Professional Documents
Culture Documents
1999 Ductile Iron Production - A Comparison of Alternative Treatment Methods PDF
1999 Ductile Iron Production - A Comparison of Alternative Treatment Methods PDF
by
Manager of Research
N-4675 Kristiansand
Norway
Tel. + 47 38 01 70 00
Fax. + 47 38 01 74 94
Ductile Iron Production - A Comparison of
Alternative Treatment Methods
Dr. T. Skaland
Abstract :
The list of variables that influence recovery includes: metal temperature, type and
size of nodularizing material, quantity of metal being treated, rate of tapping, and
treatment method. This paper will focus mainly on those treatment methods that have
proven to be the most commercially successful. These are known as the sandwich
process, the tundish cover process, the flow through process, the in-mould process,
and the cored wire process.
1. INTRODUCTION
Magnesium treatment for nodularization of ductile iron can be carried out by several
different treatment processes. Magnesium is the most popular nodularizing agent,
and it is usually added in multi-component alloy form. These components contain
additions to reduce reaction violence, promote graphite spheroidization, neutralize
the effect of impurities on graphite morphology, and control the matrix structure. The
most common alloys for nodularizing iron are ferrosilicon containing 3 to 12 per cent
magnesium. Magnesium and its alloys possess low vapourization temperatures, and
consequently their addition to molten iron must be done with extreme care.
The treatment is a means of modifying the solidification structure so that the graphite
phase precipitates and grows as spherical particles instead of flakes, thus resulting in
a cast iron with significantly improved mechanical properties. The nodularizing action
can be regarded as a simultaneous desulphurization and deoxidation treatment,
where elements having strong affinity to both sulphur and oxygen are added. When
dissolved sulphur and oxygen are removed from the liquid, graphite growth will
proceed as nodules according to its crystallographic features.
2. PROCESS HISTORY
During the years of ductile iron history from the early 1950s a range of different
treatment processes has been developed. Some were found acceptable while others
are rejected a long time ago. Some of the early processes involve lance injection and
stirring of the ladle, but most of these are not seen in foundries any more. The most
widespread processes today are the group involving overpouring of an MgFeSi-alloy
in some kind of treatment ladle. Various modified ladles to improve magnesium
recovery has been designed. Some of the best alternatives also involve application of
ladle cover lids. /1/
Finally, the last process to join the family is the cored wire process. This process has
been gaining popularity during the last 10 years, and it represents a kind of process
where the magnesium material (metal or alloy) is packed in a hollow steel tube that is
injected into the liquid iron. The tube dissolves in contact with the hot metal, and the
magnesium reaction goes on for as long as it takes to inject the required length of
wire.
Overpour or sandwich
Tundish cover
In the mould
Flow through
Plunging
Converter
Cored wire
The most important characteristics, possibilities and limitations for each process are
summarized to give a total evaluation and comparison of all processes. This is an
interesting overview for foundries that want to start up ductile iron production in
helping to choose the right process concept for a specific purpose. Also, the overview
gives a good basis for foundries that want to improve the efficiency of an existing
ductile iron process in showing which alternatives that easily can be adopted and
what improvements and investments this may involve. /2,3,4/
The left column of Table 1 lists the various characteristic features that can be
explained as follows. The treatment agent suited to process shows which
processes that apply a Ni-base alloy, an MgFeSi-alloy or a Mg-metal addition. The
Mg-content shows the normal Mg in each treatment agent that is suited to the
respective processes. Mg-recovery shows the range of Mg-yield expected for the
various processes. The relatively wide recovery-range for some processes will of
course also result from affection by various other process parameters such as
treatment temperature, vessel design, alloy selection and base metal composition.
The cost of plant shows whether any investments are required to adopt the
process. The basis is a plain ladle, and any costs will come in addition to this. For the
overpour processes no investments are necessary, while for the converter process a
completely new treatment unit with associated fume extraction system at high costs
is required. The general process flexibility describe how much freedom to alter
treatment size, location, preparations, iron composition, alloy addition rates, etc. that
each process allows. Fume emitted shows how much magnesium fume that will be
released from each process. A ranking of processes reveals that those applying high
Mg-containing materials (as Mg-metal) generate far more fume than those applying
MgFeSi-alloys with about 5% Mg. This also means that a separate fume extraction
system is required for violent reactions. The only two processes that, according to
present strict environmental regulations, most likely not will require separate fume
extraction is the tundish cover process and the in-mould process.
