Advanced High Temperature Alloys PDF

Advanced High
Temperature Alloys
Prof. Dr.-Ing.
g Uwe Glatzel
Metals and Alloys
University Bayreuth
SS 2010
University Bayreuth, Advanced High Temperature Alloys 1 Uwe Glatzel, Metals and Alloys
Lecturer:
Prof Dr
Prof. Dr.-Ing.
Ing habil
habil. Uwe Glatzel
• born Dez. 1960
• Physik-Diplom (B.Sc.
(B Sc and M.Sc)
M Sc) in Tübingen
(exchange year in Corvallis, Oregon, USA)
• PhD thesis at the Institute for Metals Research, Technical
University Berlin, Prof. Monika Feller-Kniepmeier
• post-doc (1 Jahr) at Stanford University
• Habilitation TU-Berlin
• Gerhard-Hess award of the German Science Foundation
(DFG) for young scientist (400.000 €)
• 1996-2003 full professor for Metals and Alloys, Jena
• since April 2003 Bayreuth (Chair for Metals and Alloys)
postal address:
Ludwig-Thoma-Str.
d i h 36bb phone:
h +49 (0)
( ) 921 - 55-5555
D-95447 Bayreuth, Germany e-mail: uwe.glatzel@uni-bayreuth.de
Literature
• R. Bürgel
R Bürgel, Handbuch Hochtemperatur-Werkstofftechnik
Hochtemperatur Werkstofftechnik, Vieweg
• R.C. Reed, The Superalloys - Fundamentals and Applications, Cambridge Univ. Press
• H. Frost, M. Ashby, Deformation-Mechanism Maps, Pergamon Press
• G M
G. Meetham,
th M
M. VVan dder V
Voorde,
d Materials
M t i l ffor Hi
High
h Temperature
T t Engineering
E i i
Applications, Springer
• J. Betten, Creep Mechanics, Springer
• Askeland: Materialwissenschaften, Spektrum Lehrbuch; 1994
• Callister: Materials Science and Engineering - An Introduction, Wiley, New York, 1999
• H. Schumann, Metallographie, Deutscher Verlag für Grundstoffindustrie, Leipzig
• F. Vollertsen, S. Vogler, Werkstoffeigenschaften und Mikrostruktur, Hauser Verlag
• P. Haasen, Physikalische Metallkunde, Springer-Verlag, Berlin
• H -JJ. Bargel
H. Bargel, G.
G Schulze,
Schulze Werkstoffkunde,
Werkstoffkunde VDI
VDI-Verlag
Verlag, Düsseldorf
• P. Sarrazin, A. Galerie, J. Fouletier, Mechanisms of High Temperature Corrosion, Trans
Tech Publications
lecture notes: http://www.metalle.uni-bayreuth.de then "Lehre" then "Vorlesungen",

you will find the link to this lecture notes 3and three review talks we
University Bayreuth, Advanced High Temperature Alloys
will do at the end.
Uwe Glatzel, Metals and Alloys
What You Should Know:
• basic thermodynamics
• introduction to diffusion
• introduction to dislocations
• phase diagrams
• theory of elasticity
• ...
• basic materials science courses
Contents
11. Introduction
Introduction, Basics
2. Stability of Microstructure
3. Mechanical Properties
a)) Static
b) Cyclic (Fatigue)
4 High Temperature Corrosion
4.
5. High Temperature Alloys
6. Lost Wax Investment Casting
p
7. Depending g on Time: Lectures on
a) SX Ni-Base Superalloys b) LEK 94 c) Pt-Base Superalloys
Introduction
• only alloys will be looked at (no ceramics,

ceramics no
polymers).
• no coatings (BUT : practically all high
temperature systems are coated!), simply not
enough time
time.
Maximum Temperatures for
Applications
li i off Different
iff Materials
i l
i
maximum service
i temperature
t t
Group deformation/damage mechanism
[°C]
Polymer up to 300 melting, decomposing (pyrolyze)
Glass up to 800 viscous flow
Fe-Basis (coated) up to 1100
Fe-ODS
Fe ODS up to 1300
Ni-base up to 1200
creep, dislocation climb,
Metals Pt-base up to 1600
grain boundary sliding
refractory metals in inert
atmosphere
h above
b 1600
MoSi2 up to 1800
viscous flow, glass transition
Ceramics SiC up to 1600 temperature grain boundary
temperature,
sliding
p
Composits ((SiC/C)) up
p to 1600 complex
p
usable strength
u
Overview Materials
source:
Plansee AG,
AG
Reutte,
Tirol,
Austria
500 temperature [°C] 1500 2000
usable strength
u
Taking Density into Account
usable strength
u
Oxidation Resistance
Refractory Metals:
wider
definition Most common definition of
of
refractory refractory metals (refractory =
metals widerspenstig, halsstarrig):
two elements of the 5. and
three elements of the 6. period
with melting points higher
Tm of platinum
than Pt. Processing in general
by powder metallurgy.
Density
Os, Ir Pt
Re
R
W Au
Ta
Hf
Ru, Rh, Pd
Tc
Pd
Mo
Ag
Nb
Abundance of Elements
to find 1 atom Rh
within a bunch of Si-
atoms is comparable to
find one individual
person within the word
population
Material Choice
• temperature
• environment
• moving/non-moving part
• design complexity (how to manufacture)
• price constrictions (depending on application
of system). Reduction of 1 kg in weight:
– car ~ 0 - 5 €
– plane ~ 100 – 500 €
– aerospace ~ 100.000 - 500.000 €
Influence of ... on ...
• temperature:
– phase transitions, volume fractions, ...
– diffusion (Æ recrystallization, dislocation climb, diffusional creep, ... )
– thermal fatigue (TF)
• mechanical:
– creep
– fatigue (low cycle, LCF, high cycle fatigue, HCF)
• environment:
i t
– oxidation
– corrosion
• combinations:
– thermo
thermo-mechanical
mechanical fatigue (TMF)
– stress corrosion cracking, stress oxidation, ...
Basics
Thermodynamics ↔ Kinetics
Boltzmann-statistics: energy of
movement increases with temperature
p 3
u kin atom
= kB ⋅T
2
3 Q
u total atom = 2 ⋅ u kin = 2 ⋅ kB ⋅ T = 3⋅ kB ⋅ T −
ε& = ε& 0 ⋅ e
atom
2 R ⋅T
U total mol = 3 ⋅ R ⋅ T Æ 0,33 eV, bzw. 32 kJ/mol bei 1000°C Arrhenius-plot

