Current Research on NO₂ Removal from Flue Gas Using Adsorption Methods

Abstract

Nitrogen dioxide (NO₂) is a major atmospheric pollutant and also a recoverable nitrogen resource, for which adsorption offers a promising technical pathway. This review systematically summarizes the recent progress in the removal of NO₂ from flue gas by adsorption methods, with a focus on material-level and process-level advancements. From the material perspective, three representative adsorbents—zeolites, activated carbons, and metal oxides—are comparatively evaluated in terms of their physicochemical properties, active sites, and adsorption mechanisms. Emphasis is placed on their adsorption capacity, selectivity, and hydrothermal stability, supported by both experimental and theoretical insights. From the process perspective, four adsorption-based technologies—Pressure Swing Adsorption (PSA), Temperature Swing Adsorption (TSA), Vacuum Pressure Swing Adsorption (VPSA), and Vacuum Temperature Swing Adsorption using multiple Gas circulations (GVTSA)—are analyzed regarding their principles, operational workflows, and engineering applications, with particular attention to the process intensification potential of GVTSA. The review identifies existing challenges in terms of material stability under complex conditions and process scalability, especially for severe environments such as nuclear reprocessing tail gases. Finally, future research directions are proposed toward developing multifunctional composite adsorbents with high capacity, strong environmental tolerance, and excellent regenerability, along with optimized and integrated adsorption processes, to achieve efficient NO₂ abatement and high-value recovery.

1. Introduction

Nitrogen oxides (NO_x) primarily consist of NO and NO₂, commonly found in industrial emissions, motor vehicle exhaust, biomass combustion exhaust, thermal power plant emissions, and nuclear fuel reprocessing exhaust. They represent one of the major sources of atmospheric pollution. In 2024, China’s mobile sources emitted 9.019 million tons of NO_x, while non-road mobile sources emitted 4.867 million tons. Research indicates that prolonged exposure to nitrogen oxides significantly increases the risk of respiratory and cardiovascular diseases, as well as mortality rates [1]. Furthermore, nitrogen oxides contribute to smog formation, exacerbate ozone layer depletion [2], promote acid rain formation [3], and cause photochemical smog pollution [4]. Consequently, effective nitrogen oxide removal has become a critical research focus. NO readily oxidizes to NO₂ in the atmosphere, while NO₂ is a high-value chemical feedstock, with high-purity liquid NO₂ products in China priced at approximately $6000 per ton. It finds applications in sulfuric acid production catalysts, nitric acid synthesis, and plutonium reprocessing in nuclear fuel backend operations. Consequently, large-scale NO₂ recovery can effectively treat exhaust gases, reduce treatment costs, and yield significant economic benefits.

To control NO₂ emissions from stationary sources, treatment strategies such as liquid absorption, Selective Catalytic Reduction (SCR), Selective Non-Catalytic Reduction (SNCR), biological denitrification, plasma activation, and adsorption can be employed. Selective Catalytic Reduction (SCR) [5] requires a catalyst to reduce NO₂ using reducing gases within the 250–450 °C range. This method is widely adopted as a commercial denitrification technology [6], but suffers from high costs and susceptibility to catalyst poisoning. Selective Non-Catalytic Reduction (SNCR) [7] operates without catalysts, promoting the reaction between NO₂ and reducing agents to form nitrogen gas at temperatures exceeding 950 °C, which is characterized by extremely high energy consumption. Liquid absorption methods [8] are primarily categorized into acid absorption and alkali absorption based on the type of absorbent solution. While treating large volumes of NO_x with low oxidation states, these methods exhibit low treatment efficiency and impose stringent requirements on the absorption cycle equipment. Biological denitrification [9] converts NO₂ into N₂ or HNO₃ through microbial metabolic activity. Its drawback lies in the sensitivity of microorganisms to environmental conditions, resulting in significantly reduced denitrification efficiency under high-temperature, high-dust, and high-load industrial scenarios, which limits its large-scale application. Plasma method [10] generates high-energy active particles via a plasma generator, which chemically react with NO₂ to convert it into HNO₃ or N₂. However, this approach consumes substantial electrical energy under high-frequency and high-voltage conditions. Increased oxygen concentration reduces denitrification efficiency, and equipment costs remain high. Among the above technologies, SCR and SNCR are most commonly applied in industrial settings. Their fundamental principle involves converting NO₂ into harmless N₂, aiming for complete pollution elimination. These methods are suitable for scenarios with stringent emission standards but no NO₂ recovery requirements. Meanwhile, NO₂ serves as a core intermediate in producing chemical raw materials such as nitric acid and nitrates. By recovering it and converting it into high-value-added products, we can reduce environmental treatment costs while achieving the circular utilization of nitrogen resources. Therefore, adsorption-based NO₂ recovery has emerged as a key technological pathway for efficient capture and resource utilization of NO₂ in industrial exhaust gases [11]. This approach leverages advantages such as mild operating conditions, recyclable adsorbents, and strong adaptability to low-concentration NO₂ systems.

Current research on NO₂ treatment primarily focuses on industrial waste gases and automotive exhaust, with extensive studies conducted under diverse operational conditions. In contrast, the purification of NO₂ generated from nuclear fuel reprocessing remains underexplored, despite its distinct and severe challenges. These tail gases typically contain not only moisture and CO₂, but also radioactive species such as RuO₄ and UO₃ particulates, leading to a highly corrosive and radiolytic environment [12]. Conventional adsorbents often exhibit inadequate radiation resistance and acid stability under such conditions, while existing NO₂ separation processes lack validated engineering parameters for radioactive service. Furthermore, the propensity of NO₂ to react with water to form corrosive acids, combined with its low boiling point and tendency to liquefy, complicates the recovery process in mixed gas streams. Currently, systematic studies on the development of radiation-resistant adsorbents and the adaptation of adsorption processes for nuclear off-gas treatment are scarce, highlighting a critical research gap that must be addressed to enable safe and efficient NO_x management in the nuclear fuel cycle.

This paper systematically reviews recent advancements in the performance evaluation of NO₂ adsorption materials and the process configurations of mainstream NO₂ recovery technologies, with a targeted focus on their application in the treatment of nuclear fuel reprocessing off-gases. The study specifically examines three major categories of adsorbents alongside four key adsorption processes. Through a detailed analysis of the physicochemical properties and adsorption behaviors of these materials under conditions such as high humidity, and the presence of corrosive species—this work elucidates the molecular-level mechanisms governing NO₂ capture, including the roles of active sites, surface modifications, and structural defects. Emphasis is placed on evaluating material performance based on critical metrics such as stability, adsorption capacity, and regenerability. Furthermore, the review identifies existing challenges in the long-term stability, selectivity, and engineering adaptability of adsorbents in radioactive and chemically aggressive gas streams. Finally, it outlines future research directions aimed at developing tailored adsorption materials and optimized processes for the efficient recovery and reuse of NO₂ from nuclear reprocessing tail gases, thereby contributing to safer and more sustainable waste management.

2. Research on NO₂ Adsorbent Materials

Currently, various promising NO₂ adsorbent materials have been investigated, including zeolites [11], activated carbons (ACs) [13] and metal oxides [14]. Each adsorbent material exhibits considerable NO₂ adsorption capacity. The selection of adsorbents requires comprehensive consideration of NO₂ concentration, temperature, flue gas composition, and regeneration costs. In the future, multifunctional composites represent a key direction for enhancing performance. The adsorption mechanism of adsorbents is primarily categorized into physical adsorption and chemical adsorption. Physical adsorption primarily relies on van der Waals forces to preferentially adsorb polar molecules such as NO₂ and H₂O. Chemical adsorption achieves adsorption by forming coordination bonds between acidic sites on the adsorbent surface and NO₂, or through reactions between metal ions and NO₂.

