Embed Size (px)
Citation preview
This presentation does not contain any proprietary, confidential, or otherwise restricted information
Design and Synthesis of AdvancedHigh-Energy Cathode Materials
Guoying ChenLawrence Berkeley National Laboratory
June 11, 2015
Project ID: ES225
Overview
Timeline• Start date: October, 2012
• End date: September, 2016
• Percent complete: 70%
Budget• Total project funding
- FY2013 $500K- FY2014 $500K
Partners• Collaborations: LBNL, UCB,
Cambridge, ORNL, PNNL, NCEM, ALS, SSRL
• Project lead: Venkat Srinivasan
Barriers Addressed• Energy density
• Cycle life
• Safety
Objectives – Relevance
• Obtain fundamental understandings on phase transition mechanisms, kinetic barriers, and instabilities in high-energy cathode materials.
• Control cathode-electrolyte interfacial chemistry at high operating voltages and minimize solid-state transport limitations through particle engineering.
• Develop next-generation electrode materials based on rational design as opposed to the conventional empirical approaches.
Milestones
December 2014Characterize Ni/Mn spinel solid solutions and determine the impact of phase transformation and phase boundary on rate capability (Completed)
March 2015
Complete the investigation on crystal-plane specific reactivity between Li-rich layered oxides and the electrolyte. Determine morphology effect in side reactions (Completed)
June 2015
Develop new techniques to characterize reactions and processes at the cathode-electrolyte interface. Evaluate the effect of surface compositions and modifications on side reactions and interface stability (On schedule)
September 2015
Go/No-Go: Continue the approach of using synthesis conditions to vary surface composition if significant structural and performance differences are observed (On schedule)
Approach
10 µm
1 µm 1 µm
1 µm
1) Remove the complexity in high-energy cathode materials – synthesize crystal model systems with defined attributes for the investigation of solid state chemistry, kinetic barriers and instabilities.
3) Design and synthesize optimized electrode materials based on the structural and mechanistic understandings.
2) Perform advanced diagnostics for insights – ex situ and in situ studies to characterize crystal-plan specific transport properties and interfacial chemistry. Establish direct correlations between physical properties, performance, and stability.
Li OMnNi
3D compositional mapping (APT/PNNL)
10 nm
Phase distribution (TXM/SSRL) and atomic imaging (HRTEM/NCEM)
1 µm
Technical accomplishments: overview1) Determined structural make up of the entire Li and Mn rich NMC (LMR-NMC)
crystal2) Revealed the contribution of key surface properties to the material challenges
facing LMR-NMC, including:• First cycle irreversibility and rate capability (kinetics)• TM reduction and dissolution during cycling• Capacity retention and side reactions with the electrolyte• DC resistance increase• Voltage fade
3) Investigated the kinetic implication of solid-solution vs. biphasic reaction pathways in intercalation cathode materials • Synthesized and characterized room-temperature LixMn1.5Ni0.5O4 solid
solution phases for the first time• Evaluated the role of phase boundaries and phase transformation in the
kinetics of materials with first-order transition4) Diagnostic techniques developed for the use of single-particle based investigation
relevant to cathode performance and stability.
This presentation mainly focuses on 1) and 2).
Synthesis of LMR-NMC crystal samples• Molten-salt method: high-temperature solution based synthesis
promotes crystal nucleation and growth in the flux.
a b c
d
50 nm50 nm
e f
1 µm 1 µm 1 µm
1 µm 1 µm 1 µm
Plate Needle L-Poly
S-Poly (∼ 10x more SA) Box Octahedron
• Morphology of LMR-NMC (Li1.2Ni0.13Mn0.54Co0.13O2) crystals varied by adjusting reaction precursors, flux, heating temperature and time.
Our crystals are monoclinic single phase
Needle
10 nm
[100] [110][1-10]
• Domains of three monoclinic variants in random distribution in the entire crystal.• Not a composite with R-3m and C2/m nano-domains.• Crystal structure independent of morphology.
HAADF STEM imaging (Alpesh Shukla, LBNL)
Atomic resolution EELS mapping (Alpesh Shukla, LBNL)
Pristine oxide has reduced TM on surface
HAADF O Mn Co Ni Mn,Co,Ni
• Bulk Mn, Co and Ni at 4+, 3+ and 2+, respectively. • Mn and Co reduced but Ni remains at 2+ on the surface layer (about 2 nm thick).
