Solvating additives drive solution-mediated electrochemistry and enhance toroid growth in non-aqueous Li–O2 batteries



Given their high theoretical specific energy, lithium–oxygen batteries have received enormous attention as possible alternatives to current state-of-the-art rechargeable Li–ion batteries. However, the maximum discharge capacity in non-aqueous lithium–oxygen batteries is limited to a small fraction of its theoretical value due to the build-up of insulating lithium peroxide (Li2O2), the battery’s primary discharge product. The discharge capacity can be increased if Li2O2 forms as large toroidal particles rather than as a thin conformal layer. Here, we show that trace amounts of electrolyte additives, such as H2O, enhance the formation of Li2O2 toroids and result in significant improvements in capacity. Our experimental observations and a growth model show that the solvating properties of the additives prompt a solution-based mechanism that is responsible for the growth of Li2O2 toroids. We present a general formalism describing an additive’s tendency to trigger the solution process, providing a rational design route for electrolytes that afford larger lithium–oxygen battery capacities.


The recent surge in activity in the search for batteries with energy densities surpassing those possible with Li-ion intercalation technology has been fuelled by the goal of developing mass-market electrification of road transportation. The non-aqueous Li–air battery has attracted the most attention to date because of its very high theoretical specific energy1,2. In this battery, the net electrochemical reaction is 2Li + O2 Li2O2, with the forward reaction describing discharge of the battery and the reverse reaction describing charge3,4,5,6,7,8. The high theoretical specific energy arises from the use of Li metal as the anode and ambient air as the source of O2. However, there are still substantial technical obstacles to developing a practical Li–air battery9,10. Perhaps the most significant challenges arise from parasitic chemistry and electrochemistry during battery cycling11,12,13,14,15,16,17,18 and electrical passivation of the cathode during discharge19,20,21,22, with the former limiting rechargeability and the latter limiting the capacity to less than theoretically possible, especially at higher current densities (implying a poor capacity–power tradeoff in the battery20). The electrical passivation is caused by the build-up of Li2O2, a wide-bandgap insulator, during discharge3,19,20,21,22,23. This inhibits charge transfer from the cathode to the Li2O2–electrolyte interface where the discharge electrochemistry occurs.

With ethereal electrolytes, many authors24,25,26,27,28,29 have reported that large Li2O2 toroids of variable sizes (100 nm–1 µm) are produced during discharge at low currents. However, during the course of our studies with nearly anhydrous ethereal electrolytes, we never observed toroid formation at any current, only seeing the apparent formation of thin conformal coatings of Li2O2 on the cathode surface. Understanding the origin of the large toroid features and ultimately controlling the morphology of Li2O2 growth during discharge is extremely critical to achieving high discharge capacities, because these toroids circumvent the Li2O2 charge transport limitations.

In this Article, we combine experimental measurements with theoretical modelling to show that there are two possible paths for Li2O2 growth on the cathode. One involves a surface electrochemical mechanism (as previously described8,30) that produces conformal Li2O2 cathode surface coatings with thicknesses limited by charge transport through the Li2O2. We propose that the second path is a solution-mediated electrochemical process driven by LiO2 partial solubility, where O2 acts as a redox mediator and ultimately promotes the growth of Li2O2 toroids at low currents. Although solution-mediated processes have previously been suggested for the growth of Li2O2 (refs 4, 5, 24, 25), neither the detailed electrochemical origins nor the conditions that favour the growth of large (~1 µm) Li2O2 structures have been described. In ethereal solvents we demonstrate that toroidal growth and a concomitant increase in discharge capacity are only observed when trace quantities of H2O are added to the electrolyte. However, the presence of water also results in Faradaic inefficiencies during battery operation. As a result, understanding how H2O enhances the solubility of LiO2 provides insight into the development of appropriate solvents/additives that may induce solution-growth of Li2O2 (thereby enhancing the capacity of Li–O2 batteries), but without the drawbacks that added H2O introduces.

Results and discussion

Figure 1 presents scanning electron microscopy (SEM) images of Vulcan XC72 carbon cathodes (XC72) extracted from Li–O2 batteries after galvanostatic discharge to 1 mAh capacity at a current of 50 µA with 0–4,000 ppm H2O added to the electrolyte (1 M lithium bis(trifluoromethane sulfonyl)imide (LiTFSI) in dimethoxyethane, DME). The Li–O2 batteries are identical to those used for differential electrochemical mass spectrometry (DEMS) experiments discussed previously11,12. Experimental procedures are described in the Supplementary Methods. With no added H2O, the XC72 cathode after discharge is indistinguishable from the pristine cathode before discharge. This suggests that the Li2O2 is only deposited as thin conformal films on the cathode surface. However, after the addition of 500 ppm H2O, small thin toroids (100–200 nm in size) become visible in the SEM images. As the H2O content increases, increasingly larger toroids and platelets are formed, and layering within the toroids becomes more evident.

