Dynamic Coercivity of Advanced Ferritic Steel

during Long-Term Isothermal Ageing

C. S. Kim1;*, Cliff J. Lissenden1, I. K. Park2 and K. S. Ryu3

1Engineering Science and Mechanics, The Pennsylvania State University, University Park, PA 16802, USA2Mechanical Engineering, Seoul National University of Technology, Seoul, 139-743, Korea3Korea Research Institute of Standards and Science, Daejeon 305-600, Korea

The object of this study is to characterize the microstructural evolution in advanced ferritic steel for power plant during long-termisothermal ageing by measuring the reversible permeability. Ageing was observed to coarsen the tempered carbide (Cr23C6), generate the Laves(Fe2W) phase, and reduce the mechanical strength. The dynamic coercivity decreased monotonously during isothermal ageing. The decrease incoercivity physically depends on the domain wall movement, related with the domain wall pinning by nonmagnetic particles and thedislocations. The experimental results show that the dynamic coercivity is very sensitive to the damage accumulations due to isothermal ageingof advanced ferritic steel. [doi:10.2320/matertrans.M2009170]

(Received May 8, 2009; Accepted August 18, 2009; Published October 7, 2009)

Keywords: advanced ferritic steel, isothermal ageing, reversible permeability, dynamic coercivity

1. Introduction

Advanced ferritic steels are usually based on temperedmartensite and their high temperature strength is sustained byprecipitation hardening of MX carbides or carbonitrides aswell as by solid solution hardening of Mo, V, Nb, W and Ti.Previously extensive studies have shown two types ofdominant carbides in as-tempered state, MC and M23C6, inhigh Cr ferritic steels.1–5) The M23C6 carbides are present asCr rich carbide (Cr23C6) and all of the others are the Nb/Vrich carbide (NbC/VC). Recently, the Laves (Fe2W, Fe2Mo)phases have drawn a growing interest as strengthener toimprove the creep resistance of the steels.6,7) Even thoughthese high strength steels have been widely used for turbinerotor in power plants, material degradation is obviouslyknown to occur due to their continuous exposure to hightemperatures (e.g., thermal ageing and creep). Therefore,the characterization of damage mechanism is crucial foraccessing the life and structural integrity of the components.

The objective of this work is to evaluate the micro-structural changes in advanced ferritic steel due to long-termisothermal ageing, using metallurgical techniques, and alsoto measure the reversible permeability using applicablenondestructive evaluation technique. The influence of themicrostructural features on the reversible permeability is alsoinvestigated.

2. Experimental Procedure

The ferritic 12Cr steel was fabricated by the vacuuminduction melting process, and subsequently forged at1100C. The forged specimens were tempered at 700Cafter normalizing at 1100C. The chemical composition isshown in Table 1. The specimens were aged at 650C for upto 8000 h. Microstructural observation was performed usingscanning electron microscope (SEM) after etching withVilella’s reagent. The thin film specimen was carefully

prepared by the twin jet polishing. The lath was observedusing transmission electron microscope (TEM). In addition,the structure and composition of precipitates were identifiedby energy dispersive spectroscopy (EDS) and selected areadiffraction (SAD) pattern after preparing thin film by carbonreplication technique. X-ray diffraction (XRD) patterns ofelectrolitically extracted residues were analyzed using a highresolution X-ray diffractometer (Cu K radiation).

In order to measure the reversible permeability on thesurface of specimens, a ferrite yoke was used as shown inexperimental setup of Fig. 1.8) The yoke was wound by pick-up coils, an AC perturbing coils for modulating the magneticfield, and a DC magnetizing coils for magnetizing thespecimens. The specimen with a plate shape of 30mm length,5mm width, and 1mm thickness was magnetized by amaximum magnetic field of 12 kA/m with a sinusoidalwaveform of 0.05Hz. A perturbation field was applied as areference signal of 80A/m at 40Hz. The induced firstharmonic voltage in the pick-up coil was obtained by a lock-in amplifier using the voltage across a 1 shunt resistor.

I/O acquisition board

Power amplifier

Wave form generator

f Hz

Digital lock−in amplifier

ref. in

In

out

specimen

pick−up coil

Power amplifier

out

Shunt resistor1Ω

Computer (Lab−View)

yoke ac driving coil

dc driving coilV

inI

I

Fig. 1 Experimental set up for measuring the reversible permeability.

