1. Introduction

The formation of holes through a silicon substrate is needed in the manufacturing and packaging of many silicon microsystems. Holes through a silicon substrate are, for example, required in the manufacturing of through-silicon vias (TSVs), which provide electrical connections through the substrate. Through-silicon holes are typically created using deep reactive ion etching (DRIE). DRIE relies on photomasks and photolithography to define the outline of the holes, making it inefficient and expensive for the production of small manufacturing volumes and for situations that require only a small number of TSV holes per substrate. In addition, photolithography usually demands a flat substrate surface, and DRIE can only create holes perpendicular to the substrate surface. In contrast to DRIE, laser processing allows short design and manufacturing cycles, can tolerate uneven substrate surfaces, and as demonstrated in this article, can create inclined holes, which are not perpendicular to the substrate surface. However, direct laser drilling of holes in silicon suffers from debris accumulation at the hole sidewalls and deformations of the hole shape [1–7].

Debris inside the holes can be removed either after drilling using wet etching or during drilling by immersing the sample in water. For wet etching, hydrofluoric acid is used to remove the oxidized silicon debris inside the holes [8–12]. However, wet etching adds a processing step, and hydrofluoric acid is a hazardous chemical, which can also damage other structural layers on the substrate. Alternatively to etching, immersing the substrate in water during drilling causes debris to be suspended in water, thus preventing the debris from depositing on the surfaces of the substrate [13,14]. However, maintaining a smooth and thin water layer on the substrate is challenging. Any bubbles, waves, or other thickness variations formed in the water layer will distort the laser beam as it passes through the water.

It is possible to avoid the problems caused by placing a water film on top of the substrate if water is instead allowed to enter the drilling hole from the backside of the substrate, which is in contact with the water. The backside of the substrate is defined as the side that is not facing the oncoming laser beam. One variation of this approach is applicable to transparent substrates (e.g., glass), which allow focusing laser light through the substrate to its back surface, thus utilizing multiphoton absorption available for ultrashort laser pulses. In this case, the laser focus is moved from the water-substrate interface on the backside of the substrate towards the substrate’s frontside, while ablation debris is ejected to the water [15–18]. In contrast to transparent glass substrates, the nonlinear properties of silicon in interaction with ultrashort laser pulses prevent precise, selective damaging inside silicon, even for infrared wavelengths with weak linear absorption in silicon [19–22]. Longer nanosecond laser pulses in a unique focusing geometry have been demonstrated to be able to selectively create damage inside silicon [23]. However, this selective damaging method works only inside the silicon and cannot be extended to the surfaces of the silicon substrate, thus making it impossible for water to get into contact with the damaged silicon during the laser processing step in order to remove debris. A recently reported approach allows utilizing water behind a completely opaque substrate and is based on the water entering the hole from the backside of the substrate after an initial hole through the substrate has been formed [24]. This approach was demonstrated for a ceramic substrate where it allowed reducing the taper angle of a hole during trepanning (i.e., moving the laser focus in a circular pattern) using laser pulses with a length of 12 picoseconds. However, silicon’s response to short laser pulses is similar to other semiconductors and metals [25], unintended damage to the surrounding silicon is reduced if shorter sub-picosecond laser pulses are used [26], and trepanning with the associated increase in hole diameter is unnecessary in silicon where simple percussion drilling allows forming a hole through the substrate.

This article investigates the effects of introducing water through the backside of a substrate into a silicon percussion drilling process using sub-picosecond laser pulses. The presence of water on the backside of the substrate is found to significantly reduce the amount of debris in the drilled holes and to help in creating a regular hole shape by removing deformations. The effects of water become visible after the hole reaches through the substrate, thereby allowing the water to enter the hole. The drilling process is demonstrated to allow the formation of inclined through-silicon holes, which could find applications such as TSVs in radio-frequency devices. In order to investigate the effects and mechanisms of the laser drilling process, silicon substrates were exposed to the laser, both with and without water in contact with the backside of the substrate, using a varying number of laser pulses. The states of the holes were observed by X-ray computed tomography (CT) and by scanning-electron-microscope (SEM) imaging of the cross sections and exit openings of the holes.

2. Methods

The silicon substrates used in the experiments were double-side polished, 300(±10) µm thick, with the crystal orientation of <100>, and boron doped to a resistivity between 1 Ω · cm and 10 Ω · cm. The laser used in the experiments was a diode-pumped solid-state laser with a fundamental wavelength of 1040 nm and a second harmonic wavelength of 520 nm (Spirit 1040-4-SHG, Spectra Physics). All the experiments were performed using the 520 nm wavelength due to the more reliable drilling performance achieved in initial testing with this wavelength in comparison to the 1040 nm wavelength. The reason for the performance difference between the wavelengths was not further investigated. However, it is possible that the difference is linked to the three orders of magnitude smaller linear absorption coefficient in water for the 520 nm wavelength in comparison to the 1040 nm wavelength [27]. For the 520 nm wavelength, the duration of a pulse was 293 fs, measured as the full width at half maximum (FWHM). The repetition rate of the laser was 100 kHz throughout the experiments (except in Appendix A), and the exposure time was used to control the number of pulses impinging on the substrate, with millisecond accuracy. The laser beam was focused onto the sample through a microscope objective (RMS10X, Olympus) with a numerical aperture (NA) of 0.10. The laser power was measured using a thermopile sensor (919P-010-16, Newport), which gave the maximum single-pulse energy of 23 µJ for the setup. This energy was used throughout the experiments, because low pulse energies did not always produce a hole through the substrate, that is, too low energies typically resulted in blind holes. The diameter of the laser spot on the substrate was measured to be between 5 µm and 9 µm defined as the $1/e^2$ Gaussian beam diameter [28].

