Supported Pd-Cu catalysts in the water phase reduction of nitrates: Functional resin versus alumina

Article abstract of DOI:10.1016/j.molcata.2006.08.081

Pd-Cu bimetallic catalysts supported on the cationic resin Dowex 1 × 4 (gel type poly(styrene-co-divinylbenzene) with -N(CH₃)₃⁺Cl^- groups) and basic γ-Al₂O₃ were prepared by ion exchange and the incipient wetness impregnation method, respectively, and tested in the liquid-phase hydrogenation of nitrates in water. Various methods of the reduction of metal precursors were used. The effect of the preparation method, the temperature, the reduction conditions and the support on the properties of palladium-copper composites was investigated by means of X-ray powder diffraction (XRPD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and temperature programmed reduction (TPR). The Pd-Cu catalysts prepared by liquid-phase reduction of anionic chloro-complexes of palladium(II) and copper(II) immobilised in a resin, mainly isolated domains of metallic palladium and copper, if any, contained. The reduction using hydrogen joined with calcination of palladium and copper salts supported on γ-alumina led to well formed crystallites of a palladium-copper alloy. The Pd-Cu catalysts supported on γ-Al₂O₃ exhibited higher activity of nitrate removal under given reaction conditions in a batch reactor. However, with the Pd-Cu resin-based catalysts a better final selectivity (selectivity to products other than nitrites and ammonia) at a comparable conversion of nitrates has been obtained in comparison with alumina based catalysts. The catalysts exhibited different stability with respect to the leaching of metals.

Full text of DOI:10.1016/j.molcata.2006.08.081

Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
Supported Pd–Cu catalysts in the water phase reduction of nitrates:
Functional resin versus alumina
a
a,b,∗
, Milan Hroneca,
Dana Ga sˇ parovi cˇ ov a´ , Milan Kr a´ lik
a
c
d
Zuzana Vallu sˇ ov a´ , Hannelore Vinek , Benedetto Corain
a
Institute of Organic Chemistry, Catalysis and Petrochemistry, Division of Organic Technology, Slovak University of Technology,
Radlinskeho 9, SK-812 37 Bratislava, Slovak Republic
b
VUCHT j.s.c., Nobelova 34, SK-836 03 Bratislava, Slovak Republic
Institute of Materials Chemistry, Vienna University of Technology, Veterin a¨ rplatz 1, A-1210 Vienna, Austria
c
d
Dipartimento di Scienze Chimiche, Via Marzolo 1, 35131 Padova, Italy
Received 29 March 2006; received in revised form 25 August 2006; accepted 29 August 2006
Available online 3 September 2006
Abstract
+
−
Pd–Cu bimetallic catalysts supported on the cationic resin Dowex 1 × 4 (gel type poly(styrene-co-divinylbenzene) with −N(CH
3 3
) Cl groups)
and basic ␥-Al were prepared by ion exchange and the incipient wetness impregnation method, respectively, and tested in the liquid-phase
2 3
O
hydrogenation of nitrates in water. Various methods of the reduction of metal precursors were used. The effect of the preparation method, the
temperature, the reduction conditions and the support on the properties of palladium–copper composites was investigated by means of X-ray powder
diffraction (XRPD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and temperature programmed reduction (TPR).
The Pd–Cu catalysts prepared by liquid-phase reduction of anionic chloro-complexes of palladium(II) and copper(II) immobilised in a resin, mainly
isolated domains of metallic palladium and copper, if any, contained. The reduction using hydrogen joined with calcination of palladium and copper
2 3
salts supported on ␥-alumina led to well formed crystallites of a palladium–copper alloy. The Pd–Cu catalysts supported on ␥-Al O exhibited
higher activity of nitrate removal under given reaction conditions in a batch reactor. However, with the Pd–Cu resin-based catalysts a better final
selectivity (selectivity to products other than nitrites and ammonia) at a comparable conversion of nitrates has been obtained in comparison with
alumina based catalysts. The catalysts exhibited different stability with respect to the leaching of metals.
©
2006 Elsevier B.V. All rights reserved.
Keywords: Pd–Cu catalyst; Cationic resin; ␥-Alumina; Nitrates; Water treatment
−
2
tive conversion of NO3 to N and water, without formation of
1
. Introduction
−
+
NO2 (partial reduction) and NH4 (overreduction). The limit
for ammonia in drinking water is 0.5 mg/l. The formation of
ammonia strongly depends both on chemical properties of the
catalyst used and on mass-transport phenomena.
Removal of low-concentrated nitrates from water represents
a serious ecological problem. The content of nitrates in water
can be lowered by several methods, such as ion exchange, elec-
trodialysis, reverse osmosis or microbiological treatment [1].
The catalytic reduction of nitrates to nitrogen with hydrogen
is an alternative way to obtain drinking water with the content
of nitrates fulfilling hygienic limit (50 mg/l). Therefore, it is a
subject of extensive investigation [2–17]. The target is the selec-
For this purpose, bimetallic Pd–Cu and Pd–Sn catalysts on
inorganic supports (e.g. Al O , pumice) apparently offer the
2
3
most promising prospects [4–6]. Pintar and Kajiuchi [4] have
achieved the highest selectivity of nitrogen formation (91% at
total conversion of nitrate) over the Pd–Cu/␥-Al O catalyst.
2
3
These authors found a strong dependence of selectivity of nitro-
gen formation on the preparation method [5]. Pr u˝ sse et al. [7]
have reported that Pd–Sn/Al2O3 catalysts exhibit comparable
activity with that of Pd–Cu/Al2O3 catalysts and higher selectiv-
ity towards nitrogen. Opposite results have been found for the
zirconia and titania supported Pd–Sn and Pd–Cu catalysts [8].
