Using eScience to probe structure and bonding in metal complexes: Database mining and computation

1
Using e-Science to probe structure and bonding in
metal complexes Database mining and computation

• Jonathan Charmant, Frederik Claeyssens, Natalie
  Fey,
• Mairi Haddow, Stephanie Harris, Jeremy Harvey,
  Tom Leyssens, Ralph Mansson, A. Guy Orpen and
  Athanassios Tsipis

CombeDay 2005 Southampton
2
Reactivity
e-Science
Structure
Properties
3
Metal-ligand binding
Tsipis, Orpen and Harvey, Dalton Trans. submitted
4
Metal-ligand binding 2
Database mining correlation between Oxidation-R
eduction and MPX angle and PX distance Cause
p back-bonding?
Leyssens, Orpen, Peeters and Harvey, to be
submitted
5
Metal-ligand binding a systematic approach
61 ligands, ca. 10 calculations on each ? Ligand
Knowledge Base
6
Map of Chemical Space
NR2 OR Hal Ar R
7
Model Building
Fey, Tsipis, Harris, Harvey, Orpen Mansson, to
be submitted

• Predict experimental data from calculated
  variables.
• Multiple linear regression

Solid State Rh-P Distance (Rh(I), CN4)
Tolman Electronic Parameter
8
Adding value to the structural database
Query Geometry Library for User-Defined Fragment
retrieval of matching data
Output of Statistical Data
apply outlier criteria
Outliers
Fey, Harris, Harvey and Orpen, to be submitted
DFT geometry optimisation
Optimised Geometries
compare with crystal structures
Crystal Structure and DFT agree
Crystal Structure and DFT disagree
9
Adding value to the structural database 2
10
Conclusions

• Structural database is full of data
• Data Mining already known to yield valuable
  insight
• Combine database with computation to yield more
  insight
• Probe structure and reactivity of individual
  species
• Generate ligand knowledge base
• Probe structural trends and outliers