Secondly, structure solution routines are now quite powerful and can generally
provide an appropriate starting point for structure refinement – which in a large
number of cases cannot be satisfactorily completed. There is therefore a requirement
to devote effort to improving and developing approaches to modelling and refining
such structures. Structure refinement software has for many years been driven by the
few and taken for granted by the many. The community now faces a grave problem
in that there is dwindling expertise in the area of crystal structure refinement software
development. This is in part due to the ‘golden age’ of crystallographers who
revolutionised service crystallography and solid-state structure characterisation in
the 1960s, 1970s, and 1980s being at the end of their careers. There has been little or
no business or academic reward incentives driving software development in the
intervening decades, resulting in a degree of stagnation. There have been some
advances in structure refinement, e.g. rigid body and new constraints or restraints,
with Olex2 [84] and CRYSTALS [85] introducing some notable tools, but much
more development is required and particularly for the newer challenges that single
diffraction is generating. In the modern digital era, the need for software developers
has been recognised in society and in academia to an extent – and our subject area is
no exception.
3 The Database Revolution
3.1 Nature of the CSD
Several databases of crystallographic data exist, including the three detailed here.
The Cambridge Structural Database (CSD) [14] contains the world’s collection of
organic and metal-organic small molecule crystal structures collected and shared by
the CCDC (Cambridge Crystallographic Data Centre). Inorganic structures are
collected in the Inorganic Crystal Structure Database (ICSD) [69] provided by FIZ
Karlsruhe. Protein structures are contained in the Protein Data Bank (PDB) [86]. As
the majority of chemical crystallographers concentrate primarily on molecular materials, this article concentrates for the most part on the CSD.
The CSD was founded in Dr. Olga Kennard’s group at the University of Cambridge in 1965 following inspiration from JD Bernal that the ‘collective use of data
would lead to the discovery of new knowledge; transcending the results of individual
experiments’ [87]. This has been seen in the growth of the CSD, with the database
doubling in size from half a million entries in 2009 to one million in 2019. The
cumulative growth of the CSD can be seen in the stacked columns illustrated in
Fig. 2. An analysis of the percentage of organic and metal-organic structures shows
that the number of metal-organic structures (38,131) first overtook the number of
purely organic structures (37,733) in the database in 1989. Although the ratio of
organic to metal-organic structures across the whole CSD has not changed greatly
since 2009, metal-organic structures currently account for 56.9% of the CSD.
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
105
provide an appropriate starting point for structure refinement – which in a large
number of cases cannot be satisfactorily completed. There is therefore a requirement
to devote effort to improving and developing approaches to modelling and refining
such structures. Structure refinement software has for many years been driven by the
few and taken for granted by the many. The community now faces a grave problem
in that there is dwindling expertise in the area of crystal structure refinement software
development. This is in part due to the ‘golden age’ of crystallographers who
revolutionised service crystallography and solid-state structure characterisation in
the 1960s, 1970s, and 1980s being at the end of their careers. There has been little or
no business or academic reward incentives driving software development in the
intervening decades, resulting in a degree of stagnation. There have been some
advances in structure refinement, e.g. rigid body and new constraints or restraints,
with Olex2 [84] and CRYSTALS [85] introducing some notable tools, but much
more development is required and particularly for the newer challenges that single
diffraction is generating. In the modern digital era, the need for software developers
has been recognised in society and in academia to an extent – and our subject area is
no exception.
3 The Database Revolution
3.1 Nature of the CSD
Several databases of crystallographic data exist, including the three detailed here.
The Cambridge Structural Database (CSD) [14] contains the world’s collection of
organic and metal-organic small molecule crystal structures collected and shared by
the CCDC (Cambridge Crystallographic Data Centre). Inorganic structures are
collected in the Inorganic Crystal Structure Database (ICSD) [69] provided by FIZ
Karlsruhe. Protein structures are contained in the Protein Data Bank (PDB) [86]. As
the majority of chemical crystallographers concentrate primarily on molecular materials, this article concentrates for the most part on the CSD.
The CSD was founded in Dr. Olga Kennard’s group at the University of Cambridge in 1965 following inspiration from JD Bernal that the ‘collective use of data
would lead to the discovery of new knowledge; transcending the results of individual
experiments’ [87]. This has been seen in the growth of the CSD, with the database
doubling in size from half a million entries in 2009 to one million in 2019. The
cumulative growth of the CSD can be seen in the stacked columns illustrated in
Fig. 2. An analysis of the percentage of organic and metal-organic structures shows
that the number of metal-organic structures (38,131) first overtook the number of
purely organic structures (37,733) in the database in 1989. Although the ratio of
organic to metal-organic structures across the whole CSD has not changed greatly
since 2009, metal-organic structures currently account for 56.9% of the CSD.
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
105