Promotional effect of Ce on Cu-SAPO-34 monolith catalyst for selective catalytic reduction of NOx with ammonia

Promotional effect of Ce on Cu-SAPO-34 monolith catalyst for selective catalytic reduction of NOx with ammonia

Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311 Contents lists available at ScienceDirect Journal of Molecular Catalysis A: Chemical j...

2MB Sizes 0 Downloads 2 Views

Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

Contents lists available at ScienceDirect

Journal of Molecular Catalysis A: Chemical journal homepage: www.elsevier.com/locate/molcata

Promotional effect of Ce on Cu-SAPO-34 monolith catalyst for selective catalytic reduction of NOx with ammonia Yi Cao a , Sha Zou a , Li Lan a , Zhengzheng Yang b , Haidi Xu a , Tao Lin a,∗ , Maochu Gong a , Yaoqiang Chen a,b,c,∗ a

Key Laboratory of Green Chemistry and Technology of the Ministry of Education, College of Chemistry, Sichuan University, Chengdu 610064, PR China College of Architecture and Environment, Sichuan University, Chengdu 610064, Sichuan, PR China c Sichuan Province Engineering Center of Environmental Catalytic Materials, Chengdu 610064, PR China b

a r t i c l e

i n f o

Article history: Received 10 November 2014 Received in revised form 21 December 2014 Accepted 27 December 2014 Available online 30 December 2014 Keywords: Cu-SAPO-34 Hydrothermal stability Cerium NH3 -SCR Isolated Cu2+ ion

a b s t r a c t The activity and hydrothermal stability of Cu-SAPO-34 and CuCe-SAPO-34 for selective catalytic reduction of NOx with ammonia (NH3 -SCR) were investigated systematically. The catalysts were prepared by wet-impregnation method, and characterized by N2 adsorption, X-ray diffraction (XRD), X-ray photoelectron spectrum (XPS), Ultraviolet-visible diffuse reflectance spectrum (UV–vis-DRS), H2 -temperature programmed reduction (H2 -TPR) and NH3 -temperature programmed desorption (NH3 -TPD). The experimental results of fresh catalysts suggested that Ce mainly existed on the surface of the catalyst and was well dispersed in the form of Ce3+ , and its interaction with copper could improve the dispersion of copper species and increase the amount of isolated Cu2+ ions, so that CuCe-SAPO-34 performed better NH3 -SCR activity than Cu-SAPO-34. After hydrothermal aging at 800 ◦ C for 12 h, the characterization results indicated that the introduction of cerium effectively improved the textural and structural stability of SAPO-34 since more cations at the exchange site of SAPO-34 could decrease the concentration of Si O(H) Al bond which was closely related to the damage of the SAPO-34 framework. Moreover, the addition of Ce could prevent the decrease of acid densities, promote the redistribution of CuO during hydrothermal aging and provide higher amount of isolated Cu2+ ions, leading to superior hydrothermal stability of CuCe-SAPO-34 catalyst. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Diesel engine is attractive because of its low fuel cost and low CO2 emission [1]. However, nitrogen oxides (NOx ) emitted from the diesel engine are the major air pollutants, as they may cause severe environmental issues, such as acid rain, photochemical smog, ozone depletion and haze events [2–4]. NH3 -SCR is one of the most effective methods to remove NOx exhausts [5,6]. Although mature commercial V2 O5 /WO3 –TiO2 catalyst has been used for several decades, it is not applicable in the after-treatment system equipped with diesel particulate filter (DPF). The regeneration temperature of DPF will reach above 700 ◦ C [7,8], while commercial V2 O5 /WO3 –TiO2 catalyst performs poor hydrothermal stability,

∗ Corresponding authors at: Key Laboratory of Green Chemistry and Technology of the Ministry of Education, College of Chemistry, Sichuan University, Chengdu, 610064, PR China. Tel.: +86 28 85418451; fax: +86 28 85418451. E-mail addresses: [email protected] (T. Lin), [email protected] (Y. Chen). http://dx.doi.org/10.1016/j.molcata.2014.12.020 1381-1169/© 2015 Elsevier B.V. All rights reserved.

narrow reaction window, moreover, V2 O5 has toxicity for the environment [9,10]. Thus, new catalysts are in demand to solve these problems, Cu-ZSM5 and Cu-beta catalysts are employed as their activity, selectively and hydrothermal stability was better than that of V2 O5 /WO3 –TiO2 catalyst. However, their hydrothermal stability is still too poor to integrate with DPF [11]. Therefore, a SCR catalyst with better hydrothermal resistance should be developed. Recently, Cu-SSZ-13 and Cu-SAPO-34 have received much attention due to their excellent NH3 -SCR activity, N2 selectivity and hydrothermal stability [11–13]. Until now, many researchers are focused on the active site of Cu-CHA for NH3 -SCR reaction, Fickel and Lobo [14] found that the isolated copper species located in the cages coordinated to three oxygen atoms of the six-membered rings were the active sites of Cu-SSZ-13. Furthermore, Wang et al. [15] reported that the isolated copper species at the exchange sites were the active sites. In addition, Xue et al. [16] proposed that the Cu2+ species associated with the six-ring window and displaced into the ellipsoidal cavity of SAPO-34 were the active sites. Moreover, the