When evaluating the performance of NO₂ adsorbents, adsorption capacity, regeneration capability, stability, and selectivity are key indicators. These metrics reflect the feasibility and efficacy of the adsorbent in practical applications from different dimensions. Adsorption capacity serves as a fundamental indicator for evaluating adsorbent performance, defined as the maximum amount of NO₂ that can be adsorbed per unit mass or volume of adsorbent under specified conditions, typically measured in or mmol/g. It is influenced by the adsorbent’s specific surface area, pore structure, and the quantity and nature of active sites [15]. A high specific surface area and abundant pore structure provide greater attachment space for NO₂, resulting in superior adsorption capacity. Active sites can undergo chemical reactions with NO₂. In certain metal oxide-supported adsorbents, the metal active centers can form chemical bonds with NO₂, thereby enhancing adsorption capacity. Adsorption capacity for NO₂ is highly temperature-dependent and varies between physisorption and chemisorption mechanisms. For physisorption, which is an exothermic process governed by van der Waals forces, increased temperature typically reduces equilibrium adsorption capacity due to enhanced molecular desorption. In contrast, chemisorption, involving chemical bond formation, may exhibit an increase in adsorption capacity within a certain temperature range if thermal energy aids in overcoming activation barriers for surface reactions. However, excessively high temperatures can lead to adsorbent degradation or undesirable side reactions. Regeneration refers to the process of restoring the adsorption capacity of saturated adsorbents through physical or chemical methods [16]. This process must consider energy consumption, regeneration efficiency, and the effects on the adsorbent’s structure and properties. An efficient regeneration method should achieve a good recovery of adsorption performance and structural stability while minimizing energy consumption. Stability refers to an adsorbent’s ability to maintain its performance over extended periods under complex operational conditions, encompassing thermal stability, hydrothermal stability, and resistance to poisoning [17]. Thermal stability denotes the adsorbent’s capacity to retain its structure and properties without significant degradation under high temperatures. Elevated temperatures may cause structural collapse and sintering of active sites, thereby diminishing adsorption performance. Hydrothermal stability is critical for adsorbents treating waste gases containing water vapor. Water vapor presence may interfere with the NO₂ adsorption process or even react with the adsorbent, altering its structure and performance. Selectivity refers to the property of adsorbents to preferentially adsorb NO₂, while adsorbing minimal amounts of other gases such as O₂ and CO₂ present in exhaust gases [15]. By modifying the adsorbent surface or loading specific active components, its selectivity toward NO₂ can be adjusted, making it more suitable for applications with varying exhaust gas compositions.

2.1. Zeolites

Zeolite is a crystalline silicoaluminate material featuring a regular pore structure and strong adsorption properties. It primarily consists of SiO₂, Al₂O₃, and metal cations (such as Na⁺, K⁺, and Ca²⁺), represented by the chemical formula (M_x·[(AlO₂)m(SiO₂)n]·pH₂O), where M represents the metal ion. The channels are filled with water molecules; upon heating and dehydration, a porous structure forms, enabling selective adsorption of substances based on molecular size and polarity. Zeolite synthesis typically employs hydrothermal, solvothermal, or ionothermal methods [18]. The shape of the zeolite framework and pore size can be adjusted by varying the type of structural directing agent used. Additionally, the number of framework cations can be adjusted by altering the Si/Al ratio, and the types of these cations can be modified through ion exchange processes.

Zhengming Wang and co-workers [19] systematically investigated the adsorption and separation performance of A-type, FAU-type, and MOR-type zeolites for low-concentration CO₂-NO₂ mixtures in 1998, establishing the structure-activity relationship between molecular sieve structure and adsorption performance under adsorption and TPD experimental conditions at 333 K. Low-concentration CO₂ adsorption primarily depends on the polarity of the molecular sieve surface, while NO₂ adsorption is synergistically regulated by the number of cation sites and the pore geometry. Among these, 13X exhibits the highest NO₂ adsorption capacity due to its abundant cation sites and large pore size, but its thermal stability of adsorption sites is insufficient, leading to a decline in cycling performance. Zeolite A exhibited limited NO₂ adsorption capacity due to its smaller pore size; HM, obtained by protonation modification of mordenite, not only effectively suppressed CO₂ adsorption at low concentrations but also maintained high NO₂ adsorption capacity. It featured a mild desorption temperature of 409 K and excellent thermal stability, effectively preventing nitrate species formation at elevated temperatures. Considering separation selectivity, desorption energy consumption, and cycling stability, HM is proven to be the optimal adsorbent for separating low-concentration CO₂-NO₂ mixtures via temperature-swing adsorption. This study provides crucial experimental evidence and theoretical support for adsorbent selection in separating low-concentration acidic gases, such as those from nuclear waste treatment exhaust.

F. Delachaux and co-workers [20] investigated the effects of oxygen content, water vapor, and other factors on NO₂ adsorption by NaY zeolite. The presence of Al in the NaY zeolite framework imparts a negative charge to its surface, with Na⁺ serving as the charge-compensating cation. The high charge density of Na⁺ generates a strong Coulombic field around the pores, enabling the robust electrostatic adsorption of the polar molecule NO₂, resulting in physical adsorption. The chemical adsorption of NO₂ relies on Lewis acidic sites of Na⁺ on NaY, forming coordinate bonds with the lone pair electrons of NO₂. This promotes the chemical transformation of NO₂: on one hand, some NO₂ directly combines with Na⁺ to form stable NO₃⁻ species; on the other hand, NO₂ molecules form dimers N₂O₄ mediated by Na⁺, further enhancing adsorption stability. This process is particularly pronounced in experiments where dry NaY adsorbs NO₂. For fresh, undried NaY samples, NO₂ adsorption capacity reached 3.31 mmol/g without O₂ presence, with breakthrough occurring at 400 min. In the presence of 8.5% O₂, adsorption capacity was 2.89 mmol/g, with breakthrough at 350 min. Residual moisture causes partial conversion of NO₂ to NO, yielding 1.06–0.89 mmol/g and reducing NO₂ yield. For dried NaY samples, NO₂ adsorption capacity reached 2.03 mmol/g at 8.5% O₂ with a breakthrough time of 250 min, and almost no NO was generated. Drying reduced adsorption capacity by 38% but eliminated reverse conversion of NO₂, enabling high-purity NO₂ enrichment and better suitability for industrial regeneration scenarios.

In 2003, Joël Desprès’s team synthesized ion-exchange-modified Cu-ZSM-5 zeolite to investigate the adsorption behavior of NO₂ under dry and wet conditions at elevated temperatures [21]. Results indicated that at 200 °C, NO₂ undergoes a disproportionation reaction (3NO₂ + [CuO] ⇌ NO + Cu(NO₃)₂) to form nitrate species, enabling NO storage and release. This process exhibited a broad storage temperature range and was effective under both dry and wet conditions. The presence of 5% H₂O reduced both NO₂ storage capacity and NO₂ desorption, with the desorption decreasing from 0.116 mmol/g in the dry state to 0.078 mmol/g in the wet state. The NO₂ adsorption capacity of Cu-ZSM-5 decreases with increasing temperature. This is attributed to the thermal stability of the Cu(NO₃)₂ species formed via the disproportionation reaction, which remains stable below 380 °C, but decomposes into NO₂ above 400 °C. Consequently, NO₂ adsorption and storage become impossible above 400 °C. After adsorption-desorption cycles, the crystal structure remains stable with no significant degradation.

Mingzhe Sun and colleagues prepared CuO, Cu⁺, and Cu²⁺-containing Cun+SSZ-13 zeolite by reducing Cun+SSZ-13 with H₂, aiming to develop an adsorbent for efficient NO₂ removal at ambient temperatures [22]. NO₂ adsorption performance tests were conducted on samples prepared via reduction at different temperatures. Results indicated that at 190 °C, Cu⁺ dominated in Cu⁺-SSZ-13. The smaller Cu²⁺ ion radius (0.72 Å) imposes stronger d-orbital electron confinement, hindering coordination with NO₂ π* orbitals. At 250 °C, Cu²⁺ ions (0.96 Å) exhibit weaker d-orbital electron confinement, enabling efficient π-coordination. Consequently, Cu²⁺ demonstrates strong affinity for NO₂, synergizing with residual H⁺ sites to enhance NO₂ adsorption. However, when the reduction temperature is elevated to 250 °C, Cu⁰ becomes predominant in Cun+SSZ-13. After NO₂ adsorption, increased NO release indicates that Cu⁰ catalyzes the decomposition of NO₂ into NO, significantly reducing adsorption efficiency. Mingzhe Sun’s team similarly investigated the NO₂ adsorption performance of various transition metal ion-exchanged (Cu²⁺, Co²⁺, Ni²⁺, etc.) chlorophyllite (CHA) at ambient temperature [11]. They found that Co²⁺-exchanged low Si/Al ratio CHA (Co²⁺-2-L) exhibited outstanding adsorption capacity, with a capacity of 4.65 mmol/g. This reversible adsorption was achieved through π-bond coordination, resulting in NO₂ release below 14%.

Zeyu Tao and co-workers investigated the NO₂ adsorption performance of various divalent transition metal-modified molecular sieves [23]. Under dry and humid (RH = 70%) conditions, Ca-LTA-3 (Si/Al = 3) demonstrated optimal performance, achieving an NO₂ adsorption capacity of 3.94 mmol/g with NO release as low as 0.08%—significantly lower than the 20–30% observed in conventional adsorbents. Employing thermal regeneration, the regeneration efficiency remained above 97% after 10 cycles. Mechanistically, divalent cations stabilize the NO₂–H₂O reaction intermediate HNO₂, inhibiting its decomposition into NO. This strategy is applicable across zeolite frameworks including CHA, MOR, and FAU. Notably, under 500 ppm NO₂ and 70% RH humidification conditions, the modified material achieves 132 min of zero NO_x emissions, providing a practical and economical solution for NO₂ purification.

