- Email: [email protected]
Tribology and sliding electrical contact resistance of e-beam hard Au: Effects of annealing
Wear 376-377 (2017) 1662–1672
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
Tribology and sliding electrical contact resistance of e-beam hard Au: Effects of annealing J.E. Mogonye, N. Argibay, R.S. Goeke, P.G. Kotula, T.W. Scharf n,1, S.V. Prasad n Materials Science and Engineering Center, Sandia National Laboratories, Albuquerque, NM 87185-0889, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 1 September 2016 Received in revised form 19 January 2017 Accepted 20 January 2017
Nanocomposite Au-ZnO thin films in the dilute oxide (o 5.0 vol%) regime were synthesized by electron beam (e-beam) evaporation, as alternatives to electroplated Au hardened with Ni. Tribological measurements of e-beam hard Au were made while passing current through sliding contacts; electrical contact resistance (ECR) and friction data were simultaneously acquired during the test. The friction, wear and ECR behaviour were studied for the as-deposited film condition, and after annealing at 250 °C and 350 °C in air. The study revealed that the 250 °C annealed Au-2 vol% ZnO film exhibited the lowest, stable friction coefficient s (m 0.25) and ECR ( 35 mΩ) during sliding. Furthermore, the wear rate of this 250 °C annealed ZnO hardened Au nanocomposite film was an order of magnitude lower at 1.5 10 5 mm3/N m than for a typical Ni hardened, electroplated Au film at 1.3 10 4 mm3/N m. Crosssectional transmission electron microscopy studies inside the wear surfaces revealed that the extremely stable, low friction coefficients and wear rate of annealed Au-2 vol% ZnO film was due to partial coverage of the wear surface with a ZnO tribofilm that reduced the adhesive contact contribution to wear with minimal impact on ECR. The potential implications of this study in the search for an environmentally friendly alternative to widely used electroplated hard Au are discussed. & 2017 Elsevier B.V. All rights reserved.
Keywords: Electroplated hard Au Au-ZnO film Sliding electrical contact resistance Sliding wear Tribofilm Transmission electron microscopy
1. Introduction The necessity to provide efficient transfer of electrical current between moving parts remains an engineering challenge. With increasing implementation of electrical motors and generators and further increase in electrical contacts with the expanding electronics industry, the technological applications of sliding electrical contacts has become widespread. Sliding electrical contacts can be found in DC motor and generator current collector slip rings, printed circuit board edge connectors, data cable pin connectors and receivers, microelectromechanical systems (MEMS), circuit relays and switches, instrument signal transfer slip rings and many more applications. The main engineering challenges associated with these applications, especially long service life slip rings, is minimized wear and stable wear performance, acceptable friction to minimize mechanical energy losses, low and stable electrical contact resistance (ECR) to minimize Joule heating, and thermal stability of the sliding materials. The use of gold in the electronics industry has reached over n
Corresponding authors. E-mail addresses: [email protected] (T.W. Scharf), [email protected] (S.V. Prasad). 1 On faculty sabbatical from the Department of Materials Science and Engineering, The University of North Texas, Denton, TX 76203-5310, USA. http://dx.doi.org/10.1016/j.wear.2017.01.080 0043-1648/& 2017 Elsevier B.V. All rights reserved.
300,000 kg per year in 2010 [1]. According to an evaluation by Goodman in 2002, when gold use in the electronics industry was 280,000 kg per year, the majority of this gold was used in electrical contacts, most frequently electroplated gold [2]. Gold is used in sliding electrical contacts due to its high conductivity and excellent corrosion and oxidation resistance. Although electroplated hard gold (gold alloyed with typically 1–2 wt% of cobalt or nickel) has produced an improvement over pure gold plating in friction and wear performance, it has reliability issues with ECR over time and at elevated temperatures [3]. Additionally, there are growing concerns with environmental hazards associated with the use and disposal of cyanide and arsenic baths used in plating chemistries. The reliability issue is related to an increase in ECR due to the formation of surface oxide films that can cause system failure in low voltage and high frequency applications, especially systems which have expected service lives on the order of years. Improvement of mechanical and tribological performance in hard Au has been realized by co-depositing with ZnO to synthesize ZnOhardened Au nanocomposite films [4–7]. It has been determined that the room temperature sliding friction and ECR in the dilute oxide regime (o 5 vol% ZnO) are lower than Ni hardened gold films [4]. However, it is unknown if annealing these films results in any improvements in the sliding electrical contact properties. As such, this paper presents results on friction, ECR and wear properties of as-deposited and annealed composite Au-ZnO films with
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1663
comparisons made to as-deposited and annealed Ni hardened Au films. The sliding-induced structural and chemical changes of the worn surfaces and subsurfaces that control these properties will also be presented and discussed.
2. Experimental methods 2.1. Electron-beam film deposition A Thermionics Vacuum Products Co. 10 kV Triad e-beam evaporation system was used for the deposition of both pure Au and composite Au-ZnO films. The electron beam is generated by a field emission source and magnetically deflected to the crucible to provide localized heating resulting in the sublimation of the film material. Once sublimated in a vacuum, the film material deposits on the target substrate in line-of-sight from the crucible. Prior to all film depositions the vacuum chamber was pumped to a base pressure of less than 1 10 6 Torr. The film material source to substrate distance was 305 mm and all depositions were conducted at room temperature. A rotating substrate mount was used to achieve compositional and thickness uniformity of the films. Prior to e-beam deposition, the Si substrates were first sputter coated with a nominally 300 nm thick Ti layer for adhesion followed by a nominally 300 nm thick Pt layer to act as a diffusion barrier. These layer thicknesses were determined by cross-sectional microscopy. For comparison purposes, electroplated hard Au films were obtained from a commercial plating company (Theta Plating Inc. Albuquerque, NM, USA). Nickel was used as the hardening element and plated per ASTM Type I specifications (ASTM B488-11 standard), also comprising of a 5 μm thick electroplated Ni diffusion barrier, deposited on Alloy 52 substrates (nominally a 50–50 wt% Fe-Ni alloy). Thickness of electroplated Ni hardened gold (Ni HG) film was 1.5 μm. The pure e-beam Au films, nominally 2 mm thick, were deposited by evaporation of high purity Au pellets (99.999% purity) sourced from Materion Advanced Chemicals. The resultant Au film average root-mean-square (RMS) surface roughness was determined by scanning white light interferometry (SWLI) to be 3 nm. The composite Au-ZnO films, nominally 2 mm thick, of varying composition were also deposited onto Si substrates sputter coated with Ti/Pt layers. The e-beam deposition source materials were pure Au pellets, mentioned above, and ZnO tablets (99.9% purity) from Materion Advanced Chemicals. The deposition rates were varied between 0.00 and 2.00 nm/sec to achieve the desired volume fraction of ZnO in the film in the dilute oxide (o 5.0 vol%) regime. The resultant film average RMS surface roughness of composite Au-ZnO films was determined by SWLI to be 4 nm. Prior to tribological testing, some of the pure Au and composite Au-ZnO films were ex situ annealed in an open air furnace to either 250 °C or 350 °C for 24 h dwell time. 2.2. Sliding electrical contact testing A custom built linear tribometer equipped with an Agilent Technologies B2911A digital source/meter, photograph and accompanying schematic, shown in Fig. 1, was used to conduct all simultaneous friction and ECR measurements. The tribometer independently measures applied normal force and lateral force through independent double leaf flexures, which constrain the displacement to in plane linear motion, coupled to load cells. The load cells were independently calibrated using dead weights and calibrated scales throughout the expected experimental loads, thus incorporating the leaf flexure spring rate into the calibration. Continuous ECR measurements are made by operating the source/ meter in voltage-regulated remote sensing mode and measuring
Fig. 1. Photograph and accompanying schematic of custom-built linear ECR tribometer.
the voltage drop and current using a four-point probe bridged across the film surface and counterface pin. Both the substrate of the film and the counterface pin are electrically isolated from the rest of the rig by custom fixtures machined from polycarbonate and polyether ether ketone, respectively. All tribological friction ECR tests were conducted in unidirectional sliding motion by translating the bottom fixture by the 2 mm track length while recording data then lifting the upper load cell assembly and translating the bottom fixture to the start position and repeating. A normal force of 100 mN and a linear speed of 1 mm/s were used for a duration of 100 sliding cycles. Prior to each sliding ECR experiment, a static normal force of 100 mN was applied and a voltage set to produce approximately 100 mA of direct current through the ECR circuit. All tests were conducted at room temperature in lab air (15–30% relative humidity). The
1664
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
counterface pin was made from Neyoro G alloy wires, a hard, primarily Au-Cu alloy with nominal composition of 72Au-14Cu8Pt-5Ag % by weight, sourced from Deringer-Ney, Inc. The Au-Cu alloy wires were first annealed to a hardness of approximately 1.96 GPa and then hand machined on a lathe to produce a hemispherical tip with a 1.59 7 0.05 mm radius of curvature, in order to produce a nominally circular (Hertzian) contact geometry. The machined hemispherical tips were subsequently hand lapped using progressively finer diamond paste to a center line average roughness (Ra) of 100 nm or less measured by SWLI. Prior to sliding experiments, the polished riders were ultrasonically cleaned in isopropyl alcohol and blown dry with dry nitrogen. The top insulating fixture holds the Au-Cu alloy counterface pin at a canted angle of 20°, as shown in Fig. 1, so that multiple tests could be conducted on a single rider by rotating it about its cylindrical axis to a unique contact area. At least three repeat tests were conducted for each film.
