Addition of Pd on La0.7Sr0.3CoO3 Perovskite To Enhance Catalytic

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

Addition of Pd on La0.7Sr0.3CoO3 perovskite to enhance catalytic removal of NOx Dongyue Zhao, Zhongnan Gao, Hui Xian, Lingli Xing, Yuexi Yang, Ye Tian, Tong Ding, Zheng Jiang, Jing Zhang, Li-Rong Zheng, and Xingang Li Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.7b04399 • Publication Date (Web): 21 Dec 2017 Downloaded from http://pubs.acs.org on December 21, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Addition of Pd on La0.7Sr0.3CoO3 Perovskite to Enhance Catalytic Removal of NOx Dongyue Zhao1, Zhongnan Gao1, Hui Xian2, Lingli Xing1, Yuexi Yang1, Ye Tian1, Tong Ding1, Zheng Jiang3, Jing Zhang4, Lirong Zheng4, Xingang Li1*

1 Collaborative Innovation Center of Chemical Science & Engineering (Tianjin), Tianjin Key Laboratory of Applied Catalysis Science & Technology, School of Chemical Engineering & Technology, Tianjin University, Tianjin, 300072, P. R. China 2 School of Continuing Education, Tianjin Polytechnic University, Tianjin, 300387, P. R. China 3 Shanghai Synchrotron Radiation Facility, Shanghai Institute of Applied Physics, Chinese Academy of Sciences, Shanghai 201800, PR China 4 Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, PR China

* Corresponding author: Prof. Xingang Li EMAIL: [email protected]

1

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Herein, we report the function of Pd species in the Pd-added La0.7Sr0.3CoO3 perovskite for removal of NOx (NO+NO2) in the lean-burn NOx traps (LNT) process. The addition of Pd significantly improves the De-NOx activity of the perovskite-based catalysts. The Pd-supported catalyst has an excellent NOx removal efficiency of 90.4 %, which is higher than the Pd-doped catalyst (74.4 %) and pure La0.7Sr0.3CoO3 (51.6 %). Our results show that the addition of Pd species enhances both the NOx desorption behaviors to regenerate the NOx storage sites and NOx reduction ability of the catalysts in the fuel-rich period, which induces the improved NOx trapping performance in the followed lean-burn period. These achievements result in the enhanced De-NOx activity of the Pd-added catalyst in the alternative NOx storage and reduction operations. Compared with the Pd-doped catalyst, the Pd-supported catalyst possesses the higher utilization efficiency of the Pd, thus exhibiting a higher catalytic activity.

Keyword: Perovskite; Pd; Reduction of NOx; Lean-burn; NOx trapping

2


Page 2 of 43

Page 3 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction Three-way catalysts (TWC) can not meet the increasingly strict exhaust regulation of NOx for lean-burn engines, where excessive oxygen is fed into to improve the energy efficiency and reduce CO2 emission.1,

2

Among feasible

alternative methods, lean NOx traps, also named as NOx storage and reduction (NSR), is presently an attractive technique for removal of NOx emitted from lean-burn engines.3-7 The mechanism over LNT catalysts can be briefly described as follows: in a long lean-burn period, the reaction process commonly involves the oxidation of NO to NO2 and then the formation of nitrate or nitrite.6, 8 In a subsequent short fuel-rich period, the trapped NOx is desorbed and then reduced with assistance of reductants.9, 10

The cyclic operations between the above-mentioned two phases realize the

reduction of toxic NOx. Conventional LNT catalysts contain Pt as the main active component, which, however, is quite expensive.8, 11 Within the last two decades, researchers found that ABO3 perovskite materials were a promising substitute to the noble metal based catalysts because of the former’s higher NO oxidation ability, lower cost and higher thermal stability.12-16 They are also highly flexible for the composition, by partial substitution of A and/or B site cations.17-19 We previously reported the LaxSr1-xCoO3 perovskite as a promising LNT catalyst, which exhibited the high NO to NO2 conversion and N2 selectivity.20 In the study, strontium is doped into the A-site to increase the NO oxidation ability and acts as the NOx storage sites of the catalysts. However, the De-NOx efficiency of the pure perovskite catalysts is still lower than the 3



Pt-based LNT catalysts.20, 21 Zhang’s report revealed that the addition of precious metal Pd into perovskite system significantly enhanced the NO reduction behaviors in the NOx selective catalytic reduction (SCR) process.22 Moreover, Pd is much cheaper and exhibits higher thermal stability than Pt.23 Thus, adding Pd may be a promising way to enhance the De-NOx activity of the perovskite catalysts in the LNT system. Aiming to achieve the higher NOx removal efficiency, many researchers explored the feasible combination ways of Pd and perovskite. Y. Nishihata et al.24 doped Pd into perovskite lattices, and observed the self-regeneration of Pd0/Pd2+ in and out of perovskite lattices when switching between oxidizing and reducing atmospheres, which prevents from agglomeration of Pd during reactions. Our previous work25 revealed that the Pd-doped La0.7Sr0.3CoO3 perovskite LNT catalyst displayed a high De-NOx activity in alternative lean-burn/fuel-rich cycles, as well as a high sulfur tolerance. The better thermal stability of Pd than Pt allows supporting to be the other feasible Pd-introducing way for perovskite catalysts. For example, the strong interactions between the supported Pd species and perovskite surface result in a remarkable catalytic performance in decomposing N2O.26 According to some recent literature reports, the different Pd-addition ways will lead to discrepancy in the existence form of Pd species: the supported Pd mainly lies on the surface of perovskite as PdO species, while the doped Pd enters the B-site of perovskite with a higher valence (Pd+3 and/or Pd+4).26-29 Zhou et al.30 made a comparison of the above two Pd-addition ways for the perovskite. They found that the Pd-supported La0.8Fe0.2CoO3 catalyst showed the higher activity than the Pd-doped La0.8Fe0.2CoO3 4


Page 4 of 43

Page 5 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


catalyst in three-way catalysis. Presently, there is still a lack of similar comparison in the LNT system. Thus, a more comprehensive understanding of the contribution of the different Pd species is indispensable to develop the highly efficient LNT catalysts. Herein, we reported a comprehensive comparison of two kinds of the Pd-contained perovskite-based catalysts, i.e. Pd-doped La0.7Sr0.3Co0.97Pd0.03O3 and Pd-supported Pd/La0.7Sr0.3CoO3, in the aspects of the NO oxidizability and NOx storage capacity in lean-burn atmosphere, and NOx reduction ability in fuel-rich atmosphere. The effects of the two kinds of the Pd species on the De-NOx activity in the LNT process were investigated and demonstrated in detail. Several experiments were designed to evaluate the NOx trapping performance in the lean-burn periods and NOx desorption and reduction performance in the fuel-rich periods over the perovskite catalysts after the addition of Pd. A new insight is provided into the functions of the Pd species in the La0.7Sr0.3CoO3 based perovskite in the LNT process.

2. Experimental 2.1 Preparation of the catalysts. The Pd-free La0.7Sr0.3CoO3 and Pd-doped La0.7Sr0.3Co0.97Pd0.03O3 catalysts, denoted as LSC and LSCP, respectively, were prepared via a sol-gel process.31, 32 The citric acid (CA), ethylene diamine tetraacetic acid (EDTA), La(NO3)3·6H2O, Sr(NO3)2, Co(NO3)2·6H2O and Pd(NO3)2·2H2O (supplied by Shanghai Aladdin Industrial Corporation) was added into 300 mL deionized water as the stoichiometric amount. The molar ratio of metal cations: EDTA: CA is 1:1:1.5. A certain amount of

5



Page 6 of 43

ammonium hydroxide was also added, enabling the rapid dissolution of EDTA. The slurry was stirred for 8 h at 80 ºC with a constant pH value of 7 to slowly evaporate the liquid water and form a viscous gel. The obtained precursor was then dried at 80 ºC in the vacuum oven for 12 h. The formed powder was calcined at 300 ºC in a muffle furnace for 2 h to decompose the CA and nitrate species. Subsequently, a further calcination was complemented at 700 ºC for 6 h in air flow to form the perovskite structure. The Pd-supported Pd/La0.7Sr0.3CoO3 catalyst, denoted as P/LSC, was synthesized via the wet impregnation to support Pd on the as-prepared LSC catalyst. The loading amount of Pd was the same with the LSCP catalyst at 1.4 wt.%. The powder was dried in a heat oven at 120 ºC for 12 h, and then calcined at 500 ºC for 4 h in air flow.

2.2 Characterizations of the catalysts. The measurements of specific surface areas (SBET) of the catalysts were conducted through nitrogen physisorption at -196 ºC using a Quantachrome QuadraSorb SI instrument. The contents of Pd in the final catalysts were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, VISTA-MPX). The X-ray diffraction (XRD) analysis was carried out on an X’pert Pro diffractometer (PANAlytical Company) using Cu Kα as the radiation source (λ = 0.15418 nm), with a step size of 0.02°. The average grain size of the perovskite phase was calculated by using Sherrer equation: D = K λ / B cosθ

(1) 6


Page 7 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


where D is the grain size, K is the Scherrer constant, λ is the wavelength, B is the half width of the diffraction peak (rad) and θ is the diffraction angle. The X-ray absorption fine structure (XAFS) data of the Pd and Co k-edge were collected at the 14W1 beamline of Shanghai Synchrotron Radiation Facility (SSRF) (300 mA and 3.5 GeV) and 1W1B beamline of Beijing Synchrotron Radiation Facility (BSRF) (120 mA and 2.5 GeV), respectively. The radial structure functions (RSFs) of Pd and Co K-edge were calculated by using a Fourier transforming method. The X-ray photoelectron spectroscopy (XPS) tests were conducted on a PHI-1600 ESCA spectrometer. Mg Kα was the radiation source (1253.6 eV), keeping the base pressure at 5×10−8 Pa. The binding energies (BE) were adjusted by the characteristic BE peak of C1s locating at 284.6 eV. The H2 temperature programmed reduction (H2-TPR) experiments were conducted on a TPDRO (TP-5080, Xianquan Co.) equipped with a thermal conductivity detector (TCD). 30 mg of the catalyst powder was used and heated from room temperature to 900 ºC, with a heating rate of 10 ºC min-1. A gaseous mixture of 8 vol.% H2/N2 was introduced into the quartz tube at a flow rate of 30 mL min-1. The NOx temperature-programmed desorption (NOx-TPD) experiments were conducted in the fixed bed continuous flow reactor (i.d. = 7.4 mm). Before the NOx-TPD experiments, the catalysts has been saturated with NOx in the lean-burn phase (400 ppm NO/5 vol.% O2/N2, 400mL min-1). Then, the desorption was undertaken in pure N2 flow with a flow rate of 400 mL min-1. The temperature range covered from 100 to 700 ºC with a heating rate of 5 ºC min-1. 7



Page 8 of 43

2.3 Catalytic activity measurements. NOx storage capacity (NSC) measurements were carried out within a fixed bed continuous flow reactor at 300 ºC, which is the optimal temperature for the La0.7Sr0.3CoO3 based perovskite in our previous work.33 Before the test, 200 mg of the fresh catalyst particles (40-60 mesh) was pretreated in the 5 vol.% O2/N2 flow (400 mL min-1) at 300 ºC for 0.5 h to remove the absorbed CO2 and H2O. Afterwards, the gas mixture containing 400 ppm NO/5 vol.% O2/N2 was introduced into the fixed bed reactor, with a flow rate of 400mL min-1 (space velocity: 120,000 mL g-1 h-1) for 1.5 h. The on-line chemiluminescence NOx analyzer (model 42i-HL, Thermal Scientific) was adopted to monitor the outlet concentration of NO, NO2 and NOx. The cyclic NOx storage and reduction tests with a lean-burn phase of 50 s and a fuel-rich phase of 10 s were conducted at 250-400 ºC in the reactor loaded with 200 mg of the catalyst. The lean-burn phase gas included 400 ppm NO/5 vol.% O2/N2, while the fuel-rich phase gas contained 1000 ppm C3H6/N2. The flow rate was kept at 400 mL min-1 in both the lean-burn and fuel-rich phases (space velocity: 120,000 mL g-1 h-1). The main by-product N2O was measured by an N2O Modular Gas Analyzer (S710, SICK MAIHAK). The NOx storage efficiency, NOx reducing efficiency and NOx removal percentage (NRP) in the 21th cycle were calculated by the equation (2) (3) (4), respectively. 50

NOx storage efficiency = (1- ∫0 [NOx]out ×t dt/50×[NOx]in)×100 %

(2)

NOx reducing efficiency = 60

50

(1- ∫50 [NOx]out ×t dt/(50×[NOx]in- ∫0 [NOx]out ×t dt) )×100 % 8


(3)

Page 9 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


NPR = NOx storage efficiency × NOx reducing efficiency ×100 %

(4)

where [NOx] out is the outlet NOx concentration (ppm), [NOx] out is the inlet NOx concentration (ppm) and t is time (s).

2.4 Kinetic Tests. The kinetic tests were conducted in the same micro-reactor as mentioned in Section 2.3, using 100 mg of the catalyst. A gas mixture, composed of 400 ppm NO/5 vol.% O2/N2, was fed into the quartz tube at a flow rate of 400 mL min -1. A space velocity of 240,000 mL g-1 h-1 was achieved to eliminate the impact of internal and external diffusion.34 The NO-to-NO2 conversion was maintained below 15 % to keep the reaction in the kinetic regime. The reaction rates for NO oxidation were calculated based on the equation (5): r (μmol g-1 s-1) = 106×[NO2] out×Q /(NA×m)

(5)

where Q is the volumetric flow rate (mL s-1), [NO2] out is the outlet concentration of NO2 (ppm), NA is the Avogadro constant and m is the weight of the catalyst (g).

3. Results and discussions 3.1 Structure properties of the synthesized catalysts. Figure 1 shows the XRD patterns of the LSC, LSCP and P/LSC catalysts. In Figure 1, the perovskite is the main phase for all the catalysts [JCPDS 48-0137]. The XRD patterns between 31 ºand 35 ºare zoomed in and presented in Figure 1b, where a 0.2 ºshift to lower degrees can be clearly observed for the LSCP catalyst compared with the LSC catalyst. However, no obvious shift is observed for the P/LSC catalyst.

9



The change in the 2θ value of the strongest diffraction peak of the LSCP catalyst is mainly due to the lattice expansion during the substitution of Co3+ cations by Pdn+ cations (n≥2) with a larger ionic diameter, which confirms the successful doping of Pdn+ into B-site of the perovskite lattice.26 Besides the perovskite phase, a weak diffraction peak belonging to SrCO3 is also detected in Figure 1a [JCPDS 74-1491], whose formation results from the combustion of EDTA and CA in the catalyst synthesis process.33 A little Co3O4 phase is observed with the weak diffraction peaks appearing at 2θ = 36.9 ºand 44.8 º[JCPDS 42-1467]. However, no diffraction peak belonging to Pd or PdO is detected in the XRD pattern of any catalyst. Probably, the particle size of the Pd specie is below the XRD detection limit (5 nm) or the Pd is incorporated into the perovskite lattice. The average grain size of the perovskite phase was calculated by using Sherrer equation, and Table 1 lists the relevant results. It shows that the addition of Pd does not bring an apparent influence on the surface area or grain size of the catalysts.

10


Page 10 of 43

Page 11 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1 XRD patterns of the fresh LSC, LSCP and P/LSC catalysts: (a) full patterns; (b) zoom-in of the squared area in (a).

