Structural feature and catalytic performance of Cu species distributed over TiO2 nanotubes

Article abstract of DOI:10.1021/jp064209w

Copper oxide was deposited on tubular TiO₂ via Cu²⁺ introduction into a titanate nanotube aggregate followed by calcination. The titanate has a layered structure allowing Cu intercalation and can readily transform into anatase TiO₂ via calcination for condensation of the constituting layers. The activity of the tubular catalysts, with a Cu content of 2 wt %, in selective NO reduction with NH₃ was compared with those of other 2 wt % Cu/TiO₂ catalysts using TiO₂ nanoparticles as the support. The Cu species supported on the nanotubes showed a higher activity than those supported on the nanoparticles. X-ray absorption near-edge structure (XANES) analysis showed that the Cu species on all the TiO₂ supports are in the +2 state. Extended X-ray absorption fine structure (EXAFS) investigations of these catalysts reflected higher degrees of CuO dispersion and Cu²⁺ dissolution into the TiO₂ lattice for the tubular Cu/TiO₂ catalysts. Absence of CuO bulk detection by a temperature-programmed reduction analysis for the tubular catalysts confirmed the high CuO-dispersion feature of the tubular catalysts. The dissolution of Cu²⁺ to form a Cu_x Ti_1-x O₂type of solid solution was improved by using an in-situ ion-intercalation method for Cu deposition on the nanotubes. A fraction as high as 40% for Cu²⁺ dissolution was obtained for,the tubular catalysts while only 20% was obtained for the particulate catalysts. The Cu_xTi_1-xO₂ species were considered one form of the active sites on the Cu/TiO₂ catalysts.

Full text of DOI:10.1021/jp064209w

J. Phys. Chem. B 2006, 110, 25817-25824
25817
Structural Feature and Catalytic Performance of Cu Species Distributed over TiO2
Nanotubes
Jun-Nan Nian, Shin-An Chen, Chien-Cheng Tsai, and Hsisheng Teng*
Department of Chemical Engineering, National Cheng Kung UniVersity, Tainan 70101, Taiwan
ReceiVed: July 5, 2006; In Final Form: October 14, 2006
via Cu2 introduction into a titanate nanotube aggregate followed
+
Copper oxide was deposited on tubular TiO
2
by calcination. The titanate has a layered structure allowing Cu intercalation and can readily transform into
anatase TiO via calcination for condensation of the constituting layers. The activity of the tubular catalysts,
with a Cu content of 2 wt %, in selective NO reduction with NH was compared with those of other 2 wt %
Cu/TiO catalysts using TiO nanoparticles as the support. The Cu species supported on the nanotubes showed
a higher activity than those supported on the nanoparticles. X-ray absorption near-edge structure (XANES)
analysis showed that the Cu species on all the TiO supports are in the +2 state. Extended X-ray absorption
fine structure (EXAFS) investigations of these catalysts reflected higher degrees of CuO dispersion and Cu
dissolution into the TiO lattice for the tubular Cu/TiO catalysts. Absence of CuO bulk detection by a
temperature-programmed reduction analysis for the tubular catalysts confirmed the high CuO-dispersion feature
2
3
2
2
2
2+
2
2
2
+
x 2
of the tubular catalysts. The dissolution of Cu to form a Cu Ti1-xO type of solid solution was improved
by using an in-situ ion-intercalation method for Cu deposition on the nanotubes. A fraction as high as 40%
for Cu2 dissolution was obtained for the tubular catalysts while only 20% was obtained for the particulate
+
x 2 2
catalysts. The Cu Ti1-xO species were considered one form of the active sites on the Cu/TiO catalysts.
Introduction
and even distribution for ruthenium ions over this tubular titanate
have been reported.
2
c
TiO2, which is chemically stable but active in promoting
catalytic effects, has been used as a support for metal-oxide
catalysts in many reactions including oxidation of CO and
hydrocarbons and selective catalytic reduction (SCR) of NO
because of the high activity of the supported species at relatively
low temperatures.1 For a support, a large surface area is the
In the present study, we used copper oxides as the active
species to explore the capability of the nanotubes in improving
the dispersion of active metal oxides and thus the resulting
activity, which was examined with SCR of NO. Copper oxides
supported on different porous materials are active in catalyzing
-4
4,11-14
selective NO reduction in the presence of O2,
although
5
essential feature to disperse the active species. In addition, the
their stability is one of the important challenges. We extensively
studied the dispersion, as well as the coordination with the TiO2
support, of the Cu species with X-ray absorption fine structure
spectroscopy and other techniques, attempting to give correla-
tions between the structural feature of Cu species and the
catalytic activity. Here, we report the structure of Cu species
supported on titanate-derived TiO2 nantubes and commercially
available TiO2 nanoparticles and show that the Cu species
supported on the nanotubes are more highly dispersed and
microstructure or morphology of the support framework or the
method by which the active species is introduced affects not
only the dispersion but also the structure and chemical environ-
6
ment of the active species. In the present work, we intended
to explore the possibility using titanate-derived TiO2 nanotubes
to improve the metal-oxide dispersion over TiO2 and thus the
activity of the resulting catalysts.
Titanate-based nanotube aggregates prepared from hydro-
thermal treatment on TiO2 in alkali solutions have been shown
to possess a large surface area.7 Upon mild calcination (<400
dissolved into TiO and are more active in catalyzing NO
2
-10
reduction.
°
C), the nanotubes exhibit a well-defined phase of the anatase
8d
TiO2 while still retaining the tubular structure. The titanate
nanotubes are believed to form through scrolling the sheets
peeled off from the titanate matter.9 This leads to a multilayer
feature for the tube wall. By using this titanate nanotube
aggregate as a support, it is expected that the layered structure
would provide sites for the intercalation of metal ions, thus
assisting in accommodation of the active species as well as
preventing agglomeration of the species in the subsequent
calcination for phase transformation into anatase TiO2. The
Experimental Section
b,d
Synthesis. The nanotube TiO2 aggregates were synthesized
by using a commercially available P25 TiO2 (Degussa, Ger-
2
many; surface area ) 50 m /g) as the precursor. The preparation
9
c
7,10e
was analogous to those previously reported.
