Groove Formation in Glass Substrate by a UV Nanosecond Laser

: This study reports the laser grooving of glass. In this study, an ultraviolet (UV) nanosecond laser is used to produce grooves in soda lime glass by varying laser power, scan speed, pulse repetition rate, and number of scans. The threshold power density to generate damage on glass surface is found to be ~6.37 × 10 8 W / cm 2 . According to the result of a single laser scan, at a high laser power and low scan speed, material removal does not nearly occur, and only volume melting underneath the glass surface and large thermal crack are induced. With the decreasing laser power and increasing scan speed, a groove with a smooth surface is formed owing to the melting-dominant material removal process. The groove of rough surface morphology is produced with the further increase of the scan speed due to shattering induced by shock wave. Under this condition, it is found that either small cracks are produced or a crack-free groove is obtained. In this study, the fabrication of the microchannel for bio-chip application is also attempted by multiple laser scans. The near rectangular-shaped and crack-free groove (width × depth: ~220 µ m × ~500 µ m) with quite a smooth surface is fabricated by the multiple laser scans with a non-zero scan spacing.


Introduction
The importance of glass as an engineering material is found in various fields of the industry. The high optical transparency and excellent dielectric property of the glass make it a suitable material for many optical and electrical applications. In addition, it has been widely used in biomedical areas owing to good chemical and thermal stability.
Commonly, the processing techniques (diamond [1], roll [2], and wheel cutting [3]) using a physical contact of tool bits were used to machine glass materials. However, these conventional methods generally produce random cracking and rough-cut surfaces in material. Moreover, the precision machining of the glass with these methods is very difficult.
A laser technique is greatly attractive in the processing of brittle glass materials because of its non-contact nature [4]. It can also provide a precise and clean process in a fast and automated way. Over the past several decades, many researchers have studied laser-based glass processing. In the early days, a CO 2 laser attracted great attention in the glass cutting process because the absorption rate of the CO 2 laser beam with 10.6 µm wavelength is very high in glass materials. However, in most cases, glass processing by the CO 2 laser beam involves a large thermal effect that induces various defects such as cracking, chipping, and heat affected zone (HAZ) [5][6][7][8]. This may require extra time and cost for cleaning and polishing of the cut-surface after the processing.

