We report on laser micro-welding of materials based on a localized heat accumulation effect using an amplified femtosecond Er-fiber laser with a wavelength of 1558 nm and a repetition rate of 500 kHz. We demonstrated the welding of non-alkali alumino silicate glass substrates, resulting in a joint strength of 9.87 MPa. We also welded a non-alkali glass substrate and a silicon substrate using the 1558-nm laser pulses, resulting in a joint strength of 3.74 MPa. Our laser micro-welding technique can be extended to welding of semiconductor materials and has potential for various applications, such as three-dimensional stacks and assembly of electronic devices and microelectromechanical system devices.
© 2006 Optical Society of America
Micro-joining and micro-welding are important techniques in various industries, including the precision machinery, electrical, and healthcare industries. In fact, several joining techniques, including adhesive bonding, arc bonding, anodic bonding, and soldering, have been used for realizing micro optical, mechanical, electronic and fluidic devices. Laser micro-welding is a superior method due to its advantages of high speed, high precision, consistent weld intensity, and low heat distortion . Because conventional laser micro-welding techniques are based on linear absorption, the materials to be joined should be opaque at the laser wavelength used. When the materials to be joined are transparent at the laser wavelength, a light-absorbing intermediate layer should be introduced at the interface between the transparent materials.
In contrast, we have recently demonstrated a laser micro-welding technique based on the nonlinear absorption of focused femtosecond laser pulses. Due to localized properties of the nonlinear absorption, such as multiphoton and/or tunnel absorption and avalanche ionization, this technique can realize highly space-selective welding without inserting a light-absorbing intermediate layer [2, 3]. By focusing femtosecond laser pulses at the interface between transparent materials, the energy deposited by the laser pulse melts the materials in the neighborhood of the focal point . The melted material, called a liquid pool, created at the interface fills up the original gap between the two materials . By the subsequent resolidification dynamics, the liquid pool is resolidified, thus joining the two materials.
Two different regimes can be distinguished when conducting femtosecond micromachining in bulk glass, such as waveguide writing, depending on whether the pulse period is longer or shorter than the time required for heat to diffuse away from the focal volume: the low repetition-rate regime (kilohertz order), in which material modification is produced by a single pulse, and the high repetition-rate regime (megahertz order), in which cumulative thermal effects occur [5, 6]. In the low repetition-rate regime, the laser pulses are separated by milliseconds, which far exceeds the time required for heat to diffuse out of the focal volume for a typical glass. The focal volume thus returns to room temperature before the next pulse arrives . In the high repetition-rate regime, the laser pulses collectively act as a point source of heat at the focal volume within the bulk material when the time interval between successive pulses is much shorter than the time scale for diffusion of heat out of the focal volume . High repetition-rate micromachining offers several advantages , including: (i) much higher processing speeds; and (ii) the possibility of controlling the structural modifications by changing the writing speed, because of thermal accumulation effects.
A first demonstration of femtosecond laser welding was conducted in the low repetition-rate regime. Tamaki et al. first demonstrated the welding of silica glass substrates using 1-kHz, 85-fs, 800-nm laser pulses . Watanabe et al. then reported on the joining of dissimilar transparent materials, such as borosilicate glass and fused silica, and investigated the joint strength and the transmittance through the joint volume . The possibility of femtosecond laser welding of borosilicate glass substrates in the high repetition-rate regime using 1-MHz, 360-fs, 1045-nm laser pulses has been recently reported [8, 9]. However, there has been no quantitative evaluation of the resulting joint strength in the high repetition-rate regime.
In this paper, we demonstrate the laser micro-welding of materials by a localized heat accumulation effect using an amplified femtosecond Er-fiber laser system with a wavelength of 1558 nm and a repetition rate of 500 kHz. We report on welding of non-alkali glass substrates and demonstrate that the resulting joint strength was 9.87 MPa. We also succeeded in welding a non-alkali glass substrate and a silicon substrate using 1558-nm laser pulses, at which wavelength silicon is transparent. We could obtain a joint strength as large as 3.74 MPa between the non-alkali glass substrate and the silicon substrate.
