[sdpd] Re: Minerals with unknown structure (>300)

[Robin Shirley <R.Shirley...@surrey.ac.uk>]

(1)

> and indexing  in monoclinic from 6 diffraction lines and a figure of
> merit F(6) = 1 (0.102, 118), wow ! Robin, don't laugh ! 

I don't understand - what's going on here?  A figure of merit F(6) of 1
would appear to mean a totally random fit (actually worse than most
random fits!).

But if 118 calc lines were used with just 6 obs lines, then the cell
must presumably have been derived from some external source, so the FoM 
would then essentially be irrelevant, and only included because of some 
ICDD rule and custom that you have to report one.

Otherwise, with only 6 lines and 4 free metrical variables in monoclinic
to fit them with, one should easily be able to generate numerous cells
with figures of merit >100 (though all essentially meaningless).

(2)

> F(30)= 2 (0.037, 496) in the 22-0711 ICDD entry (re-wow !).

Your first example might have some reasonable underlying explanation, 
though clumsily expressed, but this one looks like sheer junk.

How would anyone in the 21st century even consider publishing a general
powder indexing with a FoM of 2?  And no wonder it's so low if 496 calc
lines had to be generated to engineer a fit to just 30 obs ones - de
Wolff must be turning in his grave.

> My conclusion is that increasing the ICDD database content with
> redundant bad data is continuing.

I'm afraid it does rather look that way.

Robin

-------------------------------------------

To:            sdpd...@yahoogroups.com
From:          alb...@cristal.org
Date:          Thu, 13 Sep 2001 13:37:36 -0000
Reply-to:      sdpd...@yahoogroups.com
Subject:       [sdpd] Re: Minerals with unknown structure (>300)

> A starting list of > 300
> minerals with apparently unknown structure is available at :
> http://www.cristal.org/iniref/sdpd-minerals.html

In order to avoid simultaneous efforts, that list
will be updated with some results.

First try : nepouite Ni3Si2O5(OH)4, 
ICDD entries : 15-0580, 20-0791, 25-0524
Well, if you see the search-match result available
at the above web page, you will see that very probably,
another mineral with a different name (pecoraite), but 
same chemical formula is possibly the same material.
Pecoraite corresponds to the ICDD recent entry 49-1859, data
from a publication in Clay Miner., 1995, and indexing  
in monoclinic from 6 diffraction lines and a figure of
merit F(6) = 1 (0.102, 118), wow ! Robin, don't laugh ! 

Anyway, a structure determination of nepouite/pecoraite
from powder diffraction data appears an impossible mission,
you may get the pattern at the above web site too.

My conclusion is that increasing the ICDD database content 
with redundant bad data is continuing.

The studied nepouite sample sample appears to contain a bit 
of willemseite, (Ni,Mg)3Si4O10(OH)2, apparently with unknown 
structure too (nothing in ICSD), and F(30)= 2 (0.037, 496)
in the 22-0711 ICDD entry (re-wow !).

Armel Le bail
http://www.cristal.org/course/



 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 




------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get your FREE credit report with a FREE CreditCheck
Monitoring Service trial
http://us.click.yahoo.com/MDsVHB/bQ8CAA/ySSFAA/UIYolB/TM
---------------------------------------------------------------------~->

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/