Surface topography data of the films and counterface pins pre and post sliding ECR tribometry were collected using a Veeco NT1100 SWLI with either a 50 or 10 objective lens. The topographical height data was then analyzed using Bruker Vision64 version 5.41 software. Wear volumes of the pins and films, post sliding ECR testing, were measured as the missing volume displaced below a reference plane fitted to the unworn pin and film surface outside the wear track, respectively. These wear volumes were used to calculate the specific wear rates of the pins and films using Archard's equation and are reported in units of mm3/N m. 2.3. Structural and chemical characterization techniques Scanning electron microscopy (SEM) images inside and outside the wear track surfaces of the e-beam deposited Au and composite Au-ZnO films were collected using a FEI Helios 450 SEM. Transmission electron microscopy (TEM) specimens were also prepared
Fig. 2. Planar SEM images (left column) and cross-sectional STEM images (right column) of (a) as-deposited (ASD), (b) 250 °C annealed, and (c) 350 °C annealed Au-2 vol% ZnO films. Near surface SADP are shown in inset of STEM images. A higher magnification STEM image is shown in inset of (c).
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
using the same system that is equipped with a focused ion beam (FIB) column. A site specific Pt layer was deposited in the SEM to protect the film surfaces from ion-induced damage. The FIB milled TEM specimens were then removed from the bulk of the film using an ex-situ microprobe and transferred to C film coated Cu grid TEM specimen holders. The FIB was also used to mill cross-sections to the surface normal of the as-deposited, 250 °C and 350 °C annealed composite Au-ZnO films using standard ion milling techniques with the revealed cross-section surface polished by low energy (2 keV Ga þ ) ions to enhance grain boundary contrast for line-intercept measurement of grain sizes. Annular dark field scanning TEM (ADF-STEM) was used for Z-contrast imaging of these cross-sections along with acquiring selected area diffraction patterns (SADP). Site specific FIB samples were made close to the center of the film wear tracks parallel to the sliding direction. Sliding-induced changes to the structure and chemistry in the worn subsurface regions were analyzed with cross-sectional STEM and energy dispersive X-ray spectrometry (EDS) analyses, respectively. X-ray spectral images were collected from wear subsurfaces and then analyzed using Sandia's Automated eXpert Spectral Image Analysis (AXSIA) software [8]. AXSIA quickly reduces large raw spectral images to a compact solution consisting of a small number of linearly independent component image/ spectrum pairs. For display purposes, the component images were combined into a color overlay.
3. Results and discussion 3.1. Surface and subsurface structure Fig. 2 shows both planar SEM images and cross-sectional STEM images of (a) as-deposited (b) 250 °C annealed, and (c) 350 °C annealed Au-2 vol% ZnO films. From the planar SEM image in Fig. 2 (a), it is evident that the as-deposited Au-2 vol% ZnO film surface grains appear to be homogenously distributed with a measured grain size of 75 nm. There is no observable grain coarsening with annealing to 250 °C shown in the Fig. 2(b) planar SEM image. However, the planar SEM image in Fig. 2(c) shows that when the film was annealed to 350 °C, there are numerous grains that have coarsened to a measured grain size of 140 nm. Cross-sectional STEM imaging was also performed to further investigate the grain structure in the as-deposited and annealed Au-2 vol% ZnO films. The cross-sectional STEM image of the asdeposited film in Fig. 2(a) shows that the nanocrystalline grains are columnar and exhibit some growth twins. The inset SADP image shows that the Au grains are randomly orientated, i.e., have no in plane o 1114 growth texture. Prior high resolution STEMEDS studies on Au-2 vol% ZnO film in as-deposited condition revealed that ZnO appears to be well distributed, concentrated as a thin film at the Au grain boundaries and intergranularly dispersed as ZnO particles that are on the order of single of nm's in diameter [6]. The STEM and SADP images in Fig. 2(b) show no changes in the near surface grain structure and orientation with annealing to 250 °C. However, it was previously determined that there is Au grain coarsening, but only at the film/Pt layer interface since ZnO diffused to this interface to form a 5 nm thick film [6]. With absence of ZnO at the Au grain boundaries there was corresponding Au grain growth. In contrast to the as-deposited and 250 °C annealed films, the 350 °C annealed film exhibited more faceted, roughened surface Au grains, shown in Fig. 2(c), although the Au grains did not exhibit coarsening from the STEM image or texturing from the SADP image. Instead, surface crystallites formed in many locations at the Au grain boundaries that showed a much darker contrast in higher magnification STEM images, such as the one shown in the inset of Fig. 2(c). These crystallites were
1665
determined to contain Zn according to EDS analysis and appear darker in STEM since they are lower atomic number than the brighter Au grains. This darker contrast is similar to the aforementioned Zn at the Au grain boundaries. The higher magnification STEM inset image in Fig. 2(c) shows these surface crystallites are located at the Au grain boundaries suggesting intergranular diffusion of Zn to the surface. It is likely these crystallites are ZnO, which is supported by the many larger hexagonal-shape surface crystallites shown in the planar SEM image in Fig. 2(c) that correspond to the hexagonal ZnO Wurtzite crystal structure. Therefore, ZnO diffused to the surface to form crystallites up to 140 nm in size when annealed to a higher temperature of 350 °C. 3.2. Friction and ECR behavior Typical friction coefficient and sliding ECR results are shown in Fig. 3 for as-deposited pure Au and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films as well as type I (499.7 mass % Au) commercial Ni HG. The Ni HG started at μ 0.3 friction coefficient and experienced a short run in period of 20 cycles of increasing friction until entering a region of periodic friction response from cycles 30–50, and finally reaching a steady state friction coefficient value of μ 0.75 after cycle 80. The sliding ECR of the Ni HG was highly variable with values consistently exceeding 20 mΩ until approximately cycle 75 when the values stabilized at 10 mΩ. These exceptionally high ECR values in conjunction with the friction behavior prior to cycle 80 were repeatable over several experiments. High variability in ECR can be explained as being due to contact separation between the rider and film associated with the presence of wear debris in the track. In contrast to the Ni HG film, the pure e-beam Au film exhibited a rapid increase in friction coefficient to μ 1.5, commensurate with a highly adhesive and large area of contact until approximately cycle 45 before the friction coefficient variance increased. The initial cycles (o45) of ECR for the Au film retained a value of 16 mΩ and low variance in conjunction with the same relatively low variance region in the friction coefficient before increasing and becoming more variable per cycle. This change in friction coefficient and ECR beyond cycle 45 is due to adhesive wear throughout the thickness of the Au layer. The friction coefficient with increasing ZnO concentration from 0.1–2.0 vol% resulted in the trend of decreasing average friction coefficient and variance as well as more stable regimes of friction throughout the 100 cycles of sliding ECR testing. The average stable ECR values for the 0.1– 2.0 vol% films slightly increased with increasing ZnO concentration from 16 mΩ to 22 mΩ, respectively, and is likely due to the increase in resistivity of the films with increasing ZnO concentration discussed in the previous section. Overall the variance of the ECR has also significantly decreased for all of the dilute ZnO concentrations shown with the exception of a few points during the testing of the 2.0 vol% film. In the first 30 sliding cycles of the 2.0 vol% film, there are a few large spikes in ECR indicative of a large fraction of contact separation similar to the Ni HG. This ECR response was previously reported to be a result of ZnO third body debris generated at the onset of wear before being displaced from the wear track as seen in SEM and characterized by EDS and SWLI maps of the same ZnO 2.0 vol% concentration film [4]. In addition, the lowering of friction coefficient with increasing ZnO content may be affected by the nanoindentation hardness increase from 1 to 2.5 GPa as well as the elastic modulus increase from 76 to 83 GPa with increasing ZnO concentrations from 0.1 vol% to 2 vol%, respectively [6]. In summary, while the ECR values are relatively low, the friction coefficients remain high (μ 40.5) for all films in asdeposited condition. Table 1 summarizes the average and standard deviation of the friction coefficient and ECR values for the as-deposited films.
1666
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 3. Friction coefficients and ECR data per sliding cycle for as-deposited commercial Ni HG, pure Au, and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films sliding against a Au-Cu alloy rider. Shaded areas correspond to the average value, solid line, 71s (standard deviation) of data collected at 50 Hz.