The inoculation effect from treatment is also related to the violence of reaction,
and in this respect also to the degree of sulphur removal. For processes starting out
with a high base sulphur level (i.e. converter, wire) and removing most of it during
treatment, the iron will be hard to inoculate. This is because most of the potential
sulphide nuclei particles are removed from the iron during the violent reaction. Less
violent reactions (tundish cover, in-mould) will virtually not remove any of the base
iron sulphur but rather tie-up sulphur as a high number of small sulphide nuclei
particles giving a good basis for highly effective subsequent inoculation.
There will be a possible risk of Si built-up for all processes applying a Si-containing
treatment agent (MgFeSi-alloys). Foundries using a high fraction of cast iron returns
and other Si-containing charge materials may experience problems with their silicon
balance. For such foundries a Mg-metal process can be an alternative (e.g. cored
wire) or it may also help to improve the existing MgFeSi-process (significantly lower
alloy addition) and thereby lower the total addition of Si-bearing alloys to the iron. In
turn, this may also give a cleaner metal (less slag and dross), less rejected castings,
and consequently less returns to disturb the silicon balance.
The size of production unit suited to process means the size of foundry that
normally will be suited to adopt a certain process. Generally, most MgFeSi-processes
can be applied in both small and large foundries, while investments associated with
the converter and core wire requires a larger foundry to accept the extra costs.
Finally, licence/royalty has to be paid both for the flow through and converter
processes to the parent companies that have developed and patented them.
4. THE SANDWICH COVER PROCESS
The term sandwich is derived from the fact that a covering material or sandwich layer
is added on top of the spheroidizing alloy prior to the addition of molten iron to retard
the start of alloy reaction. A 1% to 2% addition of steel punchings is most widely used
as cover material although other materials as ferrosilicon is also used. In addition to
the delaying action, the cover material cools the metal in the immediate vicinity of the
magnesium reaction. This cooling effect takes advantage of the fact that magnesium
recovery increases as the metal temperature decreases. Recovery of magnesium is
normally fairly good with the sandwich treatment and may range up to 50-60 per
cent. Advantages are short treatment times, simplicity, flexibility, and the improved
recovery with less slag and fumes. The disadvantages are slightly greater
temperature loss due to the additional heat required to melt the 1% to 2% cover
materials. Figure 1 shows an example of a sandwich ladle design including alloy
pocket at the bottom. /1/
The first published information regarding the use of a covered ladle to produce ductile
iron was presented in 1978, although the precise origin can be traced back almost 15
years prior to this. Magnesium losses to MgO are greatly reduced because no new
oxygen can enter the ladle once the filling orifice of the tundish is covered with metal.
Magnesium recoveries can be as high as 70-75%, a considerable improvement over
the plain sandwich method.
In the early development stages of the tundish ladle, production foundries were
encouraged by the improved magnesium efficiency but discouraged by the need to
continually replace the cover with each treatment. This drawback was overcome by
incorporating a semi-permanent, bolted-on cover or a separate lifting lug system that
did not have to be removed after each treatment. Figure 2 shows an example of a
simple tundish cover ladle design. /5,6/
Preferentially, the height:width ratio of the ladle should be at least 2:1, with larger
ratios improving the process even more. The optimum treatment vessel for a tundish
cover process will be like a long tube divided vertically for lining maintenance. The
tundish cover process is suitable for all treatment sizes ranging from 100 kgs to 10
tonnes.
The tundish cover lid may be designed to suit a range of different foundry conditions.
Examples are fixed lid, automatic lift or manual removable lid. The metal outlet from
the lid should be sized to suit the amount of iron being treated. When a divider wall or
alloy chamber is applied, the metal stream must fall into the part of the bottom not
containing the alloy. By such a process, very high recoveries and good process
consistency can easily be obtained provided the right alloy is applied. Consequently,
low levels of fume and smoke will escape from the vessel, giving a good foundry
environment.
