- Email: [email protected]
Phase equilibria in the Al–Si–V system: The vanadium rich part
Intermetallics 19 (2011) 369e375
Contents lists available at ScienceDirect
Intermetallics journal homepage: www.elsevier.com/locate/intermet
Phase equilibria in the AleSieV system: The vanadium rich part Beatrix Huber, Klaus W. Richter* University of Vienna, Department of Inorganic Chemistry/Materials Chemistry, Währinger straße 42, 1090 Wien, Austria
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 September 2010 Received in revised form 21 October 2010 Accepted 22 October 2010
The V-rich part of the AleSieV phase diagram was determined by a combination of optical microscopy, powder X-ray diffraction (XRD), differential thermal analysis (DTA) and electron probe microanalysis (EPMA). Phase equilibria were investigated at two isothermal sections at 850 and 1300 C. High temperature DTA was performed to identify the ternary invariant reactions yielding a ternary reaction scheme and the vertical section at 50 at.% V. As cast samples were investigated in order to gain additional information about primary crystallization fields. A liquidus surface projection was constructed for the entire ternary system by combining our experimental data with those from literature. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: A. Aluminides miscellaneous B. Phase diagram B. Phase transformations
1. Introduction and literature review A detailed description of the ternary phase equilibria in the Vpoor region of the AleSieV phase diagram (i.e. between 0 and 50 at.% V) was given recently by Huber et al. [1]. The current investigation is a continuation of our previous studies in the Al-rich part of the ternary AleSieV system and covers the V-rich part. Phase equilibria at 850 and 1300 C will be presented here as well as the vertical section at 50 at.% V. An entire liquidus surface projection and a reaction scheme are given. The binary sub systems AleV, SieV and AleSi have been investigated repeatedly and are given in Massalskis Handbook of Binary Alloy Phase Diagrams [2]. A detailed literature survey is given in [1], based on the work of Richter and Ipser [3] and Gong [4] for AleV and of Jorda and Muller [5], Savickij [6] and Zhang [7] for SieV. Recently, Huber and Richter [8] discovered a low temperature phase in the binary AleV system with the composition Al50.1(5)V46.9(5). There is indication that the phase crystallizes in a tetragonal structure (possible space group P42212), but so far it was not possible to determine the structure. The invariant binary reactions from literature relevant to the current work are listed in Table 1 together with the respective references. An early assessment of the AleSieV system was given by Haußmann [9] and is based primarily on the experimental investigation of Gebhardt and Josef [10] and Müller [11]. Müller’s [11] investigations in the V-rich part of the ternary system concern the superconducting binary phase SiV3 and its ternary solid * Corresponding author. E-mail address: [email protected] (K.W. Richter). 0966-9795/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.intermet.2010.10.022
solution Al1xSixV3. A partial isothermal section at 1000 C shows the maximal solubility of Al in SiV3 to be 3.5 at. % and an equilibrium of the three phases SiV3 þ Si3V5 þ (V). Our recently published study [1] includes three isothermal sections at 500 C and 620 C for alloys up to 20 at.% V and at 850 C for alloys with more than 20 at.% V. The results of this study differ considerably from those of Gebhardt and Josef concerning the solubility of the binary phases in the ternary system. Apart from that the ternary phase Al0.6Si1.4V was discovered and its crystal structure was determined (Pearson symbol oF24, space group Fddd, TiSi2-type). In Table 2 the invariant ternary reactions reported in our previous study [1] are listed. 2. Experimental The samples were prepared from aluminum slug (99.999%, Alfa Aesar, Karlsruhe, Germany), silicon lump (99.9999%, Alfa Aesar, Karlsruhe, Germany) and vanadium pieces (99.7%). To remove oxides at the vanadium surface it was treated with diluted hydrochloric acid for some minutes in the ultrasonic bath. After rinsing it with water and acetone it was dried for some minutes at 120 C. Calculated amounts of the elements were weighed to an accuracy of 0.05 mg and arc melted on a water-cooled copper plate under an argon atmosphere. Zirconium was used as a getter material within the arc chamber. The reguli with a total mass of about 1000 mg were remelted four times for good homogenization and weighted back after each melting step to control the mass losses. The total mass losses were less than 1% and can be considered to be negligible. A series of samples were placed in alumina crucibles, which were sealed into evacuated quartz glass ampoules and then annealed at 850 C for three weeks. Putting the samples into alumina
370
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
Table 1 Binary invariant phase equilibria relevant to the present study. Phase reaction
Composition of the involved phases/at.%
T/ C
Ref.
