Cu-zeolite NH3-SCR catalysts for NOx removal in the combined NSR–SCR technology

Chemical Engineering Journal 207–208 (2012) 10–17

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej

Cu-zeolite NH3-SCR catalysts for NOx removal in the combined NSR–SCR technology Unai De La Torre, Beñat Pereda-Ayo, Juan R. González-Velasco ⇑ Departamento de Ingeniería Química, Facultad de Ciencia y Tecnología, Universidad del País Vasco, UPV/EHU, Campus de Leioa, P.O. Box 644, ES-48080 Bilbao, Bizkaia, Spain

h i g h l i g h t s " 1–6 wt.% Cu exchanged BETA and ZSM-5 zeolites are tested in SRC, NSR and dual NSR–SCR systems. " 2.1% Cu/BETA and 1.4% Cu/ZSM-5 achieved higher activity and wider windows for NH3-SCR reaction. " ZSM5 based catalysts resulted in higher activity than BETA, related with reducibility of Cu. " Improvement above 30% in NOx conversion selectively to N2 is achieved when operating dual NSR–SCR systems.

a r t i c l e

i n f o

Article history: Available online 23 July 2012 Keywords: SCR NSR NSR–SCR NOx removal Diesel engine Cu-zeolite

a b s t r a c t The challenge of efficient NOx removal from diesel and lean-burn engine exhaust gas by combining NSR and SCR catalyst is studied. Several Cu exchanged zeolites have been prepared, varying the preparation method (ion exchange and impregnation), the copper content (1–6%) and the zeolite (BETA and ZSM5). The prepared catalysts have been characterized, and acidity, surface area, crystallinity and metal reducibility have been compared. SCR experiments under 750 ppm NO, 750 ppm NH3 and 9.5% O2 (Ar to balance) discriminated low copper loading, prepared by ion exchange catalyst (Z-IE-1.4 and B-IE-2.1) as the most active for NOx conversion (>95%) in ample temperature range (280–450 °C). These active SCR catalysts were placed downstream a monolith NSR Pt–BaO/Al2O3 catalyst, running under cycled lean–rich conditions, and the improvement on NOx removal and selectivity to only N2 were determined. In an ample range of temperature, from 200 to 400 °C, NOx conversion was increased in more than 30%, also notably increasing the production of nitrogen, and reducing production of ammonia and N2O below 3% and 2%, respectively, when comparing the combined NSR–SCR configuration versus the single NSR catalyst. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction It is now well recognized that the use of diesel and lean burn engines decreases the fuel consumption and thereby reduces the CO2 emissions. However, conventional three way catalysts (TWCs) are not capable to reduce nitrogen oxides (NOx), due to the excess of oxygen in the environment. In the last decade, two main approaches towards NOx reduction have been proposed: the NOx storage and reduction (NSR) technology and the NOx selective catalytic reduction (SCR). SCR was originally developed for stationary emission sources, mainly power plants [1]. However, it soon turned out to be a promising technology for the NOx removal in automobile applications as well [2]. In 2005 it was introduced for commercial heavy-duty vehicles in Europe, and more recently also for passenger cars [3]. The NH3-SCR converter needs an external source of the selective ⇑ Corresponding author. Address: Dept. of Chemical Engineering, University of the Basque Country, P.O. Box 644, ES-48980 Bilbao, Bizkaia, Spain. E-mail address: [email protected] (J.R. González-Velasco). 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.06.092

reducing agent, e.g. urea. The urea solution is injected in a controlled way into the exhaust line, where it is thermally decomposed into NH3 and CO2. The ammonia then reacts selectively with NOx under lean (oxidizing) conditions, giving N2 as the final product [4,5]. Non-noble metals like Cu, Fe and Ce supported ZSM5 and BETA, are among the most active catalysts for the urea/NH3-SCR process [6–9]. The NSR catalysts (also called lean NOx traps, LNT) consist of a cordierite monolith washcoated with a porous alumina on which an alkali-earth oxide (e.g. BaO) and a noble metal (Pt) are deposited [10,11]. These catalysts operate alternatively under lean and rich conditions [12]. During the lean period, when oxygen is in excess, the platinum oxidizes NO to a mixture of NOx (NO + NO2), which is adsorbed (stored) on Ba as various species (nitrite, nitrate). Before an unacceptable amount of NOx slips through the catalyst, the engine switches to rich condition (reducing) for a short period where the stored NOx are released and reduced into N2 over Pt. Different types of reducing agents such as hydrocarbon, CO and H2 have been used in NSR catalyst studies [13], and hydrogen has been found to be the most effective.

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

11

On a commercial NSR system, the efficiency to transforming the emitted NOx to N2 should be as high as possible. This is remarkable, as the Pt itself is selective for the formation of N2 from NO and H2 only in a narrow range of NO to H2 ratio. Nova et al. [14] concluded that the ammonia formation over Pt/Ba/Al was dependent on the amount of stored NOx, temperature and hydrogen concentration. The reduction by H2 of nitrates stored proceeds according to a two-step mechanism in which the first step is the fast reaction of hydrogen with nitrates producing ammonia, followed by the slower reaction of the latter with nitrate species leading selectivity to N2 [15–17]. Accordingly a good tuning of the operating conditions of both the adsorption rate and the reduction phases can drive selectively to N2 and/or NH3 [18,19]. Consequently, LNTs generate NH3 during the fuel-rich purge period, and SCR catalysts, especially those based on zeolites store significant quantities of NH3 under reaction conditions [20]. Hence, combining the LNT with a downstream zeolite SCR catalyst offers a potential means of capturing NH3 generated by the LNT and using it to convert NOx that slips through the NSR catalyst. We are referring to these systems as combined NSR–SCR technology. Corbos et al. [7,21] showed that the NOx removal efficiency can be greatly improved under lean–rich atmosphere if a NSR model catalyst is physically mixed with CuZSM-5; this effect was ascribed to an increase in the formation of NCO species, their formation being promoted by Cu/ZSM-5 catalysts. The control of NOx storage and reduction in LNTs for designing combined NSR–SCR systems has been reported elsewhere [22]. We proposed the use of N2/ NH3 production surfaces in response to operational variables, including temperature and H2 concentration during the rich period, to run efficiently a combined NSR–SCR system with Fe-BETA zeolite catalyst placed downstream a Pt–BaO/Al2O3 monolith. However, only a 2% Fe-BETA catalyst was tested in [22]. Cu2+ ion-exchanged ZSM5 (Cu-ZSM5) zeolites were first showing high NO decomposition rates and NOx SCR activities [23,24]. More recently, Cu2+-exchanged beta zeolites (Cu-BETA) have been shown to have good activity in the NH3-SCR of NOx, and metal-exchanged beta zeolites are generally found to have better hydrothermal stability than similar ZSM5 catalysts [25]. In this paper, several Cu exchanged ZSM5 and BETA zeolite catalysts have been prepared with copper loadings between 1 and 6 wt.% and their NH3-SCR behavior has been compared in relation to physico-chemical properties, including physical structure, redox properties and acidity. The powder catalysts were placed downstream of a Pt–BaO/Al2O3 monolith (previously synthesized and characterized elsewhere [10]) and the significant improvement in NOx removal efficiency to N2 without practical NH3 slip through the combined NSR–SCR system, running under cycled lean–rich atmosphere, is demonstrated. The H2 concentration during the rich period of the NSR cycle should be adequately tuned for the required intermediate production of ammonia.

