Mold binder decomposition: prime source of cast iron gas defects.

Gas binder byproducts are a problem, but gas release rate, composition, volume, mold permeability are among defect-causing factors.

It is generally accepted that the gases present at the mold-metal interface are important determinants in the severity of gas defects found in a metal casting. For example, resins containing nitrogen have long been associated with the production of nitrogen pinholes in iron. Unfortunately, due to the difficulties in designing experimental methods to measure consistently the effects of these gases, it has been impossible to fully understand the relationships between the binder, casting conditions and resulting pinhole severity.

Almost all current binder systems produce significant amounts of gases during casting--it would be difficult to design a binder that did not do so. At casting temperatures, organic compounds are unstable and rapidly decompose into a mixture of CO, C|O.sub.2~ and (usually volatile) hydrocarbons. Since oxygen and/or nitrogen are essential components of polymeric resins, a mixture of |N.sub.2~, |H.sub.2~, N|O.sub.x~, N|H.sub.3~ and |H.sub.2~O gases frequently associated with pinholing also will be produced during casting.

Additional sources of nitrogen include sand additives and polymerization catalysts (urea, hexamethylene tetramine, triethylamine, ammonium chloride, flour and seacoal). Even inorganic binders such as silicates or bentonites depend on water or organic molecules such as esters to form necessary chemical bonds. Despite the gas evolution problems associated with current resins, however, a nondecomposable resin would be even less desirable because sand removal would become nearly impossible.

Despite the inevitability of binder-produced gases, the fact that pinholes only occur in certain cases indicates other factors must be important. Factors that have been investigated include:

* the volume and rate at which gases are released;

* composition of the gases;

* effect of gases already dissolved in the melt;

* casting temperature;

* the surface tension of the metal.

Apart from obvious relationships, the results are not always straightforward. Complicated combinations of factors are responsible for the presence or absence of defects.

One example would be the amount of gas released by the resins. As expected, large gas volumes and high evolution rates tend to promote defects. However, a small change in the permeability of the mold can change all the results, or the combined effects of two or more gases could also add to the problem.

For example, pinholes may form when two or more gases with individual vapor pressures less than 1 atmosphere combine to produce a total vapor pressure great enough to produce a bubble. Thus, studies attempting to determine the effect of nitrogen on pinholing might not be valid unless the hydrogen content also is known. The role of hydrogen is especially important because it has 14 times the volume of an equivalent mass of nitrogen.

The gases most frequently associated with casting defects are CO, C|O.sub.2~, |H.sub.2~ and |N.sub.2~, CO and C|O.sub.2~ are produced when carbon in the resin (represented by C|H.sub.x~) or dissolved in the iron reacts with atmospheric oxygen, the resin or |Fe.sub.2~|O.sub.3~.

|O.sub.2~ + C|H.sub.x~ |right arrow~ C|O.sub.2~ + CO + |H.sub.2~O (1)

|Fe.sub.2~|O.sub.3~ + C|H.sub.x~ |right arrow~ CO + |H.sub.2~O + Fe (2)

The relative amounts of CO and C|O.sub.2~ and residual volatile organic compounds depend on several factors. The two most important are the degree of unsaturation (double and triple carbon-carbon bonds) of the resin and the amount of oxygen present. High oxygen content and low unsaturation tend to produce more C|O.sub.2~. Thus silicate binders give smaller CO/C|O.sub.2~ ratios than organic resins.

The major sources of hydrogen are water and organic materials.

Fe + |H.sub.2~O |right arrow~ FeO + |H.sub.2~ (3a)

2Fe + 3|H.sub.2~O |right arrow~ |Fe.sub.2~|O.sub.3~ + 3|H.sub.2~ (3b)

Fe + C|H.sub.x~ |right arrow~ dissolved C + |H.sub.2~ (4)

Water reacts readily with molten iron to produce hydrogen gas and is probably the largest contributor of hydrogen. Liquid iron also is capable of reducing organic materials directly to |H.sub.2~ (as shown in equation 4), but it may be that equation 1, the oxidation of the organic compound by oxygen, occurs first with the resulting water converted into |H.sub.2~ by reactions shown in 3a or 3b. The difference between these two routes might not seem significant, but because pinhole severity is apparently affected by the mold atmosphere (reducing instead of oxidizing), it may be an important point.