Vacancy Concentration
T S Æ non-zero vacancy concentration
F = U - T·S t ti is i
in thermodynamic equilibrium
Q vac
−
cv = e R ⋅T Qvacnickel = 1,36 eV (energy necessary to create one vacancy)
T[°C] 20 300 450 800 1000 1200 1454
T/Tm 0.17 0.33 0.42 0.62 0.74 0.85 1.00
cv 10-23 3·10-12 10-9 10-6 10-5 7·10-5 3·10-4
equilibrium vacancy concentration for nickel

Nickel Vacancy Concentration
100
10-55
vaccancy concentration
10-10
10-15
ncentration [10 ]
1,00
-4
10-20
vacancy con
10-25
0 200 400 600 800 1000 1200 1400 1600
temperature [°C] Tm 0,10
0,01
0 200 400 600 800 1000 1200 1400 1600
temperature [°C] Tm
Diffusion
v r
j = − D ⋅ ∇c 1. Fick's law
) · m-22 · s-11
[j] = (atoms)
(
[D] = m2 · s-1
[c] = (atoms) · m-3
vacancy diffusion
diff i or
volume diffusion
Coefficient of Diffusion
Qvac energy to create a vacancy
Qmigration
i ti activation energy to migrate a vacancy
QSD activation energy for volume diffusion
QSD = Qvac + Qmigration
− ( Q vac + Q migration ) − Q SD
D = D0 ⋅ e k ⋅T
= D0 ⋅ e k ⋅T
QSD ≈ 17 ·kB ·Tm Æ QSDnickel ≈ 2.5 eV = 244 kJ/mol

(for a perfect crystal; defects will lower the activation energies)
Dependence Melting Point and
Enthalpy of Vacancy Creation
Tm Qvac crystal
t l
element 17·R·Tm
[°C] [eV] structure
Pb 327 0 88
0.88 0 57
0.57 fcc
Al 660 1.36 0.68 fcc
Cu 1 085 1.99 1.29 fcc
Ag 1 235 2.21 1.12 fcc
Ni 1 455 2.53 1.78 fcc
Pt 1 768 2 98
2.98 1 32
1.32 fcc
Mo 2 623 4.23 3.00 bcc
W 3 422 5.40 4.00 bcc
QSD versus Tm
400 kJ/mol
0 13 kkJ/(mol·K)
0.137 /( l )
≈ 17 · kB ·NA
Coefficient of Diffusion
Steepp slope
p indicates a
high activation energy.
Small
S ll elements
l t diffuse
diff
faster.
Diffusion in fcc crystals
slower than in bcc crystals.
y
Coefficient of Diffusion with Defects
surface diffusion Coefficient of diffusion of Th

i W.
in W
grain boundary diffusion
Overall velocity for diffusion
depending on grain boundary
volume diffusion
thickness, grain size and
pipe
diffusion dislocation density.
Pipe Diffusion
Deff = DSD + adisl. · ρ · Ddisl.
adisl. area of dislocation core
volume
ol me diffusion
diff sion ( ≈ 5 b2 ≈ 00.33 nm2)
dominant
ρ dislocation density
pipe diffusion
dominant Ddisl. pipe
i diffusion
diff i along
l
increasing
dislocation core
atom
t flux
fl ~ D·area
D
decreasing
⎛ atoms ⎞ ⎛ atoms ⎞
⎜ ⎟ ~ DSD ⋅ d grain
2
⎜ ⎟ ~ D disl. ⋅ b ⋅ n
2
⎝ time ⎠ grain ⎝ time
i ⎠ disl.
dashed line:
identical atom fluxes if:
diffusion in crystal by the velocity of pipe diffusion n
DSD ⋅ d grain = D disl. ⋅ b 2 ⋅ = D disl. ⋅ b 2 ⋅ ρ
d grain
Grain Boundary Diffusion
Deff = DSD + π · δ / d · Dgrain bound.
volume diffusion
dominant with:
grain boundary diffusion
dominant
δ effective grain boundary
fine
grain
thickness ( ≈ 2 b ≈ 0.5 nm)
coarse
grain
d grain size
Ddisl. pipe diffusion along
dislocation core
dashed line: diffusion in crystal
y by
y the velocity
y
of grain boundary diffusion
Activation Energies Indicating
Mechanism Changes
~ QSD
Single crystal aluminium, oriented such that <110>{111} slip is activated.
Lytton, Shepard and Dorn, Trans. AIME 212 (1958) 220
Diffusion in Ordered Structures
(
(Intermetallic
lli Phases)
h )
• High binding energies Æ high activation
energies Æ low coefficient of diffusion
• For example NiAl: very low enthalpy of ordered
B2 structure Æ low enthalpy outweighs entropy
Æ ordered up to melting
t
temperature
t
TmNi = 1454°C
TmAl = 660°C
TmNiAl = 1638°C
Second Fick
Fick'ss Law
C be
Can b concluded
l d d directly
di tl from
f first
fi t Fick's
Fi k' law.
l
∂c
= D ⋅ Δc
∂t Similar in heat transfer systems,
systems electrical
potential, ... .
f1 ( x ) = 1 − Γ(x )
1
⎛ x ⎞
0.8 f 2 ( x ) = 1 − Γ⎜ ⎟
⎝ 0.5 ⎠
0.6
f1(x) ⎛ x ⎞
0.4 f 3 ( x ) = 1 − Γ⎜ ⎟
f2(x) ⎝ 0.05 ⎠
0.2
⎛ x ⎞ f3(x)
solution to these
c( x , t ) = c1 − (c1 − c 0 ) ⋅ Γ⎜ ⎟
⎜2 Dt ⎟
0.5 1 1.5 2
boundary conditions: ⎝ ⎠
Thermal Conductivity
The most simple, stationary case: no heat radiation, constant
temperatures in front and back of component.
λ … coefficient of heat (or thermal)
conductivity: λ = a · cp · ρ
a … coefficient of temperature conductivity

