En 353

92
Chapter 4
RESULTS AND DISCUSSION

The results obtained by performing various experiments, as described in Chapter 3,
for different steel grades are presented in this chapter. On the basis of results obtained, an
effort has been made to understand the effect of WC/C PVD coating on the fatigue
behaviour of different low-alloy steels investigated in this study.
4.1
Mechanical properties
Tensile and hardness tests were conducted on various steel specimens according to
the procedures described in section 3.3. The results obtained are presented and discussed in
the following sub-sections.
4.1.1 Tensile properties

Standard tensile test specimens of various steel grades were tested on a universal
testing machine to obtain their stress-strain diagrams, the values of yield and ultimate
strengths. All the steels tested exhibited similar stress-strain behaviour but with slightly
different values of yield and ultimate strengths. A representative stress-strain diagram
obtained for SAE8620 steel in green state is depicted in Figure 4.1.
In order to determine the elastic coefficients for various grades of steels in the green
state, tensile tests were conducted on a tensile testing machine by gradually applying the
load in small installments and recording the strain with the help of a mechanical
extensometer, as explained in section 3.3.1. The results obtained by conducting tensile tests
on various grades of steels are summarized in Table 4.1.
93
Figure 4.1: Stress-Strain graph for SAE8620 steel in green state
Table 4.1: Results of tensile tests conducted on various steel grades in green state
Steel Grades
Mechanical Properties
SAE8620
20MnCr5
EN353
SCM420
218
268
190
225
Elastic Constant (GPa)
390
375
320
365
Yield Strength (MPa)
659
682
580
665
Ultimate Strength (MPa)
4.1.2 Microhardness
Following the procedure outlined in section 3.3.2, micro-hardness tests were
conducted on transverse cut-sections of the case carburized (uncoated) and case carburized
- WC/C coated specimens, referred to as uncoated and coated specimens, respectively
hereafter. The micro-indentations were made on the cut cross-section under a load of 200gf
for a dwell time of 15s. The micro-hardness values were estimated by measuring diagonals
of the indentations. The variation of hardness, observed for different steel specimens in the
uncoated and coated states, on moving from the surface towards the core, is shown in
Figure 4.2.
94
Figure 4.2: Effect of coating on microhardness profiles of various steels
95
As ASTM E 384 [263] recommends availability of sufficient backup material at the

site of indentation, the hardness measurements for the outer surfaces of the uncoated and
coated specimens were taken on their flat ends, rather than transverse cut-sections.
Hardness of the coating is estimated through microindentation measurements performed on
the flat coated surface under a load of 10gf and a dwell time of 15s. The hardness of the
coating is found to be HV1504174.
It is evident from Figure 4.2 that in all the steels, the coating leads to a loss in
hardness of around 35HV near the surface, which diminishes gradually on moving towards
the core. This is attributed to the processing of case carburized specimens during PVD
coating at a temperature of 220C, which is on the order of temperatures encountered
during low temperature tempering of low-alloy steels [268]. In general, beyond a depth of
around 1100 m from the surface, the loss of hardness becomes almost negligible. It
appears that owing to relatively low processing temperature during the coating process, the
loss in hardness of the steel specimens is marginal.
4.1.3 Characterization of WC/C coating through Raman Spectroscopy

The Raman spectra of the WC/C coating, obtained through procedure described in
section 3.8, are shown in Figure 4.3. The two graphs correspond to the spectra recorded at
two different locations on the flat end of a single specimen.
The Raman spectra of various allotrophic forms of carbon are well-documented in
literature [269,270]. The peak observed at 1580 cm-1 for both the locations corresponds to
the first order graphite peak (G) of sp2 bonded carbon atoms, while the peak at 1360 cm-1,
observed at location-2 refers to the disorder-induced peak (D). The peak observed at 2726
cm-1 (location-1) corresponds to the second order graphite peak (G ). The sharpness of G
and G peaks signifies the presence of highly oriented pyrolytic graphite (HOPG), which
undergoes significant broadening in case of disordered films [Refer location-2]. The
96
difference in spectra recorded at different locations in a single specimen is known to result

from the variations in local deposition and annealing conditions [108].
Figure 4.3: Laser Raman spectra of WC/C coating at two different locations.
Both the Raman spectra reveal the presence of sp2 bonding among carbon atoms
within the PVD layer, thereby confirming the presence of graphitic carbon in the deposit.
4.2
Testing and analysis of SAE8620 steel specimens

The effect of material processings, viz. case-carburization and deposition of WC/C
coating, on the fatigue performance of SAE8620 steel is presented in the following subsections. The results obtained by conducting fatigue tests are correlated with the influence
of these processes on various properties, viz. hardness, residual stresses etc., and
fractographic features.
97
4.2.1 Metallographic observations

Metallographic observations of the case carburized and tempered SAE8620 steel
specimens reveal the presence of martensitic structure of plate morphology in the case, as
shown in Figure 4.4 (a). Internal oxidation along boundaries of prior austenitic grains can
be seen extending to a depth of about 8-10 m in this figure. The carbon-rich case is also
dotted with carbide precipitates that appear bright in the dark anodic film deposited by
Klemms I reagent, as indicated by arrow marks in Figure 4.4 (b). Tempering of the
carburized steel is responsible for the formation of these spherodized carbides, which inturn leads to a reduction in the carbon content of austenite. The carbide particles help in
maintaining fine grains by inhibiting their growth. Besides, lower amounts of carbon in
austenite facilitate greater martensitic transformation by lowering the martensitic start
temperature (Ms). At some locations, chunks of carbide precipitates, segregated along prior
austenitic grain boundaries, can also be observed in Figure 4.4 (c) and (d). The core was
dominated by lath martensite [Refer Figure 4.5].
Based on the composition of SAE8620 steel given in Table 3.1 the martensitic start
temperature (Ms) of this steel can be calculated from the following equation, as suggested
by Payson and Savage [271]:
M S ( F ) = 930 570C 60Mn 50Cr 30 Ni 20Si 20Mo
Equation 4.1
From this relationship, the Ms temperature of SAE8620 steel works out to be

728.7 F, or 387 C. The estimated value of Ms temperature corresponds to the core of
carburized specimens, where chemical composition remains essentially unaltered. Once Ms
temperature is known, the expected volume fraction of martensite ( vM ) can be calculated
from the following relationship, which was put forth by Koistinen and Marburger [272]
for low-alloy steels:
vM = 1 e 0.011( M s T )
where T is the quenching temperature.
Equation 4.2
98
Figure 4.4: Metallographs of case carburized SAE8620 specimens cross-section:

(a): Martensitic structures in the case, revealed by etching in 3% Nital for 10s;
(b): Chunks of carbides at a depth of 300m, (c) and (d): Pictures reproduced
in true colour to reveal carbide segregates along prior austenitic grainboundaries. Specimens in (b), (c) and (d) etched face-up for 3 min in
Klemms - I reagent;
Figure 4.5: Metallographs showing the presence of lath martensite in the core of SAE8620
specimen etched with 3% nital for 5 seconds, viewed using: (a): 10X
objective and (b): 40X objective.
It is evident from Equation 4.2 that near complete martensitic transformation would
occur when SAE8620 steel having Ms temperature of 387C is quenched from austenitic
99
range to 120C. Further, it also appears from the above discussion that substantial amounts
of retained austenite would be present within the outermost layers of carburized case.
However, contrary to this, the metallographic observations made on the carburized case
reveal a high volume fraction of plate martensite within the outermost 80m layer [Refer
Figure 4.4 (a)]. The low fraction of retained austenite observed in the outermost layers is
attributed to the following two factors. First, an increase in Ms temperature of the case,
resulting from loss in carbon from the outermost surface layers due to decarburization,
internal oxidation [273,274] and precipitation of carbides. The second factor responsible for
this phenomenon is the occurrence of macrostraining during quenching of SAE8620 steel
specimens from austenitic temperature. Due to negative carbon gradient within the
carburized case, the Ms temperature will increase with depth. As a consequence, the
martensitic transformation would first commence at some depth below the surface during
quenching, thereby leading to development of tensile residual stresses within the surface
layers, where the phase is still austenitic. These residual tensile stresses would in-turn
stimulate a more complete martensitic transformation near the surface [275,276].
The gradual variation in microstructure upon moving from the surface towards the
core is captured by means of a number of metallographs recorded at successive depths and
compiled in the form of a mosaic, shown in Figure 4.6. It can again be observed that up to a
depth of around 80m, the microstructure is largely dominated by plate martensite,
accompanied by small amounts of retained austenite. Upon moving further inwards, we
encounter a mixed plate and lath morphology of martensite. Beyond a depth of about
700m from the outermost surface, which corresponds to the effective depth of case, welldefined packets of lath martensite can be observed, which become more orderly upon
moving further towards the core. The high volume fraction of lath martensite observed in
the core of SAE8620 steel specimens is commensurate with the predictions made by using
Equation 4.2 [272]. Lath martensite is generally the preferred microstructure for the lowcarbon core, since it is known for its better properties of strength and toughness in
comparison to the ferrite-pearlite microstructure [277].
100
Figure 4.6: Mosaic of metallographs showing variation in microstructure with depth in

case carburized and tempered specimens made of SAE8620 specimen, etched
in 3% Nital for 10 seconds.
In some of the specimens, small amounts of lower bainite [278] were also observed in
some regions of the core [Refer Figure 4.7 (a)], as revealed from the micrographs of case
carburized and WC/C coated specimens shown in Figure 4.7 (b). During the etching of
101
polished specimen with 3% nital for 2 hours, the etchant attached the martensitic matrix at a
higher rate than the lower bainitic phase, thereby leading to formation of raised
microstructures [Figure 4.7 (b)], which are further highlighted by means of an additional
source of light placed towards lower left side of the specimen, during the capture of
micrograph. The presence of small blow-holes in the material is also revealed through
scanning electron micrograph shown in Figure 4.7 (c).
The EDAX elemental analysis of the specimen [Figure 4.7 (d)], recorded at the
position marked with cross-hair in Figure 4.7 (c), revealed the dominance of Fe, Mo and C,
with their respective wt. percentages as 75.87%, 13.7% and 7.87%.
Figure 4.7: (a): Cross-section of case carburized, tempered and coated specimen etched in
3% nital for 2 h, followed by light polishing; (b): Magnified view of region A
in figure (a) showing the presence of lower bainite in martensitic matrix; (c):
Magnified view at location B in figure (a); and (d): Result of EDAX
elemental analysis performed at location marked with cross-hair in (c).
102
4.2.2 Estimation of residual stresses

The residual stress-field present in the material plays an important role in determining
its fatigue performance. The residual stresses in the specimens were estimated through Xray diffraction technique, following standard procedures described in sections 2.5.2 and 3.7.
On the basis of location and intensity, the peak corresponding to (211) plane (ref. Figure
3.26) was found to be suitable for the purpose of estimating lattice strain. The peak
exhibited broadening, which is a commonly observed phenomenon in most of the hardened
and fine-grained steels [234]. The X-ray diffraction pattern was recorded at an interval of
0.0098 with a total of 255 readings in the 2 range from 80.99 to 83.50 in the tilt range
() from 40 to +40, while the divergent slit was kept at 2mm. Parabolic curve fitting
was employed for determining the position of peaks at individual tilts. The results of X-ray
diffraction analysis, performed on uncoated and coated specimens, are summarized in Table
4.2 and shown graphically in Figure 4.8. Psi-splitting has been observed in the d vs sin2
plot [Refer Figure 4.8] for both the specimens, which signifies the presence of tri-axial
stresses near the surface of specimens. The shape of both these plots, however, rules out the
presence of texturing in the material [220,221].
Table 4.2: Determination of lattice strain in SAE8620 steel specimens through X-ray
diffraction
S. No.
Uncoated Specimen
Coated Specimen
Tilt,
sin2
d-spacing
Tilt,
sin2
d-spacing
1
-40.00
0.4132
1.17060
-40.00
0.4132
1.17113
2
-35.09
0.3305
1.17079
-35.09
0.3305
1.17152
3
-29.86
0.2479
1.17098
-29.86
0.2479
1.17195
4
-22.99
0.1525
1.17122
-22.99
0.1525
1.17239
5
-16.71
0.0827
1.17148
-16.71
0.0827
1.17282
6
0.00
0
1.17184
0.00
0
1.17334
7
16.71
0.0827
1.17186
16.71
0.0827
1.17303
8
22.99
0.1525
1.17175
22.99
0.1525
1.17268
9
29.86
0.2479
1.17162
29.86
0.2479
1.17228
10
35.09
0.3305
1.17147
35.09
0.3305
1.17189
11
40.00
0.4132
1.17129
40.00
0.4132
1.17151
103
Residual stress () corresponding to the recorded peak positions is estimated from

Equation 3.4, as reproduced below:
1 d
E
2
1 + ( hkl ) d 0 sin
To estimate the value of material constant, K
[= E / (1 + )], of the martensitic case,
we have used E = 208 GPa and = 0.29 from the reported literature [266]. The estimated
value of K (= 161.2 GPa) is in fair agreement with the value (167.5 GPa) reported in the
literature for 4820 steel under carburized condition [220]. Using these constants, the value
of residual stress () is estimated as -305.62.9 MPa for uncoated and -684.63.1 MPa for
coated specimens. The fairly high magnitude of residual compressive stresses observed in
the case-carburized condition is attributed to the high degree of martensitic transformation
near the surface, as reported in section 4.2.1. After coating the carburized specimen with
WC/C layer, the magnitude of residual compressive stresses becomes more than double.
Figure 4.8: d vs sin2 plot for uncoated and coated specimens made of SAE8620 steel
104
4.2.3 Fatigue testing

