Combinatorial approach to materials discovery

No impurity Ti (3 Å) Ti (6 Å) Ti (9 Å) Cu (3 Å) Cu (6Å) Cu (9 Å)

5 Å

10 Å

15 Å

20 Å

25 Å

35Å

45 Å

55 Å

ti(Å)

t s(Å)

Permanent magnet library

Ferroelectric library

Superconductor library

Ichiro TakeuchiUniversity of Maryland

Combinatorial Approach to Materials Discovery

• Introduction to the combinatorial approach:brief history, tools and strategies

• Integrated materials discovery engine

• Recent examples: combinatorial search of rare-earth-free permanent magnets; superconductors

Outline

University of MarylandTieren GaoSean FacklerKui JinR. Greene

SLACA. Mehta

US DOE, ONR, AFOSR

Support

Acknowledgement

Duke UniversityS. Curtarolo

Ames LabM. J. Kramer

NISTA. G. KusneM. Green

Chemical &EngineeringNews, August 2001

Combinatorial Libraries of Inorganic Materials

Luminescent materials libraries,Science 279, 1712 (1998)

Semiconductor gas sensor library, “electronic nose”,Appl. Phys. Lett. 83, 1255 (2003)

Magnetic shape memory alloy library,Nature Materials 2, 180 (2003)

Fabrication of libraries and spreadsCombinatorial PLD systems – metal oxides Combinatorial UHV sputtering system – metallic alloysCombinatorial multigun e-beam evaporator system – metal Combinatorial laser MBE – metal oxides

Rapid characterization toolsScanning SQUID microscopes – magnetic propertiesScanning microwave microscopes – resistive, magnetic, dielectricScanning X-ray microdiffractometerMagneto-optical Kerr effect (MOKE) system – magnetic propertiesScanning 4-point probe station – transportNovel device libraries incorporating MEMS, etc.

Major Facilities for CombinatorialMaterials Research at Maryland

Focus: Functional Thin Film Materials

Correlation between materials complexityand physical properties

HgNb3Ge

La2CuO4

YBa2Cu3O7

HgBa2CaCu3O7

Crit

ical

Tem

p. (K

)

306090

120150180210240

1Number of Elements2 3 4 5 6 7

1IA 2IIA 3IIIB 4IVB 5VB 6VIB 7VIIB 8VIII 9VIII 10VIII 11IB 12IIB 13IIIA 14IVA 15VA 16VIA 17VIIA 180H1 He2Li3

Be4

B5

C6

N7

O8

F9

Ne10

Na11

Mg12

Al13

Si14

P15

S16

Cl17

Ar18

K19

Ca20

Sc21

Ti22

V23

Cr24

Mn25

Fe26

Co27

Ni28

Cu29

Zn30

Ga31

Ge32

As33

Se34

Br35

Kr36

Rb37

Sr38

Y39

Zr40

Nb41

Mo42

Tc43

Ru44

Rh45

Pd46

Ag47

Cd48

In49

Sn50

Sb51

Te52

I53

Xe54

Cs55

Ba56

La57

Hf72

Ta73

W74

Re75

Os76

Ir77

Pt78

Au79

Hg80

Tl81

Pb82

Bi83

Po84

At85

Rn86

Fr87

Ra88

Ac89

Unq104

Unp105

Unh106

Uns107

Uno108

Une109

Uun110

Ce58

Pr59

Nd60

Pm61

Sm62

Eu63

Gd64

Tb65

Dy66

Ho67

Er68

Tm69

Yb70

Lu71

Th90

Pa91

U92

Np93

Pu94

Am95

Cm96

Bk97

Cf98

Es99

Fm100

Md101

No102

Lr103

Binary compounds have the form AB. (e. g., MgF, SiC, ZnO,…..)60 x 59 x different combinations: most are known.

Ternary compounds have the form ABC. (BaTiO, NiMnGa, HgCdTe,…)60 x 59 x 58 x different combinations: ~3 % of all possible known

Quaternary compounds have the form ABCD. (YBaCuO, AlNiCOFe,…)60 x 59 x 58 x 57 x different combinations: 0.01% of all possible known

Beyond??

