Age-hardening Mechanisms in a Commercial Dental Gold Alloy Containing Platinum and Palladium

Transcription

1 Age-hardening Mechanisms in a Commercial Dental Gold Alloy Containing Platinum and Palladium T. TANI, K. UDOH, K. YASUDA2, G. VAN TENDELOO1, and J. VAN LANDUYT' Department ofdental Materials Science, Nagasaki University School ofdentistry, 7-1 Sakamoto-machi, Nagasaki 852, Japan; and Center for High Voltage Electron Microscopy, University ofantwerp (RUCA), Groenenborgerlaan 171, B2020 Antwerp, Belgium The age-hardening mechanism of a commercial dental gold alloy containing platinum and palladium (in wt.%, 15 Cu, 6 Ag, 5 Pt, 3 Pd, 3 Zn, with the balance as gold) was elucidated by means of electrical resistivity, hardness tests, x-ray and electron diffraction and electron microscopy, as well as high-resolution electron microscopy. The sequence of phase transformations during isothermal aging below the critical temperature, Tc = 825 K, was described as follows: disordered solid solution a0 (FCC) metastable AuCu I' ordered - - phase (FCT) metastable a 2' disordered phase (FCC) equilibrium AuCu I ordered phase (FCT) + equilibrium a2 disordered phase (FCC). The hardening was due to the introduction of coherency strain at the interface between the AuCu I' platelet and the matrix. These ordered platelets had mutually perpendicular c-axes to compensate for the strain introduced by their tetragonality. A loss of coherency at the interface brought about softening of the alloy, i.e., over-aging. J Dent Res (70)10: , October, 1991 Introduction. It is well-known that the Types III and IV dental gold alloys develop significant changes in mechanical properties-such as hardness, strength, elasticity, and ductility-during appropriate heat treatments. These alloys were developed largely by empirical methods, which resulted in the production of what appear to be extremely complex combinations of alloys, frequently containing five or more constituents. Yet very little in-depth work has been carried out on these complex alloys, aimed at establishing their constitution and composition, the precise age-hardening mechanisms, or other physico-metallurgical aspects. By 1932, however, Wise et al. had investigated the improvements in strength, hardness, and color that could be obtained by replacing some of the gold in a typical gold-copper-silver alloy with various amounts of platinum or palladium (Wise et al., 1932; Wise and Eash, 1933). They showed that the addition of either platinum or palladium to a gold-copper-silver alloy resulted in a considerable increase in strength after an appropriate heat treatment. Furthermore, they studied a wide variety ofsuch alloys to identify phase changes associatedwithage-hardeningandemployedmetallographic and x-ray diffraction techniques. X-ray diffraction studies, however, do not give direct information on the correlation between structural and morphological changes during the aging process. Although transmission electron microscopy (TEM) has proven to be an indispensable technique for unraveling the age-hardening mechanisms and related structural and morphological changes in dental gold alloys, papers dealing with Received for publication April 20, 1990 Accepted for publication April 29, To whom correspondence and reprint requests should be addressed This work was supported by a Grant under the Monbusho International Scientific Research Program (Joint Research; ) and a Grant-in-Aid for Scientific Research (B) ( ), both sponsored by the Japanese Ministry of Education, Science and Culture. Portions ofthis paper were presented at the International Congress on Dental Materials on November 4, 1989, at Honolulu, Hawaii. these mechanisms from the standpoint of crystallography did not appear until 1975 (Kanzawa et al., 1975; Prasad et al., 1976). They showed that the age-hardening in a dental gold alloy was brought about by the formation of AuCu I face-centered tetragonal (FCT) ordered platelets on the {1001 planes of the surrounding matrix (Yasuda and Ohta, 1980). However, the configuration ofthe interface between them could not be deduced from the conventional TEM technique used in that investigation, this information being obtainable only by a high-resolution electron microscopy (HREM) study (Yasuda et al., 1986). The aim of the present study was, therefore, to elucidate agehardeningmechanisms andthe phase transformationofa commercial dental gold alloy. High-resolution electron microscopy was also performed to reveal the atomic configuration ofthe interface between the transformed ordered phase and the surrounding matrix down to the scale of inter-atomic distance. Materials and methods. The alloy used in the present study was a typical commercial gold dental casting alloy (Type IV) ofnominal composition in weight percent, 15 Cu, 6 Ag, 5 Pt, 3 Pd, 3 Zn, with the balance as gold (Platinum Gold Alloy M.C., G-C Dental Industrial Corp., Tokyo, Japan). A sheet 0.1 mm thickwas prepared by cold-rolling, for study of electrical resistivity and for electron microscopy. Billets and powder specimens were used for hardness tests and x-ray diffraction, respectively. These specimens were annealed at 1073 K in evacuated silica tubes so that a single phase of solid solution would be achieved, then quenched into ice brine. Resistivity was measured to study characteristics of phase transformation by a potentiometric method on a quenched strip during a continuous heating and cooling excursion between room temperature and 1073 K at a constant rate of 1.7 x 10-3 Ks-'. Specimens for hardness tests were isothermally aged in the temperature range of K for various periods of time. Discs of3- mm or 2.3-mm diameters punched out from the sheets were electrothinned by a double-jet technique for TEM and HREM examinations at 200 kv after subjecting the specimens to aging treatments. The electrolyte was prepared by dissolving 35 g of chromium trioxide in a solution of 200 ml of acetic acid and 10 ml of distilled water. Results. Electrical resistivity.-in Fig. 1(a), the curves show the normalized resistivity changes (p/pt) represented by the ratio ofresistivity at a measured temperature (p) and at the solution-treated temperature (1073 K) (pt). The heating curve demonstrates that the resistivity increased with temperature, started to decrease at about 350 K, went through a slight swelling at around 500 K, reached a minimum value at about 700 K, and then rose to that of the solid solution at 840 K. During the cooling process, the resistivity dropped rapidly at about 810 K and then decreased in proportion to the temperature decrease. although there was hysteresis, the inflecticz.points of 840 K and 810 K suggest these were the critical temperatures for phase transformation during the heating and 1350

2 Vol. 70 No. 10 AGE-HARDENING MECHANISMS INA DENTAL ALLOY 1351 C,) f. 0/).- a) n I.- ir (a) 810K K 0) E z n 0_ Un I- L- C.) I o 773 K * 673 K a 573 K o 473 K 7/ S.T. 10o Aging Period (s) Fig. 2-Changes in hardness during isothermal aging. appearance ofthe 110 superlattice reflection, andthe 001 superlattice CL t /reflection was found shortly after this, as seen in Figs. 3(a) and 3(b). Although the fundamental 200 reflection exhibited a change similar tothatfor ahomogeneous mechanism, the fundamental 111 reflection ~l corresponding to the disordered FCC phase split into three peaks on aging at 573 Kfor 3 x 105 s. Moreover, the 001 superlattice reflection split into two peaks asis shown in Fig. 3(a). These three peaks ofthe 111 reflectionwere identified as the 111 for an a 2 FCC phase, for Temperature (K) AuCu I' and a phases, and for the AuCu I phase as shown in Fig. Fig. 1-Electrical resistivity changes during continuous heating and 3(a). As was indicated by arrows, the apparent slight shift in the cooling (a) and the associated temperature derivative curves (b). A heating Bragg angle (26) of the diffraction peaks of the AuCu I' phase and and cooling rate of 1.7 x 10-3 Ks-' was employed. The open circles exhibit the the 2' phase was caused by the variation in composition of these continuous heating process, while the solid circles represent cooling from phases during aging, while the diffraction peaks of the AuCu I and 1073 C 2 phases were not shifted during aging. Thus, the AuCu I' ordered phase and the (a2' FCC phase are thought to be metastable phases cooling processes, respectively; therefore, the mean value of 825 K formed within a grain interior. The ordered AuCuI and thecx2 FCC was determined to