Influence of fusing on the uniformity of the distribution of Yb<sup>3+</sup> ions and the formation of clusters in silica with phosphorus admixture synthesized by SPCVD

E.A. Savelev; K.M. Golant

doi:10.1364/OME.5.002337

1. Introduction

The uniformity of the distribution of an activator in a host material is of great importance for the characteristics of solid-state lasers. The problem of doping uniformity becomes especially important in the case of large concentrations of a poorly soluble activator incorporated into the host material due to increases in the cooperative quenching effects of the excited laser level.

The solubility of rare earth oxides in silica is low. However, due to the ability to manufacture silica-based optical fibers and fiber lasers, the application of amorphous silicon dioxide as a host for rare earth oxides is of considerable interest. Adding aluminum and/or phosphorous to silica increases the solubility of rare earth elements [1]. Even in this case, the maximum concentration of rare-earth activators, which would not lead to degradation of the laser parameters, remains significantly small compared to phosphate and tellurite glasses [2–4]. The latter have much lower softening temperatures compared to silica but are less suitable for fiber fabrication. For instance, P₂O₅ has a melting point 562 °C and TeO₂ has a melting point of 733 °C, but the melting point of SiO₂ is 1713 °C [5].

Among the rare earth activators, ytterbium (Yb) has received the most attention. Since the first examples of Yb³⁺ fiber lasers, their maximum output power demonstrated in experiments has constantly grown and has reached ~10 kW for single mode operation [6]. The possibility of this success is due to the relative simplicity of the energy spectrum of the 4f-electrons in Yb³⁺ ions, which eliminates many unwanted ways for electron relaxation from a metastable to the ground state [7] that are intrinsic to other rare earth ions. The absence of alternatives for upconversion transitions in the energy spectrum as well as the small value of the quantum defect in the photoexcitation of the metastable state enable the characteristics of Yb-doped laser materials [8]. At the same time, to create more powerful fiber and waveguide lasers, it is desirable to further increase the concentration of ytterbium in glass, which would reduce the length of the active part of the waveguide and thus reduce the negative influence of the nonlinear effects, which are significant at high densities of optical power.

In thermodynamic equilibrium technologies, the increase of rare earth oxide concentrations in silica is limited by the process of clustering and the increasing probability of cooperative relaxation processes, which in most cases have adverse effects [9]. Therefore, the search for technology to allow a more homogeneous distribution of laser ions in silica at higher concentrations is a topical task.

The plasmachemical technology of silica synthesis by surface-plasma microwave discharge at low pressure, SPCVD [10], allows transparent, uniformly-doped amorphous silicon dioxide. The glass network formation in the SPCVD technology is the result of chemosorption of diatomic molecules of oxides synthesized in the plasma from halides at substrate temperatures below the temperature of the corresponding glass melt. In SPCVD, a glassy layer is formed, avoiding the stage of melting. It has been shown previously [11, 12] that for erbium ions embedded in silica by SPCVD, there is a subsequent fusing stage, which leads to the formation of the clusters, which was confirmed in experiments on the impact of fusing on the spectra and kinetics of the main (²I_13/2 → ²I_15/2) and upconversion luminescence of Er³⁺ ions.

In this paper, we present and discuss experiments on the impact of the fusing stage on the light scattering and absorption in an Yb-doped, silica host with phosphorus, synthesized by SPCVD. A critical point for the type of structure formation of phosphates in general and ytterbium phosphate in particular is the relative concentrations of oxygen and phosphorus [13]. If the oxygen-to-phosphorous ratio is greater than 4, YbPO₄ is formed. Previous study of silica fabricated by MCVD showed that the relative content of Yb and P does not affect the structure of clusters formed during the heat treatment of preforms [14]. In all cases, the crystals formed were YbPO₄.

The study of nanocrystals $L n P O_{4} \cdot x H_{2} O$ (Ln = Y, La-Nd, Sm-Lu) fabricated by the hydrothermal method showed that the increase of the $P O_{4}^{3 -} / L n^{3 +}$ ratio, including the case where Ln = Yb, leads only to acceleration of the process of $L n P O_{4} \cdot x H_{2} O$ crystal structure formation [15].

