**9. Local potential measurement**

Figure 12 depicts the scheme of an experimental setup for conducting LECD by a MAGE system. The microanode was driven to move by a step motor through an interface controlled by a computer, and the electroplating current was measured using a galvanometer. In addition, we set up a microelectrode, coupled with the saturated calomel electrode in connection with a potentiostat (Princeton EG&G Model 273 A), to oversee the potential at the location where the LECD proceeded. At least three runs had been carried out, and the standard deviation was presented in the error bar. Before setting out the deposition, the open-circuit potential was recorded (i.e. at -491.0 mV versus the SCE). The microanode was descended to touch the cathode, then drawn back to keep an initial gap at 10 μm to start the intermittent MAGE process as mentioned earlier. Once the power switched on, the potential decreased suddenly within several tenths of a second and level off to different levels depending on the electric voltages applied. This potential drop implies the occurrence of electroplating and the deposition rate could be estimated from the magnitude of current. Variation of the current with time has been discussed so we concentrate on the local potential in this section.

Figure 13 depicts the variation of local potential in the intermittent MAGE process against the voltages. Prior to electrochemical deposition, the open circuit potential (OCP) was stabilized at -491.0 mV. It dropped suddenly to a variety of levels in few tenths of a second once the electroplating setting out. The potential level decreased (in the range from -550.0 to -635.0 mV) with increasing the voltages from 3.2 to 4.6 V. The difference between the OCP

1200, 2000) were used in the wet grinding, and subsequently slurries with fine powders of Al2O3 (1.0 and 0.3 μm in diameter, respectively) were employed to polish the cross-sectional surface to a mirror. The mirror surface was pickled in a 0.1% HF solution for 30 s, rinsed with water and dried ready for the SEM examination. Figure 11 displays the SEM morphologies and their transverse section at the position marked with a line across the micrometer. Ni columns deposited at 3.2V (Fig. 11(a)), 3.4V (Fig. 11(b)), 3.6 V (Fig. 11(c)), 4.4 V (Fig. 11(d)) and 4.6 V (Fig. 11(e)). Obviously, the surface morphology and transverse structure of the columns revealed a big difference depending on the biases. The micrometer columns deposited at 3.2 V depicted a smooth surface and a regular circular transverse (Fig. 11(a)). Checking the micrographs shown in Fig. 11 (from 11(a)–(e)),we found that by increasing the electrical voltages, the columns grew into shapes with higher irregularity and less smoothness on their surface. The columns deposited at higher voltages (e.g., 4.4 V) displayed an uneven circular profile around the transverse with the surface in nodular morphology. The columns deposited at much higher voltages (e.g., 4.6 V) appeared to have a branched coral with irregular transverses shape. We were concerned with the internal compactness of the columns, which could be estimated by examining their transverse using the SEM. The compactness was found to vary to different extents, depending on the electrical biases employed. Full compactness was observed in the transverse of the columns deposited at 3.2 V (Fig. 11(a)). Less compact were the columns, with porosity in the center of their transverse (Fig. 11(c)), deposited at little higher voltages (e.g., 3.6 V). The compactness was much less for the columns (Fig. 11(c)) deposited at much higher voltages (e.g., 4.4 V), because of radial expansion of the porosity from the transverse center resulting from coarsening and combination of the voids. The interior of the transverse was almost empty and remained a coral shell for the columns fabricated at an extremely high voltage (i.e.

4.6V). The deposit looked like a branched coral with a hollow interior.

Figure 12 depicts the scheme of an experimental setup for conducting LECD by a MAGE system. The microanode was driven to move by a step motor through an interface controlled by a computer, and the electroplating current was measured using a galvanometer. In addition, we set up a microelectrode, coupled with the saturated calomel electrode in connection with a potentiostat (Princeton EG&G Model 273 A), to oversee the potential at the location where the LECD proceeded. At least three runs had been carried out, and the standard deviation was presented in the error bar. Before setting out the deposition, the open-circuit potential was recorded (i.e. at -491.0 mV versus the SCE). The microanode was descended to touch the cathode, then drawn back to keep an initial gap at 10 μm to start the intermittent MAGE process as mentioned earlier. Once the power switched on, the potential decreased suddenly within several tenths of a second and level off to different levels depending on the electric voltages applied. This potential drop implies the occurrence of electroplating and the deposition rate could be estimated from the magnitude of current. Variation of the current with time has been discussed so we concentrate on the local

Figure 13 depicts the variation of local potential in the intermittent MAGE process against the voltages. Prior to electrochemical deposition, the open circuit potential (OCP) was stabilized at -491.0 mV. It dropped suddenly to a variety of levels in few tenths of a second once the electroplating setting out. The potential level decreased (in the range from -550.0 to -635.0 mV) with increasing the voltages from 3.2 to 4.6 V. The difference between the OCP

**9. Local potential measurement** 

potential in this section.

Fig. 11. SEM morphologies for the micrometer Ni columns deposited at various voltages and their corresponding transverse section. The columns were deposited at (a) 3.2V,(b) 3.4 V, (c) 3.6 V, (d) 4.4V and (e) 4.6 V with the gap between the electrodes initially set at 10 μm

Mass Transfer Within the Location Where Micro Electroplating Takes Place 223

Figure 14 shows the dependence of the average growth rate (in the right ordinate) on the voltages employed to grow a micrometer Ni column, 1000 ± 10 μm in height. The average growth rate was calculated by dividing the height of the columns (i.e. 1000 ± 10 μm) by the growth duration. The time taken by the step motor should be deducted from the total duration of the process. From Fig. 14, the growth rate at 3.2 V is 0.114 μm s-1 and it increases from 0.114 ± 0.004 to 1.76 ± 0.06 μm s-1 with increasing the voltages from 3.2 to 4.6 V. The standard deviation increases with an increase of voltages. Voltages less than 3.2 V or higher than 4.6 V were ignored because of impractical tiny rate in the former and unsatisfactory appearance for the deposits obtained in the latter. In Fig. 14, the average current responsible for LECD was also measured and plotted (in the left ordinate) against the voltages. It increases from 0.225 ± 0.021 to 1.881 ± 0.046 mA with increasing the voltage from 3.2 to 4.6 V. The current is almost nine fold for the MAGE conducted at 4.6 V in comparison to that at 3.2 V. With respect to error bars in Fig. 14, the standard deviation increases with voltages. The growth rate estimated from the data of average current is consistent with that evaluated

Fig. 14. A plot of the average current and average growth rate for the columns against the

As shown in Fig. 14, the stabilized local potential was in the range from -550 to -635 mV for the intermittent MAGE conducted at voltages ranging from 3.2 to 4.6V (vs. SCE). The

**11. Concentration of nickel ions in the location proceeding LECD** 

**10. Average growth rate for the columns with a height of 1000 μm** 

from column height divided by the electroplating duration.

electrical bias employed in the MAGE

and local potential measured under various voltages was of concern. The difference is much greater (i.e. 144.0 ± 1.0 mV) for the MAGE conducted at 4.6 V than that (59.0 ± 1.0 mV) conducted at 3.2 V.

Fig. 12. Schematic diagram of the experimental setup for LECD conducted with the MAGE system and the local potential at the location near the top of the micrometer column measured by a microelectrode coupled with the SCE connected with a potentiostat

Fig. 13. A plot of open-circuit potential and local potentials against the voltages employed in the MAGE with the initial gap at 10 μm between the electrodes

and local potential measured under various voltages was of concern. The difference is much greater (i.e. 144.0 ± 1.0 mV) for the MAGE conducted at 4.6 V than that (59.0 ± 1.0 mV)

Fig. 12. Schematic diagram of the experimental setup for LECD conducted with the MAGE system and the local potential at the location near the top of the micrometer column measured by a microelectrode coupled with the SCE connected with a potentiostat

Fig. 13. A plot of open-circuit potential and local potentials against the voltages employed in

the MAGE with the initial gap at 10 μm between the electrodes

conducted at 3.2 V.
