You are on page 1of 11

Catalysis Today 254 (2015) 7282

Contents lists available at ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

AuCu on Nb2 O5 and Nb/MCF supports Surface properties and


catalytic activity in glycerol and methanol oxidation
Izabela Sobczak , ukasz Wolski
A. Mickiewicz University, Faculty of Chemistry, Umultowska 89b, 61-614 Poznan, Poland

a r t i c l e

i n f o

Article history:
Received 3 September 2014
Received in revised form 28 October 2014
Accepted 30 October 2014
Available online 23 December 2014
Keywords:
Nb2 O5
Nb/MCF
AuCu system
Glycerol and methanol oxidation

a b s t r a c t
Nb2 O5 and Nb/MCF were used as supports for gold and copper (AuCu system). The activity of these
catalysts in the oxidation of glycerol in the liquid phase and methanol in the gas phase was compared
with that of monometallic gold catalysts. The oxidation states of metals in the samples prepared were
characterised by XRD, TEM, UVVis, XPS and H2 -TPR. It was found that gold was present in metallic form,
whereas copper in the form of CuO, oligonuclear [Cu+ O Cu+ ]n clusters and isolated cations. The
introduction of copper changed the electronic state of gold in AuNb2 O5 and AuNb/MCF and the reduction of copper was found easier in bimetallic AuCu catalysts, which was interpreted as a result of the
interaction between gold and copper. Nb2 O5 supported catalysts were active in the glycerol oxidation.
Monometallic catalysts show high selectivity to glyceric acid, whereas an increase in the selectivity to
glycolic acid was found over bimetallic samples due to the presence of copper species active in the oxidative dehydrogenation of glyceric acid. AuCu catalysts exhibit very high activity in methanol oxidation
towards CO2 as a result of the presence of basic oxygen from CuO-like species on the catalyst surface.
Superior performance of bimetallic catalysts in oxidation reactions is due to the synergy between gold
and copper.
2014 Elsevier B.V. All rights reserved.

1. Introduction
In recent years, in the eld of heterogeneous catalysis increasing
attention has been paid to noble metal-supported catalysts because
of their potential applications in industrially important chemical
reactions [13]. Gold and bimetallic systems containing gold are
nowadays the focus of attention [1,2,4,5]. Among different bimetallic nanoparticle catalysts (like AuPt, AuPd, AuAg, AuCu systems),
AuCu modied catalysts have recently been tested in catalytic oxidation reactions [2,4]. In the AuCu bimetallic catalysts, copper is
much more prone to oxidation than gold and phase segregation
tends to occur upon treatment in oxidising atmosphere (in either
reaction or pretreatment) [2]. As a consequence, copper is enriched
on the surface and may form copperoxygen species and therefore
it can act as a promoter to provide reactive oxygen in a variety of
gold catalysed oxidation reactions. Supported goldcopper catalysts have shown promising catalytic performance in oxidation of
CO [6,7], propene [8,9] and alcohols [10,11].
Selective oxidation of alcohols to the corresponding aldehydes,
ketones and acids involving the use of molecular oxygen as an

Corresponding author. Tel.: +48 61 8291305.


E-mail address: sobiza@amu.edu.pl (I. Sobczak).
http://dx.doi.org/10.1016/j.cattod.2014.10.051
0920-5861/ 2014 Elsevier B.V. All rights reserved.

environmentally friendly oxidant is one of the most important


tasks of the ne chemicals industry [12]. Selective oxidation of benzyl alcohol to benzaldehyde with oxygen, catalysed by bimetallic
AuCu/SiO2 catalysts, has been reported in the literature [10,11].
The results have shown that AuCu bimetallic catalyst exhibited
high activity and selectivity to benzaldehyde (96% of benzyl alcohol conversion and 100% of selectivity at 353 K for Au/Cu molar
ratio = 4.5/1) due to a strong synergistic effect between Au and
Cu that arises from electronic interaction. The AuCu/SiO2 alloy
catalysts show synergistic effect not only for the benzyl alcohol
oxidation, but also for oxidation of a variety of structurally diverse
alcohols such as cinnamyl alcohol, cyclohexanol, 1-octanol, 1butanol and 3-phenyl-1-propanol [11]. Gold and copper supported
on silica exhibited much improved conversion and selectivity to
aldehydes than the monometallic Au/SiO2 catalyst.
As shown elsewhere [13,14], AuCu modied MCM-22 and
CeO2 /SBA-15 catalysts were highly active in total oxidation of
methanol in the temperature range 423523 K. It was shown that
the presence of gold on AuCu-MCM-22 surface facilitates the reduction of oxidative CuO-like species and increases the mobility of
oxygen, promoting total oxidation of methanol. Cerium species on
the surface of SBA-15 interact with gold, whereas CuAu interaction leads to the electron transfer from Cu+ to Au0 , enhancing the
redox properties of the system and increasing the total oxidation

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

activity. AuCu-mesoporous sieves appeared useful for the removal


of methanol (by total oxidation) from gases emitted from automotive devices and during a variety of industrial operations.
Recently, the oxidation of glycerol for the production of valuable
oxygenated derivatives (e.g. glyceric, tartronic, glycolic, hydroxypyruvic acids, dihydroxyacetone) has attracted much attention
[3,1517]. Selective oxidation of glycerol with oxygen has been
tested mainly with carbon supported monometallic Au/C and
bimetallic AuPt/C and AuPd/C catalysts that showed much higher
activity in the liquid phase oxidation of glycerol than metal oxides
(e.g. TiO2 , CeO2 , MgO) supported catalysts. Nevertheless, our previous studies indicated [18] that Au/Nb2 O5 catalyst in the crystalline
form is only slightly less active than Au/C, but more selective to
glyceric acid due to strong interaction between gold and niobium
in the support. In the eld of bimetallic systems it has been found
that the use of a combination of gold and palladium or platinum
as supported alloy nanocrystals resulted in a signicant enhancement in the catalyst activity to the desired product [15]. It has been
proved that both electronic and geometrical effects are involved
in a strong positive synergetic effect between gold and the second
noble metal [e.g. 19]. Moreover, it has been found that the distribution of products could be controlled by using AuPd or AuPt catalysts
[20]. AuPd catalysts showed in general better selectivity to glyceric
acid than AuPt catalysts, with Pd mainly to promote the formation of tartronic acid and Pt to glycolic acid. The overall selectivity
to glyceric acid increased when using bimetallic AuPd/C catalysts
than monometallic.
The idea of this work was to apply Nb2 O5 as well as Nb/MCF
mesostructured cellular foams as new supports for goldcopper
active phase and to investigate the effect of copper on the properties, activity and selectivity of gold catalysts in the oxidation
of glycerol and methanol (AuCu vs Au systems). It is known that
niobium compounds exhibit special properties like strong metal
support interaction (SMSI) or unique reversible interaction with
several reagents which are very important for the design of catalysts [21]. Moreover, Nb2 O5 itself is an active catalyst in oxidation
reactions [21]. That is why our interest was also to study the
inuence of metalsupport interaction depending on the type of
niobium-containing support (bulk Nb2 O5 and Nb/MCF with Nb dispersed on the surface) on the activity and selectivity in alcohols
(glycerol and methanol) oxidation.
2. Experimental
2.1. Synthesis of MCF material
MCF materials were prepared by a one-pot synthesis according
to [22]. At rst, Pluronic 123 (Poly(ethylene glycol)-blockPoly(propylene glycol)-block-Poly(ethylene glycol)-block) (8 g,
1.4 mmol) was dissolved in 300 g of 1.6 M HCl solution at 308313 K.
Then 1,3,5-trimethylbenzene (Aldrich) (8 g, 66.56 mmol) and NH4 F
(Aldrich) (0.0934 g, 2.52 mmol) were added under vigorous stirring.
Following 1 h of stirring, TEOS (Fluka) (17.054 g, 81.99 mmol) was
added. The nal mixture was stirred at 308313 K for 24 h and then
transferred into a polypropylene bottle and heated at 373 K under
static conditions for 24 h. The solid product was recovered by ltration, washed with distilled water and dried at room temperature.
The template was removed from the as-synthesized material by
calcination at 773 K for 8 h under static conditions.
2.2. Post-synthesis modication of MCF with niobium
For the preparation of Nb/MCF, a portion of MCF material was
treated by incipient wetness impregnation with an aqueous solution of C4 H4 NNbO9 complex (Aldrich). The amount of C4 H4 NNbO9

73

used for the impregnation was calculated to achieve the loading


of niobium equal to 1 wt%. The impregnated Nb/MCF was dried at
353 K and then calcined at 773 K for 4 h.
2.3. Functionalisation of Nb/MCF and Nb2 O5 with organosilane
Nb/MCF and Nb2 O5 (CBMM-Brazil, calcined at 673 K) were
grafted with 3-aminopropyl-trimethoxysilane (APMS) (Aldrich) in
order to functionalise the support before gold and copper modication. The grafting procedure was carried out according to [23] and
was as follows: 5 g of the support powder were reuxed in a dry
toluene solution (200 mL) containing 12.5 mL of APMS at 373 K for
18 h. The catalysts (NH2 -Nb/MCF and NH2 -Nb2 O5 ) were recovered
by ltration followed by washing in dry toluene (200 ml), water
(100 ml) and acetonitrile (20 ml). The powder was dried in an oven
at 353 K.
2.4. Modication of NH2 -MCF and NH2 -Nb2 O5 samples with gold
and copper
The functionalised NH2 -Nb/MCF and NH2 -Nb2 O5 materials were stirred for 1 h in aqua solution of chloroauric acid
(HAuCl4 4H2 O Johnson Matthey UK-USA) (2 wt% of Au as
assumed) (the modied procedure proposed by Mou et al. [24]).
After ltration and washing the recovered solid was stirred with
40 ml of 0.1 M NaBH4 (Aldrich) solution used for reduction of gold.
After 20 min, the solid was recovered by ltration and washed with
water.
For the preparation of AuCu catalysts, in the second step, Cu
was deposited on the Au-modied NH2 -Nb/MCF and NH2 -Nb2 O5
using Cu(NO3 )2 (Aldrich) as the precursor (1 wt% of Cu as assumed)
with the same procedure as that for Au deposition in the rst step.
The mono- and bimetallic samples were obtained after drying at
373 K and calcination at 773 K for 4 h. The gold and copper content
was analysed by ICP-MS.
The catalysts obtained were labelled as AuNb/MCF and
AuNb2 O5 , AuCuNb/MCF and AuCuNb2 O5 and CuNb/MCF and
CuNb2 O5 .
2.5. Catalysts characterisation
The materials prepared were characterised using XRD, N2
adsorption/desorption, TEM, UVVis, XPS, H2 -TPR and test reaction
2-propanol decomposition.
The XRD patterns were obtained on a D8 Advance diffractometer
(Bruker) using CuK radiation ( = 0.154 nm), with a step size of
0.05 in the 2 = 660 range.
The N2 adsorption/desorption isotherms were obtained at 77 K
using a Quantachrome Autosorb iQ apparatus. The samples were
pre-treated in situ under vacuum at 573 K. The surface area was
calculated by the BET method, whereas the volume and the diameter of mesopores were estimated according to Broekhoffde Boer
method with the FrenkelHalseyHill equation (BdB FHH) for MCF
materials [25] and BJH method for Nb2 O5 oxides.
UVVis spectra were recorded using a Varian-Cary 300 Scan
UVVisible spectrophotometer. Powdered samples were placed in
a cell equipped with a quartz window. The spectra were recorded in
the range from 800 to 190 nm. Spectralon was used as the reference
material.
For transmission electron microscopy (TEM) measurements the
powders were deposited on a grid covered with a holey carbon
lm and transferred to JEOL 2000 electron microscope operating at
80 kV.
X-ray photoelectron spectroscopy (XPS) was performed on an
Ultra-high vacuum photoelectron spectrometer based on Phoibos
150 NAP analyzer (Specs, Germany). The analysis chamber was

74

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

operated under vacuum with a pressure close to 5 109 mbar and


the sample was irradiated with a monochromatic Al K (1486.6 eV)
radiation (15 kV; 10 mA). Binding energies were referenced to the
C1s peak from the carbon surface deposit at 284.6 eV.
The temperature-programmed reduction (TPR) of the samples was carried out using H2 /Ar (10 vol.%) as a reducing agent
(ow rate = 40 cm3 min1 ). A sample (0.02 g of Nb/MCF or 0.04 g of
Nb2 O5 ) was packed in a quartz tube, treated in a ow of helium at
623 K for 2 h and cooled to room temperature (RT). Then the sample was heated at a rate of 10 K min1 to 1273 K in the presence
of a reducing mixture. Hydrogen consumption was measured by a
thermal conductivity (TCD) detector.
The 2-propanol conversion (dehydration and dehydrogenation) was performed, using a microcatalytic pulse reactor inserted
between the sample inlet and the column of a CHROM-5 chromatograph. The catalyst bed (0.02 g for Nb/MCF and 0.05 g for
Nb2 O5 ) was rst activated at 623 K for 2 h under helium ow
(40 cm3 min1 ). The 2-propanol (Aldrich) conversion was studied at 423, 473, 523 and 573 K using 3 l pulses of alcohol
under helium ow (40 cm3 min1 ). The reactant and the reaction
products: propene, 2-propanone (acetone) and diisopropyl ether
were analysed using CHROM-5 gas chromatograph on line with
microreactor. The reaction mixture was separated on 2 m column
lled with Carbowax 400 (80100 mesh) at 338 K in helium ow
(40 cm3 min1 ) and detected by TCD.
2.6. Glycerol oxidation
The glycerol oxidation experiments were performed in a 50 cm3
batch reactor from Parr. The oxidation reactions were carried out
with oxygen under pressure of 6 atm, at 333 and 363 K. NaOH
(NaOH/glycerol molar ratio = 2) and 0.05 g of gold catalyst were
added to a 1 M aqueous solution of glycerol. The reaction mixture was stirred for 5 h. For selected samples the reaction was
performed without NaOH in the reaction mixture. Quantitative
analyses of the reaction mixtures were made by high performance liquid chromatography (HPLC). The analysis was carried
out using HPLC chromatograph (Waters) equipped with ultraviolet (UV) and refractive index (RI) detectors. The reactant and the
products were separated on an ion exclusion column (IC-Pak Ion
Exclusion 7.8 300 mm Waters) heated at 343 K. The eluent was
a solution of H2 SO4 (0.0004 M). The samples were taken at the end
of the reactions, 1 ml of the sample was diluted in 10 ml of water
and 5 l of this solution was analysed.
2.7. Methanol oxidation
The methanol oxidation reaction was performed in a xed-bed
ow reactor. A portion of 0.02 g (Nb/MCF) or 0.1 g (Nb2 O5 ) of the
catalyst of the size fraction of 0.5 < < 1 mm was placed in the reactor. The difference in catalysts weight between MCF and Nb2 O5 is
caused by keeping the same volume of both types of catalysts allowing the same contact time. The samples were activated in argon ow
(40 cm3 min1 ) at 623 K for 2 h (the rate of heating was 15 K min1 ).
Then, temperature was decreased to that of the reaction. The reactant mixture of Ar/O2 /MeOH (88/8/4 mol%) was supplied at the
rate of 40 cm3 min1 . Methanol (Chempur, Poland) was introduced
to the ow reactor by bubbling argon gas through a glass saturator lled with methanol. The reactor efuent was analysed using
two online gas chromatographs. One chromatograph, GC 8000 Top
equipped with a capillary column of DB-1, operated at 313 K with
a FID detector was applied for analyses of organic compounds and
the second one, GC containing Porapak Q and 5A molecular sieves
columns for analyses of O2 , CO2 , CO, H2 O and CH3 OH, had a TCD
detector. The columns in the second chromatograph with TCD were

heated according to the following programme: 5 min at 358 K, temperature increase to 408 K (heating rate 5 K min1 ), 4 min at 408 K,
cooling down to 358 K (for the automatic injection on the column
with 5A), 10 min at 358 K, heated to 408 K (heating rate 10 K min1 ),
11 min at 408 K. Argon was applied as a carrier gas. The outlet
stream line from the reactor to the gas chromatograph was heated
at about 373 K to avoid condensation of the reaction products.
3. Results and discussion
3.1. Catalysts characterisation
3.1.1. Chemical composition of Nb2 O5 and Nb/MCF samples
modied with Au and Cu
Two series of catalysts were prepared: (i) supported on Nb2 O5
amorphous bulk oxide and (ii) supported on Nb/MCF material prepared by the impregnation of MCF with niobium source. In both
bimetallic samples (AuCuNb2 O5 , AuCuNb/MCF), the assumed
loading of gold and copper was the same, 2 wt% of Au and 1 wt%
of Cu. For monometallic catalysts (CuNb2 O5 , CuNb/MCF and
AuNb2 O5 , AuNb/MCF) the excess of copper was applied (3 wt%),
whereas the initial content of Au was 2 wt%.
Table 1 shows the real metals (Au and Cu) loading in all the
catalysts studied. Gold loading in mono- and bimetallic samples modied by grafting with organosilanes is very close to the
assumed value (1.9 wt%). It shows that the interaction of gold (in
the form of AuCl4 ions [24]) with NH2 groups (in the form of
NH3 + ions [24]) is effective for gold incorporation. Contrary, the
real content of copper is lower than assumed, especially in the copper monometallic samples. It is due to the electrostatic repulsion
between Cu2+ and NH3 + ions. The deposition of gold onto the support in the rst step facilitates the loading of Cu2+ cations [7,24].
The atomic Au:Cu ratio in the nal materials is 0.9 for AuCuNb2 O5
and 0.7 for AuCuNb/MCF (Table 1).
3.1.2. Textural properties of the supports
The texture parameters of Nb2 O5 and Nb/MCF materials used
as supports for gold and copper, calculated from N2 adsorption/desorption measurements, are shown in Table 2. The BET
surface area of Nb/MCF mesoporous sieve is much higher than
that of Nb2 O5 . This value signicantly decreases for the samples
modied with gold and copper. Simultaneously, a decrease in pore
volumes is also observed. The decrease in surface areas and pore
volumes in Au and AuCu modied catalysts suggests the pore
blocking as a result of metals incorporation into the pores of Nb2 O5
and Nb/MCF. The average pore size in Nb/MCF support is 16.7 and
27.7 nm for windows and cells, respectively. Nb2 O5 exhibits much
smaller average pore size (ca. 5 nm). Pore sizes do not change significantly after modications with Au and Cu in both cases. However,
it is worthy of notice that Nb/MCF is mesoporous material with
narrow pore size distribution (SD1), whereas Nb2 O5 shows mesomacroporosity as deduced from its N2 adsorption isotherm (the
hysteresis loop is observed in the p/po close to the saturation pressure (p/po = 0.91.0) indicating a presence of macroporosity) and
pore size distribution (SD1).
3.1.3. The state, coordination and location of Nb in the supports
UVVis spectroscopy was used to verify the coordination and
location of Nb in the supports. The UVVis spectrum of Nb2 O5
(Fig. 1) is typical of amorphous niobia [26,27] and exhibits an
intense wide band in the range 250350 nm (with a maximum at
around 280 nm) and a lower intensity band at 222 nm. These bands
are characteristic of octahedral and tetrahedral Nb species, respectively [28]. They are attributed to the charge transfer transitions
O2 to Nb5+ , which can be associated to the energy gap between
the O 2p-valence band and the Nb4d-conductance band [28]. The

