Leaving all fossil fuels in the ground is the best way to keep the carbon sequestered, but the transition to a completely fossil-fuel free many be slower that many of us are hoping. And as a consequence, we are too rapidly consuming our global carbon budget. The amount of CO2 we can still emit to keep the increase in average global temperature below a 2º, is much more limited than we realize. We will reach a point, that, even in a world without fossil fuels, we need to capture CO2 molecules from any source; a cement factory, the production of biogas, a waste incineration plant, or even directly from the air if we have exceeded our budget. The CO2 can then be permanently stored in a geological formation; or if we want to close the carbon loop for the chemical industry the CO2 can replace fossil fuels as the source for carbon; or to close the loop for transportation we can use renewable energy to upgrade the CO2 to synthetic fuels for airplanes and ships. 

In such a world, one can envision that we need an optimal separation for any source of CO2, and the requirements/demands of the separation process will depend on the destiny for CO2. The current technology is focused on liquid amine-based absorption; “one-amine-based-fits-all” solution. Our research is motivated by the idea that we can do better; by capturing CO2 with a solid adsorbent we will be able to design a material in which every atom is exactly at the right place to capture CO2. Metal organic frameworks (MOFs) are ideal for realizing this dream of designing exactly the right material. By combining metal nodes with organic linkers one can envision millions of possible nanoporous crystalline materials, e.g. MOFs. With so many possible materials it is, however, beyond the imagination to synthesize and test them all. 

The way out is to rely on computational methods, materials genomics, in which we generate in silico hundreds of thousands of MOFs and screen them for their CO2 adsorption capacity and selectivity. The problem is, however, that these libraries are completely ignorant on the most important step; these MOFs need to be synthesized. Synthetic chemists have collected much knowledge about the synthesis of MOFs and the challenge is to use this knowledge in our screening studies. Most previous studies generated many top performing structures, but left us completely in the dark on how to make them. 

The solution we found was inspired by drug design, in which libraries of known chemicals are computationally screened for those chemicals that strongly bind to a target site of a protein. A strong binding energy is important for a drug molecule to work, but there are many additional factors that are not included in the screening (e.g., toxicity, stability, etc.). Therefore, the approach is to identify the common molecular feature (i.e., the pharmacophore) of these chemicals that bind strongly. It is then that synthetic groups are able to design a new drug based on the pharmacophore but now with the knowledge on those additional factors that make it a successful drug molecule.

In our case, we have our drug molecule, CO2, but we generated over 300,000 MOFs to analyze for the common features of the top performing materials that can separate wet flue gasses. These common features, which we call adsorbaphores, where then combined with the knowledge on how to synthesize MOFs to arrive at a set of synthesizable top performing materials. These MOFs were subsequently synthesized and performance tests showed that they not only behaved as predicted, they even outperformed commercial adsorbents.

Berend Smit (@SmitBerend), École Polytechnique Fédérale de Lausanne (EPFL), Switzerland https://www.epfl.ch/labs/lsmo/

Susana Garcia (@sgarclopez), Heriot-Watt University, Edinburgh, UK. https://rccs.hw.ac.uk/

Jeffrey Reimer, University of California, Berkeley, Berkeley, CA, USA. http://india.cchem.berkeley.edu/~reimer/

P. Boyd, A. Chidambaram, E. García-Díez, C. P. Ireland, T. Daff, R. Bounds, A. Gładysiak, P. Schouwink, S. M. Moosavi, M. M. Maroto-Valer, J. Reimer, J. Navarro, T. Woo, S. Garcia, K. Stylianou, and B. Smit, Data-driven design of metal-organic frameworks for wet flue gas CO2 capture. Nature (2019) Doi: 10.1038/s41586-019-1798-7