Content uploaded by Vitalii Petranovskii
Author content
All content in this area was uploaded by Vitalii Petranovskii on Aug 20, 2014
Content may be subject to copyright.
Available via license: CC BY-NC 3.0
Content may be subject to copyright.
RESEARCH Revista Mexicana de F´
ısica 59 (2013) 170–185 MARCH–APRIL 2013
Formation of copper nanoparticles in mordenites with variable
SiO2/Al2O3molar ratios under redox treatments
V. Petranovskiia,∗, E. Stoyanovb, V. Gurinc, N. Katadad, M.-A. Hernandeze, M. Avalosa, A. Pestryakovf,
F. Ch´
avez Rivasg, R. Zamorano Ulloag, and R. Portilloh
aCentro de Nanociencias y Nanotecnolog´
ıa, Universidad Nacional Aut´
onoma de M´
exico,
km 107 carretera Tijuana-Ensenada, Ensenada 22800, B.C., M´
exico.
∗Postal address: CNYN-UNAM, P.O. Box 439036, San Ysidro, CA 92143, USA
Tel.: +52(646)-174-4602; fax: +52(646)-174-4603.
e-mail address: vitalii@cnyn.unam.mx
bDepartment of Chemistry, University of California, Riverside, California 92521.
cResearch Institute for Physical Chemical Problems, Belarusian State University, Minsk 220080, Belarus.
dDepartment of Chemistry and Biotechnology, Tottori University, Tottori 680-8552, Japan.
eDepartamento de Investigaci´
on en Zeolitas, Universidad Aut´
onoma de Puebla, Puebla, M´
exico.
fTomsk Polytechnic University, Tomsk 634050, Russia
gDepartamento de F´
ısica ESFM-IPN,
Zacatenco, 07738. M´
exico, D.F.
hFacultad de Ciencias Qu´
ımicas, Universidad Aut´
onoma de Puebla, Puebla, M´
exico.
Received 14 August 2012; accepted 13 December 2012
A series of protonated copper-containing mordenites with different SiO2/Al2O3molar ratios (MR) in the range of 10≤MR≤206 was prepared
by ion exchange in copper nitrate aqueous solution. The electron paramagnetic resonance of hydrated copper Mordenites series testifies of
several Cu2+ ions sites. Hydrogen reduction of copper ions incorporated into the mordenites was shown to lead to different reduced copper
species including small metallic particles inter alia. The structural properties and acidity of mordenites were characterized. The optical
appearance of the copper particles showed strong but nonmonotonic dependence on the MR value, in line with the variation in acidity of this
series of mordenites. Correlations between mordenite properties and the formation of different reduced copper species are discussed.
Keywords: Mordenite; SiO2/Al2O3molar ratio; Copper; Nanoparticles; Plasmon resonance
Un conjunto de zeolitas mordenitas protonadas e intercambiadas con cobre y con diferentes relaciones molares (RM) de SiO2/Al2O3en el
intervalo 10 ≤MR≤206 ha sido preparado por intercambio i´
onico en soluci´
on acuosa de nitrato de cobre. La resonancia paramagn´
etica
electr´
onica del conjunto de mordenitas con cobre en su estado hidratado muestra varios sitios de iones Cu2+. La reducci´
on por hidr´
ogeno
de los iones de cobre incorporados en las mordenitas ha demostrado que la reducci´
on produce diferentes especies de cobre, incluyendo
la reducci´
on de peque˜
nas part´
ıculas met´
alicas inter alia. Se han caracterizado las propiedades estructurales y la acidez del conjunto de
mordenitas intercambiadas. La se˜
nal ´
optica de nanopart´
ıculas de cobre mostr´
o fuerte dependencia monot´
onica, pero no con el valor RM, en
l´
ınea con la variaci´
on de la acidez de este conjunto de mordenitas. Las correlaciones entre las propiedades de las mordenitas y la formaci´
on
de diferentes especies reducidas de cobre son discutidas.
Descriptores: Mordenita; relaci´
on molar SiO2/Al2O3; Cobre; EPR; Nanopart´
ıculas; Resonancia Plasm´
onica.
PACS: 78.67.Sc; 76.30.-v; 78.40.-q
1. Introduction
Transition metals incorporated into solid matrices are widely
used in the preparation of catalysts and nanocomposites with
unusual optical, electrical and magnetic properties. Such
materials have been extensively studied during recent years.
The presence of metal in ionic form or as small clusters
and nanoparticles in the host matrices drastically changes the
properties of these composite materials even if the dopants
are present at low concentrations. As regards zeolites, their
ion exchange properties permit introduction of metal into
zeolite voids, while subsequent reduction leads to different
reduced metal species [1-3]. The multicomponent metal-
dielectric systems produced by incorporation of metal into
zeolites are of special interest concerning the metal-support
interaction, and to understand the mutual contribution of dif-
ferent components to the properties of the whole system. Ze-
olites possess molecular sieving properties due to the well-
defined size of their pore openings. This feature can lead to
additional stabilization of small clusters whose size matches
the inner space of zeolite channels [4-9]. A great variety of
natural and synthetic zeolites with controllable properties and
different crystal structures are known [10,11]. Unlike some
matrices used for incorporation of metals, zeolites not only
provide mechanical support for topological reasons, but are
also an active ionic medium which promotes a rich chemistry
of metal ions. These are incorporated at specific sites in zeo-
lite crystals, together with other accompanying reagents, like
water for example, and interact with other exchangeable ions
and/or with the different Brønsted and Lewis acid sites [10].
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 171
Further chemical transformations with metal-zeolite systems
(redox processes, dehydration, metal particle aggregation,
etc.) generate new nanocomposites with complicated struc-
tures in which the properties of matrix, incorporated metal
and metal-matrix interactions all play a role.
Copper-containing zeolites are of particular interest due
to their catalytic activity in deNOxreactions [12-14]. Al-
though these compounds received considerable attention dur-
ing the last decade, there is still ambiguity about the interpre-
tation of certain copper species in different zeolites. Metal
ion distribution within these matrices strongly depends on the
different conditions of metal incorporation, and the effect of
subsequent processes such as heat treatment of Cu-zeolites in
different atmospheres adds to the complex final structure of
these composites.
It has been stated that the SiO2/Al2O3molar ratio (MR)
is the key-factor governing formation and relative stabil-
ity of silver clusters in mordenites. Changing the chemical
composition of the mordenite framework (i.e., changing the
MR, and, consequently, the acidity that depends on the MR)
leads to non-monotonic dependence of cluster stabilization
on MR [15,16]. The MR value of Cu-mordenite is one of
the crucial characteristics enhancing the hydrothermal sta-
bility of NO reduction catalysts under lean NOxwet con-
ditions [17]. Analysis by the Monte Carlo procedure shows
that the distinct distribution of Al atom pairs in -O-Al-O-(Si-
O-)nAl-O- fragments in mordenite creates sites that provide
favorable binding environments for extra-framework Cu2+
cations; this distribution of Al atoms is determined by the
SiO2/Al2O3ratio [18]. Thus, the coordination, localization
and stabilization of copper ions in zeolitic materials strongly
depend on the structure and composition of the zeolite ma-
trix. The mordenite framework possesses several positions
for the localization of cations in general, as well as for cop-
per ions [10,19,20]. In this way, the solid zeolite matrix plays
the role of a substrate with a range of polydentate ligands for
Cu2+. Since the redox behavior of the Cu2+/Cu+/Cu0sys-
tem is very sensitive not only to the medium but also to the
complexation of copper by different ligands, this should lead
to the appearance of different copper states in zeolites.
The aim of the present work was to investigatehow acid-
ity, secondary porosity and local structure features of mor-
denites depend on the MR, and how reduction temperature
during treatment in hydrogen influences the optical appear-
ance of reduced copper species in these matrices.
The copper state in mordenite samples was detected by
means of electron paramagnetic resonance (EPR) absorp-
tion and UV/Vis absorption using diffuse reflectance spec-
troscopy (DRS), the EPR technique gives information mainly
of spin transitions of Cu2+ ions and the UV/Vis-DRS tech-
nique is a convenient method to obtain optical information
on opaque (powder-like) materials. The contribution of dif-
ferent copper states is known to be rather pronounced in the
UV/Vis-DRS spectra, which allowed us to identify their ap-
pearance in the series of copper ionexchanged and reduced
mordenites.
The plan of this work is as follows. Section 2 shortly de-
scribes the experimental procedures and devices. Section 3.1
provides a structural characterization of the used set of mor-
denites, and explains the method to determine their acid prop-
erties. ESR spectra of hydrated samples are discussed in the
Secs. 3.2. From this EPR study we have found different Cu2+
ions sites in the set of the unreduced samples with a rela-
tive estimation of the amount of Cu2+ ions in each sample.