Experimental Details
The laser system (EVLASER Co., Ltd., Gyeonggi-do, Korea) used in this study consists of a UV (355 nm) Nd:YVO 4 laser source (INNGU LASER, Pulse 355-3A) and scanner module. The fundamental wavelength of the Nd:YVO 4 is 1064 nm. However, in this study, the Nd:YVO 4 laser with third-harmonic generation is used to give a laser beam of 355 nm wavelength. The Nd:YVO 4 laser has a pulse width of 12-22 nanoseconds and is operated at a 30-80 kHz pulse repetition rate. The pulse width information at various pulse repetition rates is shown in Table 1. A laser beam generated from the laser source is delivered onto the scanner. The laser scanner has a galvano mirror system and f-theta lens (110 mm × 110 mm scan filed), so that it can move the laser beam of uniform intensity on a workpiece in fast and precise way. The effective focal length of the f-theta lens is 160 mm, and this gives a focused beam spot size of~20 µm on the workpiece. In this experiment, the direction of the laser beam polarization is horizontal to the surface of glass. The schematic diagram of the UV nanosecond laser processing is shown in Figure 1. In this study, the experiments using both single laser scan and multiple laser scans are carried out. The multiple laser scanning methods used in the experiment are presented in Figure 2.   Laser scans are also carried at various laser average powers. In this study, the laser average power is measured using the digital laser power meter with a thermal head detector (ADM-1000, lasermet, Bournemouth, UK) before the grooving experiment. During the measurement of the laser power, a defocused laser beam continuously moves along the circular track with a 3 mm diameter on the detector (to avoid damage of the detector) until the laser power value displayed is stabilized. The stabilized power value is used as a process condition.
In this work, soda lime glass from a local supplier is used for laser processing. The thickness of the soda lime glass used in the experiment is ~1.1 mm. The material properties and chemical composition of the soda lime glass are listed in Tables 2 and 3, respectively. The surface morphology of the laser-irradiated glass substrates is analyzed by an optical microscope (Dino Lite, AM4013MZT) and a field emission scanning electron microscopy (FE-SEM, TESCAN MIRA LMH).   Multiple scanning methods. Method 1 indicates multiple laser scans with zero-scan spacing. Methods 2 and 3 are for multiple laser scans with non-zero scan spacing. In methods 2 and 3, ① and ② denote a scan sequence (① is the first part of the scan sequence).
Laser scans are also carried at various laser average powers. In this study, the laser average power is measured using the digital laser power meter with a thermal head detector (ADM-1000, lasermet, Bournemouth, UK) before the grooving experiment. During the measurement of the laser power, a defocused laser beam continuously moves along the circular track with a 3 mm diameter on the detector (to avoid damage of the detector) until the laser power value displayed is stabilized. The stabilized power value is used as a process condition.
In this work, soda lime glass from a local supplier is used for laser processing. The thickness of the soda lime glass used in the experiment is ~1.1 mm. The material properties and chemical composition of the soda lime glass are listed in Tables 2 and 3, respectively. The surface morphology of the laser-irradiated glass substrates is analyzed by an optical microscope (Dino Lite, AM4013MZT) and a field emission scanning electron microscopy (FE-SEM, TESCAN MIRA LMH). Laser scans are also carried at various laser average powers. In this study, the laser average power is measured using the digital laser power meter with a thermal head detector (ADM-1000, lasermet, Bournemouth, UK) before the grooving experiment. During the measurement of the laser power, a defocused laser beam continuously moves along the circular track with a 3 mm diameter on the detector (to avoid damage of the detector) until the laser power value displayed is stabilized. The stabilized power value is used as a process condition.
In this work, soda lime glass from a local supplier is used for laser processing. The thickness of the soda lime glass used in the experiment is~1.1 mm. The material properties and chemical composition of the soda lime glass are listed in Tables 2 and 3, respectively. The surface morphology of the laser-irradiated glass substrates is analyzed by an optical microscope (Dino Lite, AM4013MZT) and a field emission scanning electron microscopy (FE-SEM, TESCAN MIRA LMH).

Result of Damage Threshold Test
The minimum laser energy to induce a visible damage on glass substrate is found through a damage threshold test. In this test, the surface of the glass is scanned with the UV laser beam of 0.33-1.76 W. The scanning speed and pulse repetition rate used for this experiment are 5 mm/s and 30 kHz, respectively. Figure 3 shows the optical scope images of the glass surface irradiated by the laser beam. As seen in the figure, below the laser power of 0.72 W, any visible damage is not observed on the surface of the glass substrate. The calculated threshold power density corresponding to this power is 6.37 × 10 8 W/cm 2 . At this power density, the width of the heat affected zone (HAZ) is very narrow; however, the width of the HAZ and crack propagation significantly increase with the laser power, as shown in the figure.

Result of Damage Threshold Test
The minimum laser energy to induce a visible damage on glass substrate is found through a damage threshold test. In this test, the surface of the glass is scanned with the UV laser beam of 0.33-1.76 W. The scanning speed and pulse repetition rate used for this experiment are 5 mm/s and 30 kHz, respectively. Figure 3 shows the optical scope images of the glass surface irradiated by the laser beam. As seen in the figure, below the laser power of 0.72 W, any visible damage is not observed on the surface of the glass substrate. The calculated threshold power density corresponding to this power is 6.37 × 10 8 W/cm 2 . At this power density, the width of the heat affected zone (HAZ) is very narrow; however, the width of the HAZ and crack propagation significantly increase with the laser power, as shown in the figure.

Analysis of Glass Surface Morphology Produced by Single Laser Scan
The experiment using single laser scan is carried out by varying the laser power, scan speed, and pulse repetition rate. Figure 4 shows the optical images of the glass surface damaged by the single laser scan at various laser powers and scan speeds under the fixed pulse repetition rate of 30 kHz. It is seen that cracks are generated along the direction of scanning owing to the thermal stress induced by rapid heating and cooling during the process. It is clearly observed that the size of the crack increases with decreasing scan speed and increasing laser power owing to greater laser energy deposition on the glass. The size of cracks produced is several tens of microns to several hundreds of microns depending on the process condition. The crack has a seagull shape, and its one end propagates across the direction of scanning.

