3D Microfabrication Using Emulsion Mask Grayscale Photolithography Technique

Recently, the rapid development of technology such as biochips, microfluidic, micro-optical devices and micro-electromechanical-systems (MEMS) demands the capability to create complex design of three-dimensional (3D) microstructures. In order to create 3D microstructures, the traditional photolithography process often requires multiple photomasks to form 3D pattern from several stacked photoresist layers. This fabrication method is extremely time consuming, low throughput, costly and complicated to conduct for high volume manufacturing scale. On the other hand, next generation lithography such as electron beam lithography (EBL), focused ion beam lithography (FIB) and extreme ultraviolet lithography (EUV) are however too costly and the machines require expertise to setup. Therefore, the purpose of this study is to develop a simplified method in producing 3D microstructures using single grayscale emulsion mask technique. By using this grayscale fabrication method, microstructures of thickness as high as 500μm and as low as 20μm are obtained in a single photolithography exposure. Finally, the fabrication of 3D microfluidic channel has been demonstrated by using this grayscale photolithographic technique.


Introduction
Rapid growth of technology has increase the efficiency and quality of our life. Therefore, in order to cope with the pace of technology development, micro and nanofabrication skill has becoming more and more demanding over each year. These days, technologies such as biochips, microfluidic device, micro-optical devices and micro-electromechanical-systems (MEMS) are designed in more complex three-dimensional (3D) micro-profiles. The beneficial of designing 3D microstructure is to maximize the functional efficiency as well as to minimize the space and materials consumptions.
The key of microfabrication process is the patterning process where it determines the feature size of microstructures. Among all patterning techniques, lithography is the most commonly used method especially in the fabrication of integrated circuit (IC) [1]. The word lithography comes from Greek, 'lithos' means stones, and 'graphia' means write [2]. As suggested by the name, lithography literally means writing on stones. In most cases of microfabrication, stones are referred to substrates and the patterns are written on substrates by using a light sensitive polymer called photoresist.
Recently, there are many different forms of lithography technique employed by the industry. The most common form of lithography used is the photolithography technique. In the IC industry, patterns are transferred from photomask onto substrate via photolithography [3]. However, due to the endless demand of critical feature size reduction, traditional photolithography technology becomes insufficient to produce sub-nano structures for the industries. Therefore, a number of alternative technologies called Next Generation Lithography (NGL) are developed. NGL includes technologies such as extreme ultraviolet lithography (EUVL), electron beam lithography (EBL) and focused ion beam lithography (FIB) [4]. These next generation lithography techniques are capable of producing high resolution 3D micro and nanostructures. However, these high end equipments may cost many millions of dollars. Besides the high capital cost, the equipment requires frequent service in order to maintain in good condition [4,5]. Additionally, the main disadvantage of these technologies is the low production speed. That is due to the present of vacuum chamber and also slow pixel by pixel writing speed [6]. Therefore, it is difficult to move to mass production with this low throughput and high capital and maintenance cost equipment.
On the other hand, the traditional photolithography process can only create binary structure per layer. Therefore, it often requires multiple photomasks and lithographic exposure to form 3D pattern from several stacked photoresist layers [7]. This fabrication method is extremely time consuming, low throughput and complicated to conduct for high volume manufacturing scale. Hence, the aim of this study is to develop a simple and low cost method to produce 3D profile microstructure by using single-step grayscale emulsion mask exposure technique.

Method
This section discusses briefly about the process of fabricating 3D microstructures by using emulsion grayscale mask. The basic steps of photolithographic process are schematically illustrated in figure 1.

Software mask design
The fabrication of grayscale photomask starts with software pattern designing process. The desired pattern can be drawn by using any graphic design software such as AutoCAD, Adobe Illustrator, CorelDraw, and so on. In this work, CorelDraw was used due to its user friendly and ease of grayscale colour conversion. First, a clear field mask with 2 % grayscale interval from the range of 50% grayscale concentration to 100 % grayscale concentration was drawn. Due to the 5:1 shrinking rate of the mask fabrication equipment, the designed pattern was drawn 5 times larger than the original dimension. The designed grayscale mask is shown in figure 2. The percentage of grayscale concentrations were used to control the height of developed photoresist obtained after the UV exposure process.

Grayscale emulsion mask fabrication
The software designed grayscale mask was printed on a transparent PET film by using imagesetter technique. The printed transparent film is called a master mask film. Then, a Simple Mask Fabrication Machine MM605 (Nanometric Technology lnc.) was used to project the image from the printed master mask film in a 5:1 reduction scale onto a High Precision Photo Plate (Konica Minolta, Inc.) also known as emulsion mask. The emulsion mask is coated with a light sensitive silver halide coating on a quartz glass base. Therefore, the entire mask exposure process was carried out in a dark room condition due to the high light sensitivity of emulsion mask. Prior to the emulsion mask exposure process, the silver halide coated surface was placed facing towards the light source and the exposure time was adjusted to 8 s.
After the exposure process, the exposed emulsion mask was immersed into an emulsion mask developer at room temperature for 2 minutes. The emulsion mask developer is a mixture of 1 part of high resolution plate developer (CDH-100) from Konica Minolta Opto, Inc. and 4 parts of distilled water. The immersed emulsion mask was stirred continuously to ensure uniform development process. During this development process, the previously exposed silver halide had form a high optical density metallic silver that can be functioned as an excellent optical filter on the emulsion mask. Therefore, a dark field emulsion mask was created during this develop process. Figure 3 shows an image of the fabricated dark field emulsion mask after the development process.

