Propane Dehydrogenation over Cobalt Aluminates: Evaluation of Potential Catalytic Active Sites

Round 1

Reviewer 1 Report

Manuscript: catalysts-2686615

Title: Propane dehydrogenation over cobalt aluminates: Evaluation of 2 potential catalytic active sites

The aim of the study was to clarify the nature of the active sites of cobalt–alumina catalysts in the non-oxidative dehydrogenation of propane (PDH). A series of CoOx-Al₂O₃ catalysts were prepared by co-precipitation and were characterized using XRD, H2-TPR, TEM and XPS after calcination and regeneration and TEM-EDX-STEM analysis was provided for selected spent catalysts. As characterization was performed ex situ, results are not conclusive and discussion is rather speculative.  Nevertheless, the attempt was honest, introduction and discussion try to present the different points of view reported in literature and the results provided are interesting. Summarizing, the manuscript merits to be published after a thorough revision.

Specific points which may be improved:

Results

1. Table 1

1a. As Co containing and Al containing solutions have the same concentration, the second column of Table 1 represents also molar ratio. So, it is better to express it in this way.

1b.Columns 2 (Co:Al)  and 3 (Co wt. %) do not match.  Taking into account the 3^rd column, Co:Al molar ratio should be higher in all catalysts. Please verify data given in Table 1.  

2. lines 66-71 and Table 1 –Textural properties

In the text, it is mentioned that textural properties “are quite comparable despite the varying cobalt content”. This is not totally accurate. CoAl-0.1 exhibits a specific surface area about 30% higher that that of CoAl-0.25. In addition, there is a progressive decrease in total pore volume as cobalt content increases. Therefore, it would be better to revise.

3. Textural properties

It would be interesting to also show the N₂ adsorption –desorption isotherms and the pore size distribution. In such materials, these can provide useful pieces of information. Please, do.

4. lines 80-75 and Figure 1.

4a. It is not accurate to conclude that ‘the crystallinity level of Co₃O₄ in the CoAl-0.25-R sample has increased, whereas in the case  of CoAl-0.1-R catalyst the opposed tendency takes place’. From the XRD patterns (shown in Fig.1), it is obvious that in case of CoAl-0.1-R catalyst, the peak at ~31,3  degrees also increases. Please, check the calculations of data shown in Table S1.

4b. In Fig. 1, only the peaks attributed at Co₃O₄ and Al₂O₃ are indicated.  Although the peaks due to CoAl₂O₄ spinel appear at angles very close to the peaks of alumina, it would be better to also indicate the position of CoAl₂O₄ spinel peaks.

5. lines 89-98, Figure 2-H₂-TPR profiles

Regarding the comments on the H₂-TPR profiles of CoAl-0.5, CoAl-0.25 and Co/Al₂O₃ catalysts, it is not clear whether the peak at 500 °C is attributed to the reduction of Co₃O₄ to CoO and the hydrogen consumption at higher temperatures is due to further reduction of cobalt to metallic phase. Taking into account the catalysts’ synthesis procedure, one would expect that cobalt aluminate is formed in all catalysts. This phase is difficult to reduce. Yet, no reduction occurs at any degree in catalysts with low Co content? Please, clarify the text.

6. lines 100-103, Fig. 3, TEM analysis

6a. Considering the synthesis method of CoAl-0.1 sample, TPR profile and Fig. 3b, one would expect CoAl₂O₄ to be the dominant phase in this sample and not CoOx particles on alumina.  Please, comment.

6b. Fig.3c is unclear and not convincing regarding the existence of CoOx particles. Please, provide a better proof.

7. Catalytic tests

Since selectivity is below 95%, which are the other products? Specifically, in the case of CoAl-0.25 and during the first hour of reaction, selectivity is less than 85%. Was there any analysis attempted? Please, explain in the text.

8. lines 143—148, Effect of the treatment of the CoAl-0.1 and CoAl-0.25 catalysts,

Figs. 7 & 8

8a. Trying to explain the results in Fig. 7, it seems that during the induction period, the catalyst is reduced by the active hydrogen produced in situ. Please, comment.

