the effective partition coefficients of microalloing
Transkript
the effective partition coefficients of microalloing
METAL 2004 Hradec nad Moravicí THE EFFECTIVE PARTITION COEFFICIENTS OF MICROALLOING ELEMENTS IN CAST STEELS EFEKTIVNÍ ROZDELOVACÍ SOUCINITELE MIKROLEGUJÍCÍCH PRVKU V LITÝCH OCELÍCH Petr Hásek a Karel Macek b a b BUT Brno, Faculty of Mechanical Engineering, IMSE, Technická 2, 616 69 Brno, CR CTU in Prague, Faculty of Mechanical Engineering, DME, Karlovo nám. 13, 121 35 Prague 2, CR Abstract The effective partition coefficients were applied to assessment of heterogeneity of distribution of microalloying elements in cast steels. Both the small castings from laboratory heats and the heavy castings from pilot plant heats were investigated. The evaluation of heterogeneity was based on the model of ideal dendrite. The effective partition coefficients and the indexes of heterogeneity were derived from concentration profiles of aluminium, titanium and vanadium. The effect of mean concentration of the above mentioned chemical elements on their average partition coefficient was also studied. Close correlation between partition coefficients and indexes of heterogeneity was confirmed. Small differences in characteristics for laboratory heats or pilot plant heat were determined. Abstrakt Efektivní rozdelovací soucinitele byly použity pro hodnocení homogenity rozložení legujících prvku v mikrolegovaných litých ocelích. Pri experimentu byly hodnoceny jednak malé vzorky laboratorních taveb a jednak masivnejší vzorky z poloprovozních taveb. Pro posouzení heterogenity byla použita metodika vycházející z modelu ideálního dendritu. Z koncentracních profilu titanu, hliníku a vanadu byly stanoveny prubehy jejich efektivních rozdelovacích soucinitelu a indexy heterogenity. Byl studován vliv prumerné koncentrace legujících prvku na velikost jejich prumerného rozdelovacího soucinitele. Experimenty potvrdily velmi tesnou závislost indexu heterogenity na efektivním rozdelovacím souciniteli. Výsledky hodnocení laboratorních a poloprovozních taveb vykazují pouze malé rozdíly. 1. INTRODUCTION The crystallization of steel is accompanied by segregation processes which cause various concentration of any chemical element on different spots in the casting. This is due to the fact that the solute element i, whether present as alloying element or impurity, is more soluble in the liquid phase l than in the solid phase s. At definite temperature the relation between the involved concentrations ci defines the partition coefficient ki = cis/cil. Because the liquid becomes progressively richer in the solute as crystallization proceeds, the solute concentrations in casting tend to rise in the areas that solidify last, i.e. in the centre of the casting. This long-range concentration variations fall in the classification of macrosegregation. In contrast to it, localized concentration variations on a scale smaller than the crystal size are called microsegregation; it is caused by dendritic solidification in alloys and therefore also called dendrite segregation. Segregation is an undesirable process that may lead to complete impairment of the casting owing to foundry defects and to non-effective heat treatment. 1 METAL 2004 Hradec nad Moravicí An original method for assessment of dendritic segregation has been submitted in the work ?1?. The concept of this work is based on the model of ideal dendrite that is an imaginary dendrite characterizing the population of all actual dendrites in the casting. The course of concentration ci of any chemical element i can be composed from a large number of local concentrations of this element as measured in a sample by means of their rowing into nondecreasing series. If the partition coefficient k i is less than unity, then the minimum measured value of concentration cimin corresponds to the centre- line of the ideal dendrite, whereas the maximum concentration cimax corresponds to its last frozen interface. These critical locations