Revealing the Roles of Cu/Ba on Ce-Based Passive NO_x Adsorbers

Abstract

At present, passive NO_x adsorbers (PNAs) represent one of the most effective technologies for addressing NO_x emissions from diesel engines during cold-start periods. Conventional PNAs, which primarily consist of noble metals (such as Pt, Pd, and Ag) loaded on metal oxides or zeolites, share the common drawback of high production costs. Consequently, developing low-cost PNAs with outstanding NO_x storage performance remains a significant challenge. In this study, a series of CuxBa5Ce adsorbents were synthesized using the impregnation method, and a monolithic adsorbent was employed to evaluate NO_x storage and release performance. Techniques such as XRD, UV-Vis DRs, H₂-TPR, XPS, and in situ DRIFTs confirmed the crucial roles of Cu and Ba in NO_x storage and release. Specifically, the incorporation of Cu into CeO₂ enhanced NO_x storage performance. Moreover, in the Cu3Ba5Ce adsorbent, the addition of Ba not only introduced new storage sites and altered the stability of NO_x adsorption species but also helped prevent the aggregation of CuO, thereby prolonging the complete NO_x storage duration and satisfying desorption temperature requirements. The Cu3Ba5Ce adsorbent exhibited the most favorable NO_x storage performance, including a complete NO_x storage time of 135 s and a NO_x storage efficiency exceeding 50% at 80 °C over a 10 min period. While PNAs loaded with noble metals, such as Pd/CeO₂ and Pt/CeO₂, exhibited NO_x storage efficiencies below 50% after adsorbing for 5 min at 80 °C. Therefore, this research offered a crucial strategy for developing non-noble-metal-loaded, Ce-based PNAs.

1. Introduction

Nitrogen oxides (NO_x), comprising NO and NO₂ emitted from diesel engines, represent a significant source of air pollution. The selective catalytic reduction in NO_x with ammonia (NH₃-SCR) is among the most effective technologies for mitigating NO_x emissions. The operational temperature range of NH₃-SCR catalysts typically exceeds 180 °C; however, the warm-up time from a cold start is usually 100–120 s [1,2]. With increasingly stringent emission regulations, further reducing NO_x emissions from diesel engines will face a greater challenge, especially during the cold-start period. A new concept of passive NO_x adsorbers (PNAs) was first proposed to store NO_x by the Ford Motor Company in 2001 and was reintroduced as a “Low Temperature NO_x Adsorber (LTNA)” in 2015. The adsorbent would adsorb and store NO_x as nitrate/nitrites species during the cold-start period; once the downstream SCR catalyst temperature increased above 180 °C, the stored NO_x would be released and reduced to N₂ by NH₃-SCR [3]. Hence, the PNAs should meet the following requirements: NO_x should be adsorbed during the cold-start period and have enough complete storage time (at least 100 s); and NO_x is desorbed in the operational temperature range of NH₃-SCR catalysts [4].

Two types of PNAs have recently garnered significant interest in both academia and industry: noble metal (Pt, Pd, and Ag)-supported metal oxides (Al₂O₃, TiO₂, CeO₂, or their solid solutions) and Pd-supported zeolites (Pd/SSZ-13, Pd/Beta, and Pd/ZSM-5) adsorbents [5,6,7,8,9,10,11]. Among these, CeO₂ is a candidate for PNA support because of its excellent redox properties and abundant oxygen vacancies, which facilitate the adsorption and storage of NO_x [12,13,14,15,16,17]. Additionally, noble metals supported on CeO₂ will further improve the NO_x adsorption capacity, and the desorption temperature is about 250–460 °C [18]. However, the NO_x adsorption capacity of Pt/CeO₂ is usually larger than that of Pd/CeO₂ [19], and both of the adsorbents exhibit NO_x storage efficiencies below 50% after adsorbing for 5 min at 80 °C. Additionally, the related characterizations revealed that NO was trapped primarily on the Pt and Pd sites in the intake air, and then the trapped NO was transferred to the surface of CeO₂ and stored as nitrate or nitrite species. And, transition metals (e.g., Pr, Y, Zr, etc.) are doped into CeO₂ to generate solid solution oxides [20,21], which are beneficial to the formation of surface oxygen vacancies (Ce³⁺-X) and active oxygen (Ce⁴⁺-O*) and then improve the NO_x storage efficiency [22,23]. It has been demonstrated that CuO can increase the adsorption and oxidation performances of NO [24]. Zeng et al. discovered that CuO highly dispersed on TiO₂ displayed excellent catalytic oxidation activity of NO rather than CuO particles [25]. In addition, loading alkali (Na, K) and alkaline earth metals (Ba) on the adsorbents can also enhance the basicity and thus improve NO_x storage capacity, e.g., BaO + NO + O₂→Ba(NO₃)₂/Ba(NO₂)_2. This stored nitrate species will decompose at elevated temperatures to release NO_x. Importantly, the nitrate and nitrite species formed on Ba were more stable than those stored on CeO₂, thus contributing to NO_x desorption in the operational temperature range of NH₃-SCR catalysts [26,27].

