Propene Poisoning on Three Typical Fe-zeolites for SCR of NOx with

Article pubs.acs.org/est

Propene Poisoning on Three Typical Fe-zeolites for SCR of NOx with NH3: From Mechanism Study to Coating Modified Architecture Lei Ma,† Junhua Li,*,† Yisun Cheng,‡ Christine K. Lambert,‡ and Lixin Fu† †

State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Tsinghua University, Beijing 100084, China ‡ Research and Innovation Center, Ford Motor Company, Dearborn, Michigan 48121, United States S Supporting Information *

ABSTRACT: Application of Fe-zeolites for urea-SCR of NOx in diesel engine is limited by catalyst deactivation with hydrocarbons (HCs). In this work, a series of Fe-zeolite catalysts (Fe-MOR, Fe-ZSM-5, and Fe-BEA) was prepared by ion exchange method, and their catalytic activity with or without propene for selective catalytic reduction of NOx with ammonia (NH3-SCR) was investigated. Results showed that these Fe-zeolites were relatively active without propene in the test temperature range (150−550 °C); however, all of the catalytic activity was suppressed in the presence of propene. Fe-MOR kept relatively higher activity with almost 80% NOx conversion even after propene coking at 350 °C, and 38% for Fe-BEA and 24% for Fe-ZSM-5 at 350 °C, respectively. It was found that the pore structures of Fe-zeolite catalysts were one of the main factors for coke formation. As compared to ZSM-5 and HBEA, MOR zeolite has a onedimensional structure for propene diffusion, relatively lower acidity, and is not susceptible to deactivation. Nitrogenated organic compounds (e.g., isocyanate) were observed on the Fe-zeolite catalyst surface. The site blockage was mainly on Fe3+ sites, on which NO was activated and oxidized. Furthermore, a novel fully formulated FeBEA monolith catalyst coating modified with MOR was designed and tested, the deactivation due to propene poisoning was clearly reduced, and the NOx conversion reached 90% after 700 ppm C3H6 exposure at 500 °C.

1. INTRODUCTION Nitrogen oxides (NO, NO2, N2O) in the exhaust gases from combustion of fossil fuels are a major cause of photochemical smog, acid rain, and ozone depletion.1−3 Removal of NOx from the exhaust of diesel engines is a major challenge to fulfill future restrictive emission standards.4 Selective catalytic reduction (SCR) of NOx to N2 using NH3 as a reductant is considered to be an effective technology for the removal of NOx from diesel engine emissions. The required ammonia is produced on board by decomposition and hydrolysis of aqueous urea in an additional tank.5,6 The most common commercial catalyst for this process is V2O5/TiO2 (anatase) promoted with either WO3 or MoO3 in stationary sources.7,8 The thermal stability of SCR catalysts is critical to withstand the harsh environment due to the high temperatures of the diesel exhaust during very high load mode or regeneration of the diesel particulate filter (DPF), while the loss of activity of V2O5−WO3/TiO2 is unavoidable due to the phase change of TiO2 from anatase to rutile around 600 °C. Iron supported zeolite catalysts have attracted much attention due to their remarkable catalytic activity for SCR of NOx with ammonia and greater thermal stability than vanadium catalysts. Because of their high activity and durability, Fe-exchanged zeolite catalysts,9 especially Fe-ZSM-5,10,11 Fe-MOR,12 and FeBEA,13,14 were studied extensively in the past decade. In recent research, it was found that the major problem of Fe-zeolite in © 2012 American Chemical Society

NH3-SCR reaction for practical applications is deactivation in the presence of hydrocarbons (HCs), which are in the exhaust from fuel combustion in the engine, especially during cold start conditions when oxidation catalysts are inactive.15,16 Inhibition of DeNOx efficiency by C3H6 competitive adsorption was observed by the temperature programmed desorption (TPD) study of coadsorption of NH3 and C3H6 on Fe-ZSM-5 catalysts,17 while the coadsorption of ammonia and propene was carried out at room temperature. This indicated that inhibition by HCs competitive adsorption might mainly occur at low temperature. Meanwhile, coke formation was demonstrated to be the main reason led to rapid deactivation of catalytic activity using HCs as reductants in SCR, on which NO oxidation was seriously inhibited, and then caused noticeable reduction of the standard SCR reaction efficiency.18−20 Up to now, there are few reports about HCs poisoning on different zeolite SCR catalysts. Therefore, it is necessary to investigate how HCs affect the activity of different Fe-zeolites in NH3-SCR of NOx for practical application in diesel engine emission control. Received: Revised: Accepted: Published: 1747

September 2, 2011 November 17, 2011 January 12, 2012 January 12, 2012 dx.doi.org/10.1021/es203070g | Environ. Sci. Technol. 2012, 46, 1747−1754

Environmental Science & Technology

Article

5% O2 was introduced into the reactor. In the tests, the total flow rate was fixed at 500 mL/min, which corresponded to a GHSV (gas hourly space velocity) of 160 000 h−1. The performance of the catalysts is presented in terms of conversion of NO x (X(NO x )) and conversion of NH3(X(NH3)) as defined by eqs 1 and 2.

In this work, the activity and deactivation behavior of FeMOR, Fe-ZSM-5, and Fe-BEA catalysts for NH3-SCR were studied, and the reason for deactivation on different zeolite catalysts was proposed. Fe-MOR kept relatively high NOx reduction activity as compared to other Fe-zeolites. On the basis of the results, we devised a strategy to modify a conventional Fe-BEA SCR catalyst in ways that could improve the HCs tolerance. The results suggested that Fe-BEA or even other kinds of Fe-zeolite SCR catalysts could have a potential application for diesel emission control by adopting such a modification strategy.