washed twice in deionised water, dried during all night and calcined at 550 °C for 4 h. On the other hand, the impregnation method consisted in adding slowly the required amount of the precursor dissolved in water (1.5 wt.%) at 40 °C and 3 mm Hg to some grams of H-beta or HZSM5, under continuous rotation until the solvent was evaporated. The samples were dried and later calcined at 550 °C for 4 h. The actual amount of Cu in the prepared catalysts was determined by ICP-AES from the solid sample. All the catalysts were then pelletized, crushed and sieved to 0.3–0.5 mm to avoid mass transfer limitations, which was checked in some previous experiments carried out with different particle sizes. In order to characterize the SCR catalyst, several techniques were employed, such as BET surface area analysis, H2-TPR, NH3-TPD, and XRD. The prepared catalysts, with the precursor dissolution concentration and the actual copper content, are shown in Table 1. The Pt–BaO/Al2O3 NSR monolith catalyst was prepared according to our previously reported procedure [10]. In summary, a cordierite monolith, 20 mm in length and diameter, with a cell density of 400 cells per square inch and a wall thickness of 150 lm was washcoated with c-alumina (163 m2 g1 after stabilization at 700 °C, 4 h) by several immersions of the monolith into the alumina slurry until 1 g Al2O3 was deposited in the monolith structure. The incorporation of platinum was carried out by adsorption from tetraammine platinum (II) nitrate solution and the excess of liquid remaining in the channels was blown out with compressed air. After calcination in air (500 °C, 4 h) and subsequent reduction of the metallic phase in a 5% H2/N2 stream (500 °C, 1 h), the barium was incorporated by immersion of the monolith in a barium acetate solution. Finally, the catalyst was calcined again (500 °C, 4 h).

2. Experimental

2.2.3. XR diffraction The change in crystalline structure of the Cu modified zeolite samples were analyzed by XRD (Philips PW1710 diffractometer). The samples were finely ground and were subjected to Cu Ka radiation in continuous scan mode from 5° to 80° of 2h with 0.02° per second sampling interval. PANalytical X’pert HighScore specific software was used to data treatment. JCPDS database was used to confirm the spectrum.

2.1. Catalysts preparation The SCR catalysts consisted of Cu-supported zeolites. Fresh zeolites were supplied by Zeolyst International, namely CP414E (BETA, Si/Al = 25) and CBV5524G (ZSM5, Si/Al = 50). Zeolites were first calcined at 550 °C for 4 h to get the protonic form. The catalysts were prepared by two different conventional procedures, namely ion exchange (IE) and impregnation (IM). Metal ion exchange was carried out by dissolving the required amount of Cu(COOCH3)2 (Panreac, 98%) in water. Then, 12 g H-ZSM5 or H-BETA were added to 1.5 l of this solution and it was stirred for 24 h at 65 °C. The ion exchanged samples were then filtered,

2.2. Catalyst characterization 2.2.1. Ammonia temperature programmed desorption (NH3-TPD) Eighty milligrams sample (hydrate state) was placed in a Ushaped quartz reactor connected to a Micromeritics AutoChem 2910 instrument. The sample was pretreated in nitrogen flow at 550 °C for 15 min, cooled down to 100 °C and treated with helium for 60 min. Then, the sample was flushed with 10% NH3/He until saturation and the TPD was started using helium as carrier gas (500 ml/min, STP). The material was heated to 550 °C at the rate of 10 °C/min, while the NH3 desorption was continuously monitored with a TCD detector. The amount of ammonia desorbed at some given temperature range was taken as the acid site concentration, whereas the temperature range at which most of the ammonia was desorbed indicated the acid strength distribution. 2.2.2. Surface area The BET surface areas of the zeolite samples were determined by N2 adsorption–desorption at 196 °C using a Micromeritics ASAP 2020 equipment.

2.2.4. Hydrogen temperature programmed reduction (H2-TPR) Reducibility of Cu in the catalyst was investigated by temperature-programmed reduction (TPR) using H2. The sample was pretreated in 30 ml/min of 10% O2/He mixture gas flow at 550 °C for 45 min and then cooled down to 30 °C and flushed with helium

12

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

Table 1 Characteristics of the prepared catalysts. Support

Si/Al

Metal incorporation methodology

Cu initial conc. (ppm)

Exchanged amount (%)

Catalyst content Cu (wt.%)