An advantage for this process concept is the design of two symmetric alloy chambers
at the bottom of the ladle. By such dual alloy chambers, the ladle can be more or less
self cleaning with respect to slag built-up. The self cleaning is done by shifting
between alloy charging and pouring in both chambers. As the alloy chamber is
growing with slag, the lid is shifted 180 and the clean pouring chamber is then
applied for the alloy. The slag contaminated alloy chamber will then be cleaned out
again by the hot and turbulent metal stream when located directly under the tundish
hole.
Figure 3 shows some alternative ways of designing tundish cover ladles with fixed lid.
A fixed lid requires an alloy charging port for introduction of MgFeSi-alloy. Also, it is
recommended to put the alloy in a pocket at the opposite side of the ladle bottom to
where the metal stream hits during filling. Ladles with fixed lid can generally not be
deslagged, and therefor slag accumulation may be a problem during time. Hence,
low slag forming alloys (low Ca, Al and RE) is recommended for such ladles to
minimize slag formation inside the ladle. /7/
The tea pot ladles offers an advantage in that liquid iron can be filled through the tea
pot spout. This means that the tundish cover basin is eliminated. Both fixed and
removable lids can be applied for tea pot ladles. For ladles operating with a single
spout, there is a possible risk that the last iron will not pick up magnesium. This iron
will stay in the spout until teeming, and if used directly into a mould it may be
undertreated or even grey iron. To avoid such problems a double teapot with a
separate filling and pouring spout can be designed. Tea pot ladles will generally
supply cleaner metal into pouring ladles and autopours due to the slag skimming
effect of the spout. Figure 4 shows examples of tea pot ladles. /7/
The lifting or removable tundish cover design is probably the most flexible and
easiest to maintain. Lifting covers can be either integrated to the ladle with a
separate lifting lug or constructed as a fully removable lid by using for instance a fork
lift truck. The lid can also be mounted onto the furnace spout. The major advantages
with removable lids are easy deslagging and cleaning of the ladles and also that the
same lid can be used for several ladles. Charging of MgFeSi-alloy is also easier
when the lid is removed from the ladle top.
The disadvantage of a lower casting yield is apparent because the reaction chamber
is a part of each runner system. The increased possibility of slag and other reaction
products to enter the casting is also a disadvantage. There probably is a greater
need for testing nodularity of every casting produced due to the fact that each mould
is a separate treatment. An advantage is elimination of the need for subsequent
inoculation addition since the process itself gives a very good nucleation effect. /8,9/
Over the last few years, cored wire has increased its popularity for nodularization of
ductile iron. Normally, the hollow wire tube is filled with either magnesium metal or a
high magnesium containing alloy. This wire is then injected into the liquid iron by a
feeding machine. Advantages are reported to be predictable Mg recovery, availability
to desulphurize and magnesium treat in one ladle, tailor made wire chemistry to fit
individual applications, greater silicon level flexibility in base iron. Normal magnesium
recoveries are found to be between 30-50%. Disadvantages are large amounts of
slag and fume associated with a very violent reaction. Also, the inoculation effect may
be largely reduced by the violent reaction since many of the nuclei are separated.
/12,13,14/
9. ANALYSIS OF MAGNESIUM
Most ductile iron foundries apply either a spectrograph or spectrometer to measure
the so-called residual magnesium content. From experience, a certain lower limit of
magnesium is found at each foundry, that provides the required nodularity in their
castings. This limit may vary significantly from one foundry to another. While one
foundry experience excellent nodularity at 0.025 per cent magnesium, another may
need 0.06 per cent magnesium to obtain good nodule structures. These differences
result from certain variables that will be explained in the following.
First, the total analytical or residual magnesium content in ductile iron is comprised of
three different contributions. Both dissolved magnesium, magnesium containing
micro-particles, as well as occasional slag or dross particles will contribute to the total
analytical magnesium found. There exist no good analytical method separating
between these three contributions, and hence magnesium analyses in foundries will
represent both dissolved Mg as well as combined Mg. The fraction of micro-particles
and slags in a sample will vary significantly from one treatment condition to another,
resulting in the variations in analytical magnesium observed between foundries.
