copper species, acid–base properties and

catalytic performance of Cu/ZrO catalysts

2

should be correlated with the strong interaction of the copper sur-

face with oxygen species deposited at the metal-support interface,

as well as on the copper particles. This is in accordance with the

observations by Knapp et al. [45]. The authors reported that the

formation of oxide species can occur on the surface of copper nano-

particles, as well as at the interface between Cu and oxide support.

The oxygen atoms are located to a large extent on the accessible

surface of the copper metal particles, since the fraction of the po-

tential interface between the agglomerated metal particles and

the oxide support is much smaller than the fraction of the oxidized

Cu [45].

The results of catalytic test of ethanol reaction conclude that the

product distribution is strongly inﬂuenced by the structure and

electronic properties of the particles and the surface acid–base

properties of the catalysts as a function of the kind of ZrO

Additionally, ethanol reaction gives an opportunity to obtain infor-

mation about the electronic properties of Cu/ZrO catalysts.

2

phase.

2

It is well known that ethanol is converted via four main groups

of reaction; dehydrogenation, dehydrogenation coupling, dehydra-

tion, and hydrogenolysis. Dehydrogenation of ethanol forms acetal-

dehyde as a primary product. The consecutive aldol condensation of

acetaldehyde yields n-butanol, crotonaldehyde, and ketones. Dehy-

dration reaction leads to formation of diethyl ether and ethylene. It

was reported [65] that the strong acidic sites are responsible for the

dehydration reaction, whereas dehydrogenation requires metallic,

moderate acid sites, and strong basic sites.

The surface oxygen species (O

ways in Cu/ZrO system, especially under reaction conditions,

which will be discussed in detail in Section 4.2. It should be noted

that the oxygen species on the copper particles in Cu/m-ZrO is

caused by the oxygen transfer from m-ZrO substrate to copper

s

) might be formed by different

2

surface, due to the high oxygen-rich interface generated by spill-

over mechanism between the copper and the zirconia during the

reduction process. The high concentration of oxygen anions, char-

The surface and acid–base properties derived from the XPS

analysis and DRIFTS of CO adsorption show appreciable informa-

tion about the surface anionic defects associated with the creation

d+

acterizing the m-ZrO

tion (Fig. 6E). This is similar to the observation for TiO

Cu [62], where it was suggested that the oxygen migration to the

2

surface, is revealed by the FTIR of CO adsorp-

of Zr species on the surface of reduced materials. Three different

2

-supported

types of basic sites are detected over the surface of ZrO

copper catalysts: strongly basic surface O ions, medium-strength

2

-supported

2ꢀ

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

15

Zr⁴⁺–O pairs, and low-strength OH groups. Their relative concen-

tration or density depends on the kind of ZrO phase. The bridged-

bonded OH groups for am- and m-ZrO phases can be easily

2ꢀ

the oxygen resistance on the copper surface, its consumption by

ethanol leads to the increase in acetaldehyde formation (Table 4

and Fig. 9).

2

decomposed under increasing the reaction temperature, and more

It was suggested by Inui et al. [4,5] that the coupling of alde-

hyde with alcohol or ethoxide species for ethyl acetate formation

occurs over the mixed metal oxides surface not over the Cu metal

surface in Cu catalysts supported on Zn–Zr–Al–O system. The

dehydrogenated ethanol or the acetaldehyde was adsorbed on

the acid sites, while the hydrogen of the OH group in ethanol

was abstracted by the surface base sites forming surface ethoxide

[4,5]. Then, a hemiacetal between these products is formed, which

dehydrogenates very fast and produces ethyl acetate. It means that

more active sites, like Cu⁰and Cu , will accelerate the acetal

adsorption.

oxygen anions and/or oxygen vacancies are formed, whereas, cus

d+

4+

Zr and Zr cations attributed to strong acid sites of Lewis type

are stabilized for Cu/t-ZrO

The high rate of acetaldehyde formation for Cu/t-ZrO

Fig. 9) can be related to the presence of ‘‘label’’ oxygen on the

2

.