The results obtained by conducting 4-point rotating bending fatigue tests on standard
fatigue specimens made from SAE8620 steel in the green state are reported in Table 4.3.
The average number of cycles to failure corresponding to a given stress level is denoted by
Nf , which is estimated by taking the average of N1 and N2. The results obtained by
performing fatigue tests on case carburized specimens in uncoated and coated conditions
are reported in Table 4.4 and Table 4.5 respectively.
Table 4.3: Results of fatigue tests conducted on specimens made of SAE8620 steel in
green state
Stress (MPa)
Cycles to failure
Average
Std. Dev.
Green
(Nf
)
( Green)
N1
N2
425
10164
13472
11818
2339
396
14136
20163
17150
4262
365
40357
52341
46349
8474
327
109370
149867
129619
28636
300
236783
352208
294496
81618
279
356783
542760
449772
131506
262
583675
882430
733053
211252
Table 4.4: Results of fatigue tests conducted on case-carburized and tempered (uncoated)
specimens made of SAE8620 steel
Stress (MPa)
Cycles to failure
Average
Std. Dev.
Uncoated
(N
)
( Uncoated)
N1
N2
f
1000
21341
26178
23760
3420
970
36756
51451
44104
10391
930
78824
112570
98697
23862
902
116280
162092
139186
32394
850
292365
412878
352622
85216
810
547443
813357
680400
188030
800
665382
946726
806054
198940
Table 4.5: Results of fatigue tests conducted on case-carburized, tempered and WC/C
coated specimens made of SAE8620 steel
Stress (MPa)
Cycles to failure
Average
Std. Dev.
(Nf WC/C)
( WC/C)
N1
N2
995
10502
15325
12913
3412
975
17783
25578
21681
5512
960
30451
36562
33507
4321
935
58749
78637
68693
14063
910
98834
152016
125425
37605
885
240679
336284
288482
67603
860
561146
838254
699700
195945
105
In order to quantify the change in fatigue life as a function of maximum alternating

stress, a linear model of the form: Y = A + BX [Refer Equation 2.5] has been fitted
separately to the fatigue data obtained for specimens in three different states (i) green, (ii)
case-carburized and (iii) case carburized, followed by WC/C coating. The term Y in the
model equation represents logarithm of the number of cycles to failure (Nf) while X denotes
the maximum alternating stress (S) in MPa. The parameters A and B in the linear model are
estimated through regression analysis, as prescribed in ASTM E 739 standard [121]. The
estimated values of parameters are summarized in Table 4.6.
Table 4.6: Values of parameters in the linear model for fatigue life of SAE8620 steel
Material & State
Parameter A
Parameter B
SAE8620 (Green)
8.839
-0.0114
SAE8620 (Case carburized)
11.923
-0.0075
SAE8620 (WC/C coated)
16.659
-0.0126
The experimental fatigue data, along with the fitted linear models and 95%
confidence bands, are plotted on semi-log scale, i.e. Log (Nf) vs Stress (S), in Figure 4.9 for
SAE8620 steel specimens in green, case-carburized (uncoated) and case carburizedWC/C
coated states. The parameters A and B appearing in the linear model are dependent on the
bulk and surface properties of the material apart from the testing conditions. These
parameters can be of prime importance while estimating the fatigue life of SAE8620 steel
in green, uncoated and coated states, processed in a fashion similar to that employed in the
present work.
The endurance limit of the specimens in various states was estimated through
staircase tests conducted for 2106 cycles. The results obtained are summarized in Table
4.7. Taking the number of less frequent event (i.e., survival or failure) as the basis for
statistical analysis of these staircase tests, the data reduction is carried out according to the
Dixon-Mood method [Refer section 2.3.2.4]. The intermediate calculations and final results
thereof are given in Table 4.8.
106
Table 4.7: Results of staircase tests conducted on specimens made of SAE8620 steel
S. No.
Green
Case Carburized
WC/C Coated
Stress
(MPa)
Cycles
Stress
(MPa)
Cycles
Stress
(MPa)
Cycles
279
356783
810
813357
885
240679
262
2000000*
800
665382
860
838254
279
542760
790
2000000*
835
2000000*
262
882430
800
2000000*
860
2000000*
245
2000000*
810
547443
885
336284
262
583675
800
946726
860
561146
245
2000000*
790
879452
835
2000000*
---
---
780
2000000*
---
---
* No failure; Test suspended.
Table 4.8: Results of Dixon-Mood statistical analysis of staircase test data obtained for
SAE8620 steel specimens
i
Green
Case Carburized
WC/C Coated
0 S0 = 245 MPa nDM,0 = 2 S0 = 780 MPa nDM,0 = 1 S0 = 835 MPa nDM,0 = 2
S1 = 262 MPa
nDM,1 = 1
S1 = 790 MPa
nDM,1 = 1
S1 = 860 MPa
nDM,1 = 1
S2 = 279 MPa
nDM,2 = 0
S2 = 800 MPa
nDM,2 = 1
S2 = 885 MPa
nDM,2 = 0
---
---
S3 = 810 MPa
nDM,3 = 0
---
---
Stress interval (d)= 17 MPa
Stress interval (d)= 10 MPa Stress interval (d)= 25 MPa
ADM = (i )(n DM ,i ) = 1
ADM = (i )(n DM ,i ) = 3
ADM = (i )(nDM ,i ) = 1
BDM = (i 2 )(n DM ,i ) = 1
BDM = (i 2 )(n DM ,i ) = 5
BDM = (i 2 )(n DM ,i ) = 1
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
( n )
DM ,i
( n )
DM ,i
( n )
DM ,i
= 0.22 (<0.3)
= 0.67 (0.3)
= 0.22 (<0.3)
Mean stress (s) = 259 MPa
Mean stress (s)= 795 MPa
Std. dev. (s) = 9.01 MPa
Std. dev. (s)= 13.25 MPa
107
An interesting phenomenon observed in Figure 4.9 is the cross-over in S-N curves of

the uncoated and coated specimens at a stress cycle of around 105, which implies that the
coated specimens show poor life in comparison to their uncoated counterparts in the low
cycle fatigue regime, but exhibit better performance in the high cycle regime. The lower
slope of S-N curve observed in Figure 4.9 for the coated specimens is also apparent from
the lower value of parameter B observed for WC/C coated specimens [Refer Table 4.6]. Su
and co-workers [10,11] in their studies on fatigue behaviour of various PVD coatings
applied to normalized AISI 1035 steel, also reported better performance of the coated
specimens in comparison to their uncoated counterparts for stress cycles higher than 5105.
Figure 4.9: S-N graphs for SAE8620 steel specimens in green, case carburized (uncoated)
and case carburized - WC/C coated states.
108
The effect of case carburization as well as WC/C coating on the fatigue strength of
SAE8620 steel has been estimated by comparing the fatigue strength of case carburized
(uncoated) and case carburized WC/C coated specimens with that of the specimens in
green state. The percentage change in fatigue strength at a given number of stress cycles is
computed from the relationship given below:
% Change in fatigue strength =
S f S Green
f
S Green
f
100
Equation 4.3
/C
where S f refers to the fatigue strength of either uncoated ( S Uncoated
) or coated ( S WC
)
f
f
specimens. It denotes the magnitude of maximum alternating stress at which the specimens
would fail after the given number of cycles (Nf). The value of S f has been computed from
the linear fitted model, given in Equation 2.5, developed for steel specimens in different
states, viz. green, case carburized (uncoated) and case carburized WC/C coated. To
estimate S f , Equation 2.5 has been re-arranged as:
Sf =
log(N f ) A
B
Equation 4.4
The values of parameters A and B for different types of steel specimens are
substituted from Table 4.6. The percentage change in fatigue strength has been computed
for different number of stress-cycles, ranging from 103 to 106. The results obtained are
summarized in Table 4.9 and depicted graphically in Figure 4.10.
Table 4.9: Effect of case carburization and subsequent WC/C coating on the fatigue
strength of SAE8620 steel.
No. of
SfGreen
SfUncoated
SfWC/C
%Change in Fatigue Strength
Cycles to
A=11.9234 A=16.6587
Case
WC/C
failure, Nf A=8.8392
Carburized
Coated
B=-0.0114
B=-0.0075
B=-0.0126
vs Green
vs Green
1000
512
1187
1080
132
111
10000
424
1054
1001
148
136
100000
337
921
922
174
174
1000000
249
788
843
216
239
109
It can be observed from Figure 4.10 that both uncoated and coated specimens show
substantial improvement in fatigue performance as compared to specimens in green state.
This enhancement in fatigue performance is attributed to the presence of residual
compressive stresses within the surface layers of the uncoated as well as coated specimens.
Therefore, the sequence of martensitic transformation [Refer section 4.2.1], which affects
the magnitude of residual stresses in the outermost surface [250], in-turn, also influences
the resulting fatigue properties. The presence of residual compressive stress delays the
crack propagation within the uncoated and coated specimens, thereby increasing their
fatigue performance in comparison to steel specimens in green state. For both types of
specimens, the percentage gain in fatigue strength increases with increasing number of
cycles. The increase observed for the coated specimens is steeper than that observed for
uncoated specimens. For bending stresses higher than 922 MPa, which corresponds to a
fatigue life of 105 cycles, the case carburized (uncoated) specimens exhibit better fatigue
performance than their WC/C coated counterparts. The presence of WC/C PVD coating on
the case carburized specimens results in improved fatigue performance for stress levels
below 922 MPa (Nf > 105 cycles). The gain observed in fatigue strength of the WC/C
coated specimens increases with increasing number of cycles upto the endurance limit of
the coated specimens.
The beneficial effect of PVD coating on fatigue performance is also reported by
Puchi-Cabrera et al. [197]. They reported 3 to 20 times enhancement in finite fatigue life of
AISI 316L stainless steel specimens when they were coated with PVD layer of TiN. In their
study, the fatigue strength of the coated specimens was observed to be 22% higher than that
of the uncoated specimens.
110
Figure 4.10: Percentage change in fatigue strength of uncoated and coated SAE8620 steel
specimens with respect to specimens in green state.
4.2.4 Fractographic observations

In order to identify the mode of fatigue failure and reveal the topography of fatigue
fracture surfaces, fractographic observations were made on the fatigue specimens in various
states, viz. green, case carburized and WC/C coated, tested at different stress levels. The
results obtained are presented and discussed in the following sub-sections.
4.2.4.1
Fractographic observations on specimens in green state
Fractographic observations on specimens in green state, failed at different stress

