Open Access
Add to BibSonomyAbstract
Techniques to control the colors and properties of glasses based on doping of the glasses with various metals and nanoparticles are widely used. In this paper, we demonstrate the migration of a nickel sphere in silica glass caused by laser illumination accompanied by nickel nanoparticle precipitation in the sphere migration trajectory. During migration, the diameter of the nickel sphere decreased. Precipitated nanoparticles with diameters of several hundred nanometers were observed in areas of up to 50 μm in radius and these nanoparticles formed four cylindrical coaxial layers with stripes at 10–20 μm intervals in the migration direction.
© 2015 Optical Society of America
1. Introduction
Techniques to control the colors and properties of glasses based on doping of the glasses with various metals and nanoparticles are widely used. Dopants are usually added to color entire glass sheets during glass fusion. Selective doping is also possible, e.g., by dipping glasses in molten salts [1] or by applying an electric field [2,3] to enable ion exchange around the surface of the glass only, and can be applied to waveguide forming [1] or strengthening of the glass [4]. These methods are commonly used to strengthen smartphone displays. Femtosecond laser illumination changes the valences of ions [5] or precipitates nanoparticles [6,7] around the focus of the beam. However, ion exchange processes can only perform doping from the surface, and dopants must be doped in advance when using a femtosecond laser. We previously reported that laser illumination caused metal sphere migration in glass [8,9]. The migration mechanism can be explained as follows: laser illumination heats and melts the metal spheres, and heat radiation and conduction from the heated sphere softens the surrounding glass, which then enables sphere migration. The surface tension of the glass decreases with increasing temperature. The surface tension is therefore relatively lower on the laser illuminated side and migration toward the light source occurs.
In this paper, we demonstrate the migration of a nickel sphere (with a diameter of a few tens of micrometers) caused by laser illumination accompanied by nickel nanoparticle precipitation in the sphere migration trajectory. Precipitated nanoparticles with diameters of several hundred nanometers were observed in areas with a radius of up to 50 µm and these particles formed four cylindrical coaxial layers with stripes at 10–20 µm intervals in the migration direction.
2. Experimental
Glass plates and nickel foils were stacked and locked together using a jig, which was then placed on an X-Y-Z stage, as shown in Fig. 1.The glass plates were stacked in the following order: silica glass with a thickness of 10 mm (AQ series, AGC Asahi Glass Co., Ltd., Tokyo, Japan), borosilicate glass with a thickness of 5 mm (Pyrex, Corning 7440, Corning Inc., Corning, NY, USA), nickel foil with a thickness of 10 µm (Purity 99 + %, NI-313173, Nilaco Corp., Tokyo, Japan), and silica glass with a thickness of 10 mm to ensure good contact between the borosilicate glass and the nickel foil. The description of the experimental apparatus is similar to that given in our previous reports [8,9]. However, some of the equipment was changed, and it is described briefly here. A continuous wave laser beam (fibre laser, RFL-C020/A/2/A, WuhHan Raycus Fiber Laser Technologies Co., Ltd., Hubei, China) was used to illuminate the nickel film through the silica glass and the borosilicate glass. The laser beam was then focused using a convex lens with a focal length of 40 mm (NYTL-30-40PY1, Sigma Koki Co., Ltd., Saitama, Japan). Side-view shadowgraph images under white-light illumination were obtained for in situ process monitoring. A band-pass filter (10BPF10-440, Newport Corp., CA, USA) was placed in front of the charge-coupled device (CCD) camera to prevent detection of scattered laser light and thermal emissions.
When the laser focus was adjusted on the light propagation side of the sphere, as shown in Fig. 2(a), i.e., when the sphere was located between the focus and the light source, the sphere stopped after moving to a certain distance. This was because, as the sphere migrates, defocusing causes the laser power density to decrease. To observe this behavior with a CCD camera while avoiding moving out of the CCD’s field of view, the laser power was adjusted to stop migration on the light source side of the CCD view. When the sphere moved out of the focus and stopped (Fig. 2(b)), then the laser illumination stopped. This migration process is called a “single process” in this paper. The stage was then moved by ~150 µm in the light propagation direction and the sphere was located on the focus side of the CCD view (Fig. 2(c)). The laser illumination then restarted.
