Copper Immersion Deposition on Magnesium Alloy: The Effect of Fluoride and Temperature

The magnesium alloy substrates used throughout the experiments were AZ91 magnesium alloy (including AZ91D and AZ91E). Their nominal compositions are shown in Table I. The substrate was cut into pieces. The pretreatment of the substrate was subsequently carried out using the following precedure

In this pretreatment precedure the substrate of magnesium alloy was glass-beaded for 10 s at a pressure of 450 kPa. The stream of glass beads bombarded the substrate in the normal direction to the surface from a distance of about 10 cm. The substrate was then cleaned in isopropanol for 6 min with sonication. Finally, alkaline degreasing was performed in a solution containing at 75°C for 6 min, followed by thoroughly rinsing in deionized water.

After the pretreatment of the substrate, two parallel processes, including copper immersion deposition and HF etching was followed (see below).

Following the pretreatment, the copper immersion deposition was immediately performed by transferring the pretreated substrate into a bath containing in various concentrations, unless specified otherwise. The immersion-coated samples were then examined by SEM using the mode of backscattered electron (BSE). The resultant BSE images were processed with Image–pro Plus software (Media Cyberhetics, Inc.) to obtain the surface coverage of the copper deposition.

After pretreatment, the HF etching process was performed by immersing the pretreated substrate into a solution that only contains hydrofluoric acid in various concentrations. It is believed that during the HF etching process, a surface film containing fluoride is formed. The characterization of such film is expected to benefit the understanding of copper immersion deposition.

OCP measurements were performed during both the copper immersion deposition and the HF etching process, while electrochemical impedance spectroscopy (EIS) was only carried out during the HF etching process. For both OCP and EIS measurements, the substrate surfaces were sealed using epoxy resin, except for a effective working area exposed to solutions. The working areas were polished using 320 grit wet emery paper followed by the pretreatment procedure as specified previously. After thoroughly rinsing in deionized water the specimens were immediately transferred into solutions for the OCP or EIS measurements.

OCPs were recorded vs. time in a cell with two electrodes, a working electrode and a reference electrode. For EIS a cell fitted with three electrodes, working electrode, counter electrode (Pt gauze with large surface area of ca. ), and reference electrode, was employed. The reference electrode used was (3 M), 0.21 V vs. a reference hydrogen electrode (RHE) at 25°C, connected to the electrolyte through a salt solution bridge of 3 M KCl. The distance between the reference electrode and the working electrode was maintained at . In this paper, all potential values reported are referenced against the (3 M), unless specified otherwise.

The EIS measurements were performed in a frequency range from 100 kHz to 1 Hz, with a bias potential of 0.0 V (vs. OCP) and a 5 mV ac perturbation. An EG&G 273A potentiostat was used, coupled with a M5210 lock-in amplifier.

XPS analyses were carried out using a Kratoes Axis Ultra instrument. Monochromatic Al X-rays were used and the pressure in the spectrometer was . Binding energy was referenced to the C 1s peak at 285 eV and the spectrometer was calibrated with Cu at 932.62 eV and Au at 83.96 eV. Spectral deconvolutions were performed after a Shirley background subtraction.

All the chemicals [analytical reagent (AR) grade], including NaOH, , , HF, and , were supplied by Sigma. In this study, deionized water prepared with a Millipore Elix 10 water deionization system was used for solution preparations.

The change in potential with time was monitored to study the electrochemical behavior of magnesium alloy during the HF etching process and the copper immersion deposition, respectively.

HF etching process

Figure 1 illustrates the results obtained during the HF etching process in solutions only containing HF in various concentrations between 2.2 and 5.5 M. Generally, after immersing the sample into solution, the measured OCP gradually shifts to a more noble value (the plateau potential), attributed to the buildup of a surface film on magnesium alloy.¹⁹ While the time for the potential to reach the plateau value is little affected by the HF concentration, the plateau potential is increased with the increase of HF concentration. In the higher hydrofluoric-containing bath (e.g., 5.5 M HF) the plateau potential was . In 2.2 M HF, however, it only attained approximately . Shown in Fig. 2 is the effect of temperature on OCP for 3.3 M hydrofluoric acid. It is observed that the plateau potential during the HF etching process increases with increasing bath temperature.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Copper immersion deposition

OCP–time curves were also collected during copper immersion deposition in solutions containing in various concentrations, as shown in Fig. 3, 4 and 5. Similar to the HF etching process, the OCP during copper immersion deposition shifts to a more noble value with increasing deposition time at the initial stage and subsequently plateaus. However, the time for the potential to reach the plateau value during copper immersion deposition (e.g., Fig. 3) is much shorter than that recorded during the HF etching process (Fig. 1). More significantly, the plateau potentials for copper immersion deposition jumped to approximately 0.1 V vs. or vs. RHE, close to a potential of redox reaction ( vs. RHE),²⁰ indi cating the presence of surface coatings.^{9, 21}