levels, reveal that different crack propagation mechanisms operate at different levels of
remote loading stress. The fracto-macrographs of the fracture-surfaces of the specimens,
arranged in increasing order of applied stress, are shown in Figure 4.11(a-g).
111
Figure 4.11 (a) and (b) show the fracture surfaces obtained at relatively low stress
levels of 262 and 279 MPa respectively. On a macroscopic scale, surfaces in both the
figures are oriented normal to the main principal stress (longitudinal axis of the specimen)
and exhibit planar topography with gradually increasing roughness as the crack propagates.
The region of crack propagation under stage-II [94] is characterized by smooth and bright
fractured surface, vividly captured in Figure 4.11 (b), where a bright, crescent-shaped
region opposite to the zone of final fracture has been demarcated with a dotted curve. The
portion between this dotted line and the region of final fracture exhibits higher roughness,
characteristic of rapid crack propagation (Stage-III). A few white patches appearing in this
zone are the result of rubbing that took place between the two halves of the specimen
during the course of failure. The area of final fracture is rather small in these cases owing to
relatively lower applied stress. The geometry and macroscopic orientation of the fractured
surface implies mode-I (tensile) crack propagation [279].
The examination of fractured specimen at low magnification revealed that the
outermost region opposite to the site of final failure exhibited near specular finish in a band
of approximately 0.5mm width, which is attributed to the initial, slow pace of crack
propagation. The same region also exhibited a number of ratchet marks [Refer Figure 4.12],
which indicate the occurrence of multiple fatigue crack nucleation sites [280]. The
examination of the specimen under SEM [Figure 4.13] reveals the formation of wedgeshaped extrusion sliver (marked with an arrow), resulting from slip bands within the
crystallites, typical of stage-I crack propagation [93,281,282].
112
Figure 4.11: Fatigue fracture surfaces of SAE8620 steel specimens in green state, tested at
(a): 262 MPa (b): 279 MPa (c): 300 MPa (d): 327 MPa (e): 365 MPa (f): 396
MPa (g) and (h) 425 MPa.
113
Figure 4.12: Optical micrograph of specimen tested at 279MPa, showing ratchet marks and
transgranular crack propagation.
Figure 4.13: SEM image of fatigue specimen tested at 279 MPa, showing transgranular
crack propagation, along with ratchet mark and extrusion sliver on the outer
surface (identified with arrow-mark).
114
Figure 4.14: SEM image of fractured specimen tested at 279 MPa, showing fatigue
striations on multiple plateaus
At some places in the region of mode-I crack propagation, fatigue striations could be
observed on the fractured surface, as revealed in the scanning electron micrograph given in
Figure 4.14. The striations are found to be bowed out roughly in the direction of crack
propagation and their topography consists of multiple plateaus. The SEM image is taken in
the stage-II region at a place midway between the crack initiation site and the region of
rapid fracture. Unlike stage-I, where striation spacing can be much more than the actual
crack growth rate [98], the stage-II striations are generally formed during each loading
cycle. The observed striation spacing, which is on the order of 0.5m, indicates a relatively
short duration of stage-II crack propagation phase (5,000-10,000 cycles) in comparison to
the total fatigue life of the specimen. Thus, it can be concluded that in case of specimens
tested under high cycle fatigue, stage-I crack propagation phase constitutes the major
proportion of fatigue life.
As the applied stress is increased beyond 279 MPa, the topography of fractured
surfaces transforms to that shown in Figure 4.11 (c) and (d), which were observed in
115
specimens tested at stress levels of 300 and 327 MPa respectively. The fractured surfaces
reveal formation of shear lips, inclined at an angle of approximately 45 with the transverse
section of the specimens. Therefore, with the increase in stress level beyond 279 MPa, the
mode of failure shifts from tensile (mode-I) to shear (mode-II), thereby leading to formation
of shear lips. Though, the formation of shear lips is not very well understood, but their
occurrence in metals having FCC or BCC lattice structure is already known and has been
extensively reported for flat, plate-shaped fatigue specimens made of aluminium alloys
[283,284] and to a lesser extent for steel plates [285]. However, the literature consulted so
far does not reveal the occurrence of shear lips in case of cylindrical steel specimens. In
general, the formation of shear lips is attributed to the fact that owing to lesser restraint on
the material at the free surface in comparison to that in the bulk, a large plastic deformation
occurs in the region where the crack front meets the free surface, which in turn promotes
the formation of shear lips [109]. The width of shear lips in the present study does not reach
an equilibrium state but continues to grow, which is in agreement with the model proposed
by Zuidema and Blaauw [286].
The advancing intersection of the crack front with the free surface is marked with
arrows in Figure 4.11 (c). As one moves in the direction of arrows, a very well defined start,
followed by a gradual increase in width of the shear lips is observed. A similar feature can
also be noticed in Figure 4.11 (d), where two mutually perpendicular shear lips are
observed at +45 and -45, starting point of one of which is identified with an arrow.
Double shear lips are known to cause a deviation in the direction of crack propagation, as
described by Schijve and co-workers [109,287]. However, in the present study, owing to
cylindrical shape of the specimens, the ridge formed between the two shear lips naturally
follows a helical path, instead of the straight-line path, as reported for flat specimens by
Schijve and co-workers.
Figure 4.11 (e) represents a more complex fractured surface of the specimen tested at
365 MPa, where crack nucleation and macroscopic growth occurred on multiple planes.
Although, the cracks could not be observed during the course of tests, but the final fractured
116
surface reveals cracks originating on multiple planes, which subsequently merged with each
other. At various sites, the crack propagation has occurred through mixed modes I and II,
while the region of stable crack growth has been reduced significantly in the wake of
relatively larger applied stresses.
When the stress level is further increased to 396MPa, which is on the order of yield
strength of the SAE8620 material [Refer Table 4.1], an interesting topography, resembling
a helical staircase emerges, as revealed in Figure 4.11 (f). As the material near the outer
surface of the specimen is subjected to severe plastic deformation, the transition of crack
propagation from modeI to modeII also initiates at the outer periphery. It appears that the
unstable crack front has led to the formation of a helical staircase topography on the
fractured surface. Observations under scanning electron microscope [Refer Figure 4.15]
reveal fibrous appearance of the fractured surface, which is attributed to modeII
transgranular crack growth [288]. Fatigue striations could not be observed anywhere on the
fractured surface, though it exhibited tire-tracks at a few places, as shown in Figure 4.16.
A further increase in stress level to 425MPa, which is well above the static yield
strength of SAE8620 steel, causes the entire material within the gauge length of the
specimen to undergo plastic deformation. The extension of plastic zone from the surface
right up to the center of the cylindrical specimen is confirmed from the topography of the
fractured surface shown in Figure 4.11 (g). The region exhibiting unstable crack front,
which was restricted to the outer periphery in case of specimen tested at an applied stress of
396 MPa [Refer Figure 4.11 (f)], now gets extended up to the neutral axis. The specimen
tested at 425 MPa also exhibited the formation of multiple macroscopic cracks, some of
which coalesced and led to the final fracture, as revealed in the side view of the specimen in
Figure 4.11 (h). For the specimens tested at stress levels of 396 and 425MPa, the plastic
straining was also accompanied by hysteresis losses within the material, thereby leading to
a considerable rise in temperature of the specimens.
117
Figure 4.15: SEM image of crack geometry under mode-II in fatigue specimen tested at
396 MPa.
Figure 4.16: SEM image of fractured specimen tested at 396 MPa, showing presence of tire
tracks
118
4.2.4.2
Fractographic observations on case-carburized specimens
The fracture surfaces of the case carburized (uncoated) and case carburized WC/C
coated specimens failed at different levels of alternating stresses were examined optically as
well as under scanning electron microscope [Refer Figure 4.17 (a) - (h)]. The examinations
were carried out to investigate the crack propagation mechanisms operative in uncoated and
coated specimens and to develop an understanding of the effect of coating on fatigue
behaviour of the case carburized SAE8620 steel.
Figure 4.17 (a) shows the optical fractograph of uncoated fatigue specimen, tested at
810 MPa, where arrow points the region of crack initiation. The crack is observed to
propagate nearly normal to the direction of principal stress (which is also the direction of
specimen axis). Almost diametrically opposite to the crack initiation site, a small step is
formed on the fractured surface due to merging of the two crack-fronts, which were slightly
out-of-plane with respect to each other. Fractographic studies conducted at higher
magnification revealed that the cracks originated either at some defect on the outermost
surface [Refer Figure 4.17 (b)] or through intergranular decohesion along prior austenitic
grain boundaries [Refer Figure 4.18]. Crack initiation through intergranular decohesion is a
commonly observed fracture mode in the carbon rich case of the carburized steels and
occurs when the applied stress exceeds the sum of residual compressive macro-stresses and
cohesion strength of the material at prior austenitic grain boundaries [239].
The crack thus initiated, however, can not always continue to progress in the same
intergranular fashion and gets arrested as a result of small plastic zone (smaller than the
grain size) at the tip of these sharp intergranular cracks. The presence of plastic zone leads
to strain-induced transformation of retained austenite, present near the outer surface of the
carburized specimen, into martensite, which further induces compressive stresses at the
crack tip [289]. As a result, the crack now propagates through stable transgranular mode,
till it reaches a critical size, whereupon it propagates as unstable fracture.
119
The dominance of transgranular mode of crack propagation is clearly revelaed in

Figure 4.17 (a) and (b). In Figure 4.17 (b), the region of stable crack propagation appears in
distinct contrast to the remaining region of unstable crack propagation and can be employed
for estimating fracture toughness of the material [Refer section 2.6].
It is interesting to note that despite the presence of large residual compressive stresses
on the order of 306 MPa, the crack initiation still occured in the outermost case-carburized
layer and not at the root of the carburized case. Similar observations regarding crack
initiation are noticed by other researchers [151, 290]. The initiation of cracks at the surface
reveals the importance of surface conditions in deciding the fatigue life despite the presence
of large residual compressive stresses at the surface.
The fractographs recorded on uncoated fatigue specimen tested at a relatively higher
stress of 1000 MPa reveal that the fractured surface is again dominated by the region of
transgranular mode of crack growth, as is evident from Figure 4.17 (c), where the crack
initiation site is marked with a circle. The regions marked with rectangles exhibited fatigue
striations and tire tracks, as is evident from the scanning electron micrograph shown in
Figure 4.17 (d), which is recorded at some location within the regions marked with
rectangle.
The top and side views of the fractured surface of the coated fatigue specimen tested
at 860 MPa are shown in Figure 4.17 (e). The fracture surface of the coated specimen is
also observed to be dominated by transgranular mode of crack propagation, similar to that
observed in Figure 4.17 (c) for uncoated specimen tested at 1000 MPa. The magnified view
of the crack initiation region marked in Figure 4.17 (e) is shown in Figure 4.18, where the
dominance of transgranular crack propagation mechanism is quite apparent. The crack
originates through intergranular decohesion, but immediately switches to transgranular
mode, whereupon it propagates circumferentially in both the directions and finally meets at
the diametrically opposite side, where it reaches along two slightly different planes, marked
as 1 and 2 in Figure 4.17 (e).
120
Figure 4.17: Fractographs of specimens made of SAE8620 steel: (a): Optical fractograph of
uncoated fatigue specimen tested at 810 MPa, showing crack initiation site; (b):
Scanning electron micrograph of the location marked by arrow in (a); (c): Optical
fractograph of uncoated specimen tested at 1000 MPa, showing crack initiation
site; (d): Scanning electron micrograph showing striations and tire tracks at a
location within the regions marked by rectangles in (c); (e): Top and side views of
coated fatigue specimen tested at 860 MPa; (f): Scanning electron micrograph
showing adherence of coating at the failed section marked with rectangles in (e);
(g): Optical fractrograph of coated fatigue specimen tested at 910 MPa, showing
ratchet marks and crack initiation site; (h): Scanning electron micrograph at a
location within the region marked with rectangle in (g), showing tire tracks.
121
Scanning electron micrograph of the region marked with a rectangle in Figure 4.17
(e), is shown in Figure 4.17 (f). The micrograph reveals that the coating remains bonded
even at the thin tip of the failed specimen. The examination of fractured surfaces of the
other specimens also indicates that no delamination of coating from the substrate wall
occurs. Earlier investigation reported by Carvalho [291] also confirms the absence of
delamination in case of WC/C coatings applied to steel substrates.
The fractured surface of the coated fatigue specimen tested at a relatively high stress
of 910 MPa is characterized by the regions of unstable crack growth, along with a large
number of ratchet marks, as identified by arrows 1 and 2 in Figure 4.17 (g). The arrow
marked 3 in Figure 4.17 (g) shows the crack initiation site. The presence of tire tracks is
clearly revealed in scanning electron micrograph shown in Figure 4.17 (h), which is taken
in the region marked with a rectangle in Figure 4.17 (g).
The comparison of crack-initiation sites observed for uncoated and WC/C coated
specimens, as shown respectively in Figure 4.17 (b) and Figure 4.18, reveals that similar
crack initiation and growth mechanisms operate in both the specimens. The overall crack
propagation is found to be transgranular in nature, while the region of final ductile fracture
exhibits microvoid coalescence morphology. Measurements made on the stable crack
propagation region [Refer section 2.6] shown in Figure 4.17 (b) provide depth, a = 322m
and width, c = 911 m, on the basis of which, fracture toughness [Refer Equation 2.21]
works out to be KIC = 25.59 MPa m , which is quite high for the steel in case carburized
condition [240].
An important observation is that the crack progresses largely through transgranular
fracture mode even through the high hardness case. As a result, there is no change in the
mode of propagation when the crack progresses to regions that are low in carbon content. In
materials, which are susceptible to intergranular cracking, a short-term, secondary stage of
fatigue crack propagation, characterized by transgranular mode of fracture, observable
fatigue striations and secondary cracking is often reported to occur prior to the final ductile
122
overload fracture [240,292]. However, no such transition from intergranular to

transgranular mode took place for case carburized specimens made of SAE8620 steel, as
the dominant mode of crack propagation was always transgranular.
Figure 4.18: Magnified view of crack initiation region of specimen shown in Figure 4.17
(e), depicting intergranular initiation and transgranular propagation.
For the coated specimen shown in Figure 4.17 (g), which was tested under low cycle
fatigue, the fracture surface exhibited crack-initiation on multiple planes, as shown in
Figure 4.19. Crack-initiation on multiple planes is also confirmed by the formation of
multiple plateus [Refer Figure 4.20] near the outermost edge of fractured surface.
These observations imply that formation of multiple cracks is encouraged under low
cycle fatigue when the specimen is in coated condition. This can be attributed to substantial
straining of the substrate under high applied stresses, which lead to formation of multiple
cracks within the coating, some of which grow to macroscopic scale. This phenomenon
seems to be responsible for relatively poor performance of the coated specimens under low
cycle fatigue. Otherwise, there is no significant difference in the surface topography as well
as associated crack propagation mechanisms for the uncoated and WC/C coated specimens,
particularly for specimens that failed at comparable number of cycles. This is in agreement
with the observation reported by Baragetti and Tordini [188]. The presence of coating
123
modifies the level of stress at which the failure would occur after a given number of load
cycles. In other words, the modes of failure remain more or less the same for both, uncoated
and WC/C coated specimens. However, the stress level for their occurrence gets modified
in the presence of WC/C coating.
Figure 4.19: Optical fractograph taken on side-wall of coated specimen shown in Figure
4.17 (g), revealing the formation of multiple cracks under low-cycle fatigue.
Figure 4.20: Close-up view of specimen shown in Figure 4.17 (g), revealing crack
formation on multiple planes.
124
4.2.5 Effect of WC/C coating on fatigue behaviour of SAE8620 steel

The observed change in the fatigue performance of the coated specimens, in
comparison to their uncoated counterparts, can be explained on the basis of residual stress
measurements reported in Section 4.2.2. The factor which seems to be responsible for
reducing the low-cycle fatigue strength of the coated specimens is the development of
multiple cracks on the hard WC/C layer, when the substrate undergoes considerable
deflection due to high applied loads under low cycle fatigue [Refer Figure 4.19 and Figure
4.20]. Some of these cracks propagate into the case-carburized layer of the substrate and
lead to early fatigue failure.
However, when it comes to high cycle fatigue, the coating seems to provide an
effective protection to the surface by suppressing the crack propagation, thereby leading to
considerable improvement in fatigue performance. The observed improvement in high
cycle fatigue performance of WC/C coated specimens is therefore attributed to higher
residual compressive stresses (685 MPa) in the coated specimens than those observed in the
case carburized (uncoated) specimens (306 MPa).
125
4.3
Testing and analysis of 20MnCr5 steel specimens

This segment of the study investigates the effect of case carburization and WC/C
coating on hardness, residual stress and fatigue performance of 20MnCr5 steel. For this
purpose, a comparative study has been performed on specimens made of 20MnCr5 steel in
green, case carburized (uncoated) and case carburized WC/C coated states, in a manner
similar to that followed for SAE8620 steel. The results obtained are presented and
discussed in the following sub-sections.

