[IUCr Home Page]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Phase ID proposal 4



Dear Colleagues,

     Michael Berndt has made some valuable suggestions for what
we should include in our phase identifier based on his hands-on
experience.  This has mostly been with inorganic phases which may
be the most difficult problem, but we need some input from those
with experience with organic and biological compounds.  Perhaps
Sam Motherwell or John Westbrook have some suggestions.  In this email I
present an updated proposal (number 4).  Can you please check it over and
reply with your comments to the group list.

General comments
----------------
     Michael's suggested identifiers can be divided into those
that provide chemical characterization, and those that provide
crystallographic characterization.  For inorganic compounds a sum
formula is sufficient for a chemical characterization, but for
organic molecules we need to have some way of identifying the
different isomers since the sum formula by itself is not
sufficient.  The crystal structure can be best identified by the
space group number and Wyckoff sequence as Michael suggests
though for organic compounds the Wyckoff sequence may less
helpful since usually only general positions are occupied.

     The crystal system and lattice centring are implicit in the
space group number but could be useful in cases where the space
group is not known.

     There are cases where different Wyckoff sequences correspond
to the same structure.  For example in P-1 the sequences a,b;
c,g; d,f and e,h all describe the same structure and differ only
in the choice of origin.  We could require that where a choice of
Wyckoff positions is possible, the lowest letter in the alphabet
should be chosen.  Michael, have you run across this problem, and
if so, how do you deal with it in your program?

     Misassigned space groups will be a major problem
particularly with older studies.  We might expect that a wrong
space group would be a sub- or super-group of the correct one so
the probability of a match would be higher if we could identify
the sub- and super-groups of the target.  This could be done
using a matrix such as that shown in Table 1 below.  It would not
however be possible to compare the Wyckoff sequences in such
cases without a considerably more sophisticated program.  If the
two materials were characterized under the same conditions of
pressure and temperature and had related space groups, there is a
good chance that at least one of the space groups was misassigned
and the phases are the same.  If they were characterized under
different conditions there would be a higher chance that the
phases really are different.  It would therefore be helpful to
have information on the temperature and pressure at which the
material was characterized.

     We could prepare a table (see Table 1) showing all the sub-
and super-space groups in a matrix form to test whether two
phases reported with different space groups might be the same.
The use of such a matrix would have additional advantages.  It
could contain the crystal system and lattice type for use in
cases where the space group was not known, and enantiomorphic
space groups could be equivalenced so that, e.g., 76 (P41) would
be considered the same as 78 (P43).  If it were necessary to
distinguish optical enantiomers, this should be done using a
different item, since in many cases the choice between P41 and
P43 is made arbitrarily.

                         TABLE 1
                         -------
Partial matrix of sub- and super-groups of each space group with
the crystal system and lattice type shown
-----------------------------------------------------------
     1    2    3    4    5    6    7    8    9    10   11   12
     aP   aP   mP   mP   mE   mP   mP   mE   mE   mP   mP   mE
1    0    x    x    x    x    x    x    x    x    x    x    x
2    x    0                                       x    x    x
3    x         0         x                        x         x
4    x              0    x                             x
5    x         x    x    0                                  x
6    x                        0         x         x    x    x
7    x                             0         x
8    x                        x         0
9    x                             x         0
10   x    x    x              x                   0
11   x    x         x         x                        0
12   x    x    x         x    x                             0


     Michael points to the possibility of two different phases
having the same Wyckoff sequence:
          'In rare cases we find compounds with the same chemical
     composition and the same Wyckoff sequence but nevertheless
     describing different modifications (Example: ICSD entries
     24674 and 27116; both H7 N O6; Wyckoff sequence: 19, a7).'
This should not be a serious concern.  This particular example
consists of two different structures proposed for the same phase
so they ought to match in a database search.  The structures were
proposed around 50 years ago when there was concern about
homomorphic structures, structures that could never be
distinguished because they give rise to identical diffraction
patterns.  Both the structures quoted above are published by the
same author but only one is likely to be correct.  True
homomorphs are very rare and only one of them will have a
reasonable crystal chemistry.  We are more likely to run into the
problem of structures with identical Wyckoff sequences in organic
compounds where all the atoms occupy general positions.  A method
of distinguishing different isomers would help here.

     We will never be able to get a perfect score using
identifiers based on the observed properties of the phases
because it is first necessary that each material must be fully
and correctly characterized, which is not always the case.  We
should aim to ensure as high a success rate as possible with the
minimum of false matches.  We should be able to achieve near
perfect matches for correctly and fully characterized phases, but
in other cases we should be content if we can retrieve all the
poorly characterized phases that are consistent with the target
phase.  We should allow for matching programs to include user-set
tolerances for matches, e.g., 10% variation in composition,
acceptance or non-acceptance of sub- or super-groups of the
target space group, etc.

PROPOSED CONTENTS OF THE IDENTIFIER
-----------------------------------

Chemical characterization:
--------------------------
Primary:       *Sum formula. For non-molecular (e.g. inorganic)
               materials only the relative abundances would be
               meaningful, for molecular materials (e.g. organic)
               the absolute abundances would be needed to aid in
               identifying the molecule.
Secondary:     *Some method is needed to identify the topology
               (connectivity) particularly for molecular
               materials.  This might also be useful for some
               non-molecular materials.  Sam, do you have any way
               of dealing with this problem in the CSD?

