Optimization and performance analysis of copper immersion coating on AZ91 magnesium alloy.

An orthogonal experimental methodology was used to study a copper immersion coating on AZ91 magnesium alloy from a hydrofluoric acid-containing bath. Factors such as hydrofluoric acid concentration, sonication time, chemical etching, and activation process were considered in the design. The analysis of variance on the orthogonal experimental results was performed, resulting in an optimal condition for achieving a uniform copper coating with high coverage. Hydrofluoric acid was found to be an essential component in the copper immersion coating bath and a possible mechanism was suggested to explain its significance in terms of magnesium film destruction and magnesium dissolution.

Keywords: Cooper immersion coating, magnesium alloy, orthogonal design

**********

Magnesium and its alloys have the highest specific strength of the structural metals. They are regarded as a solution for mass reduction of vehicles such as automobiles and aircrafts. However, challenges exist due to their low corrosion and wear resistances, although various coating technologies have been applied. Many reports (1-5) have suggested that magnesium alloys can be electroless or electro-plated; but satisfactory process sequences have yet to be developed to meet various industrial requirements.

Magnesium is highly reactive. It forms oxide on its surface even more readily than aluminum. Special procedures are therefore required to obtain an adherent and corrosion-resistant coating. Generally, the procedure for plating magnesium alloys is similar to that employed for aluminum alloys. (2,6) A chemical etching process is usually applied to remove the oxide film, followed by an activation step to activate the surface. Immersion zinc coating is then applied to prevent oxide from reforming. (2) However, the zinc immersion pretreatment process has been criticized for the precise control required to ensure adequate adhesion. In many cases nonuniform coverage on the surface is seen with spongy nonadherent zinc deposits on the intermetallic phase of the base alloys. A copper cyanide strike that must follow has also been of concern for a number of reasons. (7)

Our recent research focuses on the exploration of a simple copper immersion coating on AZ91 magnesium alloy in a hydrofluoric acid bath. Challenges encountered during preliminary studies include low surface coverage. Given the complex effect of different parameters, it was necessary to consider a systematic approach to optimize the process. In any experimental setup, a systematic analysis of process parameters is essential to find out the significant variables, which have direct impact on the process quality. Orthogonal experimental design is known to be an effective tool for the purpose, and thereby optimizing the process outputs. (8-10) This article, therefore, applies orthogonal design for the optimization of the copper immersion coating process. The influence of different parameters such as chemical etching and activation processes, as well as immersion bath conditions, have been analyzed and the optimal conditions have been established.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

EXPERIMENTAL

The substrate used in this study is AZ91D magnesium alloy, supplied by Norsk Hydro Canada, with a nominal composition of Al (8.5-9.5%), Zn (0.45-0.9%), Mn (0.17-0.40%), Si (<0.08%), Cu (<0.025%), Ni (<0.001%), Fe (<0.004%), and Be (0.0005-0.003%). The specimens were prepared with dimensions of 10 X 8 X 1.5 mm. All chemicals of ACS reagent used in this work were provided by SIGMA. Deionized water (>15 M[ohm] c[m.sup.-1]) used throughout the experiments was prepared with the Millipore Elix 10 Water Deionization System (Millipore Corporation).

Pretreatment of the samples was conducted using glass beading or mechanical polishing. For mechanical polishing, the sample was polished to 400 grits using wet emery papers. During the glass beading process, the sample was blasted with glass beads for 10 sec, using an ECONO-Finish Blast Cabinet (EF-2436, Empire Abrasive Equipment Company). Air pressure applied for the glass beading process was about 450 kPa. The stream hit the sample in a normal direction from a distance of about 10 cm. After glass beading, loosely sticking particles were removed from the blasted surface with compressed air.

Samples were subsequently degreased in an alkaline solution containing 60 g/l NaOH + 10 g/l N[a.sub.3]P[O.sub.4] at 75[degrees]C for six minutes and thoroughly rinsed in deionized water. Rinsed samples were then immediately immersed into a hydrofluoric acid bath for copper immersion coating. Prior to the immersion coating, some samples were subjected to chemical etching and followed by an activation process. The conditions of the chemical etching and activation processes are specified in Table 1.

