Atomic Layer Deposition for Preparation of Highly Efficient Catalysts for Dry Reforming of Methane
Round 1
Reviewer 1 Report
This review article describes the use of an innovative procedure for the preparation of heterogeneus catalysts. The procedure consists in the atomic layer deposition (ALD) described as a method to prepare nickel catalysts with TiO2 or MgO as supports for dry reforming of methane.
The paper is clearly written, however as a review it misses some informations . In particular the following issues should be considered:
The loading of Ni is important and it is not mentioned in the manuscript.
The experimental conditions of the mentioned catalytic dry reforming tests should be specified.
In order to help the reader understanding the peculiarity of the procedure a better comparison could be the catalytic dry reforming behavior between ALD prepared samples and analogous Ni catalyst prepared by conventional means such as impregnation or co-precipitation.
For the figures 3 and 5, in the figure captions the temperature of the catalytic tests should be added, since in the text it is only mentioned a temperature under 800 °C.
Line 166 which can contribute to removal……
Author Response
<Response to reviewers' comments.>
Reviewer 1.
This review article describes the use of an innovative procedure for the preparation of heterogenous catalysts. The procedure consists in the atomic layer deposition (ALD) described as a method to prepare nickel catalysts with TiO2or MgO as supports for dry reforming of methane. The paper is clearly written, however as a review it misses some information. In particular the following issues should be considered:
1. The loading of Ni is important, and it is not mentioned in the manuscript.
Author reply: The amount of Ni loading determined by ICP-OES (inductively coupled plasma-optical emission spectrometry) is additionally mentioned in the revised version of the manuscript. Please see line 307-308.
2. The experimental conditions of the mentioned catalytic dry reforming tests should be specified.
Author reply: As the reviewer suggested, more details of experimental conditions of DRM reactions with ALD-prepared catalysts, e.g., reactor type, reactant gas flow, mixing ratio of CO2and CH4, are additionally mentioned in the revised version of the manuscript.
Please see line 160-163 and line 309-314.
3.In order to help the reader understanding the peculiarity of the procedure a better comparison could be the catalytic dry reforming behavior between ALD prepared samples and analogous Ni catalyst prepared by conventional means such as impregnation or co-precipitation.
Author reply: This is additionally discussed in the revised version of manuscript. Please see line 322-347.
4. For the figures 3 and 5, in the figure captions the temperature of the catalytic tests should be added, since in the text it is only mentioned a temperature under 800 °C.
Author reply: The experimental conditions including reaction temperature (800 oC) are added in the respective figure caption as suggested by the reviewer.
Please refer to the figure captions of figure 3 and 5 of the revised version of the manuscript.
5. Line 166 which can contribute to removal……
Author reply: It is corrected to “which can contribute to removal….” from “which can attribute to removal....”, as suggested by the reviewer.Please see line 207.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Please find my review enclosed
Comments for author File: Comments.pdf
Author Response
<Response to reviewers' comments.>
Reviewer 2.
The paper is a review of the use of ALD method of preparation for making catalysts in the dry reforming of methane (DRM). In that sense it is sound, and the literature is correctly quoted. Schematics are clear, as well as Figures. It deserves publication after major revision.
1. I understand that to Ni were added MgO, TiO2(which form, anatase or rutile?), and ZnO (appears at line 171), and that ALD was also used to disperse Ni in mesoporous silica (which one?) and in MIL-101(Cr) (not used in catalysis I guess). All this should be stated clearly at the end of Introduction.
Author reply: Two different types of structures prepared by employing ALD as heterogeneous catalysts for DRM reaction are discussed in this article; 1) shell/core type metal-oxide/metal catalyst prepared by depositing additional metal-oxide thin films (MgO, TiO2, ZnO) using ALD on Ni particles, and 2) Ni nanoparticles confined in the mesoporous supporting templates prepared by ALD deposition of NiO on mesoporous materials (silica, alumina, MIL-101) and subsequent thermal annealing in a reductive atmosphere. NiO/MIL-101 (Cr) was not used as DRM catalyst, and it was mentioned in this article to show trench-fill capability of ALD and structural uniqueness of metal oxide layers prepared by ALD. This is additionally mentioned at the end of Introduction as suggested by the reviewer. Please refer line 69-76 of the revised version of the manuscript.