8b. The loss in activity of the regenerated catalysts (Figs 7 & 8) is probably due to catalyst’s sintering and the important decrease of the specific surface area (Table 1).

8b. Why does the decrease of the activity enhance  selectivity to propylene?

The paragraph presents the catalytic results but there is no attempt to explain them. Please, consider revising.

9. lines 157-163

 The catalytic properties of the “reference” Co/Al₂O₃ catalyst may be comparable to that of CoAl-0.05. Yet, it is not accurate to state that these catalysts have similar characteristics.  None of the data presented herein (textural properties in Table 1, XRD, TPR, XPS) supports this claim. Please, consider revising.

10. lines 166-174, Fig. 9

10a. According to this paragraph, the rapid deactivation oft CoAl-0.1 catalyst is due to destruction of cobalt aluminate and the sintering of metal particles (shown in Fig. 9e). However, EDX provides an elemental analysis and does not distinguish the oxidation state of the element.

10b. Does HAADF-STEM image in Fig. 9c allow to distinguish the nature (i.e. phase) of the indicated particles? Does magnification permit this distinction?

10c. Fig. 9b shows a cobalt-containing particle with metallic Co in the core, covered by CoOx. How can this be formed? Please, explain.

Please, address the above issues.

11. lines 175-177 regeneration of CoAl-0.1 catalyst

It is really remarkable that the metallic cobalt nanoparticles, during regeneration at 600 °C in the O₂ flow for 1.5 h, form again cobalt aluminate. Temperature and treatment duration seem rather low. Please, comment.

12. lines 180-184

It is not surprising that CoAl-0.25-R is not so easily reduced. As it can be observed in Table 1, it has sintered during regeneration and cobalt phase may have been covered by alumina. Please, comment

13. characterization of spent and regenerated CoAl-0.1 and CoAl-0.25 catalysts.

It would be interesting to provide the XRD patterns of spent catalysts. Is it possible?

Discussion

14. lines 193-206

14a. It is correct that in calcined CoAl-0.1 sample Al2O3 and cobalt aluminate are present and the presence of small CoOx particles is suggested.  What happens during the induction period? Which of these phases changes forming active sites? Which is the reason of the fast drop in activity?

14b. Comparing CoAl-0.25 catalyst to CoAl-0.1, the lower selectivity to  propylene and the longer induction period has been attributed to the easily reduced Co3O4 phase, present in CoAl-0.25 and absent from CoAl-0.1.  CoAl-0.25-R is less reducible (Fig. 2), has larger Co3O4 crystallites, exhibits better selectivity to propylene and has shorter induction period. So, the same question arises: What happens during the induction period?

Discussion should clarify (as possible) these points.

15. lines 235-238

Co in cobalt oxides has an octahedral coordination, whereas tetrahedral coordination is observed in the case of cobalt aluminate. On the other hand, why should oxygen vacancies (OVs) be present only in tetrahedral coordination?  Please, consider revising.

DRS analysis could shed light on the abundance of each type of phases in the catalysts.

16. lines 253-268,

The creation of OVs and i.e. the surface reduction of the cobalt oxidic phase is an interesting hypothesis. The rate of OVs creation could explain the variations in induction periods. On the other hand, in situ pre-reduced with hydrogen catalysts also exhibit an induction period. Still, this is just a hypothesis as it is not the direct conclusion of any of the applied techniques.  Perhaps, at attempt to “titrate” these OVs by the adsorption of probe molecule (e.g. before the test and at the higher point of activity) could give a more solid conclusion.

Conclusions

17. lines 359-360

See comment 15 and revise accordingly.

18.  lines 359-366

Speculations are allowed in discussion but not in conclusions, as there is no solid proof of the  existence of OVs, provided by the results of the study. This paragraph should be revised.