Typical friction coefficient and sliding ECR results are shown in Fig. 4 for 250 °C annealed pure Au and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films as well as type I commercial Ni HG. The Ni HG exhibited a significantly reduced steady state friction coefficient of μ 0.25 and ECR of 18 mΩ in comparison to the as-deposited film. The run in period of slightly higher friction coefficient and ECR lasted for only the first 10 cycles before obtaining the low variance and steady state regime. It is likely that annealing of the
Ni HG at 250 °C generated greater stability in the Au microstructure by diffusion of Ni underlayer through the Au film along grain boundaries, a well-known phenomenon [9]. In addition, it is unclear if annealing resulted in significant Au grain growth or the formation of a NiO surface film, however the results indicate that the surface wear evolution led to a smooth and conformal contact without large third body wear debris that can cause the contact separation ECR response shown previously in Fig. 3. The pure
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1667
Table 1 Average and standard deviation, 1s, of coefficient of friction, CoF, and electrical contact resistance, ECR, values for each film composition and annealing condition. Averages were taken from every sliding ECR test and across all 100 sliding cycles. Composition
Ni HG 0.0% ZnO 0.1% ZnO 0.5% ZnO 1.0% ZnO 2.0% ZnO
250 °C annealed
As-deposited (ASD)
350 °C annealed
Avg. CoF 7 1s
Avg. ECR (Ω)7 1s
Avg. CoF 71s
Avg. ECR (Ω) 7 1s
Avg. CoF 7 1s
Avg. ECR (Ω) 7 1s
0.57 70.13 2.17 70.95 1.99 7 0.74 1.45 7 0.36 0.62 7 0.14 0.56 7 0.12
1.00E þ 027 1.82E þ 02 2.37E-027 2.23E-02 2.14E-027 1.05E-02 2.20E-02 7 4.54E-03 2.07E-027 1.30E-04 3.99E-02 7 3.67E-02
0.29 7 0.08 1.647 0.45 1.217 0.37 1.137 0.20 0.36 7 0.07 0.277 0.03
1.61E-02 7 2.36E-03 1.14E-02 7 5.65E-03 1.18E-02 71.11E-03 2.06E-02 71.62E-03 4.75E-02 7 7.77E-03 3.60E-02 7 4.09E-03
0.39 7 0.08 1.517 0.41 1.25 7 0.35 1.23 7 0.24 0.87 7 0.20 1.09 7 0.26
2.50E-017 2.53E-01 1.98Eþ 007 5.72E þ 00 1.55E-02 7 6.98E-03 1.99E-02 7 2.25E-03 2.46E-02 72.51E-03 2.25E-02 73.20E-03
e-beam Au film friction coefficient began in excess of μ 3.0 while the ECR remained stable near 15 mΩ throughout the test, and is a result of the Au film being completely (visible to the naked eye) detached from the Pt/Ti adhesion layers in the first few cycles. This is the result of gross adhesive wear and film adhesion failure likely due to large grain formation from annealing. The 0.1 vol% ZnO film exhibited high friction coefficient values although slightly lower than the pure Au film at the 250 °C annealed condition, and did not completely fail as did the pure Au film. The high friction is likely a result of Au grain growth due to a small percentage of ZnO in the composite film. The ECR values remained low similar to the as-deposited film. The 0.5 vol% ZnO film exhibited similar results to the as-deposited film condition although the variance in friction coefficient decreased and as in the other films the ECR values remained low. In addition, the trend of decreasing friction coefficient with increasing ZnO content continued. Both the 250 °C annealed 1.0 and 2.0 vol% ZnO films showed significantly reduced and stable friction coefficient values of μ 0.25. The ECR response for both films was also similar, increasing during a run in period over the first 10 cycles and slowly stabilizing to a value near 45 mΩ and 35 mΩ, respectively, over the duration of the tests. These ECR values are nearly twice that of the as-deposited values and could be a result of a larger concentration of ZnO at the surface due to annealing that is not forming large third body debris resulting in contact separation, as seen in the ECR behavior of the as-deposited 2.0 vol% film in Fig. 3. The extremely stable and low friction coefficient of these films corroborates the hypothesis that ZnO has migrated to the surface and is acting to minimize the adhesive contact contribution to friction. In summary, both the friction coefficient and ECR values are relatively low for the Ni HG and 1.0 and 2.0 vol% ZnO composite films. Table 1 summarizes the average and standard deviation of the friction coefficient and ECR values for the 250 °C annealed films. Typical friction coefficient and sliding ECR results are shown in Fig. 5 for 350 °C annealed pure Au and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films as well as type I commercial Ni HG. The commercial Ni HG did maintain stable friction coefficients of μ o0.5, however the ECR values of the first 12 cycles of sliding are two orders of magnitude greater than the 250 °C annealed film indicative of a thick and uniform NiO film. The ECR did reduce after initial wear since the surface oxide was likely worn away, although there were large spikes due to third body contact separation. Overall, these friction coefficient and ECR values for Ni HG are not as low and stable as the 250 °C annealed film. In the case of pure Au, the film completely failed and delaminated from the Pt/Ti adhesion layers in the first few cycles, similar to the 250 °C film. Both the 0.1 and 0.5 vol% ZnO films exhibited similar friction coefficient and ECR values as the corresponding 250 °C annealed films. However, the 1.0 and 2.0 vol% ZnO films exhibited a large increase in friction coefficients from μ 0.25 to μ 0.8 when compared to their respective 250 °C annealed films, albeit
slightly more stable ECR values near 25 mΩ, although there were a few points of contact separation during cycle 19 for the 2.0 vol% ZnO film. In summary, further annealing from 250 °C to 350 °C resulted in very high friction coefficient values for all the films. Table 1 summarizes the average and standard deviation of the friction coefficient and ECR values for the 350 °C annealed films. Overall when comparing the friction and ECR response in Figs. 3–5, the 2.0 vol% ZnO composite film was very comparable in performance to the Ni HG and had a more stable ECR in the asdeposited condition. In the annealed condition of 250 °C, the 2.0 vol% ZnO film exhibited a slightly lower and more stable friction coefficient with comparable ECR values. In addition, the 2.0 vol% ZnO film exhibited more stable ECR at 350 °C in comparison to Ni HG that showed the response of a thick homogeneous NiO film. 3.3. Wear behavior of films and counterfaces In addition to friction and ECR response, the wear rates of the films and counterface riders after the sliding tests were measured to determine overall tribological performance. Fig. 6 summarizes the specific wear rates of as-deposited, 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films. It is evident that the commercial Ni HG exhibited the lowest wear rate for the as-deposited condition with an average value of 9.5 10 5 mm3/N m. This is likely a result of a conformal oxide film acting to minimize adhesive wear, which is supported by the high ECR values in Fig. 3. The 0.1 and 0.5 vol% ZnO films exhibited similar, high wear rates to that of the large grain pure e-beam Au for all the as-deposited and annealed films tested. This result is interesting in that although the hardness of these films was greater than double that of the pure e-beam Au ( 0.8 GPa), the wear rates were of the same magnitude, suggesting that the higher area fraction of large grains, determined previously for these very dilute ZnO films [6], yields similar adhesive wear to the pure e-beam Au. In contrast, the 1.0 vol% ZnO film exhibited similar wear rates to the Ni HG except for an increase by an order of magnitude for the 350 °C annealed condition indicating a change in wear mechanisms for that film. In addition, the as-deposited 2.0 vol% film showed fairly high wear rates in comparison to what was reported previously [4]. This could be due to uncertainties in the ZnO concentration arising from the variability in the e-beam deposition process or other unknown factors. However, the 2.0 vol% film did exhibit a drastically different wear rate for the 250 °C annealed condition (1.5 10 5 mm3/N m), which was the lowest wear rate of all films, even an order of magnitude lower compared to 1.3 10 4 mm3/N m for Ni HG annealed at 250 °C. After annealing the 2.0 vol% film to 350 °C, the wear rate significantly increased by two orders of magnitude, and was again similar to the other ZnO films suggesting that there is a change in wear mechanism(s) between these two annealing temperatures that is likely surface structure and/or chemistry dependent. In
1668
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 4. Friction coefficients and ECR data per sliding cycle for 250 °C annealed commercial Ni HG, pure Au, and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films sliding against a Au-Cu alloy rider. Shaded areas correspond to the average value, solid line, 71s of data collected at 50 Hz.
another study on these films, the nanoindentation hardness of the 2.0 vol% film cycled to an annealing temperature of 350 °C five times was virtually unchanged from the film hardness of 3.5 GPa in as-deposited condition [7]. However, these hardness values may not allow a direct comparison to the films in the present study since the total anneal time was much lower, i.e., 2 h for five total cycles compared to the present 24 h dwell time. The negligible effect of the shorter annealing time on hardness is further substantiated by no noticeable grain growth or other
microstructural/chemical changes, which is in contrast to the formation of ZnO crystallites at 350 °C after 24 h of annealing shown in Fig. 2(c). Fig. 7 summarizes the specific wear rates of the Au-Cu alloy counterfaces sliding against as-deposited, 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films. Overall, the wear rates of the counterfaces are lower than the respective films, but the trends are in excellent agreement with the wear rates of the films for each condition tested
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1669
Fig. 5. Friction coefficients and ECR data per sliding cycle for 350 °C annealed commercial Ni HG, pure Au, and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films sliding against a Au-Cu alloy rider. Shaded areas correspond to the average value, solid line, 71s of data collected at 50 Hz.
with the exception of the counterface sliding against the 350 °C annealed Ni HG. This is likely due to wear being localized to the counterface as the ECR data in Fig. 5 indicated a thick and continuous NiO film. The strong correlation of wear rate response between the counterfaces and films suggests that wear mechanisms for both materials are shared and highly dependent on the interfacial material structure and chemistry, and less so on the subsurface evolution, which will be presented and discussed now.
3.4. Wear surface and subsurface analysis Surface and subsurface structural and compositional analyses were performed to determine the mechanisms responsible for the large differences in friction coefficients, ECR values, and wear rates of the as-deposited and annealed 2.0 vol% ZnO films. Fig. 8 shows low and higher magnification planar SEM images of the wear track surfaces for the (a) as-deposited, (b) 250 °C annealed, and
1670
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 6. Specific wear rates of as-deposited (ASD), 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films.
Fig. 7. Specific wear rates of Au-Cu alloy counterface rider sliding against as-deposited (ASD), 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films.