be the critical temperature in the present alloy. phases were generated at grain boundaries as equilibrium phases. Fig. 1(b) also shows the derivative curves of the heating and The lattice parameter and the axial ratio were determined to be a = cooling processes that were obtained by plotting values of dp/dt nm, c/a = for the AuCu I' phase and to be a = nm, between pairs of adjacent points in curves in Fig. 1(a) against the c/a = for the AuCu I phase. temperatures midway between each pair. Clearly visible is the X- On the other hand, the metastable phases ofthe AuCu I' and (a2' shaped peak which is characteristic of order-disorder transforma- phases were difficult to detect in the diffraction patterns on aging at tion in the temperature derivative curves as seen in Fig. 1(b). 773 K as seen in Fig. 3(b). The lattice parameter and the axial ratio Age-hardening behavior.-because the critical temperature, Tc, were found to be a = nm, c/a = for the AuCu I ordered was determined from the electrical resistivity to be 825 K, age- phase on aging at 773 K for 1 x 106 s. hardening behavior was measured at several constant tempera- Conventional TEM examination.-fig. 4(a) is a typical brighttures below Tc. Fig. 2 demonstrates the variation of hardness field TEM micrograph taken from the specimen aged at 573 Kfor 3 during isothermal aging. Specimens reacted with a rapid increase x 102s corresponding to the initial stage ofhardening, and Figs. 4(b) in hardness from the initial value of 180 VHN for solution-treated and 4(c) are dark-field micrographs produced by using the 001 and (S.T.) to a maximum of 320 VHN by aging at 573 K for 3 x 105 s. A 110 superlattice reflections, respectively. A selected area electron marked decrease in hardness, however, was found after long aging diffraction (SAED) pattern corresponding to the central area offig. periods. At higher aging temperatures, the maximum value of 4(a) is shown in Fig. 4(d). A mottled contrast of fine scale is found hardness was lower, and the hardness peak appeared for lengthy in the bright-field TEM micrograph. This is direct evidence for the aging periods. existence ofcoherency strains in the specimen associated with aging X-ray diffraction.-to elucidate the phase changes during iso- (Nicholson and Nutting, 1958). In the dark-field TEM micrograph thermal aging, x-ray diffraction patterns were taken from the [Fig. 4(b)], the ordered phase appears as a bright fine scale of powder specimens aged for various periods in a K range. platelets parallel to the [010] direction. The bright regions in Fig. Figs. 3(a) and 3(b) show the 001 and 110 superlattice reflections and 4(c) are the ordered phase which is embedded parallel to the (001) the 111 and 200 fundamental reflections for the specimens aged at plane. This plate-like ordered phase was identified as the AuCu I' 573 and 773 K, respectively, for various periods indicated. FCT ordered domains that were nucleated and grown on the matrix The solution-treated specimen gave an x-ray diffraction pattern planes, as will be shown later. typical ofa disordered solid solution ofthe face-centered cubic (FCC) By prolonged aging, the fine mottled contrast became distinlattice with lattice parameters = nm, because no superlattice guishable as transformed ordered platelets. Fig. 5(a) exhibits a reflections were found in the diffraction pattern. bright-field TEM micrograph that was taken from a specimen aged An initial change in the diffraction pattern during aging was the at 573 K for 3 x 105 s. Figs. 5(b) and 5(c) represent dark-field TEM 41

3 1352 TANI et al. J Dent Res October 1991 AuCu I+ 2 (a) Oton AuCu 200 axo 200 AuCu I + ot Auu'001 AuCu AuCu ct 1AuuAuu (b) 111 a2 AuCu + 2 AuCu ao AuCu AuCul 110 AuCul AuCul X105 1X105S 3x104s, 3x103 3x103S 1X10=S 3x102' 1x10 s (deg) 20 (deg) Fig. 3-Changes in x-ray diffraction patterns during aging at 573 K (a) and 773 K (b) for various periods. Fig. 4-Electron micrographs and a selected area electron diffraction pattern taken from the specimen aged at 573 K for 3 x 102 s. (a) is a bright-field image; (b) and (c) are dark-field images using the 001 and 110 superlattice reflections, respectively; and (d) is a SAED pattern corresponding to the central part of (a). Zone axis: [001].