cp … heat capacity
ρ … density
compare:
r r r r
j = − D ⋅ ∇c q& = − λ ⋅ ∇T
∂c ∂T
= D ⋅ Δc = a ⋅ ΔT
∂t ∂t
Temperature Distribution with
Thermal
h l Barrrier
i Coating
i (TBC)
( )
cooling air
hot air
TBC
Wärmedämm- bond coat
Haftvermittlerschicht substrate
Grundwerkstoff
schicht
In case of transients, the temperature should reach a stable distribution as fast as possible in
order to reduce thermal stresses (Æ temperature conductrivity as high as possible).
I case off stationary
In t ti circumstances,
i t heat
h t conductivity
d ti it leads
l d tot heat
h t flow
fl into
i t the
th solid.
lid
Material Parameters at RT
heat cond.
cond heat cap.
cap density temp. cond.
temp cond
material/property λ cp ρ a
⎡ W ⎤ ⎡ J ⎤ ⎡ g ⎤ ⎡ −6 m 2 ⎤
⎢⎣ m ⋅ K ⎥⎦ ⎢ kg ⎥ ⎢10 ⎥
⎣ g⋅K⎦ ⎣⎢ cm ⎥⎦ s ⎦
3
⎣
ferritic steel 45 460 7.8 13.0

austenite steel 15 500 8.0 3.8
Ni-base alloys 11 450 8.2 3.0
Mo 145 240 10.2 59.0
Ti alloys (α-rich) 7 530 4.5 2.9
Al 210 890 2.7 87.0
Al2O3 bei RT 25 800 3.9 8.4
( Al2O3 bei 1000°C ) ( 6) source: Bürgel
Attention: Heat conductivity strongly depends on alloy composition,
composition see steels and pure
Ni with 91 W/(m⋅K) in comparison to Ni-base alloys with 11 W/(m⋅K)
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.
7. Depending on Time: Lectures on
Microstructure is NOT stable
annealed deformed
stress-relieved recrystallized
Recrystallization
time dependence of
recrystallization can be
approximated by
Avrami Johnson Mehl
Avrami-Johnson-Mehl
function:
n
−⎛⎜ t ⎞⎟
⎝ t0 ⎠
fr = 1− e
Grain Coarsening
• driving force: reduction of grain boundary
gy
energy
• T > 0.7 · Tm
• no pre-deformation
d f i necessary
• se
self-similar
s sys
system
e
• Ostwald ripening d ~ t1/3 (big grains eat up
smallll grains)
i )
• new g grains have low dislocation densityy
Grain Coarsening
monomodal
bimodal (some grain

boundaries are pinned,
e.g. by precipitates)
Precipitate Hardening
Requirements:
• solid solution at higher
t
temperatures
t (ability
( bilit to
t
homogenization heat
treatment)
• during cooling a two-phase
region
g should be reached
• in general: cooling rate as
g as possible,
high p thereafter
annealing (in the two-phase
region) to let grow the
precipitates
i i
Thermodynamic ↔ Kinetic
Example: Al-Cu
Al Cu Alloy
Guinier-Preston
G i i P t
Zones leading to
θ-Precipitates
(Al2Cu) have
paved the way
to the success of
Al-alloys
Other Examples of
precipitate
i i hardening:
h d i
Al2Cu in AlCu alloy:
platinum-base superalloy nickel-base superalloy

Time Dependence of
Precipitation Hardening
nucleation growth,
nucleation, growth coarsening
T = const.
dT precipitate size λT distance between precipitates

fT volume fraction of precipitates
Coherent - Semicoherent - Incoherent
(mit Orientierungsbezug) (ohne Orientierungsbezug)
aT − aM a T − a M a T − a M Δa
misfit δ := ≈ ≈ ≈
1 (a + a ) aM aT a
2 T M
Energy Consideration
ΔGtotal = ΔGvol + ΔGboundary + ΔGstrain + ΔGdefect
total change in free enthalpy
strain enthalpy (elastic energy + dislocation line energy)

reduction of enthalpy by precipitation coupled with a defect
enthalpy
py of pphase boundary
y ((scales with surface))
enthalpy of formation of matrix to precipitate (scales with volume)
Heterogeneous Nucleation
TEM-Micrograph of TiC Precipitates at
Di l
Dislocations
i in
i an Austenitic
A i i Steel
S l
Ostwald-Ripening
Ostwald Ripening of Precipitates
d3 - d03 ~ D⋅t here for T/Tm ≈ 0.74
γ'' particle
ti l size
i ini IN 738 LC att
T = 920°C.
particle coarsening constant of

