Screening for selectivity

Metal–organic frameworks are promising adsorbents for CO2 capture from flue gas, but many perform poorly when exposed to flue gas containing water. Now, a computational screening approach identifies MOFs with preserved CO2/N2 selectivities in wet flue gas and experiments confirm their outstanding CO2 capture performance.

One of today’s biggest challenges is to limit the increase of CO2 in the atmosphere. Carbon capture and storage offers a potential route to help with this1. To that end, a large number of porous adsorbents have been developed to effectively capture CO2 from flue gas emitted by anthropogenic sources2. One class of adsorbents that have great potential in CO2 capture are metal–organic frameworks (MOFs): crystalline materials composed of metal ions connected by organic linkers3,4. However, a limitation of MOFs for CO2/N2 separation is that they do not perform well when they are exposed to wet flue gas mixtures, as would be typical in real applications. This is because water competes with CO2 for the same adsorption sites of MOFs and causes a significant reduction in their CO2 selectivity. Drying the flue gas before using MOFs is thought to make the process prohibitively expensive. Therefore, design and development of MOFs that can efficiently capture CO2 from wet flue gas has been a long-standing goal. Writing in Nature, Tom Woo, Susana Garcia, Kyriakos Stylianou, Berend Smit and co-workers across Europe and North America now report large-scale computational screening of materials, combined with experiments, for the data-driven design and synthesis of MOFs with high CO2 capture performance from wet flue gas5.

Perhaps the most important advantage of MOFs over traditional porous materials is that their pore sizes and shapes, as well as the affinity of the pores towards different gases, can be tuned by changing the combination of linkers and metals. A diverse set of MOFs have been synthesized to date and a theoretically unlimited number of MOFs can be created. The existence of thousands of MOFs having different physical and chemical features offers an excellent opportunity to achieve high performance CO2 separation.

At the same time, this large materials space imposes a challenge for the materials search, as it is not possible to test CO2 separation performances of all available MOFs using purely experimental techniques. Therefore, computational screening methods play a critical role in examining a large number of MOFs in a time-effective manner to identify the most promising materials for target gas separations6. Experimental efforts can then be focused on those candidates.

In their work, Woo, Garcia, Stylianou, Smit and co-workers first generated a library of 325,000 hypothetical (computer-generated) MOFs. Using molecular simulations, they predicted the CO2/N2 selectivity and CO2 working capacity of all materials at conditions that mimic post-combustion flue gas separation processes. From these calculations, 8,325 MOFs were identified as top-ranked materials exceeding the performance of commercial zeolite 13X under dry flue gas separation conditions (CO2 working capacity >2 mmol g–1 and CO2/N2 selectivity >50).

The researchers introduce the term ‘adsorbaphore’ to describe the common pore shape and chemistry of a binding site in the MOF that provides optimal interactions to preferentially bind to CO2. This is inspired by the rational design of drug molecules where pharmacophores describing common features of binding sites for particular molecules are identified. CO2 binding sites of the top-ranked MOFs were identified and categorized into three different classes of adsorbaphores. MOFs having the parallel aromatic rings adsorbaphore for CO2 (termed A1 as shown in Fig. 1) were calculated to have a low affinity for H2O.

Fig. 1

a,b, Structural representation of the two MOFs identified for CO2 capture from wet flue gas mixtures, Al-PMOF (a) and Al-PyrMOF (b). Red box, Adsorbaphore A1, which represents the planar aromatic systems where CO2 binds in between. Pink atom, Al. Grey atom, C. Blue atom, N. Red atom, O. Pale yellow atom, H. Images are reproduced from ref. 5, Springer Nature Ltd.

Next, the researchers undertook reverse-searching for the A1 adsorbaphore in the MOF database. Molecular simulations on a subclass of MOFs that contain this preferred adsorbaphore showed that predicted materials maintain an excellent CO2 selectivity at low pressures. Moreover, selectivities of many MOFs were not affected by the presence of H2O at flue gas conditions.

Based on these computational predictions, the researchers synthesized two MOFs (Al-PMOF and Al-PyrMOF) and demonstrated that experimentally-measured CO2 and N2 adsorption isotherms of these materials are in good agreement with the simulated ones. Breakthrough experiments showed that the CO2 capture capacity of Al-PMOF is minimally influenced by humidity, while an improvement of the performance was reported for Al-PyrMOF. Excitingly, the CO2 working capacity of Al-PMOF under humid conditions outperforms that of zeolite 13X and activated carbon. This finding smartly demonstrates the data-driven design of a new MOF material with extraordinarily high CO2 capture potential from wet flue gas.

The most exciting aspect of this study is linking computational screening and synthesis of the corresponding materials through the identification of structural motifs called adsorbaphores, which enhances the synthetic viability of the approach. This work offers a foundation for future studies to identify adsorbaphores in MOFs that provide optimal interactions to preferentially bind different guests such as CH4 and H2, which can eventually accelerate the design of MOFs for numerous gas storage applications. Extending the results of this work to various gas separation processes could also be fruitful in the discovery of targeted adsorbent materials. For example, MOFs examined in this study may also have the potential to capture CO2 from natural gas mixtures. Natural gas in reservoirs contains CO2 and H2O, which reduce the energy content of natural gas and cause corrosion in pipelines.

Finally, it would be interesting to test the MOFs synthesized in this study under industrial operating conditions to accurately assess their real-world performance and stability. The future holds great promise for the large-scale applications of MOFs, but further technical and economical evaluation of the full CO2 capture process using the new MOFs will be crucial in order to lead to a new MOF-based CO2 separation technology.


  1. 1.

    Bui, M. et al. Energ. Environ. Sci. 11, 1062–1176 (2018).

    Article  Google Scholar 

  2. 2.

    D’Alessandro, D. M., Smit, B. & Long, J. R. Angew. Chem. Int. Ed. 49, 6058–6082 (2010).

    Article  Google Scholar 

  3. 3.

    Schoedel, A., Ji, Z. & Yaghi, O. M. Nat. Energy 1, 16034 (2016).

    Article  Google Scholar 

  4. 4.

    Trickett, C. A. Nat. Rev. Mater. 2, 17045 (2017).

    Article  Google Scholar 

  5. 5.

    Boyd, P. G. et al. Nature (2019).

    Article  Google Scholar 

  6. 6.

    Boyd, P. G., Lee, Y. & Smit, B. Nat. Rev. Mater. 2, 17037 (2017).

    Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Seda Keskin.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Keskin, S. Screening for selectivity. Nat Energy 5, 8–9 (2020).

Download citation

Further reading


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing