Chemical etching has been the most general method for microchannel fabrication. This method needs a masking process for the selective removal of material. As this method is based on a chemical reaction, undercuts or overcuts can be generated depending on the density of the chemical material and the duration of etching. Therefore, etching cannot be used for the fabrication of a microchannel with precision.

Powder blasting is a kind of mechanical etching process and can be applied with micro abrasives for machine micro shapes that are below 100 μm in size. In the process, micro abrasives are accelerated by highly compressed air or gases and collide with the workpiece with very high velocity and density. Thus, this process is capable of micromachining through the integration of brittle mode fractures that are based on the propagation of micro cracks. The removal of material can be accomplished by scanning the blasting nozzle along pre-defined paths on the workpiece that is covered with a mask. Since the shape of the machined workpiece can be determined by the mask pattern, very complex and/or micro shapes can be easily obtained [4–9]. Recently, CASIO Co. (Japan) developed a micro fuel processor of glass with microchannels using powder blasting [10]. Fuelcell Power Co. (Korea) has successfully fabricated microchannels on a bipolar plate [11]. It was shown that the microfabrication of channels through powder-blasting could enable small and light-weight bipolar plates.

With the development of high-precision machine technology, mechanical micro-machining has been studied for the fabrication of ultra-micro parts with high precision and high functionality [12]. It was shown that this microfabrication technology could be effectively applied to the miniaturization of fuel cells. For the microfabrication of a small fuel cell, MEMS technology based on the conventional silicon wafer technology have been used. Mechanical micromachining usually use CNC milling machine with ultra-precision position control. Precision endmilling can directly machine microchannels on the material(channel plate) without any preprocessing needed for chemical and mechanical etching processes.

The material of the samples used for the channel-fabrication experiments was stainless steel 316 L. The shape and dimensions of the microchannel plate are shown in Figure 1 and Table 1. As a preliminary process for chemical etching and surface treating by powder-blasting, pattern design and pattern film masking processes were carried out. Through the masking process, channels, which were to be removed by chemical etching or endmilling, were determined. The optimum masking conditions were determined on the basis of the required experimental results.

For the microchannel fabrication experiments, five samples were prepared, as shown in Table 1. Cases 1, 2, 3, and 4 were fabricated through both chemical etching and endmilling, while Case 3-rl and 3-r2 were fabricated by only endmilling. Endmilling was performed through a CNC machining center. The depth of endmilling was setup to be up to 4–5 times the total channel depth to minimize machining-induced deformations due to the cutting forces.

For the surface treatment of channels that are fabricated by chemical etching and endmilling, a microblaster (Sintoblaster, MB1 type) was used. Channels with masking film that was not removed were scanned by the blasting nozzle along the path, “ ”. Before the main experiments, a series of preliminary experiments were performed to obtain the optimum blasting conditions by varying the required process parameters, as shown in Table 2. Based on the preliminary experiments, the most appropriate process parameters were decided. If the blasting pressure was higher than 0.2 MPa, experimental samples were twisted. The abrasive, WA#220, yielded a rougher surface than WA#600, as shown in Figure 2. Therefore, in the following, main experiments, WA#220 was used, and the blasting pressure was fixed at 0.15 MPa.

Figure 3 shows the shape and surface roughness of the etched samples as well as those of the surface-treated samples. The shape of the channel profile was similar to the letter “U”. The experimental results are summarized in the 3^rd and 5^th columns of Table 3. The channel width, which was originally 0.8 mm in the mask pattern, increased by about 0.28 mm after etching. The maximum etching depth was about 0.337 mm.

The surface roughness of the etched channel was in the range of Ra 0.36 μm∼Ra 0.45 μm. For these etched samples, when surface treatment was carried out using powder-blasting, the channel depth, channel width, rib width, and channel straightness did not show remarkable changes after surface treatment. The channel flatness increased a little, which was thought to be due to the twisting deformation of the channel plate due to the blasting pressure. The surface roughness roughly doubled. This means the channel surface became uneven and rough, which would be a great help in increasing the coating efficiency of the catalyst.

Figure 4 shows the shape and surface roughness of an etched channel and those of a surface-treated channel. The experimental results are summarized in the last two columns of Table 3. When channels were machined by endmilling, the dimensional errors of channels and ribs were about 0.02 mm each. The channel width, which was originally 0.8 mm in the mask pattern, increased by about 0.02∼0.0427 mm after endmilling.

The maximum milling depth was about 0.318 mm. The surface roughness of these channels was in the range of Ra 0.06 μm∼Ra 0.08 μm. The channel straightness was remarkably different in comparison with those of the etched samples. These experimental results imply that the endmilling method can remarkably improve dimensional precision, when compared with the chemical-etching method. Following surface-treatment, the channel flatness increased considerably, which is due to the large twisting deformation of the channel plate by the cutting force and the residual stress. The surface roughness was in the range of Ra 0.55∼Ra 0.78 μm. Through the powder-blasting of channels that were formed by endmilling, the surface roughness increased to about 10 times. This meant the channel surface became uneven and rough, which would be a great help in increasing the coating efficiency of the catalyst.

For endmilling and blasting of channels, fabrication time was about 5 minutes and 1 minute each, respectively. On the other hand, it was about 20 minutes for chemical etching. The cost of fabrication using endmilling and blasting is thought to be very low, considering time-consuming and high cost masking preprocess for chemical etching.