For analysis of the implanted sphere, the glass was cut and then polished to intercept the sphere. The polished samples were observed using a transmission optical microscope (ECLIPSE ME600, Nikon Corp., Tokyo, Japan) and a confocal microscope (OLS4000, Olympus Corp., Tokyo, Japan; laser wavelength of 405 nm, spatial resolution of ~1.3 μm). After observation using these optical microscopes, the surface was given a carbon coating to provide electrical conductivity and the surface was observed with an energy-dispersive X-ray spectroscopy system (EDX, JED-2300, JEOL Ltd., Tokyo, Japan) and an electron probe microanalyser (EPMA, JXA-8230, JEOL Ltd.)
3. Results and discussion
After laser illumination, a nickel sphere with a diameter of ~40 µm was observed in the borosilicate glass. Laser illumination with adequate laser power density moved the sphere backwards (toward the light source), in the same manner as reported in our previous work [8,9]. When the sphere migration stopped, the illumination was stopped (a single process). Then, the focal position of the laser beam was changed and the illumination was restarted to induce migration of the sphere again. The migration of the sphere in a single process is shown in Media 1 and photographs of the process are presented in Fig. 3.During illumination, the laser focus did not scan or produce any relative motion. By repeating the process, the nickel sphere migrated through the borosilicate glass and reached the interface with the silica glass. The nickel sphere migrated into the silica glass after briefly stopping at the interface. Then, the sphere continued to move in the same manner as it had in the borosilicate glass.
Fig. 3 Time-lapse photographs taken during nickel sphere migration in borosilicate glass (see Media 1; playback speed is twice as fast as actual speed). The images here were taken (a) 24 s (12 s in Media 1) and (b) 54 s (27 s in Media 1) after the laser illumination. The laser power was 2.85 W, and the spot diameter was 50–60 µm.
The nickel sphere also migrated in the silica glass, and the migration of this sphere in a single process is shown in Media 2, while photographs of the migration are shown in Fig. 4.Unlike the migration of the sphere in borosilicate glass, stripes were observed as the sphere migrated. The lines in Fig. 4 were drawn to indicate the progress of the migration. The right line was drawn at the left side of the sphere in Fig. 4(a), and the left and center lines are located at the stripes. After laser illumination began (~0.5 s in Media 2), a white emission was observed from the nickel sphere, while the area around the sphere itself became black. Then, the sphere moved towards the light source (~1.5 s in Media 2), the emission became weaker and the migration of the black sphere was observed (~2 s in Media 2). Then, the sphere migrated towards the light source while forming black arc-shaped stripes at ~15 µm intervals located ~30 µm in front of the sphere. These arc-shaped stripes are formed by multiple particles. The nickel sphere migrated towards the light source by passing through the arc-shaped stripes, and the centers of the stripes were shifted backwards such that their shapes changed into wave forms. In the movie, the black particles moved around the sphere because of the flow of the glass. The speed of migration gradually decreased and then finally stopped (~25 s in Media 2).
Fig. 4 Time-lapse photographs taken during nickel sphere migration in silica glass (see Media 2; playback speed is twice as fast as actual speed). The images were taken (a) 16 s (8 s in Media 2), (b) 18 s (9 s in Media 2), (c) 20 s (10 s in Media 2), (d) 30 s (15 s in Media 2) after laser illumination. The laser power was 7.4 W, and the spot diameter was 160–170 µm.
The intervals between the stripes were 10–20 µm in size, and no relationship with the migration speed was observed. However, the density of the stripe depends on its location. The stripe is thicker at the stage where the sphere is accelerating than at the stage where the sphere is at a steady speed or is decelerating.
