TECHNICAL FIELD
This invention relates to metal halide oxyanion electrodes and batteries including the electrodes.
BACKGROUND
The improvement of the positive electrode remains a significant challenge for improving the gravimetric and volumetric energy density of batteries. The current state-of-art positive electrodes are based on the intercalation of lithium ions into and out of transition metal oxides during discharge and charge, respectively. Efforts to improve lithium ion positive electrode materials have been based on trying to increase the amount of mobile lithium per given amount of stationary transition metal oxide. This approach is proving difficult as removing more and more of the cations in the structure during charge results in a structure which is weakly held together this can lead to the irreversible release of oxygen gas from the lattice and a subsequent loss of the electrode's capacity. This requirement to maintain a stable structure in both the fully lithiated and delithiated states imposes some form of upper limit on the maximum achievable capacity for conventional intercalation based electrode materials (although how close we are to this fundamental limit is unclear).
SUMMARY
In one aspect, an electrode can include a halogen oxyanion salt and a conductive material.
In another aspect, a battery can include a metal electrode, a halogen oxyanion electrode, and a separator between the metal electrode and the halogen oxyanion electrode. In certain circumstances, the halogen oxyanion electrode can include a halogen oxyanion salt and a conductive material.
In certain circumstances, the halogen can be chlorine, bromine or iodine. For example, the halogen can be iodine.
In certain circumstances, the halogen oxyanion salt can be an alkali metal salt. For example, the alkali metal salt can be a lithium salt, a sodium salt or a potassium salt. In certain circumstances, the halogen oxyanion salt can be a lithium iodate, a sodium iodate or a potassium iodate.
In certain circumstances, the halogen oxyanion salt can be formed by oxidation of a metal hydroxide salt in the presence of a halogen or halide. For example, the halogen oxyanion salt can be formed by oxidation of a metal hydroxide salt by a halogen, such as iodine, or a halide, such as iodide.
In certain circumstances, the conductive material can be a conductive carbon material. For example, the conductive carbon material can include carbon black, graphene, carbon nanotubes, or graphite.
In certain circumstances, the electrode can further include a binder.
In certain circumstances, the halogen oxyanion can be iodate.
In certain circumstances, the metal electrode can include an alkali metal or metal ion negative electrode. For example, the alkali metal can include lithium, sodium or potassium. In specific examples, the metal electrode can include lithium. Alternatively, the metal ion negative electrode can be lithiated graphite or silicon.
In another aspect, a method of generating electricity can include creating an electronic connection to a battery described herein.
Other aspects, embodiments, and features will be apparent from the following description, the drawings, and the claims.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1A is a schematic depiction of an alkali metal halogen oxide electrochemical system.
FIG. 1B is a graph depicting cyclic voltammograms of solutions of 0.5M LiTFSI+10 mM LiI collected at 100 mVps under argon environment in each of the considered solvents with a Pt working electrode, either Li metal (DME, DMSO) or lithium titanium oxide (DMA) counter electrode and Ag/Ag+ reference electrode. Currents were normalized based on the maximum current observed. Potentials were converted to a Me10Fc scale based on its half-wave potential measured at the end of the experiment by adding 2 mM Me10Fc to the electrolyte (as detailed in FIG. 18).
FIG. 2 is a set of graphs depicting (panel a) Color changes when solutions of 50 mM I3− (0.2 M LiI+50 mM 12) are added to 0.1 M synthetic Li2O2 (panel b) UV-vis spectra of the liquid phase before and after the reaction with Li2O2 confirm the consumption of I3− in DMSO, but that I3− remains in DME (panel c) The change in concentration of I3− when adding a 50 mM I3− solution to a two times excess of synthetic Li2O2 (left axis, black filled symbols). I3− concentrations were determined through UV-Vis Spectroscopy. Full consumption of I3− was found in DMA, DMSO and Me-Im with differences in the plot stemming from different initial concentrations. Error bars were estimated based on the accuracy of the mass balance used during preparation of diluted samples of +/−0.5 mg. Calibration curves for each solvent can be found in FIG. 12-14. The difference between the I3−/I− and Li+, O2/Li2O2 redox potentials (right axis, open grey symbols).
FIG. 3 is a graph depicting gas chromatography of the gaseous products during the reaction between 50 mM I3− (0.2 M LiI+50 mM 12) and synthetic Li2O2. (panel a) the change in O2 concentration in the Argon carrier gas stream at 2, 22, 42 and 62 minutes following the injection of the I3− solution; (panel b) The GC sensor outputs for each of the four measurements. N2 and O2 signals were calibrated using a 2500 ppm O2+17000 ppm N2 in Argon gas mixture. The absence of N2 indicates O2 came from the reaction and not a leak in the cell.
FIG. 4 is a graph depicting a voltage profile and corresponding O2 (filled symbols) and CO2 (open symbols) evolution during charge at 0.1 mA/cm2 in 0.5 M LiTFSI in G2 (panel a, panel c) or DMSO (panel b, panel d), both with 0.1 M LiI (lighter colors) and with an additional 0.1 M LiTFSI (darker colors) to keep the overall [Li+] constant. Cells constructed with a Li metal counter electrode and a solid Li-conducting separator to prevent shuttling of oxidized iodide species from the positive electrode to the Li metal electrode where they can be chemically reduced and diffuse back to the positive electrode. See, for example, Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett 1, 747-756 (2016), which is incorporated by reference in its entirety. Potentials referenced against lithium metal counter electrode. Cells were first discharged for 20 hours at 0.05 mA/cm2 under O2 environment, and then the cell headspace was evacuated and purged with Argon gas five times. The added 0.1 M LiI could provide a theoretical maximum of 33 mM I3− and 50 mM I2, accounting for a maximum of 0.25 mAhr/cm2 of capacity.
FIG. 5A is a set of drawings showing a graph depicting (panel a) The change in concentration of I3− when adding a 50 mM I3− (0.2 M LiI+50 mM I2) solution to a two times excess of LiOH (left axis, black filled symbols). I3− concentrations were determined through UV-Vis Spectroscopy. Full consumption of I3− was found in DMA, DMSO and Me-Im with differences in the plot stemming from different initial concentrations. Error bars were estimated based on the accuracy of the mass balance used during preparation of diluted samples of +/−0.5 mg. Calibration curves for each solvent can be found in FIGS. 12-14. The difference between the I3−/I− and Li+,O2,H2O/LiOH redox potentials (right axis, open grey symbols) and Li+,IO3−,H2O/LiOH,I− redox potentials (right axis, filled grey symbols); (panel b) The change in concentration of 12 when adding a 50 mM I2 solution to a two times excess of LiOH (left axis, black filled symbols). The difference between the I2/I3− and Li+, O2, H2O/LiOH redox potentials (right axis, open grey symbols) and Li+,IO3−, H2O/LiOH, I− redox potentials (right axis, filled grey symbols; (panel c) Raman spectra of the solid participate which was separated and washed after reacting an excess of 12 with LiOH in DMSO and three reference spectra (LiIO3, LiOH and LiOH—H2O). The solid precipitate has only peaks consistent with LiIO3 and no erroneous peaks; measurement is representative of three separate locations in the solid.
FIG. 5B is a set of drawings showing a graph (panel a) showing the consumption of I3− when adding 200 μmol of commercial LiOH to 1 mL of 50 mM I3− solution (50 μmol I3−, LiOH:I3−=4:1)) (left axis, black filled symbols) measured after 48 hours. I3− concentrations were determine through UV-Vis Spectroscopy. Full consumption of I3− was found in DMA, DMSO and Me-Im with differences in the plot stemming from different initial concentrations. Error bars were estimated based on the accuracy of the mass balance used during preparation of diluted samples of +/−0.5 mg. Calibration curves for each solvent can be found in FIGS. 12-14. The difference between the I3−/I− and Li+,O2,H2O/LiOH redox potentials (right axis, open grey symbols) and Li+,IO3−,H2O/LiOH,I− redox potentials (right axis, open black symbols). Panel b shows the consumption of 12 when adding 200 μmol of commercial LiOH to 1 mL of 50 mM I2 solution (50 μmol 12, LiOH:I3−=4:1)) (left axis, black filled symbols) measured after 48 hours. The difference between the I2/I3− and Li+,O2,H2O/LiOH redox potentials (right axis, open grey symbols) and Li+,IO3−,H2O/LiOH,I− redox potentials (right axis, open black symbols). Panel c shows Raman spectra of the solid participate which was separated and washed after reacting an excess of I2 with LiOH in DMSO and three reference spectra (LiIO3, LiOH and LiOH—H2O). The solid precipitate has only peaks consistent with LiIO3 and no erroneous peaks; measurement is representative of three separate locations in the solid.
FIG. 6 is a graph depicting (panel a)1H NMR spectrum of pure DMSO, DMSO after exposure to LiOH, and DMSO after the reaction between 50 mM I2/I3− and 0.2 M commercial LiOH. After the reactions between I2/I3− and LiOH, two new peaks appear; one at ˜2.95 ppm (based on the DMSO peak being assigned to 2.5 ppm) corresponding to DMSO2 and one at ˜3.3 ppm corresponding to H2O. All 1H NMR samples were prepared by mixing 0.5 mL of the sample+0.1 mL of DMSO-D6 (for NMR locking)+10 μL of 1,4-dioxane internal reference (for quantification) and (panel b) full quantification of detected liquid and solid products after reactions between I2/I3− and LiOH. LiIO3 was quantified using iodometric titration, DMSO2 and H2O were quantified using 1H NMR with an internal standard of 1,4-dioxane.
FIG. 7 is a graph depicting voltage profile and corresponding O2 (filled symbols) and CO2 (open symbols) evolution during charge of LiOH preloaded electrodes at 0.1 mA/cm2 in 0.5 M LiTFSI in G2 (panel a, panel c) or DMSO (panel b, panel d), both with 0.1 M LiI (lighter colors) and with an additional 0.1 M LiTFSI (darker colors) to keep the overall [Li+] constant. Cells constructed with a Li metal counter electrode and a solid Li-conducting separator to prevent shuttling of oxidized iodide species from the positive electrode to the Li metal electrode where they can be chemically reduced and diffuse back to the positive electrode. See, for example, Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016), which is incorporated by reference in its entirety. Potentials referenced against lithium metal counter electrode. The added 0.1 M LiI could provide a theoretical maximum of 33 mM I3− and 50 mM I2, accounting for a maximum of 0.25 mAhr/cm2 of capacity.
FIG. 8 is a drawing depicting a battery.
FIG. 9 is a graph depicting Raman spectroscopy of commercial Li2O2, anhydrous LiOH after additional drying at 170° C. under vacuum for 24 hrs, LiOH—H2O and LiIO3.
FIG. 10 is a graph depicting cyclic voltammograms of solutions of 0.5M LiTFSI+10 mM LiI collected at 100 mVps under argon environment in each of the considered solvents with a Pt working electrode, either Li metal (G4, DME, DMSO) or lithium titanium oxide (pyridine, DMA, Me-Im) counter electrode and Ag/Ag+ reference electrode. Currents were normalized based on the maximum current observed. Potentials were converted to a Me10Fc scale based on its half-wave potential measured at the end of the experiment by adding 2 mM Me10Fc to the electrolyte (as per FIG. 18).
FIG. 11 is a graph depicting calibration curves relating UV-vis absorbance at (panel a) 364 nm and (panel b) 293 nm to the concentration of I3− in prepared DME solutions.
FIG. 12 is a graph depicting calibration curves relating UV-vis absorbance at 366 nm and 297 nm to the concentration of I3− in prepared DMSO solutions.
FIG. 13 is a graph depicting calibration curves relating UV-vis absorbance at 367 nm and 296 nm to the concentration of I3− in prepared G4 solutions.
FIG. 14 is a graph depicting a calibration curve relating UV-vis absorbance at 368 nm to the concentration of I3− in prepared pyridine solution.
FIG. 15 is a graph depicting UV-vis absorbance of pure solvents, with I3− peaks marked. The solvent's inherent absorbance interferes with the observation of 293 nm I3− peak in pyridine and Me-Im.
FIG. 16 is a graph depicting XRD reference spectra of commercial Li2O2, LiOH, LiOH—H2O, LiI, LiIO3 and DMSO2.
FIG. 17 is a photograph depicting observed corrosion of 316 stainless steel current collector following cycling of a cell with 0.1M LiI in DMSO.
FIG. 18 is a graph depicting details on how the Ag+/Ag reference electrode scale was converted to a Me10Fc scale. All measurements were made on the potentiostat against the Ag+/Ag reference electrode. The Me10Fc half-wave potential was obtained by adding 2 mM Me10Fc to the electrolyte at the end of the experiment and measuring its CV at 100 mVps. The half-wave potential of Me10Fc was determined based on taking the average potential of the anodic and cathodic peaks. The position of this Me10Fc potential on the Ag+/Ag scale was then used to determine the positions of other redox transitions on the Me10Fc scale.
FIG. 19 is a graph depicting possible correlations between solvent DN and measured half-wave potentials for I−/I3− (squares), I3−/I2 (Xs) and Li/Li+ (open circles).
FIG. 20 is a graph depicting possible correlations between solvent AN and measured half-wave potentials for I−/I3− (squares), I3−/I2 (Xs) and Li/Li+ (open circles).
FIG. 21 is a graph depicting possible correlations between solvent dielectric constant and measured half-wave potentials for I−/I3− (squares), I3−/I2 (Xs) and Li/Li+ (open circles).
FIG. 22 is a graph depicting: (panel a) color changes when solutions of 50 mM I3− are added to 0.1 M synthetic Li2O2; (panel b) UV-vis spectra of the liquid phase before and after the reaction with Li2O2 confirm the consumption of I3−; and (panel c) the concentrations of I3− before and after adding the solution to a two times excess of Li2O2. I3− concentrations were determined through UV-Vis Spectroscopy. Error bars were estimated based on the accuracy of the mass balance used during preparation of diluted samples of +/−0.5 mg. Calibration curves for each solvent can be found in FIG. 11-14.
FIG. 23 is a graph depicting full, unscaled adsorption spectra of the liquid phase retrieved after the reaction between I3− and Li2O2 in DMA, DMSO and Me-Im.
FIG. 24 is a graph depicting Raman spectra of commercial Li2O2 and LiTFSI for reference, as well as the solid recovered after the reaction between synthesized Li2O2 and I3− in G4, DME, Pyridine, DMA, DMSO and Me-Im.
FIG. 25 is a graph depicting details of GC experiments measuring the O2 generation due to the reaction between I3− and Li2O2 in DMSO.
FIG. 26 is a graph depicting the concentration of I2 present before and after the reaction between I2 and Li2O2 in DME, DMA and DMSO.
FIG. 27 is a graph depicting measured extent of reaction determined by UV-Vis of solutions of I2 in DME with different starting concentrations of LiI with both commercial Li2O2 and Li2O2 formed through disproportionation. Calculated values based on the reaction stopping when only I3− remains are shown as dashed gray lines.
FIG. 28 is a graph depicting Raman of solutions of mixtures of I2 and LiI in DME taken before and after reaction with Li2O2.
FIG. 29 is a graph depicting 1H NMR spectra of pure solvent as well as the liquid phase recovered following the reaction with Li2O2 and LiOH. All 1H NMR samples were prepared by mixing 0.5 mL of the sample+0.1 mL of DMSO-D6 (for NMR locking)+10 μL of MeCN internal reference (for quantification).
FIG. 30 is a graph depicting reference 1H NMR spectra of commercial dimethyl sulfone (DMSO2) added to DMSO-D6 with contaminate water.
FIG. 31 is a graph depicting 1HNMR spectra showing the change in proton exchange dynamics of Me-Im after creating a solution with 50 mM I2 and 0.2M LiI.
FIG. 32 is a graph depicting iodination of Me-Im led to loss of the I3− peak in UV-vis.
FIG. 33 is a graph depicting XRD of electrodes discharged in 0.1M LiI+0.5M LiTFSI solutions in G2 and DMSO. G2 electrode was discharged for 20 hours at 0.05 mA/cm2 which corresponded to its total capacity before sudden death, DMSO electrode was discharged for 40 hours at 0.05 mA/cm2 to allow more easy identification of the discharge product.
FIG. 34 is a graph depicting I3− concentration of solutions of 50 mM I3− (0.2 M LiI+50 mM I2) in a range of solvents before and after reaction with 0.2 M LiOH.
FIG. 35 is a graph depicting UV-vis spectra of solutions of I3− in DME (left) and DMSO (right) before and after reaction with 0.2 M LiOH, confirming the consumption of I3− in DMSO, but that I3− remains in DME.
FIG. 36 is a graph depicting Raman spectra of LiOH and LiOH—H2O powder compared with the solid recovered after the reaction between I3− and a two times excess of LiOH in G4, DME, pyridine, DMA, DMSO and Me-Im.
FIG. 37 is a graph depicting GC measurements during the reaction between LiOH and I3− in DMSO show no detectable quantity of O2 generated.
FIG. 38 is a graph depicting XRD of solid recovered after the reaction between LiOH and a two times excess of I3− in DMSO.
