Inorganic nanoparticles in fiber laser technology

979-8-3315-9777-1/25/$31.00 ©2025 IEEE © 2025 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, incl. reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. This is the accepted version of the following article: Jan Mrázek et al., "Inorganic nanoparticles in fiber laser technology, " in *Proc. IEEE ICTON 2025* The final version is available at: https://doi.org/10.1109/ICTON67126.2025.11125185 Inorganic nanoparticles in fiber laser technology Jan Mrázek Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic [email protected] Michal Kamrádek Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic [email protected] Jan Aubrecht Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic [email protected] Jana Proboštová Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic [email protected] Ivan Kašík Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic [email protected] Ondřej Podrazký Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic podrazký@ufe.cz Jana Novotná Department of Material Engineering Faculty of Textile Engineering, Technical University of Libere Liberec, Czech Republic [email protected] Ivo Bartoň Fiber lasers and non-linear optics Institute of Photonics and Electronics of the Czech Academy of Sciences Prague, Czech Republic [email protected] Viktor Puchý Division of Metallic Systems Institute of Materials Research of the Slovak Academy of Sciences Košice, Slovakia [email protected] Abstract— Rare earth doped yttrium aluminium garnet represents a key photonic material for the construction of active optical fibers and scintillators. The successful impregnation of garnet nanoparticles in the fibers is a major technological step that is usually covered by nanoparticle deposition method. We present the correlations between the morphology of the silica frit, nanoparticle size and final concentration of the dopants in the preform. Silica frit exhibited a wide pore size distribution with the smallest pore size of 138 nm and the maximum of 321 nm. Thermal processing of sol-gel precursors followed by controlled sedimentation allowed the preparation of a set of colloidal solutions containing holmium doped yttrium aluminium garnet with tailored nanoparticle sizes. The nanoparticles of the size of 205 nm were successfully impregnated in the frit and processed in the preform showing the refractive index difference of 0.009. The results impact the technology of active optical fibers which can be used for construction of fiber lasers and distributed sensors of harmful radiation. Keywords— optical fiber, laser, garnet, nanocrystals I. INTRODUCTION The invention of fiber lasers changed the modern world, and fiber lasers have become established in many scientific and industrial applications [1], [2]. Their unique position in telecommunications as a key component of amplifiers is unquestionable today [3]. Their outstanding parameters, including repetition frequency and output beam stability allowed the construction of advanced metrological devices including LIDARs [4]. Intensive research has pushed their output power in continuous wave mode to theoretical limits of several hundred kilowatts and thus secured their irreplaceable place in machining [5]. Fiber lasers usually benefit from the emission of rare earth (RE) elements [6]. However, the implementation of the RE in the optical fiber, that is the heart of the fiber laser, presents a technological challenge. Common laboratory scale fiber optic technology use modified chemical vapor deposition (MCVD) method to prepare highly pure and homogeneous glass and the range of RE compounds, which can be used in MCVD process, is limited. For a long time, RE-doped fibers were prepared by the solution doping process [7], [8], which consisted of impregnating solutions of RE salts in a porous frit prepared by MCVD process. The high ionic strength of the applied solutions readily caused damage and cracking of the frit resulting in background attenuation of the fiber. Another disadvantage of this method was the common clustering of REs during the drying process, which reduced the luminescence efficiency. Chelate deposition technique made huge progress in the technology of active optical fibers [9], [10], [11]. It applies volatile chelate precursors of inorganic compounds including RE to dope the fiber core. High price and low long-term stability of this class of precursors are the main drawback of this technique. Therefore, this method is very useful for large-scale applications, but its use on a laboratory scale is limited. The use of colloidal solutions as a source of RE and other non-volatile compounds gave rise to a nanoparticle deposition method that prevents unwanted RE clustering and contributed to increased laser efficiency [12], [13], [14]. Another advantage of this method is the low operating costs and the wide range of nanomaterials which can be used. Although the nanoparticle deposition provided novel optical fibers with phenomenal properties and variety of applications [15], [16], [17] the studies focused on the interaction of the nanoparticles with the frit are limited. The basic physicochemical properties that have been evaluated for solution doping [18], [19] methods are missing. In this paper, we study the interaction of yttrium aluminium garnet (YAG) nanoparticles with silica frit. We show the relationship between the size of the nanoparticles and the morphology of the silica frit, which has a crucial influence on the impregnation of the nanoparticles into the frit. The results extend the fundamental knowledge of the processes involved in nanoparticle deposition technology and can be used to prepare active optical fibers with improved optical properties for fiber lasers and scintillators.. II. EXPERIMENTAL The holmium-doped garnet nanoparticles were prepared by sol-gel approach presented elsewhere [20], [21]. The amorphous powders were thermally treated at 850 and 950 °C to induce the crystallization and tailor the nanocrystal size. The amorphous powders were crystallized at 850 and 950 °C to induce the crystallization and tailor the nanocrystal size. The nanopowders were grinded in a planetary mill for 24 hours. The totals of 5 g of the grinded nanopowders were dispersed in 50 ml of anhydrous ethanol (VWR) and sonicated for 30 minutes to form the suspensions. The suspensions were sedimented for 6 hours. The stable colloidal solutions were decanted from the sediments and used for nanoparticle doping of the preforms. We used the common modified chemical deposition (MCVD) approach to prepare the preforms . To generate the porous silica frits in an inner wall of the silica substrate tube (Heraeus), the combustion reaction of SiCl4 with oxygen ran at 1400 °C. The flow rate of SiCl4 was 100 sccm. The colloidal solutions of the garnet nanoparticles were impregnated in the frits and dried out for 24 hours at an ambient temperature. The dried doped frits were sintered and collapsed in the preforms at 2100 °C. The pore size distribution was measured on mercury intrusion porosimeter AutoPore IV 9500 (Micromeritics). The evacuation pressure and time were 80 µm Hg and 5 min, respectively. Maximum intrusion volume was 100 l∙g-1. Scanning electron microscope (SEM) Lyra 3 XMU (Tescan) was used to visualize the morphology of the nanoparticles in the sintered powders. The samples were sputtered by thin carbon layer before analyses to prevent surface charging. We used Gwyddion 2.55 data visualization and analysis software to analyze the nanocrystal size and pore size distributions in SEM scans. Particle size analyzer ZetaPals (Brookhaven) was used to measure the nanoparticle size dispersed in the colloidal solutions. The refractive index profile of the preforms was measured on A2600 analyzer (Photon Kinetics). III. RESULTS AND DISCUSSION The chemical and structural properties of the silica frit play a key role in the impregnation process including the nanoparticle deposition. The chemical composition of the frit can change the affinity of the frit for dopants and reduce the sintering temperature of the doped frit into the glass. To minimize the multiplicity of effects we used pure silica frit for the experiments. Fig. 1 shows a visualization of the deposited frit using SEM. Considering the silica soot exhibits the ideal spherical shape, the frit is composed of spherical silica nanoparticles with a mean size of approximately 190 ± 10 nm. Where the nanoparticles touch each other, small necks typical of sintering processes can be observed. The arrangement of the nanoparticles has a major influence on the overall porosity of the frit and several types of pores can be recognized. Nanoparticles arranged in close touch form a body of chains and larger aggregates. There is no ideal arrangement between the nanoparticles; as approximated in crystal structure models; but small interspaces can be recognized in the structure. The irregular arrangement of chains and aggregates creates the walls of larger randomly shaped pores that increase in size at the surface of the frit and decrease in size at the depth of the frit. The largest size of the pores on the surface reached the value of 2330 and 1850 nm in diameter but sizes about 1200 nm were not exceptional. The sizes of the pores located deeper from the surface significantly decreased. These pore diameters did not exceed 519 nm and most of the observed pores were less than 350 nm in diameter. The explanation can be found in the mechanism of the deposition process. Heat is transferred from the outside of the substrate tube to the inner wall where a frit is formed, causing a temperature gradient from the outside to the axis of the substrate tube. The deposition and growth of the silica soot take place