L þ (V) ¼ Al8V5 L ¼ Si2V þ Si5V6 L þ Si3V5 ¼ Si5V6 L ¼ SiV3 þ (V) L ¼ SiV3 þ Si3V5 L ¼ SiV3 L ¼ Si3V5 L ¼ Si2V Si5V6 ¼ Si2V þ Si3V5
w33 V w59 Si w57 Si w13 Si w29 Si w25 Si w37.5 Si 66.7 Si e
1408 1640 1670 1870 1895 1925 2010 1677 w460
[3] [14] [14] [2] [2] [2] [14] [2] [15]
w50 V 66.7 Si w37.5 Si w7 Si w25.5 Si w25 Si w37.5 Si 66.7 Si e
w38.5 V 45.4 Si w45.4 Si w19 Si w37.5 Si e e e e
crucibles was necessary to avoid reaction of the Al containing samples with the quartz glass during the heat treatment. After the annealing time the samples were quenched in cold water. Another series of samples were sealed into evacuated Ta-crucibles, annealed at 1300 C for one week in a permanent Ar-flow and then quenched in a water-cooled container. All samples were separated into several pieces, which were investigated by X-ray powder diffraction (XRD), electron probe microanalysis (EPMA) and metallographical analysis. Initial sample characterization was performed by powder XRD using a Bruker D8 Advance Diffractometer operating in reflection mode (Cu Ka radiation, Lynxeye silicon strip detector) and the patterns were analyzed using the TOPAS software [12]. To gain information about ternary phase reactions selected samples were measured by DTA and high temperature DTA, respectively. Polished pieces of the annealed samples were investigated by optical microscopy using a Zeiss Axiotech 100 microscope equipped for operation under polarized light. Selected samples were analyzed by EPMA in order to determine reliable phase compositions. The EPMA measurements were carried out on a Cameca SX electron probe 100 (Cameca, Courbevoie, France) using wavelength dispersive spectroscopy (WDS) for quantitative analysis. Pure aluminum, silicon and vanadium were employed as standard materials. The measurements were carried out at 15 kV using a beam current of 20 nA. Conventional ZAF matrix correction was used to calculate the final composition from the measured X-ray intensities. To examine as cast samples, SEM measurements were carried out on a Zeiss Supra 55 VP equipped with an EDX detector. DTA measurements up to a maximum temperature of 1490 C were performed on a DTA 404 PC (Netzsch, Selb, Germany) using open alumina crucibles and employing a slow permanent argon flow. A sample mass of approximately 150 mg was used for the experiments and the samples were checked routinely for possible mass changes during the DTA investigations. No relevant mass changes were observed. Two heating- and cooling-curves were recorded for each sample using a heating rate of 5 K min1. The Pt/Pt10%Rh thermocouples (type S) of the DTA instrument were calibrated at the melting points of pure Al, Au and Ni. High temperature DTA measurements (above 1490 C) were carried out on a Setaram Setsys Evolution 2400 using covered alumina crucibles within a slow permanent argon flow. A sample mass of Table 2 Ternary invariant phase reactions in AleSieV reported earlier [1]. Ternary invariant reaction
T/ C
E1: L ¼ (Al) þ (Si) þ Si2V U1: L þ Al45V7 ¼ (Al) þ Si2V U2: L þ Al21V2 ¼ (Al) þ Al45V7 U3: L þ Al3V ¼ Al23V4 þ Al45V7 P1: L þ Al3V þ Si2V ¼ Al45V7 U4: Lþs ¼ Al3V þ Si2V U5: L þ Si3V5 ¼ Al3Vþs P2: L þ Si2V þ Si3V5 ¼ s U6: L þ Al8V5 ¼ Al3V þ Si3V5
577 643 w664 733 745 866 976 1140 1224
Fig. 1. Isothermal section of the AleSieV system at 850 C. Diamonds: nominal composition of the samples investigated by EPMA. The V-poor part is drawn according to [1] (bright gray).
approximately 100 mg was used for the experiments and no relevant mass changes before and after the measurements were observed. One heating-curve was recorded for each sample using a heating rate of 5 K min1. The W5%Re/W26%Re thermocouple (type W5) was calibrated at the melting points of pure Au, Ni and Pt. 3. Results and discussion 3.1. Isothermal sections at 850 and 1300 C The isothermal section at 850 C constructed based on XRD and EPMA measurements, is shown in Fig. 1. In the V-rich part of the Table 3 Phase composition in various samples annealed at 850 and 1300 C determined by EPMA. Phase field Nominal sample T/ C of annealing composition Al15Si23V32
850
[Al8V5 þ SiV3 þ Si3V5]
Al70Si10V20
850
[Al8V5 þ SiV3 þ (V)]
Al30Si5V65
850
[SiV3 þ (V)]
Al10Si10V80
850
[SiV3 þ (V)]
Al5Si10V85
850
[SiV3 þ (V)]
Al40Si10V50
1300
[Al8V5 þ Si3V5 þ (V)]
Al10Si25V65
1300
[SiV3 þ Si3V5 þ (V)]
Al24Si8V68
1300
[SiV3 þ (V)]
Al21Si7V72
1300
[SiV3 þ (V)]
Al10Si10V80
1300
[SiV3 þ (V)]
– not measured due to fine microstructure.
Phase composition by EPMA/at.%
Al8V5 SiV3 Si3V5 Al8V5 SiV3 (V) SiV3 (V) SiV3 (V) SiV3 (V) Al8V5 Si3V5 (V) SiV3 Si3V5 (V) SiV3 (V) SiV3 (V) SiV3 (V)
Al
Si
V
55.3(4) 3.9(4) 1.8(1) 55.9(2) 6.6(3) 44.9(3) e e e e e e 56.4(5) 3.5(1) 45.2(4) 4.7(1) 4.1(1) 42.2(3) 6.2(1) 32.0(1) 6.1(6) 24.6(1) 5.3(4) 13.5(1)
1.7(1) 23.0(1) 35.2(1) 1.2(1) 20.0(8) 0.3(1) e e e e e e 2.0(2) 33.7(1) 1.6(2) 21.5(2) 32.8(4) 1.4(2) 19.1(1) 1.7(1) 18.4(9) 1.8(1) 17.8(9) 2.1(1)
43.0(5) 73.1(5) 63.0(1) 42.9(2) 73.4(3) 54.8(2) e e e e e e 41.6(4) 62.8(2) 53.2(7) 73.8(1) 63.1(1) 56.4(2) 74.7(2) 66.3(1) 75.5(8) 73.6(1) 76.9(7) 84.4(1)
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
371
Fig. 2. Determination of the Al-concentration in (V) for two-phase samples using lattice parameter interpolation according to Vegard’s law.
AleSieV phase diagram at 850 C we observed the three phase fields [Al8V5 þ SiV3 þ Si3V5] and [Al8V5 þ SiV3 þ (V)] and the large two phase region of [SiV3 þ (V)]. The corresponding compositions in the different phase fields are given in Table 3. Earlier results in the Al-richer part are drawn brighter and the exact compositions are listed in [1]. As reported previously [1], beside tetragonal Si3V5 we found traces (1e5%) of hexagonal Si3V5 (Mn5Si3-type) in the samples and again we assume that oxygen impurities are responsible for the occurrence of this phase. The maximal solubility of Al in SiV3 is 6.6 at.% while Si3V5 dissolves only 1.8 at.% Al. The maximal solubility of Si in Al8V5 was measured to be 1.7 at.%. Samples located in the two phase field [SiV3 þ (V)] exhibit a very fine microstructure which makes the measurement of the composition of the respective phases by means of EPMA impossible. In XRD-measurements the presence of the phases SiV3 and (V) were measured unambiguously and the cell parameters of the phases were deterimend by Rietveld refinement. Vegard’s law was applied to get information about the composition of the solid solution of Al in (V) in order to fix the tie lines shown in Fig.1. For this purpose the cell parameters of pure vanadium [13] and of Al45V55 (sample Al30Si10V60) were used for linear interpolation. Since the solubility of Si in (V) is very small the variation of the cell parameter can be attributed only to the substitution of the V-atoms by Al-atoms. In Fig. 2 the corresponding graph can be seen. In the binary SieV system Savickij et al. [6] measured the solubility of SiV3 to be 24.3e24.7 at.% Si at 800 C. Recently Zhang
Fig. 3. Isothermal section of the AleSieV system at 1300 C. Diamonds: nominal composition of the samples investigated by EPMA.
