See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/308004421

Implications of 4 e- Oxygen Reduction via Iodide Redox Mediation in Li-O2 Batteries

Article in ACS Energy Letters · September 2016 DOI: 10.1021/acsenergylett.6b00328

CITATIONS READS 3 206

7 authors, including:

Vincent Giordani Dan Addison Liox Power, Inc. Independent Researcher

30 PUBLICATIONS 769 CITATIONS 30 PUBLICATIONS 400 CITATIONS

SEE PROFILE SEE PROFILE

Linda F. Nazar Bryan Mccloskey University of Waterloo University of California, Berkeley

264 PUBLICATIONS 19,366 CITATIONS 66 PUBLICATIONS 4,941 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Vincent Giordani on 15 September 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. Letter http://pubs.acs.org/journal/aelccp

Implications of 4 e− Oxygen Reduction via − Iodide Redox Mediation in Li O2 Batteries Colin M. Burke,†,‡,# Robert Black,§,∥,# Ivan R. Kochetkov,§,∥ Vincent Giordani,⊥ Dan Addison,*,⊥ Linda F. Nazar,*,§,∥ and Bryan D. McCloskey*,†,‡

† Department of Chemical and Biomolecular Engineering, University of California, Berkeley, California 94720, United States ‡ Energy Storage and Distributed Resources Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States § University of Waterloo, Department of Chemistry, Waterloo, Ontario N2L 3G1, Canada ∥ Waterloo Institute for Nanotechnology, Waterloo, Ontario N2L 3G1, Canada ⊥ Liox Power Inc., 129 North Hill Avenue, Pasadena, California 91106, United States

*S Supporting Information

− − ABSTRACT: The nonaqueous lithium oxygen (Li O2) electrochemistry has garnered significant attention because of its high theoretical specific energy compared to the state-of-the-art lithium-ion battery. The common − active nonaqueous Li O2 battery cathode electrochemistry is the formation (discharge) and decomposition (charge) of lithium peroxide (Li2O2). Recent reports suggest that the introduction of lithium iodide (LiI) to an ether-based electrolyte containing water at impurity levels induces a 4 e− oxygen reduction reaction forming lithium hydroxide (LiOH) potentially mitigating instability issues related to typical Li2O2 formation. We provide fl quantitative analysis of the in uence of LiI and H2O on the electrochemistry − in a common Li O2 battery employing an ether-based electrolyte and a carbon cathode. We confirm, through numerous quantitative techniques, ffi − that the addition of LiI and H2O promotes e cient 4 e oxygen reduction to LiOH on discharge, which is unexpected given that only 2 e− oxygen reduction is typically observed at undoped carbon electrodes. Unfortunately, LiOH is not reversibly oxidized to O2 on charge, where instead a complicated mix of redox shuttling and side reactions is observed.

he ever increasing demands for sustainable, low carbon- scientific challenges inhibit the achievement of reversible, ’ ffi emission energy in today s society requires that we e cient Li2O2 formation and oxidation, with two being develop electrochemical energy storage systems that − T particularly pressing. First, Li2O2 and its intermediates (O2 will exceed those of current technologies. They are necessary, and LiO ) are known to undergo parasitic side reactions with for example, to deliver reliable, low-cost dispatchable power 2 organic electrolytes and carbon cathodes during battery from renewable energy and to power zero CO2-emission 3−15 vehicles. While lithium-ion is currently the battery chemistry of operation. When a realistically small electrolyte volume is − choice to meet these needs for vehicular transport, this used, these reactions limit Li O2 cell rechargeability to only a technology is approaching its theoretical energy storage limits, few cycles. Additionally, the products of these parasitic dictated by intercalation processes at the positive and negative reactions include solid carbonate/carboxylate species, which electrodes. New chemistries are urgently needed in hopes of accumulate at the electrode surface, thereby blocking O2 meeting the electrochemical energy storage needs of future evolution from the Li O −electrolyte interface on charge.16 generations. Because of its high theoretical specific energy, the 2 2 − As a result, the potential on charge necessarily increases to Li air (or O2) battery is currently being explored as a possible 1,2 − values substantially higher (>1 V higher) than the equilibrium alternative to lithium-ion. In typical nonaqueous Li O2 cell potential for O evolution from Li O , creating an energy compositions, lithium peroxide (Li2O2) formation and 2 2 2 oxidation has been identified as the reversible electrochemical inefficient battery. With these parasitic reactions seemingly reaction at the cathode: exacerbated at high potentials, this challenge perpetuates itself. +− 2Li++ O 2e ↔ Li O (1) 222 Received: August 3, 2016 The forward reaction occurs on discharge, and the reverse Accepted: September 9, 2016 reaction occurs on charge. Unfortunately, many unsolved

© XXXX American Chemical Society 747 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

