1. Introduction

Laser shock forming (LSF) or laser shock peening is a laser processing method that uses laser-induced shock wave to perform high-precision geometrical adjustments or surface treatments of a work pieces. The LSF process uses high-intensity laser pulses to vaporize an opaque sacrificial film coated on the sample surface [1–3]. The ablation of the sacrificial materials exerts a strong recoiling mechanic shockwave into the surface of the materials, which leads to plastic deformation. Compressive stress can be induced on desired surfaces of the work piece to improve the mechanic strength. This is impossible for any other microfabrication technique. Forming of work pieces with extraordinary curvatures have also been achieved [2–4]. This process can effectively reduce defect density near the surface, which improves the surface quality. The LSF has also been widely used to adjust geometrical shapes of work pieces. As a contact-free and maskless process, the LSF can form components without any mechanical tools. Compared with laser thermal forming, laser shock micro-forming is a non-thermal method that only changes materials around the laser focused area, which makes it possible to form work pieces at the room temperature with high precision.

The first LSF works were performed in the late 1960s, while high intensity laser beam was used to form and to enhance aeronautical aluminum alloys [5,6]. Since then, various laser-induced plasma confinement schemes were studied to enhance the laser-induced shock waves to improve the laser processing outcome.

The LSF has been widely used in manufacturing and material processing. Although the LSF process has been extensively explored for large work pieces since 1980s, its applications for structures at millimeter scales or below have been recent endeavors. This has been driven by the flourish of miniaturized mechanic components, microelectromechanic systems (MEMS), and even nanomechanic structures. For example, during the fabrication of MEMS devices, micro-mechanic structures such as free-standing structures or suspension structures are susceptible to the residual stress accumulated within materials. The LSF process could be the effective approach to adjusting their geometrical shapes even after the release of the micromechanical structures. Over the last decade, the LSF processes have been further extended to the micrometer scale to handle smaller work pieces in MEMS devices, micro-optical components, and even nano-mechanic structures [7–9].

However, the adaption of the LSF to process small work pieces at micrometer scales is not a straightforward transition. Although the laser shock micro-forming has been extensively studied over several decades, most of experimental researches have been performed using a laser beam with Gaussian shape. The existing LSF process is a point-by-point laser processing scheme, where a single focused beam is scanned through the targeted working area. When handling specimen with large sizes, this point-by-point laser process could form workpieces with required precision. This is because the laser focal size and laser-induced shockwave profiles is much smaller than the size of the specimen. When the LSF process is used to handle small workpieces in micrometer scales, the laser-induced shockwave profile, even produced by a highly focused Gaussian beam, is comparable to the size of the workpiece itself. The circular shockwave profile might not be able to yield deformation with required profiles and precision.

To address these challenges, this paper presents an adaptive optical technique for laser shock micro-forming process. Using a computer-controlled spatial light modulator (SLM), the on-target laser foci can be flexibly formed in arbitrary shapes to excite laser shockwave with suitable size, shape, and intensity. Shockwaves at multiple locations by multiple laser beams can be simultaneously excited. Using this flexible pulse laser beam forming tool, this paper presents studies shockwaves with noncircular profiles to deform free-standing MEMS structures on micrometer scales. Shockwaves simultaneously induced by multiple laser beams and their effects on the mechanic deformations on MEMS structures were also presented. This paper shows that the adaptive optics pulse laser beam forming tool can significantly improve the manufacturing flexibility, precision, and throughput.

2. Experiments setup

The experimental setup of the laser micro-forming process is schematically presented in Fig. 1. The laser used in this study was a Nd:YAG Q-switched laser (Quanta-Ray GCR-11). The laser operated at a wavelength of 532 nm with a repetition rate of 10 Hz. The pulse duration is between 6 and 7 ns. The Gaussian-shaped beam size is 6.4 mm (FWHM) in diameter at the laser output. The laser was first expanded to 13 mm in diameter using a pair of spherical lenses as the beam expander. The pulse energy is controlled by an attenuator which consisted of a half-wave plate and a polarizing beam splitter. The polarization beam splitter also ensured that the output laser polarization matches the primary axis of the spatial light modulator (SLM) to maximize the modulation efficiency. The SLM (Boulder Nonlinear System, P512-0532) is a liquid-crystal-on-silicon SLM with 512 by 512 pixels operated in reflective mode. The SLM is controlled via a 16-bit DVI controller which gives 65535 resolvable phase levels between 0 and 2π at 532-nm. A 7.3-mm square aperture was used to select the laser beam to cover 90% of the active area of SLM in the center. The SLM generated a hologram with a virtual image behind itself. An adaptive Gerchberg-Saxton algorithm was used to calculate the pattern on the hologram [10]. This virtual image was then relayed and focused onto a sample surface through lenses 1 (focal length of 400 mm) and 2 (focal length of 20 mm). The Rayleigh length in our optical set up is largely determined by lens (L2) with a focus length of 20 mm, as shown in Fig. 1. The Rayleigh length of a Gaussian beam is estimated to be ~17 μm, which is comparable to the sacrificial layer thickness.

