Fracture and Fracture Toughness of Cast Irons: W. L. Bradley and M. N. Srinivasan

Fracture and fracture toughness of cast irons
W. L. Bradley and M. N. Srinivasan
relatively wear resistant because the graphite acts as

The present state of knowledge on the a self-lubricating system. Second, the graphite
microscopic factors affecting the fracture and attenuates acoustic wave propagation, making cast
the fracture toughness of different types of iron ideal for machinery bases and other
cast irons: namely, grey, white, ductile,
components where vibration damping is desirable.
malleable, and compacted graphite
(vermicular) cast irons, is reviewed.
Despite the desirable physical properties men-
Information on composition and smooth tioned above and the economic adv~ntages of
tensile properties is also included, where manufacturing with cast iron, its use in many
necessary, better to explain the fracture quarters is limited by concern about its fracture
behaviour. Information available on the toughness. The higher carbon concentration
fracture toughness of grey and ductile irons is responsible for the desirable physical properties and
critically analysed and the reasons for ease of manufacturing with cast iron is, unfortun-
discrepancies observed in the literature ately, also responsible for the degradation of the
discussed. It is demonstrated that the fracture ductility and fracture toughness. The carbon, usu-
behaviour and the fracture toughness of cast
ally present principally as graphite, serves to
irons depend on the graphic morphology as
well as the matrix microstructure. IMR/208
nucleate the fracture process. at relatively low strain
levels, significantly impairing the fracture toughness
© 1990 The Institute of Metals and· ASM of cast iron compared with cast or wrought steel.
INTERNATIONAL. The authors are in the Mechanical Considerable effort has been expended during
Engineering Department, Texas A & M University, the past ten years to characterise better the fracture
College Station, TX, USA. toughness of cast iron as well as to understand
better the micromechanisms that control the frac-
ture process in these systems. The purpose of this
paper is to summarise and criticise these results on
Introduction the fracture behaviour and fracture toughness of
Cast iron has a long and illustrious history. Parts the various types of cast iron as a function of
made in China from the oldest member of the cast chemical composition, matrix microstructure, and
iron family, grey cast iron, have been dated to the graphite morphology.
6th century Be.1 By contrast, one of the newest
members of the cast iron family, nodular cast iron Five types of cast iron: Composition
(also known as ductile iron), was not developed and. microstructure
until 1948. What is it about cast iron that accounts The five types of cast iron produced commercially
for its widespread use in industry? First, it has today are white, grey, malleable, ductile, and
melting temperatures (and therefore pouring tem- compacted graphite cast iron. Typical compositions
peratures) that are 300-350 K lower than cast steel. of these alloys are presented in Table 1.2 The basic
Second, because of a larger concentation of free differences in composition between cast irons and.
carbon, graphitic (grey and ductile) cast iron has cast steels (or for that matter, wrought steels) is
the greatest fluidity and the least shrinkage of any that the former contain a much higher carbon
ferrous metal. For example, cast steel will generally content (1.8-4.0%). Furthermore, all cast irons
experience a volume shrinkage of more than 4% except white cast iron also have sufficient silicon to .
during solidification, whereas grey and ductile cast cause most of the carbon to form graphite rather
iron shrinkage can be less than 1%, depending on than cementite (or other iron carbides), the silicon
the composition and the processing conditions. This serving as a graphitiser. The primary effect of alloy
difference in shrinkage allows products to be made composition on the mechanical. properties of cast
to exact dimensions much more easily using grey irons (other than alloy cast irons) is indirect; that is,
and ductile cast iron than cast steel, with very little through the effects the various alloying elements
problem in obtaining pressure tightness as a result have on graphite morphology and matrix micro-
of reduced interdendritic shrinkage. A third advan- structure. To a lesser degree the presence of the
tage of graphitic cast iron compared with cast steel elements other than carbon in solution or as second
is its greater machinability. The combined effects of phase particles may also influence the mechanical
the above three factors generally make it 30-40% properties of cast iron.
cheaper to manufacture with grey and ductile cast With the exception of white cast iron, all casf
iron than with cast steel. irons have in common a microstructure that consists
In addition to being cheaper materials with which of a graphite phase (usually 8-14 vol._%) in a
to manufacture, cast irons have some physical matrix which may be ferritic, pearlitic, bainitic, or
properties that give them a competitive advantage tempered martensitic. A combination of chemistry
when compared with cast steel. First, they are and thermal· history is used .to control the two
International Materials Reviews 1990 Vol. 35 NO.3 129
130 Bradley and Srinivasan Fracture of cast irons
Table 1 Range of compositions for typical

unalloyed cast irons, wt·% (Ref. 2)
Type C Si Mn S P
White 1·8-3·6 0·5-1·9 0·25-0·8 0·06-0·2 0·06-0·2

Malleable
(cast white) 2·2-2·9 0·9-1·0 0·15-1·2 0·02-0·2 0·02-1·0
Grey 2·5-4·0 1·0-3·0 0·2-1·0 0·02-0·25 0·02-1·0
Ductile 3·0-4·0 1·8-2·8 0·1-1·0 0·01-0·03 0·01-0·1
Compacted
graphite 2·5-4·0 1·0-3·0 0·2-1·0 0·01-0·03 0·01-0·1
microstructural constituents which determine the

mechanical properties of cast iron, namely, graphite
morphology and matrix microstructure. Graphite in
cast iron appears in several morphologies including
leaf-like flakes, 'pop corn' shaped globules,
spheres, and winding tubes, with the graphite
morphology used as the basis for classifying the Cast iron containing graphite
various types of cast iron. The iron-iron carbide-silicon phase diagram is
presented in Fig. 3.5 The graphite phase, when it
Cast iron containing no graphite: forms instead of iron carbide, will assume a variety
White cast iron of shapes which depend on the composition, cool-
If the composition of the iron is suitably low in ing rate, and melt treatment (types of inoculants
silicon and/or the cooling rate sufficiently fast, added), or in the case 'of malleable iron, the
carbon will react with iron to form carbides such as post-solidification heat treatment. The various types
cementite (rather than the equilibrium phase, which of graphite which may form in cast irons have been
is graphite), resulting in a hard alloy commonly classified by the American Society for Testing and
known as white cast iron. The iron-iron carbide Materials (Specification A247)6 and are shown
phase diagram shown in Fig. 1 (Ref. 3) indicates schematically in Fig. 4.7 (For classification of the
that a carbon content greater than 2·0% would type of graphite in actual castings, the standard
result in a microstructure of pro eutectic cementite photographs available from ASTM should be used.)
in a matrix of eutectic cementite and austenite. On Type I graphite is essentially spherical and is the
subsequent cooling, the austenite will transform to usual form that occurs in ductile iron. Type II
pearlite, giving the microstructure shown in Fig. 2,4 graphite is a degenerate form of Type I graphite,
namely, pro eutectic cementite (white needles) in a being somewhat less spherical, and as a result,
eutectic matrix of cementite (white) and unresolved giving a nodular cast iron with somewhat less
pearlite (dark). It is readily seen that the micro- desirable mechanical properties. Type III gr~phite
structure of white iron is 'rich' in cementite, is globular and irregular in shape, resembling pop
resulting in a very hard, wear resistance alloy which corn. It is the form of graphite present in malleable
is used to advantage in applications requiring iron, which is formed from annealed white iron.
abrasion resistance. However, the fracture tough- The microstructure of a malleable cast iron is
ness is very low, as would be expected, rendering presented in Fig. 5,8 where the white matrix is
white iron unacceptable for most structural ferrite and the dark, second phase is graphite. Type
applications. IV graphite in Fig. 4 is typical of compacted
1500 1500
1400 1400
1300
~ () ~ 1300
()
•.. 0 0
...
1200
(1)
:5 ~
:J
~:J
1200 (1)
:5
co 1100 co co 1100 co
CD CD CD CD
a. a. a. a.
E 1000 E 1000
E E
(1) Austenite & Carbide (1) (1) (1)
l- I- I- I-
900 900
800 800
Eutectoid Temperature
Ferrite & Austenite
700 & Carbide 700
900 Ferrite & Carbide Ferrite & Carbide
600 600
1.0 2.0 3.0 4.0 5.0 2.0 3.0 4.0 5.0
Carbon Content, wt - °/0 Carbon Content ,wt-o/o
1 Iron-iron carbide phase diagram (Ref. 3) 3 Iron-iron carbide-silicon phase diagram (Ref. 5)
International Materials Reviews 1990 Vol. 35 No.3

Bradley and Srinivasan Fracture of cast irons 131
6 Microstructure of compacted graphite cast iron,

unetched (Ref. 9)
4 The seven types of graphite as established by nucleation. Fine flakes form initially but recales-
ASTM specification A247 (Refs. 6, 7) cence raises the eutectic graphite temperature,
resulting in coarse flakes surrounding them. Type C
flakes are otherwise known as 'kish' graphite and
graphite iron, which has a structural constituent of form in hypereutectic castings. They are undesir-
graphite which forms in very long tubular shapes. A able for good mechanical behaviour but can
cast iron microstructure containing primarily of increase the thermal conductivity of the casting.
compacted graphite is presented in Fig. 6.9 The Type D graphite, sometimes called undercooled
graphite shapes typically observed in Types V and graphite, forms when solidification occurs at a large
VI graphite as shown in Fig. 4 are sometimes' undercooling. Although it is not detrimental to
observed in ductile iron, where they are considered mechanical behaviour in itself, the mechanical
degenerate, and keep one from obtaining the best properties in sand castings with Type D graphite
mechanical properties. tend to be low because of the associated ferritic
Type VII is the flake graphite form that occurs in
grey iron. A more detailed classification of Type
VII graphite (flake graphite) in A247 has also been
made by ASTM, as shown schematically in Fig. 7.10
(As with Fig. 4, the type of flake in an actual
casting should be determined with reference to the
standard photographs available from ASTM.) Type
A is a random distribution of flakes of uniform
size and is preferred for mechanical applications.
Correct inoculation of liquid iron with adequate
graphitisation potential favours the formation of Type B
this type of graphite. Type B graphite forms a
rosette pattern as a result of a low degree of
Type C Type 0
Type E
5 Microstructure of malleable cast iron, etched in 7 The five types of flake graphite, a subclassifi-
40/0picral (Ref. 8) 'cation of Type VII in Fig. 4 (Ref. 10)

matrix. Type E graphite forms in strongly hypo-

eutectic cast irons of low carbon equivalent
that form a
strong primary austenitic dendritic
structure before eutectic solidification. The prefer-
red orientation can result in directional properties.
It may be surmised that the size, shape, and
volume fraction of the graphite in cast iron play a
dominant role in determining the strength and
fracture toughness of these alloys. This is clearly
demonstrated below .
. Producing graphite and controlling its shape
In cast iron
Malleable iron Orily in malleable cast iron is the
graphite formed during a post-solidification heat
treatment, the graphite in grey, nodular, and
compacted graphite all forming from the melt. To
make malleable iron, a white iron casting is initially
made (carbon present entirely as iron carbide) and
subsequently given an extended heat treatment at
temperatures above 900 C. This heat treatment
D
allows graphitisation in the solid state. The rapid. 9 SEMmicrograph of ductile iron showing graphite
solidification necessary to form white iron limits the nodule (Ref. 12)
section thickness in the casting that is practical for
the malleable iron process. Furthermore, the low
shrinkage and good castability usually associated iron and its derivative malleable iron) requires a
with grey and ductile iron are not realised in suitably high silicon content (silicon acting as a
malleable iron because it solidifies as white iron graphitiser) and an appropriately slow cooling rate.
(i.e. with no graphite formation). Depending on the The mechanical properties of grey iron are
cooling rate following the critical anneal, the matrix determined by the size, amount, and distribution of
surrounding the graphite may be ferritic or the graphite flakes and by the hardness of the metal
pearlitic. matrix surrounding the graphite flakes. These fac-
Grey iron Carbon is present in grey iron prin- tors are in turn controlled by the silicon and carbon
cipally as separate graphite flakes formed from the content and the cooling rate. Higher silicon and
melt during solidification (Fig. 8).11 When a grey carbon content along with slower cooling rate
iron casting is broken, most of the fracture occurs favour the formation of larger flakes of graphite
along the graphite flakes, giving a very greyish and a softer matrix, leading to a weaker alloy.
appearance to the fracture surface, and thus, the Because the graphite forms from the melt, the net
name grey iron. The formation of graphite from the shrinkage of grey cast iron during solidification is
melt rather than cementite (as is the case for white less than 1010, making it very attractive for the
production of pressure tight castings. It is also much
easier to produce components to exacting dimen-
sions'in grey cast iron than can be done with cast
steel, as noted above. The primary liability of grey
iron is that the flaky graphite extracts a very high
penalty in mechanical properties, lowering both
strength and fracture toughness.
Ductile iron Ductile iron is a modern improve-
ment on the ancient grey iron. Like grey iron, the
carbon forms as graphite during solidification.
Unlike grey iron, the graphite in ductile iron is
present in nearly spherical globules, as seen in Fig.
9,12 rather than leaf-like flakes - hence the name
spheroidal or nodular cast iron.
There is no completely adequate theory that
explains fully the complete solidification features of
cast iron with nodular graphite.13 However, as this
change .i~ the graphit~ shape is. achieved through
the addItIon of magneSIum or cenum to the melt it
is generally believed that either of these eleme~ts
neutralises the effect of surface active elements
such as sulphur and oxygen, leading to suppression
of the preferential growth of graphite in its basal
8 SEM micrograph of cast iron with flake graphite planes.14
(Ref. 11) Like grey iron, ductile iron experiences very little
International Materials Reviews 1990 Vol. 35 NO.3
Bradley and Srinivasan Fractu re of cast irons 133
shrinkage during solidification, making it a very per square millimetre, as counted in a photomicro-
easy alloy to cast. Because it combines the ease of graph at a magnification of 100. Second, the nodule
manufacturing typical of grey iron with tensile shape is characterised as the percentage of nodular-
properties more near1y like those of cast steel, ity. Here the percentage. of graphite nodules with
ductile iron has experienced a dramatic increase in an aspect ratio less than 2 are reported. Third, the
its use in industry during the past 25 years.15 volume fraction of graphite may be calculated from
Compacted graphite iron This newest· member of area ratios in a photomicrograph.
the cast iron family has most of its carbon in the Under practical conditions, the nodule count is
form of irregular tubes that are interconnected in normally controlled using a post-inoculation
each cluster, as shown in Fig .. 10.16 This graphite treatment with a silicon bearing material. In
structure and the resulting rp.echanical properties commercial castings the nodule count may vary
are intermediate between those of grey iron and from as low as 50 mm -2 in sand castings to as much
ductile iron. Sometimes called vermicular graphite as 300 mm-2 in permanent mould castings. Inad-
cast iron, compacted graphite iron retains much of equate inoculation has been recognised to be a
the cast ability 'of grey iron, but has higher tensile primary factor for poor nodularity in ductile iron
strength and ductility. Its production process is castings.18 It has also been suggested that fluid flow
similar to that of ductile iron with regard to melt due to thermal convection also results in loss of
chemistry and nodularising additions of magnesium. nodidarity.19,20 In malleable iron, graphite growth
However, it also requires an antispheroidising occurs along both basal and prism planes. The
element such as titanium to restrict the spherical presence of small amounts of hydrogen, mag-
graphite formation. Scanning electron microscopy nesium, and cerium encourages growth along the
of deeply etched compacted graphite irons (i.e. prism plane, promoting spherical growth. 18
irons with insufficient spheroidising agent or irons
with spheroidising and antispheroidising agents Types of matrices present in cast iron
being present) has shown that the graphite is Depending' on the alloy composition and thermal
interconnected within the eutectic cell but that history, a variety of matrices can be produced in
growth occurs in the c direction as well as the a cast iron.
direction. The successful growth of compacted Ferritic matrix A ferritic matrix can be obtained in
graphite requires a balance between flake promot- the as cast condition in cast iron, but it is generally
ing elements, such as sulphur and oxygen, spheroid- necessary to anneal the material after casting to
ising elements, such as magnesium or cerium~ and obtain a fully ferritic microstructure. To obtain a
antispheroidising elements, such as titanium and fully ferritic matrix in the as cast state, it is
aluminium. 17 necessary to cool very slowly a cast iron with
sufficiently high silicon content. Under these condi-
tions, part of the carbon in the melt is transformed
Quantitative metallography of cast iron to graphite during solidification and the carbon that
To characterise quantitatively the graphite shape remains in solid solution in the austenite will have
and distribution in cast iron, several parameters adequate time to diffuse to existing graphite par-
have been designated. First, the nodule count is ticles during the cool down, avoiding the formation
determined as the number of nodules of graphite of matrix carbides. Post-solidification ferritisation
of the matrix may be done either above the critical
temperature (called a critical anneal) for a shorter
time or below the critical temperature for a longer
time (called a sub critical anneal). Where a mixed
matrix of ferrite and pear1ite is present, the· ferrite
will usually surround the graphite and can precipi-
tate on the existing graphite with a minimum of
diffusion required.
Pearlitic matrix When the solidification rate and
the subsequent cooling rate leave inadequate
opportunity for the carbon to form the equilibrium
graphitic structure exclusively, some carbon may
form a pear1itic structure. Although it may re-
semble the pear1itic structure in steels, a pear1itic
matrix in cast iron usually contains less than the
0·8% carbon present in steels. Faster solidification
and post-solidification cooling rates favour the
formation of pearlite in preference to ferrite in the
matrix by limiting the diffusion of the carbon in
solution in the matrix to the second phase graphite
which formed during solidification.
Bainitic and tempered martensitic matrices An
acicular or bainitic microstructure can be produced
10 SEM micrograph of compacted graphite cast by quenching into a hot salt bath and austempering.
iron (Ref. 16) . A bainitic microstructure may also be produced in
as cast iron that has been appropriately alloyed with

