Using thermal analysis practically in iron casting.

The various phenomena that occur along the cooling curve play major roles in the quality of the final casting.

With the arrival of new 16-bit analog-to-digital converters, computers and enhancement software, it is now possible to see and measure events within the solidification of iron samples. This data can be used to examine the iron's chemistry, soundness, chill and microstructure.

The tool for obtaining this information is the thermal analysis cooling curve--a graphic manifestation of the cooling of the sample as measured in degrees per second. This article discusses the different segments of the cooling curve and how they relate to the final casting. Chronologically, the major parts of the curve are preliquidus, austenitic arrest, dendritic growth, eutectic solidification, end of freezing point and the austenitic transformation area.

For this discussion, a tellurium cup sample has been poured of a hypo-eutectic iron commonly used in malleable, gray and ductile iron castings. In the graphic display of the curve, the vertical axis represents temperature in Fahrenheit, while the horizontal axis is time in seconds.

The smaller graph at the bottom of the screen displays the first and second derivatives. The first derivative is a picture of the actual solidification, and the area it contains is the energy being released by the crystallization process. The second derivitive is used to pick out and magnify certain points on the overall curve that mark the beginning or end of a cooling phase.

Preliquidus

Preliquidus is the section of the cooling curve starting from the highest temperature (just after pouring) down to just above austenitic arrest, when dendrites begin to form. During the preliquidus stage, slow-downs in the cooling rate are called liquidus arrests. On the screen, a liquidus arrest is shown when the second derivative (the bottom-most line) passes through zero in the positive direction. Different from arrests, abnormality reactions are much weaker, with shorter and less energetic durations.

Any temporary arrests in cooling that occur in this stage are oxygen- or sulfur-related. Cupola shops will see more arrests when the coke bed gets low and the slag indicates oxidizing conditions.

In electric melting, these arrests have been seen when melting rusty scrap in coreless induction furnaces. Common temperatures for these reactions are 2155-2162F (1195-1200C) and 2138-2156F (1170-1180C).

Austenitic Arrest

During the cooling curve period of austenitic arrest, austenitic dendrites begin to form. The strength of this reaction depends on the number of nuclei available to start crystals. Good nucleation is important because the growth of these cells contributes to the final physical properties of the iron.

The liquidus temperature determines the carbon equivalent. That temperature is that of the lowest rate of cooling during the austenitic arrest. Mathematically, this is when the cooling rate is within the expected range and the second derivative passes through zero.

This has been found to be correct for base and treated gray iron using either tellurium or nontellurium cups. The duration of the liquidus can also be used in predicting nodularity in ductile iron. Some research indicates that with newer thermal analysis equipment, an accuracy of within 5% nodularity may be achieved.

Dendritic Growth

When seen on the rate of cooling graph, the area above the rate of cooling and the base line is proportional to the amount of solidification. The area before the maximum rate of cooling is austentite and the area after is eutectic.

During this time period, the austenitic dendrites are growing into the sample. This growth generates heat, so the higher the cooling rate, the lower the growth rate. Because austentite dendrite is an important factor in tensile strength, a lower rate of cooling will promote a stronger iron (though it may also promote D flake).

Eutectic Arrest

Solidus is the common name for this arrest, though, as will be shown later, the end of freezing is where the casting becomes actually solid. In a tellurium cup during this phase, iron carbide and possibly a small amount of austentite are formed. Without the presence of tellurium, this phase forms graphite, ferrite and iron carbide.

Sometimes there will be strong spikes or dips in the cooling curve during the eutectic arrest period that cannot yet be fully explained. In the case of ductile iron, Carl Loper, University of Wisconsin-Madison, suggests "the shape of the cooling curve following the solidus can be identified with the degeneration of the normal spheroidal growth process into one causing nonspheroidal forms."

A system of visual checks developed in the 1970s relates the release of heat during the first part of solidification to the percentage nodularity. An early recalescence (glowing on the metal surface) indicates nodularity, while a late recalescence at a lower temperature indicates flake graphite. This effect can also appear in gray iron.

Occasionally, there will be an unexpected recalescence in a tellurium cup. This can happen if there is undissolved graphite in the molten iron from a late carbon addition, from a graphite washed sampling spoon or from a cold melter that has not yet reached high enough temperatures to dissolve the carbon addition.

