Selective Low-Temperature Oxidative Dehydrogenation of Propane over Alumina-Supported Copper Nanoparticles with O2 and CO2 as Oxidants

TL;DR

This paper studies how copper nanoparticles on different alumina supports affect propane conversion to propylene at low temperatures using oxygen or carbon dioxide.

Contribution

The study reveals that copper nanoparticles on nano-alumina support achieve higher propylene selectivity and stability with CO2 as an oxidant.

Findings

Cu/nanoAl2O3 achieves 35–48% propylene selectivity at 250–300 °C with O2.

Using CO2 as oxidant boosts propylene selectivity to 100% for Cu/nanoAl2O3.

Copper particle size differences explain performance variations between the two catalysts.

Abstract

This study reports on the performance of alumina-supported copper-based catalysts in the oxidative dehydrogenation of propane, with copper dispersed on two distinct commercial aluminium oxide supports made of micro- and nanosized alumina, respectively. The activity and selectivity of the two catalysts was investigated at temperatures between 250 and 550 °C. At a propane-to-O2 ratio of 1:1, Cu/nanoAl2O3 achieves propylene selectivity of 35–48% at low temperatures (250–300 °C), while Cu/Al2O3 only exhibits activity starting at 350 °C with about 40% propylene selectivity. Altering the propylene-to-oxygen ratio to 3:1 enhances selectivity towards propylene in both catalysts, up to about 64% on Cu/Al2O3 at temperatures of 250–350 °C. The switch to the mild oxidant CO2 boosts propylene selectivity to 100%. In case of Cu/nanoAl2O3, the rate of propylene formation doubles that of the obtained with O2 used as oxidant. While with CO2 the Cu/nanoAl2O3 catalyst retains 100% propylene selectivity up to 500 °C, on the less active Cu/Al2O3 cracking sets off already at 400 °C. The different size of copper particles in the two catalysts is seen as a primary factor determining the observed differences in the performance of the studied catalysts.

Figures

Full text

1. Introduction

Accounting for the vast majority of global olefin output, steam cracking [1] is the dominant technology, typically using naphtha and natural gas components as feedstock. These techniques, however, have high energy demand, are plagued by high CO_2_ emissions, and yield relatively low selectivity for a specific olefin product. As a result, there has been a surge of interest in dehydrogenation (DH) and oxidative dehydrogenation (ODH) reactions, which offer a promising path toward more energy-efficient and environmentally friendly alternatives.

Non-oxidative dehydrogenation is widely applied for producing olefins due to the high selectivity of the target product [2]. For the production of propylene, a commodity chemical in steadily rising demand, propane can be used as the feedstock [3]. However, breaking two C-H bonds requires a high energy input (i.e., high temperature), which makes DH challenging even with the use of catalysts because of factors hampering performance; for example, sintering of the active metal particles and thermal cracking of products leading to coke formation, both eventually leading to the deactivation of the catalyst [4]. In propane DH, the commonly used catalysts are Pt nanoparticles [5,6,7], chromium particles [8,9], ZnO_x_ supported on dealuminated zeolites [10], and Ge, Co and V oxides [11,12,13,14]. Only a limited number of studies have examined copper-based catalysts for propane DH. For example, supported Cu particles on ZrO_2_ were reported to achieve approximately 15% propane conversion with nearly 99% propylene selectivity at 550 °C [15]. In addition, Pt/Cu single-atom alloy catalysts supported on Al_2_O_3_ have been reported to achieve propane conversion of up to 40% with propylene selectivities reaching around 85% at 600 °C, outperforming Pt and Cu alone [16]. At lower temperatures (350 °C), Cu/Al_2_O_3_ catalysts were reported to have a conversion of around 3% with 84% propylene selectivity [17].

Propylene can also be produced via the oxidative dehydrogenation (ODH) route at lower temperatures instead of via DH thanks to the exothermicity of the oxidative process [18]. Moreover, the presence of oxygen in the feed reduces coke formation. On top of that, at lower reaction temperatures, sintering of the catalytic particles may be suppressed as well. Most of the catalysts used in ODH are based on Pt (e.g., supported nanoparticles) and other metals, such as Cr, Ga, and V, as well as their oxides [19,20,21,22,23,24,25]. Pt and CoO_x_ clusters [26,27,28] and Pd-Pt catalytic films [29] have been studied as well, exhibiting high yields of the target dehydrogenated product. Specifically in the ODH of propane, the cluster catalyst, consisting of 10 platinum atoms deposited on an Al_2_O_3_ layer prepared by atomic layer deposition (ALD), possessed high activity and high selectivity to propylene, the latter reaching up to 84%, depending temperature and promotion with SnO [26]. The abovementioned catalysts perform efficiently at high temperatures, typically above 500 °C. The incorporation of small amounts of Cu into Pt/Al_2_O_3_ catalysts has been reported to enhance propylene selectivity, suppress deactivation, and improve resistance to coke formation; for example, an optimised Cu loading increased propylene selectivity to 91% at propane conversion of about 43% [30]. Cu-Cr mixed oxide systems have emerged as promising catalysts for low-temperature propane ODH, with Cu-Cr double-layer thin films exhibiting up to 50% propane conversion and propene selectivities exceeding 90% at 400 °C [31]. Although copper-based catalysts have been much less studied for propane ODH, early work on CuO catalysts in the presence of tetrachloromethane at 450 °C exhibited 10% propane conversion with 55% propylene selectivity [32].

