Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 86 (2014) 662 668

1st International Conference on Structural Integrity, ICONS-2014

Creep Crack Growth Behavior of a P91 Steel Weld

S. Venugopal*, G. Sasikala and Yatindra Kumar
Metallurgy & Materials Group, Indira Gandhi Centre for Atomic Research, Kalpakkam-603102, India
*
E-mail ID: svenu@igcar.gov.in

Abstract

Modified 9Cr-1Mo steel (P91) weld joints operating at elevated temperatures are well known to be prone to premature failure due
to cracking in the heat affected zone because of the gradients in microstructure, popularly referred to as Type IV cracking. A
campaign was undertaken to study the creep crack growth behaviour of modified 9Cr-1Mo weld joints. Creep crack growth
(CCG) tests were carried out on compact tension (CT) specimens machined from P91 weld joints prepared using multipass
shielded manual metal arc welding procedure. Specimens with two notch locations have been employed, (i) within the weld
metal, between the centreline and the fusion line and (ii) in the heat affected zone. Constant load CCG tests were carried out at
different applied loads at 798 and 898 K. The C* - da/dt correlations (da/dt =A C*m) were established for both notch locations.
At 898 K, a higher A (0.064) and lower m (0.533) for the case with notch placed in the HAZ, compared to those for the case of
notch in the weld (0.0399 and 0.75 respectively) were observed indicating the higher creep crack growth in HAZ, confirming
type IV cracking. The difference increases at lower C* levels which correspond to long term behaviour. Heavy creep damage was
observed in the HAZ region even for the sample with notch in the weld, whereas the weld metal regions showed relatively less
damage. As the crack grew, a change in its course to follow the HAZ region was observed.
2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
2014 The Authors. Published by Elsevier Ltd.
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Indira Gandhi Centre for Atomic Research.
Peer-review under responsibility of the Indira Gandhi Centre for Atomic Research
Keywords: Creep crack growth, P91 steel, weld joint, heat affected zone

1. Introduction

Creep crack growth (CCG) is a principal failure mechanism of components operating at elevated
temperatures. The failure assessment codes and standards consider the CCG behaviour of materials, and more
importantly those for the welds, in order to assess the reliability of such components. A modified 9Cr-1Mo steel
(P91) has been chosen for the steam generator applications in the prototype fast breeder reactor (PFBR) designed by
Indira Gandhi Centre for Atomic Research (IGCAR) and now in an advanced stage of construction at Kalpakkam.
Type IV cracking in heat affected zone (HAZ) of weldments has long been recognised as a problem for advanced
high Cr ferritic steels, see e.g. [1] for a review. Therefore, it is important to study initiation and growth of creep

1877-7058 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Indira Gandhi Centre for Atomic Research
doi:10.1016/j.proeng.2014.11.067
 S. Venugopal et al. / Procedia Engineering 86 (2014) 662 668 663

voids and cracks in weldment. A study was initiated with the aim of characterizing the creep crack initiation and
growth properties of different regions of P91 weld joints in the temperature range 798 to 898 K and also to
characterize the effect of notch location on these properties. The preliminary results of this campaign are presented
in this paper.

For materials creeping in the power law creep regime, the steady state stress field and strain rate
distribution at any point (r,) can be expressed using the fracture mechanics parameter C* as follows [2].
1 n +1
C*
ij = 0 ~ij ( , n ) (1)
I n 0 0 r
n n +1
C*
ijc = 0 ~ij ( , n ) (2)
I n 00 r
where r is the radial distance from, and is the angle at the crack tip, n is the Norton exponent.

 ) data as [3]
C* can be determined experimentally from the load line displacement rate (
P
C* = H (3)
B (W a )
where P is the applied load, BN is the net thickness between side grooves, W is the width, a is the crack length, H
and are specimen geometry dependent constants. For CT specimens, H = n/n+1 and = 2.2 according to ASTM
E1457 [3]. The CCG rate ( a ) can be correlated to C* as

da dt = A C *m (4)

The values of A and m characterise the CCG behaviour of the material.

