Understanding Structural Changes in LMR-NMC Materials Project ID: ES194

Jason R. Croy Voltage Fade Team Annual Merit Review Washington DC, June 16-20, 2014

This presentation does not contain any proprietary, confidential, or otherwise restricted information.

2

Overview

Timeline • Start: October 1, 2012 • End: Sept. 30, 2014 • Percent complete: 75% Budget • Voltage Fade project

Barriers Development of a PHEV and EV batteries that meet or exceed DOE/USABC goals. Partners • ORNL • NREL • ARL • JPL

Relevance • Lithium- and manganese-rich (LMR-NMC) composite cathodes offer considerable gains over current state-of-the-art chemistries. • Voltage fade and hysteresis represent significant challenges to the commercialization of these oxides. • An atomic level understanding of the mechanisms driving voltage fade and hysteresis is necessary for the design of novel, robust LMR-NMC cathodes.

Material Voltage (vs. Li/Li+)

Capacity (mAh/g)

Sp. En. (Wh/Kg)

LiCoO2 3.8 150 570 LiNi1/3Mn1/3Co1/3O2 3.7 170 629 LiNi0.8Co0.15Al0.05O2 3.7 185 685 LiMn2O4 4.0 110 440 LiFePO4 3.4 160 544

Parameters of currently available Li-ion cathodes

Composite cathode energy densities

3

Li2MnO3 LiMO2

Li

MO6

ΔV

xLi2MnO3•(1-x)LiMn0.5Ni0.5O2

Approach

• Take advantage of DOE national user facilities to gain insights into the factors affecting voltage fade and hysteresis.

• Develop an atomic-level model that captures the essential electrochemical observations associated with voltage fade and hysteresis. • Provide experimental data to the theory component of the voltage fade team to further evaluate the model. • Provide feedback to the synthesis component of the voltage fade team.

• Design and carry out experiments to validate and refine our understanding of voltage fade and hysteresis. • Use the understanding/information gained to aid the design of more robust cathode structures.

4

Progress - What We Know About LMR-NMC

• Local structure is driven by charge ordering giving regions of two types: TM-rich (LiMO2) – High M-M coordination giving “standard” electrochemical behavior. Mn-rich (Li2MnO3) – Li/Mn rich regions which show strong tendencies for Li-Mn ordering. Low Mn-M coordination, electrochemically different than bulk Li2MnO3.

• Activation of the “Li2MnO3 component” (LiMn6-type ordered regions) is necessary to induce voltage fade – concomitant with a structural hysteresis. • Voltage fade and hysteresis increase with increasing Li and Mn ordering

increasing x in xLi2MnO3•(1-x)LiMO2.

• Mn-rich regions undergo more severe structural changes relative to TM-rich regions Li/Mn ordering plays a key role in VF and hysteresis. • Hysteresis and voltage fade are correlated and depend on Li utilization, voltage, rate, and temperature (cycling is worse than high voltage ageing). • A model has been developed to help us understand these observations.

5

Li

Mn O

0 1 2 3 4 5 6 7 8

(d) Ni K

FT m

agni

tude

(a.u

.)

R (Å)

x = 0.5

x = 0.1x = 0.3

x = 0

Ni-O

Ni-M

Ni-O-Ni

0 1 2 3 4 5 6 7 8

FT m

agni

tude

(a.u

.)(c) Mn K

R (Å)

Li2MnO3

x = 0.5

x = 0.1x = 0.3

x = 0

Mn-M

Mn-O

Charge Ordering Dictates Composite Nature

• Mn-M coordination decreases with increasing x. Local environment tends towards that of pure Li2MnO3

(no peak at ~4 Å).

• Changes to local nickel environment are relatively small (peaks at ~4 Å).

• Charge ordering is a dominating factor dictating local structure. Most Li/Ni exchange associated with MnNi-rich component (Ni-O-Ni).

Croy et al., JES, 161 A318 (2014) 6

xLi2MnO3•(1-x)LiMn0.5Ni0.5O2

Synthesis and Control Over Local Ordering

0 1 2 3 4 5 6 7 8

FT m

agni

tude

(a.u

.)

R (Å)

QuenchedControlled(c)

Mn K

Quenched Slow cooled

0 1 2 3 4 5 6 7 8

R (Å)

QuenchedControlled

Co K

(d) Quenched Slow cooled

1500 1000 500 0 -500Frequency (ppm)

ControlledQuenched

6Li NMR Quenched Slow cooled

LiTM

LiLi

6540 6550 6560 6570

Nor

mal

ized

Inte

nsity

(a.u

.)

