Influence of Ce addition on the catalytic behavior of alumina-supported Cu–Co catalysts in NO reduction with CO

Journal of Colloid and Interface Science 354 (2011) 777–784

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Influence of Ce addition on the catalytic behavior of alumina-supported Cu–Co catalysts in NO reduction with CO I. Spassova a,⇑, N. Velichkova a, D. Nihtianova b, M. Khristova a a b

Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria Central Laboratory of Mineralogy and Crystallography, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria

a r t i c l e

i n f o

Article history: Received 14 July 2010 Accepted 11 November 2010 Available online 9 December 2010 Keywords: NO reduction with CO Ceria Copper–cobalt catalysts Alumina-supported catalysts

a b s t r a c t The effect of Ce addition to alumina-supported copper, cobalt, and copper–cobalt oxides with low loadings on the catalysts efficiency in NO reduction with CO was studied. The attention was focused on varying the impregnation procedure in the ternary-supported catalysts in order to determine the best catalyst as well as the reasons for the enhanced catalytic activity. Ternary Co–Cu–Ce and binary Co–Ce, Cu–Ce, and Cu–Co-supported alumina were prepared and characterized by ICP, XRD, TEM, adsorption studies, XPS, H2-TPR, and catalytic investigations. The high activity of the ternary and the binary catalysts was determined by the favorable influence of the added cerium on the dispersion of the copper and cobalt active phases. The presence of ceria contributes to the formation of appropriate active phases, resulting in catalytic sites on the surface of the samples that promote the reduction of NO with CO. Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction NOx removal from exhaust gases still remains one of major challenges in the area of environmental catalysis [1–3]. The most common approach is the reaction with residual reductants (unburned or partially burned hydrocarbons and CO) in the exhaust. Various catalysts have been extensively studied containing noble metals, ion-exchanged zeolites, and metal oxides [4–9]. The metal-supported alumina catalysts have received some of the most attention due to their high activity and stability. It has been demonstrated that alumina is good support for catalysts due to its additional ability for dispersion of transition metal cations [10,11]. One of the more efficient catalysts found is Cu/Al2O3 [12–14]. While investigating Cu/Al2O3 (0.5–10 wt.%) the authors [15] have found that the active sites for the NO–CO reaction are the isolated Cu2+ species supported on c-Al2O3, and the coexistence of aggregated species drastically reduces the activity of NO reduction. In the case of a loading amount of more than 3 wt.%, most of the supported Cu species is in a highly dispersed form, but the residues are aggregates. Aggregated species increased with the loading amount. Another efficient alumina-based catalyst testified for NO reduction with various reductants –Co/Al2O3 has attracted much attention recently due to its good activity and selectivity to nitrogen [10,16]. It is well known that the deNOx activity is greatly related to the dispersion of cobalt species that are usually influenced by the preparation method, the Co concentration, and the ⇑ Corresponding author. Fax: +359 2870 50 24. E-mail address: [email protected] (I. Spassova). 0021-9797/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2010.11.028

calcination temperatures. Some methods have been suggested to improve the dispersion of Co2+ ions and to prevent the aggregation of cobalt particles. The decrease of cobalt loading (typically <2–5 wt.%) is one of the ways to reduce the possibility for the formation of large particles [10,17]. As an abundant, nontoxic, and inexpensive material, ceria (CeO2) has attracted much attention recently in catalysis mainly due to its prominent ability to store/release oxygen as an oxygen reservoir via the redox shift between Ce4+ and Ce3+ under oxidizing and reducing conditions, respectively [18]. Inclusion of ceria has been reported to be beneficial for the NO + CO reaction [19,20] and, due to its known oxygen storage function, this oxide may help to favor NO in its competition with oxygen for reaction with hydrocarbons. Pure ceria is poorly thermostable and undergoes sintering; hence the problem could be overcome by incorporation of another metal oxide into the ceria lattice, thus facilitating the formation of mixed oxides or solid solutions [21,22]. The mixing of two different oxides offers an opportunity not only to improve the performance of the involved metal oxide but also to form new stable compounds that may lead to totally different physicochemical properties and catalytic behavior from the individual components. Doping of ceria by divalent or trivalent ions can increase the concentration of oxygen vacancies or improve its thermal stability [23]. Cu/CeO2 has been used as a model catalyst to investigate the synergetic mechanism between copper species and ceria for NO reduction by CO. It has been suggested that the relatively free movement of oxygen from ceria to supported copper, caused by the three copper species, leads to oxygen vacancies in ceria, significantly enhancing the NO conversion of Cu/CeO2

