Use of Bismuth in Ductile Iron
Carola Mox, Grede Foundries, Inc., Reedsburg Division

Introduction
Many things have been published about the use of bismuth in ductile iron, especially in inoculants for ductile iron. The literature about this is quite confusing due to varying sampling procedures, sample castings, magnesium treatments and of course the addition of the bismuth itself.

The reported results include increase in nodule count, smaller nodules and better distribution, reduced chill, decreased magnesium recovery, and bad nodularity (exploded nodules).

This study was conducted primarily to investigate the behavior of bismuth treated iron and resulting microstructure under production conditions in a highly automated foundry.

We were looking at bismuth addition as part of our continuing efforts of creative iron improvements. The literature showed some interesting potentials of bismuth and our study confirmed all the positive effects. In the time during which bismuth was used, other improvement projects continued, and we learned that some of the effects that bismuth had on our iron can be accomplished by other operational changes. To eliminate this one extra step of adding the bismuth into the ladles, the decision was made to make operational changes other than continuing efforts on automating bismuth addition. At that point the addition of bismuth was discontinued.

Conclusion
A controlled addition of pure bismuth to ductile iron in conjunction with a magnesium treatment using MgFeSi with cerium is effective in increasing nodule count, improving nodule distribution, and eliminating carbides.

The correct Total Rare Earth (TRE)/Bi ratio is key to avoid degenerated graphite and will vary from one foundry to another.

Study
The main part of the study was conducted to compare results between commercially available bismuth and rare earth containing inoculant, and pure bismuth added in the form of tablets to the rare earth containing magnesium alloy used in the Tundish treatment.

During the second part of the study we experimented with different levels of pure bismuth addition and investigated the effects on the microstructure.

All experiments were carried out using a test casting with the configuration shown in Figure 1, cupola base iron treated with 5%MgFeSi (containing 1.35% Total Rate Earth) and instream inoculation (75%FeSi). Results were confirmed by looking at actual production castings.

The bismuth was added in tablet form into the Tundish ladle after adding the MgFeSi, so it was placed on top of the alloy. The bismuth tablets were about 15mm in diameter and 15mm long and weigh 18 grams. We sampled some different sizes and weights in the process of determining what addition rate to use, but concluded the above size and weight to be optimal for production.

Chill tendency was measured on chill wedges (see Figure 1), and nodule count was determined by counting in a defined field.

Fiqure 1. Chill wedge

Results
Several trials were conducted using different jobs with 0.005% bismuth addition to the Tundish ladle. All showed very good nodule count, good nodularity, significant reduction of chill in the chill wedges, decrease in magnesium recovery, and no difference in mechanical properties.

Table 1 shows the results of a parallel test on both Disa 2013’s. Both Disas are supplied with the same base metal which is treated in Tundish ladles with MgFeSi alloy. To Disa II 0.0055% bismuth was added before treatment, while Disa I operated without bismuth addition. Every half hour a chill wedge that was mounted on the pattern plate was cut, polished and the section size of the wedge containing the last carbide was reported. The average depth of carbides in chill wedges without bismuth (from Disa I) was 4.5mm, compared to the average depth of carbides in chill wedges with bismuth addition of 1.9mm.

When bismuth was added with the magnesium treatment alloy, the final magnesium was about 10% lower than usual. The produced castings didn’t show any effects in nodularity, and we didn’t change addition rates.

The Right Rare Earth/Bismuth Ratio
Following suggestions from the literature, the first trials in the Tundish ladles with magnesium alloy used 0.02% Bi, which resulted in a TRE/Bi ratio of 0.74 (the assumption was lower bismuth recovery, this is why the addition rate was doubled). This resulted in poor nodularity (vermicluar) and very low magnesium recovery. The next test only used 0.01% bismuth, the next month we ran with 0.0055% addition, and from there the addition rate was gradually decreased to 0.002%. All samples in that range showed good results within the mentioned parameters (nodule count, chill, nodularity). See table 2 for an overview of the results.

The commercially available inoculant sampled had a range of TRE/Bi ratio from 0.3 to 0.9. It resulted in poor nodularity (vermicular), higher nodule count than pure bismuth, and higher chill tendency than with pure bismuth plus some carbides. This trial is not included in the following table due to the poor results, for which we hold the high bismuth content responsible.

Mechanical Properties
Increased attention was paid to mechanical properties of the bismuth treated iron versus iron without bismuth. Extensive testing on separately cast tensile bars as well as casting sections found no significant difference within the range of the before mentioned addition rates (except in the cases of poor nodularity, e.g. vermicular graphite present).

Bismuth Carryover
We calculated bismuth recovery and theoretical buildup of bismuth in the returns. It wouldn’t pose a problem in the cupola with 50% returns in the charge. Bismuth was added to 25% of the production, that is both small Disas only, for eight months. During that time nodule count increased shop-wide suggesting that the carried over bismuth is still potent enough to help inoculation.

After the bismuth addition was stopped, it took about six months to become noticeable in the routinely measured chill wedges. This means that the nodule count decreased somewhat after the bismuth was eliminated and also the depth of the last carbides increased somewhat, but not to the same level as before. The higher level of iron quality can be accounted for by other improvements implemented during that time. 

Back to top