A lkibological Study of Electrodeposited Gold-CoppercCadmium by Benedetto Bozzini ,a Carla Martini,b Ameriga Fanigliulo,a and Francesco BoganP aINFM, Dipartimento di Engegneria bInstituto di Metallurgia, Universita
dell’Innovazione, Universita di Lecce, v. Arnesano, di Bologna, vl. Risorgimento, Bologna, Italy
he electrodeposition of gold-copper-cadmium alloys was proposed1 as a process allowing the plating of thick layers of gold-alloys with sound mechanical properties, which could not be achieved with free-cyanide gold-copper baths. Electroforming of gold-copper-cadmium alloys currently holds an appreciable share (40%) of the global amount of gold electrodeposited for jewelry applications (or about 6% of the total amount of electrodeposited gold) world-wide.2 Electroplated layers of these alloys were described as promising materials for electric/electronic applications for contacts subjected to sliding, fretting, or repeated insertions.3 For obvious reasons, tribological behavior is critical for the functional performance in both kinds of applications. Notwithstanding, the relatively extensive literature on this system, structural and functional characterization of the electrodeposited material is rather inaccurate and naYve.1>4y5 In addition to this problem, it is necessary to stress that, even though the process met considerable industrial acceptance in the relevant niche, many process control/optimization problems still remain unsolved and render the electrodeposition of these alloys of commercial quality difficult and subject to considerable empirism. As far as experience in our group with this alloy is concerned,6 we have reasons to doubt that the samples investigated thus far have a controlled and reproducible structure and compositional distribution. The present investigation rests on previous work on the electrochemical fundamentals of this process and on the analysis of the structure of the deposits,6 which allowed the preparation of compositionally homogeneous and reproducible samples. In this article we report on mechanical ad tribological characterization of 14-, 18-, and 23-kt gold thick layers. The mechanical properties were measured by Vickers indentation (VI) and by recording hardness measurements (RHM) with a Vickers indenter and load/displacement monitoring. The tribological behavior was studied by dry sliding tests with a slider-on-cylinder apparatus and TiN countermaterial. Tribological failure mechanisms were assigned and correlations proposed between mechanical properties and wear damage.
Gold-copper-cadmium alloys were deposited galvanostatically from the following bath: Gold [as Kau(CN)21 2.5 g/L, copper [as K&u(CN)~I 60 g/L, cadmium [as KCd(CN),l 2.5 g/L, KCH 25 g/L, 70°C pH 11 at current densities (c.d.12 mA/cm2 (23 kt), 10 mA/cm2 (18 kt), 20 m&m2 (14 kt). Cathode current efficiencies were in the range 75 to 83% (anticorrelated with cd.). The coatings were not subjected to heat treatments since? (i) the correct content of cadmium ensures the absence of incorporated hydrogen; (ii> correct compositional control makes metallurgical homogenization unnecessary Prismatic brass samples of dimensions 0.5 x 0.5 x 0.4 cm3 were precoated with 5 to 7 pm of Watts nickel and plated with gold-copper-cadmium in a flowcell. The flow-cell, whose sample holder is sketched in Figure 1, allows optimal mass transport,6 and c.d. distribution homogeneities. The composition of each investigated sample was checked by EDS and the compositional scatter of samples prepared under the
Figure 1. Sample holder for the alloy electrodeposition flowcell. a: 6 cm; b: 0.75 cm; c: 0.5 cm; d: 4 cm; e: 9 cm; f: 1 cm; g: 0.5 cm; h: 1.5 cm; 1, 3: PVC screens; 2: samples; 4: anodes; 5: cathodic sample holder; A: front view; B: side view.
p l a t e d eelSt r distribution aeress your elrenlt beard.* With a S E R F I L C O Ser-Ductor ® air-free agitation system in your acid copper, nickel and gold plating tanks, you can turn up the power without fear of burning! same conditions gave rise to carat standard deviations of less than 2%. The crystallographic structure of all samples was studied by XRD. The roughness was measured by laser interferometry. Samples of two kinds of thickness were deposited: (i) about 10 pm for mechanical measurements by microindentation; (ii) a b o u t 100 ~m for crack-arrest fracture toughness and tribological testing. Mechanical properties were evaluated both by Vickers microhardness and RHM. Crack-arrest fracture toughness was estimated by the modified Palmqvist method. 7 The tribological behavior was investigated by dry sliding tests with a slider-on-cylinder tribometer. The tests were carried out in laboratory air at room temperature against a PVD-TiN-coated H S S cylinder. Friction was continuously measured, using a bending load cell and recorded as a function of the sliding distance. Values of the wear scar depths and widths on the slider and the cylinder were evaluated by stylus profilometry. Wear Tracks were examined by SEM and OM.
