Hard Pure-Gold and Gold-CNT Composite Plating Using Electrodeposition Technique with Environmentally Friendly Sulfite Bath

2.1.1 Material and method

There is a report [13] in which a sulfite complex is used for a non-cyanic plating bath. In this section, the base sulfite bath was used as a non-cyan gold plating bath. Table 1 shows the composition of the plating bath and the plating conditions.

The condition of rectangular pulse current was optimized at the average current density of 50 A/m², the pulse period of 100 ms, and the duty ratio of 0.1. The Ni-P plated film was used as an experimental substrate (amorphous alloy as a base film) with thickness of about 3 μm on Cu plate. A plating bath manufactured by Meltex was used for plating the base film. The substrate used for the experiment was insulated with a masking tape excepting the deposition surface (10 mm × 20 mm), alkaline degreasing and acid treatment were carried out as plating pretreatment. Further, for comparison, a hard gold plating film with Co-content was similarly prepared, which was plated on the probe, used a commercially cyan bath (by Meltex: Auronal 44 BC). The characteristics of the plated films were evaluated in terms of surface morphology, Vickers hardness, X-ray diffraction (XRD), and specific resistance value, by a field emission scanning electron microscope (FE-SEM: S-4100, Hitachi), a dynamic indentation tester (DUH-201, SHIMADZU), X-ray diffractometer (RINT2200V, Rigaku), a four probe method, respectively.

2.1.2 Prototype probe

Figure 1 shows the photograph of the probe with both ends moved (spring movable stroke of 0.65 mm). A prototype was fabricated by plating the developed plating film on the movable part at both ends of the probe (Cu alloy, and the Ni-P plating as base) shown in Figure 1.

The repeated durability of the contact resistance was compared and evaluated for the prototype probes (using the same parts) of the current plating film and the developed plating film. Both gold plates were used for the contact resistance measurement of the probe, and a four-terminal method was used for the electric circuit measurement. Here, the measurement was performed at room temperature (RT, the temperature: 24°C, the humidity: 50% [14]) with the probe stroke of 0.65 mm, the load of about 0.24 N, and the measuring current of 10 mA.

2.1.3 Surface observation by FE-SEM

The three types of electrodeposited films, (a) a hard gold plating film with Co-containing (AuCo film) from a cyanide bath, (b) a direct current (DC) gold plating film (DC-Au film) from an original sulfite bath, (c) a pulsed current plating film from the same original sulfite bath (PC-Au film), were prepared and observed the surface morphology by FE-SEM are shown in Figure 2, respectively.

Compared with the surface morphology of AuCo film from Figure 2(a), it can be confirmed that the DC-Au film has larger angular crystals (Figure 2(b)). The surface of PC-Au film (Figure 2(c)) shows the small crystals without corners by comparing with that of DC-Au film (Figure 2(b)).

2.1.4 Vickers hardness (Hv) measurement

Figure 3 shows the average value, the maximum value, and the minimum value of Hv obtained from three kinds of plated film surfaces by measuring 10 times at RT using a dynamic indentation tester. Here, the measurement conditions using a diamond indenter were carried out at the test load of 10 gf, the load speed of 0.675 gf/s, and the holding time of 20 s.

From this result, (c) PC-Au film is about twice as hard as (b) DC-Au film, which is harder. The Hv value is close to that of commercially available (a) AuCo film. It is presumed that the improved hardness of PC-Au film is due to the crystal size. Next, the crystal structure of the plating film was investigated by XRD.

2.1.5 Relationship between crystallite size and Vickers hardness (Hv)

Figure 4 shows the results of XRD measurement of the three plating films, (a) AuCo film, (b) DC-Au film, and (c) PC-Au film, using Cu tube (CuKα1: wavelength 1.5405 Å), at 40 kV and 20 mA.

From Figure 4, gold peaks of face-centered cubic (FCC) crystal were confirmed in all the samples. The crystallite size Dhkl of the (hkl) face (Miller Index) was calculated by the Scherrer equation of the following Eq. (1) [15].

The apparent crystallite size (Da) was determined using a weighted average of the crystallite sizes from the XRD peaks on each face. In this calculation, the following Eq. (2) was used.

As a result, the values of the crystallite size (Da) from (a) AuCo, (b) DC-Au, and (c) PC-Au films were (a) 19.1 nm, (b) 28.9 nm, and (c) 17.0 nm, respectively. In the original bath, (c) PC-Au film obtained by the pulse electrodeposition has wider peak and smaller Da than that of (b) DC-Au film obtained by the direct current electrodeposition. The reason for these results is related to the fact that the critical radius r_c of the crystal nucleus according to the following Eq. (3) is decreased by pulse electrodeposition with a large overvoltage (η) [16].

It is possible to form small crystal nucleus in this case. Here, γ is the intrinsic surface energy, V is the atomic capacity (the volume occupied by 1 g atom in the solid state), z is the valence, F is the Faraday constant, and η is the overvoltage. In general, for metals, the Hall-Petch rule expressed by the following Eq. (4) is established with mechanical yield strength σ [6, 11].

