Ablation of silica glass induced by laser plasma soft X‐ray irradiation

Abstract

Silica glass can be ablated using focused laser plasma soft X‐rays. The ablation technique enables us to fabricate trenches with a width as narrow as 50 nm. In the present paper, we have investigated the nano‐ablation process. The soft X‐ray irradiation causes a silica surface broken into almost atomic species. Ionic species have kinetic energies higher than that gained by heating to the boiling point. We measured the ablation depth as a function of the soft X‐ray fluence. The depth analysis revealed that soft X‐rays are absorbed in a silica surface with an effective absorption depth of 10 nm. The result indicates that the energy density of the soft X‐rays per unit volume at the threshold fluence is comparable to that required for breaking silica glass into atomic species. Further, the results suggest that ablation occurs before diffusion of absorbed energy into the surrounding region. In addition to energy absorption, repulsive forces between ionic species may cause ablation of the silica surface by soft X‐ray irradiation. These properties of soft X‐ray ablation may make possible the nano‐ablation of silica glass. © 2011 Wiley Periodicals, Inc. Electron Comm Jpn, 94(9): 30–35, 2011; Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/ecj.10354

1. Micromachining of Transparent Materials

High‐throughput low‐cost methods for micro‐ and nanomachining of transparent organic materials may become the key technologies for fabrication of chemical microanalyzers and chemical microreactors in the field of biomedical engineering, as well as that of functional optical elements. There are several methods for machining of silica glass and other inorganic transparent materials, such as optical lithography [1], electron beam lithography [2], [3], ion implant etching, focused ion beam etching [4], and direct optical machining by laser light or synchrotron radiation. Nanomachining has already been implemented by electron beam lithography, optical lithography, and focused ion beam etching.

Compared to these technologies, direct optical machining is an important technique that offers high throughput in a single process (and hence requires only simple equipment). However, direct optical machining using laser light is complicated by the transparency of the workpieces. Specifically, single‐photon absorption does not occur, and hence the optical energy cannot be applied efficiently to the workpiece. Thus, previous research was focused on alternative methods of inducing absorption, or on methods using femtosecond lasers. The alternative methods of inducing absorption include irradiation by vacuum UV light [5] or soft X‐rays [6], contact with a laser plasma [7], and the application of solutions to the workpiece [8].

When transparent materials are irradiated by femtosecond laser light, machining is made possible by multiphoton absorption, or by a strong electric field produced by a femtosecond laser [9–11]. When a transparent material is exposed to femtosecond laser light, repetitive irradiation of the same spot results in the formation of an absorption layer (incubation layer) due to defect generation, and this layer provides efficient absorption in further irradiation.

It is difficult to obtain arbitrary nanoscale structures by direct optical machining of inorganic transparent materials using existing lasers, and hence new processes are required. Nanoscale machining is often restricted by the diffraction barrier which depends on the wavelength of the light. Therefore, soft X‐rays in the vicinity of 10 nm are needed. Furthermore, one can expect a relatively wide range of machinable materials because processing is possible so long as light is absorbed. Using synchrotron radiation as a source of soft X‐rays makes possible the etching of Si and SiO₂ [12–22] and the chemical vapor deposition of SiO₂ [23]. However, in previous studies using synchrotron radiation, direct etching at room temperature proved impossible, and it was necessary either to heat the workpiece [14-16, 19, 21] or to use reactive gases [12, 17, 18, 20, 22]. In the former case, the etching rate improves as the temperature is raised to about 700 °C [16, 19, 21]. Thus, etching of SiO₂ and Si is possible using synchrotron radiation. In practice, however, there are problems of high photon cost, and the need for high‐temperature heating or reactive gases.

To develop a practical method for micro‐ and nanomachining using soft X‐rays, we considered a nanosecond laser plasma as a source of soft X‐rays. In addition, we used an optical system with efficient focusing of soft X‐rays in the vicinity of 10 nm. Thus, we implemented ablation of transparent organic materials using only soft X‐rays at room temperature, without any reactive gas [24], [25]. In particular, we investigated the micromachining of silica glass, and found that ablation was possible at 0.2 to 150 nm/shot by means of fluence control. Typically, a surface roughness below 1 nm is obtained when ablation is implemented by 10 irradiation shots at a rate of 50 nm/shot, which promises very smooth machining. Furthermore, we fabricated 50‐nm‐wide trenches using a 50‐nm‐wide mask, thus demonstrating the feasibility of nanomachining. Based on reported data on the thermal diffusion coefficient at room temperature, the thermal diffusion length is about 80 nm. It should be noted that finer structures can be fabricated.

