MATERIALS TRANSACTIONS
Online ISSN : 1347-5320
Print ISSN : 1345-9678
ISSN-L : 1345-9678
Optimizing Control of Nodularization of Ductile Iron Based on Oxygen Activity Measurement
Zhenyu XuDayong LiXuliang MaDequan ShiChaowei HanLihua Wang
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2018 Volume 59 Issue 12 Pages 1928-1934

Details
Abstract

Dissolved oxygen in molten iron is one of the most important metallurgical factors affecting nodularization effect and casting properties. In present paper, the metallurgical behaviors of active oxygen in molten iron are discussed during nodularization through a ZrO2 (MgO partly stabilized) solid electrolyte battery for ultra-low oxygen condition of nodularized iron liquid. Combining with the results of thermal analysis obtained almost after active oxygen content measurement, the optimization control of nodularization process for ductile iron is studied. Results show that nodularization extremely disturbs the metallurgy equilibrium in molten iron, over-nodularization promotes this non-equilibrium degree. Inoculation can decrease the non-equilibrium trend and improve the solidification model by graphite nucleation. Nodularization effect degeneration began with the inoculation fading, with the active oxygen content slight increase, the volume fraction of nodular graphite decreased with rounded morphology. Followingly, as the active oxygen content continuous increase, the nodularity began to decrease. Therefore, it is promising to achieve more accurate control of nodularization by this combined on-line measurement technique.

1. Introduction

Ductile iron casting has been widely used in modern mechanical equipment because of its excellent mechanical properties, low production cost and convenient manufacturing. The variety of application and production conditions makes it an urgent problem to control the production process of ductile iron. Casting thermal analysis can effectively reflect the effects of nodularization on molten iron, but more detailed microscopic mechanism about solidification is only a kind of speculation lacking essential experiment verification.14) According to the modern metallurgy theory of ductile iron, oxygen content is a major factor to modify the graphite shape and matrix structure formation of iron castings in the process of solidification.59) At present, the dissolved oxygen content in molten iron can only be obtained directly by the solid electrolyte concentration cell technology.5,912) The obtained results found that the best nodularizing effect and the optimal mechanical properties of the casting met at the same oxygen activity value, vermicularized molten iron had a higher oxygen activity than that of ductile iron. Therefore, it is possible to distinguish between ductile iron and vermicular cast iron by using oxygen activity measurement on line.5,8) However, the study on the optimization of nodularization by oxygen activity measurement has not been reported.

In this paper, through a solid electrolyte concentration cell used in an ultra-low oxygen content range combined with the thermal analysis, the graphite nodularization and metallurgy optimization control for molten iron was discussed in detail. It can provide the basis for nodularization effect evaluation and dynamic regulation of the metallurgical state.

2. Experimental Procedures

2.1 Oxygen probe

Oxygen concentration cell is the key component of the oxygen probe. The solid electrolyte ceramic tube is made of high density ZrO2 stabilized with MgO. The raw material ratio is 90 mol% ZrO2, 5 mol% MgO, 0.3 mol% Y2O3, 0.3 mol% CaO, 0.3 mass% (25 mol% Al2O3, 40 mol% SiO2, 40 mol% TiO2, 10 mol% Fe2O3) agglomerating accelerator. The preparation process of solid electrolyte tube is described in detail in reference.13)

A certain Cr+Cr2O3 mixture is elaborately employed as the reference electrode in cell, of which the oxygen partial pressure is nearly close to that in molten iron investigated. A detailed description about the preparation process can refer the patent.14) The designed probe has high reliability, with the testing oxygen potential accuracy of ±1 mV and testing temperature accuracy of ±0.2°C. The oxygen activity αo is calculated according to a correctional Nernst equation, which won’t discussed in detail here.

2.2 Laboratory trials

Three heats of 20 kg base iron were melted in a 30 kg medium frequency induction furnace, with overheating temperature at 1500°C. The controlled ranges of chemical compositions are 3.60–3.75 mass% C, 1.45–1.55 mass% Si, 0.30–0.50 mass% Mn, P ≤ 0.05 mass%, S ≤ 0.03 mass%. The covered tundish method was adopted with high quality pig iron, ferromanganese, 45# steel, 75SiFe and nodularizer FeSiRE6Mg7 as furnace charge. 1.0 mass% 75SiFe was covered on the nodularizer at the ladle bottom as first inoculation. After slag removal, 0.2 mass% 75SiFe was mixed into molten iron for secondary inoculation. Removing slag again, some plant ash was covered consciously on the melt surface in order to avoid direct contact with air environment. During holding in treated ladle, oxygen measurement and thermal analysis were carried out with comfortably temperature reduction. Metallographic examination about nodularity and graphite nodular count were carried out from the thermal center of the thermal analysis samples. Addition of nodularizer was 1.5 mass%, 1.2 mass% and 1.0 mass% base iron melt, respectively. Except for the nodularization state, the preparation processes of three heats were essentially the same. Table 1 shows the chemical compositions of furnace charge and treating agent.