Therefore, the avoidance of parasitic reactions is necessary to fully consumed during discharge, the discharge electro- improve both rechargeability and energy efficiency. chemistry reverts back to the more common 2 e− oxygen Second, Li2O2 is a wide bandgap insulator and is insoluble in reduction reaction to form Li2O2. On charge, the concen- all studied organic electrolytes. Therefore, as it deposits on the trations of LiI and H2O and the identities of the discharge cathode during discharge, it passivates the electrode surface, products all have a dramatic influence on the observed − fi limiting Li O2 cell discharge capacity to a minor fraction of electrochemistry and resultant voltage pro les. Isotopic labeling that necessary to compete with the specific energy of Li-ion confirmed that O evolution is observed to occur only from − 2 batteries.17 23 Recent reports have elucidated strategies to Li O , although reactions between Li O and oxidized iodo 2 2 − 2 2 partially circumvent this passivation issue by triggering a species (I3 ) can result in other product formation at low 24−27 ∼ solution-based mechanism for Li2O2 formation. However, potentials ( 3 V). No O2 evolution was observed from this mechanism results in Li2O2 deposition in large toroids and electrochemically formed LiOH. Instead, LiOH reacts with I2, 28 ∼ − perhaps Li2O2 deposition away from the cathode surface. A which electrochemically forms at 3.5 V from I3 oxidation, to fraction of Li2O2 therefore becomes electronically isolated from form H2O and LiIO3, the latter of which can be solubilized in the conductive cathode surface, likely contributing to the water-contaminated 1,2-dimethoxyethane (DME). Thus, while − observed large charge overpotentials. the 4e oxygen reduction activity on carbon electrodes in the − Substantial research into redox mediators is being pursued to presence of H2O impurities and I is intriguing, the addition of − address these charge transport limitations. In the context of Li− LiI and H2OtoaLi O2 battery does not promote a reversible O2 batteries, redox mediators are small, soluble molecules with oxygen electrochemistry in the cell compositions studied. a reversible redox potential slightly above the equilibrium We initially investigated the impact of various H2O and LiI − potential of the Li O2 battery (experimentally observed as electrolyte concentrations on the discharge and charge voltage 29 fi − ∼2.85 V). During charge, the mediator oxidizes at electroni- pro les of the Li O2 cell. The voltage response for a cally accessible sites on the cathode, diffuses to electronically Ketjenblack positive electrode discharged and charged (100 μ 2 2 ff isolated Li2O2, and accepts charge via an outer sphere electron A/cm to a capacity of 1 mAh/cm )withdierent transfer from the Li2O2, allowing the Li2O2 to be oxidized while concentrations of LiI and H2OinDME-basedaprotic regenerating the redox mediator.30 The battery therefore electrolyte is shown in Figure 1. Only small changes are charges near the redox potential of the mediator, which, if appropriately selected, is only slightly above the open-circuit potential of the battery, rather than reaching the high potentials required to presumably drive charge transport through Li2O2. The lower charge potential also provides a route to reduce parasitic processes observed at high potentials. Numerous redox − mediators have now been studied in Li O2 batteries, including tetrathiafulvalene (TTF),31 2,2,6,6-tetramethylpiperidinyloxyl (TEMPO),32 tris[4-(diethylamino)phenyl]amine (TDPA),33 iron phthalocyanine (FePC),34 and LiI.35,36 LiI is a redox mediator of particular interest in a cell containing minor H2O impurities. Recent reports indicated that in a cell containing LiI and H2O, Li2O2 was not the ultimate discharge product, but rather LiOH, possibly providing a route to reducing parasitic reactions typically associated with Li2O2 formation and oxidation.35,36 Furthermore, an incredibly low voltage gap (∼200 mV) between the discharge and charge voltage plateaus was observed, indicating that the cell achieved Figure 1. Voltage profiles of Ketjenblack (KB) electrodes (4 mg of ffi 36 KB) discharged and charged in DME/LiTFSI electrolyte with a remarkable energy e ciency. LiOH oxidation to O2 and − − ff H O through the action of the I /I redox couple (2.95 V vs di erent LiI concentrations (0.05 and 0.2 M) and H2O 2 3 μ 2 Li/Li+) was reported as the active charge electrochemistry, concentrations (500 and 5000 ppm) at 100 A/cm . The total − − salt concentration (LiI + LiTFSI) was kept at 0.3 M. The total although the I /I3 redox potential lies below the standard electrolyte volume used is 180 μL. At low LiI concentrations (0.05 potential for LiOH oxidation at standard conditions (3.45 V vs fi + M), regardless of H2O concentration, the charge pro le gradually Li/Li ). Liu et al. indicate in their motivating work that the slopes to higher voltage. At the higher LiI concentrations, an initial equilibrium of water, iodide, and oxygen is complex, and the low-voltage charge plateau (∼2.95 V) is exhibited. Only at the end role of polyanions (iodo-oxygen species) on the LiOH of the charge is a rise in potential observed for 5000 ppm H2O, 36−38 oxidation mechanism is poorly understood. Therefore, while the low charge plateau continues for the low (500 ppm) H2O − − while the reported electrochemistry of the I /I3 couple for concentration. − − Li O2 batteries is intriguing, the exact activity of the I and its − fi oxidized states, I3 and I2, toward Li2O2/LiOH oxidation is observed in the discharge voltage pro le between cells with unclear. varying LiI and H2O concentrations, although the electro- In this work, we present quantitative insight into the chemistry changes substantially, as will be discussed below. On fl fi fl in uence of LiI and H2O on the electrochemistry of a typical charge, the potential pro le is clearly in uenced by the LiI and − nonaqueous Li O2 cell employing a carbon cathode. LiOH is H2O concentrations. Generally, the higher LiI concentration observed as the primary discharge product, along with minor cells show lower charging voltages compared to their lower LiI parasitic side products, when LiI and H2O are both present in concentration counterparts. In fact, the 0.2 M LiI, 500 ppm ∼ the cell, with H2O consumed in the oxygen reduction reaction. H2O cell maintains a 3 V plateau for the duration of charge, 35,36 If H2O (which is present in impurity quantities in our study) is consistent with previous reports. As a result, the energy

748 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

Figure 2. High [LiI] DEMS results. Galvanostatic discharge and charge profiles (100 μA/cm2) with oxygen consumption and gas evolution for lithium−oxygen batteries employing a lithium anode, a cathode of ∼1 mg Ketjenblack with PTFE binder on stainless steel mesh (12 mm diameter), and 160 μL of DME with 0.25 M LiTFSI and 0.2 M LiI as the electrolyte. Each column presents a battery with different added − − − electrolyte water content: nominally anhydrous (a c), 500 ppm D2O(d f), and 5000 ppm H2O(g i). The batteries were discharged under a headspace of ∼1.5 atm oxygen and charged under a constant pressure headspace of ∼1.2 atm argon. Panels b, e, and h show oxygen consumption on discharge and evolution on charge. The asterisk (*) denotes a value adjusted for a hydrogen evolution rate at open-circuit potential, which arises from water reacting with the lithium anode. Gas evolution on charge was quantified via DEMS. Panels c, f, and i show quantitative gas evolution throughout the charge. efficiency of this cell appears outstanding, as only a ∼0.25 V gap for the direct correlation of observed oxygen consumption and between the discharge and charge voltages is observed. High evolution to the observed voltage profiles of our galvanostatic H2O concentrations, in addition to low LiI concentrations, cycling. appear to result in higher final charging voltages as well. Other Considering the discharge processes of the cells presented in carbon-based electrodes (e.g., those made from graphene Figure 2, all three cells exhibit oxygen consumption with rates ff − − materials) exhibit only small di erences in overpotentials varying between a pure 2 e /O2 process and a pure 4 e /O2 ’ (Figure S1) compared to Ketjenblack electrodes and were process depending on the cell s total H2O content. The − not studied further. nominally anhydrous battery exhibits a clear 2 e /O2 process The electrochemical reactions at various LiI and H2O for the duration of the discharge, indicative of Li2O2 formation fi ff − − concentrations were quanti ed via a combination of di erential as is typically observed in nonaqueous Li O2 cells (Figure 2a 39 − electrochemical mass spectrometry (DEMS), pressure decay/ c). The battery containing 500 ppm H2O(Figure 2d f) − fi rise measurements in a closed cell headspace, and acid/base and exhibits a 3.2 e /O2 process for the rst 0.24 mAh of discharge − peroxide titrations. Details of each measurement are provided and a 2 e /O2 process for the remainder of the 1 mAh − in the Supporting Information. discharge. The initial 3.2 e /O2 process indicates a mix of Figure 2 shows pressure decay (discharge) and DEMS LiOH formation (reaction 2), a 4 e−/O process, and Li O − 2 2 2 (charge) results for batteries employing 0.2 M LiI and H2O formation (reaction 1), a 2 e /O2 process. concentrations corresponding to nominally anhydrous (<20 +− ppm), 500 ppm, and 5000 ppm. The 0.2 M LiI was selected for 4Li+++ 4e O22 2H O ↔ 4LiOH (2) these measurements because, at modest current densities and − − with the given overall cell geometry, this concentration, coupled The switch from a 3.2 e /O2 process to a 2 e /O2 process ∼ with a low H2O concentration, enables the low 3 V potential likely occurs once H2O is completely consumed, as 500 ppm μ throughout the entire duration of charge, as observed in Figure H2O in the electrolyte volume employed (160 L) corresponds 1 and in previous reports.35,36 Pressure decay and DEMS allow to a capacity of 0.24 mAh for LiOH formation (as calculated

