We report on the fluorine doping and codoping of powder-based silica bulk glasses fabricated using REPUSIL (Reactive Powder Sintering of Silica) technology. The maximum doping level of 1.5 mol% of SiF4 is associated with a refractive index decrease by −8 x 10−3 compared to undoped silica and an essential decrease in the transition temperature by about 200 K. Fluorine codoping is eminently suitable for the direct refractive index adjustment of actively-doped silica glass materials (e.g., Al/Yb or Al/Tm). The fluorine codoping of Al/Yb silica glass significantly reduces parasitic photodarkening processes.
© 2015 Optical Society of America
During the past two decades, fiber-based light sources have been widely developed and used in broad application fields . However, the use of high power fiber lasers for surface processing, cutting, welding, drilling, etc. places tremendous demands on the laser host material with regard to refractive index value and homogeneity, as well as dopant concentration and distribution. In times of increasing laser output power, large mode area (LMA)  and extra-large mode area (XLMA)  concepts in particular require a careful adjustment of the refractive index or, more specifically, the numerical aperture in order to implement single-mode (SM) propagation or a beam parameter product (BPP) of at least 5 mm x mrad, respectively [4–6].
There are two main technological approaches to accomplishing large doped cores and low numerical apertures in combination with a pump cladding in so-called double clad fiber designs. On the one hand, the refractive index (RI) of the cladding can be raised by doping the cladding to form a pedestal and to establish a small index gap between the core providing SM behavior and the pump cladding . On the other hand, the cladding may consist of undoped silica, and the active core refractive index has to be reduced by codoping with index decreasing dopants (e.g., boron, fluorine, aluminum/phosphorous at a molar ratio of 1:1) [8, 9]. Another way to decrease the core refractive index is to use the filamentation technique, which can be done, for example, e.g. in large-pitch rod-type fiber lasers. Here, MCVD (modified chemical vapor deposition)-based, Al/Yb-doped silica rods are stacked and drawn down together using index-decreasing doped silica (mainly fluorine-doped silica) [10–12]. The capability of this approach has already been demonstrated . However, in addition to accurate index control of the initial materials, this approach requires several additional fabrication steps because the single filaments have to be smaller in diameter than the application wavelength of the fiber.
The direct incorporation of RI decreasing codopants (e.g. boron, fluorine) applying MCVD technique is limited due to the high volatility of fluorides (e.g. AlF3, REF3, boron oxides etc.). In particular at high temperature steps during the preform consolidation (collapsing) a high material loss can be observed .
Here, we describe an approach to directly decrease the refractive index of undoped and doped silica through the use of REPUSIL technology  via subsequent gas phase doping using SiF4. REPUSIL technology is a method of fabricating a large volume of actively-doped (Yb/Al, Tm/Al, Ce) and passively-doped (Al) silica bulk glasses with a high purity and homogeneity level [3, 6, 14, 15]. With regard to the refractive index, the fluorination of doped silica is a form of counter-doping. The option of counter-doping opens up new possibilities for large and extra-large mode area (LMA and XLMA) fibers with higher rare earth (RE) and Al doping  and a homogeneous dopant distribution. Even the core filamentation technique (separate Al/Yb and F doped rods) becomes easier when homogeneously-doped bulk materials become available.
Another important aspect of the incorporation of fluorine in silica materials is the change in thermochemical properties, such as the decrease in the transformation temperature and the viscosity at elevated temperatures  compared to undoped silica. Our own investigations have shown that, for instance, in a fiber containing 1 mol% of SiF4 the drawing temperature has to be reduced by approximately 200 K.