During the decomposition of N|H.sub.3~ or nitrogen-containing compounds at high temperatures, |N.sub.2~ and N|O.sub.x~ are produced.

2N|H.sub.3~ |right arrow~ |N.sub.2~ + 3|H.sub.2~ or 2N + 6H (5)

N|H.sub.3~ + |O.sub.2~ |right arrow~ N|O.sub.x~ + |H.sub.2~O (6)

Few reports have mentioned the nitrogen oxides (N|O.sub.x~), but they might be significant, depending upon the amount of oxygen present. Major sources of nitrogen are organic resins and their catalysts. Mold additives, such as seacoal and wood flour, typically contain nitrogen-bearing impurities.

Resins: Chemistry/Decomposition

There are several classes of resin systems, each with its own chemistry and decomposition characteristics. Heat-cured resins, such as shell and furan hotbox binders, use heat (and occasionally catalysts) to drive the polymerization reactions that produce the resins. Because the shell binder requires an organic amine catalyst, nitrogen will always be present in the mold.

The furan resin does not require a nitrogen catalyst, but in many cases still contains significant amounts of urea (a nitrogen compound) as a part of its furfuryl alcohol solvent system. In addition, the furan resin depends upon condensation (water-producing) reactions to produce the furan polymer.

Unless the mold is baked well, water will be present during casting when the carbon, hydrogen and oxygen atoms, the bulk of these organic resins, will produce a variety of volatile organic decomposition products such as CO, C|O.sub.2~ and |H.sub.2~O.

Nobake binder systems are based on compounds that react at room temperature (aided usually by a catalyst) to produce polymers. Unfortunately, nitrogen is usually an important atom when forming chemical bonds at lower temperatures.

For instance, isocyanate binders work by reacting a nitrogen-containing polyisocyanate with the hydroxyl group of either a large aliphatic molecule (alkyd resin) or a phenolic polymer (phenolic urethane resin). These reactions may be catalyzed with pyridine or triethylamine, which add even more nitrogen to the mold.

Another nobake binder, the furan/acid system, is a combination of nitrogen-containing urea formaldehyde polymers dissolved in furfuryl alcohol. They are catalyzed into further polymerizations by strong acids, usually TSA (toluene sulfonic acid), BSA (benzene sulfonic acid) or |H.sub.3~P|O.sub.4~ (phosphoric acid). These acids contain no nitrogen, but all contain a large proportion of oxygen that can lead to higher C|O.sub.2~ levels. The phosphorus and sulfur are not known to affect gas defect formation.

In a continuing effort to increase the quality of castings, some nitrogen-free nobake systems have been developed. These include the ester-cured phenolic, phenol formaldehyde and silicate systems. Silicate resins rely on condensation reactions between silica and either C|O.sub.2~ or organic esters. Condensation leaves water in the mold that, in some cases, is necessary for the chemical bonding upon which the resin depends.

Phenol formaldehyde resins are polymers produced from phenols and formaldehyde and catalyzed by TSA, a reaction that also produces water. The ester-cured phenolic is unusual in that it does not produce water as a byproduct. However, the alcohol and acid salt that it does form can react at high temperatures to produce water at the time of casting.

Due to the unpredictability of defect formation, it is difficult to compare the different resins. Comparisons, however, have been attempted. For example, some studies have found the following:

* furan hot-box binders were less prone to defects than shell binders;

* urethane binders produced less defects than furan acid resins, with the phenolic urethanes showing the lowest amount of gases and the smallest evolution rates;

* the silicate binders produced the most gas (probably water), the highest gas evolution rate, and in some cases, the most defects.