Sections 3.3 and 3.4 offer a detailed analysis of the exper-
imental data obtained by DRS of the series of samples un-
der study. We show the original mordenites, the copper-ion-
exchanged composites and the reduced mordenites with cop-
per. The behavior of mordenites with different MR appears to
be markedly different, and we therefore consider them sepa-
rately, summarizing obtained data in Sec. 3.5 and incorporat-
ing into the discussion our data on theoretical estimation of
the spectral appearance of reduced copper [21,22]. Section 4
concludes the Discussion and sums up the main results.
2. Experimental
Protonated forms of mordenites with MR varying from 10 to
206 were supplied by TOSOH Corporation, Japan. Copper
ion exchange was carried out at ambient temperature from
0.1 N Cu(NO3)2aqueous solution for 24 h under constant
stirring in excess of Cu2+. The samples were filtered, washed
and dried under ambient conditions followed by heating in
a dry H2flow at temperatures from 150 to 450◦C for 4 h.
The copper content in the prepared samples was determined
by atomic absorption spectrometry using Varian model 1475
equipment.
Surface area SBET and pore volume VΣwere deter-
mined by high-resolution nitrogen adsorption measurements
on an Autosorb-1 Quantachrome equipment in the range of
10−6<p/p0<1. The crystallinity and the cell constants
were monitored by X-ray diffraction (XRD) measurements
on a Philips diffractometer, model X’Pert, equipped with a
curved graphite monochromator, using Cu Kαradiation. Dif-
fuse reflectance spectra (DRS) were collected on a Varian
Cary 300 equipped with a standard diffuse reflectance unit
using a barium sulfate reference in the wavelength range 190-
850 nm. The spectra were obtained in air just after cooling
the hydrogen-treated samples.
The Brønsted acidity of the mordenite surface was stud-
ied using the method of non-aqueous potentiometric titra-
tion of the suspension of desiccated mordenites in anhy-
drous dimethylformamide medium by ethyl alcohol solution
of C2H5OK, using a pH673M potentiometer with platinum
and glass electrodes.
Temperature Programmed Desorption (TPD) measure-
ments with ammonia were performed using a TPD-1-AT Bell
Japan Inc. equipment and analyzed according to [23]. The
zeolite sample (0.1 g) was packed into a quartz cell. It was
evacuated at 773 K for 1 h, followed by ammonia adsorp-
tion (13.3 kPa) at 373 K. Weakly held ammonia was removed
by water vapor treatment [24]. Finally, the zeolite bed was
Rev. Mex. Fis. 59 (2013) 170–185
172 V. PETRANOVSKII et al
heated at 10 K·min−1under flowing helium (0.044 mol·s−1)
at reduced pressure (13.3 kPa), and the desorbed ammonia
was detected by mass spectrometer (ULVAC UPM-ST-200
or ANELVA M-QA 100 F). After the measurement, the peak
intensity was calibrated using a known amount of ammonia.
EPR spectra were measured with a Jeol Jes-Re3x spec-
trometer (JEOL Co., Ltd., Tokyo, Japan) equipped with a PC
using Esprit-425 software for the spectrometer control and
data acquisition at X-band (9.0 GHz), at a power of 1 mW
and modulation frequency of 100 KHz. The spectra were
measured at 300 and 77 K. The sample charges were around
20 mg placed in EPR-grade quartz tubes with i. d. of 2 mm.
The g values were determined in comparison with a small
signal of the standard DPPH (g = 2.0036). Experimental
EPR spectra were also simulated with the Esprit-425 Jeol
software, which uses a second-order perturbation theory for
the determination of the axial and rhombic A and g hyper-
fine parameters and the wide σof associated absorption-like
Gaussian lines. This program does not include the broaden-
ing effects of strains.
Throughout the text and figures, samples are abbreviated
as HMor for the initial protonated form of mordenite, and Cu-
Mor for Cu-exchanged forms, followed by the MR value and
temperature of hydrogen reduction in ◦C, if applicable (e.g.,
CuMor15-250).
3. Results and discussion
3.1. Characterization of H-mordenites
A set of mordenites, including deeply dealuminated ones
with MR values, ranging to 206 were used in this study. As-
suming that the crystal structure could be significantly influ-
enced during this treatment, crystallinity of the samples was
tested. XRD patterns for the set of HMor with varying MR
are shown in Fig. 1. The unit cell parameters obtained from
the experimental spectra by Rietveld refinement for HMor,
and reference data for NaMor from IZA Structure Commis-
sion [11] are shown in Table I. Parameters “a” and “c” were
less affected, while “b” changed significantly in comparison
with the reference data. There were no significant differences
in the values of “a”, “b” and “c” for the series of HMor sam-
ples. Reduction of the parameter “b” led to a small decrease
in channel aperture, and to an increase in its elliptical cross-
section aspect ratio, since the parameter “a” varied less. The
TABLE I. Lattice parameters for mordenites.
Sample a, ˚
A b, ˚
A c, ˚
A
NaMor [10] 18.11 20.53 7.528
HMor10 18.04 20.29 7.465
HMor15 18.15 20.29 7.487
HMor30 18.14 20.25 7.470
HMor206 18.07 20.27 7.470
FIGURE 1. XRD patterns of the set of protonated mordenites.
FIGURE 2. Isotherms of N2adsorption of the HMor set.
change in the “c” parameter evidences, that the lattice be-
came slightly denser in [001]-direction. Variation of XRD
patterns of H-mordenites was observed with increasing MR.
Thus, several reflections in the range of 2θ=13-18◦changed
their appearance and relative intensities, as did doublets at
23-24◦and 28◦. Remarkably, in spite of the variation in XRD
patterns, a systematic dependence of the lattice parameters on
MR cannot be inferred. Detailed discussion and refinement
of crystalline structure of this mordenite series is beyond the
scope of the present work and will be reported elsewhere.
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 173
TABLE II. Characteristics of HMor porosity and copper content measurements
HMor10 HMor15 HMor30 HMor206
SBET , m2/g 359 380 480 493
St, (external surface) m2/g 45 33 51 68
VΣ, cm3/g 0.200 0.174 0.238 0.261
Mesopore volume, cm3/g (Pore radii 2-50 nm) 0.07 0.05 0.06 0.05
Al concentration derived from MR, mol/kg 2.56 1.83 1.01 0.16
TPD-derived conc. of acid sites, mol/kg 0.44 1.24 0.49 0.13
Copper content after ion exchange, wt % 0.79 0.26 0.26 0.44
Maximum (theoretical) ion-exchange capacity
on the basis of MR, with respect to Cu2+, wt % of Cu 8.2 5.9 3.2 0.5
% of ion exchange 9.6 4.4 8.1 86.3
The most important conclusion for the discussion of this
paper is that sample crystallinity was not affected by the dea-
lumination treatment and neither the degradation of the crys-
talline pattern nor appearance of an amorphous phase was
observed.
The results of HMor porosity characterization and copper
content measurements are presented in Table II. The SB ET
value increased with MR growth; a maximum increment oc-
curred in the passage from HMor15 to HMor30. Meanwhile,
the total pore volume VΣreached a minimum for HMor15. It
is important to note that in the case of ideal mordenite crys-
tals possessing regular channels with elliptical cross-section
0.65×0.70 nm, the material had a micropore volume equal
to 0.14 cm3/g [10]. Reduction of the “b” and “c” parameters
(Table I) can lead to a decrease in this volume of not more
than ∼1%. The higher values of the observed micropore vol-
ume refer to perfect mordenite crystal, and the mesopore vol-
ume (with non-structural radii of up to 50 nm) can be asso-
ciated with defects in the microcrystals, probably originating
from the synthesis conditions employed by the manufacturer,
TOSOH Corporation of Japan (not reported). These provide
the additional contribution to VZdue to irregular porosity in-
side the crystal volume.
Variation of the total observed surface area showed that
the microstructure of these matrices depends on MR, al-
though the crystal structure only showed a subtle variation
with MR (Fig. 1). Some correlation of the data on the total
surface area and pore volumes exists, and the least “porous”
HMor15 was characterized by the lowest values of mesopore
volume. The mesopores can be associated with cleaved ar-
eas of microcrystals, defects, uneven surface, etc., and seem
to be nearly independent from the amount of Si substituted
by Al in the ideal tetrahedral TO4coordination. Meanwhile,
according to these data, the HMor15 showed the minimum
concentration of defects and other above-mentioned structure
features providing the increase in porosity. Thus, HMor15 is
expected to have the least amount of defects of crystal struc-
ture in the given series. This may be a reason of the further
featured behavior of copper ions after incorporation and re-
duction in this mordenite.
Isotherms of N2adsorption at 76 K (Fig. 2) confirm
these conclusions. The HMor15 sample, unlike the other
mordenites, exhibits a saturation zone in the interval of
0.1<p/p0<0.8indicating that the micropores are homoge-
neous, and only a mild increase in the capacity of adsorption
for p/p0>0.9was observed. For the samples HMor30 and
HMor206 the cycles of hysteresis type III [25] were better de-
fined, being highest for HMor10, and indicating the greater
contribution of mesopores in these samples.