Table 1 The physico-chemical properties of the catalysts.

Pd 3d5/2 (eV)

Pd/La (atomic ratio)

D (nm)b

SBET (m2 g-1)

-

-

-

15.7

18.3

LSCP

1.4

337.7

0.02

17.8

18.5

P/LSC

1.4

336.8

0.07

15.0

17.1

Catalyst

Pd content (wt.%)a

LSC

a

Determined by ICP-OES.

b

D: the average grain size calculated by the Scherrer equation with the (1 0 4) facet.

Figure S1 shows the XPS spectra of Pd 3d core levels of the LSCP and P/LSC catalysts, and Table 1 gives the relative data. For the P/LSC catalyst, the binding energy (BE) position of Pd 3d5/2 peak is centered at 336.8 eV, close to the BE value for Pd2+.35-37 For the LSCP catalyst, the BE position of Pd 3d5/2 peak appears at 337.6 eV, in the range of Pd3+ and/or Pd4+,35-37 suggesting that Pd mainly dissolves into the crystal lattices of the LSC perovskite. Additionally, the Pd/La atomic ratio of the P/LSC catalyst (Pd/La: 0.07) is much higher than that of the LSCP catalyst (Pd/La: 0.02) in Table 1. It also indicates that the Pd2+ species lies on the surface of the perovskite in the P/LSC catalyst, while the Pd3+ and/or Pd4+ species majorly locates in the bulk of the perovskite in the LSCP catalyst. Figure 2 presents the radial distribution functions (RDFs) of Pd K-edge gained 11



by the Fourier transform of the EXAFS data, with the aim of identifying the existence state of the doped and supported Pd species. Only one coordination peak appearing at 2.46 Å is observed in the case of Pd-foil, which is attributed to the first metallic Pd-Pd coordination shell and does not appear in the RDFs of the LSCP and P/LSC catalysts. The RDFs of the reference catalyst PdO show two major coordination peaks at 1.57 Å and 2.90 Å, referred to the first Pd-O shell and Pd-Pd shell, respectively. The Pd-supported P/LSC catalyst shows two peaks at the exactly same coordination distances of PdO, confirming that PdO is the main existence form of the Pd species in the P/LSC. However, the intensity of its first Pd-Pd shell apparently shrinks compared with the reference PdO, suggesting that the PdO species on the surface of the P/LSC catalyst is highly dispersed. For the LSCP catalyst, the first Pd-Pd shell is not observed. The disappearance of the Pd-Pd coordination can be regarded as a direct evidence for the successful doping of Pd into the perovskite lattices.27, 29 Figure S2 presents the Co k-edge XANES spectra and RDFs profile of the catalysts to determine the valence and coordination environment of Co ions. The results reveal that the addition of Pd does not make an observable difference with the LSC on the valance and coordination environment of Co for both the P/LSC and the LSCP catalysts.

12


Page 12 of 43

Page 13 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2 Pd K-edge radial distribution functions of the fresh catalysts.

3.2 Redox behaviors and NOx storage capacities of the catalysts. To illuminate the impacts of the different Pd species (doped Pd and supported Pd) on NO oxidation in lean-burn atmospheres, the kinetic tests for NO oxidation were conducted on the LSC, LSCP and P/LSC catalysts to determine the apparent activation energy (Ea), as presented in Figure 3. Clearly, the LSCP catalyst exhibits a higher NO oxidation rate in the kinetic regime, and lower Ea than the other two. It is consistent with the previous findings that doping with Pd leads to an enhancement on NO and CO oxidation.38 According to Zhou’s report30,the improved NO oxidation ability of the LSCP catalyst is probably due to the non-stoichiometric compositions and the generation of oxygen vacancies resulting from the doping of Pd and Sr. However, it will not lead to the change of the Co oxidation state.13 As a result, no obvious valence change is observed from the Co k-edge XANES spectra in Figure S2. Differently from the LSCP catalyst, the supported PdO species in the P/LSC catalyst does not bring an enhancement in the NO oxidation ability. Herein, the Ea and NO oxidation rates of the three catalysts have no significant difference, which indicates 13



that addition of Pd does not change the pathway of the NO oxidation reaction. Accordingly, perovskites are still the main active sites for NO oxidation.20

Figure 3 Arrhenius plots for NO oxidation over the fresh LSC, LSCP and P/LSC catalysts at 240,000 mL g-1 h-1. Feeding gas: 400 ppm NO/5 vol.% O2/N2.

Figure 4 (a) NOx storage profiles of the fresh catalysts at 300 ºC in lean-burn phase for 90 min; (b) the subsequent NOx-TPD profiles. Lean-burn phase: 400 ppm NO/5 vol.% O2/N2, 400 mL min-1.

The NOx storage tests for the catalysts were carried out at 300 ºC in the lean-burn 14


Page 14 of 43

Page 15 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


gas flow. Figure 4a shows the NOx storage profiles of the catalysts. Interestingly, the addition of trace amount of Pd makes an apparent contribution to the NOx storage capacity (NSC), which is consistent with the previous report.39 The NSC values are 967.3, 1200.6 and 1440.7 μmol g-1 for the LSC, P/LSC and LSCP, respectively. It has been reported that NO2 can be more easily stored than NO on basic materials.34 Therefore, the LSCP catalyst with the highest NO oxidation ability displays the longest “completely trapping duration” (8 min), which is crucial to the NOx trapping performance in the lean-burn period. Figure 4b presents the NOx-TPD profiles of the catalysts after the above storage period. The largest amount of NOx desorbs from the LSCP catalyst, which is consistent with the results of the NOx storage tests. Nevertheless, no obvious difference is observed in the desorption temperature range and the shape of the desorption peaks of the three catalysts. It indicates that the types of the trapped NOx species on the catalysts have little difference. Figure 5 presents the H2-TPR profiles of the fresh LSC, LSCP and P/LSC catalysts. The H2-TPR profile of the LSC catalyst shows two reduction peaks at 316 ºC and 507 ºC, which can be attributed to the reduction of Co3+ to Co2+ and Co2+ to Co0, respectively.20,

40, 41

After the addition of Pd, the obvious shifts to lower

temperatures are observed for the reduction peaks of Co. For example, the reduction peak of Co3+ of the P/LSC catalyst shifts from 316 ºC to 134 ºC, and that of the LSCP catalyst moves from 316 ºC to 198 ºC. It indicates that the addition of Pd enhances the reducibility of the pure LSC perovskite. The shift of the reduction temperature can be explained by the classic hydrogen spillover theory that hydrogen molecules dissociate 15



on the Pd clusters on the surface, derived from the reduction of lattice Pd ions or surface PdO particles, and then spillover to the perovskite.29 The similar temperature shift is also observed for the Co2+ reduction peaks of the LSCP and P/LSC catalysts. The reduction of Pd species of the LSCP and P/LSC catalysts happens at low temperatures. The reduction peak of the supported PdO lies at 90-110 ºC, while the reduction peak of the doped Pd3+ and/or Pd4+ locates at 100-150 ºC, in line with previous findings.27, 30 Clearly, the supported PdO on the surface of the LSC is more easily reduced than the Pd dissolved in the perovskite lattice of the LSCP, which further strengthens the H2 spillover on the perovskite for the P/LSC. It explains why the P/LSC catalyst exhibits a larger temperature shift than the LSCP in the reduction peaks of Co3+ and Co2+ towards lower temperatures, even to the extent that the reduction peak of Co3+ overlaps with that of the surface PdO.28 Whereas, the prominent reducibility of PdO in the P/LSC catalyst does not bring an increase in the NO oxidation ability, according to the Ea values (LSC: 100±8 KJ mol-1, P/LSC: 109 ±5 KJ mol-1) in Figure 3. Therefore, we can conclude that the Pd species is not the main active sites for the NO oxidation.