et al. [7] gave a thermodynamic reassessment of the entire SieV system and calculated the maximum solubility of SiV3 to be 20e25 at.% Si at 1837 C. At a temperature of 900 C the solubility decreases completely. Our results indicate that also the binary phase SiV3 exhibits certain extension at 850 C, and not to be so narrow like Zhang et al. [7] calculated in their assessment. So we used the phase boundaries given by Savickij et al. [6] in our sections at 850 C and 1300 C as binary reference. The results of our measurements at 1300 C are shown in Fig. 3 and the corresponding compositions determined by means of EPMA are also listed in Table 3. In contrast to the isothermal section at 850 C one can see, that the three phase fields [Al8V5 þ Si3V5 þ (V)] and [SiV3 þ Si3V5 þ (V)] exist. This result is in accordance with the work of Müller [11], who determined the Vrich part of the ternary AleSieV system at 1000 C. Annealing experiments were done to confirm that our results at 850 and at 1300 C are equilibrium situations and that a solidesolid transition reaction Si3V5 þ (V) ¼ Al8V5 þ SiV3 takes place above 850 C. On the one hand the sample Al15Si23V62 annealed at 850 C exhibiting the phases Al8V5, SiV3 and Si3V5 was annealed at 1180 C for three weeks. Subsequent EPMA measurements proved
Fig. 4. Back scattered electron image of the samples Al15Si23V62 and Al40Si10V50 annealed at 850 and 950 C, respectively.
372
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375 Table 4 Ternary invariant phase reactions. Reaction
T/ C
Phase Composition/at.%
U7: L þ (V) ¼ Al8V5 þ Si3V5
1390
L (V) Al8V5 Si3V5
w54. 0 w8.0 44.5 0.5 55.6 1.4 1.7 35.3
w38.0 55.0 43.0 63.0
U8: L þ SiV3 ¼ Si3V5 þ (V)
1574
L SiV3 Si3V5 (V)
w35.0 5.0 1.7 44.5
w10.0 22.0 35.5 0.5
w55.0 73.0 63.0 55.0
U9: L þ Si5V6 ¼ Si2V þ Si3V5
1635
L Si5V6 Si2V Si3V5
w9.0 0.0 0.2 0.5
w54.0 45.5 65.8 37.5
w37.0 54.5 34.0 62.5
55.5 45.0 4.0 0.5
1.5 0.5 23.0 62.5
43.0 54.5 73.0 37.5
Al
Fig. 5. Vertical section in AleSieV at 50 at.% V. Circles: invariant thermal effects; Small circles: other thermal effects.
U: Si3V5 þ (V) ¼ Al8V5 þ SiV3 850 < T < 950 Al8V5 annealing experiments; (V) not measured by DTA SiV3 Si3V5
Fig. 6. Partial ternary reaction scheme (Scheil Diagram) for AleSieV.
Si
V
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
373
Fig. 7. Liquidus surface projection of AleSieV including fields of primary crystallization. Solid lines: liquidus valleys; dotted lines: isotherms; circles: invariant points; triangles: investigated as cast samples.
that the three phases SiV3, Si3V5 and (V) were formed. On the other hand a sample with the same overall composition annealed at 1300 C and therefore exhibiting the phases SiV3, Si3V5 and (V) in large grains, was annealed at 850 C for seven weeks. XRDmeasurements of this sample however did still show the same phases. With the assumption that the respective reaction Si3V5 þ (V) ¼ Al8V5 þ SiV3 is hindered kinetically due to long diffusion distances in the course microstructure of the sample, the same sample was powdered, pressed to a pellet and again annealed at 850 C for five weeks. After this procedure the phase Al8V5 was formed and (V) was not longer found. In an attempt to further specify the temperature of the solid-state transition reaction, the sample Al40Si10V50 was annealed at 950 C directly after arc melting. It showed the phases Al8V5, SiV3 and (V), indicating the equilibrium above the transition reaction. The two situations after annealing at 850 C and 950 C are shown in Fig. 4. By means of DTA measurements we tried to determine the exact reaction temperature but since solidstate reactions generally show only small heat effects and the transition is presumably very slow, this was not possible. Considering our measurements and the results given in [11] we conclude that the reaction takes place between 850 and 950 C.
