Advances in Designing Efficient La-Based Perovskites for the NO_x Storage and Reduction Process

Abstract

To overcome the inherent challenge of NO_x reduction in the net oxidizing environment of diesel engine exhaust, the NO_x storage and reduction (NSR) concept was proposed in 1995, soon developed and commercialized as a promising DeNO_x technique over the past two decades. Years of practice suggest that it is a tailor-made technique for light-duty diesel vehicles, with the advantage of being space saving, cost effective, and efficient in NO_x abatement; however, the over-reliance of NSR catalysts on high loadings of Pt has always been the bottleneck for its wide application. There remains fervent interest in searching for efficient, economical, and durable alternatives. To date, La-based perovskites are the most explored promising candidate, showing prominent structural and thermal stability and redox property. The perovskite-type oxide structure enables the coupling of redox and storage centers with homogeneous distribution, which maximizes the contact area for NO_x spillover and contributes to efficient NO_x storage and reduction. Moreover, the wide range of possible cationic substitutions in perovskite generates great flexibility, yielding various formulations with interesting features desirable for the NSR process. Herein, this review provides an overview of the features and performances of La-based perovskite in NO oxidation, NO_x storage, and NO_x reduction, and in this way comprehensively evaluates its potential to substitute Pt and further improve the DeNO_x efficiency of the current NSR catalyst. The fundamental structure–property relationships are summarized and highlighted to instruct rational catalyst design. The critical research needs and essential aspects in catalyst design, including poisoner resistance and catalyst sustainability, are finally addressed to inspire the future development of perovskite material for practical application.

1. Introduction

The diesel engines operated under a high air-to-fuel (A/F) ratio in the range of 20 to 65 achieve higher fuel economy and better driving performance than the stoichiometric gasoline engines [1,2,3], but the higher combustion temperature and pressure in the diesel cylinder lead to increased NO_x emissions, and the net oxidizing environment of the exhaust brings huge challenges for NO_x reduction [2]. The conventional three-way catalysis (TWC) system developed for the removal of NO_x in gasoline engine exhaust is not as efficient as the A/F ratio shifts towards the lean-burn region. Thereafter, three alternative DeNO_x techniques were developed, that is, the direct decomposition of NO, selective catalytic reduction (SCR), and NO_x storage and reduction (NSR) or lean NO_x trap (LNT). NO decomposition in the absence of reductant was proposed as an ideal DeNO_x concept with thermodynamic feasibility, but was demonstrated to be kinetically sluggish, showing insufficient NO_x conversions that do not satisfy the requirements of real automotive emission control systems [4,5]. However, the other two techniques have been fully developed and commercialized over the last several decades.

An SCR technique with ammonia from urea decomposition as the reductant was proposed in 1979, which was demonstrated to be efficient and selective for NO_x reduction to N₂ in net oxidizing atmosphere, especially in the temperature range of 250–450 °C [6,7,8,9,10]. However, the complex hardware system (including heated tanks and tubes, injector nozzle, and mixing device, to ensure storage and complete decomposition of urea, homogeneous flow distribution, and accurate dosing of NH₃) means that NH₃-SCR is not a suitable NO_x abatement technology for vehicles with space constraints and cost considerations [11]. The intrinsic risk of ammonium salt [(NH₄)₂SO₄] deposition and NH₃ slip also drove the exploration of novel NO_x removal techniques. Several research organizations then focused on the selective reduction of NO_x with hydrocarbons as the reductant in the 1990s, but this technique posed a couple of problems, including inadequate NO_x conversion and a narrow temperature window [12,13]. In this context, the NSR technique was proposed by Toyota in 1995 and is characteristic of incorporating basic components in DeNO_x catalyst formulation for NO_x storage, adopting alternating lean-burn/fuel-rich operations, and not requiring external reductants [14]. In the lean-burn periods of a typical NSR process, NO is oxidized to NO₂ and diffuses to the neighboring basic sites to form nitrates or nitrites species [15,16]. Then, the fuel or fuel-rich pulse is injected into the catalytic converter chamber to reduce the stored NO_x to N₂, thereby regenerating the occupied NO_x storage sites and preparing them for the next lean-burn period [15]. Adopting this alternating operation is a smart innovation to avoid the inherent challenge of NO_x reduction in the net oxidizing atmosphere.

An NSR system is ~50% space saving compared to NH₃-SCR [11], which exhibits qualified NO_x removal efficiency as well. NSR can also be applied to lean-burn gasoline engine vehicles or coupled with SCR, hence removing the urea injection system of SCR and making it more cost effective [17,18]. However, the application of NSR also suggests technical and cost concerns: (1) basic sites are irreversibly occupied by SO, so that long-term usage leads to continuous deactivation of the catalyst; and (2) high loadings of noble metals (mainly Pt) are demanded in commercial NSR catalysts to ensure sufficient NO oxidation and NO_x reduction efficiency in frequently switching operation mode [11,16]. To date, the former constraint has been greatly alleviated with the execution of increasingly stringent legislation to reduce automotive SO_x emissions, realized by the deep desulfurization of diesel by hydrotreating in refineries [19,20], while the challenges of high cost and low thermal stability of Pt have become the bottleneck for further application of NSR, and the attempts to find economical and durable alternatives have never stopped [11,16].

Perovskite-type oxide ABO₃ has been proven to be a promising candidate, with the advantage of high thermal stability, prominent redox property, and low cost [21,22,23,24,25,26]. In a cubic perovskite-type oxide ABO₃ with space group type Pm-3m, the A-site cation is 12-fold coordinated by oxygen anions, while the B-site cations are octahedrally coordinated (6-fold) [23]. A-site cations are large and generally basic, serving as NO_x storage sites; and B-site cations are small and chemically active, functioning as the redox center for NO oxidation and NO_x reduction [27,28,29,30]. The perovskite-type oxide structure of perovskite allows the integration of redox and storage components with homogeneous distribution and strong mutual interaction that substantially reduces the distance of NO_x diffusion, which is desirable for achieving a high lean-burn NO_x trapping rate and preventing fuel-rich NO_x slip, as compared to other NSR catalysts that are typically composed of multiple active metals/oxides and basic components with limited interface [31].

The perovskite possesses excellent compositional versability, as it is able to accommodate nearly 90% of metals in the periodic table and form thermally stable oxides and able to endure partial cationic substitution, enabling flexible tuning of physio-chemical properties of the catalyst according to the reaction features [23,32,33,34]. Notably, any cation alternation or substitution can result in lattice expansion/shrinkage or a change in the extent of lattice distortion. For a specific formula, the selection of cations can refer to the tolerance factor criterion, that is, 0.8 < (r_A + r_O)/[√2(r_B + r_O)] < 1, which is reliable in most cases for predicting the structural stability of perovskite-type oxides [35]. Most of the research before 2010 investigated the perovskite simply as an NO_x storage material, as the substitute for BaO in a commercial NSR formulation. Alkaline-earth-metal-based perovskites, such as BaCeO₃, BaSnO₃, BaFeO₃, CaTiO₃, and SrTiO₃, were predominantly reported in this stage [36,37,38,39,40,41,42]. The main strategy was to use the perovskite-type oxide structure to modify sulfur- and sinter-resistant properties of the basic metals, whereas most of these formulations exhibited relatively inert redox properties and still required the addition of noble metals to facilitate NO oxidation and NO_x reduction. Over the last decade, La has gradually been accepted as an ideal option for A-site cation after extensive formulation screening. It was found to provide excellent structural thermal stability [3,16]. Most importantly, the proper ion radius of La³⁺ favors the accommodation of the transitional metals with prominent redox properties, such as Mn and Co. With the development of these perovskite formulations, noble metals have appeared less indispensable in NSR formulation.

