连铸保护渣结晶分析

ISIJ International, Vol. 48 (2008), No. 3, pp. 277–285

277

2008ISIJ

Table 1.

Properties of mold flux used in this study.

Table 2. Chemical composition of mold flux used in this

study.

Fig. 1.

XRD profiles of quenched glassy specimen.

ture range was from 500 to 900°C. After heat treatment,each specimen was cut into two pieces. Half of them wasmolded in the polyester resin and was polished by waterproof paper and by alumina slurry. Then it was etched withsaturated picric acid for a few seconds and the structurewas observed by an optical microscope. The other half wascrushed into powder for XRD analysis.

Furthermore, in order to analyze the change of structurewith time, the interruption tests were carried out at sometemperatures of heat treatment.

2.2.2. In-situ Observation by a Laser Microscope

The surfaces of quenched specimens were polished bywater proof paper and by alumina slurry. After that, the sizeof specimen was adjusted so as to be put in an alumina cru-cible for a laser microscope. The inner diameter of the cru-cible was 5mm. Then the specimen was heated in an in-frared image furnace in argon atmosphere and directly ob-served by a laser confocal microscope. The specimens wereheat-treated at various temperatures for 20min. The tem-perature range was from 600 to 900°C. The heating ratewas set to be 500°C/min. R-type thermocouple was at-tached to the sample holder to control the temperature. Dur-ing the experiments, the events that took place on the sur-face were observed and were recorded by a DVD recorder.3.

Experimental Results

Fig. 2. XRD profiles of the specimen heat-treated at 520°C for

180

min.

Fig. 3. XRD profiles of specimen heat-treated at 650°C for

180min.

3.1. Characterization of Raw Materials

In order to characterize the raw materials in this study,the quenched specimens were analyzed by XRD. One of theresults is shown in Fig. 1. There was no specific peak inXRD profile. The broad peak appeared from 20°and 30°was due to the sample holder which was made of Pyrexglass. The quenched specimens were confirmed to beglassy state.

3.2. Heat Treatment with Electric Furnace3.2.1. XRD Analysis

The XRD profile of the specimen which was heat-treatedat 520°C for 180min is shown in Fig. 2. In this case, nospecific peak was identified and was the same as that shownin Fig. 1.

The XRD profile of the specimen which was heat-treated2008ISIJ

278

at 650°C for 180min is shown in Fig. 3. In this case, somany sharp peaks were recognized and it was found thatthese peaks corresponded to cuspidine (3CaO·2SiO 2·CaF 2).

The specimens heat-treated at 500°C and 520°C for180min remained glassy. However, the specimens heat-treated at other temperature range (over 550°C) crystallizedto cuspidine, though there were some differences in peakheight in XRD profiles.

3.2.2. Analysis of Microstructure

The microstructures of the cross section of specimens,which were heat-treated at relatively high temperature for180min, are shown in Fig. 4. One can recognize that thespecimens were composed by many grains. In case of heat-

Fig. 4. Comparison of grain structure after heat-treated for

180min in the electric furnace observed by the optical

microscope.

Fig. 6. Experimental results of interruption tests. Cross section

of mold flux heat-treated at 650°C in the electric furnace

observed by the optical microscope.

Fig. 5. Comparison of grain structure after heat-treated for

180min in the electric furnace observed by the optical

microscope.

Fig. 7. Experimental results of interruption tests. Cross section

of mold flux heat-treated at 800°C in the electric furnaceobserved by the optical microscope.

treated at 800°C (Fig. 4(a)), columnar grains were ob-served. These grains might grow from the surface of thespecimen. On the other hand, in case of heat-treated at700°C (Fig. 4(c)), only equiaxed grains were observed. Theaspect ratio of these grains is almost unity and it is hardlyrecognized to grow in the specific direction. In case of750°C, columnar grains as well as equiaxed grains were ob-served (Fig. 4(b)). It was found that equiaxed grains wereobserved at relatively low temperature of heat treatment.In order to compare the size of equiaxed grains, the mi-crostructures of the specimens heat-treated at relatively lowtemperatures for 180min have been analyzed. These resultsare shown in Fig. 5. It was found that the equiaxed grainsize increased with rise of temperature of heat treatment.In order to analyze the change in equiaxed grain sizewith time, the interruption tests have been carried out at650°C and 800°C. Figure 6shows the changes in mi-crostructure of the specimens heat-treated at 650°C for var-ious heat treatment times. In case of 15min (Fig. 6(a)), nograin structure was recognized by the optical microscope.