It is even more difficult to explain the reasons behind these trends. Since a complex series of gases is generated by the pyrolysis of resins, it is not easy to determine the identity of the gas or gases contributing to a particular defect.

One common assumption concerning the nature of these gases is that |H.sub.2~ and |N.sub.2~ defects produce graphite-lined cavities surrounded by a carbon-depleted zone. CO bubbles have no graphite and usually show an oxide lining. Because the annealing of cast iron in an inert atmosphere or vacuum has similar graphite linings, this analysis is not very consistent or dependable. In addition, it is even more useless if more than one gas is responsible for the defect.

Summarizing the Conflicts

Gases detected in a mold during casting consist of some or all of the following: |N.sub.2~, |H.sub.2~, CO, C|O.sub.2~, |H.sub.2~O, N|H.sub.3~, HCN, formaldehyde and a variety of small organic compounds. The composition of the evolved gases falls roughly within the following ranges: 40-60% |H.sub.2~, 1-3% |O.sub.2~, 10-20% |N.sub.2~, 15-30% CO, 3-10% C|O.sub.2~ and 1-20% total hydrocarbons.

Due to the obvious experimental difficulties involved, the uncertainty of these values makes comparisons of resins difficult, especially between different studies. General trends, however, are observable. Although the oxygen and nitrogen values were mostly due to air in the mold prior to casting, nitrogen levels were significantly higher in cases where they were present in the resin.

Hydrogen levels were fairly similar for the organic resins, but were lower (about 30%) in the inorganic silicate esters. The CO/C|O.sub.2~ ratios, considered to be an indicator of the "reducing" nature of the mold atmosphere, varied significantly between studies.

Generally, the resins showed a decreasing CO/C|O.sub.2~ ratio in the following order: shell furan/hotbox |is greater than~ furan/acid |is greater than~ phenolic urethane |is greater than~ silicates. Total hydrocarbon production decreased in the following order: shell |is greater than~ phenolic urethane, silicate ester |is greater than~ furan/acid and furan/hotbox |is greater than~ silicate/C|O.sub.2~. It is interesting that very little water (less than 3%) is observed, even when the mold is wet. Molten iron is a good medium for dissociating water into hydrogen and oxygen.

In addition to the gas composition, defect formation is directly affected by such gas variables as the total volume released, formation rates and release time.

The total volume of evolved gases decreases in the following order: silicate esters |is greater than~ shell furan/hotbox and furan/acid |is greater than~ phenolic isocyanate. Increasing the amount of resin in the mold will increase the volume of gas evolved. Defect rates decrease on average as the gas volume decreases.

There is also a strong dependence on pouring temperature, with higher molten metal temperatures yielding greater volumes of gas. This dependence is affected by such factors as the thermal conductivity of the sand and the physical configuration of the casting. Generally, the resins producing the largest gas volumes also had the highest gas evolution rates that typically peaked 15-30 seconds after pouring. It is not known how important gas evolution rates are in defect formation, although one would expect that the rapid production of gases would promote pinholes.

The third and possibly the most critical factor may be the moment when each gas is evolved. The gas composition ranges given above are averages over the first 15 minutes after casting. Individual gas concentrations can vary widely over this time, producing a range of compositions.

For example, the silicate-ester binder, usually described as a producer of oxidizing gases and which begins to decompose with a low CO/C|O.sub.2~ ratio, gradually increases this ratio over a 10-minute period, eventually reaching a value similar to organic resins. If defects are dependent upon the composition of the mold atmosphere, then the time at which these defects are initiated could be very important.

Additional Factors

It is clear that the formation of gases during casting is a necessary, but not sufficient, condition for the production of defects. There are several known (although not necessarily understood) factors that appear to play a part in defect severity.

