Applied Catalysis A: General 411– 412 (2012) 131– 138
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
j ourna l ho me page: www.elsev ier .com/ locate /apcata
elective catalytic oxidation of ammonia to nitrogen over ceria–zirconia mixedxides
hong Wang, Zhenping Qu ∗, Xie Quan, Hui Wangey Laboratory of Industrial Ecology and Environmental Engineering, School of Environmental Sciences and Technology, Dalian University of Technology, Dalian, 116024, China
r t i c l e i n f o
rticle history:eceived 11 July 2011eceived in revised form 17 October 2011ccepted 19 October 2011vailable online 25 October 2011
eywords:
a b s t r a c t
The selective catalytic oxidation of ammonia to nitrogen (NH3-SCO) has been studied over ceria–zirconiamixed oxides (Ce1−xZrxO2). The addition of Zr into ceria leaded to the phase transition from the cubicfluorite structure to the tetragonal structure and the generation of oxygen vacancies. The Ce1−xZrxO2
(0.2 � x � 0.8) mixed oxide catalysts exhibited certain amount of moderate acid sites. Particularly, theCe0.4Zr0.6O2 catalyst achieved the largest amount of moderate acid sites and highest proportion of Ce3+
(oxygen vacancies). Meanwhile, it showed the best NH3 oxidation activity, and the complete conversion◦
elective catalytic oxidation of ammoniae1−xZrxO2 catalystshase transitioncid sites
temperature was about 360 C. In addition, numerical results also indicated that the higher zirconiumcontent in Ce1−xZrxO2 catalysts improved the N2 selectivity which was about 100% with x > 0.4. The for-mation of N2O was the main reason for low N2 selectivity over these catalysts (x � 0.4) proved by TPSRexperiments. Meanwhile, the reaction mechanism of NH3 with lattice oxygen was observed to be differ-ent from the gaseous oxygen, and the gaseous oxygen species was much more active than lattice oxygenfor NH3 oxidation.
. Introduction
The treatment of ammonia from waste streams is becoming anmportant issue due to ever increasing environmental concerns.t is known that there are a lot of chemical processes that usemmonia as a reactant or produce ammonia as a by-product. Theyll act as potential sources of NH3 slips [1]. Thus, the removal ofmmonia from waste streams is becoming an increasingly impor-ant issue. In order to control the ammonia slip, several reviewsf different techniques used for the elimination of ammonia haveeen published, such as adsorption, absorption, chemical treat-ent, catalytic decomposition, and selective catalytic oxidation
SCO) [2]. Selective catalytic oxidation of ammonia to nitrogens potentially an ideal technology for removing ammonia fromxygen-containing waste gases, and consequently it is of increasingnterest in recent years [1–11].
So far, a variety of studies have been conducted to seek efficientatalysts with low cost and enhanced catalytic capability for NH3xidation. Several types of material, such as noble metal (Ag, Pt, Pd,u, Ir) [2,4,6,12], transition metal oxides (CuO, Fe2O3, MnO2, Co3O4,
oO3, NiO) [3,6,10,13,14], and rare earth or mixed oxides (CeO2,g–Al–O) [15,16], have been developed for SCO of ammonia. Usu-
lly noble metal catalysts show superior catalytic performance at
relatively low temperatures (200–350 ◦C). However, the selectivityfor N2 on these catalysts are relatively low (typically �80%). Transi-tion metal oxides and mixed oxides catalysts with low cost exhibithigher N2 selectivity, but the operation temperature is significantlyhigher (300–400 ◦C).
Cerium oxide was well known to have a high oxygen exchangecapacity, which was related to the capacity of cerium to changeoxidation states reversibly between Ce4+ and Ce3+ by receivingor giving up oxygen [17]. Ceria (CeO2) was thus rendered a veryeffective catalyst to operate the reaction that involved oxygen. Louet al. have reported that ∼99.2% NH3 conversion was achieved dur-ing catalytic oxidation over Cu–Ce (6:4, molar/molar) catalyst at400 ◦C and the overall selectivity of N2 production varied from 19%to 82% [15]. However, pure ceria would result in a significant effi-ciency decrease under thermally harsh environment. Therefore, anew generation of mixed oxides containing CeO2 and ZrO2 has beendeveloped. The ceria–zirconia mixed oxides have been applied ina number of different fields, for example, as active supports, “oxy-gen buffers” in the three-way catalysis elimination of CO, NOx andhydrocarbons, water–gas shift reaction, and so forth. Comparedwith pure ceria, in the ceria–zirconia mixed oxides, Ce4+ was par-tially substituted for Zr4+ in the lattice of CeO2. Hori et al. [18] foundthat the beneficial effects of ZrO2 were pronounced in Ce1−xZrxO2
catalysts of which oxygen storage capacity was three to five timeshigher than that of the pure CeO2. Moreover, the doping effectof zirconium can maintain the reversible Ce3+/Ce4+ redox prop-erty even after exposure to the reduction condition above 900 ◦C
19]. Adamowska et al. [20] found that incorporating zirconiumnto CeO2 improved the thermal resistance, redox property of theatalyst and its catalytic activity. Wang et al. [21] have also notedhat the catalytic activity of CuO/Ce0.8Zr0.2O2 was higher than thatf CuO/CeO2 because of the insertion of zirconium into CeO2.
In this paper, the ceria–zirconia mixed oxides as potential cata-ysts for the SCO of ammonia was first studied. The primary aim ofhe present research was to acquire such information by studyinghe correlation of crystal structure, acid sites characteristics withhe catalytic performance of ceria–zirconia mixed oxides for SCOf ammonia. These catalysts were characterized by N2 adsorption,-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS),aman spectroscopy, H2 temperature-programmed reduction (H2-PR), NH3 temperature-programmed desorption (NH3-TPD) andemperature-programmed surface reaction (TPSR) techniques.
