Unlocking Catalytic Efficiency: How Preparation Strategies and Copper Loading Enhance Hydroxyapatite Catalysts for NH₃ Oxidation

Abstract

The selective catalytic oxidation of ammonia (NH₃-SCO) is gaining attention due to the hazardous nature of NH₃ and its inclusion in emission reduction frameworks such as the National Emission Ceilings Directive and the Gothenburg Protocol (1999). Copper-based hydroxyapatite (Cu/HAP) catalysts have emerged as a promising solution, offering high activity and cost-effectiveness. This study evaluated two preparation methods: a one-pot co-precipitation technique and post-synthesis copper deposition, varying both the contact time and copper concentration. The influence of copper loading and preparation method on catalyst performance in NH₃-SCO was investigated in a continuous flow reactor over a temperature range of 200–500 °C, with a fixed gas hourly space velocity (GHSV) of 120,000 h⁻¹ and an NH₃/O₂ ratio of 0.03. X-ray diffraction and DR-UV spectroscopy confirmed the high crystallinity of HAP and provided insights into copper speciation. X-ray photoelectron spectroscopy revealed that Cu/HAP catalysts prepared via one-pot co-precipitation predominantly contained isolated Cu²⁺ species, which were associated with high catalytic activity in selective NH₃-SCO. Conversely, a higher degree of copper structuring was observed in catalysts prepared by post-synthesis deposition, particularly at higher Cu loadings. These findings highlight the potential to tailor Cu structuring on HAP to enhance performance in NH₃-SCO through optimized preparation strategies.

1. Introduction

Hydroxyapatite (HAP, Ca₁₀(PO₄)₆(OH)₂), a bioavailable and biocompatible inorganic material, has emerged as a highly versatile catalyst support due to its unique properties including chemical and thermal stability, tunable surface acidity, and the ability to accommodate various active metal species, such as Cu²⁺ and Fe³⁺, within its framework through ionic exchange properties [1,2,3,4]. These features, combined with its eco-friendly nature, make HAP a suitable choice for a wide range of catalytic applications, particularly in environmental catalysis for addressing pressing challenges such as the mitigation of greenhouse gases and air pollutants [5].

In this field, copper-modified hydroxyapatite (Cu/HAP) has gained significant attention in various catalytic environmental reactions due to the unique redox and Lewis acid properties of Cu(II) and its good dispersibility on the HAP surface. Cu/HAP has shown promising performance in the selective catalytic reduction of NO_x by NH₃ (NH₃-SCR) [6,7,8,9,10], the decomposition of N₂O (De-N₂O) [11,12], and the abatement of volatile organic compounds (De-VOCs) [13,14,15]. Previous studies have demonstrated that the method of Cu incorporation into the HAP significantly affects the speciation and distribution of copper, thereby influencing catalytic activity and selectivity. Copper species on HAP can adopt various forms depending on the preparation method, conditions used, and copper loading. Qu et al. [16] identified five distinct locations for Cu species in HAP: (i) substitution of surface-exposed Ca²⁺ ions leading to dispersed Cu(II) clusters, (ii) deposition and structuring as bulk CuO or libethenite (Cu₂(OH)PO₄) on the surface, (iii) substitution at bulk Ca(2) or (iv) Ca(1) sites, and (v) location within the c-axis oriented channels as CuO particles. These different copper configurations vary in terms of accessibility and reactivity, influencing their suitability for different catalytic applications. In addition, approaches such as co-precipitation and the post-synthesis functionalization of supports result in varying levels of Cu dispersion and structuring.

Co-precipitation has emerged as a particularly effective method for preparing Cu/HAP catalysts with a good dispersion of copper species. In this approach, copper ions are incorporated into the hydroxyapatite lattice during the precipitation process, often resulting in isolated Cu(II) species that are surface-complexed with HAP groups or re-place calcium ions in the surface layers of the HAP structure. Recently, an innovative one-pot synthesis method based on co-precipitation followed by hydrogen treatment for copper exsolution was developed by Costentin et al. [10] This approach was shown to be highly effective in producing Cu/HAP catalysts with highly dispersed copper phases. The process promotes the migration of copper from the bulk to the surface under reducing conditions, forming highly active and accessible Cu species, which significantly enhance catalytic performance, particularly in NH₃-SCR. Such exsolution processes, inspired by similar phenomena in perovskite-based systems, allow for the precise tuning of Cu species while maintaining the structural integrity of the HAP support.

In post-synthesis functionalization, the metal speciation is particularly sensitive to the loading and contact time during preparation [12]. At low loadings (<3 wt.%) and short contact times (up to ca. 4 h), copper predominantly forms isolated species that interact strongly with the surface HAP groups or substitute surface-exposed calcium ions, resulting in highly dispersed CuO nanoclusters. These isolated species are often associated with higher catalytic activity for reactions such as VOC oxidation or NH₃-SCR due to their increased accessibility and optimal dispersion. At moderate loadings (up to 6 wt.%), slight copper aggregation into nanoparticles (~1.6 nm) begins to occur, while still maintaining significant catalytic activity. This behavior has recently been observed in reactions such as N₂O decomposition, where vicinal copper sites are essential for achieving high N₂O decomposition activity. Higher loadings (>6 wt.%) and extended contact times (48 h) lead to the aggregation of copper into larger CuO clusters, eventually resulting in the formation of the libethenite phase (Cu₂(OH)PO₄); this phase is catalytically inactive for most reactions. Each preparation method and set of conditions has distinct advantages, underscoring the importance of preparation strategies in tailoring Cu/HAP catalysts for specific applications.