(50 nm)3/h
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.
Room Temperature (RT) versus
Hi h Temperature
High T (HT) Deformation
D f i
• most alloy properties at room temperature are time and

rate independent (elastic constants, tension stress, ... ),
tension stress experiment.
• For T > 0.4 · Tm the pproperties
p (deformation)
( ) will be time
and rate dependent, creep experiment.
solid solution precipitate

deformation hardening fine grain hardening
strengthening hardening
cold deformation (RT) strong medium medium to strong medium to strong
temporary hardening, reduced strength with
reduced creep rupture fine grain material
creep (HT) medium medium to strong
strength, may lead to Î coarse grain, ideally
recrystallization single crystal
Change in Materials Properties
with Temperature
Material properties of steel and

Ni alloys at elevated
Ni-alloys
temperatures. Comparison
b t
between short-term
h tt andd long-
l
term parameters.
Tension ↔ Creep Experiment
Elastic (E-)Modulus and
Poisson's
i Ratio
i
E
shear modulus G G=
2 ⋅ (1 + ν)
Anisotropy and Temperature Dependence of
El i Constants
Elastic C in
i Ni-base
Ni b Superalloys
S ll
D. Siebörger, H. Knake, U. Glatzel, Mat. Sci. Eng. A298 (2001)
Orientation dependence of
Young’s modulus E of matrix
phase Distance from the center to
phase.
the surface indicates the
magnitude of the Young’s modulus
i this
in thi direction.
di ti
High Temperature Deformation
• dislocation glide (Peierls stress, in fcc and hcp very small and for T >
0.15 Tm negligible)
• cross slip
li off screw dislocations
di l ti andd dislocation
di l ti interactions
i t ti (for
(f a low
l
stacking fault energy Æ larger dislocation spacing Æ thermal
activation necessary,
necessary T > 0.2
0 2 Tm, influence on deformation rate)
• climb of edge dislocations to overcome obstacles:
diffusion at complete
p
dislocation line
Æ T > 0.4 Tm
Dislocation Climb
climb of edge dislocations to
annihilate each other.
arrangement in low energy

configurations (sub-grain
boundaries), climbing around
b l (leaving
abstacles (l i the
h glide
lid
plane)
movement of screw
dislocations with
ith kink
Internal Back Stress
Dislocations
il i climb
li b allows
ll annihilation
ihil i off dislocations
di l i
and to establish a constant dislocation density,
resulting in an internal back stress of:
σ int . = α ⋅ G ⋅ b ⋅ ρ
G⋅b 1 1
σdislocation = ⋅ and ρ=
2⋅π r r
G shear modulus, α constant 0.3 - 1, b magnitude of Burgers vector
Creep Experiment
behavior of pure metals:
primary secondary tertiary:
Creep Experimental Setup
up to 1400°C
Constant
temperature
aandd stress
st ess oor
load
Creep Experimental Setup for
Electrical Conductivityy Material
up to Melting Temperature
Pyrometer from left, optical strain

measurement from right,
g , both contact-free.
Interrupted creep tests
single crystal (SX) nickel base superalloy (habilitation thesis Glatzel)
7 8x10-6
6 [001] orientation,
orientation 1123K,
1123K 650MPa
5 6x10-6 [001] orientation, 1123K, 650MPa
straiin rate [1/s]

sstrain [%]
4
3 4x10-6
2
2x10-6
1
0 0
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70
time [h] time [h]
10-5
1123K,, 650 MPa

logarithm of strain rate versus strain
strain rate [1/s]]
10-6
(most valuable information for
10-7
materials
i l scientist)
i i )
0 1 2 3 4 5 6
University Bayreuth, Advanced High
strainTemperature
[%] Alloys 61 Uwe Glatzel, Metals and Alloys
Different Creep Stages
• primary creep: strain rate dε/dt decreases Æ
material hardens
• secondary creep stage: strain rate constant Æ
hardening and softening are in equilibrium Æ
dislocation multiplication and annihilation in
equilibrium Æ disl. density ρ = const.
• tertiary creep: necking (creep pores) develop
Æ local stress and strain rate increases
drastically
drastically.
Modelling of Primary and
Secondary
d Creep Stage
density Æ Å velocity
l i
ε& = ρ ⋅ b ⋅ v Æ
Problem with Low Creep Rates
Life time of stationary gas turbines > 20 years. years Assuming a

maximum deformation of 3%, this leads to an assumed
steady
d state strain
i rate (neglecting
( l i primary i andd tertiary
i
creep) of about ε&steady state = 5·10-11 s-1. Reliable data in labs
can only be obtained down to 1·10-9 s-1 (1 μm change with
l0 = 25 mm after 10 h Æ one creep experiment with 3.5%
strain per year!).
Therefore
Th f within
ithi university
i it llabs
b we are ttwo andd more
orders of magnitude too fast than real life in a stationary
gas turbine!
Engineering Creep Curves
raw data creep curves:
time to failure: Å time - strain
isochrone time to failure: Å isochrone strain
Natural Creep Law
ε& steady state = ρ ⋅ b ⋅ v
2
⎛ σ external ⎞
ρ≈⎜ ⎟
⎝ G⋅b ⎠
v~σ 1
external
σ 3
Î ε& ~ external
natural creep law
G ⋅b 2
Norton Creep Law (Empirical)
− Q creep
with the Norton creep exponent "n" and
ε& = A ⋅ σ n
external ⋅e R ⋅T
Qcreep ≈ Qself diffusion
power law
l break
b k
down (plb) stress dependence
of the stationary
T = const. creep rate of the
austenitic steel 800
dislocation
H at 900°C and
climb
1000°C:
diffusional creep
Diffusional Creep
• Nabarro-Hering creep (pure volume diffusion)
D self diffusion σ⋅Ω
ε& NH = 2
h 2
k ⋅T
• Coble creep (grain boundary diff.)

