The study of porous liquids is a relatively young field. It caught our attention through the groundbreaking work of Stuart James, Andy Cooper and Becky Greenaway, who have made use of covalent organic cages to define pores within bulky molecular solvents to yield numerous new Type II porous liquids. In the Nitschke group, much of our work focuses on developing metal-organic coordination cages. We started to wonder whether we could design new porous liquids based on coordination cages, and theorized that basing porous liquids on a different type of nanocapsule would facilitate new types of function. We also saw an opportunity to eliminate the need for a solvent, since our coordination cages can be built in modular fashion, allowing the external derivatization of the cage to be easily tuned.  Multiple design iterations were necessary to yield a coordination cage that was actually liquid at room temperature, and we were extremely pleased when our hard work paid off with the present design.

Figure 1. Liquid coordination cage 2 was used to sequester different CFCs in its neat state at room temperature.

Porous liquids are often studied as bulk materials, using techniques like DSC/TGA, rheology and PALS. However, the coordination cages known to the Nitschke group are usually studied in solution using NMR techniques, yielding information at the molecular level. This prompted us to attempt to extend the use of this technique to the realm of porous liquids. However the high viscosity of our material presented several challenges to overcome, starting with the not-so-simple question of how to load our liquid cage into an NMR tube. After much deliberation, we found that scraping the viscous material into the opening of a 3 mm NMR tube and then spinning the sample in a centrifuge at 3000 RPM was the best technique for transferring the material into the bottom of a tube. We then pushed a 2 mm glass capillary containing D₂O into the tube to increase the height of the sample and to provide a deuterated signal for locking (you can see the liquid cage flowing up the walls of the NMR tube in our video). The second challenge we faced was that the high viscosity of the sample meant that the spectra we collected were very broad, probably due to slow tumbling. However, we found that analyzing the samples with a probe temperature of 70 °C, allowed us to observe identifiable peaks consistent with the porous liquid as well as its host-guest complexes. While the obstacles we encountered during this process were very challenging, we were given the opportunity to flex our creative muscles and apply our problem-solving skills. We hope that our experiments will convince others to add NMR onto their list of analytical techniques for analyzing challenging materials! 

Figure 2. a), Partial solution-phase ¹H NMR spectra (298 K, 400 MHz, CD₃CN) of the complexes of cage 2 incorporating the alcohols shown. Host-guest peaks in the aromatic region are labelled with an asterisk (*) and all peaks below 0 ppm are assigned to the internally-bound guest. b), Partial solvent-free variable temperature ¹H NMR spectra (343 K, 400 MHz) of complexes of neat liquid cage 2 with 30 wt% of the alcohols shown in d, taken with the tube assembly depicted in c (30 wt% guest). c), In order to record NMR spectra under solvent-free conditions, neat liquid cage 2 was loaded into a 3 mm NMR tube containing a capillary filled with D₂O to provide an NMR lock signal. d), Cage 2 bound a range of different propanol and butanol isomers; a better shape match between guest and host cavity led to stronger binding. e), Neat liquid cage 2 selectively encapsulated t-butanol over n-butanol.

While NMR was a difficult hurdle to overcome, it was one that we anticipated. Other problems we encountered while working with our porous liquid cage were more surprising. For example, during the latter half of the project, we noticed that occasionally the ¹H NMR spectrum of our material taken in CD₃CN solution would show that 10-20% of the material had been converted to an unidentifiable host-guest complex instead of the desired empty cage. We were able to isolate the empty cage only 50% of the time with the unwelcome host-guest complex appearing the remaining 50% of the time.  We concluded that one of our starting materials was introducing an impurity into the sample, and that this impurity happened to be a good guest for the cage. We therefore ran an extensive series of experiments to find the source of this impurity, including re-synthesizing and re-purifying our starting materials (including commercially available reagents), distilling the solvents used for synthesis/ purification, and varying the reaction vessels used, temperature and reaction time – all to no avail. We finally discovered that the mysterious host-guest complex only appeared if the cage was analyzed in CD₃CN from a particular supplier. The impurity was never identified, and (being NMR silent) could only be detected after it had been bound inside our coordination cage. While the process of arriving at this conclusion was rife with difficulty, our lab now uses this cage as an impurity detector for commercially purchased deuterated solvents – an unexpected but not unwelcome application of this work!

Although our first attempts to combine coordination cages with porous liquids was not all smooth sailing, we’re proud to see coordination cages join the porous liquids family and excited to see what new directions the field takes next.

If you would like to see our whole porous liquids journey, the full article can be found here!

Link for video of liquid coordination cage.

Lillian Ma, University of Cambridge, LinkedIn: /lillianma7

Cally J.E. Haynes, University College London, Twitter: @CallyHaynes