2

catalyst

(

well-dispersed metallic copper nanoparticles, which facilitates

the formation of ethoxide intermediate via O–H scission in ethanol.

It was shown that the presence of oxygen facilitates the catalyst

dehydrogenation activity [63]. Similar to methanol activation on

oxidized single-crystal Cu surfaces [66], the reactivity of ethanol

to acetaldehyde can be greatly enhanced on the oxygen covered

copper clusters, because the O–H bond scission is facilitated by

the adsorbed oxygen. Surface science studies have shown that

the co-adsorbed oxygen on the surface of transition metals act as

a Brönsted base to facilitate the O–H bond cleavage [63]. Subse-

quent C–H bond cleavage leads to the formation of acetaldehyde.

This step can be also catalyzed by the presence of surface oxygen.

The surface ethoxide intermediate can either loss hydrogen to pro-

duce acetaldehyde or pick up hydrogen to form ethanol and then

decomposes to acetaldehyde. The dependence of the rate of alcohol

dehydrogenation on the oxygen coverage may be interpreted in

terms of Langmuir type dependence on the oxygen surface cover-

age [67].

+

The relatively invariable value of the water/PCD ratio for Cu/

am-ZrO suggests that the consumed hydroxyls in the processes

2

of acetaldehyde condensation and ethanol dehydration are in equi-

librium with the regeneration of OH sites and/or oxygen sites of

am-ZrO

the oxygen species at ZrO

on the surface of am-ZrO

As was mentioned above, a great part of the copper surface in

Cu/m-ZrO is in interaction with the oxygen being revealed by

2

at all reaction temperatures. It can be concluded that

2

tunes the electron density of Cu species

2

.

2

+

the high stability of Cu species, which are in some equilibrium

with Cu . The presence of Cu species with different electron densi-

ties and the high concentration of basic sites (O ) deﬁned by

DRIFTS are responsible for the high selectivity to ethyl acetate of

0

2

ꢀ

The catalytic performance of Cu/t-ZrO

the presence of strong Lewis acid sites (Zr ) and by the low basic

sites density deﬁned by the dominant Cu . The b-aldolization of

2

4

is also determined by

Cu/m-ZrO

tive copper species over Cu/m-ZrO catalyst differs from copper

2

species of Cu/t-ZrO and Cu/am-ZrO (Table 5).

2

. TOFs of ethyl acetate formation highlight that the ac-

+

2

0

2

acetaldehyde is strongly promoted, since the aldolization reaction

It should be noted that the lowest apparent rate of acetaldehyde

formation for Cu/m-ZrO catalyst at short contact time (Table 5 and

requires basic sites to abstract

a-hydrogen atom and strong Lewis

2

acid sites to bind the two molecules of acetaldehyde [68]. Acetal-

dehyde condensation toward n-butanol and crotonaldehyde is a

bimolecular reaction between adjacent adsorbed acetaldehyde

species that requires acid–base pair sites [58], which is a character-

Fig. 9) could be masked by its fast consumption for ethyl acetate

formation. In reality, the acetaldehyde takes part in the formation

of ethyl acetate formation, which means that its participation in

the reaction could be revealed by spillover and coupling reaction

between the ethoxide and acyl adsorbates. The acetaldehyde co-

feed with ethanol leads to the formation of products of aldol con-

densation [4,71], suggesting that the ethyl acetate reaction pro-

ceeds via Langmuir–Hinshelwood mechanism [72]. The diffusion

rate of adsorbed acetaldehyde should be determined by the elec-

istic of t-ZrO

2

surface.

catalyst, the selectivity to methyl ethyl ketone

For Cu/t-ZrO

2

and crotonaldehyde decreases with increasing the reaction tem-

perature at high contact time (Fig. 8). In general, the aldolization

process is accompanied by a high amount of water released under

reaction conditions (Table 4). The remarkably amounts of PCD for

tronic properties of copper nanoclusters in interface with m-ZrO

2

.