How many different compounds are there?

Take 60 “useful” elements.

There are about~100,000 known inorganic compounds.

Quaternary Masks

A B

C ED

Quaternary Masking

Ba

Quaternary Masking: 1st mask, 1st position

Ca

Quaternary Masking: 1st mask, 2nd position

Sr

Quaternary Masking: 1st mask, 3rd position

Pb

Quaternary Masking: 1st mask, 4th position

BaPb

Sr Ca

Ba

Quaternary Masking: after 1st mask

BaPb

Sr Ca

BaZr

Zr

Zr

Zr

Quaternary Masking: 2nd mask, 1st position

BaPb

Sr Ca

BaTa

Ta

Ta

Ta

Quaternary Masking: 2nd mask, 2nd position

BaPb

Sr Ca

Ba

Nb

Nb Nb

Nb

Quaternary Masking: 2nd mask, 3rd position

BaPb

Sr Ca

BaTi

Ti Ti

Ti

Quaternary Masking: 2nd mask, 4th position

BaZrO3

CaNb2O6

CaTiO3

BaNb2O6

BaTiO3

CaTa2O6

CaZrO3

BaTa2O6PbTa2O6

PbZrO3PbTiO3

PbNb2O6

SrTiO3

SrNb2O6 SrTa2O6

SrZrO3

A B

C ED

# depositions: 4 x n# combinations: 4n

5 masks: 4 x 5 = 20 depo’s45 = 1024 samples

(Right) Luminescent image of the same library after thermally processed under UV excitation.

Science 279, 1712 (1998)

Library of luminescent materials made w/ quaternary masking

Various combinatorial experimental designs:

discrete libraries vs composition spreadsComposition A B

B

A

C

• Composition spreads allow continuous mapping of physical properties and phase boundaries

• Run to run variation in ordinary experiments is removed

Library synthesis under epitaxial growth conditions

F. Tsui (UNC)H. Koinuma, M. Lippmaa, T. Chikyow (COMET)

Materials are of the same quality as single composition depositions

Fabrication of epitaxial composition spread of oxides via laser MBE

Takeuchi et al., Applied Physics Letters 79, 4411 (2001)

Scanning Microwave Microscope:a rapid characterization tool

Originally developed for rapid screening of libraries of superconductors, dielectric materials, etc.

SampleTip

Coaxial ¼ resonator

x-y-z stage Motioncontroller

Computer

f0

Q

Microwave source

Review Article: Gao, et al., Measurement Science and Technology 16, 248 (2005)

-250

-200

-150

-100

-50

0

50

100

150

200

880 890 900 910 920 930 940 950 960 970

Magnetic field (Oe)

FMR

sign

al (a

rb. u

nit)

2.45 GHz

SrTiO3

BaTiO3

CaTiO3

500400

100

300

0

200

r

Different physical properties can be mapped using an existing microwave microscope

Dielectric constant mappingof a (Ba,Sr,Ca)TiO3 pseudo-ternary library at 1 GHz

Appl. Phys. Lett. 74, 1165 (1999)

Ferromagnetic resonance (FMR)signal taken at a spot

Dielectric property

Magnetic property(spin resonance)

composition plot

Mode/materials[reference]

Physical parameter/phenomenon

Spatialresolution

Dielectric [12-14] Complex dielectric constant

100 nm

Metal [13] Impedance/resistivity 100 nm

Non-lineardielectric [15,18]

Non-linear dielectric constant

1 nm

FMR Ferromagneticresonance

100 nm*

STM-ESR Electronspin resonance

Atomicresolution

Capabilities of Multiscale Microwave Microscope

Mapping of various physical properties can be obtained at macroscopic scale (~ 1 cm) down to the listed spatial resolution

Atomic resolution microwave microscope/STM

Tunneling current resonant f

HOPG

Au(111)Atomic resolution images obtained with STM disabled –

surface approachedusing microwave feedback

DC FieldMagnet

STM Tip Built‐In

Microwave Resonator(2.5 GHz)