*Corresponding author, E-mail: chs2865@gmail.com

Materials Transactions, Vol. 50, No. 11 (2009) pp. 2691 to 2694#2009 The Japan Institute of Metals RAPID PUBLICATION

3. Results and Discussion

Figure 2 depicts the typical bright field TEM images, SADpattern, and EDS analyses of the as-tempered and 8000 haged specimens. The SAD patterns depict the primaryprecipitates of each specimen, that is, Cr23C6 (h112i zone),(NbC) (h112i zone), and Fe2W (h001i zone). The EDSprofiles indicate that the chemical composition of eachprecipitate consisting several dominant elements. The as-tempered specimen exhibited a tempered martensitic struc-ture with a high dislocation density in the lath interior andfine secondary phases, which were clearly identified as(Nb,V)(C,N) and Cr23C6 by TEM EDS and SADP analysesas shown in Fig. 2. The dislocation density in the martensitelath interior decreased as indicated by arrow in Fig. 2(b) andthe lath width increased during the isothermal ageing processas shown in TEM bright field micrographs. The Laves(Fe2W) phases were observed to precipitate primarily aroundlath boundaries, and their volume fraction increased. TheLaves phase grow faster than the Cr23C6 carbides followingthe diffusion controlled coarsening phenomena, that is,so-called Ostwald ripening.9) The coarsening rates of theCr23C6 carbide and Fe2W phases were calculated 1:251028 m3s1 and 6:84 1028 m3s1, respectively.

Figure 3 shows SEM micrographs showing precipitate onthe martensite lath boundary in as-tempered and 6000 h aged

specimen. The precipitate coarsening is already shown inTEM observation and the composition and crystal structureof precipitate analyzed precisely. The fine precipitates on themartensite lath boundaries and prior austenite grain (PAG)boundaries were observed to coarsen rapidly. The width ofmartensite lath also increased during long-term isothermalageing.

The backscattered electron (BSE) image was used todistinguish between the Cr23C6 and Fe2W phases, due tothe atomic number contrast provided by the distinct yields ofthe BSE signals from the various phases.10) Figure 4(a)shows the BSE image of 2000 h aged specimen in the vicinityof prior austenite grain (PAG) boundary. As indicatedby arrows, the precipitates could be readily distinguishedbetween the gray Cr23C6 and white Fe2W particles.Figure 4(b) depicts the XRD profiles of the extracted residuesat each aged specimen showing the phase identification ofthe precipitates. From the XRD profiles, it is found that theparticles were only identified as M23C6 and MX for theas-tempered specimen. In addition, the Fe2W phase (indexedby solid triangle) in XRD profile shows weak intensityat 1000 h ageing time, and then the peak intensity quiteincreased as the ageing time increase deducing the increaseof volume fraction of Fe2W.

NbC

Cr23C6

NbC

Cr23C6

1 µµm

Fe2W

Fe2W

1 µm

Fig. 2 TEM micrographs showing dislocation substructures and precip-

itates; (a) TEM images of the fine (Nb,V)(C,N) and Cr23C6 precipitates of

as-tempered specimen. The bright field image shows well developed lath

martensite structure (lath width is 0.2mm). EDS and diffraction patterns

show the chemical composition and crystal structure of precipitates, and

(b) TEM images showing the coarsened Fe2W precipitate of 8000 h aged

specimen. The bright field image shows recovery of lath structure.

(a)

(b)

5 µm

5 µm

Fig. 3 SEM micrographs (secondary electron image) showing precipitate

coarsening on the martensite lath boundary and increase in width of

martensite lath; (a) as-tempered and (b) 6000 h.

Table 1 Chemical composition of advanced ferritic 12Cr steel (mass%).

C Si Mn P S Ni Cr Mo W Co V N Nb Fe

0.19 0.06 0.13 0.013 0.01 0.52 11.01 0.09 3.45 0.21 0.06 0.03 0.01 bal.

2692 C. S. Kim, C. J. Lissenden, I. K. Park and K. S. Ryu

Figure 5 depicts the typical reversible permeability pro-files and position of maximum permeability as a function ofageing time in isothermally aged ferritic 12Cr steel. Theposition of maximum permeability in the relationship ofpermeability (r) versus applied field (Ha), corresponds tothe dynamic coercivity of the specimen considering that themagnitude of modulation field (Ho) is negligible comparedwith Hc (i.e., Ho Hc).

