3.1 Chemical, structural and textural properties of the catalysts
The XRD results confirmed that the precipitation of Co and Mn nitrates under hydrothermal conditions in the presence of hydrolyzing urea resulted in crystallization of Co and Mn carbonates on the supports. After calcination, the supported catalysts containing approximately 1–5 wt. % of active oxides deposited on the meshes were obtained (Table 1). Determination of the oxides composition by chemical analysis was difficult due to very low Co and Mn contents in the catalysts and simultaneous presence of Mn in the used stainless steel meshes. Therefore, the Co and Mn concentrations were determined by EDX in small amounts of the deposited oxides removed from the meshes by brushing. The Co/(Co + Mn) molar ratios in the prepared catalysts were calculated; these values corresponded approximately to the nominal ones, i.e., to the Co/(Co + Mn) molar ratios in the nitrate solutions used for the catalysts preparation (Table 1).
Table 1
Physicochemical characteristics of the catalysts.
Co/
(Co + Mn)a
|
Oxides
content
(wt.%)
|
Weight
Lossb
(%)
|
Coc
(at.%)
|
Mnc
(at.%)
|
Co/
(Co + Mn)c
|
Particle
sized
(µm)
|
SBET
(m2 gcat−1)
|
SBET
(m2 goxides−1)
|
H2–TPRe
(mmol goxides−1)
|
0
|
4.97
|
22.8
|
-
|
-
|
0
|
16
|
0.71
|
14.3
|
3.4
|
0.2
|
3.28
|
22.5
|
8.1
|
31.6
|
0.20
|
15
|
0.38
|
11.6
|
3.0
|
0.4
|
2.55
|
6.8
|
14.6
|
20.4
|
0.42
|
9–15
|
0.27
|
10.6
|
3.5
|
0.5
|
1.14
|
34.1
|
8.4
|
8.7
|
0.49
|
13
|
0.27
|
23.7
|
5.3
|
0.6
|
2.95
|
29.2
|
23.7
|
19.1
|
0.55
|
4
|
0.48
|
16.3
|
5.1
|
0.8
|
1.91
|
21.0
|
29.8
|
8.2
|
0.78
|
7–20
|
0.49
|
25.7
|
6.0
|
1.0
|
1.34
|
3.6
|
-
|
-
|
1
|
7
|
0.16
|
11.9
|
12.9
|
Com
|
97.0f
|
–
|
61.5g
|
14.0g
|
0.8
|
–
|
85
|
88
|
5.2
|
a Nominal molar ratio |
b Weight loss after treatment in an ultrasonic bath related to the initial weight of the deposited oxides |
c Determined by EDX |
d Determined from SEM images |
e Hydrogen consumption in the temperature range 25–500°C, per gram of active phase |
f The Com catalyst contained 3 wt.% of graphite, which was added before pelletizing |
g Determined by chemical analysis |
Adhesion of the deposited Co-Mn oxides to the stainless steel supports was examined by a simple test, when the catalysts samples immersed in acetone were treated in an ultrasonic bath for 10 min. After drying, the weight loss was determined and related to the initial weight of the oxides in the catalysts. The obtained values are summarized in Table 1; the deposited oxides showed quite good adhesion to the support; weight losses varied from 4 to 34 %, the highest being found for the 0.5 sample and the lowest for the 1.0 sample containing only cobalt oxide.
Powder XRD analysis of the supported samples showed the diffraction peaks at about 51 and 60° 2θ that corresponded to a Fe0.65Cr0.1Ni0.25 phase (PDF 04–019–2390). They were ascribed to the stainless steel support. In the samples obtained after hydrothermal treatment of the stainless steel meshes in the solutions containing only Co or Mn nitrates, powder XRD analysis showed presence of CoCO3 (spherocobaltite, PDF 04–008–8312) or MnCO3 (rhodochrosite, PDF 04–011–3971). In the catalyst precursors prepared in the solutions containing mixtures of Co and Mn nitrates, both CoCO3 and MnCO3 were found as distinct phases. Only MnCO3 diffraction lines were detected in the precursors with higher Mn content prepared using solutions with Co/(Co + Mn) molar ratios of 0.2 and 0.4. It is known that the Co and Mn cations are precipitated at slightly different pH values; pH of 7.5 and 8.5 were reported for precipitation of Co and Mn hydroxides in 0.01 mol l− 1 solutions, respectively (Cavani et al., 1991). Therefore, former precipitation of Co cations to create CoCO3 could be expected during urea hydrolysis resulting in gradual pH increase in the solutions.