X(NOx ) =

with[NOx ] = [NO] + [NO2 ]

2. EXPERIMENTAL SECTION 2.1. Preparation of Catalysts. The starting materials used for preparation of the catalysts are as follows. Mordenite (MOR, SiO2/Al2O3 = 13) was obtained from Zeolyst International Co., and H-ZSM-5 (SiO2/Al2O3 = 25) and HBEA (SiO2/Al2O3 = 25) were supplied by Nankai University Catalyst Co., Ltd. Fe-zeolite catalysts were prepared by using the conventional ion-exchange procedure. Three types of zeolite, mordenite, H-ZSM-5, and HBEA zeolite, were employed as the support. For each catalyst, 2 g of zeolite was added to 200 mL of 0.05 M FeCl2 solution with constant stirring at room temperature. After 24 h, the mixture was filtered and washed several times with deionized water. The obtained solid was dried at 110 °C overnight and then calcined at 550 °C for 4 h in air. Fe2+ in the catalysts was oxidized to Fe3+.21,22 Finally, the catalyst was palletized and crushed to 40− 60 mesh for evaluation. The obtained catalysts are denoted as Fe-MOR, Fe-ZSM-5, and Fe-BEA, respectively. The chemical compositions of these prepared samples were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) with a IRIS Intrepid II XSP apparatus from Thermo Fisher Scientific Inc. The iron exchange extent was calculated by 3 × (number of iron ions)/(number of aluminum

X(NH3) =

SiO2/ Al2O3

Fe-MOR Fe-ZSM-5 Fe-BEA

13 25 25

Fe content (wt %)a iron exchange extent (%)b 2.28 2.11 1.93

[NH3]inlet − [NH3]outlet × 100% [NH3]inlet

(1)

(2)

Catalyst activity tests were carried out in the temperature range of 150−500 °C. To avoid the impact of gas adsorption on the catalyst samples, the test data were recorded after the reactions had maintained stable states for 120 min. The monolithic Fe-BEA catalysts before and after the modification were aged and tested in a laboratory-scale flow reactor. NOx performance was evaluated by ramping the oven temperature at 5 °C/min from 150 to 750 °C in a flowing gas of 350 ppm NO, 350 ppm NH3, 14% O2, 5% CO2, 5% H2O, and balance N2 at 30 000 h−1 based on the geometric volume of the monolith core. The steady-state NOx performance was measured in the flow reactor connected to an FTIR instrument from MKS Instruments, Inc., with a heated sample cell system for gas analysis. 2.3. Characterization of Catalyst. A Quantachrome Nova Automated Gas Sorption System was used to measure the N2 adsorption isotherms of the samples at liquid N2 temperature (−196 °C). The specific surface area was determined from the linear portion of the BET plot. The pore size distribution was calculated from the desorption branch of the N2 adsorption isotherm using the HK method. Prior to the surface area and pore size distribution measurements, the samples were degassed in a vacuum at 300 °C for 4 h. Temperature programmed desorption (TPD) was conducted by using 0.1 g of catalyst in a quartz reactor. The adsorption was performed by passing a gas mixture, containing 500 ppm NH3, or 500 ppm NO and 5% O2 with N2 as balance gas, through the sample bed at 25 °C for 2 h with the total flow rate of 100 mL/min. Next, the adsorption gas was purged with N2 until no NH3 or NOx was detected in the effluent. TPD measurements were carried out up to 750 °C with a heating rate of 10 °C/min in flowing N2 with the flow rate of 300 (NH3-TPD) or 100 mL/min (NOx-TPD). The concentration of NH3, NO, and NO2 was continuously monitored by a FTIR spectrometer (Gasmet FTIR DX4000) equipped with a heated, low volume multiple-path gas cell (5 m). Pyridine adsorption infrared spectrum measurements (PyIR) characterization was performed in a NICOLET 6700 FTIR apparatus equipped with a DTGS detector. The spectra were recorded at a resolution of 4 cm−1 and with a scan number of 16. The samples were pressed into self-supporting film for 15− 20 mg, and were pretreated at vacuum condition (450 °C, 1.0 × 10−3 Pa) for 2 h. Next, the temperature was cooled to 90 °C, and pyridine was completely adsorbed at the same temperature. After that, the temperature was increased to the setting point (200 or 350 °C) for vacuum treatment of 20 min (1 × 10−3 Pa).

Table 1. Description of Different Fe-zeolite Catalysts sample

[NOx ]inlet − [NOx ]outlet × 100% [NOx ]inlet

94.6 133.0 118.4

Analyzed by ICP-AES. bCalculated by 3 × (number of iron ions)/ (number of aluminum ions).

