Você está na página 1de 9

Applied Catalysis A: General 456 (2013) 96104

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General


journal homepage: www.elsevier.com/locate/apcata

Inuence of the synthesis route on the catalytic oxidation of 1,2-dichloroethane over CeO2 /H-ZSM5 catalysts
B. de Rivas a , C. Sampedro a , E.V. Ramos-Fernndez b , R. Lpez-Fonseca a , J. Gascon b , M. Makkee b , J.I. Gutirrez-Ortiz a,
a Chemical Technologies for Environmental Sustainability Group, Department of Chemical Engineering, Faculty of Science and Technology, University of the Basque Country UPV/EHU, PO Box 644, E-48080 Bilbao, Spain b Catalysis Engineering, Department of Chemical Engineering, Faculty of Applied Sciences, Delft University of Technology, Julianalaan 136, NL-2628 BL Delft, The Netherlands

a r t i c l e

i n f o

a b s t r a c t
The performance of supported CeO2 /HZSM-5 catalysts with a nominal CeO2 loading of 10 wt.% was evaluated for the oxidation of one of the most common chlorinated pollutants found in waste streams, namely 1,2-dichloroethane. The inuence of the preparation method, such as impregnation in different media (water and ethanol), precipitation and ion exchange, was examined. Structural, morphological and physico-chemical changes caused as a function of the synthesis procedure were analysed by atomic emission spectroscopy, X-ray diffraction, BET measurements, transmission electronic microscopy, Xray photoelectron spectroscopy, NH3 -temperature-programmed desorption, adsorption of CO at low temperature followed by infrared spectroscopy, temperature-programmed reduction with hydrogen, energy dispersive X-ray spectroscopy and dynamic thermogravimetry coupled to mass spectrometry. The enhancement of the catalytic behaviour of the resulting ceria loaded samples with respect to plain H-ZSM5 could be explained on the basis of the synergetic effects of oxygen mobility and acid sites. In particular, the procedure based on impregnation with ethanol led to a highly dispersed ceria catalyst with a larger amount of oxygen vacancies. As a result, this catalyst required a temperature lower than 200 C for attaining 50% conversion. 2013 Elsevier B.V. All rights reserved.

Article history: Received 11 January 2013 Received in revised form 20 February 2013 Accepted 26 February 2013 Available online 6 March 2013 Keywords: Cl-VOCs Catalytic oxidation 1,2-dichloroethane Supported cerium oxide H-ZSM-5 zeolite Synthesis procedure

1. Introduction Chlorinated VOCs have been produced commercially and used for many purposes by the chemical industry including the manufacture of herbicides, plastics, and solvents. Uses outside the chemical industry include solvent degreasing in the automotive and aerospace industries, dry cleaning solvents in the garment industries and precision solvent cleaning in the electronic industries. Due to the presence of chlorine atoms in the VOC molecule, a number of negative effects exist, namely a higher thermal stability and the unavoidable conversion of chlorine atoms into hydrogen chloride or molecular chlorine. Moreover, in some cases by-products that are more toxic and hazardous than the original chlorinated VOC may be formed. Therefore, development of lowtemperature processes for chlorinated waste disposal is of great importance [1]. Supported noble-metal catalysts as well as (supported) metal oxide catalysts are commercially applied for the oxidation process.

Corresponding author. Tel.: +34 94 6012683; fax: +34 94 6015963. E-mail address: joseignacio.gutierrez@ehu.es (J.I. Gutirrez-Ortiz). 0926-860X/$ see front matter 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apcata.2013.02.026

Alumina is most frequently employed as a support [2]. Noble metals are powerful oxidation catalysts, but can easily form inorganic chlorides, with deactivation as a likely consequence. Furthermore, noble-metal chloride species could also cause chlorination of organic compounds besides its oxidation [3]. Additional disadvantages of this type of catalysts are associated with their high cost and limited availability. Metal oxide catalysts such as chromium, cobalt, vanadium and copper oxides, are reported to be suitable for this process as well, although the lower activity and volatilisation of metaloxychlorides are serious drawbacks [3]. Thus substantial efforts are currently being made to develop alternative catalysts with a comparable activity. Recently, ceria-based bulk oxides have shown promising results in the oxidation of chlorinated VOCs and other pollutants. It is due to their remarkable redox properties (based on its high oxygenstorage capacity and facile cycle of Ce4+ /Ce3+ ), thermal stability and resistance to Cl-poisoning [49]. Hence, it seems reasonable to expect that the overall catalytic behaviour of the resulting Ce catalyst could be noticeably improved by suitably supporting this active phase on a carrier. It must be taken into consideration that the support itself may play a relevant role in enhancing the conversion of the chlorinated hydrocarbons [10]. This promoting effect has

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104

97

been typically associated with the acid properties. In this paper the use of protic zeolites as appropriate potential supports is proposed. There is a wide range of zeolite structures available with variable Al/Si ratios, enabling zeolites with specic properties to be selected as catalyst support. The selection of H-ZSM5 zeolite for this work is justied on the basis of previous reports where a number of zeolites including H-Y, H-ZSM5, H-MOR and H-BETA were examined in the oxidative decomposition of chlorinated hydrocarbons [11,12]. The better catalytic behaviour was related to its relatively high internal surface area, strong acidity and three-dimensional channel system. Moreover, when compared with conventional (alumina and silica) supports Gutierrez-Ortiz et al. [13] found an enhanced behaviour for Mn2 O3 supported on H-ZSM-5 catalysts in the oxidation of dichloromethane and trichloroethylene. Accordingly, Scire et al. [14] also observed a better catalytic behaviour for supported Pt catalysts when H-ZSM5 zeolite was used as carrier with respect to H-BETA and alumina in the combustion of chlorobenzene. Rachapudi et al. [15] reported the good performance of Cr/H-ZSM5 catalysts in the combustion of vinyl chloride and trichloroethylene although a partial loss of chromium was noticed. In this work the catalytic performance of H-ZSM5 supported ceria catalysts for the gas-phase oxidation of chlorinated VOCs has been examined. On the basis of previous reports, an optimum cerium loading of 10 wt.% was found for CeO2 /H-ZSM5 catalysts [16]. Particularly, this study evaluates the different synthesis routes for CeO2 /H-ZSM5 and how it affects the catalytic nal behaviour in combustion of 1,2-dichloroethane, which was chosen as a model chlorinated VOC. 1,2-Dichloroethane (C2 H4 Cl2 , DCE), also known as ethylene dichloride, is probably one of the most important chlorinated VOC emitted in industrial ue gases, since it is used as an intermediate for the production of polyvinyl chloride. Additional uses are as a solvent in textile cleaning and metal degreasing and paint remover, a starting material for paint, varnish, and nish removers, a cleaner for upholstery and carpets, a fumigant, a lead scavenger in antiknock gasoline, and as a dispersant for plastics and elastomers such as synthetic rubber.

