Fabricating a Micro-Lens Array Using a Laser-Induced 3D Nanopattern Followed by Wet Etching and CO₂ Laser Polishing

Abstract

Many techniques have been proposed and investigated for microlens array manufacturing in three-dimensional (3D) structures. We present fabricating a microlens array using selective laser etching and a CO₂ laser. The femtosecond laser was employed to produce multiple micro-cracks that comprise the predesigned 3D structure. Subsequently, the wet etching process with a KOH solution was used to produce the primary microlens array structures. To polish the nonoptical surface to the optical surface, we performed reflow postprocessing using a CO₂ laser. We confirmed that the micro lens array can be manufactured in three primary shapes (cone, pyramid and hemisphere). Compared to our previous study, the processing time required for laser processing was reduced from approximately 1 hour to less than 30 seconds using the proposed processing method. Therefore, micro lens arrays can be manufactured using our processing method and can be applied to mass productionon large surface areas.

1. Introduction

As a micro-optical element, micro-lens arrays have been widely used for beam shaping, steering, and improving light extraction efficiency in display and imaging fields, such as digital projectors and three-dimensional (3D) imaging[1-3]. Manufacturing such a micro-lens array has been studied as a representative method, such as direct writing, photolithography, and photoresist reflow. Among the micro-lens array manufacturing methods, direct writing has a high production cost and low productivity compared to other methods; however, it has advantageous strengths in constructing 3D micro-structures with arbitrary geometric structures. Laser direct writing has been used as a rapid prototyping method for manufacturing 3D structures on glass because it is faster than other direct writing methods, such as focusedionbeam milling[4]. However, for structures constructed using laser direct writing, post-process polishing is used to achieve surface roughness of several nanometers. Various post-processing methods, such as chemical etching and heat treatment, have been proposed to overcome the limitations of laser direct writing and construct high-quality 3D micro-structures[4-7]. Recently, our group developed CO2 laser-assisted polishing of fused silica that had undergone femtosecond laser machining [8,9]. The tip of the silica-based optical fiber was machined conically using a femtosecond laser and treated with millisecond CO2 laser pulses to reduce the surface roughness[8]. Upon irradiating the CO2 laser, the optical fiber surface melted and resolidified, resulting in a surface sufficiently smooth to permit a total internal reflection.

Using a similar technique, we present manufacturing a micro-lenticular lens array using femtosecond and CO2 lasers by applying femtosecond laser machining and CO2 laser polishing processes developed in our earlier study[9]. However, our previous technique has some limitations. The previous study indicated methods for manufacturing a pre-designed shape by direct writing using a femtosecond laser and post-processing a rough surface formed by laser processing to secure surface roughness through the reflow method. However, a considerable amount of time is required in the process of producing a large 3D micro-lens array or a master mold using femtosecond laser direct writing. Although it was a small 3D structure (less than 1 ),direct writing of the femtosecond laser took more than a few hours. In the current study, we propose a method to replace direct writing using a selective laser etching (SLE) method to overcome the shortcomings of the previous study. Marcinkeviciu et al. developed the SLE technique in 2001 for manufacturing 3D micro-structures in glass materials[10-12]. This technology uses ultra-short pulse (femtosecond or picosecond) lasers, which is becoming increasingly useful in processing transparent materials, and it is based on the selective modification of materials using a focused laser beam[13]. Compared to direct writing technology, SLE technology has the advantage of producing at a fast scan speed. Since direct writing requires processing based on laser ablation, it is challenging to process at high scan speeds; however, in SLE technology, using laser irradiation, only the etching rate is different in wet etching, a subsequent process rather than ablation processing. Therefore, it is possible to process at a high scan speed.

In this study, micro-lens array fabrication comprises three steps. The first step is to selectively modify the pre-designed shape using a femtosecond laser. When a tightly focused femtosecond laser is irradiated onto the glass to modify it into a pre-designed shape using a 3D design program, microcracks are created in the focal volume region perpendicular to the laser polarization direction.