Complete exchange of 2H+ions in HMor for Cu2+ ion
leads to the composition Cu2+
xAl2xSiyO96, where y/x=MR;
the amount of copper is expected to change definitely with
a change in MR. Copper content was measured for ion ex-
changed samples and the degree of the ion exchange was cal-
culated (Table II). Experimental results demonstrate that not
all the samples attained the complete exchange in spite of the
long contact (a day) of zeolite samples with the solution in
which the amount of Cu2+ ion was in large excess. Samples
HMor10, HMor15 and HMor30 were far away from com-
pletion of the ion-exchange (Table II), and only HMor206
approximated it to ∼86%. That is, almost all acid sites in
HMor206 participate in the copper incorporation, while in
the other samples only few of the sites are active in this pro-
cess. In fact, less than 10% of total occupancy is attained
for HMor10, HMor15, and HMor30, and Cu2+ ions are ex-
pected to occupy the more energetically favorable centers in
these mordenites. The lowest degree of exchange with re-
spect to copper ions occurred in the HMor15 sample, the
next larger being in HMor10. Such dependence of copper
ion-exchangeability vs. MR, with an extreme lying within
the MR range, can be associated with appearance only in the
HMor10 samples extra-framework Al atoms in 5-coordinated
state. This Al moiety has been characterized by 27Al MAS-
NMR [26] and shows the chemical shift δ=27 ppm (es-
sentially different from the value for the tetrahedrally co-
ordinated framework Al atoms, δ= 57 ppm [27]). Also,
Rev. Mex. Fis. 59 (2013) 170–185
174 V. PETRANOVSKII et al
all samples contain some amount of extra-framework octa-
hedrally coordinated Al atoms; these form additional active
Lewis sites with δ∼0ppm and can incorporate more Cu2+-
ions [26].
Thus, the regular MR increase of mordenite samples did
not result in a variation of the amount of exchangeable cations
only. Much more difference was observed in structure disor-
der both at the levels of micro- and meso-structure of these
zeolites (Table II). Not only ion-exchangeability, but also
other properties showed an extreme for HMor15.
Acid properties of H-mordenites with 10 ≤MR ≤206
were investigated by IR-spectroscopy [16] with adsorption
of probe molecules (CO and pyridine). It has been estab-
lished that concentration and strength of both Lewis and
Brønsted sites depend on MR in a complicated way. Two
types of Brønsted acid sites with corresponding IR bands,
νOH = 3610 and νOH = 3720 cm−1, were observed, their
concentration reaching a maximum for intermediate MR.
Also, H-mordenites possess four types of Lewis acid sites.
Their concentration is minimal for mordenites with interme-
diate MR, and it increases for higher and lower MR [16].
In the present work we studied the acid properties of
the selected mordenites in greater detail. Fig. 3 shows the
results of the ammonia TPD study. The measurements ar-
gue for a significant difference between the mordenites used,
and a non-monotonic dependence of the TPD pattern is ob-
served with increasing MR. There are at least three main
maxima at the TPD curves in Fig. 3, the relative contribu-
tion of which varied with MR. The lowest-temperature peak,
ca. 490 K, which dominates for the HMor10 sample, be-
longs to the weakly bonded ammonia probably adsorbed on
extra-framework Lewis sites. Intensity of this peak falls with
increasing MR. The TPD curve for HMor10 also shows a
shoulder from which a maximum at ca. 660 K can be re-
solved. The well-pronounced maximum at ca. 730 K occurs
for the other three mordenites with little shift to the lower
temperatures while changing from HMor15 to HMor206.
FIGURE 3. Ammonia TPD spectra of HMor samples.
FIGURE 4. Amount and strength of acid sites on H-mordenite with
various aluminum concentrations. Values for mordenites with in-
termediate MR (not applied in this work) are shown also. MR val-
ues for filled circles (right to left) are 10, 15, 20, 24, 30, 72, 110,
128 and 206 respectively.
These maxima are not symmetrical, and evident contributions
from some species developing a desorption peak at 660 K can
be evaluated.
The above three ammonia thermodesorption peaks, 450-
490 (very likely Lewis), 660 and 730 K (Brønsted), are asso-
ciated with the three different groups of acid sites, the amount
of which is very variable for different MR. On the basis of the
above-mentioned NMR data on the existence of the extra-
framework pentacoordinated Al in HMor10, accompanied
by the growth of ion-exchangeability with respect to copper
ions, we may suppose that these unusual Al atoms are respon-
sible for the acid properties with the lowest strength, result-
ing in the 450-490 K peak. Note also that, for the HMor15
sample, some amount of ammonia is desorbed even at tem-
peratures ≥900 K, so a population of extremely strong acid
sites probably exists for this sample and is absent for the other
three mordenites.
The amount of acid sites was calculated from the in-
tensity of NH3desorption peaks, and plotted against the
aluminum concentration in Fig. 4, solid line. From the
point of view of zeolite structure the ratio of [Al]:[OH] must
be 1:1 (Fig. 4, dashed line). Experimentally determined
concentration of acid sites is slightly lower than the the-
oretically expected concentration in the range from 0.1 to
∼1.5 molkg−1, and decreases with further increase in alu-
minum content. The maximum deviation is observed for
HMor10. As previously reported [28], at aluminum concen-
tration higher than 1.5 molkg−1, destruction of the crystal
structure of H-mordenite is induced by humidity, resulting in
the generation of Lewis acidity (TPD peak at 490 K, Fig. 3).
The acid strength was calculated by the curve-fitting
method as ammonia adsorption heat (Fig. 4). The acid
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 175
TABLE IV. Specific concentration and pKaof Brønsted acid sites for the set of mordenites.
HMor10 HMor15 HMor30 HMor206
BAS pKaCBAS,µmol/m2BAS pKaCBAS,µmol/m2BAS pKaCBAS,µmol/m2BAS pKaCBAS,µmol/m2
1.5 0.92 1.5 0.32
2.0 1.00 1.8 0.50 1.8 0.34
2.8 0.18 2.3 0.76 2.9 0.92 2.4 0.36
3.9 0.22 3.4 0.52 3.3 0.24
5.0 0.84 5.5 0.18 5.0 0.20 4.0 0.06
5.7 0.24 6.1 0.28 5.6 0.10 5.4 0.06
strength starts from ca. 180 kJ·mol−1for HMor206. With in-
creasing aluminum content acid strength falls in the range of
0.1-0.5 mol·kg−1, and finally in the range of 0.5-2 mol·kg−1
reaches a constant value, ca. 150 kJ·mol−1, and becomes
independent of aluminum concentration, in agreement with
Katada et al. [23]. The TPD curve for the HMor10 sample
manifests a bimodal distribution of acid sites, and it is hard to
define a mean value of their acid strength by this method. The
point corresponding to HMor10 sample in Fig. 4 is marked
for the high-temperature shoulder at ca. 660 K.
The extra-framework Al is expected to form less acid hy-
droxyl than other aluminum atoms in zeolites. On the other
hand, the main peak, 730 K, corresponds to the acid sites
with the highest concentration for HMor15 and HMor30.
Its amount in HMor15 is ∼2.5 times higher (Fig. 4) com-
pared with HMor30, which approximately correlates with
their MR values. However, comparing these amounts with
that of HMor10, the increase of Al concentration resulted in
a significant fall in the total amount of acid sites. The increase
in copper concentration in the CuMor10 and CuMor206 sam-
ples compared with the CuMor15 and CuMor30 samples can-
not be explained by the additional involvement of active sites
formed with the extra-framework Al since the total amount
of acid sites does not grow in these cases. Thus, we propose
that copper incorporation and ammonia bonding do not show
a unique quantitative correspondence, and hence, affinity of
the protonated forms of mordenite with respect to Cu2+ is
controlled in a more complicated way than indicated by the
acid-base ammonia interaction.
In the case of the HMor206 sample, the TPD curve
showed a peak at 660 K, similar to that for HMor10 and a
high temperature peak at 730 K. Thus, the HMor206 sam-
ple should have the rather high strength of acid sites (which
coincides with the high average adsorption heat of ammonia,
Fig. 4), but with low total amount.
As the ammonia TPD curve shows the results of the
whole set of acid sites integrated in a sample, it was of in-
terest to detect individual acid sites with different strength.
Surface concentration of nonequivalent Brønsted acid sites
(BAS) with different pKawas determined by non-aqueous
titration. Results are listed in Table III. All samples have a
number of different types of BAS in a wide range of pKa.
The HMor10 sample showed four types of BAS with weak
and moderate strength (2.8<pKa<5.7) that correlated
with the TPD data mentioned above. A large quantity of
strong BAS (1.5<pKa<2.3) was present in HMor15,
although some amount of sites of weak and medium strength
(5.5<pKa<6.1) were present on its surface in smaller
amounts. It is worthy of note that HMor15 also showed
the weakest acid sites with pKa=6.1, in line with the high
concentration of the BAS with the most acid properties
(pKa=1.5). The same BAS with pKa=1.5 were present at
HMor206 in relatively high concentration, while very low
amounts of other types of BAS with pKa>3.3were present
on the surface of this sample. Such distribution of acid cites
concentration can explain the high integrated heat of ammo-
nia desorption for HMor206.