Among the many reactions catalyzed by Cu/HAP, the selective catalytic oxidation of ammonia (NH₃-SCO) has been less investigated despite its relevance. NH₃-SCO is a critical process for removing excess ammonia from industrial exhaust streams, converting it to nitrogen and water without forming undesirable by-products such as NO_x or N₂O [17]. In the context of the NH₃-SCO reaction, copper-containing zeolites have demonstrated superior performance due to their high catalytic activity and selectivity for nitrogen (N₂) production [18,19]. The efficiency of these catalysts is strongly dependent on the speciation of copper, which includes various forms such as isolated Cu²⁺ ions, CuO clusters, and Cu_xO_y nanoparticles. Several studies have shown that the isolated Cu²⁺ and Cu⁺ species are the primary active sites in this reaction, with CuO clusters also contributing to certain conditions. The activity of Cu/zeolites, such as Cu/SSZ-13, is primarily attributed to the redox cycling of Cu²⁺ and Cu⁺, which facilitates the oxidation of ammonia to nitrogen and water [18,20,21] through an internal selective catalytic reduction (i-SCR) mechanism. In this pathway, NH₃ is first oxidized to NO, which is then rapidly consumed by the SCR reaction, leading to the formation of N₂ and H₂O. This mechanism is supported by observations that NO is barely detected in the effluent at low temperatures, indicating its swift consumption in the reaction sequence [18,22,23]. For this mechanism, the presence of copper species appropriately coordinated to a support and with suitable electronic properties is crucial.

This study explores the potential of Cu/HAP catalysts for the ammonia oxidation reaction, with a focus on how different preparation methods influence structural and catalytic properties. Specifically, one-pot co-precipitation and the post-synthesis functionalization of HAP were employed to synthesize Cu/HAP with varying Cu loadings. The relationship between Cu dispersion at the surface, its oxidation state, and catalytic performance in NH₃-SCO was thoroughly examined.

2. Results

Two series of catalysts were prepared by introducing copper into hydroxyapatite via the one-pot co-precipitation (OP) or post-synthesis deposition (D) methods, as described in detail in Section 4.1 (Materials and Catalyst Preparation). Scheme 1 illustrates the experimental conditions for the synthesis of the Cu/HAP catalysts and the variations in operating parameters among the different samples, particularly the copper loading. In brief, the one-pot co-precipitation synthesis involved the incorporation of copper ions into the HAP lattice during its formation, giving rise to the expected homogeneous copper distribution throughout the final particles, from the bulk to the surface. In contrast, in post-synthesis deposition, copper was loaded onto the surface of preformed HAP powder, making, in principle, all of the added copper accessible on the surface. These two different preparative strategies are expected to impact copper dispersion and structuring.

Table 1. Copper content, surface area, and porosity of the prepared Cu/HAP catalysts.

Catalyst	Cu Loading a (wt.%)	Surface Area b (m2g−1)	Pore Volume c (cm3g−1)	Average Pore Radius d (nm)
Cu/HAPOPS	4.5	52	0.18	9.9
Cu/HAPOPC	5.6	95	0.27	4.9
Cu/HAPOPP	5.7	57	0.22	12.5
Cu/HAPD,L e	3.1	63	0.32	9.4
Cu/HAPD,H e	6.2	68	0.34	9.3
Cu/HAPD,2L e	6.8	76	0.20	8.4
Cu/HAPD,F e	4.6	75	0.34	8.5

^a Obtained by ion chromatography analyses (double-check analysis, see Section 4.2 Experimental section); ^b Specific surface area determined by the 2-parameter BET equation; ^c At p/p⁰ = 0.95; ^d Determined by the BJH model from the desorption branch of the isotherm (0.3 < p/p⁰ < 0.95); ^e Catalysts prepared starting from the synthetic and calcined HAP (Ref [24]) with a 57 m²g⁻¹ surface area, 0.28 cm³g⁻¹ pore volume, and 9.4 nm mean pore radius.

summarizes the copper loading, surface area values, and porosity features of the Cu/HAP catalysts studied. In general, copper was incorporated quantitatively into the catalysts using both preparation methods. The actual copper loading, determined by ion chromatography (IC) analysis, closely matched the nominal values (3 wt.% or 6 wt.%) for all catalysts, except for Cu/HAP_OPS and Cu/HAP_D,F, which exhibited slightly lower Cu contents (4.5–4.6 wt.%) compared with the nominal value (6 wt.%). The lower Cu loading in Cu/HAP_OPS was likely due to the high concentration of ammonium hydroxide in the synthesis medium. Since copper was introduced only at the final stage, it encountered an ammonia-rich environment where strong complexation with ammonia likely stabilized the copper ions, limiting their incorporation into HAP (see Scheme 1). Concerning Cu/HAP_D,F, its lower loading compared with the nominal value was expected due to the short contact time (0.25 h) between the Cu-salt precursor and the HAP powder (Scheme 1), which limited copper adsorption onto the HAP surface.

The adsorption/desorption isotherms and pore size distributions of the Cu/HAP samples, presented in Figure S1, showed that all samples exhibited a profile and a hysteresis loop typical of mesoporous materials. Surface area values ranged from 55 to 75 m² g⁻¹, while the average pore size and pore volume values were in the range of 8.5–12 nm and 0.2–0.3 cm³ g⁻¹, respectively. These values align with those of bare and calcined hydroxyapatite, indicating that copper modification did not significantly affect the pristine mesoporosity of HAP, as reported in our previous works [6,7]. The only exception was the Cu/HAP_OPC sample, synthesized via conventional co-precipitation, which exhibited a higher surface area (95 m² g⁻¹) and the smallest average pore size (4.9 nm) among all samples. This suggests that conventional co-precipitation produced smaller Cu particles compared with the other synthesis/preparation methods.