There has been a growing focus on utilising CO_2_, which has served for decades as a carbon source in various chemical syntheses and as an oxidative agent in the ODH of propane. This process has been reported for alkanes, specifically propane via ODH on metal oxide-supported In_2_O_3_ or zeolite-supported chromium oxide and Ga_2_O_3_, with high conversions reaching 84% and propylene selectivity around 90% [33,34,35]. Among catalysts for CO_2_-assisted oxidative dehydrogenation of propane, palladium doped with rare metals (Ga, Sn, In) has shown promising performance with 80–94% propane conversion and up to 20% propylene selectivity [36]. In later studies, copper-based catalysts have been explored for CO_2_-assisted oxidative dehydrogenation of propane; for example, a Cu-doped CeZn catalyst was reported to achieve approximately 47% propane conversion with 57% propylene selectivity at 550 °C [37].

While platinum remains one of the most effective catalysts in this reaction, its high price and scarcity limit its practical application. The use of copper that is 3700 times cheaper as a catalyst for ODH is garnering considerable interest in balancing catalyst cost and effectiveness. Copper catalysts, doped with other metals, such as Pt, Cs, or Pd, achieve product selectivities of 50% to 99% in various oxidative dehydrogenation reactions [38,39,40,41,42], with some copper-based catalysts showing noticeable activity at temperatures of 230 to 300 °C as well in the dehydrogenation of methanol or even propane [39,40], thus further accenting the potential of copper as the catalyst for the ODH of propane. In the oxidative dehydrogenation of, e.g., cyclohexane, in comparison with Pd clusters deposited on ultrananocrystalline diamond (UNCD) layer, Cu clusters exhibit higher selectivity towards a partially dehydrogenated product, cyclohexene (35%), while Pd primarily produces benzene (40%) at 300 °C [28]. In another dehydrogenation reaction, specifically that of benzyl alcohol, copper oxide nanoparticles have shown 99% conversion and 99% selectivity at temperatures as low as 100 °C in the ODH [38]. Similarly, atomically precise Cu_x_ (such as Cu_4_, Cu_12_, and Cu_20_) clusters supported on Al_2_O_3_ featured efficient production of propylene between 400 and 550 °C, topping propylene selectivity between 79 and 84% at 550 C, depending on the cluster size, in a reactant mixture consisting of 2% propylene and 2% oxygen in helium, while also enabling the direct formation of propylene oxide at 150–300 °C [43,44,45].

Alumina is widely used as a support material, particularly in ODH reactions [6,27,41,42]. Its popularity stems from its high surface area, thermal stability, and beneficial chemical properties, which are closely linked to its physical characteristics, such as porosity and density. In particular, an Al_2_O_3_ film deposited by atomic layer deposition was reported to stabilise nanoparticles deposited on its surface via strong metal-support interactions, defect site anchoring, and electronic effects that reduce nanoparticle mobility and prevent sintering [26,42]. In addition to its morphology, the density and distribution of surface hydroxyl groups on the surface of alumina play a crucial role in anchoring metal species, facilitating higher and more uniform dispersion of the catalytic metal while also supressing agglomeration under reaction conditions [46,47].

In this work, we employ copper-based catalysts supported on two distinct alumina substrates (microsized and nanosized Al_2_O_3_) to investigate their performance in the oxidative dehydrogenation of propane, especially at lower temperatures, below 400 °C. Alumina was chosen as the support due to its high thermal stability, large surface area, and ability to stabilise dispersed metal nanoparticles. At the same time, copper was selected as a cost-effective alternative to noble metals with proven activity in various ODH reactions. Special emphasis is placed on how differences in the size of the alumina support particles can influence the dispersion of copper, which can, hand-in-hand, affect the number and size of copper particles on the surface, along with the nature of the active sites available on the copper particles, reflected through catalytic performance. Thus, particular attention is paid to the synergistic interactions between copper and alumina, elucidated by systematically comparing the performance and stability of Cu/Al_2_O_3_ and Cu/nanoAl_2_O_3_ in the dehydrogenation of propane_,_ depending on support morphology, oxidant type and reaction conditions. Furthermore, the effect of feed composition was assessed by varying the propane-to-oxygen ratio and by employing CO_2_ as a soft oxidant.