3.2. EPR of H-mordenites and Cu-exchanged morden-
ites
EPR spectroscopy was employed to characterize the set of
hydrated HMor and CuMor with varying MR. In the case
of the set of hydrated HMor series, only HMor-10 present
a weak absorption at g=4.3, which can be assigned to tetra-
hedral Fe3+ impurities, substituting for tetrahedral silica
(Fig. 5, curve a). No EPR signal was observed at 300 K for
other HMor samples for low and high fields up to 600 mT
(Fig. 5). The EPR radical signal at g = 2.0036 of Fig. 5
comes from the DPPH marker.
The EPR spectrum of hydrated CuMor10 (Fig. 6a) shows
a resolved low field hyperfine interaction site, site H1, in its
hydrated ERP spectrum at room temperature; see Fig. 6a, and
Table 4. The setting of this hyperfine interaction H1, reflects
certain degree of immobilization of the Cu2+ cations possi-
bly due to the coordination with network oxygen. The room
temperature EPR spectra of hydrated CuMor15 (Fig. 7a), Cu-
Mor30 (Fig. 8a) and CuMor206 (Fig. 9a) are broad and struc-
tureless, with isotropic absorptions at gef f = 2.148, 2.140
and 2.148, respectively, indicating typical motional broad-
ening, that is, mobility of hydrated Cu2+ complex such as
Cu(H2O)2+
6in mordenite channels with bonding character-
istics similar to those of Cu(H2O)2+
6in solution, commonly
observed in hydrated Cu-exchanged zeolites [29,30].
Rev. Mex. Fis. 59 (2013) 170–185
176 V. PETRANOVSKII et al
FIGURE 5. EPR spectra of hydrated: a) HMor10, b) HMor15, c)
HMor30 and d) HMor206, measured at 300 K.
FIGURE 6. EPR spectra of CuMor10. a) Measured at 300 K. b)
Measured at 77 K. c) simulated EPR rhombic spectra of hydrated
CuMor10 measured at 300 K, the EPR parameters of this simula-
tion are shown in Table IV.
The liquid nitrogen temperature EPR spectra of the hy-
drated samples show the resolved low field features, belong-
ing to slightly different sites H2, H3 and H4 for CuMor10,
CuMor15 and CuMor30 respectively (Figs. 6b, 7b, 8b and
Table IV). These hyperfine low field features are due to the
freezing of motional effects, and are commonly observed in
hydrated Cu-exchanged zeolites below 120 K [30,31]. Only
the hydrated CuMor206 sample presents two hyperfine sites
H5 and H6 at 77 K. The four hyperfine lines in the low field
region come from the coupling to the copper nucleus with
I=3/2.
The experimental amplification of the low hyperfine re-
gion is shown below the each EPR spectra, characterized by
the four absorption-like lines showing a spectral broadening
increase from low to high fields, attributed to correlated g-
and A-strain effects [31,32]. This broadening dependence is
not observed in the four absorption-like lines of the EPR of
CuMor10 at 300 K (see EPR amplification of Fig. 6a).
The best EPR rhombic and axial hyperfine values at 300
and 77 K of paramagnetic Cu2+ ions of the CuMor set, ob-
tained by simulation (Table IV) have been attributed to mo-
tional freezing of hydrated Cu2+ complex in octahedral sites.
Listed in the Table IV simulated hyperfine EPR parame-
ters for sample CuMor10 at 300 K are assigned to a rhombic
site with gzz = 2.362, gxx = 2.135 and gyy = 2.077. The spec-
trum and its simulation (see Fig. 6c) show that Cu2+ ions in
CuMor10 present hyperfine features without strained effects.
This fact can be related to the lack of mobility of hydrated
Cu2+ complex [33], probably due to complex bonding to the
mordenite lattice [31].
The obtained hyperfine values of hydrated CuMor10, Cu-
Mor15 and CuMor30samples at 77 K have been simulated
FIGURE 7. EPR spectra of CuMor15. a) Measured at 300 K.
b) Measured at 77 K. c) Simulated EPR rhombic spectra of hy-
drated CuMor15 measured at 77 K, the EPR parameters of this
simulation are shown in Table IV.
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 177
TABLE IV. EPR parameters for hydrated Cu-Exchanged mordenites.
Sample Site gzz (gll)gxx (g⊥)gyy Azz (All )Axx(A⊥)Ay y σzz(σll )σxx(σ⊥)σy y Relative
(mT) (mT) (mT) (mT) (mT) (mT) intensitya
CuMor10 H1b2.362 2.135 2.077 13.0 1.5 0.0 7.7 10.5 5.8 0.09
H2 2.388 2.094 2.072 13.3 0.45 0.0 3.7 5.3 3.0 0.40
CuMor15 H3 2.400 2.097 2.073 12.9 0.05 0.0 3.5 5.7 2.9 1.00
CuMor30 H4 2.401 2.096 2.076 13.0 0.02 0.0 3.5 6.9 3.0 0.68
CuMor206 H5 2 .410 2.080 - 12.0 - 0.0 0.15
H6 2.373 2.080 - 14.1 - 0.0
Estimated errors gs=±0.003, As=±0.2mT, σs=±0.2mT, the sub-indices smeans the tensor components x, y and zof each EPR parameter.
aEstimated errors in relative intensity = 0.15.
bEPR parameters for spectrum measured at 300 K, all other EPR parameters belong to spectra measured at 77 K.
FIGURE 8. EPR spectra of CuMor30. a) Measured at 300 K.
b) Measured at 77 K.
as an axial site with gzz = 2.388, 2.400 and 2.401,
gxx = 2.094, 2.097 and 2.096 and gyy = 2.072, 2.073 and
2.076, respectively, shown in table IV. Finally, the EPR simu-
lation of hydrated CuMor206 sample at 77 K has been carried
out only for the gzz (gll)region with two axial sites, one with
gll = 2.410 and the second axial site with gll = 2.373.
Homogeneity of the copper state at 77 K in the samples
CuMor10, CuMor15 and CuMor30 demonstrates, that at low
Cu exchange levels (less than 10 %, see Table II) copper ion
in these samples occupies the only one possible site. The sec-
ond site begins to be populated only in the case of CuMor206,
with 86% Cu ion exchange (Table II).
In the last column of the Table IV, it is the EPR intensity
of the Cu2+-EPR signals of the CuMor set, normalized to the
intensity of the CuMor15 sample. This intensity is calculated
from the area of EPR adsorption as the second integral of the
measured EPR spectra and is proportional to the number of
resonant items in each sample. The founded intensity values
for CuMor10, CuMor15, CuMor30 and CuMor206 samples
FIGURE 9. EPR spectra of CuMor206. a) Measured at 300 K.
b) Measured at 77 K. c) Simulation of two axial sites of hydrated
CuMor206 spectra measured at 77 K, the EPR parameters of this
simulation are shown in Table IV.
are 0.40, 1.00, 0.68 and 0.15, respectively. These rela-
tive intensities are plotted as a bar graph in Fig. 10 and
it is shown that the amount of Cu2+ ions follows the order
CuMor15>CuMor30>CuMor10>CuMor206. As we have
observed in resent studies about copper reduction in Cu ex-
changed in Cu-Erionite [34] and in Cu-ZSM5 [35], one of the
precursors of metallic copper formed under reduction upon
hydrogen atmosphere are the Cu2+ ion observed by EPR. In
the case of Cu-Erionite we have found that the intensity of
the broad andstructureless ERP signal at room temperature
Rev. Mex. Fis. 59 (2013) 170–185
178 V. PETRANOVSKII et al
FIGURE 10. Bar graph of the relative intensities (second deriva-
tive of the EPR signal) for the set of CuMor samples. Intensities
normalized to sample CuMor15.
FIGURE 11. Absorption spectra of H-mordenites (solid lines) and
their Cu-exchanged forms (dotted lines) with MR 10 (a), 15 (b),
30 (c) and 206 (d).
decreases as the temperature of the reduction treatment in-
creases, then the copper EPR intensity can be used to moni-
tor the changes in Cu2+ population in a reduction process, an
ERP study over the CuMor series under reduction conditions
is actually in progress.
3.3. DRS of Cu-exchanged mordenites
The absorption spectra of the initial HMor and of the CuMor
samples prepared by ion exchange are shown in Fig. 11, a-d.
The mordenite matrix without copper showed the expected
absorption in the range λ < 350 nm. A clear maximum at
∼210 nm was observed for all mordenites under study. Other
less expressed features showed different appearances in the
samples: HMor10 had a weak feature at about 300 nm, which
was transformed into the shoulders for HMor15 and HMor30
samples, while HMor206 showed practically no absorption
at range λ > 280 nm. The absorption bands in this short-
wavelength range are inherent to zeolites of different types,
FIGURE 12. Subtractive absorption spectra, showing the copper
contribution in CuMor spectra. Inset displays the differential ab-
sorption at λ=620 nm for the sample CuMor10 (∆=D(CuMor10)-
D(CuMor15)) and much weaker ones for CuMor30 and CuMor206
samples.
and they can originate from the charge transfer O2−→Al3+
with participation of aluminum atoms at specific locations
(surface, corners, defects, etc.) [36].