The XRD patterns of the samples (Figure 1) confirmed the presence of the hydroxyapatite phase (JCPDS: 00–09–0432) in a hexagonal space group, as indicated by the characteristic peaks at 2θ values of 26.1°, 28.2°, 32.0°, 33.2°, 40.1°, 46.9°, and 48.5°, corresponding to the (002), (012), (211), (300), (310), (222), and (320) planes, respectively. Notably, no CuO or libethenite (Cu₂PO₄OH) phases were detected, indicating that copper was highly dispersed within the HAP matrix without segregation into the crystalline copper-containing phases. However, given the relatively low copper content, the presence of minor Cu-containing secondary phases could not be entirely ruled out. All of the collected diffractograms displayed sharp peaks, except for Cu/HAP_OPC, which had broader peaks, suggesting the presence of smaller particles, which also agreed with the surface area measurements (Table 1).

By applying the Scherrer equation, an approximate evaluation of the crystalline domain size along the c-axis (D_{(0 0 2)}) and within the a–b plane (D_{(3 0 0)}) could be obtained. The data in Table S1 indicate that all samples, with the exception of Cu/HAP_OPC, possessed comparable dimensions and displayed a characteristic elongated rod-like morphology, with a predominant exposure of {010} facets, as evidenced by the significantly high D_{(0 0 2)}/D_{(3 0 0)} ratio. In contrast, the Cu/HAP_OPC sample generally exhibited smaller dimensions and a rod-like morphology.

To gain insights into the coordination and aggregation state of copper in all samples, UV–Vis-DR spectra were recorded at room temperature over the 200–2000 nm wavelength range. The resulting spectra (Figure 2) exhibited multiple complex bands that were not straightforward to assign; a tentative interpretation, obtained through sub-band decomposition, is shown in Figure S2 and Table S2 [25]. A broad absorption band between 590 and 1000 nm, centered around 850 nm, was observed in all Cu/HAP samples, along with a distinct narrow band at approximately 200 nm. The broad band can be attributed to d–d transitions of Cu²⁺ ions in a pseudo-octahedral oxygen ligand environment, suggesting copper species occupying substitutional lattice positions or existing as dispersed nanoparticles. The narrow band near 200 nm corresponded to the O → Cu²⁺ ligand-to-metal charge transfer (LMCT) transition, which involves lattice oxygen and isolated mononuclear Cu²⁺ centers.

Additionally, a pronounced shoulder at 420 nm was observed exclusively in the samples prepared via post-synthesis functionalization with a high copper loading (6 wt.%). This feature is likely to be associated with Cu–O–Cu complexes, indicative of partial copper structuring in these samples, made possible by the extended contact time and elevated copper concentration employed in the post-synthesis method.

XPS analysis was carried out to elucidate the surface composition and chemical states of copper on the sample surfaces.

The XPS survey spectra confirmed the presence of Ca, P, O, and Cu as the major elements present at the surface of the Cu/HAP samples consistently, with the expected composition of Cu/HAP catalysts. The additional presence of carbon was due to the incorporation of carbonate (CO₃²⁻).

Table 2. Surface composition of the Cu/HAP samples as the atomic percent of elements and relevant elemental ratios obtained by XPS.

Code	Cu	Ca	P	C	O	${\displaystyle \frac{\mathit{C}\mathit{a}}{\mathit{P}}}$	${\displaystyle \frac{(\mathit{C}\mathit{a}+\mathit{C}\mathit{u})}{\mathit{P}}}$
	Atomic %
Cu/HAPOPS	2.91	16.0	13.8	17.2	50.1	1.16	1.37
Cu/HAPOPC	2.32	13.1	10.9	27.8	45.9	1.20	1.41
Cu/HAPOPP	3.53	11.5	11.5	30.5	42.9	1.00	1.30
Cu/HAPD,L	1.55	16.9	12.9	16.8	50.8	1.31	1.43
Cu/HAPD,2L	1.89	16.0	14.3	20.2	47.7	1.12	1.25
Cu/HAPD,H	1.65	17.1	13.7	20.8	47.9	1.25	1.37
Cu/HAPD, F	2.50	16.3	16.1	16.3	48.8	1.01	1.17

presents the atomic percentage (at%) of the elements observed on the Cu/HAP catalysts.

Regarding the surface copper concentration, interesting evidence has emerged. Specifically, the highest surface copper concentration was found in the samples prepared via one-pot co-precipitation, with atomic percentages of Cu between 2.3% and 3.5%, whereas those prepared via post-synthesis deposition exhibited lower values, ranging from 1.5% to 1.9%, except for Cu/HAP_D,F (2.5%), as discussed below. This may seem counterintuitive, as post-synthesis deposition should lead to copper being primarily accommodated on the surface, thus resulting in higher Cu surface concentrations. However, the extended contact time between the copper solution and the HAP solid during the deposition process likely facilitated the formation of structured copper species. As a result, despite being present uniquely at the surface, copper was not highly dispersed but rather gave rise to CuO nanoaggregates, as indicated by DR-UV spectroscopy. This interpretation is further supported by the observation that the sample prepared by the post-deposition of copper but with a shorter contact time (Cu/HAP_D,F) exhibited a high copper percentage (2.5%) that was comparable to that of the samples prepared through the co-precipitation method. In contrast, co-precipitation appeared to facilitate the gradual incorporation of copper into the HAP lattice, leading to a more homogeneous and dispersed distribution including in the surface layers.

Moreover, from the data in Table 2, it appears that the XPS Ca/P and (Ca + Cu)/P ratios were consistently lower than the bulk stoichiometric values in all of the samples. This discrepancy has been widely reported in the literature for hydroxyapatite and other calcium phosphate materials, and it might be due to the instability of the materials under prolonged exposure to the X-ray source as well as to surface reconstruction and segregation phenomena.

Figure 3 reports the high-resolution Cu 2p spectra of the Cu/HAP samples, which show spin–orbit splitting into Cu 2p_3/2 (930–935 eV) and Cu 2p_1/2 (950–960 eV) peaks. Since both peaks conveyed similar chemical information, our attention only focused on the 2p_3/2 region.

A clear difference in the Cu 2p characteristic peaks could be seen by the fitting procedure between the catalysts prepared by one-pot co-precipitation and those obtained by post-synthesis deposition.