16


Page 16 of 43

Page 17 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5 H2-TPR profiles of the fresh catalysts. The low reduction temperature of Pd species in the H2-TPR experiments indicates the easy formation of Pd0, which has been proven to play a significant role in the NOx reduction in the fuel-rich phase.25 To study the impact of Pd species on the fuel-rich performance, the NOx reduction behaviors of the catalysts were explored in a lean-burn period (400 ppm NO/5 vol.% O2/N2) at 300 ºC for 5 min for NOx storage and a subsequent long fuel-rich period (1000 ppm C3H6/N2) at 300 ºC for 25 min. The outlet NOx and CO2 concentrations were recorded in the whole storage and reduction process for the LSCP, LSC and P/LSC catalysts. The switch from lean-burn to fuel-rich atmosphere was achieved by solenoid valves without time interval. Figure 6 displays the corresponding outlet concentrations of NOx and CO2 as the function of time. In terms of the NOx profiles of the catalysts (Figure 6a), all of the three catalysts completely capture and store NOx in the lean-burn process (0-5 min), suggesting the amounts of the stored NOx in the three catalysts are the same. The stored NOx reacts with C3H6 after switching to the fuel-rich atmosphere. A part of the stored NOx directly escapes without being reduced as shown in Figure 6a. Obviously, the LSC 17



catalyst displays the largest escaping amount of NOx. Unlike the LSC catalyst, only a little NOx escapes from the two Pd-contained catalysts, exhibiting the improved De-NOx activities. In Figure 6b, a part of carbonate species decomposes and releases CO2 with the formation of nitrite/nitrate species in the lean-burn NOx storage process (0-5 min). After the switch to the fuel-rich atmosphere, CO2, i.e. the product of the reaction between the stored NOx and C3H6, is rapidly produced and released. According to the total amount of the CO2 production under the fuel-rich conditions (5-30 min), the De-NOx activities of the catalysts are identified to follow the order: P/LSC > LSCP > LSC. The difference in the De-NOx activities of the Pd-contained catalysts is probably associated with the different spatial location of Pd species in the catalysts. According to the previous reports, most of reduction and exsolution of B-site ions occurs within the bulk perovskite rather than on its surface,42, 43 whereas the reduced Pd0 within the bulk is nearly inaccessible for catalysis. As a result, it decreases the overall utilization efficiency of the Pd species. To confirm this assumption, the XPS experiments were conducted over the LSCP and P/LSC catalysts pretreated in the 1000 ppm C3H6 flow at 300 ºC for 2 min. Figure 7 shows the XPS profiles of the reduced catalysts, and Table S1 lists the relevant data. In Figure 7, Pd species in the LSCP and P/LSC catalysts are both reduced to metallic Pd after the reducing treatment by C3H6 for 2 min. However, the surface Pd/La ratio of the LSCP catalyst does not change with the reduction, which suggests that the lattice Pd still remains within the bulk of perovskite after the reduction in the fuel-rich periods. On the contrary, the supported Pd is highly 18


Page 18 of 43

Page 19 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


dispersed on the surface of the perovskite and reduced to Pd0, which will readily contact with the reactants in the fuel-rich atmosphere. Compared with the LSCP, there are more superficial Pd species on the P/LSC catalyst, as listed in Table S1. The apparent difference in the De-NOx performance of these catalysts suggests that Pd is the main active sites for NOx reduction in the fuel-rich phase. It is worth noting that the whole NOx reduction process lasts more than 10 min at 300 ºC, much longer than the storage duration (5 min). The shorter reduction period can not completely regenerate all the NOx storage sites under the real operation conditions. It is one of the major reasons for NOx accumulation on the catalysts in the successive LNT process, where a fuel-rich period is much shorter than a lean-burn period (50 s lean-burn/10 s fuel-rich in this case). As a result, only the most active NOx storage sites can be recycled in the LNT process.

19



Figure 6 (a) NOx concentrations and (b) CO2 concentrations in a lean-burn period (400 ppm NO/5 vol.% O2/N2, 400 mL min-1) for 5 min and a fuel-rich (1000 ppm C3H6/N2, 400 mL min-1) for 25 min at 300 ºC; (c) the subsequent NOx-TPD profiles.

20


Page 20 of 43

Page 21 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 7 XPS spectra of Pd 3d of the LSCP and P/LSC catalysts after the reduction in the 1000 ppm C3H6/N2 flow at 300 ºC for 2 min.

Another interesting phenomenon is observed for the two Pd-contained catalysts: the NOx escaping period lasted for only 10 min in Figure 6a, while CO2 is continuously released in the followed periods and then gradually drops to zero after 15 min in Figure 6b. The difference in the NOx and CO2 profiles indicates that there are more than one kind of nitrate/nitrite species formed in the catalysts. Here, for the simplicity of the discussion, the stored NOx is roughly divided into the unstable nitrate/nitrite species and the stable nitrate/nitrite species. The unstable nitrate/nitrite species formed on the weak NOx storage sites are preferentially regenerated after switching to the fuel-rich atmosphere. Nevertheless, the high decomposition rate over these sites makes the desorbed NOx hardly be completely reduced. Therefore, the NOx signal is observed at the beginning of the fuel-rich period in Figure 6a. When the unstable NOx storage sites have been regenerated, the reaction moves onto the strong NOx storage sites, which corresponds to the decomposition of the stable nitrate/nitrite 21



species. Then, the NOx signals detected in the exhaust gradually drop to zero as a result of the declining decomposition rate. Taking into account that the stored NOx might not be completely desorbed and reduced in the tests, the TPD experiments were conducted afterwards, with a starting temperature of 300 ºC (Figure 6c). The results show that only the P/LSC is able to completely remove the stored NOx from the catalyst at 300 ºC with assistance of C3H6. There is still a certain amount of the NOx species remaining on the LSC and LSCP catalysts. Obviously, compared with the LSCP, the P/LSC achieves a more complete regeneration of the NOx storage sites in the fuel-rich atmosphere.

3.3 Catalytic activity of the catalysts. Figure 8 shows the NOx removal percentage of the catalysts in the successively lean-burn/fuel-rich alternations in the temperature range of 250-400 ºC. For the two Pd-contained catalysts, a more remarkable increment in the De-NOx activity is observed in the whole temperature range, compared with the LSC, and the P/LSC has the highest De-NOx activity among the three catalysts. Additionally, the highest NRP of the LSCP and P/LSC catalysts are achieved at 300 ºC, i.e. 74.4 % and 90.4 %, respectively. The N2O profile and the LNT profiles of the catalysts at 300 ºC are presented in Figure S3. The N2O selectivity is around 2 %, in line with the previous reports. 3, 44 Their NRP begins to drop when the reaction temperature further increases, which can be ascribed to the limited thermal stability of nitrates/nitrites at high temperatures. It is also worth noting that a dramatic decline in the De-NOx activity is observed for all the catalysts when the reaction temperature decreases to 250 ºC, at 22


Page 22 of 43

Page 23 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


which the decomposition of nitrites/nitrates happens with a low rate in the fuel-rich phase and therefore limits the recycling of the NOx storage sites.

Figure 8 NOx removal percentage of the catalysts at different temperatures.