Nomination

BETA

25

Ion exchange Ion exchange Ion exchange Ion exchange Ion exchange Impregnation

160 320 640 960 2000 –

75 53 36 38 23 –

1.5 2.1 2.9 4.5 5.8 1.3

B-IE-1.5 B-IE-2.1 B-IE-2.9 B-IE-4.5 B-IE-5.8 B-IM-1.3

ZSM-5

50

Ion exchange Ion exchange Ion exchange Ion exchange Ion exchange Impregnation

160 320 640 960 2000 –

70 40 33 28 20 –

1.4 1.6 2.6 3.4 4.9 1.2

Z-IE-1.4 Z-IE-1.6 Z-IE-2.6 Z-IE-3.4 Z-IE-4.9 Z-IM-1.2

for 60 min. Then samples were heated from room temperature to 600 °C with 10 °C/min ramp in a 60 ml/min of 5% H2/Ar mixture gas flow. The water formed during reduction with H2 was trapped using a cold trap and the hydrogen consumption was continuously monitored with a TCD detector. 2.3. Activity tests 2.3.1. SCR experiments The SCR experiments were performed in a downflow stainless steel reactor. The reactor tube, with 1 g of 0.03–0.05 mm pelletized Cu-zeolite SCR catalyst inside, was located into a 3-zone tube furnace. The temperature was measured by a thermocouple at the top of the catalyst bed. The reaction temperature was varied from 100 to 500 °C. The composition of the feed gas mixture was 750 ppm NO, 750 ppm NH3 and 9.5% O2 using Ar as the balance gas. Gases were fed via mass flow controllers and the total flow rate was set at 3000 ml min1, which corresponded to a space velocity (GHSV) of 90,000 h1. Previous experiments made with GHSV of 22,500 and 45,000 h1 achieved almost 100% conversion in a wide range of temperatures (220–460 °C) with all the catalysts prepared, which made difficult comparison of behavior and election of the best candidate for the double NSR–SCR configuration. The NO, NO2, NH3 and N2O concentration at the reactor exit were monitored every 40 °C, once the analysis has been stabilized for at least 10 min, by online FTIR multigas analyzer (MKS 2030). The NOx and NH3 conversions were calculated as

X NO ¼

F in NO

X NH3 ¼



F out NOx

F in NO

pN2 ¼  100

ð1Þ

out F in NH3  F NH3

 100

F in NH3

2F out N2 F in NH3 X NH3

SN 2 O ¼

SNO2 ¼

þ F in NO X NO

þ

F in NO X NO

F out NO2 in F in NH3 X NH3 þ F NO X NO

in

out

ðNO Þlean þ ðNO

ð2Þ

 100

2F out N2 O F in NH3 X NH3

2Nout 2

pNH3 ¼

pN2 O ¼

and the N2, N2O, and NO2 selectivities were calculated as

SN 2 ¼

2.3.2. NSR–SCR experiments The NOx storage–reduction experiments for the single NSR configuration were performed in a vertical downflow stainless steel reactor, inside which the Pt–BaO/Al2O3 monolith was placed. When testing the double NSR–SCR configuration, the SCR zeolite powder catalyst was packed in a second reactor connected downstream the NSR reactor. Temperature was measured by thermocouples at the top and the bottom of the NSR monolith, and in the bed of SCR catalyst. Streams from either the exit of the NSR monolith or the exit of the double NSR–SCR system can be addressed to analyzers. This system was run under cycled NSR conditions between long lean and short rich periods. The composition of the lean gas mixture for NOx storage was 750 ppm NO and 9.5% O2 using Ar as balance gas. During the rich period oxygen was replaced by hydrogen (4% H2) maintaining 750 ppm of NO in the feedstream. The duration of the lean and rich period (tL = 150 s and tR = 20 s, respectively) was maintained constant and controlled by two solenoid valves. Gases were fed via mass flow controllers and the total flowrate was set at 3000 ml min1, which corresponded to a space velocity of 28,620 h1 for the NSR and 90,000 h1 for the SCR. NOx storage and reduction tests were carried out varying the catalyst temperature from 100 to 400 °C. The objective was to maximize the production of N2, with minimum NH3, NO and N2O at the exit. Formation of NO2 was practically negligible. Following are the expressions to calculate productions:

Þrich

 100 ¼

ðNOin Þlean þ ðNOout Þrich 2N2 Oout ðNO Þlean þ ðNO

R tl þtr

0 F in NO ðt l

F out N2 dt þ tr Þ

 100

ð6Þ

 100

ð7Þ

R tl þtr

NHout 3

in

2

out

Þrich

 100 ¼

 100 ¼

F out NH3 dt 0 in F NO ðt l þ t r Þ R tl þtr

F out N2 O dt 0 in F NO ðtl þ tr Þ

2

 100

ð8Þ

ð3Þ 3. Results

 100

 100

ð4Þ

ð5Þ

which should close the nitrogen balance SN2 þ SN2 O þ SNO2 ¼ 1, as no more products were formed.

3.1. Catalyst characterization Table 1 resumes the preparation procedure of the SCR catalysts, where the incorporation methodology, the initial concentration of cooper in the precursor solution, the percentage of Cu adsorbed, the actual Cu content in the zeolites and the nomenclature for each catalyst is detailed. As it can be observed, increasing the Cu concentration in the initial solution, higher loading of Cu was achieved in the zeolites, increasing from 0.015 g Cu (g BETA)1 to a

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

maximum of 0.058 g Cu (g BETA)1. However, the incorporation of Cu to the zeolite is limited at some extent. While 75% of the Cu present in the solution was adsorbed when using low concentrated solutions (160 ppm Cu), the percentage of Cu adsorbed decreased to 23% using high concentrated solutions (2000 ppm Cu). Comparing the supports, BETA and ZSM-5, slightly higher loadings of Cu were achieved in BETA, which can be related to the lower Si/Al ratio, and consequently to the higher number of sites which can be exchanged with Cu. For each support, one sample was prepared through conventional wetness impregnation, leading to a Cu content of 1.3% and 1.2% for BETA and ZSM-5, respectively. XRD spectra of the protonic zeolites (H-BETA and H-ZSM-5) and those exchanged with Cu (not shown) were almost identical, which means that the crystalline nature of the zeolites was not modified after the incorporation of Cu. On the other hand, as cooper ions were introduced into the zeolite as exchangeable cations, only well dispersed and isolated Cu ions interacting with the zeolite framework via one or two oxygen-bridging bonds may be expected [26], and consequently no peaks corresponding to copper in any form is expected, as these species are undetectable by XRD. However, it has been also reported that due to hydrolytic/ thermal transformations and local precipitation of cooper hydroxide due to local changes in pH, the formation of higher nuclearity cooper complexes inside the zeolite channels is possible [26]. In this sense, we search for CuO and Cu in the zeolite support. CuO generally exists as monoclinic crystal with the maximum intense diffraction peaks at 35.5° and 38.7° (referred to 2h scale) corresponding to [1, 1, 1] and [1, 1, 1] planes, whereas Cu0 generally exists in the cubic crystal system with the maximum intense diffraction peak at 43.3°, corresponding to [1, 1, 1] plane. No peaks could be assigned to the presence of CuO or Cu0 because the concentration of this species was not high enough to observe well defined diffraction peaks and were probably overlapped by the zeolite signal. Nevertheless, the presence of CuO and Cu cannot be ruled out. In order to characterize the redox properties of the Cu-zeolite catalysts, TPR experiments were carried out over fresh and after SCR reaction catalysts (Fig. 1). According to a blank run with HZSM5 sample, the zeolite itself did not contain any reducible ions and no H2 uptake was noticed. For all catalysts TPR plots consisted of two partially overlapped signals; the first peak had its maximum between 260 and 350 °C for BETA supported catalysts and between 200 and 360 °C for ZSM5 supported catalysts, whereas the second peak had its maximum between 310 and 680 °C for BETA supported catalysts and between 320 and 490 °C for ZSM5 supported catalyst. Thus, Cu-ZSM5 catalysts were more easily reduced than Cu-BETA, due to higher amount of reducible copper at lower temperatures. The presence of two distinct reduction peaks in Cu-zeolite catalysts has been largely reported [27–30] and attributed to two sequential reduction steps, from Cu2+ to Cu+, and then from Cu+ to Cu0.

