All-Soluble All-Iron Aqueous Redox-Flow Battery

Transcription

1 Supporting Information All-Solule All-Iron Aqueous Redox-Flow Battery Ke Gong [a], Fei Xu [a], Jonathan B. Grunewald [a], Xiaoya Ma [a], Yun Zhao [a], Shuang Gu* [], and Yushan Yan* [a] [a] Department of Chemical and Biomolecular Engineering, University of Delaware, 150 Academy Street, Newark, DE 19716, United States [] Department of Mechanical Engineering, Wichita State University, 1845 Fairmount St, Wichita, KS 67260, United States * 1

2 Experimental Methods 1. Cyclic voltammetry A three-electrode cell configuration was used for the cyclic voltammetry (CV) test. Glassy caron (dia. = 5 mm, Pine Instrument), platinum wire and a Ag/AgCl electrode were used as the working, counter, and reference electrodes, respectively. The cell was tested y a multichannel potentiostat (VMP2, Princeton Applied Research). The working electrode was polished to a mirror finish prior to the use in each set of CV test. 2. RFB cell assemly The all-solule all-fe RFB (Fig. S11) consists of two locks of electrolyte frames, two current conductors and collectors, and two Viton ruer gasket (0.4 mm thickness per piece). One piece of caron paper (2.2 cm 2.3 cm, 0.37 mm thickness Toray-H-120) was used for the negative electrode. Two pieces of caron paper (2.2 cm 2.3 cm, 0.19 mm thickness for one caron paper, Toray-H-60; 0.37 mm thickness for the other, Toray-H-120) were used for the positive electrode. The purpose of using different caron papers on each side is to keep the total thickness of caron paper only slightly larger than the total gasket thickness to ensure good sealing and firm contact etween caron paper and current collector. A piece of Nafion212 memrane was used as the cation-exchange memrane in the cell. The Nafion212 memrane was soaked in 1 M NaCl solution for 48 hours and was thoroughly washed with DI water to remove residual NaCl solution from the surface efore use. Eight olts were torqued to 16 l ft to tighten the cell and provide firm compression etween the electrode and the current collector. The electrolyte was pumped y a peristaltic pump (Cole Parmer, Masterflex L/S 600 rpm) through PTFE-lined ruer tues (Cole Parmer, ChemDurance #16). A flow rate of 100 ml min 1 was used in all tests. The all-solule all-fe RFB was tested y a commercial attery test station (Arin, BT2000). All tests were conducted at room temperature. 3. TEOA, [Fe(TEOA)OH], [Fe(TEOA)(OH)] 2, Fe(CN)6 3 and Fe(CN)6 4 crossover test 2

3 The RFB cell was set up with 20 cm 2 memrane exposure, an enriched side filled with solution with the test species, and the deficient side filled with a salt solution without the test species. For TEOA, [Fe(TEOA)OH], and [Fe(TEOA)(OH)] 2 crossover tests, the enriched solution contained 0.2 M FeCl3 or FeCl2, 1 M TEOA, and 1.5 M NaOH (30 ml). 1 M NaCl solution (15 ml) was used on the deficient side to minimize the difference in osmotic pressure and volume change of electrolyte. Every hour, an amount of 300 ul electrolyte from the deficient side was sampled and diluted with 400 µl deuterated water (D2O) for 13 C NMR analysis (Bruker AV-400), and an amount of 600 µl electrolyte was sampled and diluted to 3 ml with 2315 µl 1 M hydrochloric acid and 85 µl 70% nitric acid for ICP-OES test (Optima 7300DV). The same cell setup was used for the Fe(CN)6 3 or Fe(CN)6 4 crossover tests. 30 ml 0.2 M Na4Fe(CN)6 in 0.5 M NaOH or 0.2 M K3Fe(CN)6 in 0.5 M KOH were used on the enriched side. 0.5 M NaCl or KCl were used on the deficient side. The deficient-side solution was sampled and measured y the same method as for the 13 C NMR analysis. The 13 C NMR spectra of coordinated TEOA and free TEOA are indistinguishale, therefore the total TEOA crossover rate for oth [Fe(TEOA)OH] and [Fe(TEOA)(OH)] 2 was measured y 13 C NMR test. The crossover rates of [Fe(TEOA)OH] and [Fe(TEOA)(OH)] 2 were otained y measuring the iron element crossover with ICP-OES. The crossover rate of free TEOA in each case can then e calculated y sutracting coordinated TEOA from the total TEOA. The concentration of respective crossover species is shown in Fig. S5. It can e derived that the concentration of crossover species follows Eq. 5. V B dcb( t) P A ( CA CB) Eq. 5 dt L Where CB is the concentration of test species in the deficient side, VB is the volume of the deficient side, A is memrane area, L is memrane thickness, and CA is the concentration of test species in the enriched side. 3

