New Eutectic Alloys and Their Heats of Transformation

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

New Eutectic Alloys and Their Heats of Transformation

DIANA FARKAS and C. E. BIRCHENALL Eutectic compositions and congruently melting intermetallic compounds in binary and multicomponent systems among common elements such as A1, Ca, Cu, Mg, P, Si, and Zn may be useful for high temperature heat storage. In this work, heats of fusion of new multicomponent eutectics and intermetallic phases are reported, some of which are competitive with molten salts in heat storage density at high temperatures. The method used to determine unknown eutectic compositions combined results of differential thermal analysis, metallography, and microprobe analysis. The method allows determination of eutectic compositions in no more than three steps. The heats of fusion of the alloys were measured using commercial calorimeters, a differential thermal analyzer, and a differential scanning calorimeter.

I.

INTRODUCTION

T H E R E are various ways in which heat storage can reduce costs and make more efficient use of power generating equipment. A thermal storage unit would eliminate the difference between peak and off-peak hours, storing heat when demand is less than the generating capacity and returning it as needed. An efficient way to store heat is to utilize congruent phase transformations. The storage density depends upon the latent heat of transformation, which is the entropy change times the transformation temperature. This entropy change is small for solid-solid transformations, but much larger on melting. The first materials considered for high temperature heat storage were inorganic salts. ~ 5 The problems found in molten salts were associated with their low thermal conductivity, their highly corrosive nature, and their large volume changes during melting. The lithium and fluoride compounds having the best heat storage densities are relatively expensive materials. Metals diminish these problems, and can therefore be strongly competitive even if they have somewhat lower storage densities. The use of metallic systems for heat storage was first analyzed by Birchenall and Telkes.6 pryor 7 calculated the heats of fusion of many binary and ternary eutectics, and the most promising of these were measured by Birchenall and Riechman. 8 Materials were selected in the range of 300 to 1000 ~ on the basis of cost, availability, and storage density. Analysis of the results obtained by Riechman shows that alloys have the best observed heat storage densities if their melting temperatures lie between 700 and 850 K. Better alloys are sought in the melting ranges 900 to 1000 K and 700 K by further alloying. Therefore, it is necessary to determine congruently melting compositions in multicomponent systems which are estimated to melt in the desired temperature ranges, and evaluate the heats of fusion. In ternary systems the phase diagrams often are not known, and few thermal data are available. In some cases the formation of ternary intermetallic compounds is reported. With this lack of information any theoretical apDIANA FARKAS, formerly a Graduate Student, University of Delaware, is now Assistant Professor, Department of Materials Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061. C.E. BIRCHENALL is Distinguished Professor of Metallurgy, Department of Chemical Engineering, University of Delaware, Newark, DE 19711. Manuscript submitted June 13, 1983. METALLURGICAL TRANSACTIONS A

proach to the problem is a crude approximation. Nevertheless, such an approach can be the useful starting point of a trial-and-error method for experimental determination. An experimental method for finding and characterizing eutectics was needed that would work with no information at all on the system but would be able to use available information to converge more rapidly. The combined results of DTA, electron microprobe, and metallography provide such a trial-and-error method. Any theoretical approximation or empirical data can be used to improve the initial estimate of eutectic composition. II. CALCULATION OF E U T E C T I C POINTS AND HEATS OF FUSION For the ideal solution model all calculations are simple, and no thermal data are needed except heats of fusion of the components. For the regular solution case, the binary equations have been extended to ternary systems, an approximation that allows calculation of ternary free energies from binary data. 10-16 In this work, both ideal and regular solution models were applied to the determinations of eutectic compositions, and the results compared. Both models were applied using the intermetallic phases known to be present as components. This yielded a set of free energy equations, one for each component phase. The condition that all the fractional compositions sum to unity yields n + 1 equations in n unknown compositions and the temperature. This set of nonlinear equations was solved by computer fitting. The predicted eutectic compositions and temperatures for some alloy systems are shown in Table I. The ideal solution model can be applied to predict heats of fusion using either the participating elements or the constituent phases as components. The two approaches give different results. Both assume complete ordering in the solids, but different mixing in the liquids. Mixing of elements is likely to be a better approximation for thoroughly disordered liquids, while mixing of intermediate phases should give a better approximation when short range order is high. Regular solution equations for the latent heat of eutectic transformations were derived by Birchenall and Riechman. 8 They calculated the latent heat either from the entropy change or from the enthalpy change, assuming random mixing in the liquid. For truly regular solutions, using correct data, both approaches should be equivalent. The regular
VOLUME 16A, MARCH 1985--323

solution model can be based on either the elements or the intermetallic phases as the components of the system, and calculations were performed with both methods. Thermodynamic data for the calculations were taken from Hultgren et al. 17 except for the case Mg-Si, where they are from Eldridge et al. 18 The results are given in Table I. In addition, Table I gives the measured values for comparison with the calculations. Ideal mixing from the participating phases usually is most accurate as well as simplest. All of these methods tend to overestimate the heat of fusion, so the calculated value should be regarded as an upper bound.
III. EXPERIMENTAL P R O C E D U R E

