*2.2. Characterization of the Catalysts*

Transmission electron microscopy (TEM) characterization was conducted with a JEOL TEM-2100 multipurpose electron microscope. For scanning electron microscopy (SEM) analysis, a HITACHI SU6600 scanning electron microscope was used. Scanning transmission spectroscopy with high-angle annular dark-field scanning transmission electron microscopy (STEM/HAADF), equipped with an energy-dispersive X-ray spectroscope (EDX), was used for the elemental mapping and obtaining the STEM/HAADF images of the prepared catalysts.

Powder X-ray diffraction (XRD) analysis was performed in a Malvern Panalytical Empyrean diffractometer and the quantification of the individual Fe components in the fresh and used catalysts was completed by using Rietveld analysis of the obtained XRD data, with the High Score Plus (Malvern Panalytical) software utilizing the PDF-4+ and ICSD databases.

Atomic absorption spectroscopy (AAS) was conducted with an Analytik Jena AG ContrAA 300 spectrometer with flame ionization. The sorption of gas was measured on the surface area analyzer Autosorb iQ-C-MP (Quantachrome Anton Paar), using ASiQWin software at 77 K up to the saturation pressure of N2. For calculating the specific surface area, the Brunauer-Emmett-Teller (BET) model was used. The points for multipoint BET were determined using Roquerol's method and are within the standard range of *p*/*p*<sup>0</sup> = 0.05 to 0.3. Before the surface analysis, all catalysts were treated at 130 ◦C for 12 h under vacuum.

X-ray photoelectron spectroscopy (XPS) was conducted on a PHI 5000 VersaProbe II spectrometer with monochromatic Al*K*α radiation.

UV-visible spectroscopy was performed using a UV-Vis-NIR spectrometer Perkin-Elmer Lambda 950.

#### *2.3. Catalytic Testing*

Tests of the catalysts were performed in a Microactivity Reactor System, (PID Eng&Tech/ Micromeritics) coupled to a quartz tube reactor of 320 mm in length and 10 mm inner diameter. A total of 200 mg of a catalyst was placed onto 20 mg of quartz wool in the middle of the reactor and the reactor was conditioned at 250 ◦C in He with a flow of 30 mL/min for 40 min. No other pre-treatment of the catalysts was performed. The reaction mixture used contained CO2, H2, and He at the ratio of 1:5:4 giving 11% and 49% of CO2 and H2 in He, respectively. A total flow of 25 mL/min was used at a pressure of 1 bar. The reaction products were analyzed on an Agilent 6890 gas chromatograph equipped with TCD (column HP-PLOT/Q) and FID (column Al2O3/KCl) detectors, using an injection after 20 min of reaching the given temperature. In the range of 250 ◦C to 410 ◦C, the temperature was raised in steps of 30 ◦C, at a rate of 5 ◦C/min (see Figure S1 for the double temperature ramp applied). After reaching the highest temperature in the first ramp, the reactor was cooled down to 250 ◦C under He. Next, the catalytic test was repeated using an identical heating ramp as the first one.

#### **3. Results and Discussion**

SEM micrographs of the as-prepared catalysts are shown in Figure 1, revealing two distinctly different morphologies. FeOx (Figure 1a) and 1%-Cu/FeOx (Figure 1b) show up as rods, with up to about 5 μm in length and thinner (ca. 500 nm) 1%-Cu/FeOx than for FeOx (ca. 1.3 μm), both with a coarse surface structure. The morphology of 3%-Cu/FeOx and 5%-Cu/FeOx (Figure 1c,d, respectively) is very different, reminding us of a structure of sponge or wool. We hypothesize that the observed differences in morphology reflect the differences in the composition of the reaction mixtures used in the synthesis of this set of catalysts. One parameter is the presence of hydrazine, which has reducing and basic properties; the other cause affecting morphology is the variable concentration of SO4 2, which is introduced to the system by adding CuSO4 solution in increasing amounts with a growing Cu loading. With CuSO4 being acidic, the low pH could impact the level of etching of FeOx, [51] as well as the chemical composition.

**Figure 1.** SEM images of fresh (**a**) FeOx, (**b**) 1%-Cu/FeOx, and (**c**) 3%-Cu/FeOx, (**d**) 5%-Cu/FeOx. For additional images, see Figure S9.

