Visualizing hydrogen-induced reshaping and edge activation in MoS2 and Co-promoted MoS2 catalyst clusters

Hydrodesulfurization catalysis ensures upgrading and purification of fossil fuels to comply with increasingly strict regulations on S emissions. The future shift toward more diverse and lower-quality crude oil supplies, high in S content, requires attention to improvements of the complex sulfided CoMo catalyst based on a fundamental understanding of its working principles. In this study, we use scanning tunneling microscopy to directly visualize and quantify how reducing conditions transforms both cluster shapes and edge terminations in MoS2 and promoted CoMoS-type hydrodesulfurization catalysts. The reduced catalyst clusters are shown to be terminated with a fractional coverage of sulfur, representative of the catalyst in its active state. By adsorption of a proton-accepting molecular marker, we can furthermore directly evidence the presence of catalytically relevant S–H groups on the Co-promoted edge. The experimentally observed cluster structure is predicted by theory to be identical to the structure present under catalytic working conditions.


Reviewer #3 (Remarks to the Author):
This is a very interesting work which reported the HDS induced restructuring and the related analysis at atomic scale. Authors observed restructuring of edge of MoS2 in HDS and the difference between MoS2 and MoCoS. DFT calculations were used to understand their findings. The following points need to be addressed before considering acceptance.
(1) The title should be specific. STM, the main technique used here should be added to distinguish from TEM observation. Particle is not a specific term. Nanocluster is more appropriate. It should be included in the title. Particles were used twice in the title.
(2) Line 85, "in the range of 10^-4 mbar at 400C". The pressure is not specific. I believe it is at 10^-4 mbar (3) Early studies of restructurings such as bimetallic nanoparticles and vicinal surface Pt were not put in context in the discussion of restructuring of edge of MoS2 nanoclusters.
(4) The statement of restructuring of s-MoS2 to r-MoS2 is clearly supported by STM images. However, the lack of restructuring from s-CoMoS to r-CoMoS cannot be clearly seen from Figure 2b. (5) The number of nanoclusters were used in the statistical account in Figure 2c should be described for each of the four statistical accounts in the caption. What are the confidence level?. (6) Details and methods of statistical accounts should be added to methods. (7) STM is the main techniques used in this work. XPS is sensitive for the change of binding from qualitative point and also good for uncovering the formation of sulfur vacancies from quantitative point. Photoemission feature would be important information to support the conclusion.
The authors study the shape and termination of (Co-promoted) MoS2 nanoparticles supported on gold, using scanning tunneling microscopy and density functional theory.
According to their results, the authors show a selective structural transition of completely sulfided, Ausupported, MoS2 (s-MoS2) nanoparticles upon exposure to reductive conditions. This structural transition, from triangular to hexagonal nanoparticles, has been already documented in previous literature by the authors. However, in this case, they also demonstrate that the structural transition seems to be specific to MoS2 only, while edge activated CoMoS nanoparticles retain a hexagonal symmetry when in a sulfided (s-CoMoS) or reduced state (r-CoMoS).
The authors thoroughly discuss the STM characterization of the r-MoS2 and r-CoMoS nanoparticle edges. They also give atomic-level details on the presence of sulfur vacancies, through a proper use of STM line scans. For MoS2, experiments and DFT seems to be generally in good agreement and in agreement with the past literature. The case of CoMoS seems to be more complex and discrepancies between experiments and DFT-calculated phase diagram arise.
In particular, the authors show the presence of "quenched" edges in r-CoMoS (i.e. zones of attenuated intensity in Fig. 3c) that were explained as arising from the S interaction with the gold support. The authors suggest that the presence of adsorbed H at the edge restores the original unquenched state, due to the weakened S-Au interaction. This explanation seems to be sound, as hydrogen is effectively found to be present on the edge through pyridine titration. However, this finding is in contrast with the edge composition predicted by the phase diagram (S50H0 in Fig. 5). According to the latter, the structure should not show any adsorbed hydrogen. The authors try to justify this behavior invoking kinetic effects that prevent hydrogen to desorb as H2.
It should be noticed that the same S-Au interaction responsible for the quenching effect could be expected also on the unactivated r-MoS2 (as stated by the authors on line 302); yet in this case no quenched edges were found. This could mean that these differences may be explained in terms of the intrinsically different chemical properties of Co and Mo that here is not analyzed in depth.
The manuscript is well-written and the methodology employed is solid. The topic addresses an important aspect for modelling hydrodesulfurization on Co/MoS2 particles, which is of great interest for the catalysis community, even though real catalysis does not utilize Au as the support. The article should be publishable in NATURE COMMUNICATIONS, provided the authors address the following points: 1. The authors should provide additional information regarding the size of their surface models.
The distance between periodically repeated MoS2 nanoparticles is of particular interest to exclude artificial image-image interactions. 2. The second equation in the SI should read "del n = 2n_Mo -n_S". The authors should check, whether this is a typing error or whether it affects the results of their study. 3. To potentially resolve the discrepancy between DFT and experiments about the presence of kinetically trapped hydrogen on the edge, it would be interesting to perform calculations for evaluating the diffusion barrier of hydrogen along the edge. An easy diffusion would substantiate the DFT prediction of the S50H0 structure. High diffusion barriers would instead substantiate the hypothesis of kinetically trapped hydrogen, difficult to desorb as H2. 4. The STM collected after pyridine titration is discussed by the authors ~ line 267. In particular, they say that: "The outermost edge features in the STM images may reflect a complex convolution of electronic and geometric structure of the adsorbed molecule, but the symmetry and size agrees well with pyr-H + ". In order to have an unambiguous answer on the presence of pyr-H + , it could be interesting to perform explicit DFT calculations on the adsorption of pyridine on the CoMoS hydrogenated edge. In other words, I would confirm the adsorption structure proposed in the empirical sketch of Figure 6b by means DFT. This would allow also to understand if an effective S-Au interaction occurs in presence of pyridine/pyr-H + , confirming the explanation proposed by the authors in lines 269-271. 5. Provide charge density plot to better elucidate the interaction between S and Au, which according to the authors, is responsible for the quenched edge effects. 6. In Figure 1d, the white text is very difficult to read. The authors might want to use a darker outline to increase readability.
Reviewer #1: 1. The structures of s-MoS2 and s-CoMoS are, with no doubt, clearly presented and carefully analyzed, also in their previous studies. The analysis in Fig. 2c seems difficult to comprehend because the authors did not give detailed structural models for both r-MoS2 and r-CoMoS. At first sight, it looks contradictory that the hydrogenation reaction produces more S edges. This should be explained in more details along with atomic STM images for r-MoS2 and r-CoMoS before coming to statistical analysis.
Reply: We refer to question 2 which addresses both question 1 together with question 2.

