Open Access
Add to BibSonomy
Abstract
In the present work, we report on the preparation of silicate glass containing crystals by means of melting a mixture of YbPO4 xenotime structured crystals and SiO2 nanoparticles. This nanoparticle mixture is used for preparation of large volume core preforms for laser active optical fiber. Temperature dependent sintering and fiber drawing experiments at temperatures up to about 2000 °C were conducted in order to assess the integrity of the crystals in the preform and fiber, respectively. The survival of YbPO4 crystalline particles in silica was investigated by X-ray diffraction (XRD), electron probe microanalysis (EPMA), Raman spectroscopy as well as static and time resolved fluorescence measurements. It was found that the particles withstand the high-temperature steps during the fiber fabrication process. XRD and spectroscopic measurements suggest that the Yb ions are located in a crystalline but also in an amorphous silica-dominated surrounding in the fiber, suggesting the partial decomposition of the crystals during the fiber fabrication.
Published by Optica Publishing Group under the terms of the Creative Commons Attribution 4.0 License. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI.
1. Introduction
Optical materials doped with rare-earth (RE) ions play a crucial role as a laser active medium. Silica derived glasses dominate the class of optical fiber laser materials because of their outstanding thermal properties, excellent optical transmission, electrical performance and chemical resistance [1,2]. Intense research on fabrication and lasing behavior of Yb-doped fibers has been published over the past three decades and today, Yb-doped silica fibers are among the most used laser materials for fiber lasers operating in the 1 µm-range [3,4,5].
Conventional doping strategies make use of watersoluble precursor salts that are oxidized to the respective oxides upon thermal treatment. The most common processeses are modified chemical vapor deposition (MCVD), powder based processes, e.g. reactive powder sintering technique (REPUSIL) or melting [2,6,7,8,9,10]. However, the preforms prepared by these processes are usually limited in achievable material combination, amount of dopant and doping distribution homogeneity. Especially, diffusion controlled processes and phase equilibria sensitively impact unwanted phase separation, segregation and dissolution reactions as well as RE ion clustering that eventually have to be dealt with during fabrication of Yb-doped, laser active materials and fibers. To circumvent clustering of RE in silica, codoping with aluminum and/or phosphorous as solubilizer is employed [11]. In case of RE-doped silica glasses, any non-dissolved RE compound can initiate unwanted crystallization of RE silicates [12] apart from the possible formation of cristobalite [13].
The quest for higher power levels of fiber lasers with excellent beam quality spurs the search for new fiber fabrication concepts pushing those limits, enabling the preparation of fibers with larger absorption and emission cross sections, less photodarkening and new emission wavelengths. One strategy is to replace the traditional silica-based amorphous environment of the RE-ions of the core region by ceramic, crystalline, or glass-ceramic environment [10,14]. The challenges in preparing such composite fibers are manifold: available preform and fiber drawing technology designed for operation at high temperature to accommodate SiO2 drawing requirements has to be adapted to allow processing of non-silica material confined in silica cladding. Furthermore, the target crystalline phases of the core region that contain the desired RE ions need to survive the high temperatures used during the preform preparation and fiber drawing processes. Crystalline particles in a glass matrix, i.e. glass-ceramics (GC), are usually fabricated from a glass using a two-step heat-treatment process that leads to crystallization. However, high silica GC shows a tendency of unfavorable formation of cristobalite [15], while phases with specific crystalline forms and such doped with RE are hard to produce.
A modified powder sinter method was already used to obtain silica glass containing crystalline microparticles of SrAl2O4:Eu2+,Dy3+ showing persistent luminescence (PeL) [16]. Recently, phosphate fibers with similar PeL were successfully drawn [17]. It was demonstrated, that these microparticles partially survived the fiber drawing process at temperatures lower than the one used to draw pure SiO2 glass. This encouraging result motivated the search for other suitable RE doped crystals which are thermally and chemically stable during the high temperature silica fiber drawing process.
Possible candidates are RE orthophosphates (REPO4) which are thermally and chemically more stable than RE oxides in silica even up to high temperatures of about 2000 °C. For RE orthophosphates, melting temperatures of about 2000 °C or above are reported [18]; for YbPO4 a value of 2150 °C is reported [19] which is only slightly higher than typical preform vitrification and fiber drawing temperatures of silica fibers. The superior thermal stability of RE phosphates over RE oxides was evidenced in previous work using MCVD to fabricate Al, P, Yb-doped silica based fiber core material where the formation of YbPO4 rather than the expected Yb2O3 was found [20].
Recently, the existence of aligned elongated YbPO4 particles in 1 mm thin phosphosilicate canes fabricated using sol-gel technique with an initial molar P/Yb-ratio ≥ 2.8 was reported [21]. The phosphosilicate core material powder was prepared by a sol-gel method with subsequent Yb(NO3)3 impregnation, dehydration and sintering at 1200 °C. The obtained rods were drawn into 1.6 mm canes without cladding at about 2000 °C. Depending on the Yb-content, the rods showed optical losses in the order of hundreds of dB/m. The formation of nano-crystals of YbPO4 was also evidenced in P/Al/RE-doped silica glass prepared by MCVD [22]. This work demonstrated the preference of Yb to form highly coordinated compounds rather than entering crystal matrices with lower coordination (Yb2O3) [21]. P and RE- codoped silica layers prepared by surface plasma CVD evidenced the formation of crystalline YbPO4 clusters leading to an increase in scattering loss by about one order of magnitude [23]. The formation of YPO4 particles after dissolution of YAG crystals that were incorporated into a fiber preform by a MCVD process using a suspension doping procedure is described in [24]. The influence of the particle size on scattering losses is described in [25].
