Active Sites in Low-Loaded Copper-Exchanged Mordenite: Spectroscopic and Stability Study for Methane Oxidation Using Mild Conditions

Abstract

Low-loaded copper-exchanged mordenite samples (3 wt.% of copper) were prepared by a solid-state milling method using controlled conditions. The milled samples were then submitted to a calcination process where trimeric copper active species were formed, according to XPS, EPR, IR, and UV–vis recorded spectra. To verify the interaction of the active site with methane at mild conditions, a test experimental design was developed in a batch reactor configuration using mild two-step conditions: (1) activation temperature at 400 °C in an air atmosphere, and (2) isothermal conversion process at 200 °C with 6 bar methane. The analyzed samples were active in methanol conversion in batch conditions, nonetheless less efficient than the usually reported copper mono $\mu$-oxo sites using harder experimental conditions. The herein reported copper active sites are as follows: a trinuclear copper active cluster [Cu₃($\mu$-O)₃]²⁺ and a possible intermediate during methane contact detected as bis($\mu$-oxo) dicopper species were identified and studied on each reaction step. This study revealed that trinuclear copper active sites can be obtained through grinding. Nonetheless, they stabilize after a calcination stage in an air atmosphere. Their stability is then maintained during the whole cyclic experimental test, suggesting their potential use for multicyclic processes.

1. Introduction

The development of new technologies for utilizing methane for various purposes stems from its high global availability and the potential for being produced from renewable sources, yet it remains a challenge. In this context, the use of novel materials able to transform methane into easily condensable energy carriers, useful for the chemical and manufacturing industries, is unpostponable [1,2,3,4,5,6]. Methanol, as a chemical feedstock and an energy vector, has gained significant importance and is sought to be produced from methane, ruling out the standard known conversion methods: syngas production followed by the use of zeolite catalysts and direct hydrocarbon synthesis employing Fischer–Tropsch processes [1,7]. Methane displays a higher stability compared to any of its partial oxidation products; it requires considerable activation energy for the homolytic (105 kcal/mol) or heterolytic (400 kcal/mol) cleavage of the C-H bond, and it has a high pKa (>40) [8], making it a challenge to obtain high selectivity toward methanol at significant conversion rates (up 5% rate conversion with 80% of selectivity) [9,10]. Moreover, methanol is more reactive than methane, resulting in its overoxidation into formic acid and carbon oxides [8,11,12,13,14,15]. However, in nature, methanotrophs can convert methane to methanol at room temperature and under aerobic conditions. These organisms own a membrane-bound metalloenzyme with a catalytic copper center, known as particulate methane monooxygenase (pMMO), which is capable of catalyzing the oxygen insertion in the methane C-H bond at elevated rates [16]. These enzymes are proposed to contain copper dimers and trimers as active clusters, which are the responsible sites for the conversion process in which reactions of interest occur. The clusters in the pMMO enzymatic structure are keynotes for the development of bio-mimetic catalysts useful for methane activation [11,12,16,17].

Mordenites are widely employed as catalyst supports due to their porous open framework, large surface area, and ionic exchange properties, with industrial-level applications [18,19,20,21]. The metal insertion by the ionic exchange method modifies the physical and chemical properties of mordenite [22]. This insertion can mainly be produced in a liquid or gaseous phase or under a mechanochemical reaction [12,19]. However, the election of the insertion route also affects the methane-to-methanol conversion ratio [23]. Solid-state ion-exchange methods carried out far from thermodynamic equilibrium conditions produce local high-energy hot points by microcrystal collisions where the occurrence of local phase transitions, redox reactions, formation of solid solution, and metastable phases are possible [5,24]. Particles fracturing under milling, specifically those that can stabilize metallic clusters, can induce solid activation through the appearance of structural defects and active sites [19], enhanced by the dehydrated synthesis environment.