are then labelled as gs = 0 (the beginning of solidification) and gs = 1 (the end of solidification), where gs represents the fraction of the solidified phase. According to the above mentioned concentration profile, it is possible to assess the course of effective partition coefficient during the solidification. The latter has been therefore applied to the evaluation of the degree of microsegregation in our castings. Chemical homogeneity is of the utmost practical importance in cast low carbon microalloyed steels since their yield strength and guaranteed weldability through carbon equivalent are dependent on uniform distribution of small amounts of alloying elements. Niobium, titanium and vanadium are commonly used in concentrations of hundredth percent by weight. Aluminium also plays an important role, however, it is not usually considered as a typical microalloyer. Sufficient level of microalloying is necessary in order to promote precipitation strengthening and to control the grain growth. On the other hand, toughness is reduced if microalloying is excessive. It follows that the optimisation of valid chemical composition in this class of steels is relevant. The aim of this contribution is to evaluate dendritic segregation of aluminium, titanium and vanadium in cast microalloyed steels and work out an appraisal of optimum concentration of these elements with respect to their uniform distribution. 2. EXPERIMENTAL MATERIAL Chemical composition of chosen experimental steels is given in tab. 1. Laboratory heats of steels 15MnTi4 and 14MnNb52 were cast into small ingots having mass aprox. 1,5 kg. After solidification these ingots were normalized at 900 °C for 3 h and cooled with rate 100 °C/h. Pilot-plant heats of the other steels were cast into blocks with dimensions 400 x 400 x 250 mm. The blocks were cut into prismatic balks 250 x 100 x 100 mm and homogenized at 1050 °C for 8 h, cooled down 100 °C/h and then normalized like ingots of laboratory heats. Tabulka 1. Chemické složení sledovaných taveb stanovené chemickou analýzou [hm.%] Steel 15MnTi4 14MnNb52 27MnTiV4 12MnTi4 26MnTi4 14MnTi8 28Mn8 C 0,15 0,14 0,27 0,12 0,26 0,14 0,28 Mn 1,11 1,16 1,20 1,12 1,12 2,00 2,00 Si V Nb 0,36 0,42 0,19 0,30 0,13 0,28 0,27 0,28 0,29 - Ti 0,035 0,003 0,017 0,025 - Al 0,042 0,070 0,003 0,005 0,006 0,002 0,002 S 0,009 0,012 0,013 0,009 0,010 0,010 0,010 P 0,018 0,019 0,020 0,016 0,018 0,017 0,019 Table 1. Chemical composition of investigated heats measured by chemical analysis [wt.%] 2 METAL 2004 Hradec nad Moravicí 3. METHODS OF MEASUREM ENT AND COMPUTATION Measurements of local concentrations of chemical elements were conducted by wave dispersion microanalysis by means of apparatus CAMEBAX-MICRO. Lineal method was applied when the electron beam was directed perpendicularly to the prime arms of dendrites. The number of analysed spots n varied from 30 to 50, as well as spacing between neighbouring spots that ranged from 1 to 9 ? m. The programme of quantitative analysis ran with simultaneous correction of measured values regarding atomic number Z, absorption A and fluorescence F. Energy-dispersion analyser of characteristic X-ray radiation EDAX was used in connection with analytical electron microscope JEOL CM200. Local concentrations c of definite chemical elements were sequentially rowed into nondecreasing series, where the i-th number of the progression was labelled as c(i). In accordance with the procedure introduced in ?1?, concentrations c(i) were converted into the sequential effective partition coefficient k ef(i) using the formula (1): c( i) k ef (i ) ? (1) n 1 ?? c j n ? i ? 1 j? i Arithmetic mean of all values k ef(i) was called as effective partition coefficient k ef. Heterogeneity index was determined as a quotient of standard deviation sn-1 of the population and the mean concentration cm : s I H ? n? 1 (2) cm Tabulka 2. Strední koncentrace cm vypoctené z merení mikrosondou [hm.