(c) 350 °C annealed 2.0 vol% ZnO films. Since the wear rates of the films followed a similar trend to the counterfaces, only the films were analyzed in detail. Furthermore, when sliding against the asdeposited 2.0 vol% ZnO film it was determined previously that transfer films do adhere to the Au-Cu counterface exhibiting the same composition as the wear tracks [4]. The SEM images of the as-deposited 2.0 vol% ZnO film wear track shown in Fig. 8(a) reveal relatively uniform wear occurred inside the track with some striations along the sliding direction indicative of abrasive wear. In addition, there are a few darker contrast patches that are similar in appearance to those identified as Zn-rich wear debris [4]. An increased amount and size of these dark Zn-rich wear debris areas are shown inside the 250 °C annealed 2.0 vol% ZnO film wear track in Fig. 8(b). This wear track also exhibits a much narrower width in comparison to the film wear track in as-deposited condition, i.e., 50 μm to 30 μm, and does not contain the larger abrasive wear striations. These surface features are reflected in the much lower wear rate of the 250 °C annealed 2.0 vol% ZnO film. In contrast, the 350 °C annealed 2.0 vol% ZnO film wear track shown
in Fig. 8(c) exhibits increased wear, 56 μm track width, with a higher degree of plastic deformation and plowing of the film to the edges of the wear track. The higher magnification image in Fig. 8 (c) shows evidence of larger scale micro-abrasive striations as well as regions of adhesive pullout transfer. This change in wear mechanism is likely a result of the pre-worn roughened surface and faceted ZnO crystallites present on the surface of the 350 °C annealed 2.0 vol% ZnO film shown in Fig. 2(c). It is likely that these coarse ZnO crystals are facilitating wear by acting as third body abrasive particles. All of these features contribute to the highest wear rates for the 350 °C annealed 2.0 vol% ZnO film and counterface. To further investigate the changes in sliding ECR friction and wear mechanisms, the sliding-induced subsurface deformation structures were characterized by preparing FIB cross-sections inside the wear track (FIB-cut locations shown by boxes in Fig. 8) followed by STEM analysis. Fig. 9 shows cross-sectional STEM images and an EDS X-ray spectral image map acquired inside the as-deposited Au-2.0 vol% ZnO film wear track. From the low magnification Fig. 9(a) image, it is evident that there are several wear-induced zones when compared to the unworn film shown in Fig. 2(a). Moving from the bottom to the top of the image, there is a transition from unworn Au columnar grains to grains that bend from left to right along the sliding direction indicating the Au columnar grains plastically deform/flow along the sliding direction. On top of these bent grains, there is another wear zone of heavily refined grains that extend from the surface to a depth of 200 nm. The higher magnification STEM image in Fig. 9 (b) shows that some of these near surface, equiaxed grains are refined down to 20 nm in size. Also interesting is the clustering of ZnO, shown in the corresponding X-ray spectral image map in Fig. 9(b), which reveals a very thin ( 10 nm thick) ZnO surface film and even thinner subsurface ZnO intergranular films. These ZnO films, like the ZnO surface crystallites shown in Fig. 2(c), correspond to the darker areas in the STEM images in Fig. 9(a,b). The ZnO surface film shown in the X-ray spectral image does not extend across the entire FIB-cut suggesting it forms as islands in the wear track, in agreement with the planar SEM image shown in Fig. 8(a). Fig. 10 shows two cross-sectional STEM images acquired inside the 250 °C annealed Au-2.0 vol% ZnO film wear track. In comparison to the film wear track in as-deposited condition, there is less plasticity of the Au grains as evident by the reduction of bent columnar grains and refined near surface grains. Fig. 10(a) shows an area with near surface Au grain refinement in which the grains are orientated parallel to the sliding direction. It appears that some of these refined grains evolve into ejected wear debris; for example, the platelet-shaped features denoted by the wear surface arrows in Fig. 10(a). In addition, a very thin darker contrast ZnO film exists at some locations along this wear track surface. This ZnO film was much thicker in another section of the FIB-liftout, evident by the STEM image in Fig. 10(b) that shows an EDS determined 80 nm thick, continuous ZnO film at the wear surface. The presence of ZnO would reduce the adhesive contact and thus friction as was determined in the sliding ECR tests in Fig. 4. These thicker ZnO films also result in the slight increase of the sliding ECR from 22 mΩ in the as-deposited film, shown in Fig. 3, to 32 mΩ in the 250 °C film, shown in Fig. 4. In addition, the thicker, continuous ZnO film is protective in that it inhibits the underlying Au wear debris from forming the platelet-shaped debris shown in Fig. 10(a). Lastly, these thicker ZnO films were not observed across the entire wear surface, in agreement with the planar SEM image in Fig. 8(b). In contrast to the 250 °C annealed film, the 350 °C annealed Au-2.0 vol% film cross-section in Fig. 11 showed significant wear with only 400 nm of the Au-ZnO film remaining on top of the
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1671
Fig. 8. Low (left column) and higher (right column) magnification planar SEM images of the wear track surfaces for (a) as-deposited (ASD), (b) 250 °C annealed, and (c) 350 °C annealed Au-2.0 vol% ZnO films. The dashed boxes in the higher magnification images indicate the locations of the cross-sectional FIB liftouts.
Pt and Ti layers. The residual Au grain structure is no longer columnar but instead exhibits more heavily refined grains elongated along the sliding direction. There is also no evidence of the protective ZnO film on the wear surface across the entire FIB-liftout, in agreement with the planar SEM image in Fig. 8(c). The significantly reduced film thickness and highly deformed microstructure of the cross-section agree with the high film wear rate and suggest that a significant portion of the film thickness experienced a migration of the strengthening ZnO at the Au grain boundaries due to the elevated annealing temperature. This is supported by the Fig. 2(c) STEM image that shows ZnO diffused through the Au grain boundaries to form the ZnO surface crystallites that are not beneficial in mitigating friction and wear since they are not present in the sliding contact. Instead, they are likely ejected as loose wear debris. Therefore, the sliding-induced evolution of the surface and subsurface structure and chemistry, shown by the images in Figs. 8–11, support the experimental sliding ECR friction and wear results and mechanisms.
4. Summary and conclusions The sliding ECR friction and wear behavior of as-deposited, 250 °C, and 350 °C annealed pure Au films and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO composite films has been determined with comparisons made to commercial Ni hardened Au. Of the films tested, the 250 °C annealed Au-2.0 vol% film exhibited the best overall performance and thus is an environmentally friendly alternative to the widely used electroplated hard Au, which relies on electrochemical bath chemistries that are difficult to dispose of and recycle. Specifically, the 250 °C annealed Au-2 vol% ZnO film exhibited a low stable friction coefficient (μ 0.25), sliding ECR ( 35 mΩ), and film wear rate (1.5 10 5 mm3/N m), with the latter value being an order of magnitude lower compared to the Ni hardened Au film. Planar SEM and cross-sectional STEM studies inside the wear surfaces revealed that the extremely stable, low friction coefficients, ECR values, and wear rates were due to the formation of thick and continuous ZnO tribofilms on the surface
1672
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 11. Cross-sectional STEM image inside wear track of 350 °C annealed Au2.0 vol% ZnO film taken from FIB-cut location in Fig. 8(c). SD ¼ sliding direction.
that minimized the adhesive contact contribution to wear and mitigated abrasion in the wear track. Fig. 9. (a) Low and (b) higher magnification cross-sectional STEM images inside wear track of as-deposited Au-2.0 vol% ZnO film taken from FIB-cut location in Fig. 8(a). The AXSIA-analyzed X-ray spectral image displayed as color overlay was acquired from the box in image (b). SD ¼ sliding direction.
Acknowledgments The authors would like to acknowledge Rand Garfield for design and construction of the sliding ECR wear tester. We also thank Lisa Marie Lowery for SEM imaging and preparing FIB samples. The authors also acknowledge John Curry for a critical review of the manuscript. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000.
References
Fig. 10. Cross-sectional STEM images inside wear track of 250 °C annealed Au2.0 vol% ZnO film taken from FIB-cut location in Fig. 8(b). Images (a) and (b) correspond to brighter and darker areas of the wear track in Fig. 8(b), respectively. SD ¼ sliding direction.
[1] C. Hageluken, C.W. Corti, Recycling of gold from electronics: cost-effective use through ‘Design for Recycling’, Gold. Bull. 43 (2010) 209–220. [2] P. Goodman, Current and future uses of gold in electronics, Gold. Bull. 35 (2002) 21–26. [3] M. Antler, Gold-plated contacts: effect of thermal aging on contact resistance, in: Proceedings of the Forty-Third IEEE Holm Conference on Electrical Contacts, 1997, pp. 121–131. [4] N. Argibay, S.V. Prasad, R.S. Goeke, M.T. Dugger, J.R. Michael, Wear resistant electrically conductive Au–ZnO nanocomposite coatings synthesized by e-beam evaporation, Wear 302 (2013) 955–962. [5] R.L. Schoeppner, D.F. Bahr, H. Jin, R.S. Goeke, N.R. Moody, S.V. Prasad, Wear behavior of Au–ZnO nanocomposite films for electrical contacts, J. Mater. Sci. 49 (2014) 6039–6047. [6] N. Argibay, J.E. Mogonye, J.R. Michael, R.S. Goeke, P.G. Kotula, T.W. Scharf, M. T. Dugger, S.V. Prasad, On the thermal stability of physical vapor deposited oxide-hardened nanocrystalline gold thin films, J. Appl. Phys. 117 (2015) 145302. [7] R.L. Schoeppner, R.S. Goeke, N.R. Moody, D.F. Bahr, Mechanical and electrical performance of thermally stable Au–ZnO films, Acta Mater. 91 (2015) 1–9. [8] P.G. Kotula, M.R. Keenan, J.R. Michael, Automated analysis of SEM X-ray spectral images: a powerful new microanalysis tool, Microsc. Microanal. 9 (2003) 1–17. [9] N. Argibay, M.T. Brumbach, M.T. Dugger, P.G. Kotula, Grain boundary diffusivity of Ni in Au thin films and the associated degradation in electrical contact resistance due to surface oxide film formation, J. Appl. Phys. 113 (2013) 114906–114908.
