FATIGUE CRACK GROWTH OF Fe-0.85Mo-2Ni-0.6C STEELS WITH ...

FATIGUE CRACK GROWTH OF Fe-0.85Mo-2Ni-0.6C STEELS WITH A HETEROGENEOUS MICROSTRUCTURE

G. Piotrowski, X. Deng, and N. Chawla

Department of Chemical and Materials Engineering Arizona State University

Ira A. Fulton School of Engineering Tempe, AZ 85287-6006

K.S. Narasimhan and M. Marucci

Hoeganaes Corporation Cinnaminson, NJ 08077

ABSTRACT

Powder metallurgy processing of steel alloys typically results in a material with heterogeneous microstructure and residual porosity. The fatigue crack growth behavior of these materials is strongly affected by the nature of porosity and heterogeneous microstructure. Notched fatigue specimens were prepared from a Fe-0.85Mo prealloy mixed and binder-treated with 2%Ni and 0.6%C. The alloys were tested at three different densities: 6.98 g/cm3, 7.36 g/cm3, and 7.53 g/cm3. The microstructure at each density was characterized to determine the porosity, microconstituents, and phase fractions. Fatigue testing was performed at various R-ratios, ranging from –2 to 0.8. Increasing porosity and increasing R-ratio resulted in a decrease in ∆Kth. In situ observation of crack growth showed that the cracks propagated through Ni-rich regions. It appears that pearlite regions, and, to some extent bainite regions, however, contributed to toughening and crack deflection.

INTRODUCTION

Steel products made by press and sinter powder metallurgy (P/M) processing are replacing products made by more conventional procedures, such as casting and forging, due to the cost savings associated with near-net shape processing [1-4]. Unfortunately, P/M parts usually have some degree of residual porosity after sintering, which can adversely affect the mechanical properties of these materials [4-8]. The nature of porosity is determined by several processing variables, including the type and amount of alloying additions, powder size distribution, green density, sintering temperature, and sintering time [1,5].

In general, an increase in porosity is associated with more irregular pores, and a greater fraction of interconnected pores. Increased pore interconnectivity and pore clustering leads to increased strain localization and damage, thus reducing the strength and ductility of the steel [5,9]. Porosity has also been shown to significantly influence the fatigue response of sintered P/M steels [4-8,10-13]. Danniger et al. [5] found that pores and pore clusters act as sites of crack initiation, while Polasik et al. [8] have shown

that small cracks that nucleate at pores coalesce to form large fatigue cracks leading to fracture. A comprehensive understanding of the effects of porosity and microstructure on the fatigue crack growth behavior of sintered steels is required. In this study, we have examined the effect of porosity and microstructure on fatigue crack growth behavior of a Fe-0.85Mo-2Ni-0.6C at three different sintered densities: 6.98 g/cm3, 7.36 g/cm3, and 7.53 g/cm3. Quantitative characterization of the heterogeneous P/M microstructure was performed to determine the degree of porosity and nature of microstructure at all densities. With increasing density the degree of pore interconnectivity decreased, although density did not significantly affect the fraction of the microconstituents in the microstructure. Fatigue crack growth experiments were performed at constant R-ratio, ranging from -2 to 0.8. It will be shown that increasing porosity and R-ratio decrease the fatigue crack growth resistance of the material, and that the heterogeneous nature of the microstructure also influences the fatigue crack growth behavior.