279

At 30min (Fig. 6(b)), many but quite fine grains were ob-served. Further, at 60min (Fig. 6(c)), many grains were rec-ognized. Not only the number of grains but also grain sizeincreased, comparing the microstructure at 30min. At180min (Fig. 6(d)), the grain size increased and the crosssection of the specimen was completely covered withgrains. In case of heat treatment at 15min, cuspidine wasidentified by XRD analysis, though no grain was recog-nized by the optical microscope. The XRD profiles of thespecimen, the grain structure of which were shown in Fig.6, were almost the same and only cuspidine was recognizedas crystalline. Figure 7shows the changes in microstruc-ture of the specimens heat-treated at 800°C for various heattreatment times. In case of 5min (Fig. 7(a)), no grain wasobserved. At 30min (Fig. 7(b)), columnar grains were ob-served near the surface. At 60min (Fig. 7(c)), columnargrains grew toward the center of the specimen and someequiaxed grains were also observed. In case of 180min, columnar grains which grew more were observed.

2008

ISIJ

Fig. 8. Morphological change on the surface during heating

from room temperature to 800°C observed by the laser

microscope (Heating rate was 500°C/min).

Fig. 10. Morphological change on the surface during holding at

800°C observed by the laser microscope. This time in-cludes the heating time from room temperature to800°C.

same moment to reach the aimed temperature.

3.3.2. Morphological Change during Holding

Here again, the experimental results on 800°C and 600°Care presented as examples. The time, when the heating wasstarted, was defined to be t ϭ0.

The morphological changes on the surface during hold-ing at 800°C are shown in Fig. 10. At t ϭ100s (Fig. 10(a)),the grains started to be observed. These grains grew asshown in Figs. 10(b) and 10(c). At t ϭ240s (Fig. 10(d)), thesurface was almost covered with grains. Further, at t ϭ301s (Fig. 10(e)), the morphology on the surface was almost thesame as that at 240s.

The morphological changes on the surface during hold-ing at 600°C are shown in Fig. 11. At t ϭ65s (Fig. 11(a)), itwas just before reaching the aimed temperature, no grainwas observed. At t ϭ70s (Fig. 11(b)), it was almost thesame moment to reach the aimed temperature, many grainsimmediately precipitated on the surface and filled the sur-face. As shown in Figs. 11(c) and 11(d), the grains grewand grain boundaries became clear. At t ϭ800s (Fig. 11(e)),the morphology on the surface was almost the same as thatat 600s. 4.

Discussions

Fig. 9. Morphological change on the surface during heating

from room temperature to 600°C observed by the laser

microscope (Heating rate was 500°C/min).

3.3. Experimental Results of In-situ Observation 3.3.1. Morphological Change during Heating

Here the experimental results on 800°C and 600°C arepresented as examples.

The morphological changes on the surface during heat-ing from room temperature to 800°C are shown in Fig. 8. At 71°C (Fig. 8(a)), no change was observed. At 613°C(Fig. 8(b)), the surface became smooth. At 662°C (Fig.8(c)), the smooth surface turned to be rough in a moment.Many tiny waves were observed on the surface. At 804°C(Fig. 8(d)), this was almost the same moment to reach theaimed temperature, many black rounded particles suddenlyemerged on the surface.

The morphological changes on the surface during heat-ing from room temperature to 600°C are shown in Fig. 9. At 87°C (Fig. 9(a)), no change was observed. At around572°C (Fig. 9(b)), the surface became smooth. At 590°C(Fig. 9(c)), the surface turned to be rough in a moment.However the wavy pattern was weaker than that indicated inFig. 8(c). Then, at 608°C (Fig. 9(d)), many small grainsabruptly precipitated on the surface. This was almost the2008

ISIJ

280

4.1. Characterization of Crystallization

The XRD analysis of the quenched specimens has beenperformed. As shown in Fig. 1, there was no specific peakin the XRD profile. Therefore, it can be concluded that thequenched specimens were available for the study of crystal-lization via glassy state by heat treatment.