3.2. Ternary phase reactions and liquidus surface projection The ternary phase reactions were studied by DTA at medium and high temperature. For regular DTA measurements two heating- and cooling- cycles were performed in each measurement in order to test if equilibrium conditions can be reestablished within the samples at the selected heating/cooling rate of 5 K min1. Starting temperature was 850 C i.e. the respective annealing temperature. The maximum temperature reached in these experiments was 1490 C. Metastable thermal effects were observed in most samples in the second heating cycles. The experimental data points for thermal effects shown in Fig. 5 thus represent mainly DTA results obtained on first heating of the annealed samples. In the vertical section at 50 at.% V four invariant reactions were measured. Two transition reactions at 976 and 1224 C and one peritectic reaction at 1140 C were already established earlier (Table 2). The fourth invariant reaction at 1390 C was interpreted to be the transition reaction: L þ (V) ¼ Al8V5 þ Si3V5. In DTA measurements of the sample Al30Si20V50 we found only one invariant effect at 1390 C. The effect corresponding to the invariant reaction U6 [L þ Al8V5 ¼ Al3V þ Si3V5] at 1224 C is missing, although it should be present according to our interpretation. Our XRD results show only
374
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
Table 5 Primary crystallization investigated from as cast samples. Nominal sample composition
Primary crystallization
Al75Si5V20 Al76Si14V10 Al1.5Si44.7V53.8 Al2.5Si44.5V53 Al15Si35V50 Al47.5Si2.5V50 Al30Si10V60 Al15Si23V62 Al30Si15V55 Al40Si5V55 Al10Si10V80 Al21Si7V72 Al28Si42V30 Al22Si48V30 Al5Si56V39 Al43Si7V50 Al38Si12V50 Al37Si10V53 Al32Si11V57 Al21Si17V62 Al7Si13V80
Si3V5
s
Si3V5 Si3V5 Si3V5 (V) SiV3 Si3V5 Si3V5 (V) (V) SiV3 Si3V5 Si3V5 Si5V6 (V) Si3V5 Si3V5/(V) SiV3 SiV3 SiV3
a small amount (approximately 8%) of the phase Al3V in the sample, so we think that the thermal effect of U6 was too small to be detected in this sample. The reaction U6 was measured in different samples in the vertical sections at 20 and 35 at.% V, which can be seen in [1] as well as in the sample Al20Si30V50 (see Fig. 5), so the existence of the reaction is well established. High temperature DTA was used to measure the temperatures of invariant phase reactions higher than 1490 C. For this purpose samples in the respective three phase fields [SiV3 þ Si3V5 þ (V)] (Al5Si23V62) and [Si2V þ Si3V5 þ Si5V6] (Al0.7Si44.3V55) were heated in high temperature DTA with a heating rate of 5 K min1. Starting temperature was 1400 C. In order to avoid a possible reaction of the liquid with the alumina crucibles, the measurements were
stopped after the first visible effect and not heated further to the liquidus temperatures. These measurements allowed to determine the invariant reactions in the V-rich part at elevated temperatures. A summary of these invariant four-phase equilibria including the reaction temperatures and the approximate compositions of the involved phases, which were estimated based on a combination of all available experimental data from XRD, EPMA and DTA, is given in Table 4. In Fig. 6 the reaction scheme (Scheil diagram) in the Vrich part of the system is shown. The numbering of the invariant reaction given in Huber et al. [1] is continued. In Fig. 7 the entire liquidus surface projection is shown. The Alrich part of the liquidus projection and also the isotherms from 600 to 1400 C can be seen in detail in [1]. In the current figure the isotherms from 1400 to 2000 C are shown as dotted lines. It should be emphasized, that the isotherms are only estimates, as no liquidus temperatures above 1490 C were determined. The isotherms reflect information on the binary melting temperatures and information on invariant ternary reaction temperatures. The liquidus valleys that divide the different fields of primary crystallization are shown as solid lines with arrows indicating the direction toward lower temperature. The compositions of the liquid phase in the various ternary invariant reactions listed in Table 2 and Table 4 are shown as circles in Fig. 7. In order to reach additional information about the compositions of the liquid phase taking part in the invariant four phase reactions and to construct the liquidus surface projection, as cast samples at various compositions were investigated for primary crystallization. and the sample compositions are shown in Fig. 7 as triangles. An overview of all investigated as cast samples is given in Table 5 together with information about primary crystallization. For illustration BSE images of samples in four different primary crystallization fields are shown in Fig. 8. The liquidus surface projection is dominated by the huge primary crystallization field of Si3V5. In Fig. 8a, a sample is shown near the liquidus valley which leads from P2 to U9. The white crystals were measured to be Si3V5. In Fig. 8b the white grains of Si5V6 are the
Fig. 8. Back scattered electron image of as cast samples with the nominal compositions a: Al22Si48V30, b: Al5Si56V39, c: Al43Si7V50 and d: Al32Si11V57.
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
primary crystallized phase and the narrow crystallization field of this high temperature phase is represented by the sample Al5Si56V39. Fig. 8c shows a sample in the primary crystallization field of (V). The as cast sample Al32Si11V57 shown in Fig. 8d exhibits SiV3 as primary crystals. Together with the as cast sample Al37Si19V53 located directly on the liquidus valley between U7 and U8 it offers valuable information about the composition of the transition reaction U8. Accidentally, the composition of the liquid phase taking part in the peritectic reaction P2 at 1140 C was listed incorrectly in [1]. The composition was determined to be Al73Si16V11 and not Al74Si14V12 as indicated in [1]. Furthermore minor differences in the isopleths at 20 and 35 at.% V given in [1] are the results of the current study. Summing up, the current research results concerning the V-rich part of the ternary AleSieV phase diagram fit very well to the Alrich part published earlier [1] and the entire phase diagram is described by the combination of both works. Acknowledgements Financial support from the Austrian science foundation (FWF) under the project number P19305 is gratefully acknowledged. The
375
authors wish to thank Stephan Puchegger, Adela Zemanova and Ales Kroupa for their help in SEM measurements and Mario Ecker for experimental support.
References [1] Huber B, Effenberger HS, Richter KW. Intermetallics 2010;18:606. [2] Massalski TB, Okamoto H, Subramanian PR, Kacprzak L. Binary alloy phase diagrams. ASM; 1990. [3] Richter KW, Ipser H. Zeitschrift fue Metallkunde 2000;91:383. [4] Gong W, Du Y, Huang B, Schmid-Fetzer R, Zhang C, Xu H. Zeitschrift fuer Metallkunde 2004;95:978. [5] Jorda JL, Muller J. J Less-common Met 1982;84:39. [6] Savitskii EM, Efimov YV, Eichler K, Paufler P. Wissenschaftl Z d Univ Dresden 1978;27:675. [7] Zhang C, Du Y, Xiong W, Xu H, Nash P, Ouyang Y, Hu R. CALPHAD 2008;32:320. [8] Huber B, Richter KW. J Alloys Compd 2009;493:L33. [9] Schmid-Fetzer R. Vol. In: Petzow GaE G, editor. Aluminum-Cobalt-Silicon. Weinheim: VCH; 1991. [10] Gebhardt E, Joseph G. Zeitschrift Fur Metallkunde 1961;52:310. [11] Müller. Z. Naturforsch. 1969;24a:1134. [12] TOPAS. Version 3.0. Karlsruhe, Germany: Brucker AXS Inc.; 1999. [13] Villars P, L-DC. 1991;1e4. [14] Parthe E, Nowotny H, Schmid H. Monatsh Chem 1955;86:385. [15] Zhang C, Wang J, Du Y, Zhang W. J Mat Sci 2007;42:7046.
