selective oxidation

0 downloads 0 Views 2MB Size Report
under isothermal conditions at 100 C are: 0.20 В 10А3 sА1 for the Co3O4 and 0.08 В 10А3 sА1 for NiO and CuO. Therefore, the activity of the Co3O4, is 3 times ...
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

Available online at www.sciencedirect.com

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

Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation Jairo A. Go´mez-Cuaspud, Martin Schmal* Federal University of Rio de Janeiro, Chemical Engineering Program, NUCAT/PEQ/COPPE e Centro de Tecnologia, Bl. G 128, C.P. 68502, CEP. 21941-914 Rio de Janeiro, Brazil

article info

abstract

Article history:

We investigated the synthesis of nanosized Co3O4, NiO and CuO oxides by the

Received 7 December 2012

polymerization-combustion method and evaluated in the selective oxidation of CO. These

Received in revised form

materials were characterized before and after catalytic test under specific conditions by

22 March 2013

X-ray diffraction, scanning electron microscopy, temperature programmed reduction,

Accepted 4 April 2013

Raman spectroscopy and Diffuse reflectance infrared Fourier transform spectroscopy. For

Available online xxx

isoconversion the activity follows the order: Co3O4>NiO > CuO. The intrinsic activities under isothermal conditions at 100  C are: 0.20  103 s1 for the Co3O4 and 0.08  103 s1

Keywords:

for NiO and CuO. Therefore, the activity of the Co3O4, is 3 times higher than for NiO and

Nano-oxides

CuO. The XRD and Raman analyses after catalytic test confirm the existence of a partial

SELOX

reduction of oxides to the metallic phase and no evidence of carbon deposition over surface

DRIFTS

materials. In the case of nickel oxide the analyzed sample shows the metallic phase of

Fuel cells

nickel and in the case of copper oxide the sample presents a high grade of reduction. XRD results confirm the presence of Cu2O and metallic Cu. Finally, results showed that these oxide catalysts present a high resistance to carbonaceous formation under present reaction conditions and confirm the effectiveness of polymerization-combustion technique for synthesis of active high catalyst in the selective oxidation of CO. Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.

1.

Introduction

Fuel cells are devices that convert chemical energy directly into electricity from hydrogen and oxygen feedings. Among various advantages this kind of devices operates at low temperatures, present high power density and easy start-up [1]. However, the principal requirement for the ideal performance is promoting the on-board reforming, removing the residual CO to a hydrogen-rich stream, because CO can damage the electrochemical performance of anodic components severely, even at

ppm levels [2]. It is generally recommended CO concentrations less than 10 ppm [2,3]. For the CO removal the catalyst is important and depends on the syntheses of these materials and improvements of specific catalytic properties. The most conventional method is based on a solidestate reaction, where metal oxides are mechanically mixed, resulting in finely divided powders. However, these materials present low surface areas, need high calcination temperatures and long reaction times [4]. On the other hand, the co-precipitation method can easily produce materials with high surface area, around

* Corresponding author. Tel.: þ55 2125628352; fax: þ55 2125628360. E-mail addresses: [email protected] (J.A. Go´mez-Cuaspud), [email protected] (M. Schmal). 0360-3199/$ e see front matter Copyright ª 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijhydene.2013.04.024

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

2

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

100 m2 g1, but an enormous effort is necessary to ensure a homogeneous material with uniform particle sizes and composition [1]. Sol-gel routes have been used to synthesize a lot of interesting materials with large surface areas, but this method requires expensive metal alkoxide precursors [2,5]. The combustion method has been proposed to synthesize nanosized materials and is particularly useful in the production of ultrafine ceramic powders with small average particle size. This is an easy and fast method, with the advantage of using inexpensive precursors, producing homogeneous nanosized crystallites and highly reactive materials [6]. The combustion synthesis technique needs the contact of a saturated aqueous solution of a desired metal salt and a suitable organic fuel boiling until the mixture ignites and a selfsustaining and fast combustion reaction takes off, resulting in dry crystalline fine oxide powders [7e9]. The large amount of gases released during the reaction produces a flame that can reach temperatures above 1000  C, which help in the consolidation of desired crystalline phases. Kingsley et al. [10] described the preparation of oxides with various metal nitrates (Mg, Ca, Sr, Ba, Mn, Co, Ni, Cu and Zn) using urea or carbohydrazide as fuel. The surface area depends on both the metal nitrates and fuel. The surface areas of the prepared materials using carbohydrazide are higher, varying from 45 to 85 m2 g1, compared with the materials prepared with urea (1e20 m2 g1). Mimani [11], evaluated different fuels (hexamethylenetetramine, urea, carbohydrazide and glycine) for production of MAl2O4 (M ¼ Mn, Cu and Zn). In the case of manganese aluminate, urea was the best fuel, producing porous materials with high surface area. The copper and zinc aluminates prepared with carbohydrazide and glycine presented the best results for mesoporous materials. The fuel/oxidizer ratio was evaluated by Alinejad et al. [12] in the combustion synthesis of MgAl2O4 by using sucrose and PVA solution as fuels. The increase of fuel/oxidizer ratio caused agglomeration of materials with average crystallite sizes in the range of 12.7e17.5 nm [13]. Chandradass et al. [13] synthesized alumina zirconia nanopowder oxides with high surface area by means of citrate auto-combustion method, evaluating the effect of nitrate/citrate (N/C) ratio and the use of chelant agent. The use of a stoichiometric U/N ratio resulted in a material with high surface area and homogeneous nanocrystallites, while the excess of citrate resulted in a non-porous material with low surface areas. The formation of superficial metal nanoparticles resulted in an excellent stability for low temperature CO oxidation [14]. Thus, the aim of this paper is to investigate the synthesis of nanosized Co3O4, NiO and CuO oxides by the polymerizationcombustion method and evaluated in the selective oxidation of CO, which permit to identify some of the most important characteristics in the performance of nanostructured fuel cell components.