Fig. 1. TPR profiles of the prepared SCR catalysts. AR refers after SCR reaction.

CuO þ H2 ! Cu0 þ H; O 2þ

þ

13

ð9Þ þ

þ 0:5H2 ! Cu þ H

ð10Þ

Cuþ þ 0:5H2 ! Cu0 þ Hþ

ð11Þ

Cu

The hydrogen consumption signal was deconvoluted into two peaks identifying the temperature at which the H2 consumption was maximum. Besides, the area under the peaks was integrated in order to evaluate the total amount of hydrogen required for each reduction step (Table 2). In general, ZSM5 supported catalysts have a lower reduction temperature than BETA supported catalysts. On the other hand, there is a significant effect of Cu-loading on the reducibility of the Cu-sites, as it has been previously reported [31]. From our data it can be confirmed that increasing the Cu loading, the reduction temperature decreases. This is consistent with the formation of larger amounts of dimeric Cu-species at increasing Cu loading [32]. These dimeric Cu-species appearing at higher loadings contain bridging oxygen atoms that can react with H2 at comparably lower temperatures than isolated Cu-sites. Thus, it can be suggested that Cu dispersion decreases for high copper ladings. In the last two columns of Table 2 the total amount of H2 consumed and the H2/Cu ratio can be observed. As it was expected, the consumption of H2 increased with increasing the loading of Cu. According to the stoichiometry of reactions (9)–(11), a H2/Cu ratio close to 1 would mean that all copper in the catalyst is in the state of Cu2+ and that it has been reduced to Cu0, while a H2/Cu ratio close to 0.5 would mean that all copper is in the state of Cu+ or it has not been completely reduced during the TPR experiment. The majority of the prepared catalysts showed a H2/Cu ratio close to 1, revealing that Cu2+ was the most abundant species in the catalyst. Regarding the preparation method, no significant changes were noticed in the reduction temperature for samples prepared by liquid ion exchange or impregnation. In the same line, the H2TPR experiment carried out over catalysts already submitted to SCR reaction showed no significant changes compared to the fresh catalysts, indicating that the oxidation state of Cu remains practically unaltered after reaction. Acidity of the zeolite is one of the important parameter that determines the extent of NOx reduction with ammonia over zeolite based catalyst [33]. The acidity of all prepared catalysts was determined by NH3-TPD and the values are given in Table 3. All the samples exhibited two major desorption peaks; for ZSM5 based catalysts the first desorption peak was situated in the range 175– 210 °C corresponding to weak acid sites, whereas the second desorption peak, corresponding to strong acid sites, was detected in the range 220–370 °C. For BETA based catalyst, the low temperature desorption peak was coincident with that observed for ZSM5, but the high temperature desorption peak was observed in a much more narrow window, i.e. 250–280 °C. The high temperature desorption peak, or strong acid sites, can be related to the presence of Brønsted acid sites in zeolite catalysts [34–36]. The quantity of desorbed ammonia during TPD experiment can illustrate the number of acid sites in the sample, which generally increases with the copper content, the strong acidity in more extension. It can be also observed that the amount of desorbed ammonia at lower temperature is in general lower than the desorbed ammonia at higher temperatures, which means a superiority of Brønsted acid sites in comparison with Lewis acid sites. For ZSM5 supported catalysts, the total acidity grows gradually with the copper content, from 420 lmol NH3 (g cat.)1 corresponding to the bare zeolite to 555 lmol NH3 (g cat.)1 for Z-IE-4.9, when the copper was incorporated by L.I.E. On the other hand, the incorporation of copper by impregnation reduced the total acidity of the catalyst to 298 lmol NH3 (g cat.)1 for Z-IM-1.2, probably due to the blockage of the zeolite pores by copper aggregates which quan-

14

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

Table 2 TPR experiments. H2 consumption for Cu2+ ? Cu+, Cu+ ? Cu0, reduction temperatures and H2/Cu ratio. H2 consumption , lmol H2 (g cat.)1

Catalyst

Cu

2+

? Cu

+

+

T (°C)

0

Cu ? Cu

Total

H2/Cu ratio

T (°C)

Z-IE-1.4 F Z-IE-1.4 AR Z-IE-2.6 F Z-IE-4.9 F Z-IM-1.2 F

43 38 28 100 45

344 355 200 210 241

42 33 84 57 47

487 460 328 359 405

85 71 112 157 92

1.248 1.147 0.982 0.697 1.128

B-IE-1.5 F B-IE-2.1 F B-IE-2.1 AR B-IE-2.9 F B-IE-5.8 F B-IM-1.3 F

29 50 35 58 75 30

347 365 350 293 265 346

14 12 22 67 243 21

450 500 679 500 314 450

43 62 57 125 318 51

0.625 0.982 0.858 1.354 1.187 0.838

F = fresh catalyst; AR = after SCR reaction.