Contents lists available at ScienceDirect
Intermetallics journal homepage: www.elsevier.com/locate/intermet
Phase equilibria in the AleSieV system: The vanadium rich part Beatrix Huber, Klaus W. Richter* University of Vienna, Department of Inorganic Chemistry/Materials Chemistry, Währinger straße 42, 1090 Wien, Austria
a r t i c l e i n f o
a b s t r a c t
Article history: Received 2 September 2010 Received in revised form 21 October 2010 Accepted 22 October 2010
The V-rich part of the AleSieV phase diagram was determined by a combination of optical microscopy, powder X-ray diffraction (XRD), differential thermal analysis (DTA) and electron probe microanalysis (EPMA). Phase equilibria were investigated at two isothermal sections at 850 and 1300 C. High temperature DTA was performed to identify the ternary invariant reactions yielding a ternary reaction scheme and the vertical section at 50 at.% V. As cast samples were investigated in order to gain additional information about primary crystallization fields. A liquidus surface projection was constructed for the entire ternary system by combining our experimental data with those from literature. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: A. Aluminides miscellaneous B. Phase diagram B. Phase transformations
1. Introduction and literature review A detailed description of the ternary phase equilibria in the Vpoor region of the AleSieV phase diagram (i.e. between 0 and 50 at.% V) was given recently by Huber et al. [1]. The current investigation is a continuation of our previous studies in the Al-rich part of the ternary AleSieV system and covers the V-rich part. Phase equilibria at 850 and 1300 C will be presented here as well as the vertical section at 50 at.% V. An entire liquidus surface projection and a reaction scheme are given. The binary sub systems AleV, SieV and AleSi have been investigated repeatedly and are given in Massalskis Handbook of Binary Alloy Phase Diagrams [2]. A detailed literature survey is given in [1], based on the work of Richter and Ipser [3] and Gong [4] for AleV and of Jorda and Muller [5], Savickij [6] and Zhang [7] for SieV. Recently, Huber and Richter [8] discovered a low temperature phase in the binary AleV system with the composition Al50.1(5)V46.9(5). There is indication that the phase crystallizes in a tetragonal structure (possible space group P42212), but so far it was not possible to determine the structure. The invariant binary reactions from literature relevant to the current work are listed in Table 1 together with the respective references. An early assessment of the AleSieV system was given by Haußmann [9] and is based primarily on the experimental investigation of Gebhardt and Josef [10] and Müller [11]. Müller’s [11] investigations in the V-rich part of the ternary system concern the superconducting binary phase SiV3 and its ternary solid * Corresponding author. E-mail address: [email protected] (K.W. Richter). 0966-9795/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.intermet.2010.10.022
solution Al1xSixV3. A partial isothermal section at 1000 C shows the maximal solubility of Al in SiV3 to be 3.5 at. % and an equilibrium of the three phases SiV3 þ Si3V5 þ (V). Our recently published study [1] includes three isothermal sections at 500 C and 620 C for alloys up to 20 at.% V and at 850 C for alloys with more than 20 at.% V. The results of this study differ considerably from those of Gebhardt and Josef concerning the solubility of the binary phases in the ternary system. Apart from that the ternary phase Al0.6Si1.4V was discovered and its crystal structure was determined (Pearson symbol oF24, space group Fddd, TiSi2-type). In Table 2 the invariant ternary reactions reported in our previous study [1] are listed. 2. Experimental The samples were prepared from aluminum slug (99.999%, Alfa Aesar, Karlsruhe, Germany), silicon lump (99.9999%, Alfa Aesar, Karlsruhe, Germany) and vanadium pieces (99.7%). To remove oxides at the vanadium surface it was treated with diluted hydrochloric acid for some minutes in the ultrasonic bath. After rinsing it with water and acetone it was dried for some minutes at 120 C. Calculated amounts of the elements were weighed to an accuracy of 0.05 mg and arc melted on a water-cooled copper plate under an argon atmosphere. Zirconium was used as a getter material within the arc chamber. The reguli with a total mass of about 1000 mg were remelted four times for good homogenization and weighted back after each melting step to control the mass losses. The total mass losses were less than 1% and can be considered to be negligible. A series of samples were placed in alumina crucibles, which were sealed into evacuated quartz glass ampoules and then annealed at 850 C for three weeks. Putting the samples into alumina
370
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
Table 1 Binary invariant phase equilibria relevant to the present study. Phase reaction
Composition of the involved phases/at.%
T/ C
Ref.
