Retrofitting MOFs

What is Retrofitting?
Originally coming from construction, retrofitting describes the installation of cross linkers (CLs) into an existing framework or structure. In the context of metal-organic frameworks (MOFs), retrofitting of a MOF relates to the post-synthetic installation of CLs such as molecules with dinitrile or dicarboxylate functionalities between undercoordinated metal nodes or open metal sites (OMSs) of a MOF.[1] Over the past years, retrofitting of MOFs has shown to be a powerful concept to post-synthetically manipulate the physicochemical properties of MOFs such as their response to hydrostatic pressures, their structural flexibility and the thermal expansion behaviour.[1],[2] It should be noted that retrofitting of a MOF currently leads to a defect-state – breaking the periodicity of the parent framework – a symmetry-breaking mechanisms that is unique to porous functional systems, drawing a clear line to pure solid-state materials.

What is RetroFit?
RetroFit is a python based open-source program that is set up to guide experimentalists in the selection of promising CL@MOF systems for their labwork to avoid time-consuming trial-and-error experiments. At the heart of RetroFit is a geometrical optimization routine, which assesses the fit of a CL to the OMSs of a MOF. Obtained is a list of energy penalties for a given CL@MOF system, providing experimentalists with a guideline for setting up synthesis experiments. The approach is inspired by molecular docking, which is an important tool in structural molecular biology to assess ligand-to-protein interactions. The python file including a how-to-guide can be found on and has been published in 2019.[3] In the current version of RetroFit v1.0, a library of 20 CLs exist which can be used to screen MOFs with Cu paddlwheel-based OMSs. In the future, it is planned to further extend the model interaction potential to other metals, additional metal-node motifs and to go beyond dinitrile-based CLs such as amines, isonitriles or thiocyanates. We are very happy about any feedback / suggestions / potential collaboration regarding the program both experimentally and computationally.

Figure 1. A schematic of three different CLs fitting between the OMSs of Cu3BTC2 with BTC = 1,3,5-benzentricarboxylate. The coloured surface relates to the MIP (Model Interaction Potential) which is used as basis for the optimization routine in RetroFit (blue = favourable position of the nitrile functionality, red = unfavourable position). Visually, TCNE shows the best fit, with the dinitrile functionality in a advantageous spatial orientation for the maximisation of the underlying coordination bonds (TCNE = tetracyanoethylene).

Open scientific questions
Retrofitting MOFs is a relatively new concept with only a limiting number of synthetic examples existing. In turn, a relatively large number of scientific questions are outstanding, mostly related to the impact of the CL and the underlying ordering mechanism. Some of the most intriguing questions are, ‘how does the CL order within the pores of the parent MOF?’’is there an impact of the chemistry of the CL, i.e. the CL-to-OMSs bond strength on the properties?’ and ‘is it possible to extend the concept to CLs with some sort of functional backbone?’. These and closely related questions are currently being investigated in our research group, ulitimately leading us to a more fundamental understanding of host-guest interactions in MOFs and its underlying structure dynamics.

[1] E. A. Kaputstin, S. Lee, A. S. Alshammari, O. M. Yaghi J. Am. Chem. Soc. 2017, 3, 662.
[2] C. Schneider, D. Bodesheim, M. G. Ehrenreich, V. Crocellà, J. Mink, R. A. Fischer, K. T. Butler, G. Kieslich J. Am. Chem. Soc. 2019, 141, 10504.
[3] C. Schneider, D. Bodesheim, J. Keupp, R. Schmid, G. Kieslich Nat. Commun. 2019.