New evidence for GP zones in binary Al Li alloys

Scripta METALLURGICA Vol. 20, pp. 201-206, 1986 Printed in the U.S.A. Pergamon Press Ltd. All rights reserved NEW EVIDENCE FOR GP ZONES IN BINARY AI-Li ALLOYS J.M. Papazian , C. Sigli + and J. M. Sanchez+ + Corporate Research Center, Gruman Corporation, Bethpage, NY 11714 Henry Krumb School of Mines, Columbia University, New York, NY 10027 (Received September 30, 1985) (Revised November 20, 1985) Introduction Silcock originally proposed ( I ) that precipitation in binary AI-Li alloys follows the sequence: supersaturated solid solution 6' (Al3Li) 6 (AILi) Subsequent TEM and X-ray work has generally confirmed this proposal, as described in two recent reviews (2,3). Although TEM observations often show that 6' is present immediately after quenching and f a i l to show any other phases (2,3,4), calorimetric studies led Nozato and Nakai to propose that two GP zone phases and a short-range ordered 6' phase precede the f u l l y ordered ~' phase (5). Subsequent thermal analysis experiments by Balmuth showed some low-temperature dissolution peaks similar to those observed by Nozato and Nakai, but were interpreted on the basis of the reversion of fine a' particles rather than the dissolution of GP zones (6). Currently, AI-Li alloys are the object of major commercial development efforts, and the presence or absence of a a' precursor phase may be significant in attempts to achieve optimum precipitate distributions by means of multiple stage aging or thermomechanical treatments. Thus, i t is the purpose of this communication to present new theoretical and experimental results which support the idea that a metastable phase other than a' is present after room temperature aging of binary AI-Li alloys. Theoretical Results Recently, a phenomenological model based on the Cluster Variation Method (CVM) was developed to study solid-solid and s o l i d - l i q u i d phase e q u i l i b r i a in binary alloys (7). The approach, emphasizes the description of order-disorder reactions in the solid phases, and i t was used in Ref. 7 to accurately describe the thermodynamic potentials and phase diagram of the NiAl system. Subsequently, the same phenomenological model was used to investigate the thermodynamic properties of the stable and metastable phases observed in AI-Li binary alloys (8). The CVM has been shown in the past to be a very convenient approach to study long- and short-range order in binary and multicomponent systems (8). The model is particularly useful for the description of ordered compounds that deviate appreciably from stoichiometry and, s i g n i f i c a n t l y , i t can be used to make reliable predictions concerning the free energies of metastable phases. The details of application of the model to the AI-Li system are discussed in Refs. 7 and 8 and w i l l not be repeated here. The results pertinent to this paper are shown in Fig. 1, which is a plot of the calculated stable and metastable phase boundaries. Also included are experimental data points taken from the l i t e r a t u r e for the ~-6' metastable equilibrium. Good agreement is obtained above 500 K between the calculated and experimental (4, 9) a-6' solvus concentrations; below 500 K, however, the calculated a-6' two-phase region is wider than calculated by Ceresara et al. (10) using low-angle X-ray scattering measurements. The behavior of this two-phase region can be explained in our thermodynamic description by the tendency of 201 0036-9748/86 $3.00 + .00 Copyright (c) 1986 Pergamon Press Ltd. 202 GP ZONES the solid solution to segregate below 400 K. subject of this paper. IN AL-Li Vol. 20, No. 2 This additional level of metastability is the The existence of a metastable m i s c i b i l i t y gap below 400 K is predicted by the model for AILi fcc solutions containing up to approximately 17 at.% (5 wt%) l i t h i u m . The computed m i s c i b i l i t y gap is shown in Fig. 1 as the dot-dash l i n e . This segregation tendency is metastable relative to the a-6 and a-6' e q u i l i b r i a . Consequently, the m i s c i b i l i t y gap f a l l s inside the a-6' two-phase region. This result indicates that AI-Li alloys in the indicated concentration range, when quenched fast enough, may be expected to segregate and form characteristic Guinier-Preston zones. Experllental Results Precipitation and dissolution reactions in several AI-Li alloys were examined by differential scanning calorimetry (DSC). The alloys ranged from pure binaries to quaternary alloys of commercial purity. The