In this study, we investigated the ablation process of silica glass using laser plasma soft X‐rays in terms of its application to nanomachining. First, we estimated the energy of soft X‐rays per unit area (fluence), and determined its relation to the ablation depth. Then we investigated the depth at which the energy of soft X‐rays is absorbed, and the energy dissipation. In addition, we identified the emitted species and measured their kinetic energy. Based on these results, we discussed the process of ablation.

2. Irradiation System for Pulsed Laser Plasma Soft X‐rays

An irradiation system using laser plasma soft X‐rays is shown in Fig. 1. Soft X‐rays (X) are emitted by a plasma generated when the target (T) is irradiated by pulsed laser light (L). In this study, a Ta target was irradiated by a pulsed Nd:YAG laser (wavelength: 532 nm, per pulse energy: 700 mJ/pulse, pulse width: 7 ns, repetition frequency: 10 Hz) in a vacuum chamber at 10^–4 Pa. These soft X‐rays were focused on silica glass (S) using an elliptic mirror (M).

Figure 1

Schematic diagram of micromachining apparatus by laser plasma soft X‐ray technique.

This focused irradiation was achieved by using an elliptic mirror made of silica glass and coated with gold. The mirror was designed so that the intensity of soft X‐rays on the workpiece surface reached its maximum at a wavelength of about 10 nm. The intensity of soft X‐rays depended on (a) the solid angle at which the mirror was seen from the light source, and (b) the grazing angle at which soft X‐rays were incident on the mirror. As the grazing angle increases, the solid angle too increases, and the total intensity of the incident soft X‐rays increases, while the reflection factor decreases. It was shown that when the angle of incidence with respect to the mirror is 200 mrad, soft X‐rays with a wavelength of about 10 nm (about 100 eV) can be focused efficiently on the workpiece. However, soft X‐rays with energies of about 200 eV and higher cannot be reflected by the mirror to irradiate the workpiece surface. The fluence of soft X‐rays was adjusted by shifting the sample (S) from the focal position of the elliptic mirror. The fluence was estimated using the reported efficiency of conversion from 532‐nm Nd:YAG laser light to soft X‐rays [26], the focusing efficiency of the mirror (0.006) [24], and the beam cross‐section area at the sample, which location was found geometrically from the positions of the sample and the elliptic error. We assumed that soft X‐rays with energies of 100 ±20 eV contribute to ablation. The details of fluence estimation are given in another paper [24].

Neutral and charged particles emitted from the surface of silica glass due to irradiation by soft X‐rays were detected, respectively, by a quadrupole mass spectrometer Q and an electrode E. Thus, we measured the mass of neutral particles as well as the kinetic energy and mass of charged particles. However, the temporal resolution of the detection system for neutral particles was not sufficient, and we only identified neutral species.

When a Ta target is irradiated by laser light to obtain soft X‐rays, plasma generation is accompanied by fast electrons, fast atoms, fast ions, and droplets about 1 μm in size emitted from the target. Thus, a shield was installed to prevent these particles from reaching the sample surface and detection system.

3. Surface of Silica Glass After Ablation

Figure 2 shows a confocal laser microphotograph of silica glass irradiated by soft X‐rays. Here a contact mask was used to confirm the effect of irradiation by soft X‐rays. The contact mask was implemented as a grid of square apertures in Ni thin film, and was applied to the sample surface during irradiation by soft X‐rays. The surface ablation depth was 470 nm in 10 irradiation shots. A surface roughness of 1 nm was obtained, which is a significant result in terms of machining performance.

Figure 2

Microstructures on silica glass fabricated by laser plasma soft X‐ray technique.

