November 15, 2013 / Vol. 38, No. 22 / OPTICS LETTERS
4613
Etching rate enhancement by shaped femtosecond pulse train electron dynamics control for microchannels fabrication in fused silica glass Pengjun Liu,1 Lan Jiang,1 Jie Hu,1,* Xueliang Yan,1 Bo Xia,1 and Yongfeng Lu2 1
Laser Micro/Nano Fabrication Laboratory School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China 2
Department of Electrical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0511, USA *Corresponding author: [email protected] Received August 1, 2013; revised September 23, 2013; accepted October 3, 2013; posted October 7, 2013 (Doc. ID 195103); published November 6, 2013
The dependence of the etching rate on the ultrafast pulse shaping is observed when microchannels are fabricated in fused silica glass using the method of femtosecond laser irradiation followed by chemical etching. In comparison with the conventional femtosecond pulses, the temporally shaped pulse trains can greatly enhance the etching rate under the same processing conditions. The enhancement is mainly attributed to the localized transient electron dynamics control by shaping the ultrafast pulse, resulting in higher photon absorption efficiency and uniform photomodification zone. Furthermore, processing parameters, including pulse delay and pulse energy distribution ratio, have also been investigated to optimize microchannels fabrication. © 2013 Optical Society of America OCIS codes: (140.3390) Laser materials processing; (220.4000) Microstructure fabrication. http://dx.doi.org/10.1364/OL.38.004613
The microchannel is one of the most important structures in a lab-on-a-chip device, which is designed with a goal of miniaturizing an entire biological or chemical laboratory onto a single glass substrate [1]. A femtosecond (fs) laser is a promising tool for microchannel fabrication. Generally speaking, there are two kinds of fabrication methods: fs laser direct microchannel drilling and fs laser irradiation followed by chemical etching (FLICE). The former method is based on ablationinduced material removal. Hollow microchannels are directly fabricated using fs pulses [2], temporally shaped fs pulses based on electron dynamics control [3], spatially shaped fs pulses [4], and other assisted processes, e.g., liquid [5] and ultrasonic wave [6], aimed at fabricating high-aspect-ratio microchannels. However, the microchannel length is limited because of the debris deposition. Liao et al. [7] fabricated arbitrary length microchannels by using porous glass. But such a process requires specific materials and needs an additional postannealing process to remove the pores. FLICE is an ideal method for longer microchannels fabricated in fused silica, which mainly consists of two steps: (1) irradiation of the fused silica by focused fs laser pulses and (2) etching of the laser irradiation zone by a hydrofluoric acid (HF) solution [8]. For FLICE, a main challenge is to improve the morphologies of the microchannels. Helical microchannels of arbitrary length and uniform diameter were fabricated by introducing extra access ports [9]. The cross section shape of the microchannels could be controlled by suitable wobbling of the glass substrate or using astigmatically shaped fs pulses during the irradiation process [10,11]. Another significant challenge is to improve the etching rate or etching selectivity. Etching selectivity of the laser irradiation zone could be increased by 2 times using KOH solution instead of HF solution [12]. Hnatovsky et al. [13] investigated the influence of laser polarization on the etching rate, revealing that enhanced etching selectivity can be obtained when the laser scanning direction 0146-9592/13/224613-04$15.00/0