• STEM imaging in multiple zone axes is essential for determine the structures.• Bulk has monoclinic structure while surface has spinel structure with reduced TM.• Spinel formation on pristine surface is directional/morphology dependent –
minimal spinel formation in the TM layer stacking direction.
Reduced surface TM in spinel structure
[111]s and [103]M [110]s and [010]M [121]s and [001]M
Multiple zone axes HAADF STEM imaging (Alpesh Shukla, LBNL)
PrisCne surface TM reducCon morphology dependent
Fluorescence Yield (FY, 50 nm depth)
Auger Electron Yield (AEY, 1–2 nm depth)
Total Electron Yield (TEY, 2–5 nm depth)
Sol XAS L edge spectra collected on composite electrodes with
carbon and binder (SSRL beamline 10-‐1)
Ni Mn Co
L2 L3 L2 L3
• Depth-‐resolved XAS confirms reduced Mn and Co on the top surface (a few nm) while Ni remains at 2+ in en,re par,cle.
• Effects of surface facet and surface area – TM reduc,on least on Plate and L-‐Poly samples with predominantly TM layer surface while most on S-‐Poly and Box samples.
Mn3+/Mn2+ Co2+
Surface TM reducCon increases with cycling
• Mn XAS spectra show lower valent Mn content increases with cycling (2.5-‐4.6 V).
• Cycling-‐induced surface Mn reduc,on occurs on all samples but most extensive on the Box sample.
Cycling-‐induced TM reducCon surface dependent 45 cycles
Pris,ne
• Cycling-‐induced TM reduc,on progresses from the surface to the bulk. • Effects of morphology and ini,al spinel content – most TM reduc,on during cycling
occurred on Box while least on L-‐Poly sample.
• Transmission mode imaging on LMR-‐NMC crystals at a spa,al resolu,on of about 20 nm (single pixel).
• Mn and Ni are 4+ and 2+, consistent with the measurement on the bulk sample.
• No varia,on in oxida,on state from the center to the edge of the plate crystal, consistent with minimum TM reduc,on on the pris,ne plates.
Chemical distribuCon of TM at parCcle level – prisCne
1 2 3 4
0.4 µm
STXM, BL 11.0.2 (with T. Tyliszczak, ALS)
• Some Mn and Co reduc,on on par,cle surface but significantly less than the large amount detected by XAS on the electrodes.
• Are the reduced Mn and Co observed on cycled electrodes structural to the crystal or surface deposits resul,ng from the TM dissolu,on/migra,on/precipita,on process?
Chemical distribuCon of TM at parCcle level – cycled
0.2 µm
1 2 3
STXM
Sol XAS
Cycling-‐induced TM dissoluCon/migraCon/precipitaCon Separators recovered from the cycled half cells:
• The side facing lithium covered with black deposits which increase with cycling.
• Surface proper,es of pris,ne oxide, both morphology and ini,al spinel content, affect TM dissolu,on.
Mn Co Ni
• Sol XAS (L edge, TEY mode) shows presence of Mn, Co and Ni on the separator, all at 2+ oxida,on state.
a b c d
• Both size and surface facet have major impact on structural evolu6on and first-‐cycle ac6va6on kine6cs during chemical delithia6on with NO2BF4/CH3CN.
• Size cri6cal for kine6cs – best performance on S-‐Poly sample with the smallest par6cles.
Surface proper+es impact first-‐cycle ac+va+on kine+cs
• Effect of particle size – best activation kinetics observed on S-Poly sample. • Effect of surface facet – among large crystals, Plate easiest while Box most difficult
to activate upon electrochemical charging.• Diffusion coefficient nearly two orders magnitude smaller at the activation plateau.
3.8 4.0 4.2 4.4 4.6 4.8
1E-16
1E-15
1E-14
Plate Needle L-Polyhedron S-Polyhedron
D (c
m2 s
-1)
Voltage (V)
Surface properties impact first-cycle activation kinetics
0 50 100 150 200 250 300 3502.0
2.5
3.0
3.5
4.0
4.5
5.0
Volta
ge (V
)
Capacity (mAh/g)
Plate Needle L-Poly S-Poly Box Octahedron
Surface impact rate capability and capacity retention
• Effect of particle size – best rate capability and most capacity from the S-Poly sample.