Haiyang Tao and colleagues systematically investigated the effects of zeolite Si/Al ratio, compensating cations, and topological structures on NO₂ adsorption performance for deep purification and recovery of NO_x in humid gases, elucidating the adsorption mechanism and optimizing strategies [24]. Mechanistically, Brønsted acid sites (BAS) on H-type decagonal linear zeolites (e.g., MFI-topology HZSM-5, FER topology ZSM-35) drive the NO-NO₂ catalytic cycle. Their one-dimensional linear channels (pore diameter ≈ 0.55 nm) eliminate diffusion bottlenecks, facilitating rapid NO and O₂ diffusion to active sites and ensuring continuous adsorption-oxidation cycles. The resulting NO₂ dimer (N₂O₄, size ≈ 0.58 nm) exhibits high channel compatibility. Combined with the strong NO₃⁻ binding derived from BAS, this significantly enhances adsorption stability and capacity. In contrast, cage-type zeolites like FAU-topology HY are prone to blockage of their supercage-window structures by H₂O or NO_x agglomerates. Their dispersed BAS distribution hinders continuous catalytic cycles, markedly reducing adsorption efficiency. Conclusions indicate: HZSM-5 adsorption capacity increases with decreasing Si/Al ratio, while HY achieves synergistic optimization of capacity and NO₂ recovery purity at Si/Al ≈ 5.5. H-type linear zeolites (HZSM-5, ZSM-35) exhibit outstanding water resistance, regenerability, NO_x breakthrough capacity (0.038–0.045 mmol/g), and NO₂ enrichment ratios (8.3–12.6) even at 90% humidity, outperforming Na-type and cage-type zeolites. Accordingly, HZSM-5 with a low Si/Al ratio is preferred when prioritizing adsorption capacity, while HY molecular sieves with a Si/Al ratio ≈ 5.5 are recommended when balancing capacity and NO₂ recovery purity.

Table 1. Adsorption capacity of NO₂ by zeolites with different Si/Al ratios and frameworks.

Zeolite	Si/Al Ratio	Framework	Adsorption Test Conditions	NO2 Adsorption Capacity (mmol/g)	Refs.
4A	1	LTA	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	0.36	[19]
5A	1	LTA	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	0.87
13X	1.6	FAU	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	2.99
HM	5	MOR	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	1.24
NaM	5	MOR	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	1.78
NaY	2.4	FAU	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	1.78
USY	6.1	FAU	800 ppm NO2, 950 ppm CO2, 0.11 L/min, 333 K	0.43
NaY(wet)	2.55	FAU	350 ppm NO2, 8.5% O2, N2, 1.17 L/min, 303 K	3.31	[20]
NaY(wet)	2.55	FAU	350 ppm NO2, N2, 1.17 L/min, 303 K	2.89
NaY(dry)	2.55	FAU	350 ppm NO2, 8.5% O2, N2, 1.17 L/min, 303 K	2.03
CuZSM-5	40	MFI	500 ppm NO2, 10% O2, N2, 5.0 L/min, 473 K	0.165	[21]
CuZSM-5	40	MFI	500 ppm NO2, 5% H2O, 10% O2, N2, 5.0 L/min, 473 K	0.139
Cun+SSZ-13-190-R	12	CHA	1000 ppm NO2, He, 0.06 L/min, 298 K	1.79	[22]
HSSZ-13	12	CHA	1000 ppm NO2, He, 0.06 L/min, 298 K	1.17
Ca-LTA	3	LTA	500 ppm NO2, Air, 0.2 L/min, 298 K	3.94	[23]
Mn-LTA	3	LTA	500 ppm NO2, Air, 0.2 L/min, 298 K	3.13
Na-LTA	3	LTA	500 ppm NO2, Air, 0.2 L/min, 298 K	3.87

shows the NO₂ adsorption capacities of different zeolite types. Zeolite crystal structures are highly defined, enabling the determination of ideal configurations and the optimal types and quantities of cations for achieving adsorption and separation. The adsorption mechanism of NO₂ on zeolites primarily relies on its coordination with metal cations within the pore channels. Zeolites exhibit outstanding adsorption performance for NO₂ in dried exhaust gases. However, in practical applications, ambient humidity significantly impacts their efficiency—water molecules compete with NO₂ for metal active sites, leading to reduced adsorption capacity. Consequently, developing zeolite materials with hydrophobic modifications or water-stable metal sites that maintain high NO₂ adsorption capacity and selectivity under humid conditions represents a key challenge in current research.

2.2. Activated Carbons

Activated carbons (ACs), as a typical porous carbonaceous material, has been extensively developed and applied as a filtration material in air purification fields [25]. Its preparation primarily involves two methods: physical activation and chemical activation [26,27], with specific process characteristics as follows. Physical activation is achieved by pyrolyzing carbon precursors under oxygen-limited or anaerobic conditions. Within the 200–600 °C temperature range, the precursor decomposes and volatilizes its volatile components. As temperature increases, chemical bonds between carbon atoms in the organic material rearrange, forming SP²-hybridized carbon atom layers—these layered structures resemble graphite and can stack via π bonds to form a layered architecture. Unlike high-purity graphite, the graphite-like layers in activated carbon often exhibit structural defects and distortions. Subsequent high-temperature controlled gasification of the pyrolyzed material under oxidizing agents (such as steam, carbon dioxide, or air) is required to refine the pore structure and adjust chemical properties. Chemical activation involves impregnating the precursor with chemical reagents, commonly including NaOH, KOH, ZnCl₂, and H₃PO₄, which can be applied either before or after pyrolysis. The interaction between these reagents and the carbonaceous material promotes pore formation and expansion. Compared to physical activation methods, chemical activation offers greater control over pore size distribution and surface chemical properties.

Zhongwei Li [13] and colleagues investigated the characteristics and mechanisms of NO₂ adsorption by activated carbon. NO₂ binds to saturated carbon atoms via chemical adsorption, forming C-NO₂ structures; another portion adsorbs physically, primarily attaching to saturated carbon atoms, hydroxyl groups, ether bonds, and ester groups. When the temperature was lowered from 30 °C to −20 °C, the adsorption capacity for 200 ppm NO₂ increased from 1.26 mmol/g to 2.16 mmol/g. Li Zhongwei also studied the adsorption capacity for NO at the same concentration and found that its value was consistently lower than that of NO₂. Density functional theory calculations indicate that the physical adsorption energy of NO₂ on activated carbon surfaces ranges from −38.5 to −82.7 kJ/mol, while NO ranges from −5.2 to +44.5 kJ/mol. This demonstrates that NO₂ exhibits significantly stronger binding affinity to activated carbon surfaces than NO.

Zohra Belala [28] and colleagues prepared activated carbon using date pits as precursors via CO₂ physical activation (activated at 850 °C for 30, 60, and 120 min, respectively) and investigated its NO₂ adsorption performance at low temperatures in a fixed-bed reactor. Results indicated that extended activation time increased the charcoal’s specific surface area and micropore volume. The C120 sample activated for 120 min exhibited an adsorption capacity of 2.32 mmol/g, outperforming multiple traditional woody lignocellulosic biomass-derived activated carbons. Adding 10% O₂ to the feed gas slightly increased adsorption capacity, while raising the temperature from 20 °C to 60 °C caused a significant decrease in adsorption capacity. TPD and TGA analyses revealed that NO₂ could be directly adsorbed or reduced to NO. The adsorption mechanism was related to the microporous structure and surface oxygen-containing functional groups, confirming that date pit-based activated carbon is an efficient NO₂ adsorbent.

Imen Ghouma with co-workers utilized industrial waste olive pomace as a precursor to prepare three types of activated carbon—AC-H₃PO₄, AC-CO₂, and AC-H₂O—via three activation methods: CO₂ physical activation, steam physical activation, and H₃PO₄ chemical activation [29]. Adsorption performance testing was conducted under low NO₂ concentrations of 5 ppm. Among these, AC-H₂O exhibited the highest adsorption capacity due to its abundant basic surface groups and mesoporous structure. AC-H₃PO₄, possessing more acidic surface groups, demonstrated the strongest activity in reducing NO₂ to NO but the lowest adsorption capacity. This demonstrates that the surface chemistry and pore structure of activated carbon jointly determine its adsorption performance: basic groups promote NO₂ adsorption, while acidic groups tend to reduce NO₂ to NO, which has weaker affinity for activated carbon than NO₂. Mesoporous structures enhance NO₂ diffusion efficiency within the carbon, thereby improving adsorption efficiency.