Almost all specimens which were heat-treated for

Fig. 12. Cross section of mold flux heat-treated at 800°C in the

furnace observed by the polarizing microscope.

Fig. 11. Morphological change on the surface during holding at

600°C observed by the laser microscope. This time in-cludes the heating time from room temperature to

600°C.

180min crystallized and the species of crystalline was cus-pidine. However, the specimens, which were heat-treated at500°C and 520°C for 180min, did not crystallize as shownin Fig. 2. This indicates that the specimens heat-treatedunder these conditions were still in glassy state and there-fore it is supposed that the specimens were still in the incu-bation period for crystallization.

4.2. Morphology of Grains

As shown in Figs. 4 and 7, the columnar grains seem togrow unidirectionally from surface to center of the speci-men. Further, the equiaxed grains seem to grow radially asshown in Figs. 4, 5, 10 and 11.

In order to clarify the morphology of grain during heattreatment, the thin samples for polarizing microscope havebeen prepared and observations with polarizing microscopehave been made. One example of these observations ofcolumnar grains is shown in Fig. 12. This specimen washeat-treated at 800°C for 180min. Here typical dendriticpatterns were clearly observed and this indicated that thegrain precipitated on the surface grew towards the center ofthe specimen.

The equiaxed grains observed with the laser microscopewere analyzed by the FE-SEM. One example of these ob-servations is shown in Fig. 13. This specimen was heat-treated at 600°C for 20min and corresponded to the speci-men shown in Fig. 11. The grain exhibited the typical den-dritic pattern. The diameter of the grain was approximately20mm and this agreed with the observation results shownin Figs. 11(d) and 11(e).

In short, it is certain that grains on the surface and in thespecimen exhibited the dendritic pattern. Since the under-281

Fig. 13. The grain precipitated on the surface holding at 600°C

for 20min in the infrared furnace observed by the FE-SEM.

cooling for grain growth is necessary for dendritic growthof metals,15) these grains which exhibited the typical den-dritic patterns grew with large undercooling.

4.3. Number and Growth Rate of Grains

In order to analyze the crystallization process from theglassy state, the number of grains and growth velocity ofgrains were characterized in heat treatment. Furthermore,the relation between number and growth velocity of grainsas a function of temperature were discussed.

4.3.1. Number of grains

(1)Number of Grains as a Function of Time

The changes in grain morphologies which were heat-treated at 650°C for various times was shown in Fig. 6. Thenumber of grains as a function of time was evaluated. Here,the number of grains per unit area, which is equal to densityof grains (rgrain ) was characterized. The result is shown inFig. 14. Since these were the results of interruption tests,there were some scatters in these results. However, it can beconcluded that rgrain increased abruptly and after that it wasalmost constant during heat treatment.

Based on the experimental results of in-situ observation shown in Fig. 10, the change in rgrain with heat treatmenttime was analyzed. In this case, the temperature of heattreatment was 800°C. The result is shown in Fig. 15. rgrain rapidly increased in the initial stage and then gradually in-creased during heat treatment. Therefore, it can be safelyconcluded that the generation of grain took place in a mo-ment and rgrain did not increase much during heat treat-2008

ISIJ

Fig. 14. Relation between heat treatment time and density of

grains obtained by the interruption tests. The specimen

was heat-treated with electric furnace at 650°C.

Fig. 16. Relation between temperature of heat treatment and

density of grains.

Fig. 15. Relation between heat treatment time and density of

grains obtained by in-situ observation. The specimen

was heat-treated with infrared furnace at 800°C.

Fig. 17. Relation between temperature of heat treatment and

grain size heat-treated for 180min.

ment. This finding holds true in other conditions of heattreatment within the range of this study.