Since the ALD process operates generally below 300 oC, metal oxide layers prepared by ALD show amorphous structures.Please refer line 134-136.
2.Only XRD and SEM are used as characterization methods which is quite poor for such a paper. At least the composition of catalysts (Ni/MgO and the like) should be given. A Table with composition, (number of cycles, see caption to Figure 3, thickness, mean size (SEM), etc. would be of great interest.
Author reply: We added a table (table 1) summarizing the specific area, pore volumes of the samples (bare Ni, MgO/Ni, TiO2/Ni) determined by Brunauer-Emmett-Teller (BET) and Barret-Joyner-Halenda (BJH) method, respectively. The thickness of ALD deposited metal oxide films (MgO or TiO2) on Ni particles can be determined by TEM analysis for some samples (200 cycled MgO/Ni and 500 cycled TiO2/Ni) and they are also summarized in table 1. Please refer line 139-142 with table 1 in the revised version of the manuscript.
3.As any experiment may be reproduced by other researchers, experimental details should be provided, at the least when the authors consider their own work. This is mandatory for catalytic experiments (line 127) whose operating conditions are never detailed (reactor, loading, contact time, methane to carbon oxide ratio, and the like). CH4conversion should be shown on Figure 3.
Author reply: As the reviewer suggested, more details of experimental conditions of DRM reactions with ALD-prepared catalysts, e.g., reactor type, reactant gas flow, mixing ratio of CO2and CH4, are additionally mentioned in the revised version of the manuscript.
Please see line 160-163 and 309-314.
Also, CH4conversion also additionally shown in Figure 3. Please refer to the revised Figure 3.
4.Several imprecisions. The “commercially available mesoporous silica”, which provenance (Figure 9, caption line 243)? Same question for line 101, the Ni particles, and the “precursor vapors”. Caption to Figure 3, meaning of “cycles”, which refers probably to ALD technique (written “thickness”in caption)? Line 218, “mesoporous substrate”. Is it the MOF?
Author reply: Mesoporous silica with a mean pore size of ~ 12 nm, particle size 200-425 mesh and pore size distribution from 2-20 nm was purchased from Sigma Aldrich and used as mesoporous supporting materials. This is additionally mentioned. Please see line 303-305.
Ni powder with a diameter of 0.5-1 mm was purchased from Sigma Aldrich.Please see line 122.
Ti precursor (Ti[OCH(CH3)2]4) bought from Sigma Aldrich was used for TiO2deposition, while Mg precursor (Mg(Cp)2) purchased from EG Chem Co., Ltd used for MgO deposition. For both cases (TiO2and MgO deposition), H2O vapor was used as oxidizing agent.Please see line 126-129.
“Cycle” in each graph legend of Figure 3 refer to the number of ALD cycles used for deposition of each metal oxide (MgO or TiO2) on Ni particle. Its meaning is additionally provided in Figure caption. Please see the revised caption of Figure 3.Determination of the thickness of metal oxide layers (MgO and TiO2) prepared by small number of ALD cycles, e.g., 50 ALD cycled MgO and 100 ALD cycled TiO2layers, based on TEM images was very challenging, since the thickness of the respective metal oxide layer was too thin. Therefore, we keep the notation for each metal oxide/Ni sample of the number of ALD cycles used for metal oxide deposition.
We did not specify “mesoporous substrate” in line 273-274 to MOF(MIL-101(Cr)) intentionally in order to discuss the general situation of layer-by layer deposition of thin films on the internal wall of pores.
5. Line 100, better introduce Figure1b, tell why a structured alumina and what for “nanostructured TiO2”(it seems very interesting…).
Author reply: AAO membrane with a regularly ordered pore structure has been considered as a very promising material for separation filter and gas sensor since particles having larger size than the pores can be selectively filtered. AAO membrane can be fabricated by anodization of Al foil and its pore size can be controlled by in few nanometer scale (generally few tens of nanometers’ regime) by changing anodization conditions. However, in conjunction with ALD process, pore size can be controlled in atomic scale, moreover, surface chemical functionally can be controlled. The ALD-prepared nanostructure TiO2with a regularly ordered pore structure showed superior toluene adsorption efficiency; the toluene adsorption capacity of TiO2nanostructure prepared by applying 300 ALD cycles was 3.8 toluene molecules/nm2. This is additionally mentioned in the revised version of the manuscript.Please see line 103-106.