. Experimental
.1. Catalyst preparation
The ceria–zirconia mixed oxides (Ce1−xZrxO2) catalysts with = 0, 0.2, 0.4, 0.6, 0.8 and 1.0 were prepared via a surfactant-emplated method using cerium (III) nitrate (Ce(NO3)3·6H2O)nd zirconium nitrate (Zr(NO3)4·5H2O) as precursors. For a typ-cal synthesis, an aqueous solution containing cetyltrimethylmmonium bromide (CTAB) was added to an mixed aqueous solu-ion of Ce(NO3)3·6H2O and Zr(NO3)4·5H2O. The molar ratio ofTAB/([Ce] + [Zr]) was 0.4. The mixture was stirred for 0.5 h andhen aqueous ammonia (25%) was slowly added with stirring tillhe pH equaled 10. The above mixture was stirred for another 12 hnd then the homogeneous slurry mixture was hydrothermallyreated at 90 ◦C for 24 h in a teflon-lined autoclave vessel. Afterydrothermal treatment, the precipitates were filtered, washedith deionized water and absolute ethanol until Br species wasndetectable by an AgNO3 solution. Finally, the obtained solidsere dried at 100 ◦C overnight and then calcined at 550 ◦C for 3 h
n air.
.2. Catalyst characterization
The N2 adsorption/desorption isotherms at about −196 ◦Cere measured using a NOVA 4200e Surface Area & Porenalyzer. The specific surface area was calculated using therunauer–Emmett–Teller (BET) model. The pore size distributionsere determined by the Barrett–Joyner–Halenda (BJH) methodsing the desorption branch of the isotherms.
Routine X-ray powder diffraction experiments were carried outn a Rigaku D/max-�b X-ray diffractometer with monochromaticetector. Copper K� radiation was used, with a power setting of0 kV and 100 mA.
UV Raman spectra were recorded on a homemade DL-2 UVaman spectrograph with He–Cd laser of 325 nm excitation wave-
ength. The laser power measured at the samples was below.0 mW under 325 nm radiation.
X-ray photoelectron spectroscopy was measured using an X-ay photoelectron spectrometer (AMICUS, Shimadzu, Japan) with
monochromatic X-ray source of Al K� under ultra-high vacuum3–2 × 10−6 Pa). The XPS data from the regions related to the C 1s,
1s, Zr 3d, Ce 3d core levels were recorded for each sample. Theinding energies were calibrated internally by the carbon deposit
1s binding energy (BE) at 284.8 eV. The deconvolution method of
PS spectra is fitted by Gaussian function.
H2 temperature programmed reduction and the temperaturerogrammed desorption of ammonia was performed on a ChemET TPR/TPD Chemisorptions Analyzer. The production gaseous
neral 411– 412 (2012) 131– 138
during NH3-TPD experiments, such as N2 (m/e = 28), N2O (m/e = 44),NO (m/e = 30), NO2 (m/e = 46), H2 (m/e = 46) was recorded usingquadrupole mass spectrometer (QMS). Typically, 100 mg of sam-ple was pretreated in a flowing stream of He at 120 ◦C for 1 h. Afterthat, H2–Ar mixture (10% H2 by volume) was introduced into theinstrument and the temperature was ramped to 900 ◦C at a heatingrate of 10 ◦C/min. The H2 consumption was monitored by a thermalconduction detector (TCD). For the NH3-TPD experiments, follow-ing the pretreatment step, the catalyst was saturated by a flowof NH3–He mixture (10% NH3 by volume). The reactor was thenpurged with He for a further 2 h to remove weakly adsorbed NH3.Then He was passed through the reactor and the temperature wasramped from room temperature to 700 ◦C at a rate of 10 ◦C/min.
Temperature-programmed surface reaction experiments wereperformed in a fixed-bed reactor using a quadrupole mass spec-trometer as detector. Typically, 100 mg of sample was pretreated byheating in a flowing stream of He from room temperature to 120 ◦Cat 10 ◦C/min. The temperature was held at 120 ◦C for 1 h and thencooled to room temperature. Then, the catalyst was saturated by aflow of NH3/N2/He. The reactor was then purged with He for a fur-ther 2 h to remove weakly adsorbed NH3. Following the ammoniaadsorption step, O2 was passed through the reactor and the temper-ature was increased from room temperature to 800 ◦C at a rate of10 ◦C/min. Simultaneously, the mass signals for N2 (m/e = 28), N2O(m/e = 44), NO (m/e = 30), NO2 (m/e = 46) were recorded and usedfor the calculation of the desorption profiles.
2.3. Catalytic activity tests
The selective catalytic oxidation of ammonia was carried out ina fixed-bed quartz reactor (8 mm in interior diameter). The com-position and amount of the inlet gas mixture was set by mass flowcontrollers. The typical reactant gas composition was as follows:1000 ppm NH3, 10 vol.% O2, and balance He. Each test was car-ried out by loading 200 mg of catalyst, and the bed volume wasabout 0.15 cm3. The total flow rate of the reaction mixture was100 ml/min, and the gas hourly space velocity (GHSV) was about40,000 h−1. The inlet and outlet gas were analyzed by Gas Chro-matograph using a 5A column with a TCD detector for N2 and theNH3 analyzer (GXH-1050, Beijing) to monitor the concentration ofammonia. The signal of all reactants and possible products weremeasured step by step after stabilization of the signals at a giventemperature.