Table 1 Chemical compositions of the furnace charge and various treating agents, mass%.

3. Results and Discussion

3.1 Non-equilibrium characteristics of nodularized molten iron

Figure 1 shows the typical changing curves about oxygen activity αo, oxygen potential Eo with tested temperature T during molten iron preparation, the holding times before testing were marked close to the corresponding oxygen potential point. After nodularization, the molten iron was covered with plant ash on the surface of melt for protection. As shown in Fig. 1, the oxygen activity αo of base iron is strongly dependent on the temperature change. Within the measured range of 1485°C up to 1560°C, it presents a approximate linear relationship between αo and T. This correlation is confirmed in other heats with different slop and value. The αo values struggle to maintain a low level for a time after nodularization, but there is a rebounded trend within prolonging holding time. Nodularization (about at 1450°C) made oxygen activity αo and oxygen potential Eo sharply reduced, which corresponded with the intense deoxidization function by spheroidizing elements Mg and RE; The oxygen potential Eo of base iron stays at a stable level, increasing significantly after nodularization in the period of holding time. For three different nodularized melts, there are same changing trends about characteristics of measured oxygen activity. However, the increase of oxygen activity αo and nodularizing degeneration are more obvious for uncovered liquid iron.

Fig. 1

Changing relation of oxygen activity αo and oxygen potential Eo in the preparation of nodularized melt.

Combining the changing trends of oxygen activity αo and oxygen potential Eo can evaluate the metallurgical treatment state. In smelting process, the dissolved oxygen is in equilibrium with the mass of SiO2 like products existed in melt system. Hence, the liquid iron system is in a low energy state. After nodularization and inoculation, severe de-oxygenation and de-sulfurization break the metallurgical equilibrium of the melt system. The oxygen combined in some low stable oxides will decompose and produce more stable MgO with the Mg. Subsequently, oxygen converts to be with Mg to maintain the chemical equilibrium. However, the solubility of Mg in molten iron is extremely low, and the Mg vapor pressure in carbon-saturated molten iron is relatively high. Consequently, Mg vapor will continuously evaporate from the melt, resulting in nodularization effect decline. At the same time, the molten iron is in a state of extreme disequilibrium, resulting in severely absorption oxygen from surrounding atmosphere, furnace wall environment and decomposed oxides to maintain the equilibrium of metallurgical reaction according to the chemical principles. In measurement process, oxygen activity value is determined by oxygen potential Eo and liquid iron temperature T. On the one hand, as the temperature reducing, the oxygen activity decreases in the liquid iron. On the other hand, oxygen activity will increase by taking oxygen from the surrounding environment. The oxygen potential is mainly influenced by the dissolved oxygen content, while, the temperature has relatively small effect on this value. From the experimental results, the oxygen uptake is significantly higher in nodularized molten iron than in base iron. In melting phase, base iron is relatively saturated with oxygen, and the dependence of oxygen activity αo in liquid iron on testing temperature is exponential. The oxygen potentials maintain at a relatively stable level; However, the oxygen content is quickly reduced to an extremely low value during nodularization. When oxygen activity decrease affected by fall of testing temperature is almost comparable to the positive influence of dissolved oxygen increase, the oxygen activity will not change significantly in the short holding period. But, the oxygen potential Eo remains increase. In small ladle without protection on the liquid iron surface, the specific surface area is relatively larger than in big ladle with huge volumetric melt, so the oxygen uptake and the nodularization effect decline are more significant. In a closed environment, the oxygen activity decreases with the reduction of iron liquid temperature. As shown, Compared with oxygen activity and oxygen potential results, it is distinct to present the metallurgy state transition of the nodularized liquid iron.