Materials and methods

19. lines 277-289

A few questions regarding catalysts preparation:

a. Which was the rhythm of NH₄OH addition?

b. Was there an aging period?

c. Why is it necessary to remove ammonia? How was this verified?

Abstract

20. lines 12-27

It is not clear in the text how were cobalt oxide, metallic cobalt nanoparticles and tetrahedral Co2+ species in the CoAl2O4 spinel ‘evaluated as potential active-site ensembles based on the obtained kinetic data’. First, there are no kinetic data. Recording changes in conversion as a function of time-on-stream is not considered as a kinetic study.  Second, there is no link between propane conversion or selectivity with any of these phases, as, according to the text, all of them are present in the active CoAl-0.1 catalyst. Please, consider revising.

Author Response

Issue raised by the reviewer:

1a. As Co containing and Al containing solutions have the same concentration, the second column of Table 1 represents also molar ratio. So, it is better to express it in this way.

Changes in the manuscript: We have made the suggested changes in Table 1.

1b. Columns 2 (Co:Al)  and 3 (Co wt. %) do not match.  Taking into account the 3rd column, Co:Al molar ratio should be higher in all catalysts. Please verify data given in Table 1.

Answer:  For pure cobalt aluminate CoAl₂O₄, the maximal (theoretical) content of Co is 33.3%.  That value could be expected for CoAl-0.5 sample when the taken molar ratio of Co and Al is 1:2. However, the value determined by ICP is 28.8%, likewise the actual values of the Co contents are lower for CoAl-0.25 catalysts. This can be explained by the presence of CoO_x in the oxidation state higher than (II). In addition, the deviations can be attributed to incomplete and none stoichiometric precipitation of hydroxides during the catalysts preparation. On the other hand, the contents of Co = 6.7 and 6.9 found for CoAl-0.1 and CoAl-0.1-R catalysts match well with the theoretical value of 6.9% for 10% CoAl₂O₄ and 90% Al₂O₃.  

Issue raised by the reviewer:

2. lines 66-71 and Table 1 –Textural properties

In the text, it is mentioned that textural properties “are quite comparable despite the varying cobalt content”. This is not totally accurate. CoAl-0.1 exhibits a specific surface area about 30% higher that that of CoAl-0.25. In addition, there is a progressive decrease in total pore volume as cobalt content increases. Therefore, it would be better to revise.

Changes in the manuscript: The text is revised accordingly:

The BET surface area is the highest for as-prepared CoAl-0.1 (291 m² g⁻¹), which is comparable to that of CoAl-0.05 (274 m² g⁻¹), despite the varying cobalt content, namely 6.7 and 4.3%, respectively. In addition, the total pore volume is gradually increasing as the cobalt content decreases. However, the BET surface areas likewise the pore volumes of spent-regenerated samples CoAl-0.25-R and CoAl-0.1-R have reduced markedly.

3. Issue raised by the reviewer:

Textural properties

It would be interesting to also show the N2 adsorption –desorption isotherms and the pore size distribution. In such materials, these can provide useful pieces of information. Please, do.

Changes in the manuscript:  We have added this data in the Supplementary Materials.

4. Issue raised by the reviewer:

lines 80-75 and Figure 1.

4a. It is not accurate to conclude that ‘the crystallinity level of Co3O4 in the CoAl-0.25-R sample has increased, whereas in the case of CoAl-0.1-R catalyst the opposed tendency takes place’. From the XRD patterns (shown in Fig.1), it is obvious that in case of CoAl-0.1-R catalyst, the peak at ~31,3  degrees also increases. Please, check the calculations of data shown in Table S1.

Answer:   Indeed, in case of CoAl-0.1-R catalyst, one can see an increased peak at ~31.3° in the XRD pattern. This, however, is mainly due to enhanced contribution of the CoAl2O4 and Al2O3 peaks at ~31.3-32.0 degrees. In fact, this narrow and overlapping area is not convincing for the purpose of comparison. Therefore, determination of average crystallite sizes and weight ratios of Co₃O₄/Al₂O₃ was carried out for a range of 2θ = 63−70°, which is more probative. This is mentioned in the experimental section 4.3.    

