HOMEWORK

From Dieter
13-5

Module #23

CREEP & SUPERPLASTICITY

READING LIST
►DIETER: Ch. 13, all

• Chapter 13 in Meyers & Chawla

• T.G. Nieh, J. Wadsworth and O.D. Sherby,
Superplasticity in metals and ceramics, (Cambridge
University Press, Cambridge, 1997).
 High Temperature Deformation
of Crystalline Materials
• T < 0.5 Tmp design so  < ys
• T > 0.5 Tmp design so  < ys

We must also consider

HIGH permanent deformation
TEMPERATURE occurring by CREEP over
DEFORMATION long periods of time.
MODES

CREEP = time-dependent plastic deformation
YIELDING = time-independent plastic deformation
• We can incorporate these
things into Deformation Stress YIELDING

Mechanism Maps (DMM)

[Frost and Ashby]
 glide
/G
• They are useful in design
 creep
applications – have DMM
for each alloy – but to elastic creep
really know/predict high T
useful life of alloy, still Temperature
must do testing – usually T/Tmp
accelerated!
Phenomenological
Description of Creep
STRAIN CREEP
TIME DEPENDENT
Elastic 

CONSTANT
LOAD!
YIELDING
TIME INDEPENDENT
Elastic 

Plastic 

t1 time

When stress is removed

 Constant Load Creep Test
d
 
  dt
III
III
II
II INC. ,T II
II INC. ,T
III
II
I II III
I
0
time tf time tf

Stage II:   constant AND minimum, II

Difference between constant load
and constant stress creep tests

Fracture
Constant X
Load
Constant
d Stress
SS  Tertiar
dt
d y
Strain

dt creep

E
Secondary creep

time
Primary creep
 STAGES OF CREEP
STAGE 0: Initial Strain on Loading (Elastic???)

STAGE I: Transient (Primary) Creep

  with  time & strain
   and subgrains form  TEM
Cause   Analyses
 to  strength of material 

STAGE II: Steady State Creep Constant

microstructure  constant minimum
creep rate, II
  
rate of    due rate of   
Cause  
 to deformation due to recovery
 STAGES OF CREEP – cont’d

STAGE III: Tertiary Creep

  with  time & strain
fracture is soon to occur after Stage III begins

recrystallization

Cause coarsening of 2nd phase particles
formation of internal cracks & voids

 CREEP
Design Considerations

Usually either II or t f

This parameter is usually addressed in

scientific analyses as it is most readily
correlated with microstructures and
creep mechanisms.

tf: Use when failure must not occur but

deformation/creep is tolerable.
 Ex., Casing on rocket combustion chamber

II : Use when long life is necessary and dimensional

tolerances are critical.
 Ex., Jet engine turbine blade
Steady-State Creep Rate (Monkman-Grant Relationship)

Where steady-state creep dominates creep life (some materials)

min  t f  constant

Sometimes it needs to be modified as follows:

min

tf  constant;   1
 Steady-State Creep Rate (Dorn Relationship)

Qc 
II  ss  A n exp  
 RT  When T  0.5Tmp
n Qc  activation energy for
   Q  self-diffusion
 o   exp  c 
o   RT 

A & n are material constants Vary with

Qc is the creep activation energy &T
due to changes in the
creep mechanism

This equation also holds for many materials and establishes

the temperature and stress dependence of the creep rate.

Correlates with the activation energy for the

mechanisms causing creep. We’ll address the Qc  R
ln   
1
2

mechanisms shortly (in great detail). 1 1

T  T 
 1 2 
 For many materials:
When T ≥ 0.5Tmp
Q is the approximate activation energy for self diffusion of elements

Activation Energy for Self Diffusion, QSD (MJ/mole)

100
Nb
α-Fe UO2
W
Ni
MgO
NaCl
Al O
Cu γ-Fe 2 3
Al β-Co
AgBr
Zn Mg Ag
Au
10-1 α-Ti β-Ti
Pb
In Ca
Sn

Na Li

Qc = QSD
10-2
10-2 10-1 100

Activation Energy for Creep, Qc (MJ/mole)