In this review, we summarize the recent advances of the perovskites (mainly La-based) applied in the NSR field. Apart from highlighting the promising catalyst formulations and activity results, the review provides insights into the structure–property relationship that would instruct the rational design of perovskite-based NSR catalysts. Moreover, we disassemble the complex NSR reaction process into three general steps, that is, NO oxidation, NO_x storage/trapping, and NO_x desorption and reduction, and separately review the performance of perovskite catalysts in each step to provide deep and comprehensive perspectives on the advantages and potential problems of La-based perovskite employed in NSR reaction. The critical research needs are finally addressed for further development of this type of material.

2. NO Oxidation over La-Based Perovskites

NO oxidation to NO₂ as the first step of NO_x storage substantially impacts the overall storage efficiency in lean-burn periods [43,44], especially at low temperatures, when kinetic limitations dominate NO oxidation. This necessitates the incorporation of active oxidizing components, such as noble metals and perovskites, in NSR catalyst formulation. Over the Pt-based catalysts, NO oxidation was reported to proceed following the Langmuir–Hinshelwood (L-H) and/or Eley–Rideal (E-R) mechanism [45,46,47]. Both mechanisms describe O₂ adsorption and activation on noble metal clusters as the rate-limiting step of NO oxidation, but there is a divergence of opinion on whether the adsorbed NO on acid sites or gaseous NO takes part in the reaction. Weiss et al. [48,49] and Mulla et al. [50] pointed out that NO oxidation activity cannot be simply interpreted to be in inverse proportion to the cluster size of noble metal. Instead, a slight increase in cluster size decreased the binding energy of oxygen on coordinatively unsaturated Pt atoms, and thus increased the concentrations of oxygen vacancies and enhanced NO oxidation activity. However, long-term usage inevitably leads to the severe growth of Pt clusters and a dramatic activity decline. This motivates the later development of more thermally stable active centers for NO oxidation, such as perovskite.

Compared to noble metals, fewer mechanism studies have carried out on perovskites. Constantinou et al. [51] reported the reaction orders for NO₂, NO, and O₂ over LaMnO₃ to be around −1, 1, and 1, respectively, in accordance with the data collected over Pd- and Pt-based catalysts [49,52], which might indicate that NO oxidation over LaMnO₃ perovskite and noble metal catalysts share the same kinetic step. The negative reaction order of NO₂ implies a product inhibition effect. NO₂ strongly absorbs on NO oxidation sites with a high sticking coefficient [50,53,54,55,56], thereby inhibiting NO adsorption and oxidation. Considering the compositional versability of perovskite-based catalysts, the NO oxidation mechanism might vary as a function of the formulations. Besides L-H and E-R, the Mars−van Krevelen (MvK) mechanism might also function with lattice oxygen participating in the reaction.

As a general trend, perovskites possess comparable or even higher NO oxidation activity and much better durability than noble metal catalysts. He et al. [57] investigated the total substitution of Pt by LaCoO₃ perovskite in an NSR formulation and obtained an enhanced NO-to-NO₂ conversion of 38% over K₂CO₃/LaCoO₃/ZrTiO₄, compared to 18% over K₂CO₃/Pt/ZrTiO₄. The promising NO oxidation activity of the perovskite-based formulation was attributed to the low electron density in the d-orbit of Co³⁺, which was beneficial for accepting a lone pair on the N atom in NO. Afterwards, Chen et al. [58] also reported that the high oxidation state of B-site cations would be favorable for NO oxidation and identified the lattice oxygen bonded to Mn⁴⁺ as the active species. In this study, a series of nonstoichiometric La_xMnO₃ (0.9 < x < 1.11) catalysts were prepared. As the ratio of La/Mn decreased, the transformation of Mn³⁺ to Mn⁴⁺ was triggered to compensate for the charge imbalance caused by nonstoichiometry. As a concomitant of that, the NO-to-NO₂ conversion was promoted. As revealed in these studies, the high oxidation states of B-site cations, such as Co³⁺ and Mn⁴⁺, can be stabilized by a perovskite-type structure, which is considered a major reason for the superior oxidation activity of perovskite compared to single-transition metal oxides [59].

However, the B-site oxidation state does not always correlate well with activity trends. As a typical example, Li and coworkers [60] studied NO oxidation over Sr-doped LaCoO₃ and LaMnO₃ perovskites. The transformation of Mn³⁺ to Mn⁴⁺ induced by low-valence Sr²⁺ doping did not lead to a notable activity enhancement in this case. In contrast, the NO conversion over the LaCoO₃ increased from 69% to 86% at 300 °C after Sr doping. This was associated with the formation of oxygen vacancies in La_0.9Sr_0.1CoO₃, on which the weakly adsorbed oxygen species were highly active for NO oxidation. These observations were confirmed in a recent study by Onrubia et al. [61], which looked deeply into the effects of Sr doping on NO oxidation activity (Figure 1). They revealed that the amount of oxygen vacancies in the La_1−xSr_xMnO₃ did not further increase as Sr substitution reached 10%, and the charge imbalance caused by further Sr doping was compensated by Mn⁴⁺ formation, which, however, did not promote the NO oxidation activity. On the other hand, the generation of oxygen vacancy appeared to be the only mechanism for charge neutralization of La_1−xSr_xCoO₃. This leads to their conclusion that the amount of oxygen vacancies is the principal factor governing the NO oxidation activity of perovskites. In addition, Sr doping causes notable increases in the specific surface areas (SSA) of LaMnO₃ and LaCoO₃ perovskites.

Sr doping also alleviates the [CoO₆] octahedral distortion, and in some cases leads to the transition of the crystal structure of La-based perovskites from rhombohedral to cubic [62]. The Co³⁺ electronic configuration correspondingly switched from high/intermediate spin state to low spin state with empty e_g filling. This favors the adsorption of electrophilic absorbates, and hence enhances NO oxidation.

Ag⁺ doping induced a more substantial charge imbalance than Sr²⁺ doping, which is beneficial to the formation of oxygen vacancies. With 10% of Ag doped into the A site of the LaMnO₃ perovskite, the activation energy for NO oxidation declined from 66.7 ± 2.1 to 47.9 ± 1.8 kJ mol⁻¹ [63]. However, the maximum substitution amount of Ag in LaMnO₃ was only 20% at the A site.

Compared to A-site substitution, the alternation of B-site redox cations has a more direct impact on NO oxidation activity. Chen et al. [64] first investigated the activity of different B-site cations and provided the NO oxidation activity order of LaCoO₃ > LaMnO₃ > LaFeO₃, in accordance with the reducibility of the B-site cation. The results suggest that the adsorbed NO may react with lattice oxygen species. However, further examination reveals the inadequacies of such a simple correlation. For example, LaNiO₃ shows a higher reduction temperature than LaCoO₃ in H₂-TPR, but a lower NO-to-NO₂ conversion [65]. In effect, when lattice oxygen participates in NO oxidation, this reaction is based on the cycling of reversible redox couples, such as Co³⁺/Co²⁺ and Mn⁴⁺/Mn³⁺, corresponding to the cyclic lattice oxygen storage and reaction process. Reducibility is apparently not enough to describe the cycling property, which is also determined by other factors, including the ability of the perovskite to host oxygen vacancies, filling rate of gaseous O₂ to oxygen vacancies, and lattice oxygen mobility [61]. Furthermore, in some reported perovskite systems, the adsorbed oxygen species appear more active for NO oxidation than lattice oxygen. For example, Zhong et al. [65] prepared a series of Co-doped LaNiO₃ perovskites by a co-precipitation approach. As a general trend, Co doping increased the NO and O₂ adsorption amount, and the adsorption strength of O₂. The highest NO oxidation activity was observed over the LaNi_0.7Co_0.3O₃ catalyst, and this was ascribed to moderate adsorption strength and appropriate amount of adsorbed O₂.