749 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

− − Figure 3. Acid/base and iodometric titration results for cells identical to those studied in Figure 2d f (500 ppm H2O) (a) and Figure 2g i fi (5000 ppm H2O) (c). The acid/base titration quanti es lithium hydroxide and other decomposition products present; these are distinguished by using the expected value of basic decomposition products formed during lithium peroxide formation. On charge, the relative fates of the decomposition products and the electrochemically formed LiOH are unknown, so they cannot be decoupled. The iodometric titration gives * ’ the lithium peroxide results. I2 formation during the 3.5 V plateau in panels c and d interfered with the iodometric titration s ability to μ “ ” μ properly quantify Li2O2 at the end of charge (7.3 mol of Li2O2 was titrated at the end of charge, even though only 0.3 mol was produced during discharge). from the reaction 2 stoichiometry), and the discharge potential Liu et al.,35,36 LiOH formation in a cell of this composition is a − fi − noticeably drops to a lower plateau at the onset of the 2 e /O2 fascinating nding, because the 4 e /O2 oxygen-reduction − Li2O2 formation process. The battery containing 5000 ppm reaction (ORR) reaction is uncommon in nonaqueous Li O2 − − H2O(Figure 2g i) exhibits a near 4 e /O2 process for the batteries and on carbon electrodes in general. duration of the 1 mAh discharge, indicating LiOH formation On charge, the three cells presented in Figure 2 exhibit very dominates at high H2O content. Theoretically, as calculated little oxygen evolution compared to the oxygen consumed from reaction 2, 5000 ppm H2O in the electrolyte volume during discharge. The total oxygen evolution from the employed (160 μL) corresponds to a capacity of 2.4 mAh for nominally anhydrous battery, for example, is 2.4 μmol, or less ffi LiOH formation; therefore, there is su cient H2O to allow a 1 than 13% of what would be expected assuming full oxidation of mAh discharge even after considering (the discernible) H2O the discharge products. As the nominally anhydrous and 500 consumption at the lithium anode. The presence of LiOH as a ppm H2O electrolytes exhibit Li2O2 as their dominant discharge fi ff discharge product was further con rmed via X-ray di raction product and the 5000 ppm H2O electrolyte exhibits LiOH as its (XRD) studies of the cathode, as shown in Figure S2 for a cell dominant discharge product, the limited O2 evolution during ’ ∼ ∼ with high LiI and H2O content. the cells 3 V charge plateaus shows that the 3 V charge fi Quanti cation of both LiOH and Li2O2 was performed using plateau is predominantly related to electrochemical processes established titrations on extracted cathodes from cells with other than O2 evolution, irrespective of the discharge product. identical composition as those studied in Figure 2 (Figure 3).40 Additionally, while the 5000 ppm battery exhibits a small ∼ For a battery containing 500 ppm H2O and 0.2 M LiI (Figure amount of oxygen evolution during the step from the 3V − μ μ ∼ 2d f), 5.8 mol of LiOH and 15.8 mol of Li2O2 are expected plateau to the higher 3.5 V plateau, DEMS shows negligible to form during the 1 mAh discharge given the amount of O2 oxygen evolution otherwise. Therefore, LiOH is not reversibly − − μ consumed during the 4 e /O2 and 2 e /O2 processes; 5.1 mol oxidized to produce O2 at these potentials. The cell chemistry μ of LiOH and 11.1 mol of Li2O2 were titrated (Figure 3a,b), on charge is extremely complex, being a function of the − giving yields of 87% LiOH and 70% Li2O2. The Li2O2 yield is concentrations of I and H2O, as well as the ratio of discharge − consistent with previous reports for nominally anhydrous Li capacity to total H2O content in the cell (and hence the 40 fi O2 cells employing Ketjenblack cathodes. Therefore, LiOH product distribution formed during discharge). The speci c ffi formation is more e cient than Li2O2 formation in these cells, features of the charge chemistry are now discussed. although parasitic processes during LiOH formation still persist. Numerous processes occur during charge at the ∼3V fi As identi ed previously by both Aurbach and co-workers and plateau. If substantial Li2O2 is formed, as is the case in the low

