Understanding hydrogen redistribution and designing a new hydrogen extraction method

IntroductionThe Model

Application to real processesConclusion

Understanding hydrogen redistribution anddesigning a new hydrogen extraction method

Daniel Gaude-Fugarolas, Ph.D, FCPS ([email protected])

dgaude Prime Innovation SLUIndependent Research in Physical Metallurgy and Engineering

3rd UK-China Steel Research ForumRutherford Appleton Laboratory (UK)

10-11 July 2014

Daniel Gaude-Fugarolas, Ph.D, FCPS ([email protected]) Understanding hydrogen redistribution and its extraction



1 Introduction

2 The Model

3 Application to real processesAnalysis of Casting (or any cooling process)Development of a new hydrogen extraction methodAnalysis of BakingFuture work

4 Conclusion




Embrittlement prevention methods

At the design stage (material & componentengineering)

Extraction from liquid metal during refining(Vacuum, AOD, et c.)

Extraction from solid at high temperature (veryslow cooling & directional cooling:METAL2010, HSLA2011, Euromat2011, Steel &Hydrogen2014, et c. )

Extraction by treating after cooling (Bakingtreatment: METAL2014)




Embrittlement prevention methods

At the design stage (material & componentengineering)

Extraction from liquid metal during refining(Vacuum, AOD, et c.)

Extraction from solid at high temperature (veryslow cooling & directional cooling:METAL2010, HSLA2011, Euromat2011, Steel &Hydrogen2014, et c. )

Extraction by treating after cooling (Bakingtreatment: METAL2014)

Can we still do anything else to reduce the incidenceof hydrogen embrittlement in metals?




The Model

Let’s start at the beginning:Let’s try and understand the redistribution of

interstitial elements




The Model

Let’s start at the beginning:Let’s try and understand the redistribution of

interstitial elements

An interstitial diffusion model




Interstitial diffusion: Hydrogen




Model: Diffusion

Characteristics of the model:(METAL2010-2014, HSLA2011, EUROMAT2011, Steel& Hydrogen2014, et c.)




Model: Diffusion


Thermal evolution and T gradients (Heat Equation)




Model: Diffusion



Phase transitions from Liquid to α BCC (Thermodynamic Model)




Model: Diffusion




Hydrogen diffusion as random walk, driven by chemical activationgradient




Model: Diffusion




Hydrogen diffusion as random walk, driven by chemical activationgradient (... think of it as following a partial saturation gradient)




Model: Diffusion




Hydrogen diffusion as random walk, driven by chemical activationgradient (... think of it as following a partial saturation gradient)

Interstitial solubility and saturation as function of temperature andmatrix phase or trap type and distribution




Model: Trapping

Each trap type characterisedby its characteristic releaseenergy barrier

Interaction of each of thetrap sites with lattice

Exchange with atmosphereat free surfaces: localequilibrium across thesurface (Sievert’s law)

E

E t

d

��

��

Trap_1 Trap_i Trap_n.. ..

Atm

osp

her

e

Lattice




Analysis of Casting (or any cooling process)Development of a new hydrogen extraction methodAnalysis of BakingFuture work

Application to real processes

What can we do with this model? (Very briefly)

Analysis of Casting (or any cooling process)

Development of new H extraction method

Analysis of Baking

... and much more!!






0

1

2

3

4

5

6

-12.5 -5 0 5 12.5 0

1

2H c

onte

nt /p

pm

Thickness /cm

Fast cooling Slow cooling

----

----

----

-cen

ter

piec

e---

----

-

0s 30s

150s 600s

1200s1700s

0h1h6h

24h42h 0

5

10

15

20

25

30

35

40

45

-12.5 -5 0 5 12.5 0

5

10

15

H p

artia

l sat

urat

ion

Thickness /cm


----

----

---c

ente

r p

iece

----

----

---

0s 600s 900s

1200s1500s1700s

0h1h6h

24h42h

Steel: 25cm thick with 2ppm

Effect of cooling rate:

Fast cooling vs. Slow cooling






0

1

2

3

4

5

6

-12.5 -5 0 5 12.5 0

1

2H c

onte

nt /p

pm

Thickness /cm


----

----

----

-cen

ter

piec

e---

----

-

0s 30s

150s 600s

1200s1700s

0h1h6h

24h42h 0

5

10

15

20

25

30

35

40

45

-12.5 -5 0 5 12.5 0

5

10

15

H p

artia

l sat

urat

ion

Thickness /cm


----

----

---c

ente

r p

iece

----

----

---

0s 600s 900s

1200s1500s1700s

0h1h6h

24h42h


Effect of cooling rate:

Fast cooling vs. Slow cooling

Effect of thickness: (5-50cm)

Effect of FCC to BCC transformationtemperature:

Steel A: 700oC vs. Steel B: 450oC 1.5

2

2.5

3

3.5

4

4.5

5

5.5

6

5 10 20 30 40 50

Max

H c

onte

nt /p

pm

Thickness /cm

Alloy AAlloy BStart H





Development of a new hydrogen extraction method


“Understanding hydrogen redistribution during steel casting, and its effective extraction bythermally induced up-hill diffusion”

D. Gaude-Fugarolas, in: Journal of Iron and Steel Research International 18 supl.1.1 (2011)159–163.

Also at proceedings: High Strength Low Alloy (HSLA2011) International Conference, Beijing,China, 2011.






A severe temperature gradient forces hydrogen to flow towards the coreregion of a component, where it can reach severe supersaturation







Actually, NO!!. The temperature gradient forces hydrogen to flowtowards higher temperature regions!!







Actually, NO!!. The temperature gradient forces hydrogen to flowtowards higher temperature regions!!

Then, why don’t we try instead to redirect the hydrogen fluxtowards the surface?

(and eventually get rid of it)






Standard casting operation: Interstitialelement flux creates enriched regions atthe core of the piece

Pouring cup

Feeder

Casting

Riser






Standard casting operation: Interstitialelement flux creates enriched regions atthe core of the piece

Pouring cup

Feeder

Casting

Riser

Modified casting operation with a severethermal gradient towards the surfacebeing imposed, eliminating interstitials

Pouring cup

Feeder

Casting

Riser

Heating Element






0

1

2

3

4

5

-12.5 -5 0 5 12.5

H c

onte

nt /p

pm

Thickness /cm

Hea

ted

Surf

ace

0s60s

300s1200s3600s8400s 0.01

0.1

1

10

-12.5 -5 0 5 12.5

H p

artia

l sat

urat

ion

Thickness /cm

Hea

ted

Surf

ace

0s60s

300s1200s3600s7200s7600s8000s8400s


Surface I: Fast cooling

Surface II: Kept at high temperature(i.e. 1500oC)

Temperature gradient maintained 2h






0

1

2

3

4

5

-12.5 -5 0 5 12.5

H c

onte

nt /p

pm

Thickness /cm

Hea

ted

Surf

ace

0s60s

300s1200s3600s8400s 0.01

0.1

1

10

-12.5 -5 0 5 12.5

H p

artia

l sat

urat

ion

Thickness /cm

Hea

ted

Surf

ace

0s60s

300s1200s3600s7200s7600s8000s8400s


Surface I: Fast cooling

Surface II: Kept at high temperature(i.e. 1500oC)

Temperature gradient maintained 2h

Final H content 0.99ppm!!i.e 50% Reduction!!

Partial Saturation duringtreatment below 1.0 !!





Analysis of Baking

Analysis of Baking

“On the effectiveness of baking as hydrogen embrittlement reduction treatment”D. Gaude-Fugarolas, in: Proceedings of METAL2014, 21-23 May, Brno, Czech Republic, 2014.