The incorporation of copper led to the appearance of an
intensive band at ∼208 nm for all CuMor samples. This
band corresponds to the ligand-to-metal charge transfer tran-
sition from oxygen to Cu2+ [37]. Simultaneously, a weak
band with a maximum at ∼810 nm appeared (Fig. 11). This
band is typical for spin-allowed d−dtransition 2Eg→2T2g
of Cu2+ ion in pseudo-octahedral oxygen coordination, e.g.
Cu(H2O)2+
6[37,38] and Cu2+ in solid Al2O3matrix [39].
Differences in the spectra of CuMor and HMor allowed us
to extract the proper copper absorption in the CuMor spectra
(Fig. 12). Outlines of d-d bands for differential CuMor spec-
tra are in general the same; the shape of this band is somewhat
dissimilar only for CuMor10. This appearance of the band of
Cu2+ d−dtransitions can provide qualitative information
on the state of copper ions, as the acid properties of differ-
ent mordenites may influence their surrounding coordination.
The main part of the six oxygen ligands in the first coordina-
tion shell of Cu2+ belongs to water molecules, but a variable
minor part of them is from the zeolite framework. The num-
ber of framework oxygen atoms included in the coordination
shell must depend on the properties of mordenite and on the
different treatments applied to the sample. For the freshly
prepared samples the experimental maxima of d-d transitions
are shifted in the range 810-830 nm (Fig. 12). Blue shift
of this band in comparison with ∼900 nm for Cu(H2O)2+
6
ion in aqueous solutions indicates the increase of ligand field
while the water molecule is replaced by framework oxygen.
Short-wavelength slope of the d-d bands of CuMor30 and
CuMor206 samples shows a very weak band at ∼620 nm,
which is absent in the spectrum of the CuMor15 sample. So,
differences in the spectra of other samples compared to the
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 179
CuMor15 spectrum allows to extract the band at ∼620 nm
(Fig. 12, inset). This band shows the highest intensity in the
spectrum of the CuMor10 sample. The state of Cu2+ with
band at 620 nm corresponds to the distorted complexes, like
the square pyramidal or square planar complexes [40], ex-
treme cases of tetragonal distortion of the oxygen octahedron
due to the Jahn-Teller effect [41]. Thus, at least two states
of Cu2+, differing in coordination polyhedra are presented in
the samples CuMor10, CuMor30 and CuMor206 (Fig. 12, in-
set). Their appearance testifies to the difference in interaction
of the ions with mordenites depending on MR.
3.4. DRS of reduced Cu-mordenites
The DRS data of the reduced samples contain many new fea-
tures. Under reduction conditions the d-d band starts to disap-
pear, which indicates a degree of completeness of the reduc-
tion. The mordenite samples reduced in temperatures from
150 to 450◦C reveal that four new features that appear in the
spectra as well as strong changes in the two bands associated
with Cu2+ cited in the previous paragraph (∼810-830 nm
and 208 nm). Altogether, the following six features are seen
in spectra, as shown by arrows in Figs. 13-16:
(i) the charge transfer band of Cu2+ at ∼210 nm;
(ii) the small maximum at ∼250 nm;
(iii) the feature at 300-350 nm;
(iv) the feature at 400-500 nm;
(v) the band peaking at 550-600 nm with complex depen-
dence on the line shape while MR is changing;
(vi) the band of d-d transition of Cu2+ at ∼810-830 nm.
The maximum (i) of the charge transfer band diminished
with temperature increase. This temperature dependence al-
lows the assumption that the reduction process is not com-
plete at 150◦C but goes on up to 450◦C. Overlapping of this
band with the proper absorption of the mordenite matrix it-
self complicates an unambiguous attribution of this peak, es-
pecially in the case of its low intensity. Some small copper
clusters can also be proposed as species responsible for the
absorption in this region [21]. The charge transfer band cer-
tainly makes a contribution to absorbance with a maximum
at ∼210 nm, as long as the d-d band of Cu2+ (vi) is simulta-
neously present in the spectra.
The weak band (ii) at 250 nm coincides with a band of
isolated or poorly associated Cu+ions [42]. In general, in
contrast with the peaks (i) and (vi), as temperature rose, the
intensity of this absorption band (ii) was higher.
Features (iii), (iv) and (v) at 300-350 nm, 400-500 nm
and 550-600 nm respectively can be discussed together as be-
longing to the reduced Cu(0) products (copper nanoparticles
in the size range <10 nm) interpreted earlier in more detail
in [22]. The less pronounced but insistent bands (iii) and (iv)
FIGURE 13. Absorption spectra of CuMor10 reduced at different
temperatures.
FIGURE 14. Absorption spectra of CuMor15 reduced at different
temperatures.
appear in line with the main maximum (v), usually with in-
creasing temperature. It is known that the spectrum of copper
Rev. Mex. Fis. 59 (2013) 170–185
180 V. PETRANOVSKII et al
nanoparticles includes additional structure together with the
classical Mie-theory plasmon resonance due to features of the
copper band structure.
The band (v) at λ=550-600 nm appears and increases with
temperature for all mordenites (Figs. 13, 15, 16) except for
CuMor15 (Fig. 14). Temperature of appearance and shape
of this feature strongly depend on MR. It comes off in the
peak for reduced CuMor10-150, CuMor10-250 (Fig. 13) and
CuMor206-350, CuMor206-450 (Fig. 16) samples, or as a
step-like shoulder in the same range of wavelengths for Cu-
Mor30 starting from 250◦C (Fig. 15). It is of interest to note
that starting from the peak at low temperatures it changes to a
shoulder at higher temperatures only in the case of CuMor10
samples (Fig. 13). The shape of this feature for CuMor30
- CuMor206 samples does not depend on the temperature of
reduction, but its intensity (either peak or shoulder) varies
significantly (Fig. 15 and 16).
Simulation of the same spectral features (iii-v) for cop-
per nanoparticles based on the familiar relations from the
Mie theory allowed us to conclude [22] that size of the parti-
cles and the medium dielectric function ε0significantly affect
their intensity and shape, while the position of the band is
much less affected. The larger size of particles results in the
higher intensity of the plasmon resonance (v). The increase
of ε0also has an absorption-enhancing effect. The shoulder
in the range 550-600 nm can be attributed to small copper
particles located predominantly at the outer surface of mor-
denite microcrystals. The pronounced maximum at the same
range may be due to larger copper particles and to a higher di-
electric function of the medium. Since plasmon band appears
for the particles larger in size than regular mordenite channels
of ∼0.7 nm diameter, the copper particles can be formed in-
side mesopores existing in mordenite crystals. These could
be cleaved areas in which the mordenite matrix surrounds a
copper particle. In fact, the mordenites used in this work have
porosity at different levels (Table II). It is worth to note that
the use of non-crystallographic pores in crystals is currently
considered as a new approach in catalyst design [43,44].
Thus, the spectra in Figs. 13-16 evidence that formation
of the larger particles is more favorable in CuMor10 and Cu-
Mor206 and less favorable in CuMor30. In spite of the non-
monotonic variation of this phenomenon vs. MR, this fact
correlates with the rise in porosity and the decrease in acid-
ity of mordenite. The highest acidity of HMor15 leads to the
less favorable conditions for copper reduction and complete
absence of the plasmon resonance band (v).
In line with the above-mentioned bands, other features
like an increase in background absorption (Fig. 13, 16), or a
broad descent in absorption from the UV region to the end of
the available range (Fig. 13) may be seen in the spectra. Their
appearance, as well as the development of the all aforemen-
tioned peaks to their full extent depends on MR (Figs. 13-16),
and we consider them separately for different mordenites.
3.4.1. DRS of the reduced CuMor10
Peak intensity (i) significantly decreased with increasing
reduction temperature; absorbance at this wavelength re-
turned to the absorbance of the HMor10 sample (Fig. 13 and
Fig. 11). Simultaneously, the d-d band of Cu2+ (vi) was com-
pletely absent. Hence, the complete reduction of Cu2+ ions
was achieved in the CuMor10-450 sample and only the dif-
ferent reduced copper species were observed.
Unlike most other samples, CuMor10 demonstrates a
well-resolved maximum (ii) at 250 nm that develops from
the shoulder of CuMor10-150 to the well-pronounced peak
of CuMor10-250-CuMor10-450 (Fig. 12). Feature (iii) con-
tributes approximately in equal measure to all the spectra,
while feature (iv) is better resolved for the CuMor10-450
sample. The plasmon resonance band (v) shows a compli-
cated development, increasing in intensity from CuMor10-
150 to CuMor10-250, followed by a fall in intensity and
then another increase from CuMor10-350 to CuMor10-450,
in line with a remarkable change in the shape of this maxi-
mum, degenerating from a peak into a broad step-like feature.