%], efektivní rozdelovací soucinitele k ef a jejich smerodatné odchylky sn-1 (cm ) a sn-1 (k ef) Steel cm 15MnTi4 sn-1(cm) kef sn-1(kef ) cm 27MnTiV4 sn-1(cm) kef sn-1(kef ) cm 26MnTi4 sn-1(cm) kef sn-1(kef ) cm sn-1(cm) 28Mn8 kef sn-1(kef ) V Al Ti Steel 0,041 0,02 cm ±0,041 ±0,015 14MnNb52 sn-1(cm) 0,416 0,528 kef ±0,358 ±0,267 sn-1(kef ) 0,194 0,004 0,068 cm ±0,039 ±0,007 ±0,022 12MnTi4 sn-1(cm) 0,841 0,232 0,778 kef ±0,078 ±0,336 ±0,193 sn-1(kef ) 0,017 0,003 0,013 cm ±0,014 ±0,004 ±0,013 14MnTi8 sn-1(cm) 0,487 0,262 0,394 kef ±0,360 ±0,380 ±0,268 sn-1(kef ) 0,012 0,004 0,004 ±0,013 ±0,006 ±0,007 0,363 0,25 0,241 ±0,325 ±0,368 ±0,287 V Al Ti 0,108 0,006 ±0,030 ±0,010 0,791 0,222 ±0,116 ±0,324 0,015 0,005 0,014 ±0,015 ±0,008 ±0,016 0,404 0,228 0,354 ±0,320 ±0,286 ±0,285 0,009 0,003 0,088 ±0,013 ±0,004 ±0,016 0,282 0,248 0,875 ±0,359 ±0,330 ±0,108 Table 2. Mean concentrations cm calculated from microanalyzer measurement [wt.%], effective partition coefficients k ef and their standard deviations sn-1 (cm ) or sn-1 (k ef) 4. RESULTS AND DISCUSSION 3 METAL 2004 Hradec nad Moravicí 1 0,8 0,6 0,4 Ti 0,2 0 0 0,2 0,4 0,6 0,8 1 S. e. p. coefficient kef(i) [-] Solidified phase part Podíl ztuhlé fáze gs g[-]s [-] S. e. p. coefficient kef(i) [-] S. e. p. coefficient kef(i) [-] Heterogenity index I H [-] Heterogeneity index IH [-] In tab.2, there are collected mean concentrations, effective partition coefficients, and their standard deviations of investigated chemical elements that were computed as arithmetic means from experimental data obtained by microanalyzer measurements. By comparing the Ti mean concentrations 2,0 V given in tab.1 and Al tab.2, one can see 1,5 15MnTi4 fairly good agreement 14MnNb52 between 1,0 27MnTiV4 corresponding values IH = -1,087.ln(k ef ) + 0,0306 2 12MnTi4 within standard R = 0,9983 0,5 26MnTi4 deviations. There are 14MnTi8 only two exceptions 0,0 28Mn8 concerning the content 0,1 0,2 0,3 0,4 0,5 0,6 0,8 1 0,7 0,9 of aluminium in steel Effective partition coefficient kef [-] 14MnNb52 and vanadium in steel Obr. 1. Korelace indexu heterogenity IH a efektivního rozdelovacího 27MnTiV4, which soucinitele k ef might be ascribed to macrosegregation in Fig. 1. Correlation between heterogeneity index IH and effective the ingot. partition coefficient k ef Fig.1 represents the relation of heterogeneity index and effective partition coefficient. Drawing symbols distinguish both the chemical elements (by shape of symbol) and steels (by colour of symbol). Even at first glance one can see very tight correlation between the two characteristics of microsegregation. It also follows that there are no relevant differences between laboratory heats and pilot-plant heats. 1 0,8 0,6 0,4 V 0,2 0 0 0,2 0,4 0,6 0,8 Solidified phase part Podíl ztuhlé fáze gs g[-]s [-] 1 0,8 Obr. 2. Závislost sekvencního efektivního rozdelovacího soucinitele k ef(i) na podílu ztuhlé fáze gs 0,6 Al 0,4 0,2 0 0 0,2 0,4 0,6 0,8 1 Solidified Podíl ztuhlé phase fáze part gs g[-]s [-] 4 Fig. 2. Dependence of sequential effective partition coefficient k ef(i) on solidified phase part gs 1 METAL 2004 Hradec nad Moravicí 1,0 S. e. p. coefficient kef(i) [-] However, it is worth to note, that effective partition coefficients, Increasing mean [-] 0,8 complying the applied method of Ro concentration of zde computation and heat treatment of chem. (i) lov ací 0,6 samples, involve not only element [1]so uci ef nit processes running during k el k0,4 d solidification, but also c a concentration changes related to b 0,2 cooling of the casting and its heat treatment. Smaller ingots of 0,0 0 0,2 0,4 0,6 0,8 1 laboratory heats should exhibit g s fáze [1]g [-] Podíl ztuhlé lower heterogeneity than massive Solidified phase part gs [-] blocks of pilot-plant heats, but the Obr. 3. Závislost sekvencního efektivního were subjected to rozdelovacího soucinitele k ef(i) na podílu ztuhlé fáze gs – latter homogenisation treatment and the obecné schéma former did not. Fig.2 shows the dependences Fig. 3. Dependence of sequential effective partition