P25 (3 g) was
hydrothermally treated with 10 N NaOH (100 mL) in a Teflon-
lined autoclave at 130 °C for 24 h. The slurry, which contained
9,10
a titanate composed of lamellar layers, was then washed with
0.1 N HCl solution (1 L) for several times to reach a pH value
of 1.6. After filtration, the obtained solid was dried at 110 °C
for 12 h to give the nanotubes.
highly dispersed metal oxide over a support is believed to be
the active phase for heterogeneous catalysis.11 A high capacity
The nanoparticle TiO2 supports used in the present study were
P25 and other commercially available large-surface-area prod-
*
Author to whom correspondence should be addressed. Tel: 886-6-
2
385371; fax: 886-6-2344496; e-mail: hteng@mail.ncku.edu.tw.
1
0.1021/jp064209w CCC: $33.50 © 2006 American Chemical Society
Published on Web 11/30/2006

2
5818 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Nian et al.
2
ucts, UV100 (Sachtleben, Germany; surface area ) 340 m /g)
reduction experiments were conducted at temperatures ranging
between 100 and 350 °C. The NO concentration was analyzed
by a chemiluminescence NO/NOx analyzer. Gas chromatography
(Hewlett-Packard HP6890, United States; molecular sieve 5A
column, TCD, He carrier) was used to analyze the product
stream, which was principally comprised of NO, NH3, O2, N2,
and H2O. Nitrogen was found to be the main product of NO
reduction in the present work, and only negligible amounts of
N2O and NO2 can be detected
The surface reducibility of the catalysts was studied by
temperature-programmed reduction (TPR). Prior to each TPR
experiment, the catalyst was pretreated in O2 at 400 °C for 1 h.
TPR was carried out under a flow of 10 vol % H2/He mixture
from room temperature to 750 °C at a heating rate of 10 °C/
min. A thermal conductivity detector was used to continuously
monitor the consumption of H2 during the TPR. The volume of
hydrogen taken up by the catalysts was calibrated against pure
CuO. The decomposition of the experimental convoluted TPR
peaks was performed by using a nonlinear least-squares fitting
program with a symmetric Gaussian function.
2
and ST01 (Ishihara, Japan; surface area ) 320 m /g). These
large-surface-area products had average particle sizes less than
1
0 nm.
The introduction of Cu into the tubular or particulate TiO2
aggregates was conducted by using incipient-wetness impregna-
tion of Cu(NO3)2 with stirring and sonication at the ambient
temperature and then vacuum drying at 110 °C for 12 h. The
concentration of Cu(NO3)2 solutions was adjusted to prepare
catalysts of 2 wt % Cu-loading, which was also confirmed with
atomic absorption spectrometry. Calcination of the Cu-loaded
specimens at 400 °C for 1 h was conducted to give the final
Cu-loaded catalysts for further analysis and examination. The
catalysts were designated as Cu/NT for Cu supported on the
nanotubes and Cu/P25, Cu/UV, and Cu/ST for Cu supported
on the P25, UV100, and ST01 nanoparticles, respectively.
Taking advantage of the layered structure of the titanate species,
the degree of Cu dispersion over the nanotubes is expected to
be improved by carrying out an in-situ Cu-incorporation during
the period of titanate nanotube formation. To prepare this type
of Cu-loaded nanotubes, an appropriate amount of Cu(NO3)2
was added to the TiO2/NaOH mixture for the hydrothermal
treatment, which was also followed by the HCl washing for
nanotube formation. After vacuum drying at 110 °C for 12 h,
the Cu-loaded nanotube specimen was also calcined at 400 °C
for 1 h to give the final catalyst (Cu/NT1; 2 wt % in Cu).
Results and Discussion
Figure 1 shows the TEM images of the specimens derived
from the consecutive treatments, including the hydrothermal
autoclaving, the posttreatment acid washing, and finally the
calcination at 400 °C, on the P25 TiO2 precursor for synthesis
of nanotube aggregates. The corresponding XRD patterns are
depicted in Figure 2. After the hydrothermal treatment in NaOH,
the TiO2 precursor transformed into titanate, as indicated by
the XRD pattern in Figure 2a. This titanate material obtained
from the hydrothermal treatment in NaOH was suggested to be
Na2-xHxTi2O5‚H2O because its diffraction pattern was consistent
with that of H2Ti2O5‚H2O (bottom of Figure 2) documented in
the powder diffraction files of the Joint Committee on Powder
Characterization. The phase identification of the specimens
was conducted with powder X-ray diffraction (XRD) using a
Rigaku RINT2000 diffractometer equipped with Cu KR radia-
tion. The data were collected for scattering angles (2θ) ranging
between 5 and 70° with a step size of 0.01°. The microstructures
were explored with a transmission electron microscope (TEM;
Hitachi FE-2000, Japan). The Brunauer-Emmett-Teller surface
areas of the samples were determined by N2 adsorption at -196
°
C using an adsorption apparatus (Micromeritics ASAP 2010,
17
Diffraction Standards. The TEM image of this titanate in Fig-
United States). The amount of N2 adsorbed at relative pressures
near unity (p/p0 ) 0.96 here) was used to give the total pore
volume. The average pore diameter was determined from the
surface area and pore volume, under the assumption of cylindri-
cal-pore configuration.15
ure 1a shows a condensed phase containing multilayered struc-
17
ture, which is characteristic of H2Ti2O5‚H2O. Upon acid wash-
9d
ing, the titanate sheets would peel off and scroll into tubes,
accompanied by Na substitution with proton.
+
10a,e
However,
after the acid washing to reach a pH of 1.6, the material not
only transformed to a tubular configuration (Figure 1b) but also
showed a dramatic change in the crystalline structure from the
Structural feature and chemical environment of the Cu species
on the supports were characterized by X-ray absorption fine
structure spectroscopy, including X-ray absorption near-edge
structure (XANES) and extended X-ray absorption fine structure
titanate to a hybrid of the titanate and anatase TiO (Figure 2b).
2
The XRD pattern shows an obvious weakening of the peaks at
2θ ) 9.2, 24, and 28°, which respectively correspond to the
characteristic diffraction of (200), (110), and (310) faces of the
(
EXAFS). The spectra were measured on the Wiggler beam line
of the Taiwan Synchrotron Radiation Research Center. A
double-crystal Si(111) was used to monochromatize X-rays from
the 1.5 GeV electron storage ring with a current range of 120-
titanate, while the anatase TiO (101) diffraction at 2θ ) 25°
2
can be seen to show up. The readiness of this phase transition
2
00 mA. Photon energy was calibrated by characteristic preedge
from the titanate to anatase TiO was attributed to the fact that
2
peaks in the absorption spectrum of Cu foil (8979 eV). The
EXAFS data were analyzed using the UWX-AFS 3.0 and FEFF
both phases own a zigzag arrangement for their constituting
1
0e,f
TiO -octahedron layers (top of Figure 2).
