Selective hydrogenation of nitrate to nitrite in water over Cu-Pd bimetallic clusters supported on active carbon

Article abstract of DOI:10.1016/j.molcata.2006.01.041

Hydrogenation of nitrate (200 ppm) in water with H₂ over Cu-Pd clusters supported on active carbon (AC) was investigated at 333 K using a gas-liquid co-current flow system. Two types of Cu-Pd bimetallic clusters, stabilized with either poly(vinylpyrrolidone) (PVP) or sodium citrate (SC), revealed that the catalysts possessed similar activity (per unit weight of Pd) and high selectivity toward nitrite when pH was 10.5 at the outlet of the reactor. The high selectivity toward nitrite on PVP-stabilized cluster/AC was minimally influenced by the atomic ratio of Cu/Pd (=0.5-4.0); activity was maximal at a ratio of 1:1. Increasing pH to 12.4 by addition of NaOH enhanced the selectivity toward nitrite to 93% over SC-stabilized Cu_0.63-Pd cluster/AC, but caused a decrease in the reaction rate. Over Cu_0.63-Pd cluster/AC, hydrogenation of nitrite as an intermediate occurred much more slowly than that of nitrate at pH 10.5, suggesting that high selectivity toward nitrite is attained by OH^- inhibiting adsorption of nitrite. XRD and STEM gave the size of the Cu_0.63-Pd cluster on AC as 4 nm; the structure of the cluster remained almost unchanged during the reaction. The activity and selectivity of the Cu_0.63-Pd cluster/AC was superior to those of the Cu_0.63-Pd cluster on oxides such as TiO₂, Al₂O₃, and ZrO₂. In addition, the Cu_0.63-Pd cluster/AC was more active and selective than conventionally prepared Cu_0.63-Pd/AC, indicating that the Cu-Pd cluster is an excellent precursor for selective catalysts in the hydrogenation of nitrate to nitrite.

Full text of DOI:10.1016/j.molcata.2006.01.041

Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
Selective hydrogenation of nitrate to nitrite in water over Cu-Pd
bimetallic clusters supported on active carbon
Yoshinori Sakamoto, Yuichi Kamiya, Toshio Okuhara∗
Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan
Received 13 October 2005; received in revised form 19 January 2006; accepted 21 January 2006
Available online 28 February 2006
Abstract
Hydrogenation of nitrate (200 ppm) in water with H2 over Cu-Pd clusters supported on active carbon (AC) was investigated at 333 K using a
gas–liquid co-current flow system. Two types of Cu-Pd bimetallic clusters, stabilized with either poly(vinylpyrrolidone) (PVP) or sodium citrate
(SC), revealed that the catalysts possessed similar activity (per unit weight of Pd) and high selectivity toward nitrite when pH was 10.5 at the outlet
of the reactor. The high selectivity toward nitrite on PVP-stabilized cluster/AC was minimally influenced by the atomic ratio of Cu/Pd (=0.5–4.0);
activity was maximal at a ratio of 1:1. Increasing pH to 12.4 by addition of NaOH enhanced the selectivity toward nitrite to 93% over SC-stabilized
Cu0.63-Pd cluster/AC, but caused a decrease in the reaction rate. Over Cu0.63-Pd cluster/AC, hydrogenation of nitrite as an intermediate occurred
much more slowly than that of nitrate at pH 10.5, suggesting that high selectivity toward nitrite is attained by OH− inhibiting adsorption of nitrite.
XRD and STEM gave the size of the Cu0.63-Pd cluster on AC as 4 nm; the structure of the cluster remained almost unchanged during the reaction.
The activity and selectivity of the Cu0.63-Pd cluster/AC was superior to those of the Cu0.63-Pd cluster on oxides such as TiO2, Al2O3, and ZrO2. In
addition, the Cu0.63-Pd cluster/AC was more active and selective than conventionally prepared Cu0.63-Pd/AC, indicating that the Cu-Pd cluster is
an excellent precursor for selective catalysts in the hydrogenation of nitrate to nitrite.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Hydrogenation; Nitrate; Nitrite; Cu-Pd; Cluster
1. Introduction
(AC) possessed enhanced activity and stability for this reac-
tion compared to other Cu-Pd-supported catalysts [8]. However,
these catalysts remain unsuitable for practical usage, because
Cu-Pd/Al2O3 is an effective solid catalyst for the hydro-
genation of nitrate (NO3−) [1]. Nitrate is potentially harmful to
human health, causing conditions such as blue baby syndrome
and cancer [2], and should be reduced to below the maximum
allowable level of 50 ppm in drinking water. The hydrogena-
tion of nitrate using a Pd or Pt bimetallic catalyst has been
studied extensively [3–15]. The Cu-Pd, Sn-Pd, and In-Pd cat-
alysts possess high activity for the hydrogenation of nitrate with
high selectivity toward nitrogen [5]. Strukul et al. reported that
ZrO2 and SnO2 can serve as effective supports for Cu-Pd in this
reaction system [6]. Pintar and co-workers have achieved the
conversion of nitrate to nitrogen over Cu-Pd/Al2O3 with high
selectivity using a single-flow fixed-bed reactor [7]. One of the
authors and co-workers demonstrated that Cu-Pd/active carbon
the concentration of ammonia (11.4 ppm over 5.6 wt% [Cu0.2
-
Pd]/AC) produced is above the allowable level of 0.5 ppm in
drinking water. Cu-Pd/glass fiber cloth [9] and Cu-Pd/pumice
[10] produce low levels of ammonia of about 0.5 ppm from the
hydrogenation of nitrate, but the catalytic activity is insufficient.