The second step is to etch the modified area using laser irradiation through wet etching. To etch the laser-modified zone, a KOH solution was used. In this etching process, a difference occurs in the etch rate between the laser-modified and non laser-modified zones, and the pre-designed shape is selectively removed. The microcracks generated in the laser-modified zone allow the etching solution to diffuse deeper into the glass.

In the last step, the reflow post treatment using a CO2 laser is performed to secure the surface roughness of the micro lens array fabricated in the previous two steps. The surface roughness of the sample constructed in the SLE step is poor ; thus, it is unsuitable for use as an optical device. CO2 laser irradiation melts and resolidifies the glass, and the glass is machined into a smooth surface sufficient for use as an optical device

2. Selectvie Laser-Induced Ehtcing Process

The sample used in this study to produce a microlens array was an optically polished, 25 mm × 25 mm (width × length), 1-mm thick fused silica glass (VIOSIL-SQ, Shin-Etsu Chemical Co., Ltd, Chiyoda-ku, Tokyo, Japan). Fused silica glass is widely used as an optical component due to its excellent optical properties, chemical resistance, and heat resistance.

2.1 Femtosecond Lasers Induce Modification (Microcracks) on Fused Silica Glass

Fabricating micro-lens arrays on fused silica was performed in three steps. The first is modifying the material into a pre-designed

shape using a femtosecond laser. The second is etching the laser-modified zone by placing the laser-modified sample into the KOH solution to proceed with wet etching. The last step is a post treatment process using a CO2 laser to create a smooth surface on the micro-lens array manufactured in the previous steps.

Laser modification processing was conducted referring to the results of previous study by this study group on nanopatterning technology by irradiating at regular intervals using the femtosecond laser[14].

Fig. 1 shows the pre-designed shape used for micro-lens array fabrication and a diagram of the modified procedure using the femtosecond laser. The modified process by

Fig. 1 Predesigned shape by three-dimensional computer-aided design for femtosecond laser processing: (a) cone type, (b) pyramid type, and (c) hemispherical type. Diagram of femtosecond laser-induced modification

femtosecond laser irradiation is processed layer-by-layer, similar to the processing method of additive manufacturing. Three pre-designed types (cone type (Fig. 1(a)), pyramid type (Fig. 1(b)), and hemispherical type (Fig. 1(c)) were selected for the design to realize the convex micro-lens array production efficiency. The micro-lens array was designed to have an 80 µm width, 50 µm height, and 90 µm pitch.

The wavelength of the infrared (IR) femtosecond laser source (Satsuma Amplitude) used in the laser modification process was 1030 nm, its pulse repetition rate was 760 kHz, and the pulse width was 400 fs. The laser modification condition was at 165  fluences, and the diameter of the laser beam

that focused on the sample surface was around 2 µm. The femtosecond laser was integrated with a 2-axis(XY) galvanometer mirro scanner (Itelli SCAN10, ScanLabs) and an air-bearing 3-axis (XYZ) servo motion stage with a controller(A3200, Aerotech). The laser beam was focused on the sample using a 20× microscopic objective with a numerical aperture of 0.45 (LCPLANN20× /0.45IR, Olympus) equipped with a collar for spherical aberration correction. The laser scan speed was 200mm/s, and the slice layer thickness (distance between layers during layer-by-layer processing using femtoseconds) was set to 5 µm. When processing using the IR femtosecond laser, the samples were equilibrated using a vacuum chuck and a tilting stage. The modified sample was observed under a confocal microscope (OLS4100,Olympus).

Fig. 2 shows the microlens array predesign (Fig. 2(a)) and a representative image (Fig.

Fig. 2 (a) Representative three-dimensional computer-aided design data for femtosecond laser modification and (b) representative image of observation under a confocal microscope after sample modification using a femtosecond laser (hemispherical type)

2(b)) of an observation result of the sample modified using the femtosecond laser under a confocal microscope (hemisphere type). The confocal microscope image shows modification marks on the laser-irradiated area. On the glass surface, a diagonal pattern (a scan line generated when modifying using a femtosecond laser) at regular intervals was observed. In addition, on the inner space of the glass, the basic shape of the lens (hemisphere) was observed.