MATERIALS AND EXPERIMENTAL PROCEDURE An Fe-0.85Mo prealloy powder was blended and binder treated with 2 wt.% Ni and 0.6wt.% graphite [14,15]. All powders were pressed into rectangular blanks and sintered at 1120ºC for 30 minutes in a 90% N2–10% H2 atmosphere. The samples were pressed and sintered to obtain three different sintered densities: 6.98 g/cm3, 7.36 g/cm3, and 7.53 g/cm3. The samples with 7.53 g/cm3 density were made by a double-press/double-sintering process. Porosity was determined by image analysis of several representative micrographs using both optical and scanning electron microscopy (SEM). The micrographs were segmented into black and white images, and the porosity determined by an automated procedure. Quantitative characterization of the phase fractions of the microstructure was performed by both image analysis of the optical and SEM images and Vickers hardness tests. The phase area fractions of each density were determined using a segmentation procedure with various colors, whereby each phase was assigned a specific color. The segmented images were then analyzed based on color to determine the phase fractions. Hardness measurements were conducted using a Vickers hardness indenter with a 50 g applied load. A minimum of 20 measurements were conducted in each phase. Fatigue tests were performed on a servo-hydraulic load frame equipped with a Questar (Questar Corp., New Hope, PA) traveling microscope. The traveling microscope was used for in situ measurement and observation of fatigue crack growth. This was particularly important because it allowed visualization of the interactions between the crack and the microconstituents in the microstructure. All fatigue tests were performed using a single edge notch axial fatigue configuration [16-18]. The samples were machined by electrodischarge machining (EDM) to the following dimensions: height of 45 mm, width of 11.5 mm, thickness of 7.48 mm. The edge notch was machined by EDM to a length of 4.5 mm. Fatigue tests were performed at constant R-ratio, ranging from 0.8 to -2, in load control at a frequency of 30 Hz. A decreasing ∆K procedure was used until ∆Kth was reached. At this point, the ∆K was raised to a value in the Paris law regime, and the ∆K increased to obtain the higher ∆K behavior. Samples were pre-cracked to a length of about 600 µm. The typical growth increment for a given ∆K was 200 µm, such that the increment encompassed several phases in the heterogeneous microstructure and was sufficiently larger than the plastic zone from the previous ∆K increment. RESULTS AND DISCUSSION

Microstructure Characterization The porosity at each density was determined from several optical micrographs using image analysis, Table 1. At the lowest density, 6.98 g/cm3, the measured porosity was 9.5%, while at the highest density 7.53 g/cm3, the porosity was about 3.9%. The values obtained from the image analysis technique

Table 1. Porosity Measurements versus Density

correlated well with those obtained from calculation of the sintered density from the pore-free density The microstructure at the three densities showed noticeable differences in porosity, Figure 1. At the highest porosity, larger, more irregular, and interconnected pores were observed. With decreasing porosity, the overall pore size was smaller, and the pore shape was more regular.

Etched microstructures showed the heterogeneous nature of the microconstituents, in addition to the porosity, Fig. 2. These included coarse pearlite, fine pearlite, Ni-rich regions (likely Ni-rich austenite, surrounding the pores), and bainite, which appears at the periphery at the Ni-rich regions. Ni-rich regions likely formed as a result of incomplete diffusion of the elemental Ni into the surrounding matrix upon sintering.

Quantitative characterization of the phases and phase fractions in the microstructures, at all three densities, was also conducted, Fig. 3. The phase fractions were determined from both optical and SEM micrographs. Each phase was segmented in the micrograph and indexed by a specific color. A similar analysis has been conducted by Komenda et al. [19] to characterize a heterogeneous steel microstructure. The analysis showed that the phase fractions in the microstructure were relatively uniform at all densities. As expected, the microstructure consisted primarily of coarse pearlite, with smaller amounts of fine pearlite and Ni-rich phase, and minor amounts of bainite. The bainite microconstituent was observed primarily at the periphery of the Ni-rich regions. It is interesting to note that the results from optical micrographs overestimated the amount of

coarse pearlite and Ni-rich regions, at the expense of the amount of fine pearlite. This is understandable, since the fine pearlite is more difficult to resolve in the optical microscope. Thus, the true phase fraction measurements were obtained from the higher magnification SEM images.

Sintered Density (g/cm3)

Porosity from Sintered Density (%)

Porosity from Image Analysis (%)

6.98 10.3 9.5 ± 0.8 7.34 4.5 4.5 ± 0.6 7.53 3.2 3.9 ± 0.2

50 µm50 µm50 µm

(a) (b) (c)

Figure 1. Microstructure of Fe-0.85Mo-2Ni-0.6C steels at: (a) 6.98 g/cm3, (b) 7.34 g/cm3, and (c) 7.53 g/cm3. At higher porosity the pores are larger, more irregular, and more interconnected.

Figure 2. Characteristic heterogeneous Microstructure of p/m steel consisting of porosity, coarse and fine pearlite, bainite, and Ni-rich phase (likely austenite).