4b. In Fig. 1, only the peaks attributed at Co3O4 and Al2O3 are indicated.  Although the peaks due to CoAl2O4 spinel appear at angles very close to the peaks of alumina, it would be better to also indicate the position of CoAl2O4 spinel peaks.

Answer:  The peaks related to CoAl₂O₄ are practically coinciding with those of Co₃O₄ as the latter also forms spinel crystals.

Changes in the manuscript:  We indicated positions of peaks related to CoAl₂O₄ spinel by green dash lines in Fig. 1. 

5. Issue raised by the reviewer:

lines 89-98, Figure 2-H2-TPR profiles

Regarding the comments on the H2-TPR profiles of CoAl-0.5, CoAl-0.25 and Co/Al2O3 catalysts, it is not clear whether the peak at 500 °C is attributed to the reduction of Co3O4 to CoO and the hydrogen consumption at higher temperatures is due to further reduction of cobalt to metallic phase. Taking into account the catalysts’ synthesis procedure, one would expect that cobalt aluminate is formed in all catalysts. This phase is difficult to reduce. Yet, no reduction occurs at any degree in catalysts with low Co content? Please, clarify the text.

Answer:  Indeed, the peak at 500 °C is attributed to the reduction of Co3O4 to CoO. The hydrogen consumption at higher temperatures is due to further reduction of cobalt to metallic phase. In accord with the given literature [35,36], cobalt aluminate indeed is difficult to reduce,  and reduction is quite negligible for the fresh catalysts with low Co content.    

Changes in the manuscript:  To clarify that, one sentence was altered accordingly:

\\For these catalysts, the reduction temperature of the first peak, attributed to the reduction of Co₃O₄ to CoO, is about 500 °C, that is notably higher than that of bulk Co₃O₄ (200–400 °C [35,36]) due to strong interaction of Co₃O₄ with Al₂O₃, which prevents the catalyst reduction.//

6. Issue raised by the reviewer:

lines 100-103, Fig. 3, TEM analysis

6a. Considering the synthesis method of CoAl-0.1 sample, TPR profile and Fig. 3b, one would expect CoAl2O4 to be the dominant phase in this sample and not CoOx particles on alumina.  Please, comment.

Answer:  Yes, indeed,CoAl2O4 is the dominant phase in this sample.  

Changes in the manuscript:  To clarify that, the text was altered accordingly:

//According to TEM data (Figure 3), CoAl-0.1 sample consists mainly of a separate phase of cobalt aluminate CoAl₂O₄ (Figure 3b). Besides, the catalyst contains Al₂O₃ crystals, which are uniformly covered by CoO_x particles having a size of about 1–2 nm (Figure 3с).//

6b. Fig.3c is unclear and not convincing regarding the existence of CoOx particles. Please, provide a better proof.

Answer: A feature of the HAADF method is the dependence of the received signal on the mass of the chemical element. Since cobalt is more than twice heavier than aluminum, the white particles can be attributed to the cobalt signal. In addition, according to EDX data of this area of the sample, the amount of cobalt in comparison with aluminum is rather low, so it is logical to assume that it is cobalt oxide that is observed in the image.  

Changes in the manuscript:  Figure 3 c was replaced by another with better resolution. In addition, new text is added:

// Notably, in Figure 3 the white particles are attributed to cobalt based on the relative mass (atomic number) of scattering Co and Al atoms, which affect the HAADF signal differently. //

7. Issue raised by the reviewer:

Catalytic tests

Since selectivity is below 95%, which are the other products? Specifically, in the case of CoAl-0.25 and during the first hour of reaction, selectivity is less than 85%. Was there any analysis attempted? Please, explain in the text.

Answer:  Yes, of course, monitoring of the impurity of such possible by-products as methane, ethane, ethylene, and C4-hydrocarbons was also performing.