(2)Number of Grains as a Function of Temperature

As shown in Figs. 14 and 15, the numbers of grains atcertain temperatures of heat treatment were almost constantduring heat treatment, or in final stage of heat treatment.Similarly, rgrain at another temperature of heat treatmenthave been analyzed. The relation between temperature ofheat treatment and rgrain is shown in Fig. 16. Here, all ex-perimental data obtained in this study are included. The ab-solute values of rgrain were not equal each other even in thesame temperature of heat treatment. There may be follow-ing three reasons. They are: the difference of furnace, thedifference of diffusion of atoms or ions (bulk or surface)and the difference of location for observation. However, itwas found that rgrain decreased with rise of temperature ofheat treatment. The reason of this will be discussed in thefollowing Sec. 4.3.3.

4.3.2. Size of Grains

(1)Size of Grains as a Function of Temperature of Heat

Treatment

The cross sections of the specimens, which were heat-treated at various temperatures for 180min, are shown inFig. 5. Almost all grains found in Fig. 5 were spherical andthen the average diameters of grains were evaluated. Theresults of in-situ observation by the laser microscope areshown in Figs. 10 and 11. The grains on the surface werealso spherical and the average diameters of grains were2008

ISIJ

282

evaluated. The relation between temperature of heat treat-ment and average grain size is shown in Fig. 17. The aver-age grain size increased with rise of temperatures of heattreatment in the range of this study. This is attributed to thecoupled effect of nucleation rate and growth velocity as afunction of temperature, which will be discussed in Sec.4.3.3.

(2)Size of Grains as a Function of Heat Treatment Time

and Growth Velocity of Grains

From the interruption tests of heat treatment in the elec-tric furnace, the morphologies of cross section were ob-tained as shown in Fig. 7. The average diameters of grainswere evaluated and the relation between heat treatment timeand average diameter of grain was indicated in Fig. 18. There were some scatters in the experimental results be-cause of interruption test. However, it can be said that thegrains grew at a constant velocity by approximately 5000s and after that the grain diameter did not increase much, be-cause the grains touched to another grains. Here the growthvelocity of grains (V grain ) was obtained to be 1.3ϫ10Ϫ2mm/s.

Based on the experimental results of in-situ observation as shown in Fig. 10, the change of average diameter ofgrains in heat treatment time is shown in Fig. 19. Hereagain, the grain diameter increased linearly in the initialstage and the slope became gentle gradually. Finally, thegrain size became a constant value, approximately 120mm. The growth velocity of grain (V grain ) was obtained from theslope indicated by solid line in Fig. 19. V grain at 800°C was

Fig. 18. Relation between heat treatment time and grain size at

650°C obtained by interruption tests using electric fur-

nace.

Fig. 19. Relation between heat treatment time and grain size at

800°C obtained by in-situ observation. This heat treat-ment time includes the heating time from room temper-

ature to 800°C.

Fig. 20. Enlarged view of morphological change on the surface

during holding at 600°C observed by the laser micro-

scope.

determined to be 0.95mm/s.

When the temperature of heat treatment was low, thegrains were small as shown in Fig. 11. It was impossible todirectly measure the change of grain size with time. There-fore, in this case following two methods were adopted tomeasure the grains size.

Firstly, a certain grain, that generated after many graingrew, was used to measure the size as a function of timewith enlarged view. Morphological changes with time areshown in Fig. 20. The grain paid attention was indicated bya white open circle. The experimental result is shown inFig. 21. The grain linearly increased its size from 140s to320s. After that, the grain size increased slowly because itcame into contact with the neighboring grains. V grain was obtained from the slope in the initial stage of this figure andwas 2.9ϫ10Ϫ2mm/s.

Secondly, the growth velocity of grain was estimatedwithout characterizing the change of grain size with time.According to the in-situ observation, many grains generatedat t ϭ70s in a moment as shown in Fig. 11(b). But it wasnot possible to evaluate the grain size at that time. Thus, thegrain diameter at t ϭ70s was assumed to be 0mm. In con-trast to this, at t ϭ600s, the shape of grains was clearly rec-ognized and the average grain size was also evaluated. As-suming that the growth velocity was constant from t ϭ70s to t ϭ600s, V grain was obtained to be 4.1ϫ10Ϫ2mm/s. Thisvalue is close to the estimated value from Fig. 21,2.9ϫ10Ϫ2mm/s.