Over the years, different mononuclear, binuclear, and trinuclear copper active sites have been proposed to participate in the activation of methane to produce methanol. However, there is no definitive consensus on their structure nor the involved mechanisms for methane activation [25,26]. For instance, Rosenzweigh et al. proposed a catalytic complex of a copper dimer [16,27], while Chan et al. suggested the existence of trimer copper species [28,29]. From extensive DFT studies and spectroscopic data, monocopper [30,31,32], dicopper [10,33,34,35], and tricopper [36,37] Cu-oxo species have been studied and proposed as active sites for methane activation using Cu-exchanged mordenite. The proposed species enables methane activation through the homolytic cleavage of the C-H bond, utilizing radical-like oxygen atoms to form methoxy intermediates. These intermediates subsequently rebound to the hydroxyl species, forming methanol. This methanol production is achieved through cyclic chemical looping procedures, which provide high control over material oxidation and reaction with methane. The above, and the effect of the amount of copper loaded within the zeolite, could explain the high reaction selectivity towards the methanol formation [38,39,40]. However, the conversion ratio is affected by many factors, as it is known to depend on the Cu salt used [41], Cu/Al ratio [5], interchange [42], activation [6] processes, and zeolite type [22], among others.

This contribution discusses the stability of copper active sites through an experimental test to validate their performance at mild reaction conditions. The Cu-exchanged mordenite samples were prepared through a solid-state mechanochemical ionic exchange reaction using copper(II) acetylacetonate as the copper source, followed by calcination in an air atmosphere at 400 °C. The copper-containing active sites were identified through UV–vis, XPS, IR, and EPR spectra, complemented by ICP and thermogravimetric data. These techniques provided information on the activation reaction effects on the identified copper species and their changes at the different stages of the methane reaction. The results discussed herein revealed that trinuclear copper active sites appear to be active in methanol conversion. However, they are less efficient than the usually reported copper mono $\mu$-oxo sites using harder experimental conditions. These trinuclear copper active clusters, [Cu₃($\mu$-O)₃]²⁺, can be obtained through the grinding process, and they stabilize after a calcination process using an air atmosphere. Such stability is maintained throughout the entire cyclic experimental test, suggesting their potential for use in multicyclic processes, although the material was not tested for robustness over multiple cycles.

2. Materials and Methods

2.1. Synthesis of Materials

All the chemicals used were purchased from reagent providers and employed without further purification. Commercial NH₄-mordenite (NH₄-Mor) was employed with a Si/Al mole ratio of 20 and a simplified molecular formula SiO₂:AlO₃ from Alfa Aesar (ThermoFisher Scientific, Waltham, Massachusetts, USA, CAS: 1318-02-1). Copper (II) acetylacetonate (Cu(C₅H₇O₂)₂, Cu(acac)₂) from Sigma-Aldrich (St. Louis, Missouri, USA, CAS: 13395-16-9) with 99.9% purity was used as the cupric ion source.

The exchange percentage and calcination conditions were chosen based on the data reported in similar works [38,40,43,44,45,46] and on our previous analysis [47]. The cation exchange between NH₄ and Cu²⁺ was performed using solid-state milling. An amount of 900 mg of NH₄-Mor was mixed with 126.9 mg (Cu 3 wt.%) of Cu(acac)₂ and then ground using a ball mill during an hour (Laboratory Mini Mill Pulverisette 23 from Laval Lab), 40 Watt power, in an air atmosphere. Stoichiometric calculations determined the amounts of copper. Homogeneous samples were obtained.

Afterwards, the samples were calcinated to form the active sites within the mordenite porous framework; the composition containing Cu 3 wt.% was calcined for two hours using an air atmosphere and a convection muffle at 400 °C (heating rate of 10 °C min⁻¹), resulting in the Cu3%400T sample.

2.2. Reaction Test: Testing of Activity for Selective Oxidation of Methane

The interaction of the active sites with methane was tested using a 500 mL Batch Reactor System, following a cyclic reaction procedure as reported by Tomkins et al. [38]. The method was set up as follows: 250 mg of the synthesized powdered copper-exchanged mordenite was colocated inside the batch reactor. (1) A calcination activation step was applied in an air atmosphere, as described in Section 2.1. (2) Immediately after, the air atmosphere was removed by a flow displacement and replaced with 6 bar of gaseous methane, obtaining the Cu3%400T6P sample. The reaction experiments with methane involved an isothermal heating process at 200 °C, which was applied for two hours to avoid methane combustion [33,48]. (3) Subsequently, the reactor was cooled to room temperature at a rate of 10 °C min⁻¹. Then, 50 mg of the recovered reacted samples was dispersed in 1 mL of deionized water and stirred vigorously for 30 min to extract the reaction products. (4) Finally, after centrifugation and filtration (2000 rpm, 2 $\mathsf{\mu }$m filter), the liquid phase was analyzed by gas chromatography–mass spectrometry (GC-MS) to verify the ability of the studied copper cores to convert methane to methanol. During this water desorption step, water may act as both an extraction solvent and as a source of oxygen in the direct conversion of methane to methanol over the copper-exchanged mordenite [49,50].

