ODS Alloy Development...ODS Alloy Development Ian Wright , Bruce Pint, Claudette McKamey Oak Ridge...
Transcript of ODS Alloy Development...ODS Alloy Development Ian Wright , Bruce Pint, Claudette McKamey Oak Ridge...
1
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
ODS Alloy Development
Ian Wright , Bruce Pint, Claudette McKameyOak Ridge National Laboratory
17th Annual Fossil Energy Materials ConferenceBaltimore, Maryland, April 22-24, 2003
2
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Barriers:• Joining• Highly-directional properties: for tubes, transverse strength << axial• Unusual mechanical behavior; strain-rate sensitivity/mode of failure• Cost
Options:• Unconventional joining approaches• Innovative processing to obtain the desired microstructure• Improved quantification of alloy properties and characteristics so that there are no surprises
Scientific approach:• Understand and quantify all available routes for joining• Develop mechanistic understanding for understanding how to control the alloy microstructure; and of the oxidation behavior
Goal: Facilitate the Exploitation of ODS Alloys
3
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
• Creep strength to temperatures > conventional high-temperature alloys• Potential for use to temperatures where typically ceramics are considered • Excellent oxidation resistance• Resistance to sulfidation; steam oxidation• Current Focus: ODS-FeCrAl alloys
Related work:• Special Metals Inc: ODS tubing• European COST programs• SBIR at MER Corp.• ARM programs:
–Foster Wheeler–UCSD–U. Liverpool
• ORNL: ‘nano-clusters’
Why ODS Alloys?
1
10
100
1000
600 700 800 900 1000 1100 1200
Ave
rage
100
kh S
tress
Rup
ture
Stre
ngth
, MP
a
Metal Temperature,°C
CurrentAverage
TheoreticalMaximum
ODS-FeCrAls(longitudinal)
USC
HT-HX
Afte
r Sta
rr &
Shi
bli,
2002
4
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Alloys of Interest
ORNL developmentY2O3-Al2O3Ti,Si15.92.2BalODS-Fe3Al
oxidation comparator‘ZrO2-Al2O3’Ti,Si5.520BalKanthal APM
Dour MetalY2O3-Al2O3Ti4.516.5BalODM751
PlanseeY2O3-Al2O3Ti5.520BalPM2000
956 modificationY2O3-Al2O3Ti,Si5.721.6BalMA956H
Special Metals Inc.Y2O3-Al2O3Ti4.520BalMA956
REotherAlCrFeRemarksComposition, weight percentAlloy
5
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Presentation Content
• Joining
• Temperature limits– fireside/steam-side compatibility
• Mechanical properties– transverse (hoop) strength
• ODS-specific issues– strain-sensitivity/mode of failure
6
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Joining of ODS Alloys•Joining must avoid
– redistributing the Y2O3 dispersed phase– changing the grain structure size/shape/orientation
•Challenges:– fusion processes: probably a last resort
brazing in COST-522 program– friction/inertia welding: distortion of microstructure– diffusion bonding
TLP: successfully demonstrated on other ODSplasma-assisted diffusion bonding: MER Corp
– others:explosive bonding: successfully demonstrated (COST-501) pulsed magnetic weldingmechanical--threading + brazing
7
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Tested Configurations for ODS Joints
Bayonet tubes (COST-522)
ODS tube
Ceramic tube
Joints
HeaderAIR INAIR OUT
ODS tube
Header
‘Safe’ end
Explosive bond
British Gas ‘Harp’ Joint
Conventional weld
ODS Corner block
ODS tubes
TLP bond
TLP joint: HiPPS program
8
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Other Possible Configurations for ODS Joints
Reinforced joints
Threaded & brazed Overlapped and brazed Inertial welded
tube-to-flange joint
9
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Inertia welding of MA-956 tubes
• 63.5 mm diam. x 7 mm wall thickness, unrecrystallized MA-956 tube
• mechanically robust joints have be produced in using inertia welding
• process window was determined based on the integrity of the joint in bend tests in coupons cut from joined tubes
