The Nanoscale World

Fractal analysis of islands in ultrathin organic films

rated by 0 users
Answered (Not Verified) This post has 0 verified answers | 2 Replies | 3 Followers

Top 75 Contributor
14 Posts
Points 158
Bruker Employee
DigestMigrator posted on Mon, Sep 13 2010 2:32 PM

Dear SPM users,

I am searching for an image analysis software that allows to determine
the fractal dimension (df) of islands of the first few layers in organic
thin film growth.
The islands are two dimensional objects (the z dimension is fixed by the
length of the molecule) and exhibit round to dendritic shapes.

So far I have used the coastline fractal dimension analysis exploiting
the relationship" perimeter= c* area^ (df/2)" and the gwyddion fractal
dimension analysis (cube counting value-1), getting contradictory
results aswell between the two measures and already published experiments.

In literature people use a box-counting method described in Vicsek's
textbook "Fractal growth phenomena" to determine the Hausdorff fractal
dimension but do not give any information of how they extract this
measure from SPM images.

If anybody has software recommendations or other useful tips, please
post it or contact me.

Thank you in advance and good luck for your research!

Andreas Straub

--
Andreas Straub,PhD student in Physics
CNR-ISMN Bologna
Nanotechnology of Multifunctional Materials Research Division

Via Gobetti 101
I-40129 Bologna
Italy

e-mail: a.straub@bo.ismn.cnr.it
web page: www.bo.ismn.cnr.it
tel: +39 051-639-8526
fax: +39 051-639-8540

  • | Post Points: 14

All Replies

Top 10 Contributor
75 Posts
Points 3,652
Suggested by Ben Ohler

Hi Andreas,

I did use the box-counting method to analyze some lipid monolayer domain images for my dissertation. We only looked at a couple images and we didn't consider it a very critical analysis, so we didn't put any work into automating the analysis or otherwise making it easy. I forget the exact details, but we literally just drew different size grids on printouts of the AFM images and counted boxes manually. In our case the boundary was not so well defined either, which introduced more uncertainty.

Here's what I wrote about it in my dissertation:

Normal 0 false false false MicrosoftInternetExplorer4

        There exist several ways to estimate the fractal dimension, D , of an experimentally observed shape. We used the box counting technique described by Birdi (1993). This technique is applied by taking a grid of box length "l" and defining x = 1/l as the “precision” of the grid. The grid is placed over the shape being analyzed and the number of boxes, N , containing any portion of the border of the shape are counted. This is repeated for several grids of different box sizes. One finds that the number of non-empty boxes asymptotically approaches the dependence: N= K * x^D, for some constant K. The fractal dimension, D, can then be determined by regression of the line: log(N) = log(K) + D*log(x) . This technique was validated on a known fractal, the Koch snowflake, for which the fractal dimension can be calculated analytically to be 1.262 (Mandelbrot, 1983).

The Birdi reference is: Birdi, K. S. 1993. Fractals in chemistry, geochemistry, and biophysics : an introduction. Plenum, New York

Note the idea to check your analysis on a fractal of known fractal dimension. That might help resolve the discrepancy between the two methods you have already used.

I wish I had a better answer for you. Hopefully someone else will be able to direct you to a more automated analysis.

Regards,

-Ben

  • | Post Points: 10
Not Ranked
1 Posts
Points 10

Dear Andreas,

You can try WSxM software. You can download it for free at www.nanotec.es. There is an option called Flooding that allows you to detect the highest (islands) or lowest  (holes) points of an image. Then it allows to get different information about perimeters, areas, volumes... of the islands, including a fractal analysis, where you can get the fractal dimension.

 

Greetings,

 

Pablo Ares.

 

  • | Post Points: 10
Page 1 of 1 (3 items) | RSS
Copyright (c) 2011 Bruker Instruments