2.2 Wet Etching of the Laser- Modified Area

The wet etching process proceeded for the modified sample for chemical etching ability, and 8 mol/L KOH was used as the etching solution. A previous study that has used KOH and glass for wet etching was referenced[15]. The temperature was set to 80°C in an ultrasonic bath and kept for 24 h. Once the process was completed, the processed shape was observed under a scanning electron microscope (SEM) (CX-200TA, COXEM).

Fig. 3 Confocal image observed during wet etching on samples processed using data designed in various shapes, scanning electron microscope image observed post-wet etching. (a), (d) cone type, (b), (e) pyramid type and (c), (f) hemisphere type

Fig. 3 shows the second step of SLE, the four-hour wet etching process under a confocal microscope (four hours; top), and results under an SEM (bottom). In addition, the predesign post-SLE results were observed After four hours of immersion in a KOH solution, the areas modified by laser began to etch (Fig. 3, top). The SEM images (Fig. 3, bottom) observed after 24-h immersion confirm the predesign wet etching process. Rough patterns and surfaces were observed on the surface.

3. Post-Treatment Process Using Co2 Laser

A micro-lens array should have a smooth surface; however, as shown at the bottom of Fig. 3, the surface is extremely rough because of the previous fabricated process. CO2 laser treatment was conducted to obtain a smooth lens surface using plano-convex micro-lens arrays.

Fig. 4 shows a diagram of the CO2 laser system and details of the scanning scheme for the process. The CO2 laser (Diamond G-100i, Coherent) had a wavelength of 9.4µm and a pulse repetition rate of 5 kHz. The diameter of the beam that focused on the sample surface was 200 µm. The sample was processed after being equilibrated using a vacuum chuck and a tilting stage. The processing conditions for the CO2 laser were referenced from previous studies[9]. Subsequently, the CO2 laser was irradiated in zigzags with a scanning pitch of 20 µm using a 2D scanner. (Fig. 4(b)) When the beam’s scanning speed was set to 50 mm/s, the repetition rate of the CO2 laser was 5 kHz, and the duty cycle was 40. The calculated fluence from the beam spot size and irradiated power was 4.46J/cm2 on the sample plane.

Fig. 4 (a) Post-treatment focus position using a CO2 laser and (b) scanning scheme details for CO2 laser processing

The mid-IR wavelength of the CO2 laser has numerous applications in glass processing because of the strong linear absorption in the glass. Damage thresholds for melting and ablation in CO2 lasers are related to critical temperatures, such as softening or melting points of fused silica[16]. Simultaneously, by irradiating the glass surface using the CO2 laser, the surface temperature of the glass rapidly rises above the softening temperature due to the interaction with the CO2 laser, resulting in the melting of the glass surface. Considering the laser beam size and shift pitch, the nonuniformity stress in the fused silica after CO2 laser treatment is minimized, enabling uniform post-treatment processing.

SEM shows the melted surface of the CO2 laser-irradiated fused silica microlens array, and Fig. 5 shows typical surface images of fused silica after CO2 laser irradiation. After CO2 laser posttreatment, the surface becomes smooth, regardless of the predesigned shape, and is processed into a plano-convex microlens array shape. The pores visible on the surface of the microstructure before CO2 laser irradiation disappeared, and the surface roughness was significantly decreased. Treatment conditions were applied differently according to the predesigned shape. Reflow processing conditions (defocusing distance and number of scans) using a CO2 laser did not show a common trend on the entire predesigned shape, resulting in random processing results.