2. Experimental

The chemical composition of the P91 steel chosen for this study is given in Table 1. This was received in
the form of plates of thickness 25 mm, in normalized (at 1323K for 30 min) and tempered (at 1053K for 75 min)
condition. Welded joints were fabricated from the plates using multi-pass shielded metal arc welding (SMAW)
process. Welding consumable used was of matching composition (Table 1) developed indigenously by IGCAR in
collaboration with Indian industry.

Table 1: Chemical composition (in mass %) of the P91 steel and the weld

Element C Mn Si P S Cr Ni Mo V Nb N Al Ta Fe

P91 0.097 0.37 0.31 0.018 0.005 9.29 0.38 0.92 0.26 0.08 0.057 Bal

Weld 0.1 0.39 0.48 0.021 0.009 9.5 0.14 1.00 0.25 0.10 0.065 0.024 Bal.

Table 2: The details of preparation of weld pads

Consumable : P91 Electrode, Basic coated, Number of passes : 50

3.2 mm
Joint Design : Single V-groove Current : 90-110 A
Process : Shielded Metal Arc Voltage : 20-24 V
Preheat temp : 523 K Travel Speed : 2.75-3.0 mm/s
Inter-pass temp. : 523 -573 K Heat input rate : 0.75-1.04 kJ/mm
664 S. Venugopal et al. / Procedia Engineering 86 (2014) 662 668

The weld pads were subjected to radiography to ensure the absence of any defects. Post-weld heat treatment of
1033K for 180 min was imparted to the weld pads. Compact tension (CT) type specimens of 20 mm thickness were
fabricated from these pads. Two notch tip locations were used as schematically shown in Fig. 1; (A) within the weld
metal about 2 mm from the fusion boundary (designated as type A specimen) and (ii) in the HAZ (designated as
type B specimen). Fatigue pre-crack of ~ 5 mm was introduced at room temperature. The constant load CCG tests
were completed for type A specimens in the load range 5 kN to 10 kN at 898 K, and for type B specimens, in the
load range 20 to 30 kN at 798 K. Limited testing of type B specimens was carried out for comparison of CCG
behaviour at the same stress intensity factor levels as type A at 898 K. Creep crack length was measured by D.C.
electrical potential drop method. The load-line displacement was measured using single arm rod-and-tube type creep
extensometer at the integral knife edges on the specimen.

a b
Base metal
Base metal

Weld metal Weld metal

Base metal Base metal

Fig. 1: Schematic of the compact tension specimens indicating the notch location with reference to the
weld deposit.

3. Results and Discussion

Typical plots of the variation of load line displacement and a with time are presented in Fig. 2 and 3
respectively for Type A specimens and Fig. 4 and 5 for Type B specimens. The data were analysed following the
guidelines of ASTM E1457 [2]. Crack initiation time, t0.2, i.e., the time to attain 0.2 mm creep crack growth as a
function of initial value of CCG parameter (C*) is presented in Fig. 6. It is observed that initiation time for both
weld and HAZ follow the same scatter band. It should be noted that the tests on specimens with weld notch were
carried out at 898 K, whereas those for HAZ notch were carried out at 798 K. However, considering the scatter
typical of the initial CCG data, comparison of these should be done with caution. The crack growth rate (da/dt) as a
function of C* for P91 weld joint specimens with the notch in weld and HAZ regions are presented in Fig. 7. In spite
of the lower testing temperature for the specimen with HAZ notch, data for both notch locations fall in the same
scatter band, indicating lower resistance to creep crack growth in HAZ.
 S. Venugopal et al. / Procedia Engineering 86 (2014) 662 668 665

5 6
P91Weld, 898 K 8 kN
4.5 P91Weld, 898 K, 5 kN
Notch in the w eld region
Notch in the w eld region 5 10 kN
4

3.5

D is pla c e m e nt, m m
Displacem ent, m m

4
3
2.5 3

2
2
1.5

1
1
0.5

0 0
0 1000 2000 3000 4000 5000 6000 7000 0 50 100 150 200 250 300 35
Time, h Time, h

Fig. 2: Variation of creep displacement as a function of time for Type A specimens.