Cu/t-ZrO

water. Cu/t-ZrO

tios compared to those for Cu/am-ZrO

This suggests some regeneration of the oxygen sites by consump-

2

do not correspond to the respective amounts of released

catalyst is characterized by lower water/PCD ra-

and Cu/m-ZrO (Fig. 11).

The Cu catalysts supported on am- and t-ZrO exhibit a rela-

tively high stability of the selectivity to ethyl acetate formation

at all reaction temperatures. A certain decrease in the ethyl acetate

2

formation at temperature above 513 K is observed for Cu/m-ZrO

2

tion of the hydroxyls of t-ZrO

vations in Ref. [69].

2

under reaction, similar to the obser-

catalyst at high contact time (Figs. 7B and 10). The latter is accom-

panied with a strong increase in the selectivity to propanone for-

mation on Cu/m-ZrO at 525–550 K, being is seen Fig. 10. The

relationship between the propanone formation and the loss of

The surface oxygen linked to copper nanoclusters in Cu/am-

ZrO affects strongly the electron density of copper species. The

dominant Cu species and the low basic sites density determined

by the low strength of OH groups of am-ZrO seem to be responsi-

2

+

ethyl acetate selectivity of Cu/m-ZrO

gen mobility may be involved in the performance of the catalyst.

It was shown [73] that CuO/CeO exhibits a high capacity for

2

seems to indicate that oxy-

2

ble for the high selectivity to acetaldehyde (Table 4, Fig. 7B and C).

It has been shown [70] that isolated copper ions catalyze the dehy-

drogenation of ethanol to acetaldehyde formation, while multiple

Cu ensembles show high yields of ethyl acetate. The absence of

strong Lewis acid sites does not provoke the strong acetaldehyde

condensation similar to that observed for Cu/t-ZrO

the temperature of reaction, probably, causes a thermal decompo-

sition of the hydroxyls at am-ZrO that leads to the formation of

2

propanone formation in the ethanol reaction, where ceria is an

oxygen supplier due to its high oxygen storage capacity. However,

the present reaction over Cu/ZrO

tions employing anhydrous ethanol as reactant and in reductive

atmosphere. Similar to CeO , bulk m-ZrO is also characterized by

2

occurs in non-oxidative condi-

2

. Increasing

2

oxygen-rich surface. The increase in the reaction temperature

above 513 K could cause a loss of the oxygen from the bulk m-

2

more oxygen anions or oxygen vacancies and to the change in

the electron density of copper species caused by the oxygen migra-

tion. The last provokes the ethyl acetate formation, being revealed

by the increase in the rate of ethyl acetate formation at tempera-

ture P525 K (Figs. 7B and 9), as well as by the lower apparent acti-

vation energy value in Table 5. On the other hand, depending on

ZrO

species active for ethyl acetate formation. Propanone can be

formed by an oxidation of acetaldol with the surface oxygen (O

adsorbed on Cu/ZnO/Al catalyst [74]. The oxygen consumed

during the acetaldol oxidation could be regenerated by the water

decomposition: H O ? H + O or by the CO dissociation, pro-

2

, leading to a change in the electron density of surface copper

s

)

2 3

O

2

s

2

1

6

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

2

Fig. 12. Schematic illustration of the proposed mechanisms for the formation of atomic oxygen on the surface of copper and of m-ZrO .

duced during propanone formation: CO

according to Nakamura et al. [75], the possibility for CO

tion on the surface of Cu(110) under the present reaction condi-

tions should be very small. Also, the direct dissociation of H O is

2

? CO + O

s

[74]. However,

tion is determined by the high oxygen mobility from the bulk m-

ZrO phase to copper species causing a high density of basic sites

and a more heterogeneous distribution of the surface copper spe-

cies (Cu /Cu ). The increase in reaction temperature at short con-

tact time of 1.2 min leads to the increase in the rate of ethyl

2

dissocia-

2

0

+

2

kinetically rather difﬁcult at the Cu sites. A coupling of OH groups

by proton-transfer channel according to Henderson [76] could be

predominantly responsible for the surface oxygen formation on

acetate formation over Cu/am-ZrO

gen transfer from Cu to bulk am-ZrO

2

catalyst, provoked by the oxy-

and establishment of oxygen

2

ꢄ

2

Cu/m-ZrO

2

catalyst: 2OH ? O + H

O

(Fig. 12). DFT studies of

vacancies, which subsequently are diffused to copper species by re-

dox mechanism. The last one provokes the highest rate of ethyl

copper surface also showed that the disproportion of OH groups

is a preferable way for the water dissociation [64].