750 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter − H2O content cells (Figure 2a g), it can disappear during this without any shuttling, a capacity of only 0.16 mAh would be plateau, as observed from iodometric titrations presented in expected. Also, analogous cells that employ higher concen- Figure 3a,b. A small fraction of Li2O2 disappearance is ascribed trations of H2O, such as 5000 ppm, exhibit the two-plateau to oxidation that evolves O (Figure 2c,f) while the remainder behavior as shown in Figure 2g, indicating water’s ability, 2 − appears to chemically react with I3 to form a soluble product. perhaps through its reaction with the lithium metal electrode, − − ’ − − I3 was likely formed from I oxidation during charge in Figure to inhibit the system s ability to sustain the I /I3 redox 3a, as the battery separator after charge was a faint yellow-red in shuttling between electrodes. − − color, indicative of I3 in aprotic polar media. We note that I3 When LiOH is the dominant discharge product, i.e., cells formed during charge can result in a false positive in our with high H2O content, it disappears only in cells that reach the ff ∼ − − iodometric titration, so that the di erence in the Li2O2 titrated high 3.5 V plateau, suggesting the I /I3 redox shuttle between points i and ii in Figure 3a (9.4 μmol difference, or dominates at the ∼3 V plateau. As shown in Figure 3a, acid/ 85% of the Li2O2 formed during discharge) represents a lower base titrations indicate that after a full discharge−charge cycle, ∼ limit of the actual Li2O2 that reacted on charge. cells with low H2O content, and thus those that exhibit the 3 The Li2O2 disappearance cannot account for the entirety of V plateau for the duration of charge, still contain all LiOH the electrochemical charging capacity, implying that the ∼3V formed during discharge. This is consistent with SEM images in − − plateau is related in part to I /I3 redox shuttling between the Figure S4 showing that solid discharge products remain after anode and cathode. Figure 4a shows that indeed a cell charge, indicating that there is no activity for LiOH oxidation at this low charge potential. Cathodes discharged and charged in ∼ cells with high H2O content that reach the 3.5 V plateau on charge, however, exhibit (after charge) only a minor fraction of the basic products formed during discharge. The ∼3.5 V − plateau is likely related to the I3 /I2 redox couple, which is active at these potentials, and I2 formation is clearly visible in the glass fiber separators of cells that reach this plateau. ∼ Recalling the minimal O2 evolution exhibited at even the 3.5 V plateau (Figure 2g−i), this suggests that the dominant chemistry occurring at this plateau is a non-O2 evolving LiOH − oxidation reaction involving the I3 /I2 redox couple, which will be discussed later. fi Isotopic labeling of O2 and H2O nds that no evolved O2 originates from LiOH formed during discharge. O2 evolution from LiOH during charge requires O−O bond reformation that occurs between statistically random oxygen atoms contained in the LiOH. LiOH forms during discharge through O2 bond splitting and reaction with H2O. Thus, a cell discharged under 16,16 18 O2 and containing a H2 O impurity would be expected to fi 16,18 evolve signi cant O2 on charge (i.e., the evolved O2 should be enriched in 18O). The oxygen evolution of a cell that was charged after being discharged to a capacity of 1 mAh in 0.2 M 18 16,16 LiI/5000 ppm H2 O under O2 is shown in Figure 5.No − Figure 4. Li O2 batteries charged under 1 atm argon gas with no prior discharge as a function of current density. Electrolyte is 0.25 M LiTFSI and 500 ppm H2O in DME with 0.05 M LiI (a) and no added LiI (b). Cyclic voltammograms of these cells are also presented in Figure S3. These cells were constructed with two Whatman GF/D glass fiber separators, 300 μL of electrolyte, and cathodes containing 2 mg of a 95:5 mass ratio Super P carbon to PTFE on 10 mm stainless steel mesh. The interelectrode distance was ∼500 μm.

containing just 0.05 M LiI and 500 ppm H2O (with our cathode composition and cell geometry) that has yet to be discharged can be “charged” at low current densities at ∼3Vin perpetuity, while Figure 4b shows that an attempted charge of an identical cell without LiI exhibits an immediate sharp rise in voltage. This ability to “charge” without formation of any discharge product is evidence that an I−/I − redox shuttle 3 Figure 5. Voltage profile and corresponding oxygen evolution current can be maintained between the electrodes at current fi μ 2 2 pro le for a cell discharged and charged (100 A/cm ) in 0.2 M densities of less than roughly 100 μA/cm , which is the current 18 LiI/DME with 5000 ppm H2 O. On discharge, a reasonable density used in Figures 1 and 2. Generally, the shuttling process 18 fraction of total produced LiOH is O labeled (some Li2O2 is also fl 16 is known to be strongly in uenced by mediator concentration, produced). On charge, only O2 is observed from the oxidation of ff ffi 18 di usion coe cient, current density, and interelectrode Li2O2. No signal from O2 (LiOH oxidation) is observed at either 41 − − − − − distance. Of note, if all I were converted to I3 in this cell low potential (I /I3 ) or high potential (I3 /I2).

751 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

18 O-enriched O2 is evolved at any point during the entire observed on discharge, and O2 evolution is observed on charge. − charge process. The transition from I3 to I2 is accompanied by The O2 evolution coincides with a rising charge voltage above 16 ∼ the evolution of a small volume of isotopically pure O2, 3.2 V, and it corresponds to 63% of the expected yield based − − indicating that the O ObondinO2 remained intact on the O2 consumed during the 2 e /O2 discharge process. throughout the discharge−charge process, as expected with This oxygen-evolution reaction (OER)/ORR value is similar to − Li2O2 formation and oxidation. This result is also consistent (though slightly lower than) the yield in aprotic Li O2 cells in 4 ff with Figure 2f (500 ppm H2O cell), where discharge under a the absence of LiI. A battery discharged to a cuto capacity of 18 16 − pure O2 headspace (with H2 O as the water impurity) results less than 0.5 mAh, where only a 4 e /O2 process is observed in pure 18O evolution during charge. This confirms that the (Figure 6d−f), exhibits only minor oxygen evolution on charge. 2 fi oxidation of LiOH to form O2 does not occur either chemically These results further con rm that only Li2O2, and not LiOH, or electrochemically, and the small amount of O2 observed can be oxidized to evolve O2 in aprotic electrolytes on carbon from cells containing LiOH as the primary discharge product cathodes. Batteries employing lower iodide concentrations are ∼ can be ascribed to the oxidation of Li2O2. not able to maintain the desirable 3 V charge plateau but do Interestingly, lowering the iodide concentration enables exhibit more oxygen evolution on charge than the higher LiI higher rates of oxygen evolution on charge when Li2O2 is the concentration cells, with the oxygen originating exclusively primary discharge product. For example, Figure 6 presents from Li2O2. The DEMS and titration data clearly indicate that the addition of LiI to a water-contaminated, nonaqueous lithium− oxygen battery enables 4e− oxygen reduction involving the consumption of H2O to form LiOH on discharge. However, on charge, LiOH is not reversibly oxidized back to oxygen. Any Li2O2 formed during discharge is only partially oxidized to evolve oxygen on charge, with higher oxygen evolution efficiencies observed at lower iodide concentrations. Con- sequently, the low 3 V plateau on charge obtained in Figure 1 is not due to reversible oxygen electrochemistry, but rather to a complex iodo-oxygen electrochemistry that is highly dependent on the exact cell composition and testing conditions (e.g., LiI and H2O concentration, current density, cathode composition, and cell geometry). This still leaves open the question of the fate of LiOH at the high, ∼3.5 V charge plateau (i.e., where it is observed to disappear). Contrasting reports exist that demonstrate the removal of LiOH from the electrode surface during charge of a − − Li O2/LiI cell. Liu et al. reports that I3 oxidizes LiOH to ∼ 36 evolve O2 during the 3 V plateau. Our own results, as well as those from Aurbach and co-workers,35 demonstrate that LiOH does not react at the low potential plateau but can be removed from the cathode surface at the higher (∼3.5 V) voltage plateau, but without O2 evolution. One possible explanation for this behavior is that LiOH oxidation results in LiIO3 formation. In alkaline aqueous media, the following reaction occurs:42