nickel or molybdenum.21 A tempered martensitic 250
Class 50
microstructure in cast iron is· produced with a heat
treatment similar to that used for steel, including a Class 40
quench and temper. It is shown below that a (\j 200
bainitic matrix gives a more desirable combination E Class 35
z
of strength and toughness than tempered martensite :2
for cast iron. en
~150 Class 30
Primary carbides While white iron has a matrix Ci)
that is dominated by the presence of carbides, the ~
other cast irons may also contain primary carbides. (f)
~100
Faster cooling rates present near the end of r-
solidification. may trap carbon in the form of
primary carbides that would otherwise have solidi- 50
fied as graphite. These carbides are generally more
blocky in shape than the thin lamellar carbides
found in pearlite. They are detrimental to mechani-
cal properties, especially to machinability, and thus, 0.1 0.2 0.3
are to be avoided where possible. They can Strain, %
generally be eliminated by austenitising followed by 11 Tensile stress-strain behaviour of several grey
slow cooling. cast iron samples (Ref. 23)
Tensile properties of the various cast irons tensile strengths rather than their respective alloy
The tensile properties of the various cast irons are chemistry. Thus, a Class 20 grey cast iron is
summarised in Table 2.22 guaranteed to have a UTS of at least 138 MN m-2
(20 ksi). The 0·2% offset yield strength is not
Grey cast iron
related to matrix deformation but rather to some
Tensile test results for several classes of grey cast
initial amount of opening of the graphite flakes
irons are shown in Fig. 11.23 Itshould be noted that
which gives the required 0·2% offset. The ultimate
the grey cast iron does not have a well defined
compressive strength (UCS) of grey iron is usually
linear region in its tensile stress-strain relationship.
much higher than the ultimate tensile strength
At relatively low stress levels, the graphite flakes (UTS) since the graphite flakes are much more
begin to open up as small microcracks resulting in a
detrimental to strength when loaded in tension
non-linear load-displacement curve. Grey cast irons
than when loaded in compression. The modulus,
are usually classified by their minimum guaranteed
measured as the -initial tangent to the stress-strain
curve, is seen to be quite low compared with pure
Table 2 Tensile properties*- of cast irons (Ref. 22) iron or steel and reflects the. significant deflection
occurring even at low loads that results from the
YS, UTS, UCS, E, EI.,
Materialt MN m-2 MN m-2 MN m-2 GN m-2 %
opening of graphite flakes. In contrast with steel,
the tensile modulus is seen to increase with increas-
Grey irons
114 <1
ing strength, a stronger matrix reducing the ease
Class 25 as cast 206 759
Class 30 as cast 232 893 117 <1 with which graphite flakes can open and reduce the
Class 30 annealed 142 576 100 effective moduli in grey irons.
Class 35 as cast 240 869 124 <1
Class 40 as cast 289 1070 126 <1 Malleable cast iron
Malleable irons The designations for malleable iron in Table 2 refer
M3210 224 345 157 >10
to the minimum 0·2% yield strength in ksi (first two
M4504 310 448 165 >4
M5503 379 517 171 >3
numbers) and the minimum guaranteed elongation
M7002 483 621 177 >2 in a 2 in gauge section (last two numbers). The pop
Ductile (nodular) irons corn shaped graphite (Fig. 12)24 is seen to be less
60-40-18 (ferritic) 276 414 172 >18 detrimental than the flaky graphite in grey iron to
80-60-03 (pearlitic-
ultimate tensile strength and elongation. The modu-
ferritic) 414 552 172 >3
100-70-02 (pearlitic) 483 690 172 >3
lus is also seen to be both much greater than that
120-90-02 (tempered for grey iron and relatively independent of the
martensitic) 621 828 172 >2 tensile strength.
Compacted graphite iron
Type 1 (pearlitic) 379 448 1212 ,45 1 Ductile iron
Type 2 (pearlitic) 276 345 145 1 The six number designations for ductile iron give
Type 3 (ferritic) 193 276 145 4
the guaranteed minima for ultimate tensile strength
White iron
ASTM A532 I-A
(in ksi), the 0·2% offset yield strength (in ksi), and
(VHN-550) 276-345 ... 166-179 <1 the percentage elongation. The higher strength
noted for ductile iron compared with malleable iron
* YS, 0·2% offset yield strength; UTS, ultimate tensile strength; UCS,
ultimate compressive strength; E, Young's modulus; EI, elongation. is due to the greater solid solution strengthening by
t The material identification numbers are described in the text. silicon and to the better graphite' nodule shape.
150
125 Unnotched
E
Z
~100
~
~ 0>
Q)
Lij 75
~
a.
~ 50 Notched
.c:
o
25
o 120 160 200 240 280 320

Test Temperature, K
13 Fracture energy curve for, standard Charpy V-
notch and unnotched specimens of malleable
12 SEM micrograph of malleable cast iron showing iron (Ref. 26)
irregular (pop corn) type graphite (Ref. 24)
specimens or fatigue precracked Charpy specimens

Compacted graphite iron instead of the standard blunt notched Charpy
The tensile properties for the ferritic compacted specimen (0·254 mm root radius). When this is
graphite cast iron are seen to fall between those of done both the energy absorbed and the transition
grey iron and those of ductile iron, as one would temperature indicated are affected. This behaviour
expect. The effect of microstructure on the tensile is illustrated with the experimental results presen-
properties of the various cast irons can be seen in ted in Fig. 13 (malleable iron)26 and Fig. 14 (ductile
Table' 2 by comparing the tensile properties of cast iron),27 where comparisons of energy absorption of
iron in any given family of cast irons where the Charpy specimens with standard notches are
graphite shape would be common and only the compared with results obtained on Charpy speci-
matrix changes (e.g. ferritic ductile iron to tem- mens which were unnotched and on Charpy speci-
pered martensitic ductile iron). Unlike the effect of mens which were notched and then fatigue
graphite shape, the effect of the matrix is such that precrackecl before testing.
tensile strength and ductility generally change in When dynamic tear tests are made, on cast iron,
opposite directions, with higher strength matrices the standard approach of introducing a crack
producing lower tensile elongations of the through the use of a brittle weld bead has been
respective· alloys. replaced by a machined-in notch of root radius
0·025 mm. Since dynamic tear specimens can be
Methods of evaluating fracture made in various sizes but are generally larger than
standard Charpy specimens and also have a sharper
toughness of metals and alloys
The test methods used to assess the fracture
behaviour of metals and alloys may be broadly 0.22
divided into two groups: (a) those which measure 0.20 •
the transition temperature where the fracture pro- Charpy V-Notch
o
cess changes from ductile to brittle, and (b) those ~ 0.18 •
which measure the resistance to flaw propagation ~ 0.16 o
using a fracture mechanics approach. cd"
~ 0.14
«
:g 0.12
o
Transition temperature approach ::>
In the transition temperature approach, the energy >. 0.10
C>
o
required to initiat~ and propagate a crack thr0.ugh a ~ 0.08
specimen, giving complete fracture of the specImen, w
Q) 0.06
is measured over a range of temperatures. The total ~ o
energy required to fracture the specimen is con- ~ 0.04
sidered a relative measure of the fracture toughness u: 0.02
of the material.
Dynamic tear. tests (ASTM £604-83 )25 and ~60 -40 -20 0 20 40 60 80
Charpy impact tests have been used to measure the Temperature,OC
transition temperature in cast irons. In characteris- 14 Fracture energy curves for standard Charpy
ing the fracture behaviour of cast iron, some V-notch and fatigue precracked specimens of
investigators have chosen to use unnotched Charpy ductile iron (Ref. 27)
International Materials Reviews 1990 Vol. 35 No.3'

(ASTM E399-81);29 (b) the I-integral elastic-

't
0
plastic approach (ASTM E813-81);30 and (c) the
X 10 crack tip opening displacement, or CTOD elastic-
plastic approach (BS 5762).31 Each of these ap-
I"
E • • proaches allows one to determine a fracture
Z
8 toughness parameter which relates flaw size, nom-
ci inal stress at which that flaw will propagate, and
~ Q)
~.•..• fracture toughness of the given material. Compari-

c 6 sons of fracture toughness measurements based on
::>
.•.•..... anyone of these three approaches should allow one
>-
0) to compare quantitatively flaw tolerance in various
~ Impact
Q)
c 4 materials, preferably at some common fraction of
w their respective yield strengths. It is important to
~ Q)
note that some of the significant work reviewed in
::J
'0 2 this paper was performed before the final adoption
~ of the elastic-plastic standards for the I-integral and
at
CTOD methods. The use of these tentative stand-
o 160 180 200 220 240 260 280. 300 ards which were later revised has in several cases
Test Temperature, K led to some erroneous results being reported in the
literature. Where possible, these experimental
15 Influence of loading rate on fracture energy of results have been reanalysed using the various
annealed ductile iron with 2'95% 5; samples current standards for fracture mechanics testing.
were standard Charpy V-notch specimens with Since a number of authors in the mid 1970s
fatigue precrack of 0·76 mm: closed symbols
reported their fracture mechanics results on cast
indicate ductile fracture whereas open symbols
correspond to brittle fracture (Ref. 28)
iron in terms of Ko and/or Kmax, whether, any
useful information can be gleaned from such results
should be considered. Bradley32 argued that for
notch, neither the transition temperature nor the ductile material behaviour, Ko values will always
energy absorbed in fracturing specimens for these underestimate K1c• However~ Kmax can be larger or
two tests are comparable. The greater constraint in smaller than the actual K lc depending on the
the dynamic tear test will always give a lower degree of non-linearity in the load-displacement
ductile to brittle transition temperature than that curve and the amount of stable crack growth that
observed in a Charpy test. It should be added that occurs before maximum load. The I-integral may
Charpy tests measure principally the crack initiation be calculated using the standard relationship30
energy whereas <;lynamictear tests measure mainly
I= rxA/Bb (1)
the propagation energy, giving yet another reason'
why results' from these tests are not directly where A is the amount of work done, b the
comparable. uncracked ligament width, B the specimen
A comparison of the transItion temperature thickness, and rx a geometry dependent scaler; rx
measured in dynamic tests such as dynamic tear or assumes a value of 2·0 for three point bend
Charpy with the transiton temperature measured in specimens with deep cracks. The associated stress
a quasistatic fracture mechanics test is also inappro- intensity can be calculated from I using the
priate. The yield strengths of cast irons have a relationship
similar rate dependence to that of ferritic steels,
and thus give a transition temperature that is quite
1 = K2/E' (2)
rate dependent. For example, a ferritic ductile iron where E' is the effective modulus (depending on
is seen in Fig. 15 to have a shift in its ductile to degree of constraint). In terms of critical values,
brittle transition temperature of over 50 K when equation (2) may be written as
the loading rate is changed from slow bend to (3)
llc = K]c/E'
impact.
In summarising this section, it should be empha- Combining equations (1) and (3) above gives
sised that determining the effect of composition and KJc = (rxE'A/Bb)1/2 (4) .
microstructure on the fracture toughness of cast
iron using comparisons of transition temperature Figure 16 is a schematic of a load-displacement
must be done with great care to avoid drawing curve for a ductile material. The use of the low
incorrect conclusions. Only where the same loading energy fracture mechanics (LEFM) relationships' in
rate, specimen size, and notch acuity are used can ASTM E399 in combination with Po or Pmax' gives
valid comparisons be made. values for Ko and Kmax that could also be calcu-
lated from equation' (3) using the areas ABC anq
AB' C' respectively, noting B' is at the same load
Fracture mechanics approach to value as P max. The true value of KJc may be
measuring fracture toughness calculated using the areas AB"C", assuming that
There are three fracture mechanics approaches to crack growth begins at a displacement of C". It is
measuring fracture toughness in metals and alloys: . clearly seen that the value of KJc will always be
(a) the linear elastic fracture mechanics approach greater than or equal to Ko for ductile material
size, shape, and volume fraction of graphite; and on

the matrix microstructure. In this section, the direct
and indirect effects of composition on the transition
temperature and upper shelf fracture toughness of
cast iron are reviewed. It is often the case that
changes in alloy chemistry or microstructure that
t increase the upper shelf fracture toughness also
increase the ductile to brittle transition tem-
perature. Such changes (e.g. decreasing, the nodule
count) while increasing the fracture toughness at
higher temperatures may actually lower the fracture
toughness in the temperature range of interest for a
given application.
The principal elements that play a role directly in
determining the fracture toughness of ferritic duc-
A e e' e" G'"
tile iron are silicon, phosphorus, and carbon.
Displacement (6) --+ Copper, tin, and manganese play a role indirectly
through their influence on the tendency to form a
16 Schematic load-displacement curve for ductile pearlitic matrix. Magnesium, lead, sulphur, and
material (Ref. 32)
antimony concentrations must be carefully con-
trolled because they affect the shape of the graphite
behaviour. At the same time, KJc may be less than phase. Chromium forms chromium carbides which
or greater than Kmax calculated using P max in the decrease toughness and machinability. Sulphur also
LEFM relationship in ASTM E399 that relates P to combines with manganese to form manganese
K. This exact location of B"C" will determine sulphides, which degrade the upper shelf fracture
whether this area (and therefore K Ic) is larger than toughness. A more detailed discussion of the effects
AB'C' (and Kmax). of these elements on the transition· temperature and .
In summary, KQ values reported in the literature upper shelf fracture toughness follows.
may be considered lower bound values of KI~ for
ductile material behaviour. Kmax values may be Silicon
greater than or lower than Kic. Reported values of Silicon, an essential element in all cast irons except
lic and Dc in the literature must be carefully white cast irons, facilitates the nucleation of graph-
reviewed to determine how they were calculated ite, reducing the amount of carbon that is present
and corrections must be made if the reported values as iron carbides (cementite). Silicon is a potent
were calculated with early versions of the standards solid solution strengthener as evidenced by the
for CTOD and ldetermination. recent work of Salzbrenner, 34who determined that
the yield strength of ductile iron increases linearly
with increasing silicon content, a 1 increase in
%
Dynamic fracture toughness:
silicon content giving a 105 MN m-2 increase in the
A fracture mechanics approach
0·20/0 offset yield strength. Since the ductile to
Dynamic fracture ~oughness parameters, KId or lId, brittle transition temperature (IT) usually increases
are widely measured using fatigue precracked with increasing yield strength (except when the
Charpy bars that are broken in a drop tower or increase in yield strength is via grain size refine-
Charpy machine with an instrumented tup. The ment), one would expect increasing the silicon
recorded load-time record can be used to calculate content to increase the TT through silicon's effect
a load-displacement history for the test. If the on the yield strength. Figure 17,35 which gives
moment of crack extension can be determined, then Charpy V-notch energy v. temperature for several
the measured data can be used to determine KId or ferritic ductile irons with a wide range of silicon
lId' Quasistatic tests on nodular cast· iron have contents indicates an "-'5 K shift in the TT for each
shown that the moment of crack extension for IT 0·1 % increase in silicon content. A lower silicon
compact tension specimens (ASTM E399-81 )29 as content has also been found to give a slightly higher
well as for fatigue precracked Charpy specimens is upper shelf fracture toughness, again probably
essentially at the point of maximum loading during through the lower yield strength34 and the larger
the test. 33 Thus, one may reasonably assume that plastic zone size that forms at the crack tip.
crack growth during a dynamic test of cast iron also The above observations suggest that it is desir-
begins when the maximum load is reached. This is able to keep the silicon content as low as possible if .
the approach that has been used in work reported one desires good low temperature fracture
in the literature thus far. toughness. However, it should be emphasised that
lower silicon content increases the probability of
Effect of alloy composition on fracture matrix pearlite formation in as. cast ductile iron,
toughness of cast irons necessitating a subsequent heat treatment to obtain
a fully ferritic microstructure to give good low
General temperature toughness.32 An alternative is to
The fracture toughness of. various cast irons substitute some nickel for silicon, since nickel is
depends on their respective compositions; on the also a good graphitiser and a less· potent solid
25
Silicon
20 1.88%
E
Z 20
~ 15
0) 3.63% Phosphorus
E
CD
c
z
ill ~
0)
~10 CD 15
m
..c
c
ill
() >-
a.
5 m
..c
() 10
0200 220 240 260 280 300 320 340 360 380 400
Temperature, K
5
17 Effect of silicon content on ductile to brittle
transition of ferritic ductile iron, as measured
using standard Charpy V-notch specimens 200 220 240 260 280 300 320 340· 360 380 400
(Ref. 35) Temperature, K
19 Effect of phosphorus content on Charpy V-notch

impact energy for ferritic ductile iron (Ref. 35)
solution strengthener than silicon when used as an
alloy in ductile iron.34 Thus, one can achieve an
essentially ferritic microstructure in the as cast fracture toughness. Like silicon, the effect of
condition, but with the economic penalty associated phosphorus on fracture toughness is primarily on
with higher cost nickel substitution for the lower the ductile to brittle transition temperature as seen
priced silicon. in Fig. 19, which gives Charpy energy. v. temper-
ature for ferritic ductile iron when the phosphorus
Phosphorus
content is varied. An increase in phosphorus
Phosphorus is a common impurity element found in
content from 0·05 to 0·10% can increase the
cast iron. In small concentrations it is present in
transition temperature by 60 K. Thus, maintaining
solution, but above about 0·05% it forms a very
the phosphorus content below 0·05% is essential to
brittle eutectic structure which occurs at the cell
achieve good low temperature fracture toughness.
boundaries. The eutectic structure is called steadite,
as seen in Fig. 18,36 and is very detrimental to the The eutectic phosphide structure also has a detri-
mental effect· on the upper· shelf fracture energy.
Carbon
Most of the carbon in a ferritic cast· iron is present
as graphite nodules. Increasing the carbon content
can result in a higher nodule count and/or larger
nodules (or flakes). To the extent that increasing
the carbon content results in a higher nodule count,
it will lower the transition temperature and the
upper shelf fracture toughness. Details of the effect
of nodule size, shape, and volume fraction on
fracture toughness are discussed in a subsequent
section ..
Magnesium
Magnesium is thought to act as a scavenger of other
tramp impurities which selectively absorb on certain
crystallographic planes of the graphite allowing
growth in only two dimensions, thus assisting the
formation of a' graphite phase which is essentially
spherical in shape. The amount of magnesium
required to produce spherically shaped graphite
varies with the concentration of tramp impurities
present in the melt. Typically, magnesium concen-
trations vary from 0·02 to 0·080/0. Details of the
effect of magnesium and other elements on the
formation of spherical graphite nodules have been
summarised by Lux.37 Wallace38 and Dong39 have
18 Microstructure of grey cast iron showing indicated that cerium, barium, and calcium may be
steadite (angular white islands), etched with 20/0 effective as supplemental inoculants to magnesium
picral (Ref. 36) particularly in dealing with specific impurities.
Sulphur Table 3 Comparison of Charpy type impact values

Sulphur should normally be maintained below for various grey cast irons (Ref. 42)
0·02% to avoid lowering the fracture toughness of Type* Charpy impact value,t N m
ductile iron. The effect of sulphur. is principally on
Undercooled graphite 7·25
the upper shelf fracture toughness, since it causes 0·056%P 16·81
graphite to form non-spherical shapes and forms 0·125%P 12·95
coarse second phase particles that are poorly 0·5%P 8·81
bonded to the matrix, facilitating early void 0·77%P 8·88
nucleation.40 Acicular
Acicular
26·30
35·12
* %P refers to phosphorus content.