6
After calcining
1
6
3
7
.0 programs. Fourier transformation was performed on k -
the nanotubes at 400 °C, the tubular structure shrank and was
damaged (Figure 1c), although the contour of the tubular
framework seemingly remained. The TEM image of the Cu-
containing tubular TiO2 (see Supporting Information) was
similar to that shown in Figure 1c. The well-defined anatase
phase is exclusively present in the XRD pattern of the calcined
specimen (Figure 2c), however. The calcination may have
caused dehydration of the layered titanate on the walls into
anatase TiO2, thus leading to the rupture of the tubular structure.
The physical characteristics of the TiO2 supports determined
from N2 adsorption are summarized in Table 1. The surface
-
1
weighted EXAFS oscillation in the range of 2.8-14 Å .
Multiple shell fitting of the EXAFS data was conducted in the
R-space.
Catalytic Measurement. Selective reduction of NO with NH3
was performed over a catalyst bed packed within a Pyrex tube
with 1 cm in diameter. The fixed bed consisted of 0.2-g catalyst
particles with a type K thermocouple placed at the center of
the bed for temperature measurements. Reactant gases (640 ppm
NO, 640 ppm NH3, and 1 vol % O2) were fed to the reactor
through mass flow controllers, and helium was added to make
3
2
up a total flow rate of 300 cm /min. The gas hourly space
areas of UV100, ST01, and NT are large (>300 m /g), except
velocity (GHSV) was about 10 000/h for this system. The NO
that of P25. The pore volume of the commercially available

Structure of Cu Species on TiO Nanotubes
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25819
Figure 2. Powder XRD patterns of the specimens obtained after the
consecutive treatments on the P25 TiO precursor: the hydrothermal
2
autoclaving in NaOH (a), the posttreatment acid washing (b), and finally
the calcination at 400 °C (c).
TABLE 1: The Physical Characteristics of the TiO
Supports
2
aVerage pore
diameter
(nm)
surface area
pore Volume
2
3
samples
(m /g)
(cm /g)
P25
50
340
320
390
0.13
0.35
0.60
1.2
14
UV100
ST01
NT
5.6
7.4
12
and to minimize the possible textural variation during the SCR
experiments, which were conducted at 100-350 °C. Because
the calcination caused agglomeration of the TiO2 supports, Cu/
UV, Cu/ST, and Cu/NT showed surface areas of ca. 100, 110,
2
and 190 m /g, respectively, and pore volumes of ca. 0.25, 0.54,
3
and 0.86 cm /g. Supported active oxides generally sinter upon
thermal treatment. However, no copper oxide peaks can be
observed in the XRD patterns of the Cu/TiO2 catalysts shown
in Figure 3. This indicates that Cu species were well-dispersed
to a certain extent and that the crystalline size of copper oxides
was too small to be detected by the diffraction. Only the anatase
TiO2 phase can be detected by XRD for all the Cu/TiO2 catalysts
shown in Figure 3. As to the Cu/P25 catalysts, catalytic
measurements showed that Cu/P25 exhibited a relatively low
SCR activity (Figure 4), and its structure was not further
explored in the following discussion.
Figure 1. TEM images of the specimens obtained after the consecutive
treatments on the P25 TiO precursor: the hydrothermal autoclaving
2
in NaOH (a), the posttreatment acid washing (b), and finally the
calcination at 400 °C (c).
samples is mainly contributed by the interstices between the
nanoparticles. The pore size, which corresponds to the interstice
width, would increase with the size of the constituting nano-
particles. Accordingly, an order of P25 > ST01 > UV100 for
the particle size is thus expected. The pore volume of the
nanotubes, on the other hand, is contributed by the internal
volume inside the tubes as well as the interstices between the
tubes. This large-surface-area feature for these TiO2 materials
is advantageous for dispersion of active species.
The catalysts from the incipient-wetness impregnation, Cu/
ST, Cu/UV, and Cu/NT, were all subjected to activity analysis
with the SCR reaction. The temperature dependence of the
steady-state NO conversion with NH over the Cu/TiO catalysts
3
2
is shown in Figure 4. Preliminary experiments have revealed
that the activity of all the bare TiO supports in NO reduction
2
was negligibly low, and the data are not presented. The activity
increased significantly upon Cu loading. Figure 4 shows that,
for these catalysts with 2 wt % Cu content, a conversion
maximum can be reached at temperatures of ca. 300 °C and
All the Cu-loaded catalysts prepared from the Cu(NO3)2
impregnation had been calcined at 400 °C to remove the nitrate

25820 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Nian et al.
Figure 3. Powder XRD patterns of the Cu-loaded (2 wt %) TiO
2
catalysts.
Figure 5. XANES spectra (a and c) and their first derivatives (b and
2 2
d) of Cu, Cu O, CuO, and the Cu-loaded (2 wt %) TiO catalysts.
and the lack of an intense peak below 8985 eV both are the
2
+
18-20
characteristic features for the Cu species absorption.
The
2
+
dipole-allowed 1s f 4p transition for Cu can be observed as
an intense peak at ca. 8986 eV, while those for Cu and Cu
can be observed at ca. 8982 and 8983 eV, respectively. To
make the XANES features more distinct, the first derivatives
of the XANES spectra have been taken and are shown in Figure
0
1+
18
1
9
2
+
5
b. The derivative spectrum of Cu in CuO shows a more
resolved 8978 eV feature for the 1s f 3d transition and a main
peak, which corresponds to the edge energy for Cu in the
Figure 4. Variation of steady-state NO conversion with temperature
for reactions over the Cu-loaded (2 wt %) TiO catalysts. The reacting
, 1 vol % O , and
2+
2
tetragonal symmetry, positioned at 8983.5 eV for the 1s f 4p
stream compositions: 640 ppm NO, 640 ppm NH
He as the diluent.
