Abstract
There are many different treatment methods to produce ductile iron. All of them involve the introduction of magnesium in some form; either alone or in combination with cerium or other rare earth metals. Successful production of high quality ductile iron depends on the successful control of magnesium. Operators are constantly on the vigil in controlling and maintaining sufficient amount of magnesium in the iron to promote satisfactory nodularization.

In a batch process of production such as in ladle, sandwich or tundish operation or in a semi-continuous operation of treatment-pressure pour combination, it is difficult to maintain the level of magnesium. It is also equally difficult to know the exact percent of magnesium in the melt before pouring. A major concern expressed frequently by the ductile iron producers is, is there a way to rapidly and accurately test and analyze the final metal not only to ensure the success of the magnesium treatment but also to know the exact percentage of magnesium to make up adjustments to charge additions? One possible way is to apply the principle of thermal analysis to analyze final iron. BCIRA applied this principle as far as to accept or reject a batch of metal, with 0.04%Mg as the reference. Present authors have extended the applicability of thermal analysis to predict exact percentage of magnesium in the final iron. The analysis, made in a tellurium/sulfur cup is displayed in less than three minutes of pouring.

Introduction
Ductile iron also known by the names of spheroidal graphite and nodular iron is made by treating the liquid cast iron with special "spheroidising" elements to promote the precipitation of graphite in the form of spheroids rather than as interconnected flakes. Alkaline earth metals that include magnesium and calcium and the rare earth metal such as cerium, lanthanum, and yttrium all have been proved to be the special elements.

In most American foundries, treatment with magnesium (Mg) is used in the production of ductile iron. Magnesium is a very highly reactive metal with a low vapor pressure and this combination results in a considerable loss of the element during treatment. In practice, excess amount of magnesium is added to the base metal to retain a certain level of residual magnesium in the treated metal. Residual magnesium level which produces completely spheroidal graphite is observed to be between 0.02-0.06%. A very low magnesium can result in insufficient spheroidisation and a very high magnesium can result in porosity and carbides. Under both conditions, the produced castings will be of inferior quality. The critical "low-high" window for magnesium, though universally accepted as 0.02-0.06%, has been trimmed and narrowed by individual foundries to suit their production and quality standards.

It is not an exaggeration when we say that the addition and control of magnesium is the single most important step in the production of ductile iron, and the production of quality castings is determined by how well a foundry can precisely control the level of magnesium in their melt. This becomes doubly crucial when the foundry runs a Converter treated iron with a Presspour system and a DISA set up.

Photo 1

Fisher converter Presspour combination In this setup, the base metal is treated in a Fisher converter (Photo 1) and then transferred to the presspour pouring furnace before pouring the inoculated metal into the molds (Photo 2) automatically. This operation is ideally suited for a high production foundry that can handle a large tonnage of hot metal every hour.

Photo 2

Fisher Converter: Here the treatment involves the use of pure magnesium in a closed chamber. In the chamber very violent reaction takes place. A high percentage of magnesium recovery is achieved in this process. Reaction only takes a few minutes but the preparation of the chamber takes about 10-15 minutes with the possibility of only 3-5 taps per hour. Converter treatment produces a cleaner iron with less slag, permits the use of high silicon returns in the charge material, and tolerates a high sulfur base iron. It is not easy to predict the recovery of magnesium in the converter process. Many factors influence this; the extent of reaction; the consistency of operation, reaction of magnesium whether complete or insufficient, plugging of the holes in the chamber plate, over tapping of metal, build up of slag inside the reactor, improper amount of added magnesium or any other inadvertent operator negligence.

Even though the time of reaction is monitored as a precautionary check, it is not always possible to guess the success of the operation or to have a knowledge of the magnesium in the liquid metal. It is important to know this for two reasons:

Presspour Pouring Furnace:
The function of the pouring furnace is to act as a reservoir of ready source of treated iron. In addition to this, the furnace helps to maintain the temperature of iron. Both magnesium and temperature can be closely controlled here. The treated iron from the converter is immediately transferred to the furnace fill spout in batches of 2-4 tons, depending on the capacity of the converter. Therefore, there is a close control in the cycle of metal flow right from the time the base metal is dispatched to the converter for the treatment to its storage in the holding furnace and drain through the pouring spout to the casting molds. Photo 2 shows the entrance of the metal in the fill spout into the furnace and the flow through the pour spout.

