Platinum Metals Rev., 2009, 53, (4), 189
Precious Palladium-Aluminium-Based Alloys with High Hardness and Workability
PROMISING POTENTIAL FOR APPLICATION IN JEWELLERY AND WATCHMAKING
- Julien Brelle
- Andreas Blatter**
- PX Holding SA,
- R&D, Boulevard des Eplatures 42, CH-2300 La Chaux-de-Fonds, Switzerland
- René Ziegenhagen
- Cartier Horlogerie, Branch of Richemont International SA,
- 10 Chemin des Aliziers, CH-2300 La Chaux-de-Fonds, Switzerland
New palladium-aluminium-based alloys with promising potential for application in the areas of jewellery and watchmaking are presented. A particular emphasis is placed on the mechanical behaviour of ternary palladium-aluminium-ruthenium (PdAlRu) alloys with 95 wt.% Pd. The new alloys combine high plasticity with high hardness relative to common Pd alloys. The low work-hardening rate enables cold working in excess of 95% reduction without intermediate annealing. The hardness (Vickers pyramid indentation) ranges from 100 HV to 300 HV in the annealed condition, depending on the Al:Ru ratio. Their whiteness in terms of colour coordinates is compared with platinum and white gold. The feasibility of porcelain fusion to PdAlRu for decorative purposes is also demonstrated.
Palladium is not widely recognised as a precious metal in jewellery and watchmaking. Yet, with the price evolution of precious metals over the last few years, use of palladium in these markets has seen renewed interest (1–3). For illustration, gold was roughly double the price of palladium in 2006 and about four times the price of palladium at the end of 2008 (4). In addition, palladium alloys for jewellery, which usually contain 95 wt.% Pd (950 Pd), have a lower density (around 12 g cm−3) than 18 carat white gold (close to 15 g cm−3) and 950 platinum (about 21 g cm−3). Hence, an item of volume 1 cm3 in 950 Pd contains 11.4 g Pd. By comparison, the same item made in 18 carat white gold (750 Au) or 950 Pt will contain, respectively, 11.3 g Au or 20 g Pt.
Furthermore, the 950 Pd alloys approach the ‘ideal’ white colour of platinum without requiring rhodium plating like most white gold alloys. Unlike for gold alloys, the white sheen will therefore not wear off, eliminating the bother and expense of re-plating. 950 Pd alloys also satisfy the general requirements for jewellery and watch alloys: they are nickel-free, malleable, easy to polish, and have desirable setting and forming characteristics. Their high Pd content also confers good corrosion and tarnishing resistance, a crucial aspect in jewellery and watchmaking.
The inherently low hardness of Pd alloys is, however, an important technical limitation for their use in jewellery and particularly in watchmaking. The hardness of existing 950 Pd alloys, with alloying metals such as ruthenium (PdRu), gallium (PdGa) or copper (PdCu), falls between 70 HV (PdCu) and 120 HV (PdGa) in the annealed state, and between 145 HV (PdCu) and 200 HV (PdGa) after 75% strain hardening. These values are substantially lower than those typical of platinum or white gold alloys (≥ 130 HV annealed, ≥ 250 HV work hardened).
In an attempt to develop a 950 Pd single-phase alloy with substantially higher hardness, comparable with platinum and white gold, while maintaining the favourable colour and workability of conventional Pd alloys, we investigated PdAl-based compositions, in particular the PdAlRu system. This paper describes the background to the alloy development, presents the main characteristics of 950 PdAlRu alloys in terms of mechanical properties and workability, and addresses the possibility of fusing coloured ceramic material to the alloy for decorative purposes.
Background to the Development of the PdAlRu Alloys
While precipitation hardening may also be of interest to further increase the rigidity and wear resistance of finished components, solid solution strengthening is the mechanism that must provide the base hardness of the alloy in the annealed state.
Solid solution strengthening is the result of strain produced in the crystal lattice, mainly by the size misfit between matrix and solute atoms. Since size misfit also limits the terminal solid solubility, as described by the Hume-Rothery rules (5), it must be kept within certain limits to ensure a solubility of at least 5 wt.%, which is necessary for a 950 Pd single-phase alloy. For a given solute, the strength increases with its atomic fraction (at.%). Higher atomic fractions are achieved when alloying with light elements. In 950 Pd, for illustration, 5 wt.% aluminium corresponds to 17.2 at.%. A 950 Pd alloy may include several alloy additions, which must be fully soluble and must add up to a total of 5 wt.%.
The effects of a great number of solute elements on the hardness of palladium have been compiled previously (6, 7). Germanium, silicon and boron have a strong hardening effect. However, B is difficult to alloy and Ge and Si both exhibit nearly zero solubility. As a result, when added in sufficient concentrations to give a hardness above 150 HV, these elements tend to precipitate at the grain boundaries and thereby render the alloy too brittle for practical use. For those elements which are more practical in terms of alloying, such as other precious metals or 3d transition metals, the hardness values attained at concentrations of 5 wt.% barely exceed 100 HV. Ru is one of the elements showing the most pronounced effect on the hardness of Pd alloys. Hardness values in the range 150 HV to 200 HV can be achieved with rare earth metals such as cerium (8).
With respect to the light elements, 5 wt.% titanium raises the hardness to 150 HV (9), while Al boosts the value to 320 HV (440 HV after 80% cold work), according to our own experiments. While a hardness of 150 HV is at the lower edge of the target range, a value of 320 HV may be inconveniently high for many conventional jewellery manufacturing operations such as stamping or setting.
The present study therefore focused on PdAl-based alloys, incorporating ternary additions of Ru, Ti and magnesium in order to moderate the hardness. Ru was chosen because it is a noble metal, a good solution hardener, and commonly used in 950 Pd alloys. Ti and Mg were chosen because they are lightweight, good solution hardeners, and may possibly provide a mechanism of precipitation hardening by the formation of tiny AlTi or AlMg compounds upon ageing – in similarity with superalloys. Table I shows the hardness values obtained for various ternary alloys in the annealed and 80% work hardened conditions, respectively. This shows that the hardness values in the annealed state lie in the target range and that cold work generates substantial hardening. The values displayed are those for ‘low’ and ‘high’ concentrations of the ternary additions; intermediate concentrations gave intermediate values for hardness. All alloys were single phase and sufficiently malleable for a rolling reduction of 80% without cracking. It is interesting to note that there have been two independent patent applications for 950 PdAl-based alloys (10, 11).
The advantage of Ru as the ternary element is that unlike Mg or Ti, it does not cause a violent reaction with Al upon alloying. In this paper, we focus on our development work on the 950 PdAlRu system.
Two alloys of nominal composition (in wt.%) Pd95.5Al0.9Ru3.6 and Pd95.5Al2.8Ru1.7 were prepared in a vacuum induction melting unit. The unit chamber was evacuated and purged with argon several times before backfilling with argon to 600 mbar. The elemental metals were melted in a zirconia crucible. The Al flakes were wrapped in Pd sheets to avoid any reaction with the crucible and also to alloy the Al with the higher melting point Pd without significant expulsion of Al. Plate-like ingots of 5 kg each were cast into an oxidised copper mould to constitute the feedstock for the various tests. After a first flat rolling, rods were cut off the plate and further rolled to adequate size for tensile testing while the rest of the plate was used for workability tests and microstructural investigations.
The X-ray diffraction pattern in Figure 1 reveals that the ternary alloys are essentially single phase, face centred cubic (f.c.c.) solid solutions. In Ru-rich alloys, a new diffraction peak appears, and its intensity increases with increasing Ru content. Additional peaks, too weak to be seen in Figure 1, become visible when zooming into the data. These peaks, located at scattering angles, 2θ, of 44.0°, 58.3°, 78.2°, 84.8°, and 104.8°, respectively, are close to those of pure Ru and enable the second phase to be assigned to a Ru-rich PdRu hexagonal close packed (h.c.p.) solid solution.
The measured lattice constants, a, of the ternary f.c.c. matrix are accurately reproduced with a hard sphere approximation by the linear combination of the atomic sizes, Sj, defined as the minimum interatomic distance in the unit cell of element j (Equation (i)):
where the coefficients cj correspond to the atomic fraction of element j.
The atomic sizes for Pd, Al and Ru are, respectively, 2.750 nm, 2.863 nm and 2.650 nm (12). Since Al has a greater atomic size than Pd by 4.1%, whereas Ru is smaller by 2.6%, the apparent shift in a is marginal among different ternary alloys. In other words, strengthening due to lattice distortion is not apparent through a significant shift of the diffraction peaks. In particular, there exists a ternary composition for which the lattice constant almost matches that of elemental Pd. The lattice constant derived from the diffraction pattern of Pd95.5Al0.9Ru3.6 is 3.888 nm, compared to 3.886 nm for Pd.
The hardness of the PdAlRu alloys in the annealed state (1000°C for one hour), measured using a Vickers hardness tester with a 1 kg load (HV1), approximates to a linear function of the Al content, as shown in Figure 2. Therefore, the hardness can be tuned to any value from about 100 HV (PdRu) to 320 HV (PdAl). The hardness values in the work-hardened condition range from 165 HV (PdRu) to 440 HV (PdAl). Corresponding values for two ternary compositions are given in Table I. Upon ageing of annealed samples at 700°C for twenty minutes, the hardness of those alloys with higher Ru content increases slightly, indicating a mechanism of precipitation hardening. The intensity of age hardening remains modest, however: it approaches but does not exceed the increase of 25 HV observed at the highest Ru content, i.e. for the binary 950 PdRu alloy. This strengthening is also evident in an increase in yield strength of 50 MPa for Pd95.5Al0.9Ru3.6 (Figure 3).
Age hardening is accompanied by a substantial increase in electrical resistivity, ρel, as measured by means of the four-point probe technique on discs of thickness 2 mm and diameter 27 mm (13). For Pd95.5Al0.9Ru3.6, ρel reversibly switches from 25.5 μΩ cm in the annealed state to 91 μΩ cm in the age hardened state. By comparison, the values for Pd95.5Al2.8Ru1.7, which does not exhibit age hardening, are 32.1 μΩ cm and 35.6 μΩ cm, respectively. Since an increase in electrical resistivity is caused by additional scattering of electrons at crystal imperfections, such as a lattice distortion or the presence of precipitates, and since age hardening occurs with the appearance of a PdRu phase as discussed above, it is tempting to correlate age hardening with PdRu precipitates. However, the main diffraction peak of the PdRu phase persists upon annealing, suggesting that the precipitates are not fully solubilised.
Figure 4 shows the variation of conventional yield strength, Rp0.2, ultimate tensile strength, Rm, and fracture strain, A50, with cold work. The alloys undergo significant strain hardening only upon initial cold working. Beyond about 30% of cold work, yield strengths (Figure 4) and hardness values (Figure 5) remain essentially constant. It is worth mentioning that the tensile properties of the Ru-rich alloy after standard annealing (1000°C for 1 hour) depend on its thermomechanical history. This is no longer the case after cold working. We attribute this memory behaviour to a variable evolution and dissolution of the PdRu precipitates.
where σt is the true stress, σ is the engineering stress, ɛt is the true strain and ɛ is the engineering strain.
When applied to the stress-strain curves of annealed samples in the plastic domain (Figure 3), this approximation returns n = 0.29 for Pd95.5Al2.8Ru1.7 and n = 0.25 for Pd95.5Al0.9Ru3.6, values that are typical of low stacking-fault energy alloys such as Al alloys.
Figure 6 depicts the mechanical properties of PdAlRu alloys in comparison with those of common Pd, Pt or Au alloys. The ternary 950 PdAlRu alloys exhibit higher strength and hardness than conventional 950 PdRu, but lower fracture strains. Tensile strengths and hardness values are similar to those of Pt or Au alloys. Fracture strains are comparable or somewhat lower in the annealed state, whereas they are higher after 75% cold work. Yield strengths may be somewhat lower or higher in the annealed condition, depending on the Al content. After cold working, however, the yield strengths of the PdAlRu alloys remain largely below those of the Pt or Au alloys – another clear manifestation of the low work-hardening rate, i.e. the high plasticity, of these ternary alloys.
Young's modulus, E, and Poisson's ratio, ν, are listed in Table II. These elastic properties were deduced from measurements of the longitudinal and transverse sound velocities (15). The pulse-echo measurements were performed on plates 2 mm to 3 mm thick, using appropriate transducers to excite either the longitudinal (at 10 MHz) or the transverse (2.5 MHz) acoustic mode. The Poisson's ratio of 0.37 is typical for precious metals. The Young's modulus of 139 GPa to 145 GPa is comparable to that of the conventional PdRu alloy (148 GPa). It lies between those of 18 carat Au alloys (90 GPa to 110 GPa) and 950 Pt alloys (approximately 170 GPa to 210 GPa). Regarding specific stiffness, E/ρ, the PdAlRu alloys with densities, ρ, in the range 10.5 g cm−3 to 11.6 g cm−3 slightly exceed Pt alloys (ρ ≥ 20 g cm−3) and clearly outperform Au alloys (ρ ≥ 15 g cm−3).
The CIELab colorimetric indices L* (lightness), a* (red-green chromaticity index) and b* (yellow-blue chromaticity index) of polished samples were determined using a spectrophotometric colorimeter (Konica Minolta CM-3610d spectrophotometer) (16). The measurement was carried out in a standard configuration with D65 illumination, a 10° observer, and in specular component included (SCI) mode. The ideal white would return (L*/a*/b*) indices of (100/0/0). The measured values are (84/1/4.5) for Pd95.5Al2.8Ru1.7 and (86/0.9/4.1) for Pd95.5Al0.9Ru3.6. These values are comparable to those of standard 950 PdRu, and closer to the colour of the platinum alloy 950 PtRu (87.7/0.7/3.4) than to ‘premium’ 18 carat white gold (82/> 1.5/> 6). However, the colour indices of the two PdAlRu alloys suggest that the effect of aluminium is to add a slight yellowish tinge and to somewhat diminish the brightness.
For the classification of white gold alloys, a simple colour grading system based on the ASTM D1925 (1988) yellowness index (YI) has recently been proposed (17). Within this system, the lower the YI the whiter the alloy. The whitest metals and alloys such as silver or 950 PtRu have values of YI ≈ 8. The PdAlRu alloys attain YI ≈ 10, which is comparable to pure Pd or Pt. White gold alloys, in contrast, have substantially higher indices, at YI ≥ 15.
Figure 4 shows another characteristic feature of the two PdAlRu alloys: their yield strengths do not steadily approach their tensile strengths with increasing cold work. Rather, the gap between the two parameters remains relatively large, thus facilitating the forming of complex and asymmetric shapes.
The good workability of the two alloys was confirmed by the fabrication of watch cases, backs and bezels by employing rolling, stamping and annealing operations. Plate ingots with surfaces machined to eliminate possible microcracks and flaws were used as a starting material. The plates were easily rolled to over 90% reduction without intermediate annealing. In accordance with the data in Figure 5, the hardness rose by only about 40 HV upon increasing the cold work from 23% to 95% reduction. Blanks were then roughly punched out from bands of thickness 8.8 mm, followed by fine punching for improved surface finish and dimensional tolerances. The final shaping of larger series of components by progressive die stamping is in progress.
The most intriguing observation during these operations was the pronounced tendency of both alloys to heat up considerably during plastic work. While heating to a certain extent is usual for rolling processes, the heating up of a disc during punching to temperatures so high that it cannot be touched by hand is extraordinary. This significant temperature rise during plastic deformation might be related to the low work hardening by promoting dynamic recovery.
Use of the Ceramic Fusion Technique with PdAlRu Alloys
Inspired by the dental technique of ceramic veneering of precious metals, the feasibility of fusing coloured ceramic overlays on PdAlRu alloys for decorative purposes was investigated. In the dental technique, Pd-containing alloys are in fact preferred. Pd oxidises more readily than Au or Pt, which guarantees a better bonding to the ceramic. The presence of Al in the PdAlRu alloys is certainly favourable in this respect. The dental sector commercialises a broad range of ceramic materials with coefficients of thermal expansion (CTE) in the range 8 × 10−6 K−1 to 14 × 10−6 K−1 (18, 19). The CTE of Pd is 12 × 10−6 K−1, and alloying with a total of 5 wt.% Al and Ru is not expected to change this value significantly. A close match of the ceramic CTE to the metal CTE is important to avoid cracking, notably during cooling after the firing process.
Two types of commercial dental ceramics were tested: VITA VM®13 veneering material for intensive or translucent colours, and the VITA Akzent® stain powder for pitch black dyeing (19). The ceramics were either applied as overlays or filled in to trench patterns machined into the PdAlRu discs.
The ceramic-to-metal fusion was performed following the directions of the ceramic supplier (19). In short, it consists of preparing the metal surface by sandblasting and controlled thermal oxidation. Different ceramic layers are then applied and fired one after the other at 890°C for one to two minutes: a first layer to promote cohesion, a second opaque layer, and a final glass-ceramic coloured layer. Additional layers may be necessary in order to fill in possible gaps produced upon firing.
Figure 7 exemplifies PdAlRu discs prepared and polished by the methods described above, with differently coloured ceramic inlays. The colours are uniform and no pores or cracks are apparent.
In search of 950 Pd alloys with improved mechanical properties, PdAlRu alloys proved particularly promising. The PdAlRu alloys presented in this paper possess beneficial characteristics for applications in jewellery, and in particular in watchmaking. The palladium content of 95 wt.% is common to most countries. The PdAlRu alloys are whiter than most 18 carat white gold alloys. Furthermore, they are compatible with the dental veneering technique, which opens up the potential for decorating articles with ceramic ornaments in appealing colours. The PdAlRu alloys exhibit excellent workability and forming characteristics, similar to those of commonly used 950 Pd alloys. At the same time, they exhibit much higher strength and hardness, more comparable to those of gold or platinum alloys. Moreover, the mechanical properties can be tuned in an extended range by varying the Al:Ru ratio. Upon cold working, for a given strain, the yield stress increases much less than it does in other precious metals, while the tensile strength increases in broadly similar fashion. This characteristic imparts to the alloys enhanced plasticity and excellent workability.
- S. A. Forrest and B. Clarke, ‘End-Users, Recyclers and Producers: Shaping Tomorrow's PGM Market and Metal Prices’, in“International Platinum Conference ‘Platinum Surges Ahead’”, Sun City, South Africa, 8th–12th October, 2006, Symposium Series S45, The Southern African Institute of Mining and Metallurgy, Johannesburg, South Africa, 2006, p. 307 LINK http://www.saimm.co.za/events/0610platinum/downloads/307-316_Forrest.pdf
- B. Libby, ‘Palladium Premieres’, MJSA Journal, March 2006, p. 35
- “The Santa Fe Symposium on Jewelry Manufacturing Technology 2008”, ed.E. Bell, Proceedings of the 22nd Symposium in Albuquerque, New Mexico, U.S.A., 18th–21st May, 2008, Met-Chem Research Inc, Albuquerque, New Mexico, U.S.A., 2008 LINK http://www.santafesymposium.org
- Kitco, Inc, Past Historical London Fix: http://www.kitco.com/gold.londonfix.html(Accessed on 3rd July 2009)
- W. Hume-Rothery, R. E. Smallman and C. W. Haworth, “The Structure of Metals and Alloys”, 5th Edn., The Metals and Metallurgy Trust, London, U.K., 1969, 407 pp
- G. Beck, in“Edelmetall-Taschenbuch”, 2nd Edn., ed.A. G. Degussa, Hüthig-Verlag, Heidelberg, Germany, 1995, p. 217
- The PGM Database:http://www.pgmdatabase.com(Accessed on 3rd July 2009)
- J. R. Hirst, M. L. H. Wise, D. Fort, J. P. G. Farr and I. R. Harris, J. Less-Common Met., 1976, 49, 193 LINK http://dx.doi.org/10.1016/0022-5088(76)90035-7
- J. Evans, I. R. Harris and L. S. Guzei, J. Less-Common Met., 1979, 64, (2), P39 LINK http://dx.doi.org/10.1016/0022-5088(79)90186-3
- A. Blatter, J. Brelle and R. Ziegenhagen, PX Holding SA,‘Alliage à Base de Palladium’,Swiss Appl.CH00032/08; 2008
- P. Battaini, 8853 SpA, ‘High-Hardness Palladium Alloy for Use in Goldsmith and Jeweller's Art and Manufacturing Process Thereof’,Italian Appl.TO2006/0086; U.S. Appl.2008/0,063,556
- H. W. King, Bull. Alloy Phase Diagrams, 1982, 2, (4), 527 LINK http://dx.doi.org/10.1007/BF02876187
- F. M. Smits, Bell Syst. Tech. J., 1958, 37, 711
- R. Hill, “The Mathematical Theory of Plasticity”, Oxford Classic Texts in the Physical Sciences, Oxford University Press Inc, New York, U.S.A.,1998, 366 pp
- “Nondestructive Testing Handbook”, Volume 7, “Ultrasonic Testing”,eds.A. S. Birks, R. E. Green Jr. and P. McIntire, American Society for Nondestructive Testing, Columbus, Ohio, U.S.A., 2007, 600 pp
- “Precise Color Communication: Color Control from Perception to Instrumentation”, Product Applications, Konica Minolta Sensing Inc, Japan, 1998: http://www.konicaminolta.com/instruments/knowledge/color/pdf/color_communication.pdf(Accessed on 3rd July 2009)
- S. Henderson and D. Manchanda, Gold Bull., 2005, 38, (2), 55 LINK http://www.goldbulletin.org/assets/file/goldbulletin/downloads/Henderson_2_38.pdf
- Wieland Dental online: Products: Veneering Ceramic:http://www.wieland-dental.de/produkte/verblendkeramik/page.html?L=1 (Accessed on 3rd July 2009) LINK http://www.wielandprecisiontechnology.de/de/ips-emax-cad
- VITA Zahnfabrik website:http://www.vita-zahnfabrik.com/(Accessed on 3rd July 2009)
Julien Brelle graduated in Materials Science and Engineering from the École Polytechnique Fédérale in Lausanne, Switzerland (2005), with a specialisation in metal matrix composites. He is now working as a Research Engineer at PX Group, a producer of metal products for the watch, jewellery and medical sectors. He is mainly involved in the development of speciality alloys and related processing.
After his Ph.D. in Physics (1986), Andreas Blatter led a research group at the Institute of Applied Physics in Berne, Switzerland, and spent a year at the IBM Almaden Research Center, U.S.A, as a Visiting Scientist. His research was focused on non-equilibrium laser processing, thin films and metallic glasses. Since 1996, he has been the R&D Director of PX Group. His main research topics include precious metals and speciality alloys and their related technologies, as well as corrosion and biocompatibility studies.
René Ziegenhagen received his degree in Materials Science and Engineering from the École Polytechnique Fédérale in Lausanne, Switzerland (1986). He was then involved in the research of precious metals and the development of new industrial processes, such as metal injection moulding and forging, before joining Cartier in the Richemont Group as a Senior Project Manager. At Cartier, his main concerns include the quest for new materials and new production technologies to meet requirements and regulations on biocompatibility and ecotoxicity.