DSC technique was typical of current practice and identical to that previously described (11). A typical result is shown in Fig. 2, where DSC curves from a pure A1-2.5 wt% Li alloy are plotted. The solid curve is from .063 i n . thick by .250 i n . diameter discs of materiai that had been solution treated for 10 min. at 520°C, quenched in ice brine and aged at room temperature for approximately six months. The dotted line is from the same alloy after aging 24 h at 200°C. Since the 6' solvus for this composition is approximately 290°C (2), the endothermic peak in the 220-320°C temperature range is assumed to represent 6' dissolution, and the endotherm at 425-500°C is assumed to represent 6 dissolution. These endotherms appear in both of the curves. However, in the low temperature region, the room temperature aged material shows additional events: a pronounced endotherm (I00-150°C) and a subsequent exotherm (150-200°C). Both of these events are absent in the material aged at 200°C. Comparison to previous work shows similar low temperature peaks in 1.7% and 2.0% alloys aged at 55°C (5) and in a 3.0% alloy aged at 72°C (6). Additional alloys were examined, most of them displayed a low temperature dissolution peak similar to that in Fig. 2. Results from these a11oys, which ranged from pure binaries to commercial a11oys are summarized in Fig. 3, where the observed dissolution peak temperatures are plotted against the lithium content of the a l l o y . Results from previous work are also included (5,6). Also plotted in Fig. 3 is the m i s c i b i l i t y gap phase boundary predicted by the CVM calculation. The observed peak temperatures agree reasonably well with the calculation. The dissolution peak temperature is a kinetic parameter and is a function of the heating rate and precipitate size as well as being related to the phase boundary (12). Therefore, the comparison in Fig. 3 is not meant to imply that the phase boundary has been experimentally established; rather, i t is meant to indicate a general agreement between the predicted l i m i t of m i s c i b i l i t y and the observed dissolution behavior. These results indicate that there is some cluster or zone present after low temperature aging, that t h i s zone dissolves in the I00-150°C temperature range, that i t appears in many AI-Li alloys, and that i t s behavior is similar to that predicted by the CVM calculations. I t is also noteworthy that the enthalpy of the dissolution reaction, approximately 5 J/g, is similar to that observed for the dissolution of GP zones in AI-Cu alloys (5-8 J/g) (11). Nozato and Nakai (5) assumed that the dissolution of AI-Li GP zones gave rise to this peak, whereas Balmoth (6) thought i t was due to the reversion of fine 6' precipitates. The sequence assumed by Balmuth was that small, preexisting 6' precipitates were taken above their solvus by rapid heating in the DSC; thus they became unstable and dissolved giving rise to the lowtemperature endotherm. In order to examine t h i s hypothesis, DSC scans were made at a series of heating rates. The results are plotted in Fig. 4 and show that, apart from the expected s h i f t in peak temperature, the reaction is substantially the same even at the slowest heating rate of 1K/min. This indicates that the peak is not caused by reversion of 6', since at slow heating rates, 6' would be expected to grow rather than dissolve. In addition, recent work has examined the effects of c a p i l l a r i t y on the ~' solvus (13) and predicts that the solvus for a 1,0 nm particle of 6' in our 2.5% Li alloy would be at 170°C, s i g n i f i c a n t l y higher than the observed dissolution peak. A final experiment was performed to examine the precipitation kinetics of the unknown phase. A series of samples of the 2.5% Li alloy was solution treated and quenched, then held Vol. 20, No. 2 GP ZONES IN AI-Li 203 for various lengths of time at room temperature before calorimetric analysis. The results are plotted in Fig. 5 and show that immediately after quenching the dissolution peak is not present rather, there is a small exotherm (formation peak) followed by an equivalent dissolution. This indicates that the zones were not present at the start of the DSC run, but some formed in the 50-85°C interval and subsequently redissolved. The other curves show that after 2 h aging at room temperature some small zones are present, and after 8 h significantly more have appeared. Six months of aging results in a significant increase in the volume fraction (peak area) and th~ average size (peak temperature). Additional results from samples aged for times between 2 hour and 6 months have been omitted from Fig. 5, but show that the increase in peak temperature and enthalpy with aging time is a continuous and regular process. Thus, the kinetics of appearance of the zones at room temperature are very similar to those of other binary aluminum alloys. Sumr~ The DSC results described above are thought to present a r e l a t i v e l y complete and persuasive picture of the characteristics of an AI-Li phase which is metastable with respect to 6'. The phase is not present immediately after quenching; i t appears at room temperature with kinetics that are similar to those of GP zones in other A1 base alloys, i t dissolves in the 75-150°C temperature range with an enthalpy typical of GP zones, and the kinetics of i t s dissolution are also similar to those of other GP zones. The DSC measurements are not an a r t i f a c t caused by the dissolution of ~', which is significantly more stable and which is also probably present. Finally, the existence of GP zones was predicted independently on sound theoretical grounds. According to the above results, optimum conditions for observation of the zones would be after long-term aging of a 2 to 2.5 wt% alloy in the 50 to 75°C temperature region; TEM examinations of such samples are currently underway. .Ac.knowledgment s The authors would like to acknowledge helpful technical discussions with Drs. J.H. Driver and P.N. Adler and the laboratory assistance of Mr. H. Baker. Mr. A.P. Divecha of the Naval Surface Weapons Center, Dr. J.H. Driver, and Dr. B. Dubost of Cegedur-Pechiney kindly provided the AI-Li alloys. This work at Grumman was performed as part of the Grumman Corporation Internal Research and Development Program and the work at Columbia was p a r t i a l l y supported by the NSF under grants DMR-82-06195 and DMR-85-10594. References 1. 2. 3. 4. 5. 6. 7. 8. g. 10. 11. 12. 13. J . M . Silcock, J. Inst. Metals, 88, 357, (1959-60). D.B. Williams in Aluminum-Lithium, edited by T.H. Sanders, Jr. and E.A. Starke, Jr., TMSAIME, 89-100, (1981). T . H . Sanders, Jr. and E.A. Starke, Jr. in Aluminum-Lithium I I , edited by E.A. Starke, Jr. and T.M. Sanders, J r . , TMS-AIME, 1-15, (1984). D.B. Williams and J.W. Edington, Metal Science, 9, 529-32, (1975). R. Nozato and G. Nakai, Trans. J.I.M., 18, 679-80, (1977). E. Balmuth, Scripta Met, 18, 301-304, (1984). C. Sigli and J.M. Sanchez, Acta Metall, 33, 1097-1104, (1985). C. Sigli and J.M. Sanchez, Acta Metall, to be published. B. Noble and G.E. Thomson, Met. Sci. J., 5, 144, (1975). S. Ceresara, G. Cocco, G. Fagherazzi, G. Giarda, and L. S c h i f f i n i , La Metall. Italiana, I , 20, (1978). J . M . Papazian, Metall. Trans., 12A, 269-80, (1981). J . M . Papazian, Metall. Trans., 13A, 761-69, (1982). S . F . Baumann and D.B. Williams, Acta Metall, 6, 1069-78, (1985). 204 GP ZONES 1100 I I I IN AI-Li I i Vol. i i i I LIQUID 900 T 700 500 ~ ~ LIQUID //<o' rl. .12 III~ xl 300 |/ 0.0 t "\ 0.1 , I 0.2 , 1 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 ATOMIC CONCENTRATION OF Li FIG. 1. Calculated phase boundaries in the binary AI-Li system. The metastable 6' field is denoted by dashed lines; the predicted metastable miscibility gap is shown in dot-dashed lines. The experimentally determined 6' solvus data points also are shown. 1.6 I I I ! I l I I ~' DISSOLUTION I I DISSOLUTION o (z w -lkO E3 1.4 [ e z o~ 1.2 L) 1.0 /--- 24 h at 200°C I I 100 I I 200 I I 300 I I 400 I I 500 T.C FIG. 2. DSC results obtained from a 2.5% Li alloy at 10 K/rain. 20, No. 2 Vol. 20, No. 2 GP ZONES 200 I IN AI-Li I 205 | I Q 150 L) I00 50 A REF. 6 Q THIS WORK CALCULATED PHASE BOUNDARY 1 I I , I 1 2 3 4 LITHIUM CONCENTRATION,WEIGHT% FIG. 3. Dissolution peak temperatures observed in various AI-Li alloys including results from Refs. 5 & 6. The solid line is the predicted metastable miscibility gap calculated using the CVM. 5 K/min,..--.,~.. ' . ,.-~. /" ,'" ;~.V',,.~IO ~,/ ,." '.?', K/ram • ". v ~" o 0 0 't ,~., ~, K/rain ',t'~---20K/min\ ",k .," ,,, \ ., - u_ <3 f: -.1 I -.2 25 I 75 I I 125 I I 175 T.°C FIG. 4. The effect of DSC heating rate on dissolution in AI2.5% Li. The material had been solution treated, quenched, and aged for 6 mo. at room temp. 206 GP ZONES .2 I I 8h IN AI-Li I --~ I 2h--- I I /"- '~--- ~-~_/" /." . .. " ; "~. Vol. ! =~ 6 mo ,,\ '" "!, PI~Z) ,~ 0 oo0.a. -,1 -.2 25 I I 75 I I 125 I 175 T,~C FIG. 5. The effect of delay time at room temp. between quenching and DSC analysis of AI-2.5% Li. 1~ 20, No. 2