Latest from Issues and Ideas

Jill Lehmann | Dreamstime
Sands through and hourglass.

Permanent Mold Casting of Ductile Iron ? Part II

Nov. 28, 2003
In this overview, the permanent mold process to cast ductile iron is examined in its various aspects, including process automation, heat treatment, gating system design, production of large castings, and more. This is the second of three parts.
Heat Treatment

One operational aspect of the permanent mold casting of DI is the required high temperature heat treatment of castings. The typical as-cast microstructure of PM-cast DI includes iron carbides and a large amount of small graphite nodules resulting from rapid cooling. Heat treatment is necessary to decompose the iron carbides and to ensure and regulate the reproducibility of the microstructure and mechanical properties of the casting. It is also essential for attaining the good machining properties of finished castings. Table 1 gives typical heat treatment cycles for PM-cast ASTM standard grades of DI to obtain a desirable combination of microstructural and mechanical properties.

As can be seen, common for all grades of ductile iron is austenization at 920-940º C (1688-1722º F) for 1-4 hrs. (the lower limit refers to thin-walled castings; the upper to thick-walled parts), and then cooling at a different rate. Figure 12 illustrates the typical graphite and metallic structure in the center of cylindrical sample of 0.75 in. dia. after high temperature normalizing. Due to graphitization during austenization, as a result of decomposition of iron carbides, the quantity of graphite nodules and their diameter are slightly increased. In comparison with the as-cast microstructure, with increasing casting modulus the casting cooling rate in the furnace will decrease, and ferrite content may dominate even after normalizing, without the use of pearlite stabilizers as shown in Figure 13 (right). Note the decarburized layer of ferrite on the surface of the specimen [Figure 13 (left)] resulting from heat treatment in the furnace without a protective atmosphere.

For ferritic grade DI 60-40-18 (FDI), a two-stage ferritizing anneal is typically applied, which also includes austenization at 920-940º C (1688-1722º F) for 1-4 hrs.; cooling in the furnace to 720-740º C (1329-1364º F) and soaking at this temperature for 1-4 hrs.; slow cooling in the furnace to 600-620º C (1111-1148º F) and further cooling in air.

When castings are slow-cooled in the furnace after annealing through the range of 600º to 400º C (1112 and 752º F) or castings are cooled in the mold after pouring, a significant reduction in Charpy impact strength at room temperature, commonly known as the temper embrittlement (TE) of FDI, can result. Typically, TE does not affect tensile strength and elongation, but significantly impairs impact strength of FDI. This phenomenon is similar to that observed for years in ferritic malleable iron, ferritic chromium steels, and in all austenitic chromium-nickel steels.


Chemistry Optimization

In order to reduce chill tendency and shrinkage-related defects, it is recommended that a relatively high carbon equivalent (CE) in PM cast ductile iron be maintained. Increasing CE also reduces the time needed for austenization and the overall length of the two-stage annealing cycle.

Increasing CE may be done in two ways: increasing either the carbon or the silicon content. The silicon in ferritic ductile cast irons is primarily an alloying element. If silicon content is increased above a certain range, ductility and toughness decrease. The metallic matrix becomes a silicoferrite, which is less ductile than ordinary ferrite, and failure occurs in regions with the highest silicon concentration.

In contrast, many researchers associate TE with the concentration of elements such as silicon, phosphorous, and manganese on grain boundaries. A special study was performed to optimize Si and P content in PM cast DI.

Standard 25mm Y- blocks cast by the PM process, as shown in Figure 14 prior to cutting, were subject to two stages of annealing. After rough machining, the Charpy impact bars from each heat were heat treated at 600º C (1112 F) for 2 hrs. and cooled in the furnace to 400º C (752º F) at three different velocities of 30, 85, and 150º C/ hr.

Test specimens then were machined to a size slightly larger than that of the final impact test bar. After heat treatment, the bars were machined to their final dimensions. The Charpy impact tests were performed on the standard un-notched rectangular (10 x 10 x 55 mm) bars fractured with a pendulum hammer. The fracture energy and fracture appearance were recorded as a function of phosphorous or silicon content in specimens at various cooling rates after tempering. Each point on the graphs is a mean value of 8-12 tests.

Fractured Charpy bars were used for microstructural investigations. Standard metallographic techniques were employed to reveal the nodule count, nodularity, and microstructural constituents. Scanning electron microscopy (SEM), along with X-ray microprobe analysis, was used to study fracture surfaces, as well as the distribution and concentrations of phosphorous, silicon, and manganese in grain boundaries and within the grains.

The first series of experiments was designed to investigate the level of silicon content that would not make FDI susceptible to TE, while maintaining relatively high phosphorous content of approximately 0.11 percent. In other words, to exclude the factors related to silicoferrite brittleness. Two sets of samples with gradually increased silicon content from approximately 2.66 to 3.86 percent and with approximately 0.11 percent phosphorous were subjected to cooling through the 600-400º C (1112-752° F) range in air (series 1) and in the furnace at a velocity of about 150º C/hr. (series 2).