Besides transition metals, the perovskite structure allows the substitution of noble metals into the B site as well, but this is always limited to a trace amount due to cost considerations and ionic size disparity [66,67]. Pt doping generally promotes NO oxidation of perovskite since Pt itself is a highly active oxidation center. For example, LaCo_0.92Pt_0.08O₃ shows a higher NO-to-NO₂ conversion of 66.7% at 300 °C than the Pt-free counterpart of 42.0% [66], but the high Pt loading (~6.0 wt.%) makes the results less promising. Pd was demonstrated in our study to be less active than Pt in NO oxidation [67]. Only a slight increase in NO oxidation rate could be observed for Pd-doped La_0.7Sr_0.3CoO₃ perovskite compared to the Pd-free counterpart in the kinetic regime, and the Pd-supported perovskite even showed a lower NO oxidation activity. The kinetics show that the incorporation of Pd in either way does not alter the NO oxidation pathway. Perovskite continues to be the dominant oxidation center.

3. NO_x Storage

Upon NO being oxidized to NO₂ and diffused to neighboring basic sites, NO_x storage takes place on the catalyst surface, which is a unique behavior of NSR catalysts. Considering the essential impact of NO_x storage efficiency in lean-burn periods on overall DeNO_x activity, great efforts have been devoted to developing functional basic material that satisfies NO_x storage capacity (NSC) in a wide temperature window. Most of the studies before the year 2010 focused on single oxides with varying basicity, morphology, and interaction with supports, especially on the Pt/BaO/Al₂O₃ model catalyst [11,16]. Perovskite accommodates basic metals in a complex oxide lattice with synergetic interaction with B-site cations, thus exhibiting new features for NO_x storage, distinct from the conventional basic components, which will be highlighted in this section with Pt/BaO/Al₂O₃ as the benchmark.

Table 1 summarizes the NSC of La-based perovskite catalysts reported in the recent literature. Note that NSC cannot be simply described as an inherent attribute of the basic component, which is comprehensively affected by multiple factors [51]. As a general trend, NSC varies with the temperature in a volcanic shape, restricted by insufficient NO-to-NO₂ conversions and sluggish NO_x diffusion at low temperatures (<250 °C), and by the loss of thermal stability of nitrates in the high-temperature region (>450 °C) [68]. With prominent NO oxidation ability, and a negligible NO_x diffusion barrier determined by the nature of the perovskite-type oxide structure, La-based perovskites generally exhibit higher NSC at low temperatures than the Pt/BaO/Al₂O₃ model catalyst, as observed over LaMn_0.9Fe_0.1O₃ and alumina-supported La_0.5Sr_0.5Fe_1−xTi_xO_3-δ (Figure 2) [59,69], showing the potential to reduce the NO_x slip during cold start of the diesel engine, a long-lasting headache for automotive NO_x removal. Along this line, it should be mentioned that a prominent NO oxidation and NO_x storage performance was recently observed over Sr-based perovskites, such as Mn, Fe, Co-doped SrTiO₃ and SrFeO_3−δ, and Ruddlesden−Popper (RP)-type layered Sr₃Fe₂O_7−δ perovskite, in the low-temperature range [70,71,72], which was ascribed to the high mobility and reactivity of lattice oxygen. Among them, the Mn-doped SrTiO₃ exhibits a unique NO adsorption behavior at 323 K, that is, NO is adsorbed on the perovskite surface using lattice oxygens without the occurrence of NO oxidation, and this enables the catalyst to show significant NO_x storage capacity in a kinetically restricted temperature region for NO oxidation. The excellent NO_x storage performance at low temperatures makes them not only promising NSR catalysts, but also potential passive NO_x adsorber (PNA) materials.

At medium temperatures (250–450 °C), the NSC of Pt/BaO/Al₂O₃ substantially increases as a consequence of improved NO-to-NO₂ conversions (Figure 2). In comparison, the NSC of La-based perovskites appears less promising, especially for the LaCoO₃ and LaMnO₃ without cationic substitution, the NSCs of which were reported to be lower than 100 μmol g⁻¹ at 300 °C [73,74], as compared to ~140 μmol g⁻¹ for Pt/BaO/Al₂O₃ with 20% of Ba loadings, reported in [59]. Say et al. [75] investigated the NO_x adsorption behavior of these two perovskites using in situ infrared spectroscopy. The NO_x adsorption yielded multiple types of N-bearing species, including monodentate nitrite and nitrate, bidentate nitrate, bridging nitrate, and a small amount of bridging/chelating nitrites, but the free nitrate ions were not detected, suggesting that only surface La sites in perovskite are functional for NO_x storage. The inaccessibility of bulk storage sites can be ascribed to the high structural stability of La-based perovskite, which hinders the bulk phase transformation during NO_x adsorption and desorption [40]. This constitutes the major difference between NO_x trapping on perovskites and on single basic oxides, that is, the latter allows the inward NO_x diffusion with phase transformation to nitrate [11]. The surface adsorption nature, to a large extent, accounts for the inadequate NSC of LaCoO₃ and LaMnO₃ perovskites at medium temperatures, even with the La-enriched surface [75], especially considering the poor texture property resulting from crystal growth of oxides at high calcination temperature. The surface areas of most proposed perovskites lie in the range of 5–20 m² g⁻¹.

Aiming to improve NSC, additional alkali or alkaline earth metals are always incorporated in the perovskite formulation. For La-based perovskites, Sr is a common choice, since the doped Sr not only provides a stronger NO_x storage site on the surface, but also, as stated earlier, enhances NO oxidation activity and increases the surface area of the perovskite [61]. Our study [76] first investigated NO_x storage on Sr-doped LaCoO₃ perovskite through FT-IR and demonstrated three possible pathways (Figure 3): (1) NO₂ combines with surface lattice oxygen and forms monodentate nitrate; (2) NO₂ is directly trapped onto the oxygen vacancies and forms the “nitrate on perovskite” species; (3) NO₂ is trapped by segregated SrCO₃ and forms free nitrate ions. The presence of segregated SrCO₃ greatly increases NSC, and the nitrates formed on SrCO₃ show higher thermal stability than those on perovskite sites. After NO_x storage, a surface enrichment of Sr was observed, indicating possible exsolution of Sr²⁺ from the perovskite lattice and migration to the catalyst surface owing to the interaction with NO_x.

Dong et al. [77] substituted 30% of La with Sr and prepared La_0.7Sr_0.3MnO₃ via a sol-gel procedure under optimized synthesis conditions. The calefactive velocity of 2 °C min⁻¹, calcination temperature of 700 °C, and pH of 8 were identified to favor the formation of pure perovskite phase and reduce particle aggregation, yielding the best NO oxidation and NO_x storage performance. Nonetheless, the best NSC of La_0.7Sr_0.3MnO₃ was only 170.4 μmol g⁻¹ at 350 °C (Table 1). As a comparison, the La_0.7Sr_0.3CoO₃ catalysts prepared via the same procedure present NSC in the range of 700.0–967.3 μmol g⁻¹ [67,76]. In effect, the type of B-site cations affects the existing state of Sr. Compared to Mn-based perovskite, Co-based perovskite is less prone to accommodate Sr into perovskite lattice. The segregated SrCO₃ phase can always be detected in XRD patterns of La_1−xSr_xCoO₃ [67,76].