To enhance the adsorption performance of activated carbon, Kiman with co-workers prepared Co₃O₄-loaded activated carbon [30]. Activated carbon loaded with 15% Co₃O₄ proved optimal for NO₂ adsorption, exhibiting a saturation capacity of 4.06 mmol/g. After five adsorption-desorption cycles, its capacity retention remained at 89.2%, with significantly higher selectivity toward NO₂ compared to the ACs matrix and other loading samples. Eleni’s team modified wood-based activated carbon with dimethylamine to investigate the effect of organic amines on the denitrogenation performance of activated carbon materials [31]. Results indicate that samples pre-oxidized with nitric acid followed by dimethylamine modification (BOA) achieved optimal performance at 1.07 mmol/g, demonstrating a fivefold increase in NO₂ adsorption capacity compared to oxidized but unmodified carbon (BO). Characterization via elemental analysis and FTIR confirmed that pre-oxidation promoted the introduction of nitrogen-containing functional groups—such as amide groups and amine carboxylates—into the carbon matrix through acid-base reactions. These functional groups react with NO₂ to form nitrosamines and other products, while the carbon matrix undergoes re-oxidation by NO₂ and releases NO. This reveals the core mechanism of reactive NO₂ adsorption by the modified activated carbon.

Aleksandra B.W. and co-workers prepared nitrogen-rich activated carbon using direct coking coal as raw material through the following process: hydrochloric acid-hydrofluoric acid demineralization → pyrolysis (500 °C) → NaOH activation (800 °C) → ammonia oxidation (300/350 °C, ammonia-air mixture) [32]. They conducted a systematic study on the physicochemical properties of the adsorbent and its NO₂ adsorption performance. Results indicate that the ammonia oxidation stage significantly influences nitrogen content and adsorption performance. Samples undergoing ammonia oxidation at 350 °C after activation exhibited optimal performance, achieving dry NO₂ adsorption capacities of 0.51 mmol/g and wet (70% relative humidity) capacities of 1.63 mmol/g. The material simultaneously exhibits a high specific surface area of 2153 m²/g and abundant basic functional groups. In the wet environment, the reaction between NO₂ and water to form nitric acid (3NO₂ + H₂O → 2HNO₃ + NO) significantly enhances adsorption capacity. The Langmuir model (R² > 0.989) indicates monolayer adsorption of iodine, confirming the material’s outstanding adsorption properties and enabling high-value utilization of coal-based waste.

Table 2. Adsorption capacity and stability of NO₂ by activated carbon prepared from different precursors and different preparation methods.

Precursor	Preparation Method	Adsorption Test Conditions	NO2 Adsorption Capacity (mmol/g)	Refs.
/	Degassing at 200 °C for 2 h	200 ppm NO2, N2, 0.5 L/min, 303 K	1.26	[13]
/	Degassing at 200 °C for 2 h	200 ppm NO2, N2, 0.5 L/min, 253 K	2.16
Date pits	Activated for 2 h at 850 °C	500 ppm NO2, N2, 0.33 L/min, 293 K	2.32	[28]
Date pits	Activated for 2 h at 850 °C	500 ppm NO2, N2, 0.33 L/min, 333 K	1.85
Olive pomace	50% phosphoric acid solution stirred at 110 °C for 9 h	5 ppm NO2, 50%RH air, 23.5 L/min, 296 K	0.183	[29]
Olive pomace	Pure CO2 gas flow, 750 °C activation for 6 h	5 ppm NO2, 50%RH air, 23.5 L/min, 296 K	0.313
Olive pomace	Water vapor flow, activated at 750 °C for 6 h	5 ppm NO2, 50%RH air, 23.5 L/min, 296 K	0.348
Wood	Dimethylamine (DMA)-modified commercial BAX-1500	1000 ppm NO2, air, 0.45 L/min, 296 K	1.07	[31]
Orthocoking coal	Chemical activation with NaOH at 800 °C, accompanied by ammonia oxidation	1000 ppm NO2, air, 0.45 L/min, 298 K	0.51	[32]
Orthocoking coal	Chemical activation with NaOH at 800 °C, accompanied by ammonia oxidation	1000 ppm NO2, air, 0.45 L/min, 298 K	1.63

shows the NO₂ adsorption capacities of different activated carbons. Activated carbons, as a widely available porous material with high specific surface area and ease of functionalization, exhibits excellent adsorption capacity for NO₂ under low-temperature, atmospheric-pressure conditions. Research indicates that its adsorption performance is jointly determined by physical structure and surface chemistry. However, current research still faces bottlenecks. Addressing the insufficient water resistance and interference tolerance of activated carbon, its susceptibility to poisoning in complex industrial flue gases, and the design and development of materials tailored for specific scenarios represent key challenges that need to be overcome in the future.

2.3. Metal Oxides

Metal oxides serve as adsorbents primarily due to their highly reactive surface atoms, which readily interact with other substances. This capability stems from key characteristics: first, the metal cations and oxygen anions on the surface form Lewis acid-base pairs. Their acidity or basicity depends on the metal ion’s electronegativity, valence state, and coordination environment. Basic sites are mainly provided by surface crystalline O₂⁻ or OH⁻. These sites are electron-rich and exhibit strong chemical affinity for acidic gases like NO₂, forming stable surface species through chemisorption. Appropriate synthesis methods enable the preparation of metal oxides with high specific surface areas and controllable pore structures. High specific surface area ensures sufficient exposure of active sites, while suitable pore structures facilitate gas molecule diffusion and transport, enabling effective utilization of internal active sites. Oxygen vacancies represent intrinsic, ubiquitous point defects on metal oxide surfaces. They serve not only as high-energy adsorption sites but also significantly alter the electronic structure of the material surface. The formation of oxygen vacancies increases the electron density around surrounding metal ions, enhancing their electron-donating ability and thereby strengthening surface basicity. These vacancies themselves act as highly active sites, capable of directly activating small molecules like O₂ and H₂O or serving as initiation sites for adsorption reactions, effectively capturing and dissociating target gas molecules. Metal oxides typically exhibit excellent thermal stability and mechanical strength, maintaining structural integrity under high temperatures and harsh chemical environments—a critical factor for sustained operation in industrial adsorption processes. Furthermore, through elemental doping, composite oxide construction, or morphology control, their surface acidity/basicity, redox capabilities, and pore structures can be precisely tuned to achieve highly selective, high-capacity adsorption for specific adsorbates. Table 3 shows the NO₂ adsorption capacities of different metal oxides.

Yaying Ji together with co-workers prepared Pt/Al₂O₃ and Pt/La-Al₂O₃ using γ-Al₂O₃ as the substrate via the impregnation method [33]. They investigated their NO_x adsorption and desorption performance in the low-temperature range of 80–160 °C to evaluate their suitability as passive NO_x adsorbents (PNA). The addition of 1 wt% La created new NO_x storage sites, nearly doubling the initial NO_x storage efficiency (NSE). However, the NO_x desorption efficiency (NDE) below 250 °C was slightly lower than that of Pt/Al₂O₃. After five adsorption-desorption cycles, the enhancing effect of La disappeared because the strong storage sites could not be regenerated at 250 °C. In contrast, the equilibrium adsorption and desorption performance of unreduced Pt/Al₂O₃ exhibits a maximum NO_x desorption amount of 0.029 mmol/g at 250 °C, making it a preferred material for low-temperature NO_x adsorption.

Xinyu Li and colleagues investigated the low-temperature desorption performance of CeO₂/γ-Al₂O₃ materials with Ce loading ranging from 0.5 to 5.0 wt% as passive NO_x adsorbents (PNA) [34]. Characterization via XRD, TEM, and UV-vis spectroscopy confirmed Ce dispersion as CeO₂ nanoparticles on the γ-Al₂O₃ surface, with particle size increasing with Ce loading (reaching 8.0–9.3 nm at 4.0–5.0 wt%). PNA performance testing and in situ NO-DRIFTS indicate two NO_x adsorption pathways: rapid direct NO capture (forming nitrite, 1204–1225 cm⁻¹) and slow NO₂ capture (forming nitrates, 1350–1605 cm⁻¹), where nitrites decompose at 200–300 °C to release NO, meeting PNA requirements; The 3.0 wt% Ce-loaded 3.0 CA sample exhibited optimal comprehensive performance, achieving a NO_x storage capacity of 0.0297 mmol/g at 120 °C. After 16 h of hydrothermal aging at 750 °C, the capacity retention reached 79.8%, and after 5 CO cycles, it retained 94%. Although the specific surface area of the 3.0 wt% Ce-loaded sample exhibited a specific surface area (122.3 m²/g) slightly lower than pure Al₂O₃ (142.1 m²/g), but it decreased by only 11% after hydrothermal aging, significantly less than the 27% reduction observed in pure Al₂O₃. This indicates that Ce stabilizes the pore structure of Al₂O₃, preventing site failure due to structural collapse during adsorption.