Two major problems are encountered here:

As a result of these two factors, the magnesium percentage in the furnace continuously fluctuates, more so when the melt is held for a considerable length of time. Therefore, it is very difficult to even guess the percentage of magnesium in the holding furnace. For a good control of magnesium, a knowledge of %Mg in the furnace is important for the following two reasons:

Pressure Pouring Spout:
The metal in the furnace is pressurized to fill the pouring basin which has the stopper rod auto pouring mechanism. The metal in the basin is inoculated by wire treatment process and is further aided by instream inoculation during filling of the molds. Precise quantity of treated and inoculated metal is delivered by the automatic pouring system into the continuous array of molds generated. Level of magnesium at the pouring basin near the pouring spout depends on

The Problem:
So the metal in the pouring basin ready to be cast is the latest point at which it is essential to maintain magnesium within the limits for acceptable nodularity. Magnesium content in the pouring basin is a cumulative effect of the previous two operations, what happens in the converter and holding furnace. Magnesium content decreases in an unpredictable way with any delay in any one of the processing steps. Given the fluctuation of magnesium every minute, it is absolutely essential to actually know the exact percentage of magnesium at this point of production. This is important because

To know the exact percentage of magnesium, the sample has to be analyzed spectrographically. It takes time. Any time spent in waiting for the analysis result will cause

The decision to pour or treat has to be made on the shop floor, rather quickly on the spot, otherwise the production will be delayed, with a net loss in revenue.

So in a high production foundry there is a grave need for a quick control tool which would inform the operator about the quality of the molten metal before the metal is poured. There is a considerable cost savings if a procedure is established where by quick analysis of residual magnesium is made available. Is there a way?

Magnesium Analysis:
One direct way to estimate the percent magnesium is spectrometer analysis of the sample cast in a special mold. The poured sample has to be transported from the pouring basin to the location of the spectrometer, where it is ground and analyzed spectrographically for the elements. There is an understandable delay in obtaining the results for a feed back. In a busy foundry this may take any where from 15-30 minutes or more.

An indirect way to insure that there is enough magnesium to cause spheroidisation is metallographic examination of the cooled sample poured from the metal to be cast. This is a widely used method for checking the graphite structure, as spectrometer instrument is expensive for small foundries to afford. However, this method like spectometer is equally time consuming and labor intensive. Besides, the metallographic method only gives the go ahead for the poured sample, but does not indicate where the exactly the magnesium is in the allowable window. Without this knowledge the operator still cannot make a decision as to how to adjust his charge material for treatment to continue maintaining the required magnesium.

Metal cannot be held for the results from spectrometer or metallographic examination to be relayed as this holding will lead to further magnesium loss.

BCIRA in collaboration with L&N had developed a system called MgLab that would indicate whether the melt has sufficient magnesium of 0.04% (or (0.35%). The prediction is based on the cooling curve characteristics obtained in a specially formulated cup containing controlled quantity of sulfur (to combine with magnesium) and tellurium (to make the iron solidify as white iron after removing magnesium as MgS). The results are displayed as Pass/Fail message within three minutes of pouring the sample. A fail sample can have a magnesium of any where from 0 to 0.04 percentage (or 0 to 0.035%) and a Pass sample can have magnesium anywhere above 0.04% (or 0.035%).

This is a very good, quick and easy alternate method. But the most important question in the production floor, particularly when you have the metal in the holding furnace with a constant fluctuation of magnesium, is not just a Pass or a Fail, but what is the exact amount of magnesium in the melt. The exact amount is important because the entire operation depends on the control of magnesium and this information is needed to make decisions to adjust the additions at the starting point of the cycle. This is a vicious circle where one has to be extremely vigilant and make decisions readily to keep up the production of quality castings.

Drive towards a method of rapid determination of magnesium
The converter-presspour system started operation in the beginning of 1997. At that time for magnesium control an old MgLab was used to check the Pass/Fail condition and a procedure was set as to how to react to the message. Many times this wasn’t enough for the operators. A Fail sample could be anywhere from 0.02 to 0.042 and still they had to wait for the spectrometer analysis. For casting in the molds it is okay to go with the metallographic examination; all it implies is that the particular metal in the pouring basin at that instant is good. However, to sustain the magnesium in the holding furnace, they needed to know the precise amount of it in the final metal, so they can adjust their treatment charge.