acetate formation over Cu/am-ZrO

P525 K. The high density of Lewis acid sites deﬁned by cus Zr

centers and the predominance of superﬁcial Cu⁰species over t-

ZrO leads to the formation of by-products by aldol condensation.

2

at reaction temperature

4+

2

The slight decrease in ethyl acetate formation for Cu/m-ZrO at

temperature range 498–523 K is accompanied with a strong in-

crease in the water/PCD ratio from 3.2 to 5 independent of the in-

crease in the products of the aldol condensation and ethanol

dehydration (Figs. 10 and 11B). Probably, the coupling mechanism

of OH groups on the copper surface at that temperature range is

responsible for the formation of water and adsorbed oxygen on

Cu/m-ZrO , as well as for the enhanced propanone formation.

2

The readily consumption of the oxygen for ethyl acetate and prop-

anone formation would cause a loss of the electron transfer from

2

Acknowledgments

The authors gratefully acknowledge by support ﬁnance for this

research from CNPq and FAPESP and the Brazilian Synchrotron

Light Laboratory (LNLS) for the XAS beamline experiments

(

D06A-DXAS 10156) and (D04B-XAFS1). S.D. is gratefully acknowl-

2

the bulk m-ZrO justifying the ethyl acetate selectivity decrease.

The subsequently increase in the reaction temperature from 523

up to 548 K leads to a decrease in the water/PCD ratio value from

edged to Bulgarian Ministry of Education, Science and Youth (Pro-

ject DTK-02/36).

5

to 3, as well as of ethyl acetate formation, which suggests that

Appendix A. Supplementary material

other mechanism is involved in the formation of adsorbed oxygen.

In this case, the surface oxygen can be formed by the interaction of

the hydroxyls, near Zr cations, with a hydrogen anion at cus Zr

4

+

3+

Supplementary data associated with this article can be found, in

the online version, at http://dx.doi.org/10.1016/j.jcat.2013.06.022 .

ꢀ

2ꢀ

[

77], acting as a Lewis base: OH + H ? H

2

" + O (Fig. 12). This

leads to the increase in propanone formation, since acetaldol also

reacts with the oxygen generated by this mechanism on the sup-

port surface. In Fig. 12, it is attempted to illustrate the proposed

mechanism for the formation of oxygen on the surface of Cu nano-

particles and of m-ZrO under ethanol reaction at reaction temper-

2

ature range of 498–548 K.

References

[

1] V.D. Bastos, Etanol Alcoholquimica e biorreﬁnarias, BNDES Set. 25 (2007) 5–38 .

2] P. Chladek, E. Croiset, W. Epling, R.R. Hudgins, Can. J. Chem. Eng. 85 (2007)

917–924 .

[

3] A.B. Sánchez, N. Homs, J.L.G. Fierro, P.R.d. La-Piscina, Catal. Today 107–108

(

2005) 431–435 .

It should be noted that the catalysts showed high ethanol con-

version stability during 12 h of work at all reaction temperatures

4] K. Inui, T. Kurabayashi, S. Sato, N. Ichikawa, J. Mol. Catal. A: Chem. 216 (2004)

147–156 .

[

5] K. Inui, T. Kurabayashi, S. Sato, Appl. Catal. A: Gen. 237 (2002) 53–61 .

6] S.W. Colley, J. Tabatabaei, K.C. Waugh, M.A. Wood, J. Catal. 236 (2005) 21–33 .

7] A.B. Gaspar, A.M.L. Esteves, F.M.T. Mendes, F.G. Barbosa, L.G. Appel, Appl. Catal.

A: Gen. 363 (2009) 109–114 .