Although noble-metal-loaded PNAs exhibit high NO_x storage capacities at low temperatures [5,6,7,8,9,10,11,28], developing cost-effective, non-precious metal PNAs with high NO_x storage capacities is essential. In this study, a series of CuxBa5Ce adsorbents supported on CeO₂ with improved NO_x storage capacities and storage stabilities were prepared by the step-impregnation method. Various characterizations, including X-ray diffraction, UV-Vis diffuse reflection spectroscopy, X-ray photoelectron spectroscopy, H₂ temperature-programmed reduction, and in situ diffuse reflectance infrared Fourier transform spectra, were employed to investigate the presence of Cu and Ba on CeO₂ and elucidate their roles on NO_x adsorption and storage over CeO₂-based PNAs. This work aims to provide a novel strategy for designing cost-effective, non-noble PNAs.

2. Results and Discussion

2.1. Results of NO_x Adsorption/Desorption

The NO_x adsorption and NO_x storage efficiencies (NSE) of CeO₂, CuxCe, Ba5Ce, and CuxBa5Ce adsorbents are shown in Figure 1. Among the investigated adsorbents, pure CeO₂ exhibited a relatively poor NO_x storage efficiency of 10.1% and no complete storage time. NO_x storage performance was significantly improved after impregnated with 1 wt.% Cu on CeO₂, and the NSE of Cu1Ce was increased to 20.8% from 10.1% in CeO₂, but there was no change on the complete storage time. Increasing the loading of Cu, the NSE was increased to 43.4% in Cu3Ce and 40.7% in Cu5Ce, where the complete storage time of NO_x was extended to 68 s in Cu3Ce and 74 s in Cu5Ce. The above results indicated that the NSE and complete storage time of NO_x were both effectively improved via increased loading of Cu on CuxCe. To further improve the PNA performance of the adsorbents, 5 wt.% Ba was impregnated on CeO₂. The NO_x storage performance of Ba5Ce was not enhanced after loading Ba on CeO₂; however, the NSE of the CuxBa5Ce adsorbents were evidently improved after introducing Cu into Ba5Ce. The NSEs of Cu3Ba5Ce and Cu5Ba5Ce increased to 52.2% and 49.7%, respectively, surpassing the NO_x storage efficiencies of Ce-based PNAs loaded with Pd/Pt. And, their complete storage times were both extended to 135 s. According to the above phenomena, the NSEs and complete storage times of CuxCe adsorbents were mainly affected by the content of Cu, whereas the complete storage time would be further extended by introducing Ba into CeO₂.

