Two questions from a returning user

[Tony Bell]

Hello all

Last year I got a lot of help from this group in analysing PDF data
using PDFgui, I was able to get some reasonable results out. Now I
have finally got round to getting these results ready for publication,
we have a draft paper nearly ready to be submitted, just a few things
to get tidied up with the results.

One of the samples that I analysed was an iron oxide, I was able to
refine atomic coordinates for 3 Fe and 4 O sites. However, for some
reason I only get esd errors on the lattice parameters, temperature
factors and the Fe coordinates, no errors are given on the O sites?
Has anyone seen something like this before in their data analyses?

The second question is that I would like to run PDFgui on my home
computer. I have been using a PC version on the work computer but at
home I have a Mac PowerBook G4 with Mac OS X version 10.4.11. Forgive
the simplicity of question but I really am a newbie at linux stuff. I
see it is possible to install a unix version of PDFgui for the Mac. Is
there anything special I would need to do to my Mac to install linux
or is it already there?

thanks tony....

[Pavol Juhas]

On Thu, Apr 16, 2009 at 08:37:24AM -0700, Tony Bell wrote:
...

> One of the samples that I analysed was an iron oxide, I was able to
> refine atomic coordinates for 3 Fe and 4 O sites. However, for some
> reason I only get esd errors on the lattice parameters, temperature
> factors and the Fe coordinates, no errors are given on the O sites?
> Has anyone seen something like this before in their data analyses?

Hi Tony, are you sure that parameters constraining the O sites are
not fixed in the last step? It is hard to tell why this happens
- please, check the "Refinement parameters" section in the Results
page, whether the oxygen related parameters have any ESDs.

> The second question is that I would like to run PDFgui on my home
> computer. I have been using a PC version on the work computer but at
> home I have a Mac PowerBook G4 with Mac OS X version 10.4.11. Forgive
> the simplicity of question but I really am a newbie at linux stuff. I
> see it is possible to install a unix version of PDFgui for the Mac. Is
> there anything special I would need to do to my Mac to install linux
> or is it already there?

Mac OS X is actually a Unix-like operating system so PDFgui should run
on it directly, there is no need to install Linux. However, there are
several external tools and libraries that need to be installed before
to get PDFgui working. Please, follow the instructions at

http://danse.us/trac/diffraction/wiki/MACBuildInstructions

Hope this helps. Cheers,

Pavol

--
Dr. Pavol Juhas
Applied Physics & Applied Mathematics
Columbia University
Room 200 Mudd Building, MC 4701
500 West 120th Street
New York, NY 10027

[Tony Bell]

Hello folks

Pavol replied to my question last week about PDFgui for Mac OS X

>
> > The second question is that I would like to run PDFgui on my home
> > computer. I have been using a PC version on the work computer but at
> > home I have a Mac PowerBook G4 with Mac OS X version 10.4.11. Forgive
> > the simplicity of question but I really am a newbie at linux stuff. I
> > see it is possible to install a unix version of PDFgui for the Mac. Is
> > there anything special I would need to do to my Mac to install linux
> > or is it already there?
>
> Mac OS X is actually a Unix-like operating system so PDFgui should run
> on it directly, there is no need to install Linux.  However, there are
> several external tools and libraries that need to be installed before
> to get PDFgui working.  Please, follow the instructions at
>
>    http://danse.us/trac/diffraction/wiki/MACBuildInstructions
>

I followed the URL which Pavol had given,

I downloaded the appropriate version of MacPorts for my version of OS
X

then I tried installing gsl and this is what happened

unknown000a95f2dc84:~ ynotlleb$ sudo port install gsl
Password:
Sorry, try again.
Password:
Portfile changed since last build; discarding previous state.
---> Configuring expat
Error: Target org.macports.configure returned: configure failure:
shell command " cd "/opt/local/var/macports/build/
_opt_local_var_macports_sources_rsync.macports.org_release_ports_textproc_expat/
work/expat-2.0.1" && ./configure --prefix=/opt/local --mandir=/opt/
local/share/man " returned error 77
Command output: checking build system type... powerpc-apple-
darwin8.11.0
checking host system type... powerpc-apple-darwin8.11.0
checking for gcc... /usr/bin/gcc-4.0
checking for C compiler default output file name... configure: error:
C compiler cannot create executables
See `config.log' for more details.

Error: The following dependencies failed to build: texinfo gettext
expat libiconv gperf ncurses ncursesw lzmautils
Error: Status 1 encountered during processing.

