High-intensity laser-irradiated metal spheres in glass move toward a light source while leaving the doping metal in their trajectories. A method for controlling the trajectory length, which can be used to produce new optical devices in glass, has not been proposed yet. In-situ observations clarified the relationship, wherein the trajectory length increased with the increasing laser power and irradiation duration; the maximum and minimum being 2.0 and 0.1 mm, respectively. Microscopic observations, elemental analysis, and counting the number of metal particles revealed that the maximum speed metal sphere generated the most metal-containing area with the highest number of metal particles.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Glass is a highly optically transparent, durable, and heat resistant material. Owing to these properties, it is used for manufacturing many products, such as electronic equipment displays , optical communication fibers , and gamma-radiation shielding materials containing a significant amount of lead and bismuth . Glass containing nanoparticles of a metal (Au, Ag, and Cu) exhibit various optical effects. For example, the spectra of Au particles depend on the particle size, deviation from the spherical shape, and volume fraction. The center absorbance spectra shifted from 515 to 575 nm by changing the sphere diameter from 9 to 99 nm. This phenomenon—known as the surface plasmon effect—occurs when the electrons in nanoparticles absorb a certain wavelength resonating with light . Both erbium-doped glass and neodymium-doped glass can enhance the specific wavelength light via stimulated emission. An optical amplifier that amplifies the light attenuated by propagation  and a glass laser that emits light of a specific wavelength strongly  are the known applications. Glass containing crystallites of the mixed semiconductor CdSxSe1-x is a material used for long-wavelength-pass filters; it changes its color from yellow to deep red depending on the values of composition, x . Besides these, optical memory with chalcogenide glass  and optical switching by applying a nonlinear effect  have been reported. First mixing fine metal particles into the entire volume of glass is the conventional method to produce metal-doped optical glass devices. Adding optical effects only in a micro area or adding two or more metals is difficult; thus, creating different optical effects in glass is also challenging.
As a solution for this problem, we propose continuous wave laser backside irradiation (CW-LBI). CW-LBI implants a metal sphere with a diameter of approximately 0.1 mm into glass, manipulates it, and modifies its optical properties in a micro diameter trajectory . CW-LBI uses the interfacial tension between the softened glass and melted metal sphere to move the sphere toward the laser. When the temperature around the metal sphere is uniform, the metal sphere does not move. However, laser illumination generates a temperature difference between the laser absorbing surface of the sphere and the other side. The interfacial tension of the metal sphere in the laser absorbing surface decreases more than it does on the other side of the sphere because the interfacial tension decreases with temperature. The metal sphere moves toward the laser source due to a resultant interfacial tension force . The metal sphere movement velocity was approximately 0.1 mm/s. The authors discovered a new metal sphere movement which the metal sphere diffused into glass under higher power density laser illumination . This is referred to as fast movement because the velocity was significantly higher, approximately 100 mm/s.
For fast movements, in-situ observations and spectroscopy revealed the occurrence of plasma around the metal sphere. Similar to other plasma generating methods in a solid material, multiphoton absorption has been reported . In this method, laser illumination inside glass causes optical absorption, including multiphoton absorption, and the glass is translated into a plasma in a short time to change the refractive index around the focal spot . In an optical glass fiber with high power laser illumination, a high-intensity emission sometimes occurs and moves rapidly in it toward the laser source; this phenomenon is called fiber fuse [15,16]. Laser absorption by a crack or dirt heats the optical glass and generates plasma, leading to high-intensity emissions, which follow the high-intensity emission point and then move toward the laser source [15,16]. An optical absorption mechanism can be defined as follows. First, SiO2 decomposes pyrolytically into SiOx (x ≤ 1) when the optical glass fiber is heated by laser illumination with a wavelength of 1064 nm. SiO starts to absorb the laser illumination efficiently at high temperatures above 2300 K, and heats itself and the surrounding glass. After the temperature exceeds the threshold, SiO decomposes pyrolytically into plasma; an emission point moves at a velocity of hundreds of millimeters per second in glass as fiber fuse proceeds . Similar to fiber fuse, laser illumination heats a metal sphere as a starting point to heat bulk glass. High-intensity laser illumination increases the temperature of the glass and metal sphere above the plasma threshold. Its emission moves toward the laser source. Fiber fuse is a high-speed emission phenomenon in silica glass; additionally, a composition of metal and glass also moved fast with emission in this research. To the best of our knowledge, this new high-intensity plasma movement, generated using different materials, has not been extensively investigated so far. Additionally, the metal element diffuses into a trajectory of the metal sphere during fast movements. Fast movement generates a micro metal additive trajectory with a diameter of approximately 0.1 mm. Moreover, spheres with different compositions can be used to diffuse two or more elements into the glass simultaneously. Controlling the trajectory refractive index may result in an optical communication path. Er- or Nd-doped trajectories can generate an amplifier system in glass [5,6]. Implanting Au  or crystallites of mixed semiconductor CdSxSe1-x  in trajectories and changing the volume fraction can generate the color filter. Additionally, a metal sphere in chalcogenide glass might result in optical memory  because the metal sphere can heat and quench the glass with laser switching.
The applications require a control on the trajectory length and an addiction. For a fast-moving metal sphere with high temperature and higher power density, changes in the power density may lead to the desired trajectory length and addiction, and long illumination may lengthen the trajectory length.
We controlled the trajectory length and elucidated the relationship between the metal sphere speed and metal addiction in this study. First, the trajectory was observed with a high-speed camera, which revealed its dependence on laser power and illumination time. Higher laser power generated a longer trajectory and diminished the metal spheres. Additionally, longer illumination generated a longer trajectory. The trajectory length can be controlled stepwise from 0.1 to 2.0 mm by combining the laser power and illumination time. Furthermore, the results with energy-dispersive X-ray spectrometry (EDS) showed the metal diffusion distribution of the fast movement trajectories. Finally, we discussed the relationship between the metal additive amount, metal sphere velocity, and heat input.
Experimental apparatus used here was the same as described in . A sample includes a piece of borosilicate glass (TEMPAX Float, SCHOTT AG, Mainz, Germany), steel foil (# 223171, Nilaco Corp. Tokyo, Japan), heat insulator layer (Grade ES, Tosoh Corp., Tokyo, Japan), and a jig that clamped them together. Continuous wave laser illumination (RFL-C020/A/2/A, Wuhan Raycus Fiber Laser Technologies Co., Ltd., Hubei, China), focused onto the sample from the direction of the borosilicate glass, implanted a metal sphere into the glass. The wavelength of the laser was 1064 nm. The spot diameter and power in the metal sphere was 300 µm and 12 W, respectively. The metal sphere and its surrounding glass were observed with a high-speed camera (Phantom V7.3, Vision Research Inc., NJ, USA). A metal halide lamp (LS-M250 Sumita Optical Glass Inc., Saitama, Japan) was used to light up the sphere and its surroundings. A band-pass filter with a center wavelength of 440 nm (10BPF10-440, Newport Corp., Irvine, CA, USA) filtered the scattered laser light and heat radiation from the metal sphere. It was placed between the high-speed camera and the sample. After cross-sectioning a sample, scanning electron microscope (SEM) images (JSM-6510A, JEOL Ltd., Tokyo, Japan) and transmitted optical images with an optical microscope (Eclipse ME600, Nikon Corp., Tokyo, Japan) showed the trajectory surfaces of slow and fast movements. Each SEM image was a backscattered electron image. Additionally, detection with EDS revealed the elemental analysis of the trajectories. After the metal sphere was implanted into the glass, it moved toward the direction of the laser source while continuously illuminated and stopped at a certain spot . Low- and high-power intensity illumination generated transparent and colored trajectories, respectively . The main subject was a colored trajectory for the fast movements in this study. The diameters of the metal spheres were 80 ± 10 µm. A three-axis stage controlled the laser spot diameter at the metal sphere to be 120 µm. High-speed images with a high-speed camera showed the changing trajectory length while changing the laser power and illumination time.
