Characterization of Internal Oxidation - NISTcecamp/TMS2005/TMS 2005...0.05 0.2 0.5 0 B S O C K β=...

Post on 03-Aug-2021

2 views 0 download

Transcript of Characterization of Internal Oxidation - NISTcecamp/TMS2005/TMS 2005...0.05 0.2 0.5 0 B S O C K β=...

1

TMS 2005 --- Diffusion Symposium

Characterization of Internal Oxidation

Yali Li 1 and John E. Morral2

1 University of Connecticut, Storrs CT 06269 USA2 Ohio State University, Columbus OH 43210 USA

2

Internal Oxidation Characterization

Cu-11%NiO2

unoxidized region

Internal oxidation region

Phases present --- XRDOxide fraction profile --- Image analysisConcentration profile --- EDXThickness of internal oxidation region

3

Oxide Fraction Profile

Simulation results for saturated case in literature :

Rel

ativ

e ox

ide

frac

tion

Reduced distance

0

β→1

0.05

0.2

0.5

0B

SOCCK

* E. K. Ohriner and J. E. Morral, Scripta Metallurgica, 13 (1979) 7.

4

Oxide Fraction Profile

Experimental results in literature:

Fe-5%Cr

Fe-Cr alloys internally oxidized at 1000 °C for 20 hours with Po2=8.7×10-17atm.** O. Ahmed and D. J. Yong, Electrochemical Society Proceedings, 38 (1999) 77.

5

Oxide Fraction Profile

Cu-7% Ni internally oxidized at 900° C in Rhines Pack

0

1000

2000

3000

4000

5000

10 30 50 70 90 110

Inte

nsity

NiOMatrix

XRD patternsSEM

6

Oxide Fraction Profile

Image analysis: oxide area fraction profile

2µm

80µm

grey scale binary mode

7

Oxide Fraction Profile

NiO area fraction versus distance

0 10 20 30 40 50 60 70 800

5

10

15

20

25

Distance from surface (micron)

Prec

ipita

te a

rea

perc

ent

Cu-7% Ni alloys internally oxidized at 900° C for 1 hour

8

Oxide Fraction Profile

Simulation results: saturated case

0.00

0.25

0.50

0.75

1.00

0 0.4 0.8 1.2 1.6

=0.22 =0.03β β

Rel

ativ

e ox

ide

frac

tion

Reduced distance

DICTRA results0

β→1

0.050.2

0.5

Ohriner and Morral

9

Oxide Fraction Profile

Simulation results: unsaturated case

Effect of β on relative oxide fraction profiles

x/x*

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

=0.001

=0.01

=0.1

=0.3

=0.5

=0.9

β

Rel

ativ

e ox

ide

frac

tion

β

β

ββ

β

10

Oxide Fraction Profile

Summary:

When alloys are saturated with oxygen, oxide mole

fraction decreases asymptotically to zero and the position

of moving boundary cannot be defined.

When alloys are unsaturated with oxygen, the moving

boundary between internal oxidation region and

unoxidized region is well defined.

11

Concentration Profiles

The classic model:

Moving boundary

Internal oxidation regionα(A+ dissolved O)+ BOν

Unoxidized regionα (A+ B)

Oxi

dizi

ngat

mos

pher

e

Assumptions:All the solute B is consumed to form BOv in the internal oxidation region.

The concentrations of dissolved B and O at the moving boundary are zero.

* C. Wagner, Z. Electrochem, 63 (1959) 772.

12

Concentration Profiles

The classic model:O

xidi

zing

atm

osph

ere

Moving boundary

Internal oxidation regionα(A+ dissolved O)+ BOν

Unoxidized regionα (A+ B)

Distance from surface (x)

Con

cent

ratio

n

x*

αOC

αBC

BC

OC

0BC

SOC

ααBO CCK ×= = 0

* C. Wagner, Z. Electrochem, 63 (1959) 772.

13

Concentration Profiles

Experimental results in literature:

Distance from surface (cm)

Haf

nium

( w

t% )

α+HfO2α

Nb-Hf alloy

* D. L. Corn et al. Oxidation of Metals, 35 (1991) 139

14

Concentration Profiles

Concentration profiles under local equilibrium conditions

distance

α + β α

* W.J. Boettinger et al. Acta Metall.Mater. 48 (2000) 481-492.* J. E. Morral and H. Chen, Scripta Mater. 43 (2000) 699-703.

Matrix concentration profile, Ca,

is continuous in slope and value.

Average profile, , and

matrix profile, Cα, meet at the

moving boundary.