Kim and the co-workers also used γ-Al₂O₃ as the support and employed the co-impregnation method to prepare a Cu/Ba bimetallic modified adsorbent (Cu10Ba10), investigating its NO_x storage and regeneration performance at 150 °C [35]. Results indicate that Cu10Ba10 exhibits a NO storage capacity (NSC) of 0.078 mmol/g (5.6 times higher than the Ba-modified adsorbent Ba20) and superior initial NO₂ storage efficiency (NSE) compared to Ba20. Regarding desorption performance, the peak desorption temperature for NO₂ on Cu10Ba10 decreased to 350 °C (lower than Ba20’s 500 °C), and it retained 79% of its initial NSC after 5 cycles (compared to only 57% for Ba20). Mechanistically, DRIFTS and DFT confirm that NO preferentially adsorbs onto the CuO surface (forming nitrite/nitrate) before migrating to the BaCO₃ surface to convert into stable double nitrate (1315–1400 cm⁻¹). while NO₂ strongly adsorbs at CuO oxygen vacancies, promoting the formation of stable nitrate ions on BaCO₃. This synergistic mechanism enables efficient storage and low-temperature regeneration.

Gupta and the co-workers developed a low-cost, environmentally friendly sodium manganese oxide (Na_xMnO₂, abbreviated as NMO) material [14]. This material exhibits outstanding chemical adsorption performance for acidic gases at room temperature under humid conditions, with an adsorption capacity of 0.53 mmol/g for NO₂. It can be regenerated by soaking in a NaOH-H₂O₂ mixed solution for 8 h. Characterization via XRD, XPS, FTIR, and other techniques confirmed that adsorbed NO₂ is converted into NO₂⁻ and NO₃⁻. The adsorption process relies on Mn³⁺/Mn⁴⁺/Mn²⁺ cycling and the synergistic interaction between lattice oxygen and surface hydroxyl (-OH) groups. One year later, orthorhombic Na_0.4MnO₂ microrods were synthesized via solid-phase calcination (using sodium acetate and manganese acetate precursors, calcined at 800 °C for 12 h) [36]. With a specific surface area of 7.8 m²/g, it exhibited outstanding chemical adsorption performance toward 100 ppm NO₂ under humid conditions at room temperature, achieving a maximum adsorption capacity of 0.92 mmol/g, whereas under dry conditions, the capacity was only 0.62 mmol/g. XPS and DFT calculations confirmed that NO₂ is converted into 43.7% nitrite (404.1 eV) and 56.3% nitrate (407.1 eV), with the adsorption process relying on surface hydroxyl neutralization and Mn²⁺/Mn³⁺/Mn⁴⁺ cycling. DFT calculations further validated strong interactions between gas molecules and Mn-O/Na-O sites on the Na_0.4MnO₂ surface (with adsorption energies as high as −4.03 eV). This study provides theoretical and experimental support for the room-temperature removal of low-concentration acidic gases. Gupta further utilized black powder extracted from discarded Zn-Mn alkaline batteries (primarily containing ZnO, MnO₂, ZnMn₂O₄, and other phases) as an adsorbent, achieving efficient removal of 100 ppm SO₂ and NO₂ at 20 °C and 80% relative humidity [37]. At room temperature 20 °C and 80% relative humidity, the optimal material (Ene_R from Energizer batteries) exhibited an adsorption capacity of 0.26 mmol/g for 100 ppm NO₂, and 0.22 mmol/g when dry. Regeneration of the saturated adsorbent using a NaOH-H₂O₂ mixed solution maintained stable NO₂ adsorption capacity after three adsorption-desorption cycles, with an average adsorption capacity of 0.30 mmol/g, demonstrating its potential as a renewable adsorbent.

Toso and colleagues investigated key performance parameters of 1 wt% Pd/CeO₂ materials as passive NO_x adsorbents [38]. Samples with distinct structural characteristics were synthesized via different methods including isovolumetric impregnation, dry grinding, and solution combustion. The study revealed that specific surface area is the core factor determining NO_x storage capacity (exhibiting a linear positive correlation, ranging from 0.017 to 0.387 mmol/g), while pore structure (pore size, pore volume) and carrier morphology showed no significant influence. Pd loading mode dominates NO_x desorption kinetics rather than specific surface area. Samples synthesized via dry grinding and initial wet impregnation exhibited strong Pd-Ce interactions (Pd-O-Ce structure), enabling over 60% of NO_x desorption above 523 K. In contrast, impregnation-synthesized samples showed higher desorption rates between 373–523 K, exceeding 40%. As shown in Figure 1, NO desorption is significantly lower on the milled sample, and NO₂ becomes dominant above 523 K. In situ DRIFTS confirmed that strong Pd-Ce interactions promote the oxidation of nitrite (1170–1310 cm⁻¹) to more stable nitrate (1500–1590 cm⁻¹). This study provides a clear direction for structural optimization of Pd/CeO₂ PNA: high specific surface area enhances storage capacity, while strong Pd-Ce interactions optimize the desorption temperature window.

Yeji’s team employed a hydrothermal method using urea as a precipitant to directly synthesize Co-Mg-Al layered double hydroxides (LDHs) on spherical γ-Al₂O₃ surfaces, investigating changes in adsorption performance [39]. At 150 °C under 400 ppm NO + 10% O₂ (N₂ equilibrium), NSE reached 85% (initial 10 min) with a storage capacity of 0.218 mmol/g. In the presence of 10% H₂O, storage capacity decreased by only 12%, demonstrating excellent water resistance. During thermal desorption, all samples exhibited NO_x desorption peaks below 400 °C (meeting PNA requirements). MMO/Al₂O₃-1U showed two desorption peaks: 250 °C corresponding to nitrite decomposition and 380 °C corresponding to nitrate decomposition. The high-temperature peak area accounted for 65%, indicating increased formation of stable nitrates. After five cycles, MMO/Al₂O₃-1U retained 79% of its initial storage capacity, whereas the conventional Ba-based material retained only 57%. This result provides novel insights for NO_x adsorption.

Table 3. Adsorption capacity of NO₂ by different metal oxides.

Metal Oxides	Weight Fraction (wt%)	Adsorption Test Conditions	NO2 Adsorption Capacity (mmol/g)	Refs.
Pt/Al2O3	1wt% Pt	300 ppm NO, 5% O2, Ar, 0.12 L/min, 393 K	0.16	[33]
Pt/La-Al2O3	1wt% Pt+1wt% La	300 ppm NO, 5% O2, Ar, 0.12 L/min, 393 K	0.11
NaxMnO2	/	100–500 ppm NO2, N2, 0.1 L/min, 298 K	0.53	[14]
Na0.4MnO2	/	100 ppm NO2, wet air, 0.2 L/min, 298 K	0.92	[36]
Na0.4MnO2	/	100 ppm NO2, air, 0.2 L/min, 298 K	0.62
ZnO+MnO2+ZnMn2O4+Mn3O4	/	100 ppm NO2, air, 0.2 L/min, 293 K	0.22	[37]
ZnO+MnO2+ZnMn2O4+Mn3O4	/	100 ppm NO2, 80%RH air, 0.2 L/min, 293 K	0.26
Pd/CeO2	1wt% Pd	100 ppm NO, 100 ppm NO2, 8% O2, 10% CO2, 5% H2O, N2, 0.4 L/min, 373 K	0.39	[38]

Table 3. Adsorption capacity of NO₂ by different metal oxides.

Metal Oxides	Weight Fraction (wt%)	Adsorption Test Conditions	NO2 Adsorption Capacity (mmol/g)	Refs.
Pt/Al2O3	1wt% Pt	300 ppm NO, 5% O2, Ar, 0.12 L/min, 393 K	0.16	[33]
Pt/La-Al2O3	1wt% Pt+1wt% La	300 ppm NO, 5% O2, Ar, 0.12 L/min, 393 K	0.11
NaxMnO2	/	100–500 ppm NO2, N2, 0.1 L/min, 298 K	0.53	[14]
Na0.4MnO2	/	100 ppm NO2, wet air, 0.2 L/min, 298 K	0.92	[36]
Na0.4MnO2	/	100 ppm NO2, air, 0.2 L/min, 298 K	0.62
ZnO+MnO2+ZnMn2O4+Mn3O4	/	100 ppm NO2, air, 0.2 L/min, 293 K	0.22	[37]
ZnO+MnO2+ZnMn2O4+Mn3O4	/	100 ppm NO2, 80%RH air, 0.2 L/min, 293 K	0.26
Pd/CeO2	1wt% Pd	100 ppm NO, 100 ppm NO2, 8% O2, 10% CO2, 5% H2O, N2, 0.4 L/min, 373 K	0.39	[38]

3. Application Research on the NO₂ Purification Process Using the Adsorption Method

NO exhibits low chemical reactivity and is difficult to capture directly and effectively by most adsorbents. Conversely, NO₂ is more readily physically adsorbed or chemically fixed due to its stronger polarity and reactivity. Furthermore, liquid NO₂ is a high-value chemical product (nitric acid production and rocket propellants) with a boiling point of 21.2 °C, whereas NO possesses comparatively lower commercial value. Consequently, in adsorption recovery processes, catalytic oxidation of NO to NO₂ is typically prioritized. This approach not only enhances overall removal efficiency but also enables high-value resource recovery. NO₂ common adsorption processes include Pressure Swing Adsorption (PSA), Temperature Swing Adsorption (TSA), and Vacuum Pressure Swing Adsorption (VPSA). PSA achieves separation through periodic pressure changes with relatively short cycles; TSA relies on temperature regulation for adsorption and desorption, with heating and cooling taking longer; VPSA incorporates a vacuum environment to reduce desorption pressure and enhance efficiency.