3.2 Oxygen activity in base iron

Figure 2 shows the corresponding relationship between the oxygen activity αO of three heats base iron and the reciprocal of temperature 1/T. The regression coefficients of measured oxygen activity values, liquidus temperature TL and white width D of the wedge sample are shown in Table 2.

Fig. 2

Dependence of oxygen activity on inverse temperature with a changing Si content in base iron.

Table 2 Characteristic temperatures in various characteristic cooling curves of a base iron.

For three heats, only the addition amount of the nodularizer was different, and the other process parameters kept consistent as much as possible. From the experimental results, the TL has a minuscule reduction, indicating the carbon equivalent CE difference due to the ingredient principle. However, in the three cases, the oxygen activity values measured at equivalent temperature have no changing trends with Si content in ingredient. The CE changes can’t reflected by the oxygen activity values. It also indirectly shows that ordinary oxygen cell can’t be used for the dissolved Si content determination in molten iron, because there exists only the oxygen reduction reaction on the interface of electrolyte and molten iron, where Si and O balance does not established. In this case, SiO2 solid phase can be introduced into the electrolyte interface as auxiliary electrode to balance O in molten iron, by which the Si activity can be measured correctly in the melt.

Nevertheless, the slope of the regression line increases as the Si content or CE increase in base iron. Since this slope value represents the statistical law of multiple oxygen activity measurement points, so it is not accidental. According to the thermodynamics.5,8) The above regression slope is proportional to the standard enthalpy change that is equal to the heat evolved by an exothermic reaction. Therefore, it corresponds directly to the de-oxygenation products (FeO+MnO) in liquid iron. In this experiment, the FeO contents in the three heats were calculated by interpolation method, of which between 25 and 50% consistent with Orths et al.15) The difference in slope of the regression curve reflects the fluctuation of the oxide percent in these base irons. In experimental condition, the Si content increase indicates the reduction of the addition of scrap steel and FeSi alloy, and the increase of pig iron addition. The SiO2 like oxides contained in pig iron are more than other charge, so the tendency of under-cooling is decreased, which correspond with reduction of white width in wedge sample and better eutectic nucleation performance. In addition, the number of various oxide particles in liquid iron is very large, chemical equilibrium is maintained between the dissolved oxygen and the oxide particles in the melt. In nodularization and inoculation, Mg with stronger affinity with oxygen will first take oxygen from low stable oxides and form more complex oxides. At this moment, the change of dissolved oxygen content in molten iron is hard to detect. Subsequently, Mg continues to react with dissolved oxygen. This change of dissolved oxygen content and oxide proportion leads to the fluctuation of initial oxygen activity content, which has important influence on the establishment of a new equilibrium relationship in the process of Mg treatment.16,17)

3.3 Metallurgical quality state of nodularized liquid iron

3.3.1 Optimal nodularization effect

Figure 3 is the oxygen activity curve depending on decreasing measurement temperature. Measured oxygen activities and characteristic temperatures of thermal analysis are shown in Table 3. As shown from Fig. 3, measured oxygen activity αO is presented as an exponential function of the temperature. By the fitting curve, the data are extrapolated toward the liquidus temperature, from which the predicted oxygen activity values of the three heats with eutectic composition tend to different values. For 1.5 mass% and 1.2 mass% nodularizer addition, the initial oxygen activity values are extremely low, and decrease with temperature decline.

Fig. 3

Dependence of oxygen activity on inverse temperature with a changing Si content in base iron.

Table 3 Measured oxygen content data in experiment.

Moreover, initial oxygen activity in 1.5 mass% nodularizer nodularized liquid iron is higher, but with a sharper reducing rate, and the final oxygen activity value still go beyond that of 1.2 mass% nodularized molten iron. The corresponding oxygen activity values in 1.0 mass% nodularized liquid iron are much more than these above two situations. The main reason may be that for the former two, the more the amount of nodularizer is added, the lesser the SiO2 like products will present in the molten iron. Additionally, some SiO2 class oxides will separate from the liquid iron system with the rise of MgO like products to the surface of the molten iron. In the chemical equilibrium system, decrease of the oxide productions makes the oxygen activity value decline depending on the temperature. Indicating from the lowest eutectic temperature TEU, the under-cooling tendency of molten iron will increase. While, the eutectic under-cooling temperature of molten iron with 1.0 mass% nodularizer addition possesses typical solidification characteristic of compacted cast iron. Its minimum eutectic under-cooling temperature TEU is comparable with that of ductile iron and the eutectic recalescence temperature is higher than ductile iron, with a larger eutectic temperature range for compacted cast iron. It is because of the greater non-equilibrium state, the final oxygen activity in the over-nodularized molten iron inversely exceeds that of relatively less-nodularized molten iron. This conclusion is consistent with the decline of the nodularization effect verified by metallographic results. In the initial state of nodularization, the graphite nodularity of over-nodularized molten iron is slightly inferior to that of the less-nodularized molten iron. In any case, the nodularity of the two molten irons both increase with the holding time extending. However, the nodularizing effect declines faster in over-nodularized molten iron. The oxygen activity values are relatively high in the liquid iron added with 1.0 mass% nodularizer, and the flake graphite appears in the later stage of holding. While, the relationship between log αO and temperature T presents linear, of which the slope corresponds to the heat associated with MgO formation.5)