Antimony
t Unnotched 16 x 16 x 126 mm specimens.
Antimony can have a very detrimental effect· on
fracture toughness by preventing the formation of a
spherical. graphite shape that would otherwise shelf fracture toughness are discussed in a sub-
occur. 40It also assists in the formation of a pearlitic sequent section.
matrix at concentration levels of 0·05-0·10% (Ref.
Chromium
40). Since fracture toughness of ductile iron is
Chromium is a very potent primary carbide former.
adversely affected by pearlite (as discussed in detail
Chromium carbides both degrade the upper shelf
below), good fracture toughness is facilitated by
fracture toughness and also reduce the machin-
keeping the antimony below 0.05%, as is common
ability. Thus, chromium is typically kept below
in good commercial practice.
0·04 % for ferritic and 0·10% for pearlitic grades of
ductile iron. Slightly higher amounts are allowed in
Lead and bismuth grey iron.
Lead and bismuth have a very potent effect on the
fracture toughness of ductile iron through the
degenerate graphite nodule shapes they produce. Summary of effects of alloy composition
Thus, they should be maintained below concen- on fracture toughness
trations of 0·003% and 0·01 %, respectively, if In this section the significant effects that even trace
optimal fracture toughness properties are to be amounts of many elements can have on the
obtained.4o The quantitative effects of nodule mechanical properties of cast iron, particularly
degeneracy caused by excessive lead, bismuth, ductile iron, have been emphasised. Wallace38 has
sulphur, or antimony are considered in a sub- noted that the mechanical properties of ductile cast
sequent section. iron are more sensitive to trace element effects than
almost any other commercial material. This is
Manganese .because trace elements greatly affect the shape of
Manganese works against good fracture toughness the graphite phase as well as the amount of carbon
in ductile iron .in two ways. First, it favours the that is likely to be retained in the matrix as pearlite
formation of a pearlitic matrix. Second, manganese or bainite. The graphite morphology and the matrix
forms intercellular carbides when present in con- microstructure in turn essentially determine the
centrations greater than 0.70 %
Maximum
• fracture toughness of the cast iron. The effects of
toughness, which is achieved in ferritic grade cast other elements on the fracture toughness of ductile
iron, required that the manganese content be less iron can be inferred (through their effect on the
than 0·20% (Ref. 40). microstructure) from the 'Iron castings handbook,41
(e.g. as summarised in Table 3) or from an
Tin excellent datasheet published in Metals Progress. 40
Tin has a very potent influence on the tendency of
the matrix to form pearlite. In concentrations Effect of microstructure on fracture
greater than 0.10%, it also promotes non-spherical process and fracture toughness
graphite particle morphology. 40Thus, good fracture
toughness requires the maintenance of tin at con-
of cast irons
centrations less than 0.10%. Effect of graphite
The primary effect of graphite on cast iron fracture
Copper is on the upper shelf fracture toughness, with a
Copper also has a very strong influence on the more modest (and probably indirect) effect on the
tendency of the matrix to form pearlite, though it is ductile to brittle transition temperature. Direct
less potent than tin in this regard. However, it does observations in the scanning electron microscope of
not adversely affect graphite shape as tin does.38 the microstructural features of crack growth and
Thus, where strength is to be developed through fracture surface morphologies in cast iron fracture
formation of a. pearlitic matrix, copper is the clearly indicate the dominating influence graphite
alloying element of choice. However, optimal has on crack propagation in both grey and ductile
toughness requires a ferritic matrix, in which case cast iron. For example, Fig. 20 shows the details of
the copper concentration should be kept below SEM observation of the polished surface of a
0·100/0. The quantitative effects of pearlite in the ferritic ductile iron sample. The sample was sub-
matrix on the transition temperature and upper jected to incremental strains through bending,
before SEM observation. (The incremental strain ahead of the main crack front; the latter may be
units shown in Fig. 20 are arbitrary and each unit seen at the upper right corner in Fig. 20b and c.
represents one degree revolution of a threaded The degree to which graphite nodules facilitate
screw used to apply bending load to the specimen.) . premature void nucleation and coalescence in duc-
As the strain increases, debonding begins at the top tile iron is seen in the results of Sandoz et al., 35 who
and/or bottom of· the nodule and then propagates found the upper shelf Charpy toughness of a 2·250/0
around the periphery until debonding is completed. Si steel to be 90 N m whereas the addition of 3.6°10
Debonding is followed by microcracking between carbon in the form of graphite nodules (""'12
adjacent nodules, giving void coalescence well vol.-olo) reduced this value to 15 N m.
A further reduction from 15 N m to values as low
as 1 N m of Charpy V-notched impact energy
results if the graphite particles have a flake shape
rather than a spherical shape.44 Figure 21 is a direct
observation in the SEM of fracture of a ferritic grey
cast iron Charpy specimen. The blunt notch with
initial damage developing at the graphite flakes
adjacent to the notch is seen in Fig. 21a. Figure 21b
shows an enlarged view of this region. Note that all
a of the microcracking is at the graphite/ferrite
interface, with crack advance by coalescence of
these graphite induced microcracks. The graphite
flake shape in grey iron causes debonding to occur
at much lower local strain levels than those
f
20 SEM micrographs of ferritic ductile iron
photographed after crack growth in a straining
fixture (as polished, unetched); observed surface 21 SEM micrograph showing crack path in pearlitic
is normal to crack path and general direction of grey cast iron. Fracture path zone (1) in a is
cra'ck movement is from right to left (Ref. 43) shown magnified ,in b (Ref. 45)

23 SEM observation of fractured surface of ferritic

ductile iron compact tension specimen· in stable
crack growth region (Ref. 45)
random nature of fracture in these systems.

Cracking goes primarily through the graphite/ferrite
debonds, resulting in a large fraction of the frac-
tured surface as the graphite phase.
Effect of matrix
As the microstructure of cast iron is changed from a
soft, ductile" ferritic constituent to a stronger, but
more brittle constituent such as pearlite or bainite,
the fracture toughness is affected in two ways: (a)
the ductile to brittle transition temperature is
b sharply increased, and (b) the upper shelf fracture"
is reduced. These two factors in combination
22 SEM micrograph of grey cast iron showing
fractured surface, highlighted using glass filler.
significantly reduce the room temperature Charpy
Zone marked (1) in a is shown magnified in b impact strength of ductile iron, from 20 to 3 N m,
(Ref. 45) even though the upper shelf fracture toughness is
only reduced from 20 to 10 N m.
In the following sections, the fracture toughness
of cast irons is reviewed by commercial classes (i.e.
required to nucleate debonding in nodular cast grey iron, malleable iron, ductile iron, compacted
iron, thus giving a much lower fracture toughness. iron, and white iron). A comparison of the fracture
Figure 22 is a photomicrograph of a fractured piece toughness of the ferritic and pearlitic grades of
of grey iron. The polished plane observed in this these various classes of cast iron gives a direct
figure is perpendicular to the fracture surface, indication of the effect of graphite particle shape on
allowing the crack path to be delineated. Again, it fracture toughness, since each class is distinguished
is clearly seen that fracture begins with failure at by its distinctive graphite particle morphology
the graphite/ferrite interfaces and proceeds with (except for white cast iron which has no graphite,
coalescence of this damage by cracking across the only cementite plus ferrite).
matrix bridges between adjacent graphite flakes.
For better clarity, the fracture zone marked (1) in
Fig. 22a is shown at a higher magnification in Fracture toughness of cast irons
Fig. 22b.
The fracture surface of a ductile iron specimen is Grey cast iron
shown in Fig. 23. The volume fraction of graphite Dynamic fracture toughness of grey cast iron
in this specimen is --12%
, but the areal fraction of Grey cast irons (with flake graphite) have Charpy
voids including the spaces left by deposed graphite V-notch upper shelf energy values of 1·4-6·8 N m
nodules ·is greater than 500/0, indicating the non- (Refs. 46-49) with higher matrix strength grey cast
irons usually giving higher values for the upper

shelf toughness, whether measured by Charpy or Pmax
fracture mechanics tests. This unusual behaviour

10
(most ferrous alloys h~ve a decreasing upper shelf
fracture toughness with increasing strength) is a Po
result of the fact that fracture occurs through the /
/
graphite flakes with very little matrix deformation. /
It appears that a stronger matrix is more resistant to ./ \
\
/ \
the deformation that must accompany micro crack ~5% Secant Offset \
\
formation by debonding at the graphite/matrix I
'I \
\
interface or by graphite cracking. It should be r/ \
\
'I
added that no apparent ductile to brittle transition f \
\
temperature is noted in grey cast iron because the \
o
fracture process dissipates so little energy via matrix 0.0 0.2 0.4 0.6
deformation, even at temperatures where the Load-line displacement, mm
matrix is ductile. 24 Load-displacement curve during fracture test of

The deleterious effect of phosphorus on the compact tension specimen of grey cast iron
fracture toughness noted above is seen in the results (Ref. 32) .
of Gilbert42 in Table 3, where an increase in
phosphorus concentration from 0·056 to 0·77%
resulted in a decrease in the room temperature
Charpy energy from 16·8 to 8·9 N m. Furthermore, K1c lies between 19 and 40 MN m-3/2. Determin-
the table also indicates the beneficial effect of a ation of the exact moment of crack extension is
higher strength matrix on the impact toughness of necessary to specify further K1c for grey cast iron.
grey cast iron, with the bainitic matrix (acicular cast Owing to the non-linear behavior of grey cast
iron) giving over three times the impact strength of iron caused by the extension and contraction of
a ferritic matrix (undercooled graphite cast iron). It graphite flakes during load-unload cycles,
should be noted that the tests results in Table 3 compliance measurements cannot be used to
were measured using oversized unnotched bars determine the moment of crack extension. Multiple
(16 x 16 x 126 mm), which explains the higher specimen testing as per ASTM E813-81 can also be
than expected impact energy values observed. somewhat ambiguous in that distinguishing between
Flake graphite is very deleterious to the upper fatigue cracking and crack extension under mono-
shelf fracture toughness (as well as strength) tonic loading during the fracture mechanics tests is
because the flakes act like internal cracks, due to not straightforward in grey cast iron; neither is heat
the ease with which they debond from the matrix as tinting. Therefore, Bradley32 used a combination of
well as their poor intrinsic strength. Not all graphite heat tinting and dye penetrant analysis and
structures in Fig. 7 are equally damaging to the determined for the class 35 grey iron, whose
upper. shelf fracture toughness. Medium size Type load-displacement curve is shown in Fig. 24, that
A ~raphite is preferred to Type D or coarse Type crack advance occurred at approximately Ko so
A. 0 A finer eutectic cell size and larger amounts of K1c = 19 MN m-3/2.
that Ko ::::::
transformed primary austenitic dendrites improve The values of fracture toughness reported in the
the toughness.51 literature for grey cast iron almost always give
either Ko or Kmax, rather than K1c, because of the
Quasistatic fracture toughness of grey cast iron ambiguity in determining accurately the moment of
Several attempts have been made to measure the crack extension. A question then arises as to which
fracture toughness· of grey iron using a fracture of the two is the better measure of the fracture
mechanics approach. 32,42,49,52-59 A load- toughness. Bradley observed more than 1·6 mm of
displacement curve taken by Bradley32 on a IT crack advance before reaching maximum load in
compact tension specimen in an attempt to measure class 35 grey iron.32 In view of this result and the
K1c for a class 35 grey iron is presented in Fig. 24. discussion presented above in this paper, it is
This load-displacement record is fairly typical of considered that the Ko values in the· literature are
results obtained for all grades of grey cast iron. reasonable and conservative estimates of K1c. Kmax
One may measure to Po and Pmax in Fig. 24 and values would. tend to overestimate the fracture
use these· in the standard LEFM relationships in toughness of grey cast iron.
ASTM E399 to calaculate Ko and Kmax values of 19 All of the results from the literature32,42,49,52-59
and 24 MN m -3/2, respectively. If one calculates for Ko and Kmax for grey cast iron are presented in
Imax, or J at P max, and then calculates KJ , a value Fig. 25 as a function of UTS. The increase in
of 40 MN m -3/2 is obtained. Clearly, K;:xhas no toughness with increasing strength previously
physical meaning when taken from a load- observed for dynamic fracture in grey cast iron is
displacement curve such as the one given in Fig. 24. also seen to be true for the quasistatic fracture
By the time the maximum load has been reached in toughness of grey cast iron. Higher strength in turn
the fracture toughness test of grey cast iron (e;g. is obtained by strengthening the matrix,52,53 by
Fig. 24), the material behaviour is so non-linear reducing the eutectic cell size,52,53 or by reducing
that using P max in an LEFM formula makes no the amount of graphite by reducing the carbon
physical sense. One may conclude, therefore, that content.42 It is interesting to note that in the work
o Ka (Grover and POllard)55 ° Kmax (Tait and Spencer)56 200

c
6 Kmax (Venkatasubramanian and • Ka (Tait and Spencer)56
180
Baker-SENB specimens)52
* KJC (Bradley)32 E c
• Kmax (Venkatasubramanian and Z 160
Baker-SENT specimens)52
40 ~~
Q)
140
c 120
N w
M 34
IE ° rn 100
Q)
z r- 80
~_28 ° ()
x
ca
E E
co
60 c c
~ 22
(5 ~ 40
a o o Ferritic malleable iron
~ 16 20 c Fully ferritized nodular iron
B
EI o
10 200 220 240 260 280 300 320 340 360 380
100 140 180 220 260 300
Ultimate Tensile Strength J MN m-2 Temperature, K
25 Relations'hip between tensile strength and frac- 26 Dynamic tear energy v. test temperature for
ture toughness parameters Ka and Kmax for grey ferritic malleable iron and ferritic ductile iron
cast iron (Refs. 32, ~2, 49, 52-59) (Ref. 65)
of Venkatasubramanian and Baker,52 the Kmax gives them a lower yield strength. In Fig. 26 the
values were much lower for single edge notched energy absorbed in a dynamic tear test on ferritic
tensile (SENT) specimens than for single edge malleable cast iron is compared with dynamic tear
notched bend (SENB) specimens. As shown in Fig. results on nodular cast iron.65 The nil ductility
24, the Kmax values in SENT specimens approach temperature (NDT) for ferritic malleable iron is
the KQ values. On the other hand, Kmax values in seen to be -96°C whereas ferritic nodular cast iron
SENB specimens are much higher than the KQ is seen to have a NDT of -56°C.
values. This difference between the two types of Some fracture mechanics results have been
specimens is mainly beause SENB specimens are published on malleable iron based on LEFM
more stable than the SENT specimens, particularly considerations. Either Kmax or Ko values have been
when the former are tested under displacement reported65-67 and these values indicate that upper
control. shelf fracture toughness actually increases with both
increasing strength and/or lowering of the tem-
Malleable cast iron perature of the test. While this may be a reasonable
The graphite phase in malleable cast iron is called correlation for grey cast iron, it is not so for
tempered graphite since it is formed by tempering malleable cast iron. The reason for this may be best
white cast iron. The shape of the graphite is seen in understood by examining the, actual load-
Figs. 4 (Type III), 5, and 12 to be irregular displacement data taken on compact tension speci-
glo bular , or pop corn shaped. As expected this mens of two grades of malleable cast irons, M3210
form of graphite is less damaging to the upper shelf and M5503, as presented in Fig. 27. (The former
fracture toughness than is the flake form. iron has a minimum yield strength of 220 MN m-2
Ferritic malleable cast iron has an upper shelf and a minimum elongation of 10%, while the
Charpy energy of 13 N m compared with 9 N m for corresponding values for the latter are 380 MN m-2
pearlitic malleable cast iron.6o Other investiga- and 3%, respectively.) The (room temperature)
61
tors ,62 have reported a lower transition temper- load-displacement curves are clearly seen to be
ature and a somewhat higher upper shelf Charpy quite non-linear. The initial non-linearity in both
eQergy of 19-20 N m for ferritic malleable cast iron curves is associated with plastic deformation rather
and 12 N m for pearlitic malleable cast iron. than crack extension, and thus, the higher yield
Shulte61 has also presented Charpy V-notch data strength material (M5503) is seen to have a higher
showing the dramatic reduction in upper shelf Po value. It has recently been determined by
toughness and increase in the ductile to brittle Bradley,45 using a 30% unload compliance tech-
transition temperature that result from phosphorus nique to monitor crack growth, that crack ,growth
concentations of 0·15% or greater in both ferritic begins at or just before maximum load in fracture
and pearlitic malleable cast iron. Similar results mechanics tests of IT compact tension specimens of
have been reported by others.63,64 The phosphorus malleable and nodular cast iron with a/W = 0·06.
not only forms a very low melting. point phosphide Thus, an appropriate fracture mechanism analysis
but it also degenerates the graphite particle shape. of test results in Fig. 27 would be to use a I-integral
Malleable cast irons are generally regarded as approach to calculate the critical 1. value at or just
having the best low temperature properties of any before maximum load where crack extension
of the cast iron family, having a much lower ductile begins. This 'procedure leads to to a proper assess-
to brittle transition temperature than nodular cast ment of the fracture toughness of malleable iron,
iron because they have about 1% less silicon, which see below.