• Effect of surface facet – Plate best while Box worst among large-sized samples.
5 10 15 20 25 300
50
100
150
200
250
Plates Needles Large polyhedrons Small polyhedrons BOXOct
Cap
acity
(mA
h/g)
Cycle number
10 20
40 80
150 200mA/g
0 15 30 45 60 7550
100
150
200
250
Cap
acity
(mA
h/g )
Cycle number
2.5-4.6 V, 10 mA/g
0 15 30 45 60 75
60
80
100
120
Cap
acity
rent
entio
n (%
)
Cycle number
• Worst capacity retention on Box sample with the most TM reduction during cycling.
Surface impact on DC resistance rise
20 40 60 80 1000
200
400
600
800
1000
Res
ista
nce
(Ωcm
2 )
SOC ( %)
Plate Needle L-Poly S-Poly Octahedron
Cycle 520 40 60 80 100
0
200
400
600
800
1000
Cycle 15
SOC (%)20 40 60 80 100
0
200
400
600
800
1000
Cycle 25
SOC (%)
• DC resistance rise at low SOC (<50%) reduces usable energy of the material. • DC resistance rise occurs on all our samples. • Effect of particle size – much improved on S-Poly sample with smaller particles.
Surface impact on voltage fade
• Voltage fade occurs on all our samples. • Voltage fade is associated with TM reduction during the cycling – Box sample
has the most voltage fade while L-Poly has the least.
0 20 40 60 80 1003.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
Volta
ge (V
)
SOC (%)
Cyc 5 Cyc 15 Cyc 25 Cyc 35 Cyc 45 Cyc 55 Cyc 65 Cyc 75 Cyc 85 Cyc 95 Cyc 105 Cyc 115 Cyc 125 Cyc 135 Cyc 145
Voltage fade on plate sample
0 20 40 60 80 1003.0
3.2
3.4
3.6
3.8
4.0
4.2
4.4
Cycle 25
SOC (%)
Plate Needle L-Poly S-Poly Box Octahedron Commercial
2.5-4.6 V, C/5
Phase boundaries and kinetic behavior in materials with first-order phase transition
• Conventional wisdom says the access to solid-solution reaction pathways increases rate capability and cyclability.
• Is there correlation between kinetics and the extent of solid solution transformation in two-phase systems? Kinetics as a function of solid solution transformation?
Ex situ
In situ
48 49 50
(331)
58 59 60 61
(333)
67 68 69 70
(531)
Li0.01Li0.10Li0.20Li0.29Li0.39Li0.49Li0.59Li0.69Li0.78
Li0.95Li0.84Li0.74Li0.63Li0.53Li0.42Li0.32Li0.21Li0.10
• But solid solutions are metastable and difficult to isolate for their physical and kinetic properties evaluation.
Synthesis/isolation and characterization of solid solutions
• The cubic phases merge into a single solid-solution phase at elevated temperatures which remains phase pure upon cooling to RT.
• Thermal behavior is Li content dependent.
TXM, BL 6.0.3 (SSRL) with Jordi Cabana (UIC)
2 µm
• Particle level distribution of phases monitored. • Physical properties of solid solution similar to the pristine.
Collaborations • Drs. Marca Doeff and Phil Ross (LBNL), Drs. Ethan Crumlin and Tolek Tyliszczak (ALS),
Drs. Apurva Mehta and Yijin Liu (SSRL) – synchrotron in situ and ex situ XRD, XAS, XPS, STXM and TXM studies
• Dr. Robert Kostecki (LBNL) – Raman and FTIR characterization of electrode materials
• Prof. Clare Grey (Cambridge) – NMR studies
• Prof. Bryan McCloskey (UC Berkeley) – gassing under high-voltage operation of cathodes
• Prof. Jordi Cabana (UIC) – synchrotron TXM studies
• Prof. Shirley Meng (UCSD) – synchrotron coherent X-ray diffractive imaging (APS)
• Dr. Ashfia Huq (ORNL) – neutron diffraction
• Dr. Chongmin Wang (PNNL) – TEM
• Dr. Arun Devaraj (PNNL) – atom probe tomography
• Dr. Jagit Nanda (ORNL) – new cathode material synthesis and characterization
Future Work• Further investigate the impact of synthesis, particle morphology,
native and artificial surface modification on the rate performance, cycling and thermal stabilities of Li-TM-oxides.