778

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784

during the NO + CO reaction [24]. Also, it has been found that metal/ceria interactions induced by the establishment of contacts between the two components strongly affect their redox properties and, as a consequence, their catalytic properties [25]. Recently, a Cu–Ce/Al2O3 catalyst has shown promising properties for NO reduction with CO [26]. Cobalt-promoted Cu/CeO2 and Cu–Ce/ Al2O3 catalysts have been studied in oxidation reactions [27,28]. It has been found that a small amount Co significantly improves the activity in CO oxidation. However, such investigations for NO reduction with CO are not found. In the present work, we investigate the effect of cerium addition to alumina-supported copper, cobalt, and copper–cobalt oxides with low loadings on the catalysts efficiency in NO reduction with CO. We suggest that the combination of the supported metal oxides will lead to the formation of active phases with physicochemical properties suitable for the studied reaction. The attention is also focused on varying the impregnation procedure in the ternary-supported catalysts in order to determine the best catalyst as well as the reasons for the enhanced catalytic activity. 2. Materials and methods 2.1. Catalyst preparation The catalysts were prepared by successive impregnation of the support with nitrate solutions of copper, cobalt, and cerium. The impregnation was performed with nitrate metal solutions Cu(NO3)2, Co(NO3)2, and (NH4)2Ce(NO3)6 with concentrations of 8 g Me (Me = Cu, Co, Cu + Co with Cu:Co = 1:2) for 100 ml water and 3.6 g Ce for 100 ml water. A 0.3–0.8-mm fraction of AlO(OH) Rhône Poulenc (Al) with ABET = 271 m2/g and Vt = 0.45 cm3/g was used as a support. Copper–cobalt–cerium-supported catalysts were prepared by altering the succession of the impregnation with two kinds of solutions—cerium solution and copper–cobalt solution. The impregnation took place in 16 h, and after this the samples were dried for 2 h at 120 °C and calcined for 6 h at 450 °C. In this way the following ternary-supported samples were prepared: Al(Cu + Co + Ce), simultaneous impregnation with the three metals in the impregnating solution; AlCe(Cu + Co), successive impregnation first with cerium-containing solution and then with copper– cobalt-containing solution; Al(Cu + Co)Ce, successive impregnation first with copper–cobalt-containing solution and then with cerium-containing solution. Binary-supported Al(Co + Ce), Al(Cu + Ce), and Al(Co + Cu) catalysts with metals deposited simultaneously were prepared. Single AlCo, AlCu, and AlCe-supported ones were obtained for comparison.

pressure of 1  108 Pa. The photoelectron spectra were excited using unmonochromatized Al Ka1,2 radiation (hm = 1486.6 eV), C1s line at 284.1 eV used as reference one. The composition and chemical surrounding of samples were investigated on the basis of the areas and binding energies of C1s, O1s, Cu2p, Co2p, Ce3d, Al2p photoelectron peaks and Scofield’s [29] photoionization cross sections. H2-TPR measurements were made on a Chem BET TPR/TPD apparatus Quantachrome (USA) using 10% H2 in Ar, at a gas flow of 20 ml/min and heating rate of 10 °C/min. 2.3. Catalytic tests The catalytic experiments on NO + CO reduction were carried out in a conventional flow apparatus in the temperature range 25–300 °C. The catalytic tests were performed with a gas mixture so that a redox index [CO]inlet/[NO]inlet  1 was maintained, where the gas mixture contained approximately 1200 ppm NO and 1200 ppm CO; GHSV was 26,000 h1. After a catalytic test at 25 °C and isothermal desorption in Ar flow, temperature-programmed desorption (TPD) was carried out in the same catalytic apparatus at a heating rate of 13 °C min1 with an Ar flow in a range of 25–300 °C. The concentrations of NO and CO were continuously measured by infrared gas analyzers. The outlet concentrations of NO and CO were monitored by a ‘‘UNOR 5-Maihak’’ (Germany) and the CO2 by an ‘‘Infralyt 2106’’ (Germany). A Specord 75 IR (Germany) spectrophotometer with a 1-m folded path gas cell (Specac) was used for determination of the outlet N2O content. A thermal converter was applied to NOx (NO + NO2) analysis. The N2 concentration in the outlet gas was determined on the basis of the material balance with respect to NO consumption. The transient response method [30] was used to study the interaction of the gas phase with the catalyst surface. The turnover frequency (TOF) values for NO conversion to N2 (converted NO per surface unit of active metal) over the investigated catalysts per copper and per cobalt content were calculated in order to evaluate the cerium promotion on the studied catalysts:

TOF ¼

V CMg 22:4  P  A

where V is space velocity (cm3/h), C is inlet concentration of NO (vol.%), M is molecular mass of NO, g is a conversion of NO (%), PMe (g) is the active metal content, Me = Cu, Co, and A is specific surface area (m2/g). 3. Results and discussion

2.2. Catalyst characterization

3.1. Chemical analysis and adsorption characteristics

The Cu, Co, and Ce concentrations in the impregnation solution were determined by chemical analysis. The Cu, Co, and Ce content in the supported catalysts after extraction with HCl was determined by means of ICP-AES analysis. X-ray diffraction (XRD) data were obtained using a Bruker D8 Advance diffractometer with Cu Ka radiation and SolX detector. The textural characteristics were determined by low-temperature (77.4 K) nitrogen adsorption in a conventional volume apparatus. The specific surface area ABET was calculated by the BET method. The total pore volume (Vt) was determined at a relative pressure p/po = 0.95. The mesopore-size distribution was made according to the Pierce method. The TEM investigations were performed by TEM JEOL 2100 with an accelerating voltage of 200 kV. The XPS measurements were done in the UHV chamber of an ESCALAB-Mk II (VG Scientific) electron spectrometer with a base

Table 1 shows the results for the chemical analysis of the obtained catalysts, the BET specific surface areas, the total pore volumes, and the phase compositions determined by XRD. Irrespective of the fact that the impregnating solutions are with the same Cu, Co, Ce, and (Cu + Co) concentrations the samples have different mass% copper, cobalt, and cerium, depending on the nature of the active component and on the sequence of impregnation. The succession of the impregnation affects the metal content due to the different selective adsorption of the metal ions on the support [31]. The total metal content for binary and ternary catalysts varies from 2.38 to 3.90 mass%. The data for texture parameters show that the deposition of the active phase leads to a decrease of the specific surface area and of the total pore volume as compared to that of the initial AlO(OH). This diminution depends on the metal content and on the nature of the components of the active phases.