During drilling, a silicon substrate was mounted on a custom-built holder while the laser focus was kept fixed on the top surface of the substrate (see Fig. 1). The holder was designed to keep water in contact with the backside of the substrate. Simultaneously, the water level was prevented from rising above the top surface of the substrate, which would be detrimental for the focusing of the laser to the top surface of the substrate. Maintaining the water in contact with the backside of the substrate allowed removing ablation debris and air bubbles from the drilled holes. The debris and air bubbles were removed from the drilling area by water circulation through the holder. The inclined holes were manufactured by tilting the substrate while keeping the laser beam orientation unchanged.

 

Fig. 1. Setup for water-assisted laser drilling. The silicon substrate is kept in contact with the holder using vacuum. Water is circulated through the holder using a pump. A suitable water level is maintained in the holder by a holder wall that allows the overflow of excess water.

Download Full Size | PDF

The samples were characterized using CT and SEM imaging of the sample surfaces and cross sections. Unless stated otherwise, the samples were not cleaned or otherwise treated before sample characterization. The CT and SEM images presented in this article are from separate holes and thus do not represent the development of the same hole in different stages of the drilling process. In order to obtain cross-sectional SEM images, the single-crystalline silicon substrates were cleaved along a row of holes aligned to a crystal plane. Cleaving was used instead of dicing in order to minimize the formation of additional debris and to avoid flushing away the debris created in laser drilling. However, the exact position of the cross-sectional cleaving plane inside a hole was not controllable. This lack of control resulted in difficulties in exposing the entire hole in a cross-sectional image, especially for the narrow and bent holes. The CT images produced information that was complementary to the cross-sectional SEM images because they reliably captured the entire volume of a hole, albeit not with the resolution of the SEM images.

3. Results

Both SEM and CT imaging were used to inspect the holes formed in the silicon substrate. A key element in producing high-quality holes using ultrashort laser pulses was having water in contact with the backside of the silicon substrate during laser drilling. The high-quality holes were debris-free and symmetrical, and they were demonstrated in both vertical and inclined orientations (see Fig. 2).

 

Fig. 2. Cross-sectional SEM images showing debris-free and well-defined holes through a silicon substrate drilled with water in contact with the backside of the substrate. (a) A vertical hole drilled using 100 000 laser pulses. (b) An inclined hole drilled with a 30° angle to the surface normal using 1 000 000 laser pulses. The tenfold increase in the number of pulses was required for the inclined holes because they are longer than the vertical holes. This increase did not disproportionately widen the hole.

Download Full Size | PDF

In order to see the effects of water on the drilling process, drilling was conducted at two locations on the same sample, first as a dry process without water and then as a wet process with water on the backside of the substrate. Before the number of laser pulses reached 7 500, none of the holes had penetrated through the substrate, and there were no apparent differences between the dry and wet drilling processes (see Fig. 3). After the number of laser pulses exceeded 7 500, the bottom parts of the wet-drilled holes started to expand. This expansion gradually extended upwards towards the hole entrances. In contrast, the dry-drilled holes stayed largely unchanged even with increasing numbers of laser pulses. This result indicates that the water changes the drilling process after the hole reaches through the entire substrate.

 

Fig. 3. X-ray computed tomography (CT) images of holes drilled in a silicon substrate (a) without water and (b) with water. When the number of applied laser pulses reaches around 10 000 or more, the holes drilled with water behind the substrate develop differently from the ones drilled without water. Each image is from a separate set of holes.

Download Full Size | PDF

Dry-drilled holes suffer from defects such as branching, bending, and debris accumulation. Branching refers to the formation of multiple partially parallel hole channels and is visible, together with hole bending, in the CT images (see Fig. 3). Debris and the signs of branching are visible on the backside of the substrate (see Fig. 4(a)). Debris covers the exit holes, and the removal of the debris reveals poorly defined openings with multiple exit points and areas of damaged silicon. The appearance of the multiple exit points is attributed to the branching of the holes inside the silicon (see Fig. 5). Branching is present even before the holes penetrate through the substrate and does not disappear even when a large number of laser pulses is used in dry drilling (see Fig. 5). In addition to debris accumulation around the hole openings, debris also builds up on the hole sidewalls, especially in the upper part of the holes (see Fig. 5).

 

Fig. 4. SEM images of the backsides of silicon substrates showing the exit openings of dry-drilled and wet-drilled holes. (a) Debris covers the dry-drilled exit holes. The insets show the same holes after removing the debris by mechanical grinding and wet etching for approximately 15 minutes in a 4.9 wt% solution of potassium hydroxide in deionized water. The exit openings of the dry-drilled holes have an irregular shape and show evidence of branching of the holes into separate channels. (b) The wet-drilled holes are free from debris even though the substrate was not cleaned after wet drilling. The shape of the exit opening after 40 000 pulses is not regular, but nonetheless the opening is clearly larger than the opening of the dry-drilled hole with the same amount of laser pulses. Increasing the number of laser pulses to 100 000 increases the opening diameter and brings the opening shape closer to a regular elliptical shape. The images of the wet-drilled holes were created by stacking SEM images with different focuses in order to create composite images with a greater depth of field.