The survival of YbPO4 particles as doping material in various bulk glass melts had been found to strongly depend not only on the thermal stability of the crystals themselves but rather on the chemical nature of the host glass melt [26,27]. Recently, YbPO4 crystals embedded via MCVD suspension doping process in germanosilica glass host has been reported [28]. A review article on spectral properties of RE-doped silica optical fibers including the effects of processing methods like all-glass fibers vs nanoparticle doped fibers is given by Varak [5].
In the present work, we report on the survival of YbPO4 nanoparticles during silica preform material fabrication by a powder sinter method for laser active optical fibers as a function of processing temperature. The particles are embedded in pure SiO2 without any further compounds like GeO2 [28] or Al2O3 from former YAG [24] that may lead to differences in the viscous behavior of the matrix. In comparison to the MCVD based methods including suspension doping of porous layers that are used to prepare a fiber preform [24,28], the sinter method allows the preparation of large volume bodies rather than layers, i.e. much bigger fiber cores. While in MCVD derived preforms, typical preform core diameters are around 1…3 mm, preform core diameters of > 11 mm can be achieved using the sinter method. The reported core diameters of the MCVD-derived fibers are about 10 µm [24] and 4 µm [28], respectively, whereas using the powder sinter method diameters above 100 µm are achievable (see paragraph 3.2). In general, higher core diameters allow a power scaling at the expense of beam quality. However, the higher core diameter relates usually to longer exposure times upon preform sintering or fiber drawing at high temperatures affecting the crystal formation or survival. While the duration and temperature of thermal treatment at high temperature is determined by the collapsing steps during MCVD fabrication, the thermal treatment can be chosen without limits in case of powder sinter method. Therefore, in the present work we discuss in detail the survival of crystals observed upon different thermal conditions used in our powder-based sinter method. Crystal survival is evaluated based on spectroscopic properties measured on test fibers fabricated from drawing the material. Two types of material treatment for fiber fabrication are compared: (I) the powder-route via powder-in-tube processing, and (II) the vitrified material-route. Sample characterisation was done using X-ray diffraction (XRD), electron probe microanalysis (EPMA), Raman spectroscopy as well as static and time resolved fluorescence. We demonstrate that the particles obtained from max. 3.5 wt% xenotime YbPO4 nanoparticles mixed with high purity silica nanoparticles withstand the high temperature steps of the fiber fabrication process.
2. Materials and methods
2.1. Materials
2.1.1. YbPO4 nanoparticle synthesis
The xenotime YbPO4 nanoparticles were prepared by a precipitation procedure using commercially available YbCl3·6H2O (Sigma-Aldrich, 99.9%) and (NH4)2HPO4 (Sigma-Aldrich, > 99%) according the recipe given in Ref. [26]. The obtained precipitate had been calcined at 1200 °C for 2 h (sample YbPO4-1200 in [26]) and ground in a mortar using ethanol for about 10 minutes followed by drying for at least 1 hour in an oven at 60 °C to remove the solvent. XRD analysis of the resulting particles proved >97% xenotime crystal structure and a small amount (<3 wt%) of Yb(PO3)3 metaphosphate. EPMA investigations revealed an equivalent molar composition in the powder of 50.1 mol% P2O5 and 49.9 mol% Yb2O3. The crystallite size is about 200 nm as obtained from XRD measurements [26].
2.1.2. Preparation of YbPO4-SiO2 nanoparticle powder mixtures
Silica powder mixtures containing YbPO4 crystals were derived from aggregates of high-purity silica particles (kindly provided by Heraeus Quarzglas GmbH, Germany) with a specific surface area of about 30 m2/g. SEM analysis of the silica particles indicated an average size in the order of 100 nm and an aggregate size of upto 500 µm.
All powder mixtures were obtained by means of mixing the as prepared YbPO4 crystalline xenotime nanoparticles with SiO2 nanopowder, similar to the method described in [16]: batch sizes of about 100 g each were prepared in an agate mortar by gradually adding the SiO2 nanoparticles to the synthesized YbPO4 particles starting with a 1:1 weight ratio.
Mixtures with upto 3.5 wt% of YbPO4 were obtained, see Table 1. The (equivalent) Yb2O3 content of samples for drawing experiments (≤ 1.5 wt%, see Table 1) was chosen so as to be comparable with existing all-glass MCVD derived fiber core material (0.11 and 0.16 mol% Yb2O3 reference fibers fabricated by MCVD solution doping [29]). Table 1 contains an additional MCVD reference fiber with a comparable Yb-content but with Al2O3 codoping, typical for laser fibers.
Table 1. Overview of samples including their YbPO4 doping amounts and equivalent oxides doping concentrations
2.1.3. Preparation of YbPO4-SiO2 bodies via thermal sintering under different conditions
Green bodies with a relative powder density of about 50% were prepared by pressing the powder mixtures (samples A) to cylinders using a cold isostatic press (KIP 200E, Paul-Otto Weber GmbH, Remshalden, Germany, pressure: 50 MPa) and then machining them to a diameter of 11 mm and a typical length of about 2 cm. Sample cylinders were inserted into a fused silica jacketing tube and sintered at atmospheric pressure for 2 min. Different temperatures in the range from 1450 °C to max. 1850 °C were realized using an inductive furnace (HR-80, Thermal Technology GmbH, Bayreuth, Germany). The heating rate was in the order of 100 K/min [30]. Temperatures above 1850 °C were realized using a typical MCVD lathe equipped with a H2/O2 burner.
2.1.4. Fiber drawing using YbPO4-SiO2 dense bodies sintered on MCVD lathe
For fiber drawing experiments, cylindrical bodies from the as prepared powder mixtures (samples B) having a diameter of 20 mm and a length of at least 4 cm were prepared by pressing method as described above. The cylinders were thermally dried at a temperature up to 900 °C in order to reduce the risk of bubble formation during fiber drawing. Then, a sample cylinder was inserted in a silica tube and assembled on a MCVD lathe for vitrification at a temperature of about 2050 °C. This sintering was carried out at about 250 K higher temperatures than reported in Refs. [24] and [28].