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

75

Table 1
Metals loadings and average size of gold particles in the catalysts.
Catalyst

Au, wt% as assumed

Au, wt% ICP

Cu, wt% as assumed

Cu, wt% ICP

atomic Au/Cu ratio

Au (average size), nm TEM

AuNb2 O5
AuCuNb2 O5
CuNb2 O5
AuNb/MCF
AuCuNb/MCF
CuNb/MCF

2
2

2
2

1.9
1.9

1.9
1.9

1
3

1
3

0.9
0.7

0.7
1.4

1/0
0.7/1
0/1
1/0
0.9/1
0/1

4.9
5.7

4.3
3.6

2.2
2.2
1.3
1.0

Table 2
Texture parameters of the catalysts.
Sample

Surface area, BET [m2 /g]

Pore volume [cm3 /g]

Mean pore diameter [nm]

Nb2 O5
AuCuNb2 O5
Nb/MCF
AuCuNb/MCF

85
67
607
395

0.11
0.10
2.40
1.90

5.1a
5.0a
16.7b
16.8b

Pore diameter determined from adsorption branches of N2 isotherms (BJH method).


Windows diameter determined from desorption branches of N2 isotherms (BdBFHH method).
Cells diameter determined from adsorption branches of N2 isotherms (BdBFHH method).
Table 3
XPS data for the catalysts studied.

UVVis spectra of MCF mesoporous silica modied with niobium


shows two bands at 220 and ca. 260 nm. The shift of the band from
280 to 260 nm in comparison to Nb2 O5 suggests the presence of
distorted octahedral Nb species on the surface of MCF [28]. These
octahedra are predominantly corner-sharing [29]. Moreover, the
band at 220 nm indicates the presence of tetraedrally coordinated
niobium species. Nishimura et al. and Ziolek et al. [30,31] assigned
the band at 220230 nm to tetrahedral niobium linked to the silica
surface. It suggests that part of Nb introduced by impregnation has
been localised in the walls of MCF as a result of calcination process. Thus, it can be concluded that niobium in Nb/MCF is located
both in the extra-framework positions of MCF in the form of Nb2 O5
dispersed on the surface and in the framework as isolated Nb
species.
XPS results clearly indicate that Nb in Nb2 O5 and Nb/MCF
supports is totally in the pentavalent state (Table 3). However,
the binding energy of Nb3d5/2 in Nb/MCF is signicantly higher
(208.1 eV) than that in the bulk Nb2 O5 (207.1 eV) [32]. The shift
in BE of Nb towards higher energy conrms the interaction of Nb5+
with Si in the framework of the molecular sieve as well as the interaction of small clusters of highly dispersed Nb2 O5 (Nb in distorted
octahedral coordination) with the surface of mesoporous MCF
silica.

Catalyst

Au4f [eV]

Cu2p [eV]

Nb3d [eV]

AuNb2 O5
AuCuNb2 O5

83.2
84.0

207.1
207.2

AuNb/MCF
AuCuNb/MCF

83.1
83.9

932.6
934.6

933.2
935.3

208.1
207.9

3.1.4. The state of copper and gold and the size of gold crystallites
The state of gold and copper in the catalysts prepared was studied by XRD, TEM, UVVis, XPS and H2 -TPR techniques. XRD, TEM
and UVVis results clearly indicate that metallic gold crystallites
are formed on both Nb2 O5 and Nb/MCF materials. In XRD patterns
(SD2) of the calcined Nb2 O5 and Nb/MCF samples containing Au the
reections characteristic of metallic gold (at 2 = 38.2 and 44.4 )
[33,34] are well visible. The metallic state of gold is conrmed by
UVVis spectra (Fig. 1). The observed intense ultravioletvisible
(UVVis) band at ca. 500 nm is typical of metallic gold [35]. TEM
images of the catalysts and gold particle size distributions are
shown in Fig. 2 and the average gold particle sizes of mono and
bimetallic catalysts are listed in Table 1. The results obtained clearly

258

283

506

Au-Nb/MCF
512

AuCu-Nb/MCF

Cu-Nb/MCF

0.25

Cu-Nb2O 5

222

253

Kubelka-Munk

Kubelka-Munk

261

Kubelka-Munk

220

220

AuCu-Nb2O5
272

2
*20

540

0.25

Au-Nb2O5

200 300 400 500 600 700 800

Wavelengh t, nm

Nb2O5
200

300

400

500

600

Wavelength, nm

700

800

Kubelka-Munk

a
b

27.7c
27.9c

200

300

400

500

600

700

800

Wave leng th, nm

Fig. 1. UVVis spectra of Nb/MCF and Nb2 O5 catalysts modied with Au and AuCu.

200 300 400 500 600 700 800

Wave lengh t, nm

76

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

Fig. 2. TEM images of catalysts containing gold and Au particles size distributions.

indicate that the particle size is determined by the chemical composition of the active phase and the type of support. Among the
catalysts containing gold supported on Nb/MCF, the smallest average metal particles size is noted for AuCuNb/MCF. The analysis
of the particle size distributions of AuNb/MCF and AuCuNb/MCF
materials shows that the introduction of copper increases the number of gold particles with diameters in the range 13 nm (this
fraction increases from 19 to 46%) and decreases the number of
particles with the size above 5 nm (this fraction decreases from 81
to 54%). It indicates the stabilising effect of Cu on gold nanoparticles
during calcination. A similar effect of copper was observed earlier
for AuCuSiO2 catalysts [6,15]. A coreshell model to interpret the
key role of Cu in limiting the aggregation of gold particles has been
proposed [6]. The typical coreshell structure was not formed in
the case of MCF samples studied. Copper oxide identied by H2 -TPR
measurement (described below) deposited on the gold nanoparticles in the form of some patches on the surface of gold core inhibits
the aggregation of gold particles during high-temperature calcination. The above effect of copper species, deposited in the same way,
on the gold particles size is not observed for Nb2 O5 based catalysts.
In this case the introduction of copper into Au/Nb2 O5 increases the
fraction of nanoparticles with the size above 6 nm (from 32 to 50%)
indicating the agglomeration of gold. Most probably it results from
much lower surface area (and as a consequence a lower dispersion of metal) of Nb2 O5 than Nb/MCF. Another reason may be the
presence of macropores in the structure of Nb2 O5 , which do not
limit the growth of the particles sizes promoting their agglomeration (in both AuNb2 O5 and AuCuNb2 O5 ). It is in agreement with
the results reported in [36] documenting that the nal gold particle size in porous materials increases with the pore size. Moreover,
one cannot exclude that the larger fraction of smaller gold crystallites (in the range 15 nm Fig. 2) on the surface of mono- and
bimetallic catalysts based on Nb/MCF than Nb2 O5 is also a result of
the interaction of highly dispersed niobium species on MCF surface
with gold.
Copper species on the surface of Nb2 O5 and Nb/MCF supports
were identied by XRD, UVVis and H2 -TPR techniques. In XRD
patterns of copper containing samples the small intensive reection assigned to CuO at 2 = 38 [JCPDS 45-0937] [37] is present
only for Cu/NbMCF material (SD2). However, one should remember about the XRD analysis limitations. This method gives reliable
results for relatively large crystals and high enough concentration
of species (the loading of copper is the highest on Cu/NbMCF
1.4 wt%). Moreover, the reection from CuO could be hidden in XRD
patterns of AuCu samples (due to similar position of the reexions characteristic of metallic gold and copper oxide). Therefore,
UVVis spectroscopy was applied as a complementary technique.