The metallographic observations performed on 20MnCr5 steel in green state revealed
the presence of fine pearlitic microstructures [Refer Figure 4.21] with interlamellar spacing
on the order of 0.5 m. The occurrence of fine pearlitic microstructure is attributed to the
presence of chromium and manganese in this steel [293], as revealed from Table 3.1.
Figure 4.21: Metallograph of 20MnCr5 steel (etched in 3% Nital for 4s) in green state
revealing the presence of pearlitic microstructure.
The microstructure of case carburized and tempered specimen of 20MnCr5 steel is

depicted in Figure 4.22. The microstructures are recorded across depth of the specimen and
compiled in the form of a mosaic.
126
Figure 4.22: Mosaic of metallographs, showing variation of microstructure with depth in

case carburized and tempered specimens made of 20MnCr5 steel, etched in
3% Nital for 8 seconds.
127
The microstructure within the outermost 400m is dominated by plate martensite,

dotted with segregates of carbide particles, which appear as white dots, scattered upto a
depth of 300m. These precipitates are vividly revealed upon etching with Klemms - I
reagent, as shown in Figure 4.23.
Figure 4.23: Metallograph of case carburized and tempered specimen made of 20MnCr5
steel, etched in Klemms - I reagent for 3 minutes, revealing the presence of
carbide particles in the carbon-rich case.
Apart from plate martensite, some amount of retained austenite can also be observed
upto a depth of around 400m in Figure 4.22. The amount of retained austenite reduces
significantly on moving from the outermost surface to a depth of around 400m, beyond
which, its decrease becomes marginal. This reduction in retained austenite with increase in
depth is attributed to a rise in Ms temperature, as a consequence of negative gradient in
carbon content.
The amount of retained austenite observed in case carburized and tempered
specimens made of 20MnCr5 steel is relatively higher than that observed in SAE8620 steel
specimens processed in similar fashion [Refer Figure 4.6]. It may be attributed to relatively
lower Ms temperature of 368C for 20MnCr5 steel, as estimated from Equation 4.1 [Refer
Table 3.1 for chemical composition] in comparison to that estimated for SAE8620 steel
128
(Ms = 387C). Due to lower Ms temperature, the transformation in the core of specimens
made of 20MnCr5 specimens would have commenced at a later stage, whose time would be
overlapping with the start of martensitic transformation at the surface. As a result of this
alteration in the sequence of martensitic transformation, the advantage of strain-induced
martensitic transformation would be harnessed to a lesser extent for specimens made of
20MnCr5 steel, in comparison to those made of SAE8620 steel. The significantly higher
amounts of retained austenite in the outermost layers, in comparison to those observed for
SAE8620 steel, provide an explanation for the considerably lower values of hardness
observed for 20MnCr5 steel specimens [Refer Figure 4.2].
Similar to SAE8620 steel, the microstructure of 20MnCr5 specimens also reveals a
mixed morphology of plate and lath martensite at a depth of 400 m to 600 m below the
surface. Beyond a depth of around 600 m, the microstructure is dominated by lath
martensite, which becomes more and more orderly with further increase in depth.

The residual stress in the uncoated and WC/C coated specimens made of 20MnCr5
steel was estimated through X-ray diffraction technique, as described earlier for SAE8620
steel (Ref. section 4.2.2). The results obtained are summarized in Table 4.10 and depicted
graphically in Figure 4.24. Psi-splitting is observed to occur to a greater extent than that
observed earlier for SAE8620 steel [Refer Figure 4.8], thereby implying higher magnitude
of triaxial stresses in the surface layers of 20MnCr5 steel specimens.
Residual stresses corresponding to the recorded peak positions are estimated from the
relationship given in section 3.7.2. From the recorded shift in peaks, the value of residual
stress () is estimated as -184.83.0 MPa for the uncoated and -449.23.0 MPa for the
coated specimens. The magnitude of residual compressive stresses in the uncoated
specimens made of 20MnCr5 steel is about 40% lower than that observed in the
corresponding specimens made of SAE8620 steel [Refer section 4.2.2]. This is attributed to
the presence of significantly higher amounts of retained austenite in the surface layers of
129
20MnCr5 steel specimens, as compared to that observed in SAE8620 steel specimens. The
deposition of WC/C coating, again led to enhance the magnitude of residual compressive
stresses, but they still remained 35% lower than the compressive residual stresses observed
in WC/C coated SAE8620 steel specimens.
Table 4.10: Determination of lattice strain in 20MnCr5 steel specimens through X-ray
diffraction
S. No.
Uncoated Specimen
Coated Specimen
2
d-spacing
Tilt,
sin2
d-spacing
Tilt,
sin
1
-40.00
0.41318
1.17081
-40.00
0.41318
1.17089
2
-35.09
0.33047
1.17092
-35.09
0.33047
1.17117
3
-29.86
0.24789
1.17105
-29.86
0.24789
1.17148
4
-23.99
0.16530
1.17120
-23.99
0.16530
1.1718
5
-16.71
0.08267
1.17137
-16.71
0.08267
1.17213
6
0
0
1.17169
0
0
1.17260
7
16.71
0.08267
1.17177
16.71
0.08267
1.17256
8
23.99
0.16530
1.17174
23.99
0.16530
1.17238
9
29.86
0.24789
1.17166
29.86
0.24789
1.17217
10
35.09
0.33047
1.17158
35.09
0.33047
1.17194
11
40.00
0.41318
1.17148
40.00
0.41318
1.1717
Figure 4.24: d vs sin2 plot for uncoated and coated specimens made of 20MnCr5 steel
130

fatigue specimens made of 20MnCr5 steel in various states, viz. green, case carburized
(uncoated) and case carburized WC/C coated, are reported in Table 4.11, Table 4.12 and
Table 4.13 respectively.
Table 4.11: Results of fatigue tests conducted on specimens made of 20MnCr5 steel in
green state
Stress (MPa)
Cycles to failure
Average
Std. Dev.
Green
Nf
Green
N2
N1
398
11783
17245
14514
3862
361
25498
38673
32086
9316
315
67672
118260
92966
35771
283
157356
273353
215355
82022
255
283679
622208
452944
239376
Table 4.12: Results of fatigue tests conducted on case-carburized and tempered
(uncoated) specimens made of 20MnCr5 steel
Stress (MPa)
Cycles to failure
Average
Std. Dev.
Nf Uncoated
Uncoated
N2
N1
1000
7407
10164
8786
1949
935
11661
17140
14401
3874
920
18766
26673
22720
5591
900
324294
782463
553379
323974
*
880
640758
2000000
----840
775436
2000000*
----*
772
2000000
------* No failure; Test suspended.
coated specimens made of 20MnCr5 steel
Stress (MPa)
Cycles to failure
Average
Std. Dev.
WC/C
Nf
WC/C
N2
N1
941
5604
8807
7206
2265
888
6963
10315
8639
2370
842
11773
19618
15696
5547
765
26115
43525
34820
12311
720
59872
101734
80803
29601
131
The data obtained from the fatigue experiments is plotted on a semi-log scale, i.e.,
Log (Nf) vs Stress (S), in Figure 4.25 along with the results of statistical analysis. For
specimens in green and WC/C coated states, the data points are observed to lie on a straight
line. The extent of finite life portion of the S-N curve for uncoated specimens was, however,
too short to be confidently ascertained as linear on semi-log scale. Nevertheless, for the
sake of comparison with other types of steels, a linear curve based on the model Y = A +
BX, described in ASTM E 739 standard [121], has been fitted to the experimental data. The
estimated values of parameters A and B in the linear model, obtained for 20MnCr5 steel
specimens in various states, are summarized in Table 4.14. The values of parameters A and
B for 20MnCr5 steel in the case carburized (uncoated) state have been computed by taking
into account the data given in only first three rows of Table 4.12. However, it was not
possible to work-out the 95% confidence bands for the range of data available for uncoated
specimens.
Table 4.14: Values of parameters in the linear model for fatigue life of 20MnCr5 steel
Material & State
Parameter A
Parameter B
20MnCr5 (Green)
8.210
-0.0102
20MnCr5 (Uncoated)
8.504
-0.0046
20MnCr5 (WC/C coated)
8.235
-0.0048
Unlike SAE8620 steel, the S-N graph for case carburized (uncoated) specimens made
of 20MnCr5 steel appears to be indicative of double-knee formation. The first knee forms at
around 20,000 cycles, after which, the curve remains asymptotic almost up to 106 cycles.
But failure of some specimens in the vicinity of 106 cycles implies a further dip in the
fatigue performance. The behaviour beyond this point couldnt be ascertained with
confidence, as testing for higher number of cycles, on the order of 107, was beyond the
scope of present study. Accordingly, this portion of the graph is represented by an
extrapolated, dashed line. For the case carburized WC/C coated specimens, the kneeformation was observed to occur at 105 cycles.
132
As expected, the fatigue performance of 20MnCr5 steel specimens in case carburized

and tempered (uncoated as well as WC/C coated) states is significantly superior to that
noticed for specimens tested in green state. But unlike SAE8620 steel, the coated specimens
made of 20MnCr5 steel exhibit inferior fatigue behavior than their uncoated counterparts in
both low as well as high cycle fatigue regimes.
The endurance limit of the specimens in various states was estimated through
staircase tests conducted for 2106 cycles. The results of these tests are summarized in
Table 4.15. Taking the number of less frequent event (i.e. survival or failure) as the basis
for statistical analysis of staircase tests, the data reduction is carried out according to the
Dixon-Mood method [refer section 2.3.2.4]. The intermediate calculations of the statistical
analysis along with the final results obtained for endurance limits, pertaining to specimens
in various states, are presented in Table 4.16.
Table 4.15: Results of staircase tests conducted on specimens made of 20MnCr5 steel
S. No.
Green
Case Carburized
WC/C Coated
Stress
(MPa)
Cycles
Stress
(MPa)
Cycles
Stress
(MPa)
Cycles
255
283679
920
26673
765
43525
230
564563
900
324294
720
59872
205
2000000*
880
2000000
675
2000000*
230
2000000*
900
782463
720
2000000*
255
622208
880
640758
765
26115
230
2000000*
860
2000000*
720
101734
---
---
880
2000000
675
2000000*
---
---
900
2000000*
---
---
---
---
920
18766
---
---
133
20MnCr5 steel specimens
i
Green
Case Carburized
WC/C Coated
S0 = 205 MPa
nDM,0 = 0
S0 = 860 MPa
nDM,0 = 1
S0 = 675 MPa
nDM,0 = 2
S1 = 230 MPa
nDM,1 = 1
S1 = 880 MPa
nDM,1 = 2
S1 = 720 MPa
nDM,1 = 1
S1 = 255 MPa
nDM,2 = 2
S1 = 900 MPa
nDM,2 = 1
S2 = 765 MPa
nDM,2 = 0
---
---
S3 = 920 MPa
nDM,3 = 0
---
---
ADM = (i )(n DM ,i ) = 5
ADM = (i )(n DM ,i ) = 4
ADM = (i )(nDM ,i ) =
BDM = (i 2 )(n DM ,i ) = 9
BDM = i 2 (n DM ,i ) = 6
( )
BDM = (i 2 )(n DM ,i ) =
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
( n )
DM ,i
( n )
DM ,i
( n )
DM ,i
= 0.22 (<0.3)
= 0.5 (0.3)
= 0.22 (<0.3)
Mean stress, (s) = 234 MPa
134
Figure 4.25: S-N graphs for 20MnCr5 steel specimens in green, case carburized
(uncoated) and case carburized - WC/C coated states.
135

Fractographic observations were recorded to study the topography of surfaces
fractured under fatigue and to identify the associated fracture mechanisms at different stress
levels. The topography of specimens failed in green state at different level of stress was
found to be similar to that observed for SAE8620 steel specimens in green state [Figure
4.11]. However, the examination of fractured surfaces of case carburized and tempered
specimens in uncoated and coated states revealed the fracture surfaces to exhibit
significantly different topography and associated crack propagation mechanisms, in
comparison to their counterparts made of SAE8620 steel.
Figure 4.26 shows macrographs of fractured surfaces of uncoated as well as WC/C
coated specimens made of 20MnCr5 steel. It can be observed, that on a macroscopic scale,
the fracture surfaces of the uncoated as well as WC/C coated specimens tested at various
stress levels generally exhibit three distinct regions: (i) Outermost region exhibiting grainy
appearance and having a number of ratchet marks, indicating crack propagation through
intergranular decohesion cracking and multiple crack origin sites respectively,
(ii) Intermediate, donut shaped region, exhibiting river-patterns, characteristic of secondarystage transgranular fracture, generally observed in low-carbon regions [240,292] and
(iii) Innermost region, featuring final ductile fracture due to overload. The region of final
(ductile) fracture is almost concentric with the cross-section of the specimen, which is a
typical feature of failure under pure rotating-bending fatigue [280]. The extent of ductile
fracture is observed to shrink with the increasing magnitude of maximum alternating stress
encountered during the fatigue test.
Similar to SAE8620 steel specimens, the crack in 20MnCr5 specimens (in uncoated
as well as WC/C coated states) was always found to initiate at the surface. The fracture
surfaces of the uncoated and coated specimens exhibited similar features, except that the
coated specimens tested under low cycle fatigue regime exhibited a mild tendency to form
shear lips [Figure 4.26 (e) and (f)], which was not observed in the uncoated specimens.
136
Figure 4.26: Optical fractographs showing macroscopic features on uncoated / coated

