Catalyzed Nano-Framework Stabilized High Density Reversible Hydrogen

Storage Systems

E. RönnebroT. Boyle

Sandia National Laboratories

F.-J. WuJ. StricklerAlbemarle

Corporation

S. Arsenault, D. Mosher, S. Opalka, X. Tang, T. Vanderspurt, B. Laube, R. BrownUnited Technologies Research Center

DOE Hydrogen Program Annual Merit Review

Arlington, VAJune 9, 2008

Project ID STP16

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

2

Overview

Timeline7/1/07 Start (signed 9/4/07)7/1/10 End

Budget$1.26M Total Program DOE: $1.01M

SNL: $360kAlbemarle: $90kCost share: 20% (31% UTRC $)

FY07: $80kFY08: $480k

Barriers AddressedP. Lack of Understanding of Hydrogen Physisorption and ChemisorptionA. System Weight and VolumeE. Charging/Discharging Rates

www.eere.energy.gov/hydrogenandfuelcells/mypp

Partner ParticipationSandia National LaboratoriesAlbemarle CorporationAspen Aerogels

3

ObjectivesDesign & synthesize hydride / nano-framework combinations to improve:

Reversible capacity Desorption temperature Cyclic life

Build upon successes previously demonstrated in the community and extend to a wider range of doped, functionalized and catalyzed framework chemistries to:Advance the understanding of behavior modification by nano-frameworks Obtain / maintain nano-scale phase domainsTune hydride / framework interactions to

Decrease desorption temperature for highly stable compoundsStabilize high capacity compounds – ligand eliminationInfluence desorption product formation

Activate H2 dissociation on highly dispersed catalytic sites

4

High Level Approach

Framework Synthesis

Hydride Synthesis

-15

-10

-5

0

5

10

Hea

t Flo

w (W

/g)

0 100 200 300 400T em perature (°C )

2007-091D SC .03 LiBH 4 0 .5gr Spex m il led 15m in2007-091D SC .07 Silica 0.25gr L iBH 4 0 .25gr

E xo U p U niversa l V 4.4A TA Ins

Screening CharacterizationChemical reactivities: DSC, TGAFramework morphology: BET

Hydride Incorporation

Structure & PerformanceSEM, TEM, Sievert’s

Atomistic & Thermodynamic ModelingFramework design, Material compatibility

Down-select

HSCVASP

DSCReaction Testing

0 100 200 300 400 5000.00

0.05

0.10

0.15

0.20

C

kmol

Temperature

H2(g)

ZrO2 BLiBH4

Li2ZrO3ZrH2

0.0

0.1

0.2

0.3

0.4

0.5

0 1 2 3 4 5Time (Hours)

Pres

sure

(bar

)

0

100

200

300

400

Tem

pera

ture

(°C

)

LiBH4 + SiO2 (aerogel) Desorption, 290°C, 1bar

5

Detailed ApproachPhase

TaskPhase I

Years 1 & 2Phase IIYear 3

First Principles Modeling

Examine hydride / framework interactions.Screen doped / catalyzed / functionalized frameworks.

Evaluate mechanisms of down-selected hydride / framework systems.

Thermodynamic Modeling

Assess chemical compatibility of hydride / framework combinations.

Comparison of bulk scale & nano scale properties.

Hydride Development

Synthesize high capacity hydride materials.

Refine high capacity hydride synthesis methods.

Hydride Incorporation

into NFS

Incorporate hydrides into designed frameworks.Characterize properties.

Maximize hydride incorporation into framework for improved capacity.

Nano-Framework

Development

Synthesize and characterize uncatalyzed and catalyzed frameworks.

Optimize doped / catalyzed / functionalized framework for down-selected systems.

Iterative design and synthesis of high H2 capacity systems.

6

Milestones

Date Milestone or Go / No Go Decision

2008 Q1 Select advanced nano-framework structure

2008 Q1 Demonstrate synthesis of desired nano-framework structure

2008 Q2 Synthesize top UTRC / Albemarle candidate hydride material

2008 Q4 Evaluate relative performance of catalysts

2008 Q4 Synthesize optimal catalyzed nano-framework structure

2009 Q1 Demonstrate loading of UTRC / Albemarle hydride into catalyzed framework

2009 Q2 Demonstrate loading of Sandia hydride into catalyzed framework

2009 Q3Go / No Go on whether to proceed with original plan or redirect based on:> 50% hydride deposition into the CFSReasonable absorption/desorption behavior for at least one cyclePerformance relative to the state-of-the-art material shows promise

2008 Q1 Synthesize top Sandia candidate hydride material

7

Material Focus & Partner Roles

A&TM: Atomistic and Thermodynamic ModelingNFS: Nano Framework Structure

HSM

HSMCSMNFS

CSM

NFS

A&TM

Hydrogen Storage Material (HSM)SNL: Ca(BH4)2 (stable borohydride)UTRC / Albemarle: NaTi(BH4)4*ligand, …Compatibility screening / Baseline modeling: LiBH4

HSM: Hydrogen Storage MaterialCSM: Combined Structure & Material

8

Capabilities – Nano-Framework Development

Autoclave System2200 PSI 300°C

Wet Chemistry Laboratory

Process DevelopmentRange of chemistries– Oxides– Non-oxides– Polymers, …

Large Scale Manufacturing– Future cost reduction

9

Capabilities – Hydride Synthesis / Incorporation

Autoclave System- Solvated incorporation- 2200 PSI- 300°C

Solid State Processing- Very rapid, low cost

screening- Limited conditions- High cost for high

volume production

Solution Based ProcessingChemical Design & Synthesis- Excellent control- High purity products- Expensive processing- Cost- effective high

volume production

High-Pressure station- Solid state reactions- Wide range of P and T <20,000psi, <500°C- Autoclave with six samples capability

Solvated Hydride Incorporation- Solvents selected for ease of

removal

10

Experimental Methods – Characterization

XRD of SBP LiMg(AlH4)3

Monoclinic structurerefinement from PND and SR-PXD ongoingInstitute for Energy, NO.

Inte

nsity

2θ (degrees)

XRD of SBP LiMg(AlH4)3

Monoclinic structurerefinement from PND and SR-PXD ongoingInstitute for Energy, NO.

Inte

nsity

2θ (degrees)

Crystalline structure & phase

X-Ray DiffractionDifferential Scanning Calorimetry

Assess reversibility potential, phase behavior

-6

-4

-2

0

2

4

6

8

10

0 100 200 300Temperature (°C)

Hea

t Flo

w (W

/g)

168.2°C

140.3°C

192.6°C

132.3 J/g10.5 J/gExo

Endo 158.0 J/g

Thermogravimetry-Mass Spectroscopy

Desorption temperature & species

0.E+00

1.E-09

2.E-09

3.E-09

4.E-09

5.E-09

0 100 200 300 400 500 600Temperature(˚C)

Ion

coun

t /m

g H2

NH3

( )No co-reactant

0.E+00

1.E-09

2.E-09

3.E-09

4.E-09

5.E-09

0 100 200 300 400 500 600Temperature(˚C)

Ion

coun

t /m

g H2

NH3

( )No co-reactant

* Additional characterization support will be provided by other MHCoE partners

Morphology, catalyst dispersion and size, hydride loading

TEM / High Res SEM

Quantify total metal loading

Inductively CoupledPlasma

BET Nitrogen Porosimetry

Surface area, average pore size, and pore size distribution

11

Hydride

~5 nm diameter

Pore

Concept Input Models RoleAtomic Modeling of Hydride-NFS Interactions

Al2O3NFSWallSlab

LiBH4HydrideCluster

NFS wall~ 0.6 nm

thick

Interface

Mechanistic simulations guide NFS down-selection, design and modification.

Hydride thermodynamics

Interfacial physi-/chemisorptionreactions influence on hydride stability and dehydrogenation

NFS stability and modification:a) Doping to tune reactivity b) Loading with H2 activation

catalysts for reversibilityc) Surface functionalization

SchematicFilled NFS

Density Functional TheoryGround State Minimization

Direct Method Lattice DynamicsThermodynamic Property Prediction

Conduct atomic modeling toinvestigate and prescreen:

Baseline System

12

Adhesion to NFSΔHadh (kJ/mole LiBH4)

(- =favorable)-37 -35 -18

Screening of NFS Stability and Hydride Interactions

ZrO2 NFS predicted to have low reducibility in an H2 atmosphere and weaker tunable interfacial associative interactions with LiBH4.

(strong binding weak binding)

Strategy:Guide experiments by determining promising as well as unfavorable system characteristics.

Baseline System:LiBH4 is highly stable.Balance reversibility by increasing dehydrogenation product interactionwith NFS.

LiBH4 LiBH4 LiBH4

Al2O3SiO2 ZrO2

LiBH4 on Al2O3 LiBH4 on ZrO2LiBH4 on SiO2

13

Hydride Physi-/Chemisorption Interactions w/ NFS

• Dehydrogenation influenced by product adsorption on NFS surface. • Ca(BH4)2/NFS under investigation.

Fully Charged DischargedGround State Dehydrogenation Reaction Predictions

ΔHdeh

kJ/mole LiBH4

8LiBH4 w/o NFS 8 LiH + 8B + 12H2 +1148LiBH4 / Al2O3

Al2O3∗2Li∗2H + 6LiH+8B +12H2 +1008LiBH4 / ZrO2

ZrO2∗Li+ 7LiH + 8B + 12.5H2 +112Equally strong Li-Obond in charged and

discharged states

LiBH4 on Al2O3

Adsorbed Li & H enhance discharge

LiBH4 on ZrO2

2Li & 2H on Al2O3

∗ = AdsorbedExperimental thermodynamic ΔHdeh at 298 K:100 kJ/(mole LiBH4), HSC v. 5.1.104 kJ/(mole LiBH4), Smith and Bass,

J. Chem. Eng. Data 8 (1963) 8.

Li on ZrO2

14

NFS Doping Alters Hydride DehydrogenationLiBH4 on

Sc-doped ZrO2

LiBH4 onTi-doped ZrO2

LiBH4 onV-doped ZrO2

Ground State LiBH4Dehydrogenation Reaction Predictions on NFS

ΔHdeh

kJ/mole LiBH4

Zr0.92Ti0.08O2 +125

ZrO2 +112Zr0.92Sc0.08O2 +85

Zr0.92V0.08O2 +120

NFS dopants used to balance both lattice stability and electronic NFS/hydride interfacial interactions. Sc predicted to enhance LiBH4 dehydrogenation.

Dopants substituted in most favorable position. Increased hydride interactions with increased typical formal dopant valence.

Calculated for favorable discharge to single adsorbed *Li on surface.

15

Thermodynamic Modeling - NFS / Hydride Compatibility

Objective: Use thermodynamic modeling in combination with experimentation to guide design of stable framework compositions.

Predicted reaction from thermodynamics motivates additional modeling and experimental efforts.

0 100 200 300 400 5000.00

0.05

0.10

0.15

0.20

F

C

Temperature

H2(g)

ZrO2 B

LiBH4Li2ZrO3

ZrH2

Km

ol

Examined additional compositions to determine NFS stability and system reversibility with borohydrides:

5 Oxides3 Carbides1 Nitrides

16

Thermodynamic Modeling – Potential Oxide Reduction

Only LiBH4 thermodynamic data initially available in HSC SoftwareOxides may be susceptible to reductionSiO2 was reported as catalyst for LiBH4 dehydrogenation1, but has shown thermodynamic instability

Currently conducting thermodynamic evaluations of Ca(BH4)2 / NFS interactions

Methods to prevent oxide reduction and examine the possibility of boronic acid formation are being explored.

4LiBH4 + 3SiO2 = 2Li2O*SiO2 + Si + 4B + 8H2(g)ΔG298K = -49.286 kJ

Example: Possible hydride reactivity with NFS

1. A. Züttel, S. Rentsch, P. Fischer, P. Wenger, P. Sudan, Ph. Mauron, Ch. Emmenegger, “Hydrogen Storage Properties of LiBH4”, J. Alloys and Compounds 356 (2003)

17

XerogelInitial focus on oxide and carbon materials:

Xerogels (SiO2, Al2O3, ZrO2): 5nm & >300m2/g

Cryogels (Al2O3): 5 - 20nm & >300m2/g

SiO2 Aerogels: 17nm & > 550m2/g

Carbon Aerogel: 10 - 25nm & >600m2/g

Nano-Framework Materials Development

Cryogel

SiO2Aerogel

CarbonAerogel

18

Hydride DevelopmentMetal borohydrides, Alkaline (Ak)-Transition metal (Tm)-B-H, were developed under contract DE-FC36-04GO14012.Multistep reactions significantly lower dehydrogenation onset temperatures and improved kinetics. Only trace B2H6/B3H9detected in the outgas.

400ºC

2E-10

2.5E-10

3E-10

3.5E-10

4E-10

4.5E-10

5E-10

0 100 200 300 400 500 600Temperature (˚C)

H2

ion

Cur

rent

/mg

of A

kBH

4

Ak-Tm-B-H-1

AkBH4

Ak-Tm-B-H-2

Ak-Tm-B-H-4

TGA/MS

For compounds with limited reversibility, the NFS will inhibit irreversible segregation thus improving reversibility.

H2 Desorption

0.0%

4.0%

8.0%

12.0%

0 1 2 3 4 5 6Time (hour)

H2

wt%

Ak-Tm-B-H-2b, 400˚C

Ak-Tm-B-H-4, ˚

Ak-Tm-B-H-1, 500˚C

Ak-Tm-B-H-6, 400˚C

H2 Desorption

19

NaTi(BH4)4∗ligand – Up to 7.3 wt% H2 Endothermic

Down-selected NaTi(BH4)4 based on potential for reversibility, low desorption temperature and possible solution based incorporation.

Wavenumbers (cm-1)

Starting complex

Starting complex (fully charged)

60ºC Discharge, 20ºC Recharge100ºC Discharge

B-H peak range (1900 – 2500 cm-1)

*

*

*Starting complex (fully charged)

60ºC Discharge, 20ºC Recharge100ºC Discharge

B-H peak range (1900 – 2500 cm-1)

*

*

*

Abs

orba

nce

Diffuse Reflectance Infrared Fourier Transform Spectra

Starting complex (fully charged)60ºC Discharged, 20ºC recharged100ºC Discharged

20

11B MAS NMR and XRD show re-formation of Ca(BH4)2 after re-hydriding at 330°C and 100 bar

Bowman, Hwang,

Kim, Reiter, ZanRönnebro

Ca(BH4)2

Ca(BH4)2 as made

Desorbed at 450°C

Desorption at 320°C

Re-hydrided at 330°C

ppm

+H2 CaB6

a-B

* * likely B12H12

Hydride Material – Ca(BH4)2

See more details on Ca(BH4)2 in ST36, Ronnebro, Majzoub

CaB6 + 2CaH2 + 10 H2 → 3Ca(BH4)2 @700bar, 400°C, 48hours

21

Sc

B

N

H

C

Solvated Synthesis of Metal Borohydridessolv

MXn = ScCl3, TiCl4, ZrBr4,CaI2;; solv = THF, DME or MeOH

MXn + n NaBH4 M(BH4)n(solv)x

THF

Solution deposition of MBH onto silica substrate illustrates that deposition by this method will be possible. However, contamination with NaX by product appears to be

problematic. Schemes to further purify the MBH from the NaX are underway.

Ca

BOH

C

Ca(BH4)2(DME)2Sc(BH4)3(py)3

CaI2 +2NaBH4 Ca(BH4)2(solv)xsolv

• single crystal X-ray diffraction confirms Ca(BH4)2(DME)2 was synthesized. Used for deposition studies, with THF adduct

looking the most promising

DME

MeOH

22

AlBr3 + 3 NaBH4 Al(BH4)3 + 3 NaBr

2 Al(BH3)3 + 3 Ca(NR2)2 3 Ca(BH4)2 + 2 Al(NR2)3

2 Al(BH3)3 + 3 MgR2 3 Mg(BH4)2 + 2 AlR3

tol

tol

tol

Zanella et al. Inorganic Chemistry 2007 (ASAP).

Solvent-free Synthesis of Metal Borohydrides

Identification of final materials not consistent with known PXRD patterns of Mg(BH4)2or Ca(BH4)2. Contamination by Br indicates additional work to purify final product

necessary. Sequential stepwise characterization of intermediates underway.

AlBr3 + 3 LiBH4 + 2/3 Ca(NR2)2 in tol/reflux

20 25 30 352-Theta(°)

x10^3

5.0

10.0

15.0

Inte

nsity

(Cou

nts)

[MgBH4.xy]

2 Al(BH3)3 + 3 MgR2 in tol

23

Hydride / Framework Compatibility Screening

Compatibility screening ⇒ DSC

-1.5

-1.0

-0.5

0.0

0.5

Rev

Hea

t Flo

w (W

/g)

0 100 200 300 400 500Temperature (°C)

2008-014DSC.003 LiBH4_ZrO2 50-502008-014DSC.001 LiBH4

Exo Up Universal V4.5A TA Ins

ZrO2 + LiBH4

LiBH4

LiBH4

Xerogel

Ball Mill

Press Pellet Melt LiBH4(290°C, 190Bar)

LiBH4 use for initial compatibility screening because of high reactivity, availability, and existing thermodynamic data.

DSC screening to determine reaction compatability of framework with hydride.

24

Compatibility Screening – SiO2 + LiBH4

Possible non-reversibility suggested via thermodynamic assessment. DSC shows split peak at 280-300°C suggesting reaction event.

Potential Reaction Pathway4LiBH4 + 3SiO2 → 2Li2O*SiO2 + 4B + Si + 8H2(g)

ΔG298K = - 49.286 KJ

-10

-8

-6

-4

-2

0

2

Hea

t Flo

w (W

/g)

0 50 100 150 200 250 300 350 400 450Temperature (°C)

2008-045DSC.002 Silica sol-gel (Xerogel)+LiBH42008-045DSC.007 LiBH4

Exo Up Universal V4.5A TA In

Hea

t Flo

w (W

/g)

Temperature (C)

SiO2 + LiBH4

LiBH4

25

Compatibility Screening – SiO2 + LiBH4

After discharge LiBH4 is decomposed but unidentified peaks exist.Conclusion: SiO2 is reactive with hydride and unsuitable for NFS.

10 15 20 25 30 35 40 450

500

1000

1500

50 55 60 65 70 75 80 85Two-Theta (deg)

0

500

1000

1500

[ ] g g y pp97-000-0853> Li 2SiO 3 - Lithium Catena-silicate

97-001-5414> Li 2(Si 2O5) - Dilithium Phyllo-disilicate97-006-1753> LiD - Lithium Deuteride

97-009-0849> Li 4(B4Si8O24) - Tetralithium Tetraborooctasilicate

Inte

nsity

(Cou

nts)

theta cal

Bkd, Ka2, DSeal Subtd

XRD data after 290°C, 1 bar, 6 hours

No LiBH4 (RT, HT)Unidentified Peaks

Li2SiO3Li2(Si2O5)LiHLi4(B4Si8O24)

26

Compatibility Screening – ZrO2 + LiBH4

DSC of ZrO2 + LiBH4 is similar to LiBH4, suggesting that ZrO2 is stable and non-reactive in the presence of this strongly reducing hydride.

-10

-8

-6

-4

-2

0

2H

eat F

low

(W/g

)

0 100 200 300 400 500 600Temperature (°C)

2008-045DSC.007 LiBH42008-045DSC.005 Zirconia (HS) (Xerogel)+LiBH4

Exo Up Universal V4.5A TA In

Hea

t Flo

w (W

/g)

Temperature (C)

ZrO2 + LiBH4

LiBH4

27

Solution coating of ceramic (SiO2 initially) substrate with metal borohydride.

Solvent selection is critical:Stability in synthesis and depositionBinding to surface is appropriate

Analyze air-sensitive materials.

base

O-ring

Berylliumdome

Specimenplate assemblyVacuum port

(sealed)

Specimenplate

BeD-XRD

Solvated Hydride Incorporation

Volatile enough to be removed prior to BH4 decomposition

Solution Deposition

28

Project Summary

ModelingSimulations show interfacial NFS interactions can alter stability of hydride and discharged products.Dopants balance both NFS lattice stability and electronic NFS/hydride interfacial interactions.

Framework and HydrideInitial nano-framework structures have been synthesized (ZrO2, Al2O3, SiO2, TiO2, Carbon).Multiple suitable oxide candidates have been identified. ZrO2 selected because of low reducibility.Compatibility screening reactions with LiBH4 have been performed.UTRC / Albemarle focus on ligand stabilized: NaTi(BH4)4∗ligand Sandia focus on stable borohydride: Ca(BH4)2

Improve reversibility of high capacity hydride candidates by developing advanced NFS chemistries through combined modeling and experimentation.

29

Atomistic ModelingSimulate Ca(BH4)2 and NaTi(BH4)4∗ligand hydride interactions with ZrO2 NFS. Virtually tune doped NFS to balance hydride stability and dehydrogenation. Virtually develop doped, functionalized, catalyzed NFS to enhance reversibility.

Framework and HydrideEvaluate initial oxide (ZrO2) aerogel.Continue to assess oxide, modified carbon and other alternative framework materials. Examine support interactions with selected NaTi(BH4)4∗ligand and Ca(BH4)2.Evaluate doped, heterogeneously catalyzed and functionalized nano-frameworks.

Future Plans

Ca(BH4)2 on ZrO2