is perpendicular to the laser polarization direction, which is due to the long range, periodic, polarization-dependent nanostructures formed in the laser irradiation zone. In this study, we experimentally demonstrate the dependence of the etching rate on the pulse shaping, revealing that the etching rate of the laser irradiation zone can be greatly enhanced by temporally shaping the fs pulse. Two possible reasons are proposed to explain the etching rate enhancement, and the underlying physical mechanisms are also discussed. In addition, the processing parameters, e.g., pulse delay, and pulse energy distribution ratio have also been investigated for optimal microchannel fabrication. The fs laser pulse is linearly polarized, with a central wavelength of 800 nm, pulse duration of 50 fs, and repetition rate of 1 kHz. A 4f pulse shaper (BSI MIIPS BOX 640) is utilized to compensate for the phase distortion automatically, generating transform-limited pulse. The fs pulse can also be shaped to be pulse train with 50 fs subpulse duration by the pulse shaper. The sample is a 1.0 mm thick, all-surfaces polished fused silica glass. The laser beam is focused beneath the rear surface of the sample by a 10× microscope objective (NA 0.3). The sample is moved along the z direction at different speeds, resulting in the photomodification of the laser irradiation zone in a single step, as shown in Fig. 1. In our experiments, conventional pulses (single pulse per train) and double pulses (double pulses per train) are investigated. The total energies of the conventional pulses and the double pulses are always kept the same by the combination of a half-wave plate and a polarizer. The sample is irradiated under different scanning speeds ranging from 5 μm∕s to 300 μm∕s; and then, the sample is etched in ∼8% HF aqueous solution for 5 h. The ultrasonic bath is used to refresh the etching agent during the etching process every half an hour. The microchannels are characterized by an optical microscope (Olympus BX51). © 2013 Optical Society of America
4614
OPTICS LETTERS / Vol. 38, No. 22 / November 15, 2013
Fig. 1. Schematic diagram of the experimental setup. The inset is the partially enlarged view of the laser irradiation zone in the sample. HWP, half-wave plate; P, polarizer.
In the first set of experiments, the delay time of the double pulse is fixed at 500 fs, and the energies of the two subpulses are equal. Figure 2 shows the microscopic images of the etched microchannels by (a) conventional pulses and (b) double pulses at irradiation energy of 0.5 μJ. The section shape of the etched microchannels is conical in the two cases, which is because the etching rate of the laser irradiation zone is higher than that of the substrate material [14] and the entrance etching time is longer with respect to the buried regions. When the laser scanning speed is 5 μm∕s, the etching lengths of the microchannels irradiated by the conventional pulses and double pulses are about 60 and 600 μm, respectively. Here, we define the etching rate as the etching length per hour. In this case, the etching rate of the zone irradiated by double pulses is about 10 times greater than that by the conventional pulses. Although the etching lengths both decrease slightly at the scanning speed of 10 μm∕s, the enhancement of the etching rate by the double pulses is more obvious. In order to quantitatively explore the etching rate enhancement, we carefully calculate the etching rates under different irradiation energies and laser scanning speeds. In Fig. 3(a), for double pulses, when the irradiation pulse energy is 0.5 μJ and the laser scanning speed is increased from 5 μm∕s to 20 μm∕s, the etching rate is about 120 μm∕h and keeps nearly constant. However, for conventional pulses, the etching rate is less than 15 μm∕h within this scanning speed interval. When the scanning speed is increased further, the etching rate by the double pulses drops down dramatically. In Fig. 3(b), the irradiation pulse energy is increased to 1.5 μJ. We divide the scanning speeds into three regimes. In Regime I, the laser
Fig. 2. Microscopic images of the etched microchannels irradiated by (a) conventional pulses and (b) double pulses. The irradiation energy is fixed at 0.5 μJ, 1 kHz repetition rate.