From the vitrified samples, acrylate coated silica fibers were drawn using a vertical fiber drawing tower at temperatures adjusted near to the softening point of the SiO2 glass (∼ 2000 °C, drawing speed 9 m/min).
2.2. Methods
2.2.1. XRD analysis
The crystalline phases in the samples were determined on crushed powdered probes by X-ray diffraction using CuKα radiation (XRD, device X’pert Pro, Malvern Panalytical, Almelo, The Netherlands) and identified using the ICDD database.
Microarea XRD measurements were performed on a general purpose diffractometer (Rigaku Smartlab, Japan) with 0.5 mm beamwidth (in y) and ∼1.5 mm (in x – footprint dependent on angle of incidence). Measurements were taken each 0.5 mm across the whole sample diameter of the as-vitrified sample slice.
2.2.2. EPMA analysis
Composition of crystalline phases and the surrounding glass matrix of the vitrified material was determined by wavelength dispersive electron probe microanalysis (WD-EPMA) using a JXA8800L (JEOL company) microprobe analyzer. Compositions are given as equivalent oxides in mol%, that is YbPO4 is expressed as equivalent amounts of Yb2O3 and P2O5 at ± 5% relative uncertainty.
2.2.3. Raman analysis
A Raman microscope (Renishaw inVia) was used in confocal mode with a 100x objective at 532 nm and 785 nm laser excitation. A linescan with a resolution of 0.5 µm was conducted in order to determine the profile of the interface crystalline particle – glass matrix.
Raman mapping on one sample was performed using the same Renishaw spectrometer type but using an Ar-laser with 514 nm excitation wavelength. Spectra were collected with an 100x-0.85NA lens with 3x 60 s accumulation at each measurement position. An area of ∼10 × 10 µm2 was mapped with 0.3 µm spacing resulting in 34 × 34 individual spectra. The spectra were analyzed at each individual measurement spot for the signal contribution by crystalline YbPO4 and amorphous SiO2 (assuming signal superposition), using spectra at positions obviously within the glassy matrix and directly at a particle as references.
2.2.4. Fluorescence analysis
Fluorescence spectroscopy was done on the sintered bulk samples (samples of group A) and the test fibers (samples of groups B-D) using a 906 nm excitation source (BWT laser, China), and a collection fiber (600 µm, NA 0.2, CeramOptec). A spectrometer was used for detection (Spectro 320D, Instrument Systems, detection range 190 nm to 2150 nm). Spectrometer settings were: excitation 906 nm, bandpass 5 nm, LP-filter 950 nm for pump light suppression.
The fluorescence decay as a measure of lifetime of the corresponding excited energy state (2F5/2) was measured on fiber samples (group B-D, see Table 1). The measurement was done using a pulsed laser source around 976 nm (SDL-MOPA) with 20 Hz repetition rate, 2 ms pulse width and an accuracy of ±5 µs. An excitation pulse width of 2 ms was chosen to ensure a sufficiently strong signal. The signal was detected using an InGaAs-detector and displayed with a Tektronix 250 MHz oscilloscope. A filter was used to ensure that no pump light reached the detector.
To avoid any disturbing influences by reabsorption of emitted fluorescence or self-trapping by multiple scattering events, the twice-perpendicular method was used as described in [31]. In order to reduce any parasitic effects, the decay measurement was done using a very thin powder layer and a long-pass filter for pump light suppression.
3. Results and discussion
In this section, results from sample sintering and sample characterisation using XRD, EPMA, Raman spectroscopy as well as static and time resolved fluorescence will be presented and discussed.
3.1 Preform
3.1.1. Sintering results
The sintering procedure yielded mechanically stable bodies with decreasing porosity upon increasing sintering temperatures, as indicated by shrinkage measurement results. The linear shrinkage in radial direction (change in sample diameter before and after treatment) during the 2 min oven heating experiments was 17-20%.
Figure 1(a) shows that sintering at 1950 °C resulted in a bubble free, opaque body containing YbPO4 particles. The microscopic image of a polished sample piece reveals a random distribution of the particles and agglomerates (Fig. 1(b)) but no pores. In contrast to this, an undoped SiO2 reference sample was completely homogeneously vitrified (transparent, Fig.1a) under the same conditions.
Fig. 1. (a) Images of preform sintered on a MCVD lathe at 1950 °C containing 3.5 wt % of YbPO4 (opaque) between two parts of undoped sintered silica as illuminated by daylight, (b) microscopic image of cross section of a part of sample slice.
A bubble-free opaque preform could also be obtained for a similar treatment at 2050 °C which covers well the temperature range of silica fiber drawing process.
3.1.2. EPMA results of sintered samples
The spatial distribution of YbPO4 particles can be seen in a backscattered electron (BSE) contrast image in Fig. 2(a). An EPMA measurement across a YbPO4 particle in silica matrix is shown in Fig. 2(b) and (c). Please note that the composition is expressed as equivalent Yb2O3 and P2O5 rather than the actual YbPO4.
Fig. 2. BSE images (a, b) and EPMA analysis (c) of a sample from a preform treated at 2050 °C. (The marked grey area represents the particle aggregate region as crossed by the red line in Fig. 2 (b).)
Fig. 2(b) shows a large particle cluster with a size of about 25 µm. Apparently, there is no P and Yb dissolved in the matrix at a large distance away from the cluster, as supported by the EPMA line scan results presented in Fig. 2(c). The cluster size derived from the red line in the BSE image corresponds well to the dimensions obtained from the inflection points of the EPMA linescans. There is no obvious dissolution or dispersion by flowing of YbPO4 particles. A diffusion zone of a few microns beyond these points is evidenced. The measured SiO2 content at the particle region most likely results from EPMA probing at deeper sample distance.