Air-free agitation provides better, more even distribution of the deposit without introducing air bubbles that result in mouse bites and costly rejects. Dirt particles are kept in suspension where they can be more efficiently removed by the filtration system.
RESULTS A N D D I S C U S S I O N
Ser-Ductor air-less agitation: • Provides increased plating speed • Improves throwing power • Prevents chemical and temperature stratification • Permits focused directional agitation
Composition and Structure Wide-scan (2,500-1,000 ~m 2) EDS analyses are reported in Table I (three i n d e p e n d e n t m e a s u r e ments on the surface of each sample). Quite acceptable compositional homogeneity within each sample 10
* Prove it to yourself. Call us today for details on SERFILCO's Performance Guaranty Program. Call us todav for details and see our catalog "on-line" at www.serfilco.com
!, i /'
08 10mAcrn'2 ~¢: .
':: }i [;
")'~,: O0 37
800-323-5431 • FAX: 847-559-1995 e-mail: sales @serfilco.com
20' (Cu Ka) D Figure 2. X-ray diffractograms for Au-Cu-Cd alloys electrodeposited at different current densities.
A-10D Circle 052 on reader card or go to www.thru.to/webconnect
(cross-sectional EDS measurements) and for replicated samples were obtained. Results of XRD analyses of the investigated samples are shown in Figure 2 and Table II. The structure consisted of a series of solid solutions, coherently with w h a t was observed in Ref. 6. P h a s e content was estimated from average composition by EDS and V6gard-law composition, s Crystallite dimensions D were e s t i m a t e d with the Scherrer equation, s No significant differences were observed between thin and thick samples. Average (R a) and root-mean-square (Rrms) values of roughness, as measured by laser interferometry, are reported in Table III. The overall morphology of these samples confirms what was also observed by SEM and AFM, 6 a positive correlation b e t w e e n roughness and thickness can be noticed, as expected for the onset of morphological instabilities for prolonged electrodeposition.
Mechanical Properties Mechanical properties were m e a s u r e d by Vickers microindentation with loads of 25 and 100 g applied for 10 sec and by RHM, by applying a linear ramp of load up to 1 N in 30 sec and holding at the maxim u m load for 10 sec in order to measure strain-rate sensitivity at room temperature. Crack-arrest fracture toughness was estimated with a Vickers indenter with a load of 15.6 g for 14- and 18-kt alloys, no cracking of 23-kt alloys could be observed with loading up to 30 kg; therefore, no quantification of Kic0 for this composition can be given. The results are reported in Table IV. Young's modulus (Y) values were obtained by RHM with an Oliver fit of the starting portion of the
unloading curve. 9 It can be observed the R H M is sensitive to the roughness of the sample (variances are positively correlated with roughness data) and might lead to underestimates because the Oliver fit does not take into account the finite curvature of the relevant part of the unloading curve and assigns a lower slope by performing a linear interpolation. Nevertheless, Y data are in the literature range for metallurgical gold-copper alloys. 1° No reliable estimates of Y could be obtained from the available VI data, because the standard deviation was so high (relatively high u n c e r t a i n t y of indentation depth estimate n) that no statistically meaningful difference among the alloys could be assessed, the estimates tend to group around 120 Gpa. Hardness data were obtained by VI and RHM. The results obtained with the two techniques do not show significant differences. A predictable increase of hardness, due to solid solution hardening, is noticed as a function of alloying element content; quantitative values are close to literature ones for similar alloys. 3-5 The plasticity index 6 = eel/(~el + epl) was evaluated by VI as described in Ref. 11 and RHM by directly measuring the penetration depth and applying the form function for the relevant indenter. 12 Quantitative differences emerge between the two methods. The reasons for this difference, which does not affect the observed trend of 6 versus composition, can be twofold: (i) the estimation of indentation depth by VI is definitely less accurate than by RHM; (ii) the extent of plasticization is much larger in VI. Stain-rate sensitivity was estimated by fitting the Metal Finishing
