[RMCProfile-users] 1. about the loops. 2. Maximums and minimums of the distance window constraints

Tue May 23 19:56:34 BST 2017

Thanks for Helen's kind reply, I am doing synchrotron X-ray PDF for the RMC
simulations. The X-ray scattering has a disadvantage about the lighter atoms
comparing with neutron, so I also got many problems when I do RMC to
synchrotron X-ray PDF.
Some question. 

1.  I got a very good fitting with just fitting F(Q) and G(r), but why the
RMC fitting go to terrible noise again after I add Bragg data, with an
increased weighting of the F(Q)? Below is the weight setting for my job:

BRAGG :: 
  > WEIGHT ::   0.01
XRAY_RECIPROCAL_SPACE_DATA :: 
  > DATA_TYPE :: F(Q)
  > WEIGHT :: 0.5

2. For multiple RMC boxes refinement, can I do the the multiple RMC
refinement based on the structure which has a good RMC refinement already?

Bests
Bo

>-----Original Message-----
>From: rmcprofile-users [mailto:rmcprofile-users-bounces at rmcprofile.org] On
>Behalf Of helen.playford at stfc.ac.uk
>Sent: Monday, May 22, 2017 11:02 AM
>To: rmcprofile-users at rmcprofile.org
>Subject: Re: [RMCProfile-users] 1. about the loops. 2. Maximums and
minimums
>of the distance window constraints
>
>Hi Bo,
>
>1. You could also do this by just not including the distance window
constraint
>block at all! But anyway, there's no reason that it shouldn't be a reliable
result,
>but it does depend on whether the partial functions (i.e. things like bond
lengths
>and distributions) make sense for the system you're looking at. This is
where you
>have to use your judgement to determine if a result makes sense.
>
>2. An extreme lower limit on the number of moves generated is that it must
be
>greater than the number of atoms in the box! But again there isn't a rule
for this.
>If the refinement has converged (i.e. the chi^2 is no longer going down)
then
>you've probably done "enough" moves. Usually this isn't a problem because
>modern computers generate so many moves so quickly. In terms of running
time,
>depending on my system I've needed everything from 6 to 72 hours.
>
>Best wishes,
>Helen
>
>-----Original Message-----
>From: rmcprofile-users [mailto:rmcprofile-users-bounces at rmcprofile.org] On
>Behalf Of youqin5912
>Sent: 22 May 2017 09:53
>To: rmcprofile-users at rmcprofile.org
>Subject: Re: [RMCProfile-users] 1. about the loops. 2. Maximums and
minimums
>of the distance window constraints
>
>Thanks very much for your kind reply! I still have two questions about the
RMC.
>1. For the second question, I try to set 0 to all the maximums and minimums
of
>the distance window constraints, which means a full relax fitting RMC. I
still get a
>good RMC fitting result, is that a reliable result?
>2. How many moves should be generated for each RMC box at least?
>
>Thank you very much!
>
>Bo Jiang
>Postdoc
>NTNU, Department of Materials Science and Engineering Cell phone: (+47) 94
>24 98 88
>
>>-----Original Message-----
>>From: rmcprofile-users [mailto:rmcprofile-users-bounces at rmcprofile.org]
>>On Behalf Of helen.playford at stfc.ac.uk
>>Sent: Monday, May 22, 2017 10:22 AM
>>To: rmcprofile-users at rmcprofile.org
>>Subject: Re: [RMCProfile-users] 1. about the loops. 2. Maximums and
>minimums
>>of the distance window constraints
>>
>>Hi Bo,
>>
>>Thanks for your email.
>>
>>1. Because there are a finite number of atoms in an RMC box, some of
>>the
>partial
>>pair correlation functions may have only a very small weighting with
>>regard
>to the
>>total PDF. If those small correlations are of interest then it is
>>necessary
>to have
>>multiple RMC boxes that have been refined against the same data which
>>can basically be added together to get partial functions with better
>statistical quality.
>>Furthermore, bond angle distributions calculated from a single box are
>likely to be
>>very noisy, so the same technique of adding together multiple boxes is
>>used
>to
>>improve this. As a rule of thumb if you want to get good partial
>correlation
>>functions then something like 20 - 50 boxes will probably work, but for
>bond
>>angles you will need closer to 100. This is obviously system dependent
>>as
>well.
>>
>>2. Usually you'll base the distance window values on where the peaks in
>>the
>PDF
>>go to the baseline, and you'll choose the edges of the peak. If you're
>using the
>>distance window constraint on overlapping peaks then you will have to
>>use
>your
>>judgement to estimate a "width" around the average positions. There's
>>no
>one
>>rule that will work for everything - it might need a bit of trial and
>error!
>>
>>I hope this is helpful!
>>
>>Best wishes,
>>Helen
>>
>>-----Original Message-----
>>From: rmcprofile-users [mailto:rmcprofile-users-bounces at rmcprofile.org]
>>On Behalf Of Bo Jiang
>>Sent: 22 May 2017 08:48
>>To: rmcprofile-users at rmcprofile.org
>>Subject: [RMCProfile-users] 1. about the loops. 2. Maximums and
>>minimums of the distance window constraints
>>
>>Hi everyone,
>>        I am working with PDF data collected from X-ray synchrotron
>>source
>from
>>ESRF, and I have got the final RMC model. I have two questions
>>
>>1.      Why the author said that 'the modelling was performed 120 times to
>>improve statistics' in Adv. Funct. Mater. 2013, 23, 185-190. So I
>>should do
>more
>>than hundreds of times to get the average structure?
>>
>>2.      How to decide the Maximums and minimums of the distance window
>>constraints? Can I  just plus or minus 0.5A or 0.6A base on the average
>structure
>>atoms distance?
>>
>>Thank you very much!
>>
>>
>>Bo Jiang
>>Postdoc
>>NTNU, Department of Materials Science and Engineering Cell phone: (+47)
>>94
>24
>>98 88
>>
>>
>>-------------- next part -------------- An HTML attachment was
>>scrubbed...
>>URL: <http://lists.rmcprofile.org/pipermail/rmcprofile-
>>users/attachments/20170522/0f9bb1cb/attachment.html>
>>_______________________________________________
>>rmcprofile-users mailing list
>>rmcprofile-users at rmcprofile.org
>>http://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users
>>
>>_______________________________________________
>>rmcprofile-users mailing list
>>rmcprofile-users at rmcprofile.org
>>http://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users
>
>_______________________________________________
>rmcprofile-users mailing list
>rmcprofile-users at rmcprofile.org
>http://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users
>
>_______________________________________________
>rmcprofile-users mailing list
>rmcprofile-users at rmcprofile.org
>http://lists.rmcprofile.org/mailman/listinfo/rmcprofile-users