scanning speed is between 5 μm∕s and 20 μm∕s, and the etching rates of the two cases are almost the same. In Regime II, the laser scanning speed is between 50 μm∕s and 100 μm∕s, and the etching rates of the zones irradiated by double pulses are significantly enhanced. While in Regime III (the laser scanning speed > 100 μm∕s), the enhancement effect by double pulses is gradually reduced. For double pulses, Regimes I and II are the saturation regimes because the etching rate keeps nearly its maximum. The larger saturation regime embodies the advantages of the pulse train technique in micromachining. Similar etching rate changes can be observed when the irradiation pulse energy is increased to 2 μJ, as shown in Fig. 3(c). In addition to etching rate enhancement, an optimum laser scanning speed also exists [see Figs. 3(a)–3(c)], above which the etching rate begins to drop down dramatically. The optimum scanning speeds of the double pulses and the conventional pulses are plotted against the irradiation energy, as shown in Fig. 3(d). It can be seen that the optimum scanning speeds of double pulses are about 3–5 times greater than that of the conventional pulses. A higher optimum scanning speed indicates that the sample can be photomodified more efficiently, which is beneficial for waveguide and microchannel fabrication. The experimental results show the influences of the pulse shaping on the etching rate, which is closely related with the photomodification properties of the laser irradiation zone. As for the enhancement by the double pulses, there may be two possible reasons for this phenomenon. First, the photomodification of the laser irradiation zone can be enhanced by the double pulses because of its higher photon absorption efficiency. Figure 4 shows the optical microscope image of the internal photomodification at a focusing depth of (a) 100, (b) 150, and (c) 200 μm below the surface, by double pulses and conventional pulses. The total irradiation pulse energies and laser shot numbers of the two cases are kept the same. The zones irradiated by conventional pulses appear
Fig. 3. Etching rate of conventional pulses and double pulses as a function of the laser scanning speed at energy of (a) 0.5 μJ, (b) 1.5 μJ, and (c) 2 μJ. (d) Shows the plot of optimum laser scanning speed against the irradiation energy.
November 15, 2013 / Vol. 38, No. 22 / OPTICS LETTERS
Fig. 4. Optical microscope image of internal photomodification at focusing depth of (a) 100 μm, (b) 150 μm, and (c) 200 μm below the surface. The scale bar is 50 μm.
bright because of refractive index change [15], and this kind of photomodification is weak in etching selectivity. However, disruptive photomodification can be observed at the double pulses irradiation zones, which exhibit a higher etching selectivity. Actually for double pulses irradiation, free electrons are excited by the first subpulse through nonlinear ionization (photoionization and impact ionization). These free electrons couple with the lattice over a time scale ranging from less than a picosecond to several tens of picoseconds depending on the material properties. For fused silica, the free electron relaxation time is typically less than 500 fs [16], which is even much shorter after strong ionizations by fs pulses. Our previous simulations demonstrate that it is possible to manipulate transient free electron density distributions inside fused silica by shaping the pulse train to achieve better ablation quality and higher photon efficiencies [17]. The manipulability of energy distribution by the pulse train technology makes it possible to control the transient (fs time scale), localized (nanometer length scale) free electron densities of the irradiation area, which in turn determines the optical and thermal properties. Especially for our experiments, the overall reflectivity decreases as the pulse separation increases from 0 fs to several hundreds of femtoseconds. The lower reflectivity increases the photon absorption efficiencies, which leads to stronger photomodification of the laser irradiation zone, resulting in a higher etching rate [18]. Second, the laser irradiation zone can also be photomodified more uniformly with less density fluctuation by the double pulses. This is because the first subpulse acts as a preconditioner, and then the subsequent pulse effectively induces the uniform modification of the laser irradiation zone along the laser scanning direction [19], which can promote the chemical etching process, hence increasing the etching rate further. The aforementioned experiments indicate that the etching rate of the laser irradiation zone could be greatly enhanced by the temporally shaped double pulses. To determine the optimal parameters for microchannel fabrication, another set of experiments are conducted. The irradiation energy is fixed at 1.5 μJ, 1 kHz, and the laser scanning speed is fixed at 30 μm∕s.
4615
Fig. 5. Plot of the etching rates versus the pulse delay at a pulse energy of 1.5 μJ, the laser scanning speed is fixed at 30 μm∕s, 1 kHz repetition rate.