2.3. Materials Characterization

Powder X-ray diffraction (XRD) powder patterns were recorded with CuK$\alpha$1 ($\lambda$ = 1.54056 Å) radiation in a D8 Advance diffractometer from Bruker, equipped with a Lynx Eye Detector, using a Bragg-Brentano geometry. The induced coupled plasma-optical emission spectroscopy (ICP-OES) technique was employed to corroborate the copper weight percent inside the synthesized calcined materials. For this analysis, a Perkin Elmer ICP-OES Optima 8000 model was used. Winlab software was used for data analysis. An emission line of 327.393 nm was used for the copper analysis, with a calibration curve in the 0–70 mg/L range. UV–Visible spectroscopy (UV–vis) spectra were collected in a diffuse reflectance mode with a Perkin Elmer spectrometer equipped with an integrating sphere. Spectra were recorded from 200 to 800 nm. Results are presented in the Kubelka–Munk function plot of the acquired reflectance data. Electron Paramagnetic Resonance (EPR): A Bruker equipment model ELEXSYS II E500 with an X band (9.5 GHz) continuous wave was employed. The samples were analyzed at a temperature of 90 K, with a sweep time of 30 s and a power of 1 mW. Infra-Red (IR) absorption spectra were recorded in a PerkinElmer RX-1 spectrometer using an ATR accessory. Thermogravimetric (TG) data were recorded under a nitrogen flow (1 L/min) in a high-resolution mode where the heating rate was controlled by the sample weight change, using a TA Instruments IR 500 equipment. X-Ray Photoelectron Spectroscopy (XPS) measurements were performed using a Thermo Scientific K-Alpha X-ray photoelectron spectrometer working at 72 W and equipped with a hemispherical analyzer and a monochromatic AlK$\alpha$ X-ray source (1486.6 eV) in the Constant Analyzer Energy (CAE) mode. The base pressure in the analyzer chamber was 1 × 10⁻⁹ mBar. Survey scans were recorded using a 400 $\mathsf{\mu }$m spot size and a fixed pass energy of 200 eV, whereas high-resolution scans were collected at 20 eV of pass energy. The charge corrections of all spectra were referenced to the position of the C1s adventitious peak at 284.8 eV. Gas Chromatography/Mass Spectrometry (GC-MS): Aqueous samples analysis was performed in a Clarus SQ 680 gas chromatograph coupled to a Clarus SQ 8T mass spectrometer (Perkin Elmer, Boston, MA, USA) using a PE-WAX (Pekin Elmer) capillary column (50 m × 0.25 mm × 0.25 $\mathsf{\mu }$m phase thickness). As an internal standard, 1 $\mathsf{\mu }$L of isopropyl alcohol (IOH) was used to identify methanol or other alcohols as by-products. The temperature program was set as follows: 40 °C for 8 min, 20 °C/min up to 80 °C, 40 °C/min up to 90 °C, and 50 °C/min up to 180 °C. The injection temperature was 150 °C, with helium used as the carrier gas at a flow rate of 1.5 mL min⁻¹. The mass spectrometer was operated in an electron impact mode (70 eV), with selective ion monitoring (m/z 29, 31, 43, 45, 46, 58, 59, and 60) at a dwell time of 0.05 s, while maintaining the chromatograph interface and source temperature at 280 °C. Methanol quantitation (m/z 29 + 31) and IOH (m/z 45) were made using a five-point calibration curve (R > 0.99) and the areas of specific ions mass-chromatograms.