FIG. 39 is a graph depicting Raman spectra collected before and during the reaction between I3− and LiOH in DMSO. Measurements were taken by directly measuring the solution phase of a 50 mM I3− DMSO solution in a vial either with (red) or without (blue) LiOH present (during the initial stages of the reaction.
FIG. 40 is a graph depicting XRD patterns of preloaded electrodes following charging in G2 and DMSO compared with references for LiOH, LiOH—H2O and LiIO3. Electrodes show only peaks present on XRD taken on the pristine carbon paper (CP) and LiOH.
FIG. 41 is a graph depicting (panel a) Calculated thermodynamics of the oxidation of Li2O2(black) and LiOH(blue) to O2 (solid) and LiIO3 (dotted). These values are overlayed with the measured half-wave potentials of the I−/I3− and I3−/I2 redox couples in DME, DMA and DMSO. (panel b) Predicted selectivity towards O2 and LiIO3 formation based on the minimum O—O distance in the crystal lattice.
FIG. 42 is a graph depicting DEMS from cells with and without KO2 in between glass fiber separators in 0.1M KI in G2. I3−/I2 is formed at the positive carbon paper electrode can diffuse towards the negative electrode where it can be chemically reduced by the K metal plated onto the Cu film creating shuttling between the electrodes as has been shown previously. If KO2 is present between the separators (electronically isolated from both electrodes), it can also be oxidized by the I3−/I2 in the electrolyte and give off O2 gas and enhance capacity.
FIG. 43 is a graph depicting solid phase after the reaction between Li2O and I3− in DMSO shows clear evidence of LiIO3 through Raman (top) and XRD (bottom).
FIG. 44 is a graph depicting comparison of the color of 12 solutions in hexane (left) and DME (right). The purple color indicates the absence of I3− and therefore no Solvent-I+ complexes.
FIG. 45 is a graph depicting the liquid phase (left) and solid phase (right) following the reaction between hexane and commercial Li2O2.
FIG. 46 is a graph depicting Raman spectra of the solid recovered after the reaction between Li2O2 and I2 in hexane. Peaks are consistent with Li2O2 and solid LiI3 (which may have a slightly shifted peak compared with I3− in solution.
FIG. 47 is a graph depicting Raman spectra of the solution recovered after the reaction between I2 and LiOH in DME compared to reference spectra for solutions of I3− and I2 in DME. Spectra show on I3− remains after the reaction.
FIG. 48 is a graph depicting Raman spectra of LiOH synthesized via the disproportionation of KO2 in a two times excess of LiTFSI in MeCN with added water. Spectra indicates the anhydrous phase of LiOH was formed.
FIG. 49 is an image of vials following the reaction between 200 μmol of synthetic LiOH and 1 mL of 50 mM I3− solution (50 μmol I3−, LiOH:I3−=4:1)) after 96 hours. Similar to the commercial LiOH, the DMSO solution became colorless after ˜1 hour, the DMA solution became colorless after ˜96 hours and the DME solution did not change color.
FIG. 50 is a graph depicting XRD and SEM of the commercial LiIO3 used to construct LiIO3 battery electrodes.
FIG. 51 is a drawing depicting a schematic of cell used to test discharge process.
FIG. 52 is a graph depicting sample discharge profile (discharged at C/40) of composite electrodes constructed with commercial LiIO3, carbon and a PvDF binder.
FIG. 53 is a graph depicting sample discharge profile (discharged at 0.1 mA/cm2) of electrodes prepared by drop casting LiIO3 and Vulcan carbon onto carbon paper substrate.
FIG. 54 is a graph depicting Raman spectra on electrodes after discharge show the clear formation of anhydrous LiOH.
FIG. 55 is a graph depicting XRD on electrodes after discharge show the clear formation of anhydrous LiOH.
FIG. 56 is a micrograph depicting SEM of a pristine composite electrode made from LiIO3, carbon and PvDF binder.
FIG. 57 is a micrograph depicting SEM of a discharged composite electrode made from LiIO3 shows clear morphological changes.
FIG. 58 is a graph depicting discharge profile (bottom) and results of titrations to quantify the amount of formed LiOH and LiIO3 (top) for cells discharged in different solvents with O0V % added H2O.
FIG. 59 is a. graph depicting (left) XRD and (right) Raman characterization on the discharged electrodes demonstrate the consumption of LiIO3 and formation of LiOH on discharge
FIG. 60 is a graph depicting discharge profile (bottom) and results of titrations to quantify the amount of formed LiOH and LiIO3 (top) for cells discharged in DME with different amounts of added H2O.
FIG. 61 is a graph depicting (left) XRD and (right) Raman characterization on the discharged electrodes demonstrate the consumption of LiIO3 and formation of LiOH on discharge.
FIG. 62 is a set of micrographs depicting SEM images of pristine (left) and discharged (right) electrodes demonstrate substantial morphological changes during discharge, consistent with dissolution of reactants and precipitation of reaction products.
FIG. 63 is a graph depicting solubilities of LiOH and LiIO3 measured with inductively coupled plasma (ICP) indicate solubility in the mM region with added H2O in the electrolyte. Viscodensity measurements indicate high viscosity and a linear relationship between the volume of added water and its resulting concentration in the mixed electrolyte.
FIG. 64 is a graph depicting overpotentials during discharge are consistent with mass transport limitations as evidenced by the linear relationship between overpotential and log(viscosity/LiIO3 solubility).
FIG. 65 is a graph depicting discharge process is proposed to go through a three step process of 1) dissolution of LiIO3 2) electrochemical reduction of IO3− to OH− and 3) precipitation of LiOH.
DETAILED DESCRIPTION
Lithium oxygen batteries (a potential alternate positive electrode chemistry) work by reacting oxygen in its gaseous form with lithium ions to form lithium peroxide on discharge and then reforming oxygen gas and lithium ions on charge. The lithium oxygen approach can be thought of as the opposite approach to lithium ion as the entire solid structure of lithium peroxide formed on discharge is decomposed into ions and gaseous oxygen on charge. While this approach leads to a significantly higher theoretical energy density, it poses significant challenges such as the reactivity of reaction intermediate and the challenges associated with such an extreme state change between the discharged and charged forms.
The positive electrode chemistry developed herein takes an intermediate approach to those of lithium ion and lithium oxygen batteries. During discharge (reaction 1), solid lithium iodate reacts with water in the electrolyte to form solid lithium hydroxide. On charge, the process is reversed (reaction 2 and 3) and lithium iodate is regenerated. This is achieved by exploiting the fact that the soluble lithium iodide forms on discharge acts as a soluble redox mediator, allowing the process to happen in solution. Since the solid structure is different between the charged and discharged states, the fundamental concern of structural stability in both the lithiated and delithiated states is potentially circumvented. Additionally, since the phase transition is not as extreme as lithium oxygen batteries, some of these challenges may also be mitigated.
Discharge:
LiHO3+3H2O+5Li++6e−→I−+6LiOH (1)
Charge:
2I−→I2+2e− (2)
6LiOH+3I2→5Li++5I−+3H2O+LiO3 (3)
FIG. 8 schematically illustrates a rechargeable battery 1, which includes anode 2, cathode 3, electrolyte 4, anode collector 5, and cathode collector 6. The battery can include a housing including an electrolyte (not shown). The battery can be a lithium battery, for example, a lithium-halogen oxyanion battery.
The electrolyte can include an aprotic solvent. The aprotic solvent can be 1,2-dimethoxyethane (DME), pyridine, DMA, Me-Im, G2 or G4. The electrolyte can include a salt, such as, both lithium bis(trifluromethane sulfonyl)-imide (LiTFSI) or lithium iodide (LiI).
The halogen oxyanion electrode, or positive electrode, can be a mixture of a halogen oxyanion salt and a conductive material. The conductive material can include carbon black, graphene, carbon nanotubes, or graphite. A binder, such as a polymer, can be applied hold the components of the electrode together, for example, a poly(carboxylic acid), poly(carboxylic acid), poly(acrylic acid) (PAA), poly-(methacrylic acid) (PMAA), poly(ethylene oxide) (PEO), poly(vinyl alcohol), or poly(vinylpyrrolidone).
The active material of the electrode is the halogen oxyanion salt. The halogen oxyanion salt can be a lithium salt, a sodium salt or a potassium salt. The halogen oxyanion salt can be a chlorate salt, a bromate salt or an iodate salt. Alternatively, the halogen oxyanion salt can be formed during a charging cycle by reaction of a metal hydroxide and a metal halide salt, for example, lithium hydroxide in the presence of lithium iodide.
The negative electrode can include lithium, sodium or potassium. In specific examples, the metal electrode can include lithium metal or a lithium compound, such as a lithium metal oxide (e.g., a lithium cobalt oxide or a lithium manganese oxide), lithiated graphite or silicon, or other metal ion complex. The term “battery” as used herein includes primary and secondary (rechargeable) batteries.
The separators that directly contact on the electrode can include porous organic polymers or porous glass separators. The separator permits ionic conduction but not electrical conduction between the electrodes.
In one implementation of this chemistry, an electrode resembling a conventional lithium ion composite electrode (active material+carbon+binder) is made using either lithium iodate or lithium hydroxide as the active material (depending on if the battery is assembled in its charged or discharged state). This chemistry is found to be highly sensitive to the electrolyte composition. In initial studies, an electrolyte based on 1,2-dimethoxyethane (DME) with both lithium bis(trifluromethane sulfonyl)-imide (LiTFSI) and lithium iodide (LiI) salts as well as added water (5-10 weight percent) has been used. The combined effect of a weakly interacting aprotic solvent (DME) and strong water-iodide interactions has been found to enhance the protonation of water (necessary for improving the discharge process (reaction 1)). The charge process has already been identified as a parasitic process on charging on lithium hydroxide based battery chemistries with lithium iodide present. It is unclear whether this occurs based on forming iodine (I2) as shown in reaction 2, or whether a triiodide (I3−) or pentaiodide (I5−) intermediate is instead formed.
A cell based on this chemistry would need sufficient electrode porosity to allow water and iodide to reach all active material in the positive electrode, however, since both discharged and charged states are either solid/insoluble in the electrolyte, or soluble in the electrolyte, an open, gas positive electrode (such as those used in lithium oxygen batteries) is not needed—improving volumetric energy density. The cell would likely have to be kept free of molecular oxygen to avoid parasitic reactions. This positive electrode could be paired with any negative electrode which is based on lithium ions (lithiated graphite/silicon, lithium metal, etc). Some protection of the negative electrode from water in the electrolyte is likely necessary.
While the above chemistry is based on lithium and iodide oxyanions, similar chemistries may be possible based on substituting lithium for other alkali metals such as sodium and potassium or substituting iodine for other halides such as bromine, chlorine or fluorine.
Calculated Theoretical Gravimetric Energy Density (Positive Electrode Active Material Only)
|
Lithium ion (LCO)
875-1000 Whr/kg
|
Lithium oxygen
3250-3450 Whr/kg
|
Lithium iodate/hydroxide
1550-1750 Whr/kg
|
|
Lithium iodide (LiI) has been extensively studied as a soluble redox mediator in Li—O2 batteries in order to catalyze the charging process. Despite some promising initial results, serious ambiguities exist in the literature regarding the reactivity between oxidized iodide species (I3−/I2) and the products formed during discharge (Li2O2/LiOH). In this work, we systematically examined the solvent-dependence of the oxidizing power of I3−/I− and I2/I3− towards Li2O2/LiOH through the use of ex-situ chemical reactions where the liquid reaction products were examined using UV-vis spectroscopy and 1H NMR, the solid reaction products were studied by Raman spectroscopy and XRD and the gaseous products were assessed using gas chromatography. In addition, the role of I− on the charging of Li—O2 batteries and LiOH pre-loaded cells was examined using DEMS, where the amount of oxygen release was quantified. Stronger solvation of Li+ and I− ions can lead to an increase in the oxidizing power of I3−, which allowed I3− to oxidize Li2O2/LiOH in stronger solvents, such as DMA, DMSO and Me-Im, whereas in weaker solvents (G4, DME), the more oxidizing I2 was needed before a reaction could occur. It was observed that Li2O2 was oxidized to O2, whereas LiOH was oxidized to an IO− intermediate, which could either disproportionate to LiIO3 or attack solvent molecules. Based on observed reactions with KO2/Li2O, we propose that while LiIO3 formation is thermodynamically favored, O2 gas evolution dominates in the oxidation of Li2O2 due to a kinetic barrier to O—O bond dissociation in the formation of LiIO3.
In general, lithium-oxygen batteries offer considerably higher gravimetric energy density than commercial Li-ion batteries (up to three times). Despite this promise, rechargeable nonaqueous Li—O2 batteries suffer from considerable fundamental issues relating to cycle life, parasitic reactions and poor round trip efficiency. Some of the most significant issues stem from the poor kinetics of Li2O2 oxidation on charge, which leads to high overpotential and considerable parasitic reactions. Soluble redox mediators, such as LiI, have been proposed as a potential solution to this problem, however, despite a number of promising initial results, there exists considerable discrepancy in the literature regarding the oxidizing power of I3−/I2 (the oxidized species formed during charge) against both Li2O2 and LiOH (potential discharge products of the Li—O2 chemistry), as well as the product formed by their oxidation. Some studies have suggested that I3− can oxidize Li2O2/LiOH, while others suggest the more oxidizing I2 is needed to react with Li2O2/LiOH and others still have claimed that LiOH is inactive in the presence of I3−/I2. There are studies that claim LiOH is oxidized reversibly to O2, whereas others claim it is irreversibly oxidized to LiIO3. In this study, we use detailed quantifications, a wide range of characterization techniques and cells constructed with a solid Li-conducting separator to eliminate shuttling in order to resolve these ambiguities. We show that the oxidizing power of I3− is solvent-dependent and can oxidize Li2O2/LiOH in stronger solvents (DMA, DMSO and Me-Im), but the more oxidizing I2 is required in weaker solvents like DME and G4. Furthermore, we show that Li2O2 is oxidized to O2, whereas LiOH is irreversibly oxidized to IO− which can either disproportionate to form LiIO3 or attack solvent molecules.