Table 3 Acidity of Cu-zeolite catalysts, determined by NH3-TPD. Sample

Weak acidity

Strong acidity

Total

lmol NH3 g1

T (°C)

lmol NH3 g1

T (°C)

lmol NH3 g1

H-ZSM5 Z-IE-1.4 Z-IE-1.6 Z-IE-2.6 Z-IE-3.4 Z-IE-4.9 Z-IM-1.2

183 161 122 46 94 176 50

197 210 197 188 183 200 185

237 256 295 380 359 379 248

371 246 220 274 269 235 273

420 417 417 426 453 555 298

H-BETA B-IE-1.5 B-IE-2.1 B-IE-2.9 B-IE-4.5 B-IE-5.8 B-IM-1.3

388 304 251 205 312 328 174

183 183 210 194 174 191 182

151 307 395 464 444 746 433

278 264 258 255 266 257 257

539 611 646 669 756 1074 607

Table 4 Surface area of the prepared SCR catalysts. ZSM5 samples

Surface area (m2 g1)

BETA samples

Surface area (m2 g1)

H-ZSM5 Z-IE-1.4 Z-IE-1.6 Z-IE-2.6 Z-IE-3.4 Z-IE-4.9 Z-IM-1.2

398 385 380 361 374 350 298

H-BETA B-IE-1.5 B-IE-2.1 B-IE-2.9 B-IE-4.5 B-IE-5.8 B-IM-1.3

585 562 493 481 463 459 427

tity is expected to be less for liquid ion exchange catalysts. Similar trends can be observed for Cu-BETA catalysts. The incorporation of cooper to the zeolite support also affected notably the surface area of the catalysts. The fresh ZSM5 zeolite presented a surface area of 398 m2 g1 which gradually decreased with increasing cooper loading. In fact, when 4.9% of Cu (Z-IE-4.9) was incorporated through liquid ion exchange to the zeolite, the surface area decreased to 350 m2 g1 (Table 4). On the other hand, the incorporation of Cu through wetness impregnation led to a much larger decrease in the exposed surface area (298 m2 g1, ZIM-1.2), which was related to the presence of cooper aggregates blocking the pores. In the case of BETA supported catalysts, a similar trend was observed. The surface area of the bare zeolite (BETA, 585 m2 g1) gradually decreased to 459 m2 g1 (B-IE-5.8) when Cu was incorporated by liquid ion exchange, whereas the surface area of the B-IM-1.3 was reduced to 428 m2 g1 when Cu was added by impregnation.

Fig. 2. Conversion of NOx (filled symbols) and NH3 (empty symbols) for (a) BETA and (b) ZSM5 supported catalysts for SCR reaction. Copper content: j low IE, N intermediate IE, . high IE and d low IM.

3.2. Ammonia SCR activity tests The NOx selective catalytic reduction activity (SCR) tests of the prepared catalysts was carried out under a feedstream with the following composition: 750 ppm NO, 750 ppm NH3, 9% O2 and Ar to balance. The total flow rate was set at 3000 ml min1, which corresponded to a space velocity (GHSV) of 90,000 h1. This GHSV, higher than used in practice, allowed best comparing the behavior of the prepared catalysts, as makes the effect of temperature more relevant. The NO, NO2 NH3 and N2O concentrations at the reactor exit were monitored from 100 to 500 °C, every 40 °C, once the analysis had been stabilized for 10 min. Fig. 2a and b shows the conversions of NOx and NH3 as a function of the reaction temperature. For each support, ZSM5 and BETA, three catalysts were chosen as representative of low, intermediate and high Cu content, prepared by ion exchange. The copper low content catalysts prepared by impregnation are also included in Fig. 2. The conversion trends are typical for NOx SCR reactions, reaching a maximum in activity for an intermediate temperature [37,38]. For the samples studied, the NOx conversion maximum was reached between 350 and 450 °C for BETA supported catalysts and between 250 and 350 °C for ZSM5 supported catalysts. The higher activity of ZSM-5 based catalysts at lower temperatures is probably related with the higher reducibility of Cu at low temperature observed in TPR experiments.

15

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

Table 5 shows the maximum NOx conversion for each catalyst and the temperature at which it has been reached (three first columns). In the case of BETA supported catalysts, B-IE-2.1 showed the maximum NOx conversion of 95.8% at 420 °C. For catalysts with higher copper loadings the maximum NOx conversion decreased to 87.3% and 82.7% for B-IE-4.5 and B-IE-5.8, respectively, which instead were achieved at lower temperature, 340 °C. It is suggested that the increase of Cu loading promotes the oxidation of NO to NO2 at lower temperature which activates the fast SCR reaction (2NH3 + NO + NO2 ? 2N2 + 3H2O), and consequently shifts the maximum NOx conversion to lower temperature. The fourth column in Table 5 shows that ammonia conversion reached almost 100% for NOx maximum conversion. Only N2 and N2O (no NO2) were detected at the reactor exit as deduced from selectivities shown in columns fifth and sixth, which close the mole balance with N2 and N2O. It can be noted that higher copper loadings enhance selectivity to N2O, but always maintained below 6.5%. The last column in Table 5 indicates the amplitude of the temperature windows for maintaining NOx conversion higher than 70%. The wider amplitude was achieved with low copper content (B-IE-2.1) and it was decreased with the copper loading. Thus, it can be concluded that among the BETA catalysts prepared by ion exchange, the low Cu content B-IE-2.1catalyst resulted the most active as it combines the highest NOx conversion and the wider temperature window. Although the increase in the copper content has a beneficial effect on the activity at low temperature, however the maximum NOx conversion and temperature window amplitude are significantly reduced due to some ineffective use of copper. The catalyst B-IM-1.3, prepared by impregnation, achieved the lowest maximum conversion and the narrowest temperature window, as it was expected due to the low Cu dispersion obtained by this preparation method. In the case of ZSM5 supported catalysts prepared by ion exchange, as the copper content increases the NOx conversion curve shifted to lower temperatures whereas the catalyst achieved lower maximum NOx conversions. This trend is similar to that previously described for BETA catalysts; however, all Cu-ZSM5 catalysts presented very similar amplitudes of temperature windows. Thus, the best ZSM5 catalyst has to be chosen only based on the maximum NOx conversion, resulting in the low Cu loading Z-IE-1.4. Again, the impregnated Z-IM-1.2 can be disregarded due to its worse behavior, with 75.2% of maximum NOx conversion and narrow temperature window amplitude of 75 °C.