These activity values obtained in experiments are hardly comparable considering the temperature effect. Measured oxygen activity were recalculated to reference values at 1420°C using the fitting liner function. From the recalculated oxygen activity and under-cooling temperature of thermal analysis, it can be seen that the recalculated oxygen activity value increases with the extension of holding time. The optimal nodularity corresponds to the same oxygen activity value as the maximum graphite nodular count. After nodularization, oxygen activity and oxygen potential reach the lowest value. However, the nodularizing effect of the graphite particle is not the best. Graphite particles are roundness but with a small volume fraction. Even the distribution of graphite particles appears orientation, and there is free carbide at grain boundaries. With in-depth study, it is found that there is a certain amount of austenite occurring in the matrix, and the graphite particles are arranged among the austenite dendrites. In this case, the liquid iron is hypoeutectic solidification from the thermal analysis results with severe under-cooling, and the large shrinkage characteristic in the late of solidification leads serious shrinkage cavity and shrinkage porosity. Because nodularization makes the liquid iron in extreme under-cooling state, even if implementing second inoculation, a good heterogeneous nucleation state is not formed within a short period of time. Solidification mode of the molten iron transits to hypoeutectic, a small amount of austenitic crystal nucleus grows up and develops, which affects graphite eutectic solidification into a nodular morphology. As the time of nodularization and incubation extending, there is significantly change in the solidification characteristics of nodularized iron liquid, from original hypoeutectic into eutectic solidification, and after a period of time to hypereutectic solidification. The precipitated graphite also experiences shape transition of nodular, vermicular and final the flake, and the oxygen activity is restored to a higher level. Figure 4 shows the change of the nodularization effect with time at this experiment conditions, the nodularity achieves optimum as high as 85% at the 6 min oxygen activity measurement point (about 1396°C of liquid iron) with optimal graphite count. This conclusion testifies the nodular formation-accelerating function of the inoculation from another point of view.

Fig. 4

Effect of holding time on the morphology of the graphite.

3.3.2 Solidifying mode transition of the nodularized liquid iron

The above experimental results show that the solidification mode of molten iron is a dynamic process, and the non-equilibrium metallurgical state of graphite nodularization increases this dynamic solidification effect. Strong deoxidizing and desulfurizing functions of nodularization eliminate the guiding role of oxygen and sulfur atoms to carbon precipitation (Sulfur and oxygen absorb on prismatic plane of the graphite crystal to induce graphite growth as flake shape), which makes the high temperature graphite phase difficult to nucleation and growth, under-cooling tendency of the liquid iron increases. At this moment, if the graphite nucleation is favorable, the surface active atom Mg or RE in the liquid iron will guide the graphite nucleation to branch and promote the graphite growth as nodular shape. When the number of nucleation is relatively small, too much nodularizing atom Mg or RE will cause the over-nodularization, and the graphite nodular will become un-rounded. If there were relatively large volumes of nucleation in liquid iron, when Mg or RE contents are constant, the number of Mg or RE on the unit nucleation surface is reduced, which is conducive to the stable growth of the graphite nodular and the graphite shape is obtained as roundness. According to the Fe–C phase diagram, if the under-cooling degree of the molten iron increasing, the carbon equivalent at real eutectic point will be shifted to the right and increased, and the actual solidification of molten iron presents as hypoeutectic model. And then the graphite potential is going down. The eutectic growth rate increases rapidly once the eutectic graphite nodule begins nucleation in large volumes. This reduction of under-cooling caused by solidification kinetics increasing the crystallization latent heat released leads to decrease of under-cooling tendency of the liquid iron. Along with the development of solidification, the carbon equivalent at real eutectic point will be shifted to the left and reduced, and the molten iron is transformed into eutectic and even hypereutectic solidification mode. As the eutectic growth rate decreasing, if the beneficial effects of the increase of the nucleation point can’t be able to make up for the negative impact of decrease of the graphite growth rate, the under-cooling tendency of the liquid iron will increase again. The appearance of the secondary under-cooling in the early stage of the eutectic solidification can lead to the deficiency of graphite expansion in the late of eutectic stage, and the shrinkage tendency will increase. On the one hand, with nodularization effect decline, the graphite will transform into worm and flake growth, and the under-cooling tendency of the liquid iron decreases. On the other hand, the inoculation effect declines faster, and the Si in the 75SiFe alloy inoculant will dissolve into liquid iron as alloying element. The development of the two factors will eventually promote the reduction of the carbon equivalent of the eutectic point and the increase of the real carbon equivalent, which results in eutectic solidification for the molten iron.