Thus, based on the results of previous studies, the formation of YbPO₄ crystals in our samples is expected. The important point, however, is to estimate the role of the fusing stage in this process.

Section 2 summarizes technological details on sample fabrication and the methods used for their characterization. Rayleigh scattering analysis of the measured loss spectra of slab lightguides before and after fusing is presented in Section 3. Evolution of clusters size and crystallinity caused by material fusing is considered in Section 4.

2. Experimental

For the experiments, we fabricated samples in the form of channel lightguides. To this end, using the SPCVD method, double- and triple-layer structures were synthesized on the inner surface of a substrate silica tube with an 18 mm in outer diameter and a wall thickness of 1.5 mm. Samples of the channel lightguides were obtained by longitudinal cutting of slabs from the tube wall together with three (for type ‘#1’ samples) or two (for type ‘#2’samples) layers of amorphous silicon dioxide successively deposited on the inner surface. The first and the third layers of 40 μm in thickness contain fluorine to reduce the refractive index of the glass. The second silica layer of ~150 μm in thickness contains ytterbium and phosphorus. Plasmachemical synthesis of the layers was conducted with the help of a plasma column scanning at a frequency of 8 Hz along the section of a substrate tube ~25 cm in length. A vapor-gas mixture of SiCl₄ + CF₄ + O₂ and SiCl₄ + POCl₃ + YbCl₃ + O₂ was used in the synthesis of the layers of the reflecting claddings and core, respectively. The temperature of the substrate tube wall during the deposition was approximately 1150 °C.

Two opposite side planes and edges of the slabs cut from the tubes (Fig. 1) were polished and refinished to better than 0.1 μm, and the length and width of the obtained channel lightguides were 2 ± 0.1 cm and 150 ± 15 µm, respectively. The first group of channel lightguides (hereinafter designated “uf” for unfused) were cut from the tube with a non-fused layer structure of glass, which was formed immediately in the course of the plasmachemical deposition. The second group of samples (hereinafter referred to as “f” for fused) was obtained from the same tube, but after outside heat penetration at a surface temperature of 1500-1700°C while spinning in the flame of a longitudinally moving hydrogen-oxygen burner.

Fig. 1 Fabrication of the channel waveguide with an activated core (sample of type # 1). D = 18 mm, d = 1.5 mm, L = 20 mm, x = 150 μm, a = 40 μm, b = 150 μm.

Download Full Size | PDF

The transmission spectra of the channel lightguides were collected using an LS-1 (Ocean Optics) halogen lamp in the visible and near-infrared regions and a DDS-30 (Lomo) deuterium discharge lamp in the UV part of the spectrum. Radiation from the lamps was transferred by lenses in a section of standard optical fiber with a pure silica core 105 μm in diameter and fluorine-doped silica reflection cladding, face-to-face connected to the edge of a lightguide. Radiation from an opposite edge of the lightguide was delivered to the Avantes CCD spectrometer using a piece of a KU-type pure-silica core optical fiber with a core diameter of 20 μm.

X-ray microprobe analysis and SEM images of the samples were obtained with a Quanta-200 electron microscope equipped with an energy dispersion EDAX spectrometer. The X-ray structure analysis and TEM images of the specimens were obtained with a JEOL JEM-2100 electron microscope. The work was performed in the core facility of the Moscow Institute of Physics and Technology.

In Figs. 2(a)-2(c) and 3(a) and 3(b), the SEM images of the side lightguide planes for samples of types #1 and #2, respectively, are presented. Lighter strips in the figures correspond to a higher concentration of Yb with respect to the relatively darker areas. The surfaces of the type #1 samples (particularly the “uf”) have white dots approximately 0.1 μm in size due to surface contamination associated with the process of polishing.

