Zeolite Y modified with palladium as effective catalyst for selective catalytic oxidation of ammonia to nitrogen

Journal of Catalysis 316 (2014) 36–46

Contents lists available at ScienceDirect

Journal of Catalysis

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

Zeolite Y modified with palladium as effective catalyst for selectivecatalytic oxidation of ammonia to nitrogen

http://dx.doi.org/10.1016/j.jcat.2014.04.0220021-9517/� 2014 Elsevier Inc. All rights reserved.

⇑ Corresponding author. Fax: +48 12 6340515.E-mail address: [email protected] (K. Góra-Marek).

Magdalena Jabłonska a, Anna Król b, Ewa Kukulska-Zajac b, Karolina Tarach a, Lucjan Chmielarz a,Kinga Góra-Marek a,⇑a Faculty of Chemistry, Jagiellonian University in Kraków, Ingardena 3, 30-060 Kraków, Polandb Oil and Gas Institute, Lubicz 25A, 31-503 Kraków, Poland

a r t i c l e i n f o

Article history:Received 8 January 2014Revised 22 April 2014Accepted 28 April 2014

Keywords:PalladiumZeolitesSCOIR spectroscopy

a b s t r a c t

Zeolites HY modified with palladium (0.05–2.5 wt.%) were found to be active and selective catalysts inthe process of the selective oxidation of ammonia to nitrogen (NH3-SCO). Parent zeolite and its modifi-cations with palladium were investigated with regard to their structural, textural, and redox properties.The IR spectroscopy, X-ray diffraction, and BET methods were employed for these purposes. Correlationbetween loading and aggregation of Pd-species deposited on zeolites and their activity and selectivity inthe NH3-SCO process was found. Studies of adsorption forms of ammonia and product of its conversiondone by IR spectroscopy as well as catalytic tests performed with various space velocities gave insightinto the reaction mechanism. Zeolite Y modified with palladium is an interesting catalyst of selectiveammonia oxidation for potential application in diesel car exhaust gas purification system.

� 2014 Elsevier Inc. All rights reserved.

1. Introduction

An efficient technology for NOx (=NO + NO2) abatement frompower plant flue gases and mobile sources is its selective catalyticreduction (NH3-SCR, DeNOx) with ammonia (4NO + 4NH3 + -O2 ? 4N2 + 6H2O). Although the stoichiometry of the reactioninvolves a 1 to 1 consumption ratio of NH3 to NO, majority of theDeNOx installations operate under the conditions of NH3/NO < 1(�0.9–0.95) to control ammonia slip [1].

A strategy for improving the abatement of NOx emissiontogether with keeping the goal of low NH3 slip could be a connec-tion in series, downstream of the DeNOx converter, additional reac-tor for the NH3-SCO process, in which ammonia unreacted in theDeNOx process will be converted into nitrogen and water vapor:

4NH3 þ 3O2 ! 2N2 þ 6H2O ð1Þ

An alternative could be adding of the additional catalytic bedinto the DeNOx converter, downstream to NH3-SCR catalyst,responsible for the selective ammonia oxidation to nitrogen. Thistechnology seems to be very promising because it does not needsupplying of any additional reactant for the NH3-SCO process (oxy-gen, which plays a role of oxidizing agent, is a typical component of

flue gases). The problem is only the development of suitable cata-lyst, which will be able to selectively convert ammonia to nitrogenat relatively low temperatures (T < 400 �C) in order to avoid addi-tional heating of flue gases. Moreover, the catalyst should be stableunder reaction conditions and resistant for poisoning by all compo-nents present in flue gases. Various types of materials have beenstudied as potential catalysts for the selective ammonia oxidationto N2. Among them the most important are as follows: (a) noblemetals [e.g. 2–4], (b) transition metal oxides [e.g. 5,6], and (c)ion-exchanged zeolites [e.g. 7,8]. An early review by Il’chenko [9]on NH3 oxidation compared catalytic performance of various zeo-lite-based catalysts for the low-temperature process. It was shownthat zeolite Y modified with different transition metals by ion-exchange method was active in the NH3-SCO process. The follow-ing order of decreasing activity of transition metal modified zeo-lites was reported: CuY > CrY > AgY > CoY > FeY > NiY �MnY. Inall cases the selectivity to N2 was relatively high (>95%) with anexception of AgY and CrY. Li and Armor [10] studied ZSM-5 dopedwith Pd, Rh, and Pt as catalysts for the NH3-SCO process. Amongthem, relatively high activity and selectivity to N2 in the low-tem-perature range (6300 �C) were found for the Pd-containing cata-lysts. Similar results, proving high activity of the Pd-ZSM-5catalyst under different conditions, were also reported by Longand Yang [11]. For both Y and ZSM-5-based catalysts, it was shownthat loading, aggregation, and location of noble metal speciesstrongly influence their catalytic performance [e.g., 12,13]. In spite

Table 1

M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46 37

of numerous studies focused on this problem, it has not been fullyexplained in the case of the NH3-SCO process, and therefore, fur-ther studies are necessary. Also studies of the mechanism ofammonia oxidation over zeolites modified with noble metals havenot given full explanation; however, various possible mechanismswere proposed in the scientific literature. One of them, which isoften called the imide mechanism, was proposed by Zawadzki[14]. In the first step, ammonia is oxidized with the formation ofimide (NH); then, imide species react with molecular oxygen (O)to form the nitrosyl (HNO) intermediate, which in the next stepreacts with imide resulting in the formation of N2 and H2O. N2Ois formed by reaction of two nitrosyl species, while reactionbetween NH and O2 leads to the formation of NO. According to thismechanism, dissociation of O2 and formation of active atomic oxy-gen species (O) is the key reaction step determining reaction rate.

Another mechanism of selective ammonia oxidation, called thehydrazine mechanism [15,16], involves, in the first step, oxidationof ammonia to amide (NH2) species by atomic oxygen (O) and sub-sequently the formation of hydrazine-type intermediate (N2H4). Inthe next step, N2H4 is oxidized by O2 to N2 and N2O. It should benoted that also in case of this mechanism, active oxygen speciesplay an important role in ammonia conversion.

The i-SCR (internal selective catalytic reduction) [17–19] isanother mechanism of ammonia oxidation. This mechanism con-sists of two main steps. In the first step, part of ammonia moleculesis oxidized to NO:

4NH3 þ 5O2 ! 4NOþ 6H2O ð2Þ

where N2 and N2O are formed in the subsequent reactions betweenNO and ammonia (reactions 3a and 3b, respectively), which was notconsumed in the process (reaction 2):

4NOþ 4NH3 þ O2 ! 4N2 þ 6H2O ð3aÞ

4NOþ 4NH3 þ 3O2 ! 4N2Oþ 6H2O ð3bÞ

Thus, in this case, the effective catalyst should be active in twoprocesses, ammonia oxidation to NO (reaction 2) and reduction ofNO by NH3 with the formation of N2 (reaction 3a).

