Basics and application of

phase-field modeling in
materials science
5 : miscellaneous subjects;
summary, perspectives
Mathis Plapp

Laboratoire de Physique de la Matière Condensée

CNRS/Ecole Polytechnique, 91128 Palaiseau, France
 Eutectic solidification

Al-Cu (Trivedi et al)

CBr4-C2Cl6 (S. Akamatsu, G. Faivre)

 Eutectic phase-field model

- Multi-phase-field method: introduced by Access group (1996)

- Each phase is represented by a phase field pi with i=a,b,L
- Constraint : pa + pb + pL = 1

R. Folch and M. Plapp, Phys. Rev. E 72, 011602 (2005)

« Ideal » interfaces

p3 = 0

p1=1-p2

One variable
Generic situation

p3 varies
(third-phase
« adsorption »)

Two variables
Visualization in the Gibbs simplex

One independent variable

p1  p2  p3  1
Two independent variables
 Construction of the interpolation
functions

Original Access model: f TW   p i2p 2j

i j
fTW   p 1  p i  2 2
Present model: i works !

     p 9p  5/ 4
i
g i  p i2 151  p i  1  p i  p j  p k
2 2
Tilting function: i i

Together with two-phase generalization of the antitrapping current:

Reduces exactly to the quantitative alloy model on each SL-interface
Benchmark: equal volume fractions

W : thickness of the diffuse interface

l : lamellar spacing ( = width of a lamella pair)
For comparison: qualitative model

For largest W: 20 % error !

Eutectic solidification : composites
From lamellae to rods and back
The composition is
slowly varied:
c  0.3  c  0.3

Liu,Lee,Enlow,Trivedi
Multi-component multi-phase solidification

• Binary alloy models can be generalized to multiple components

• Multi-phase approach can be generalized to arbitrary number of phases
• Expressions for multi-component antitrapping currents are available
• Coupling to thermodynamic databases done by several groups
• Commercial (MICRESS) and open source (OpenPhase) codes available
• Polycrystals: two approaches (multi-phase-field or orientation-field)
• Input parameters needed: free-energy functions (~ OK), diffusion
coefficients (???), interface properties (???!!?!?)

There is a need for extensive benchmarking !

 Polycrystals

Experiments: Qin et al., J. Cryst. Growth (2005) Simulations: Warren et al., Acta Mat. (2003)
 Models: « top down »
Sharp interface models

Need as input:
- Equations of motion for the boundaries
- Surface energies (of solid-liquid interfaces and grain boundaries)
- Mobilities

Multi-phase-field models (Steinbach et al., Chen et al.)

- Use a different phase field for each grain

- The couplings between the fields allows to specify surface tensions
and mobilities
- Looks computationally costly, but can be implemented efficiently
 Models: « bottom up »

Molecular dynamics

Need as input:
- Interatomic potentials
- Surface energies and mobilities are « emergent properties »
- Short length and time scales

Phase-field crystals (Elder et al., many others …)

- Continuum theory (temporal average of MD or simplified DFT)

- Keeps atomic length scales, but allows to reach larger time scales
- Allows to describe solidification and grain boundary dynamics
Models with orientation fields

q
 Basic formulation
Free energy functional (pure substance):
2

F
K
V 2
f   Hf f   Lugf   fori f, q 
T  Tm
Equations of motion: u
F L / cp
 t f   f
f
F
 t q   q
q
 t u  Dfu   t f
 Orientational free energy
Replace gradient square by absolute value (Kobayashi, Warren et al.):

fori f, q  K qsf q (Model I)

Introduce singular coupling function (Mellenthin, Henry, Plapp):

fori f, q  K qsfq

2
(Model II)

7f  6f
3 4
sf   s0  0 s1  
1  f2
Branched growth structures
 Origin of branched growth

Interplay

between an

instability that favors branching

and

stabilizing factors (of microscopic origin)

 Laplacian instability
 u0
2

Laplacian
field
+
Growth ~
Flux
vn  J u=0
u=1
 Viscous fingering in a Hele-Shaw cell
With S. Nguyen, V. K. Verma (IIT Guwahati), R. Folch (U. Tarragona)

W y
x V0
b

   
Incompressible, immiscible fluids :   V1    V2  0
12 1   12 2  
Darcy’s law :  2 V1  P1  2 V2  P2
b b
Laplace’s law : P1  P2  
 Finger patterns and fluid properties

Growth Selection λ=L/W

Shear-thinning Normal-stress fluid Yield-stress fluid

 From fingers over dendrites to cracks

PEO solution (Ignès-Mullol et al.) Liquid crystals (Buka et al.)