L þ (V) ¼ Al8V5 L ¼ Si2V þ Si5V6 L þ Si3V5 ¼ Si5V6 L ¼ SiV3 þ (V) L ¼ SiV3 þ Si3V5 L ¼ SiV3 L ¼ Si3V5 L ¼ Si2V Si5V6 ¼ Si2V þ Si3V5
w33 V w59 Si w57 Si w13 Si w29 Si w25 Si w37.5 Si 66.7 Si e
1408 1640 1670 1870 1895 1925 2010 1677 w460
[3] [14] [14] [2] [2] [2] [14] [2] [15]
w50 V 66.7 Si w37.5 Si w7 Si w25.5 Si w25 Si w37.5 Si 66.7 Si e
w38.5 V 45.4 Si w45.4 Si w19 Si w37.5 Si e e e e
crucibles was necessary to avoid reaction of the Al containing samples with the quartz glass during the heat treatment. After the annealing time the samples were quenched in cold water. Another series of samples were sealed into evacuated Ta-crucibles, annealed at 1300 C for one week in a permanent Ar-flow and then quenched in a water-cooled container. All samples were separated into several pieces, which were investigated by X-ray powder diffraction (XRD), electron probe microanalysis (EPMA) and metallographical analysis. Initial sample characterization was performed by powder XRD using a Bruker D8 Advance Diffractometer operating in reflection mode (Cu Ka radiation, Lynxeye silicon strip detector) and the patterns were analyzed using the TOPAS software [12]. To gain information about ternary phase reactions selected samples were measured by DTA and high temperature DTA, respectively. Polished pieces of the annealed samples were investigated by optical microscopy using a Zeiss Axiotech 100 microscope equipped for operation under polarized light. Selected samples were analyzed by EPMA in order to determine reliable phase compositions. The EPMA measurements were carried out on a Cameca SX electron probe 100 (Cameca, Courbevoie, France) using wavelength dispersive spectroscopy (WDS) for quantitative analysis. Pure aluminum, silicon and vanadium were employed as standard materials. The measurements were carried out at 15 kV using a beam current of 20 nA. Conventional ZAF matrix correction was used to calculate the final composition from the measured X-ray intensities. To examine as cast samples, SEM measurements were carried out on a Zeiss Supra 55 VP equipped with an EDX detector. DTA measurements up to a maximum temperature of 1490 C were performed on a DTA 404 PC (Netzsch, Selb, Germany) using open alumina crucibles and employing a slow permanent argon flow. A sample mass of approximately 150 mg was used for the experiments and the samples were checked routinely for possible mass changes during the DTA investigations. No relevant mass changes were observed. Two heating- and cooling-curves were recorded for each sample using a heating rate of 5 K min1. The Pt/Pt10%Rh thermocouples (type S) of the DTA instrument were calibrated at the melting points of pure Al, Au and Ni. High temperature DTA measurements (above 1490 C) were carried out on a Setaram Setsys Evolution 2400 using covered alumina crucibles within a slow permanent argon flow. A sample mass of Table 2 Ternary invariant phase reactions in AleSieV reported earlier [1]. Ternary invariant reaction
T/ C
E1: L ¼ (Al) þ (Si) þ Si2V U1: L þ Al45V7 ¼ (Al) þ Si2V U2: L þ Al21V2 ¼ (Al) þ Al45V7 U3: L þ Al3V ¼ Al23V4 þ Al45V7 P1: L þ Al3V þ Si2V ¼ Al45V7 U4: Lþs ¼ Al3V þ Si2V U5: L þ Si3V5 ¼ Al3Vþs P2: L þ Si2V þ Si3V5 ¼ s U6: L þ Al8V5 ¼ Al3V þ Si3V5
577 643 w664 733 745 866 976 1140 1224
Fig. 1. Isothermal section of the AleSieV system at 850 C. Diamonds: nominal composition of the samples investigated by EPMA. The V-poor part is drawn according to [1] (bright gray).
approximately 100 mg was used for the experiments and no relevant mass changes before and after the measurements were observed. One heating-curve was recorded for each sample using a heating rate of 5 K min1. The W5%Re/W26%Re thermocouple (type W5) was calibrated at the melting points of pure Au, Ni and Pt. 3. Results and discussion 3.1. Isothermal sections at 850 and 1300 C The isothermal section at 850 C constructed based on XRD and EPMA measurements, is shown in Fig. 1. In the V-rich part of the Table 3 Phase composition in various samples annealed at 850 and 1300 C determined by EPMA. Phase field Nominal sample T/ C of annealing composition Al15Si23V32
850
[Al8V5 þ SiV3 þ Si3V5]
Al70Si10V20
850
[Al8V5 þ SiV3 þ (V)]
Al30Si5V65
850
[SiV3 þ (V)]
Al10Si10V80
850
[SiV3 þ (V)]
Al5Si10V85
850
[SiV3 þ (V)]
Al40Si10V50
1300
[Al8V5 þ Si3V5 þ (V)]
Al10Si25V65
1300
[SiV3 þ Si3V5 þ (V)]
Al24Si8V68
1300
[SiV3 þ (V)]
Al21Si7V72
1300
[SiV3 þ (V)]
Al10Si10V80
1300
[SiV3 þ (V)]
– not measured due to fine microstructure.
Phase composition by EPMA/at.%
Al8V5 SiV3 Si3V5 Al8V5 SiV3 (V) SiV3 (V) SiV3 (V) SiV3 (V) Al8V5 Si3V5 (V) SiV3 Si3V5 (V) SiV3 (V) SiV3 (V) SiV3 (V)
Al
Si
V
55.3(4) 3.9(4) 1.8(1) 55.9(2) 6.6(3) 44.9(3) e e e e e e 56.4(5) 3.5(1) 45.2(4) 4.7(1) 4.1(1) 42.2(3) 6.2(1) 32.0(1) 6.1(6) 24.6(1) 5.3(4) 13.5(1)
1.7(1) 23.0(1) 35.2(1) 1.2(1) 20.0(8) 0.3(1) e e e e e e 2.0(2) 33.7(1) 1.6(2) 21.5(2) 32.8(4) 1.4(2) 19.1(1) 1.7(1) 18.4(9) 1.8(1) 17.8(9) 2.1(1)
43.0(5) 73.1(5) 63.0(1) 42.9(2) 73.4(3) 54.8(2) e e e e e e 41.6(4) 62.8(2) 53.2(7) 73.8(1) 63.1(1) 56.4(2) 74.7(2) 66.3(1) 75.5(8) 73.6(1) 76.9(7) 84.4(1)
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
371
Fig. 2. Determination of the Al-concentration in (V) for two-phase samples using lattice parameter interpolation according to Vegard’s law.
AleSieV phase diagram at 850 C we observed the three phase fields [Al8V5 þ SiV3 þ Si3V5] and [Al8V5 þ SiV3 þ (V)] and the large two phase region of [SiV3 þ (V)]. The corresponding compositions in the different phase fields are given in Table 3. Earlier results in the Al-richer part are drawn brighter and the exact compositions are listed in [1]. As reported previously [1], beside tetragonal Si3V5 we found traces (1e5%) of hexagonal Si3V5 (Mn5Si3-type) in the samples and again we assume that oxygen impurities are responsible for the occurrence of this phase. The maximal solubility of Al in SiV3 is 6.6 at.% while Si3V5 dissolves only 1.8 at.% Al. The maximal solubility of Si in Al8V5 was measured to be 1.7 at.%. Samples located in the two phase field [SiV3 þ (V)] exhibit a very fine microstructure which makes the measurement of the composition of the respective phases by means of EPMA impossible. In XRD-measurements the presence of the phases SiV3 and (V) were measured unambiguously and the cell parameters of the phases were deterimend by Rietveld refinement. Vegard’s law was applied to get information about the composition of the solid solution of Al in (V) in order to fix the tie lines shown in Fig.1. For this purpose the cell parameters of pure vanadium [13] and of Al45V55 (sample Al30Si10V60) were used for linear interpolation. Since the solubility of Si in (V) is very small the variation of the cell parameter can be attributed only to the substitution of the V-atoms by Al-atoms. In Fig. 2 the corresponding graph can be seen. In the binary SieV system Savickij et al. [6] measured the solubility of SiV3 to be 24.3e24.7 at.% Si at 800 C. Recently Zhang
Fig. 3. Isothermal section of the AleSieV system at 1300 C. Diamonds: nominal composition of the samples investigated by EPMA.
