High-Temperature Mechanical Behavior and Fracture Analysis of a

High-Temperature Mechanical Behavior and Fracture Analysisof a Low-Carbon Steel Related to Cracking

BEGONA SANTILLANA, ROB BOOM, DMITRY ESKIN, HIDEO MIZUKAMI,MASAHITO HANAO, and MASAYUKI KAWAMOTO

Cracking in continuously cast steel slabs has been one of the main problems in casting fordecades. In recent years, the use of computational models has led to a significant improvementin caster performance and product quality. However, these models require accurate thermo-mechanical properties as input data, which are either unreliable or nonexistent for many alloysof commercial interest. A major reason for this lack of reliable data is that high-temperaturemechanical properties are difficult to measure. Several methods have been developed to assessthe material strength during solidification, especially for light alloys. The tensile strength duringsolidification of a low carbon aluminum-killed (LCAK; obtained from Tata Steel MainlandEurope cast at the DSP plant in IJmuiden, the Netherlands) has been studied by a technique forhigh-temperature tensile testing, which was developed at Sumitomo Metal Industries in Japan.The experimental technique enables a sample to melt and solidify without a crucible, makingpossible the accurate measurement of load over a small solidification temperature range. In thecurrent study, the tensile test results are analyzed and the characteristic zero-ductility and zero-strength temperatures are determined for this particular LCAK steel grade. The fracture sur-faces are investigated following tensile testing, which provides an invaluable insight into thefracture mechanism and a better understanding with respect to the behavior of the steel duringsolidification. The role of minor alloying elements, like sulfur, in hot cracking susceptibility isalso discussed.

DOI: 10.1007/s11661-012-1331-1� The Minerals, Metals & Materials Society and ASM International 2012

I. INTRODUCTION

CRACKING at the surface of continuously cast steelslabs has been one of the main problems in casting formany years. Many of the cracks that occur duringsolidification are hot tears.

In recent years, the use of computational models hasled to a significant improvement in insight in casterperformance and product quality. However, these mod-els require accurate thermomechanical properties asinput data, which are often unreliable or even nonex-istent for many alloys of commercial interest.[1,2] More-over, the thermomechanical properties of steels are stillpoorly known at high temperatures, close to the solidus.

A major reason for this lack of reliable data is thathigh-temperature mechanical properties are difficult tomeasure. Several methods developed to assess thematerial strength during solidification and close to thesolidus, particularly for aluminum alloys,[3] could beadapted for steel.On the other hand, the problem of hot tearing during

continuous casting of steels has been extensively studiedin the literature from the viewpoint of the castingprocess. Numerous studies refer to the influencingprocess parameters, such as chemical composition,cooling rate, or strain rate. The treatment of hot tearingin steels has some peculiarities as compared to alumi-num alloys. This is mostly related to the specificsolidification structure that is formed during steelsolidification.The solidification process in steel is conventionally

divided into three different stages[3,4] schematicallyshown in Figure 1:Stage 1: Formation of dendrites during solidification

and mass feeding. As the steel starts to solidify, it does soin a dendritic structure. While cooling, secondarydendrite arms form at a short distance behind theprimary dendrite tips. These dendrites have no mechan-ical bond with each other, as they are separated byliquid steel. Consequently, the shell at this location haslittle or no strength as both liquid and solid are free tomove. At this stage, the name ‘‘mushy zone’’ ismisleading, as it is actually a slurry with the newlyformed dendrites suspended in the liquid. Only after the

BEGONA SANTILLANA, Principal Researcher, is with Tata SteelEuropeResearchDevelopment&Technology, P.O. Box 10000, 1970CAIJmuiden, The Netherlands. Contact e-mail: [email protected] ROB BOOM, Emeritus Professor and Senior AdvisorM2i, is with theDepartment ofMaterials Science andEngineering,DelftUniversity of Technology and Materials innovation institute M2i,Mekelweg 2, 2628CD, Delft, The Netherlands. DMITRY ESKIN,Professor, is with the Brunel Centre for Advanced SolidificationTechnology, Brunel University, Uxbridge UB8 3PH, U.K. HIDEOMIZUKAMI, Senior Research Engineer, and MASAHITO HANAO,Researcher, are with the Corporate Research & Development Labora-tories, Sumitomo Metal Industries, LTD, 16-1 Sunayama, Kamisu,Ibaraki-Pref. 314-0255, Japan. MASAYUKI KAWAMOTO, GeneralManager, is with the Sumitomo Metal Industries, LTD, Triton SquareOffice Tower Y 8-11, Harumi 1-Chome, Chuo-Ku, Tokyo 104-6111,Japan.

Manuscript submitted January 25, 2012.Article published online August 11, 2012

5048—VOLUME 43A, DECEMBER 2012 METALLURGICAL AND MATERIALS TRANSACTIONS A

temperature has dropped below the coherency point, areal mush is formed.[3,4]

If thermal strain is imposed on the semisolid materialat this early stage of solidification, the resulting openingin the structure can be filled by intergranular orinterdendritic liquid.[5]

Stage 2: Coarsening and impingement of dendritesduring solidification. This stage can be divided into twosubstages, as follows:

(a) Coarsening of dendrites during solidification andinterdendritic feeding: Secondary dendrite arms be-gin to reach out and interlock with each other andthe individual dendrites start to connect with eachother giving the solidified shell some strength.However, at this stage, the permeability of the net-work is still large enough to prevent defect forma-tion by melt feeding.

(b) Coarsening of dendrite arms and interdendriticseparation: Dendrites reach the stage when theinterlocked secondary dendrite branches becomeindistinguishable. As the temperature drops fur-ther, the structure starts to resemble columns with-out visible dendrite branching. With increasingsolid fraction, liquid is isolated in pockets orimmobilized by surface tension.[3,4]

During these substages, as liquid is trapped betweeninterlocking dendrites, the free passage for liquid isblocked, transforming continuous liquid films intoisolated liquid droplets. As a result, the strength of thematerial is very low due to the existence of thisnoncontinuous liquid film between the primary den-drites. If an external stress is applied to the material,then hot cracking can easily occur. This type of crackingis defined as solidification cracking, interdendriticcracking, or hot tearing.

Stage 3: Grain formation at the end of solidification andbridging or solid feeding. At this stage, the boundaries ofthe primary dendrites become invisible on polished

sections. This is called the transition from the dendriticto the grain structure. A thin liquid film can still bepresent at the grain boundaries due to the presence ofsegregated elements in the liquid, lowering the meltingpoint of this film. Hot cracks formed at this stage arenamed grain-boundary or intergranular cracks, or hotcracks to distinguish them from the hot tears formedduring stage 2. Solid-state creep is the only way toaccommodate solidification shrinkage and thermalstresses.[3,4]

Grain structures behave somewhat differently fromthe columnar structures of dendrites. The arrangementof grain boundaries is random, causing the topology ofany remaining liquid films to be very intricate. Thegradual transition of a continuous liquid with isolatedgrains or clusters to a solid network with isolateddroplets or films of liquid is called ‘‘percolation.’’[5] Atthis point, if strain is applied and separation of grainsoccurs, then the isolated liquid droplets can form liquidbridges observed at the fractures.The development of mechanical strength of the

semisolid steel can be characterized by two temperaturesthat also determine the hot cracking susceptibility of aparticular steel.The zero strength temperature (ZST) is defined as the

temperature during cooling at which forces can first betransmitted perpendicular to the solidification direc-tion.[6] This ZST corresponds to a fraction of solid (fS) inthe range of 0.65 to 0.80. At this range (stage 2), becausea segregated liquid film still exists between the dendrites,a small strain leads to cracking. This results in a crack orsegregate-rich zone depending on whether or not liquidsteel can still flow between the dendrites.As the steel solidifies and cools further, the strength

and ductility of the material increases and at some pointthe material acquires plasticity.[6] The temperature atwhich the transition occurs from brittle to ductilebehavior is known as the zero ductility temperature(ZDT) and is commonly associated with a fraction of

Fig. 1—Solidification structure with characteristic temperatures in hot tearing.

METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 43A, DECEMBER 2012—5049

solid (fS) between 0.98 and 1. In Figure 1, a schematicrepresentation of the ZST, ZDT and the solidificationstructure is shown.

