How to generate a solid form landscape without a crystal structure

Back To Discover

Written by

Sophie Bryant

Posted on

December 6, 2021

A solid form landscape shows you the possible polymorphs and stoichiometries your solid form system could take. This allows you to identify risks and find opportunities to impact stability and properties. So how can you foresee solid forms before you have a crystal structure? Here we see how the CSD Landscape Generator in Mercury makes it possible.

What is a solid form landscape, and how can it help my R&D?

When we talk about the solid form landscape, we mean the possible forms a single- or multi-component system could take. Different polymorphs, different stoichiometries, different packing… each form will have different stability and different properties.

In pharmaceuticals, this could mean a less active or even a harmful form. In functional materials it might mean a less accessible, less effective, or non-functional form. In all solid-state materials, the properties can directly impact handling and processing.

One famous example is Ritonavir, which was pulled from production when an alternative polymorph was found—at great financial and human patient cost. Estimates predict that 15–45% of small molecule drugs on the market could show similar issues.

Generating a solid form landscape therefore means that you can spot risks to mitigate, and opportunities to exploit, before they impact your product.

How can I generate a solid form landscape?

Current methods rely on crystal structure prediction, or CSP. These generate large landscapes and predict the energy of each system. However, they are computationally expensive and require a lot of expertise. CSP as a field continues to grow, but challenges remain. (See CCDC Senior Research Fellow, Jason Cole’s blog: What is CSP? And why is it so difficult?)

The CSD Landscape Generator offers a rapid, simple way to generate a solid form landscape without prior expertise or experience. Using fingerprinting and matching molecular shapes (described by Cole et al. here), the system compares your entered molecule to the experimentally confirmed structures in the Cambridge Structural Database and generates a solid form landscape. This is all done in a few clicks in the well-known program Mercury.

You can enter a 1D SMILES or a 2D sketch and get this landscape of possible crystal structures in minutesall based on empirical data.

Watch the 3-minute video showing how it works here.


Can I generate a solid form landscape without a crystal structure?

Yes. You can enter a 1D SMILES string or a 2D sketch of a molecule that has not even been synthesized. Using the empirical data in the Cambridge Structural Database (CSD) or your own proprietary database, the system generates a 3D molecular structure and then putative 3D crystal structures. This is all done in a few clicks in the program Mercury or is also accessible in the CSD Python API. 


Does the CSD Landscape Generator tell me the energy of a system?

No. These are just possible structures to guide and inform further experimental or CSP work, and the energy is not calculated accurately. You will get a simple score that correlates with sublimation enthalpy from known structures in the CSD. An example of the method in use is described in the publication relating to the sixth CSP Blind Test here.


How can I access the CSD Landscape Generator?

You need a licence for CSD-Theory, which also gives access to tools to manage and visualize your CSP data.

Contact us to ask a question or request a demonstration.


Learn more

Learn more about CSD-Theory here.

See what else is new in the latest CSD release here. 



CSD-Theory (3)

Polymorph (11)

Release 2021.3 (5)

Solid Form Informatics (7)