2.

Experimental

2.1.

Preparation of catalysts

The nano-crystallite Co3O4, NiO and CuO oxides were prepared by the polymerization-combustion technique starting

from the corresponding nitrates Co(NO3)2.6H2O 99.9%, Ni(NO3)2.6H2O 99.9%, and Cu(NO3)2.9H2O 99.99%. Similarly, solid citric acid monohydrate 99.99% was used. Stoichiometric quantities of each solid salt (0.01 mol) were added to 20 mL of deionized water in different glass vessels equipped with magnetic stirring (150 rpm), reflux system and controlled temperature. Once reached the total dissolution of each nitrate precursor, we added solid citric acid solution at 0.5:1 M ratio with respect to the total concentration of metal cations in solution. These mixtures were kept under reflux at 120  C for 12 h, obtaining viscous liquids which were heated at 150  C under air flux in an oven until complete solvent evaporation. The solid foams were treated in an alumina crucible at 300  C for 30 min under air flow, using a ramp of 10  C min1 to start the auto-combustion stage. The obtained solids were maintained in a furnace at 350  C for 6 h under oxygen flow (60 mL min1), to eliminate carbonaceous residues and then grounded and sieved. The stoichiometric composition of the redox mixture was calculated based on the total oxidizing and reducing valences [15]. The stoichiometric ratio of nitrate to citrate (N/C), assuming complete combustion is described in reaction 1. Finally, the samples were calcined in flowing air at 400  C for 3 h to clean the surface from carbon residues. ðNO3 Þ þ C6 H18 O7 /6CO2 þ N2 þ 9H2 O þ oxides

2.2.

(1)

Characterization of catalysts

The chemical composition was obtained using an X-ray fluorescence (XRF) apparatus, Rigaku Model RIX 3100. Samples were isostatically pressed in pellets and analyzed quantitatively to verify the purity of obtained samples. The specific area BET was evaluated by nitrogen adsorption isotherms at 196  C, using the ASAP-2020 apparatus (Micromeritics). All samples were degassed at 300  C overnight to remove humidity. The crystalline structure was determined by X-ray diffraction, in a Miniflex Rigaku diffractometer, using Cu Ka radiation (l ¼ 1.54186  A) between 10 and 90 with steps of 0.05 and a speed analysis of 0.15 min1. The refinement, indexing and the simulation of the structures were done with Cellref3.0 and Rietveld software that allowed us to establish the chemical composition and crystallographic structure of the oxides. The crystallite size estimation was done using the highest diffraction signals, using the DebyeeScherrer equation, taking the value of half peak width set by a Lorentzian function and using a constant of 0.89 as reference. Temperature programmed reduction (TPR-H2), was performed in a Micromeritics Pulse Chemisorb model 2705 equipment. The sample was heated at 200  C for 2 h, flowing pure helium and then reduced with a mixture of 5% H2/He (30 mL min1), rising the temperature up to 700  C at 10  C min1 and the consumption of H2 was measured in a thermal conductivity detector. Analysis by scanning electron microscopy (SEM) was carried out in a LEO 440 microscope (Leica-Zeiss), equipped with an electron gun. The images were obtained with a focus distance of 10e25 mm, an accelerating voltage of 20 kV and a current of 100e200 pA, measurement time of 100 s and

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

3

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

counting rate 1.2 kcps. The samples were placed on a sticker attached to a graphite holder and aluminum objects shaded with gold to obtain a better contrast in images. Raman spectroscopy measures were carried out in a HR-UV 800 infinity microprobe (Jobin-Yvon) spectrometer equipped with CCD detector (70  C) and a laser power of 10.7 mW. The Raman spectra of the catalysts were collected between 800 and 1800 cm1 before and after catalytic tests projecting a continuous wave laser of HeeNe supplying the excitation line at 632 nm through the samples exposed to air at room temperature. Diffuse reflectance infrared Fourier transform spectroscopy of CO adsorption (DRIFTS-CO), was carried out on a Nicolet spectrometer (Nexus 470 model), with a MCT detector and equipped with a diffuse reflectance chamber (Spectra Tech) for high temperature treatment. The in situ treatment was carried out using a feed composition of H2 þ CO þ O2 (60:1:1 mL min1), which was analyzed before and after evacuation.