Download Full Size | PDF

 

Fig. 5. SEM images of the cross sections of dry-drilled holes after substrate cleaving. A number above a hole depicts the amount of laser pulses used to drill the hole. Each image is taken from a separate hole. Branching of the holes is visible. Debris builds up on the sidewalls of the holes when the number of laser pulses increases. Even increasing the number of laser pulses to 50 000 does not widen the dry-drilled holes significantly.

Download Full Size | PDF

Wet drilling removes any defects created in the holes during the initial phases of the drilling. Bending and branching of the holes are visible after the drilled holes have penetrated through the substrate, but they disappear when the holes start to laterally expand with further exposure to laser pulses (see Figs. 3(b) and 6). The expansion of the hole diameter removes also the debris from the hole sidewalls (see Fig. 6). The removal of the debris progresses from the bottom of the holes (i.e., from the backside of the substrate) toward the top (see Figs. 3(b) and 6). After debris removal, the sidewalls of the holes are left clean with sub-micrometer-scale roughness (see Fig. 6). The exit openings typically have an irregular shape at first, but the shape progresses toward a regular elliptical shape when the number of laser pulses is increased (see Fig. 4(b)).

 

Fig. 6. SEM images of the cross sections of wet-drilled holes after substrate cleaving. A number above a hole depicts the amount of laser pulses used to drill the hole. Each image is taken from a separate hole. Branching is visible at the bottom of the holes, especially after exposure to 7 500 and 10 000 laser pulses, but increasing the number of laser pulses expands the holes, thus eventually removing branching. Debris built inside the holes is removed, progressing upwards from the bottom of the holes as demonstrated in the inset. The removal of debris leaves a clean sidewall surface with sub-micron structures. Exposure to 2 500 laser pulses leaves small marks at the lower part of the substrate, which indicates that the damaged volume in silicon extends deeply into the substrate at this stage, a conclusion supported by the CT images in Fig. 3.

Download Full Size | PDF

The holes obtained in this work have dimensions suitable for use in TSVs. The initial hole channels at the start of the drilling process can be as narrow as 3 µm (see Fig. 6), but when drilling continues and the holes are exposed to 100 000 laser pulses, the diameters of the holes grow towards 100 µm. For example, the diameter of the hole in Fig. 2(a) varies along the vertical axis from 55 µm to 97 µm. The variation of the hole diameter within a single hole originates from a slightly tapered hole shape where the exit opening is the widest part of the hole. The length of the hole in Fig. 2(a) is 296 µm, and thus the hole features an aspect ratio between 1:5 and 1:3. In addition, inclined holes were produced with a 30° angle to the surface normal. This angle extends the hole length to around 350 µm (see Fig. 2(b)). The extended length of the inclined hole required a tenfold increase in the number of laser pulses to one million. The diameter of the inclined hole in Fig. 2(b) ranges from 88 µm to 153 µm, thus featuring an aspect ratio between 1:4 and 1:2. Using the inclined holes to realize inclined TSVs could improve both the insertion and return loss in radio frequency applications as indicated by the simulation results in Appendix B.

4. Discussion

The results show that laser drilling of a silicon substrate with water in contact with the backside of the substrate allows for the creation of clean and well-defined holes through the substrate. Contrary to wet drilling, dry drilling resulted in poorly defined holes with a significant amount of debris accumulated on the hole sidewalls. Wet drilling does not need wet etching after drilling nor does it suffer from the challenges associated with placing water on top of the substrate during laser drilling in order to improve the hole quality.

The defects in dry drilling result from interactions between the high-aspect-ratio holes and the laser light. Laser drilling of silicon typically starts with a phase of a rapid hole-depth increase before slowing down significantly [25,29]. This description is supported by the results in the present study: Laser-induced damage extended deep inside the silicon after only 2 500 laser pulses whereas further extension with more laser pulses was slow (see Figs. 3 and 6). The slowing down of the drilling process typically coincides with the random development of the deformations of the hole shape [25,29]. These deformations, meaning bending and branching, were also observed in the present study (see Figs. 3, 4(a), and 5). Due to the random nature of the development of the deformations, the initial formation of the holes does not progress steadily. As a result, Figs. 4(a) and 5 show only examples of the appearance of holes drilled with the specified number of laser pulses. The shape of an already formed hole affects the energy distribution of the laser pulses propagating in the hole [30], and the formation of bends has been attributed to asymmetrical reflections from hole sidewalls [25,31]. It could be possible to affect the propagation of the laser pulses in the hole by using a tailored Bessel beam instead of a Gaussian beam, and this approach has been demonstrated for drilling of straight, 100 µm long holes in silicon [32]. The formation of bends in the holes have also been explained by interactions with aerosols, which are ablation products created by previous laser pulses [33]. The formation of the hole branches has been explained with two processes: (1) straightening of a bend in the hole so that drilling along the original hole direction continues [7,25] and (2) formation of a bypass channel around a debris plug blocking the hole [34]. Deformations, especially branching, could be minimized and longer holes formed by using higher fluence, which can be achieved by increasing single-pulse energy [5,25,35,36]. The present study used the largest available single-pulse energy of 23 µJ allowed by the experimental setup. The continuation of dry drilling after initial penetration builds more and more debris on the hole sidewalls (see Fig. 5). Eventually, debris buildup will stop the drilling process entirely because an insufficient amount of laser light passes through the debris to the bottom of the hole to continue the drilling action [6,7,25]. In order to use the holes for applications such as TSVs, the removal of the defects created by dry drilling is necessary, but the continuation of dry drilling cannot accomplish this removal.