3.1. Pressure Swing Adsorption

Pressure Swing Adsorption (PSA) is a physical adsorption separation technology that utilizes the differences in equilibrium adsorption capacity or adsorption kinetics of various components within a mixed gas, based on porous solid adsorbents, and achieves separation through periodic pressure changes. Its core mechanism is founded on the thermodynamic principle that the adsorption phase equilibrium undergoes significant changes with system pressure, combining the selective separation function of the adsorbent with a cyclic pressure regulation operating mode.

A typical PSA unit consists of two or more adsorption towers, which perform strictly sequenced cyclic operations through program-controlled valves. A basic PSA process cycle includes the following stages:

• Adsorption: Under elevated pressure, feed gas is introduced into the adsorption bed. Strongly adsorbed (or rapidly diffusing) components are selectively captured, while weakly adsorbed components elute as product. An adsorption front forms within the bed and advances over time.
• Pressure Reduction: Upon adsorption saturation, the bed pressure is reduced to a lower level through pressure equalization recovery or direct venting. Partially adsorbed components are desorbed. This stage may incorporate co-current venting, counter-current venting, or multi-stage pressure equalization steps to recover energy.
• Desorption/Regeneration: Under low pressure, the partial pressure of the adsorbed components is further reduced through countercurrent flushing or vacuum extraction, enabling complete desorption and restoring the adsorbent’s adsorption capacity.
• Pressure Build-up: The bed pressure is raised to adsorption pressure using feed gas or product gas to prepare for the next cycle. In multi-bed systems, pressure is often initially increased using vent gas from other beds to enhance energy efficiency.

A. Sultana and colleagues employed NaY zeolite in 2004 to separate NO_x-containing gases with SO₂ using pressure swing adsorption (PSA) [40]. This study systematically elucidates the microscopic mechanism of NaY zeolite PSA for NO_x removal. Through engineering parameter optimization, it successfully establishes an operational window that effectively resists SO₂ poisoning. The core approach involves leveraging a rapid reversible process dominated by physical adsorption of N₂O₃ at elevated temperatures. By controlling oxygen concentration and shortening contact time, irreversible chemical adsorption and oxidation of SO₂ are suppressed, enabling efficient NO_x adsorption and material regeneration. This study simulated lean-burn exhaust gas using 500 ppm NO, 500 ppm NO₂, 200 ppm SO₂, 5% H₂O, 5% CO₂, 10% O₂, and He balance. It compared operating conditions at various temperatures (215 °C, 255 °C) and cycle durations (adsorption/desorption cycles of 30 min/30 min, 10 min/10 min, 20 min/15 min, and 45 min/30 min). Through experimental comparison, process parameters were determined as shown in

Table 4. Process parameters for PSA separation of NO_x resistant to SO₂ poisoning [40].

Conditions	Temperature (℃)	Adsorption/Desorption Cycle (min)	Adsorption/Desorption Space Velocity (h−1)	Regenerated Gas
Parameters	255	10/10	30,000/15,000	5%H2O+ 10%O2+ He

.

Under the aforementioned process conditions, NaY zeolite exhibits outstanding resistance to sulfur poisoning. After cumulative exposure to a dose of 360 mg/g SO₂, its NO_x adsorption capacity remains stable at approximately 6.0 mg/g, comparable to its performance in the absence of SO₂. The coexistence of SO₂ competes with NO_x for adsorption sites and reaction-required water molecules by forming species such as bisulfite.

PSA process demonstrates potential for flexible operation and relatively low energy consumption in treating low-concentration, complex-component NO₂ waste gases due to its rapid adsorption/desorption characteristics based on pressure cycling. This section demonstrates that precisely controlling operating temperature, cycle duration, and regeneration gas composition can effectively optimize the performance of adsorbents like zeolites in PSA processes. Particularly in complex flue gases containing competitive gases such as SO₂, process parameter design can suppress adsorbent poisoning to some extent and maintain selective NO₂ adsorption capacity. However, traditional PSA processes still face challenges due to limited desorption driving forces. To enhance desorption efficiency and product recovery concentration, further reducing the partial pressure of the adsorbate beyond pressure reduction has become a key direction for process intensification. This directly leads to vacuum pressure swing adsorption (VPSA) technology, which integrates a vacuum desorption step into the PSA process.

3.2. Vacuum Pressure Swing Adsorption

Vacuum Pressure Swing Adsorption (VPSA) is a gas separation and purification process that builds upon Pressure Swing Adsorption (PSA) by incorporating a vacuum regeneration step to significantly enhance separation efficiency and product recovery rates. Compared to PSA, the distinction in VPSA lies solely in the desorption cycle phase, where the pressure within the desorption tower is reduced to vacuum. This enables gases adsorbed by the adsorbent to be released more thoroughly under vacuum conditions, achieving highly efficient separation applications.

The Rokkasho Reprocessing Plant in Japan employed VPSA technology in 2003 to recycle NO_x generated at the facility [41]. The plant’s exhaust gas composition comprised 8.9% NO_x, 24.8% H₂O, 4.8 ppb RuO₄, with air making up the balance. The NO_x was recovered in the form of NO₂. The pilot plant operated for 100 days without failure [12]. At 338 K, the denitrification unit achieved over 95% recovery of nitrogen oxides from the exhaust gas, with a processing capacity of 500 kg/d. The recovered NO_x exhibited a purity exceeding 99.8% (with less than 0.2% H₂O). The process flow diagram for the nitrogen oxide unit is shown in Figure 2.

The parameters of the adsorbent used in this process are shown in

Table 5. Adsorbents Used in Rokkasho Nuclear Fuel Reprocessing Plant [12].

Route	Adsorbents	Specifications
Remove RuO4	USY200	Si/Al ratio = 200
Desiccation	Silica-gel + H-type Mordenite	Si/Al ratio = 20
NOx Recovery	Pentasil (3000) zeolite	Si/Al ratio = 3000

. For the ruthenium tetroxide adsorbent, USY200 was selected due to its strong hydrophobicity and excellent acid resistance. It exhibits the highest RuO₄ retention rate, with a breakthrough of only 2%, a decontamination factor >5000, and an adsorption capacity matched to the RuO₄ half-life (1 year), eliminating the need for frequent replacement. For the water adsorbents, silica-gel underwent preliminary dehydration, while synthetic mordenite zeolite underwent deep drying. The final product contained <0.2% H₂O. The synthetic mordenite zeolite demonstrated high stability and acid tolerance. For NO_x adsorbents, Pentasil (3000) zeolite was selected for its high NO₂ adsorption selectivity, balanced hydrophobicity and acid resistance, and suitability for VPSA cycling conditions.

Table 6. Specific process parameters for VPSA NO_x recovery at Rokkasho Fuel Reprocessing Plant [12].

Chemical Engineering Device	Adsorption/Desorption Pressure (kPa)	Temperature (K)
Ruthenium removal Tower	120	343
Desiccation tower	120/60	343
NOx recovery tower	120/5	298
Cooler	110	260

lists the parameters for several adsorption towers in the VPSA process. In this study, NO_x gas reacts with condensate to form HNO₃, which corrodes pipelines and equipment. Therefore, SUS 304 stainless steel is used for the adsorption towers, making it suitable for this scenario. The adsorbents loaded are all monolithic adsorbents, replacing traditional spherical adsorbents to reduce the risk of adsorbent fragmentation clogging pipelines. The operating temperature is 343 K, maintained above the gas dew point of 338 K to prevent condensation of HNO₃. During desorption, dry air is used for countercurrent purging to rapidly remove moisture or NO₂ from the adsorbate phase, minimizing acid residue.

During this process, the adsorbent can be recycled. The vacuum pressure swing adsorption process flow for NO_x treatment at Rokkasho Reprocessing Plant is shown in Figure 3.

Step 1: Drying the exhaust gas using acid-resistant adsorbents, with no significant loss of NO_x. Given the overall humid atmosphere containing NO_x, acid formation is highly likely. To prevent degradation of the adsorbent’s activity, the required adsorbent must possess strong acid resistance. The mixed atmosphere generated at this plant has a dew point of 338K. To minimize acid’s impact on adsorbent activity, adsorption is conducted at 343K. This adsorption process primarily dries the gas by adsorbing water vapor, with a small amount of NO_x also adsorbed. It is essential to maintain the NO_x concentration above a certain threshold to ensure minimal loss of this target collection material.