• reproducibility of joining parameters is excellent
B. Kad/Interface Welding
10
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Very sharp demarcation of deformed microstructure after inertia welding
JOINT
Tube outer surface
B. K
ad, 2
003
11
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Plasma-Assisted Diffusion Bonding
MA956
• MER Corp-SBIR-II
• ‘clean’ joint
• thin joined zone
• apparent grain continuity
Bond line
12
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Bond line
AlTi
Y
O
BSE
SE
• EPMA suggests an accumulation of Ti, Al, Y, and O along the bond line
• predominantly Ti• alloy contains approx. 0.5 %Ti
Bond-line exhibits Ti-rich precipitates
13
TEM indicates discrete particles
Grain/Side 1
Imag
e K
arre
n M
ore
(853
)
BoundaryParticles alongboundary
0.5 µm
Grain/Side 2
14
Particles are TiC and Al2O3
Al2O3 particles at boundary
TiC particles at boundary
Al O
Ti C
0.25 µmImage Karren More (858 + maps 859…)
15
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
TLP Joining: Collaborative Effort
J. Hurley/N.A. Bornstein, 2003
MA956
•EERC-N.A. Bornstein-ORNL
•TLP approach based on concepts demonstrated for HiPPS joints
•special considerations for application to an alumina scale-forming alloy
•proof-of-concept TLP alloy
Bond line
Extent of interdiffusion
16
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Temperature LimitsThe basis for modeling the oxidation-limited lifetimes of these alloys is
relatively straightforward, since:
• they form essentially Al2O3 scales that are uniform in thickness
• there is negligible internal attack (life can’t be equated to section thinning)
• the Al concentration gradient in the alloy is flat until very near the end of life
As a result, it is possible simply to equate the oxidation lifetime to the rate of consumption of the available Al to form the alumina scale:
Life = Al available for oxidation/oxidation rate
17
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
The oxidation kinetics of these alloys have a characteristic form
P1 P2
0
5
10
15
20
0 200 400 600 800 1000 1200 1400 160
MA956H1300°C
Tota
l Mas
s Cha
nge,
mg/
cm2
Time (hr, in 100 hr cycles)
'Breakaway'(end of life)
Parabolic oxidation 'stage 2'
Linear oxidation'stage 3'
18
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Current ModelThe current expression of the model is:
tb = {[S*10-4*ρA*Aτ*e-Qτ/(1.987*T)]2/(3600*A2*e-Q2/(1.987*T))} + {[1/(3600*M)*(V/A)*(ρM/(A3*e-Q3/(1.987*T)))*[(CBo-CBb) – M*S*10-4*(A/V)*(ρA/ρM)*Aτ*e-Qτ/(1.987*T)]} hours
Input required is:1. Alloy data: ρM; CBo; CBb (need to measure CBb)2. Oxide data: ρA; M; S; (constants based on oxide/alloy stoichiometry)3. Alloy oxidation descriptors: Arrhenius data A2, Q2; A3, Q3, and Aτ, Qτ
4. The metal temperature (T), and the component size (V/A)
19
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Summary of Oxidation Kinetics
P1 P2
-13.5
-13
-12.5
-12
-11.5
-11
-10.5
-10
0.6 0.65 0.7 0.75 0.8
log
k 2
1000/T°K
ODS-Fe3Al
MA956
MA956: k2 = 53 x e -86,600/RT
ODS-Fe3Al: k
2 = 28 x e -85,554/RT
13001250 1200 1100 1000Temperature, °C
Parabolic oxidation ‘stage 2’
-10.5
-10
-9.5
-9
-8.5
-8
0.6 0.65 0.7 0.75 0.8
log
k 3
1000/T°K
ODS-Fe3Al
MA956H
MA956H: k3 = 0.2 x e -56,900/RT
ODS-Fe3Al: k
3 = 3 x e -62,436/RT
13001250 1200 1100 1000Temperature, °C
Linear oxidation ‘stage 3’
Some alloys haven’t run long enough to establish the Stage 3 oxidation rate
20
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Calculated Lifetimes
P1 P2
100
1000
104
105
0.2 0.4 0.6 0.8 1
V/A, mm
1200°C
1300°C
1100°C
1000°C
1250°C
Oxi
datio
n Li
fetim
e, h
Dashed Lines: predicted lives for MA956H using 2-stage model
21
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Calculated vs Observed Lifetimes
P1 P2
100
1000
104
105
0.2 0.4 0.6 0.8 1
V/A, mm
Dashed Lines: predicted lives (2-stage model )
1200°C
1300°C
Data Points: observed lifetimes
1100°C1000°C
1250°C
Data for MA956H
22