Even though NSC is a significant descriptor for evaluating or predicting NO_x storage efficiency in lean-burn periods of NSR reaction, it only reflects the quantity of storage sites, rather than the quality. Indeed, the NSR reaction does not even demand huge NSC, but enough “reversible NO_x storage sites” that can rapidly trap NO_x and be effectively regenerated [67]. In this sense, compared to those sites embedded inside bulk, the surface sites are more desirable, since they skip the sluggish NO_x diffusion in the radial direction [31]. Along this line, efforts were devoted to exposing more surface storage sites of perovskite, and two strategies have proven effective.

(1) Preparing porous perovskite material. The construction of the porous structure of perovskite requires a hard or soft template. For example, Ye et al. [73] reported a nano-casting method for preparing mesoporous perovskite, with Santa Barbara Amorphous (SBA) as the hard template that was finally removed by NaOH solution. The as-prepared LaCoO₃ shows a high SSA of 75.0 m² g⁻¹ and NSC of 252 µmol g⁻¹, as compared to 93 µmol g⁻¹ for the catalyst synthesized by the conventional sol-gel method. In a recent study, Xie et al. [78] also used the nano-casting method to synthesize mesoporous LaCoO₃ with KIT-6 as the hard template. The as-prepared LaCoO₃-meso exhibits an even higher SSA of 246 m² g⁻¹ and two times higher NSC compared to the counterpart prepared via the sol-gel method. More oxygen defects were produced on the surface of LaCoO₃-meso with the confinement of the porous structure of the KIT-6 template, and this is revealed to improve oxygen mobility and favor fast NO_x storage and desorption. The majority of NO_x adsorbed on oxygen defects desorbed at low temperatures (<240 °C), indicating the weak basicity of such storage sites. Polymethyl methacrylate (PMMA) microspheres are the soft template generally used for synthesizing three-dimensionally ordered macroporous (3DOM) perovskite, and they can be completely removed by calcination. The as-prepared 3DOM perovskites are widely used for the simultaneous removal of soot and NO_x, since the macropores (>50 nm) can accommodate soot particles (<25 nm). However, for the application of porous perovskite materials, the mechanical strength and cost issues should be considered [23]

(2) Dispersing perovskite on oxide supports. This strategy has a simpler preparation procedure and is more commonly used. The properties of the support, to a large extent, affect the NO_x storage performance of perovskite. Supporting LaCoO₃ on ZrTiO₄ was found to be effective in improving NSC and reducing perovskite sintering [57]. Afterwards, You et al. [79,80] loaded 10% LaCoO₃ perovskite on CeO₂ and ternary (Y,Ce,Zr)O₂ oxide supports. Despite the low loading, the catalyst achieved promising NO-to-NO₂ conversions and NSC, which was attributed to the high dispersion of perovskite and abundant surface-adsorbed oxygen species on the Ce-based support. More recently, Onrubia-Calvo et al. [81,82,83] conducted a series of studies comprehensively investigating the La_1−xSr_xCoO₃ perovskite supported on Al₂O₃, a cheap sinter-resistant support with a large surface area. At low loadings, Co preferably combined with Al₂O₃ and induced a solid phase transformation to CoAl₂O₄; high loadings benefited the formation of perovskite phase, but the less active bulk perovskite phase was also generated. Moreover, 30% perovskite loading on alumina was identified as an efficient formulation to balance dispersion and amount of active component (Figure 4), achieving a high NO_x storage capacity of 306 μmol (g perovskite)⁻¹ at 400 °C, compared to the bulky counterpart of 115 μmol (g perovskite)⁻¹ [83]. The particle size of perovskite decreased from 50–150 nm to 20–50 nm in diameter after being supported on alumina. In a later study by Ecker et al. [59], alumina was again examined as support for perovskite. Fe was adopted as the B-site cation, partially doped by Nb, Ti, or Zr, to enhance sulfur resistance and phase stability. These catalysts, particularly La_0.5Sr_0.5Fe_1−xTi_xO_3-δ, exhibited promising hydrothermal stability, and the NSC decrease was less than 20% after aging at 750 °C. To date, the most promising results were reported by an early study conducted by Ding et al. [84], obtained by loading La_0.7Sr_0.3CoO₃ on the mesoporous SiO₂. The small crystalline size of perovskite confined in the mesoporous SiO₂ increased the surface area. The as-prepared catalyst has an SSA of 237.6 m² g⁻¹, more than ten times higher than the perovskite prepared by the conventional sol-gel process (18.7 m² g⁻¹). The mesoporous SiO₂-supported perovskite achieved an NSC of 5269.2 μmol g⁻¹ in 3 h. Since NO oxidation activity was not substantially improved with the increase in SSA, in this case, the enhanced NSC was mainly related to the high dispersion of perovskite particles. Additionally, owing to the interactions between perovskite and mesoporous SiO₂, the catalyst exhibited higher stability in the reducing atmosphere than perovskite prepared by the conventional sol-gel method.

At high temperatures (>450 °C), the storage sites on La-based perovskite surface are basically not functional, due to the relatively weak basicity, which is unable to stabilize nitrates [59]. This further necessitates the incorporation of strong basic components into perovskite-based catalyst formulation [81]. Besides the aforementioned Sr-containing perovskites, various formulations were proposed along this line, such as 5% K₂CO₃ impregnated on mesoporous LaCoO₃ [73], Ba/Al₂O₃ ball-milled with LaMnO₃ [85], and La_xBa_1−xCoO₃ prepared via sol-gel procedure [86,87,88]. The results suggest that the high-temperature NSR performance was enhanced in the presence of these strong basic components. In fact, the Ba-based perovskite exhibits better high-temperature NO_x storage performance than the La-based counterparts, such as the BaTi_0.8Cu_0.2O₃ proposed in Albaladejo-Fuente’s study [89]. However, the Ba-based perovskite shows lower chemical and structural stability, as indicated by the phase segregation and transformation during the NO_x storage reaction [89]. Ba²⁺ in the A site is also less resistant to CO₂ in the atmosphere than La³⁺ due to the strong basicity, and hence it is more prone to exsolution from perovskite matrix to form a surface BaCO₃ layer [90]. These issues require special consideration in future studies, and the potential competitive or synergetic effect between different storage sites in NO_x adsorption or desorption at various temperatures needs to be specifically investigated.

Table 1. Review of NO_x storage capacity of the La-based perovskites.