A delay generator controlled the laser illumination time and synchronized the timing of illumination and the high-speed camera photography. The power varied from 5 to 17 W and the illumination time τ was 100 ms in the first experiment, revealing the relationship between the trajectory length L and power P. The powers were 9, 11, and 13 W and illumination time τ varied from 5 to 30 ms in 5 ms steps in the second experiment, thereby revealing the minimum trajectory length LMin. The minimum illumination time τMin required for the metal sphere movements depended on P. With shorter τ, a shorter trajectory was expected; hence, the experiments were conducted using the illumination time τ` = τMin − 5 [ms], where a metal sphere did not move, and τ=τMin + 5n [ms] (n = 0, 1, 2, 3) where a metal sphere moved.
3. Results and discussion
3.1 In-situ observation
The following method revealed the relationship between trajectory length and laser power. First, a metal sphere moved to a position 2 mm away from the metal foil with low-intensity laser illumination. After that manipulation, high-intensity laser illumination of the metal sphere generated fast movements. The experiment was performed three times to measure the trajectories. Figure 1 shows the behavior of the metal sphere. Figures 1(1), (2), and (3) show images at P = 9, 11, and 13 W, respectively. Figures 1(a), (b), and (c) show the images immediately before laser illumination (t = 0 ms), at t = 40 ms, and immediately after laser illumination (t = 100 ms). With no laser illumination after slow movement, some cracks occurred owing to the thermal strain. Consequently, Fig. 1(a) shows a black belt around a metal sphere. The black belt disappeared in Figs. 1(b) and (c) after laser illumination. The metal sphere was illuminated via laser from left to right in Fig. 1 and the metal sphere moved toward the laser source, to the left, with emission. The trajectory length L increased to 0.57, 0.98, and 1.26 mm with the increasing laser power P = 9, 11, 13 W, respectively. Additionally, the metal sphere diameter d diminished from 90 µm in Fig. 1(a) to 63 µm in Fig. 1(c) for the 13 W case. Fast movement consequently decreased the metal sphere diameter by 27 µm.
Figure 2 shows the relationship between the trajectory length L and laser power P. The X-axis denotes the laser power P [W] and Y-axis denotes the trajectory length L [mm]. Here, cross mark, triangle, square, and circle are used to plot the first, second, and third experimental values, and the average value, respectively. In each trial, the trajectory length L tended to increase with the increasing P. At P = 5 W, the metal sphere neither moved nor generated any trajectory in all the trials. The average value increased monotonically from L = 0.3 mm to L = 2.0 mm at the maximum.
Figure 3 shows the relationship between the changes in the metal sphere volume ΔV and laser power P. The X-axis denotes the laser power P [W] and Y-axis denotes the amount of metal sphere volume change ΔV [×105 µm3]. In each trial, the changes in the metal sphere volume ΔV tended to decrease with increasing P. For P = 5 and 7 W, ΔV did not change in all the trials. The average value decreased monotonically from ΔV= - 0.1 ×105 µm3 to ΔV= - 0.72 ×105 µm3 at the minimum.
As a result of Fig. 2 and Fig. 3, the colored trajectories tended to increase in length L and decrease in the amount of the volume with increasing the laser power P, which is related to the heat input into the metal sphere Q. However, the heat input decreased as the metal sphere moved toward the laser due to defocus. The laser power P increased, along with the heat input to a metal sphere Q and the distance of the metal sphere movement; hence, the metal sphere moved for a longer duration in the molten state. The volume of the dissolving metal in the trajectory from the metal sphere increased. When the laser power P increases, the trajectory length L increases, and the amount of metal sphere volume decreases.
3.2 Relationship between laser illumination time τ and trajectory length L
To investigate the minimum trajectory length under these experimental conditions, the laser illuminated the metal sphere with an illumination time of τ = 5–30 ms in steps of 5 ms. Here, with shorter τ, the metal sphere did not separate from the trajectory, and the exact trajectory length was unclear. To solve this problem, lower power, which did not generate a colored trajectory, illuminated the metal sphere to separate it from the colored trajectory, as shown in Fig. 4 and Visualization 1. Here, τ and T denote the laser illumination time of high and low intensity, respectively. Figures 4(a)–(c) and (c)–(e) show the phases of low and high intensity, respectively. First, a high-intensity laser at P = 13 W illuminated a static metal sphere in glass (as shown in Fig. 4(a)). Then, the metal sphere moved and generated a colored trajectory with emission (as shown in Fig. 4(b)). After the high-intensity laser illumination, the metal sphere stopped moving and stayed inside the colored trajectory (as shown in Fig. 4(c)). Low-intensity laser illumination at P = 7 W separated the metal sphere from the trajectory slowly (as shown in Fig. 4(d)). Figure 4(e) shows the metal sphere and trajectory separating at T = 400 ms.