room-temperature creep curves under 1 N with the equation: dh/dt = (0.041 P/h2)” where h is the indentation depth and the constants are derived from the form function. The stain-rate sensitivity is not significantly affected by the composition. A slight trend can be obtained towards lower values for lower carat alloys. Film-substrate adhesion energies can be estimated from RHM loading curves displaying portions with a flat first derivative corresponding to indenter penetration depths of about 0.10 to 0.14 of the film thickness.g Lower-carat alloys tend to adhere more strongly to the Watt-nickel substrate. An approximate value for the crack-arrest fracture toughness was obtained with the modified Palmqvist technique,7 the rigorous procedure could not be applied, since it would have implied the production of several thick alloy samples. Tribological Behavior Two kinds of tests were carried out: (i) at low-loads
(2.5 and 5 N, sliding speed 0.3 m/set, sliding length 1,000 m) in order to obtain a reliable wear-scar depth and wear track morphology without digging through the depth of the coating; (ii) higher loads (7.5-17.5 N, same sliding speed, varying sliding lengths chosen in order not to wear the coating down to the substrate, typically about 100 m> to check for possible load-dependence of the wear mechanism. Results of low-load tests are reported in Table V. In Figure 3 we report typical SEM micrographs of wear-scars. The wear-tracks are homogeneously ploughed by sliding against the asperities of the counter-material (Fig. 3A), a limited amount of debris are left on the wear-tracks (Fig. 3B), while a considerable amount tends to be compacted at the entrance of the wear-scar. The morphology of the wear-tracks, characterized by the presence of debris, is suggestive of a microcutting mechanism. The fric-
New Surface Technologies manufactures a broad line of high-quality electroplating chemistries... manufactured to stringent quality control standards. We are specialists in zinc and chromate conversion coatings, particularly our new trivalent high corrosion variety. What’s more, NST offers technical support for every product we sell.
Among the products we offer are: SURFACE PRETREATMENT PRODUCTS Soak Cleaners Electra-Cleaners . Spray Cleaners l
PLATING PROCESSES, FUNCTIONAL Cyanide Zinc Chromate Conversion Coatings-Conventional Alkaline Non-Cyanide Zinc Trivalent Conversion Chloride Zinc Coatings Alloy Plating Processes Topcoats Cadmium Processes Acid Tin l
PLATING PROCESSES, Cyanide Copper Alkaline Non-Cyanide Copper Acid Copper
DECORATIVE Semi-Bright Nickel Bright Nickel Chrome Plating Additives
PHOSPHATE CONVERSION COATINGS Iron Phosphate, Spray and Immersion Zinc Phosphate, Spray and Immersion Rust Preventive Oils Corrosion Resistant Paints
WASTE TREATMENT PRODUCTS Polyelectrolytes both Anionic and Cationic Coagulants Defoamers l l
NewSurfaceTechn~l&i& 28 Industry Drive
Phone: (440) 439-5790
Bedford, Ohio 44146 l
Fax: (440) 439-5794
Circle 041 on reader card or go to wwwthru.to/webconnect January 2003
tion of these effects, but does not follow the overall trend of wear data. Y values are not correlated with other mechanical properties of the investigated materials (CC in the range 0.51-0.61), while KIcO correlates with plastic properties (CC in the range 0.95-l.OO), but display a different trend, An acceptable correlation of the trned of wear scar depth can be obtained by including KIco in an Archard-type relation W = A x P/(HV x K,,,), even though a quantitative analysis cannot be performed because of the lack of values for high-gold alloys. No improvements are obtained by taking into account Y. Similarly, critical load data for adhesive failure Lcrit at higher loads can be correlated (Lcrit = eAN, D 2 = 0.98) with the ratio of the system quantity substrate-film adhesion energy SEE MS to the materials p roperty Young modulus, expressing the coating/substrate load transfer efficiency. CONCLUSIONS
f3 Figure 3.Typical SEM micrographs of wear tracks at low loads (Au-CL&d alloy electrodeposited at 10 mA/cm*; load, 2.5 N; sliding speed, 0.3 m/set; sliding length, 1,000 m); A: exit slide showing wear scars; B: close-up of A showing wear debris.