While the molar ratio of Yb:P is 1:1 in case of the as-prepared YbPO4 particles, after thermal treatment it is increased to about 1.25:1. This is most likely due to the evaporation of gaseous POx species at high temperatures. A similar effect of reduced P content via evaporation was previously reported for silica fiber preforms prepared by MCVD [32].
3.1.3. XRD results of sintered samples
XRD was carried out to prove the survival of the crystalline particles upon temperature treatment. XRD patterns are shown in Fig. 3, they evidence the existence of both SiO2 cristobalite and YbPO4 xenotime depending on sintering temperature. In all samples, XRD patterns revealed that YbPO4 xenotime (ICDD file 00-045-0530) survived at sintering temperatures up to 1950 °C. In the pattern of the fiber sample consisting of a particle doped core region (0.25 wt% YbPO4) and amorphous silica cladding, additional peaks were observed on the amorphous signal, which indicate that at least some of the crystalline structure of the particles is preserved in the fiber. For all samples sintered between 1550 °C and 1700 °C formation of cristobalite was obvious, mostly the α-phase (ICDD file 00-039-1425). However, the XRD patterns of the samples prepared at 1550 °C and at 1650 °C indicate the presence of a cubic SiO2 phase (β-cristobalite, ICDD file 00-027-0605). In samples sintered at temperatures above 1700 °C, i.e. above the melting temperature of cristobalite [33], no crystalline SiO2 was found. The sintered samples without a high cristobalite content (i.e. sintering temperatures <1600 °C and >1700 °C, respectively) reveal an increased amorphous XRD background.
Fig. 3. XRD pattern of samples prepared with 3.5 wt% of YbPO4 sintered at different temperatures as indicated, of the as-prepared YbPO4 particles, and of cristobalite for comparison. Also shown: peak pattern of a sample of a fiber drawn at 2050 °C. It is worth mentioning that the fiber sample contains only 0.25 wt% YbPO4 in the core and is surrounded by amorphous SiO2 cladding material. For a better presentation of the results, the curves are plotted with offsets and arbitrarily scaled intensities.
The relative amount of formed cristobalite can be estimated taking into account the ratio of peak intensities of cristobalite <111 > over YbPO4 < 020 > at 2θ=21.48° and 26.12°, respectively. However, the reliability of quantification of the crystalline phases is limited because of the predominant amorphous phase. Figure 4 shows the ratio of the calculated relative amount of formed cristobalite as a function of sintering temperature. There is no indication of formation of Yb-silicate crystals from the XRD pattern of any sintered samples.
Fig. 4. Ratio of peak intensities of cristobalite <111 > over YbPO4 < 020 > as a function of sintering temperature, normalized to the maximum value. The dashed line is drawn to guide the eyes.
To assess an areal image of crystallinity, microarea XRD was performed on a sample slice treated at 1950 °C. Data was background subtracted prior to plotting as different levels of background signal are present across the measurements (particularly high background close to the edges of the sample, where the steel sample holder is irradiated). The analysis results are shown in Fig. 5.
Fig. 5. Microarea XRD results of sample prepared with 3.5 wt% of YbPO4 and sintered at 1950 °C as a function of position across the sample diameter (see photograph, the bars show the approximate measurement areas). The left graph reveals the average intensity of the amorphous hump (a-SiO2) and the highest intensity crystal peak (YbPO4) around 26° as a function of y-coordinate as determined from the (heat-)map. The graph at the top shows exemplary patterns collected from the samples center, Pos [0,0], and around the interface between opaque core and undoped silica in the photograph.
The sample shows similar silica intensities across the whole sample diameter except for the edges where the X-ray footprint starts irradiating the sample holder due to the round sample shape. The crystal phase is present more or less uniformely in the scattering white (opaque) part in the center of the sample, which underlines the aforementioned somewhat random particle distribution as observed in microscopy investigation (Fig. 1(b)). The signal intensity of the crystalline YbPO4 phase rapidly decreases near the interface between the doped core and the undoped SiO2 cladding region, as also visible in the microscopic image.The averaged intensities around the amorphous silica hump (∼21°) and the highest intensity peak of YbPO4 (∼26°) reveal the homogeneous occurence of the glassy phase across the whole sample whereas the crystal phase is rather randomly distributed in high amounts within the opaque, doped core, with rapidly decreasing signal intensities around the transition region between doped core and undoped cladding. The variations of the YbPO4 curve are a result of different distances of the nearest YbPO4 crystals from the sample surface that can be bigger than the penetration depth of the X-ray (order of magnitude: 101 µm) and influences the signal intensity. The artefacts around ∼44° 2θ are signals of the steel sample holder which marks the edges of the sample disk.
3.1.4. Raman results of sintered samples
Raman microscopy was used to get local information on the amorphous and crystalline characteristics of the samples, especially at the interface between the crystalline YbPO4 particle and the glass matrix. In order to determine the profile of the interface between crystalline particle – glass matrix, Raman spectra were recorded every 0.5 µm. A randomly chosen particle in the preform sintered at 1950 °C that was clearly visible by optical microscopy was investigated. Raman spectra were recorded starting at the particle and then proceeding to the surrounding glass matrix, similar to the inset cross section shown by the red line in Fig. 6 (a). The Raman spectrum when measured in the particle is similar to the one presented in [26] confirming the survival of the particles during the sintering process. The spectrum exhibits the primary bands at 490 cm−1 and 1004 cm−1 which can be related to bending and symmetrical stretching modes of Yb-xenotime, respectively [26]. When measured in the pure glass, the Raman spectrum exhibits the typical Raman bands of silica glass [34]: The main band around 450 cm-1 can be related to the symmetric stretching Si -O -Si bond vibration and the shoulders at 485 and 600 cm−1, to the so-called defects lines D1 (four-membered rings) and D2 (three-membered rings), respectively [34]. The band at ∼600 cm-1 can be attributed to rocking vibration and the signal at 1000–1200 cm−1 to the Si – O – Si asymmetric stretching related to Q4 units. However, some additional peaks can be seen between 1000 and 1200 cm-1 taking the Raman spectra in the Yb-containing crystals. They can be thus attributed to symmetrical stretching modes of Yb-xenotime [26].