283

Fig. 21. Relation between heat treatment time and grain size at

600°C. This heat treatment time includes the heatingtime from room temperature to 600°C.

4.3.3. Relation between rgrain and V grain with TemperatureCrystallization from glassy state is qualitatively ex-plained on nucleation rate and crystal growth rate as a func-tion of temperature for ordinary glass.16) One example isshown in Fig. 22. The driving force for nucleation increaseswith decrease of temperature. On the contrary, the mobilityof atoms or ions decreases with decrease of temperature.Therefore, the nucleation rate exhibits maximum at a cer-tain temperature. The growth rate also exhibits maximum ata certain temperature in a similar way. The nucleation rateis high and the crystal growth rate is low at relatively lowtemperature, for example at T 1in this figure. On the otherhand, the nucleation rate is low and crystal growth rate ishigh at relatively high temperature, for example at T 2. Inte-2008

ISIJ

Fig. 22. Schematic curves of the nucleation rate and the crystal

growth rate as a function of temperature. Temperature

range in this study is also indicated by hatched zone.

Fig. 24. Relation between rgrain and V grain with temperature of

heat treatment. These were the result of in-situ observa-tion using the laser microscope.

Fig. 23. Relation between rgrain and V grain with temperature of

heat treatment. These were the result of interruption

tests using the electric furnace.

gration of nucleation rate for a certain period should be thenumber of grains. Thus, nucleation rate in Fig. 22 directlycorresponds to rgrain in this study. Crystal growth rate isequal to V grain . Therefore, it is possible to qualitatively com-pare present experimental results with this figure.

The experimental results on rgrain and V grain as a functionof temperature of heat treatment using the electric furnaceare shown in Fig. 23. They were discrete data because ofthe interruption tests. The data points are interpreted withsmooth curves. According to this figure, rgrain decreases and the V grain increases with rise of temperature of heattreatment within the range of this study.

In a similar way, the experimental results on rgrain and V grain as a function of temperature of heat treatment on thesurface obtained by the laser microscope are shown in Fig. 24. Here again, the data points are interpreted with smoothcurves. rgrain decreases monotonously with rise of tempera-ture of heat treatment. On the other hand, V grain increases with rise of temperature and exhibits maximum at around800°C. Then, V grain decreases with rise of temperature ofheat treatment.

The experimental results presented in Figs. 23 and 24 aredifferent quantitatively. As mentioned previously, there maybe three reasons. The difference of methods of heat treat-ment, the difference of mobility of ions or atoms and thedifference of location for observation may influence the ab-solute value of rgrain and V grain . However, the fashion of2008

ISIJ

284

change in rgrain and V grain with temperature are quite similar.These experimental results agree with the qualitative repre-sentation as shown in Fig. 22. The range of temperature ofheat treatment in this study is indicated with hatched zonein Fig. 22. Therefore, it can be concluded that the crystal-lization of mold flux from the glassy state is similar to thatof ordinary glass.

The grain size increased with rise of temperature of heattreatment as shown in Fig. 17. The reason of this is that thenumber of grains is low due to low nucleation rate at hightemperature. Further, high growth velocity also leads tolarge grain size at high temperature.

The columnar grains were observed when the tempera-ture of heat treatment was relatively high, as shown in Figs.4(a), 7 and 12. The reason of this is not clear for the mo-ment. However, this may be concerned with nucleation rateas a function of temperature. At high temperature, e.g. T 2in Fig. 22, the number of equiaxed grains is low because oflow nucleation rate. Therefore, columnar grains which crys-tallize on the surface can continue to grow without encoun-tering the equiaxed grains.4.4.