Figure 5 shows the movements of a metal sphere at various laser illumination time τ. The metal sphere did not move at all for certain values of τ due to inadequate heat input. Experiments were conducted in conditions of minimum τMin, τMin + 5 ms, τMin + 10 ms, and τMin +15 ms. Metal spheres were separated from the colored trajectories, as described in Fig. 4, to measure the exact trajectory length. Figure 6 shows the relationship between the trajectory length L and laser illumination time τ. X-axis denotes the laser illumination time τ [ms] and Y-axis denotes the trajectory length L [mm]. Here, cross mark, triangle, and square indicate P = 9, 11, and 13 W, respectively. At P = 9, 11, and 13 W, the minimum illumination times required for fast movement were τMin = 15, 10, and 5 ms, respectively. Additionally, at τMin, the trajectory length was minimum, L = 0.12, 0.30, and 0.10 mm for P = 9, 11, and 13 W, respectively. Under the condition that τ is less than or equal to τMin, the metal sphere did not move or only transformed its shape. Under the condition that τ is greater than or equal to τMin, the trajectory length L increased monotonically with increasing τ; hence, the minimum trajectory length obtained in this study was LMin = 0.10 mm at P = 13 W and τ = 5 ms. It is necessary to discuss whether LMin = 0.10 mm was a true minimum or not. As a result of Figs. 1 and 5, the trajectory length L tended to be shorter with lower power P and shorter illumination time τ. This showed that the trajectory length L tended to be shorter when the heat input Q into the metal sphere decreased. However, a significantly lower Q could not generate fast movements, such as for P = 9 W and τ < 10 ms and for P = 11 W, τ < 5 ms. The decreasing heat input Q with adequate heat input required for fast movement can decrease the trajectory length L.
In this study, the stable minimum illumination time τ was 5 ms and the only experiment at τ > 5 ms is discussed here. The minimum trajectory length LMin = 0.10 mm was the true minimum trajectory. The trajectory length L can be controlled from 0.10 mm to 2.0 mm by using the experimental conditions described in this study.
3.3 Microscopic observation, elemental analysis, and relationship between Fe intensity peak and metal sphere velocity
After cross-sectioning, SEM and transmitted optical images showed the surface of the trajectory of fast movement, and the detection of elements with EDS revealed the relationship between the Fe signal intensity in the trajectory and the fast movement phenomenon. Figure 7 shows the cross-sectional observation and the result of point analysis with EDS for two fast movement trajectories. Figures 7(1) and (2) show the results for P = 9 and 13 W, respectively. Additionally, Figs. 7(a), (b), (c), (d), (1-e), and (1-d) show the transmission image of an overall trajectory, an SEM image, an enlarged image of the head part of the trajectory, an enlarged image of the tail part, an enlarged image of (1-d), and the result of point analysis with EDS, respectively. The targets of point analysis with EDS indicated via cross marks are a glass substrate (A), metal sphere (B), metal particles on the trajectory (C), and a trajectory without metal particles (D). The polishing process removed the metal sphere shown in Fig. 7(2). Fe was not detected in (A); however, it was further detected strongly in the order of (D) < (C) < (B). The trajectory contained metal sphere element, Fe, and highly condensed Fe formed metal particles. Area (C) also contained the main glass elements, Si and O. The spatial resolution of the EDS used in this study was 2 µm in the depth direction . Because the diameter of metal particles, 0.5 µm, was smaller than the spatial resolution, the result of (C) contained the surrounding glass signals simultaneously.
Figure 8 shows the SEM images of a cross-section of a slow-moving trajectory for P = 7 W. Figures 8(a) and (b) show an optical microscope and SEM micrograph of the transparent areas. The EDS only detected the glass elements, Si and O, at the trajectory positions indicated by cross marks, and little Fe. Hence, Fe seemed to diffuse on the trajectory during fast movements.
