The Effect of the Metal Oxide as the Support for Silver Nanoparticles on the Catalytic Activity for Ammonia Ozonation

Abstract

Ammonia is one of the common inorganic pollutants in surface waters. It can come from a wide range of sources through the discharge of wastewater (industry, agriculture, and municipal waters). Catalytic ozonation reaction can efficiently remove ammonia nitrogen without introducing other pollutants and improve the nitrogen selectivity of reaction products by controlling the reaction conditions. Catalysts based on silver nanoparticles (Ag NPs) have shown excellent O₃ decomposition performance; therefore, they are promising catalysts for catalytic ammonia ozonation due to their high reactivity, stability, and selectivity to N₂. In this study, we synthesized well-defined silver nanoparticles (Ag NPs) using a modified alkaline polyol method and then dispersed them on solid oxide supports (Fe₃O₄, TiO₂, and WO₃). Before being deposited on the oxide support, the silver nanoparticles were characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), and UV-VIS spectroscopy. The obtained catalysts, Ag_Fe₃O₄, Ag_TiO₂, and Ag_WO₃ were characterized using X-ray diffraction (XRD), scanning electron microscopy (SEM), BET surface area analysis, UV-VIS spectroscopy, temperature-programmed reduction (H₂-TPR), and temperature-programmed desorption (TPD) of CO₂ and NH₃. It has been demonstrated that the nature of the support significantly influences the physicochemical properties of the catalysts, as well as their catalytic performance in ammonia ozonation reaction.

1. Introduction

Ammonia has become a significant concern as a major contaminant in wastewater due to its warming effects. Excessive ammonia levels in natural water bodies can cause eutrophication, threatening both human health and aquatic life. Biological treatments are often used for ammonia-contaminated wastewater because of their cost-effectiveness and high efficiency. However, these treatments are sensitive to fluctuations in pH, dissolved oxygen, temperature, and other water parameters [1,2,3].

For effective environmental protection, the maximum ammonia concentration should not exceed 0.5 mg L⁻¹ [4]. Chemical oxidation methods, using ozone and other oxidants, offer several advantages over biological approaches [5,6,7]. Ozone (O₃) is commonly used in water treatment for oxidation and disinfection. As a strong oxidizing agent, ozone selectively reacts with amino groups, leaving some contaminants untreated. Additionally, when ozone decomposes in water, it can produce free radicals, such as hydroxyl radicals (•OH), resulting in a quicker but less selective reaction pathway. To improve ozone decomposition and enhance advanced oxidation processes (AOPs), various techniques, such as adding chemical agents (H₂O₂), using light irradiation (like UV), or employing catalysts, are typically employed.

Catalytic ozonation, recognized as an advanced oxidation process, has gained considerable attention in recent years. In this process, the catalyst promotes ozone decomposition, generating active oxygen free radicals that increase the organic pollutants’ degradation and mineralization rates [8].

Ammonia exists in two primary forms: when water pH is below 9, ammonium (NH₄⁺) is the predominant form; when pH exceeds 9, free ammonia (NH₃) predominates. The non-ionized form is particularly toxic due to its lipid solubility [9,10]. Ammonium is commonly found in industrial wastewater and requires pre-treatment before being released into municipal water systems. Over the last decade, numerous ammonia removal methods have been researched, including biological treatment [11], biofiltration [12], air/steam stripping [13], breakpoint chlorination [14], chemical precipitation [15], ion exchange [16], and catalytic ozone oxidation [17,18].

Recent studies [17] have shown that catalytic ozone oxidation effectively removes NH₄⁺, particularly when using oxide-based catalytic systems composed of transition metals (MO_x, where M includes Co, Ni, Mn, Sn, Cu, Mg, and Al). Magnesium oxide (MgO) exhibits high catalytic activity but low selectivity for N₂ gas, while cobalt oxide (Co₃O₄) demonstrates good N₂ selectivity but lower activity. A notable catalyst for NH₄⁺ removal with selective nitrogen gas production was reported by Chen et al. [19], using a MgO/Co₃O₄ composite catalyst with a molar ratio of 2:8. Additionally, Nguyen et al. [20] achieved 96% ammonia removal and approximately 78% N₂ selectivity using Fe₂O₃-Co₃O₄ modified dolomite.

Noble metals supported on metal oxides present a promising alternative for the catalytic ozonation of ammonium-containing waters, as ammonium can be selectively oxidized into safe gaseous compounds [21,22]. Silver nanoparticles (Ag NPs) have attracted attention for their unique physical, chemical, optical, electrical, and biological properties. However, separating solid Ag NPs from colloidal solutions can be challenging once they are formed. To overcome this, Ag NPs have been extensively immobilized on solid supports, where they attach to substrates through hydrogen bonding or electrostatic interactions [23]. The metal oxide support can either be inert or catalytically active. When using catalytically active supports, the immobilization of silver nanoparticles creates a synergistic effect. For a catalyst to be effective in the ozonization reaction, its surface must be capable of exchanging chemical species with the reaction medium, allowing for the adsorption and dissociation of ozone, as well as the regeneration of chemical bonds. Modification of the catalytic surface can occur through the generation of Lewis acidic and basic sites, along with Brønsted sites, which may sometimes function cooperatively.