Figure 5 shows the etching rate versus the double pulse delay ranging from 0 fs to 4 ps (a longer pulse delay is limited by the equipment). Each data point represents an average of two experimental data, and the error bars are shown. It is worth noting that double pulse with a delay time of 0 fs is actually the conventional pulse. It can be clearly seen that the etching rate of the laser irradiation zone irradiated by double pulses is enhanced by about 2 times with respect to the conventional pulses. Within the delay time of 0–500 fs, the etching rate increases, and a small peak is observed at around 500 fs. According to our discussion, when the pulse delay is within the electron relaxation time (
4613
Etching rate enhancement by shaped femtosecond pulse train electron dynamics control for microchannels fabrication in fused silica glass Pengjun Liu,1 Lan Jiang,1 Jie Hu,1,* Xueliang Yan,1 Bo Xia,1 and Yongfeng Lu2 1
Laser Micro/Nano Fabrication Laboratory School of Mechanical Engineering, Beijing Institute of Technology, Beijing 100081, China 2
Department of Electrical Engineering, University of Nebraska-Lincoln, Lincoln, Nebraska 68588-0511, USA *Corresponding author: [email protected] Received August 1, 2013; revised September 23, 2013; accepted October 3, 2013; posted October 7, 2013 (Doc. ID 195103); published November 6, 2013
The dependence of the etching rate on the ultrafast pulse shaping is observed when microchannels are fabricated in fused silica glass using the method of femtosecond laser irradiation followed by chemical etching. In comparison with the conventional femtosecond pulses, the temporally shaped pulse trains can greatly enhance the etching rate under the same processing conditions. The enhancement is mainly attributed to the localized transient electron dynamics control by shaping the ultrafast pulse, resulting in higher photon absorption efficiency and uniform photomodification zone. Furthermore, processing parameters, including pulse delay and pulse energy distribution ratio, have also been investigated to optimize microchannels fabrication. © 2013 Optical Society of America OCIS codes: (140.3390) Laser materials processing; (220.4000) Microstructure fabrication. http://dx.doi.org/10.1364/OL.38.004613
The microchannel is one of the most important structures in a lab-on-a-chip device, which is designed with a goal of miniaturizing an entire biological or chemical laboratory onto a single glass substrate [1]. A femtosecond (fs) laser is a promising tool for microchannel fabrication. Generally speaking, there are two kinds of fabrication methods: fs laser direct microchannel drilling and fs laser irradiation followed by chemical etching (FLICE). The former method is based on ablationinduced material removal. Hollow microchannels are directly fabricated using fs pulses [2], temporally shaped fs pulses based on electron dynamics control [3], spatially shaped fs pulses [4], and other assisted processes, e.g., liquid [5] and ultrasonic wave [6], aimed at fabricating high-aspect-ratio microchannels. However, the microchannel length is limited because of the debris deposition. Liao et al. [7] fabricated arbitrary length microchannels by using porous glass. But such a process requires specific materials and needs an additional postannealing process to remove the pores. FLICE is an ideal method for longer microchannels fabricated in fused silica, which mainly consists of two steps: (1) irradiation of the fused silica by focused fs laser pulses and (2) etching of the laser irradiation zone by a hydrofluoric acid (HF) solution [8]. For FLICE, a main challenge is to improve the morphologies of the microchannels. Helical microchannels of arbitrary length and uniform diameter were fabricated by introducing extra access ports [9]. The cross section shape of the microchannels could be controlled by suitable wobbling of the glass substrate or using astigmatically shaped fs pulses during the irradiation process [10,11]. Another significant challenge is to improve the etching rate or etching selectivity. Etching selectivity of the laser irradiation zone could be increased by 2 times using KOH solution instead of HF solution [12]. Hnatovsky et al. [13] investigated the influence of laser polarization on the etching rate, revealing that enhanced etching selectivity can be obtained when the laser scanning direction 0146-9592/13/224613-04$15.00/0
