Inducing skyrmions in ultrathin Fe films by hydrogen exposure

Magnetic skyrmions are localized nanometer-sized spin configurations with particle-like properties, which are envisioned to be used as bits in next-generation information technology. An essential step toward future skyrmion-based applications is to engineer key magnetic parameters for developing and stabilizing individual magnetic skyrmions. Here we demonstrate the tuning of the non-collinear magnetic state of an Fe double layer on an Ir(111) substrate by loading the sample with atomic hydrogen. By using spin-polarized scanning tunneling microscopy, we discover that the hydrogenated system supports the formation of skyrmions in external magnetic fields, while the pristine Fe double layer does not. Based on ab initio calculations, we attribute this effect to the tuning of the Heisenberg exchange and the Dzyaloshinsky–Moriya interactions due to hydrogenation. In addition to interface engineering, hydrogenation of thin magnetic films offers a unique pathway to design and optimize the skyrmionic states in low-dimensional magnetic materials.

That being said, I am not convinced that this study constitutes a breakthrough that deserves publication in Nature Communication, for two main reasons: (i) I don't see how the present work advances the understanding of chiral magnetism at the surface of transition metals. Explaining the skyrmion radius enhancement due to H-intercalation is reasonable but not unexpected. Considering that this enhancement is essentially due to lattice distortion (increasing distance between the two Fe layers), such distortions are not surprising.
(ii) I am not convinced the present work could open new avenues for the engineering of skyrmion lattices. Upon H exposure, both H1-Fe and H2-Fe phases appear in a non-controlled manner, which renders the skyrmion engineering using this technique quite questionable. In addition, the presence of skyrmion lattice seems to be very sensitive to structural distortions, omnipresent in ultrathin films, which makes it hardly adapted to engineering purposes.
In summary, this is a nice study performed in a rigorous manner; I also acknowledge that this is probably the first observation of skyrmion modification under H exposure. That being said, I do not see how disruptive this observation is and how is can open "novel avenues" for skyrmion engineering. In my view, this manuscript deserves publication in a specialized journal, but not in a broad-audience journal such as Nature Communication.

Reviewer #3 (Remarks to the Author):
In their manuscript "Inducing skyrmions in ultrathin Fe films by hydrogen exposure", Dr. Hsu and co-workers describe spin-STM experiments and a theoretical interpretation of significant effects of hydrogen absorption on the magnetic properties of two monolayer thick Fe/Ir(111) films. Magnetic skyrmions are currently a hot topic and a substantial community is interested in materials and properties. This paper is timely, and the finding that hyrdogen absorption allows stabilization of magnetic skyrmions in films that do not support skyrmion formation in absence of hydrogen (at least not at the available field strength) is interesting to the community. The STM experimental data looks convincing and the theoretical interpretation seems reasonable. The paper is written well and it will be interesting to the audience of nature communications, I recommend publication.
Review of "Inducing skyrmions in ultrathin Fe films by hydrogen exposure" by Dr. Hsu et al. submitted to Nature Communications.
The manuscript describes the influence of hydrogen exposure on the magnetic spin texture in an Fe double layer (Fe-DL) on an Ir(111) single crystal substrate. Using STM topography, they revealed two kinds of superstructures on the Fe-DL. One is termed as H1-Fe while the other is termed as H2-Fe. Using SP-STM technique, they found that H1-Fe is a spin spiral ground state at zero magnetic field, while H2-Fe is ferromagnetic. With applied magnetic field, they demonstrated a transition from the spin spiral state to a magnetic skyrmion state. Furthermore, they tried to reproduce their experimental results by simulations based on Heisenberg exchange interaction and DMI parameters determined from ab initio calculations. Their calculations suggested that the incorporation of H into different vertical positions and in different concentrations leads to the two different magnetic states. As far as I know, the result (hydrogen-induced skyrmion state on a surface) is novel and their experimental technique (SP-STM) is well established. Of course, the control of the microscopic magnetic interactions responsible for creating and stabilizing individual skyrmions in ultrathin films is an important topic for the future development of skyrmion-based devices. Basically, I think the manuscript is appropriate for publication on Nature Communications. I should point out, however, a concern to be solved before acceptance of the manuscript. In Figure 1b&c, a lot of defects appear to be introduced in the hydrogenated Fe double layer. Different kinds of defects are observed for H1-Fe (protrusions) and H2-Fe (pits). Since the magnetic spin texture of a thin film specimen is likely to be affected by such defects, the authors must describe the nature of the two kinds of defects and discuss their influence clearly. In addition, from a viewpoint of practical applications, it would be helpful for readers to comment on the stability of the hydrogen-induced magnetic state. Furthermore, it appears strange that Discussion contains only concluding sentences. So, I suggest to the authors including the two topics -the nature and influence of defects and the stability of the hydrogenated magnetic state -in Discussion before concluding sentences.

