Open Access
Add to BibSonomyAbstract
A novel microwelding procedure to join Si-to-glass using ps-laser pulses with high repetition rates is presented. The procedure provides weld joint with mechanical strength as high as 85 MPa and 45 MPa in sample pairs of Si/aluminosilicate (Si/SW-Y) and Si/borosilicate (Si/Borofloat 33), respectively, which are higher than anodic bonding, at high spatial resolution (< 20 µm) and very high throughput without pre- and post-heating. Laser-matter interaction analysis indicates that excellent weld joint of Si/glass is obtained by avoiding violent evaporation of Si substrate using ps-laser pulses. Laser welded Si/glass samples can be singulated along the weld lines by standard blade dicer without defects, demonstrating welding by ps-laser pulses is applicable to wafer-level packaging.
© 2015 Optical Society of America
1. Introduction
Multifunctional microsystems such as MEMS and MOEMS are experiencing rapid growth over the last decades. With this growth, joining of dissimilar materials such as semiconductor-to-glass (semiconductor/glass) is becoming increasingly important, since it encompasses the issue of packaging and encapsulation of devices such as accelerometers and pressure sensors to maintain a controlled environment or prevent contamination and damage.
Anodic bonding has been widely used for Si/glass bonding in different applications [1] since its invention [2]. However, anodic boning has some disadvantages that long time and high temperature heating with the aid of high electric field is needed providing no spatial resolution of bonding, and only alkali-contained glasses are employable. Laser-based microwelding of Si/glass has a possibility of Si/glass joining at high spatial resolution without pre- and post-heating and electric field.
A plenty of papers have been published on welding Si/glass, using ns- [3,4] and fs-laser pulses [5,6]. However, the mechanical strength and the throughput attained in laser-microwelding of Si/glass so far are not high enough to compete with anodic bonding [7,8]. This is because it is not easy to optimize process parameters in laser welding dissimilar materials such as Si and glass having large difference in optical and thermal properties. Although laser-based eutectic bonding [9] and selective laser bonding (SLB) [10] are also developed to lower the bonding temperature, their joint strength and process throughput are not high enough again.
In this paper, the laser-matter interaction in microwelding of Si/glass is studied with the focus on the molten Si behavior, and thereby a novel fusion welding procedure is developed using ps-laser pulses at high pulse repetition rates. Crack-free conditions are analyzed based on thermal stress model [11,12], since prevention of cracks is one of the key issues in the welding brittle materials [13,14]. Our welding procedure provides joint strength higher than anodic bonding at high throughput and high spatial resolution. In the present study, the mechanical strength of the weld joints is determined by a shear test. Welded Si/glass samples are singulated along the weld lines by standard blade dicer to demonstrate that the developed microwelding procedure is applicable to wafer-level packaging.
2. Si/glass microwelding process
2.1 Experimental procedures
For joining Si/glass samples, fiber laser (XLASE, Corelase) is used, which provides pulse duration τp = 20 ps, wavelength λ = 1,060 nm, M2 = 1.5 and maximum pulse repetition rate f = 4 MHz. The focused laser beam is irradiated from the glass side using a lens of NA0.1 to provide a diameter of d ≈10 µm at the interface of Si/glass samples. The samples are translated by a 3-D stage (LD-10, Corelase) with maximum translation speed of v = 2 m/s, and a homemade linear stage with maximum translation speed of v = 5 m/s. Nd:YAG lasers with 532 nm are also used to study the molten Si behavior in welding process.
Si substrate is welded to glass substrates including borosilicate glass (Borofloat 33 and D263, Schott), and aluminosilicate glass (SW-Y, AGC). Borofloat 33 is widely used for MEMS in combination with silicon, since coefficient of thermal expansion (CTE) is close to silicon. SW-Y has CTE much closer to Si wafer in a wide temperature range. Although D263 is not the suitable counterpart of Si for joining due to its large difference in CTE from Si, D263 is used for exploring the microwelding process, because laser-absorption properties are well studied [15–17]. In welding process, Si/glass sample pairs with optical contact [18–20] are used, which prevents cracks by producing embedded molten pool to suppress shrinkage stress in welding brittle material [11,12]. For sample preparation with optical contact, Si and glass substrates are cleaned by the standard method [18]. Optical contact samples are laser-welded without applying external contact pressure unlike the case of existing procedures [3–6]. Mechanical strength of the Si/glass weld joint is evaluated by a shear test.