8,11) Therefore, dynamic coercivity(i.e., the position of maximum reversible permeability) couldbe successfully obtained from the reversible permeabilityprofile, and monotonously decreased as a function of ageingtime as shown in Fig. 5(b). Tensile strength also linearlydecreased with ageing time, which means that the materialswere mechanically and magnetically softened as well. Thetensile strength shows well linear relationship with dynamiccoercivity, which gives a wonderful potential to evaluatethe materials degradation using reversible permeabilitymeasurement.

The dynamic coercivity was observed to decrease duringisothermal ageing, but in the initial ageing time does notnoticeable change. Physically, the magnetization process isclosely related to domain wall movement, and the coercivitystrongly depends on the degree of pinning of domainwall.12,13) Microstructural impurities, such as inclusion,dislocation and grain boundary may hinder the domain wallmovement in ferromagnetic materials, which are potentials ofpinning sites of domain wall. A dislocation, which is a typicallattice defect evolving local strain due to an atomic mismatchwith the matrix, imparts a force on the domain wall, and thus,could be an effective obstacle to domain wall movement. Inthis study, the test materials are tempered martensite phase,consisted of dislocation tangles and lath martensite (high

dislocation density, 1012 cm2).4) The dislocation density wasmeasured by the Hall-Williamson method.14) As the ageingtime increased, the dislocation density decreased steadilyduring the whole ageing time as shown in Fig. 6(a), but inthe initial ageing time dislocation density decreased little.15)

The dislocation density of as-tempered specimen is 2:11015 m2, and continuously decreases to 1:58 1015 m2 for8000 h ageing. As dislocation density decrease, the latticestain might be also decrease, thus causing the coercivity todecrease during the whole ageing time.

Nonmagnetic particles are effective obstacles to domainwall movement, as they vary the area of the wall boundary.The decrease in wall energy due to the pinning phenomenonis related to the particle size and concentration. Themechanisms by which wall movement is impeded byinclusions are different by the relative size of inclusionscompared to the wall thickness.16) Inclusions, smaller than1 mm, easily adhere to the wall, and thus, decrease in energyof the wall by the pinning phenomenon. Inclusions are mosteffective mechanism in pinning the domain wall when thesize is equivalent to the domain wall thickness (e.g., 40 nmfor iron). In this study, the mean size of precipitates is severalhundred nano-meter and smaller than a few micro-meter.Therefore, precipitate could be an effective pinning site. Theprecipitate growth following the Ostwald ripening mecha-nism resulted in a decrease in concentration of precipitate asshown in Fig. 6(b), and hence, decreased the number ofpinning sites. As the ageing time increase, the precipitate arecoarsening as shown in previous explains (see Fig. 2 and

Cr23C6

Fe2W

Cr23C6

Fe2W

10 µm

35 40 45 50 55 60 65 70 75

as-tempered

8000 h

4000 h

2000 h

1000 h

2θ

Inte

nsi

ty (

arb

itra

ry u

nit

)

Fe2WNbCCr

23C

6

(b)

(a)PAGB

Fig. 4 Backscattered secondary electron image and X-ray diffraction

profiles; (a) discrimination between Cr23C6 and Fe2W precipitate, and (b)

XRD profiles of extracted residues showing the phases of precipitate at

each aged specimens.

-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12

500 h

2000 h

4000 h

as-tempered

Rev

ersi

ble

per

mea

bili

ty, µ

rev /

a.u

.

Magnetic field, HC / kA m-1

8000 h

0 1500 3000 4500 6000 7500 90001.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7CoercivityTensile strength

Ageing time, t / h

Dyn

amic

co

erci

vity

, HC /

kA m

-1

700

750

800

850

900

950

1000

1.2 1.3 1.4 1.5 1.6750

800

850

900

950

1000

Ten

sile

str

eng

th,

σ / M

Pa

Coercivity, HC

/ kA*m-1

Ten

sile streng

th, σ

/ MP

a

(a)

(b)

Fig. 5 Reversible permeability profiles, dynamic coercivity, and tensile

strength; (a) Reversible permeability profiles showing the variation in the

peak poison at each aged specimen, and (b) dynamic coercivity and tensile

strength as a function of ageing time, and the linear relationship between

tensile strength and dynamic coercivity.

Dynamic Coercivity of Advanced Ferritic Steel during Long-Term Isothermal Ageing 2693

Fig. 3), but the number density decreased due to the Ostwaldripening. This means that the pinning sites (i.e., precipitates)decrease. Inclusions generally reduce the area of domainwall boundary following the Kersten theory (i.e., inclusionreduces the wall energy).17) This theory yields the followingrelation for the coercivity.