is perpendicular to the laser polarization direction, which is due to the long range, periodic, polarization-dependent nanostructures formed in the laser irradiation zone. In this study, we experimentally demonstrate the dependence of the etching rate on the pulse shaping, revealing that the etching rate of the laser irradiation zone can be greatly enhanced by temporally shaping the fs pulse. Two possible reasons are proposed to explain the etching rate enhancement, and the underlying physical mechanisms are also discussed. In addition, the processing parameters, e.g., pulse delay, and pulse energy distribution ratio have also been investigated for optimal microchannel fabrication. The fs laser pulse is linearly polarized, with a central wavelength of 800 nm, pulse duration of 50 fs, and repetition rate of 1 kHz. A 4f pulse shaper (BSI MIIPS BOX 640) is utilized to compensate for the phase distortion automatically, generating transform-limited pulse. The fs pulse can also be shaped to be pulse train with 50 fs subpulse duration by the pulse shaper. The sample is a 1.0 mm thick, all-surfaces polished fused silica glass. The laser beam is focused beneath the rear surface of the sample by a 10× microscope objective (NA 0.3). The sample is moved along the z direction at different speeds, resulting in the photomodification of the laser irradiation zone in a single step, as shown in Fig. 1. In our experiments, conventional pulses (single pulse per train) and double pulses (double pulses per train) are investigated. The total energies of the conventional pulses and the double pulses are always kept the same by the combination of a half-wave plate and a polarizer. The sample is irradiated under different scanning speeds ranging from 5 μm∕s to 300 μm∕s; and then, the sample is etched in ∼8% HF aqueous solution for 5 h. The ultrasonic bath is used to refresh the etching agent during the etching process every half an hour. The microchannels are characterized by an optical microscope (Olympus BX51). © 2013 Optical Society of America
4614
OPTICS LETTERS / Vol. 38, No. 22 / November 15, 2013
Fig. 1. Schematic diagram of the experimental setup. The inset is the partially enlarged view of the laser irradiation zone in the sample. HWP, half-wave plate; P, polarizer.
In the first set of experiments, the delay time of the double pulse is fixed at 500 fs, and the energies of the two subpulses are equal. Figure 2 shows the microscopic images of the etched microchannels by (a) conventional pulses and (b) double pulses at irradiation energy of 0.5 μJ. The section shape of the etched microchannels is conical in the two cases, which is because the etching rate of the laser irradiation zone is higher than that of the substrate material [14] and the entrance etching time is longer with respect to the buried regions. When the laser scanning speed is 5 μm∕s, the etching lengths of the microchannels irradiated by the conventional pulses and double pulses are about 60 and 600 μm, respectively. Here, we define the etching rate as the etching length per hour. In this case, the etching rate of the zone irradiated by double pulses is about 10 times greater than that by the conventional pulses. Although the etching lengths both decrease slightly at the scanning speed of 10 μm∕s, the enhancement of the etching rate by the double pulses is more obvious. In order to quantitatively explore the etching rate enhancement, we carefully calculate the etching rates under different irradiation energies and laser scanning speeds. In Fig. 3(a), for double pulses, when the irradiation pulse energy is 0.5 μJ and the laser scanning speed is increased from 5 μm∕s to 20 μm∕s, the etching rate is about 120 μm∕h and keeps nearly constant. However, for conventional pulses, the etching rate is less than 15 μm∕h within this scanning speed interval. When the scanning speed is increased further, the etching rate by the double pulses drops down dramatically. In Fig. 3(b), the irradiation pulse energy is increased to 1.5 μJ. We divide the scanning speeds into three regimes. In Regime I, the laser
Fig. 2. Microscopic images of the etched microchannels irradiated by (a) conventional pulses and (b) double pulses. The irradiation energy is fixed at 0.5 μJ, 1 kHz repetition rate.