Figure 2 shows the desorption curves and the desorbed amount of NO_x in different temperature ranges. The two desorption peaks located at ca. 140 °C and 340 °C were discovered on CuxCe, respectively. Cu3Ce possessed the maximum NO_x desorption amount among the investigated CuxCe adsorbents, but the desorption amount of NO_x would not be increased by further enhancing the loading of Cu to 5 wt.%, which was consistent with the NSE results. The above phenomena illustrated that the NSE and desorption amount of CuxCe were positively correlated. The desorption amounts of CuxBa5Ce were not changed by further loading Cu on Ba5Ce, but the desorption peak temperature of CuxBa5Ce shifted to higher temperatures, where the desorption temperature range was shifted from 150–340 °C of Cu3Ce to 200–460 °C of Cu3Ba5Ce, respectively. Although the desorption temperature of Cu3Ba5Ce was lower than those of Pt/Pd-loaded Ce-based PNAs (250–460 °C), it still met the practical requirements. The higher desorption temperature of CuxBa5Ce would be attributed to increase the storage sites of NO_x and/or changing the adsorption species of NO_x and its stability by introducing Ba into CeO₂ [26,27].

2.2. Textural and Structural Properties

The XRD results showed the effect of the addition of Cu and Ba on the structural properties of CeO₂. The main diffraction peaks were assigned to the fluorite cubic structure of CeO₂ (JCPDS 34-0394) for all samples, as displayed in Figure 3. In the CuxCe adsorbents, no diffraction peaks attributed to CuO were observed on Cu1Ce, whereas the intensity increased with increasing the content of Cu from 3 wt.% to 5 wt.%. A new diffraction peak attributed to BaCO₃ (2θ = 23.9°) was observed after further loading 5 wt.% Ba onto CeO₂. According to the report [29], the “low temperature LT-BaCO₃” was formed when the loading of Ba was 5 wt.%–6 wt.% via the reaction between BaO generated by the thermal decomposition of the precursor Ba(OAc)₂ and CO₂ in air, which displayed excellent NO_x storage performance but was easily decomposed to BaO and CO₂ between 400–800 °C. However, it was noteworthy that the intensity of the diffraction peak assigned to CuO was decreased after loading Cu onto Ba5Ce, and no CuO diffraction peak was observed on Cu5Ba5Ce. This result indicated that the loading of Ba onto CeO₂ would effectively inhibit the aggregation of CuO.

It has been reported that a decrease in specific surface area leads to a decrease in NO_x storage performance when CeO₂ is used as a NO_x storage site [30]. According to the textural and structural data in

Table 1. The main properties of studied samples.

Samples	SBET (m2/g)	Total Pore Volume (mL/g)	Average Pore Radius (nm)	Crystalline Size of CeO2 a (nm)
CeO2	152.3	0.25	3.3	7.6
Cu1Ce	143.6	0.21	3.1	7.6
Cu3Ce	136.9	0.22	3.2	7.6
Cu5Ce	133.9	0.23	3.4	7.6
Ba5Ce	119.8	0.22	3.3	7.6
Cu1Ba5Ce	116.1	0.20	3.5	7.7
Cu3Ba5Ce	104.7	0.19	3.7	7.8
Cu5Ba5Ce	104.1	0.17	3.3	7.7

^a Crystallite size of CeO₂ determined from the XRD diffraction peak by Scherrer’s equation.

, the crystalline size of CeO₂ almost did not change, but its specific surface area decreased after loading Cu and Ba. The specific surface area of Cu3Ce was 136.9 m²/g and declined by 23.5% to 104.7 m²/g of Cu3Ba5Ce, but the NSE of Cu3Ba5Ce was 7% higher than that of Cu3Ce, and the complete storage time was extended by 67 s, which indicated that the loading of Ba provided new storage sites and effectively compensated for the loss of CeO₂ storage sites.