4. Conclusions

In this paper, the evolution of the dissolved oxygen contents during melting and preparation of the liquid iron for ductile iron were researched by combining oxygen activity measurement and thermal analysis. The most important conclusions from this study may be summarized as follows:

  1. (1)    Experiment results prove that the nodularized liquid iron is in a serious non-equilibrium state from the point of view of physical chemistry metallurgy about oxygen.
  2. (2)    In the melting temperature range of the base iron, oxygen and silicon maintain the chemical reaction balance. However, because solid electrolyte interface of the probe does not establish this Si–O reaction equilibrium, so the measured oxygen activity can not reflect the change of Si content in the molten iron.
  3. (3)    The non-equilibrium state of physical and chemical metallurgy for nodularized liquid iron seriously affects the solidification structure and graphite nodularization effect. It should optimize the nodularization treatment from oxygen behavior and solidification dynamics.

Acknowledgment

The authors gratefully acknowledge the financial support by the National Natural Science Foundation of China (Grant No. 51474082).

REFERENCES
  • 1)  Thermal analysis of cast iron. Product overview of Heraeus Electro-Nite, https://www.heraeus.com/media/media/hen/media_hen/products_hen/iron/thermal_analysis_of_cast_iron.pdf, (cited 2018-09-24).
  • 2)  V. Anjos: PhD thesis, (2015) University of Duisburg-Essen, DE.
  • 3)   D.M.  Stefanescu: Int. J. Metalcast. 9 (2015) 7.
  • 4)  C. Andersson, P.P. Katrincholm and T.A. Hills: U.S. Patent: US003783570B1 (2004).
  • 5)   F.  Mampaey,  D.  Habets,  J.  Plessers and  F.  Seutens: Int. Foundry Res./Giessereiforschung 1 (2008) 2.
  • 6)   J.  Hampl and  T.  Elbel: Arch. Foundry Eng. 4 (2010) 55.
  • 7)  F. Mampaey: Proc. I-SUP2008, (Sustainable Production, Bruges Belgium, 2008) p. 176.
  • 8)   F.  Mampaey,  D.  Habets,  J.  Plessers and  F.  Seutens: Int. J. Metalcast. 4 (2010) 25.
  • 9)   J.  Ekengård,  A.  Diószegi and  P.G.  Jönsson: Int. J. Metalcast. 10 (2016) 500.
  • 10)  T. Kusakawa, X. Xu and S. Okimoto: Report casting. research lab., (Waseda University, 1988).
  • 11)   L.  Elmquist,  J.  Orlenius and  A.  Diószegi: AFS Trans. 115 (2007) 625.
  • 12)   J.  Hampl,  T.  Elbel and  T.  Valek: Metalurgija 2 (2015) 323.
  • 13)  X.G. Pan and M.J. Yang: China Patent: CN102863213A (2013).
  • 14)  X.G. Pan and M.J. Yang: China Patent: CN103586455A (2014).
  • 15)   K.  Orths,  W.  Weis and  M.  Lampic: Giessereiforschung 4 (1975) 113.
  • 16)   Y.  Igarashi and  O.S.  Senri: Int. J. Cast. Met. Res. 11 (1998) 83–88.
  • 17)  T. Skaland: Proc. AFS Cast Iron Inoculation Conference, (Schaumburg, Illinois, 2005) pp. 13–30.
 
© 2018 The Japan Institute of Metals and Materials
feedback
Top