The range between the ZST and ZDT is defined as thebrittle temperature range (DTB) and gives a qualitativeguide to hot crack sensitivity.[3,7]

DTB ¼ ZST� ZDT ½1�

The wider the range, the longer the period of time (ata constant cooling rate) that the semisolid steel stays inthis critical range, and the larger the stress concentrationresulting from the thermal contraction restriction. It canbe stated that the tendency to crack formation dependson the brittle temperature range of the alloy.[3,7]

The steel first solidifies in the form of dendrites. Thesedendrites when growing will reject the solute elementswith the partition coefficient<1 to the liquid. As a resultof incomplete equilibrating diffusion in the solid, theconcentration of the solutes in the liquid phase willprogressively increase and result in the formation oflow-melting phases. This process is known as microseg-regation. The movement of the solidification front andflow motion may translate this microsegregation to thescale of the casting (sample), forming macrosegregation(uneven chemical composition on larger scale) in thesolidified alloy.[8]

Below the equilibrium solidus, the liquid phase rich instrongly segregating elements, such as sulfur and phos-phorus, is located interdendritically and at the grainboundaries; therefore, when the quantity of liquid phasedrops, the sulfur and phosphorus concentration raisesrapidly.[9] As a result of the strong segregation, thesolidification temperature of this film is much lower thanthe solidus calculated from the initial composition of thesteel.[4,9,10]

Therefore, the segregating elements in the steel caninfluence the hot tearing susceptibility as they widen theDTB, displacing its lower limit to lower tempera-tures.[4,9–11]

The upper temperature of brittleness range dependsmainly on the carbon concentration; the influence ofsulfur or phosphorus will be more sensitive for theminimum ductility temperature (MDT). Other elementsin the steel, e.g., Mn, may adjust this temperature andfraction of solid through reactions with the segregatingelements. The brittle–ductile transition temperature willbe controlled by minor elements and its tendency towardsegregation.

The application of this or any other hot tearingcriterion is hindered by the lack of factual data on themechanical properties of solidifying steels.

The ability to predict at least the tendency of hottearing of a solidifying steel is of great interest due to thecontinuous trend towards increased casting productivityby increasing casting speed, combined with higherquality demands and the development of new steelgrades.

This article is aimed at demonstrating the possi-bility of high-temperature tensile testing of steels andits application to the analysis of the hot tearingphenomenon.

II. MATERIALS AND METHODS

A. LCAK Steel Grade

The tensile strength during solidification of an LCAKsteel (chemical composition listed in Table I; cast at theDirect Sheet Plant of Tata Steel Mainland Europe inIJmuiden, the Netherlands) has been studied by atechnique developed at Sumitomo Metal Industries,Japan for high-temperature tensile testing.[12,13] Theexperimental technique enables a sample to melt andsolidify without a crucible, resulting in accurate mea-surement of load over a small solidification temperaturerange.Two sets of experiments were both done with the

same LCAK steel grade but from different heats. Thefirst set of experiments was done to evaluate thepossibility to gather relevant information for crackingsusceptibility for this steel grade and the elongation wasnot necessary at this point of the investigation; there-fore, it was not recorded. Considering that the results ofthis first test were very promising, it was decided to carryout a second set of tests but with elongation to analyzethe brittle temperature range not only visually but alsowith the actual data.Nevertheless, the two sets of tests as illustrated

here show the reproducibility of the mechanical prop-erties of a commercial steel grade and the accuracy ofthe testing results, considering that the samples comefrom two different heats with same nominal chemicalcomposition.Two sets of tests have been carried out. In the first set,

one measurement was taken every 20 deg and someuncertainties arose, leading to the second set of testswith one sample tested every 10 deg starting fromequilibrium solidus temperature. In addition, in thesecond set of tests, the elongation of the samples wasestimated from the recorded displacement of the sampleduring testing.

B. Tensile Test Device

Tensile tests were performed in a temperature rangefrom 1373 K (1100 �C) to 1793 K (1520 �C) using theapparatus shown in Figure 2.[12,13]

The testing rig is placed in a chamber filled withAr gas and kept during testing at 1 atmosphere ofpressure. The high-temperature tensile test apparatus is

Table I. Critical Temperatures and Composition

(Weight Percentage) of LCAK Steel

Parameter Value

T liquidus (equilibrium) 1806 K (1533 �C)T solidus (equilibrium) 1785 K (1512 �C)C 0.045Mn 0.210P 0.005S 0.005Si 0.010N 0.0025Ca 0.0030


an assembly of a high-frequency induction coil with afrequency of 100 Hz and an output of 15 kW, a loadingcell, and a two-color optical pyrometer.[12,14,15]

The diameter of the sample is 10 mm and the length is100 mm.[12,14,15] The central region of 5 mm in length ofthe sample is melted by high-frequency inductionheating, and the load is applied and measured afterreaching the prescribed conditions. The gravity forcedoes not influence the measured load because the shapeof the molten zone is maintained by the surface tensionand the electromagnetic force.