In the catalysts prepared via one-pot co-precipitation, the primary peak appeared at a binding energy of approximately 933 eV. While this binding energy is consistent with both Cu(I) and Cu(II) oxidation states, the presence of a strong satellite peak between 941 and 944 eV, along with the observed spin–orbit splitting of around 20 eV between the Cu 2p_3/2 and Cu 2p_1/2 peaks, suggests that copper was predominantly present as isolated Cu(II). Additionally, a smaller shoulder peak (934.9 eV) was observed in the Cu/HAP_OPC and Cu/HAP_OPP samples; this signal is typical of the Cu(OH)₂ environment and can be attributed to copper in interstitial positions along the hexagonal channel, which during calcination forms [O^2–– Cu²⁺– O^2–] moieties [26,27].

In contrast, the analysis of samples prepared via post-synthesis deposition revealed the absence of the 933 eV component, and thus Cu species were not present as isolated centers on the catalyst surfaces. Instead, the primary peaks were centered at approximately 934 eV for Cu/HAP_D,L and Cu/HAP_D,2L, corresponding to Cu(OH)₂ and ascribable to oligomeric [O^2–– Cu²⁺– O^2–] moieties, and at 935.7 eV for Cu/HAP_D,H and Cu/HAP_D,F, which can be attributed to nanostructured CuO/CuCO₃. Regarding this difference in copper speciation, among the post-synthesis samples, an interesting correlation seemed to emerge between copper speciation and copper concentration in the precursor salt solution used for deposition. Specifically, in the Cu/HAP_D,L and Cu/HAP_D,2L samples prepared using a less concentrated copper nitrate solution (0.08 M, see Scheme 1), the predominant species was interstitial Cu²⁺, whereas when using higher concentration solutions (0.16 M, see Scheme 1), the formation of copper nanostructures in Cu/HAP_D,H and Cu/HAP_D,F was observed. This phenomenon could be attributed to the differing acidity of the precursor salt solutions, which correlates with the copper concentration: the higher the copper concentration, the more acidic the environment. The increased acidity may partially dissolve HAP, leading to the reprecipitation of nanoaggregates containing copper.

The high-resolution XPS spectra of the O 1s region are shown in Figure S3. In Cu/HAP samples prepared via the one-pot method (Figure S3a), the O 1s peak exhibited a main component centered around 531 eV, accompanied by asymmetric broadening toward higher binding energies. Deconvolution of the O 1s signal revealed three distinct contributions at 531.1 eV, 532.5 eV, and 533.3 eV, which could be assigned to the lattice oxygen from HAP (the dominant component), oxygen vacancies, and hydroxyl groups/bridging oxygen species, respectively [28,29,30]. Additionally, in the Cu/HAP_OPS sample, a minor peak appeared at 534.5 eV, which may be attributed to surface carbonate oxygen or physisorbed hydroxyl ions.

In contrast, the O 1s spectral profiles of the samples prepared via post-synthesis deposition exhibited a more complex and heterogeneous distribution. In all of these samples, the contribution attributed to lattice HAP oxygen (ca. 531 eV) was noticeably less intense than in the one-pot series, possibly indicating a more disordered and less-apatitic surface layer. Specifically, Cu/HAP_D,L and Cu/HAP_D,2L showed a predominant peak at 532.8 eV, which is commonly interpreted as indicative of oxygen vacancies or adventitious hydroxyl species derived from water adsorption [28,31].

Conversely, in Cu/HAP_D,H and Cu/HAP_D,F, the dominant feature appeared between 533.7 and 534.1 eV. This signal might be associated with P–OH species or adsorbed water; however, given the similar pre-treatment conditions applied to all samples, the former attribution (P–OH) is more plausible.

It should also be noted that potential contributions from surface copper carbonates or copper oxides may partially overlap with the signal of lattice oxygen, further complicating their identification.

Transmission electron microscopy (TEM) coupled with energy-dispersive X-ray spectroscopy (EDS) analyses were performed on two representative samples: Cu/HAP_OPS, selected from the one-pot series, and Cu/HAP_D,H from the post-synthesis deposition series, in which a more aggregated copper phase should be present.

TEM images (Figure 4) confirmed the presence of hydroxyapatite (HAP) nanorods in both samples. These structures exhibited an elongated morphology, with lengths ranging from 50 to 150 nm and widths between 10 and 30 nm. The observed dimensions were in good agreement with the crystallite sizes estimated from the X-ray diffraction data using the Scherrer equation (Table S1). In the case of Cu/HAP_D,H, EDS data collected in multiple regions of the sample (Table S3) showed large local variations in copper content, which could suggest non-uniform dispersion, potentially indicating aggregation or clustering of metal particles, in agreement with the UV–Vis-DR and XPS observations. Conversely, a uniform copper distribution across multiple areas was observed in Cu/HAP_OPS, with an average value close to the one determined by IC, which might be consistent with a more dispersed state for copper. Nonetheless, a definitive assessment of the copper aggregation state would require dedicated characterization techniques specifically designed to elucidate the spatial distribution and chemical environment of copper species at the atomic scale.

The catalytic performances of bare HAP and Cu/HAP samples in NH₃-SCO reaction were evaluated by feeding a gaseous mixture containing ammonia (ca. 300 ppm), oxygen (ca. 10,000 ppm), and nitrogen (as inert) in a fixed bed reactor at a GHSV of 120,000 h⁻¹. Along with NH₃ consumption and N₂ formation, even the formation of NO, N₂O, and NO₂ from the side reactions was monitored through FTIR spectrophotometry.

The catalytic results are shown in Figure 5 as plots of NH₃ conversion as a function of the temperature, while the values of selectivity to N₂ at fixed temperature (350 °C) are listed in

Table 3. NH₃-SCO selectivity of the Cu/HAP catalysts evaluated at a fixed reaction temperature (350 °C).