lean period (150 s); and 750 ppm NO, 4% H2 in Ar during the rich period (20 s). Concentrations of N2/NH3/N2O were monitored at the intermediate position, i.e. after NSR but before SCR, where production of ammonia as reactant for subsequent SCR should be important, and also they were monitored at the exit of the double NSR–SCR reactor. Fig. 3 shows the evolution of NOx conversion and the NH3 and N2 production with temperature, for the simple NSR configuration and for the combined NSR–SCR configuration, the latter with B-IE2.1 and Z-IE-1.4 SCR catalysts. It can be observed that the addition of an SCR catalyst downstream the NSR catalyst improves significantly the NOx conversion, mainly for temperatures between 200 and 250 °C. The maximum NOx conversion achieved with the single NSR resulted in 50%, and it was increased up to 76% with the double NSR–SCR configuration, independently of the SCR catalyst. In concordance with the SCR activity previously analyzed, the Z-IE1.4 catalyst made the double NSR–SCR configuration more active at lower temperature (200 °C) than the B-IE-2.1 (250 °C). It is worthy to note that the combined NSR–SCR system achieved efficient NOx reduction at lower temperatures than NSR or SCR single systems with the aid if intermediate ammonia formed during the rich

3.3. NSR–SCR activity tests The most active catalyst for SCR reaction, B-IE-2.1 and Z-IE-1.4, were used to accomplish the double NSR–SCR configuration, by adapting the SCR catalyst downstream the NSR catalyst. The monolith NSR catalyst used has been the same for all experiments [10]. The experiments in the double system were made with the following feedstream composition: 750 ppm NO, 9.5% O2 in Ar during the

Fig. 3. (a) NOx conversion, (b) NH3 production, and (c) N2 production for single NSR and double NSR–SCR configurations, from 150 to 400 °C. (Z for ZSM5, B for BETA).

Table 5 Maximum NOx conversion and corresponding temperature; NH3 conversion and N2/N2O at this temperature. Sample

Maximum NOx conversion

Temperature window for XNO > 70%

XNO (%)

T (°C)

X NH3 (%)

SN2 (%)

SN2 O (%)

B-IE-2.1 B-IE-4.5 B-IE-5.8 B-IM-1.3

95.8 87.3 82.7 82.7

420 340 340 380

98.6 99.2 98.4 99.8

98.4 95.5 93.5 97.6

1.3 4.5 6.5 2.4

280–520 °C 275–440 °C 250–395 °C 320–450 °C

(240) (165) (145) (130)

Z-IE-1.4 Z-IE-3.4 Z-IE-4.9 Z-IM-1.2

97.9 93.6 89.3 75.2

340 340 300 300

99.7 99.3 99.7 98.9

99.5 96.1 94.9 98.6

0.5 3.9 5.1 1.1

280–440 °C 235–425 °C 205–375 °C 260–355 °C

(160) (190) (170) (75)

16

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

Table 6 Production of N2, NH3 and N2O for single NSR and double NSR–SCR systems at different temperatures. Temperature (°C)

150 200 250 300 350 400

Single NSR

NSR + Z-IE-1.4

NSR + B-IE-2.9

pN2

pNH3

pN2 O

pN2

pNH3

pN2 O

pN2

pNH3

pN2 O

7.6 9.8 27.1 32.4 35.0 29.4

29.2 28.8 19.5 17.2 12.2 9.4

2.5 2.3 0.7 0.4 0.3 0.3

10.0 70.8 72.7 71.6 64.2 50.4

28.6 1.7 0.2 0.1 0.05 0.04

2.1 2.0 1.2 0.6 0.5 0.5

22.8 48.7 73.0 70.3 63.0 51.7

19.3 12.8 1.8 0.3 0.2 0.1

2.2 2.5 1.4 0.8 0.7 0.7

When operating the double NSR–SCR configuration with B-IE2.1 SCR catalyst, the NOx concentration at the reactor exit decreases notably in comparison with the single NSR. As it can be observed in Fig. 4b most of the NH3 formed during the NSR catalyst regeneration is adsorbed in the B-IE-2.1 catalyst. Then, in the subsequent lean period, the NOx slipping the NSR reacts with the ammonia stored following the SCR reaction. This is the reason why the NOx concentration at the reactor exit is much lower for the double configuration. The occurrence of the SCR reaction was verified by the increase in the N2 signal detected by MS, which was not observed for the single NSR. As it can be observed in Fig. 3, at 200 °C the NOx conversion resulted in 66% with N2 as the principal product, i.e. with a production of 49% but 13% of NH3. The Z-IE-1.4 catalyst in the double configuration makes more evident the enhancement in NOx conversion and N2 production, resulting in values of 76% and 71%, respectively, with ammonia no more than 2%. The fact that this catalyst adsorbs more ammonia at 200 °C (Fig. 4b) makes the catalyst more active for the SCR reaction.

4. Conclusions

Fig. 4. NOx and NH3 concentration profiles, and N2 MS-signal profiles for single NSR and combined NSR–SCR configurations, at 200 °C.

period of the NSR process in the SCR placed dwonstream, but no need of external NH3 feed as in the case of the single SCR system. The product distribution at the exit of the reactor is markedly influenced by the configuration used (Table 6). The N2O production was always below 3% and even decreases with temperature below 1%. The double NSR–SCR configuration decreases the NH3 production (Fig. 3b) in favor of higher N2 production (Fig. 3c), in the whole studied temperature range. This behavior improvement can be explained from data shown in Fig. 4. Fig. 4 shows the NOx and NH3 concentration profiles determined by online FTIR when the reaction was conducted at 200 °C, for single and double configurations. Also the N2 signal profiles determined by MS are shown. When operating with the single NSR system, the typical results were obtained [18]. At the beginning of the lean period practically all the NO fed was stored in the NSR Pt–BaO/Al2O3 catalyst in the form of nitrates and nitrites [39]. Then as the lean period time increases the adsorption sites become progressively saturated and the NOx concentration at the NSR outlet is increasing. Afterwards, during the rich period, the injected H2 produced the release and reduction of the stored NOx to a mixture of N2, NH3 and N2O [19]. In Fig. 4b it can be noted a sharp peak corresponding to ammonia concentration at the exit of the single NSR. As it can be observed in Fig. 3, at these operating conditions (200 °C) ammonia is the principal product at the reactor exit (29%), while the N2 production only accounts for 10% which was also detected as a sharp peak in Fig. 4c.