This change of shape can provide evidence that the copper
particles formed at low reduction temperatures in the inte-
rior of mordenite crystals, are formed on their exterior parts
at higher temperatures. This variation of particle positions
can be explained by the increased mobility of copper atoms
at high temperatures, which favors their aggregation into the
particles on the outer surface of microcrystals. These areas
FIGURE 15. Absorption spectra of CuMor30 reduced at different
temperatures.
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 181
FIGURE 16. Absorption spectra of CuMor206 reduced at different
temperatures.
provide an environment with lower dielectric constants than
those of mesopores in the interior of zeolite [22].
It should here be mentioned that this complicated evo-
lution of the Cu(0) plasmon band (v) at different reduction
temperatures occurs only in CuMor10. Other mordenites
result just in one type of feature, either a maximum (Cu-
Mor206, Fig. 16) or a shoulder (CuMor30, Fig. 15), and
in some cases (CuMor15, Fig. 14) the feature does not ap-
pear at all. Thus, this peculiarity of the CuMor10 sample
may be associated with a specific distribution of acid sites:
only this sample demonstrates clearly the bimodal peak of
ammonia TPD (Fig. 3). It seems that the sites with lower
acidity favor the reduction of copper into Cu(0) located in-
side the mordenite matrix, while higher acidity give rise to
migration of reduced copper onto the outer surface. The
highest acidity (CuMor15) prevents particle formation com-
pletely. In fact, the highest contribution of acid sites in the
pKarange 5.0-5.7 (1.08 µmol/m2)is observed for HMor10,
while for HMor15 and HMor30 the greatest contributions are
due to BAS with pKa1.5-2.3 (2.68 µmol/m2)and pKa1.8-
3.4 (1.84 µmol/m2), respectively (Table III). The two ammo-
nia TPD peaks at 490 K and 660 K (Fig. 3) were recorded
for HMor10, which corresponds to the two distinct acid sites.
The weak acid sites in HMor10 provides for more efficient
copper incorporation and reduction starting at 150◦C, while
the strong acid sites assist the appearance of intermediate re-
duced Cu+species.
3.4.2. DRS of the reduced CuMor15
The spectra of the CuMor15 samples (Fig. 14) did not show
efficient reduction of copper. Indeed, noticeable bands (i)
and (vi) belonging to Cu2+ ions were observed for all Cu-
Mor15 samples except for CuMor15-150. The band (ii) at
250 nm due to Cu+species discussed above (Sec. 3.3) was
not detected clearly; nevertheless, it may have been hidden as
a weak shoulder interfering with the strong peak at 210 nm,
especially for CuMor15-250.
The sample reduced under mild conditions, CuMor15-
150, revealed a very broad descent down to the end of the
available scale (Fig. 14). Similar features in absorption of
solids are sometimes observed when a number of absorp-
tion mechanisms exist, and an absorption coefficient usu-
ally grows with photon energy. These absorption mecha-
nisms in our case may be attributed to the presence of dif-
ferent intermediates of reduced copper. However, increas-
ing temperature suppressed the formation of these species.
Their features are not apparent in the spectra of other Cu-
Mor (Figs. 13, 15, 16). Nevertheless, this type of absorption
may be extracted for CuMor10-150 and to a lesser extent for
CuMor30-150 samples. Since initial HMor15 zeolite has the
maximum number of acid sites with the higher acidity, high
acidity probably prevents the efficient reduction of Cu(II) to
produce copper nanoparticles.
It is of interest to include in the interpretation of this
broad absorption the possibility of formation of small copper
clusters. The mild reduction conditions can favor formation
of these less stable species as compared with that of larger
nanoparticles. For example, 8-atomic clusters Ag8are eas-
ily stabilized in the case of Ag-exchanged mordenites, and
the mild reduction conditions correspond to the appearance
of Ag8clusters rather than of the larger silver nanoparti-
cles [15]; however, in the case of copper in zeolites no un-
ambiguous indication has been made to this date [45]. Suc-
cessful stabilization of the clusters of definite nuclearity is
conditioned by their fitting into the mordenite channels with
cross-section of 0.65×0.70 nm [4,21]. In the case of copper
this good fitting requirement of Cu8isomers is violated since
the possible size of corresponding copper clusters should be
less than the size of silver clusters of the same nuclearity due
to the smaller size of copper atom. Probably, copper clusters
of the higher nuclearity could fit into the mordenite chan-
nel. In the CuMor15-150 spectrum there is a weak feature in
the range 300-350 nm, attribution of which should be other
than species (iii) enumerated above for CuMor samples. This
weak maximum (CuMor15-150, Fig. 14) may be considered
as the absorption of one of the possible Cuncopper clusters;
however, the data on direct assignment are still disputable.
We can only note that the bands at 280, 375, 450 nm [46] and
355, 375, 410 nm [47] were assigned to different low-stable
copper species like Cux+
n, formed in Cu2+-containing solu-
tions, and a number of maxima in the range 300 – 550 nm
were associated with Cunclusters stabilized in matrices of
rare gases [48,49].
Rev. Mex. Fis. 59 (2013) 170–185
182 V. PETRANOVSKII et al
3.4.3. DRS of the reduced CuMor30
The further increase of MR leads to a modification of the
optical appearance of the copper reduced forms (Fig. 15).
First of all, the plasma resonance peak (v) corresponding to
the copper nanoparticles is developed; however, in this case
it has a step-like shape rather than a well-developed maxi-
mum. This shoulder has been assigned to very small copper
particles (with radii less than 10 nm) [22,50]. In the case of
CuMor30 this feature appears at the reduction temperature
of 250◦C and regularly rises with increasing temperature. In
the case of CuMor30-450, the spectrum in the range 300-550
nm has an absorption plateau with remarkably complex struc-
ture that is difficult to interpret (see Fig. 15) starting from
the plasmon resonance shoulder. For the samples CuMor10
(Fig. 13) and CuMor206 (Fig. 16) weak maxima (iii) and (iv)
related to the copper interband transitions are observed in this
range in line with the plasmon resonance [51]. Meanwhile,
for this sample all spectra except for CuMor30-450 include
noticeable bands (i) and (vi) of Cu2+ ions. The feature (ii)
at ∼250 nm related to the Cu+state is hidden as a shoulder
on the slope of the intense 210 nm peak (i). Thus, this com-
position of the mordenite does not favor easy and complete
reduction of Cu(II) to Cu(0).
The region (iii) 300-350 nm shows some weak features
in the curves corresponding to all reduction temperatures
(Fig. 15). For the samples reduced under high tempera-
tures they look very similar to satellites to the plasmon res-
onance band due to interband transitions in copper. How-
ever, the CuMor30-150 sample that reveals no plasmon res-
onance shows a feature of particular interest in this region.
The inset in Fig. 15 with the subtracted spectrum undoubt-
edly evidences that, for this sample, the maximum occurs at
322 nm. A similar peak is observed in the CuMor15-150
sample (Fig. 14) but with higher intensity. This peak may
be associated with small copper clusters formed under mild
reduction conditions, as well as with the CuMor15-150 sam-
ple. These samples, HMor15 and HMor30, are similar in
the strength of active acid sites and copper ion exchange ca-
pacity. The lower concentration of the acid sites and their
lower pKafor HMor30 than for HMor15 explains the less
expressed spectral feature at 322 nm. This observation also
supports the view that cluster formation may be connected
with the acid sites of high strength.
3.4.4. DRS of the reduced CuMor206
Spectra of this sample can be divided into two distinct types:
(A) 150 and 250◦C and (B) 350 and 450◦C (Fig. 16). These
types are clearly distinguished throughout the whole spec-
tral range. In the spectra of the A-type absorption band of
Cu2+d−dtransitions the feature (vi) is developed at the
trace level. The minor feature (v) from the copper nanopar-
ticles only appears weakly in the CuMor206-150 spectra.
The weak shoulder (ii) at ∼250 nm is developed in the
CuMor206-150 spectrum but is absent from the CuMor206-
250 sample. The feature (i) at 210 nm can be distinguished as
a shoulder in the spectra, and absorption continues to increase
to the limit of the spectral range. It is possible to infer from
the absence of the typical features both for Cu(II) and for any
reduced species that practically all Cu(II) is reduced in the A-
type samples, but these reduced forms have absorbance that
does not appear in the accessible spectral range, probably at
λ < 190 nm.
Spectra of the B-type show the well-resolved plasmon
resonance band (v) accompanied by two bands (iii) and (iv).
The weak peak (ii) of Cu+at 250 nm is well resolved. Hence,
a significant part of Cu(II) in B-type of samples is reduced
to metal nanoparticles. An important difference between
B- and A-type spectra is the intense background absorbance
throughout the entire spectral interval under study. This back-
ground may be associated with the structureless absorbance
by large metal particles and aggregates of larger size [52].