In UVVis spectra of Cu and AuCu modied supports (Fig. 1),


the absorption band at 258 nm is observed. The higher intensity
of that band and the shift towards lower value of wavelength in
comparison to the spectrum of AuNb/MCF results from copper
introduction. According to the literature, the bands in the range
240260 nm are assigned to the charge transfer transition between
the ligand O2 and metal centre Cu2+ of copper oxide (O2 Cu2+ )
(LMCT ligand to metal charge transfer) [38,39]. Moreover, oligonuclear [Cu+ O Cu+ ]n clusters are present on the surface of
monometallic CuNb2 O5 and CuNb/MCF giving rise to the band
in the region 400700 nm [40,41]. In the spectra of AuCuNb2 O5
and AuCuNb/MCF this band is hidden in the tail of the spectrum.
It is worth of notice that the above mentioned band is wider in the
spectra of CuNb/MCF material suggesting longer chains of clusters
due to the presence of highly dispersed Nb species on MCF support.
More information on the copper state on the surface of the
catalysts comes from the H2 -TPR results (Fig. 3). The TPR proles of copper containing Nb2 O5 and Nb/MCF materials present
signals characteristic of copper reduction at a temperature below
800 K. It is known from literature that the reduction of CuO dispersed on the surface of Nb2 O5 occurs in the range 693733 K
[42,43]. The TPR proles of Cu/Nb2 O5 and CuNb/MCF catalysts
show well distinguished peaks at 694 and at 734 K, respectively.
For the bimetallic AuCu/Nb2 O5 and AuCuNb/MCF samples these
peaks are shifted towards lower temperatures, 683 and 704 K, indicating easier copper oxide reduction, as compared to monometalic
materials. It conrms the interactions between gold and copper.
Similar behaviour has been found earlier for PdCu/Nb2 O5 catalysts [43]. Sa et al. [44] have assigned the copper reduction at lower
temperatures to the formation of bimetallic PdCu species and to
spillover of hydrogen.
There is a signicant difference in the low-temperature H2 -TPR
proles of the catalysts based on bulk niobia and on mesoporous
cellular foams (Nb/MCF). The reduction proles of Nb/MCF samples
containing copper reach two maxima in the range of 400600 K.
The rst peak ca. 450 K is due to the reduction of bulk CuO, while
the second peak at ca. 580 K is related to the partial reduction of
isolated Cu2+ ions generating Cu+ ions [45]. Moreover, one cannot
exclude that the peaks at ca. 700 K (Nb/MCF type catalysts) and ca.
690 K (Nb2 O5 type catalysts) cover also the reduction of Cu+ ions to
metallic copper.
In the high temperature (HT) range of TPR proles (>900 K) the
peaks assigned to the reduction of niobium species appear. The rst
peak at 1158 K in the proles of Nb2 O5 based catalysts is attributed
to the reduction of Nb5+ to Nb4+ in the form of Nb2 O5 to NbO2 , while
the second at 1255 K to the reduction of NbO2 to lower oxidation
state of niobium. The low intensity peak at around 940 K would be

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

assigned to the reduction of smaller Nb2 O5 clusters [46]. The TPR


proles of Nb/MCF based samples in the HT range show a major
peak at around 1160 K conrming the presence of Nb2 O5 clusters
on the MCF surface. Most probably this broad peak covers also the
reduction signal of Nb localised in the framework of MCF [47]. It is
important to note that the reduction of the dispersed CuO is easier
for the catalysts based on Nb2 O5 than Nb/MCF, pointing to different
CuNb interactions in both types of catalysts.
The XRD, UVVis and H2 -TPR complementary results indicate,
that in both AuCuNb2 O5 and AuCuNb/MCF samples, copper is
present mainly in the form of CuO and [Cu+ O Cu+ ]n clusters
and in smaller part as copper ions (Cu2+ and/or Cu+ ).
More precise information on the oxidation states of gold and
copper species in Nb-containing catalysts can be obtained from XPS
study (Table 3). XPS data conrm the results obtained from XRD,
UVVis and TEM measurements clearly indicating that gold in the
metallic state is present on the surface of all gold-containing samples. The binding energies of Au 4f7/2 for monometallic samples are
83.2 eV and 83.1 eV for AuNb2 O5 and AuNb/MCF, respectively.
The shift of Au 4f7/2 peaks towards a lower binding energy than that
of bulk Au0 (84.0 eV) [48] indicates that negatively charged gold
particles (Au ) are formed on the surface of momometallic catalysts. Such electronegative gold species have also been observed
earlier for Au/TiO2 [49,50] and their presence has been attributed to
the changes in electronic structure of the gold particles as a result of
the electron transfer from the titania support to the Au particles. In
the case of catalysts used in this work this behaviour is most probably a result of the modication procedure applied. In this procedure
the negatively charged AuCl4 ions are adsorbed on NH3 + ions from
organosilane and after the reduction with NaBH4 and the removal of
HCl on washing, the negatively charged gold particles are produced
[24]. The introduction of Cu2+ ions changes the electronic state of
gold. The binding energy of Au increases (84.0 for AuNb2 O5 and
83.9 eV for AuNb/MCF) and it is similar to the energy of metallic
Au described in the literature [48]. The shift of XPS peak from Au4f
in relation to the position characteristic of monometallic gold samples conrms the synergetic interactions between gold and copper.
Bimetallic samples (AuCuNb2 O5 and AuCuNb/MCF) are characterised also by the presence of copper in two oxidation states, Cu2+
and Cu+ (Table 3). It is in agreement with XRD, UVVis and H2 -TPR
results and conrms the presence of CuO and [Cu+ O Cu+ ]n
copper species on the surface of catalysts. However, it is important
to note that BE of copper ions depends strongly on the nature of the
support. It has been reported that the BE of Cu2p3/2 characteristic
of Cu+ species varies between 932.3 and 933.6 eV [9,51], whereas
that characteristic of Cu2+ between 933.5 and 935.0 eV [39,5254].
For the AuCuNb2 O5 catalyst, the Cu2p3/2 XPS peaks corresponding to Cu2+ and Cu+ appeared at 934.6 eV and 932.6 eV, respectively.
For AuCuNb/MCF, the BE of Cu2p3/2 is shifted to the higher values,
935.3 eV for Cu2+ and 933.2 eV for Cu+ . For the last sample the BE
characteristic of Cu2+ is signicantly higher than that of bulk CuO
(933.7 eV) [55]. Such a shift towards higher energy may result from
the better dispersion of copper oxide on MCF mesoporous silica
than on bulk Nb2 O5 [56].
3.1.5. Acidity/basicity of the catalysts 2-propanol
decomposition test reaction
The acidity/basicity of the catalysts applied in this work was
determined in the test reaction of 2-propanol dehydration and
dehydrogenation [56]. Dehydration of alcohol to propene and/or
di-isopropyl ether requires acidic centres (Lewis or Brnsted),
whereas the dehydrogenation to acetone occurs on the basic sites.
Some authors [e.g. 57] have reported that acetone formation takes
place on redox centres. As shown in Table 4 on all catalysts the
main reaction product of 2-propanol decomposition is propene,
indicating the domination of acidic properties of the samples. Both,

77

Table 4
The results of 2-propanol decomposition reaction.
Catalyst

Nb/MCF
AuNb/MCF
AuCuNb/MCF
Nb2 O5
AuNb2 O5
AuCuNb2 O5

i-PrOH conv. [%]