Catalyst	Selectivity (%)
	N2	N2O	NO	NO2
Cu/HAPOPS	84.2	10.9	4.4	0.5
Cu/HAPOPC	79.4	10.2	9.2	1.2
Cu/HAPOPP	88.8	11.0	-	0.2
Cu/HAPD,L	85.3	8.6	3.9	2.2
Cu/HAPD,2L	84.9	9.4	4.6	1.1
Cu/HAPD,H	87.5	6.7	5.1	0.7
Cu/HAPD,F	81.1	13.8	4.9	0.2

TEM characterization of the most active catalyst, Cu/HAP_OPS (Figure 4c,d), revealed that the sample morphology remained unaltered after the reaction, with no evidence of copper aggregation even at high magnification.

.

Catalytic tests confirmed that bare HAP exhibited negligible oxidation activity (Figure S4), highlighting the crucial role of copper in enabling efficient NH₃ oxidation. Regardless of the preparation procedure, all Cu-containing catalysts exhibited significant catalytic activity in the 300–400 °C range, with good selectivity toward N₂, the desired product. Overall, the best catalytic performance was observed at 350 °C, where the NH₃ conversion surpassed 90% (Figure 5) while maintaining relatively high N₂ selectivity. However, excessive oxidation beyond this temperature led to the overproduction of nitrogen oxides (in particular, N₂O and NO), limiting the practical catalyst efficiency.

Variations in Cu loading and preparation method influenced both the conversion efficiency and the selectivity trends. Among the Cu/HAP one-pot catalysts (Figure 5a), Cu/HAP_OPS demonstrated the best catalytic performance, with an onset temperature of 200 °C and NH₃ conversion exceeding 90% at 350 °C. In contrast, Cu/HAP_OPC and Cu/HAP_OPP showed lower catalytic activity, requiring temperatures of 500 °C to achieve similar NH₃ conversion levels. Despite this, these two catalysts exhibited higher selectivity toward N₂, particularly Cu/HAP_OPP (Figure S5 and Table 3). The high N₂ selectivity at 350 °C suggests that these catalysts may favor a reaction pathway that minimizes the formation of undesired nitrogen oxides. However, above 350 °C, the N₂ selectivity declined sharply, indicating the increased formation of nitrogen oxides such as N₂O, NO, and NO₂. This behavior reflects the well-known mechanism of ammonia oxidation, where at higher temperatures, excessive oxidation promotes the formation of oxidized nitrogen species [32,33].

For catalysts prepared via deposition, Cu/HAP_D,L exhibited the highest catalytic activity, achieving >90% NH₃ conversion before 400 °C. Cu/HAP_D,2L and Cu/HAP_D,H displayed comparable performance (Figure 5b), with conversion efficiencies approaching those of Cu/HAP_D,L. On the other hand, Cu/HAP_D,F exhibited significantly lower activity, reaching ~90% NH₃ conversion only at 500 °C (Figure 5b).

Based on catalytic performance, evaluated as NH₃ conversion to N₂, the following ranking, computed at 350 °C, can be established:

Cu/HAP_OPS > Cu/HAP_D,L > Cu/HAP_D,2L > Cu/HAP_D,H ≈ Cu/HAP_D,F > Cu/HAP_OPC > Cu/HAP_OPP.

This ranking highlights that an optimal NH₃-SCO catalyst could be designed by enhancing NH₃ conversion, which requires high copper dispersion on the surface, or by achieving high N₂ selectivity through a judiciously structured copper surface.

3. Discussion

The present study aimed to provide significant insights into the impact of the preparation method and copper loading on the structural and catalytic properties of Cu-modified hydroxyapatite (Cu/HAP) catalysts for the NH₃-SCO reaction. The comparative analysis of Cu features in the samples obtained via the one-pot co-precipitation and post-synthesis deposition methods highlights critical differences in copper dispersion, structure, and catalytic efficiency.

A key finding is that the one-pot co-precipitation method facilitated higher copper surface exposure, as evidenced by the XPS results showing an increased surface copper concentration (2.5–3.5 at.%) compared with post-synthesis deposition (1.5–1.9 at.%). This was attributed to the gradual incorporation of Cu²⁺ into the HAP lattice during synthesis, preventing the formation of large CuO clusters. In contrast, prolonged contact time in the post-synthesis deposition method promoted copper aggregation into larger nanoaggregates, as indicated by the UV–Vis, and reduced the effective catalytic surface area. These results corroborate findings from previous studies on metal-doped hydroxyapatites and other inorganic supports, where uniform metal dispersion enhances catalytic activity [34,35,36].

XPS analysis further confirmed that the Cu/HAP catalysts prepared by one-pot co-precipitation predominantly contained isolated Cu²⁺ species, which are known to enhance catalytic activity in NH₃-SCO due to their ability to participate in redox cycles.

The post-synthesis catalysts, particularly those prepared with extended deposition times, exhibited Cu(OH)₂ and nanoaggregates of CuO-like species, which are often associated with lower catalytic performance, in particular, low selectivity. The presence of a Cu 2p peak at ~935.7 eV in the post-synthesis samples suggests an increased contribution from the CuO/CuCO₃ phases, which are less catalytically active in NH₃ oxidation.

Regarding the catalytic performance, Cu/HAP_OPS emerged as the most effective catalyst, achieving >90% NH₃ conversion at 350 °C with high N₂ selectivity (~84%). This superior performance aligns with the hypothesis that well-dispersed Cu species promote a selective oxidation pathway, favoring N₂ formation over undesired over oxidized by-products such as NO and N₂O. Cu/HAP_OPP, despite having the highest Cu surface concentration (3.53%), exhibited lower overall conversion while maintaining high N₂ selectivity (~89%), suggesting that additional factors, such as the oxidation state and local coordination environment as well as accessibility, influence the overall catalytic efficiency. In particular, considering the HR-XPS peak in the Cu 2p_3/2 region (933–935 eV) and its decomposition in sub-bands (Figure 3), Cu/HAP_OPS presented a single contribution centered at ~933.6 eV, typical of high electron density copper-species, while Cu/HAP_OPP exhibited two different contributions for the 2p_3/2 peak (932.9 and 934.9 eV), indicating a more complex coordinative environment for copper. Additionally, due to differences in the synthesis conditions (Scheme 1), Cu species might have been preferentially located on the surface in Cu/HAP_OPS, while in Cu/HAP_OPC and Cu/HAP_OPP, they were present in both the bulk and surface regions.