Figure 6. Low [LiI] DEMS results. Panels a and d show 3I232+↔++ 6LiOH LiIO 5LiI 3H O (3) galvanostatic discharge and charge profiles (200 μA/cm2) from lithium−oxygen batteries employing a lithium anode; a cathode of Hence, it is possible that at high charge voltages, when the ∼ − 2 mg XC72 carbon with PTFE binder on stainless steel mesh (12 I3 /I2 redox couple is actively forming I2, LiOH reacts to form mm diameter); and 80 μL of an electrolyte of 0.25 M LiTFSI, 0.05 small amounts of LiIO , given that a 6 LiOH:1 LiIO ff 3 3 M LiI, and roughly 2000 ppm H2O in DME. The only di erence ff stoichiometric ratio is expected from reaction 3. To prove between the batteries is the discharge capacity cuto , as labeled. this hypothesis, a modified version of the iodometric titration Panels b and e show oxygen consumption on discharge and was used to detect small amounts of LiIO3 present on various evolution on charge versus capacity. Panels c and f show ff quantitative gas evolution throughout the charge. The asterisk positive electrodes charged to di erent capacities. A detailed (*) denotes a value adjusted for a hydrogen evolution rate at open- description of this method is in the Supporting Information. circuit potential, which arises from water reacting with the lithium The fraction of LiIO3 was determined for electrodes stopped at anode. various states of charge in either 0.05 M LiI or 0.2 M LiI/5000 ppm H2O/DME electrolyte (Figure 7). This H2O concen- tration was chosen to ensure LiOH was formed as the pressure decay (discharge) and DEMS (charge) results for overwhelmingly dominant product on discharge. For low LiI batteries discharged with 0.05 M LiI (instead of 0.2 M LiI, as concentrations (0.05 M), it is clear that a small quantity of μ μ was used in Figure 2) and 2000 ppm H2Oin80 Lof LiIO3 is produced (<0.5 mol) on cell charge. The small electrolyte (corresponding to 0.5 mAh of possible LiOH magnitude is understandable, because at 0.05 M LiI complete formation). When discharged beyond 0.5 mAh (Figure 6a−c), conversion of I− to I equates to a maximum of 1.5 μmol of − − 2 a switch from a 4 e /O2 process to a 2 e /O2 process is LiIO3 produced via reaction 3, assuming the discharge product

752 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

Figure 7. Electrochemical curves for batteries containing (a) 0.05 M LiI/5000 ppm H2O/DME and (c) 0.2 M LiI/5000 ppm H2O/DME and fi corresponding LiIO3 for (b) 0.05 M and (d) 0.2 M LiI at the points speci ed. At low LiI concentrations, little LiIO3 is generated because of V the small concentration of I2. At high LiI concentrations, an abundance of LiIO3 is observed at > 3.5 because of the generation of large μ 2 amounts of I2, which reacts with the LiOH to form LiIO3. Batteries were cycled at 100 A/cm , and cathode carbon loading was 4 mg of KB. fi is 100% LiOH. Nonetheless, the presence of LiIO3 is con rmed prepared with nominally anhydrous electrolytes (<50 ppm) on extracted positive electrodes with both EDX analysis, which with three different concentrations of LiI (0 M, 0.05 M, and 0.2 ’ shows LiIO3 s characteristic 3:1 ratio of O:I (Figure S7), and M). Figure 8a demonstrates the decrease of the charge XRD analysis, which shows diffraction patterns characteristic of overpotential with increasing LiI concentration. Of note, a LiIO3 (Figure S8). A similar quantity of LiIO3 is also observed distinct charge plateau is not observed for a cell with either 0.05 for the cell discharged/charged in 0.2 M LiI at voltages below M (magenta curve) or 0.2 M (blue curve) LiI. Instead, the 3.4 V (<0.5 μmol). However, at high voltages and significant charge overpotential (which is lower than that of the typical cell capacity overcharge, and thus high concentrations of I2, without LiI) exhibits a steady rise throughout the entire charge μ approximately 3.0 mol of LiIO3 is observed on the electrode process. surface, clearly indicating a higher reactivity of LiOH with I2 Although the presence of LiI leads to reduced overpotential − than I3 . Reaction between LiOH and I2 (and the lack of a during charge, it both reduces Li2O2 yield on discharge and − fi ffi reaction between LiOH and I3 ) was con rmed with mock oxygen evolution e ciency on charge. Increasing the LiI chemical experiments, as shown in Figure S5. LiIO3 is very concentration leads to decreasing yields of Li2O2 on discharge, soluble in H2O (76 g/100 mL) and hence can easily dissolve in expressed as a fraction of the theoretical value expected for a 2 the 5000 ppm H2O electrolyte. Dissolution is further increased mAh discharge capacity (assuming complete conversion in the with the production of H2O accompanying LiIO3 formation. absence of side-reactions; Figure 8b, blue curve). While this This is a plausible explanation for the removal of LiOH from behavior could arise from several sources, we note that in the the electrode surface, but this does not fully explain the removal absence of H2O, LiI could potentially promote the formation of of LiOH at low voltages by Liu et al.36 We note that the total LiOH through superoxide attack on glyme, as demonstrated by 35 amount of H2O contained in our cells is small compared to the Kwak et al. Apart from the lowered Li2O2 conversion, H2O included in cells studied by Liu et al. (e.g., 5000 ppm H2O increased LiI concentrations also lead to enhanced redox μ μ in 160 L in our cells versus 45 000 ppm H2O in 700 L in Liu shuttling, requiring a higher Coulombic input to evolve all of et al.). This distinction could give rise to the discrepancy the oxygen from the peroxide, and thus dramatically decreasing observed at low voltages in each study, because both LiOH the oxygen evolution efficiency on charge. Figure 8b shows that − − solubility and the chemical reaction rate between LiOH and I3 the 3.7 e /O2 process which characterizes charging in the 37 increase with increasing H2O content. Other factors that absence of LiI, which already exceeds the ideal of 2.0 owing to could play a role in this discrepancy are depth of discharge (i.e., side-reactions, further increases with greater LiI content (red amount of product formed) and the time delay between the curve). Thus, oxygen continues to evolve when overcharging − end of discharge and the characterization of the discharge the cell, resulting in an e /O2 value as high as 6.0 for 0.2 M LiI products. owing to the I− shuttle. fi ff − It is very clear that the H2O concentration has a signi cant In conclusion, the e ect of LiI and H2O on the Li O2 − impact on the chemistry of the Li O2 cell when LiI is present, chemistry can be summarized as follows (see the graphic because it plays a significant role in the formation of LiOH as provided in the abstract). On discharge, the presence of both ffi − the dominant discharge product. To isolate the e cacy of LiI as LiI and H2O promote a 4 e /O2 process (LiOH production) − a redox catalyst for Li2O2 formation and oxidation, cells were instead of the normal 2 e /O2 process (Li2O2 production)