Zoom In Zoom Out Reset image size

The effect of HF concentration is significant in terms of both the plateau potential and the time for the potential to reach the plateau value, as observed by comparing the plot for 2.2 HF concentration bath with the curves for the other concentrations (Fig. 3). In addition, the plateau potential attained during copper immersion deposition is also related to the copper ion concentration in the bath, as shown in Fig. 5. It seems that the OCP is not significantly affected by temperature (Fig. 4).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Overall, for both HF etching process and copper immersion deposition, the OCP results show significant changes in the plateau potential under various conditions. These changes reflect information originating from the surface film formation and deposition process. However, a realistic interpretation of such information requires input from other experiments (e.g., EIS, XPS, SEM, etc.).

EIS measurements were conducted to obtain more information during the HF etching process. Shown in Fig. 6 are the impedance spectra in a solution containing 3.3 M HF at various temperatures between 25 and 70°C. The spectra are characterized by a capacitive loop in the frequency range studied. A simple, one time constant equivalent circuit (insert, Fig. 6) may be applied to represent the capacitive loop, i.e., a solution resistance, , in series with a parallel RC circuit with a resistance, , and a capacitance, . The capacitive spectra were fitted using a simple Powersine EIS software (EG&G Instruments, Inc.). The fitting process only gives the values of resistance with the fitted circle. The capacitance values were therefore obtained by , where is the angular frequency at the maximum point of the Nyquist plot. The values of and are plotted against HF concentration or bath temperature, as shown in Fig. 7 and 8.

Zoom In Zoom Out Reset image size

Capacitances obtained in this study, range between 8 and (Fig. 7 and 8). The values are below that anticipated for a double-layer capacitance (around ).²² Generally, such capacitance values were attributed to both a charge transfer and a film effect.^{22, 23} In this study, the impedance measurements were performed after the magnesium samples were immersed in the solution for 90 s. The period for the impedance measurement coincided with the plateau stage on the OCP plots (Fig. 1 and 2). During this period, the surface film had already built up. The capacitance obtained was therefore mostly attributed to the film effect and the associated essentially represents the film resistance. As such, the data show that with increase of both temperature (Fig. 7) and HF concentration (Fig. 8), the film resistance, , is increased.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The EIS analysis indicates a possible film formation on the magnesium alloy surface during the HF etching process, consistent with OCP results.

General survey spectra from XPS measurements of the magnesium surface treated before and after the HF etching process are shown in Fig. 9. It detected magnesium, aluminum, fluorine, and oxygen as the main composition of the surface film. The presence of sodium, chlorine, sulfur, and silicon is attributed to a glass beading process or to possible surface contamination. From the general survey spectra, it is obvious that before HF etching, the magnesium surface consists of magnesium and aluminum oxide/hydroxide. After the HF etching process, a fluorine peak appeared at approximately 686 eV, indicating the presence of fluoride. The surface concentration of the different elements measured from the XPS is listed in Table II.

Zoom In Zoom Out Reset image size

In order to analyze the chemical states in the surface film, peak fittings for the high-resolution XPS spectra of Mg 2p, O1s, and F 1s core level regions were carried out as shown in Fig. 10. To perform the fitting process, only one peak was needed to fit the spectra for F 1s, but three peaks of Mg 2p BE(I), Mg 2p BE(II), and Mg 2p BE(III) were required for Mg 2p, and two peaks of O1s BE(I) and O1s BE(II) were used for O 1s.²⁴ Table III summarizes the fitting results, which give not only the absolute binding energy (BE) of different core levels at various conditions, but also the difference of BEs between the levels. By using the difference of BEs, many problems in choosing an appropriate reference level can be avoided when comparing the results from various experiments and the literature.

Zoom In Zoom Out Reset image size

Table III. Chemical shifts (in electron volts) for Mg 2p, O 1s, and F 1s core level regions for samples 02, 03, and 04 (see Table II for the sample treatment conditions).