Fig. 2 SEM images of the active layer of type #1 samples: a) - before fusing (uf), b) - after fusing (f), c) - the line of dopant concentrations for sample #1f (the full image width corresponds to 1 μm) d) - concentration profiles along the white line.

Download Full Size | PDF

Fig. 3 SEM image of the active layer of type #2 samples: a) – before fusing (uf), b) - after fusing (f).

Download Full Size | PDF

Table 1 summarizes the results of the X-ray microprobe analysis of the samples performed in the areas corresponding to the enlarged images of Figs. 2 and 3. Despite the significant difference in the concentrations of Yb and P in the different samples, the ratio of their molar concentrations in all samples is close to unity. Some of the differences in the chemical composition of samples before and after fusing are caused by variation of the dopant contents along the substrate tube during deposition.

Table 1. The chemical composition of the samples

View Table

In our experiments, samples of type #1, with a higher concentration of ytterbium, were used for the X-ray analysis, whereas samples of type #2, which have a slightly thicker lightguiding core, were used for measuring the optical characteristics.

Figure 2(d) shows a profile of the relative changes in the concentrations of Yb and P measured along the line marked in Fig. 2(b). This measurement was conducted only for the type #1f sample. Because the diameter of the electron beam probe exceeds the cluster sizes, the X-ray microprobe analysis of the composition on a scale <100 nm does not give accurate values of the elements at each particular point but allows one to observe their signatures along the sample. When building the curves of Fig. 2(d), the intensity of the characteristic lines for Yb and P was normalized to an average corresponding to the tested area, which can be considered to be proportional to the average concentration of the dopants in the sample.

Figure 2(d) shows that the changes in the concentrations of Yb and P are correlated, and their maximum values correspond to the center of the cluster. As mentioned above, the chemical composition (Table 1) indicates that the average molar concentrations of Yb and P in the deposited layer are similar. From the trace of Fig. 2(d), one can conclude that the 1:1 relationship between ytterbium and phosphorus contents remains in the cluster, so it would be logical to assume that the cluster composition is YbPO₄.

3. Loss spectra, Rayleigh scattering

Figure 4 shows the spectral dependence of the attenuation coefficients, $α$ expressed over the logarithm of the ratio of the intensities in the output (I) and in the input (I₀) of the lightguide:

α = - (10 / L) \cdot \lg (I / I_{0}),

where L is the length of the lightguide. In the 320-1050 nm wavelength band, there are two main sources of loss: an absorption band of Yb³⁺ ions and scattering. Fusing significantly increases the scattering loss, and this increase is associated with an increase in size of the cluster inclusions in the glass, as shown in Figs. 2(b) and 3(b). After fusing, the shape of the absorption spectrum of the Yb³⁺ ions is slightly changed, indicating a change in the nearest neighbor surroundings of the ions.

Fig. 4 Loss spectra of type #2 samples in the spectral band from 320 to 1050 nm.

Download Full Size | PDF

We now evaluate the change of the cluster size caused by fusing. Assume all clusters to be homogeneous spherical particles of identical composition, and their sizes on average are significantly less than 320 nm: the smallest wavelength of the investigated optical band. Then, to estimate the sizes of the scattering centers, we can use the Rayleigh approximation [16], in which:

С_{s} = \frac{8}{3} π k_{m e d}^{4} {| f |}^{2},

f = \frac{3}{4 π} \frac{m^{2} - 1}{m^{2} + 2} V,

where

С_{s}

is the scattering cross-section,

f

is the polarizability of the scattering cluster, m is the relative refractive index of a transparent substance of the scattering cluster with respect to the environment,

k_{m e d} = 2 π / λ_{m e d}

(the wavenumber),

λ_{m e d}

is the wavelength of light in the surrounding matter in which the scattering cluster is located, and V is the volume of the scattering cluster. In this approximation, the attenuation coefficient can be expressed as follows:

α = γ 10 \lg (e), γ = С_{s} N = 24 π^{3} n_{m e d}^{4} {(\frac{m^{2} - 1}{m^{2} + 2})}^{2} \frac{N V^{2}}{λ^{4}},

where

n_{m e d}

is the refractive index,

λ

is the wavelength of light in a vacuum, and

N

is the concentration of clusters. By setting

N = N_{Y b} / n_{0} = N_{Y b} V_{0} / V,

where

n_{0} = V / V_{0}

is the number of ions in one cluster,

V_{0}

is the volume per one Yb³⁺ ion in a cluster, and

N_{Y b}

is an average concentration of the Yb³⁺ ions in the material, we obtain:

α = 240 \lg (e) π^{3} n_{m e d}^{4} {(\frac{m^{2} - 1}{m^{2} + 2})}^{2} V_{0} \frac{N_{Y b}}{λ^{4}} V,

or:

α = \frac{С}{λ^{4}},

where

С = 240 \lg (e) π^{3} n_{m e d}^{4} {(\frac{m^{2} - 1}{m^{2} + 2})}^{2} V_{0} N_{Y b} V .

In Fig. 5, the loss spectra of the lightguides and their fitting by Eq. (6) are presented in the logarithm-logarithmic scale in the wavelength band from 320 to 750 nm. The contribution of the absorption by Yb³⁺ ions to loss is negligible for this wavelength band. According to Eq. (5), the attenuation coefficients for samples with similar concentrations of Yb³⁺ ions are directly proportional to the volume of the scattering clusters. This means that the growth of clusters leads to an increase of the attenuation coefficient of the Rayleigh scattering, which can be used to estimate the cluster size. If $m,$ $n_{m e d}$ and $V_{0}$ in the #2f and #2uf samples are similar, i.e., the clusters vary only in volume but do not have significant differences in the chemical composition or the structure, Eq. (7) for fused and unfused can be rewritten as follows:

С_{u f} = 240 \lg (e) π^{3} n_{m e d}^{4} {(\frac{m^{2} - 1}{m^{2} + 2})}^{2} V_{0} N_{Y b}^{u f} V_{u f}

С_{f} = 240 \lg (e) π^{3} n_{m e d}^{4} {(\frac{m^{2} - 1}{m^{2} + 2})}^{2} V_{0} N_{Y b}^{f} V_{f}

Dividing (8) by (9) it is easy to obtain:

Fig. 5 Loss spectra of type #2 samples and their fitting by Eq. (6). Symbol size rightly indicate the uncertainties of the measurement.

Download Full Size | PDF

\frac{V_{f}}{V_{u f}} = \frac{С_{f} N_{Y b}^{u f}}{С_{u f} N_{Y b}^{f}}

4. Results and discussion

Figure 6 shows a TEM image of sample #1f. Clusters with a high concentration of ytterbium and phosphorus are clearly seen, with the size of the largest of them exceeding 130 nm. As expected, some clusters are crystallized (Fig. 6(b)). This is evidenced by a set of periodically-placed, straight, dark and light lines separated by a distance of ~5 $Å$ on the TEM image. YbPO₄ has a tetragonal crystal system with lattice parameters ‘a’ and ‘b’ equal to 6.816 $Å$ and 5.966 $Å$ respectively [17]. The second value is close to the period estimated from the image of Fig. 6(b). Even in the “crystallized” clusters, the lattice in the TEM image can be seen only when the crystal planes are “rightly” (in parallel) positioned with respect to the electronic beam. Therefore, we cannot conclude whether all of the clusters or only a part of them are in the crystalline phase. No correlation was found between the size of the observed clusters and the presence of crystalline phases. According to the boundary lines distinguishable in the TEM image of Fig. 6(b), there are at least 8-10 individual crystalline components that participated in the formation of the cluster.

Fig. 6 TEM - images of sample #1f: a) - a region with the largest concentration of Yb, b) – the crystallized cluster (the straight, white line emphasizes the periodic structure in the image, which confirms the crystallinity).

