All cast iron materials are affected by this defect, whereas nodular graphite cast iron is most likely to be affected. This defect only occurs on thickwalled parts, i.e. on medium-weighted to heavy castings; however it also occurs sometimes on light-weighted but very compact castings with a very large casting modulus. The most affected parts are always molding parts subjected to high thermal loads such as sand edges, hot spots etc., i.e. parts where the material remains liquid for quite a time thus causing long reaction times and the intensive heating of the mold material.
This defect occurred both on furan resin mold materials (PTS as well as phosphoric acid hardened) and on synthetic bentonite-bound mold materials as well as on dry sand castings. The defect was most likely to occur on regenerated furan resin sands with high annealing loss with phosphoric acid hardening showing the highest rate.
According to practical experience reports, this defect occurs to an increased extent on bentonite-bound mold materials in case of the excessive enrichment of exothermal riser residues in the mold material, in particular the increased fluorine content caused by these risers. Up to now, it is not clear whether the fluorine content is the actual cause or only the indicator for this defect.
The defects are pockmark-like wrinkles and recesses over the whole surface of the casting and/or in the casting skin. They are covered with a white/bluish film made of fiber-like SiO2 which can be removed by blasting the castings leaving the pockmarked casting surface only (Fig. 1 to 4).
It is quite remarkable that, depending on the material, flake graphite, vermicular graphite or nodular graphite clearly maintained in its original form and order is deposited in the SiO2 cover (Fig.5). This means, that the defect is not merely caused by the melt/mold material reaction but also to a certain extent in the solidus range of solidification.
Fibrous SiO2 is formed between 1200 and 1400 °C of SiO (which is a product of SiO2 reduction) either by careful oxidation or by disproportionation according to the equations 1 to 3:
It is unstable and transforms to cristobalite when tempered or to amorphous SiO2 when exposed to air, however always maintaining its fibrous structure. Besides quartz sand, surface oxides and slag rich in SiO2 can be used as reducing agents for the cast iron / mold material system, but also magnesium in the iron and hydrocarbons in the mold material (defects very likely to occur on nodular graphite cast iron in resin-bound mold materials!). However, it is still unknown which “contaminants” in the mold material act as reducing agents or only as stimulating flux, e.g. for decreasing the sintering point.
Another interpretation would be that at the beginning of the solidification process the surface of the iron is oxidized by the oxygen available. As is generally known, oxide phases rich in SiO2 (e.g. fayalite) occur during the formation of liquid oxidation products (slag) and solid scale layers. After the available oxygen has been used up during the formation of silicates, these silicates may be reduced to SiO under reducing conditions which prevail in the mold due to the pyrolysis of the mold material additives. During reoxidation, the fibrous SiO2 is formed of SiO.
G. Levelink has another explanation: The white cover is formed by the oxidation of easily oxidizable elements (aluminum, magnesium, silicon); an explanation for their formation is that the gas milieu in the boundary between casting and mold (furan resin mold) is in general easily oxidized. Although the gases released from the furan resin only contain very little oxygen, it can be oxidized by elements with high oxygen affinity; it is, however, not available for iron.
Due to the low availability of oxygen, the oxidation process is very slow. Thus the involved elements have the chance to diffuse from the casting to the boundary. In case of thick-walled castings, the resulting cover may reach a layer thickness of several millimeters.
The low density of the oxidation products plays a role for this phenomenon. Thus the layer remains permeable for the oxygen from the mold materialand can easily continue to grow as long as oxygen is available and oxidizable elements move from the liquid iron to the boundary. This low density also implies the large volume of the formed oxide layer which can be deposited in thick layers. The growth of the oxidation products originates in “nuclei”, with pocks being formed at first which in turn cause recesses in the casting. After blasting the casting, the white cover is removed from the pocks, and only recesses are visible on the casting.
According to M. Schrod and H. J. Wojtas, the holes on the surface of the casting are obviously caused by the reaction layer. Highly viscous phases are formed in this reaction layer which are blown up from the released crystal water of the involved mixtures or CO-CO2 mixtures resulting both from oxidation and reduction processes of the involved mineral mixtures. Due to the obviously high viscosity of these phases, the inside pressure of the “foams” and/or the reaction layers formed around the quartz grains may massively increase and press the components of the reaction layer or the foam/reaction layer into the solidifying casting body. After demolding and/or blasting, these reaction layers are removed, and the recesses/holes remain on the surface of the blasted specimens which contain oxides and mineral mixtures.
However, even this theory cannot explain the graphite inclusions in the SiO2 fiber layers observed here and there. A possible explanation is that iron is displaced from the mold surface by SiO2 growing into the surface, because fibrous SiO2 requires a higher volume than quartz.
Measures for prevention:
Although the causes are not yet completely clear, the following measures taken in practice usually help avoiding this casting defect:
1. Reducing the casting temperature as far as this is compatible with the casting program and the occurrence of slag defects (see Slag inclusion).
2. Increased addition of new sand (see Addition of new sand) in particular if regenerated furan resin sands are used and/or if the quality of the reclaimed material is improved.
3. Checking the reclaimed material for annealing loss (< 1.0 %), electrical conductivity (< 200 mS/cm), nitrogen content (< 0.02 %) and for phosphate acid hardening also for the phosphate content (< 0.5 %).
4. Increasing the permeability to gas.
5. Minimizing the influx of resin-bound core residues which have not been cast.
6. Replacing exothermal riser tools by insulating risers.
7. Adding less hardener to furan resin-bound molds and using PTS acid hardener instead of hardeners containing phosphoric acid.
8. Using face coatings, if possible “plastic” face coatings, for the formation of a separating layer between iron and mold material.
9. Minimizing elements with oxygen affinity, checking the melt additives.
10. Reducing the silicon content, if any, to below 3.0 %.
11. Allowing for the optimal cooling of the endangered casting areas, especially for thick-walled parts.
Mold sand-based casting defect, Casting defect
Chill wedge test piece