3. Results and Discussion

3.1. Sample Characterization—Copper Species Evidence

ICP-OES analysis confirms that, within the expected error of this technique, most of the copper added to the milled NH₄-Mor + Cu(acac)₂ mixture was incorporated into the zeolite after calcination. The assumption of the ion exchange process (3.1% by ICP-OES) was confirmed by XRD, UV–vis, IR, TG, XPS, and EPR data (discussed below).

XDR showed no significant structural changes relative to the original NH₄-Mor (Figure S1, in the Supplementary Materials) during the ionic exchange or after heat treatment. This means the mordenite framework was preserved for the activated samples of copper-exchanged mordenite. Furthermore, no appearance of additional peaks related to copper oxides or sintered copper species is evident, which suggests that the formed copper species have a size smaller than 3 nm and/or are well dispersed throughout the mordenite framework [12]. The relatively small effective ionic radius of the copper(II) ion favors the solid-state ionic exchange reaction, allowing its diffusion throughout the material framework. This means that Cu reaches the structural positions where charge balancing occurs and where active sites are formed [51,52].

The TG curves, IR spectra, and XPS spectra also confirm the stability of the calcined-exchanged sample and the methane-reacted sample. IR spectra for the parent NH₄-Mor, exchanged (NH₄+Cu(acac)₂), activated (Cu3%400T), reacted (Cu3%400T6P), and washed (Cu3%400T6PH2O) samples (Figure 1a) are very similar. Distinctive bands corresponding to symmetric stretching from a single 4-membered ring (at 634 cm⁻¹) and double-ring (at 585 and 568 cm⁻¹) vibrations suggest that mordenite crystallinity is preserved. The evolution from a doublet to a single broad band in the IR spectra between the NH₄-Mor+Cu(acac)₂ mixture and the calcined Cu3%400T sample is consistent with the thermal decomposition of Cu(acac)₂ and subsequent incorporation of Cu into the zeolitic framework. This process reduces the local asymmetry of the Cu coordination environment, likely due to the formation of $\mu$-oxo bridges or framework-bound Cu species, leading to a spectrally merged signal in the observed region (Figure 1a, inset). Moreover, the presence of a slight intensity decrease and red shifts (∼5 cm⁻¹) from methane-reacted samples suggests the appearance of an environment distortion related to the copper insertion (see Figure 1a, inset). Likewise, as depicted in Figure 1a, two principal subtle changes in IR spectra are evident at the different stages of the test conversion cycle (see Section 2.2 and Section 3.2). These changes correspond to the region near 1400–1450 cm⁻¹ and 3300–3700 cm⁻¹, which are related to the ammonia cation, and terminal silanols and O-H vibrations disturbed by hydrogen bond (intrazeolite H-bonds) mordenite groups [53], respectively. After ion exchange, the two small bands near 1400 and 1450 cm⁻¹ disappear, suggesting a destabilization of the ammonia groups belonging to the parent mordenite [54,55]. Similarly, an evident decrease from bands at 3726 and 3472 cm⁻¹ is noticed, advising that some of the protons from the bridged Si-OH groups were replaced by copper ions [52].

The TG curve from the Cu3%400T sample (Figure 1b) is also in favor of the material stability after air oxidation treatment with almost no residual acetylacetonate (C₅H₇O₂); compared to the parent mordenite and the NH₄-Mor + Cu(acac)₂ mixture (see Figure S2, Supplementary Materials), a slightly smaller hydration degree (∼1.7%) on the exchanged samples is observed as expected for low-copper loaded species. To identify those copper species, the nature of the Cu species ligands can be determined using XPS, where such information can be obtained via the binding energy (BE).

Figure 2 shows the XPS spectra for the parent mordenite, exchanged mordenite Cu3% (NH₄-Mor + Cu(acac)₂ mixture), and the activated/calcined Cu3%400T sample, where a lack of copper species is evident for the parent mordenite. This is in agreement with the expected BE corresponding to the O1s region composed by two main signals: the primary distribution in 531.2 eV associated with the oxygen atoms from mordenite and a minor contribution related to OH species located at 532.6 eV (Figure 2a).