The appearance of the laser-irradiated samples is observed by an optical microscope. The cross-section of welded samples at selected experimental conditions are observed by scanning electron microscope (SEM) of field-emission type (JSM-5310LV, JOEL), and the element distribution is analyzed by EPMA (ISIS3.1-L300, Oxford).
2.2 Microwelding characteristics of Si/glass by ns-laser pulses
Prior to studying microwelding process using ps-laser pulses, Si/glass welding process by ns-laser pulses is analyzed. Figure 1 shows the appearance of the laser-irradiated Si sample without and with glass substrate. In the sample without glass substrate, minute Si droplets spread randomly around the molten region with a diameter of DSi ≈10 µm. In Si/glass sample with optical contact, the molten Si spreads radially along the interface over a much larger diameter of Doc ≈60 driven by the recoil pressure of evaporation [21], because the molten Si pressurized by Si vapor is confined in the narrow space and directed into radial direction by the substrates. The value Doc/DSi, which is a measure of evaporation recoil pressure, is as large as ≈3, suggesting the recoil pressure of evaporation is very large in welding Si/glass using ns-laser pulses. It is also observed that the Si droplets are splashed radially, indicating some clearance is formed between the substrates.
Fig. 1 Appearance of samples irradiated with ns-laser pulses (τp = 7 ns, λ = 532 nm, Q0 = 7.5 µJ, d = 13 µm). (a) Si without glass substrate. (b) Si/Borofloat 33.
In order to observe the behavior of molten Si in more detail, larger pulse energy is irradiated to Si/glass sample. As the cross-section is shown in Fig. 2, the molten Si is radially flowed out of the laser-irradiated region to leave a vacancy near the laser axis (A-A). Beside the vacancy, clearance as large as ≈1 µm is found, which is caused by the insertion of the pressurized molten Si layer between the substrates. In the surrounding area, the clearance is partly filled by the molten Si. It is noted that cracks are produced in the solidified region (B-B), since thus produced molten region has free surface, reducing the joint strength in ns-laser welding [4].
Fig. 2 Cross-section of laser-irradiated Si/SW-Y near the laser axis (A-A) and surrounding area of the laser spots (B-B). Cracks are produced in the molten Si layer, since molten region contains free surface. (τp = 25 ns, λ = 532 nm, Q0 = 63 µJ, laser spot diameter d ≈34 µm)
It is known that cracks are developed by the shrinkage stress (tensile stress) after solidification in welding of brittle material like Si and glass [13,14], if the molten region contains free surface [11,12]. We suppose that the surface temperature of Si substrate rises during laser pulses so that the laser absorption concentrates to the thin surface layer of Si substrate, when ns-laser pulses are irradiated to Si substrate, because laser absorption coefficient of Si increases with increasing temperature [22,23]. As the result, high recoil pressure of evaporation is produced and lasts for longer laser pulses of ns, and thereby the molten Si is flowed out of the laser-irradiated region to produce molten pool with free surface, which causes cracks by the shrinkage stress. Thus it is concluded that ns-laser pulses are not suitable for welding Si/glass.
2.3 Molten Si behavior at low repetition rate with ps-laser pulses
It is expected that the temperature rise of Si substrate during laser pulses can be suppressed using fs- to ps-laser pulses [24] and the working time of recoil pressure can be shortened. While several papers have been published on microwelding Si/glass using fs-laser pulses [5,6], the attained joint strength is lower than 10 MPa [5]. This is considered to be because the laser beam is tightly focused in the glass substrate using high NA lens and most laser energy is consumed in the glass substrate before reaching the Si/glass interface to be welded.