2. Results and Discussion

2.1. ODH of Propane with a 1:1 Ratio of Propane and Oxygen

Blank tests with nanoAl_2_O_3_ and Al_2_O_3_ were conducted to assess the activity of the supports. During data analysis, the activity of the support was subtracted from that of the copper-containing catalyst (see Figure S1) to distinguish the effect of metal loading.

On the blank supports, the conversion is lower, staying below 6%, with nanoAl_2_O_3_ and Al_2_O_3_ alone compared to their counterparts decorated with copper nanoparticles. NanoAl_2_O_3_ shows increased activity from 400 °C onwards, while Al_2_O_3_ exhibits measurable activity mainly at higher temperatures (500–550 °C). Propylene selectivity was predominant for both substrates, reaching 75% and 80% for nanoAl_2_O_3_ and Al_2_O_3_, respectively, at 500 °C. The ethylene fraction increases with temperature for both substrates, while the production of methane is minimal.

The addition of copper nanoparticles significantly enhances the catalytic activity for both nanoAl_2_O_3_ and Al_2_O_3_, resulting in an approximately tenfold increase in conversion at 500 °C. The addition of copper also significantly lowers the temperature at which the catalysts become active. Figure 1 shows the evolution of propane conversion and product selectivity (propylene, ethylene, methane, and CO_2_) on the studied catalysts, Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3,_ during the applied temperature ramp. For Cu/nanoAl_2_O_3_ (Figure 1a), propane conversion starts at 5% at 250 °C and gradually increases with the temperature, reaching 29% at 500 °C. The selectivity to propylene on Cu/nanoAl_2_O_3_ (Figure 1b) is the highest at 250 °C at 48% and it decreases with increasing temperature, dropping to a minimum of 5.8% at 450 °C. Product-wise, CO_2_ formation reaches 92% at 450 °C, making it the most abundant byproduct of the propane ODH. Nevertheless, the propane cracking products begin to appear at higher temperatures: ethylene at 400 °C and methane at 450 °C. The conversion and selectivity under different conditions are shown in Table S2.

When using Cu/Al_2_O_3_ as a catalyst, propane conversion increases throughout the entire temperature ramp, reaching a maximum conversion of 25% at 550 °C (Figure 1c); however, it is about 5% lower through the entire ramp than on Cu/nanoAl_2_O_3_ (Figure 1a). Propylene selectivity on Cu/Al_2_O_3_ (Figure 1d) has a maximum of 37% at 350 °C, then decreases significantly with rising temperature. Other products, primarily CO_2_, become more prevalent, with CO_2_ making up around 85% at 550 °C. Additional ethylene begins forming at 500 °C, followed by methane at 550 °C. The conversion and selectivity for Cu/Al_2_O_3_ at different conditions are shown in Table S3.

Comparing the two catalysts, Cu/nanoAl_2_O_3_ shows onset of activity at 250 °C, at a temperature 100 °C lower than Cu/Al_2_O_3_ with propylene selectivities of 48 and 37%, respectively, at the lowest temperatures of their activity (see Figure 1 and Table 1 for details.)

The total rate (r) of product formation per deposited copper atom for both catalysts during the temperature ramp is plotted in Figure 2a, showing the about 40% higher rate for Cu/nanoAl_2_O_3_. Figure 2b depicts the evolution of the rates of formation of the individual products on Cu/nanoAl_2_O_3_. The CO_2_ rate is very high in the ODH reaction, with propylene rates being around 5 mmol·g_Cu_^−1^·h^−1^ throughout the temperature ramp and 8.5 mmol·g_Cu_^−1^·h^−1^ at 550 °C. Figure 2c shows the evolution of products using the catalyst Cu/Al_2_O_3_.

2.2. ODH of Propane with a 3:1 Ratio of Propane and Oxygen

To moderate the oxidation environment, experiments with a propane-to-O_2_ ratio of 3:1 were conducted, which is sufficient to facilitate dehydrogenation at low temperatures while minimising unwanted side reactions. The reaction gas concentration for this experiment was 0.75% propane and 0.25% oxygen. The results of the experiment with Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3_ are shown in Figure 3 and Figure 4, respectively (for a detailed list of conversion and selectivities, see Tables S2 and S3).

Blank tests with nanoAl_2_O_3_ and Al_2_O_3_ were performed using the same conditions, and the results are shown in Figure S2. Both supports show minimal activity, with conversions remaining below 3% for nanoAl_2_O_3_ and 2% for Al_2_O_3_ up to 550 °C. NanoAl_2_O_3_ begin to show measurable conversion from 400 °C, whereas Al_2_O_3_ become active between 500 and 550 °C. At 500 °C, propylene dominates the product distribution with selectivities of 70% on nanoAl_2_O_3_ and 56% on Al_2_O_3_. Minor amounts of methane and ethylene is detected in the 500–550 °C range; nanoAl_2_O_3_ produces up to 30% ethylene, whereas Al_2_O_3_ generates a comparatively higher, 34% fraction of methane.

The catalyst activity decreases with decreasing O_2_ concentration; however, a significant increase in propylene selectivity is observed from 250 to 550 °C (Figure 3) with trends similar to those in the previous experiment for the Cu/nanoAl_2_O_3_ activity (see Figure 3a). The propane conversion reaches the maximum of 12% at 500 °C, while the propylene selectivity (shown in Figure 3b) is the highest at 250 °C with 48.3%. The propylene selectivity decreases to a minimum at 400–450 °C and increases again with increasing temperature (500–550 °C). CO_2_ is also the most abundant byproduct in these experiments, with selectivity reaching 87% at 400 °C. The CO_2_ selectivity is lower than in the experiments with a 1:1 reactant ratio. Ethylene emerges at 450 °C with 5% selectivity, while methane appears at 500 °C with 1.5% selectivity.

The conversion using Cu/Al_2_O_3_ (Figure 3c) is approximately half of that obtained with the 1:1 reactant ratio, with a maximum at 500 °C. Figure 3d shows the product selectivity of propylene, CO_2_, ethylene, and methane. The propylene selectivity at 250 °C is 54.7%; with increasing temperature, it decreases to 26% at 450 °C and then increases again to 38% at 550 °C. Meanwhile, the CO_2_ selectivity reaches 68% at 450 °C. Cracking starts at 450 °C, and the selectivity to cracking products (ethylene and methane combined) reaches around 18% at 550 °C.

The rates of product evolution for propane ODH with a 3:1 ratio of propane to oxygen are shown in Figure 4a for both catalysts. The Cu/nanoAl_2_O_3_ catalyst shows up to about 3% higher activity during the ramp in comparison with Cu/Al_2_O_3_. The product rates using Cu/nanoAl_2_O_3_ (Figure 4b) show a higher CO_2_ yield than propylene, with the cracking products (ethylene, methane) emerging at 450 °C. Compared to a reaction with abundant oxygen, cracking occurs at higher temperatures. Cu/Al_2_O_3_ gives the highest CO_2_ yield at 450 °C, accompanied by increased ethylene and methane (Figure 4c).

The experiment with a 3:1 propane-to-oxygen ratio demonstrated that Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3_ can effectively facilitate propane dehydrogenation under these conditions, even at 250 °C, while relatively minimising complete combustion in comparison with a feed of 1:1 propane and oxygen. Cu/nanoAl_2_O_3_ produces here smaller fractions of ethylene and methane than Cu/Al_2_O_3_. Overall, the propylene selectivity for both catalysts improves with respect to the 1:1 reactant ratio, especially notably for Cu/Al_2_O_3_ with an increase from 37% to around 60% in the low-temperature range.

We hypothesise that the observed differences in the ODH performance of the copper-based catalysts on the micro- and nanosized alumina supports can have their origin in differences in their surface area (i.e., facilitating different dispersion of copper which can yield different-sized copper particles with active sites of different nature), defect density, surface termination, metal-support interactions, and acid–base properties between the micro- and nanosized support.

2.3. Use of CO2 as a Mild Oxidant

Both Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3_ produce significant to prevailing fractions of CO_2_ during ODH. To assess the effect of the nature of the oxidant in the dehydrogenation of propane, the performance of these two catalysts using CO_2_ as a soft oxidant was addressed. For a complete list of conversions and selectivities at different temperatures, see Tables S2 and S3. The gas mixture contained 0.75% propane and 0.75% CO_2_ in He, giving a 1:1 reactant ratio.

First, the conversion and selectivity of the blank substrates was determined, with the results plotted in Figure S3, showing very low activity for both supports. NanoAl_2_O_3_ reaches only 1.1% conversion at 550 °C, with detectable activity starting no earlier than 500 °C. At 550 °C, the product distribution consists of 66% propylene, 8% methane, and 27% ethylene. Al_2_O_3_ shows similarly limited performance, achieving 1.4% conversion at 550 °C, with the same selectivity pattern (66% propylene, 34% ethylene).

Figure 5 represents propane conversion and product selectivity on the copper-containing catalysts. Compared to using O_2_ as an oxidant, the propane conversion is about an order of magnitude lower, reaching around 2.4% above 400 °C for Cu/nanoAl_2_O (Figure 5a) and about 2% at the highest temperatures with Cu/Al_2_O_3_ (Figure 5c).

The conversion of CO_2_ follows a similar trend to that of propane. The combustion route is supressed, and both catalysts primarily produce propylene, in case of Cu/nanoAl_2_O_3_, with 100% selectivity up to 500 °C (Figure 5b). Using Cu/nanoAl_2_O_3_, cracking occurs at 550 °C with ethylene and methane making up a 30% fraction of products combined. For Cu/Al_2_O_3_, cracking sets off already at 400 °C, producing primarily ethylene (Figure 5d) up to about 39% at 500 °C, and a small fraction of methane occurring at 550 °C at the expense of ethylene.