There has been considerable interest in nonaqueous Li—O2 batteries in the past decade due to their high theoretical gravimetric energy density (potentially up to 3 times that of commercial lithium ion batteries). See, for example, Aurbach, D., McCloskey, B. D., Nazar, L. F. & Bruce, P. G. Advances in understanding mechanisms underpinning lithium-air batteries. Nat. Energy 1, 16128 (2016); Bruce, P. G., Freunberger, S. A., Hardwick, L. J. & Tarascon, J.-M. Li—O2 and Li—S batteries with high energy storage. Nat. Mater. 11, 19-29 (2011); and Kwabi, D. G. et al. Materials challenges in rechargeable lithium-air batteries. MRS Bull. 39, 443-452 (2014), each of which is incorporated by reference in its entirety. This large theoretical improvement in gravimetric energy density stems from the fundamentally different reactions of the Li—O2 battery chemistry, which relies on reducing gaseous oxygen to form solid lithium peroxide (2Li+O2═Li2O2, E0=2.96 VLi) or lithium oxide (2Li+O2═Li2O, E0=2.91 VLi). Previous work shows that the discharge of nonaqueous Li—O2 batteries can produce Li2O2 with low overpotential, the morphology of which is dependent on the solvent, counter anion and potential/rate. See, for example, Lu, Y.-C. et al. Lithium-oxygen batteries: bridging mechanistic understanding and battery performance. Energy Environ. Sci. 6, 750 (2013); Viswanathan, V. et al. Li—O2 Kinetic Overpotentials: Tafel Plots from Experiment and First-Principles Theory. J. Phys. Chem. Lett. 4, 556-560 (2013); Kwabi, D. G. et al. Experimental and Computational Analysis of the Solvent-Dependent O2/Li+—O2− Redox Couple: Standard Potentials, Coupling Strength, and Implications for Lithium-Oxygen Batteries. Angew. Chem. Int. Ed. 55, 3129-3134 (2016); Johnson, L. et al. The role of LiO2 solubility in O2 reduction in aprotic solvents and its consequences for Li—O2 batteries. Nat. Chem. 6, 1091-1099 (2014); Sharon, D. et al. Mechanistic Role of Li+ Dissociation Level in Aprotic Li—O2 Battery. ACS Appl. Mater. Interfaces 8, 5300-5307 (2016); Burke, C. M., Pande, V., Khetan, A., Viswanathan, V. & McCloskey, B. D. Enhancing electrochemical intermediate solvation through electrolyte anion selection to increase nonaqueous Li—O2 battery capacity. Proc. Natl. Acad. Sci. 112, 9293-9298 (2015); Kwabi, D. G. et al. Controlling Solution-Mediated Reaction Mechanisms of Oxygen Reduction Using Potential and Solvent for Aprotic Lithium-Oxygen Batteries. J. Phys. Chem. Lett. 7, 1204-1212 (2016); and Mitchell, R. R., Gallant, B. M., Shao-Hom, Y. & Thompson, C. V. Mechanisms of Morphological Evolution of Li2O2 Particles during Electrochemical Growth. J. Phys. Chem. Lett. 4, 1060-1064 (2013), each of which is incorporated by reference in its entirety. Unfortunately, charging rechargeable Li—O2 batteries with nonaqeuous electrolytes requires a high overpotential to liberate molecular oxygen and this reaction is considerably more irreversible at high potentials as shown by McCloskey et al., leading to poor round-trip efficiency and cycle life resulting from parasitic side reactions. See, for example, McCloskey, B. D. et al. Combining Accurate O2 and Li2O2 Assays to Separate Discharge and Charge Stability Limitations in Nonaqueous Li—O2 Batteries. J. Phys. Chem. Lett. 4, 2989-2993 (2013); Aurbach, D., McCloskey, B. D., Nazar, L. F. & Bruce, P. G. Advances in understanding mechanisms underpinning lithium-air batteries. Nat. Energy 1, 16128 (2016); Bruce, P. G., Freunberger, S. A., Hardwick, L. J. & Tarascon, J.-M. Li—O2 and Li—S batteries with high energy storage. Nat. Mater. 11, 19-29 (2011); and Kwabi, D. G. et al. Materials challenges in rechargeable lithium-air batteries. MRS Bull. 39, 443-452 (2014), each of which is incorporated by reference in its entirety. Therefore, considerable efforts have been placed on attempting to catalyze the charging process in Li—O2 batteries. See, for example, Lim, H.-D. et al. Rational design of redox mediators for advanced Li—O2 batteries. Nat. Energy 1, 16066 (2016); Lim, H.-D. et al. Superior Rechargeability and Efficiency of Lithium-Oxygen Batteries: Hierarchical Air Electrode Architecture Combined with a Soluble Catalyst. Angew. Chem. Int. Ed. 53, 3926-3931 (2014); Liu, T. et al. Cycling Li—O2 batteries via LiOH formation and decomposition. Science 350, 530-533 (2015); Bergner, B. J., Schurmann, A., Peppler, K., Garsuch, A. & Janek, J. TEMPO: A Mobile Catalyst for Rechargeable Li—O2 Batteries. J. Am. Chem. Soc. 136, 15054-15064 (2014); Kwak, W.-J. et al. Li—O2 cells with LiBr as an electrolyte and a redox mediator. Energy Env. Sci 9, 2334-2345 (2016); Kwak, W.-J. et al. Understanding the behavior of Li-oxygen cells containing LiI. J Mater Chem A 3, 8855-8864 (2015); Chen, Y., Freunberger, S. A., Peng, Z., Fontaine, O. & Bruce, P. G. Charging a Li—O2 battery using a redox mediator. Nat. Chem. 5, 489-494 (2013); Sun, D. et al. A Solution-Phase Bifunctional Catalyst for Lithium-Oxygen Batteries. J. Am. Chem. Soc. 136, 8941-8946 (2014); Feng, N., He, P. & Zhou, H. Enabling Catalytic Oxidation of Li2O2 at the Liquid-Solid Interface: The Evolution of an Aprotic Li—O2 Battery. ChemSusChem 8, 600-602 (2015); Kundu, D., Black, R., Adams, B. & Nazar, L. F. A Highly Active Low Voltage Redox Mediator for Enhanced Rechargeability of Lithium-Oxygen Batteries. ACS Cent. Sci. 1, 510-515 (2015); Torres, W. R., Herrera, S. E., Tesio, A. Y., Pozo, M. del & Calvo, E. J. Soluble TTF catalyst for the oxidation of cathode products in Li—Oxygen battery: A chemical scavenger. Electrochimica Acta 182, 1118-1123 (2015); Wu, S., Tang, J., Li, F., Liu, X. & Zhou, H. Low charge overpotentials in lithium-oxygen batteries based on tetraglyme electrolytes with a limited amount of water. Chem Commun 51, 16860-16863 (2015); Zhu, Y. G. et al. Dual redox catalysts for oxygen reduction and evolution reactions: towards a redox flow Li—O2 battery. Chem Commun 51, 9451-9454 (2015); Pande, V. & Viswanathan, V. Criteria and Considerations for the Selection of Redox Mediators in Nonaqueous Li—O2 Batteries. ACS Energy Lett. 2, 60-63 (2016); Yao, K. P. C. et al. Utilization of Cobalt Bis(terpyridine) Metal Complex as Soluble Redox Mediator in Li—O2 Batteries. J. Phys. Chem. C 120, 16290-16297 (2016); Zeng, X. et al. Enhanced Li—O 2 battery performance, using graphene-like nori-derived carbon as the cathode and adding LiI in the electrolyte as a promoter. Electrochimica Acta 200, 231-238 (2016); Zhang, W. et al. Promoting Li2O2 oxidation via solvent-assisted redox shuttle process for low overpotential Li—O 2 battery. Nano Energy 30, 43-51 (2016); and Zhang, T., Liao, K., He, P. & Zhou, H. A self-defense redox mediator for efficient lithium-O2 batteries. Energy Env. Sci 9, 1024-1030 (2016), each of which is incorporated by reference in its entirety.
While solid-state catalysts have been employed to reduce the overpotential during charge, including metal oxides, modified carbon and metals/metal alloys, these catalysts rely on good electrical contact between Li2O2 and the catalyst throughout the entire charging process and do not suppress side reactions during charging. See, for example, Xu, J.-J. et al. Synthesis of Perovskite-Based Porous La0.75Sr0.25MnO3 Nanotubes as a Highly Efficient Electrocatalyst for Rechargeable Lithium-Oxygen Batteries. Angew. Chem. Int. Ed. 52, 3887-3890 (2013); Yao, K. P. C. et al. Solid-state activation of Li2O2 oxidation kinetics and implications for Li—O2 batteries. Energy Environ. Sci. 8, 2417-2426 (2015); Yin, Y.-B., Xu, J.-J., Liu, Q.-C. & Zhang, X.-B. Macroporous Interconnected Hollow Carbon Nanofibers Inspired by Golden-Toad Eggs toward a Binder-Free, High-Rate, and Flexible Electrode. Adv. Mater. 28, 7494-7500 (2016); Li, L. & Manthiram, A. O- and N-Doped Carbon Nanowebs as Metal-Free Catalysts for Hybrid Li-Air Batteries. Adv. Energy Mater. 4, 1301795 (2014); Shui, J. et al. Nitrogen-Doped Holey Graphene for High-Performance Rechargeable Li—O2 Batteries. ACS Energy Lett. 1, 260-265 (2016); Xu, J.-J., Wang, Z.-L., Xu, D., Zhang, L.-L. & Zhang, X.-B. Tailoring deposition and morphology of discharge products towards high-rate and long-life lithium-oxygen batteries. Nat. Commun. 4, (2013); Kwon, H.-M. et al. Effect of Anion in Glyme-based Electrolyte for Li—O2 Batteries: Stability/Solubility of Discharge Intermediate. Chem. Lett. 46, 573-576 (2017); and Wong, R. A. et al. Critically Examining the Role of Nanocatalysts in Li—O2 Batteries: Viability toward Suppression of Recharge Overpotential, Rechargeability, and Cyclability. ACS Energy Lett. 3, 592-597 (2018), each of which is incorporated by reference in its entirety. An alternative approach is the use of soluble redox mediators to promote electron transfer to the surface of the electronically insulating Li2O2, where the redox mediator is first electrochemically oxidized at the electrode surface and then the oxidized form of the redox mediator chemically oxidizes Li2O2 to form Li+ ions and molecular oxygen and regenerate the reduced form of the redox mediator. See, for example, Radin, M. D. & Siegel, D. J. Charge transport in lithium peroxide: relevance for rechargeable metal-air batteries. Energy Environ. Sci. 6, 2370 (2013), which is incorporated by reference in its entirety. Many organic molecules like TEMPO, TDPA and TTF as well as inorganics like LiI and LiBr have been proposed as redox mediators. Lithium iodide (LiI) has received considerable attention owing to a number of studies suggesting high cycling performance14,15. See, for example, Lim, H.-D. et al. Superior Rechargeability and Efficiency of Lithium-Oxygen Batteries: Hierarchical Air Electrode Architecture Combined with a Soluble Catalyst. Angew. Chem. Int. Ed. 53, 3926-3931 (2014); Liu, T. et al. Cycling Li—O2 batteries via LiOH formation and decomposition. Science 350, 530-533 (2015); Bergner, B. J., Schurmann, A., Peppler, K., Garsuch, A. & Janek, J. TEMPO: A Mobile Catalyst for Rechargeable Li—O2 Batteries. J. Am. Chem. Soc. 136, 15054-15064 (2014); Kwak, W.-J. et al. Li—O2 cells with LiBr as an electrolyte and a redox mediator. Energy Env. Sci 9, 2334-2345 (2016); Kwak, W.-J. et al. Understanding the behavior of Li-oxygen cells containing LiI. J Mater Chem A 3, 8855-8864 (2015); Bergner, B. J. et al. Understanding the fundamentals of redox mediators in Li—O2 batteries: a case study on nitroxides. Phys. Chem. Chem. Phys. 17, 31769-31779 (2015);
Bergner, B. J. et al. How To Improve Capacity and Cycling Stability for Next Generation Li—O2 Batteries: Approach with a Solid Electrolyte and Elevated Redox Mediator Concentrations. ACS Appl. Mater. Interfaces 8, 7756-7765 (2016); Lee, D. J., Lee, H., Kim, Y.-J., Park, J.-K. & Kim, H.-T. Sustainable Redox Mediation for Lithium-Oxygen Batteries by a Composite Protective Layer on the Lithium-Metal Anode. Adv. Mater. 28, 857-863 (2016); Kundu, D., Black, R., Adams, B. & Nazar, L. F. A Highly Active Low Voltage Redox Mediator for Enhanced Rechargeability of Lithium-Oxygen Batteries. ACS Cent. Sci. 1, 510-515 (2015); and Torres, W. R., Herrera, S. E., Tesio, A. Y., Pozo, M. del & Calvo, E. J. Soluble TTF catalyst for the oxidation of cathode products in Li—Oxygen battery: A chemical scavenger. Electrochimica Acta 182, 1118-1123 (2015), each of which is incorporated by reference in its entirety. Lim et al. have suggested stable cycling with low overpotential over 900 cycles using LiI as a soluble redox mediator in a tetraglyme (G4) electrolyte with a CNT fibril electrode. In addition, Liu et al. have claimed to achieve 2000 cycles using LiI in a 1,2-dimethoxyethane (DME)-based electrolyte containing ˜5 v % H2O with a reduced graphene oxide electrode and LiOH as the dominant discharge product. See, for example, Lim, H.-D. et al. Superior Rechargeability and Efficiency of Lithium-Oxygen Batteries: Hierarchical Air Electrode Architecture Combined with a Soluble Catalyst. Angew. Chem. Int. Ed. 53, 3926-3931 (2014); and Liu, T. et al. Cycling Li—O2 batteries via LiOH formation and decomposition. Science 350, 530-533 (2015), each of which is incorporated by reference in its entirety. However, ambiguities exist in the influence of LiI on both the discharge and charge processes. See, for example, Kwak, W.-J. et al. Understanding the behavior of Li-oxygen cells containing LiI. J Mater Chem A 3, 8855-8864 (2015); Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016); Tulodziecki, M. et al. The role of iodide in the formation of lithium hydroxide in lithium-oxygen batteries. Energy Env. Sci (2017), Qiao, Y. et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Lett. 1869-1878 (2017); and Li, Y. et al. Li—O2 Cell with LiI (3-hydroxypropionitrile)2 as a Redox Mediator: Insight into the Working Mechanism of I− during Charge in Anhydrous Systems. J. Phys. Chem. Lett. 4218-4225 (2017), each of which is incorporated by reference in its entirety.
LiI addition in the electrolyte can change the dominant discharge product from Li2O2 to LiOH, LiOH.H2O or LiOOH.H2O by decreasing the pKa of water in the electrolyte. See, for example, Liu, T. et al. Cycling Li—O2 batteries via LiOH formation and decomposition. Science 350, 530-533 (2015); Kwak, W.-J. et al. Understanding the behavior of Li-oxygen cells containing LiI. J Mater Chem A 3, 8855-8864 (2015); Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016); Tulodziecki, M. et al. The role of iodide in the formation of lithium hydroxide in lithium-oxygen batteries. Energy Env. Sci (2017); and iao, Y. et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Lett. 1869-1878 (2017); and Zhu, Y. G. et al. Proton enhanced dynamic battery chemistry for aprotic lithium-oxygen batteries. Nat. Commun. 8, 14308 (2017), each of which is incorporated by reference in its entirety. Adding water to DME-based electrolytes up to 5000 ppm still results in Li2O2 on discharge when no LiI is present. On the other, while the discharge of a Li—O2 battery forms Li2O2 in LiI-containing DME-based electrolytes without added water, the dominant discharge product can become LiOH (H2O>˜500 ppm) or LiOOH (H2O>˜5 v %) instead of Li2O2 with water addition in the electrolytes. See, for example, Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016); Tulodziecki, M. et al. The role of iodide in the formation of lithium hydroxide in lithium-oxygen batteries. Energy Env. Sci (2017); Qiao, Y. et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Lett. 1869-1878 (2017); Li, Y. et al. Li—O2 Cell with LiI (3-hydroxypropionitrile)2 as a Redox Mediator: Insight into the Working Mechanism of I− during Charge in Anhydrous Systems. J. Phys. Chem. Lett. 4218-4225 (2017); Zhu, Y. G. et al. Proton enhanced dynamic battery chemistry for aprotic lithium-oxygen batteries. Nat. Commun. 8, 14308 (2017); and Kwabi, D. G. et al. The effect of water on discharge product growth and chemistry in Li—O2 batteries. Phys Chem Chem Phys 18, 24944-24953 (2016), each of which is incorporated by reference in its entirety. Specifically, at low H2O:LiI ratios (lower than 5), LiOH instead of Li2O2 has been observed, which is accompanied by the oxidation of iodide to triiodide, while at high H2O:LiI ratios, a mixture of Li2O2, LiOOH—H2O and LiOH—H2O has been observed with no triiodide detected44. The formation of LiOH and relevant products upon discharge is promoted by the lowered deprotonation energy of water due to the stronger solvation of water molecules by organic solvent molecules such as MeCN (Kwabi et al.) and the interactions between water molecules and anions such as I−. See, for example, Tulodziecki, M. et al. The role of iodide in the formation of lithium hydroxide in lithium-oxygen batteries. Energy Env. Sci (2017); and Kwabi, D. G. et al. The effect of water on discharge product growth and chemistry in Li—O2 batteries. Phys Chem Chem Phys 18, 24944-24953 (2016), each of which is incorporated by reference in its entirety. These studies have shown that the major proton source for the formation of LiOH/LiOH—H2O/LiOOH—H2O is added water and not the decomposition of solvents such as DME, which is supported by a subsequent computational work showing water as a more energetically favorable proton source for the formation of LiOH than DME. On the other hand, Qiao et al. and Kwak et al. have suggested iodide-catalyzed decomposition of G4 to promote the formation of LiOH based on the observation of greater LiOH with increasing amounts of LiI added to the electrolyte. This apparent discrepancy can be explained by the addition of water associated with the LiI used (Qiao et al. with dried LiI having >98%, Sigma Aldrich under vacuum at 80° C. overnight and Kwak et al. with anhydrous LiI, Sigma-Aldrich with no mention of drying).