The temperature of the molten zone is controlled byan R-type thermocouple welded to the sample surface at5 mm from the molten zone. In addition, a two-coloroptical pyrometer is used to control the temperaturetogether with the R-type thermocouple.

The temperature of the molten zone is estimated usingthe regression line obtained in advance from a series ofseparate experiments where the relation between thetemperature of molten zone and the control point hasbeen determined.[12,14,15] The regression line is calculatedusing a set of samples equipped with a thermocouple inthe center of the sample, a thermocouple in the normalcontrol location that is linked to the thermal controller,and the pyrometer. The procedure is to measure with allthese devices every 100 K (100 �C) during heating thesamples from 1173 K (900 �C) to 1673 K (1400 �C) andduring cooling to 1173 K (900 �C). For the actualtesting at different temperatures, the controller is setincluding the error according to the regression line. Theregression formula used is given in Eq. [2].

ST ¼ 1:017CTþ 7:4907 R2 ¼ 0:9988� �

½2�

where ST is the sample thermocouple temperature andCT is the controller temperature.

Since the regression line is calculated using a set ofsamples equipped with a thermocouple in the center ofthe sample, it is not possible to reach melting temper-atures with this method without avoiding melting anddissolution of the thermocouple itself into the liquidpool. Therefore, the regression line is determined up to a

temperature that assures no melting of the sample.Consequently, the regression line for the temperaturesaround the molten zone is extrapolated.A pyrometer is used as a backup means to measure

the temperature in case the thermocouple welded in thesample’s surface would fail during testing. In addition,the pyrometer’s data are also included in the determi-nation of the regression line.Each sample was heated, melted, and solidified

sequentially according to the control conditions ofhigh-temperature tensile test. The temperature of thecontrolled zone was raised from room temperature to30 K below the solidus temperature and maintained atthis temperature for 30 seconds. Then, the temperaturewas raised to 30 K above the liquidus temperature andmaintained for 120 seconds to establish the steady state.Afterwards, the sample was cooled at a cooling rate of0.17 (Ks�1). When the temperature of the molten zonereached a preset temperature, the sample was pulled at astrain rate of 1 9 10�2 s�1.[12,14,15]

After testing, the samples were cooled by turning offthe heating coil until reaching room temperature insidethe chamber. The purpose of the high-temperature-range test is to measure the material strength in thesolidification region, thus mimicking the mold region ofa continuous caster.From the tensile test data, the measured forces can be

converted into nominal stress, if the area of the load isknown, with the following equation:

r ¼ F

A½3�

where r is the nominal tensile stress (Pa), F is the forceto pull the sample (N), and A is the loaded area (m2).Following the findings of Nakagawa et al.,[10] the

loaded area is taken as the cross-section area of thesamples. It was not possible to evaluate the instanta-neous area during testing (reduction of area). However,considering that there is no ductility in the brittletemperature range, the area of the sample will be nearlythe same until failure.This instantaneous gauge length (elongation with

respect to the starting set point measured by displace-ment of the sample during testing), gives a nonzerostarting point for the measurement; consequently for thetest results, the point of ZDT has nonzero values.Instead, it is better to name this point as MDT for thetest results analyzed in this study.

III. MECHANICAL TEST RESULTS

The results of tensile testing are considered to berepresentative based on the previous experience ofSumitomo Metal Industries.[12] The fracture surfacesof tested specimens were preserved after testing andstudied by naked eye and by scanning electron micro-scopy (SEM). The results were interpreted in terms ofpossible hot tearing mechanisms. With the test data andthe postmortem study under SEM, the brittle temper-ature range defined is within the temperatures typical ofthe continuous casting mold.

Fig. 2—Scheme of the hot tensile test apparatus used in this study.


A. Visual Analysis of the Tensile Samples

Figure 3 shows the tested tensile specimens from thesecond set of test samples. From this figure, we candeduce that the MDT lies in the range from 1663 K(1390 �C) to 1673 K (1400 �C) and the ZST lies in therange from 1733 K (1460 �C) to 1743 K (1470 �C). Theelongation at fracture (ductility) increase is clearlyvisible in the sample tested at 1663 K (1390 �C)(Figure 3).

There is a difference between both sets of samples forthe testing temperature 1673 K (1400 �C). In the firstset, this sample already shows some ductility andelongation at fracture; however, in the second test, thissample seems to be brittle with a flat fracture.