3
2
2
+
transition. The Cu species in the octahedral symmetry (e.g.,
2+
Cu in Cu(OH)2) would have an edge energy ca. 4.5 eV larger
2
+
19a
than that of Cu in the tetragonal symmetry. By contrast,
the conversion drops with further increase of the temperature.
The drop in NO conversion at high temperatures was attributed
0
the edge-energy features of the 1s f 4p transition for the Cu
and Cu derivative spectra appear at ca. 8979 and 8980 eV,
respectively (Figure 5b).
1
+
to the activated oxidation of NH3 (i.e., nonselective reduction
of O2 with NH3) to NO.12 We have also observed a lowering
of the maximum-conversion temperature by increasing Cu
content beyond 2 wt %. Therefore, 2 wt % Cu/TiO2 is the
content used for the structure and activity studies to avoid a
possible narrowing of the temperature window.
The XANES spectra of the catalysts and their first derivatives
are shown in Figure 5c and d, respectively. It can be seen,
especially from the derivative spectra, that the 1s f 4p transition
0
1+
features of Cu and Cu are not observed for all the three Cu-
containing catalysts, Cu/ST, Cu/UV, and Cu/NT. Both the 1s
As to the effect of the support type, Figure 4 shows that the
NO conversion over the different Cu/TiO2 catalysts has an order
of NT > UV100 > ST01. To have such a difference in
conversion, there must be a significant difference in the activity
of the catalysts (i.e., the specific rate or turnover frequency),
especially at high conversions with reactions conducted at high
temperatures. There was little difference in the crystalline phase
of the TiO2 supports. The chemical structure of the active Cu
species must have played a key role determining the catalyst
activity.
To analyze the structural difference of the Cu species, the
catalysts were subjected to X-ray absorption fine structure
spectroscopic measurements (see Supporting Information for the
absorption spectra). In Figure 5a, Cu K-edge XANES spectra
of CuO, Cu2O, and Cu metal are shown. The presence of a
weak feature at 8978 eV (a dipole-forbidden 1s f 3d transition)
2
+
f 3d and 1s f 4p features for Cu , however, are present in
the spectra of the catalysts, indicating that the Cu species on
the catalysts are in the +2 state. A detailed inspection on the
derivative spectra of CuO and the catalysts reveals that the
2+
feature for the 1s f 4p transition for the Cu species dispersed
in TiO2 has a positive energy shift of ca. 2 eV relative to the
2
+
energy for Cu in the tetragonal symmetry (i.e., CuO). This
2
+
indicates that a fraction of Cu ions in the catalysts are in a
1
8c
4+
site distorted toward the octahedral symmetry. Because Ti
in TiO2 is in the octahedral symmetry and the bond distance of
19,21
Ti-O is smaller than that of Cu-O in the CuO lattice2,
it is
+
suggested that there should be a certain amount of Cu doped
4
+
into the Ti
sites of the TiO2 lattice for the catalysts.
Furthermore, the edge energy of the tubular Cu/NT is slightly
larger than those of the particulate Cu/ST and Cu/UV (Figure

Structure of Cu Species on TiO Nanotubes
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25821
TABLE 3: Structural Parameters, Bond Distance (R), and
Coordination Number (CN) of the Cu Species in the
Cu-Loaded (2 wt %) TiO
2
Catalysts Obtained from EXAFS
Analysisa
2
2
catalyst
Cu/ST
shell
R (Å)
CN
σ (Å )
Cu-O
Cu-Cu
Cu-Cu
Cu-Ti
Cu-O
Cu-Cu
Cu-Cu
Cu-Ti
Cu-O
Cu-Cu
Cu-Cu
Cu-Ti
Cu-O
Cu-Cu
Cu-Cu
Cu-Ti
1.932
2.947
3.315
3.837
1.944
3.026
3.173
3.802
1.936
3.013
3.205
3.807
1.934
3.071
3.296
3.751
3.57
5.00
1.81
1.36
3.52
4.94
1.90
1.40
3.60
5.05
0.37
2.01
3.53
4.94
0.29
2.80
0.005
0.026
0.025
0.009
0.005
0.013
0.006
0.015
0.006
0.023
0.007
0.017
0.006
0.023
0.004
0.020
Cu/UV
Cu/NT
Cu/NT1
3
Figure 6. The Fourier transformed k ø(k) EXAFS of the reference
2
CuO and Cu-loaded (2 wt %) TiO catalysts. The dashed line curves
denote the best fitting of the EXAFS spectra.
TABLE 2: Structural Parameters, Bond Distance (R), and
Coordination Number (CN) of CuO and TiO
2
Obtained
a
σ is the Debye-Waller factor.
from EXAFS Analysisa
2
2
catalyst
CuO
shell
R (Å)
CN
σ (Å )
tion shells were used to fit the EXAFS data of the catalysts.
The first and second shells have similar structural parameters
Cu-O
Cu-O
Cu-Cu
Cu-Cu
Cu-Cu
Cu-O
Cu-Cu
Ti-O
1.974
2.781
2.905
3.082
3.182
3.428
3.432
1.962
3.022
3.781
3.880
4
2
4
4
2
2
2
6
4
4
8
0.007
0.003
0.006
0.006
0.021
0.007
0.004
0.009
0.009
0.006
0.001
2
+- 2+
2+-
2+
and could be identified with Cu
O
and Cu Cu ion
correlations at ca. 1.9 and 3.0 Å, respectively, although the
coordination numbers are distinctly different from those in CuO.
The third shells at ca. 3.3 Å could be assigned to the contribution
from bonds corresponding to the Cu-Cu (3.18 Å), Cu-O (3.43
Å), and Cu-Cu (3.43 Å) coordination shells in pure CuO.
However, the coordination number varies significantly with the
catalyst type, showing few neighbors for Cu/NT compared to
ca. two coordinated atoms for Cu/ST and Cu/UV. The fourth
shell at ca. 3.8 Å cannot be accounted for using the CuO model.
This suggests, on the basis of structural parameters of pure TiO2
TiO
2
Ti-Ti
Ti-Ti
Ti-O
a
σ is the Debye-Waller factor.
2+
5
d), an indication of a higher fraction of Cu2+ doped into TiO2.
given in Table 2, a certain fraction of Cu dissolved into the
This aspect will be discussed further by using the EXAFS
analysis.
TiO2 lattice, in agreement with the interpretation by the XANES
analysis.