a

ions), and the results of iron content are summarized in Table 1. The coating modified Fe-BEA SCR catalysts used in this study were fully formulated wash-coated cordierite monoliths taken from a major catalyst supplier. Fe-BEA cores of 1” diameter × 1” long were coated by a layer of MOR or ZSM-5. 2.2. Activity Test of Catalyst. Catalytic activity tests were performed in a fixed-bed quartz tube reactor of 9 mm internal diameter containing 100 mg of catalyst powder (40−60 mesh). The concentration of NH3 and NOx (NO, NO2, and N2O) in the inlet and outlet gas was measured by a FTIR spectrometer (Gasmet FTIR DX4000) made in Finland. Propene was selected as the HCs in this study due to its high abundance in diesel exhaust gas and relative ease for introduction into the reactor as a gas. At steady state, a flowing N2 mixture containing 500 ppm C3H6 (when used), 500 ppm NO, 500 ppm NH3, and 1748

dx.doi.org/10.1021/es203070g | Environ. Sci. Technol. 2012, 46, 1747−1754


Article

The IR spectrum of the catalyst was recorded from 1700 to 1400 cm−1 at each temperature. The in situ DRIFTS experiment was recorded with a Nicolet Nexus spectrometer equipped with a liquid nitrogen-cooled MCT detector. The samples were purged in a flow of N2 at 500 °C for 30 min, and then cooled to 200 °C. The background spectrum was recorded in flowing N2 and was subtracted from the sample spectrum that was obtained at the same temperature. Thus, the IR absorption features that originated from the structural vibrations of the catalyst were eliminated from the sample spectra. In the experiment, the IR spectra were recorded by accumulating 100 scans at a spectral resolution of 4 cm−1. The gas mixtures contained 500 ppm C3H6, 500 ppm NO, 5% O2, and balance of N2. The total gas flow rate was 100 mL/min (ambient conditions). Thermal analysis was performed in a TGA/DSC 1 system from Mettler-Toledo. The samples were heated in flowing 20 mL/min N2 as protective gas from 100 to 400 °C at a rate of 10 °C/min. When temperature was stable at 400 °C, the weight signals were recorded for fresh catalysts. After 1000 ppm C3H6 as reactive gas was introduced into the reactor for 2 h, the weight signals were recorded for poisoned catalysts.

3. RESULTS 3.1. SCR Performance of Different Catalysts. 3.1.1. Catalytic Activity. Figure 1a shows the catalytic activity of different zeolite catalysts for NH3-SCR. The NOx conversions of Fe-MOR and Fe-BEA were higher than 80% in a wide temperature range of 300−500 °C, and reached nearly 100% at 350−500 °C. In comparison, the Fe-ZSM-5 catalyst showed relatively low catalytic activity and reached the maximum NOx conversion of 74% at 450 °C. The decrease order of NOx conversion was Fe-BEA > Fe-MOR > Fe-ZSM-5. Figure 1a also shows the catalytic activity for NH3-SCR of NOx over different Fe-zeolite catalysts after propene coking. In Figure S1, the NOx conversion of these three Fe-zeolite catalysts becomes stable under the propene atmosphere within 60 min. To avoid the impact of gas adsorption and accumulation of coke on the catalyst samples, NO x concentrations of test data were recorded when the reaction was stable for 120 min. When 500 ppm propene was added to the mixture of reaction gas, the NOx conversion decreased to some extent for these three Fe-zeolite catalysts at various temperatures. It can be seen that NOx conversion of the poisoned catalysts ranked in the sequence of Fe-MOR > FeBEA > Fe-ZSM-5. Fe-MOR has still kept high activity with almost 80% NOx conversion even after C3H6 coking for 120 min, while the activity clearly decreased on the other two catalysts. Although poisoned Fe-ZSM-5 showed the lowest SCR activity, Fe-BEA was most seriously affected by the propene. NOx conversions were all decreased to some extent as propene was injected in the feed gas. On these three catalysts, the SCR activity showed a sharp decrease in the high temperatures (>400 °C), and a small decrease at low temperatures (400 °C). Hence, NOx conversion was significantly affected by propene at high temperatures. The original color of these three Fe-zeolite catalysts was light brown, but it all changed to black after NH3-SCR of NO test

Figure 1. Activity comparison of SCR of NOx over different Fe-zeolite catalysts with or without C3H6. Reaction conditions: 0.1 g of catalyst, 500 ppm NH3, 500 ppm NO, 5% O2, 500 ppm C3H6 (when used), and balance N2, GHSV = 160 000 h−1. (a) NOx conversion on different catalysts; (b) NH3 conversion on different catalysts.

with propene. Separate experiments showed that the catalyst activity of Fe-zeolite catalysts could completely rebound if the coked catalysts were heated at 550 °C in 5% O 2 /N 2 atmosphere. After the regeneration, the color of Fe-zeolite catalysts would change to light brown again. This result indicates that the carbonaceous deposit or coke could be burned off by excess oxygen in the reaction gas. NH3 conversions with and without propene were also summarized during the NH3-SCR reaction on these three catalysts. Figure 1b showed that Fe-BEA and Fe-MOR yielded almost the same NH3 conversion values as NOx conversion values in the whole temperature range, and Fe-ZSM-5 yielded a little higher NH3 conversion than NOx conversion, especially at the temperature above 450 °C, which might be related to the ammonia oxidation at higher temperature.7,23 On the other hand, these three Fe-zeolite catalysts yielded some different features of NH3 conversions in the presence of propene, which showed almost the same NH3 conversion values as NOx conversion values at low temperatures (350 °C). This feature might be due to ammonia oxidation or ammoxidation reactions at high temperatures.17 3.1.2. Coating Modified Fe-zeolite SCR Catalysts with Improved Propene Tolerance. The above results showed that Fe-MOR kept relatively high activity after propene poisoning, 1749