catalyst pellets with a 0.3 to 0.5 mm diameter were prepared by a process of compressing the oxide powders into akes in a hydraulic press (Specac), crushing and sieving. All the samples obtained prior to catalytic activity and selectivity experiments were characterised using several analytical techniques. 2.2. Characterisation techniques The metal content in the catalysts was measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) in ARL Fisons 3410 ICP equipment. Prior to analysis, the samples were dissolved in an acid solution (HCl+HNO3 ) with small amounts of HF. XPS measurements were performed with a Phoibos 150 1DDLD analysis system from Specs. The samples were excited by the non-monochromatised Al K source (1486.6 eV). Survey spectra were measured at 40 eV pass energy. For the individual peak energy regions a pass energy of 20 eV was used, with an electron angle of 90 . The sample powders were deposited onto a KBr wafer. The constant charging of the samples was corrected by referencing all the energies to the C 1s peak energy (set at 284.6 eV), arising from adventitious carbon. The binding energies of the catalysts were corrected with respect to the Ag 3d 5/2 peak energy set at 368.28 eV, taken as internal reference peak. The peaks were tted by CasaXPS 2.3.16 software, using a properly weighted sum of Lorentzian and Gaussian component curves after background subtraction according to Shirley [17]. Transmission electron microscopy (TEM) investigations were performed using a Philips CM200 microscope equipped with LaB6 crystal as electron source and operating at 200 kV. Bright eld images were acquired using a high resolution CCD camera. Drops of emulsions, created by sonication of the powder samples in ethanol, were deposited on C coated Cu grids and left in air top dry. X-ray diffraction (XRD) studies were carried out on a XPERT-MPD X-ray diffractometer with Cu K and Ni lter. The X-ray tube was operated radiation (=1.5406 A) at 30 kV and 20 mA. Samples were scanned from 5 <2<80 and the X-ray diffraction line positions were determined with a step size of 0.02 and counting time of 2.5 s per step. Textural properties were evaluated from the nitrogen adsorption-desorption isotherms, determined at -196 C with a Micromeritics ASAP 2010 apparatus. The samples (20 mg) were outgassed at 350 C for 16 h. The adsorption data recorded from the amount of N2 physisorbed at different relative pressures were treated with the full BET equation. Temperature-programmed desorption (TPD) of ammonia was performed on a Micromeritics Autochem 2920 instrument equipped with a quartz U-tube coupled to a thermal conductivity detector. Prior to adsorption experiments, the samples (40 mg) were pre-treated in a 5%O2 /He stream at 550 C. Then, they were cooled down at 100 C in a He ow (20 cm3 min1 ). Later, the NH3 adsorption step was performed by admitting a ow of 10%NH3 /He at 100 C up to saturation. Subsequently, the samples were exposed to a ow of helium (50 cm3 min1 ) for 1 h at 100 C in order to remove weakly physisorbed ammonia from the surface. Finally, desorption was carried out from 100 to 550 C at a heating rate of 10 C min1 in a He stream (50 cm3 min1 ). This temperature was maintained for 120 min until the adsorbate was completely desorbed. The amount of gases desorbed was determined by time integration of the TCD curves. The acid properties of the samples were also characterised by CO adsorption at low temperature followed by infrared spectroscopy. To this purpose 2.5 mg of sample were pressed into self-supported wafers. The catalyst was placed in an IR quartz cell equipped with CaF2 windows. A movable sample holder allowed the sample to be placed in the infrared beam for the measurements or into a furnace for thermal treatments. The cell was connected to a vacuum line for the pretreatment. The specimen was activated under vacuum at 500 C overnight to remove adsorbed species. After

2. Experimental 2.1. Catalysts preparation The parent ZSM-5 zeolite (Si/Al = 27.3 as determined by XRF analysis [17]) was supplied by Zeolyst Corp. The supported samples with a 10 wt.% nominal Ce content (12 wt.% nominal CeO2 content) were prepared using several synthesis methodologies including impregnation with excess of solvent, precipitation and ion exchange. When the impregnation (10 g of zeolite support) was carried out two different solvents (100 cm3 ) were employed, namely water and ethanol. The selected metallic salt was cerium (III) nitrate hexahydrate (Aldrich). The molar concentration was 0.07 M. The solvents were removed under reduced pressure by using a rotary evaporator. The resulting catalysts were denoted as CeZ/IM-W and CeZ/IM-E, respectively. The sample prepared by precipitation was contacted with an 0.007 M aqueous solution (1 L) of ammonium cerium (IV) nitrate (Aldrich) and stirred at room temperature for 3 h. An aqueous solution of ammonia, added dropwise to the suspension until pH 8.5. The sample was labelled as CeZ/PP. Finally, an ion-exchanged zeolite catalyst (CeZ/IE) was obtained. The solution was prepared by dissolving cerium (III) nitrate hexahydrate in deionised water (0.007 M, 1 L). The ion exchange (9 g of zeolite) was carried out at 80 C under vigorous stirring for 24 h. After ltration, the obtained solid was washed to remove nonexchanged ions. All the prepared samples were dried in an oven at 110 C overnight followed by calcination at 550 C for 4 h in static air. Next,