573 K
67
13
18
96
74
92

Selectivity [%] towards


Propene

Acetone

100
96
71
100
97
50

0
4
29
0
3
50

dispersed Nb2 O5 and Nb+ species in the framework are the source
of Lewis acidity which comes from the Nb/MCF support. On the
other hand, Brnsted and Lewis acid centres are present on amorphous Nb2 O5 support [27]. The modication with gold decreases
the acidity of catalysts (a decrease in activity is observed). However, the introduction of Cu species into Au-monometallic materials
enhances the activity by generating basic/redox centres as deduced
from an increase in the selectivity to acetone. The acetone formation indicates the dehydrogenating properties of bimetallic
catalysts. The oxygen from CuO is a source of basicity, whereas
copper ions are responsible for redox activity.
3.2. Glycerol oxidation with oxygen in the liquid phase
Catalytic oxidation of glycerol with oxygen allows the production of a number of new derivatives, among them glyceric, tartronic
and glycolic acids that have attracted recently signicant interest
[3,15,16,58,59]. The crucial research target is to control selectivity
for the desired product. It is known that the chemical composition
of the active phase, the temperature of the reaction, the pH of the
reaction mixture and the particle size of the metal are crucial for
the formation of a desired product.
3.2.1. Effect of gold and copper
The prepared mono and bimetallic AuCu catalysts based on
Nb2 O5 were tested in glycerol oxidation processes. The catalyst
activity was estimated by glycerol conversion at 333 and 363 K and
the results are given in Table 5 and Fig. 4.
For the reactions on AuNb2 O5 , CuNb2 O5 and AuCuNb2 O5
under identical reaction conditions (333 K), an increase in glycerol
conversion is observed for bimetallic catalyst indicating the synergetic effect between Au and Cu species (Fig. 4 A). A similar effect
of bimetallic active phase on the activity in glycerol oxidation has
been found earlier for AuPt, AuPd, PtCu catalysts supported on
carbon [e.g. 17, 19, 20]. Although negatively charged gold particles in monometallic Au catalysts (present also on Au/Nb2 O5 ) have
been found active in aerobic oxidation of alcohols [6062], the synergetic effect between Au0 and Cu species on Nb2 O5 seems to be
more important factor to obtain high performance of Nb2 O5 catalysts in glycerol oxidation. Moreover, it has been reported [15]
that the size effect i.e., the small particle size is responsible for
the improved activity of AuCu/SiO2 (a decrease in gold particles
size for bimetallic catalysts was observed) in the benzyl alcohol
oxidation. The theoretical calculations have also revealed that the
activity for the alcohol oxidation increases with decreasing Au coordination number [63]. Taking into account that the average size of
gold crystallites in AuCuNb2 O5 is larger than in AuNb2 O5 , no
correlation can be made between the increase in the activity of
bimetallic sample studied in this work with the size of gold particles.
Depending on the type of active phase (Au vs AuCu) a different distribution of products is observed for each catalyst (Table 5,
Fig. 4 B). AuNb2 O5 activates the reaction mainly towards glyceric

78

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

Table 5
Catalytic activity and selectivity of Au and AuCu-supported catalysts in liquid phase oxidation of glycerol.
Catalyst

Temp., K

Glycerol conv. %

AuNb2 O5
AuNb2 O5
AuCuNb2 O5
AuCuNb2 O5
CuNb2 O5
AuCuNb2 O5 a
AuCuNb/MCFa

333
363
333
363
363
363
363

54
93
65
90
51
23
15

Selectivity, %
Glyceric acid

Tartronic acid

Glycolic acid

Formic acid

Oxalic acid

Glyoxylic acid

44
45
24
13
10
traces
1

traces
1
traces

1
11
14
27
25
3
1

traces
1
7
16
13
traces

traces
1
1
1
1

traces
8

Reaction conditions: glycerol 100 mmol/kg, NaOH/glycerol molar ratio = 2, pO2 = 6 bar.
a
Reaction mixture without NaOH.

1158
50

1255

H2 consumption. a.u.

H2 consumption, a.u.

250

694

Cu-Nb2O5
683

AuCu-Nb2O5
400

935

600

800

1000

734

1162

597

448

Cu-Nb/MCF
704

583

AuCu-Nb/MCF

453

400

1200

600

800

1000

1200

1400

Temperature, K

Temperature, K

Fig. 3. H2 -TPR proles of Au and AuCu modied catalysts.

acid (44% of selectivity at 54% of glycerol conversion). The selectivity to other products of oxidation, including tartronic, glycolic
and formic acids is very low. It indicates that at low (333 K) reaction temperature the over-oxidation of glyceric acid is limited. The

introduction of copper leads to glycolic and formic acids formation


(the selectivity increases up to 14% and 7%, respectively). It suggests
that the copper species are active in the oxidative dehydrogenation process of glycerol. The initial dehydrogenation of glycerol

Fig. 4. The effect of copper species and the reaction temperature on the activity (A) and selectivity (B) of AuNb2 O5 catalysts in glycerol oxidation. GLYCEA = glyceric acid,
GLYCOA = glycolic acid, FORA = formic acid.

OH
O

OH

glyceraldehyde
[O]

[O]
[O]

OH
HO

OH
glycerol

[-H2]

OH
HO

OH
O
glyceric
acid

[O]
[-H2]

OH
HO

OH
O
O
tartronic
acid

[-CO2]

HO

[O]

O
OH

glycolic
acid

Scheme 1. Possible reaction pathway for glycerol oxidation.

[ - H2]

[-CO2] formaldehyde

O
H

OH

formic
acid

towards glyceraldehyde is proposed as the rst step in the oxidation process [64]. The presence of OH is essential to observe
any glycerol oxidation when using gold containing catalysts. In the
presence of NaOH, the proton is readily abstracted from one of
the primary hydroxyl groups of glycerol. The formation of AuCusystem increases the dehydrogenating properties of AuNb2 O5
catalysts, in line with the results of 2-propanol decomposition reaction. Moreover, the copper-containing catalysts are known to be
active in dehydrogenation of alcohols [6567]. The formation of
glycolic acid on AuCuNb2 O5 suggests that copper species catalyse the dehydrogenation of glyceric acid towards tartronic acid
which is next transformed to glycolic acid (via decarboxylation)
and subsequently to formic acid (via degradation of glycolic acid
to formaldehyde and CO2 and next oxidation of formaldehyde)
(Scheme 1). Copper species can also be involved in generation of
active oxygen on the catalyst surface. It has been shown that in
CO oxidation reaction Cu played an important role in activating
oxygen [6,7]. The presence of copper in contact with gold contributes much to the activation of oxygen.
The oxidation of glycerol is a typical consecutive reaction
proceeding according to the rake model [18]. In such a process the
selectivity is strongly determined by the adsorption and desorption
rate of each product. Thus, the relation between the rate of desorption of glyceric acid and the rate of its transformation to tartronic
and glycolic acids determines the selectivity to one or the other
product. The faster the desorption of glyceric acid, the lower the
selectivity to tartronic and glycolic acids. Therefore, it cannot be
excluded that copper species in AuCuNb2 O5 catalyst strengthen
the chemisorption of glyceric acid and thanks to it the selectivity
to glycolic acid via formation of tartronic acid increases.
As shown in Fig. 4 A, the increase in the reaction temperature
from 333 K to 363 K causes a signicant increase in the glycerol conversion, for both, AuNb2 O5 and AuCuNb2 O5 catalysts. Along with
increasing activity caused by growing temperature, an increase in
glyceric acid yield on AuNb2 O5 (from 24 to 42%) and glycolic acid
yield on AuCuNb2 O5 (from 9% to 25%) is observed (Fig. 4B). Similarly as at 333 K, glyceric acid is the main product of the reaction
at 363 K on AuNb2 O5 . The increase in temperature promotes the
formation of glycolic acid (an increase in selectivity observed on
Au/Nb2 O5 ) but a comparison of the selectivity of AuNb2 O5 and
AuCuNb2 O5 obtained for similar glycerol conversion (93 and 90%,
respectively) conrms that copper species are active in glycolic acid
formation on AuCuNb2 O5 (an increase in selectivity from 11 up
to 27% when copper is present on the catalyst surface). Moreover,
glycolic acid is also the main reaction product on Cu/Nb2 O5 . However, it is worth noting that when using AuCuNb2 O5 , the yield to
glycolic acid is higher than when using CuNb2 O5 (25% and 13%,
respectively) (Fig. 4B).
The sum of the selectivities towards products of glycerol oxidation is lower than 100%. It could be due to the polymerisation
of the reaction products to compounds not detected by the HPLC,

MeOH conversion, %

473 K

16

37

Au-Nb/MCF

With NaOH
Without NaOH

90

AuCu-Nb/MCF

23

15

AuCu-Nb2O5

AuCu-Nb/MCF

oxidation of the reaction products to CO2 or the adsorption of products on the supports surface as suggested in [68].
3.2.2. Effect of pH of the reaction mixture
As proved in the literature [64,69], glycerol oxidation on gold
catalysts is favoured by high-pH conditions, since the hydroxide
species adsorbed on gold are necessary to activate both C H and
O H bonds of glycerol (also adsorbed on the surface). However,
Prati et al. [70] and Hutchings et al. [71] have reported the basefree oxidation of glycerol to glyceric acid with good conversion
and selectivity obtained using AuPt catalysts supported on carbon, mordenite and MgO. High activity and selectivity to glyceric
acid have been also obtained over a PtCu/C [17]. In all cases the
improved performance of bimetallic catalysts in oxidation of glycerol under base-free conditions was attributed to the formation of
AuPt or PtCu alloyed phases.
In this study, the glycerol conversion was examined using
the reaction mixture without NaOH (pH 8) in the presence of
AuCuNb2 O5 and AuCuNb/MCF catalysts. Mesoporous silica based
catalysts were not studied in the presence of NaOH because of
solubility of silica in basic solutions. The results are shown in
Fig. 5. It can be clearly observed that the activity of AuCuNb2 O5
signicantly decreases when the reaction is performed without
NaOH. It indicates that the formation of alloy is necessary to obtain
high activity of catalysts in this reaction conditions. When copper species exist separately, in the neighbourhood of gold, NaOH
is necessary for the initiation of glycerol oxidation reaction, similarly as for monometallic gold catalysts. The observed selectivity to
tartronic and glycolic acid conrms that copper species are active in
oxidative dehydrogenation of glycerol to tartronic acid. The activity of AuCuNb/MCF catalyst is lower than that of AuCuNb2 O5 . It
indicates the role of AuNb interactions when bulk Nb2 O5 is used
as a support (SMSI between gold and niobium). The AuNb interaction in bulk Nb2 O5 is stronger than when niobium is dispersed
on the mesoporous MCF support.