DTA runs in the process of determination of eutectic compositions were done in an apparatus specially built for this purpose. It was designed to work with a large sample, in order to allow subsequent metallography; EDAX and XRD were performed on the same sample. The final characterization of eutectic and intermetallic compounds as heat storage media was perlbrmed in commercial differential scanning calorimeters.

IV.

RESULTS

A. New Eutectics and Intermetallic Compounds

The experimental iterative procedure was initiated by preparing a sample of an estimated composition. The melting behavior was determined by DTA. Metallographic inspection of the alloy was conducted optically and by scanning electron microscopy (SEM), and the observed phases are correlated with the DTA results. The sample was then searched for a small eutectic region, and the composition of any small zone suspected of being a eutectic mixture was obtained by energy dispersive X-ray analysis (EDAX). If some doubt remains about the number or identity of the phases present, X-ray diffraction measurements were made. If there was a eutectic region with a melting temperature in a desirable range, a new sample was prepared with the composition measured by EDAX in the eutectic region of the previous sample. The procedure is repeated until a completely eutectic alloy was obtained. New intermetallic compounds found in the process were also studied, especially in higher order systems where the probability of finding a desirable eutectic was lower. Alloys were prepared in a mercury spark gap induction furnace of relatively low frequency (50 khz), which provided heat and vigorous stirring to the melt inside a completely sealed crucible. Metallic materials were melted in alumina crucibles, and semiconductors were melted in graphite; both types of crucibles were sealed with Sauereisen No. 78 cement. The crucible was placed inside a mullite tube, with a thermocouple touching it at the bottom and argon flowing over it from the top. This procedure prevented vapor losses and provided stirring to the melt.
Table I.

A search for new eutectics and intermetallic compounds was conducted in several ternary systems and two quatemaries. The search method converged in three steps or less, in two cases to negative results. Table II summarizes the results step-by-step. Table III presents chemical and physical data on the new eutectics and compounds. Especially interesting microstructures are shown in Figures 1 through 3. Table III includes X-ray data for the new intermetallic compounds. These data are based on diffraction patterns of

Fig. l--Microstructure of Mg-Cu-Ca eutectic, SEM. Magnification 410 times.

Measured and Calculated Eutectic Composition, Temperature, Heat of Fusion

Alloy

Temperature, K Exp. Ideal* Reg.*

Composition, Mole Pct Exp. Ideal* Reg.* 47.1 12 17.5 17 16.2 37.5 29 52 0 15 18 12 36 32 53 7 17 17 i1 40 33

AHt K j o u l e / K g

Exp. 757 519 545 351 422 264 360 254 310 310 180

Ideal 1255 573 443 561 431

Ideal* 1071 573 573 410 376 410 402 376 230

Reg.

Reg.*

Si-Mg (Si-Mg2Si) 1219 1183 1289 A1-Si 852 933 834 A1-Mg-Si 833 A1-Cu (A1-A12Cu) 821 654 790 A1-Cu-Si 844 Mg-Ca (Mg MgzCa) 790 A1-Cu-Mg 779 546 541 Cu A1-AlzCu-A12MgCu Mg Mg-Cu-Zn 725 A1-Mg (A1-AI3Mg2) 724 555 676 A1-Mg-Zn 716 Mg-Zn (Mg-Mg2Zn) 613 592 655 *Participating phases, instead of elements, treated as components.
324--VOLUME 16A, MARCH 1985

584 360 406

448 381 402

456 247

477 464

METALLURGICAL TRANSACTIONS A

Table II.

Steps in the Search for New Eutectics and Intermetallic Compounds First Alloy Second Alloy Thermal Analysis 2 peaks Microscopic Examination Thermal Analysis 2 peak Third Alloy Microscopic Examination Mg6Si7Cul6, Mg2Si, Si; 3-eutectic Thermal Analysis 1 peak 1043K

System ~ Cu-Si-Mg 56 37 17 Mg-Cu-Ca 60 25 15

Microscopic Examination

Mg6Si7Cul6 + segregated Mg2Cu


3-eutectic + segregated MgzCu

Mg6Si7Cul6 + eutectic
Mg2Ca, Mg, ternary compound 3-eutectic

Cu-Zn-P 69 17 14 Cu-Si-P 83 10 7 Cu-Zn-Si 74 19 7 Si-Mg-Ca

Cu3P, brass, ternary compound; 3-eutectic single phase, probably Cu2PSi ternary compound + 3-eutectic
CaSi2, Si

1 peak, 726 K, with a bump 1 peak, 993 K 1 peak, Ill3 K 2 peaks

1 peak 726 K

2 peaks

Cu4Si, two ternary phases; 3-eutectic CaSi2, Si

1 peak, 1038 K 2 peaks Mg2Si, Si, CazMgSi3; 3-eutectic 1 peak 1138 K

Si-Cu-Mg-A1 34 27 25 14 of quaternary compound A1-Cu-Zn-Mg 57 31 8 4 A1-Mg-Si-Zn

Al2Cu, Si, Mg2Si, quaternary compound

4 peaks

same ~--

4 peaks

2 quat. 2 peaks 4-eutectic AI-Mg-Zn eutectic 2 peaks same + Mg2Si + Si ~-~Elements are listed in the order of decreasing weight percentage in the final composition.

1 peak, 793 K 2 peaks

Table III. Elements* Si Mg Ca 49 30 21 Cu P S i Cu Si Mg 56 27 17 Cu Zn 74 19 Cu Zn 69 17 A1 Cu 5731 Si 7 P 14 Zn Mg 8 4

Chemical and Physical Data on the New Eutectics and Intermetallic Compounds Transformation Temperature, K 1138 Identified Phase** Cu2PSi: 1.97, 1.93, 1.68, 2.12, 1.60, 1.53, 1.13 ,~ Si, Mg2Si, MgCavSi3: 2.56, 2.23, 1.41 A Si, Mg2Si, Mg~Cu~6-Si7 Cu4Si Cu:~P, /3 brass A1 Unidentified Phases (Comments) (This ternary phase, tentatively identified, appears to melt at the temperature given.) (eutectic microstructure in Fig. 1) (nonlamellar eutectic similar to A1-Si) two ternary phases ternar~ compound: 1.30, 1.85, 1.09, 2.19 A (lamellar eutectic) two quaternary intermetallics, and a phase rich in Cu or Zn (eutectic microstructure in Fig. 2) ternar~ compound: 1.40, 2.24, 2.63, 1.59 A (eutectic microstructure in Fig. 3, too fine to determine composition of phase) quaternary phase: 2.59, 3.60, 2.01, 5.55 A

lll3 1043 1038 993 793

Mg Cu Ca 60 25 15

726

Mg, Mg2Ca

Si Cu Mg AI 34 27 25 14 (Composition of quaternary phase)

Si, A12Cu, Mg2Si

*In order of weight percentages, as given beneath the symbol, for the last composition tested, except for Si-Cu-Mg-Al where the SEM analysis of the quaternary phase is given. **The phase symbol is underlined if the X-ray lines are shifted from the tabulated positions. Interplanar spacings, in order of decreasing intensity, refer to the preceding phase. METALLURGICALTRANSACTIONSA VOLUME 16A,MARCH 1985 325

(a)

(b)

Fig. 2--Microstructure of Mg-Si-Ca eutectic, SEM. Magnification 475 times.

Fig. 3--Microstructure of AI-Cu-Mg-Zn eutectic, SEM. (a) Magnification 133 times, (b) magnification 530 times.

multiphase alloys, and no crystal structures were determined. In most alloys it was found that the peaks corresponding to known compounds were sometimes shifted markedly from the positions recorded for the pure phases, owing to the dissolution of additional elements or to an excess of one or more of the constituent elements of the phase. In Table 1II, which summarizes the observations on these ternary and quaternary alloys, the phases with significantly shifted X-ray diffraction lines are underlined. Chemical analyses of the last investigated composition, usually the eutectic composition, are given in the order of decreasing weight percentage. Any unidentified phases are noted, and the corresponding unassigned X-ray lines are given in the order of decreasing intensity.