specimens made of 20MnCr5 steel, fatigued at various stress levels: (a):
Uncoated specimen tested at 1000 MPa, (b): Uncoated specimen tested at
935 MPa, (c): Uncoated specimen tested at 920 MPa (d): Uncoated
specimen tested at 900 MPa (e): WC/C coated specimen tested at 941 MPa
(f): WC/C coated specimen tested at 888 MPa (g): WC/C coated specimen
tested at 842 MPa and (h): WC/C coated specimen tested at 765 MPa.
137
Figure 4.27 shows the magnified views of regions marked with rectangles in Figure
4.26. These images are formed as mosaics covering entire swaths on the cross-section,
constructed by digitally stitching together individual images captured at successive
locations on the fractured surface. It is revealed from these images that intergranular
decohesion along prior austenitic grain boundaries is the dominant mode of crack initiation
and propagation.
The regions of stable, transgranular crack propagation are too small and difficult to
observe in most of the specimens. One such region is identified with the help of a dashed
line in Figure 4.28. Quantitative measurements taken on the region of stable crack
propagation [Refer section 2.6] provide depth, a = 79m and width, c = 415 m, for which,
fracture toughness, KIC = 16.62 MPa m , as estimated from Equation 2.21. This value is
quite low in comparison to the value of KIC (= 25.59 MPa m ) estimated for SAE8620
steel [Refer section 4.2.4.2], where transgranular fracture was observed to be the dominant
mode of crack advancement.
Owing to the limitations on depth-of-field in optical microscope, the topography of
surface exhibiting intergranular decohesion is better revealed by scanning electron
micrograph shown in Figure 4.29. This type of fracture is generally known to occur within
the high hardness region [294] and is also promoted by grain-boundary precipitates as well
as environmental assisted corrosion [295]. The intermediate region showing river-pattern
topography was found to exhibit cleavage-like transgranular fracture [Refer Figure 4.30].
A closer examination of the edges of fractured surfaces under optical microscope
revealed the presence of small loops of fracture zones in the vicinity of outermost layers,
that could be observed in both the specimens (uncoated and coated), as shown in Figure
4.31 (a-e). At some places, these cracks led to chipping-off of some material from the edge,
as shown in Figure 4.31 (d). This is in stark contrast to the observations made on SAE8620
steel specimens and is attributed to the formation of complex crack geometry under
intergranular crack propagation. As multiple crack fronts continue to progress while
138
changing their direction of propagation, some portions of material near the outermost
surface gets chipped-off.
Figure 4.27: Magnified views of regions marked with rectangles in the corresponding
fractographs given in Figure 4.26.
139
Figure 4.28: Micrograph indicating intergranular crack initiation and small region of
stable transgranular growth (marked by dashed-line), followed by
dominantly intergranular propagation in the specimen shown in Figure
4.26 (d).
Figure 4.29: Scanning electron micrograph showing intergranular cracking in the

specimen shown in Figure 4.26 (b).
140
Figure 4.30: Magnified view of specimen shown in Figure 4.26 (a), depicting three
different regions of crack propagation: (A): Region dominated by
intergranular fracture, (B): Region of cleavage-like transgranular fracture,
characterized by river pattern and (C): Region of ductile fracture.
As such, no significant difference was observed in the surface topography as well as

the associated crack propagation mechanisms among the uncoated and WC/C coated
specimens, particularly for specimens that had failed at comparable number of cycles. The
formation of shear lips near the outermost region of the coated specimens subjected to low
cycle fatigue [Figure 4.26 (e) and (f)] can be attributed to the increase in magnitude of
residual compressive stresses in the outermost layers of these specimens, which might have
hindered the occurrence of failure under mode I.
141
Figure 4.31: Magnified views showing the presence of curved-cracks in various

specimens whose fractographs are shown in Figure 4.26: (a): Non-radial
cracks in uncoated specimen shown in Figure 4.26 (d); (b): Looping cracks
in coated specimen shown in Figure 4.26 (e); (c): Completely looped cracks
in coated specimen shown in Figure 4.26 (f); (d): Composite micrograph
showing chipped-off material in specimen shown in Figure 4.26 (c), and (e):
Curved-cracks in coated specimen shown in Figure 4.26 (h).
142
4.3.5 Investigation of double-knee behaviour in 20MnCr5 steel

It is revealed from Figure 4.25 that while the WC/C coated specimens made of
20MnCr5 steel exhibit a well-defined endurance limit of 713MPa, the behaviour of
uncoated specimens is suggestive of double-knee formation, with estimated mean
endurance limit of 890MPa. These observations are commensurate with those reported by
Naito [162] and Parrish [276], where such behaviour is attributed to the presence of either
high temperature transformation products (HTTP) or internal oxidation [Refer section
2.4.1].
With a view to investigate the possible reason behind the double-knee behaviour, a
detailed study was performed on the failed uncoated specimens. The influence of
microstructure in the wake of crack was studied by preparing the fracture surface for
metallographic analysis. When viewed at high magnification, deep-penetrating grainboundary oxides could be observed in the outermost layers of polished and etched
specimens, as shown in Figure 4.32 (a) and (b). In order to ascertain the depth of oxide
formation, the specimens were polished lightly after etching with Nital. The resulting
micrographs, given in Figure 4.32 (c) and (d) reveal the presence of extensive network of
grain-boundary oxides. The oxides were found to be present up to a depth of about 1825m, which is substantially greater than the critical depth of 15m [276,296], thereby
affecting fatigue behaviour of the specimens.
In order to further confirm the presence of oxides within the surface layers, the
specimens were viewed under scanning electron microscope and EDAX elemental analysis
was performed over the region exhibiting oxidation. Figure 4.33 (a) shows the appearance
of oxides. The EDAX plot of the region marked with rectangle in Figure 4.33 (a) is shown
in Figure 4.33 (b). The results obtained through EDAX quantitative analysis [Refer Table
4.17] also confirm the presence of oxygen, thereby implying grain boundary oxidation in
the outermost surface of the specimens.
143
Figure 4.32: Micrographs indicating the presence of internal oxidation in 20MnCr5

specimens: (a) and (b): Specimens etched in 3% Nital for 5 seconds;
(c): Specimen polished after etching to reveal the depth of penetration of
oxides; (d): Micrograph taken in the vicinity of chipped-off portion
appearing towards left end in Figure 4.27 (b); the image is digitally
processed for enhancing depth of field by stitching together portions of
various photographs taken by shifting the focal plane of the microscope.
Figure 4.33: (a): Scanning electron micrograph showing the presence of oxide precipitates
within a grain near the surface of a case-carburized 20MnCr5 steel specimen;
(b): EDAX spectrum, confirming the presence of oxygen in the region
identified with a rectangle in (a).
144
Table 4.17: Results of quantitative EDAX elemental analysis

Element
Weight %
C
10.83
O
1.17
Cr
1.04
Mn
1.16
Fe
85.79
Atomic %
35.33
2.87
0.79
0.83
60.19
Dowling et al. [297] suggested that the presence of tensile residual stresses at the
surface of fatigue specimen may indicate the presence of high temperature transformation
products (HTTP), which is also responsible for double knee behaviour, as observed in this
study for uncoated 20MnCr5 steel specimens. However, in the present study, HTTP could
not be observed in the specimens. In order to rule out the absence of tensile residual stresses
in the outermost surface of the specimen, the residual stress measurements were made
across the depth of uncoated specimens.
In order to perform these X-ray diffraction residual stress measurements, a number of
specimens were cut from the ends of uncoated fatigue specimens. These specimens were
mounted face-down on a magnetic bed and their sectioned sides were ground flat, so as to
achieve equal thickness among all the specimens. The thickness of these specimens after
grinding was measured to high degree of accuracy using a digital micrometer.
Subsequently, the carburized faces of these specimens were polished using SiC papers of
various grades, so as to remove material upto a certain, pre-determined depth from each
specimen. Difference between the initial thickness (immediately after grinding the back
face) and final thickness after polishing with SiC paper served as a measure of depth below
surface, up to which the material had been removed by the SiC paper. After polishing, the
specimens were etched in 3% Nital and finally washed with ethanol. The final depth below
original surface was determined to an accuracy of 1 m before characterizing the specimens
for estimation of residual stresses. The results of residual stress analysis are given in Table
4.18 and shown graphically in Figure 4.34.
145
Table 4.18: Determination of lattice strain and associated residual stresses at various
depths in case carburized specimens made of 20MnCr5 steel.
S. No.
sin2
d-values () obtained at various depths below surface

1 m
24 m
85 m
219 m
320 m
0.41318
1.17082
1.17065
1.17085
1.17081
1.17116
0.33047
1.17102
1.17082
1.17100
1.17092
1.17123
0.24789
1.17125
1.17098
1.17116
1.17105
1.17131
0.16530
1.17150
1.17118
1.17129
1.17120
1.17139
0.08267
1.17174
1.17136
1.17148
1.17137
1.17148
1.17218
1.17158
1.17175
1.17169
1.17163
0.08267
1.17219
1.17149
1.17178
1.17177
1.17162
0.16530
1.17209
1.17137
1.17169
1.17174
1.17157
0.24789
1.17195
1.17124
1.17159
1.17166
1.17152
10
0.33047
1.17180
1.17108
1.17148
1.17158
1.17146
11
0.41318
1.17165
1.17096
1.17135
1.17148
1.17138
-319.6
-266.4
-221.9
-184.8
-118.4
Residual Stress
(MPa)
Figure 4.34: d vs sin2 plots for case-carburized, uncoated specimens made of 20MnCr5
steel at various depths below surface.
146
The variation of residual stress across depth of the case carburized and tempered
(uncoated) specimens made of 20MnCr5 steel is plotted on semi-logarithmic scale in Figure
4.35. As expected, the magnitude of residual compressive stress decreases with increasing
depth. The absence of tensile residual stresses at the surface of the specimens rules out the
chances of formation of HTTP in the present study.
These observations thus imply that the observed anomalous double-knee behaviour of
case carburized and tempered, uncoated specimens made of 20MnCr5 steel is due the
formation of internal oxides during carburization. These oxides provide sites of
discontinuities, or material inhomogenieties, which increasingly govern the fatigue crack
initiation and propagation [298]. Similar observations concerning shift in the knee of S-N
curve for specimens having internal oxidation are also reported by Krauss [239].
Figure 4.35: Residual stress profile in case carburized 20MnCr5 steel showing variation of
residual stress with depth below surface.
With a view to further investigate the effect of internal oxidation on the observed
fatigue behavior of case-carburized and tempered (uncoated) specimens made of 20MnCr5
steel, the fatigue tests were conducted on a specimen after removing the surface layers
affected by internal oxidation through grinding with successive grades of SiC paper.
147
Though the resulting specimen had a neck diameter of 5.906mm, which is outside the
prescribed tolerance of 60.05 mm, the specimen exhibited significantly improved fatigue
performance. When subjected to a stress of 1014 MPa, it survived 5,29,574 cycles, which is
much higher in comparison to the average life of 8,786 cycles, exhibited by regular
uncoated specimens tested at 1000 MPa [Refer Table 4.12].
Another interesting observation was that the crack now had a sub-surface origin,
thereby forming a fish-eye, which appears as a bright spot towards the top of Figure 4.36
(a). Fish eye features appear bright because the freshly fractured surface does not come in
contact with air till the crack-front reaches the outermost surface. When viewed at higher
magnification, radial marks emanating from the crack initiation site can be observed [Refer
Figure 4.36 (b)]. Examination at still higher magnification reveals the optically dark area
(ODA) at the site of crack origin, as shown in Figure 4.37. ODAs are known to appear only
for the specimens fractured after large number of cycles and are absent in case of specimens
failed at small number of cycles [294]. This area appears dark because of different fracture
features, which give rise to a very rough morphology.
Figure 4.36: Optical fractographs of uncoated specimen, polished to remove surface layers
affected by internal oxidation, fatigue tested at 1014 MPa. (a): Fracture
macrograph showing the formation of fish-eye; (b): Magnified view of the
fish-eye appearing in (a).
148
Figure 4.37: Optical fractograph of ODA at the crack initiation site within the fish-eye.
All these observations pertaining to sub-surface crack origin imply that a defect-free
surface, obtained by complete removal of oxidation-affected layers, leads to a remarkable
improvement in fatigue performance of the case carburized and tempered specimens made
of 20MnCr5 steel.
4.3.6 Effect of WC/C coating on fatigue behaviour of 20MnCr5 steel

It is revealed from the investigation that the fatigue behaviour of case carburized
WC/C coated specimens is quite inferior to that observed for case carburized (uncoated)
specimens made of 20MnCr5 steel. Though the magnitude of residual compressive stress
among the coated specimens is considerably higher than the residual stress in the uncoated
specimens, the dominance of crack propagation through intergranular decohesion cracking,
coupled with grain-boundary oxidation, seems to be responsible for inferior fatigue
behaviour of the coated specimens. It appears that since the hardness of steel gets
149
compromised to some extent during the coating process [Refer Figure 4.2], the cracks find
an easy passage along weak boundaries of prior austenitic grains, especially in the vicinity
of oxidized layer.
The observed fatigue limit of 713 MPa for coated specimens in the present study is
around 5% higher than the value of 680 MPa reported by Brugger [161] for 20MnCr5 steel
that was free from HTTP formation; except that this limit is found to occur at 105 cycles in
our case. For the uncoated specimens, the first knee is found to form at a considerably
higher stress level of around 890 MPa. However, after remaining horizontal for some
extent, the graph again exhibits decline in the vicinity of 106 cycles. Though the fatigue
behaviour beyond this point could not be ascertained with confidence, as tests beyond
2106 cycles were not possible due to experimental limitations, still, the double-knee
formation is clearly implied by the results obtained.
150
4.4
Testing and analysis of EN353 steel specimens

The influence of case carburization, followed by subsequent coating with WC/C layer
on the fatigue performance of EN353 steel is presented in the following sub-sections.