The UV-Vis diffuse reflectance spectrum was employed to investigate the coordination and oxidation states of Cu. Three bands were observed on all samples, as shown in Figure 4; the strong band around 275 nm would be attributed to O²⁻→Cu²⁺ ligand to the metal charge transfer (LCMT) transition. The band at 340–450 nm could be assigned to the charge transfer involving (Cu-O-Cu)²⁺ cluster-like species, and the broad band (black dashed box) at 600–800 nm could be ascribed to the ²E_g→²T_2g spin-allowed transition of Cu²⁺ situated in a distorted octahedral symmetry. The spectra of CuxCe and CuxBa5Ce revealed that all the adsorbents showed the 340–450 nm band, indicating that loading Cu on CeO₂ would result in the formation of (Cu-O-Cu)²⁺ clusters, and similar results were reported by Shi et al. [31]. It has been demonstrated that the adsorption and storage of NO_x were improved by the formation of (Cu-O-Cu)²⁺ clusters [32,33]. A red shift was observed on the absorption edge at 340–450 nm, which may have been due to an increase in the number of (Cu-O-Cu)²⁺ clusters caused by an increase in Cu loading [34,35]. Additionally, the larger CuO particles were formed on the Cu5Ce due to a clear edge adsorption at 750–800 nm, but the adsorption intensities of Cu5Ba5Ce were weaker than those of Cu5Ce, which indicated that the loading of Ba inhibited the aggregation of CuO, in agreement with the result of XRD [36].

2.3. TPR

H₂-TPR was performed to investigate the reducibility of the different adsorbents, as shown in Figure 5. There were two or three reduction peaks between 100 °C and 250 °C after loading Cu onto CeO₂, but the reduction peak at 500–600 °C could have been attributed to the reduction in the surface active oxygen on CeO₂ and Ba5Ce. The low-temperature peak (α) was attributed to the reduction peak of highly dispersed Cu species, and the medium-temperature peak (β) was attributed to the reduction in (Cu-O-Cu)²⁺ clusters, which would have been consistent with the results of the UV-Vis DRs [37]. Two reduction peaks were observed on Cu1Ce, whereas a new (γ) peak that appeared on Cu3Ce and Cu5Ce would be assigned to the reduction in CuO particles with increasing Cu loading, which could be accordance with the results of XRD [31]. On the contrary, both (α) and (β) peaks could be identified, and no reduction peak assigned to the CuO particles was observed on CuxBa5Ce with an increase in Cu loading, which further demonstrated that the loading of Ba inhibited the formation of CuO particles, in agreement with the results of UV-Vis DRs and XRD.

2.4. XPS

XPS was performed to identify the chemical states of elements on adsorbents. The XPS spectra of Cu 2p, O 1s, and Ce 3d are shown in Figure 6, and the concentrations of surface Cu²⁺, Ce³⁺, chemically adsorbed oxygen, and the atomic ratios are listed in

Table 2. Surface compositions of the adsorbents.

Samples	Cu2+/Cu	Oα/O	Ce3+/Ce	O/Ce	Cu/Ce	Ba/Ce	Cu/Ba	Ba 3d a
CeO2	-	26.5	19.5	4.82
Cu1Ce	71.6	26.2	16.8	4.77	0.24
Cu3Ce	76.2	25.8	15.4	5.27	0.36
Cu5Ce	81.7	25.6	14.2	5.29	0.46
Ba5Ce	-	39.3	19.1	5.52		0.31		3.51
Cu1Ba5Ce	61.9	38.9	15.8	5.78	0.27	0.29	0.86	3.09
Cu3Ba5Ce	66.2	37.9	14.4	5.88	0.32	0.25	1.30	2.54
Cu5Ba5Ce	71.2	37.9	13.4	6.11	0.39	0.25	1.53	2.49

^a Surface composition (at.%).

.