2. Laser micro-welding technique using femtosecond laser pulses
Figures 1(a), 1(b), and 1(c) show the laser micro-welding procedure using femtosecond laser pulses. Two substrates (Sample 1 and Sample 2) were carefully cleaned, stacked one on another [Fig. 1(a)], and pressed together by three bolts and a lens held in a fixture [Fig. 1(b)] to eliminate an air gap between the two substrates in order to avoid laser ablation at the interface. We performed laser welding in the region where the gap was below approximately λ/4 (λ: wavelength) before the welding by observing white-light fringe pattern (Newton’s rings pattern) in the sample from above (along the z-axis). The normal force (W), the average contact pressure (Pmean), and the maximum contact pressure (Pmax) between the two substrates were estimated using Hertz′s formula :
where a is the diameter of a black spot in the Newton’s rings pattern, r is the radius of curvature of the pressing lens, E1 and E2 are the Young′s moduli of Sample 2 and the pressing lens, and v1 and v2 are Poisson′s ratios of Sample 2 and the pressing lens, respectively. In this experiment, the pressing lens was made of fused silica. Table 1 shows Young′s moduli and Poisson′s ratios for the materials used. In order to weld two substrates, femtosecond laser pulses were focused at the interface between the two substrates [Fig. 1(c)]. The focal region was elongated along the optical axis due to nonlinear propagation, such as filamentation. The filamentary propagation of femtosecond laser pulses bridges the two substrates along the laser propagation axis (z-axis) [11, 12]. This filamentary propagation creates a liquid pool due to localized melting, and the liquid pool at the interface then resolidifies, welding the two substrates. Filamentary propagation is superior for laser welding because the elongated liquid pool means that it is not necessary to translate the focal spot along the z-axis. The characteristic time for thermal diffusion in borosilicate glass is approximately 2 μs  and that of non-alkali glass is almost the same as that of borosilicate glass; therefore, repetition rates above 500 kHz lead to accumulation of thermal energy in the focal region.
To create a welding volume, the filament was first translated by 100 μm along the y-axis, and was subsequently translated in steps of 1 μm along the x-axis. When the length of the filament was fz μm, a welding volume of 100 μm × 100 μm × fz μm was produced. By repeatedly displacing by 20 μm along the x-axis and y-axis, a 3 × 3 array of welding volumes was produced. Figures 2(a) and 2(b) show schematic diagrams of the welding volumes.
To estimate the joint strength, we performed a simple tensile test after welding the substrates. A schematic diagram of the tensile tester is illustrated in Fig. 3. The front face of Sample 1 was joined to a string and the rear face of Sample 2 was joined to a base with an adhesive. The load was increased by adding weights until the welded sample was cleaved into two substrates. When the sample was cleaved, we determined the joint strength by dividing the load by the welding areas.
3. Experimental setup
The optical setup used for structural modification inside the non-alkali glass was almost the same as that used in Refs. [15, 16]. An amplified femtosecond Er-fiber laser system (IMRA America, FCPA μJewel B-250) generated 1558-nm laser pulses. It should be noted that the repetition rate was set at 500 kHz. The pulse duration was 947 fs. The sample was mounted on a two-dimensional translation stage with 100-nm resolution (Physik Instrumente V102.2L). The laser pulses were focused at a depth of 200 μm beneath the surface by a 20× objective lens (Olympus LMPlan 20×IR) with a numerical aperture (NA) of 0.40. The pulse energy was controlled by rotating a half-wave plate in front of a Glan-laser polarizer. The maximum input energy was 0.80 μJ/pulse (400 mW) in front of the objective lens. Images of the structural modification produced in the non-alkali glass were observed in the xy-plane and in the xz-plane by optical transmission microscopes with white-light illumination.
4. Experimental results
4.1 Heat accumulation effect in non-alkali glass
First, we investigated the structural modification in non-alkali alumino silicate glass (HOYA CANDEO OPTRONICS NA35 , size: 7 mm × 20 mm × 0.7 mm). In order to investigate the heat accumulation effect, femtosecond laser pulses were focused with static exposure for various exposure times. In this experiment, we set the input pulse energy at 0.40 μJ/pulse (200 mW). Figure 4(a) shows optical images of the refractive-index change in the xy-plane when increasing the exposure time (the number of pulses) from 1 to 1000 s by factors of 10. Figure 4(b) shows the dependence of the diameter on the exposure time. The diameter of the first dark ring of the refractive-index change expanded from approximately 4 μm to 7 μm from 1 to 1000 s. The size of the refractive-index change in the x-direction was larger than the focal spot size (approximately 2.4 μm). The 1.7-fold increase in radial extent was relatively small considering the 1000-fold increase in exposure time. Furthermore, the size of the refractive-index change increased as the number of laser pulses increased, and the optical images in Fig. 4(a) reveal that a concentric ring was eventually generated [4, 6, 7, 17–19]. We confirmed that the heat accumulation effect was induced at the repetition rate of 500 kHz.
