Environmental Embrittlement of Ductile Irons - A Review of Available Data

by Martin Gagné, Manager - Sorelmetal Technical Services
Rio Tinto Iron & Titanium
and Kathy L. Hayrynen, Technical Director
Applied Process Inc.

1.   INTRODUCTION
The generic definition of embrittlement is given by Krauss⁽¹⁾ as the “Reduction in the normal ductility of a metal due to a physical or chemical change”. Most common examples of such a phenomenon are blue brittleness, hydrogen (in situ) embrittlement and temper brittleness. While the phenomena listed above referred to changes, chemical or physical, occurring within the metal structure, this definition can be extended to embrittlement resulting from changes generated by the interaction between the metal and its environment. These types of embrittlement are identified as hydrogen embrittlement (hydrogen originating from the environment), stress-corrosion cracking and liquid metal embrittlement.

Ductile Irons have been recently reported to be prone to environmentally assisted embrittlement. The objective of this article is to review the data published on this subject^{(2, 3, 4, 5, 6)}, with a particular emphasis on “Austempered Ductile Irons”.

2.   REVIEW OF ENVIRONMENTALLY ASSISTED EMBRITTLEMENT MECHANISMS
As will be seen in the following sections, it is difficult to pinpoint what mechanism (or interaction of mechanisms) causes the embrittlement of Ductile Irons in liquid environment. A quick review of the known mechanisms is presented⁽⁷⁾ to allow the reader to evaluate the possible relationship of such phenomena with the reported data.

2.1   Hydrogen Embrittlement
Embrittlement may occur in many metals in presence of a very small amount of hydrogen. Hydrogen may be introduced during melting and entrapped during solidification, may be picked-up during solid state processing (heat treatment, plating, welding…) or introduced by a cathodic reaction during corrosion. Hydrogen is present in the metal as monoatomic hydrogen due to the dissociation of molecular hydrogen by chemisorption at the surface (or as a result of oxidation reaction (corrosion)).

Back to top

As illustrated in Figure 1, hydrogen causes delayed fracture at a stress level significantly lower than the strength of the metal. The fracture process may be cleavage, intergranular or transgranular and no single fracture mode is characteristic of hydrogen embrittlement. There is no univocal mechanism for hydrogen embrittlement. One of them proposed that hydrogen in relatively small amount decreases the cohesive strength in regions of high stresses.

Fig. 1. Delayed Fracture Curve Caused by Hydrogen Embrittlement⁽⁷⁾

2.2   Stress-Corrosion Cracking
Stress-corrosion cracking (SCC) is the failure of an alloy from the combined effects of a corrosive environment and a static tensile stress (applied or residual). The chemical environment causing SCC does not produce chemical corrosion of the alloy and the species causing SCC need not be present in large concentration. The formation and rupture of a passive layer at the crack tip is an important mechanism. Moreover, it is widely believed that electrochemical dissolution plays a major role in the crack initiation and propagation. There is a possibility of the adsorption of damaging ions that weaken the atomic bonding at the crack tip. If hydrogen is generated as a result of corrosion species, it is then able to enter the metal, diffuse to the crack tip and cause crack propagation. There is growing evidence of a link between SCC and hydrogen embrittlement.

Figure 2 shows a schematic of crack growth under constant stress in a corrosive environment. Note that the threshold stress is lower than yield strength and that the resulting fracture pattern may be intergranular, transgranular or a mixed mode.

Fig. 2.  Schematic of the Three-Stages Environmentally Assisted Cracking under Constant Load in an Aggressive Environment⁽⁷⁾.

2.3   Liquid Metal Embrittlement
Liquid metal embrittlement (LME) occurs when a solid metal surface is wetted by a lower melting point metal and results from the direct interaction of the liquid metal atoms with the highly strained atoms at a crack tip. Adsorption of atoms from liquid metal greatly reduces the surface energy and though this, the fracture stress is reduced. The prerequisite for the occurrence of LME are: i) a good intimate contact or wetting between the surface of the solid metal and the liquid metal; ii) an applied or residual stress; iii) some measure of plastic flow and some obstacle to dislocation motion (plastic flow) at the solid-liquid interface. Other factors reported to promote LME are the presence of a sharp notch, a high strain rate, coarse grain size and temperature⁽⁸⁾.

Back to top

3.   ENVIRONMENTALLY ASSISTED EMBRITTLEMENT OF DUCTILE IRONS
Table 1 lists the various Ductile Iron grades investigated for environmentally assisted embrittlement by various authors. Also included are unpublished data recently obtained by Hayrynen and Boeri⁽⁹⁾. All information presented in the following sections were sourced from these references.