The sphere diameter of ~28 µm decreased to ~14 µm after migration for a distance of ~350 µm (Fig. 5). The sphere became smaller as the migration progressed, and finally could not migrate because laser absorption by the glass occurred in a similar manner to the phenomenon that was described in the literature [10,11]. The phenomenon is explained as follows: only the heated spot of the glass (and not the metal sphere) moved toward the light source as a result of thermal radiation and conduction, because the absorption of the glass was enhanced by the laser heating. Because the power density required for migration increased as the sphere size decreased, the power density subsequently became larger than the laser absorption threshold of the glass. In the borosilicate glass, no stripes or particles were observed, as shown in Fig. 3, and no size difference was observed, even after migration of 1,800 µm.
Fig. 5 Ni sphere size. Micrographs are shown of the Ni sphere with a diameter of 28 μm (a) that decreased in size as migration progressed in the silica glass, and the sphere reached a final diameter of 14 μm (b) after migration of 350 μm. (c) Ni sphere diameter at various migration distances.
For analysis of the stripes, the glass was cut and polished to expose the sphere (see the Experimental section). The polished surface was then coated with carbon and was observed by scanning electron microscopy with energy-dispersive X-ray spectrometry (SEM-EDX). Figure 6 shows the SEM micrographs and the X-ray maps. Figure 6(a) shows a backscattered electron composition image of the sphere in the borosilicate glass. A sphere with a diameter of ~35 µm was observed and no contrast was observed around this sphere. X-ray maps from the same area are also shown in Fig. 6. In the borosilicate glass, only nickel was detected from the sphere. In the area of the glass, silicon was detected but no nickel was observed. These mapping results indicate that the nickel sphere and the borosilicate glass have a clear boundary.
Fig. 6 Scanning electron micrographs (a)–(e), and X-ray maps (f)–(k) of the cross-sections of the sphere: the Ni sphere is shown in borosilicate glass (a) and in silica glass (b)–(e), while (c) shows a magnified image at the sphere, and (d) and (e) show the particles that were observed around the areas marked in (b). X-ray maps are shown of (f) Si-Kα and (g) Ni-Lα in borosilicate glass and of (h) Si-Kα and (i) Ni-Lα in silica glass.
In the silica glass, a sphere with a diameter of ~12 µm and particles were observed in the area where the stripes were observed with the optical microscope. Figure 6(b) shows the complete images around the sphere, Fig. 6(c) shows a magnified image of the sphere, and Fig. 6(d) and Fig. 6(e) show the particles that were observed in the area indicated in Fig. 6(b). These particles were a few hundred nm in diameter. From the sphere, nickel was observed and no silicon was detected, in a similar manner to the borosilicate glass case. Nickel was detected where the particles were observed. However, nickel was not detected by EDX with sensitivity of ~1 wt.% in the glass where the sphere passed through. Nickel particles were detected in areas away from the sphere; thus, nickel appeared to have diffused into the silica glass. More sensitive electron microprobe analysis measurements with sensitivity of 0.1 wt.% were therefore performed ~3 µm away from the sphere; however, no nickel was detected.
Figure 7 shows optical micrographs of the same samples that were observed in Fig. 6 before the carbon coating process. The nickel spheres and the particles were observed to be black during transmission optical microscopy (Fig. 7(a), (b)); however, they were observed to be white during confocal microscopy (Fig. 7(c)–(g)). Figure 7(e)–(g) present confocal micrographs that were focused within the glass at depths of (e) ~15 µm, (f) ~30 µm and (g) ~37 µm below the surface, which are determined based on the real depth multiplied by the refractive index (which is 1.47 at 400 nm) [12]. In Fig. 7(d)–(g), the white dots shown are the nickel particles with diameters of a few hundred nanometers that were observed in Fig. 6.
Fig. 7 Particle distributions in silica glass. (a,) (b) Micrographs of the metal sphere after polishing under transmission illumination, and (c)–(g) reflection images taken with the confocal microscope. A sphere is shown in borosilicate glass in (a) and (c) and in silica glass in (b) and (d)–(g). The images are focused at the surface in (c) and (d), and at depths of (e) 15 µm, (f) 30 µm and (g) 37 µm under the surface. An illustration of the nickel particle distribution in silica glass is shown in (h).