Figure 1: Li2O2 discharge product morphology control.

af, SEM images of a Vulcan XC72 carbon cathode without any discharge (a) and of similar cathodes discharged to a capacity of 1 mAh at a rate of 50 µA using nominally anhydrous (<30 ppm) 1 M LiTFSI in DME as the electrolyte (b) and with water contents of 500 ppm (c), 1,000 ppm (d), 2,000 ppm (e) and 4,000 ppm (f) in the electrolyte. The size of the Li2O2 toroids increases with the amount of water in the electrolyte. The thin platelets observed in d and f probably increase in size and number of layers to form toroids. All scale bars, 1 µm.

Figure 1 clearly shows that, in ethereal electrolytes, the existence of Li2O2 toroids and platelets (as well as their shape, size and abundance) depends on the concentration of added H2O in the electrolyte. Previous reports24,26,27,28,29,31,32 have described toroids and platelet structures with a range of shapes and sizes in ethereal electrolytes, and we suggest that this is probably due to varying levels of water contamination in the cells and different discharge currents. Supplementary Fig. 1 shows a diminishment in toroid particle diameter with increasing current at a fixed H2O content, with no toroids and only small platelet deposits present at currents greater than 1 mA for an H2O concentration of 4,000 ppm. At low H2O content, although toroids are still observed at very low currents, they disappear at currents higher than 1 mA, in general agreement with previous observations24,29,33. Figure 1 and equivalent experiments at a discharge current of 250 µA (Supplementary Fig. 6) show that the final discharge product morphology depends on both the discharge current and the H2O content in the electrolyte. At high H2O contents, toroid and platelet formation occurs up to quite high currents, whereas at low H2O content, toroid formation disappears at quite low current rates. Only in the nominally anhydrous DME/LiTFSI electrolyte does toroid/platelet formation disappear altogether, regardless of the current rate. Supplementary Fig. 7 shows that toroid formation is not observed in the nominally anhydrous electrolyte at current rates as low as 10 µA.

We also observed similar discharge morphology changes with increasing H2O content in other cathodes, such as TiC and AvCarb P50 carbon paper (P50) (Supplementary Figs 2 and 3). Similar Li2O2 morphology is observed when comparable quantities of water are added to other electrolyte solvents, for example, tetraethylene glycol dimethyl ether (TEGDME) and dimethylsulfoxide (DMSO) (Supplementary Figs 4 and 5), suggesting that the presence of H2O in the electrolyte also causes drastic Li2O2 morphology changes in these electrolytes.

The discharge capacity also increases as H2O concentration increases (Fig. 2a), in agreement with previous studies27,34,35,36. Even at high currents (3 mA) where no toroid formation is apparent, an increase in discharge capacity is observed when water is present (Supplementary Fig. 14). We argue that the improvement in capacity at both low and high currents arises due to a solution-mediated mechanism for Li2O2 formation (discussed later) that overcomes the charge transport limitations inherent in the surface growth of Li2O2. It is well known12 that Li2O2 reacts quantitatively with pure H2O to form LiOH and H2O2. However, this reaction cannot account for the full increase in capacity with added H2O, as shown in Supplementary Fig. 15, suggesting that toroid/platelet formation from a solution process is responsible for the increase in capacity. Although it is clear that trace H2O has a positive impact on capacity, it is also critical to understand its effect on the battery chemistry and rechargeability.

Figure 2: Discharge capacity increase with increasing water content in the electrolyte.

a, Experimental gravimetric discharge capacities for batteries with XC72 carbon cathodes and 1 M LiTFSI in DME with varying water content in the battery electrolyte. The experimental discharge capacities were obtained from galvanostatic discharges to a reductive potential of 2.3 V (versus Li/Li+) at a discharge rate of 250 µA. Error bars represent uncentainty in the gravimetric capacity calculation and are computed from the standard deviation of carbon loading in the cathodes. b, Theoretically predicted discharge capacities from the developed electrochemical model. A cathode surface area of ~200 cm2 was assumed for the capacity calculation. The model predicts an approximately fivefold enhancement due to the addition of water through triggering of the solution mechanism. Dotted lines are guides to the eye.