∗
Correspondingauthorat:InstituteofOrganicChemistry, CatalysisandPetro-
chemistry, Division of Organic Technology, Slovak University of Technology,
Radlinskeho 9, SK-812 37 Bratislava, Slovak Republic. Tel.: +421 2 5249 5242;
fax: +421 2 5249 3198.
E-mail addresses: milan.kralik@stuba.sk, mkralik@vucht.sk (M. Kr a´ lik).
1
381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.08.081

9
4
D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
Previous studies also showed that an additional factor strongly
influencing the properties of the catalyst is the type of the support
from an aqueous solution of Cu(NO3)2 and Pd(NO3)2 (0.7 ml
of distilled water/g ␥-Al2O3). In the first step, the support was
◦
[
8–13].
Palladium and copper can form alloys in the whole con-
impregnated with copper nitrate, dried 12 h at 50 C and 12 h
◦
at 100 C and reduced pressure (10 kPa). Then it was calcined
in air at 500 C for 1 h, gradient 10 C/min, and subsequently
impregnated with palladium nitrate. The obtained material was
dried 12 h at 50 C and 12 h at 100 C and reduced pressure
(10 kPa) and calcined in air at 500 C for 3 h. The concentrations
◦
◦
centration range; however, formation of an alloy depends on
the procedure for the generation of metal atoms. Temperature,
reduction conditions, interactions with a support and spatial
hindrance against migration of atoms are the main parameters
influencing the formation of alloys and/or isolated pure metal
domains, which can be of various size and at various distances
◦
◦
◦
of Pd and Cu metals were 3.90 and 0.98 wt.%, respectively. The
catalysts precursors were reduced with hydrogen in the liquid
◦
[
18]. The resulting structure of the palladium–copper composite
phase in methanol for 1 h at 25 C and 0.5 MPa (cat 3) or in the
◦
strongly affects the catalytic properties.
gas phase for 1 h at 500 C (cat 4).
We report in this paper the effect of (i) the support, (ii)
the preparation method and (iii) the composition of the Pd–Cu
catalyst on the activity and selectivity in the liquid-phase hydro-
genation of nitrates. The study was focused on a gel type cationic
resin and a generally used classical support such as ␥-alumina.
The prepared catalysts were characterised by means of X-
ray microprobe analysis (XRMA), X-ray powder diffraction
(XRPD), scanning electron microscopy (SEM), transmission
electron microscopy (TEM), nitrogen adsorption/desorption
isotherms(BET)andtemperatureprogrammedreduction(TPR).
X-ray powder diffraction patterns were obtained using the
Diffractometer PHILIPS (HP generator PW 1730/goniometer
PW 1050) with an X-ray source Cu K␣; λ = 0.154 nm and with
a Ni filter. To prevent the surface oxidation of copper in con-
tact with air [14], samples of the reduced catalysts were dried
2
. Experimental
2
.1. Catalysts preparation and characterisation
◦
in nitrogen atmosphere at 60 C and embedded in a poly(methyl
All chemicals were of analytical grade purity, supplied by
methacrylate) matrix. X-ray microscope analysis and scanning
electron microscopy were performed by a Cambrige Stereoscan
either Fluka (Dowex 1 × 4, CuCl2·2H2O, Pd(NO3)2·2H2O,
Cu(NO3)2·3H2O), Sigma Aldrich (␥-Al2O3), Safina Jesenice
250 EDXPW 9800. Image analysis was carried out with an
(
PdCl2, 40 wt.% solution) or Lachema Brno (others). The
Image Pro Plus Program. Transmission electron micrographs of
catalysts were obtained with a JEOL 2010 electron microscope
with GIF. Water swollen samples of the analysed catalysts were
ground with pestle and mortar. A drop of the obtained suspen-
sion was deposited onto a carbon grid and the resin let to dry
in air before the analysis. BET nitrogen adsorption/desorption
Dowex 1 × 4 microgel resin (particle size 100–200 mesh,
tetraalkylammonium groups 4.1 mmol/g ion exchange capacity)
2
and basic ␥-Al2O3 (ALOX560C; BET surface area 155 m /g,
average pore diameter 5.8 nm, particle size 150 mesh) were
employed as carriers of active components.
The Dowex 1 × 4 gel-type resin was washed with 1 M KCl,
◦
◦
isotherms were measured at −196 C with the Sorptomatic 1900
distilled water, and cold methanol and finally dried at 50 C
apparatus. TPR measurements were carried out in a quartz flow
at reduced pressure (10 kPa). The ion exchange capacity of the
resinwas4.18 mmol/g. ThechlorideformofDowex1 × 4anion-
exchanger (3.8 g) was metallated by ion exchange (3 h, mild
stirring) using 20 ml of a solution of PdCl2 (1.1109 g, 40 wt.%
solution in HCl/H2O) and CuCl2 (0.1789 g) in 4 M aqueous HCl
reactor using a gas feed stream of 5% H2 in He, total flow
◦
3
0 ml/min, with a temperature ramp 10 C/min. The hydrogen
consumption was measured with a quadrupole as an analyser.