B. Measurement of Heats of Transformation


Heats of eutectic transformation were measured in the Perkin-Elmer DSC II when the melting temperature was lower than 700 ~ For higher temperatures the duPont D T A model 990 was used with its high temperature cell. Two of the previously measured eutectics 8 were repeated, Mg-Zn because the previous result was believed to be too low and Mg-Si as a reference. For Mg-Zn the new value was larger, and for Mg-Si excellent agreement was found with the previous measurement. The results for the known alloys are listed in Table IV, together with those newly determined. The room temperature densities used in calculating the volumetric storage densities were measured with a pycnometer bottle.

Table IV.

Summary of Heat of Fusion Measurements

System

Composition' (Wt Pct)

Temperature (K)

AH~ (kJ/kg)

Density (kg/m 3) 1900 2250 6880 6600 1380 4150 7170 7000 5600 8670 2730 3140 2000 2800 2380 2260

a757 SiMg 44 Mg 1219 305 Si-Mg-Ca 30 Mg, 21 Ca 1138 243 Mg2Cu -1114 92 Cu-P-Si 10 P, 7 Si 1113 197 Cu-Si 20 Si 1076 314 Mg-Si-Zn 38 Si, 15 Zn 1073 272 Mg-Ca 16 Ca 1063 422 Cu-Si-Mg 27 Si, 17 Mg 1043 125 Cu-Zn-Si 19 Zn, 7 Si 1038 368 Cu-Zn-P 17 Zn, 14 P 993 134 Cu-P 9P 988 176 Zn-Cu-Mg 45 Cu, 6 Mg 978 230 ZneMg -861 422 A1-Cu-Si 30 Cu, 5 Si 844 305 A1-Cu-Mg-Zn 22 Cu, 18 Mg, 6 Zn 793 184 Mg-Cu-Ca 25 Cu, 23 Ca 726 254 Mg-Cu-Zn 25 Cu, 15 Zn 725 310 A1-Mg-Zn 35 Mg, 6 Zn 716 146 Mg-Ca-Zn 28 Ca, 17 Zn 673 180 Zn-Mg 48 613 JThe most plentiful element, listed first, supplies the balance of the mass.

AHj (kJ/m 3) 144 104 68.6 104


63 130 38 x x 104 104

Cp. (kJ/kg-K) Solid 0.79 --0.50 --

Liquid ------

M 104

175 x 104 88 x 104 258 x 104


75 X x x x 104 104 104 104 10 4 104

0.75 -0.54
--

-----

152 115 96 37
71

0.42 1.30 1.51


---

-1.20 1.13
---

74 x 104 33 1 0 4

1.63
--

1.46
--

326

V O L U M E 16A, M A R C H 1985

METALLURGICAL TRANSACTIONS A

V.

DISCUSSION

The search method exceeded expectations by converging in three steps in the most complicated case. However, in some cases it was not possible to distinguish easily all the phases present. In a number of systems the phases were identified by composition or X-ray diffraction measurements. It was also difficult to recognize some eutectic microstructures. In addition to metallography, comparison of the overall composition to that of the fine, clearly eutectic region was helpful. Size of the phases was not by itself an adequate criterion, since extreme cases were found. Several samples exhibited widely varied spacings in different regions, probably as a consequence of different freezing rates. The performance of the DTA for large samples was entirely adequate for qualitative measurements. A 30 ~ difference between the measured and actual transition temperatures was found consistently for all calibration runs (Sn, A1, Ag). The corrected DTA results for multicomponent alloys were useful in the identification of the phases present when the melting temperatures of the compounds were previously known. The characterization of congruently melting alloys as heat storage materials poses no great problems. The commercial calorimeters performed well within the expected limits, with uncertainties of 5 pct for DTA and 3 pct for DSC. Conversion of heat storage densities from a mass to a volume basis requires improved density measurements. The present uncertainty in the room-temperature densities is due mainly to the porosity of the alloys. Measurement of densities of some of the alloys at the transition temperature is being performed in a related study. As reported previously) the heats of transformation are systematically well below the calculated values as summarized in Table I. The predictions of the ideal model deviate to a large extent from the true eutectic temperatures, but the compositions are a useful approximation. The regular solution model improves the values obtained for the temperatures, but not appreciably those for the compositions. In the ternary case there is no improvement whatsoever due to the assumptions that had to be made in estimating the ternary activities from binary data. Further alloying of the promising binary eutectics does lower the transformation temperature as expected, but it does not improve the heat storage density as expected. The deviation of the actual values from those calculated is erratic and unpredictable. The main reason for this variation probably is the different degrees of ordering that take place in the liquids. Since the presence of stable intermetallic compounds was sought to increase the heat of transformation, the probability of ordering in the liquid is high. To take this complication into account additional thermodynamic data are needed. The fact that in many cases the addition of a third component did not improve the heat of transformation as expected suggests that ordering in the liquid is more likely for multicomponent systems than for binary systems. Statistically it is more likely that one pair of elements will exhibit strong affinity in the liquid and therefore contribute to short range ordering. The practical implication of these conclusions is that it will be difficult to find alloys that have storage densities substantially greater than the ones presently reported, alMETALLURGICAL TRANSACTIONS A