Another important factor affecting the analytical magnesium values is the applied
sampling technique. If a sample is taken from the top of a waiting ladle and from the
last liquid to be poured, different Mg-levels will be found. Micro-particles and slag
particles will float in the liquid according to Stokes' law, the larger particles floating
much faster than the smaller. This will result in an inhomogeneous distribution of non-
metallics in the iron during time, and the analytical magnesium will vary accordingly
as a function of time and sampling position. These phenomena will also contribute to
the variations observed from foundry to foundry, since time factors and sampling
techniques may be quite different.
If a tundish cover lid is employed in addition to the sandwich cover process, fume and
flare will be eliminated and the process largely improved. A reduction to about 1.3
wt% alloy addition may be possible still obtaining good ductile iron. If the tundish
cover process is kept unchanged and the alloy sizing is changed from the initial 10-
40 mm to a 1-20 mm, further improvements are possible. A finer and wider sizing will
give a denser alloy bulk packing in the sandwich and by that the reactivity will be
reduced. A further reduction to 1.2 wt% addition may be possible. Finally, if the
calcium content of the alloy is increased from 0.5% to for instance 1.5% the reactivity
will be significantly lowered. Additions can be further reduced to the very low level of
about 1.1 wt% still producing excellent ductile iron structures.
The Mg2Si phase is a reactive compound in liquid iron, although its reactivity is
relatively low compared to pure Mg-metal in contact with liquid iron. Mg2Si will
dissolve during evolution of a certain portion of magnesium vapour bubbles. The
majority of magnesium will, however, due to the locally high Si-concentration dissolve
and diffuse to the nearest S and O atoms to form sulphide and oxides.
The ferrosilicon phases (FeSi and FeSi2) will dissolve without any violence of
reaction, giving a local enrichment of silicon in the iron close to dissolving alloy
particles. The silicon concentration will drop from 45% at the alloy-particle interface
down to about 2.5% in the bulk iron at a certain distance from the particle. The local
high silicon will significantly increase magnesium solubility in the iron during
dissolution of the alloy. This will, in turn, enhance the recovery of magnesium since a
large portion of the magnesium can be absorbed by the high-Si iron for diffusion to
nearest sulphur and oxygen atoms rather than forming Mg-vapour bubbles.
Most MgFeSi-alloys contain about 45% silicon. This is the minimum level required for
the highly reactive elements as Mg, Ca, Ce, etc. to form compounds with silicon
(silicides). Silicides are the only stable combinations for Mg, Ca, and Ce in FeSi-
based alloys. If silicon is lowered to levels below 43%, there will be a lack of Si
available to produce such silicides. Since the Fe-containing phases (i.e. FeSi, FeSi2)
do not combine or dissolve either Mg, Ca or Ce, these elements would be lost during
alloy production at silicon levels below 43%. The small amount of free Si-phase in the
microstructure indicates that a small excess of Si is present.
12. SUGGESTED TARGET ANALYSIS
According to the ISO-standard No. 1083 (1976) there are six main grades of as-cast
ductile iron denoted as 800/2, 700/2, 600/2, 500/7, 400/12, and 370/17, where the
first figure is minimum tensile strength (in MPa) and the second minimum elongation
to fracture (in per cent). Table 2 gives recommended target analysis for the main
elements C, Si and Mn to obtain specified properties. Also, variations in optimum
composition as a function of casting section size ranging from 13 to 100 mm is given.
It is also important to note that for the ISO-grade 370/17, silicon should be lowered to
a maximum of 2.5% if also impact properties are important for the castings. For the
Grades 800/2 and 700/2, copper or tin may also be added to obtain a fully pearlitic
matrix structure with accompanied properties. For optimum ductile iron treatment,
base metal sulphur content should be restricted to maximum 0.020% giving a final
ductile iron sulphur of maximum 0.015%. Also, phosphorus in all grades of ductile
iron should be maintained below 0.03%.
Most cored wires contain either a pure Mg-metal powder or a high Mg-containing
silicon alloy (>30% Mg). This is to reduce the total length of expensive wire needed
for a full ductile iron treatment. As described previously, the high Mg-alloy or Mg-
metal gives a much more violent reaction than a normal MgFeSi-alloy, and hence the
best Mg-recoveries reported in the literature for a cored wire is about 50%. For an
optimized tundish cover process, Mg-recoveries may be as high as 75-80% which
means that the yield of Mg for this alternative is close to the double of that normally
experienced with the wire.