2.2. Catalytic tests
(
13.33 g Pd/l and 3.33 g Cu/l) according to the following reac-
tions:
Catalytic tests were carried out in a thermostatic batch reac-
tor, four-necked flask (volume, 500 cm ) equipped with a KPG
+
−
2−
2
(P)-N(CH3)3 Cl (s) + [PdCl4]
3
{(P)-N(CH3)3 }2[PdCl4]2−(s) + 2Cl−
+
stirrer, inlets for hydrogen-nitrogen (45 mol% of H ) and 0.1 M
2
→
(1)
(2)
HCl supply, a capillary for sampling and off-gas. The flow rate
of the H2–N2 mixture was 64 ml/min and the total pressure
bar. Hydrochloric acid was fed throughout the catalytic exper-
+
−
−
(
P)-N(CH3)3 Cl (s) + [CuCl3]
1
+
−
−
◦
iments to maintain the pH at 5–6. A typical starting reaction
→
(P)-N(CH3)3 [CuCl3] (s) + Cl
3
mixture consisted of 0.25 dm KNO3 aqueous solution contain-
3
−
The obtained material was dried at 50 C under reduced pres-
ing 100 mg/dm NO3 and 0.15 g of catalyst. The catalytic runs
◦
sure to constant weight. The loading of resins was 4.10 and
.95 wt.% of palladium and copper, respectively. The cationic
were carried out at 25 C. Samples were periodically collected
0
with sampling intervals ranging from 5 to 30 min. After the cat-
alytic test the used resin supported catalyst was treated with
0.25 M NaCl to displace any nitrate and nitrite ions entrapped
inside the support. In the solution recovered after the restora-
metalled resins were reduced with hydrogen in 2.5% water solu-
tion of Na2CO3 for 1 h at 25 C and 0.5 MPa (cat 1) or with
◦
0.066 M solution of sodium borohydride in ethanol for 1 h at
−
ambient temperature under occasional stirring (cat 2).
Pd–Cu catalysts on the inorganic support—␥-Al2O3 were
prepared by incipient wetness successive impregnation, starting
tion of the catalyst Cl form, the contents of nitrates and nitrites
trapped inside the catalyst were determined in the same way as
for the reaction mixture.

D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
95
2
.3. Analysis
The determination of nitrates, nitrites and ammonia was per-
Table 1
Codes and relevant properties of the prepared Pd–Cu catalysts
Catalyst
Support
Reduction
Distribution
dXRPD
(nm)
−
of metalsa
formed by VIS absorption spectrophotometry [19]. NO3 ions
were determined with salicylic acid in the presence of sulfuric
acid (maximum absorption at 410 nm), NO2 ions via reaction
with sulfanilic acid and N-(1-naphtyl)ethylenediamine dichlo-
ride(maximum absorption at550 nm)and NH4 ions by reacting
with Nessler’s reagent (maximum absorption at 425 nm). The
samples of water were analysed for palladium and copper by
atomicabsorptionspectroscopy(AAS)onaCarlZeissJenaAAS
3
3
3
cat 1
Dowex 1 × 4
H2, 0.5 MPa,
Na CO ,
P
3.8
−
2
3
b
H2O
+
cat 2
cat 3
Dowex 1 × 4
␥-Al2O3
NaBH4, EtOH
H2, 0.5 MPa,
MeOH
E
VP
n.d.
8.2
◦
cat 4
␥-Al2O3
H2, 500 C
VP
14.2
a
E: even; P: peripheral; VP: very peripheral.
atomic absorption spectrometer.
b
2.5 wt.% aqueous solution of Na2CO3.
. Results and discussion
XRMA analysis of the Pd–Cu resin supported catalysts
.1. Catalysts preparation and characterisation
(Fig. 3) reveals either a non-uniform (peripheral) distribution of
metals and higher concentration of metals (mainly palladium)
on the surface (reduction with hydrogen in the presence of 2.5%
aqueous solution of Na2CO3), or an uniform distribution (reduc-
tion with NaBH4 solution in ethanol). A schematic description
of the properties of the prepared catalysts is given in Table 1.
The metal deposition on ␥-Al2O3 during the preparation of
the catalyst (cat 4) causes a decrease in the specific surface area
The ion-exchange resin Dowex 1 × 4 was supplied in the
form of regular spherical, ca. 100 ␮m in diameter beads. The
ion-exchange process and reduction with NaBH4 did not cause
any change of the shape of the particles, as shown by SEM
micrographs. On the contrary, a morphological change in the
material after the reduction with hydrogen at higher pressure
2
(
0.5 MPa) (Fig. 1a) was observed. Light areas on the scanning
electron micrograph of the Pd–Cu alumina supported catalyst
Fig. 2a) reveal heavy elements. It is notable that besides large
of the material (BET method) from 163 to 140 m /g.
TPR measurements of the alumina supported catalysts were
carried out using 5 vol.% H2 in He. Due to a low difference
in thermal conductivity combination of hydrogen with helium
as a carrying gas is supposed to be less appropriate than with
(
agglomerates, little uniformly distributed metal particles on the
alumina surface are also present.
−
◦
Fig. 1. Scanning electron micrographs of the Pd–Cu catalyst supported on Dowex Cl reduced with H2 in aqueous Na2CO3 (2.5%, w/w) at 25 C and 0.5 MPa (cat
1
◦
) (a), and of the Pd–Cu alumina supported catalyst reduced with H2 at 500 C (cat 4) (b).
◦
Fig. 2. Scanning electron micrograph (a) and EDX analysis (b) of the Pd–Cu alumina supported catalyst reduced with hydrogen at 500 C (cat 4).

9
6
D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
◦
Fig. 3. XRMA analysis of Pd–Cu catalysts supported on Dowex 1 × 4, reduced with (a) H2 in aqueous Na2CO3 (2.5%, w/w) at 25 C and 0.5 MPa (cat 1); (b) sodium
borohydride (0.066 M solution in ethanol) (cat 2).
other inert gases, e.g. argon or nitrogen [14,20–24]. However,
using MS analyser the TPR carried out with H2/He can pro-
vide quite reliable results. Combination of H2/He was also used,
e.g. by Fernadez-Garcia et al. [25] for studying Pd–Cu alloys.
