HI Fellows,
thanks for the responses. Of course, my question was not posed without
sinister motivation....
Given the strict definition
http://reference.iucr.org/dictionary/Isomorphous_crystals
as opposed to the context sensitivity pointed out by Gerard, I just needed a
paper trail to deflect
potential nitpicker's wrath and distribute the blame equally. Thank you for
playing.
The non-isomorphism (now strict) is, as mentioned often anisotropic, i.e.
here the unit cell expanded
not equally in all directions. The degree to with reflections become
independent is I think given by
the reciprocal space overlap function G, and of course the (absolute)
differences are most pronounced
at higher resolution, where the reflections become more or less independent.
I still decided to use the
same R-free set for both data sets, just to deflect potential smart
comments, although am not convinced
that it really matters as my resolution is 1.9 A for both, meaning that
most reflections will be independent.
So, given the strict meaning of isomorphism eo ipso, I think 'almost
isomorphous' is too strong for a 5% cell change,
and para fails on grounds of linguistic multitude... proxymorph has a nice
ring, but is
also linguistically nonparsimonious.
Given that one model is much better than the other, putromorphism or,
linguistically consistent, scatomorphism comes to mind....
Cheers, BR
------------------------------------------------------
Bernhard Rupp
Crystallographiae Vindicis Militum Ordo
Vista, CA 82084
http://www.hofkristallamt.org/
[log in to unmask]
+1 925 209 7429
+43 767 571 0536
------------------------------------------------------
Many plausible ideas vanish
at the presence of thought
------------------------------------------------------
First, I just need a paper trail in case some brilliant
-----Original Message-----
From: CCP4 bulletin board [mailto:[log in to unmask]] On Behalf Of Gerard
Bricogne
Sent: Dienstag, 31. Mai 2016 13:27
To: [log in to unmask]
Subject: Re: [ccp4bb] AW: [ccp4bb] Not isomorphous but similar is...?
Dear all,
It seems to me that the real question is not one of terminology but of
operational reality.
Operationally speaking, then, the fact that two crystals are
isomorphous means that one can meaningfully compare their structure factors,
or their amplitudes, in a reflection-wise manner, i.e.
F1(hkl) with F2(hkl), without having to involve any other reflections.
If this is at the limit of making sense but one somehow gets away with it
for certain purposes, then I guess one would talk about "poorly isomorphous"
crystals. If reflection-wise comparison breaks down, then a change of
lattice is required and one has a case of NCS, more specifically of distinct
crystal forms, requiring masking of the contents of the ASU in one crystal
form, rearrangement into an ASU for the second one, and Fourier
transformation. Writing this in the way it was done in the early literature
on NCS, this implements (up to fussy
details) a Shannon interpolation between the two transforms, which reflects
the fact that the structure factors or their amplitudes can no longer be
compared individually, hkl by hkl.
It therefore seems to me that this operational distinction is more
important than the search for just the right terminology. Once two crystals
are not sufficiently isomorphous to allow a meaningful reflection-wise
comparison of structure factors, then they are non-isomorphous, with "poorly
isomorphous" in the twilight zone where those paired comparisons are still
possible but only to a resolution significantly lower than that of the data.
The reckoning of the degree of isomorphism by cell parameters only
should not be based of percentages of cell lengths, by the way:
it is the absolute differences of lengths, and their magnitude as a fraction
of the resolution limit, that determine whether pairwise comparisons make
sense or not.
With best wishes,
Gerard.
-
On Tue, May 31, 2016 at 10:59:20AM +0000, [log in to unmask] wrote:
> Being an optimist, I would use "almost isomorphous"
>
> My 2 cts,
> Herman
>
> Von: CCP4 bulletin board [mailto:[log in to unmask]] Im Auftrag von
> Bernhard Rupp (Hofkristallrat a.D.)
> Gesendet: Dienstag, 31. Mai 2016 11:11
> An: [log in to unmask]
> Betreff: [ccp4bb] Not isomorphous but similar is...?
>
> Here is one for the definition buffs:
>
> What do I call the relation between 2 structures of same SG that pack
> similarly but have too different cell constants/volume (5%) to be
> called isomorphous, but are still - what? Isostructural? Near-isomorphous?
>
> Thx, BR
>
--
===============================================================
* *
* Gerard Bricogne [log in to unmask] *
* *
* Global Phasing Ltd. *
* Sheraton House, Castle Park Tel: +44-(0)1223-353033 *
* Cambridge CB3 0AX, UK Fax: +44-(0)1223-366889 *
* *
===============================================================
|