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Calculated vs Observed Lifetimes
P1 P2 Reasonable predictions for 1100-1300°C
100
1000
104
105
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Oxi
datio
n-Li
mite
d Li
fetim
e, h
r
V/A, mm
Dashed Lines: predicted lives (2-stage model)
1200°C
1300°C
Data Points: Observed Lifetimes
1100°C
1000°C
1250°C
Data for MA956
23
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Calculated Lifetimes at 1100°C
P1 P2TUBE 1in OD x 0.1in (25.4 x 2.54 mm), V/A = 0.64 mm
6000
8000
104
3 104
5 104
0.2 0.4 0.6 0.8 1
V/A, mm
ODM751
MA956MA956H
ODS-Fe3Al
Data for 1100°C/2012°F
PM2000
KanthalAPM
Based on CBb values:
FeCrAls--0.001
Fe3Al--0.056
• CBb is difficult to measure
• don’t know if it is T-dependent
24
Cross section of scale on MA956H
Karren More (875 + 876)
Al2O3 scale
1 µm
Alloy substrate
voids in Al2O3 scale(not many)
100h, 1200°C, air
25
Alloy-oxide interface on MA956H
Ti-enriched at interface
Ti
Y-containing particles in alloyThere is also Y-enrichment at Al2O3
grain boundaries (not shown)
Y
0.25 µm
Karren More (879 + maps 880…)
26
Scale cross section on MA956
thicker Al2O3 scale than on 956H
Karren More (881 + 882)
1 µm
voids in Al2O3 scale(many more )
TiC particleat interface
elongated Y-Ti-Si-C grain(no oxygen!) ~4 mm longbetween Al2O3 grains (red bars show ends of grain)
Alloy substrate
100h, 1200°C, air
27
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Current approach under-predicts oxidation lifetime
• Predictions should be conservative!
• Are lab results overly affected by specimen shape?
• V/A is a ‘shape factor,’ but doesn’t discriminate among parallelepipeds
• Other shapes:
0.80
0.64
1.13
0.67
V/A mm
452400523Cylinder
46068612.5231.6Standard parallelepiped
28043512.5141.6Parallelepiped-2
283
Volume mm3
353
Surface area mm2
151.6Disc
Width mm
Length mm
Thickness mm
Shape
28
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Effect of Specimen Shape
P1 P2
0
5
10
15
20
25
30
35
0 500 1000 1500 2000 2500 3000
MA956, 1250°C
Tota
l Mas
s Gai
n, m
g/cm
2
Time, h
Large p'ped
P'piped(square)
Disc
Cylinder
PLAN
CROSS SECTION
some shapes make less efficient use of the Al reservoir?
29
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Possible Shape Factors
Shortest diffusion path from center of specimen thickness to outer surface (dmin) is given by locus of surface of a sphere of radius = d/2
dDe = diffusion length to an edge
dDc = diffusion length into a corner
Longest diffusion path is from the center of mass (= dDcm)
ddDc
dDcmdDe
center of mass
l
w
dmin = d/2d/2
30
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Shape Factors: Diffusion Lengths
0.707 x d
—
0.707 x d
EdgedDe
0.5 x sqrt(d2 + diam2)0.707 x dDisc
0.5 x sqrt(l2 + diam2)0.707 x diamCylinder0.866 x d
CornerdDc
0.5 x sqrt(l2 + w2 + d2)
Center of MassdDcm
P’piped
Shape
d = specimen thickness I = specimen lengthw = specimen width
31
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Summary
• Joining–3 routes are being evaluated
– inertia welding--controlled microstructural distortion– plasma-assisted diffusion--clean joints (?); examining reinforced design– TLP bonding--questions about amount of residual elements and their effects
–significant effort on other techniques in the Vision 21-SM project–Temperature limits
–generating data for all available ODS-FeCrAls–have an initial working model for life prediction
– continuing long-term exposures to validate and improve the model– issues of over-prediction, and the influence of shape
– lifetimes of 30-70kh at 1100°C in air for typical tube sizes–continuing to generate data in steam, other environments
32
OAK RIDGE NATIONAL LABORATORYU. S. DEPARTMENT OF ENERGY
Sensitivity to CBb
P1 P2Don’t know if CBb is T-dependent
104
105
0.2 0.4 0.6 0.8 1
V/A, mm
Data for ODS-Fe3Al, 1100°C/2012°F
PM2000
KanthalAPM
CBb
1%2%3%4%5%
Tube2.5 mm (0.1 in)wall thickness
PM2000, APM: CBb
= 0.1%