 

Fig. 1 Sketch of the proposed adaptive laser system for the LS-µF process. WP: Half-wave plate; PBS: Polarizing beam splitter; A: Aperture; L: Lens; SLM: Spatial light modulator; M: Mirror; BS: Beam splitter; MS: Motion stage; WLS: White light source; CC: CMOS camera.

Download Full Size | PDF

An optical imaging setup was also implemented to observe the LS-μF process on sample surfaces in real time. This imaging system consists of a CMOS camera and an imaging lens. A white light source was used to illuminate sample surfaces using a beam splitter cube. The overall adaptive laser beam forming system projects 30% of the laser power on the target sample surfaces. This is largely limited by the diffraction efficiency of the SLM. The peak laser intensity is limited by the damage threshold of the SLM at 3.5 MW/cm² for the laser pulses [11]. In this experiment, the peak laser intensity was capped below 1 MW/cm² to avoid equipment damage.

Figure 2(a) shows four examples of the laser beam shaping using the SLM. These beam shapes were projected on the target surface and recorded by the CMOS camera. The SLM was firstly calibrated to accurately control the beam shape and size on the target. This was accomplished by a calibration experiment. A pair of bars with the width of 1 pixel was projected onto a flat aluminum plate to generate two ablation marks. By measuring the ablated line features. The correlation between SLM input image in term of number of pixels can be related to physical separation on targets. Figure 2(b) shows the calibration results using two bars separated by 64, 128, 192, 256, 320, and 384 pixels, respectively. The minimum on-target laser focus produced by the SLM shaped laser beam was estimated to be 10 μm × 10 μm, while the largest on-target beam spot is approximately 400 μm x 400 μm. The on-target beam shape can be adjusted into arbitrary shapes. The on-target pulse energy is less than 4 mJ/pulse, which is limited by the damage threshold of the spatial light modulator.

 

Fig. 2 (a) Various laser beam shapes projected on an aluminum surface recorded by the CCD camera, (b) the SLM calibration results between the imaging pixel size and the actual on-target size of the laser projection.

Download Full Size | PDF

3. Results

In this paper, free-standing MEMS structures were fabricated in 8111 aluminum alloy foils with a 20-μm thickness. The free-standing structures were bridges with 750 μm in length and 125 -µm in width. They were fabricated via the laser ablation process. Figure 3(a) shows an SEM micrograph of the MEMS structures. The metal surface was spray coated with KRYLON^® Colormaster^TM paint as a sacrificial layer. The thickness of the sacrificial layer was measured by an optical profolometer (Zygo NewView 8000) to be ~10-µm. Liquid water was used as plasma confinement media. During the experiment, the thickness of the water confined layer was optimized at 3 mm to produce the best forming outcomes [12–15]. Although the additional water layer induces aberration in the focused laser beam, it can be corrected used the imaging system in the laser setup shown in Fig. 1 through calibration experiments. After the LSF process, the remained sacrificial layer on the sample was removed carefully with acetone, and then cleaned with IPA and deionized water. The Zygo NewView 8000 optical profiling system was used to monitor the surface profile before and after the LS-µF process. SEM (JSM6510 Scanning electron microscope) images were also observed for more detailed surface profile results.

 

Fig. 3 (a) Topview of SEM image of the free-standing bridges fabricated by laser ablation, (b) demonstration of different laser beam profiles used in the LS-µF process, (c) schematic of comparative tests between simultaneous and asynchronous LS-µF processes.

Download Full Size | PDF

Two sets of experiments were carried out. They are schematically depicted in Figs. 3(b) and 3(c). The first set of the experiments studied effects of the laser beam profiles on the shockwave generations and resulted deformations. In this experiment, shown in Fig. 3(b), an SLM-shaped laser beam was compared with a Gaussian beam. The second set of the experiments studied laser shock micro-forming using multiple laser beams generated by the SLM. In this experiment, shown in Fig. 3(c), two rectangular shape beams were simultaneously projected on free-standing structures. The micro-forming results were the compared with one shaped laser beam with the same beam energy projected on the same locations sequentially.