99

79

Fig. 5. The effect of the base in the reaction mixture on the activity of AuCuNb2 O5
and AuCuNb/MCF catalysts in glycerol oxidation at 363 K.

523 K

98

100
90
80
70
60
50
40
30
20
10
0

AuCu-Nb2O5

100
90
80
70
60
50
40
30
20
10
0

MeOH conversion, %

A
100
90
80
70
60
50
40
30
20
10
0

Glycerol conversion, %

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

373 K

423 K

97

22

99

37

Au-Nb2O5

AuCu-Nb2O5

Fig. 6. The effect of copper species on the activity of Au-catalysts in methanol oxidation: Nb/MCF supported catalysts (A), Nb2 O5 supported catalysts (B).

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

MeOH molecues*min-1*nm-2

80

373 K
45
40
35
30
25
20
15
10
5
0

25

423 K

473 K

41

40

34

25
10

Fig. 7. The effect of the support on the activity of Au- and AuCu-containing catalysts in methanol oxidation (activity as the number of MeOH molecules reacted per
minute per nm2 of the surface area of gold crystallites).

3.3. Methanol oxidation with oxygen in the gas phase


Methanol oxidation is another important target process in
industry [72,73] leading to various products depending on the
nature of catalyst and temperature of reaction. Formaldehyde (FA)
is nowadays the most desirable industrial product of MeOH oxidation, because it is an important intermediate in the synthesis
of many chemicals. Selective oxidation of methanol to formaldehyde over gold catalysts has recently been studied on Au/Al2 O3 [74],
Au/ZnO, Au/TiO2 [75] and Auzeolites [76]. Moreover, the catalytic
combustion of methanol is one of the most promising processes
for volatile organic compounds abatement. Supported noble metals
(Pt, Pd, Rh) [77], Au/iron oxides and Au/CeO2 [78,79] are among the
catalysts employed in this process.
The results of MeOH oxidation at 373523 K on catalysts studied in this work are shown in Table 6 and Figs. 68. It is clear that
the activity (shown as MeOH conversion [%] and MeOH conversion rate [mmol/g/min]) of AuCu-catalysts is signicantly higher
than those of monometallic gold-catalysts, on both Nb2 O5 and
Nb/MCF supports. It clearly indicates that the presence of copper improves the catalytic activity of monometallic samples in
MeOH oxidation as a result of synergetic interactions between Au
and Cu species (enhancement of the redox properties of the system). Cu-O like species play the role of active centres promoted
by the presence of gold. A similar effect of copper introduction
has been found earlier for AuCu-MCM-22 [13], AuCuCeO2 /SBA-15
and AuCuCeO2 /ZrO2 /SBA-15 catalysts [14]. For the reaction performed on AuCuNb2 O5 and AuCuNb/MCF, 97% (373 K) and 98%
(473 K) of methanol conversion was reached, respectively (Fig. 6A
and B). The lower reaction temperature, in which almost total conversion of methanol on AuCuNb2 O5 is observed (Fig. 6B), indicates
higher activity of the catalyst based on Nb2 O5 . It again indicates the
role of AuNb interactions when bulk Nb2 O5 is used as a support
(SMSI). The higher activity of Nb2 O5 samples supporting Au and Cu
in MeOH oxidation was conrmed by the calculations expressing
the activity as the number of MeOH molecules reacted per minute
per nm2 of the surface area of gold crystallites (Fig. 7). Moreover,
it is worth of notice that the small size of gold nanoparticles on
the surface of Nb/MCF and Nb2 O5 (smaller on AuCuNb/MCF than
AuCuNb2 O5 ) catalysts does not play an important role in MeOH
oxidation. It is in line with our earlier studies [13,80].
The introduction of copper into gold catalysts inuences not
only the activity, but also the selectivity in methanol oxidation.
As shown in Table 6 and Fig. 8, the modication of Au/Nb2 O5
and AuNb/MCF with copper signicantly changes the distribution

of products. For AuCuNb/MCF, the increase in the selectivity to


methyl formate in comparison to AuNbMCF, measured for a similar MeOH conversion, is observed (from 77 to 98% Fig. 8A). At
a higher temperature (523 K) on AuCuNb/MCF methyl formate is
transformed to CO2 (64% of selectivity) (Fig. 8A). It suggests that
basic oxygen from CuO is responsible for CO2 formation. However, it is important to add that CuNb/MCF material shows low
activity in MeOH oxidation (5% of conversion at 523 K) with high
selectivity to formaldehyde. Similarly as for AuCuNb/MCF, also for
AuCuNb2 O5 an increase in the selectivity to CO2 in comparison to
AuNb2 O5 is observed (from 25 to 97%, measured for similar MeOH
conversion Fig. 8 B). However, it is important to stress that the
ability to total oxidation of MeOH is much higher for Nb2 O5 than
Nb/MCF based catalysts (64% and 97% of selectivity to CO2 at 99%
of MeOH conversion for AuCuNb/MCF (523 K) and AuCuNb2 O5
(423 K), respectively) (Fig. 8). At a lower temperature (373 K) 74%
of selectivity to CO2 is observed (Table 6). It shows the usefulness
of AuCuNb2 O5 catalysts for the removal of VOC.
Moreover, it worth of notice that dimethyl ether appears
among the products of methanol oxidation reaction on AuNb2 O5
(Table 6). This product is obtained by a bimolecular dehydration
of MeOH on acid centres of the support. Gold and copper loading
signicantly diminishes the formation of dimethyl ether, because
of the blockade of acid centres on the surface of the support.
According to the mechanism of methanol oxidation proposed in literature, methanol oxidation involves the formation
of chemisorbed methoxy groups on acidic sites [73,81,82] with
participation of nucleophilic oxygen (basic sites) or redox centres [82,83]. It was shown [75] that the support is a reservoir of
adsorbed methoxy species and the activation of oxygen occurs at
the perimeter of Au nanoparticles. Only methoxy species adsorbed
on the support in the direct neighbourhood of the Au nanoparticles can act as intermediates. It suggests the relevance of the
goldsupport interface in initiation of the catalytic reaction. This
feature was found also for other reactions (e.g. CO oxidation [84]
and WGS process [85]). Methoxy species adsorbed on the support are further transformed to formaldehyde (FA) resulting from
the extraction of hydrogen [73,81]. If FA is chemisorbed strongly
enough on nucleophilic sites, it can interact with the next MeOH
molecule and form methyl formate (MF). The oxidation of methanol
to CO2 can proceed in the radical reaction on basic centres or step by
step with the oxidation to formaldehyde and formic acid on redox
centres. The results described above indicate that the adsorption of
formaldehyde on Au-catalysts is strong and that is why a formaldehyde molecule interacts with another methanol molecule to form
methyl formate (with high selectivity) with the participation of the
pairs of acid and basic centres (from the surface of the support).
Copper oxide species (CuO and oligonuclear [Cu+ O Cu+ ]n
clusters) present in AuCu containing catalysts exhibit acidicbasic
properties, whereas copper ions are the source of Lewis acid sites.
The presence of MF in the reaction products on AuCu-samples
implies the participation of LAS in the reaction pathway towards
CO2 . Lewis acid sites are strong enough and do not allow CH2 O
molecules to desorb from the surface of AuCu-catalysts to the gas
phase. Strong LAS favour the reaction with the other methanol
molecules towards formic acid and further to MF. Synergetic interaction of gold with copper species enhances the redox properties
of CuO-like species (indicated by H2 -TPR study) with an overall
increase in the mobility of oxygen which favours total oxidation of
methanol (higher on Nb2 O5 supported catalysts). With increasing
reaction temperature (Table 6) the selectivity to CO2 becomes more
important. Taking into account the production of MF at lower temperatures of the reaction on bimetallic samples, one can conclude
that the total oxidation proceeds through formate intermediate. In
the consecutive process of CO2 formation, the basic oxygen plays an
important role. It comes from copper-oxide species. The stronger

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

81

Table 6
Catalytic activity and selectivity in MeOH oxidation reaction carried out at different temperatures on Nb/MCF and Nb2 O5 supported catalysts.
Catalyst

Temp. [K]

MeOH conv. [%]

MeOH conv.
rate 102
[mmol/g/min]

HCHO

HCOOCH3

CH3 OCH3

C2 H4

CH3 OCH2 OCH3

CO2

AuNb/MCF
AuNb/MCF
AuCuNb/MCF
AuCuNb/MCF
AuCuNb/MCF
AuCuNb/MCF
CuNb/MCF
CuNb/MCF
AuNb2 O5
AuNb2 O5
AuNb2 O5
AuCuNb2 O5
AuCuNb2 O5
CuNb2 O5
CuNb2 O5