753 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

scenario. Hence, the low charge overpotential seen in LiI- − containing Li O2 batteries is primarily due to iodo-oxygen electrochemistry, rather than reversible oxygen evolution. It should be noted, given our results, that other cell compositions (e.g., those with cathodes comprising active materials other than carbon, higher water or LiI concentrations than studied here, other halide salts, etc.) may result in a substantially different electrochemistry, and more research is certainly necessary to fully understand the complex charge chemistry observed in this system. ■ ASSOCIATED CONTENT *S Supporting Information The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsenergy- lett.6b00328. Materials, experimental methods, additional figures (PDF)

■ AUTHOR INFORMATION Corresponding Authors *E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Author Contributions fi # Figure 8. (a) Electrochemical pro les on discharge and charge of C.M.B. and R.B. contributed equally to this work. nominally anhydrous Li−O batteries (<50 ppm H O) containing 2 2 Notes varying concentrations of LiI. (b) Discharge yield of Li2O2 from fi titration of cathodes discharged to 2.0 mAh (blue squares); LiI The authors declare no competing nancial interest. ffi concentration dependence of O2 evolution e ciency (red circles), expressed as the charge capacity (in moles of electrons) divided by ■ ACKNOWLEDGMENTS the amount of O2 evolved up to the maximum charge capacity Helpful discussions with Prof. Clare Grey are gratefully (integrated over charge; in moles). Batteries were cycled at 100 μ 2 acknowledged. C.M.B., B.D.M., V.G., and D.A. gratefully A/cm , and the cathode carbon loading was 4 mg of KB. Dashed acknowledge support from the U.S. Department of Energy, lines are to guide the eye. Energy Efficiency and Renewable Energy Vehicle Technologies Office under Award No. DE-0006869, and C.M.B and B.D.M. − typical of an aprotic Li O2 cell employing an ether-based acknowledge previous support that enabled this work through electrolyte and undoped carbon cathode. High H O concen- −2 the Laboratory Directed Research and Development Program trations result in almost stoichiometric 4 e /O2 LiOH of Lawrence Berkeley National Laboratory under U.S. Depart- formation, whereas intermediate or low H2O concentrations ment of Energy Contract No. DE-AC02-05CH11231. C.M.B. result in a mixture of Li2O2 and LiOH formation. On charge, a gratefully acknowledges support from a NASA Space fi variety of processes occur which are dependent on the nal Technology Research Fellowship under Award No. discharge product composition and the concentrations of LiI − NNX16AM56H. L.F.N. is grateful for funding for this work and H2O. I oxidation, in the absence of any discharge product, from NRCan through their ECO-EII program and acknowl- features two redox couples, I−/I − (E ∼ 3.0 V) and I −/I (E ∼ − − 3 3 2 edges NSERC for partial support through the Discovery and 3.5 V). The I /I3 shuttle between the anode and cathode Canada Research Chair programs. substantially contributes to the cell electrochemistry at low charge potentials (2.95−3.10 V). LiOH does not oxidize when ■ REFERENCES − − this I /I3 couple is active, although Li2O2 can react to either (1) Christensen, J.; Albertus, P.; Sanchez-Carrera, R. S.; Lohmann, evolve O or form unidentified iodo-oxygen species. The I −/I 2 − − 3 2 T.; Kozinsky, B.; Liedtke, R.; Ahmed, J.; Kojic, A. A Critical Review of redox is observed after the initial I /I3 redox in cells with high Li/Air Batteries. J. Electrochem. Soc. 2012, 159,R1−R30. water content in the electrolyte. This transition results in O2 (2) Luntz, A. C.; McCloskey, B. D. Nonaqueous Li−Air Batteries: A evolution from Li O , presumably promoted by the formed I . Status Report. Chem. Rev. 2014, 114, 11721−11750. − 2 2 2 The I3 /I2 redox is active during a second charge plateau that (3) McCloskey, B. D.; Burke, C. M.; Nichols, J. E.; Renfrew, S. E. takes place at ∼3.5 V, and LiOH is observed to disappear Mechanistic Insights for the Development of Li-O2 Battery Materials: Addressing Li O Conductivity Limitations and Electrolyte and during this process. LiOH reacts with I2 to form a small amount 2 2 fi Cathode Instabilities. Chem. Commun. 2015, 51, 12701−12715. of LiIO3, which is soluble in the water- lled electrolyte via reaction 3, but does not evolve O . On the basis of our studies, (4) McCloskey, B. D.; Bethune, D. S.; Shelby, R. M.; Mori, T.; 2 − Scheffler, R.; Speidel, A.; Sherwood, M.; Luntz, A. C. Limitations in while a low charge overpotential makes the use of I seem Rechargeability of Li-O2 Batteries and Possible Origins. J. Phys. Chem. appealing, the lack of O2 evolution from Li2O2/LiOH oxidation Lett. 2012, 3, 3043−3047. in LiI-containing cells makes it a poor additive for efficient Li− (5) Garcia, J. M.; Horn, H. W.; Rice, J. E. Dominant Decomposition O2 cycling. Of the oxygen consumed on discharge, less than Pathways for Ethereal Solvents in Li-O2 Batteries. J. Phys. Chem. Lett. 13% is recovered on charge under the best-case ∼3 V charge 2015, 6, 1795−1799.