Activation energies for steady state creep and self diffusion for a select group of crystalline materials.
Adapted from O.D. Sherby and A.K. Miller, J. Eng. Mater. Technol., v. 101 (1979) p. 387.
Several deformation Single deformation
mechanisms are operative mechanism is operative
• Slip (Diffusional flow)
• Climb
• Cross slip
• Diffusion???
• etc.
Q = variable Q = constant

Acta Metall., v. 5, p. 219 (1957)

 Generalized form of the Dorn Equation
n p
ADGb     b 
      , where D  Do exp(Q / kT )
kT  G   d 

D = diffusion coefficient
d = grain size
b = Burgers vector
k = Boltzmann’s constant
T = the absolute temperature (degrees Kelvin)
G = the shear modulus
 = applied stress
n = stress exponent
p = inverse grain size exponent
A = a dimensionless constant.

This form of the Dorn equation applies for all creep mechanisms.
CREEP MECHANISMS
T
Ideal shear strength
It is convenient to discuss
deformation (and creep) Dislocation glide
mechanisms in terms of
(low temp)
temperature and/or applied Dislocation creep
stress.
(high temp)
/G
Deformation mechanism maps or
Coble (g.b.)
(DMMs) * /G Diffusion creep

NH (lattice)
Elastic
regime
* Adapted from: M.F. Ashby, Acta
Metall., v. 20 (1972) p. 887-897.

T/Tmp

WE’LL ADDRESS THE DETAILS IN A LITTLE WHILE

 CREEP MECHANISMS
Low T or High : Dislocation Glide (DG) Creep

• Dislocation () motion is assisted by thermal activation.

• Does it involve diffusive processes?

Energy
Energy

Eo Eo
E

Obstacle to motion Obstacle to motion

Position Position

• Eo is the activation that must be overcome by the

dislocation to cause it to move from one lattice location
to another.
• Thermal activation supplies energy (in absence of
stress) which reduces the activation energy barrier
by E.

  Eo   E   Eo    bas 
dg  o exp   exp    o exp   exp  
 kT   kT   kT   kT 

• NOTE: for BCC metals, T has a major influence on

plastic deformation
 Diffusional Flow Creep Mechanisms
Nabarro-Herring (NH) Creep
• Occurs solely by diffusional mass transport (i.e., diffusion).

• Is important for much higher T’s and lower σ’s than was the case
for dislocation glide creep.

• Can occur in crystalline and amorphous materials.

 Applied stress After flux

Arrows denote
vacancy flux
Original
Contraction shape
“Poisson effect”


 Diffusional Flow Creep Mechanisms – cont’d
Nabarro-Herring (NH) Creep
• Applied stress creates tensile and compressive regions within a
grain.
– Concentration of vacancies in tensile regions is greater than that in
compressive region.
– Vacancy concentration gradient  diffusion
– Diffusion leads to
shape change
 Applied stress After flux

Arrows denote
vacancy flux
Original
Contraction shape
“Poisson effect”


 2
ANH DLGb  b    
NH     
kT d  G
ANH  10  16
DL  lattice diffusion coefficient

NH creep is competitive with  glide creep

NH creep  with T and σ
 Diffusional Flow Creep Mechanisms
Coble Creep
• Also driven by stress-induced vacancy
concentration gradient,
BUT
• Diffusional mass transport occurs along:
– Grain boundaries in polycrystalline materials
– Surfaces in single crystal materials.
3
A D Gb  b    
C  C GB      Shape after flux
kT  b  d   G 
AC  30  50
DGB  grain boundary diffusion coefficient
 = effective width of grain boundary

• NH and Coble creep are really parallel processes

(i.e., they operate in tandem).
Vacancy Flow Due to NH and Coble Creep Mechanisms
Vacancy sink

Vacancy source

Nabarro-Herring Coble
2 3
A D Gb  b     A D Gb  b   
NH  NH L     C  C GB     
kT d  G kT  b  d  G
 Grain Boundary Sliding

Voids are created

 during diffusion-
Grain boundary
sliding occurs to
assisted flow. keep material from
falling apart

GB sliding occurs in conjunction with the NH & Coble creep

mechanisms

GB sliding is also thought to be the mechanisms that allows a

material to extend in length with no net change in grain size
during superplastic forming operations
 Dislocation Climb-Glide Creep
• RECALL: Crystals always contain defects
• Certain point defects, vacancies, can promote plastic deformation
by helping ’s overcome obstacles. This is “vacancy assisted climb”