Changes in the manuscript:  To clarify that, new text is added:

//It should be noticed that under the specified reaction conditions, especially during the induction period, propane underwent concurrent cracking reactions yielding mainly methane, along with carbonaceous material (see below), which is typical for the high temperature PDH [2–10].//

8. Issue raised by the reviewer:

lines 143—148, Effect of the treatment of the CoAl-0.1 and CoAl-0.25 catalysts, Figs. 7 & 8

8a. Trying to explain the results in Fig. 7, it seems that during the induction period, the catalyst is reduced by the active hydrogen produced in situ. Please, comment.

8b. The loss in activity of the regenerated catalysts (Figs 7 & 8) is probably due to catalyst’s sintering and the important decrease of the specific surface area (Table 1).

8c. Why does the decrease of the activity enhance  selectivity to propylene?

The paragraph presents the catalytic results but there is no attempt to explain them. Please, consider revising.

Answer:  All questions raised are actually discussed in section 3 – Discussion.

9. Issue raised by the reviewer:

lines 157-163

The catalytic properties of the “reference” Co/Al2O3 catalyst may be comparable to that of CoAl-0.05. Yet, it is not accurate to state that these catalysts have similar characteristics.  None of the data presented herein (textural properties in Table 1, XRD, TPR, XPS) supports this claim. Please, consider revising.

Answer:  We have revised this.

Changes in the manuscript:  The text is changed:

\\This is not surprising, taking into account the close content of cobalt (Table 1), while other characteristics of these catalysts are somewhat different.\\

10. Issue raised by the reviewer:

lines 166-174, Fig. 9

10a. According to this paragraph, the rapid deactivation oft CoAl-0.1 catalyst is due to destruction of cobalt aluminate and the sintering of metal particles (shown in Fig. 9e). However, EDX provides an elemental analysis and does not distinguish the oxidation state of the element.

10b. Does HAADF-STEM image in Fig. 9c allow to distinguish the nature (i.e. phase) of the indicated particles? Does magnification permit this distinction?

10c. Fig. 9b shows a cobalt-containing particle with metallic Co in the core, covered by CoOx. How can this be formed? Please, explain.

Please, address the above issues.

Answer:  Fig. 9e illustrates the sintering, while Fig 9b discloses the structure of the particles formed. The metallic Co nanoparticles are normally oxidized to CoOx due to contact with air even at room temperature [33]. The nature (i.e. phase) of the indicated particles in Fig. 9c is actually the same as in Fig. 3c. See, please, the answer to question 6.   

11. Issue raised by the reviewer:

lines 175-177 regeneration of CoAl-0.1 catalyst

It is really remarkable that the metallic cobalt nanoparticles, during regeneration at 600 °C in the O2 flow for 1.5 h, form again cobalt aluminate. Temperature and treatment duration seem rather low. Please, comment.

Answer:  The metallic Co nanoparticles undergo actually quick oxidation followed by reaction of CoOx with Al2O3. The reaction time appeared to be enough, while the conditions are comparable with those for the synthesis of CoAl-0.1 (from cobalt and aluminum hydroxides in air at 650 ^oC within 4 h).   

12. Issue raised by the reviewer:

lines 180-184

It is not surprising that CoAl-0.25-R is not so easily reduced. As it can be observed in Table 1, it has sintered during regeneration and cobalt phase may have been covered by alumina. Please, comment

Answer:  We have actually commented this (though in other words) in section 3: “…CoAl-0.25-R, showing better selectivity to propylene and shortened induction period (Figure 8), differs from CoAl-0.25 by increase in Co3O4 particles size (up to 100 nm, Figure S4) and their increased crystallinity (Figure 1, Table S1). Alongside, CoAl-0.25-R became practically non-reducible with hydrogen until 700 oC (Figure 2). Furthermore, the high-temperature treatment with oxygen both of the spent CoAl-0.1 and CoAl-0.25 catalysts resulted in a marked decrease in specific surface area (Table 1), that may be one of the reasons of the partial loss of their activity.”