Thus, impression of lower quality of the obtained TPR curves
mina and together with relatively large particles (about 47 nm)
enables a huge formation of beta hydride. The reduction of the
monometallic Cu catalyst was possible only above 300 C. In
◦
comparison with other authors, e.g. Gauthard et al. [27], only
(
Fig. 4), especially that for palladium, could be ascribed to pecu-
liar features of the studied system (metal support interaction)
26]. This generates a TPR curve from which it is not possible
[
to distinguish the reduction stage of palladium oxide (it starts
at room temperature) and the decomposition of beta-hydride (a
◦
typical temperature is about 60 C [21,22]). Such complicated
behaviours were also observed by Fernandez et al. [22] in case
of Pd/alumina catalysts. When the calcined palladium catalyst
◦
was treated by He at 300 C, cooled and immediately reduced
◦
a small inverse peak at 70 C was registered. This observation
allows us to state that a strong interaction between metallic-
oxide species and support still exists. Relatively small particles
(
about 14 nm, see bellow) also do not favour a large formation
of palladium beta-hydride. A nice example of identification of
beta-hydride has been given by Karski et al. [23] who oxidised
the reduced Pd/SiO2 catalyst and then again measured the TPR
curves. Inthiscasesilicarepresentsamoreinertsupportthanalu-
Fig. 4. TPR profiles of the ␥-alumina supported catalyst precursors: PdO–CuO
(
a), CuO (b), PdO (c).

D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
97
one peak was registered, i.e. the reduction of copper oxide, and
no reduction of cuprous oxide. The Pd–Cu catalyst precursor
◦
showed one large band centred at 170 C that can be assigned to
the reduction of Cu oxides promoted by the presence of the Pd
metal [8,14,21,24].
Similarly with the conclusions reported by Roveda et al. [28],
owing to the low thermal stability of the Dowex supports (max-
◦
imally 150 C), TPR measurements were not possible for cat 1
and cat 2.
The presence of crystalline Pd- or/and Cu-containing
−
microphases within the Pd–Cu/Dowex Cl (cat 1 and cat 2) and
Pd–Cu/␥-Al2O3 (cat 3 and cat 4) samples was investigated by
-
XRPD (Figs. 5 and 6). The Pd–Cu/Dowex Cl sample reduced
with NaBH4 (cat 2) is very well dispersed; the particle sizes of
the Pd- or Cu-containing phases were too small to be detected
by XRPD, or they were amorphous (Fig. 5). This agrees with
a similar observation reported in the paper by Palomares et al.
−
[
15]. In the Pd–Cu/Dowex Cl sample reduced with H2 (cat 1),
the X-ray powder diffraction analysis showed a peak of metal-
◦
lic palladium(1 1 1) at 40.1 and no characteristic peak for the
Cu-containing phase was observed. The size of palladium crys-
tallites in this catalyst is estimated to be 3.8 nm (Table 1).
In the Pd–Cu ␥-alumina supported samples (cat 3) and (cat
◦
4
) (Fig. 6), the shift of the Pd(1 1 1) diffraction peak (40.1 )
to higher angles (in comparison with the position exhibited by
palladium nanoclusters) gives a clear evidence of Pd–Cu alloy
formation [29,30]. As the found diffraction signals do not fit with
any of those signals corresponding to Pd–Cu ordered alloys, it
may be concluded that disordered alloys are formed during the
reduction procedure. This statement is in agreement with reports
given in the literature [25,29]. The shift in the peak position is
consistent with Vegard’s Law that states that the lattice param-
eter of a binary solid solution is directly proportional to atomic
percentage of the alloy and can be used to monitor the relative
composition of an alloy [30]. As it is apparent from Figs. 5 and 6,
Fig. 6. X-ray powder diffraction patterns for ␥-alumina (a), ␥-alumina support-
ing palladium (b) and palladium–copper species reduced with H2 at 500 C
◦
(cat 4) (c); ␥-alumina supporting palladium (d) and palladium–copper species
◦
reduced with hydrogen in methanol at 25 C and 0.5 MPa (cat 3) (e).
◦
the calcination at 500 C after the deposition of metal precursors
allowed obtaining of the Pd–Cu alloy despite the reduction in
◦
the liquid phase at 25 C (cat 3).
−
TEM analysis of the Pd–Cu/Dowex Cl catalyst cat 1
(Fig. 7a) reveals that the nanometer sized metal crystallites
appears to be “glued” by a metal-containing matrix. We specu-
late that Pd nanoclusters are embedded inside the amorphous
II,I
20–100 nm-sized globular structures of Cu
oxo-hydroxi
species, whose structure might be including also subnanometric
0
Cu particles.
On the contrary, TEM analysis of the ␥-alumina supporting
palladium–copper catalyst cat 4 reveals larger bimetallic par-
ticles 21 ± 17 nm (Figs. 7b and 8), which is comparable with
an average size of 14 nm calculated from the XRPD patterns
(Table 1). It is worth to note that in spite of partial leaching of
metals from the Pd–Cu/␥-Al2O3 catalysts (see below) no sig-
nificant change has been monitored either from TEM or XRPD
measurements.
−
Fig. 5. X-ray powder diffraction patterns for the Dowex Cl support-
ing palladium–copper species reduced with hydrogen in aqueous Na2CO3
3.2. Catalytic tests
◦
−
(
2.5%, w/w) at 25 C and 0.5 MPa (cat 1) (a), the Dowex Cl support-
ing palladium–copper species reduced with sodium borohydride (0.066 M in
ethanol) (cat 2) (b). (A diffraction line at 60 is from a metal sample holder.
◦
Results from catalytic tests in the reduction of nitrates to
nitrite, nitrogen and ammonium ion are reported in Figs. 9–12.
The marked decrease in the concentration of nitrate at the
Intensity increases with increasing of transmittance of a specimen and higher
counting of signal.)

9
8
D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
◦
Fig. 7. Transmission electron micrographs of the Pd–Cu/Dowex 1 × 4 catalyst reduced with H2 in aqueous Na2CO3 (2.5%, w/w) at 25 C and 0.5 MPa (cat 1) (a)
◦
and the Pd–Cu/␥-Al2O3 catalyst reduced with H2 at 500 C (cat 4) (b) after catalytic tests (scale for (b) is approximately two times larger).