2.3.

Catalytic test

The CO selective oxidation was measured in a continuous fixed bed U-shaped reactor under atmospheric pressure. The samples (200 mg in each case) were dried at 250  C for 30 min prior catalytic tests under He flow (30 mL min1). After cleaning and cooling, the gas valve was switched to the gas mixture and adjusted until stabilization. The feed composition was 1% CO, 1% O2 60% H2 (vol%) in He balance at 120 mL min1, GSHV ¼ 36,000 h1 and a contact time of 0.1 s. Preliminary tests with constant W/F presented similar conversion, discarding mass transfer effects. The reaction was studied at different temperatures from 50 to 300  C. Exit gases were analyzed by gas chromatography (micro GC Varian CP 3800) equipped with a column Poraplot Q and molecular sieve and a TCD and FID detectors, permitting to evaluate the conversion of CO, O2 (XCO, Xo2 ) and the selectivity of reaction (SCo2 ) as shown in following equations: XCO ð%Þ ¼

½COin  ½COout  100 ½COin

(2)

XO2 ð%Þ ¼

½O2 in  ½O2 out  100 ½O2 in

(3)

SCO2 ð%Þ ¼ 0:5

 ½COin  ½COout   100 ½O2 in  ½O2 out

3.

Results and discussion

3.1.

Composition and surface areas

(4)

The chemical composition derived from X-ray fluorescence for the determination of major and trace elements confirm an excellent correlation between the proposed and obtained compositions as shown in Table 1. These results are closely related to the effectiveness of the synthesis method to obtain highly pure nanostructured materials, suggesting that self-

Table 1 e Elemental composition of metal oxides derived from X-ray fluorescence analysis. Catalyst

Nominal compositions

Real composition

Co3.0O4.0 Ni1.0O1.0 Cu1.0O1.0

Co2.98O3.99 Ni0.97O0.99 Cu0.97O0.96

Co3O4 NiO CuO

combustion process was completed successfully without the presence of carbonaceous deposits. The specific surface areas (SBET, m2 g1) of pure oxides solids are presented in Table 2, and show that the surface area of CuO is the lowest and presents low porosity compared to NiO and Co3O4. The maximum surface area was obtained for Co3O4 and could be attributed to the lowest auto-combustion temperature reached during the synthesis process, since, all samples were pretreated at 300  C. The cobalt sample started the auto-ignition first, followed by the nickel and copper oxides, respectively, releasing gases due to the thermal dissociation of nitrate precursors and thus, decreasing the surface area of NiO and CuO solids. It was probably caused during the thermal treatment, which might induce the growth of the oxide particles.

3.2.

X-ray diffraction

The XRD patterns of Co3O4, NiO and CuO oxides after polymerization-combustion process are shown in Fig. 1aec. Results indicate crystalline structure of the corresponding calcined oxides with the corresponding reflection lines indicating pure structures of Co3O4 (JCPDS: 42e1467), CuO (JCPDS: 45e0937) and NiO (JCPDS: 89e5881) oxides in accordance with the literature [16]. Results were analyzed using Rietveld refinement, as presented in Fig. 1 and Table 3, showing structures of the oxides; mean crystallite sizes, volume and lattice parameters. Crystallite sizes were estimated by using the highest diffraction signals and DebyeeScherrer equation, taking the value of half peak width (b), adjusted to a Lorentzian function and using a constant of 0.89 as reference, resulting in nanometric crystallite sizes. From the ICCD databases we calculated for each oxide, space groups, crystal systems and cell parameters as shown in Table 3. The crystallite sizes are quite different; dc ¼ 16.4 nm; 25.2 nm and 23.9 nm for Co3O4, NiO and CuO, respectively. The nanocrystallites obtained by means polymerization-combustion route promoted the stabilization of the active phases in contrast with the same catalysts prepared by the traditional

Table 2 e Textural properties: surface area and pore volume of the metal oxides. Catalyst Co3O4 NiO CuO

Surface area (m2/g)

Pore volume (cm3/g)

Pore radius ( A)

54.1 27.9 11.7

0.67 0.23 0.01

230.0 188.4 163.5

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

4

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

Fig. 1 e Rietveld refinement of XRD patterns for a. Co3O4, b. NiO and c. CuO oxides, before catalytic test.

thermal treatment of nitrate precursor salts [7,16,17]. Notwithstanding is the different lattice parameter a ¼ 0.8056 nm; a ¼ 0.4177 nm and a ¼ 0.4690 nm for the Co3O4, NiO and CuO, respectively, which are in agreement with the reported values in the literature [18e21].

3.3.