In this study, 20 ps-laser pulses are used for microwelding Si/glass, and is focused by a lens of NA = 0.1 to provide the laser intensity below the threshold for photoionization of the glass [25] to deposit laser energy preferentially to the Si/glass interface. In this section, the experimental results obtained at rather low pulse repetition rate of f = 0.25 MH are shown. Single laser pulse with energy of Q0 = 4 µJ, which provides nearly the same fluence as ns-laser pulse shown in Fig. 1, is irradiated to the Si/glass sample. A vacancy with diameter ≈5 µm and submicron thickness is also formed in the Si substrate, while little change is found in glass substrate, as shown in Fig. 3. The size of the vacancy is much smaller than that of ns-laser pulse (Fig. 2), indicating the recoil pressure of evaporation produced by ps-laser pulses is significantly smaller than ns-laser pulses, as expected. Micro-cracks are also found in the Si substrate, because free surface is formed due to the existence of the vacancy in the molten region.
Fig. 3 Cross-sections of laser-irradiated optical contact sample at Q0 = 4 µJ, which provides the fluence nearly the same as Fig. 1 (f = 0.25 MHz, v = 5 m/s). Vacancy and micro-cracks are found.
Figure 4 shows the appearance of the Si substrate without and with glass substrate irradiated by 20 ps-pulses at Q0 = 4 µJ at different translation speeds. In the sample without glass substrate, clear-edged melt circles M without spatter of Si droplets are produced unlike the case of ns-laser pulses (Fig. 1). In the sample with glass substrate, the melt circle G is significantly smaller than the case of ns-laser pulses without accompanying melt spatter. In Fig. 4(b), M’ and E correspond to the trace of M and the vacancy (Fig. 3), respectively.
Fig. 4 Laser-irradiated samples at different translation speeds in (a) Si sample without glass substrate and (b) Si/glass (Borofloat 33) sample with optical contact. M, G and E show separated melt circle without glass substrate, molten region with optical contact sample, and the evaporated region shown in Fig. 3 (f = 0.25MHz, Q0 = 4 µJ).
Figure 5 shows the widths of the molten Si region without (DSi) and with (Doc) glass substrate plotted vs. translation speed v. Corresponding number of the laser pulse in the laser spot N ( = fd/v) is also shown in the horizontal axis. It is seen DSi ≈14 µm is unchanged except the low speed region of v < 0.5 m/s. At v = 5 m/s, Doc ≈24 µm, which corresponds to Doc/DSi ≈1.7, is significantly smaller than the case of ns-laser pulses, indicating the removal of the molten Si by the recoil pressure of evaporation in ps-laser pulses is reduced in comparison with the case of ns-laser pulses again. As v decreases, Doc decreases unexpectedly to approach DSi despite the fact that input laser energy per unit welding length increases, suggesting the laser energy tends to concentrate near the laser axis as N increases, as is described below. Obviously this is the opposite tendency to glass/glass welding [20].
Fig. 5 Width of molten region of silicon with (G: Doc) and without (M: DSi) glass substrate plotted vs. translation speed v. N is the number of pulse in the laser beam spot given by fd/v (f = 0.25MHz, Q0 = 4 µJ).
In Fig. 6, which is the magnification of Fig. 4(b), the molten Si flow is schematically illustrated. At v = 5 m/s where each melt region G is produced independently, the point-symmetric melt flow (white arrows) is produced by the recoil pressure of evaporation originated from E. At v = 3 m/s where nearly half of G is overlapped, while the molten Si in the front half G spreads radially in the same manner as v = 5 m/s, the spread of the rear half G’ (yellow arrows) is restricted by the solidified Si layer produced in the previous pulse, resulting in the asymmetric pattern along the translation direction. As N increases, Doc decreases to approach DSi at N ≈10, since the spread of the molten Si is increasingly restricted by the Si layer produced in previous pulses. In contrast, in ns-laser pulses molten Si flows out almost freely, since a large clearance between the substrates is produced by the spreading molten Si having high temperature and large kinetic energy.
Fig. 6 Two-dimensional melt Si flow in the magnified picture of Fig. 4(b). E, M and G represent evaporation region and melt circle without and with glass substrate, respectively. Arrows show the melt flow (a) v = 5 m/s and (b) v = 3 m/s. (Q0 = 4µJ, f = 0.25MHz)
2.4 Molten Si behavior at high repetition rate with ps-laser pulses
Figure 7 shows the cross-section of Si/D263 at higher pulse repetition rate of f = 2 MHz at Q0 = 2 µJ and N = 10. The distributions of Si and O are also shown in the figure. The interface shows the curved structure consisting of a pit near the laser axis and wings on its both sides. It is noted that the defects such as the cracks and the vacancy observed at single pulse irradiation (Fig. 3) are not found any longer. In this process, the convection and diffusion play an important role, as described below.
Fig. 7 Cross-section of welded Si/D263 by 20-ps laser pulses at f = 2 MHz, Q0 = 2 µJ and v = 2 m/s. Element analysis of (a) O and (b) Si, (c) backscattered electron image, and (d) schematic illustration showing how the recoil pressure of evaporation provides curved structure containing a pit and wings.
The pit is produced by the recoil pressure of evaporation depressing the molten Si region, because thickness of the molten layer increases by heat accumulation effect at high pulse repetition rates. Then the convective melt flow along the solid wall is created to provide the vertical flow component across the interface, and the wings are produced in the glass region, as illustrated in Fig. 7(d). As the result, the laser energy absorbed in the Si substrate is transferred to the glass region by the convection. It is expected that the joint strength is increased by the anchor effect of the curved structure.
Si fragments are separated from the Si wings in the glass region, and the contours of the wings and the fragments are increasingly blurred as the laser axis is approached, because the Si diffuses into the glass. As the result, the laser energy is directly absorbed in the glass through linear process, since the bandgap of the glass is narrowed by mixing with Si. The fact that Si wings curve toward the laser axis indicates higher temperature is reached by direct laser absorption as the laser axis is approached.
It should be noted that no vacancy is found near the interface any longer. This is considered to be because the violent evaporation of Si is changed to milder evaporation due to the diffusion and convective flow of glass into Si substrate. Cracks are also prevented by the disappearance of the vacancy, because the molten region has no free surface [11,12].
In Fig. 7(c), narrow clearances less than 200 nm are found between the substrates beside the molten Si region, and are produced by the insertion of the spreading molten Si between the substrates even in ps-laser pulses. Strictly speaking, the molten region does have free surface. However, the clearance is much smaller than the case of ns-laser pulses due to smaller energy contained in the molten Si layer in ps-laser pulses. Thus we consider that the shrinkage stress caused by the plastic deformation of the molten region [11,12] is negligible. It should be also emphasized that crack-free microwelding of the dissimilar materials having large difference in CTE is realized in this study. This is because shrinkage stress caused by difference in CTE can be compensated when embedded molten pool is produced.
Experiment is performed at different pulse energies Q0 and translation speeds v. Figure 8 shows the cross-sections and element distributions of welded Si/glass samples at different values of Q0 at f = 2 MHz and v = 2 m/s. While cracks are found in the figure, we confirmed no cracks are produced after welding, indicating the cracks are produced not in welding but in sample preparation accidentally. The curved structure is produced in Q0 2 µJ, and the depth of the pit and the height of the mixed region increase as Q0 increases, because the recoil pressure of evaporation increases.
Fig. 8 Cross-section (backscattered electron image) and element analysis of O and Si in laser welded Si/D263 by 20-ps laser pulses at different values of Q0 at f = 2 MHz and v = 2 m/s. Note that cracks are produced not by welding but by sample preparation accidentally.
Figure 9 shows the results at different translation speeds v at f = 2 MHz and Q0 = 2 µJ. As the translation speed v decreases, the sizes of the wings and the fragments of Si decrease. This is because the wings and the fragments are diffused out at slower translation speed due to longer diffusion time.