35 32
P91Weld, 898 K 8 kN
P91Weld, 898 K, 5 kN 31 Notch in the w eld region
33 10 kN
Notch in the w eld region
30
C ra c k le ngth, m m

31 Crack len gth, m m 29

28
29
27

27 26

25
25
24

23 23
0 1000 2000 3000 4000 5000 6000 7000 0 50 100 150 200 250 300 350
Time, h Tim e, h

Fig. 3: Variation of crack length as a function of time for Type A specimens.

3.5 7.0
Crack extension, m m

30 kN P91 weld-798 K
Displacement, mm

P91 weld-798 K
3.0 25 kN
6.0 Notch in HAZ
Notch in HAZ
2.5 20 kN 5.0
2.0 4.0
1.5 3.0
1.0 2.0
30 kN
0.5 1.0 25 kN
20 kN
0.0 0.0
0 500 1000 1500 2000 0 500 1000 1500 2000
Time, h Time, h
Fig. 4: Variation of creep displacement as a Fig. 5: Variation of crack length as a
function of time for Type B specimens. function of time for Type B specimens.
666 S. Venugopal et al. / Procedia Engineering 86 (2014) 662 668

10000
Notch in HAZ
Notch in weld
Initiation time, h
1000

100

10

1
0.01 0.1 1 10
C*, kJ/m^2h

Fig. 6: Crack initiation time (t0.2) as a function of initial C* for the notch in weld and HAZ
regions.
1.E+01
Open symbols: Notch in Weld, 898 K
1.E+00 Solid symbols: Notch in HAZ, 798 K

1.E-01
da/dt, mm/h

1.E-02

1.E-03

1.E-04

1.E-05
0.001 0.01 0.1 1 10 100

C*, kJ/m^2h

Fig. 7: The crack growth rate (da/dt) as a function of CCG parameter C* for P91 weld joint specimens
with the notch in weld and HAZ regions.

Comparison of the CCG Characteristics for the two Notch Location

For a clearer comparison, tests were conducted with the same initial value of stress intensity factor K for
both types of specimens at the same temperature, namely, 898 K. The loads applied were chosen based on the
initial crack lengths determined on the surface to yield a similar level of K. However, the initial crack lengths
determined optically after the test were slightly different from those at the surface. Therefore, the type A
specimens had a slightly higher initial K than type A specimen (19.8 and 17.9 MPa.m1/2 respectively). The
variations of displacement and crack growth with creep time are presented in Fig. 8.
 S. Venugopal et al. / Procedia Engineering 86 (2014) 662 668 667

3.5 33
HAZ HAZ
Displacement, mm 3
31 weld

Crack length, mm
weld
2.5

2 29

1.5 27
1
25
0.5

0 23
0 100 200 300 400 0 100 200 300 400

Time, h Time, h

Fig. 8: Variations of displacement and crack length as a function of time.

Even at a slightly higher K level, the crack growth rate in type A (weld) specimen was found to be
consistently lower than that in type B (HAZ) specimen. The displacement rates were similar for both types in the
steady state, but increased for the type B specimen towards the tertiary stage which set in earlier. The C* vs da/dt
plots for the two types of specimen are shown in Fig. 9. It may be noted that the crack growth rate at the HAZ was
higher than that in the weld metal at similar C* levels, especially at low C* levels indicative of long term effects.

Creep crack growth in the HAZ region is characterized by extensive creep void formation exclusively in
the fine-grained heat-affected zone. This HAZ region is highly susceptible to creep damage accumulation due to the
fine (average grain size of < 5 m) equiaxed grain structure with no martensitic lath structure [4]. Extensive creep
damage was observed in the HAZ region of type A specimens (Fig. 10) though this region is away from the highest
stress concentration regions. In type B specimens, damage in this region is expected to be more extensive due to
closer proximity to the notch tip and lead to a higher crack growth rate at longer durations corresponding to lower
end of C* (Fig. 9).