Water allows wet drilling to remove debris and to correct shape deformations by widening the hole. Wet drilling consists of two stages. The first stage is a dry process where the initial hole cavity is formed in the substrate without the hole being open on the backside of the substrate (i.e., a blind hole is formed). The second stage initiates when drilling is continued and the hole reaches through the substrate. In the second stage, water enters the hole from the backside of the substrate and starts to enhance the drilling action. Because the water comes in contact with laser-irradiated material first at the bottom of the hole, this is also where the lateral expansion of the hole starts (see Figs. 3(b) and 6). A similar enhanced material removal at a water-solid interface during laser drilling has been reported in a situation where a laser filament simultaneously excites the entire depth of a transparent substrate material [37]. After water has entered a hole, the exit opening on the backside of the substrate has at first an irregular shape (see Fig. 4(b)). The initial irregular shape of the hole can originate from the deformations created before the hole penetrated through the substrate and came in contact with water. However, if laser drilling continues, the shape of the hole approaches a more regular elliptical shape (see Fig. 4(b)).

The enhanced material removal in wet drilling is caused by interactions between water, solid material, and laser light. Water has been reported in the literature to decrease the ablation threshold of silicon and to increase the rate of silicon removal [38,39]. The enhanced material removal in water during laser drilling can be attributed to at least the following three possible effects. Firstly, laser-induced plasma generated on a solid surface in water confinement has a higher pressure in comparison to that generated in air confinement and thus creates a stronger impact on the solid surface when expanding [40,41]. Secondly, the collapse of cavitation bubbles in water can cause an even stronger impact on the solid surface than that originating from the expanding plasma [41]. Thirdly, debris created in water confinement is suspended in the water, thus preventing the debris from attaching to the sidewalls of the hole. Some of the phenomena related to laser drilling, such as the collapsing of the cavitation bubbles, water flow, and heat dissipation, are dynamic and short-lived, taking place between laser pulses. A change in the hole shape and in the dynamics of the hole formation was observed when the time delay between the laser pulses was increased from 10 µs to 10 ms (see Appendix A). The most significant change was that the holes started to expand first from the top parts of the holes. The observed changes indicate that the phenomena taking place between the laser pulses play a role in the wet drilling process.

Laser drilling provides an alternative to DRIE in forming holes for TSV applications. As a serial process, laser drilling is slower in comparison to the parallel DRIE, but laser drilling does not require physical photomasks and the related cleanroom tools and processes. The pulse repetition rate used in the present study was 100 kHz, which means that a hole that required 100 000 laser pulses to drill had a drilling time of one second. This means that approximately 3 600 holes per hour can be manufactured. As an example of a low-TSV-density application, the number of TSVs in a single solar cell is on the scale of tens of thousands [4]. As a result, laser drilling could be an alternative to DRIE for solar cells or other applications with low TSV densities, such as micro-electro-mechanical systems (MEMS) and their packages. However, it would be beneficial to increase the speed of the laser drilling process and thus its economic efficiency. This increase in speed could be achieved using the following four methods. The first method is to increase the repetition rate of the laser, causing the laser to deliver the required number of pulses in a shorter time period. A significant increase in the repetition rate could, however, change the laser drilling process, for example through heat buildup in the sample [42,43]. The second possible method to increase the drilling speed is to parallelize the laser drilling process using a spatial light modulator to split the laser beam into multiple beams [44]. The third method is to use a higher single-pulse energy. This could allow initial penetration through the substrate with a lower number of laser pulses. The higher energy could also lead to fewer deformations of the hole channel, thus reducing the amount of laser drilling in water required to remove these deformations. The fourth method is to use thinner substrates, which could allow shorter drilling times as compared to drilling the holes through the standard 300 µm thick substrates used in this study. Applications suitable for this approach include solar cells due to the limited diffusion length of the charge carriers [4] and cap wafers for MEMS devices due to the reduced package size. In addition to drilling speed, hole diameter is a significant parameter for most applications. A small hole diameter reduces the area the hole occupies on the substrate. Initial laser-drilled holes are small, only a few micrometers in diameter, but the hole diameter has to be expanded to remove debris and damaged material on the hole sidewalls and to form a regularly shaped hole. The resulting expansion of the hole diameter could be decreased if the scale of the deformations of the initial hole could be reduced. As mentioned above, one possible strategy to reduce the deformations is using higher laser fluence [5,25,35,36]. The diameters of the laser-drilled holes in the present study are approximately 100 µm, which is adequate for applications such as TSVs.