[FIGURE 3 OMITTED]

All immersion coating processes were carried out in a hydrofluoric acid bath containing copper sulfate at room temperature for 10 min, unless specified otherwise. The volume of the solution placed in the immersion bath was 100 ml. The immersion baths were classified as a "still bath" and a "sonicated bath," depending on whether sonication was applied. The sonication system used in this experiment is the Model 9333 ultrasonic bath (Lab-Line Instruments Inc.) with a frequency of 40 kHz.

A Hitachi S3500N Variable Pressure SEM was employed to examine the surface morphologies of the immersion coating. Copper coating coverage was measured by processing the SEM images, analyzed with Image-Pro Plus software (Media Cyberhetics Inc.).

RESULTS

Preliminary Experimental Results

In order to apply copper deposits on magnesium alloy using an immersion coating process, a survey experiment was conducted in various immersion baths. The preliminary results revealed that it was difficult for copper to deposit and anchor on the magnesium alloy due to its chemical reactivity. In order to enable the formation of copper coating, hydrofluoric acid was employed in the immersion coating bath and was found to be an essential component. Shown in Figure 1 are the back-scattered electron images of copper immersion coatings in the hydrofluoric acid bath. In a 2.2 M HF immersion bath, nodular copper deposit was formed on the magnesium surface (Figure 1a). When the immersion coating was performed in a 5.5 M HF bath, however, only a small amount of copper was deposited (Figure 1b). It is therefore concluded that the copper deposition is directly related to the HF concentration under the investigated range. Moreover, it was observed during our experiments that HF concentration lower than 2 M would result in violent magnesium dissolution and spongy nonadherent copper deposits on the magnesium alloy.

Furthermore, back-scattered electron images of the copper coatings from the hydrofluoric acid bath under various conditions were also obtained (Figures 2-4). After analyzing these images the following observations can be made:

[FIGURE 4 OMITTED]

(1) Chemical etching, in combination with the subsequent activation process, has a significant effect on the copper immersion coating. Copper coverage on the sample etched by a chromium oxide solution with a subsequent alkaline activation (Figure 2b) is much higher than that etched in an oxalic acid, followed by a fluoride activation (Figure 2a).

(2) Copper coverage on the glass-beaded surface (Figure 3a) is much higher than that on the polished sample (Figure 3b). The images also indicate that the uniformity of the copper coating on the former is much better than on the latter.

(3) A comparison of Figures 4a and 4b indicates that the application of sonication to the immersion coating bath promotes copper growth significantly. The effect of sonication on the copper immersion coating and its mechanism will be discussed in a separate article.

In addition, the effect of immersion coating time on the copper coverage was studied, as shown in Figure 5. The copper coverage increased with increasing immersion coating time at the initial stage (within two minutes) and then plateaued with an extended coating time.

The preliminary studies of the copper immersion coating addressed multiple effects of different factors. To find the significant variables affecting the immersion coating quality, it is desirable to apply an orthogonal design for optimization of the coating process. From the above discussion, the immersion coating time has little influence on the surface coverage beyond two minutes of deposition and the glass beading surface pretreatment is superior to the polishing. As such, the immersion coating time was fixed at 10 min and glass beading was conducted on all the samples throughout the orthogonal experiments. Four factors--HF concentration, sonication time, chemical etching, and activation processes--were considered in the design.

Orthogonal Experiment Results

Based on the preliminary experimental results, four factors, at four levels each, were selected for the orthogonal experiment, as shown in Table 1. The experimental arrangements using an orthogonal array are shown in Table 2. In this table, each process parameter was assigned to a column and 16 parameter combinations were available.