As shown in Figure 6a, the spectra of Cu 2p were centered at 932.28, 933.78, 951.88, and 953.38 eV and represented the photoelectrons emitted from Cu⁺/Cu⁰ 2p_3/2, Cu²⁺ 2p_3/2, Cu⁺/Cu⁰ 2p_1/2, and Cu²⁺ 2p_1/2, respectively. The shift to higher binding energy with the increase in Cu loading amount indicated that the interaction between Cu and Ce was enhanced. The concentration of Cu²⁺ on the adsorbents was calculated as (Cu²⁺(%) = A(Cu²⁺)/Σ[A(Cu²⁺) + A(Cu⁺)] × 100 (A represents the peak area of each element)). As listed in Table 2, the concentrations of surface Cu²⁺ increased in the following order: Cu1Ce (71.6) < Cu3Ce (76.2%) < Cu5Ce (81.7%) and Cu1Ba5Ce (61.9%) < Cu3Ba5Ce (66.2%) < Cu5Ba5Ce (71.2%), and the ratios of Cu/Ce, Cu/Ba, and O/Ce gradually increased with the increase in Cu loading and the decrease in Ce content. The results illustrated that Cu²⁺ was the main chemical valence of Cu on all adsorbents, and its concentration increased with the increase in Cu content. However, it is noteworthy that the concentration of Cu²⁺ was approximately 10% lower after loading 5 wt.% Ba onto CeO₂. The possible explanation was that the electron density around Cu was affected by the strong electron donation from doped Ba species [38], thus reducing the concentration of Cu²⁺. Additionally, the ratio of Ba/Ce decreased with increasing Cu loading, which may have been due to the fact that excessive Cu loading would cover the Ba on the surface, leading to a decrease in Ba concentration, as evidenced by the decreased surface content of Ba, listed in Table 2.

The O 1s XPS spectra of the adsorbents are shown in Figure 6b. The band at the higher binding energy (531.48 eV) is ascribed to the surface oxygen O₂²⁻ and O⁻ (O_α; yellow color); the peak at the lower binding energy (529.23 eV) is assigned to the lattice oxygen O²⁻ (O_β; green color) [39]. The concentration of Oα as (Oα(%) = A(Oα)/Σ[A(Oα) + A(Oβ)] × 100 (A represents the peak area of each element)). For CuxCe and CuxBa5Ce adsorbents, there was a slight decrease in surface oxygen (O_α) concentration with the increase in Cu content; however, it is noteworthy that the concentration of O_α on the CuxBa5Ce adsorbents is 1.5 times that of the CuxCe adsorbents. For CuxBa5Ce, more surface oxygen is related to the maintenance of charge balance after doping Ba [40]. Surface oxygen, acting as active oxygen species, is favorable to boost the formation of NO₂, and then the NO_x storage is advanced.

The Ce 3d XPS spectra of the adsorbents are shown in Figure 6c, and the corresponding assignments are defined [41]. Two groups of spin orbital multiplets attributed to 3d_3/2 and 3d_5/2 are denoted as “u” and “v” in the binding energy range of 880–920 eV. It is widely reported that the u₁ and v₁ belong to the orbital peaks of the 3d¹⁰4f¹ electron state Ce³⁺, and u₀, u₂, u₃, v₀, v_2, and v₃ belong to the orbital peaks of 3d¹⁰4f⁰ electron state Ce⁴⁺. The concentration of Ce³⁺ on the adsorbents is calculated as (Ce³⁺(%) = A(v1) + A(u1)/Σ[A(v) + A(u)] × 100 (A represents the peak area of each element)). The concentration of Ce³⁺ decreases in the following order: CeO₂ (19.1%) > Cu1Ce (16.8%) > Cu3Ce (15.4%) > Cu5Ce (14.2%) and Ba5Ce (19.5%) > Cu1Ba5Ce (15.8%) > Cu3Ba5Ce (14.4%) > Cu5Ba5Ce (13.4%). The results indicate that the concentration of Ce⁴⁺ is increased with the increase in Cu content. According to the literature [42], the increased amounts of Cu²⁺ and Ce⁴⁺ are beneficial to generate more crystal lattice oxygen (O_β) species, providing a material basis for the formation of highly reactive oxygen species. This may have been the reason for the slight decrease in the concentration of surface oxygen (O_α) species.

In short, the XPS results show that Cu²⁺and Ce⁴⁺ are the main chemical valences of Cu and Ce for all adsorbents_, which can provide a material basis for reactive oxygen species and the formation of (Cu-O-Cu)²⁺ clusters. Additionally, the increase in surface oxygen species is favorable for the improvement of oxidation activity after the addition of Ba to CeO₂ and then improving the NSE of CeO₂.