though some aspects of their performance may continue to improve. Investigation of quaternary alloys may put metals in a Competitive position at lower temperatures. Further alloying may also yield a better accommodation of the melting temperature to the heat sources and help to control the volume change of the material on melting.

VI.

SUMMARY

New eutectics and intermetallic compounds in several multicomponent systems were characterized as heat storage media. A search method for congruently melting compositions was developed based on the combined results of differential thermal anaylsis, metallography, and microprobe analysis. The method converges in no more than three steps. Eutectic compositions and heats of fusion were calculated using ideal and regular solution theory. The results of the calculations were found to be useful as starting points for the experimental search procedure. The characterization as heat storage media was performed using commercial calorimeters. The results show that further alloying of promising binary materials lowers the melting temperature, but does not improve the heat storage density substantially. Nevertheless, metals may be in a competitive position as heat storage materials, especially if further alloying is used to lower the melting temperatures.

ACKNOWLEDGMENTS
Financial support of this work was provided by the National Aeronautics and Space Administration, Lewis Research Center, under Grant NSG-3184 for the United States Department of Energy. Frequent discussions with program monitors Richard Vernon and Joseph P. Joyce were greatly appreciated.

REFERENCES
1. J. Schroder: J. Eng. Ind., 1975, vol. 97, pp. 893-96. 2. R.P. Tye, J.G. Bourne, and A.O. Desjarlais: Molten Salt Thermal Energy Storage, NASA Report No. NASA-CR-135098, August 1976, Dynatech Corp. 3. A.A. Ferrara, R. A. Haslett, and J. P. Joyce: 12th Intersociety Energy Conversion Engineering Conference, ASME, New York, NY, 1977, pp. 547-54. 4. H.C. Haru, J.F. Dullea, A. Kardas, L. Paul, L. (3. Marianowski, E. Ong, V. Sampath, V.M. Huang, and J.G. Wolak: Molten Salt Thermal Energy Storage Systems, NASA Report CR-135419, IGT, March 1978. 5, J.L. Clayton: Research on Energy Storage for Solar Thermal Conversion, NASA Report CR-159427, October 1974. 6. C.E. Birchenall and M. Telkes: Sharing the Sun, K.W. Boer, ed., Pergamon Press, New York, NY, 1976, vok 8, pp. 138-54. 7. N. Pryor: unpublished research, University of Delaware, 1976. 8. C.E. Birchenall and A. Riechman: Metall. Trans. A, 1980, vol. llA, pp. 1415-20. 9. L.S. Darken: J. Am. Chem. Soc., 1950, vol. 72, pp. 2090-914. 10. G.W. Toop: Trans. TMS-A1ME, 1965, vol. 233, pp. 850-55. 11. F.D. Richardson: Physical Chemistry of Melts in Metallurgy, Academic Press, London, 1974, vol. 1, pp. 143-46. 12. N.J. Olson and G.W. Toop: Trans. TMS-A1ME, 1966, vol. 236, pp. 590-92. VOLUME 16A, MARCH 1985--327

13. J.L. Meijering: Acta Metall., 1957, vol. 5, pp. 257-64. 14. N.J. Olson and G.W. Toop: Trans. TMS-AIME, 1969, vol. 245, pp. 905-10. 15. M. Hillert: Phase Transformations, ASM Publication, Metals Park, OH, 1970, pp. 184-216. 16. I. Ansara, C. Bernard, L. Kaufman, and P. Spencer: CALPHAD,

1978, vo1. 2, pp. 1-15. 17. R. Hultgren, P.D. Desai, D.T. Hawkins, M. Gleiser, and K.K. Kelley: Selected Values of the Thermodynamic Properties of Binary Alloys, ASM, Metals Park, OH, 1973. 18. J.M. Eldridge, E. Miller, and K.L. Komarek: Trans. TMS-A1ME, 1967, vol. 239, pp. 775-81.

328--VOLUME 16A, MARCH 1985

METALLURGICAL TRANSACTIONS A

You might also like