The mechanism of nitrate hydrogenation in the presence of
conventional Cu-Pd catalysts has been proposed to have a nitrite
intermediate [3,8,11]. The Cu-Pd pairs of the Cu-Pd catalyst
supported on AC accelerate the hydrogenation of nitrate to nitrite
(NO2−) (the first step); subsequent hydrogenations of nitrite
to nitrogen and ammonia (the second step) proceed mainly on
Pd aggregates in the Cu-Pd catalyst [8]. A one-stage process
using a single catalyst with acceptable activity, selectivity, and
stability for the practical treatment of drinking water has not
been developed.
Recently, nano-sized monometallic or bimetallic clusters
have attracted attention because of their unique properties [16].
∗
Corresponding author. Tel.: +81 11 706 4513; fax: +81 11 706 4513.
E-mail address: oku@ees.hokudai.ac.jp (T. Okuhara).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.01.041

Y. Sakamoto et al. / Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
81
These metal clusters have found applications as catalysts [17],
chemical sensors [18], and devices [19]. Toshima et al. [20–22]
have synthesized a series of polymer-protected bimetallic clus-
ters and analyzed the structures. Cu-Pd bimetallic colloids with
different atomic ratios were prepared in the presence of poly(N-
vinyl-2-pyrrolidone) and the formation of a bimetallic alloy
phase was confirmed [20]. These clusters selectively catalyzed
hydration of acrylonitrile and hydrogenation of cyclooctadiene
[21]. Extended X-ray absorption fine structure (EXAFS) analy-
sis led to proposal of a heterobond-philic structure for the Cu-Pd
nanoclusters, involving shorter Cu-Pd bond distances and large
coordination numbers of Cu around Pd and Pd around Cu [22].
These Cu-Pd bimetallic clusters may provide crucial informa-
tion about the roles of the Cu-Pd sites in the hydrogenation of
nitrate.
We reported preliminary findings that Cu-Pd bimetallic clus-
ters supported on AC possess high activity and selectivity for
hydrogenation of nitrate to nitrite at alkaline pH, but are mostly
inactive for hydrogenation of nitrite [12]. In addition, some
monometallic Pd catalysts are selective for the hydrogenation
of nitrite to N2 and N2O [3,10,23]. Thus, a two-stage process
involving hydrogenation of nitrate to nitrite followed by nitrite
hydrogenation can be proposed for removal of nitrate in water.
5 wt% In-Pd (4:1)/AC (54% NO2−-selectivity at 87% conver-
sion)[13]and0.50 wt%Cu-Pd(1.7:1)/Al2O3 [14]wereselective
for nitrate–nitrite hydrogenation, but the selectivity was not sat-
isfactory.
Recently, an anamox reaction (anaerobic oxidation of ammo-
nia) using bacteria (involving nitrite + ammonia) has attracted
attention for the treatment of drain blockages and septic sys-
tems [24]. Thus, the selective hydrogenation of nitrate–nitrite
may also find application to the purification of sewage.
The present study focuses on the catalytic performance of
supported Cu-Pd clusters for the hydrogenation of nitrate. A
gas–liquid co-flow reaction system was used to assess catalytic
performance at the steady state. Effects of stabilizer for clus-
ter, support, loading of cluster, and the atomic ratio of Cu/Pd
on catalytic activity and selectivity were investigated. XRD
and STEM were used to characterize the Cu-Pd bimetallic
particles.
adjustment of pH to 9.5–10.5 using aqueous NaOH (Wako Pure
Chemical Co., 1 mol dm−3). The resulting CunPd(OH)x was
immediately reduced using ethylene glycol and then allowed
to stand at 471 K for 3 h under a nitrogen atmosphere. Acetone
(Wako Pure Chemical Co., about 6 dm3) was then added to the
suspension. The resulting solid was isolated by centrifugation
to afford the Cu-Pd clusters. We obtained five clusters with
Cu/Pd ratios of n = 0, 0.5, 1, 2, and 4 in [Cun-Pd]PVP. Support of
the Cu-Pd cluster on AC (Wako Pure Chemical Industries, Ltd.,
1155 m2 g−1) was achieved by incipient wetness impregnation
using an aqueous colloidal solution (4.68 mmol dm−3) of the
Cu-Pd clusters. The resulting solid was dried overnight at
373 K. Note that loading amounts of the clusters were limited
to 0.4 wt% (sum of Cu and Pd) because the sample powder
aggregated during impregnation. The total amount of metals in
these catalysts was adjusted to 23.4 ␮mol g-cat−1
The second Cu-Pd cluster, stabilized with SC ([Cu0.63-Pd]SC
.
)
was prepared according to a previously reported method [25]. A
solution of Pd(NO3)2·nH2O (Wako Pure Chemical Co., 7.4 g)
and Cu(NO3)2·3H2O (Wako Pure Chemical Co., 4.9 g) in water
(100 cm3) was added to a solution of sodium citrate (Wako
Pure Chemical Co., 30 wt%) in water (295 g). The mixture
was added to a solution of FeSO4 (Wako Pure Chemical Co.,
25 wt%) in water (129 g). Cu2+ and Pd2+ were immediately
reduced by FeSO4 to form the Cu-Pd cluster. The resulting
suspension was stirred at room temperature for 20 h under a
nitrogen atmosphere. The resulting solid was isolated by cen-
trifugation to afford the Cu-Pd clusters. AC, TiO2 (Aerosil
P-25, 46 m2 g−1), ZrO2 [135 m2 g−1, obtained by calcination
of Zr(OH)4 at 723 K prepared from ZrOCl2·8H2O (Wako Pure
Chemical Co.) and an aqueous solution of NH3 (Wako Pure
Chemical Co.)], and ␥-Al2O3 (JRC-ALO-1, Reference Catalyst
of Catalysis Society, Japan, 152 m2 g−1) were used as supports.