Fig. 9 Cross-section (backscattered electron image) and element analysis of O and Si in laser welded Si/D263 by 20-ps laser pulses at different values of v at f = 2 MHz, and Q0 = 2 µJ.
3. Characterization of weld joint
3.1 Mechanical strength of weld joint
In evaluating the mechanical strength of the Si/glass weld joint using optical contact samples, the contribution of the attracting force between the optical contact faces by van der Waals force [19] has to be taken into account. We determine the shear strength of the weld joint σw with optical contact sample by [18,26,27]
where Sw is the welded area, and Foc is the rupture load of the optical contact sample without weld bead (Foc = σocSoc; Soc = optical contact area and σoc = optical contact force per unit area). In order to minimize the influence of Foc in evaluating σw, it is desirable to reduce Soc. In this study, optical contact samples with a width as small as 1 mm is prepared by HF-etching and masking [27], and five weld beads (width 15 ~18 µm) were made parallel in the optical contact area, as shown in Fig. 10.Fig. 10 (a) Sample used for shear test having the optical contact face with a width of 1 mm prepared by HF-etching and masking. (b) Appearance of laser-weld beads at f = 1 MHz, Q0 = 2 µJ and v = 2 m/s where five parallel beads are made in the optical contact area.
It is known that the value of the weld joint σw estimated by Eq. (1) tends to decrease as the cross-sectional area Sw increases, even if the weld bead has the same metallurgical property. This is because the stress propagates not uniformly in the weld joint but more or less concentrates in some part of the weld joint [13], making the comparison of the strength of the weld joint having different geometry difficult. Thus the mechanical strength σw shown in this paper is considered to be more or less underestimated, because Sw in multi-line welding is five times larger than that of single-line welding. However, the effect of welding parameter on the shear strength can be justly evaluated using the sample shown in Fig. 10(b).
The effect of f and v on shear strength σw is examined at constant pulse energy Q0 = 3 µJ. Figure 11(a) shows the effect of f on σw at translation speed of v = 2 m/s in Si/SW-Y sample pairs. The corresponding values of N ( = fd/v) are also shown in the horizontal axis. Data points are the average values of five samples. The shear strength σw is seen to increase with increasing f, because the molten pool becomes thicker and hence produces more pronounced curved structure due to larger heat accumulation effect. Unexpectedly, however, the value σw at f = 4 MHz is significantly lower than the value at f = 2 MHz, presumably because the laser energy deposited in the unit weld length by average laser power fQ0 is too high at rather slow welding speed of v = 2 m/s. At f = 4 MHz, σw is expected to increase to reach the comparable value to f = 2 MHz at N = 10 by doubling v. Figure 11(b) shows the effect of translation speed v on σw at f = 1 MHz in Si/SW-Y samples, indicating maximum value of σw is reached at v = 1 m/s. This welding speed corresponds to N = 10 again, suggesting that N = 10 is the optimum value at Q0 = 3 µJ, assuming f is large enough to provide thick molten pool by heat accumulation effect. Thus it is expected strong weld joint can be obtained at higher welding speeds in proportion to f at N = 10.
Fig. 11 Shear strength of laser-welded Si/SW-Y sample at Q0 = 3 µJ. (a) Effect of pulse repetition rate at v = 2 m/s. (b) Effect of translation speed v at f = 1 MHz.
Figure 12 shows σw plotted vs. v at a low pulse repletion rate of f = 0.25 MHz. It is seen σw is significantly lower than the value at f = 1 MHz, although it is much higher than the values obtained by laser based joint strength reported before. No increase in σw is obtained with increasing N, suggesting that f = 0.25 MHz is not high enough to produce thick molten pool by heat accumulation effect. These results support our model that the shear strength of the weld joint is increased by the anchor effect of the curved structure, which is produced by the recoil pressure of evaporation, if f is large enough to provide thick molten pool.