Nitrogen: Although nitrogen-containing resins are a recognized problem, there are instances that intensify these problems. The use of incorrect stoichiometry when mixing binder components is one such factor. An incorrect ratio of the components (including any nitrogen-containing catalysts) can leave reactive amines in the mold. This produces |N.sub.2~ and N|H.sub.3~ more readily than the resin and promotes nitrogen pinholing and can be caused by infrequent calibration or a mechanical malfunction of binder pumps.

Incomplete reaction between the binder components, due to poor mixing or improper curing temperatures and times, will also produce the same effect. Even when the components are perfectly balanced and mixed, excess resin in the mold produces extra nitrogen. Additional nitrogen also may be generated by the use of reclaimed sands containing traces of leftover resin.

Water: Due to the many possible sources of water, hydrogen pinholing is often a major problem. Resins, dependent on condensation reactions for their formation, leave water inside the mold. Resins may also be hygroscopic, a serious problem during humid weather. Some resins, such as phenolic urethanes, have been shown not to be hygroscopic, but water vapor in the air can still react slowly with them. This not only weakens the resins by attacking the polymeric linkages, but it also produces amines that promote nitrogen defects during casting.

Even if the cured resin is not hygroscopic, the original binder components usually are. Thus, the components can be attacked by water vapor (either in their storage containers prior to making the mold or afterwards) if some of the components are left unreacted due to incorrect component ratios, poor mixing, etc. Baking the mold prior to casting will help alleviate some of these problems.

Mold Atmospheres: The more reducing mold atmospheres (those rich in |H.sub.2~ and having a higher CO/C|O.sub.2~ ratio) tend to promote pinholing. No satisfactory explanation for this phenomenon exists, due in part to a less than obvious correlation between defects and mold additives that affect the "reducing" nature of the mold.

For example, the addition of |Fe.sub.2~|O.sub.3~ has been shown to produce a more oxidizing atmosphere and decrease the incidence of defects. Pitch and seacoal additions, however, produce reducing atmospheres and help alleviate pinholes, even though seacoal actually increases hydrogen pickup by the iron. Cereal additions increase the amount of |H.sub.2~ present, leave the CO/C|O.sub.2~ ratios unaffected and increase defects.

Part of the problem is due to an inexact definition of a "reducing" atmosphere. A truly reducing atmosphere should not only have a high CO/C|O.sub.2~ ratio, but also high |H.sub.2~ and hydrocarbon values. Yet, a comparison of the pyrolysis products of different resins shows that all three values can vary independently. Which, if any, of these gases actually affects pinholing tendencies remains to be determined.

Indeed, since time profiles of resin pyrolysis experiments show that the relative amounts of these gases are constantly changing with time, a so-called "reducing" resin may actually be producing an "oxidizing" atmosphere during the critical time in which the pinholes are forming. This important area should be a subject for future study.

Temperature: The effects of temperature are varied, although higher temperatures typically lead to more defects. The composition of the resin pyrolysis products is temperature dependent, with higher temperatures generally producing more reducing atmospheres as well as increasing the amounts of gases produced. Higher pouring temperatures also lead to longer solidification times that promote pinholing in thin castings and decrease them in thick castings. These effects depend on factors such as the thermal conductivity of the sand and the particular chemical reactions that each resin must undergo.

Dissolved gases: Gases already present in the liquid iron increase the likelihood of defects. Molten iron already saturated with nitrogen, for example, will be less likely to tolerate the influx of additional nitrogen from the resin. In fact, depending upon the conditions, this same iron might not be able to tolerate the influx of any gas. Defects dependent on a combination of gases to overcome low individual vapor pressures might occur. Thus, hydrogen from the resin, in combination with hydrogen and nitrogen already present in the liquid iron, might be required to form hydrogen/nitrogen defects.

Aluminum: Aluminum can be a major contributor to hydrogen pinholing. It has the ability to dissociate water into hydrogen and oxygen, although liquid iron is able to perform this conversion itself. Evidence has shown, however, that aluminum increases the amount of hydrogen pickup by the iron, which can dramatically exacerbate the effect of water in the mold. Its ability to lower the surface tension of the liquid iron has also been demonstrated. A lower surface tension would allow gases from the mold to penetrate more easily into the metal, increasing pinholing.