2. Material preparation and characterization
The fabrication of REPUSIL-based doped or undoped silica glasses has already been described in [14, 15]. The starting material is a high-purity silica nano powder, which is subsequently doped and purified. An intermediate step in the following technology route is the preparation of porous green bodies, which can be directly used for fluorine doping. In this work, we used porous silica green bodies, undoped and pre-doped with Al or Al/Yb. Fluorination can be distinguished into two specific process routes: i) low-temperature SiF4-based fluorination (LTF) applying a temperature of 750°C – 1000°C with a long time process and ii) high-temperature SiF4-based fluorination (HTF) applying a temperature of 1000°C – 1400°C. At higher temperatures, sintering becomes too strong and vitrification sets in. An effective fluorination is no longer possible for bulk materials and is only achievable for thin layers (µm range, commonly deposited during modified chemical vapor deposition (MCVD) [18, 19]). In all cases the samples have been pre-heated to reduce the surface OH, to realize a comparable OH concentration for all samples, and to ensure a constant temperature (no temperature gradient). A change of the sample morphology has not been observed.
Both processes allow the addition of oxygen, which makes for better redox control of the dopants (e.g., Yb2+/Yb3+) .
The investigation of both fluorine concentration and fluorine distribution of vitrified glass samples was done by wavelength dispersive electron probe microanalysis (WD-EPMA) using a JXA8800L (JEOL company) microprobe analyzer. To do this, thinly-polished silica slices were prepared and coated with approximately 20 nm of carbon. Exact radial concentration profiles were taken across the sample diameter for each element contained in the material.
The refractive index of the bulk samples was measured using the Metricon prism coupling device model 2010M (633 nm and 1300 nm). Here, refractive indices are determined from the measurement of the angle incident to the prism at which total reflection on the prism base breaks. The accuracy is about 10−4.
The refractive index profile of the preform samples was measured nondestructively by a deflection angle method with a spatial resolution of about 6 µm.
If fibers were drawn from the preforms, their spectral loss/attenuation behavior was determined using the well-established cut-back method. For this purpose, the light of a combined halogen/deuterium white light source was launched into the fiber. The light transmitted by the fiber was then analyzed using an Instrument Systems Spectro 320D fiber spectrometer. By comparing the transmitted intensity spectra of a larger and a shorter fiber length, the loss/attenuation spectrum was able to be obtained.
3. Results and discussion
3.1 Precursor and process evaluation
All investigations are based on porous SiO2 bulk materials fabricated by the reactive powder sintering process (REPUSIL).
The fluorination of silica can be categorized, for example, by the usage of specific fluorine precursors . In addition to SiF4 or NF3 and sulfur-containing precursors such as SF4 or SF6, carbon-containing precursors such as CF4 or C2F6 can be applied . Irrespective of precursor type, it was shown that SiF4 is always an intermediate product . Therefore, it was obvious to apply SiF4 directly as a fluorination agent. A comparison of different fluorination agents regarding their fluorination capability of undoped silica is demonstrated in Table 1.
Figure 1 shows the influence of fluorination on the change in the refractive index. The index change can be linearly correlated to the SiF4 concentration measured, which is of great importance to material characterization . The SiF4 concentration can be measured by electron probe microanalysis and the refractive index profile estimated.
It becomes clear that SiF4 is the most efficient fluorinating agent. Another aspect is the absence of carbon or sulfur, which can act as a source of contamination (e.g., carbon soot). Furthermore, the application of SiF4 is procedurally easy to control and allows fluorination under different redox conditions. One drawback is the presence of an additional Si carrier, which complicates the exact adjustment of the fluorine concentration. However, the presence of Si in the precursor suppresses a gas phase etching process. A joint characteristic of all fluorination processes is the formation of volatile fluorine-containing compounds. This can lead to an out-diffusing and evaporation of dopants (network modifier) or even matrix material (network former) and requires an exactly-defined process control concerning temperature, gas flow volume, time, partial pressure of the precursor in the carrier gas, and impurities like OH and Cl2, respectively. Table 2shows the achievable refractive index decrease of silica depending on different processing temperatures. At temperatures higher than 900°C, an opposing process of de-fluorination has a remarkable influence, and the fluorine concentration decreases significantly.