QuenchedControlled

(a) Mn K

Quenched Slow cooled

850°C for 12 hour followed by: Quenched – LN2 cooled plates, Slow cooled – 16 hours to RT

Mn-M Co-M

local “composite” structure persists regardless of cooling rates

Long et al. submitted

Quenched 4 6 8 10 12 14 16 18

Inte

nsity

(a.u

.)

2-Theta

QuenchedSlow Cooled

(003

)

(101

)

(104

) (0

15)

(107

) (1

10)

(113

)

R3m

Energy (eV) 7

Li1.2Mn0.4Co0.4O2

6540 6550 6560 6570

Li1.2Mn0.4Co0.4O2Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

Li1.2Mn0.6Ni0.2O2

(a) Mn K

0 1 2 3 4 5 6 7 8

FT m

agni

tude

(a.u

.)

R (Å)

(b)

Mn K

Li1.2Mn0.4Co0.4O2Li1.2Mn0.6Ni0.2O2

Mn – M Mn – O

6540 6550 6560 6570Energy (eV)

QuenchedLi2MnO3

• Variations between compositions due to Mn4+/Ni2+

interactions and Li+/Ni2+ exchange

• Li and Mn ordering dominates , Mn-M CN ~4 in both compositions regardless of cooling rates. Related to the low-temp formation of Li2MnO3.

• Both samples quenched from 850°C

Croy et al. unpublished

4.8 4.9 5.0 5.1 5.2 16.2 16.4 16.6 16.8 17.0

Norm

alize

d In

tens

ity

2θ, degree

450 oC 550 oC 650 oC 750 oC 850 oC

003

108 110

(003)

(108) (110)

~400-450°

C

Y. Kan et al. HRXRD

Synthesis and Control Over Local Ordering

8

6540 6550 6560 6570 6580

Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

Mn K

Li2MnO3

(a)

0 1 2 3 4 5 6 7 8

(b)

FT m

ag. (

a.u.

)

R (Å)

Mn K

Li2MnO3

5.1 V100 mAh/g200 mAh/g2.0 V

Mn-O

Mn-Mn

0 1 2 3 4 5 6 7 8

(b)

FT m

ag. (

a.u.

)

R (Å)

Mn K

Li2MnO3

5.1 V2.0 V

• Mn3+ and extreme damping of EXAFS JT distortions.

• Single O-bond gives Mn-O coordination ~3 (1.9 Å).

• JT distortion gives coordination of ~6 at (1.9/2.3 Å).

• Likely two phases present (Li2MnO3, ??).

Croy et al. unpublished

6540 6550 6560 6570 6580

5.1 V ~300 mAh/g

Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

Mn K

Li2MnO3

(a)

6540 6550 6560 6570 6580

5.1 V

200 mAh/g100 mAh/g

2.0 V

Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

Mn K

Li2MnO3

(a)

6540 6550 6560 6570 6580

1 cycle Li2MnO3

Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

Mn K

Li2MnO3

(a)

Fresh 1 cycle

Li2MnO3 Vs. “Li2MnO3 component” (XAS)

9

0 1 2 3 4 5 6 7 8

(b)

FT m

ag. (

a.u.

)

R (Å)

Mn K

1 cycle Li2MnO3

1 cycle HE5050

Li2MnO3

6540 6550 6560 6570 6580

1 cycle Li2MnO3

1 cycle HE5050Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

Mn K

Li2MnO3

(a)

• No evidence for significant Mn4+ reduction to Mn3+

• Single O-bond gives Mn-O coordination of ~6 (1.9 Å) (See ES193, H. Iddir)

• No damping of Mn-M peaks • Integrated LMO is clearly different than pure LMO

Li2MnO3 Vs. “Li2MnO3 component” (XAS) Croy et al. unpublished

0.5Li2MnO3•0.5LiMn0.375Ni0.375Co0.25O2

6537 6540 6543

Der

ivat

ive

of N

orm

. Abs

orpt

ion

Energy (eV)

Li2MnO3 - 1 cycleHE5050 - 1 cycle

Li2MnO3

10

0 1 2 3 4 5 6 7 8

Mn - M

FT

mag

nitu

de (a

.u.)

R (Å)

(a) Mn K

Li2MnO3_Ni

3 cycles

50 cycles

Mn - O

0 1 2 3 4 5 6 7 8

R (Å)

Li2MnO3_Ni

3 cycles

50 cycles

Ni KNi - M

Ni - O

(b)

FT m

agni

tude

(a.u

.)• Clear changes observed in the local manganese environment in early cycles.

• Very small changes to the nickel environment.

• Mn-Ni interactions in MnNi-rich regions stabilize Mn on edges of domains.

• Interior of Li/Mn-rich regions are most effected by cycling.

Li2MnO3 Component Most Effected by Cycling

Croy et al., JES, 161 A318 (2014) 11

0.5Li2MnO3•0.5LiMn0.5Ni0.5O2

• Charge ordering, especially between Li and Mn, is the dominant factor dictating local, nanocomposite nature of xLi2MnO3•(1-x)LiMO2 oxides. • Li/Mn rich domains form at low temperatures early in synthesis resulting in two different average environments for Mn relative to other TMs. • Li/Mn rich domains are locally similar to Li2MnO3 ; however, the observed electrochemistry is different than pure Li2MnO3.

• Li/Mn ordering plays a critical role in structural changes on cycling.

Part 1 Summary

12

x [xLi2MnO3•(1-x)LiMn0.5Ni0.5O2] 0.1 0.3 0.5

Voltage Fade vs. x

0

10

20

30

40

50

60

70

80

3.7

V C

harg

e C

apac

ity (m

Ah/

g)

VF is a Function of x in xLi2MnO3•(1-x)LiMO2 Croy et al., JES, 161 A318 (2014)

2.0 2.5 3.0 3.5 4.0 4.5 5.0

Cell Voltage (V)

4.7 - 2.0 V cycles

cycles 4, 18, 30

x = 0.5 dQ

/dV

• Activation and high voltage cycling (~4.4 V) result in low voltage capacity (<3.5 V) due to structural changes – VF configuration.

• Magnitude of “VF capacity” is proportional to lithium in the transition metal layers of the initial composite – e.g., Li and Mn ordering.

13

0.0 0.2 0.4 0.60

150

300

450

600

∆V a

t 50%

SO

C (m

V)

x [xLi2MnO3_(1-x)LiMn0.5Ni0.5O2]

Hysteresis vs. x

x [xLi2MnO3•(1-x)LiMn0.5Ni0.5O2]

Hysteresis is a Function of x in xLi2MnO3•(1-x)LiMO2

0 50 100 150 200 250 300

2

3

4

5

∆V = 540 mV

Capacity (mAh/g)

4.7 V

3.7 V

4.1 V

x = 0.5

cycle 2, 3, 4

Cell

Vol

tage

(V

)

Croy et al., JES, 161 A318 (2014)

2.0 2.5 3.0 3.5 4.0 4.5 5.0

Cell Voltage (V)

4.7 - 2.0 V cycles

cycles 4, 18, 30

x = 0.5 dQ

/dV

3.0-3.3 V

• Lithium removed above ~4.0 V cannot be entirely re-accomodated until ~3.2 V on discharge. • Represents a ~1.0 V hysteresis for some fraction of the overall lithium content. • Magnitude depends on x in xLi2MnO3•(1-x)LiMO2.

14

2 3 4 5

2

0

2

1

2

dQ/d

V

Cell voltage (V)

3

4

5 6 7

8

• Study the structure and TM oxidation states at equivalent voltages and SOCs on charge and discharge (XAS and XRD) after activation. • Charge/discharge to each point followed by 12 hour hold or rest. • Cathodes were prepared and sealed in aluminized Mylar pouches under helium atmosphere for ex-situ XAS and XRD.

0 50 100 150 200 2502.0

2.5

3.0

3.5

4.0

4.5

5.0

76

5

4

3

2

1,8

Capacity (mAh/g)

Cell

vol

tage

(V

) After Activation

Croy et al., J. Phys. Chem. C, 6525 (2013) X-ray Studies on Hysteresis in Composites Toda HE5050

15

0 1 2 3 4 5 6 7 8

FT m

agni

tude

(a.u

.)

R (A)

3.6 Vd

3.9 Vd

4.2 Vd

3.6 Vc

4.2 Vc

4.7 Vc

3.1 V

(a) Ni K EXAFS

• Same trend for Mn and Co K-edge data Spectroscopic evidence of hysteresis. • Lithium does not have access to the same sites on charge and discharge even at

equivalent lithium contents.

8340 8350 8360

3.6 Vd

3.9 Vd

Nor

mal

ized

Inte

nsity

(a.u

.)

Energy (eV)

3.6 Vc4.2 Vc

(21% SOC) (77% SOC) (75% SOC) (51% SOC)

(100% SOC) = 257 mAh/g

Ni K XANES

Croy et al., J. Phys. Chem. C, 6525 (2013) Structural Hysteresis Confirmed by XAS

Toda HE5050

16

VF and Hysteresis Are Correlated

• Cells cycled in truncated windows with decreasing, lower cutoff voltages (x axis).

• Upper cutoff voltage constant at 4.7 V.

• Main graph shows the average fade in discharge OCV between cycles 2 and 23 as a function of lower cutoff voltage (-----). • Inset shows calendar time plot – Fade in OCV as a function of time on test – along with 4 cells held at 4.7 V for several, equivalent times on test ( ).