For a comparison of rates obtained in ODH with molecular oxygen shown in Figure 2 and Figure 4, Figure 6a shows the rates of product evolution during propane ODH with CO_2_ on both catalysts.

Remarkably, the rate of propylene formation with CO_2_ exceeds the rate observed with oxygen on Cu/nanoAl_2_O_3_ and makes up about 50% of the rate obtained on Cu/Al_2_O_3_, while at the same time, the formation of byproducts with CO_2_ is suppressed, the latter severely hampering selectivity of these catalysts when using O_2_ as the oxidant.

2.4. Characterisation of the Catalysts

The composition of Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3,_ was analysed using atomic absorption spectroscopy (AAS), determining the total copper loading as 2.2 wt% in Cu/nanoAl_2_O_3_ and 2.7 wt% in Cu/Al_2_O_3_ (Table 1).

The SEM image of the as-prepared Cu/nanoAl_2_O_3_ is presented in Figure 7a, showing a wool-like morphology of nanoAl_2_O_3_. TEM and STEM images, shown in Figure 7c,e, indicate that the nanoAl_2_O_3_ substrate itself is not homogeneous. The TEM images depict various morphologies (rods, pellets, and spheres) of support particles with diameters of 15–50 nm, spanning a wide range of alumina phases, and the Cu NPs form surface structures on the nanoAl_2_O_3_ that are similar to chains, with individual particles connected. Figure 7b presents the SEM images of the spent Cu/nanoAl_2_O_3_ catalyst, showing that the morphology of the catalyst is not changed during the experiment. The nanoAl_2_O_3_ particles (around 180–500 nm in diameter) are visible alongside the flake-like structures in the Cu/Al_2_O_3_ catalyst, which are around 500 nm in diameter. The STEM images (Figure 7e,f) show the same nanoAl_2_O_3_ morphology as in fresh catalysts. The Cu nanoparticles appear to form into chain-like structures of individual Cu NPs on the surface of the nanoAl_2_O_3_. The fresh and spent catalysts are visually very similar in terms of the agglomerates of Cu NPs. The slight changes in the surface distribution of copper and further agglomeration might remain undetected by this method.

Figure 8a shows the morphology of Cu/Al_2_O_3_, where Al_2_O_3_ exhibits a distinctive flake-type structure in contrast to the nanoAl_2_O_3_, featuring smaller particles with a diameter of approximately 70 nm. The TEM (Figure 8c) images reveal the presence of smaller particles as well—roughly spherical CuO particles—measuring approximately 10 nm in diameter. The STEM image (Figure 8e) shows that the copper nanoparticles are distributed on the surface of the powder Al_2_O_3_ (the Cu NPs typically appear with higher contrast than Al_2_O_3_). As seen in Figure 8b, the SEM image of the spent Cu/Al_2_O_3_ reveals the transformation of flake and spherical particulates on the surface of the 100–150 μm powder particles into roughly spherical particulates approximately 150 nm in diameter. Flake-like particles, which are present in the fresh catalyst, can also be found. In Figure 8d,f, the STEM-HAADF and TEM figures depict alumina morphologies similar to those of the fresh catalyst. The TEM image depicts a particle covered with copper particulates, which have agglomerated into sharp star-like formations during the reaction. For the Cu/Al_2_O_3_ samples, the visual changes to the Cu NPs are more pronounced. However, these changes may be limited to a few localised areas and not represent the entire sample.

The SEM and TEM images and the selected-area electron diffraction pattern (SAED) of both supports (nanoAl_2_O_3_ and Al_2_O_3_) showed a similar structural composition, irrespective of the introduction of copper nanoparticles. They are presented in the Supplementary Materials (Figures S4, S5, and S6, respectively).

STEM-HAADF images and EDX mapping of fresh and spent Cu/nanoAl_2_O_3_ (Figure 9) provide additional insights into the morphology and Cu NP distribution. The STEM image of the fresh Cu/nanoAl_2_O_3_ (Figure 9a) shows that the nanoAl_2_O_3_ particles are of various shapes and sizes, including rods, flakes, and spherical structures. The EDX mapping reveals that although copper nanoparticles, around 3 nm in diameter (see inset in Figure 9a), are to a great extent uniformly distributed on the surface of nanoAl_2_O_3_ (Figure 9c), during the course of the reactions they may also form larger aggregates, up to about 114 nm in size (Figure S7). In the spent catalyst (Figure 9b), a similar “woolly” morphology of nanoAl_2_O_3_ is observed, along with the agglomeration of copper nanoparticles into larger particles (up to around 270 nm in diameter).