The objective of this work is to assess the influence of the support on the oxidation reaction of ammonium with ozone. Three catalysts, designated Ag_Fe₃O₄, Ag_TiO₂, and Ag_WO₃, were developed and tested in the catalytic ozonation of ammonia. Several characterization techniques, including X-ray diffraction (XRD), hydrogen temperature-programmed reduction (H₂-TPR), termoprogrammed desorption NH₃-TPD, and CO₂-TPD, were employed to evaluate the relationship between the catalyst’s chemical and physical properties and its activity, as well as the interaction between silver nanoparticles (Ag NPs) and the oxide support.

2. Results

2.1. Synthesis of the Catalysts

Ag nanoparticles were obtained by optimizing a modified polyol method described in our previous paper [24]. Therefore, the synthesis of silver nanoparticles was achieved by combining 50 mL of a 0.02 M AgNO₃ solution in ethylene glycol (EG) with equivalent volumes of a 0.2 M PVP/EG solution and a 0.15 M NaOH solution, resulting in a molar metal ratio of PVP to silver of 1:10. The synthesis was conducted under an argon atmosphere to ensure inert conditions. The nucleation and reduction in metallic precursors with ethylene glycol took place at a temperature of 130 °C for 60 min.

The most effective separation method for recovering over 95% of the metallic nanoparticles involved a double volume of acetone to the colloidal suspension and cooling the solution to −16 °C for 24 h. After this period, metal nanoparticles were washed several times with acetone to eliminate any remaining ethylene glycol. The resulting solid phase containing silver nanoparticles was then dried at 80 °C.

TiO₂ and WO₃ were purchased from Aerosil, Japan and Alfa Aesar, respectively, while the magnetite particles were synthesized using chemical co-precipitation, following the methodology outlined in the literature [25]. Specifically, 40 mL of 25% NH₄OH was added dropwise to an aqueous solution containing 3 g of FeCl₃ × 6H₂O and 1 g of FeCl₂ × 4H₂O dissolved in 200 mL of deionized water. This resulted in the immediate formation of a black magnetic powder. After heating the mixture to 80 °C for 1 h, the black magnetic powder was magnetically separated, thoroughly washed with deionized water, and then dried in an oven at 80 °C for 2 h.

The silver nanoparticles could be easily dispersed in ethanol and deposited onto oxide supports such as Fe₃O₄, TiO₂, and WO₃, with a metal loading of 1%. The nanoparticles and oxide support, dissolved in ethanol, were subjected to ultrasound treatment for 30 min. The catalyst suspension was then placed in a stirring bath and maintained at 85 °C for 2 h to facilitate solvent evaporation. The resulting catalyst was subsequently dried at 100 °C for 4 h.

Finally, calcination was performed at 350 °C, with a temperature increase of 10 °C per minute, for a total duration of 2 h to remove the organic polymer.

2.2. Characterisation of the Catalysts

2.2.1. Transmission Electron Microscopy (TEM) and Scanning Electron Microscopy (SEM)

Figure 1 presents the TEM image of silver nanoparticles and their size distribution.

As observed in Figure 1, it is clear that silver nanoparticles were obtained in the proposed synthesis. Figure 1a shows relatively uniform silver nanoparticles with most sizes ranging from 3 to 6 nm. This particle range is confirmed in the histogram from Figure 1b.

Figure 2 shows the SEM micrograph for the catalyst samples Ag_WO₃, Ag_TiO₂, and Ag_Fe₃O₄. As can be seen, the morphology of the catalysts is given by the morphology and structure of the oxide support. The SEM micrographs of Ag_WO₃ and Ag_TiO₂ (commercial oxides) show particles with uniform distribution, spherical morphology, and slight agglomeration, while the SEM micrograph of the Ag_Fe₃O₄ catalyst shows a higher degree of particle agglomeration.

2.2.2. X-Ray Diffraction (XRD)

Figure 3 shows the XRD spectra of the obtained catalysts. The observed diffraction lines are characteristic of each support used: Fe₃O₄, TiO₂, and WO₃. We can also observe the specific diffraction lines of Ag (ICDD-00-900-8459) at 2θ = 38.11; 44.30; 64.44; and 77.40.