Change of Mold Flux Accompanying Crystalliza-tion

The results of in-situ observation with the laser confocalmicroscope indicated that the surface of the specimen atfirst became smooth and then many waves appeared in amoment as shown in Figs. 8 and 9. The reason that the sur-face of the specimen became smooth is that the specimenexpanded with rise of temperature due to thermal expan-sion. When the waves appeared on the surface during heat-ing, the specimen was rapidly cooled and XRD analysiswas carried out. This confirmed that cuspidine precipitatedat this moment. Therefore, it is safely concluded that theshrinkage due to crystallization made the surface of thespecimen rough. Furthermore the rate of crystallization ofcuspidine is quite high because many waves appeared in amoment.

Formation of wavy surface due to crystallization maylead to mild cooling in the mold as pointed out in the litera-tures. 11,12)

maker , (1994), October, 55.

S. Hiraki, M. Kawamoto, T. Murakami, M. Hanao and T. Watanabe:CAMP-ISIJ , 11(1998), 185.

T. Matsumiya, T. Saeki, J. Tanaka and T. Ariyoshi: Tetsu-to-Hagané, 68(1982), 1782.

T. Saeki, S. Ooguchi, S. Mizoguchi, T. Yamamoto, H. Misumi and A.Tsuneoka: Tetsu-to-Hagané, 68(1982), 1773.

Y . Sugitani and M. Nakamura: Tetsu-to-Hagané, 65(1979), 1702.T. Mizoguchi, N. Ogibayashi and T. Kajitani: Tetsu-to-Hagané, 81(1995), 971.

H. Shimizu, R. Yamaguchi, M. Suzuki, M. Nakada, R. Nishimachi and H. Murakami: CAMP-ISIJ , 9(1996), 848.

A. Yamauchi, S. Itoyama, N. Bessho and K. Sorimachi: CAMP-ISIJ , 8(1995), 1115.

H. Sakai, K. Kawashima, T. Shiomi, H. Okamoto and K. Watanabe:CAMP-ISIJ , 8(1995), 1015.

T. Kanazawa, S. Hiraki, M. Kawamoto, K. Nakai, K. Hamazaki andT. Murakami: Tetsu-to-Hagané, 83(1997), 701.

K. Watanabe, M. Suzuki, K. Murakami, H. Kondo, A. Miyamotoand T. Shiomi: Tetsu-to-Hagané, 83(1997), 115.

A. Yamauchi, K. Sorimachi, T. Sakuraya and T. Fujii: Tetsu-to-Hagané, 79(1993), 37.

M. Kawamoto, K. Hakari and T. Watanabe: CAMP-ISIJ , 10(1997),890.

M. Kawamoto, T. Tsukaguchi, N. Nishida, T. Kanazawa and S. Hi-raki: ISIJ Int., 37(1997), 134.

W . Kurz and D. J. Fisher: Fundamentals of Solidification, Trans TechPublication, Aedermannsdorf, Switzerland, (1992), 63.

M. Yamane: Hajimete-Glass-Wo-Tukuru-Hito-No-Tameni, UchidaRokakuho Publishing Company, Tokyo, (1999), 173.

5. Conclusions

2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15) 16)

To analyze the crystallization process of mold flux fromglassy state, a mold flux, which has high tendency for crys-tallization, has been used in this study. Quenched speci-mens have been heat-treated in various conditions and char-acterized. Following conclusions have been derived.

(1)Glassy specimen heat-treated in the electric furnaceover 550°C for 180min crystallized. Cuspidine was con-firmed by XRD analysis.

(2)Grains precipitated and grew during heat treatment.Both columnar and equiaxed grains grew dendritically.

(3)The number of grains per unit area (rgrain ) andgrowth velocity of grains (V grain ) were characterized. rgrain decreased with rise of temperature of heat treatment. V grain increased and decreased after the maximum with rise oftemperature. This relationship qualitatively agrees with thetheoretical explanation for crystallization of ordinary glass.(4)In-situ observation carried out with the laser micro-scope revealed that the glassy specimen crystallized in amoment over 600°C and the surface of the specimen turnedto be rough. This may lead to mild cooling in the mold.

REFERENCES

1) C. A. Pinheiro, I. V. Smarasekera and J. K. Brimacombe: Iron Steel-

2852008ISIJ