On the other hand, the photoemission profiles of the XPS spectrum from the sample Cu3% Cu2p_3/2 (Figure 2b) were built by three different copper species, which were identified based on information from similar studies. The first one corresponds to Cu₂O and Cu(I)-Cu(II) from Cu(acac)₂, having different coordination, and the last one is attributed to an emergent $\mu$-oxo species type formation at 932.5 eV [56], 932.7 eV [57], 933.8 eV [58], 934.2 eV [57], 936.3 eV [57,59], and 936.44 eV [19]. To estimate the shift of the electron BE as a function of the copper oxidation state and local structure, Artiglia et al. carried out theoretical calculations of the electron BE of Cu 2p electrons in various possible active species. The calculated BE of copper in different reference bulk compounds (932.3 eV for metallic copper, 932.5 eV for Cu(I) oxide, and 934.2 eV for Cu(II) oxide) is in good agreement with the literature [60]. Furthermore, the simulated BE of mono($\mu$-oxo) dicopper sites, having a formal oxidation state of +2, is 934.3 eV. In contrast, it shifts to 936.1 eV for bis($\mu$-oxo) dicopper sites. On the other hand, the tricopper sites exhibited two BE values, i.e., 934.8 eV and 935.9 eV [61]. It is crucial to note the emergent formation of $\mu$-oxo species related to the grinding process, where local high-energy hotspots may arise from microcrystal collisions, which may also be associated with an unidentified intermediary species, explaining the shift observed from the reported Cu(acac) species. Furthermore, the Cu3% O1s XPS spectra verified the parent mordenite stability, showing the 532.2 eV peak slightly shifted from the one depicted in Figure 2a presumably because of the low-loaded copper incorporation as evidenced by the small contribution at 530.6 eV attributed to Cu₂O species corresponding to UV–vis and EPR characterizations (see below). Figure 2c shows the XPS spectra from the Cu3%400T sample after temperature treatment. In this spectrum, in the Cu2p region, the Cu(acac) attributed contribution vanishes as suggested by thermogravimetry, in which very low acetylacetonate residuals were detected. Herein, the remanent peaks are attributed to a possible $\mu$-oxo species type at 935.9 eV. Moreover, the principal peak located at 933.2 eV is ascribed to three different possibilities: (1) Cu(I) species that belong to Cu₂O, as no satellite associated to Cu(II) species is observed, or (2) reduced $\mu$-oxo Cu sites due to high-vacuum experimental procedures, or (3) isolated Cu²⁺ ions coordinated to the 8-membered ring (MR) cage of MOR. In this case, the peak shows a negative shift of the BE (933.5 eV) by 0.8 eV compared to that of mono($\mu$-oxo) dicopper sites. This is explained by the absence of extra-framework Cu-O bonds (bridging oxygen in the dicopper species that facilitates partial charge transfer onto oxygen atoms) [57]. The literature reports that Cu(I) may oxidize to Cu(II) at temperatures above 350 °C in air, and the potential overlap of this BE with CuO, so this signal may also include contributions from Cu(II) species. Nonetheless, the absence of satellite peaks typically associated with CuO suggests a non-bulk, likely highly dispersed Cu(II) environment or mixed-valence $\mu$-oxo type clusters. On the other hand, in the O1s region, a new emergent contribution is identified at 534.1 eV, typically associated with adsorbed water molecules. The appearance of this signal for the calcined sample is attributed to the re-adsorption of ambient water vapor onto newly formed Lewis acid sites or unsaturated Cu centers exposed during calcination. These sites are highly hydrophilic and can bind water molecules upon exposure to room conditions. According to UV–vis (Figure 3), the peak located at 935.9 eV could be ascribed to the presence of mono($\mu$-oxo) dicopper or trimeric (II) species where an extra-framework oxygen atom participates because the spectroscopic UV–vis region in which extra-framework oxygen copper species usually appear is in 20,000–33,000 cm⁻¹ [8,36,62]. In this context, the mono-($\mu$-oxo) dicopper(II) complex [Cu₂($\mu$-O)₃]²⁺ has been identified in Cu-mordenite as a broad band located from 22,200 to 22,700 cm⁻¹ [12,19,48,57,63,64], and interpreted as a charge transfer ${\mathrm{O}}_{bridge}$ → Cu²⁺ transition [19]. Nevertheless, Grundner et al. [9,36] recently proposed a broad band near 31,000 cm⁻¹ associated with a low-symmetry single trimeric site [Cu₃($\mu$-O)₃]²⁺ different from those of all species proposed before [63,64,65] and formed in Cu-mordenite, corroborating the idea that species formed in Cu-mordenite have not been fully understood [12].