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

presence of two distinguishable Cu2+ sites in the framework of SSZ13 was demonstrated by Kwak et al. [17], on the one hand, Kim et al. [18] proposed that both isolated Cu2+ ions in CHA cage and in D6R site were the active sites, and the Cu2+ ions in D6R site were slightly more active than those in CHA cage; on the other hand, Szanyi et al. [19] suggested that the isolated copper species interacted with strongly adsorbed molecules (H2 O or NH3 ) could facilely migrate from D6R site to CHA cage, so it has been widely accepted that the isolated Cu2+ ions of Cu-CHA were the active sites for NH3 -SCR reaction. Meanwhile, the changes of Cu-CHA during hydrothermal aging process were explored by many groups. Wang et al. [20] and Vennestrøm et al. [21] reported that CuO could redistribute into the pores of SAPO-34 to generate more isolated copper ions. Kwak et al. [11] and Schmieg et al. [7] proposed that for zeolite catalyst, dealumination, collapse of structure and loss of active sites during hydrothermal led to decrease NH3 -SCR activity. Kim et al. [18] demonstrated that during hydrothermal aging process, the unstable copper species in CHA cage facilely agglomerated to CuO cluster, and the growth of CuO cluster could eventually destroy the crystal structure of CHA zeolite. Many groups have proposed that Cu-SAPO-34 catalyst was stable enough to encounter the 800 ◦ C hydrothermal treatment. However, the textural and structural properties of SAPO-34 significantly decreased, and the maximum NO conversion of Cu-SAPO-34 was lower than 90% after hydrothermal aging [22–24]. Thus, it is necessary to further improve the hydrothermal stability of Cu-SAPO-34. The Ce additive used to improve the hydrothermal stability of catalyst has been extensively investigated. For instance, Yan et al. [25] and Iwasaki and Shinjoh [26] proposed that the addition of Ce could improve the hydrothermal stability of ZSM-5 and beta, moreover, Zhang and Flytzani-Stephanopoulos [27] and Pang et al. [28] demonstrated Ce addition improve the activity and hydrothermal stability of Cu-ZSM5. In addition, Gao et al. [2] proposed that introduction of appropriate second cations would promote the activity of Cu-CHA catalysts. Herein, in this study, the effect of the addition of Ce on NH3 -SCR reaction and hydrothermal stability of Cu-SAPO-34 were investigated. In addition, the N2 adsorption, XRD and NH3 -TPD measurements were applied to determine the effect of cerium on the textural, structural properties and acid densities of the SAPO-34, respectively. Furthermore, the interaction between copper and cerium was demonstrated by XPS, and the copper species in the bulk of catalysts were measured by UV–vis-DRS and H2 -TPR.

2. Experimental 2.1. Catalyst preparation Cu-SAPO-34 was prepared by a wet impregnation method, firstly, 1.325 g of Cu(NO3 )2 ·4H2 O (99%, Kelong) was dissolved into 5.2 ml of deionized water, then commercial NH4 -SAPO-34 (NanKai University, Si/(Si + Al + P) = 0.1) was dropped into the obtained Cu(NO3 )2 solution, and the paste was stirred for 1 h to ensure that Cu was dispersed well. The as-prepared Cu-SAPO-34 was dried at 100 ◦ C for 6 h and calcined at 550 ◦ C for 3 h. And the H-SAPO-34 (HSA) was obtained by calcining the NH4 -SAPO-34 at 550 ◦ C for 3 h. CuCe-SAPO-34 was prepared by two-step method. Firstly, 100 g of commercial NH4 -SAPO-34 was ion-exchanged with 1 l of 0.1 M Ce(NO3 )3 (99%, Jinshan) solution at 80 ◦ C for 6 h. It was then filtered and washed with distilled water. After that the paste was dried at 100 ◦ C for 12 h to obtain Ce-SAPO-34 (CeSA). Secondly, the Cu(NO3 )2 solution was impregnated on CeSA using the same method as above-mentioned. Finally the prepared catalyst powders were coated on honeycomb cordierites (400 cell per inch2 ,

305

Yixing Jiangsu) as described in our previous studies [29,30], so that the monolith catalysts with a loading of 120 g/l could be obtained. To investigate their hydrothermal stability, the monolith catalysts were hydrothermal aged at 800 ◦ C for 12 h with 7 vol% H2 O/air at gas hourly space velocity (GSHV) of 30,000 h−1 . The fresh Cu-SAPO34 and CuCe-SAPO-34 were denoted as Cu and CuCe, respectively, and the corresponding aged samples were denoted as Cu-800 and CuCe-800, respectively. 2.2. Steady-state reactor experiments The catalytic activity measurement was carried out in a fixed bed quartz flow reactor. Reactant gases were regulated by means of mass-flow controllers before entering the reactor. The concentrations of the simulated gases were as follows: 1000 ppm NO, 1000 ppm NH3 , 5% O2 , 10 vol% H2 O, balanced with N2 , the gas hourly space velocity (GHSV) was 30,000 h−1 , the total flow rate was 1250 ml min−1 . The concentrations of NOx (NO and NO2 ), N2 O and NH3 in the inlet and outlet gases were continually analyzed by a FT-IR (Antaris IGS, Nicolet, Thermo Fisher Scientific). The catalysts were pretreated in the simulated gases at 550 ◦ C for 2 h before the activity test. Catalytic activity tests were carried out over the temperature range of 150–550 ◦ C. The NO conversion is calculated as the following formula: NO(

conversion ([NO]in − [NO]out ) × 100% )= % [NO]in

2.3. Catalyst characterization A QUADRASORB SI automatic surface analyzer system (Quantachrome Corporation) was used to measure the N2 adsorption isotherms of the samples at the temperature of liquid N2 (−196 ◦ C). The specific surface areas were determined from the linear portion of the BET plot. Prior to adsorption N2 , the samples were evacuated at 300 ◦ C for 3 h. The chemical content of each sample was determined by inductively coupled plasma atomic emission spectrometry (ICP-AES) with an IRIS Intrepid II XSP apparatus (Thermo Fisher Scientific Inc.). XRD data were collected on Rigaku D/MAX-rA diffractometer with Cu radiation ( = 0.15406 nm). The X-ray tube was operated at 45 kV and 25 mA. The samples were investigated in the 2 range of 5–50◦ at a scanning speed of 0.02◦ S−1 . XPS experiments were carried out on a spectrometer (XSAM800, KRATOS Co.) with Al K␣ radiation under ultra-high vacuum (UHV). The C 1 s peak (284.5 eV) was used for the calibration of binding energy values. The pressure in the analytical chamber was about 10−9 Pa. UV–vis-DRS were taken at room temperature on a “Shimadzu” UV-3600 PC spectrophotometer equipped with a diffuse reflectance accessory, and barium sulfate was used in background scans. The test sample was the mixture of 250 mg catalyst and 1000 mg BaSO4 . Spectra were collected from 200 to 1100 nm. The H2 -TPR was carried out on a commercial instrument with a TCD detector (Xianquan, TP5076), and 100 mg sample was applied for each measurement. The sample was first pretreated in N2 (30 ml min−1 ) at 400 ◦ C for 1 h to remove the H2 O and other impurities adsorbed on the surface of sample. Then, the sample was cooled to 50 ◦ C in N2 flow, the H2 -TPR was carried out at a linear heating rate of 10 ◦ C/min from 50 ◦ C to 1000 ◦ C under 5 vol% H2 /N2 (30 ml/min) flow. NH3 -TPD was carried out on TP5076, and 100 mg sample was applied for each measurement. The sample was first pretreated in He (30 ml min−1 ) at 400 ◦ C for 1 h to remove the H2 O and other impurities adsorbed on the surface of sample. Then, the sample was cooled to 120 ◦ C, the catalyst was exposed in NH3 (2% NH3 /N2 ,