3.1 SLM-shaped laser beam shock micro-forming

To study effects of laser beam profiles on the outcomes of the laser-induced deformation on MEMS structures, comparative experiments were carried out. The laser micro-forming experiment was first carried out using an unshaped laser beam while the output from the frequency-double YAG laser was directly focused onto the surface of the free-standing structures as depicted in Fig. 3(b). In this experiment, the pulse laser was focused on the target using a fused silica convex lens with a focal distance of 20 mm. The focal size (FWHM) was estimated to be ~10 μm in diameter. The pulse energy on the target is 0.3 mJ. The laser-induced deformation was then examined using SEM as shown in Fig. 4(a). The shockwave deforms the entire bridge. The largest deformation, which is located at the focal point of the laser, was measured as 50 μm below the original location. The profile of the laser-induced deformation is presented in Fig. 4(b). Although the depth of the laser-induced deformation can be changed by the laser pulse energy as shown in Fig. 4(b), the profile of the laser-induced deformation remained the same. The other important feature of the laser-induced deformation exerted by the Gaussian beam profile is the non-uniformity across the width of the bridge around the focal spot. This is evident in the zoom-in SEM micrograph shown in Fig. 4(c). Although the laser excited shockwave induced deformation with sizes many times larger than that of the laser focal size (10 μm), the Gaussian energy profile still produced a circular deformation around the laser focal spot. An atomic force microscope was used to study the surface morphology across the laser focal spot. It reveals surface morphology similar to those areas receiving no laser irradiation. This suggests that the crate shape shown in Fig. 4(c) was induced by mechanic deformation rather than laser ablation or laser melting. Using the optical profilometer (Zygo NewView 8000), 3D surface image of the deformed region in the free-standing bridge is presented in Fig. 4(d).

 

Fig. 4 (a) Birdview SEM images of laser micro-formed free-standing bridge structures by a Gaussian laser beam, (b) the deformation depth vs on-target laser energy using the Gaussian laser beam, (c) the close-up image around laser-impact region, and (d) 3D surface profile plot of the deformed bridge by the Gaussian beam.

Download Full Size | PDF

Results presented in Fig. 4 reveal the limitation of the laser shock micro-forming using Gaussian beams at micrometer scales. When the feature size is comparable to sizes of the laser-shock induced mechanic deformations, the processing non-uniformity due to the nonuniform laser energy distribution becomes significant. This is evidently shown in Figs. 4(c) and 4(d). To improve the uniformity at micron scales, an SLM-shaped laser beam was used to generate shockwave. Using the SLM laser projection system, a rectangular laser beam profile with a size of 10 µm × 125 µm was projected onto free-standing structures to generate shockwaves. Even if the laser intensity distribution inside the shaped laser beam might not be uniform, the shaped laser beam produced significantly improved deformation uniformity.

A laser-deformed freestanding structure is shown in Fig. 5(a). This feature was produced by a laser pulse with an energy of 0.3 mJ on the target surface, which is the same as what used to produce the results shown in Fig. 4 using the Gaussian beam. The maximum deformation produced by the shaped laser beam was measured by the optical surface profolimeter as 24 μm, this is about half of what was achieved using the focused Gaussian laser beam presented in Figs. 4(a) and 4(b). This is probably due to the fact that the shaped laser beam is much larger (>10 times) on the target surface. The recoil pressure induced by the laser ablation becomes weaker. Figure 5(b) shows the deformation profiles across the length of the freestanding bridge produced by a single shaped laser pulse with 0.1, 0.2, 0.3, 0.4, and 0.7 mJ pulse energies, respectively. Along the length of the free-standing bridge, the deformation profiles produced by the SLM-shaped beam were similar to those produced by the Gaussian beam as shown in Fig. 4(b). However, along the width of the bridge, the shaped laser beam produced much more uniform deformation. This is shown in a zoom-in SEM micrograph in Fig. 5(c). The surface profile measurement results are presented in the inset of Fig. 5(d). Both reveal a uniform deformation across the 125 μm width of the bridge. Figure 5(d) compares the deformation profiles produced by the shaped laser beam and an unshaped Gaussian beam with identical pulse energy of 0.3 mJ on target. The crate depth produced by the Gaussian beam exceeds 12 μm, while the SLM-shaped laser beam yielded far less surface height variation ~2.2 μm. The focal area using SLM-shaped laser beam (10 μm × 125 μm) is substantially larger than that of the Gaussian beam (FWHM ~10 μm). This is probably the reason that causes the difference of the maximum deformation at 0.1 mJ as shock pressure produced by the shaped laser beam barely exceed the yielding strength of the workpiece. When larger laser energy was used, both the Gaussian beam and the shaped laser beam produce laser shock pressure that is substantially larger than the yielding strength, leading to comparable deformation in term of size and depths.

 

Fig. 5 (a) Birdview SEM images of laser micro-formed free-standing bridge structures by a rectangular-shape laser beam, (b) the deformation depth vs on-target laser energy using the shaped laser beam, (c) the close-up image around laser-impact region, (d) surface profile measurements along the width of the free-standing bridge using a Gaussian Beam and a sharped laser beam. The 3D surface profile plot of the deformed bridge by the shaped laser beam is shown as inset.