When compared with Fig. 6, many more particles were observed in Fig. 7. This was because the particles that were exposed at the surface and those that were exposed at the focal depth were all observed using the confocal microscope; however, only the particles that were at the surface were detected by SEM-EDX.
These observations are explained as follows. At the surface (Fig. 7(d)), particles were observed at the center up to a width of ~10 µm (zone A), and no particles were observed (zone B) in the adjacent zone outside zone A. Large particles were seen 20–35 μm from the center (zone C), and smaller particles were recognized in the zone 35–50 µm from the center (zone D). At planes located ~15 μm below the surface (Fig. 7(e)), no particles were observed at the center at a width of ~30 μm (zone B), but larger particles and smaller particles were observed at widths of ~20 μm (zone C) and ~15 μm (zone D), respectively. At the planes ~30 μm below the surface (Fig. 7(f)), large particles were observed at widths of up to ~45 μm (zone C) and stripes were also observed. Outside zone C, small particles were observed up to a width of ~20 μm (zone D). At the planes ~37 μm below the surface (Fig. 7(g)), small particles were observed at widths of up to ~70 μm (zone D). In Fig. 7(d), fewer particles were observed in zone D when compared with the other figures, because the brightness was set to be low because of the surface reflection properties of the glass. Therefore, any particles that were present in zone D could not be observed. The surface reflection caused the right side of Fig. 7(d) to appear to be white. The existence of small particles in zone D in Fig. 7(d) was confirmed by enhancement of the gain.
The process where the nickel sphere migration caused nickel particle precipitation must also be considered. Glass coloration using metal colloids, e.g., gold, silver and copper, is a commonly performed process. The formation processes of these colloids are as follows [13]: 1. Metal ion dissolution in glass from the metal compound; 2. Reduction from metal ions to nonvalent metals; and 3. Nucleation and growth. Metal ion dissolution occurs under high temperature conditions, and the reduction, nucleation, and growth processes occur around the temperatures of the glass transition point and the glass softening point. Notably, two nickel spheres would become a single sphere if they made contact. Therefore, the sphere temperature must be higher than the melting point of nickel (1,728 K).
The solubility and diffusivity of nickel in silica glass at temperatures higher than 1,700 K are important properties for the discussion of this process. Ghoshtagore [14] measured diffusivity in amorphous SiO2 deposited by chemical vapor deposition and reported that the nickel diffusion was Fickian, that nickel diffusivity obeys the Arrhenius equation, and that the diffusivity of nickel below 1,200°C was low. Mayer et al. [15] reported that nickel diffusion through 1.6-nm-thick SiO2 occurred in the 700–1,050 K temperature range. In our experiments, the nickel temperature was higher than 1,728 K, and thus was certainly high enough for diffusion.
There have been many reports on the solubility of nickel in silicate glass. For example, Borisov [16] reported nickel solubility in silicate melts with various SiO2 contents in the 1,300 to 1,430°C temperature range, and the NiO concentration in the silicate melt ranged from a few ppm to 13 wt.%. However, to the best of the authors’ knowledge, there have been no reports to date on nickel solubility in SiO2. Amekura et al. [17] reported the implantation of nickel ions in amorphous SiO2 and nickel nanoparticle formation in SiO2. After ion implantation with an acceleration energy of 60 keV followed by annealing at 800°C, nickel was observed in a ~100-nm-thick surface layer. The total nickel content in the SiO2 was evaluated by Rutherford backscattering spectrometry and the amount of nickel that was forming metallic nanoparticles was estimated using optical absorption. The total nickel atom content was 5.2 × 1016 atoms/cm2, while that of the metallic nanoparticles was ~4.9 × 1016 atoms/cm2 [18]. Here, we hypothesized that the difference between the total nickel atom content and that of the metallic nickel nanoparticles, which was 0.3 × 1016 atoms/cm2, was caused by dissolution in the silica glass. The weight of SiO2 was calculated to be 22 µg in an area of 1 cm2 with a thickness of 100 nm, by assuming that the SiO2 density was 2.2 g/cm2. A quantity of 3 × 1016 nickel atoms, i.e., 0.29 µg of nickel, was dissolved in the layer, and the nickel concentration was calculated to be 1.3 wt.%. The difference between the total number of nickel atoms and that of the nickel nanoparticles decreased with increasing annealing temperature. The temperature was more than 1,728 K in our experiment, and the solubility in our experiment was therefore considered to be higher than 1.3 wt.%.