et al. [7] gave a thermodynamic reassessment of the entire SieV system and calculated the maximum solubility of SiV3 to be 20e25 at.% Si at 1837 C. At a temperature of 900 C the solubility decreases completely. Our results indicate that also the binary phase SiV3 exhibits certain extension at 850 C, and not to be so narrow like Zhang et al. [7] calculated in their assessment. So we used the phase boundaries given by Savickij et al. [6] in our sections at 850 C and 1300 C as binary reference. The results of our measurements at 1300 C are shown in Fig. 3 and the corresponding compositions determined by means of EPMA are also listed in Table 3. In contrast to the isothermal section at 850 C one can see, that the three phase fields [Al8V5 þ Si3V5 þ (V)] and [SiV3 þ Si3V5 þ (V)] exist. This result is in accordance with the work of Müller [11], who determined the Vrich part of the ternary AleSieV system at 1000 C. Annealing experiments were done to confirm that our results at 850 and at 1300 C are equilibrium situations and that a solidesolid transition reaction Si3V5 þ (V) ¼ Al8V5 þ SiV3 takes place above 850 C. On the one hand the sample Al15Si23V62 annealed at 850 C exhibiting the phases Al8V5, SiV3 and Si3V5 was annealed at 1180 C for three weeks. Subsequent EPMA measurements proved
Fig. 4. Back scattered electron image of the samples Al15Si23V62 and Al40Si10V50 annealed at 850 and 950 C, respectively.
372
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375 Table 4 Ternary invariant phase reactions. Reaction
T/ C
Phase Composition/at.%
U7: L þ (V) ¼ Al8V5 þ Si3V5
1390
L (V) Al8V5 Si3V5
w54. 0 w8.0 44.5 0.5 55.6 1.4 1.7 35.3
w38.0 55.0 43.0 63.0
U8: L þ SiV3 ¼ Si3V5 þ (V)
1574
L SiV3 Si3V5 (V)
w35.0 5.0 1.7 44.5
w10.0 22.0 35.5 0.5
w55.0 73.0 63.0 55.0
U9: L þ Si5V6 ¼ Si2V þ Si3V5
1635
L Si5V6 Si2V Si3V5
w9.0 0.0 0.2 0.5
w54.0 45.5 65.8 37.5
w37.0 54.5 34.0 62.5
55.5 45.0 4.0 0.5
1.5 0.5 23.0 62.5
43.0 54.5 73.0 37.5
Al
Fig. 5. Vertical section in AleSieV at 50 at.% V. Circles: invariant thermal effects; Small circles: other thermal effects.
U: Si3V5 þ (V) ¼ Al8V5 þ SiV3 850 < T < 950 Al8V5 annealing experiments; (V) not measured by DTA SiV3 Si3V5
Fig. 6. Partial ternary reaction scheme (Scheil Diagram) for AleSieV.
Si
V
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
373
Fig. 7. Liquidus surface projection of AleSieV including fields of primary crystallization. Solid lines: liquidus valleys; dotted lines: isotherms; circles: invariant points; triangles: investigated as cast samples.
that the three phases SiV3, Si3V5 and (V) were formed. On the other hand a sample with the same overall composition annealed at 1300 C and therefore exhibiting the phases SiV3, Si3V5 and (V) in large grains, was annealed at 850 C for seven weeks. XRDmeasurements of this sample however did still show the same phases. With the assumption that the respective reaction Si3V5 þ (V) ¼ Al8V5 þ SiV3 is hindered kinetically due to long diffusion distances in the course microstructure of the sample, the same sample was powdered, pressed to a pellet and again annealed at 850 C for five weeks. After this procedure the phase Al8V5 was formed and (V) was not longer found. In an attempt to further specify the temperature of the solid-state transition reaction, the sample Al40Si10V50 was annealed at 950 C directly after arc melting. It showed the phases Al8V5, SiV3 and (V), indicating the equilibrium above the transition reaction. The two situations after annealing at 850 C and 950 C are shown in Fig. 4. By means of DTA measurements we tried to determine the exact reaction temperature but since solidstate reactions generally show only small heat effects and the transition is presumably very slow, this was not possible. Considering our measurements and the results given in [11] we conclude that the reaction takes place between 850 and 950 C.