Interestingly, the performance of catalysts prepared via post-synthesis deposition varied significantly with copper loading and contact time. Cu/HAP_D,L, with a moderate copper content (3.1 wt.%), exhibited high NH₃ conversion (>90% at 350 °C) and N₂ selectivity (85.3%), which were comparable with the best co-precipitated sample. However, increasing the Cu loading (Cu/HAP_D,H and Cu/HAP_D,2L) did not proportionally enhance the performance, implying the existence of an optimal dispersion threshold beyond which aggregation counteracts the catalytic benefits. This was even more evident in the case of Cu/HAP_D,F. Indeed, although this catalyst exhibited the highest surface copper concentration among the post-synthesis series, according to XPS, it ultimately resulted in the poorest NH₃ conversion and a high production of undesired nitrogen oxides. This behavior may be attributed to the non-optimal dispersion of copper, which, despite being located on the surface, was present in the form of copper oxide nanoaggregates, as suggested by the HR Cu 2p₃/₂ peak analysis.

These findings reinforce the importance of synthesis strategies in tailoring Cu/HAP catalysts for optimal NH₃-SCO performance. Figure 6 summarizes the relationship between the copper phase properties and catalytic performance. Specifically, the observed catalytic ranking can be explained by considering four key factors: copper structuring, surface concentration, accessibility, and catalyst surface area.

The one-pot co-precipitation method proved to be more effective in generating highly dispersed and catalytically active Cu species, although some of these species may have not been accessible as they were accommodated in the bulk. Post-synthesis deposition requires the precise control of the loading and contact time to avoid detrimental aggregation.

The performance of the Cu/HAP catalysts developed in this work was compared with a selection of Cu-based catalysts reported in the literature, as summarized in Table S4. Although a direct comparison between the catalysts listed was challenging due to the differing reaction conditions (such as NH₃ concentration, O₂ content, gas balance, and GHSV/WHSV), some general trends could still be observed. Notably, the Cu/HAP catalysts showed a remarkably promising performance. With a T₁₀₀ (temperature of 100% aNH₃ conversion achievement) ranging from 300° to 450 °C and N₂ selectivity exceeding 80%, these samples appeared highly competitive when compared to state-of-the-art Cu-based catalysts for NH₃-SCO.

For instance, Cu/TiO₂ and Cu/Al₂O₃ showed good T₁₀₀ values (250–300 °C and 400 °C, respectively, Table S4) and high N₂ selectivity (>95%), but under different conditions (e.g., GHSV 50,000 h⁻¹). Other oxide systems, such as CuOx/La₂Ce₂O₇ and CuO–Fe₂O₃, demonstrated even lower T₁₀₀ values (225–425 °C) with selectivity above 80%, but typically operated at a lower GHSV compared with Cu/HAP (e.g., 20,000 h⁻¹ vs. 120,000 h⁻¹).

Therefore, considering the high GHSV, excellent N₂ selectivity, and competitive T₁₀₀ values, Cu/HAP catalysts could represent a highly efficient and viable alternative to benchmark Cu-based systems for NH₃-SCO, even under more demanding reaction conditions.

4. Materials and Methods

4.1. Materials and Catalyst Preparation

Ammonium hydrogen phosphate ((NH₄)₂HPO₄, ≥98% purity), ammonium hydroxide solution (NH₄OH, 28–30% w/w), copper(II) nitrate trihydrate (Cu(NO₃)₂∙3H₂O, >99% purity), and oxalic acid (50 mM, >99% purity) were obtained from Sigma Aldrich (St. Louis, MO, USA).

Calcium nitrate tetrahydrate (Ca(NO₃)₂∙4H₂O, ≥99% purity), lithium hydroxide (LiOH, >95 mM, >95% purity), and hydrochloric acid solution (HCl, 37 wt.%) were purchased from Carlo Erba (Milano, Italy). Ammonium dihydrogen phosphate ((NH₄)H₂PO₄, ≥98% purity) was obtained from Fine Chemicals Inc. (Baytown, TX, USA). Milli-Q^® was obtained from a Merck Millipore purification system (Burlington, MA, USA). Barium sulfate, extra pure reagent, was obtained from Nacalai Tesque Inc. (Kyoto, Japan).

A complexometric indicator solution composed of 4-(2′-pyridylazo)-resorcinol free acid (96% purity from Sigma-Aldrich, St. Louis, MO, USA) at 0.4 mM in ammonium hydroxide (NH₄OH, 30% purity, from Carlo Erba, Italy) and acetic acid (CH₃COOH, pure reagent, from Carlo Erba, Milano, Italy) was employed

Stoichiometric bare HAP with Ca/P = 1.67 was synthesized via a conventional co-precipitation method using diammonium hydrogen phosphate ((NH₄)₂HPO₄) and calcium nitrate tetrahydrate (Ca(NO₃)₂∙4H₂O) as the phosphate and calcium precursors, respectively. Each preparation batch produced approximately 4 g of HAP. In brief, 250 cm³ of a 0.167 M solution of Ca(NO₃)₂∙4H₂O was added dropwise through a peristaltic pump to 250 cm³ of a 0.100 M (NH₄)₂HPO₄ solution, which had been previously stirred and maintained at 80 °C in a 5-necked flask.