3.5. General analysis of the samples with different MR
The preparation procedure and composition of mordenite de-
fines the nature and concentration of acid sites, secondary
porosity and total surface area of zeolite matrices. Closer
examination proved that with respect to variations in MR
of mordenites, non-monotonic dependence is observed for:
a) secondary porosity; b) total surface area; c) total acidity;
d) relative amount of different types of acid sites, i.e. their
pKaand surface concentrations.
Depending on these variations of mordenite properties
changes are observed for: 1) the amount of incorporated cop-
per ions via ion-exchange; 2) the amount of Cu(II) ions in
the set of hydrated samples measured by EPR; 3) initial tem-
perature of Cu(II) reduction with formation of nanoparticles;
4) the shape of plasmon resonance band of copper nanoparti-
cles; and 5) the appearance of other reduced copper species.
The reduction temperature strongly influences the species
of copper formed. All the above-mentioned factors provide
changes in the final products. The nature of active acid sites
can affect the positions of copper ions, their reactivity, and
the behavior of final reduction products.
The TPD measurements (Fig. 3) allow us to separate acid
sites into three different types, participating, presumably, in
the process of copper incorporation and reduction. Surface
concentration and pKaof the Brønsted acid sites revealed by
titration measurements (Table III) correlate with TPD of am-
monia and demonstrate complex dependence on MR regard-
ing their appearance and acidity. In general, the following
correlations may be found between the properties of acid sites
occurring at different MR in the mordenites and the spectral
appearance of the copper species formed as a result of hydro-
gen reduction:
In the samples CuMor15 and CuMor30 a large amount of
strong acid sites (730 K peak in the TPD data) is present.
A small fraction of these is occupied by copper ions af-
ter ion exchange (Table 2). This environment promotes re-
duction of Cu2+ ions either to small copper clusters under
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 183
low-temperature (CuMor15-150, CuMor30-150), or to cop-
per nanoparticles within the size range ∼1-2 nm under high-
temperature reduction (CuMor30-350, CuMor30-450). The
high acid strength of these sites prevents the reduction of cop-
per to metal, and the sample CuMor15 appears to be almost
unreduced even at 450◦C. Lower concentration of strong acid
sites in HMor206 leads only to prevention of Cu2+ reduction
at temperatures lower than 350◦C.
Intermediate strength acid sites (660 K peak in the TPD
data) bind copper ions in the CuMor10 and CuMor206 sam-
ples with less strength. The low temperatures are sufficient
only for copper reduction to Cu(I), rather than to Cu(0), while
with higher temperatures reduced copper species aggregate
rapidly to form nanoparticles of ∼5-10 nm. Plasmon reso-
nance of the particles of this size range appears as a well de-
veloped band. The fine details of its contour are determined
by particle location, environment, shape factor, etc. [53].
The weakest acid sites (490 K peak in the TPD data), pre-
sumably of a different nature, occur in considerable amounts
in the CuMor10 sample only. Their nature is possibly con-
nected with extra-framework aluminum atoms; this may ex-
plain their peculiar behavior. These sites bind copper ions
weakly and copper nanoparticles with a well-pronounced
plasmon resonance maximum are formed along the whole
temperature range.
These three types of acid sites exist in protonated mor-
denite and, according to the findings above, contribute in very
different manner to copper reducibility. The spectroscopic
data on the state of Cu2+ ions in the Cu-exchanged forms of
mordenites collected within the framework of this work allow
us to separate different copper adsorption sites only partially.
One sub-band at ∼620 nm may be extracted from the broad
optical absorption band related to d−dtransition in Cu2+. It
occurs in the case of CuMor10 possessing the featured acid
sites and this sub-band may be associated with these weakest
acid sites, which do not occur in Cu-exchanged mordenites
with other MR. The other two acid sites of higher strength
are associated with the framework T-positions of aluminum
and probably produce two main positions for copper ion ab-
sorption. Two copper sites were determined by Cheetham
et al. using direct X-ray analysis of a single crystal of mor-
denite with MR=28.8 (close to our MR=30) [19]. One site
coordinated with oxygen in the elliptical 8-rings, the other
(octahedral) within the main 12-ring channel.
Since the crystal structure of mordenites with
15≤MR≤206 is practically the same (Sec. 3.1), we pro-
pose that similar positions can be realized for other mor-
denites, not only for MR∼30 (except for MR=10 which
has, as was stated above, the possible additional position
for ion-exchanged ions). However, under our experimental
conditions mordenites are capable of exchanging only low
percentages of Cu (Table II). These low values mean that
only one of the two copper positions can be active, and only
one type of acid center develops the main activity in HMor30.
The second type makes a much lower but still evident con-
tribution, as seen from the change in the low-temperature
side of the TPD curve discussed in Sec. 3.1. Thus, it may
be proposed that the positions within the 8-member ring are
responsible for Cu absorption in CuMor15 and CuMor30.
HMor206, in contrast, makes its main contribution to the
TPD at intermediate temperatures. Thus, an increase of the
MR value from 30 to 206 changes the type of copper site
that is most active. In Mor206 with its lower Al content,
this site is more stable and provides higher ion-exchange abil-
ity. The site is probably located in the 12-member ring. The
less the strength of the active centers the more complete the
ion-exchange. 12-ring positions are weaker than 8-ring po-
sitions. This is to say, a strong acid keeps cations worse.
This fact is common in the chemistry of aqueous solutions.
Due to the higher degree of ion-exchange, both types of cen-
ters are probably occupied, at least partially. The variabil-
ity in Cu2+ bonding is manifested in the reduction process:
strong acidity and the weaker bond of copper ions with the
zeolite framework (8-rings) prevent effective reduction, but
lower acidity promotes the reduction (for copper ions sitting
within 12-rings).
The significant effect of hydrogen reduction temperature
on copper reducibility may also be explained from this view-
point. Dehydration of mordenites used in this work may be
extended to up to 200-300◦C depending on the MR [10,22]
and makes additional changes in the copper absorption cen-
ters. Instead of two copper positions within the 8- and 12-
rings, in dehydrated mordenite (up to 510◦C) three positions
were detected [19]: the position within the 8-ring equal, but
the 12-ring can now contain two positions. On the basis of
these crystallographic data for original and dehydrated mor-
denites with copper, it may be speculated that the reduction
temperature has little effect when 8-ring positions are ac-
tive, but the higher effect of temperature would be for cop-
per located within the 12-ring. Variable reactivity of copper
in other zeolites depending on its crystallographic positions
has also been noted in other studies [20,54-56]. Probably,
HMor206 incorporates copper into 12-rings, HMor30 and
HMor15 incorporate it into the 8-ring, and HMor10, in addi-
tion to the positions in the 12-ring, has the extra-framework
copper (see above, Sec. 3.1 an Ref. 57). The maximum re-
ducibility from experimental data considered in Secs. 3.4 can
be attained when copper ions are in the more open 12-ring,
and the minimum would be for copper in the 8-ring.
4. Conclusions
Copper incorporated by ion exchange into mordenites with
various SiO2/Al2O3values (MR), and therefore variable acid
properties, can be reduced by heat treatment in hydrogen with
the concomitant production of a number of species: Cu(I)
states, copper nanoparticles of different size and localization
in the mordenite, and small copper clusters. The spectral ap-
pearance of these species in the samples recorded by EPR
and DRS of the Cu-containing hydrated, as well as reduced
mordenites strongly depends on the value of MR and on the
Rev. Mex. Fis. 59 (2013) 170–185
184 V. PETRANOVSKII et al
reduction temperature. MR serves as the key factor that reg-
ulates both concentration and strength of the acid sites. The
following correlations were obtained:
(i) the mordenite sample with the lowest acid strength,
MR=10, in part due to the presence of extra-framework
aluminum, provided efficient copper reduction; and the
size of the nanoparticles produced grows with the re-
duction temperature;
(ii) the most acid sites (both in concentration and strength)
for MR=15 and 30 prevented copper reduction, but,
only in these cases, under mild conditions, the possi-
bility of small copper cluster formation was noted in
the optical spectra;
(iii) in the case of acid sites of medium strength in the
highly dealuminated mordenite with MR=206, copper
reduction proceeded effectively only at high tempera-
tures, and nanoparticles with the pronounced plasmon
resonance band were formed.
The clear difference in behavior of MR-variable morden-
ites with respect to copper reducibility was interpreted as be-
ing due to exchange of Cu2+ ions into different sites, depend-
ing upon the amount of aluminum present in the mordenite
(controlled by MR ratio). The more active (and, correspond-
ingly, the more reducible) copper enters the positions in the
8-rings, and the less active copper appears in the 12-rings.
Secondary porosity plays a significant role among the factors
regulating locations where copper nanoparticle formation oc-
curs, in media with varying dielectric functions.
The material of this work contributes to the understand-
ing of copper behavior in zeolites and catalysts by allowing
us to propose MR as an efficient tool in metal state regu-
lation, which is an important factor in catalytic activity and
selectivity. Various optical features of zeolites with copper
nanoparticles may be proposed for the design of new optical
materials with absorption tuned by the ionic properties of the
matrix.