4 Given the experimental conditions, VB and CA can e assumed as constant. Since CB is negligile compared to CA, the driving force ΔC (i.e., CA CB) can also e treated as constant at CA. The permeation coefficient of oth free TEOA and Fe-TEOA complex can then e otained y a linear fitting of CB vs. time for total TEOA, coordinated TEOA, respectively. 4. TEOA degradation study in the presence of Fe(CN)6 3 The chemical reaction etween Fe(CN)6 3 and TEOA was studied y analyzing the 1 H NMR spectra of deuterated water solutions containing 0.2 M TEOA, 0.5 M K3Fe(CN)6, and 0.5 M KOH. The electrolyte was sampled at different times and measured y 1 H NMR spectroscopy using the same solvent and dilution ratio as in the TEOA crossover test. The 1 H NMR spectra for predicted degradation products were otained y MestReNova 6.1. The rate of reaction etween Fe(CN)6 3 and TEOA was also studied y CV. It can e derived that a typical CV curve for a catalytic reaction shows a limiting value of current i, given y Eq. 6 1 i nfac Eq. 6 * * 1/ 2 3 C ( Dk) Fe( CN ) 6 TEOA Where n is the numer of electrons transferred in electrochemical reaction, F is Faraday s constant, C is the concentration of Fe(CN)6 3, * 3 Fe( CN ) 6 * C TEOA is the concentration of TEOA, D is the diffusion coefficient of TEOA 2, k is the kinetic constant of reaction etween TEOA and Fe(CN)6 3. The reaction rate constant can therefore e calculated y a linear fitting of TEOA concentration as a function of limiting current. CV studies at different TEOA concentrations and the linear fitting of TEOA concentration with limiting current are shown in Fig. S7. Since TEOA will further undergo direct electrochemical reaction at higher potentials, the limiting current was chosen using the plateau region at 0.4 V vs. Ag/AgCl. 5. Impedance test Impedance for the all-solule all-fe full RFB cell was measured at 50% SOC. To study the impact of TEOA on the Nafion memrane, RFB cell was set up with three solution concentrations: a) 3 M NaOH 4

5 for oth sides; ) 3 M NaOH and 2 M TEOA for oth sides; and c) 3 M NaOH, 2 M TEOA and 0.4 M FeCl3 for oth sides. The impedance was measured y a potentiostat (Solatron A1287) and impedance analyzer (Solatron A1260). 6. Cycle test RFB cell was set up with the negative electrolyte y the solution containing 0.2 M FeCl3, 1 M TEOA, and 1.5 M NaOH and the positive electrolyte y the solution containing 0.2 M Na4Fe(CN)6 and 3 M NaOH. These concentrations were chosen to keep similar osmotic pressure of the electrolytes and thus to minimize water transport during long-term cycle test. The current density was set at 40 ma cm 2 with cut-off voltages of 0.5 V (discharge) and 1.6 V (charge). 7. Polarization test The RFB cell was set up with the same electrolytes as in cycle test. The attery was first charged to 70% SOC; then, alternating charge and discharge current was applied. Cell voltage was recorded at each current density. Reference (1). Allen J. Bard, L. R. F. Electrochemical Methods: Fundamentals and Applications; Wiley: New York; (2). Chang, L. C.; Lin, T.; Li, M. H. Mutual Diffusion Coefficients of Some Aqueous Alkanolamines Solutions. J Chem Eng Data 2005, 50,