Treatment after cooling: Baking

Storage of the metal components in over at low warmtemperature (typically 150-230 oC) for a long period,up to 24-48 hours.

The treatment aims to reduce internal stresses and toreduce hydrogen content.

If treatment is not performed immediately aftercasting and cooling, it might become ineffective.

Effectiveness of treatment varies.

Why does effectiveness of baking vary so much?





Analysis of Baking

Hydrogen content: 1 ppm

Baking: 12h at 190oC or 300oC

Sites: Lattice, Dislocation, Grain boundary,Precipitate (& Desorption to atmosphere)

Steel A: Allotriomorphic Ferrite (725oC), largegrain, low dislocation density

Steel B: Bainitic/Martensitic Ferrite (450oC),small grain, high dislocation density





Analysis of Baking

Hydrogen content: 1 ppm

Baking: 12h at 190oC or 300oC

Sites: Lattice, Dislocation, Grain boundary,Precipitate (& Desorption to atmosphere)

Steel A: Allotriomorphic Ferrite (725oC), largegrain, low dislocation density

Steel B: Bainitic/Martensitic Ferrite (450oC),small grain, high dislocation density

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(a0)

Steel ALATTICE

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(a0)

Steel ALATTICE

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(b0)

Steel B

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(b0)

Steel B

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(a1)

DISLOCATION

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(a1)

DISLOCATION

0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(b1)0.001

0.005

0.025

0.125

0.625

0 6 h 12 h

H c

onc

/ppm

Time /s

(b1)

0.0000

0.0001

0.0010

0.0100

0.1000

0 6 h 12 h

H c

onc

/ppm

Time /s

(a2)

Steel AGRAIN

0.0000

0.0001

0.0010

0.0100

0.1000

0 6 h 12 h

H c

onc

/ppm

Time /s

(a2)

Steel AGRAIN

0.0000

0.0001

0.0010

0.0100

0.1000

0 6 h 12 h

H c

onc

/ppm

Time /s

(b2)

Steel B

0.0000

0.0001

0.0010

0.0100

0.1000

0 6 h 12 h

H c

onc

/ppm

Time /s

(b2)

Steel B

0.35

0.40

0.45

0.50

0.55

0 6 h 12 h

H c

onc

/ppm

Time /s

(a3)

PRECIPITATE

0.35

0.40

0.45

0.50

0.55

0 6 h 12 h

H c

onc

/ppm

Time /s

(a3)

PRECIPITATE

0.35

0.40

0.45

0.50

0.55

0 6 h 12 h

H c

onc

/ppm

Time /s

(b3)0.35

0.40

0.45

0.50

0.55

0 6 h 12 h

H c

onc

/ppm

Time /s

(b3)

“On the effectiveness of baking as hydrogen embrittlement reduction treatment”D. Gaude-Fugarolas, in: Proceedings of METAL2014, 21-23 May, Brno, Czech Republic, 2014.





Conclusions on Baking

The success of baking depends of the interaction of hydrogencontent, microstructure and trap distribution.

Some microstructures already saturate with low hydrogen contents,making baking useless.

In general, baking is only effective when the microstructure is farbelow saturation.

The baking temperature needs to be tailored to trap type and Hdistribution.

Some defects may even increase their H content during baking.





Future work

This is still work in progress...

Surface defects

Crack initiation

H absorption from atmosphere

To be presented at, may be, METAL2015, Brno, Czech Republic, May 2015.




Conclusion

Conclusion




Conclusion

To summarise...

Small amounts of hydrogen can endanger the integrity of criticalmetal components.

A physical model has been presented offering a correct description ofhydrogen redistribution during manufacturing operations.

Several methods exist to reduce hydrogen embrittlement in metal,but not always successful.

A new method has been presented to reduce hydrogen content usingof imposed temperature gradients.

(Patent filed in US, Europe, China, et c., Already awarded or inprocess, and open for licensing).




Thanks

Thank you for your attention!!

For more information, please visit (or email):[email protected]