In this paper, a series of the Pd-catalysts was prepared by wet-ness impregnation of palladium species from PdCl2 solution onzeolite Y. The resulting materials were examined by XRD, FTIRspectroscopy, and nitrogen adsorption techniques. The sorptionof NH3 allowed to follow in quantitative manner the changes inconcentration of Brønsted and Lewis acid centers, resulted fromPd deposition on zeolite. Furthermore, the experiments of ammo-nia and carbon monoxide sorption led to quantitative interpreta-tion of Pd-forms. The effects of both the Pd content and theactive component distribution on the catalytic performances ofthe studied materials were discussed. Additionally, the preliminarystudies of the reaction mechanism by analysis of adsorbed speciesformed on the catalyst surface under reaction conditions by FTIRspectroscopy and catalytic tests performed with various spacevelocities were done.

Conditions of Pd ion-exchange for studied zeolites HY and results of chemicalanalysis.

Zeolite Chemical analysisof Pd content(lmol g�1)

Volume ofPdCl2 solution(cm3)

Supportmass (g)

Time ofexchange(h)

HY – – – –0.05 wt.% Pd/HY 5 2.10 1.00 240.1 wt.% Pd/HY 10 4.20 1.00 240.25 wt.% Pd/HY 24 10.75 1.00 240.5 wt.% Pd/HY 47 21.10 1.00 241 wt.% Pd/HY 94 42.23 1.00 241.5 wt.% Pd/HY 141 63.10 1.00 242.5 wt.% Pd/HY 235 105.00 1.00 24

2. Experimental

2.1. Catalyst preparation

The synthesis of catalysts containing palladium was conductedbased on previously developed methods [20–22], using aqueoussolution of PdCl2 (2.23 � 10�3 M). The pH of solutions was keptin the range of 3–4 by the addition of suitable amounts of HCl solu-tion (0.1 M). The molar ratio of PdCl2:HCl was 1:5.

The synthesis of catalysts was conducted as follows: to specifiedvolume of PdCl2 solution – an appropriate amount of steamed

zeolite Y (Zeolyst, CBV500, NH4-form, Si/Al = 2.6) was added, andthe obtained suspension was stirred at room temperature usingmagnetic stirrer until the complete adsorption of Pd2+ ions. Com-pleteness of adsorption was colorimetric monitored by the reactionwith thiourea. Then, the solid was filtered and washed with dis-tilled water until Cl� ions were absent in a filtrate. Finally, the sam-ples were dried at room temperature.

Solutions of precursor were prepared in situ. Time necessary tocomplete adsorption of Pd2+ ions was 24 h. Detailed conditions ofthe impregnation procedure of NH4Y zeolite are given in Table 1.

2.2. Characterization methods

It should be mentioned that all characterization methods dis-cussed below were applied to the non-reduced samples (withoutany treatment under air or hydrogen conditions). The only excep-tion was the CO sorption experiment, in which the hydrogen trea-ted samples were used.

2.2.1. Chemical analysis of metal content, low-temperature N2

sorption, and X-ray diffraction studiesPd content was determined by ICP AES analysis (Table 1) using a

high-performance sequential plasma spectrometer (Model ARL3410 ICP). Palladium was marked at wavelengths of 360.955 and340.458 nm. In the first step, the solid samples were digested (ina mixture of HF, HCl and HNO3) in a microwave oven for specialtemperature program. Then, aliquots of solution were diluted toa volume of 100 ml using deionized water. The Pd content in theresulting samples varied from 0.05 to 2.5 wt.%. The sample codesinclude information about the Pd loading.

The BET surface area and pore volume of the samples weredetermined by N2 sorption at �196 �C using a 3Flex (Micromeri-tics) automated gas adsorption system. Prior to the analysis, thesamples were degassed under vacuum at 250 �C for 24 h. The spe-cific surface area (SBET) was determined using Brunauer–Emmett–Teller (BET) model according to Rouquerol recommendations [23].The micropore volume and specific surface area of microporeswere calculated using the Harkins–Jura model (t-plot analysis).

The X-ray diffraction (XRD) patterns of the samples wererecorded with a D2 Phaser diffractometer (Bruker) using Cu Karadiation (k ¼ 1:54060 Å, 30 kV, 10 mA).

2.2.2. IR spectroscopy studies with probe moleculesPrior to FTIR studies, all the samples were pressed into the form

of self-supporting wafers (ca. 5 mg/cm2) and in situ thermally trea-ted in an IR cell at 550 �C under vacuum for 1 h. For carbon mon-oxide (Linde Gas Poland, 3.7) sorption experiment, afterpretreatment, the samples were cooled down to 200 �C and thencontacted with hydrogen (80 Tr in gas phase). Reduction was per-formed at 200 �C for 2 h. After this time, the samples were evacu-ated at 500 �C for 1 h and cooled down to room temperature.

h

g

f

e

d

c

b

a

B

30 35 40 455 15 25 35 45

INTE

NSI

TY /

a.u.

2Θ/degrees 2Θ/degrees

h

g

f

e

d

c

b

a

A

Fig. 1. The XRD patterns in range of 5–45� (A) and 30–45� (B) of native zeolite HY (a) and zeolites modified with different Pd contents 0.05 wt.% Pd (b), 0.1 wt.% Pd (c),0.25 wt.% Pd (d), 0.5 wt.% Pd (e), 1 wt.% Pd (f), 1.5 wt.% Pd (g), 2.5 wt.% Pd (h).

Table 2The textural parameters of zeolites HY and its modifications with palladium.

Zeolite SBET (m2 g�1) Smeso (m2 g�1) Smicro (m2 g�1) Vmicro (cm3 g�1)

HY 1037 95 942 0.3640.05 wt.% Pd/HY 993 93 900 0.3350.1 wt.% Pd/HY 938 88 840 0.3240.25 wt.% Pd/HY 938 88 850 0.3170.5 wt.% Pd/HY 937 87 850 0.3141 wt.% Pd/HY 926 86 840 0.3071.5 wt.% Pd/HY 865 89 776 0.2972.5 wt.% Pd/HY 813 85 728 0.239

38 M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46

Sorption of ammonia (PRAXAIR, 3.6) and nitrogen monoxide(Linde Gas Poland 99.5%) were performed on the vacuum pre-trea-ted but non-reduced samples.

Spectra were recorded with a Bruker Tensor 27 spectrometerequipped with an MCT detector. The spectral resolution was of2 cm�1. The IR spectra were normalized to the same mass of sam-ple (10 mg).