After calcination, Co3O4 (04–005–4386) and Mn2O3 (PDF 04–008–6383) were identified in the powder XRD pattern of the 1.0 and 0 catalysts, which contained only single Co and Mn oxides (Fig. 1). However, the Co–Mn mixed oxides were formed after calcination though the Co and Mn carbonates crystallized as separated phases on the surface of stainless steel meshes during their interaction with aqueous solutions of Co and Mn nitrates in presence of hydrolyzing urea. In the samples with lower Mn content, an incorporation of Mn cations into Co3O4 spinel lattice or direct CoMnO3 formation could be anticipated. The pattern of the calcined 0.8 catalyst having relatively low Mn content contained Co3O4 accompanied by a cubic Co–Mn mixed oxide corresponding to Mn1.9Co1.1O4 with spinel structure (PDF 04–009–3198) whereas Co3O4 together with CoMnO3 (PDF 03–065–3696) were identified in the 0.6 catalyst. The powder XRD patterns of the 0.5 and 0.4 catalysts showed only CoMnO3 as a crystalline mixed oxide deposited on the stainless steel support. Calcination of the precursors with low Co content prepared using solutions with Co/(Co + Mn) molar ratios of 0.4 and 0.2, in which Co and Mn carbonates were distinguished, led to transformation into various mixed oxides in dependence on Co and Mn contents. Only CoMnO3 was found in the 0.4 sample obtained from the precursor, in which CoCO3 diffraction lines were not detected; excess of Mn could be incorporated into an XRD–amorphous part. Tetragonal Mn2CoO4 spinel (PDF 04–007–4368) and Mn5O8 (PDF 04–008–3979) were found in the powder XRD pattern of the 0.2 catalyst with high Mn content. The powder X-ray pattern of the pelletized commercial catalyst (Com) is shown in the Supplementary material (Fig. S1). The catalyst contained a Co–Mn–Al mixed oxide with spinel structure. Slight amount of graphite (3 wt.%) was added to the mixed oxide before pelletizing.
Summarizing, the CoMnO3 phase with rhombohedral ilmenite structure, indicating the presence of tetravalent Mn cations, was found in the 0.4, 0.5, and 0.6 catalysts (Fig. 1). Mixed oxides with spinel structure were detected in other Co-Mn catalysts; a non-perfect structure ordering with Co and Mn cations in various oxidation states could be expected in mixed oxides emerging after calcination of carbonate precursors at 500°C in oxidation environment (air).
In literature, slightly different composition was reported for the Co–Mn mixed oxides with Co:Mn molar ratio of 1:1 prepared by various methods. For example, Tang et al. prepared the (Co,Mn)(Mn,Co)2O4 phase by decomposition of oxalates (Tang et al., 2014). Other authors reported a mixture of tetragonal Mn2CoO4 and cubic MnCo2O4 oxides when applied ball milling (Masi et al., 2015) or sol–gel combustion method (Lee et al., 2014). Difference in the phase compositions of the obtained oxides could be caused by thermal treatment; aforementioned authors used higher temperatures and longer times during thermal treatment, resulting in the formation of different oxide phases.
The surface morphology of the as–obtained catalysts is documented by SEM images in Fig. 2. The stainless steel meshes were covered with rather big particles of Co–Mn oxides originating from crystals of carbonate precursors formed during hydrothermal reaction. The crystals of carbonate precursors had well–defined, roughly cubic shape (not shown here), which remained preserved after calcination. Sizes of the oxide particles changed in dependence on Co and Mn content (Table 1). The biggest particles with size of about 15 µm were found in the 0 and 0.2 catalysts with high Mn content and decreased with increasing Co content in the deposited oxides. The particles with different sizes of about 9 and 15 µm were observed in the SEM images of the 0.4 catalyst. Powder XRD pattern of this sample before calcination contained diffraction lines corresponding to MnCO3 only (Fig. 1) and thus, presence of an amorphous Co component, making the precursor crystallization more difficult, could be expected. The 0.5 catalyst contained more uniform particles with size of 13 µm whereas the 0.6 sample showed the smallest particles (4 µm). In comparison with the 0.6 catalyst, the 0.8 and 1.0 ones with high Co content showed larger particles (about 7 µm) but some considerably larger particles were also observed in the 0.8 catalyst.