For the silver NPs, the diffractograms present the characteristic diffraction lines of Ag at 2θ = 38.11 (111); 44.34 (200); 64.7 (220); and 77.40 (311), which is consistent with the standard card (ICDD-00-900-8459) [26] with crystalline cubic structure and lattice constants of a = b = c = 4.0904 nm).

The diffraction-specific lines for Ag_WO₃ are observed, which correspond to the WO₃ crystallite phase (ICDD-01-083-0950) [27]. Also, diffraction-specific lines of silver are observed at 38.11; 44.3; 64.7; and 77.4 2θ.

The XRD diffractogram of Ag_TiO₂ presents both the specific diffraction line for anatase (ICDD-00-021-1272) [28] and rutile (ICCD-00-021-1276) [29], along with the specific lines of silver.

In the case of Ag_Fe₃O₄, characteristic diffraction lines of magnetite (ICDD-00-019-0629) [30] are present, as well as specific diffraction lines of silver.

The unit cell parameters (a = b = c = 4.094 nm) of silver suffer a slight contraction when immobilized on the support. This contraction (

Table 1. Lattice parameter, crystallite size, and phase composition from XRD analysis.

Samples	Unit Cell Parameters (nm) Silver a = b = c	Crystallite Size (nm) *	Phase Composition **
			Ag%	Support%
Ag NPs	4.094	7.3	100	-
Ag_WO3	4.092	20.1	1.12	WO3-98.84
Ag_TiO2	4.082	5.8	0.98	Anatase-84.7 Rutile-11.7
Ag_Fe3O4	4.090	10.1	1.19	Fe3O4-98.81

* Crystallite size determined from XRD diffraction using the Williamson–Hall method. ** Quantitative Analysis determined using The Reference Intensity Ratio (RIR) method.

) can be attributed to the interaction between the silver nanoparticles and the oxide support. The crystallite size determined from XRD diffraction using the Williamson–Hall method varies due to this interaction. Quantitative analysis, conducted using the Reference Intensity Ratio (RIR) method (Table 1), confirms the presence of silver at approximately 1% in the obtained catalysts.

2.2.3. Temperature-Programmed Reduction (H₂-TPR)

The metal–support interaction and reducibility of the catalysts were investigated through H₂-TPR measurements. The interaction between the noble metal and the support can influence the chemical state of silver nanoparticles and the support material. The H₂-TPR profiles are presented in Figure 4 and

Table 2. The structural and optical characteristics of catalysts.

Catalyst	Specific Surface Area (m2/g)	H2 Consumed by TPR Analysis (µmole/g)	Band Gap by UV-Vis Spectra (eV)
Ag_WO3	13	1020	2.97
Ag_TiO2	51	56	3.16
Ag_Fe3O4	185	2690	1.97

. Also, the strong reducibility of the catalysts can create significant oxygen vacancies.

The reduction temperature of silver oxide, formed during calcination in air, shifts towards higher values depending on the type of support material. A higher reduction in temperature indicates a stronger interaction between silver nanoparticles and the support. Among the catalysts studied, the Ag_WO₃ catalyst exhibits the strongest interaction with its support.

For Ag_TiO₂, the reduction peaks at low temperatures (~135 °C) are attributed to the reduction in small Ag₂O clusters with weak interaction (Ag₂Ow). The highest temperature reduction peaks (~400 °C) correspond to Ag₂O particles that exhibit a stronger interaction with the TiO₂ support (Ag₂Os) [31,32].

For Ag/Fe₃O₄, the reduction peak observed at 170–400 °C is ascribed to the reduction in Ag₂O, which occurs simultaneously with the reduction in Fe₃O₄ to FeO. The peak at 600–700 °C corresponds to the further reduction in FeO to metallic Fe. Adding an Ag promoter favors the oxidization of Fe₃O₄ to FeO and facilitates the reduction of iron oxide to Fe [33,34].

In the case of Ag_WO₃, the large peaks that appear at 625 and 800 °C from the H₂-TPR profile are attributed to the reduction in the W⁶⁺ species W⁶⁺/W³⁺ and W³⁺/W²⁺, respectively [3]. The reduction in silver from Ag_WO₃ takes place starting from the temperature of 296 °C. The reduction temperature is higher than in the case of TiO₂ and Fe₃O₄ samples, which indicates a stronger interaction of the Ag NP with the support WO₃.

From the quantitative analysis of the TPR results for Ag_TiO₂ presented in Table 2, it can be seen that only 52% of the silver is in oxidized form, the rest being in metallic, Ag⁰ form. Also, the ratio Ag⁺/Ag⁰ (1.08) can be accurately estimated from H₂-TPR for the Ag_TiO₂ catalyst only, as silver is the only species that is reduced in the analyzed temperature range. The XPS analysis of Ag NPs confirms that silver exists in both metallic and oxidized form (Figures S1 and S2). A higher concentration of metallic Ag⁰ enhances ammonium oxidation [34].