TABLE 1
Ductile Iron Grades Investigated for Environmentally Assisted Embrittlement

Figure 3 compares the stress-strain behavior of normal and embrittled materials. The material property mainly affected is the ductility which is significantly reduced when embrittlement occurs. Fracture toughness behaves similarly to elongation and was also used to quantify the embrittlement phenomenon. Data using both properties will be used in the following sections.

Fig. 3. Typical Stress-Strain Behavior of Normal and Embrittled Materials⁽²⁾.

Note that the different authors may have utilized different experimental parameters and/or investigated the effect of different variables on the embrittlement. The presented analysis tried to take these differences into consideration. However, please refer to the original papers if more details are needed.

3.1  ASTM-A536 Ductile Iron Grades

· Ferritic Ductile Iron
Figure 4 presents the results of the investigation carried out by Druschitz et al⁽⁶⁾ on 65-45-12 ferritic Ductile Iron tested at slow strain rate (1% per minute) at room temperature, in contact with various liquid environments. Under these test conditions, no embrittlement was detected. Three other studies^{(2, 3, 4)} reported similar results, either by measuring tensile properties or fracture toughness. It has been clearly shown by Komatsu et al⁽²⁾ that the occurrence of residual pearlite (~10%) does not affect the behavior of ferritic irons vis-à-vis the embrittling effect of liquids.

Back to top

Fig. 4. Effect of Liquid Environments on the Elongation of Ferritic Ductile Iron ⁽⁶⁾ (strain rate 1% per minute).

· Pearlitic Ductile Irons
The data reported on the embrittlement of pearlitic Ductile Irons in contact with water is presented in Table 2.

TABLE 2
Effect of Water Environment on the Elongation of Pearlitic Ductile Irons (low strain rate)

It is seen that embrittlement of pearlitic Ductile Iron is reported by Shibutani et al ⁽³⁾ and Martinez et al ⁽⁴⁾, but not by Druschitz ⁽⁶⁾. It is worth noting that Komatsu ⁽²⁾ also observed the embrittlement phenomenon when measuring fracture toughness parameters.

A review of the process route used to produce the pearlitic Ductile Irons tested revealed that Druschitz ⁽⁶⁾ tempered his specimens at 565^oC (1050^oF) for 2 hours while the other investigators ^{(3, 4)} tested them as normalized. The tempering is carried out to reduce the hardness of the iron as shown in literature ⁽¹⁰⁾ and is accompanied by a higher ductility of the structure. Druschitz reports 7.5% elongation after tempering, a high value when compared to normalized/non tempered or as-cast pearlitic castings. The description of the microstructure by Druschitz (fine pearlite) does not allow to conclude on the microstructural changes taking place during tempering that would explain the increase in ductility. However, it appears unlikely that significant decomposition of pearlite to ferrite + graphite occurred at the tempering temperature according to data reported in literature ⁽¹¹⁾. However, pearlite spheroidization, which is the precursor step to pearlite decomposition and results in a quasi-continuous ferrite network, is known to occur in less than two hours at 600-620^oC in 3.5% C, 2% Si Ductile Iron and may explain the different embrittlement behavior reported by Druschitz. Spheroidization of pearlite during tempering for two hours at 565oC is considered probable by Mullins ⁽¹³⁾.

·  Quenched and Tempered Ductile Iron
Figures 5 a) and b) present the results obtained by Druschitz ⁽⁶⁾ and Komatsu ⁽²⁾ on quenched and tempered Ductile Irons. While Druschitz did not report the embrittlement of Q&T Ductile Iron, whatever the fluid used, severe embrittlement (measured by fracture toughness) was observed by Komatsu ⁽²⁾. Note that Shibutani did not report embrittlement when carrying out tensile test, probably because of the brittle behavior of the Q&T specimens used (no plasticity).

Back to top

Fig. 5 a)

Fig. 5 b)

Fig. 5. Effect of Liquid Environments on the Properties of Quenched and Tempered Ductile Iron

a) Elongation (Druschitz ⁽⁶⁾)
b) Fracture toughness (Komatsu ⁽²⁾)

The heat treatments used in the two studies were, however, different, the major difference being the tempering treatment, Table 3. The treatment used by Druschitz resulted in a tempered martensite structure in which carbides are fully spheroidized in a ferritic matrix while the one used by Komatsu probably partly retained the original plate-like martensitic structure.