Table 2 The catalytic activity of the catalysts in the 21th LNT cycle. Catalyst

NOx storage efficiency (%)

NOx reducing efficiency (%)

NRP (%)

N2O selectivity (%)

LSC

63.4

81.5

51.6

1.8

LSCP

88.4

84.2

74.4

2.1

P/LSC

98.4

91.9

90.4

1.9

To further elucidate the role of the Pd species in the LNT process, Figure 9 shows the evolution curves of the outlet NOx concentration during the steady 21th lean-rich cycle at the optimum temperature (300 ºC), and Table 2 lists the relevant results. In Figure 9, after switching from the lean-burn to fuel-rich atmosphere, fewer amounts of NOx escape from the P/LSC catalyst than that from the other two catalysts. Among the three catalyst, the P/LSC catalyst exhibits the highest NOx reducing 23



efficiency in the fuel-rich period and NRP in the alternative operations, although the LSCP catalyst possesses the higher NO oxidation ability (Figure 3) and NOx storage capacity (Figure 4) than the P/LSC. It can be deduced that the De-NOx activity of the perovskite system in the LNT process is restricted by the inadequate NOx reduction ability in the fuel-rich periods.

Figure 9 The 21th cycle of the LNT process at 300 ºC. Lean-burn phase: 50 s, 400 ppm NO/5 vol.% O2/N2; fuel-rich phase: 10 s, 1000 ppm C3H6/N2 (space velocity: 120,000 mL g-1 h-1).

Furthermore, an interesting phenomenon should be noted that the addition of Pd also lead to an obvious distinction in the lean-burn performance of the three catalysts. NOx species can be completely captured and stored on the P/LSC catalyst in the lean-burn period. However, there are 3.5 μmol g-1 and 11 μmol g-1 NOx (calculated from Table 2) escaping from the LSCP and LSC catalysts in a single lean-burn period, respectively. When switching back to the lean-burn period in the next cycle, the P/LSC catalyst shows a high efficiency for recapturing NOx in the lean-burn 24


Page 24 of 43

Page 25 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


atmosphere. Among the three catalysts, the LSC exhibits the worst NOx trapping performance in the LNT process.

Figure 10 Effect of water vapor and carbon dioxide on the De-NOx activities of the catalysts in the LNT process at 300 ºC. Lean-burn phase: 50 s, 400 ppm NO/5 vol.% O2/N2, fuel-rich phase: 10 s, 1000 ppm C3H6/N2; the space velocity is 120,000 mL g-1 h-1; the wet feed includes 4 vol.% H2O, 5 vol.% CO2 in both lean-burn and fuel-rich phase.

Figure 10 presents the De-NOx activity of the catalysts in the LNT process with or without H2O and CO2 in feeding gas. Figure S4 gives the LNT profiles of the catalysts at 300 ºC with H2O and CO2 in feeding gas. The NOx conversion drops for all the catalysts after the addition of H2O and CO2. The decline of NRP is within reasonable range based on the previous literature report.45 The Pd-contained catalysts exhibit much better resistance to water vapor and CO2 than the LSC catalyst, especially for the P/LSC catalyst with the NRP of about 60 %.

25



Page 26 of 43

3.4 Influence of Pd species on NOx trapping performances. Beside the contributions to the NOx reduction behaviors, the addition of Pd also leads to an improvement of the NOx trapping performance in the lean-burn periods in Figure 9. The previous discuss of the function of Pd in the lean-burn atmosphere alone, as conducted in Figures 3 and 4, is insufficient to explain the disparity in the NOx trapping ability of the catalysts in the LNT process in Figure 9. To further illustrate the influence of the Pd species on the NOx trapping performance, the isothermal NOx adsorption tests were carried out over the spent catalysts after the LNT tests. Figure 11 displays the NOx profiles of the catalysts recorded during NOx trapping, and Table 3 gives the relevant data. In Figure 11, compared with the fresh catalysts, the NOx storage amounts of all the spent catalysts apparently decrease, because of the accumulation of the trapped NOx in the catalysts during the NOx storage and reduction cycles. The irreversible NOx storage sites on the catalysts that can not regenerate themselves in the LNT process are inactive in the lean-burn operation. The reversible NOx storage sites that can realize the trapping/releasing behaviors in the LNT process determine the NOx trapping efficiency in the lean-burn atmospheres.

Table 3 The NSCs of the fresh and spent catalysts and the ratios of the regenerated NOx storage sites. Catalyst

NSC

26


Ratio of the regenerated NOx

Page 27 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


-1

-1

storage sites a

Fresh (μmol g )

Spent (μm g )

LSC

963.6

354.5

0.37

LSCP

1446.7

737.5

0.51

P/LSC

1200.6

1001.9

0.83

a

the regenerated ratio =NSCspent/ NSCfresh The NOx-TPD experiments were also conducted over the spent catalysts to

explore the thermal stability of the NOx species trapped on the irreversible trapping sites after the NOx storage and reduction cycles. Figure 12a shows the NOx desorption profiles over the spent catalysts, and Table S2 provides the amounts of the desorbed NOx from the catalysts. In Figure 12a, the area of the desorption peaks of the spent catalysts follows the order of LSCP > LSC > P/LSC. The amounts of the desorbed NOx in Table S2 are close to the lost NSC of the spent catalysts in Table 3. The spent P/LSC catalyst only desorbs a small amount of the residual NOx, suggesting most of the stored NOx is removed from the storage sites during the fuel-rich periods. Although the doped Pd in the LSCP brings an increase in the NOx storage capacity in Figure 4a, unfortunately, the extra NOx storage sites resulting from the Pd doping can not be effectively regenerated in the followed fuel-rich periods.

27



Figure 11 NOx storage profiles of the spent catalysts in the lean-burn phase (400 ppm NO/5 vol.% O2/N2, 400 mL min-1) at 300 ºC.

Figure 12a also shows the initial desorption temperatures as an auxiliary illustration, which reflects the thermal stability of the residual NOx. Compared with the initial desorption temperatures of the fresh catalysts in Figure 4b (275 ºC), those of the spent catalysts shift to higher temperatures, demonstrating the remained NOx species in the spent catalysts is trapped by the strong irreversible NOx storage sites that can not be regenerated in the fuel-rich phase. The spent P/LSC catalyst shows the higher initial desorption temperature than the spent LSCP, suggesting that the NOx species remained in the former is more stable than that in the latter one.

28


Page 28 of 43

Page 29 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 12 (a) NOx-TPD profiles of the spent catalysts; (b) NOx-TPD profiles for the fresh P/LSC catalyst after trapping NOx in the lean-burn phase for 3 min, 5 min and 13 min. Lean-burn phase: 400 ppm NO/5 vol.% O2/N2, 400 mL min-1.

The results of Figure 4b show that the types of the trapped NOx species on the three fresh catalysts have little difference. Therefore, the difference in the amounts of the reversible and irreversible trapping sites of the catalysts is determined by the regeneration ability of the NOx storage sites in the fuel-rich atmosphere. It has been reported that the addition of Pt can improve the NOx desorption behaviors, which majorly happens at the interface of the NOx storage sites and metallic Pt reduced in the fuel-rich atmosphere.46,

47

Similarly to Pt, our results reveal that Pd can also

accelerate the NOx desorption behavior in the short fuel-rich period, which has seldom been reported before. Therefore, Table 3 provides the ratio of the regenerated NOx storage sites of the catalysts (P/LSC: 0.83, LSCP: 0.51, LSC: 0.37), as a significant parameter to reveal the function of Pd on the NOx desorption performance. Based on the results, the addition of Pd restrains the NOx accumulation and benefits and the 29



regeneration of the NOx storage sites during the NOx storage and reduction cycles. Compared with the doped Pd, the supported Pd on the surface more closely contacts with the C3H6 and the NOx storage sites, and gets reduced more easily according to the H2-TPR results in Figure 5. It makes a larger contribution to the NOx desorption behaviors in the fuel-rich period. Thus, the fewer amounts of NOx remains in the spent P/LSC, and the remained NOx species is also more stable than that in the spent LSCP catalyst. Accordingly, more reversible NOx storage sites on the P/LSC take part in the NOx trapping process than those on the LSCP, which is the reason for the outstanding lean-burn performance of the P/LSC catalyst in the LNT cycles. In turn, the effective regeneration of the NOx storage sites will promote the De-NOx activity of the P/LSC catalyst in the whole LNT process. Compared with the Pd-contained catalysts, only a smaller amount of the reversible NOx storage sites on the LSC catalyst works in the LNT process, indicating the limited regeneration ability of the NOx storage sites without the addition of Pd. The doping of Pd enhances the NO oxidation ability and NOx storage capacity of the LSCP catalyst in Figures 3 and 4. However, compared with the LSCP, the P/LSC catalyst with the lower NO oxidation ability and NOx storage capacity exhibits the better NOx trapping performance in the LNT process, because of the effective regeneration of the NOx storage sites in the fuel-rich periods. Our previous studies show that the LSC catalyst possesses the sufficient NO oxidation ability and NOx storage capacity in lean-burn atmospheres, which are even higher than the Pt-based catalyst.20, 25 Therefore, in this case the NO oxidation ability and NOx storage capacity 30