473
523
373
423
473
523
473
523
373
423
523
373
423
473
523

16
37
8
19
98
99
1
5
22
37
99
97
99
20
38

57
132
29
68
350
353
4
18
16
26
71
69
71
14
27

17
27

traces
31
traces
60

2
40

28
22

77
61
100
98
18
5

11
57
67

25
2
8
5

traces
traces

2
traces
traces
24
1
1
4
23

traces
traces
11
traces
traces
traces
traces

traces

1
1

3
traces

6
12
traces
2
82
64
traces
27
42
31
25
74
97
57
50

16

FA 17

Au-Nb/MCF
473 K

MF
98

19
AuCu-Nb/MCF
423 K

99

CO2
64

FA
31
AuCu-Nb/MCF
523 K

100
90
80
70
60
50
40
30
20
10
0

100
90
80
70
60
50
40
30
20
10
0

conversion

selectivity

EN 11
DME
24
99

FA
40

99

CO2
97

CO2
25
Au-Nb2O5
523 K

AuCu-Nb2O5
423 K

100
90
80
70
60
50
40
30
20
10
0

Selectivity, %

MF
77

selectivity
MeOH conversion, %

conversion

Selectivity, %

MeOH conversion, %

A
100
90
80
70
60
50
40
30
20
10
0

Selectivity [%]

Fig. 8. Effect of copper species on the selectivity of AuCuNb/MCF (A) and AuCuNb2 O5 (B) catalysts in methanol oxidation. FA = formaldehyde, MF = methyl formate,
EN = ethene, DME = dimethyl ether.

metalsupport interaction in bulk Nb2 O5 based catalysts leading


to easier reducibility of copper species makes the total oxidation of
MeOH on AuCuNb2 O5 faster.
4. Conclusions
Nb2 O5 and Nb/MCF supports were modied with Au and Cu in
order to study the effect of copper species on the properties, activity
and selectivity of gold catalysts in oxidation processes (AuCu vs Au
systems). Copper present on the surface of supports in the form of
CuO, oligonuclear [Cu+ O Cu+ ]n clusters and isolated cations
changes the electronic structure of gold. The negatively charged
gold particles (Au ) are formed on the surface of momometallic catalysts, whereas the introduction of copper species generates
metallic gold (the BE of Au increases). The presence of gold in metallic state facilitates the reduction of copper. Higher surface area
of Nb/MCF support and the interaction of highly dispersed niobium species on MCF surface with gold result in a larger fraction
of smaller gold crystallites (in the range 15 nm) on the surface of
mono- and bimetallic catalysts based on Nb/MCF in comparison to
Nb2 O5 . The addition of copper enhances this effect.
The obtained results have shown that the AuCu system based
on Nb2 O5 is highly active in glycerol and methanol oxidation processes. The modication of AuNb2 O5 with copper increases the
activity in comparison with monometallic gold catalyst due to the
synergy between gold and copper. The activity of Nb/MCF catalysts
in methanol oxidation is lower than that of Nb2 O5 based catalysts.
The strong metalsupport (bulk Nb2 O5 ) interactions are responsible for the higher activity of AuCuNb2 O5 system.

The introduction of copper into gold catalysts inuences not


only the activity, but also the selectivity in oxidation reactions.
Depending on the type of active phase (Au vs AuCu) the distribution of products is different. AuNb2 O5 activates the glycerol
oxidation mainly towards glyceric acid, whereas the addition of
copper signicantly increases the selectivity towards glycolic acid.
Copper species are active in the oxidative dehydrogenation process
of glyceric acid towards tartronic acid which is next transformed
to glycolic acid (via decarboxylation) and subsequently to formic
acid (via degradation of glycolic acid to formaldehyde and CO2
and next formaldehyde oxidation). Bimetallic AuCuNb/MCF and
AuCuNb2 O5 catalysts exhibit high activity in low temperature
total oxidation of methanol due to the presence of basic oxygen
from CuO on the catalyst surface. The ability to total oxidation of
MeOH is much higher over Nb2 O5 than over Nb/MCF based catalysts. It indicates the usefulness of AuCuNb2 O5 catalysts for the
removal of VOC.
The results obtained suggest that by the use of monometallic
AuNb2 O5 and AuNb/MCF or by the proper choice of the second metal introduced on the surface of Au-containing Nb2 O5 and
Nb/MCF materials one can obtain the catalysts selective to the
desired products of oxidation of glycerol and methanol.
Acknowledgements
The authors are grateful to Prof. Maria
Mickiewicz University in Poznan for very
of the results. National Science Centre in
2013/10/ST5/00642) is acknowledged for the

Ziolek from Adam


helpful discussion
Poland (Grant No.
nancial support of

82

I. Sobczak, . Wolski / Catalysis Today 254 (2015) 7282

this work. Thanks are also due to Johnson Matthey (UK-USA) and
CBMM (Brazil) for the kind gifts of HAuCl4 and Nb2 O5 , respectively.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.cattod.
2014.10.051.
References
[1] G.C. Bond, C. Luis, D.T. Thompson, Catalysis by Gold, Imperial College Press,
2006.
[2] A. Wang, X. Yan Liu, Ch.-Y. Mou, T. Zhang, J. Catal. 308 (2013) 258271.
[3] M. Besson, P. Gallezot, C. Pinel, Chem. Rev. 114 (2014) 18271870.
[4] Ch.L. Bracey, P.R. Ellis, G.J. Hutchings, Chem. Soc. Rev. 38 (2009) 22312243.
[5] H. Wang, W. Fan, Y. He, J. Wang, J.N. Kondo, T. Tatsumi, J. Catal. 299 (2013)
1019.
[6] X. Liu, A. Wang, T. Zhang, D.-S. Su, Ch.-Y. Mou, Catal. Today 160 (2011) 103108.
[7] X. Liu, A. Wang, L. Li, T. Zhang, Ch.-Y. Mou, J.-F. Lee, J. Catal. 278 (2011) 288296.
[8] R.J. Chimentao, F. Medina, J.L.G. Fierro, J. Llorca, J.E. Sueiras, Y. Cesteros, P.
Salagre, J. Mol. Catal. A: Chem. 274 (2007) 159168.
[9] J. Llorca, M. Dominguez, C. Ledesma, R.J. Chimentao, F. Medina, J. Sueiras, I.
Angurell, M. Seco, O. Rossell, J. Catal. 258 (2008) 187198.
[10] C. Della Pina, E. Falletta, M. Rossi, J. Catal. 260 (2008) 384386.
[11] W. Li, A. Wang, X. Liu, T. Zhang, Appl. Catal. A: Gen. 433434 (2012) 146151.
[12] R.A. Sheldon, J.K. Kochi, Metal-Catalysed Oxidation of Organic Compounds,
Academic Press, New York, 1981.
[13] P. Kaminski, I. Sobczak, P. Decyk, M. Ziolek, W.J. Roth, B. Campo, M. Daturi, J.
Phys. Chem. C 117 (2013) 21472159.
[14] P. Kaminski, M. Ziolek, J. Catal. 312 (2014) 249262.
[15] Y. Zhang, X. Cui, F. Shi, Y. Deng, Chem. Rev. 112 (2012) 24672505.
[16] E.G. Rodrigues, S.A.C. Carabineiro, J.J. Delgado, X. Chen, M.F.R. Pereira, J.J.M.
rfao, J. Catal. 285 (2012) 8391.
[17] D. Liang, J. Gao, J. Wang, P. Chen, Y. Wei, Z. Hou, Catal. Commun. 12 (2011)
10591062.
[18] I. Sobczak, K. Jagodzinska, M. Ziolek, Catal. Today 158 (2010) 121129.
[19] L. Prati, A. Villa, F. Porta, D. Wang, D. Su, Catal. Today 122 (2007) 386390.
[20] C.L. Bianchi, P. Canton, N. Dimitratos, F. Porta, L. Prati, Catal. Today 102103
(2005) 203212.
[21] I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 36033624.
[22] P. Schmidt-Winkel, W.W. Lukens Jr., P. Yang, D.I. Margolese, J.S. Lettow, J.Y.
Ying, G.D. Stucky, Chem. Mater. 12 (2000) 686696.
[23] X. Zhang, E. Sau Man Lai, R. Martin-Aranda, K.L. Yeung, Appl. Catal. A: Gen. 261
(2004) 109118.
[24] X. Liu, A. Wang, X. Yang, T. Zhang, Ch.-Y. Mou, D.-S. Su, J. Li, Chem. Mater. 21
(2009) 410418.
[25] W.W. Lukens Jr., P. Schmidt-Winkel, D. Zhao, J. Feng, G.D. Stucky, Langumir 15
(1999) 54035409.
[26] M. Ziolek, I. Sobczak, P. Decyk, L. Wolski, Catal. Commun. 37 (2013) 8591.
[27] D. Prasetyoko, Z. Ramli, S. Endud, H. Nur, Mater. Chem. Phys. 93 (2005) 443449.
[28] T. Armaroli, G. Busca, C. Carlini, M. Giuttari, A.M.R. Galletti, G. Sbrana, J. Mol.
Catal. A 151 (2000) 233243.
[29] M. Wiegel, M. Hamoumi, G. Blasse, Mater. Chem. Phys. 36 (1994) 289293.
[30] M. Nashimura, K. Asakura, Y. Iwasawa, J. Chem. Soc. Chem. Commun. 15 (1986)
16601661.
[31] B. Kilos, A. Tuel, M. Ziolek, J.C. Volta, Catal. Today 118 (2006) 416424.
[32] M. Trejda, A. Wojtaszek, A. Floch, R. Wojcieszak, E.M. Gaigneaux, M. Ziolek,
Catal. Today 158 (2010) 170177.
[33] C. Kan, W. Cai, Z. Li, G. Fu, L. Zhang, Chem. Phys. Lett. 382 (2003) 318324.
[34] A.C. Gluhoi, X. Tang, P. Marginean, B.E. Nieuwenhuys, Top. Catal. 39 (2006)
101110.
[35] D.L. Feldheim, C.A. Foss, Metal Nanoparticles. Synthesis, Characterization and
Applications, Basel Marsel Dekker, Inc., New York, 2002.
[36] M.T. Bore, H.N. Pham, E.E. Switzer, T.L. Ward, A. Fukuoka, A.K. Datye, J. Phys.
Chem. B 109 (2005) 28732880.
[37] T. Tsoncheva, J. Rosenholm, M. Linden, Appl. Catal. A: Gen. 318 (2007) 234243.
[38] A. Gallo, T. Tsoncheva, M. Marelli, M. Mihaylov, M. Dimitrov, V. Dal Santo, K.
Hadjiivanov, Appl. Catal. B: Environ. 126 (2012) 161171.
[39] A. Kong, H. Wang, X. Yang, Y. Hou, Y. Shan, Microporous Mesoporous Mater.
118 (2009) 348353.
[40] Y. Jiang, Q. Gao, Mater. Lett. 61 (2007) 22122216.
[41] C.M. Chanqua, A.L. Cnepa, J. Bazn-Aguirre, K. Sapag, E. Rodrguez-Castelln, P.
Reyes, E.R. Herrero, S.G. Casuscelli, G.A. Eimer, Microporous Mesoporous Mater.
151 (2012) 212.