Figure 3a presents X-ray diffractograms (XRDs) near the Li2O2 (100) and (101) peaks from Avcarb P50 paper cathodes extracted from batteries that were otherwise similar to those studied in Fig. 1. The only additional H2O-induced XRD feature is a small peak at 30.65° that has tentatively been identified as Li2NH (Supplementary Fig. 9), a likely decomposition product of the electrolyte salt, LiTFSI. These results confirm that the majority of the crystalline discharge product is Li2O2, regardless of electrolyte water content. Notably, no crystalline LiOH is observed in the XRDs, Raman spectroscopy (Supplementary Fig. 17 and Supplementary Section 2) or Fourier-transform infrared (FTIR) spectroscopy1 of the cathodes. The Li2O2 diffractograms clearly show a decreasing peak width as a function of increasing water content in the electrolyte solution, implying that the Li2O2 crystallite size increases, in agreement with the SEM images shown in Fig. 1. Other authors have shown a decrease in XRD linewidth with current (presumably at fixed H2O content) as the toroid size increases24,32,37. For the nominally anhydrous battery, there was no apparent change in crystallite size with current (Fig. 3b) at an equal discharge capacity of 2 mAh. Therefore, XRD provides no evidence for a transformation to an amorphous Li2O2 deposit at higher current rates as suggested by others24,38.

Figure 3: Ex situ XRD measurements on discharged cathodes.

a, θ −2θ X-ray diffractograms near the Li2O2 (100) and (101) resonance peaks on P50 carbon cathodes discharged at a rate of 250 µA in cells using 1 M LiTFSI in DME with varying water contents (shown in the Legend) as electrolytes. The diffractograms show a strong narrowing of the (100) and (101) peaks with increased water content in the electrolyte, which is consistent with the increasing Li2O2 crystallite size evident from the SEM images in Fig. 1. b, Similar diffractograms collected on cathodes discharged at different discharge rates (shown in the legend) using the nominally anhydrous 1 M LiTFSI in DME electrolyte. No changes in peak width were observed, even with a nearly two orders of magnitude change in the discharge current. This suggests that the Li2O2 remains crystalline and the changes in crystallite size, if any, are below the instrumental resolution in all samples discharged under anhydrous conditions, irrespective of the discharge rate. All cathodes were discharged to a discharge capacity of 2 mAh and all the curves are normalized to the carbon cathode's (002) XRD peak (not shown).

To quantify the effects of added H2O on the electrochemistry and possible parasitic reactions, both quantitative DEMS and a Li2O2 titration were performed11,12. If no parasitic electrochemistry (or chemistry) occurs during discharge, the expected yield of Li2O2 produced relative to the theoretical yield from the discharge capacity is unity (). Similarly, two electrons are ideally utilized for each O2 consumed, (e/O2)dis = 2.00, where O2 consumption is monitored using a pressure decay measurement. Supplementary Figs 11 and 12 show (e/O2)dis and , respectively, for cells with varying water content. The values without added H2O are consistent with previous measurements12 for these low current conditions, that is, for an XC72 cathode and (e/O2)dis = 2.02. However, both deviate further from their ideal values as H2O is added, demonstrating that the added H2O induces parasitic processes during discharge. We also note that a titration of both solvent and cathode yields a higher total peroxide content than that from the cathode itself, which we believe is indirect evidence for soluble H2O2 formation that can form via superoxide disproportionation in the water-contaminated electrolytes (Supplementary Section 2).

Because H2O induces parasitic (electro)chemistry during discharge, it is not surprising that H2O impurities also affect the charging potential. Galvanostatic discharge–charge cycles as a function of H2O content are presented in Supplementary Fig. 13. These profiles demonstrate that although the initial charging potential Uchg is nearly identical in all cases, the rate of increase in Uchg with charging capacity Qchg is strongly dependent on H2O content. We have previously suggested that the initial Uchg is indicative of the low Li2O2 fundamental kinetic overpotentials30, but that the increase in Uchg with Qchg is related to the role of parasitic products in charging39. This interpretation for Supplementary Fig. 13 is entirely consistent with the enhanced parasitic chemistry associated with added H2O (Supplementary Figs 11 and 12). Comparing the performance and (electro)chemistry of the anhydrous cells with those with trace amounts of H2O raises two questions. First, how does H2O increase capacity and induce toroid formation? Second, is it possible to find another additive that can give the positive benefits of added H2O without its drawbacks?