Analysis of Glass Surface Morphology Produced by Single Laser Scan
The experiment using single laser scan is carried out by varying the laser power, scan speed, and pulse repetition rate. Figure 4 shows the optical images of the glass surface damaged by the single laser scan at various laser powers and scan speeds under the fixed pulse repetition rate of 30 kHz. It is seen that cracks are generated along the direction of scanning owing to the thermal stress induced by rapid heating and cooling during the process. It is clearly observed that the size of the crack increases with decreasing scan speed and increasing laser power owing to greater laser energy deposition on the glass. The size of cracks produced is several tens of microns to several hundreds of microns depending on the process condition. The crack has a seagull shape, and its one end propagates across the direction of scanning.  Figure 5 shows the glass surface damaged by the single laser scan with higher pulse repetition rate. The aspect of the crack propagation is similar up to 50 kHz pulse repetition rate; however, the significant increase of the crack size is found from the pulse repetition rate of 60 kHz, as shown in the figure. Owing to the large growth of the cracks, neighboring cracks are linked to each other like in a chain, and their size reaches several millimeters. At a very high pulse repetition rate (80 kHz), the long crack of a line shape is eventually produced along the direction of scanning. The detailed morphology of the glass surface is investigated with SEM. Figure 6 shows SEM images of the glass surface scanned by a laser beam at a pulse repetition rate of 30 kHz. At the laser power of 3.32 W and scan speed of 5 mm/s, the glass surface irradiated by the laser beam shows a smooth wave morphology. This indicates that material removal does not nearly occur, and the interaction between the laser beam and glass only induces volume melting underneath the glass surface. The evidence of the volume melting is clearly observed at the end facet of the glass plate (the end point of the scanning path, see Figure 6i). As the laser power decreases and scan speed increases, the morphology of the glass surface is turned into a groove shape through the material removal process, as shown in the figure. It is found that the groove with a very smooth surface is obtained  Figure 5 shows the glass surface damaged by the single laser scan with higher pulse repetition rate. The aspect of the crack propagation is similar up to 50 kHz pulse repetition rate; however, the significant increase of the crack size is found from the pulse repetition rate of 60 kHz, as shown in the figure. Owing to the large growth of the cracks, neighboring cracks are linked to each other like in a chain, and their size reaches several millimeters. At a very high pulse repetition rate (80 kHz), the long crack of a line shape is eventually produced along the direction of scanning.  Figure 5 shows the glass surface damaged by the single laser scan with higher pulse repetition rate. The aspect of the crack propagation is similar up to 50 kHz pulse repetition rate; however, the significant increase of the crack size is found from the pulse repetition rate of 60 kHz, as shown in the figure. Owing to the large growth of the cracks, neighboring cracks are linked to each other like in a chain, and their size reaches several millimeters. At a very high pulse repetition rate (80 kHz), the long crack of a line shape is eventually produced along the direction of scanning. The detailed morphology of the glass surface is investigated with SEM. Figure 6 shows SEM images of the glass surface scanned by a laser beam at a pulse repetition rate of 30 kHz. At the laser power of 3.32 W and scan speed of 5 mm/s, the glass surface irradiated by the laser beam shows a smooth wave morphology. This indicates that material removal does not nearly occur, and the interaction between the laser beam and glass only induces volume melting underneath the glass surface. The evidence of the volume melting is clearly observed at the end facet of the glass plate (the end point of the scanning path, see Figure 6i). As the laser power decreases and scan speed increases, the morphology of the glass surface is turned into a groove shape through the material removal process, as shown in the figure. It is found that the groove with a very smooth surface is obtained The detailed morphology of the glass surface is investigated with SEM. Figure 6 shows SEM images of the glass surface scanned by a laser beam at a pulse repetition rate of 30 kHz. At the laser power of 3.32 W and scan speed of 5 mm/s, the glass surface irradiated by the laser beam shows a smooth wave morphology. This indicates that material removal does not nearly occur, and the Appl. Sci. 2020, 10, 987 6 of 11 interaction between the laser beam and glass only induces volume melting underneath the glass surface. The evidence of the volume melting is clearly observed at the end facet of the glass plate (the end point of the scanning path, see Figure 6i). As the laser power decreases and scan speed increases, the morphology of the glass surface is turned into a groove shape through the material removal process, as shown in the figure. It is found that the groove with a very smooth surface is obtained when a relatively low scanning speed is used. Under this process condition, the groove with thermal cracks is generated as seen in Figure 6b,e,f,j. With the further increase of the scanning speed, the groove produced has a very rough morphology; however, this process condition does not nearly induce cracks around the groove.