30 80
T=297 K
Pmax
20
z
~~ o
"0 ~ Km~ ----~
co
0
-J o ~
..,....,.,"""""
10
.,.........,""'"
..,....,.,
Ka ..,....,.,"""""
~~~
..,....,., .
..,....,.,
~~..,....,.,
20
0.2 0.3 0.4 0.5 0.6 0.7
250 300 350 400 450 500 550
Load-line Displacement, mm Yield Strength. MN m-2
27 Load-displacement curves from fracture
28 ~ariation of critical stress intensity KJc with
mechanics tests on compact tension specimens
Yield strength for malleable iron (Ref. 45)
of M3210 and M5503 malleable iron: A, and A2
refer to the area under the load-displacement
curve up to Pmax for M5503 and M3210 spe-
cimens, respectively (Ref. 32)
Very little work has been done to determine the
effect of nodule count or nodularity on the upper
shelf fracture toughness or the ductile to brittle
The earlier work of Bradley32 purported to give transition temperature of malleable iron.
Moore 72,73 found no effect of nodule count on the
llc results for maIlable iron. Unfortunately, Bradley
used the 10% unload-reload compliance approach ductile to brittle transition temperature of ferritic
recommended in the tentative ASTM E813 stan- malleable iron. Askeland and Fleischman,74 using
dard .for llc determination published in 1977 (the unnotched Charpy specimens, found an increase in
current standard allows up to 40% unload). Sub- impact energy of about 10% at room temperature
sequently, work by Salzbrenner et al.68 and for an increase in nodule count from 30 to
Bradley45 indicated that at least a 20% unload is 300 mm -2. It will be seen below that with ductile
necessary in unload compliance measurements of iron a high nodule count tends to give a somewhat
crack length in cast iron to avoid a premature lower value for KJc on the upper shelf and a slightly
indication of crack extension because of time lower ductile to brittle transition temperature. It is
dependent plastic deformation that occurs in cast believed that the same trends should be true in
irons, both during unloading and reloading within malleable iron, though there are insufficient data at
5% of the previous monotonic load-displacement present to draw such a conclusion.
curve. Thus, the llc values reported by Bradley and In summary, ferritic malleable iron can have
Mead in Refs. 32, 69-71 underestimate the true llc
values for all of the materials studied except for
grey iron which was characterised using a multiple • Sidegrooved, Fatigue Precracked Charpy
specimen approach. A reanalysis of the data in • V-Notched Charpy
Refs. 45 and 69-71 has been made, with the ~ o V-Notched 3/8" Square
assumption (based on the additional experimental 300
work cited in Refs. 45 and 68) that crack extension ~
begins very near the maximum load in the test in ~ 280
malleable and ductile iron compact tension speci- Q)
Q.
mens (a/W = 0,6). The results are presented in E 260

Q)
Table 4 and in Fig. 28 where KJrnax values (assumed ~

c
to equal KJc) are plotted as a function of yield
2·00240
strength ~or ~he same malleable irons previously c
reported In FIg. 27. For comparison, KQ and Kmax ~ 220
o
are also presented in Fig. 28. The very reasonable ~
13
result of a monotonically decreasing' KJc value with ~200
~ncreasing yield. strength (and decreasing ductility) E
IS observed, whIle Ko and Kmax both underestimate 180 • 0
the real K1c value and wrongly imply that a higher
strength matrix in malleable iron gives a better 100 140 180 220 260 300
upper shelf fracture toughness. These results . Hardness. HB
illustrate why the use of Ko or Kmax to give even a 29 Variation of impact transition temperature with
toughness based ranking for malleable iron is hardness (and therefore strength) of malleable
inadvisable. iron (Ref. 26)

Table 4 KJ values recalculated from Ref. 32 and figure that debonding at the graphite/matrix inter-
assumed to equal KJc face followed by link up of microcracks through
KJc the. matrix provides the path for crack propagation
KJmax (previously in ductile iron.
Sy,t (new), reported), One would expect ferritic ductile iron to have a
Material* MN m-2 Temp., °C MN m-3/2 MN m-3/2
somewhat better upper shelf fracture toughness
M3210 230 24 76 44 than does ferritic malleable iron because the spher-
(ferritic malleable -19 76 41
ical graphite nodule should extract a smaller pen-
iron) -57 71 43
alty in ductile fracture toughness than would the
M4504 359 24 70 55
(ferritic/pearlitic -19 72 48
somewhat irregular, globular graphite nodule shape
malleable iron) -59 30 30 of ferritic malleable iron. A spherical particle shape
M5503 411 24 57 45 will have the minimum interfacial stress for a given
(tempered martensitic -18 53 52 nominal stress before interfacial failure which
malleable iron) -58 34 29 initiates ductile rupture by void coalescence.75
M7002 540 24 50 50 Since the growth and coalescence of voids involves
(tempered martensitic -19 39 39 a plastic deformation process, it is to be expected
malleable iron) -58 39 39
that the fracture energy is related to the size of the
8(}-6Q-03 433 24 33 27
(pearlitic/ferritic -19 25 25
dimples thus formed.76 Ferritic ductile iron has an
ductile iron) -48 26 26 upper shelf Charpy V-notched toughness of
DO & T:t: 717 24 48 52 14-24 N m (Ref. 77) compared with ferritic mal-
(tempered martensitic -19 51 48 leable iron with a range of 12-19 N m and grey iron
ductile iron) -59 50 51 with a range of 1-6 N m. The upper shelf dynamic
D5B:t: 304 24 89 64 tear impact energy for standard 15·9 mm DT
(austenitic ductile -47 85
iron) -59 91
specimens of ferritic ductile iron are 175 N m
67
compared with 100 N m' for ferritic malleable
* Material identification
t Yield strength.
numbers are described in the text.
iron.66 Upper shelf KJc values in excess of
:t: SAE designations. 100 MN m -3/2 are possible for ferritic ductile iron 78
compared with 75 MN m -3/2 for ferritic malleable
iron. As discussed above, malleable iron will
upper shelf Charpy impact values of 19 N m and typically have a lower transition temperature than
KJc values of 75 MN m -3/2. These very good ductile iron because of its lower silicon content and
fracture toughness properties are retained down to the resultant lower strength.
extraordinarily low service temperatures, as shown
in Fig. 29, where the transition temperature meas- Upper shelf, quasistatic fracture toughness
ured on different types of specimens (side grooved of ferritic ductile iron
fatigue precracked Charpy, V-notched Charpy, and The values of upper shelf fracture toughne~s K1c of
V-notched 10 mm square) are ~resented as a ferritic ductile cast iron reported in the literature
function of hardness (or strength). 6 Increasing the have shown a systematic increase over time that
strength of malleable iron by increasing the strength might seem to have resulted from increasing the
of the matrix reduces the upper shelf fracture quality of this cast material. In reality, the early
toughness and increases the transition temperature. measurements of ambient temperature fracture
As an example, increasing the yield strength of toughness of ferritic ductile iron were almost all
malleable iron from 220MN m-2 minimum (fully incorrect for reasons explained below. Over the
ferritic) to 380 MN m -2 minimum (quenched and past several years, a more consistent pattern has
tempered) increases the NDT temperature .from emerged which has allowed a clear definition of the
177 to 211 K (Refs. 65, 66), decreases the upper upper shelf fracture toughness of ferritic ductile
shelf dynamic tear energy from 100 to 80 N m, iron to be established.
decreases the upper shelf Charpy values from 16 to
12 N m (Fig. 26),60 and decreases KJc from 73 to Fracture toughness measures of K,c:
53 MN m-3/2 (Fig. 27 and Table 4). LEFM approaches
Nanstad et al.78,79 were among the first to study the
fracture toughness of ductile iron using a LEFM
Ductile iron (also known as nodular iron approach (1974). Using IT or smaller compact
or spheroidal graphite cast iron) tension specimens, they measured almost
A much greater. amount of research has been exclusively KQ values except for pearlitic iron at
conducted to determine the fracture toughness of very low temperature. Interpreting KQ as a fracture
cast iron with spherical nodules of graphite (ductile toughness parameter rather than as an early indi-
iron) than for cast iron with graphite flakes (grey cation of crack tip plasticity, they concluded that
iron) or irregular globules of graphite (malleable the fracture toughness of ferritic and pearlitic irons
iron). Types I and II graphite (ASTM A are 247)6 were similar, the fracture toughness of ductile iron
found in well made ductile iron with degenerate increased with thickness (unlike other materials),
Types V and VI sometimes observed (Fig. 6). The and that vermicular graphite gave a similar fracture
significant role graphite plays in determining crack toughness to high nodularity ductile iron, with a KQ
growth in cast irons has already been discussed with value of about 40 MN m -3/2 indicated for high
reference to Fig. 20. It may be recalled from this nodularity, ferritic ductile iron. In "reality, their
results are related to yielding rather than the Table 5 Fracture toughness measurements of
fracture toughness of ductile iron. These results ferritic ductile iron
indicate that permanent deformation (yielding) in Specimen size Temp., Ka, Kmax,
ductile iron is not a sensitive function of nodularity Ref(s). and type* K MN m-3/2 MN m-3/2
(though fracture in ferritic ductile iron is) and that 65 1T-CT 298 65
yielding is more difficult in thicker specimens 80 4T-CT 298 82
because they have greater constraint.32 The primary 81,82 Double torsion 298 64-74
(not fatigue precrack)
lesson from this study is that LEFM are not very
78, 79 1T-CT 40
well-suited for measuring the upper shelf fracture 83 6T or 8T-CT 233-298 87-93 115-135
toughness in ferritic ductile iron unless very large
* CT compact tension.
specimens are used.
In a subsequent study of ferritic ductile iron
conducted by Worzala et al., 65 IT compact tension with larger specimens gIVIng larger values of Ko
specimens were used, and again, no valid K1e and Kmax as one would expect. For smaller compact
values were obtained. However, a Kmax value of tension specimens (i.e. IT), significant non-linearity
65 MN m - 3/2 was reported. Ostensson, 80 using 4T in the load-displacement curve is observed and
compact tension specimens, measured a Ko value crack growth begins at (or near) maximum load in
of 82 MN m-3/2 in a ductile cast iron with a matrix the test. Therefore, Kmax is much less than K j
consisting of 92°10 ferrite and 8°1o pearlite. Here (and thus, K1e) for IT specimens. max
again valid K1e, based on LEFM, was not obtained. For much larger specimens (i.e. 6T), a relatively
Ten Haagen and Berry81,82 attempted to measure small amount of non-linearity is observed in the
the upper shelf fracture toughness of ferritic ductile loa~-displacement curve, crack growth begins well
iron using the double torsion method with a linear below P max, and Kmax is an upper bound for K1e•
elastic analysis. They measured Ko values of 64-74 Thus, Kle should be bounded by Ko and Kmax
MN m -3/2, which are higher than _the Ko values values that have been measured on sufficiently large
measured by Nanstad et al.78,79 undoubtedly specimens so as to allow only limited non-linearity.
because they did not fatigue precrack their speci- Thus, the results presented in Table 5 would
mens. Neither specimen size nor root radius of their suggest an ambient temperature K Ie value between
machined in notches were indicated in their paper. 87 and 115 MN m-3/2• Later results using elastic-
Motz83 using 6T and 8T compact tension speci- plastic fracture mechanics discussed in the next
mens of ferritic ductile iron measured Ko and Kmax section confirmed this observation.
values over the temperature range -40 to 20°C,
with K1e values measured at lower temperatures. Fracture toughness measurements using CTOD:
Selected results from Motz83 are presented in Fig. an elastic-plastic fracture mechanics approach
30. Of particular interest are the ambient and near Recognising the inherent elastic-plastic nature of-
ambient temperature results reported for Ko and the fracture behaviour in ferritic ductile iron on the
Kmax• Over the temperature range -40 to 22°C, Ko upper shelf, several groups of investigators have
values of 87-93MN m-3/2 are noted with measured attempted to measure the fracture toughness using
Kmax values of 115-135 MN m-3/2. Only at tem- CTOD.84-89 The principal results of these studies
peratures of -60°C or lower were K1e values are presented in Figs. 31-33. Figure 31 shows the
measured. variation of critical crack opening displacement as a
The results of these five studies which attempted function of the yield strength for ferritic ductile iron
to measure K1c for ferritic ductile iron at ambient samples as obtained by Luyendijk and Nieswaag.89
temperature are summarised in Table 5. The These investigators have compared the CTOD at
geomery dependence of Ko and Kmax are noted, initiation of stable crack (Fig. 31a) and at maximum
load (Fig. 31b) obtained using DD 19 and BS 5762
approaches, as shown in Fig. 31. Figure 32 shows
the variation of critical crack opening displacement
~
(")
140
as a function of temperature for ferritic ductile iron
0 0
IE 0 samples as obtained by Holdsworth and Jolley. 84
120
z The effects of nodule count (Fig. 32a) and carbon-
2: content (Fig. 32b) on the relationship between
en 100
en
<D ~ ()
() CTOD and temperature were studied by these
c 80 0 ()
investigators. The effect of specimen size on the
.c
0>
:3 relationship between CTOD and temperature for
0 ~~
•....
e
60
~
(J:)
~
., • ferritic ductile iron samples as studied by Niskanen
t>m
:3 40
• -i o Kmax
is shown in Fig. 33. Results on Iron 3 from
Holdsworth and Jolley84-86 (Fig. 322 may be
u: 20
()Ka
• K,c
compared with the results of Niskanen 7 (Fig. 33)
and Luyendijk and Nieswaag88,89 (Fig. 31), since
0 these studies were conducted principally on ferritic
-140 -120 -100 -80 -60 -40 -20 0 20
ductile iron with nominally the same silicon content
Test Temperature, C
0
(2.4°10) and 0.2°10 offset yield strength
30 Fracture toughness v. temperature for various (260 MN m-2). Room temperature results from
grades of ductile iron used by Motz (Ref. 83) these three studies are compared in Table 6, and
0.20..-------------------,
(0) Iron 1
(a) At initiation of stable crack growth
0.15
ACTOD using
DO 19 Material Nodule Count
0.15 - Iron 1 37 mrri2
• CTOD using Iron 2 107 mrri2
A BS 5762 0.10
E Iron 3 52 mrri2
A
E E
00.10
0
- • A
E
limit of fibrous crack growth
C 0.05
l-
0 • Q)
E
0.05~
• Q)
u
• CO
Q.
CJ)
o
-200 -150 -100 -50 0
(5 Temperature, 0 C
I I I • 0)
c
c 0.20
(b) At maximum load Q)
c-
(b)
•• ••
ACTOD using O Material Carbon
0.20~
A
DO 19 ~u
0.15 Iron 3
Content
3.63%
••• Iron6
Iron3
co
•
• CTOD using
BS 5762 o Iron
Iron
4
5
2.58%
3.99%
Iron5
E mu Iron 6 3.23%
E :;:;
A 0.10
0' 0.15- A
(5
o
I-
• A
o • • 0.05
limit of fibrous crack growth
0.10 - •
o
I I I" I -200 -150 ~100 -50 o
0.05
200 250 300 350 400 450 Temperature, 0C
Yield Strength MN m-2
J
32 Crack tip opening displacement v. temperature
31 Yield strength v. crack tip opening displacement for ferritic ductile irons in which a nodule count
for a crack initiation and b unstable crack and b carbon content (and therefore, volume
growth (at maximum load) (Refs. 88, 89) fraction of graphite) have been systematically
varied (Ref. 84)
serious discrepancies are noted (e.g. Dc values specimens will experience crack growth well before
reported at room temperature range from 0·09 to maximum load is reached.9o By taking note of the
0·43 mm). above factors, the wide variations in Dc presented in
The variations in the calculated values for CTOD Table 6 can be reconciled, see below.
summarised in Table 6 may be attributed to three The significant size effect of the CTOD values
factors: (a) whether DD 19 or BS 5762 was used to measured at maximum load as seen in Table 6
calculate the critical CTOD from the measured and Fig. 33 needs to be addressed. Cayard90 has
mouth opening displacements, (b) what value of r is recently tested fatigue precracked Charpy bars and
assumed (or what location is assumed for the plastic bend bars that were 25 x 19 x 62·5 mm and has
hinge point), and (c) how accurately the moment of determined by compliance measurements that crack
crack extension is identified. With regard to the growth begins at or just before maximum load.
third factor, the specimen size is critical. Smaller Luyendijk and Nieswaag88,89 measured a critical
specimens (10 x 20 x 50 mm) will have crack CTOD for crack initiation that was, on an average,
growth beginning at maximum load whereas larger only about 65% of CTOD at maximum load on
their 40 x 20 x 100 mm sized specimens as can be
seen by comparing the data in Fig. 31a and b.
Table 6 Critical crack tip opening displacement for
Niskanen87 indicated larger values for CTOD meas-
ferritic ductile iron fracture at room tem- ured at maximum load on 150 x 75 x 375 mm
perature* specImens than on 50 x 25 x 125 mm specimens,
which in turn had a larger value than the results
r value Standard Specimen
measured on 20 x 10 x 50 mm specimens. In
Ref(s). used used size, mm Dc,mm Dmax,mm
combination these results clearly indicate that fer-
88,89 0·48 DD19 20 x 40 x 100 0·16 0·20 riticductile iron specimens much larger than
88,89 0·48 BS 5762 20 x 40 x 100 0·11 0·17
87 0·33 DD19 75 x 150 x 375 0·43
20 x 10 x 50 mm will exhibit some stable crack
25 x 50 x 125 0·11 extension before maximum load. Since the amount
·10 x 20 x 50 (0·11)t 0·09 (0·15)* of crack extension before maximum load is geome-
85,86 0·48 DD19 10 x 20 x 50 (0·11)t 0·15 try dependent, meaningful geometry independent
* Yield strength 260 MN m-2• measurements of fracture toughness in ferritic
t Recalculated using r = 0·48 rather than r = 0·33 originally assumed ductile iron cannot be obtained from the CTOD
in Ref. 87.
+ Recalculated using BS 5762 rather than DD 19 and assuming
values at maximum load on specimens larger than
Dc = Dmax for this size specimen. about 20 x 10 x 50 mm. With larger specimens, a
where
a effective crack length
~ 0.4 E elastic modulus
K stress intensity factor .
+-'
,C r plastic rotational factor (= 0·48 in
Q)
BS 5762)
E Sy yield stress
Q)
u 0.3 Vp plastic component of clip gauge
ctS
Q. displacement
CJ)
VV testpiece width
o z distance of clip gauge location from
0)
c testpiece surface
c 0.2 b critical crack opening displacement
Q)
Q. v Poisson's ratio.
o
.:::t::. Table 6 indicates the significant magnitude of the
U
ctS discrepancy between CTOD values calculated from
~
o the same experimental data using DD 19 and
ctS 0.1 BS 5762 (equations (5a) and (5b)), particularly
u at low values of CTOD. For example, for a
10x20x50 mm
+-'
.;:: 60 MN m-2 yield strength, be (at Pmax) = 0·14 and
o 0·11 mm according to DD 19 and BS 5762,
respectively.
The geometry dependence of be (at Pmax) is
200 300 clearly seen in Table 6. Based on the discussion
Temperature, K above, Pmax would only correspond t.o crack ~n-
itiation in the 10 x 20 x 50 mm speCImens, WIth
33 Variation of crack tip opening displacement v.
crack growth beginning before crack advance in the
temperature for ferritic ductile iron tested with
several sizes of three point bend bars (Ref. 87) larger specimens. This is why the CTOD (at P max)
varies significantly with specimen size as well as
temperature. If the 10 x 20 x 50 mm specimen
does exhibit crack growth beginning at maximum
compliance or potential drop metho~ .s~ould be
load, the room temperature value of 0·09 mm
used to determine the moment of InItIal crack
(calculated using equation (5a) and assuming r =
extension. Otherwise, one may greatly overestimate
0·33) will give a DD 19 CTOD value of 0·153-
the crack initiation CTOD for ferritic ductile iron.
0·158 mm for r = 0·48. This DD 19 (r = 0·48)
Holdsworth and Jolley's CTOD calculated from
value of ---0·145 mm can be seen in Luyendijk and
measured mouth opening displacement at P max (see
Nieswaag's work to correspond to a BS 5762 be
Fig. 32) may now be reconci~ed with the fin~~ngs of value of 0·11 mm.
Luyenkijk and Nieswaag (FIg. 31). Luye~d~J~ ~nd In summary, the results from Figs. 31~33, ~s
Nieswaag using DD 19 indIcate that crack InItIatIon summarised for room temperature behaVIour In
occurs at a CTOD value of 0·145 mm with unstable
Table 6, have been interpreted to give a consistent
crack growth at 0·19 mm for their 40 x 20 x
initiation be value of 0·11 mm, which can be
100 mm specimens. Holdsworth and Jolley, also
obtained from all the three studies when a proper r
using DD 19, found a critical CTOD at max~mum
value (0·48) is used at true initial crack extension
load for their iron no. 3 to be 0·148 mm, conSIstent
(maximum load in small specimens and load for
with the idea that the CTOD at maximum load
initiation determined by potential drop or unload
gives a reasonable estimate of the initiation fracture
compliance otherwise) to calculate the CTOD from
toughness in 20 x 10 x 50 mm specimens. Holds-
the measured mouth opening displacement.
worth and Jolley reported a lower CTO D initiation
The critical CTOD values can be related to
value of 0·05 mm which is considered unreliable
critical I-integral values using the relationship
because their r value varied from 0·1 to 0·48 over
this interval. Ideally, r should be a constant. Their lIe = mSybe (6)
apparent variation in r is associated with the
where m, the constraint constant, has been
assumption implicit in DD 19 that all mouth open
determined to be 1·6 for ferritic ductile iron.66
displacement is plastic, when in reality, very little is
Combining equation (2) with the standard
plastic in the initial load up.
relationship for converting lIe into an equivalent
The equations for b in DD 19 and BS 5762 are
KJe value using equation (3) cited above (i.e.
given by
, KJe = (IteE)OOS) allows the conversion. of the
b = r(VV - a) (Sa) CTOD value of 0·11 mm into an equivalent KJe
DD 19:
r(VV ..,- a) + a + z value of 88 MN m -3/2, which is remarkably
consistent with the results inferred from LEFM
BS 5762:
b = K2(1 - v2) + 0·4(VV - a)Vp (87 MN m-3/2 < Klc < 115 MN m-3/2) and with
, 2SyE 0·4VV + 0·6a + z results of 90 MN m-3/2 determined using a I-integral
(5b) approach reported recently. 68,90
Fracture mechanics measurements of J1c: Table 7 KJmax values recalculated from Refs. 69-71
an elastic-plastic fracture mechanics approach and assumed to equal KJc for various
The first attempt to measure the upper shelf ductile irons arranged in increasing order of
fracture toughness of ferritic ductile iron was made yield strength Sy .
by Bradley and Mead69.7o in 1979. Using the KJc (previously
tentative ASTM standard for I-integral (which Sy, K (new),
Jmax reported),
2 MN m-
Item no. MN m- Temp.,oC MN m-
3/2 3/2
became ASTM E813-81), they ran single specimen
tests on compact tension specimens, using a 10% 269 24 98 45
unload compliance to monitor crack extension. -73 83 64
Owing to some limitations in their function gener- 2 283 -51 84 57
-73 89 59
ator described elsewhere,71 their actual compliance
3 317 24 80 52
measurements were made during the reload portion -51 98-112 65
of the unload-reload· cycle over a load range of -73 49 54
3-60/0 below the previous maximum load. As P max 4 331 24 81 52
was approached, time dependent inelastic deforma- -12 81 67
tion gave a change in the reload compliance that -51 57 59
resulted in premature indications of crack extension 5 345 24 97 55
and gave KJc values of only 45-55 MN m-3/2. -12 88 67
-51 49 54
Bradley32 also used the tentative ASTM standard
6 352 24 88 58
noted above and obtained KJc values in the range of -12 96-105 75
25-67 MN m -3/2 for grades of ductile iron listed -51 48 40
in Table 4. (The ASTM nomenclature 80-60-03 has 7 386 24 77 54
been described above: DQ&T and D5-B are the -12 70-77 55
designations of the Society of Automotive Engin- -51 46 50
eers (SAE) for tempered martensitic and austenitic 8 483 24 80 53
-12 48 55
ductile irons.)
. Subsequently, work by Salzbrenner et· al. b~ in
1983 (also using a single· specimen approach with
multiple unload compliance, but run with 20% shown that they can be more properly interpreted
unload to avoid any misleading compliance meas- to indicate an upper shelf fracture toughness of
urements due to time dependent, inelastic defor- 88-90 MN m -3/2. The most recent measurements of
mation) indicated that crack extension occurred I1c confirm this conclusion.77,90.97
approximately at maximum load for IT compact Relationships from the literature for lower
tension specimens of ferritic ductile iron. These ex- strength steels which correlate Charpy impact
periments gave the upper shelf fracture toughness of energy and K Ic would suggest that the fracture
90 MN m-3/2 alluded to above. Furthermore, more toughness of ductile iron should be less than
recent work by McKinney and co-workers,91-93 40 MN m -3/2, or alternatively, the impact energy
Bradley and Andler, 27,94 Tanner and Bradley, 33,95 should be closer to 50 N m (relationship in Ref. 98
Bradley,96 Cayard and Bradley, 90,97and Salzbren- used in conjunction with data in Ref. 94). In fact, it
ner77 have confirmed .the finding that the upper is the relatively low value of Charpy impact energy
shelf fracture toughness of ferritic ductile iron with compared with cast steel (16 N m compared with
'"'--'2'2%Si and with a nodule count of '"'--'60 mm-2 is 80 N m) that has caused engineers to view .ductile
around 90 MN m -3/2. A reanalysis of the data of iron as a low toughness material unsuited fof the
Bradley and Mead69-71 assuming that crack growth substitution of cast steel. In the next section work
in their IT compact tension specimens began at conducted to address and answer this question is
maximum load further supports the contention that reviewed.
KJc for ferritic ductile iron is '"'--'90MN m -3/2 (Table
7). The summary of reanalyses carried out for Reconciling low Charpy impact energy with
different malleable and ductile iron samples, as high KJc values of ductile cast iron
presented in Tables 4 and 7, further indicates that, Bradley and co-workers91-93.96 conducted a com-
in general, K values are higher than KJc values . parative study of the fracture toughness of ferritic
Jrnax
calculated previously, 32,69-71 but the difference ductile iron to a cast steel with very similar tensile
tends to get reduced as the yield strength increases properties; namely, ASTM A216-82.99 This study
and at lower temperatures. This is to be expected in was conducted using IT compact tension specimens
view of the reduction in crack tip plasticity in the andCharpy specimens. The Charpy specimens were
latter cases. prepared in several ways: with the standard V-
notch, V-notch plus fatigue precrack, V-notch plus
Summary of upper shelf, quasistatic fracture side grooving, and V-notch with fatigue precrack
toughness results for ferritic ductile iron and side grooving. The compact tension specimens
In summary, early indications of the upper shelf, were fractured quasistatically on a tensile testing
quasistatic fracture toughness of ferritic ductile iron (MTS) machine whereas the Charpy specimens
ranged from 40 to 185 MN m-3/2, the largest values were fractured either quasistatically on an MTS
implied from the reported CTa D ·measurements machine or dynamically on an instrumented Charpy
and the smallest values from LEFM values of KQ. machine. The critical value of K as predicted from
However, a careful reanalysis of this data has I is represented as KJd for dynamically fractured
200 in energy absorbed under this condition more

180 Cast Steel - S G & FPC accurately reflected the difference in the intrinsic
toughness. It may be noted that the cast steel
160 Fracture Toughness
C\(
....... samples considered in Fig. 34 are provided with
(")
140
IE side grooves with a view to suppress the formation
120
z of shear lips during fracture. This helps to make the
2: 100 o comparison between cast steel and ductile iron
~ 80 Ductile Iron - FPC more meaningful, as the latter does not experience
~ 60 shear lip formation.
e-.e e
A second surprising conclusion from the study of
40
Bradley et al.92 was that the transition temperature
20 for the ferritic ductile iron was much lower than for
200 220 240 260 280 300 320 340 360 380 400 420
the cast steel, resulting in ductile iron having a
Temperature, K
greater value of KJd than the ASTM A216 cast steel
from room temperature down to about -40°C (Fig.
34 Dynamic fracture toughness for ductile iron and 34). The strengthening of the ferritic ductile iron is
cast steel as determined from instrumented principally through solid solution strengthening by
Charpy impact tests (Ref. 92)
silicon, whereas the cast steel achieves a similar
strength via the presence of 20-25% pearlite, which
specimens. This study indicated that the Charpy is much more prone to nucleate cleavage cracks. In
V-notch energy value of a cast steel (ASTM summary, the fracture mechanics indication of the
A216-82) is about 5 times that of ductile iron with fracture toughness of ductile iron in relation to cast
similar tensile properties; however, the upper KJd steel is seen to be much more realistic than the
value of steel is only about 1·8 times greater (Fig. exaggerated difference indicated by the Charpy
34). Figure 35 helps to explain this apparent test. .
anomaly. It is seen that in the absence of a fatigue
precrack, the plastic zone size in cast steel is· much Experimental techniques for determination of
greater than in ductile iron, but when a sharp crack dynamic fracture toughness of ductile iron
is present the difference in plastic zone size is using fracture mechanics approach
greatly reduced, with most samples showing rela- Most of the attempts to measure the dynamic
tively small plastic zone sizes. This implies that fracture toughness of ductile iron have involved the
fatigue precracked cast steel and ductile iron use of standard Charpy, V-notched, fatigue pre-.
samples experience initial crack extension under cracked Charpy, or dynamic tear specimens. In what
condition of full constraint and thus, the difference follows, the appropriateness of each of these
methods to arrive at realistic values of dynamic
fracture toughness of ferritic ductile iron are
discussed.
Use of standard Charpy V-notched specimens
Kobayashi, based on his investigations, 100-102 has
Blunt notched Charpy specimen of cast steel
implied that· standard Charpy V-notched specimens
may be used to determine the fracture toughness of
grey and ductile iron. However, Caironi et al.103
have found that the maximum allowable root radius
to get meaningful J 1c values in ductile iron is
between 0·067 and 0·149 mm, well below the radius
--~~'------ of 0·25 mm provided in standard Charpy V-notched
Blunt notched specimen of ductile iron specimens. Furthermore, Andler94 has clearly
shown that fatigue precracked Charpy specimens
give very different initiation energies compared
with standard Charpy V-notched specimens. And,
as shown in Fig. 14, the fracture energy at a given
temperature is significantly higher for a standard
Fatigue precracked specimen of cast steel Charpy V-notched specimen than for a fatigue
precracked specimen.
It thus seems that the use of the standard Charpy
V-notched specimen is generally inappropriate to
determine the dynamic fracture toughness of ductile
__ A _ iron. As graphite debonding dominates the fracture
process in ductile iron, standard Charpy V-notched
Fatigue precracked specimen of ductile iron specimens may be acceptable only when the
35 Plastic zone size at initiation of cracking for internodular spacing is less than the standard notch
blunt notched (V-notched) and fatigue pre- root radius of 0·25 mm. This would require a
cracked Charpy specimens of ductile iron and nodule count of less than 20 mm -2; commercial
cast steel (as inferred from microhardness grades of ductile iron have nodule counts generally
testing) (Ref. 91) in excess of 20 mm - 2•
Use of dynamic tear specimens A few investigators maximum load in fatigue precracked Charpy speci-
have tried to obtain fracture mechanics type results mens" for ferritic ductile iron, giving an unambig-
from dynamic tear (DT) measurements taken with uous indication of the moment of crack extension.
a drop tower provided with an instrumented It is thus clear that the use of the fatigue
tup.33,104 The DT specimens had pressed in notches precracked Charpy specimens is best suited for
of 0·025 mm root radius. Bradley and Tanner33 . determining the dynamic fracture toughness of
compared the fracture behaviour of ferritic ductile ferritic ductile iron. In using this test, however, two
iron DT specimens with those of fatigue precracked factors merit attention. First, it is desirable to limit
Charpy and IT compact tensile specimens. They the impact velocity to 2 m S-1 rather than the use
found that stable crack growth occurred before of the standard Charpy impact velocity of 5·4 m S-l
maximum load in DT specimens (a12 = 0·31) as and employ an inertial correction factor, as orig-
opposed to stable crack growth occurring at inally suggested by Turner.105,106 Second, the load-
maximum load in fatigue precracked Charpy displacement record from an instrumented Charpy
(a12 = 0·6) and IT compact tensile (alw = 0·5) test should be corrected to take into account the
specimens. This resulted in KJrnax values being stiffness of the Charpy impact te'sting machine.107
significantly greater for the DT specimens than for Failure to observe these procedures may result in
the other two, as shown in Fig. 36. Kobayashi and the measured lower shelf fracture toughness
NishilOO have also observed a similar stress intensity exceeding the actual values for the material by
based on LEFM. But, as discussed in the section on 50-75% (Ref. 33).
the fracture toughness of grey cast iron above, Dynamic fracture toughness measurements on
stable crack growth occurs before maximum load in ductile iron The results obtained by several
grey cast iron and thus, Kmax does not provide a investigators on the dynamic fracture toughness of
meaningful measure of the fracture toughness of ductile iron are reviewed in the light" of the
grey cast iron. discussion on experimental needs in the previous
It may thus be concluded that the use of DT section. Most of the published dynamic fracture
specimens is also inappropriate to determine the studies on ductile iron have been done by three
dynamic fracture toughness of ferritic ductile iron groups: Kobayashi's group in Japan,46,10o-102 Wor-
and grey cast iron, since stable crack growth occurs zala, Heine, and Loper's group at the University of
in both these materials before maximum load in.a Wisconsin, 65,66,104,10~,109and Bradley's group at
DT test. At present, there is no reliable method Texas A&M University.27,91-96 Tuler et al.l1O have
available to measure the moment of crack extension also done a significant amount of work on the
in this test. dynamic fracture of ductile iron, most. of which has
Use of fatigue precracked CharfY specimens Brad- not been published.
ley and Tanner33 and Cayard9 have compared the The studies of Kobayashi's and Bradley's groups
KJc values for ferritic ductile iron obtained with IT have been made primarily on Charpy specimens.
compact tension specimens and fatigue precracked Kobayashi's use of P max in a LEFM analysis
Charpy specimens, both tested quasistatically. They resulted in an underestimation of KJd, giving a
found that the values compared reasonably well. As value of -----60 MN m -3/2 (Ref. 101). This value