Temperature programmed reduction (TPR)

The TPR-H2 profiles of Co3O4, NiO and CuO oxides are shown in Fig. 2. The maximum reduction peak of Co3O4 occurred at 390  C with a shoulder around 310  C. According to Khodakov [22] the reduction of unsupported Co3O4 could be ascribed to successive reductions, at low temperature between 100 and 350  C which are commonly assigned to either partial reduction of Co3O4 or reduction-decomposition in hydrogen of residual cobalt precursors and the reduction at higher temperature attributed to the reduction of Co2þ species to metallic Co0. Enache [23] observed different reduction peaks between 400 and 600  C on supported catalysts which do not occur in this sample prepared by means polymerization-combustion technique. The reduction occurs as following: Co3 O4 þ H2 /3CoO þ H2 O

(5)

3CoO þ 3H2 /3Co0 þ 3H2 O

(6)

Fig. 2 e TPR-H2 profiles for Co3O4, NiO and CuO oxides (10% H2eHe ramping rate 10  C minL1).

TPR-H2 profile of NiO oxide displays also a maximum reduction at 359  C with a shoulder around 300  C. The reducibility of the nickel based-catalysts has been extensively studied by TPR. Based on the reduction temperature, different nickel species can be found. Up to 330  C it is attributed the reduction of bulk nickel oxide to metallic Ni0. The shift of the main peak to lower temperatures can be associated with the maximum temperature achieved during the combustion. Comparing the reduction of NiO prepared by the conventional method it turns out that the maximum peak is shifted to lower temperatures from 502  C to 359  C. The reduction at lower temperature is attributed to the reduction of non-stoichiometric Ni2O3 species to NiO, which

Table 3 e Crystalline properties of oxides. Catalyst Co3O4 NiO CuO

Average crystallite size (nm)

Space group

Crystal system

Cell parameters

16.4 25.2 23.9

Fd-3m (227) Fm-3m (225) C2/c (15)

Cubic Cubic Monoclinic

a ¼ 8.056  A a ¼ 4.177  A a ¼ 4.690  A b ¼ 3.420  A c ¼ 5.131  A b ¼ 99.54

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

resulted from the preparation method. TPR-H2 profile for CuO sample exhibits one and wide peak at 230  C, which is related to the reduction of bulk CuO, to Cu0 as been reported previously [24].

3.4. Scanning electron microscopy analysis in emission field (FEG-SEM) Scanning electron microscopy analysis in field emission (FEG-SEM) was performed in a Quanta 200 microscope (FEI), with maximum operating voltage of 20 kV. The images were acquired using EDT detector. Details of operating conditions for images acquisition, such as spot size and working distance

5

(WD), and sample region extension observed are available in the micrographic bar presented here. SEM observations displayed in Fig. 3 show different magnifications the morphological and surface characteristics. At microscopic level, the solids are composed of irregular aggregates, distributed with particle sizes between 0.5 and 200 mm. In principle, these images are related with the texture and the relief created by the elimination of volatile substances that were produced in the combustion of organic compounds during thermal treatment. In the same sense it is notable that these materials present some degree of densification, favoring a compact morphology, which is reflected in the intensity of the principal diffraction signals.

Fig. 3 e Scanning electron microscopy images for a. b. Co3O4, c. d. NiO and e. f. CuO oxides. Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

6

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

3.5.

Activity and selectivity

Fig. 4 shows the conversion of CO and O2 and the selectivity of the catalysts at different temperatures for the selective oxidation of CO under H2 rich atmospheres. The CO and O2 conversions for cobalt oxide started at 50  C and reached 100% at 250  C and at 200  C, respectively. The selectivity of CO2 was 100% at 50  C. These results agree very well with our previous reported results for the perovskite LaCoO3 catalyst [25]. However, the results for NiO and CuO catalysts presented lower CO conversions, (70% at 200  C and 38% at 250  C, respectively), and the conversion of O2 (93%e200  C and 100% at 250  C, respectively) while the selectivity of CO2 (81%e 100  C and 63% at 100  C respectively). The selectivity decreased with increasing temperature, which may be attributed to the competition of CO and H2 occurring above the maximum conversion of CO, favoring the oxidation of H2, thus affecting the conversion of CO [26]. In general terms, the CO coverage decreases above 200  C, leading to a decreasing selectivity due to the increasing H2 oxidation rate [27]. Manasilp and Gulari [28] verified a similar behavior for Pt/Al2O3 catalyst, prepared by solegel technique. The maximum CO conversion (80%) was obtained at 170  C, and a feed composition of 1% CO, 1% O2 and 60% H2. Indeed, the SELOX reaction is characterized by competitive adsorption between CO and dissociated H2. The positive effect of temperature on both conversion and selectivity to CO2 is noticed up to the moment where CO conversion decreases. The CO2 selectivity decreases, while the H2 conversion increases, and CO conversion decreases drastically [26]. Indeed, in this temperature range there is methane formation with the formation of water, according the reactions: CO þ 3H2 /CH4 þ H2 O DH ¼ 217 kJ=mol CO2 þ 4H2 /CH4 þ 2H2 O DH ¼ 178 kJ=mol However, methane and CO are totally or partially oxidized. The possible reactions are: CO þ 1=2O2 /CO2

DH ¼ 284 kJ=mol

CH4 þ 1=2O2 /CO þ 2H2

DH ¼ 36 kJ=mol

All these are very exothermic reactions. Besides, there are other reactions due to the formation of water and CO2: the “shift” and methane reforming reactions.