6 Tale S1. Electrode potentials and standard rate constants of iron(iii)/iron(ii)-ligand redox pairs with different phs Ligand Areviation φ vs. SHE [a] k 0 (cm/s) [] ph of test condition Tripyridinetriazine tpt (from SCE) 0 5-Nitro-o-phenanthroline np (from SCE) 0 5-Chloro-o-phenanthroline clp (from SCE) 0 Bipyridine py (from SCE) o-phenanthroline phen (from SCE) ,9-Dimethyl-o-phenanthroline 2,9-dmp (from SCE) 0 Terpyridine tp (from SCE) Cyanopyridine cp (from SCE) 0 4-methy-o-phenanthroline mp (from SCE) ,7-Dimethyl-o-phenanthroline 4,7-dmp (from SCE) Pyridinecaroxaldehyde p-cph (from SCE) 0 malic acid (from Ag/AgCl) xylitol (from Ag/AgCl) glycerol (from Ag/AgCl) Dimethylsulfoxide dmso (from Ag/AgCl) malonic acid (from Ag/AgCl) glycine (from Ag/AgCl) Cyanide [c] CN (from SCE) Nitrilotriacetic acid nta (from SCE) Diethylenetriaminepentaacetic acid dtpa (from SCE) ethylenediaminetetraacetic acid edta (from SCE) oxalate ox (from SCE) 2.2 citrate (from Ag/AgCl) Triethanolamine [c] TEOA (from Ag/AgCl) [a] φ : formal redox potential. Potential was converted to standard hydrogen potential (SHE) y relationships: SCE = V vs. SHE, Ag/AgCl = V vs. SHE. [] k 0 : standard rate constant of redox reaction. k 0 was either taken directly from literature or converted from exchange current density reported in literature. [c] This work. 6

7 a Figure S1. (a) Cyclic voltammetry of [Fe(TEOA)OH] /[Fe(TEOA)OH] 2 at different scan rates on glassy caron electrode. The electrolyte contains 0.2 M Na[Fe(TEOA)OH] and 0.2 M Na2[Fe(TEOA)OH] prepared y mixing 0.2 M FeCl2, 0.2 FeCl3, 2 M TEOA, and 3 M NaOH. () Cyclic voltammetry of Fe(CN)6 3 /Fe(CN)6 4 at different scan rates on glassy caron electrode. The electrolyte contains 0.1 M Na4Fe(CN)6, 0.1 M K3Fe(CN)6, and 3 M NaOH. Inset: linear fitting of peak current density versus square root of scan rate. a Figure S2. (a) Cell voltage curve of a charge-discharge test of all-solule all-fe RFB at 40 ma/cm 2, and () Zoomed-in cell voltage, positive and negative potential vs. Ag/AgCl during the end of the discharge process, as indicated in the green rectangle in Figure S2 (a). The decrease of positive potential indicates the positive side is the dominating cause for the cell voltage drop and therefore responsile for the coulomic efficiency loss. 7

8 Figure S3. Cell voltage curve of Fe-CN symmetrical cell. The CE remains at 100%, which excludes the possiilities of oth oxygen evolution of side reaction and the crossover of Fe(CN)6 3 species. a Figure S4. 13 C NMR spectra of electrolytes efore and after cycle test. (a) Negative electrolyte and () Positive electrolyte. A numer of new peaks showed up in positive electrolyte after 20 cycles, as enclosed in red circle, whereas no new peak was oserved for negative electrolyte after test. 8