The employed method of stainless steel coating with Co and Mn carbonate precursors led to crystallization of big particles with roughly cubic shape. Gradually increasing pH in the solution due to urea hydrolysis resulted in precipitation of Co and Mn cations and crystallization of tiny particles well adhering to the support surface, followed by growth of these particles with increasing reaction time; finally, well–defined big crystals were obtained. The conditions of hydrothermal reaction affect strongly size and morphology of the prepared products; for example, formation of urchin–like hierarchical CoMn2O4 nanostructures (Venkateswarlu et al., 2017), pseudo–spherical particles of Co3O4 and MnxCo3−xO4 core-shell nanoparticles (Bian, 2018), or three–dimensional mesoporous Co3O4 flowers constructed by many nanoneedles (Liao et al., 2018) was reported.
All supported catalysts showed low surface area ranging from 0.16 to 0.71 m2 g− 1 (calculated per gram of the catalyst, i.e. including the stainless steel meshes). For better comparison of the catalysts, their surface area was recalculated per gram of the active oxides (Co, Mn, and/or Co–Mn ones). The surface area of the active oxides changed from 11.6 to 25.7 m2 g− 1 (Table 1).
The results of H2–TPR measurements obtained in the temperature range from 25 to 900°C are shown in Fig. 3 and Table 1. TPR profiles (Fig. 3) reflected various phase composition and Co and Mn contents in the catalysts. The 1.0 sample containing only Co3O4 showed one main reduction peak with Tmax at 395°C and a shoulder at somewhat lower temperature. This peak was ascribed to the reduction of Co3+ to Co2+ and Co2+ to Co0 (Calgaro &Perez-Lopez, 2017, Lin &Chen, 2004). Reduction of the 0 catalyst containing only Mn2O3 proceeded in two steps with temperature maxima at 355 and 442°C. The reduction process included Mn3O4 as an intermediate; reduction of MnO2/Mn2O3 to Mn3O4 followed by reduction of Mn3O4 to MnO was reported (Kapteijn et al., 1994, Wu et al., 2016). The reduction of MnO2/Mn2O3 can proceed also in one step without Mn3O4 intermediate formation (Arena et al., 2001) depending on the conditions of TPR experiments. Incorporation of Mn into cobalt oxide manifested in broadening of temperature range, in which the reduction of the catalysts proceeded. The TPR experiments revealed reduction of some other mixed oxide components formed from carbonate precursors during calcination. The highest number of such components was observed with the 0.4 and 0.5 catalysts. Four distinguished peaks were observed in TPR profile of the 0.4 catalyst with Tmax at 335, 400, 490, and 593°C. In the TPR profile of the 0.5 catalyst, the reduction peaks were overlapping and a hint of new low–temperature peak can be seen at 227°C. This temperature maximum can be ascribed to the reduction of adsorbed oxygen species (Chen et al., 2017, Luo et al., 2018) and/or to Co4+ reduction, which can be formed by oxidation of surface Co3+ according to equation ½ O2 + Co3+ → O− + h (Co4+), where h is a positive hole, i.e., orbital without electron (Kaczmarczyk et al., 2016). The finding agrees well with the published values of Tmax for Co4+ reduction at about 200°C (Jung et al., 2019). Reduction of the commercial Co–Mn–Al mixed oxide catalysts (the Com sample) proceeded in two temperature regions, the first between 100 and 500°C and the other between 500 and 900°C. Both regions consist of some overlapping reduction peaks. Li et al. observed a similar TPR profile of Co–Mn–Al mixed oxide obtained at 500°C from a coprecipitated precursor (Li et al., 2019).