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
Tribology and sliding electrical contact resistance of e-beam hard Au: Effects of annealing J.E. Mogonye, N. Argibay, R.S. Goeke, P.G. Kotula, T.W. Scharf n,1, S.V. Prasad n Materials Science and Engineering Center, Sandia National Laboratories, Albuquerque, NM 87185-0889, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 1 September 2016 Received in revised form 19 January 2017 Accepted 20 January 2017
Nanocomposite Au-ZnO thin films in the dilute oxide (o 5.0 vol%) regime were synthesized by electron beam (e-beam) evaporation, as alternatives to electroplated Au hardened with Ni. Tribological measurements of e-beam hard Au were made while passing current through sliding contacts; electrical contact resistance (ECR) and friction data were simultaneously acquired during the test. The friction, wear and ECR behaviour were studied for the as-deposited film condition, and after annealing at 250 °C and 350 °C in air. The study revealed that the 250 °C annealed Au-2 vol% ZnO film exhibited the lowest, stable friction coefficient s (m 0.25) and ECR ( 35 mΩ) during sliding. Furthermore, the wear rate of this 250 °C annealed ZnO hardened Au nanocomposite film was an order of magnitude lower at 1.5 10 5 mm3/N m than for a typical Ni hardened, electroplated Au film at 1.3 10 4 mm3/N m. Crosssectional transmission electron microscopy studies inside the wear surfaces revealed that the extremely stable, low friction coefficients and wear rate of annealed Au-2 vol% ZnO film was due to partial coverage of the wear surface with a ZnO tribofilm that reduced the adhesive contact contribution to wear with minimal impact on ECR. The potential implications of this study in the search for an environmentally friendly alternative to widely used electroplated hard Au are discussed. & 2017 Elsevier B.V. All rights reserved.
Keywords: Electroplated hard Au Au-ZnO film Sliding electrical contact resistance Sliding wear Tribofilm Transmission electron microscopy
1. Introduction The necessity to provide efficient transfer of electrical current between moving parts remains an engineering challenge. With increasing implementation of electrical motors and generators and further increase in electrical contacts with the expanding electronics industry, the technological applications of sliding electrical contacts has become widespread. Sliding electrical contacts can be found in DC motor and generator current collector slip rings, printed circuit board edge connectors, data cable pin connectors and receivers, microelectromechanical systems (MEMS), circuit relays and switches, instrument signal transfer slip rings and many more applications. The main engineering challenges associated with these applications, especially long service life slip rings, is minimized wear and stable wear performance, acceptable friction to minimize mechanical energy losses, low and stable electrical contact resistance (ECR) to minimize Joule heating, and thermal stability of the sliding materials. The use of gold in the electronics industry has reached over n
Corresponding authors. E-mail addresses: [email protected] (T.W. Scharf), [email protected] (S.V. Prasad). 1 On faculty sabbatical from the Department of Materials Science and Engineering, The University of North Texas, Denton, TX 76203-5310, USA. http://dx.doi.org/10.1016/j.wear.2017.01.080 0043-1648/& 2017 Elsevier B.V. All rights reserved.
300,000 kg per year in 2010 [1]. According to an evaluation by Goodman in 2002, when gold use in the electronics industry was 280,000 kg per year, the majority of this gold was used in electrical contacts, most frequently electroplated gold [2]. Gold is used in sliding electrical contacts due to its high conductivity and excellent corrosion and oxidation resistance. Although electroplated hard gold (gold alloyed with typically 1–2 wt% of cobalt or nickel) has produced an improvement over pure gold plating in friction and wear performance, it has reliability issues with ECR over time and at elevated temperatures [3]. Additionally, there are growing concerns with environmental hazards associated with the use and disposal of cyanide and arsenic baths used in plating chemistries. The reliability issue is related to an increase in ECR due to the formation of surface oxide films that can cause system failure in low voltage and high frequency applications, especially systems which have expected service lives on the order of years. Improvement of mechanical and tribological performance in hard Au has been realized by co-depositing with ZnO to synthesize ZnOhardened Au nanocomposite films [4–7]. It has been determined that the room temperature sliding friction and ECR in the dilute oxide regime (o 5 vol% ZnO) are lower than Ni hardened gold films [4]. However, it is unknown if annealing these films results in any improvements in the sliding electrical contact properties. As such, this paper presents results on friction, ECR and wear properties of as-deposited and annealed composite Au-ZnO films with
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1663
comparisons made to as-deposited and annealed Ni hardened Au films. The sliding-induced structural and chemical changes of the worn surfaces and subsurfaces that control these properties will also be presented and discussed.
2. Experimental methods 2.1. Electron-beam film deposition A Thermionics Vacuum Products Co. 10 kV Triad e-beam evaporation system was used for the deposition of both pure Au and composite Au-ZnO films. The electron beam is generated by a field emission source and magnetically deflected to the crucible to provide localized heating resulting in the sublimation of the film material. Once sublimated in a vacuum, the film material deposits on the target substrate in line-of-sight from the crucible. Prior to all film depositions the vacuum chamber was pumped to a base pressure of less than 1 10 6 Torr. The film material source to substrate distance was 305 mm and all depositions were conducted at room temperature. A rotating substrate mount was used to achieve compositional and thickness uniformity of the films. Prior to e-beam deposition, the Si substrates were first sputter coated with a nominally 300 nm thick Ti layer for adhesion followed by a nominally 300 nm thick Pt layer to act as a diffusion barrier. These layer thicknesses were determined by cross-sectional microscopy. For comparison purposes, electroplated hard Au films were obtained from a commercial plating company (Theta Plating Inc. Albuquerque, NM, USA). Nickel was used as the hardening element and plated per ASTM Type I specifications (ASTM B488-11 standard), also comprising of a 5 μm thick electroplated Ni diffusion barrier, deposited on Alloy 52 substrates (nominally a 50–50 wt% Fe-Ni alloy). Thickness of electroplated Ni hardened gold (Ni HG) film was 1.5 μm. The pure e-beam Au films, nominally 2 mm thick, were deposited by evaporation of high purity Au pellets (99.999% purity) sourced from Materion Advanced Chemicals. The resultant Au film average root-mean-square (RMS) surface roughness was determined by scanning white light interferometry (SWLI) to be 3 nm. The composite Au-ZnO films, nominally 2 mm thick, of varying composition were also deposited onto Si substrates sputter coated with Ti/Pt layers. The e-beam deposition source materials were pure Au pellets, mentioned above, and ZnO tablets (99.9% purity) from Materion Advanced Chemicals. The deposition rates were varied between 0.00 and 2.00 nm/sec to achieve the desired volume fraction of ZnO in the film in the dilute oxide (o 5.0 vol%) regime. The resultant film average RMS surface roughness of composite Au-ZnO films was determined by SWLI to be 4 nm. Prior to tribological testing, some of the pure Au and composite Au-ZnO films were ex situ annealed in an open air furnace to either 250 °C or 350 °C for 24 h dwell time. 2.2. Sliding electrical contact testing A custom built linear tribometer equipped with an Agilent Technologies B2911A digital source/meter, photograph and accompanying schematic, shown in Fig. 1, was used to conduct all simultaneous friction and ECR measurements. The tribometer independently measures applied normal force and lateral force through independent double leaf flexures, which constrain the displacement to in plane linear motion, coupled to load cells. The load cells were independently calibrated using dead weights and calibrated scales throughout the expected experimental loads, thus incorporating the leaf flexure spring rate into the calibration. Continuous ECR measurements are made by operating the source/ meter in voltage-regulated remote sensing mode and measuring
Fig. 1. Photograph and accompanying schematic of custom-built linear ECR tribometer.
the voltage drop and current using a four-point probe bridged across the film surface and counterface pin. Both the substrate of the film and the counterface pin are electrically isolated from the rest of the rig by custom fixtures machined from polycarbonate and polyether ether ketone, respectively. All tribological friction ECR tests were conducted in unidirectional sliding motion by translating the bottom fixture by the 2 mm track length while recording data then lifting the upper load cell assembly and translating the bottom fixture to the start position and repeating. A normal force of 100 mN and a linear speed of 1 mm/s were used for a duration of 100 sliding cycles. Prior to each sliding ECR experiment, a static normal force of 100 mN was applied and a voltage set to produce approximately 100 mA of direct current through the ECR circuit. All tests were conducted at room temperature in lab air (15–30% relative humidity). The
1664
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
counterface pin was made from Neyoro G alloy wires, a hard, primarily Au-Cu alloy with nominal composition of 72Au-14Cu8Pt-5Ag % by weight, sourced from Deringer-Ney, Inc. The Au-Cu alloy wires were first annealed to a hardness of approximately 1.96 GPa and then hand machined on a lathe to produce a hemispherical tip with a 1.59 7 0.05 mm radius of curvature, in order to produce a nominally circular (Hertzian) contact geometry. The machined hemispherical tips were subsequently hand lapped using progressively finer diamond paste to a center line average roughness (Ra) of 100 nm or less measured by SWLI. Prior to sliding experiments, the polished riders were ultrasonically cleaned in isopropyl alcohol and blown dry with dry nitrogen. The top insulating fixture holds the Au-Cu alloy counterface pin at a canted angle of 20°, as shown in Fig. 1, so that multiple tests could be conducted on a single rider by rotating it about its cylindrical axis to a unique contact area. At least three repeat tests were conducted for each film.