Article

pore diameter did not decrease significantly. This indicated that coke mainly formed on the surface without blocking pores of zeolite. After the poisoned catalysts were regenerated at 550 °C in the 5% O2/N2, the physical properties were almost recovered, and their surface area, pore volume, and pore diameter were all similar to the fresh catalyst. The physical structure of the catalysts were not destroyed with propene poisoning, and the properties of physical structure kept almost the same values on the regenerated catalyst as compared to the fresh catalyst. MOR, ZSM-5, and HBEA have different pore structures. MOR is comprised of two channel types: the larger channels, also called main channels, accessible through 12-ring aperture with an opening of 6.5 × 7.0 Å, and the smaller channels, often referred to as side-pockets, which include 8-ring aperture of 2.6 × 5.7 Å. ZSM-5 and HBEA are three-dimensional high-silica zeolite and are comprised of 10-ring aperture (5.1 × 5.5 Å and 5.3 × 5.6 Å) and 12-ring aperture (6.6 × 6.7 Å and 5.6 × 5.6 Å), respectively.24 Because of the specific pore structure, channels of MOR allow three-dimensional diffusion for small molecules (e.g., N2, NO, NH3, etc.) and monodimensional for larger molecules (e.g., C3H6, etc.).25 As compared to MOR zeolite, ZSM-5 and HBEA, with three-dimensional structures, are much more susceptible to deactivation, because of coke blockage of channels during acid catalytic reactions. Fe-ZSM-5 and Fe-BEA catalysts were more seriously affected by propene in the NH3-SCR reaction (Figure 1a). Hence, the pore structures of these Fe-zeolite catalysts are one of the main factors in the HCs poisoning. 3.2.2. TPD. The results of the NH3-TPD analysis of fresh and poisoned Fe-zeolite powder catalysts are given in Figure S2. Three peaks of NH3 adsorption were observed on Fe-MOR catalysts located at 120, 170, and 470 °C, respectively. The peak at 110 °C could be assigned to physical adsorbed ammonia on Fe-MOR.26 Assignment of the peak at 170 °C was somewhat controversial. It could be attributed to physical adsorbed ammonia,27 weakly adsorbed ammonia on acid sites,26,28 or ammonia associations.29 We carefully speculated that this feature might be attributed to ammonia weakly adsorbed on acid sites. The peak at 470 °C was attributed to the adsorption of NH3 on strong acid sites including Lewis and Brönsted acid sites.16,26,30 It could not reveal any difference between the acidity strengths of Lewis and Brönsted acid sites by the NH3TPD tests,27,29 which would be further explained by the pyridine adsorption section. When Fe-MOR was poisoned, the temperature of the peak intensity slightly decreased, and the total amount of NH3 desorption decreased from 385.1 to 352.5 μmol/g (by 8.5%). Fe-ZSM-5 catalyst also showed three NH3 desorption peaks similar to Fe-MOR. The peak at 140, 220, and 440 °C could be assigned to physical ammonia adsorption, ammonia adsorbed on weak acid sites, and on strong Lewis and Brö nsted acid sites, respectively. When Fe-ZSM-5 was poisoned, the total amount of NH3 desorption decreased from 494.8 to 469.2 μmol/g (by 5.2%). In addition, Fe-BEA showed that two large desorption peaks centered at 210 and 350 °C, which could be assigned to ammonia desorption on weak acid sites and strong Lewis and Brönsted acid sites, respectively. The intensity of NH3 desorption peak decreased from 613.8 to 475.6 μmol/g (by 22.5%) over poisoned Fe-BEA. The above results showed Fe-BEA was most affected by propene, and ammonia could adsorb on the acid sites even in the presence of propene, which might not seriously affect their catalytic activity. Because desorption of NH3 on strong acid

while the NOx activities clearly decreased over the other Fezeolite catalysts. However, the initial activity of fresh Fe-BEA was the highest of all three catalysts. It inspired us to consider that a zeolite, especially MOR, could be used as a protective layer to reduce HCs deactivation of Fe-BEA SCR catalysts. The wash-coated Fe-BEA monolith cores before and after the modification by a top layer coating of ZSM-5 or MOR zeolite were tested for HCs tolerance. Figure 2 shows the NOx

Figure 2. NOx conversion over different Fe-zeolite catalysts before and after 700 ppm C3H6 exposure at 500 °C. Reaction conditions: 350 ppm NO, 350 ppm NH3, 14% O2, 5% CO2, 5% H2O, and balance N2, GHSV = 30 000 h−1.

conversion before and after 700 ppm C3H6 exposure at 500 °C on these catalysts. It is important to note that the fresh activity of Fe-BEA with and without the extra zeolite layer was the same. In the presence of propene, the SCR activity of FeBEA coated by ZSM-5 or MOR decreased by 16.2% and 9.1% at 500 °C, while Fe-BEA without a coating decreased by 32.3%. These results suggest that by adopting such a modification strategy, the potential for Fe-zeolite SCR catalysts for diesel vehicle applications could be increased. 3.2. Characterization of Catalysts. 3.2.1. BET and Pore Structure Analysis of Catalysts. BET surface area and pore structure results are summarized in Table 2. For the Fe-MOR, Table 2. Comparison of BET Surface Area and Pore Structure Results of Fresh, Propene Poisoned, and Regenerated Fe-zeolite Catalysts sample Fe-MOR Fe-MOR-HC Fe-MORregenerated Fe-ZSM-5 Fe-ZSM-5-HC Fe-ZSM-5regenerated Fe-BEA Fe-BEA-HC Fe-BEAregenerated

SBET (m2/ g)

pore volume (cm3/g)

average pore diameter (nm)

420 398 406

0.21 0.18 0.20

0.46 0.46 0.46

399 337 366

0.19 0.16 0.18

0.48 0.47 0.48

510 405 506

0.23 0.20 0.22

0.49 0.49 0.49

Fe-ZSM-5, and Fe-BEA powder catalysts, the surface area was reduced with propene coking, which decreased from 420 to 398 m2/g, 399 to 337 m2/g, and 510 to 405 m2/g, respectively. For these three catalysts, pore volume decreased somewhat, but 1750