For the support of [Cu0.63-Pd]SC, the aqueous colloidal solution
of the Cu-Pd cluster was subjected to the same procedure as
that used for the [Cun-Pd]PVP/AC. We obtained three types of
[Cu0.63-Pd]SC/AC with loading amounts of 0.5, 1.0, and 2.0 wt%
Cu + Pd.
2.2. Conventional Cu-Pd catalysts
2. Experimental
Conventional Cu-Pd/AC as well as Pd/AC and Cu/AC were
prepared by incipient wetness impregnation using PdCl2 and
2.1. Preparation of Cu-Pd cluster catalysts
Cu(NO3)2. An aqueous solution of PdCl2 (0.113 mol dm−3
)
was added dropwise to AC at room temperature. After dry-
ing overnight at 373 K, an aqueous solution of Cu(NO3)2
(0.116 mol dm−3) was introduced to the solid at room temper-
ature. The resulting solids were dried overnight at 373 K in an
oven. Then, theyweretreatedunderaflowofHefor1 h, followed
by a flow of H2 at 573 K for 2 h to yield 2.0 wt% Cu0.63-Pd/AC,
5.0 wt% Pd/AC, and 3.0 wt% Cu/AC.
Two types of Cu-Pd clusters, stabilized with
poly(vinylpyrrolidone) (PVP) and sodium citrate (SC),
were synthesized. First, Cu-Pd clusters stabilized with PVP
(denoted [Cun-Pd]PVP; n represents Cu/Pd atomic ratio) were
prepared according to the method of Toshima et al. with
some modifications [20]. A solution of Pd(OAc)2 (Wako
Pure Chemical Co., 21–105 mg) in 1,4-dioxane (Wako Pure
Chemical Co., 2 cm3) was added to a solution of ethylene glycol
(Wako Pure Chemical Co., 150 cm3) containing PVP (Wako
Pure Chemical Co., 1 g) and CuSO4 (Wako Pure Chemical Co.,
0–93 mg, Cu + Pd = 0.468 mmol). The mixture was stirred for
1 h until the solids were completely dissolved, followed by the
2.3. Catalytic hydrogenations of nitrate and nitrite in water
−
−
The hydrogenation of NO3 or NO2 with H2 was per-
formed using a gas–liquid continuous up-cocurrent flow reactor
(Pyrex tube, 10 mm i.d.). Temperature of the reactor was main-

82
Y. Sakamoto et al. / Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
−
the cluster particles for [Cu2-Pd]PVP was approximately 2 nm,
as determined by TEM [20].
tained at 333 K in a water bath. An aqueous solution of NO3
or NO2−, prepared from NaNO3 (Wako Pure Chemical Co.)
at a concentration of 200 ppm (3.22 mmol dm−3) or NaNO2
(Wako Pure Chemical Co.) at the concentration of 148 ppm
(3.22 mmol dm−3), and gas [H2 or some cases H2 + CO2 (1:1)]
was fed into the reactor under atmospheric pressure through a
Fig. 1 shows a typical time course for the catalytic hydro-
−
genation of NO3 over 0.18 wt% [Cu2-Pd]PVP/AC at 333 K.
−
The % conversion is defined by 100 × (concentration of NO3
−
at inlet − concentration of NO3 at outlet)/(concentration of
−
−
NO3 at inlet). At the initial stage of the reaction, the conver-
pre-heated zone (Pyrex tube, 4 mm i.d.). The NO3 flow rate
was varied from 0.02 to 2.34 mmol h−1 and that of H2 was
sion increased gradually and stabilized after 23 h. As discussed
later, since the structure of the cluster itself remained unchanged
during the reaction, the changes in the conversion and selectiv-
ities are probably due to elimination of the stabilizer such as
adjusted to 3.5 mmol h−1. The pH of the NO3 and NO2
−
−
solution was approximately 5.5 or 5.8, respectively. The pH
at the reactor outlet was intermittently monitored with a pH
meter (HORIBA, pH METER F-22). Gas at the reactor out-
let was sampled by microsyringe and analyzed by TCD-GC
(Shimadzu, GC-8A) with a Molecular Sieve 5A column (for
N2 and O2) or a Porapak Q column (for N2O). Concentrations
−
PVP. The selectivity (on the basis of N-atoms) toward NO2
also gradually increased to approximately 83%. In contrast, the
selectivity toward N2O and N2 decreased with time, eventu-
ally reaching very low values. The formation of NH3 gradually
decreased. Other [Cun-Pd]PVP/AC samples exhibited similar
time courses, while changes at the initial stage were more rapid
over [Cu0.63-Pd]SC/AC. Catalytic activity and selectivity were
estimated from data obtained after 23 h. Note that [Pd]PVP was
inactive.
of NH3, NO2−, and NO3 in the aqueous phase at the reac-
−
tor outlet were measured using a flow injection analysis system
(FIA) (consisting of a Soma Optics S-3250 detector and a Sanuki
Industry FI-710 analyzer equipped with a RX-703T pump and
a R-5000C reactor). For these measurements, detection limits
of NO3−, NO2−, and NH3 were 0.001, 0.001, and 0.1 ppm,
respectively.