Page 30 of 43

Page 31 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


of the catalysts are not the key factors of the LNT process. Combining the results in Figures 9 and 11, we can conclude that the De-NOx activity of the catalysts is limited by the regeneration rate of the NOx storage sites and the followed NOx reduction ability in the fuel-rich period. The results of Figures 11 and 12a give a reasonable explanation for the difference in the NOx trapping performance in Figure 9. Among the three spent catalysts, only the spent P/LSC catalyst can still exhibit a “completely trapping duration” for about 4 min at the initial NOx trapping process in Figure 11. Thus, only the P/LSC catalyst completely traps NOx in the lean-burn period in Figure 9. Compared with the LSCP and LSC catalysts, the P/LSC catalyst, benefiting from its higher regeneration ability of the NOx storage sites, can regenerate a part of the strong NOx storage sites and make them reversible. The regeneration of these sites results in the fast NOx recapturing rate of the P/LSC catalyst at the beginning of the lean-burn period in Figure 9. Based on above analysis, it can be concluded that the supported Pd species plays a more significant role in improving the lean-burn activity in the LNT process than the doped Pd species.

There are three points that can be confusing in the results of the above experiments. (1) Although there are sufficient unoccupied NOx storage sites, a part of NOx still escapes from the LSC and LSCP catalysts in the lean-burn period of the LNT process in Figure 9. It is because the remained NOx storage sites can not be fast enough to completely capture the inlet NOx, since the strong NOx storage sites have already been occupied. (2) The NOx storage sites of the P/LSC catalyst can not be 31



completely regenerated in the fuel-rich phase of the LNT process at 300 ºC, as indicated by Figure 12a. It does not conflict with the prior results in Figure 6c that the NOx stored on the P/LSC catalyst is totally desorbed and effectively reduced at 300 ºC. In Figure 6, the fuel-rich period in the De-NOx tests is 25 min, but the duration of the fuel-rich phase in each LNT cycle is only 10 s in Figure 9. Such a short reduction period makes the strong NOx storage sites hardly be completely regenerated. (3) The LNT activities of the LSC and LSCP catalysts are limited by the insufficient NO x desorption behaviors at 300 ºC. The thermal stability of the nitrates will decrease with the increase of temperature, leading to higher regeneration efficiency at higher temperatures. However, the catalytic activities of the LSC and LSCP do not increase when the temperature rises in Figure 8. It should be noted that LNT is a complex process that is influenced by more than one factor. In this case, the LSC and LSCP catalyst do not have the excellent NOx reducing efficiency as the P/LSC. The quicker NOx desorption rate brought by further increasing the reaction temperature actually makes it harder for the LSC and LSCP catalysts to reduce the desorbed NOx. There will be more amount of NOx directly escaping from the catalyst after desorption without being reduced at higher temperatures than at 300 ºC. This explains why the catalysts exhibit the highest LNT activity at 300 ºC, even though not all of the NOx storage sites are effectively regenerated at this temperature. In order to further explore the NOx trapping order in the lean-burn atmosphere, the NOx-TPD experiments after the NOx storage for 3, 5 and 13 min were carried out over the fresh P/LSC catalyst in Figure 12b, and Table S3 lists the quantity of the 32


Page 32 of 43

Page 33 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


stored NOx. In Figure 12b, only a small amount of NOx desorbed after NOx storage for 3 min, in the temperature range of 381 to 529 ºC. When the NOx storage period is prolonged to 5 min, the amount of desorbed NOx slightly increases and the initial desorption temperature shifts from 381 ºC to 338 ºC. The initial desorption temperature further moves to 299 ºC after the NOx storage prolonged to 13 min. However, no apparent shift is observed in the ending temperatures of the desorption peaks. Based on the above facts, it can be deduced that the NOx species gives priority to adsorb on the strong absorption sites, which desorbs at the relatively higher temperatures, that is, a high thermal stability. After the strong absorption sites are occupied, the relatively weaker absorption sites have more chance of trapping NOx. It is interesting to find that in Figure 12b the NOx-TPD profile of the fresh P/LSC catalyst after the NOx storage for 5 min is similar with that of the spent P/LSC catalyst in Figure 12a. The NOx desorption peaks of these two profiles locates in the exactly same temperature range and the peak shapes are highly similar. There is even little difference in the amount of the desorbed NOx. It suggests that the “irreversible NOx storage sites” in the LNT process are exactly the strong NOx storage sites that are preferentially occupied in a lean-burn period. These storage sites are readily occupied in the lean-burn phase of the LNT process, but they are too stable to be regenerated in the fuel-rich phase. Once trapping NOx in the lean-burn atmosphere and releasing/reducing NOx in the fuel-rich atmosphere reach the balance, the LNT process gets the steady state.

33



4. Conclusion Two Pd-contained catalysts (LSCP and P/LSC) were investigated to reveal the influence of the different Pd species on the De-NOx activity of the LSC based perovskite in the LNT process. Pd is proven to be introduced into the B-site of the crystal lattices of the LSC perovskite in the LSCP catalyst and onto the surface of the perovskite in the P/LSC catalyst, through sol-gel and wet impregnation methods, respectively. The results of the cyclic NOx storage and reduction operations display a significant improvement of the De-NOx activity after the addition of Pd, compared with the LSC. Among the three catalysts, the highest NRP belongs to the P/LSC catalyst (90.4 %). The LSCP catalyst performs the higher NO oxidation ability and NOx storage capacity than the P/LSC catalyst in the lean-burn atmosphere. However, it does not show a qualified De-NOx activity in the whole LNT process. It is because the NO oxidation ability and NOx storage capacity of the Pd-free LSC is already high enough. Thus, these two factors do not obviously influence the De-NOx activity of the catalysts. The high De-NOx activity of the Pd contained catalysts in the LNT process mainly derives from two aspects, i.e. the enhanced NOx desorption and NOx reduction abilities of the catalysts in the fuel-rich periods after adding Pd. The effective desorption of the stored NOx benefits the NOx trapping performance and avoids the direct escape of NOx in the followed lean-burn period. The improved NOx reduction ability of the catalysts after the addition of Pd makes the desorbed NOx be effectively reduced, leading to an excellent De-NOx activity in the LNT process. According to the 34


Page 34 of 43

Page 35 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


H2-TPR and XPS results, the supported Pd species on the surface of the LSC perovskite easily contacts with the NOx storage sites and reducing agents in the fuel-rich atmosphere; while the doped Pd species remains in the bulk of the perovskite, which is inaccessible for catalysis. Compared with the LSCP catalyst, the P/LSC catalyst exhibits a higher overall utilization efficiency of the Pd species. Thus, the P/LSC catalyst achieves the more effective regeneration of the NOx storage sites and higher NOx reducing efficiency, which are proven to be the key factors that determine the De-NOx activity over the catalysts. All of these findings demonstrate the enhancement effect of Pd on the catalytic performances in both the lean-burn and fuel-rich atmospheres, and provide new insights into the roles of the Pd species in the LNT process.