98

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104

this step the samples were cooled down to -140 C and CO was dosed up to 39 mbar. Transmission spectra were recorded in the 22502000 cm1 range at 4 cm1 resolution on a Nicolet Nexus spectrometer equipped with an extended KBr beam splitting and a mercury cadmium telluride (MCT) cryodetector. Redox behaviour was examined by temperature-programmed reduction experiments (TPR). These experiments were conducted on a Micromeritics Autochem 2920 instrument as well. Firstly, all the samples were pre-treated in an oxygen stream (5%O2 /He) at 550 C for 1 h, and then cooled down to room temperature. The reducing gas used in all experiments was 5%H2 in Ar, with a ow rate of 50 cm3 min1 . The temperature range explored was from room temperature to 950 C with a heating rate of 10 C min1 . This temperature was maintained for 0.5 h so as to complete the reduction process. The water produced by reduction was trapped into a cold trap. The consumption of H2 was quantitatively was quantitatively measured by time integration of TPR proles. The amount of coke present in the used samples was determined by means of dynamic thermogravimetry using Setaram Setsys Evolution apparatus under atmospheric pressure coupled to a quadrupole mass spectrometer (Pfeiffer Prisma). The mass loss and the sample temperature were continuously recorded by a computerised data acquisition system. After an isothermal step at 150 C the combustion of coke was carried out from 150 to 900 C at a constant heating rate of 5 C min1 . The oxidant stream was 5%O2 /He (50 cm3 min1 ) owing downwards onto the cylindrical sample holder. The quantication of chlorine content was evaluated by energy dispersive X-ray (EDX) spectroscopy using a JEOL JSM-6400 scanning electron microscope coupled with analysis software INCA Energy 350 from Oxford Instruments. 2.3. Catalyst activity determination Catalytic tests were performed in a bench-scale xed bed reactor (Microactivity modular laboratory system provided by PID Eng&Tech S.L.) operated at atmospheric pressure and fully monitored by computer. The reactor was made of quartz with an internal diameter of 10 mm and a height of 300 mm, in which the temperature is controlled with a thermocouple place in the catalyst bed. Typically 0.85 g of catalyst in powdered form (0.30.5 mm) was loaded. The reaction feed consisted of 1000 ppm of DCE in dry air with a total gas ow of 500 cm3 min1 (15,000 h1 ). Catalytic activity was measured over the range 150550 C and conversion data were calculated by the difference between inlet and outlet concentrations. Conversion measurements and product proles were taken at steady state, typically after 30 minutes on stream. Product selectivity was calculated based on either chlorine or carbon atoms present in that product divided by the total chlorine or carbon atoms present in the product stream (expressed as %), according to the following equations (Cx Hy Clz corresponds to the chlorinated by-products that may result from the incomplete combustion of the feed molecule). SHCl = SCl2 = SCO2 = [HCl] 100 [HCl] + 2[Cl2 ] + z [Cx Hy Clz ] 2[Cl2 ] 100 [HCl] + 2[Cl2 ] + z[Cx Hy Clz ] [CO2 ] 100 [CO] + [CO2 ] + x[Cx Hy Clz ] (1)

Fig. 1. XRD patterns of the CeZ samples.

Further details on analytical procedures are described elsewhere [16]. 3. Results and discussion 3.1. Characterisation of the catalysts Thermogravimetric analysis was preliminary conducted in order to determine the temperature for the oxidative decomposition of the precursor salt. As a result, the activation procedure could be dened to fully oxidise the cerium precursor. Several weight losses were recorded from 110 to 300 C (not shown here), though no further change in weight was observed at higher temperatures. Thus, calcination at a temperature of at least 300 C was required to convert cerium nitrate into cerium oxide in all the samples. ICPAES analysis (Table 1) revealed that the ceria content of the solid samples obtained by impregnation (CeZ/IM-W and CeZ/IM-E) and precipitation (CeZ/PP) was around 1113 wt.%. These values were relatively close to the ceria nominal content. However, the cerium content of CeZ/IE was found to be noticeably lower, around 3 wt.%. In fact, for the given Si/Al ratio, this loading almost coincided with the maximum exchange capacity. Table 1 also summarises the textural properties of the zeolite catalysts in terms of specic surface area and pore volume. A slight decrease (1015%) in the specic surface area with respect to the bare zeolite was observed. According to these values and assuming that the ceria crystallites are non-porous, it could be concluded that the metallic particles were preferentially deposited on the external surface [18]. The preparation route apparently had no relevant inuence on the textural properties. Fig. 1 shows the XRD patterns of the prepared CeO2 -zeolite catalysts. The patterns of the reference bulk CeO2 (JCPDS 898436) and H-ZSM5 (JCPDS 440003) samples were also included for the sake of comparison. The presence of ceria on CeZ/IM-E, CeZ/IM-W and CeZ/PP samples was assessed, more notably on CeZ/IM-W. Expectedly, the XRD pattern of the ion exchanged sample was virtually identical to that of the support. The mean diameter of the crystallite of the CeO2 could only be estimated according to the Scherrer equation for CeZ/IM-W (about 8 nm, slightly lower than 9 nm estimated for bulk ceria). In case of the CeZ/IM-E and CeZ/PP samples, the much better dispersion of

(2)

(3)

The feed and efuent streams were analysed using an on-line 7980A Agilent Technologies gas chromatograph equipped with a thermal conductivity (CO and CO2 ) and an electron capture detector (chlorinated hydrocarbons). Analysis of HCl and Cl2 was carried out by means of ion selective electrode and titration, respectively.

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104

99

Fig. 2. TEM images of the CeZ samples.

the active phase (resulting in smaller crystallites) did not allow the calculation of a mean CeO2 crystallite diameter [19]. In order to gain further insight into the size and distribution of ceria crystallites the samples were characterised by TEM. The images included in Fig. 2 revealed that the cerium oxide added to the support was preferentially located at the external surface of the zeolite irrespective of the catalyst synthesis route. The ne particles were relatively spherical in shape and each particle was found to be an aggregate of nanocrystallites [20]. The crystallite size varied between 7.6 nm for CeZ/IM-W and 3.54 nm for CeZ/IME and CeZ/PP (Table 1). These values were comparatively smaller that the size of bulk ceria (9 nm as estimated by XRD). On the other hand, it must be pointed out that ceria particles were not observed on the sample synthesised by ion exchange (CeZ/IE). As for the samples prepared by impregnation a remarkable observation was that the crystallite size signicantly varied depending on

the solvent employed. Thus, the use of ethanol led to the formation of smaller crystallites upon calcination compared to samples prepared using water. In an effort to understand the effect of the solvent on the observed differences, the eluted gas phase species during the calcination were analysed (Fig. 3). The thermal protocol included two isothermal steps at 30 and 110 C. The temperature was ramped at 10 C min1 . The mass signals corresponding to H2 O (m/z = 17 and m/z = 18), CO (m/z = 12 and m/z = 28), CO2 (m/z = 44), NO (m/z = 30), NO2 (m/z = 46) and ethanol (m/z = 31 and m/z = 45) were monitored. It was noticed that while the decomposition of the cerium precursor to nitrogen oxides occurred at about 250 C for the CeZ/IM-W sample, it required signicantly higher temperatures (325 C) for the CeZ/IM-E sample. It is thought that the higher afnity of ethanol for the zeolite prevented the mobility of the Ce particles during calcination. As a result, the resulting crystallite size was smaller. Alternatively, as proposed by Ho and Su [21], this