Catalyst Formulation (Preparation Method)	Reaction Conditions	NOx Storage, μmol g−1 (Temperature, °C)	Ref.
K2O/LaCoO3/ZrTiO4 (impregnation)	400 ppm NO/5% O2/N2; GHSV = 60,000 mL g−1 h−1	325.0 (350)	[57]
Pt/La0.5Sr0.5Fe0.5Ti0.5O3/Al2O3 (impregnation)	500 ppm NO/500 ppm CO/200 ppm C3H6/9% H2O/9% CO2/6.5% O2; GHSV = 80,000 h−1	117.0 (250) 150.0 (350) 50.0 (450)	[59]
La0.7Sr0.3CoO3 (sol-gel)	400 ppm NO/ 5% O2/ balanced N2; GHSV = 120,000 mL g−1 h−1	967.3 (300)	[67]
LaMn0.9Fe0.1O3 (combustion)	500 ppm NO/8% O2/N2; GHSV = 30,000 h−1	371.6 (100) 392.3 (300) 306.3 (400)	[69]
LaCoO3 (sol-gel)	400 ppm NO/5% O2/N2; GHSV = 80,000 mL g−1 h−1	93.0 (350)	[73]
Mesoporous LaCoO3 (nano-casting)	400 ppm NO/5% O2/N2; GHSV = 80,000 mL g−1 h−1	981.0 (350)	[73]
La0.7Sr0.3CoO3 (sol-gel)	800 ppm NO/5% O2/ N2; GHSV = 80,000 mL g−1 h−1	700.0 (300)	[76]
La0.7Sr0.3MnO3 (sol-gel)	800 ppm NO/5% O2/ N2; GHSV = 80,000 mL g−1 h−1	170.4 (350)	[77]
K2O/LaCoO3/(Y,Ce,Zr)O2 (impregnation)	400 ppm NO/5% O2/N2; GHSV = 45,000 h−1	639.0 (350)	[79]
LaCoO3/3% K2O/CeO2 (impregnation)	400 ppm NO/5% O2/N2; GHSV = 45,000 h−1	512.4 (350)	[80]
30% Pd/La0.7Sr0.3CoO3/Al2O3 (impregnation)	500 ppm NO/6% O2/Ar; GHSV = 123,500 h−1	97.3 (350)	[83]
La0.7Sr0.3CoO3/mesoporous SiO2 (deposition)	500 ppm NO/ 5% O2/ N2; GHSV = 80,000 mL g−1 h−1	5269.2 (300)	[84]
La0.7Sr0.3CoO3 (sol-gel)	500 ppm NO/5% O2/ N2; GHSV = 80,000 mL g−1 h−1	2405.0 (300)	[84]

Table 1. Review of NO_x storage capacity of the La-based perovskites.

Catalyst Formulation (Preparation Method)	Reaction Conditions	NOx Storage, μmol g−1 (Temperature, °C)	Ref.
K2O/LaCoO3/ZrTiO4 (impregnation)	400 ppm NO/5% O2/N2; GHSV = 60,000 mL g−1 h−1	325.0 (350)	[57]
Pt/La0.5Sr0.5Fe0.5Ti0.5O3/Al2O3 (impregnation)	500 ppm NO/500 ppm CO/200 ppm C3H6/9% H2O/9% CO2/6.5% O2; GHSV = 80,000 h−1	117.0 (250) 150.0 (350) 50.0 (450)	[59]
La0.7Sr0.3CoO3 (sol-gel)	400 ppm NO/ 5% O2/ balanced N2; GHSV = 120,000 mL g−1 h−1	967.3 (300)	[67]
LaMn0.9Fe0.1O3 (combustion)	500 ppm NO/8% O2/N2; GHSV = 30,000 h−1	371.6 (100) 392.3 (300) 306.3 (400)	[69]
LaCoO3 (sol-gel)	400 ppm NO/5% O2/N2; GHSV = 80,000 mL g−1 h−1	93.0 (350)	[73]
Mesoporous LaCoO3 (nano-casting)	400 ppm NO/5% O2/N2; GHSV = 80,000 mL g−1 h−1	981.0 (350)	[73]
La0.7Sr0.3CoO3 (sol-gel)	800 ppm NO/5% O2/ N2; GHSV = 80,000 mL g−1 h−1	700.0 (300)	[76]
La0.7Sr0.3MnO3 (sol-gel)	800 ppm NO/5% O2/ N2; GHSV = 80,000 mL g−1 h−1	170.4 (350)	[77]
K2O/LaCoO3/(Y,Ce,Zr)O2 (impregnation)	400 ppm NO/5% O2/N2; GHSV = 45,000 h−1	639.0 (350)	[79]
LaCoO3/3% K2O/CeO2 (impregnation)	400 ppm NO/5% O2/N2; GHSV = 45,000 h−1	512.4 (350)	[80]
30% Pd/La0.7Sr0.3CoO3/Al2O3 (impregnation)	500 ppm NO/6% O2/Ar; GHSV = 123,500 h−1	97.3 (350)	[83]
La0.7Sr0.3CoO3/mesoporous SiO2 (deposition)	500 ppm NO/ 5% O2/ N2; GHSV = 80,000 mL g−1 h−1	5269.2 (300)	[84]
La0.7Sr0.3CoO3 (sol-gel)	500 ppm NO/5% O2/ N2; GHSV = 80,000 mL g−1 h−1	2405.0 (300)	[84]

4. NO_x Desorption and Reduction

The engines are periodically shifted to fuel-rich conditions by introducing reductant pulses to regenerate the saturated storage sites, and NO_x is desorbed and reduced on the redox center. Since the desorption and the reduction occur in series in a rather short time, these two steps are basically inseparable and always examined in combination over the catalysts. Four factors mainly determine the efficiency of fuel-rich regeneration: (1) reaction temperature; (2) intrinsic reducing ability of the reductant (i.e., H₂ > CO > propene > propane); (3) stability of nitrate species, determined by the basicity of storage site; (4) redox property of the catalyst. In an early study [75], perovskite materials were found to be capable of activating N-O bonds at ca. 277 °C (even in the absence of reductants), with the formation of N₂ and N₂O as the NO_x decomposition products. This observation has long been considered to be an important indicator of La-based perovskites being promising catalysts for fuel-rich reactions. However, more detailed examinations revealed the deficiency of perovskite in NO_x reduction.

Table 2 summarizes the reaction conditions and DeNO_x activity of La-based perovskite catalysts in the recent literature. Our previous study [76] reported 71.4% DeNO_x activity and nearly 100% N₂ selectivity over the La_0.7Sr_0.3CoO₃ perovskite under alternating 3 min lean-burn/1 min fuel-rich conditions. However, only the NO_x species stored on perovskite sites were reduced in fuel-rich periods, while the saturated SrCO₃ site could hardly be regenerated, especially at low temperatures. Considering that such irreversible storage sites do not contribute to DeNO_x activity and may even cover the active sites on the surface, Peng et al. [91] used the diluted HNO₃ to remove the bulk SrCO₃ on La_0.5Sr_0.5CoO₃ perovskite, and observed a DeNO_x activity enhancement from 11.9 to 42.4% at 250 °C. These results indicate the mismatch between the strong basicity of SrCO₃ and relatively weak NO_x reduction ability of La-based perovskites. A-site cationic deficiency can also inhibit the formation of SrCO₃, which serves as an extra driving force for Sr doping into the perovskite structure, as observed with the La_0.5Sr_0.3MnO₃ catalyst in our study [31]. However, the DeNO_x activity of 43.2% over this catalyst at 350 °C was still insufficient for practical application, and the absence of strong basic sites inevitably leads to activity loss in the high-temperature region. It should be mentioned that the cationic deficiency affects the stability of the ABO₃ perovskite structure, and it can result in the formation of a complex layered intergrown structure and impurity phases [92], which should be carefully examined.

Shi [69] demonstrated that even with the excellent NO oxidation ability, the overall DeNO_x activity of LaMn_0.9Fe_0.1O₃ was still less than half that of conventional Pt/BaO/Al₂O₃ catalyst in the whole temperature range (30–400 °C), due to the weak NO_x reduction ability. In this case, a non-thermal H₂-plasma was employed to assist in activation of reductants in the fuel-rich periods (Figure 5a), and then an excellent DeNO_x activity of around 90% was achieved over perovskite with temperatures as low as 30 °C. Interestingly, the plasma did not significantly improve the DeNO_x activity of the conventional Pt/BaO/Al₂O₃ catalyst, over which the rate-limiting step was NO oxidation and NO_x storage (Figure 5b,c).