From the liquidus and solidus temperatures of a tellurium-treated sample, the carbon and silicon can be estimated. The equations for these values are affected by the amount of phosphorous, manganese, tin and chrome present. Individual foundries should adjust the slope and offset the silicon equation for each major type of iron.

Recalescence in an untreated sample tells a great deal about the inoculation of the iron. The greater the recalescence, the more inoculation the iron requires. Well-inoculated iron will typically have a recalescence from 6-10F- (3-5C). Uninoculated, electric-melted iron will range from 14-18F (7-9C). Regardless of the source of iron, its present condition and inoculation requirements can be determined by the recalescence.

Chill

Chill is another important development at this stage of the curve. Chill occurs when the iron becomes undercooled enough to form carbides instead of graphite and pearlite. The temperature at which carbides form moves with silicon and is the same temperature at which the tellurium solidus occurs. In degrees, undercooling is generally a negative number, and is the distance between the minimum recalescence of the untreated curve and the carbidic (tellurium induced) solidus arrest.

Since the depth of chill depends on chemistry, casting thickness and inoculation, an effective way to control chill is by manipulating the degree of undercooling. This is done by pouring both a treated and nontreated cup. From the treated cup, which reports on silicon, the carbide eutectic can be calculated. From the untreated cup, the undercooling temperature is determined. From both results, the degree of undercooling can be calculated.

Carbides can form in a casting in two ways. First, if the initial undercooling of the solidus dips below the carbide eutectic temperature, carbides can freeze out the liquid. Chill wedges force this by extreme cooling of one edge of the sample. Thin castings with sharp corners will generally have at least a small amount of chill in these corners, despite good practice.

The second type of chill happens when the temperature at the end of the solidus reaction starts to fall off and the end of freezing point has not been reached. This accounts for the small intercellular carbides sometimes found in micros.

While under normal plant production conditions it may be impossible to keep the tail of the solidus curve above the carbide eutectic temperature, good foundry practice will keep the carbides microscopically small and few enough to prevent any adverse effect in machining the casting. A curve that is well above the carbide eutectic and then rapidly falls off in temperature at the end of the solidus is less likely to produce significant carbides because most of the metal has completed solidification.

This is a crucial area of the cooling curve because in the final product any shrinkage not related to foundry practices or materials is generally caused by the dendritic austentite blocking feeding paths, as well as the time required for eutectic solidification.

End of Freezing Point

A common mistake is thinking that solidus is the actual point of final solidification. The final freeze of the grain boundaries, however, can occur 40-120F (20-60C) later in the curve. The difference is important to the casting's soundness and strength.

After most of the heat of solidification has been used up, the low melting phases of the iron--phosphorous, sulfur, tin, antimony and lead--are still liquid. These solutes will have been forced to the grain boundaries, where they finally freeze in concentrations more than 100X their normal levels. If the phase concentrations are high, the end of freezing temperature will be lower.

Therefore, the difference between the solidus temperature and the end of freezing temperature indicates the concentration of these low melting point phases.

This is important for two reasons. First, feeding cannot occur at this point. The dendritic austentite and eutectic crystals occupy more than 99.9% of the casting. If the delta freezing temperature is too large, then microshrinkage voids can occur in the grain boundaries. In addition, these low melting phases are not good for strong physical properties. A large delta freezing may lead to lower tensile strength, yield strength and percentage elongation.

Austenitic Transformation

This is the point between 1450-1400F (788-760C) when the austentite is transformed into the form of iron carbide. It may become pearlite, martinite, bainite or mixtures of all three. The transformation is as yet hard to detect. In a normal sample, this process may take as long as 10-15 minutes to occur.

Problems arise when the cup disintegrates and falls away, exposing the sides of the sample to air. This causes a noticeable change in the rate of cooling and can confuse the analysis if it happens at an inopportune time. Current cup designs will have to be rethought for the austenitic transformation number to be a useful tool.

The information presented here is related to relative changes in the thermal analysis and is not an absolute predictor. No foundry should expect numbers from another to apply, and each must do its own homework to determine equations and operating ranges.

As foundries strive to produce a more consistent product, thermal analysis provides the tools necessary to manipulate casting properties. While it can help control the carbon, silicon and CE contents, it also can aid in determining the levels and consistency of oxidation, inoculation, chill and microshrinkage in iron.