2.2.4. Temperature-Programmed Desorption (CO₂₋TPD and NH₃-TPD)

The surface acid-base properties of Ag_Fe₃O₄, Ag_TiO₂, and Ag_WO₃ were evaluated using CO₂-TPD and NH₃-TPD. The normalized desorption curves are shown in Figure 5a,b, while detailed information on the types and quantities of basic and acid sites is provided in

Table 3. Distribution of the strength of acidity from NH₃-TPD and basicity from CO₂-TPD.

Catalyst	Acid Strength Distribution %			Total Acidity (mmole/g)	Base Strength Distribution %			Total Basicity (mmole/g)
	<200 °C Weak	200–400 °C Medium	>400 °C Strong		<250 °C Weak	250–450 °C Medium	>450 °C Strong
Ag_WO3	11.79	37.38	50.83	0.125	26.97	43.88	29.15	0.949
Ag_TiO2	5.03	11.65	83.32	0.064	19.62	42.91	37.47	0.893
Ag_Fe3O4	3.14	86.32	10.54	0.410	8.35	29.62	62.03	0.417

.

The temperature of CO₂ desorption is specific for identifying three types of basic species: weak basic sites (hydroxyl groups remaining on the oxidic surface after calcination, or Brønsted basic sites), medium-strength Lewis basic sites, and strong Lewis basic sites (incompletely coordinated O²− ions) [35]. Thus, the desorption peaks below 400 °C correspond to weak and medium basic sites, while peaks above 400 °C indicate stronger Lewis basic sites.

For Ag_WO₃ and Ag_TiO₂, optimal CO₂ adsorption occurs at weak and medium basic sites within the temperature range of 100–400 °C. Desorption from strong basic sites occurs at higher temperatures (400–600 °C). In contrast, Ag_Fe₃O₄ shows a significantly lower density of basic sites compared to Ag_WO₃ and Ag_TiO₂.

Figure 5b illustrates the acidity profiles of the catalytic systems. The acidity of Ag_Fe₃O₄, Ag_TiO₂, and Ag_WO₃ plays a crucial role in ammonia adsorption/desorption, which involves N–H bond dissociation followed by nitrogen and nitrate desorption from the catalyst surface. Desorption peaks observed up to 400 °C provide insights into their catalytic behavior. Peaks between 90 and 260 °C are attributed to physically adsorbed NH₃ and weakly adsorbed NH₄⁺ on Brønsted acid sites. These peaks reflect the release of NH₄⁺ from Brønsted acid sites and strongly adsorbed NH₃ from Lewis acid sites [36,37,38]. The desorption peak located between 260 and 420 °C corresponds to NH₄⁺ adsorbed on Brønsted acid sites and NH₃ coordinated to Lewis acid sites. The peaks at 420–650 °C are attributed to strongly adsorbed NH₃ on Lewis acid sites. Notably, the pronounced peak at ~350 °C for Ag_Fe₃O₄ highlights its strong catalytic efficiency and significant concentration of Lewis acid sites.

Recent studies have shown that Lewis acid sites contribute to ammonia dissociation, and NH₄⁺ on Brønsted acid sites is more reactive. Strong acid sites may be linked to the formation of hydroxyl groups on the support surface and the NH₃ dissociation capacity of the catalysts [37].

The acidity of Ag_WO₃, Ag_TiO₂, and Ag_Fe₃O₄ is directly correlated with their catalytic activity in ammonia ozonation. Ammonium is adsorbed on Brønsted acid sites of the support, while ozone decomposes on the silver active sites.

Additionally, previous studies have shown that the basic nature of ozone promotes its adsorption and dissociation at the Lewis acid sites of the catalyst, leading to the formation of atomic oxygen [38,39].

2.2.5. UV-Vis Spectroscopy

The Kubelka–Munk absorption curves derived by UV-Vis analysis of studied Ag_WO₃, Ag_TiO_2, Ag_Fe₃O₄, and Ag NPs catalysts are presented in Figure 6.

All catalysts exhibit light absorption around 320 nm (UV range). For the Ag_Fe₃O₄ catalyst, a strong adsorption is observed in the range 380–800 nm, while for Ag_TiO₂ a strong adsorption took place in the 400–1100 nm range (UV-Vis range), confirming the plasmonic effect of Ag in the visible spectrum. The absorption bands at wavelengths between 300 and 600 nm are given by the presence of metal ions. The existence of silver particles having various dimensions leads to the existence of an extended absorption band. This optical feature gives us information about the efficient use in advanced oxidation reactions (photocatalysis or ozone oxidation). Also, all catalysts show distinct adsorption maxima in the range of 250–400 nm, which may be attributed to intrinsic structural defects, such as metal vacancies or oxygen vacancies [40].