Fig. 2 demonstrates−the contact time (W/F) dependence in
hydrogenation of NO3 in the presence of 0.18 wt% [Cu2-
Pd]PVP/AC, where W is the catalyst weight and F the flow rate
of the reaction solution. The conversion increased with contact
time; the selectivity was almost constant within this time range.
2.4. Characterization
−
X-ray diffraction patterns of Cu-Pd clusters dispersed in
water and supported on AC, and conventionally prepared Pd/AC,
Cu/AC, and Cu-Pd/AC, were obtained by X-ray diffractometry
(Rigaku, Mini Flex) using Cu K␣ radiation (λ = 0.154 nm). For
measurement of XRD, Cu-Pd clusters dispersed in water (the
colloidal suspension) were placed into a glass holder at room
temperature and covered with a membrane to prevent bleed-
ing. The crystallite size of the particles was calculated from
the half width of the diffraction peak (2θ = around 40◦) using
Scherrer’s equation, D = 0.9λ/β cosθ, where λ is the X-ray wave-
length (Cu K␣), θ the diffraction angle, and β half line width.
The amounts of dissolved Pd and Cu were measured by ICP
(Shimadzu ICPS-7000) using the solution at the reactor out-
let. Scanning transmission electron microscope (STEM) images
were taken with a Hitachi S-4800 instrument equipped with a
STEM unit operating at 25 kV.
Interestingly, 0.18 wt% [Cu2-Pd]PVP/AC gave mainly NO2
with high selectivity (>85%) nearly independently of W/F under
the reaction conditions. Amounts of N2 and N2O formed were
negligibly small, while NH3 was formed with a selectivity of
15%.
Fig.
3
shows the influence of the Cu content,
[100 × Cu/(Cu + Pd)] of [Cun-Pd]PVP/AC, on the reaction
rate and selectivity for hydrogenation of NO3−. The reaction
rate was determined from the W/F dependence of the conversion
and selectivity from data obtained by the conversion between
37% and 92%. As Fig. 3 indicates, the catalytic activity
was dependent on the Cu content, the maximum activity
3. Results
3.1. Hydrogenation of nitrate over [Cun-Pd]PVP/AC
Cu-Pd clusters stabilized with PVP were systematically char-
acterized using UV–vis spectroscopy, XRD, XPS, TEM, and
EXAFS by Toshima et al. [20,21,22]. When colloidal solutions
of [Cun-Pd]PVP were examined by UV–vis spectroscopy, an
absorption peak at 526 nm was not observed as reported by
Toshima and Wang [20]. Since the peak at 526 nm is due to
plasma oscillations characteristic of large particles of Cu, the
lack of this peak in the spectra of PVP-protected Cu-Pd clusters
indicates that these clusters were generated without formation
of large Cu aggregates. XRD of [Cu2-Pd]PVP exhibited patterns
similar to those reported in the literature [20,21]. The size of
Fig. 1. Time course of hydrogenation of nitrate over 0.18 wt% [Cu2-Pd]PVP/AC.
−
(᭹) Conversion of NO3 and selectivity toward (ꢀ) NO2−, (ꢁ) N2, (ꢀ)
N2O, and (ꢁ) NH3. Reaction conditions: catalyst weight, 1.0 g; temperature,
−
333 K; reactant, NO3 (200 ppm from NaNO3), 0.13 mmol h−1; H2 (1 atm),
3.5 mmol h−1
.

Y. Sakamoto et al. / Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
83
Table 1
Effects of support in Cu-Pd clusters on activity and selectivity for hydrogenation of nitratea
Catalyst
Reaction rate
Conversion (%)
Selectivityb (%)
N2 N2O
2.4
16.7
8.9
0
WF−1 (liq)
(mmol h−1(g-cat)−1
)
(g h dm−3
)
−
NO2
NH3
1.0 wt% [Cu0.63-Pd]SC/AC
1.0 wt% [Cu0.63-Pd]SC/TiO2
1.0 wt% [Cu0.63-Pd]SC/ZrO2
1.0 wt% [Cu0.63-Pd]SC/Al2O3
2.017
0.356
0.376
0.355
86.1
68.6
72.4
65.0
0
0
0
0
89.2
56.3
69.8
72.3
8.4
27.0
21.3
27.7
1.8
6.3
6.2
5.9
a
Reaction conditions: NO3− 200 ppm (NaNO3 3.22 mmol dm−3); H2 (1 atm), 3.5 mmol h−1; catalyst, 0.02–1.0 g; reaction temperature, 333 K; pH at outlet of the
reactor, about 10.5.
b
On the basis of N atom.
3.2. Hydrogenation of nitrate over [Cu0.63-Pd]SC/support
Table 1 shows the reaction rate and selectivity for hydrogena-
−
tion of NO3 over supported [Cu0.63-Pd]SC, where AC, TiO2,
ZrO2, and Al2O3 were employed as a support. Table 1 demon-
strates that 1.0 wt% [Cu0.63-Pd]SC/AC possessed higher activity
for the hydrogenation of NO3− than [Cu0.63-Pd]SC supported on
TiO2, ZrO2, andAl2O3. Inaddition,−[Cu0.63-Pd]SC/ACexhibited
the highest selectivity toward NO2 even at higher conversion.
In the solution of the reactor-outlet the metals, including Cu
and Pd, were not detected by ICP during the reaction over the
catalysts prepared from the Cu-Pd clusters.