STEM–HAADF images and EDX mapping of the fresh and spent Cu/Al_2_O_3_ catalyst (Figure 10) provide additional insights into the morphology and copper nanoparticle distribution in this catalyst. The alumina support exhibits the flake-type structure observed in the TEM images. In contrast to Cu/nanoAl_2_O_3_, the Cu nanoparticles/particle assemblies on Cu/Al_2_O_3_ are much less uniformly dispersed and are significantly larger, between 30 and 70 nm, on the fresh catalyst. Upon reaction, these particles further sinter, forming even larger agglomerates of up to 500 nm in the spent sample.

Similar structural effects have been reported for other supported metals on γ-Al_2_O_3_, where differences in support morphology and surface chemistry profoundly influence metal dispersion. The abundant terminal hydroxyl groups and associated surface defect structures on nanosized γ-Al_2_O_3_ have been shown to stabilise atomically dispersed metal species more effectively than on microsized supports, which tend to favour formation of larger particles, whereas supports with fewer terminal hydroxyl groups are more prone to facilitate metal aggregation [46,47].

The electron diffraction patterns were analysed to identify the alumina phases (see Figure 11). The SAED of Cu/nanoAl_2_O_3_ corresponds to the θ-Al_2_O_3_ polymorph, which typically forms at high temperatures (over 900 °C) and is considered one of the most thermodynamically stable transition alumina phases [48,49]. Similarly, the SAED of Cu/Al_2_O_3_ shows the presence of the theta phase. Additionally, the typical Cu^II^O scattering pattern is observed for both catalysts.

Nevertheless, these SAED patterns may not represent the bulk material. For this reason, the XRD was performed to analyse the composition of the alumina substrates and to determine whether it changes with the addition of Cu nanoparticles. The XRD patterns of the catalysts are shown in Figure 12 (the XRD patterns of the substrates are shown in Figure S8). The patterns indicate the presence of α-phase and θ-Al_2_O_3_. The changes before and after the reaction are noticeable for the nanoAl_2_O_3_ catalysts, both with (Figure 12a) and without (Figure S8a) copper. In the fresh catalysts, aluminium hydroxide phases were present (their main contribution was in the 2θ ranges 17.5–19° and 20–22°), which disappear after the reaction. The aluminium hydroxides presence can be the result of the commercial colloidal alumina being dispersed in water. At the same time, their subsequent decomposition can be attributed to temperature-induced effects during the propane conversion reaction. The powder Al_2_O_3_ catalysts with (Figure 12b) and without copper (Figure S8b) seem to remain relatively unchanged after the reaction. In addition to the phases mentioned above, the samples also contain the κ-phase. Compared to α-Al_2_O_3_, the κ-phase typically has a lower pore density and a smaller grain size [50]. The copper, identified as Cu^II^O in the fresh catalyst, is reduced to Cu^0^ after the reaction, as evidenced by the appearance of peaks at 50° and 74°. Adding copper nanoparticles also changes the intensity of peaks attributed to alumina phases.

The Rietveld refinement shows good agreement in the peak positions for all phases. However, it yields poorer results for the intensities of those peaks with the higher contribution from the theta phase. The Pawley approximation for this phase was employed to determine the associated error. The process involved treating the intensity as a free parameter to be refined rather than calculating its contribution from the atomic positions. There were several potential reasons for the discrepancy in intensities, including induced preferred orientation due to sample preparation, the possibility that the theta phase was non-stoichiometric (i.e., incorrect atomic positions were used in the phase model), or the presence of an additional undetected phase that matches the peak positions. The latter is the most probable explanation, and it is supported by the presence of a peak (in Cu/Al_2_O_3_ and Al_2_O_3_) or a shoulder (in Cu/nanoAl_2_O_3_ and nanoAl_2_O_3_) at around 45.6°, which cannot be attributed to any specific phase. This fact, along with strong peak broadening and very high peak overlapping, worsens the accuracy of the refinements, resulting in compositions that exhibit errors of up to 8%, depending on the case and phase considered. The refinement results are depicted in Table 2 for the copper-containing catalysts and Table S4 for the substrates. The table values represent the average between quantifications with and without Pawley’s approximation. At the same time, the errors correspond to the standard deviation. The fresh nanoAl_2_O_3_ samples contain approximately 15% aluminium hydroxide. The higher hydroxyl content in nanoAl_2_O_3_ likely contributes to the superior dispersion of copper nanoparticles, as surface hydroxyls act as anchoring sites that stabilise small metal clusters and prevent agglomeration, consistent with previous studies on metal–alumina interactions [46,47]. The rest of the composition comprises mainly θ-phase and α-phase alumina. On the other hand, the powder Al_2_O_3_ samples are composed solely of the α, θ, and κ-phases. We hypothesise that the lack of resolution of a Cu particle size by XRD in the spent Cu/nanoAl_2_O_3_ catalyst may hint towards amorphous and/or (sub)nanometre-sized highly dispersed copper particles after the reaction.