−
beginning of the reaction over Pd–Cu/Dowex Cl catalysts
(
fact is sintering of the small metal clusters into larger agglom-
erates which is favoured at elevated reduction temperature (see
Table 1).
Fig. 9) is caused by intraparticle ion exchange [31].
Higher activities in nitrate hydrogenation were obtained with
Pd–Cu/␥-Al2O3 supported catalysts. As outlined above, forma-
tion of Pd–Cu alloy was observed for this kind of catalysts. In
accordance with the hypothesis about the reaction mechanism
Cat 1 was the best one from the Dowex type catalysts tested.
A higher catalytic activity was exhibited by the Pd–Cu/␥-Al2O3
obtained by the reduction in the liquid phase with hydrogen
◦
[
11,13,14,32] the interfaces between the two metal phases, or the
at 25 C (cat 3). The reason for this higher catalytic activity
−
presence of an alloy, play a key role in the reduction of NO3 .
The other factor important for an increase in the catalytic activity
is a higher accessibility of metal centres due to their location on
the outer surface of the support particles. The Pd–Cu/␥-Al2O3
is probably a better accessibility of catalytic centres, which are
located mainly on the outer surface. Apparently, the effect of the
higher resistance against mass transport in the case of uniformly
◦
bimetallic catalyst reduced at 500 C (cat 4) exhibits the highest
activity among all tested catalysts (Fig. 12). A not negligible
◦
Fig. 9. Nitrate concentration in water vs. time for nitrate reduction (25 ) over the
Pd–Cu catalysts (4 wt.% Pd, 1 wt.% Cu), 0.15 g of a catalyst; (ꢀ) Pd–Cu/Dowex
◦
1
× 4 reduced with H2 in aqueous Na2CO3 (2.5%, w/w) at 25 C and 0.5 MPa
Fig. 8. Histogram of particle sizes of the Pd–Cu/␥-Al2O3 catalyst (cat 4) deter-
mined from TEM pictures (250 particles, an average of maximum and minimum
dimensions of a particle was used).
(cat 1), (᭹) Pd–Cu/Dowex 1 × 4 reduced with sodium borohydride, 0.066 M in
◦
ethanol (cat 2), (ꢀ) Pd–Cu/␥-Al2O3 reduced with H2 in methanol at 25 C and
◦
0.5 MPa (cat 3) and (ꢀ) Pd–Cu/␥-Al2O3 reduced with H2 at 500 C (cat 4).

D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
99
respectively. Table 2 reveals data of experiments obtained with
the Pd–Cu/Dowex 1 × 4 catalyst reduced with hydrogen (at
0.5 MPa at the presence of Na2CO3), i.e. catalyst cat 1, and
data on the experiment with the Pd–Cu/Dowex 1 × 4 catalyst
cat 2 (reduction with NaBH4).
As seen from Table 2, replacing of HCl with CO2 leads to
a drop of the apparent catalytic activity, in both cases, which is
accompanied by a higher N2 selectivity. The decrease in activ-
ity in the experiments with CO2, especially in the case cat 1
(larger metal nanoclusters) is supposed to be caused by the for-
mation of a scarcely soluble CuCO3·Cu(OH)2 (stoichiometry of
malachite), which lowers the number of active copper sites. The
increase of the partial pressure of hydrogen in the presence of
CO2 has no effect on the experiments with cat 1. However, in
◦
Fig. 10. Nitriteconcentrationinwatervs. timefornitratereduction(25 ) over the
the case of cat 2 at relatively high H pressure over-reduction is
2
Pd–Cu catalysts (4 wt.% Pd, 1 wt.% Cu), 0.15 g of a catalyst; (ꢀ) Pd–Cu/Dowex
favoured [31]. Decrease in the catalytic activity is also connected
◦
1
× 4 reduced with H2 in aqueous Na2CO3 (2.5%, w/w) at 25 C and 0.5 MPa
−
with the above-discussed competitive adsorption of HCO3 and
(
cat 1), (᭹) Pd–Cu/Dowex 1 × 4 reduced with sodium borohydride, 0.066 M in
−
◦
NO3 , because they have a very similar geometric structure and
ethanol (cat 2), (ꢁ) Pd–Cu/␥-Al2O3 reduced with H2 in methanol at 25 C and
◦
0
.5 MPa (cat 3) and (ꢀ) Pd–Cu/␥-Al2O3 reduced with H2 at 500 C (cat 4).
distribution of electronic charge [33].
The substitution of CO2 for HCl in experiments with alumina
changed the activity, but not significantly. Opposite directions
of these changes (increase in the case of cat 3 and decrease in
the case of cat 4) are caused probably by a different composi-
tion of alloys and size of the agglomerates as was deduced from
XRPD measurements (see above). Taking into account adsorp-
tion of the both HCO3 and NO3 anions, a lower covering
of the catalytic surface with hydrogen has to be considered.
This contributes to a lower number of hydrogen moieties with
potential to “over-reduction”. Consequently the formation of
ammoniaisdecreasedandtheselectivitytonitrogenisincreased.
An increase in the partial pressure of hydrogen had a positive
effect on the activity of the catalyst supported on alumina, but a
negative influence on the final selectivity (Table 3). These obser-
vations are in accordance with a concept about the mechanism
of catalytic nitrate reduction, in which a higher concentration of
chemisorbed hydrogen, increases the rate of nitrate removal. On
the other hand, the higher concentration of chemisorbed hydro-
distributed metal particles throughout the catalyst (cat 2) results
in an appreciably lower catalytic activity.
