To address this challenge, a UV-crosslinkable polymer is combined with an ionic liquid and a silver salt. The crosslinkable polymer provides mechanical structure, while the ionic liquid serves as a non-volatile plasticizer to increase ion transport. Poly(ethylene glycol) diacrylate (PEGDA) is the crosslinkable polymer, 1-butyl-3-methylimidazolium tetrafluoroborate is the ionic liquid, and AgBF4 is the silver salt. Ionic liquids are chemically and thermally stable with negligible vapor pressure, high ionic conductivity, and a wide electrochemical window. 100 nm thick films of 80/20 and 60/40 wt% PEGDA/IL are spin-coated, annealed, and measured by AFM inside an argon-filled glovebox where O2 and H2O are controlled to < 0.1 ppm. Using a conducting AFM tip as one electrode and a silver substrate as the other, automated measurement of hundreds of filament formation/dissolution events are made at predefined locations in a single experimental session. Filament formation occurs at a 3 V forward bias over timescales that range five orders of magnitude (Fig. (a)), while dissolution occurs at a 3 V reverse bias on the timescale of hundreds of milliseconds.
As shown in Fig. (a), the formation time distributions for both 60/40 and 80/20 PEGDA/IL samples are bimodal: the fast formation times occur within a few milliseconds while the slow can require several hundreds of seconds. This is likely a consequence of IL-rich and poor regions within the film, as supported by AFM modulus measurements showing regions of high and low elastic modulus (Fig. (b)). Filaments are expected to form faster in regions of low modulus, and slower in regions of high modulus. This explanation is further supported by comparing the samples containing 60 versus 80 wt% PEGDA: the distribution shifts towards increasing formation times with increasing polymer concentration. These results show that the formation/dissolution kinetics can be controlled by controlling the microstructure of the film, and a more homogeneous microstructure will be required to narrow the distribution of formation/dissolution times.
Acknowledgement: This worked was supported by the Defense Advanced Research Projects Agency (DARPA) Atoms to Products (A2P) program, grant #FA8650-15-C-7546.