The treatment of porous silica with chlorine for cleaning (e.g., to remove 3d elements) and drying (e.g., to remove OH) is a valid method in established preform fabrication techniques . However, our investigations indicate that chlorination and residual Cl2 and/or OH have a remarkable effect on the refractive index decrease that is finally observed (Table 3).
In the process flow of low and high temperature fluorination (LTF and HTF, respectively) we distinguish between two significant mechanisms. The LTF process comprises an addition of both silicon and fluorine, as well as the substitution of oxygen, and has to be considered in the refractive index adjustment. The addition reaction becomes noticeable by a sample weight increase. The substitution of oxygen in the HTF process is superimposed by parallel sintering and increased fluorine diffusion. This results in a faster process, possibly larger sample volumes, and reduced losses of codopants such as Al and Yb. However, in both cases (LTF and HTF) the fluorine concentration in the boundary area (either interface or outer surface) deviates from the sample core area caused by small amounts of fluorine evaporation during vitrification. This process step is separated from the fluorination and is not performed under a complementary SiF4 addition. However, the evaporation of volatile fluorine containing compounds can lead to a parabolic refractive index profile, which will be discussed later.
Each type of fluorination has its specific advantages but calls for detailed process knowledge and controllability: LTF is more energy efficient, suitable for small samples, and allows better and faster process studies; HTF is appropriate for cladding tube vitrification, has a lower tendency for contamination due to cladding tube processing, is suitable for easier preform fabrication, and can be integrated into the process (allows the integration of fluorination as a direct sub-process in preform fabrication). In Table 4 the main characteristics of both fluorination process types are summarized.
3.2 Fluorination of undoped and codoped silica
3.2.1 The SiO2 / SiF4 system
All samples in the SiO2/SiF4 system were treated by applying the LTF (900°C, 3 h). However, from specific conditions different refractive index levels arise. Figure 3 shows the refractive index profiles of several fluorinated silica glass samples fabricated at the same temperature but with specific additional treatments (supplementary O2, no carrier gas, and varying pretreatment of the amorphous silica powder). A refractive index (RI) change down to −8·10−3 was able to be achieved. Depending on the fluorination conditions, the samples show a reverse parabolic RI gradient or a parabolic RIP with an overall refractive index lower than that of undoped silica glass (see Fig. 3). The former is caused by partial de-fluorination, notably in the outer surface zone. The porous samples are additionally preloaded with SiF4 (sample evacuation and refilling with SiF4), which results in a more homogeneous distribution of SiF4. Together with a processing time of about 3 h, we obtained the RI profile shown in Fig. 3 (three lower curves). However, if a more accurate step-index profile is necessary, the outer parts of the preform cylinders with the strongest fluorine gradient can be mechanically removed (grinding, polishing). The amount of material removed depends on the required RI accuracy or the RI deviation allowed, respectively.
Parabolic RIP shaping is caused by a time-wise effect (e.g., 3 h vs. 45 min. of fluorination). If the fluorination process is diffusion based, the reaction of SiF4 with the SiO2 matrix starts from the outer surface. The SiF4 diffuses into the porous body and the fluorination in the center is delayed. A timely cancellation of the process leads to the specific RI shape described.
This effect can be exploited to fabricate gradient index preforms/fibers. This opens up the possibility for the fabrication of bulk gradient index materials (preforms/fibers) without the need for stepwise layer deposition.
The optical attenuation of such a gradient index fiber based on fluorination is shown in Fig. 4.A light-propagating structure originates from the fact that there is a lower fluorine concentration in the fiber center and a higher one in the outer areas. The minimum loss is about 10 dB/km at a wavelength of 1100 nm. Absorptions between 800 and 1000 nm and >1400 nm arise from fiber coating absorptions as the fluorinated material is directly coated with a low refractive index acrylate.