The FTIR spectra of the examined catalysts are presented in Fig. 4. The spectra were taken in the range from 4000 to 400 cm− 1 but their analysis was possible only in the range from 400 to 900 cm− 1 due to interferences from the meshes forming artificial bands at wavenumbers higher than 900 cm− 1. In the range from 400 to 900 cm− 1, the stretching vibrations of metal–oxygen bonds can be found. Intensity of the bands is very low due to low concentration of active components in the supported catalysts; nevertheless, some interesting features can be recognized in the obtained spectra. Stretching vibrations at 554 and 659 cm− 1 are typical for Co3O4 with spinel structure (Bentley et al., 1968, Jirátová et al., 2017). In the FTIR spectrum of the 1.0 catalyst, the bands 580, 648, and 667 cm− 1 were observed. These bands correspond to Co3O4, with variations between 560–574 cm− 1 for the first band and 659–661 cm− 1 for the second one. The first band ν1 at 580 cm− 1 was attributed to the spinel lattice vibrations connected with Co cations in octahedral sites. However, it seems that the first band is composed of more than one component. The second band ν2 at about 658 cm− 1 was ascribed to vibrations of bonds between oxygen and Co in tetrahedral sites. The FTIR spectrum of the 0 catalyst containing only Mn oxide showed very broad but distinct band with maximum at 719 cm− 1. This band can be attributed to the Mn–O bond (Dubal et al., 2010, Wu et al., 2017). In the Co–Mn mixed oxide catalysts, similar stretching vibrations of Co–O and Mn–O bonds at 568 and 656 cm− 1 were observed (Habibi &Bagheri, 2017) and ascribed to the vibrations of Co cations in octahedral and tetrahedral oxygen environment related to the Co–O and/or Mn–O bonds in spinels, respectively (Martin de Vidales et al., 1995, Ramachandran et al., 2015). The band at 860 cm− 1 can be attributed to δ(C–O) vibrations, coming likely from the rest of MnCO3 remaining in the catalysts even after calcination (Udayabhanu et al., 2017).
The surface composition in the near–surface region of the catalysts and chemical state of the elements were determined by XPS. As a carbon tape was used for fixing of the samples to a holder, it could manifest itself in a higher concentration of C. Thus, the calibration of the spectra was carried out according to a metal bound oxygen (529.6 eV). Binding energies of core level electrons are shown in the Supplementary material (Table S1) and the surface concentrations of the components are summarized in Table 2. In addition to the main components (Co, Mn, O) given in Table 2, Fe was also detected on the catalysts surfaces, likely coming from a slight erosion of stainless steel meshes during hydrothermal treatment in the solutions containing hydrolyzing urea.
Table 2
Surface composition of the catalysts (at. %) determined by XPS.
Co/(Co + Mn)a
|
C 1s
|
O 1s
|
Co 2p
|
Mn 2p
|
Fe 2p
|
0
|
47.77
|
37.97
|
0.00
|
12.76
|
1.50
|
0.2
|
53.00
|
35.71
|
2.40
|
7.60
|
1.30
|
0.4
|
56.92
|
32.51
|
4.35
|
4.41
|
1.82
|
0.5
|
52.38
|
37.70
|
3.74
|
2.18
|
4.01
|
0.6
|
58.38
|
33.10
|
5.18
|
2.73
|
0.62
|
0.8
|
50.49
|
37.56
|
7.02
|
1.79
|
3.14
|
1.0
|
66.04
|
28.41
|
3.38
|
0.00
|
2.16
|
a Nominal molar ratio |
Deconvolution of the Co 2p, Mn 2p and O 1s peaks is reported in Table 3. The Co 2p peak consisted of the two maxima, which corresponded to Co 2p1/2 and Co 2p3/2 orbitals. The spin-orbital splitting (15.2 ± 0.05 eV) as well as broad satellite structure (Biesinger et al., 2011) points to the presence of Co2+ and Co3+ (Hagelin-Weaver et al., 2004). The peak at 780 eV was assigned to tetrahedral Co2+, while the peaks at 781.3 and 783.5 eV were ascribed to octahedral Co3+. This is in line with the literature data (Langell et al., 2000) and our earlier works (Chromčáková et al., 2015, Obalová et al., 2013, Obalová et al., 2007), although a reverse approach was also reported (Iablokov et al., 2015).
Table 3. Deconvolution of Co 2p, Mn 2p and O 1s peaks, molar ratios of Co2+/Co3+, Mn3+/Mn4+, and surface oxygen vacancies Ov/(Co+Mn) from XPS spectra.

The Mn 2p3/2 peak centered at 642.3–641.9 eV was ascribed to Mn3O4 and MnO2 (Kim &Shim, 2010). Determination of the manganese oxidation state was done as reported previously (Obalová et al., 2013, Pacultova et al., 2019). The Mn 2p3/2 spectrum was fitted by two peaks corresponding to Mn3+ (the component with lower binding energy) and Mn4+ (the component with higher binding energy). The Mn3+ ion is a strong oxidizing agent and its disproportionation to give Mn2+ and MnO2 is reported. Summarizing, based on the XPS results (Table 3) it can be concluded that Mn3+ is prevailing on the surface of the examined catalysts. The Co2+ is prevailing on the surface of most catalysts, with the exception of the 0.2 and 0.5 samples, where the concentration of Co3+ is higher by 20 or 10 %, respectively. Table 4shows a relation between the surface and bulk concentrations of the catalyst components. Compared to the bulk, surface of the examined catalysts is enriched with cobalt, which is mostly pronounced in the 0.5 catalyst. Surface concentration of manganese is suppressed, again mostly in the 0.5 catalyst.