Acknowledgments
We greatly benefited from many discussions and collabora-
tions with Dr. Yoshihiro Sugi, Dr. Karl Seff, Dr. Robert
Marzke and Dr. Sergey Romanov. We wish to thank Jose
Victor Tamariz Flores for his experimental work on measure-
ments of Cu content. The authors would also like to ex-
press their gratitude to Francisco Ruiz, Eloisa Aparicio, Eric
Flores and Juan Peralta for the invaluable technical support.
The research reported in this paper was supported by grant
IN114603 3 from UNAM-PAPIIT, grant 102907, CONA-
CYT, Mexico, and grant RFBR 090300347-a, Russia. The
authors F. Chavez Rivas and R. Zamorano Ulloa acknowl-
edge support from COFAA-IPN-Mexico.
1. B. C. Gates, Chem. Rev. 95 (1995) 511.
2. K. Seff, and T. Sun, Chem. Rev. 94 (1994) 857.
3. G. Stucky, and J. Mac Dougall, Science 247 (1990) 669.
4. V.S. Gurin, N.E. Bogdanchikova, and V.P. Petranovskii, J.
Phys. Chem. B 104 (2000) 12105.
5. P.P. Edwards, P.A. Anderson, and J.M. Thomas, Acc. Chem.
Res.29 (1996) 23.
6. Yu.A. Alekseev, V. Bogomolov, T. Zhukova, V. Petranovskii,
and S. Kholodkevich, Sov. Phys. Solid State 24 (1982) 1384.
7. T. Sun, K. Seff, N.H. Heo, and V. Petranovskii, Science 259
(1993) 495.
8. Y. Park, Y.S. Lee, and K.B. Yoon, J. Am. Chem. Soc. 115 (1993)
12220.
9. J. Ogden, N. Bogdanchikova, J. Corker, and V. Petranovskii,
Eur. J. Phys. D 9(1999) 605.
10. D. W. Breck, Zeolite Molecular Sieves. Structure, Chemistry
and Use (A Wiley-Interscience Publication, John Wiley &
Sons, New York, 1974).
11. Atlas of Zeolite Structure Types, 5th revised edition, (Ch. Baer-
locher, W. M. Meier and D. H. Olson, Eds. 2000). Available
on-line: http://www.iza-structure.org/databases/
12. H. Yahiro, and M. Iwamoto, Appl. Catal. A 222 (2001) 163.
13. C. Torre-Abreu, C. Henriques, F.R. Ribeiro, G. Delahay, and
M.F. Ribeiro, Catal. Today 54 (1999) 407.
14. V.I. Parvulescu, P. Grange, and B. Delmon, Catal. Today 46
(1998) 233.
15. N. Bogdanchikova, V. Petranovskii, R. Machorro, Y. Sugi, V.M.
Soto, and S. Fuentes, Appl. Surf. Sci. 150 (1999) 58.
16. N. Bogdanchikova, V. Petranovskii, S. Fuentes, E. Paukshtis,
Y. Sugi, and A. Licea-Claverie, Mater. Sci. Eng. A 276 (2000)
236.
17. S.Y. Chung, S.H. Oh, M.H. Kim, I.S. Nam, and Y.G. Kim,
Catal. Today 54 (1999) 521.
18. B.R. Goodman, K.C. Hass, W.F. Schneider, and J.B. Adams,
Catal. Lett.68 (2000) 85.
19. M.P. Attfield, S.J. Weigel, and A.K. Cheetham, J. Catal. 170
(1997) 227.
20. P.A. Jacobs, W. de Wilde, R.A. Shoonheidt, and J.B. Uytter-
hoven, J. Chem. Soc. Faraday Trans. I 72 (1976) 1221.
21. V. Gurin, N. Bogdanchikova, and V. Petranovskii, Mater. Sci.
Eng. C 18 (2001) 37.
22. V. Petranovskii, V. Gurin, N. Bogdanchikova, A. Licea-
Claverie, Y. Sugi, and E. Stoyanov, Mater. Sci. Eng. A 332
(2002) 174.
23. N. Katada, H. Igi, J.-H. Kim, and M. Niwa, J. Phys. Chem. B
101 (1997) 5969.
Rev. Mex. Fis. 59 (2013) 170–185
FORMATION OF COPPER NANOPARTICLES IN MORDENITES WITH VARIABLE SiO2/Al2O3. . . 185
24. H. Igi, N. Katada, M. Niwa, in: M.M.J. Treacy, B.K. Marcus,
M.E. Bisher, J.B. Higgins, Materials Research Society 4(1999)
2643.
25. K. S. W. Sing et al., Pure Appl. Chem. 603 (1985) 57.
26. V. Petranovskii et al., Stud. Surf. Sci. Catal.142 (2002) 815.
27. J. W. Akitt, Progress in NMR Spectr. 21 (1989) 1.
28. N. Katada, Y. Kageyama, and M. Niwa, J. Phys. Chem. B 104
(2000) 7561.
29. S. C. Larsen, A., Aylor, A. T., Bell, and J. A. Reimer, J. Phys.
Chem.98 (1994) 11533.
30. P. J. Carl, and S. C. Larsen, J. Catal. 182 (1999) 208.
31. P. J. Carl, and S. C. Larsen, J. Phys. Chem. B.104 (2000) 6568.
32. W. Froncisz, and S. H. James, J. Chem. Phys. 73 (1980) 3123.
33. C. Oliva, et al., J. Chem. Soc., Faraday Trans. 93 (1997) 2603.
34. F. Chavez Rivas, V. Petranovskii and R. Zamorano Ulloa, Rev.
Mex. F´
ıs. 56 (2010) 328.
35. E. Casta˜
neda Miranda, Tesis Licenciatura “Caracterizaci´
on de
la zeolita Cu-ZSM5 por EPR y DRX para razones molares de
20 y 30”. Febrero del 2005. ESFM-IPN.
36. A.B.P. Lever, Inorganic Electronic Spectroscopy, 2nd Edition,
Elsevier, Amsterdam, 1984.
37. F.S. Hadzhieva, V.F. Anufrienko, T.M. Yurieva, V.N. Vorobiev,
T.P. Minyukova, React. Kinet. Catal. Lett.30 (1986) 85.
38. G.T. Palomino, P. Fisicaro, S. Borgida, A. Zecchina, E. Gi-
amello, and C. Lamberti, J. Phys. Chem. B 104 (2000) 4064.
39. R.M. Friedman, J.J. Freeman, and F.W. Lytle, J. Catal.55
(1978) 10.
40. B.J. Hathaway, Coord. Chem. Rev. 52 (1983) 87.
41. C.J. Ballhausen, Introduction to ligand field theory. (McGraw-
Hill Book Company Inc., New York, 1963).
42. R. Debnath and K.S. Das, Chem. Phys. Lett. 155 (1989) 52.
43. P.D. Yang et al., Science 282 (1998) 2244.
44. H.T. Wang, Z.B. Wang, L.M. Huang, A. Mitra, B. Holmberg,
and Y.S. Yan, J. Mater. Chem.11 (2001) 2307.
45. Y. Kuroda, S. Konno, Y. Yoshikawa, H. Maeda, Y. Kubozono,
H. Hamano, R. Kumashiro, and M. Nagao, J. Chem. Soc.-
Faraday Trans. 93 (1997) 2125.
46. B.G. Ershov, E. Janata, and A. Henglein, Radiat. Phys. Chem.
39 (1992) 123.
47. J. Khatouri, M. Mostafavi, J. Amblard, and J. Belloni, Chem.
Phys. Lett. 191 (1992) 351.
48. G.A. Ozin, H. Huber, D. McIntosh, S. Mitchell, J.G. Norman
Jr., and L. Noodleman, J. Am. Chem. Soc.101 (1979) 3504.
49. G.A. Ozin, S. Mitchell, D.F. McIntosh, S.M. Mattar, and J.
Garcia-Prieto, J. Phys. Chem. 87 (1983) 4651.
50. I. Lisiecki, and M.P. Pileni, J. Phys. Chem. 99 (1995) 5077.
51. J.F. Janak, A.R. Williams, V.L. Moruzzi, Phys. Rev. B 11 (1975)
1522.
52. Y.I. Petrov. Clusters and Small Particles, Nauka, Moscow
(1986).
53. U. Kreibig, M. Vollmer, Optical Properties of Metal Clusters
(Springer-Verlag, Berlin, 1995).
54. M.P. Attfield, S.J. Weigel, and A.K. Cheetham J. Catal. 172
(1997) 274.
55. B. Wichterlova, J. Dedecek, Z. Sobalyk, A. Vondrova, and K.
Kliery, J. Catal. 169 (1997) 194.
56. Y. Kuroda, Y. Yoshikawa, S. Emura, R. Kumashiro, and M. Na-
gao, J. Phys. Chem. B 103 (1999) 2155.
57. M. Niwa, K. Suzuki, N. Katada, T. Kanougi, and T. Atoguchi,
J. Phys. Chem. B 109 (2005) 18749.
Rev. Mex. Fis. 59 (2013) 170–185