Surface topography data of the films and counterface pins pre and post sliding ECR tribometry were collected using a Veeco NT1100 SWLI with either a 50 or 10 objective lens. The topographical height data was then analyzed using Bruker Vision64 version 5.41 software. Wear volumes of the pins and films, post sliding ECR testing, were measured as the missing volume displaced below a reference plane fitted to the unworn pin and film surface outside the wear track, respectively. These wear volumes were used to calculate the specific wear rates of the pins and films using Archard's equation and are reported in units of mm3/N m. 2.3. Structural and chemical characterization techniques Scanning electron microscopy (SEM) images inside and outside the wear track surfaces of the e-beam deposited Au and composite Au-ZnO films were collected using a FEI Helios 450 SEM. Transmission electron microscopy (TEM) specimens were also prepared
Fig. 2. Planar SEM images (left column) and cross-sectional STEM images (right column) of (a) as-deposited (ASD), (b) 250 °C annealed, and (c) 350 °C annealed Au-2 vol% ZnO films. Near surface SADP are shown in inset of STEM images. A higher magnification STEM image is shown in inset of (c).
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
using the same system that is equipped with a focused ion beam (FIB) column. A site specific Pt layer was deposited in the SEM to protect the film surfaces from ion-induced damage. The FIB milled TEM specimens were then removed from the bulk of the film using an ex-situ microprobe and transferred to C film coated Cu grid TEM specimen holders. The FIB was also used to mill cross-sections to the surface normal of the as-deposited, 250 °C and 350 °C annealed composite Au-ZnO films using standard ion milling techniques with the revealed cross-section surface polished by low energy (2 keV Ga þ ) ions to enhance grain boundary contrast for line-intercept measurement of grain sizes. Annular dark field scanning TEM (ADF-STEM) was used for Z-contrast imaging of these cross-sections along with acquiring selected area diffraction patterns (SADP). Site specific FIB samples were made close to the center of the film wear tracks parallel to the sliding direction. Sliding-induced changes to the structure and chemistry in the worn subsurface regions were analyzed with cross-sectional STEM and energy dispersive X-ray spectrometry (EDS) analyses, respectively. X-ray spectral images were collected from wear subsurfaces and then analyzed using Sandia's Automated eXpert Spectral Image Analysis (AXSIA) software [8]. AXSIA quickly reduces large raw spectral images to a compact solution consisting of a small number of linearly independent component image/ spectrum pairs. For display purposes, the component images were combined into a color overlay.
3. Results and discussion 3.1. Surface and subsurface structure Fig. 2 shows both planar SEM images and cross-sectional STEM images of (a) as-deposited (b) 250 °C annealed, and (c) 350 °C annealed Au-2 vol% ZnO films. From the planar SEM image in Fig. 2 (a), it is evident that the as-deposited Au-2 vol% ZnO film surface grains appear to be homogenously distributed with a measured grain size of 75 nm. There is no observable grain coarsening with annealing to 250 °C shown in the Fig. 2(b) planar SEM image. However, the planar SEM image in Fig. 2(c) shows that when the film was annealed to 350 °C, there are numerous grains that have coarsened to a measured grain size of 140 nm. Cross-sectional STEM imaging was also performed to further investigate the grain structure in the as-deposited and annealed Au-2 vol% ZnO films. The cross-sectional STEM image of the asdeposited film in Fig. 2(a) shows that the nanocrystalline grains are columnar and exhibit some growth twins. The inset SADP image shows that the Au grains are randomly orientated, i.e., have no in plane o 1114 growth texture. Prior high resolution STEMEDS studies on Au-2 vol% ZnO film in as-deposited condition revealed that ZnO appears to be well distributed, concentrated as a thin film at the Au grain boundaries and intergranularly dispersed as ZnO particles that are on the order of single of nm's in diameter [6]. The STEM and SADP images in Fig. 2(b) show no changes in the near surface grain structure and orientation with annealing to 250 °C. However, it was previously determined that there is Au grain coarsening, but only at the film/Pt layer interface since ZnO diffused to this interface to form a 5 nm thick film [6]. With absence of ZnO at the Au grain boundaries there was corresponding Au grain growth. In contrast to the as-deposited and 250 °C annealed films, the 350 °C annealed film exhibited more faceted, roughened surface Au grains, shown in Fig. 2(c), although the Au grains did not exhibit coarsening from the STEM image or texturing from the SADP image. Instead, surface crystallites formed in many locations at the Au grain boundaries that showed a much darker contrast in higher magnification STEM images, such as the one shown in the inset of Fig. 2(c). These crystallites were
1665
determined to contain Zn according to EDS analysis and appear darker in STEM since they are lower atomic number than the brighter Au grains. This darker contrast is similar to the aforementioned Zn at the Au grain boundaries. The higher magnification STEM inset image in Fig. 2(c) shows these surface crystallites are located at the Au grain boundaries suggesting intergranular diffusion of Zn to the surface. It is likely these crystallites are ZnO, which is supported by the many larger hexagonal-shape surface crystallites shown in the planar SEM image in Fig. 2(c) that correspond to the hexagonal ZnO Wurtzite crystal structure. Therefore, ZnO diffused to the surface to form crystallites up to 140 nm in size when annealed to a higher temperature of 350 °C. 3.2. Friction and ECR behavior Typical friction coefficient and sliding ECR results are shown in Fig. 3 for as-deposited pure Au and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films as well as type I (499.7 mass % Au) commercial Ni HG. The Ni HG started at μ 0.3 friction coefficient and experienced a short run in period of 20 cycles of increasing friction until entering a region of periodic friction response from cycles 30–50, and finally reaching a steady state friction coefficient value of μ 0.75 after cycle 80. The sliding ECR of the Ni HG was highly variable with values consistently exceeding 20 mΩ until approximately cycle 75 when the values stabilized at 10 mΩ. These exceptionally high ECR values in conjunction with the friction behavior prior to cycle 80 were repeatable over several experiments. High variability in ECR can be explained as being due to contact separation between the rider and film associated with the presence of wear debris in the track. In contrast to the Ni HG film, the pure e-beam Au film exhibited a rapid increase in friction coefficient to μ 1.5, commensurate with a highly adhesive and large area of contact until approximately cycle 45 before the friction coefficient variance increased. The initial cycles (o45) of ECR for the Au film retained a value of 16 mΩ and low variance in conjunction with the same relatively low variance region in the friction coefficient before increasing and becoming more variable per cycle. This change in friction coefficient and ECR beyond cycle 45 is due to adhesive wear throughout the thickness of the Au layer. The friction coefficient with increasing ZnO concentration from 0.1–2.0 vol% resulted in the trend of decreasing average friction coefficient and variance as well as more stable regimes of friction throughout the 100 cycles of sliding ECR testing. The average stable ECR values for the 0.1– 2.0 vol% films slightly increased with increasing ZnO concentration from 16 mΩ to 22 mΩ, respectively, and is likely due to the increase in resistivity of the films with increasing ZnO concentration discussed in the previous section. Overall the variance of the ECR has also significantly decreased for all of the dilute ZnO concentrations shown with the exception of a few points during the testing of the 2.0 vol% film. In the first 30 sliding cycles of the 2.0 vol% film, there are a few large spikes in ECR indicative of a large fraction of contact separation similar to the Ni HG. This ECR response was previously reported to be a result of ZnO third body debris generated at the onset of wear before being displaced from the wear track as seen in SEM and characterized by EDS and SWLI maps of the same ZnO 2.0 vol% concentration film [4]. In addition, the lowering of friction coefficient with increasing ZnO content may be affected by the nanoindentation hardness increase from 1 to 2.5 GPa as well as the elastic modulus increase from 76 to 83 GPa with increasing ZnO concentrations from 0.1 vol% to 2 vol%, respectively [6]. In summary, while the ECR values are relatively low, the friction coefficients remain high (μ 40.5) for all films in asdeposited condition. Table 1 summarizes the average and standard deviation of the friction coefficient and ECR values for the as-deposited films.
1666
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 3. Friction coefficients and ECR data per sliding cycle for as-deposited commercial Ni HG, pure Au, and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films sliding against a Au-Cu alloy rider. Shaded areas correspond to the average value, solid line, 71s (standard deviation) of data collected at 50 Hz.
