Goodenough-JB: central claims
Documents: one paper (the only file in the manifest), Braga, Grundish, Murchison and Goodenough, "Alternative strategy for a safe rechargeable battery", Energy & Environmental Science (2016). Read through its text layer, a digest, and Opus deep reads of Fig. 2 (p.4) and Figs. 3-9 (pp.5-7) from rendered page images; Fig. 1 (p.3) was read as an image by earlier reviewers. Values quoted from figures are digitised (about +-2 % of full scale) unless the paper states them.
Source gap: The foundational works the paper relies on were not provided and not read: Braga et al. 2014 (J. Mater. Chem. A, Li3ClO glasses), Braga et al. 2016 (Energy Environ. Sci. 9, 948, symmetric cells) and Braga et al. 2017 (J. Electrochem. Soc. 164, A1, electric dipoles in Na glass). The conductivity, wetting and dipole claims are asserted in this paper by citation only. Raw data, any independent replications, and published critiques or replies (a comment and reply in Energy Environ. Sci. are recalled [memory, medium] but not verified) are also missing. Obtain those three papers and the comment and reply. Later steps must say plainly when a conclusion rests on [memory] alone.
- A glass electrolyte based on Li or Na, oxygen and chlorine (A2.99Ba0.005O(1+x)Cl(1-2x)) conducts the alkali ion at above 1e-2 S/cm at 25 C (the text; Fig. 9 is annotated 5-10 mS/cm, at or below that) with a very low barrier (0.06 eV) and a wide stability window (above 9 eV in the text, "E_g > 8 V" in Fig. 4).
- That glass is wetted by lithium or sodium metal, forms no passivating interphase (SEI), and allows reversible, dendrite-free plating and stripping of the metal anode; the authors' only direct evidence is a symmetric cell run at +-3 mA/cm2 (1 h half-cycles, about 79 cycles over about 160 h of a 200 h trace) with an overpotential of a few mV, which does not distinguish good wetting from a soft short.
- Its electric dipoles, after about 10 days of ageing, form ferroelectric condensates giving a very large dielectric constant and large double-layer capacitances at both electrodes, and the cell voltage and its drift (rise to about 2.65 V at 10 days, decay to 1.65 V by about 30 days, fitted with free time constants of 94 h and 346 h) are explained by two such capacitors in series (Vcell = Q/Ceq) rather than by ordinary electrode potentials and diffusion; no capacitance was measured.
- In a "plating" cell the anode metal is deposited on a copper/carbon cathode current collector during discharge (shown by photos and back-scatter SEM; no usable EDS, since EDS cannot detect Li), so capacity is set by the anode metal and not by a host cathode; a redox center sets the voltage without being consumed (Li-S about 2.5-2.7 V, Na-ferrocene about 1.9-2.1 V discharge and 2.4-2.6 V charge vs Na, Li-MnO2 about 3.05-3.15 V), and with none present the voltage is about 3.5 V; the paper also admits limiting voltage to avoid electrostatic energy storage.
- Coin cells built this way cycled tens to hundreds of times (Li-S 46 cycles inside a 2.50-2.67 V window; Na-ferrocene 200 cycles; Li-MnO2 170, "over 250" or 270 cycles depending on caption, abstract or figure label), with 90.1 % of the Li anode capacity (10.8x the sulfur capacity; 10.5 Wh/g of Li) delivered in one 28-day Li-S discharge, supporting a safe, low-cost, high-energy-density, long-life battery for electric vehicles. The cycling protocols are fixed-time and cannot show capacity fade; the Li-S cycles use only about 24 % of sulfur capacity per cycle if "g" means g of S, and the Li-MnO2 "3x" excess (cycle 34) is about 1.6x the one-electron theoretical MnO2 capacity.