**3. Current measured in the continuous MAGE**

Figure 4 exhibits the variation of electroplating current against time in MAGE process where the microanode was ascended continuously at a constant rate of 2.0 μm s−1. In Fig. 4, we found that continuous MAGE could only possibly be performed in the voltage range from 3.0 to 5.0 V. Otherwise, as the continuous MAGE is conducted at voltages less than 2 V, the current responsible for growing the column is too tiny (in a range from 325 to 25 μA) to be

(i.e., 3(c) in Fig. 3(a)) up to 1000 μm. In higher magnification, we are able to distinguish between Fig. 3(b) and (c). The upper segment is covered by greater (65 μm in diameter) nodular particles but the lower one is covered by finer particles (44 μm in diameter). The

(a)

Fig. 3. (a) SEM morphology of the nickel microcolumn resulted from intermittent MAGE under two different biases in a sulfate bath where the microanode was kept at a separation of 10 μm from the cathode to start electroplating in each step. The microcolumn was deposited at a bias of 3.2 V on the copper surface to a height of 500 μm, then the bias was

morphology of the lower portion (formed at a bias of 3.2 V) and (c) Magnified morphology

Figure 4 exhibits the variation of electroplating current against time in MAGE process where the microanode was ascended continuously at a constant rate of 2.0 μm s−1. In Fig. 4, we found that continuous MAGE could only possibly be performed in the voltage range from 3.0 to 5.0 V. Otherwise, as the continuous MAGE is conducted at voltages less than 2 V, the current responsible for growing the column is too tiny (in a range from 325 to 25 μA) to be

switched to 4.0 V to continue the deposition from 500 to 1000 μm. (b) Magnified

of the upper portion (formed at a bias of 4.0 V) for the micrometer nickel column

(b) (c)

**3. Current measured in the continuous MAGE** 

deviation of particle sizes is higher for the upper (±3.5 μm) than the lower (±0.5 μm).

used in practice. To the contrary, as the continuous MAGE is performed at voltages higher than 6.0 V, the current rose rapidly (in 18 s) to reach a critical value (i.e., 20mA), then the circuit is shut off on the purpose to protect the apparatus. As shown in Fig. 4, the current rises fast and it fluctuates profoundly with the continuous MAGE conducted at higher voltages (e.g., 4 and 5 V) than at lower voltages (e.g., 3 V). Higher current may result from abundant reduction of nickel ions and hydrogen ions; current fluctuation especially happened at higher voltages is ascribed to gas –bubbling (evolution of oxygen and hydrogen gas) from both electrodes.

Fig. 4. A plot of current against the electroplating time for the MAGE process conducted at various biases and the microanode moved continuously at a constant rate (e.g., 2.0 μm s-1). The initial gap between the microanode and the Cu-substrate was 20 μm

Fig. 5. Current against the electroplating time in the process of continuous MAGE at a bias of 3.0 V with the motion rate of the microanode varying in the range from 0.3 to 3.0 μm s-1. The initial gap between the microanode and the Cu-substrate was 20 μm

Mass Transfer Within the Location Where Micro Electroplating Takes Place 213

Fig. 6. Variation of current with the electroplating time for the micrometer nickel column fabricated via intermittent MAGE at 3.2 and 4.2 V during (a) the initial stage (in 20 s) and (b) the final stage to reach a column height of 500 μm. The initial gap is at 10 μm in each

A schematic model is demonstrated in Fig. 7(a) to illustrate the growth of the column fabricated by continuous MAGE. Prior to electrochemical reaction, the microanode was ascended to keep an initial gap of 20 μm from the cathode. As soon as the electrochemical deposition started, the microanode was driven to ascend at a constant rate (V). In response to stages 1, 2, ... and n, as shown in Fig. 7(a1), (a2) and (a3), the micrometer column was growing to various heights (i.e., at h1, h2, ... and hn) with the separation between the microanode and microcolumn at d1, d2, ... and dn, respectively. The dashed region in Fig. 7(a3) was re-plotted in Fig. 7(an) for detailed investigation. Supposedly a column established continuously from the (n-1)th stage to nth. The top surface of the column

**5. Models for column growth in continuous MAGE** 

intermittent cycle

Another plot is given in Fig. 5 to show the current variation against time for MAGE conducted at 3.0 V by controlling the microanode to ascend continuously at a variety of rates from 0.3 to 3.0 μm s−1. In Fig. 5, on the curve responsible for continuous MAGE with ascending rate at 3.0 μm s−1, the current rises abruptly to 2.5 mA, drops to 0.25 mA in 60 s and levels off subsequently. This implies that the nickel column grows very fast in the initial period (less than 60 s) but the growth rate in the subsequent stages decays to very slow. Even the duration of this process lasted for 240 s, the column grew rapidly to a height of 25.1 μm almost within the initial 60 s. In contrast to the case where the microanode ascended continuously at 0.3μm s−1, the current led to a sudden rise in 10 s. The growth rate of the column is much faster than the ascending rate of the microanode. As a result, the column grows so swiftly that facilitates its top to contact the microanode. This short-circuit contact may ruin the apparatus. Therefore, on purpose to protect the apparatus, we designed an automatic switch into the system. Once the current exceeding 20 mA the power of the system is shut off. According to Fig. 5, the ascending rate of the microanode is better controlled in the range from 0.5 to 2.0 μm s−1 to ensure longer duration for column