• Decreasing the lower cutoff voltage clearly increases voltage fade with a maximum fade found between 3.0 – 3.3 V. • Cycling between 4.7 – 3.2 V accelerates voltage fade more than any other electrochemical exposure. Same window giving significant hysteresis.

Toda HE5050

Gallagher et al., Electrochem. Comm., (2013)

17

Li/TM

Li/TM

Li/TM

Discharged

Charged

Discharged

Hysteresis

Voltage Fade

Proposed Mechanism of Voltage Fade and Hysteresis

• VF/hysteresis are related, structure and charge/discharge energetics differ.

• Any model for this class of materials must account for both.

Our Conceptual Interpretation of VF and Hysteresis

Gallagher et al., Electrochem. Comm., (2013)

• Charging to ~3.8 V and above induces migration to tetrahedral sites

• Cations are ‘stuck’ in that site until a critical Li

content is reached on discharge (~3.2 V) • At the critical lithium content cations can:

• migrate back to original site (hysteresis)

• continue on to the lithium layer (voltage fade)

• remain ‘stuck’ – capacity loss, impedance rise

18

Why Propose This Model? • ‘Dumbbell’ formation (tetrahedral migration) in structural transformations of layered materials.

Ma et al., JES, 160 , A279 (2013)

• Now have spectroscopic evidence for hysteresis mechanism (B. Key – ES187)

(H. Iddir – ES193)

Observed for other layered systems

• LiVO2 (Thackeray) • LiMnO2 (Ceder, Bruce) 3.2 V processes • Li1.2Mn0.4Cr0.4O2 (Balasubramanian) • LiMn0.5-xCr2xNi0.5-xO2 (Karan)

ES187, B. Key

6Li NMR

0 10 20 305.0

5.2

5.4

5.6

5.8

6.0

6.2

Firs

t-She

ll O

xyge

n C

oord

inat

ion

First-Cycle Charge (hours)

Mn-ONi-O

ES193, H. Iddir XAS

19

ORNL neutron data

2.0 2.5 3.0 3.5 4.0 4.5 5.0

Cell Voltage (V)

4.7 - 2.0 V cycles

cycles 4, 18, 30

2.0 2.5 3.0 3.5 4.0 4.5 5.0

Cell Voltage (V)

4.7 - 2.0 V cycles

cycles 4, 18, 30

2.0 2.5 3.0 3.5 4.0 4.5 5.0

dQ

/dV

Cell Voltage (V)

4.7 - 2.0 V cycles

cycles 4, 18, 30

x = 0.1

x = 0.3 x = 0.5

Increase in capacity in 3.7 V region from cycle 4 to 30 (2.0 – 4.7 V windows) ‘Voltage Fade config.’

Decrease in capacity in 3.8 – 4.7 V region from cycle 4 to 30 (minus irr. cap.) ‘Hysteresis config.’

Hysteresis Voltage Fade

Decrease in ‘hysteresis capacity’ leads to a concomitant increase in ‘voltage fade capacity’

x [xLi2MnO3•(1-x)LiMn0.5Ni0.5O2] 0.1 0.3 0.5

0

5

10

15

20

25

Cap

acity

(mA

h/g)

Croy et al., JES, 161 A318 (2014)

EC models agree – See ES189, D. Dees

Testing Predictions of Proposed Model

20

Li/TM

Li/TM

Li/TM

Discharged

Charged

Discharged

Hysteresis

Voltage Fade

Proposed Model of LMR-NMC Materials

• Future studies will be aimed at direct verification, from experiment and theory, of cation occupancies (tetrahedral/octahedral) and oxidation states at different SOCs.

Gallagher et al., Electrochem. Comm., (2013)

21

Summary • Charge ordering drives local “composite” nature of LMR-NMC materials.

– Activation of the Li2MnO3 component is necessary for both – Both increase in prominence with increasing x – Cycling above ~4 V extracts lithium that is not accommodated on discharge until 3.0 – 3.3 V – Window studies show that a lower cutoff of 3.0 – 3.3 V accelerates Voltage fade faster than any other electrochemical exposure

• Proposed model involves tetrahedral migration of lithium and TMs on charge creating a barrier to lithium insertion until a sufficient driving force is established at some critical lithium content on discharge (~3.2 V). • Driving force triggers one of three possibilities:

1) TM/Li returns to its original octahedral position in the TM layer Hysteresis 2) TM/Li migrates to a new octahedral position (e.g. in the Li layer) Voltage Fade 3) TM/Li becomes trapped in the tetrahedral site Loss of capacity, increased impedance

• Voltage fade and hysteresis are related and depend on x in xLi2MnO3•(1-x)LiMnO2.

23