Nevertheless, this visual examination situates thebrittle temperature range in almost the same tempera-ture range for both sets, confirming good reproducibilityof the test.

B. Tensile Strength vs Temperature

The comparison of tensile data for the two testsperformed is shown in Figure 4 (squared points: first setof tests; rhomboid points: second set of tests). It can benoted that the accuracy and consistency of the data isvery good.

The two sets of points for the tensile data follow anonlinear function of temperature (interpolation curvein Figure 4), following a Sigmoidal logistic equation asfollows:

r ¼ a

1þ e�bðT�T0Þ½4�

where r is the nominal tensile stress (Pa); a is a constantequal to 28, also referred as amplitude; b is a coefficientequal to –0.03; and T0 is the inflexion point of the curve,in this case equal to 1613 K (1340 �C).Having the displacement data as an instantaneous

gauge length for each temperature, it is possible to buildthe elongation vs temperature diagram and to determine

Fig. 3—Second set of sample after testing.

Fig. 4—Stress vs temperature results for both sets of tests and thecorresponding elongation (only for the second test) used in the deter-mination of zero strength temperature and minimum ductilitytemperature.


not only visually with the broken samples (Figure 3) butalso with the actual data the minimum ductility involvedin the brittle temperature range (see triangle points inFigure 4).

The elongation in the solidification range is minimumat the MDT (which can be around the nonequilibriumsolidus), and it increases at higher temperatures and toinfinity at the liquidus and above. Note: The stress athigh temperatures can be too low to be detected and yetenough to strain weak and highly ductile material.Actually, this is not really elongation but fluidity (as thereciprocal of viscosity). Since in some tests the force isvery low or even nil, fracture is detected visually througha glass window in the chamber.

C. Zero Strength Temperature

The ZST should correspond to a fraction of solid (fS)of around 0.65 to 0.80, which means that in the currentexperiment the solidification continues below the equi-librium solidus. At this stage of solidification, a segre-gate-rich liquid film still exists between the dendrites;therefore, even a small strain may lead to cracking.

At this temperature, the stress starts to build up, butdue to the lack of strain in the material, the cracking willbe brittle. Therefore, ZST can be accurately measuredby knowing the moment when the stress has a valuehigher than zero but the strain remains low.

Having the possibility of accurate measurements ofstress and strain during tensile testing, the two charac-teristics of this particular temperature are easily deter-mined by combining the data on temperaturedependences of nominal stress and total elongation(Figure 4).

It is concluded that for the LCAK steel grade, theZST is between 1713 K (1440 �C) and 1733 K(1460 �C).

D. Minimum Ductility Temperature

The temperature at which the transition occurs frombrittle to ductile behavior is known as the ZDT and iscommonly associated with a fraction of solid (fS)between 0.98 and 1.

For the LCAK steel grade, by combining the temper-ature dependences of nominal stress and total elonga-tion temperature and the dependence of true stress ontrue strain, the MDT could be determined around1673 K (1400 �C) (Figure 4), which is in good agree-ment with the visual assessment.

IV. FRACTURE SURFACES

To get insight into the mechanisms of hot tearingsusceptibility, the fracture surfaces of specimens tested atthe limits of the brittle temperature range (see Figure 3)were analyzed under SEM. All the samples tested abovethe ZST show a smooth fracture, which is typical of thefracture in the presence of a continuous liquid layerseparating dendrites, testifying for the nonequilibriumsolidification that ends at some temperature below the

equilibrium solidus. The results are shown for thespecimen tested at 1733 K (1460 �C) considered to bearound ZST, the specimen tested at 1713 K (1440 �C) asit is in the middle of the DTB, and the specimen testedaround the MDT at 1693 K (1420 �C).

A. ZST Sample at 1733 K (1460 �C)The sample tested at 1733 K (1460 �C) (from the first

set) clearly shows brittle behavior, being close to theZST. In Figure 5(a), a top view of the sample fracture isshown.Towards the center of the sample, there is an

area with dendrites growing in several directions(Figure 5(b)). These dendrites could be growing in voidsoriginated during testing and filled by the liquid.The central part of the fractured sample exhibits

smooth grain boundaries and bridges typical of thegrain separation through liquid film, as shown inFigure 5(b). This confirms the previous observationsby Rogberg,[16] Hansson,[4] and Dantzig and Rappaz.[5]

The amount of liquid present in the center of the sample(as compared with the expected amount from theequilibrium phase diagram) is increased due to the

Fig. 5—(a) Surface fracture of the 1733 K (1460 �C) specimen show-ing also the liquid pool in the center. (b) Detail with bridges anddendrites on the 1733 K (1460 �C) fracture surface.


segregation as a result of progressive solidification fromthe surface to the center.

