Metallic Glasses Seminar - A. L. Greer

7/24/2019 Metallic Glasses Seminar - A. L. Greer

http://slidepdf.com/reader/full/metallic-glasses-seminar-a-l-greer 1/61

Amorphous Bulk Metals ―

Metallic Glasses

A. L. GreerDept. of Materials Science & Metallurgy

University of Cambridge

Materials on the HorizonCambridge, 9 December 2008



Crystal

• regular atomicarrangement

• slip planes forplastic deformation



Glass

• no periodicity, but

• density ~ same ascrystal

• local configurations~ same as crystal

• plastic flow is difficult



How to make a glass?

• a glass forms if crystallization is avoided on cooling• the density of the glass depends on cooling rate



The Glassy State

— is found for all classes of material:

• oxide (e.g. SiO2)• ionic (e.g. ZnF2)• polymeric• metallic• carbohydrates



Metallic Glasses

• metals and alloys are naturally crystalline

• pure metals cannot form glasses — their simplestructure crystallizes too easily on cooling the liquid

• alloying can stabilize the liquid, and aids glassformation (“confusion principle”)

• for a binary alloy such as Fe80

B20

(atomic %), thecritical cooling rate for glass formation is

105 to 106 K s –1







Bulk Metallic Glasses

• multicomponent compositions aid glass formation

• the critical cooling rate is much lower (~1 K s –1)

• glasses can be formed in bulk



Metallic Glasses

• are now understood to be true glasses

Outline

• new insights on atomic-level structure

• metallic glasses as structural materials

• current progress on understanding plastic flow

• improving mechanical properties, and new horizons



from The Times Higher Education Suppl . 3 Feb. 2006

Structure

John Desmond Bernal1901-1971

The dense random packing model forthe structure of liquid metals.



Short-to-medium-range order in metallic glasses:― solute-centred clusters

H.W. Sheng, W.K. Luo, F.M. Alamgir, J.M. Bai & E. Ma, Nature 439 (2006) 419.



Interpenetrating clusters in the efficient cluster packing model ofMiracle et al.

D B. Miracle, Acta Mater . 54 (2006) 4317.



Dynamics in metallic glasses― atoms in icosahedral clusters are the least mobile

YQ Cheng et al ., Appl. Phys. Lett . (2008) 93, 111913



YQ Cheng et al., Phys. Rev. B (2008) 78, 014207

Degree of icosahedral order increases on cooling, and is also differentfrom composition to composition, correlating with glass-forming ability.



M.F. Ashby & A.L. Greer: Scripta Materialia 54 (2006) 321.(in Viewpoint Set on Mechanical Behavior of Metallic Glasses , edited by T.C. Hufnagel)

Elastic limit σ y plotted against density ρ for 1507 metals, alloys,metal-matrix composites and metallic glasses. The contoursshow the specific strength σ y / ρ .

Metallic glasses for structural applications



Ce70Al10Cu20 — T g = 338 K, T x = 390 K

B. Zhang, D.Q. Zhao, M.X. Pan, W.H. Wang & A.L. Greer:“Amorphous metallic plastic”, Phys. Rev. Lett. 94 (2005) 205502.





J Schroers et al ., Scripta Mater . (2007) 57, 341





Microformability of BMGs

• of interest for micro-& nano-imprinting ofsurfaces

AFM and SEMimages of a patterned(100) Si die and a Pt-based BMG imprintedwith the die (10 MPa,550 K, 300 s)

Y. Saotome et al. “The micro-nanoformability of Pt-based metallic glass and thenanoforming of three-dimensional structures”, Intermetallics 10 (2005) 1241.



J. Schroers: “The superplastic

forming of bulk metallic glasses”,JOM 57(5) (2005) 35.




Fracture toughness and elastic limit for metals, alloys, ceramic,glasses, polymers and metallic glasses. The contours show theprocess-zone size d in mm.



The world’s smallest motor



from Materials Selection in Mechanical Design (2nd ed.)

M. F. Ashby, Butterworth-Heinemann, 1999

metallic glasses

— compared to metalsand alloys in general,

the glasses have highstrength σ and lowstiffness E , that is,unusually high elasticstrain —

σσσσ/E



J.H. Tregilgas, “Amorphous titaniumaluminide hinge”Adv. Mater. Proc . 162 (Oct. 2004) 40.

MEMS Applications of Metallic Glasses

The Texas Instruments Digital

Light Processor (DLP) data

projector technology is based on

mirrors supported by amorphous

Ti-Al hinges. DLP devices with

>1.3 x 106

addressable mirrorsare in production, and the hinges

still show no fatigue failures after

1012 cycles.



Pressure Sensors

Diaphragms

Annual production now nearly 50 million units





metallic glasses

materials for elasticenergy storage —

want to maximize

σσσσ2/E



Strain→

S t r e s s →

Within the elastic (reversible) regime ―

σ σσ σ y

E

area = σ σσ σ 2/2E = elastic energystored per unit volume



Strain→

S t r e s s →

to increase the elastic stored energy―

E

increase the yield stress, σ σσ σ y



Strain→

S t r e s s →

E

decrease the Young modulus, E

to increase the elastic stored energy―



T Fukushige & S Hata, J. Microelectro. Syst . (2005) 14, 243

MEMS Applications

A conical spring microactuator

with a long stroke of 200 mmnormal to the substrate. The

spring is a 7.6 µm thick film of

Pd76Cu7Si17 metallic glass.



Golf clubs …. and tennis-racket frames, baseball bats, skis …





metallic glasses

materials for elasticenergy storage —

want to maximize

σσσσ2/E




Fracture toughness and Young’s modulus for metals, alloys,ceramic, glasses, polymers and metallic glasses. The contoursshow the toughness G c in kJ m –2.




Fracture toughness and Young’s modulus for metals, alloys,ceramic, glasses, polymers and metallic glasses. The contoursshow the toughness G c in kJ m –2.



• the plastic flow stress in shear is proportional to the elastic shearmodulus — thus the shear modulus is a measure of the difficulty ofplastic flow

• similarly the bulk modulus is a measure of the difficulty of cracking

• thus high values of the shear-to-bulk modulus ratio µ / B should favourbrittleness and vice versa

• proposed by Pugh in 1954, and developed by others —

S.F. Pugh, Philos. Mag . 45 823 (1954).

A. Kelly, W.R. Tyson and A.H. Cottrell, Philos . Mag . 15 567 (1967).

J.R. Rice and R. Thomson, Philos. Mag . 29 73 (1974).

A.H. Cottrell, in Advances in Physical Metallurgy , edited by J.A. Charles and

G.C. Smith (Institute of Metals, London, 1990), pp. 181–187.

Metals: Plasticity or Brittleness?



Compilation of all relevant and available data on as-cast(unannealed) metallic glasses (mostly, but not all BMGs)

J.J. Lewandowski, W.H. Wang & A.L. Greer, “Intrinsic plasticity or brittleness ofmetallic glasses”, Philos. Mag. Lett. 85 (2005) 77.

D f ti f M t lli Gl



F. Spaepen: “A microscopic

mechanism for steady state

inhomogeneous flow in metallic

glasses”, Acta Metall . 25 (1977)

407.

Deformation of Metallic Glasses

Ambient temperature / high stress-- flow localization in shear bands

High temperature / low stress

-- homogeneous viscous flow



Plastic deformation of a thin plate of a thin plate of Pd77.5Cu6Si16.5

glass in tension. Shear bands are consistent with work-softening.

H. Kimura, PhD Thesis (1978) Tohoku Univ.



L.A. Davis & S. Kavesh, J. Mater. Sci . 10 (1975) 453.

Fracture surface of Pd77.5Cu6Si16.5 — characteristic vein pattern,

formed by Saffman-Taylor fingering in a liquid-like layer.

The thickness of the liquid-like layer must be at least severalµm.

2500



0

500

1000

1500

2000

2500

0 0.01 0.02 0.03 0.04 0.05 0.06

True Strain

T r u e S t r e s s [ M P a ]

Vitreloy

0.1 MPa Hydrostatic Pressure

Yield/Fracture Strength = 1986 MPa

εεεεf = 0%

J.J. Lewandowski

At ambient temperature, metallic glasses in tension can appear

macroscopically brittle, despite extensive local deformation in theshear bands.



Thickness of shear bands?

TEM studies consistently suggest a shear-band thickness of ~10 nm



M. Chen, A. Inoue, W. Zhang & T. Sakurai: “Extraordinary plasticity of ductile bulk

metallic glasses”, Phys. Rev. Lett. 96 (2006) 245502.

Molecular dynamics simulations



N.P. Bailey, J. Schiøtz &K.W. Jacobsen, Phys. Rev. B 73 (2006) 064108.

Molecular-dynamics simulations

— also show a shear-band thickness of ~10 nm

Q.-K. Li & M. Li, Appl. Phys. Lett . 88(2006) 241903.



Operation of Shear Bands

TEM shows that the shear is sharply localized — — thickness of shear band = 10 to 20 nm

The origins of localization remain controversial — structural change, ortemperature rise?

Measurements of temperature rise 0.4 K to 1000 K

Predictions of temperature rise 40 K to 1000 K

Th f ibl ti th d



The fusible-coating method

The operation of a shear band in a BMG generates a

hot plane and melts the coating (of tin). The total workdone by shear is proportional to the offset δ .