3.2.2 The SiO2 / Al2O3 / SiF4 system
The Al-codoped silica has also been treated according to LTF. A set of samples with up to 4.7 mol% of Al2O3 doping levels was used for fluorination with SiF4. Under the given conditions the results can be stated as follows:
- − There is a small decrease in aluminum concentration during fluorination. The strength of this decrease depends on the absolute Al content (Fig. 5). It is remarkable that under the same fluorination conditions the Al loss decreases with an increasing absolute Al concentration. However, the final F concentration does not depend on the Al content.
- − A stronger decrease in Al on the sample edges (surfaces) can be observed.
- − The final fluorine concentration does not depend on the preset Al level.
As mentioned above, the main motivation for the F doping of Al-codoped silica glass is the fine adjustment of the RI and the compensation of RI increasing caused by codopants such as Al. An extreme example of index compensation is when the index is made to match the pure silica glass. In this case, knowledge about the exact dopant concentration and the changes observed during the fluorination process is of major importance (Fig. 5). Figure 6(a) shows the refractive index evolution of fluorinated Al-codoped silica glass. It can be observed that at up to 2.7 mol% of Al2O3 the Al2O3 caused refractive index change can be completely compensated by fluorination of the material. At higher Al concentrations the refractive index of the material is decreased by fluorination but remains higher than for pure silica. Taking into account a slightly varying SiF4 concentration the coefficient of determination expresses a better linear dependency of Δn on the Al2O3-SiF4 ratio compared to the consideration of only Al2O3 in an Al2O3-SiF4 doped silica glass (Fig. 6(b)).
3.2.3 The SiO2 / Al2O3 / Yb2O3 / SiF4 system
One of the most interesting codoped silica glass materials with regard to high power fiber laser applications is the SiO2-Al2O3-Yb2O3 system. All samples in the system described here are fabricated according to high temperature fluorination (HTF).
The specific adjustment of fluorination conditions allows a high flexibility regarding the RI profile (RIP) and the absolute RI values even for Al2O3-Yb2O3-SiF4 codoped silica glass (Figs. 7(a)-7(d).Whereas the absolute RI value is caused by an incremental contribution of the matrix material (SiO2) and the codopants (Al2O3, Yb2O3, and SiF4), the RI shape is basically dominated by fluorine distribution. By varying the RI profile the guiding properties of the core in LMA fibers may be changed to reduce – for example, using a gradient profile – the bending loss or to maximize the mode field diameter and to generate a flat-top profile in the case of a non-bent fiber using an inverse parabolic RI profile.
Starting from the same initial material (3 mol% of Al2O3 and 0.1 mol% of Yb2O3), different RI shapes and distribution can be implemented under varying (mainly time effected) fluorination conditions (see Fig. 8).For a sample with the codopant concentration described, the absolute maximum RI can be reduced to a third of the initial value (see Fig. 8, which shows long time fluorination).
The Al concentration profile depicted in Fig. 7(c) expresses a stronger signal noise than samples with a lower Al concentration (Fig. 7(a)). Obviously, at Al concentrations of >3 mol%, the Al tends toward phase separation resulting in an increased inhomogeneity. However, the homogeneity can be significantly improved by a subsequent thermal homogenization process which leads to a diffusion-based offset of dopants (Fig. 9).
Several samples of fluorine-doped and fluorine-free materials have been processed into optical fibers. This opened up the possibility for an estimation of optical losses based on absorption, scattering, and RI effects (e.g., bending sensitivity).
Figure 10 shows the optical loss of two selected multimode fiber samples fabricated from the same starting material (3 mol% of Al2O3 and 0.1 mol% of Yb2O3, a non-fluorinated and a fluorinated fiber core, respectively; the latter shows a parabolic RIP diagramed in Fig. 8). It becomes clear that fluorination has a significant effect on the UV edge (blue shift), the OH content (decrease by about one order of magnitude), and the basic loss. This loss increases because of a higher bending sensitivity which origins from a lower numerical aperture of the fluorinated sample (0.09 vs. 0.14, values based on equivalent step index (ESI) profile).