2

(Fig. S6). It can be summarized that am- and m-ZrO phases are

suitable carriers for supported Cu catalysts for ethyl acetate

formation.

[

8] L. Wang, W. Zhu, D. Zheng, X. Yu, J. Cui, M. Jia, W. Zhang, Z. Wang, Reac. Kinet.

Mech. Catal. 101 (2010) 365–375 .

9] N. Iwasa, N. Takezawa, Bull. Chem. Soc. Jpn. 64 (1991) 2619–2623 .

5

. Conclusions

[

10] J. Agrell, H. Birgersson, M. Boutonnet, I. Melián-Cabrera, R.M. Navarro, J.L.G.

Fierro, J. Catal. 219 (2003) 389–403 .

Copper catalysts supported on different ZrO

2

polymorphs are

[11] M. Shimokawabe, H. Asakawa, N. Takezawa, Appl. Catal. 59 (1990) 45–58 .

[12] F.W. Chen, H.C. Yang, L.S. Roselin, W.Y. Kuo, Appl. Catal. A: Gen. 304 (2006) 30–

found to be active catalysts for ethanol dehydrogenation to acetal-

dehyde or ethyl acetate depending on the kind of zirconia phase.

The catalytic properties of Cu catalyst supported on am-, m-, or t-

39 .

[

13] K.T. Jung, A.T. Bell, Catal. Lett. 80 (2002) 63–68 .

14] M.D. Rhodes, A.T. Bell, J. Catal. 233 (2005) 198–209 .

ZrO

2

phase are due the speciﬁc electron density of supported cop-

[15] R.A. Koeppel, A. Baiker, A. Wokaun, Appl. Catal. A: Gen. 84 (1992) 77–102 .

[16] X.R. Zhang, L.C. Wang, C.Z. Yao, Y. Cao, W.L. Dai, H.Y. He, K.N. Fan, Catal. Lett.

0

+

per species (Cu and Cu ) deﬁned by the particle size and the inter-

face at copper metal oxide support. The active surface sites for Cu/

m-ZrO catalyst show that the higher rate of ethyl acetate forma-

2

102 (2005) 183–190 .

[

17] K. Takeshita, S. Nakamura, K. Kawamoto, Bull. Chem. Soc. Jpn. 51 (1978) 2622–

2627 .

A.G. Sato et al. / Journal of Catalysis 307 (2013) 1–17

17

[

18] A.B. Gaspar, F.G. Barbosa, S. Letichevsky, L.G. Appel, Appl. Catal. A: Gen. 380

2010) 113–117 .

19] H. Oguchi, H. Kanai, K. Utani, Y. Matsumura, S. Imamura, Appl. Catal. A: Gen.

93 (2005) 64–70 .

20] I. Ritzkopf, S. Vukojevic, C. Weidenthaler, J.-D. Grunwaldt, F. Schüth, Appl.

Catal. A: Gen. 302 (2006) 215–223 .

21] G. Águila, S. Guerrero, P. Araya, Catal. Commun. 9 (2008) 2550–2554 .

22] J.P. Espinós, J. Morales, A. Barranco, A. Caballero, J.P. Holgado, A.R. González-

Elipe, J. Phys. Chem. B 106 (2002) 6921–6929 .

23] A. Caballero, J.J. Morales, A.M. Cordon, J.P. Holgado, J.P. Espinos, A.R. Gonzalez-

Elipe, J. Catal. 235 (2005) 295–301 .

[47] D.F. Cox, K.H. Schulz, Surf. Sci. 249 (1991) 138–148 .

[48] D. Bianchi, T. Chaﬁk, M. Khalfallah, S.J. Teichner, Appl. Catal. A: Gen. 112 (1994)

57–73 .

(

2

[49] K.T. Jung, A.T. Bell, J. Mol. Catal. 163 (2000) 27–42 .

[50] G. Cerrato, S. Bordiga, S. Barbera, C. Morterra, Surf. Sci. 377–379 (1997) 50–55.

[51] K. Pokrovsky, K.T. Jung, A.T. Bell, Langmuir 17 (2001) 4297–4303 .