The ablation depth can be varied in the range of 0.2 nm/shot to 150 nm/shot by fluence control. There is a threshold of F_th = 60 mJ/cm², and ablation occurs only at higher fluences. Straight‐line fitting can be obtained in a single logarithmic plot, and the ablation depth D is related to the fluence F as follows:

Here α can be interpreted as the effective absorption coefficient, and F_th as the fluence threshold. This result suggests that ablation begins before the energy of the absorbed soft X‐rays diffuses beyond the absorption area. We may assume that when soft X‐rays are absorbed by the silica glass surface, this energy diffusion from the absorption area proceeds by thermal diffusion. As explained in the introductory section, the thermal diffusion length corresponding to the pulse duration of soft X‐rays is estimated as 80 nm. On the other hand, fine trenches 50 nm wide, that is, smaller than the thermal diffusion length, were obtained. This would be hardly possible if melting occurred within the pulse duration. Therefore, we may infer that a nonthermal process is predominant. The effective absorption coefficient in the above equation can be estimated as α = 9 × 10⁵ cm^–1. Since 1/α = 10 nm, machining is possible with a depth resolution of 10 nm. These results indicate that the energy density ϵ of soft X‐rays absorbed per unit area is ϵ = F_th · α = 55 kJ/cm³. This is close to the energy required in order to decompose silica glass into atoms: SiO₂ + 77 kJ/cm³ → Si(gas) + 2O (gas). Thus, we may conclude that irradiation by soft X‐rays above the threshold F_th provides sufficient energy to decompose the surface of silica glass into atoms. It has already been reported that the ablation depth varies linearly with the number of irradiation passes using soft X‐rays [24], which also confirms the proposed model of the ablation process. That is, irradiation by soft X‐rays produces direct ablation in one shot. On the other hand, we may hardly assume that soft X‐rays are absorbed in an affected surface layer produced by multiple irradiation, and that ablation occurs due to energy diffusion into this affected layer. Furthermore, ablation is not likely to be related to surface contamination or surface polish.

Previous research using synchrotron radiation considered only the etching region where the depth grows directly with the fluence. But the efficient ablation process investigated in this study occurs in the region with high energy of soft X‐rays per unit time and unit area.

4. Ablation Particles

Figure 3 presents results observed with the quadrupole mass spectrometer for particles ablated when silica glass was irradiated by laser plasma soft X‐rays. In order to detect neutral particles, they were ionized by electrons emitted from a filament; however, the neutral particles were detected even without such ionization. Therefore, the fast components observed within 1 ms after irradiation by soft X‐rays are ions. The components observed after that are neutral particles. Here ²⁸Si atoms are difficult to distinguish from ¹⁴N₂ molecules (residual gas), and hence we measured ²⁹Si atoms and applied a normalization based on the natural abundance so as to maintain the total amount of Si natural atoms. Thus, we detected Si atoms, SiO molecules, O atoms, and Si₂ molecules as the main neutral particles. Larger neutral molecules such as Si₃ were more difficult to detect. These results show that the neutral particles produced by irradiation of silica glass by soft X‐rays were mostly decomposed into atoms.

Figure 3

Neutrals ejected from silica surface by focused laser plasma soft X‐rays. (a) Si, (b) SiO, (c) O, and (d) Si₂ neutrals are dominant among neutral species.

Figure 4 shows the ion current measured at 300 mm from silica glass irradiated by laser plasma soft X‐rays at a fluence of about 1 J/cm². The peak time of the signal observed at 10 to 40 μs varies directly with the measurement distance in the range of 50 to 300 mm. From this it follows that the average ion speed is v̄ = 1.4 × 10⁶ cm/s. This result confirms that the signal represents ions emitted by the sample after irradiation by soft X‐rays. Conversely, the signal does not represent ions emitted by the Ta target during the generation of soft X‐rays, or to ions arriving at the electrode after collision with the chamber wall.

Figure 4

Ion current measured at 300 mm from silica surface after irradiation with pulsed laser plasma soft X‐rays.

In order to estimate the shares of ions and neutral particles among the emitted particles, we used a confocal laser microscope to measure the volume ablated from the silica glass surface after irradiation by soft X‐rays. In addition, we found the number of emitted ions as a function of the electrode solid angle, assuming that ions are uniformly emitted from the silica glass surface in a hemisphere. As a result, we found that 10 ±5% of all the particles were ions, and the rest were neutral particles.