Discontinuity between the TSVs and a coplanar waveguide leads to impedance mismatch and degradation of radio-frequency performance [45,46]. Inclined TSVs could decrease this discontinuity and thus improve the radio-frequency performance in comparison to vertical TSVs. To analyze the feasibility of the inclined holes for this purpose, simulations were used to determine the insertion and return loss of TSVs with different inclinations. The simulations showed that increasing a TSV angle to the surface normal improves both the insertion and return loss (see Appendix B). The inclined holes manufactured in this work have a 30° angle to the surface normal. Increasing the angle to 45° would further improve the radio-frequency properties (see Appendix B). For this reason, drilling a hole with a 45° angle was also attempted, but no penetration through the substrate was achieved. A possible reason for the lack of penetration is the increase of the hole length when the angle to the surface normal is increased; already a 30° angle required a tenfold increase in the number of pulses in comparison to vertical holes. Larger angles might be achieved using thinner substrates or by utilizing approaches that can allow longer holes such as increasing laser fluence [5,25,35,36] or lowering the pressure of surrounding gas [33,47]. It is worth noting that metal deposition into the inclined holes could be challenging because there is no line-of-sight for metal particles to land on all the surfaces inside the inclined holes. The problem of getting the metallic conductor into the hole could be solved, for example, by using magnetic assembly to pull metal rods into the holes [48–51].

In addition to inclined holes, laser drilling of holes through silicon opens also other possibilities for new types of devices and process flows in comparison to DRIE. Since DRIE relies on an etch mask produced using photolithography, it is challenging to apply on substrates with high surface topography as is typical for MEMS substrates. In contrast to DRIE, laser drilling is not dependent on the surface elevation or orientation, thus in principle allowing it to form holes through substrates with high surface topographies. In addition, reactive-ion-etching techniques require separate etching processes for the removal of layers consisting of different materials, whereas ultrashort laser pulses can penetrate different material layers. These material layers can be, for example, the buried oxide layer in a silicon-on-insulator (SOI) substrate or insulator and metal layers on a silicon substrate. The behavior of silicon is similar to that of metals under ultrashort-pulse laser drilling [25], and metal drilling suffers from similar shape deformations as the drilling of silicon [52]. These similarities indicate that the laser drilling method used in this work could also be applied to other semiconductor and metal substrates. The applications of the laser-drilled holes are not only limited to electrical feedthroughs, but the holes could, for example, also be used as fluid channels in microfluidic applications.

5. Conclusion

Placing water in contact with the backside of a silicon substrate was found to improve the quality of the holes that are percussion drilled through the silicon substrate using ultrashort laser pulses. After laser light has formed an initial hole through the substrate, the water enters the hole from the backside of the substrate, and the interaction between the water and the laser light expands the hole laterally. This expansion removes debris from the hole sidewalls and the shape deformations of the hole. Both the debris and the deformations occur also in holes drilled without water but are not removed with further dry drilling. Therefore, wet drilling improves the quality of the laser-drilled holes in comparison to dry drilling. The drilling speed achieved in this work was one hole per second, making the laser drilling a viable alternative to DRIE in applications that require low hole densities or in applications that present challenges for DRIE. Possible strategies to further increase the drilling speed were also discussed. The laser-drilled holes demonstrated in this work feature diameters between 50 µm and 100 µm and extend through a 300 µm thick standard silicon substrate, making them suitable for TSV applications. Furthermore, laser drilling of inclined holes was demonstrated, an achievement that would be impractical to produce using DRIE. The inclined holes can, for example, find applications in efficiently guiding radio-frequency signals as indicated by the simulation results. The laser drilling method used in this article could also be applied for drilling high-quality holes through other semiconductor and metal substrates in the future.

A. Appendix A: Effects of Low Repetition Rate

The present study investigated the effects of the laser repetition rate on the drilling process by drilling holes using a 100 Hz repetition rate with water in contact with the backside of the substrate. This repetition rate is significantly lower than the 100 kHz used in the rest of this work. The resulting holes were typically expanded at the upper part of the hole, while the bottom part of the hole was often only a narrow channel (see Fig. 7). Some of the holes showed signs of mechanical rupture of the silicon at the bottom of the hole. These results differ significantly from the results obtained with a 100 kHz repetition rate where the expansion of the hole started from the bottom of the hole and continued upwards (see Figs. 3(b) and 6).

 

Fig. 7. SEM images of the two sides of a single hole after cleaving the silicon substrate. The hole was drilled with water in contact with the backside of the substrate. The repetition rate of laser pulses was 100 Hz, and the number of pulses was 8 000. The hole significantly differs in shape from those drilled using a 100 kHz repetition rate.

Download Full Size | PDF

B. Appendix B: Simulation of Radio-Frequency Performance of Inclined TSVs

The laser drilling method used in this article is capable of producing inclined holes through silicon substrates. The inclined holes are challenging to realize using traditional methods, such as DRIE. A TSV produces a discontinuity in a radio-frequency transmission path, which leads to mode conversion and degradation of the return loss (S11) and insertion loss (S21) [45,46]. In the literature, coaxial TSVs have been shown to achieve a better impedance match by adjusting the diameters of the coaxial conductors together with the dielectric constant and the thickness of the insulating liner [53]. In the present study, simulations are used to determine whether using inclined TSVs instead of conventional vertical TSVs could produce a less abrupt discontinuity between a coplanar-waveguide (CPW) and the TSVs, thus improving the loss performance of the TSVs.

