Data Mining with DDView and the PDF4 Databases

1
Data Mining with DDView and the PDF-4 Databases

• Carbamazepine Polymorphs

Some slides of this tutorial have
sequentially-layered information that is best
viewed in Slide Show mode
2
This is one of three example-based tutorials for
using the data mining capabilities of DDView
with the PDF-4 database and it covers the
following topic

• Carbamazepine Polymorphs
• a PDF-4/Organics application
• investigating polymorphic forms of an active
  pharmaceutical ingredient (API)

Two other similar tutorials for data mining exist
and cover the following topics

• CIGS Photovoltaics
• solid solution / cell parameter relationship
• FeO Non-stoichiometric Oxides
• sorting out temperature and stoichiometric
  effects on cell parameters

3
Carbamazepine Polymorphs

• An example for the PDF-4/Organics 2008 database
  that answers the following questions
• How many polymorphs are known to exist?
• How do I distinguish these polymorphs by XRPD?

4
Carbamazepine Polymorphs

• Search for PDF entries
• Empirical Formula C15H12N2O
• Note that this is not unique to carbamazepine
• Name carbamazepine
• By itself, this name search will also include
  hydrates, solvates, derivatives
• Elements Only C, H, N, O
• This could be performed, but really is a
  redundancy if the empirical formula search given
  above is performed

5
Results Table Preferences

• Useful information for sorting out polymorphic
  forms should be in the results table for
  retrieved entries
• A plot of space group vs. reduced cell volume can
  often provide relevant groupings for polymorph
  differentiation
• Another useful plot would be any of the reduced
  cell edge lengths vs. reduced cell volume
• Suggested fields for results table

• International Space Group
• Space Group Number
• Reduced cell a
• Reduced cell b
• Reduced cell c
• Reduced cell volume

• PDF
• Empirical Formula
• Quality Mark
• Compound Name
• Common Name

6
Specifying Results Table Preferences for
Carbamazepine Polymorphs Search
Selected Fields Use these buttons to move a
selected item up or down in the listed order for
the results table. Prepare the Selected Fields
list for this exercise as illustrated here.
Available Fields Use these buttons to move
selected items between the Available Fields
list of 60 items and the Selected Fields list
of items that will be displayed in the results
table.
7
Entering the Empirical Formula Search Criterion
Empirical Formula criteria is entered on the
Elements tab of the Search window.
There should be no space between each element and
its subscript, and one space before entering
each subsequent element. These elements can be
entered in any order using the Contains
Elements option, but must match the order of the
formula on the PDF card if the Contains Phrase
option is used.
8
Entering the Carbamazepine Name Search Criterion
The name criterion is entered on the Names tab
Since carbamazepine may appear as either a
Compound Name or Common Name, the criterion
is best entered in the All Names category.
Because it is a single word, it does not matter
if Contains Words or Contains Phrase is
selected.
The search can now be performed.
9
Carbamazepine Polymorphs

• 17 hits in PDF-4/Organics 2008 database
• 12 have cell parameter and space group
  information
• 5 different space group designations
  (1,2,14,15,148)

10
Carbamazepine Polymorphs

• To illustrate groupings related to space groups
  and reduced cell volumes given in the results
  table, the following is performed to create the
  appropriate graph

Enter RedCellVol for the X-axis field, and SG
for the Y-axis field
11
Space Group / Reduced Cell Volume Groupings for
12 Carbamazepine PDF Entries
Note the groupings have an approximate 234
Reduced Cell Volume ratio. The Z values for the
unit cells in each group will almost assuredly
have the same ratio.
R-3
The groupings suggest that the likely number of
known polymorphs for carbamazepine is 3, 4, or 5.
P-1
C2/c
P1
P21/n
12
Carbamazepine Another Viewa-axis Length vs.
Reduced Cell Volume
Beta (6)
Form IV (1)
Alpha (1)
Gamma (4?)
Here, the groupings have been labeled based on
information from the Compound Name or
Comments field from at least one member of the
group.
13
Carbamazepine Comparing XRD Data

• To sort out all the entries, including those
  without unit cell information, the X-ray powder
  diffraction (XRPD) patterns themselves can be
  used
• With DDView, the user can overlay simulated XRPD
  patterns based on individual PDF entry data
• Example overlay simulated XRPD patterns for
  entry 00-033-1566 (reported as alpha form but
  without cell parameters) and entry 00-043-1998
  (reported as alpha form with cell parameters
  included)

Double click anywhere on these Results table
rows to open the corresponding PDF entries (or
right-click and select Open PDF Card)
14
Overlaying Simulated XRPD Patterns for Comparison
This icon, when clicked, displays an XRPD pattern
simulated from the peak list of the current PDF
entry.