[52] B. Bachiller-Baeza, I. Rodrigues-Ramo, A. Guerrero-Ruiz, Langmuir 14 (1998)

3556–3564 .

[53] B.A. Raich, H.C. Fohy, Ind. Eng. Chem. Res. 37 (1998) 3888–3895 .

[54] A. Barranco, F. Yubero, J.A. Mejias, J.P. Espinos, A.R. Gonzalez-Elipe, Surf. Sci.

482–485 (2001) 680–686 .

[

24] G.K. Chuah, S.H. Liu, S. Jaenicke, J. Li, Microporous Mesoporous Mater. 39

[55] J.A. Mejias, V.M. Jimenez, G. Lassaleta, A. Fernandez, J.P. Espinos, A.R. Gonzalez-

Elipe, J. Phys. Chem. 100 (1996) 16255–16262 .

(

2000) 381–392 .

[

25] K.J. Soerensen, N.W. Cant, Catal. Lett. 33 (1995) 117–125 .

26] Y. Okamoto, H. Gotoh, H. Aritani, T. Tanaka, S. Yoshida, J. Chem. Soc., Faraday

Trans. 93 (1997) 3879–3885 .

[56] E. Wahlstrom, N. Lopez, R. Schaub, P. Thostrup, A. Ronnau, C. Africh, E.

Laegsgaard, J.K. Norskov, F. Besenbacher, Phys. Rev. Lett. 90 (2003) 026101–

026104 .

[

27] A. Frenkel, Z. Krystallogr. 222 (2007) 605–611 .

[57] M. Rønning, F. Huber, H. Meland, H. Venvik, D. Chen, A. Holmen, Catal. Today

100 (2005) 249–254 .

[58] W.P. Dow, T.J. Huang, J. Catal. 160 (1996) 171–182 .

[59] J.A. Rodriguez, J.Y. Kim, J.C. Hanson, M. Pérez, A.I. Frenkel, Catal. Lett. 85 (2003)

247–254 .

[60] W. Grüner, N.W. Hayes, R.W. Joyner, E.S. Shpiro, H. Raﬁq, M. Siddiqui, G.N.

Baeva, J. Phys. Chem. 98 (1994) 10832–10846 .

[61] O.L. Blajiev, A. Hubin, Electrochim. Acta 50 (2005) 4297–4307 .

[62] K. Varazo, F.W. Parsons, S. Ma, D.A. Chen, J. Phys. Chem. B 108 (2004) 18274–

18283 .

[63] K.Y. Guan, E.J.M. Hensen, Appl. Catal. A: Gen. 361 (2009) 49–56.

[64] Q.L. Tang, Z.P. Liu, J. Phys. Chem. C 114 (2010) 8423–8430 .

[65] M.E. Manriquez, T. Lopez, R. Gomez, J. Navarrete, J. Mol. Catal. A: Gen. 220

(2004) 229–237 .

[66] J.N. Russel, S.M. Gates, J.T. Yates, Surf. Sci. 163 (1985) 516–540 .

[67] B.A. Sexton, A.E. Hughes, N.R. Avery, Surf. Sci. 155 (1985) 366–386 .

[68] A. Yee, S.J. Morrison, H. Idriss, Catal. Today 63 (2000) 327–335 .

[69] K.G. Azamm, I.V. Babich, K. Seshan, L. Lefferts, J. Catal. 251 (2007) 153–162 .

[70] J.L. Gole, M.G. White, J. Catal. 204 (2001) 249–252 .

[71] A.G. Sato, P. Diogo Volanti, I.C. de Freitas, E. Longo, J.M.C. Bueno, Catal.

Commun. 26 (2012) 122–126.

[72] I. Chorkendorff, J.W. Niemantsverdriet, Concepts of Modern Catalysis and

452, ISBN: 3-527-30574-2.

[73] T. Nishiguchi, T. Matsumoto, H. Kanai, K. Utani, Y. Matsumura, W.J. Shen, S.

Imamura, Appl. Catal. A: Gen. 279 (2005) 273–277 .