As the dimensions of the Li2O2 toroids and platelets are significantly larger than the charge transport-limited dimensions of 1–10 nm, we argue that they must be formed by a solution-mediated mechanism that also contributes to the electrochemistry, as also suggested by others24,27. Thus, the net battery discharge is the sum of the two different contributions. Figure 4a shows discharge linear scan voltammograms (LSVs), both with and without added H2O in DME. The anhydrous DME discharge LSV is similar to that observed previously8 and was assigned to the surface electrochemical process8,30 producing Li2O2 (the peak at ~2.5 V in Fig. 4a). The discharge LSV curve with 4,000 ppm water (Fig. 4b) shows an additional peak at potentials lower than ~2.5 V. We suggest that this additional peak in the LSV is related to the solution-mediated growth of Li2O2. This additional peak represents an electrochemical process principally forming Li2O2, as indicated by a peroxide titration ( following the LSV with 4,000 ppm added H2O in DME; for the nominally anhydrous DME). Note that the onset potential in the LSV curve with 4,000 ppm added H2O is shifted to lower potentials. Accordingly, current from both surface and solution processes is possible under typical galvanostatic conditions (where a potential plateau ranges from 2.4 to 2.7 V). Hence, both the surface and solution electrochemical processes contribute to galvanostatic discharges.

Figure 4: The two pathways for Li2O2 formation.

a,b, Discharge LSVs performed at 0.05 mV s−1 with a Vulcan XC72 carbon cathode and lithium anode using 1 M LiTFSI based electrolyte solutions in nominally anhydrous DME (a) and DME with 4,000 ppm water (b). The anhydrous DME sample shows a single sharp peak in the LSV at ~2.5 V, whereas the LSV curve for the cell with 4,000 ppm water exhibits a distinct second peak at ~2.3 V. We attribute the first peak to surface electrochemical growth of Li2O2 and the second peak to the solution-mediated growth of Li2O2, where O2 acts as a redox mediator. c, Theoretically predicted discharge LSV curves for the two independent mechanisms using the developed electrochemical model. The peak currents and the relative potential differences between the two peaks from theory are in good agreement with experiment. The differences in the absolute potentials from theory and experiment could be due to the cell impedance, which is not subtracted from the experimental LSV curves.

The schematic in Fig. 5 summarizes the two electrochemical paths for Li2O2 crystal growth. The surface electrochemical growth is given, as before30, by

where * refers to a surface-adsorbed species, (s) indicates solid and sol indicates a species in solution. Both Li+ + e charge transfer or disproportionation (2LiO2* → Li2O2(s) + O2) can contribute to the second step of growth8,40. We also consider a second slower possible route for the growth of Li2O2 crystals induced by the generation of soluble reduced oxygen species in the presence of H2O. The dominant Li2O2 equilibrium surface produced by reaction (1) is the O-rich (0001) surface, that is, a Li2O2 surface with half a monolayer of LiO2 adsorbed23,30,41. Its solubility is given as

Figure 5: Proposed mechanism for the growth of Li2O2 toroids in the presence of water.

a, The deposition of Li2O2 in a Li–O2 cell is shown schematically to proceed via a surface electrochemical growth process that occurs on a nucleated film of Li2O2 through the sequential transfer of solvated Li+ (Li+(sol)) and an electron (e) to the intermediate species LiO2*, and eventually forming Li2O2. The electron must therefore tunnel through the nucleated Li2O2 film as indicated, and this process is limited by the electronic conductivity of Li2O2. The presence of a solvent (water in our experiments) that solvates LiO2* to Li+(sol) and solvated O2 (O2(sol)) triggers a solution pathway leading to the growth of toroids, as shown schematically. O2(sol) adsorbs as LiO2* on the growing toroidal particle, ultimately disproportionating to form Li2O2. Thus, O2(sol) acts as a redox shuttle and leads to the formation of large particles, thereby circumventing the conductivity limitations in the surface electrochemical growth. b,c, Li2O2 toroidal particle size predicted by the particle growth model for an electrolyte containing 1,000 ppm of water (b) and 4,000 ppm of water (c). Larger-sized discharge products are observed at higher water contents, consistent with the experimental observations.

The equilibrium is governed by the stability of the Li+ and O2 ions in solution relative to the LiO2* adsorption energy on Li2O2. As described in detail in the Supplementary Section 4, the stability of Li+ and O2 ions in solution is related to the Gutman donor number (DN) and acceptor number (AN), respectively, so the LiO2* solubility depends on these parameters. The addition of water triggers this dissolution process by solvating O2 efficiently due to its very high AN of ~55. We believe that the solution soluble O2 undergoes subsequent reaction on a growing Li2O2 toroid through the generic mechanism

Many different detailed mechanisms could contribute to reaction (3). However, the key point is that Li2O2 solution growth uses O2(sol) as a redox shuttle. The most likely mechanism (shown in Fig. 5a) is that (a) LiO2* solvates in an equilibrium fashion from the O-rich (0001) Li2O2 that conformally coats the cathode produced via the surface process (1) (because this is the dominant surface area of O-rich (0001) Li2O2 formed), (b) O2(sol) diffuses to a growing particle where it again forms LiO2*, (c) two such LiO2* disproportionate to form Li2O2 on a larger growing particle, and (d) LiO2* regenerates via reaction (1) at the empty site on the conformal layer. It is worth highlighting that the anhydrous DME has an AN of ~10 and therefore does not induce enough solubility of LiO2* for the solution growth of toroids/platelets, even at discharge currents as low as 10 µA.