3.3. Characterization and hydrogenation of nitrate over
[Cu0.63-Pd]SC/AC
Fig. 2. Contact time dependence of the hy−drogenation of nitrate over 0.18 wt−%
[Cu2-Pd]PVP/AC. (᭹) Conversion of NO3 and selectivity toward (ꢀ) NO2
,
(ꢁ) N2, (ꢀ) N2O, and (ꢁ) NH3. Reaction conditions: temperature, 333 K; reac-
Fig. 4 illustrates XRD patterns of colloidal [Cu0.63-Pd]SC
(dispersed in water) and 2.0 wt% [Cu0.63-Pd]SC/AC, conven-
tant, NO3−, 200 ppm from NaNO3; H2 (1 atm), 3.5 mmol h−1
.
tional 5.0 wt% Pd/AC, 3.0 wt% Cu/AC, and 2.0 wt% Cu0.63
-
Pd/AC. Two peaks (2θ = 40.4 and 47.0◦) were observed in the
being obtained at a Cu content of 50 mol% (Cu/Pd = 1). The
selectivity toward NO2− was high for all [Cun-Pd]PVP/AC, and
[Cu4−-Pd]PVP/AC showed the highest selectivity (86.4%) toward
region of the diffraction angle for [Cu0.63-Pd]SC in the col-
NO2
.
Fig. 3. Effect of atomic ratio of Cu/Pd in Cu-Pd clusters on activity and selectiv-
ity of hydrogenation of nitrate over [Cun-Pd]PVP/AC (Cu + Pd = 0.0234 mmol g-
cat−1). (᭹) Reaction rate and selectivity toward (ꢀ) NO2−, (ꢁ) N2, (ꢀ) N2O,
and (ꢁ) NH3. Reaction conditions: temperature, 333 K; NO3−, 200 ppm from
Fig. 4. XRD patterns of CuPd cluster dispersed in water and supported on AC,
conventional Pd/AC, Cu/AC, and Cu-Pd/AC. (a) [Cu0.63-Pd]SC dispersed in
water, (b) AC, (c) 2.0 wt% [Cu0.63-Pd]SC/AC after H2 treatment at 333 K for
23 h, (d) 5.0 wt% Pd/AC, (e) 3.0 wt% Cu/AC, and (f) conventional 2.0 wt%
Cu0.63-Pd/AC.
NaNO3; H2 (1 atm), 3.5 mmol h−1
.

84
Y. Sakamoto et al. / Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
Table 2
Hydrogenation of nitrate or nitrite with H2 over various Cu-Pd catalystsa
Entry
Reactant
Catalyst
pHb
Reaction ratec
Conversion (%)
Selectivityd (%)
WF−1 (liq) (g h dm−3
)
−
N2
N2O
NO2
NH3
−
1
2
3
4
5
6
NO3−
0.19 wt% [Cu-Pd]PVP/ACe
1.0 wt% [Cu0.63-Pd]SC/ACe
1.0 wt% [Cu0.63-Pd]SC/ACe
1.0 wt% [Cu0.63-Pd]SC/ACe
2.0 wt% [Cu0.63-Pd]SC/ACe
2.0 wt% Cu0.63-Pd/ACf
10.5
6.5
10.5
12.4f
10.5
10.5
0.203
2.395
2.017
0.206
4.333
1.652
90.9
100
93.6
94.4
91.6
91.4
4.3
0
24.0
0
0
0
82.9
0
87.2
92.3
84.8
71.9
12.8
35.0
10.3
6.4
13.7
22.0
23.4
9.9
3.2
16.2
1.3
4.6
NO3−
NO3−
NO3−
NO3−
NO3
41.0
2.5
1.3
1.5
6.1
0
−
7
NO2
1.0 wt% [Cu0.63-Pd]SC/ACg
10.5
0.072
59.1
12.9
0
–
87.1
27.4
a
−
−
Reaction conditions: NO3 200 ppm (NaNO3 3.22 mmol dm−3) or NO2 148 ppm (NaNO2 3.22 mmol dm−3); H2 or CO2, 3.5 mmol h−1; catalyst, 0.02–1.0 g;
reaction temperature, 333 K.
b
pH at outlet of the reactor. To adjust pH to 6.5, CO2 (3.5 mmol h−1) was fed to the H2 flow.
c
mmol h−1 (g-cat)−1
.
d
e
f
On the basis of N atom.
Prepared with Cu-Pd bimetallic cluster.
Aqueous solution of NaOH was added to the reaction mixture.
Prepared by a conventional impregnation using aqueous Cu(NO3)2 and PdCl2.
g
loidal state (Fig. 4a). The crystallite size of the cluster was
approximately 4 nm, determined from the half-width of the
diffraction peak at 2θ = 40.4◦ due to the [1 1 1] plane, which
is consistent with the results from STEM. The 2.0 wt% [Cu0.63
-
Pd]SC/AC powder after treatment with H2 flow at 333 K showed
two peaks (2θ = 40.4 and 46.8◦) (Fig. 4c), which were con-
sistent with those of the colloidal cluster. Note tha−t the XRD
pattern was unchanged after hydrogenation of NO3 for 22 h.