Table 1 Physico-chemical properties the CeZ samples. Catalyst H-ZSM5 CeO2 CeZ/IM-W CeZ/IM-E CeZ/PP CeZ/IE
a b c

CeO2 (wt.%)a 10.8 12.9 13.4 3.1

CeO2 (wt.%)b 13.1 13.2 8.6 0.1

Surface area (m2 g1 ) 464 99 422 406 396 459

Pore volume (cm3 g1 ) 0.243 0.210 0.237 0.231 0.269 0.239

CeO2 crystallite size (nm)c 9.0* 7.6 3.5 4.0

Ce4+ /Ce3+ ratio 2.7 6.1 2.4

Determined by ICP-AES. Determined by XPS. Determined by TEM (*determined by XRD).

100

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104

Ce4+
m/z=18 m/z=17

Ce3+ v v

600
CeZ/IM-W

MS Signal, a.u.

Intensity, a.u.

m/z=28 m/z=44

500

u u

u0 v v v0

m/z=12 m/z=46 m/z=30 m/z=31 m/z=45

400

CeZ/PP

Temperature (C)

300
m/z=18 m/z=17

CeZ/IM-E

MS Signal, a.u.

200
m/z=28 m/z=44

930

920

910

900

890

880

Binding Energy, eV
100
Fig. 4. Ce 3d peaks XPS spectra of the CeZ samples.

m/z=12 m/z=46 m/z=30 m/z=31 m/z=45

20

40

60

80

100

120

140

Time (min)
Fig. 3. Thermal decomposition followed by mass spectrometry of the zeolite catalysts prepared by impregnation with water (CeZ/IM-W) and ethanol (CeZ/IM-E).

relatively reduced size could be also explained by the presence of ethoxy groups on the surface which hindered the agglomeration of CeO2 by physically interfering. The Ce 3d XPS spectra of the different supported catalysts were examined. With this technique we could determine the surface cerium content of the catalysts as well as the oxidation state Ce(IV) or Ce(III). In general, a good correlation was found between the compositions given by either ICP-AES or XPS for the samples prepared by impregnation. In contrast, major discrepancies were found for the CeZ/PP and CeZ/IE samples. In the case of CeZ/PP this was related to a more probable presence of cerium inside the pores which cannot be detected by XPS. As for the CeZ/IE sample only traces of ceria were found on the catalyst surface, thereby evidencing the incorporation of this element into the zeolite lattice. Fig. 4 displays the XPS spectra of the Ce 3d core level peaks of CeO2 in the region of 930870 eV and its deconvolution. The deconvoluted peaks consisted of two series of distinguishable peaks of the Ce4+ and Ce3+ ionic states (3d5/2 and 3d3/2 ). The peaks labelled as v and v were assigned to a mixture of Ce 3d9 4f2 Ln2 and Ce 3d9 4f1 Ln1 Ce(IV) nal states, and the peak v was attributed to Ce 3d9 4f0 Ln Ce(IV) nal state. Moreover, the peaks v0 and v were associated with Ce 3d9 4f2 Ln-1 and Ce 3d9 4f1 Ln states, respectively, the states of Ce(III). Similarly, the u, u, and u bands were the states of Ce4+ while u0 and u were the states of Ce3+ [2224]. The abundance of Ce3+ could be evaluated, after tting the Ce 3d peaks, from the ratio between the area corresponding to peaks assignable to Ce3+ and the total area in the Ce 3d region. The obtained values are listed in Table 1 and revealed that both cerium states coexisted in the supported zeolite catalysts after calcination in air at 550 C. Noticeable differences in Ce4+ /Ce3+ distribution were noted as a function of the synthesis procedure (Table 1). Hence, smaller amounts of Ce3+ (at around 14%) were found for the impregnated sample in ethanol media while substantially higher amounts were noticed for CeZ/IMW (at about 27%) and CeZ/PP (at about 29%) samples. It is believed

that trivalent Ce3+ can be either in regions of sesquioxide Ce2 O3 or surrounded oxygen vacancies of CeO2 [24,25]. Temperature-programmed desorption of ammonia was used in order to characterise not only the overall acidity but also the acid strength distribution of the catalysts (Fig. 5). The area under the curve gave an estimation of the amount of acid sites in each sample (Table 2). In general, the TPD curves displayed by all the zeolite supported ceria catalysts presented two distinct desorption peaks. The rst peak that appeared at approximately 200 C, was indicative of the presence of weak sites, whereas the second one, located at around 380 C, could be assigned to the presence of strong acid sites. In order to quantify the number of acid sites of each acid type, the prole was deconvoluted into two Gaussian-like peaks. Results shown in Table 3 indicated that the supported samples exhibited a slightly higher overall acidity (413%) than the bare support, probably due to the contribution of the intrinsic acidity of the deposited metallic phase. Note that pure ceria was characterised by a substantial overall acidity. The strength of the acid sites after ceria incorporation was not apparently affected since no signicant change in desorption peak temperatures was observed [26]. Moreover, the relative acid strength distribution revealed a roughly identical presence of both weak and strong sites. It found that the CeZ/IM-E catalyst presented the largest overall acidity (0.56 mmol

H-ZSM5

500

TCD signal, a.u. g-1

Temperature, C

CeZ/IE

400

CeZ/PP

300
CeZ/IM-E

200
CeZ/IM-W CeO2

100

20

40

60

80

time, min
Fig. 5. NH3 -TPD proles of the CeZ samples.

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104 Table 2 Acid properties of the CeZ samples. Catalyst H-ZSM5 CeO2 CeZ/IM-W CeZ/IM-E CeZ/PP CeZ/IE Total acidity (mol NH3 g1 ) 490 112 541 556 475 508 Weak acidity (mol NH3 g1 ) 265 (54) 39 (35) 281 (52) 278 (50) 266 (56) 274 (54)

101

Strong acidity (mol NH3 g1 ) 225 (46) 73 (65) 260 (48) 278 (50) 209 (44) 234 (46)

Values in brackets indicate the relative abundance (%) of each type of acid sites.