Concerning the possibility that cobalt phase may have been covered by alumina, this is not supported by the TEM data (Figure S4).

13. Issue raised by the reviewer:

characterization of spent and regenerated CoAl-0.1 and CoAl-0.25 catalysts.

It would be interesting to provide the XRD patterns of spent catalysts. Is it possible?

Answer:  We did this, but the quality of the XRD patterns was poor due to strong carbonization of the samples. In general, the same peaks were observed as for the fresh and regenerated samples, but in poor quality. At least the crystalline phase of the metallic Co (which we tried to see) did not appear, as well as any carbon crystals were not detected due to their amorphous nature. 

14. Issue raised by the reviewer:

lines 193-206

14a. It is correct that in calcined CoAl-0.1 sample Al2O3 and cobalt aluminate are present and the presence of small CoOx particles is suggested.  What happens during the induction period? Which of these phases changes forming active sites? Which is the reason of the fast drop in activity?

14b. Comparing CoAl-0.25 catalyst to CoAl-0.1, the lower selectivity to propylene and the longer induction period has been attributed to the easily reduced Co3O4 phase, present in CoAl-0.25 and absent from CoAl-0.1.  CoAl-0.25-R is less reducible (Fig. 2), has larger Co3O4 crystallites, exhibits better selectivity to propylene and has shorter induction period. So, the same question arises: What happens during the induction period?

Discussion should clarify (as possible) these points.

Answer:  Both these points - 14a and 14b – comprise the key questions of this and other studies on PDH catalyzed by metal oxides.  We have considered them in detail in the Discussion, and it is difficult to add something else.

During the induction period, the tetrahedral Co²⁺ species should be formed exhaustively. After that, all the forms of CoOx containing Co²⁺ species may be partially reduced to metallic Co via the intermediacy of OVs. According to previous (cited) points of view, only tetrahedral Co²⁺ species or C⁰ could be the active catalytic sites in PDH. In contrast, we suggest the involvement of OVs as the key active sites.  They appear to be largely formed during the induction period.

15. Issue raised by the reviewer:

lines 235-238

Co in cobalt oxides has an octahedral coordination, whereas tetrahedral coordination is observed in the case of cobalt aluminate. On the other hand, why should oxygen vacancies (OVs) be present only in tetrahedral coordination?  Please, consider revising.

DRS analysis could shed light on the abundance of each type of phases in the catalysts.

Answer:   The Co₃O₄ oxide had a spinel structure where Co²⁺ occurs in the tetrahedral position 8a (1/8, 1/8, 1/8), and Co³⁺ occurs in the octahedral position 16d (1/2, 1/2, 1/2). In cobalt aluminate Co²⁺ also occurs in the tetrahedral position 8a (1/8, 1/8, 1/8). So, the main form of Co in fresh and especially in partially reduced catalysts with low content of Co is tetrahedral.  

16. Issue raised by the reviewer:

lines 253-268,

The creation of OVs and i.e. the surface reduction of the cobalt oxidic phase is an interesting hypothesis. The rate of OVs creation could explain the variations in induction periods. On the other hand, in situ pre-reduced with hydrogen catalysts also exhibit an induction period. Still, this is just a hypothesis as it is not the direct conclusion of any of the applied techniques.  Perhaps, at attempt to “titrate” these OVs by the adsorption of probe molecule (e.g. before the test and at the higher point of activity) could give a more solid conclusion.

Answer:  We agree, that we suggested only hypothesis without solid experimental proofs. One of them may be titration of OVs, as well as other known techniques [50-58]. However, to be credible, these experiments must be performed under high temperature conditions, which is a challenge. This cannot be addressed within the frames of one paper but which may be thought worthy of further study.

17. Issue raised by the reviewer:

Conclusions

lines 359-360

See comment 15 and revise accordingly.