The loss performances of the TSVs were simulated using CST Microwave Studio for different variations of the TSV inclination. The simulation model consists of a 50 Ω top CPW redistribution layer (RDL) (a signal line width of 160 µm and a gap of 26 µm), two sets of CPW TSVs (a via diameter of 35 µm and a substrate thickness of 300 µm), and a 50 Ω bottom CPW RDL (a line width of 160 µm and a gap of 26 µm) with a length of 1.7 mm (see Fig. 8(a)). Inclining the TSVs was found to have a significant positive effect on the TSV loss performance (see Fig. 8(b)). The simulated insertion and return loss is the worst for the conventional straight (i.e., 0°) TSV where S21 is −0.27 dB and S11 is −13.97 dB, at 10 GHz frequency. For an inclined TSV with an angle of 30° to the surface normal, S21 is −0.15 dB and S11 is −18.26 dB, at 10 GHz. An inclined TSV with an angle of 45° to the surface normal achieves the best insertion and return loss performance with S21 being −0.13 dB and S11 being −20.36 dB, at 10 GHz. These results indicate that the inclined TSVs could have an improved radio-frequency performance in comparison to the traditional straight TSVs produced using DRIE.

 

Fig. 8. A simulation of the radio-frequency performance of TSVs with different inclinations. (a) An illustration of the coplanar-waveguide model used in the simulation. (b) The simulation results show an improvement in the return loss S11 and insertion loss S21 when the TSV angle to the surface normal is increased

Download Full Size | PDF

Funding

Stiftelsen för Strategisk Forskning (GMT14-0071); Knut och Alice Wallenbergs Stiftelse (Working on Venus project).

Acknowledgments

Mikael Bergqvist is thanked for manufacturing the sample holder used in laser drilling and Daniel Brevemark for his contributions in collecting the CT imaging data.

Disclosures

The authors declare no conflicts of interest.

References

1. B. Tan, “Deep micro hole drilling in a silicon substrate using multi-bursts of nanosecond UV laser pulses,” J. Micromech. Microeng. 16(1), 109–112 (2006). [CrossRef]  

2. B. Tan, S. Panchatsharam, and K. Venkatakrishnan, “High repetition rate femtosecond laser forming sub-10 µm diameter interconnection vias,” J. Phys. D: Appl. Phys. 42(6), 065102 (2009). [CrossRef]  

3. F. Brandi, N. Burdet, R. Carzino, and A. Diaspro, “Very large spot size effect in nanosecond laser drilling efficiency of silicon,” Opt. Express 18(22), 23488 (2010). [CrossRef]  

4. S. Ahn, D. J. Hwang, H. K. Park, and C. P. Grigoropoulos, “Femtosecond laser drilling of crystalline and multicrystalline silicon for advanced solar cell fabrication,” Appl. Phys. A 108(1), 113–120 (2012). [CrossRef]  

5. S. Döring, J. Szilagyi, S. Richter, F. Zimmermann, M. Richardson, A. Tünnermann, and S. Nolte, “Evolution of hole shape and size during short and ultrashort pulse laser deep drilling,” Opt. Express 20(24), 27147 (2012). [CrossRef]  

6. S. Döring, S. Richter, F. Heisler, T. Ullsperger, A. Tünnermann, and S. Nolte, “Influence of ambient pressure on the hole formation in laser deep drilling,” Appl. Phys. A 112(3), 623–629 (2013). [CrossRef]  

7. S. Döring, T. Ullsperger, F. Heisler, S. Richter, A. Tünnermann, and S. Nolte, “Hole Formation Process in Ultrashort Pulse Laser Percussion Drilling,” Phys. Procedia 41, 431–440 (2013). [CrossRef]  

8. C. W. Tang, H. T. Young, and K. M. Li, “Innovative through-silicon-via formation approach for wafer-level packaging applications,” J. Micromech. Microeng. 22(4), 045019 (2012). [CrossRef]  

9. A. Pan, J. Si, T. Chen, Y. Ma, F. Chen, and X. Hou, “Fabrication of high-aspect-ratio grooves in silicon using femtosecond laser irradiation and oxygen-dependent acid etching,” Opt. Express 21(14), 16657 (2013). [CrossRef]  

10. B. Gao, T. Chen, Y. Chen, and J. Si, “Fabrication of Micro-Hole Arrays in Silicon Using Femtosecond Laser Irradiation and Acid Etching,” in 2015 Int. Conf. Comput. Intell. Commun. Networks, (IEEE, 2015), pp. 1338–1340.

11. Y. Li, T. Chen, A. Pan, C. Li, and L. Tang, “Parallel fabrication of high-aspect-ratio all-silicon grooves using femtosecond laser irradiation and wet etching,” J. Micromech. Microeng. 25(11), 115001 (2015). [CrossRef]  

12. T. Chen, A. Pan, C. Li, J. Si, and X. Hou, “Study on morphology of high-aspect-ratio grooves fabricated by using femtosecond laser irradiation and wet etching,” Appl. Surf. Sci. 325, 145–150 (2015). [CrossRef]  

13. J. Kaakkunen, M. Silvennoinen, K. Paivasaari, and P. Vahimaa, “Water-Assisted Femtosecond Laser Pulse Ablation of High Aspect Ratio Holes,” Phys. Procedia 12, 89–93 (2011). [CrossRef]  

14. K. Choo, Y. Ogawa, G. Kanbargi, V. Otra, L. Raff, and R. Komanduri, “Micromachining of silicon by short-pulse laser ablation in air and under water,” Mater. Sci. Eng., A 372(1-2), 145–162 (2004). [CrossRef]  