TABLE 3
Tempering Treatments Used for Q&T Specimens

3.2 Austempered Ductile Irons
Various grades of Austempered Ductile Iron were studied for environmentally assisted embrittlement. Table 4 identifies them and the investigators.

TABLE 4
ADI Materials Evaluated

· ASTM-897M Grades
All ASTM-897M Austempered Ductile Iron grades (1 to 4) were embrittled when tested, either for tensile properties or fracture toughness. As examples, data reported by Martinez et al ⁽⁴⁾, Druschitz et al (6) and Hayrynen et al ⁽⁷⁾ are presented in Figures 6 and 7 and in Table 5. Table 6 compares the various relative reduction in elongation measured by each author.

Back to top

Fig. 6. Effect of Liquid Environments on the Elongation of ASTM-897M Grade 1 (Druschitz et al ⁽⁶⁾)

Fig. 7. Effect of Water Environment on the Elongation of ASTM-897M Grades 2, 3 and 4 (Martinez et al ⁽⁴⁾).

TABLE 6
Effect of Water Environment on UTS and Elongation of ASTM 897M Grades 1 and 3
(Hayrynen and Boeri ⁽⁸⁾)

TABLE 7
Relative Reduction of Elongation for ASTM-897M ADI Materials when in Contact with Water

Back to top

It is seen in Table 7 that the reduction of ductility, as measured by tensile elongation, is about 70 to 80 %. The only data differing from the others is the one reported by Hayrynen and Boeri for ADI grade 3. However, the initial elongation value was 8%, while data reported by Martinez et al mention about 11% as “non-embrittled value”. The rupture in wet environment occurred at the same value, i.e. 4%, in both studies.

It is interesting to note that all grades fractured at 3-4% elongation in all studies, with only a few exceptions. Finally, it is also worth noting that the embrittlement by fluids other than water reported by Druschitz (6) was also reported by Martinez et al (5).

· SAE-J-2477-750 Grade
This grade of ADI, also referred as machinable ADI, was investigated by Druschitz et al ⁽⁶⁾ and Hayrynen and Boeri ⁽⁷⁾. Druschitz investigated SAE-J-2477-750 materials with different hardnesses; for comparison purpose with Hayrynen’s data, the data on the more ductile material (247 BHN) is presented in Figure 8, while Hayrynen’s results are listed in Table 8.

Fig. 8. Effect of Fluid Environments on the Elongation of SAE-J-2477-750 ADI (Druschitz et al ⁽⁶⁾)

TABLE 8
Effect of Water Environment on the Tensile Properties of SAE-J-2477-750 ADI (Hayrynen and Boeri ⁽⁷⁾)

When in contact with water, both studies report embrittlement of the material. However, as shown in Table 9, the “embrittled” value measured by Druschitz is higher and the relative loss of ductility significantly less than those reported by Hayrynen. Such a difference is believed to be related to differences in microstructures. These irons being austenitized in the inter-critical temperature zone, the matrix obtained prior to austempering is a mixture of ferrite and austenite, the ferritic phase being not affected by the austempering reaction. According to Druschitz ⁽¹²⁾, pro-eutectoid ferrite appeared mainly as a network in the final structure of its specimen, while it was more fragmented in the Hayrynen’s samples ⁽⁹⁾. This may explain the different results reported by the two authors.

Back to top

TABLE 9
Comparison of Data Reported on the J-2477-750 ADI Grade

It is worth noting that Druschitz ⁽⁶⁾ reported no embrittlement when in contact with liquids other than water base fluids.

4. DISCUSSION
The review of the data available on the environmentally induced embrittlement of Ductile Irons clearly shows that all grades, with the exception of ferritic irons, are sensitive to the phenomenon. The mechanisms by which this phenomenon occurs are, however, not clearly identified. Nevertheless, the following observations from the published papers, together with the previously reported data, may help to understand the phenomenon.

1. The phenomenon is not a grain boundary embrittlement or related to the presence of graphite. Ferritic Ductile Irons, which exhibit a large area of grains boundaries and the highest fraction of graphite are not sensitive to the phenomenon.

2. The phenomenon is independent of the time spent in contact with the liquid and disappears when the liquid is removed, as shown by Komatsu ⁽²⁾ and Martinez ⁽⁴⁾.

3. Increasing strain rate ^{(2, 4, 6)} reduces the extent of the phenomenon, which is time sensitive and probably involves time controlled phenomena (diffusion, chemical reaction, adsorption …).