On the other hand, a comparison of the behaviour of cata-
lysts at approximately 44% (and/or 65%) conversion of nitrates
reveals that the resin supported Pd–Cu catalysts are more selec-
tive than the alumina supported catalysts. In fact, the Pd–Cu/␥-
Al2O3 catalysts cat 3 and cat 4 provide selectivity to products
other than nitrite and ammonia (equal to 78.8 and 79.2%). The
−
−
−
Pd–Cu/Dowex Cl catalyst cat 2 exhibits selectivity to the for-
mation of nitrogen equal to 93.8% (at nitrate conversion 44%).
−
+
The selectivities to N2 (σN ), NO2 (σNO2
−
) and NH4 (σ
+
)
2
NH4
formation in the hydrogenation of nitrates were quantified on
the basis of the expressions described in the literature [31].
The effect of various hydrogen pressures and ways for main-
tainingpH(CO2 insteadofHCl)isdocumentedinTables2and3,
◦
Fig. 11. Ammonia concentration in water vs. time for nitrate reduction (25 )
over the Pd–Cu catalysts (4 wt.% Pd, 1 wt.% Cu), 0.15 g of a catalyst; (ꢀ)
◦
Pd–Cu/Dowex 1 × 4 reduced with H2 in aqueous Na2CO3 (2.5%, w/w) at 25 C
and 0.5 MPa (cat 1), (᭹) Pd–Cu/Dowex 1 × 4 reduced with sodium borohydride,
Fig. 12. Final nitrate conversion and selectivity to N2, NO2 and NH4+ ions
−
0
2
.066 M in ethanol (cat 2), (ꢁ) Pd–Cu/␥-Al2O3 reduced with H2 in methanol at
◦
◦
◦
5 C and 0.5 MPa (cat 3) and (ꢀ) Pd–Cu/␥-Al2O3 reduced with H2 at 500 C
in the hydrogenation of nitrates in water (25 C) after experiments with cat 1–4
(
cat 4).
catalysts.

1
00
D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
Table 2
Final nitrate (mNO3
−
) and nitrite content (m
−
) in the Pd–Cu catalysts used, concentration of Cu dissolved in the treated water (cCu), nitrate conversion (x
−
)
NO3
NO2
−
+
a
and selectivity (σ) to N2, NO2 and NH4 over the Pd–Cu/Dowex 1 × 4 catalysts (4 wt.% Pd, 1 wt.% Cu)
Catalyst; reductant; adjustment of pH
m
−
(mg)
m
−
(mg)
cCu (mg/l)
x
−
NO3
(%)
σ (%)
NO3
NO2
−
NH4+
N2
NO2
b
cat 1; H2–N2 ; 0.1 M HCl
2.5
10.5
10.9
7.2
9.7
11.7
0.227
0.036
0.016
0
0
0
0
0
0
0
0
0
83
50
49
61
52
43
71
89
89
94
96
82
2.2
0.4
0.2
0
0
0
26.8
10.6
10.8
6
4
18
b
c
cat 1; H2–N2 ; CO2
c
cat 1; H2; CO2
b
cat 2; H2–N2 ; 0.1 M HCl
b
c
cat 2; H2–N2 ; CO2
cat 2; H2; CO2c
a
−
◦
2
4
50 ml of distilled water with 100 mg/l NO3 (KNO3), 0.5 g of catalyst, H2–N2 (H2) 64 ml/min, 1 bar, 25 C, pH 6, time 135 min.
3 mol% of H2.
b
c
Separate inlet of CO2.
gen contributes to isolation of nitrogen atoms, which are reacting
with neighbouring hydrogen atoms and form ammonia.
An important feature of the catalytic process is the lifetime of
the catalyst and leaching of metals. As to the lifetime, one could
expect that little metal particles inside the polymeric support
would be more easily dissolved than larger Pd–Cu compos-
ite particles located on the surface of ␥-alumina. However, the
opposite is true. Analysis of copper and palladium in the reac-
tion mixture from catalytic tests with resin supported catalysts
were under the detection limit. Such a stability of the resin type
catalysts could be assigned to the protective effect of the resin
framework. This protective effect implicates an intimate coex-
istence of metallic palladium and copper, as well as an oxidised
form of copper, which are supposed to be intermediates in the
reduction of nitrates. Gao et al. [13] have reported the following
equation:
concentration of hydroxyl anions, which are one of the products
in course of the reduction of nitrates. Hence, circumstances for
the reduction of copper(I) to copper(0) are favoured.
The alumina supported catalysts undergo a low, but apprecia-
ble leaching of copper into the treated water (Table 3). No leach-
ingof palladium is observed. In comparisonwiththe Dowex type
catalysts (cat 1 and cat 2), a different environment of catalytic
centres (higher acidity, especially in the case of CO2 as a buffer-
ing agent) causes worse conditions for the reduction of copper(I)
to copper(0) and copper dissolves at higher extent. Although the
substitution of CO2 for HCl can improve the selectivity to nitro-
gen, the application of CO2 buffering in the experiments with
alumina catalysts has a negative impact for the leaching of cop-
per (see Table 3). It may be stated, that the addition of CO2
lowers the concentration of hydrogen on the surface of the alu-
mina catalyst (decrease in partial pressure of hydrogen, as well
−
as chemisorption of HCO3 anions) and the potentiality for the
NO3 + 2Cu ꢁ Cu2O + NO2−, E = 0.361 V
−
(3)
reduction of copper(I) to copper(0) is lowered. When a pure
hydrogen stream is applied instead of the mixture with nitrogen,
i.e. the partial pressure of hydrogen is higher, consequently the
concentration of hydrogen on the surface of alumina catalyst is
also higher. A higher concentration of hydrogen on the catalysts
surface enhances the reduction of copper(I) to copper(0) and the
leaching of copper has a lower extent (Table 3).