Having to play with their gut instinct to keep the magnesium under control, this was raised as the foremost important issue in every meeting. So it was decided to devote a person to look into the methods of predicting or analyzing magnesium in the final metal at the pouring station. After a quick study it was decided to explore the same technique of "cooling curve" analysis, which is also the basis of MgLab. Same specially formulated Te/S coated 0.04 commercially available cup was considered. The MgLab unit was first generation unit about twenty years old and had no display of cooling curve or any parameters.

Experiments were run in parallel to MgLab. Molten metal from the pouring basin was poured into the cup and data collected by means of a data acquisition system. Extensive testing on a wide range of compositions was made. Numerous cooling curves were analyzed for special features. Critical parameters were found which were very sensitive to magnesium variation. These parameters reflect the way in which sulfur and tellurium present in the cup affect the nature of solidification of the iron. Based on these parameters, a set up has been established in the foundry, which analyses the cooling curves, measures the parameters and displays the percent magnesium on the screen. The analysis is made within three minutes of pouring.

Table 1 shows the data collected randomly over a period of three months during production (only 25 points listed).

 The predictions are compared against the magnesium analysis by spectrometer. The data shows the results are in close agreement, accurate and reliable particularly in our critical range of 0.025-0.048%. Table 2 shows that most of the results were within 0.004 points from the spectrometer values. 

Comparisons were also made with magnesium analyzed by wet chemical methods and magnesium analyzed by an external lab which is routinely done in the foundry. In all cases the magnesium prediction has been very satisfactory.

The predictions were possible because of the rigid control in chemistry of other elements, manganese and sulfur. Manganese and sulfur can influence the prediction. But in a converter process where there is a natural desulfurisation occurs, sulfur is reduced to a minimum value. Sulfur hardly exceed 0.007% during production, except during start up. Manganese is controlled by the choice of a good quality steel as a charge make up in the base melting. Other tramp elements Mo, B, Sn, Ni are controlled within limits. From the data collected over a period of three months, the following variation in chemistry in the final iron was recorded: C=3.60-3.81, Si=2.55-2.78, Mn=0.262-0.282, Cr=0.027-0.031, Sn=0.007-0.009, Ni=0.024-0.026.

Other elements that fluctuate a little bit are carbon and silicon. Silicon variation mainly comes from the inoculation treatment, which can vary from one casting geometry to another. The combined effect of carbon and silicon as CE may be one restricting factor in the prediction. It is important to have a eutectic or hyper eutectic alloy composition and the prediction. It is important to have a eutectic or hyper eutectic alloy composition and the prediction has been seen to be good for the CE of 4.45 or higher.

Total vs. Active Magnesium
The magnesium analysis given by spectrometer method is total magnesium whereas the magnesium analysis given by cooling curve method is active magnesium.

Magnesium is a very powerful deoxidizer and a desulfuriser. Oxygen combines with magnesium to form magnesium oxide and complex silicates and sulfur in the metal combines with magnesium to form magnesium sulfide. Analysis given by spectrometer is referred to as the total magnesium which includes the dissolved magnesium and magnesium bound as compounds present in the sample. Analysis given by spectrometer does not differentiate how magnesium is present in the sample.

Total Mg = Mg dissolved + Mg compounds (oxide, sulfide, silicate)

Magnesium present as compounds will reduce the available quantity of magnesium to promote spheroidisation. Therefore actual amount of magnesium that is truly contributing to the formation of spheroids is that which is in excess of that which combines with sulfur and oxygen. This excess magnesium is referred as the active or free magnesium

Active Mg = Total Mg – Mg compounds (oxide, sulfide, silicate)

It is the active magnesium that controls the structure of graphite. Analysis of active magnesium is very difficult compared to the analysis of total magnesium by spectrometer.

Correlation of total magnesium to active magnesium can be made, such as the one in the present work, provided

Photo 3

Further developments In the last two years after the initial setup was made in the foundry, STAS Canada, has developed and built a compact thermal analysis unit to measure percent magnesium in iron. Figure shows the photo of the ProcessLab (Photo 3) unit for magnesium. The unit is simple, automatic and has standard features of storing and exporting results. In addition to giving magnesium within three minutes, the unit also gives a Go/noGo message (Photo 4) within two minutes of pouring. Photo shows the Go message (or a noGo message as the case may be) and the percent of magnesium 0.047 in this case. The number 2 seen on the display indicates that the results are from input or channel No. 2.