Fig. 5 Scanning electron microscope images before and after post-treatment (a),(d) cone type, (b),(e) pyramid type, and (c), (f) hemisphere type

To induce the optimal polishing conditions, the defocusing distance of the CO2 laser and the number of scans should be adjusted, depending on the predesigned shapes. For cone and pyramid types, treatment was possible under the same conditions due to geometric similarity. We fixed the focal position 1 mm away from the focal plane, and the laser beam was scanned three times over the entire field. For the hemispherical type, the number of scans was fixed once, and the experiment was conducted while changing the focal position 1–4 mm away from the focal plane. After performing the optimal CO2 polishing process for each type, the nonoptical surface was uniformly melted

and resolidified, making it possible to fabricate a microlens array with a smooth surface. Since the hemispherical type is similar to the microlens array to be fabricated, the manufacturing treatment was the easiest, and the quality of the manufactured lens was satisfactory. In other words, the hemispherical shape is the most efficient for processing. The reflow processing condition using the CO2 laser in this study achieved the ideal lens shape when the predesign before the reflow was hemispherical, indicating that it is vital to initially create the optimal shape for producing the lens. However, it was restrictive to produce a well-controlled lens using the pyramid predesign with sharp edges.

Fig. 6 Microscopic image of the light concentration ability of each micro lens array, (a) cone type, (b) pyramid type, (c) hemisphere type

To examine the focusing ability of the fabricated lens, we evaluated the focusing power of the microlens. CCD showed that the fabricated microlens array could focus the transmitted light to a single point (Fig. 6). Fig. 6 shows the results of the light concentration of the CO2 reflow-processed microlens array of the cone, pyramid, and hemispherical types.

The major difference between this study and our previous studies is using laser induced microcrack prestructuring instead of the full laser ablation technique. The prestructured crack uses postprocess chemical etching to form a designed macrostructure. This new technique significantly reduced the laser processing time. The laser scan speed used in laser ablation-based technique of the previous study was 8 mm/sec, whereas the laser scan speed the current study was 200 mm/sec, which is approximately 20 times faster. This increase in the scan speed drastically reduced the time required to manufacture a structure of the designed shape from approximately an hour (in the

previous study) to about 30 s or less (in the present study). Therefore, the spot size of the laser beam in the present and the previous study is the same. In addition, the processing method in this study can drastically reduce the processing time because of the smaller structure of the previous study. The processing method using SLE compared to the direct writing method that uses laser ablation can significantly reduce the laser scanning time, and the manufacturing speed should be highlighted in the future for applications on large surfaces.

Using the proposed production method, only the plano-convex spherical shape confirmed fabricating the microlens array.

In future studies, various microlens arrays, such as plano-concave spherical, planoconvex cylindrical, and plano-concave cylindrical should be fabricated. Various microlens arrays can be used in the beam shaping field through microoptical elements and laser diode array collimation. Additionally, shapes that are challenging to realize as commercialized products should be

fabricated using the method presented by this team of scholars in the fields of microoptical elements and lab on a chip, which is an application field of microfluid dynamics. Therefore, applying our processing technique in various fields beyond microlens array fabrication is expected through optimal design.

4. Conclusion

In this study, we proposed a microlens array fabrication technique using SLE and a CO2 laser to overcome the shortcomings of previous study conducted by this group. The microlens array fabrication using our technology was conducted in three steps. First, several microcracks were created in fused silica glass with a predesigned shape 3D structure using a femtosecond laser. Second, a primary microlens array structure was created through wet etching using a KOH solution on the sample prepared in step one. Finally, a reflow postprocessing process was performed using a CO2 laser to polish the sample from a nonoptical surface to an optical surface. The major difference between this study and our previous studies is using laser-induced microcrack prestructuring instead of the full laser ablation technique. This technique significantly reduced the laser processing time from 1 h to 30 s using the femtosecond laser of the proposed method. The plano-convex type microlens array can be fabricated with three primary shapes

(cone, pyramid, and hemisphere) using the new technique. In the reflow postprocessing process according to the primary shape, the trend could not be confirmed, as random processing results were shown according to the processing parameter. Fabricating the hemispherical shape was the easiest, and the other two shapes have similar fabricating challenges. We expect that the fabricating method proposed in this study will enable mass production of microlens arrays and dramatically reduce the processing time.