SEM images with EDX analysis of the highlighted part are depicted in Figure S2. These images show the elemental composition on the surface of the fresh catalysts. The chemical composition and elemental distribution of FeOx and Cu/FeOx catalysts were characterized by EDX mapping and their EDX spectra. The STEM/HAADF images are depicted for FeOx in Figure S3a, 1%-Cu/FeOx in Figure S4a, 3%-Cu/FeOx in Figure S5a, and 5%-Cu/FeOx in Figure S6a. The images are depicting the same nanorod morphology as shown by the TEM images. As shown, the two dominant elements observed were Fe and O, which were distributed evenly in the catalysts. EDX elemental mapping is shown in Figures S3–S6. Elemental maps, including copper, iron, and carbon, shown in Figures S4–S6, for the copper-containing catalysts, present uniformly dispersed copper in these catalysts. The detected copper particulates appear to be around 2 nm in diameter. In these maps, carbon is shown to be present as the common contamination from exposing the samples to air or detected on the sample grid.

The effect of the composition of the synthesis solution on the composition of the iron component was also confirmed by XRD (Figure S7), with compositions listed for the individual catalysts in Table 1.


**Table 1.** Fractions of various forms of iron oxide components in the fresh catalysts, obtained from XRD data using Rietveld refinement of analysis and specific surface area of fresh catalysts, as determined from adsorption/desorption of nitrogen on the surface of the fresh catalysts. Multipoint BET was assessed using Roquerol's method.

<sup>a</sup> Obtained from XRD. <sup>b</sup> Obtained from atomic absorption spectroscopy.

It is important to note, that the mean size of coherent domains obtained from XRD data can underestimate the particle size, because it does not take into account the possibly present non-crystalline surface layer, and the mean X-ray coherence length (MLC) in the case of a multiphase system can be less accurate as well. The data are summarized in Table S1. FeOx consists of a coherent domain length of 9 nm for α-Fe2O3, 12 nm for α-FeO(OH), and 11 nm for Fe3O4. Whereas, 1%-Cu/FeOx has 12 nm for α-Fe2O3 and 11 nm for α-FeO(OH) and (not in the table) 56 nm for the copper phase. The 3%-Cu/FeOx catalyst has 19 nm for α-Fe2O3 and 9 nm for α-FeO(OH). The 5%-Cu/FeOx catalyst has 19 nm α-Fe2O3 and 10 nm α-FeO(OH). Diffraction peaks of copper were detected only in 1%-Cu/FeOx, at 50.3◦, 58.8◦, and 87.9◦, due to too small a particle size and/or their amorphicity in the other samples [52]. The chemical composition of iron and copper in the Cu/FeOx samples was determined from the analysis of XPS spectra, as shown in Figures S8–S10, including the deconvoluted spectra and the results from fitting the spectra. The broad Fe 2p3/2 peak indicates the presence of Fe in different valence states. The deconvolution of this peak shows the presence of a mixture of Fe2+ and Fe3+ in the fresh catalysts, where all spectra can be deconvoluted by two main components and one satellite feature. The spectral features above 933.0 eV are related to Cu2+ in copper oxides [53,54]. See also Table S2 for a summary of the XPS analysis.

Adsorption–desorption isotherms (Figure S11) were recorded at 77 K up to the saturation pressure of nitrogen to determine the surface area of the prepared catalysts using the Brunauer–Emmett–Teller (BET) theory. The hysteresis of the two curves shows that the prepared catalysts are porous. The specific surface area of the catalysts was calculated to be 114 m2/g for FeOx, 114 m2/g for 1%-Cu/FeOx, 93 m2/g for 3%-Cu/FeOx, and 92 m2/g for 5%-Cu/FeOx. See Table 1 for the data summary, reflecting on the difference in the morphologies of the catalysts. The UV-vis spectra of the fresh catalysts (Figure S12) show an absorption peak at 275 nm, associated with that of iron oxides [55].

#### *3.1. Catalyst Testing*

The catalysts did not undergo any reduction pre-treatment before the tests. Thus, the first temperature ramp can be considered as a pre-treatment step of the catalyst directly under the reactants which have a reducing character due to the excess hydrogen in it. Figure 2 shows the evolution of the CO2 conversion (left column) and the product selectivities (right column) of the studied catalysts during the applied double temperature ramp. The drop in conversion between the individual ramps (Figure 2c) and the constantly evolving selectivity observed for FeOx (see Figure 2b and Table S3) indicate that this particular catalyst did not converge to its final state during the applied double temperature ramp. On the contrary, the identical CO2 conversion observed for the Cu/FeOx catalysts at the highest temperatures of the first and the second temperature ramp, along with the experimental accuracy comparable product selectivities, provide a hint about the completed annealing of 1%-Cu/FeOx, 3%-Cu/FeOx, and 5%-Cu/FeOx during the first temperature ramp.