On commercial NOx removal diesel automobile converters, increasing the efficiency to transform all emitted NOx only to N2 is challenging. Recently, combined NSR–SCR systems are been investigated as most promising alternative. In this work several Cu exchanged zeolites have been prepared, varying the preparation method (ion exchange and impregnation), the copper content (1– 6%) and the zeolite (BETA and ZSM5). The prepared catalysts were tested in the SCR reaction, the NSR reaction and the dual NSR–SCR configuration. The catalysts prepared by ion exchange and low copper content, i.e. B-IE-2.1 and Z-IE-1.4 resulted the most active in the SCR reaction, as gave maximum conversion and wider temperature window amplitude. Cu2+ was the most active species in the catalysts, as deduced from the H2/Cu ratio close to unity in H2-TPR experiments, for both the fresh and after reaction samples. Higher metal loading decreased the maximum NOx conversion although this was achieved at lower temperature, probably due to promotion of NO to NO2 oxidation and subsequent activation of fast SCR reaction. Higher Cu loadings enhanced selectivity to N2O, but it was always maintained below 6%, and NO2 was not detected at the temperature of maximum NOx conversion. Comparing the support, ZSM5 resulted active at lower temperature than BETA, this related with the higher reducibility of Cu observed in the TPR experiments. The low metal loading catalysts prepared by impregnation achieved much lower NOx conversion and narrower temperature window, due to the formation of copper aggregates during this preparation methodology, which decrease metal dispersion on the catalyst. In experiments with the combined NSR–SCR system, ammonia was not fed to the reactor. Instead, the ammonia formed during the rich period in the NSR catalyst was used for NOx reduction in the downstream SCR catalyst. Improvement above 30% in NOx

U. De La Torre et al. / Chemical Engineering Journal 207–208 (2012) 10–17

conversion selectively to N2, has been achieved when operating the combined NSR–SCR system in comparison with operation in the single NSR system, while decreasing the production NH3 and N2O at the exit at very levels below 3% and 2%, respectively. In accordance with the SCR activity of the catalysts, the Z-IE-1.4 catalyst enhanced NOx conversion and N2 production from 200 °C and above, while this improvement was detected with the catalyst BIE-2.1 from 300 °C and above. Above 300 °C the behavior of both catalysts can be considered very similar. 5. Final remarks The dual NSR–SCR configuration has been tested with a previously optimized NSR monolith catalyst [10] followed by different powder SCR catalysts prepared as reported in the present paper. In real automobile application also the SCR catalyst should be used in monolithic form. Once concluded here that B-IE-2.1 and Z-IE-1.4 have resulted best potential SCR catalyst candidates for the combined NSR–SCR systems, we are synthesizing them on monolith structures and their kinetic behavior and operating conditions will be reported in a near future paper. Acknowledgements Authors wish to acknowledge the financial support provided by the Spanish Economy and Competitivity Ministry (CTQ200912517) and the Basque Government (GIC 07/67-JT-450-07). One of the authors (UDLT) wants to acknowledge to the Basque Government for the PhD Research Grant (BFI-2010-330). References [1] P. Forzatti, L. Lietti, E. Tronconi, Nitrogen oxides removal, in: I.T. Horvath (Ed.), Industrial Encyclopedia of Catalysis, Wiley, New York, 2002. [2] R.H. Heck, R.J. Farrauto, S.T. Gulati, Catalytic Air Pollution Control, second ed., John Wiley & Sons, New York, 2002. [3] C. Enderle, G. Vent, M. Paule, F. Duvinage, BLUETEC Diesel Technology – Clean, Efficient and Powerful, SAE Tech. Paper 2008-01-1182, 2008. [4] P. Forzatti, L. Lietti, I. Nova, E. Tronconi, Diesel NOx aftertreatment catalytic technologies: analogies in LNT and SCR catalytic chemistry, Catal. Today 151 (2010) 202. [5] R. Bonzi, L. Lietti, L. Castoldi, P. Forzatti, NOx removal over a double-bed NSR– SCR reactor configuration, Catal. Today 151 (2010) 376. [6] L. Lietti, I. Nova, E. Tronconi, P. Forzatti, Transient kinetic study of the SCR– DeNOx, reaction, Catal. Today 45 (1998) 85. [7] E.C. Corbos, M. Haneda, X. Courtois, P. Marecot, D. Duprez, H. Hamada, Cooperative effect of Pt–Rh/Ba/Al and CuZSM-5 catalysts for NOx reduction during periodic lean–rich atmosphere, Catal. Commun. 10 (2008) 137. [8] N. Wilken, K. Wijayanti, K. Kamasamudram, N.W. Currier, R. Vedaiyan, A. Yezerets, L. Olsson, Mechanistic investigation of hydrothermal aging of Cubeta for ammonia SCR, Appl. Catal. B 112 (2012) 58. [9] A. Corma, V. Fornés, E. Palomares, Selective catalytic reduction of NOx on Cubeta zeolites, Appl. Catal. B 11 (1997) 233. [10] B. Pereda-Ayo, R. López-Fonseca, J.R. González-Velasco, Influence of the preparation procedure of NSR monolithic catalysts on the Pt–Ba dispersion and distribution, Appl. Catal. A: Gen. 363 (2009) 73. [11] B. Pereda-Ayo, D. Duraiswami, R. López-Fonseca, J.R. González-Velasco, Influence of Pt and Ba precursors on the NSR behaviour of Pt–Ba/Al2O3 monoliths for lean-burn engines, Catal. Today 147 (2009) 244. [12] W.S. Epling, L.E. Campbell, A. Yezerets, N.W. Currier, L.E. Parks, Further evidence of multiple NOx sorption sites on NOx storage/reduction catalysts, Catal. Today 96 (2004) 21. [13] H. Abdulhamid, E. Fridell, M. Skoglundh, Influence of the type of reducing agent (H2, CO, C3H6 and C3H8) on the reduction of stored NOx in a Pt/BaO/Al2O3 model catalyst, Topics Catal. 30–31 (2004) 161. [14] I. Nova, L. Castoldi, L. Lietti, E. Tronconi, P. Forzatti, How to control the selectivity in the reduction of NOx with hydrogen over Pt–Ba/Al2O3 lean NOx trap catalysts, Topics Catal. 42–43 (2007) 21.