(Lee et al., APL 97, 183111 (2010))

Composition Spreads of Ternary Metallic Alloy Systems

Co-sputtering scheme Ni

Mn

Al

3” spread wafer

Ni Al

Mn

Phase diagram

Composition is mapped using an electron probe (WDS)

RT Scanning SQUID microscope(Magma, Neocera)

SQUID assembly inside vacuum

leveling probe and scanning stage

Room temperature samples are measured

z-SQUID is used tomeasure Bz distribution

Tip-sample distance is typically 100~200 microns

15 20 25 30 35 40 45 50

60

50

40

30

20

col

ro

w

-2.50e+007 0.00e+000 2.50e+007

rho1_25_x

100-150 emu/cc

50-70 emu/cc30-40 emu/cc

10-20 emu/cc

Scanning SQUID image of a Ni-Mn-Ga spread wafer (room temperature)

0 13 25 38 50 63 75

80

60

40

20

0

col

ro

w

-2.50e+007 0.00e+000 2.50e+007

rho1_25

Mn rich

Ni rich Ni2Ga3 rich

Combinatorial Search of Ferromagnetic Materials

GaNi 0 1 2 3 4 5 6 7 8 9 10

Mn

50 100 150 200 250

M (emu/cc)

Ni2Ga3

Nature Materials 2, 180 (2003)

Rapid detection of shape memory alloy compositions by visual inspection

Composition spread deposited onmicromachined cantilever array

Film thickness ~0.5 m

Detection of martensitic phase transformation

Functional phase diagram of Ni-Mn-Ga

20 40 80

20

40

60

80

60

80

20

Mn

40

20 40 80

20

40

60

80

Ni

60

80

20

40

Ni2Ga3 Ga

Increasing transitiontemperature

Ferromagnetic regions

Most strongly magnetic

Martensites

Nature Materials 2, 180 (2003)

Integration of theory and high-throughput experiments

Step 2 Step 3Step 1

Integrated materials discovery engine

ExperimentalTrack

TheoreticalTrack

Step 2 Step 3Step 1

ExperimentalTrack

TheoreticalTrack

Advantages of this approach:

Predictions are sometimes “off” by stoichiometric variations.


Step 2 Step 3Step 1

ExperimentalTrack

TheoreticalTrack

Advantages of this approach:

Predictions are sometimes “off” by stoichiometric variations.

Large number of data points in combinatorial experiments suitable for building models.


Consortium of QM calculations

41api

http://aflowlib.org/Curtarolo,et al (Duke)

Step 2 Step 3Step 1

ExperimentalTrack

TheoreticalTrack

Example:Rare‐earth‐free permanent magnetsAPL 102, 022419 (2013);Scientific Reports 4, 6367 (2014)


Rare-earth (Nd, Dy, Sm, etc.)-free magnets are needed due to their fluctuating prices

Search for new permanent magnet materialsw/o rare-earth elements

The prices of many rare‐earth metals have increased by more than 10 fold in the past few years

Permanent magnets for: direct drive wind turbines

Current magnets: Nd‐Fe‐B, Sm‐Co

Advanced electric drive motors

History of development of permanent magnets

Best magnets contain rare-earth elements: Nd, Dy, Sm

Nd-Fe-B

Sm-Co

Year

How to design new permanent magnets• Need high energy product (BH)max

• Need high magnetization M ‐ need Fe and/or Co• Need high uniaxial anisotropy K – coercive field Hc

‐ e.g. NdFeB: 5 x 106 J/m3; SmCo5 2 x 107 J/m3

Paths to intrinsic anisotropy (without rare‐earth):• Modify FeCo: cubic to tetragonal, electronic structure• MnBi/MnAlX• Atomically ordered phase of FeNi

Modify/distort FeCo: add 3rd element X (d elements):spin‐orbit coupling; high anisotropy high coercive field‐> Make Fe‐Co‐X composition spreads

1852 C

2150 C 3412 C

2617 C

Melting points Melting points

1536 C 1495 C

Identification of composition with enhanced coercive field: Fe‐Co‐Mo Scientific Reports 4, 6367 (2014)