Even though the plasma technique provides great reinforcement for La-based perovskites, its application in a real aftertreatment system is restricted by cost constraints and product selectivity considerations. The noble metal is still required in the perovskite-based formulation to facilitate NO_x reduction. Wen et al. [93] synthesized a series of Pt/LaCoO₃/K₂O/Al₂O₃ with different Pt loadings (0, 0.3%, and 1.0%). On one hand, the addition of Pt enhanced the NO_x reduction ability of LaCoO₃/K₂O/Al₂O₃. On the other hand, the presence of LaCoO₃ increased the NO oxidation ability and NSC of Pt/K₂O/Al₂O₃, thereby reducing the Pt loading from 1.0% to 0.3% for achieving the equivalent DeNO_x activity in a wide temperature range (Figure 6). The combination of perovskite and a trace amount of noble metal seems to be a promising and economical formulation for efficient NO_x removal.

Compared to Pt, Pd is more commonly used together with perovskite, due to the lower activation temperature for hydrocarbons and better thermal stability [94]. The inadequate NO oxidation ability of Pd will not be problematic when perovskite is present in a catalyst formulation. Nishihata et al. [95,96] proposed a famous intelligent Pd-doped perovskite catalyst (LaFe_0.57Co_0.38Pd_0.05O₃) in 2002 characterized by the ‘‘self-regeneration” of Pd, which showed the behavior of repeatedly moving into and out of the perovskite lattice in response to lean/rich atmospheric alternations, and in this way the particle growth could be suppressed. Along this line, Ueda et al. [97] added a basic component and prepared an La_0.7Ba_0.3Fe_0.776Nb_0.194Pd_0.03O₃ catalyst, which showed a DeNO_x activity of 47% at 250 °C. Afterwards, we reported [98] a La_0.7Sr_0.3Co_0.97Pd_0.03O₃ perovskite formulation prepared via the sol-gel method with comparable activity and selectivity to Pt-based catalyst. A combination of characterizations, including XRD, XPS, and EAXFS, were used to identify the dissolution of Pd into the perovskite lattice.

With the aim of achieving higher DeNO_x activity, in a more recent study, we investigated the optimal combination methods of perovskite and Pd by comparing the DeNO_x performance of doped Pd cation and PdO impregnated on perovskite surface [67]. As a general trend, both Pd species significantly promote the DeNO_x activity of La_0.7Sr_0.3CoO₃ perovskite by accelerating NO_x desorption and reduction. The impregnated PdO on the surface more effectively accelerated these steps due to the higher accessibility, and the efficient regeneration in turn increased NO_x storage efficiency in the lean-burn periods. The effects of the spatial locations of Pd species were confirmed in a later study by Onrubia et al. [84], which also demonstrated the poorer accessibility of Pd cations accommodated in perovskite structure. In this study, alumina was used as the support to disperse Pd and perovskite. The formulated 1.5% Pd/30% La_0.7Sr_0.3CoO₃/Al₂O₃ catalyst shows a prominent DeNO_x activity of 86.2% at 350 °C.

Besides the effect of spatial location, there is also debate concerning the active chemical state of Pd on perovskite, which is more complicated, because the chemical state may change dynamically during lean-burn/fuel-rich alternations. By properly tuning metal–support interactions, an interesting in situ activation behavior was observed over the Pd/La_0.7Sr_0.3MnO₃ catalyst in our study (Figure 7a) [99]. The DeNO_x activity substantially increased from 56.1% to 90.1%, with N₂O production simultaneously suppressed. This behavior resulted from the transformation of Pd²⁺ to Pd⁰ in the reaction atmosphere. The generated Pd⁰ showed a stronger ability to activate C₃H₆, and thus achieved eight-fold higher TOF for NO_x reduction than Pd²⁺ (Figure 7b). Notably, the excessive strength of interactions suppressed the in situ formation of Pd⁰, thus lowering the De-NO_x activity even with high dispersion of Pd. These studies provide important information for designing efficient Pd-perovskite NSR catalysts.

Table 2. Review of the reaction conditions and DeNO_x activity of the La-based perovskites.

Catalyst (Preparation Method)	Reaction Conditions	DeNOx Activity, % (Temperature, °C)	Ref.
La0.5Sr0.3MnO3 (sol-gel)	Lean (50 s): 400 ppm NO/5% O2/ N2; rich: phase (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	43.2 (350)	[31]
LaCo0.92Pt0.08O3 (sol-gel)	Lean (120 s): 280 ppm NO/8% O2/8% CO2/N2; rich (30 s): 280 ppm NO/ 8% CO2/3.5% H2/ N2, GHSV = 72,000 h−1 (SO2 treatment: 100 ppm SO2/8% O2/ N2 for 45 min; regeneration: 3.5 vol% H2 at 500 °C for 12 h)	90.9 (350) 25.5 (350, SO2 treatment) 75.5 (350, regeneration)	[66]
La0.7Sr0.3CoO3 (sol-gel)	Lean (50 s): 400 ppm NO/5% O2/ N2; rich (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	51.6 (300)	[67]
La0.7Sr0.3Co0.97Pd0.03O3 (sol-gel)	Lean (50 s): 400 ppm NO/ 5% O2/N2; rich (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	74.4 (300)	[67]
1.4% Pd/La0.7Sr0.3CoO3 (impregnation)	Lean (50 s): 400 ppm NO/5% O2/N2; rich (10 s): 1000 ppm C3H6/ N2; GHSV = 120,000 mL g−1 h−1	90.4 (300)	[67]
LaMn0.9Fe0.1O3 (sol-gel)	Lean (10 min): 500 ppm NO/8% O2/N2; rich (2 min): 1% H2/ N2; GHSV = 10,000 h−1, plasma assisted	>80% (30–300)	[69]
LaCoO3 (sol-gel)	Lean (3 min): 500 ppm NO/ 1000 ppm C3H6/6.7% O2/N2; rich (1 min): 500 ppm NO/ 1000 ppm C3H6/ N2; GHSV = 80,000 mL g−1 h−1	71.4 (300)	[76]
1.5% Pd/30% La0.7Sr0.3CoO3/Al2O3 (impregnation)	Lean (150 s): 500 ppm NO/6% O2/Ar; rich (20 s): 512 ppm NO/ 3% H2/ Ar; GHSV = 123,500 mL g−1 h−1	86.2% (350)	[82]
La0.7Sr0.3CoO3 (sol-gel, acid wash)	Lean (3 min): 500 ppm NO/5% O2/ N2; rich (1 min): 500 ppm NO/1000 ppm C3H6/ N2; GHSV = 120,000 mL g−1 h−1	55.0% (300)	[91]
0.3% Pt/LaCoO3/K2O/Al2O3 (impregnation)	Lean (120 s): 500 ppm NO/ 8% O2/ Ar2, rich (120 s): 500 ppm NO/3.5%H2/Ar; GHSV = 72,000 mL g−1 h−1	80.0 (300)	[93]
La0.7Ba0.3Fe0.776Nb0.194Pd0.03O3 (glycine-nitrate method)	Lean (54 s): 512 ppm NO/200 ppm C3H6/10% O2/Ar; rich (6 s): 512 ppm NO/200 ppm C3H6/4% CO/Ar; GHSV = 60,000 mL g−1 h−1	47.0 (250)	[97]
La0.7Sr0.3Co0.97Pd0.03O3 (sol-gel)	Lean (2 min): 0 or 100 ppm SO2/ 500 ppm NO/ 6.7% O2/ N2; rich (1 min); 0 or 100 ppm SO2/ 500 ppm NO/ 1000 ppm C3H6/ N2; GHSV = 32,000 h−1	100 (325) 99.2 (325, with SO2)	[98]
2.1%Pd/La0.7Sr0.3MnO3 (sol-gel)	Lean (50 s): 400 ppm NO/5% O2/ N2; rich: phase (10 s): 1000 ppm C3H6/ N2; GHSV = 120,000 mL g−1 h−1	90.1 (350)	[99]
2.1%Pd/La0.7Sr0.3MnO3 (impregnation)	Lean (50 s): 400 ppm NO/5% O2/N2; rich: phase (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	72.0 (350)	[99]

Table 2. Review of the reaction conditions and DeNO_x activity of the La-based perovskites.