which describes a redox cycle in the alkaline solution. Accord-
ing to the proposed mechanism of the catalytic nitrate reduc-
tion [13,14], the formed copper(I) oxide can be reduced with
hydrogen chemisorbed on palladium adjacent to copper(I) into
metallic copper species:
The importance of a non-strong acid environment for the
stability of copper–palladium catalytic centres was evidently
proved by a precise analysis of the reaction mixture using
similar catalysts like those reported in paper [10]. The tests
Cu2O + H2(Pd) ꢁ 2Cu(Pd) + H2O
(4)
In acid solution, copper oxides are less stable and can dissolve
[
13]. Intraparticlespaceofthecationicresinensuresaratherhigh
Table 3
The highest concentration of nitrite (NO2 max), concentration of Cu dissolved in the treated water (cCu), activity at 50% conversion of nitrates (APd), nitrate conversion
−
+
(
x
) and selectivity (σ) to N2, NO2 and NH4 over Pd–Cu/␥-Al2O3 catalysts (4 wt.% Pd, 1 wt.% Cu)a
−
−
NO3
−
−
Catalyst; reductant; adjustment of pH
NO2 max (mg/l)
cCu (mg/l)
APd (mgNO3 /min gPd)
x
−
(%)
σ (%)
NO3
−
NH4+
N2, NOx
NO2
b
cat 3; H2–N2 ; 0.1 M HCl
10.5
0.9
7.3
1.3
2.7
0.014
0.018
0.013
0.019
0.014
46
62
79
58
139
91
95
98
95
35
70.4
40
78.5
36
5
0.1
1
0.2
0
60
29.5
59
21.3
64
b
c
cat 3; H2–N2 ; CO2
b
cat 4; H2–N2 ; 0.1 M HCl
b
c
cat 4; H2–N2 ; CO2
cat 4; H2; CO2c
100
a
−
◦
2
4
50 ml of distilled water with 100 mg/l NO3 (KNO3), 0.15 g of catalyst, H2–N2 (H2) 64 ml/min, 1 bar, 25 C, pH 6, time 220 min.
3 mol% of H2.
b
c
Separate inlet of CO2.

D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
101
were performed with an acid form of a catalyst prepared from
a gel type strong anionic resin supporting 2 and 0.5 wt.% of
palladium and copper, respectively. Leaching of copper equal
to 0.72% with respect to the original content of copper was
registered after 220 min of duty. Such a strong dissolution of
copper results from the strong acid environment created by
4. pH in the neighbourhood of catalytic centres. If it is too
low the reduction process of copper in the catalytic cycle
is slower, even metals may dissolve. A proper value of pH
seems to be about 5 [11,33].
5. A proper pressure of hydrogen, which has to be “tuned” in
such a way to ensure a sufficient rate of the reduction of
nitrogen species, stability of palladium–copper composite
and minimising the over-reduction of ammonia. The pressure
of hydrogen, about 0.05–0.1 MPa seems to be appropriate.
–
SO3H pendant groups. Hence, conditions for the reduction
process of copper(I) to copper(0) [13] are not satisfactory.
6
. Masstransportphenomena, whichinfluencethecatalyticeffi-
ciency, selectivity and leaching of metals.
4
. Conclusions
Bimetallic Pd–Cu catalysts based on the gel-type resin
Dowex 1 × 4 and on ␥-alumina as supports can be obtained
according to protocols that enable the formation of bimetallic
supported nanodomains (nanoalloy). The so obtained materials
Very complicated is just the last item. A certain optimal value
of the pressure of hydrogen is tightly connected with mass trans-
port from the bulk liquid to the catalytic surface and diffusion
inside a porous catalytic particle [31]. The combination of these
effects on the concentration of reaction species around the cat-
alytic centre exclude a simple evaluation of a process. In other
words, consideration of experimental data has to be presented
with respect to applied catalysts and reaction conditions. We
have tried to express such interpretation of data in this paper.
−
−
+
catalyse the hydrogenation of NO3 to N2, NO2 and NH4
with moderate activity and various selectivity in favour of nitro-
gen. In this connection, the most selective catalysts (71–96%)
are those based on the functional resin. In spite of smaller metal
particles generated inside the polymer framework, this type of
catalyst proved to be more stable than that supported on ␥-
alumina. The latter fact is explained by the effects of a special
environmentinsidetheswollenmatrix, wherethepalladiumcop-
per catalytic centres are located.
Acknowledgements
Despite a lot of work done in the field of the catalytic reduc-
tion of nitrates over bimetallic catalysts, it is rather difficult to
make some unambiguous conclusions about the role of alloys
and pure metals. Importance of an intimate contact between a
noble and not noble metals, e.g. palladium and copper has been
nicely demonstrated by Yoshinaga et al. [11]. Virtually 100%
conversion was obtained using an alloy Pd–Cu catalyst in com-
parison with only a 53% conversion obtained with a mechanical
mixture of metal palladium and metal copper catalysts. Simi-
larly, Epron [14] has stated that nitrates are totally and rapidly
reduced on Pt–Cu bimetallic catalysts when both metals are in
close contact. Thus, the interaction between copper and plat-
inum is of major importance, in order to maintain copper in the
metallic state by way of hydrogen adsorbed on platinum. Gau-
thard [27] also stresses the effects of particle size, the electronic,
and/or geometric aspects induced by the promoter, which could
explain the differences in selectivity as a function of the bimetal-
lic catalysts. Results obtained in our laboratories and discussed
above are in accordance with the cited authors. The following
list of factors positively influencing the catalytic activity and
selectivity to nitrogen can be put down:
This work was supported by the funds of the Slovak VEGA
projects No. 1/9142/02 and No. 1/2459/05. We acknowledge
Prof. Konrad Hayek from University Innsbruck for the TEM pic-
ture of the Pd–Cu alumina supported catalyst and Vladim ´ı r Jor ´ı k
from Institute of Inorganic Chemistry, Technology and Materi-
als, Slovak University of Technology for XRPD measurements.
References
[1] A. Kapoor, T. Viraraghavan, J. Environ. Eng. 123 (1997) 371.