Download Full Size | PDF

The SLM-shaped laser beam yields more uniformed outcome along the width of the bridge, the forming outcome along the length of the bridge is similar to those formed by the Gaussian-shaped beam. This is supported by the FWHM measurement of the laser-induced deformation shown in Fig. 6(a). Both laser beams yield ~400 μm FWHM deformations along the length of the free-standing bridge structures. This size is largely determined by the material properties and thickness of the free-standing bridges. The pulse energy and pulse shape do not have significant impacts on the forming size. The effects of repeated LSF are presented in Fig. 6(b) using a pulse energy of 0.4 mJ. The first laser pulse always has the larger effects than sequent pulses. This is probably due to the fact that the first laser pulse had deformed the sample out of focus, so the subsequent laser was focused in the water layer (thickness ~3 mm), instead of on the sample surface to induce maximum shockwave. Due to smaller on-target laser fluence, the SLM-shaped laser pulse produce smaller deformation as shown in Fig. 6(b). On the other hand, repeated laser shocks do not have significant effects on the size of the deformation, since the size of the on-target laser spot are much smaller than that of the laser-induced deformation.

 

Fig. 6 (a) Relationship of laser-induced deformation FWHM size vs laser pulse energy and (b) Laser-induced deformation depth vs. pulse numbers using 0.4-mJ pulse energy.

Download Full Size | PDF

3.2 SLM-shaped multi-laser beam shock micro-forming

To control the shape and magnitude of the LS-μF, the conventional laser process scheme uses a point-by-point approach to scan across the entire area of the workpiece. This process is slow. More importantly, the laser-induced deformation at a given location is susceptible to the deformation produced at adjacent locations by previous laser pulses, which might have significant impacts on the laser processing accuracy for MEMS devices. This problem can be effectively addressed using the adaptive optics technique.

An SLM can generate multiple laser beams to deform a workpiece at multiple locations simultaneously. To study the effects of parallel laser processing, the same free-standing metal bridge described in the former sections were used for the parallel LSF experiments. The SLM first generated two 10 μm × 125 μm bar-shaped laser beams which were projected on the free-standing structures. The total pulse energy is 0.4 mJ (0.2 mJ per location) on the target. The on-target spacing of two laser beam was set as 324 μm centered at the middle of the bridge. The surface profile of the deformed structure is shown in Fig. 7. As a comparison, single-shot experiments were also carried out by one shaped laser beam in sequence. In this case, the SLM generated one 10 μm × 125 μm bar-shaped laser beam to shock the free-standing structures at the same location of the bridge of the simultaneous shock experiments but in sequence. The on-target pulse energy was set at 0.2 mJ. The profiles for the cross-section of the bridge from both the simultaneous and sequential laser processing are shown in Fig. 7(b).

 

Fig. 7 (a) Surface profile of the free-standing bridge deformed by two laser bars simultaneously. The spacing between two laser bars is 324 μm and (b) 1D surface profile of free-standing bridge deformed by simultaneous two bars and separate two bars.

Download Full Size | PDF

In the sequential laser processing, the first laser pulse deposited the energy at a location of 213 μm, followed by the second pulse at 537 μm. The sequential process yields a non-uniform profile between the two laser incident points. The second laser pulse induced a deformation apparently influenced by the outcome of the first laser pulse at 213 μm, yielding a deeper deformation and leading to a non-uniform deformation profile. In a sharp contrast, the simultaneous laser process while two shaped laser beams deposited laser energy at the same time at the 213 μm and 537 μm positions yielded a flatter depth profile between two incident locations. The simultaneous laser processing also produced deeper deformation, probably due to large impact force produced collectively. The roughness at the flat bottom between two laser impact locations are the intrinsic surface roughness from the metal sheets.

An interesting perspective also arise from the adaptive laser shock scheme. Previous research work shows that shockwaves generated by multiple laser beams at the same time could interfere with each other if they are sufficiently close, leading to magnitude increases of the pressure waves at intersect locations [16]. The adaptive optics technique offered intriguing opportunities to explore interaction of shockwaves and their potentials to improve outcomes of laser processing.

4. Conclusion

In this work, SLM-assisted adaptive laser shock micro forming is presented as an efficient method with high throughput to enhance surface qualities and to improve geometric shape accuracy for microscale fabrication. In contrast to traditional point-by-point LSF, the SLM was used to precisely shape the laser beam profile to yield highly uniform and more controllable micro deformations. Shockwaves simultaneously excited by multiple laser beams generated by the spatial light modulator and its effects on the micro-forming process were also studied. The results presented in this paper show that the adaptive-optics laser beam forming is an effective and flexible method to generate shockwaves with various shapes and sizes of wavefronts and at multiple locations for laser processing at microscales.