4. The phenomenon is observed in structures with extended phase interfaces (pearlite, tempered martensite, ausferrite).

5. Shibutani ⁽³⁾ reported that similar embrittlement occurs when tensile test are carried out in a hydrogen atmosphere.

6. Forming a ferritic layer at the surface of the specimens prevents embrittlement ⁽³⁾, as does painting the component ⁽³⁾.

Two mechanisms for environmentally assisted embrittlement of Ductile Irons have been proposed in the literature: Hydrogen embrittlement ⁽³⁾ and Chemisorption ^{(5, 6)} . At this stage, there is no clear evidence that would favour one theory via-à-vis the other. However, whatever the mechanisms involved, the presence of a high density of phase interfaces appears unavoidable for the phenomenon to occur. These phase boundary zones are privileged sites for hydrogen pick-up and diffusion or for adsorption of atoms or molecules, both phenomena known to cause embrittlement. It is worth noting, however, that chemisorption (or Liquid Metal Embrittlement) is favored by high strain rate, while the opposite was observe for embrittlement of Ductile Irons.

The presence of ferrite, either as a fully ferritic matrix, as a quasi-continuous phase in a composite matrix (spheroidized pearlite, fully tempered martensite) or as a continuous network (J-2477-750 ⁽⁶⁾) probably prevents the embrittlement phenomenon by deforming plastically under the applied stress, avoiding the formation and propagation of cracks.

5. CONCLUDING REMARKS
The review of available data on the environmentally induced embrittlement of Ductile Irons indicates that high strength Ductile Iron grades (pearlitic, Q&T and ADI) are sensitive to the phenomenon. When these grades are treated in order to get a continuous (or quasi-continuous) ferritic matrix (by tempering or inter critical austenitization of ADI), the phenomenon fades.

The review of the data also evidences that three factors have to be present to observe the environmentally assisted embrittlement of high strength Ductile Irons:

1. Presence of a liquid in contact with the material;

2. Applied stress approaching yield strength of the material;

3. Low strain rate.

When designing a part, the “design safety factor” usually ensures that the component will not be loaded near the yield strength of the material. Nevertheless, when working in a liquid environment, additional precautions should be taken to ensure that the three above-mentioned factors are not found at the same time. Painting of the components should also be considered.

Back to top

REFERENCES

1. George Krauss, Steels: Heat Treatment and Processing Principles, ASM International, Materials Park, 1989.

2. S. Komatsu, C.Q. Zou, S. Shibutani and Y. Tanaka, “Embrittlement Characteristics of Fracture Toughness in Ductile Iron by Contact with Water”, Int. J. Cast Metals Res., vol. 11, 1999, pp. 539-544.

3. S. Shibutani, S. Kamatsu and Y. Tanaka, “Embrittlement of Austempered Spheroidal Graphite Cast Iron by Contact with Water and Resulting Preventing Method”, Int. J. Cast Metals Res., vol. 11, 1999, pp. 579-585.

4. R.A. Martinez, R.E. Boeri and J.A. Kora, “Embrittlement of Austempered Ductile Iron Caused by Contact with Water and Other Liquids”, Int. J. Cast Metals Res., vol. 13, 2000, pp. 9-15.

5. R.A. Martinez, S.N. Simison and R.E. Boeri, “Environmentally Assisted Embrittlement of ADI – Current Understanding”, Proceedings of the World Conference on ADI, Louisville, 2002, pp. 91-96.

6. A.P. Druschitz and D.J. tenPas, “Effect of Liquids on the Tensile Properties of Ductile Iron”, SAE Paper no. 2004-01-0793, SAE Conference, Detroit, 2004.

7. G.E. Dieter, Mechanical Metallurgy, 3rd Edition, McGraw Hill, New York, 1986.

8. A.K. Sinha, Ferrous Physical Metallurgy, Butterworths, Boston, 1989.

9. K. Hayrynen and R. Boeri, Unpublished results, May 2005.

10. B.V. Kovacs, “Heat Treatment” in Ductile Iron Handbook, American Foundrymen’s Society, Des Plaines, 1999.

11. M. Gagné, “The Combined Effect of Chromium and Manganese on the Thermal Stability of Eutectic and Eutectoid Cementite in Ductile Irons”, Canadian Metallurgical Quarterly, vol. 25, no. 3, 1986.

12. A. Druschitz, personal communication, June 2005.

13. J.D. Mullins, personal communication, July 2005.

Back to top