Article

sites located at 470 °C on Fe-MOR, at 440 °C on Fe-ZSM-5, and at 350 °C on Fe-BEA, we speculated that the acid strength of these three catalysts ranked in the sequence of Fe-MOR > Fe-ZSM-5 > Fe-BEA. The results of TPD profiles of NO and NO2 for fresh and poisoned Fe-zeolite powder catalysts are given in Figure S3. For the fresh Fe-MOR catalyst, two peaks of NO2 adsorption were observed in the entire temperature range: one large peak was at low temperature of 180 °C, and the other small one was at high temperature of 380 °C. While NO desorbed with only one peak at 330 °C, the peaks of NOx desorption located at low temperatures (300 °C) are attributed to the decomposition of monodentate nitrates and bidentate nitrates, respectively.31 For the poisoned Fe-MOR catalyst, only one peak was observed for NO2 and NO desorption located at the low temperatures of 160 and 190 °C, respectively. These peaks at low temperatures can be ascribed to decomposition of monodentate nitrates. As compared to fresh Fe-MOR catalyst, the desorption amount of NO for poisoned Fe-MOR actually increased, while the desorption amount of NO2 for poisoned Fe-MOR was 92.2 μmol/g much lower than 366.2 μmol/g on fresh Fe-MOR. Additionally, the total desorption amount of NOx of Fe-MOR in the entire temperature range was 271.2 μmol/g for poisoned catalysts and 422.9 μmol/g for fresh catalyst. This indicated that there was almost no chemical adsorption of bidentate nitrates formed on the poisoned Fe-MOR catalysts, and NO was difficult to be activated or oxidized on the poisoned catalyst. For fresh and poisoned Fe-ZSM-5, desorption peaks of NOx were similar to Fe-MOR, which indicated the same mechanism of coke deposition might have occurred. The total desorption amount of NOx of Fe-ZSM-5 in the entire temperature range was 159.5 μmol/g for poisoned catalysts and 303.0 μmol/g for fresh catalyst. For fresh Fe-BEA, only a high temperature desorption peak of NO and NO2 at 400 °C was observed, which can be also assigned to bidentate nitrates decomposition. For poisoned Fe-BEA, the desorption peak of NO2 and NO was observed at 170 and 240 °C, respectively. This illustrated that monodentate nitrate and bridging nitrate species formed on the poisoned FeBEA, and there was no bidentate nitrate formation on the poisoned Fe-BEA.31 This meant NO was completely activated and oxidized on the Fe-BEA surface during the NO + O2 adsorption process. Among these three Fe-zeolite catalysts, the Fe-BEA catalyst had the strongest oxidation ability (Figure S4). NO might be easy to oxidize and transform to nitrate species on the catalyst surface. The total desorption amount of NOx of Fe-BEA in the entire temperature range was 170.4 μmol/g for poisoned catalysts and 249.9 μmol/g for fresh catalyst. 3.2.3. Py-IR. Figure 3 presents the pyridine adsorption infrared spectra of different Fe-zeolite after vacuum treatment at 200 and 350 °C, respectively. On the basis of the literature,32 when pyridine is adsorbed on the Lewis and Brönsted acid sites of the catalysts, vibration peaks will emerge at 1450 and 1540 cm−1, respectively. These two peaks distinguish the types (Lewis or Brönsted), relative amounts, and strength of the acids on the surface of catalysts. On Fe-MOR catalyst, the peaks of Brönsted and Lewis acid simultaneously appeared, and both of them did not significantly change with temperature rising to 350 °C. On Fe-ZSM-5 and Fe-BEA catalysts, the peaks of Brönsted and Lewis acid also simultaneously appeared, and the peaks of Lewis acids decreased, while the peaks of Brönsted acids changed slightly when increasing the temperature to 350 °C. This indicated that the acid strength of the Lewis sites on

Figure 3. Py-IR spectra of Fe-MOR (a), Fe-ZSM-5 (b), and Fe-BEA (c) after vacuum treatment at 200 and 350 °C.

Fe-ZSM-5 and Fe-BEA catalysts was weak, while that of Brönsted sites was strong. The peak strength of Fe-MOR and Fe-BEA presented a greater amount of Lewis acids sites than Brönsted sites at 200 and 350 °C, but the peak strength of FeZSM-5 presented the opposite. The above results implied that 1751