I2 (which is more oxidizing than I3−) is required to oxidize Li2O2 and generate molecular O2 in anhydrous DME and G4. Ambiguities exist in what oxidized iodide species can oxidize Li2O2 and LiOH and what oxidation products, such as O2, are formed. For example, Qiao et al. have reported that I3− can oxidize peroxide-like species (in part Li2O2) to form O2 with water addition up to 30 v % in G4. See, for example, Qiao, Y. et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Lett. 1869-1878 (2017), which is incorporated by reference in its entirety. On the other hand, Zhu et al. discuss that the oxidation of Li2O2 requires the formation of I2 while LiOOH—H2O formed in diglyme (G2) and DMSO with 9.1 v % water can be oxidized to form O2 by I3−. See, for example, Zhu, Y. G. et al. Proton enhanced dynamic battery chemistry for aprotic lithium-oxygen batteries. Nat. Commun. 8, 14308 (2017), which is incorporated by reference in its entirety. Moreover, Liu et al15 have suggested that I3− can oxidize LiOH formed in DME and G4 with the addition of ˜5 v % water to generate O2. In addition, Zhu et al. have suggested that LiOH was oxidized to O2 by 12. See, for example, Zhu, Y. G. et al. Proton enhanced dynamic battery chemistry for aprotic lithium-oxygen batteries. Nat. Commun. 8, 14308 (2017), which is incorporated by reference in its entirety. However, the concept of LiOH oxidation to O2 by I3− is rebutted by Viswanathan et al. arguing the oxidation of LiOH by I3− as thermodynamically uphill in DME and Shen et al. suggesting that observed charge capacity is from iodine redox and inactive LiOH is accumulated. See, for example, Viswanathan, V. et al. Comment on ‘Cycling Li—O2 batteries via LiOH formation and decomposition’. Science 352, 667-667 (2016); and Shen, Y., Zhang, W., Chou, S.-L. & Dou, S.-X. Comment on ‘Cycling Li—O2 batteries via LiOH formation and decomposition’. Science 352, 667-667 (2016), each of which has been incorporated by reference in its entirety. Furthermore, Qiao et al. argued that LiOH was inactive in the presence of I3− and I2, and Burke et al. have proposed that LiOH is oxidized irreversibly to lithium iodate (LiIO3) by I2 in DME, which is in agreement with Liu et al. noting LiIO3 formation from LiOH formed in a 3 wt % water/DME solution. See, for example, Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016); Qiao, Y. et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Lett. 1869-1878 (2017); and Liu, T. et al. Response to Comment on “Cycling Li—O2 batteries via LiOH formation and decomposition”. Science 352, 667-667 (2016),
The discrepancies found for the oxidation of Li2O2 and LiOH by I3−/I2 in previous work may result from several factors. First, some previous claims of O2 evolution have not been supported by quantification of reaction products to ensure the amount of oxygen detected as the dominant path of the reaction but not from cell leakage or H2O2 contamination of the solvents. Second, the oxidizing power of I3− and I2 against Li2O2 or LiOH can be solvent-dependent. Generally speaking, I− ions go through two distinct redox transitions during oxidation in aprotic electrolytes, having first iodide anions (I−) oxidized to form triiodide (I3−) and I3− oxidized to form iodine (I2), where the potentials of the I−/I3− and I3−/I2 redox transitions can be significantly influenced by solvent. While it has been previously suggested that changes in these redox potentials may be important for the performance of LiI as a redox mediator in Li—O2 batteries, this effect has not been studied systematically. This concept is supported by a very recent study, where Nakanishi et al.54 have shown that the thermodynamic shifts in the iodide redox on a lithium scale due to the effect of solvent and lithium concentration can change the oxidizing power of I3− against Li2O2 in 1 M and 2.8 M LiTFSI electrolytes in DMSO and G4 with 0.1 M LiI. See, for example, Nakanishi, A. et al. Electrolyte Composition in Li/O2 Batteries with LiI Redox Mediators: Solvation Effects on Redox Potentials and Implications for Redox Shuttling. J. Phys. Chem. C (2018), which is incorporated by reference in its entirety.
The role of LiI on the charging process of Li—O2 batteries can be examined by systematically studying the solvent-dependent oxidizing power of I3−/I− and I2/I3− towards Li2O2 and LiOH. The oxidizing power of I3−/I− and I2/I3− towards Li2O2 and LiOH was examined chemically by examining the consumption of I3− upon addition of synthetic Li2O2 (from disproportionation), where the liquid reaction product was examined using UV-vis spectroscopy and 1H NMR, the solid reaction products were studied by Raman spectroscopy and XRD and the gaseous products were assessed using gas chromatography. In addition, the role of I− on the charging of Li—O2 batteries and LiOH-pre-loaded cells was examined using DEMS, where the amount of oxygen release was quantified. It is shown here that I3−/I− potentials increase with greater solvent AN, suggesting stronger solvation of I− while I2/I3− redox potentials are largely solvent independent. Therefore, stronger solvation of Li+ and I− ions in solvents such as DMA, DMSO and Me-Im can increase the oxidizing power of I3−/I−, allowing I3− to effectively oxidize Li2O2 to generate O2, which was supported by chemical and electrochemical experiments. On the other hand, in solvents where both I− and Li+ are weakly solvated such as glymes, I3−/I− redox potentials are not high enough to oxidize Li2O2 and the more oxidizing I2 is required for the oxidation of Li2O2 to O2 to proceed. The oxidation of LiOH by I3− was also found to be solvent dependent, where no reaction was observed in G4, DME and pyridine while the reaction proceeded to completion in DMA, DMSO and Me-Im where the I3−/I− redox potential was above ˜3.1 VLi. It is shown here that the reaction between LiOH and oxidized iodide species produced water and a hypoiodite (IO−) intermediate, which could either disproportionate to form LiIO3 or attack solvent molecules and result in decomposition products such as dimethyl sulfone (DMSO2). From GC of ex-situ reactions and DEMS during the charging of pre-loaded LiOH electrodes, no O2 gas evolution was observed during the reaction between LiOH and oxidized iodide species. The selectivity between O2 and the thermodynamically preferred LiIO3 can be governed by a kinetic barrier relating to O—O bond dissociation and this kinetic barrier prevents IO− formation, allowing for the evolution of gaseous O2 when oxidizing Li2O2, which was supported by reactions between oxidized iodide species and KO2/Li2O.
Experimental
I. Chemicals
High purity dimethyl sulfoxide (“DMSO”, Sigma Aldrich, anhydrous, >99.9%), diethylene glycol dimethyl ether (“G2”, Sigma Aldrich, anhydrous, 99.5%), N,N-dimethylacetamide (“DMA”, Sigma Aldrich, anhydrous, 99.8%), 1-methylimidazole (“Me-Im”, Sigma Aldrich, 99%), pyridine (Sigma Aldrich, anhydrous, 99.8%) and tetraethylene glycol dimethyl ether (Sigma Aldrich, >99%) were purchased and dried over molecular sieves for at least a week before use. 1,2-dimethoxyethane (DME) was purchased from Acros and was degassed and dried using a Glass Contour Solvent Purification System built by SGWater USA, LLC. Lithium bis(trifluoromethanesulfonyl)imide (“LiTFSI”, 99.99% extra dry grade from Solvay) was used as received. High purity LiI (ultra dry, 99.999% pure), I2 (99.9985% pure), Li2O2 (90%), Li2O (99.5%) and decamethylferrocene (“Me10Fc”, 99%) chemicals were ordered from Alfa Aesar and were used as received. LiOH (anhydrous, 99.995%) was purchased from Alfa Aesar and was further dried under vacuum for 24 hrs at 170° C. to ensure only the anhydrous phase remained (see FIG. 9). KO2 (99% pure) powder was purchased from Sigma Aldrich and was used as received.
All chemicals were stored in an argon-filled glovebox (MBraun, USA) with H2O and O2 content of <0.1 ppm. Electrolytes were prepared by dissolution of a desired amount of salts in the solvent with molarity determined by the volume of solvent added. The total H2O content in the solvents and electrolytes was checked using a C20 compact Karl Fisher coulometer from Mettler Toledo and for the dry solvent it was <20 ppm for ˜2 g of sample. A 20 wt % solution of LiTFSI in DME was found to have a slightly higher water content of 21 ppm (compared with 3.0 ppm for the pure DME solvent). Solutions of 0.2 M LiI in all solvents were clear, indicating the absence of H2O2 contamination, which can be of particular concern in glymes.
Due to the low purity of commercially available Li2O2 (90%), for most experiments, Li2O2 was first synthesized through the well-known disproportionation reaction between KO2 and Li-containing salt4:
2LiTFSI+2KO2→2KTFSI+Li2O2+O2 (1)
In all experiments, a two times excess of LiTFSI was used, the reaction occurred in the solvent being studied and the reaction was allowed to proceed for one hour with stirring to ensure complete production of Li2O2. The resulting solution of unconsumed LiTFSI and produced KTFSI as well as the precipitated Li2O2 was used directly without additional processing/washing. The presence of LiTFSI/KTFSI was assumed to have a negligible influence on subsequent reactions.
II. Redox Potential Measurements of I3−/I− and I2/I3− Redox Couples Using Cyclic Voltammograms
Cyclic voltammograms (CVs) were collected of solutions of 0.5M LiTFSI+10 mM LiI collected at 100 mVps under argon environment in each of the considered solvents. Electrolytes were prepared in an Argon-filled glove box (MBraun, <0.1 ppm H2O, <0.1 ppm O2) and transferred to a second Argon-filled glovebox directly through a shared antechamber (MBraun, <0.1 ppm H2O, <0.10% O2). The electrolyte was bubbled with Argon for at least 30 minutes prior to beginning electrochemistry. Due to the volatility of DME, for the DME experiment, the argon was first saturated with DME vapor by bubbling the Argon through pure DME prior to going to the electrolyte. The working macroelectrode was platinum and either a Li metal (G4, DME, DMSO) or lithium titanium oxide (pyridine, DMA, Me-Im) counter electrode was used. A fritted Ag/Ag+ reference electrode (0.1M TBAClO4+10 mM AgNO3 in MeCN) was used and following collection of CVs, 2 mM Me10Fc was added to the solution and CVs were collected to determine the Me10Fc half-wave potential. Li+/Li potentials were determined in G4, DME, DMA and DMSO using a piece of Li metal. See Table 2.
TABLE 2
|
|
Estimated half-wave potentials of I−/I3− and I3−/I2 vs Me10Fc
|
Solvent
I−/I3− vs Me10Fc
I3−/I2 vs Me10Fc
|
|
TEGDME
0.003
0.653
|
DME
0.018
0.632
|
Pyridine
0.286
0.496
|
DMA
0.069
0.648
|
DMSO
0.228
0.629
|
Me-Im
0.299
0.748
|
|
In addition to the two expected peaks associated with the I−/I3− and I3−/I2 redox transitions, both pyridine and Me-Im exhibit additional redox features (FIG. 10). Pyridine is known to in literature to form stable complexes with oxidized forms of iodide as well as adsorb strongly on platinum surfaces. One can therefore attribute the small peak at ˜−0.35V vs Me10Fc to an adsorption/desorption process (total charge passed ˜1.2×10−7 C) and the additional features in the I3−/I− and I2/I3− redox peaks to the formation of iodine-solvent complexes. Given the similarities in structure between Me-Im and pyridine, we suggest that similar iodine-solvent complexes are also possible in Me-Im and would account for the additional feature observed in the anodic sweep of the Me-Im CV. Since neither pyridine nor Me-Im are likely solvent candidates for lithium oxygen batteries due to instability issues, the precise origin and implications of these additional redox features in the presented CVs was not investigated further.
III. Using I−/I− and I3−/I2 for Chemical Li2O2 and LiOH Oxidation
In an argon-filled glovebox (MBraun, O2, H2O<0.1 ppm), solutions of I3− (0.2M LiI+50 mM I2) and I2 (50 mM I2) were first prepared in each solvent and allowed to fully dissolve under stirring. For studies of Li2O2, a two times excess of Li2O2 was first synthesized through disproportionation using 1 mL of the solvent to be studied and the reaction was allowed to proceed under stirring for ˜1 hour. For studies of LiOH, a two times excess of LiOH powder was added to 1 mL of solvent and allowed to reach equilibrium under stirring for ˜1 hour. Next, 1 mL of the I3−/I2 solution was added to the vial with Li2O2/LiOH and 1 mL of solvent. The reaction was allowed to take place under stirring for 24 hours, following which, the solid product was allowed to settle for 1 hour and the liquid and solid phases were separated. This ex-situ, chemical analog approach has been used extensively previously and has been very effective at isolating a chemical reaction to enable its independent study.
IV. Physical Characterization of Reaction Liquids, Solids and Gases
UV-Vis was performed using a Perkin Elmer Lambda 1050 UV/VIS/NIR Spectrophotometer. The pure solvent (e.g. G4, DME, etc.) was used as the blank solution, except in assessments of the pure solvent absorbance where no blank was used. Solutions were prepared in an Argon glovebox and sealed in a quartz cuvette used for data collection, preventing air exposure. Due to high molar absorptivity of I3−, the solutions with I3− were diluted in pure solvent so that the intensity of I3− absorption signals (at ˜293 nm and ˜364 nm) were within the calibration range (FIG. 12-14). The concentrations of triiodide were calculated based on the absorption intensity at the wavelength of the peak absorbance of the calibration curves. In the case where both peaks were distinguishable above the solvent's inherent absorbance (FIG. 15), the average of the concentration determined by both peaks was used. The absorption spectra in the figures are rescaled (arbitrary units) in order to visualize the difference in I3− concentration for different solutions. Thus, a high concentration of I3− corresponds to high absorption at wavelengths 293 nm and 364 nm and vice versa. The scale factors and the calculation of I3− concentrations are summarized in Table 1. Dilutions were calculated based on a mass balance of the added solvent and I3− solution. Error bars for the diluted samples were estimated based on an error of +/−0.5 mg in each weight measurement (+/−0.1 mg from the accuracy of the balance with additional error incurred due to a small amount of evaporation). In the case of determining the concentration of I2 in solution, the solution was first mixed with a ˜4 times excess of LiI to chemically form I3− in solution through the association of I2 and I−. The resulting I3− concentration was then determined using the as described above.
TABLE 1
|
|
UV-vis scaling factors applied in manuscript figures
|
Scale in
|
FIG.
Data
Dilutions
FIG.
|
|
22
G4
D1 - 51.61 mg into 1983.49 mg pure G4
1
|
before
D2 - 43.98 mg into 2967.27 mg pure G4
|
22
G4
D1 - 50.48 mg into 1991.72 mg pure G4
1
|
after
D2 - 49.75 mg into 2976.53 mg pure G4
|
22
DME
D1 - 43.71 mg into 1751.37 mg pure DME
0.955
|
before
D2 - 41.1 mg into 2614.28 mg pure DME
|
22
DME
D1 - 90.79 mg into 1750.39 mg pure DME
0.671
|
after
D2 - 69.02 mg into 2611.43 mg pure DME
|
22
Pyridine
D1 - 49.3 mg into 1953.19 mg pure Pyridine
1.254
|
before
D2 - 46.18 mg into 2930.22 mg pure Pyridine
|
22
Pyridine
D1 - 49.26 mg into 1960.9 mg pure Pyridine
0.762
|
after
D2 - 44.68 mg into 2932.65 mg pure Pyridine
|
22
DMA
Starting I3− concentration from mass balance
2.34
|
before
|
22
DMA
0.4 mL into 2.6 mL pure DMA
0.00417
|
after
(scaled to reflect deviation from usual dilution)
|
22
DMSO
Starting I3− concentration from mass balance
0.641
|
before
|
22
DMSO
0.4 mL into 2.6 mL pure DMSO
0.00417
|
after
(scaled to reflect deviation from usual dilution)
|
22
Me-Im
Starting I3− concentration from mass balance
1.468
|
before
|
22
Me-Im
0.4 mL into 2.6 mL pure Me-Im
0.00417
|
after
(scaled to reflect deviation from usual dilution)
|
|
Iodometric titration was performed with a prepared 5 mM thiosulfate solution (anhydrous 99.99% Sigma-Aldrich, stored in desiccator) using a 50 mL burette (Class A, graduation 0.10 mL, tolerance ±0.05 mL from VWR) and starch indicator (1% w/v of Amylodextrin) in aqueous solution (18.2 MΩ·cm, Millipore). The thiosulfate solution was first standardized with a KIO3 (99.995% pure from Sigma Aldrich) solution of a known concentration in three separate probes. 10 mL of KOI3 solution was added to Erlenmeyer flask (250 ml), to which ˜100 mg of KI (Bioultra >99.5% TA from Sigma Aldrich) and 2 mL of 6 M H2SO4 was added. The obtained I3− solution was immediately titrated with thiosulfate solution. Just before the end-point, indicated by a light straw-like color, the starch solution was added, resulting in a change of color to a dark red/brown (this color change is due to branched Amylodextrin rather than blue when using straight chain amylose). The thiosulfate solution was prepared fresh the same day as the titration experiment. Titrations to determine LiIO3 formed through reactions in I3− were performed by allowing the reaction to reach completion (indicated by the complete consumption of I3− based on the solution becoming colorless). The entirety of the solid and liquid phases were then transferred to an Erlenmeyer flask (rinsing the reaction vial three times with DI H2O) and then titrated as per above.
Raman spectroscopy was performed on a LabRAM HR800 microscope (Horiba Jobin Yvon) using an external 20 mW He:Ne 633 nm laser (Horiba, Jobin Yvon) and, focused with a 50× long working distance objective and a 10-0.3 neutral density filter. A silicon substrate was used to calibrate the Raman shift. An air-tight cell was used for powders, and all samples preparation was done in an argon-filled glovebox. Liquid samples were tightly sealed in a 3 mL vial and assessed using a 10× working length. Reference spectra of Li2O2, LiOH, LiOH—H2O and LiIO3 are available in FIG. 9.
XRD of discharged products and powders was performed on a Rigaku Smartlab diffractometer in Bragg-Brentano geometry. A domed air tight XRD cell holder from Panalytical was used to prevent exposing the electrodes to ambient atmosphere. Reference spectra for LiOH, LiOH—H2O, Li2O2, LiI, DMSO2 and LiIO3 are available in FIG. 16.
1H NMR was performed on a Bruker AVANCE and Bruker AVANCE III-400 MHz nuclear magnetic resonance (NMR) spectrometers. Samples were prepared by mixing 0.5 mL of the sample+0.1 mL of DMSO-D6 (for NMR locking)+10 μL of internal reference (either MeCN (Acetonitrile anhydrous, 99.8%, Sigma-Aldrich dried over molecular sieves) or 1,4-dioxane (anhydrous, 99.8%, Sigma-Aldrich, dried over molecular sieves) chosen to avoid overlap with peaks of interest).
