MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Liquid Metal Embrittlement

Dominique Gorse (presented by Vassilis Pontikis)

CEA-IRAMIS / CNRS, Laboratoire des Solides Irradiés91191 Gif-sur-Yvette, FRANCE

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Outline• Environment & Mechanical behavior• What is LME ?• Surfaces, Liquids & Wetting• Experimental facts• Empirical criteria• Modeling• Conclusive remarks

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Surfaces and mechanical behavior

Surfaces control the nucleation of cracks and dislocations and the annihilation of dislocations as well. However, bulk phenomena are predominent in plasticity.

“If dislocations are nucleated principally at the surface, we may expect that irradiation with alpha particles, which penetrate only a short distance, will affect the mechanical properties. The evidence is contradictory”

F. R. N. Nabbarro in “Theory of Crystal Dislocations ”, Clarendon, 1967 P. 281

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Joffe effect, irradiated KCl, Rebinder et al., 1944

air

60 s H2O, dried, air

H2O

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

QuickTime™ et undécompresseur codec YUV420

sont requis pour visionner cette image.

Steel 316L + Hg, Medina-Almazan et al. (2005)

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

good wetting <90°

bad wetting >90°

LV

SV SL L

V

S

GB /2

SL

SL

GB

S1L

S2L

S1P + S2P = GB

2 SP cos/2 = GB

(SV, LV, SL)mechanical equilibriumchemical equilibrium

SV - SL = LV cos(Young)

Grain boundaries

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Liquid Metal Embrittlement: what is it ?

Ensemble of phenomena, including wetting, whereby materials suffer a marked decrease in their ability to deform (loss of ductility) or in their ability to absorb energy during fracture (loss of toughness), with little change in other mechanical properties.

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

• Cd: ductile in inert environment is brittle in presence of Ga at room T.• Zn: brittle at ≈150 °C in presence of Hg or Ga

(Rozhanski et al.,1957, Goryunov et al. 1978).• Al: can be cleaved in presence of Bi (Lynch, 1985).

LME: Single crystals or polycrystals ?

Al-T4/2024 Hardrath & McEvily (1961)

The liquid metal may invade grain & interphase

boundaries

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Macroscopic features - I

Ratio of fracturestresses of 4145 steel with & without Pb(0.3%)

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Macroscopic features - II

Room T

400 < T < 620 °F

Steel 4145Mostovoy & Breyer, (1968)

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LME&Chemistry The effect of impurities

Westwood et al., (1971)

Polycrystalline Al

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LMEDelayed failure

2024-T4 Al alloy + Hg amalgam

Rostoker et al. (1960)

Static fatigue

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LME and Grain Boundaries: Al/Ga

Ludwig et al. (2006)

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LME and Grain Boundaries

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LME: consensus about facts• Instantaneous failure under applied or residual

stresses• Delayed failure at a static stress level below

the tensile strength• Microstructure is important BUT LM’s embrittle

amorphous alloys too ! Ashok et al. (1981)• Plasticity is often present• Stress independent (?) grain boundary

penetration• High temperature “corrosion”

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

From consensus: prerequisites & empirical criteria of occurrence

However factors determining which liquid metal will embrittle which solid metal still remain unclear

!

• Presence of an external stress• A pre-existing crack or plasticity (at least limited) & presence of obstacles to dislocation motion e.g. grain boundaries, twins, precipitates• Adsorption of the active species at the obstacles and at the tip(s) of propagating crack(s)•  Limited mutual solubility of the liquid and the solid

•  Little or no tendency of stable high Tm compounds

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Is modeling understanding ?

An heuristic approach (sometimes a random walk)

Experiments

Analytic theory Numerical approaches

LME ?

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Modeling LME & Crack propagation: Elasticity

QuickTime™ et undécompresseur TIFF (LZW)sont requis pour visionner cette image.

dU =πadaσ 2

E

G =dUda

=πaσ 2

EGc =2σ

σ c =2γ sEπ a

(Griffith −1921)

Adsorption of the liquid at the crack tipdecreases s and thus σc

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LME & Crack propagation: ....adding plasticity

Griffith’s model is elastic (Orowan, Stoloff & Johnson, …)

σ c =2γ eff E

π a(Orowan −1950)

eff = γ p + γ s (Al : γ p ≈ 4200γ s , Kargol et  al., 1977)But…

eff = Aγ s , A ≈ σ 0 / σ y ≈ 100  for  steel (Gilman, 1960)However…

Small changes in s large influence on eff

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

“Thermodynamic” model: Successes & limitations

• Cu/Pb-melt Eborall et al. (1956)• Al bicrystals/HgGa solution Kargoll et al. (1977)• Al 6061/inclusions Bi, Cd, Pb Roth et al. (1980, 1982)

• No mechanisms• No microscopic description of the SL interface• No kinetics• No understanding of plasticity at the crack tip

BUT …

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

LME & Cracks & DuctileBrittleTransitionTemp

System s(J/m2)

Al 1.0

Al-Pb 0.344

Al-Cd 0.298

Al-Bi 0.288Old & Trevena (1979)

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Limitations of this approach

• Lack of surface energy data (ab-initio ?)• Physical parameters influencing LME do not

appear in the model (e.g. microstructure)• No insight into the mechanisms (atomistic scale)

Consensus: reduction in s is necessary but not sufficient

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

plane

Slipplane

Slipplane

σa

σa

σm =γ sEx0

- Bond breaking

σ =2σ aaρ

- Crack length 2a  Stress at the tip

-=x0 σ prop =γ sE4a

σ prop(blunted ) =Eγ fracture

4a=

Eγ sξ4a

AIRC: the atoms of the liquid (black) reduce/increase the bond strength

AIRC Adsorption Induced Reduction in Cohesion

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

• Hardening is absent• Enhanced plasticity at the crack tip and

HT LME are not explained• Transport mechanisms are not considered• No realistic prediction of the crack velocity

AIRC: present situation

•  AIRC is widely accepted•  Consistent with several experimental findings (Zn-Hg, Zn-Ga,        hydrogen, …)•  Explains impurity action

BUT ...

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

AIRC variants & improvements

Wetting effects ?Rabkin et al, (19..)

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

[010]

[001]

Modeling:Crystal-Liquid

interface

Geysermans et al. (2005) %(x) = δ x − xi( )

i∑

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Conclusive remarks• LME can affect several structural materials and be

a major cause of damage

• Understanding of LME is still limited• No general agreement exists about mechanisms• There is lack of atomic scale description• No predictions can be made i.e. which liquid

metal embrittles which solid metal

• Research on LME could result in decisive advances in the physics & chemistry of interfaces

MATGEN-IV.2 February 2 - 7, 2009, Kiruna, Sweden

Origins of Illustrations1. http://www.tech.plym.ac.uk/sme/Interactive_Resources/

tutorials/FailureAnalysis/Fractography/Fractography_Resource4.htm