Figure 9 shows an SEM image (a), the line elemental analysis of two trajectories with EDS (b), the relation between the position of the sphere on the trajectory z (c), and the metal sphere velocity (d). Figures 9(1) and (2) show the results for P = 9 and 13 W, respectively. The gray circle in the v–z diagram shows the position of the metal sphere. The EDS graph in Figs. 9(1) and (2) showed a Si signal with intensity one fifth times as strong as the original data to facilitate the discussion on Fe signals. The detection conditions for Fig. 9(1) were the same as that for Fig. 1(2). After laser illumination, a metal sphere accelerated to reach the maximum velocity (fast movement) and then decelerated (slow movement). Finally, it stopped moving after the laser was switched off. The phenomenon of transformation from fast movement to slow movement resulted from the heat input into the metal sphere decrease as the diameter of the laser spot increased . The three areas of the trajectory were defined as follows: the tail part where the metal sphere was located before illumination (a), the belly part where the metal sphere moved fast (b), and the head part where fast movement transitioned into slow movement. In Fig. 9(1), Fe was detected at z = 0.00-0.56 mm and z = 0.62-0.70 mm. The area of z = 0.00-0.56 mm corresponded to the area of fast movement. In particular, the signal intensity of Fe at z = 0.19 mm was the strongest in the trajectory. The area of z = 0.62-0.70 mm shows the ending position of the metal sphere, and thus, the Fe intensity peak was much higher than that of the trajectory. Similarly, in Fig. 9(2), Fe was detected at z = 0.00-0.97 mm. In both Figs. 9(1) and (2), the Fe intensity peak in the trajectory lowered gently as z increased from the tail to the head part, and oscillated in Fig. 9(b). The EDS detection area contained some metal particles simultaneously owing to an inadequate spatial resolution of EDS compared to the diameter of metal particles. Fe was not detected in the area of transparent trajectory.
Next, the relationship between the position on trajectory z and the metal sphere velocity v is discussed to calculate the velocity from Fig. 1, which was the second trial experiment for P = 13 W. The X-axis denotes the metal sphere movements from the area where the metal sphere was at before illumination z [mm], and the Y-axis denotes the metal sphere velocity v [mm/s] in Fig. 9(c). Both Figs. 9(1) and (2) show the strongest intensity peak of Fe in the area before the metal sphere reached the maximum velocity of 22 mm/s and 44 mm/s, respectively. As z increases further, the Fe intensity peak decreased below the detection limit. Fe diffuses into the glass more when the sphere velocity is higher.
Finally, the discussion on the variations in heat input into a metal sphere is presented below. In Fig. 9(c), the X-axis denotes the distance from the starting point to the metal sphere z [mm]. The Y-axis denotes the heat input per 1 ms ΔQ [J], calculated using the laser illumination area only for a metal sphere. Furthermore, the calculations assumed the following four conditions: defocus with a metal sphere movement, decrease the metal sphere volume, absorption of only the laser illumination area equivalent to an area of the metal sphere diameter, and a Gaussian profile of laser distributions. Although the metal spheres were black just after illumination, the diameter of the metal sphere appeared to increase with strong emission during fast movement. Then, after fast movement, the emission disappeared, and the diameter became smaller than that for t = 0 ms. As t increased, ΔQ decreased. At P = 9 W, ΔQ values were 59, 48, and 50 mJ in the area where metal sphere was going to move, just before it reached the maximum velocity, and when it finished fast movement (no Fe diffusion), respectively. Here, ΔQ increased from 48 mJ to 50 mJ temporarily because the measured diameter of the metal sphere increased from 78 µm to 98 µm due to emission. Similarly, for P = 13 W, ΔQ was 90 mJ and 52 mJ in the area where metal was going to move and where it finished fast movement and Fe diffusion, respectively. Heat input per 1 ms, ΔQ > 50∼52 mJ, seemed to satisfy the condition that Fe from a metal sphere diffuses into a trajectory. Maintaining the laser intensity ΔQ > 50–52 mJ could generate longer trajectories than the maximum trajectory length reported here (LMax = 2.0 mm) by scanning the focus during the metal sphere movement.