End of Review
Reviewer #1 (Remarks to the Author): We thank the Reviewers for their careful evaluation of our manuscript as well as for the useful comments. Our replies are typeset in blue. The changes in the manuscript are summarized at the end of the reply.  (111)  Communications. I should point out, however, a concern to be solved before acceptance of the manuscript." Reply: We would like to thank the Reviewer for identifying the essence of our work. We have replied to all comments in a point-by-point fashion below. Figure 1b&c, a lot of defects appear to be introduced in the hydrogenated Fe double layer. Different kinds of defects are observed for H1-Fe (protrusions) and H2-Fe (pits). Since the magnetic spin texture of a thin film specimen is likely to be affected by such defects, the authors must describe the nature of the two kinds of defects and discuss their influence clearly." "In addition, from a viewpoint of practical applications, it would be helpful for readers to comment on the stability of the hydrogen-induced magnetic state. Furthermore, it appears strange that Discussion contains only concluding sentences. So, I suggest to the authors including the two topics -the nature and influence of defects and the stability of the hydrogenated magnetic state -in Discussion before concluding sentences." Reply: We thank the Reviewer for this comment. Following his/her suggestion, we expanded the Discussion section by the paragraphs about the role of defects and the stability of the magnetic structure. We also expanded Supplementary Note 2 with the identification of different types of defects, since we found that this is connected to repeated dosages of hydrogen. See items 1, 4, 5, 10, 11, 12, 13 in the list of changes. However, we found that the defects do not significantly influence the period of spin spirals and the size of skyrmions, or the fact that the latter can only be observed in the H1-Fe structure among the systems studied in the manuscript, in the 3-5 T field range. This indicates that the magnetic phases are robust against local modifications, and can only be influenced significantly together with the atomic structure by forming different hydrogen-induced phases.

Comment (3): "(1)
In Supplementary Note 2: Page 3, Line 11: " temperature und subsequent" should be "temperature and subsequent"." Reply: We thank the Reviewer for pointing out this typo, which we have corrected in the resubmitted version, see item 8 in the list of changes.
The reply and corresponding changes to the questions of Reviewer #2: Reply: We thank the Reviewer for his/her conclusion that "This is an overall beautiful and consistent study that demonstrates the impact of hydrogenation on the presence of skyrmions." However, we feel that the Reviewer did not fully appreciate the importance of our work, and we would like to clarify the significant advancements as described in our manuscript by emphasizing the following points: 1. What we report on is not a "skyrmion radius enhancement" as judged by Reviewer 2. As it is clearly stated in the abstract and repeated on page 2 in the last paragraph of the Introduction, "the hydrogenated system supports the formation of skyrmions in external magnetic fields, while the pristine Fe double layer does not", which indicates a fundamental difference between the two systems. The ground state of the Fe-DL system is a spin spiral state with approximately 1.2 nm period, which is not influenced by an external magnetic field up to 9 T. The H1-Fe structure displays a spin spiral ground state with 3.5 nm period, and this is the only system studied in the manuscript where the formation of isolated skyrmions is reported. Concerning the sizes of magnetic skyrmions, as shown in Fig. 3, the average diameter of the skyrmions is around 3 nm at 3 T, which decreases upon increasing the applied field. On the other hand, the H2-Fe structure supports the ferromagnetic state. These results clearly demonstrate that hydrogenation of the pristine Fe-DL leads not only to a modulation of the magnetic structure, but magnetic phase transitions as well.
To further emphasize this point in the manuscript, we expanded the Discussion section by explaining that the Fe-DL stays in the spin spiral state, while the H1-Fe structure transforms into the skyrmion state under external field. We also added a sentence about the field dependence of the Fe-DL based on spin dynamics simulations at the end of the theoretical section. See items 3, 5 in the list of changes. Figure 4 demonstrates that the effects range from a slight decrease in the magnetic period to the formation of a ferromagnetic state. From the experimental side, two structures with different magnetic properties are reported. Taking all of these facts into account, we find it very difficult to judge what would be the "expected" behaviour of the system when exposed to hydrogen in the absence of a combined rigorous experimental and theoretical investigation.

In our opinion, the fact that "distortions are not surprising" does not mean that the presented hydrogen effect is trivial. It is known that the microscopic magnetic interactions delicately depend on the electronic structure of the material, leading to a wide variety of types of magnetic orderings observable in nature.
Although it is expected that hydrogen adsorption influences the electronic structure, its effect on the magnetic ordering seems to be very difficult to predict. As an example, Supplementary Fig. 5b demonstrates that tuning the hydrogen concentration while keeping the interlayer distances fixed also leads to a modulation of the spin spiral period, also mentioned in the 4 th paragraph on page 4 of the main text. However, since interatomic hybridization processes directly depend on the distances between the atoms, despite these complications it is possible to observe a strong correlation between the period of the magnetic ground state and the interlayer distances as discussed in the manuscript, which leads to a relatively simple but qualitatively correct explanation of the effects. Although In order to further clarify the role of the electronic structure in the formation of the magnetic ground state, of which the interlayer distances are only a single but certainly important aspect, we extended the theoretical discussion in the second paragraph on page 2, see item 2 in the list of changes. We also comment on the importance of hybridization effects in the last paragraph of the main text, already present in the previous version.