Catalyst (Preparation Method)	Reaction Conditions	DeNOx Activity, % (Temperature, °C)	Ref.
La0.5Sr0.3MnO3 (sol-gel)	Lean (50 s): 400 ppm NO/5% O2/ N2; rich: phase (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	43.2 (350)	[31]
LaCo0.92Pt0.08O3 (sol-gel)	Lean (120 s): 280 ppm NO/8% O2/8% CO2/N2; rich (30 s): 280 ppm NO/ 8% CO2/3.5% H2/ N2, GHSV = 72,000 h−1 (SO2 treatment: 100 ppm SO2/8% O2/ N2 for 45 min; regeneration: 3.5 vol% H2 at 500 °C for 12 h)	90.9 (350) 25.5 (350, SO2 treatment) 75.5 (350, regeneration)	[66]
La0.7Sr0.3CoO3 (sol-gel)	Lean (50 s): 400 ppm NO/5% O2/ N2; rich (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	51.6 (300)	[67]
La0.7Sr0.3Co0.97Pd0.03O3 (sol-gel)	Lean (50 s): 400 ppm NO/ 5% O2/N2; rich (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	74.4 (300)	[67]
1.4% Pd/La0.7Sr0.3CoO3 (impregnation)	Lean (50 s): 400 ppm NO/5% O2/N2; rich (10 s): 1000 ppm C3H6/ N2; GHSV = 120,000 mL g−1 h−1	90.4 (300)	[67]
LaMn0.9Fe0.1O3 (sol-gel)	Lean (10 min): 500 ppm NO/8% O2/N2; rich (2 min): 1% H2/ N2; GHSV = 10,000 h−1, plasma assisted	>80% (30–300)	[69]
LaCoO3 (sol-gel)	Lean (3 min): 500 ppm NO/ 1000 ppm C3H6/6.7% O2/N2; rich (1 min): 500 ppm NO/ 1000 ppm C3H6/ N2; GHSV = 80,000 mL g−1 h−1	71.4 (300)	[76]
1.5% Pd/30% La0.7Sr0.3CoO3/Al2O3 (impregnation)	Lean (150 s): 500 ppm NO/6% O2/Ar; rich (20 s): 512 ppm NO/ 3% H2/ Ar; GHSV = 123,500 mL g−1 h−1	86.2% (350)	[82]
La0.7Sr0.3CoO3 (sol-gel, acid wash)	Lean (3 min): 500 ppm NO/5% O2/ N2; rich (1 min): 500 ppm NO/1000 ppm C3H6/ N2; GHSV = 120,000 mL g−1 h−1	55.0% (300)	[91]
0.3% Pt/LaCoO3/K2O/Al2O3 (impregnation)	Lean (120 s): 500 ppm NO/ 8% O2/ Ar2, rich (120 s): 500 ppm NO/3.5%H2/Ar; GHSV = 72,000 mL g−1 h−1	80.0 (300)	[93]
La0.7Ba0.3Fe0.776Nb0.194Pd0.03O3 (glycine-nitrate method)	Lean (54 s): 512 ppm NO/200 ppm C3H6/10% O2/Ar; rich (6 s): 512 ppm NO/200 ppm C3H6/4% CO/Ar; GHSV = 60,000 mL g−1 h−1	47.0 (250)	[97]
La0.7Sr0.3Co0.97Pd0.03O3 (sol-gel)	Lean (2 min): 0 or 100 ppm SO2/ 500 ppm NO/ 6.7% O2/ N2; rich (1 min); 0 or 100 ppm SO2/ 500 ppm NO/ 1000 ppm C3H6/ N2; GHSV = 32,000 h−1	100 (325) 99.2 (325, with SO2)	[98]
2.1%Pd/La0.7Sr0.3MnO3 (sol-gel)	Lean (50 s): 400 ppm NO/5% O2/ N2; rich: phase (10 s): 1000 ppm C3H6/ N2; GHSV = 120,000 mL g−1 h−1	90.1 (350)	[99]
2.1%Pd/La0.7Sr0.3MnO3 (impregnation)	Lean (50 s): 400 ppm NO/5% O2/N2; rich: phase (10 s): 1000 ppm C3H6/N2; GHSV = 120,000 mL g−1 h−1	72.0 (350)	[99]

5. Resistance to Poisoners

An inherent defect of NSR catalyst is the susceptibility to sulfur poisoning, due to the close binding affinity between basic components and SO₂ that easily leads to the formation of thermally stable sulfate, which causes the severe loss of NSC and continuous deactivation of the catalyst [11]. The sulfur tolerance is a major concern for the application of an NSR catalyst. Our recent studies illustrated that La-based perovskites had stronger sulfur tolerance than conventional Pd/BaO/Al₂O₃ and Pt/BaO/Al₂O₃ catalysts, which was tentatively assigned to the relatively weak basicity of perovskite storage sites and the stability of perovskite structure [31,99]. The incorporation of Fe into the B site of perovskite could further enhance the sulfur tolerance of La-based perovskite [100]. After the presulfation, the NSC of the La_0.7Sr_0.3CoO₃ dropped by 43.6%, while that of the La_0.7Sr_0.3Co_0.8Fe_0.2O₃ only dropped by 6.4%, since SO₂ preferentially bonded with Fe to form Fe₂(SO₄)₃, which circumvented the poisoning of the Sr site.

The regeneration of sulfated catalyst would be enhanced in the presence of noble metals. For example, Pd⁰ species was revealed to be effective in reducing sulfur poisoning through promoting hydrogen spillover to regenerate adjacent sulfated storage sites before they developed into the irreducible bulk phase [98]. The DeNO_x activity of La_0.7Sr_0.3Co_0.97Pd_0.03O₃ declined from 100% to 70.2% after the presulfating treatment but was fully restored by H₂ reduction at a mild temperature (325 °C). In contrast, the DeNO_x activity of Pt/BaO/Al₂O₃ catalyst declined from 99.2 to 54.9% and could not be recovered by H₂ reduction. The study also reveals that the sulfur species are less likely to deposit on the catalyst in the co-feeding condition, as compared to long-term sulfur pretreatment. However, once the bulk sulfate species form on the basic sites, they can hardly be regenerated in H₂ under 450 °C. The influence of sulfur poisoning was also comparatively studied with Pt/BaO/Al₂O₃ and LaCo_0.92Pt_0.08O₃ [66]. Both catalysts suffered a drastic loss of NSC after a long-term presulfation (perovskite: 55.5 to 6.0 μmol g⁻¹, Pt/BaO/Al₂O₃: 53.9 to 2.9 μmol g⁻¹), as well as the decline of NO-to-NO₂ conversion. The recovery of NSC and NO oxidation activity of the catalysts by 3.5% H₂ reduction at 350 °C was time dependent. A long regeneration time benefited the removal of sulfate species from the catalysts but led to the damage to perovskite structure. The perovskite catalyst exhibited better regenerability (Figure 8). This was assigned to the formation of reducible cobalt sulfate on perovskite, which is less stable than bulk BaSO₄ formed on Pt/BaO/Al₂O₃.