4.2 Laser micro-welding of transparent materials
We investigated the laser micro-welding of transparent materials using a femtosecond fiber laser at 1558 nm based on the heat accumulation effect. These experiments were performed at room temperature in an air atmosphere.
4.2.1 Laser micro-welding of non-alkali glass substrates
First, we demonstrated the laser welding of non-alkali glass substrates. The laser welding was performed in accordance with the procedure described in Section 2.
The normal force (W), the average contact pressure (Pmean), and the maximum contact pressure (Pmax) between non-alkali glass substrates were calculated using Eq. (1) to Eq. (3) and the values in Table 1. The diameter of the black spot in the Newton’s rings pattern was approximately 50 μm. The radius of curvature of the pressing lens was 6.9 mm. By substituting these values into Eq. (1) to Eq. (3), we calculated W, Pmean, and Pmax to be 0.9 N, 115 MPa, and 172 MPa, respectively.
The input energy was 0.8 μJ/pulse (400 mW), and the length of the filament was approximately 25 μm at this energy. As a result, a welding volume of 100 μm × 100 μm × 25 μm was produced. As described in Section 2, a 3 × 3 array of welding volumes was produced. Figures 5(a) and 5(b) show optical images of the welding volumes produced at the scan speed of 20 μm/s in the xy-plane and in the xz-plane, respectively. Figure 5(a) shows that the welding volume was formed by the refractive-index change. Figure 5(b) shows that the welding volume was formed around the interface between the non-alkali glass substrates.
In the regime where heat accumulation occurs, we first welded the non-alkali glass substrates and then evaluated the joint strength. The joint strengths were measured to be 9.87 MPa and 6.81 MPa at the scan speeds of 100 μm/s and 200 μm/s, respectively. The joint strengths were weaker than that in Ref. . The possible reason of weaker joint strength is attributed to a total fluence, which is determined by multiplying the single-pulse fluence by the number of pulses in the focal spot. When pulse energy increases and/or the scan speed decreases, the joint strength will become larger because the joint strength is proportional to the total fluence . In our experiments, the maximum pulse energy at a repetition rate of 500 kHz is 0.8 μJ/pulse, which is less than that used in previous report . Laser systems with higher output energy will increase the joint strength. Another possibility is repulsion that is released when the sample holder is set aside. The released repulsion may decrease the joint strength.
4.2.2 Laser micro-welding of non-alkali glass substrate and silicon substrate
To verify that this technique can be extended to semiconductor materials, we demonstrated the welding of a silicon substrate with a non-alkali glass substrate.
Silicon is transparent in the wavelength range of 1.3 μm to 14 μm. Note that non-alkali glass and silicon are both transparent at 1558 nm. The coefficient of thermal expansion of silicon (2.6 × 10-6/°C)  is the same order of that of non-alkali glass (3.7 × 10-6/°C).
The welding procedure was the same as that used between the non-alkali glass substrates. That is, the non-alkali glass substrate was first mounted on the silicon substrate, and femtosecond laser pulses were then focused at the interface between the substrates. The input energy was 0.8 μJ/pulse (400 mW). Figure 6 shows an optical image of the welding volume in the xz-plane produced at a scan speed of 200 μm/s and a repetition rate of 500 kHz.
The joint strength was 3.74 MPa, which was on the same order as that between non-alkali glass substrates (9.87 MPa). We investigated the morphology of the cleaved surfaces of the two substrates after laser welding. Observation using an optical microscope revealed that the silicon adhered to the non-alkali glass substrate only in the welding volume.
For femtosecond laser welding of glass substrates, previous reports used a 1-kHz Ti:sapphire laser at a wavelength of 800 nm [2, 3]. Compared with the Ti:sapphire laser, the fiber laser we used has some advantages, such as (i) stable output power, (ii) high repetition rate, (iii) excellent beam quality, (iv) high conversion efficiency of optical energy (ratio of excitation light to output light), (v) compactness, and (vi) ease of handling. As a result, the fiber laser will be an important tool for laser welding in industrial applications. The joining of non-alkali glass substrates was performed at the scan speeds of the 100 μm/s and 200 μm/s using laser pulses at 500 kHz. In the repetition-rate regime higher than 500 kHz, we expect that the processing speed will be greatly improved.