Associated Content Supporting Information Comparison of Co K-edge XANES (Figure S1); XPS profiles of surface Pd species (Figure S2); NSR profiles with or without H2O and CO2 (Figures S3-S4).

Acknowledgement This work was supported by the National Natural Science Foundation of China (U1232118, 21476159), the 973 program (2014CB932403), and the Natural Science Foundation of Tianjin, PR China (15JCZDJC37400, 15JCYBJC23000). Authors are also grateful to the Program of Introducing Talents of Disciplines to China Universities (B06006). 35



References (1) Bai, Z. F.; Zhang, Z. S.; Chen, B. B.; Zhao, Q.; Crocker, M.; Shi, C. Non-Thermal Plasma Enhanced NSR Performance over Pt/M/Ba/Al2O3 (M = Mn, Co, Cu) Catalysts. Chem. Eng. J. 2017, 314, 688-699. (2) Parks, J. E. Less Costly Catalysts for Controlling Engine Emissions. Science 2010, 327, 1584-1585. (3) Abdulhamid, H.; Fridell, E.; Skoglundh, M. The Reduction Phase in NOx Storage Catalysis: Effect of Type of Precious Metal and Reducing Agent. Appl. Catal. B: Environ. 2006, 62, 319-328. (4) Chen, X.; Schwank, J.; Li, J.; Schneider, W. F.; Goralski, C. T.; Schmitz, P. J. Thermal Decomposition of Dispersed and Bulk-Like NOx Species in Model NOx Trap Materials. Appl. Catal. B: Environ. 2005, 61, 164-175. (5) Lietti, L.; Nova, I.; Forzatti, P. Role of Ammonia in the Reduction by Hydrogen of NOx Stored over Pt-Ba/Al2O3 Lean NOx trap Catalysts. J. Catal. 2008, 257, 270-282. (6) Nova, I.; Castoldi, L.; Lietti, L.; Tronconi, E.; Forzatti, P.; Prinetto,; Ghiotti, F. G. NOx Adsorption Study over Pt-Ba/Alumina Catalysts: FT-IR and Pulse Experiments. J. Catal. 2004, 222, 377-388. (7) Amberntsson, A.; Fridell, E.; Skoglundh, M. Influence of Platinum and Rhodium Composition on the NOx Storage and Sulphur Tolerance of a Barium Based NOx Storage Catalyst. Appl. Catal. B: Environ. 2003, 46, 429-439. (8) Epling, W. S.; Campbell, L. E.; Yezerets, A.; Currier, N. W.; Parks, J. E. 36


Page 36 of 43

Page 37 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Overview of the Fundamental Reactions and Degradation Mechanisms of NOx Storage/Reduction Catalysts. Catal. Rev. 2004, 46, 163-245. (9) Mráček, D.; Kočí, P.; Choi, J. S.; Partridge, W. P. New Operation Strategy for Driving the Selectivity of NOx Reduction to N2, NH3 or N2O during Lean/Rich Cycling of a Lean NOx Trap Catalyst. Appl. Catal. B: Environ. 2016, 182, 109-114. (10) López-Suárez, F. E.; Illán-Gómez, M. J.; Bueno-López, A.; Anderson, J. A. NOx Storage and Reduction on a SrTiCuO3 Perovskite Catalyst Studied by Operando DRIFTS. Appl. Catal. B: Environ. 2011, 104, 261-267. (11) Roy, S.; Baiker, A. NOx Storage-Reduction Catalysis: from Mechanism and Materials Properties to Storage-Reduction Performance. Chem. Rev. 2009, 109, 4054-4091. (12) Shi, C.; Zhang, Z. S.; Crocker, M.; Xu, L.; Wang, C. Y.; Au, C.; Zhu, A. M. Non-Thermal Plasma-Assisted NOx Storage and Reduction on a LaMn0.9Fe0.1O3 Perovskite Catalyst. Catal. Today 2013, 211, 96-103. (13) Onrubia, J. A.; Pereda-Ayo, B.; De-La-Torre, U.; González-Velasco, J. R. Key Factors in Sr-Doped LaBO3 (B= Co or Mn) Perovskites for NO Oxidation in Efficient Diesel Exhaust Purification. Appl. Catal. B: Environ. 2017, 213, 198-210. (14) He, X.; Meng, M.; He, J.; Zou, Z.; Li, X.; Li, Z.; Jiang, Z. A Potential Substitution of Noble Metal Pt by Perovskite LaCoO3 in ZrTiO4 Supported Lean-Burn NOx Trap Catalysts. Catal. Commun. 2010, 12, 165-168. (15) Yoon, D. Y.; Lim, E.; Kim, Y. J.; Kim, J. H.; Ryu, T.; Lee, S.; Cho, B. K.; Nam, I. S.; Choung, J. W.; Yoo, S. NO Oxidation Activity of Ag-doped Perovskite Catalysts. 37



J. Catal. 2014, 319, 182-193. (16) Chen, J.; Shen, M.; Wang, X.; Wang, J.; Su, Y.; Zhao, Z. Catalytic Performance of NO Oxidation over LaMeO3 (Me= Mn, Fe, Co) Perovskite Prepared by the Sol-Gel Method. Catal. Commun. 2013, 37, 105-108. (17) Li, X. Z.; Yan, X. Y.; Zuo, S. X.; Lu, X. W.; Luo, S. P.;. Li, Z. Y.; Yao, C.; Ni, C. Y. Construction of LaFe1−xMnxO3/Attapulgite Nanocomposite for Photo-SCR of NOx at Low Temperature. Chem. Eng. J. 2017, 320, 211-221. (18) Li, Z. Q.; Meng, M.; Li, Q.; Xie, Y. N.; Hu, T. D.; Zhang, J. Fe-Substituted Nanometric La0.9K0.1Co1−xFexO3-δ Perovskite Catalysts Used for Soot Combustion, NOx Storage and Simultaneous Catalytic Removal of Soot and NOx. Chem. Eng. J. 2010, 164, 98-105. (19) Yang, W.; Zhang, R. D.; Chen, B. H.; Bion, N.; Duprez, D.; Royer, S. Activity of Perovskite-Type Mixed Oxides for the Low-Temperature CO Oxidation: Evidence of Oxygen Species Participation from the Solid. J. Catal. 2012, 295, 45-58. (20) Li, X. G.; Dong, Y. H.; Xian, H.; Hernández, W. Y.; Meng, M.; Zou, H. H.; Ma, A. J.; Zhang, T. Y.; Jiang, Z.; Tsubaki, N.; Vernoux, P. De-NOx in Alternative Lean/Rich Atmospheres on La1-xSrxCoO3 Perovskites. Energy Environ. Sci. 2011, 4, 3351-3354. (21) Wen, W.; Wang, X. Y.; Jin, S.; Wang, R. H. LaCoO3 Perovskite in Pt/LaCoO3/K/Al2O3 for the Improvement of NOx Storage and Reduction Performances. RSC Adv. 2016, 6, 74046-74052. (22) Yang, W.; Zhang, R. D.; Chen, B. H.; Duprez, D.; Royer, S. New Aspects on the 38