Answer:  In Conclusions, we have stated that “The structural motif such as tetrahedral Co2+ species located in the CoAl2O4 and Co3O4 spinel forms as well as in the small CoOx particles tightly covering alumina can be regarded as the feasible primary active-site ensembles.” That means that tetrahedral Co2+ species are most actual, but not exclusive.

18. Issue raised by the reviewer:

lines 359-366

Speculations are allowed in discussion but not in conclusions, as there is no solid proof of the  existence of OVs, provided by the results of the study. This paragraph should be revised.

Answer:  We have revised this.

Changes in the manuscript:  One part of the former text in lines 359-366 was placed in Discussion. Another part was changed:

 // However, their subsequent evolution under PDH reaction conditions, consisting in partial reduction to metallic cobalt through the intermediacy of surface OVs, may provide further rationale for the origin of the catalytic activity of cobalt aluminates in PDH. \\

It should be noted that the intermediacy of OVs in the reduction of metal oxides to metals is well known and does not require proof.

19. Issue raised by the reviewer:

lines 277-289

A few questions regarding catalysts preparation:

1. Which was the rhythm of NH4OH addition?
2. Was there an aging period?
3. Why is it necessary to remove ammonia? How was this verified?

Answer:  a. The rate of NH4OH addition was determined by watching the level of pH = 7.5 of the reaction mixture (as it is said in section 4.2). 

1. No special aging period was employed.
2. Not ammonia only. It was necessary to remove the residual (none precipitated) hydroxides. This was easier to control by controlling the presence of ammonia with pH meter.

20. Issue raised by the reviewer:

Abstract

lines 12-27

It is not clear in the text how were cobalt oxide, metallic cobalt nanoparticles and tetrahedral Co2+ species in the CoAl2O4 spinel ‘evaluated as potential active-site ensembles based on the obtained kinetic data’. First, there are no kinetic data. Recording changes in conversion as a function of time-on-stream is not considered as a kinetic study.  Second, there is no link between propane conversion or selectivity with any of these phases, as, according to the text, all of them are present in the active CoAl-0.1 catalyst. Please, consider revising.

Answer: The catalytic behavior of the phases was evaluated based on their relative content in catalysts vs to catalytic efficiency towards PDH.

Changes in the manuscript:  We have corrected the Abstract by replacing //obtained kinetic data\\ with \\ catalytic performance data //.

The authors express their gratitude to the reviewer for the evaluating of the work and valuable advice and comments.

Author Response File: Author Response.doc

Reviewer 2 Report

This manuscript presents a study of propane dehydrogenation (PDH) on a variety of cobalt catalysts supported on alumina, aiming to unravel the activity contribution from different species including CoOx ensembles, crystalline Co3O4, metallic Co, and oxygen vacancy. The evolution between these species was also investigated and the induced impacts on activity and selectivity were discussed. The materials were also sufficiently characterized. The research results are of interest for PDH Co catalyst design and this work is systematic and rigorous. My concerns mainly lie in its presentation since I found the logic of the discussion and conclusion may be improved to make this manuscript be considered for publication. 

My specific comments:

1. In the introduction section, the authors mentioned a bunch of studies about PDH on supported oxide catalysts, “namely CoOx, VOx, GaOx, MoOx, FeOx, WOx, InOx, ZnO, ZrO2 and TiO2, deposited on silica, alumina, zeolites and other oxide carriers [2–10]”. Citations should be added individually for each catalyst rather than citing several review articles to cover all these catalysts. Recent publications below should be cited as the representative reports for PDH on VOx, GaOx, ZnOx.

ChemCatChem 2020, 13, 882–899

J. Am. Chem. Soc. 2022, 144, 15079–15092

ACS Catal. 2023, 13, 4971–4984

2. The presentation of the discussion is not clear to understand the focus of each paragraph. It reads to me like the first four paragraphs are about the roles of CoOx, Co3O4, and metallic Co, while the latter part discusses the role of oxygen vacancy. However, it is rather difficult to follow and understand what is the exact impact of each species. The authors may improve the logic by discussing one point in one graph, possibly including the impact of Co3O4 particle size, the impact of metallic Co, and the evolution from Co2+ to metallic Co.