Responses to Reviewer Comments:
We do not see any significantly truncated s-MoS2 particles, which can be seen from the histogram in Figure 2c. Here s-MoS2 is generally rather perfect triangular (f~1  or f~0 (big s-MoS2) and there are no counts at or around f=0.5.
Action: To avoid the risk of confusion we added the following description to the main text: 3. In this study, the authors increased the H2 partial pressure to 10-4 mbar, which is two orders of magnitude higher than their previous studies. This brought the question that whether the reduced structure observed is relevant to the structure under industrial conditions and a high pressure of a few torr H2 will bring new changes after reduction. How stable is CoMoS?
Thank you for this relevant question. Overall, the conclusion of our paper is indeed that the absolute pressure strongly defines which edge structures are exposed for both MoS2 and CoMoS. Reviewer #2: 1. The authors should provide additional information regarding the size of their surface models.
The distance between periodically repeated MoS2 nanoparticles is of particular interest to exclude artificial image-image interactions.
We agree that this is important information and we have provided a more detailed description in the method section.
2. The second equation in the SI should read "del n = 2n_Mo -n_S". The authors should check, whether this is a typing error or whether it affects the results of their study.
Thank you for pointing this out, the typo is now corrected.    Rangarajan et al. (ACS Catal. 2016, 6, 2904−2917 in which the author determines the energies and activation barriers for forming the pyridinium ion upon pyridine adsorption on a hydrogenated 50%S Co S edge (Figure 2 below). Here they first of all find that the formation for the adsorbed pyridinium ion is thermodynamically favorable and that the pyridinium formation has no energy barrier. Hence these calculation clearly concludes that pyridine adsorption on the hydrogenated edges indeed form the pyridinium ion as proposed in our empirically based ball model in the original manuscript.  5. Provide charge density plot to better elucidate the interaction between S and Au, which according to the authors, is responsible for the quenched edge effects.
Reply: Reviewer 1 made a similar point in his/her question 5 which further emphasizes the need of such simulated STM data (charge density plot), and we refer to the replies and actions taken in our reply number 5 to reviewer 1.
6. In Figure 1d, the white text is very difficult to read. The authors might want to use a darker outline to increase readability.
Reply: Thank you for pointing this out. We have changed the figure so the text is now outside the basal plane of MoS2.

Reviewer #3:
(1) The title should be specific. STM, the main technique used here should be added to distinguish from TEM observation. Particle is not a specific term. Nanocluster is more appropriate. It should be included in the title. Particles were used twice in the title.

Reply:
We have now changed the title to "Visualizing hydrogen-induced reshaping and edge activation in MoS2 and Co-promoted MoS2 catalyst clusters using Scanning Tunneling Microscopy" and changed the word "particle" to "cluster" throughout the paper.
(2) Line 85, "in the range of 10^-4 mbar at 400C". The pressure is not specific. I believe it is at

10^-4 mbar
Reply: We agree and we have changed the text as suggested.
(3) Early studies of restructurings such as bimetallic nanoparticles and vicinal surface Pt were not put in context in the discussion of restructuring of edge of MoS2 nanoclusters.
Reply: Thank you for pointing this out, we have now added citations to some of the early work in the field of gas induced reconstructions, noting that the effect for compound nanoclusters such as sulfides are less well understood than corresponding metallic systems.
(4) The statement of restructuring of s-MoS2 to r-MoS2 is clearly supported by STM images.
However, the lack of restructuring from s-CoMoS to r-CoMoS cannot be clearly seen from Figure 2b.