3.2. Test Reaction: Methane Oxidation Using the Materials Under Study, the Performance of the Active Sites Using Mild Conditions

As described in the previous section, using mild reaction conditions, a test reaction assessed the performance of the active sites (species described in Section 3.1) within the mordenite framework as reactive species for methane partial oxidation. The herein discussed cyclic reaction conversion process, with batch reaction conditions, involves three main steps: (i) the activation of Cu-mordenite at 400 °C (already discussed), (ii) reaction with methane at 200 °C (isothermal process), and (iii) washing with liquid water at room temperature to extract the formed methanol [12,39,43]. The samples were analyzed at each stage during the reaction process.

Using the GC-MS technique, the water used to wash the samples that reacted with gaseous methane was analyzed using the methanol standard and program temperatures described in Section 2.3 of the experimental part, with the measurements focused on methanol quantitation. They yielded a very low conversion rate of 0.0632 $\mathsf{\mu }$mol/grCu, showing that Cu-mordenite can stabilize different dehydrated copper species that are less reactive than the $\mu$-oxo dicopper(II) core, which requires a higher methane pressure to achieve the reduction of copper oxide. Such species, under mild conditions, could hinder the reaction cycle by blocking access to the structural region where the active copper species are formed (mouth-side pocket, eight-member ring [36,63,66]). The relatively low-obtained methanol in this study (24,745–29,840 $\mathsf{\mu }$moles per gram of catalyst) [47] could be ascribed to the low quantity of active species and the used static reaction conditions, which had already been reported for different experimental configurations [12,38,39,43] but also because of the Si/Al ratio, which is known to chance the catalytic properties of mordenite [67].

3.3. Active Sites Characterization After Test Reaction

Electronic spectroscopy diffuse reflectance spectra were recorded following the steps of the cyclic process for the herein considered experiments, as established in the reaction test. After sample activation at 400 °C, a well-defined band in the 31,000 cm⁻¹ region related to the formation of Cu(II) trimeric species is evident for the activated/calcined sample (Figure 3), suggesting the formation of a population of these active sites within the material framework.

This band is maintained after methane contact, and the same intensity ∼23,000 cm⁻¹ band is also evident during this stage. These results are similar to the bands detected in previous UV—vis spectra ascribed to the ELO (Extra Lattice Oxygen) → Cu CT transitions of the bis($\mu$-oxo)dicopper [Cu₂-($\mu$-O)₂]²⁺ core [68] but not detected in other steps during the cyclic test procedure. This result suggests that the interaction of methane with the Cu-activated sample causes a partial conversion of Cu(I) to the bis($\mu$-oxo) dicopper core. Nevertheless, the amount of the bis($\mu$-oxo) dicopper complex vanishes after washing with liquid water at room temperature to extract the formed products, suggesting that this site is an active or intermediate core during methane interaction, as indicated by the XPS’s small shift in the reacted samples. After methane contact, a tiny band located at 38,000 cm⁻¹ is depicted in Figure 3, whose behavior corresponds to the conversion cycle as follows: it tends to increase during the methane contact process and disappears during the washing process. The bands starting at 35,000 cm⁻¹ are assigned to mordenite → Cu(II) CT transitions, in agreement with UV–vis data in the literature [19].

The bands assigned to a mono ($\mu$-oxo) dicopper(II) complex were not observed in the samples under study, which were prepared using an air atmosphere during the activation step process. The region beyond 15,000 cm⁻¹ was assigned to an octahedral [19] Cu(II) with water molecules as ligands, suggesting the possibility of formation of different copper (II) species within the same material, and uncompleted dehydration of the samples, and no d-d transitions associated with the mono-($\mu$-oxo) dicupric site were observed, which are expected for samples with a Cu/Al ratio > 0.20 [69]. The proposed Cu(II) trimeric species [Cu₃($\mu$-O)₃]²⁺ seemed to be highly stable under dry and room conditions due to the synthesis process through all the cyclic process, which was also confirmed by XPS-reacted samples (see Figure S4 and Table S1 in the Supplementary Materials), where no significant changes are observed compared to the spectra shown in Figure 2c. After the methane reaction, the XPS spectra obtained from the Cu3%400T6P sample showed a slight 0.1 eV shift in the Cu2p region, likely due to an intermediate species or molecule added to a free copper bond, also evident in the IR (Section 3.1). Nevertheless, this assumption is difficult to maintain due to the ultra-high vacuum inside the chamber, where intermediate adsorbed molecules (if available) could be desorbed. A similar behavior is observed for the Cu3%400T6P washed sample, where the most significant change is noted in the O1s region, indicating an increase in water contribution resulting from the washing process.