Fig. 12 Shear strength of laser-welded Si/SW-Y sample plotted vs. translation speed v at f = 0.25 MHz (Q0 = 3 µJ).
The joint strength is also evaluated in the weld joint of Si/Borofloat 33 at Q0 = 3 µJ at different translation speeds v. As the results are shown in Fig. 13, the shear strength σw = 45 MPa is obtained. This value is somewhat smaller than the highest value in Si/SW-Y, presumably due to the influence of the local residual stress caused by the slight mismatch in CTE between silicon and Borofloat 33. It should be noted, however, that the value σw = 45 MPa still exceeds anodic bonding. σw in Si/Borofloat is independent on N at f = 0.25 MHz and 2 MHz unlike the case of Si/SW-Y. We do not understand the mechanism responsible for this tendency. Further study is needed to find the exact mechanism of Si/glass microwelding and to optimize the process parameters.
Fig. 13 Shear strength of laser-welded Si/Borofloat 33 sample plotted vs. N at f = 0.25 MHz and f = 2 MHz (Q0 = 3 µJ).
The maximum shear strength σw obtained in the present study is as high as σw = 85 MPa, which is significantly higher than anodic bonding [7,8], although we have not optimize the process parameters including Q0, f and v. In addition higher shear strength is expected, taking into consideration that the σw is of underestimation because large value of Sw is used in the shear test.
3.2 Space-selective microwelding
In addition to the excellent mechanical strength, Si/glass microwelding by ps-laser pulses has an advantage of excellent space selectivity in contrast to anodic bonding. The spatial resolution of the weld of 15 ~18 µm is available using the lens of NA0.1, as seen in Figs. 7–9. The advantage of the space selective microwelding is demonstrated by performing pattern welding where a grid structure of a size of 3 mm with a street width of 400 μm is “drawn” by the microwelding lines for joining Si/Pyrex sample, as shown in Fig. 14. Each grid line (width: 180 µm) consists of 20 weld beads. No defects are found in the weld joint by ultrasonic test and the optical microscope. Then the laser-welded samples are singulated by a standard blade dicer along the weld streets. No defects are found in the magnified picture shown in Fig. 14(c), and no coolant passage through the welded line by dicing is observed. The result indicates the weld joint has excellent mechanical strength and hermetic properties for wafer level packaging in MEMS, sensors and so on.
Fig. 14 Laser welded grid pattern in Si/Pyrex at f = 2 MHz and Q0 = 1 µJ. Weld line of width 180 µm consists of 20 weld lines. (a) Appearance of sample singulated by a standard dicer. (b) Ultrasonic examination. (c) Magnified picture of diced sample.
3.4 Throughput of microwelding
In addition to high joint strength and high spatial resolution, microwelding of Si/glass using ps-laser pulses is also characterized by high throughput. Figure 15(a) shows an example of cross-section of two-dimensionally welded Si/Borofloat obtained by scanning focused laser spot with a slight overlapping. This indicates two-dimensional microwelding of desired geometry can be obtained at high process speeds.
Fig. 15 (a) Typical cross-section at N = 20 with line separation of 14 µm (f = 2 MHz, Q0 = 2 µJ, v = 1 m/s). (b) Throughput of two-dimensional welding of Si/glass using ps-laser pulses at N = 5 and N = 10 where A and B correspond to the condition shown in Fig. 11(b).
Assuming the laser pulse energy Q0 is irradiated at the pulse repetition rate f and the laser spot size d, the areal joining rate Φ and average laser power Wab are given by
where D is width of weld bead, f is pulse repetition rate and Δ is pulse-to-pulse distance given (Δ = d/N), d = focused spot size and β = D/d. Figure 15 shows Φ plotted vs. f for N = 5 and N = 10, assuming d = 10 µm and Q0 = 3 µJ. In the figure the average laser power Wab is also plotted, and A and B indicate areal joining rates of N = 10 (A) and N = 5 (B) at f = 1 MHz shown in Fig. 11(b), which correspond to Φ = 15 mm2/s and Φ = 30 mm2/s, respectively. Assuming 6-inch Si wafer is welded two-dimensionally to glass wafer using laser pulses with Q0 = 3 µJ and f = 6 MHz (Wab = 18 W), the joining is completed within approximately 3 min at N = 10 and 1.5 min with N = 5 including pre- and post-heating.4. Summary
A novel microwelding procedure of Si-to-glass using 20 ps-laser pulses at high pulse repetition rates has been developed with spatial resolution of 15 ~18 µm at very high throughput without pre- and post-heating. Microwelding is performed using the laser beam focused by a lens with NA0.1 to provide laser intensity in glass below the threshold of photoionization.