6. Line 101, it is strange –at first sight –to wrap the active phase for DRM (Ni0) by an oxide and not the contrary, “supporting”Ni onto an oxide (MgO, ZnO, TiO2). What was the reason for this strategy? Fortunately, line 123, there are cracks for Ni to appear and act as a catalyst. Pictures of those cracks are welcome…
Author reply: Generally, catalytic active metal nanoparticles on metal oxide supports are considered as heterogeneous catalysts. And their catalytic activity can be varied upon the choice of metal-oxide materials as supports by providing dissimilar metal/oxide interface sites as well as altering the geometrical structures of the supported metal particles. Alternatively, metal oxide can be deposited on metal particles, which are often referred as inverse-catalyst. The study of those inverse catalysts allows us to elucidate the contribution of metal-oxide interface on catalytic activity, and some inverse catalysts exhibited even higher activity than regular structure catalysts (metal/metal oxides) consist of same compounds. Owing to excellent trench-fill capability and controllability of metal oxide films thickness of ALD as mentioned above, ALD technique can be very useful in study of inverse catalysts by fabricating various thin metal oxide films with different thicknesses on metal particles. In addition, the prepared thin metal oxide films on metal nanoparticles can also improve the catalytic stability by preventing aggregation of metal nanoparticles during the catalytic reaction. In this review article, the experimental results showing utilization of ALD in fabrication of various metal oxide thin films (MgO, TiO2, ZnO) on Ni particles are summarized including their catalytic behaviors towards DRM reaction.
These are additionally mentioned in the revised manuscript. Please see line 107-121.
SEM images of cracks existed on ALD-prepared metal thin films (MgO and TiO2) on Ni particles were not easy to be obtained which was probably due to their small dimensions. However, increase of pore volumes upon the ALD-deposition of metal oxide films (MgO and TiO2) implies the existence of cracks on metal oxide films. In addition, the fact that catalytic activity of Ni particles was improved by depositing MgO and TiO2layers using ALD also indicates the crack formation on metal oxide films; the surface of Ni particles was still available for CH4and CO2molecules since their surface was not completely blocked by those metal oxide films due to existence of cracks.Please see line 142-153 and line 179-186.
7.Some sentences have to be revised in accordance with experiments: Don’t forget that 11 (72 hrs fig.3) hrs are not enough to decide if a catalyst is efficient or not
Author reply: We carefully revised sentences in accordance with experimental data shown in this article.
8.Line 142 “For Ni particles wrapped by TiO2or MgO using ALD, one can find improvement in both initial catalytic activities and sustainability of the catalytic activity with respect to those of bare Ni”, assertion true for MgO, not for TiO2.
Author reply: Those sentences were revised as follow, “For Ni particles wrapped by MgO using ALD, one can find improvement in both initial catalytic activity and sustainability of the catalytic activity with respect to those of bare Ni. On the other hands, initial catalytic activities of TiO2/Ni catalysts were slightly higher or lower than those of bare Ni depending on the number of ALD cycles applied for TiO2deposition (100 cycles or 500 cycles). However, the ALD deposition of TiO2thin layers (either 100 or 500 cycles) always resulted in improved sustainability of catalytic activity of Ni particles. These results indicate that TiO2or MgO deposited on Ni do not completely block the active Ni sites, which can be attributed to the aforementioned crack formation within TiO2or MgO layers wrapping Ni particles.”. Please see line 179-186.
9. Line 175 “The primary role of MgO and TiO2layers for enhancing DRM catalytic activity and selectivity seems to be analogous: No, TiO2does not promote activity (as written in following lines), its carbonate is not so stable as Mg is. ZnO is even more acidic (line 181), so it favors carbon growth.
Author reply: Promotion of catalytic activity of Ni particles by ALD-deposition of TiO2films was depending on the thickness of TiO2, however, the ALD-deposition of metal-oxide thin layers (MgO and TiO2) always resulted in the catalytic stability of Ni particles. Therefore, one can conclude that the primary role of MgO and TiO2layers for enhancing DRM catalytic stability seems to be analogous, which is geometric perturbation of coke formation by inducing carbon filament growth. These sentences were revised and please see line 179-186, 216 and 219-223.