Due to the different Mg-content and reactivity, difference in fume and flare emission
for the two alternatives will be quite substantial. The cored wire process will, since at
least 50% of the Mg is lost, generate large amounts of fume (and slag) which means
that a separate fume extraction unit is absolutely required. Such units can be fairly
expensive and must be included in the total investment costs for the process. The
tundish cover process, on the other hand, generating much less fume will normally
not require a fume extraction unit.
The main advantage for the cored wire process is the option of treating high sulphur
base iron directly without a separate desulphurizing addition. The tundish cover
process will be most efficient if base iron sulphur is restricted to maximum 0.03% S.
Closely connected to sulphur removal is the inoculation effect from the treatment
process. As explained previously, the removal of most sulphur (cored wire) gives a
treated iron that can be hard to inoculate, especially if thin sections are to be poured.
An optimized tundish cover process that virtually not remove any sulphur at all (base
and final sulphur is equal), gives a much stronger basis for a highly effective
subsequent inoculation. Hence, if effective inoculation is a critical parameter this
must be taken into consideration when choosing treatment process.
Another possible advantage with the cored wire process is the small silicon addition
made. If the silicon balance is a problem even at low MgFeSi-alloy additions the wire
can be an interesting alternative.
The cored wire process will normally not be used for treatments smaller than 500 kgs
due to low Mg-recoveries and high process costs. For larger treatments, the Mg-yield
will improve somewhat and the wire may become more feasible. The tundish cover
process has a high degree of flexibility in treatment weight and can easily be adopted
for all sizes of treatment from a few kilograms to several tonnes. Due to all the above
considerations, the cored wire process is best suited for medium to large foundry
units while the tundish cover process can be applied for all sizes of foundries. /16/
REFERENCES
1. Ductile Iron - Molten Metal Processing, 2nd edition, American Foundrymens
Society Publication, Des Plaines, IL, 1986.
3. T.Bex: Ductile iron: one of the centurys metallurgical triumphs, Modern Casting,
February, 1991.
5. T.L.Forshey, et al.: Modification of, and production experience with the tundish
cover for ductile iron treatment, AFS Transactions, pp. 53-57, 1982.
9. C.E.Dremann: New alloys for making ductile iron in the mold, AFS Transactions,
pp. 263-268, 1983.
11. R. Steel: The use of the Flotret process for the production of large castings,
BCIRA Conference - S.G. Iron The Next 40 Years, Warwick, UK, 1987.
12. H.Mair: Advancements of cored wire applications within the steel and cast
metals industries, 5th Int. Ferroalloys Congress, New Orleans, April 23-26, 1989.
13. R.Cairns: Manufacture of ductile iron from cupola melted iron using cored-wire
technology, 89th IBF Annual Conference (Castcon 92), Stratford-upon-Avon, 11-12
June, 1992.
14. J.Rotella and R.Mickelson: Using cored wire in the production of ductile iron,
AFS Transactions, 1991.
13 mm 3.6-3.8 2.6-2.8 0.5 3.6-3.8 2.6-2.8 0.3 3.6-3.8 2.6-2.8 0.2 3.6-3.8 2.6-2.8 0.1
13-25 mm 3.5-3.6 2.2-2.5 0.6 3.5-3.6 2.2-2.5 0.35 3.5-3.6 2.2-2.5 0.25 3.5-3.6 2.2-2.5 0.15
25-50 mm 3.5-3.6 2.1-2.3 0.7 3.5-3.6 2.2-2.4 0.4 3.5-3.6 2.2-2.4 0.3 3.5-3.6 2.2-2.4 0.15
50-100 mm 3.4-3.5 1.9-2.1 0.8 3.4-3.5 2.0-2.2 0.5 3.4-3.5 2.0-2.2 0.35 3.4-3.5 1.8-2.0 0.2
100 mm 3.4-3.5 1.8-2.0 0.8 3.4-3.5 1.8-2.0 0.6 3.4-3.5 1.8-2.0 0.4 3.4-3.5 1.8-2.0 0.25