In the second set of test samples, the fracture surfaceis similar, with the same type of dendrites and inclu-sions, particularly showing segregated interdendriticliquid, probably formed as a result of segregation of

alloying elements (Mn, S, etc.) as can be seen inFigure 6.The fracture of the samples tested at 1733 K

(1460 �C) shows stage 2 of hot crack formation, withthe dendrites covered with a thin liquid film. At thisstage, the material has just passed (or is in the vicinityof) the point when the strength starts to build up (ZST)and liquid films are still trapped between the interlock-ing dendrites; therefore, the resistance to externalstresses is very low, and hot cracking can easily occur.Consequently, the LCAK steel considered in this studywill show an interdendritic cracking behavior in tem-peratures around 1733 K (1460 �C).

B. 1713 K (1440 �C) Samples

Figures 7(a) and (b) clearly show the presence of aeutectic pattern on the fractures of the samples fromboth sets of tests, which testifies for the presence of aliquid film separating the grains at this stage ofsolidification.The eutectic pattern not only demonstrates that the

liquid phase exists between the grains but also attests forthe high solute concentration of this liquid due tosegregation. In this case, the liquid phase may beFig. 6—Detail of the 1733 K (1460 �C) sample showing segregated

liquid between dendrite branches.

Fig. 7—(a) Eutectic structure features and voids between dendritebranches in the specimens tested at 1713 K (1440 �C). (b) Eutecticpatch at the dendrite face.

Fig. 8—(a) Fracture surface of the 1693 K (1400 �C) specimens:grain facets with wavy liquid patterns. (b) Bridges at the grain sur-face and void between two grains with a liquid film that seems to bepercolating through the void.


preserved to temperatures considerably lower than theequilibrium solidus, which is one of the reasons for lowductility.[10,17]

Similar eutectic structures have been observed byNakagawa et al.[10] when conducting comparable hottensile testing on steel grades with higher carboncomposition. They reported that at temperatures about100 K below the equilibrium solidus, the samples didstill have liquid, and the displacement during testingremained small. This difference has been also attributedto the segregation in the interdendritic liquid, and thetemperature difference along the radius directionof the sample may be related to this macroscopicsegregation.[10]

It is important to point out that in this type ofexperiments the temperature gradient always exists inboth axial and radial directions. The gradient is obviousfrom the fracture surfaces that show different structuresat the surface and in the center of the sample. As aresult, the assignment of temperature to a particular setof mechanical properties is always a result of ‘‘volume-averaged’’ procedure. The setup of the experiment isdesigned and optimized in such a way as to minimize thegradients and make the temperature measurementduring the testing accurate.

C. MDT Samples at 1673 K (1400 �C)The samples tested at 1673 K (1400 �C) are close to

the MDT, showing in the center of the sample

as-solidified faceted ferritic grains (Figure 8(a)). Thefracture is intergranular and brittle with some signs ofthe liquid present at the facets (wavy patterns inFigure 8(a) and liquid film in Figure 8(b)). In the secondset of samples tested, bridges were found in the grainssurface (Figure 8(b)).The fracture of these specimens show stage 3 of hot

crack formation; only the grains were observed and nodendrite morphology can be seen in the fracture surface(except for the coarse dendrites close to the surface dueto testing conditions). As a result of segregation ofalloying elements, a thin liquid film, apparently with alow-melting point, is present between grains. Aroundthis temperature, the transition between brittle andductile fracture begins (MDT), changing the fracturebehavior from interdendritic to intergranular.These findings are similar to a previous study carried

out by Hansson[4] for INVAR alloys. Rogberg[16] hasalso found lines of solidified melt, which appeared oncrystal surfaces, supporting the theory of the existenceof thin films causing brittle behavior above MDT.