The laser-based microwelding process of Si/glass is studied with the focus on analyzing molten Si behavior, and it is found that weld defects cannot be prevented using ns-laser pulses. This is considered to be because larger recoil pressure of evaporation works on the molten Si pool for longer duration of the laser pulse. In ps-pulse laser welding, defect-free welding is realized at high pulse repetition rates by suppressing the violent evaporation with the aid of diffusion and convective flow between Si and glass. Further study is needed to understand the exact mechanism of Si/glass microwelding.
The maximum shear strength of the weld joint attained in this study is 85 MPa in Si/SW-Y sample, and 45 MPa in Si/Borofloat 33. These values are significantly higher than existing laser-based welding procedures, and exceed anodic bonding. Laser welded Si/glass samples can be singulated along the weld streets by a standard blade dicer without defects, demonstrating microwelding by ps-laser pulses is applicable to wafer-level packaging.
Acknowledgments
This work is the extension of the technology, which was developed by the collaboration between Osaka University and Corelase in 2006. The research work described in this paper is in part supported by Finnish Funding Agency for Technology and Innovation TEKES.
References and links
1. P. R. Younger, “Hermetic glass sealing by electrostatic bonding,” J. Non-Cryst. Solids 38–39, 909–914 (1980). [CrossRef]
2. G. Wallis and D. Pomerantz, “Field assisted glass–metal sealing,” J. Appl. Phys. 40(10), 3946–3949 (1969). [CrossRef]
3. C. Luo and L. Lin, “The application of nanosecond-pulsed laser welding technology in MEMS packaging with a shadow mask,” Sens. Actuators A 97–98, 398–404 (2002). [CrossRef]
4. J. S. Park and A. A. Tseng, “Transmission laser bonding of glass with silicon wafer,” Proc. 2004 JUSFA Japan US Symposium on Flexible Automation, Denver (2004).
5. T. Tamaki, W. Watanabe, and K. Itoh, “Laser micro-welding of transparent materials by a localized heat accumulation effect using a femtosecond fiber laser at 1558 nm,” Opt. Express 14(22), 10460–10468 (2006). [CrossRef] [PubMed]
6. A. Horn, I. Mingareev, A. Werth, M. Kachel, and U. Brenk, “Investigations on ultrafast welding of glass-glass and glass-silicon,” Appl. Phys., A Mater. Sci. Process. 93(1), 171–175 (2008). [CrossRef]
7. A. Cozma and B. Puers, “Characterization of the electrostatic bonding of silicon and Pyrex glass,” J. Micromech. Microeng. 5(2), 98–102 (1995). [CrossRef]
8. J. Wei, H. Xie, M. L. Nai, C. K. Wong, and L. C. Lee, “Low temperature wafer anodic bonding,” J. Micromech. Microeng. 13(2), 217–222 (2003). [CrossRef]
9. 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(2), 550–561 (2005). [CrossRef]
10. M. J. Wild, A. Gillner, and R. Poprawe, “Locally selective bonding of silicon and glass with laser,” Sens. Actuators A 93(1), 63–69 (2001). [CrossRef]
11. I. Miyamoto, K. Cvecek, and M. Schmidt, “Crack-free conditions in welding of glass by ultrashort laser pulse,” Opt. Express 21(12), 14291–14302 (2013). [CrossRef] [PubMed]