Gas chromatography (GC) was performed using Argon (5.0, Praxair) as a carrier gas flowing at ˜12 sccm, through a glass cell. Cell was purged with Ar for 1 hour, during the last 15 minutes of which, a background spectrum was taken. The reaction compartment contained 15 mL of DMSO with either Li2O2 formed from disproportionation or commercial LiOH suspended in solution with active stirring. 2 mL of I3− solution (0.2 M LiI+50 mM I2 in DMSO) was injected using a syringe which was sealed onto a port of the glass reaction cell prior to purging without exposure to the ambient. 1 mL of gas sample was injected into a gas chromatograph (GC, SRI 8610C in the Multi-Gas #3 configuration). Samples were injected after 2, 22, 42 and 62 minutes of reaction. GC was calibrated using a 2500 ppm O2+17000 ppm N2 in Argon gas mixture.
V. Li—O2 Cell Assembly and Tests
Li—O2 cells consisted of a lithium metal negative electrode (Chemetall, Germany, 15 mm in diameter) and a carbon paper with gas diffusion layer positive electrode (FuelCellsEtc, F2GDL, LOT: TST008, 12.5 mm diameter). The carbon paper was dried for 24 hours at 90° C. under vacuum and transferred to a glove box (H2O<0.1 ppm, O2<0.1 ppm, Mbraun, USA) without exposure to ambient air. Glass fiber (Whatman, GF-A/GF-F, 17 mm diameter) was dried at 150° C. under vacuum overnight and was transferred to the glove box without exposure to the ambient. Lithium-ion conducting glass-ceramic electrolyte (19 mm diameter, 150 μm thick, LICGC, Ohara Corp) was dried at 80° C. under vacuum overnight. Cells were constructed by placing a single piece of glass fiber separator on top of the lithium, adding 120 μL of liquid electrolyte, followed by the lithium-ion conducting glass-ceramic electrolyte, another piece of glass fiber separator, another 120 μL of liquid electrolyte and finally the carbon paper positive electrode. No 316 stainless steel current collector was used to avoid a reaction which was observed between the 316 stainless steel and iodine formed during charge in some cells (see FIG. 16). The origin of this corrosion is not fully understood and is worthy of further investigation as it poses challenges for the practical implementation of LiI as a redox mediator, however, adequate performance over a single charging cycle was acquired by simply avoiding the use of the current collector and restricting the electrolyte contact with the 316 stainless steel spring as much as possible. For cells not analyzed using DEMS, following assembly, cells were transferred to a connected second argon glove box (Mbrau2n, USA, H2O<0.1 ppm, O2<1%) without exposure to air and pressurized with dry O2 (Airgas, 99.999% pure, H2O/CO/CO2<0.5 ppm) to 25 psi (gauge) to ensure that an adequate amount of O2 was available. The oxygen pressure in the cell was measured using a pressure gauge during the experiments to confirm proper cell sealing. Electrochemical tests were conducted using a Biologic VMP3.
LiOH preloaded electrodes were prepared by drop casting a slurry (70% wt Vulcan Carbon, 20% wt LiOH, 10% wt PTFE) onto neat carbon paper (Toray TGP-H-60, 12.5 mm diameter). The vulcan carbon (VC), PTFE and carbon paper were dried at 80° C. under vacuum for 24 hours and transferred to a glovebox (Mbraun, USA, H2O<0.1 ppm, O2<1%) without exposure to the ambient. The LiOH and VC were ground into a homogenous mixture using a mortar and pestle then added to a suspension of PTFE in DME. After allowing the mixture to stir for 1 hour, the slurry was drop cast 50 μL at a time until the desired mass loading was achieved. Individual 12.5 mm pieces of carbon paper were weighed before and after drop casting to determine the amount of mixture deposited. Typical loadings of the VC/LiOH/PTFE mixture were 3.9-5.0 mg per electrode (1.267 cm2). Electrodes were additionally dried under vacuum for ˜15 minutes to remove residual DME.
VI. Differential Electrochemical Mass Spectroscopy of Cells During Charging
A custom-made DEMS setup based on a design by McCloskey et al., which has been reported previously, was used for assessing gas evolution during the charging process. O2, CO, CO2, H2 and H2O evolution during charge was quantified at 20 minute intervals using a mass spectrometer coupled with pressure monitoring. Details of DEMS and cell technical construction are available online. Argon (Airgas, 99.999% pure, O2, H2O, CO2<1 ppm) was used as a carrier gas. In all cells, no detectable quantities of CO, H2 and H2O were detected, so these values are omitted from all plots. Cells were prepared as described above. Li—O2 cells were first discharged under O2 environment for 20 hours at 0.05 mA/cm2. The cell environment was then changed to Argon by evacuating the cell and refilling it with Argon five times and charged at 0.1 mA/cm2 to a cut-off voltage of 4.5 VLi. LiOH− preloaded electrodes were charged under argon environment at 0.1 mA/cm2 to a cut-off voltage of 4.5 VLi.
Results and Discussion
VII. Solvent-Dependent Potential of I3−/I−
The redox potential of I3−/I− was shown to shift positively against the solvent-insensitive reference potential of decamethylferrocene (Me10Fc) from DME, DMA and DMSO while that of I3−/I2 remained nearly constant, as shown in FIG. 1B. The reduction and oxidation peaks of the I3−/I− (centered between 0.02 and 0.23 VMe10Fc) and I3−/I2 (centered at ˜0.64 VMe10Fc) couples were observed in cyclic voltammograms (CVs), from which the redox potential of I3−/I− was obtained by averaging the reduction and oxidation peak centers (FIG. 1, FIG. 10). These measurements were collected using a Pt macroelectrode as the working electrode and Ag+/Ag as the reference electrode in solvents containing 10 mM LiI with 0.5 M LiTFSI under an argon environment, where the Ag+/Ag reference electrode potential scale was converted to that of Me10Fc following previous work−6 (FIG. 18) for each solvent. The shifts in the potential of the I3−/I− redox were plotted against reported Guttmann acceptor number (AN), Guttmann donor number (DN) and dielectric constant for these solvents (FIGS. 19-21). Here the positive shift in the potential of I3−/I− can be attributed to increasing thermodynamic stability of the I− ion through solvation via higher Guttmann acceptor number (AN), dielectric constant and possibly through the formation of ion pairs with Li+47. Solvation effects can more significantly influence I− ions as there are three I− ions for each I3− and the larger I3− ions (which are more charge diffuse) might interact with the solvent less. Following the same argument, the solvent-insensitive redox potential of I2/I3− can be attributed to the weak solvation of I3− and I2 in the considered solvents.
Such positive shifts in the potential of I3−/I− increases its oxidizing power towards Li2O2 (or the thermodynamic driving force to oxidize Li2O2) to evolve O2 (Li2O2=>2Li++O2+2e−), with a trend of G4<DME<Pyridine<DMA<DMSO<Me-Im. Considering that the Li+/Li potential decreases from DME, DMA to DMSO on the Me10Fc scale due to stronger lithium solvation with high Guttmann donor number (DN) and high dielectric constant (the Born model), and that the free energy of O2 and Li2O2 are solvent independent, the redox potential of Li+,O2/Li2O2 would follow the same trend as the Li+/Li potential, decreasing from 0.00 VMe10Fc in DME to −0.11 VMe10Fc in DMA and −0.31 VMe10Fc in DMSO. Therefore, as the potential of I3−/I− shifts to higher values from DME, DMA to DMSO and that of Li+,O2/Li2O2 moves to lower values on the Me10Fc scale, the oxidative power of I3−/I− towards Li2O2 oxidation increases, from −0.04 eV in DME, to −0.36 eV in DMA and −1.08 eV in DMSO. Using the linear free energy relationship that links thermodynamics and kinetics, one would expect that the kinetics of I3− against Li2O2 oxidation would significantly increase from DME, DMA to DMSO.
VIII. Solvent-Dependent Oxidizing Power of I3−/I− and I2/I3− Towards Li2O2
The solvent-dependent oxidizing power of I3/I− towards Li2O2 was examined by adding Li2O2 (0.1 M, Li2O2:I3−=2:1) to 50 mM I3− (50 mM of I2+0.2 M of LiI) in different solvents. The brown-colored solution became clear in DMA (<24 hours), DMSO (˜1 minute) and Me-Im (˜10 seconds), as shown in FIG. 2, panel a. On the other hand, the brown color became less pronounced for pyridine while no visible color change was found for DME and G4 after 24 hours. The color change observed for DMA, DMSO and Me-Im can be attributed to the reduction of I3− (dark brown) to I− (colorless). This hypothesis is supported by UV-vis spectroscopy of the liquid phase decanted from the reaction mixture after 24 hours, where characteristic peaks for I3− at 293 nm and 364 nm disappeared for DMA, DMSO and Me-Im while those for DME, G4 and pyridine remained, as shown in FIG. 2, panel b and FIGS. 22 and 23. The concentration changes of I3− during the reaction (24 hours) quantified using the absorbance of I3− solutions with known concentration (as detailed in FIGS. 12-14 increased with greater redox potentials of I3−/I− from G4, DME, pyridine to DMA (DMSO or Me-im) in FIG. 2, panel c. All the I3− was consumed in DMA, DMSO and Me-im and Raman spectra of the solid recovered after the reaction between Li2O2 and I3− revealed Li2O2 remained after the reaction as Li2O2 was 2 times over stoichiometric (FIG. 24). This trend is in agreement with the greater kinetics of Li2O2 oxidation by I3/I− with higher potentials in solvents such as DMA and DMSO which renders a higher thermodynamic driving force relative to Li+,O2/Li2O2(FIG. 1). Further support for Li2O2 oxidation by I3− in DMSO came from oxygen evolution as detected by gas chromatography (GC) accompanied with color changes of the solution after addition of Li2O2. GC measurements of oxygen evolution from solutions of DMSO (FIG. 3, FIG. 25) show that the amount of oxygen detected was comparable to that expected for oxidation of I3− by I3−+Li2O2→2Li++3I−+O2. Therefore, having solvents not only with higher AN to increase the potential of I3−/I− but also with higher DN to lower the potential of Li,O2/Li2O2 in solvents such as DMA and DMSO promotes the oxidizing power of I3−/I− towards Li2O2 as opposed to solvents like G4 and DME.
I3− in DME can oxidize Li2O2 in small part considering experimental uncertainty while previous studies showing that I3− cannot oxidize Li2O243,45-47 in DME. The more oxidizing I2 could fully oxidize synthetic Li2O2 in DMA and DMSO and I2 in DME reacted until all I2 was reduced to I3− via I2+Li2O2→2Li++2I−+O2 and I2+I−<I3− (FIG. 26). Reactions between commercial Li2O2 and I3−/I2 in DME proceeded to a lesser extent than synthetic Li2O2 as shown in FIG. 27. I3− in DME did not react at all with commercial Li2O2, whereas solutions of I2 (50 mM I2) and I5− (50 mM I2+25 mM LiI) proceeded until all higher order polyiodide species were reduced to I3−. Raman on the solution phase after the reaction revealed that only I3− species remained (FIG. 28). We attribute this difference in oxidizing power of polyiodide species against commercial and synthetic Li2O2 to the oxygen-rich, defective surface of Li2O2 formed through disproportionation which has been reported previously and is anticipated to be more readily oxidized than bulk Li2O2. A discussion of higher-order polyiodide species (such as I5− and I7) is presented below.
No solvent decomposition was detected for G4, DME and DMA while decomposed species from pyridine, DMSO and Me-Im were found in presence of Li2O2 and/or I3−. 1H NMR measurements of the solution phase decanted from the reaction mixture after 24 hours was used to detect protonated species produced after the addition of synthetic Li2O2. No changes were observed in G4, DME and DMA (FIG. 29, indicating no detectable solvent decomposition. On the other hand, a small peak at ˜2.95 ppm appeared for DMSO, which can be attributed to the presence of ˜6 mM dimethyl sulfone (DMSO2) (FIG. 30). This observation is in agreement with previous work showing that DMSO is chemically unstable in the presence of Li2O2-like and LiO2 species. In addition, changes were found for the 1H NMR peaks of Me-Im at ˜7.1 ppm and ˜7.7 ppm, splitting into two peaks and shifting downfield, respectively (FIG. 31), As comparable changes were found when a solution of I3− was prepared in Me-Im (without addition of Li2O2), we attribute this to changes in Me-Im proton exchange dynamics caused by the introduction of a Brønsted base (I−/I3−). Interactions between iodide species and Me-Im are known to be strong and lead to the iodination of Me-Im, which is supported by observed color-fading of I3− solutions of Me-Im over time, which was most pronounced in diluted samples for UV-Vis analysis (FIG. 32). Considering that the short duration of Li2O2 oxidation (<10 seconds) and long iodination reaction time (>weeks for a 50 mM solution without Li2O2) to render colorless solutions, the oxidation of Li2O2 by I3− to form I− dominates.
Discharging and charging of a Li—O2 battery with and without LiI as a redox mediator was performed, where the added I− was electrochemically oxidized to I3− and/or I2 during charge. DEMS cells were assembled using 0.5 M LiTFSI in diglyme (G2) or DMSO, with and without 0.1 M LiI, where added 0.1 M LiI could provide a theoretical maximum of 33 mM I3− and 50 mM I2, accounting for a maximum of 0.25 mAhr/cm2 of capacity. G2 was selected as an analogous solvent to DME with a lower vapor pressure, but lower viscosity than G4. Cells were first discharged at 0.05 mA/cm2geo for 20 hours to yield 1 mAhr/cm2 in capacity and only Li2O2 was detected by XRD (FIG. 33). While the addition of 0.1 M LiI did not lead to significant changes the discharge voltage, it markedly reduced the charging potential in both solvents. The entire charging voltage profile for G2 was sloped and the reduction of charging voltage can be attributed to electrochemical oxidation of I− to I3− (>2.98 VLi) and I3− to I2 (>3.59 VLi), with minute oxidation of Li2O2 by I3−. This hypothesis is supported by the DEMS measurements in G2, which showed no greater oxygen evolution rate with addition of I− upon charge, as shown in FIG. 4, panel c. On the other hand, the charging voltage profile with DMSO exhibited a plateau below the oxidation potential of I3− to I2 (3.90 VLi), which was accompanied with a significantly enhanced (˜two times) rate of oxygen evolution during charging relative to that without I−. The result confirmed that having the I3−/I− redox potential equal to or lower than Li+,O2/Li2O2 in solvents such as G2 and DME was unable to promote the oxidation of Li2O2 while those with greater potentials than that of Li+,O2/Li2O2 in solvents such as DMSO can facilitate Li2O2 oxidation kinetics to evolve O2 and lower charging overpotential of Li—O2 batteries.
IX. Solvent-Dependent Oxidizing Power of I3−/I− and I2/I3− Towards LiOH
The solvent-dependent oxidizing power of I3−/I− towards LiOH was examined by adding LiOH (2 times excess) to 50 mM I3− (50 mM of I2+0.2 M of LiI) in different solvents. The presence of water and I− can lead to the formation of LiOH on discharge, where water is consumed in this reaction. Thus we examine how oxidized iodide species can promote the oxidation of LiOH, beginning from anhydrous conditions. The brown-colored solution became clear in DMA (˜48 hours) and DMSO (<1 hour) (Me-Im with ˜10 minutes). On the other hand, no clearly visible color change was found for pyridine, DME and G4 after 48 hours. The color change observed for DMA, DMSO and Me-Im can be attributed to the reduction of I3− (dark brown) to I− (colorless). This hypothesis is supported by UV-vis spectroscopy of the liquid phase decanted from the reaction mixture after 48 hours, where characteristic peaks for I3− at 293 nm and 364 nm disappeared for DMA, DMSO and Me-Im while those for DME, G4 and pyridine remained, as shown in FIG. 5, panel a and FIG. 34 and FIG. 35. The concentration changes of I3− during the reaction (48 hours) was quantified using the absorbance of I3− solutions with known concentration (as detailed in FIGS. 12-14) increased with greater redox potentials of I3−/I− from G4, DME, pyridine to DMA (DMSO or Me-im) in FIG. 4, panel a. All the I3− was consumed in DMA, DMSO and Me-im while nearly no I3− was consumed in G4, DME and Pyridine. Raman spectra of the solid recovered after the reaction between LiOH and I3− revealed only LiOH as expected from LiOH being 2 times overstoichiometric (FIG. 36). Similarly, the addition of LiOH to the more oxidizing I2 in DMA and DMSO led to complete consumption of I2 while the consumption of I2 stopped when the reaction had proceeded to a point when all I2 would be expected to be converted to I3− in DME, as shown in FIG. 5, panel b.