779

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784 Table 1 Chemical composition, specific surface area, total pore volume, and phase composition. Sample

Al(Cu + Co + Ce) AlCe(Cu + Co) Al(Cu + Co)Ce Al(Cu + Co) Al(Cu + Ce) Al(Co + Ce) AlCu AlCo AlCe

Metal content in sample (mass%) Cu

Co

Ce

0.66 0.52 0.65 0.88 2.65 – 2.60 – –

1.00 1.44 0.90 1.50 – 2.50 – 3.00 –

1.10 0.98 0.92 – 0.98 1.40 – – 1.05

Total metal (mass%)

ABET (m2 g1)

Vt (cm3 g1)

Phase composition (XRD)

2.76 2.94 2.47 2.38 3.58 3.90 2.60 3.00 1.05

160 164 175 184 196 202 180 235 188

0.31 0.32 0.32 0.34 0.39 0.28 0.40 0.34 0.42

Co3O4 CuCo2O4, CeO2 Co3O4 CuCo2O4, CeO2 Co3O4, CeO2 Co3O4, CuCo2O4 CeO2, trace CuAl2O4 Co3O4, CeO2 CuO, trace CuAl2O4 Co3O4 CeO2

The XRD data of the support indicate the presence of orthorhombic AlO(OH) only. All diffraction peaks are relatively broad and reveal the fine crystalline nature of the support. The XRD data do not present a satisfactory picture for the phase composition of the catalysts, because the metal content of the active metals is quite low and the supported oxides are with a high dispersity. When Ce is deposited alone, the lines of CeO2 are predominant, those of the support being negligible in intensity. The Al(Co + Ce) sample shows small peaks for Co3O4 along with CeO2. Peaks for CoAl2O4 are undistinguishable with those for Co3O4. When only Cu is deposited, there are pronounced peaks for CuO. The diffraction patterns of the Al(Cu + Ce) show peaks for CeO2 and a trace of CuAl2O4, with CuO not being observed. Similar findings on Cu–Ce–O samples which do not show CuO reflection peaks have been reported by [32,33]. The same is the situation with the ternary-supported samples where no signs of CuO are detected along with the peaks for CeO2 and Co3O4. Peaks of CuCo2O4 are present in a binary Al(Cu + Co) and in ternary Al(Cu + Co + Ce) and AlCe(Cu + Co). 3.2. TEM data A more complete picture of the phase composition of the catalysts gives the TEM analysis. TEM data show that all catalysts are heterogeneous in their morphology. A bright-field micrograph of Al(Cu + Co + Ce) is presented in Fig. 1. Two types of forms are registered by TEM—one shape is elongated, and the other is spherical. The spherical particles have smaller sizes than the elongated particles. The rest of the investigated catalysts have similar morphology, Al(Co + Ce) being an exception where particles with a spherical shape are dominant. SAED (selected area electron diffraction) data show that all catalysts have a complex phase composition. The ternary-supported samples differ in composition, according to the sequence of preparation. Fig. 1 presents the electron diffraction diagram of the catalyst Al(Cu + Co + Ce). Here, CeAlO3, CuO, and CeO2 are predominant phases, but impurities of Cu2O and spinels CuCo2O4 and CuAl2O4 are observed also. SAED patterns of AlCe(Cu + Co) show—CuCo2O4, Co3O4, CeO2, and CuAl2O4, Al(Cu + Co)Ce—mainly CeO2, with the addition of Co3O4 and CuCo2O4. The data from TEM analysis are in good agreement with those of XRD. 3.3. Adsorption studies The mesopore distribution curves for all supported catalysts and for the initial AlO(OH) are presented in Fig. 2. The distribution curve of AlO(OH) shows an ill-resolved peak at 20–30 nm, probably due to the nonhomogeneity of the mesoporous texture of the support. This nonhomogeneity is determined by the globular nature of AlO(OH). The active phase deposition leads to redistribution of the part of the mesopore space of the support due to blocking and/or partial

Fig. 1. TEM image of Al(Cu + Co + Ce).

pore filling. Cerium plays an important role in this redistribution as a strong texture modificator. It influences in different ways the adsorption of Cu and Co ions in the binary-supported catalysts. More strongly cerium assists the adsorption of copper in sample Al(Cu + Ce). As a result the pores are filled to a greater extent and the maximum in the distribution curve is shifted to the fine pores. The promotion of Ce on Co adsorption is weaker (sample Al(Co + Ce)). In the binary Al(Cu + Co) copper and cobalt are competitors for the same adsorption centers of oxide support. Depending on their content a better resolved peak in the distribution curve could be observed. The adsorption of Cu and Co in the ternary-supported catalysts depends on the succession of copper–cobalt and Ce deposition. When Ce is deposited finally Cu and Co are able to enter deeper in the pores and the peak in the distribution curve is more pronounced. The distribution curve in shape and position resembles that of the binary Al(Cu + Co). When cerium is deposited first the distribution curve changes. Cerium occupies the adsorption centers of Cu and Co and the maximum in the distribution curve is shifted to the pores with larger size. This is a result of blockage of the pores

780

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784 Table 2 Surface composition from XPS.

11

10

9

2

Sample

Ce (at.%)

Co (at.%)

Cu (at.%)

Al(Cu + Co + Ce) AlCe(Cu + Co) Al(Cu + Co)Ce Al(Cu + Co) Al(Cu + Ce) Al(Co + Ce)

0.6 0.2 0.8 – 0.6 0.8

0.3 0.4 0.2 0.3 – 0.3

0.1 0.2 – 0.3 0.4 –

8

3.4. XPS data 7

dVp/dRp

1 6 3 5 6 4

3

2

7 4 5

1

0 2

4

6

8

Rp, nm Fig. 2. Poresize distribution 1 – AlO(OH), 2 – Al(Cu + Ce), 3 – Al(Co + Ce), 4 – Al(Cu + Co), 5 – AlCe(Cu + Co), 6 – Al(Cu + Co + Ce), 7 – Al(Cu + Co)Ce.

because of the dominating Ce adsorption. During the simultaneous deposition (sample Al(Cu + Co + Ce)) the active metals are competitors for the same adsorption centers but Ce is not able to cope with Cu and Co. However, it affects their disposition in the mesoporous space and hinders copper and cobalt to penetrate deeper to the pores.

To gain a better understanding of the chemical state of all the elements on the catalysts surface the samples were investigated by XPS techniques. Fig. 3 shows the XPS spectra of Cu 2p, Co 2p, and Ce 3d for the binary and ternary samples. Table 2 presents the surface metal content in the studied catalysts. The Cu 2p XP spectrum in Fig. 3 shows two main peaks of Cu 2p3/2 and Cu 2p1/2 at about 934.2 and 954.1 eV. The Cu 2p3/2 spectrum in terms of both presence of intense satellite structure and binding energy value indicates that the major part of copper is Cu2+; the main peaks are asymmetric to the side of the lower energies, probably due to the presence of Cu+ and Cu [34,35]. The Co 2p3/2 transition in the catalysts is characterized by a main peak and a satellite one on the higher binding energy side. The Co 2p3/2 main peak of the catalyst is centered at a binding energy of about 780.2 eV. On the other hand, the Co 2p3/2–2p1/2 spin– orbit splitting is visible and equal to 15.4 eV, similar to [36] for the XPS spectra of Co3O4, and in addition to XRD these results point to a simultaneous presence of Co2+ and Co3+ species in the spinel phase on the surface of the samples, not excluding the presence of CuxCo3xO4 [37] in the ternary samples. The Ce 3d spectrum displays the Ce 3d5/2–Ce 3d3/2 spin–orbit splitting. The oxidation state of cerium could be ascertained from the intensity of the satellite peak at 914.0 eV, characteristic for the Ce4+ state. In CeO2, this peak bears 14% of the total integrated intensity of the Ce 3d transitions [38,39]. In all cerium-supported catalysts this value is lower than 10%, thus indicating the presence of some Ce3+ in addition to Ce4+. The surface composition of the ternary-supported catalysts shows differences in copper, cobalt, and cerium content, depending

Cu2p

Co2p

Ce3d

6

6

6

5

5

5

4

4 4

2 3 3 1 22

1

930

940

950

960

780

790

800 880

900

920

Binding energy, eV Fig. 3. XP spectra of : 1 – Al(Cu + Co), 2 – Al(Cu + Ce), 3 – Al(Co + Ce), 4 – Al(Cu + Co)Ce, 5 – AlCe(Cu + Co), 6 – Al(Cu + Co + Ce).

781

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784

on the sequence of deposition. When cerium is deposited first its surface content is quite low. For the rest of the ternary-supported catalysts the surface cerium content is comparable. On other hand, copper is not observed on the surface of the sample Al(Cu + Co)Ce; hence it penetrates deep in the bulk of the support, resembling the effect reported in [40] and explained with the higher selective adsorption of copper ions on the alumina support in comparison to cobalt. These data confirm the conclusions made from analysis of the textural characteristics of the samples. When the metals are simultaneously deposited cerium favors the relatively even metal disposition on the surface of the support.

peak at a region 650–670 °C is attributed to the reduction of the bulk ceria. In the case when cerium is deposited first its influence on the reduction of Co3O4 and CuO–Co3O4 is less pronounced. When Ce is deposited after cobalt and copper a broadening on the high-temperature side of the second peak appears as a shoulder. So, the reducibility of the ternary catalysts is affected by the impregnation sequence of the metals. These results indicate that the addition of Ce to the samples could improve the dispersion and enhance the reducibility of copper and cobalt species. The existence of the copper species could also favor the reducibility of the cobalt species. 3.6. Catalytic studies

3.5. H2-TPR data The TPR profiles of Al(Cu + Co + Ce), Al(Co + Ce) and Al(Cu + Ce) samples are displayed in Fig. 4. For the sample Al(Co + Ce) we observe five peaks—280, 361, 445, 556, and 665 °C. According to [36,41], we suggest a stepwise reduction of cobalt oxide via Co3+ ? Co2+ ? Co0. The peak at 280 °C is attributed to the reduction of Co3+ to Co2+; the peak at 361 °C and the shoulder one at 400 °C are ascribed to the reduction of Co2+ to Co0. The very broad and weak in intensity maximum around 650–670 °C is registered for all samples. It is attributed to the reduction of bulk CeO2. The peaks at 556 and 665 °C, we suggest to be connected with the reduction of CeO2 according to [42,43]. We observe five peaks, at 121, 211, 253, 449, and 650 °C, for the sample Al(Cu + Ce). Composite TPR signals have been found for alumina-supported CuO by many authors. It was hypothesized that small CuO clusters and/or isolated Cu2+ ions are reduced at lower temperatures than larger CuO particles. So, we deduce the reduction of Cu particles strongly interacting with ceria to be reproduced by the very weak peak at 121 °C as in [44], while the peaks at 211 and 253 °C are attributed to the larger Cu particles less or not associated with ceria. Peaks at 449 and 650 °C as noted for Al (Cu + Ce) are connected with the reduction of the CeO2 particles. Four reduction peaks are observed at 120, 215, 325 °C, and a broad shoulder at 650–670 °C for the ternary Al(Cu + Co + Ce) sample, similar to that found in [45]. In our case, the reduction of Cu particles interacting with ceria appears at the same position as with the binary sample. The reduction of cobalt species occurred at a relatively lower temperature, which is ascribed to the strong synergistic effect of copper and cobalt. The peaks for nonassociated Cu particles are overlapped with those for Co reduction. A broad

Fig. 5 illustrates the dependence of the conversion degree of NO on temperature for the studied catalysts. Pure alumina support shows low activity itself (of about 16% at 300 °C). In the present experiments, N2O was detected in the reaction system up to 130 °C over the ternary-supported catalysts and up to 150 °C over the binary-supported samples. When the temperature was higher than these temperatures, no N2O was measured and the reaction proceeds to N2 only. It is obvious that the ternary-supported catalysts are more active than the binary-supported ones in the studied reaction. The Al(Cu + Co + Ce), prepared by simultaneous impregnation with the three active metals, is more active than the AlCe(Cu + Co) and Al(Cu + Co)Ce and at 130 °C it shows over 50% conversion of NO. The conversion of NO for all other samples at the same temperature is below 50%. From Fig. 5 is evident that AlCu, AlCo, and AlCe express very low activity and at 300 °C the NO conversion is about 20%. The addition of Ce in the binary systems affects the activity at temperatures over 130 °C, whereas the activity of the ternary-supported catalysts is enhanced even at room temperature. The higher activity of the ternary and the binary catalysts is specified probably by the favorable influence of the added cerium on the dispersion of the copper and cobalt active phases and on the formation of appropriate active phases, resulting in catalytic sites for reduction of NO with CO.

1

100

2 80

3

H2 consumption

3

2

1 200

400

600

800

Temperature, ºC Fig. 4. H2-TPR profiles of: 1 – Al(Co + Cu + Ce), 2 – Al(Co + Ce), 3 – Al(Cu + Ce).

NO conversion, %

4 5 60 6

40

7 8 9

20

0 0

100

200

300

Temperature, ºC Fig. 5. Temperature dependence of NO conversion degree on samples: 1 – Al(Cu + Co + Ce), 2 – AlCe(Cu + Co), 3 – Al(Cu + Co)Ce, 4 – AlCu + Co, 5 – AlCu + Ce, 6 – AlCo + Ce, 7 – AlCu, 8 – AlCo, 9 – AlCe.

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784

1 2 3

Intensity (a.u.)

4

5

6

7 8 9

50

100

150

200

250

300

Temperature, ºC Fig. 6. TPD spectra of NO for 1 – Al(Cu + Co + Ce), 2 – AlCe(Cu + Co), 3 – Al(Cu + Co)Ce, 4 – Al(Cu + Co), 5 – Al(Cu + Ce), 6 – Al(Co + Ce), 7 – AlCu, 8 – AlCo, 9 – AlCe.

penetrate deeper to the pores, so they stay more on the surface, facilitating the NO reduction with CO. The catalytically active sites are a complex of ions from the active phases and from the support’s surface. The sequence of preparation strongly affects the formation of various kinds of catalytically active sites. In the case of Al(Cu + Co + Ce) an important role is probably played by the availability of Ce in two oxidation states Ce3+ and Ce4+, proved by XPS and TEM analysis. A transient response method was applied to obtain information about the processes that occur on the surface of the studied catalysts and about the mechanism of the reduction of NO with CO. Fig. 7a and b presents the response curves of the reagents NO and CO and of the products N2O, N2, and CO2 at temperatures 80, 130, 170, and 250 °C. The figures are for Ce-promoted Al(Cu + Co + Ce) and for nonpromoted Al(Cu + Co). The differences in the curves reveal the different rate-controlling steps of the reaction over the catalysts. The change in the rate-controlling step is associated with a change in the reaction mechanism. The response curves for NO and CO for Al(Cu + Co) are of a monotonically growing type at 80 °C and of a monotonically decreasing type at 130 and 170 °C. The monotonically type

Concentration of NO, CO, CO2 , N2O and N2, ppm

The sequence of the metal oxide deposition in the ternary catalysts also influences the activity in NO reduction, as the simultaneous deposition of the three metals favors the appropriate disposition of metals and promotes their interaction. Fig. 6 shows the TPD spectra of NO for all samples. It is evident that all samples have desorption peaks for NO. The most intense one at low-temperature is the NO desorption of Al(Cu + Co + Ce)—the catalysts with the highest activity (Fig. 5). The mathematical analysis of the desorption curve establishes four temperatures of NO desorption—three clearly defined low-temperature desorption peaks at 80, 100, and 150 °C and one high-temperature desorption peak at 250 °C with a low intensity. These peaks represent probably four forms of NO adsorption. The TPD curves of the rest of the ternary-supported catalysts show only two desorption peaks with a low intensity at 150 and at 250 °C; i.e., only two NO adsorption forms are observed. The TPD spesctra of the binary-supported catalysts present three forms of NO adsorption at 80, 150, and 250 °C, where the most intense are for Al(Co + Ce) at 150 and 250 °C. Clearly defined are the low-temperature NO desorption peaks for the single supported AlCu, AlCo, and AlCe. All catalysts do not show desorption peaks for CO. The catalyst with the highest activity Al(Cu + Co + Ce) shows four desorption peaks, thus revealing four NO adsorption forms. Hence, the highest efficiency could be explained with the availability of these forms, which are deduced to correlate with the catalytically active sites. We suppose that the interaction between Cu–Ce and Co–Ce for the binary- and the ternary-supported samples leads to formation of new catalytically active sites responsible for NO reduction. The simultaneous deposition has advantages for the influence of cerium on copper and cobalt, as evident from textural data. These results show that cerium affects the metal disposition in the mesoporous space and hinders copper and cobalt to

o

o

o

130 C

80 C

o

170 C Ar

Ar NO+CO+Ar

NO+CO+Ar

250 C Ar

NO+CO+Ar

Ar

NO+CO+Ar

1200 CO CO

1000

NO

CO NO

800

NO

CO

600 NO CO2

400 CO2

CO2

200

N2

N2O

CO2

N2

120

90

60

30

CO2

N2

N2O

0

150

180

Time, min

(a) Concentration of NO, CO, CO2, N2O and N2, ppm

782

1200

o

o

o

80 C Ar

o

170 C

130 C

NO+CO+Ar

Ar

NO+CO+Ar

250 C Ar

NO+CO+Ar

NO+CO+Ar

Ar

1000 CO2

800 NO

CO2

600

CO

CO CO2

CO

CO CO2

400 200

CO2

NO NO N2

N2

N2O

N2

NO

N2

0 30

60

90

120

150

180

Time, min

(b) Fig. 7. Transient response curves of NO, CO, CO2, N2O and N2 for (a) – Al(Cu + Co), (b) – Al(Cu + Co + Ce).

783

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784 Table 3 TOF values at 150 °C. Sample

Conversion degree (%)

NO inlet (vol.%)

TOFCu104

TOFCo104

TOFCe104

Al(Cu + Co + Ce) AlCe(Cu + Co) Al(Cu + Co)Ce Al(Cu + Co) Al(Cu + Ce) Al(Co + Ce) AlCu AlCo AlCe

60 46 33 27 24 21 15 12 5

0.11 0.10 0.11 0.12 0.13 0.13 0.11 0.11 0.11

33.27 24.02 14.88 9.40 2.97 – 1.30 – –

21.96 10.75 8.67 5.51 – 2.81 – 1.17 –

19.96 13.02 9.88 – 6.57 4.54 – – 0.86

response, according to Kobayashi’s classification, indicates that the rate-limiting step in the reaction mechanism could be the surface reaction or desorption of the products. As desorption of the products is absent in the stop stage the surface reaction seems to be the rate-determining step. The responses of N2O and N2 are of the instantaneous type at 80 and 130 °C; N2O is not detected at 170 °C. The instantaneous response deduces that the rate-limiting step could be the surface reaction or the adsorption of the reagents. The response curves at 250 °C show that only the response for CO is of the monotonically growing type, while those for NO, CO2, and N2 are of the overshot type, revealing that the rate-limiting step is the creation or regeneration of the catalytically active sites. Thus, at higher temperatures the mechanism of the reaction proceeds in a different route. The response curves for NO and CO at 80 °C for Al(Cu + Co + Ce) are quite different—they are of the overshot type. Those of CO2, N2O, and N2 are of the monotonically growing type. These responses give an illustration of the promoting action of the ceria on the reaction at relatively low temperatures. A sharp increase of the catalyst’s activity toward NO at temperatures above 80 °C (Fig. 5) is observed. At 130 °C the response curves of the NO and CO2 are of the overshot type and those of CO and N2 are of the monotonically growing type. No N2O is detected; i.e., the reduction of NO proceeds to N2. Hence, at about 100 °C the mechanism of NO reduction with CO probably changes. The response curves for NO, CO, CO2, N2O, and N2 at 170 °C are of the instantaneous type. The absence of desorption curves for NO and CO at the stop stage shows that the rate-limiting step is the surface reaction. The response curves for NO and CO at 250 °C are of the overshot type and those of CO2 and N2 are of the monotonically growing type that determines the rate-limiting step as creation or regeneration of the new catalytically active sites. The response curves show that depending on the temperature the catalytic reaction proceeds on various types of active centers where the reduction is via different mechanisms. It is known, that there are three possible routes for the reaction NO + CO:

As noted above in the catalyst characterization section, Cu, Co, and Ce exist in various oxidation states in the ternary samples. Recently, some authors have stated that the Ce3+ + Cu2+ M Ce4+ + Cu+ redox couple plays an important role in the studied reaction [24,46] and this is the reason why the addition of Ce promotes the NO reduction by CO on copper–cerium-supported catalysts. In this case the existence of Cu+ and its redox action (Cu+ M Cu2+) are established to facilitate the NO adsorption and reduction by CO. In our case, cobalt exists on the surface of the ternarysupported samples in addition to copper. XPS and TEM data evidence the availability of Co2+ and Co3+ as well. We could propose an additional redox couple, involving cerium and cobalt ions (Ce3+ + Co3+ M Ce4+ + Co2+), the presence of which would facilitate the investigated reduction. The catalytic results show that cerium strongly promotes the activity of Cu and Co with the binary and ternary samples. Probably, these reversible interactions between copper–cerium and cobalt–cerium are responsible for the enhanced activity of the ternary-supported catalysts. To evaluate the activity of the Cu and Co ions depending on the cerium promotion, it is necessary to compare the TOF values for copper and cobalt of the investigated catalysts. The data for TOF (according to Eq. (1)) at 150 °C are presented in Table 3. The obtained TOF values of Co and Cu show that the ternary-supported catalysts are more efficient than the binary ones as their TOF values are much bigger than others. On the other hand, the converted quantities NO from the copper and cobalt active components in the ternary and binary-supported catalysts are several times higher than the same for the single supported samples. So, the efficiency of the noted catalysts is not simply additive to the same for the single oxides. This endorses our suggestion that the addition of cerium to copper and cobalt supported on alumina promotes the formation of new active centers for NO reduction with CO and our assumption for the reversible interactions between copper–cerium and cobalt–cerium as the reason for the enhanced activity of the ceria added catalysts.

2NO þ 2CO ! N2 þ 2CO2

The ceria-promoted alumina-supported copper and cobalt catalysts with metal loadings up to 4% total metal are effective in the reduction of NO with CO. The ternary-supported samples are more active than the binary ones, where the catalyst prepared by simultaneous impregnation with the three active metals is the most active in the temperature range. The addition of cerium in the binary systems affects the activity at temperatures over 130 °C, whereas the activity of the ternary-supported catalysts is enhanced even at room temperature. The higher activity of the ternary and the binary catalysts is determined by the favorable influence of the added cerium on the dispersion and the oxidation states of the copper and cobalt in the active phases. The presence of ceria contributes to the formation of appropriate active phases, resulting in catalytic sites on the surface of the samples that promote the reduction of NO with CO.

2NO þ CO ! N2 O þ CO2 NO ! N2 O þ 1=2O2

ð1Þ and N2 O þ CO ! N2 þ CO2

and N2 O ! N2 þ 1=2O2 :

ð2Þ ð3Þ

In the present experiments, N2O was detected up to 130 °C over the ternary-supported catalysts and up to 150 °C over the binary samples. When the temperature was higher than these temperatures, no N2O was measured in the reaction stream, which indicates that the route according to Eq. (1) is dominant at temperatures higher than 130 °C for the ternary and 150 °C for the binary catalysts. At low temperatures the reaction between NO and CO proceeds possibly via the routes according to Eqs. (2), (3) because of the amount of N2O detected.