Any ideas? I did this in my top directory, is there a specific
directory where I need to be running this command?

thanks tony....

[Pavol Juhas]

On Mon, Apr 27, 2009 at 08:13:49AM -0700, Tony Bell wrote:
...

> Pavol replied to my question last week about PDFgui for Mac OS X

...

> > Mac OS X is actually a Unix-like operating system so PDFgui should run
> > on it directly, there is no need to install Linux.  However, there are
> > several external tools and libraries that need to be installed before
> > to get PDFgui working.  Please, follow the instructions at
> >
> >    http://danse.us/trac/diffraction/wiki/MACBuildInstructions
> >
>
> I followed the URL which Pavol had given,
>
> I downloaded the appropriate version of MacPorts for my version of OS
> X
>
> then I tried installing gsl and this is what happened
>
> unknown000a95f2dc84:~ ynotlleb$ sudo port install gsl

...

> C compiler cannot create executables

Hi Tony, a Google search for the error message suggests the problem
is in missing or outdated "Xcode tools". Please, make sure Xcode
tools are installed and try again. To my knowledge the port install
command should work from any directory. If you obtained Xcode tools
from the system DVD and still get the error, update to the latest version
from the Mac website. You will need to register with Apple
Development Central, but otherwise the download is free.

Good luck. Please, let me know if this helped and especially if it
is necessary to update the MACBuildInstructions wiki. Thank you,

[Tony Bell]

A question for all you diffpy experts out there.

I am writing a paper which compares the results of Rietveld and PDF refinements taken from the same data sets. Rietveld can give crystallite size information, it is possible to get crystallite/particle size information from the output from PDFgui?

thanks

tony....

-- 
Tony Bell - ynot...@mac.com
Appleton, Warrington, Cheshire.

http://tinyurl.com/38uybu

[chris farrow]

Hi Tony,

This depends on the size of the particles. If the range of the PDF is on the order of the particle size, you can use the spdiameter parameter to get an approximate particle size (which assumes the particle is a sphere). If the PDF range is much shorter than the particle diameter, you will probably get uncertain results, at best. Note that spdiameter is very correlated with qdamp, since they both attenuate the PDF. You'll need a good measurement of qdamp in order to get a meaningful for spdiameter.

Good luck!

Chris

Christopher L. Farrow
Associate Research Scientist
Applied Physics and Applied Mathematics Department
Columbia University
Office: 4265 Biomedical and Physical Sciences Building
       Michigan State University
       East Lansing, MI 48824-2320
Phone: (517)884-5677
Fax: (517)353-4500
email: clf...@columbia.edu
web: http://www.columbia.edu/~clf2121

[Katharine Page]

PDFgui is currently capable of applying a spherical shape factor for a
particular particle diameter to any phase in the refinement. The
parameter is called 'spdiamter' and values are given in Angstroms (It
is not applied when the value is given as zero.) This value can be
refined.

However, from the PDFgui manual:

"PDFgui is capable of modeling the effect of the finite nanoparticle
size using a spherical
shape factor. Relevant PDF parameter is spdiameter which is the
diameter of the nanoparticle. This parameter is highly correlated with
various other parameters one would like to refine, such as anisotropic
ADPs, scale factors, correlated motion parameters and so on.
The refinement procedure is therefore rather delicate and the
solutions are not as robust as
we are used to in cases of crystalline materials."

I have personally found that estimated crystalline correlation length
from Scherrer broadening of the diffraction pattern is similar to the
refined spdiameter parameter from PDFgui for particles smaller than
~10 nm. However, this is only true if 'reasonable' values are applied
for other variables. "Reasonable' can be quite subjective! I believe
one should use care in calibrating the Gaussian dampening due to the
instrument (qdamp in PDFfgui).

I personally feel it is preferable (for discrete particles) to use
complementary techniques to provide an idea of particle size and at
least initially set this in your PDF refinement. This will improve
the accuracy of any parameters which may be correlated (as per the
warning in the manual). When I refine spdiameter, I usually use a
large r range, and then set the value when I refine using lower r
ranges. For a larger r range, the 'qbraod' parameter should be well
know from calibration.

Others may have their own strategy, but at least this has worked well
for me in studying oxide and metal nanoparticles....

Cheers,
Kate

--
********************************************************************
Katharine Page
Phone: (505) 665 4310 kpa...@gmail.com
********************************************************************

Science does not know its debt to imagination.
~Ralph Waldo Emerson