Table 4
Relative concentration (in %) of Co and Mn in the catalysts determined by EDX and XPS and relation of surface to bulk concentrations of individual components.
Co/(Co + Mn)a
|
EDX
|
XPS
|
XPS/EDX
|
Co
|
Mn
|
Co
|
Mn
|
Co
|
Mn
|
0
|
-
|
-
|
0.00
|
100.00
|
-
|
-
|
0.2
|
20.48
|
79.52
|
24.02
|
76.08
|
1.17
|
0.96
|
0.4
|
41.51
|
58.49
|
49.71
|
50.40
|
1.20
|
0.86
|
0.5
|
49.32
|
50.68
|
63.28
|
36.89
|
1.28
|
0.73
|
0.6
|
55.31
|
44.69
|
65.57
|
34.56
|
1.19
|
0.77
|
0.8
|
78.34
|
21.66
|
79.68
|
20.32
|
1.02
|
0.94
|
1.0
|
-
|
-
|
99.71
|
0.00
|
-
|
–
|
a Nominal molar ratio |
The deconvolution of oxygen spectra revealed three peaks with binding energies of about 529.8, 531.5, and 533.0 eV. The first peak at 529.8 eV (Ol) was attributed to the lattice oxygen in the metal oxides, the second Ov centered at 531.5 eV corresponded to the oxygen vacancies, and the third Ow around 532.3 eV was assigned to the hydroxyl species of surface adsorbed water (Feng et al., 2016, Wang et al., 2017). The number of oxygen vacancies belonging to Co and Mn oxides (Table 3) was calculated as the ratio of the oxygen vacancies Ov concentration to the sum of Co and Mn concentrations (in at. %).
3.2 Activity and selectivity in the ethanol total oxidation
The prepared catalysts were tested in the total oxidation of ethanol. Ethanol conversions over the examined catalysts in dependence on reaction temperature are demonstrated in Fig. 5 and characteristic values describing catalytic activity and selectivity of the prepared catalysts are summarized in Table 5. Temperatures T50, at which 50 % conversion of ethanol was achieved, varied in the range from 185°C for the 0.2 catalyst to 207°C for the 1.0 catalyst. The deposition of Co–Mn oxide thin layer enabled its high utilization in ethanol oxidation due to practically negligible influence of internal diffusion. Despite more than 85 times lower content of Co–Mn oxides in the catalytic bed, the 0.2 catalyst exhibited practically the same catalytic performance in terms of T50 as the commercial pelleted Co–Mn–Al catalyst (Table 5). Specific catalytic activity R200, expressed as mmol of ethanol converted per gram of oxides and hour at 200°C, was chosen to compare better the performance of examined catalysts containing various amounts of active Co–Mn oxides. The highest specific activity was found for the 0.5 catalyst, despite it showed slightly higher T50 temperature than the 0.2 sample. In order to evaluate the stability of the supported catalysts with time on stream, four consecutive catalytic experiments were performed. While the R200 of 34.2 mmol·goxides−1·h− 1 was achieved for the 0.5 catalyst in the first experiment, this value reached 35.9 mmol·goxides−1·h− 1 in the fourth experiment and the deactivation of the catalyst was not observed.
Table 5
Catalytic properties of the catalysts with various Co/(Co + Mn) molar ratios deposited on stainless steel meshes compared to commercial Co–Mn–Al mixed oxide pellets (Com). Reaction conditions: 8 meshes in the bed, 770 ppm of ethanol in air, reaction temperature increase 2°C min− 1 from 50 to 400°C, GHSV 20 l g− 1h− 1.