3.4. EPR Spectra on the Copper Detected Sites

The EPR spectra shape line recorded at 90 K for all the Cu3% family samples obtained during exchange and test with methane reaction is shown in Figure 4. After the thermal activation process (Figure 4c), a characteristic signal associated with Cu(II) attributed to [Cu₃($\mu$-O)₃]²⁺ species is depicted as suggested by XPS and UV–vis spectra in Section 3.1. These signals are consistent with an S = 1/2 system that exhibits axial anisotropy with well-defined values for g_‖, g_⊥ and A_‖ (

Table 1. EPR g-value and hyperfine constants of c, d, and e spectra in Figure 4; corresponding to activated, reacted, and washed samples of Cu3% family samples.

EPR Spectrum	Experimental Sample	g‖ Value	g⊥ Value	A‖ Hyperfine Constant (cm−1)
c	Cu3%400T (activated)	2.385	2.099	138.63 × 10−4
d	Cu3%400T6P (reacted with methane)	2.379	2.097	138.46 × 10−4
e	Cu3%400T6P + H2O (washed with water)	2.391	2.114	136.00 × 10−4

). This kind of EPR spectrum is consistent with an antiferromagnetically coupled trinuclear group formed by Cu ions with 1/2 and 3/2 electronic and nuclear spin, respectively, corresponding with the XPS spectra from the calcined Cu3%400T sample. The exchange interaction between Cu(II) ions is crucial in explaining the EPR spectra of this cluster. The effect of exchange interaction on the EPR properties of the ground state S = 1/2 for this type of cluster, and similar clusters, has been well documented in the literature [70,71,72]. The exchange interaction induces an effective anisotropy in the g tensor of the ground state. In the X band, the EPR spectra are characterized by an absorption peak in g_‖ and another with g_⊥. Considering the values for g_‖, g_⊥, and A_‖ from Table 1, we can associate the EPR spectra obtained as coming from trinuclear Cu(II) S = 1/2 clusters, as has been performed before [73,74,75]. Assuming that it is possible to indistinctly label the Cu(II) ions of the trimer as Cu(1), Cu(2), and Cu(3), the exchange coupling between Cu(2) and Cu(3) forces these ions to align themselves in an antiparallel array, producing a total spin ${\mathrm{S}}_{Cu\left(2\right)-Cu\left(3\right)}$ = 0 and resulting in the Cu(1) ion as “free”. In this sense, the complete cluster behaves like an isolated Cu(II) ion [74].

Analysis of hyperfine lines in the g_‖ provides valuable information on the relative magnitude of the exchange coupling constants. In the present case, well-resolved hyperfine features are obtained in the samples studied at 90 K. The hyperfine coupling was not detected in NH₄-Mor, a result expected due to the absence of copper in the original mordenite sample (Figure 4a). Neither the NH₄-Mor + Cu(acac)₂ mixture (Figure 4b) shows peaks corresponding to the hyperfine interactions, except undefined shoulders, attributed to an effect of water molecules and slight traces of copper from Cu(acac)₂, more evident with the KBT increase, which causes a broadening in these signals.

A hyperfine contribution is depicted in the Cu3%400T activated sample (Figure 4c) as the appearance of four signals associated with the interaction between S = 1/2 and I = 3/2. After the calcination treatment of the sample, the hyperfine signals increase compared to the behavior observed for the CuMor3%400T6P reacted sample (Figure 4d), corresponding to a stronger nuclear spin interaction with the electronic spin, suggesting that the covalence degree is lower in the reacted sample since the electron is more probably found in the copper neighborhood as indicated by XPS and UV–vis spectra from reacted samples. The behavior of the hyperfine features could mean that the unpaired spin is primarily located in a single Cu center within the trinuclear unit.

