Goodenough-JB · Source catalogue

A 7-page PDF (ResearchGate cover page plus the 6-page journal article). The authors describe an all-solid-state lithium or sodium 'metal-plating' cell. The electrolyte is a glass of nominal composition A2.99Ba0.005O(1+x)Cl(1-2x) (A = Li or Na), made as a slurry in ethanol, soaked into fiberglass or recycled paper and heated above 130 C. It is claimed to have cation conductivity above 1e-2 S/cm at 25 C, motional enthalpy about 0.06 eV, an energy-gap window above 9 eV, and a surface wetted by metallic Li/Na, so that no SEI forms and no dendrites grow. The glass is said to contain electric dipoles that, after about 10 days of room-temperature ageing, coalesce into ferroelectric 'molecular condensates' and give a very large dielectric constant and large electric-double-layer capacitances (EDLCs) at both electrode/electrolyte interfaces. On discharge, lithium (or sodium) from the anode is claimed to be plated onto a copper/carbon cathode current collector instead of being inserted into a host cathode, so capacity is set by the amount of alkali metal in the anode. The cell voltage is said to be set by a 'redox center' (S8, ferrocene or MnO2 particles) in the cathode that is not itself reduced (for sulfur, as long as V stays above 2.34 V). Without a redox center the voltage is said to be about 3.5 V. The open-circuit and discharge voltages are rationalised with two EDLCs in series: Vcell = Q/Ceq, Ceq = CA*CC/(CA+CC), with Q treated as constant, so voltage rises as Ceq falls. Evidence shown: (Fig. 1) a Li-S coin cell with 1.99 mg sulfur (S:glass:C = 47:43:10 wt%) discharged over 28 days gave a capacity far above the sulfur capacity, equal to 90.1% of the lithium anode capacity, with energy density 10.5 Wh per g of Li (8.5 Wh/g for V > 2.34 V). (Fig. 2) After full discharge, photos and SEM back-scatter images of the disassembled cell are said to show no lithium left at the anode and lithium on the cathode side; the deep read found no EDS spectrum or element map in the figure at all, only dark crusty deposits and by-label zone assignments. (Fig. 3) A Li-S cell cycled 46 times (10 h each way at 40 mA/g, with a seal that was admitted to be imperfect). (Fig. 5) Open-circuit voltage versus time after assembly, rising then falling, typical initial VOC about 1.65 V. (Fig. 6) A Na-ferrocene cell cycled 200 times at 40 mA per g of ferrocene: discharge about 1.9-2.1 V and charge about 2.4-2.6 V vs Na, half-cycles of about 2 h (about 80 mAh/g, 56 % of ferrocene's one-electron capacity), charge-discharge gap growing from about 0.2 V to 0.7 V by cycle 200. (Fig. 7) A Li-MnO2 cell, voltage band 3.0-3.2 V (mean about 3.05-3.15 V) at 25 mA/g; cycle labels on record are 170 (caption), 'over 250' (abstract) and 270 (figure label); the 34th (20 h) cycle gave about 3x the 'measured' MnO2 capacity but only about 1.6x the one-electron theoretical capacity (308 mAh/g), and the routine 10 h cycles lie within it. (Fig. 8) LEDs lit by a Li-MnO2 coin cell (V > 3.0 V) and a Li-S cell (VOC 2.61 V, 2.55 V under a blue LED). (Fig. 9) Chronopotentiometry on a symmetric Cu//Li/glass/Li//Cu cell over 200 h, annotated I = +-3 mA/cm2 and 'sigma_optimized = 5-10 mS/cm at 25 C'; digitised: 1 h half-cycles (3 mAh/cm2), about 79 cycles over about 160 h, overpotential about 5-6 mV (ASR about 2 ohm cm2), after an early +75 mV spike; the later drop to a small flat voltage fits either better wetting or a soft short. Tests used a Fluke voltmeter and LANHE or Arbin cyclers. Thickness of the electrolyte membrane was 0.06 mm. Authors state that membrane thickness, cathode composition and interfaces were not optimised, and no rate performance is quantified. The Broader-context abstract claims a 3.08 V discharge cycled over 250 times with no loss of energy density. Cell-level energy density, power, temperature range and mass or volume of the whole cell are not reported. Cited background works (refs 1-3, Braga et al. 2014 J. Mater. Chem. A; Braga et al. 2016 Energy Environ. Sci. 9, 948; Braga et al. 2017 J. Electrochem. Soc. 164, A1) hold the glass-electrolyte and dipole claims; these were not available and were not read. Deep-read additions (pp.4-7 figures). Fig. 1: voltage starts near 1.65 V open-circuit and jumps to 2.68 V under discharge (about 1 V above the prior VOC, unexplained by ordinary electrochemistry). Fig. 3: Li-S cycling window only 2.50-2.67 V (never near 2.34 V); 400 mAh/g per cycle, about 24 % of sulfur capacity if g means g of S, so the 46 cycles need no excess capacity; fixed-time cycling cannot show capacity fade. Fig. 4: paper gives mu_A = -3.05 V, mu_C = -0.40 V vs SHE (E_F = -mu plotted positive; earlier files dropped the signs); energy gap shown as 'E_g > 8 V' versus text '> 9 eV'; the capacitor cartoon's own numbers do not satisfy its 'C_C << C_A' condition; mu_C has no stated measurement method and 2.65 V is partly circular. Fig. 5: VOC fits V = V0 + (Vmax-V0)(1-exp(-t/tau_Ci)), tau_Ci = 94 h, rising to 2.65 V at 240 h, then V = (Vmax/2)(1+exp(-t/tau_Cf)), tau_Cf = 346 h (assumed halving to 1.325 V), reaching 1.65 V at about 726 h (about 30 days); this resolves the 1.65 V versus 2.61 V puzzle (different ages). Fig. 8: meter readings 3.037 V (Li-MnO2), 2.607 V and 2.552 V (Li-S without and with blue LED; LED current not given). Text p.4 admits a voltage limit 'to avoid electrostatic energy storage at a high voltage' (ref. 2). Electrolyte is cast from ethanol into paper or fiberglass and heated only above 130 C, so residual ethanol, water and cellulose are unaddressed candidate reactants for extra charge. The Fig. 6 reading in QE item 3 (minute-long cycles) was corrected by the deep read: the axis is time per cycle, half-cycles are about 2 h.