scanning speed is between 5 μm∕s and 20 μm∕s, and the etching rates of the two cases are almost the same. In Regime II, the laser scanning speed is between 50 μm∕s and 100 μm∕s, and the etching rates of the zones irradiated by double pulses are significantly enhanced. While in Regime III (the laser scanning speed > 100 μm∕s), the enhancement effect by double pulses is gradually reduced. For double pulses, Regimes I and II are the saturation regimes because the etching rate keeps nearly its maximum. The larger saturation regime embodies the advantages of the pulse train technique in micromachining. Similar etching rate changes can be observed when the irradiation pulse energy is increased to 2 μJ, as shown in Fig. 3(c). In addition to etching rate enhancement, an optimum laser scanning speed also exists [see Figs. 3(a)–3(c)], above which the etching rate begins to drop down dramatically. The optimum scanning speeds of the double pulses and the conventional pulses are plotted against the irradiation energy, as shown in Fig. 3(d). It can be seen that the optimum scanning speeds of double pulses are about 3–5 times greater than that of the conventional pulses. A higher optimum scanning speed indicates that the sample can be photomodified more efficiently, which is beneficial for waveguide and microchannel fabrication. The experimental results show the influences of the pulse shaping on the etching rate, which is closely related with the photomodification properties of the laser irradiation zone. As for the enhancement by the double pulses, there may be two possible reasons for this phenomenon. First, the photomodification of the laser irradiation zone can be enhanced by the double pulses because of its higher photon absorption efficiency. Figure 4 shows the optical microscope image of the internal photomodification at a focusing depth of (a) 100, (b) 150, and (c) 200 μm below the surface, by double pulses and conventional pulses. The total irradiation pulse energies and laser shot numbers of the two cases are kept the same. The zones irradiated by conventional pulses appear
Fig. 3. Etching rate of conventional pulses and double pulses as a function of the laser scanning speed at energy of (a) 0.5 μJ, (b) 1.5 μJ, and (c) 2 μJ. (d) Shows the plot of optimum laser scanning speed against the irradiation energy.
November 15, 2013 / Vol. 38, No. 22 / OPTICS LETTERS
Fig. 4. Optical microscope image of internal photomodification at focusing depth of (a) 100 μm, (b) 150 μm, and (c) 200 μm below the surface. The scale bar is 50 μm.
bright because of refractive index change [15], and this kind of photomodification is weak in etching selectivity. However, disruptive photomodification can be observed at the double pulses irradiation zones, which exhibit a higher etching selectivity. Actually for double pulses irradiation, free electrons are excited by the first subpulse through nonlinear ionization (photoionization and impact ionization). These free electrons couple with the lattice over a time scale ranging from less than a picosecond to several tens of picoseconds depending on the material properties. For fused silica, the free electron relaxation time is typically less than 500 fs [16], which is even much shorter after strong ionizations by fs pulses. Our previous simulations demonstrate that it is possible to manipulate transient free electron density distributions inside fused silica by shaping the pulse train to achieve better ablation quality and higher photon efficiencies [17]. The manipulability of energy distribution by the pulse train technology makes it possible to control the transient (fs time scale), localized (nanometer length scale) free electron densities of the irradiation area, which in turn determines the optical and thermal properties. Especially for our experiments, the overall reflectivity decreases as the pulse separation increases from 0 fs to several hundreds of femtoseconds. The lower reflectivity increases the photon absorption efficiencies, which leads to stronger photomodification of the laser irradiation zone, resulting in a higher etching rate [18]. Second, the laser irradiation zone can also be photomodified more uniformly with less density fluctuation by the double pulses. This is because the first subpulse acts as a preconditioner, and then the subsequent pulse effectively induces the uniform modification of the laser irradiation zone along the laser scanning direction [19], which can promote the chemical etching process, hence increasing the etching rate further. The aforementioned experiments indicate that the etching rate of the laser irradiation zone could be greatly enhanced by the temporally shaped double pulses. To determine the optimal parameters for microchannel fabrication, another set of experiments are conducted. The irradiation energy is fixed at 1.5 μJ, 1 kHz, and the laser scanning speed is fixed at 30 μm∕s.
4615
Fig. 5. Plot of the etching rates versus the pulse delay at a pulse energy of 1.5 μJ, the laser scanning speed is fixed at 30 μm∕s, 1 kHz repetition rate.
Figure 5 shows the etching rate versus the double pulse delay ranging from 0 fs to 4 ps (a longer pulse delay is limited by the equipment). Each data point represents an average of two experimental data, and the error bars are shown. It is worth noting that double pulse with a delay time of 0 fs is actually the conventional pulse. It can be clearly seen that the etching rate of the laser irradiation zone irradiated by double pulses is enhanced by about 2 times with respect to the conventional pulses. Within the delay time of 0–500 fs, the etching rate increases, and a small peak is observed at around 500 fs. According to our discussion, when the pulse delay is within the electron relaxation time (