2.3. Catalytic Ozonation Reaction of Ammonia on Ag_WO₃, Ag_TiO₂, and Ag_Fe₃O₄

The catalytic process of ozonation can be represented schematically by the following equations:

(1)

Hoigné and Bader [41] studied the kinetics of the ozonation reaction with ammonia and found that when the ozone-to-ammonium (O₃:NH₄⁺) stoichiometry was set at 4:1, the reaction followed a second-order rate law with a rate constant K₁ = 5 M⁻¹ s⁻¹. The reaction rate for ammonium oxidation increased with rising pH (around 9.5) due to the deprotonation of NH₄⁺ into NH₃.

In aqueous solutions, ammonia exists in two forms depending on the pH: at pH < 7, the ammonium ion (NH₄⁺) predominates, whereas at pH > 7, the concentration of free ammonia (NH₃) increases [42]. In alkaline conditions, where NH₃ is the dominant form, the ozone decomposition capacity increases, promoting the formation of hydroxyl radicals (•OH) with a high oxidation potential (E⁰ = 2.80 V) [40,43].

The oxidation of ammonia and the selectivity of catalysts for nitrogen (N₂) production were evaluated using an initial NH₄⁺ (NH₄Cl) concentration of 50 mg/L, an ozone dose of 2 mg/L, a catalyst dose of 1 g/L, and a solution pH of 9.5. The catalytic efficiency is presented in Figure 7, which highlights the influence of the support materials (Fe₃O₄, TiO₂, and WO₃) on the ammonium oxidation reaction. The analysis of the results from the catalytic tests reveals that changes in ammonium concentration and nitrate formation primarily occur within the first 100 min. After this period, the evolution of these parameters slows significantly. For this reason, the catalytic test was concluded after 120 min.

After 120 min of reaction, the Ag_Fe₃O₄ catalyst exhibited the highest catalytic efficiency in ammonium ozonation, achieving a removal efficiency of ~71% and a nitrogen (N₂) selectivity of 68.43%. This fact can be attributed to the acidic site on the Ag_Fe₃O₄ surface that facilitates NH₃ adsorption.

Control experiments conducted in the absence of a catalyst with O₃ showed a removal efficiency of only 5.74%.

The activity and selectivity of the ammonia ozonation reaction occur through various kinetic processes. These processes are linked to generating reactive oxygen species, the surface adsorption and desorption rates of ammonia, and ozone decomposition. These rates are closely related to the acid-base properties of the catalyst surface.

Figure 8 shows the efficiency of the catalysts, (a) Ag_WO₃; (b) Ag_TiO₂; and (c) Ag_Fe₃O₄ in the ozonation of the ammonia reaction. As can be seen from Figure 7, nitrogen selectivity follows the same trend for all catalysts, it increases in the first 30 min and then decreases as NO₃⁻ is formed. The catalytic efficiency followed the order Ag_Fe₃O₄ >Ag_TiO₂ >Ag_WO₃. During the ozonation process, nitrate (NO₃⁻) was the major byproduct and very small amounts of nitrites were formed due to the fast oxidation of nitrite (NO₂⁻) in the ozone-rich environment.

Catalytic ozonation represents a combined process between ozone decomposition and catalytic reactions that facilitate the generation of active oxidative species.

In an alkaline environment, the ozone decomposition reaction takes place according to the equations [40].

O₃ + H₂O + e⁻ = O₂ + 2OH⁻

(2)

O₃ + OH⁻→HO₂⁻ + O₂

(3)

HO₂⁻ + O₃→O₃⁻ + HO₂

(4)

The alkaline environment promotes the formation of OH⁻ in solution. OH⁻ acts as an initiator that accelerates ozone decomposition to generate reactive oxygen species such as •OH [38]. Ammonium is efficiently removed by oxidation of •OH. On the other hand, the reaction rate of O₃ with NH₃ is about 20 times higher than that of NH₄⁺ [43,44], so the alkaline environment also makes ozonation more efficient. The efficiency of an oxidant in oxidation reactions is determined by its redox potential. It is important to note that the redox potential of ozone is affected by pH levels, measuring 2.07 V in acidic conditions and 1.24 V in alkaline conditions.

The nature of the support and the interaction of Ag–support (Fe₃O₄, TiO₂, and WO₃) influences the formation of acid-base sites, free oxygen sites, and surface hydroxyl groups. This fact can promote the decomposition of ozone and can achieve higher ammonia removal rates and nitrogen selectivity.

A low band gap is advantageous for the activation of reactants, such as O₃ and ammonia, resulting in the highest catalytic performance of Ag NPs supported on Fe₃O₄.

Figure 9 shows the product distribution at the end of the reaction after 120 min, NH₄⁺, NO₃⁻ (mg/L), and O₃ (2 g/Nm³) concentrations and the evolution of the studied catalysts.