Appl. Sci. 2019, 9, x 6 of 12 when a relatively low scanning speed is used. Under this process condition, the groove with thermal cracks is generated as seen in Figure 6b,e,f,j. With the further increase of the scanning speed, the groove produced has a very rough morphology; however, this process condition does not nearly induce cracks around the groove. In the high power density laser-matter interaction, physical phenomena such as melting, vaporization, and plasma formation are involved, and material removal is mostly achieved through vaporization and melt ejection driven by high vapor pressure [24,25]. According to Figure 7a,b, it is seen that molten materials removed from the irradiation regions splash out on the surface near grooves. This indicates that melting is the dominant material removal mechanism. At a relatively high scanning speed (see Figure 7c,d), many molten glass powders not ejected are observed inside the groove owing to less laser energy deposition on the glass surface, which in turn makes a rough surface morphology at the irradiation zone. Under a relatively high scanning speed, the shattering of the surface is also seen at the edge region of the groove. This is attributable to the shock wave induced by plasma breakdown during the process. The sharp and non-melt glass pieces found around the groove in Figure 7d are the outcome of shattering. Therefore, in this case, high temperature vaporization inducing the shock wave as well as the melting contributes to material removal. The generation of micron bubbles is seen in HAZ (see Figure 8). During the interaction between the laser beam and glass, thermal energy inside the glass could induce thermal dissociation and other chemical reactions of the components shown in Table 3. This may generate the oxygen gas to be the source of the bubbles inside the glass. In the high power density laser-matter interaction, physical phenomena such as melting, vaporization, and plasma formation are involved, and material removal is mostly achieved through vaporization and melt ejection driven by high vapor pressure [24,25]. According to Figure 7a,b, it is seen that molten materials removed from the irradiation regions splash out on the surface near grooves. This indicates that melting is the dominant material removal mechanism. At a relatively high scanning speed (see Figure 7c,d), many molten glass powders not ejected are observed inside the groove owing to less laser energy deposition on the glass surface, which in turn makes a rough surface morphology at the irradiation zone. Under a relatively high scanning speed, the shattering of the surface is also seen at the edge region of the groove. This is attributable to the shock wave induced by plasma breakdown during the process. The sharp and non-melt glass pieces found around the groove in Figure 7d are the outcome of shattering. Therefore, in this case, high temperature vaporization inducing the shock wave as well as the melting contributes to material removal. The generation of micron bubbles is seen in HAZ (see Figure 8). During the interaction between the laser beam and glass, thermal energy inside the glass could induce thermal dissociation and other chemical reactions of the components shown in Table 3. This may generate the oxygen gas to be the source of the bubbles inside the glass.
On the basis of the results for single laser scan tests, some important findings could be described as follows. The interaction between the UV laser and material at a high laser power and low scan speed just induces volume melting underneath glass surface and large thermal crack without material Appl. Sci. 2020, 10, 987 7 of 11 removal. With the decreasing laser power and increasing scan speed, a groove with a smooth surface starts to form owing to the melting-dominant material removal process. A crack-free groove is only obtained at a relatively low laser power and high scanning speed. Under this process condition, both shattering induced by shock wave and melting contribute to the material removal. Process characteristics depending on the laser scanning condition are summarized in Table 4.  On the basis of the results for single laser scan tests, some important findings could be described as follows. The interaction between the UV laser and material at a high laser power and low scan speed just induces volume melting underneath glass surface and large thermal crack without material removal. With the decreasing laser power and increasing scan speed, a groove with a smooth surface starts to form owing to the melting-dominant material removal process. A crack-free groove is only obtained at a relatively low laser power and high scanning speed. Under this process condition, both shattering induced by shock wave and melting contribute to the material removal. Process characteristics depending on the laser scanning condition are summarized in Table 4.   On the basis of the results for single laser scan tests, some important findings could be described as follows. The interaction between the UV laser and material at a high laser power and low scan speed just induces volume melting underneath glass surface and large thermal crack without material removal. With the decreasing laser power and increasing scan speed, a groove with a smooth surface starts to form owing to the melting-dominant material removal process. A crack-free groove is only obtained at a relatively low laser power and high scanning speed. Under this process condition, both shattering induced by shock wave and melting contribute to the material removal. Process characteristics depending on the laser scanning condition are summarized in Table 4.