**Figure 2.** Hydrogenation of CO2. CO2 conversion and selectivities by (**a**,**b**) FeOx, (**c**,**d**) 1%-Cu/FeOx, (**e**,**f**) 3%-Cu/FeOx, (**g**,**h**) 5%-Cu/FeOx. A total of 200 mg of catalyst and CO2/H2/He (1:5:4, total flow 25 mL/min).

Since the catalysts evolve during the first temperature ramp, their performance during the first temperature ramp is not discussed here. The comparison of CO2 conversion shows that the addition of copper boosts conversion from around 28% up to about 37%, with respect to the copper-free reference catalyst. While all three interrogated Cu/FeOx catalysts primarily produce methane along with a small amount of carbon monoxide, depending on the temperature and copper content, these catalysts produce up to about a total of 43% of C2-C5 olefins and paraffins. As reported in the literature, copper plays a dual role, activating hydrogen for the reaction with CO2 itself and also for the reduction of iron oxide via hydrogen spillover, increasing the overall activity of iron-based catalysts.

Using the 1%-Cu/FeOx catalyst (Figure 2c,d), a relatively steady fraction of CH4 is produced in the applied temperature range, with methane fraction amounting to around 75% at the highest temperature of 410 ◦C, and about 66% at lower temperatures. With an increasing temperature, this catalyst exhibits primarily a temperature-dependent switch from paraffins to olefins. In the case of the 3%-Cu/FeOx (Figure 2e,f) and 5%-Cu/FeOx (Figure 2g,h) catalysts, methane formation dominates at 410 ◦C with a methane selectivity of 82% and 86%, respectively. At lower temperatures in the presence of a small amount of C5+ (in the order of about ~2%) up to 37% and 39% C2-C4 olefins and paraffins are produced in the temperature region of 250–370 ◦C. The formation of the relative fractions of C2-C4 olefins and paraffins is dependent on the copper contents in Cu/FeOx catalysts. As the concentration of copper increases from 1% to 5%, the selectivity of paraffin hydrocarbons also increases, accompanied by the decrease in the olefins fraction. Significantly fewer olefins are produced on 3%-Cu/FeOx in comparison with 1%-Cu/FeOx, and an even smaller amount is produced on 5%-Cu/FeOx. In terms of selectivity, we hypothesized the central role of the morphology of the catalyst because the catalysts of comparable composition synthesized differently turned out to convert CO2 primarily into methanol or benzene [37], some others produce methanol and small hydrocarbons [56], while the currently presented catalysts solely produce hydrocarbons. The presence of Cu plays an important role in both enhancing CO2 conversion and the selectivity toward higher hydrocarbons. Cu not only acts as an RWGS catalyst but also reduces the iron oxides into metallic iron (Fe2O3 → Fe3O4 → FeO → Fe), which is converted into Fe5C2 due to the carbonation process [18] to the active phase that produces higher hydrocarbons [57]. As the concentration of Cu increases, the fraction of Fe5C2 also increases, as confirmed by the XRD of the used catalysts (Table 2).

**Table 2.** Fractions of various forms of iron oxide and iron carbide components in the spent catalysts, obtained from XRD data using Rietveld refinement of analysis and specific surface area of spent catalysts, as determined from adsorption/desorption of nitrogen on the surface of the spent catalysts. Multipoint BET was assessed using Roquerol's method.


#### *3.2. Characterization of the Catalysts after Catalytic Tests*

Spent catalysts were characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), high-angle annular dark-field scanning transmission electron microscopy (STEM/HAADF), energy-dispersive X-ray spectroscopy (EDX), X-ray diffraction (XRD), sorption of N2 with the BET model and X-ray photoelectron spectroscopy (XPS).

SEM micrographs of spent catalysts (Figure 3) show a massive restructuring of all the catalysts, featuring intertwined polymorphs which indicate the presence of Fe5C2 or Fe3C [37,58]. Figure 4 shows TEM images of the spent catalysts, which for FeOx, reveal particles with a mean diameter of 68 ± 26 nm, and for the Cu/FeOx catalysts, particle sizes increase with copper content: 1%-Cu/FeOx—61 ± 19 nm, 3%-Cu/FeOx—75 ± 25 nm, and 5%-Cu/FeOx—81 ± 35 nm to 150 nm. Additional TEM images can be found in Figure S13.

**Figure 3.** SEM images of (**a**) spent FeOx, (**b**) spent 1%-Cu/FeOx, (**c**) spent 3%-Cu/FeOx, and (**d**) spent 5%-Cu/FeOx showing major restructuring at the micrometer and sub-micrometer scale. Additional images are shown in Figure S10.