Obstacle

     Slip plane

vacancy

     Slip plane
ADGb   
5

CG   
kT  G 

    Slip plane

 
    Slip plane
 Solute Drag Creep
• Occurs in some metallic alloys (i.e. those that don’t display short-time creep
behavior).
• At low temperatures, solute atoms are immobile.
• At higher temperatures, solute atoms may diffuse, albeit slowly.
• From the Taylor-Orowan equation:    bv
dislocations move at a velocity v.
• Provided the dislocation velocity is near the velocity at which the solutes
diffuse, the solutes can move along with the dislocations.
Dsolute 3  1   Dsolute   kT     
2

SD     b 2   G  kT  G 
 b co
2
 64 2
b o 
c    
OR
( = atomic volume)
2
D      
SD  ASD  solute
2   
 b   kT  G 
• This leads to serrated flow (i.e., the Portevin-Le Chatelier effect).
 Creep in two phase alloys
• The creep rate of a material can be greatly reduced by the incorporation
of a fine dispersion of non-deforming particles at grain boundaries.

• The particles effectively inhibit grain boundary sliding.

Mass
transfer

GRAIN 1

GRAIN 2 Particle

For grain 1 to slide relative to grain 2, there must be mass transport to

“move” one portion of the particle relative to the other.

• Larger particles are more effective at inhibiting grain boundary sliding.

• ROLE OF PARTICLES: they inhibit recrystallization; they “pin” grain
boundaries.
 TABLE
Representative values of the parameters n, p and approximate
values of the constant A for ss

Mechanism Favored by Description A n p

Nabarro-Herring High temperature, Vacancy diffusion 10-15 1 2
Creep low stress and through the crystal
large grain size lattice

Coble Creep Low stress, fine Vacancy diffusion 30-50 1 3

grain sizes and along grain
temperatures less boundaries
than those for
which NH creep
dominates

Grain Boundary Same range as NH Sliding 2 2 or 3

Sliding and Coble creep accommodated by
vacancy diffusion
through the crystal
lattice (p=2) or along
grain boundaries
(p=3)

Dislocation creep High stress, lower Dislocation motion, 3-8 0

temperatures in with climb over
comparison to microstructural
Coble creep, and obstacles.
large grain sizes

n p
ADGb     b 
     
kT  G   d 
where D  Do exp(Q / kT )
COMBINED CREEP MECHANISMS
 diffusion  NH  C

Controlling parameters:

 ANH vs. AC AND grain size (d)

 NH:   d 2

 Coble:   d 3  grain size is more important here

Effect of Grain size on  mechanism

C
 diffusion  NH  C
Ln 
NH
NH  C

Coble creep
dominates if d < dc
dC

Ln d
 Diffusion & Dislocation Creep
• In this case,  is critical for determining the relative proportion
that each mechanism contributes to the overall creep rate.
total  diffusion  CG
 climb/glide processes
total  sum of mechanisms as shown below
UO2 d  10 m
GC
diffusion  glide

 glide 
ln (strain rate)

GC  diffusion Diffusion

Dominated
  (const) m

c
Faster process
dominates
ln (stress) deformation

[Courtney]
 DEFORMATION MECHANISM MAPS
• Deformation mechanism maps (DMMs) show combinations of σ and T for
specific creep mechanisms.

• DMMs have been developed

T from constitutive equations for
Ideal shear strength
alloys and are available for only
Dislocation glide a small fraction of engineering
(low temp)
Dislocation creep alloys.
(high temp)
• DMM: shows the dominant &
/G
or active creep mechanism for a
particular  & T
Coble (g.b.)
/G Diffusion creep

NH (lattice)
Elastic Visual aids
regime
*We can also superimpose
Normalized stress strain rates onto DMM. Can
Maximum  /G = T/Tmp
0.1 see how this changes the
dominant creep mechanism.
 Deformation Mechanism Maps for Al, W, and MgO

 [Courtney]
Constant  10 5

G
Al W MgO
Elastic at Elastic at Elastic at
T=200°C T=2000°C T=2000°C

• Does grain size influence DMMs?