Groove Formation by Multiple Laser Scans
In order to produce the groove or channel structures with various widths and depths in the glass substrate, multiple laser scans are necessary. In this section, the groove formation by the multiple laser scans is described. One of the process conditions to produce the crack-free groove is used in the experiment for multiple laser scans. The process condition selected for the experiment is a laser power of 2.28 W, scan speed of 25 mm/s, and pulse repetition rate of 30 kHz. Figure 9 shows the result of multiple laser scans with zero scan spacing. In this experiment, a laser beam scans the same scan track several times. From Figure 9, it is confirmed that the increase of the scan number does not induce cracking around the irradiation region.
Appl. Sci. 2019, 9, x 8 of 12 a Type I: material removal does not nearly occur. b Type II: material removal mostly occurs through melt ejection. c Type III: material removal occurs through both melt ejection and shattering induced by shock wave.

Groove Formation by Multiple Laser Scans
In order to produce the groove or channel structures with various widths and depths in the glass substrate, multiple laser scans are necessary. In this section, the groove formation by the multiple laser scans is described. One of the process conditions to produce the crack-free groove is used in the experiment for multiple laser scans. The process condition selected for the experiment is a laser power of 2.28 W, scan speed of 25 mm/s, and pulse repetition rate of 30 kHz. Figure 9 shows the result of multiple laser scans with zero scan spacing. In this experiment, a laser beam scans the same scan track several times. From Figure 9, it is confirmed that the increase of the scan number does not induce cracking around the irradiation region.  Figure 10 shows the measured depth and width of the groove as a function of the number of scans. The optical images (cross sectional view of the grooves) used to measure the depth and width of the groove are shown in Figure 11. The width of the groove does not nearly change as the scan number increases, whereas the depth of the groove tends to increase with the scan number, as seen in the figure. In high power density laser-material interaction, the absorption of the laser beam in the material dramatically increases once a laser-induced crater is produced as a result of the multiple reflection of the laser beam in the crater [24]. This could help the material removal in multiple laser scans, although the intensity of the laser beam decreases owing to continuous beam defocusing with the depth of the groove.   Figure 11. The width of the groove does not nearly change as the scan number increases, whereas the depth of the groove tends to increase with the scan number, as seen in the figure. In high power density laser-material interaction, the absorption of the laser beam in the material dramatically increases once a laser-induced crater is produced as a result of the multiple reflection of the laser beam in the crater [24]. This could help the material removal in multiple laser scans, although the intensity of the laser beam decreases owing to continuous beam defocusing with the depth of the groove.    The multiple laser scans with a non-zero scan spacing is also tested to produce a wide groove. Two different scanning schemes are tested to evaluate the efficiency in material removal (see Figure  2). Figure 12 shows the cross section of the groove produced with two different scanning schemes using the same laser parameters. It is found that more material can be removed with method 3. In method 3, a rough surface morphology is uniformly produced for the whole scanned area after the first set of the laser scan is finished. This contributes to the acceleration of material removal owing to more laser energy absorption through the multiple reflection of the laser beam from the second set of the laser scan. The multiple laser scans with a non-zero scan spacing is also tested to produce a wide groove.
Two different scanning schemes are tested to evaluate the efficiency in material removal (see Figure 2). Figure 12 shows the cross section of the groove produced with two different scanning schemes using the same laser parameters. It is found that more material can be removed with method 3. In method 3, a rough surface morphology is uniformly produced for the whole scanned area after the first set of the laser scan is finished. This contributes to the acceleration of material removal owing to more laser energy absorption through the multiple reflection of the laser beam from the second set of the laser scan.  The multiple laser scans with a non-zero scan spacing is also tested to produce a wide groove.
Two different scanning schemes are tested to evaluate the efficiency in material removal (see Figure  2). Figure 12 shows the cross section of the groove produced with two different scanning schemes using the same laser parameters. It is found that more material can be removed with method 3. In method 3, a rough surface morphology is uniformly produced for the whole scanned area after the first set of the laser scan is finished. This contributes to the acceleration of material removal owing to more laser energy absorption through the multiple reflection of the laser beam from the second set of the laser scan.  Figure 13 shows the groove produced by multiple laser scans using method 3. A near rectangularshaped groove is produced to imitate the microchannel of a bio-chip. It is seen that the groove has a quite smooth inner surface. The width and depth of the groove are~220 µm and~500 µm, respectively. No visible crack is observed around the groove wall along the direction of scanning. This result may suggest that a UV nanosecond laser could be one of the potential tools for the fabrication of bio-chip channels. Figure 13 shows the groove produced by multiple laser scans using method 3. A near rectangular-shaped groove is produced to imitate the microchannel of a bio-chip. It is seen that the groove has a quite smooth inner surface. The width and depth of the groove are ~220 μm and ~500 μm, respectively. No visible crack is observed around the groove wall along the direction of scanning. This result may suggest that a UV nanosecond laser could be one of the potential tools for the fabrication of bio-chip channels.