754 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

(6) Sharon, D.; Afri, M.; Noked, M.; Garsuch, A.; Frimer, A. A.; (24) Aetukuri, N. B.; McCloskey, B. D.; Garcia, J. M.; Krupp, L. E.; Aurbach, D. Oxidation of Dimethyl Sulfoxide Solutions by Electro- Viswanathan, V.; Luntz, A. C. Solvating Additives Drive Solution- chemical Reduction of Oxygen. J. Phys. Chem. Lett. 2013, 4, 3115− Mediated Electrochemistry and Enhance Toroid Growth in Non- − 3119. Aqueous Li-O2 Batteries. Nat. Chem. 2015, 7,50 56. (7) Ottakam Thotiyl, M. M.; Freunberger, S. A.; Peng, Z.; Bruce, P. (25) Johnson, L.; Li, C.; Liu, Z.; Chen, Y.; Freunberger, S. A.; Ashok, − G. The Carbon Electrode in Nonaqueous Li O2 Cells. J. Am. Chem. P. C.; Praveen, B. B.; Dholakia, K.; Tarascon, J.-M.; Bruce, P. G. The − Soc. 2013, 135, 494 500. Role of LiO2 Solubility in O2 Reduction in Aprotic Solvents and its − − (8) Freunberger, S. A.; Chen, Y. H.; Drewett, N. E.; Hardwick, L. J.; Consequences for Li O2 Batteries. Nat. Chem. 2014, 6, 1091 1099. Barde, F.; Bruce, P. G. The Lithium-Oxygen Battery with Ether-Based (26) Burke, C. M.; Pande, V.; Khetan, A.; Viswanathan, V.; Electrolytes. Angew. Chem., Int. Ed. 2011, 50, 8609−8613. McCloskey, B. D. Enhancing Electrochemical Intermediate Solvation − (9) Bryantsev, V. S.; Giordani, V.; Walker, W.; Blanco, M.; Zecevic, through Electrolyte Anion Selection to Increase Nonaqueous Li O2 S.; Sasaki, K.; Uddin, J.; Addison, D.; Chase, G. V. Predicting Solvent Battery Capacity. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 9293−9298. Stability in Aprotic Electrolyte Li-Air Batteries: Nucleophilic (27) Giordani, V.; Tozier, D.; Tan, H.; Burke, C. M.; Gallant, B. M.; Substitution by the Superoxide Anion Radical (O2(center dot-)). J. Uddin, J.; Greer, J. R.; McCloskey, B. D.; Chase, G. V.; Addison, D. A Phys. Chem. A 2011, 115, 12399−12409. Molten Salt Lithium−Oxygen Battery. J. Am. Chem. Soc. 2016, 138, (10) Bryantsev, V. S.; Uddin, J.; Giordani, V.; Walker, W.; Addison, 2656−2663. D.; Chase, G. V. The Identification of Stable Solvents for Nonaqueous (28) Adams, B. D.; Radtke, C.; Black, R.; Trudeau, M. L.; Zaghib, K.; Rechargeable Li-Air Batteries. J. Electrochem. Soc. 2013, 160, A160− Nazar, L. F. Current Density Dependence of Peroxide Formation in A171. the Li-O2 Battery and its Effect on Charge. Energy Environ. Sci. 2013, 6, (11) Adams, B. D.; Black, R.; Williams, Z.; Fernandes, R.; Cuisinier, 1772−1778. M.; Berg, E. J.; Novak, P.; Murphy, G. K.; Nazar, L. F. Towards a (29) McCloskey, B. D.; Scheffler, R.; Speidel, A.; Girishkumar, G.; − Stable Organic Electrolyte for the Lithium Oxygen Battery. Adv. Energy Luntz, A. C. On the Mechanism of Nonaqueous Li O2 Electro- Mater. 2015, 5, 1400867. chemistry on C and Its Kinetic Overpotentials: Some Implications for (12) Kundu, D.; Black, R.; Adams, B.; Harrison, K.; Zavadil, K.; Li−Air Batteries. J. Phys. Chem. C 2012, 116, 23897−23905. Nazar, L. F. Nanostructured Metal Carbides for Aprotic Li-O2 (30) Chase, G. V.; Zecevic, S.; Walker, W. T.; Uddin, J.; Sasaki, K. A.; Batteries: New Insights into Interfacial Reactions and Cathode Giordani, V.; Bryantsev, V.; Blanco, M.; Addison, D. D. Soluble Stability. J. Phys. Chem. Lett. 2015, 6, 2252−2258. Oxygen Evolving Catalysts for Rechargeable Metal-Air Batteries. U.S. (13) Black, R.; Oh, S. H.; Lee, J. H.; Yim, T.; Adams, B.; Nazar, L. F. Patent 20120028137 A1, 2012. Screening for Superoxide Reactivity in Li-O2 Batteries: Effect on (31) Chen, Y. H.; Freunberger, S. A.; Peng, Z. Q.; Fontaine, O.; − Li2O2/LiOH Crystallization. J. Am. Chem. Soc. 2012, 134, 2902 2905. Bruce, P. G. Charging a Li-O2 Battery Using a Redox Mediator. Nat. (14) McCloskey, B. D.; Scheffler, R.; Speidel, A.; Bethune, D. S.; Chem. 2013, 5, 489−494. Shelby, R. M.; Luntz, A. C. On the Efficacy of Electrocatalysis in (32) Bergner, B. J.; Schurmann, A.; Peppler, K.; Garsuch, A.; Janek, J. − Nonaqueous Li-O2 Batteries. J. Am. Chem. Soc. 2011, 133, 18038 TEMPO: A Mobile Catalyst for Rechargeable Li-O2 Batteries. J. Am. 18041. Chem. Soc. 2014, 136, 15054−15064. (15) Walker, W.; Giordani, V.; Uddin, J.; Bryantsev, V. S.; Chase, G. (33) Kundu, D.; Black, R.; Adams, B.; Nazar, L. F. A Highly Active V.; Addison, D. A Rechargeable Li-O2 Battery Using a Lithium Low Voltage Redox Mediator for Enhanced Rechargeability of Nitrate/N,N-Dimethylacetamide Electrolyte. J. Am. Chem. Soc. 2013, Lithium-Oxygen Batteries. ACS Cent. Sci. 2015, 1, 510−515. 