The objective of the orthogonal experiment was to determine the optimal combination of the immersion coating process variable so as to obtain a desired copper immersion coating on the magnesium alloy. Such a copper coating can act as an intermediate layer to protect the magnesium surface from dissolving and oxidizing during subsequent electroless deposition or electroplating. Our recent studies found that a uniform copper immersion coating with an adequate coverage is essential for achieving the desired electroless deposition on the immersion-coated magnesium surface. The coverage of the copper immersion coating is, therefore, the most important quality characteristic to be measured. Experiments according to the orthogonal arrangement were conducted. To obtain more reliable results, each treatment in the orthogonal array was repeated three times, with the results shown in Table 2.

[FIGURE 5 OMITTED]

The mean level response of each factor, [R.sub.pl], was calculated as:

[R.sub.pl] = [1/k] [k.summation over (i=1)] [bar.[eta].sub.i] (1)

where k (= 4) is the number of treatments involving level, l (l = 1 to 4), for factor, p (parameter, p = A, B, C, and D) and [bar.[eta].sub.i] is the average coverage corresponding to each treatment.

According to equation (1), the calculated results are listed in Table 2 and plotted in Figure 6. These mean level response curves suggest that, within the investigated level ranges, the optimal combination of the coating process parameter levels to obtain the high coverage is [A.sub.1], [B.sub.3], [C.sub.3], and [D.sub.4] (Figure 6). It is noted that further decreasing HF concentration (Factor A) or increasing sonication time (Factor D) may further increase the coating coverage. However, such conditions would deteriorate the immersion coating quality, as will be discussed in the Discussion.

Analysis of Variance

The main purpose of the analysis of variance (ANOVA) is to test the significance of design factors in affecting the quality characteristics. The total sum of square, S[S.sub.T'] is given by:

S[S.sub.T] = [n.summation over (i=1)][[bar.[eta]].sub.i.sup.2] - [1/n][[n.summation over (i=1)][bar.[eta].sub.i]][.sup.2] (2)

where n is the total number of treatments, e.g., n = 16. The sum of squares from the tested factor, S[S.sub.p'] can be calculated as:

S[S.sub.p] = [1/t][t.summation over (l=1)]([k.summation over (i=1)][bar.[eta].sub.i])[.sup.2] - [1/n][[n.summation over (i=1)][bar.[eta].sub.i]][.sup.2] (3)

where t (= 4) is the number of levels in factor, p.

The remainder of the sum of squares that represents the error term, S[S.sub.e'] is equal to

S[S.sub.e] = S[S.sub.T] - S[S.sub.A] - S[S.sub.B] - S[S.sub.C] - S[S.sub.D] (4)

The variances for a factor, [V.sub.p'] and for the error term, [V.sub.e'] are calculated by dividing the corresponding sum of squares by the relevant degrees of freedom, [D.sub.p] and [D.sub.e']

[V.sub.p] = S[S.sub.p]/[D.sub.p] [V.sub.e] = S[S.sub.e]/[D.sub.e] (5)

where [D.sub.p] = t-1 and [D.sub.e] = (n-1)-4(t-1). Finally, to examine the significance of the factors, values for the F-test are obtained by dividing the variance of a factor by that of the error term,

[F.sub.p] = [V.sub.p]/[V.sub.e] (6)

The results of the analysis of variance based on equations (2-6) are shown in Table 3. According to the analysis of variance, the effect of HF concentration on the coating coverage is statistically significant at the confidence level of 0.25.

DISCUSSION

Both level response (Figure 6) and the analysis of variance (Table 3) suggest that the effect of HF concentration on the coating coverage is the most significant factor. As has been discussed in the Results section, hydrofluoric acid is an essential component in the immersion bath to enable the formation of copper coating on the magnesium alloy. Within the HF concentration range studied, the mean level response indicates that the lower the HF concentration, the higher the coverage (Figure 6). Examination of surface morphologies of the copper immersion coatings obtained under different experimental conditions showed that higher coverage was usually accompanied with higher magnesium dissolution. For example, in the immersion bath containing 5.5 M hydrofluoric acid, very little material was dissolved from the magnesium surface during the immersion coating process, as indicated by the remaining erosion marks in the uncoated area, which were caused by the glass beading process (11) (Figure 1b); correspondingly, only a small amount of copper was deposited on the magnesium surface. In contrast, in a 2.2 M hydrofluoric acid bath, significant magnesium dissolution occurred, as indicated by the rough morphologies in the uncoated area (Figure 1a); the magnesium dissolution led to the formation of clusters of nodular copper deposits on the magnesium surface with a high surface coverage. The dissolution of magnesium as an anodic reaction provides the driving force for the cathodic reaction of copper reduction and deposition:

[FIGURE 6 OMITTED]

Anodic reaction: Mg = M[g.sup.2+] + 2e (8)

Cathodic reaction: C[u.sup.2+] + 2e = Cu (9)

According to the Pourbaix diagram, (12) magnesium dissolves readily as M[g.sup.++] in aqueous solutions with a pH below 11. Above pH 11, its dissolvability gradually decreases with the increase of pH due to the formation of a stable magnesium hydroxide film. (13)

In hydrofluoric acid solution, however, pH only varies slightly with HF concentration under the investigated range. The dissolvability of magnesium in hydrofluoric acid is, therefore, mainly determined by the HF concentration rather than the pH. At low HF concentrations, dissolution of magnesium is significant. As the HF concentration increases, the dissolvability of magnesium decreases and becomes less visible. It was observed during our experiments that, at HF concentrations greater than 5 M, very little dissolution of magnesium occurred. This is presumably due to the formation of a relatively insoluble and protective film of magnesium fluoride, Mg[F.sub.2'] according to the following reaction:

Mg + 2[F.sup.-] = Mg[F.sub.2] + 2e (10)

Based on the previous discussion, in order to achieve a uniform coating with high surface coverage, the hydrofluoric acid concentration can be adjusted to control the formation of the fluoride film and the dissolution rate of magnesium. On one hand, the formation and growth of the surface film is needed to protect the magnesium from being violently dissolved, which, in turn, would create difficulty for the reduction and anchoring of copper. A very concentrated HF bath (e.g., > 5 M), on the other hand, will lead to a low dissolution rate in magnesium and, hence, a very low copper coverage (Figure 1b). However, although higher coating coverage in a low hydrofluoric acid concentration bath (e.g., 2.2 M) can be achieved, a spongy nonadherent copper deposit with voids/grooves is unavoidable (Figure 1a). As such, a concentration range of 3-4 M in hydrofluoric acid should be employed to achieve a uniform and adherent copper immersion coating.

The application of sonication to the immersion-coating bath is also an important factor. The back-scattered electron images (Figures 4a and b) showed that copper growth was significantly promoted by applying sonication to the immersion-coating bath. According to the level response, it seems that the effect of sonication on the coating coverage is relatively low at the short times from 0.5 to 2 min (Figure 6). However, there was an apparent increase in coating coverage when the sonication time was increased to 10 min. In this study, the degreasing process was performed in the alkaline solution, which possibly caused the formation of magnesium hydroxide film on the sample surface. (12) During the subsequent immersion coating, an electrochemical process including the hydroxide film destruction and/or transformation, accompanied with magnesium dissolution, could occur. Such a process presumably predominated the immersion coating process at the initial stage, as indicated in the coverage-time curve obtained in a still bath (Figure 5). The curve shows that initially the coating coverage increases with increasing coating time and then reaches the plateau, possibly due to the formation and subsequent thickening of the hydroxide-fluoride film (14) on the magnesium surface. In the sonicated bath, however, the sonication might have produced jets or cavitations, which continually attack the magnesium surface film, leading to continued magnesium dissolution and, in turn, copper deposition. The copper coverage was, therefore, increased with increasing immersion coating time in the sonicated bath, as indicated from the mean level response curve (Figure 6). However, it was noticed during our experiment that prolonged sonication to the immersion bath would result in large voids. Thus, the sonication time should not exceed 10 min.

For the other two factors investigated, the optimal conditions for obtaining better coverage are obvious, i.e., [B.sub.3] for etching bath (E[B.sub.3]) and [C.sub.3] for activation bath (A[B.sub.3]). From the discussions, the optimal experiment conditions for obtaining a uniform copper immersion coating with an adequate coverage can be identified as: [A.sub.2], [B.sub.3], [C.sub.3], and [D.sub.4]. For example, the sample is chemically etched using the chromium oxide solution ([B.sub.3]), followed by the alkaline activation ([C.sub.3]), and then the copper immersion coating is performed in the 3.3 M hydrofluoric acid bath ([A.sub.2]) with a sonication time of 10 min ([D.sub.4]) from the beginning of the immersion coating process.

[FIGURE 7 OMITTED]

According to the optimal combination of factor levels ([A.sub.2][B.sub.3][C.sub.3][D.sub.4]), immersion coating on the magnesium alloy was conducted. Figure 7 shows the back-scattered electron image. It is apparent in the figure that the coating coverage is much higher than those obtained from experiments prior to the optimization using the orthogonal design, i.e., Figure 1.

It should be pointed out that the present study only investigated the main effects of the four factors, while their interactions have not been considered. The analysis of variance shown in Table 3 indicates that the significance levels (F-values) for the tested factors are low; even for the most significant factor (HF concentration), it is significant only at a confidence level of 0.25. However, experimentally, it has been observed that the coating coverage did show very significant changes as the factor levels changed. For example, a comparison of copper coatings obtained from the sonicated bath (Figure 1a) and the still bath (Figure 1b) clearly indicated that sonication significantly promoted the copper deposition. However, such a very significant effect was not fully reflected from the analysis of variance. The low significant contribution of the sonication factor from the analysis of variance might be attributed to interactions among the experiment factors. Further studies are required to clarify these interaction effects using a different experiment array.

CONCLUSIONS

Optimization of the copper immersion coating on magnesium alloy using orthogonal methodology was performed. After analyzing the effect of the tested parameters on the copper immersion coatings, an optimal experiment condition was identified as: [A.sub.2], [B.sub.3], [C.sub.3], [D.sub.4], i.e., the sample was chemically etched using chromium oxide solution ([B.sub.3]), followed by an alkaline activation ([C.sub.3]). The copper immersion coating was performed in a bath containing 3.3 M hydrofluoric acid ([A.sub.2]) with sonication for 10 min ([D.sub.4]) from the beginning of the immersion coating process. The coating coverage obtained under the optimal experimental conditions was significantly increased, as compared to those obtained from experiments prior to the optimization. Furthermore, the hydrofluoric acid concentration in the immersion bath was identified as possibly the most significant factor. A possible mechanism was suggested to explain the significance of hydrofluoric acid in terms of magnesium film destruction and magnesium dissolution.

Table 1 -- Coating Process (a) Parameters and Their Levels

                  Factors                      Levels
Symbol  Coating Parameter     Unit    1           2           3

A       HF concentration (b)  Molar   2.2         3.3         4.4
B       Etching bath (c)              E[B.sub.1]  E[B.sub.2]  E[B.sub.3]
C       Activation bath (d)           A[B.sub.1]  A[B.sub.2]  A[B.sub.3]
D       Sonication time (e)   Minute  0           0.5         2

        Levels
Symbol  4

A       5.5
B       E[B.sub.4]
C       A[B.sub.4]
D       10

(a) General coating procedure: Substrate--glass beading--degreasing in
60 g/l NaOH + N[a.sub.3]P[O.sub.4] 10 g/l at 75[degrees]C for 6 min--
chemical etching--activation--immersion coating (room temperature, 10
min).
(b) HF concentration is the hydrofluoric acid concentration in the
immersion coating bath (0.67 M CuS[O.sub.4] + xM HF: x = 2.2, 3.3, 4.4,
5.5 M).
(c) Etching baths (EB) were chosen as: E[B.sub.1]--without etching;
E[B.sub.2]--[H.sub.3]P[O.sub.4] (85%) 380 ml/l + [H.sub.2]S[O.sub.4]
(98%) 16 ml/l; E[B.sub.3]--Cr[O.sub.3] 180 g/l + Fe[(N[O.sub.3]).sub.3]
9[H.sub.2]O 40 g/l + KF 3.5 g/l; and E[B.sub.4]--Oxalic acid ([C.sub.2]
[H.sub.2][O.sub.4]) 10 g/l + wetting agent (62A from Entrone-Ont Inc.,
New Haven, CT) The etching process was performed at room temperature for
30 sec.
(d) Activation baths (AB) were chosen as: A[B.sub.1]--without;
A[B.sub.2]--100 g/l N[H.sub.4]H[F.sub.2] + 200 ml/l [H.sub.3]P[O.sub.4]
(75%); A[B.sub.3]--[K.sub.4][P.sub.2][O.sub.7] 100 g/l + N[a.sub.2]
C[O.sub.3] 30 g/l + NaF 5g/l; and A[B.sub.4]--340 ml/l HF (38%
hydrofluoric acid). The activation process was performed at room
temperature for 30 sec.
(e) Sonication was applied to the immersion bath, just before the sample
was put into the bath for copper coating, and then stopped at the
predetermined time (Factor D) in Table 1, leaving the immersion coating
process to continue until a total of 10 min had elapsed.

Table 2 -- Orthogonal Experimental Arrangement and Results

                                                                D
No. of              A            B             C           Sonication/
Treatment       HF conc./M  Etching Bath  Activation Bath  min

 1               2.2        E[B.sub.2]    A[B.sub.1]          0.0
 2               2.2        E[B.sub.3]    A[B.sub.2]          0.5
 3               2.2        E[B.sub.4]    A[B.sub.3]          2.0
 4               2.2        E[B.sub.1]    A[B.sub.4]         10
 5               3.3        E[B.sub.2]    A[B.sub.2]          2.0
 6               3.3        E[B.sub.3]    A[B.sub.1]         10
 7               3.3        E[B.sub.4]    A[B.sub.4]          0.0
 8               3.3        E[B.sub.1]    A[B.sub.3]          0.5
 9               4.4        E[B.sub.2]    A[B.sub.3]         10
10               4.4        E[B.sub.3]    A[B.sub.4]          2.0
11               4.4        E[B.sub.4]    A[B.sub.1]          0.5
12               4.4        E[B.sub.1]    A[B.sub.2]          0.0
13               5.5        E[B.sub.2]    A[B.sub.4]          0.5
14               5.5        E[B.sub.3]    A[B.sub.3]          0.0
15               5.5        E[B.sub.4]    A[B.sub.2]         10
16               5.5        E[B.sub.1]    A[B.sub.1]          2.0
[R.sub.p1] (b)  50          36            33                 23
[R.sub.p2]      34          24            25                 27
[R.sub.p3]      19          37            40                 28
[R.sub.p4]      14          23            21                 41
X (c)           36          14            19                 18

                Coverage (a)%,
No. of          [[eta].sub.i]   Average
Treatment       1   2   3       coverage [bar.[eta].sub.i]

 1              47  29  35               37
 2              40  34  55               43
 3              61  58  54               58
 4              65  71  48               61
 5              22  30  24               26
 6              60  65  61               62
 7               6   5   4                5
 8              40  37  57               45
 9              23  30  31               28
10              14  12  13               13
11              18  18  11               16
12              20  20  20               20
13               4   4   3                4
14              28  36  23               29
15              12  18  11               14
16              13  20  16               16
[R.sub.p1] (b)
[R.sub.p2]
[R.sub.p3]
[R.sub.p4]
X (c)

(a) Coating coverage was obtained by processing the back-scattered
electron images, using Image-Pro Plus software (Media Cyberhetics Inc.).
Three groups of data listed here were obtained from the three repeated
tests for each treatment in the orthogonal array.
(b) [R.sub.p1] is the mean level response for factor, p, at level, l
(l = 1-4, p = A, B, C or D).
(c) X is the extreme level difference (X = [R.sub.max] - [R.sub.min]).

Table 3 -- Result of the Analysis of Variance

Symbol  Coating Parameter  df  SS    [V.sub.p]  [F.sub.p]

A       HF concentration    3  2911  970        7.22
B       Etching bath        3   665  222        1.65
C       Activation bath     3   845  282        2.09
D       Sonication time     3   778  259        1.93
Error                       3   403  134        [F.sub.0.25](3,3) = 2.36
Total                      15  5603

Note: df is the degree of freedom, SS is the sum of squares, [V.sub.p]
(= S[S.sub.p]/df) is the variance of the parameter tested, and [F.sub.p]
= [V.sub.p]/[V.sub.e] [V.sub.e] is the variance of the error term.

ACKNOWLEDGMENTS

The present work was sponsored by the Natural Science & Engineering Research Council (NSERC) of Canada. The authors are grateful to John Nagata for his contribution to experimental work, Mike Meinert for his work on SEM measurement, and William Wells for the sample preparations. Thanks are also given to Mahmud-UI Islam for his internal review.

References

(1) Chen, J.H., Chang, C.C., and Lee, T.S., "Pretreatment for Plating on Magnesium Alloys," AESF SUR/FIN' 91, (1991).

(2) Dennis, J.K., Wan, M.K.Y.Y., and Wakes, S.J., "Plating on Magnesium Alloy Die Castings," Trans. Inst. Met. Finish, 63, 74 (1985).

(3) Fairweather, W.A., "Electroless Nickel Plating of Magnesium," Trans. Inst. Met. Finish, 75, 113 (1997).

(4) Sharma, A.K., Suresh, M.R., Bhojraj, H., Narayanamurthy, H., and Sahu, R.P., "Electroless Nickel Plating on Magnesium Alloy," Metal Finishing (USA), 96, 10 (1998).

(5) Xiang, Y., Hu, W., Liu, X., Zhao, C., and Ding, W., "Initial Deposition Mechanism of Electroless Nickel Plating on Magnesium Alloys," Trans. Inst. Met. Finish, 79, 30 (2001).

(6) ASTM Standard (B 480 - 88), "Standard Guide for Preparation of Magnesium and Magnesium Alloys for Electrodepositing" (1988).

(7) Sakata, Y., "Electroless Nickel Plating Directly on Magnesium Alloy Die Castings," 74th AESF Technical Conference, 15 (1987).

(8) Han, Q., Liu, K., Chen, J. and Wei, X., "Hydrogen Evolution Reaction on Amorphous Ni-S-Co Alloy in Alkaline Medium," Int. J. Hydrogen Energy, 28, 1345 (2003).

(9) Lin, T.R., "Experimental Design and Performance Analysis of TiN-Coated Carbide Tool in Face Milling Stainless Steel," J. Mater. Processing Technology, 127, 1 (2002).

(10) Shaji, S. and Radhakrishnan, V., "Analysis of Process Parameters in Surface Grinding with Graphite as Lubricant Based on the Taguchi Method," J. Mater. Processing Technology, 141, 51 (2003).

(11) Possart, W., Bockenheimer, C., and Valeskem B., "The State of Metal Surfaces After Blasting Treatment: Part I: Technical Aluminium," Surf. Interface Anal. (UK), 33, 687 (2002).

(12) Pourbaix, M., Atlas of Electrochemical Equilibria in Aqueous Solution, NACE, Houston, 1974.

(13) Ambat, R., Aung, N.N., and Zhou, W., "Studies on the Influence of Chloride Ion and pH on the Corrosion and Electrochemical Behavior of AZ91D Magnesium Alloy," J. Appl. Electrochem. 30, 865 (2000).

(14) Cowan, K.G. and Harrison, J.A., "The Automation of Electrode Kintiks III. The Dissolution of Mg in [Cl.sup.-], [F.sup.-], and O[H.sup.-] Containing Aqueous Solutions," Electrochim. Acta., 25, 899 (1980).

Lianxi Yang, ([dagger]) Ben Luan, Woo-Jae Cheong, and Jiaren Jiang -- Integrated Manufacturing Technologies Institute*

* National Research Council Canada, 800 Collip Circle, London, Ont., Canada N6G 4X8.

([dagger]) Author to whom correspondence should be addressed: voice: 519.430.7133; fax: [phone]; email: [e-mail].