9 a Figure S5. 13 C NMR spectra of electrolytes of the deficient side of a permeation cell with the enriched side eing (a) Fe(CN)6 4 and () Fe(CN)6 3. No sign of Fe(CN)6 4 or Fe(CN)6 3 was detected over five days. The lowest concentration that could e detected y NMR spectroscopy in this test is 0.01 M. The permeation coefficient was then estimated to e less than cm 2 /s, y assuming the concentration after 5-day test is elow 0.01 M. 9

10 a c d Figure S6. 13 C NMR spectra of electrolyte of the deficient side of a permeation cell with the enriched side eing (a) [Fe(TEOA)(OH)] 2 and () [Fe(TEOA)(OH)]. The concentration of TEOA is otained y comparing the peak integral with those in spectra of standard TEOA solution having known TEOA concentrations. (c) Linear fitting of total TEOA concentration and [Fe(TEOA)(OH)] 2 concentration vs. test time. (d) Linear fitting of total TEOA concentration and [Fe(TEOA)(OH)] concentration vs. test time. 10

11 a Figure S7. (a) CV scan on glassy caron electrode at 5 mv/s from 0 V to 0.5 V vs. Ag/AgCl of 0.01 M Na4Fe(CN)6 (lue curve), 0.2 M TEOA (green curve), 0.01 M Na4Fe(CN) M TEOA (red curve). All electrolytes are supported with 1 M NaOH. The rising current at Ew > 0.4 V vs. Ag/AgCl confirms the electrochemical oxidation of TEOA on electrode surface (green curve). () CV scan on glassy caron electrode at 5 mv/s from 0 to 0.5 V vs. Ag/AgCl of 0.01 M Na4Fe(CN)6 with different TEOA concentrations from 0.2 M to 0.6 M. The current density at 0.4 V vs. Ag/AgCl of forward scan was chosen as limiting current density. Inset: linear fitting of limiting current density vs. TEOA concentration. 11

12 a Figure S8. (a) 1 H NMR spectra of solutions containing 0.5 M K3Fe(CN)6, 0.2 M TEOA and 0.5 M KOH with test time. Inset: zoomed-in figure of 1 H NMR etween chemical shifts of 4.0 and 3.2 ppm. Peaks a and are attriuted to α and β hydrogen in TEOA. Peaks c, d and e are possily due to the oxidation at α-caron and the formation of 1-(is(2-hydroxyethyl)amino)ethane-1,2-diol and 2-hydroxy- N,N-is(2-hydroxyethyl)acetamide. () Predicted 1 H NMR spectrum of 1-(is(2- hydroxyethyl)amino)ethane-1,2-diol (upper) and 2-hydroxy-N,N-is(2-hydroxyethyl)acetamide (lower). It is possile that the oxidation takes place at α-caron and generates 1-(is(2- hydroxyethyl)amino)ethane-1,2-diol. Diol can e further oxidized y Fe(CN)6 3 and produces 2- hydroxy-n,n-is(2-hydroxyethyl)acetamide. The shape and relative peak position of the otained degradation product match the predictions for the proposed product. 12

13 a Figure S9. (a) Impedance of all-solule all-fe RFB showing ohmic resistance (Rohm, depicted in lue color) and charge transfer resistance (Rc, red). The ohmic resistance accounts for the major part of total resistance. () Experimental (dot) and predicted ir-free (dash) polarization curve of all-solule all-fe RFB at 70% State of Charge (SOC). The ir loss is indicated in the orange-shaded area. Figure S10. Impedance test of a symmetrical cell with three different electrolytes: 3 M NaOH (dark lue), 2 M TEOA in 3 M NaOH (lue), and 0.4 M FeCl3 + 2 M TEOA in 3 M NaOH (green). The area specific resistance (ASR) difference shows that the contamination of TEOA and Fe-TEOA to Nafion memrane is responsile for large internal resistance. 13

14 Figure S11. Diagram of cell configuration for all-solule all-fe RFB. The electrolyte is flowing into the system through the side inlet of electrolyte frame. The positive and negative electrolytes flow through the small holes on electrolyte frame, conductor, and current collector into the chamer encompassed y gasket and memrane. N and P in parentheses stand for negative and positive, respectively. 14