flakes. Inset: broken grinder in as-received condition. 1.5 (C.-A. Baer, California Polytechnic State University)

Fig. 30 Fig. 31

Fig. 32 Fig. 33

Fig. 34 Fig. 35

Effect of graphite morphology on fracture of permanent mold cast, hypereutectic gray iron (4.5 to4.8% carbon equivalent). Permanent mold castings are typically small--15 kg (30 lb) max--and thustheir as-cast mechanical properties are particularly sensitive to solidification structure and size, shape,and distribution of graphite. In this study, cylindrical samples were cast in a cold, 230 C (450 F)mold and allowed to air cool. Typical microstructure: undercooled (type D) graphite in a ferritematrix. Initially cracked (nil ductility) and crack-free (normal ductility) samples were compared. Fig.30 and 31: Fracture surface of nil-ductility sample that cracked during casting or machining. Fractureinitiated at and propagated from graphite-ferrite interfaces. SEM, 190 and 1900. Fig. 32 and 33:Solidification structure and graphite morphology of permanent mold cast, gray iron sample with nilductility. The crack propagated along the preferred orientation of solidification. Dendrite spacing wasa narrow 50 m. Graphite was fine tipped, roughly cylindrical, and isolated from ferrite cells--amorphology that apparently has an adverse effect on ductility and is also a major contributor tographite-ferrite interface cracking. 2% nital, 82 and 330. Fig. 34 and 35: Fracture surface of samplehaving normal ductility. Fracture was artificially generated by impact. Note how material resistedgraphite-ferrite interface cracking. Microstructure (not shown) was more isotropic than that of the nil-ductility casting, with a wider dendrite spacing (85 m). Graphite was medium sized, interconnected,

and penetrated ferrite cells. SEM, 200 and 1000 (D.C. Wei, Kelsey-Hayes Company)

Ductile Irons: Atlas of Fractographs

Ductile Irons

Fig. 36 Fig. 37

Fig. 38 Fig. 39 Fig. 40

Brittle cleavage fracture of ductile iron spur gear (ASTM A536, grade 100-70-03) due to improperheat treatment. Tensile strength was 544 MPa (78.9 ksi), much less than the 690 MPa (100 ksi)required by the specification. Elongation was nil; the specification called for 3% min. The induction-hardened case on the teeth was shallower and harder than specified (50 + HRC versus 46 HRC), andthe martensitic microstructure had not been tempered as specified. Direct cause of fracture was thepresence of inverse chill (carbides in thick sections of the casting) associated with microporosity. Thiscarbidic material, which formed at thermal centers due to segregation of carbide-forming elements,increased hardness, decreased tensile strength, and promoted brittle fracture. Proper heat treatmentswould have corrected the deficiencies in case structure and properties and would also have preventedthe occurrence of inverse chill. Fig. 36: Fracture surface at core of gear directly below origin. Notetranscrystalline (cleavage) mode of fracture. SEM, 200. Fig. 37: Photomicrograph of core. Matrix is100% pearlite. Note presence of inverse chill (carbides) and associated porosity. 2% nital, 200. Fig.38: Fracture face at casting surface and in the induction-hardened case. SEM, 20. Fig. 39: Boxed areain Fig. 38. Note cleavage fracture appearance and nodule surrounded by cracked material believed tobe carbidic. SEM, 200. Fig. 40: Microstructure directly below fracture surface and in case. There is acarbide-appearing envelope around each nodule of temper carbon that formed as the inverse chill

decomposed. The two nodules at upper left also have retained austenite around them. Note thesecondary crack extending between the casting surface and the carbide envelope of a temper nodule.2% nital, 200 (G.M. Goodrich, Taussig Associates Inc.)

Fig. 41 Surface of a fatigue-test fracture in an experimental crankshaft of induction-hardened 80-60-03 ductileiron with a hardness of 197 to 225 HB. Fatigue-crack origin is at arrow A. Porosity at arrow B was unrelated tofracture initiation. 2.5

Fig. 42 Surface of a fatigue-test fracture in an experimental crankshaft of ductile-iron with a hardness of 241to 255 HB. Note the multiple fatigue-crack origins at the journal edge (at right). Fatigue beach marks areevident, which is unusual in cast iron. Actual size

Fig. 43 Surface of a fatigue fracture in an experimental crankshaft broken in a fatigue test. The material isductile iron with a hardness of 241 HB. The origin of the fatigue crack is at the edge of the journal, at arrow.Actual size

Fig. 44

Fig. 45

Fig. 46

How fatigue cracks propagate through ductile irons. Fig. 44 and 45: In an as-cast, commercial pearliticductile iron, the crack changes direction from grain to grain, usually following nodule-matrixinterfaces. Fig. 45 etched in 2% nital, both at 80. Fig. 46: Crack propagation through the temperedmartensite of a heat-treated ductile iron is more matrix controlled than in either a pearlitic (Fig. 44and 45) or ferritic microstructure. 100 (F.J. Worzala, University of Wisconsin)

Fig. 47 Fig. 48 Fig. 49

Fatigue fracture surfaces of pearlitic and ferritic ductile irons. Compositions of pearlitic irons: 3.63 to3.80% C, 0.34% Mn, 2.02 to 2.66% Si. Compositions of ferritic iron: 3.75 to 2.82% C, 0.34% Mn, 2.30to 2.66% Si. Striations noted in all cases. Fig. 47 and 48: Mixture of striations and fractured pearlitelamellae on fracture surfaces of commercial pearlitic ductile irons. Striations are the fine steplikefeatures, not the macroscopic waviness or undulations. SEM, 207 and 198. Fig. 49: A high loadfatigue fracture surface of a ferritic ductile iron. SEM, 375 (F.J. Worzala, University of Wisconsin)

Fig. 51Fig. 50

Comparisons of fatigue and monotonic (tension, bending, impact) fracture surfaces for various ferriticductile iron microstructures. Fatigue fractures are characterized by striations (see Fig. 47, 48, and 49),by relatively little opening up or stretching of matrix material around nodules, and by nodules thatappear to have had large pieces of graphite broken off of them. Mode of crack propagation inmonotonic fractures is very ductile with considerable stretching of nodule-bearing cavities. Thedemarcation line between fatigue (FAT) and monotonic (FRA) fracture is noted in each of the threefractographs. Fig. 50: Commercial ferritic iron (2.30 to 2.66% Si) tested at room temperature. SEM,90. Fig. 51: Same material as in Fig. 50, but during tensile portion of test at -40 C (-40 F). SEM,90. Fig. 52: High-silicon ferritic ductile iron (3.5% Si). SEM, 90 (F.J. Worzala, University of