Unfortunately, the consumption of I3− by reacting with LiOH in solvents such as DMSO did not yield oxygen evolution as shown from GC measurements after the addition of LiOH (FIG. 37). To assess reaction products generated during the reaction between I3− and LiOH, the reaction between an excess of I3− (2 times) and LiOH in DMSO was allowed to carry out for more than one week. Raman (FIG. 5, panel c) and XRD (FIG. 38) revealed LiIO3 without LiOH and LiOH—H2O. The formation of LiIO3 can come from the following reaction: 3I3−+6LiOH→8I−+5Li++3H2O+LiIO3. The reaction between I3− and LiOH (I3−+2LiOH→2I−+2Li++H2O+IO−) has been well studied in aqueous systems as well as the disproportionation of hypoiodite (IO−) to IO3−71-73 (3IO−→2I−+IO3−). In-situ Raman spectroscopy revealed a vibration at 430 cm−1 previously attributed to IO−74 (FIG. 39), which provided support to the proposed mechanism involving IO−. The formation of LiIO3 is also supported by previous observations by Burke et al.43 based on charging of a cell with LiI in DME with LiOH formed during discharge. The trend in FIG. 5, panel a can be attributed to the greater kinetics of LiOH oxidation by I3−/I− with higher potentials in solvents such as DMA and DMSO to render higher thermodynamic driving force relative to Li+,IO3−,H2O/LiOH,I− (FIG. 5, panel a).
1H NMR analysis and iodometric titration of the solution phase before and after reaction with 50 mM I3−/I2 further confirmed the proposed reaction mechanism for the formation of LiIO3. A H2O peak became visible following the addition of LiOH to DMA (FIG. 27), DMSO (FIG. 6, panel a, FIG. 27) and Me-Im (FIG. 27) with 50 mM I3− for 48 hours. No H2O was detected in the liquid solution from mixing DMSO (without oxidized iodide species) with LiOH for 48 hours while an H2O peak at 3.36 and 3.30 ppm was detected after reacting LiOH with 50 mM I3− and 50 mM I2 in DMSO, respectively. The upfield shift of H2O found for I2 compared to I3− can be attributed to the larger quantity of LiI in the I3− solution as shown previously for 1H NMR chemical shift of H2O induced by interaction with I− in DME44. In contrast, no H2O and other changes were observed in the 1H NMR spectra of G4 and DME (FIG. 27) following the reaction between Li2O2 and I3− while pyridine showed the emergence of some small peaks (FIG. 27) which we attribute to solvent decomposition. The concentration of H2O was quantified (±5 mM) using an internal MeCN reference, where ˜80 mM was found in DMA and ˜60 mM was found in DMSO and Me-Im. The amount of water detected was greater than that (50 mM) expected from the proposed reaction, 3I3−+6LiOH→8I−+5Li++3H2O+LiHO3, where the difference might be attributed to solvent decomposition. DMSO2 at 2.95 ppm of ˜18 mM was found for DMSO and the peak changes of Me-Im (FIG. 27, FIG. 29) can be attribute to the previously discussed interactions with I−/I3−70. The amount of the iodate species detected with iodometric titration (8.2 mM and 6.4 mM for reactions with I3− and I2 in DMSO, respectively) was close to that (16.7 mM) expected for 3I3−+6LiOH→8I−+5Li++3H2O+LiHO3, as shown in FIG. 6, panel b. The difference can be attributed to the decomposition of DMSO by IO− via IO−+(CH3)2SO→I−+(CH3)2SO2, having 18.5 mM and 24.9 mM DMSO2 (18.5/24.9 mM IO− consumed) in this decomposition reaction for reactions with I3− and I2, respectively. A similar oxidation of DMSO to DMSO2 was reported by Liu et al. in the ruthenium-catalyzed oxidation of LiOH. Therefore, combined spectroscopic data from 1H NMR, Raman, GC and iodometric titration show that the oxidation of LiOH by oxidized iodide species such as I3− leads to the formation of an IO− intermediate, which can disproportionate to form LiIO3 as the major product and attack solvent molecules to form species such as DMSO2. This reaction mechanism does not lead to the formation of O2 gas as some have reported previously.
The proposed oxidation mechanism of LiOH in the presence of oxidized iodide species is supported by galvanostatic charging and DEMS measurements (FIG. 7) of preloaded LiOH electrodes with a solid Li-conducting separator to eliminate shuttling, charged in 0.5 M LiTFSI G2 (FIG. 7, panel a, panel c) and DMSO (FIG. 7 panel b, panel d) with and without 0.1 M LiI addition (in cases where no LiI was added, an additional 0.1 M LiTFSI was added to fix the total Li+ concentration at 0.6 M). Of significance, there was no observable oxygen generation in either G2 or DMSO, which supports the proposed oxidation of LiOH by oxidized iodide species to form lithium iodate. The majority of the charging plateau took place above the I3−/I2 redox transition in G2 (comparable to that in DME), indicating that I3− could not oxidize LiOH in glymes but I2 could (FIG. 5, panel a, panel b). On the other hand, significant capacity was noted below the I3−/I2 redox transition in DMSO, corresponding to the formation of LiIO3 from I3−. XRD of the electrodes after charging (FIG. 40) indicated that not all LiOH was removed, which is consistent with the calculated charging capacity based on the mass of deposited LiOH (7.3 and 5.2 mAhr/cm2 for G2 and DMSO, respectively). However, the observed capacity is significantly larger than the maximum calculated capacity based on the oxidation of LiI (˜0.25 mAhr/cm2), indicating consumption of LiOH during charge. We postulate the incomplete oxidation of LiOH in-situ may relate to either slow kinetics of oxidation of LiOH by iodide species (shown to be much slower than the oxidation of Li2O2 in ex-situ experiments) and/or the passivation of the LiOH surface by insoluble LiIO3. Leftover LiOH after charging is consistent with the observations of Qiao et al., however, using ex-situ reactions and a solid Li-conducting separator to eliminate shuttling, we are able to demonstrate that LiOH is still active during the charging process and not inactive as suggested by Qiao et al.
The reaction of Li2O2 by oxidized iodide species leads to O2 gas evolution whereas LiOH is oxidized to IO−, which can then either disproportionate to form LiIO3 or attack solvent molecules. We estimated the free energy of formation for LiIO3 by combining the computed enthalpy of LiIO3 formation from Huang et al. with approximated entropy estimated from KIO3, where full derivation is available in the Supporting Information. The formation of LiIO3 from Li2O2 and LiOH was found thermodynamically at 2.21 and 2.97 VLi, respectively as shown in FIG. 41. The oxidation of Li2O2 to O2 in the presence of oxidized iodide species is anticipated to occur when the polyiodide equilibrium potential is higher than 2.96 VLi. Similarly, LiOH is expected to be oxidized to O2 by polyiodide species with redox potentials above 3.35 VLi. The estimated thermodynamics predicting oxidation of LiOH to LiIO3 at voltages greater than ˜3 VLi is in good agreement with measured values in this work. Since I3− is unable to oxidize LiOH in DME, but is able to oxidize LiOH in DMA, we would anticipate the formation of LiIO3 from LiOH to occur in the range of 2.98-3.14 VLi. Since LiIO3 is thermodynamically favorable to form from both Li2O2 and LiOH, it is proposed that the evolution of O2 without the formation of LiIO3 upon oxidation of Li2O2 by oxidized iodide species can be attributed to slow kinetics of O—O dissociation needed to form IO− and subsequently LiIO3. Further support to this hypothesis came from the experiments with KO2 and Li2O. The oxidation of KO2 by I3− in G2 was found to readily evolve O2 using DEMS, and have a color change from brown to colorless in the ex situ chemical reaction (FIG. 42) while the chemical reactions between Li2O and I3− in DMSO led to LiIO3 (FIG. 43).
The formation of LiIO3 from the oxidation of LiOH by I3−/I2 indicates a significant incompatibility between LiI as a redox mediator for Li—O2 batteries and any water present in the electrolyte. As previous demonstrated, the presence of LiI in DME-based electrolytes decreases the deprotonation energy of water, leading to the formation of LiOH (even with H2O content <40 ppm). In this work, we have demonstrated that the oxidation of LiOH leads to the formation of LiIO3, and not the reversible formation of O2, as well as the regeneration of water. Therefore, with cycling in the presence of any water (even contaminate levels of water), through the action of consuming water to form LiOH on discharge and oxidizing it to LiIO3 and reforming water on charge, LiI will be converted to LiIO3, leading to the irreversible loss of the redox mediator. The full implication of this reaction on the cycle life of cells with LiI as a redox mediator is beyond the scope of this work, but this work highlights a significant issue that needs to be addressed in order for LiI to be practically implemented as a redox mediator in a Li—O2 battery.
The role of LiI on the charging process of Li—O2 batteries was examined by systemically studying the solvent-dependent oxidizing power of I3−/I− and I2/I3− towards Li2O2 and LiOH. The oxidizing power of I3−/I− and I2/I3− towards Li2O2 and LiOH was examined chemically by examining the consumption of I3− upon addition of synthetic Li2O2, where the liquid reaction product was examined using UV-vis spectroscopy and 1H NMR, the solid reaction products were studied by Raman spectroscopy and XRD and the gaseous products were assessed using gas chromatography. In addition, the role of I− on the charging of Li—O2 batteries and LiOH pre-loaded cells was examined using DEMS, where the amount of oxygen generated was quantified. We have shown that I3−/I− shifts towards higher potentials in solvents with higher dielectric constant and AN, suggesting stronger solvation of I− ions, whereas the I2/I3− potential was observed to be largely solvent independent in the considered solvents. This strong solvation of I− ions, coupled with a strong solvation of Li+ ions in solvents like DMA, DMSO and Me-Im was found to increase the oxidizing power of I3−/I−, allowing I3− to effectively oxidize Li2O2 to generate O2, which was supported by chemical and electrochemical experiments. In solvents with weaker solvation of I− and Li+ (such as DME and G4), the more oxidizing I2/I3− redox couple was needed before Li2O2 could be fully oxidized to O2. The oxidation of LiOH by I3− was also found to be solvent dependent, where no reaction was observed in G4, DME and pyridine while the reaction proceeded to completion in DMA, DMSO and Me-Im where the I3−/I− redox potential was above ˜3.1 VW. No O2 was detected from the oxidation of LiOH by I3− using gas chromatography and the charging of pre-loaded LiOH electrodes in DEMS, but instead, the oxidation of LiOH was found to produce water and a hypoiodite (IO−) intermediate, which could either disproportionate to form LiIO3 or attack solvent molecules and result in decomposition products such as dimethyl sulfone (DMSO2). The selectivity between O2 and the thermodynamically preferred LiIO3 can be governed by a kinetic barrier relating to O—O bond dissociation and this kinetic barrier prevents IO− formation, allowing for the evolution of gaseous O2 when oxidizing Li2O2, which was supported by reactions between oxidized iodide species and KO2/Li2O. This work highlights a significant incompatibility between LiI as a redox mediator for Li—O2 batteries and even trace amounts of water in the electrolyte, which may lead to the consumption of the LiI redox mediator to form LiIO3 with cycling.
Demonstrating the Charging Reaction of the Proposed Chemistry
The reaction between I3− and LiOH was found to be solvent-dependent, with I3− being fully consumed in DMA, DMSO and Me-Im but little to no reaction occurring in G4, DME and pyridine. The solvent-dependent oxidizing power of I3−/I− towards LiOH was examined by adding commercial LiOH (0.2 μmol, LiOH:I3−=4:1) to 1 mL of 50 mM I3− (50 mM of I2+0.2 M of LiI, I−:I2=4:1) in different solvents. The brown-colored solution became clear in DMA (˜48 hours), DMSO (˜1 hour) and Me-Im (˜10 minutes). This color change could be attributed to the reduction of I3− (dark brown) to I− (colorless) as revealed by UV-vis spectroscopy of the liquid phase decanted from the reaction mixture after 48 hours (FIG. 5B, panel a, FIG. 35). On the other hand, no color change was found for pyridine, DME and G4 after 48 hours as evidenced by the characteristic peaks for I3− at 293 nm and 364 nm remaining after the reaction with LiOH (FIG. 5B, panel a, FIG. 35). The consumption of I3− after the reaction for 48 hours was quantified using the absorbance of I3− with known concentrations (FIGS. 11-14). All the I3− was consumed in DMA, DMSO and Me-Im while nearly no I3− was consumed in G4, DME and Pyridine, as shown in FIG. 1A. Similarly, as shown in FIG. 1B, the addition of LiOH to the more oxidizing I2 in DMA and DMSO led to complete consumption of I2 while in DME, the reaction stopped after only I3− remained (FIGS. 5B, panel b, FIG. 47), resulting from the previously discussed association between I− generated by the reaction and the remaining I2 via I2+I−↔I3−. Anhydrous LiOH synthesized via the disproportionation of KO2 in a two times excess of LiTFSI in MeCN with added water (FIG. 48) was found to exhibit similar reactivity to commercial anhydrous LiOH in the presence of I3−, with a brown-colored 50 mM I3− solution becoming clear in DMA (˜96 hours) and DMSO (˜4 hour), but no visible color change in DME after 96 hours (FIG. 49).
The reaction between I3− and anhydrous LiOH in solvents such as DMSO did not yield oxygen evolution as shown from GC measurements with commercial LiOH (FIG. 37). As expected due to the excess of LiOH (LiOH:I3−=4:1), Raman spectra of the solid recovered after the reaction between LiOH and I3− in all solvents revealed anhydrous LiOH as the dominant phase remaining after the reaction (FIG. 36). To further probe potential solid reaction products between I3− and LiOH, the I3− excess reaction with commercial anhydrous LiOH (LiOH:I3−=1:1) in DMSO was performed for more than one week. Raman (FIG. 5B, panel C) and XRD (FIG. 38) of the solid recovered revealed LiIO3 only without LiOH remaining. The presence of LiIO3 has been previously reported by Burke et al., Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Letters 2016; 1:747-56, which is incorporated by reference in its entirety. upon charging of cells having LiOH formed during discharge with LiI in DME. The formation of LiIO3 can come from the following reaction: 3I3−+6LiOH→8I−+5Li++3H2O+LiIO3, which can include the reaction between I3− and LiOH to generate hypoiodite (I3−+2LiOH→2I−+2Li++H2O+IO−) and the disproportionation of hypoiodite (IO−) to iodate IO3− (3IO−→2I−+IO3−). See, Gerritsen C M, Gazda M, Margerum D W. Non-metal redox kinetics: hypobromite and hypoiodite reactions with cyanide and the hydrolysis of cyanogen halides. Inorganic Chemistry 1993; 32:5739-5748, Lengyel I, Epstein I R, Kustin K. Kinetics of iodine hydrolysis. Inorganic Chemistry 1993; 32:5880-5882; and Xie Y, McDonald M R, Margerum D W. Mechanism of the Reaction between Iodate and Iodide Ions in Acid Solutions (Dushman Reaction). Inorganic Chemistry 1999; 38:3938-40, each of which is incorporated by reference in its entirety. The presence of IO− is supported by the presence of a vibration at 430 cm−1 previously attributed to IO− (FIG. 39) by in-situ Raman spectroscopy of a solution of commercial anhydrous LiOH with I3− in DMSO. See, Wren J C, Paquette J, Sunder S, Ford B L. Iodine chemistry in the +1 oxidation state. II. A Raman and uv-visible spectroscopic study of the disproportionation of hypoiodite in basic solutions. Canadian Journal of Chemistry 1986; 64:2284-96, which is incorporated by reference in its entirety. The thermodynamic driving force to form LiIO3 from LiOH (3I3−+6LiOH→8I−+5Li++3H2O+LiIO3) is much greater than that for oxygen evolution (2I3−+4LiOH→6I−+4Li++2H2O+O2), and increases with greater redox potentials of I3−/I− on the Li+/Li scale from G4/DME, to DMA, to DMSO, to Me-Im (FIG. 10, FIG. 18). Of particular significance is the case of DMA, where full consumption of I3− was observed and the reaction to form LiIO3 is predicted to be spontaneous (Erxn=−ΔGrxn/6F=+0.17V) whereas the reaction to form O2 is not (Erxn=−ΔGrxn/4F=−0.21V), which further supports the preference for LiIO3 formation instead of O2 evolution. Similar trends were found for I2/I3− where increased thermodynamic driving force correlated with increased consumption of I2 in DME, DMA and DMSO (FIG. 5B, panel b).