Typical friction coefficient and sliding ECR results are shown in Fig. 4 for 250 °C annealed pure Au and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films as well as type I commercial Ni HG. The Ni HG exhibited a significantly reduced steady state friction coefficient of μ 0.25 and ECR of 18 mΩ in comparison to the as-deposited film. The run in period of slightly higher friction coefficient and ECR lasted for only the first 10 cycles before obtaining the low variance and steady state regime. It is likely that annealing of the
Ni HG at 250 °C generated greater stability in the Au microstructure by diffusion of Ni underlayer through the Au film along grain boundaries, a well-known phenomenon [9]. In addition, it is unclear if annealing resulted in significant Au grain growth or the formation of a NiO surface film, however the results indicate that the surface wear evolution led to a smooth and conformal contact without large third body wear debris that can cause the contact separation ECR response shown previously in Fig. 3. The pure
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1667
Table 1 Average and standard deviation, 1s, of coefficient of friction, CoF, and electrical contact resistance, ECR, values for each film composition and annealing condition. Averages were taken from every sliding ECR test and across all 100 sliding cycles. Composition
Ni HG 0.0% ZnO 0.1% ZnO 0.5% ZnO 1.0% ZnO 2.0% ZnO
250 °C annealed
As-deposited (ASD)
350 °C annealed
Avg. CoF 7 1s
Avg. ECR (Ω)7 1s
Avg. CoF 71s
Avg. ECR (Ω) 7 1s
Avg. CoF 7 1s
Avg. ECR (Ω) 7 1s
0.57 70.13 2.17 70.95 1.99 7 0.74 1.45 7 0.36 0.62 7 0.14 0.56 7 0.12
1.00E þ 027 1.82E þ 02 2.37E-027 2.23E-02 2.14E-027 1.05E-02 2.20E-02 7 4.54E-03 2.07E-027 1.30E-04 3.99E-02 7 3.67E-02
0.29 7 0.08 1.647 0.45 1.217 0.37 1.137 0.20 0.36 7 0.07 0.277 0.03
1.61E-02 7 2.36E-03 1.14E-02 7 5.65E-03 1.18E-02 71.11E-03 2.06E-02 71.62E-03 4.75E-02 7 7.77E-03 3.60E-02 7 4.09E-03
0.39 7 0.08 1.517 0.41 1.25 7 0.35 1.23 7 0.24 0.87 7 0.20 1.09 7 0.26
2.50E-017 2.53E-01 1.98Eþ 007 5.72E þ 00 1.55E-02 7 6.98E-03 1.99E-02 7 2.25E-03 2.46E-02 72.51E-03 2.25E-02 73.20E-03
e-beam Au film friction coefficient began in excess of μ 3.0 while the ECR remained stable near 15 mΩ throughout the test, and is a result of the Au film being completely (visible to the naked eye) detached from the Pt/Ti adhesion layers in the first few cycles. This is the result of gross adhesive wear and film adhesion failure likely due to large grain formation from annealing. The 0.1 vol% ZnO film exhibited high friction coefficient values although slightly lower than the pure Au film at the 250 °C annealed condition, and did not completely fail as did the pure Au film. The high friction is likely a result of Au grain growth due to a small percentage of ZnO in the composite film. The ECR values remained low similar to the as-deposited film. The 0.5 vol% ZnO film exhibited similar results to the as-deposited film condition although the variance in friction coefficient decreased and as in the other films the ECR values remained low. In addition, the trend of decreasing friction coefficient with increasing ZnO content continued. Both the 250 °C annealed 1.0 and 2.0 vol% ZnO films showed significantly reduced and stable friction coefficient values of μ 0.25. The ECR response for both films was also similar, increasing during a run in period over the first 10 cycles and slowly stabilizing to a value near 45 mΩ and 35 mΩ, respectively, over the duration of the tests. These ECR values are nearly twice that of the as-deposited values and could be a result of a larger concentration of ZnO at the surface due to annealing that is not forming large third body debris resulting in contact separation, as seen in the ECR behavior of the as-deposited 2.0 vol% film in Fig. 3. The extremely stable and low friction coefficient of these films corroborates the hypothesis that ZnO has migrated to the surface and is acting to minimize the adhesive contact contribution to friction. In summary, both the friction coefficient and ECR values are relatively low for the Ni HG and 1.0 and 2.0 vol% ZnO composite films. Table 1 summarizes the average and standard deviation of the friction coefficient and ECR values for the 250 °C annealed films. Typical friction coefficient and sliding ECR results are shown in Fig. 5 for 350 °C annealed pure Au and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films as well as type I commercial Ni HG. The commercial Ni HG did maintain stable friction coefficients of μ o0.5, however the ECR values of the first 12 cycles of sliding are two orders of magnitude greater than the 250 °C annealed film indicative of a thick and uniform NiO film. The ECR did reduce after initial wear since the surface oxide was likely worn away, although there were large spikes due to third body contact separation. Overall, these friction coefficient and ECR values for Ni HG are not as low and stable as the 250 °C annealed film. In the case of pure Au, the film completely failed and delaminated from the Pt/Ti adhesion layers in the first few cycles, similar to the 250 °C film. Both the 0.1 and 0.5 vol% ZnO films exhibited similar friction coefficient and ECR values as the corresponding 250 °C annealed films. However, the 1.0 and 2.0 vol% ZnO films exhibited a large increase in friction coefficients from μ 0.25 to μ 0.8 when compared to their respective 250 °C annealed films, albeit
slightly more stable ECR values near 25 mΩ, although there were a few points of contact separation during cycle 19 for the 2.0 vol% ZnO film. In summary, further annealing from 250 °C to 350 °C resulted in very high friction coefficient values for all the films. Table 1 summarizes the average and standard deviation of the friction coefficient and ECR values for the 350 °C annealed films. Overall when comparing the friction and ECR response in Figs. 3–5, the 2.0 vol% ZnO composite film was very comparable in performance to the Ni HG and had a more stable ECR in the asdeposited condition. In the annealed condition of 250 °C, the 2.0 vol% ZnO film exhibited a slightly lower and more stable friction coefficient with comparable ECR values. In addition, the 2.0 vol% ZnO film exhibited more stable ECR at 350 °C in comparison to Ni HG that showed the response of a thick homogeneous NiO film. 3.3. Wear behavior of films and counterfaces In addition to friction and ECR response, the wear rates of the films and counterface riders after the sliding tests were measured to determine overall tribological performance. Fig. 6 summarizes the specific wear rates of as-deposited, 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films. It is evident that the commercial Ni HG exhibited the lowest wear rate for the as-deposited condition with an average value of 9.5 10 5 mm3/N m. This is likely a result of a conformal oxide film acting to minimize adhesive wear, which is supported by the high ECR values in Fig. 3. The 0.1 and 0.5 vol% ZnO films exhibited similar, high wear rates to that of the large grain pure e-beam Au for all the as-deposited and annealed films tested. This result is interesting in that although the hardness of these films was greater than double that of the pure e-beam Au ( 0.8 GPa), the wear rates were of the same magnitude, suggesting that the higher area fraction of large grains, determined previously for these very dilute ZnO films [6], yields similar adhesive wear to the pure e-beam Au. In contrast, the 1.0 vol% ZnO film exhibited similar wear rates to the Ni HG except for an increase by an order of magnitude for the 350 °C annealed condition indicating a change in wear mechanisms for that film. In addition, the as-deposited 2.0 vol% film showed fairly high wear rates in comparison to what was reported previously [4]. This could be due to uncertainties in the ZnO concentration arising from the variability in the e-beam deposition process or other unknown factors. However, the 2.0 vol% film did exhibit a drastically different wear rate for the 250 °C annealed condition (1.5 10 5 mm3/N m), which was the lowest wear rate of all films, even an order of magnitude lower compared to 1.3 10 4 mm3/N m for Ni HG annealed at 250 °C. After annealing the 2.0 vol% film to 350 °C, the wear rate significantly increased by two orders of magnitude, and was again similar to the other ZnO films suggesting that there is a change in wear mechanism(s) between these two annealing temperatures that is likely surface structure and/or chemistry dependent. In
1668
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 4. Friction coefficients and ECR data per sliding cycle for 250 °C annealed commercial Ni HG, pure Au, and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films sliding against a Au-Cu alloy rider. Shaded areas correspond to the average value, solid line, 71s of data collected at 50 Hz.
another study on these films, the nanoindentation hardness of the 2.0 vol% film cycled to an annealing temperature of 350 °C five times was virtually unchanged from the film hardness of 3.5 GPa in as-deposited condition [7]. However, these hardness values may not allow a direct comparison to the films in the present study since the total anneal time was much lower, i.e., 2 h for five total cycles compared to the present 24 h dwell time. The negligible effect of the shorter annealing time on hardness is further substantiated by no noticeable grain growth or other
microstructural/chemical changes, which is in contrast to the formation of ZnO crystallites at 350 °C after 24 h of annealing shown in Fig. 2(c). Fig. 7 summarizes the specific wear rates of the Au-Cu alloy counterfaces sliding against as-deposited, 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films. Overall, the wear rates of the counterfaces are lower than the respective films, but the trends are in excellent agreement with the wear rates of the films for each condition tested
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1669
Fig. 5. Friction coefficients and ECR data per sliding cycle for 350 °C annealed commercial Ni HG, pure Au, and Au-(0.1, 0.5, 1.0, and 2.0 vol%) ZnO films sliding against a Au-Cu alloy rider. Shaded areas correspond to the average value, solid line, 71s of data collected at 50 Hz.
with the exception of the counterface sliding against the 350 °C annealed Ni HG. This is likely due to wear being localized to the counterface as the ECR data in Fig. 5 indicated a thick and continuous NiO film. The strong correlation of wear rate response between the counterfaces and films suggests that wear mechanisms for both materials are shared and highly dependent on the interfacial material structure and chemistry, and less so on the subsurface evolution, which will be presented and discussed now.
3.4. Wear surface and subsurface analysis Surface and subsurface structural and compositional analyses were performed to determine the mechanisms responsible for the large differences in friction coefficients, ECR values, and wear rates of the as-deposited and annealed 2.0 vol% ZnO films. Fig. 8 shows low and higher magnification planar SEM images of the wear track surfaces for the (a) as-deposited, (b) 250 °C annealed, and
1670
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 6. Specific wear rates of as-deposited (ASD), 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films.
Fig. 7. Specific wear rates of Au-Cu alloy counterface rider sliding against as-deposited (ASD), 250 °C, and 350 °C annealed commercial Ni HG, pure Au and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO films.
