Development of Theories on Graphite Formation in Ductile Cast Iron.
By:Cees van de Velde
Last revision:January 15, 2004
1. INTRODUCTION
Conducting a survey on the subject of solidification of nodular cast iron is no easy task as the number of publications in this field, along with the different opinions on the nodule formation process, appear endless. However, the outcome of much of my research work, has proved so fascinating and totally different from anything previously published that such a survey has become more that justified. Although it has not been practical to involve all of the available source material in this study, care has been taken to make selections objectively in order to give a fair representation of the theories put forward over the past 43 years. As much use as possible has been made of the original publications and quotations are numbered successively and placed directly following the main text. Examining ideas in chronological order has helped in understanding the reasoning behind today's thinking.
2.0 THE YEARS PRECEEDING THE DISCOVERY OF NODULAR CAST IRON.
Patent rights for the production of cast iron with spheroidal graphite were granted to Adey in 1948 and to Millis, Gagnebin and Pilling in 1949. These dates can, therefore, be regarded as heralding the commercial production of nodular cast iron. Earlier experiments had, however been made with a view to increasing the mechanical properties of normal gray iron. As far back as 1926, for example, a Meehanite-type cast iron with fine flake graphite was developed, offering tensile strengths up to 450 N/mm2 and elongation properties up to 2 % as is shown in fig.1.
Figure 1,Advertisement for Meehanite Metal(1926)
Most experiments were carried out at extremely high superheating temperatures, by using basic linings, strong inoculation and/or a vacuum melting process and, by accident, some produced more or less spherolitic graphite inclusions. The significance of these observations was largely overlooked, meaning that the discovery of nodular cast iron had actually taken place some 20 years prior to the granting of patent rights [Quot.1]. More detailed information on the period 1900-1948 can also be found at this Website. For years, spherolitic graphite separations were regarded as a unique phenomenon present only in carbon-containing iron, nickel and cobalt alloys. It is, therefore, surprising to find some spherolitic inclusions in a nickel alloy without carbon. Photographs in "Metallkunde" by Sachs (1933), depict such structures before and after a desulfurisation treatment of a high sulfur containing nickel alloy. As indicated in fig. 2 one can see that before desulfurisation, the nickel sulfide is concentrated on the cell-boundaries but, after the addition of magnesium, nickelsulfide inclusions are concentrated in ball-shaped form within the nickel crystals as is shown in fig.3 ! Figure 2 shows the structure of cast nickel with 0,8% sulfur content.
Note the nickel-sulfide inclusions along the cell boundaries. Figure 3 shows the changed structure after magnesium addition.
Note the spheroidal nickel-sulfide inclusions.
Other experiments during this early period are also noteworthy. Piwowarsky [2] and Wittmoser [3,4] produced interesting publications. An article by Milles [5] gives insight on developments in America. Keßler and Reuter [6] present an overall view on the complexity of the, now long forgotten, patent situation which followed the discovery of nodular cast iron. A recent article in Modern Castings [6a] by K.D.Millis tells about all the facts that lead to this magnificent invention.
3.0 GENERAL IDEAS ON THE FORMATION OF NODULAR GRAPHITE OVER THE LAST 30 YEARS.
Reasons for the formation of nodular graphite have given rise to numerous hypotheses and speculation over the past 30 years, with various well-known authors conducting surveys.
1965. A review of the formation of spheroidal graphite in cast iron. S.Banerjee [quot.2,3],[7] The literature has been critically examined by listing about 150 experimentally observed facts which are pertinent to the formation of spheroidal graphite. The discussion seems to indicate that the nucleation of a spheroid takes place necessarily in the liquid state even in a Mg-treated cast iron, and its growth takes place by the diffusion of carbon atoms from the source (which could be the melt, austenite or decomposition of cementite) to the spheroid nucleus. This idea is consistent with all the listed observations.
1970 Zur Theorie der Bildung von Kugelgraphit im Gusseisen. B.Lux [quot.4].[8]
The fundamental work of E.Scheil et al by now leaves little doubt as to the formation of nodular graphite. In general, graphite nodules nucleate directly in the melt, a fact that has been clearly demonstrated in centrifugal experiments, in which nodules were separated from the melt. Floating graphite spherolites are often found in heavy castings and the amount of dual-nodules found in the metal structure is higher then statistically could be expected.
1990 On the solidification kinetics of Spheroidal Graphite Cast iron. Stefanescu [quot.5,6],[9].By 1990 it has been widely accepted that growth of the eutectic starts with nucleation and growth of graphite in the liquid followed by early encapsulation of these graphite spheroids in austenite shells. Further growth takes place by the diffusion of carbon from the melt through this shell. It has recently been shown that austenite dendrites play a significant role in eutectic solidification of nodular cast iron and that it is possible for austenite dendrites to grow partially independent of graphite spheroids.
4.0 THE DEVELOPMENT OF THE BASIC THEORIES.
As soon as the possibilities and advantages of the "new" material were recognized, developments followed swiftly with cerium being introduced as a nodularizing element by Morrogh and NiMg proposed by the International Nickel Company. While reliable production of cast iron with nodular graphite appeared to be within the reach of every well-equipped foundry, little was, however, known about the mechanism behind the actual nodule formation. Some theories were developed in the early fifties based on experience gained during the production of the new material but also strongly influenced by the existing ideas about the mechanisms of graphite formation in gray iron.
4.1 Cementite-theory.
In the early years, heat-treatment was essential to obtaining a nodular structure that was fully free of cementite. It is, therefore, not surprising that one of the first theories was based on the view that nodules were formed by the actual decomposition of cementite. This theory had, after all, been in existence for more than 40 years with regard to the formation of flake graphite in cast iron. Such thoughts prevailed even after a nodular structure was produced directly in the as-cast condition. According to Morrogh [10], one of the pioneers in this field, a graphite nodule was never formed in the melt, but always by the decomposition of cementite [quot.7]. Dunphy and Pellini [11] stated that nodules are formed in the neighborhood of existing dendrites, immediately before the start of the eutectic reaction. Local regions of supersaturated melt are formed close to the dendrite-arms and it is in these local liquid regions that the nodules form. Further growth of the nodules takes place by the decomposition of the cementite, build up around the nodule (see fig.4). Figure 4, nodule formation acc.to Dunphy.
4.2 The Austenite-theory.
De Sy [ref.12] observed that dendrites, found in shrinkage cavities, showed almost the same chemical composition as the base-material [quot.8]. For this and many other reasons, he believed that nodular iron solidified without eutectic, but in a mixed crystal system, the latter being formed by supersaturated austenite. Precipitation of graphite particles, therefore, occurred within the austenite dendrites with further growth taking place by a peritectic transformation or reaction. Formation of nodules within a supersaturated austenite crystal according to De Sy is shown in fig.5. Figure 5, nodule formation within (super)saturated austenite.
Figure 6, nodule formation within (super)saturated austenite acc. to Wittmoser.
4.3 The Melt-theory.
Scheil [17] is convinced that nodules form in the melt, independent of existing austenite dendrites [quot.11]. The nucleation and initial growth takes place in the melt and upon reaching a certain dimension tey are encased in a shell of austenite. Further growth takes place by diffusion of carbon from the melt, via the austenite shell to the existing nodules.
If an austenite encased nodule collides with an austenite dendrite, both structures melt together to form a new structure in which the original shell and dendrite cannot be detected. Fig. 7. shows nodule formation according to Scheil as explained by Wlodawer [69]. Figure 7 nodule formation acc. to Scheil. Figure 8 preferred nucleation sites for graphite nodules between austenite dendrite arms (Patterson).
4.4 The Bubble Theory.
In the early days many experiments, using a variety of alloying elements, were conducted in order to obtain a spheroidal graphite precipitation. The only similarity between these alloying additions was the fact that they all could be converted into a gaseous form by the molten iron. This fact lead to the development of the so-called Bubble-Hypothesis, first proposed by v.Nieuwland [19], but better publicized by Karsay [20,21]. The Bubble Theory expounds that the tiny bubbles in the liquid metal, created by the gas evolution, are ideal sites for nuclei which will give rise to the growth of graphite nodules. In this theory, the graphite grows radially from the outside into the bubble as is indicated in fig.9. Figure 9 Gasbubbles as a template for the final graphite shape.
4.5 Graphite-morphology.
Other researchers disregarded the question of when and where nodules form, but looked instead for the reason why graphite separates in a spheroidal shape. Marincek [22] for instance, supposed that high surface tension was responsible for the ball shape [quot.13], fig.10. Hillert en Lindblom [23], on the other hand, assumed that the crystal growth of the graphite was influenced by the presence of disturbing atoms , derived from the nodularising elements, in the crystal structure. Figure 10 The typical ball-shape is caused by high surface tension acc. to Marincek.