135, 2076−2079. (34) Sun, D.; Shen, Y.; Zhang, W.; Yu, L.; Yi, Z. Q.; Yin, W.; Wang, (16) McCloskey, B. D.; Speidel, A.; Scheffler, R.; Miller, D. C.; D.; Huang, Y. H.; Wang, J.; Wang, D. L.; Goodenough, J. B. A Viswanathan, V.; Hummelshøj, J. S.; Nørskov, J. K.; Luntz, A. C. Twin Solution-Phase Bifunctional Catalyst for Lithium-Oxygen Batteries. J. − − Problems of Interfacial Carbonate Formation in Nonaqueous Li O2 Am. Chem. Soc. 2014, 136, 8941 8946. Batteries. J. Phys. Chem. Lett. 2012, 3, 997−1001. (35) Kwak, W.-J.; Hirshberg, D.; Sharon, D.; Shin, H.-J.; Afri, M.; (17) Viswanathan, V.; Thygesen, K. S.; Hummelshoj, J. S.; Norskov, Park, J.-B.; Garsuch, A.; Chesneau, F. F.; Frimer, A. A.; Aurbach, D.; J. K.; Girishkumar, G.; McCloskey, B. D.; Luntz, A. C. Electrical Sun, Y.-K. Understanding the Behavior of Li-Oxygen Cells Containing − Conductivity in Li2O2 and its Role in Determining Capacity LiI. J. Mater. Chem. A 2015, 3, 8855 8864. Limitations in Non-Aqueous Li-O2 Batteries. J. Chem. Phys. 2011, (36) Liu, T.; Leskes, M.; Yu, W.; Moore, A. J.; Zhou, L.; Bayley, P. 135, 214704. M.; Kim, G.; Grey, C. P. Cycling Li-O2 Batteries via LiOH Formation (18) Albertus, P.; Girishkumar, G.; McCloskey, B.; Sanchez-Carrera, and Decomposition. Science 2015, 350, 530−533. R. S.; Kozinsky, B.; Christensen, J.; Luntz, A. C. Identifying Capacity (37) Liu, T.; Kim, G.; Carretero-Gonzalez,́ J.; Castillo-Martínez, E.; “ Limitations in the Li/Oxygen Battery Using Experiments and Grey, C. P. Response to comment on Cycling Li-O2 batteries via Modeling. J. Electrochem. Soc. 2011, 158, A343−A351. LiOH formation and decomposition. Science 2016, 352, 667−667. (19) Hojberg, J.; McCloskey, B. D.; Hjelm, J.; Vegge, T.; Johansen, (38) Liu, T.; Kim, G.; Carretero-Gonzalez,́ J.; Castillo-Martínez, E.; K.; Norby, P.; Luntz, A. C. An Electrochemical Impedance Bayley, P. M.; Liu, Z.; Grey, C. P. Response to Comment on “Cycling Spectroscopy Investigation of the Overpotentials in Li-O2 Batteries. Li-O2 Batteries via LiOH Formation and Decomposition. Science 2016, ACS Appl. Mater. Interfaces 2015, 7, 4039−4047. 352, 667−667. (20) Knudsen, K. B.; Luntz, A. C.; Jensen, S. H.; Vegge, T.; Hjelm, J. (39) McCloskey, B. D.; Bethune, D. S.; Shelby, R. M.; Girishkumar, Redox Probing Study of the Potential Dependence of Charge G.; Luntz, A. C. Solvents’ Critical Role in Nonaqueous Lithium− − − Transport Through Li2O2. J. Phys. Chem. C 2015, 119, 28292 28299. Oxygen Battery Electrochemistry. J. Phys. Chem. Lett. 2011, 2, 1161 (21) Meini, S.; Piana, M.; Beyer, H.; Schwammlein, J.; Gasteiger, H. 1166. A. Effect of Carbon Surface Area on First Discharge Capacity of Li-O2 (40) McCloskey, B. D.; Valery, A.; Luntz, A. C.; Gowda, S. R.; Cathodes and Cycle-Life Behavior in Ether-Based Electrolytes. J. Wallraff, G. M.; Garcia, J. M.; Mori, T.; Krupp, L. E. Combining − Electrochem. Soc. 2012, 159, A2135 A2142. Accurate O2 and Li2O2 Assays to Separate Discharge and Charge − (22) Gerbig, O.; Merkle, R.; Maier, J. Electron and Ion Transport In Stability Limitations in Nonaqueous Li O2 Batteries. J. Phys. Chem. − − Li2O2. Adv. Mater. 2013, 25, 3129 3133. Lett. 2013, 4, 2989 2993. (23) Adams, B. D.; Black, R.; Radtke, C.; Williams, Z.; Mehdi, B. L.; (41) Narayanan, S. R.; Surampudi, S.; Attia, A. I.; Bankston, C. P. Browning, N. D.; Nazar, L. F. The Importance of Nanometric Analysis of Redox Additive-Based Overcharge Protection for Passivating Films on Cathodes for Li-Air Batteries. ACS Nano 2014, 8, Rechargeable Lithium Batteries. J. Electrochem. Soc. 1991, 138, 12483−12493. 2224−2229.

755 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756 ACS Energy Letters Letter

(42) Kulyukhin, S. A. Fundamental and Applied Aspects of the Chemistry of Radioactive Iodine in Gas and Aqueous Media. Russ. Chem. Rev. 2012, 81, 960−982.

756 DOI: 10.1021/acsenergylett.6b00328 ACS Energy Lett. 2016, 1, 747−756

View publication stats