D. Nonmetallic Inclusions on the Fracture Surfaces

The fracture surfaces of the examined samples showsome interesting features that may help in understandingthe nature of the last liquid remained during solidifica-tion in this steel grade.In support of the mechanisms of microsegregation

described in the introduction, manganese sulfides (MnS)

Fig. 9—(a) Dendrites with MnS particles in their surface in the 1713 K (1440 �C) sample Note: Figures have different magnification. (b) MnSparticles in the dendrite surface in the 1673 K (1400 �C) sample. (c) Type III MnS particles on the fracture surface of the 1653 K (1380 �C)sample. (d) Type III MnS particles on the 1633 K (1360 �C) sample, showing also a segregated liquid bridge.


were found at the interdendritic fracture surfaces of allexamined specimens of the LCAK hot tensile samples.Figure 9 gives representative examples of these inclu-sions at some of the test temperatures, i.e., 1713 K(1440 �C), 1673 K (1400 �C), 1653 K (1380 �C), and1633 K (1360 �C). Similar MnS particles have beenfound previously by other authors.[4,10,18,19] The sulfideparticles in the samples at lower temperature have therosette shape typical of divorced eutectics.

Traditionally, sulfides can be classified as types I, II,and III.[20] Type I particles appear as individual spher-oids of a wide size distribution scattered randomly in theinterdendritic spaces. Type II particles appear as adegenerate eutectic network made of irregular cylindri-cal rods, spread out in the interdendritic spaces, delin-eating the dendrite arm boundaries. The classical‘‘beads’’ and ‘‘stringers’’ are in reality intersections ofa eutectic network by the polishing plane. Type IIIparticles appear as idiomorphic crystals scatteredthrough the whole dendritic structure but frequentlyare situated in interdendritic spaces. They often projectarms that join type II particles.[20]

These findings correlate very well with the type ofparticles found in this research. Type I can be found athigh temperatures (Figure 9(a)). The type II particlesare clearly visible in the sample at 1673 K (1400 �C)(Figure 9(b)). Rogberg[16] and Hansson[4] encounteredthe same type of particles (interconnected stringers as asort of degenerate eutectic) also on fracture surfaces oftensile specimens (even though their test setup wasslightly different).

In a close study on the morphology of type IIIinclusions, Flemings and Bigelow[20] found that theseparticles show a cubic pattern often embedded in‘‘dendritic’’ projections of the same inclusion and thatthey often surround regions of type II inclusions, withprojecting arms that join the interdendritic network. Theparticles found on the fracture surfaces of lowertemperature samples in this research compare very wellwith the findings of Flemings and Bigelow (Figures 9(c)and (d)).

In the analyzed specimens, the sulfide particles growin amount and volume with decreasing temperature ofthe test, which concurs with the assumption that theyare formed from the segregated liquid. The liquid phasein the samples tested at a lower temperature has moretime to segregate the sulfur; therefore, the MnS precip-itates are bigger and more ‘‘eutectic’’ in appearance.After fracture, the induction heating system is turnedoff; accordingly, the solidification goes on until fracturewith the same established cooling rate. Thus, it isobvious that the samples tested at lower temperaturesare for longer time at the given cooling rate, but thesamples tested at higher temperatures the solidification/cooling rate are changed after fracture.

This observation is in good agreement with Flemingsand Bigelow,[20] who showed a relation between themorphology and the size and type of inclusions with thedistance from the ingot cooling surface. By increasingthe distance, there is an increase in the average diameterof the type I inclusions and an increase of the eutecticrod diameter and in average eutectic spacing in type II

inclusions. Even when the cooling rate of the tensilesamples is not the same as in this research, each samplecould be related with distance from the surface of an

Fig. 10—(a) MnS rosette particle with microscopic particles contain-ing Al and Ti from the 1633 K (1360 �C) specimen. (b) MnS rosetteparticle with microscopic particles containing Al and Ti from the1603 K (1330 �C) specimen.

Fig. 11—Energy-dispersive X-ray spectroscopy analysis of the1603 K (1330 �C) specimen.


ingot/cast slab; therefore, the mechanism of particlesgrowing does fit very well with the reported one.

It is interesting to note that microscopic particlescontaining Al (Al2O3 or AlN) and Ti have been found inthe MnS rosette matrix (Figure 10). However, it is notpossible to determine the exact composition (TiN orTiO2) of these ‘‘microparticles’’ due to their size, whichare smaller than a micron (Figure 11). The occurrence ofthese particles may indicate the complex nature of theeutectics.