Key claims

• A glass electrolyte A2.99Ba0.005O(1+x)Cl(1-2x) (A = Li or Na) has cation conductivity above 1e-2 S/cm at 25 C with motional enthalpy about 0.06 eV.
• The glass is wetted by metallic lithium or sodium, forms no SEI, and permits dendrite-free plating and stripping of the alkali-metal anode.
• The glass has an energy-gap window above 9 eV (Fig. 4 says 'E_g > 8 V'), so it is stable in contact with both an alkali-metal anode and a high-voltage cathode.
• After about 10 days of ageing at room temperature the glass's electric dipoles form ferroelectric molecular condensates that give a very large dielectric constant and large electric-double-layer capacitances at both interfaces.
• During discharge, lithium or sodium from the anode is plated onto a Cu/carbon cathode current collector, so cell capacity is set by the amount of anode metal rather than by a host cathode.
• A Li-S cell with 1.99 mg sulfur delivered a discharge capacity of 90.1% of the lithium anode capacity, far above the sulfur capacity, with the sulfur not reduced above 2.34 V.
• A cathode redox center (S8, ferrocene, MnO2) sets the plating-cell voltage (roughly 2.5-2.7 V for S8, about 2 V vs Na for ferrocene, about 3.05-3.1 V for MnO2) without being consumed, and with no redox center the voltage is about 3.5 V.
• Cell voltage equals Q/Ceq with Ceq the series combination of the two interfacial double-layer capacitances and Q constant, so open-circuit and discharge voltage changes follow capacitance changes.
• Such cells cycle reversibly: Na-ferrocene 200 cycles (discharge about 2 V vs Na), Li-MnO2 170, over 250 or 270 cycles (caption, abstract, figure label) at about 3.05-3.1 V, Li-S 46 cycles, with no loss of energy density.
• The resulting cell is safe, low-cost, and has high volumetric energy density and long cycle life suitable for an all-electric vehicle.
• The symmetric Cu//Li/Li-glass/Li//Cu cell ran at +-3 mA/cm2 for about 200 h with overpotentials of a few mV, showing negligible interfacial impedance (Fig. 9).
• The Li-S cell open-circuit voltage rises to about 2.65 V in 10 days then decays toward V_max/2 with tau = 346 h, reaching 1.65 V after about 30 days (Fig. 5), and this is attributed to changing double-layer capacitance.

Relevance: This is the only document in the folder and is the central statement of the Goodenough-Braga 'glass electrolyte / plating cathode' battery concept, the project's subject. It combines three separable ideas: a super-ionic glass electrolyte, dendrite-free alkali-metal plating, and a capacitor-based explanation of cell voltage. The last is a departure from conventional electrochemistry.

Mainstream conflict: Substantial, per the article's own account. (1) Conventional electrochemistry sets cell voltage by the difference in electrochemical potential of the two electrodes (Li metal against whatever is reduced or oxidised at the cathode), not by an EDLC series formula with constant Q; the paper uses a capacitor voltage Q/Ceq with Q 'constant' without a demonstrated charge balance. (2) A Li||Cu cell that plates Li on copper normally sits near 0 V vs Li/Li+ at the plating step, so a stated discharge near 2.5-3.5 V with Li plating on the cathode side, and a sulfur or MnO2 'catalytic relay' that is not consumed, requires a mechanism not given in standard theory. (3) Ionic conductivity above 1e-2 S/cm with an 0.06 eV barrier, a gap above 9 eV, and 'ferroelectric molecular condensates' with huge permittivity in an amorphous glass are unusual relative to known solid electrolytes. (4) The capacity-versus-sulfur argument (capacity beyond what the sulfur can supply) could equally be explained by electrolyte or other side reactions, which the paper does not exclude; no independent replication is cited. [memory, medium] Other groups publicly questioned the claims after publication and in later years, including on thermodynamic grounds, but this has not been checked on the web at this step. Additional points from the figure reads: a Fermi-level (work-function) difference between two metals is a contact potential that equalises on connection and cannot drive sustained current; discharge voltage rising about 1 V above the preceding open-circuit voltage is the opposite of ordinary IR loss; EDS cannot detect lithium [memory, high], so the Fig. 2 verification does not verify Li.