[2] K.D. Vorlop, T. Tacke, Chem. Ing. Technol. 61 (1989) 836.
[
3] U. Pr u¨ sse, K.D. Vorlop, J. Mol. Catal. A: Chem. 173 (2001) 313.
[4] A. Pintar, T. Kajiuchi, Acta Chim. Slovenica 42 (1995) 431.
[5] J. Batista, A. Pintar, M. Cech, Catal. Lett. 43 (1997) 79.
[6] F. Deganello, L.F. Liotta, A. Macaluso, A.M. Venezia, G. Deganello, Appl.
Catal. B: Environ. 24 (2000) 265.
[
[
7] U. Pr u¨ sse, M. H a¨ hnlein, J. Daum, K.D. Vorlop, Catal. Today 55 (2000) 79.
8] G. Strukul, R. Gavagnin, F. Pinna, E. Modaferri, S. Perathoner, G. Centi,
M. Marella, M. Tomaselli, Catal. Today 55 (2000) 139.
[
9] A.J. Lecloux, Catal. Today 53 (1999) 23.
[10] D. Ga sˇ parovi cˇ ov a´ , M. Kr a´ lik, M. Hronec, Collect. Czech. Chem. Commun.
4 (1999) 502.
6
[
[
11] Y. Yoshinaga, T. Akita, I. Mikami, T. Okuhara, J. Catal. 207 (2002) 37.
12] A.E. Palomares, J.G. Prato, F. Marquez, A. Corma, Appl. Catal. B: Environ.
41 (2003) 3.
1
. Combination of domains of pure palladium (suitable for the
reduction of nitrites) and palladium–copper composites—
even alloys (necessary for the sufficient reduction of nitrates
to nitrites). The size of these domains in nanometers.
. Structure of a support, which prevents migration of metal
particles, i.e. either mesoporous structure, e.g. advised by
Vorlop et al. [2,3], or polymer network.
[
13] W. Gao, N. Guan, J. Chen, X. Guan, R. Jin, H. Zeng, Z. Liu, F. Zhang,
Appl. Catal. B: Environ. 46 (2003) 341.
[
[
[
14] F. Epron, F. Gauthard, C. Pin e´ da, J. Barbier, J. Catal. 198 (2001) 309.
15] A.E. Palomares, J.G. Prato, F. Rey, A. Corma, J. Catal. 221 (2004) 62.
16] G. Centi, S. Perathoner, Appl. Catal. B: Environ. 41 (2003) 15.
2
3
[17] J. S a´ , D. Ga sˇ parovi cˇ ov a´ , K. Hayek, E. Halwax, J.A. Anderson, H. Vinek,
Catal. Lett. 105 (2005) 209.
[
18] V. Ponec, G.C. Bond, in: B. Delmon, J.T. Yates (Eds.), Catalysis by Metals
and Alloys, Elsevier Science B.V, Amsterdam, 1995.
. Location of catalytic centres. The catalytic centres have to be
located either close to the surface in case of a porous material,
or non-uniform distribution throughout the catalytic particle.
[
19] M. Hor a´ kov a´ , P. Lischke, A. Gr u¨ nwald, Chemick e´ a fyzik a´ lne met o´ dy
anal y´ zy vod, SNTL-ALFA, Praha, 1989.

1
02
D. Ga sˇ parovi cˇ ov a´ et al. / Journal of Molecular Catalysis A: Chemical 264 (2007) 93–102
[
[
[
[
[
[
[
20] J.L. Lemaitre, in: F. Delannay (Ed.), Characterization of Heterogenous
Catalysts, Marcel Dekker, Inc., New York, 1984, p. 32.
21] R. Gavagnin, L. Biasetto, F. Pinna, G. Strukul, Appl. Catal. B: Environ. 38
[27] F. Gauthard, F. Epron, J. Barbier, J. Catal. 220 (2003) 182.
[28] A. Roveda, A. Benedetti, F. Pinna, G. Strukul, Inorg. Chim. Acta 349 (2003)
203.
[29] A. Renouprez, K. Lebas, G. Bergeret, J. Mol. Catal. A: Chem. 120 (1997)
217.
[30] T.J. Klemer, N. Shukla, C. Liu, X.W. Wu, E.B. Sveberg, O. Mryasov, R.W.
Chantrell, D. Weller, Appl. Phys. Lett. 81 (2002) 2220.
[31] D. Ga sˇ parovi cˇ ov a´ , M. Kr a´ lik, M. Hronec, A. Biffis, M. Zecca, B. Corain,
J. Mol. Catal. A: Chem. 244 (2006) 258.
[32] O.M. Ilinitch, L.V. Nosova, V.V. Gorodetskii, V.P. Ivanov, S.H. Trukhan,
E.N. Gribov, S.V. Bogdanov, F.P. Cuperus, J. Mol. Catal. A: Chem. 158
(2000) 237.
(
2002) 91.
22] M.B. Fernandez, C.M. Piqueras, G.M. Toneto, G. Crapiste, D.E. Damiani,
J. Mol. Catal. A: Chem. 233 (2005) 133.
23] S. Karski, I. Witonska, J. Rogowski, J. Goluchowska, J. Mol. Catal. A:
Chem. 240 (2005) 155.
24] J. Batista, A. Pintar, D. Mandrino, M. Jenko, V. Martin, Appl. Catal. A:
Gen. 206 (2001) 113.
25] M. Fernand e´ z-Garcia, J.A. Anderson, G.L. Haller, J. Phys. Chem. 100
(
1996) 16247.
26] D. Uzio, in: J. Lynch (Ed.), Physico-Chemical Analysis of Industrial Cat-
alysts, Technips, Paris, 2003, p. 27.
[33] A. Pintar, M. Setinc, J. Levec, J. Catal. 174 (1998) 72.