3.2. Ternary phase reactions and liquidus surface projection The ternary phase reactions were studied by DTA at medium and high temperature. For regular DTA measurements two heating- and cooling- cycles were performed in each measurement in order to test if equilibrium conditions can be reestablished within the samples at the selected heating/cooling rate of 5 K min1. Starting temperature was 850 C i.e. the respective annealing temperature. The maximum temperature reached in these experiments was 1490 C. Metastable thermal effects were observed in most samples in the second heating cycles. The experimental data points for thermal effects shown in Fig. 5 thus represent mainly DTA results obtained on first heating of the annealed samples. In the vertical section at 50 at.% V four invariant reactions were measured. Two transition reactions at 976 and 1224 C and one peritectic reaction at 1140 C were already established earlier (Table 2). The fourth invariant reaction at 1390 C was interpreted to be the transition reaction: L þ (V) ¼ Al8V5 þ Si3V5. In DTA measurements of the sample Al30Si20V50 we found only one invariant effect at 1390 C. The effect corresponding to the invariant reaction U6 [L þ Al8V5 ¼ Al3V þ Si3V5] at 1224 C is missing, although it should be present according to our interpretation. Our XRD results show only
374
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
Table 5 Primary crystallization investigated from as cast samples. Nominal sample composition
Primary crystallization
Al75Si5V20 Al76Si14V10 Al1.5Si44.7V53.8 Al2.5Si44.5V53 Al15Si35V50 Al47.5Si2.5V50 Al30Si10V60 Al15Si23V62 Al30Si15V55 Al40Si5V55 Al10Si10V80 Al21Si7V72 Al28Si42V30 Al22Si48V30 Al5Si56V39 Al43Si7V50 Al38Si12V50 Al37Si10V53 Al32Si11V57 Al21Si17V62 Al7Si13V80
Si3V5
s
Si3V5 Si3V5 Si3V5 (V) SiV3 Si3V5 Si3V5 (V) (V) SiV3 Si3V5 Si3V5 Si5V6 (V) Si3V5 Si3V5/(V) SiV3 SiV3 SiV3
a small amount (approximately 8%) of the phase Al3V in the sample, so we think that the thermal effect of U6 was too small to be detected in this sample. The reaction U6 was measured in different samples in the vertical sections at 20 and 35 at.% V, which can be seen in [1] as well as in the sample Al20Si30V50 (see Fig. 5), so the existence of the reaction is well established. High temperature DTA was used to measure the temperatures of invariant phase reactions higher than 1490 C. For this purpose samples in the respective three phase fields [SiV3 þ Si3V5 þ (V)] (Al5Si23V62) and [Si2V þ Si3V5 þ Si5V6] (Al0.7Si44.3V55) were heated in high temperature DTA with a heating rate of 5 K min1. Starting temperature was 1400 C. In order to avoid a possible reaction of the liquid with the alumina crucibles, the measurements were
stopped after the first visible effect and not heated further to the liquidus temperatures. These measurements allowed to determine the invariant reactions in the V-rich part at elevated temperatures. A summary of these invariant four-phase equilibria including the reaction temperatures and the approximate compositions of the involved phases, which were estimated based on a combination of all available experimental data from XRD, EPMA and DTA, is given in Table 4. In Fig. 6 the reaction scheme (Scheil diagram) in the Vrich part of the system is shown. The numbering of the invariant reaction given in Huber et al. [1] is continued. In Fig. 7 the entire liquidus surface projection is shown. The Alrich part of the liquidus projection and also the isotherms from 600 to 1400 C can be seen in detail in [1]. In the current figure the isotherms from 1400 to 2000 C are shown as dotted lines. It should be emphasized, that the isotherms are only estimates, as no liquidus temperatures above 1490 C were determined. The isotherms reflect information on the binary melting temperatures and information on invariant ternary reaction temperatures. The liquidus valleys that divide the different fields of primary crystallization are shown as solid lines with arrows indicating the direction toward lower temperature. The compositions of the liquid phase in the various ternary invariant reactions listed in Table 2 and Table 4 are shown as circles in Fig. 7. In order to reach additional information about the compositions of the liquid phase taking part in the invariant four phase reactions and to construct the liquidus surface projection, as cast samples at various compositions were investigated for primary crystallization. and the sample compositions are shown in Fig. 7 as triangles. An overview of all investigated as cast samples is given in Table 5 together with information about primary crystallization. For illustration BSE images of samples in four different primary crystallization fields are shown in Fig. 8. The liquidus surface projection is dominated by the huge primary crystallization field of Si3V5. In Fig. 8a, a sample is shown near the liquidus valley which leads from P2 to U9. The white crystals were measured to be Si3V5. In Fig. 8b the white grains of Si5V6 are the
Fig. 8. Back scattered electron image of as cast samples with the nominal compositions a: Al22Si48V30, b: Al5Si56V39, c: Al43Si7V50 and d: Al32Si11V57.
B. Huber, K.W. Richter / Intermetallics 19 (2011) 369e375
primary crystallized phase and the narrow crystallization field of this high temperature phase is represented by the sample Al5Si56V39. Fig. 8c shows a sample in the primary crystallization field of (V). The as cast sample Al32Si11V57 shown in Fig. 8d exhibits SiV3 as primary crystals. Together with the as cast sample Al37Si19V53 located directly on the liquidus valley between U7 and U8 it offers valuable information about the composition of the transition reaction U8. Accidentally, the composition of the liquid phase taking part in the peritectic reaction P2 at 1140 C was listed incorrectly in [1]. The composition was determined to be Al73Si16V11 and not Al74Si14V12 as indicated in [1]. Furthermore minor differences in the isopleths at 20 and 35 at.% V given in [1] are the results of the current study. Summing up, the current research results concerning the V-rich part of the ternary AleSieV phase diagram fit very well to the Alrich part published earlier [1] and the entire phase diagram is described by the combination of both works. Acknowledgements Financial support from the Austrian science foundation (FWF) under the project number P19305 is gratefully acknowledged. The
375
authors wish to thank Stephan Puchegger, Adela Zemanova and Ales Kroupa for their help in SEM measurements and Mario Ecker for experimental support.
References [1] Huber B, Effenberger HS, Richter KW. Intermetallics 2010;18:606. [2] Massalski TB, Okamoto H, Subramanian PR, Kacprzak L. Binary alloy phase diagrams. ASM; 1990. [3] Richter KW, Ipser H. Zeitschrift fue Metallkunde 2000;91:383. [4] Gong W, Du Y, Huang B, Schmid-Fetzer R, Zhang C, Xu H. Zeitschrift fuer Metallkunde 2004;95:978. [5] Jorda JL, Muller J. J Less-common Met 1982;84:39. [6] Savitskii EM, Efimov YV, Eichler K, Paufler P. Wissenschaftl Z d Univ Dresden 1978;27:675. [7] Zhang C, Du Y, Xiong W, Xu H, Nash P, Ouyang Y, Hu R. CALPHAD 2008;32:320. [8] Huber B, Richter KW. J Alloys Compd 2009;493:L33. [9] Schmid-Fetzer R. Vol. In: Petzow GaE G, editor. Aluminum-Cobalt-Silicon. Weinheim: VCH; 1991. [10] Gebhardt E, Joseph G. Zeitschrift Fur Metallkunde 1961;52:310. [11] Müller. Z. Naturforsch. 1969;24a:1134. [12] TOPAS. Version 3.0. Karlsruhe, Germany: Brucker AXS Inc.; 1999. [13] Villars P, L-DC. 1991;1e4. [14] Parthe E, Nowotny H, Schmid H. Monatsh Chem 1955;86:385. [15] Zhang C, Wang J, Du Y, Zhang W. J Mat Sci 2007;42:7046.