EXAFS data for CuO, Cu2O, and Cu metal and the catalyst
samples were collected, and the normalized spectra (see
Supporting Information) reveal that the local environment of
the Cu ion in the catalysts resembles that in CuO, rather than
that in Cu2O or in Cu metal. The Fourier transforms (FTs) of
There are four-coordinated Ti-Ti at 3.78 Å (the third shell)
and eight-coordinated Ti-O (the fourth shell) at 3.88 Å in pure
TiO2. These bond distances are quite close to those of the fourth
2+
shell in the catalysts. If most of the Cu ions dissolved in TiO2
4
+
are situated at the surface, the magnitude of third-shell Ti
3
2-
the k ø(k) EXAFS for the reference CuO and catalysts are shown
and fourth-shell O scattering should decrease because of the
Ti on the TiO2
surface has a coordination of 3 with third-shell Ti and of 4
18,19
4+
in Figure 6, with the radial distance ranging within 4 Å. The
spectra in the figure have been corrected for phase shifts. The
FTs for the catalysts are somewhat different from that of CuO,
especially exhibiting the lack of the strong outer-shell features
possessed by the FT for CuO (see Supporting Information). As
to the influence imposed by the supports, a noticeable difference
is that there shows a peak at ca. 3.3 Å in the FTs of Cu/ST and
Cu/UV while the tubular Cu/NT does not exhibit this feature.
In pure CuO, Cu is surrounded by Cu and O ions of outer shells
with bond distances near 3.3 Å (see Table 2). This difference
observed for the catalysts indicates a larger grain size of CuO
for the particulate Cu/ST and Cu/UV than for the tubular
Cu/NT.
The bond distance (R) and coordination number (CN) of
different shells obtained by the FEFF simulation are summarized
in Table 3 for the catalysts. The fitting range for all the samples
is 1-4 Å, and the bond distances given in the text and tables
are phase-corrected values. Because the XANES analysis of the
catalysts reflected that the chemical environment of the Cu ions
is similar to that of CuO, we have used CuO as the primary
model for fitting of all the catalysts. The structural parameters
of CuO model compound are given in Table 2. Four coordina-
lower coordination number of surface sites.
4+
2
-
2+
with fourth-shell O . Therefore, Cu ions of the catalysts
4
+
stabilized in Ti sites should have a total coordination number
of 7 for the fourth shell at ca. 3.8 Å if the solid solution at
surface is assumed to be in the form of CuxTi1-xO2 by assuming
that the oxygen ion vacancy caused by Cu dissolution is
negligible. Taking Cu/ST as an example, the fourth shell of
Cu/ST was found to contain 1.36 (Table 3) rather than the 7
ions expected for pure TiO2 surface. This indicates that only
19% ()1.36/7) of the Cu in the catalyst is dissolved in TiO2,
on the basis of a linear-combination principle, with 81% of
the Cu in CuO or CuO-like phases. We subjected these
percentage values of Cu distribution to further examination. For
the first shell of the catalysts, the contribution cannot be ascribed
solely to first-shell O of CuO at 1.96 Å. Contribution from
first-shell O scattering of TiO2 at 1.95 Å should be taken into
account because of the presence of Cu ions dissolved in TiO2.
Cu should have a coordination number of 5 (with O ) if
Cu ions are substituted for surface Ti in TiO2. The first
coordination of Cu in small-size CuO or CuO-like species is
3 (O ). By adopting the preceding Cu-distribution percentages,
2
+
2
+
1
8
2
+
2
-
2-
2+
2
+
2-
2
+
4+
2
+
2-

25822 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Nian et al.
the first-shell coordination number of Cu/ST is calculated to
have a value of 3.38 ()5 × 19% + 3 × 81%), which is close
to the value obtained from the EXAFS analysis (i.e., 3.57 from
Table 3).
The larger fourth-shell CN for Cu/NT than for Cu/ST and
2+
Cu/UV reflects a higher degree of Cu dissolution into TiO2
lattice. On the basis of the analysis for Cu distribution shown
above, the percentage of Cu2+ dissolution in TiO2 was found
to be 20% for Cu/UV and 29% for Cu/NT. The first-shell
coordination numbers calculated from these percentages, 3.40
and 3.58, respectively, for Cu/UV and Cu/NT, agree well with
the EXAFS data shown in Table 3. As to the second shell of
the catalysts at a distance near 3.0 Å, the contribution can be
assigned, on the basis of the data in Table 2, to second-shell
2
-
2+
O
and third- and fourth-shell Cu scattering of CuO as well
2
+
as second-shell Ti scattering of TiO2. Because of the more
multitudinous existence of possible neighbors at a distance near
3
.0 Å, the larger coordination numbers observed for the second
shell of the catalysts, in comparison with those for the first shell,
are as expected. The third shell (at a distance of ca. 3.3 Å) of
the catalysts has been ascribed to contribution from the outer
shells of CuO (fifth to seventh shells). It is of interest to observe
the low-degree occupancy of this shell for the nanotube-
supported Cu species while Cu/ST and Cu/UV have ca. 2
coordination for this shell. This observation clearly reflects that
the CuO or CuO-like clusters present on the nanotube support
are much smaller in size than those supported on ST01 and
UV100, thus leading to few neighbors detected at this higher
coordination shell. Intercalation of Cu2 into the multilayer tube
wall may have prevented the agglomeration of CuO over the
NT surface during calcination. The ion-exchange function of
Figure 7. H
) TiO catalysts. Peak assignments: R for highly dispersed CuO, â
2
2
consumption profiles by TPR for the Cu-loaded (2 wt
%
2
+
for Cu in TiO lattice, and γ for bulk CuO.
2
between the tubular and particulate catalysts. There was an
obvious peak observed at ca. 320 °C for the particulate Cu/ST
and Cu/UV, while the spectra of the tubular catalysts do not
show the presence of peaks at temperatures above 300 °C. The
H2-consumption patterns can be decomposed into three peaks
+
(
dashed lines in Figure 7) responsible for the reduction of
2c,9c
the titanate nanotubes, as a matter of fact, has been reported.
2
+
different Cu species. The 320 °C peak (γ) can be assigned to
reduction of bulk CuO.
