User login

You are here

A strained film grown on a vicinal substrate: Steps bunch or not to bunch?

When a strained film is grown on a vicinal substrate, the steps advance like a train when the deposited atoms have sufficient mobility to reach the step edges. However, as the steps advance, the strain-induced force monopoles associated with the steps cause the steps to attract to each other (J. Tersoff, PRL 74, 4962, (1995)), resulting in a thermodynamic instability of the steps in the form of step bunching (J. Tersoff, et al., PRL 75, 2730 (1995)).

Recently, it was shown that this bunching instability can be suppressed with proper control of the growth kinetics, taking advantage of the standard step-edge barrier effect (commonly known as the Ehrlich-Schwoebel (ES) barrier effect), and a morphological phase diagram was constructed to obtain smooth films in the regime of persistent step-flow growth (Wei Hong et al., PRL  95, 095501 (2005); see also an earlier post by Wei at node/313).

A more recent twist on this topic was the report by Mina Yoon et al. that, even in the unstable bunching regime with the ES barrier effect already taken into account, the bunching instability won't be triggered unless the system reaches a critical film thickness; therefore, smooth films could still be obtained in an otherwise unstable regime, or effectively in a broader volume/area of the phase space defined by the growth parameters (M. Yoon et al., PRL 99, 055503 (2007)). It was further shown that one could exploit the well-defined scaling properties of the critical film thickness to extract intrinsic energetic parameters for a given system. 

Mina-PRL07.pdf269.42 KB


Dear Zhenyu: 

Thank you very much for posting this paper, which reminded me of the time you spent at Harvard. 

I learned a lot from you in the process of writing several papers with you.  Your idea of "information quantum" and the style of telling a story were particularly illuminating.  Would you be willing to post a note on how to write a PRL, in the way that you explained to us?   Even though few of us mechanicians will  ever write a PRL, many of us do feel that too many papers in our community just don't have a story to tell, or tell a story so badly. And many papers are just too long, too unfocussed and too boring. 

A short note from you would inject a breath of fresh air into the community.  What do you say?


Hi Zhigang,

What a flattering challenge!:)

I believe I learned quite some from you on how to write concisely and precisely through the preparation of the few joint papers over the years.

Anyway, I don't think I am really positioned to respond meaningfully to your challenge/request. But since PRL is more of a physics journal and I am still largely a physicist, I will share my two cents of thoughts based on my experience with the journal, more for the purpose of continuing the exchanges.

The notion that the story line of a PRL paper should be built around a single, fresh (or new), and soundly validated information quantum was first expressed to me by Prof. George Comsa, a towering figure in surface science (retired a few years ago from Julich Research Center of Germany), when we were working on a joint paper in his office back in summer 1995. He used the term "informon", and I believe that in his opinion a good PRL should try to avoid packaging too many less essential points that potentially could distract people from the central message of the informon. Over the years, I have found this notion quite instrumental when wrapping up papers intended for PRL, though whether a submission would eventually be successful also could depend on other factors (for example on highly resonant or very different perspectives of the authors and the referees on otherwise the same idea).

Different authors have their own preferred PRL styles. For people in my group, we more or less adopt the following style:

Paragraph 1: broad interest paragraph, telling the editors and referees why the paper should in principle draw the interest of a broad readership of PRL. This paragraph could of course be somewhat subjective, but a decent effort is needed here even to avoid an editorial rejection without in-depth review (nowdays PRL rejects a significant fraction of manuscripts without review based on factors such as lack of general interest and poor presentation).

Paragraph 2: setting-up paragraph. Ideally, here one narrows down from the broad area of paragraph 1 to a more specific subject in the field, and after reading this paragraph, one should walk away with the compelling impression that something has to be done on the subject.

Paragraph 3: flag-ship paragraph. This is the place to announce what has been done in this paper in response to the appeal in paragraph 2. The claims have to be significant enough without overstating, otherwise a referee might judge that the paper is not worthy for a PRL even if the remaining part of the paper is all soild.

Paragraphs 4 to N: Detailed results to soundly validate the claims made in paragraph 3.

Paragraph N+1: Concluding paragraph. If space permits, add a brief and concise conclusion to refresh on the central findings of the paper, as after diving into paragraphs 4 to N a referee/reader might have already been exhausted and forgotten what the paper is about:).

Admittedly, it is a bit risky for me to post such personal thoughts in an open forum, but if anyone who entered research only recently would find it at least slightly helpful when wrapping up his or her first PRL ms, the risk is worth taking.

I welcome critical comments on the above thoughts.



Zhigang Suo's picture

Dear Zhenyu: 

When you first explained this idea to me a few years ago, I was so impressed that I asked you to write it down in my notebook.  Ah, the problem with notebooks is that they are not searchable.