Table 3 Redox properties of the CeZ samples. Catalyst H2 consumption (mol H2 gcat 1 ) 350 C H-ZSM5 CeO2 CeZ/IM-W CeZ/IM-E CeZ/PP CeZ/IE 40.0 27.0 36.3 17.6 7.3

H2 consumption (mol H2 gCe 1 ) 350 C 158.0 249.8 329.6 160.0 146.2

Degree of reduction (%Ce3+ ) 350 C 1.4 8.6 11.3 5.5 4.1

%OV 350 C 0.3 2.2 2.8 1.4 1.0 550 C 3.0 16.0 17.4 15.6 6.1

550 C 344.1 167.0 223.1 175.1 43.2

550 C 1357.3 1516.9 2028.3 1591.8 863.7

550 C 11.8 64.2 69.8 54.8 24.2

NH3 g1 ) and the highest amount of strong acid sites (0.28 mmol NH3 g1 ). The nature (Brnsted or Lewis) of the acid sites present on the examined samples was characterised via FTIR by assessing the interaction between a probe molecule (CO) and the catalyst surface at low temperature. FTIR spectra in the C-O stretching region (22502100 cm1 ) of the H-ZSM5 supported catalysts after CO adsorption are displayed in Fig. 6 at -120 C and 0.8 mbar. As for bulk CeO2 two bands located at 2171 cm1 (non-specic interactions of CO with the surface) and 2155 cm1 (CO linearly adsorbed on surface-exposed metal ions, e.g. Lewis sites) were clearly observed [27]. Their intensities increased with the CO pressure, and

63 21 55 21

38 21

71 21
22 30 22 10

CeZ/IM-E CeZ/IM-W CeZ/PP CeZ/IE HZSM-5 CeO2

2250

2200

2150

2100

Wavenumber,

cm-1

Fig. 6. FTIR spectra of CO adsorption over the CeZ samples.

disappeared quickly after evacuation at room temperature. Previous works on pyridine adsorption followed by FTIR also revealed the Lewis character of the cerium oxide [28]. In contrast, the spectrum of the bare zeolite was characterised by the presence of two intense absorption bands centred at 2138 cm1 , which were assigned to the weakly adsorbed CO (band strongly dependent on the CO pressure). The band centred at 2163 cm1 was associated with CO adsorbed on the protons of the bridging Si(OH)Al [2931]. Two additional weak bands at 2138 and 2117 cm1 were also noticed. These vibrations corresponded to O-bonded CO [32]. Similar bands were found in the various supported ceria catalysts. In addition, for the samples containing the higher amount of ceria, two bands at 2171 and 2210 cm1 are clearly visible as shoulders at the high and low frequency tail of the proton-adsorbed CO peak, in good agreement with the presence of Lewis acidity from ceria crystallites at the zeolite surface. This revealed the presence of Lewis acidity attributable to the ceria crystallites deposited on the zeolite surface. This result is line with the ndings reported by Xiaoning et al. [33] and Bi et al. [34] who observed an increase in the ratio Lewis/Brnsted acid sites after addition of ceria to a HZSM-5 zeolite. Also, two new bands at 2230 and 2215 cm1 were noticed on the CeZ/IM-E sample (not discernible for the parent zeolite and only barely visible for the other ceria catalysts upon magnication). These bands were assigned to strong Lewis acid sites [35], being a prove of real ion exchange. Cerium oxide is characterised by its oxygen storage capacity as a result of the formation of redox (Ce4+ /Ce3+ ) couples. The redox properties of the catalysts were thus examined by temperatureprogrammed reduction (TPR) using hydrogen. Fig. 7 shows the TPR proles of different CeO2 /H-ZSM5 catalysts along with that of pure ceria. Firstly, it should be highlighted that the pure CeO2 sample displayed a broad reduction prole with a low-temperature H2 peak centred at around 480 C, which was attributed to the reduction of the uppermost layers of Ce4+ , and a high-temperature H2 peak centred at 870 C approximately, which was associated with the bulk ceria reduction [4]. Note that no appreciable hydrogen consumption was observed for the parent H-ZSM-5 zeolite (not shown in Fig. 7). The shape of the TPR prole for any of the ceria supported catalysts was remarkably different since the reduction process took place in the whole temperature range. Note that irrespective of the CeZ catalyst the complete reduction of ceria occurred at 950 C. These curves were deconvoluted into three reduction bands (except

Absorbance, a.u.

21 17

102

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104


1000
CeO2 CeZ/IE

100

900 800

TCD signal, a.u. g-1

80
700

Temperature, C

CeZ/PP

600 500 400

Conversion, %

60

CeZ/IM-E

300 200

40

CeZ/IM-W

100

20

40

60

80

100

120

20

time, min
Fig. 7. H2 -TPR proles of the CeZ samples.

CeZ/IM-E CeZ/IM-W CeZ/PP CeZ/IE CeO2 H-ZSM5 Homog.


200 300 400 500

for CeZ/IM-W whose prole was deconvoluted into four peaks). These bands were tentatively assigned to the facile surface reduction of ceria (350 C), an intermediate reduction of the surface and the bulk (460510 C), and the complete bulk reduction of the remaining crystallites (650950 C) [16,36,37]. Interestingly, the reduction of the surface ceria on the supported samples was considerably shifted to lower temperatures (from 480 to 350 C), and occurred to a larger extent on the samples prepared by impregnation in contrast with the CeZ/PP sample whose extensive reduction required higher temperatures. Two additional observations from the H2 -TPR proles deserve an explanation. On one hand, it was observed that the H2 consumption for the reduction of CeZ/PP sample was lower. This can be attributed to the lower metal oxide content and to the fact that, as shown by XPS analysis, a notable fraction of the cerium atoms on this catalyst was present as Ce3+ . On the other hand, it is believed that the notable hydrogen uptake at temperatures higher than 750 C for the impregnated samples may be correlated to the formation of cerium aluminate (or cerium silicate) when ceria was well dispersed onto the support [19,38]. In contrast to the other samples, a single peak (located at 380 C) was observed in the case of the ion exchanged sample. The amount of H2 consumed, which is a direct measurement of the amount of oxygen evolved from the sample, was used to calculate the amount of Ce3+ formed during reduction (nCe3+ ). Table 3 includes the values of the reduction degree (%Ce3+ ) referred to the cationic sublattice according to the following equation % Ce3+ = nCe3+ 100 nCeT (4)

Temperature, C
Fig. 8. Light-off curves of DCE oxidation over the CeZ samples (reaction conditions: 1000 ppm DCE in dry air; 500 mL min1 ; W = 0.85 g; 635 g h molDCE 1 ).