• Why does Coble creep occur at low T relative to NH creep?
 Incorporation of
strain rate
considerations
and ignoring the
elastic
component
modifies DMMs

Why does deformation rate shift

the various regimes as illustrated?

[Courtney]
GRAIN SIZE EFFECT

Constant T

[Courtney]

 How does grain size influence DMMs?

 Effects of Structure & Properties on Creep Resistance

II  with:
   applied stress
 diffusivity (  activation energy)
 grain size

 For a given application T, use of higher Tmp material will 

diffusivity and correspondingly  II .

 In general:
 Diffusivities in bcc metals >> diffusivities in fcc metals
  if we have a bcc metal and an fcc metal with the same Tmp,
bcc   fcc
  grain size,   for diffusion controlled creep but does not have any
influence on the creep mechanism.

 For glide mechanism:  grain size will  

 Grain Size Effect
• MarM 200  Ni-based superalloy used in turbine blades

[Courtney]

  109 sec 1
d  100  m  / G  103
T / Tmp  0.5  is a Coble creep process
d  1 cm  / G  103  GS is very important
T / Tmp  0.8
 A few important comments
• Recall, that one way to engineer a material to exhibit
higher creep resistance is to add second phase
particles.

• If the particles restrict grain boundary sliding, they will

decrease the steady state creep rate.

• Most creep resistant alloys utilize this technique.

Single crystals are also used in conjunction with
second phase particles.
 Estimates of Creep Life
Have critical correlations between
properties and performance

SS or time to creep Want to know creep

behavior over extended
fracture (tf) are known period. Can be
for short times ≥ 20 – 30 years
(accelerated tests)

Design Lives: Jet turbine ---------------- 10,000 hr (1 year)

Stationary turbine ------- ~10 years
Nuclear reactor ---------- ~40 years

MUST: be able to extrapolate properties measured over short (accelerated) time

to predict performance over extended time.

ONE
WAY: Larson-Miller (L-M) Method
Correlates temperature T (°R)
with the time for failure tr at a
constant engineering stress .
 Larson-Miller (LM) Equation
T (log tr  c)  m

c=constant, depends on alloy

m=parameter that depends on stress = LM parameter

Experimentally the L-M

• Known “c” for alloy, run a test at a approach gives optimistic
life predictions
specific T and measure tr (time to
rupture … previously I may have
used tf to denote the same thing).
2
log tr
• Now, if we keep stress constant,
m = constant. 1
m2
Determine tr
• We can change T and can m1
calculate the time to rupture
(failure).
1/T
Intersection of Pick T
lines gives
value of c
Effects of creep on microstructure and fracture morphology
• Microstructure:
– During creep microstructural changes occur. Some are significant.
– Atomic scale
• Dislocations are created and forced to move through the material leading to
work hardening (Stage I).
• Combined hardening and recovery processes can lead to the formation of
subgrains or dislocation networks (Stage II).

– Micro-/Macro-structures:
• Slip band formation
• Grain boundary sliding
• Cavity formation and growth
• Cracking
• Grain growth
• Precipitate/second phase coarsening (Ostwald ripening)
• Etc.
 Effects of
creep on
microstructure
and fracture
morphology

Images from ASM

Specialty Handbook,
Heat-Resistant
Materials, J.R. Davis
editor, (ASM
International, Materials
Park, OH, 1997) p.24
Effects of creep on microstructure and fracture morphology
• Fracture:
– Intergranular creep ruptures occur by (i) triple point cracking or (ii)
grain boundary cavitation. Which process occurs depends upon strain
rate and test temperature.

Triple
point

σ cracking
σ
σ σ

Cavitation

– Intermediate T, high strain rate:

• Triple-point (or wedge) cracking
• GBS can produce stress concentrations at grain boundary triple points to
induce and propagate wedge cracks.
• Etc.
Effects of creep on microstructure and fracture morphology
• Fracture (cont’d):

Triple
point

σ cracking
σ
σ σ

Cavitation

– High T, low strain rate:

• GBS promotes cavity formation. This is a diffusion-controlled
process which should not be confused with microvoid formation
during slip.
 Examples


σ
Triple point
cracking

σ

Images from ASM Specialty Handbook, Heat-Resistant Materials, J.R.

Davis editor, (ASM International, Materials Park, OH, 1997), p.25
 Examples


σ
Cavitation

σ

Images from ASM Specialty Handbook, Heat-Resistant Materials, J.R.