Hc /f 2=3E

Msrð1Þ

where E is the domain-wall energy density, f is the volumefraction of inclusions, r is the radius of inclusions, and Ms isthe saturation magnetization.This equation shows that the coercivity is proportional tothe volume fraction of inclusions. Physically, coercivity isrelated to the pinning force of magnetic domain wall anddefined the maximum pinning force against domain wallmovement in ferromagnetic materials. Therefore, as thenumber of precipitates decreased, the dynamic coercivitydecreased.

The dynamic coercivity in the initial ageing time doesnot change as shown in Fig. 5(b). In this initial ageing,dislocation density decreased little, and moreover numberdensity of precipitate does not noticeable decrease. As it isabove mentioned, in the initial ageing time the pinning sites

of domain wall changed little, and thus causing the dynamiccoercivity did not any change.

4. Conclusions

In the present study, the microstructural evolution ofadvanced ferritic 12Cr steel for long-term isothermal ageingwas characterized and interpreted in relation to the reversiblepermeability. The dynamic coercivity was obtained from thereversible permeability profiles, which is measured by theyoke type probe. Long-term ageing caused coarsening of theprecipitate (i.e. Cr23C6, Fe2W), and also recovering disloca-tion resulting in mechanical softening. The decrease indynamic coercivity with ageing time was clearly related tothe decrease in the number of pinning sites, dislocationdensity and precipitate concentration.

Acknowledgement

This work was supported by the Korea Science andEngineering Foundation (KOSEF) grant funded by the Koreagovernment (MOST) (No. 2007-00467).

REFERENCES

1) J. M. Vitek and R. L. Klueh: Metall. Trans. A 14 (1983) 1047–1055.

2) B. A. Senior: Mater. Sci. Eng. A 119 (1989) L5–L8.

3) J. Hald: Steel Res. 67 (1996) 369–374.

4) C. S. Kim, S. I. Kwun, B. Y. Ahn, S. H. Nahm and S. S. Lee: Solid State

Phenom. 118 (2006) 475–478.

5) C. S. Kim, S. I. Kwun and I. K. Park: J. Nucl. Mater. 377 (2008) 496–

500.

6) Z. Kubon and V. Foldyna: Steel Res. 66 (1995) 389–393.

7) V. Foldyna, Z. Kubon, M. Filip, K. H. Mayer and C. Begner: Steel

Res. 67 (1996) 375–381.

8) K. S. Ryu, J. S. Park, S. H. Nahm, K. M. Yu, Y. B. Kim and D. Son:

J. Magn. Magn. Mater. 231 (2001) 294–298.

9) T. Wang, G. Sheng, Z. K. Liu and L. Q. Chen: Acta Mater. 56 (2008)

5544–5551.

10) C. S. Kim, I. H. Kim, I. K. Park and C. Y. Hyun: Key Eng. Mater. 326–

328 (2006) 1201–1204.

11) K. S. Ryu, S. H. Nahm, J. S. Park, K. M. Yu, Y. B. Kim and D. Son:

J. Magn. Magn. Mater. 251 (2002) 196–201.

12) C. H. Worner and J. E. Valdes: J. Appl. Phys. 63 (1999) 4324–4326.

13) C. S. Kim and S. I. Kwun: Mater. Trans. 48 (2007) 3028–3030.

14) G. Williamson and W. Hall: Acta Metall. 1 (1953) 22–31.

15) C. S. Kim: Ph. D thesis, (Korea University, Seoul, Korea 2007) (in

English).

16) B. D. Cullity: Introduction to Magnetic Materials, 2nd ed., (Addison-

Wesley, New York, USA, 1972) pp. 317–333.

17) C. W. Chen: Magnetism and Metallurgy of Soft Magnetic Materials,

(North-Holland, Amsterdam, Netherlands, 1977) pp. 132–135.

0 1500 3000 4500 6000 7500 90001.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

Dis

loca

tio

n d

ensi

ty, ρ

/ 10

15m

-2

Ageing time, t / h

(a)

(b)

0 1500 3000 4500 6000 7500 90002

3

4

5

6

7

8

9

Nu

mb

er d

ensi

ty, N

/ µm

-2

Ageing time, t / h

Fig. 6 Variation in dislocation density and number density of precipitates

as a function of ageing time.

2694 C. S. Kim, C. J. Lissenden, I. K. Park and K. S. Ryu