2.5. NO_x-DRIFTs

NO_x adsorption behaviors of CuxCe samples were investigated by in situ DRIFTs measurements, as shown in Figure 7. The DRIFTs bands would be attributed to bridging bidentate nitrate (1577, 1562 cm^–1), monodentate nitrite (1457 cm⁻¹), monodentate nitrate (1290–1243 cm^–1), and bridging bidentate nitrite (1303, 1186 cm^–1) species found on pure CeO₂ [43,44,45]. The band of nitrite species decreased, whereas the band of nitrate species increased after injecting NO + O₂ for 10 min, which meant that the nitrite species were oxidized to nitrate species. In addition, the intensity of the adsorption bands would not be further strengthened by extend the injection time, indicating that the pure CeO₂ was saturated after NO_x adsorption for 10 min. For Cu1Ce, the main adsorption bands would be attributed to the bidentate nitrite species (1186 cm⁻¹), and the corresponding intensities were gradually increased after injecting NO + O₂ for 5 min. However, the adsorption intensities of the bidentate nitrite species (1211 cm⁻¹) on the Cu3Ce adsorbent were significantly higher than those of Cu1Ce within 10 min. For Cu5Ce, the adsorption intensity of the bidentate nitrite species (1211 cm⁻¹) was not further enhanced within 10 min, and the main species were changed to bridged bidentate nitrate (1567 cm⁻¹), monodentate nitrite (1457 cm⁻¹), and chelated bidentate nitrate (1278 cm⁻¹) species after 10 min, which may have been that the oxidation activity of NO was enhanced after loading excess Cu to facilitate the formation of various nitrate species [32]. The intensities of these nitrate and nitrite species on all CuxCe adsorbents were significantly higher than those of pure CeO₂. The above results suggested that the NO_x adsorption species were mainly in the presence of nitrite species on all CuxCe adsorbents within 10 min, and increasing the loading of Cu was advantageous to promote NO_x adsorption.

Figure 8 shows the NO_x adsorption behaviors of CuxBa5Ce samples. For pure Ba5Ce, the DRIFT bands at 1565, 1196, and 1126 cm^–1 were detected after injecting NO + O₂ for 30 min, which were attributed to bridging bidentate nitrate species on Ce sites, monodentate nitrites, and monodentate nitrates on Ba sites, respectively [46]. Loading 1 wt.% Cu on Ba5Ce, the monodentate nitrite species (1196 cm^–1) were found in the initial exposure stage and obtained their maximum intensities. Then, the monodentate nitrite species (1196 cm^–1) decreased slowly after 10 min, with plenty of bridging bidentate nitrites (1234 cm^–1) forming on Ba sites, and bridging bidentate nitrates (1565, 1604 cm^–1) forming on Ce sites. For Cu3Ba5Ce, the adsorption intensity of monodentate nitrite species (1196 cm^–1) formed on Ba sites was higher than those of Cu1Ba5Ce within 10 min. Moreover, the nitrate species would not be changed with further extended the adsorption time, and a weak shift to high frequency was found. After further increasing the loading of Cu to 5 wt.%, the adsorption intensities of bridging bidentate nitrates (1565, 1604 cm^–1) and monodentate nitrites (1024 cm^–1) formed on Ce sites were significantly higher than Cu1B5Ce and Cu3B5Ce. The monodentate nitrite species (1196 cm^–1) showed a significant red shift with increased the adsorption time of NO_x and transformed into bridging bidentate nitrites (1234 cm^–1) after 30 min, indicating that 5 wt.% Cu loading favored the formation of bidentate nitrite on Ba sites and monodentate nitrites on Ce sites.

The adsorbed nitrate/nitrite species on the Ce sites displayed poor thermal stability than those adsorbed on the Ba sites, which was probably attributed to the covalent or ionic bonds between Ba and nitrogenous species [47]. And, it was known that there were significant differences among the thermal stabilities of different nitrate/nitrite species (nitrite < bidentate nitrates < monodentate nitrates < ionic nitrates) [47]. Therefore, the thermal stabilities of the adsorbed nitrate/nitrite species increased in the following order: bidentate nitrite species (1211, 1186 cm⁻¹) formed on Ce sites < bidentate nitrites (1234 cm^–1) formed on Ba sites < monodentate nitrites (1196 cm^–1) formed on Ba sites. The above results demonstrated that the loading of Cu was beneficial to improve the amount of NO_x adsorption without the change in NO_x adsorbed species, whereas the loading of Ba provided new adsorption sites and changed the NO_x adsorption species to strengthen thermal stability.