For the spectrum of Figure 4c, the hyperfine constant (A = 138.63 × 10⁻⁴ cm⁻¹) can be associated with the trimer copper active sites bonded to framework and extra-framework oxygen atoms [36] as suggested by XPS and UV–vis spectra in Section 3.1, present in the activated CuMor3%400T sample, remaining the electron in the vicinity of copper atoms. In the sample reacted with methane (CuMor3% 400T6P, spectrum in Figure 4d), the decrease in the g_‖ and g_⊥ values are associated with a change in the first neighbors’ surroundings as predicted by XPS, IR, and UV–vis spectroscopies; specifically, it can be related to methane-to-methanol intermediate species located on the active sites (directly bonded or stabilized by extra-framework oxygen). This organic species locally alters the active site’s electronic structure, as suggested by UV–vis, where the bis($\mu$-oxo) dicopper core is detected. A reduction in the hyperfine signal suggests that one electron leaves the vicinity of the copper atom, approaches the oxygen atom, and modifies the chemical environment of the active site. Finally, the washed sample (CuMor3% 400T 6P H₂O, spectrum in Figure 4e) shows an important decrease in the hyperfine coupling signal and an increase for the g_‖ and g_⊥ values, which can be associated with a higher structural deformation produced by the remaining water molecules because of the washing process.

According to UV–vis spectra (see Figure 3), the presence of water molecules through the material would allow superficial methoxy species to become hydrolyzed into molecular methanol during the step of reaction with methane [69]. Active sites with high g-values (2.37–2.40) have been observed in Cu-mordenite and have been associated with the presence of square pyramidal and square planar Cu²⁺ sites [64,68,76]. Owing to the low symmetry of the trimeric copper cluster identified by UV–vis spectroscopy and the copper species found by EPR, it is probable that Cu atoms in the cluster are not equivalent to each other, in concordance with reported works using DFT theory [36,66,77]. There is no reason to believe that some silent EPR species of copper(I) cannot be formed during that process. An exhaustive EPR-UV–vis study, conducted at different temperatures and frequencies, could shed light on corroborating the existence of distinct, previously unreported Cu species.

4. Conclusions

The present contribution discusses the stability and activation of copper active sites in low-loaded copper-exchanged mordenite during the partial oxidation of methane to form methanol, based on experimental tests performed under mild conditions in a batch reactor. A principal contribution of trimer copper species is formed within the mordenite framework during the activation (calcination) process, as identified from the recorded UV–vis, XPS, and EPR spectra. Their modification and stability were then tested through interactions with methane. By the solid-state ionic exchange reaction used herein, combined with subsequent calcination of the milled mixture, the formation of trimeric active copper species and a possible intermediate bis($\mu$-oxo) dicopper core during methane reaction was detected, exhibiting significant stability at room conditions. The formation of the usually reported [Cu₂($\mu$-O)]²⁺ species was not observed in the activation step within the temperature herein applied, presumably due to the Si/Al ratio for the mordenite used and the considered experimental conditions. The experimental evidence indicates that solid-state ionic exchange and an activation step of 400 °C produce [Cu₃($\mu$-O)₃]²⁺ species as active sites, which is corroborated by the appearance of a UV–vis band at 31,000 cm⁻¹, a solely axial EPR signal (g_‖ = 2.385, g_⊥ = 2.099, A_‖ = 0.013863 cm⁻¹) and the Cu2p and Cu2p_3/2 XPS spectra. Nevertheless, according to the spectroscopic results, during the methane reaction stage, other copper species may participate in methane oxidation to methanol under isothermal conditions; a probable intermediate suggested is the bis($\mu$-oxo) dicopper core. Nonetheless, the principal active trimer copper sites are obtained during grinding. They stabilize after a calcination process using an air atmosphere. This stability is maintained throughout the entire cyclic experimental test, suggesting their suitability for multicyclic processes. A better understanding of the active sites at different low loading percentages, analyzing all the resulting products after methane reaction, and studying different multicycle test reactions are necessary to improve the existing knowledge of methane oxidation under mild conditions.