Sample	M 2p BE(I)	M 2p BE(II)	M 2p BE(III)	O 1s BE(I)	O 1s BE(II)	F 1s	O 1s BE(I)-Mg 2p BE(I)	F 1s-Mg 2p BE(II)	F 1s-Mg 2p BE(III)
02	50.8	51.9	53.1	532.2	533.8	686.3	481.4	634.4	633.2
03	50.9	52	53.3	532.3	534.1	686.3	481.4	634.3	633
04	50.8	52	53.2	532.1	534	686.3	481.3	634.3	633.1
Data from Ref. 24							481.3 to 481.7	634.5 to 634.7	632.9
Ascribed to

According to Verdier et al. ,²⁴ the Mg 2p BE(I) is related to O 1s BE(I). The difference, , measured between the O 1s BE(I) and Mg 2p BE(I), ranges from 481.3 to 481.4 eV (Table III), which is in good agreement with the previously reported values²⁴ for a magnesium hydroxide, . This suggests that instead of MgO, was present at the surface of different samples. For the higher BE, Mg 2p BE(II), its binding energy difference from the fluorine core level, , is between 634.3 and 634.4 eV (Table III). They could be ascribed to -type magnesium hydroxyfluoride, consistent with Verdier et al.²⁴ Furthermore, another peak, Mg 2p BE(III), was measured, having a BE of 53.1-53.3 eV. Its BE difference from fluorine core level, , is 633-633.2 eV, suggesting the presence of at the surface film according to the BE difference reported for a standard .²⁴

The oxygen spectra (Fig. 10) also shows a high BE peak, O1s BE(II), at approximately 534 eV for the different samples, which could be ascribed to oxygen atoms bound to carbon atoms in the contamination layer.^{25, 26}

HF etching process

SEM was employed to examine the surface morphology of magnesium alloy treated under various conditions. Shown in Fig. 11 and 12 are the secondary electron images obtained after the HF etching process. For comparison, the image obtained from a sample before the HF etching process is also presented (Fig. 11f). It shows that the surface morphologies formed by the HF etching process exhibit a typical "mud-like" texture. "Mud-like" texture is usually produced by precipitation of corrosion products. As corrosion products dehydrate, the precipitate contracts, forming cracks.

Zoom In Zoom Out Reset image size

In this study the observed mud-like texture was formed by the precipitation of magnesium-fluoride species generated during the HF etching process. This may give an indication of the significance of magnesium dissolution. Thus, the effect of temperature on the surface texture of magnesium after the HF etching process is significant: more dissolution of magnesium is observed at lower temperature (e.g., 25°C) than higher temperature (e.g., 70°C). In particular, there is little difference between the images of Fig. 11e, HF-treated at 70°C, and 11f, before HF treatment, which is evident that magnesium dissolution at 70°C rarely occurs. With regard to the HF concentration effect (Fig. 12), it is also clear that the dissolution of magnesium in the higher HF concentration (e.g., 5.5 M HF) is less observable than that in lower HF concentration (e.g., 1.1 M HF).

Zoom In Zoom Out Reset image size

Copper immersion deposition

The BSE images of copper immersion deposition are presented in Fig. 13 and 14, showing the effects of HF concentration and bath temperature, respectively. The surface coverage of copper deposition derived from these images is shown as the inserts in the corresponding figures. Figure 13 indicates that surface coverage first increases with increasing HF concentration and then decreases after reaching a maximum point at 2.2 M HF. With regard to temperature effect it is clear that the copper surface coverage decreases with increasing bath temperature (Fig. 14).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

 From the experimental results, it is clear that both temperature and HF concentration significantly affect the film formation on the magnesium surface during the HF etching process.

According to the OCP results (Fig. 1 and 2), the value of plateau potential ranges from to , depending on HF concentration and bath temperature. The standard potential of electrode reaction for is vs. the saturated hydrogen electrode (SHE).²⁰ In chloride aqueous solutions the actual corrosion potential of magnesium is usually increased to to vs. SHE, attributable to the formation of a surface film of or MgO.^{27, 4} In this study, the further increase in OCP in the presence of fluoride instead of chloride is due to the fluoride-containing species in the film.

In the literature, a film formation on magnesium alloy surface is frequently reported^{3, 28} although the exact film structure and composition may be different under different surface treatment conditions. In this study, the XPS analysis results show that the surface films formed by the HF etching process are rich in fluorine (Table II) and presumably have a structure of (Table III). The aluminum content in the film is relatively low, consistently for all samples, after the HF etching process. This may be attributed to the stability of magnesium and aluminum in hydrofluoric acids. In HF solutions, aluminum is reported to be unstable and may form complexes, dissolving into the solution.^{29, 30} In contrast, magnesium may form to develop a surface film.^{19, 27} The resultant may also be followed by a exchange or hydration process, leading to the formation of a more stable surface film containing .²⁷

SEM observations confirmed such a film formation on the magnesium surface. The films are more compact at higher temperatures (Fig. 11) or in solutions with higher HF concentration (Fig. 12). Both trends are also reflected by the OCP results, in terms of the plateau potential (Fig. 1 and 2), and the EIS results where the resistance, , increases as HF concentration or temperature increases (Fig. 7 and 8). It is notable from Table II that fluorine content in the surface film was increased with the increase of either HF concentration (sample 02 to 03) or temperature (sample 02 to 04). The higher fluoride content in the film could be one reason for the observed compact surface film. Based on the above discussion it is expected that the HF concentration or bath temperature could render control of the magnesium dissolution for achieving a good quality coating during the copper immersion deposition.

