Metastable and glassy ionic conductors MAGIC Nancy Dudney, ORNL A. - - PowerPoint PPT Presentation

Nancy Dudney, ORNL

• A. Westover, A. Kercher, S. Kalnaus, G. Veith, M. Palmer – ORNL
• V. Lacivita, G. Ceder – LBNL
• W. West – JPL
• S. Wang, W. Tenhaeff – Univ. Rochester
• J. Christensen, B. Kozinsky, M. Kornbluth, J. Mailoa, G. Bucci – Bosch

Find a “MAGIC” glass electrolyte with

• the stability of Lipon and
• higher conductivity and
• practical alternative to sputtering

Total project cost: $3.75M Length 27 mo.

Project Vision

Examine our hypothesis

Do thin-film, glassy, metastable electrolytes inhibit formation of Li filaments that limit performance of ceramic electrolytes?

Metastable and glassy ionic conductors “MAGIC”

20µm

The Concept

• Synthesize a conductive N-stabilized glassy electrolyte as nanopowders

Conventional sputtered film ~1µm Advanced manufacturing of membranes

Models and films to down select

nucleate gas quench

Support with polymer mesh Li3 P O4  LiSiPON “MAGIC” Si B Nx Sy also LLZO, LATP amorphous * Handle safely.

Plasma torch glassy nano powders

The Team

2

ORNL Make it, test it! Jet Propulsion Lab Engineer one-step to a stable Li interface! LBNL What is the role of N? Bosch (Cambridge) What compositions remain glassy from melt  quench?

• Univ. of Rochester

Polymer mesh can support thin glass. Bosch (Sunnyvale) How do dendrites form, propagate?

• 1. MAGIC electrolyte  efficient Li cycling, high rate, high capacity
• 2. Plasma torch processing  stable, conductive MAGIC powders
• 3. Mechanical properties of MAGIC  high K1C by indentation
• 4. Hypothesis test. Will glasses be best at high current?

3 MAGIC metastable and glassy ionic conductors IONICS review 2019 Nancy Dudney June 7, 2019

2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 10

• 12

10

• 11

10

• 10

10

• 9

10

• 8

10

• 7

10

• 6

10

• 5

10

• 4

Ionic Conductivity (S/cm) 1000/T (K

• 1)

Li4P2O7

Li4SiO4 Li2Si0.7P0.3O3N0.18 Li2.8Si0.7P0.3O3.2N0.22

5 full cells of 5 cm2, 0.3 mAh/cm2 1 mA/cm2, 100 cycles 30% degrade 2 MPa-m1/2, K1c 200 MPa, flex $10/m2

9th Q end 10th Q

Project Objectives

full cells thin films

10 1.0 0.1

Results (using sputtered MAGIC thin films)

MAGIC thin films by sputtering Theory of glass structure and transport

• Understand structure, and Li+

mobility

• Kinetic passivation at Li

4 Insert Presentation Name June 7, 2019

What about Li shorts?

• No shorts to 10mA/cm2
• Li dendrites form at artificial

boundary, not through

• Edges and grain boundaries

 non uniform current

• New targets,

– long life, – reproducible, – lower cost

JACS 2018 140 11029; Chem Mater, 2018 30, 7077 ACS EnergyLetters, 4 (2019) 651

Results (for Induction Plasma Torch prepared MAGIC nanopowders)

MAGIC nano-powders and membranes – limited to Si rich, need more N Theory of melt/quench of glass

Structure for 200 atoms, 3x to average Mean square displacement (D) for Li motion

• Ortho Li3PO4 crystallizes with slow quench
• Pyro and meta phosphate share corners,

form glasses

5 MAGIC metastable and glassy ionic conductors IONICS review 2019 Nancy Dudney e June 7, 2019

With 525-625°C anneal:

• Partial crystallization
• No loss of N or Li !

2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 10-12 10-11 10-10 10-9 10-8 10-7 10-6 10-5 10-4

Ionic Conductivity (S/cm) 1000/T (K-1)

Li2Si0.7P0.3O3N0.18 cold pressed Li2.8Si0.7P0.3O3.2N0.22 cold pressed Li2.8Si0.7P0.3O3.2N0.22 hot pressed

Li2Si0.7P0.3O3N0.18 sintered

800°C 350°C

LiSiPON

Challenges and Potential Technical Partnerships

• Need more control of gases through induction plasma torch. (tool design Tekna or other)
• How to consolidate glass powders so no boundaries? (glass processing Saint Gobain)
• Both sputtering and plasma torch are expensive, but will come down

nucleate gas quench

tuned for Ar

Expense ($$ being 1-10 $/m2) Sputter thin film IPT for powder Precursors / targets $$  $ $ Process – Watt : Watt $$ $$ Yield when optimized 1090% 1585% Direct film on cathode Self-support sheet ($ TBD)  Find MAGIC composition that is ductile.

• Watt for watt, both processes ≈ rate
• Material loss needs to be eliminated
• Then both viable price at scale - if

electrolyte is very thin.

T2M

• Final goals – by end of project

– Larger cells (with sputtered MAGIC) – Mechanical and thermal properties for full composition range

• Beyond – Nano-powders* promising, further pre-pilot work needed

– Modified plasma torch  complete range, LiPON to LiSiON – Larger lot of powder  Li+ conductive – Explore advanced manufacturing

• *provisional patent

7 MAGIC metastable and glassy ionic conductors IONICS review 2019 Nancy Dudney June 7, 2019

5 full cells of 5 cm2, 0.3 mAh/cm2 1 mA/cm2, 100 cycles <30% degrade 2 MPa-m1/2, K1c 200 MPa, flex $10/m2

9th Q end 10th Q

Large scale sputtering is a viable option. Powders of oxynitride glass provide alternative processing. Will a dense, sintered glass membrane effectively resist Li filaments and dendrites?

Images from ORNL brochure