The Fe signal intensity increased when the diffusion areas of Fe and the absolute value of the metal sphere volume change increased with higher laser power (Fig. 1–Fig. 6). In Fig. 3, at P = 9 and 13 W, the metal sphere volume changed by ΔV = -3.0×102 µm3 and -1.1×104 µm3, respectively; thus, the absolute value increased with higher laser power. Additionally, in Fig. 9, at z = 0.00–0.20 mm, the Fe intensity peak for P = 13 W (Fig. 9(2)) was higher than that for P = 9 W (Fig. 9(1)). With higher laser power, a longer metal-containing trajectory with highly condensed Fe area in the tail part required more volume of the metal sphere.
The heat input per ms ΔQ for P = 13 W was higher than that for P = 9 W. The glass heated more and its viscosity reduced with higher laser power. Hence, higher laser power, P = 13 W, increased the ability of the metal sphere to diffuse into the trajectory.
Figure 10 shows enlarged SEM and binarization images. The binarization was processed with the ImageJ image processing software. Figure 11 shows the number of metal particles N, the average particle diameter φ [µm], and the metal area S [µm2] in each part of the trajectory (the tail, belly, and head) calculated using enlarged images of Fig. 7(1-c)-(1-e) and (2-c)-(2-e). Here, the metal area was defined as S = N × π (φ/2)2. Circles, triangles, and cross marks show the number of metal particles N, the average particle diameter φ, and the metal area S, respectively. Furthermore, red and black marks show the data for P = 9 and 13 W, respectively. N increased near the head part at P = 9 W. In contrast, at P = 13 W, the maximum N was in the belly part and decreased from the belly toward the head part. At both P = 9 and 13 W, the maximum φ at the tail part (φ = 12.1 µm at P = 9 W, φ = 46 µm at P = 13 W) decreased toward the head part. The maximum S at the tail part (S = 22 µm2 at P = 9 W, S = 157 µm2 at P = 13 W) decreased similar to φ. The minimum S at P = 9 and 13 W were 22 µm2 and 157 µm2 at the head part, respectively. Thus, the Fe signal intensity in the tail part was the strongest in the trajectory and tended to decrease toward the head part (Fig. 9) owing to larger S in the tail part and increased φ.
The reason why the metal particles and the colored glass without metal particles contained Fe was as follows. The first hypothesis is that fast movement generates plasma in high-temperature surrounding glass; some amount of the metal sphere dissolved into glass as solid particles and liquid metal, and they peeled from the metal sphere to combine with the glass. The second hypothesis is that only liquid Fe combined with the glass and quenched to form fine particles because the distance from the heated liquid Fe and the heat source, i.e., the metal sphere, increased with time. Proving these hypotheses will be a subject of our future work. Dynamic consideration of the flow of solid particle and liquid metal from a metal sphere, Fe density change during fast movement, and the temperature field of the metal sphere and the trajectories will elucidate the problem.
In summary, we reported a method for controlling the trajectory length generated by a fast-moving metal sphere in glass, and the relationship between the metal sphere speed and metal addiction in this paper. First, in-situ observations revealed the relationship between the trajectory length, changes in the metal sphere volume, and laser power, as well as that between the trajectory length and laser illumination time. Second, the cross-sectional surfaces of metal sphere trajectories were observed with SEM and their elements were detected with EDS. This revealed the location dependence of the intensity of Fe signals in the trajectory. Finally, the discussions regarding the metal sphere velocity, input heat, and the metal additive amount suggested the following conclusions. The trajectory length L increased and the metal sphere volume V decreased when the laser power P increased. Then, L also increased with the increasing laser illumination time τ. Furthermore, L depended on the heat input into the metal sphere ΔQ and could be controlled from 0.10 to 2.0 mm. Thus, we conclude that maintaining ΔQ > 50-52 mJ by focusing the laser light or increasing the laser power during fast movements could generate trajectories longer than 2.0 mm in our future work.
Additionally, the metal particles embedded Fe in the fast movement trajectory. The number of metal particles and the intensity peak of Fe were maximum around the area where the metal sphere moved the fastest. Consequently, changing the metal sphere speed can control the Fe concentration added to the fast movement trajectory.
Japan Society for the Promotion of Science (18J22593, 24656096, 19H02035).
The authors declare no conflicts of interest.