Article

4. DISCUSSION The NH3-SCR reaction mechanism on iron-exchanged zeolite was deeply studied in the past decade.9,12,36,37 On Fe-MOR and Fe-ZSM-5, Yang et al.12,36 suggested that ammonia was adsorbed on the Brönsted acid sites and activated to NH4+. NO was oxidized to NO2 on the iron active sites. Next, NO2 reacted with two adjacent NH4+ to form NO2(NH4+)2, which further reacted with NO and decomposed to the SCR products N2 and H2O. On Fe-BEA, Kureti et al.37 proposed that a dual site mechanism implying the adsorption and reaction of NO and NH3 on neighboring Fe3+ sites. NH3 undergoes several adsorption/desorption cycles on the HBEA substrate before adsorbing and reacting on the Fe3+ sites. The uptake of NH3 results in partial reduction of Fe3+ sites, which are finally recycled by O2. In summary, it is accepted that the SCR reaction over iron-exchanged zeolite needs two kinds of sites: the acidic sites of zeolite support for ammonia adsorption and activation, and the iron sites for NOx activation. The activation species of ammonia and NOx then form some complex, which finally yields nitrogen and water. To deeply understand the difference of propene poisoning effect of these three Fe-zeolite catalysts, these two kinds of sites should be considered and discussed. Acidity amount and strength are two main factors for the zeolite catalyst. On the basis of the pyridine adsorption and NH3-TPD results (Figure 3 and Figure S2), the total amount of zeolite acidity ranked as Fe-BEA > Fe-ZSM-5 > Fe-MOR, and the strength of zeolite acidity just showed the opposite, and ranked as Fe-MOR > Fe-ZSM-5 > Fe-BEA. The reason was that the number of acid sites is determined by the aluminum tetrahedrally coordinated in the framework, and the strength of acidity is independent of the number of acid sites, and determined by the zeolite structure.38 In TG test with propene (Figure S6), Fe-MOR, Fe-ZSM-5, and Fe-BEA catalysts showed a completely different amount of carbon deposition, which ranked in the sequence of Fe-BEA > Fe-ZSM-5 > FeMOR. The sequence was the same as the sequence of total acidity number. Considering the hydrocarbons were most activated on the acid sites, we speculated that the acidity numbers of Fe-zeolite were most related to the carbon deposition. The strength of acidity played a role of adsorption and activation ability and was not related to total carbon deposition number. It can be seen that both pore structure and acidity of the zeolite have effects on the carbon deposition. Guisnet et al.39 found that the roles of strength and density of active acid sites (often protonic sites) were generally limited in comparison to the roles played by size and shape of cavities (or channel intersections) and by the size of their apertures. This might be the reason that Fe-MOR was not seriously affected by the propene. First and foremost, Fe-MOR catalyst only has onedimensional pore structure, which is not conducive to propene diffusion. Second, Fe-MOR catalyst has relatively lower acidity number, which is an important factor in the propene activation. In our results, Fe3+ active sites on which the NO was oxidized to NO2 were possibly poisoned in the SCR reaction (Figures S3 and S4). The ability of NO oxidation to NO2 was inhibited on propene poisoned catalyst below 350 °C, and the oxidation of NO to NO2 recovered and reached the activity of fresh catalyst above 400 °C (Figure S4). This further demonstrated that the Fe3+ active sites were hindered by carbonaceous deposits, and nitrate species had difficulty forming on the poisoned Fe3+ sites,

the acid strength of Fe-MOR was strongest among these three catalysts, which further demonstrated the speculation in NH3TPD. In addition, the total acid amount of Fe-zeolites, based on the total peak area of Lewis and Brönsted sites, ranked in the sequence of Fe-BEA > Fe-ZSM-5 > Fe-MOR. The carbonaceous deposits on Fe-zeolite might be related to the existence of Fe-zeolite acid sites, which played an important role in the coking process.19 3.2.4. DRIFTS. In our previous study,16 the band at 2250 cm−1 was formed when the NO + O2 + C3H6 were introduced on the Fe-ZSM-5 catalyst, which represented isocyanate (−NCO) generation. To demonstrate propene effect on the SCR reaction, DRIFTS spectra were recorded as a function of time after 500 ppm C3H6 + 500 ppm NO + 5% O2/N2 was introduced into the cell at 200 °C. Transient reaction results (Figure S5) showed formate species (around 3127 and 2937 cm−1) and isocyanate (around 2261 cm−1) mainly formed on these three Fe-zeolite catalysts.33,34 The negative bands with absorbance appeared at around 3603 cm−1 during exposure to NO + C3H6 + O2 could be interpreted as that the formation of nitrates and acetate resulted in the consumption of OH group.35 In addition, the nitrate species bands were very weak on the these three Fe-zeolite catalysts when the C3H6, NO, and O2 were introduced at the same time, so wavenumber below 2000 cm−1 was not given here. Figure 4 summarizes the results

Figure 4. Formation of −NCO (indicated by band area around 2250 cm−1) at 200 °C upon passing 500 ppm C3H6 + 500 ppm NO + 5% O2/N2 over Fe-MOR, Fe-ZSM-5, and Fe-BEA catalyst.

of the formation of isocyanate (indicated by band area around 2250 cm−1) at 200 °C. After C3H6 + NO + O2/N2 was passed over Fe-MOR, the catalyst surface was gradually dominated by the adsorbed isocyanate. The IR band due to isocyanate increased and became stable until 15 min. By comparison, the increased trend of the isocyanate band of Fe-ZSM-5 was slightly higher than that of Fe-MOR when the same gases were passed over the Fe-ZSM-5. Yet for Fe-BEA catalyst, isocyanate peak intensity continued to increase, and the isocyanate content on the surface was significantly higher than for the other two catalysts after 30 min. These results prove that NO + O2 are relatively inactive in reacting with propene, and less isocyanate forms during the reaction on Fe-MOR. 1752