2.3. Catalytic tests

Modified zeolites were tested in the role of catalysts for theselective catalytic oxidation of ammonia (NH3-SCO). The experi-ments were performed under atmospheric pressure in a fixed-bed flow microreactor system (i.d., 7 mm; l., 240 mm). The reactantconcentrations were continuously monitored using a quadrupolemass spectrometer RGA 200 (PREVAC) connected to the reactorvia a heated line. Prior to the reaction, each sample of the catalyst(50 mg diluted with 50 mg of SiO2 (POCH, BET 192 m2 g�1, particlesize of 120–355 lm)) was outgassed in a flow of pure helium atconstant heating rate of 10 �C/min up to 500 �C and then cooleddown to room temperature. The composition of the gas mixtureat the reactor inlet was [NH3] = 0.5 vol.%, [O2] = 2.5 vol.% and[He] = 97 vol.%. For the selected catalysts, additional catalytic testswith the gas mixture containing water vapor were done. In thiscase, the reaction mixture with the following composition of wasused: [NH3] = 0.5 vol.%, [O2] = 2.5 vol.%, [H2O] = 3.2 vol.% and[He] = 93.8 vol.%. Total flow rate of the reaction mixture was40 cm3/min, while a space velocity was 15,400 h�1. Moreover, forthe selected samples, additional catalytic tests with space veloci-ties of 30,800 and 61,600 h�1 were done. The reaction was studiedat temperatures ranging from 100 to 500 �C with the linear tem-perature increase of 10 �C/min. The signal of helium line servedas an internal standard to compensate possible fluctuations ofthe operating pressure. The sensitivity factors of analyzed lineswere calibrated using commercial mixtures of gases.

3. Results and discussion

3.1. Physicochemical properties of the catalysts

The zeolite samples modified with various amounts of palla-dium were characterized with respect to their structure, texture,

surface acid properties, and speciation as well as dispersion of cat-alytically active component.

XRD studies were carried out for both native zeolite HY andmodified zeolites Pd/HY to confirm the stability of the zeolitestructure during the impregnation procedure (Fig. 1A). The pres-ence of reflections typical of the FAU structure in diffractogramsof modified zeolites, which are characterized by almost identicalintensities compared to native zeolite, indicates that impregnationprocedure does not disturb the zeolite Y structure. Fig. 1B depictsthe XRD patterns of native zeolite HY and the Pd-based catalysts.Apart from the diffraction peaks of the zeolite matrix, any reflec-tion characteristic of PdO species was identified in the diffracto-grams of the Pd/HY catalysts, regardless of the different Pdcontents (0.05–2.5 wt.%). Palladium species are supposed to be ofa small particle size; therefore, their reflections cannot be discrim-inated from the zeolite pattern [24]. Most likely, their growth isrestricted by the pore size of zeolite Y. All these findings indicatethat Pd-species are highly dispersed on the support surface or theyare characterized by the amorphous nature [25,26].

Table 2 shows the textural parameters of the zeolite supportand its modifications with palladium. The micropore volume(0.36 cm3 g�1) and BET specific surface area (1037 m2 g�1) deter-mined for HY zeolite are in the range typical of faujasite structure[27]. All textural parameters were preserved after impregnationprocedure. The mesopore’s surface is retained on the same level,whereas a consecutive decrease in both micropore’s surface andvolume clearly evidences the location of Pd-species in microporesof zeolite. For zeolite with the highest Pd loading (2.5 wt.%), boththe micropore volume and specific surface area were significantlyreduced, which can be attributed to the plaguing of micropore

3800 3700 3600 3500 3400 3300 3200

0,0

0,3

0,6

3525

3550

3600

36253670

360536703745f = e- a

e

d

c

bABSO

RBA

NC

E

ν/cm-1

a

Fig. 2. The OH groups spectra of native zeolite HY (a) and zeolites modified withdifferent Pd contents 0.25 wt.% Pd (b), 0.5 wt.% Pd (c), 1 wt.% Pd (d), 2.5 wt.% Pd (e).Difference spectrum (f) represents the loss of hydroxyl groups after impregnation.

Table 3The concentration of Brønsted and Lewis acid sites determined in quantitative IRstudies of ammonia sorption in studied zeolites.

Zeolite CBrønsted (lmol g�1) CLewis (lmol g�1)

HY 960 3800.05 wt.% Pd/HY 960 3850.1 wt.% Pd/HY 930 3900.25 wt.% Pd/HY 900 4040.5 wt.% Pd/HY 850 4271 wt.% Pd/HY 800 4741.5 wt.% Pd/HY 720 5212.5 wt.% Pd/HY 750 615

M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46 39

mouths by agglomerated Pd-species. The formation of the low dis-persed Pd-forms has been evidenced by both IR and catalyticresults (Sections 3.2.2 and 3.3).

As mentioned earlier, the growth of Pd-species could be limitedby the pore size of zeolite Y; thus, any reflection characteristic ofPdO and any other the Pd-containing phases were discriminatedfrom XRD patterns.

3.2. IR studies

As mentioned, the IR spectra were normalized to the same massof sample. It should be noted that two ways of IR spectra normaliza-tion have been developed: to the same sample’s mass and the sameintensity of the overtones’ bands in the 2100–1500 cm�1 frequencyregion (internal standard method) [28]. The internal standard meth-odology is helpful when either the amount of water in the non-acti-vated sample had not been determined before IR studies or thematerials of the evident difference in size of crystals are willing tobe compared. Applied impregnation procedure influenced neitherthe size of zeolite grains nor the zeolite structure; therefore, we stillconsider the same kind of zeolite of the same textural properties (theidentical structure, the same density of T-atoms). Thus, both normal-ization procedures are equivalent. It has been verified by the identi-cal intensity of the overtones’ bands in the IR spectra of Pd-faujasitesnormalized to the same mass of sample.

It should be emphasized that the internal standard method isalso limited and should be carefully applied. For instance, it isnot valid for purely amorphous and composite materials with notfully retained structure. Other examples are dealuminated orsteamed zeolites: the intensity of overtones bands is reported toincrease strongly with the Al atoms extraction. In the later cases,the main condition of the same density of T-atoms in studiedmaterials, required of the internal standard methodology, cannotbe obeyed. Therefore, the use of the position and the shape ofthe 2100–1500 cm�1 bands appear to be doubtful.

3.2.1. The spectra of OH groupsFurther information on the differences between the catalysts

was driven from the IR spectroscopy. Fig. 2 shows the spectra inthe region characteristic of hydroxyl groups in the native zeoliteHY (spectrum a) and zeolites of various Pd loadings: 0.25 wt.% Pd(b), 0.5 wt.% Pd (c), 1 wt.% Pd (d), 2.5 wt.% Pd (e). The spectrumof native zeolite consists of six bands originating from the hydrox-yls groups of different nature. The 3745 and 3670 cm�1 bands areassigned to the isolated Si-OH silanols and to the Al-OH groups inthe extra framework material (EFAl), respectively. The 3550 and3625 cm�1 bands are assigned to the Si(OH)Al groups in sodalitecages and supercages of the zeolite support [29]. These freely oscil-lating bridging hydroxyls are not involved into interaction withextra framework electron-acceptor aluminum species, EFAl. Thebands at 3525 and 3600 cm�1 originate from the vibrations of Si(O-H)Al groups of the same location, but they are perturbed by theinteraction with EFAl species [30].

The analysis of the OH groups’ spectra of the catalysts pointsto a gradual decline in the intensity of the bands both sila-nols (3745 cm�1) and the bridging Si(OH)Al groups (3650–3500 cm�1). The difference spectrum (f) showing the changes inthe intensities of the hydroxyl bands of native zeolite (support)and zeolite with the highest Pd loading points to a preference inthe hydroxyl groups’ consumption. The drop of the intensity ofthe 3600 cm�1 Si(OH)Al bands evidences the reduction in the num-ber of strongly acidic Brønsted centers; thus, protons balancing thenegative framework of zeolite are believed to be substituted for thepositively charged PdxOy

+ forms. These conclusions were con-firmed by the quantitative studies of ammonia sorption (nextchapter).