Table 3 presents the calculated specific surface area and pore volume for the catalysts and supports in both their as-made and spent forms based on the adsorption isotherms shown in Figure S9. The surface area and pore volume of the supports (nanoAl_2_O_3_ and Al_2_O_3_) remain unchanged after the reaction, suggesting that copper was primarily responsible for the changes in the spent catalysts. For Cu/nanoAl_2_O_3,_ the specific surface area increases slightly. At the same time, the pore volume stays nearly the same before and after the reaction (shown in Table 3). In contrast, Cu/Al_2_O_3_ shows a slight decrease in the surface area of the spent catalyst, likely due to pore collapse or blockage by side products from the propane ODH reaction at high temperatures.

3. Experimental Section

Copper sulphate pentahydrate (CuSO_4_·5H_2_O), colloidal aluminium oxide (Al_2_O_3_, 20% w/w in water, particle size 30–60 nm), and sodium borohydride (NaBH_4_) were purchased from Sigma Aldrich (St. Louis, MO, USA). Powder aluminium oxide (Al_2_O_3_, powder) was purchased from Penta (Prague, Czech Republic). The gases C_3_H_8_ (3% in helium), O_2_ (1% in helium), CO_2_ (1% in helium), and He (99.99%) were purchased from Messer (Prague, Czech Republic). Deionised water (purity 0.05 μS·cm^−1^, AQUAL 29, Merci, Brno, Czech Republic) was used to prepare the catalysts and to wash the catalysts. Sonicator SONOPULS HD 4400 Ultrasonic homogeniser (Bandelin electronic GmbH, Berlin, Germany) was used to mix the solution. Eppendorf Centrifuge 5702 (Hamburg, Germany) was used to separate the solid products.

3.1. Preparation of the Catalysts

The catalysts used in this study were synthesised by reducing copper sulphate with sodium borohydride in the presence of colloidal (nanosized) or powdered (macrosized) Al_2_O_3_, leading to the deposition of the resulting Cu nanoparticles onto the alumina surface. For Cu/nanoAl_2_O_3_, 196 mg of copper sulphate was dissolved in 145 mL of deionised water, and 4.7 mL of colloidal aluminium oxide (nanoAl_2_O_3_) was added for the catalyst. For the catalyst denoted Cu/Al_2_O_3_, 150 mL of deionised water and 1 g of aluminium oxide (Al_2_O_3_) were added instead. The solutions were stirred thoroughly using a magnetic stirrer at room temperature for 10 min. The reducing agent, 50 mL of borohydride solution (NaBH_4_, 59.2 mg in 50 mL water), was added at 1 mL·s^−1^ while using sonication pulses and thorough stirring. Employing ultrasonic pulses during Cu NP preparation ensured high nanoparticle dispersion [51,52]. After mixing the reactants, the solutions were mixed for another 10 min. The prepared mixtures were isolated by centrifugation for 10 min at 4400 rpm and washed with deionised water. This cleaning procedure was repeated twice. The powder samples were dried overnight in an oven at 60 °C and 1 bar. Dried colloid nanoAl_2_O_3_ and powder Al_2_O_3_ were used as blanks for the catalytic test for Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3_, respectively.

3.2. Catalyst Characterisation Techniques

Scanning electron microscopy (SEM) was performed using a HITACHI SU6600 (Hitachi High-Tech, Tokyo, Japan). Transmission electron microscopy (TEM) was performed on a FEI Tecnai TF20 X-twin (200 kV) (Thermo Fisher Scientific, Hillsboro, OR, USA), with a field-emission gun (FEG) and a point resolution of 2.5 Å, equipped with an Energy-Dispersive X-ray (EDX) detector (EDAX, Mahwah, NJ, USA). The microscope was operated in scanning mode with a High-Angle Annular Dark Field Detector (STEM-HAADF) (Thermo Fisher Scientific, Hillsboro, OR, USA). TEM images and selected-area electron diffraction (SAED) patterns were recorded on a Gatan UltraScan CCD camera (Gatan, Pleasanton, CA, USA) with a resolution of 2048 × 2048 pixels. SAED patterns were evaluated using the Process Diffraction software package V_7.8.1 Q [53]. The EDX spectra/maps were processed using the FEI TIA software version 4.2 sp1 build 816. Powder samples were dispersed in distilled water, and the suspensions were subjected to a 5 min ultrasound treatment. The diluted suspension was dropped on a holey-carbon-coated copper grid, and the sample was dried at ambient temperature. High-Resolution Transmission Electron Microscope (HRTEM) images were obtained using FEI Titan 60–300 kV (Thermo Fisher Scientific, Hillsboro, OR, USA) with a resolution of 1.4 Å and an Energy-Dispersive X-ray spectroscope (EDX) for elemental mapping. The microscope was also used in STEM-HAADF mode. Atomic absorption spectroscopy (AAS) was performed using a ContrAA 300 (Analytik Jena AG, Jena, Germany) spectrometer with flame ionisation. X-ray diffraction (XRD) characterisation was performed on a SmartLab SE Multipurpose Rigaku diffractometer (Rigaku Corporation, Tokyo, Japan) with Cu Kα radiation and a HyPix-3000 2D detector (Rigaku Corporation, Tokyo, Japan). All patterns were acquired in Bragg–Brentano geometry with the detector operating in 1D mode. Data processing involved using X’Pert HighScore Plus software V5.2 (Malvern Panalytical B.V., Almelo, The Netherlands) for phase identification and TOPAS V3 for Rietveld refinement. The specific surface area of the catalyst was determined by gas sorption measured on the surface area and catalyst analyser 3 flex (Micromeritics, Norcross, GA, USA) at 77 K up to the saturation pressure of N_2_. The Brunauer–Emmett–Teller (BET) model was used to calculate this area. The multipoint BET values were determined using the Rouquerol method and were within the standard range of p/p_0_ = 0.05–0.3. Before the surface analysis, all catalysts were treated at 200 °C for 4 h under vacuum, followed by 130 °C for 12 h under vacuum.