However, at high doping levels (for shorter fibers and the suppression of nonlinear processes), the photodarkening (PD) issue becomes more relevant. The high Al doping is necessary for a PD decrease as the PD characteristics are significantly changed by a variation in the Yb-Al ratio . The fluorination of highly Al and Yb-doped silica glass makes it possible to compensate the increased refractive index and to adjust the numerical aperture for an optimized beam quality. Even F doping itself has a significant effect on PD losses (see Fig. 11). To the best of our knowledge there is no evidence to date for this effect. However, this observation is a first verification and needs additional proofs which are currently under investigation. Another advantage is that the absorption cross section remains constant compared to pure SiO2/Al2O3/Yb2O3 systems but is significantly reduced in phosphorous or aluminum/phosphorous codoped materials without any fluorine addition [9, 25].
It was shown that the fluorination of pure or Al2O3 and Al2O3/Yb2O3-codoped silica using SiF4 is easily accessible for powder-based materials. Two process routes have been distinguished: low temperature fluorination (LTF) and high temperature fluorination (HTF). The main process parameters (temperature, time, type of precursor, carrier gas and gas flow) have been investigated and the repercussion on the final fluorine content, the refractive index value, and distribution were discussed. Depending on the fluorination conditions, the materials/preforms can exhibit a parabolic or inverse parabolic refractive index profile.
As the passively-doped materials are available as bulk materials they can be used for further processing in fiber fabrication by means of stack-and-draw technology. Fluorinated, actively-doped silica glass is also available as a laser material. The latter will be a matter of future investigations.
The financial and organizational support by the Free State of Thuringia and the European Social Fund (ESF) under the contract 2011 FGR 0104 (FG FaserTech) is gratefully acknowledged.
The work was also supported by the Federal Ministry of Education and Research (BMBF) under the contracts 13 N 9555 (FALAMAT) and 13 N 12712 (REMILAS).
References and Links
1. D. J. Richardson, J. Nilsson, and W. A. Clarkson, “High power fiber lasers: current status and future perspectives,” JOSA B 27(11), B63–B92 (2010).
2. A. Tünnermann, H. Zellmer, H. W. Schöne, A. Giesen, and K. Contag, “New concepts for diode pumped solid state lasers in high power diode lasers,” in High-Power Diode Lasers, R. Diehl, ed. (Springer 2000).
3. A. Langner, M. Such, G. Schötz, F. Just, M. Leich, A. Schwuchow, S. Grimm, H. Zimer, M. Kozak, B. Wedel, G. Rehmann, C. Bachert, and V. Krause, “Multi-kW single fiber laser based on an extra large mode area fiber design,” Proc. SPIE 8237, 82370F (2012). [CrossRef]
4. J. Limpert, T. Clausnitzer, A. Liem, T. Schreiber, H.-J. Fuchs, H. Zellmer, E.-B. Kley, and A. Tünnermann, “High-average-power femtosecond fiber chirped-pulse amplification system,” Opt. Lett. 28(20), 1984–1986 (2003). [CrossRef] [PubMed]
6. A. Langner, M. Such, G. Schötz, F. Just, M. Leich, S. Grimm, J. Dellith, M. Jäger, H. Zimer, M. Kozak, B. Wedel, G. Rehmann, C. Bachert, and V. Krause, “Design evolution, long term performance and application tests of extra large mode area (XLMA) fiber lasers,” Proc. SPIE 8601, 86010G (2013). [CrossRef]