3.2.2. IR quantitative studiesAmmonia molecule, together with pyridine, is usually used as a

probe for the quantification of both Brønsted and Lewis sites insolid acid catalysts [31]. Small kinetic diameter of ammonia pro-vides opportunity for detecting each acid site in zeolite Y eventhose hidden is cubooctahedra and hexagonal prisms. Interactionof ammonia with Brønsted and Lewis acid sites results in thedevelopment of the 1450 cm�1 and 1635 cm�1 band, respectively.The concentrations of both Brønsted (NH4

+) and Lewis (NH3L) acidsites were calculated based on the maximum intensities of theNH4

+ and NH3L bands as well as corresponding values of the extinc-tion coefficients which were determined as described in Ref. [32].

Table 3 gathers the concentration of Brønsted and Lewis acidsites in zeolite HY and in its Pd-modified forms determined in IRquantitative measurements of ammonia sorption. A decrease inthe amount of Brønsted centers is clearly visible. For zeolite withthe highest Pd content, a number of protonic sites were reducedby 22% in comparison with the support. This decrease in the con-tent of strongly acidic Brønsted centers is in line with the drop inthe intensity of the 3600 cm�1 Si(OH)Al bands previously detected.Along with a decline in the amount of protonic centers, an increasein Lewis acid centers concentration is observed. These electron-acceptor centers originate from the oxide forms of palladium.Again, it can be concluded that the positively charged PdxOy

+ spe-cies replaced the protons previously balancing the negative chargeof the zeolite framework.

Comparison of the surface concentration of Pd-centers (inlmol g�1) determined in quantitative IR measurements of ammo-nia sorption and total Pd concentration in the samples, obtainedfrom chemical analysis (Fig. 3), allows calculating the contributionof Pd-centers accessible for the catalytic process, and consequently,measuring the dispersion of Pd-species. Taking into account ahigh stability of steamed zeolite Y and its resistance against

0 5 1024

47

94

141

235

0 0 1030

5070

120

160

Concentration of Pd /μmolg-1

ICPPy sorption (IR)

Fig. 3. The comparison of the concentration of Pd-centers driven form IR exper-iments of NH3 sorption and chemical analysis for native zeolite HY and zeolitesmodified with different Pd contents.

2200 2100 2000 1900

0,0

0,2

0,4

2130

f

1940

ed

c

ba

1995

1988

2101

2108

2121

ABSO

RBA

NC

E

ν /cm-1

2172

Fig. 4. The CO spectra adsorbed (in �140 �C) on native HY zeolite (a) and 2.5 wt.%Pd/HY zeolite non-reduced (b) and zeolites after H2 reduction performed at 200 �C Kfor 2 h (c) 0.5 wt.% Pd/HY, (d) 1 wt.% Pd/HY, (e) 1.5 wt.% Pd/HY, (f) 2.5 wt.% Pd/HY.

40 M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46

dealumination, it was assumed that impregnation of zeolite withPd did not result in the generation other Lewis centers than thoseoriginating from Pd-forms. The difference in the concentrations ofLewis centers in Pd-modified zeolites and native support is relatedto various amounts of palladium centers accessible for ammoniamolecules. The Pd concentrations calculated from the chemicalanalysis and those obtained from IR measurements (Fig. 3) arenearly identical for the catalysts with metal content in the rangefrom 0.05 to 1.5 wt.%, which clearly evidences very high dispersionof Pd-oxide forms deposited on the zeolite support. Only for thecatalyst with the highest surface Pd enrichment (2.5 wt.%), adecline in the quantities of centers available for ammonia moleculewas observed: the number of available Pd-centers provides 68% ofall possible adsorption centers. This finding indicates the formationof palladium agglomerates that are responsible for the significantdrop of micropore volume detected for zeolite of the highest Pdloading (Section 3.1). Moreover, this low Pd dispersion is supposedto be reflected in the catalytic performance of studied zeolites.

3.2.3. Carbon monoxide sorptionCarbon monoxide is a probe molecule widely employed to elu-

cidate the electron donor species speciation, both in a qualitativeand in a quantitative aspect [33–40]. Interaction of the carbonmonoxide molecules with the acid centers present in native zeoliteHY leads to the appearance of the 2172 cm�1 band associated withthe formation of OH-CO adducts (Fig. 4, the spectrum a). Sorptionof CO on as-made 2.5 wt.% Pd/HY, presented here as an example ofwhole series of the as-made Pd-catalysts (spectrum b), showedthat the only 2130 cm�1 band attributed to Pdn+ ions in oxideforms. Due to the method applied for palladium deposition, onlyPd-centers of this nature tend to occur on the surface of zeoliteHY. The bands corresponding to the presence of metallic Pd-spe-cies on the surface of the non-reduced catalysts were not found.

To evaluate the population of Pd-centers on catalysts’ surface,the reduction of Pd-oxide species to metallic Pd0 was carried outwith hydrogen at 200 �C for 2 h. For the reduced catalysts contain-ing 0.5–2.5 wt.% Pd (Fig. 4, spectra c–e), interaction of CO with Pd-centers resulted in the 2120–1988 cm�1 bands’ development thatare typically assigned to the metallic Pd0 forms. In the literature,the bands in the 2120–2000 cm�1 frequency region have beenattributed to linear Pd0(CO) carbonyls [41], whereas the featuresat 1995 and 1988 cm�1 have associated with Pd0

2(CO)2 adducts[41,42]. Broad bands of low intensities observed at 1940 cm�1 havebeen assigned to CO molecules adsorbed in the bridged forms on Pdmetallic species [43].

The intensity of the Pd0(CO) band (2120–1940 cm�1) was usedfor the estimation of Pd0 species dispersion. Up to 1.5 wt.% Pd,

the integral intensities of all the bands ascribed to the presenceof Pd0 forms increased with the amount of Pd impregnated onthe zeolite matrix. For the 1.5 wt.% Pd/HY catalyst, the highest sur-face Pd enrichment is suggested. Zeolite with the highest Pd con-tent (2.5 wt.%) is characterized by the Pd carbonyl bands of theintensity comparable to the total intensities of bands in the1.5 wt.% Pd/HY catalyst. This proves lowered dispersion of Pdmetallic particles for the catalyst with the highest Pd content dueto their agglomeration. With increasing Pd amounts, the natureof the metallic centers was also changed. For zeolites HY with0.5–1.5 wt.% Pd loadings, the most typical were isolated linearforms of Pd0(CO) carbonyls of the band frequencies higher thanfor zeolite with the highest Pd content (2.5 wt.%). This evidencesthe stronger electron-acceptor properties of Pd0 forms in the 0.5–1.5 wt.% Pd-catalysts than those found in zeolite 2.5 wt.% Pd/HY.Consequently, for 2.5 wt.% Pd/HY zeolite, not only a decrease indispersion of Pd metallic species but also drop in electron-acceptorproperties of metallic centers were observed, probably due toagglomeration and the formation of larger clusters of metallic pal-ladium. For Pd contents lower than 0.5 wt.%, the presence of Pd0

forms was not observed, pointing to the location of these formsin positions inaccessible to CO molecules, i.e. in cubooctahedra orhexagonal prisms.