28] S. Ardizzone, C.L. Bianchi, J. Electroanal. Chem. 465 (1999) 136–141 .

29] S. Velu, K. Suzuki, C.S. Gopinath, H. Yoshida, T. Hattori, Phys. Chem. Chem.

Phys. 4 (2002) 1990–1999 .

30] S. Ardizzone, C.L. Bianchi, Surf. Interf. Anal. 30 (2000) 77–80 .

31] M.K. Dongare, A.M. Dongare, V.B. Tare, E. Kemniz, Solid State Ionics 455 (2002)

[

1

52–153 .

32] J.A. Navio, M.C. Hidalgo, G. Colon, S.G. Botta, M.I. Letter, Langmuir 17 (2001)

02–210 .

33] C. Morant, J.M. Sanz, L. Galan, L. Soriano, F. Rueda, Surf. Sci. 218 (1989) 331–

45 .

[

2

3

34] J. Sloczynski, R. Grabowski, A. Kozlowska, P.K. Olszewski, J. Phys. Chem. Chem.

Phys. 5 (2003) 4631–4640 .

35] C.Z. Yao, L.C. Wang, Y.M. Liu, G.S. Wu, Y. Cao, W.L. Dai, H.Y. He, K.N. Fan, Appl.

Catal. A: Gen. 297 (2006) 151–158 .

36] K. Shimizu, H. Maeshima, H. Yoshida, A. Satsuma, T. Hattori, Phys. Chem.

Chem. Phys. 3 (2001) 862–866 .

37] D.M. Pickup, I. Ahmed, V. FitzGerald, R.M. Moss, K.M. Wetherall, J.C. Knowles,

M.E. Smith, R.J. Newport, J. Non-Cryst. Solids 352 (2006) 3080–3087 .

38] L.S. Kau, D.J. Spira-Solomon, J.E. Penner-Hahn, K.O. Hodgson, E.I. Solomon, J.

Am. Chem. Soc. 109 (1987) 6433–6442 .

39] M.H. Groothaert, J.A. van Bokhoven, A.A. Battiston, B.M. Weckhuysen, R.A.

Schoonheydt, J. Am. Chem. Soc. 125 (2003) 7629–7640 .

40] L.A. Grunes, Phys. Rev. B 27 (1983) 2111–2131 .

41] Q. Wang, J.C. Hanson, A.I. Frenkel, J. Chem. Phys. 129 (2008) 234502–234507 .

42] Y. Okamoto, T. Kubota, H. Gotoh, Y. Ohto, H. Aritani, T. Tanaka, S. Yoshida, J.

Chem. Soc., Faraday Trans. 94 (1998) 3743–3752 .

[

[74] D.J. Elliot, F. Pennella, J. Catal. 119 (1989) 359–367 .

[75] J. Nakamura, J.A. Rodriguez, C.T. Campbell, Does J. Phys.: Condens. Matter 1

(1989) SB149–SB160 .

[76] M.A. Henderson, Surf. Sci. Rep. 46 (2002) 1–308 .

[77] A. Trunschke, D.L. Hoang, H. Lieske, J. Chem. Soc., Faraday Trans. 91 (1995)

4441–4444 .

[

43] F. Yang, Y. Choi, P. Liu, J. Hrbek, J.A. Rodriguez, J. Phys. Chem. C 114 (2010)

1

7042–17050 .

44] J. Engeldinger, C. Domke, M. Richter, U. Bentrup, Appl. Catal. A: Gen. 382

2010) 303–311 .

(

[

45] R. Knapp, S.A. Wyrzgol, A. Jentys, J.A. Lercher, J. Catal. 276 (2010) 280–291 .

46] A. Dandekar, M.A. Vannice, J. Catal. 178 (1998) 621–639.

Article Doi

Effect of the ZrO2 phase on the structure and behavior of supported Cu catalysts for ethanol conversion

DOI: 10.1016/j.jcat.2013.06.022

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.jcat.2013.06.022

Full text of DOI:10.1016/j.jcat.2013.06.022

Products guided by the article

Relevant to this article

Hot Product