Letter pubs.acs.org/JPCL
Robust NaO2 Electrochemistry in Aprotic Na−O2 Batteries Employing Ethereal Electrolytes with a Protic Additive Iwnetim I. Abate, Leslie E. Thompson, Ho-Cheol Kim, and Nagaphani B. Aetukuri* IBM Research Almaden, 650 Harry Road, San Jose, California 95120, United States S Supporting Information *
ABSTRACT: Aprotic metal−oxygen batteries, such as Li−O2 and Na−O2 batteries, are of topical research interest as high specific energy alternatives to state-of-the-art Li-ion batteries. In particular, Na−O2 batteries with NaO2 as the discharge product offer higher practical specific energy with better rechargeability and round-trip energy efficiency when compared to Li−O2 batteries. In this work, we show that the electrochemical deposition and dissolution of NaO2 in Na−O2 batteries is unperturbed by trace water impurities in Na−O2 battery electrolytes, which is desirable for practical battery applications. We find no evidence for the formation of other discharge products such as Na2O2·H2O. Furthermore, the electrochemical efficiency during charge remains near ideal in the presence of trace water in electrolytes. Although sodium anodes react with trace water leading to the formation of a high-impedance solid electrolyte interphase, the increase in discharge overpotential is only ∼100 mV when compared to cells employing nominally anhydrous electrolytes.
B
charge-transfer-limited and does not passivate the cathode surface. This is surprising because NaO2, which is the Na−O2 battery’s primary discharge product, is also an electronic insulator24 like Li2O2 and electrode passivation due to NaO2 deposition should have limited the ultimate discharge capacity. A microscopic understanding of the crystal growth mechanism of large NaO2 crystals in Na−O2 batteries can reveal details about possible electrochemical pathways to the formation and dissolution of insulating products in metal− oxygen batteries. Such an understanding will be essential for improving rechargeability of Na−O2 batteries and potentially also for designing electrolytes that afford large discharge capacities for Li−O2 and other metal−oxygen battery systems. Consequently, many studies have focused on understanding NaO2 crystal growth mechanisms in Na−O2 batteries. There are two main mechanisms that have been proposed: (1) a solution-mediated mechanism for NaO2 crystal growth possibly resulting from a non-negligible solubility of NaO2 in glyme electrolytes23 and (2) a proton phase-transfer catalysis11 mediated NaO2 growth wherein charge transfer to the growing NaO2 crystal is mediated by a mobile hydroperoxy anion (HO2−). In the latter, the presence of trace quantities of a protic impurity such as water is essential for NaO2 crystal growth. In a previous study on Li−O2 batteries, we have shown that trace water in battery electrolytes enhances discharge capacities by promoting solution-mediated growth of Li2O2 crystals.25 A similar mechanism might also enhance discharge capacities in Na−O2 batteries. Therefore, irrespective of the
atteries with specific energies and energy densities comparable to gasoline are essential for mass adoption of electric automobiles.1,2 Metal−oxygen batteries, Li−O2 and Na−O2 batteries in particular, offer the highest theoretical specific energy among the known battery types.3−5 Li−O2 and Na−O2 batteries that use nonaqueous aprotic electrolytes enable the use of metallic lithium and sodium negative electrodes required for the practical realization of high theoretical energy density and specific energy.2,5,6 The overall electrochemical reaction in an aprotic Li−O2 cell is given by 2Li+ + 2e− + O2 ↔ Li2O2, where the forward reaction represents discharge and the reverse reaction represents charge.7−10 The analogous electrochemical reaction in an aprotic Na−O2 cell is given by Na+ + e− + O2 ↔ NaO2.5,11 Among the two battery types, Li−O2 offers a higher theoretical specific energy.2,5,11 This is because (1) lithium has a higher charge-to-mass ratio than sodium (1/7 for Li compared to 1/23 for Na) and (2) Li2O2, the discharge product of a Li−O2 cell, has a charge-to-oxygen ratio higher than that of NaO2, which is the discharge product in a Na−O2 cell (2e−/O2 for Li2O2 versus 1e−/O2 for NaO2). Consequently, aprotic Li− O2 cells are researched extensively.4,12−15 However, Li−O2 cells suffer from several limitations: poor practical specific energy, high charge overpotentials, and poor rechargeability, in particular.16−18 The poor practical specific energy in Li−O2 cells is suggested to be due to cathode passivation by the electrically insulating Li2O2, which is the batteries’ primary discharge product.18,19 Na−O2 cells, by contrast, have been shown to have relatively higher practical specific energy, low charge overpotentials, and better rechargeability.20,21 Upon discharge of a Na−O2 battery, NaO2 cubes with sides >1 μm in length are routinely observed.5,11,22,23 This is evidence that the electrochemical deposition of NaO2 in Na−O2 cells is not © XXXX American Chemical Society
