<<

SOLID ELECTROLYTE BATTERIES

John B. Goodenough and Yuhao Lu

Texas Materials Institute The University of Texas at Austin

Project ID: ES060

DOE Vehicle Technologies Annual Merit Review Meeting May 9-13, 2011

This presentation does not contain any proprietary or confidential information. 1 The University of Texas at Austin Overview

Timeline Barriers

• Project Start Date-Oct. 2009 • - solid electrolyte. • Project End Date- Sept. 2010 • Choice of redox couples. • Percent complete: 100% complete • Flow-through- cell design

Budget Partners

• Funding received in FY09-FY10 – $315K • Karim Zaghib of Hydro Quebec • Funding received in FY10-FY11 – $315K

2 The University of Texas at Austin Milestones

 Develop and test a suitable test cell. (Jan. 10) Completed

 Optimize components of the cell. (Apr. 10) Completed

 Develop a lithium/Fe(NO3)3 cell. (July. 10) Completed

 Design of a new cell (Sept. 10) Completed

 Test flow-through cell (Jan. 11) Completed

3 The University of Texas at Austin Typical Lithium Ion Battery (LIB) Cell

e- e-

- +

Li+

Anode Li+ Cathode () (Oxidizing Agent)

Li+

SEI layer Electrolyte

 Capacity limited by Li solid in cathode and loss in SEI layer

limited by Eg of electrolyte.

4 The University of Texas at Austin Why Oxides?

E E

Co: 4s0 Co: 4s0

Co(II): t4e2 µLi

4.0 eV Co(IV): t5e0 EFC

EFC (pinned) Co(III): t6e0 Co(III): t6e0

O: 2p6 O: 2p6 N(E) N(E)

Li CoO (0 0.55 or H inserted from electrolyte. 2- 2-  At surface, 2(O2) =2O +O2↑ 0  Note: Li1-x(Co0.15Ni0.85)O2 (0

0.2 eV 0.5 eV 0 µ µLi LiC6 µAlloy 1.5 eV 1.1 3.45 eV LUMO µLTO 4.0 eV 4.75 eV µLFP

µLNC 4.3 HOMO µLNM 5.0 Decomposition  Li0 - e-=Li+; dentrites on recharge because of SEI layer. - +  LiC6 - e = 6C + Li ; SEI layer prevents rapid charge, consumes Li from cathode.  Alloy undergo displacement reactions, which give large volume change.  Can be buffered by C and have allowing fast charge, but SEI layer if V<1 V. +  Need µC matched to HOMO or, with Li -permeable SEI layer, to 6 decomposition voltage. The University of Texas at Austin Existing Insertion Compound

HOMO

Organic electrolyte

LUMO

. Lithium metal has the highest specific energy of 3860 mAh/g. No cathode materials can match the capacity. . Can we use lithium and develop a cathode with a capacity to match?

7 The University of Texas at Austin Existing Electrolytes

6 1) High Li+ ion conductivity

/Li 5

+ 2) Wide electrochemical window

4 3) Retention of /electrolyte interface

3 4) Chemical stability InorganicSolid

2 Organic Liquid 5) Safety (nonflammable) Potential (V) vs. Li Organic Liquid Ionic Liquid + Polymer Ionic Liquid + Polymer + Polymer + OrganicPolymer Liquid 1 Ionic Liquid Organic+ Liquid

0 0.02 0.03 0.18 0.81 1 2 4 7

+ -3 8 LiLi-ion ion Conductivity Conductivity ( ×10 S/cm) (mS/cm) The University of Texas at Austin Objectives: Li Batteries Using a Solid Electrolyte Separator

 Concept Air or S Organic Liquid Inorganic 8 Li Aqueous (Fe3+/Fe2+) flow Or nothing Solid through

 Present Practice (Visco, PolyPlus Battery Company)

1M LiPF in Li 6 Li Ti Al (PO ) Sea EC/DEC 1.3 1.7 0.3 4 3

Li3N  Proposed Practice 1M LiPF in Fe(CN) 3-/Fe(CN) 4- flow Li 6 Li Zr R O 6 6 EC/DEC 7 2 3 12 through