We think that another reason why our work "advances the understanding of chiral magnetism" is the role played by frustrated exchange interactions in the formation of non-collinear ground states, see Table 1. A large number of previous investigations on skyrmions, including the overwhelming majority of the experimental studies, is based on a micromagnetic description only containing ferromagnetic exchange interaction, Dzyaloshinsky-Moriya interaction and magnetocrystalline anisotropy, where only the Dzyaloshinsky-Moriya interaction may favour the formation of skyrmions. This way, our work once again establishes a connection between recent theoretical investigations discussing frustrated exchange interactions (see e.g., reference [11] and the references therein) and experimental observations.
To emphasize this point, we expanded the second paragraph on page 6 of the Discussion section, see item 5 in the list of changes.

Comment (2): "(ii) I am not convinced the present work could open new avenues for the engineering of skyrmion lattices. Upon H exposure, both H1-Fe and H2-Fe phases appear in a non-controlled manner, which renders the skyrmion engineering using this technique quite
questionable. In addition, the presence of skyrmion lattice seems to be very sensitive to structural distortions, omnipresent in ultrathin films, which makes it hardly adapted to engineering purposes." Reply: We cannot agree that the hydrogenated phases appear in a completely "noncontrolled manner". This question is discussed in detail in Supplementary Note 2. In a series of systematic growth studies, it is demonstrated in Supplementary Fig. 2

that repeated H adsorption and post-annealing cycles lead to the appearance of well-defined, extended and connected H1-Fe areas where skyrmions can be observed, because the H2-Fe areas coalesce.
It is true that the appearance of magnetic skyrmions is linked to the structural changes induced by the hydrogen adsorption; however, Fig. 3 clearly demonstrates the shape and size of skyrmions inside the H1-Fe structure is completely well-defined, despite the presence of local defects. Furthermore, we do not agree that structural distortions and defects themselves may prohibit the engineering of non-collinear spin structures, given the significant number of observations of magnetic skyrmions even in amorphous systems such as in references [8,13,14,15]. Therefore, we believe that hydrogen-induced structural distortions are not a limiting factor, but an alternative route for developing and stabilizing individual magnetic skyrmions in ultrathin magnetic films.
We expanded the Discussion section on pages 5-6 to emphasize the stability of the observed non-collinear spin structures despite the presence of defects and structural distortions. In Supplementary Note 2 (page 3), we added a sentence explaining how the H2-Fe and consequently the H1-Fe areas coalesce during repeated annealing cycles. See items 5, 9 in the list of changes.

Comment (3): "In summary, this is a nice study performed in a rigorous manner; I also acknowledge that this is probably the first observation of skyrmion modification under H exposure. That being said, I do not see how disruptive this observation is and how is can open "novel avenues" for skyrmion engineering. In my view, this manuscript deserves publication in a specialized journal, but not in a broad-audience journal such as Nature Communication."
Reply: As already discussed above, we would like to emphasize that the key outcome of our studies is not simply a "skyrmion modification under H exposure" as commented by the Reviewer. On the contrary, the major finding of our work is that adsorbing a low-Z element such as atomic hydrogen on an ultrathin magnetic film can strongly modify non-collinear magnetic order, even resulting in the emergence of skyrmions which are absent in the pristine system without hydrogen adsorption. We thus expect that this effect should stimulate further investigations well beyond the specific material choice of H, Fe and Ir as discussed in the present manuscript. We believe that the presentation of the results is appropriate for a broader readership, while the content is in agreement with the mission statement "Papers published by the journal represent important advances of significance to specialists within each field", thereby meeting the publication criteria of Nature Communications. 13. Supplementary Fig. 4, page 6, the white triangles were added to mark the hydrogen vacancies in panel a and solid red circles were added to infer possible missing H atoms from the proposed structural models in panels b and c. The caption has been revised accordingly.

Reviewers' comments:
Reviewer (1) In connection with reviewer 2's comment (ii), I suggest to the authors that they add notes in the main text on their prospect to create pure H1-Fe phase without coexistence of H2-Fe phase to engineer skyrmion states in a fully controlled manner.
(2) In connection with the nature of the surface defects (hydrogen vacancy or extra hydrogen atoms on the surface), I also suggest to the authors that they prepare Figure  4c to evaluate the influence of the H concentration on the surface (above Fe2). The figure should be similar to Figure 4b evaluating the influence of the H concentration on the interlayer (between Fe1 and Fe2).

Reviewer #2 (Remarks to the Author):
The authors have replied carefully and convincingly to my comments. I particularly appreciate that they took the time to develop their arguments and added appropriate discussion to the manuscript. I indeed missed the point of H-driven spin-spiral-to-skyrmion transition, and I recognize the stability of the process and the in-depth analysis provided in the paper. I therefore support publication in Nature Communications as it is.

Reviewer #3 (Remarks to the Author):
As I indicated in my earlier report, I believe that this paper will be interesting to the audience of nature communications. In this revised version the authors addressed all comments provided by all reviewers, I recommend publication of this manuscript.
We thank the Reviewers for their careful evaluation of our manuscript as well as for the positive recommendation. Our replies are typeset in blue. The changes in the manuscript are highlighted in blue and also summarized at the end of this reply. Reply: We thank the Reviewer for assessing the essence of our work and appreciate his/her positive comments.