4 Vol. 70 No. 10 AGE-HARDENING MECHANISMS INA DENTAL ALLOY 1353 Fig. 5-Electron micrographs and a selected area electron diffraction pattern taken from the specimen aged at 573 K for 3 x 105 s. (a) is a bright-field image; (b) and (c) are dark-field images using the 001 and 110 superlattice reflections, respectively; and (d) is a SAED pattern corresponding to the central part of (a). micrographs formed by using the 001 and 110 superlattice reflections, respectively. Both of these micrographs are taken from the same area as Fig. 5(a). It can be seen that the ordered platelets have grown inboth their lengths and thicknesses. The size ofthe ordered platelets varied from 5.8 nm to 11 nm in length with increasing aging period of 3 x 102 s through 3 x 105 s. Although the Miller indices are used in Fig. 5(d) in terms of an FCC lattice, the SAED pattern showed that the FCT AuCu I' superlattices have their c- axes in the three cube-edge directions of the disordered matrix, because the simultaneous presence ofthe 001 and 110 superlattice reflections in the SAED pattern indicates that the AuCu I' ordered platelets were arranged with their c-axes perpendicular to each other (Kanzawa et al., 1975; Yasuda and Kanzawa, 1977; Yasuda and Ohta, 1980). In Figs. 4(d) and 5(d), the SAED patterns show the presence of diffuse scattering and streaks along the <100> directions in the 001 and equivalent superlattice reflections, as well as in fundamental reflections owing to coherency strains and to the thinness ofthe ordered platelets. Actually, characteristic striation contrast was found parallel to <110> directions in Fig. 5(b). The growth rate ofthe AuCu I' ordered platelets, however, seemed very slow during aging at this temperature. When a specimen was aged at 773 K for 1 x 103 s, corresponding to just prior to the occurrence of the hardening, TEM micrographs showed formation of the AuCu I platelets developing in length but not in thickness. As seen in Fig. 6(a), a bright-field TEM micrograph shows well-oriented thin platelets alternating regularly. It can be seen that these alternating thin platelets abut one another. Figs. 6(b) and 6(c) show dark-field TEM micrographs produced using the 001 and 110 superlattice reflections, respectively. The dark-field micrographs indicate that the alternating thin platelets are the AuCu I' ordered phase that is formed on the disordered matrix planes. The SAED pattern shown in Fig. 6(d) shows thatthe 001 and equivalent superlattice reflections are accompanied by diffuse scattering owing to coherency strains at the interface between the AuCu I' ordered platelet and the surrounding matrix. By lengthy aging, these alternating thin platelets have grown in both their lengths and thicknesses, as seen in Fig. 7(a), which exhibits atypicalbright-fieldtemmicrographtakenfrom a specimen aged at 773 K for 1 x 105 s. Figs. 7(b) and 7(c) are dark-field TEM micrographs formed by using the 001 and 110 superlattice reflections, respectively; the SAED pattern is shown in Fig. 7(d). In the SAED pattern, the 001 type superlattice reflections are not accompanied by any diffuse scattering and streaks along the <100> directions on this stage. Moreover, the AuCu I' ordered platelets have a smooth interface with the surrounding matrix, as seen in Fig. 7(a). The length of the AuCu I' ordered platelets grew from 52 nm to 150 nm in mean values with increasing aging at 773 K for periods of 1 x 103 s through 1x 105 s. High-resolution electron microscopy.-to elucidate the atomic arrangement at the interface between the AuCu I' ordered platelet

5 1354 TANI et al. J Dent Res October 1991 Fig. 6-Electron micrographs and a selected area electron diffraction pattern taken from the specimen aged at 773 K for 1 x 103 s. (a) is a bright-field image; (b) and (c) are dark-field images using the 001 and 110 superlattice reflections, respectively; and (d) is a SAED pattern corresponding to the central part of (a). and the surrounding matrix, HREM observation was performed so that the atomic configurations at the interfaces could be analyzed. Fig. 8 exhibits a HREM image taken from a specimen aged at 573 K for 1 x 106 s. This image was produced by using a larger size of aperture which included superlattice reflections as well as FCC fundamental reflections up to 220FCC with the incident electron