306

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

Fig. 1. The NO conversion and N2 O generation of Cu and CuCe catalysts before and after hydrothermal aging. Feed condition: 1000 ppm NH3 , 1000 ppm NO, 5% O2 , 10 vol% H2 O and N2 balance, GSHV: 30,000 h−1 ; hydrothermal aging condition: 7 vol% H2 O and air balance, GSHV: 30,000 h−1 at 800 ◦ C for 12 h.

Fig. 2. XRD patterns of the Cu, CuCe, Cu-800 and CuCe-800 catalysts.

3.2. Textural and structural properties 30 ml min−1 ) for 30 min at 120 ◦ C, after that the catalyst was purged by He (30 ml min−1 ) for 60 min to remove physical adsorbed NH3 . Finally NH3 -TPD was performed at a linear heating rate of 8 ◦ C/min from 120 to 650 ◦ C under He flow (30 ml/min).

3. Results and discussion 3.1. NH3 -SCR activity test The NO conversions of catalysts before and after hydrothermal treatment for standard SCR reaction were shown in Fig. 1. For the fresh catalysts, CuCe catalyst displayed better performance than Cu catalyst over the whole temperature range. In addition, after hydrothermal treatment at 800 ◦ C for 12 h, the activity of CuCe800 showed only slight decrease and its maximum NO conversion still maintained above 95%. Compared with CuCe-800 catalysts, the activity of Cu-800 decreased obviously with the maximum NO conversion decreased by ca. 15%. The activities of the support materials were shown in Fig. S1, and the results revealed that Ce in SAPO-34 was not active site for NH3 -SCR reaction, and the addition of Ce had little influence on the NH3 -SCR activity of the pure support. N2 O was not only a greenhouse gas which is much more serious than CO2 , but also an ozone depletion catalyst, so the generation of N2 O was one of important aspects in practical use to evaluate the performance of SCR catalyst and the results were depicted in Fig. 1. The concentrations of byproduct N2 O in outlet were very low for the fresh catalysts (<5 ppm), and they were still lower than 10 ppm even after hydrothermal aging. In addition, CuCe-800 yielded less N2 O than Cu catalyst. The results above inferred that the addition of Ce could improve the activity and hydrothermal stability of Cu catalyst.

The BET surface areas and pore structure results of the support materials and catalysts were summarized in Table 1. The surface areas and pore volumes of SAPO-34 decreased slightly after Ce ion exchanged or Cu impregnated, suggesting that the introduction of Ce or Cu did not obviously affect the textural properties of the support. However, hydrothermal treatment had a tremendous impact on the BET surface areas and pore structure of these catalysts, as shown in Table 1, BET surface areas and pore volumes of Cu-800 and CuCe-800 decreased evidently. Compared to the BET surface areas (523.7 m2 /g) and pore volumes (0.26 cm3 /g) of Cu catalyst, both surface areas and pore volumes of Cu-800 catalyst decreased by 26%, while those of CuCe-800 decreased by only 19%, revealing that the addition of Ce had promoting effect on the textural stability of SAPO-34. To further confirm the effect of hydrothermal aging on the phase of Cu species and the structure of SAPO-34 of Cu and CuCe catalysts, XRD measurement was used and the XRD patterns were depicted in Fig. 2. All the patterns in Fig. 2 showed diffraction peaks typical of SAPO-34 structure. In addition, for the fresh Cu and CuCe catalysts, two diffraction peaks located at 35.58◦ and 38.88◦ which were assigned to CuO phase (PDF, 48-1548) could be observed. The average crystallite sizes of CuO of Cu and CuCe catalysts, based on the Sherrer equation, were calculated to be 37 nm and 27 nm, respectively, also the intensities of CuO peak of Cu spectrum was higher than those of CuCe spectrum. Therefore, it could be concluded that CuO existed in both Cu and CuCe catalysts, and copper species in CuCe catalyst dispersed better than that in Cu catalyst. CuO particle would catalyze the reaction of NH3 with O2 and even generate NOx [31,32], so that there was not enough NH3 react with NO, thus, making the CuCe (with less CuO) perform better high temperature NO activity than Cu catalyst. After hydrothermal treatment at 800 ◦ C, the peaks of CuO phase disappeared, and the ICP results listed in Table 2 showed no decline of the contents of

Table 1 BET surface areas, pore volumes, acid amounts and acid densities of Cu and CuCe catalysts before and after hydrothermal aging sample

BET surface area (m2 /g)

Pore volume (ml/g)

Adsorbed NH3 amounts (mmol/g)

Acid densities (␮mol/m2 )

HSA CeSA Cu CuCe Cu-800 CuCe-800

554.6 532.2 523.7 515.2 386.7 417.0

0.27 0.26 0.26 0.25 0.18 0.21

0.68 0.72 0.83 0.77 0.39 0.54

1.2 1.4 1.6 1.5 1.0 1.3

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

307

Table 2 Surface composition and chemical composition of Cu and CuCe catalysts before and after hydrothermal aging Concentration of surface elements (wt%)a