δ ⋅D grain boundary σ⋅Ω
ε& C = 2
h 3
k⋅ T
h grain size, δ thickness of grain boundary
Combined NH and Coble Creep:
σ ⋅ Ω ⎛ Dself diffusion π ⋅ δ⋅ D grain boundary ⎞ σΩ D eff

ε& diffusion
diff i creep = ε& NH + ε& C = 2 ⋅ ⋅ ⎜⎜ + ⎟⎟ ~ ⋅ 2
k⋅ T ⎝ h 2
h 3
⎠ kT h
π ⋅ δ⋅ D grain boundary
D eff = Dself diffusion +
h
real geometry (non-cuboidal grains)
Temperature Dependence of
Stationary
i Creep Rate
σ = 28 MPA = const.
fcc alloys:
n − Qc
3, 5 ⎛ σ ⎞
ε& s = A ⋅ γ SF ⋅ ⎜ ⎟ ⋅ e R ⋅T
⎝E⎠
Austenitischer Stahl 800H

Activation Energy for Creep
slope = 1
Constant Load ↔ Constant Stress
n n
⎛ F⎞ ⎛ F ⋅ (1 + ε) ⎞
ε& = ε& 0 ⋅ σ = ε& 0 ⋅ ⎜ ⎟ = ε& 0 ⋅ ⎜⎜
n
⎟⎟ = ε& 0 ⋅ σ 0n ⋅ (1 + ε )n
⎝A⎠ ⎝ A0 ⎠
failure
in case the gauge length
deforms uniform with
constant volume
This method is applicable to

determine the stress exponent "n"
only,
y, if the secondaryy creepp state
lasts to at least 10%
ln ε& = ln ε& 0 + n · ln σ0 + n · ln (1+ε) = const. + n · ln (1+ε)

Ashby Deformation
Mechanism Maps
n=3
Ashby Deformation
Mechanism Maps
Versetzungsklettern !
dislocation climb !
Deformation Mechanisms:
Elastic Deformation: Spontaneous and reversible deformation. In the elastic region: σ = E·ε (rule of
thumb: εe, max ≈10-3, but definitely << 1%). Plastic or non-reversible deformation achieves way higher
strains. Coble-creepp (g
(grain boundary y diffusion)) is in theory
y possible
p even at 0 K.
Dislocation Glide: … without significant time dependent recovery (climb). Is dominant in the complete
temperature regime from 0 K up to the melting point Tm at moderate and higher stress levels. At low
temperatures (< 0.4⋅T
0 4 Tm) dislocation glide has the lower boundary in the range of the elastic stress limit
(typically 10-3⋅E).
Dislocation Climb: At higher temperatures (> 0.4⋅Tm) and lower stress levels dislocation climb plays the
major role => time dependent constant strain rate (dε/dt)ss ~ σn, with a Norton stress exponent in-between 3
und 8.
Diffusional Creep: InIn-between
between 0 K und 0.8⋅T
0.8 Tm and very low stress levels: Coble
Coble-creep
creep (grain boundary
diffusion). Below 0.4⋅Tm not measurable. For geological times a time dependent deformation can be
determined. Transition to Nabarro-Herring creep (volume diffusion) is dependent on grain size and grain
boundary thickness
thickness. The transition temperature from coble to Nabarro-Herring
Nabarro Herring creep can be explained by
the different activation energies of volume and grain boundary diffusion.
Creep of Alloys
a) interaction dislocation σi = α ⋅ G ⋅ b ⋅ ρ + σsolid solution
and impurity (low temp.)
b) stationary dislocation
pinned by impurities
(C
(Cottrell
ll clouds)
l d)
c) pulled off Cortrell clouds
(Lüd bands)
(Lüders b d)
d) gliding dislocation trails
i
impurities
iti behind
b hi d (viscous
( i glide)
lid )
e) impurities faster than dislocation (very high temp., no hardening)
f) annihilation due to dislocation climb
Precipitation Hardening
σi = α ⋅ G ⋅ b ⋅ ρ + σsolid solution + σprecipitate
threshold stress concept (with n ≈ 3 - 4 and Qcreep = Qself diffusion):

n − Qc
σ − σ0 ⎞
&ε ss = A ⋅ ⎛⎜ ⎟ ⋅ e R ⋅T
⎝ E ⎠
coherent and semi-
in-coherent phase
mechanism temperature coherent phase
boundaries
boundaries
cutting 0 K up to Ts yes no
bypass by Orowan 0 K up to Ts yes yes
climb
li b over obstacles
b l > 0.4⋅T
0 4 Ts yes no
Hardening Mechanisms as
Function of Precipitate Size
dT0 initial precipitate size
σ1 and σ2 arbitrary external stress levels