βα +C

Con

cent

ratio

n

Cα+β

15

Concentration Profiles

The classic model versus Local equilibrium model

x*

0BC

Distance from surface (x)

Con

cent

ratio

n

αOC

Moving boundaryOxidized region Unoxidized region

αBC

BC

The classic model: BC αBC

K = 0

BC αBCLocal equilibrium model:

16

Concentration Profiles

DICTRA simulation results: saturated case

0.00

0.05

0.10

0.15

0.0 1.0 2.0 3.0 4.0 5.00.000

0.005

0.010

0.015

Distance x ( ×10-4m )

Con

cent

ratio

n

CB (x)

CO (x)

Simulation conditions: β =0.96, DB/DO=10-3

2% oxide

Internal oxidation region

Unoxidized region

17

Concentration Profiles

Concentration profiles of Cu-10% Ni alloys oxidized at 950 °C

-50 0 50 100 150 200 250 300 350 400 450 5000

2

4

6

8

10

12

3 hours 10 hours 30 hours

Ni%

in

mat

rix

Distance from surface (micron)

Estimated position of oxidation frontier

18

Concentration Profiles

Error Function Model simulation results: unsaturated case

8.5

9.0

9.5

10.0

10.5

0.00 0.02 0.04 0.06 0.08 0.100.05

0.10

0.15

0.20

0.25

CB (x)

CO (x)

CB (x)

Con

cent

ratio

n

Reduced distance from surface ( λ )

moving boundary

Simulation conditions: β =0.96, DB/DO=10-3

19

Concentration Profiles

Summary:

Concentration profiles in the classic model are in error because

they are not supported by experimental results in the literature

or in this study and have no theoretic basis.

When K = 0, there is no long-range diffusion of solute B, thus

the enrichment phenomena proposed by the classic model

doesn’t occur.

Near the boundary between oxidized and unoxidized region,

matrix solute concentration approaches initial solute

concentration rather than zero as assumed by the classic model.

20

Thickness of Internal Oxidation Region

Cu-10%Ni alloys oxidized at 950 °C

3 hours

1 hour

10 hours

30 hours

Oxidation time (hour)

Thi

ckne

ss2

(×10

4µm

2 )

0

2

4

6

8

10

12

0 5 10 15 20 25 30 35

21

Thickness of Internal Oxidation Region

Cu-Ni alloys oxidized at 950 °C for 3 hours2.5% Ni

5.0% Ni

10.0% Ni

20.0% Ni0

2

4

6

8

0 10 20 30 40 50

Thi

ckne

ss2

(×10

4µm

2 )

01 NiC

22

Conclusions (1)

For saturated cases:

(a) Because oxide fraction decreased asymptotically to zero, there

was no distinct boundary between the oxidized and

unoxidized regions.

(b) For DO>>DB, when oxide fraction approaches zero, matrix

solute concentration is close to initial solute concentration.

(c) DICTRA predicted internal oxidation under local equilibrium

for 0.03 < β < 0.22 for alloys saturated with oxygen.

23

Conclusions (2)

For unsaturated cases:

(a) The moving boundary between internal oxidation region

and unoxidized region is well defined.

(b) For DO>>DB, solute concentration in matrix at the

moving boundary is close to initial solute concentration.

(c) When β → 1, Error Function Model can be used to

model internal oxidation for unsaturated alloys.

24

Acknowledgements

Major advisor: Dr. John E. Morral

Dr. Ronald N. Caron at Olin Corporation

Dr. Carelyn E. Campbell at NIST

Dr. Caian Qiu at Ques Tek

UConn staff: Mary Anton, Fred Massicotte

All members of UConn diffusion group

National Science Foundation

25

26

Applications

Design alloys with improved oxidation resistant

Design intermetallic alloys capable of protective scale formation ( e. g. Al2O3, Cr2O3 and SiO2)*

Synthesize functional ceramic surface structures (e.g. nitride and carbide catalysts Co3Mo3N or Co3Mo3C) by gas-metal reactions (oxidation, nitridation, carburization, etc.)*

* M. P. Brady and P.F. Tortorelli, Intermetallics, 12(2004) 779-789

27

Approach: Experimental studies

Rhines Pack: Cu (shot) + Cu2O (powder) in a stainless steel tube

Sample size : 10×10 ×20 (mm)

The tube was welded at the ends in argon atmosphere

Cu-Ni alloy

Cu2O powder

Cu shot

Stainless steel tube

28

Experimental results: Optical microstructures

---- Cu atom --- Cu+1 --- O-2

Cu shot +Cu2O powder

Stainless steel tube

Cu-Ni alloy

(1) The mixture of Cu and Cu2O in a Rhines Pack

Two Cu sources for the new Cu layer:

Ni : 6.59 cm3/mole NiO: 11.15 cm3/mole

(2) Internally oxidized Cu-Ni alloys

29

Experimental results: EDX

Concentration profiles of Cu-10% Ni alloys oxidized at 950° C

Estimated position of oxidation frontier

0 50 100 150 200 2500

2

4

6

8

10

12

14

16N

i %

Distance from surface (micron)

Ni% in matrix Average Ni%

30

Experimental results: EDX

New Cu layer at the surface:

4000

4500

Quantitative results:

Cu --- 99.69 wt%Ni --- 0.31 wt%

Kα1 Kβ1 Lα1 Lβ1

Cu 8.046 8.904 0.93

0.851

0.95

Ni 7.477 8.263 0.863

31

Experimental results: Phase identification (XRD)

XRD patterns of Cu-20%Ni – before oxidation

0

500

1000

1500

2000

2500

3000

3500

10 20 30 40 50 60 70 80 90 100 110

2 theta

Inte

nsity

(1 1 1)

(2 2 0)

(2 0 0)

(3 1 1)(2 2 2)

Cu-20%Ni

32

Experimental results: Phase identification (XRD)

XRD patterns of Cu-20%Ni – after oxidation

0

1000

2000

3000

4000

5000

10 30 50 70 90 1102 theta

Inte

nsity

(1 0 1) (1 0 4)

(0 1 2)

(1 1 3)(0 2 4)(2 0 2)

NiO

33

Experimental results: Phase identification (XRD)

NiO crystal structure:

34

Experimental results: EDX

Concentration profiles of Cu-Ni alloys oxidized at 950° C for 3 hours

-50 0 50 100 150 200 250 300 350 4000

4

8

12

16

20

24

Ni%

in m

atrix

Distance from surface (micron)

20 wt% Ni 10 wt% Ni 5 wt% Ni

Estimated position of oxidation frontier

35

Background: Existing models of internal oxidation

(a)

(b)

(c)

(a) C. Wagner, Z. Electrochem, 63 (1959) 772. (b) J. A. Nesbitt, Oxidation of Metals, 44 (1995) 309. (c) G. Bohm et al. Acta Metall.,12 (1964) 641

36

Background: Venn diagram

A:Zero solubility

productK=0

B: DB = 0

D:Non-Local equilibrium

The classic mode = )()( BADA ∩∪∩

CDLocal equilibrium model =

37

Background: Local equilibrium assumptions

The compositions and amount of

each phase are given by the local

average composition, the phase

diagram, and the lever rule.

Nucleation and growth rate of

precipitates are so rapid that only

long range diffusion need be

considered.

βCC

αCC

αBC β

BC

α

β

A B

C

CC

BC

α+β

38

Simulation results: EFM

Diffusion paths:

0.00

0.05

0.10

0.15

0.20

0.25

0.30

9.6 9.7 9.8 9.9 10.0 10.1

Diffusion path in alpha

Diffusion path in alpha+oxidealpha/alpha+oxide phase boundary

tie line

CO

CB

α

α + oxide

Simulation conditions: , , , ==−=5106.9 −×=K

39

Simulation results: DICTRA

Diffusion path:

α + oxide

α

Simulation conditions: , , , ==−=5106.9 −×=K

40

DICTRA simulation

An ideal A-B-O solid solution system

Three elements: A - solventB - soluteO - fast diffuser (DO>>DB)

Two phases:matrix phaseline compound (BO)

41

Simulation results: DICTRA

Effect of on oxide mole fraction:0BC

Distance x ( m)410−×

f oxi

de

0.00

0.10

0.20

0.30

0.40

0.50

0.0 1.0 2.0 3.0 4.0 5.0

%100 =BC

%150 =BC

%200 =BC

Simulation conditions: , , =−=5106.9 −×=K

42

Simulation results

Comparison of EFM and DICTRAEFM DICTRA

Application Unsaturated case Saturated case

Boundary conditions

= Constant = Constant

[Deff] Constant Concentration dependent

Phase boundary Linear

β β →1 0.03 < β < 0.55

Local equilibrium conditions Satisfied Satisfied

KCC OB =

SOC

0=BJx

CDJ BBB ∂∂

−=

SOC

43

Interesting diffusion paths

two-phase region

Zigzag shapeSingle-phase region

Serpentine shape

C1→

C2→

α

βα + β

C2→

C1→

⎥⎦

⎤⎢⎣

2221

1211

DDDD

Major eigen vector direction

Gibbs Phase law: f = c – p + 2