Article

for SCR of NOx with propene under lean-burn conditions. Appl. Catal., B 2009, 90, 416−425. (5) Johnson, T. Diesel engine emissions and their control: An overview. Platinum Met. Rev. 2008, 52, 23−37. (6) Eichelbaum, M.; Farrauto, R. J.; Castaldi, M. J. The impact of urea on the performance of metal exchanged zeolites for the selective catalytic reduction of NOx: Part I. Pyrolysis and hydrolysis of urea over zeolite catalysts. Appl. Catal., B 2010, 97, 90−97. (7) Busca, G.; Lietti, L.; Ramis, G.; Berti, F. Chemical and mechanistic aspects of the selective catalytic reduction of NOx by ammonia over oxide catalysts: A review. Appl. Catal., B 1998, 18, 1− 36. (8) Heck, R. M. Catalytic abatement of nitrogen oxides-stationary applications. Catal. Today 1999, 53, 519−523. (9) Brandenberger, S.; Kröcher, O.; Tissler, A.; Althoff, R. The state of the art in selective catalytic reduction of NOx by ammonia using metal-exchanged zeolite catalysts. Catal. Rev.-Sci. Eng. 2008, 50, 492− 531. (10) Ma, A. Z.; Grunert, W. Selective catalytic reduction of NO by ammonia over Fe-ZSM-5 catalysts. Chem. Commun. 1999, 1, 71−72. (11) Long, R. Q.; Yang, R. T. Superior Fe-ZSM-5 catalyst for selective catalytic reduction of nitric oxide by ammonia. J. Am. Chem. Soc. 1999, 121, 5595−5596. (12) Long, R. Q.; Yang, R. T. Selective catalytic reduction of NO with ammonia over Fe3+-exchanged mordenite (Fe-MOR): Catalytic performance, characterization, and mechanistic study. J. Catal. 2002, 207, 274−285. (13) Balle, P.; Geiger, B.; Kureti, S. Selective catalytic reduction of NOx by NH3 on Fe/HBEA zeolite catalysts in oxygen-rich exhaust. Appl. Catal., B 2009, 85, 109−119. (14) Frey, A. M.; Mert, S.; Due-Hansen, J.; Fehrmann, R.; Christensen, C. H. Fe-BEA zeolite catalysts for NH3-SCR of NOx. Catal. Lett. 2009, 130, 1−8. (15) He, C.; Wang, Y.; Cheng, Y.; Lambert, C. K.; Yang, R. T. Activity, stability and hydrocarbon deactivation of Fe/Beta catalyst for SCR of NO with ammonia. Appl. Catal., A 2009, 368, 121−126. (16) Li, J.; Zhu, R.; Cheng, Y.; Lambert, C. K.; Yang, R. T. Mechanism of propene poisoning on Fe-ZSM-5 for selective catalytic reduction of NOx with ammonia. Environ. Sci. Technol. 2010, 44, 1799−1805. (17) Heo, I.; Lee, Y.; Nam, I.-S.; Choung, J. W.; Lee, J.-H.; Kim, H.-J. Effect of hydrocarbon slip on NO removal activity of CuZSM5, FeZSM5 and V2O5/TiO2 catalysts by NH3. Microporous Mesoporous Mater. 2011, 141, 8−15. (18) Isarangura na ayuthaya, S.; Mongkolsiri, N.; Praserthdam, P.; Silveston, P. L. Carbon deposits effects on the selective catalytic reduction of NO over zeolites using temperature programmed oxidation technique. Appl. Catal., B 2003, 43, 1−12. (19) Bartholomew, C. H. Mechanisms of catalyst deactivation. Appl. Catal., A 2001, 212, 17−60. (20) Devarakonda, M.; Tonkyn, R.; Tran, D.; Lee, J.; Herling, D. Modeling species inhibition of NO oxidation in urea-SCR catalysts for diesel engine NOx control. J. Eng. Gas Turbines Power 2011, 133, 092805. (21) Feng, X.; Keith Hall, W. FeZSM-5: A durable SCR catalyst for NOx removal from combustion streams. J. Catal. 1997, 166, 368−376. (22) Long, R. Q.; Yang, R. T. Characterization of Fe-ZSM-5 catalyst for selective catalytic reduction of nitric oxide by ammonia. J. Catal. 2000, 194, 80−90. (23) Brandenberger, S.; Krocher, O.; Tissler, A.; Althoff, R. The determination of the activities of different iron species in Fe-ZSM-5 for SCR of NO by NH3. Appl. Catal., B 2010, 95, 348−357. (24) http://www.iza-structure.org/databases/. (25) de Macedo, J. L.; Dias, S. C. L.; Dias, J. A. Multiple adsorption process description of zeolite mordenite acidity. Microporous Mesoporous Mater. 2004, 72, 119−125. (26) Lobree, L. J.; Hwang, I.-C.; Reimer, J. A.; Bell, A. T. Investigations of the state of Fe in H-ZSM-5. J. Catal. 1999, 186, 242−253.

and there was very little desorption of NO2 over the poisoned catalyst at high temperatures. As a result, the activity considerably decreased due to the poor activation of NO, which is the rate determining step in the NH3-SCR of NOx.12,36 Carbonaceous deposition on zeolites is a very complex chemical reaction and occurs through oligomerization of the olefinic cracking products followed by cyclization, etc., and finally yields biaromatics, triaromatics, etc.19,40 In the process of NH3-SCR with propene on Fe-zeolite, the final carbonaceous deposit product obviously inhibited the NOx activation. Moreover, by introducing propene into NH3-SCR reaction, some nitrogenated byproducts, like isocyanate (−NCO), etc., would be formed during the coke forming process and NH3SCR reaction. These species probably deposited on the catalyst surface and then hindered the SCR reaction. Some researchers41,42 considered that −NCO was also an important intermediate, which could increase NOx conversion in SCR reaction. Yet this phenomenon was always observed on the catalysts that were active in HC-SCR reaction. All of these three Fe-zeolite catalysts showed almost no activity in the C3H6-SCR reaction at all temperatures (figure not given), which further demonstrated that −NCO was not a key intermediate in NH3SCR in the presence of propene. Although almost all kinds of Fe-zeolite SCR catalysts deactivated in the presence of propene, Fe-MOR kept relatively higher SCR activity as compared to the other two Fe-zeolites. On the basis of these results, we devised a strategy to modify a conventional Fe-BEA SCR catalyst in ways that could improve the HCs tolerance. The results showed that the deactivation of coating modified Fe-BEA SCR catalyst due to HCs poisoning was reduced and then the DeNOx efficiency was improved. The results suggest that Fe-BEA or even other kinds of Fe-zeolite SCR catalysts could have a potential application for diesel emission control by adopting such a modification strategy.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional figures. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +86 10 62771093; fax: +86 10 62771093; e-mail: [email protected].