1
Notch in HAZ
Notch in Weld
1

0.1
da/dt, mm.h

0.53
da/dt = 0.064C*

0.75
0.01 da/dt = 0.0399C*

P91 weld joint

898 K
1E-3
0.1 1 10
2 1
C*, kJ.m .h

Fig. 9: The C* vs da/dt plots for the two Fig. 10: Extensive creep damage observed in the
types of specimens at similar initial K heat affected zone of the weld joint.
levels.

The microstructural reasons for this may be explained as in the following. As mentioned earlier, the creep
properties of weld deposits and the heat affected zones (HAZ), specifically for the class of 9Cr-1Mo steels, differ
significantly from those of the base material since the welding process introduces very complex metallurgical and
mechanical heterogeneities in to this region due to the solid state phase transformations, residual stresses and heat
treatments [5]. The intercritical zone experiences peak temperature between Ac1 and Ac3 and only partial
austenitization takes place here. The microstructure consists of fresh transformation products and re-tempered
martensite. Also, coarse M23C6 carbides have been reported to be present in the intercritical zone in comparison to
the parent metal P91 [6]. Creep deformation and damage evolution are significantly influenced by the presence of
fine precipitates and second phase particles, in addition to the solid solution hardening. Various types of carbides
and carbonitride such as M23C6, M6C, M3C, M7C3, M5C2, M2X, MX etc and phases (Laves) have been reported to
668 S. Venugopal et al. / Procedia Engineering 86 (2014) 662 668

be present in the steel of this class [7]: and Laves phases. Grain boundary sliding and grain boundary migration
which accompany the creep deformation are also influenced by the precipitates. Damage nucleation and growth
rates are enhanced by the stress concentration near these precipitates [8]. Also, coarsening of M23C6 or Laves phase
(Fe,Cr)2Mo due to long term thermal exposure leads to a decrease in the creep resistance [9] of the HAZ. Formation
of Z phase, Cr(V,Nb)N, at the expense of VN-precipitates is another factor that leads to a long term degradation
effect.

Conclusions

The creep crack growth behavior of P91 weld joint has been characterized using the fracture mechanics
parameter C* for two different notch locations, and da/dt-C* correlations were established. Crack initiation time was
found to be a function of initial C* for both location and for HAZ notch at 798 K was similar to that for weld notch
at 898K. The CCG rates were found to be higher in the HAZ region, especially for low C* levels as indicated by a
higher A value in the da/dt-C* correlation.

References

1. Francis J. A., Mazur W., and Bhadeshia H. K. D. H., Type IV cracking in ferritic power plant steels,
Materials Science and Technology, 2006 VOL 22 NO 12, 1387.
2. Webster GA, Ainsworh RA. High temperature component life assessment, London, Chapman & Hall;
1994.
3. ASTM E1457-07, Standard Test Method for Measurement of Creep Crack Growth Times in Metals. ASTM
International, 100 Barr Harbor Dr., P.O. box C700 West Conshohocken, Pennsylvania 19428-2959, United
States, 2007.
4. Peter Mayr, Stefan Mitsche, Horst Cerjak, Samuel M. Allen, The Impact of Weld Metal Creep Strength on
the Overall Creep Strength of 9% Cr Steel Weldments, Trans ASME Journal of Engineering Materials and
Technology, Vol. 133, 2011,021011-1 to 7
5. Bhadeshia, H. K. D. H., ISIJ International, 41, 2001, p. 626.
6. Laha K.,Chandravathi K. S., Bhanu Sankara Rao K., Mannan S. L., Sastry D. H, Trans. Indian Inst. Met.,
53, 2000, p. 217.
7. Hald, J., In: Advanced Creep Data for Plant Design & Life Extension (Proc. of Int Seminar),. Praha,
SVM 2003, p. 58.
8. Ashby M. F.,Gandhi C., Taplin D. M. R., Acta Met., 27, 1979, p. 699.
9. Panait C, Bendick W, Fuchsmann A, Gourgues-Lorenzon A-F, Besson J. Study of the microstructure of the
grade P91 steel after more than 100,000h of creep exposure at 600C, Inter. J. Press. Vess. Pip. 87, 326-
335.