5.0 COMPARISON OF TODAYS VIEWS WITH THE BASIC-THEORIES.
Theories so far examined, give a picture of early thoughts on the way nodules were formed in cast iron. If today, more than 40 years later, we compare recent statements by Stefanescu [9] with the basic-theories of the past, we can distinguish two important facts:
A) Of the original basic ideas, only one -the Melt-Theory- has survived.
B) Austenite dendrites have only recently been believed to play an important role in the formation of graphite nodules.
Conclusion (A) is easy to accept as it is quite natural for the "most suitable" theory to survive.
Conclusion (B) is, however, much more difficult to understand as the majority of the basic-theories, proposed in the fifties, emphasized the importance of austenite dendrites on nodule formation.
Why, therefore, do we now rediscover in the nineties, what was seen as a starting-point in the fifties? What happened in the intervening years?
6.0 THE STRUGGLE FOR THE RIGHT THEORY.
6.1 Compromises.
The coexistance of theories all attempting to explain the same phenomenon in completely different ways was bound to break down. However, at first, theorists tried to unify the different theories. Morrogh [24], for example assumed that depending upon the chemical composition and nodule-count, formation of nodules could result from decomposition of cementite as well as directly from the melt. By 1961, however, Morrogh [25] has changed his views and accepted the melt-theory . The serious rivals, Scheil and Wittmoser, tried to reach an agreement in a publication of 1958 [26]. However, it turned out to be a very poor compromise and in the section of interest for the foundry industry ( C%>2.5), differences in opinion remained. Motz [27] also supposed different formation mechanisms, depending on the composition of the material.
6.2 Preference for the Melt-theory.
In general, a lot of faith was placed in the Melt theory which closely resembled ideas on the formation of flake graphite in gray iron. Furthermore, the theory was easy to understand and difficult to disprove. The pioneers of the Melt theory, Scheil and Patterson, strengthened their theory in various publications [28,29,30]. Loper en Heine [31] pinpointed the start of graphite formation as well above the eutectic temperature and thus, evidently, in a liquid state. Loper [32] established the fact that at the eutectic solidification, the amount of nodules remains constant, thus determining that nucleation must take place before this point is reached. He also reconfirmed the growth of nodules by solid state diffusion. According to Oldfield en Humphreys, nodules grow directly in the melt, without the interaction of carbides [33]. Kempers studied the flotation of nodules when separated in the melt [ref.34]. Schöbel reports on results obtained after centrifuging nodular cast iron and his research projects clearly reveal the presence of nodules in the melt [35]. Henke describes the formation of nodules from a nucleus with further growth taking place by diffusion and reconfirms the already almost generally accepted view on the solidification mechanism [36]. O.Tombrock proved the existence of free nodules in the liquid melt [37].
6.3 The Melt-theory, the Winner.
Support for the Melt theory increased, albeit not as a result of new discoveries but rather because alternative theories were actively disproved. Amongst the latter, were the cementite and austenite theories.
Cementite-theory.
It was soon shown that cementite was not necessary for formation of a spherolite as graphite formation never took place in the carbon-rich cementite, but always in the austenite phases.
Austenite-theory.
The formation of nodules from supersaturated austenite had already been doubted during the "compromise-periode", but lost support quickly and logically when it was proven that supersaturated austenite did not in fact exist.
By elimination, therefore, the melt-theory remained and could be regarded as the "winner" especially following Lux's publication, which gave a good overall view on the thoughts of the period. The remaining theories e.g. bubble and morphology theories, fit in well with the Melt-theory as, although the bubble-theory has always been viewed apprehensively, it is almost impossible to disprove [quot.14].
7.0 CONCEQUENCES OF ACCEPTANCE OF THE MELT-THEORY.
General acceptance of the melt-theory has, without doubt, strongly influenced further research.
7.1 The disappearance of dendrites.
Once it was generally accepted that nodules formed directly in the melt, other structural components, such as dendrites, were no longer needed for nodule formation. Soon, they were to disappear from all theories. Scheil [29], for example remarks: "To keep it simple, we will speak about an austenite region, although it actually is an austenite dendrite." [quot.15]. And: "Dendrites are deformed by the nodules, in such a way that they cannot be recognized afterwards."[quot.16].
Dendrites also gradually disappear from drawings, as can be noted in Morrogh's schematic of different stages in the solidification of nodular iron, as can be seen in the following figures:
Fig.11. Successive stages in the solidification of nodular cast iron. Primary dendrites are left out. Morrogh,1961 [25]. Fig.12. The solidification of nodular iron (schematic). Henke,1967,[36]. Fig.13. Various stages of the eutectic in nodular cast iron. Jolley,Holdsworth,1971,[38]. Fig.14. The solidification of the graphite-austenite eutectic in nodular cast iron. Marincek,1980, [39].
Note the changes in subscripts.
Dendrites in gray cast iron have always suffered from a rather negative association with inferior mechanical properties. Aside from this, most of the chemical compositions used for the production of nodular cast iron are almost eutectic, so dendrites should not be present in the first place.
However, as dendrites are lost from sight, one possible influence on nodule formation, is likewise dismissed.
7.2 Increased interest for Graphite Morphology.
One question put forward by many researchers and foundry-men enquired as to why graphite crystallized in the melt sometimes in the shape of flakes and, at others as nodules. Answers were provided mostly in the form of very complex theories which were unintelligible to the average practically minded foundryman.
In 1954 Marincek [22] put forward his ideas about the influence of surface tension on nodule formation. He proposed that as a result of a high surface tension, the nucleus tries to minimize its volume/surface ratio, giving rise to a ball-shaped graphite. Hillert en Lindblom [23] believed that the typical spheroidal shape was the result of screw dislocations in the graphite lattice, caused by encapsulation of foreign atoms. Stadelmeier [40] assumed that nodules form in a gas bubble which is created by the evaporation of the nodularising elements. Patterson [30] emphasizes the importance of interfacial energy during the formation of graphite nodules. Minkoff [41] states that the growth mechanism of graphite is influenced by the absorption of foreign atoms. Herfurth [42] found out that interfacial energy plays an important role in graphite morphology. Kalvelage en Pötschke [43] propose that growth of graphite is altered by the absorption of nodularising elements on to the prism plane of a graphitecrystal. Liu et al. [44] concluded their SEM research by stating that graphite precipitation can gradually change from one type to another with the definite shape dependent, more or less, on growth in one preferred direction. Sy-Sen et al. [45] assume a branched spiral growth mechanism for graphite nodules. Zhu Peiyue et al. [46] state that a twin/tilt mechanism during the growth of graphite is responsible for the definite graphite shape. Purdy en Audier [47] used the electron microscope for their research into graphite nodules. They discovered many crystal defects, built-in to the lattice which were probably due to impurities. Minkoff [48] reports on an instability mechanism as the reason for the differences in graphite shape. Fig.15 shows the crystalline structure of graphite giving rise to many growth mechanisms.
Fig.16 gives an impression of the changes in graphite growth, caused by the absorption of foreign atoms. According to Minkoff.
Even today, however, hypothesis in this field continue.
7.3 Increasing interest for graphite nucleation.
Whilst it is agreed that a nodule forms directly in the melt, questions arise as to where exactly growth begins. It is generally accepted that the starting point for solidification takes place on a nucleus and abundant inclusions of various chemical compositions, have been examined in order to identify the necessary nuclei. De Sy [49] suspects that compositions based on magnesium, formed by the decomposition of a complex magnesium-silicon-carbide, form the basis for the nucleation in nodular cast iron. Bakkerus en Rosenstiel [50] analyzed the metallic inclusions that are often found within nodules but have discovered no difference in chemical composition between these inclusions and the base material. A definite conclusion on the nucleating effect of an mg-si core could not, however, not be reached from results available. Weis [51] came to the conclusion that no one specific substance is responsible for nucleation, but rather that all suspensions present in the melt might have a nucleating effect. The only requirement is that substances must supply the energy levels for the growth of graphite. Lalich en Hitchings [52] point out that Magnesium-calcium sulfide inclusions, complex silicates, sulfides and sulfides of rare earth metals are potential nuclei. Which composition becomes active, depends on the applied melting-techniques and the type of nodularising alloy used. Wallace en Maier [53] propose sulfides of rare earth, strontium, calcium and barium as effective nuclei. Ueda et al [54] reach the conclusion that the nucleation effect increases the more the graphite and the potential nucleus resemble their crystal structure. Tartera [55] proves that sulfur is essential in the nucleation of gray iron. Kayama en Suzuki [56] show that undissolved graphite particles play an important role during the nucleation of gray iron whilst sulphur delays the break-down of graphite nuclei. Wang en Fredriksson [57] suspect a possible homogeneous graphite crystallization, caused by an inhomogeneous silicon distribution when inoculants are dissolved. Itofuji en Uchikawa [58] examine the formation mechanism of chunky-graphite. Inclusions of the type Mg-Ca-S-Si-O are found in the center of nodules. The authors consider that the occurrence of these inclusions is purely accidental and does not constitute proof of any nucleating effect. Nakae et al [59] are convinced that by using a calcium containing inoculant, the calcium acts as a catalyst for the formation of cristobalite, which on its turn act as a nuclei for graphite precipitation. Azzam et al [60] have analyzed the reactionproducts of impurities obtained from melts treated with rare earth metals. The substances were found to be suitable substrates for graphite separations. With research ongoing, a list such as this requires constant updating as, at regular intervals, new substances are found which may prove to play a role in the nucleation of graphite spherolites.