We also measured the voltage on a grid installed in front of the electrode used for detection of the ion current, as shown in Fig. 1, so as to identify the ion species. Ions with energy greater than applied voltage V can be detected selectively. In addition, ion species can be identified from the relation V = 1/2mv̄². Thus, we determined Si⁺ ions with a kinetic energy of 25 eV and O⁺ ions with a kinetic energy of 13 eV as the main components.

Considering the emission of such ions with high kinetic energy, we may infer that ablation is driven by repulsive forces between ions generated due to irradiation by soft X‐rays. That is, irradiation by soft X‐rays weakens chemical bonds, ions are emitted together with neutral particles, and ablation occurs because of Coulomb explosion.

5. Conclusions

We have investigated the ablation process occurring in silica glass due to focused irradiation by laser plasma soft X‐rays. In such irradiation, ions make up to 10 ±5% of all emitted particles, and the rest are neutral particles. We found that both the emitted ions and neutral particles are mostly decomposed into atoms. In particular, the emitted ions have a greater kinetic energy than the energy of evaporation. In addition, the relation between the ablation depth D and the soft X‐ray fluence F can be determined from the effective absorption coefficient α and the ablation threshold F_th, namely, D = –(1/α)ln(F/F_th). This means that ablation starts before the absorbed energy has diffused. We also showed that when soft X‐rays at the fluence threshold are absorbed with an effective depth of 1/α = 10 nm from the surface, the absorbed energy F_th·α is close to that required for atomic decomposition of silica glass. We assumed that, in addition to this energy, repulsive forces between emitted ions contribute to ablation. The above results make it clear that when silica glass is irradiated with laser plasma soft X‐rays, the rays are absorbed effectively within about 10 nm of the surface, and energy dissipation from the absorption area can be neglected. Thus, ablation processing of silica glass using laser plasma soft X‐rays offers advantageous characteristics in terms of practical nanomachining.

Acknowledgements

Professor K. Asakawa of the University of Tsukuba helped us with the experimental facilities. This study was subsidized in part by the NEDO Industrial Technology Research Grant Program 2001–2005, MEXT Grants‐in‐Aid for Scientific Research (15656183, 17360348), and by the Nippon Sheet Glass Foundation for Materials Science and Engineering. We take this occasion to express our gratitude.

AUTHORS (from left to right)

Tetsuya Makimura (member) received a bachelor's degree from Nagoya University (physics) in 1989, completed the doctoral program in 1994, and became a research associate at the University of Tsukuba. He is now a lecturer. His research interests are nanomachining of inorganic transparent materials using laser plasma soft X‐rays and light emission of Si‐based light‐emitting materials using laser ablation. He holds a D.Sc. degree, and is a member of JSAP and LSJ.

Takashige Fujimori (nonmember) received a bachelor's degree from the University of Tsukuba (fundamental engineering) in 2006 and is now in the doctoral program at the Graduate School of Pure and Applied Science. His research interest is the ablation mechanism of silica glass using laser plasma soft X‐rays.

Shuichi Torii (nonmember) (photograph not available) is enrolled in the first half of the doctoral program in electronic engineering and applied physics in the Graduate School of Pure and Applied Sciences, University of Tsukuba. His research focuses on the mechanism of ablation of silica glass by laser plasma soft X‐rays.

Hiroyuki Niino (member) received a bachelor's degree from Kyushu University (applied chemistry) in 1986. Since 1987 he has been a researcher at the National Institute of Advanced Industrial Science and Technology. His research interests are laser photochemistry and laser ablation. He holds a D.Eng. degree, and is a member of LSJ, JSAP, and CSJ.

Kouichi Murakami (nonmember) received a bachelor's degree from Osaka University (electrical engineering) in 1970 and completed the M.E. and doctoral programs. After serving as a researcher at the Japan Technology Foundation and RIKEN, a research associate at Osaka University, a lecturer at the University of Tsukuba (1980), and then an associate professor, he has been a professor since 1999. He is the leader of the Nanoscience Project, involved in implementation of the 21st Century COE Program. His present research interests are semiconductor engineering and nanoscience.