Table 4. Comparative results of the studied catalysts with previous works reported in the literature.

Catalyst	Reaction Conditions	NH4+ Conversion (%)	N2 Selectivity (%)	Observations	Ref.
MgO	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 9.3; reaction time = 6 h	>95%	~20%	High conversion but low selectivity toward gaseous products at 60 °C	[17]
	[NH4+], 50.00 mg L−1 from NH4Cl; reaction solution volume, 250 mL; reaction temperature, 10 °C; O3 concentration, 4.56 mg L−1; pH = 9.0; (catalyst dose, 0.10 g·L−1); pH after the reaction = 10.3; reaction time = 60 min	77.5%	0	Relatively high ammonia removal	[45]
NiO	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 6.3; reaction time = 6 h	~80%	~30%	Good conversion but low selectivity toward gaseous products at 60 °C	[17]
Co3O4	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 2.1; reaction time = 6 h	74%	88%	Good conversion and high selectivity toward gaseous products at 60 °C	[17]
CuO	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 4.6; reaction time = 6 h	~50%	~50%	Low conversion and low selectivity at 60 °C	[17]
ZnO	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 6.2; reaction time = 6 h	~40%	~30%	Low conversion and low selectivity at 60 °C	[17]
Mn3O4	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 2.1; reaction time = 6 h	~30%	~60%	Very low conversion, medium selectivity at 60 °C	[17]
Fe2O3	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 2.6; reaction time = 6 h	~20%	~75%	Very low conversion, high selectivity at 60 °C	[17]
SnO2	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 2.7; reaction time = 6 h	<20%	~70%	Very low conversion, high selectivity at 60 °C	[17]
Al2O3	0.1 g catalyst; O3/O2 (1.88 mmol L–1 as O3, total flow rate = 100 cm3 min–1); initial pH = 5.4; pH after the reaction = 3.1; reaction time = 6 h	<20%	~60%	Very low conversion, medium selectivity at 60 °C	[17]
γ-Al2O3	100 g L−1 catalyst dosage; initial NH4+-N 50 mg/L; pH = 9; ozone concentration 50 mg/L; reaction time = 60 min	62.7%	0	Medium NH4+-N removal	[43,46]
MgO-Co3O4	1 g L−1 catalyst; initial ammonia concentration 50 mg/L; pH = 9; flow rate of ozone 12 mg/min; contact time = 2 h	82.5%	44.8%	Good conversion, medium selectivity towards N2 at 50° C	[19]
Fe2O3-Co3O4 dolomite	1 g L−1 catalyst; initial ammonia concentration 50 mg/L; pH = 9; flow rate of ozone 12 mg/min; reaction temperature 50 °C; reaction time = 60 min	96%	78%	High conversion, good selectivity towards N2 at 50 °C	[20]
Ag_Fe3O4	1 g L−1 catalyst; 50 mg L−1 NH3 (initial concentration); 2 mg/L O3; solution pH= 9.5; reaction time = 120 min	72%	69%	Good conversion and high selectivity toward gaseous products at 20° C	This work
CaO2	80 g L−1 catalyst dosage; initial NH4+-N 10 mg/L; pH = 10.68; ozone concentration 2.5 g/h; reaction time 15 min	95%	0	Very good NH4+-N removal	[43,47]
Co/Mg(OH)2	1 g L−1 catalyst; initial NH3 concentration of 10 mg/L; ozone dose of 1.2 mg/L; pH = 9.0; reaction time = 120 min	90.2%	63.4%	This catalyst can efficiently remove NH3 and form gaseous N	[40]

provides a comparison of the catalytic performance of the studied catalysts (Ag_WO₃, AgTiO₂, and Ag_Fe₃O₄) in ammonia ozonation with other catalytic systems reported in the literature.

The superior advantages of our catalyst consist in the fact that it can operate at ambient temperature, it has good stability, a low-cost price, and it is very easy to recover with the help of a magnet. These advantages recommend the use of this catalyst in processes of decontamination of waters containing ammonia. It can also be successfully used in a combined technology for nitrate reduction in which ammonium results as an unwanted product.

In Figure 10, we propose a scheme of the mechanism for ammonia oxidation on the Ag_Fe₃O₄ catalyst.

As illustrated in the diagram above, the oxidation reaction of ammonium with ozone in an alkaline aqueous medium using the Ag-Fe₃O₄ catalyst involves two combined processes. The alkaline pH (9.5) promotes the formation of hydroxide ions (OH⁻) in the solution. In the first stage, OH⁻ acts as an initiator that accelerates the decomposition of ozone, generating reactive oxygen species such as hydroxyl radicals (•OH), hydroperoxide (HO₂¯), or oxygen (O₂), according to Equations (2) and (3). During this stage, silver nanoparticles serve as active centers, facilitating the transfer of electrons to ozone. As a result, Ag⁺ is reduced to Ag⁰, which contributes to the electrostatic equilibrium within the support network. In the second stage, a chain reaction occurs involving HO₂⁻, which generates more hydroxyl radicals. The oxygen produced from the decomposition of ozone fills the oxygen vacancies in the Fe₃O₄ support lattice. Additionally, silver significantly enhances the surface and redox properties of the catalytic ozonation process. Silver facilitates the electron transfer between Ag⁺, Ag⁰, and back to Ag⁺, thereby accelerating the overall reaction kinetics. Also, Cl⁻ ions (present in solution from NH₄Cl -ammonium source) can participate in NH₄⁺ ozonation, so Cl⁻ is first oxidized by O₃ on Fe₃O₄ to form ClO⁻ [17], which oxidizes NH₄⁺. On the other hand, a significant concentration of Lewis acid sites of Ag_Fe₃O₄ can be correlated with the high catalytic efficiency of this catalyst.

Additional catalytic tests were carried out to evaluate the stability of the Ag_Fe₃O₄ catalyst that showed the highest catalytic activity for ammonia oxidation. The four experiments were performed using a catalyst dosage of 1 g L⁻¹, with an initial ammonium concentration of 50 mg L⁻¹. Each experiment has a duration of 120 min. After 120 min of reaction, the catalyst was recovered, cleaned with distilled water, dried at 100 °C for 2 h, topped at a dosage of 1 g L⁻¹, and reused in the next cycle. The results, after four reusability tests, are reported in Figure 11, in terms of conversion of ammonia removal vs. time.

Table S1 shows the ammonium conversion, the selectivity to NO₃, and NO₂ after four reaction cycles. An increase in the ammonium removal conversion can be observed in the second reaction cycle, followed by a slight decrease while the N₂ selectivity increases.

The structural changes in the catalysts during the stability tests were realized by the X-ray diffraction characterization of the Ag_Fe₃O₄ catalyst after the reaction (Figure S3. XRD diffraction for fresh Ag_Fe₃O₄ and used Ag_Fe₃O₄ catalysts). From the X-ray diffractogram for the spent catalyst (see Table S2), the stability of the catalyst can be observed even after four reaction cycles. The lattice parameter of silver (a = b = c = 4.093) nm remains constant compared to fresh Ag_Fe₃O₄. The crystallite size determined from XRD diffraction using the Williamson–Hall method is 12.5, while the concentration of silver on the surface calculated by the Reference Intensity Ratio (RIR) method is 1.02% Ag compared to 1.19% for the fresh catalyst (Table S2).

3. Materials and Methods

X-ray powder diffraction (XRD) measurements of silver nanoparticles (Ag Np) supported on Fe₃O₄, TiO₂ (Aerosil P25), and WO₃ (Alpha Aesar. Haverhill, MA, USA) were conducted using a Rigaku Ultima IV X-Ray Diffractometer (Tokyo, Japan), operating at 40 kV and 30 mA with Cu Kα radiation (Kα = 1.54 Å). The diffraction patterns were recorded in the 2Θ range of 30–90° at a scan rate of 5°/min and a step size of 0.02°. Crystalline phases were identified using Rigaku’s PDXL software (version 1.8.0.3) in conjunction with the ICDD PDF-2 database.

Transmission electron microscopy (TEM)analysis of the Ag NP was performed with an FEI Tecnai G2-F30 S-Twin microscope (Beaverton, OR, USA) operating at 300 kV. Small quantities of the colloidal nanoparticles were drop-cast onto holey carbon-coated copper grids and air-dried before TEM analysis.

The morphology of the samples (Ag_WO₃, Ag_TiO₂, and Ag_Fe₃O₄) was investigated by scanning electron microscopy (SEM) using a high-resolution microscope, Zeiss EVO HD15 Scanning Electron Microscope (Jena, Germany), operating at 20 kV, equipped with QUANTAX EDS spectrometer—Brucker XFlash 6/30 (Billerica, MA, USA).

The distribution of basic and acidic sites on the surface of silver-supported oxides (WO₃, TiO₂, and Fe₃O₄) was analyzed using temperature-programmed desorption (TPD) with pre-adsorbed CO₂ or NH₃ as probing gases, performed on a CHEMBET 3000-Quantachrome instrument (Boynton Beach, FL, USA). A 50 mg sample, dried and calcined, was placed in a quartz tube and heated to 300 °C in a 90 mL·min⁻¹ CO₂ or NH₃ atmosphere for 60 min, then cooled to room temperature. The sample was subsequently purged with a 90 mL·min⁻¹ stream of argon to remove physically adsorbed CO₂ or NH₃. Once a stable baseline was achieved, desorption was carried out from 50 °C to 800 °C at a heating rate of 10 °C·min⁻¹. Desorption peaks were deconvoluted to quantify the basic and acidic sites (classified as weak, medium, and strong) present in the sample. The thermal conductivity detector (TCD) signal was calibrated using a known amount of CO₂ or NH₃.

Hydrogen temperature-programmed reduction (H₂-TPR) was performed to evaluate the reducibility of the catalysts (Ag_Fe₃O₄, Ag_TiO₂, and Ag_WO₃), and to examine the interaction between the active phase (Ag nanoparticles) and the support surface. A 50 mg sample was heated from room temperature to 1073 K at a rate of 10 K·min⁻¹ in a stream of 5% H₂/Ar.

Diffuse reflectance UV-Vis spectroscopy was conducted using a Perkin Elmer Lambda 35 spectrophotometer (Shelton, CT, USA) equipped with an integrating sphere. Spectra were recorded at room temperature over the wavelength range of 200–1100 nm, using Spectralon as the reference. Reflectance data were converted to absorption spectra using the Kubelka–Munk function, F(R).

Catalytic tests were performed in a quartz reactor setup, including a pH meter, stirring system, temperature maintenance through recirculation at 20 °C, and an ozone generator. Before testing the catalytic materials, optimal experimental parameters for the ozone oxidation reaction were established. Aqueous ammonium solution from NH₄Cl at a concentration of 50 mg/L was introduced into the reactor along with the catalyst sample. Ozone was bubbled into the reactor from the ozone generator while maintaining the 1 g/L catalyst in suspension using a magnetic stirrer. Every 30 min, a 1 mL sample was withdrawn from the reaction solution for analysis. The concentrations of NO₃⁻, NO₂⁻, and NH₄⁺ were determined using two Dionex ICS 900 ion chromatographs (Thermo Fisher Scientific, Waltham, MA, USA). An anion-exchange resin column, along with an aqueous mixture of NaHCO₃ (4.5 mM) and Na₂CO₃ (1.4 mM) as the eluent and H₂SO₄ (10 mM) as the regenerant solution, was used for anion analysis. For cation analysis, a cation-exchange resin column was utilized with a methane sulfonic acid (20 mM) solution and TBAOH (40 mM) as the eluent and regenerant phases, respectively. The conversion of ammonia and selectivity for N₂ gas (including NO_x and N₂) and NO₃⁻ selectivity were calculated using Equations (5)–(7).

$$X_{{NH}_4^{+}}\%={\displaystyle \frac{\left[{NH}_4^{+}\right]\mathrm{t}}{\left[{NH}_4^{+}\right]i}}\times 100$$

(5)

$${S_N}_{2-gas}\%=100-{S_{NO}}_3$$

(6)

$${S_{NO}}_3\%={\displaystyle \frac{\left[{NO}_3\right]\mathrm{t}}{\left[{NH}_4^{+}\right]c}}\times 100$$

(7)

where:

• $X_{{NH}_4^{+}}$ is the ammonia conversion,
• $\left[{NH}_4^{+}\right]\mathrm{t}$ is the ammonia concentration at time t and $\left[{NH}_4^{+}\right]i$ is initial ammonia concentration,
• $\left[{NH}_4^{+}\right]c$ is the consumed ammonium at the end of the reaction.

4. Conclusions

Silver nanoparticles with dimensions of 3–6 mm were successfully synthesized using the alkaline polyol method. These nanoparticles were then utilized as catalysts by depositing them onto various oxide materials, resulting in the formation of three distinct catalytic systems: Ag_Fe₃O₄, Ag_TiO₂, and Ag_WO₃.

The characterization of these catalysts involved several analytical techniques, including transmission electron microscopy (TEM), scanning electron microscopy (SEM), X-ray diffraction (XRD), UV-Vis spectroscopy, H₂-TPR, NH₃-TPD, CO₂-TPD, and specific surface area by BET method. These characterization methods provided insight into the structural and optical properties of the synthesized catalysts.

The materials were tested in the ammonium oxidation reaction to evaluate their catalytic performance, using ozone as the oxidizing agent.

The results revealed notable differences in catalytic efficiency among the catalysts, clearly ranking their effectiveness: Ag_Fe₃O₄ outperformed Ag_TiO₂ and Ag_WO₃. This compelling evidence underscores the critical importance of selecting the right oxide material, as it plays a significant role in enhancing catalytic activity during the ozonation process.

The high dispersion of Ag, along with an optimal content of Ag⁰ and an adequate quantity of acidic sites, were essential for enhancing the ammonia ozonation reaction. Notably, the Ag_Fe₃O₄ catalyst showed remarkable stability, demonstrating minimal leaching of their active components.