8.0 RETURN OF THE DENDRITES.
According to the basis-rules of the melt-theory, every graphite nodule is considered to be an individual structure and, as such, is not deemed to have any functional relationship with his surroundings other than by pure accident. For many years, this idea has held fast when the structure of nodular cast iron has been observed.
8.1 Dendritic fractures.
Fractures, sometimes found on knocked off risers, reveal a far more regular structure than could be occasioned by a random solidification pattern. Such fractures can clearly be seen, often with the naked eye, to have a definite crystalline structure, as described for the first time by Haik [ref.61]. Some years later, Loper and Heine [62] conducted further examinations and described the phenomenen extensively, referring to the process as "spiking", as it is comparable with a characteristic found in malleable iron. Parks and Loper [63] also concluded that certain oxidizing circumstances during melting and storage of metal strongly promote a dendritic growth.
8.2 A re-examination of the Fe-C-Si diagrams.
Following renewed research in this field, Olen and Heine [64] reached the conclusion that something was wrong with the existing Fe-C en Fe-C-Si diagrams. An urgent need to re-examine these basic diagrams exists.
8.3 Spiking in the eighties.
Many publications on spiking appeared during the Eighties, indicating a renewed interest in this subject.
Verlinden and Kikkert [65] claim that numerous factors contribute to the occurrence of a dendritic structure and for the first time in many years, dendrites reappear in the well-known solidification picture as can be seen in fig. 17. Figure 17 Dendrite formation during the solidification of nodular cast iron acc. to Verlinden and Kikkert [65].
8.4 Back again!.
Aside from examination of the abnormal spiking structure, other researchers had been looking out for the lost dendrites. Wlodawer [69] was probably the first to establish the fact that in virtually all normal specimens many nodules were seen to show a certain mutual direction. Nodule connections seemed to occur in lines, circles, arcs or as positioned under certain angles. In the case of mutual orientations such as these, dendrites play an important role [quot.18]. The spherical conglomerates, comprising a large nodule, surrounded by many smaller nodules, is viewed as being the nodular equivalent of the eutectic cell that is recognized in gray iron and is represented in fig.18. Fig.18 Conglomeration of nodules acc. to Wlodawer [69].
Engler en Ellerbrok [70] found nothing but dendrites. Their idea on the eutectic solidification of nodular cast iron is shown in fig.19. Fig.19 Eutectic solidification of nodular cast iron acc. to Engler and Ellerbrock [70].
Nieswaag [71] re-introduces dendrites in the well-known solidification picture in his explanation of the solidification process of nodular cast iron. Fig.20 Solidification of nodular cast iron acc. to Nieswaag [71]: 1=liquid;2=primary austenite dendrites;3=nodule formed in the liquid;4=austenite-shell around the ndule.
Using the so-called tracer method, Rickert and Engler [72] established the fact that austenite dendrites play a dominant role in the solidification of eutectic nodular graphite iron (fig.21). Fig.21 Solidification morphology of eutectic nodular cast iron acc. to Rickert and Engler [72].
Zhou,Schmitz and Engler [73] remark that the formation of dendrites in hypereutectic compositions is common and can be regarded as normal. Fig.22 Nodule formation in hypereutectic nodular cast iron acc. to Zhou et al [73]:1=flotation of primary graphite nodules;2=austenite dendrites;3=nodule formation;4=carbon-rich region;5=direction of heat flow;6=interdendritic region;7=growth direction.
Motz and Wolters [ref.74] think that dendrites, as well as austenite shells, can encapsulate nodules. When solidification has taken place no difference between either constituent can be discerned as indicated in fig.23. Fig.23 Schematic of the eutectic solidification in nodular cast iron acc. to Motz and Wolters [74]. (a)Primary crystallisation, occurrence of convectional flows in the melt;(b)Crystallization of the other eutectic phase, particularly in supersaturated regions, gravitational segregation;(c)Progression of the eutectic solidification, gravitational segregations;(d)formation of solidified, completed and related residual melt regions.
In their research project on directionally solidified nodular iron, Stefanescu and Bandyopadhyay [ref.9] reach the following conclusions:
1) At the eutectic temperature, austenite dendrites and nodules form independently of each other in the liquid.
2) Growth of nodules in contact with the liquid is limited.
3)Graphite nodules collide with dendrites either by means of flotation or convection.
4)Nodules can be encapsulated in austenite either before or after collision takes place.
5)Growth of nodules by carbon diffusion through the enveloping austenite shell occurs only after nodules have become attached to the austenite dendrite.
6)Austenite dendrites grow partially by carbon diffusion, but primarily because of undercooling and supersaturation of the melt.
An artist representation of the solidification front with various graphite shapes according to Stefanescu is shown in fig.24. Fig.24 Artist impression of variations of the solidification front with various graphite shapes.
9.0 Back., but with a difference.
Although it has taken nearly 30 years, the once-important dendrites are back in the picture. There is, however, a significant difference. Whilst in early theories dendrites played an important and even an active, in modern thinking they play a passive part as today the melt theory dictates the formation of nodules in the melt. Dendrites are in fact considered to be something of a nuisance as they consume nodules, dictate their orientation and therefore obstruct regular distribution. Obvious, it is necessary to critical re-examine the proofs upon which the Melt-theory is based. By so doing, it may be easier to determine the reason for its popularity and general acceptance.
10.0 THE REAL PROOF OF THE MELT-THEORY.
10.1 The actual proof.
Schöbel's publication "Precipitation of Graphite during the Solidification of Nodular Cast Iron "[35] in 1964 constituted an international breakthrough in recognition of the Melt-theory. The paper, which gave an overview of the experimental work of Scheil and Schöbel, was based on the two following assumptions:
1) At the start of the eutectic solidification, nodular graphites precipitates. Shortly after that, the nodule is enveloped by an austenite shell and further growth occurs by diffusion of carbon from the melt, through the solid austenite to the existing nodule. By measuring the radii of the austenite envelopes and the associated graphite nodules in samples, whose solidification is interrupted by quenching during the eutectic solidification, one can predict the nodule precipitation from the melt [quot.19].
2)Graphite nodules formed directly in the melt will float to the surface because of the difference in density. For hypereutectic compositions this is well-known. In the case of hypoeutectic melts, austenite crystals will show a tendency to sink towards the bottom of the melt while the graphite nodules will try to float towards the surface. The effect of flotation can strongly be increased by increasing the gravitational forces .
The entire basis of the Melt-theory has been based on the above points. Schöbel proved that the solidification of nodular cast iron when exposed to extremely high gravitational forces (270 G) is strongly influenced. He noted that graphite nodules could be traced beneath the ( with molten glass covered ) surface of hypereutectic nodular iron but offered no comment as to the nature or composition of these carbon separations. Similar experiments applied to hypoeutectic compositions revealed a concentration of graphite nodules in the direction of the gravitational forces. The frequency distribution of graphite shapes (small deviations in spheroidal shape) showed differences between centrifuged and non-centrifuged samples. At an earlier date, Wittmoser [14] had also examined graphite particles obtained by means of centrifugal separation and had come to the conclusion that the majority of these nodules were in fact, hollow shells. Experiments undertaken in 1960 by O.Tombrock [37], also reconfirmed the existence of individual nodules in liquid iron. In this research, a small amount of molten nodular iron was intimately mixed with molten silver. As the mixture solidified, the silver -owing to its low solubility in cast iron- separated. The test was based upon the theory that if independent graphite nodules had existed in the liquid, some would have been captured in the solidified silver. One can, therefore, conclude from the above observations that free "nodules" have only been observed under testing circumstances where high mechanical forces play an important role. It is questionable whether or not these outcomes can be related to normal conditions.