As mentioned earlier, nickel solubility in silicates is a minimum of a few ppm and the concentration of ~1.3 wt.% was estimated from the reports of Amekura et al. [17, 18] In our results, nickel was not detected by electron probe microanalysis with a sensitivity of 0.1 wt.%. However, nickel diffusion and solubility in silica glass have proven to be indispensable when explaining this phenomenon. To describe this phenomenon precisely, the nickel in the silica glass must be analyzed with high sensitivity in situ under high temperature conditions of more than 1,728 K, and this analysis is a challenging task. In addition, the softened SiO2 flowed around the sphere. This effect must also be considered as part of the analysis of this phenomenon, and thus many fundamental data are required.
Amekura et al. [19] reported nickel nanoparticle oxidation during the annealing of nickel ion-implanted amorphous SiO2. They reported that 85% of the Ni atoms form metallic nanoparticles in their as-implanted state. After annealing at 800°C in a vacuum, this ratio increased to 90%. In contrast, after annealing in oxygen, the metallic nickel was oxidized to produce NiO. Our experiment was conducted in an air atmosphere, but this diffusion occurred in the glass and could not be affected by the oxygen in the atmosphere. The particles were therefore metallic nickel.
The process of nickel diffusion from the sphere and particle formation was considered as follows: nickel was diffused into the silica glass around the laser-illuminated sphere at a temperature above the melting point of nickel, and nickel ions or nonvalent nickel diffused in the silica glass. When, apart from the sphere, the temperature decreased [9], the nickel was precipitated, and particles with diameters of several hundred nanometers were grown. These particles did not grow any larger than several hundred nanometers, because the particles were heated by the laser illumination and began to dissolve into the glass, until precipitation and dissolution were finally balanced at a diameter of several hundred nanometers. The nickel particles did not migrate backwards like the nickel spheres. This was considered to be because the small particles were not heated sufficiently to migrate towards the light source. Nickel solubility and diffusivity in silica glass at this temperature are important properties for the discussion of this process; however, these properties have not yet been clarified sufficiently.
4. Summary
We discovered a phenomenon where nickel particles with diameters of several hundred nanometers were precipitated by nickel sphere migration in silica glass. The observed particles had radii of ~50 µm and formed four cylindrical coaxial layers with stripes at intervals of 10–20 µm in the migration direction. Many mechanisms remain to be clarified in this process, including the precipitation and stripe formation mechanisms. No nickel dopants were detected from the silica glass.
Acknowledgments
The authors gratefully acknowledge T. Kishi and S. Itoh for their helpful comments. The support of the Japan Society for the Promotion of Science through a Grant-in-Aid for Scientific Research (no. 24656096) is also gratefully acknowledged.
