As it has been shown before, COFs can display excellent and diverse properties and a few methods to process them have been developed. However, examples of applications are still scarce and those reported generally do not show competitive performances and durability. The reason behind this is that, beside surface supported films, all these processing strategies are very recent, in fact, the earliest works described in the previous section date from 2015. This implies that there has not been enough time to fully combine into devices the existing knowledge of how to design functional COFs and how to process them, since the new question of to what extent processing affects some properties also needs to be addressed.
Some of the first applications were those using COF thin-films. For example, a COF with a structure featuring electron-donor and acceptor groups was prepared in the form of 50 nm thick oriented films on an ITO electrode with a 10 nm molibdenum oxide layer that acts as a hole extractor. Over the COF film, an aluminium electrode with a 20 nm hole- blocking layer of zinc oxide was deposited. This device contains all the components of a solar cell, being the COF the active layer. Upon irradiation, the device shows efficient charge separation, however, the performance is hindered by the high rate of recombination124.
Another example of COF thin-film adapted to application is found in the field of electrical energy storage. An anthraquinone containing COF thin film was prepared on gold because its redox behaviour shows potential for charge storage. However, the poor conductivity of the network precluded most of the redox active moieties from participating. For this reason, the conductive polymer PEDOT was polymerised inside the channels of the COF (Figure 1.33a). The beneficial effect of PEDOT in distributing the
Figure 1.33. a) Depiction of the incorporation of PEDOT in the pores of a DAAQ-TFP COF
film by electropolymerisation. b) Cyclic voltammetry response in 0.5 M H2SO4 of a DAAQ-
TFP COF film before (red) and after (blue) PEDOT modification. The inset presents the cyclic voltammetric response for the unmodified film using an expanded current scale. c) Photographs of a PEDOT-modified DAAQ-TFP COF working device powering a green LED. Adapted from reference 125.
charge throughout the whole COF and allowing all the anthraquinones to be reduced was clearly seen by the huge current enhancement in cyclic voltammetries (Figure 1.33b). The full electrochemical studies showed
good performance over a range of charge-discharge rates and decent stability. As a proof-of-concept, two coin cells with COF electrodes were fabricated and succeeded in lighting a LED (Figure 1.33c-d)125.
Moving onto another area, the preparation of monodisperse spheres and the anchoring of COFs to other materials has also been used to facilitate COF application in chromatographic separations. In the simplest case, a spherical COF was charged in a gas chromatography column and used as the stationary phase. The results obtained for alkane, alcohol and pinene isomers separation showed good resolution and selectivity103. However, the full potential of COFs in chromatography was shown in another example in which a chiral COF was synthesised in situ in a fused-silica capillary column. In order to retain the COF and steer its growth to the column walls, the silica was previously functionalised with amino groups. As a result, a chiral capillary column was obtained that even performed better than some commercially available ones in the separation of certain enantiomeric mixtures126.
A last example of application made possible by the availability of a suitable processing technique is the use of COFs for nanofiltration. On the one hand, COF films were formed at the dichloromethane-water interface (see Section 1.4.3) with a thickness of a few micrometres to ensure its mechanical stability. The films were placed between two macroporous supports in order to study their performance when a solution is passed through them in a dead-end cell (Figure 1.34a). The solvent permeance values were better than those of other materials used for nanofiltration and the rejection of solutes was high showing good size selectivity95.
On the other hand, cross-flow nanofiltration (Figure 1.34b) has also been achieved with COFs. To reach this goal, an alumina tube for this type of filtration was functionalised with amino groups that, as has been described several times, allowed the grafting of an imine-based COF to its surface.
The coverage with a 400 nm thick COF layer was complete and uniform (Figure 1.34c-d). As in the previous example, good rejection of dyes dissolved in water was found (Figure 1.34e-f), hinting towards its potential for water purification127.
Figure 1.34. a) Setup of nanofiltration with a COF thin film as membrane in a dead-end
cell. As can be seen, the filtered solution is colourless. b) Scheme of cross-flow filtration. The mixture is fed from the right and flows around the tubular membrane. Only part of the solvent and small molecules can cross the membrane and are collected at the bottom, while larger molecules and the rest of the solvent exit the system from the left. c) Photographs of an untreated (left) and COF functionalised (right) alumina tube for cross- flow filtration. d) EDX mapping of the cross-section of a COF functionalised tube. C is the tracer for the COF layer and Al for the alumina support. e) Water permeance and rejection rates of different dyes with tubular COF membranes. f) Stability test of the tubular COF membranes. g) Comparison of the performance of COFs in nanofiltration with that of MOFs and other porous materials and composites. Adapted from references 95 and 127.