12. I. Miyamoto, K. Cvecek, Y. Okamoto, and M. Schmidt, “Internal modification of glass by ultrashort laser pulse and its application to microwelding,” Appl. Phys., A Mater. Sci. Process. 114(1), 187–208 (2014). [CrossRef]
13. M. Watanabe and K. Satoh, “Welding mechanics and its applications,” 270–274 (Tokyo: Asakura, 1965)
14. T. Terasaki, “Welding distortion and residual stress,” J. Jpn. Welding Soc. 78(2), 139–146 (2009). [CrossRef]
15. I. Miyamoto, A. Horn, and J. Gottmann, “Local melting of glass material and its application to direct fusion welding by ps-laser pulses,” J. Laser Micro/Nanoeng. 2(1), 7–14 (2007). [CrossRef]
16. I. Miyamoto, A. Horn, J. Gottmann, D. Wortmann, and F. Yoshino, “Fusion welding of glass using femtosecond laser pulses with high-repetition rates,” J. Laser Micro/Nanoeng. 2(1), 57–63 (2007). [CrossRef]
17. I. Miyamoto, K. Cvecek, and M. Schmidt, “Evaluation of nonlinear absorptivity in internal modification of bulk glass by ultrashort laser pulses,” Opt. Express 19(11), 10714–10727 (2011). [CrossRef] [PubMed]
18. K. Cvecek, I. Miyamoto, J. Strauss, M. Wolf, T. Frick, and M. Schmidt, “Sample preparation method for glass welding by ultrashort laser pulses yields higher seam strength,” Appl. Opt. 50(13), 1941–1944 (2011). [CrossRef] [PubMed]
19. V. Greco, F. Marchesini, and G. Molesini, “Optical contact and van der Waals interactions: the role of the surface topography in determining the bonding strength of thick glass plates,” J. Opt. A, Pure Appl. Opt. 3(1), 85–88 (2001). [CrossRef]
20. I. Miyamoto, “Laser welding of glass,” in Handbook of Laser Welding Technologies, S. Katayama, ed. (Woodhead Publishing, 2013), pp. 301–331.
21. I. Miyamoto and G. A. Knorovsky, “Laser microwelding,” in Microjoining and Nanojoining, Y. Zhou, ed. (Woodhead Publishing, 2008), pp. 345–417.
22. G. E. Jellison Jr and D. H. Lowndes, “Optical absorption coefficient of silicon at 1.152µ at elevated temperatures,” Appl. Phys. Lett. 41(7), 594–596 (1982). [CrossRef]
23. J. Geist, A. Migdall, and H. P. Baltes, “Analytic representation of the silicon absorption coefficient in the indirect transition region,” Appl. Opt. 27(18), 3777–3779 (1988). [CrossRef] [PubMed]
24. B. N. Chichkov, C. Momma, S. Nolte, F. von Alvensleben, and A. Tünnermann, “Femtosecond, picosecond and nanosecond laser ablation of solids,” Appl. Phys., A Mater. Sci. Process. 63(2), 109–115 (1996). [CrossRef]
25. B. C. Stuart, M. D. Feit, S. Herman, A. M. Rubenchik, B. W. Shore, and M. D. Perry, “Nanosecond-to-femtosecond laser-induced breakdown in dielectrics,” Phys. Rev. B Condens. Matter 53(4), 1749–1761 (1996). [CrossRef] [PubMed]
26. I. Miyamoto, K. Cvecek, Y. Okamoto, M. Schmidt, and H. Helvajian, “Characteristics of laser absorption and welding in FOTURAN glass by ultrashort laser pulses,” Opt. Express 19(23), 22961–22973 (2011). [CrossRef] [PubMed]
27. Y. Okamoto, I. Miyamoto, K. Cvecek, A. Okada, K. Takahashi, and M. Schmidt, “Evaluation of molten zone in micro-welding of glass by picosecond pulsed laser,” J. Laser Micro/Nanoeng. 8(1), 65–69 (2013). [CrossRef]