= cutting
passing by: ε& ~ d T
1
climbing:
li bi ε& ~
d T2
Cutting is relevant only for coherent
precipitates
Å Dependence of stationary creep rate on

initial p
precipitate
p size for two different
external stress levels
Pinning of Dislocations by
Carbides
bid ini Austenitic
i i Steell
T = 1000°C,
1000°C σ = 25 MPa,
MP carbides
bid off th
the ttype TiC undd M23C6
Very High Volume Fractions
Volume fractions of 70% are only achievable with non-spherical
non spherical precipitates.
precipitates
Spacing between precipitates is getting smaller Æ Orowan stress
σOrowan
O ≈GG·b/L
b/L necessary. For small strains precipitates are not cut by
dislocations. With G = 90 GPa, b = 0.25 nm, L ≈ 75 nm => σOrowan ≈ 300 MPa
nickel base superalloys
ODS alloys:
ll
G ⋅ b ⋅ f vol.
σOrowan ≈
d part.
Dispersion Hardening
(oxide dispersion strengthened alloys (ODS-alloys))
yield
precipitate strengthened
stress
dispersion strengthened
temperature
back-side p
pinningg of dislocation byy
ODS-particle (Rössler + Arzt)
Summary:
Hardening
d i Mechanisms
h i
Internal back stress in steady state regime: σi = α ⋅G ⋅b⋅ ρ
Orowan stress in case of precipitates or particles: σOrowan ≈ G·b/L

ΔΩ
Solid solution strengthening: σ solid solution ≈ const. ⋅
Ω
Δa
In case of coherent precipitates: σ coherency ≈ E
a
Creep Damage
creation of a creep pore in poly-
crystalline material due to disloction glide:
a) cracks at grain boundaries b) cavities (micropores) at grain boundaries

Creep Damage
fracture
nucleation, not detectable with OM
micropore, difficult to detect

micro cracks
pear necklace like chain of

micropores (easy detectable)
Extrapolation of Time-to-Fracture Data
(L
(Larson-Miller
Mill plot,
l Larson-Miller
L Mill parameter))
M k
Monkmann-Grant
G t relation
l ti with
ith constant
t t K andd exponentt m ≈ 1:
1
K
tf = m or: ln(t f ) = K − m ⋅ ln(ε& ss )
ε& ss
Q creep
− 1
ε& ss = B ⋅ e R ⋅T or: ln((ε& ss ) = B1 − B2 ⋅
T
1 1
ln(t f ) = K − m ⋅ B1 − m ⋅ B2 ⋅ = −C + P ⋅
T T
with material dependent constants C and P
Larson Miller Plot
Larson-Miller-Plot
stationary
t ti gas turbine,
t bi about
b t 20 years off service
i ~ 130.000
130 000 h
Comparison of CMSX-6,
LEK 94 andd CMSX-4,
CMSX 4
patent Wöllmer, Glatzel,
Mack, Wortmann P = T⋅[20 + ln(tf)]⋅10-3 (T in K, tf in h)
Comparison LEK 94 with
CMSX-4 andd CMSX-6
CMSX-6 [Wortmann 88] 8.0 g/cm3
500 CMSX-4 [Erickson 94] 8.7 g/cm3
CMSX-4 [Frasier 90] 8.7 g/cm3
3
LEK-2 8.5 gg/cm
3
24 K LEK-4 8.2 g/cm
LEK-5 8.2 g/cm3
MPa]
LEK-3 8.1 g/cm3

stress [M
LEK-66
LEK 3 g/cm3
88.3
3
230 LEK-1C 8.4 g/cm
3
LEK-1B 8.3 g/cm
3
LEK-1A 8.2 g/cm
ΔT = 10 K
29 K
120
10 K
Not corrected
25 26 27 28 29 30 31 32
regarding density!
Larsen Miller parameter
Larsen-Miller-parameter
-3
P = T (20+log tB) 10
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.
Time Dependent Variation of Stress
and/or
d/ Temperature
T and/or
d/ ...
Wöhler di
Wöhl diagram for
f T < 0.4·T
0 4 Tm. Z time
ti fatigue
f ti limit,
li it D endurance
d
fatigue limit
a) type I metal (bcc) b) type II metal (fcc) endurance limit at 2·107
Change in Wöhler Diagram with
Temperature and
d Holding
ldi Time
i
Thermal Fatigue
Thermal breathing of turbine blade:

a) heating phase: edges reach high temperatures faster than interior
b) cooling phase: edges cool faster than interior
c) repeated thermal cycles lead to thermal fatigue cracks at edges
Thermal Strains and Stresses :
εthermal = αthermal · ΔT, or: σthermal = E · εthermal
Æ σthermal = E · αthermal · ΔT
Lower E-Modulus
E Modulus is Helpful:
Æ orientation of single crystals in <100> direction reduces thermal stresses
TMF and many other Time
Dependent Test Techniques
Can not be covered in this lecture!
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.
High Temperature Corrosion
• oxidation: external and internal, passivation

• carburization (internal carbides)
• nitration: internal, seldom nitrite passivation
• sulfurization: external (sometimes
passivation),
p ss v o ), seseldom
do internal
e
Worldwide 1 ton iron per minute corrodes to rust (low

temperature aqueous corrosion).
corrosion)
Ellingham Richardson Diagram
Ellingham-Richardson-Diagram
right hand and lower axes