15. Y. Li, K. Itoh, W. Watanabe, K. Yamada, D. Kuroda, J. Nishii, and Y. Jiang, “Three-dimensional hole drilling of silica glass from the rear surface with femtosecond laser pulses,” Opt. Lett. 26(23), 1912 (2001). [CrossRef]  

16. D. Hwang, T. Choi, and C. Grigoropoulos, “Liquid-assisted femtosecond laser drilling of straight and three-dimensional microchannels in glass,” Appl. Phys. A 79(3), 605–612 (2004). [CrossRef]  

17. R. An, Y. Li, Y. Dou, H. Yang, and Q. Gong, “Simultaneous multi-microhole drilling of soda-lime glass by water-assisted ablation with femtosecond laser pulses,” Opt. Express 13(6), 1855 (2005). [CrossRef]  

18. X. Zhao and Y. C. Shin, “Femtosecond laser drilling of high-aspect ratio microchannels in glass,” Appl. Phys. A 104(2), 713–719 (2011). [CrossRef]  

19. M. Chanal, V. Y. Fedorov, M. Chambonneau, R. Clady, S. Tzortzakis, and D. Grojo, “Crossing the threshold of ultrafast laser writing in bulk silicon,” Nat. Commun. 8(1), 773 (2017). [CrossRef]  

20. V. V. Kononenko, V. V. Konov, and E. M. Dianov, “Delocalization of femtosecond radiation in silicon,” Opt. Lett. 37(16), 3369 (2012). [CrossRef]  

21. A. Mouskeftaras, A. V. Rode, R. Clady, M. Sentis, O. Utéza, and D. Grojo, “Self-limited underdense microplasmas in bulk silicon induced by ultrashort laser pulses,” Appl. Phys. Lett. 105(19), 191103 (2014). [CrossRef]  

22. H. Kämmer, G. Matthäus, S. Nolte, M. Chanal, O. Utéza, and D. Grojo, “In-volume structuring of silicon using picosecond laser pulses,” Appl. Phys. A 124(4), 302 (2018). [CrossRef]  

23. O. Tokel, A. Turnalı, G. Makey, P. Elahi, T. Çolakoğlu, E. Ergeçen, Ö. Yavuz, R. Hübner, M. Zolfaghari Borra, I. Pavlov, A. Bek, R. Turan, D. K. Kesim, S. Tozburun, S. Ilday, and F. Ö. Ilday, “In-chip microstructures and photonic devices fabricated by nonlinear laser lithography deep inside silicon,” Nat. Photonics 11(10), 639–645 (2017). [CrossRef]  

24. Q. Ma, H. Zhu, Z. Zhang, K. Xu, X. Dai, S. Zhu, and A. Wang, “An Investigation into Picosecond Laser Micro-Trepanning of Alumina Ceramics Employing a Semi-Water-Immersed Scheme,” Materials 12(11), 1812 (2019). [CrossRef]  

25. S. Döring, “Analysis of the Hole Shape Evolution in Ultrashort Pulse Laser Drilling,” Ph.D. thesis, Friedrich-Schiller-University, Institute of Applied Physics (2014).

26. R. Rieske, R. Landgraf, and K. J. Wolter, “Novel method for crystal defect analysis of laser drilled TSVs,” in Proc. - Electron. Components Technol. Conf., (2009), pp. 1139–1146.

27. G. M. Hale and M. R. Querry, “Optical Constants of Water in the 200-nm to 200-μm Wavelength Region,” Appl. Opt. 12(3), 555 (1973). [CrossRef]  

28. A. Ben-Yakar and R. L. Byer, “Femtosecond laser ablation properties of borosilicate glass,” J. Appl. Phys. 96(9), 5316–5323 (2004). [CrossRef]  

29. S. Döring, S. Richter, S. Nolte, and A. Tünnermann, “In situ imaging of hole shape evolution in ultrashort pulse laser drilling,” Opt. Express 18(19), 20395 (2010). [CrossRef]  

30. L. S. Jiao, E. Ng, H. Y. Zheng, and Y. L. Zhang, “Theoretical study of pre-formed hole geometries on femtosecond pulse energy distribution in laser drilling,” Opt. Express 23(4), 4927 (2015). [CrossRef]  

31. H. Hidai, Y. Kuroki, S. Matsusaka, A. Chiba, and N. Morita, “Curved drilling via inner hole laser reflection,” Precis. Eng. 46, 96–103 (2016). [CrossRef]  

32. F. He, J. Yu, Y. Tan, W. Chu, C. Zhou, Y. Cheng, and K. Sugioka, “Tailoring femtosecond 1.5-μm Bessel beams for manufacturing high-aspect-ratio through-silicon vias,” Sci. Rep. 7(1), 40785 (2017). [CrossRef]  

33. B. Xia, L. Jiang, X. Li, X. Yan, and Y. Lu, “Mechanism and elimination of bending effect in femtosecond laser deep-hole drilling,” Opt. Express 23(21), 27853 (2015). [CrossRef]  

34. K. V. Gubin, K. V. Lotov, V. I. Trunov, and E. V. Pestryakov, “Modification of narrow ablating capillaries under the influence of multiple femtosecond laser pulses,” J. Appl. Phys. 120(11), 113103 (2016). [CrossRef]  

35. T. Matsumura, T. Nakatani, and T. Yagi, “Deep drilling on a silicon plate with a femtosecond laser: experiment and model analysis,” Appl. Phys. A 86(1), 107–114 (2006). [CrossRef]  

36. S. Döring, S. Richter, A. Tünnermann, and S. Nolte, “Evolution of hole depth and shape in ultrashort pulse deep drilling in silicon,” Appl. Phys. A 105(1), 69–74 (2011). [CrossRef]  

37. D. J. Hwang, K. Hiromatsu, H. Hidai, and C. P. Grigoropoulos, “Self-guided glass drilling by femtosecond laser pulses,” Appl. Phys. A 94(3), 555–558 (2009). [CrossRef]  

38. H. Liu, F. Chen, X. Wang, Q. Yang, H. Bian, J. Si, and X. Hou, “Influence of liquid environments on femtosecond laser ablation of silicon,” Thin Solid Films 518(18), 5188–5194 (2010). [CrossRef]  

39. J. Ren, M. Kelly, and L. Hesselink, “Laser ablation of silicon in water with nanosecond and femtosecond pulses,” Opt. Lett. 30(13), 1740 (2005). [CrossRef]  

40. A. Dupont, P. Caminat, P. Bournot, and J. P. Gauchon, “Enhancement of material ablation using 248, 308, 532, 1064 nm laser pulse with a water film on the treated surface,” J. Appl. Phys. 78(3), 2022–2028 (1995). [CrossRef]  

41. J. Lu, R. Q. Xu, X. Chen, Z. H. Shen, X. W. Ni, S. Y. Zhang, and C. M. Gao, “Mechanisms of laser drilling of metal plates underwater,” J. Appl. Phys. 95(8), 3890–3894 (2004). [CrossRef]  

42. L. L. Taylor, J. Qiao, and J. Qiao, “Optimization of femtosecond laser processing of silicon via numerical modeling,” Opt. Mater. Express 6(9), 2745 (2016). [CrossRef]  

43. L. L. Taylor, R. E. Scott, and J. Qiao, “Integrating two-temperature and classical heat accumulation models to predict femtosecond laser processing of silicon,” Opt. Mater. Express 8(3), 648 (2018). [CrossRef]  

44. M. Silvennoinen, J. Kaakkunen, K. Paivasaari, and P. Vahimaa, “Parallel femtosecond laser ablation with individually controlled intensity,” Opt. Express 22(3), 2603 (2014). [CrossRef]  

45. B. J. Kim, H. A. Man-Lyun, and Y. S. Kwon, “New through-wafer via interconnections with thick oxidized porous silicon sidewall via,” Jpn. J. Appl. Phys. 45(8A), 6141–6145 (2006). [CrossRef]  

46. L. Yu, H. Yang, T. T. Jing, M. Xu, R. Geer, and W. Wang, “Electrical characterization of RF TSV for 3D multi-core and heterogeneous ICs,” in IEEE/ACM Int. Conf. Comput. Des. Dig. Tech. Pap. ICCAD, (2010), pp. 686–693.

47. B. Xia, L. Jiang, X. Li, X. Yan, W. Zhao, and Y. Lu, “High aspect ratio, high-quality microholes in PMMA: a comparison between femtosecond laser drilling in air and in vacuum,” Appl. Phys. A: Mater. Sci. Process. 119(1), 61–68 (2015). [CrossRef]  

48. A. C. Fischer, S. J. Bleiker, T. Haraldsson, N. Roxhed, G. Stemme, and F. Niklaus, “Very high aspect ratio through-silicon vias (TSVs) fabricated using automated magnetic assembly of nickel wires,” J. Micromech. Microeng. 22(10), 105001 (2012). [CrossRef]  

49. S. J. Bleiker, A. C. Fischer, U. Shah, N. Somjit, T. Haraldsson, N. Roxhed, J. Oberhammer, G. Stemme, and F. Niklaus, “High-Aspect-Ratio Through Silicon Vias for High-Frequency Application Fabricated by Magnetic Assembly of Gold-Coated Nickel Wires,” IEEE Trans. Compon., Packag., Manuf. Technol. 5(1), 21–27 (2015). [CrossRef]  

50. M. Asiatici, M. J. Laakso, A. C. Fischer, G. Stemme, and F. Niklaus, “Through Silicon Vias With Invar Metal Conductor for High-Temperature Applications,” J. Microelectromech. Syst. 26(1), 158–168 (2017). [CrossRef]  

51. M. J. Laakso, S. J. Bleiker, J. Liljeholm, G. E. Mårtensson, M. Asiatici, A. C. Fischer, G. Stemme, T. Ebefors, and F. Niklaus, “Through-Glass Vias for Glass Interposers and MEMS Packaging Applications Fabricated Using Magnetic Assembly of Microscale Metal Wires,” IEEE Access 6, 44306–44317 (2018). [CrossRef]  

52. S. Bruneau, J. Hermann, G. Dumitru, M. Sentis, and E. Axente, “Ultra-fast laser ablation applied to deep-drilling of metals,” Appl. Surf. Sci. 248(1-4), 299–303 (2005). [CrossRef]  

53. S. W. Ho, S. W. Yoon, Q. Zhou, K. Pasad, V. Kripesh, and J. H. Lau, “High RF performance TSV silicon carrier for high frequency application,” in Proc. - Electron. Components Technol. Conf., (2008), pp. 1946–1952.