3. Materials and Methods

3.1. Preparation of Adsorbents

A series of CuO/CeO₂ adsorbents were prepared by the conventional incipient wetness impregnation method from the precursors Cu(NO₃)₂·3H₂O (AR grade, Keshi, China) and the commercial CeO₂ powder (Sinocat, Chengdu, China) with different mass percentages (x = 1, 3, and 5) of CuO, labeled as CuxCe. Next, the as-prepared powders were achieved after drying at 105 °C for 12 h and calcining at 550 °C for 5 h.

In total, 5 wt.% BaO (denoted as Ba) from the precursor Ba(OAc)₂ (AR grade, Keshi, China) was supported on the commercial CeO₂ powder to prepare Ba5Ce as the same prepared process of CuxCe. Then, different mass percentages (x = 1, 3, and 5) of CuO from the precursor Cu(NO₃)₂·3H₂O were impregnated on the as-prepared Ba5Ce powder to prepare CuxBa5Ce (x = 1, 3, and 5) and followed the same dried and calcined process as CuxCe and Ba5Ce.

The resulting seven adsorbent powders were subsequently coated on honeycomb cordierites (cylinder, diameter 11 mm, length 26 mm, bulk 2.5 cm³, 62 cell/cm², Corning Ltd., Corning, NY, USA) to achieve the monolithic adsorbents. The detailed preparation process of the monolithic adsorbents referred to our previous works [48].

3.2. Catalytic Activity and Characterization

The PNA performances of adsorbents were tested using a continuous flow fixed-bed quartz reactor. The testing process procedures are shown in Figure 9. The NO, O₂, and N₂ flows were controlled by a 50 mL/min, 100 mL/min, and 1 L/min mass flowmeter before entering the reactor, respectively. The simulated reaction conditions were as follows: 500 ppm NO, 8 vol.% O₂, N₂ as balance gas, and a gas concentration error of about 10 ppm. The total space velocity of the gas was 30,000 h⁻¹, corresponding to a total flow rate of 1250 mL/min. The outlet concentrations of NO, NO₂, N₂O, and NH₃ were monitored by a Gas Fourier transform infrared spectrometer (Nicolet Antaris IGS-6700, Thermo Fisher Scientific, Waltham, MA, USA). NO_x was provided by the China National Testing Institute (Chengdu, China), and O₂ was provided by the Dongfeng Industrial Gas Company (Chengdu, China).

The NO_x storage efficiency (hereafter denoted as NSE) and NO_x desorption amounts (μmol) were calculated according to the following formula:

$$\mathrm{NSE}=\left(1-\frac{{\int }_0^t\left(\left[{\mathrm{NO}}_{\mathrm{x}}\right]\mathrm{out}\right)\mathrm{dt}}{{\int }_0^t\left(\left[\mathrm{NO}\right]\mathrm{in}\right)\mathrm{dt}}\right)\times 100\%$$

$${\mathrm{NO}}_{\mathrm{x}}\mathrm{desorption}\mathrm{amounts}=\frac{V\times {\int }_{t\left(\mathrm{T}0\right)}^{t\left(\mathrm{T}\right)}\left(\left[{\mathrm{NO}}_{\mathrm{x}}\right]\mathrm{out}\right)\mathrm{dt}}{22.4\times 1000}$$

where t is the NO_x storage time; V is the total flow rate; [NO]_in is the inlet NO concentration during NO_x storage; [NO_x]_out is the outlet NO_x concentration during either NO_x storage or the subsequent NO_x desorption period; t(To) is the start time of NO_x-TPD corresponding to the NO_x storage temperature before the temperature is raised; and t(T) is the end time of NO_x-TPD corresponding to the desired NO_x desorption temperature.