Co/
(Co + Mn)a
|
T50
(°C)
|
T90(CO2)
(°C)
|
R200b
(mmol goxides−1h− 1)
|
AcAmaxc
(ppm)
|
Tmaxd
(°C)
|
COmaxe
(°C)
|
Tmaxf
(°C)
|
Relative activityg
|
0
|
200
|
248
|
6.7
|
485
|
226
|
37
|
245
|
22
|
0.2
|
185
|
244
|
14.5
|
566
|
217
|
95
|
236
|
48
|
0.4
|
195
|
251
|
14.7
|
601
|
227
|
149
|
249
|
49
|
0.5
|
195
|
253
|
34.2
|
609
|
225
|
163
|
256
|
114
|
0.6
|
194
|
264
|
15.0
|
506
|
221
|
108
|
235
|
50
|
0.8
|
189
|
254
|
23.6
|
619
|
221
|
156
|
247
|
79
|
1.0
|
207
|
260
|
20.6
|
579
|
243
|
76
|
261
|
69
|
Com
|
189
|
353
|
0.3
|
20
|
164
|
0
|
-
|
1
|
a Nominal molar ratio |
b Catalytic activity in mmol of ethanol converted per gram of oxides and hour at 200°C |
c Maximum concentration of acetaldehyde during the experiment |
d Temperature at which the maximum concentration of acetaldehyde was achieved |
e Maximum concentration of carbon monoxide during the experiment |
f Temperature at which the maximum concentration of carbon monoxide was achieved |
g Catalytic activity in terms of R200 related to that of the Com catalyst |
Typical evolution of reaction byproducts during ethanol oxidation is shown in Fig. 6. Acetaldehyde was detected as the main byproduct of ethanol oxidation over all supported catalysts. Maximum acetaldehyde concentrations over the supported catalysts were achieved at the temperatures ranging from 217 to 243°C and varied from 485 ppm for the 0 catalyst to 619 ppm for the 0.8 one (Table 5). Evolution of acetaldehyde over the pelleted Co–Mn–Al mixed oxide catalyst (the Com sample) was considerably lower (up to 20 ppm only) and the maximum of its formation occurred at lower temperature (164°C). The maximum evolution of CO was found over the most active 0.5 catalyst (163 ppm), while no CO was detected over the Com catalyst. Importantly, the T90(CO2) temperature needed to attain 90% conversion of all carbon components to carbon dioxide was much lower with the supported catalysts (from 244 to 260°C) than with the Com sample (353°C).
A strong interaction between Co and Mn components in the Co–Mn mixed oxide catalysts was confirmed by H2–TPR (Fig. 3). Incorporation of Mn changed reduction behavior of the cobalt oxide. While the 1.0 catalyst exhibited the reduction maximum at 395°C and the 0 catalyst containing only Mn oxide showed low–temperature maximum at 355°C, enhanced reducibility was observed for the supported Co–Mn mixed oxides. The H2–TPR profiles of 0.2, 0.4, 0.6, and 0.8 samples revealed shift of the low–temperature maximum reduction to 335°C. Furthermore, the 0.5 catalyst exhibited a shoulder at about 227°C, which was tentatively ascribed to reduction of Co4+ components (Mitran et al., 2020). Thus, the high catalytic activity of the 0.5 catalyst could be connected with its improved reduction behavior. It was reported recently that the reduction of Co-Mn mixed oxides occurs in two steps (Bulavchenko et al., 2018). In the first step, (Mn,Co)3O4 is reduced to form (Mn,Co)O solid solutions; Co2MnO4 oxide starts to be reduced to Co0.6Mn0.4O first and then Mn–rich Mn2CoO4 is transformed into Mn0.6Co0.4O. In the second step, the (Mn,Co)O solid solutions are reduced to metallic cobalt and MnO. Similar behavior can be observed for other Co–Mn oxide catalysts.
We reported previously a good correlation between H2–TPR profiles and performance of the catalysts in ethanol oxidation (Gaálová et al., 2019, Topka et al., 2019, Topka &Klementová, 2016). The above-mentioned correlation of catalysts reducibility with catalytic performance of Co–Mn mixed oxide catalysts points to the redox (Mars – van Krevelen) reaction mechanism. This idea is supported by the XPS results shown in Fig. 7. They indicate that, regardless of the catalyst composition, the catalytic activity (shown here as relative to the Com catalyst) correlates with the amount of surface oxygen vacancies. Surface oxygen vacancies have been recently reported to play a crucial role in VOC oxidation (Lin et al., 2018, Yang et al., 2019). These defect sites on the surface of metal oxides adsorb oxygen from the air and transform it into catalytically active species. Similarly to our results, Mitran et al. recently observed that replacing of Co3+ with Mn4+ in Co–Mn oxides up to Mn0.15Co2.85 leads to the formation of active sites for ethanol oxidation (Mitran et al., 2020). Interaction of Co3+ and Mn4+ with lattice oxygen created oxygen vacancies, which served as catalytically active sites.