3.3. Catalytic activity

The modified zeolite samples were tested in the role of the cat-alysts for the selective oxidation of ammonia to N2 and water vapor(reaction 1). Fig. 5 presents the results of activity experiments per-formed without catalyst (empty reactor) and over pure SiO2. Inboth cases, ammonia conversion started at temperatures as highas 375 �C and at 500 �C did not exceed 20%. The results of the cat-alytic tests performed for native zeolite HY and its derivatives withpalladium are shown in Fig. 6. The ammonia conversion profile forHY was similar to that obtained for SiO2, indicating that this sam-ple is not active in the NH3-SCO process. Nitrogen was the mainproduct of ammonia oxidation; however, also significant contribu-tion of NO and N2O in the reaction products was found. Deposition

0

20

40

60

80

100

Temperature /°C

NH3 conversion N2 selectivity NO selectivity N2O selectivity

B

0

20

40

60

80

100

CO

NVE

RSI

ON

/SEL

ECTI

VITY

/%

Temperature /°C

A

100 200 300 400 500100 200 300 400 500

Fig. 5. The results of activity tests for the NH3-SCO process performed with an empty reactor (A) and reactor containing pure SiO2 (B).

M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46 41

of palladium on HY resulted in its catalytic activation in the NH3-SCO process. It should be noted that an increase in palladium load-ing resulted in the activation of the catalysts at lower temperaturesand additionally gradually decreased selectivity of ammonia oxida-tion to nitrogen in the high-temperature range. The contribution ofnitrogen oxides, which were the side reaction products, increasedwith an increasing noble metal loading.

The ammonia conversion curves for all the studied samples arecompared in Fig. 7. As it was mentioned above, NH3 conversionincreased with an increase in palladium loading. The catalysts con-taining 1.5 and 2.5 wt.% of palladium exhibited nearly the sameNH3 conversion profile and in particular maintained a high levelof activity. Ammonia oxidation measured in the presence of thesecatalysts started above 125 �C and temperature of 250 �C was suf-ficient for its complete conversion in the reaction mixture, exhibit-ing 89–95% selectivity to N2 below 250 �C. The same catalyticbehavior evidenced for two zeolites of the highest Pd loadings(1.5 and 2.5 wt.%) can be related to lower dispersion of active spe-cies in the 2.5 wt.% Pd/HY sample as it was previously anticipatedfrom quantitative IR studies and low-temperature nitrogensorption.

The results of the catalytic tests are summarized in Table 4,which presents temperatures needed for 50% and 100% of ammo-nia conversion as well as selectivity to N2 at these temperatures.High activity of palladium modified zeolites in selective ammoniaoxidation was previously reported; however, in majority of thesestudies, zeolite ZSM-5 was tested in the role of noble metal support[1,10,11]. These catalysts effectively operated at temperaturesabove 250 �C with selectivities to nitrogen about 70–90%. More-over, catalytic systems based on metal oxides and noble metaldeposited on metal oxide supports, active in the process of selec-tive ammonia oxidation, were reported in the scientific literature.Gang et al. [3] studied various noble metals deposited on Al2O3

in the role of the catalysts for the NH3-SCO process. High activityof the Ag/Al2O3 catalyst in the low-temperature range wasreported; however, selectivity to nitrogen was not satisfactory(100% of NH3 conversion with selectivity to N2 of 82% at 160 �C).Copper deposited on titania (Cu/TiO2) was another very active cat-alytic system for selective ammonia oxidation [44]. In the presenceof this catalyst, ammonia was completely oxidized in the reactionmixture, with high selectivity to N2 (95%) at 250 �C. Interesting cat-alytic properties in the process of selective ammonia oxidationwere reported for CuO deposited on RuO2 [45] and carbon nano-tubes [46]. In case of the former catalyst, 100% of ammonia conver-sion with selectivity to nitrogen of 97% was obtained at 160 �C.While In the presence of the second catalyst, ammonia was com-pletely oxidized in the reaction mixture with selectivity to N2 of99% at 190 �C. However, it must be pointed out that these catalysts

are relatively expensive, and therefore, their commercialization ishindered. Extended review of various catalytic systems, tested inthe process of the selective ammonia oxidation, is presented inSupplementary File. Comparison of the results obtained for themost active and selective catalysts, reported in the scientific liter-ature and zeolite Y doped with palladium, which was the subject ofour research, shows that the studied catalysts belong to the groupof the most active and selective catalytic systems for the low-tem-perature ammonia oxidation.

For the 1.5 wt.% Pd/HY and 2.5 wt.% Pd/HY catalysts, additionalcatalytic tests with the gas mixture containing water vapor weredone. Results of these studies are presented in Fig. 8. As it can beseen, an introduction of water vapor into the reaction mixturedecreases ammonia conversion, but this effect is not very strongand total NH3 oxidation in wet atmosphere can be obtained at300 �C. On the other hand, the comparison of selectivities obtainedfor the process performed in dry (Fig.6) and wet (Fig. 8) conditionsevidences that ammonia is more selectively oxidized to N2 even ifwater vapor was introduced into the gas mixture. Thus, introduc-tion of water into reaction mixture decreased ammonia conversionbut increased selectivity to nitrogen, which is a desired reactionproduct.

In general, it can be concluded that an increase in palladiumloading decreased temperature of effective catalyst operation, pos-sibly by increased number of active sites, but also decreases selec-tivity to N2, which is a desired product of ammonia oxidation. As itwas shown by NH3-sorption studies and BET analysis, more aggre-gated noble metal species were formed in the samples with higherpalladium loading. It seems possible that such species are respon-sible for deeper ammonia oxidation to nitrogen oxides. Anotherexplanation could be related to the assumption that the processof ammonia oxidation proceeds according to the i-SCR mechanism.It seems possible that for the catalysts with higher palladium load-ing, ammonia is nearly completely oxidase to NO (reaction 2), andtherefore, there is not enough ammonia for the conversion of NO toN2 (reaction 3a) and N2O (reaction 3b).