In summary, the study of gas defects is hindered not only by its many complicated variables and lack of analytical techniques, but also by the large number of possible defect mechanisms. Defects appearing similar in different castings may have formed for entirely different reasons. They may not even contain the same gases, but to date, the ability to analyze the gases trapped within defects is unknown.

Therefore, one important goal in arriving at the genesis and the certain identification of the gases inside defects remains a prime goal of metallurgical study and research.

Bibliography

W.D. Scott, P.A. Goodman, R.W. Monroe, "Gas Generation at the Mold-Metal Interface," AFS Transactions, p 145-156 (1978).

C.E. Bates, W.D. Scott, "Decomposition of Resin Binders and the Relationship Between the Gases Formed and the Casting Surface Quality--Part 3," AFS Transactions, p 209-226 (1977).

R.L. Naro, R.L. Pelfrey, "Gas Evolution of Synthetic Core Binders: Relationship to Casting Blowhole Defects," AFS Transactions, p 365-376 (1983).

W.D. Scott, C.E. Bates, "Decomposition of Resin Binders and the Relationship Between Gases Formed and the Casting Surface Quality," AFS Transactions, p 519-524 (1975).

C.E. Bates, W.D. Scott, "The Decomposition of Resin Binders and the Relationship Between Gases Formed and the Casting Surface Quality--Part 2," AFS Transactions, p 793-804 (1976).

R.L. Naro, "Variables Affecting the Formation of Porosity Defects in Iron Castings Prepared with Urethane Binder Systems," AFS Transactions, p 257-266 (1974).

H.G. Levelink, F. Julien, H. DeMan, "Gas Evolution in Molds and Cores as the Cause of Casting Defects," AFS International Cast Metals Journal, p 56 (1981).

J.F. Wallace, P.F. Wieser, "Pinholes in Gray Iron," AFS Research Report 6514, p 48-61 (1965).

J.V. Dawson, J.A. Kilshaw, A.D. Morgan, "The Nature and Origin of Gas Holes in Iron Castings," AFS Transactions, p 17-33 (1965).

C.E. Bates, R.W. Monroe, "Mold Binder Decomposition and Its Relation to Gas Defects in Castings," AFS Transactions, p 85-100 (1981).

R.W. Monroe, "Use of Iron Oxide in Mold and Core Mixes for Ferrous Castings," AFS Transactions, p 355-364 (1985).

R.S. Lee, "Nitrogen Contamination in Green Sand," AFS Transactions, p 875-882 (1987).

P. Frazier, R. Boudreau, "Ester-Cured Phenolic Cold Box Process," AFS Transactions, p 381-384 (1986).

S.F. Carter, W.J. Evans, J.C. Harkness, J.F. Wallace, "Factors Influencing the Formation of Pinholes in Gray and Ductile Iron," AFS Transactions, p 245-268 (1979).

M. Svilar, J.F. Wallace, "Removal of Aluminum from Gray Cast Iron to Reduce Pinholes," AFS Transactions, p 421-430 (1978).

R.V. Naik, J.F. Wallace, "Surface Tension--Nucleation Relations in Cast Iron Pinhole Formation," AFS Transactions, p 203-224 (1980).

G.S. Lukacek, R.E. Fontaine, N.J. Ruffer, "Humidity--Its Effect on Nobake Binders," AFS Transactions, p 455-462 (1983).

J.V. Dawson, L.W.L. Smith, "Pinholing in Cast Iron and Its Relationship to the Hydrogen Pick-up from the Sand Mold," BCIRA Journal of Research and Development, vol 6, p 2-9 (1956).

"Thermal Decomposition of Furfuryl Alcohol Resins," NASA Technical Report N63-22125 (1963).

B. Hernandez, J.F. Wallace, "Mechanisms of Pinhole Formation in Gray Iron," AFS Transactions, p 34-47 (1979).