Download Full Size | PDF

The electron transmission diffraction pattern of a region with clusters is shown in the inset of Fig. 7. The diffraction maxima in the form of bright points specific to a crystalline phase, along with a bright wide halo characteristic of the amorphous phase of silicon dioxide, are clearly seen. Using this electron transmission diffraction pattern, we calculated the angular positions for the major diffraction maxima of the scattered electrons (Fig. 7). Data for the X-ray diffraction for the YbPO₄ crystal from [18] are added to the same plot. The angular position of the main peaks within the margin of error is the same as in the electron transmission diffraction pattern. There are noticeable differences in the relative intensities of the bands. This discrepancy may be due to following:

Fig. 7 Angular dependence of the diffraction intensity for sample #1f and YbPO₄ crystal [18]. Inset: electron transmission diffraction pattern of sample #1f.

Download Full Size | PDF

1. Maxima in the directions of 20° and 26° overlap with a broad region associated with amorphous SiO₂, which prevents the identification of the diffraction peaks.
2. The peak in the direction closest to 35° includes 3 diffraction peaks of YbPO₄: 33°, 35° and 37°, which are difficult to resolve because of their small diffraction intensity.
3. In the calculations, we took into consideration not the intensity but the total number of diffraction maxima close to a particular angular direction. For simplicity, we considered the intensities of all the diffraction maxima to be equal. The latter, generally speaking, does not correspond to reality and adds uncertainty to the estimation of the diffraction peak intensities.

The resolving power of the apparatus used is insufficient to provide reliable data for unfused sample, so the question remains whether the clusters in unfused material are in a crystalline or an amorphous phase. We now proceed to estimate the cluster size changes caused by fusing.

To do this, we use Eq. (10) and fit the results for the plots in Fig. 5 (values of $С_{f}$ and $С_{u f}$ ) and the data from the chemical analysis of our samples (values of $N_{Y b}^{u f}$ and $N_{Y b}^{f}$ form Table 1). We obtain $V_{f} / V_{u f} = 13.3 \pm 1.3.$ The present evaluation shows that the process of fusing leads to a 12- to 14- fold increase of the average cluster volume. As noted above, according to the TEM image (Fig. 6(b)), crystallized clusters consist of 8-10 parts; this finding is in an agreement with our estimates performed on the basis of the measured Rayleigh scattering spectra before and after fusing.

5. Conclusion

Using SPCVD technology, channel optical lightguides with a silica core doped with ytterbium and phosphorus were fabricated. The immediate plasma chemical deposition of such glass in the conditions of the SPCVD process yields an amorphous layer with inclusions in the form of clusters containing ytterbium and phosphorus. The fusing of this layer leads to a thirteen fold increase in the average cluster size, causing a sharp rise in the attenuation coefficient in the short wavelength part of the spectrum. At least some of the clusters are not amorphous and have a crystalline structure. Analysis of the electron transmission diffraction pattern and chemical composition suggests that the primary substance in these clusters is crystalline YbPO₄.

Acknowledgment

This work was supported by Russian Foundation for Basic Research, project 14-29-08170.

References and links

1. K. Arai, H. Namikawa, K. Kumata, T. Honda, Y. Ishii, and T. Handa, “Aluminum or phosphorus co-doping effects on the fluorescence and structural properties of neodymium-doped silica glass,” J. Appl. Phys. 59(10), 3430–3436 (1986). [CrossRef]

2. M. J. Weber, J. D. Myers, and D. H. Blackburn, “Optical properties of Nd³⁺ in tellurite and phosphotellurite glasses,” J. Appl. Phys. 52(4), 2944–2949 (1981). [CrossRef]

3. J. Wang, E. Vogel, and E. Snitzer, “Tellurite glass: a new candidate for fiber devices,” Opt. Mater. 3(3), 187–203 (1994). [CrossRef]

4. G. Turri, V. Sudesh, M. Richardson, M. Bass, A. Toncelli, and M. Tonelli, “Temperature-dependent spectroscopic properties of Tm³⁺ in germanate, silica, and phosphate glasses: A comparative study,” J. Appl. Phys. 103(9), 093104 (2008). [CrossRef]

5. D. R. Lide, W. M. Haynes, G. Baysinger, L. I. Berger, M. Frenkel, R. N. Goldberg, H. V. Kehiai, K. Kuchitsu, D. L. Roth, and D. Zwillinger, CRC Handbook of Chemistry and Physics, 90th Edition (CRC Press, 2010).