The results presented above suggest the capability of the
19,22
For temperatures below 300 °C, the
2
+
layered structure of the titanate in accommodating Cu ions
inside the layers during the course of impregnation. This leads
to higher degrees of CuO dispersion and CuxTi1-xO2 formation
and thus to the higher activity of the resulting catalyst. To further
improve the dispersion of Cu2 over the nanotubes, we
employed the in-situ ion-intercalation method, in which Cu ions
can intercalate into the interlayer of the titanate species during
titanate nanotube formation. Under this circumstance, a large
fraction of Cu2 could be contained in the layered tube wall
after the peeling-off of the Cu-containing layers and subsequent
scrolling into tubes. After calcination, the thus-synthesized
catalyst is designated as Cu/NT1. The SCR measurement was
also conducted over Cu/NT1 and a very high activity, in
comparison with those of the other catalysts, can be observed
in Figure 4. The XANES and EXAFS spectra shown in Figures
lower-temperature peak, R, can be assigned to reduction of
highly dispersed CuO while the higher-temperature â peak to
2+
22
Cu in the TiO2 lattice because of metal-support interaction.
The negligible presence of the γ peak for Cu/NT and Cu/NT1
+
reflected the higher degree of Cu dispersion on the TiO2
2+
nanotubes than on the nanoparticles. The Cu ions in the
tubular catalysts were either situated in highly dispersed CuO
or dissolved in TiO2. The relative intensities of the â peaks were
+
2
0% for Cu/ST, 22% for Cu/UV, 28% for Cu/NT, and 42% for
2+
Cu/NT1, which were analogous to the fractions of Cu
dissolution in TiO2 estimated by the EXAFS analysis. As to
the peak temperature, the R peak for Cu/NT1 had a higher peak
temperature than those for the others. This may suggest that
2+
the R-Cu in Cu/NT1 was, as a matter of fact, situated in a
transition phase between the highly dispersed CuO and TiO2
lattice.
5
and 6 reflect that the Cu species contained in Cu/NT1 is
2+
principally Cu , a situation the same as those for the other
catalysts. Similar to Cu/NT, the lack of the peak at 3.3 Å in the
The above characterizations reflected a higher degree of Cu
dispersion obtained by using the titanate nanotubes as the
support. The hypothetical scheme for the formation of these
Cu/TiO2 catalysts is depicted in Figure 8, in which the tube is
presented as the projection along the tube axis. The intercalation
2+
EXAFS reflects again a high dispersion of Cu . The bond
distance and coordination number of different shells obtained
from the EXAFS simulation for Cu on Cu/NT1 are also shown
in Table 3. On the basis of the Cu-distribution analysis shown
above, Cu/NT1 was found to have 40% of Cu2 dissolution in
TiO2. This in-situ ion-intercalation method is indeed more
efficient in producing CuxTi1-xO2 solid solution for Cu sup-
ported on TiO2. The dissolved Cu may have served as one form
of the active sites for the SCR reaction,19 thus leading to the
high activity of Cu/NT1.
2+
of Cu ions into the interlayer space on the tube wall may
+
have inhibited agglomeration of the formed oxide species during
calcination. To form the anatase phase, the figure shows that
condensation of the titanate layers and thus rupture of the tubular
10f
framework are essential. EXAFS and TPR both reflected the
negligible presence of bulk CuO species on the nanotube-
supported Cu catalysts. The ion-intercalating feature of the
2+
H2 uptake profiles over the catalysts during temperature-
programmed reduction (TPR) are shown in Figure 7. The TPR
spectra show a distinct difference in the H2-consumption pattern
support also promoted the dissolution of Cu ions into the TiO2
lattice during calcination. By contrast, the agglomeration was
easier for CuO supported on nanopartices and led to the

Structure of Cu Species on TiO Nanotubes
J. Phys. Chem. B, Vol. 110, No. 51, 2006 25823
3
EXAFS spectra, and Fourier transformed k ø(k) EXAFS of Cu,
Cu2O, CuO, and Cu-loaded TiO2 catalysts. This material is
available free of charge via the Internet at http:/pubs.acs.org.
References and Notes
(
1) Zanella, R.; Giorgio, S.; Henry, C. R.; Louis, C. J. Phys. Chem. B
002, 106, 7634.
2) (a) Hinz, A.; Larsson, P.-O.; Skåman, B.; Andersson, A. Appl. Catal.
2
(
B 2001, 34, 161. (b) Perkas, N.; Wang, Y.; Koltypin, Y.; Gedanken, A.;
Chandrasekaran, S. Chem. Commun. 2001, 988. (c) Bavykin, D. V.; Lapkin,
A. A.; Plucinski, P. K.; Friedrich, J. M.; Walsh, F. C. J. Catal. 2005, 235,
10.
(
3) (a) Long, R. Q.; Yang, R. T.; Chang, R. Chem. Commun. 2002,
4
52. (b) Qi, G.; Yang, R.T. Appl. Catal. B 2003, 44, 217. (c) Pe n˜ a, D.A.;
Uphade, B. S.; Smirniotis, P. G. J. Catal. 2004, 221, 421. (d) Popa, A. F.;
Mutin, P. H.; Vioux, A.; Delahay, G.; Coq, B. Chem. Commun. 2004, 2214.
(e) Ciardelli, C.; Nova, I.; Tronconi, E.; Chatterjee, D.; Bandl-Konrad, B.
Chem. Commun. 2004, 2718. (f) Bourikas, K.; Fountzoula, Ch.; Kordulis,
Ch. Appl. Catal. B 2004, 52, 145.
Figure 8. Scheme for the formation of the tubular and particulate Cu/
(4) (a) Komova, O. V.; Simakov, A. V.; Rogov, V. A.; Kochubei, D.
I.; Odegova, G. V.; Kriventsov, V. V.; Paukshtis, E. A.; Ushakov, V. A.;
Sazonova, N. N.; Nikoro, T. A. J. Mol. Catal. A 2000, 161, 191. (b) Su a´ rez,
S.; Jung, S. M.; Avila, P.; Grange, P.; Blanco, J. Catal. Today 2002, 75,
2+
TiO
structures. The Cu
for NO reduction with NH
2
catalysts from calcination on the Cu -loaded supports of different
x
Ti1-xO
2
is considered one form of the active sites
3
.
331.