V. CONCLUSIONS

The hot tensile tests have been carried out atSumitomo Metal Industries, Japan for an LCAK steelgrade cast at the DSP plant in IJmuiden at Tata SteelMainland Europe, The Netherlands. The tensile testdata and fracture surfaces of samples have been ana-lyzed. The following results were found:

1. The steel under given testing conditions solidifiesunder nonequilibrium conditions.

2. The MDT of the LCAK steel is in the vicinity of1673 K (1400 �C).

3. The ZST of the LCAK steel is between 1713 K(1440 �C) and 1733 K (1460 �C).

4. At temperature around MTD this alloy has inter-granular failure; however, in the vicinity of theZST, cracks propagate interdendritically.

5. The structural features indicating the presence ofthe liquid have been observed on fracture surfacesof samples tested in the brittle temperature range.

6. Sulfur segregates to the liquid in the course of non-equilibrium solidification and forms MnS from thesolute-enriched liquid located at dendrite and grainboundaries. This nonequilibrium liquid may in-crease hot cracking sensitivity of the steel, therebylowering the MDT.

ACKNOWLEDGMENTS

This research was carried out under project numberM41.5.08320 within the framework of the Research

Program of the Materials innovation institute M2i(www.m2i.nl). The authors would like to thank thecontribution of Prof. Andrew Howe in the fracto-graphic analysis and the discussion of the results ofthis study. The extensive SEM studies were performedtogether with MSc Sabri Sengo and Dr. Jean Campa-niello from Tata Steel RD&T.

REFERENCES1. K.O. Yu: Modelling for Casting and Solidification Processing, 1st

ed., Marcel Dekker, New York, NY, 2009. .2. C. Monroe and C. Beckermann: Mater. Sci. Eng. A, 2005,

vols. 413–414, pp. 30–36.3. D. Eskin, Suyitno, and L. Katgerman: Progr. Mater. Sci., 2004,

vol. 49, pp. 629–711.4. K. Hansson: Ph.D. Dissertation, Royal Institute of Technology

KTH, Stockholm, Sweden, 2001.5. J.A. Dantzig and M. Rappaz: Solidification, 1st ed., CRC Press,

Lausanne, Switzerland, 2009, pp. 429–72, 519–63.6. Y.M. Won, T.J. Yeo, D.J. Seol, and K.H. Oh: Metall. Mater.

Trans. B, 2000, vol. 4B, 31, pp. 779–94.7. D. Eskin and L. Katgerman: Metall. Mater. Trans. A, 2007,

vol. 38A, pp. 1511–19.8. K. Fisher and W. Kurz: Fundamentals of Solidification, 4th ed.,

Trans Tech Publications, Ueticon-Zuerich, Switzerland, 1998, pp.63–92, 117–32.

9. A. Chojecki, I. Telejko, and T. Bogacz: Theor. Appl. Fract. Mech.,1997, vol. 27, pp. 99–105.

10. T. Nakagawa, T. Umeda, J. Murata, Y. Kamimura, and N. Niwa:ISIJ Int., 1995, vol. 6 (35), pp. 723–29.

11. R. Pierer, C. Bernhard, and C. Chimani: Rev. Metall., 2007, vol. 2,pp. 72–83.

12. H. Mizukami, S. Hiraki, M. Kawamoto, and T. Watanabe: Tetsu-to-Hagane, 1998, vol. 84 (11), pp. 763–68.

13. H. Mizukami, A. Yamanaka, and T. Watanabe: ISIJ Int., 2002,vol. 9 (42), pp. 964–73.

14. N. Baba, K. Ohta, Y. Itoh, and I. Kato: Rev. Metall., 2006, vol. 4,pp. 174–79.

15. T. Kato, M. Kawamoto, A. Yamanaka, and T. Watanabe: ISIJInt., 2003, vol. 43 (11), pp. 1742–50.

16. B. Rogberg: Scand. J. Metall., 1983, vol. 12, pp. 51–66.17. T. Emi and A.W. Cramb: Surface Defects on Continuously Cast

Strands—Casting Volume, 11th ed., The AISE Steel Foundation,Pittsburgh, PA, 2003, pp. 4–6.

18. ASM Handbook, Volume 12: Fractography, 11th ed., ASM Inter-national, Materials Park, OH, 1987, pp. 99, 620–70, 1012.

19. G. Van der Voort: ASMHandbook, Volume 09: Metallography andMicrostructures, 8th ed., ASM International, Materials Park, OH,2004, pp. 220–21.

20. M.C. Flemings and L. Bigelow: Metall. Mater. Trans. B, 1975,vol. 6B, pp. 275–83.


http://www.m2i.nl

Documents

High-Temperature Mechanical Behavior and Fracture Analysis of a