beam being parallel to the [001] direction in the reciprocal lattice. In Fig. 8, one finds regions consisting ofalternating black and white lines and regions composed ofbright dots. The former is identified to be the AuCu I' ordered platelet, where only one-dimensional resolution is observed, because the alternating black and white lines indicate the alternative stacking of (001) planes containing copper (white lines) or gold (black lines) atoms; the direction ofthe c-axis is depicted by arrows. The latter regions were found to consist of arrangements of bright dots distinguishable as two different areas, indicated A and B. The region A is identified to be the AuCu I' ordered platelet that has the c-axis normal to the surface of the HREM image, the region B is a disordered FCC structure that will be the 0x2 phase, as is indicated by the optical diffraction pattern shown in Figs. 9(a) and 9(b). Discussion. The isothermal age-hardening curves ofthe present alloy showed a drastic reduction in hardening by prolonged aging after showing considerable increase in hardness at the initial stage ofaging below the critical temperature, T = 825 K, as seen in Fig. 2. Changes in x- ray diffraction patterns indicated the formation ofmetastableaucu I' FCT ordered and t2, disordered FCC phases at the initial stage of isothermal aging by a homogeneous mechanism. Prolonging the aging period caused the formation of the equilibrium AuCu I ordered and ct2 disordered phases by a heterogeneous nucleation mechanism (Hisatsune et al., 1982; Yasudaet al., 1983; Udoh et al., 1984). At present, it is not clear which atomic sites (Au site or Cu site in the unit cell of the AuCu I' and AuCu I ordered phases) will be occupied by minor constituents such as Ag, Pt, and Pd in the present alloy. It is, however, undoubtedly true that the minor constituents are soluble in the ordered phase within the solubility limit imposed by the miscibility of each constituent in the Au-Cu alloy. This is because the lattice parameter and axial ratio ofthe present ordered phase (a = nm, c/a = for the AuCu I' and a = nm, c /a = for the AuCu I) differ from those of the AuCu I formed in the equiatomic composition of a Au-Cu binary alloy (Pearson, 1964). The hardening was attributed to the metastable AuCu I' ordered platelets that were formed on the {100l planes of the disordered matrix. These ordered platelets had mutually perpendicular c-axes to compensate for the strain introduced bytetragonality ofthe AuCu I' ordered phase, as seen in the TEM micrographs (Figs. 4-7). The

6 Vol. 70 No. 10 AGE-HARDENING MECHANISMS INA DENTAL ALLOY 1355 Fig. 7-Electron micrographs and a selected area electron diffraction pattern taken from the specimen aged at 773 K for 1 x 105 s. (a) is a bright-field image; (b) and (c) are dark-field images using the 001 and 110 superlattice reflections, respectively; and (d) is a SAED pattern corresponding to the central part of (a). disordered matrix and the AuCu I' platelets have the following orientation relationship: [O01]AuCuF 11 {100L1di, namely, the AuCu I' ordered platelets are formed parallel to one ofthe three cube planes of the disordered matrix, with the planes as a contact plane (Fig. 8), as was observed in a commercial dental gold alloy (Yasuda et al., 1986). In this situation, the lattice misfit, 5 a, between the two lattices is givenby 6= 2(d -d2)i(d1+d2) (d1-d2)d1,ifd1 d2 (Hirsch et al., 1971), where d and d are the spacings of the relevant lattice planes, i.e., d1 and d2 correspond to the lattice parameters of the ordered phase and the matrix, respectively. The lattice parameters ofthe AuCu I' platelet were determined to be a = nm, c = nm, c /a= 0.962, and that of the disordered FCC matrix was found to be a = nm from x-ray diffraction patterns. If the AuCu I' ordered platelet remains coherent with the matrix, the lattice misfit along the [100]AuCu F direction is calculated from the above formula to be as compressive strain and that along the [001]Aucul direction is found to be as tensile strain. Thus, it is clear that the misfit along the c- axis of the AuCu I' platelet is larger than that of the a-direction; therefore, the strain field introduced by the misfit is observed as strain contrast at the periphery of the AuCu I' platelets as seen in Fig. 8. To compensate for these strains, the