3.2. Catalytic activity, product distribution and stability Typically catalytic activity for chlorinated VOC oxidation was characterised by monitoring the rise in conversion as a function of temperature under given test conditions. A characteristic curve, referred as light-off or ignition curve was obtained. T50 (temperature at which 50% conversion was reached) was used as an indicative of the relative reactivity of the catalysts. Fig. 8 shows the light-off curves recorded at constant WHSV (635 g h molDCE 1 ) and feed concentration (1000 ppm in air) over the various Ce/Z catalysts. Also the curves corresponding to the bare support, HZSM5, and the pure CeO2 are included for the sake of comparison. These samples were previously subjected to a calcination procedure similar to that of the supported catalysts. The possibility of homogeneous gas phase reactions at the reaction temperature range studied was checked through an experiment placing crushed quartz into the reactor tube. It was found that all the catalysed reactions needed noticeable lower reaction temperatures than the homogenous reaction irrespective of the catalyst examined. The bare zeolite exhibited a remarkable better conversion with respect to the pure cerium oxide (with a shift of about 100 C). This high activity suggested that acidity indeed plays a major role in the activation of the chlorinated molecule due to its potential to cleavage the H-Cl bond [40]. Independently of the synthesis route employed the supported ceria signicantly increased the activity for DCE oxidation. These results demonstrate that the reaction is accelerated by the presence of both acid and redox sites: the chlorinated compound is efciently adsorbed on the acid site and the oxidation reaction occurs at the redox site in the close vicinity [28,41]. As can be observed in Fig. 9, the catalytic activity follows the trend CeZ/IM-E > CeZ/IM-W > CeZ/PP > CeZ/IE, demonstrating that samples prepared via impregnation are the most active. The availability of labile oxygen, expressed as the %OV, can be used to rationalise the catalytic results. Fig. 9 indicates that there is a clear relationship between %OV and the activity of the ceria catalysts expressed in terms of T50 value. The sample prepared by impregnation with ethanol (CeZ/IM-W) exhibits a T50 value as low as 195 C (Table 4). This sample was characterised by the highest value of the %OV parameter. On the other hand, total acidity can be considered to be a catalytic key property of secondary importance since no signicant differences were found among the various

where nCeT is the amount of cerium ions in one mole of the oxide (nCeT is equal to 1 for CeO2 ). On the other hand, since the stoichiometry of the reduction process establishes that the formation of two Ce3+ ions corresponds to the simultaneous formation of one oxygen vacancy [5,39], the percentage of oxygen vacancies (%OV) referred to the total amount of oxygen atoms initially present in the oxide can be obtained as follows %OV = %Ce3+ 4 (5)

Table 3 lists the %OV values after the reduction process up to two reduction temperatures, namely 350 and 550 C. It was found that this parameter signicantly varied as a function of the preparation method. In this way, the highest value corresponded to the sample prepared by impregnation with ethanol (CeZ/IM-E). This behaviour was in agreement with the smaller ceria crystallite size and the lower amount of Ce3+ present in this catalyst after calcination as revealed by XPS.

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104


3.0

103

2.5

CeZ/IM-E

2.0

CeZ/IM-W
%OV
1.5

1.0

CeZ/PP

CeZ/IE

0.5

0.0
0 19 0 20 0 21 0 22 0 23 0 24 0 25

T50, C
Fig. 9. Relationship between the %OV parameter and the T50 value of the CeZ samples (reaction conditions: 1000 ppm DCE in dry air; 500 mL min1 ; W = 0.85 g; 635 g h molDCE 1 ). Table 4 T50 values ( C) of DCE oxidation over the CeZ samples. Catalyst H-ZSM5 CeO2 CeZ/IM-W CeZ/IM-E CeZ/PP CeZ/IE T50 ( C) 255 320 215 195 225 245 T90 ( C) 295 395 260 245 265 285

Reaction conditions: 1000 ppm DCE in dry air; 500 mL min1 ; W = 0.85 g; 635 g h molDCE 1 .

procedures used for preparing the ceria zeolite catalysts. Note that after ceria deposition the overall acidity of the resulting catalysts slightly increased with the generation of new Lewis acid sites. Results on the effect of DCE concentration on catalytic conversion are not available. Typically it is assumed that light-off curves are identical for DCE concentration in the 2502000 ppm range (pseudo-rst order kinetics in DCE, zero order for oxygen due to its large concentration). High conversion is not the only criterion for selection of good chlorinated VOC destruction catalysts. It is indeed useful to provide some basic information about the nature and the amount of reaction products, as the combustion of chlorinated VOCs may be accompanied by the concomitant production of chlorinated byproducts, sometimes more toxic and recalcitrant than the starting compound. DCE catalytic oxidation gave rise to CO2 , CO, HCl and Cl2 as major products. Nevertheless, its decomposition was accompanied by the generation of methyl chloride and vinyl chloride at mild temperatures (between 175350 C). According to Eqs. (6)(9) vinyl chloride is formed by dehydrochlorination of the feed molecule, and this by-product can be further partially oxidised to methyl chloride. It is important to point out that the formation of vinyl chloride may cause deactivation by coking, as it can readily polymerise to form carbonaceous residues [42]. The peaks concentrations were 140220 ppm (200 C) for methyl chloride and 100120 ppm (250 C) for vinyl chloride. C2 H4 Cl2 C2 H3 Cl + HCl C2 H3 Cl + 5/2O2 2CO2 + HCl + H2 O C2 H3 Cl + O2 CH3 Cl + CO2 CH3 Cl + 3/2O2 CO2 + HCl + H2 O (6) (7) (8) (9)