Step 2: Concentrate NO_x in the dry exhaust gas using a suitable adsorbent. Since atmospheric NO rapidly converts to NO₂, the NO_x in this process is primarily NO₂. Given the presence of moisture in the system, the adsorbent must be acid-resistant. This step adsorbs a significant amount of NO_x, facilitating the subsequent cooling and liquefaction process to enhance NO_x yield and improve purity.

Step 3: The desorbed gas containing high concentrations of NO₂ is sent to the condenser. Since NO₂ boils at 21.2 °C, it is recovered as a liquid product at this stage. Over 95% of the NO₂ in the feed gas is recovered as a liquid product.

During desorption, over half of the NO₂ is liquefied in the cooler, while the unliquefied NO₂ is returned to the denitrification tower for recovery. The volume fraction of lost NO₂ is <1%, achieving a NO_x recovery rate exceeding 95%. Regarding dynamic adsorption capacity, PSA typically achieves only a fraction of the capacity of TSA. This characteristic enables significant downsizing of PSA units for equivalent treatment requirements, effectively conserving space and reducing capital investment. However, current research remains confined to laboratory scale, employing simulated gases under non-radioactive conditions, with neither scale nor parameters meeting industrial application standards. For nuclear reprocessing operations, process variations result in significant differences in NO_x tail gas composition and concentration. Consequently, adaptive optimization of this PSA process based on actual operating conditions is urgently needed.

3.3. Temperature Swing Adsorption

Temperature Swing Adsorption (TSA) is a cyclic adsorption process that utilizes the characteristic of adsorbents where their adsorption capacity for gas components changes significantly with temperature. By periodically altering the temperature of the adsorption bed, it achieves gas separation and purification. Its thermodynamic foundation lies in the inherently exothermic nature of the physical adsorption process of gases, where the adsorption equilibrium constant decreases as temperature increases. TSA is typically configured with dual or multiple beds to enable continuous processing. A complete cycle comprises the following core steps:

• Adsorption: The pretreated feed gas passes through the adsorption bed at ambient or lower temperatures. Strongly adsorbable impurities (such as water, organic solvents, and high-boiling-point components) are selectively retained, with the purified gas discharged as product. This stage constitutes an isothermal or near-isothermal exothermic process, requiring careful consideration of the bed temperature rise’s impact on adsorption performance.
• Heating Desorption: When the adsorption front approaches the bed outlet, the process switches. High-temperature regeneration gas (such as hot nitrogen, steam, or superheated air) is introduced into the saturated adsorption bed. Heat is transferred to the adsorbent, providing the energy required for desorption. This imparts sufficient kinetic energy to the impurity molecules to dislodge them from the adsorption sites. The desorbed impurities are carried out of the bed by the regeneration gas stream.
• Cooling: After thermal regeneration, the bed temperature remains significantly higher than the adsorption temperature. Clean gas at ambient or low temperatures must be introduced to cool the bed down to the required adsorption temperature. This cooling process also serves to purge residual desorbed components.
• Pressure Adjustment and Standby: Following cooling, pressure adjustment may be necessary to match adsorption conditions. The bed then enters standby mode, awaiting transition to the next adsorption cycle.

Zhiqiang Wang and colleagues developed a closed-loop deep purification process for high-concentration, highly volatile NO_x tail gas (270–2700 ppm, the concentration ratio of NO to NO₂ is approximately 9:1) emitted from nuclear facilities [42]. This process centers on ZSM-5 zeolite temperature-swing adsorption technology coupled with urea absorption pretreatment, and its engineering feasibility was validated through a 200 Nm³/h pilot plant. The study proposed an innovative adsorption-concentration-return-reabsorption closed-loop process, as illustrated in Figure 4.

• Pre-treatment: High-concentration, fluctuating raw tail gas undergoes primary denitrification in a urea absorption tower with >90% efficiency, significantly reducing inlet concentration.
• Deep purification: Post-urea-treated tail gas, still at elevated concentration, undergoes cooling and deep drying using 3A molecular sieves before entering a ZSM-5 zeolite adsorption tower for deep adsorption purification.
• Regeneration and Closed-Loop: After adsorption saturation, the ZSM-5 bed undergoes optimized variable-temperature desorption at 200 °C with system pressure reduced to −50 kPa, yielding high-concentration NO_x desorption gas (>40,000 mg/m³).

Material accounting indicates that the recycled desorbed gas is mainly NO₂, accounting for only 2% of the total original NO_x. After a 24-h stable operation test in a 200 Nm³/h pilot-scale system, the breakthrough time reached 12.5 h, demonstrating the adsorption bed’s excellent pollutant-holding capacity and long-term operational stability. The NO_x concentration at the outlet of the adsorption tower remained stably ≤ 10 ppm, with a purification efficiency > 99.88% and a desorption efficiency ≥ 95%. The dynamic adsorption capacity of ZSM-5 zeolite reached as high as 0.36 mmol/g, and after 16 adsorption-desorption cycles, the adsorption capacity decay rate was <2%, entering a stable phase. However, for the pilot-scale stability test, the 24-h duration is excessively short, and the feed gas was simulated tail gas after urea absorption, which failed to fully replicate the complexity and harshness of real nuclear tail gas. Although 16 cycles can illustrate short-term stability, key engineering issues such as the long-term decay trend of adsorbent performance, adsorbent pulverization caused by high-temperature regeneration, and the increase in bed pressure drop have not been verified for nuclear industry devices that require continuous operation for months or even years. This verification experiment was conducted in a non-radioactive environment, and the irradiation effects and contamination of the adsorbent by trace radioactive aerosols, iodine isotopes, etc., which may be present in real nuclear tail gas, were not mentioned—these are also urgent research gaps that need to be addressed in domestic and international studies.

TSA process exhibits unique advantages in treating NO_x in exhaust gas due to its specific operational mode. However, during the temperature swing adsorption process, the heating and cooling of the adsorbent bed are slow, so a complete TSA cycle may take several hours. Additionally, repeated heating and cooling cycles may lead to adsorbent aging and deactivation. These factors have somewhat limited the widespread industrial application of temperature swing adsorption.

3.4. Temperature-Pressure Synergistic Strategy

Liu’s team proposed a novel temperature-and-pressure swing cyclic adsorption process, namely Vacuum Temperature Swing Adsorption using multiple Gas circulations (GVTSA), which integrates dual regulation of temperature and pressure to optimize the efficiency and performance of the adsorption process [43]. Employing self-prepared granular Na-ZSM-5 zeolite as the adsorbent, the authors conducted adsorption-recovery experiments and scale-up studies of NO_x from sintering flue gas. The core innovation of the GVTSA process lies in the combination of vacuum temperature swing adsorption (VTSA) and closed hot gas circulation technology, coupled with the innovative adoption of a segmented gas chromatography (GC) analytical strategy. Field tests were performed using dual-tower setups at two scales (small tower, ST; large tower, LT) to validate the process feasibility. Based on the experimental results, a conceptual triple-tower process design for continuous production was further proposed, laying a foundation for the industrial application of the GVTSA technology in NO_x capture and recovery from sintering flue gas. The dual-tower GVTSA system incorporates controlled multiple hot gas recirculation cycles supplemented by purging steps, as shown in Figure 5.

• Adsorption: Flue gas passes through the adsorption tower, where Na-ZSM-5 catalyzes the oxidation of NO to NO₂ and adsorbs it.
• Heating: Close the tower inlet and outlet, activate the closed-loop system, and heat the circulating gas via the heat exchanger to raise the bed temperature to the target desorption temperature of 493K.
• Gas Circulation: During the first gas circulation (GC), the thermal circulation gas flows continuously while desorption occurs at a constant standard pressure of −50 kPa within the system. When the NO₂ concentration in the desorption gas exceeds 2%, the desorbed NO₂ gas passes through a buffer tank and heat exchanger for temperature adjustment. The high-concentration NO₂ exceeding 2% is then recovered as desorption gas. For the nth GC, fresh NO_x-free gas from another tower in the adsorption phase is introduced to initiate a new cycle, further driving deep desorption.
• Purge: Slowly purge the bed with a small volume of dry, cold purified gas to cool and carry out residual high-concentration NO_x gas for recovery.
• Purification: Rapidly purge with a large volume of purified gas to cool the bed to the adsorption temperature of 308 K. The purge gas is directly vented.