For some time, whenever students and I made a paper too complex, I would tell them how Zhenyu would write.  Now the notion of focussing on a single good idea sounds so familiar to my group that even my students start to lecture me when I slip back to my old way of packaging a lot of ideas in one paper. 

The students do so without using your name.  Consider the plagiarism the compliment of highest order.  Thank you for taking the risk to post the idea, setting it free from the notebook.  We can now search and forward. 

Yanfei Gao's picture

Zhenyu and Zhigang:

This is a very nice discussion. A simple idea delivered well is, however, difficult to achieve. Very often, I found out that it's hard to balance conciseness and detailed explanation. I guess the paper should be written in terms of a well targeted audience.

In fact, in contrast to Rui's comment, I prefer reading long papers than short papers, because long papers are usually easier to underand.


Yanfei Gao,

Rui Huang's picture

I agree with Zhenyu on writing a short paper around one new idea, and his style of writing will certainly be helpful for many authors pursuing publications in high impact-factor journals like PRL (including myself). On the other hand, I hope that Zhigang won't just stop writing more conventional, long papers for mechanics journals of whatever impact factors. Personally I have enjoyed reading many long papers by Zhigang and other mechanicians. It is not always as easy as reading a short paper, but it is a learning process, which I think is very important for young researchers like myself. The short papers often keep me up to date on new experimental observations or acclaimed theoretical developments. The latter often takes a companion long paper to fully convince the readers or simply leaves them suspicious about the details. As an author, writing a long paper forces me to do a more thorough job in both reviewing the literature and researching the new ideas, which are often rushed through in a short paper.  


Zhigang Suo's picture

Dear Rui: 

We'll never stop writing papers that put several ideas into a perspective.  I have just read a 3-page essay by Wlater Noll, the author (with Truesdell) of the influential text The Nonlinear Field Theories of Mechanics.  He made an interesting point that an individual may play three roles:  teacher, researcher and professor.  His characterization of professor is very close to what you are saying.

But I just feel that as a community we are writing too many long papers that synthesize old ideas, and too few papers that create new ideas.

Rui Huang's picture

Dear Zhigang,

Thanks for the pointer to Walter Noll's essay. I enjoyed reading it and agree with him on most of it. Clearly, the role of a university faculty member is perfected with a balance of teacher, researcher, and professor. However, the distinction between the roles of a researcher and a professor is not always obvious. Likewise, whether an idea is new or old often depends on the person who is judging it. As a relatively young faculty member, I have encountered many ideas that appeared to be new to me but turned out not later. It is painful sometimes, but we have to live with the fact that both physics and mechanics, in particular, have had a long history with many brilliant contributors. On the other hand, the "priority" is now so important in academics that we see many rushing into print prematurely. 


Yanfei Gao's picture


I agree with you on the point of information explosion. It appears that quite often a simple idea can be used to explain some "new" experiments, which may turn out to be done. Keeping up with these short experimental papers is certainly a good suggestion. But considering the time...


N. Sukumar's picture

Excellent comments and suggestions, Zhenyu. I think possibly some journals in mechanics that already are reputed (e.g., JMPS) can also encourage short papers that have `something noteworthy to say.'  The incentive c'd be a much shorter time-to-publication. In many papers, the parts that are new are only a fraction of the total and hence it might not be all bad to encourage short-and-focused papers that can only be fully discerned by the specialist working in the same field. 

Pradeep Sharma's picture

Zhenyu, this is a very interesting post. I regularly read PRL and find myself in agreement with your thoughts. In particular, it is worth emphasizing that a very large fraction of the papers indeed never get sent for review and are rejected at the editors desk. In one of the APS meetings, the PRL editor showed us the dramatic increase in submissions to PRL thus requiring such measures.

I do appreciate long mechanics style journal papers but I believe there is room for a letters type forum as well in the mechanics area. I agree with Sukumar that it would be nice if journals such as JMPS would open up a section that incorporates "letters". Journal of Applied Mechanics has a brief notes section however unfortunately, unlike the physics communities, such "brief notes" are not considered prestegious.

Thanks Zhenyu for the nice points about how to write a short paper. I have also forwarded the post to my students and asked them to read. I believe they will learn a lot. 

I think it is good idea to have a short and quick journal in the mechanics domain. Maybe we can move one step further, a virtual journal. I think now days many people just download papers from the website and print them out when needed. So a paperless journal can satisfy the needs and dramatically reduces the operational cost. Maybe iMechanics can take the lead to create something like “iMechanics Bulletin” or “iMechanics Letters.” Particularly there is already a broad body of readers who regularly visit the website.  

Wei Lu

Very nice discussion about journal papers. However as a phd student I think a scientific paper should be instructive enough to be easily understood by graduate students like me, so the supporting paragraph and related or old ideas might be cited referred to make the new ideas clear enough.



Subscribe to Comments for "A strained film grown on a vicinal substrate: Steps bunch or not to bunch?"

More comments


Subscribe to Syndicate