Fig. 9 presents turnover number (TON) of ammonia conversionrelated to surface palladium atoms (determined by NH3-adsorp-tion studies), which were assumed to be active sites in the studiedprocess. It can be seen that TON values decrease with increasingloading of noble metal. As it was shown by NH3-adsorption studiesand BET analysis, an increase in the palladium content resulted in agradual aggregation of noble metal species. Thus, the results pre-sented in Fig. 9 show that more aggregated palladium species wereless catalytic active than monomeric noble metal species intro-duced into zeolite Y. These results are supported by the FTIR stud-ies of the CO-pre-adsorbed samples, which show that aggregatedpalladium species are characterized by lower electron-acceptor

0

20

40

60

80

100

NH3 conversion N2 selectivity NO selectivity N2O selectivity

0.5 wt.% Pd/HY

0

20

40

60

80

100

CO

NVE

RSI

ON

/SEL

ECTI

VITY

/% HY

0

20

40

60

80

100

CO

NVE

RSI

ON

/SEL

ECTI

VITY

/% 0.05 wt.% Pd/HY

0

20

40

60

80

100

1 wt.% Pd/HY

0

20

40

60

80

100

0.1 wt.% Pd/HY

CO

NVE

RSI

ON

/SEL

ECTI

VITY

/%

0

20

40

60

80

100

1.5 wt.% Pd/HY

0

20

40

60

80

100

CO

NVE

RSI

ON

/SEL

ECTI

VITY

/%

Temperature /°C

0.25 wt.% Pd/HY

100 200 300 400 500100 200 300 400 500

100 200 300 400 500 100 200 300 400 500

100 200 300 400 500 100 200 300 400 500

100 200 300 400 500 100 200 300 400 5000

20

40

60

80

100

Temperature /°C

2.5 wt.% Pd/HY

Fig. 6. The results of activity tests for the NH3-SCO process performed over x wt.% Pd/HY catalysts.

42 M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46

properties. Thus, it is possible that catalytic activation of ammoniaadsorbed on such centers is less effective than for monomeric pal-ladium species.

3.4. Studies of the reaction mechanism

In order to determine the reaction mechanism, additional stud-ies were performed. For the most promising samples (1.5 wt.% Pd/

HY and 2.5 wt.% Pd/HY), catalytic tests with various space veloci-ties were done. Results of these studies are presented in Fig. 10.As it was expected, an increase in space velocity resulted in adecrease of ammonia conversion. Moreover, it was observed thatfor the catalytic tests performed with increased space velocity,the selectivity to N2 decreased and selectivity to NO increased,especially at higher temperatures. These results fully support theproposed i-SCR mechanism. For the tests performed with a rela-

0

20

40

60

80

100N

H3 C

ON

VER

SIO

N /%

Temperature /°C

0.05 wt.% Pd 0.1 wt.% Pd 0.25 wt.% Pd 0.5 wt.% Pd 1 wt.% Pd 1.5 wt.% Pd 2.5 wt.% Pd

Fig. 7. The comparison of the results of catalytic tests performed over the x wt.% Pd/HY catalysts.

Table 4Comparison of the results of catalytic tests (T50% and T100% temperatures needed for50% and 100% of NH3 conversion, respectively).

Zeolite T50%

(�C)N2 selectivityat T50% (�C)

T100%

(�C)N2 selectivityat T100% (�C)

HY – – – –0.05 wt.% Pd/HY 304 100 450 950.1 wt.% Pd/HY 289 100 400 950.25 wt.% Pd/HY 258 97 400 930.5 wt.% Pd/HY 236 97 350 951 wt.% Pd/HY 220 98 300 961.5 wt.% Pd/HY 176 95 250 902.5 wt.% Pd/HY 174 95 250 95

0 50 100 1500

1

2

3

4

TON

/s-1

Concentration of surface Pd species /μmolg-1

Fig. 9. Turnover number (TON) of ammonia conversion over surface palladiumatoms determined at 225 �C.

M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46 43

tively low space velocity, the contact time of reactants with cata-lyst surface is long enough for both reaction steps – ammonia oxi-dation to NO (reaction 2) and NO reduction by unreacted ammonia(reaction 3a). In contrast, for experiments performed withincreased space velocity, the contact time was too short for effec-tive reduction of NO by ammonia, and therefore, selectivity toNO increased, while selectivity to N2 decreased.

Moreover, for zeolites 1.5 wt.% and 2.5 wt.% Pd/HY, additionalstudies of ammonia oxidation reaction monitored by IR spectros-copy were done. Fig. 11A depicts the spectrum of ammonia sorbedat 130 �C in native zeolite HY (spectrum a) and 1.5 wt.% Pd/HY zeo-lite (spectrum b). In the next step, the catalyst with pre-adsorbedammonia was heated to 250 �C (spectrum c) and 300 �C (spectrum

0

20

40

60

80

100

NH3 conversion (dry) N2 selectivity (wet) NO selectivity (wet) N2O selectivity (wet) NH3 conversion (wet)

CO

NVE

RSI

ON

/SEL

ECTI

VITY

/%

Temperature /°C

1.5 wt. % Pd/HY

100 200 300 400 500

Fig. 8. The results of catalytic tests performed in wet atmosphere for 1.5 wt.% Pd/HY anddry atmosphere are shown – open symbols.

d) for 30 min. and then cooled to 130 �C for the detection ofadsorbed species formed by ammonia oxidation. In the spectrumof ammonia adsorbed on native zeolite HY (Fig. 9, the spectruma), the 1430 cm�1 band of ammonium ions NH4

+ appeared as themost characteristic moiety attributed to reaction of ammonia withBrønsted sites (strongly acidic Si(OH)Al groups). A small intensityof the 1635 cm�1 band is associated with the presence of a certainamount of ammonia chemisorbed on Lewis centers (electron-acceptor aluminum species, EFAl). Thus, framework Brønsted acidsites and and/or acid centers associated with extra framework alu-mina are responsible for the increased activity of zeolite Y in com-parison with silica (cf. Figs. 6 and 7).

Impregnation of zeolite with Pd resulted in further develop-ment of the 1635 cm�1 band due to generation of some amountof electron-acceptor centers found as Pd-oxide forms. Furthermore,a consumption of the ammonium ion band (30% in the initial inten-sity of the NH4

+ band) is in line with the substitution of someammonium ions for Pd-oxide forms, which balance the negativecharge of the zeolite framework. Such Pd-oxide forms are supposedto be catalytically active in the process of selective ammoniaoxidation.

Heating of 1.5 wt.% Pd/HY with ammonia pre-adsorbed up to250 �C (spectrum c) resulted in an appearance of new bands inthe 2224–1230 cm�1 frequency region originating from the ammo-nia oxidation products. Simultaneously, a slight decrease in inten-sity of the band related to ammonia bonded to Pd-oxide specieswas detected. All these findings indicated that ammonia associatedwith Pd-centers was oxidized. Among the products of this reaction,NO (the 1875 cm�1 band), NO2 (1305 and 1270 cm�1 bands), N2O

0

20

40

60

80

100

NH3 conversion (dry) N2 selectivity (wet) NO selectivity (wet) N2O selectivity (wet) NH3 conversion (wet)

CO

NVE

RSI

ON

/SEL

ECTI

VTY

/ %

Temperature /°C

2.5 wt. % Pd/HY

100 200 300 400 500

2.5 wt.% Pd/HY. For comparison also NH3 conversions obtained in tests performed in

Fig. 10. The results of catalytic tests performed with various space velocities for 1.5 wt.% Pd/HY (left) and 2.5 wt.% Pd/HY (right).