1. A. Ellison and I. A. Cornejo, “Glass Substrates for Liquid Crystal Displays,” Int. J. Appl. Glas. Sci. 1(1), 87–103 (2010). [CrossRef]
2. T. Miya, Y. Terunuma, T. Hosaka, and T. Miyashita, “Ultimate low-loss single-mode fibre at 1.55 µm,” Electron. Lett. 15(4), 106–108 (1979). [CrossRef]
3. N. Singh, K. J. Singh, K. Singh, and H. Singh, “Comparative study of lead borate and bismuth lead borate glass systems as gamma-radiation shielding materials,” Nucl. Instrum. Methods Phys. Res., Sect. B 225(3), 305–309 (2004). [CrossRef]
4. L. M. Liz-Marzán, “Tailoring surface plasmons through the morphology and assembly of metal nanoparticles,” Langmuir 22(1), 32–41 (2006). [CrossRef]
5. W. Huang, R. R. A. Syms, E. M. Yeatman, M. M. Ahmad, T. V. Clapp, and S. M. Ojha, “Fiber-Device-Fiber Gain From a Sol-Gel Erbium-Doped Waveguide Amplifier,” IEEE Photonics Technol. Lett. 14(7), 959–961 (2002). [CrossRef]
6. J. A. Der Au, D. Kopf, M. Moser, and U. Keller, “60-fs pulses from a diode-pumped Nd: glass laser,” Opt. Lett. 22(5), 307–309 (1997). [CrossRef]
7. R. K. Jain and R. C. Lind, “Degenerate four-wave mixing in semiconductor-doped glasses,” J. Opt. Soc. Am. 73(5), 647–653 (1983). [CrossRef]
8. A. V. Kolobov and J. Tominaga, “Chalcogenide glasses as prospective materials for optical memories and optical data storage,” J. Mater. Sci.: Mater. Electron. 14(10/12), 677–680 (2003). [CrossRef]
9. R. Electrical, “Reversible Electrical Switching Phenomena in Disordered Structures,” Phys. Rev. Lett. 21(20), 1450–1453 (1968). [CrossRef]
10. 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]
11. H. Hidai, J. Wada, T. Iwamoto, S. Matsusaka, A. Chiba, T. Kishi, and N. Morita, “Experimental and theoretical study on the driving force and glass flow by laser-induced metal sphere migration in glass,” Sci. Rep. 6(1), 38545 (2016). [CrossRef]
12. N. Nishioka, H. Hidai, S. Matsusaka, A. Chiba, and N. Morita, “Clarification of fast metal sphere movement in glass,” Opt. Express 27(13), 18988–19001 (2019). [CrossRef]
13. M. Jupé, L. Jensen, A. Melninkaitis, V. Sirutkaitis, and D. Ristau, “Calculations and experimental demonstration of multi-photon absorption governing fs laser-induced damage in titania,” Opt. Express 17(15), 12269–12278 (2009). [CrossRef]
14. T. Meany, M. Gräfe, R. Heilmann, A. Perez-Leija, S. Gross, M. J. Steel, M. J. Withford, and A. Szameit, “Laser written circuits for quantum photonics,” Laser Photonics Rev. 9(4), 363–384 (2015). [CrossRef]
15. S. I. Todoroki, “In-situ observation of fiber-fuse propagation,” Jpn. J. Appl. Phys. 44(6A), 4022–4024 (2005). [CrossRef]
16. E. M. Dianov, V. E. Fortov, I. A. Bufetov, V. P. Efremov, A. E. Rakitin, M. A. Melkumov, M. I. Kulish, and A. A. Frolov, “High-speed photography, spectra, and temperature of optical discharge in silica-based fiberss,” IEEE Photonics Technol. Lett. 18(6), 752–754 (2006). [CrossRef]
17. Y. Shuto, S. Yanagi, and S. Asakawa, “Fiber fuse phenomenon in step-index single-mode optical fibers,” IEEE J. Quantum Electron. 40(8), 1113–1121 (2004). [CrossRef]
18. H. Soejima (1979). Spatial resolving power of electron prove X-ray microanalyzer (Doctoral dissertation)[in Japanese]. Retrieved from http://hdl.handle.net/11094/405