Unlike Li-ion batteries which store energy using solid Li-intercalants, redox flow-batteries (RFBs) utilize electroactive species dissolved in electrolyte to store energy.  The flowability of these species enables the independent scalability of power and energy achievable with a tank/reactor device architecture. Redox-active polymers (RAPs) have shown great promise as candidate electrolytes to mitigate the parasitic crossover of electroactive species in RFBs.  Previous work [G. Nagarjuna et al., J. Am. Chem. Soc., 136, 16309 (2014)] has shown that RAP solutions exhibit nearly ten-fold lower trans-separator permeability than their monomer counterparts. These results suggest that size-exclusion may be a viable means of reducing the crossover of electroactive species and concomitant capacity/efficiency losses in redox flow batteries (RFBs) when a separator is used in lieu of an ion-selective membrane.  Though low trans-separator permeability can be achieved with large RAP molecules, the viscosity of concentrated RAP solutions increases substantially with large RAPs.  It has been observed that RAPs exceeding 21 kg/mol have produced ten-fold higher viscosity than their monomer counterparts at the same concentration [G. Nagarjuna et al., J. Am. Chem. Soc., 136, 16309 (2014)]. Aqueous RAPs have also exhibited non-Newtonian rheology that is sensitive to the concentration of supporting salt [T. Janoschka et al., Nature, 527, 78 (2015) & T. Janoschka et al., Polym. Chem., 6, 7801 (2015)].  In addition to decreasing the flowability of RAP solutions, maintaining the mobility of the conductive ions needed for RFB power capability is likely to be a challenge for highly viscous electrolytes.  To assess the limits of RFB performance for concentrated RAP solutions, we are investigating ionic/RAP mobility and viscosity as a function of solution composition and RAP molecular weight (MW). 

Our preliminary modeling has shown that net coulombic-efficiency loss due to diffusive crossover scales in proportion to separator permeability and in inverse proportion to ionic conductivity.  Initial experiments have focused upon determining the ionic conductivity of viologen²⁺ RAPs (associated with PF₆^- anions) dissolved in acetonitrile with LiBF₄ as the supporting electrolyte.  The observed conductivity is a result of the mobility, concentration, and valence of charged species in solution (charged RAPs, PF₆^-, Li⁺, and BF₄^-).  Fig. 1(a) shows the ionic conductivity (measured via electrochemical impedance spectroscopy) as a function of LiBF₄ concentration. 0.64 mol/kg RAP solutions exhibit higher ionic conductivity than plain LiBF₄ solutions at all salt concentrations investigated.  This result suggests that the increased concentration of conductive species (charged RAPs and PF₆^-) has a more substantial effect on conductivity than any reduction in the mobility due to increased solution viscosity. Fig. 1(b) shows the variation of ionic conductivity with RAP concentration at a fixed salt concentration. Low MW RAP solutions (0.56kDa) show higher ionic conductivity than high MW RAP solutions (318kDa) in both the plots.  This result is likely due to the decreased mobility of charged RAPs with high MW.

Both of the RAP solutions investigated (with high and low MW, respectively) show ionic conductivity that is relatively insensitive to LiBF₄ concentration.  This result is important because it suggests that minimal amounts of supporting electrolyte are needed for a viologen-based cell to function. Also, once the RAP concentration is increased beyond 0.32 mol/kg, the ionic conductivity varies little with RAP concentration. This enables identification of an upper limit for active species concentration that optimizes cell performance and minimizes RAP viscosity. The currently measured samples contain viologen²⁺, and, consequently, ionic conductivity is likely to decrease when reduced to viologen¹⁺.  Therefore, on-going work will target measurement of ionic conductivity across a range of state-of-charge for the viologen-based RAPs.  Also, further testing will assess the contribution to ionic conductivity for less concentrated RAP solutions, where supporting electrolyte may be needed.

Measurement of trans-separator permeability must also be conducted to determine the composition of RAP solutions that minimizes crossover.  Furthermore, links will be established between ionic/RAP mobility and the viscosity of RAP solutions through systematic rheometric investigation under steady shear-rate conditions.  Modeling efforts will continue to link measured properties to system-level performance. Viscosity measurements will allow us to assess the flowability of RAPs by determining pumping pressure requirement through the use of rheological constitutive models.