discussed above, stable crack extension occurs at " would have been even lower but for the fact that
they used specimens with a 0·2 mm slit rather than
160r-------------------, fatigue precracked specimens.
Bradley and Andler27,94 used fatigue precracked
140
Charpy sized specimens, with the data analysed
using a I-integral approach (except on the lower
shelf, where LEFM were used). They obtained
upper shelf KJd values of 95-105 MN m-3/2, which
120
'" are very reasonable, given that the previously cited

values were 90 MN m-3/2 for quasistatic,
M
E 100 upper
z shelf fracture toughness of ferritic ductile iron (Fig.
~ 80
'C 36). It should be noted than an· impact velocity of
~ 5·4 m S-l was used in these studies and machine
cd
u
60 - compliance corrections were "not made. Thus the
~
reported upper shelf values may be slightly higher
40 -
(but not more than 5%) than the actual values.
However, the lower shelf KId values of 55-60 MN
ill -3/2 shown in Fig. 36 are ~uite high compared
20 - o Dynamic CT(FPC)
x Dynamic FPC Charpy with the values of 30 MN m - 72reported for nom-
Dynamic FPC DT
(!)
inally identical ferritic cast iron tested with an
0
-100 -65 -30 5 40 75 impact velocity of less than 2 m s -1 with data
Temperature,O C analysis correcting for machine compliance.33,95
36 Quasistatic and dynamic fracture toughness
The effect of microstructural features of ductile
values as measured using fatigue precracked iron on the dynamic fracture toughness has been
Charpy (FPC), compact tension (CT), and studied in detail by Bradley's group. The effect of
dynamic tear (DT) specimens (Ref. 94) percentage pearlite in the matrix on the upper shelf

110 0
100
0
100 80
<2% Pearlite
90 C\j
~ ~ 60
80 E
C\J 0
z
.•••....
C'1 70 • 2. 40
I
~
E 60 0
* Code C (91% nodularity)
z • 20
+ Code I (75% nodularity)
2:~ 50 # Code L (47% nodularity)
>75% Pearlite
'0 0
~ 40 -100 -70 -40 -10 20 50
Temperature, 0 C
30
39 Effect of nodularity on dynamic fracture tough-
20 ness KJd of ferritic ductile iron (Refs. 33, 95)
10
0 nodularity, nodule count, and volume fraction of
-40 o 40 80 120 graphite on the upper shelf dynamic fracture
toughness, as measured by KJd• As indicated while
Temperature,O C
discussing the quantitative metallography of cast
37 Effect of pearlite in ductile to brittle transition iron in this paper, three parameters have been
temperature of ductile iron (Ref. 94) designated to characterise quantitatively the graph-
ite shape and distribution in cast iron. The graphite
shape is considered to be nodular if the aspect ratio
fracture toughness values KJd and the transition is less than 2. The nodule count was determined as
temperature is seen in Fig. 37. It is clear that the number of nodules per square millimetre, as
pearlite both lowers the upper shelf and increases counted in a micrograph at 100 times magnification.
the transition temperature. . - The volume fraction of graphite was reported from
Bradley and co-workers91-93 found that the upper areal ratios in the micrograph. The results of
shelfKJd value was 89 MN m-3/2 (Fig. 34). This Bradley and Tanner33,95 are presented in Figs. 38
study also indicated that slag or degenerated graph- and 39. It is clear from the figures that the upper
ite greatly reduces the upper shelf KJd value of shelf fracture toughness KJd decreases as the nodule
ferritic ductile iron. It may be recalled that graphite count increases and the nodularity (percentage of
is considered to be degenerate (or irregular, vermi- nodules) decreases. These results are intuitively to
cular) if the aspect ratio (length of major axis/ be expected but what may not be so apparent is the
length of minor axis) is greater than 2. The lower observed decrease in the transition temperature
shelf KId values (35-40 MN m -3/2) reported in this that accompanies the increase in the nodule count.
stud~ are slightly higher than the value 30 MN The probable reason for this observation is dis-
m-32 previously suggested to be the best estimate cussed in a subsequent section. The above results
of lower shelf dynamic fracture toughness. This is again indicate that the upper shelf KJd for ductile
because an impact velocity of 3·3 m S-1 was used in iron with about 2·3% silicon and a nodule count of
this study, which is somewhat above the recom- less than 100- mm-2 is about 90 MN ffi-3/2. Bradley
mended maximum impact velocity of 2 m S-:I, as and Tanner33,95 also studied the effect of loading
discussed in the previous section. rate on the fracture toughness of ductile iron. These
Bradley and Tanner33,95 examined the effect of results (Fig. 40) show that the loading rate has a
100
100
80
C\J 80
"-
("'")
C\j
"-
M 60
IE c
E z 60
z :E~
2:
. 40 ~~ c
~ ~ 40
c Code A (105 nodules mrii2) C)
20 + Code 8 (183 nodules mrii2) ~

# Code 0 (515 nodules mrTi2)
20 * Quasistatic
c Dynamic
o
-100 -70 -40 -10 20 50 0
Temperature, 0 C -100 -70 -40 -10 20 50
Temperature, 0 C
38 Effect of nodule count on dynamic fracture
toughness KJd of ferritic ductile iron (Refs. 33, 40 Effect of strain rate on fracture toughness of
95) ferritic ductile iron (Refs. 33, 95)

significant effect on the ductile to brittle transition Effect of graphite volume fraction, nodularity,
temperature with a very slight effect on the upper and nodule count on quasistatic fracture
shelf fracture toughness. It is interesting to note toughness of ductile iron
that the loading rate dependent shift in the trans- It has been noted above35,86 that the addition of 12
ition temperature is similar to that noted by Rolfe vol.-olo of graphite nodules to a 2.25°10 silicon steel
and Barsomlli for ferritic steels. (thus making it a ferritic ductile iron) lowers the
upper shelf Charpy V-notch impact energy from 80
Summary of upper shelf dynamic fracture . to 16 N m. In this section the effects of graphite
toughness studies on ferritic ductile iron volume fraction, nodule count, and nodularity on
There are only a few reliable results for Kid and the quasistatic fracture toughness of ductile iron are
KJd for ferritic ductile iron published in the. litera- examined. For comparative purposes, some data
ture. These results indicate that the upper shelf involving Charpy specimens and dynamic tear
dynamic fracture toughness of ferritic ductile iron specimens are also considered.
(with 2.3°10 silicon and having less than 100 mm-2 Effect of volume fraction of graphite At least two
nodule count) as measured by KJd is studies84,112,113 have been made that indicate that
.......
90-95 MN m -3/2, this value decreasing with increasing the volume fraction of graphite decreases
increasing nodule count, decreasing nodularity, and the upper shelf fracture toughness of ductile iron,
increasing pearlite content in the matrix. with a relatively minor .effect on the ductile to
brittle transition temperature, as illustrated in Fig.
Lower shelf fracture toughness of ductile iron 32b. These results include fracture toughness as
A number of investigators33,68,83-85,95,lo8 have
measured using Charpy V-notch, dynamic tear, and
measured the lower shelf fracture toughness of CTOD specimens. As it is difficult to increase the
ductile iron and have obtained valid Kic results. It volume fraction of graphite while holding the
should be emphasised that for ferritic ductile iron nodule count constant, it is not clear whether the
the ductile to brittle transition temperature is observed decrease in the fracture toughness is
-100°C or lower, while for pearlitic ductile iron it indirectly a result of an increase -in the nodule count
may be above room temperature. while increasing the volume fraction, or whether
Typical results for fully ferritic ductile iron and increasing the nodule size alone (by increasing the
ferritic' ductile iron with about 15°10 pearlite are volume fraction at constant nodule count) would
presented in Fig. 41. The conclusion one may draw give a similar decrease in the fracture toughness.
from the data available in all the above references Effect of graphite nodularity It is clear from the
is that both ferritic and pearlitic ductile irons have results already presented that a degeneracy in
similar lower shelf fracture toughness values of graphite shape from nodule to flake gives a pre-
about 23-28 MN m -3/2. At temperatures above cipitous drop in KJc from 90 to 18 MN m -3/2.
-110°C, the fracture toughness of ferritic ductile However, the magnitude of decrease' in the upper
iron increases sharply with increasing temperature. shelf fracture toughness that results from a more
The presence of as little as 15°10 pearlite in ductile modest degeneracy as represented by vermicular
iron will cause a significant increase in ductile to graphite (Fig. 6) needs to be assessed. Work by
brittle transition temperature, ·as illustrated in Fig. Capaletti and Hornaday114 on unnotched Charpy
41. For' temperatures around O°C the fracture specimens indicated a monotonic decrease in
toughness KJc is a very sensitive function of the impact energy with decreasin~ nodularity. On the
percentage of pearlite in the matrix, the latter other hand, Nanstad et al.10 and Lazaridis and
affecting the fracture toughness by lowering the Loper115 have reported very little variation in Ko
upper shelf value, and more importantly, by values with nodularity (though they did observe a
shifting the ductile to brittle transition temperature. significant variation in dynamic tear energy). This is
not surprising in the case of ferritic ductile iron
since Ko measures only the onset of plastic defor-
mation on the upper shelf for this material. Recent
100
• • . . results by Bradley and Tanner33,95 indicated a drop

in the upper shelf fracture toughness from 90 to
about 70 MN m -3/2, as the degree' of nodularity
~
/ dropped from 91 to 47°10. Thus, it appears that
P'
/Ductile Iron With 15% Pearlite while the loss of nodularity does. not significantly
-,>-/ affect the yield strength (which is what Ko meas-
ured in Refs. 108 and 114), it does indeed lower the
stable crack growth
•• KJ1c
resistance to crack propagation, as one would
unstable crack growth expect. This aspect is discussed further below when
00 KJ Q
considering the fracture toughness of compacted
graphite cast iron.
o
-160 -120 -80 -40 o 40 Effect of nodule count A significant number of
Temperature, 0 C studies have been performed to determine the
41 Comparison of fracture toughness of fully ferritic effect of nodule count on the fracture behaviour of
ductile iron (GGG-40) with fracture toughness ductile iron, since this parameter can vary with
of a ferritic ductile iron with 15% pearlite in cooling rate during solidification. In many studies
matrix (5-45) (Ref. 95) . using both unnotched47 and V_notched83,115-l2l
Charp~ specimens as well as dynamic tear speci- 40

mens, 4 researchers have consistently found that a
higher nodule count (reduced internodular spacing)
gives a lower upper shelf fracture toughness and a
lower ductile to brittle transition temperature. For
example, Gilbert117 found that the Charpy impact ........
~ 30
energy was reduced from 20 to 12 N m as the 2
nodule count was increased from 30 to 500 mm-2. l0- -o:::~ -~__
1
Holdsworth and J 0lley84 have determined the
CTOD at maximum load as a function of the
nodule count for a constant volume fraction of
a
+-'
()
~
LL
>.
+-'
-----------E
::::--------------:.::_------
'A
. _--- .••. -6
graphite (Fig. 32a). Increasing the nodule count (f) 20

c
gave a significant drop in the upper shelf fracture Q)
+-'
toughness with a slight decrease in the ductile to c
brittle transition temperature. (f)
(f)
More recent examinations of the effect of graph~ Q)
l0-
ite nodule count on fracture tou1hness of ductile W Line No. Symbol Nodule Count,mni2
iron were made by Salzbrenner7 and by Bradley ~ 10 1 0 37
and Tanner.33,95 In these studies, KJc and KJd were ~
() 2
3
0
6.
107
52
determined for ferritic ductile irons with widely
varying nodule counts. The results from these two
·C
()
4
5
•• 21
74
studies are combined in Fig. 42-along with those of

6
• 40
Holdsworth and Jolley.84 It should be noted that o

the nodule count reported for a given microstruc- -200 -150 -100
ture can vary by as much as 50% from one Temperature,O C
laboratory to another, depending on the cut-off
43 Effect of nodule count on lower shelf fracture
criteria used for counting small graphite specks. It toughness of ductile iron (Ref. 84)
is clear from Fig. 42 that a finer graphite spacing
results in a lower KJc. In fact, the relatively linear
relationship between KJc and the square root of the
nodule spacing, dO.5 (note: dO.5 ~ N-O.25), obser- needs to be achieved over a shorter distance when a
ved in this figure is consistent with the recent high nodule count is present.
ductile fracture theory of Ritchie and Thompson. 122 There is only one clear study84 on the effect of
From a· physical viewpoint, the larger the number nodule count on the lower shelf fracture toughness
of nodules, the easier it is to achieve void coale- of ferritic ductile iron, which indicates that there is
scence at the crack tip, since the nodule spacing will a slight improvement in the lower shelf fracture
be closer. This also means that the critical strain toughness as the nodule count increases (Fig. 43).
Two studies have been reported which indicate
that a higher -nodule count improves· the fracture
toughness of pearlitic ductile iron.123,124.Holds-
Ref. S MN m-2 worth and J0lley123 fouild that in pearlitic ductile
120
•
@
77
95
Y260
290 ~ iron higher CTOD values (at maximum load) are
...~ ~ obtained with increased nodule count, both in the
77
77
330
410 •
0 84 -260 upper and lower shelves. Furthermore, Chi and
110
• 77 185
Ruizhen 124 found an increase in the fracture
• toughness of pearlitic ductile iron with nodule count
100 0
at room temperature.
(\J
~ Since there is limited crack tip plasticity for
'-
(")
90
@ ferritic ductile iron in the lower shelf and both in
E 0 the upper and lower shelves for pearlitic ductile
z
:2
... iron, the above studies seem to indicate that the
A
:s::.~ 80 @ larger the number of nodules, the greater is the
resistance to cleavage.
In summary, a high nodule count will give a
70
~ lower value of the upper shelf fracture toughness
... but a lower transition temperature and. a ·higher
60
@
~ lower shelf fracture toughness, for ferritic ductile
iron. Two investigations have shown that very large
nodules (resulting in very low nodule count) can
50
0.3 0.4 0.5 have an embrittling effect on ferritic ductile iron, as
1/N1/4 (N in nodules mrri2) evidenced by large reductions in tensile elonga-
42 Effect. of nodule count on upper shelf fracture tion 77and impact energy. 124
toughness KJc of ferritic ductile iron: d is nodule Effect of nodule count on ductile to brittle transition
spacing which is proportional to N, the nodule temperature of ductile· iron It has been noted
count (Refs. 33, 77, 84, 95) above that silicon steel with nodules (i.e. ferritic
ductile iron) has a ductile to brittle transition obtained from pearlitic ductile iron is of little
temperature that is 130 K lower than that of a practical significance because it is obtained only at
conventional cast steel. 35 The reason given for the temperatures above 200°C. At room temperature,
beneficial effect of graphite nodules in resisting the by contrast, pearlitic ductile iron has an impact
brittle fracture of both ferritic and pearlitic ductile value of only 3N m. Thus, at ambient or lower
iron is that the nodules locally give crack tip temperatures, pearlitic ductile. iron exhibits low
blunting, makin.r. it more difficult to propagate a fracture toughness.
cleavage crack.8 ,115-120 A similar argument is made For quasistatic fracture of pearlitic ductile iron,
for the further lowering of the transition tem- the mid point of the ductile to brittle transition
perature that results from increased nodule count, occurs at ----30°Cwith upper shelf values obtained at
though this effect could also result from the ----75°C(Refs. 85, 123). Since the mid point of the
correlation of higher nodule count with finer ferritic ductile to brittle transition temperature is approxi-
grain size which is known to lower the transition mately the ambient temperature, any factor that
temperature in steel. Alternatively, the presence of shifts the transition temperature for pearlitic ductile
nodules lowers the stress required for ductile iron will have a significant effect on the quasistatic
fracture, and thus ductile fracture is easier than fracture toughness measured at room temperature.
cleavage. It is believed that this accounts for the great variety
A similar advantage in the transition temperature of results reported in the literature for pearlitic
of ferritic ductile iron compared with a ferritic- ductile irons as summarised below. The wide range
pearlitic cast steel with similar tensile properties has of reported values in the upper shelf fracture
been noted above (Fig. 35). Mogford and Hull116 toughness of ferritic ductile iron has already been
found a linear relationship between the transition attributed to variations in experimental procedure
temperature and the internodular spacing. Alterna- and analysis. For pearlitic ductile iron, most of the
tively, Nishi et al.125 found that the transition reported procedures satisfy ASTM E399-8129 and
temperature decreased linearly with the square root thus, the variations in fracture toughness measured
of the internodular spacing. They have determined are real rather than a function of experimental
that the ductile to brittle transition temperature conditions.
(TT) for unnotched Charpy sized bars of ferritic . Effect of pearlitic matrix constituent· on' fracture
ductile iron is given by toughness of ductile iron Several investi-
gators32,83,123,124 have found K1cat room tem-