In the dual-tower GVTSA process research, the self-developed NaZSM-5 zeolite exhibited a saturated adsorption capacity as high as 1.496 mmol/g, with a working adsorption capacity stabilizing at approximately 0.10 mmol/g, meeting the emission standard of ≤50 mg/m³. NO_x recovery exceeded 92%, with NO₂ concentrations enriched from approximately 20 ppm at the inlet to an average of >2.5%. This represents an enrichment factor exceeding 1000-fold, meeting feed requirements for subsequent low-temperature condensation recovery processes. After 16 consecutive adsorption-desorption cycles, the adsorbent’s working capacity remained stable, demonstrating the process and adsorbent’s excellent cycling and regeneration capabilities. The dual-tower GVTSA process employs a closed-loop system to prevent significant NO₂ dilution caused by hot air/steam purging in conventional TSA systems. It maintains desorption drive through multiple fresh gas injections, achieving high desorption efficiency and high NO₂ concentration enrichment. However, the dual-tower GVTSA process inherently operates intermittently—one tower adsorbing while the other desorbs—preventing continuous operation. To meet the demand for treating high-flow, continuously emitted sintering flue gas from steel mills, a triple-tower GVTSA process was designed to enable continuous NO₂ output.

The key steps in NO_x adsorption using the triple-tower GVTSA process are as follows:

• Adsorption Stage: Flue gas (308 K) treated for desulfurization and dehydration (RH < 1%) enters the denitrification tower. Na-ZSM-5 zeolite catalyzes the oxidation of NO to NO₂ and selectively adsorbs it, maintaining outlet NO₂ concentration below ultra-low emission standards (≤25 ppm) for approximately 14 h.
• Desorption Stage: Includes heating (HT) and multi-stage gas circulation (GC), with the core being enhanced NO₂ desorption through closed-loop circulation.
• Heating (HT, 3 h): Gas circulating within the closed loop is heated via a heat exchanger to 493 K. This preheats the bed (HT1, 2 h) before maintaining temperature for initial NO₂ enrichment (HT2, 1 h).
• Gas Circulation (GC, 3 h): Comprises three 60-min cycles (G₁–G₃). Circulating gas is drawn from the exhaust of another denitrification tower in adsorption phase (NO_x-free), supplemented into the closed loop for continuous circulation. This disrupts NO_x adsorption equilibrium, promoting desorption of strongly adsorbed NO₂ from zeolite into the gas phase. Part of the desorbed NO₂ enrichment gas is diverted to the condensation recovery system, with the remainder recycled.
• Cooling stage: Includes purging (S) and purification (P), balancing bed cooling with residual NO₂ recovery.
• Purge (S, 3 h): Low-flow (1/5 feed gas volume) clean gas is introduced into the bed to gradually cool it and purge residual NO₃. Enriched gas is diverted for recovery (S1 phase), while subsequent makeup gas is used for desorption in the next tower (S2 phase).
• Purification (P, 3 h): High-flow (≥1/2 feed gas volume) clean gas rapidly cools the bed to adsorption temperature (308 K). Tail gas is directly vented, and the bed enters the next adsorption cycle.

The triple-tower GVTSA process flow diagram is shown in Figure 6. Through sequential switching between three denitrification towers (A/B/C), the system simultaneously completes three stages: adsorption (AD) → heating/circulation (HT1/HT2/GC) → purging/purification (S1/S2/P), achieving continuous NO_x removal and recovery. During the adsorption phase (AD), dehydrated flue gas enters the corresponding tower (Tower A), where the adsorbent captures NO_x. The purified dry tail gas is then sent to the dehydration tower for reuse. During the desorption phase (HT1/HT2/GC), the saturated tower (Tower C) switches to the green pathway shown in the diagram. Driven by a fan, the gas is heated via heat exchangers (HT1/HT2) while a vacuum pump maintains vacuum conditions. Gas recirculation (GC) facilitates NO_x desorption and enrichment. Finally, during the regeneration phase (S1/S2/P), the desorbed tower (Tower B) switches to the blue pathway. Purge gas cools and pressurizes the tower, completing tower cooling and adsorbent regeneration to prepare for the next adsorption cycle.

The optimized GVTSA process recovers NO_x primarily as NO₂, achieving a recovery rate of up to 92%. From the feed gas’s original 200 ppm NO_x concentration, approximately 2% NO₂ is enriched, meeting the requirements for low-temperature condensation recovery. The Na-ZSM-5 zeolite catalyzes the oxidation of NO to NO₂, achieving a NO₂ to NO ratio of 9:1 in the product stream. With 88.9% of NO converted, this process resolves the challenges of NO’s poor adsorption and recovery. Designed for steel sintering flue gas, a single triple-tower system can process 3 × 10⁴ m³/h of flue gas, recovering approximately 86.25 tons of NO₂ annually. The entire system consumes approximately 600 kW of energy.

Although this study proposes a triple-tower optimization strategy, the long-term operational data for the triple-tower process is limited to performance degradation testing over 16 cycles. The equipment scaling effect has not been fully demonstrated, and further validation is required for industrial implementation. Furthermore, this study focuses on iron ore sintering flue gas treatment, where NO_x concentrations are approximately 200 ppm. However, NO_x concentrations in nuclear fuel reprocessing tail gases can reach thousands or even tens of thousands of ppm. The desorption efficiency under higher concentration conditions, the stability of NO₂ enrichment concentrations, and the stability of the adsorbent under radioactive conditions have not been investigated. Therefore, further experimental design and research are particularly important.

The advantages and disadvantages of the four NO₂ purification processes are shown in

Table 7. Comparison of Primary NO₂ Adsorption Processes.

Process	Advantages	Disadvantages
PSA	High cycle rate and automation level	The desorption driving force is limited
VPSA	Complete desorption, high product purity, suitable for complex mixtures	Long cycle times and potential thermal aging of adsorbents
TSA	More thorough than PSA desorption with higher recovery rates	Requires a vacuum system, increasing equipment investment and energy consumption
GVTSA	High desorption efficiency	Complex manufacturing processes, challenging control requirements, and high equipment investment costs

. In this study, process selection requires comprehensive consideration of exhaust gas characteristics (such as concentration, flow rate, and composition), product purity and concentration requirements, site conditions, and investment and operating costs. For demanding scenarios like nuclear reprocessing, GVTSA demonstrates significant potential due to its high-efficiency concentration and interference resistance. For lower-concentration, smaller-scale applications, optimized VPSA or TSA may offer greater economic viability. The selection of process technology requires comprehensive consideration of exhaust gas characteristics (such as concentration, flow rate, and composition), product purity and concentration requirements, site conditions, and investment and operational costs. For demanding scenarios like nuclear reprocessing, GVTSA demonstrates significant potential due to its high-efficiency concentration and interference resistance. For lower-concentration, smaller-scale applications, optimized VPSA or TSA may offer greater economic viability.

4. Conclusions and Outlook

This study systematically reviews research progress in adsorption methods for NO₂ purification. Compared to traditional technologies such as liquid absorption, SCR/SNCR, biological methods, and plasma methods, adsorption offers significant advantages in NO₂ resource recovery, atmospheric pressure and low-temperature operation, and minimal secondary pollution. It is suitable for scenarios requiring NO₂ reuse, such as the nitric acid industry and nuclear fuel reprocessing. Regarding adsorbent materials, zeolites demonstrate outstanding performance in selective NO₂ adsorption due to their tunable pore structures and surface acidic sites; metal modification can further enhance their adsorption capacity and hydrothermal stability. Activated carbon and biochar exhibit promising potential for low-temperature NO₂ adsorption based on their high specific surface area and diverse surface functional groups; alkali modification and metal loading can improve their performance. Metal oxides achieve efficient NO₂ adsorption through surface alkaline sites, oxygen vacancies, and redox properties. At the process level, PSA and TSA are the current mainstream technologies. Engineering examples of VPSA in Rokkasho Reprocessing Plant and TSA in China have validated the feasibility of adsorption methods for industrial-scale NO_x recovery. The novel GVTSA process enhances desorption efficiency and regenerative performance by intensifying the desorption concentration gradient through multi-stage gas cycling, offering a new pathway for process optimization.

Although significant progress has been made in NO₂ purification via adsorption, high-concentration and high-volume NO₂ tail gases still pose challenges, such as the need for large quantities of adsorbent, which increases equipment investment and operating costs. Meanwhile, energy optimization during adsorbent regeneration and the efficient recovery of desorbed NO₂ have yet to yield economically viable integrated solutions. In addition, NO₂ in nuclear fuel reprocessing tail gases contains complex components such as radioactive impurities and corrosive gases, placing extremely high demands on adsorbent radiation resistance and acid tolerance. Currently, research in this field remains extremely limited, with a lack of specialized adsorbent materials and process systems tailored for this specific scenario. Therefore, future research on NO₂ purification via adsorption should focus on material innovation and process optimization. Efforts should be strengthened in the development of large-scale adsorbent production technologies, implementation of industrial demonstration projects, collection of process operation data in nuclear industrial scenarios, optimization of equipment design and operational parameters, and promotion of adsorption technology from laboratory research to large-scale engineering practice, so as to provide more competitive technical solutions for NO₂ emission reduction and resource utilization in nuclear fuel reprocessing plants.