Crystallographic characterization:
----------------------------------
Primary:       *Phase and structure type (amorphous, liquid,
               Pauling file standard structure code).
               *The space group number.
               *Wyckoff sequence.
Secondary:     *The crystal system and lattice centring could be
               given instead of the space group if the latter is
               not known.  This information can be included in
               the space-group matrix shown in Table 1 to help in
               matching this information to particular space
               groups.
               *Official mineral name for minerals.  Technically
               these names should be used only for natural
               materials but we could extend them to synthetic
               analogues since we are interested here in the
               structure and not the origin of the material.  See
               Table 2 for possible problems.

Other properties:
-----------------
Primary:       *Temperature and pressure at which the material
               was characterized.  This would not itself be part
               of the identifier but could be used to indicate
               the likelihood of a close match representing the
               same phase.  For example, if two phases had space
               groups that were closely related they would be
               more likely to be the same phase if they were
               measured at the same temperature and pressure.  We
               should allow for ranges of temperatures and
               pressures to be included in this field.
Secondary:     *These would include colour, form etc. that might
               be useful for poorly characterized materials.

                         TABLE 2
                         Example
                         -------
The following table shows how these identifiers might be assigned
in the Pb/Sb sulfide system.  I have included both the space
group number and the crystal system together with lattice
centring to show what each would look like, though only the space
group number is needed.  The mineral name has also been given
(even when the material was synthesized in the laboratory) to
show how it can be helpful for phase identification.  For an
explanation of possible problems, see the notes on each of the
entries below.

#    Proposed ID                                  Sb/(Sb+Pb)
-----------------------------------------------------------------
1    Pb-S4-Sb2,*,liq,*,*,*,900K                   0.75

3    Pb-S,*,NaCl,225cF,b-a,Galena                 0.00

4    Pb7-S13-Sb4,*,oxtl,19oP,a24                  0.36

5    Pb3-S6-Sb2,*,oxtl,62oP                       0.40

6    Pb5-S11-Sb4,*,oxtl,62oP,c20,Boulangerite     0.44
7    Bi0.3-Pb5-S10.7-Sb3.7-Se0.3,*,oxtl,14mP,e40,Boulangerite
                                                  0.43
8    Pb4.82-S11-Sb4.11,*,oxtl,62oP,c21,Boulangerite  0.46
9    Pb9-S22-Sb9,*,oxtl,62oP,c23,Boulangerite     0.50

10   Pb9-S21-Sb8,*,oxtl,15mE,f19-e2,Semseyite     0.47
11   Pb2-S5-Sb2,*,oxtl,62oP,c9                    0.50
12   Pb4-S11-Sb4,*,oxtl,55oP,h4-g5-b              0.50
13   Pb7-S19-Sb8,*,oxtl,15mE,f16-e2,Heteromorphite 0.53
14   Pb5-S14-Sb6,*,oxtl,2aP,i25                   0.55

15   Pb4-S13-Sb6,*,oxtl,12mE,i23,Robinsonite      0.60
16   Pb4-S13-Sb6,*,oxtl,1aP,a46,Robinsonite       0.60

17   Pb5-S17-Sb8,*,oxtl,15mE,f14-e2,Plagionite    0.62

18   Pb1.6-S7-Sb3.4,*,oxtl,173hP,c12,Zinkenite    0.68
19   Pb18-S81-Sb42,*,oxtl,173hP,c13-a2,Zinkenite  0.70

20   Pb3-S15-Sb8,*,oxtl,15mE,f12-e2,Fueloeppite   0.73

21   S15-Sb9.8,*,oxtl,15mE,f12-e2,(Pb-free Fueloeppite)1.0

22   S3-Sb2,*,Sb2S3,62oP,c5,Stibinite             1.0
23   S3-Sb2,*,Sb2S3,47oP,l3-k3-j3-13,Stibinite    1.0
24   S3-Sb2,*,Sb2S3,31oP,a10,Stibinite            1.0

NOTES TO TABLE 2
#    ICSD numbers and comment
---------------------------------------------------------------
1    Liquid (temperature given in K)
     3    38293, 62190, 62191-4, 63091-5, 68701, 68712, 68969, 80539.
     4    75143.
     5    This phase is not well characterized
     6    201310, 300107. Note the different space groups and
     	compositions reported for Boulangerite
     7    68663. This natural sample contains Se and Bi.
     8    41273. The increased Wyckoff count is the result of an Sb
 	 atom being displaced from its ideal position onto two nearby
  	 sites.  This could be a problem for positionally disordered
  	 structures.
     9    37441. The formula as given in not electroneutral.
     10   38838.
     11   35640, 35640, 300106.
     12   200601. This composition is not electroneutral
     13   100295.
     14   74441. Pearson symbol given as aI100 in ICSD
     15   300109. Pearson symbol given as mI92 in ICSD
     16   20171. Wrong assignment of space group in this structure
    	 determination, but it is a subgroup of SG # 12.  It is a
    	 space group not commonly found and therefore suspect.
     17   23569.
     18   61191. Zinkenite assigned different compositions and site
    	 occupancies
     19   30781. Pearson symbol given as hP71 in ICSD
     20   142, 8168, 23661, 41849.
     21   60818. Pearson symbol given as mC99 in ICSD may correspond
	 to actual cell contents
     22   15236, 22176, 26751, 30779. 41929.
     23   82871. Probably incorrect space group
     24   85302. Probably incorrect space group

     Can you please review this proposal and circulate your
comments to the group.

          I wish you all the very best for 2003.

                    David




*****************************************************
Dr.I.David Brown,  Professor Emeritus
Brockhouse Institute for Materials Research,
McMaster University, Hamilton, Ontario, Canada
Tel: 1-(905)-525-9140 ext 24710
Fax: 1-(905)-521-2773
idbrown@mcmaster.ca
*****************************************************


Reply to: [list | sender only]


Copyright © International Union of Crystallography

IUCr Webmaster