Conventional laser joining techniques require that one of the substrates to be joined be transparent at the wavelength of the laser, and that the other substrate be absorbing at that wavelength. For example, because a Nd:YAG laser (wavelength, 1064 nm) has an absorption band only in silicon, silicon-glass joints have conventionally been performed by the following process [20, 21]: (i) the laser beam is transmitted through the transparent material (glass), and (ii) is absorbed at the surface of the opaque material (silicon). Alternatively, the welding of transparent materials can be performed by using a light-absorbing intermediate layer between the substrates. In our experiments, the wavelength of the fiber laser was 1.5 μm. At this wavelength, both the silicon substrate and the non-alkali glass substrate are transparent; the welding of the silicon substrate and the non-alkali glass is thus performed by nonlinear absorption in the materials without introduction of an intermediate layer. When welding the silicon substrate and non-alkali glass substrate, the surface of the silicon substrate is melted by an optically-excited electron-hole plasma . The melting of the silicon substrate at the interface eliminates the gap between the substrates. Although further investigation of the welding mechanism is the subject of future work, this welding technique using 1.5-μm lasers has the potential to join the silicon substrates.
Watanabe et al. reported on the joining of dissimilar transparent materials, such as borosilicate glass and fused silica, whose coefficients of thermal expansion are different, by use of the welding technique with 800-nm femtosecond laser pulses. We believe that our technique can be applied to the welding of dissimilar transparent materials, such as silicon substrate and fused silica, whose coefficients of thermal expansion are different.
Tight focusing using a high-NA objective increases the intensity in the focal volume, thus compensating for the limited output energy of the laser pulses (0.8 μJ/pulse in our experiments). High-NA objectives produce shorter filaments, resulting in reduced welding volume and thus increased intensity. In addition, they have shorter working distances, meaning that they can be used to weld only samples of limited thickness because the laser pulses must propagate through the first sample to the interface. On the other hand, low-NA objectives produce longer filaments, resulting in increased welding volume and thus reduced intensity. However, they have longer working distances and can thus be used to weld thicker samples. Therefore, it is necessary to select an objective of suitable NA to weld the samples. If the first sample is thick, however, spherical aberrations should also be taken into consideration and be compensated for.
We performed laser micro-welding of materials based on a localized heat accumulation effect using an amplified femtosecond Er-fiber laser with a wavelength of 1558 nm and a repetition rate of 500 kHz. We demonstrated the welding of non-alkali glass substrates, resulting in a joint strength of 9.87 MPa. We also succeeded in welding a silicon substrate and a non-alkali glass substrate using the 1558-nm laser light, where silicon is transparent. The joint strength between the silicon substrate and the non-alkali glass substrate was 3.74 MPa. We believe that our work is an important step towards the laser micro-welding of semiconductor materials, which may have various applications, such as three-dimensional stacks, assembly of electronic devices, and fabrication of microsystems and microelectromechanical systems (MEMS). By selecting the laser wavelength, it should be possible to apply this welding technique to a wide range of materials.
The authors gratefully acknowledge H. Nagai and M. Yoshida from AISIN SEIKI CO., LTD. for the use of the femtosecond fiber laser (IMRA America, FCPA (.Jewel B-250). The authors would like to thank J. Nishii from the National Institute of Advanced Industrial Science and Technology and S. Onda and S. Sowa from Osaka University for useful discussions.
References and links
1. A. W. Y. Tan and F. E. H. Tay, “Localized laser assisted eutectic bonding of quartz and silicon by Nd:YAG pulsed-laser,” Sens. Actuators A 120, 550–561 (2005). [CrossRef]
2. T. Tamaki, W. Watanabe, J. Nishii, and K. Itoh, “Welding of transparent materials using femtosecond laser pulses,” Jpn. J. Appl. Phys. Part 2 44, L687–L689 (2005). [CrossRef]
3. W. Watanabe, S. Onda, T. Tamaki, K. Itoh, and J. Nishii, “Space-selective laser joining of dissimilar transparent materials using femtosecond laser pulses,” Appl. Phys. Lett. 89, 021106 (2006). [CrossRef]
4. C. B. Schaffer, J. F. García, and E. Mazur, “Bulk heating of transparent materials using a high-repetition-rate femtosecond laser,” Appl. Phys. A 76, 351–354 (2003). [CrossRef]
5. R. Osellame, S. Taccheo, M. Marangoni, R. Ramponi, P. Laporta, D. Polli, S. D. Silvestri, and G. Cerullo, “Femtosecond writing of active optical waveguides with astigmatically shaped beams,” J. Opt. Soc. Am. B 20, 1559–1567 (2003). [CrossRef]
6. R. R. Gattass, L. R. Cerami, and E. Mazur, “Micromachining of bulk glass with bursts of femtosecond laser pulses at variable repetition rates,” Opt. Express 14, 5279–5284 (2006), http://www.opticsexpress.org/abstract.cfm?id=90286. [CrossRef] [PubMed]
7. C. B. Schaffer, A. Brodeur, J. F. García, and E. Mazur, “Micromachining bulk glass by use of femtosecond laser pulses with nanojoule energy,” Opt. Lett. 26, 93–95 (2001). [CrossRef]
8. I. Miyamoto, A. Horn, D. Wirtmann, J. Gottmann, and F. Yoshino, “High-precision, high-throughput fusion welding of glass using femtosecond laser pulses,” presented at the 4th International Congress on Laser Advanced Materials Processing (LAMP2006), Kyoto, Japan, 16–19 May, 2006.