Fig. 6. Raman spectra at different positions (0, 4.5 and 10 µm from starting point of the red scanning line (a) λex = 532 nm, (b) λex = 785 nm. The inset is an optical microscopy image marking the position of the line profile (red line).
A Raman microscope with 780 nm excitation wavelength was also used to perform micro-luminescence measurements at selected spots (see Fig. 6(b)). It was found that the shape of the emission within the crystal agglomerate region matches the one reported [26] for pure YbPO4 crystals. The spectra obtained at the intermediate glass region exhibits an emission band that is typical of Yb3+ ions located in an amorphous material. The latter we consider as evidence that the particles dissolve (at least partially) in the glass matrix during the fabrication process.
Following the line scan analysis, a Raman mapping measurement was conducted to visualize the particle surrounding in more detail, especially the transition from crystalline particle to amorphous silica matrix. The results are shown in Fig. 7.
Fig. 7. (a) Optical micrograph of a particle embedded in silica matrix after treatment at 1950 °C investigated during Raman spectroscopic analysis. (b) Regular Raman map of the particle showing the fraction of crystalline YbPO4 signal in the individual spectra (see experimental section) at all positions of the map with 0.3 µm step size. (c) Confocal Raman-linescan for increased spatial resolution, revealing a comparibly sharp transition between crystalline YbPO4 and the surrounding glassy material.
For the analysis, the Raman spectra were separated into their individual signal contributions from pure silica-glass and YbPO4 crystal phase (compare Fig. 6). The Raman mapping supports the aforementioned sharp transitions between glass and embedded particle (Fig. 7 b,c). However, due to the irregular particle shapes, some parts might be covered by glass phase in the polished surface, which is not diretly evident from simple reflected light microscopy. E.g. while the top left part of the particle in Fig. 7 a) looks equally scattering under an optical microscope as the rest of the particle, its clearly showing substantial glass signal in the Raman map, revealing the existence of matrix material at those measurements position. Similarly the barely visible disturbance seen at the bottom edge of the optical micrograph in Fig. 7(a) seems to be crystal phase in a very fine particle distribution, as substantial crystalline YbPO4 signal is observed there (Fig. 7(b)), although the micrograph does not show white light scattering from these positions.
3.1.5. Fluorescence results of sintered samples
The normalized fluorescence emission spectra of the sintered samples are given in Fig. 8. The spectrum of the sintered samples changes with temperature. A remarkable feature is the increasing peak at about 1035 nm with increasing temperature, especially in case of samples treated above the melting temperature of cristobalite. A fully amorphous Yb-doped silica fiber (MCVD3, see Table 1) with Al2O3 codoping is also shown for comparison.
Fig. 8. Fluorescence spectra of samples prepared at different treatment temperatures as indicated. The emission range in the shaded area is prone to reabsorption and the results are ambiguous and therefore are excluded from further discussion. An Yb-doped silica fiber (MCVD-made preform) is also shown for comparison. Excitation wavelength was 906 nm.
The emission around 980 nm is valid only on a very limited scale since in granular samples the fluorescence light can undergo multiple scattering events which increase the likelihood of any form of fluorescence self-trapping. Additionally, there is an influence by reabsorption of any emitted fluorescence in Yb-doped samples. Reabsorption is particularly strong around 980 nm because in this wavelength range the absorption- and emission cross sections are similar. Therefore, the spectral shape around 980 nm will not be further discussed.
3.2 Fiber
3.2.1. Fiber drawing results
The drawing of a fiber from the higher doped material of 1.5 wt% YbPO4 resulted in very inhomogeneous and extremely fragile fiber. So a preform was prepared with 0.25 wt% YbPO4 and drawn to a fiber with a 150 µm cladding diameter (labeled as CP-fiber). This fiber was mechanically stable and thus used for further investigation. A typical cross section of the drawn fiber is given in Fig. 9. It indicates decreased light transmission in the fiber core due to scattering centers. Furthermore, larger particle clusters > 1 µm are observed revealing the partial survival of particles. The survival has been further confirmed by XRD (see Fig. 3). Despite of the low particle content, the attenuation measurement on the fiber could not be performed due to high scattering losses induced by the large particle (cluster) size being in the same order of magnitude as the wavelength.
Fig. 9. Microscopic image of the cross section of a fiber in transmitted light (a) and zoom showing clusters (b). Length of the transmitted fiber: about 2 cm, scale bar: 50 µm.
3.2.2. Fluorescence spectra of CP-fiber
Fluorescence spectra of the fiber were measured and compared to those of standard Yb2O3-doped silica and Yb2O3-doped silica fiber prepared using MCVD (labeled as MCVD1 and MCVD2, respectively, see Tab.1). The recorded and normalized (relative to peak-value) spectra are displayed in Fig. 10. In all cases, Yb-doping concentrations are assumed to be low enough not to cause any concentration quenching effects [35]. The spectral shape of the fluorescence curves reveal that the CP-fiber behaves differently compared with the MCVD2 (phosphosilicate) fiber but very similar to the MCVD1 (silicate) fiber. This indicates that the crystalline structure apparently partially dissolves during high temperature fiber drawing which results in an obviously dominating amorphous environment of the Yb-ions, free of P. P occupation on Yb sites seems unlikely due to the different sizes of the two cations, - the ionic radius of P5+ and Yb3+ being very different [36]. Instead, formation of gaseous and evaporable POx species are more likely to be generated. The parallel existence of an amorphous environment of the Yb-ions as well as of crystalline YbPO4 clusters is also obvious when comparing fluorescence spectra of the fiber to that of a sintered sample (see Fig. 10, the dark grey and dotted grey curve). This confirms the observed XRD results (see Fig. 3).