CO þ H2 O/CO2 þ H2 CH4 þ H2 O/CO þ 3H2 CH4 þ CO2 /2CO þ 2H2

DH ¼ 41kJ=mol DH ¼ 206kJ=mol DH ¼ 247kJ=mol

The last two and the reverse shift reaction are endothermic reactions. In fact, the CO2 selectivity decreases because the reverse “shift” reaction in this temperature range.

Activity For an isoconversion of 40% the temperatures are 100  C, 150  C and 200  C for the Co3O4, NiO, and CuO, respectively, indicating that the activity follows the order: Co3O4 > NiO > CuO. The intrinsic activities under isothermal conditions at 100  C are 0.20  103 s1 for the Co3O4 and 0.08  103 s1 for NiO and CuO. Therefore, the activity of the Co3O4, is 3 times higher than for NiO and CuO. The activity of the nanosized Co3O4 catalyst depends on the temperature. Fig. 5a shows that the initial conversion at 100  C is approximately 60%, but deactivates with time on stream. However, at 200  C the initial conversion is 90% and was very stable with time on stream for this reaction. These results are comparable with the literature [26]. However, Alvarez et al. [29] claim that nanosized rods and wires are very active for the CO oxidation only at low temperatures, deactivating significantly. On the other hand, in the presence of water, the CO oxidation occurred at elevated temperatures around 170  C. Comparing with our nanosized particles Co3O4 catalyst tested with a feed mixture of CO, H2 and O2, the CO conversion reached 90% at the same temperature indicating high activity and good stability of this catalyst. As reported previously [30] the presence of water decreases the CO conversion but is reversible after removing water. In fact, there are competitive adsorptions of CO molecules and dissociated O and H atoms at the surface sites. However, the CO adsorption is preferred and more effective on polycationic oxides, improving the response of intrinsic activity [31].

3.6.

Stability tests

The catalytic stability of the catalysts was tested at two different temperatures, 100 and 200  C under isothermal reaction conditions. The results are displayed in Fig. 5, showing quite different performances. At lower temperature 100  C the CO and O2 conversions decreased for the Co3O4 from 65% to

Fig. 4 e CO and O2 conversion percentage and CO2 selectivity vs temperature a. Co3O4, b. NiO and c. CuO in the selective oxidation of CO. Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

7

Fig. 5 e CO, O2 conversion and CO2 selectivity for a. and b. Co3O4, c. and d. NiO, e. and f. CuO oxides under isothermal reaction conditions.

40% until approximately 2 h, remaining constant for more than 8 h. On the contrary, the NiO and CuO exhibited lower CO and O2 conversions, but are stable for more than 8 h with time on stream. However, the CO2 selectivity for the Co3O4 decreased initially but increased significantly (90%) stabilizing for long time. The CO2 selectivity for Co3O4 and NiO at 100  C are low, suggesting that under such conditions the H2 oxidation prevails and stable, with exception of CuO catalyst. At 200  C the CO and O2 conversions increased situation attributed in all cases to the development of activated thermally reactions [32]. At higher temperatures, during the thermal activation, the cobalt oxide may proportionate more active metallic surfaces that cause a decrease in the reaction activation energy, meanwhile bulk NiO and CuO oxides are known to be insulators and contribute very little to the total active area [32,33].

Long stability test of these materials was tested under isothermal conditions with repetitive cycles (30 h), showing the CO conversion with time on stream (Fig. 6). Results confirm high stability for more than 30 h. The CO conversion of Co3O4 was initially high around 90%. The same behavior was observed for the NiO catalyst, with high initial conversion (78%). The CuO catalyst presented lower conversions but good stability with time on stream. The deactivation process is assigned to the coke deposition as discussed in the next section.

4.

Post-reaction and in situ analyses

4.1.

Raman spectroscopy

Raman spectrums before and after reaction are shown in Fig. 7a and b, respectively. These results exhibit different

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

8

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

CO conversion (%)

100 80 60 40 20 0 0

5

10

15

20

25

30

Time (hours) Fig. 6 e Stability test results of metal oxides under isothermal conditions at 200  C.

vibration modes. The Co3O4 spectrum presents a band around 670 cm1 which is attributed to the octahedral sites (CoO6) with A1 symmetry. Bands around 470 and 510 cm1 are related to the Eg and F2g symmetry modes and the band around 191 cm1 is related with tetrahedral coordination (CoO4) with F2g symmetry. These results agree with XRD results evidencing the spinel formation. The differences in morphology, particle size and sample stoichiometry affect the Raman spectra of the samples, causing band shifts and broadenings. In this case, the sample shows an intense and narrower Raman band, as well as the highest Raman shift, which is indicative of high crystalline structures. The big crystalline size samples present less lattice tension and the Raman spectra are better defined [34]. After catalytic tests, this sample do not present significant changes, and with the objective to analyze the carbon deposition, the spectrum was analyzed from 100 to 1800 cm1, however there are not evidences of carbonaceous species over oxide surface, as shown in Fig. 7b. The Raman spectrum of bulk NiO at room temperature consists of several bands; five vibrational bands; one-phonon (1P) TO at 440 cm1 and LO at 560 cm1 modes. The frequency and shape of the phonon bands do not vary with temperature or catalytic test after 500 min reaction under present reaction

