LLNL-PRES-XXXXXX This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344. Lawrence Livermore National Security, LLC

!"#$%&'(%)*+,-./0%100234%5/,67%8/96/:%

;20/%<=%!"#$%

!"#$%&'(!")*+,-.(/0+*1,-(#10-*2134($)51.6-)(%-4-1*67(&+.4+*0289(

>3:,/0?/%>69/,@-,/%A376-034%>3B-,37-,*%CD%E--+=%FDED%)/-=%;D%>//=%GD%C-0/9=%5D%5-,34/H=%FD%C32@300%

This presentation does not contain any proprietary, confidential, or otherwise restricted informationProject ID# ST129

!"#$"%#&'

•! Lack of understanding of hydrogen physisorption and chemisorption (Barrier O)

•! System weight and volume (Barrier A) •! Charge/discharge rate (Barrier E)

FY15 DOE Funding: $250K FY16 Planned DOE Funding: $735K Total Funds Received: $985K

Funded Partners: Sandia National Laboratories (lead) Lawrence Berkeley National Laboratory

Project start date: 9/17/2015

Timeline Barriers addressed

Team Budget

Phase I end date: 9/30/2018

#%

(#)#"*+,#'-'./0#,1%"#%

!! 5/734%-,.306?%I,3@/:-,JH%K5'LHM%%!! )6/,3,?N6?34%607/.,37/+%B24J%30+%

030-H?34/%@/734%N*+,6+/H%%!! )6.NOP,/HH2,/%H*07N/H6H%%!! L20?76-0346Q/+%?3,B-0%/0?3PH24307H%

30+%2.$.34'+*+.,.+5+%+6'7#8%*'!! G-,B/07%H267/%I-,%@-+/4%7/H760.%30+%

9346+376-0%

!! 1@B6/07OP,/HH2,/%RSG%I-,%.3HOH2,I3?/%607/,3?76-0H%

!! 9.:1';<$*='*/4.$21%.+'*+8'#7%44%.+'42#,1$.4,.2='

!! (4/?7,-0%@6?,-H?-P*%30+%RO,3*%HP/?7,-@6?,-H?-P*%

!! >-:O/0/,.*%6-0%H?377/,60.%I-,%H2,I3?/%N*+,-./0%+/7/?76-0%

2 nm

16001590158015701560Photon Energy (eV)

Al K-edge TEY (Surface) TFY (Bulk)

Al2O3

NaAlH4

Ti-NaAlH4

Al-foil

!"#!$%"&'!("' )%#(!*%$'+'

!! >3*+137'?.+1#'@*$).':.$'4.$/#+14'!! ABCD1:(2.202+'7.)#,3)*$'8=+*7%,4':.$'

/3)ED43$:*,#D%+1#$:*,#',F#7%41$='!! T43HH6?34%5&%U'E%+#1%,'?.+1#'@*$).'

:.$'+.+<#G3%)%/$%37'1$*+42.$1'!! HF*4#<5#)8'7.8#)%+6':.$'4.)%8<41*1#'

2F*4#'1$*+4:.$7*1%.+'E%+#1%,4'!! @.7231*1%.+*)'42#,1$.4,.2='!! @.773+%1='4.:1&*$#'-'8*1*/*4#4'

)*518T%:644%P,-96+/%,.773+%1='1..)4'30+%:.3+8*1%.+*)'3+8#$41*+8%+6'-I%PN/0-@/03%.-9/,060.%7N/,@-+*03@6?H%30+%J60/76?H%7-%/03B4/%+/9/4-P@/07%-I%H-46+OPN3H/%N*+,-./0%H7-,3./%@37/,634H%

!%

Bond chemistry

Thermodynamics &

Task 1: ThermodynamicsTask 2: TransportTask 3: Gas-surface interactionsTask 4: Solid-solid interfacesTask 5: Additives and dopantsTask 6: Materials informatics

Additives

Approach: HyMARC tasks target thermodynamics and kinetics

3

Approach: Validated multiscale modeling

10-10 10-6 10-210-8 10-4

Length (m)