The crystallite size of the Cu-Pd cluster of [Cu0.63-Pd]SC/AC
was approximately 4 nm, either after H2 treatment or after the
hydrogenation of nitrate, indicating that cluster size remained
unchanged upon impregnation and the reaction. The diffraction
peaks of monometallic 5.0 wt% Pd/AC (2θ = 40.0 and 46.5◦ in
Fig. 4d) and monometallic 3.0 wt% Cu/AC (2θ = 43.2 and 50.3◦
in Fig. 4e) were due to Pd and Cu particles, respectively [26,27].
Conventional 2.0 wt%Cu0.63-Pd/AC exhibitedthree sharp peaks
due to bimetallic particles (2θ = 40.0, 42.7, and 46.5◦ in Fig. 4f).
The middle peak (2θ = 42.7◦ in Fig. 4f) was located on the left
of the Cu particles peak (Fig. 4e), while the other peaks were
consistent with those of Pd particles (Fig. 4d). The size of the
particles on 2.0 wt% Cu0.63-Pd/AC was approximately 8 nm.
Fig. 5 shows a STEM image for [Cu0.63-Pd]SC/AC. The
STEM image demonstrates the dispersion of particles with sizes
of 2–4 nm (average of about 3.5 nm).
Fig. 5. STEM image of [Cu0.63-Pd]SC/AC.
Fig. 6 shows the influence of the loading amount of [Cu0.63
-
Pd]SC on the reaction rate and selectivity for the hydrogena-
tion of NO3−. Besides [Cu0.63-Pd]SC, 0.19 wt% [Cu-Pd]PVP is
plotted as a reference. The reaction rate increased nearly in
proportion to the loading amount of Cu-Pd clusters, includ-
ing 0.19 wt%[Cu-Pd]PVP. When compared at conversions higher
than 90%, the selectivities for products were minimally affected
by the loading amount of Cu-Pd clusters.
Table 2 summarizes the catalytic results of [Cu0.63-Pd]SC/AC
as compared with [Cu-Pd]PVP/AC, conventional Cu0.63-Pd/AC,
and those of [Cu0.63-Pd]SC/AC at different pH values. The reac-
tion rates were estimated from the W/F dependencies of the
Fig. 6. Effect of Cu-Pd cluster loading on reaction rate and selectivity in
the hydrogenation of nitrate over 0.19 wt% [Cu-Pd]PVP/AC and 0.5–2.0 wt%
[Cu0.63-Pd]SC/AC. (᭹) Reaction rate and selectivity toward (ꢀ) NO2−, (ꢁ)
−
N2, (ꢀ) N2O, and (ꢁ) NH3. Reaction conditions: temperature, 333 K; NO3
,
200 ppm from NaNO3; H2 (1 atm), 3.5 mmol h−1
.

Y. Sakamoto et al. / Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
85
conversion below 65%. The selectivities at conversion higher
than 90% are shown. The 1.0 wt% [Cu0.63-Pd]SC/AC catalyst
was about 10 times more active than 0.19 wt% [Cu-Pd]PVP/AC
(entries 1 and 3). If the activity was compared per unit weight of
For conventional 2.0 wt% Cu0.63-Pd/AC, peaks correspond-
ing to the Pd particles appeared, but the peak at 42.7◦ assignable
to Cu particles was located slightly on the left of the peak of
3.0 wt% Cu/AC. Therefore, the particles on 2.0 wt% Cu0.63
-
Pd, the activities of these catalysts were comparable. The selec-
Pd/AC must contain Cu-rich Cu-Pd particles as well as Pd
particles.
−
tivity toward NO2 was slightly higher over 1.0 wt% [Cu0.63
-
Pd]SC/AC than over 0.19 wt% [Cu-Pd]PVP/AC. Note that the
activity of 2.0 wt% [Cu0.63-Pd]SC/AC was about 3 times greater
than that of conventional 2.0 wt% Cu0.63-Pd/AC (entries 5 and
6). Furthermore, 2.0 wt% [Cu0.63-Pd]SC/AC had superior selec-
tivity toward NO2−, suppressing NH3 formation in contrast to
conventional 2.0 wt% Cu0.63-Pd/AC.
The present study demonstrated that [Cu0.63-Pd]SC/AC as
well as [Cun-Pd]PVP afforded predominantly NO2−, along with
−
small amounts of NH3 during the hydro−genation of NO3 at
alkaline pH. The selectivity toward NO2 was influenced sig-
nificantly by neither the stabilizer, the Cu/Pd ratio in [Cun-
Pd]PVP, nor the loading amount of [Cu0.63-Pd]SC on AC (Table 2
and Figs. 3 and 6). Since these cluster catalysts showed the
high selectivity for nitrate–nitrite compared to the conventional
Cu0.63-Pd/AC (Table 2), the uniformity in the active sites (Cu-
Pd) as pointed out by Toshima et al. [21] may be the reason for
the high selectivity over the cluster catalysts.
Effects of pH on the rate and selectivity for hydrogenation of
NO3− were examined using 1.0 wt% [Cu0.63-Pd]SC/AC (entries
2–4). To adjust pH to neutral (pH 6.5) at the outlet, CO2 was
introduced into the H2 flow to attain a ratio of 1:1 (total 1 atm)
(entry 2). While the reaction rate at pH 6.5 was similar to that at
pH 10.5 (entry 3), the selectivity toward NO2− was quite differ-
ent, (i.e., zero at pH 6.5; entry 2). When pH was increased to 12.4
(at the outlet) by adding NaOH to the reactant solution, selectiv-
4.2. Selective hydrogenation of nitrate to nitrite over
[Cu0.63-Pd]SC/AC
−
ity toward NO2 reached 92.3%; but the rate was significantly
decreased (entry 4).
−
It has been proposed that the hydrogenation of NO3 in the
−
The hydrogenation of NO2 (148 ppm) at pH 10.5 (at the
−
presence of bimetallic Cu-Pd catalysts occurs through NO2
outlet) demonstrates that the reaction rate of 1.0 wt% [Cu0.63
-
(Eq. (1)) [3,8,11].
Pd]SC/AC was about 1/30 that of hydrogenation of NO3− under
the same conditions, while the selectivity toward NH3 was high
(entry 7).
NO3− → NO2− → [NO]ad → N2, NH3
(1)
It is generally accepted that the bimetallic sites of Cu and Pd are
required for the adsorption of NO3− and accelerati−on its hydro-
genation. In contrast, the hydrogenation of NO2 proceeded
readily on Pd as well as on Cu-Pd sites [3,10,23]. We propose
that forconventional Cu-Pd/AC, theCu-PdandPdsitesarecoex-
istent in the particles of Cu-Pd/AC and the Cu-Pd sites accelerate
thehydrogenationofNO3− toNO2− andsubsequenthydrogena-
4. Discussion
4.1. Structure and catalytic properties of Cu-Pd clusters
The structures of Cu-Pd clusters protected with PVP were
analyzed systematically using a variety of methods [20–22].
According to Toshima, Asakura, and co-workers, Cu-Pd alloy
structures have a wide range of Cu/Pd ratio for the nanoclusters
displaying a heterobond-philic structure with preferential for-
mation of Cu–Pd bonds. Their results indicate that the Cu–Pd
bond is shorter than expected due to a strong interaction [22]. In
the present study, the Cu–Pd clusters protected with PVP ([Cun-
Pd]PVP) were characterized using only UV–vis and XRD, which
confirmed the previously reported data [20,21].
The new Cu-Pd cluster/AC, [Cu0.63-Pd]SC/AC, was analyzed
by XRD (Fig. 4) and STEM (Fig. 5). As shown in Fig. 4a, XRD
of [Cu0.63-Pd]SC gave broad peaks due to the cluster, indicating
that small particles (about 4 nm) were present in the colloid solu-
tion. Cu and Pd are known to form a miscible alloy in the entire
composition range. For the disordered fcc phase of Cu-Pd, the
lattice constant increased linearly with the Pd content [21]. Shift
of the diffraction lines of [Cu0.63-Pd]SC supported on AC to the
right of those of Pd/AC indicated the formation of Cu0.63-Pd
clusters. The peak positions of [Cu0.63-Pd]SC/AC are compat-
ible with those of well-known crystallite phases such as Cu3-
Pd and Cu-Pd [28,29]. The XRD pattern of [Cu0.63-Pd]SC/AC
did not change during hydrogenation of NO3−, indicating that
the structure of the cluster remained unchanged during the
reaction.
−
tion of NO2 occurs mainly on the Pd sites (or plane) as well
as on the Cu-Pd sites [8]. However, since Cu-Pd and supported
clusters possess nearly uniform Cu-Pd sites, the difference in
Cu-Pd particles between the cluster and conventional catalysts
likely explains the different selectivities for NO2−. Hydrogena-
tion of NO3− was about 30 times faster than that of NO2− over
1.0 wt% [Cu0.63-Pd]SC/AC (Table 2), which clearly expl−ains the
−
high selectivity for NO2 in the hydrogenation of NO3
.
−
The effects of pH on the reaction rate and selectivity of NO3
hydrogenation suggest the critical role of OH− (Table 2). An
increase in pH resulted in a significant decrease in the reac-
tion rate and increased selectivity toward NO2−. We previously
−
reported that adsorption of NO2 on Cu-Pd/AC and Pd/AC
decreased with increasing pH [15]. Therefore,−it is reasonable
that the reaction rate of hydrogenation of NO2 at alkaline pH
was low over the Cu-Pd catalysts. Kinetic studies have suggested
that NO3− was more strongly adsorbed on C−u-Pd sites than OH−
−
[15]. Therefore, NO2 formed from NO3 was released into
solution without re-adsorption, which explains the high selec-
tivity for nitrite. Fig. 7 illustrates a model for reaction on the
surface of the cluster.
The development of a catalyst with acceptable activity, selec-
tivity, and stability for practical water treatment has been dif-

86
Y. Sakamoto et al. / Journal of Molecular Catalysis A: Chemical 250 (2006) 80–86
clusters are excellent precursors for the selective catalysts for
the hydrogenation of nitrate to nitrite.
Acknowledgements
This work was partially supported by a project of Core
Research for Evolutional Science and Technology (CREST) at
the Japan Science and Technology Agency (JST). This work was
also supported by a Grant-in-Aid for Scientific Research (Grant
No. 17560678) from the Ministry of Education, Culture, Sports,
Science, and Technology, Japan.
Fig. 7. Proposed model of the surface reaction over Cu-Pd cluster at pH 10.5.
References
−
[1] K.-D. Vorlop, T. Tacke, Chem. Ing. Tech. 61 (1989) 836.
[2] World Health Organization (Ed.), Health Hazards from Nitrate in
Drinking-Water, WHO Regional Office for Europe, Copenhagen, 1985.
[3] S. Ho¨rold, K.-D. Vorlop, T. Tacke, M. Sell, Catal. Today 17 (1993) 21.
[4] F. Epron, F. Gauthard, J. Barbier, Appl. Catal. A 237 (2002) 253.
[5] U. Pru¨sse, M. Hahnlein, J. Daum, K.-D. Vorlop, Catal. Today 55 (2000)
79.
[6] G. Strukul, R. Gavagnin, F. Pinna, E. Modaferri, S. Perathoner, G. Centi,
M. Marella, M. Tomaselli, Catal. Today 55 (2000) 139.
[7] A. Pinter, J. Batista, J. Levec, Catal. Today 66 (2001) 503.
[8] Y. Yoshinaga, T. Akita, I. Mikami, T. Okuhara, J. Catal. 207 (2002) 37.
[9] Y. Matatov-Meytal, V. Barelko, L. Yuranov, L. Kiwi-Minsker, A.
Renken, M. Sheintuch, Appl. Catal. B 31 (2001) 233.
[10] F. Deganello, L.F. Liotta, A. Macaluso, A.M. Venezia, G. Deganello,
Appl. Catal. B 24 (2000) 265.
ficult. A two-stage process invo−lving NO3 hydrogenation to
−
NO2 in the first stage and NO2 hydrogenation in the second
stage has been proposed [13]. This process allows optimization
ofthecatalystsateachstageandimprovementoftheirselectivity.
Some monometallic Pd catalysts are selective in the second stage
reaction, the hydrogenation of NO2− to N2 and N2O [3,10,23].
However, few reports examine the catalytic hydrogenation of
NO3− to NO2− [12]. The present results allow the possibility of
a two-stage process for the purification of water polluted with
NO3−. The formation of NH3 as a by-product in the hydrogena-
−
tion of NO3 results from the hydrogenation of the product
−
NO2 (Table 2). Considering the sel−ective formation of NH3
(87%) in the hydrogenation of NO2 over 1.0 wt% [Cu0.63
-
[11] J. Wa¨rna˚, I. Turunen, T. Salmi, T. Maunula, Chem. Eng. Sci. 49 (1994)
5763.
−
Pd]/AC, the contribution of secondary hydrogenation of NO2
during NO3− hydrogenation cannot be ignored. While the selec-
tivity toward NH3 decreased to 6.4% (corresponds to 3.3 ppm)
at a pH of 12.4 (Table 2), further reduction of the NH3 selectiv-
ity is strictly required for the practical water treatment using the
two-stage process.
[12] Y. Sakamoto, K. Nakata, Y. Kamiya, T. Okuhara, Chem. Lett. 33 (2004)
908.
[13] L. Lemaignen, C. Tong, V. Begon, R. Burch, D. Chadwick, Catal. Today
75 (2002) 43.
[14] A.J. Lecloux, Catal. Today 53 (1999) 23.
[15] I. Mikami, Y. Sakamoto, Y. Yoshinaga, T. Okuhara, Appl. Catal. B 44
(2003) 79.
[16] R. Kubo, J. Phys. Soc. Jpn. 17 (1962) 975.
[17] G. Schmid, V. Maihack, F. Lantermann, S. Peschel, J. Chem. Soc.,
Dalton Trans. (1996) 589.
5. Conclusions
Two types of Cu-Pd bimetallic clusters, stabilized with
poly(vinylpyrrolidone) (PVP) and sodium citrate (SC), were
synthesized and used as starting materials for catalysts for the
hydrogenation of nitrate to nitrite. These clusters/activated car-
bon possessed similar activity (per unit weight of Pd) and selec-
tivity. XRD and STEM revealed that the size of the Cu0.63-Pd
cluster stabilized by SC on AC was approximately 4 nm and
remained nearly unchanged during the reaction. For the hydro-
genation of nitrate at alkaline pH, the Cu-Pd cluster/AC was
highly selective for the formation of nitrite. By comparison
with conventional Cu-Pd/AC and Cu-Pd cluster/oxide such as
TiO2 and Al2O3, specific catalysis of the Cu-Pd cluster/AC was
attributable to the uniformity of surface sites of the cluster. The
high selectivity for nitrite over Cu-Pd cluster/AC is due to OH−
strongly inhibiting adsorption of nitrite. In conclusion, Cu-Pd
[18] R. Elghanian, J.J. Storhoff, R.C. Mucic, R.L. Letsinger, C.A. Mirkin,
Science 277 (1997) 1078.
[19] G. Scho¨n, U. Simon, Colloid Polym. Sci. 273 (1995) 202.
[20] N. Toshima, Y. Wang, Chem. Lett. (1993) 1611.
[21] N. Toshima, Y. Wang, Langmuir 10 (1994) 4574.
[22] C.-R. Bian, S. Suzuki, K. Asakura, L. Ping, N. Toshima, J. Phys. Chem.
B 106 (2002) 8587.
[23] A. Pinter, J. Batista, J. Levec, Wat. Sci. Tech. 37 (1998) 177.
[24] M. Strous, J.A. Fuerst, E.H.M. Kramer, S. Logemann, G. Muyzer, K.T.
van de Pas-Schoonen, R. Webb, J.G. Kuenen, M.S.M. Jetten, Nature
400 (1999) 446.
[25] Y. Ishihara, T. Hirai, Jpn. Kokai Tokkyo Koho, 2004-57954 (2004) to
Catalysis & Chemical Ind. Co., Ltd.
[26] JCPDS Alphabetical Index, Inorg. Phases, #46–1043, 1996.
[27] JCPDS Alphabetical Index, Inorg. Phases, #4–0836, 1974.
[28] JCPDS Alphabetical Index, Inorg. Phases, #7–138, 1967.
[29] JCPDS Alphabetical Index, Inorg. Phases, #48–1551, 1998.