Æ O2 partial pressure at T = 0.
As an example pO2 of
10-15 Pa = 10-20 bar = 10-17 mbar
is shown as a dashed line.
only the oxides below this line

are thermodynamic stable.
Time Dependent Oxidation
Oxidation Mechanisms
• logarithmic (not shown) Æ low temperature oxidation
which eventually comes to a stop or no measurable increase
in oxide scale thickness (e.g. Al, Cr, Mg).
• pparabolic mass changeg (Δm/A)
( )2 ~ t. Diffusion through
g
oxidation layer (either oxygen or metal). Most favorable
oxidation behavior.
• linear mass change: oxide layer with cracks Æ continuous
contact with metal (e.g. Ta, Nb).
• mass loss: volatile oxides Æ catastrophic oxidation (e.g. V,
Mo W,
Mo, W Cr,
Cr Pt).
Pt) You can see it inside a broken light bulb.
bulb
Pilling-Bedworth
Pilling Bedworth Ratio
PB = (volume of oxide of one metal atom)/(volume of metal atom)
Oxide TiO MgO Al2O3 MgO2 Ti2O3 ZrO2 Ti3O5 NiO FeO TiO2 CoO
PB 0.70 0.81 1.28 1.34 1.50 1.56 1.65 1.65 1.70 1.73 1.86
Oxide Cr2O3 FeCr2O4 Fe3O4 Fe2O3 SiO2 Ta2O5 Nb2O5 W

PB 2.05 2.10 2.11 2.15 2.15 2.50 2.68 3.40
ideal is 1.1 to 1.3

Of course thermal expansion coefficients also play a major role for the stability of oxide scales.
Alloying Effects:
different elements have

different oxygen affinity
concentration changes
diffusion rates are different
oxide layer
y contains other
metals
Example Ni
Ni-Cr-Al
Cr Al
Å Ni Cr 10 Al 5
oxide layer
y and
internal
oxidation occurs
[MB1]noch andere Eigenschaften reinschreiben?
Observations for the

Superalloy Rene N5
Diploma thesis Bensch, 2009

and submitted paper
layer number layer composition properties

1 cover oxide layer NiO, CoO thick and porous monophase layer
2 interlayer of oxides NiAl2O4 , NiTa2O6, Cr2O3 thick and porous layer consisting of two fractions
3 third oxide layer Al2O3 dense and thin monophase layer

4 γ’-free layer see Tab. 1 Al-content of 2.2 wt. %
5 γγ’ reduced layer composition in-between

in between layer number 4 and 6 reduced Al content,
content γ’
γ morphology change
6 two-phase centre region nominal composition of René N5 (Tab. 1) regular γ’/ γ structure, see Fig. 6 f)
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.
High Temperature Alloys
T > 500°C, Application in:
• energy generation
• engines (cars, trains, airplanes, ships, ... )
• chemical industry
• metallurgy
• mechanical engineering
Overview Metals
ele struc-
struc Ttrans.Tm ρ max O-solubility
max. O solubility advantages/disadvantages
m. ture [°C] [g/cm3] [at.%]
Ti α 882 4.5 31.9 + low density

hdp 1855 4.5 8 + high melting point
β krz
k + abundant
b d t available
il bl
+ low αth. (~ 10-5 K-1)
− now alloy known with adequate strength for temperatures > 600°C
− high oxygen and nitrogen solubility > 700°C, increased brittleness
− linear oxidation > 800°C
− low thermal conductivity
− ignition hazard
V krz 1910 6.1 17 − catastrophic oxidation; Tm(V2O5) = 658°C
Cr krz 1863 7.2 0.0053 − very brittle at RT; conventionally not processable
Mo krz 2623 10.2 0.03 + very high creep strength
+ lowαth, high thermal conductivity, good thermal fatigue strength
− very brittle at RT
− catastrophic
p oxidation; Tm((MoO5) = 795°C
− no long lasting coating available
W krz 3422 19.3 ≈0 + highest melting point of metals (only C with even higher Tm)
+ very high creep strength
+ low αth, high thermal conductivity, good thermal fatigue strength
− very brittle at RT
− catastrophic oxidation > 1000°C durch hohe WO3-Abdampfrate
− no long lasting coating available
University Bayreuth, Advanced High Temperature Alloys 106− very high density Uwe Glatzel, Metals and Alloys
Overview Metals
elem. structure Ttrans. ρ max. O- advantages/disadvantages
Tm [g/cm3] solubility
[°C] [at.%]
α krz 912 7.9 0.0008 + veryy ggood corrosion resistance by
y alloying
y g with Cr or ((Cr + Al))
γ kfz 1395 7.7 0.0098 + γ-structure can be stabilized down to RT (by Ni)
Fe δ krz 1538 7.4 0.029 + very good processable and weldable
+ low cost (~ 1 €/kg)
− strength at high temperatures (> 700°C) limited
Co ε hdp 422 8.8 ≈0 + very good corrosion resistance by alloying with Cr or (Cr + Al)
α kfz 1495 8.7 0.048 + Co-alloys castable in air good weldability
− only moderate hardening available
− Ni-additions necessary to stabilize fcc structure, reduces strength
Nii kf
kfz 14
1455 89
8.9 00
0.05 + broad
b d possibilities
ibili i for
f alloying,
ll i high
hi h strengthh increase
i possible
ibl
+ very good corrosion resistance by alloying with Cr or (Cr + Al)
+ processable
− relatively low melting point
−αth.
th high,
g , low thermal conductivityy
Pt kfz 1772 21.5 ≈0 + high corrosion and oxidation resistance
+ high melting point
− very high density
− very expensive (~ 33 €/g)
Evolution of materials
used in aero-engines
The earlier approach of technology
technolog transfer from military
militar to civil
ci il is
tending to switch direction.
© www.azom.com
10 000 h Life Time
10.000
Example of Intermetallic
Phases (Ni-Al-System)
Ni-Al
Ni Al Intermetallic Phases
phase structure Ttrans. ρ advantages/disdavantages
Tm [g/cm3]
[°C]
Ni3Al L12 1383 7.5 + anomalous temperature dependence of strength
+ same structure base
b than
h Ni matrix i (fcc)
(f )
+ stable for larger Al variations > 1 wt.% Al
+ ductile as single crystal
− high density
− brittle
b ittl as polycrystal
l t l (can
( be
b hindered
hi d d byb boron
b doping
d i (grain
( i
boundary strengthener)
−Al-content not sufficient to build stable Al2O3-layer Æ reduced high
temperature oxidation resistance
NiAl L10 1638 5.85 + very good oxidation resistance, since 30 wt.% Al
+ high melting point
+ low density
+ ordered structure upp to melting
g point
p
+ high thermal conductivity
+ low coefficient of thermal expansion
− extremely brittle at temperatures below 500°C (von Mises criterion
not fulfilled))
− low strength at high temperatures
NiAl, B2 Ordered
Intermetallic Phase
• At a first sight very interesting (see
g ) but despite
advantages) p many y efforts and many
y
100 Mio. US$ research money spent, up today
no bulk usage of NiAl has been achieved.
achieved
• BUT: aluminum coatings leading to NiAl