25 µm

Bainite

Coarse Pearlite

Fine Pearlite

Pore

Ni-rich

Vickers hardness of the several microconstituents showed that the bainite regions were the hardest, followed by fine pearlite, coarse pearlite, and Ni-rich areas, Fig. 4. The relatively low hardness of the Ni-rich phase suggests that it may be retained austenite [20]. The hardness of each phase, as expected, did not vary significantly with density. Fatigue Behavior Fatigue crack growth experiments at the three densities showed that porosity had a strong effect on the crack growth rate (da/dN) and the threshold stress intensity factor, ∆Kth. Figure 5 shows the da/dN versus ∆K behavior for all densities, at various R-ratios. ∆Kth at the three densities varied between 14.7-16.4 MPa.m1/2 at R=-2, to 2.9-4.3 MPa.m1/2 at R = 0.8. These values are in the range of those found by others in similar Fe-Mo sintered steels [4,21-23]. The slope in the steady state or Paris law regime, m, was also measured [24]. Increasing R-ratio resulted in an increase in m from about 2 at low R-ratio to around 10 at R-ratio of 0.8. This increase in slope is indicative of a much higher crack growth rate, for a given ∆K, due to increasing Kmin. We now compare the fatigue crack growth behavior, at all three porosity levels for three R-ratios: -1, 0.1, and 0.8, Fig. 6. At all R-ratios, decreasing porosity resulted in a higher ∆Kth. This is particularly clear at the highest R-ratio of 0.8. The fatigue crack growth data, at all R-ratios, was analyzed using the two- parameter approach proposed by Vasudevan and Sadananda [26,27]. In the two-parameter approach there exist two critical parameters that dictate when fatigue crack propagation will take place, at a given crack growth rate, Fig. 7. The two

critical components are the cyclic component ∆K and the static component Kmax. The critical values for crack propagation then, shown in Fig. 7, are the asymptotic values in the curve, ∆K* and K*max. By measuring the values ∆K* and K*max over a range of da/dn values, “fatigue trajectory” plots can be generated. These provide a means to elucidate changes in material behavior with increasing crack growth rate. The two-parameter approach has also been used to categorize material behavior into different classes, Fig. 8 [26]. Materials which are dominated by environmental effects exhibit the behavior described by Class IV. Materials which are dominated by the static component, Kmax

*, will show a significant monotonic component to fatigue damage, similar to that shown by the Class II material. The so-called “ideal material,” described as Class III behavior in Fig. 8, exhibits a perfect “L-shaped” behavior.

100

200

300

400

500

600

700

6.9 7 7.1 7.2 7.3 7.4 7.5 7.6

Vic

kers

Har

dnes

s N

umbe

r (V

HN

)

Density (g/cm3)

Bainite

Fine Pearlite

Coarse Pearlite

Ni-rich

Figure 4. Vickers hardness measurements of individual phases. As expected, density does not affect the hardness of the individual phases.

0

20

40

60

80

100

6.9 7 7.1 7.2 7.3 7.4 7.5 7.6 7.7

Coarse Pearlite

Density (g/cm3)

Pha

se A

rea

Frac

tion

(%)

Total Pearlite

Fine Pearlite Ni-rich

Bainite

SEMOptical

Figure 3. Phase and microconstituent area fractions by optical and scanning electron microscopy (SEM). Optical microscopy underestimates the fraction of fine pearlite.

Figure 6. Effect R-ratio on fatigue crack growth, as a function of density. Increasing porosity and R-ratio resulted in a decrease in fatigue crack growth resistance.

10 -11

10 -10

10 -9

10 -8

10 -7

10 -6

2 4 6 8 10 30

6.98 g/cm 3

7.36 g/cm3

7.53 g/cm3

R = 0.8 R = 0.1

R = -1

da/d

N (m

/cyc

le)

∆K (MPa-m1/2)

10 -11

10 -10

10 -9

10 -8

10 -7

10 -6

2 4 6 8 10 30

m

R = 0.8 m = 10

R = 0.1 m = 5 R = 0.3

m = 5 R = -1m = 4

R = -2 m =2

da/d

N (m

/cyc

le)

∆K (MPa-m1/2)

6.98 g/cm3

10-11

10-10

10-9

10-8

10-7

10-6

2 4 6 8 10 30

m

R = 0.8m = 10 R = 0.1

m = 4 R = 0.3 m = 6

R = -1 m = 3

da/d

N (m

/cyc

le)

∆K (MPa-m1/2)

7.34 g/cm3

10 -11

10 -10

10 -9

10 -8

10 -7

10 -6

2 4 6 8 10 30

m

R = 0.1 m = 5 R = 0.3

m = 5 R = 0.8 m = 10

R = -1 m = 5

R = -2m = 3

da/d

N (m

/cyc

le)

∆K (MPa-m1/2)

7.53 g/cm3

Figure 5 da/dn versus ∆K for: (a) 6.98 g/cm3 (b) 7.34 g/cm3, and (c) 7.53 g/cm3. Porosity has a strong influence on the fatigue behavior of the P/M steel. Increased porosity and R-ratio decrease the ∆KTH values of the steel. The slope of the steady-state region of curves was also found to increase with R-ratio, indicative of an increased monotonic contribution to fatigue.

(a) (b)

(c)

Figure 7. Schematic of two parameter approach to fatigue (after Vasudevan and Sadananda [27]). Two critical parameters are required for crack growth, a cyclic component, ∆K, and a static component, Kmax.

∆K

Kmax

Fatigue Crack Growth Region

K*max

∆K*

We begin by examining the ∆K versus R-ratio behavior. The curves are relatively linear, although it can been seen that with increasing density, a higher ∆K is required to achieve a given crack growth rate. Also, at least up to an R-ratio of 0.8, there does not seem to be a plateau in ∆K, as described by the Class III behavior. Rather, is likely that large degree of plasticity is present at the crack tip at very high R-ratio, that makes the P/M materials fall

into the Class II category. The changes in material behavior with increasing R-ratio and with increasing crack growth rate are more easily observed in a series of plots of ∆K versus Kmax, Fig. 9. A series of curves are shown at each of the crack growth rates, shown in Fig. 10. As the degree of porosity increases, the monotonic contribution to the fatigue crack growth behavior increases. This can be seen graphically as the deviation from the horizontal asymptote in the ∆K-Kmax plots. The behavior can be rationalized by the fact that with increasing R-ratio, the localization of strain and plasticity at the crack tip are enhanced in a material with a higher degree of porosity. Thus, a lower threshold is required for crack growth to take place with a larger degree of porosity at the crack tip. This behavior is exacerbated with increasing crack growth rate, as shown by the increased deviation of the “L-shaped” curves with increasing crack growth rate. This behavior is more predominant at the lower densities, because of the enhanced effect of porosity. At the highest density, the material behavior is closer to the classical “ideal material.” To quantify the degree of monotonic superposition, a comparison between the experimental data and a theoretical pure fatigue line was made. The condition for “pure fatigue” is that ∆K* equal K*max, Fig. 11(a). Thus, any devitation from this 45o line, must be due to environmental, crack closure, or static effects. As the line moves toward the horizontal axis, this signifies an increased contribution from monotonic loading, since Kmax is the dominating component to fatigue [25]. Our data show that at the two lower densities the behavior is close to parallel to the horizontal axis, indicating significant static effects associated with strain localization between pores, Fig. 11(b). At the highest density, however, the line is close to parallel to the pure fatigue line. This indicates that the static contribution with increasing R-ratio is not as predominant. Rather, the off-set between the two parallel lines may be attributed to two components: (a) fatigue crack closure at low R-ratio and (b) static effect at higher R-ratio.

0

10

20

30

40

50

-2.5 -2 -1.5 -1 -0.5 0 0.5 1

K [M

Pa-m

1/2 ]

∆

R-ratio

K*max

∆K*

∆KTH

1e-9

1e-8

1e-7

0

10

20

30

40

50

-2.5 -2 -1.5 -1 -0.5 0 0.5 1

K [M

Pa-m

1/2 ]

∆

R-ratio

K*max

∆K*

∆KTH

1e-9

1e-8

1e-7

0

10

20

30

40

50

-2.5 -2 -1.5 -1 -0.5 0 0.5 1

K [M

Pa-

m1/

2 ]∆

R-ratio

K*max

∆K*

1e-7

1e-8

1e-9

∆KTH

6.98 g/cm3 7.34 g/cm3 7.53 g/cm3

Figure 9. ∆K versus R for: (a) 6.98 g/cm3, (b) 7.34 g/cm3, and (c) 7.53 g/cm3.

(a) (b) (c)

∆K∆K

KmaxR-ratio

V I

II III

IV

V

I

II

IIIIV

Environment effect

Strong Weak

Figure 8. Classification of different types of materialsby the two parameter approach. Class I materials exhibit limited sensitivity to environmental effects. Class III material is termed the “ideal material” behavior, while Class II materials are influence by large plasticity at the crack tip.

Figure 10. Trajectory maps of ∆K versus Kmax at various da/dN: (a) 6.98 g/cm3, (b) 7.34 g/cm3, and (c) 7.53 g/cm3. Note that for highest porosity, the horizontal asymptote is diminished due to the significant contribution of plasticity at the crack tip (from Kmax).

0

10

20

30

40

0 10 20 30 40

R = 0.8

R = 0.3

R = 0.1

R = -1 R = -2

10 -10

10 -9

10 -8

10 -7 m/cycle

Pure Fatigue R = 0

0

10

20

30

40

0 10 20 30 40

R = 0.8

R = 0.3

R = 0.1

R = -1

10-10

10-9

10-8

10-7 m/cycle

Pure Fatigue R = 0

0

10

20

30

40

0 10 20 30 40

R = 0.8

R = 0.3

R = 0.1

R = -1 R = -2

10 -10

10 -9

10 -8

10 -7 m/cycle

Pure Fatigue R = 0

Kmax (MPa-m1/2) Kmax (MPa-m1/2)

Kmax (MPa-m1/2)

∆K (M

Pa-

m1/

2 )

∆K (M

Pa-

m1/

2 )

∆K (M

Pa-

m1/

2 ) 6.98 g/cm3 7.34 g/cm3

7.53 g/cm3

(a) (b)

(c)

∆K*

Kmax*

Pure Fatigue Line

Monotonic Contribution

Increasing Crack Growth Rate

3

7

11

15

3 7 11 15

Pure Fatigue

7.55 g/cm3

7.36 g/cm3

6.98 g/cm3

∆K*

Kmax*

Experimental Results

Figure 11. Quantifying the Degree of Monotonic Superposition: (a) monotonic contribution quantified by measuring the deviation of the experimental data from the “pure fatigue” line, and (b) as porosity increases, the degree of monotonic superposition also increases. This is indicated by the increasing deviation from the pure fatigue line.

(a) (b)

In this section we describe fractographic analysis after fatigue crack growth. Since the crack path appeared quite tortuous, due to the hetrogeneous microstructure of these steels, we have also quantified the degree of crack deflection on the measured ∆K, per the model of Suresh [28]. The crack appears to be highly dependent on the phase(s) at the crack tip, Fig. 12. For the Ni-rich regions, cracks tend to propagate in a linear fashion, suggesting that the Ni-rich regions offer little to no resistance to crack propagation. This is further by the Vickers hardness data, which showed the Ni-rich phase to be very soft, indicating that it might be Ni-rich austenite phases [20]. For the pearlite regions, cracks tend to be highly deflected, with some evidence of the Fe3C particles in the ferrite matrix bridging the crack, Fig. 13. Pores inside the Ni-rich regions can also act as nucleation sites for secondary cracks. It has been demonstrated that these cracks originating at pores ahead of the crack tip, often join the main crack [8,13], Fig. 14. Cracks propagating through the coarse pearlite, fine pearlite, and bainite all show large increases in the degree of fatigue resistance due the crack deflection, Fig. 14. Crack arrest is often present, and further induces crack deflection through branching. The role of the Ni-rich areas in fatigue is still not quite clear. Quantitative measurements of the crack growth rate in the different microstructural constitutents are being conducted, and will shed some light on the role of Ni-rich areas.

(a)

20 µm 20 µm

(b)

Figure 12. Fatigue Crack Behavior through heterogeneous microstructure: (a) Ni rich, and (b) Pearlite. The crack propagates through the Ni-rich areas, but it tortuous and deflected through pearlite due to the Fe3C needles.

0

50

100

150

200

250

300

0 2 4 6 8

Energy (keV)

Cou

nts

Fe

Fe

Fe

C

Figure 13. Crack Bridging due to Fe3C: (a) Fe3C needles pulled out of the ferrite matrix during fatigue, and (b) EDS analysis showing the composition corresponding to Fe3C.

Fe3C

1 µm

Fe3C

(a) (b)

The crack deflections induced by the heterogeneous microstructure increased the fatigue resistance of the steel. The increase in the fatigue resistance of the steels can be quantified by applying the crack deflection model by Suresh [28]. The model assumes that for a given crack length there exist linear portions, S, and deflected portions, D, Fig. 15. The angle between the linear portion and the deflected portion is given by θ. By measuring S, D and θ, the true ∆K, corrected for crack deflection, can by calculated from the following equation:

[1] The values of ∆KI,app calculated from the use of the Suresh deflection model represent increases in fatigue resistance due to crack deflection. The corrected ∆K values are shown in Fig. 16. Note that the curves shift to the right, indicating some contribution from deflection, although this contribution is quite small. Also, the contribution from deflection does not appear to be noticeably different between the three densities. This may be due to the fact that all three densities exhibit relatively equivalent heterogeneous microstructures.