10.2 The mysterious austenite-shell.
Formation of a shell of austenite around an existing nodule is an essential factor in the melt-theory. The eutectic solidification meant the formation of two phases, graphite as well as austenite. The formed austenite was put away around the nodule. At a later date, this assumption became more flexible, in that the enveloping capability was also believed to be a role of primary dendrites. The first description of a shell formed around a nodule can be found in a publication of 1953 by Scheil who was trying to find a compromise between the Melt- and the Austenite-theories: "Let's assume that nucleation takes place in the melt, but that at the same time or soon afterwards, the small graphite nodule is surrounded by a shell of austenite, then we are dealing with the same mechanism." [quot.21] The formation of an austenite shell around a nodule has been studied by many researchers, most of whom have been strongly influenced by viewing the microstructure as it appeared after quenching. An example of such a quenched structure that gave rise to the enveloping mechanism is shown in fig.25. Figure 25 Quenched nodular cast iron. The formation of graphite nodules just started. Primary austenite dendrites can be seen, together with embryo nodules,each surrounded by an envelope of austenite. In the interdendritic positions, there is white iron arising from the rapid quenching of the untransformed liquid. (Morrogh [25]).
Figure 26 shows the general idea of nodule formation by a growing austenite shell. The nodule grows in contact with the melt and is encapsulated by austenite, and further growth of the nodule within the shell takes place by solid-state carbon diffusion. Fig.26.Mechanism of shell formation around a nodule:a=growth of a nodule in contact with the melt;b=encapsulation by austenite;c=growth of a nodule within the shell by solid-state carbon diffusion
>Tiller [75], however, has his doubts about such shell formation, his concern revolving around the following factor. Diffusion of carbon within the liquid is 20 times faster then in solid austenite. It is, therefore, to be expected that growth of a nodule, as long as it is in contact with the liquid, should be swifter and be preferred above a nodule surrounded by a layer of austenite.
Aside from this, ball-shaped austenite-shells growing in an undercooled melt would be extremely unstable meaning that dendrites could easily form on the shell surfaces and that a ball-shape could not be maintained. Therefore, Tiller cannot find a single advantage with respect to nodule growth in the examined concepts of shell formation.
Karsay and Campomanes [76] believe that a weak or missing shell is responsable for the formation of "chunky" graphite.
Cole [77] states that under normal cooling conditions, a shell of austenite does not exist. The shells, found after quenching, therefore, should be attributed to human intervention.
Jolley [38] discovers channels within the austenite shell and proposes that a growth mechanism by carbon transport through these channels takes place as shown in fig.27.
Figure 27 growth of a nodule through channels in the austenite shell acc. to Jolley and Holdsworth [38].
Lux,Mollard and Minkoff [ref.78] differentiate between shells of austenite noting the so-called "Halo's", that are formed during the solidification around primaire graphite, and "Shells" of ferrite, which form after solidification by carbon diffusion in the solid state. Because of the special way graphite grows particles are enveloped by austenite during the solidification process. Fig.28 shows this formation of austenite and martensite shells.
Figure 28 Formation of austenite and martensite shells acc. to Lux et al [78].
Engler and Ellerbrock [70] used displacement tests to study the morphology of the solidification front concluding that no nodules with accopanying austenite shells could have existed in de melt and that nodules are encapsulated by austenite dendrites in an early stage as shown in fig.29.
Figure 29 Stages in the encapsulation proces of nodules acc. to Engler and Ellerbrok [70].
Jiyang,Schmitz and Engler [79] examined the solidification behaviour of hypoeutectic, eutectic and hypereutectic material and discovered no differences. They point out that channels are found in the shells, meaning that austenite starts to grow at several sites around a nodule. As impurities were found in the channels, the authors proposed that graphite shape depends upon the velocity of shell formation. They differentiate between the following austenite shells: quickly formed, slowly formed and open shells (fig.30).
Figure 30 Relationship between the type of austenite and the shape of the graphite spheroids acc. to Jiyang et al [79].
Gan en Loper [80] emphasise that nodules are often found to countain a considerable amount of inherent austenite and by making comparisons with the actual matrix structure, conclude that nodules do not possess an external shell.
Itofuji et al. [58] examine vermicular graphite inclusions discovering "liquid channels" in the austenite shells.Fig.31 shows the influence of liquid channels in the austenite shell on graphite shape.
Figure 31 Influence of liquid channels in austenite shell on graphite shape acc. to Itofuji [82].
Chen, Wu, Liu and Loper [81] re-examen the above mentioned channels and conclude that they are responsible for the growth of vermicular graphite and note that liquid channels are formed by segregation of certain trace elements (fig.32).
Figure 32 Formation of vermicular graphite acc. to Chen et al [81]
Itofuji and Uchikawa [82] suppose that liquid channels in the surrounding austenite influence the formation of the different graphite shapes, the starting point for their hypothesis being the well-known Bubble theory, which is represented in fig.33. Figure 33. Formation of vermicular, nodular and chunky graphite acc. to Itofuji [82].
Boeri and Weinberg [83] propose that both nodules and austenite dendrites develop independently in the liquid and that the nodules are then encapsulated by the growing austenite dendrites.
In a recent publication [ref.83A], Minkoff states : "In the spheroidal case, it is now generally accepted that growth of a graphite spheroid by diffusion of carbon through an envelope of austenite is wrong."
11.0 CARBON SEPARATION.
11.1 Crystallinity.
In virtually every paper on the subject of cast iron, the separated carbon is given the name graphite. The typical crystal structure of graphite has given rise to many theories as to its morphology. However, as graphite is always examined at room temperature, is it accurate to deduce that it is present at the precise moment that carbon was separated from the melt? In the literature studied so far, the formation of graphite as a direct result of carbon deposits, has ever been questioned. Many researchers however, have found non-graphitic carbon within nodular and flake graphite [7,47]. Purdy and Audier [47] examined graphite nodules and flakes finding a considerable amount of amorphous carbon in both shapes on the graphite-iron borderline. The transition from amorphous carbon to graphite takes place later [quot.22] An abundance of literature exists on the conversion of non-graphitic carbon into graphite. The mechanism being strongly influenced by pressure, temperature and catalytic action of metals. In 1904, Moissan [84] succeeded in producing carbon inclusions possessing a diamond structure. Carbonated iron was superheated to 3000 degrees and quenched in water.Nowadays, these findings are highly questioned [84a], but recently Zhukov [84b] reports on new diamond-like allotropic forms of carbon, also found in cast iron. It can therefore be noted that all known carbon structures, ranging from amorphous to diamond, can be found in cast iron depending on the applied cooling conditions.
11.2 The growth mode.
The bulk of existing theories concerning formation and growth of nodules assume a growth direction starting at the center. Such a view fitted best with these theories or could even have formed their basic starting-point. An early representation of this view can be seen in fig.34.
The radial structure sometimes found in well-polished specimens, has often been regarded as an indication or even "proof" of a radial growth mechanism of a nodule. However, Sadocha [84c] proved that this radial structure was only a surface phenomenon caused by the sample preparation.
Some researchers make use of a different approach. Karsay [20,21] and Stadelmeyer [40], advocating the Bubble theorie, suppose that growth of a nodule starts at the perifery and proceeds towards the centre as shown in fig.35. Figure 35 Growth of a nodule from the periphery to the center acc. to Stadelmaier [40]. A third mechanism was proposed by Wlodawer [page 393/394], in which was put forward that the position and shape of the graphite inclusions were induced by the surrounding austenite crystals. In a recent publication, Itofuji and Uchikawa proposed their so called "Site-theorie", [82], [quot.23]. In this theory the authors state that the shape of the separated graphite is totally dependent upon the available space in which the separation takes place.
With this in mind, it is interesting to compare some microstructures in the light of this theory:
Figure 36 [85] shows a dendritic graphite pattern (malleable iron), that was formed after een high temperature heat-treatment. The original porousities have been filled up by carbon, separated during the heattreatment [Site-theory].
Figure 37 [ref.41] shows a degenerated nodule. According to the original comment, the visual branches grow out of the nodule itself [Melt-theory].
The shape of the branches in fig.37 and the graphite structures in fig.36 are strikingly resemblant.
Still other views on the growth mode of nodules exists.According to Hurum [ref.84a], a nodule is built up by coagulation of colloidal graphite particles.
Epanchintsev [ref.85b] found that every individual nodule or flake is built up from numerous tine flakelets.