Fig. 10. Normalized fluorescence spectra of CP-fiber (black), and the two MCVD-prepared fibers MCVD1 (green), and MCVD2 (orange) and the sample sintered at 1950 °C (dotted grey curve). Excitation wavelength was 906 nm.
The dominating amorphous environment present in the fiber can be attributed to the additional heat treatment of the preform during fiber drawing at a temperature of about 2100 °C. Furthermore, the drawing process exerts not only a thermal load but also a mechanical force on the crystalline particles that may influence the dissolution rate of particles in the neck-down region or may lead to mechanical stress on the particles after cooling. Comparable changes in dissolution rates have been observed in crystal growth processes due to stirring effects [37]. Changing in the particle shapes during drawing has also been observed before [38]. These effects might explain the reason the different emission spectrum from what is expected for a native CP-fiber to what would be expected for an amorphous core (MCVD derived) fiber, apart from expectation relying to the doubled heat treatment alone.
3.2.3. Fluorescence decay results of CP-fiber
Fluorescence decay curves were recorded for the CP-fiber, the MCVD1 fiber, the MCVD2 fiber, and the as-prepared YbPO4 powder using a 976 nm excitation source and monitoring emission at around 1 µm. The decay-time is dependent on several parameters as e.g. the (local) concentration of Yb-ions, the type and concentration of other ions (the surrounding matrix, e.g. P5+, Al3+) and the local structural conditions (amorphous, crystalline or both present partially).
The results of the fluorescence decay measurements are shown in Fig. 11 and summarized in Table 2.
Fig. 11. Fluorescence decay curves of the CP-fiber, the MCVD1 fiber, the MCVD2 fiber, and the as-prepared YbPO4 powder using a 976 nm excitation source and monitoring emission at ≥ 1 µm. An appropriate filter was used to separate pump from signal light. Please note, that MCVD1 fiber contains no P2O5, for details see Tab.1. The time resolution is about 5 µs caused by the pulse decay time of the excitation source. The fit data was taken with the help of Python’s numpy-Module.
Table 2. Fluorescence decay results of the different investigated samples (powder & fibers)
The fluorescence decay curve of the CP-fiber is best described by using a double-exponential fit, while the curves of the two MCVD derived fibers yield best a single-exponential fit. The two different decay constants in the CP-fiber are expected due to the two different phases (amorphous and crystalline). Moreover, these two phases differ in terms of their (local) concentration and phonon energy. This is reflected in a different decay behaviour for both phases. The second time-constant (713 µs; lower weighting) is thus attributed to the amorphous environment because the value is similar to the decay constant of the (amorphous) MCVD-1 fiber. The other time constant (356 µs, higher weighting) therefore reflects the crystalline environment indicated by the similarity to the decay-constant of the (crystalline) as-prepared powder. Furthermore, it has to be mentioned that Yb(PO3)3 being potentially present in the original powder mixture might decompose into orthophosphate and P2O5 at our sintering or drawing temperatures and therefore would not be observable in the fiber anymore. Note that the given weighting factor is a fit-result and does not necessarily represent the amount of the corresponding phase.
Comparing this second value with the decay constant for the as-prepared powder (356 µs vs. 248 µs, respectively), one clearly recognizes a significant discrepancy between the crystalline fiber-phase and the crystalline powder phase which might seem surprising. We assume the mismatch between both values stems from concentration quenching causing a reduction in fluorescence decay in the powder.
Comparing the decay behaviour of the CP-fiber to that of the MCVD-fibers, it is worth to note that the decay constant of CP-fiber of 713 µs is close to the decay time of the MCVD1 fiber with only Yb2O3 doping (760 µs, no P present). This again indicates that in both fibers, Yb3+ experiences similar structural conditions despite the different fabrication technologies. The lower time constant of 356 µs (higher weighting) of the CP-fiber most likely reflects the crystalline orthophosphate environment as also observed by XRD (see Fig. 3, as discussed above).
In the CP-fiber, particles from clusters might be partly separated by the drawing process during the viscous flow that increases the mean nearest neighbor distance in comparison to the as-prepared powder. Additionally, this separation allows the silica to have an increased contact to the YbPO4 particles including dissolution processes from the particles surface. In sum, that turns the spectral properties of our CP-fiber similar to those of an amorphous (MCVD derived) fiber.
4. Conclusions
A method to obtain and sinter nanoparticle mixtures based on SiO2 and YbPO4 has been studied. This methodology is perfectly suitable to be used for the preparation of large volume core preforms for laser active optical fibers. The preparation of silicate glass containing crystals by means of melting a mixture of YbPO4 xenotime structured crystals and SiO2 nanoparticles has been evaluated. We confirmed that the YbPO4 particles can withstand the multi-step fiber fabrication process. However, we observed that the crystals survive only partially the high temperature vitrification process and the subsequent high temperature fiber drawing process, i.e. a second high-temperature process at about 2000 °C just below the melting temperature of YbPO4. This observation corresponds well to reported results on fibers with GeO2- or Al2O3-containing glass matrix prepared by MCVD suspension doping.
The present investigation suggests that in the sintered glass-particle composites as well as in the CP-fiber, Yb3+ ions experience two different phase surroundings. Using XRD, Raman spectra and fluorescence steady-state and decay-measurement data indicate that the two different phase environments are the suspected crystalline surrounding as well as amorphous silica-dominated surrounding.
The size of surviving YbPO4 particles aggregates ranges in the order of the wavelength of light causing scattering. Therefore, the resulting fiber losses are quite high.
Funding
Deutscher Akademischer Austauschdienst (project 57405001); Deutsche Forschungsgemeinschaft (WO 1220/16-1, WO 2540/1-1); Academy of Finland (Academy Project-316483, Photonics Research and Innovation PREIN-320165).
Acknowledgement
We thank Andy Scheffel and Dr. Jan Dellith (Leibniz-IPHT) for assistance in EPMA analysis, and Dr. S. Unger (Leibniz-IPHT) for providing the reference MCVD fibers. Authors Contribution. LP, LW and KW conceived and designed of this study. JK performed the drawing and analysis of fiber, AK conducted the thermal treatment of powder samples. RM, ML, RS, AV, LK and AS collected data and performed the analysis. RM, ML, RS and KW wrote the draft of the manuscript. All authors reviewed and approved it for publication.
Disclosures
The authors declare no conflicts of interest.
Data Availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. A. Langner, M. Such, G. Schötz, V. Reichel, S. Grimm, F. Just, M. Leich, J. Kirchhof, B. Wedel, G. Köhler, O. Strauch, O. Mehl, V. Krause, and G. Rehmann, “Development, manufacturing and lasing behavior of Yb-doped ultra large mode area fibers based on Yb-doped fused bulk silica,” Proc. SPIE 7580, 75802X (2010). [CrossRef]
2. K. Schuster, S. Unger, C. Aichele, F. Lindner, S. Grimm, D. Litzkendorf, J. Kobelke, J. Bierlich, K. Wondraczek, and H. Bartelt, “„Material and technology trends in fiber optics,” Adv. Opt. Techn. 3(4), 447–468 (2014). [CrossRef]
3. M. N. Zervas, “High power ytterbium-doped fiber lasers — fundamentals and applications,” Int. J. Mod. Phys. B 28(12), 1442009 (2014). [CrossRef]
4. H. M. Pask, R. J. Carman, D. C. Hanna, A. C. Tropper, C. J. Mackechnie, P. R. Barber, and J. M. Dawes, “Ytterbium-doped silica fiber lasers: versatile sources for the 1-1.2 µm region,” IEEE J. Select. Topics Quantum Electron. 1(1), 2–13 (1995). [CrossRef]
5. P. Varak, M. Kamradek, J. Mrazek, O. Podrazký, J. Aubrecht, P. Peterka, P. Nekvindova, and I. Kasík, “Luminescence and laser properties of RE-doped silica optical fibers: The role of composition, fabrication processing, and inter-ionic energy transfers,” Opt. Mater.: X 15, 100177 (2022). [CrossRef]
6. S. Nagel, J.B. MacChesnay, and K.L. Walker, “An overview of the modified chemical vapor deposition (MCVD) process and performance,” IEEE Trans. Microwave Theory Tech. 30(4), 305–322 (1982). [CrossRef]
7. S. Unger, F. Lindner, C. Aichele, M. Leich, A. Schwuchow, J. Kobelke, J. Dellith, K. Schuster, and H. Bartelt, “A highly efficient Yb-doped silica laser fiber prepared by gas phase doping technology,” Laser Phys. 24(3), 035103 (2014). [CrossRef]
8. A. Langner, G. Schötz, M. Such, T. Kayser, V. Reichel, S. Grimm, J. Kirchhof, V. Krause, and G. Rehmann, “A new material for high-power laser fibers,” Proc. SPIE 6873, 687311 (2008). [CrossRef]
9. A. Langner, T. Kayser, G. Schötz, M. Such, J. Kirchhof, S. Grimm, and V. Reichel, “Verfahren zur Herstellung von co-dotiertem Quarzglas,” patent DE102007045097A1 (2007).
10. A. Veber, Z. Lu, M. Vermillac, F. Pigeonneau, W. Blanc, and L. Petit, “Nano-structured optical fibers made of glass-ceramics, and phase separated and metallic particle-containing glasses,” Fibers 7(12), 105 (2019). [CrossRef]
11. J. Kirchof, S. Unger, A. Schwuchow, and V. Reichel, “High power laser fibers - comparison of alumi-nium and phosphorus codoping,” OECC/ACOFT conf.2008
12. L. L. Lee and D. S. Tsai, “Ion clustering and crystallization of sol-gel-derived erbium silicate glass,” J. Mater. Sci. Lett. 13(8), 615–617 (1994). [CrossRef]
13. R. C. Breneman and J. W. Halloran, “Kinetics of cristobalite formation in sintered silica,” J. Am. Ceram. Soc. 97(7), 2272–2278 (2014). [CrossRef]
14. W. Blanc, M. Vermillac, L. Petit, A. Lukowiak, Zh. Lu, F. Mady, M. Benabdesselam, S. Chaussedent, A. Mehdi, and M. Ferrari, “Nanoparticles in optical waveguides: a toolbox to promote lasers, amplifiers and sensors,” 21st Int. Conf. on Transparent Optical Networks (ICTON), 1–6 (2019).
15. M. F. Zawrah and E. M. A. Hamzawy, “Effect of cristobalite formation on sinterability, microstructure and properties of glass/ceramic composites,” Ceram. Int. 28(2), 123–130 (2002). [CrossRef]
16. N. Ojha, T. Trautvetter, I. Norrbo, A. Kalide, M. Lastusaari, R. Mueller, and L. Petit, “Sintered silica bodies with persistent luminescence,” Scr. Mater. 166, 15–18 (2019). [CrossRef]
17. A. Lemiere, A. Szczodra, S. Vuori, B. Bondzior, T. W. Hawkins, J. Ballato, M. Lastusaari, J. Massera, and L. Petit, “Bioactive phosphate glass-based fiber with green persistent luminescence,” Mater. Res. Bull. 153, 111899 (2022). [CrossRef]
18. Y. Hikichi and T. Nomura, “Melting temperatures of monazite and xenotime,” J. Am. Ceram. Soc. 70(10), C-252–C-253 (1987). [CrossRef]
19. L. Boatner, “Synthesis, structure, and properties of monazite, pretulite, and xenotime,” Reviews in Mineralogy and Geochemistry 48(1), 87–121 (2002). [CrossRef]
20. L. D. Iskhakova, F. O. Milovich, D. Y. Erin, V. V. Velmiskin, and S. L. Semenov, “Phase separation and crystallization of phosphate–silicate glass cores of preforms of fiber optic,” Glass Phys. Chem. 44(2), 137–144 (2018). [CrossRef]
21. C. Kinowski, H. El Hamzaoui, B. Capoen, G. Bouwmans, A.-M. Blanchenet, K. Delplace, B. Prochet, and M. Bouazaoui, “YbPO4 nano-cylinders formation and alignment within optical fiber preforms using fiber-drawing process,” Mater. Res. Bull. 97, 293–299 (2018). [CrossRef]
22. C. I. Oppo, R. Corpino, P. C. Ricci, M. C. Paul, S. Das, M. Pal, S. K. Bhadra, S. Yoo, M. P. Kalita, A. J. Boyland, J. K. Sahu, P. Ghigna, and F. d’Acapito, “Incorporation of Yb3+ ions in multicomponent phase-separated fibre glass preforms,” Opt. Mater. (Amsterdam, Neth.) 34(4), 660–664 (2012). [CrossRef]
23. E. A. Savelev and K. M. Golant, “Influence of fusing on the uniformity of the distribution of Yb3+ ions and the formation of clusters in silica with phosphorus admixture synthesized by SurfacePlasma CVD,” Opt. Mater. Express 5(10), 2337–2346 (2015). [CrossRef]
24. V. Fuertes, N. Gregoire, P. Labranche, N. Hamada, B. Bellanger, Y. Ledemi, S. Larochelle, and Y. Messaddeq, “Cubic-shaped and rod-shaped YPO4 nanocrystal-doped optical fibers: implications for next generation of fiber lasers,” ACS Appl. Nano Mater. 6(6), 4337–4348 (2023). [CrossRef]
25. W. Blanc and B. Dussardier, “Formation and applications of nanoparticles in silica optical fibers,” J. Opt. 45(3), 247–254 (2016). [CrossRef]
26. A. Veber, T. Salminen, A. Matthes, R. Mueller, K. Wondraczek, and L. Petit, “Synthesis, characterization, and optical properties of ytterbium(iii) phosphates and their incorporation in different glass matrices,” J. Phys. Chem. C 125(1), 702–715 (2021). [CrossRef]
27. N. Ojha, H. Nguyen, T. Laihinen, T. Salminen, M. Lastusaari, and L. Petit, “Decomposition of persistent luminescent microparticles in corrosive phosphate glass melt,” Corros. Sci. 135, 207–214 (2018). [CrossRef]
28. Z. Lu, N. Vakula, M. Ude, M. Cabié, T. Neisius, F. Orange, F. Pigeonneau, L. Petit, and W. Blanc, “YbPO4 crystals in as-drawn silica-based optical fibers,” Opt. Mater 138, 113644 (2023). [CrossRef]
29. S. Unger, A. Schwuchow, S. Jetschke, V. Reichel, A. Scheffel, and J. Kirchhof, “Optical properties of Yb-doped laser fibers in dependence on codopants and preparation conditions,” Proc. of SPIE. 6890, 689016 (2008). [CrossRef]
30. T. Habisreuther, T. Elsmann, Z. Pan, A. Graf, H.J. Pissler, M. Rothhardt, R. Willsch, H. Bartelt, and M.A. Schmidt, “Optical sapphire fiber Bragg gratings as high temperature sensors,” Proc. 8794, Fifth Europ. Workshop on Opt. Fibre Sensors, 87940B (2013).
31. A. Schwuchow, S. Unger, S. Jetschke, and J. Kirchhof, “Advanced attenuation and fluorescence measurement methods in the investigation of photodarkening and related properties of ytterbium-doped fibers,” Appl. Opt. 53(7), 1466–1473 (2014). [CrossRef]
32. J. Kirchhof, S. Unger, A. Schwuchow, S. Jetschke, and B. Knappe, “Spatial distribution effects and laser efficiency in Er/Yb doped fibers,” Proc. SPIE 5350, 222 (2004). [CrossRef]
33. CRS Handbook of Chemistry and Physics, 95th edition, 2014/15 (CRS Press, Inc., 2014).
34. R. J. Bell, N. F. Bird, and P. Dean, “The vibrational spectra of vitreous silica, germania and beryllium fluoride,” J. Phys. C: Solid State Phys. 1(2), 304299 (1968). [CrossRef]
35. S. Unger, A. Schwuchow, J. Dellith, and J. Kirchhof, “Optical properties of ytterbium/aluminium doped silica glasses,” Opt. Mater. Express 10(4), 907–925 (2020). [CrossRef]
36. A.F. Holleman and N. Wiberg, Lehrbuch der Anorganischen Chemie (de Gruyter, 2007).
37. D. Elwell and H.J. Scheel, Crystal Growth from High-Temperature Solutions (Academic Press, 2011).
38. Z. Lu, T. Robine, M. Guzik, M. Bellec, D. Tosi, C. Molardi, F. Pigeonneau, and W. Blanc, “Shaping nanoparticles in optical fibers through thermal engineering,” Proc. SPIE 12142, 121420D (2022). [CrossRef]