3. Results and discussions
3.1. NH3 oxidation over Ce1−xZrxO2 catalysts
The catalytic performance of the SCO of NH3 on Ce1−xZrxO2 cat-alysts at various temperatures is shown in Fig. 1. It can be observedthat NH3 conversion increased with the increase of reaction tem-perature. For pure CeO2, the catalytic activity was quite poor, andthe N2 selectivity was the lowest in the 240–420 ◦C temperaturerange. The Ce1−xZrxO2 mixed oxides catalysts showed much highercatalytic activity with the increase in Zr content (x < 0.8). Moreover,the N2 selectivity also showed a fast increase at low temperature.Specifically, for x = 0.6, 0.4 and 0.2, they exhibited the higher NH3conversion and the lower complete oxidation temperature (360 ◦C,380 ◦C and 380 ◦C, respectively). However, further increasing theamount of ZrO2 addition (x � 0.8) resulted in the decrease of NH3conversion, and for samples Ce0.2Zr0.8O2 and ZrO2, the complete
NH3 conversion temperature was 400 ◦C and 420 ◦C, respectively.Comparatively, the selectivity of NH3 to N2 increased consistentlywith the addition of Zr, and pure ZrO2 showed the highest N2 selec-tivity (about 100%). In addition, the other gaseous including NO,
Z. Wang et al. / Applied Catalysis A: General 411– 412 (2012) 131– 138 133
Nwaahfoi
3
3
aradstWabt
TC
Fig. 1. NH3 conversion (a) and N2 selectivity (b) over Ce1−xZrxO2 catalysts.
2O, NO2 was also analyzed by MS. The result indicated that N2Oas the main byproduct for NH3 oxidation, at higher temperatures
certain amount of NO was formed, and no NO2 was detected atny of the temperatures studied. The Ce0.4Zr0.6O2 catalyst exhibitedighest catalytic performance for NH3 selective oxidation. There-
ore, an appropriate Ce/Zr ratio in the catalyst was necessary tobtain high activity for NH3 oxidation and the lower Ce/Zr ratiomproved the N2 selectivity.
.2. Catalyst characterization
.2.1. BET surface area and pore characteristicsThe surface areas, pore volume and average pore diameter of
ll catalysts used in this study are listed in Table 1. All the mate-ials were mesoporous (pore size was between 2.7 and 8.3 nm)ccording to the IUPAC [17,22], and presented a narrow pore sizeistribution. The pure CeO2 exhibited the lowest value of specificurface area (ca.98.9 m2/g). However, the specific surface areas of
he Ce1−xZrxO2 samples increased slightly with the addition of Zr.
hen x = 0.6, the specific surface areas reached about 121.8 m2/g,nd then subsequently decreased when x > 0.6. From Fig. 1a, it cane concluded that the higher catalytic activity could be related withhe specific surface areas of the catalyst.
able 1haracterization of Ce1−xZrxO2 catalysts.
Catalysts SBET(m2/g) Total pore volume(cm3 g−1)
CeO2 98.9 0.23
Ce0.8Zr0.2O2 100.8 0.21
Ce0.6Zr0.4O2 119.9 0.20
Ce0.4Zr0.6O2 121.8 0.21
Ce0.2Zr0.8O2 119.8 0.19
ZrO2 114.1 0.32
a Crystallite size of Ce1−xZrxO2 determined from the XRD (1 1 1) diffraction peak by Sch
Fig. 2. XRD patterns of Ce1−xZrxO2 catalysts.
3.2.2. X-ray diffraction and Raman spectroscopyThe XRD patterns of Ce1−xZrxO2 catalysts are presented in Fig. 2.
Pure CeO2 presented the typical reflections of the cubic fluoritestructure at 28.5◦, 33.1◦, 47.5◦, 56.5◦, 59.2◦ and 69.3◦,correspondingto the (1 1 1), (2 0 0), (2 2 0), (3 1 1), (2 2 2) and (4 0 0) planes, respec-tively [23]. For pure ZrO2, the characteristic peaks at 28.3◦ and 31.4◦
could be assigned to the monoclinic (M) phase of ZrO2, and thepeaks at 30.2◦ and 34.7◦ represented the tetragonal (T) phase inZrO2. The Ce0.8Zr0.2O2 sample with lowest Zr content maintainedthe fluorite structure, which was supported by the fact that theposition and shape of their peaks was similar to that of pure CeO2.However, for samples with higher Zr content in the mixed oxides,their reflections in the diffraction patterns systematically shiftedto higher diffraction angles compared with pure CeO2, which wasattributed to the shrinkage of lattice due to the lower ionic radius ofthe Zr4+ ion (0.084 A) with respect to Ce4+ (0.098 A) in agreementwith the Vegard rule [22]. Moreover, the diffraction peaks of themixed oxides were also broadened due to the distortion of the cubicfluorite structure, indicating the phase transition from the cubic flu-orite structure to the tetragonal structure [24]. For sample x = 0.4,the diffraction peaks were asymmetry, which also represented theslight structure deformation of the cubic phase to the tetragonalphase. The Ce0.2Zr0.8O2 sample showed a shoulder peak at 30◦ anda weaker peak at 62.5◦ closed to that of pure ZrO2 line, which indi-cated the presence of both the tetragonal and monoclinic zirconia.This suggested that partial ZrO2 phase has been segregated fromthe Ce0.2Zr0.8O2 mixed oxide [25].
The average crystallite size of Ce1−xZrxO2 catalysts is listed inTable 1. In comparison to the pure ceria, which had a crystalline sizeof 10.2 nm, the Ce1−xZrxO2 mixed oxides exhibited much smallercrystallite sizes. Particularly, the crystalline size decreased to onlyabout 5 nm as the fraction of Zr (x) was 0.4, 0.6 and 0.8. There-fore, the incorporation of ZrO2 into CeO2 resulted in the decrease
of crystallite size [26].
To obtain the accurate identification of distortions of sub-latticeoxygen and defects in the structure, Raman spectroscopy was per-formed due to its strong sensitivity for the oxygen atoms in the
Average pore diameter(nm) Crystallite sizea (nm)
2.7 10.28.3 10.15.4 5.14.3 5.03.4 5.77.9 7.4
errer’s equation.
134 Z. Wang et al. / Applied Catalysis A: General 411– 412 (2012) 131– 138
paRaaaroacTstt3FbmCots[dTxsotlttsqcgC
ivncttaCtp
Table 2Surface elemental composition and relative proportion for the samples derived fromXPS analyses.
Samples Surface composition (at.%) Ce/Zr ratio Ce3+ in Ce (%)
than pure CeO2. Generally, the presence of Ce3+ was assigned withthe generation of oxygen vacancies according to due to the chargecompensation [38,39]. Therefore, it was likely that the Zr doping
Fig. 3. Raman spectra of Ce1−xZrxO2 catalysts.
resence of heavy atoms such as Ce and Zr [27]. Raman spectra ofll samples prepared in the present study are shown in Fig. 3. Theaman spectra of CeO2 showed a prominent peak at ∼460 cm−1
nd two broader peaks at ∼600 cm−1 and 1180 cm−1. The bandt ∼460 cm−1 could be attributed to the characteristic F2g Ramanctive mode of the fluorite-type lattice [28] as evidenced by XRDesults, and could be viewed as a symmetric breathing mode ofxygen atoms around cerium ions [29]. The peak at 600 cm−1 wasssigned to the nondegenerate longitudinal optical (LO) mode oferia, which was linked to oxygen vacancies in the ceria lattice [30].he band at 1180 cm−1 can be assigned to the defects of fluoritetructure in ceria [31]. For pure ZrO2 sample, the spectral fea-ures at 150, 270, 331, 476 and 640 cm−1 was predominant due toetragonal ZrO2, whereas the further two bands, located at 181 and84 cm−1, indicated the presence of the monoclinic phase [32,33].or the sample with x = 0.2, no Raman lines attributed to ZrO2 coulde observed, which was in accordance with the XRD measure-ents. Thus it was thought that ZrO2 was incorporated into the
eO2 lattice to form a solid solution. However the Raman intensityf these peaks (460 cm−1, 600 cm−1, 1180 cm−1) decreased withhe amount increasing of ZrO2. It has been known that the inten-ity of Raman signals depended on the grain size and morphology27]. Thus, the above observation should be associated with theecreasing of grain size of ceria–zirconia catalysts, as shown inable 1. Additionally, an extra band at 650 cm−1 appeared when
was increased to 0.4, which was attributed to the tetragonal sub-titution of oxygen atoms from cubic fluorite structure [34]. Thebservation predicted the possibility of the existence of metastableetragonal phase or/and the association with tetragonal phase-likeattice distortion in the Ce0.6Zr0.4O2 catalyst. When x was increasedo 0.6, the band at 650 cm−1 increased obviously in intensity. Fur-hermore, the peak at 460 cm−1 corresponding to cubic fluoritetructure of ceria even disappeared. When x was increased subse-uently to 0.8, the intensity at 650 cm−1 was further strengthenedompared with that of x = 0.4 and 0.6. Thus, the above results sug-ested that there existed the tetragonal phase in the Ce0.6Zr0.4O2,e0.4Zr0.6O2 and Ce0.2Zr0.8O2 catalysts.
Considering the catalytic activity, it could be concluded that thencorporation of Zr into CeO2 remarkably improved the NH3 con-ersion compared with pure CeO2. On the other hand, it was alsooted by the results of XRD and Raman that the addition of Zr intoeria leaded to the phase transition from the cubic fluorite structureo the tetragonal structure and promoted the formation of struc-ural defects. These structural defects can adsorb NH3 molecules
nd oxygen much more strongly [32]. However, Compared withe0.4Zr0.6O2 and Ce0.6Zr0.4O2 catalysts, the Ce0.2Zr0.8O2 exhibitedhe lower NH3 conversion. This should be attributed to the partialhase segregations of ZrO2 in Ce0.2Zr0.8O2 catalysts.
3.2.3. X-ray photoelectron spectroscopyThe XPS measurement was used to verify the surface elements
and elemental oxidation states of Ce1−xZrxO2 mixed oxidation. Thesurface elemental contents and Ce/Zr atomic ratio calculated fromthe normalized peak areas of Ce 3d, Zr 3d and O 1s core level spec-tra are listed in Table 2. The detected surface Ce/Zr atomic ratio ofCe0.8Zr0.2O2 catalyst was close to the theoretical value (3.84), whichindicated that the isovalent Zr was well incorporated into the CeO2lattice to form the solid solution, as confirmed by XRD results. ForCe1−xZrxO2 (x = 0.4, 0.6) catalysts, the surface Ce/Zr atomic ratio washigher compared with the theoretical value (1.50 and 0.67). Theenrichment of Ce species on the surface of Ce–Zr mixed oxides wasobtained. Postole et al. [22] also found that the cerium enrichmenton ceria–zirconia mixed oxides produced, being the Ce/Zr surfaceatomic ratio of 0.80, 1.60 and 2.87 versus a nominal composition of0.42, 1.0 and 2.33, respectively. However, the lower Ce/Zr atomicratio (0.18) for Ce0.2Zr0.8O2 catalyst compared with the theoreticalatomic ratio (0.25) should be due to the segregation of ZrO2 on thesurface.
The complex spectrum of Ce 3d was composed into eight peakscorresponding to four pairs of spin-orbit doublets, as shown inFig. 4, indicating that the coexistence of Ce3+ and Ce4+. The label-ing of peaks was similar to those reported by Burroughs et al. [35].Moreover, the difference in Ce 3d3/2 and Ce 3d5/2 binding energies(18.4–18.5 eV) for the whole samples was in agreement with anexpected value of 18.5 eV [36,37]. In Fig. 4, the bands labeled as u(901.1 eV), u′′ (907.6 eV) and u′ ′′ (916.7 eV) were arising from Ce4+
3d3/2, while the bands labeled as v (882.7 eV), v′′ (889.1 eV) and v′ ′′
(898.3 eV) were arising from Ce4+ 3d5/2. Compared with the bandsof Ce4+ 3d, the bands labeled as u′ (903.8 eV) and v′ (885.3 eV) wasdue to the Ce3+ 3d3/2 and Ce3+ 3d5/2, respectively. The relative con-centration of Ce3+ is shown in Table 2. It can be found that the Zrdoped samples exhibited relatively higher concentration of Ce3+
Fig. 4. Ce 3d XPS spectra for Ce1−xZrxO2 catalysts.
Z. Wang et al. / Applied Catalysis A: General 411– 412 (2012) 131– 138 135
fwCCCNdwtcctt
3
rtCrnwrttr
C
t8wltCtogo[ctoa[t
2.
the intensity of moderate acid sites decreased when x > 0.6. Yet,
Fig. 5. H2-TPR profiles of Ce1−xZrxO2 catalysts.
acilitated the reduction of Ce4+ to Ce3+, and the oxygen vacanciesere also easily generated on the surface of Ce–Zr mixed oxides.ombining with the catalytic activity, it could be observed that thee1−xZrxO2 (x = 0.2, 0.4, 0.6) catalysts with higher concentration ofe3+ on the surface exhibited the better catalytic performance forH3-SCO. Interestingly, the highest proportion of Ce3+ species wasetected on Ce0.4Zr0.6O2 catalyst, and the catalytic performanceas also best. Thus the oxygen vacancies were responsible for
he activity promotion of NH3-SCO. However, when the zirconiumontent was increased to 0.8, there was a decreasing in the Ce3+
oncentration compared with x = 0.6. This should be resulted fromhe phase segregations of ZrO2 from mixed oxides, as confirmed byhe XRD and XPS observations.
.2.4. H2-TPRThe TPR profiles of the catalysts are presented in Fig. 5. No
eduction for pure ZrO2 was observed in our work, and the reduc-ion of Zr4+ will occurs above 1000 ◦C [40]. The TPR profile of pureeO2 showed two prominent broad peaks, thus suggesting that theeduction of CeO2 occurs via a stepwise mechanism [17]. The sig-al at ca.560 ◦C was attributed to the reduction of the surface shell,hile the high temperature band at ca. 860 ◦C was due to a bulk
eduction [34]. Meanwhile it has been known that the ceria reduc-ion at 560 ◦C will result in the formation of oxygen vacancies (�)hat can enhance ionic conductivity in CeO2 and play an importantole for catalytic reaction process [41]:
e4+ + H2 + O2− → Ce3+ + H2O + �
The TPR profile of Ce0.8Zr0.2O2 catalyst showed that the addi-ion of Zr decreased the bulk reduction step from 860 ◦C (ceria) to20 ◦C. The promotion of the reduction in the bulk upon dopingith ZrO2 should be attributed to that the incorporation of isova-
ent non-reducible elements like Zr4+ into the cubic CeO2 lattice ledo the stress in the lattice, which facilitated oxygen extraction fromeO2 [42]. When x was increased to 0.4, the reduction peak at highemperature disappeared, and a broad peak at about 565 ◦C wasbserved. The surface and bulk reduction of CeO2 cannot be distin-uished clearly, and the similar phenomena was also be observedn Ce0.75Zr0.25O2 and Ce0.62Zr0.38O2 catalysts reported by Guo et al.43]. What is more, with increasing the ZrO2 content (x > 0.2), the H2onsumption peak of surface cerium species was shifted to higheremperatures. These results indicated that too high concentrationf ZrO did not favor the mobility of oxygen ion and showed a neg-
2tive effect on the reduction of surface cerium. Adamowska et al.20] have also noted that the transformation of cubic structure intoetragonal structure would affect the mobility of oxygen ions in
Fig. 6. TPD profiles of ammonia on Ce1−xZrxO2 catalysts.
the catalyst. Thus, it should be pointed out that the structure wasintently associated with reduction of surface cerium.
Interestingly, for Ce1−xZrxO2 catalyst except for ZrO2 andCe0.2Zr0.8O2, the additional low-temperature reduction peak atabout 435–490 ◦C attributed to the highly dispersed ceria specieswas also observed [44]. In addition, the Ce0.4Zr0.6O2 catalyst withthe smallest particle size (5 nm) in all the catalysts showed thelowest reduction temperature and highest reduction intensity forthe species. Nevertheless, it was also observed in Table 1 thatthe particle size was close among Ce0.4Zr0.6O2, Ce0.6Zr0.4O2 andCe0.8Zr0.2O2 catalyst. Zhao et al. [45] have reported that an obviouslow-temperature reduction peak at about 370 ◦C in the H2-TPR pro-files of Ce0.67Zr0.33O2 sample prepared by coprecipitation methodwas observed compared with that of homogeneous precipitation,microemulsion and hydrothermal method. Thus it was necessaryto understand the further reason about the obvious difference ofthe Ce0.4Zr0.6O2 catalyst with respect to the others by fabricatingthe catalyst using different preparation method in the followinginvestigation. Simultaneously, it has been shown from Fig. 1 thatthe Ce0.4Zr0.6O2 catalyst exhibited the highest catalytic activity forNH3 oxidation and better N2 selectivity. However, the NH3 con-version of Ce0.2Zr0.8O2 catalyst was lower than that of Ce0.4Zr0.6O2catalyst. This should be due to partial phase segregations of ZrO2 inCe0.2Zr0.8O2 mixed oxide, which made that the reduction peak ofCe0.2Zr0.8O2 catalyst obviously shifted to higher temperature.
3.2.5. NH3-TPDThe NH3-TPD profiles on Ce1−xZrxO2 catalysts are shown in
Fig. 6. The amount of ammonia desorbed for the samples is ameasure of their acid sites concentration, whereas the shape ofdesorption spectra depends on the heterogeneity of acid sitesstrength [8]. For all catalysts, the desorption profiles of ammoniawas observed over the wide temperature range of 20–700 ◦C. Thiswas the characteristic of the presence of several adsorbed NH3species differing in thermal stability [7]. A quantitative analysisindicated that the density of acid sites followed this relative order ofCe0.8Zr0.2O2 > CeO2 > Ce0.6Zr0.4O2 > Ce0.4Zr0.6O2 > Ce0.2Zr0.8O2 > ZrOThen it indicated that the Ce/Zr ratio affected the density of acidsites. In Fig. 6, for pure CeO2 catalyst, only two desorption peaksat about 150 ◦C and 420 ◦C were observed, which indicated thepresence of weak and strong acid sites. With the increase inzirconium content, the peaks of moderate acid sites (250–370 ◦C)were appeared. Moreover, the intensity of moderate acid sitesincreased with the zirconium content when x � 0.6. Subsequently,
the Ce0.4Zr0.6O2 catalyst held the largest amount of moderate acidsites. The relationship between acid sites and catalytic activitywas shown in Fig. 7. It was observed that the Ce–Zr mixed oxides
136 Z. Wang et al. / Applied Catalysis A: General 411– 412 (2012) 131– 138
Fo
wcaCtdtssrscZ
crl0c(Tc
taodNwTp
ig. 7. Variations of NH3 conversion and percentage of acid sites with the increasingf Ce/Zr ratios.
ith higher amount of moderate acid sites showed the excellentatalytic activity. Thus, the above results confirmed the moderatecid sites favored the catalytic oxidation of NH3. Though thee0.8Zr0.2O2 sample also presented the higher moderate acid sites,he significant contribution of strong acid sites to total acidityepressed the NH3 oxidation. Wang et al. [46] have reportedhat the catalytic performance of Ce1−xZrxO2 catalysts for directynthesis of diethyl carbonate (DMC) was affected by acid–baseites, and the acid–base sites played an important role in theeaction. Akah et al. [47] have also found that the presence oftrong acid sites (425 ◦C) in H–[Al] ZSM-5 leaded to the lower NH3onversion in the SCO of ammonia compared with the Fe–H–[Al]SM-5 in absence of strong acid sites.
Combining with the XRD, Raman, XPS and NH3-TPD results, itould be concluded that the formation of moderate acid sites waselated to be the presence of metastable tetragonal phase, whicheaded to the slight modification of cubic crystal structure with.2 � x � 0.8. Moreover, the incorporation of zirconia ions into theeria framework also resulted in the change of the BET surface areaTable 1) and improved the formation of oxygen vacancies (Table 2).hus it was thought the catalytic activity for NH3 oxidation waslosely related with the crystal structure of the mixed oxide.
In parallel with the NH3 desorption, the H2 formation indicatedhat the some NH3 was further activated and transformed inton NHx type of species [2]. The production of N2 and N2O werebserved during NH3-TPD process. No trace of NO2 and NO wereetected over the whole examined temperature range. During theH -TPD experiments, the N formation on Ce Zr O catalysts
3 2 1−x x 2as observed in Fig. 8. Moreover, the results of N2-TPD and He-
PD demonstrated that the N2 formation did not originate from thehysically adsorbed N2 and the decomposing of residual nitrate on
Fig. 8. N2 production profiles during NH3-TPD on Ce1−xZrxO2 catalysts.
Fig. 9. N2 production profiles of temperature-programmed surface reaction (TPSR)for Ce1−xZrxO2 catalysts.
the surface of the Ce–Zr mixed oxides. For pure ZrO2, there was noavailable oxygen in H2-TPR experiment. In addition, only one lowertemperature peak of N2 formation was observed in Fig. 8. Thus, itcould be thought that the N2 formation at low temperatures maybe due to the combination of two NHx species [48,49]. The otherCe–Zr mixed oxides should also exhibit the same reaction route forN2 formation at low temperatures (100–260 ◦C) during NH3-TPD.At higher temperature (370 ◦C), the activated lattice oxygen of theCe–Zr mixed oxides reacted with NHx species to produce N2 [3,10].In parallel with N2 formation, the N2O formation was also observedduring NH3-TPD experiments. The main reaction routes was thatthe NHx species reacted with lattice oxygen to form HNO intermedi-ates, then two HNO species reacted with each to produce N2O [50].Furthermore, the nitrogen peak at 370 ◦C revealed the differencesamong the catalysts. Remarkably, the Ce0.4Zr0.6O2 catalyst showedthe maximum N2 production peak at 370 ◦C than other catalysts.It was reasonable to suggest that the lattice oxygen would oxidizethe adsorbed ammonia species to form N2 and N2O, and might beone of the active oxygen species for the SCO of ammonia reaction.
3.2.6. TPSRReactivity of chemisorbed ammonia on the surface of catalysts
with gas oxygen is studied by temperature-programmed surfacereaction technique. The production of N2, NO and N2O weredetected in this experiment. The N2 production profiles are shownin Fig. 9. In the presence of gaseous oxygen, a peak for N2 formationwas obviously observed at 250 ◦C. Long and Yang [10] also foundthat gaseous oxygen species may participate in the SCO reaction atlower temperatures. Moreover, for 0.2 � x � 0.6 catalysts, the pro-duction peak of N2 profiles (250 ◦C) was larger than pure CeO2 andZrO2 catalysts. Comparison with the catalytic profiles measured onthe catalysts showed that these catalysts (0.2 � x � 0.6) were muchmore active for the NH3-SCO, which was also in accordance withthe relative concentration of Ce3+ (oxygen vacancies) on the surfaceof catalysts obtained by XPS. Also, production profiles of N2O wereshown in Fig. 10. It could be observed that the formation amount ofN2O gradually decreased with Zr content from 0.4 to 1.0 at below∼340 ◦C. Therefore, the low selectivity of ceria-rich catalysts wasdue to the formation of N2O during ammonia oxidation reaction.Meanwhile, in Fig. 1b, it can be also observed that zirconium-richcatalyst favored the SCO of NH3 to N2 and improved the N2 selec-tivity. Nevertheless, the N2 selectivity started to be decreased forpure ZrO2 when the temperature was higher than 360 ◦C, whichwas related to the larger amount of N2O production at above 350 ◦C
(Fig. 10). Fornasiero et al. [51] have also found that the reductionof NO by CO strongly depended both on the content of ZrO2 inthe solid solution and the nature of the phase present. This alsoindicated that the N2 selectivity for NH3-SCO reaction was strongly
Z. Wang et al. / Applied Catalysis A: Ge
Fig. 10. N2O production profiles of temperature-programmed surface reaction(TPSR) for Ce1−xZrxO2 catalysts.
