DFT project

One of the goals of the course is to become familiar with modern computational software that solve difficult problems in solid state physics. You will use the software package OpenMX to calculate the band structure of a semiconductor, metal, or insulator. You can complement your own calculations by downloading band structures from an open source project called aflow. Decide on an interesting system (talk to the instructor and others about what is feasible). You always start by reproducing the work of others, and then go on to explore by yourself. You will calculate (at minimum for a passing grade) the dispersion relation and the total and partial densities of states of an interesting material. You will also calculate at least one other property – electron density, optical properties, spin polarization etc., for a B or higher grade. The difficulty of the project will be taken into account in assigning the grade.

Poster assignment

Summarize your results as a poster to be shown during class (date will be posted on the main calendar of the course). This is a conference of your peers, and may also be attended by interested parties such as former and prospective PH575 students, and faculty. OSU Student Media Services, located in the Valley Library, print posters for free if it is a class assignment. A typical poster size is 3 feet wide, 3.5 - 4 feet tall. Text and diagrams should explain your calculations, and provide other information about your material. During the poster session, you'll stay with your poster for 50% of the time, and circulate around other posters for 50% of the time.

Paper assignment

After the poster session, you will write a paper that describes your findings. The feedback you get from the poster should help you write a better paper.

General things I look for when grading the paper:

  • the paper distills the knowledge you have gained about electronic structure,
  • the paper would be helpful to an audience of incoming PH575 student (the instructor will put himself in the shoes of typical PH575 students when deciding if the paper is well written).
  • the paper shows evidence of your scientific curiosity, and your ability to explain essential concepts.

Specific points I look for:

  • Good discussion of what your calculations show.
  • Figures are easy to read (font size is big enough, axes are labelled, captions are sufficient to interpret the figure)
  • The paper includes real-world connections such as applications of the material, or the reason for academic interest in the material.
  • The paper cites a number of appropriate references from academic literature.

Typical papers in the past have been about 8 pages including figures and references. Please use 12 point Times or 11 point Arial. Please use 1.5 x line spacing and 1.25” margins. These are standard choices for an academic manuscript and makes grading easier. Color printing can make a huge difference for interpreting crystal structures, partial DoS etc.

If you have questions, please ask! I will almost always email the whole class with my answer, or post the answers here under FAQ.

Paper/poster FAQ


You are writing a scientific article for an audience of your peers, e.g. incoming PH575 students. You and your peers are highly educated, technically savvy people, but you do not know everything about every material that has ever been discovered. You are educating your peers about a particular material that is of scientific interest or technological importance and you want to convey what is interesting or important about that material from the point of view of a scientist. You are all capable of reading and understanding articles in any of the journals on solid state physics or materials.

How to go beyond the basics?

Here are some ideas that might stimulate your creativity

  • explore another option within OpenMX
  • pose a question that could be explored using this new tool
  • study a family of materials
  • go beyond the standard “default clicks” in OpenMX to explore something you found interesting
  • choose a more challenging material (not always recommended because this is often an all-or-nothing proposition).

What materials have people chosen before?

Here is non-comprehensive list of materials that people have chosen in the past.

Do students from previous years have useful advice for me?

Yes! Here is the advice from previous years.

What is a CIF file

The Crystallographic Information File (CIF) is the standard for crystallographic data exchange prescribed by the International Union of Crystallography . It was described in a paper in Acta Cryst. (1991). A47, 655–685 by S. R. Hall, F. H. Allen and I. D. Brown.

Here is some of the text from a CIF file describing anatase TiO2:

Ti1 Ti+4 0 0 0 1 0.0
O1 O-2 0.2821 0.2821 0 1 0.0"

Translating the above code into english: “There is a Ti atom at (x,y,z) = (0,0,0) and an O atom at (0.2821a, 0.2821b, 0), where a and b are unit cell dimensions.

It is helpful to drag a CIF file into the OpenMX Viewer (a browser based application) to see the crystal structure. You may wish to export the crystal structure as an .xyz file so that you have the actual atomic coordinates.

Can I try grading someone's paper from a previous year?

Yes. Ethan has copies of papers from previous years. However, you cannot do your paper on the same material that you read about in someone else's paper.

Collaboration or no?

Teach each other about how use the OpenMX basics. Share tips on how to use the software, especially on how to get to the DoS and the E(k) plots. This just like being in a lab where you show someone how to use a piece of equipment. This kind of collaboration is healthy, but don't do the work for the other person. In a lab, you show someone a tool and he or she decides what experiment to do with it.

Talk to others to help point you to information, but don't let the other person do all the legwork for you, and likewise, don't do the legwork for others.

Please do not perform the other person's calculation. Help him or her with specific calculational problems, but don't define the goals of the assignment for the other person.

Don't work together to construct your posters - the temptation to produce the same format is very strong. This is your poster. But critique someone's poster and offer advice if asked. Tell each other where to get poster materials or how to get access to good printing etc.

Length of paper?

Depends somewhat on the choice of topic and how many figures you include. Typical papers in the past have been about 8 pages including figures and references.

Color printing?

You should use color - it makes a huge difference in crystal structures, band structures, DoS plots etc.

Why a poster as well as a paper?

It's so that you can share your work with the other students and learn something from them, too. Science is about communicating ideas and results, so you need to practice that. So come up with something interesting to tell your peers.


Don't write a technical manual (for example, “go to this menu, and click this button”).

Beware of the “true statements” trap. Just because you say true things does not mean a paper is well constructed or a valuable learning tool. You may have gleaned a piece of information from a well-respected paper in a reputable journal, but quoting that authority without integrating the information into the flow and design of your paper is not helpful.

Don’t think of the PH575 instructor as your audience. The instructor may give the paper to a colleague in chemistry or electrical engineering, or a grad student, to see whether they can make sense of the writing.

How do I know my band structure is right?

Look in the literature - someone else is bound to have calculated it. Results will vary a bit with different methods used. Remember to reference others' work in your report.

My material has 80 atoms per conventional unit cell - is that OK?

80 atoms is too big, however, check whether the primitive unit cell has fewer atoms. By trying to do computations with too many atoms you'll use all the memory and tie up the system so no one can use it. Not good!

Can I study doping, (eg B-doped Si) or solid solutions like ZnS_(1-y)Se_y?

No, it's best to stay away from these because you will need a much larger unit cell. This is because if you replace say 1% of Si with P, you will need 100 atoms in your system (99 Si and 1 P) to define the basic unit. Refer to the question above.

Can I study graphene?

During class time we will study several properties of graphene using analytical theory. Your poster will be boring if you reproduce the same things we've already done in class. However, you might have a good idea how you could use OpenMX to answer questions that go beyond the graphene calculations we do in class.

Can I study conductivity as a function of temperature?

No. All calculations are inherently T = 0 calculations. There are no excitations of the system included.