Conclusions
In this study, the UV nanosecond laser grooving process for glass substrate is carried out, and the process characteristics including surface morphology, crack, and material removal mechanism are investigated. The important conclusions of this work are as follows.
(1) The damage threshold power density of the UV laser beam for soda lime glass surface is ~6.37 × 10 8 W/cm 2 (at 5 mm/s scan speed and 30 kHz pulse repetition rate).
(2) The groove produced by the single laser scan process can be categorized as two types based on the morphology of the groove. The first type of the groove has a smooth surface owing to the melting-dominant material removal process, and the cracks of several tens to hundreds of microns in size are usually produced along the groove of this type. The second type of the groove has a rough surface morphology due to shattering and non-ejected molten glass powders; however, in this case, the crack-free groove can be obtained when a relatively high scan speed is used. The process conditions producing the crack-free groove are shown in Table 5. (3) The near rectangular-shaped groove, imitating the microchannel for a bio-chip application, of several hundred micrometers in size and with smooth surface is successfully fabricated by multiple laser scans without crack. This result shows the feasibility of a UV nanosecond laser as a potential tool for the fabrication of bio-chip channels.
Author Contributions: J.S., conceptualization, data curation, formal analysis, funding acquisition, methodology, project administration, supervision, visualization, writing-original draft, writing-review & editing. K.N., data curation, investigation, methodology, visualization. All authors have read and agreed to the published version of the manuscript.

Conclusions
In this study, the UV nanosecond laser grooving process for glass substrate is carried out, and the process characteristics including surface morphology, crack, and material removal mechanism are investigated. The important conclusions of this work are as follows.
(1) The damage threshold power density of the UV laser beam for soda lime glass surface is 6.37 × 10 8 W/cm 2 (at 5 mm/s scan speed and 30 kHz pulse repetition rate).
(2) The groove produced by the single laser scan process can be categorized as two types based on the morphology of the groove. The first type of the groove has a smooth surface owing to the melting-dominant material removal process, and the cracks of several tens to hundreds of microns in size are usually produced along the groove of this type. The second type of the groove has a rough surface morphology due to shattering and non-ejected molten glass powders; however, in this case, the crack-free groove can be obtained when a relatively high scan speed is used. The process conditions producing the crack-free groove are shown in Table 5. (3) The near rectangular-shaped groove, imitating the microchannel for a bio-chip application, of several hundred micrometers in size and with smooth surface is successfully fabricated by multiple laser scans without crack. This result shows the feasibility of a UV nanosecond laser as a potential tool for the fabrication of bio-chip channels.