(
5) (a) Bond, G. C. Heterogeneous Catalysis: Principles and Applica-
2
+
formation of bulk CuO, as shown in Figure 8. The Cu -
substituted structure on the TiO2 supports should be assigned
one form of the active sites in the SCR because the degree of
the Cu2 dissolved had an order identical to that for the activity,
that is, Cu/NT1 > Cu/NT > Cu/UV > Cu/ST. The in-situ ion-
intercalation method to prepare Cu/NT1 was especially effective
tions; 2nd ed.; Oxford University Press: Oxford, U.K., 1987. (b) Chmielarz,
L.; Kustrowski, P.; Zbroja, M.; Rafalska-Lasocha, A.; Dudek, B.; Dziembaj,
R. Appl. Catal. B 2003, 45, 103.
(6) (a) Yonemitsu, M.; Tanaka, Y.; Iwamoto, M. Chem. Mater. 1997,
+
9
, 2679. (b) Yonemitsu, M.; Tanaka, Y.; Iwamoto, M. J. Catal. 1998, 178,
207. (c) Liu, C. C.; Teng, H. Appl. Catal. B 2005, 58, 69.
(7) (a) Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino, T.; Niihara, K.
Langmuir 1998, 14, 3160. (b) Kasuga, T.; Hiramatsu, M.; Hoson, A.; Sekino,
T.; Niihara, K. AdV. Mater. 1999, 11, 1307.
2
+
in intercalating Cu , thus leading to the high activity of the
resulting catalyst. It is expected that this tubular titanate material
can be used to prepare anatase TiO2-supported metal oxides, in
addition to CuO, with a high activity in heterogeneous catalysis.
(8) (a) Seo, D.-S.; Lee, J.-K.; Kim, H. J. Cryst. Growth 2001, 229,
428. (b) Wang, Y. Q.; Hu, G. Q.; Duan, X. F.; Sun, H. L.; Xue, Q. K.
Chem. Phys. Lett. 2002, 365, 427. (c) Yao, B. D.; Chan, Y. F.; Zhang, X.
Y.; Zhang, W. F.; Yang, Z. Y.; Wang, N. Appl. Phys. Lett. 2003, 82, 281.
(
d) Tsai, C.-C.; Teng, H. Chem. Mater. 2004, 16, 4352. (e) Bavykin, D.
V.; Parmon, V. N.; Lapkin, A. A.; Walsh, F. C. J. Mater. Chem. 2004, 14,
370. (f) Bavykin, D. V.; Gordeev, S. N.; Moskalenko, A. V.; Lapkin, A.
A.; Walsh, F. C. J. Phys. Chem. B 2005, 109, 8565. (g) Saponjic, Z. V.;
Dimitrijevic, N. M.; Tiede, D. M.; Goshe, A. J.; Zuo, X.; Chen, L. X.;
Barnard, A. S.; Zapol, P.; Curtiss, L.; Rajh, T. AdV. Mater. 2005, 17, 965.
Conclusions
3
The present study has shown that, through intercalation of
Cu2 ions in the large surface-area titanate nanotubes, a high
activity TiO2-supported CuO catalyst can be obtained by
calcining the Cu-loaded nanotubes. In comparison with the
catalysts using TiO2 nanoparticles as the support, the Cu species
on the developed tubular Cu/TiO2 was less prone to agglomerate
under calcination and exhibited a higher activity in selective
catalytic reduction of NO with NH3. X-ray absorption analysis
on the catalysts showed that the dispersion of CuO as well as
+
(9) (a) Du, G. H.; Chen, Q.; Che, R. C.; Yuan, Z. Y.; Peng, L.-M.
Appl. Phys. Lett. 2001, 79, 3702. (b) Chen, Q.; Zhou, W.; Du, G.; Peng,
L.-M. AdV. Mater. 2002, 14, 1208. (c) Sun, X.; Li, Y. Chem. Eur. J. 2003,
9, 2229. (d) Zhang, S.; Peng, L.-M.; Chen, Q.; Du, G. H.; Dawson, G.;
Zhou, W. Z. Phys. ReV. Lett. 2003, 91, 256103. (e) Zhu, H. Y.; Lan, Y.;
Gao, X. P.; Ringer, S. P.; Zheng, Z. F.; Song, D. Y.; Zhao, J. C. J. Am.
Chem. Soc. 2005, 127, 6730. (f) Pavasupree, S.; Suzuki, Y.; Yoshikawa,
S.; Kawahata, R. J. J. Solid State Chem. 2005, 178, 3110. (g) Yu, H.; Yu,
J.; Cheng, B.; Zhou, M. J. Solid State Chem. 2006, 179, 349. (h) Bavykin,
D. V.; Friedrich, J. M.; Alexei, A.; Lapkin, A. A.; Walsh, F. C. Chem.
Mater. 2006, 18, 1124. (g) Ferreira, O. P.; Souza, Filho, A. G.; Mendes,
Filho, J.; Alves, O. L. J. Braz. Chem. Soc. 2006, 17, 393.
2+
the dissolution of Cu ions into the TiO2 lattice was promoted
2+
because of this Cu -intercalation. The activity in the NO
2
+
reduction was found to increase with the degree of Cu
dissolution, suggesting that the CuxTi1-xO2 type of solid solution
should be involved in the catalytic reaction. Temperature-
programmed reduction also confirmed the high degree of Cu
dispersion on the tubular catalysts by showing the negligible
presence of the bulk CuO. A specific feature of this tubular
titanate is its structural similarity with the TiO2 anatase. Taking
advantage of this feature and the layered structure, metal ion-
loaded TiO2 catalysts with high degrees of metal-ion dispersion
and solid solution formation, and thus a high activity, can be
obtained.
(10) (a) Yang, J.; Jin, Z.; Wang, X.; Li, W.; Zhang, J.; Zhang, S.; Guo,
X.; Zhang, Z. Dalton Trans. 2003, 3898. (b) Ma, R.; Bando, Y.; Sasaki, T.
Chem. Phys. Lett. 2003, 380, 577. (c) Zhang, M.; Jin, Z.; Zhang, X. G.;
Yang, J.; Li, W.; Wang, X.; Zhang, Z. J. Mol. Catal. A 2004, 217, 203. (d)
Ma, R.; Fukuda, K.; Sasaki, T.; Osada, M.; Bando, Y. J. Phys. Chem. B
2
005, 109, 6210. (e) Tsai, C. C.; Teng, H. Chem. Mater. 2006, 18, 367. (f)
Nian, J. N.; Teng, H. J. Phys. Chem. B 2006, 110, 4193.