**Figure 4.** TEM images of spent catalysts after CO2 hydrogenation of (**a**) FeOx, (**b**) 1% Cu/FeOx, (**c**) 3% Cu/FeOx, and (**d**) 5% Cu/FeOx showing the transformation of the rod-like particles of the fresh catalysts into roughly spherical particles.

SEM images with EDX analysis are depicted in Figure S14. STEM/HAADF imagining of spent catalysts are depicted in Figures S15a, S16a and S17a, and Figure 5a for FeOx, 1%-Cu/FeOx, 3%-Cu/FeOx, and 5%-Cu/FeOx, respectively. These figures show the presence of nanoscale Fe species of about 150 nm in diameter with Cu agglomerates of 10 nm in diameter. From the EDX analysis, Fe and O appear to be dominant elements (Figures S15b, S16b and S17b and Figure 5b), forming the core of the spent FeOx particles. There is a recognizable layer of carbon (13 nm) seen on the surface of the particles, indicating the deposition of carbon during the reaction. This layer is most clearly visible in the spent 5%-Cu/FeOx (Figure 5c), and the least in the reference FeOx catalysts (Figure S15). In addition, as seen in Figures S16c and S17c and Figure 5c, copper became highly dispersed in all copper-containing spent catalysts.

**Figure 5.** (**a**) STEM-HAADF of spent 5%-Cu/FeOx, and (**b**–**d**) EDX elemental mapping of Fe, O, C and Cu.

The fractions of different iron oxides and carbides present in the used samples obtained by XRD (Figure S18) are summarized in Table 2. Spent FeOx consists of 100% Fe3O4, spent 1%-Cu/FeOx has 95% Fe3O4 and 5% Fe5C3; the 3%-Cu/FeOx catalyst contains 71.3% Fe3O4 and 28.7% Fe5C3; and 5%-Cu/FeOx has 59.8% Fe3O4, 37.2% Fe5C3 and 3% Fe3C. The XPS of the spent catalysts reveals the presence of metallic Fe along with Fe2+ and Fe3+. The components at the binding energy of around 932.6 eV are related to Cu(0).

The XPS of the spent catalysts similarly shows the presence of metallic Fe along with Fe2+ and Fe3+. The components at the binding energy of around 932.6 eV are related to Cu(0) (see Figure S8). Table S2 summarizes the chemical constituents of the spent catalysts.

With the increasing copper load, an increase can be seen in the iron carbides content, which is consistent with the role of copper in reducing iron oxide. We note that copper was not detected by XRD on the spent catalysts either, which indicates that copper is present in either an amorphous form or the size of the majority of the copper particles is below the detection limit of XRD, as indicated by STEM-HAADF (Figure 5).

XRD further confirmed the formation of iron carbides in the copper-containing catalysts. From the XRD data, we obtained the mean X-ray coherence length (MLC), which is summarized in Table S4. Spent FeOx consists of a coherent domain length of 63 nm for the Fe3O4 phase. Whereas, spent 1%-Cu/FeOx has 63 nm for Fe3O4 and 29 nm for Fe5C2, and spent 3%-Cu/FeOx has 47 nm for Fe3O4 and 40 nm for Fe5C2. Spent 5%-Cu/FeOx has 36 nm for Fe3O4, 35 nm for Fe5C2, and 23 nm for Fe3C.

Adsorption–desorption isotherms of spent catalysts (Figure S19) were obtained at 77 K up to the saturation pressure of nitrogen to determine the surface area of the spent catalysts. As opposed to the fresh catalyst, the adsorption/desorption curves have a course that suggests a smaller surface area, perhaps due to the inaccessibility of pores due to the carbon deposition. The specific surface area of the catalysts was calculated to be 18.6 m2/g for FeOx, 18.6 m2/g for 1%-Cu/FeOx, 18.0 m2/g for 3%-Cu/FeOx, and 17.4 m2/g for 5%-Cu/FeOx. Thus, this is just a mere fraction of the as-made catalyst prior to the catalytic test, with the specific surface areas of the spent catalysts being about 10 times smaller (see Table 2) than that of the fresh catalysts (see Table 1).

For the convenience of the reader, Figure S20 summarizes the CO2 conversion and product selectivity during the second temperature ramp for the investigated catalysts, compared to the individual reaction temperatures. In order to test the stability of the 5%-Cu/FeOx catalysts, four more temperature ramps were performed (Figure S21). After the second temperature ramp, the CO2 conversion decreased with each ramp, accompanied by an increase in CO generation and a decrease in the fraction of CH4 and higher hydrocarbons. After the fourth ramp, the CO2 conversion decreased from the initial 35% to 19%. This change could be caused by the accumulation of carbon on the surface of the catalysts, as confirmed by STEM-HAADF and elemental mapping (Figure 5), where a layer of 13 nm thick of carbon is seen on the surface of the particles (Figure 5c).