Composition with enhanced coercive field was identified

Magnetic hysteresis loop mapping of Fe‐Co‐Mo spread

Hc ~ 1.2 KOe(K ~ 30 eV/atom)

Hc mapping

Grouping of structures based on synchrotron diffraction

FeMo

Co

Calculated structures

Hc ~ 1.2 KOe(K ~ 30 eV/atom)

Hc mapping

Grouping of structures based on synchrotron diffraction

FeMo

Co

Identification of composition with enhanced coercive field: Fe‐Co‐MoIdentification of composition with enhanced coercive field: Fe‐Co‐Mo Scientific Reports 4, 6367 (2014)

Identified Fe8CoMo has a tetragonal structure; genetic algorithm and DFT give K values in agreement with experiment

Search for rare‐earth free permanent magnets

• High energy product (BH)max

‐ High magnetization M (Fe and/or Co)‐ High uniaxial anisotropy K – coercive field Hc

(NdFeB: 5 x 106 J/m3; SmCo5 2 x 107 J/m3)

Paths to intrinsic anisotropy (without rare‐earth):• Modify FeCo: cubic to tetragonal, electronic structure• MnBi/FeCo• Atomically ordered phase of FeNi

Optimizing MnBi/CoFe exchange coupled bilayers: soft layer thickness gradient on MnBi

BiMn annealing

MnBiCo

MnBiDeposition

of Co

Soft layer gradient 10 ‐ 0 nmglass or Si sub

MnBi thickness20 nm

(BH)max doubles from 12 to 25 MGOe by adding 3 nm of Co

(MG

Oe) 25 MGOe

Optimizing MnBi/CoFe exchange coupled bilayers: soft layer thickness gradient on MnBi

History of development of permanent magnets

Nd-Fe-B

Sm-Co

YearThis work: MnBi thin film/multilayers

Step 2 Step 3Step 1

ExperimentalTrack

TheoreticalTrack

Examples: Combinatrorial search of superconductivity in Fe‐BAPL Materials 1, 042101 (2013)


Targeting superconductors predicted by theory

Prediction:FeB4 is a superconductor with Tc ~ 15‐20 K

Fe-B phase diagram (1994)

FeB2FeB4

Not much is known in this region

Exploration of new superconductors: Fe-B composition spread

3” wafer

Fe rich B rich

16 spot 4‐terminal pogo pin arrays:

Cut wafers into 1 cm2 pieces and measure 16 spots at once

Color change tracks:composition change, crystallinity change, and metal to insulator transition

Fe-B composition spread: Fe-rich side, 16 spots on one 1 cm2 chip

Ch 113” wafer

more Bmore Fe

temperature

resistance

4.2 K 300 K

All metallic

Semiconducting to insulating

more Bmore Fe

temperature

resistance

4.2 K 300 K

Fe-B composition spread: B-rich side, 16 spots on one 1 cm2 chip

Ch 11Ch 3

Ch 13

Middle region:FeB2 – FeB4

more Bmore Fe

temperature

resistance

4.2 K 300 K

Fe-B composition spread: FeBx(x =2-4), 16 spots on one 1 cm2 chip

FeBx: we have found the superconductor

Susceptibilityshows diamagnetism

Bc2(T)= Bc2(0)[1-(T/Tc)2]/ [1+(T/Tc)2]

gives Bc2(0) = 2 T

-> Type II BCS superconductor

Partial R drop

~ 10 K?

Superconducting phase was detected in 2 spread wafers

Combinatorial Time Lapse:a day in the life of

a combinatorial materials scientist

Sean Fackler

Summary

Combinatorial experiments can be used to carry out effective mapping of large compositional phase spaces previously unexplored

This strategy has been incorporated into many technological areas; We have used this strategy to discover many new functional materials

Combinatorial strategy is the natural counterpart to the concerted theoretical efforts taking place within the Materials Genome Initiative

Review article: Green, et al.J. Appl. Phys. 113, 231101 (2013)