conditions, whereas the scattering intensities are strongly temperature dependent shifting to lower frequencies and decreasing in intensity with increasing temperature, disappearing completely close to the null point. More accurate Raman studies suggest unambiguously the presence of an additional band at about 200 cm1 which was not observed in Fig. 7. The interpretation of this band requires reassessment of the lattice dynamics calculation for bulk NiO, and is only present for the cubic Fm-3m phase [35]. The Raman spectrum of bulk CuO displays three bands located at around 300, 350, 640 cm1, both, the strong band at 300 cm1 and the weak band at 350 cm1 are ascribed to CuO. The band centered at 640 cm1 is associated with Cu2O, additionally the intensity of Raman scattering is directly proportional to the number of scattering centers present in the volume illuminated by the laser beam. Thus, we conclude that for CuO sample is possible coexistence of CuO and Cu2O phases, since Cu2O does not crystallize or cannot be detected by XRD, as reported previously [36]. The broad band located at 1310 cm1indicates the presence of carbonaceous species derived from selective oxidation reaction test, since before catalytic test this band does not appear.

4.2. Diffuse reflectance infrared spectroscopy analyses(DRIFTS) DRIFTS analysis for the selective CO oxidation reaction was performed on Co3O4, NiO and CuO catalysts and the results are shown in Fig. 8. All catalysts do not present CO adsorption band between 1610 and 1630 cm1. The absorption bands located between 3780 and 3790 cm1 are related to stretching OH groups, probably with coordinated water molecules over oxide surface derived from the oxidation reaction, which according to the literature confirm the increasing H2 and O2 consumption from room temperature up to 250  C [37]. At the same time, bands located between 3660e3670 cm1 and 3240e3270 cm1 are mainly related to stretching OeH groups that confirm the presence of water. The major bands located at 3020 cm1 and between 1300 and 1350 cm1 are related to the presence of methane and with a strong stretching signals of CeH bond, as has been reported previously [38].

Fig. 7 e Raman spectra for a. Co3O4, b. NiO and c. CuO metal oxides before and after catalytic test and obtained for polymerization-combustion method. Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

9

Fig. 8 e In situ DRIFTS spectrums of selective CO oxidation reaction for a. Co3O4, b. NiO, c. CuO and d. reaction blanks at 50 and 100  C.

The small bands between 2340 and 2350 cm1 are associated to the stretching vibration of average strength of the bonds O]C]O, suggesting the presence of carbon dioxide entrapped in the pores of the oxides and it is clear that its concentration changes along time and represent the selectivity of oxidation reaction to CO2 formation. Thus, the Co3O4 oxide presents the most important change in terms of CO2 concentration, since it has the highest CO and O2 conversion and CO2 selectivity, followed by NiO and CuO oxides, respectively. The bands between 2170 and 2180 cm1, are related to the presence of CeO and C]O bonds due to asymmetric stretching movement for CO and CO2 molecules, in the region of 1400e1550 cm1, which are important signals corresponding to strong stretching of CeO bond. Finally, the bands between 725 and 1180 cm1 correspond to the bonding movement out of plane for CeH bonds and with different signals associated with CO2 3 species over surface of the metal

cations and confirm the consumption of O2 to formation of carbonate species [39]. The XRD and Raman analyses after catalytic test confirm the existence of a partial reduction of oxides to the respective metallic phase, as observed in Fig. 9. There is no evidence of carbon deposition over surface materials. The refinement of XRD data using Rietveld confirm that in each case there are two principal phases related with reduction of metal cation and its respective oxide. In case of cobalt oxide, the XRD analysis suggests the presence of CoO and Co3O4 phases, identified with its respective diffraction lines. In the case of nickel oxide the analyzed sample shows the metallic phase corresponding to metallic nickel and in the case of copper oxide the sample presents a high grade of reduction, and XRD results confirm the presence of Cu2O and metallic copper. Finally, results showed that these oxide catalysts present a high resistance to carbonaceous formation under present

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

10

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

The XRD and Raman analyses after catalytic test confirm the existence of a partial reduction of oxides to the respective metallic phase, no evidence of carbon deposition over surface materials. In the case of nickel oxide the analyzed sample shows the metallic phase corresponding to metallic nickel and in the case of copper oxide the sample presents a high grade of reduction and XRD results confirm the presence of Cu2O and metallic Cu. Finally, results showed that these oxide catalysts present a high resistance to carbonaceous formation under present reaction conditions and confirm the effectiveness of polymerization-combustion technique for synthesis of active high catalyst in the selective oxidation of CO.

Acknowledgments J. A. Go´mez-Cuaspud thanks FAPERJ for post-doctoral scholarship. The authors gratefully acknowledge to Carlos Andre´ de Castro Perez and Marta M. M. Amorim for discussions and CNPq, CAPES, and FINEP for financial support.