11) P aˆ rvulescu, V. I.; Grange, P.; Delmon, B. Catal. Today 1998, 46,
33.
(12) (a) Jeong, S. M.; Jun, S. H.; Yoo, K. S.; Kim, S. D. Ind. Eng.
(
2
Chem. Res. 1999, 38, 2210. (b) Curtin, T.; Regan, F. O.; Deconinck, C.;
Kn u¨ ttle, N.; Hodnett, B. K. Catal. Today 2000, 55, 189.
(13) (a) Petunchi, J. O.; Hall, W. K. Appl. Catal. B 1993, 2, L17. (b)
Komatsu, T.; Ogawa, T.; Yashima, T. J. Phys. Chem. 1995, 99, 13053. (c)
Cant, N. W.; Cowan, A. D. Catal. Today 1997, 35, 89. (d) Corma, A.;
Forn e´ s, V.; Palomares, E. Appl. Catal. B 1997, 11, 233. (e) Long, R. Q.;
Yang, R. T. Ind. Eng. Chem. Res. 1999, 38, 873. (f) Kieger, S.; Delahay,
G.; Coq, B. Appl. Catal. B 2000, 25, 1.
Acknowledgment. This research is supported by the Na-
tional Science Council of Taiwan (NSC 94-2214-E-006-007).
We also thank Dr. Jyh-Fu Lee and Mr. Yu-Li Tai of the Taiwan
Synchrotron Radiation Research Center for their help on the
analysis of the XANES and EXAFS spectra.
(
14) (a) Pasel, J.; K a¨ âner, P.; Montanari, B.; Gazzano, M.; Vaccari, A.;
Makowski, W.; Lojewski, T.; Dziembaj, R.; Papp, H. Appl. Catal. B 1998,
8, 199. (b) Zhu, Z.; Liu, Z.; Liu, S.; Niu, H.; Hu, T.; Liu, T.; Xie, Y.
Appl. Catal. B 2000, 26, 25. (c) Hsu, L. Y.; Teng, H. Appl. Catal. B 2001,
5, 21. (d) Teng, H.; Hsu, L. Y.; Lai, Y. C. EnViron. Sci. Technol. 2001,
35, 2369. (e) Hsu, L.Y.; Teng, H. Appl. Catal. B 2003, 42, 69.
1
Supporting Information Available: The TEM image of the
Cu-containing tubular TiO2; the X-ray absorption spectra,
3

25824 J. Phys. Chem. B, Vol. 110, No. 51, 2006
Nian et al.
(
15) Lowell, S.; Shields, J. E. Powder Surface Area and Porosity, 3rd
ed.; Chapman & Hall: New York, 1991; pp 14-89.
16) Stern, E. A.; Newville, M.; Ravel, B.; Yacoby, T.; Haskel, D.
Physica B 1995, 209, 117.
17) (a) Powder Diffraction Files of the Joint Committee on Powder
(19) (a) Okamoto, Y.; Kubota, T.; Gotoh, H.; Ohto, Y.; Aritani, H.;
Tanaka, T.; Yoshida, S. J. Chem. Soc., Faraday Trans. 1998, 94, 3743. (b)
Bera, P.; Priolkar, K. R.; Sarode, P. R.; Hegde, M. S.; Emura, S.; Kumashiro,
R.; Lalla, N. P. Chem. Mater. 2002, 14, 3591.
(20) (a) Huang, Y. J.; Wang, H. P. J. Phys. Chem. A 1999, 103, 6514.
(b) Huang, Y. J.; Wang, H. P.; Lee, J. F. Appl. Catal. B 2003, 40, 111. (c)
Liu, C. C.; Teng, H. Appl. Catal. B 2005, 58, 69.
(21) (a) Bouchet, R.; Weibel, A.; Knauth, P.; Mountjoy, G.; Chadwick,
A. V. Chem. Mater. 2003, 15, 4996. (b) Zhao, L. Y.; Yang, P. C.; Wang,
X. K.; Xie, Y. N.; Wu, N. Z.; Xie, Y. C. Appl. Surf. Sci. 2004, 228, 257.
(22) (a) Zhou, R. X.; Yu, T. M.; Jiang, X. Y.; Chen, F. Zheng, X. M.
Appl. Surf. Sci. 1999, 148, 263. (b) Hern a´ ndez-Huesca, R.; Santamar ´ı a, J.;
Braos-Garc ´ı a, P.; Maireles-Torres, P.; Rodr ´ı guez-Castell o´ n, E.; Jim e´ nez-
L o´ pez, A. Appl. Catal. B 2001, 29, 1. (c) Jiang, X. Y.; Lu, G. L.; Zhou, R.
X.; Mao, J. X.; Chen, Y.; Zheng, X. M. Appl. Surf. Sci. 2001, 173, 208. (d)
Chary, K. V. R.; Sagar, G. V.; Naresh, D.; Seela, K. K.; Sridhar, B. J.
Phys. Chem. B 2005, 109, 9437.
(
(
Diffraction Standards; Card No. 47-0124; International Center for Diffraction
Data: Newton Square, PA, 2000. (b) Sugita, M.; Tsuji, M.; Abe, M. Bull.
Chem. Soc. Jpn. 1990, 63, 1978.
(18) (a) Kau, L. S.; Spira-Solomon, D. J.; Penner-Hahn, J. E.; Hodgson,
K. O.; Solomon, E. I. J. Am. Chem. Soc. 1987, 109, 6433. (b) Lu, K.; Wan,
J. Phys. ReV. B 1987, 35, 4497. (c) Kau, L. S.; Hodson, K. O.; Solomon,
E. I. J. Am. Chem. Soc. 1989, 111, 7103. (d) Grunet, W.; Hayes, N. W.;
Joyner, R. W.; Shpiro, E. S.; Rafiq, M.; Siddigui, H.; Baeva, G. W. J. Phys.
Chem. 1994, 98, 10832. (e) Kuroda, Y.; Yoshikawa, Y.; Konno, S.; Hamano,
H.; Maeda, H.; Kumashiro, R.; Nagao, M. J. Phys. Chem. 1995, 99, 10621.
(
f) Beutel, T.; Sarkany, J.; Lei, G. D.; Yan, J. Y.; Sachtler, W. M. H. J.
Phys. Chem. 1996, 100, 845.