Note that the deep decomposition of these by-products generated over H-ZSM5 required temperatures higher than 500 C. Other authors found acetic acid and acetaldehyde in the product stream over CeO2 /Y catalysts [41], which was consistent with the proposed mechanism for DCE oxidation [28]. However, these species were not detected in our experiments probably due to their condensation at the acid zeolite sites, likely resulting in the formation of carbon deposits. At 300325 C, where DCE decomposition was almost complete and the yield of chlorinated by-products was appreciably low (<25 ppm), the chlorine atoms present in the feed were preferentially converted into hydrogen chloride leading to HCl selectivity values of around 97% over CeZ/IM-E catalyst. On the other hand, all supported ceria catalysts deeply abated DCE to give rise COx as major carbon-containing oxidation products. The relative CO/CO2 ratio was near-unity. When increasing the reaction temperature the formation of hydrogen chloride was inhibited in favour of molecular chlorine. This conversion was a consequence of the occurrence of the so-called Deacon reaction (2HCl+1/2O2 H2 O+Cl2 ) which was promoted at high temperatures due to the presence of cerium. Hence, the selectivity values towards HCl decreased to around 60% at higher temperatures (550 C). At higher temperatures the oxidation of CO to CO2 was favoured as well. As a result, signicantly higher CO2 selectivity values could be attained (70% for CeZ/IM-E). Note that CO2 selectivity at 550 C over the plain H-ZSM5 was only 30%. This relatively low selectivity to CO2 could be considered as a major drawback for CeO2 /H-ZSM5 systems. At this point, it should be pointed out that the addition of water has been proven as an effective strategy to increase, at least partially, the selectivity to CO2 owing to the water gas shift reaction. Also the selectivity to HCl would be promoted by reversing the Deacon reaction. However, the presence of steam (or large amounts of HCl, other VOCs and other inorganic spectators, namely NOx , NH3 and CO) may involve a negative impact on catalytic activity due to competitive effects. Another alternative approach to promote CO2 selectivity could the incorporation of a promoter with a high activity in the oxidation of CO to CO2 (Co3 O4 ). The stability of the most active sample, namely CeZ/IM-E, was investigated by analysing the evolution of conversion with time on stream at 225 C for 72 h. This temperature was selected as it provoked a conversion of less than 100%, thus providing a more sensitive indication of changes on the catalyst performance with time on line. A moderate deactivation with a decrease in conversion from 70 to 55% was observed as shown in Fig. 10. The decline in conversion was evident during the initial 5 h interval while conversion remained reasonably stable at longer time on stream. This behaviour is in agreement with the results reported by Huang et al. [43] for the oxidation of DCE over CuO/CeO2 /USY zeolite catalysts. Note that the generation of chlorinated by-products as well the selectivity to CO/CO2 and HCl/Cl2 remained unaffected with extended time on stream. The used catalyst was characterised by BET measurements, dynamic thermogravimetry, XRD and EDX. The decrease in activity was attributed to the deposition of coke, about 1.6 wt.% as quantied by thermogravimetry. It is thought that the loss in activity caused by coking at longer time intervals was minimal because the amount of deposited coke did not increase with time. TGAMS analysis also revealed that these carbonaceous deposits could be completely burnt off at temperatures lower than 550 C to yield CO2 as deep oxidation product. This observation is important for the eventual regeneration step since it did not require excessively high temperatures which could induce a structural collapse of the zeolitic support and/or sintering of CeO2 crystallites. On the other hand, the BET surface area of the spent catalyst was seen to decrease by 10%. XRD analysis suggested that the framework was preserved, thereby evidencing a relatively good structural stability, even after