4. Conclusion

784

I. Spassova et al. / Journal of Colloid and Interface Science 354 (2011) 777–784

Acknowledgments The authors are indebted to the National Science Fund of Bulgaria (Contract No. DTK 02/64/2009) for financial support. The authors acknowledge also the National Centre for New Materials UNION (Contract No. DCVP 02/2/2009). References [1] M. Shelef, Chem. Rev. 95 (1995) 209. [2] R. Burch, Catal. Rev. 46 (2004) 271. [3] R. Hernández-Huesca, J. Santamaría-González, P. Braos-García, P. MairelesTorres, E. Rodríguez-Castellón, A. Jiménez-López, Appl. Catal. B 29 (2001) 1. [4] H. Ohtsuka, Appl. Catal. B 33 (2001) 325. [5] Y. Li, J.N. Armor, J. Catal. 145 (1994) 1. [6] L.L. Ren, T. Zhang, D.B. Liang, C.H. Xu, J.W. Tang, L.W. Lin, Appl. Catal. B 35 (2002) 317. [7] P.W. Park, H.H. Kung, D.-W. Kim, M.C. Kung, J. Catal. 184 (1999) 440. [8] L.F. Liotta, G. Pataleo, A. Macaluso, G. Di Carlo, G. Deganello, Appl. Catal. A 245 (2003) 167. [9] V.I. Pârvulescu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233. [10] J.Y. Yan, M.C. Kung, W.M.H. Sachtler, H.H. Kung, J. Catal. 172 (1997) 178. [11] M.C. Kung, P.W. Park, D.-W. Kim, H.H. Kung, J. Catal. 181 (1999) 1. [12] J. Shibata, K. Shimizu, A. Satsuma, T. Hattori, Appl. Catal. B 37 (2002) 197. [13] E. Seker, N. Yasyerli, E. Gulari, C. Lambert, R.H. Hammerle, J. Catal. 208 (2002) 15. [14] L. Chen, T. Horiuchi, T. Osaki, T. Mori, Appl. Catal. B 23 (1999) 259. [15] T. Yamamoto, T. Tanaka, S. Suzuki, R. Kuma, K. Teramura, Y. Kou, T. Funabiki, S. Yoshida, Topics Catal. 18 (1-2) (2002) 113. [16] F. Zhang, S. Zhang, N. Guan, E. Schreier, M. Richter, R. Eckelt, R. Fricke, Appl. Catal. B 73 (2007) 209. [17] R. Burch, J.P. Breen, F.C. Meunier, Appl. Catal. B 39 (2002) 283. [18] B.M. Reddy, A. Khan, Y. Yamada, T. Kobayashi, S. Loridant, J.-C. Volta, J. Phys. Chem. B 107 (2003) 5162. [19] R.Q. Long, R.T. Yang, Appl. Catal. B 27 (2000) 87. [20] A. Trovarelli, Catal. Rev. Sci. Eng. 38 (1996) 439. [21] B.M. Reddy, P. Lakshmanan, A. Khan, J. Phys. Chem. B 108 (2004) 16855.

[22] Z. Wu, B. Jiang, Y. Liu, H. Wang, R. Jin, Environ. Sci. Technol. 41 (2007) 5812. [23] T. Tabakova, V. Idakiev, V.J. Papavasiliou, G. Avgouropoulos, T. Ioannides, Catal. Commun. 8 (2007) 101. [24] J. Chen, Y. Zhan, J. Zhu, C. Chen, X. Lin, Q. Zheng, Appl. Catal. A 377 (2010) 121. [25] C. Hardacre, R.M. Ormerod, R.M. Lambert, J. Phys. Chem. 98 (1994) 10901. [26] M. Khristova, B. Ivanov, I. Spassova, T. Spassov, Catal. Lett. 119 (1–2) (2007) 79. [27] Z. Lendzion-Bielun, M.M. Betahar, S. Monteverdi, D. Moszynski, U. Narkievicz, Catal. Lett. 134 (2010) 196. [28] J.-W. Park, J.-H. Jeong, W.-L. Yoon, H. Jung, H.-T. Lee, D.-K. Lee, Y.-K. Park, Y.-W. Rhee, Appl. Catal. A 274 (2004) 25. [29] J. Scofield, J. Electron. Spectrosc. Relat. Phenom. 8 (1976) 129. [30] M. Kobayashi, Chem. Eng. Sci. 37 (1982) 393. [31] T. Feltes, L. Espinosa-Alonso, E. Smit, L. D’Souza, R. Meyer, B. Weckhuysen, J. Regalbuto, J. Catal. 270 (1) (2010) 95. [32] C.R. Jung, J. Han, S.W. Nam, T.H. Lim, S.A. Hong, H.I. Lee, Catal. Today 93–95 (2004) 183. [33] G.R. Rao, H.R. Sahu, B.G. Mishra, Colloid Surf. A 220 (2003) 261. [34] G. Moretti, G. Fierro, M. Lo Jacono, P. Porta, Surf. Interface Anal. 14 (1989) 325. [35] Sh.-M. Zhang, W.-P. Huang, X.-H. Qiu, B.-Q. Li, X.-Ch. Zheng, Sh.-H. Wu, Catal. Lett. 80 (2002) 41. [36] M. Voß, D. Borgmann, G. Wedler, J. Catal. 212 (2002) 10. [37] G. Fierro, M. Lo Jacono, M. Inversi, R. Dragone, P. Porta, Topics Catal. 10 (2000) 39. [38] P. Schmitz, R. Usmen, C. Peters, J. Graam, R. McCabe, Appl. Surf. Sci. 72 (1993) 181. [39] M. Fernandez-Garcia, E. Gomez-Rebollo, A. Guerrero Ruiz, J.C. Conesa, J. Soria, J. Catal. 172 (1997) 146. [40] D. Mehandjiev, P. Dimitrova, React. Kinet. Catal. Lett. 56 (2) (1995) 341. [41] C. Tang, W. Yu, C. Wang, S. Chien, Catal. Lett. 116 (2007) 161. [42] H. Yao, Y. Yao, J. Catal. 86 (1984) 254. [43] E. Moretti, M. Lenarda, L. Storaro, A. Talon, T. Montanari, G. Busca, et al., Appl. Catal. A 335 (1) (2008) 46. [44] W. Liu, M. Flytzani-Stephanopoulos, Chem. Eng. J. 64 (1996) 283. [45] Y. Su, S. Wang, T. Zhang, S. Wang, B. Zhu, J. Cao, Z. Yuan, S. Wu, Catal. Lett. 124 (2008) 405. [46] A. Leontiou, A. Ladavos, P. Pomonis, Appl. Catal. A 241 (2003) 133.