10. Line 250-251 “Ni incorporated in the mesoporous silica showed very high stability of DRM catalytic activity (Figure 9a)”. But the experiment lasted only 11 hr vs. 72 hr for others. Would it be the reason why no carbon was observed?
Author reply: The catalytic activity test of Ni/silica was conducted for 72 hours at 800 oC and we corrected an error in the time scale (min -> hours) of Figure 9a. For comparison, results of DRM reaction with Ni particles under same experimental conditions (for 72 hours at 800 oC) were added as Figure 9b. We would also like to mention that Ni/silica catalysts did not undergo noticeable deactivation towards DRM reaction at 800 oC for 7 days (168 hours). The results of DRM reaction with Ni/silica at 800 oC for 7 days were also added in Figure 9 (Figure 9 d). Please see line 315-321 with the revised Figure 9 (figure 9a, 9b, and 9d).
11. Lines 256-260: Please tell more about “It is worth mentioning that Ni nanoparticles supported by mesoporous silica, alumina (where?) and titania showed almost identical”(but this is not the case in this work) “activity and stability for DRM catalytic reaction, indicating that as far as Ni nanoparticle catalysts supported by stable mesoporous substrate are concerned, (the) chemical composition of the substrate is less important for catalytic behavior of Ni nanoparticles [58].”This comment is at variance with all catalytic experiments made with nickel, for steam reforming since 60 years, as for DRM since 30 years. So, why doing that work?
Author reply:
Ni nanoparticles supported by mesoporous TiO2and alumina were prepared by applying 50 ALD cycles, and their catalytic performance towards DRM reaction were examined under same experimental conditions used for the case of Ni/silica catalysts (CH4and CO2conversions of Ni/TiO2and Ni/alumina catalysts are additionally provided in Figure 10a and 10b). Mesoporous TiO2substrate was prepared by depositing TiO2thin films on mesoporous silica (Sigma Aldrich, a mean pore diameter of ~ 12 nm) using ALD technique, whereas mesoporous alumina with a mean pore diameter of ~ 11.6 nm was used as purchased from the company (Sasol).Ni nanoparticles (~ 10 nm) supported by mesoporous silica, TiO2and alumina showed very similar activity and stability for DRM catalytic reaction.It indicates that as far as Ni nanoparticle catalysts (~ 10 nm) supported by stable mesoporous substrate are concerned, chemical composition of the substrate is less important for catalytic behavior of Ni nanoparticles rather than the size of Ni nanoparticles confined by porous structure of supporting materials. These are additionally mentioned in revised manuscript with Figure 10. Please see line 353-366 with Figure 10.
12. Finally the influence of acidity/basicity of the added oxides, and/or of their relative reducibility at 800 °C in the presence of H2and CO is not discussed or scarcely mentioned, although the quality of the paper would be enhanced.
Author reply: The influence of the chemical nature of metal oxide supports, such as acidity/basicity and reducibility, on catalytic behaviors of metal particles towards DRM reaction has been pointed out by many researchers. For instance, basic sites existed on metal oxides such as La2O3and MgO can enhance the activation of CO2, which can reduce the carbon formation and catalysts deactivation [91]. On the other hands, some research groups demonstrated that CO2activation can also take place on acidic metal oxides via the reaction with hydroxyls on the surface of acidic supports, but the CO2activation on acidic supports is weaker than that on basic supports [28,92,93]. The catalytic DRM reaction of Rh catalysts supported on various reducible and irreducible metal oxides were examined and the irreducible metal oxide supports generally led to higher catalytic activity of the supported metal catalyst than the case of irreducible metal oxide supports [91]. However, it has been also reported that reducible CeO2can promote catalytic activity of the Ni catalysts for DRM reaction by acting as the oxygen accumulator [94]. These are additionally mentioned in the revised manuscript. Please see line 230-241.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors have greatly improved the manuscript which is now publishable in this journal.
Only a few minor corrections:
line 108 heterogeneous catalysts. Their catalytic activity......
line 113 catalyst exhibiting....
line 184 improved catalytic activity.....