In a partially protic solvent (for example, DME with added H2O), O2(sol) is known to undergo disproportionation, ultimately forming H2O2. H2O2 formation is a relatively slow step in mixed aprotic–H2O solvents with a modest H2O concentration42. However, this step, along with a reaction between H2O and Li metal, slowly consumes the H2O and eventually this reduction of water reduces the overall dissolution rate of LiO2* and ultimately terminates the solution growth mechanism.

Based on this mechanism, we developed an electrochemical model that accounts for the simultaneous surface and solution routes to the formation of Li2O2. Where appropriate we make comparisons with the experiments discussed above. This model includes the disproportionation of O2(sol) in the presence of water (Supplementary Section 4). The evolution of the different chemical species (H2O, O2(sol) and H2O2) during galvanostatic discharge for several different levels of added H2O is shown in Supplementary Fig. 22. The significant increase in the concentration of O2(sol) in the presence of H2O promotes the solution mechanism in addition to the surface electrochemical growth of Li2O2, and therefore increases the discharge capacity (Fig. 2b). However, the solution route ultimately shuts off due to a decrease in soluble O2(sol) as the H2O is consumed by conversion to H2O2 and by reaction with the Li anode. This consumption leads to a decreased solubility of O2(sol), which ultimately determines the maximum discharge capacity for a given H2O content. The maximum discharge capacity, therefore, is a function of the H2O content in the electrolyte. The results from the model are summarized in Fig. 2b and are in qualitative agreement with Fig. 2a, that is, an approximately fivefold enhancement in discharge capacity for the 4,000 ppm case relative to the anhydrous case. Using the same model, we also simulated the LSV for 4,000 ppm added H2O (Fig. 4c). This clearly demonstrates the existence of two distinct peaks associated with the surface and solution electrochemical growth routes and is in good agreement with Fig. 4b. Note that the LSV experiments are shifted to lower potentials due to resistive losses at these currents21.

Clearly, the solution mechanism allows particle sizes (and capacities) larger than the few-nanometre dimensions defined by charge transport. Understanding how particles dynamically grow in such a process is challenging to probe experimentally and model theoretically. At low currents, large layered toroids are observed. At higher currents (and/or lower H2O content), much smaller particle sizes are formed. In the Supplementary Section 4 we present a plausible kinetic model for the solution growth of these particles. It is based on assuming three coupled kinetic processes: a solution growth rate of Li2O2 on the particles, a passivation rate that helps terminate the solution growth (for example, formation of Li2NH on the Li2O2 particle surface), and a rate for formation of defects/holes in the passivation layer. With reasonable kinetic rates and at low currents, layered toroids are naturally formed via this mechanism, with sizes dependent upon the H2O concentration and overall current. Two examples are shown in Fig. 5b and c, and an animation of toroid growth is provided in Supplementary Movie 1. At higher currents, diffusion limitations of O2(sol) restrict the size of the growing particles, and at lower H2O concentrations, the shorter times (capacities) available for solution growth restrict the particle size (Supplementary Fig. 1). An extended discussion of this model and its relationship to the parameters of the solution mechanism are presented in the Supplementary Section 4.