Sample preparation
The substrate used in this work is a 3 x 1 inch microscopic glass slides (DURAN Group). Glass substrate was selected due to its transparent property in order to create 3D relief surface structure by using the back UV exposure method. The glass slides were first cleaned to remove surface contaminants by using ultrasonic bath (GT sonic VGT-1613QTD) with acetone for 5 minutes, followed by ultrasonic bath with methanol and isopropyl alcohol (IPA) each for 5 minutes before being rinsed with distilled water. Immediate after the surface cleaning process, 2 ml of SU-8 2010 photoresist (MicroChem) was dispensed and spread over the entire surface of the 3 x 1 inch glass substrate to produce a 700 µm thick SU-8 film. During this process, the coated SU-8 film will selfplanarized and result in a flat and uniform layer due to the surface tension and high mobility.

Soft baking
The previously SU-8 coated glass substrate was soft baked at the temperature of 95 •C for 10 hours by using a conventional oven. It is important to make sure that all the SU-8 coated samples were placed flatly in the conventional oven as the gravity force may affect the flow of photoresist. The flow of photoresist will cause deformation on the coating surface and leads to non-uniform coating surface.

UV exposure
The photoresist coated glass substrate was then exposed by using One side Mask Aligner LA4100_R1 (Sanei Electric Inc.) with back exposure method (see figure 4). The power density of the i-line (365 nm) mercury lamp is set to 180 W and the SU-8 coated glass substrates were exposed for 30 s.

Post exposure baking
The post exposure bake took place right after the UV exposure process. The exposed samples were first baked at temperature of 65•C for 2 minutes on a hot plate. Then the temperature of hotplate was gradually ramped up to 95 •C for 10 minutes. After 10 minutes of baking, the hotplate was switched off while the samples were still left on top. The samples were allowed to cool down slowly to a room temperature. This slow cooling process reduces stress built up in the cross-linked SU-8 and further avoids cracking and deformation of patterns during the development process.

Develop
The developing process was carried out by immersing the exposed samples into SU-8 developer solution (MicroChem). An ultrasonic cleaner was used to enhance the developing process. With the help of ultrasonic cleaning, the developing process took only 2 minutes to completely dissolve all unexposed SU-8 photoresist from the glass substrate. After the first development, the developed samples were rinsed by using another batch of clean SU-8 developer solution, then followed by isopropyl alcohol (IPA) and distilled water. Finally, the developed samples were blown dry by using stream of nitrogen gas.

Result and discussion
In this work, all samples were analyzed by using a manual coordinate measuring machine (CMM) from Mitutoyo. Table 1 shows the effect of different percentage of grayscale concentration to the result of developed photoresist thickness. For each different grayscale concentration, 30 data points were measured in order to obtain the average thickness of each grayscale concentration.   Based on the result obtained, it is concluded that the thickness of developed photoresist can be manipulated by controlling the percentage of grayscale concentration. The thickness of the developed photoresist varies even a uniform UV exposure dosage is applied during the exposure process. From the result obtained in Table 1, thicker developed photoresist was observed when a high percentage of grayscale value was designed during the software mask designing stage. This is due to the high concentration of grayscale value was inverted during the grayscale emulsion mask fabrication process. Therefore, a low grayscale concentration emulsion mask is fabricated by using the high grayscale concentration software designed master film mask.
Based on table 1, a contrast curve of SU-8 2010 negative photoresist was drawn (see figure 5). By using the power density of 180 W and 30 s of UV exposure, this emulsion mask fabrication method is capable to produce a thick developed photoresist up to a maximum of 550 µm and a thin developed photoresist down to a minimum of 16 µm in just one single exposure. Hence, by using the same fabrication method, a microfluidic channel with 3D micromixer is fabricated. The fabricated result is evaluated under Hitachi TM-1000 Table Top Scanning Electron Microscope (SEM). Figure 6 (a) shows the software design of the micromixer by using grayscale concentration at the range from 60 % to 80 %. While figure 6 (b) shows the result of the developed SU-8 photoresist by using the grayscale emulsion mask fabrication technique.

Conclusion
This work presents an alternative method of fabricating an emulsion grayscale mask that has the capabilities to fabricate 3D structures in micron scale thickness. The results presented in this paper can be used as a guide line to fabricate complicated design of 3D microstructures by adjusting the grayscale concentration during the software mask designing process. Based on the result obtained, the thickness of developed photoresist can go up to a maximum of 550 µm and minimum of 16 µm in just one single exposure. Hence, it is concluded that this emulsion mask grayscale photolithography method is a faster and cost-effective 3D fabrication method compared to traditional photolithography and NGL.