Cu Cu-800 CuCe CuCe-800 a

Concentration of bulk elements (wt%)

Cu

Ce

Al

P

Si

IS /IM

Cu

Ce

Al

P

Si

Cus /Cub

5.3 3.4 4.4 3.8

– – 24 19

14 17 9.3 13

10 14 11 13

15 9.2 6.7 4.9

0.70 0.78 0.76 0.70

3.1 3.1 3.0 3.1

– – 1.2 1.1

19.9 19.1 20.6 19.4

17.1 17.8 19.2 17.5

3.5 3.9 3.8 3.4

1.6 1.0 1.3 1.1

Calculated by the XPS results.

copper, inferring that CuO migrated into the pores of SAPO-34 or the particle size of CuO became too small to be detected by XRD measurement. However, the high temperature NO conversion of aged catalysts declined, maybe some other changes occurred during hydrothermal treatment reduces the activity. The intensities of the diffraction peaks of SAPO-34 of both Cu and CuCe catalysts decreased after hydrothermal aging as shown in Fig. 2. Furthermore, the diffraction peaks of AlPO4 (PDF, 31–0028) at 21.48◦ and 35.38◦ appeared, inferring that hydrothermal treatment led to significantly decrease crystallinity of SAPO-34. Moreover, the intensities of the AlPO4 diffraction peaks of Cu-800 spectrum were obviously higher than those of CuCe-800 spectrum, suggesting that more AlPO4 impurities generated over Cu-800 than those formed over CuCe catalyst during hydrothermal aging. In conclusion, the BET and XRD results indicated that the textural and structural stability of Cu-SAPO-34 was improved evidently by the introduction of Ce. 3.3. Surface composition and the oxidation state of Cu and Ce XPS measurements were used to examine the effect of the hydrothermal aging on chemical state of surface Ce and Cu. The Ce 3d spectra of fresh and aged CuCe catalysts were shown in Fig 3. Two sets of spin–orbit split peak of Ce 3d were labeled as u and v which was corresponding to 3d3/2 and 3d5/2 , respectively. The u3 /v3 doublet was caused by photoemission from Ce4+ , and the u1 /v1 and u2 /v2 doublets were the shakeup peaks of Ce4+ which was due to electrons of filled O 2p orbital transfer to the empty Ce 4f orbital. The u0 /v0 and u0  /v0  doublets were due to photoelectron emission from Ce3+ ion [33–38]. It should [4] be noted that the Ce 3d spectrum of CuCe catalyst in Fig. 3 showed only four peaks at around 885.8 (v0 ), 882.3 (v0  ), 904.6 (u0 ) and 901.4 (u0  ) eV, respectively, suggesting that Ce3+ existed on SAPO-34 surface. In addition, Reddy

Fig. 3. XPS of Ce 3d in CuCe and CuCe-800 catalysts.

et al. [39] and Yuliati et al. [40] proposed that with low Ce contents the Ce species were highly dispersed and interacted with the support wherein most of the cerium is in +3 oxidation state. Thus, the Ce3+ in the catalyst was highly dispersed. After hydrothermal aging, the Ce 3d spectrum showed ten peaks at 885.8 (v0 ), 882.3 (v0  ), 898.3 (v1 ), 888.3 (v2 ), 883.0 (v3 ), 904.6 (u0 ), 901.4 (u0  ), 916.7(u1 ), 907.3 (u2 ), and 901.1 (u3 ) eV, respectively. Therefore, Ce3+ and Ce4+ were co-existed in the CuCe-800 catalyst. Furthermore, the relative content of Ce4+ /(Ce3+ + Ce4+ ) of CuCe-800 catalyst was calculated by the areas of the Ce4+ divided by the areas of the total Ce 3d spectrum [34,41], the result showed that 26.5% Ce4+ generated due to Ce gathering during hydrothermal process. Fig. 4 presented the Cu 2p XPS spectra of Cu and CuCe catalysts before and after hydrothermal aging. The Cu 2p spectrum showed two sets of peaks, corresponding to Cu 2p3/2 and Cu 2p1/2 . The Cu2+ species displayed two shake-up satellite peaks at 942.2 and 961.4 eV, associated with the main peaks of Cu 2p3/2 located at around 933.8 eV and Cu 2p1/2 located at around 953.4 eV, respectively [38]. Because the shake-up satellite peaks do not exist in Cu◦ and Cu+ XPS spectra, this feature has been used to distinguish between Cu2+ and Cu+ or Cu◦ species. In addition, the relative intensity of the satellite peak to main peak (IS /IM ) of Cu 2p3/2 for pure tenorite (CuO) was 0.47 [42]. However, the IS /IM of Cu 2p3/2 of Cu, Cu Ce, Cu-800 and CuCe-800 was 0.70, 0.76, 0.78 and 0.70, respectively, which was significantly higher than that of CuO. Christensen and Langell [42] reported that increased Cu to O bond length could lead to higher value of IS /IM . Accordingly, it could be deduced that the surface copper species were mainly Cu2+ and the interaction between Cu2+ ion and SAPO-34 may result in increased Cu to O bond length, consequently higher IS /IM values were obtained. It should be noted that the Cu 2p3/2 binding energy of CuCe catalyst at 933.56 eV was lower than that of Cu catalyst at 933.82 eV. Copper core XPS peaks shift to higher binding energies due to the decrease of Cu to O bonding distance [42]. In addition, Bera et al. [43] reported that the Cu could interact with Ce to generated the Cu O Ce species, and the distance of Cu O bond of CuO (1.959 Å) was slightly shorter than that of Cu O Ce species (1.962 Å). Therefore, the generation of Cu2+ O Ce3+ species would lead to the shift of binding energy of Cu 2p3/2 to lower value. After hydrothermal aging due to redistribution of CuO [21] the Cu 2p3/2 peak of Cu-800 and CuCe-800 shifted to higher binding energy at 934.01 eV. XPS measurements also could be used to quantify the surface elements composition, and the results were listed in Table 2. For the fresh catalysts, the surface Cu concentration of CuCe (4.4 wt%) was much lower than that of Cu (5.3 wt%), three presumptions could be pointed out: (1) copper content of Cu catalyst was higher than that of CuCe catalyst, (2) the distribution of copper species in Cu catalyst was better than that in CuCe catalyst or (3) more copper species existed on the surface of Cu catalyst compared with CuCe catalyst. Meanwhile, the Cu loading of each catalyst which was tested by ICP method was analogous (3.1 wt%), so that the first presumption could be ruled out. Moreover, the XRD results showed that the particle size of CuO of CuCe was smaller than that of Cu catalyst so that the second presumption could also be eliminated. Therefore, it could be speculated that the amount of copper species existed