The addition rate was set to 1.65 cm³·min⁻¹, and the pH (~10) was kept constant throughout the synthesis by the controlled addition of a 28–30% solution of ammonium hydroxide solution using a Metrohm 736 GP Titrino titrator (Metrohm, Herisau, Switzerland). The resulting suspension was cooled to room temperature and aged under mild stirring for 20 h. It was then filtered and washed with Milli-Q^® water until a neutral pH was reached. The obtained white solid was dried at 80 °C under moderate vacuum for 8 h to remove excess ammonia and further dried at 120 °C overnight.

Seven Cu-containing hydroxyapatite (HAP) samples were prepared, employing two distinct methods:

• Three samples were synthesized via a one-pot co-precipitation procedure, and
• Four samples were obtained through a deposition process on preformed HAP.

Each preparation yielded approximately 4 g of Cu/HAP samples, with a nominal copper content of 3 wt.% or 6 wt.%.

In the one-pot co-precipitation method, a copper precursor salt was introduced during HAP synthesis, allowing copper to be incorporated into the crystal structure. A Cu(NO₃)₂·3H₂O solution and a Ca(NO₃)₂·4H₂O solution were added to an (NH₄)H₂PO₄ solution, which served as the phosphate precursor. The use of (NH₄)H₂PO₄ over (NH₄)₂HPO₄ was preferred to maintain a pH of ~7, preventing copper hydroxide precipitation.

Three variations of the co-precipitation procedure were implemented:

• One-Pot-Superficial (OPS): A 0.167 M Ca(NO₃)₂·4H₂O solution was added dropwise, followed by a 0.176 M Cu(NO₃)₂·3H₂O solution to a 0.100 M (NH₄)H₂PO₄ solution, using a peristaltic pump (1.65 cm³·min⁻¹);
• One-Pot-Co-Alimented (OPC): A single solution containing 0.151 M Ca(NO₃)₂·4H₂O and 0.016 M Cu(NO₃)₂·3H₂O (total metal cations concentration: 0.167 M) was added dropwise to a 0.100 M (NH₄)H₂PO₄ solution using a peristaltic pump (1.65 cm³·min⁻¹);
• One-Pot-Phosphate (OPP): This procedure was identical to OPS but included a final addition of 3 mL of a 2 M H₃PO₄ solution to lower the pH, preventing copper oxide aggregation or precipitation.

The suspensions were cooled to room temperature and aged under mild stirring for 20 min. The resulting solids were recovered by filtration, washed with Milli-Q^® water until a neutral pH was achieved, and dried under vacuum at 80 °C for 8 h, followed by overnight drying at 120 °C and atmospheric pressure. The powders were then calcined at 500 °C (1 °C·min⁻¹) for 1 h.

The synthesized samples were labelled as follows:

• Cu/HAP_OPS (One-Pot-Superficial),
• Cu/HAP_OPC (One-Pot-Co-alimented), and
• Cu/HAP_OPP (One-Pot-Phosphate).

Four additional Cu-containing HAP samples were prepared using a deposition procedure, in which HAP powder was brought into contact with a Cu(NO₃)₂·3H₂O solution. Two operational parameters were varied between syntheses:

• Cu-loading (3 wt.% or 6 wt.%), and
• Contact time (15 min or 48 h).

To prepare approximately 4 g of each sample, pre-dried HAP powder (120 °C, overnight) was added to 250 mL of a Cu(NO₃)₂·3H₂O solution (with concentration of 8.3 × 10⁻³ M for 3 wt.% Cu loading or 1.7 × 10⁻² M for 6 wt.% Cu loading). The mixture was stirred at 40 °C for 48 h. Afterward, the samples were filtered, washed with Milli-Q water, dried overnight at 120 °C, and calcined at 500 °C (1 °C·min⁻¹) for 1 h.

The resulting samples were labelled as follows:

• Cu/HAP_D,L (Low Cu-loading, 48 h contact time),
• Cu/HAP_D,H (High Cu-loading, 48 h contact time),
• Cu/HAP_D,2L (High Cu-loading, 48 h contact time, double deposition), prepared by performing the 3 wt.% Cu deposition procedure twice on the same sample, and
• Cu/HAP_D,F (High Cu-loading, 15 min contact time).

4.2. Catalyst Characterization

The determination of copper loading in the prepared HAP-based samples was performed using a Dionex C-120 ion chromatograph equipped with a UV–Vis detector. The system was fitted with a CS5 column containing a negatively charged resin as the stationary phase.

As the eluent, a solution of lithium hydroxide, 95 mM, purity 99%, from Carlo Erba and oxalic acid, 50 mM, purity ≥ 99%, from Sigma-Aldrich was used. A complexometric indicator solution, consisting of 4-(2′-pyridylazo)-resorcinol free acid (0.4 mM, purity 96%, from Sigma-Aldrich) in NH₄OH (purity 30%, from Carlo Erba) and acetic acid (pure reagent, from Carlo Erba), was added to the eluent to enable the detection of copper ions by the UV–Vis detector.

To obtain a calibration curve for Cu determination, a commercial multi-standard solution containing various metal ions (10 components, 1000 ppm each, from AreaChem, Naples, Italy) was employed to prepare three standard solutions with concentrations of 1, 5, and 10 ppm.

The copper concentration in the samples was double-checked by analyzing both digested solids (after mineralization with HCl 37 wt.%) and filtered solutions.

Specific surface area (SA) and porosity (pore volume and pore size) values were determined by N₂ adsorption/desorption isotherms collected at −196 °C using an automated surface area analyzer (Sorptomatic 1990 by Carlo Erba Instruments, Italy).

Prior to analysis, the sample (ca. 0.30 g, pressed, ground, and sieved to a particle size range of 80–200 mesh) was outgassed at 350 °C for 4 h under vacuum to remove water and other volatile organic compounds adsorbed on the surface. The specific surface area (SA) was calculated using the Brunauer–Emmett–Teller (BET) two-parameter equation in the p/p₀ range of 0.05–0.3. Pore volume and pore size were determined from the desorption branch (0.3 < p/p₀ < 0.95) of the collected isotherms using the Barrett–Joyner–Halenda (BJH) model equation.