Following the procedure similar to that described in sections 4.2 and 4.3 for other steels
investigated in this study, the fatigue performance of EN353 steel has also been evaluated
under various states, viz. green, case carburized (uncoated) and case carburized WC/C
coated and the results obtained have been correlated with the microstructure, hardness and
residual stress to identify and explain the prevailing crack propagation mechanisms.

A mosaic of metallographs, showing variation of microstructure across depth in a
case-carburized specimen made of EN353 steel, is shown in Figure 4.38. In order to prepare
the metallograph, the specimen was first etched in 3% Nital for 5 seconds, followed by
immersion in potassium metabisulfite solution for 20 seconds. The difference in thickness
of the film deposited on various microstructural constituents led to formation of different
colours through interference of light. The plate martensite is imparted yellow to green tint,
while the matrix phase appears deep red in this mosaic of slightly over-etched
metallographic specimen.
Some internal oxidation near the outermost edge of the specimen can be observed in
these metallographs. The oxides appear as a network of dark bands along prior austenitic
grain boundaries. The depth of penetration (~ 15 20 m) of these oxides is on the order of
austenitic grain size. Figure 4.39 (a) shows high-contrast metallograph, vividly revealing
the presence of internal oxidation along prior austenitic grain boundaries. The region
marked with rectangle in Figure 4.39 (a) is shown magnified in Figure 4.39 (b), where
inter-grain as well as intra-grain oxides can be observed separately. The formation of these
oxides generally results due to oxidation of chromium and manganese [157,158] present in
EN353 steel [Refer Table 3.1].
151
Figure 4.38: Mosaic of metallographs, showing variation in microstructure with depth in

case carburized and tempered specimens made of EN353 steel, etched in 3%
Nital for 5 seconds, followed by immersion in potassium metabisulfite
solution for 20 seconds.
152
A careful examination of Figure 4.38 reveals that martensitic structures occupy

different volume fractions at different depths below the outermost surface. The outermost
region, extending to a depth of around 50 70 m, contains high volume fraction of plate
martensite, exhibiting fine acicular morphology (etched yellow-green). This region is
followed by a sudden increase in the content of retained austenite (etched deep-red), which
extends over a band of around 100 m width. At a depth of around 160 m, the volume
fraction of plate martensite (etched green) again begins to increase and becomes highest at
depths ranging from 450 to 750 m. Beyond a depth of 750 m (which corresponds to the
effective depth of case), a mixture of plate and lath martensite can be observed, which
gradually gives way to pure lath martensite, extending right into the core of the specimen,
which is vividly revealed through nital-etch [Refer Figure 4.40 (a) and (b)].
At higher magnifications, segregates of carbide particles could be seen dispersed in
the case, as revealed in the colour metallograph shown in Figure 4.40 (c). In this
metallograph, the red matrix phase corresponds to retained austenite, while the carbide
precipitates can be seen as bright white spots, some of which have been identified with
arrow-marks.
The martensitic start (Ms) temperature, as estimated from Equation 4.1, corresponding
to the chemical composition of EN353 [Refer Table 3.1], works out to be 367C, which
corresponds to 94% volume fraction of martensitic transformation [Refer Equation 4.2] at a
quenching temperature of 120C [Refer section 3.2]. This is in agreement with the high
volume fraction of lath martensite observed in the core, where the chemical composition
essentially remains unaltered even after case carburization.
The value of Ms temperature obtained for this steel is quite comparable to that
calculated for 20MnCr5 steel in section 4.3.1, but is significantly lower than that obtained
for SAE8620 steel [Refer section 4.2.1]. Thus, during quenching of EN353 steel specimens,
the martensitic transformation in the core would have commenced at a relatively later stage
in comparison to that in the specimens made of SAE8620 steel. This altered timing would
153
have adversely affected the possibility of strain-induced martensitic transformation near the
surface of specimens made of EN353 steel. Further, the underlying band of retained
austenite (observed at a depth ranging from 60m to 160m) would have served to provide
a soft base incapable of transmitting stresses to the layer above. These arguments explain
the relatively lower volume fraction of martensite in the outermost layers of EN353 steel
specimens, as compared to those made of SAE8620 steel.
Figure 4.39: (a): Metallograph indicating the presence of internal oxidation in case
carburized and tempered specimens made of EN353. The specimen was
lightly polished after etching with nital; and (b): Magnified view of region
marked with rectangle in (a).
154
Figure 4.40: Metallographs of the cross section of case carburized and tempered specimen
made of EN353 steel. (a) and (b): Lath martensite in the core; and (c):
Composite metallograph showing martensite, retained austenite and chunks
of carbide precipitates (marked with arrows) within the carburized and
tempered case. Specimens in (a) and (b) etched with 3% nital for 5s,
specimen in (c) were etched with 3% nital for 4s, followed by immersion in
potassium metabisulfite for 12s.

The X-ray diffraction residual stress estimation was performed according to the
procedures described earlier [Refer sections 2.5.2, 3.7.2 and 4.2.2]. The results obtained are
summarized in Table 4.19 and depicted graphically in Figure 4.41. Similar to the
observations made on SAE8620 and 20MnCr5 steels, slight -splitting has also been
observed in the diffraction patterns recorded for EN353 steel.
155
Table 4.19: Determination of lattice strain in EN353 steel specimens through X-ray
diffraction
S. No.
Uncoated Specimen
Coated Specimen
Tilt,
sin2
d-spacing
Tilt,
sin2
d-spacing
-40.00
0.41318
1.17026
-40.00
0.41318
1.17043
-35.09
0.33047
1.17037
-35.09
0.33047
1.17069
-29.86
0.24789
1.17053
-29.86
0.24789
1.17091
-23.99
0.16530
1.17069
-23.99
0.16530
1.17118
-16.71
0.08267
1.17087
-16.71
0.08267
1.17144
1.17113
1.17173
16.71
0.08267
1.17111
16.71
0.08267
1.17159
23.99
0.16530
1.17102
23.99
0.16530
1.17137
29.86
0.24789
1.17092
29.86
0.24789
1.17115
10
35.09
0.33047
1.17079
35.09
0.33047
1.17090
11
40.00
0.41318
1.17068
40.00
0.41318
1.17069
Figure 4.41: d vs sin2 plot for uncoated and coated specimens made of EN353 steel
156
The residual stress is estimated to be -225.0 3.6 MPa for uncoated and -399.9 3.8
MPa for coated specimens. Thus, for this steel too, the case carburized surface contains
residual compressive stresses, which undergo considerable enhancement upon coating with
WC/C layer through physical vapour deposition. The surface residual stresses in the
uncoated and coated specimens made of EN353 steel are significantly lower than their
counterparts made of SAE8620 steel. This is attributed to relatively higher amount of
retained austenite in the vicinity of outermost surface of EN353 steel specimens [Figure
4.38] than that observed in specimens made of SAE8620 steel [Figure 4.4 (a), Figure 4.6].

fatigue specimens made of EN353 steel in various states, viz. green, case carburized
(uncoated) and case carburized WC/C coated are reported in Table 4.20, Table 4.21 and
Table 4.20: Results of fatigue tests conducted on specimens made of EN353 steel
specimens in green state
Stress (MPa)
Cycles to failure
Average
Std. Dev.
Green
Nf
Green
N2
N1
555
6730
8775
7753
1446
537
9113
21938
15526
9069
480
15675
27846
21761
8606
426
33456
66488
49972
23357
357
105672
194063
149868
62502
317
171248
353672
262460
128993
270
316893
667382
492138
247833
Table 4.21: Results of fatigue tests conducted on case-carburized and tempered
(uncoated) specimens made of EN353 steel
Stress (MPa)
Cycles to failure
Average
Std. Dev.
Uncoated
Nf
Uncoated
N2
N1
1000
8467
15264
11866
4806
925
20752
66754
43753
32528
860
69815
98762
84289
20469
765
176743
327483
252113
106589
686
289467
615463
452465
230514
663
556733
925362
741048
260660
157
coated specimens made of EN353 steel
Stress (MPa)
Cycles to failure
Average
Std. Dev.
WC/C
Nf
WC/C
N2
N1
923
4563
8727
6645
2944
843
16220
22881
19551
4710
726
122589
239346
180968
82560
680
298356
445892
372124
104324
655
566229
896546
731388
233569
Similar to the procedure followed earlier for other steels [Refer sections 4.2.3 and
4.3.3], the change in fatigue life as a function of maximum alternating stress, has been
expressed in terms of linear model Y = A + BX, described in ASTM E 739 standard [121].
This linear model has been fitted separately to the fatigue data recorded for specimens in
green, case-carburized (uncoated) and case carburized WC/C coated states. The estimated
values of parameters A and B in the linear model are summarized in Table 4.23.
Table 4.23: Values of parameters in the linear model for fatigue life of EN353 steel
Material & State
Parameter A
Parameter B
EN353 (Green)
7.304
-0.0061
EN353 (Uncoated)
9.171
-0.0050
EN353 (WC/C coated)
10.787
-0.0076
The endurance limits of the specimens in various states were ascertained by means of
staircase tests, conducted for 2106 cycles, the results of which are summarized in Table
4.24. The data obtained from the staircase tests is processed according to the Dixon-Mood
method [Refer section 2.3.2.4], by taking into account the number of less frequent event
i.e., survival or failure. The outline of analysis, along with the results obtained, is presented
in Table 4.25.
158
Table 4.24: Results of staircase tests conducted on specimens made of EN353 steel
S. No.
Green
Case Carburized
WC/C Coated
Stress
Cycles
Stress
Cycles
Stress
Cycles
(MPa)
(MPa)
(MPa)
1
270
667382
686
615463
680
445892
*
2
245
2000000
663
925362
655
2000000*
3
270
316893
640
784560
680
298356
*
4
245
754038
617
2000000
655
566229
*
5
220
1209843
640
2000000
630
789473
*
6
195
2000000
663
556733
605
2000000*
7
220
2000000*
640
2000000*
630
2000000*
8
----663
2000000*
655
896546
9
----686
289467
630
2000000*
EN353 steel specimens
i
Green
Case Carburized
WC/C Coated
0 S0 = 195 MPa nDM,0 = 1 S0 = 617 MPa nDM,0 = 1 S0 = 605 MPa nDM,0 = 1
S1 = 220 MPa
nDM,1 = 1
S1 = 640 MPa
nDM,1 = 2
S1 = 630 MPa
nDM,1 = 2
S2 = 245 MPa
nDM,2 = 1
S1 = 663 MPa
nDM,2 = 1
S2 = 655 MPa
nDM,2 = 1
S3 = 270 MPa
nDM,3 = 0
S3 = 686 MPa
nDM,3 = 0
S3 = 680 MPa
nDM,3 = 0
ADM = (i )(n DM ,i ) = 1
ADM = (i )(n DM ,i ) = 4
ADM = (i )(nDM ,i ) = 4
BDM = (i 2 )(n DM ,i ) = 1
BDM = i 2 (n DM ,i ) = 6
( )
BDM = (i 2 )(n DM ,i ) = 6
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
( n )
DM ,i
( n )
DM ,i
( n )
DM ,i
= 0.22 (<0.3)
= 0.5 (0.3)
= 0.5 (0.3)
159
The experimental fatigue data, along with the results estimated from the fitted linear
models and 95% confidence bands, is plotted in Figure 4.42. It is observed from the figure
that Log (Nf) vs Stress (S) curves for EN353 steel specimens in various states follow
straight-line paths. The case carburized specimens in uncoated as well as WC/C coated
condition exhibit much better performance than the specimens in green state. The
endurance limits for case carburized (uncoated) as well as case carburized WC/C coated
steels are nearly 180% higher than the endurance limit for steel in green state. The fatigue
performance of case carburized and tempered (uncoated) specimens deteriorates upon
application of WC/C coating. The loss observed in fatigue strength diminishes with
increasing number of cycles to failure (Nf).
The contribution of case carburization as well as application of WC/C coating on
carburized specimens, towards percentage change in fatigue strength of the specimens
subjected to a given number of stress cycles, as computed from Equation 4.3, is given in
Table 4.26 and depicted graphically in Figure 4.43.
These observations are somewhat similar to those for SAE8620 steel [Refer section
4.2.3] except that the coated specimens exhibit poor fatigue response than the uncoated
ones even at stress cycles beyond 105. However, the endurance strengths of uncoated and
WC/C coated specimens are observed to be quite identical.
strength of EN353 steel.
SfGreen
SfUncoated
SfWC/C
A= 7.3040
A=9.1708
A=10.7869
B=-0.00608
B=-0.00503
1000
708
10000
No. of
Cycles to
failure, Nf
B=-0.00763
Case
Carburized
vs Green
WC/C
Coated
vs Green
1228
1021
73
44
543
1029
889
89
64
100000
379
830
758
119
100
1000000
214
631
627
194
193
160
Figure 4.42: S-N graphs for EN353 steel specimens in green, case carburized (uncoated)
161
Figure 4.43: Percentage change in fatigue strength of uncoated and coated EN353 steel
specimens with reference to specimens in green state.
162

Fractomacrographs recorded on case carburized, uncoated as well as WC/C coated
specimens made of EN353 steel, failed at different number of cycles, are shown in Figure
4.44. As observed for 20MnCr5 steel, on a macroscopic scale, the fracture surface generally
comprises of three regions: (i) Outermost region exhibiting grainy appearance and having a
number of ratchet marks, indicating multiple sites of crack origin, (ii) Intermediate, donut
shaped region, exhibiting river-patterns, implying secondary stage transgranular mode of
crack propagation [240,292] and (iii) Innermost region, featuring final ductile fracture.
The region of final fracture is more or less concentric with the cross-section of specimen
and its extent decreases with the increase in magnitude of alternating stress applied during
fatigue test. As observed earlier for other steels, the fatigue crack in this steel too was
always found to initiate at the outer surface.
Fractographic studies conducted at higher magnification revealed that the cracks
originated at the outer surface through intergranular decohesion along prior austenitic grain
boundaries [Refer arrow mark in Figure 4.45], followed by stable transgranular
propagation for some time. As discussed earlier in section 4.2.4.2, this type of crack
initiation and propagation are commonly observed fracture modes in the carbon rich case of
the carburized steels.
In the micrograph shown in Figure 4.45, the width (c) and depth (a) of stable
transgranular crack propagation region (marked with dashed line) are found to be 363m
and 165m respectively. Using these values in Equation 2.21, the fracture toughness, KIC of
the case carburized EN353 steel works out to be 17.38 MPa m . Therefore, the fracture
toughness of EN353 is slightly better than 20MnCr5 steel, but is considerably lower than
that observed for SAE8620 steel in similar state.
163
Figure 4.44: Optical fractographs showing macroscopic features in uncoated and WC/C
coated specimens made of EN353 steel, fatigued at various loads: (a):
Coated specimen tested at 843 MPa, (b): Uncoated specimen tested at 925
MPa, (c): Coated specimen tested at 726 MPa (d): Uncoated specimen
tested at 860 MPa (e): Coated specimen tested at 655 MPa, and
(f): Uncoated specimen tested at 765 MPa.
164
It is observed that the region of unstable crack propagation contains a substantial

percentage of intergranular decohesion along prior austenitic grain boundaries [Refer
Figure 4.45 and Figure 4.46]. Upon reaching the region with relatively low carbon content,
the crack is observed to advance almost entirely by cleavage-like transgranular fracture,
where the surface morphology is characterized by river patterns [Figure 4.47]. The
comparison of Figure 4.46 and Figure 4.47 with Figure 4.29 reveals that the fraction of
intergranular mode of crack propagation observed in EN353 steel is somewhat less than
that observed in 20MnCr5 steel specimens.
Figure 4.45: Micrograph of specimen shown in Figure 4.44 (d), indicating crack initiation
by intergranular cracking (marked with arrow).
In line with the observations made on other steels investigated in the present work, no
significant differences in crack initiation and propagation mechanisms could be observed
among the case carburized uncoated as well as WC/C coated specimens made of EN353
steel, which had failed at comparable number of fatigue cycles. Only the stress level for
their occurrence gets modified in the presence of WC/C coating.
165
Figure 4.46: Scanning electron micrograph of specimen shown in Figure 4.44 (a),
indicating a mix of transgranular and intergranular cracking.
Figure 4.47: Optical fractograph of specimen shown in Figure 4.44 (c), indicating various
regions of crack propagation. The fractograph is constructed as a mosaic by
stitching together four individual fractographs.
166
4.4.5 Effect of WC/C coating on fatigue behaviour of EN353 steel

It is evident from the results obtained [Table 4.25 and Figure 4.42] that the fatigue
behaviours of case carburized (uncoated) and case carburized WC/C coated specimens
made of EN353 steel are quite identical in the high cycle regime. However, the coated
specimens exhibit slightly inferior fatigue performance in comparison to their uncoated
counterparts. Unlike 20MnCr5 steel specimens, the endurance limits are observed to be
quite comparable for the uncoated and coated specimens made of EN353 steel.
The coating process had led to a slight compromise in hardness [Figure 4.2], but
significant gain in terms of residual compressive stress [Refer section 4.4.2]. However, as
the cracks propagate through intergranular decohesion, the residual stresses induced in the
thin coating seem to have little influence on them. The presence of internal oxidation along
prior austenitic grain boundaries offers innumerable crack initiation sites on the surface,
which are not effectively plugged by the thin, hard and brittle coating. Rather, any crack
developed in the coating as a result of poor cohesion among the underlying grains would
tend to aggravate the problem by providing site for viable crack growth on the surface,
whose hardness has already been compromised somewhat by the coating process.
For the sake of comparison with other steels, a more aggravated situation of this kind
was encountered for 20MnCr5 steel, where the extent of grain boundary oxidation was
more severe, while the crack propagation was almost entirely intergranular in nature. Both
these factors seemed to severely impair the fatigue performance of 20MnCr5 steel
specimens in coated conditions. On the other hand, in specimens made of SAE8620 steel,
the benefits of enhanced residual compressive stresses upon coating with WC/C layer could
be harnessed because of largely transgranular crack propagation and insignificant amount
of internal oxidation at the surface. In the backdrop of these observations, the effect of
WC/C coating on fatigue behaviour of EN353 steel appears to lie in-between the effects
observed in SAE8620 and 20MnCr5 steels.
167
4.5
Testing and analysis of SCM420 steel specimens

The influence of case carburization and physical vapour deposition of WC/C layer on
the fatigue performance of SCM420 steel is presented in the following sub-sections. The
observed fatigue behaviour of specimens in various states is correlated with the influence of
material processing on the resulting microstructure, hardness, residual stresses and crack
propagation mechanisms.

The variation in microstructure with depth in case carburized specimens made of
SCM420 steel is captured in the mosaic of metallographs shown in Figure 4.48. The
specimen was first etched with 3% nital for revealing the microstructure, followed by
immersion in potassium metabisulfite solution [Refer Table 3.2] for deposition of anodic
film to enhance the colour contrast among various constituent phases.
In the micrograph, one can observe fine plate martensitic structure in the form of
bluish, needle-shaped morphology, along with retained austenite, which provides light
purple background. Since the martensitic start (Ms) temperature increases with depth due to
negative carbon gradient, the highest volume fraction of martensite is generally found to
occur at some depth below the surface during quenching. In Figure 4.48, the highest
volume fraction of plate martensite can be observed, first upto a depth of about 60m,
followed by another band at a depth of around 160m to 230m. The formation of outer
martensitic layer is attributed to relatively higher quenching rate at the surface, along with
facilitation of martensitic transformation due to introduction of tensile stresses in the
surface layer during the quenching of steel, when the sub-surface layer undergoes
martensitic transformation. Fine carbide precipitates within the carbon-rich case were
revealed upon colour-etching with Klemms I solution, as shown in Figure 4.49.
168
Figure 4.48: Mosaic of metallographs, showing variation of microstructure with depth in

case carburized and tempered specimens made of SCM420 steel, etched in
3% Nital for 4 seconds, followed by immersion in potassium metabisulfite
solution for 9 seconds.
169
The observed band of retained austenite at depths ranging from 60m to 160m is
similar to that reported for specimens made of EN353 steel [Refer Figure 4.38]. However,
the fraction of retained austenite observed in SCM420 steel specimens is considerably less
than that observed in EN353 specimens, which is also reflected in the corresponding
hardness measurements made on these steels [Refer Figure 4.2].
The martensitic start temperature (Ms) corresponding to chemical composition in the
core of SCM420 steel [Refer Table 3.1] is estimated to be 380C [Refer Equation 4.1],
which corresponds to around 95% volume fraction of martensite for quenching carried at
120C [272]. The metallographs recorded near the core of SCM420 steel specimens also
confirm such a high volume fraction of martensite. Further, a higher value of Ms
temperature for core is helpful in bringing-about a more complete martensitic
transformation near the surface due to introduction of tensile stresses at an early stage of
quenching. This is in line with the observations reported in the preceding paragraphs.
Figure 4.49: Metallograph of SCM420 specimen, colour etched with Klemms I reagent
for 2 minutes to reveal the presence of carbides (marked with arrows). Green
tint employed for contrast enhancement.
170

The results of X-ray diffraction residual stress measurements, recorded by following
the procedure outlined in sections 2.5.2 and 3.7.2, are summarized in Table 4.27 and shown
graphically in Figure 4.50. The residual stress is found to be -261.73.6 MPa for the
uncoated and -531.83.8 MPa for the coated specimens. Among various steels analyzed in
the present study, the magnitude of residual stress observed for case carburized specimens
made of SCM420 steel is the second highest after SAE8620 steel. This is attributed to the
presence of high volume fraction of martensite in the outermost layers of these specimens.
The deposition of WC/C coating on the surface of case carburized specimens led to an
enhancement of 103% in the magnitude of residual compressive stresses. This gain in
compressive stresses is third highest after SAE8620 and 20MnCr5 steels, among the four
steels investigated in the present work.
Table 4.27: Determination of lattice strain in SCM420 steel specimens through X-ray
diffraction
S. No.
Uncoated Specimen
Coated Specimen
Tilt,
sin2
d-spacing
Tilt,
sin2
d-spacing
-40.00
0.4132
1.17099
-40.00
0.4132
1.17110
-35.09
0.3305
1.17113
-35.09
0.3305
1.17140
-29.86
0.2479
1.17127
-29.86
0.2479
1.17172
-22.99
0.1525
1.17139
-22.99
0.1525
1.17203
-16.71
0.0827
1.17149
-16.71
0.0827
1.17231
0.00
1.17149
0.00
1.17254
16.71
0.0827
1.17118
16.71
0.0827
1.17213
22.99
0.1525
1.17096
22.99
0.1525
1.17179
29.86
0.2479
1.17076
29.86
0.2479
1.17143
10
35.09
0.3305
1.17058
35.09
0.3305
1.17111
11
40.00
0.4132
1.17043
40.00
0.4132
1.17078
171
Figure 4.50: d vs sin2 plot for uncoated and coated specimens made of SCM420 steel

fatigue test specimens made of SCM420 steel in various states, viz. green, case carburized
(uncoated) and case carburized WC/C coated, are reported in Table 4.28, Table 4.29 and
Table 4.28: Results of fatigue tests conducted on specimens made of SCM420 steel in
green state
Stress (MPa)
Cycles to failure
Average
Nf
Std. Dev.
Green
Green
N1
N2
445
8267
15672
11970
5236
410
23560
39863
31712
11528
350
82389
176732
129561
66711
310
147364
336448
241906
133703
280
269671
694063
481867
300090
172
Table 4.29: Results of fatigue tests conducted on case-carburized (uncoated) specimens
made of SCM420 steel
Stress (MPa)
Cycles to failure
Average
N1
N2
Uncoated
980
9278
17209
13244
5608
945
14790
27367
21079
8893
900
36273
98375
67324
43913
850
102703
280382
191543
125638
815
238450
616892
427671
267599
780
567392
1002832
785112
307903
Nf
Std. Dev.
Uncoated
coated specimens made of SCM420 steel
Stress (MPa)
Cycles to failure
Average
Nf
Std. Dev.
WC/C
WC/C
N1
N2
940
6748
14673
10711
5604
915
17783
32453
25118
10373
880
58749
118637
88693
42347
850
98834
252016
175425
108316
820
240679
536284
388482
209024
The change in fatigue life as a function of maximum alternating stress is represented

in terms of linear model Y = A + BX [121], obtained by fitting individually the set of
fatigue data recorded for SCM420 specimens in various states. The estimated values of
parameters A and B appearing in the model [Refer section 4.2.3] are summarized in Table
4.31. The fatigue data obtained through experiments and fitted linear models are
represented in Log (Nf) vs Stress (S) graph, along with hyperbolic bands of 95% confidence
limits in Figure 4.51. The S-N curves for the steel specimens in various states are found to
exhibit linear trend on semi-log scale.
Table 4.31: Values of parameters in the linear model for fatigue life of SCM420 steel
Material & State
Parameter A
Parameter B
SCM420 (Green)
8.2849
-0.0094
SCM420 (Uncoated)
13.0251
-0.0092
SCM420 (WC/C coated)
16.1665
-0.0129
173
The endurance limit of the specimens in various states was ascertained by conducting
staircase tests for 2106 cycles, as described earlier for other steel grades employed in this
study. The results of these tests are summarized in Table 4.32. The results obtained by
performing statistical analysis of staircase test data through Dixon-Mood method [Refer
section 2.3.2.4] are given in Table 4.33.
The endurance limit for the SCM420 steel in case carburized (uncoated as well as
WC/C coated) conditions is found to be nearly three times higher than the endurance limit
observed for SCM420 steel specimens in green state. The comparison of case carburized
(uncoated) and case carburized - WC/C coated specimens reveals around 2.6%
improvement in endurance limit upon coating.
The S-N curves shown in Figure 4.51 indicate that the coated specimens exhibit a
relatively lesser decline in the fatigue strength with increasing number of cycles (Nf) than
their uncoated counterparts. The S-N curves of the uncoated and coated specimens crossover each other at a stress cycle of around 2.4105. Therefore, the deposition of coating on
case carburized specimens improves their fatigue performance in the high cycle fatigue
regime beyond Nf = 2.4105, but for Nf < 2.4105, the uncoated fatigue specimens exhibit
better fatigue performance than their coated counterparts. The observed behaviour is quite
similar to that reported for SAE8620 steel in section 4.2.3.
Table 4.32: Results of staircase tests conducted on specimens made of SCM420 steel
S. No.
Green
Case Carburized
WC/C Coated
Stress
Cycles
Stress
Cycles
Stress
Cycles
(MPa)
(MPa)
(MPa)
1
280
269671
815
238450
820
536284
2
260
2000000*
780
1002832
795
2000000*
3
280
694063
745
2000000*
820
240679
4
260
564378
780
2000000*
795
974782
5
240
780672
815
616892
770
784560
6
220
2000000*
780
567392
745
2000000*
7
240
2000000*
745
836471
770
2000000*
*
*
8
260
2000000
710
2000000
----9
----745
2000000*
----* No failure; Test suspended.
174
SCM420 steel specimens
i
Green
Case Carburized
WC/C Coated
S0 = 220 MPa
nDM,0 = 0
S0 = 710 MPa
nDM,0 = 1
S0 = 745 MPa
nDM,0 = 1
S1 = 240 MPa
nDM,1 = 1
S1 = 745 MPa
nDM,1 = 2
S1 = 770 MPa
nDM,1 = 1
S2 = 260 MPa
nDM,2 = 1
S1 = 780 MPa
nDM,2 = 1
S2 = 795 MPa
nDM,2 = 1
S3 = 280 MPa
nDM,3 = 2
S3 = 815 MPa
nDM,3 = 0
S3 = 820 MPa
nDM,3 = 0
ADM = (i )(n DM ,i ) = 9
ADM = (i )(n DM ,i ) = 4
ADM = (i )(nDM ,i ) = 3
BDM = (i 2 )(n DM ,i ) = 23
BDM = i 2 (n DM ,i ) = 6
( )
BDM = (i 2 )(n DM ,i ) = 5
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
2
BDM nDM ,i ADM
( n )
DM ,i
( n )
DM ,i
( n )
DM ,i
= 0.69 (0.3)
= 0.5 (0.3)
= 0.67 (0.3)
175
Figure 4.51: S-N graphs for SCM420 steel specimens in green, case carburized (uncoated)
The contribution of case carburization as well as WC/C coating towards percentage