TT = 74 - 15·8do.5 - 19·5D-o.5 (4)
perature and below to be ----30MN m-3/2 or less for
where d is the internodular spacing and D the pearlitic ductile iron. However, modest amoupts of
ferritic grain size, both measured in millimetres. ferrite surrounding graphite nodules in such a
Typical values of transition temperatures for structure significantly increase the toughness at
ferritic ductile irons (with about 2·3% silicon and a room .temperature (again probably through shifting
nodule count of 60-100 mm -2) are as follows. The the transition temperature as well as increasing the
nil ductility transition temperature is about -40°C upper shelf energy). For example, Mead and
(Re.f. 66). The mid point in a Charpy V-notch Bradley (Table 7, item no. 7) found, in a nor-
energy transition is approximately about 0 to -20°C malised ductile iron (which showed small amounts
(Refs .. 27, 35, 83, 126). The mid point in the of ferrite surrounding graphite nodules in a pearlitic
transition temperature range for .KJd is between matrix), a KJc value of 54 MN m-3/2 which corre-
-10 and -40°C (Refs. 27, 33, 95). The quasistatic sponds to a recalculated K equal to 77 MN m -3/2
Jrnax
fracture toughness transition temperature is -130°C at room temperature. (Note: the amount of crack
(Ref. 93). extension was not clearly defined in this test. )
Summary of effects of graphite nodules on fracture Nanstad et al.l08 have measured KQ values of
toughness of ductile iron In summation, a low 45 MN m-3/2 and Kmax values of 68 MN m-3/2 for
nodule count, a small volume fraction of graphite, as cast pearlitic ductile iron (56% pearlite, 44%
and good nodularity are necessary to achieve the ferrite). Ostensson80 reported a KQ value of
highest possible upper shelf fracture toughness in 95 MN m-3/2 at room temperature and a valid K1c
ferritic ductile iron. The best lower shelf fracture value of 65 MN m -3/2 at O°Cfor a ductile iron with
toughness for ferritic ductile iron is obtained when 70% pearlite and 30% ferrite. Niskanen,87 for a
the nodil1e count is high. Pearlitic ductile iron will ferritic-pearlitic iron, found a K1c value equal to
have the best upper and lower shelf fracture 50 MN m -3/2 at room temperature and a value of
toughness when the nodule count is high and the 23 MN m -3/2 at -100°C. These results in total
nodularity is good. A lower ductile to brittle suggest that the upper shelf fracture toughness of a
transition temperature for both ferritic and pearlitic - ferritic-pearlitic ductile iron is not dramatically less
ductile irons results when the nodule count is than that for a ferritic ductile iron. Furthermore,
higher. the much lower room temperature fracture tough-
Effect of matrix microstructure on transItIon ness values observed for ferritic-pearlitic ductile
temperature" of ductile iron The mid. point trans- irons are due principally to their much higher
ition temperature for Charpy V-notch specimens ductile to brittle transition temperature rather than
tested dynamically shifts from -10 to 120°C as the to a significant lower of the upper shelf
matrix changes from fully ferritic to fully pearlitic, values.l22
while the .upper shelf impact energy decreases It should be noted that some of the earliest
from 21 to 11 N m (Ref. 122). A value of 11 N m CTOD values published for pearlitic ductile irons
gave artificially high figures because of the use of

DD 19. The magnitude of the error is much greater
in pearlitic ductile iron than in ferritic ductil~ iron
because most, if not all, of the CTOD values In the
former are elastic, rather than plastic as assumed in
DD 19. Thus, Holdsworth and Jolley85 reported a
CTOD value of 0·0375 mm for a pearlitic ductile
iron, while Luyendijk and Nieswaag obtained a
value of 0·012 mm for a similar cast iron using
BS 5762. Calculation of CTOD using DD 19 for the
cast iron used by Luyendijk and Nieswaag gives a
value of 0·046 mm, which is comparable to that
obtained by Holdsworth and Jolley. A CTOD value
of 0·012 mm corresponds to a K1c value of
42·2 MN m-3/2•
The results obtained by Seetharamu and Sri- 45 SEM fractograph showing stable crack growth
nivasanl26-129 for two sets of permanent mould region in permanent mould ductile iron
ductile iron samples are shown in Fig. 44. The (Ref. 128)
chemical compositions of both the sets were similar
except for the silicon content, which was 3°1o in one room' temperature K value of 48 MN ~ -3/2,
set and 2°1o in the other. It is seen from Pig. 44 that Jrnax
which was nearly the same at -19 and -59 C. In

for a given silicon content, the fracture toughness
contrast, 80-60-03 grade ferritic-pearlitic ductile
decreases with increase in pearlite content. The
iron with a much lower yield strength (433 MN
significantly lower fr,acture toughness values exhi-
m-2) showed lower K values at corresponding
bited by the samples .with 3°1o silicon may be Jrnax
temperatures. Also the K values for the }att~r ~t

attributed to the upward shift in the ductile to Jrnax
-19°C were significantly lower than at 24 C, IndI-

brittle transition temperature when the silicon
cating that the transition temperature lies in this
content is increased, as discussed above. The SEM
range.32 It may also be noted in this context that
fractograph shown in Fig. 45 is typical of 3°10 silicon Niskanen87 measured K1c values of 50-60 MN
permanent mould ductile iron castings with less m -3/2 down to -70°C, for a quenched and tem-
than 50°10 pearlite and shows a significant amount
pered ductile iron with a nominal yield strength of
of fibrous rupture, indicating that stable crack 820 MN m-2•
growth occurs in these castIngs,. 128 wh'l
I e P'I~. 46'
• IS
typical of 3°10 silicon permanent mould ductIle Iron Effect of bainitic microstructural constituent
castings with more than 50°10 pearlite and indicates on fracture toughness of ductile iron
that fracture is predominantly by cleavage in'these Considerable interest has developed in recent years
castings. 128 in austempered ductile iron for gear applications.
A tempered martensitic matrix ~i~es lo~~r This material gives a much better toughness than
Charpy im~act toughness than a ferntIc-pearl~t~c other ductile iron grades at hardness levels· suitable
matrix.125,1 0 However, tempered martensitIc for gears.131 Whereas a pearlitic ductile iron with a
matrices give a much higher strength than does a yield strength of 550 MN m -2 has an ambient
ferritic-pearlitic matrix .. If comparison is made at Charpy impact value of only 3 N m (Ref. 132), a
similar strength levels, a cast iron with tempered bainitic ductile iron with a. yield strength of
martensite will have a greater upper shelf fracture 725 MN m-2 has a Charpy value of 9 N m (Ref.
toughness and a lower ductile to brittle transition 133). Smith et al.134 have noted that at similar
temperature. For example, in Table 7 D7003 grade values of hardness, ductile iron with a bainitic
quenched and tempered martensitic ductile iron
(yield strength, 717 MN m -2) was found to have a
100
80 + +
N +
l"'"l
+
E 60
z + +
L:
u 40
:Z ~~~~ ~
+ 2% Si
~ ~ ~~
20
~ 3% Si
0
0 20 40 60 80 100
PEARLlTE,%
44 Effect of pearlite content on fracture toughness 46 SEM fractograph showing transgranular cleav-
of permanent mould ductile iron (Refs. 126-129) age in permanent mould ductile iron (Ref. 128)

matrix gives a better Charpy toughness than a 20

• CGI - Gray iron base
pearlitic or tempered martensitic ductile iron.
• CGI - Ductile iron base
Whether this was the result of a lower transition o CGI - 0.67% Cu-pearlitic
temperature and/or a greater upper shelf toughness • Ductile iron - 0.49% Cu - pearlitic
was unclear since all the tests were run at room ~ Ductile iron - ferritic/pearlitic
temperature. Gundlach and Janowak135 have stated
that much better combinations of strength and 15
ductility result in austempered ductile irons than in E
ferritic-pearlitic or tempered martensitic ductile z
irons. Barbezat and Mayer136 have recently
published a KJc value of 70 MN m -3/2 at room ~
0)
~
temperature for an austempered ductile iron with a Q)
yield strength of 740 MN m-2. They also obtained c

W
values of 58 and 56 MN m -3/2 at -20 and -40°C, ~ 10
Cl.
respectively. Thus it is clear that austempered ~
ductile iron shows the potential of obtaining good res
..c
fracture toughness even for quite high strength ()
levels at low temperatures.
--..-.---!-. .. .....
Summary of effect of pearlitic and bainitic
microstructural constituents on fracture
toughness of ductile iron
5
..• ..• ..
••• ii---.
---- " "
--~1
In summary, increasing the matrix pearlite •
penalises the ambient temperature fracture
toughness primarily by shifting the ductile to brittle
transition temperature and to lesser degree, by
o 20 40 60 80 100
lowering the upper shelf fracture toughness. As is Pearl ite 0/0
I
the case with conventional steels, tempered marten- 47 Effect of pearlite content on Charpy V-notch
sitic and bainitic matrices in ductile iron give better impact energy of as cast compacted graphite
combinations of strength and fracture toughness cast iron (CGI) compared with ductile iron: data
than pearlitic matrices. refer to compacted graphite cast iron and duc-
tile iron specimens of different compositions
(Ref. 137)
Compacted graphite cast iron
Very little research hcs been conducted in the area ambient temperature. The fracture toughness of the
of fracture toughness of compacted graphite cast latter was assessed from CTOD at initiation of
iron. Since the shape of the compacted (vermicular) stable crack .
. graphite is an intermediate between flake and The microstructural data and the associated
nodule, one would expect the tensile and fracture mechanical properties of these permanent mould
toughness properties to fall in between those of castings with different amounts of irregular (vermi-
grey cast iron and ductile iron. The most definitive cular) graphite are summarised in Table 8. It is seen
work on the fracture toughness of compacted that in the as cast state, the amount of pearlite
graphite cast iron is by Loper et al.137 These results, decreases as the percentage of non-spheroidal
presented in Figs. 47 and 48, compare the Charpy (vermicular) graphite increases. Furthermore, the
and the dynamic tear energies of compacted graph- yield stress values of these castings are significantly
ite cast iron with those of ductile iron. These lower than those of permanent mould ductile iron
comparisons for several different matrix microstruc- castings.128 A comparison of the results in Table 8
tures. indicate that the deleterious effect of vermicu- for pearlitic compacted graphite cast iron with those
lar graphite on fracture toughness is much more
pronounced for a ferritic matrix than for a pearlitic
matrix. Furthermore, the compacted graphite cast Table 8 Microstructural data and mechanical
iron seems to have a somewhat lower ductile to properties of chill free permanent mould
brittle transition temperature than does ductile magnesium treated iron castings with
iron, possibly because the ductile fracture has been irregular (vermicular) graphite: Si content
made easier relative to cleavage by the formation of 3% (Ref. 128)
compacted graphite. Pearlite, Nsp, * Sy,t Elongation, K,c,
During the. course of their investigations on the Column no. 0/0 0/0 MN m-2 % MN m-3/2
fracture toughness of ductile iron? Seetharamu and 1 50·0 33·1 454·2 5·10 37·36
Srinivasan127,128 encountered vermicular graphite in 2 42·0 24·3 471·9 4·80 38·57
different proportions with nodules in several cases. 3 38·7 16·7 449·8 5·50 37·58
4 30·2 9·5 425·8 5·25 36·70
The fracture toughness of these samples was also 5 25·1 5·5 419·9 5·30 37·14
assessed. Chill free permanent mould castings (with 6 16·0 6·1 423·8 5·63 39·34
3% silicon) with vermicular graphite gave valid K1c
* Nsp is percentage occurrence of graphite nodules in microstruc-
values as per ASTM E399 while castings with 2% ture; remainder is irregular (vermicular) graphite.
silicon exhibited considerable crack tip plasticity at t Yield strength~

00% Pearlite, annealed.2 • 0% Pearlite, annealed.2

~23% Pearlite.2 A 5% Pearlite, as-cast.2
0.40 A 44% Pearlite.2 0.40 • 35% Pearlite, normalized.2
.98% Pearlite. z::z 25-97% Pearlite, as-cast.
N
'E
J •. 0.30 --:: 0.30
~ ~
0)
0>
~ ~
(1) (1)
C c
W W
~ 0.20 ~ 0.20
CO co
Q.) (1)
r- r-
u
E 0.10
CO
c
~
o
j 0.10
o (0) o (b) .
150 200 250 300 350 150 200 250 300 350
Temperature, K Temperature, K
48 Dynamic tear energy v. temperature for b compacted graphite cast iron and a ductile iron (Ref. 137)
for pearlitic nodular graphite cast iron shown in itic grades of white cast iron had better impact
Fig. 44 indicates that the fracture toughness values strength than did martensitic gades, though no such
are similar. This means that either the fracture clear pattern was noted for Kld values.
toughness of pearlitic ductile iron is relatively Eriksson 140 measured a fracture toughness value
insensitive to graphite morphology or, that the of 22 MN m -3/2 on a white cast iron sample with
observed trend is due to the lowering of ductile to martensitic matrix corresponding to a ViCker's
brittle transition temperature by vermicular hardness value of 633. A comparison of his results,
graphite. shown in Fig. 49, which gives K1c versus hardness,
Vermicular graphite was also encountered by with those of Gahr and Scholz, indicates that both
Seetharamu and Srinivasanl28,129 in a casting with
2% silicon, made in an investment mould. In the
fully annealed (ferritised) condition the yield 34
strength was about 300 MN m-2, which was signi-
ficantly lower than the mean value of 385 MN m-2 C\l
•......
obtained for ferritic permanent mould and sand f(

castings with 2% silicon. Furthermore, the critical E 30
CTOD value of this' casting sample was 0·041 mm z
2: 31.
as against 0·055 mm obtained for samples with -
u 29 •• 33
nodular graphite. It may again be surmised from
these results that the fracture toughness of vermicu-
£
-0 26
••
53
lar graphite cast iron is significantly lower than that £ Austenitic 41&
of ductile iron. 137 en 031

en
Q) 55~27
c 22 33 29 ~9
.c
White cast iron 0)
::::s
A· number of studies have been conducted to 350 053
0 450
determine the fracture toughness of white cast l- 51 1~~ 59
Q) 10 ~. ~
iron. 138-141 White cast iron is generally used in lo- ~ 19
::::s 18 05 21 A~23 ~
applications where abrasion resistance is desirable. U 55
rn ~ 9 ~47
Abrasion resistance normally increases with l0- 39 ~ ~ 25
u. 43t 15
hardness however, a concurrent decrease in frac- Austenitic
41 11~
ture toughness is noted. Gahr and Scholz139 have Martensitic 7~ ~ 13
14
found that the dynamic fracture toughness (KId)
values, measured on 10 x 10 x 55 mm specimens 500 600 700 800
with a 0·2 mm wide slot, were in the range of 14-24 Hardness, HV50
MN m-3/2 while the quasistatic fracture toughness
49 Relationship between fracture toughness and
values, measured on fatigue precracked compact hardness for white cast irons of different micro-
tension specimens, were in the range structures, produced by heat treating in variety
25-33 MN m-3/2. Under dynamic loading, austen- of ways (Ref. 138)

100
100 fully ferritic
ferritic grades of ductile iron
15% pearlite
80 /
/---------------------
80 malleable iron
,," .•.. -------- ---------------- I
I as cast-70% pearlite
I
/
I
/ ,,-
-------
••....••...
---.-.-
I / "/' 100% pearlite
/ compacted graphite iron
/";----~- ---------- 1/'
_-J--~-- __ I __ /-- _
o , I ,/ / / /. quenched and tempered

':2. 40 I ,/.
// ,/ /' ",,"./
/',,"
--_ ...• " ~._.-""
20 _._._._._._._._._._.~~n 20
-100 -75 -50 -25 25 50 75

Temperature,OC -150 -100 -50 0 50 100 150
Temperature, °C
50 Effect Qf graphite shape on fracture toughn'ess
of ferritic grades of various cast irons: these 52 Effect of microstructure on fracture toughness
results are generally based on data from the of ductile iron: these results are generally based
literature; where data from the literature were on data from the literature; where data from the
incomplete, approximations have been included literature were incomplete, approximations
for purposes of illustration have been included for purposes of illustration
the hardness and the toughness values obtained by siders only ductile iron (with spheroidal graphite)
Eriksson are lower. This may be attributed to the and varies the matrix from, fully ferritic to fully
presence of some flake graphite present in Eriks- pearlitic, through progressive increase in pearlite
son's sample. . percentage, the upper shelf fracture toughness
Sare141 studied the quasistatic fracture toughness decreases and the ductile to brittle transition
of white cast iron using the double torsion test. His temperature increases (Fig. 52). However, in the
results are generally consistent with those of Gahr case of grey cast iron (with flake graphite) the
and co-workers. 138,139 fracture toughness perceptibly increases as the
matrix strength increases, that is, as the matrix
changes from ferritic to pearlitic.
Summary Most alloying elements or impurity elements
The fracture toughness of cast iron has been shown present in cast irons affect the fracture toughness
to be dependent on the graphite morphology, as indirectly through effect on the graphite morphol-
well as the matrix microstructure present in the cast ogy or matrix microstructure. Two exceptions to
iron. Figures SO-52 summarise the authors' obser- this general rule are siliron and phosphorus, both of
vations by comparing the fracture toughness values which significantly increase the ductile to brittle
of various cast irons. With a fully ferritic matrix, transition temperature.
the upper ,shelf fracture toughness is a very sensitive
function of graphite shape, while the lower shelf
fracture toughness is relatively insensitive to graph- References
ite shape variations (Fig. 50). For a pearlitic matrix, 1. c. F. WALTON and T. 1. OPAR (eds.): ~Iron castings hand-
the graphite morphology is again found to be book', 3-4; 1981, Des PJaines, IL, Iron Castings Society.
relatively unimportant at ambient temperature or 2. c. F. WALTON and T. J: OPAR (eds.): 'Iron castings hand-
below, which essentially corresponds to lower shelf book', 122; 1981, Des Plaines, IL, Iron Castings Society.
behaviour for these alloys (Fig. 51). If one con- 3. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand-
book', 134; 1981, Des Plaines, IL, Iron Castings Society.
4. A. R. BAILEY and L. E. SAMUELS: 'Foundry metallography -
80 annotated metallographic specimens' ,146; 1976, Bletch-
N
worth, UK, Metallurgical Services.
ductile iron
"-
M
pearlitic grades of
5. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand-
'E 60 book', 135; 1981, Des Plaines, IL, Iron'Castings Society.
malleable iron
z
2: ".•..• ---- 6. 'Specification for classification of type and size of graphite
o " I compacted graphite iron flakes in gray iron', A247; Philadelphia, PA, American
':2. 40 /---7'---- Society for Testing and Materials.
, " 7. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand-
/_--,,"
book', 145; 1981, Des Plaines, IL, Iron Castings Society.
20 _._._._._._._._~~ iron
8. A. R. BAILEY and L. E. SAMUELS: 'Foundry metallography -
annotated metallographic specimens', 169; 1976, Bletch-
-25 25 50 75 100 125 150 9. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand-
Temperature, ° C book', 129; 1981, Des Plaines, IL, Iron Castings Society.
10. R. ELLIOTI: 'Cast iron technology', 12; 1988, London,
51 Effect of graphite shape on fracture toughness
Butterworths.
of pearlitic grades of various cast irons: these 11. R. ELLIOTI: 'Cast iron technology', 224; 1988, London,
results are generally based on data from the Butterworths.
literature; where data from the literature were 12. A. R. BAILEY and L. E. SAMUELS: 'Foundry metallography -
incomplete, approximations have been included
for purposes of illustration , annotated metallographic specimens', 167; 1976, Bletch-

13. G. S. COLE: in 'Source book on ductile iron', 162; 1977, 51. B. OSTENSSON:Eng. Fract. Mech., 1972~ 4, 443-448.
Metals Park, OH, American SoCiety for Metals. 52. T. V. VENKATASUBRAMANIAN and T. 1. BAKER:Met. Techno!.,
14. w. F. SMITH: 'Structure and properties of engineering 1978, 5, 57-61.
alloys', 335; 1981, New York, McGraw-Hill. 53. T. 1. BAKER:Mater. Eng. App!., 1978, 1, 13-18.
15. K. D. MILLIS: in 'Source book on ductile iron', 37; 1977, 54. D. K. VERMA and 1. T. BERRY: Trans. ASME, 1982, 104,
Metals Park, OH, American Society for Metals. 262-266.
16. R. ELLIOTT: 'Cast iron technology', 225; 1988, London, 55. A. G. GLOVERand G. POLLARD:J. Iron Steel Inst., 1971,209,
Butterworths. 138-141.
17. R. ELLIOTT: 'Cast iron technology', 109; 1988, London, 56. R. B. TAIT andD. P. SPENCER: 'Tensile, impact and fracture
Butterworths. toughness characterization of three cast iron sheave wheel
18. R. HUMMER: in 'The metallurgy of cast iron', (ed. B. Lux materials', Research Report, Department of Metallurgy,
et al.), 147-159; 1975, St Saphorin, Switzerland, Georgi University of Witwatersrand, South Africa, 1982.
Publishing. 57. L. GRUTER:Mater. Sci. Eng., 1978, 35, 157-164.
19. G. s. COLE: in 'The metallurgy of cast iron', (ed. B. Lux 58. L. GRUTER:Dr Ing. thesis, Bochum, Ruhr University, FRG,
et al.), 673-695; 1975, St Saphorin, Switzerland, Georgi 1975.
Publishing. 59. L. GRUTER:Giessereiforschung, 1977,29,47-51.
20. M. S. RAMA PRASAD, C. S. NARENDRANATH, and M. N. 60. 'Modern pearlitic malleable castings handbook'; 1958,
SRINIVASAN:in 'Solidification technology in the foundry and Granville, OH, Malleable Research and Development
casthouse', 336-344; 1983, London, The Metals Society. Foundation.
21. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand- 61. G. Yu. SCHULTE:Metalloved. Term. Obrab. Met., Nov. 1973,
book', 132; 1981, Des PIa"ines, IL, Iron Castings Society. 35-37.
22. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand- 62. F. W. JACOBS and F. L. PRESTON: Trans. AFS, 1975, 83,
book', 237, 302, 326, 382, 449; 1981, Des Plaines, IL, Iron 263-270.
Castings Society. 63. P. B. BURGESS: Trans. AFS, 1967, 75, 665-675.
23. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand- 64. T. OKUMATOand K. KONDO: Imono (J. Jpn Foundrymen's
book', 211; 1981, Des Plaines, IL, Iron Castings Society. Soc.), 1962, 34, 34.
24. A. R. BAILEYand L. E. SAMUELS: 'Foundry metallography - 65. F. 1. WORZALA,R. W. HEINE, and Y. W. CHENG: Trans. AFS,
annotated metallographic specimens', 172; 1976, Bletch- 1976, 84, 675-682.
worth, UK, Metallurgical Services. 66. A. LITTLEand H. J. HEINE: 'The metallurgy of cast iron', (ed.
25. 'Standard test method for dynamic fracture toughness', B. Lux et at.), 767-788; 1975, St Saphorin, Switzerland,
Spec. E604-83; 1983, Philadelphia, PA, American Society Georgi Publishing.
for Testing and Materials. 67. v. I. OVCHINNIKOV,V. A. KURDYUKOV,and s. G. GIRENKOV:
26. 'Fracture resistance of malleable iron', Bull. 78, Malleable Metalloved. Term. Obrab. Met., Feb. 1982, 13-14.
Research and Development Foundation, Dayton, OH, 1969. 68. R. 1. SALZBRENNER, 1. A. VAN DEN AVYLE, T. 1. LUTZ, and
27. w. L. BRADLEYand M. R. ANDLER; in 'Fracture prevention in w. L. BRADLEY:in 'Fracture mechanics: sixteenth sympos-
energy and transport systems'" 524-535; 1984, West Mid- ium', (ed M. F. Kanninnen and A. T. Hopper), STP 868,
lands, UK, Engineering Materials Advisory Services. 328-345; 1985, Philadelphia, PA, American Society for
28. 'Gray and ductile iron castings handbook'; 1971, Cleve- Testing and Materials.
land, OH, Gray and Ductile Iron Founders Society. 69. w. L. BRADLEYand H. MEAD: in 'Cast metals for structural
29. 'Standard test method for plane-strain fracture toughness of and pressure containment applications',' MPC-11, 69-87;
metallic materials', Spec. E399-81; 1981, Philadelphia, PA, 1979, New York, American Society of Mechanical
American Society for Testing and Materials. Engineers.
30. 'Standard test method for JIe, a measure of fracture 70. w. L. BRADLEY and H. MEAD: Trans. AFS, 1980, 88,
toughness', Spec. E813-81; 1985, Philadelphia, PA, Amer- 265-279.
ican Society for Testing and Materials. 71. H. E. MEAD: MS thesis, Texas A&M University, 1979.
31. 'Method for crack opening displacement (COD) testing', 72. c. T. MOORE: Br. Cast Iron Res. Assoc. J., 1961, 9,
BS 5762; 1979, London, British Standards Institution. 385.
32. w. L. BRADLEY:Trans. AFS, 1981, 89, 837-848. 73. c. T. MOORE: Br. Cast Iron Res. Assoc. J., 1963, 11,
33. w. L. BRADLEYand M. TANNER:in 'Fracture mechanics', STP 58-60.
945, 18; 1987, Philadelphia, PA, American Society for 74. D. R. ASKELANDand R. F. FLEISCHMAN: Trans. AFS, 1978,
Testing and Materials. 86, 373-378.
34. R. J. SALZBRENNER:'Tensile behavior of ferritic ductile cast 75. G. E. DIETER: 'Mechanical metallurgy', 264; 1986, New
iron', Sandia National Laboratories, Albuquerque, NM, York, McGraw-Hill.
1986. 76. R. W. HERTZBERG:'Deformation and fracture mechanics of
35. G. SANDOZ, H. BISHOP, and W'. PELLINI: Trans. ASM, 1954, engineering materials', 252; 1983, New York, Wiley.
46,418. 77. R. SALZBRENNER:J. Mater. Sci., 1987,22,2135-2147.
36. c. F. WALTON and T. J. OPAR (eds.): 'Iron castings hand- 78. R. K. NANSTAD,F. 1. WORZALA,and c. R. LOPER: Trans. AFS,
book', 133; 1981, Des Plaines, IL, Iron Castings Society. 1974, 82, 473-486.
37. B. LUX: AFS Cast Met. Res. J., 1972, 8, 25-38; 49-65. 79. R. K. NANSTAD, F. 1. WORZALA, and c. R. LOPER: in 'The
38. 1. F. WALLACE: Trans. AFS, 1975, 83, 363-378. metallurgy of cast iron', (ed. B. Lux et al.), 789-807; 1975,
39. J. M. DONG: Trans. AFS, 1978, 86, 249-256. St Saphorin, Switzerland, Georgi Publishing.
40. Met. Prog., Oct. 1978, 125, 64-68. 80. B. OSTENSSON:Scand. J. Metall., 1973, 2, 194-196.
41. c. F. WALTON and T. J. OPAR (eds.): 'Iron castings hand- 81. c. w. TEN HAAGENand 1. T. BERRY:in 'Fracture 1977', (ed.
book', 141; 1981, Des Plaines, IL, Iron Castings Society. D. M. R. Taplin), Vol. 3, 565-572;1977, Waterloo,
42. G. N. 1. GILBERT:Br. Foundryman, 1961, 54, 5. Canada, University of Waterloo Press.
43. L. ELDOKYand R. C. VOIGT: Trans. AFS, 1985, 93, 365-372. 82. c. W. TEN HAAGEN and 1. T. BERRY: Trans. AFS, 1976, 84,
44. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand- 535-542.
book', 260; 1981, Des Plaines, IL, Iron Castings Society. 83. 1. M. MOTZ: Giesserei, 1980,67,628.
45. w. L. BRADLEY:Unpublished work, Texas A&M University. 84. s. R. HOLDSWORTHand G. JOLLEY: in 'The metallurgy of cast
46. T. KOBAYASHI: Trans. Iron Steel Inst. Jpn, 1976, 16, iron', (ed. B. Lux et at.), 809-825; 1975, St Saphorin,
138-145. Switzerland, Georgi Publishing.
47. Y. KURODAand H. TAKADA:AFS Cast Met Res. J., 1970, 6, 85. s. R. HOLDSWORTHand G. JOLLEY: Br. Foundryman, 1975,
(2),63-74. 68,169-172.
48. R. EISENSTADTand G. H. ROULSTON: in Proc. ASME Gas 86. G. JOLLEY and s. R. HOLDSWORTH:in 'Fracture 1977', (ed._
Turbine Conf., Houston, TX, Mar. 1975. D. M. R. Taplin), Vol. 2, 403-413; 1977, Waterloo,
49. A. G. GLOVERand G. POLLARD:in 'Fracture', 350-359; 1969, Canada, University of Waterloo Press.
London, Chapman and Hall. , 87. 1. NISKANEN:Int. Cast Met. J., 1977,2, 2(}'-23.
50. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand- 88. T. LUYENDIJK and H. NIESWAAG: Proc. 47th Int. Foundry
book', 144; 1981, Des Plaines, IL, Iron Castings Society. Cong., Jerusalem, 1980, 641-683.

Bradley and Srinivasan' Fracture of cast irons 161
89. T. LUYENDIJK and H. NIESWAAG: in 'Solidification technology 113. J.F. WALLACE: 'Toughness of ductile iron: literature
in the foundry and casthouse', 569-574; 1983, London, The review', 13; 1980, Mountainside, NJ, Ductile Iron Society.
Metals Society. 114. T. L. CAPELETTI and 1. R. HORNADAY: Trans. AFS, 1974, 82,
90. M. CAYARD: MS thesis, Texas A&M University, 1987. 59.
91. K. E. McKINNEY: MS thesis, Texas A&M University, 1984. 115. A. LAZARIDIS and c. R. LOPER: Trans. AFS, 1974, 82,
92. w. L. BRADLEY, K. E. McKINNEY, and P. GERHARDT: in 473-486. .
'Fracture mechanics: seventeenth volume', (ed. J. H. 116. I. L. MOGFORD and D. HULL: J. Iron Steel Inst., 1968, 206,
Underwood et al.), STP 905, 75-94; 1986, Philadelphia, 79-88. "
PA, American Society for Testing and Materials. 117. G. N. 1. GILBERT: Foundry Trade J., 1966, 121, 507.
93. K. E. McKINNEY, W. L. BRADLEY, and P. GERHARDT: Trans. 118. G. JOLLEY and G. N." 1. GILBERT: Br. Foundryman, 1967, 60,
AFS, 1984, 92, 239-250. 79.
94. M. R. ANDLER: MS thesis, Texas A&M University, 1983. 119. "·G. N. 1. GILBERT: Br. Cast Iron Res. Assoc. J., 1964, 12,590.
95. G. M. TANNER: MS thesis, Texas A&M University, 1986. 120. G. N. 1. GILBERT: Br. Cast Iron Res. Assoc."]., 1963, 11, 199.
96. w. L. BRADLEY: J. Met., Jan. 1985, 37, 74-78. 121. G. N. 1. GILBERT: Br. Cast Iron Res. Assoc. J., 1964, 12, 791.
97. M. CAYARD and w. L. BRADLEY: Eng. Fract. Mech., 1989, 33, 122. R. O. RITCHIE and A. W. THOMPSON: Metall. Trans., 1985,
(4), 121-132. 16A, 233-248.
98. D. M. NORRIS, 1. E. REAUGH, and w. L. SERVER: 'Fracture 123. s. R. HOLDSWORTH and G. JOLLEY: Br. Foundryman, 1974,
mechanics', STP 743, 207-217; 1981, Philadelphia, PA, 68,77-82.
American Society for Testing and Materials. 124. X. CHI and Y. RUIZHEN: Int. Cast Met." J., 1979, 4, (2),
99. 'Castings for fusion welding for high temperature service', 21-28.
Spec. A216-82; 1982, Philadelphia, PA, American Society 125. N. NISHI, T. KOBAYASHI, arid s. TAGA: Imono (J. Jpn
for Testing and Materials. Foundrymen's Soc.), 1976, 48, 9.
100. T. KOBAYASHI and S. NISHI: Imono (J. Jpn Foundrymen's 126. S. SEETHARAMU and M. N. SRINIVASAN: Trans. AFS, 1984,92,
Soc.), 1980, 52, 76-87. 539.
101. S. NISHI and T. KOBAYASHI: in Proc. 4th Int. Conf. on 127. S. SEETHARAMU and M. N. SRINIVASAN: Trans. AFS, 1983,91,
'Strengthening mechanisms in metals and alloys', Nancy, 867.
France, 1976, Vol. 2, 524-528. 128. s. SEETHARAMU: PhD thesis, Indian Institute of Science,
102. T. KOBAYASHI: Nihon Kinzoku Gakai Kaiho (Bull. Jpn Inst. 1981.
Met.), 1979, 18, 512-520. 129. S. SEETHARAMU and M. N. SRINIVASAN: in 'Fracture mechan-
103. G. CAIRONI, c. A. GORIA, G. SILVA, and G. TOSI: Unpublished ics in engineering applications', (ed. G. C. Sih and S. R.
work. Valluri), 165; 1979, The Netherlands, Sijthoff and
104. Y. CHENG, R. W. HEINE, and F. J. WORZALA: Trans. AFS, Noordhoff.
1979, 87, 17-3~. . / 130. T. KOBAYASHI: Tetsu-to-Hagene (J. Iron Steel Inst. Jpn) ,
105. c. E. TURNER: m 'Impact testmg of metals', STP 466; 1970, 1973, 59, 1578-1591.
Philadelphia PA, American Society for Testing and 131. N. L. CHURCH: 'Bainitic ductile irons', Technical Paper
Materials. 816- T-OP, Suffern, NY, The International Nickel Com-
106. F. SHIH: in Meeting on Nonlinear Fracture Mechanics, pany, Sept. 1972.
Knoxville, TN, 1986, American Society for Testing and 132. 1. A. DILEWIJNS: AFS Cast Met. Res. J., June 1973, 9, 64.
Materials. 133. c. F. WALTON and T. 1. OPAR (eds.): 'Iron castings hand-
107. w. L. SERVER: J. Test. Eva!., 1978, 6, 29-34. ~ook', 354; 1981, Des Plaines, IL, Iron Castings Society.
108. R. K. NANSTAD, F. 1. WORZALA, and c. R. LOPER: Trans. AFS, 134. 1. SMITH, L. BROWN, and H. L. MARCUS: Trans. AFS, 1980,
1975, 83, 245-256. 88, 427-436.
109. F."1. WORZALA and Y. W. CHENG: in Proc. Conf. on Dynamic 135. R. B. GUNDLACH and 1. F. JANOWAK: Met. Prog., July 1985,
Fracture, London, 1977, American Society for Metals and 132, 19-26.
the Welding Institute. Vol. 1, 227-235. 136. G. BARBEZAT and H. MAYER: Sulzer Tech. Rev., 1983, (2),
110. F. R. TULER, 1. ALESZKA, D. R. IRELAND, R. ABBOTT, and L. 23-28. .
JENKINS: in Proc. Conf. on Dynamic Fracture, London, 1977, 137. c. R. LOPER, M. J. LAUCH, H. K. PARK, and A. M. GYARMATY:
American Society for Metals and the Welding Institute. in Proc. 46th Int. Foundry Congress, Madrid, Spain, 1979,
Vol. 1, 217-226. " CIATF.
111. s. T. ROLFE and 1. M. BARSOM: 'Fracture and fatigue control 138. K. z. GAHR a"nd D. V. DOANE: Metall. Trans., 1980, l1A,
in structures: Applications of fracture mechanics'; Engle- 613-620.
wood Cliffs, NJ, Prentice-Hall. 139. K. Z. GAHR and w. G. SCHOLZ: J. Met., Oct. 1980,32, 38-44.
112. R. BARTON: 'Graphite nodule number and distribution - 140. K. ERIKSSON: Scand. J. Metall., 1973, 2, 197-203.
their effects on nodular cast iron', Report No. 1304, 141. I. R. SARE: Met. Techno!., 1979, 6, 412-419.
British Cast Iron Research Association, Redditch, 1966.

Fracture and Fracture Toughness of Cast Irons: W. L. Bradley and M. N. Srinivasan

Uploaded by

Copyright:

Available Formats

Fracture and Fracture Toughness of Cast Irons: W. L. Bradley and M. N. Srinivasan

Uploaded by

Document Information

Original Title

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

Fracture and Fracture Toughness of Cast Irons: W. L. Bradley and M. N. Srinivasan

Uploaded by

Copyright:

Available Formats

Fracture and fracture toughness of cast irons

W. L. Bradley and M. N. Srinivasan

relatively wear resistant because the graphite acts as

Table 1 Range of compositions for typical

White 1·8-3·6 0·5-1·9 0·25-0·8 0·06-0·2 0·06-0·2

microstructural constituents which determine the

1.0 2.0 3.0 4.0 5.0 2.0 3.0 4.0 5.0

Carbon Content, wt - °/0 Carbon Content ,wt-o/o

International Materials Reviews 1990 Vol. 35 No.3

6 Microstructure of compacted graphite cast iron,

International Materials Reviews 1990 Vol. 35 No.3

matrix. Type E graphite forms in strongly hypo-

as cast iron that has been appropriately alloyed with

o 120 160 200 240 280 320

specimens or fatigue precracked Charpy specimens

International Materials Reviews 1990 Vol. 35 No.3'

(ASTM E399-81);29 (b) the I-integral elastic-

~.•..• fracture toughness of the given material. Compari-

size, shape, and volume fraction of graphite; and on

19 Effect of phosphorus content on Charpy V-notch

Sulphur Table 3 Comparison of Charpy type impact values

* %P refers to phosphorus content.

International Materials Reviews 1990 Vol. 35 NO.3

23 SEM observation of fractured surface of ferritic

random nature of fracture in these systems.

irons usually giving higher values for the upper

fracture mechanics tests. This unusual behaviour

matrix is ductile. 24 Load-displacement curve during fracture test of

o Ka (Grover and POllard)55 ° Kmax (Tait and Spencer)56 200

International Materials Reviews 1990 Vol. 35 No.3

mens (a/W = 0,6). The results are presented in E 260

Table 4 and in Fig. 28 where KJrnax values (assumed ~

International Materials Reviews 1990 Vol. 35 NO.3

200 in energy absorbed under this condition more

'" are very reasonable, given that the previously cited

International Materials Reviews 1990 Vol. 35 No.3

20 + Code 8 (183 nodules mrii2) ~

International Materials Reviews 1990 Vol. 35 No.3

• • . . results by Bradley and Tanner33,95 indicated a drop

Charp~ specimens as well as dynamic tear speci- 40

graphite (Fig. 32a). Increasing the nodule count (f) 20

studies are combined in Fig. 42-along with those of

Holdsworth and Jolley.84 It should be noted that o

gave artificially high figures because of the use of

which was nearly the same at -19 and -59 C. In

temperatures. Also the K values for the }att~r ~t

-19°C were significantly lower than at 24 C, IndI-

International Materials Reviews 1990 Vol. 35 NO.3

matrix gives a better Charpy toughness than a 20

yield strength of 740 MN m-2. They also obtained c

International Materials Reviews 1990 Vol. 35 No.3

00% Pearlite, annealed.2 • 0% Pearlite, annealed.2

obtained for ferritic permanent mould and sand f(

of ductile iron. 137 en 031

International Materials Reviews 1990 Vol. 35 NO.3

o , I ,/ / / /. quenched and tempered

-100 -75 -50 -25 25 50 75

International Materials Reviews 1990 Vol. 35 No.3

International Materials Reviews 1990 Vol. 35 No.3