12.0 CONCLUSION.
The literature review shows that the still popular melt theory has strongly influenced research on the solidification of nodular iron. However, it would seem, that placing such faith in this theory has not been so wise. By accepting the melt theory, almost without any criticism, dendrites, although so important in the early days, were dismissed. It took almost 25 years before the basic role of dendrites was again recognized. Much energy has been put into the search for nuclei and crystal growth mechanisms. As a result the numerous and divergent theories made a unified point of view impossible. Added to this, is the fact that the proof of the melt theory is at best weak. Its basis relies upon overwhelming contradictions with regard to the austenite shells, the solid state diffusion and the growth mode of graphite and the fact that carbon is directly separated as graphite has never been proven. There remains only one serious question to put forward: Could it be possible that we have overlooked a fundamentally different, probably much simpler, mechanism that causes the typical carbon shapes in nodular cast iron?
QUOTATIONS.
Piwowarsky Hochwertiges Gusseisen. Zweiter Neudruck. Springer-Verlag, Berlin, 1961. [quot.1] page 209: "Es ist wahrscheinlich, daß bei den Schmelzüberhitzungsversuchen von E.Piwowarski(1) die dort durch hohe überhitzung angestrebten und erzielten ,,temperkohleartigen'' Ausbildungsformen des Graphits (Abb. 156 ganz rechts) bereits mehr oder weniger Spärolithenstruktur besaßen (40,43). Aber bei den schwachen Vergrößerungen, welche um jene Zeit (1920-1930) allgemein für die Beobachtungen der Graphitausbildung verwendung fanden (meistens 50 bis 100 li. Vergr.), war diese Form der Ausbildung des elementaren Kohlenstoffs den meisten Beobachtern entgangen."
A review of the formation of spheroidal graphite in cast iron. S.Banerjee. The British Foundryman September 1965. [quot.2] page 341: "The literature has been critically examined by listing about 150 experimentally observed facts which are pertinent to the formation of spheroidal graphite."
[quot.3] page 341: "The discussion seems to indicate that the nucleation of a spheroid takes place necessarily in the liquid state even in a Mg-treated white cast iron, and its growth takes place by the diffusion of carbon atoms from the source (which could be the melt,austenite or decomposition of cementite) to the spheroid nucleus.This idea is consistent with all the listed observations."
B.Lux. Zur theorie der Bildung von Kugelgraphit im Gusseisen. Giesserei-Forschung 1970, Heft 2. [quot.4] page 71: "Seit den grundlegenden Arbeiten von E.Scheil und Mitarbeitern 15)58)74)60)besteht aber heute kaum ein Zweifel mehr darüber, dass im Normalfall die Sphärolithen direct aus der Schmelze entstehen. Eindeutig bewiesen wird dies unter anderem durch die Möglichkeit, sie durch Zentrifugieren zu trennen 58), und durch das häufich beobachtete Aufschwimmen der Kugeln bei grossen Gussstücken 23)51)79). Ausserdem gibt es verhältnismassig häufig (rnd 4%) zusammengewachsene Kugeln,wie Bild 6 zeigt58)23)."
On the solidification kinetics of Spheroidal Graphite Cast Iron. D.M.Stefanescu,D.K.Bandyopadhyay. Conference Proceedings Cast Iron IV, 1990, Materials Research Society. [quot.5] page 15: "At the present time, the most widely accepted mechanism for the solidification kinetics of spheroidal graphite (SG) iron is the diffusion controlled growth of graphite through the enveloping shell. Recently, it has been shown that austenite dendrites play a significant role in eutectic solidification of SG iron, and that it is possible for austenite dendrites to grow partially independent of graphite spheroids."
[quot.6] page 18: "It has been widely accepted that growth of this eutectic starts with nucleation and growth of graphite in the liquid followed by early incapsulation of these graphite spheroids in austenite shells(enveloppes)."
Morrogh and Williams. Graphite Formation in Cast Irons. Journal of Iron and Steel Institute, 1947. [quot.7] page 367: "No evidence has been obtained in any of the alloys studied that spherulitic graphite structures form directly from liquid.It appears that a solid carbide must decompose to yield a spherulitic structure, and this confirms experimence in the malleablizing process."
A.L.De SY.
Graphite Spherulite Formation and Growth. Foundry, November 1953. [quot.8] page 103: "In a shrinkage cavity of a definitely hypereutectic nodular iron,we observed well formed macroscopic dendrites(fig2). The composition of a small dendrite itself was checked and found almost identical to that of the casting. Does a eutectic form dendrites? We don't believe it does;sometimes small dendrites of one of the phases but never dendrites of the eutectic itself.Therefore, and for many other reasons, it is believed that nodular iron solidifies without eutectic but in a mixed cristals system, forming supersaturated austenite in which graphite precipitation occurs, beginning at a certain degree of super-saturation, and spherolutic graphite growth follows by a peritectic transformation or reaction."
Graphitbildung in Eisen-Kohlenstoff-Gußlegierungen. A.Wittmoser Giesserei, April 1959. [quot.9] page 183: "Demgegenüber ist bei Ausscheidung des Graphits in Kugelform seine Lage im Primärgefüge unverandert und durch die überwiegend zentrale Position gekennzeichnet(vgl Bilder 5e bis h)... Diese ausschliesslich auf den Ergebnissen zahlreicher Versuchsreihen beruhenden Argumente sind nach Auffassung des Verfassers so überzeugend, dass über die Möglichkeit der Entstehung des Kugelgraphits in grauerstarrenden Legierungen aus dem an Kohlenstoff übersattigten Mischkristall (d.h. im festen Zustand) keine Zweifel bestehen dürften."
[quot.10] page 181: "Schon der Gefügeaufbau des Gußeisens mit Kugelgraphit unterscheidet sich ganz grundsätzlich von dem der Gußeisenlegierungen mit Lamellengraphit. Besonders die Seigerungsätzung zeigt deutlich, daß bei Gußeisen mit Kugelgraphit eine Kristallisation vorliegen muß, die im wesentlichen der Erstarrungsweise von Mischkristallegierungen entspricht(vgl. Bilder 4 und 5). In guter Übereinstimmung mit dieser Befund, lassen sich auch bei Vorliegen einer kugelformigen Graphitausbildung über den gesammten im Eisen-Kohlenstoff-Schaubild dargestellten Konzentrationsbereich keine im bisherigen Sinne "Untereutektische" und "übereutektische" Legierungen unterscheiden."
Untersuchungen über die Kristallisation des Gußeisens mit Kugelgraphit. Archiv für das Eisenhüttenwesen Mai/Juni 1953. E.Scheil , L.Hütter. [quot.11] page 244: "Auf Grund der eigenen Untersuchungen ist es am wahrscheinlichsten, dass der Kugelgraphit in der Regel auf folgende Weise entsteht. Zunächst bilden sich Graphitkeime in der Schmelze unabhängig von gegebenenfalls schon vorhandenen Dendriten aus Y-Mischkristall. Nachdem diese eine gewisse, noch nicht bekannte Große erreicht haben, entsteht um sie ein Hof aus Y-Mischkristall. Die weitere Vergrößerung der Graphitkugeln erfolgt durch Diffusion durch die Hülle aus Y-Mischkristall, die sich ihrerseits mit dem Fortschreiten der Kristallisation verdickt."
Über den Einfluss von Fremdkeimen auf die Kristallisation von Metallen und Legierungen, insbesondere auf die Ausbildung des Eutectikums in Gußeisen. W.Patterson. Giesserei Techn-Wissensch.Beihefte März 1952. [quot.12] page 377: "Die Annahme der bevorzugten Entstehung in einer Bucht wird durch die konzentrische Abdrängung des Kohlenstoffs durch die Grenzflächen der Bucht und die damit bewirkte besonders starke örtliche Übersättigung gerechtfertigt."
Beitrag zur Entstehung des kugeligen Graphits im Gusseisen. Giesserei Juni 1954, s.313/320. B.Marincek. [quot.13] "Die Graphitausscheidung eines mit Magnesium genügend behandelten Gusseisens mit Kugelgraphit erfolgt derart, dass sich die Graphitkeime wegen der hohen Oberflachenspannung der Schmelze bei niedriger Temperatur(Unterkuhlung) und in der Kugelform ausscheiden. Gleichzeitig werden die Graphitkeime vom eutectischen Austenit vollständig umgeben und von der Schmelze getrennt. Das Weiterwachsen des Graphits in der Kugelform erfolgt derart, dass der Kohlenstoff durch die graphitumgebende Austenitschicht herandiffundiert."
Graphitbildung in Eisen-Kohlenstoff-Gusslegierungen. Giesserei April 1959. A.Wittmoser. [quot.14] page 186: "1.Die Auffassung, dass submikroskopisch kleine Gasblasen mit keimbildendem Character für den Graphit vorliegen können, wird nur schwer zu beweisen und wohl überhaupt nicht zu widerlegen sein."