• To overlay an XRPD pattern from another entry,
  use the Plots drop-down menu and select Add
  Full Trace .

• To overlay an XRPD pattern from another entry,
  use the Plots drop-down menu and select Add
  Full Trace .
• Enter the PDF entry number for the pattern to be
  overlaid 00-043-1998.

• To overlay an XRPD pattern from another entry,
  use the Plots drop down menu and select Add
  Full Trace .
• Enter the PDF entry number for the pattern to be
  overlaid 00-043-1998.
• Once the PDF has been entered, click OK to
  overlay the requested XRPD pattern.

Additional XRPD patterns can be overlaid using
the Plots/Add Full Trace option
15
a-Carbamazepine Pattern Comparison
These two patterns show a similar set of peaks,
at roughly similar positions, indicating
polymorphic similarity. Intensities are somewhat
dissimilar but could be explained by sample
thickness red pattern from thin layer, blue
pattern from thick sample.
The comments section for entry 00-033-1566
indicates that this is a deleted pattern, having
been replaced by 00-043-1998.
16
Mis-assigned Polymorphic Form
Entry 00-043-1988 (red pattern) was also labeled
as the a-form by its author. However, one can
see from the comparison with a known a-form
pattern (top graph blue) and a known ?-form
pattern (bottom graph blue), that this entry is
actually for a ?-carbamazepine.
17
Overlaying Many XRPD Patterns
Using Ctrl-Click, one can select multiple
entries from the Results window and overlay
their XRPD patterns on one graph. Here, all the
reported a-form entries have been selected.
A Right-Click on one of these selected entries
brings up the menu shown here. The Open
Diffraction Pattern choice will prepare the
desired graph of overlaid patterns.
18
Overlay of Reported a-Carbamazepine Patterns on
One Graph
The result clearly shows the obvious difference
between the 00-043-1988 XRPD pattern (blue) and
the other two. One can use this technique to
determine which XRPD patterns among the 17
entries are similar, including those without
polymorphic form or space group designations.
Results of such a determination are shown on the
next slide.
19
Grouping of Similar XRPD Patterns for 17
Carbamazepine PDF Entries
One can overlay the simulated patterns for all 17
PDF entries to establish four groups of similar
patterns. Note Several entries with no space
group or polymorphic form information have been
reasonably assigned to one of these four groups.
9 entries 7 reported as ß-form 5 reported SG as
P21/n 1 reported SG as P21/c
2 entries 2 reported as a-form 1 reported SG as
R-3
5 entries 1 reported as ?-form 1 misreported as
a-form 3 reported SG as P-1 1 reported SG as P1
1 entry 1 reported as Form IV 1 reported SG as
C2/c
20
Carbamazepine - Cell vs. Space Group17 PDF
Entries (12 Experimental, 5 Calculated from
Single Crystal data)
Alpha
Gamma
Beta
Revisiting the Cell vs. Space Group graph, all
17 entries can be assigned to one of these four
groupings, based on similarity of their XRPD
patterns.
Form IV
21
Carbamazepine Polymorphs

• Crystallographic and powder diffraction data
  suggest 4 known polymorphic forms
• Data mining and display capabilities of DDView
  give users the ability to categorize the database
  entries
• The users own pattern can be compared with the
  known polymorph patterns to ascertain polymorphic
  form

22
Thank you for viewing our tutorial. Additional
tutorials are available at the ICDD web site
(www.icdd.com).
International Centre for Diffraction Data 12
Campus Boulevard Newtown Square, PA 19073 Phone
610.325.9814 Fax [phone]