Fig. 9 e XRD patterns for a. Co3O4, b. NiO ; metallic Ni and c. Cu2O and metallic Cu, obtained after catalytic test.

reaction conditions and confirm the effectiveness of polymerization-combustion technique for synthesis of high active catalyst in the selective oxidation of CO.

5.

Conclusions

This study we evaluated the effect of citric acid in the synthesis of transition catalytic oxides, and initial conditions such as pH, concentration and homogeneity, which are crucial parameters in the design and tailoring of some catalytic properties such as porosity, homogeneity and specific activity. The polymerization-combustion technique provides an appropriate method to obtain different kind of active transition oxides. XRD results showed that the nanocrystallites obtained by means polymerization-combustion route promoted the stabilization of the active phases in contrast with the same catalysts prepared by the traditional thermal treatment of nitrate precursor salts. For an isoconversion the activity follows the order: Co3O4>NiO > CuO. The intrinsic activities under isothermal conditions at 100  C are 0.20  103 s1 for the Co3O4 and 0.08  103 s1 for NiO and CuO. Therefore, the activity of the Co3O4, is 3 times higher than for NiO and CuO. The catalytic stability of the catalysts was tested at two different temperatures, 100 and 200  C under isothermal reaction conditions. Results at lower temperature 100  C showed that the CO and O2 conversions decreased until approximately 2 h, remaining constant for more than 10 h. The Raman spectra before and after reaction showed that Co3O4 presented octahedral sites (CoO6) and tetrahedral coordination (CoO4) evidencing the spinel formation.

references

[1] Litster S, McLean G. PEM fuel cell electrodes. Journal of Power Sources 2004;130:61e76. [2] Bhatia KK, Wang CY. Transient carbon monoxide poisoning of a polymer electrolyte fuel cell operating on diluted hydrogen feed. Electrochimica Acta 2004;49:2333e41. [3] Nielson FPR, Bonfim RPF, Mariana MVM, Schmal M. Investigation of activity losses of gold nanoparticles in the CO selective oxidation. Journal of Power Sources 2010;195:7386e90. [4] Choudhury S, Bhuiyan MA, Hoque S m. Effect of sintering temperature on apparent density and transport properties of NiFe2O4: synthesized from nanosize powder of NiO and Fe2O3. International Nano Letters 2012;2:6e9. [5] Veith M, Mathur S, Lecerf N, Huch V, Decker T. Sol-gel synthesis of nanoscaled BaTiO3, BaZrO3 and BaTi0:5Zr0:5O3 oxides via single source alkoxide precursors and semialkoxide routes. The Journal of Sol-Gel Science and Technology 2000;15:145e58. [6] Xu G, Zhao G, Ren Z, Shen G, Han G. PVA assisted synthesis of nanosized perovskite PZT powder by a two stage precipitation route. Materials Letters 2006;60:685e8. [7] Deganello F, Marci G, Deganello G. Citrate nitrate autocombustion synthesis of perovskite type nanopowders: a systematic approach. Journal of the European Ceramic Society 2009;29(3):439e50. [8] Chakrabarti N, Maiti HS. Chemical synthesis of PZT powder by autocombustion of citrate nitrate gel. Materials Letters 1997;30(2e3):169e73. [9] Qi X, Zhou J, Yue Z, Gui Z, Li L. Auto-combustion synthesis of nanocrystalline LaFeO3. Mater Chem Phys 2003;78(1):25e9. [10] Kingsley JJ, Pederson LR. Energetic materials in ceramics synthesis. MRS Proceedings 1992;296:296e361. [11] Mimani T. Instant synthesis of nanoscale spinel aluminates. Journals of Alloys Compounds 2001;315:123e8. [12] Alinejad B, Sarpoolaky H, Beitollahi A, Saberi A, Afshar S. Synthesis and characterization of nanocrystalline MgAl2O4spinel via sucrose process. Materials Research Bulletin 2008;43:1188e94.