Author Response
Reviewer 1.
The authorshave greatly improved the manuscript which is now publishable in this journal.
Only a few minor corrections:
1. line 108 heterogeneous catalysts. Their catalytic activity......
Author reply: The relating paragraph was re-written to address issues raised by reviewer 2. English of entire manuscript was proofread by native speakers (using proofreading service of eWorldEditing).
2. line 113 catalyst exhibiting....
Author reply: It was corrected as “catalysts exhibit”. Please refer to line 116.
3. line 184 improved catalytic activity.....
Author reply: It was corrected as “catalytic sustainability of Ni particles”. Please refer to line 189.
Author Response File: Author Response.pdf
Reviewer 2 Report
Please see attached review
Comments for author File: Comments.pdf
Author Response
<Response to reviewers' comments.>
Reviewer 2.
The paper has been revised. The answers to questions are satisfactory. Please check again the level of English (even in the added texts). Minor revision is asked for English edition and the following:
No doubt that ALD is a nice method for tailoring catalysts and that the authors’ contribution is good. But in literature 99% papers using ALD deal with metal supported on oxides (when the active phase is metallic). Therefore, the methodology (interesting by itself, I fully agree) chosen by the authors consisting in supporting MgO or TiO2 onto nickel particles has to be better introduced on the point of view of catalysis. The quality of the paper would be greatly enhanced. Moreover lines 110-113, the term “inverse catalyst” is mentioned: “Alternatively, metal oxide can be deposited on metal particles, which are often referred as inverse-catalysts...”. This term “inverse catalyst” is not appropriate: the support covers the active phase, this is completely unusual and a priori a weird idea since the active sites are masked. So, the authors must (for their paper to be better understood) show a bit of pedagogy. Moreover, writing “often” means that several papers (coming from other authors) have appeared in literature and so they must be quoted. I noticed that this expression is used in electrochemical catalysis and it would be interesting to comment on the differences with the “normal” catalysis studied here.
Author reply: Conventionally, catalytic active metal supported by metal oxides are considered as heterogeneous catalysts. Both the geometrical structures (size, shape, morphology) and electronic natures (oxidation states) of the supported metal are related to its catalytic activity. The catalytic activity of conventional metal/metal oxide catalysts can be also influenced by underlying metal oxide supports. The geometrical and electronic structures of the supported metals can be influenced by underlying metal oxides. In addition, the catalytic behaviors of metal/metal oxide catalysts can be varied upon the choice of metal oxide materials as supports by providing dissimilar metal/metal oxide interface sites as well as participation of metal oxides into the catalytic reaction. Alternatively, metal oxide can be deposited on metal particles (metal oxide/metal), which are often referred as inverse-catalyst [83-86]. These metal oxide/metal systems have been studied to elucidate the contribution of metal oxide interface on catalytic activity since 1940s [83-86]. A large number of research groups have investigated these inverse-catalysts, and it has been reported that some inverse catalysts exhibit even higher activity than regularly structured catalysts (metal/metal oxides) consists of same compounds [84,87]. More details on “inverse-catalysts” are added in the revised manuscript as suggested by reviewer. Please refer to line 105-118.
Lines 117-118: “In addition, the prepared thin metal oxide films on metal nanoparticles can also improve the catalytic stability by preventing aggregation of metal nanoparticles during the catalytic reaction.” This phenomenon is called sintering. But in the paper the authors insist on the formation or not of coke, not on sintering which is not studied.
Author reply: Those statements are removed from the manuscript. Please refer to revised manuscript.
Lines 138-139: “Only upon a proper post-annealing process of the ALD-prepared structure, stoichiometric metal oxide layers such as TiO2 and MgO can be formed [13,30]”. Please describe the operating conditions for “a proper post-annealing process”, as well as for the primary annealing process (temperature, ramp, atmosphere, etc.) as it well known that they determine the textural and structural properties.
Author reply: For the case of TiO2, 5 hrs of annealing at 800 oC under N2atmosphere led to the formation stoichiometric TiO2layers [30]. Those statements are corrected, and annealing conditions are additionally mentioned as suggested by reviewer.Please see line 141-143.
Author Response File: Author Response.pdf