tion coefficient invariably reached a high value (about 0.9) after a short transient. Short tests (50 to 100 m) were run at increasing leads (steps of 2.5N) and the surface was checked by OM for wear mechanism; no reliable wear-scar depths could be obtained with these measurements. We could observe that, irrespective of the wornthrough depth (typically 25 to 75% of the total thickness), a critical load exists above which the coating tends to detach from the substrate. We tried to correlate the two observed wear mechanisms with mechanical properties of the deposits. The starting point of the analysis of low-load data was the Archard equation. In the present case, all the plasticity-related quantities (HV, 6, n) are strongly correlated [absolute values of correlation coefficients (CC) in the range 0.98-1.00); HV in the classical Archard formula accounts for the combina46
The relationships between mechanical properties and tribological behavior of 14-, 18-, and 23-kt goldelectrodeposited alloys were copper-cadmium reported on. In order to obtain significant results, compositional homogeneity of the samples must be ensured: this was achieved by growing the alloys under controlled hydrodynamic conditions. Two tribological failure mechanisms were observed: at low loads the alloys fail by microcutting, at higher loads coatings tend to detach from the substrate. The former mechanism is typical for the bulk material, the latter one depends on the coating-substrate system. Material loss brought about by microcutting can be correlated with an Archard-type relationship corrected for toughness effects. Critical loads for coating detachment depend on deposit-substrate adhesion energy. BIOGRAPHIES
Benedetto Bozzini is Professor of Applied Physical Chemistry at the University of Lecce, Italy. He holds an M.Sc. in Nuclear Engineering (Politecnico di Milano, Italy), Ph.D. in Electrochemical Engineering (politecnico di Milano, Italy), postdoctorate in electron spectroscopies (National Physical Laboratory, Teddington, Middlesex, UK). Carla Martini is Senior Lecturer at the University of Bologna, Italy, with a M.Sc. in Industrial Chemistry (Politecnico di Milano, Italy), PhD. in Metallurgy Ameriga Fanigliulo is a Ph.D. candidate in Materials Engineering,Universityof Lecce,Italy,and holds a M.Sc.in Chemistry(universiti di IV&no, Italy). Francesco Bogani is a Research Engineer at the Metal Finishing
of Lecce. He holds a B.Sc. in Agricultural
1. Dettke, M. et al., Galucznotechnik, 62:773; 1971; 63:729; 1972 2. Bozzini, B. and P.L. Cavallotti, “Stat0 dell’arte sull’elettrodeposizione di preziosi in Italia,” Atti Seminario Assotec, Milan0 (I); 1999 3. Saxer, W., Galvanotechnik, 82~3427; 1991 4. Steinmann, S. and W. Fliimann, Metalloberfltiche, 29:x4; 1975 5. Robert, J.J., Galvano-Organo, 46:33; 1976 6. Losa, E., “Elettrodeposizione caratterizzazione di AuCu-Cd,” Tesi di Laurea in Ingegneria Nucleare, Politecnico di Milano, AA; 1998-99 7. Bozzini, B. and M. Boniardi, 2.f: Metallkunde, 88:493; 1997 8. Warren, B.E., “X-Ray Diffraction,” Dover Publications Inc., New York; 1990 9. Oliver, W.C. and G.M. Pharr, d Material Rea, 7:1564; 1992 10. “Edelmetalltaschenbuch,” Degussa AG, G. Beck et al., Eds., p. 210; Hiithig Vlg. Heidelberg; 1995 11. Bozzini, B. et al., Comp. Sci. Technol., 59:1579; 1999 MF 12. Brotzen, F.R., Intern. Mat. Rev., 39:24; 1994
A Complete Line of ’
PLATINUM CMil’ED Platinum Coated Titanium And Platinum Clad Columbium
our readersan added, on-line service for locating product information: ONE WEBSITE ADDRESS that lets you connect to MF advertiser websites. Metal Finishing’s new WebConnect service will let you link directly to the websites of suppliers with.... l l l l l
NO looking up or Typing in Multiple URLs NO Futile Searches NOSpam NO Cookies
With Mf’s WebConnect you’ll be able to research product and technology before requesting information. AND, you can connect to websites by company name or product type. One easy link to bookmark URL - www.thru.to/webconnect
FOR GOLD and RHODIUM PLATING Available in expanded mesh or solid sheet in a variety of standard stock sizes or made to order. FOR COPPER-NICKEL-CHROME Designed to meet your requirements plating thickness and long life. WE CAN RECOAT YOUR PRESENT ANODES Call us for price and delivery information.
Write for free pfoduct information on platinum coated and platinum &d anodes. www.plateworld.com/intrepid.htm
Intrepid Industries Salem Industrial Park-Bldg. 9 291 Route 22 East, Suite #3 Lebanon, NJ 08833 l
Circle 028 on reader card or go to www.thru.to/webconnect January 2003