■

ACKNOWLEDGMENTS We appreciate Amy Alice Chastain for her assistance in proofreading. This work was financially supported by the National Natural Science Foundation of China (Grant No. 51078203), the National High-Tech Research and Development (863) Program of China (Grant No. 2010AA065002), and Ford Motor Co.

■

REFERENCES

(1) Bosch, H.; Janssen, F. Preface. Catal. Today 1988, 2, v. (2) Parvulescu, V. I.; Grange, P.; Delmon, B. Catalytic removal of NO. Catal. Today 1998, 46, 233−316. (3) Granger, P.; Parvulescu, V. I. Catalytic NOx abatement systems for mobile sources: From three-way to lean burn after-treatment technologies. Chem. Rev. 2011, 111, 3155−3207. (4) Jagtap, N.; Umbarkar, S. B.; Miquel, P.; Granger, P.; Dongare, M. K. Support modification to improve the sulphur tolerance of Ag/Al2O3 1753



Article

(27) Niwa, M.; Katada, N.; Sawa, M.; Murakami, Y. Temperatureprogrammed desorption of ammonia with readsorption based on the derived theoretical equation. J. Phys. Chem. 1995, 99, 8812−8816. (28) Hunger, B.; Hoffmann, J.; Heitzsch, O.; Hunger, M. Temperature-programmed desorption (TPD) of ammonia from HZSM-5 zeolites. J. Therm. Anal. Calorim. 1990, 36, 1379−1391. (29) Lónyi, F.; Valyon, J. On the interpretation of the NH3-TPD patterns of H-ZSM-5 and H-mordenite. Microporous Mesoporous Mater. 2001, 47, 293−301. (30) Camiloti, A. M.; Jahn, S. L.; Velasco, N. D.; Moura, L. F.; Cardoso, D. Acidity of Beta zeolite determined by TPD of ammonia and ethylbenzene disproportionation. Appl. Catal., A 1999, 182, 107− 113. (31) Liu, F.; Asakura, K.; He, H.; Liu, Y.; Shan, W.; Shi, X.; Zhang, C. Influence of calcination temperature on iron titanate catalyst for the selective catalytic reduction of NOx with NH3. Catal. Today 2011, 164, 520−527. (32) Lercher, J. A.; Gründling, C.; Eder-Mirth, G. Infrared studies of the surface acidity of oxides and zeolites using adsorbed probe molecules. Catal. Today 1996, 27, 353−376. (33) Chafik, T.; Kameoka, S.; Ukisu, Y.; Miyadera, T. In situ diffuse reflectance infrared Fourier transform spectroscopy study of surface species involved in NOx reduction by ethanol over alumina-supported silver catalyst. J. Mol. Catal. A: Chem. 1998, 136, 203−211. (34) Li, J.; Zhu, Y.; Ke, R.; Hao, J. Improvement of catalytic activity and sulfur-resistance of Ag/TiO2-Al2O3 for NO reduction with propene under lean burn conditions. Appl. Catal., B 2008, 80, 202− 213. (35) Shimizu, K.-i.; Kawabata, H.; Satsuma, A.; Hattori, T. Role of acetate and nitrates in the selective catalytic reduction of NO by propene over alumina catalyst as investigated by FTIR. J. Phys. Chem. B 1999, 103, 5240−5245. (36) Long, R. Q.; Yang, R. T. Reaction mechanism of selective catalytic reduction of NO with NH3 over Fe-ZSM-5 catalyst. J. Catal. 2002, 207, 224−231. (37) Klukowski, D.; Balle, P.; Geiger, B.; Wagloehner, S.; Kureti, S.; Kimmerle, B.; Baiker, A.; Grunwaldt, J. D. On the mechanism of the SCR reaction on Fe/HBEA zeolite. Appl. Catal., B 2009, 93, 185−193. (38) Miyamoto, Y.; Katada, N.; Niwa, M. Acidity of β zeolite with different Si/Al2 ratio as measured by temperature programmed desorption of ammonia. Microporous Mesoporous Mater. 2000, 40, 271−281. (39) Guisnet, M.; Magnoux, P.; Martin, D. Roles of acidity and pore structure in the deactivation of zeolites by carbonaceous deposits. In Catalyst Deactivation 1997; Bartholomew, C. H., Fuentes, G. A., Eds.; 1997; pp 1−19. (40) Guisnet, M.; Magnoux, P. Coking and deactivation of zeolites: Influence of the Pore Structure. Appl. Catal. 1989, 54, 1−27. (41) Shichi, A.; Hattori, T.; Satsuma, A. Involvement of NCO species in promotion effect of water vapor on propane-SCR over Co-MFI zeolite. Appl. Catal., B 2007, 77, 92−99. (42) Jagtap, N.; Umbarkar, S. B.; Miquel, P.; Granger, P.; Dongare, M. K. Support modification to improve the sulphur tolerance of Ag/ Al2O3 for SCR of NOx with propene under lean-burn conditions. Appl. Catal., B 2009, 90, 416−425.

1754


Propene Poisoning on Three Typical Fe-zeolites for SCR of NOx with

Recommend Documents