Page 38 of 43

Page 39 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Mechanism of C3H6 Selective Catalytic Reduction of NO in the Presence of O2 over LaFe1–x(Cu, Pd)xO3−δ Perovskites. Environ. Sci. Technol. 2012, 46, 11280-11288. (23) Sartipi, S.; Khodadadi, A. A.; Mortazavi, Y. Pd-doped LaCoO3 Regenerative Catalyst for Automotive Emissions Control. Appl. Catal. B: Environ. 2008, 83, 214-220. (24) Nashihata, Y.; Mizuki, J.; Akao, T.; Tanaka, H.; Uenishi, M.; Kimura, M.; Okamoto, T.; Hamada, N. Self-Regeneration of a Pd-Perovskite Catalyst for Automotive Emissions Control. Nature 2002, 418, 164-167. (25) Li, X. G.; Chen, C.; Liu, C.; Xian, H.; Guo, L.; Lv, J.; Jiang, Z.; Vernoux, P. Pd-Doped Perovskite: an Effective Catalyst for Removal of NOx from Lean-Burn Exhausts with High Sulfur Resistance. ACS Catal. 2013, 3, 1071-1075. (26) Dacquin, J.; Dujardin, C.; Granger, P. Surface Reconstruction of Supported Pd on LaCoO3: Consequences on the Catalytic Properties in the Decomposition of N2O. J. Catal. 2008, 253, 37-49. (27) Guo, X.; Meng, M.; Dai, F.; Li, Q.; Zhang, Z.; Jiang, Z.; Zhang, S.; Huang, Y. NOx-Assisted Soot Combustion over Dually Substituted Perovskite Catalysts La1-xKxCo1-yPdyO3-δ. Appl. Catal. B: Environ. 2013, 142, 278-289. (28) Ziaei-Azad, H.; Khodadadi, A.; Esmaeilnejad-Ahranjani, P.; Mortazavi, Y. Effects of Pd on Enhancement of Oxidation Activity of LaBO3 (B= Mn, Fe, Co and Ni) Pervoskite Catalysts for Pollution Abatement from Natural Gas Fueled Vehicles. Appl. Catal. B: Environ. 2011, 102, 62-70. (29) Eyssler, A.; Winkler, A.; Safonova, O.; Nachtegaal, M.; Matam, S. K.; Hug, P.; 39



Weidenkaff, A.; Ferri, D. On the State of Pd in Perovskite-Type Oxidation Catalysts of Composition A(B, Pd)O3±δ (A= La, Y; B= Mn, Fe, Co). Chem. Mater. 2012, 24, 1864-1875. (30) Zhou, K. B.; Chen, H. D.; Tian, Q.; Hao, Z. P.; Shen, D. X.; Xu, X. B.; Mol. J. Pd-Containing Perovskite-Type Oxides Used for Three-Way Catalysts. Catal. A: Chem. 2002, 189, 225-232. (31) You, R.; Zhang, Y.; Liu, D.; Meng, M.; Jiang, Z.; Zhang, S.; Huang, Y. A Series of Ceria Supported Lean-Burn NOx Trap Catalysts LaCoO3/K2CO3/CeO2 Using Perovskite as Active Component. Chem. Eng. J. 2015, 260, 357-367. (32) Zhu, X. B.; Tu, X.; Chen, M. H.; Yang, Y.; Zheng, C. H.; Zhou, J. S.; Gao, X. La0.8M0.2MnO3 (M= Ba, Ca, Ce, Mg and Sr) Perovskite Catalysts for Plasma-Catalytic Oxidation of Ethyl Acetate. Catal. Commun. 2017, 92, 35-39. (33) Ma, A. J.; Wang, S. Z.; Liu, C.; Xian, H.; Ding, Q.; Guo, L.; Meng, M.; Tan, Y. S.; Tsubaki, N.; Zhang, J.; Zheng, L. R.; Li, X. G. Effects of Fe Dopants and Residual Carbonates on the Catalytic Activities of the Perovskite-Type La0.7Sr0.3Co1-xFexO3 NOx Storage Catalyst. Appl. Catal. B: Environ. 2014, 146, 24-34. (34) Constantinou, C.; Li, W.; Qi, G.; Epling, W. S. NOx Storage and Reduction over a Perovskite-Based Lean NOx Trap Catalyst. Appl. Catal. B: Environ. 2013, 134-135, 66-74. (35) Petrović, S.; Karanović, L.; Stefanov, P. K.; Zdujićd, M.; Terlecki-Baričevića, A. Catalytic Combustion of Methane Over Pd Containing Perovskite Type Oxides. Appl. Catal. B: Environ. 2005, 58, 133-141. 40


Page 40 of 43

Page 41 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(36) Moulder, J. F.; Stickle, W. F.; Sobol, P. E.; Banben, K. D. Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer: Minnesota, 1992. (37) Miquel, P.; Yamin, Y.; Lombaert, K.; Dujardin, C.; Trentesaux, M.; Gengembre, L.; Granger, P. Pd Characterization by XPS in Perovskite Catalysts for NOx Reduction: Influence of Thermal Aging. Surf. Interface Anal. 2010, 42, 545-550. (38) Ouyang, X.; Scott, S. L. Mechanism for CO Oxidation Catalyzed by Pd-Substituted BaCeO3, and the Local Structure of the Active Sites. J. Catal. 2010, 273, 83-91. (39) Say, Z.; Dogac, M.; Vovk, E. I.; Kalay, Y. E.; Kim, C. H.; Li, W.; Ozensoy, E. Palladium Doped Perovskite-Based NO Oxidation Catalysts: the Role of Pd and B-Sites for NOx Adsorption Behavior via In-Situ Spectroscopy. Appl. Catal. B: Environ. 2014, 154-155, 51-61. (40) Engelmann-Pirez, M.; Granger, P.; Leclercq, G. Investigation of the Catalytic Performances of Supported Noble Metal Based Catalysts in the NO+H2 Reaction under Lean Conditions. Catal. Today 2005, 107, 315-322. (41) Xu, J.; Liu, J.; Zhao, Z.; Xu, C.; Zheng, J.; Duan, A.; Jiang, G. Easy Synthesis of Three-Dimensionally Ordered Macroporous La1-xKxCoO3 Catalysts and Their High Activities for the Catalytic Combustion of Soot. J. Catal. 2011, 282, 1-12. (42) Neagu, D.; Tsekouras, G.; Miller, D. N.; Menard, H.; Irvine, J. T. In Situ Growth of Nanoparticles through Control of Non-Stoichiometry. Nat. Chem. 2013, 5, 916-923. (43) Katz, M. B.; Zhang, S.; Duan, Y.; Wang, H.; Fang, M.; Zhang, K.; Li, B.; 41



Graham, G. W.; Pan, X. Reversible Precipitation/Dissolution of Precious-Metal Clusters in Perovskite-Based Catalyst Materials: Bulk versus Surface Re-Dispersion. J. Catal. 2012, 293, 145-148. (44) Zhang, Y. X.; Liu, D. S.; Meng, M.; Jiang, Z.; Zhang, S. A Highly Active and Stable Non-Platinic Lean NOx trap Catalyst MnOx-K2CO3/K2Ti8O17 with Ultra-Low NOx to N2O Selectivity. Ind. Eng. Chem. Res. 2014, 53, 8416-8425. (45) Dupré, J.; Bazin, P.; Marie, O.; Daturi, M.; Jeandel, X.; Meunier, F. Effects of Temperature and Rich-Phase Composition on the Performance of a Commercial NOx-Storage-Reduction Material. Appl. Catal. B: Environ. 2016, 181, 534-541. (46) Castoldi, L.; Righini, L.; Matarrese, R.; Lietti, L.; Forzatti, P. Mechanistic Aspects of the Release and the Reduction of NOx Stored on Pt-Ba/Al2O3. J. Catal. 2015, 328, 270-279. (47) Nova, I.; Lietti, L.; Castoldi, L.; Tronconi, E.; Forzatti, P. New Insights in the NOx Reduction Mechanism with H2 over Pt–Ba/γ-Al2O3 Lean NOx Trap Catalysts under Near-Isothermal Conditions. J. Catal. 2006, 239, 244-254.

42


Page 42 of 43

Page 43 of 43 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For Table of Contents Only

43


Addition of Pd on La0.7Sr0.3CoO3 Perovskite To Enhance Catalytic

Recommend Documents