Quantifications through 1H NMR analysis of the solution phase and iodometric titration after reaction with 50 mM I3−/I2 further confirmed the proposed reaction mechanism for the formation of LiIO3. A H2O peak became visible following the addition of LiOH to DMA (FIG. 30), DMSO (FIG. 6, panel a, FIG. 16) and Me-Im (FIG. 29) with 50 mM I3− for 48 hours. No H2O was detected in the liquid phase from mixing DMSO (without oxidized iodide species) with LiOH while an H2O peak at 3.36 and 3.30 ppm was detected after reacting LiOH with 50 mM I3− and 50 mM I2 in DMSO, respectively (FIG. 6, panel a). The upfield shift of H2O found for 12 compared to I3− can be attributed to the larger quantity of I− in the solution following the reaction with I3−, as shown previously for changes in the 1H NMR chemical shift of H2O induced by interactions with I− in DME. See, for example, Tulodziecki M, Leverick G M, Amanchukwu C V, Katayama Y, Kwabi D G, Barde F, et al. The role of iodide in the formation of lithium hydroxide in lithium-oxygen batteries. Energy Environ Sci 2017; 10:1828-42, which is incorporated by reference in its entirety. In contrast, no H2O or other changes were observed in the 1H NMR spectra of G4 and DME (FIG. 29) following the reaction between LiOH and I3− while pyridine showed the emergence of some small peaks (FIG. 29) which we attribute to solvent decomposition. In addition to the formation of H2O, reactions between LiOH and I3−/I2 in DMSO resulted in DMSO2 (˜2.95 ppm, FIG. 30) with quantity of 18 μmol and 25 μmol for reactions with I3− and I2, respectively (FIG. 6, panel a and panel b) while Me-Im experienced peak changes (FIG. 29) which can be attributed to the previously discussed interactions with I−/I3−. See, for example, Schutte L, Kluit P P, Havinga E. The substitution reaction of histidine and some other imidazole derivatives with iodine. Tetrahedron 1966; 22:295-306, which is incorporated by reference in its entirety. The amount of iodate species detected with iodometric titration (8.2 μmol and 6.4 μmol for reactions with I3− and I2 in DMSO, respectively) was close to that expected (16.7 μmol) for 3I3−+6LiOH→8I−+5Li++3H2O+LiIO3, as shown in FIG. 6, panel b. The difference can be attributed to the decomposition of DMSO by a IO− intermediate via IO−+(CH3)2SO→I−+(CH3)2SO2 which accounts for 18.5/24.9 μmol of IO− consumed in reactions with I3− and I2, respectively, which otherwise could have disproportionated to form LiIO3. A similar oxidation of DMSO to DMSO2 from intermediates of LiOH oxidation was reported by Liu et al. in a ruthenium-catalyzed Li—O2 battery system. See, for example, Liu T, Liu Z, Kim G, Frith J T, Garcia-Araez N, Grey C. Understanding LiOH Chemistry in a Ruthenium Catalyzed Li—O2 Battery. Angewandte Chemie International Edition 2017; 56:16057-62, which is incorporated by reference in its entirety. Therefore, combined spectroscopic data from 1H NMR, Raman, GC and iodometric titration show that the reaction between LiOH and oxidized iodide species such as I3− leads to the formation of an IO− intermediate, which can disproportionate to form LiIO3 as the major product and attack solvent molecules to form species such as DMSO2. This reaction mechanism does not lead to the formation of O2 gas as some have reported previously. See, for example, Liu T, Leskes M, Yu W, Moore A J, Zhou L, Bayley P M, et al. Cycling Li—O2 batteries via LiOH formation and decomposition. Science 2015; 350:530-3; and Zhu Y G, Liu Q, Rong Y, Chen H, Yang J, Jia C, et al. Proton enhanced dynamic battery chemistry for aprotic lithium-oxygen batteries. Nature Communications 2017; 8:14308, each of which is incorporated by reference in its entirety.
Without being bound to any specific theory, the proposed reaction mechanism of LiOH in the presence of oxidized iodide species is supported by galvanostatic charging and DEMS measurements (FIG. 7) of pre-loaded commercial, anhydrous LiOH electrodes with a solid Li-conducting separator to eliminate shuttling, charged in 0.5 M LiTFSI G2 (FIG. 7, panel a and panel c) and DMSO (FIG. 7, panel B and panel D) with and without 0.1 M LiI addition (in cases where no LiI was added, an additional 0.1 M LiTFSI was added to fix the total Li+ concentration at 0.6 M). Of significance, there was no observable oxygen generation in either G2 or DMSO, which supports the proposed reaction with LiOH by oxidized iodide species to form LiIO3. The majority of the charging plateau took place above the I3−/I2 redox transition in G2 (comparable to DME/G4), indicating that I3− could not react with LiOH in glymes but I2 could, which is consistent with ex-situ chemical reactions (FIG. 5B, panel a and panel b). On the other hand, significant capacity was noted below the I3−/I2 redox transition in DMSO, corresponding to the formation of LiIO3 from I3−. XRD of the electrodes after charging (FIG. 40) indicated that not all LiOH was removed, which is consistent with the calculated charging capacity based on the mass of deposited LiOH (7.3 and 5.2 mAhr/cm2 for G2 and DMSO, respectively) being considerably higher than the achieved charging capacity (1.0 and 1.6 mAhr/cm2 for G2 and DMSO, respectively). However, the observed capacity is significantly larger than the maximum calculated capacity based on the oxidation of LiI (˜0.25 mAhr/cm2), indicating consumption of LiOH during charge. We postulate the incomplete oxidation of LiOH in-situ may relate to either slow kinetics of reaction with LiOH by oxidized iodide species and/or the passivation of the LiOH surface by insoluble LiIO3. Leftover LiOH after charging is consistent with the observations of Qiao et al., however, using ex-situ reactions and a solid Li-conducting separator to eliminate shuttling, we are able to demonstrate that LiOH is still active during the charging process and not inactive as suggested by Qiao et al. See, for example, Qiao Y, Wu S, Sun Y, Guo S, Yi J, He P, et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Letters 2017; 2:1869-78, which is incorporated by reference in its entirety.
Demonstrating the Discharge Process
Electrodes were prepared using commercially available LiIO3 from Sigma-Aldrich which was received and maintained in an α crystal structure (FIG. 50). After grinding by hand, the particle size obtained was 10-200 um (FIG. 50). Cells were constructed using a lithium iron phosphate (LFP) counter electrode and potentials were converted to a Li scale using the LFP potential of 3.45VLi. O'hara glass was using as a solid li-ion conducting membrane to prevent shuttling of iodide species between the electrodes (FIG. 51). Discharges were performed using both composite electrodes (FIG. 52) and drop cast electrodes (FIG. 53). Composite electrodes were prepared by grinding LiIO3 with carbon (SuperP and/or Vulcan carbon (VC)) and a polymer binder (PvDF) (FIG. 52) and deposited onto aluminum foil using a solvent (such as dimethoxymethane or NMP). Other electrodes were prepared by drop casting a slurry of LiIO3 and Vulcan carbon with a PTFE binder onto a carbon paper substrate (Toray 60). Electrodes were discharged under Argon environment in an electrolyte consisting of 5-10 w % H2O in 1,2-dimethoxyethane or acetonitrile. 0.5M LiTFSI was added as a conducting salt with an additional 0.1M LiI added to some electrolytes. Discharge profiles consisted of a sloped voltage profile from 2.2-2.7V vs Li (FIG. 52, FIG. 53). Electrodes recovered following discharge were analyzed using Raman (FIG. 54) and XRD (FIG. 55) and show the clear formation of anhydrous LiOH during the discharge process. Additional SEM images of pristine electrodes (FIG. 56) and discharged electrodes (FIG. 57) demonstrate clear morphological differences following discharge.
Additional work was carried out to understand the role of the electrolyte and water content on the discharge process. Discharges were carried out using drop cast LiIO3 electrodes in 1,2-dimethoxyethane (DME), acetonitrile (MeCN) and 1,4-dioxane (DOL), all with added 10 v % H2O. Discharge capacity and voltage increased from DOL, to DME, to MeCN, reaching the full anticipated discharge capacity of 880 mAhr/gLiIO3 (FIG. 58). Both XRD and Raman were performed on the discharged electrodes and support the anticipated removal of LiIO3 and formation of LiOH (FIG. 59). Acid base titrations were performed to quantify the amount of LiOH formed during discharged by adding 5 mL DI water to a 20 mL vial container either the discharged electrode or separator and allowing all LiOH to dissolve for 30 minutes. Titration was performed using 10 mM HCl and a phenolphthalein indicator. Immediately following the acid-base titration, ˜50 mg KI and 0.5 mL 5M H2SO4 was added to the vial and an iodometric titration was performed using 10 mM Na2S2O3 solution, adding a 1 w %/V starch solution near the end point. The iodometric titration results enabled the quantification of the remaining LiIO3 after the discharge process. The consumption of LiIO3 and formation of LiOH during the discharge process in the different solvents supports the proposed 6e− reduction of LiIO3 to LiOH (FIG. 58). Similar discharges were performed in DME with 1 v %, 5 v %, 10 v % and 20 v % H2O added (FIG. 60). Titrations (FIG. 60) and XRD/Raman characterization (FIG. 61) again support the proposed discharge mechanism.
Substantial morphological changes in the electrode were observed with SEM (FIG. 62), suggesting that LiIO3 and/or LiOH may be soluble in the electrolyte. In order to assess the solubility of LiIO3 and LiOH in the electrolytes, mixtures of DME, DOL and MeCN with added water were allowed to saturate with LiOH or LiIO3 under stirring for 3 days. The resulting mixtures were centrifuged and the decanted liquid was assessed using an inductively couple plasma technique to quantify the amount of dissolved lithium (FIG. 63). It was observed that upon the addition of water at 10 v % and 20 v %, both LiOH and LiIO3 could be solubilized up to ˜10 mM. The high overpotential during the discharge process are believed to stem from mass transport limitations. This hypothesis is supported by a linear trend between the overpotential on discharged (where the thermodynamic potential is 2.97V vs Li+/Li) and the logarithm of viscosity divided by the solubility of LiIO3 in the electrode (FIG. 64). Overall, the titration, SEM and discharge profile results are all consistent with a mechanism which starts with dissolution of LiIO3, then a 6e− reduction of IO3− to OH− and I−, followed by precipitation of LiOH onto the electrode surface and separator (FIG. 65).
REFERENCES (EACH OF THE FOLLOWING REFERENCES IS INCORPORATED BY REFERENCE IN ITS ENTIRETY)
- 1. Aurbach, D., McCloskey, B. D., Nazar, L. F. & Bruce, P. G. Advances in understanding mechanisms underpinning lithium-air batteries. Nat. Energy 1, 16128 (2016).
- 2. Bruce, P. G., Freunberger, S. A., Hardwick, L. J. & Tarascon, J.-M. Li—O2 and Li—S batteries with high energy storage. Nat. Mater. 11, 19-29 (2011).
- 3. Kwabi, D. G. et al. Materials challenges in rechargeable lithium-air batteries. MRS Bull. 39, 443-452 (2014).
- 4. Lu, Y.-C. et al. Lithium-oxygen batteries: bridging mechanistic understanding and battery performance. Energy Environ. Sci. 6, 750 (2013).
- 5. Viswanathan, V. et al. Li—O2 Kinetic Overpotentials: Tafel Plots from Experiment and First-Principles Theory. J. Phys. Chem. Lett. 4, 556-560 (2013).
- 6. Kwabi, D. G. et al. Experimental and Computational Analysis of the Solvent-Dependent O2/Li+—O2− Redox Couple: Standard Potentials, Coupling Strength, and Implications for Lithium-Oxygen Batteries. Angew. Chem. Int. Ed. 55, 3129-3134 (2016).
- 7. Johnson, L. et al. The role of LiO2 solubility in O2 reduction in aprotic solvents and its consequences for Li—O2 batteries. Nat. Chem. 6, 1091-1099 (2014).
- 8. Sharon, D. et al. Mechanistic Role of Li+ Dissociation Level in Aprotic Li—O2 Battery. ACS Appl. Mater. Interfaces 8, 5300-5307 (2016).
- 9. Burke, C. M., Pande, V., Khetan, A., Viswanathan, V. & McCloskey, B. D. Enhancing electrochemical intermediate solvation through electrolyte anion selection to increase nonaqueous Li—O2 battery capacity. Proc. Natl. Acad. Sci. 112, 9293-9298 (2015).
- 10. Kwabi, D. G. et al. Controlling Solution-Mediated Reaction Mechanisms of Oxygen Reduction Using Potential and Solvent for Aprotic Lithium-Oxygen Batteries. J. Phys. Chem. Lett. 7, 1204-1212 (2016).
- 11. Mitchell, R. R., Gallant, B. M., Shao-Horn, Y. & Thompson, C. V. Mechanisms of Morphological Evolution of Li2O2 Particles during Electrochemical Growth. J. Phys. Chem. Lett. 4, 1060-1064 (2013).
- 12. McCloskey, B. D. et al. Combining Accurate O2 and Li2O2 Assays to Separate Discharge and Charge Stability Limitations in Nonaqueous Li—O2 Batteries. J. Phys. Chem. Lett. 4, 2989-2993 (2013).
- 13. Lim, H.-D. et al. Rational design of redox mediators for advanced Li—O2 batteries. Nat. Energy 1, 16066 (2016).
- 14. Lim, H.-D. et al. Superior Rechargeability and Efficiency of Lithium-Oxygen Batteries: Hierarchical Air Electrode Architecture Combined with a Soluble Catalyst. Angew. Chem. Int. Ed. 53, 3926-3931 (2014).
- 15. Liu, T. et al. Cycling Li—O2 batteries via LiOH formation and decomposition. Science 350, 530-533 (2015).
- 16. Bergner, B. J., Schurmann, A., Peppler, K., Garsuch, A. & Janek, J. TEMPO: A Mobile Catalyst for Rechargeable Li—O2 Batteries. J. Am. Chem. Soc. 136, 15054-15064 (2014).
- 17. Kwak, W.-J. et al. Li—O2 cells with LiBr as an electrolyte and a redox mediator. Energy Env. Sci 9, 2334-2345 (2016).
- 18. Kwak, W.-J. et al. Understanding the behavior of Li-oxygen cells containing LiI. J Mater Chem A 3, 8855-8864 (2015).
- 19. Chen, Y., Freunberger, S. A., Peng, Z., Fontaine, O. & Bruce, P. G. Charging a Li—O2 battery using a redox mediator. Nat. Chem. 5, 489-494 (2013).
- 20. Sun, D. et al. A Solution-Phase Bifunctional Catalyst for Lithium-Oxygen Batteries. J. Am. Chem. Soc. 136, 8941-8946 (2014).
- 21. Feng, N., He, P. & Zhou, H. Enabling Catalytic Oxidation of Li2O2 at the Liquid-Solid Interface: The Evolution of an Aprotic Li—O2 Battery. ChemSusChem 8, 600-602 (2015).
- 22. Kundu, D., Black, R., Adams, B. & Nazar, L. F. A Highly Active Low Voltage Redox Mediator for Enhanced Rechargeability of Lithium-Oxygen Batteries. ACS Cent. Sci. 1, 510-515 (2015).
- 23. Torres, W. R., Herrera, S. E., Tesio, A. Y., Pozo, M. del & Calvo, E. J. Soluble TTF catalyst for the oxidation of cathode products in Li—Oxygen battery: A chemical scavenger. Electrochimica Acta 182, 1118-1123 (2015).
- 24. Wu, S., Tang, J., Li, F., Liu, X. & Zhou, H. Low charge overpotentials in lithium-oxygen batteries based on tetraglyme electrolytes with a limited amount of water. Chem Commun 51, 16860-16863 (2015).
- 25. Zhu, Y. G. et al. Dual redox catalysts for oxygen reduction and evolution reactions: towards a redox flow Li—O2 battery. Chem Commun 51, 9451-9454 (2015).
- 26. Pande, V. & Viswanathan, V. Criteria and Considerations for the Selection of Redox Mediators in Nonaqueous Li—O2 Batteries. ACS Energy Lett. 2, 60-63 (2016).
- 27. Yao, K. P. C. et al. Utilization of Cobalt Bis(terpyridine) Metal Complex as Soluble Redox Mediator in Li—O2 Batteries. J. Phys. Chem. C 120, 16290-16297 (2016).
- 28. Zeng, X. et al. Enhanced Li—O2 battery performance, using graphene-like non-derived carbon as the cathode and adding LiI in the electrolyte as a promoter. Electrochimica Acta 200, 231-238 (2016).
- 29. Zhang, W. et al. Promoting Li2O2 oxidation via solvent-assisted redox shuttle process for low overpotential Li—O2 battery. Nano Energy 30, 43-51 (2016).
- 30. Zhang, T., Liao, K., He, P. & Zhou, H. A self-defense redox mediator for efficient lithium-O2 batteries. Energy Env. Sci 9, 1024-1030 (2016).
- 31. Xu, J.-J. et al. Synthesis of Perovskite-Based Porous La0.75Sr0.25 MnO3 Nanotubes as a Highly Efficient Electrocatalyst for Rechargeable Lithium-Oxygen Batteries. Angew. Chem. Int. Ed. 52, 3887-3890 (2013).
- 32. Yao, K. P. C. et al. Solid-state activation of Li2O2 oxidation kinetics and implications for Li—O2 batteries. Energy Environ. Sci. 8, 2417-2426 (2015).
- 33. Yin, Y.-B., Xu, J.-J., Liu, Q.-C. & Zhang, X.-B. Macroporous Interconnected Hollow Carbon Nanofibers Inspired by Golden-Toad Eggs toward a Binder-Free, High-Rate, and Flexible Electrode. Adv. Mater. 28, 7494-7500 (2016).
- 34. Li, L. & Manthiram, A. O- and N-Doped Carbon Nanowebs as Metal-Free Catalysts for Hybrid Li-Air Batteries. Adv. Energy Mater. 4, 1301795 (2014).