AuCu I' platelets must be arranged in a configuration of three orientation variants with perpendicular c-axis at the early stage of ordering, probably at the nucleation stage. From the 001 dark-field TEM micrographs, the length of the AuCu I' ordered platelets was measured and plotted as the logarithms ofdomain size, log D, vs. aging periods, logt, as shown in Fig. 10. Although there is some dispersion, each experimental point lies on the straight line which is expressed by D = Do t, where D is domain size ofthe ordered platelet, D. is constant, n is the logarithmic growth rate, and t is the aging period. The growth rate (n = log D/log t) was determined to be 0.10 and 0.28 for aging at 573 K and 773 K, respectively. It should be noted that a straight line on the loglog scales shows a saturation curve against the aging periods on the linear scales, if n < 1. Thus, drastic growth of the AuCu I' platelets was observed by aging at 773 K for the initial stage. As the thin platelets of the AuCu I' phase grow larger, the strain field becomes too great to be accommodated simply by bending of the lattice planes. At this stage, complete coherency breaks down and the strains at the interface between the AuCu I' platelet and the matrix are accommodated by introduction ofinterfacial dislocations. Thus, the difference in lattice spacing of the two phases leads to a loss of coherency at the interface of the AuCu I' platelet and the matrix. Actually, no strain contrast was found in Fig. 7 at the peripheral regions ofthe AuCu I'platelets aged at 773 Kfor 1 x 105 s. Therefore, it is thought that the drastic decrease in hardness is brought about by the loss of coherency at the interfaces. Although electron microscopic observation was not performed for the equilibrium phases of the AuCu I ordered and the a2

7 1356 TANI et al. J Dent Res October 1991 Fig. 8-Bright-field high-resolution electron micrograph taken from the specimen aged at 573 K for 1 x 106 s. Fig. 9-Optical diffraction patterns obtained from different areas in Fig. 8: (a) is taken from the AuCu I' ordered platelet indicated as A, and (b) is taken from the disordered c,2, phase indicated as B in Fig. 8.

8 Vol. 70 No. 10 AGE-HARDENING MECHANISMS INA DENTAL ALLOY 1357 E 102.N a.a E : K K -.~~~~ I fi Il REFERENCES HIRSCH, P.B.; HOWIE, A.; NICHOLSON, R.B.; PfikSHLEY, D.W.; and WHELAN, M.J. (1971): Electron Microscopy of Thin Crystals, London: Butterworth, pp HISATSUNE, K.; OHTA, M.; SHIRAISHI, T.; and YAMANE, M. (1982): Aging Reactions in a Low Gold, White Dental Alloy, JDent Res 61: KANZAWA, K.; YASUDA, K.; and METAHI, H. (1975): Structural Changes Caused by Age-Hardening in a Dental Gold Alloy, J Less-Comm Met 43: NICHOLSON, R.B. and NUTTING, J. (1958): Direct Observation of the Strain Field Produced by Coherent Precipitated Particles in an Agehardened Alloy, Philos Mag 3: PEARSON, W. B. (1964): A Handbook of Lattice Spacings and Structures of Metals and Alloys, London: Pergamon Press, p PRASAD, A.; ENG, T.; and MUKHERJEE, K. (1976): Electron Microscopic Studies of Hardening in Type III Dental Alloy, Mater Sci Eng 43: UDOH, K.; HISATSUNE, K; YASUDA, K.; and OHTA, M. (1984): Isother mal Age-Hardening Behaviour in Commercial Dental Gold Alloys Con- Aging periods (s) training Palladium, Dent Mater J 3: WISE, E.M.; CROWELL, W.; and EASH, J.T. (1932): The Role of the Platinum Metals in Dental Alloys II, Trans Met Soc AIME 99: Fig. 10-Domain size (D) vs. aging periods (t). WISE, E.M. and EASH, J.T. (1933): The Role of the Platinum Metals in Dental Alloys III, Trans Met Soc AIME 104: YASUDA, K. and KANZAWA, Y. (1977): Electron Microscope Observation disordered, these are undoubtedly produced at girain boundaries by in an Age-hardenable Dental Gold Alloy, Trans Jpn Inst Met 18: prolonged aging; therefore, the grain boundary products do not YASUDA, K and OHTA, M. (1980): Age-hardening Characteristics of a contribute to the age-hardening in the present a lloy. Commercial Dental Gold Alloy, J Less-Comm Met 70: YASUDA, K; UDOH, K.; HISATSUNE, K; and OHTA, M. (1983): Structural Changes Induced by Ageing in Commercial Dental Gold Alloys Containing Palladium, Dent Mater J 2: YASUDA,K;VANTENDELOO,G.;VAN LANDUYT,J.;andAMELINCKX, S. (1986): High-resolution Electron Microscopic Study ofage- hardening in a Commercial Dental Gold Alloy, JDent Res 65: