Difference between revisions of "Main Page"

From ResearchWiki
(→‎simulation procedure)
(→‎Recent Results)
Line 69: Line 69:
  
 
====Recent Results====
 
====Recent Results====
 +
* [[Near Commensurate Phase]]
 
* [[Bulk Sims]]
 
* [[Bulk Sims]]
 
* [[Escaping Energy Minimum]]
 
* [[Escaping Energy Minimum]]

Revision as of 16:30, 25 April 2010

This is a wiki for collaborating on research projects.

Current Projects

Molecular Dynamics of Block-copolymers (MDBCP)

Scaling up BCP Runs on DESE computers

Division of Electrical and Software Engineering (DESE) in the Center for Devices and Radiological Health at the FDA is providing high performance computer resources and support for large-scale molecular dynamics simulations. Initial efforts will be on scaling up the moderate size runs (<100,000 beads for ~10 millions iterations) to larger size runs, meaning systems upwards of 1,000,000 particles. Simulations of films with an air interface, although scientifically of great interest, are inherently processor-inefficient and then thus difficult to maintain running without complaints from IT staff. Therefore, we will be working with the less scientifically important case of a bulk polymer material, ie one with no air interface. Two initial system sizes will be run for fairly short periods of time: Warewulf Runs:

Making HiDef Movies of Simulations

Goal: 720p movies of simulation trajectories. Procedure:

  • set povray width/height flags to 1280x720
  • collect consecutive snapshot images w/ name format as follows: image%04d.png
  • using ffmpeg ( http://ffmpeg.org/ ), create a playable movie file as follows:
 ffmpeg -b 500000 -i %04d.png  -r 12 -vcodec msmpeg4v2 -s 800x600 nice.asf

The flags may be optimized to improve movie quality in the future. The msmpeg4v2 codec was used to create .asf formatted movies, as surprisingly, this was the only combination that would provide playable movies for powerpoint on a windows machine, specifically my FDA laptop.

Manuscript Drafts

Figures

Working on perl script to call 'configuration.cpp' and perform calculations.

To be run in directory containing dump data, param_set.txt and config_dumpfiles.in. The script creates a subdirectory, EXT-FIFO, if it does not exist. The directory contains all necessary input files to do an extension. The dump output is written at a very high rate to a named pipe, which is then fed into the configuration.cpp program. Thus, no dumpfile accumulates and the lammps output only exists long enough to do configuration.cpp calculations on it.

usage: mkext-fifo-2-24-10.pl [ jobs ] where [ jobs ] are the command line arguments taken by configuration.cpp (a.out).

/home/cforrey/CODE/datedCode/mkext-fifo-2-24-10.pl

simulation procedure
polymer characterization
morphology vs substrate/temp
lamellar characterization
Next paper:


Recent Results

Progress

  • To be done:
    • Composition and surface energy phase diagram (at a particular T, like 1.4?)
  • Resources:

Plan for Paper

  1. Main points of paper:
    1. We have a valid strategy for simulating BCP thin films.
    2. We can reproduce many expected results for BCP thin films (islands and/or holes, L0 vs. T, etc.). Surface and substrate wetting play a crucial role, as already amply demonstrated in the experimental and theoretical literature.
    3. We have identified internal film pressures/tensions/stresses as having a significant role in controlling film behavior (morphology orientation, L0, islands and holes... and also defect density, ordering kinetics, etc.):
      1. A "happy" (stable, low-energy) film has a low internal tension. The formation of an island or hole is one place to "pay the energy penalty" of having incommensurate thickness. Internal film stresses is another.
        1. Currently speculative: The tradeoff between forming islands-and-holes versus building up internal stress depends on the bending energy (chain stiffness).
        2. Currently speculative: Tension builds up to a critical level before being released ("pop") with the formation of an island or hole (e.g. as T is changed and thus L0 changes).
      2. Currently speculative: The vertical state is low-energy than horizontal on neutral surfaces, because a tension mismatch between the two blocks (slightly different cohesive energy) creates an energy penalty in horizontal configuration, which can be relieved (slightly) in the vertical configuration via surface "puckering". (We have experimental data showing the puckering.)
        1. Currently speculative: A perfectly vertical state (rather than vertical with tilt) should be preferred because this can minimize chain perturbation at the film surface (and also minimizes surface area?). Thus a perfectly vertical state should appear when the bending energy is increased (and the surface tensions increase?).
  2. Figures: (tentative, of course)
    1. A graph of T vs. time to explain our general simulation protocol. Sub-panels show generic image of a spin-cast starting state, and a final state (with an expected morphology; e.g. commensurate horizontal on a strongly wetting substrate). This justifies our methodology.
    2. A few "boring" results (e.g. morphology vs. composition?), mostly to justify that our protocol is "getting the right answer".
    3. Phase diagram (images) of gamma and T for LAM. This will show many expected trends (horizontal vs. vertical), and we will point out some interesting features (islands and holes, tilt angles, etc.)
    4. L0 vs. T (and corresponding pressure vs. T?); trend is in expected direction (L0 decreases as T increases for the usual reasons) but exact scaling is informative.
    5. Phase diagram of tensions for the gamma vs. T images. We can show the "residual final tension" or the "tension in z direction" or the "tension along LAM normal" or whatever best makes our point: that internal stress and tension mismatch between the blocks drives behavior.
    6. Phase diagram (images) of composition and gamma for T=1.4 (or something). This will show the interplay between composition and film-confinement effects. There are plenty of papers showing, e.g. conversion from cylinder to sphere morphology in thin films. We can reproduce this behavior and (hopefully) show the tension/stress patterns that exist in the film (and, we argue, drive morphology changing...)
    7. And some more...

Future Directions/Projects

  • Explore role of substrate topography (roughness, channels) on controlling orientation.
  • Explore role of chemical patterning on BCP order. In particular look at BCP pattern fidelity (e.g. how are errors or LER in the pattern translated into the BCP LER), defect tolerance, persistence of templating in the z-direction, and so on. In other words, all the effects of note for using BCP chemical patterning for litho. Be aware of significant work already done: DOI: 10.1021/ma702514v, DOI: 10.1103/PhysRevLett.102.197801, DOI: 10.1039/b902283j and others by Nealey.
  • Combine MD simulations with scattering modeling, to compare with experimental data. (Possibly fit experimental data by adding an energetic bias in LAMMPS based on fit quality?)

Getting started with a Wiki

Consult the User's Guide for information on using the wiki software.

Some useful links:

Retrieved from "http://yager-research.ca/wiki/index.php?title=Main_Page&oldid=1297"