Über den Entstehungsort der Graphitkugeln im erstarrenden Gusseisen. Giesserei,Techn.-Wissensch.Beihefte, Oktober 1961 . E.Scheil, J.D. Schöbel. [quot.15] page 205: "Die in der Schmelze nach Annahme 2 entstandene Graphitkugel wird nach einer gewissen Zeit von einem y-Hof umhüllt. Der Y-Hof wandelt sich je nach den Abkühlungsbedingungen weiter in Perlit oder in Martensit um.Er soll unabhängig vom Endgefüge als Y-Hof bezeichnet werden.Ausserdem soll in diesem Abschnitt zur Vereinfachung auch dann von einem Y-Hof gesprochen werden wenn es eigenlich ein primärer Y-Mischkristall ist, in dem nach Annahme 3 eine Graphitkugel entstanden ist.Der Ausdruck 'HOF' soll also nichts über die Entstehungsart der Graphitkugel aussagen."
[quot.16] page 211: "Die Frage wo die Dendriten bleiben ist von W.Patterson sowie von E.Scheil und L.Hutter dahin beantwörtet worden,dass durch die Graphitkugeln die Dendriten so verformt werden, dass sie als solche nicht mehr erkannt werden."
A Revision of the Fe-C-Si-System. AFS Cast Metals Research Journal, March 1968. K.R.Olen, R.W.Heine. [quot.17] page 28: "A number of anomalies in the as cast microstructures of eutectic and hypereutectic commercial cast irons suggested the need to re-examine the phase diagram and phenomenology of solidification of Fe-C-Si alloys.Cooling curves for hypereutectic ductile irons frequently indicate a pronounced primary arrest preceeding the eutectic arrest,Fig.1a. When this occurs, examination of the microstructure reveals the presence of several prior dendritic areas,Fig. 1b."
Gelenkte Erstarrung von Gusseisen. Robert Wlodawer. Giesserei-Verlag G.M.B.H.,Düsseldorf, 1977. [quot.18] page 373: "Schon in üblichen ebenen Schliffbildern ist oft ersichtlich, dass Kugelgraphit bzw. die umgebenden Austenitschalen sich zu klumpigen, unregelmässigen Gebilden zusammenballen (31). Bei entsprechender Schulung des Auges und vor allem bei Betrachtung hinreichend grosser Schliff-Flächen fällt auf, dass diese Zusammenballungen regelmässig strukturiert sind: Wir sehen die Anordnung der Sphärolithen etwa in Form von Geraden, von Winkeln, von kreisähnlichen Figuren (Fig.9.135). Gerade und winkelförmige Anordnungen werden zweifellos durch Dendriten bedingt."
Precipitation of Graphite during the Solidification of Nodular Cast Iron. Seminar-Proceedings Detroit 1964. J.D.Schöbel. [quot.19] page 305: "This paper, which covers experimental work of E.Scheil and the autor (14,16) is based on the following two assumptions. 1.Nodular graphite precipitates from the melt at the beginning of the eutectic solidification. It will be surrounded by eutectic austenite thereafter and grow by diffusion of carbon through the solidified austenite. By measuring the radii of the globular austenite envelopes and the associated graphite nodules in nodular cast iron samples, whose solidification is interrupted by quenching within the range of the eutectic solidification, it is possible to predict that a graphite nodule precipitates from the melt. Such calculations were performed first by E.Scheil and l.Hütter(13)."
[quot.20] page 306: "2.Graphite nodules which precipitate directly from the melt will show a tendency to float to the surface of the melt because of the low density of carbon compared with that of liquid iron. This fact is well known for hypereutectic nodular cast iron(17). In the case of hypoeutectic melts, there will be at the beginning of the eutectic solidification a relative flotation of primary austenite crystals and graphite nodules: the austenite crystals will show a tendency to sink towards the bottom of the melt while the graphite nodules will try to float towards the surface. This relative flotation of primary austenite and graphite nodules will hardly be observable by working under normal conditions, i.e. with normal gravitational forces. Increasing the gravitational forces by centrifuging nodular cast iron melts during their solidification will increase the effect of flotation."
Untersuchungen über die Kristallisation des Gusseisens mit Kugelgraphit. Archiv für das Eisenhüttenwesen, Mai/Juni 1953. E.Scheil und L.Hütter.
[quot.21] page 243: "Nimmt man an,dass die Keimbildung in der Schmelze erfolgt, dass aber, entweder gleichzeitig oder bald darauf, die kleine Graphitkugel von einer Hülle aus Y-Mischkristall überzogen wird, so liegt der gleiche Mechanismus vor.Bei dieser Vorstellung besteht aber der Vorteil, dass die Möglichkeit der Bildung von Graphitkugeln in der Schmelze durch ihr Aufsteigen im Schwerefeld oder unter dem Einfluss von Zentrifugalkraften schon bewiesen ist."
Electron Microscopical Observations of Graphite in Cast Irons. G.R.Purdy, M.Audier. In:The Physical Metallurgy of Cast Iron. Proceedings of the Third International Symposium on the Physical Metallurgy of Cast Iron,Stockholm, Sweden, August 29-31, 1984. Editors; H.Fredriksson, M.Hillert. Uitgave: North-Holland, New York-Amsterdam, Oxford. [quot.22]: "The evidence of Figure 6 is strongly indicative of a lateral growth process. In addition, the observations of amorphous carbon at apparent 'crystallization sites ' on the surface of the spheroids are suggestive of a growth mechanism for the spheroids: In this mechanism, carbon is transported through the solid iron phase, and is deposited at the iron-carbon interface as amorphous carbon. The crystallization of the amorphous carbon to form graphite would then take place mainly at steps, or terminations of aromatic layers."
Formation Mechanism of Chunky Graphite in Heavy-Section Ductile Cast Irons. AFS Transactions , 1990. H.Itofuji en H. Uchikawa. [quot.23] page 447: "In new theory, it was defined that the graphite morphology might be dependent on the site where graphite precipitated. That is to say, graphite does not take morphology by itself but graphite is given the morphology by the site and graphite grows along the site."
REFERENCES
[1] G.Sachs, Praktische Metallkunde, Erster Teil, Berlin, Verlag von Julius Springer, 1933.
[2] Eugen Piwowarsky, Hochwertiges Gusseisen. Zweiter Neudruck. Springer-Verlag, Berlin, 1961.
[3] A.Wittmoser, Zur Systematik der Werkstoffgruppe Gusseisen. Giesserei, 1954, page 685/692.
[4] A.Wittmoser, Gedanken zur Einordnung der Eisen-Kohlenstoff-Gusslegierungen. Giesserei, 1955, page 573/580.
[5] K.D.Millis, Spheroidal graphite cast iron-its development and future. The British Foundryman, 1972, page 6/14.
[6] K.Keßler, H.Reuter. Derzeitiger Stand und Entwicklungstendenzen von Gusseisen mit Kugelgraphit. Giesserei, 1967, page 169/176.
[6a] K.D.Millis, The invention of Ductile Iron... In Millis' Own Words.Modern Castings, october 1998, page 41/44.
[7] S.Banerjee, A review of the formation of spheroidal graphite in cast iron. The British Foundryman September 1965, page 344/353.
[8] B.Lux, Zur theorie der Bildung von Kugelgraphit im Gusseisen. Giesserei-Forschung 1970, Heft 2, page 65/80.
[9] D.M.Stefanescu, D.K.Bandyopadhyay. On the solidification kinetics of spheroidal Graphite cast iron. Conference Proceedings Cast Iron IV, 1990, Materials Research Society, page 15/26.
[10] Morrogh and Williams, Graphite Formation in Cast Irons. Journal of Iron and Steel Institute 1947, page 322/370.
[11] R.P.Dunphy, W.S.Pellini, Nodule genesis and growth in magnesium-treated hypoeutectic irons. Foundry 80, 1952 page 82/86, 93, 196, 198/200.
[12] A.L.De Sy, Graphite Spherulite Formation and Growth. Foundry, November 1953, page 100
[13] Factoren die de vorming van bolvormige grafiet in gietijzer beinvloeden. Voordracht gehouden door Prof.dr.A.J.Zuithoff op 10-12-1963 te Amsterdam voor de bijeenkomst van de Kring Metalen van de Bon voor Materialenkennis en de Ned. Ver. van Gieterijtechnici.
[14] A.Wittmoser, Graphitbildung in Eisen-Kohlenstoff-Gusslegierungen. Giesserei April 1959, page 177/187.