A review of the related literature, including this study, is summarized for the convenience of the reader in Table S5 and Figure S22.

## **4. Conclusions**

Cu/FeOx catalysts with a Cu content of 0%, 1%, 3%, and 5% were found to be highly active in CO2 hydrogenation with the conversion reaching 36.8%. The observed selectivity to C1 versus C2-C4, C2-C4 =, and C5+ hydrocarbons is dependent on the catalyst's composition, morphology, and temperature. The observed range of products is diametrically different from those observed for catalysts with similar compositions but synthesized using other precursors. Depending on their copper content, the as-made particles possess rather distinct morphologies, most likely attributable to the changing ratio of the hydrazine and copper sulfate used during the synthesis of the individual samples, which affects the extent of the etching of the iron-oxide particles, as well as their crystallinity, chemical composition, and morphology at various scales. The findings presented in this paper indicate potential new ways of tuning the morphology and composition of iron-oxide-based particles, for example, by balancing the relative ratio of hydrazine and SO4 <sup>2</sup><sup>−</sup> ions or by using different precursors in the synthesis, ultimately yielding catalyst compositions and morphologies with variable catalytic performance. We note that, while some alkali (K, Na) doped Fe-based catalysts have a higher CO2 conversion, these require harsh conditions for pre-treatment under hydrogen or a CO atmosphere at temperatures exceeding 350 ◦C for extended times up to 12 h. Our catalysts have very low selectivity towards carbon monoxide as compared to other reported catalysts.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/catal12050516/s1. Figure S1: Double temperature ramp applied during catalytic testing; Figure S2: SEM/EDX of fresh catalysts; Figure S3: STEM/HAADF of fresh FeOx; Figure S4: STEM/HAADF of fresh 1%-Cu/FeOx; Figure S5: STEM/HAADF of fresh 3%-Cu/FeOx; Figure S6: STEM-HAADF of fresh 5%-Cu/FeOx; Figure S7: XRD patterns of fresh catalysts; Table S1: The mean X-ray coherence length of fresh catalysts; Figure S8: XPS of fresh and spent catalysts; Figure S9: XPS of FeOx; Figure S10: Wide scan XPS spectra of fresh and spent catalysts; Table S2: Chemical composition of Fe and Cu in catalysts; Figure S11: Adsorption and desorption isotherms of fresh catalysts; Figure S12: UV-vis spectra of fresh catalysts; Table S3: Conversion and selectivities of FeOx and Cu/FeOx catalysts; Figure S13: TEM images of fresh and spent catalysts; Figure S14: SEM/EDX of spent catalysts; Figure S15: STEM/HAADF images of spent FeOx; Figure S16: STEM/HAADF images of spent 1%-Cu/FeOx; Figure S17: STEM/HAADF images of spent 3%-Cu/FeOx; Figure S18: XRD patterns of spent catalysts; Table S4: The mean X-ray coherence length of spent catalysts; Figure S19: Adsorption and desorption isotherms of spent catalysts; Figure S20: Comparison of the performance of the catalysts; Figure S21: CO2 conversion and selectivity over the course of 6 consecutive temperature ramps; Table S5: Literature comparison of CO2 hydrogenation by different iron-based catalysts; Figure S22: Literature comparison of CH4/CO selectivity against CO2 conversion.

**Author Contributions:** K.S. synthesized the catalysts and performed their tests, analyzed data, and drafted the manuscript; M.I.Q. contributed to the catalyst testing and discussed the results; N.Ž. participated in the catalyst testing, discussed the catalytic results, and contributed to the writing of the manuscript; J.E.O. measured and interpreted the UV-vis spectra; P.S. collected and evaluated the XPS spectra; L.K. contributed to the characterization of the catalysts, the interpretation of the results and correcting the manuscript; Š.V. planned and correlated the effort, discussed and interpreted the results, and wrote and finalized the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** K.S., M.I.Q., N.Z. and S.V. gratefully acknowledge the support from the European Union's Horizon 2020 research and innovation program under grant agreement No 810310, which corresponds to the J. Heyrovsky Chair project ("ERA Chair at J. Heyrovský Institute of Physical Chemistry AS CR—The institutional approach towards ERA"). The funders had no role in the preparation of the article. The work of L.K was supported by the ERDF project "Development of pre-applied research in nanotechnology and biotechnology" (No. CZ.02.1.01/0.0/0.0/17\_048/0007323) and an internal grant of Palacky University IGA\_PrF\_2021\_032.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on reasonable request from the corresponding authors.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