Added H2O enhances the discharge capacity in ether solvents by inducing a solution mechanism for the growth of Li2O2, but it also induces enhanced parasitic chemistry. A key question is whether other additives/solvents can also induce the solution growth mechanism, but perhaps without the additional parasitic chemistry. As discussed in detail in Supplementary Section 4, the solubility of LiO2* is determined by the Gutman DN and AN of the solvent. Based on experimental measurements on the redox potential shifts for O2/O2 and Li/Li+, we developed an expression for the relative free energy of dissolution, equation (2), on any solvent (see Supplementary Section 4 for details). Figure 6 shows a contour plot of the free energy of dissolution as a function of AN and DN with several known solvents, as indicated in the plot. DME and acetonitrile (MeCN) have a limited propensity to solvate LiO2* and are therefore ineffective in promoting solution growth. Of the pure aprotic solvents, DMSO is relatively active because of its high DN. In fact, cathodic LSVs of cells with DMSO-based electrolytes (Supplementary Fig. 19) also exhibit a weak second peak ascribed to the solution-mediated Li2O2 formation mechanism. Furthermore, very small toroids are observed on cathodes that are galavanostatically discharged in nearly anhydrous DMSO (~30 ppm water content, Supplementary Fig. 5) at low currents, in agreement with Fig. 6, but these grow substantially in size (and layering) with added H2O. Another possible additive with strong solvating properties is CH3OH. The addition of 4,000 ppm CH3OH to DME does increase the maximum discharge capacity by approximately three times relative to pure DME at 100 µA current (Supplementary Fig. 14a). An LSV with methanol added to DME (Supplementary Fig. 18) also shows an additional peak (although the peak current is smaller compared to H2O) that we attribute to the solution growth of Li2O2. Unfortunately, the pKa of protic CH3OH is slightly lower than for H2O, so a galvanostatic discharge also produces H2O2 and a of ~0.5 (0.4 mAh discharge capacity).

Figure 6: Quantitative basis for solvent selection for high-capacity Li–O2 batteries.

The free energy of dissolution for LiO2* into Li+(sol) and O2(sol) in different solvents as a function of the Gutman acceptor and donor numbers (AN and DN). The free energy plot is normalized relative to that of pure DME. Dimethyl formamide (DMF), dimethyl acetamide (DMAc) and dimethyl sulfoxide (DMSO) have high DNs and are thus capable of stabilizing Li+. Water and methanol, on the other hand, have high ANs and thus stabilize O2. We predict that solvents that fall in the top-right quadrant of this plot will favour solution-mediated deposition of Li2O2, which will be essential for high-capacity Li–O2 batteries.

The enhanced discharge capacity and the growth of large Li2O2 toroids in the presence of added water is definitive evidence that the discharge capacity of Li–O2 batteries need not be limited by the surface growth route and the electronically insulating nature of Li2O2. The solubility of O2 in the battery electrolyte can activate a mechanism where O2 acts as a redox mediator for the electrochemical growth of Li2O2 that is not limited by the charge transport of Li2O2. However, the addition of water increases parasitic electrochemistry and is therefore not a good additive, even if the lithium anode is protected by a water-impermeable membrane. Fortunately, the examples above show that this solution route is not limited only to added H2O and validates the analysis that led to the predictions of Fig. 6. Solvents/additives that allow for increased solubility of LiO2* will be key to obtaining large discharge capacities, and we have developed a quantitative basis for the rational selection of solvents based on their AN and DN. The desired additive must have a high DN and/or AN while at the same time satisfying the stringent criterion of having a high pKa to avoid H2O2 formation and related parasitic processes. The rational selection criteria we have provided could pave the way for the identification of additives that can enhance the discharge capacity, while still optimizing the rechargeability of the Li–O2 battery.


  1. 1

    Imanishi, N., Luntz, A. C. & Bruce, P. G. The Lithium Air Battery (Springer, 2014).

    Google Scholar 

  2. 2

    Bruce, P. G., Freunberger, S. A., Hardwick, L. J. & Tarascon, J-M. Li–O2 and Li–S batteries with high energy storage. Nature Mater. 11, 19–29 (2012).

    CAS  Google Scholar 

  3. 3

    Hummelshoj, J. S. et al. Communications: elementary oxygen electrode reactions in the aprotic Li–air battery. J. Chem. Phys. 132, 071101 (2010).

    CAS  PubMed  Google Scholar 

  4. 4

    Laoire, C. O., Mukerjee, S., Abraham, K. M., Plichta, E. J. & Hendrickson, M. A. Influence of nonaqueous solvents on the electrochemistry of oxygen in the rechargeable lithium–air battery. J. Phys. Chem. C 114, 9178–9186 (2010).

    CAS  Google Scholar 

  5. 5

    Laoire, C. O., Mukerjee, S., Abraham, K. M., Plichta, E. J. & Hendrickson, M. A. Elucidating the mechanism of oxygen reduction for lithium–air battery applications. J. Phys. Chem. C 113, 20127–20134 (2009).

    CAS  Google Scholar 

  6. 6

    Abraham, K. M. & Jiang, Z. A polymer electrolyte-based rechargeable lithium/oxygen battery. J. Electrochem. Soc. 143, 1–5 (1996).

    CAS  Google Scholar 

  7. 7

    Ogasawara, T., Débart, A., Holzapfel, M., Novák, P. & Bruce, P. G. Rechargeable Li2O2 electrode for lithium batteries. J. Am. Chem. Soc. 128, 1390–1393 (2006).