- 35. Shui, J. et al. Nitrogen-Doped Holey Graphene for High-Performance Rechargeable Li—O2 Batteries. ACS Energy Lett. 1, 260-265 (2016).
- 36. Xu, J.-J., Wang, Z.-L., Xu, D., Zhang, L.-L. & Zhang, X.-B. Tailoring deposition and morphology of discharge products towards high-rate and long-life lithium-oxygen batteries. Nat. Commun. 4, (2013).
- 37. Kwon, H.-M. et al. Effect of Anion in Glyme-based Electrolyte for Li—O2 Batteries: Stability/Solubility of Discharge Intermediate. Chem. Lett. 46, 573-576 (2017).
- 38. Wong, R. A. et al. Critically Examining the Role of Nanocatalysts in Li—O2 Batteries: Viability toward Suppression of Recharge Overpotential, Rechargeability, and Cyclability. ACS Energy Lett. 3, 592-597 (2018).
- 39. Radin, M. D. & Siegel, D. J. Charge transport in lithium peroxide: relevance for rechargeable metal-air batteries. Energy Environ. Sci. 6, 2370 (2013).
- 40. Bergner, B. J. et al. Understanding the fundamentals of redox mediators in Li—O2 batteries: a case study on nitroxides. Phys. Chem. Chem. Phys. 17, 31769-31779 (2015).
- 41. Bergner, B. J. et al. How To Improve Capacity and Cycling Stability for Next Generation Li—O2 Batteries: Approach with a Solid Electrolyte and Elevated Redox Mediator Concentrations. ACS Appl. Mater. Interfaces 8, 7756-7765 (2016).
- 42. Lee, D. J., Lee, H., Kim, Y.-J., Park, J.-K. & Kim, H.-T. Sustainable Redox Mediation for Lithium-Oxygen Batteries by a Composite Protective Layer on the Lithium-Metal Anode. Adv. Mater. 28, 857-863 (2016).
- 43. Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016).
- 44. Tulodziecki, M. et al. The role of iodide in the formation of lithium hydroxide in lithium-oxygen batteries. Energy Env. Sci (2017). doi:10.1039/C7EE00954B
- 45. Qiao, Y. et al. Unraveling the Complex Role of Iodide Additives in Li—O2 Batteries. ACS Energy Lett. 1869-1878 (2017). doi:10.1021/acsenergylett.7b00462
- 46. Li, Y. et al. Li—O2 Cell with LiI (3-hydroxypropionitrile)2 as a Redox Mediator: Insight into the Working Mechanism of I− during Charge in Anhydrous Systems. J. Phys. Chem. Lett. 4218-4225 (2017). doi:10.1021/acs.jpclett.7b01497
- 47. Zhu, Y. G. et al. Proton enhanced dynamic battery chemistry for aprotic lithium-oxygen batteries. Nat. Commun. 8, 14308 (2017).
- 48. Kwabi, D. G. et al. The effect of water on discharge product growth and chemistry in Li—O2 batteries. Phys Chem Chem Phys 18, 24944-24953 (2016).
- 49. Torres, A. E. & Balbuena, P. B. Exploring the LiOH Formation Reaction Mechanism in Lithium-Air Batteries. Chem. Mater. (2018). doi:10.1021/acs.chemmater.7b04018
- 50. Viswanathan, V. et al. Comment on ‘Cycling Li—O2 batteries via LiOH formation and decomposition’. Science 352, 667-667 (2016).
- 51. Shen, Y., Zhang, W., Chou, S.-L. & Dou, S.-X. Comment on ‘Cycling Li—O2 batteries via LiOH formation and decomposition’. Science 352, 667-667 (2016).
- 52. Liu, T. et al. Response to Comment on “Cycling Li—O2 batteries via LiOH formation and decomposition”. Science 352, 667-667 (2016).
- 53. Bentley, C. L., Bond, A. M., Hollenkamp, A. F., Mahon, P. J. & Zhang, J. Voltammetric Determination of the Iodide/Iodine Formal Potential and Triiodide Stability Constant in Conventional and Ionic Liquid Media. J. Phys. Chem. C 119, 22392-22403 (2015).
- 54. Nakanishi, A. et al. Electrolyte Composition in Li/O2 Batteries with LiI Redox Mediators: Solvation Effects on Redox Potentials and Implications for Redox Shuttling. J. Phys. Chem. C (2018). doi:10.1021/acs.jpcc.7b11859
- 55. Reid, C. & Mulliken, R. S. Molecular compounds and their spectra. IV. The pyridine-iodine System1. J. Am. Chem. Soc. 76, 3869-3874 (1954).
- 56. Kolthoff, I. M. & Jordan, J. Voltammetry of iodine and iodide at rotated platinum wire electrodes. J. Am. Chem. Soc. 75, 1571-1575 (1953).
- 57. Grassian, V. H. & Muetterties, E. L. Electron energy loss and thermal desorption spectroscopy of pyridine adsorbed on platinum (111). J. Phys. Chem. 90, 5900-5907 (1986).
- 58. Black, R. et al. Screening for Superoxide Reactivity in Li—O2 Batteries: Effect on Li2O2/LiOH Crystallization. J. Am. Chem. Soc. 134, 2902-2905 (2012).
- 59. McCloskey, B. D., Bethune, D. S., Shelby, R. M., Girishkumar, G. & Luntz, A. C. Solvents' Critical Role in Nonaqueous Lithium-Oxygen Battery Electrochemistry. J. Phys. Chem. Lett. 2, 1161-1166 (2011).
- 60. Harding, J. R. Investigation of oxidation in nonaqueous lithium-air batteries. (Massachusetts Institute of Technology, 2015).
- 61. Harding, J. R., Amanchukwu, C. V., Hammond, P. T. & Shao-Horn, Y. Instability of Poly(ethylene oxide) upon Oxidation in Lithium-Air Batteries. J. Phys. Chem. C 119, 206947-6955 (2015).
- 62. Noviandri, I. et al. The Decamethylferrocenium/Decamethylferrocene Redox Couple: A Superior Redox Standard to the Ferrocenium/Ferrocene Redox Couple for Studying Solvent Effects on the Thermodynamics of Electron Transfer. J. Phys. Chem. B 103, 6713-6722 (1999).
- 63. Matsumoto, M. & Swaddle, T. W. The Decamethylferrocene(+/0) Electrode Reaction in Organic Solvents at Variable Pressure and Temperature. Inorg. Chem. 43, 2724-2735 (2004).
- 64. Born, M. Volumen und hydratationswarme der ionen. Z. Für Phys. 1, 45-48 (1920).
- 65. Atkins, P. W. & MacDermott, A. J. The Born equation and ionic solvation. J. Chem Educ 59, 359 (1982).
- 66. ADAMSON, A. W. Physical Chemistry of Surfaces. 12
- 67. Gallant, B. M. et al. Influence of Li2O2 morphology on oxygen reduction and evolution kinetics in Li—O2 batteries. Energy Environ. Sci. 6, 2518 (2013).
- 68. Sharon, D. et al. Oxidation of Dimethyl Sulfoxide Solutions by Electrochemical Reduction of Oxygen. J. Phys. Chem. Lett. 4, 3115-3119 (2013).
- 69. Kwabi, D. G. et al. Chemical Instability of Dimethyl Sulfoxide in Lithium-Air Batteries. J. Phys. Chem. Lett. 5, 2850-2856 (2014).
- 70. Schutte, L., Kluit, P. P. & Havinga, E. The substitution reaction of histidine and some other imidazole derivatives with iodine. Tetrahedron 22, 295-306 (1966).
- 71. Gerritsen, C. M., Gazda, M. & Margerum, D. W. Non-metal redox kinetics: hypobromite and hypoiodite reactions with cyanide and the hydrolysis of cyanogen halides. Inorg. Chem. 32, 5739-5748 (1993).
- 72. Lengyel, I., Epstein, I. R. & Kustin, K. Kinetics of iodine hydrolysis. Inorg. Chem. 32, 5880-5882 (1993).
- 73. Xie, Y., McDonald, M. R. & Margerum, D. W. Mechanism of the Reaction between Iodate and Iodide Ions in Acid Solutions (Dushman Reaction). Inorg. Chem. 38, 3938-3940 (1999).
- 74. Wren, J. C., Paquette, J., Sunder, S. & Ford, B. L. Iodine chemistry in the +1 oxidation state. II. A Raman and uv-visible spectroscopic study of the disproportionation of hypoiodite in basic solutions. Can. J. Chem. 64, 2284-2296 (1986).
- 75. Liu, T. et al. Understanding LiOH Chemistry in a Ruthenium Catalyzed Li—O2 Battery. Angew. Chem. Int. Ed. (2017). doi: 10.1002/anie.201709886
- 76. Huang, C., Kristoffersen, H. H., Gong, X.-Q. & Metiu, H. Reactions of Molten LiI with I2, H2O, and O2 Relevant to Halogen-Mediated Oxidative Dehydrogenation of Alkanes. J. Phys. Chem. C 120, 4931-4936 (2016).
- 77. Lide, D. R. CRC Handbook of Chemistry and Physics.
Derivation of the Role of Solvation Energy on Li2O2 Oxidation by I3−: For the reaction:
Li2O2+I3−→2Li++3I−+O2
Using the same approach as Kwabi et al2:
ΔGrxn=2ΔfG0Li++ΔfG0O2+3ΔfG0I−−ΔfG0Li2O2−ΔfGI3−+2ΔGLi+solv+ΔGO2solv+3ΔGI−solv−ΔGI3−solv
Assuming ΔG0O2solv≈0 and collecting all formation energy terms as ΔfG0rxn:
ΔGrxn=ΔfG0rxn+2ΔGLi+solv+3ΔGI−solv−ΔGI3−solv
Approximation of LiIO3 Formation Energy
From Huang et al3, the following reaction:
LiI+3/2O2→LiIO3
Has a reaction enthalpy of −3.0 eV. From Lide4, the S0 of KIO3 is 1.57 meV/° K and the ΔfG0 of LiI is −2.80 eV. Based on approximating the S° of LiIO3 to be the same as KIO3, at 298.15° K, the ΔfG0 of LiIO3 is calculated to be −5.05 eV.
Discussion of Polyiodide Species
While shifts in thermodynamics caused by changes in Li+ and I− solvation energy provide a clear explanation as to why the ability of I3− species to chemically oxidize Li2O2 can change with solvent, the dissociation of I3− into I2 species given by equilibrium (5), is also solvent dependent5 and must also be considered.
In addition to this equilibrium, there also exists higher order polyiodide species such as pentaiodide (I5−) and heptaiodide (I7−) that exist in other equilibriums caused by the association of I2 to Ix− (x=1, 3, 5)6-8:
I5−![custom-character]()
I2+I3− (1)
I7−![custom-character]()
I2+I5−
To further complicate matters, iodine-solvent complexes can also cause the dissociation of I2 into I− and I+ leading to further still equilibria to consider9:
I2![custom-character]()
Solvent·I++I− (3)
where the I− formed from this dissociation would then associate with another I2 to form I3− via reaction (5). The existence of these chemical equilibria considerably complicates the interpretation of reactions involving oxidized forms of iodide, as a large number of different species can be present in the solution at any given time. This begs the question; which polyiodide species are responsible for the observed oxidation of Li2O2?
In order to examine the possible role of highly oxidizing I+ species, the reaction between Li2O2 and 12 in hexane (a solvent which does not support the formation of Solvent-I+ complexes as indicated by its purple color in FIG. 44) was conducted. After reacting with commercial Li2O2, the originally purple 50 mM I2 hexane solution became clear and an orange/brown solid remained (FIG. 45). Raman spectroscopy on this solid reveals Li2O2, in addition to peaks consistent with LiI3 (FIG. 46). While stable triiodide salts can be formed with larger cations, such as Cesium10, such complexes are not typically stable with Li+. However, due to hexane's very low solubility for most salts11, the I3− remaining after the reaction of I2 with Li2O2 is likely more stable as a solid precipitate than in solution. Over time, this precipitate was found to lose its color, which would be consistent with the sublimation of I2 gas and formation of LiI. Since I2 is able to react with Li2O2 in hexane, we can conclude that the formation of I+ is not necessary for the oxidation of Li2O2.
The equilibria between I2, I3− and higher order polyiodide species cannot be untangled as effectively as there isn't an equivalent model solvent which eliminates these equilibria. Experiments were performed using solutions of 50 mM 12, I5− (50 mM I2+25 mM LiI) and I3− (50 mM I2+0.2M LiI) in DME and both commercial Li2O2 and Li2O2 formed through disproportionation (see FIG. 27). In all cases, the reaction with commercial Li2O2 was observed to stop once the reaction proceeded enough such that the solution contained 50 mM Ix (within experimental error). Raman spectra on the solution before and after the reaction (FIG. 28) indicated that while initially, the dominant species were I2, I5− and I3−, respectively, following the reaction, only the signal from I3− was visible in all solutions. Despite these results which suggest that some polyiodide species, like I5−, may be reactive with Li2O2, we note that these experiments do not clarify whether the reaction proceeds directly from I5−, or whether I5− first dissociates to I2 via reaction (1) and then the formed I2 reacts with Li2O2. In fact, the dissociation equilibria responsible for the formation of a more oxidizing and less oxidizing species (for example I3− dissociates to the more oxidizing I2 and I−), are actually unequivocally linked to the oxidizing power of the associated complex. Considering I3−, we recall from above that the increased solvation energy of I− ions increases the potential of the I−/I3− redox transition vs Me10Fc. However, the dissociation of I3− into I2 and I− will also be promoted by stronger solvation of I−. Furthermore, since this dissociation is in equilibrium, the thermodynamic driving force for the oxidation of Li2O2 by either I3− or the I2 formed from dissociation will be identical due to Nemstian shifts associated with the I2, I3− and I− concentrations present in this equilibrium. We therefore suggest that the distinction between the various polyiodide species which exist in all of the existing equilibria in solution is only relevant in discussion of reaction kinetics (and even here may not prove important unless the dissociation step is rate limiting) and that whether a reaction will occur or not is governed simply by the thermodynamic driving force for the reaction between Li2O2 and any of the polyiodide species which are in equilibria, which can be effectively understood with the framework presented above.
REFERENCES (EACH OF THE FOLLOWING REFERENCES IS INCORPORATED BY REFERENCE IN ITS ENTIRETY)
- 1. Burke, C. M. et al. Implications of 4e− Oxygen Reduction via Iodide Redox Mediation in Li—O2 Batteries. ACS Energy Lett. 1, 747-756 (2016).
- 2. Kwabi, D. G. et al. Experimental and Computational Analysis of the Solvent-Dependent O2/Li+—O2− Redox Couple: Standard Potentials, Coupling Strength, and
- Implications for Lithium-Oxygen Batteries. Angew. Chem. Int. Ed. 55, 3129-3134 (2016).
- 3. Huang, C., Kristoffersen, H. H., Gong, X.-Q. & Metiu, H. Reactions of Molten LiI with I2, H2O, and O2 Relevant to Halogen-Mediated Oxidative Dehydrogenation of Alkanes. J. Phys. Chem. C 120, 4931-4936 (2016).
- 4. Lide, D. R. CRC Handbook of Chemistry and Physics.
- 5. Bentley, C. L., Bond, A. M., Hollenkamp, A. F., Mahon, P. J. & Zhang, J. Voltammetric Determination of the Iodide/Iodine Formal Potential and Triiodide Stability Constant in Conventional and Ionic Liquid Media. J. Phys. Chem. C 119, 22392-22403 (2015).
- 6. Popov, A. I., Rygg, R. H. & Skelly, N. E. Studies on the Chemistry of Halogens and of Polyhalides. IX. Electrical Conductance Study of Higher Polyiodide Complex Ions in Acetonitrile Solutions1,2. J. Am. Chem. Soc. 78, 5740-5744 (1956).
- 7. Parrett, F. W. & Taylor, N. J. Spectroscopic studies on some polyhalide ions. J. Inorg. Nucl. Chem. 32, 2458-2461 (1970).
- 8. Svensson, P. H. & Kloo, L. Synthesis, Structure, and Bonding in Polyiodide and Metal Iodide-Iodine Systems. Chem. Rev. 103, 1649-1684 (2003).
- 9. Palomares, E., Clifford, J. N., Haque, S. A., Lutz, T. & Durrant, J. R. Control of charge recombination dynamics in dye sensitized solar cells by the use of conformally deposited metal oxide blocking layers. J. Am. Chem. Soc. 125, 475-482 (2003).
- 10. Li, Y. et al. Li—O2 Cell with LiI (3-hydroxypropionitrile)2 as a Redox Mediator: Insight into the Working Mechanism of I− during Charge in Anhydrous Systems. J. Phys. Chem. Lett. 4218-4225 (2017). doi:10.1021/acs.jpclett.7b01497
- 11. Markus von Pilgrim, Mihail Mondeshki & Jan Klett. [Bis(Trimethylsilyl)Methyl]Lithium and -Sodium: Solubility in Alkanes and Complexes with O- and N-Donor Ligands. Inorganics 5, 39 (2017).
Other embodiments are within the scope of the following claims.
Patent Application
20210273215