Here, σ₀ is the intrinsic yield strength (independent of crystal size), k is the constant depending on materials, and d is the crystal size. For the metal film, a proportional relationship is established between the yield strength and the hardness Hv. Paying attention to Eq. (4), the smaller the crystal size is, the harder the metal material is.

Various samples were prepared by changing pulse conditions in the same method, and the relationship between grain size (Da) from XRD and Hv was as shown in Figure 5 (Hall-Petch rule).

From the results in Figure 5, although the Hv of the pulsed electrodeposition film increases as the crystallite size decreases, the plot tends to decrease from around the critical point (Da^−1/2 = 0.24). The reason for this is presumed to be the result of the inverse Hall-Petch rule [17, 18, 19] that the self-weight collapse occurs in this gold electrodeposition film when the crystallite size becomes too small. Therefore, the condition around Da^−1/2 = 0.22 (D = 20.7 nm) was selected to form the plated film to obtain appropriate hardness for probe application.

2.1.6 Specific resistance value of plating film

In general, the four probe method of the following Eq. (5) is used as a standard method for measuring the resistivity ρ at RT [20]. Here, the variable t is the plating thickness, and the value (V/I) indicates the resistance value between the measuring terminals.

In order to measure the resistivity, the sample plated films were peeled. The resistivity ρ was calculated by the four probe method using the tungsten probe at intervals of 1 mm, and the results are shown in Table 2 (reference description: crystallite size (Da), average value of Hv). The plating thickness is obtained by averaging five points as measured by the fluorescent X-ray meter (SFT 3200, SII).

Compared to the resistivity of the commercially available (a) AuCo film, those of both (b) DC-Au film and (c) PC-Au film from the original pure gold plating bath were about 3 × 10⁻⁶ mΩ cm lower.

2.1.7 Prototype probe applied with gold plating

In a system in which elastic deformation is predominant, such as 0.1 N < P < 100 N (P is the contact load), the contact resistance R is expressed by the empirical formula of the following Eq. (6) [21, 22]. Here, ρ₁ and ρ₂ are the resistivity values of two materials in contact, H is the Brinell hardness, and P is the contact load.

It is suggested that the contact resistance can be lowered by the decrease of the resistivity proportional to its value. Utilizing this, the pulsed hard pure gold plating film was applied to the parts of the probe. Using two types of plated film probes, (a) AuCo film (conventional product) from commercial cyanide bath and (c) PC-Au film (fabricated product) from original bath, the results of the durability comparison during the contact resistance of 100,000 times are shown in Figure 6. In Figure 6, the vertical axis represents the contact resistance value [mΩ] of the probe; (a) AuCo plating and (c) PC-Au plating, and the horizontal axis represents every 1000 durability measurement times. The bar is maximum and minimum values at every 1000 times measurement, and the average values were plotted.

Calculating the average value of the contact resistance, (a) the resistance of the conventional product probe was 32.6 mΩ (deviation 2.07), and (c) that of the fabricated product probe was 21.6 mΩ (deviation 0.80). This fabricated probe has an advantage that the contact resistance value decreases by 11 mΩ and its variation is low, and improvement of electrical contact characteristics could be realized.

2.2.1 Carbon nanotube (CNT) and gold plating bath

A vapor-grown carbon fiber (VGCF; Showa Denko) which is one type of multiwalled CNTs, was used in this study. An Au-CNT plating bath was prepared by adding 0.05 M Na₃Au(SO₃)₂, 0.5 M Na₂SO₃, 100 pp. 2,2′-dipyridyl, and 0.2 g/l VGCF into ion-exchanged water. Moreover, 0.1 g/l trimethyl stearyl ammonium chloride (C₂₁H₄₆ClN) was added into the above solution. The mixture was stirred using an ultrasonic agitator (28 kHz) for 2 h. The volume of the plating bath was 100 cm³. The plating was performed at 60°C and pH 8 using a magnetic stirring agitator. The current density was 50 A/m² and a Pt-plated Ti mesh was used as the anode [8]. A copper (Cu) board with an exposed area of 200 mm² (10 × 20 mm²) was used as the substrate. The Cu substrate was plated using a Ni-P amorphous alloyed layer with a thickness of 3 mm. A Ni-P plating bath is a commercially available plating solution (Meltex). Before plating, the substrate was pretreated by alkali degreasing and acid treatment. For comparison, an Au plating film was prepared using an Au plating bath without adding CNTs. FE-SEM equipped with energy-dispersive X-ray spectroscope (EDX; JED-2300F, JEOL) was used to study the surface morphology and CNT content of the fabricated plating films. Hv was measured using a dynamic hardness tester (DUH-201, Shimadzu). A diamond indenter was vertically employed with a test load of 0.098 N, a load speed of 6.6 mN/s, and a hold time of 20 s at RT. The intrinsic resistance of the fabricated plating films was determined by a four-point probe method. Tungsten (W) probes were used with a probe spacing of 1 mm. Tribological properties were measured using a ball-on-plate-type reciprocating friction abrasion test machine (MMS-2419, Nissho-EW). A brass ball (8 mm in diameter) plated with 4.5 μm Ni-P and 1 μm hard-Au layers, was used as the counter surface. The test was repeatedly conducted in 50 cycles at a load of 0.5 N, a sliding length of 2 mm, a sliding speed of 0.5 mm/s at RT. The test was performed under ambient conditions without any lubricants. During the test, the friction coefficient was measured continuously using a load tester.