The copper immersion deposition on magnesium involves charge transfer between reacting chemical species of . More specifically^{20, 31}

When magnesium substrate is immersed in a solution containing copper ions, magnesium atoms (less noble) dissolve as the anodic reaction and spontaneously are replaced by copper atoms from the solution (copper deposition). As soon as the copper deposition begins, the surface of the magnesium substrate becomes a mosaic of anodic (magnesium) and cathodic (copper) sites. Theoretically, the process continues until almost the entire substrate is covered with copper.

However, it was observed that in a simple aqueous solution only containing copper ions, the copper deposition process was accompanied by violent magnesium dissolution, impeding the development of good quality coating. According to the previous discussion, hydrofluoric acid is beneficial to the surface film formation on magnesium, which protects magnesium from violent dissolution. It is therefore added to the immersion-deposition bath to render a control of the magnesium dissolution, hence copper deposition. After adding hydrofluoric acid to the copper immersion bath (e.g., ), the deposition process continues until the entire substrate is covered with copper and fluoride/hydroxide film. At this point anodic dissolution virtually ceases, resulting in the termination of immersion deposition, which can be manifested by the coverage-time plot.¹⁴ Thus, the magnesium surface after the copper deposition process is composed of two areas, i.e., coated areas (white) and uncoated areas (dark), as shown in Fig. 13 and 14.

According to the above discussion, a model (Fig. 15) is proposed for the understanding of the copper immersion deposition process. In this model, before immersion (step 1) the magnesium substrate is covered with a magnesium hydroxide film. The film generally possesses a poor stability and is not perfectly uniform.⁶ It may also contain imperfections/defects. After immersing the magnesium substrate into the copper deposition solution (step 2), anodic magnesium dissolution occurs, preferentially at those regions/sites where the protection of magnesium substrate by the film is relatively weak. Correspondingly, in the other regions/sites where the potential is nobler as compared to the anodic sites, copper deposition takes places. During the same period, a new film rich in fluoride on the uncoated areas may be formed by the reaction of ^{19, 27} and by the destruction/transformation of hydroxide film. These processes continue until the entire substrate is covered with copper (coated areas) and fluoride/hydroxide film (uncoated areas), as depicted in step 3.

Zoom In Zoom Out Reset image size

The relation between HF concentration and copper surface coverage is given in Fig. 13, which shows a peak at approximately 2.2 M HF. This can be explained according to the proposed model: in a lower HF-containing bath (e.g., 1.1 M HF), the surface film formed on magnesium possessed a loose and discontinuous feature (Fig. 12a) with poor protection. The dissolution of magnesium is therefore violent, resulting in nonadherent and spongy copper deposits that were subsequently partially removed from the substrate surface by violent hydrogen evolution. With an increase of HF concentration, the surface film formed became increasingly more dense and compact (Fig. 12) with a higher content of fluoride species (Table II). Such a compact film, simultaneously formed during copper deposition, not only prevented violent magnesium dissolution but also enabled copper to deposit as fine and adhesive crystallites (e.g., Fig. 13b). With the further increase in HF concentration, however, copper deposition was only able to take place at a very early stage, because the rapid film formation limited the available anodic sites on the surface. As a result, a decrease in copper coverage is observed (Fig. 13).

Figure 14 indicates that the surface copper coverage on the substrate decreased with the elevation of bath temperature. For an understanding of the mechanism of such temperature effects, the enthalpy change of the reactions involved in copper immersion deposition is listed below

where is the enthalpy change of reactions at 298.5 K. The formation of the corrosion product, , subsequently reacts with or water to form a surface film ^{12, 27}

As indicated, Reactions 3, 4 are exothermic, while 6, 7 are endothermic. Therefore, increasing temperature is unfavorable to both magnesium dissolution (Reaction 3) and copper reduction (Reaction 4), while the film formation is favored by either forming fluoride (Reaction 6) or oxide film (Reaction 7). As a result the copper surface coverage decreases with increasing bath temperature.