Samples for transmission electron microscopy (TEM) were prepared by mild grinding in an agate mortar, suspension in isopropanol, sonication, and deposition on Au grids covered by a holey carbon film. TEM observations were carried out on a field-emission-gun FEI Tecnai G2 F20 super twin electron microscope equipped with a Gatan Slow Scan CCD 974 camera at the Earth Science Department, University of Milan.

The structural characterization was performed by X-ray powder diffraction using an Analytical XPERT PRO powder diffractometer equipped with a copper tube (CuKα = 1.54060 Å). The X-ray source operated at 40 kV and 25 mA under the following conditions: type of scan-continuous; angle (2θ)-5–65°; step size (2θ)-0.033°; scan rate-50 s; and total acquisition time 14 min.

The obtained diffractograms were compared with reference XRPD patterns from the JCPDS-ICDD database.

The average crystal size was calculated by Scherrer’s formula:

D = k∙λ/(B∙cosθ)

where D is the crystalline diameter, k is the shape constant (≈0.9), λ is the radiation wavelength ≈1.5406 Å, θ is the Bragg angle, and B is the full width at half-maximum of the observed peak corrected for instrumental broadening using a standard monocrystalline silicon reference sample.

UV–Vis diffuse reflectance spectra (UV-DRS) were recorded in the 200–2600 nm range using a UV-3600 Plus spectrophotometer (Shimadzu, Kyoto, Japan) equipped with an ISR-603 integrating sphere for solids.

The sample powders were finely pressed into a circular sample holder, which was then positioned in a quartz cuvette and fixed in the integrating sphere. Barium sulfate was used as the reference material (100% reflectance).

Spectra were collected as reflectance (R%) and then converted into absorbance (Abs) using the following equation:

Abs = −log(R%/100)

The obtained spectra were decomposed in the 200–2000 nm range using OriginPro 8 software, employing a combination of Gaussian functions.

The surface composition of the samples was analyzed by X-ray photoelectron spectroscopy (XPS) using an M-PROBE Surface Spectrometer (Surface Science Instruments SSI, Mountain View, CA, USA) with an Al (Kα) source, featuring a spot size adjustable from 0.15 mm to 1 mm in diameter. Measurements were conducted at an applied voltage of 10 V under a vacuum of 10⁻⁷–10⁻⁸ Torr. Survey scans were recorded in the 0–1100 eV binding energy range, using a spot size of 800 μm and an energy resolution of 4 eV (scan rate: 1 eV per step). ESCA Hawk software was used for data processing. All resulting binding energy values were corrected using the C 1s peak (C–C), fixed at 285 eV, as a reference. The C 1s photopeak comprised four components: C–C, C–O, and C=O bonds as well as a specific feature corresponding to CO₃²⁻, with respective binding energies of 284.6, 286.5, 288.5, and 290.0 eV (see Figure S5).

The normalized surface concentrations of the detected species were computed by excluding the contribution of adventitious carbon (C–C at 284.6 eV).

4.3. Catalytic Tests

The catalytic performance of all samples in the NH₃-SCO reaction was evaluated in a dedicated stainless-steel continuous reaction line equipped with a set of mass flow controllers (Bronkhorst, Hi-Tech Instruments, Nijverheidsstraat, The Netherlands), a tubular vertical electric oven (Controller-Programmer type 818, Eurotherm, Como, Italy), a quartz tubular catalytic microreactor (5 mm i.d.), and an online FTIR spectrophotometer with a DTGS detector (Bio-Rad, Hercules, CA, USA) for the qualitative and quantitative determination of the fed and vented gaseous species.

In a typical experiment, a fixed amount of catalyst sample (ca. 500 mg) was pressed, ground, crushed, and sieved to obtain particles in the 45–60 mesh range (0.35–0.25 mm), then dried at 120 °C overnight. The catalyst pretreatment was then performed in situ under an air flow at 120 °C for 30 min.

The catalytic activity was studied as a function of temperature in the 120–500 °C interval, maintaining each temperature for 60 min to ensure the attainment of steady-state reaction conditions.

Ammonia was fed at a fixed concentration (300 ppm) and diluted in synthetic air containing 1000 ppm of O₂. The total flow rate was 6 NL/h⁻¹ (GHSV was 120,000 h⁻¹).

The gas mixture vented from the reactor was monitored by an online FTIR consisting of a multiple-reflection gas cell (with 2.4 m path length; resolution, 2 cm⁻¹; sensibility, 1.5; 92 scans per 180 s) to quantify the unconverted reagent and the formed products. The total absorbance of all IR-active gaseous species (Gram–Schmidt) vented by the reactor was continuously recorded as a function of time, while the reaction temperature was varied. The main detected species—NH₃, NO, N₂O, and NO₂—were quantified by considering the peak height of a selected absorbance line (Figure S6) by using the measured calibration factors. Details on the calculations are reported in the Supplementary Materials (Table S5).

5. Conclusions

This study demonstrated that the method of copper incorporation into hydroxyapatite has a profound impact on the catalyst performance, not just in terms of dispersion and structure, but also in shaping reaction pathways and efficiency in NH₃-SCO. The observed differences between one-pot co-precipitation and post-synthesis deposition highlight the importance of synthesis strategies in controlling catalytic behavior. These findings reinforce the broader principle that material design at the nanoscale can dictate macroscopic functionality, particularly in environmental catalysis.

Overall, this work underscores the viability of Cu/HAP as a sustainable NH₃-SCO catalyst, offering a balance between activity, selectivity, and cost-effectiveness. By refining preparation strategies, particularly in promoting Cu dispersion and preventing aggregation, further advancements in catalytic efficiency and durability can be achieved, contributing to more effective ammonia abatement technologies.