3.3. Catalytic Testing

The catalytic experiments were performed in a Microactivity-Reference Catalytic Reactor, PID Eng&Tech/Micromeritics, using a quartz tube reactor, 320 mm long and 10 mm inner diameter. Typically, 150 mg of a catalyst was placed on top of 20 mg of quartz wool in the reactor and conditioned at 150 °C in 40 mL/min of He for 45 min. First, blanks (nanoAl_2_O_3_ and Al_2_O_3_) were tested, and those results were treated as baselines for Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3_, respectively. The temperature ramp is shown in Figure 13, starting at 250 °C and increasing up to 550 °C. The temperature was raised in 50 °C increments at 5 °C/min in the 40 mL/min He flow. After reaching each temperature step and remaining at that temperature for 20 min, the helium was switched to the reaction mixture, which contained a 1:1 mixture of propane and O_2_ diluted with He, to 0.75% propane, 0.75% O_2_, and 98.5% He. A total flow of 40 mL/min was used at a pressure of 1 bar. The different reactant ratios used were 0.75% propane, 0.25% O_2_ and 99% He (3:1). The experiments with CO_2_ as a soft oxidant were performed with a reaction mixture containing 0.75% propane, 0.75% CO_2_ and 98.5% He (1:1). The reaction products were analysed on an Agilent gas chromatograph 6890 equipped with TCD (HP-PLOT/Q) and FID (Al_2_O_3_/KCl) detectors, injecting the gas mixture from the reactor after 10 min of dwell time within the inlet of the reactants at each temperature step of the temperature ramp showed below.

The propane conversion rate was calculated from the integrated GC peak areas of the products and reactants. The carbon-based selectivity and conversion reported herein were obtained after subtraction of the activity of the blank nanoAl_2_O_3_ and Al_2_O_3_. The rate was calculated based on the mass of the catalyst, giving the millimole of product produced by a gram of copper per hour. For comparison with cluster catalysts, we also report the rate calculated per copper atom, i.e., providing the number of product molecules per copper atom per second. Table S1 compares the Cu catalysts tested in this study with results from a previous study that used Cu and Pt clusters as catalysts.

4. Conclusions

This study discussed the catalytic performance of copper catalysts prepared via wet impregnation on commercially available aluminium oxide supports of different granularity at the nano- and microscale as potential candidates for propane dehydrogenation, based on non-precious metals, as part of an effort to identify a highly cost-effective alternative to noble-metal-based catalysts using both molecular oxygen and carbon dioxide as oxidants. Characterisation of the Cu/nanoAl_2_O_3_ and Cu/Al_2_O_3_ samples shows higher dispersion of copper on the nanoalumina, making the differences in the copper particle size the most likely driving force for the observed differences in the catalytic performance of the two catalysts. Under the applied feed conditions and using oxygen as the oxidant, propane conversions of up to about 10% and propylene selectivities reaching around 65% can be obtained at temperatures below 350 °C, depending on the catalyst and propane-to-oxygen ratio. At higher temperatures, combustion and cracking prevails. Especially promising appears to be the use of carbon dioxide as a mild oxidant, where on Cu/nanoAl_2_O_3_ a 100% selectivity towards propylene oxide is observed up to 500 °C, while on Cu/Al_2_O_3_ cracking sets off already at 400 °C. The unique selectivity showcased by Cu/nanoAl_2_O_3_ with carbon dioxide as the oxidant is accompanied by its about doubled efficacy, i.e., rate, of propylene formation in comparison with using molecular oxygen, which further articulates the potential offered by copper-based catalyst for energy- and cost-efficient dehydrogenation of propane.