simultaneously with the sintering of the frit. The part of the frit attached directly to the substrate tube is heated to a higher temperature than the surface part according to the temperature gradient. The higher temperature at the interface between the substrate tube and the frit facilitates the fusing of the created nanoparticles and sealing of the pores. Therefore, pores located deeper in the frit have a smaller diameter than pores located closer to the surface. Fig. 1. SEM visualization of the silica frit used for the impregnation of the nanoparticles. Fig. 2. Pore size distribution of the silica frit used for the impregnation of the nanoparticles. To verify visual observation by SEM, we recorded the pore size distribution on mercury intrusion porosimeter and the results are shown in Fig. 2. The normalized frequency corresponds to the number of pores of a certain diameter. The normalized frequency curve did not show the presence of pores smaller than 138 nm, reached a raising edge for the value of 162 nm and increased rapidly reaching a maximum for 321 nm. Then, the number of pores decreased rapidly to the value of 832 nm and regularly decreased for larger pore sizes. The range of the pores of the size from 230 to 776 nm with the maximum at 321 nm showed the higher normalized frequency than 50%. With respect to the general pore size distribution and established terminology, the frit can be classified as a macroporous material with pore size higher than 50 nm. The recorded data was fairly consistent with SEM’s observations. The maximum counts of pores of the size around 321 nm corresponded to the partially sintered pores located in deeper parts of the frit closer to the interface between the substrate tube and the frit. These pores were fused from the larger pores during the deposition process. The pores of the sizes of several thousands of nanometers corresponded to the pores localized closer to the surface of the frit. Two specific regions can be recognized in the chart showing pore size distribution for the diameter smaller than 160 nm and larger than 5000 nm. Although mercury intrusion porosimetry should be applicable for pore size measurements down to 10 nm, we did not observe a response for pore sizes smaller than 138 nm. On one hand, small pores should be fused faster to minimize interfacial free energy, leading to their disappearance in favor of bulk material. On the other hand, ordinary CVD-prepared silica soot exhibits mesoporous porosity that had not been observed in the record. The cause can be found in the limited peak-to-peak resolution of the device used and gas porosimetry should be applied to cover this pore size range. The pore size region greater than 5000 nm represented different specificity. Such large pores were not observed in SEM images. Their presence can be attributed to the penetration of mercury into larger areas on the frit surface, where the walls of sintered nanoparticles create relief over a large distance. Values of the normalized frequency in this region oscillated around 6.5 ± 0.5%. The information about the pore size distribution can be used to predict the maximum nanoparticle size that can be used for successful impregnation in the frit. We can assume that the concentration of the impregnated nanoparticles of defined diameter is proportional to the normalized frequency of pore size. The larger the nanoparticles, the less likely they are to impregnate inside the frit. From this point of view, the highest probability of impregnation of the nanoparticles inside the frit should be achieved for the nanoparticle smaller than 321 nm. The larger nanoparticles are adsorbed on the frit surface and washed out during the pouring of the nanoparticle solution. However, such a size limit will be further reduced for several reasons. The interaction of the nanoparticles with the surface of the pores is the first one and the probability of impregnation will increase with the affinity of the nanoparticles for the frit. Another reason is the irregular arrangement of the pores. In order for the nanoparticles to reach the deeper parts of the frit, they must pass through the throats connecting the individual pores. These throats act as slits preventing the passage of larger particles and allow smaller particles to pass through only. Both phenomena can have a negative effect on the impregnation of nanoparticles inside the frit and reduce the final nanoparticle concentration. TABLE I. OVERVIEW OF THE NANOPARTICLE SIZES Nanoparticle size Garnet nanoparticles DLS (nm) YAG 850 °C 205 YAG 950 °C 301 YAG sediment 3200 We used dynamic light scattering (DLS) to measure the size distribution of the nanoparticles and to evaluate the longtime stability of the colloidal solutions. The results are shown in Fig. 3 and the values are given in Table 1. Within 6 hours, phase separation occurred in the suspension and two