44 M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46

(2224 cm�1), and NO3� ions (the 1415 cm�1 band) were identified

[39,40,47,48]. Independently, the 1415 cm�1 band could beassigned to N–H vibrations in –NHx hydrazine-type compounds,whereas the 1560 cm�1 band was associated with either bidentatenitrates or –HNO nitroxyl species [39,40,47,48].

Additional experiment of NO adsorption on the 1.5 wt.% Pd/HYcatalyst surface monitored by FTIR spectroscopy (Fig. 9, spectrume) showed that, in the absence of oxygen, NO can be depositedon the surface in the form of NO3

� ions (the 1415 cm�1 band) point-ing to high activity of oxygen atoms in PdOx species.

Concluding, adsorbed nitrogen oxides (NO, NO2, N2O) and NO3�

ions as well as chemisorbed ammonia species were found on thecatalyst surface at temperature 250 �C. It should be noted that

catalytic test performed over the 1.5 wt.% Pd/HY sample at thistemperature resulted in a very high ammonia conversion andselectivity to N2. Thus, the i-SRC mechanism of this process, includ-ing reaction between nitrogen oxides and chemisorbed ammonia(reactions 3a and 3b), cannot be in this case excluded. On the otherhand, possible hydrazine-type species found on the catalyst sur-face cannot also exclude the hydrazine mechanism.

An increase in the reaction temperature to 300 �C (spectrum d)resulted in disappearance of the NO and NO2 bands assigned toammonia oxidation. At the same time, intensity of the bandsrelated to the presence of NO3

�and NO2� ions increased. Moreover,

intensity of the band at 1635 cm�1 related to ammonia chemi-sorbed on Pd-oxide forms was significantly reduced, which means

0,0

0,2

0,4

0,6

0,8

1,0

e

15601915

13051270

1270

1415

1430

16351875

d

c

b

a

ABSO

RBA

NC

E

ν/cm-1 ν/cm-1

2100 1800

1915

c

2224 1875

A

2250 2000 1750 1500 1250 2250 2000 1750 1500 1250

0,0

0,2

0,4

0,6

0,8

b

a

1270

1415

2100 1800

b

1915

a

2224 1875

B

Fig. 11. (A) The spectra of NH3 adsorbed (in 130 �C) on native zeolite HY (a) and 1.5 wt.% Pd/HY zeolite (b). The spectra (c) and (d) were collected for 1.5 wt.% Pd/HY zeolitewith the pre-adsorbed NH3, heated to 250 �C and 300 �C (for 0.5 h), resp., and next cooled down to RT. Spectrum e were registered at RT for 1.5 wt.% Pd/HY zeolite after thetreatment with NO at 300 �C for 0.5 h. (B) The IR spectra collected for 1.5 wt.% Pd/HY (a) and 2.5 wt.% Pd/HY (b) zeolites heated with pre-adsorbed NH3 to 300 �C (for 0.5 h),and next cooled down to RT.

M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46 45

that such ammonia forms are very reactive and were nearlycompletely converted to the reaction products. If it will beassumed that reaction proceeds according to the i-SCR mecha-nism, drop in selectivity to N2 observed at temperatures above300 �C (cf. Fig. 7) could be explained by the lack of properly acti-vated ammonia for the reduction of nitrogen oxides (reactions 3aand 3b).

Similar behavior was identified for zeolite of the highest Pdloading. Heating of 2.5 wt.% Pd/HY with ammonia pre-adsorbedup to 300 �C (Fig. 11B, spectrum b) resulted in the developmentof the bands at 1875 cm�1 (NO), 1270 cm�1 (NO2), 2224 cm�1

(N2O), and 1415 cm�1 (NO3�). However, the ammonia oxidation

product bands are of the same intensities as those detected for zeo-lite with 1.5 wt.% Pd (Fig. 11B, spectrum a). Thus, the loweredactivity of the 2.5 wt.% Pd/HY catalyst in the SCO NH3 process isidentified with the presence of less active PdOx agglomerates.

The finely dispersed PdOx species are supposed to be the mainadsorbed sites of NH3 molecules, and the NH3(ad-Lewis) could be fur-ther activated and transformed into ammonia oxidation products(NO and NO2). Nevertheless, the role of ammonia bonded toBrønsted acid sites (i.e. NH4

+ ions) as the additional reservoir ofammonia chemisorbed cannot be excluded [49–52]. Probably highselectivity of the catalysts to nitrogen, especially in the high-tem-perature range, is related to protection of ammonia chemisorbedon the Brønsted acid sites of the zeolite framework against oxida-tion and therefore its accessibility to reduction of nitrogen oxides(i-SCR). Moreover, Brønsted acidic sites are expected to play animportant role in the isomerization of H2NO2 to HNOOH [51], fol-lowed by dehydration to yield NO. The H2NO2 intermediates arestabilized on Brønsted acidic sites, which involves in proton shuf-fling to enhance the NO production. In this way, NO is the primaryproduct of ammonia oxidation in an oxidizing environment and N2

can be eventually generated via the i-SCR reaction.In the 2.5 wt.% Pd/HY sample, the plaguing of micropores by

aggregated PdOx species limited also the accessibility of Brønstedacid sites for ammonia molecules. Consequently, the additionalreservoir of chemisorbed ammonia (NH4

+ ions) was strongly

reduced; thus, catalytic efficiency improvement, expected for zeo-lite with the highest Pd loading, was not effective route.

4. Conclusions

Zeolite Y doped with palladium was found to be active andselective catalyst for the process of ammonia oxidation to N2 andH2O. The catalytic activity of modified zeolites increased with anincrease in noble metal loading. Opposite order was found forselectivity to nitrogen, which decreased with an increase in palla-dium content. Detailed studies of dispersion and aggregation ofpalladium species deposited on the zeolite surface have shown thatfor higher noble metal loading part of them exist in the form ofaggregated species, which were found to be less active and selec-tive in comparison with more dispersed palladium species. Studiesof the reaction mechanism have suggested that palladium oxidespecies (PdOx) are active sites for ammonia oxidation to nitrogenoxides, while part of ammonia is stabilized against oxidation overthe zeolite framework acid site. Therefore, it seems that the pro-cess of selective ammonia oxidation over zeolites modified withpalladium proceeds according to the i-SCR mechanism betweenNO and NO2 formed over PdOx species and ammonia chemisorbedon acid sites of the zeolite framework.

Acknowledgments

This work was financed by Grant No. 2013/09/B/ST5/00066from the National Science Centre, Poland. The research was par-tially carried out with the equipment purchased thanks to thefinancial support of the European Regional Development Fund inthe framework of the Polish Innovation Economy Operational Pro-gram (Contract No. POIG.02.01.00-12-023/08).

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.jcat.2014.04.022.

46 M. Jabłonska et al. / Journal of Catalysis 316 (2014) 36–46

References

[1] R.Q. Long, R.T. Yang, Catal. Lett. 78 (2002) 353.[2] D.P. Sobczyk, E.J.M. Hensen, A.M. de Jong, R.A. van Santen, Top. Catal. 23 (2003)

109.[3] L. Gang, B.G. Anderson, J. van Grondelle, R.A. van Santen, Appl. Catal. B 40

(2003) 101.[4] L. Chmielarz, M. Jabłonska, A. Struminski, Z. Piwowarska, A. Wegrzyn, S.