3. Are there any references that metallic Co can lead to catalyst deactivation, as mentioned: “Moreover, a slow reduction of CoOx to metallic cobalt followed by its aggregation to cobalt nanoparticles (Figure 9b) can rather be considered as the reason of the catalysts progressive deactivation.”

Author Response

To Reviewer #2

1. Issue raised by the reviewer:

In the introduction section, the authors mentioned a bunch of studies about PDH on supported oxide catalysts, “namely CoOx, VOx, GaOx, MoOx, FeOx, WOx, InOx, ZnO, ZrO2 and TiO2, deposited on silica, alumina, zeolites and other oxide carriers [2–10]”. Citations should be added individually for each catalyst rather than citing several review articles to cover all these catalysts. Recent publications below should be cited as the representative reports for PDH on VOx, GaOx, ZnOx.

ChemCatChem 2020, 13, 882–899

1. Am. Chem. Soc. 2022, 144, 15079–15092

ACS Catal. 2023, 13, 4971–4984

Answer: Thank you for the latter citations on the subject, we have added two of them (2-nd and 3-d under numbers [12] and [13], as the 1-st is related to oxidative reactions), while we did not give them individually as they do not related directly to cobalt catalysts – the main subject of our study.  

2. Issue raised by the reviewer:

The presentation of the discussion is not clear to understand the focus of each paragraph. It reads to me like the first four paragraphs are about the roles of CoOx, Co3O4, and metallic Co, while the latter part discusses the role of oxygen vacancy. However, it is rather difficult to follow and understand what is the exact impact of each species. The authors may improve the logic by discussing one point in one graph, possibly including the impact of Co3O4 particle size, the impact of metallic Co, and the evolution from Co2+ to metallic Co.

Answer: To clarify this, new text below Scheme 1 is added in Discussion.

// Summarizing the above considerations, nanocrystalline CoAl₂O₄ and Co₃O₄, as well as fine CoOx particles on alumina (all present in CoAl active catalysts) may exhibit the same catalytic properties due to similar evolution under PDH conditions, namely partial reduction to metallic cobalt (ultimately) through the intermediacy of surface OVs. Small content of cobalt in CoAl catalysts ensures its strong interaction with alumina and embedding in the spinel structure, which generally prevents deep reduction of CoOx to Со⁰ under PDH conditions and provides the catalysts relative stability. Based on these assumptions, the progress in the design of cobalt-containing catalysts and similar transition metal oxide catalytic systems will require taking into account their optimal ability to generate surface OVs along with resistance to subsequent reduction and metallization. //

3. Issue raised by the reviewer:

Are there any references that metallic Co can lead to catalyst deactivation, as mentioned: “Moreover, a slow reduction of CoOx to metallic cobalt followed by its aggregation to cobalt nanoparticles (Figure 9b) can rather be considered as the reason of the catalysts progressive deactivation.”

Answer: The metallic Co itself (at least in case of small nanoparticles) is not a cause of the catalyst deactivation, while the presence of large particles can be the reason of cracking reactions over Co-catalysts (Catal. Lett. (2015) 145:1413–1419. DOI 10.1007/s10562-015-1533-4, Ref. [21] - former [19]). In many cases, as we quoted in our manuscript, small nanoparticles of metallic Co are even regarded as the key catalytic sites in PDH. However, when the catalyst works well in an oxide form, its partial deep reduction to metallic form should be considered as undesirable side reaction, spoiling the catalyst.

However, we have replaced the mentioned phrase in the Discussion with another phrase: 

// Moreover, a slow reduction of CoO_x to metallic cobalt followed by its aggregation to cobalt nanoparticles (Figure 9b) is accompanied by the catalysts progressive deactivation, which may be related. // 

The authors express their gratitude to the reviewer for the evaluating of the work and valuable advice and comments.

Author Response File: Author Response.doc