H₂O is another poisoner for NSR catalysts and is known to impact NO_x storage performance via competitively adsorbing on basic sites, forming hydroxide [11]. Our study [99] revealed that compared with the conventional Pt/BaO/Al₂O₃ catalyst, the La-based perovskite showed slightly higher H₂O resistance, which was tentatively attributed to the higher NO-to-NO₂ conversion [99], since NO₂ could be more easily stored on basic sites than NO. Different from SO₂, H₂O poisoning is reversible and less severe; the perovskite catalyst’s activity can be gradually restored after shutting off the H₂O feed [66].

Besides SO₂ and H₂O, soot in the exhaust can also affect NSR performance. Toyota proposed a diesel particulate NO_x reduction (DPNR) filter system, which combines the functions of diesel particulate filter (DPF) and NSR and is operated under NSR cyclic conditions. Previous studies of Pt-based model NSR catalysts reveal that soot deposited on the surface of NSR catalyst reduces the NSC but promotes NO_x desorption and reduction [101]. However, there is a lack of research on perovskite-based catalysts for DPNR systems. Although some authors focus on the simultaneous removal of soot and NO_x with perovskite-based catalysts, they study the oxidation of soot and simultaneous reduction of NO_x [102,103]. This operation mode is very different from NSR.

6. Conclusions and Perspective

Diesel engines achieve lower fuel consumption and CO₂ emissions; however, the net oxidizing environment of exhaust is problematic for NO_x reduction. The conventional three-way catalysis designed for gasoline vehicles is not functional under lean-burn conditions, which drives the development of NSR and SCR techniques that basically cover different market niches. The NSR technique does not demand a complex hardware system to realize NO_x removal, so it is ideal for light-duty diesel vehicles with space constraints and cost considerations. However, NSR catalysts demand high loadings of Pt to achieve efficient NO_x removal, which adds to the cost and reduces the thermal stability of the catalysts. The La-based perovskites, considered an efficient, durable, and economical alternative to Pt, have been extensively researched over the past decade. Through comprehensively reviewing these reported formulations and activities, a global evaluation of La-based perovskite materials applied in the NSR process is provided.

Perovskite-type structure stabilizes the uncommonly high oxidation state of B-site cations, and thus exhibits significant NO oxidation activity, especially LaMnO₃ and LaCoO₃, which show even higher oxidation activity than Pt-based catalysts. This advantage is further strengthened by the partial substitution of A-site La by low-valence cations, such as Sr²⁺, Ag⁺, or Ba²⁺, since this triggers the formation of oxygen vacancies to neutralize charge imbalance. The prominent NO oxidation activity and eliminated NO_x diffusion barrier endow La-based perovskites with superior low-temperature NO_x storage capacity compared to conventional Pt/BaO/Al₂O₃, showing the potential to reduce the NO_x slip during cold start; however, at medium and high temperatures, La-based perovskites are less efficient for NO_x storage, which is ascribed to the surface adsorption nature, poor texture property, and relatively weak basicity of the La site. This motivates the adoption of various synthetic strategies to increase the exposed surface area of perovskite, and to incorporate stronger basic components into perovskite-based catalyst formulations. The major problem of La-based perovskites is the inadequate NO_x reduction performance in fuel-rich periods. This shows that the total substitution of noble metal by perovskite is not applicable. Nonetheless, a lower amount of noble metal is required in the presence of perovskite to reach the same DeNO_x performance, since perovskite covers the shortage in NO oxidation and low-temperature NSC of noble metal catalysts. The simultaneous use of perovskite and a trace amount of noble metal is proven to be a feasible strategy for achieving high DeNO_x activity in an economical way; therefore, how to increase the accessibility of noble metal and keep it in an active metallic state in a predominantly oxidizing atmosphere will be critical. Furthermore, the storage sites on La-based perovskite show strong sulfur tolerance, especially when Fe and Pd are incorporated. This increases the applicability to real automotive aftertreatment systems.

However, in-depth studies are still urgently needed for the future development of perovskite-based catalysts for NSR applications. The critical research areas and aspects that should be considered in future research mainly focus on the following aspects:

(1)

In-depth understanding of the NSR reaction mechanism in perovskites. To date, most mechanism studies concerning the NSR reaction were conducted with a Pt-based model catalyst, while for other types of catalysts, including perovskites, only vague pathways and plausible lines have been proposed. Considering the complexity of the reaction process, more in-depth experimental and theoretical studies should be conducted to unravel the detailed reaction routes, active sites, and intermediates to support the rational design of perovskite formulations with predictable properties.

(2)

Comprehensive comparison of NSR performance of La-based perovskites with others. The majority of perovskite-based materials reported for NSR applications over the last decade have been La-based. The obtained data reveals that La³⁺ provides prominent structural integrity with few impurities and little phase segregation, as well as thermal and chemical stability resistant to phase transformation during NO_x storage and reduction. Furthermore, the extensively studied LaCoO₃ and LaMnO₃ catalytic systems exhibit a good redox property, which is desirable for the NSR reaction. However, the weak basicity of A-site La³⁺ brings challenges for high-temperature NO_x storage, and the usage of La and Co increases environmental burdens from the perspective of sustainability. Some interesting data have been recently reported for Sr- and Ba-based perovskite formulations. On one hand, the stronger basicity of A-site alkaline earth metal provides better high-temperature NO_x storage performance, and the +2 oxidation state of the A-site cation increases the B-site oxidation state and/or generates oxygen vacancies desirable for NO oxidation and NO_x storage. On the other hand, these strongly basic cations are generally less resistant to CO₂, NO_x, and SO₂ in the reaction atmosphere, and hence are prone to exsolution from the perovskite matrix, leading to the risk of collapse of the perovskite structure, but this tendency may vary depending on specific formulations. At this stage, a comprehensive study of the effects of different A-site cations on the activity and stability of perovskite catalysts in NSR reactions will be rather helpful.

(3)

Synergetic/competitive relation between different storage sites. As stated earlier, a practical NSR catalyst should be capable of efficiently trapping and releasing NO_x in a wide range of temperatures. This indicates that multiple storage components with different basicity should be used in combination; however, almost no comprehensive research has been devoted to this line. Current research suggests that La-based perovskite is an excellent low-temperature NO_x storage component, but whether the combination of perovskite and stronger basic sites efficient for high-temperature NSR reaction can work as they are intended remains unclear.

(4)

Redox stability and resistance to H₂O, CO₂, and SO₂ poisoning. Most perovskite formulations are designed and tested under ideal laboratory conditions. The durability of perovskite catalysts in reducing atmosphere or when exposed to the common poisoners in automotive exhaust should be carefully considered and further clarified.

(5)

Sustainability has become an essential concern in catalyst design with the guidance of legislation and increasing environmental awareness of the community. However, this is entirely neglected in the currently reviewed publications. The environmental burden of using Pt on a per kilogram basis is calculated to be 3–4 orders of magnitude larger than base metals, in the aspects of global warming potential, cumulative energy demand, terrestrial acidification, freshwater eutrophication, and human toxicity [104], which necessitates the development of environmentally friendly base metal alternatives. Along this line, in perovskite formulation design, the metals at the lower end of the scale of environmental impacts should be preferably considered, such as manganese, iron, and titanium at B-site. These metals also result in decent NSR performance, while the usage of nickel and cobalt should be more prudent. For A-site cations, Sr and Ba are better choices than La.