    CAS  Google Scholar 

  8. 8

    McCloskey, B. D., Scheffler, R., Speidel, A., Girishkumar, G. & Luntz, A. C. On the mechanism of nonaqueous Li–O2 electrochemistry on C and its kinetic overpotentials: some implications for Li–air batteries. J. Phys. Chem. C 116, 23897–23905 (2012).

    CAS  Google Scholar 

  9. 9

    Girishkumar, G., McCloskey, B., Luntz, A. C., Swanson, S. & Wilcke, W. Lithium–air battery: promise and challenges. J. Phys. Chem. Lett. 1, 2193–2203 (2010).

    CAS  Google Scholar 

  10. 10

    Christensen, J. et al. A critical review of Li/air batteries. J. Electrochem. Soc. 159, R1–R30 (2011).

    Google Scholar 

  11. 11

    McCloskey, B. D. et al. Limitations in rechargeability of Li–O2 batteries and possible origins. J. Phys. Chem. Lett. 3, 3043–3047 (2012).

    CAS  PubMed  Google Scholar 

  12. 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).

    CAS  PubMed  Google Scholar 

  13. 13

    Freunberger, S. A. et al. The lithium–oxygen battery with ether-based electrolytes. Angew. Chem. Int. Ed. 50, 8609–8613 (2011).

    CAS  Google Scholar 

  14. 14

    Shao, Y. et al. Making Li–air batteries rechargeable: material challenges. Adv. Funct. Mater. 23, 987–1004 (2013).

    CAS  Google Scholar 

  15. 15

    Bryantsev, V. S. et al. Predicting solvent stability in aprotic electrolyte Li–air batteries: nucleophilic substitution by the superoxide anion radical (O2•–). J. Phys. Chem. A 115, 12399–12409 (2011).

    CAS  PubMed  Google Scholar 

  16. 16

    Bryantsev, V. S. et al. The identification of stable solvents for nonaqueous rechargeable Li–air batteries. J. Electrochem. Soc. 160, A160–A171 (2013).

    CAS  Google Scholar 

  17. 17

    Assary, R. S., Lau, K. C., Amine, K., Sun, Y-K. & Curtiss, L. A. Interactions of dimethoxy ethane with Li2O2 clusters and likely decomposition mechanisms for Li–O2 batteries. J. Phys. Chem. C 117, 8041–8049 (2013).

    CAS  Google Scholar 

  18. 18

    Younesi, R., Norby, P. & Vegge, T. A new look at the stability of dimethyl sulfoxide and acetonitrile in Li–O2 batteries. ECS Electrochem. Lett. 3, A15–A18 (2014).

    CAS  Google Scholar 

  19. 19

    Albertus, P. et al. Identifying capacity limitations in the Li/oxygen battery using experiments and modeling. J. Electrochem. Soc. 158, A343–A351 (2011).

    CAS  Google Scholar 

  20. 20

    Luntz, A. C. et al. Tunneling and polaron charge transport through Li2O2 in Li–O2 batteries. J. Phys. Chem. Lett. 4, 3494–3499 (2013).

    CAS  Google Scholar 

  21. 21

    Viswanathan, V. et al. Electrical conductivity in Li2O2 and its role in determining capacity limitations in non-aqueous Li–O2 batteries. J. Chem. Phys. 135, 214704 (2011).

    CAS  PubMed  Google Scholar 

  22. 22

    Radin, M. & Siegel, D. Charge transport in lithium peroxide: relevance for rechargeable metal-air batteries. Energy Environ. Sci. 6, 2370–2379 (2013).

    CAS  Google Scholar 

  23. 23

    Radin, M. D., Feng, T. & Siegel, D. J. Electronic structure of Li2O2{0001} surfaces. J. Mater. Sci. 47, 7564–7570 (2012).

    CAS  Google Scholar 

  24. 24

    Adams, B. D. et al. Current density dependence of peroxide formation in the Li–O2 battery and its effect on charge. Energy Environ. Sci. 6, 1772–1778 (2013).

    CAS  Google Scholar 

  25. 25

    Black, R., Adams, B. & Nazar, L. F. Non-aqueous and hybrid Li–O2 batteries. Adv. Energy Mater. 2, 801–815 (2012).

    CAS  Google Scholar 

  26. 26

    Zhai, D. et al. Disproportionation in Li–O2 batteries based on a large surface area carbon cathode. J. Am. Chem. Soc. 135, 15364–15372 (2013).

    CAS  PubMed  Google Scholar 

  27. 27

    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. Nature Commun. 4, 2438 (2013).

