**2. Surface morphology of micrometer columns influenced by motion modes of the micro anode**

At first, the morphology of micrometer columns fabricated in nickel sulfate bath by MAGE was of concern interest. Optical microscopy (OM) and scanning electron microscopy (SEM,

stepping motor under precise control, as shown in Fig. 1. In Fig. 1, the cathode was placed horizontally in the electrolytic cell (F) and connected with the negative pole of the dc power supply (V). The microanode assembly (H), connected with the positive pole of the power supply (V), was vertically fixed on a one-dimensional moving table. A servo microstepping motor (M) was used to drive the table through a micro-CPU (C) via the D/A converter (D) and driver (E). A relay (G) was connected with the anode assembly. Through control with dedicated software, the microanode was moved vertically with a resolution of 20 nm per step. Prior to electroplating, intimate contact between the microanode and the cathode was assured through a measurement of null electrical resistance. The microanode was then lifted from the cathode to a variety of gaps (in the range from 1 to 100 μm) to start the MAGE. In this study, a variety of dc-voltage biases (in the range from 3 to 6 V) were employed to conduct MAGE and their corresponding current was monitored with the

So far we have publish a few papers [2-11] to discuss the electrochemical kinetics with respect to MAGE process. The heterogeneous distribution of very intensive electric field in local sites was determined significantly by experimental parameters such as motion modes of the microanode, applied electric voltage, initial gap between the cathode and microanode, and etc. In the present work, we concentrate ourselves on mass balance of electrochemical active ions those which supplied via mass transport from the bulk solution and to be consumed to turn into metallic micro feature. In terms of various models, we applied the commercial software ANSYS 8.0 to simulate the systems so as to understand the

Fig. 1. Schematic diagram of the microanode-guide electroplating system in which the capital letters denote the following. (A) Current sensor, (B) A/D converter, (C) micro-CPU, (D) D/A converter, (E) driver, (F) cell, (G) relay, (H) anode,(M) micro-stepping motor and

**2. Surface morphology of micrometer columns influenced by motion modes** 

At first, the morphology of micrometer columns fabricated in nickel sulfate bath by MAGE was of concern interest. Optical microscopy (OM) and scanning electron microscopy (SEM,

current sensor (A).

(V) voltage source

**of the micro anode** 

electrochemical mechanism of the MAGE process.

S3500, Hitachi Co.) were employed to observe their surface morphology. Figure 2 depicts the OM of the micro columns fabricated under certain conditions with different motion modes of the micro anode. As the micro anode was driven to ascend continuously at a constant rate of 1.8 μm s-1 to perform MAGE at 5.0 V, a column appearing in dendrite was formed (Fig. 2(a)). In contrast, if the micro anode was driven to ascend intermittently (with an initial gap of 10 μm within each intermittent cycle) at the same voltage (i.e., 5.0 V) until reaching certain heights, a column revealing periodical nodes (Fig. 2(b)) was established. Comparing with both the columns, we found that they showed a similar diameter (roughly 50 μm), the dendrite tended to decrease the diameter with increasing its height, as shown in Fig. 2(a); however, the nodal one, depicted in Fig. 2(b), tended to vary the diameter periodically with the height. If we conducted the intermittent MAGE under a lower bias (i.e., at 3.5 V), we obtain a micrometer nickel column in uniform diameter (50 μm) with smooth morphology (as shown in Fig. 2(c)).

Fig. 2. Optical micrographs (OM) of the nickel micrometer columns fabricated by the MAGE process in a sulfate bath where the initial gap between the electrodes was the same (at 10 μm) but their dc-voltage bias and the motion mode of the microanode changed as follows: (a) bias at 5.0 V and the microanode moved continuously at a rate of 1.8 μm s-1, (b) bias at 5.0 V and the microanode moved intermittently and (c) bias at 3.5 V and the microanode moved intermittently

Figure 3(a) exhibits SEM morphologies of a nickel column consisting of two segments due to change of voltages in the intermittent MAGE. With respect to fabricating the lower segment (i.e., 3(b) in Fig. 3(a)), we conducted MAGE at 3.2 V until reaching a height of 500 μm. Then we switched the voltage to 4.0 V to continue the MAGE process to grow the upper segment

Mass Transfer Within the Location Where Micro Electroplating Takes Place 211

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

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).

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

The initial gap between the microanode and the Cu-substrate was 20 μm

hydrogen gas) from both electrodes.

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

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 switched to 4.0 V to continue the deposition from 500 to 1000 μm. (b) Magnified morphology of the lower portion (formed at a bias of 3.2 V) and (c) Magnified morphology of the upper portion (formed at a bias of 4.0 V) for the micrometer nickel column