7. S. Yao, A. S. Webb, A. J. Boy, R. J. Standish, A. Dhar, and J. K. Sahu, “Linearly polarized ytterbium-doped fiber laser in a pedestal design with aluminosilicate inner cladding,” Laser Phys. Lett. 8(6), 453–457 (2011). [CrossRef]
9. S. Jetschke, S. Unger, A. Schwuchow, M. Leich, and J. Kirchhof, “Efficient Yb laser fibers with low photodarkening by optimization of the core composition,” Opt. Express 16(20), 15540–15545 (2008). [CrossRef] [PubMed]
10. J. Kirchhof, S. Unger, J. Kobelke, K. Schuster, K. Mörl, S. Jetschke, and A. Schwuchow, “Materials and technologies for microstructured high power laser fibers,” Proc. SPIE 5951, 595107 (2005). [CrossRef]
11. C. Wirth, O. Schmidt, A. Kliner, T. Schreiber, R. Eberhardt, and A. Tünnermann, “High-power tandem pumped fiber amplifier with an output power of 2.9 kW,” Opt. Lett. 36(16), 3061–3063 (2011). [CrossRef] [PubMed]
12. H. R. Müller, J. Kirchhof, V. Reichel, and S. Unger, “Fibers for high power lasers and amplifiers,” C. R. Phys. 7(2), 154–162 (2006). [CrossRef]
13. J. Limpert, T. Schreiber, S. Nolte, H. Zellmer, T. Tünnermann, R. Iliew, F. Lederer, J. Broeng, G. Vienne, A. Petersson, and C. Jakobsen, “High-power air-clad large-mode-area photonic crystal fiber laser,” Opt. Express 11(7), 818–823 (2003). [CrossRef] [PubMed]
14. 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. Technol. 3(4), 447–468 (2014).
15. M. Leich, F. Just, A. Langner, M. Such, G. Schötz, T. Eschrich, and S. Grimm, “Highly efficient Yb-doped silica fibers prepared by powder sinter technology,” Opt. Lett. 36(9), 1557–1559 (2011). [CrossRef] [PubMed]
16. W. He, M. Leich, S. Grimm, J. Kobelke, Y. Zhu, H. Bartelt, and M. Jäger, “Very large mode area ytterbium fiber amplifier with aluminum-doped pump cladding made by powder sinter technology,” Laser Phys. Lett. 12(1), 015103 (2015). [CrossRef]
17. M. Kyoto, Y. Ohoga, S. Ishikawa, and Y. Ishiguro, “Characterization of fluorine-doped silica glasses,” J. Mater. Sci. 28(10), 2738–2744 (1993). [CrossRef]
18. J. Kirchhof and S. Unger, “Thermodynamics of fluorine incorporation into silica glass,” J. Non-Cryst. Solids 354(2-9), 540–545 (2008). [CrossRef]
19. A. N. Guryanov, M. Yu. Salganskii, V. F. Khopin, A. F. Kosolapov, and S. L. Semenov, “High-Aperture Optical Waveguides Based on Fluorine-Doped Silica Glass,” Inorg. Mater. 45(7), 823–826 (2009). [CrossRef]
20. J. Kirchhof, S. Unger, A. Schwuchow, S. Grimm, and V. Reichel, “Materials for high-power fiber lasers,” J. Non-Cryst. Solids 352(23-25), 2399–2403 (2006). [CrossRef]
21. F. Froehlich, C. Aichele, S. Grimm, and K. Schuster, “Fluorine incorporation into porous silica by gas phase doping with C2F6 in N2,” Opt. Mater. Express 3(11), 1839–1854 (2013). [CrossRef]
22. A. Langner, G. Schötz, M. Such, M. Schwerin, M. Trommer, S. Grimm, A. Kalide, M. Leich, and F. Just, Patent DE 10 2013 204 815 A1 (2014).
24. R. Clasen, “Preparation and sintering of high-density green bodies to high-purity silica glasses,” J. Non-Cryst. Solids 89(3), 334–344 (1987). [CrossRef]
25. J. Kirchhof, S. Unger, A. Schwuchow, S. Jetschke, and B. Knappe, “Dopant interactions in high-power laser fibers,” Proc. SPIE 5723, 261–272 (2005). [CrossRef]