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024

i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 3 ) 1 e1 1

[13] Chandradass J, Jun B, Bae D. Effect of different fuels on the aluminaezirconia nanopowder synthesized by solegel autocombustion method. Journals of Non-crystalline Solids 2008;354:3085e7. [14] Yang W, Zhang R, Chen B, Bion N, Duprez D, Royer S. Activity of perovskite-type mixed oxides for the low-temperature CO oxidation: evidence of oxygen species participation from the solid. Journal of Catalysis 2012;295:45e58. [15] Hong NT, Caldero´n-Moreno JM, Popa M, Crespo D, Hong L, Phuc NX. LaNiO3 nanopowder prepared by an amorphous citrate route. Journal of the European Ceramic Society 2006;26:403e7. [16] Martı´nez A, Lo´pez C, Ma´rquez F, Dı´az I. FischereTropsch synthesis of hydrocarbons over mesoporous Co/SBA-15 catalysts: the influence of metal loading, cobalt precursor, and promoters. Journal of Catalysis 2003;220:486e99. [17] Avgouropoulos G, Ioannides T. Selective CO oxidation over CuOeCeO2 catalysts prepared via the ureaenitrate combustion method. Applied Catalysis A: General 2003;244:155e67. [18] Rivas ME, Fierro JLG, Guil-Lo´pez R, Pen˜a MA, La Parola V, Goldwasser MR. Preparation and characterization of nickelbased mixed-oxides and their performance for catalytic methane decomposition. Catalysis Today 2008;133e135:367e73. [19] Patil PS, Kadam LD. Preparation and characterization of spray pyrolyzed nickel oxide (NiO) thin films. Applied Surface Science 2002;199:211e21. [20] Cornaglia LM, Carrar CR, Petunchi JO, Lombardo EA. Characterization of cobalt-impregnated VPO catalysts. Catalysis Today 2000;57:313e22. [21] Rastkar AR, Niknam AR, Shokri B. Characterization of copper oxide nanolayers deposited by direct current magnetron sputtering. Thin Solid Films 2009;517:5464e7. [22] Khodakov AY. Stoichiometric reduction of Co3O4 to metallic Co. Applied Catalysis A: General 2003;254:273e88. [23] Enache DI, Rebours B, Roy-Auberger M, Revel R. In situ XRD study of the influence of thermal treatment on the characteristics and the catalytic properties of cobalt-based fischer–tropsch catalysts. Journal of Catalysis 2002;205:346e53. [24] Xiaoyuana J, Guanglie L, Renxian Z, Jianxin M, Yu C, Xiaoming Z. Studies of pore structure, temperatureprogrammed reduction performance, and micro-structure of CuO/CeO2 catalysts. Applied Surface Science 2001;173:208e20. [25] Chagas CA, Toniolo FS, Magalhaes NR, Schmal M. Aluminasupported LaCoO3 perovskite for selective CO oxidation (SELOX). International Journal of Hydrogen Energy 2012;37(6):5022e31.

11

[26] Scheuneman R. Oxidac¸a˜o seletiva do CO com catalisadores de Pt suportados em o´xidos mistos de ferro-zirconia. Tese de Doutorado. Rio de Janeiro: Universidade Federal do Rio de Janeiro; 2009. [27] Kim DH, Lim MS. Kinetics of selective CO oxidation in hydrogen-rich mixtures on Pt/alumina catalysts. Applied Catalysis A: General 2002;224:27e38. [28] Manasilp A, Gulari E. Selective CO oxidation over Pt/alumina catalysts for fuel cell applications. Applied Catalysis B: Environmental 2002;37:17e25. [29] Alvarez A, Ivanova S, Centeno MA, Odriozola JA. Subambient CO oxidation over mesoporous Co3O4: effect of morphology on its reduction behavior and catalytic performance. Applied Catalysis A: General 2012;431e432:9e17. [30] Magalha˜es RN, Toniolo FS, da Silva VT, Schmal M. Selective CO oxidation reaction (SELOX) over cerium-doped LaCoO3 perovskite catalysts. Applied Catalysis A: General 2010;388:216e24. [31] Mooi J, Selwood PW. Structure and catalytic activity of supported manganese, copper and iron oxides. Journal of American Chemical Society 1952;74(10):2461e4. [32] Narayan RV, Kanniah V, Dhathathreyan A. Tuning size and catalytic activity of nano-clusters of cobalt oxide. Journal of Chemical Sciences 2006;118(2):179e84. [33] Cindrella L, Jeyachandran R. European nanosystems worshop, 03e04 December 2007. Paris, France; pp. 101e106. [34] Escribano R, Sloan JJ, Siddique N, Sze N, Dudev T. Raman spectroscopy of carbon-containing particles. Vibrational Spectroscopy 2001;26:179e86. [35] Escribano R, Sloan JJ, Siddique N, Sze N, Dudev T. Raman spectroscopy of carbon-containing particles. Vibrational Spectroscopy 2001;26:140e201. [36] Vedeanu N, Cozar O, Ardelean I, Lendl B, Magdas DA. Raman spectroscopic study of CuOeV2O5eP2O5eCaO glass system. Vibrational Spectroscopy 2008;48:259e62. [37] Jernigan GG, Somorjai GA. Carbon monoxide oxidation over three different oxidation states of copper: metallic copper, copper (I) oxide, and copper (II) oxide - a surface science and kinetic study. Journal of Catalysis 1994;147(2):567e77. [38] Schmal M, Cesar DV, Souza MV, Guarido CE. Drifts and TPD analyses of ethanol on Pt catalysts over Al2O3 and ZrO2partial oxidation of ethanol. Canadian Journal of Chemical Engineering 2011;89(5):1166e75. [39] Refat MS, Elsabawy KM. Infrared spectra, Raman laser, XRD, DSC/TGA and SEM investigations on the preparations of selenium metal, (Sb2O3, Ga2O3, SnO and HgO) oxides and lead carbonate with pure grade using acetamide precursors. Bulletin of Material Science 2011;34(4):873e81.

Please cite this article in press as: Go´mez-Cuaspud JA, Schmal M, Nanostructured metal oxides obtained by means polymerization-combustion at low temperature for CO selective oxidation, International Journal of Hydrogen Energy (2013), http:// dx.doi.org/10.1016/j.ijhydene.2013.04.024