Received: April 21, 2016 Accepted: May 23, 2016
2164
DOI: 10.1021/acs.jpclett.6b00856 J. Phys. Chem. Lett. 2016, 7, 2164−2169
Letter
The Journal of Physical Chemistry Letters
Figure 1. (a) Ultimate specific discharge capacity of Na−O2 cells as a function of water concentration in 1 M NaOTf in DME electrolyte. SEM images of P50 carbon cathodes after discharging to 0.6 mAh in Na−O2 cells employing (b) a nominally anhydrous 1 M NaOTf in DME as the electrolyte and (c) an electrolyte with 400 ppm water (scale bars = 10 μm). (d) e−/O2 ratio during discharge of Na−O2 cells employing electrolytes with different water concentrations as denoted in the figure. (e) Raman spectra of P50 carbon cathodes after a full discharge in cells employing the nominally anhydrous and the electrolyte with 400 ppm water. The Raman peak at 1153 cm−1 in the spectra for discharged cathodes corresponds to NaO2. The other peaks can be attributed to the P50 carbon cathode whose spectrum is shown for reference. (f) XRD pattern from cathodes discharged to 3.2 mAh in Na−O2 cells employing the nominally anhydrous electrolyte and the electrolyte with 400 ppm water (JCPDS 01-0770207). Peaks denoted by the red asterisks correspond to the P50 carbon cathode.
electrolyte. Scanning electron microscopy images (see Figure 1b,c) on discharged cathodes extracted from cells employing the nominally anhydrous electrolyte and the electrolyte with 400 ppm water show the characteristic cube morphology of NaO2 observed previously by other research groups.5,11,22,27 The presence of trace water in the electrolyte does not seem to have any measurable impact on the size of NaO2 cubes. The average lengths of the cube sides are nearly equal (∼3.9 μm; see Figure S2) for a cathode discharged with the nominally anhydrous electrolyte and for a cathode discharged in a cell employing the electrolyte with 400 ppm of H2O. These observations suggest that trace water has no measurable impact on the sizes of electrochemically deposited NaO2 cubes. This is contrary to what has been observed by us in aprotic Li−O2 cells25 and the observations of Xia et al.11 in aprotic Na−O2 cells where water concentration-dependent increases in the sizes of Li2O2 toroids and NaO2 cubes were observed, respectively. Also contrary to observations in Li−O2 batteries,25 the addition of trace quantities of water does not accentuate parasitic electrochemistry in Na−O2 cells. The e−/O2 ratio during discharge, an indicator for the ideality of the discharge process, is near unity (with only a 3% deviation across cells) for all Na−O2 cells irrespective of their water concentration (see Figure 1d). This is evidence that NaO2 is the predominant discharge product in all these cells. Consistent with this, both Raman spectroscopic investigation28 (Figure 1e) and X-ray diffraction (Figure 1f) on the discharged cathodes extracted from Na−O2 cells employing the nominally anhydrous electrolyte and the electrolyte with 400 ppm of H2O confirm
mechanism, it seems that water as an electrolyte additive might be essential for improving discharge capacities in Na−O2 batteries. However, water is known to accentuate parasitic electrochemistry leading to poor rechargeability in Li−O2 batteries.25 In this work, we investigated the impact of trace water additives in Na−O2 electrolytes on the discharge capacity and discharge and charge electrochemistry of Na−O2 cells. We found NaO2 to be the only discharge product with no evidence for any other oxides of sodium such as Na2O2, for example, irrespective of the electrolyte water concentration (≤400 ppm of H2O). Consequently, rechargeability and discharge and charge electrochemical efficiencies of aprotic Na−O2 batteries remain near ideal even in cells employing electrolytes with ∼400 ppm water. We investigate the origins and implications of the robust NaO2 electrochemistry in ethereal electrolytes containing water as a protic additive. Na−O2 cells assembled for this work are based on a Swagelok type design and are similar to the Li−O2 cells that we reported elsewhere25,26 (also see Figure S1). We used sodium as the negative electrode, 1 M NaOTf in DME as the electrolyte (with different water concentration as described below), and Avcarb P50 carbon as the positive electrode for all Na−O2 cells described in this work (see the Supporting Information for details). Figure 1a shows the specific capacity of Na−O2 cells at a current density of ∼200 μA/cm2 for a nominally anhydrous electrolyte (