The N₂ adsorption–desorption was measured on a Quadrosorb SI automated surface area and a pore size analyzer (Quantachrome, Boynton Beach, FL, USA) at 77 K, and the sample was pre-treated at 300 °C for 3 h prior to the measurement. The surface areas of the adsorbents were calculated by the Brunauer–Emmett–Teller theory.

Powder X-ray diffraction (XRD) experiments were performed on a Rigaku DX-2500 diffractometer (Rigaku, Tokyo, Japan) using Cu Ka (l = 0.15406 nm) radiation. The tube voltage and current were 40 kV and 25 mA, respectively. XRD powder diffractograms were recorded at 0.031°/s intervals in the 2θ range of 10–90°.

The X-ray photoelectron spectra (XPS) data were recorded by the Thermo Scientific K-Alpha spectrometer with 12 kV high pressure and 6 mA operating current using Al Kα radiation. C 1 s (284.8 eV) was used for the internal standard to calibrate the binding energy.

UV-Vis diffuse reflectance spectra (UV-Vis DRS) were recorded in the range of 200–800 nm on a Shimatsu Uv-3600i Plus (Shimadzu, Tokyo, Japan).

Temperature programmed reduction with H₂ (H₂-TPR) experiment was carried out on a TP-5076 TPD/TPR dynamic adsorption (Xianquan, China) with a thermal conductivity detector. The samples approximately of 100 mg were pre-treated in a quartz tubular microreactor in a flow of pure N₂ at 450 °C for 1 h and then cooled to 30 °C. The reduction was carried out in a flow of 5 vol.% H₂-95 vol.% N₂ from 30 °C to 900 °C with a heating rate of 10 °C/min.

In situ DRIFT Spectra were performed using a Nicolet Nexus 6700 FTIR spectrometer (Nicolet, Waltham, MA, USA) with a diffuse reflectance infrared Fourier transform spectroscopy (Drifts) equipped with a reaction cell and a KBr window. The reaction temperature was controlled by an Omega program temperature controller (Omega, Norwalk, CT, USA) and a heating rate of 10 °C/min. The powder sample was activated in N₂ at 450 °C for 30 min. The background spectra of the sample were recorded in pure N₂ from 450 to 80 °C. The adsorbents were exposed to 2 vol.% NO-98 vol.% N₂/5 vol.% O₂-95 vol.% N₂ at 80 °C for 30 min, and the NO_x adsorption spectra at 80 °C of adsorbents were collected.

4. Conclusions

The transition metal Cu and alkaline earth metal Ba were utilized to enhance the NO_x storage efficiency and desorption performance of CeO₂. The specific roles of Cu and Ba in this process were investigated thoroughly. Cu3Ba5Ce with 3 wt.% of Cu and 5 wt.% of Ba showed the highest NO_x storage efficiency of 52.2% at 80 °C for 10 min, surpassing the NOx storage efficiency of Ce-based PNAs loaded with Pd/Pt. The desorption temperature range was 200–460 °C. The optimized functions of Cu and Ba could be summarized as follows. First, the loading of Cu on CeO₂ was beneficial to improve the amount of NO_x adsorption and storage, where Cu was present as Cu²⁺ in the form of (Cu-O-Cu)²⁺ clusters when the loading of Cu was ≤3 wt.% and agglomerated into CuO particles with the further increasing of the content of Cu to 5 wt.%, thus reducing the storage of NO_x. Then, the formations of CuO particles were inhibited, and more surface oxygen participated in the storage of NO_x after the addition of Ba to CeO₂. Also, CuxBa5Ce provided new storage sites, and nitrate/nitrite species adsorbed on the Ba sites were more thermally stable than those adsorbed on the Ce sites, which resulted in a shift of NO_x desorption temperature in the NH₃-SCR reaction temperature window. Consequently, the NO_x storage efficiency, complete storage time, and desorption performance of CeO₂ were effectively enhanced by the introduction of the non-precious metals Cu and Ba, along with further optimizations. The economy of Ce-based PNAs could be significantly improved.