Figure 15. Suresh’s model [28] for quantifying crack deflections. The crack consists of linear regions, S, and deflected regions, D. The angle between S and D is θ.

50µm

(b)

50µm (c)

50µm

(d)

50µm

Figure 14. In situ observation of cracking during fatigue: (a) Initial, (b) 10,000 cycles, (c) 16,000 cycles, and (d) 22,000 cycles.

(a)

12

effapp,I SD

S2

cosDkK

−

+

+

θ

∆≈∆

CONCLUSIONS In this study the effect of porosity and heterogeneous microstructure on the fatigue crack growth behavior of sintered steels was systematically studied. The following conclusions can be made based on our results:

• The heterogeneous microstructure plays an important role in fatigue crack behavior, in particular, porosity significantly influences fatigue crack growth.

• Increasing porosity and R-ratio resulted in a decrease in ∆KTH values. Increased R-ratio and

porosity also increased the monotonic contribution during fatigue due to strain localization between pores at the crack tip.

• The fatigue crack path is tortuous and highly dependent on microstructure. Areas of pearlite cause

crack arrest, deflection and branching, while Ni-rich and porous areas do not appear to offer resistance to crack growth.

• Crack deflections increase the fatigue resistance of the material for all three densities, although

the contribution from crack deflection is quite small. ACKNOWLDEGMENTS The authors gratefully acknowledge Hoeganaes Corp. for financial support of this research.

Figure 16. Increase in Fatigue Resistance Due to Deflection: (a) 6.98 g/cm3, (b) 7.34 g/cm3, and (c) 7.53 g/cm3. The extent of deflection is similar for all the three densities.

10 -10

10 -11

1 10 100

10 -9

10 -8

10 -7

10 -6 R = 0.80.3 0.1 -1

∆K (MPa-m1/2)

d

a/dN

(m/c

ycle

)

7.34 g/cm3

-1

∆K (MPa-m1/2)

-2

10 -10

10 -11

1 10 100

10 -9

10 -8

10 -7

10 -6

R = 0.8 0.3 0.1

d

a/dN

(m/c

ycle

) 6.98 g/cm3

R = 0.8

0.3 0.1

∆K (MPa-m1/2)

d

a/dN

(m/c

ycle

)

10 -10

10 -11

1 10 100

10 -9

10 -8

10 -7

10 -6

-2-1

7.53 g/cm3

(a) (b)

(c)

REFERENCES

1. German, R.M., Powder Metallurgy of Iron and Steel, Wiley Intersciences, New York, 1998. 2. Narasimhan, K.S., “Sintering of Powder Mixtures and the Growth of Ferrous Powder

Metallurgy”, Mat. Chem. & Phys., Vol. 67, 2001, pp. 56-65. 3. Narasimhan, K.S., “Recent Advances in Ferrous Powder Metallurgy”, Adv. Performace Mat.,

Vol. 3, 1996, pp. 7-27. 4. Hadrboletz, A., Weiss, B., “Fatigue Behaviour of Iron Based Sintered Material: A Review”, Intl.

Mat. Rev., Vol. 42, No. 1, 1997, pp. 1-44. 5. Danniger, H., Spoljaric, D., Weiss, B., “Microstructural Features Limiting the Performance of

P/M Steels”, Intl. J. Powder Metal., Vol. 33, No. 4, 1997, pp. 43-53. 6. Chawla, N., Deng, X., Marrucci, M., Narasimhan, K.S., “Effect of Density on the Microstructure

and Mechanical Behavior of Powder Metallurgy Fe-Mo-Ni Steels”, Advances in Powder Metallurgy and Particulate Materials, Vol. 6, 2003, Metal Powder Industries Federation, pp. 7-257 – 7-269.

7. Chawla, N., Polasik, S., Narasimhan, K.S., Murphy, T., Koopman, M., Chawla, K.K., “Fatigue Behavior of Binder-Treated P/M Steels”, Intl. J. Powder Metall., Vol. 37, No. 3, 2001, pp. 49-57

8. Polasik, S.J., Williams, J.J., Chawla, N., “Fatigue Crack Initiation and Propagation of Binder Treated Powder Metallurgy Steels”, Metall. Mater. Trans. A, Vol. 33A, 2002, pp. 73-81.

9. X. Deng, G. Piotrowski, N. Chawla, K.S. Narasimhan, “Effect of Pore Clustering on the Mechanical Behavior of Powder Metallurgy (P/M) Steels”, P/M Sci. & Tech. Briefs, In Press, 2004.

10. Sonsino, C.M., “Fatigue Design for Powder Metallurgy”, Met. Powder Report, Vol. 45, No. 11, pp. 754-764.

11. Christian, K.D., German, R.M., “Relation Between Pore Structure and Fatigue Behavior in Sintered Iron-Carbon-Copper”, Intl. J. Powder Metall., Vol. 31, No. 1, 1995, pp. 51-61.

12. Holmes, J., Queeney, R.A., “Fatigue Crack Initiation in a Porous Steel”, Powder Metall., Vol. 28, 1985, pp. 231-235.

13. Gerard, D.A., Koss, D.A., “The Influence of Porosity on Short Fatigue Crack Growth at Large Strain Amplitudes”, Int. J. Fatigue, Vol. 13, No. 4, 1991, pp. 345-352.

14. F.J. Semel, “Properties of Parts Made from ANCORBOND Processed Carbon-Nickel-Steel Powder Mix”, Adv. Powder Metall. Part. Mater., Metal Powder Industries Federation, Princeton, NJ, 1989, p.9.

15. S.H. Luk, J.A. Hamill Jr., “Dust and Segregation-Free Powders for Flexible P/M Processing”, Adv. Powder Metall. Part. Mater., Metal Powder Industries Federation, Princeton, NJ, 1993, p.153.

16. Tada, H., Paris, P.C., Irwin, G.R., The Stress Analysis of Cracks Handbook, Del Research Corporation, Hellertown, 1973.

17. Marchand, N., Parks, D.M., Pelloux, R.M., “KI-Solutions for Single Edge Notch Specimens Under Fixed End Displacements”, Int. J. Fracture, Vol. 31, 1986, 54-65.

18. Ahmad, J., Papaspyropoulos, V., Hopper, A.T., “Elastic-Plastic Analysis of Edge-Notched Panels Subjected to Fixed Grip Loading”, Eng. Fracture. Mech., Vol. 38, No. 4/5, 1991, pp. 283-294.

19. Komenda, J., “Automatic Recognition of Complex Microstructures Using the Image Classifier”, Mat. Char., Vol. 46. 2001, pp. 87-92.

20. Prasad, S.N., Mediratta S.R., Sarma, D.S., “Influence of Austenitisation on the Structure and Properties of Weather Resistant Steels”, Mat. Sci. & Eng. A, Vol. A358, 2003, pp. 288-297.

21. Bertilsson, J., Karlsson, B., “Powder Metallurgy Group Meeting”, Powder Metallurgy Group, Buxton, 1986, 183.

22. Fleck, N.A., Smith, R.A., “Use of Simple-Models to Estimate Effect of Density on Fracture-Behavior of Sintered Steel”, Powder Metall., Vol. 3, 1981, pp. 121-125.

23. Mellanby, I.J., Moon, J.R., “Fatigue of Inhomogeneous Low-Alloy P/M Steels”, Powder Metall., Vol. 32, No. 3, 1989, pp. 209-214.

24. Paris, P.C., Gomez, M.P., Anderson, W.E., “A Rational Analytic Theory of Fatigue”, The Trend in Engineering, The University of Washington, Seattle, Vol. 13, No. 1, 1961, 9-14.

25. Sadananda, K., Vasudevan, A.K., Kang, I.W., “Effect of Superimposed Monotonic Fracture Modes on the ∆K and Kmax Parameters of Fatigue Crack Propagation”, Acta Mat., Vol. 51, 2003, pp. 3399-3414.

26. Vasudevan, A.K., Sadananda, K., “Classification of Fatigue-Crack Growth- Behavior”, Metall. & Mat. Trans. A, Vol. 26, No. 5, 1995, pp. 1221-1234.

27. Vasudevan, A.K., Sadananda, K., “Fatigue Crack Growth Behavior of Titanium Alumindes”, Mat. Sci. & Eng., Vol. 192/193, 1995, pp. 490-501.

28. Suresh, S., “Crack Deflection - Implications for the Growth of Long and Short Fatigue Cracks”, Metall. Trans. A, Vol. 14A, No. 11, 1983, pp. 2375-2385.