layers is heavily used.
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.
MTS-Factory
MTS Factory in Bayreuth
ground-breakingg ceremony:
g y 20.02.2008,, topping-out
pp g ceremony:
y 06.06.2008
start of production: ~ 12/2008
MTS-Factory
MTS Factory, June 2008
MTS-Factory
MTS-Factory
Processing of a Turbine
Blade
l d
FPI
X-Ray
y
F i
Feinguss, Wachsausschmelzverfahren,
W h h l f h lost
l wax investment
i casting,
i ...
Archaeological Evidence
(Bibracte) ~ 50 B.C.
ceramic mould filled with wax
cloth clip
Singly Crystal Castin in Bayreuth
at the
h Chair
h i for
f Metalsl And d Alloys
ll
Contents
11. Introduction
a) Static
b) Cyclic (Fatigue)
4.

Advanced High Temperature Alloys PDF

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Advanced High Temperature Alloys PDF

Загружено:

Авторское право:

Доступные форматы

Advanced High

lecture notes: http://www.metalle.uni-bayreuth.de then "Lehre" then "Vorlesungen",

• only alloys will be looked at (no ceramics,

500 temperature [°C] 1500 2000

500 temperature [°C] 1500 2000

500 temperature [°C] 1500 2000

U total mol = 3 ⋅ R ⋅ T Æ 0,33 eV, bzw. 32 kJ/mol bei 1000°C Arrhenius-plot

T[°C] 20 300 450 800 1000 1200 1454

T/Tm 0.17 0.33 0.42 0.62 0.74 0.85 1.00

cv 10-23 3·10-12 10-9 10-6 10-5 7·10-5 3·10-4

equilibrium vacancy concentration for nickel

QSD ≈ 17 ·kB ·Tm Æ QSDnickel ≈ 2.5 eV = 244 kJ/mol

surface diffusion Coefficient of diffusion of Th

Deff = DSD + π · δ / d · Dgrain bound.

Single crystal aluminium, oriented such that <110>{111} slip is activated.

Lytton, Shepard and Dorn, Trans. AIME 212 (1958) 220

a … coefficient of temperature conductivity

ferritic steel 45 460 7.8 13.0

bimodal (some grain

platinum-base superalloy nickel-base superalloy

dT precipitate size λT distance between precipitates

(mit Orientierungsbezug) (ohne Orientierungsbezug)

ΔGtotal = ΔGvol + ΔGboundary + ΔGstrain + ΔGdefect

total change in free enthalpy

strain enthalpy (elastic energy + dislocation line energy)

enthalpy of formation of matrix to precipitate (scales with volume)

d3 - d03 ~ D⋅t here for T/Tm ≈ 0.74

particle coarsening constant of

• most alloy properties at room temperature are time and

solid solution precipitate

Material properties of steel and

D. Siebörger, H. Knake, U. Glatzel, Mat. Sci. Eng. A298 (2001)

arrangement in low energy

primary secondary tertiary:

Pyrometer from left, optical strain

straiin rate [1/s]

1123K,, 650 MPa

Life time of stationary gas turbines > 20 years. years Assuming a

raw data creep curves:

time to failure: Å time - strain

isochrone time to failure: Å isochrone strain

• Coble creep (grain boundary diff.)

h grain size, δ thickness of grain boundary

σ ⋅ Ω ⎛ Dself diffusion π ⋅ δ⋅ D grain boundary ⎞ σΩ D eff

real geometry (non-cuboidal grains)

Austenitischer Stahl 800H

This method is applicable to

ln ε& = ln ε& 0 + n · ln σ0 + n · ln (1+ε) = const. + n · ln (1+ε)

σi = α ⋅ G ⋅ b ⋅ ρ + σsolid solution + σprecipitate

threshold stress concept (with n ≈ 3 - 4 and Qcreep = Qself diffusion):

σ1 and σ2 arbitrary external stress levels

Å Dependence of stationary creep rate on

nickel base superalloys

Internal back stress in steady state regime: σi = α ⋅G ⋅b⋅ ρ

Orowan stress in case of precipitates or particles: σOrowan ≈ G·b/L

a) cracks at grain boundaries b) cavities (micropores) at grain boundaries

micropore, difficult to detect

pear necklace like chain of

LEK-3 8.1 g/cm3

Thermal breathing of turbine blade:

εthermal = αthermal · ΔT, or: σthermal = E · εthermal

Æ orientation of single crystals in <100> direction reduces thermal stresses

Can not be covered in this lecture!