• Determine Li concentration and cycling dependent transition-metal movement in (through structural rearrangement process) and out of (through TM dissolution process) the oxide particles and examine the mechanisms.
• Investigate interfacial chemistry between high voltage cathode and electrolyte. Determining the dynamic structural and chemical changes at the interface.
• Identify key surface properties and features hindering stable cycling of Li-TM-oxides at high voltages.
• Explore other aspects of particle engineering to improve cathode performance and stability.
Summary• Thin layer of defective spinel with reduced TM exists on the surface of pristine
LMR-NMC. The amount of spinel formation is largely controlled by particle morphology (surface facet and surface area).
• Both morphology and surface spinel on pristine impact kinetics and stability, particularly:o Structural stability – pristine surface TM reduction and spinel formation,
cycling-induced TM reduction, DC resistance changes and voltage fadeo Kinetics – chemical delithiation, first cycle activation and rate capability o Reactivity – capacity retention, coulombic efficiency and TM dissolution
• TM reduction increases with cycling which progresses from the surface to bulk.• Not all reduced TM is structural. TM dissolution/precipitation largely
contributes to the reduced TM on cycled electrodes detected by surface sensitive XAS.
• RT LxMNO solid solution phases were synthesized and characterized. The roles of phase boundaries and phase transformation in the kinetics of materials with first-order transition investigated.
Technical Back-Up Slides
Standard soft XAS spectra for transition metals
Journal of Electron Spectroscopy and Related Phenomena 190, 64–74 (2013) Physical Review B 84, 014436 (2011)
Journal of Electron Spectroscopy and Related Phenomena 114–116,
855–863 (2001)
Surface Co reduction increases with cycling
• Co L-edge XAS spectra show the formation of Co2+ increases with cycling.
770 780 790 800 810
16 cycle
45 cycle
30 cycle
Inte
nsity
(a.u
.)
AEY TEY FY
Pristine
770 780 790 800 810
770 780 790 800 810
770 780 790 800 810Energy (eV)
770 780 790 800 810Energy (eV)
770 780 790 800 810
Energy (eV)
Inte
nsity
(a.u
.)Plate Needle L-Poly
S-Poly OctahedronBox
Property-controlled crystal synthesis• We utilize high-temperature and low-temperature solution based
synthesis techniques, including solvothermal and molten-salt synthesis, to prepare high-quality crystal samples.
1 µm
LiCl-KCl flux Phase Diagram
mole LiCl/(KCl+LiCl)
Tem
pera
ture
(oC
)
ASalt-liquidASalt-liquid + KCl(s)
ASalt-liquid + LiCl(s)353o
771o
610o
0.592
1 µm1 µm
Effect of flux on LiMn1.5Ni0.5O4 morphology
Thermal-driven LixMNO solid solution formation
• Formation of LixMn1.5Ni0.5O4 (LixMNO) solid solutions initiated around 150 oC and completed around 250-265oC.
• Cooling of solid solution phases follow thermal expansion behavior with no phase separation.
• Thermal behavior is Li content dependent.
On cooling
50 100 150 200 2508.16
8.17
8.18
8.19
8.20
Latti
ce p
aram
eter
(Å)
Temperature (C°)
Li1 Li0.90
Li0.82 Li0.71 Li0.51
50 100 150 200 2508.008.028.048.068.088.108.128.148.168.188.20
Latti
ce p
aram
eter
(Å)
Temperature (C°)
Li0.06Li 0
Li0.25Li0.11
Li0.40
50 100 150 200 2508.008.028.048.068.088.108.128.148.168.188.20
Li0.71Li0.82
Li0.90
Latti
ce p
aram
eter
(Å)
Temperature (C°)
Li1
Li0.51
LixMNO phase diagram on heating