[15] A.Wittmoser, A propos de la formation des nodules de graphite. Intern. Gieterij Congres, Bruxelles 1951, page 433/449.
[16] A.Wittmoser, Untersuchung der Erstarrungsvorgänge in Eisen-Kohlenstoff-Gusslegierungen mit Hilfe von kalorischen Messungen. Giesserei Techn.Wissensch. Beihefte,Oktober 1961, page 195/201.
[17] E.Scheil und L. Hutter, Untersuchungen über die Kristallisation des Gusseisens mit Kugelgraphit. Archiv fur das Eisenhuttenwesen, Mai/Juni 1953 , page 237/246.
[18] W.Patterson, Über den Einfluß von Fremdkeimen auf die Kristallisation von Metallen und Legierungen, insbesondere auf die Ausbildung des Eutectikums in Gußeisen. Giesserei Techn.-Wissensch. Beihefte, März 1952. page 355/378.
[19] P.E.A. van Nieuwland, Een theorie ter verklaring van het ontstaan van grafietnodulen in gietijzer. Metalen, Jaargang VIII, nr.18, page 333/335.
[20] S.I.Karsay, Ductile Iron, The state of the art. Uitgave QIT, 1976.
[21] S.I.Karsay, Ductile Iron, The state of the art. Uitgave QIT, 1980.
[22] B.Marincek, Beitrag zur Entstehung des kugeligen Graphits im Gusseisen. Giesserei Juni 1954, page 313/320.
[23] M.Hillert and Y. Lindblom, The Growth of Nodular Graphite.Journal of the Iron and Steel Institute. April 1954, page 388/390.
[24] H.Morrogh, The Solidification of Nodular Iron. Journal of the Iron and Steel Institute. April 1954, page 378/382.
[25] H.Morrogh, Industri Information nr.11 Uitgave Infinitas, Stockholm, 1961, page 35/72.
[26] E.Scheil und A.Wittmoser, Kritische Betrachtungen über die Entstehung des Kugelgraphits in Gusseisen. Z.Metallkde. 1958, Heft 5, page 256/267.
[27] J.Motz, Über die Kugelgraphitbldung in übereutectischen Gusseisenschmelzen. Giesserei, Dezember 1959, page 953/957.
[28] E.Scheil, Über die eutectische Kristallisation. Giesserei Techn.Wissensch.Beihefte April 1959, page 1331/1338.
[29] E.Scheil , J.D. Schobel, Über den Entstehungsort der Graphitkugeln im erstarrenden Gusseisen. Giesserei Techn. Wissensch. Beihefte . Oktober 1961, page 203/213.
[30] W.Patterson, Zur Theorie der Kugelgraphitkristallisation. Giesserei Techn. Wissensch. Beihefte . April 1961, Heft 2, page 65/70.
[31] C.R.Loper ,Jr and R.W.Heine, Graphite formation during solidification of cast iron. Trans. of the A.F.S., 1962, page 583/600.
[32] C.R.Loper,Jr., Ductile Iron solidification study using electron microscope. Trans. of the A.F.S., 1962, page 963/967.
[33] W.Oldfield,J.G.Humphreys, Formation of nodular graphite in hypo-eutectic irons. BCIRA Journal 1962, page 315/324
[34] H.Kempers, Sphärolitenflotation bei Gusseiesen mit Kugelgraphit. Giesserei, Dezember 1966, page 841/846.
[35] J.D.Schöbel, Precipitation of Graphite during the solidification of Nodular Cast Iron. Seminar-Proceedings Detroit 1964.
[36] F.Henke, Eutektische Erstarrung von Eisen-Kohlenstoff-Legierungen. Giesserei-Praxis, November 1967, page 391/407.
[37] O.Tombrock, A new technique for studying the solidification proces of cast iron. Meehanite Paper no. E.1002, 1961.
[38] G.Jolley,S.R.Holdsworth, Solidification of nodular iron containing boron. The British Foundryman 1971, page 230/235.
[39] B.Marincek, Grundlagen und Verfahren der Qualitatsprufung von Gusseisenschmelzen. Teil 1.Grundlagen. Giesserei 1980, nr.19, page 587/593.
[40] H.Stadelmaier, Über Sphärolithenbildung in Metallschmelzen. Z.Metallkunde 1960, Heft 10, page 601/604.
[41] I.Minkoff, Factors affecting growth of spheroidal graphite. Modern Castings Jan 1962, page 66/72.
[42] K.Herfurth, Einige Probleme des Kristallwachstums und ihre Bedeutung für die Graphitkristallisation in Gusseisen. Giessereitechnik 1964, Heft 10, page 298/302.
[43] L.Kalvelage, Das Wachstum von Graphit in übereutectischen Eisen-Kohlenstoff-Schmelzen. Giesserei-Forschung 1978, Heft 4, page 143/151.
[44] P.C.Liu, C.R.Loper,Jr., T.Kimura, H.K.Park, Observations on the Graphite Morphology In Cast Iron.
AFS Transactions 1980, page 97/118.
[45] Chen Sy-Sen, Yi Sun-Sheng, Wang Zu-Lun, Chen Sy-Chen, Jia Shun-Lian. The mechanism of spheroidal graphite formation during primary crystallization of cast iron. 48 Intern. Gieterij Congres, Varna, 1981, nr.8CN.
[46,47,48] The Physical Metallurgy of Cast Iron. Proceedings of the Third International Symposium on the Physical Metallurgy of Cast Iron,Stockholm, Sweden,August 29-31,1984. Editors: H.Fredriksson,M.Hillert. Uitgave:North-Holland,New York-Amsterdam,Oxford.
[46]:Zhu Peiyne, Sha Rozeng, Li Yanxiang, Effect of Twin/Tilt on the growth of Graphite . [ref.47]:G.R.Purdy, M.Audier, Electron Microscopical Observations of Graphite in Cast Irons. [ref.48]:I.Minkoff, The spherolitic growth of graphite.
[49] A.L.De Sy, The Core of Graphite Spherules in Nodular Cast Iron. Metal Progress, June 1950, page 774/775.
[50] H.Bakkerus,A.P.von Rosenstiel, Onderzoek naar de aanwezigheid van kernen in nodulair grafiet. Metalen, nr.2 1964, page 40/44.
[51] 2. Internationale Symposium over de Metallurgie van Gietijzer, Geneve, 1974. Summary by W.Weis in: Giesserei 1974, nr.15, page 475.
[52] M.J.Lalich,J.R.Hitchings, Characterization of Inclusions as Nuclei for Spheroidal Graphite in Ductile Cast Irons. AFS Transactions 1976, page 653/664.
[53] J.F.Wallace,R.D.Maier, How Inoculation Works. Proceedings of AFS-CMI Conference February 6-7,1979 ,Illinois.
[54] Y.Ueda, M.Takita, K.Ishikawa, T.Yamamoto, A study on the Nucleation of Graphite in Cast Iron. Translation of :Imono, Vol 52 (1980), No.9.
[55] I.Tartera, Cast Iron Inoculation Mechanisms. AFS International Cast Metals Journal, dec.1980, page 7/14.
[56] N.Kayama,K.Suzuki, Effects of sulfur on the nucleation and growth of graphite out of residual graphite particles in molten cast iron. Report of the castings research laboratory, Waseda University. No 32, 1981.
[57] CH.Wang,H.Frederiksson, On the mechanism of inoculation of cast iron melts. 48 Intern. Foundry Congres, Varna 1981, nr. 25S.
[58] H.Itofuji,Y.Kawano,N.Inoyama,S.Yamamoto,B.Chang,T.Nishi. The formation mechanism of compacted/vermicular graphite in cast irons AFS Transactions 83-106, 1983, page 831/840.
[59] S.A.Azzam, L.Ya.Kozlov, A.P. Vorobev, A. Fahmy Abdel-Salam. Role of impurities and rare earth metals in the formation of the structure of ductile iron. 60th World Foundry Congress 1993, The Hague, paper nr. 41.
[60] H.Nakae, K.Okauti, K.Takai, H.Koizumi, T.Oka, Role of calcium for inoculation of cast iron. 60th World Foundry Congress 1993, The Hague, paper nr. 24.
[61] R.El Haik, Fonderie nr.197, 1962, page 243/246.
[62] C.R.Loper,Jr., Dendritic Structure and Spiking in Ductile Iron. AFS transactions, 1968, page 547/554.
[63] T.W.Parks,C.R.Loper,Jr., A study of the conditions promoting dendritic growth in ductile iron. AFS transactions 1969, page 90/96.
[64] K.R.Olen,R.W.Heine, A Revision of the Fe-C-Si System. AFS Cast Metals Research Journal, March 1968, page 28/43.