10-12

10-9

10-3

10-6

100

Tim

e (s

)

MgB2

Mg(BH4)2

H2

Quantum Monte Carlo

DFT/ab initio molecular

dynamics

Classical molecular dynamics

Kinetic Monte Carlo

Phase field modeling

Modeling approach focuses on integrating techniques at multiple scales to tackle complexities of “real” (beyond-ideal) materials

Validation: X-ray spectroscopy,

NMR, FTIR

Validation: TGA, Sieverts,

PCT calorimetry

Validation: TEM & microscopy

4

LLNL contributions to HyMARC

Porous carbon synthesis

X-ray absorption/emission

Jon Lee

Ted BaumannDFT & ab initio molecular dynamics: Brandon Wood

Phase-field mesoscale kinetic modeling: Tae Wook Heo

Kinetic Monte Carlo: Stanimir Bonev

Quantum Monte Carlo: Miguel Morales

Postdocs: ShinYoung Kang, Keith Ray, Roman Nazarov, Patrick Shea

Multiscale modeling

Pat Campell

5

Progress towards milestones with key LLNL FY16 activities

Q1: (S) Synthesize library of bulk-phase model storage systems for T1-T5Task 1: Synthesize and characterize high-surface area carbon frameworks with pore sizes < 2 nm (100%)

Q2: (S) Demonstrate size control method for one prototype complex hydride nanostructureTask 1: Perform sensitivity studies to determine design rules for nanoconfined metal hydrides (50%)

Q3: (C) Demonstrate in-situ soft X-ray AP-XPS, XAS, XES tools, with sample heatingTasks 3 & 5: Performed XAS spectroscopy of hydrides and dopants at LBL/ALS, SLAC, and Canadian Light Source (75%)

Q4: (C+T) Identify hydride mobile species and diffusion pathwaysTask 2: Perform ab initio molecular dynamics of surface and bulk model hydrides (80%)Task 2: Gather defect formation energies and migration barriers for model materials (70%)Task 2: Construct initial KMC framework for interface/amorphous transport (100%)Task 3: Compute H2 dissociation energetics on model surfaces (75%)Task 3: Construct initial microkinetic model for surface reactions (100%)Task 4: Construct initial phase-field modeling code for solid-solid interface kinetics (100%)

Q4: (S+C) Synthesize library of nanoparticles: 1 – 5 nm, 5 – 10 nm, > 10 nm for one prototype hydrideTasks 1-5: Select model materials for facilitating experiment-theory feedback within each Task (100%)

6

C*4E'T'I,,.72)%4F7#+1J'I,,3$*1#'4.$/#+1'#+#$6#1%,4'

>?@'

&6-0%

-P7C<<%

-P7SC(%-P7C<$B%

FSGG%

FG%SC(%

SC([&V%SC([&!%

C>\S%

H2 p

hysi

sorp

tion

ener

gy, e

V

H2 p

hysi

sorp

tion

ener

gy, e

V

H2 loading (wt.%) 0.5 1.5 1.0

M)/1(+"*$0/N%OAJ%?#P%Q"?,*#%,9%?"/%-)*%R""<#%SG!%9,*%T2%#04)#%0/%AUG7=V=%

;)<)140/N%"33*,3*0"4)%SG!%"33*,W0+"40,/#%9,*%(0N()*73*)##8*)%<,"-0/N%X04(%+8<403<)%T2%"-#,*340,/%#04)#%

V%))'*)4.'/#'34#8'1.'2*$*7#1#$%W#',)*44%,*)'2.1#+1%*)4':.$':313$#'7*1#$%*)4'4,$##+%+6'OC*4E'UP'*+8'2$#8%,1'%4.1F#$74'1F*1',*+',.72*$#'*6*%+41'9*+8%*'7#*43$#7#+14'

O8"/48+%A,/4)%J"*<,%YOAJZ%("#%[))/%-)+,/#4*"4)-%"#%"%?0"[<)%"33*,"1(%9,*%0+3*,?0/N%"/-%[)1(+"*$0/N%4()%"118*"1.%,9%3*)-014)-%T2%#,*340,/%,/%+,-)<%AUG#%SG!%98/140,/"<#%N0?)%?)*.%-0#3"*"4)%)/)*N0)#D%[84%N),+)4*0)#%"*)%*)<0"[<)%"/-%1"/%[)%8#)-%"#%OAJ%0/384#%

]%

C*4E'T'I,,.72)%4F7#+1J'H.$.34',*$/.+'4=+1F#4%4'

I*,N*)##%0/%*)3*,-810[<)%#./4()#0#%3"4(X".#%9,*%1"*[,/#%X04(%4"0<,*)-%3,*,#04.%L%1()+0#4*.%

COOH

HOOC

HO

HOOC

HOOC

HO O

OH

OHOOHO

@2(.)<$

"R>RTDI$U?$

!"#$%&'()!!*+,-./(!&0,-,12+&3*45-'6)-2+,78&

D3*30(-*$

VR$TDI$?M8W?$+$

XY$-6-9,)-.4)$ C@$<239+,),$:3A243--,:10($

J. Am. Chem. Soc. 132, 14067 (2010); J. Phys. Chem. Lett. 2, 921 (2011); ACS Nano. 2014, 8, 11013–11022

6*"3()/)%,W0-)7-)*0?)-%#./4()#0#%

3*,1)-8*)%

T0N(7!%()"4%4*)"4+)/4%1"/%*)-81)%3,*)%#0\)%"/-%0/1*)"#)%3,*)%+,/,-0#3)*#04.%X(0<)%0+3*,?0/N%1*.#4"<<0/04.%

&VC%1,+3*)##0,/%1"/%*)+,?)%<"*N)*%3,*)#%X04(,84%-"+"N)%

K#*1<1$#*1#8'*+8',.72$#44#8'4*72)#4'8#)%"#$#8'1.'9*+8%*':.$'F%6F<2$#443$#'6*4'4.$21%.+'7#*43$#7#+14':.$'/#+,F7*$E%+6'%4.1F#$7'2$#8%,1%.+4'

<%

Pore width (nm)

Por

e vo

lum

e cm

2 /g-n

m)

Pore width (nm)

Por

e vo

lum

e cm

2 /g-n

m)

Pore width (nm) P

ore

surfa

ce a

rea

m2 /g

)

Task 1 Accomplishment: Modeling sorbent charge/field effects

Developed first-principles modeling framework for simulating charge and field effects on H2 physisorption on conductive sorbents for guiding synthesis

±qUses Effective Screening Medium Method* to compute physisorption as function of Fermi level

Bound & polarized H2

Infinite dielectric medium

Can be applied to functionalized materials and any 2D conductive sorbents, and will be used to assess polarization effects in HyMARC experiments

Unbound & unpolarized

H2 (reference)

*Otani & Sugino, Phys. Rev. B 73, 115407 (2006)

Predicted field-induced enhancement of physisorptionV

9

Task 1 Accomplishment: More accurate hydride thermodynamics

Established protocols for more accurate DFT computations of hydride thermodynamics that can improve predictive power beyond established methods by accounting for:

Mg

BH

• Dynamical (anharmonic) contributions to free energy via explicit finite-T simulations• Microstructure, including phase coexistence and interface/surface contributions• Phase morphology & nucleation: dispersed molecular defects vs. amorphous vs. crystalline

Anharmonic molecular modes in I-m42 (MgBH4)2 from ab initio molecular dynamics

Results: Anharmonic modes can contribute as much as

12 kJ/mol in some predicted phases of Mg(BH4)2, which could alter phase stability

10

Task 1&4 Accomplishment: Hydride phase fraction predictions

Demonstrated thermodynamic phase fraction prediction as a function of pressure, temperature, and size, accounting for microstructure, phase coexistence, and nucleation

DFT + Wulff construction Free energy minimization with respect to phase fractions in assumed microstructure

Predicted size-dependent phase fractions for hydrogenation of Li3N/[LiNH2+2LiH] system in core-shell microstructure

Interface-induced Li2NH phase suppression improves kinetics and reversibility

Successfully predicted Sandia-measured reaction pathway in nanoconfined Li-N-H and confirmed relevance of interfaces in determining phase stability

11

Task 1,4&5 Accomplishment: Phase nucleation in hydrides

Computed critical nucleus size for Li-N-H and Mg-B-H systems, including intermediates

Computed critical nucleus sizes for model hydrides to guide nanoscale synthesis as a means of kinetically suppressing undesirable intermediate phases

LiH nucleation during hydrogenation of the Li3N/[LiNH2+2LiH] system

Homogeneous nucleation Heterogeneous nucleation

{pij} configuration index

Interface energy approximation:

Small size: No nucleation

Large size: Nucleation

12

C*4E'TZ\-]'I,,.72)%4F7#+1J'91*1%41%,*)'*22$.*,F'1.'%+1#$:*,#4'S)+,/#4*"4)-%/,/7)g80<0[*08+%#4"40#401"<%"33*,"1(%4,%g8"/409.%)^)14#%,9%

/,/70-)"<%0/4)*9"1)#%"/-%"--040?)#%,/%*)"140,/%3"4(X".#%

Li3N

Li 2NH+L

iH Li 3N

LiNH2+LiH

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

0 0.2 0.4 0.6 0.8 1Mole fraction X

0

0.2

0.4

0.6

0.8

1

Nor

mal

ized

freq

uenc

y p Li3N

Li2NHLiHLiNH2

C24J% +a#"0@%+aY"0@% +a!"0@%

D(E(FG((

D(E(FG((

D(E(HIG(

D(E(JG((

D(E(JFFG(

D(E(KHG((

D(E(JFFG((

D(E(JFFG((

C24J% +a#"0@%+aY"0@% +a!"0@%

K=8$.6#+*1%.+'O_HZ"RR132'$'

9#+4%1%"%1='*+*)=4%4',*+'41##$'1.&*$84'7.41'2$.7%4%+6'#+6%+##$%+6'*"#+3#4'

I*,["[0<04.%I%,9%d02ET%3("#)%#833*)##0,/%0/%/"/,#1"<)%ed0ET2c2d0Tfad05E%#"+3<)-%,?)*%0/4)*9"10"<%)/)*N0)#%

0.3 ! pij ! 0.7

0.3 ! pij ! 0.7

#V%

Fraction of phase X

Pro

babi

lity

Fraction of phase X Fraction of phase X Fraction of phase X

Fraction of phase X

Pro

babi

lity

Fraction of phase X Fraction of phase X Fraction of phase X

A#F=8$.6#+*1%.+'O_H`T/*$P'

Task 2 Accomplishment: Ab initio molecular dynamics (AIMD)

AIMD is being used to elucidate transport/reaction pathways and enable finite-temperature calculations of thermodynamic and spectroscopic quantities

Established software framework for automating AIMD, to be integrated within open-source AiiDA task management engine

AIMD of an Mg(BH4)2

surface @800K showing

desorption of H2and BH3

Initialization(DFT electronic structure)

Thermalization(NVT AIMD)

dynamically set number of steps

for each task

internal error-

checking identifies successful

completion of each task

Molecular dynamics manager

Equilibration(NVE AIMD)

Production(NVT AIMD)

Observables

Atomic coordinates (e.g, from ICSD)

14

Task 2&3 Accomplishment: Multiscale surface hydrogen kinetics

Surface redistribution of atomic hydrogen Surface dissociation + diffusion

Demonstrated multiscale modeling of surface dissociation and diffusion of hydrogen by synthesizing DFT and continuum approaches

Combining with low-energy ion scattering (Sandia) to analyze surface diffusion

On model carbon surface from DFT calculations of H on graphene

15

C*4E'\'I,,.72)%4F7#+1J'9.)%8'2F*4#'1$*+4:.$7*1%.+'E%+#1%,4'

S)+,/#4*"4)-%1,83<)-%"4,+0#401%"/-%+)#,#1"<)%9*"+)X,*$%1,+[0/0/N%4()*+,-./"+01#D%+)1("/01"<%#4*)##D%"/-%3("#)%/81<)"40,/aN*,X4(%4,%+,-)<%#,<0-%3("#)%4*"/#040,/%$0/)401#%

M5E$%62--%-*-23)O%.*+-26"1."E%-*-23)O%-E"#C1.+)O%0.Y5#.D.+)%

•! ;811)##98<<.%-)+,/#4*"4)-%"33<01"40,/%4,%+,-)<%0/4)*#4040"<%(.-*0-)%YI-TWZk%/,X%"-"340/N%4,%ANT2%["#)-%,/%SG!%YddEdZ%c%1<"##01"<%AS%Y;"/-0"Z%

•! J,+3"*0#,/#%X04(%dMEd%)W3)*0+)/4#%,/%)/1"3#8<"4)-%(.-*0-)#%#(,X%3*)-0140,/#%1"348*)%#0\)7-)3)/-)/4%$0/)401%[)("?0,*%

Phase-field modeling code (LLNL)

Phase 1 or

Grain 1

Phase 2 or

Grain 2

F = dV[ f (!,X,...)! +

(gradient energy)+ (elastic energy)+...]

DFT (LLNL) + Molecular dynamics (Sandia)

!"! !"# !"$ !"% !"& !"'(!"!!)

(!"!!*

(!"!!&

(!"!!$

!"!!!

!"!!$

!"!!&

+

+

+,-+. /0012+34561 +7*08+95215:

;</=+71>

?@09<:

A 9<B9C /0/9D+7E :

F G%!!H

Phase kinetics & comparison with LBNL experiments

100 200 300 400 500Time steps

0

0.05

0.1

0.15

0.2

0.25

0.3

Vol

ume

frac

tion

of H

-ric

h ph

ase

d=74nm (X0=0.25)d=158nm (X0=0.25)

Parameters

D)!<(7")*+,-.102+.(

#+"<<)*%3"*401<)#%Y9"#4)*Z%

<"*N)*%3"*401<)#%Y#<,X)*Z%

#$%

Vol.

fract

ion

of H

-ric

h ph

ase

Time steps

Free

ene

rgy

H composition

@T=300K

C*4E'b-]'I,,.72)%4F7#+1J';I9'42#,1$.4,.2='.:'F=8$%8#4'

S)+,/#4*"4)-%C7*".%"[#,*340,/%#3)14*,#1,3.%YC>;Z%9,*%3*,[0/N%1"4"<.#4%1()+0#4*.%"/-%#8*9"1)%?#P%[8<$%(.-*,N)/"40,/%)^)14#%,/%1"4"<.#4#%X04(0/%-,3)-%(.-*0-)#%

T.-*,N)/"40,/%)^)14#%,9%!0J<5%"/-%!0J<57-,3)-%E"><T:%X04(%?"*.0/N%1"4"<.#4%<,"-0/N#%["#)-%,/%!0%d27D%d57)-N)%C>;%Y)W3)*0+)/4#%3)*9,*+)-%"4%>d;Z%

I8*)%!0J<5%8/"^)14)-%[.%T2%

)W3,#8*)%T270/-81)-%1()+01"<%1("/N)#%#0N/"<)-%[.%-*"+"401%1("/N)%0/%#0N/"<%#4*)/N4(%Y0P)PD%#8*9"1)%1,/1PZ%

;I9'$#43)14'.+'R*<I)<K'*+8'?6<^<K'4=41#74'*$#'34#8'1.'8#"%4#'7.$#'$#*)%41%,'43$:*,#'F=8$.6#+*1%.+'7.8#)4Z'&F%,F',*+'/#'8%$#,1)='"*)%8*1#8'/=',.7231%+6';I9'OQ^RQP'

#]%

TiCl3-doped NaAlH4

TiCl3

External collaborations

• NMR and borohydride chemistry: T. Autrey, M. Bowden, B. Ginovska-Pangovska(PNNL; H2 Storage Characterization Group)• Several discussions, plus site visits in November 2015 and April 2016

• HyMARC will focus on solid-state aspects and PNNL on borohydride chemistry

• Neutron diffraction/spectroscopy: T. Udovic (NIST)• DFT computations of H2 physisorption on MOFs: M. Head-Gordon (LBNL;

H2 Storage Characterization Group)• Li-N-H system: N. Poonyayant, N. Angboonpong, and P. Pakawatpanurut

(Mahidol University, Thailand)• Kinetic Monte-Carlo for solid-state diffusion: H. Kreuzer (Dalhousie U.)

Collaborations are crucial for realizing HyMARC goals

Also extensive collaborations within HyMARC

18

Remaining challenges/barriers & mitigation strategies

• Understanding internal microstructure evolution is challenging• Working with HyMARC team and collaborators to develop protocols for imaging low-Z

materials with minimal beam damage• Focusing early efforts on materials (e.g., MgH2) that can be more easily characterized

• Phase transformations of complex hydrides combine composition & phasechanges, yet we are currently treating these separately• Separate treatment allows each to be independently validated, but integration of both will

be explored in collaboration with the Mg(BH4)2 individual FOA project in FY17

• Need better validation of thermodynamics and transport for amorphous materials• Working with HyMARC team to synthesize amorphous materials for testing our amorphous

transport and thermodynamics theory framework

• Improved interface energy estimates are desirable• HyMARC data will be used to benchmark new approaches for dealing with chemical and

elastic contributions to the interface free energy• Statistical approach allows us to quantify robustness of predictions with respect to

interface energy assumptions

• QMC needs reliable input geometries for sorbent energetics• HyMARC will collaborate with M. Head-Gordon (LBNL, part of NREL-led consortium) to use

high-throughput DFT with corrected functionals on clusters to obtain geometries

19

Proposed future work

Milestone Description & LLNL subtask % complete

Q3 FY16 Demonstrate in situ soft X-ray AP-XPS, XAS, XES tools with sample heatingPerform synchrotron XAS/XES of hydrogenation on catalyzed materials

100

Q4 FY16 Identify hydride mobile species and diffusion pathwaysCompute defect formation energies and ab initio dynamics of mobile species in model systems

25

Q4 FY16 SMART

Synthesize library of nanoparticles: 1-5 nm, 5-10 nm, > 10 nm for prototype hydrideDemonstrate porous carbon nanoconfining medium synthesis with controlled pore sizes

50

Q1 FY17 Compute H2 binding curves by QMC 15

Q2 FY17 Perform sensitivity analysis of local binding and second-sphere effects 0

Q2 FY17 G/NG

Rank improvement strategies: open metal sites, acid sites, polarization effects, phase change materialsPerform analysis of potential of open metal sites (via QMC), acid sites/polarization (via DFT), and phase change materials (thermodynamic analysis)

25

20

Summary

• Integrated theory/synthesis/characterization framework of HyMARC aims toprovide foundational understanding and new tools for solid-state hydrogenstorage

• FY16 LLNL modeling tasks focused on establishing framework for multiscaleintegration and approaches for beyond-ideal materials modeling

• FY16 LLNL synthesis tasks focused on establishing key strategies for tailoredporous carbon

• FY16 LLNL characterization tasks focused on spectroscopic changes uponhydrogenation (complements LBNL activities)

• Early feedback between experiments and theory on the Pd-H and Mg-H systemsdemonstrates feasibility of multiscale (de)hydrogenation kinetics simulation

• Will be moving towards testing on complex hydrides and more complicatedsorbents in FY17

• Flexible modeling, synthesis, and characterization protocols are being developedin preparation for application to upcoming FOA projects

21

CF*+E'=.3':.$'=.3$'*11#+1%.+e'

!!%