Supporting Information

Transcription

1 Supporting Information Multielectron Cycling of a Low Potential Anolyte in Alkali Metal Electrolytes for Non-Aqueous Redox Flow Batteries Koen H. Hendriks, ab Christo S. Sevov, ab Monique E. Cook, ab and Melanie S. Sanford* ab a. Department of Chemistry, University of Michigan, 930 N University Ave., Ann Arbor, Michigan, 4809, United States. b. Joint Center for Energy Storage Research (JCESR). These authors contributed equally. Materials Compound + and were synthesized as described previously. S Tetrabutylammonium hexafluorophosphate ( 99%, electrochemical grade), potassium hexafluorophoshate ( 99%), Bis(trifluoromethane)sulfonimide lithium salt (99.95% trace metal basis) and Bis(trifluoromethane)sulfonimide potassium salt (97%) were obtained from Sigma Aldrich. Sodium hexafluorophosphate (99+%) was obtained from Alfa Aesar and was recrystallized from refluxing acetonitrile prior to use. Lithium hexafluorophosphate (battery grade) was obtained from Oakwood Chemical. All electrolyte salts were dried under high vacuum for 48 hours before being transferred to a nitrogen-filled glovebox. Acetonitrile (99.8%, anhydrous) was obtained from sigma Aldrich and used as received. M stock solutions of the supporting electrolytes in acetonitrile were prepared in a glovebox and dried over 3Å molecular sieves for at least two days prior to use. Cyclic Voltammetry Cyclic voltammetry was performed in a nitrogen-filled glovebox with a Biologic VSP multichannel potentiostat/galvanostat using a three electrode electrochemical cell, consisting of a glassy carbon disk working electrode (0.07 cm 2, BASi), a Ag/Ag + quasi-reference electrode (BASi) with 0.0 M AgBF 4 (Sigma) in acetonitrile, and a platinum wire counter electrode (23 cm, ALS). The glassy carbon disk electrode was polished in a nitrogenfilled glovebox using aluminum oxide polishing paper (9 micron and micron, Fiber Instrument) and anhydrous acetonitrile. All experiments were run at a scan rate of 00 mv/s in the stock electrolyte solutions containing 5 mm +. Ferrocene (-0 mm) was used as an internal voltage reference. Charge/discharge cycling Bulk electrolysis charge/discharge measurements were carried out in a nitrogen-filled glovebox with a BioLogic VSP galvanostat in a custom glass H-cell and reticulated vitreous carbon electrodes (00 ppi, Duocell). S A Ag/Ag + quasi-reference electrode (BASi) with 0.0 M AgBF 4 was used on the working side of the H-cell. A porous glass frit (P5, Adams and Chittenden) was used as the separator (Figure S6). The electrolyte contained mm active species and M supporting electrolyte. Both chambers of the H-cell were loaded with 5 ml electrolyte solution and were stirred continuously during cycling at a rate of 5 ma. After the first charge, the counter chamber of the cell was emptied and refilled with fresh electrolyte solution and RVC electrode after which cycling was resumed. Cycling under flow condition was performed with a zero-gap flow cell S2 comprised of graphite chargecollecting plates containing an interdigitated flow field in combination with two layers of non-woven carbon felt electrodes (Sicracet 29AA) on each side. PTFE gaskets were used to achieve ~20% compression of the felt. A Daramic 75 membrane separated the two half cells, and the exposed area of the membrane in the gasket window was used as the active area (2.55 cm 2 ). No pretreatment of the membrane or electrodes was performed. Two glass reservoirs were each filled with 5 ml, 00 mm solution of in M KPF 6 electrolyte and a peristaltic pump (Cole-Parmer) with Norprene and PFA tubing was used to circulate the electrolyte solutions at 0 ml/min through the cell. Galvanostatic charge/discharge cycling was performed using a BioLogic VSP galvanostat employing + V and V voltaic cut offs. EIS was performed from 500 khz to 0 mhz using a 0 mv sine perturbation. Polarization curves were measured by galvanostatically holding the cell at 0 50 ma and recording the voltage after a 20 s stabilization time. The cell