Results of the impact tests were plotted against silicon content depending on cooling rate after tempering and are presented in Figure 15. As can be seen, impact strength decreased when silicon content increased rapidly from 3 to 3.2 percent, then relatively slowly at higher silicon levels for both series. Based on these results, the silicon content in subsequent heats was narrowed to the range of 2.7-2.9 percent.

Figure 16 illustrates the effect of phosphorous and cooling rate on the impact strength of ductile iron containing approximately 2.8 percent silicon, tempered at 600º C and cooled through the 600-400° C range at three different velocities: 30º C/hr., 85º C/hr., and 150º C/hr. These data confirm the trend of reduced impact strength with increased phosphorous content. A slight reduction of impact strength was noticed in specimens with relatively low (0.04-0.07 percent) phosphorous content. Ductile iron containing phosphorous (0.08-0.12 percent) produced a steady reduction of impact strength. Impact strength was drastically reduced in specimens containing 0.09 to 0.12 percent phosphorous and cooled at 30º C/hr. DI containing phosphorous from 0.04 to 0.08 percent cooled at the same velocity did not reveal TE.

It is noteworthy that only a very slow cooling rate through the 600-400º C range (approximately 30º C/hr.) generated TE, while cooling at higher velocities of 85 and 150º C/hr. did not. In other words, the TE of ductile iron with high phosphorous content may not occur if the cooling rate through 600-400º C intervals are greater than approximately 85º C/hr.

The addition of calcium or cesium (Figure 17) couldn’t prevent the TE of DI with high phosphorous content; only slight reductions of impact strength have been observed in DI heats treated with Ca-containing additives.

To study the reversal of TE, eighteen Charpy DI test specimens were made containing approximately 0.11 percent P. Six of these were tested before heat treatment and twelve were subjected to embrittling heat treatment. Six of the twelve treated specimens were tested at their embrittled stage and six were repeatedly tempered at 600º C and cooled rapidly at 150º C/hr. Results of these tests (Figure 18) illustrate the reversal phenomena of TE. The impact strength after repeated tempering and rapid cooling at approximately 150º C/hr. was restored to approximately the same value as that prior to the embrittling heat treatment.

It is noteworthy that these results were obtained while maintaining relatively high silicon content (2.7-2.9 percent). This procedure may be applied by DI foundries as a corrective action when TE occurs in ultra-thin wall sand, centrifugal, and PM castings.

From an economics perspective, when ductile iron parts are not intended for use in a low temperature environment, it is not necessary to maintain a low phosphorous content in the base iron and to use expensive low phosphorus raw materials.


Mechanical Properties

The goal of this study was to statistically prove the stability of DI’s major mechanical properties and to establish a correlation between tensile strength and hardness in order to predict these values in permanent mold DI castings.

Mechanical properties were tested on standard test specimens 12.5 mm (0.5 in.) dia. with the 50 mm (2 in.) gauge length machined from standard 25 mm (1 in.) Y- blocks cast by the PM process. Prior to machining, the Y-blocks were heat treated by applying different heat treatment cycles to develop specific microstructures and mechanical properties. The Brinell hardness (HB) values and microstructural analyses were obtained from the tensile test specimens. A statistical analysis of the experimental data was performed using dispersion and correlation methods.

Figures 19 and 20 show cumulative probability plots of tensile strength and hardness for the PM-cast ductile iron specimens. Straight plot lines indicate that the distributions are normal. Tensile strength has a normal distribution when the hardness value is 207; the mean of the distribution (at the frequency of events of 0.50) is approximately 580 MPa. The standard deviation is calculated as the difference between the 0.5 and 0.159 of the frequency of events axis. Thus, the parallelism of distribution is evidence of equal deviations of studied mechanical properties.

The mean values and standard deviations for each of the hardness curves are presented in Table 2. As can be seen, the tensile strength value for the DI, which had a Brinell hardness of 207, has a mean of 582 and a standard deviation of 34.1 versus the graphically determined values of 580 and 35 MPa respectively.

Figure 21 indicates the general trend of decreasing elongation with increasing tensile strength in DI’s. Also, as the hardness increases, the elongation decreases.

It is interesting to note the influence of graphite nodularity on tensile strength and elongation of both ferritic and pearlitic irons: when graphite nodularity increases, both the tensile strength and the elongation increases, but the effect is more pronounced in pearlitic DI.

The relationships between hardness and tensile strength are presented in Table 3. The value of the correlation coefficient is relatively high, so the variance will be relatively low when using hardness to predict tensile strength.

The inaccuracy of the relationships does not permit an exact evaluation with high confidence, but for a given hardness value, a confidence level can be predicted for a range of tensile strengths. Based on information developed, two different approaches can be used. One is to calculate the probabilities using statistical relationships, and the second is to directly read the results from the charts. Both approaches will be applied to an example problem in next month’s issue.