change in the fatigue strength of specimens subjected to a given number of stress cycles
(Nf), as estimated from Equation 4.3, is given in Table 4.34 and depicted graphically in
Figure 4.52.
176
strength of SCM420 steel.
SfGreen
SfUncoated
SfWC/C
A=8.2849
A=13.0251
A=16.1665
B=-0.00937
B=-0.00916
B=-0.01289
1000
564
1094
10000
457
100000
1000000
No. of
Cycles to
failure, Nf
1021
Case
Carburized
vs Green
94
WC/C
Coated
vs Green
81
985
944
115
106
351
876
866
150
147
244
767
789
214
223
Figure 4.52: Percentage change in the fatigue strength of uncoated and coated SCM420
steel specimens with respect to specimens in green state.
It is observed from Figure 4.52 that both the uncoated and coated specimens show
substantial improvement in fatigue performance as compared to specimens in green state.
This enhancement in fatigue performance is attributed to the presence of high volume
177
fraction of martensite in the outermost layers and the associated high magnitude of residual
compressive stresses induced in the uncoated as well as WC/C coated specimens.
The percentage gain in fatigue strength of both types of specimens increases with
increasing number of cycles, with a relatively steeper increase for the coated specimens. At
stresses higher than 835 MPa, which corresponds to a fatigue life of 2.4105 cycles, the
uncoated specimens exhibit better performance than the coated specimens. However, the
coated specimens exhibit better fatigue performance at stress levels below 835 MPa (Nf >
2.4105 cycles). The observed gain in fatigue strength increases with the increase in
number of cycles until the endurance limit of the coated specimens is reached.

Fractographic observations were made on case carburized, uncoated as well as WC/C
coated specimens made of SCM420 steel, which had failed at different stress cycles. The
low magnification fractomacrographs are shown in Figure 4.53 (a) to (d). High
magnification micrographs of the crack-initiation regions (identified with arrow-marks),
recorded in the same orientation, are placed adjascent to the respective fractomacrographs.
It can be observed from the fractographs that on a macroscopic scale, the fracture
surface is comprised of an outermost, donut-shaped region, dominated by transgranular
mode of crack advancement, which is surrounding the inner region of rapid fracture. The
fracture surface topography corresponding to major portion of fatigue life of the specimens
is thus dominated by transgranular mode of crack advancement, which is quite similar to
that observed for specimens made of SAE8620 steel [Refer section 4.2.4.2], but is contrary
to that observed for 20MnCr5 [Refer section 4.3.4] and EN353 [Refer section 4.4.4] steel
specimens.
From the magnified fractographs shown in the right column of Figure 4.53, it is
revealed that the crack propagates along multiple plateaus, which have concave or convex
morphology and are at different elevations with respect to each other, joined by tear ridges.
Though such features are not apparent for the specimen shown in Figure 4.53 (d), but a
178
further magnified view of the crack initiation site shown in Figure 4.54 is indicative of
crack propagation along two different planes, with slightly different orientations.
On the basis of measurements taken on the stable crack propagation region marked
with dashed curve in Figure 4.53 (a), the value of fracture toughness, KIC [Refer Equation
2.21] for the case carburized specimens made of SCM420 steel is estimated as
21.64 MPa m . Among the four different steels investigated in this study, the fracture
toughness of SCM420 is the second highest after SAE8620.
A composite, magnified fractograph covering a complete diametric swath of the
specimen shown in Figure 4.53 (d) is provided in Figure 4.55, to reveal fracture surface
topography over the entire cross-section. The specimen survived for more than 5105
cycles in fatigue and reveals a substantial extent of stage-I crack propagation region, shown
further magnified in Figure 4.56, which is typical of specimens tested under high cycle
fatigue [295]. When viewed at still higher magnification [Refer Figure 4.57], the region was
found to be dominated by factory-roof morphology, implying crystallographic orientation
of crack-propagation [298], accompanied by a change in direction at the grain boundaries.
The fracture surface of the specimen tested under low cycle fatigue [Figure 4.53 (b)]
exhibited crack-initiation on multiple planes, which can be seen in both the views of
composite fractomicrograph in Figure 4.58. The cracks are identified by arrow-marks in the
top and side views. Crack-initiation on multiple planes for the specimens tested under low
cycle fatigue is also confirmed by formation of a double edge, as shown in Figure 4.59.
These observations imply that multiple cracks are formed when the coated specimens are
subjected to low cycle fatigue. This observation is quite similar to that reported in section
4.2.4.2 for SAE8620 steel specimens, and can be attributed to substantial straining of the
substrate under high applied stresses, which leads to formation of multiple cracks within the
coating, some of which grow to macroscopic scale. This phenomenon seems to be
responsible for relatively poor fatigue performance of the coated specimens in comparison
to the uncoated ones under low cycle fatigue.
179
Figure 4.53: Optical fractographs of SCM420 steel specimens. (a): Uncoated specimen
tested at 980 MPa, (b): Coated specimen tested at 940 MPa, (c): Uncoated
specimen tested at 780 MPa, and (d): Coated specimen tested at 820 MPa.
Fractomicrographs given in the right column provide magnified views of
crack initiation sites, captured by holding the specimen in same orientation
as in the left column.
180
Figure 4.54: Magnified optical fractograph of specimen shown in Figure 4.53 (d),
depicting crack initiation site
Figure 4.55: Optical fractograph covering entire cross-section of fractured specimen

shown in Figure 4.53 (d). The fractograph is constructed as mosaic by
joining together six individual fractographs.
181
Figure 4.56: Optical fractograph of a stage-I crack propagation site in the specimen
shown in Figure 4.53 (d).
Figure 4.57: Optical fractograph showing cleavage-like crystallographically oriented

stage-I fatigue fracture exhibiting factory-roof morphology, recorded at
the region marked with rectangle in Figure 4.56.
182
Figure 4.58: Composite optical fractograph showing formation of multiple cracks in

specimen shown in Figure 4.53 (b), tested under low-cycle fatigue.
Figure 4.59: Optical fractograph showing multiple-plane cracking for specimens tested
under low-cycle fatigue.
183
4.5.5 Effect of WC/C coating on fatigue behaviour of SCM420 steel

The volume fraction of martensite in the outermost layers, as well as the magnitude of
residual compressive stresses, observed in SCM420 steel specimens are quite similar to
those observed for specimens made of SAE8620 steel. Similar to the observations made on
all other steels examined in the present study, the coating led to substantial enhancement in
the residual stress field in case carburized SCM420 steel. However, unlike 20MnCr5 and
EN353 steel specimens, hardly any internal oxidation could be observed in the specimens
made of SCM420 steel. Further, the nature of crack propagation was found to be largely
transgranular, similar to that observed in SAE8620 steel specimens, though the region of
stable crack propagation was considerably smaller than that observed for SAE8620
specimens processed in a similar manner.
An examination of the S-N graph given in Figure 4.51 reveals that in comparison to
their case carburized (uncoated) counterparts, the case carburized WC/C coated
specimens made of SCM420 steel had poor fatigue performance under low cycle fatigue,
but exhibited improved performance under high cycle fatigue. The calculations based on
fitted linear fatigue models indicate that the two graphs intersect at a value of Nf = 2.4105.
The observed fatigue performance, along with metallographic and fractographic
observations, clearly indicates a strong similarity in the fatigue behaviour of SCM420 and
SAE8620 steels [Refer section 4.2]. Therefore, it is apparent that for case carburized and
tempered substrates made of SCM420 steel, the deposition of WC/C coating will lead to
considerable enhancement in fatigue performance under infinite life regime.
Observations concerning the influence of WC/C coating on fatigue behaviour of
SAE8620 and SCM420 steels in this study are quite similar to the results reported by Su
and co-workers [10,11] in their studies on fatigue behaviour of normalized AISI 1045 steel
substrates, coated with different types of PVD coatings, viz. Titanium Nitride (TiN),
Titanium Carbonitride (Ti(C, N)), Chromium Nitride (CrN) and Chromium Carbonitride
(Cr(C, N)). The authors observed that when the substrate is subjected to high stresses,
cracks develop within the coating, which act as stress raisers and aid in early failure of the
184
specimen. However, for specimens surviving more than 5105 cycles under fatigue, the
coating provided effective protection to the surface. Ferreira et al. [193], in their study
involving fatigue behaviour of 42CrMo4 steel coated with various types of deposits, also
reported similar observations. They attributed the deteriorated performance under low cycle
fatigue to the formation of cracks within the coating.
In the present study, the cracks were found to originate at multiple sites in specimens
tested under low cycle fatigue. This has been attributed to cracking of the hard and brittle
coating, when the substrate is subjected to large strains under low-cycle fatigue. Since the
thickness of coating (2m) is more than ten times the average surface roughness, Ra (<
0.2m) [Refer section 3.2] of the specimens, the cracks developed within the coating are
likely to act as stress raisers, thereby promoting pre-mature failure of the specimen under
fatigue. On the other hand, substantial regions of slow, stage-I fatigue crack propagation
have been observed in the specimens tested under high cycle fatigue. It is under such
conditions that the large residual compressive stresses introduced by the coating are able to
play a constructive role in enhancing the fatigue performance of the specimens. Besides,
when the applied stresses are low, the coating provides an effective seal against crack
initiation at the surface.

En 353

Uploaded by

Copyright:

Available Formats

En 353

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

En 353

Uploaded by

Copyright:

Available Formats

92

RESULTS AND DISCUSSION

4.1.1 Tensile properties

Figure 4.1: Stress-Strain graph for SAE8620 steel in green state

Figure 4.2: Effect of coating on microhardness profiles of various steels

As ASTM E 384 [263] recommends availability of sufficient backup material at the

4.1.3 Characterization of WC/C coating through Raman Spectroscopy

difference in spectra recorded at different locations in a single specimen is known to result

Testing and analysis of SAE8620 steel specimens

4.2.1 Metallographic observations

M S ( F ) = 930 570C 60Mn 50Cr 30 Ni 20Si 20Mo

From this relationship, the Ms temperature of SAE8620 steel works out to be

Figure 4.4: Metallographs of case carburized SAE8620 specimens cross-section:

Figure 4.6: Mosaic of metallographs showing variation in microstructure with depth in

4.2.2 Estimation of residual stresses

Residual stress () corresponding to the recorded peak positions is estimated from

To estimate the value of material constant, K

[= E / (1 + )], of the martensitic case,

4.2.3 Fatigue testing

In order to quantify the change in fatigue life as a function of maximum alternating

* No failure; Test suspended.

Stress interval (d)= 17 MPa

Stress interval (d)= 10 MPa Stress interval (d)= 25 MPa

Mean stress (s) = 259 MPa

Mean stress (s)= 795 MPa

Mean stress (s)= 855 MPa

Std. dev. (s) = 9.01 MPa

Std. dev. (s) = 11.27 MPa

Std. dev. (s)= 13.25 MPa

An interesting phenomenon observed in Figure 4.9 is the cross-over in S-N curves of

% Change in fatigue strength =

4.2.4 Fractographic observations

Fractographic observations on specimens in green state

Fractographic observations on specimens in green state, failed at different stress

Fractographic observations on case-carburized specimens

The dominance of transgranular mode of crack propagation is clearly revelaed in

overload fracture [240,292]. However, no such transition from intergranular to

4.2.5 Effect of WC/C coating on fatigue behaviour of SAE8620 steel

Testing and analysis of 20MnCr5 steel specimens

4.3.1 Metallographic observations

The microstructure of case carburized and tempered specimen of 20MnCr5 steel is

Figure 4.22: Mosaic of metallographs, showing variation of microstructure with depth in

The microstructure within the outermost 400m is dominated by plate martensite,

4.3.2 Estimation of residual stresses

4.3.3 Fatigue testing

20MnCr5 (WC/C coated)

As expected, the fatigue performance of 20MnCr5 steel specimens in case carburized

* No failure; Test suspended.

Stress interval (d)= 25 MPa

Stress interval (d)= 20 MPa Stress interval (d)= 45 MPa

Mean stress, (s) = 234 MPa

Mean stress (s)= 890 MPa

Mean stress (s)= 713 MPa

Std. dev. (s) = 13.25 MPa

Std. dev. (s) = 17.14 MPa

Std. dev. (s) = 23.85 MPa

4.3.4 Fractographic observations

Figure 4.26: Optical fractographs showing macroscopic features on uncoated / coated

Figure 4.29: Scanning electron micrograph showing intergranular cracking in the