104

B. de Rivas et al. / Applied Catalysis A: General 456 (2013) 96104

100

a Marie Curie Intra European Fellowship within the 7th European Community Framework Program. References
[1] R.S. Francis, in: G. Ertl (Ed.), Handbook of Heterogeneous Catalysis, vol. 5, WileyVCH Verlag, Weinheim, 2008, pp. 23852394. [2] S. Pitkaho, L. Matejova, K. Jiratova, S. Ojala, R.L. Keiski, Appl. Catal. B 126 (2011) 215224. [3] J.J. Spivey, in: G. Ertl (Ed.), Handbook of Heterogeneous Catalysis, vol. 5, WileyVCH Verlag, Weinheim, 2008, pp. 23942411. [4] J.I. Gutirrez-Ortiz, B. de Rivas, R. Lpez-Fonseca, J.R. Gonzlez-Velasco, Appl. Catal. A 269 (2004) 147155. [5] B. de Rivas, R. Lpez-Fonseca, C. Sampedro, J.I. Gutirrez-Ortiz, Appl. Catal. B 90 (2009) 545555. [6] B. de Rivas, R. Lpez-Fonseca, M.A. Gutirrez-Ortiz, J.I. Gutirrez-Ortiz, Appl. Catal. B 101 (2011) 317325. [7] Q. Dai, X. Wang, G. Lu, Appl. Catal. B 81 (2008) 192202. [8] E.V. Ramos-Fernandez, J.C. Serrano-Ruiz, J. Silvestre-Albero, A. SepulvedaEscribano, F. Rodriguez-Reinoso, J. Mater. Sci. 43 (2008) 15251531. [9] A. Goncalves, J. Silvestre-Albero, E.V. Ramos-Fernandez, J.C. Serrano-Ruiz, J.J.M. Orfao, A. Sepulveda-Escribano, M.F.R. Pereira, Appl. Catal. B 113114 (2012) 308317. [10] I. Maupin, L. Pinard, J. Mijoin, P. Magnoux, J. Catal 291 (2012) 104109. [11] R. Lpez-Fonseca, J.I. Gutirrez-Ortiz, J.R. Gonzlez-Velasco, Appl. Catal. A 271 (2004) 3946. [12] R. Lpez-Fonseca, J.I. Gutirrez-Ortiz, M.A. Gutirrez-Ortiz, J.R. GonzlezVelasco, Catal. Today 107108 (2005) 200207. [13] J.I. Gutirrez-Ortiz, R. Lpez-Fonseca, U. Aurrekoetxea, J.R. Gonzlez-Velasco, J. Catal 218 (2003) 148154. [14] S. Scir, S. Minic, C. Crisafulli, Appl. Catal. B 45 (2003) 117125. [15] R. Rachapudi, P.S. Chintawar, H.L. Greene, J. Catal 185 (1999) 5872. [16] B. de Rivas, C. Sampedro, R. Lpez-Fonseca, M.A. Gutirrez-Ortiz, J.I. GutirrezOrtiz, Appl. Catal. A 417418 (2012) 93101. [17] D.A. Shirley, Phys. Rev. B 5 (1972) 47094714. [18] C. Ramos Moreira, M. Maciel Pereira, X. Alcob, N. Homs, J. Llorca, J.L.G. Fierro, P. Ramrez de la Piscina, Microporous Mesoporous Mater. 100 (2007) 276286. [19] J. Strunk, W.C. Vining, A.T. Bell, J. Phys. Chem. C 115 (2011) 41144126. [20] T. Dhannia, S. Jayalekshmi, M.C. Santhosh Kumar, T. Prasada Rao, A. Chandra Bose, J. Phys. Chem. Solids 70 (2009) 14431447. [21] S.-W. Ho, J. Catal 168 (1997) 5159. [22] B.M. Reddy, L. Katta, G. Thrimurthulu, Catal. Today 175 (2011) 585592. [23] J. Matharu, G. Cabailh, R. Lindsay, C.L. Pang, D.C. Grinter, T. Skala, G. Thornton, Surf. Sci. 605 (2011) 10621066. [24] N. Sutradhar, A. Sinhamahapatra, S. Pahari, M. Jayachandran, B. Subramanian, H.C. Bajaj, A.B. Panda, J. Phys. Chem. C 115 (2011) 76287637. [25] N.K.N. K. Renuka, J. Alloys Compd. 513 (2012) 230235. [26] Y. Sugi, Y. Kubota, K. Komura, N. Sugiyama, M. Hayashi, J.-H. Kim, Stud. Surf. Sci. Catal. 158 (2005) 12791286. [27] M. Daturi, C. Binet, J.-C. Lavalley, A. Galtaryies, R. Sporken, Phys. Chem. Chem. Phys. 1 (1999) 57175724. [28] B. de Rivas, R. Lpez-Fonseca, J.R. Gonzlez-Velasco, J.I. Gutirrez-Ortiz, J. Mol. Catal. A 278 (2007) 181188. [29] J. Szanyi, M.T. Paffett, Microporous Mater. 7 (1996) 201218. [30] L. Chen, L. Lin, Z. Xu, T. Zhang, Q. Xin, P. Ying, G. Li, C. Li, J. Catal 161 (1996) 107114. [31] R.A. Shigeishi, B.H. Chiche, F. Fajula, Microporous Mesoporous Mater. 43 (2001) 211226. [32] T. Armaroli, L.J. Simon, M. Digne, T. Montanari, M. Bevilacqua, V. Valtchev, J. Patarin, G. Busca, Appl. Catal. A 306 (2006) 7684. [33] W. Xiaoning, Z. Zhen, X. Chunming, D. Aijun, Z. Li, J. Guiyuan, J. Rare Earths 25 (2007) 321328. [34] J. Bi, M. Liu, C. Song, X. Wang, X. Guo, Appl. Catal. B 107 (2011) 6876. [35] M.S. Holm, S. Svelle, F. Joensen, P. Beato, C.H. Christensen, S. Bordiga, M. Bjorgen, Appl. Catal. A 356 (2009) 2330. [36] A. Piras, S. Colussi, A. Trovarelli, V. Sergo, J. Llorca, R. Psaro, L. Sordelli, J. Phys. Chem. B 109 (2005) 1111011118. [37] A.S. Prakash, C. Shivakumara, M.S. Hegde, Mater. Sci. Eng. B 139 (2007) 5561. [38] L.F. Liotta, A. Longo, G. Pantaleo, G. Di Carlo, A. Martorana, S. Cimino, G. Russo, G. Deganello, Appl. Catal. B 90 (2009) 470477. [39] A. Trovarelli, Catalysis by Ceria and Related Materials, Catalytic Science Series, vol. 2, Imperial College Press, London, 2002. [40] R. Lpez-Fonseca, B. de Rivas, J.I. Gutirrez-Ortiz, A. Aranzabal, J.R. GonzlezVelasco, Appl. Catal. B 41 (2003) 3142. [41] J. Zhou, L. Zhao, Q. Huang, R. Zhou, X. Li, Catal. Lett. 127 (2009) 277284. [42] J.I. Gutirrez-Ortiz, B. de Rivas, R. Lpez-Fonseca, J.R. Gonzlez-Velasco, Appl. Catal. B 65 (2006) 191200. [43] Q. Huang, X. Xue, R. Zhou, J. Mol. Catal. A 344 (2011) 7482.

80

Conversion, %

60

40

20

0 0 20 40 60 80

time, h
Fig. 10. DCE oxidation over CeZ/IM-E at 225 C as a function of time on stream (reaction conditions: 1000 ppm DCE in dry air; 500 mL min1 ; W = 0.85 g; 635 g h molDCE 1 ).

the exposure of the catalyst to low concentrations of HCl. EDX analysis indicated the presence of relatively reduced amount of Cl on the samples, about 0.4 wt.%. 4. Conclusions The inuence of the synthesis method for preparing H-ZSM5 zeolite supported ceria catalysts on the oxidation of a model chlorinated VOC was examined. All the synthesised samples exhibited a relatively higher activity than the bare support. Among the explored synthetic routes, samples prepared by impregnation using ethanol as solvent resulted the most active. From catalyst structureactivity relationships it could be established that catalytic activity was primarily governed by a more effective oxygen mobility of the deposited ceria crystallites, which in turn was favoured by a relatively reduced crystallite size. Overall acidity and acid strength distribution appeared to be less important. In this sense, it should be highlighted that the deposition of ceria provoked the generation of new strong Lewis sites, in addition to the well known zeolite Brnsted acid sites. The formation of chlorinated by-products, mainly vinyl chloride and methyl chloride, was relatively low and these were abated at sufciently low temperatures (<350 C). As a consequence, the main chlorinated deep oxidation product was hydrogen chloride instead of molecular chlorine since at this low temperature the Deacon reaction is less important. Conversion of carbon atoms to CO was preferential, this suggesting that an optimum formulation of these ceria-based catalysts should include a second catalytic step that favours the oxidation towards CO2 . Acknowledgements The authors wish to thank the UPV/EHU and Gobierno Vasco (SAIOTEK S-PE09UN23) for nancial support. Technical and human support from SGIker (ICP-AES (J.C. Raposo), XRD (A. Larranaga), TEM (G. Lpez), EDX (A. Sarmiento) and XPS (M.B. Snchez)) is gratefully acknowledged. E.V. Ramos-Fernandez was supported by

Você também pode gostar