3- + 4- Li + Fe(CN)6 (aq) = Li + Fe(CN)6 (aq); E=3.40 V

9 The University of Texas at Austin Solid Electrolytes

• All Solid-State Cell • Solid-Electrolyte Separator (+) Would allow a lithium anode (+) All solid-state cell would Would block dendrites from a simplify and lighten packaging lithium anode or a Li/solid- electrolyte interface Would allow alternative , Sulfides offer large Eg and σLi> 3+ 2+ 10-3 S/cm e.g. air, S8, or Fe /Fe (aq) flow through (-) Insertion compounds change volume on cycling, so cycle life (-) For an aqueous cathode, need an -4 of solid/solid interfaces is oxide separator: σLi(oxide) ≈ 10 problematic S/cm fabricated thin and dense on a support; and not reduced by lithium

10 The University of Texas at Austin Approach: An Alkali-Metal/Aqueous Cathode Cell

Organic electrolyte Anode reaction: nA → nA+ + ne-; (A=Li or Na) Cathode Reaction: Mz+(aq) + ne- → M(z-n)+(aq); Overall reaction: nA + Mz+(aq) = nA+ + M(z-n)+(aq); (1≤ n < z).  No phase change takes place in the cathode.  No catalyst is needed in the cathode.  The cell is rechargeable.  The cell works at room temperature.  The cell works in the voltage range of 2.8 to 4.2 V.

11 The University of Texas at Austin Accomplishments: Selection of Aqueous Cathode with Existing Solid Electrolyte

 Aqueous electrodes will depend on electrolyte and must have: 1) proper redox potentials; 2) no side reactions; 3) good stability in water; 4) good reversibility; 5) reliable safety; 6) no toxicity and 7) low cost. 12 The University of Texas at Austin Accomplishments: A Lithium|Fe(NO3)3 (aq) Cell

5.5 7.0 First Cycle A B 5.0 6.0 5.0 4.5 Fe(OH)3 pH=4.48 4.0 3+ ∆V= 0.95 V Fe

4.0 ∆V= 1.96 V 3+ 2+ Fe2+ Fe /Fe pH 3.0 3.5 Potential (V) 2.0 FeO(OH) pH=2.0 3.0 Discharge 1.0 Charge 2.5 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 1.0 2.0 3.0 4.0 5.0 Normalized Capacity -1 Concentration (mole L )  The open circuit voltage of the battery is 3.99 V. Its initial discharge voltage is 3.74 V at a current of 0.5 mA.  After discharge Its initial charge voltage is 4.73 V at the current of 0.5 mA.  Hydrolyzation of Fe(III) results in precipitates of FeO(OH) for pH>2 and

Fe(OH)3 for pH>4.8 when pH of the aqueous electrode solution increases with cell discharge. 13 The University of Texas at Austin Accomplishments: 3- Lithium| Fe(CN)6 (aq) Cells

3.8 5.0 110 4.5 B 100 4.0 3.6 90 3.5 80 ∆ 3.0 V=0.35 V 70 3.4 2.5 60 2.0 50 3.2 0.4 40 Potential (V) 0.3 30 3-

Capacity (mAh) 0.2 3.0 3- 0.01 M Fe(CN)6 20 0.01 M Fe(CN)6 3- 3- 0.1 10 0.1 M Fe(CN) 0.1 M Fe(CN)6

6 Coulombic efficiency (%) 2.8 0.0 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0 5 10 15 20 25 Normalized capacity Cycle number  No hydrolyzation of Fe(III)  The theoretical voltage of the battery is 3.40 V.  Its effective voltage with the commercial solid electrolyte distributes between 3.33 V and 3.68 V.  Different concentrations of the redox couple in the aqueous electrode slightly affect the normalized discharge/charge curves.  The battery shows a high coulombic efficiency. 14 The University of Texas at Austin Accomplishments: 3- Power Density and Reversibility of Li|Fe(CN)6 (aq) Cells

4.0 14 0.45 120 D 12.53 mW/cm2 3.5 . 12 )

-2 0.40 100 3.0 10 0.35 80 2.5 8 2.0 0.30 60 . 6 1.5 1.77 V 0.25 40 Capacity h) (mA Cycle 3.0 V – 3.75 V

Potential (V) 4 1.0 Coulombic Efficiency (%) 0.20 20 0.5 2 Capacity Columbic Efficiency 0.0 0 densityPower cm (mW 0.15 0 0 2 4 6 8 10 12 0 200 400 600 800 1000 Current density (mA cm-2) Cycle number  The cell demonstrates a maximum power density of 12.53 mW/cm2, which is competitive with a direct methanol of similar active area.  Due to no cathode phase change in the charge/discharge, the cell shows a good reversibility. No capacity loss was observed over 1000 cycles.