F(
dtTdot
owttoraooNiot(id
4
apc
[[
[
[
[
[[
[[
ig. 11. NO production profiles of temperature-programmed surface reactionTPSR) for Ce1−xZrxO2 catalysts.
ependent on the addition of ZrO2 into CeO2. It could be observedhat the reaction temperature (250 ◦C) with gaseous oxygen duringPSR experiments was lower than that (370 ◦C) of lattice oxygenuring NH3-TPD experiments. This suggested that the gaseousxygen species was much more active than lattice oxygen at lowemperature and more essential for NH3-SCO reaction.
A certain amount of NO (Fig. 11) was also observed to be formedver all samples in the wide temperature range of 50–500 ◦C,hereas it was not detected in NH3-TPD experiment. Thus the reac-
ion pathway of NH3 with the gaseous oxygen could differ fromhe lattice oxygen. The oxidation reaction of ammonia via latticexygen obeys hydrazinium intermediate mechanism [4,9,13,50]. Aseported by Darvell et al. [9], the hydrazine decomposition mech-nism was important under conditions where is little, or limitedxygen available. Jones et al. [52] also proposed that, at very lowxygen concentrations, the route for NH3 oxidation may go via2H4 species. The lattice oxygen was just assigned to the lim-
ted oxygen/low concentration oxygen. However the SCO reactionf gaseous oxygen with ammonia might be in accordance withhe hypothesis of in situ or “internal” selective catalytic reductioniSCR) mechanism [4,12,47]. The further experiments are requiredn order to study thoroughly the mechanism of the ammonia oxi-ation over Ce1−xZrxO2 catalysts.
. Conclusions
In this study, the Ce1−xZrxO2 catalysts have been characterizednd tested for the SCO of ammonia. It was observed that the incor-oration of zirconium into the CeO2 lattice noticeably affected therystal structure, crystallite size, acid sites and catalytic activity.
[
[
neral 411– 412 (2012) 131– 138 137
Moreover, Zr doping facilitated the reduction of Ce4+ to Ce3+, andthe oxygen vacancies were also easily generated on the surface ofCe–Zr mixed oxides. As indicated by XRD and Raman, ceria–zirconiamixed oxides exhibited a cubic-tetragonal structural phase transi-tion. When x = 0.4, 0.6 and 0.8, the tetragonal crystallite appearedin the Ce1−xZrxO2 catalysts, and their crystallite size was sharplydecreased. Furthermore, the higher amount of moderate acid siteswas also observed in Ce–Zr mixed oxides compared with pureCeO2 and ZrO2, and these catalysts exhibited better catalytic activ-ity. Particularly, the Ce0.4Zr0.6O2 catalyst showed the highest NH3conversion and the lowest complete conversion temperature atabout 360 ◦C. For the Ce0.2Zr0.8O2 catalyst, the lower activity com-pared with Ce1−xZrxO2 (x = 0.2, 0.4 and 0.6) should be attributedto the partial phase segregations of ZrO2 in Ce0.2Zr0.8O2. In addi-tion, one important phenomenon was found that zirconium-richcatalyst (x > 0.4) exhibited the higher N2 selectivity (∼100%), whichsuggested that the higher Zr content played an important role forimproving the N2 selectivity. The formation of N2O was the mainreason of low N2 selectivity for these catalysts (x � 0.4) proved byTPSR experiments. It was reasonable to suggest that the catalyticproperties were closely dependent on the crystal structure and acidsites. Meanwhile, it was also found that the gaseous and lattice oxy-gen species all participated in the SCO of NH3 to N2. The reactionmechanism of NH3 with lattice oxygen differed from that of thegaseous oxygen. And the gaseous oxygen species was much moreactive than lattice oxygen at low temperature for NH3-SCO reaction.
Acknowledgements
This work was supported by the National High TechnologyResearch and Development Program of China (863 Program) (No.2009AA062604), the National Nature Science Foundation of China(No. 20807010), the Program for New Century Excellent Tal-ents in University (NCET-09-0256), the Program for ChangjiangScholars and Innovative Research Team in University (IRT0813),and the Fundamental Research Funds for the Central Universities(DUT10LK19).
References
[1] J. Phillips, Control Pollution Prevention Options for Ammonia Emissions, U.S.Environmental Protection Agency, 1995.
[2] G. Olofsson, L.R. Wallenberg, A. Andersson, J. Catal. 230 (2005) 1–13.[3] G. Lu, J. van Grondelle, B.G. Anderson, R.A. van Santen, J. Catal. 186 (1999)
100–109.[4] L. Zhang, H. He, J. Catal. 268 (2009) 18–25.[5] H.M.J. Kusar, A.G. Ersson, M. Vosecky, S.G. Järås, Appl. Catal. B: Environ. 58
(2005) 25–32.[6] M.J. Lippits, A.C. Gluhoi, B.E. Nieuwenhuys, Catal. Today 137 (2008) 446–452.[7] M. Amblard, R. Burch, B.W.L. Southward, Catal. Today 59 (2000) 365–371.[8] L. Chmielarz, P. Kustrowski, Z. Piwowarska, M. Michalik, B. Dudek, R. Dziembaj,
Top. Catal. 52 (2009) 1017–1022.[9] L.I. Darvell, K. Heiskanen, J.M. Jones, A.B. Ross, P. Simell, A. Williams, Catal.
Today 81 (2003) 681–692.10] R.Q. Long, R.T. Yang, J. Catal. 207 (2002) 158–165.11] J.M.G. Amores, V.S. Escribano, G. Ramis, G. Busca, Appl. Catal. B: Environ. 13
(1997) 45–58.12] X.Z. Cui, J. Zhou, Z.Q. Ye, H.R. Chen, L. Li, M.L. Ruan, J.L. Shi, J. Catal. 270 (2010)
310–317.13] L. Lietti, G. Ramis, G. Busca, F. Bregani, P. Forzatti, Catal. Today 61 (2000)
187–195.14] M. Amblard, R. Burch, B.W.L. Southward, Appl. Catal. B: Environ. 22 (1999)
159–166.15] J.C. Lou, C.M. Hung, S.F. Yang, J. Air Waste Manage. Assoc. 54 (2004) 68–76.16] L. Chmielarz, P. Kustrowski, A.R. Lasocha, R. Dziembaj, Appl. Catal. B: Environ.
58 (2005) 235–244.17] M.H. Yao, R.J. Baird, F.W. Kunz, T.E. Hoost, J. Catal. 166 (1997) 67–74.18] C.E. Hori, H. Permana, K.Y. Simon Ng, A. Brenner, K. More, K.M. Rahmoeller, D.
Belton, Appl. Catal. B: Environ. 16 (1998) 105–117.19] T. Daniela, T. Alessandro, L. Jordi, L. de, D. Carla, Giuliano, Catal. Today 43 (1998)
79–88.20] M. Adamowska, S. Muller, P. Dacosta, A. Krzton, P. Burg, Appl. Catal. B: Environ.
74 (2007) 278–289.
1 s A: Ge
[
[
[
[
[
[
[
[[
[
[[[
[[
[[
[[[[
[[
[
[
[[[[[50] G. Olofsson, A. Hinz, A. Andersson, Chem. Eng. Sci. 59 (2004) 4113–4123.[51] P. Fornsiero, G. Balducci, J. Di Monte, V. Kaspar, G. Sergo, A. Gubitosa, M. Ferrero,
22] G. Postole, B. Chowdhury, B. Karmakar, K. Pinki, J. Banerji, A. Auroux, J. Catal.269 (2010) 110–121.
23] I. Atribak, B. Azambre, A. Bueno López, A. García-García, Appl. Catal. B: Environ.92 (2009) 126–137.
24] Y.Z. Chen, B.J. Liaw, W.C. Chang, C.T. Huang, Int. J. Hydrogen Energy 32 (2007)4550–4558.
25] J.L. Ayastuy, A. Gurbani, M.P. González-Marcos, M.A. Gutiérrez-Ortiz, Appl.Catal. A 387 (2010) 119–128.
26] W. Khaodee, N. Tangchupong, B. Jongsomjit, P. Praserthdam, S. Assabumrun-grat, Catal. Commun. 10 (2009) 494–501.
27] B.M. Reddy, P. Bharali, Y.H. Seo, E.A. Prasetyanto, S.E. Park, Catal. Lett. 126 (2008)125–133.
28] L.J. Liu, Z.J. Yao, B. Liu, L. Dong, J. Catal. 275 (2010) 45–60.29] H. Vidal, J. Kaspar, M. Pijolat, G. Colon, S. Bernal, A. Cordón, V. Perrichon, F. Fally,
Appl. Catal. B: Environ. 30 (2001) 75–85.30] B.M. Reddy, P. Bharali, P. Saika, S.E. Park, M.W.E. Vanden Berg, M. Muhler, W.
Grünert, J. Phys. Chem. C 112 (2008) 11729.31] W.J. Shan, W.J. Shen, C. Li, Chem. Mater. 15 (2003) 4761–4767.32] K.A. Pokrovski, A.T. Bell, J. Catal. 241 (2006) 276–286.
33] A. Trovarelli, F. Zamar, J. Llorca, C. de Leitenburg, G. Dolcetti, J.T. Kiss, J. Catal.
169 (1997) 490–502.34] B.M. Reddy, G.K. Reddy, L. Katta, J. Mol. Catal. A: Chem. 319 (2010) 52–57.35] P. Burroughs, A. Hamnett, A.F. Orchard, G. Thornton, J. Chem. Soc. Dalton Trans.
17 (1976) 1686–1698.
[
neral 411– 412 (2012) 131– 138
36] A.E. Nelson, K.H. Schulz, Appl. Surf. Sci. 210 (2003) 206–221.37] J. Fan, X.D. Wu, Q. Liang, R. Ran, D. Weng, Appl. Catal. B: Environ. 81 (2008)
38–48.38] Y. Sun, P.A. Sermon, J. Mater. Chem. 6 (1996) 1025–1029.39] C.C. Zhang, J. Lin, Phys. Chem. Chem. Phys. 13 (2011) 3896–3905.40] R.D. Monte, J. Kaspar, Catal. Today 100 (2005) 27–35.41] A. Łamacz, A. Krzton, G.D. Mariadassou, Catal. Today (2011),
doi:10.1016/j.cattod.2011.01.032.42] K.G. Azzam, I.V. Babich, K. Seshan, L. Lefferts, J. Catal. 251 (2007) 153–162.43] Y. Guo, G.Z. Lu, Z.G. Zhang, S.H. Zhang, Y. Qi, Y. Liu, Catal. Today 126 (2007)
296–302.44] B.M. Reddy, P. Bharali, P. Saikia, G. Thrimurthulu, Y. Yamada, T. Kobayashi, Ind.