Clay suspensions (Van Damme et al.)

Two-phase flow: advected phase-field model
  
 t f  f  f  W  f  W (f) f   V  f
3 2 2 2

non-conserved Ginzburg-Landau f advection

(f)   
f
cancels
curvature effect


v n       n̂  V
 Model for Hele-Shaw flow
 
V  0
1 f 1 f
  1  2
L y
x V0
b 2 2
12f    
 2
V  P  f
b
Capillary effect

Principle: Obtain pressure and velocity fields with standard solvers, with
fluid properties that depend on the phase field

R. Folch et al. (1999) Hele-Shaw flow

T. Biben and C. Misbah (2003) coupling with Navier Stokes
J. Beaucourt et al. (2005) generalization to visco-elastic materials
 Fingers in shear-thinning fluids

Shear-thinning fluid :

 2   f  f  
0
2

1  as 2
f s   a  1
1  s2
Low-viscosity zone in front
of the finger makes it thinner
~  / V 
 Convective Cahn-Hilliard equation
F   K c   Hc  c1  c  c 2 
2 2 2
V

F
Chemical potential :  
c
Mass diffusion current : J   M


 t c  Vc  M 

   

  t V  VV  P  Kc  c    
Simulation of unstable two-phase flows

Didier Jamet (CEA Grenoble)

 Problem : mixing vs. diffusion

d0
At the surface of a spherical domain :   d 0  
R
Mass transport by diffusion : D  Mf c 
Simulation of a branching crack

Alain Karma and Alexander Lobkowski (Boston)

 In three dimensions …

Hervé Henry (PMC)

Nematic liquid crystals
Anisotropic molecules
can exhibit mesophases:
Orientational but not
positional ordering

Simplest case:
isotropic-nematic transition
Rod- or disc-shaped molecules

Nematic Isotropic
Deformation and surface anchoring

bend

splay

Surface anchoring
twist
 Frank-Oseen free energy

Define a director field n̂x  : local average orientation

FV  
2
 
K1  2 K 2
  n̂ 
2

  2
n̂    n̂ 
K3
2
 
n̂    n̂   2

splay twist bend

Surface anchoring:

n̂x  is a unit vector
FS   K af n̂, â 
dS
 Tensor order parameter


d 2
d
Q ij  d i d j   ij
3

Eigenvalues: Tr Q  0  l1  l 2  l 3  0
Uniaxial nematics: l 1  l 2  l 3 / 2
3   ij 
Q ij  q  n i n j    Eigenvalues: q,  q / 2,  q / 2
2  3
 Landau-de Gennes free energy
1
2
  1

F   L1  i Q jk  L 2  jQ jk 
2

2
2

1
2 3
1
4

A(T) Q ijQ ij  BT Q ijQ jk Qki  C Q ijQ ij
1
 2

ny
F

nx
Liquid crystal solidification

 Mixture of liquid
crystal and isotropic
impurities
 Growth limited by
impurity diffusion
in a temperature
gradient
 Used as analog for
solidification
(Bechhoefer et al.)
 Directional solidification
Liquid crystal alloy: phase transition driven by composition

c  ceq q, T0  x  Vt mcT0 

u T  T0  Gx  Vt  u eq x, t   lT 
cT0  lT G


A  2  6 u  ueq x, t   
B  12  12 u  ueq x, t   C  8/ 3

 tu  D u   tq
2

F
 t Q ij 
Q ij
constra int s
Directional solidification of LC
Other applications of phase-field models
• Solid-solid transformation (precipitation, martensites): includes elasticity
• Epitaxial growth
• Crack propagation
• Grain growth
• Nucleation and branch formation: includes fluctuations
• Solidification with convection: includes hydrodynamics
• Fluid-fluid interfaces, multiphase flows, wetting
• Membranes, biological structures
• Electrodeposition: includes electric field
• Electromigration

Long-term goal: connect length scales to obtain predictive capabilities

(computational materials science)