(c) 350 °C annealed 2.0 vol% ZnO films. Since the wear rates of the films followed a similar trend to the counterfaces, only the films were analyzed in detail. Furthermore, when sliding against the asdeposited 2.0 vol% ZnO film it was determined previously that transfer films do adhere to the Au-Cu counterface exhibiting the same composition as the wear tracks [4]. The SEM images of the as-deposited 2.0 vol% ZnO film wear track shown in Fig. 8(a) reveal relatively uniform wear occurred inside the track with some striations along the sliding direction indicative of abrasive wear. In addition, there are a few darker contrast patches that are similar in appearance to those identified as Zn-rich wear debris [4]. An increased amount and size of these dark Zn-rich wear debris areas are shown inside the 250 °C annealed 2.0 vol% ZnO film wear track in Fig. 8(b). This wear track also exhibits a much narrower width in comparison to the film wear track in as-deposited condition, i.e., 50 μm to 30 μm, and does not contain the larger abrasive wear striations. These surface features are reflected in the much lower wear rate of the 250 °C annealed 2.0 vol% ZnO film. In contrast, the 350 °C annealed 2.0 vol% ZnO film wear track shown
in Fig. 8(c) exhibits increased wear, 56 μm track width, with a higher degree of plastic deformation and plowing of the film to the edges of the wear track. The higher magnification image in Fig. 8 (c) shows evidence of larger scale micro-abrasive striations as well as regions of adhesive pullout transfer. This change in wear mechanism is likely a result of the pre-worn roughened surface and faceted ZnO crystallites present on the surface of the 350 °C annealed 2.0 vol% ZnO film shown in Fig. 2(c). It is likely that these coarse ZnO crystals are facilitating wear by acting as third body abrasive particles. All of these features contribute to the highest wear rates for the 350 °C annealed 2.0 vol% ZnO film and counterface. To further investigate the changes in sliding ECR friction and wear mechanisms, the sliding-induced subsurface deformation structures were characterized by preparing FIB cross-sections inside the wear track (FIB-cut locations shown by boxes in Fig. 8) followed by STEM analysis. Fig. 9 shows cross-sectional STEM images and an EDS X-ray spectral image map acquired inside the as-deposited Au-2.0 vol% ZnO film wear track. From the low magnification Fig. 9(a) image, it is evident that there are several wear-induced zones when compared to the unworn film shown in Fig. 2(a). Moving from the bottom to the top of the image, there is a transition from unworn Au columnar grains to grains that bend from left to right along the sliding direction indicating the Au columnar grains plastically deform/flow along the sliding direction. On top of these bent grains, there is another wear zone of heavily refined grains that extend from the surface to a depth of 200 nm. The higher magnification STEM image in Fig. 9 (b) shows that some of these near surface, equiaxed grains are refined down to 20 nm in size. Also interesting is the clustering of ZnO, shown in the corresponding X-ray spectral image map in Fig. 9(b), which reveals a very thin ( 10 nm thick) ZnO surface film and even thinner subsurface ZnO intergranular films. These ZnO films, like the ZnO surface crystallites shown in Fig. 2(c), correspond to the darker areas in the STEM images in Fig. 9(a,b). The ZnO surface film shown in the X-ray spectral image does not extend across the entire FIB-cut suggesting it forms as islands in the wear track, in agreement with the planar SEM image shown in Fig. 8(a). Fig. 10 shows two cross-sectional STEM images acquired inside the 250 °C annealed Au-2.0 vol% ZnO film wear track. In comparison to the film wear track in as-deposited condition, there is less plasticity of the Au grains as evident by the reduction of bent columnar grains and refined near surface grains. Fig. 10(a) shows an area with near surface Au grain refinement in which the grains are orientated parallel to the sliding direction. It appears that some of these refined grains evolve into ejected wear debris; for example, the platelet-shaped features denoted by the wear surface arrows in Fig. 10(a). In addition, a very thin darker contrast ZnO film exists at some locations along this wear track surface. This ZnO film was much thicker in another section of the FIB-liftout, evident by the STEM image in Fig. 10(b) that shows an EDS determined 80 nm thick, continuous ZnO film at the wear surface. The presence of ZnO would reduce the adhesive contact and thus friction as was determined in the sliding ECR tests in Fig. 4. These thicker ZnO films also result in the slight increase of the sliding ECR from 22 mΩ in the as-deposited film, shown in Fig. 3, to 32 mΩ in the 250 °C film, shown in Fig. 4. In addition, the thicker, continuous ZnO film is protective in that it inhibits the underlying Au wear debris from forming the platelet-shaped debris shown in Fig. 10(a). Lastly, these thicker ZnO films were not observed across the entire wear surface, in agreement with the planar SEM image in Fig. 8(b). In contrast to the 250 °C annealed film, the 350 °C annealed Au-2.0 vol% film cross-section in Fig. 11 showed significant wear with only 400 nm of the Au-ZnO film remaining on top of the
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
1671
Fig. 8. Low (left column) and higher (right column) magnification planar SEM images of the wear track surfaces for (a) as-deposited (ASD), (b) 250 °C annealed, and (c) 350 °C annealed Au-2.0 vol% ZnO films. The dashed boxes in the higher magnification images indicate the locations of the cross-sectional FIB liftouts.
Pt and Ti layers. The residual Au grain structure is no longer columnar but instead exhibits more heavily refined grains elongated along the sliding direction. There is also no evidence of the protective ZnO film on the wear surface across the entire FIB-liftout, in agreement with the planar SEM image in Fig. 8(c). The significantly reduced film thickness and highly deformed microstructure of the cross-section agree with the high film wear rate and suggest that a significant portion of the film thickness experienced a migration of the strengthening ZnO at the Au grain boundaries due to the elevated annealing temperature. This is supported by the Fig. 2(c) STEM image that shows ZnO diffused through the Au grain boundaries to form the ZnO surface crystallites that are not beneficial in mitigating friction and wear since they are not present in the sliding contact. Instead, they are likely ejected as loose wear debris. Therefore, the sliding-induced evolution of the surface and subsurface structure and chemistry, shown by the images in Figs. 8–11, support the experimental sliding ECR friction and wear results and mechanisms.
4. Summary and conclusions The sliding ECR friction and wear behavior of as-deposited, 250 °C, and 350 °C annealed pure Au films and Au-(0, 0.1, 0.5, 1.0, 2.0 vol%) ZnO composite films has been determined with comparisons made to commercial Ni hardened Au. Of the films tested, the 250 °C annealed Au-2.0 vol% film exhibited the best overall performance and thus is an environmentally friendly alternative to the widely used electroplated hard Au, which relies on electrochemical bath chemistries that are difficult to dispose of and recycle. Specifically, the 250 °C annealed Au-2 vol% ZnO film exhibited a low stable friction coefficient (μ 0.25), sliding ECR ( 35 mΩ), and film wear rate (1.5 10 5 mm3/N m), with the latter value being an order of magnitude lower compared to the Ni hardened Au film. Planar SEM and cross-sectional STEM studies inside the wear surfaces revealed that the extremely stable, low friction coefficients, ECR values, and wear rates were due to the formation of thick and continuous ZnO tribofilms on the surface
1672
J.E. Mogonye et al. / Wear 376-377 (2017) 1662–1672
Fig. 11. Cross-sectional STEM image inside wear track of 350 °C annealed Au2.0 vol% ZnO film taken from FIB-cut location in Fig. 8(c). SD ¼ sliding direction.
that minimized the adhesive contact contribution to wear and mitigated abrasion in the wear track. Fig. 9. (a) Low and (b) higher magnification cross-sectional STEM images inside wear track of as-deposited Au-2.0 vol% ZnO film taken from FIB-cut location in Fig. 8(a). The AXSIA-analyzed X-ray spectral image displayed as color overlay was acquired from the box in image (b). SD ¼ sliding direction.
Acknowledgments The authors would like to acknowledge Rand Garfield for design and construction of the sliding ECR wear tester. We also thank Lisa Marie Lowery for SEM imaging and preparing FIB samples. The authors also acknowledge John Curry for a critical review of the manuscript. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000.
References
Fig. 10. Cross-sectional STEM images inside wear track of 250 °C annealed Au2.0 vol% ZnO film taken from FIB-cut location in Fig. 8(b). Images (a) and (b) correspond to brighter and darker areas of the wear track in Fig. 8(b), respectively. SD ¼ sliding direction.
[1] C. Hageluken, C.W. Corti, Recycling of gold from electronics: cost-effective use through ‘Design for Recycling’, Gold. Bull. 43 (2010) 209–220. [2] P. Goodman, Current and future uses of gold in electronics, Gold. Bull. 35 (2002) 21–26. [3] M. Antler, Gold-plated contacts: effect of thermal aging on contact resistance, in: Proceedings of the Forty-Third IEEE Holm Conference on Electrical Contacts, 1997, pp. 121–131. [4] N. Argibay, S.V. Prasad, R.S. Goeke, M.T. Dugger, J.R. Michael, Wear resistant electrically conductive Au–ZnO nanocomposite coatings synthesized by e-beam evaporation, Wear 302 (2013) 955–962. [5] R.L. Schoeppner, D.F. Bahr, H. Jin, R.S. Goeke, N.R. Moody, S.V. Prasad, Wear behavior of Au–ZnO nanocomposite films for electrical contacts, J. Mater. Sci. 49 (2014) 6039–6047. [6] N. Argibay, J.E. Mogonye, J.R. Michael, R.S. Goeke, P.G. Kotula, T.W. Scharf, M. T. Dugger, S.V. Prasad, On the thermal stability of physical vapor deposited oxide-hardened nanocrystalline gold thin films, J. Appl. Phys. 117 (2015) 145302. [7] R.L. Schoeppner, R.S. Goeke, N.R. Moody, D.F. Bahr, Mechanical and electrical performance of thermally stable Au–ZnO films, Acta Mater. 91 (2015) 1–9. [8] P.G. Kotula, M.R. Keenan, J.R. Michael, Automated analysis of SEM X-ray spectral images: a powerful new microanalysis tool, Microsc. Microanal. 9 (2003) 1–17. [9] N. Argibay, M.T. Brumbach, M.T. Dugger, P.G. Kotula, Grain boundary diffusivity of Ni in Au thin films and the associated degradation in electrical contact resistance due to surface oxide film formation, J. Appl. Phys. 113 (2013) 114906–114908.