9. J. Bovatsek, A. Arai, and C. B. Schaffer, “Three-dimensional micromaching inside transparent materials using femtosecond laser pulses: new applications,” presented at CLEO/QELS and PhAST 2006, California, USA, 21–26 May, 2006.
10. S. P. Timoshenko and J. N. Goodier, Theory of elasticity, 3rd ed. (McGraw-Hill, New York1970).
11. K. Yamada, W. Watanabe, T. Toma, K. Itoh, and J. Nishii, “In situ observation of photoinduced refractive-index changes in filaments formed in glasses by femtosecond laser pulses,” Opt. Lett. 26, 19–21 (2001). [CrossRef]
12. S. Onda, W. Watanabe, K. Yamada, K. Itoh, and J. Nishii, “Study of filamentary damage in synthesized silica induced by chirped femtosecond laser pulses,” J. Opt. Soc. Am. B 22, 2437–2443 (2005). [CrossRef]
14. G. W. C. Kaye and T. H. Laby, Tables of physical and chemical constants, 16th ed. (Longman, New York1995).
15. T. Tamaki, W. Watanabe, and K. Itoh, “Fabrication of gratings by use of femtosecond fiber laser in glass,” JLMN-Journal of Laser Micro / Nanoengineering (submitted).
16. T. Tamaki, W. Watanabe, H. Nagai, M. Yoshida, J. Nishii, and K. Itoh, “Structural modification in fused silica by a femtosecond fiber laser at 1558 nm,” Opt. Express 14, 6971–6980 (2006). [CrossRef] [PubMed]
17. S. M. Eaton, F. Yoshino, L. Shah, A. Arai, H. Zhang, S. Ho, and P. R. Herman, “Thermal heating effects in writing optical waveguides with a 0.1 - 5 MHz rate ultrafast fiber laser,” in Microelectronics and Photonics IV, J. Fieret, P. R. Herman, T. Okada, C. B. Arnold, F. G. Bachmann, W. Hoving, K. Washio, Y. Lu, D. B. Geohegan, F. Trager, and J. J. Dubowski, eds., Proc. SPIE5713, 35–42 (2005). [CrossRef]
18. S. M. Eaton, H. Zhang, P. R. Herman, F. Yoshino, L Shah, J. Bovatsek, and A. Y. Arai, “Heat accumulation effects in femtosecond laser-written waveguides with variable repetition rate,” Opt. Express 13, 4708–4716 (2005). [CrossRef] [PubMed]
19. L. Shah, A. Y. Arai, S. M. Eaton, and P. R. Herman, “Waveguide writing in fused silica with a femtosecond fiber laser at 522 nm and 1 MHz repetition rate,” Opt. Express 13, 1999–2006 (2005). [CrossRef] [PubMed]
20. M. J. Wild, A. Gillner, and R. Poprawe, “Locally selective bonding of silicon and glass with laser,” Sens. Actuators A 93, 63–69 (2001). [CrossRef]
21. A. Gillner, J. Holtkamp, C. Hartmann, A. Olowinsky, J. Gedicke, K. Klages, L Bosse, and A. Bayer, “Laser applications in microtechnology,” J. Mater. Process. Technol. 167, 494–498 (2005). [CrossRef]
22. C. V. Shank, R. Yen, and C. Hirlimann, “Time-resolved reflectivity measurements of femtosecond-optical-pulse-induced phase transitions in silicon,” Phys. Rev. Lett. 50, 454–457 (1983). [CrossRef]