fractions were clearly observed, the white compact sediments at the bottom of the vial and the translucent solutions above it. The sediments contained the particles larger than a micrometer with broad size distribution centered around 3200 nm. The size distribution of the translucent solutions showed maxima at 205 and 301 nm for the powders thermally treated at 850 °C and 950 °C, respectively. Both size distributions were quite narrow with the half-width smaller than 28 nm. The particles in the colloidal solutions are usually aggregated in larger clusters stabilized by surface charge or surfactants. The aging of the nanopowders can cause an increase in size. The larger aggregates are no long-time stable and sediment faster than the smaller one. The nanoparticles remain dispersed in the solution for a longer time. Technology places several requirements on the stability of colloidal solutions. The colloidal solutions must be temporarily stable for successful impregnation into the frit. Poor stability may result in an inhomogeneous distribution of nanoparticles, hence dopants, as sedimentation would occur simultaneously with impregnation. To cover this point, the colloid solutions were allowed to sediment for several hours to make the colloid stable enough for homogeneous impregnation into the frit. Another point is the size of the dispersed nanoparticles. The Fig. 3. Particle size distribution of garnet nanoparticles, showing the shift in mean nanoparticle diameter with increasing processing temperature and the extreme value recorded for the sediment. The inset shows a photograph of the colloidal solution after sedimentation. nanoparticles must be small enough to pass through the pores and impregnate the frit deeper where the pores are smaller. Considering the presented values of the pore size distribution, the size of the nanoparticles should be smaller than the maximum of the pore size distribution at 321 nm, and both colloidal solutions should be successfully used to dope the fiber core. To observe the distribution of impregnated nanoparticles in the preform core, we used a refractive index profiler, and the results are shown in Fig. 4. The profiles of the refractive indices of the preforms follow a Gaussian distribution with maxima around the central axis. The basis reached a value of 1.457 that corresponds to the refractive index of pure silica glass. The preform doped with 205 nm diameter nanoparticles showed a maximum refractive index of 1.466, which corresponded to a refractive index difference between the doped core and the quartz substrate tube of 0.009. The EDS microanalysis revealed that the concentrations of the dopants were 0.203, 0.49 and 0.002 at.% for Y2O3, Al2O3 and Ho2O3, respectively. The preform doped with 301 nm diameter nanoparticles showed a maximum refractive index of 1.4573 only. With respect to the pore size distribution, the nanoparticles of this size should be impregnated in the frit. However, the impregnation was not successful. The low refractive index difference and the narrowing of the refractive index peak indicated that the nanoparticles were not impregnated deeper in the frit but were adsorbed only near the surface. This failure can be attributed to the disordered throats in the frit which prevented the passage of larger particles. Comparing the diameter of the impregnated nanoparticles with the maximum pore size, the diameter of the nanoparticles should not be greater than 63% of the maximum of the pore size distribution to be successfully impregnated in the frit. IV. CONCLUSIONS We demonstrated the correlations between the morphology of the silica frit, nanoparticle size and final concentration of the dopants in the preform. Silica frit exhibited a wide pore size distribution with the smallest pore size of 138 nm and the maximum of 321 nm. The pore size was correlated to the distance from the substrate tube. Thermal processing of sol-gel precursors followed by controlled sedimentation allowed the preparation of a set of colloidal solutions containing holmium doped yttrium aluminium garnet with tailored nanoparticle sizes. The nanoparticles of the size of 205 nm were successfully impregnated into the frit that was processed into the preform. The preform showed the refractive index difference of 0.009 and the concentration of Y2O3, Al2O3 and Ho2O3 of 0.203, 0.49 and 0.002 at.