2 was brought back to 25% SOC after every measurement point with a 0 ma galvanostatic hold of opposite polarity. NMR spectra of electrolyte solutions were obtained on a Varian MR400 spectrometer using solvent suppression techniques. Chemical shifts are reported in parts per million (ppm) relative to TMS, with the residual solvent peak used as an internal reference. Additional Tables & Figures LiPF6 NaPF6 E vs Ag/Ag + KPF6 TBAPF Time (s) Figure S. Time vs voltage profiles of e - cycling of 5mM + in the various electrolytes, cycles 5 8 are shown. Coulombic efficiencies: TBAPF 6 = 96%, KPF 6 = 98%, NaPF 6 = 97%, LiPF 6 = 97%. E vs Ag/Ag KPF6 TBAPF Time (s) Figure S2. Time vs voltage profiles of 2e - cycling of 2.5 mm + in the various electrolytes, cycles 5 8 are shown. Coulombic efficiencies: TBAPF 6 = 99%, KPF 6 = 99%.

3 TBAPF 6 KPF 6 NaPF 6 LiPF Figure S3. Normalized discharge capacity versus cycle number of 4-benzoyl-N-methylpyridinium in the various electrolytes using 5 mm redox material, M of the respective supporting electrolyte and 5 ma charge/discharge rate. A similar decay behavior is observed compared to the electron cycling of compound + in Figure a. Undried LiPF 6 Dried LiPF 6 Undried LiPF vol% DMA Figure S4. Normalized discharge capacity versus cycle number of 5 mm 4-benzoyl-N-methylpyridinium at 5 ma charge/discharge rate using M LiPF 6 supporting electrolytes. The supporting electrolyte solution was either undried or dried over 3 Å molecular sieves prior to use. A PF 5 scavenger in the form of N,N-dimethylacetamide (DMA) was added to undried electrolyte solution as to enhance cycling stability. LiTFSI KTFSI LiPF 6 KPF Figure S5. Normalized discharge capacity versus cycle number of 5 mm + at 5 ma charge/discharge rate using M LiTFSI or KTFSI supporting electrolytes, compared to the cycling studies using LiPF 6 or KPF 6 also shown in Figure a.

4 Current (ma) Before cycling After Cycling E vs. Ag/Ag + (V) Figure S6. CV of electrolyte solution before and after bulk electrolysis of + in LiPF 6, showing an oxidation peak after cycling at + V that likely originates from a compound crossing over from the counter side of the H-cell. Figure S7. Photograph of flow setup showing the flow cell, reservoirs, pump and tubing (left). Photograph of H-cell setup (right) Im[Z] (W cm 2 ) Voltage (V) Re[Z] (W cm 2 ) Polarization Depolarization Current Density (ma/cm 2 ) Figure S8. EIS spectrum (left) and polarization curves (right) of the flow cell at 25% SOC.

5 Figure S9. Photograph of partially disassembled flow cell after cycling, showing some leakage of electrolyte solution between the white PTFE gaskets causing initial capacity loss. Before Cycling AcN TMB After Cycling TMB AcN ppm Figure S0. H NMR spectrum of electrolyte solution before and after flow-cycling with added trimethoxybenzene (TMB) as internal standard. Integration of the signal at 4.8 ppm originating from paramagnetic, relative to the standard, revealed that >95% of is retained after the experiment. The small diamagnetic signals at.4, 2.8, 4.4 and 7.5 ppm after cycling are consistent with +. S No other diamagnetic signals were detected. S Sevov, C. S.; Hickey, D. P.; Cook, M. E.; Robinson, S. G.; Barnett, S.; Minteer, S. D.; Sigman, M. S.; Sanford, M. S. J. Am. Chem. Soc. 207, 39, S2 Milshtein, J. D.; Kaur, A. P.; Casselman, M. D.; Kowalski, J. A.; Modekrutti, S.; Zhang, P. L.; Attanayake, N. H.; Elliott, C. F.; Parkin, S. R.; Risko, C.; Brushett, F. R.; Odom, S. A. Energy Environ. Sci., 206, 9, 353.