L l lti f ti ti t h b d i th BMG



Local melting of a tin coating at shear bands in other BMGs

(Cu50Zr50)92Al8 La55Al25Cu10Ni5Co5



B. Yang, P.K. Liaw, G. Wang, M. Morrison, C.T. Liu, R.A. Buchanan & Y. Yokoyama:

“In-situ thermographic observation of mechanical damage in bulk-metallic glassesduring fatigue and tensile experiments”, Intermetallics 12 (2004) 1265.

Average measuredtemperature rise inshear bands = 0.4 K(for observed width of0.15 mm)



Resolution of the fusible-coating method

• temporal resolution ≈ thermal diffusion time for coating thicknessincluding latent heat of melting, the resolution ≈ 30 ps

• spatial resolution ≈ scale of islands ≈ 100 nm

In contrast for direct infrared measurements the best reported resolutioncombinations are —

• for imaging 1.4 ms ~11 µµµµm

• for single detector ~10 µµµµs 100 µµµµm

1500



0

500

1000

1500

-2 -1 0 1 2

T e m p e r a t u

r e R i s e ,

∆ ∆∆ ∆

T

( K )

Distance, x ( µµµµm)

H = 0.4 kJ m-2

H = 2.2 kJ m-2

7

50

1

0.2

10

50

167

1000∆∆∆∆T = 207 K

Minimum observedmelting half-width= 200 nm

Observed melting

half-width = 1 µµµµm

H = 0.4 kJ m

–2 H

= 2.2 kJ m

–2

Distance, x (µµµµm)

Half-profiles oftemperature at a typical

shear band evolving overtime (in nanoseconds)

― calculated fromindependently measuredthermal diffusivity

Profile when tin coating is

melted to maximum width



How fast are shear bands?

Estimates of the time of shear vary enormously―

― as short as 0.2 ns(estimated from the offset and the speed of sound)

― as long as 100 ms

(estimated from serrations on stress-strain curves)



DB Miracle, A Concustell, Y Zhang, AR Yavari & AL Greer:Thermal profiles around shear bands in metallic glasses, submitted

Mode III Shear

velocity of propagation V p

(velocity of the shear-band front)≤

90% of transverse sound speed

shear velocity V s(does work and sets local shear

time) V s

= 4γ V p

γ = 0.0267 (Johnson & Samwer)V s = 0.107 V p

Velocities less than ~50% of the maximumwould be insufficient to melt tin coatings

Length scales and times associated with shear-band heating



[measured (from TEM) thickness of shear band = 10 to 20 nm]

1000thermal diffusion length at end of shear (nm)

1990

upper-bound estimate of temperature rise atshear-band centre (K)

20.8lower-bound estimate of shear duration, δt (ns)

[assumed shear velocity = 206 m s –1]

4.3shear offset calculated from H and τ y (µµµµm)

167time at melting limit (ns)

2.2calculated heat content of shear band, H (J m –2)

1 µµµµmobserved half-width of melted zone

Extreme conditions in shear bands



Extreme conditions in shear bands

heating rate ≈ 1012 K s –1

shear rate ≈ 1010 s –1

cooling rate ≈ 109 K s –1

How to improve the mechanical properties of BMGs?



How to improve the mechanical properties of BMGs?

One approach is to make a composite —

• very encouraging results by introducing a ductile crystalline phase into

the glassy matrix

• increases ductility• lowers strength (but better compromise of properties)

Another approach is to control the operation of shear bands within

the glassy phase —

• aim to deflect shear bands

• prevent failure by one dominant band

• proliferation of bands makes deformation more diffuse and increasesenergy absorption

• also reduces shear offset at each shear band → reduces cracking

• possibility of work-hardening?



How to deflect shear bands or force their proliferation?

• any inhomogeneity in the amorphous phase

• embedded particles

• crystallites (produced by annealing)

• globules of second glassy phase (after phase separation)

• voids (bubbles / porosity)

Partially crystallized glasses



y y g

— can have crystals smaller or larger than the shear-band thickness

When the crystallites are very small, shear-band operation continueslargely unchanged.

Effects of dispersed nanocrystals



A. Inoue, W. Zhang, T. Tsurui, A.R. Yavari & A.L. Greer, “Unusual room-

temperature compressive plasticity in nanocrystal-toughened bulk copper-zirconium glass”, Philos. Mag. Lett . 85 (2005) 221.

Glassy Cu-Zr normally

shows a compressive

failure strain of ~1.5%.

As-cast Cu50Zr50, 1 mm

rod, with 10-15 %

volume fraction of 5-10

nm shows much better

plasticity.

True stress / true strain curveTest stopped at 52% strain

Record-breaking toughness in glassy-crystalline composites with



DC Hofmann et al., Nature (2008) 451, 1085

optimised microstructural scale ―



Metallic Glasses

• new insights on atomic-level structure

• metallic glasses as structural materials

• current progress on understanding plastic flow

• improving mechanical properties, and new horizons

Documents

Metallic Glasses Seminar - A. L. Greer