%, respectively. The nanoparticles of the size of 301 nm caused the refractive index difference 0.0003, by an order of magnitude lower than in the case of the smaller nanoparticles. The results impact the technology of active optical fibers which can be used for construction of fiber lasers and distributed sensors of harmful radiation. ACKNOWLEDGMENT The authors acknowledge the financial support from the Czech Grant Agency contract N° 23-05507S, inter academic research project Mobility plus contract N° SAV-25-09. And European Union and the Czech Republic's state budget under the project LasApp CZ.02.01.01/00/22_008/0004573. REFERENCES [1] W. Blanc et al. ‘The past, present and future of photonic glasses: A review in homage to the United Nations International Year of glass 2022’, Progress in Materials Science, vol. 134, Apr. 2023, doi: 10.1016/j.pmatsci.2023.101084. [2] A. Dhar, N. Choudhury, and R. Sen, ‘Advancements in fabrication technology for rare-earth doped optical fibers: A retrospective analysis’, Optics Communications, vol. 577, Mar. 2025, doi: 10.1016/j.optcom.2024.131463. [3] A. Ghosh, U. kumar Samanta, A. Dhar, S. Das, and M. chandra Paul, ‘Exploring the potential of a bismuth-erbium-vanadium co-doped optical fiber as a stable Q-switcher in the 1550 nm region’, Applied Optics, vol. 63, no. 19, pp. 5130–5136, Jul. 2024, doi: 10.1364/AO.525023. [4] X.-H. Mei et al., ‘Time and frequency domain characteristics of a quasi-continuous 2 μm thulium-doped fiber laser’, Journal of Infrared and Millimeter Waves, vol. 44, no. 2, pp. 156–162, Apr. 2025, doi: 10.11972/j.issn.1001-9014.2025.02.003. [5] M. Mendiratta and S. Prakash, ‘Investigation of nanosecond laser micro-drilling of titanium under different processing environments’, Proceedings of The Institution of Mechanical Engineers Part BJournal of Engineering Manufacture, Dec. 2024, doi: 10.1177/09544054241305167. [6] W. Blanc et al., ‘The past, present and future of photonic glasses: A review in homage to the United Nations International Year of glass 2022’, Progress in Materials Science, vol. 134, Apr. 2023, doi: 10.1016/j.pmatsci.2023.101084. [7] A. Dhar et al., ‘Multielement (P-Yb-Zr-Ce-Al-Ca) fiber for moderatepower laser application with enhanced photodarkening resistivity’, Physica Status Solidi A-Applications and Materials Science, vol. 214, no. 6, Jun. 2017, doi: 10.1002/pssa.201600655. [8] I. Kasik, P. Peterka, J. Mrazek, and P. Honzatko, ‘Silica Optical Fibers Doped with Nanoparticles for Fiber Lasers and Broadband Sources’, Current Nanoscience, vol. 12, no. 3, pp. 277–290, 2016, doi: 10.2174/1573413711666150624170638. [9] B. Lenardic and M. Kveder, ‘Advanced Vapor-Phase Doping Method Using Chelate Precursor for Fabrication of Rare Earth-Doped Fibers’, in OFC: 2009 Conference on Optical Fiber Communication, Vols 15, 2009, pp. 1538–1540. [10] T. Shi et al., ‘Ytterbium-doped large-mode-area silica fiber fabricated by using chelate precursor doping technique’, Applied Optics, vol. 53, no. 15, pp. 3191–3195, May 2014, doi: 10.1364/AO.53.003191. [11] P. Miluski et al., ‘Tm<SUP>3+</SUP>/Ho<SUP>3+</SUP> profiled co-doped core area optical fiber for emission in the range of 1.6-2.1 μm’, Scientific Reports, vol. 13, no. 1, Aug. 2023, doi: 10.1038/s41598-023-41097-2. [12] W. Blanc et al., ‘Compositional Changes at the Early Stages of Nanoparticles Growth in Glasses’, Journal of Physical Chemistry C, vol. 123, no. 47, pp. 29008–29014, Nov. 2019, doi: 10.1021/acs.jpcc.9b08577. [13] A. Eal-Junior, E. Pedruzzi, L. Macedo, and W. Blanc, ‘NP-doped optical fibers for all-fiber integrated optical sources, sensors, and actuators with self-sensing capabilities’, Optics Express, vol. 32, no. 25, pp. 45122–45132, Dec. 2024, doi: 10.1364/OE.541175. Fig. 4. Refractive index profiles of the preforms prepared by the impregnation of YAG nanoparticles of various sizes in the silica frit. [14] I. Kasik et al., ‘Erbium and Al2O3 nanocrystals-doped silica optical fibers’, Bulletin of the Polish Academy of Sciences, vol. 62, 2014. [15] J. Aubrecht et al., ‘Self-swept holmium fiber laser near 2100 nm’, Optics Express, vol. 25, no. 4, pp. 4120–4125, Feb. 2017, doi: 10.1364/OE.25.004120. [16] I. Kasik et al., ‘Fiber-optic detection of chlorine in water’, Sensors and Actuators B-Chemical, vol. 139, no. 1, pp. 139–142, May 2009, doi: 10.1016/j.snb.2008.10.064. [17] V. Matejec et al., ‘Microstructure fibers for gas detection’, Materials Science & Engineering C-Biomimetic and Supramolecular Systems, vol. 26, pp. 317–321, Mar. 2006, doi: 10.1016/j.msec.2005.10.043. [18] A. Dhar et al., ‘Characterization of porous core layer for controlling rare earth incorporation in optical fiber’, Opt. Express, vol. 14, no. 20, pp. 9006–9015, Oct. 2006, doi: 10.1364/OE.14.009006. [19] V. Petit, A. L. Rouge, F. Béclin, H. E. Hamzaoui, and L. B. and, ‘Experimental Study of SiO2 Soot Deposition using the Outside Vapor Deposition Method’, Aerosol Science and Technology, vol. 44, no. 5, pp. 388–394, 2010, doi: 10.1080/02786821003671315. [20] V. Cuba et al., ‘Radiation induced synthesis of powder yttrium aluminium garnet’, Radiation Physics and Chemistry, vol. 80, no. 9, pp. 957–962, Sep. 2011, doi: 10.1016/j.radphyschem.2011.04.009. [21] J. Mrazek et al. ‘YAG Ceramic Nanocrystals Implementation into MCVD Technology of Active Optical Fibers’, Applied Sciences-Basel, vol. 8, no. 5, Art. no. 5, May 2018, doi: 10.3390/app8050833.