15 The University of Texas at Austin Accomplishments: 3- Stability of Li|Fe(CN)6 (aq) Cells with Existing Electrolyte -2 Current density (mA cm ) -600 5.0 0.1 0.2 0.5 1.0 1.5 2.0 2.5 3.0 8 A Fresh cell C -500 4.5 7 ) -2 4.0 Rint → RSE after 6 3.5 -400 aging uncycled for 10 h )

5 2 3.0 2.8 V -300 cm

2.5 4 Ω

2.0 ( Z'' 3 -200 1.5 354.89 Hz

Potential (V) 2 1.0 -100 19.95 KHz Potential 1 MHz 2.51 Hz 0.5 1 Power density

0.0 0 density Power cm (mW 0 0 1 2 3 4 5 6 7 8 0 RS 100RF 200 300RSE 400RInt 500 600 2 Time (Hours) Z' (Ω cm ) -800 C After deep discharge A C -700

-600 RS (solution) 42.46 43.79 -500 ) 2 (charge transfer) -400 RF 120.52 115.62 cm

Z'' ( Z'' -300 (solid electrolyte) 446.68 Hz RSE 160.24 237.84 -200 22.39 KHz 0.01 Hz 1 MHz -100 RInt(interface) 143.34 346.33

0 0 R 100R 200 300 400R 500 600 700R 800 s F SE2 Int 16 Z' (Ω cm ) The University of Texas at Austin Accomplishments: Rechargeable Lithium|Cathode- (RLCFB)

 The aqueous cathode in a flow-through mode can be individually stored in a “fuel” tank, which reduces the volume of the battery and increases the design flexibility of the battery structure.  Recycling of aqueous cathode can match the large capacity of lithium metal. 17 The University of Texas at Austin Accomplishments: Behavior of RLCFB

3.65 4.0 18 3.60 3.5 16 3.55 3.0

14 ) 2 3.50 2.5 12 3.45 10 0.41 ml/min 2.0 3.40 0.29 ml/min 8

Voltage (V) 0.16 ml/min 1.5 Voltage (V) 3.35 0 ml/min 6 3.30 1.0 0.41 ml/min 4 densityPower (MW/cm 0.29 ml/min 3.25 0.5 0.16 ml/min 2 0 ml/min 3.20 0.0 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 2 4 6 8 10 12 14 16 Time (Hour) Current density (mA/cm2)  The maximum power density of the battery with existing electrolyte is about 17 mW/cm2.  Flow rates of the aqueous cathode don’t influence significantly the power density of the battery.  Development of a thinner solid-electrolyte separator with a higher Li+-ion conductivity would provide commercially viable room-temperature EES for energy sources alternative to fossil fuels.

18 The University of Texas at Austin Summary

 The study demonstrates the feasibility of a Lithium|water battery utilizing lithium metal as anode and a redox couple soluble in as the cathode.  This new strategy represents a third-generation lithium-ion battery promising lower cost than the conventional lithium-ion , safe operation, and having a columbic efficiency and voltage greater than that of a Li/air battery with, in principle a comparable capacity.  A flow-through mode of the aqueous cathode can provide a full and efficient match of the capacity of a lithium-metal anode. The flow of the aqueous solution continuously brings heat out of the battery system and keeps the battery working near a mild condition.  Flow rates of the aqueous cathode in the study don’t influence significantly the power density of the battery.  The demonstrated power output is limited by the commercially available solid electrolyte separating an organic-liquid or polymer anolyte and the aqueous cathode; the fabrication of a superior solid electrolyte is clearly needed.  The new strategy promises to be applicable to both the electric-vehicle market and the problem of electrical for the grid.

19 The University of Texas at Austin Future Work

 By stabilizing the existing electrolyte against degradation, which we have shown to be due to the presence of Ti(IV), we will obtain data on the specific capacity versus molar fraction of the Fe(III)/Fe(II) couple in the cathode solution and how capacity is retained at higher discharge/charge rates.  We will obtain data on the ability of seals to prevent water crossover into the anode compartment and whether a lithium/solid-electrolyte solid-solid interface can be maintained over a long cycle life.  We will explore different cathode current-collector configurations.  Alternative solid electrolytes will be identified and their fabrication into dense, durable films will be explored.  The use of a Na rather than a Li anode with a Na+-ion solid electrolyte will be investigated.

20 The University of Texas at Austin