17

[15] I. Nova, L. Lietti, P. Forzatti, Mechanistic aspects of the reduction of stored NOx over Pt–Ba/Al2O3 lean NOx trap systems, Catal. Today 136 (2008) 128. [16] S.S. Mulla, S.S. Chaugule, A. Yezerets, N.W. Currier, W.N. Delgass, F.H. Ribeiro, Regeneration mechanisms of Pt/BaO/Al2O3 lean NOx trap catalyst with H2, Catal. Today 136 (2008) 136. [17] B. Pereda-Ayo, J.R. González-Velasco, R. Burch, C. Hardacre, S. Chansai, Regeneration mechanisms of a lean NOx trap (LNT) catalyst in the presence of NO investigated using isotope labeling techniques, J. Catal. 285 (2012) 177. [18] B. Pereda-Ayo, D. Duraiswami, J.J. Delgado, R. López-Fonseca, J.J. Calvino, S. Bernal, J.R. González-Velasco, Tunning operational conditions for efficient NOx storage and reduction over Pt–Ba/Al2O3 monolith catalyst, Appl. Catal. B: Environ. 96 (2010) 329. [19] B. Pereda-Ayo, D. Duraiswami, J.A. González-Marcos, J.R. González-Velasco, Performance of NOx storage-reduction catalyst in the temperature–reductant concentration domain by response surface methodology, Chem. Eng. J. 169 (2011) 58. [20] G. Guo, J. Warner, G. Cavataio, D. Dobson, E. Badillo, C. Lambert, The Development of Advanced Urea-SCR Systems for Tier 2 Bin 5 and Beyond Diesel Vehicles, SAE Technical Paper 2010-01-1183. [21] E.C. Corbos, M. Haneda, X. Courtois, P. Marecot, D. Duprez, H. Hamada, NOx abatement for lean-burn engines under lean–rich atmosphere over mixed NSR–SCR catalysts: influences of the addition of a SCR catalyst and of the operational conditions, Appl. Catal. A: Gen. 365 (2009) 187. [22] B. Pereda-Ayo, D. Divakar, J.R. González-Velasco, Control of NOx storage and reduction in NSR bed for designing combined NSR–SCR systems, Catal. Today 172 (2011) 66. [23] M. Iwamoto, H. Yahiro, Y. Torikai, T. Yoshioka, N. Mizuno, Novel preparation method of highly copper ion-exchanged ZSM5 zeolites and their catalytic activities for NO decomposition, Chem. Lett. 11 (1990) 1967. [24] A. Grossale, I. Nova, E. Tronconi, D. Chatterjjee, M. Weibel, The chemistry of the NO/NO2–NH3 ‘‘fast’’ SCR reaction over Fe-ZSM5 investigated by transient reaction analysis, J. Catal. 256 (2008) 312. [25] S. Brandenberger, O. Kröcher, A. Tissler, R. Althoff, The state of the art in the selective catalytic reduction of NOx by ammonia using metal-exchanged zeolite catalysts, Catal. Rev. Sci. Eng. 50 (2008) 492. [26] G. Centi, S. Perathoner, Nature of active species in cooper-based catalysts and their chemistry of transformation of nitrogen oxides, Appl. Catal. A: Gen. 132 (1995) 179. [27] C. Torre-Abreu, M.F. Ribeiro, C. Henriques, G. Delahay, Characterisation of CuMFI catalysts by temperature programmed desorption of NO and temperature programmed reduction. Effect of the zeolite Si/Al ratio and copper loading, Appl. Catal. B. 12 (1997) 249. [28] R. Bulánek, B. Wichterlová, Z. Sobalik, J. Tichy´, Reducibility and oxidation activity of Cu ions in zeolites: effect of Cu ion coordination and zeolite framework composition, Appl. Catal. B 31 (2001) 13. [29] J.Y. Yan, G.D. Lei, W.M.H. Sachtler, H.H. Kung, The activation of Cu/ZSM5 catalysts for lean NOx reduction: characterization of changes of Cu states and zeolite support, J. Catal. 161 (1996) 43. [30] J.Y. Yan, W.M.H. Sachtler, H.H. Kung, Effect of Cu loading and addition of modifiers on the stability of Cu/ZSM5 in lean NOx reduction catalysis, Catal. Today 33 (1997) 279. [31] B. Moden, J.M. Donohue, W.E. Cormier, H.-X. Li, Effect of Cu-loading and structure on the activity of Cu-exchanged zeolites for NH3-SCR, in: Proc. 4th Int. FEZA Conference, 2008, pp. 1219–1222. [32] P. Da Costa, B. Moden, G.D. Meitzner, D.K. Lee, E. Iglesia, Spectroscopic and chemical characterization of active and inactive Cu species in NO decomposition catalysts based on Cu-ZSM5, Phys. Chem. Chem. Phys. 4 (2002) 4590. [33] A. Sultana, T. Nanba, M. Sasaki, M. Haneda, K. Suzuki, H. Hamada, Selective catalytic reduction of NOx with NH3 over different copper exchanged zeolites in the presence of decane, Catal. Today 164 (2011) 495. [34] N.A.S. Amin, S. Ammasi, Dual-bed catalytic system for direct conversion of methane to liquid hydrocarbons, J. Nat. Gas Chem. 15 (2006) 191. [35] C. Costa, I.P. Dzikh, J.M. Lopes, F. Lemos, F.R. Ribeiro, Activity–acidity relationship in zeolite ZSM-5. Application of Brönsted-type equations, J. Mol. Catal. A: Chem. 154 (2000) 193. [36] D.W. Fickel, E. D’Addioa, J.A. Lauterbacha, R.F. Lobo, The ammonia selective catalytic reduction activity of copper-exchanged small-pore zeolites, Appl. Catal. B 102 (2011) 441. [37] J.H. Kwak, D. Tran, S.D. Burton, J. Szanyi, J.H. Lee, C.H.F. Peden, Effects of hydrothermal aging on NH3–SCR reaction over Cu/zeolites, J. Catal. 287 (2012) 203. [38] J. Li, H. Changa, L. Maa, J. Haoa, R.T. Yang, Low-temperature selective catalytic reduction of NOx with NH3 over metal oxide and zeolite catalysts – a review, Catal. Today 175 (2011) 147. [39] U. Elizundia, R. López-Fonseca, I. Landa, M.A. Gutiérrez-Ortiz, J.R. GonzálezVelasco, FTIR study of NOx storage mechanisms over Pt–BaO/Al2O3 catalysts. Effect of the Pt–BaO interaction, Topics Catal. 42–43 (2007) 37.