Witkowski, M. Michalik, Appl. Catal. B 130–131 (2013) 152.[5] M. Amblard, R. Burch, B.W.L. Southward, Appl. Catal. B 22 (1999) L159.[6] R.Q. Long, R.T. Yang, J. Catal. 207 (2002) 158.[7] R.Q. Long, R.T. Yang, J. Catal. 201 (2001) 145.[8] A. Akah, C. Cundy, A. Garforth, Appl. Catal. B 59 (2005) 221.[9] N.I. Il’chenko, Russ. Chem. Rev. 45 (1976) 1119.

[10] Y. Li, J.N. Armor, Appl. Catal. B 13 (1997) 131.[11] R.Q. Long, R.T. Yang, Chem. Commun. (2000) 1651.[12] M. Dams, L. Drijkoningen, B. Pauwels, G. van Tendeloo, D.E. de Vos, P.A. Jacobs,

J. Catal. 209 (2002) 225.[13] G. Koyano, S. Yokoyama, M. Misono, Appl. Catal. A 188 (1999) 301.[14] J. Zawadzki, Discuss. Faraday Soc. 8 (1950) 140.[15] J.M.G. Amores, V.S. Escribano, G. Ramis, G. Busca, Appl. Catal. B 13 (1997) 45.[16] L.I. Darvell, K. Heiskanen, J.M. Jones, A.B. Ross, P. Simell, A. Williams, Catal.

Today 81 (2003) 681.[17] G. Qi, J.E. Gatt, R.T. Yang, J. Catal. 226 (2004) 120.[18] L. Zhang, H. He, J. Catal. 268 (2009) 18.[19] L. Chmielarz, A. Wegrzyn, M. Wojciechowska, S. Witkowski, M. Michalik, Catal.

Lett. 141 (2011) 1345.[20] A. Drelinkiewicz, M. Hasik, M. Kloc, J. Catal. 186 (1999) 123.[21] M. Hasik, A. Drelinkiewicz, M. Choczynski, S. Quilard, A. Pron, Synthetic Met.

84 (1997) 93.[22] M. Hasik, A. Bernasik, A. Drelinkiewicz, K. Kowalski, E. Wenda, J. Camra, Surf.

Sci. 507–510 (2002) 916.[23] J. Rouquerol, P. Llewellyn, F. Rouquerol, Stud. Surf. Sci. Catal. 160 (2007) 49.[24] J. Huang, C.J. Xue, B.F. Wang, X.Z. Guo, S.R. Wang, React. Kinet. Mech. Catal. 108

(2013) 403.[25] H.Y. Zhang, B. Dai, X.G. Wang, L.L. Xu, M.Y. Zhu, J. Ind. Eng. Chem. 18 (2012) 49.[26] H.Y. Zhang, B. Dai, X.G. Wang, W. Li, Y. Han, J.J. Gu, J.L. Zhang, Green Chem. 15

(2013) 829.[27] http://www.iza-structure.org/databases/.[28] J.A. Lercher, Ch. GriJndling, G. Eder-Mirth, Catal. Today 27 (1996) 353.

[29] H.G. Karge, E. Geidel, in: H.G. Karge, P. Behrens, J. Weitkamp (Eds.), MolecularSieves, Science and Technology, vol. 4, Springer-Verlag, Berlin Heidelberg,2004, p. 1.

[30] N. Malicki, P. Beccat, P. Bourges, Ch. Fernandez, A.-A. Quoineaud, L.J. Simon, F.Thibault-Starzyk, in: R. Xu, J. Chen, Z. Gao, W. Yan (Eds.), Studies and SurfaceScience and Catalysis, From Zeolites to Porous MOF Materials – The 40thAnniversary of International Zeolite Conference, vol. 170 A, 2007, p. 762.

[31] J. Datka, K. Góra-Marek, Catal. Today 114 (2006) 205.[32] K. Góra-Marek, M. Derewinski, J. Datka, P. Sarv, Catal. Today 101 (2005) 131.[33] K. Góra-Marek, B. Gil, J. Datka, Appl. Catal. A 353 (2009) 117.[34] K. Góra-Marek, A.E. Palomares, A. Glanowska, K. Sadowska, J. Datka, Micropor.

Mesopor. Mater. 162 (2012) 175.[35] K. Chakarova, M. Mihaylov, K. Hadjiivanov, Micropor. Mesopor. Mater. 81

(2005) 305.[36] K. Hadjiivanov, E. Ivanova, H. Knözinger, Micropor. Mesopor. Mater. 58 (2003)

225.[37] E. Ivanova, K. Hadjiivanov, S. Dzwigaj, M. Che, Micropor. Mesopor. Mater. 89

(2006) 69.[38] K. Góra-Marek, A. Glanowska, J. Datka, Micropor. Mesopor. Mater. 158 (2012)

162.[39] E. Ivanova, M. Mihaylov, F. Thibault-Starzyk, M. Daturi, K. Hadjiivanov, J. Mol.

Catal. A 274 (2007) 179.[40] P. Pietrzyk, Z. Sojka, in: B. Delmon, J.T. Yates (Eds.), Stud. Surf. Sci. Catal. 171

(2007) 27 (and references therein).[41] M. Primet, L.C. De Menorval, J. Fraissard, T. Ito, J. Chem. Soc. Faraday Trans. 1

(81) (1985) 2867.[42] Y. Yu, O.Y. Gutiérrez, G.L. Haller, R. Colby, B. Kabius, J.A. Rob van Veen, A.

Jentys, J.A. Lercher, J. Catal. 304 (2013) 135.[43] T. Rades, V.Y. Borokov, V.B. Kazansky, M. Polisset-Thfoin, J. Fraissard, J. Phys.

Chem. 100 (1996) 16238.[44] S. He, Ch. Zhang, M. Yang, Y. Zhang, W. Xu, N. Cao, H. He, Sep. Purif. Technol. 58

(2007) 173–178.[45] X. Cui, J. Zhou, Z. Ye, H. Chen, L. Li, M. Ruan, J. Shi, J. Catal. 270 (2010) 310–317.[46] S. Song, S. Jiang, Appl. Catal. B 117–118 (2012) 346–350.[47] T. Venkov, K. Hadjiivanov, D. Klissurski, Phys. Chem. Chem. Phys. 4 (2002)

2443.[48] K. Hadjiivanov, Catal. Rev. Sci. Eng. 42 (2000) 71 (and references therein).[49] N.-Y. Topsøe, Science 265 (1994) 1217.[50] N.-Y. Topsøe, H. Topsøe, J.A. Dumesic, J. Catal. 151 (1995) 226.[51] N.-Y. Topsøe, H. Topsøe, J.A. Dumesic, J. Catal. 151 (1995) 241.[52] R.-M. Yuan, G. Fu, X. Xu, H.-L. Wan, J. Phys. Chem. C 115 (2011) 21218.