[42] C.E.M. Guarido, D.V. Cesar, M.M.V.M. Souza, M. Schmal, Catal. Today 142 (2009)
252257.
[43] M. Pinto Maia, M. Andrade Rodrigues, F. Barboza Passos, Catal. Today 123 (2007)
171176.
[44] J. Sa, H. Vinek, Appl. Catal. B: Environ. 57 (2005) 247256.
[45] I. Sobczak, P. Decyk, M. Ziolek, M. Daturi, J.C. Lavalley, L. Kevan, A.M. Prakash,
J. Catal. 207 (2002) 101112.
[46] I. Nowak, M. Jaroniec, Top. Catal. 49 (2008) 193203.
[47] M. Ziolek, I. Sobczak, A. Lewandowska, I. Nowak, P. Decyk, M. Renn, B.
Jankowska, Catal. Today 70 (2001) 169181.
[48] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray
Photoelectron Spectroscopy, PerkinElmer Corporation, Physical Electronics
Division, Eden Prairie, 1992.
[49] P. Lignier, M. Comotti, F. Schuth, J.-L. Rousset, V. Caps, Catal. Today 141 (2009)
355360.
[50] A. Zielinska-Jurek, E. Kowalska, J.W. Sobczak, W. Lisowski, B. Ohtani, A. Zaleska,
Appl. Catal. B: Environ. 101 (2011) 504514.
[51] S. Albonetti, T. Pasini, A. Lolli, M. Blosi, M. Piccinini, N. Dimitratos, J.A. LopezSanchez, D.J. Morgan, A.F. Carley, G.J. Hutchings, F. Cavani, Catal. Today 195
(2012) 120126.
[52] Y. Huang, H. Ariga, X. Zheng, X. Duan, S. Takakusagi, K. Asakura, Y. Yuan, J. Catal.
307 (2013) 7483.
[53] C.M. Chanquia, K. Sapag, E. Rodriguez-Castellon, E.R. Herrero, G.A. Eimer, J. Phys.
Chem. C 114 (2010) 14811490.
[54] M.V. Lombardo, M. Videla, A. Calvo, F.G. Requejo, G.J.A.A. Soler-Illia, J. Hazard.
Mater. 223224 (2012) 5362.
[55] Z. Wu, Y. Wang, W. Huang, J. Yang, H. Wang, J. Xu, Y. Wei, J. Zhu, Chem. Mater.
19 (2007) 16131625.
[56] A. Gervasisni, J. Fenyvesi, A. Auroux, Catal. Lett. 43 (1997) 219228.
[57] C. Lahousse, J. Bachelier, J.C. Lavalley, H. Lauron-Pernot, A.M. Le Govic, J. Mol.
Catal. 87 (1994) 329332.
[58] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 52535277.
[59] Ch.-H. Zhou, J.N. Beltramini, Y.-X. Fan, G.Q. Lu, Chem. Soc. Rev. 37 (2008)
527549.
[60] H. Tsunoyama, H. Sakurai, Y. Negishi, T. Tsukuda, J. Am. Chem. Soc. 127 (2005)
93749375.
[61] H. Tsunoyama, H. Sakurai, T. Tsukuda, Chem. Phys. Lett. 429 (2006) 528532.
[62] H. Tsunoyama, N. Ichikuni, H. Sakurai, T. Tsukuda, J. Am. Chem. Soc. 131 (2009)
70867093.
[63] M. Boronat, A. Corma, F. Illas, J. Radilla, T. Rdenas, M.J. Sabater, J. Catal. 278
(2011) 5058.
[64] S. Carrettin, P. McMorn, P. Johnston, K. Grifn, G.J. Hutchings, Chem. Commun.
(2002) 696697.
[65] Ch.-Y. Shiau, S. Chen, J.C. Tsai, S.I. Lin, Appl. Catal. A 198 (2000) 95102.
[66] T. Mitsudome, Y. Mikami, K. Ebata, T. Mizugaki, K. Jitsukawa, K. Kaneda, Chem.
Commun. (2008) 48044806.
[67] G.S. Pozan, I. Boz, M. Ali Gurkaynak, Chem. Biochem. Eng. Q 21 (2007)
235240.
[68] S. Carrettin, P. McMorn, P. Johnston, K. Grifn, Ch.J. Kiely, G.J. Hutchings, PCCP
5 (2003) 13291336.
[69] B.N. Zope, D.D. Hibbitts, M. Neurock, R.J. Davis, Science 330 (2010) 7478.
[70] A. Villa, G.M. Veith, L. Prati, Angew. Chem. Int. Ed. 49 (2010) 44994502.
[71] G.L. Brett, Q. He, C. Hammond, P.J. Miedziak, N. Dimitratos, M. Sankar, A.A.
Herzing, M. Conte, J.A. Lopez-Sanchez, Ch.J. Kiely, D.W. Knight, S.H. Taylor, G.J.
Hutchings, Angew. Chem. Int. Ed. 50 (2011) 1013610139.
[72] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Industrial Catalysis Processes,
John Wiley & Sons, Inc., 2005.
[73] J.M. Tatibouet, Appl. Catal. A: Gen. 148 (1997) 213252.
[74] M.J. Lippits, R.R.H. Boer Iwema, B.E. Nieuwenhuys, Catal. Today 145 (2009)
2733.
[75] K. Khler, M.C. Holz, M. Rohe, A.C. van Veen, M. Muhler, J. Catal. 299 (2013)
162170.
[76] A.N. Pestryakov, V.V. Lunin, N. Bogdanchikova, O.N. Temkin, E. Smolentseva,
Fuel 110 (2013) 4853.
[77] J.J. Spivey, Ind. Eng. Chem. Res. 26 (1987) 21652180.
[78] S. Minico, S. Scire, C. Crisafulli, S. Galvagno, Appl. Catal. B: Environ. 34 (2001)
277285.
[79] S. Scire, S. Minico, C. Crisafulli, C. Satriano, A. Pistone, Appl. Catal. B: Environ.
40 (2003) 4349.
[80] I. Sobczak, K. Szrama, R. Wojcieszak, E.M. Gaigneaux, M. Ziolek, Catal. Today
187 (2012) 4855.
[81] G. Busca, A.S. Elmi, P. Forzatti, J. Phys. Chem. 91 (1987) 52635269.
[82] X. Gao, I.E. Wachs, M.S. Wong, J.Y. Ying, J. Catal. 203 (2001) 1824.
[83] K. Routray, W.Ch. Zhou, J. Kiely, I.E. Wachs, ACS Catal. 1 (2011) 5466.
[84] I.X. Green, W.J. Tang, M. Neurock, J.T. Yates, Science 333 (2011) 736739.
[85] R. Si, J. Tao, J. Evans, J.B. Park, L. Barrio, J.C. Hanson, Y. Zhu, J. Hrbek, J.A. Rodriguez,
J. Phys. Chem. C 116 (2012) 2354723555.

You might also like