of sequential effective partition coefficient k ef(i) on solidified phase part gs – generalized coefficient of microalloying scheme element on the fraction of solidified phase. The dependences differ mutually for every single steel. If we now compare the course of these dependences with mean concentration of chemical element concerned (tab.2), we obtain ge neralized scheme for influence of the latter magnitude on discussed relationship (fig.3). The effect of concentration on partition coefficient of any chemical element is able to contribute to explanation of unusual scatter for values of these coefficients reported in literature, e.g. 0,12 ? 0,92 for Al and 0,05 ? 0,62 for Ti ?2??3?. From fig.2 also follows different rate of increase of chemical homogeneity (uniformity of distribution) with increasing content of microalloying element. It can be better seen on fig.4, where effective partition coefficients are plotted against mean concentrations of elements detected by microanalyzer. Drawing 1,0 symbols used are 0,9 identical with those in 0,8 fig.1. For titanium and vanadium the curve 0,7 seems to have logarithmic 0,6 character, whereas 0,5 straight line approximates 0,4 the dependence for 0,3 aluminium. The level of 0,2 segregation of chemical 0,1 element must be assessed 0,0 not only with respect to 0,00 0,05 0,10 0,15 0,20 value of its effective Mean concentration cm [wt.%] partition coefficient, but also its mean Obr. 4. Závislost efektivního rozdelovacího soucinitele k ef concentration shell be na strední koncentraci prvku cm taken into account. If we row investigated elements Fig. 4. Dependence of effective partition coefficient k ef on according to their elements mean concentration cm e x Ef. partition coefficient k ef [-] s 5 METAL 2004 Hradec nad Moravicí concentration needed for attaining k ef = 0,5, we shall obtain lower mean concentration for titanium and vanadium as compared to aluminium. For optimisation of content of microalloying elements in cast low carbon steels from the point of view of their uniform distribution, it would be necessary to define some critical value of effective partition coefficient, e.g. 0,95. Our present results (fig.4) do not permit to apply such approach. 5. CONCLUSIONS Presented assessment of dendritic segregation of selected chemical elements in cast microalloyed steels based the model of ideal dendrite resulted in following conclusions: 1. Laboratory heats as compared to pilot-plant heats do not exhibit relevant differences in chemical heterogeneity of investigated elements. The reason for this unexpected experimental fact is probably concerned with opposite effect of solidification conditions and heat treatment. 2. Very tight correlation between effective partition coefficients and heterogeneity indexes was confirmed for all steels investigated. 3. Previously suggested generalized scheme for effect of mean concentration of chemical element in steel on its sequential effective partition coefficient was approved. 4. Titanium and vanadium attain the value 0,5 of effective partition coefficient at lower mean concentration as compared to aluminium. 5. Our present results are insufficient for optimising the contents of microalloying elements with respect to dendritic segregation. Acknowledgement Authors gratefully appreciate the financial support provided by Grant Agency of the Czech Republic to project No.106/03/0473. REFERENCES: [1] DOBROVSKÁ,J., DOBROVSKÁ,V., MILLION,B., STRÁNSKÝ,K.: Estimation of the partition coefficients of elements from the distribution curves their dendritic segregation (theory). In: Proceedings Diffusion and thermodynamics of materials. Brno: IPM AS CR, 1998, p. 25. [2] LEVÍCEK,P., STRÁNSKÝ,K.: Metalurgické vady ocelových odlitku. Prague: SNTL, 1984. [3] DÁPALA,J., KUCHAR,L.: Solidus and liquidus curves and distribution coefficients of admixtures in iron and prediction of the solidification interval in low-alloyed steels. Hutnické listy, 2000, vol. LV, is. 4-7, p. 61. 6