Davis editor, (ASM International, Materials Park, OH, 1997), p.25
 Superplasticity
• It has been observed that some materials when heated above 0.5Tmp
can elongate to extremely large strains (e.g., ~5000%).

• We take advantage of this phenomenon to form complex shapes that

cannot normally be obtained by forging, extrusion, or other
metalworking processes.

• This is called “superplasticity.”

• The ability of a material to deform superplastically is related to its

resistance to necking during deformation. This resistance is defined
by the strain rate sensitivity exponent, m.

[Meyers & Chawla]

 Superplasticity – cont’d
• Microstructural Requirements:

1. Grain size ≤ 10 μm
– This grain size must not change during straining!
– Explains why we see superplasticity so frequently in two-phase alloys that
are resistant to grain growth.

2. Grain shape remains equiaxed during superplastic deformation

– Grains “slide” and “tilt” under stress.

• Strain Rate Sensitivity:

 flow  F ( );    flow
 flow  F ( );    flow  constant
 flow  k    
m
Strain rate exponent

Depends on strain rate

 [Courtney]

[Meyers & Chawla]

 Determination of Strain Rate Sensitivity

Δ

As m increases, the material
becomes more resistant to
Increase strain rate
necking. •We can determine m
from strain rate change
and/or stress relaxation
In the neck, neck  non  neck tests.

•We can deduce m from

1.  flow stress of the stress strain curves
collected at different
neck > flow stress strain rates…not as
log  accurate.
in non-necked region
2.  deformation stops
in the neck Slope = m =
1/m’

(1) and (2) are a

result of strain rate
hardening ln (strain rate)
 m
Experimental
100
Observations
Values of m lie
between 0 and 1. For
superplastic metals, m
In superplastic 10-1 is closer to 1.

materials
m = 0.3 – 0.8 for most 10-2
total 
1 2 3 4
10 10 10 10

ln  III
Superplasticity is only
II observed in region II;
where   rapidly
with  .
I

ln 

m III
II

I
superplastic

ln 
The amount of
superplastic
deformation tends
to increase as T
increases and d
decreases.

[Courtney]
 Grain switching mechanism
M.F. Ashby and R.A. Verall, “Diffusion-accommodated flow and
superplasticity,” Acta Metallurgica, V. 21, n. 2 (1973), p. 149-163.

• The primary mechanism contributing to superplasticity is argued to be grain

boundary sliding (GBS).

• Grain shapes don’t change; grain locations change.

• Diffusion is required to allow grains

to “switch” positions.

Region I: low  -- diffusional flow &

grain switching

Region III: high  --  creep

Region II: where superplasticity occurs. A mixture of I and III.

* m is maximum where you have both diffusional flow and creep!

M.F. Ashby and R.A. Verall, “Diffusion-
accommodated flow and superplasticity,”
Acta Metallurgica, V. 21, n. 2 (1973), p.
149-163.
What types of materials deform superplastically?
• Al alloys
– 7075, 2095
– Al-6Cu-0.5Zr – (1200% strain)
– Al(6061)-20% SiC(whiskers) – (1400% strain)
– Etc…
• Ti alloys
– Ti-6Al-4V – (1000% strain)
• Ni alloys
– Alloy 718
• Others
– Bi-44%Sn eutectic – (1950% strain)
– Mg-33%Al eutectic – (2100% strain)
– Pb-62%Sn eutectic – (4850% strain)
• Ceramics
– Zirconia – (350% strain)
– Zirconia + SiO2 – (1000% strain)
 Types of superplasticity
• Fine-structure superplasticity (FSS)
– Occurs at low rates (i.e., 10-4 to 10-3 s-1. Fine grain size is
required along with second phase particles at grain
boundaries to retain grain size.

• Internal-stress superplasticity (ISS)

– Occurs when internal stresses can develop in a material
allowing tensile plasticity to occur under low, externally
applied stresses. Fine grain size is not required here.

• High strain-rate superplasticity (HSRS)

– An extension of FSS. Occurs at higher rates (i.e., 10-1 to 101
s-1, which is close to the strain rates for forging. Fine grain
size is necessary along with the presence of liquid phases at
grain boundaries.