[65] J.Verlinden,L.J.Kikkert, Dendritisches Gefüge in Gusseisen mit Kugelgraphit.Eine Zwischenbilanz. Meehanite Research report nr. 1375/G. november 1982.
[66] M.Gagne,R.Goller, Plate fracture in ductile iron castings. AFS Transactions 83-84, 1983, page 37/46.
[67] R.Barton, Graphite nodule alignment and plate fracture in nodular iron. BCIRA Report nr. 1640, november 1985.
[68] M.H.Mulazimoglu, Y.M.Yang, J.F.Wallace, N.X.Ding, J.L.Mercer, Solidification Studies of Spiking and Large-Small Nodule Formation in Ductile Cast Iron Produced by the In-the-Mold Process. AFS Transactions, 1985, page 627/642.
[69] Robert Wlodawer, Gelenkte Erstarrung von Gusseisen. Giesserei-Verlag G.M.B.H., Dusseldorf, 1977.
[70] S.Engler, R.Ellerbrok, Über die Bildung von Austenithüllen um Graphitkugeln während der Erstarrung von Eisen-Kohlenstoff-Silicium-Schmelzen. Giesserei-Forschung 1977, Heft 4, page 141/145.
[71] H.Nieswaag, Nodulair Gietijzer,Kenmerken en Mogelijkheden. NVvGT/KIVI leergang, Ontwerpen en gieten van hoogwaardige constructiedelen, oktober 1983, Delft, page 253/285.
[72] A.Rickert, S.Engler, Solidification Morphology of Cast Irons, The Physical Metallurgy of Cast Iron. Proceedings of the Third International Symposium on the Physical Metallurgy of Cast Iron,Stockholm, Sweden, August 29-31, 1984. Editors: H.Fredriksson, M.Hillert. Uitgave:North-Holland, New York-Amsterdam, Oxford.
[73] J.Zhou, W.Schmitz, S.Engler, Untersuchung der Gefügebildung von Gusseisen mit Kugelgraphit bei langsamer Erstarrung. Giesserei-Forschung, 1987, Heft 2, page 55/70
[74] J.M.Motz, D.B.Wolters, Über Erstarrungsgefüge und Eigenschaften in dickwandigen Gußstücken aus ferritischem Gußeisen mit Kugelgraphit. Teil 1: Mikroseigerungen. Giessereiforschung 40 (1988) nr.2, page 69/78.
[75] W.A.Tiller, Fe-C and Fe-C-Si Alloys. Seminar-Proceedings June 16-18 1964, Detroit.
[76] S.I.Karsay, E.Campomanes, Control of graphite structure in heavy ductile iron castings. AFS Transactions Vol 78, 1970, page. 85/92.
[77] G.S.Cole, Solidification of Ductile Iron. AFS Transactions, 1972, page 335/348.
[78] Lux, Mollard, Minkoff, On the formation of envelopes around graphite in cast iron. Proceedings of the second Int. Symposium on the metallurgy of cast iron. Geneve 1974.
[79] Z.Jiyang, W.Schmidz, S.Engler, Formation of Austenite Shell Around Nodular Graphite and its Effect on Deteriation of Graphite. AFS Transactions, 1990, page 783/786.
[80] Y.Gan, C.R.Loper,Jr., Observations on the formation of graphite in compacted and spheroidal graphite cast irons. AFS Transactions 83-184, 1983, page 781/788.
[81] J.Y.Chen, D.H.Wu, P.C.Liu, C.R.Loper,Jr., Liquid Metal Channel Formation in Compacted/Vermicular Graphite Cast Iron Solidification. AFS Transactions 1986, page 537/544.
[82] H.Itofuji, H.Uchikawa, Formation Mechanism of Chunky Graphite in Heavy-Section Ductile Cast Irons. AFS Transactions, 1990, page 429/448.
[83] R.Boeri, F.Weinberg, Microsegregation in Ductile Iron. AFS Transactions 1989, page 179/184.
[83A] I.Minkoff, How do cast irons solidify? A revieuw of micromodelling using the computer, supporting experiments and basic theory. The Foundryman Jan/febr. 1993, page 16/22.
[84] R.Meijer, P.Engelsman, Vier eeuwen Diamant. Uitgeverij Uniepers Amsterdam, 1986.
[84a] H.A.M.Snelders, Hoofdstukken uit de geschiedenis van de schikunde. Amsterdam, Rudopi 1979.
[84b] A.A.Zhukov,P.Ramachandra Rao, New findings in carbon chemistry and their relation to cast irons. Indian Foundry Journal, June 1994,page 13/18.
[84c] J.E.Guzleski, On the growth of spherulitic graphite in nodular cast iron. Carbon 1975, Vol13, page 167/73.
[85] Typical Microstructures of Cast Iron, The British Cast Iron Research Association, Alvechurch,Birmingham, uitgave 1951, page 49.
[85a] F.Hurum, A study of the formation of nodular graphite. Modern Cast 1952, page 834/48.
[85b] O.G.Epanchintsev,M.P.Shebatinov, Graphite structures as observed with a scanning microscope. Russian castings production 1971, page 386/89.
Figure 2 shows the structure of cast nickel with 0,8% sulfur content.
Figure 3 shows the changed structure after magnesium addition.
Figure 4, nodule formation acc.to Dunphy.
Figure 5, nodule formation within (super)saturated austenite.
Figure 6, nodule formation within (super)saturated austenite acc. to Wittmoser.
Figure 7 nodule formation acc. to Scheil.
Figure 8 preferred nucleation sites for graphite nodules between austenite dendrite arms (Patterson).
Figure 9 Gasbubbles as a template for the final graphite shape. 
Figure 10 The typical ball-shape is caused by high surface tension acc. to Marincek.
Fig.11. Successive stages in the solidification of nodular cast iron. Primary dendrites are left out. Morrogh,1961 [25].
Fig.12. The solidification of nodular iron (schematic). Henke,1967,[36].
Fig.13. Various stages of the eutectic in nodular cast iron. Jolley,Holdsworth,1971,[38].
Fig.14. The solidification of the graphite-austenite eutectic in nodular cast iron. Marincek,1980, [39].
Fig.15 shows the crystalline structure of graphite giving rise to many growth mechanisms.
Figure 17 Dendrite formation during the solidification of nodular cast iron acc. to Verlinden and Kikkert [65].
Fig.18 Conglomeration of nodules acc. to Wlodawer [69].
Fig.19 Eutectic solidification of nodular cast iron acc. to Engler and Ellerbrock [70].
Fig.20 Solidification of nodular cast iron acc. to Nieswaag [71]: 1=liquid;2=primary austenite dendrites;3=nodule formed in the liquid;4=austenite-shell around the ndule.
Fig.21 Solidification morphology of eutectic nodular cast iron acc. to Rickert and Engler [72].
Fig.22 Nodule formation in hypereutectic nodular cast iron acc. to Zhou et al [73]:1=flotation of primary graphite nodules;2=austenite dendrites;3=nodule formation;4=carbon-rich region;5=direction of heat flow;6=interdendritic region;7=growth direction.
Fig.23 Schematic of the eutectic solidification in nodular cast iron acc. to Motz and Wolters [74]. (a)Primary crystallisation, occurrence of convectional flows in the melt;(b)Crystallization of the other eutectic phase, particularly in supersaturated regions, gravitational segregation;(c)Progression of the eutectic solidification, gravitational segregations;(d)formation of solidified, completed and related residual melt regions.
Fig.24 Artist impression of variations of the solidification front with various graphite shapes.
Figure 25 Quenched nodular cast iron. The formation of graphite nodules just started. Primary austenite dendrites can be seen, together with embryo nodules,each surrounded by an envelope of austenite. In the interdendritic positions, there is white iron arising from the rapid quenching of the untransformed liquid. (Morrogh [25]).
Fig.26.Mechanism of shell formation around a nodule:a=growth of a nodule in contact with the melt;b=encapsulation by austenite;c=growth of a nodule within the shell by solid-state carbon diffusion
Figure 29 Stages in the encapsulation proces of nodules acc. to Engler and Ellerbrok [70].
Figure 33. Formation of vermicular, nodular and chunky graphite acc. to Itofuji [82].
Figure 35 Growth of a nodule from the periphery to the center acc. to Stadelmaier [40]. A third mechanism was proposed by Wlodawer [page 393/394], in which was put forward that the position and shape of the graphite inclusions were induced by the surrounding austenite crystals. In a recent publication, Itofuji and Uchikawa proposed their so called "Site-theorie", [82], [quot.23]. In this theory the authors state that the shape of the separated graphite is totally dependent upon the available space in which the separation takes place. 