References and links
1. R. V. Ramaswamy and R. Srivastava, “Ion-exchanged glass waveguides: a review,” J. Lightwave Technol. 6(6), 984–1000 (1988). [CrossRef]
2. S. Matsusaka, T. Kobayakawa, H. Hidai, and N. Morita, “Laser micro-machinability of borosilicate glass surface-modified by electric field-assisted ion-exchange method,” J. Phys. Conf. Ser. 379, 012035 (2012). [CrossRef]
3. S. Matsusaka, M. Naotomo, H. Hidai, A. Chiba, N. Morita, and T. Itoi, “Formation of a buried silver nanowire network in borosilicate glass by solid-state ion exchange assisted with forward and reverse electric fields,” Appl. Phys. Lett. 105(10), 103102 (2014). [CrossRef]
4. T. Haranoh, H. Ishikawa, N. Shinkai, and M. Mizuhashi, “Crack evolution in Vickers indentation for soda-lime-silica glass,” J. Mater. Sci. 17(5), 1493–1500 (1982). [CrossRef]
5. J. Qiu, K. Kojima, K. Miura, T. Mitsuyu, and K. Hirao, “Infrared femtosecond laser pulse induced permanent reduction of Eu3+ to Eu2+ in a fluorozirconate glass,” Opt. Lett. 24(11), 786–788 (1999). [CrossRef] [PubMed]
6. J. Qiu, X. Jiang, C. Zhu, H. Inouye, J. Si, and K. Hirao, “Optical properties of structurally modified glasses doped with gold ions,” Opt. Lett. 29(4), 370–372 (2004). [CrossRef] [PubMed]
7. X. W. Jiang, J. R. Qiu, H. D. Zeng, C. S. Zhu, and K. Hirao, “Laser-controlled dissolution of gold nanoparticles in glass,” Chem. Phys. Lett. 391(1-3), 91–94 (2004). [CrossRef]
8. H. Hidai, T. Yamazaki, S. Itoh, K. Hiromatsu, and H. Tokura, “Metal particle manipulation by laser irradiation in borosilicate glass,” Opt. Express 18(19), 20313–20320 (2010). [CrossRef] [PubMed]
9. H. Hidai, M. Matsushita, S. Matsusaka, A. Chiba, and N. Morita, “Moving force of metal particle migration induced by laser irradiation in borosilicate glass,” Opt. Express 21(16), 18955–18962 (2013). [CrossRef] [PubMed]
10. H. Hidai, M. Yoshioka, K. Hiromatsu, and H. Tokura, “Glass modification by continuous-wave laser backside irradiation (CW-LBI),” Appl. Phys., A Mater. Sci. Process. 96(4), 869–872 (2009). [CrossRef]
11. H. Hidai, M. Yoshioka, K. Hiromatsu, and H. Tokura, “Structural changes in silica glass by continuous-wave laser backside irradiation,” J. Am. Ceram. Soc. 93, 1597–1601 (2010).
12. W. S. Rodney and R. J. Spindler, “Index of refraction of fused quartz glass for ultraviolet, visible, and infrared wavelengths,” J. Opt. Soc. Am. 44(9), 677–678 (1954). [CrossRef]
13. C. Louis and O. Pluchery, Gold Nanoparticles for Physics, Chemistry and Biology (Imperial College Press, 2012).
14. R. N. Ghoshtagore, “Diffusion of nickel in amorphous silicon dioxide and silicon nitride films,” J. Appl. Phys. 40(11), 4374–4376 (1969). [CrossRef]
15. J. T. Mayer, R. F. Lin, and E. Garfunkel, “Surface and bulk diffusion of adsorbed nickel on ultrathin thermally grown silicon dioxide,” Surf. Sci. 265(1-3), 102–110 (1992). [CrossRef]
16. A. A. Borisov, “Experimental study of the effect of SiO2 on Ni solubility in silicate melts,” Petrology 14(6), 530–539 (2006). [CrossRef]
17. H. Amekura, N. Umeda, Y. Takeda, J. Lu, K. Kono, and N. Kishimoto, “Formation processes of nickel oxide nanoparticles in SiO2 by metal-ion implantation combined with thermal oxidation,” Nuclear Instrum. Methods B 230(1-4), 193–197 (2005). [CrossRef]
18. H. Amekura, H. Kitazawa, and N. Kishimoto, “Non-magnetic to magnetic and non-metal to metal transitions in nickel nanoparticles in SiO2 under heat treatment,” Nuclear Instrum. Methods B 219–220, 825–829 (2004). [CrossRef]
19. H. Amekura, H. Kitazawa, N. Umeda, Y. Takeda, and N. Kishimoto, “Nickel nanoparticles in silica glass fabricated by 60 keV negative-ion implantation,” Nuclear Instrum. Methods B 222(1-2), 114–122 (2004). [CrossRef]