308

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

Fig. 5. UV–vis-DRS of Cu, CuCe, Cu-800 and CuCe-800 catalysts.

ion could replace the proton of the surface Si O(H) Al bond to decrease the concentration of Si O(H) Al bond. Furthermore, the amount of Cu2+ ion of CuCe catalyst was higher than that of Cu catalyst in the exchange site of SAPO-34. Therefore, the addition of Ce could improve the structural stability of SAPO-34.

3.4. Nuclearity of the Cu species

Fig. 4. XPS of Cu 2p in Cu and CuCe catalysts before (a) and after (b) hydrothermal aging.

on the surface of Cu catalyst was larger than that of CuCe catalyst. Combined the third presumption with the content of Cu obtained from ICP results, it could be deduced that more copper species in the pore of Cu-Ce-SAPO-34. After hydrothermal treatment, the surface Cu concentration of Cu-800 and CuCe-800 decreased, and Cus /Cub of the aged catalysts was close to 1, these results were consistent with Vennestrøm et al. report [21], and they confirmed that copper species could redistributed into the bulk of SAPO-34 under high temperature treatment. It should be noted that the Ce concentration (23.9 wt%) on the surface was much higher than that (1.2 wt%) in the bulk, Kooten et al. [44] considered this might be due to the large radius of the hydrated Ce3+ cation (estimated to be 5.8 Å), compared to the channels of SAPO-34 (3.8 × 3.8 Å channels), so that it was too large to migrate into the pores of SAPO-34, resulting in Ce mainly existed on the surface of SAPO-34. According to Gao ea al. report [45], irreversible hydrolysis of the Si O(H) Al bond would cause the damage of the textural and structural properties of the SAPO-34. This phenomenon was also happened on ZSM-5 support as demonstrated by Brandenberger et al., and the Al site bearing a cation was more stable [46]. Moreover, it was proposed that the stability of the zeolite was improved by decreasing the concentration of Si O(H) Al bond [25]. During Ce3+ ion exchange procedure, Ce

UV–vis-DRS spectroscopy was used to identify the nuclearity of the Cu species in the catalysts. The spectra were displayed in Fig. 5. For the fresh Cu and CuCe catalysts, a sharp absorption band at 224 nm and a broad band between 300 and 850 nm could be observed. According to the literatures, the sharp band at 224 nm was attributed to O → Cu charge transfer of isolated Cu+ /Cu2+ species [5,20], two bands at 360 and 457 nm were assigned to O Cu O and Cu O Cu complex [20,47], respectively, and a stepshape band at 800 nm was corresponded to CuO species [48]. The band at 265 nm corresponded to the charge transfer from oxygen to Ce3+ [40,49] could not be identified clearly due to the overlapping of the band of Ce3+ and that of isolated Cu+ /Cu2+ . As the literatures reported [40], the band of Ce4+ was at around 420 nm. Since the band at 457 nm which was attributed to the Cu O Cu was abroad band and its position was close to that of the band of Ce4+ at 420 nm, and it may be influenced by Ce4+ . However, the Ce 3d XPS results showed that little Ce4+ existed in the CuCe catalyst, so that the influence of Ce4+ on the band of Cu O Cu could be eliminated. It should be noted that the intensities of the bands at 224, 360, and 457 nm in CuCe spectrum were stronger than those in Cu spectrum, indicating that more isolated Cu2+ ions, O Cu O and Cu O Cu complex existed in CuCe catalyst than that in Cu catalyst. Thus, a conclusion consistent with the XPS results could be drawn that Ce addition could improve the dispersion of copper species, so that the activity performance of CuCe was better than that of Cu catalyst. After hydrothermal aging, the spectra of Cu-800 and CuCe800 catalysts displayed two bands at 224 and 800 nm. The broad band at 800 nm was attributed to d–d transitions of the isolated hydrate Cu2+ species[5]. Compared with the spectra of fresh catalysts, the bands at 360 and 457 nm disappeared, and the step-shape band at 800 nm was replaced by a broad band which corresponded to isolated hydrate Cu2+ ion, meanwhile, the intensity of isolated Cu+ /Cu2+ ion band at 224 nm was strengthened, revealing that the O Cu O, Cu O Cu and CuO species transferred to isolated copper ions. These results were consistent with Wang ea al. report [20] that

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

309

Fig. 6. H2 -TPR profiles of the Cu, CuCe, Cu-800 and CuCe-800 catalysts.

CuO species could migrate into the pores of SAPO-34 to generate more isolated copper species. 3.5. H2 -TPR H2 -TPR was used to identify the copper species in catalysts before and after hydrothermal aging. The H2 -TPR results of fresh and aged Cu and CuCe catalysts were depicted in Fig. 6. For the fresh catalysts, a sharp peak (200–350 ◦ C) and a broad peak (350–800 ◦ C) could be found in the H2 -TPR profiles. The sharp peak could be de-convoluted into three peaks (A, B and C) due to its asymmetry. According to the literatures, peak A at around 250 ◦ C was assigned to the reduction of Cu2+ in CHA cage (Cu2+ to Cu+ ) [17,18], peak B at around 288 ◦ C was assigned to the reduction of CuO to Cu◦ [16,50], peak C at around 325 ◦ C was assigned to the reduction of Cu2+ on D6R in the sodalite cage to Cu+ [18], the broad peak D was assigned to the reduction of Cu+ to Cu◦ [15]. And the quantitative calculation results of hydrogen consumption were shown in Table 3. The sum of the hydrogen consumption of peak A and peak C (A + C), which represents the amount of isolated Cu2+ was higher for CuCe catalyst than that of Cu catalyst, indicating the larger amount of isolated Cu2+ in CuCe. After hydrothermal aging the amount of isolated Cu2+ increased for both Cu and CuCe catalysts, and the value of isolated Cu2+ of CuCe-800 was higher than that of Cu-800 catalyst as well. In addition, the reduction peak of CuO for Cu catalyst was much larger than that for CuCe catalyst and they both declined after hydrothermal aging, these results were consistent with the XPS and UV–vis-DRS results, and the reduction peak of CuO for Cu-800 was still larger than that for CuCe-800, suggesting that Ce modified the migration of CuO species. However, the H2 consumption of each catalyst was lower than the theoretical value (0.485 mmol/g) which was calculated assuming that all the Cu2+ of the sample were reduced into Cu◦ metal, since some isolated copper species were too stable to be reduced to Cu◦ below 800 ◦ C [15] or the Cu2+ underwent self-reduction during pretreatment under inert N2 flow at 400 ◦ C [51,52]. The reduction of isolated Cu2+ in SAPO-34 was proposed by a two-step mechanism [16]: Cu2+ to Cu+ (A + C) and Cu+ to Cu◦ (D). Thus, if the copper species of the pretreated sample were all Cu2+ , then (A + C) > D meant that the reduction amount of isolated Cu2+ to Cu+ was larger than that of Cu+ to Cu◦ , that is, partial Cu+ cannot be reduced to Cu◦ and when (A + C) = D, it meant that the reduction amount of isolated Cu2+ to Cu+ was equal to that of Cu+ to Cu◦ , in other words, all Cu+ was reduced to Cu◦ . On the contrary,

Fig. 7. NH3 -TPD profiles of support materials and catalysts before and after hydrothermal aging.

it could be calculated from Table 3 that (A + C) < D for all catalysts, so that self-reduction of Cu2+ occurred in all catalysts. According to the Cu 2p XPS results, all copper spices were Cu2+ , therefore, in order to avoid the influence of self-reduction of Cu2+ on calculating the real H2 -consumption, the 2 × D was used to replace the (A + C + D) and (2 × D + B) was employed to estimate the real H2 consumption without the influence of self-reduction. If all the Cu+ species were reduced to Cu◦ , (2 × D + B) should be equal to the theoretical value. However, the (2 × D + B) < 0.485 mmol/g for the fresh catalysts, while the (2 × D + B) ∼0.485 mmol/g for the aged catalysts, inferring that partial Cu+ could not be reduced to Cu◦ only happened in the fresh catalysts. In addition, the H2 consumption of peak D of both Cu-800 and CuCe-800 increased after hydrothermal aging, which is probably due to the weakened interaction between the Cu+ ion and SAPO-34 as a result of the partial destruction of zeolite structure [18]. 3.6. Acid densities NH3 -TPD was usually used to determine the acid strength and acid distribution. Fig. 7 showed the NH3 -TPD profiles of HSA and CeSA support materials, fresh Cu and CuCe catalysts and aged Cu and CuCe catalyst. The NH3 -TPD profiles of the support materials contain two distinguishable regions which could be de-convoluted into four peaks located at ca. 195 (A), 280 (B), 386 (C) and 451 ◦ C (D), respectively. According to the literatures, the peak (A) at lower temperature was assigned to the weak Brønsted acid sites at surface hydroxyls [53,54], and the peaks (B, C and D) at higher temperature were related to the structural Brønsted acid sites inferred to moderate and strong acidity [53–55]. Two new peaks located at 355 and 530 ◦ C were observed in the NH3 -TPD profiles of Cu and CuCe catalysts. The peak at 355 ◦ C (C) was ascribed to NH3 desorbed form the moderate structural Brønsted acid sites and overlapped with NH3 desorbed from the acid sites created by CuO fine particle and/or Cu O Cu cluster interact with SAPO-34 (more details were shown in supporting information), and the peak at 530 ◦ C (E) was ascribed to the NH3 desorbed from the strong Lewis sites created by copper species. After hydrothermal aging at 800 ◦ C, the distinguishable peaks C and D became weaker and transformed to broad peaks that could not be clearly identified, suggesting that the amount of structural Brønsted acid sites declined due to the partial collapse

310

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311

Table 3 Quantitative analysis of the TPR profiles of Cu and CuCe catalysts before and after hydrothermal aging sample

A (mmol/g)

B (mmol/g)

C (mmol/g)

D (mmol/g)

Total (mmol/g)

Cu CuCe Cu-800 CuCe-800

0.058 0.073 0.056 0.060

0.207 0.113 0.019 0.008

0.025 0.030 0.071 0.071

0.089 0.130 0.240 0.249

0.379 0.346 0.386 0.388

of SAPO-34 structure which was confirmed by XRD measurements, and the acid sites which was created by the interaction between CuO fine particles and SAPO-34 disappeared due to redistribution of CuO particles. Wang et al. [55] proposed that the nitrite and/or nitrate species formed on the active sites of the Cu-SAPO-34 and the NH3 adsorbed on the acid sites (Brønsted acid sites) could migrate to the active sites (Lewis acid sites) to generate NH4 NO2 and NH4 NO3 . NH4 NO2 could directly decompose to N2 and H2 O, and NH4 NO3 could react with NO to form NH4 NO2 . Thus, the acid properties of the Cu-SAPO34 catalysts were very important for NH3 -SCR reaction. In addition, Yu et al. [54] found that acid densities were related to NO conversion at low temperature and NH3 oxidation at high temperature. The catalyst with higher acid density performed better NO conversion activity at low temperature and inhibited the NH3 oxidation at high temperature. The acid amounts of each catalyst were estimated by integrating the NH3-TPD spectrum, and the acid density was the ratio of the sum of the acid amount to the surface areas for each sample, and the results were shown in Table 1. For the fresh catalysts, the acid density of Cu was analogous to that of CuCe, so that the better activity of CuCe catalyst could be ascribed to more isolated Cu2+ ions in CuCe catalyst. In addition, the acid density of CuCe-800 was slightly higher than that of Cu-800, combined with the UV–vis-DRS and H2 -TPR results, it was reasonable that CuCe-800 performed better NO conversion than Cu-800 catalyst. However, the Cu-800 and CuCe-800 with more isolated Cu2+ ions performed poorer activities than Cu and CuCe catalyst, respectively. For Cu catalyst, it should be noted that the acid density of aged catalyst decreased by nearly 40%, while the isolated Cu2+ only increased about by 20% which was calculated from the H2 -TPR results, meanwhile, the maximum NO conversion of Cu-800 decreased by 15%; and for CuCe catalyst, the acid densities of aged catalyst decreased by nearly 14%, while the isolated Cu2+ increased about 15%, and the maximum NO conversion of CuCe-800 decreased by 5%. So the decreases of acid densities influenced the NH3 -SCR activity to a larger extent compared with the increases of isolated Cu2+ ions after hydrothermal aging.

4. Conclusion The behaviors of fresh and aged Cu and CuCe catalysts for NH3 SCR were examined. The characterization results showed that the cerium highly dispersed on the surface of SAPO-34 interacted with copper species could improve the dispersion of copper species and increase the amount of isolated Cu2+ ions, so that CuCe catalyst performed better NH3 -SCR activity than Cu catalyst. After hydrothermal aging, the characterization results showed that catalyst with Ce addition held better textural properties, generated less AlPO4 , kept higher acid densities and formed larger amount of isolated Cu2+ , consequently enhanced catalytic performance was achieved for CuCe-800 catalyst. Moreover, the maximum NO conversion of CuCe-800 was higher than 95%, while the maximum NO conversion of Cu-800 was only 85%. Thus, it could be concluded that the introduction of Ce could further improve the activity and hydrothermal resistance of the Cu catalysts.

Acknowledgements This work was financially supported by the National Natural Science Foundation of China (Grant No. 21173153), the National High-Tech Research and Development (863) Program of China (Grant No. 2013AA065304). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.molcata. 2014.12.020. References [1] P. Granger, V.I. Parvulescu, Chem. Rev. 111 (2011) 3155–3207. [2] F. Gao, J. Kwak, J. Szanyi, C.F. Peden, Top. Catal. 56 (2013) 1441–1459. [3] R.-J. Huang, Y. Zhang, C. Bozzetti, K.-F. Ho, J.-J. Cao, Y. Han, K.R. Daellenbach, J.G. Slowik, S.M. Platt, F. Canonaco, P. Zotter, R. Wolf, S.M. Pieber, E.A. Bruns, M. Crippa, G. Ciarelli, A. Piazzalunga, M. Schwikowski, G. Abbaszade, J. Schnelle-Kreis, R. Zimmermann, Z. An, S. Szidat, U. Baltensperger, I.E. Haddad, A.S.H. Prevot, Nature 514 (2014) 218–222. [4] Q. Li, X. Hou, H. Yang, Z. Ma, J. Zheng, F. Liu, X. Zhang, Z. Yuan, J. Mol. Catal. A: Chem. 356 (2012) 121–127. [5] S.A. Bates, A.A. Verma, C. Paolucci, A.A. Parekh, T. Anggara, A. Yezerets, W.F. Schneider, J.T. Miller, W.N. Delgass, F.H. Ribeiro, J. Catal. 312 (2014) 87–97. [6] M. Colombo, I. Nova, E. Tronconi, Catal. Today 151 (2010) 223–230. [7] S.J. Schmieg, S.H. Oh, C.H. Kim, D.B. Brown, J.H. Lee, C.H.F. Peden, D.H. Kim, Catal. Today 184 (2012) 252–261. [8] P. Balle, B. Geiger, D. Klukowski, M. Pignatelli, S. Wohnrau, M. Menzel, I. Zirkwa, G. Brunklaus, S. Kureti, Appl. Catal. B: Environ. 91 (2009) 587–595. [9] N. Wilken, K. Wijayanti, K. Kamasamudram, N.W. Currier, R. Vedaiyan, A. Yezerets, L. Olsson, Appl. Catal. B: Environ. 111–112 (2012) 58–66. [10] G. Cavataio, J. Girard, J.E. Patterson, C. Montreuil, Y. Cheng, C.K. Lambert, SAE Int. 01 (2007) 1575. [11] J.H. Kwak, D. Tran, S.D. Burton, J. Szanyi, J.H. Lee, C.H.F. Peden, J. Catal. 287 (2012) 203–209. [12] J.H. Kwak, R.G. Tonkyn, D.H. Kim, J. Szanyi, C.H.F. Peden, J. Catal. 275 (2010) 187–190. [13] D.W. Fickel, E. D’Addio, J.A. Lauterbach, R.F. Lobo, Appl. Catal. B: Environ. 102 (2011) 441–448. [14] D.W. Fickel, R.F. Lobo, J. Phys. Chem. C 114 (2009) 1633–1640. [15] L. Wang, W. Li, G. Qi, D. Weng, J. Catal. 289 (2012) 21–29. [16] J. Xue, X. Wang, G. Qi, J. Wang, M. Shen, W. Li, J. Catal. 297 (2013) 56–64. [17] J. Hun Kwak, H. Zhu, J.H. Lee, C.H.F. Peden, J. Szanyi, Chem. Commun. 48 (2012) 4758–4760. [18] Y.J. Kim, J.K. Lee, K.M. Min, S.B. Hong, I.-S. Nam, B.K. Cho, J. Catal. 311 (2014) 447–457. [19] J. Szanyi, J.H. Kwak, H. Zhu, C.H.F. Peden, Phys. Chem. Chem. Phys. 15 (2013) 2368–2380. [20] L. Wang, J.R. Gaudet, W. Li, D. Weng, J. Catal. 306 (2013) 68–77. [21] P.N.R. Vennestrøm, A. Katerinopoulou, R.R. Tiruvalam, A. Kustov, P.G. Moses, P. Concepcion, A. Corma, ACS Catal. 3 (2013) 2158–2161. [22] R. Martínez-Franco, M. Moliner, P. Concepcion, J.R. Thogersen, A. Corma, J. Catal. 314 (2014) 73–82. [23] F. Gao, E.D. Walter, N.M. Washton, J. Szanyi, C.H.F. Peden, Appl. Catal. B: Environ. 162 (2015) 501–514. [24] L. Ma, Y. Cheng, G. Cavataio, R.W. McCabe, L. Fu, J. Li, Chem. Eng. J. 225 (2013) 323–330. [25] J.Y. Yan, W.M.H. Sachtler, H.H. Kung, Catal. Today 33 (1997) 279–290. [26] M. Iwasaki, H. Shinjoh, Chem. Commun. 47 (2011) 3966–3968. [27] Y. Zhang, M. Flytzani-Stephanopoulos, J. Catal. 164 (1996) 131–145. [28] L. Pang, C. Fan, L. Shao, K. Song, J. Yi, X. Cai, J. Wang, M. Kang, T. Li, Chem. Eng. J. 253 (2014) 394–401. [29] H. Xu, Y. Wang, Y. Cao, Z. Fang, T. Lin, M. Gong, Y. Chen, Chem. Eng. J. 240 (2014) 62–73. [30] Z. Yang, N. Zhang, Y. Cao, M. Gong, M. Zhao, Y. Chen, Catal. Sci. Technol. 4 (2014) 3032–3043. [31] J.H. Kwak, R. Tonkyn, D. Tran, D. Mei, S.J. Cho, L. Kovarik, J.H. Lee, C.H.F. Peden, J. Szanyi, ACS Catal. 2 (2012) 1432–1440.

Y. Cao et al. / Journal of Molecular Catalysis A: Chemical 398 (2015) 304–311 [32] J. Wang, Y. Huang, T. Yu, S. Zhu, M. Shen, W. Li, J. Wang, Catal. Sci. Technol. 4 (2014) 3004–3012. [33] K.N. Rao, P. Venkataswamy, B.M. Reddy, Ind. Eng. Chem. Res. 50 (2011) 11960–11969. [34] H. Xu, Q. Zhang, C. Qiu, T. Lin, M. Gong, Y. Chen, Chem. Eng. Sci. 76 (2012) 120–128. [35] B.M. Reddy, A. Khan, Y. Yamada, T. Kobayashi, S. Loridant, J.-C. Volta, J. Phys. Chem. B 107 (2003) 5162–5167. [36] X. Yao, F. Gao, Q. Yu, L. Qi, C. Tang, L. Dong, Y. Chen, Catal. Sci. Technol. 3 (2013) 1355–1366. [37] L. Lan, S. Chen, M. Zhao, M. Gong, Y. Chen, J. Mol. Catal. A: Chem. 394 (2014) 10–21. [38] J. Liu, X. Li, Q. Zhao, D. Zhang, P. Ndokoye, J. Mol. Catal. A: Chem. 378 (2013) 115–123. [39] J. Krishna Reddy, G. Suresh, C.H. Hymavathi, V. Durga Kumari, M. Subrahmanyam, Catal. Today 141 (2009) 89–93. [40] L. Yuliati, T. Hamajima, T. Hattori, H. Yoshida, Chem. Commun. (2005) 4824–4826. [41] M.S.P. Francisco, V.R. Mastelaro, P.A.P. Nascente, A.O. Florentino, J. Phys. Chem. B 105 (2001) 10515–10522. [42] G.L. Christensen, M.A. Langell, J. Phys. Chem. C 117 (2013) 7039–7049. [43] P. Bera, K.R. Priolkar, P.R. Sarode, M.S. Hegde, S. Emura, R. Kumashiro, N.P. Lalla, Chem. Mater. 14 (2002) 3591–3601.

311

[44] W.E.J. van Kooten, B. Liang, H.C. Krijnsen, O.L. Oudshoorn, H.P.A. Calis, C.M. van den Bleek, Appl. Catal. B: Environ. 21 (1999) 203–213. [45] F. Gao, E.D. Walter, N.M. Washton, J. Szanyi, C.H.F. Peden, ACS Catal. 3 (2013) 2083–2093. [46] S. Brandenberger, O. Kröcher, M. Casapu, A. Tissler, R. Althoff, Appl. Catal. B: Environ. 101 (2011) 649–659. [47] F. Giordanino, P.N.R. Vennestrom, L.F. Lundegaard, F.N. Stappen, S. Mossin, P. Beato, S. Bordiga, C. Lamberti, Dalton Trans. 42 (2013) 12741–12761. [48] N. Wilken, R. Nedyalkova, K. Kamasamudram, J. Li, N. Currier, R. Vedaiyan, A. Yezerets, L. Olsson, Top. Catal. 56 (2013) 317–322. [49] Z. Li, M. Flytzani-Stephanopoulos, J. Catal. 182 (1999) 313–327. [50] S. Fan, J. Xue, T. Yu, D. Fan, T. Hao, M. Shen, W. Li, Catal. Sci. Technol. 3 (2013) 2357–2364. [51] C. Dossi, A. Fusi, G. Moretti, S. Recchia, R. Psaro, Appl. Catal. A: Gen. 188 (1999) 107–119. [52] J. Sárkány, J. Mol. Struct. 410–411 (1997) 137–140. [53] J. Wang, T. Yu, X. Wang, G. Qi, J. Xue, M. Shen, W. Li, Appl. Catal. B: Environ. 127 (2012) 137–147. [54] T. Yu, J. Wang, M. Shen, W. Li, Catal. Sci. Technol. 3 (2013) 3234–3241. [55] D. Wang, L. Zhang, K. Kamasamudram, W.S. Epling, ACS Catal. 3 (2013) 871–881.