6. V. Fomin, M. Abramov, A. Ferin, A. Abramov, D. Mochalov, N. Platonov, and V. Gapontsev, “10 kW single-mode fiber laser,” presented at 5th International Symposium on High-Power Fiber Lasers and Their Applications, St. Petersburg, June 28-July 1, 2010.

7. C. B. Layne, W. H. Lowdermilk, and M. J. Weber, “Multihonon relaxation of rare-earth ions in oxide glasses,” Phys. Rev. B 16(1), 10–20 (1977). [CrossRef]

8. L. D. DeLoach, S. A. Payne, L. L. Chase, L. K. Smith, W. L. Kway, and W. F. Krupke, “Evaluation of absorption and emission properties of Yb³⁺ doped crystals for laser applications,” IEEE J. Quantum Electron. 29(4), 1179–1191 (1993). [CrossRef]

9. F. Auzel, “Upconversion and Anti-Stokes Processes with f and d Ions in Solids,” Chem. Rev. 104(1), 139–174 (2004). [CrossRef] [PubMed]

10. K. M. Golant, “Surface plasma chemical vapor deposition: 20 years of application in glass synthesis for lightguides (a review),” in XXI International Congress on Glass, Strasbourg, Proc. on CD ROM, paper L13 (2007).

11. A. V. Kholodkov and K. M. Golant, “Er³⁺ Ion photoluminescence in silicate glasses obtained by plasma-chemical deposition in a low-pressure microwave discharge,” Tech. Phys. 50(6), 719–726 (2005). [CrossRef]

12. A. V. Kholodkov, K. M. Golant, and L. D. Iskhakova, “Peculiarities of Er³⁺ Photoluminescence in Halogen-Doped Amorphous Silica,” Phys. Wave Phenom. 17(3), 155–164 (2009). [CrossRef]

13. H. Y.-P. Hong, “The Crystal Structure of Ytterbium Metaphosphate, YbP3O9,” Acta Crystallogr. B 30(7), 1857–1861 (1974). [CrossRef]

14. 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. 34(4), 660–664 (2012). [CrossRef]

15. Z. Huo, C. Chen, D. Chu, H. Li, and Y. Li, “Systematic synthesis of lanthanide phosphate nanocrystals,” Chemistry 13(27), 7708–7714 (2007). [CrossRef] [PubMed]

16. H. d. Hulst, Light Scattering by Small Particles (Dover Publications, 1981).

17. W. O. Milligan, D. F. Mullica, G. W. Beall, and L. A. Boatner, “Structures of ErPO₄, TmPO₄, and YbPO₄,” Acta Crystallogr. C 39(1), 23–24 (1983). [CrossRef]

18. Z. Xu, P. Ma, C. Li, Z. Hou, X. Zhai, S. Huang, and J. Lin, “Monodisperse core-shell structured up-conversion Yb(OH)CO₃@YbPO₄:Er³+ hollow spheres as drug carriers,” Biomaterials 32(17), 4161–4173 (2011). [CrossRef] [PubMed]

Specimen	Yb content, mol. %	P content, mol. %
#1uf	1.41 ± 0.04	1.59 ± 0.07
#1f	1.25 ± 0.04	1.34 ± 0.07
#2uf	0.51 ± 0.03	0.55 ± 0.03
#2f	0.45 ± 0.03	0.53 ± 0.03

Influence of fusing on the uniformity of the distribution of Yb³⁺ ions and the formation of clusters in silica with phosphorus admixture synthesized by SPCVD

Abstract

1. Introduction

2. Experimental

3. Loss spectra, Rayleigh scattering

4. Results and discussion

5. Conclusion

Acknowledgment

References and links

Cited By

Figures (7)

Tables (1)

Equations (10)

Optical Materials Express