    Google Scholar 

  28. 28

    Fan, W., Cui, Z. & Guo, X. Tracking formation and decomposition of abacus-ball-shaped lithium peroxides in Li–O2 cells. J. Phys. Chem. C 117, 2623–2627 (2013).

    CAS  Google Scholar 

  29. 29

    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).

    CAS  PubMed  Google Scholar 

  30. 30

    Hummelshoj, J. S., Luntz, A. C. & Norskov, J. K. Theoretical evidence for low kinetic overpotentials in Li–O2 electrochemistry. J. Chem. Phys. 138, 034703–034712 (2013).

    CAS  PubMed  Google Scholar 

  31. 31

    Ottakam Thotiyl, M. M. et al. A stable cathode for the aprotic Li–O2 battery. Nature Mater. 12, 1050–1056 (2013).

    CAS  Google Scholar 

  32. 32

    Gallant, B. M. et al. Influence of Li2O2 morphology on oxygen reduction and evolution kinetics in Li–O2 batteries. Energy Environ. Sci. 6, 2518–2528 (2013).

    CAS  Google Scholar 

  33. 33

    Horstmann, B. et al. Rate-dependent morphology of Li2O2 growth in Li–O2 batteries. J. Phys. Chem. Lett. 4, 4217–4222 (2013).

    CAS  PubMed  Google Scholar 

  34. 34

    Guo, Z., Dong, X., Yuan, S., Wang, Y. & Xia, Y. Humidity effect on electrochemical performance of Li–O2 batteries. J. Power Sources 264, 1–7 (2014).

    CAS  Google Scholar 

  35. 35

    Cho, M. H. et al. The effects of moisture contamination in the Li–O2 battery. J. Power Sources 268, 565–574 (2014).

    CAS  Google Scholar 

  36. 36

    Meini, S., Piana, M., Tsiouvaras, N., Garsuch, A. & Gasteiger, H. A. The effect of water on the discharge capacity of a non-catalyzed carbon cathode for Li–O2 batteries. Electrochem. Solid-State Lett. 15, A45–A48 (2012).

    CAS  Google Scholar 

  37. 37

    Jung, H-G. et al. A transmission electron microscopy study of the electrochemical process of lithium–oxygen cells. Nano. Lett. 12, 4333–4335 (2012).

    CAS  PubMed  Google Scholar 

  38. 38

    Tian, F., Radin, M. D. & Siegel, D. J. Enhanced charge transport in amorphous Li2O2 . Chem. Mater. 26, 2952–2959 (2014).

    CAS  Google Scholar 

  39. 39

    McCloskey, B. D. et al. Twin problems of interfacial carbonate formation in nonaqueous Li–O2 batteries. J. Phys. Chem. Lett. 3, 997–1001 (2012).

    CAS  Google Scholar 

  40. 40

    Peng, Z. et al. Oxygen reactions in a non-aqueous Li+ electrolyte. Angew. Chem. Int. Ed. 50, 6351–6355 (2011).

    CAS  Google Scholar 

  41. 41

    Mo, Y., Ong, S. P. & Ceder, G. First-principles study of the oxygen evolution reaction of lithium peroxide in the lithium–air battery. Phys. Rev. B 84, 205446 (2011).

    Google Scholar 

  42. 42

    Che, Y. et al. Water-induced disproportionation of superoxide ion in aprotic solvents. J. Phys. Chem. 100, 20134–20137 (1996).

    CAS  Google Scholar 

Download references


The authors thank R. Shelby for Raman measurements, D. Bethune and G. Wallraff for discussions and help with experiments and the IBM model shop for support with the DEMS system. N.B.A. acknowledges guidance from H.C. Kim and W.W. Wilcke. V.V. is supported by a faculty startup grant from Carnegie Mellon University.

Author information




All authors contributed to the design of the research. N.B.A., J.M.G. and L.E.K. performed the experimental measurements and N.B.A. performed the experimental data analysis. V.V. and A.C.L. designed the theoretical calculations, which V.V. then performed. N.B.A., B.D.M., V.V. and A.C.L. co-wrote the manuscript. All authors discussed the results and commented on the manuscript.

Corresponding authors

Correspondence to Nagaphani B. Aetukuri or Venkatasubramanian Viswanathan.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Supplementary information

Supplementary information (PDF 15472 kb)


Supplementary movie 1 (MP4 175 kb)

Supplementary movie 1

Supplementary movie 1 (MP4 175 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Aetukuri, N., McCloskey, B., García, J. et al. Solvating additives drive solution-mediated electrochemistry and enhance toroid growth in non-aqueous Li–O2 batteries. Nature Chem 7, 50–56 (2015).

Download citation

Further reading