**2.4 Approaches to high-accuracy microfabrication**

This section describes important parameters that are necessary for high-accuracy microfabrication.

#### **2.4.1 Resolution**

The effect of Fresnel diffraction on resolution is determined by wavelength and the gap between the mask and resist surface, whereas the effect of secondary electrons on resolution is determined by wavelength (Figure 12). To transfer a narrow line pattern, it is necessary to narrow the gap between the mask and the resist surface.

Fig. 12. Relationship between wavelength and pattern width

#### **2.4.2 Optimum experimental conditions**

As mentioned in Section 2.3, the processing depth is determined by the dosage and development time and increases with increasing development temperature. For example, to

processed depth at a development time of 180 min. Figures 9 and 11 show the experimental values and formula approximations. The processing depth is determined by the dosage and development time, as shown in Figures 9 and 11. Thus, it is possible to determine the

Fig. 11. Relationship between dosage and processed depth at a development time of 180 min

This section describes important parameters that are necessary for high-accuracy

The effect of Fresnel diffraction on resolution is determined by wavelength and the gap between the mask and resist surface, whereas the effect of secondary electrons on resolution is determined by wavelength (Figure 12). To transfer a narrow line pattern, it is necessary to

As mentioned in Section 2.3, the processing depth is determined by the dosage and development time and increases with increasing development temperature. For example, to

processing depth using these approximation formulas.

**2.4 Approaches to high-accuracy microfabrication** 

narrow the gap between the mask and the resist surface.

Fig. 12. Relationship between wavelength and pattern width

**2.4.2 Optimum experimental conditions** 

microfabrication.

**2.4.1 Resolution** 

increase processing depth, even when the dosage is small, the development time or temperature must increase. On the other hand, when the dosage is high, the development time can be short. However, to achieve high-accuracy microfabrication, it is necessary to determine optimum experimental conditions.

When the dosage is too high, bubbles often form inside the resist, especially near the surface. Figure 13 shows a microscopic photograph of damage to the PMMA sheet after exposure. Figure 13B shows exposed microchannel patterns. As seen in the figure, bubbles spread not only to the exposed area but also to the non-exposed area. We confirmed that bubbles influenced the structure after development. Because bubbles often form when the amount of absorbed energy exceeds 80 kJ/cm3, it is necessary to select a dosage that does not exceed this "surface-damage energy amount."

As mentioned in Section 2.3, the molecular mass of PMMA must be below a certain value in order for PMMA to dissolve in developer. The minimum energy required for dissolution is called the "development-limit energy amount." Although the development limit energy amount varies in the literature, it is approximately 1 kJ/cm3 based on previous experiments. Variation in the literature values is due to differences in the molecular mass of the PMMA. As the depth from the surface increases, the absorbed energy decreases exponentially; thus, it is necessary to consider not only the surface-damage energy amount but also the development-limit energy amount to determine the optimum dosage.

Next, we discuss the development temperature. As mentioned in Section 2.3, the etching rate increases with increasing development temperature. Because of the higher development temperature and faster etching rate, the processing depth may become large in a short amount of time. However, when the development temperature is too high, the etching rate increases, but the development-limit energy amount decreases. Thus, it is necessary to select the optimum development temperature (Fujinawa et al., 2006). If the development-limit energy amount is too low, PMMA resist develops not only in the exposed area, but also in the non-exposed areas (through the absorber). Therefore, sloped-sidewall structures are fabricated. On the other hand, when the development temperature is low, the processing depth is low, and a long development time is required. Because PMMA swells in water, a long development time is a disadvantage. For these reasons, the development process was performed at 37°C.

Fig. 13. Microscope photographs of damaged to the PMMA sheet after exposure

#### **2.4.3 Micro-loading effect**

In this section, the influence of the micro-loading effect is described. The micro-loading effect is a phenomenon that leads to different processing depths depending on the line width. If the line width is narrow, circulation of the developer worsens, it is difficult to

Fabrication of 3-D Structures Utilizing Synchrotron Radiation Lithography 327

(PCT) technique and the pixels exposure technique. When the PCT technique is used, it is possible to fabricate a microneedle array structure (Khumpuang et al., 2006). When the pixels exposure technique is used, it is possible to fabricate a 3-D structure with a complex

As mentioned in Section 2.3, a characteristic of SR lithography is that it is possible to fabricate a thick structure, and the processing depth can be controlled by dosage. Thus, it was expected that SR characteristics could be applied to the fabrication technology of 3-D structures with free-form surfaces or sloped sidewalls. Additionally, studies on 3-D processing methods using the LIGA process have been reported previously. "SLIGA" adds a sacrificial layer process to the LIGA process. Although the traditional LIGA process can only fabricate a mechanical component fixed to the substrate, the SLIGA technique enables fabrication involving sensors with moving parts (Ruzzu et al., 1998). For example, fabrication of a high-aspect-ratio microactuator was reported. However, this processing technique produces structures with sidewalls that are vertical to the resist surface. To control the inclination angle of the sidewall, the "skew exposure technique" is used to incline resists for X-rays during exposure (Tsuei et al., 1998; Ehrfeld et al., 1999). However, the structures that can be fabricated are limited, and fabrication of structures with free-form surfaces is not possible. Recently, methods that give energy distributions to the resist surface, such as grayscale mask UV lithography, have been reported. If complicated energy distributions can be given to resist surfaces, it will be possible to fabricate complicated

Fig. 16. PCT technique, black color shows X-ray absorber and corn color shows membrane, energy amount is shown by shading, the energy distribution is deposited in the resist by scanning, and a more complex energy distribution can be given by rotating the mask 90

Figure 16 shows the PCT technique. The energy distribution is deposited in the resist by scanning, as shown in the figure. Three-dimensional structures were fabricated by

surface, such as an asymmetrical structure.

**3.1 Advancement in the LIGA process** 

arbitrary 3-D structures.

**3.2 PCT technique** 

supply new developer to the bottom, and dissolved PMMA tends to remain near the bottom. Figure 14 shows the relationship between line width and processing depth using dosages of 0.03 and 0.05 Ah. Plots show experimental values for line widths of 10, 15, 20 and 25 m, and lines show the processing depth of the domain where the line is wide. Figure 15 shows the percentage of processing depths that reached the domain where the line is wide. As shown in Figures 14 and 15, processing depth decreased with decreasing line width. However, higher percentages of processing depth where the line is wide correspond to longer development time. That is, although developer circulation was reduced, dissolution progressed slowly. When the dosage was low, the percentage was high. Additionally, the processing depth was low, and thus the circulation of the developer was good. However, as shown in Figure 14B and C, when the line width was too narrow, dissolution did not progress slowly. Thus, when narrow line patterns are transferred, it is necessary to consider the influence of the micro-loading effect.

Fig. 14. Relationship between line width and processing depth; (A) is 30 min; (B) is 180 min; and (C) is 360 min

Fig. 15. Percentage of processing depths that reached the domain where the line is wide; (A) is 30 min; (B) is 180 min; and (C) is 360 min