2.2.2 Gold-CNT composite plating film

The fabricated Au-CNT composite film appeared to be relatively black. Figure 7(a) and (b) shows SEM images of the fabricated Au-CNT composite film and the Au film, respectively.

CNTs were tangled with the Au matrix and protruded from the matrix surface. The surface morphology of the Au-CNT composite film was relatively rough owing to the existence of many voids. It was previously reported that, in the composite film, metal is easily separated at the CNT apex and defects of CNTs. Before CNTs are entirely coated, the plating process is promoted at the apex or defects of CNTs, resulting in the formation of voids in the deposited film [34]. The result of EDX analysis indicated that the content of CNTs in the Au plating film was about 4 mass %. In the plating process, CNTs could not be homogeneously dissolved in the solution without C₂₁H₄₆ClN even if it had been mechanically stirred for 48 h, and consequently, the plating bath could not be prepared. From the above results, C₂₁H₄₆ClN was found to be an effective additive for Au-CNT composite film fabrication using a non-cyanide bath, which is an environmentally friendly method.

2.2.3 Gold plating bath and zeta potential

The mechanism of the eutectoid composition of the Au-CNT composite film is assumed to be based on the deposition of CNTs by electrophoresis. Normally, the zeta potential of CNTs decreases and easily becomes negative when the solution becomes basic [45, 46]. Although CNTs are modified by N-doping or heat treatment, their zeta potential becomes negative at a basic pH of 8 [47, 48]. Since the CNTs repulsed by the cathode during reduction react, it is difficult to combine CNTs in a basic solution. At this stage, although the role of the additive has not yet been clearly elucidated, the results obtained in this study apparently show that the additive is important for Au-CNT composite film fabrication.

Figure 8 shows the VGCF zeta potential of 0.05 g/L under various potential of hydrogen (pH) in three solutions; (a) conventional dispersant of 0.5 mM in water, (b) fabricated dispersant (C₂₁H₄₆ClN) of 0.5 mM in water, (c) plating solution diluted 20 times.

The hydrophobicity of the additive is likely to effectively provide a positive zeta potential for CNTs. This was also evidenced in the plating process. The zeta potential of CNTs was found to be positive in the solution, and consequently, CNTs with a positive zeta potential were attracted toward the plating cathode.

2.2.4 Hardness and electrical properties of Au-CNT composite film

Table 3 shows the Hv and resistivity of the Au-CNT composite film and the Au film, respectively. The Hv is the average measured at 10 points. In general, to measure the Hv accurately, film thickness is required to be 10 times higher than the depth of the indenter pressed into the film. Thus, it should be noted that the measured value includes the substrate effect.

The Hv of the Au-CNT composite film and the Au film were Hv 133 and Hv 95, respectively. The fabricated Au-CNT composite film was harder than the Au film by 1.4 times.

For intrinsic resistance measurement, while the film thickness is sufficiently small compared with the probe spacing, the intrinsic resistivity ρ can be expressed by ρ = (πt/ln2)/(V/I) [20], where value t is the average film thickness measured at five points by X-ray fluorescence analysis (SFT3200, SII), and (V/I) is the average resistance of the thin film measured at three points. The resistivity of the Au-CNT composite film and the Au film were 5.6 and 4.7 × 10⁻⁶ mΩ cm, respectively. The fabricated Au-CNT composite film showed a higher resistivity than the Au film by 1.2 times. The voids and the interface between Au and CNTs are likely attributed to the Vickers hardness and the resistivity.

2.2.5 Friction coefficients of the Au-CNT composite film

Figure 9 shows the friction coefficients of the Au-CNT composite film and the Au film, as a function of sliding length against Sn ball of 8 mm diameter.

In the case of the Au-CNT composite film, the friction coefficient gradually decreased toward the sliding length. After the test, at a total sliding length of 200 mm (50 repeated cycles), the friction coefficient was 0.28. On the other hand, in the case of the Au film, the friction coefficient increased to 0.58 and gradually decreased at a sliding length of 130 mm. The Au-CNT composite film showed a lower friction coefficient than the Au film.

Figure 10(a) and (b) shows SEM images of the Au-CNT composite film and the Au film after the wear test, respectively.

The worn area of (b) the Au film was relatively large compared with that of (a) the Au-CNT composite film. The track with a width of approximately 200 mm was found on the surface of the Au film. The entire surface of the worn area of the Au film was damaged, while that of the Au-CNT composite film was partly damaged. CNTs still remained in the worn area of the Au-CNT composite film, lying transversely [Figure 10(a) inset]. Adhesive wear is one of the friction phenomena. The smaller the surface area, the weaker the force of the adhesion shear [49]. The small contact area and transversely lying CNTs seem to contribute to the low friction coefficient.