Optical properties of a Bismuth-doped pure silica sol-gel core photonic crystal fiber (PCF) were investigated. We report on the absorption, CW luminescence and time resolved luminescence spectra at different excitation wavelengths at room temperature. Complex structure of the energy levels of Bismuth-connected centers in pure silica glass is put in evidence.
©2010 Optical Society of America
Reported for the first time in , the broadband Near InfraRed (NIR) PhotoLuminescence (PL) in Bismuth-doped silica-based glasses has attracted much attention due to the potential applications in fiber lasers and amplifiers in the spectral range of 1150 – 1500 nm. However, the performances of fiber lasers operating in this wavelength (WL) region reported up to now are still limited. For instance, the slope efficiency of 32% has been reported at 1160 nm  and 14% at 1500 nm  at room temperature. Obviously, the problem of laser efficiency is closely connected to the fact that the nature of Bismuth-related luminescent centers is unclear up to now. It implies that the optimization of glass composition, choice of precursor and technology for fiber fabrication remain purely empirical. NIR PL has been reported in a wide variety of glass matrices and crystals (mainly with multicomponent composition) and the nature of the luminescent center in Bismuth-doped glasses and crystals has also been largely discussed. Besides, it has been shown that even in a pure silica glass broadband NIR PL could be observed  and, furthermore, the luminescent properties in a such simple glass composition are still complex because of the probable existence of different luminescent centers . Due to the fact that pure silica glass composition relatively simplify the analysis and can favor a center emitting in the 1400 nm WL range, in present paper we investigate the luminescent properties of a pure silica Bismuth-doped photonic crystal fiber (PCF) with a core realized by the sol-gel technique. Absorption and PL properties of the fiber together with assigning of absorption and PL bands are discussed.
2. Results and discussion
The process of monolithic Bismuth-doped sol-gel preform preparation was similar to the one described in our recent article . In present experiment, Bismuth concentration in silica glass was increased up to 300 ppm. As an illustration, optical absorption of a 450 ppm Bismuth-doped sol-gel preform sintered at 1300°C is shown in Fig. 1.
As can be seen, the absorption spectrum is similar to the one presented in reference  and puts in evidence two main absorption bands centered around 380 nm and 420 nm. It appears that these bands are a complex bands composed of three sub-bands as it is suggested by the Gaussian multi-peak fit shown in Fig. 1 (here the background absorption was fitted as a polynome of third order).
In order to use this pure silica preform as a fiber core, an air/silica PCF has been realized using the conventional stack and draw process . During the fabrication, the sol-gel preform has been heated several times to high temperatures. At the first step it was fused at both ends to Suprasil F300 tubes (Heraeus Tenevo LLC) in order to make it long enough for our drawing facilities. Then it was drawn into 1.6mm diameter rods at about 2000°C. After that, the 25mm diameter stack containing this rod in its center has been drawn into 4mm canes at a furnace temperature of 2060°C. At the last step, one of these canes has been sleeved into a 8mm diameter tube and drawn into the final fiber at 2000°C.
This fiber (SEM image presented in Fig. 2) has an outside diameter of 125 µm and a core diameter of 6.4 µm defined as the distance between two diametrally opposed holes. The pitch of the periodic cladding, Λ, and the diameter of the air holes, d, are 4.0 µm and 1.6 µm, respectively. These values lead to the ratio d/Λ below 0.42 so that the fiber can be considered as endlessly singlemode .
To quantify background losses and Bismuth-related absorption, attenuation measurements have been performed using the conventional cutback method. Supercontinuum white light source (FIANIUM SC-400) and tungsten lamp-based white light source (ANDO AQ-4303B) have been alternatively used together with Optical Spectrum Analyzer (OSA, ANDO AO-6315A). A first cutback has been performed on a 10 m-long sample in order to estimate the background losses. A minimum loss value of about 0.9 dB/m was measured at 1090 nm (inset of Fig. 3), which was about 75 times larger than the value obtained at the same WL with an un-doped pure silica sol-gel core PCF. This measurement also allowed us to quantify OH content by estimating the absorption around 1245 nm. The amplitude of this OH-related band was found to be 29.5±3.9 dB/km. Taking into account the results of Humbach et al. , the content of OH groups was estimated to be 10.9±3.5 ppm. Such an OH level should lead to an absorption with a peak amplitude of 0.68±0.23 dB/m at 1383 nm.
The absorption connected to Bismuth-doping has been measured in a 3.4 m-long piece of PCF and it is presented in Fig. 3. Two main broad and intense (several dB/m) absorption bands appear with a maxima at around 810 and 1386 nm together with a week band at 910 nm clearly seen in the inset of the figure. This absorption spectrum slightly differs from the one observed in Bismuth-doped aluminosilicate fibers , though the NIR absorption band at 1400 nm was previously reported in . At the same time it is quite similar to the absorption band which has been reported in phospho-germanosilicate fibers  and in germanosilicate fibers . In Fig. 4 we show a detailed view of the absorption bands and their Gaussian decomposition (see numerical data in the figure). It clearly appears that both bands are structured. In order to give a better representation of the OH contribution in the NIR absorption band at 1400 nm, the corresponding absorption, based on our previous estimation of OH-content, is depicted in Fig. 4b as a black bold line. This absorption was taken into account in our Gaussian decomposition. Obviously the absorption due to the OH groups represents only a limited contribution to this band resulting in a sharp peculiarity at 1386 nm. The NIR absorption connected to Bismuth-doping in the range of 1400 nm seems to be constituted of three different sub-bands as it is suggested by our multi-peak fit. This experimental result denies the assigning of this NIR absorption band to Bismuth centers associated to Germanium  and puts in evidence that this band should be attributed to Bismuth centers in a pure silica subnetwork. Further evidences of our assignment were obtained in PL experiments.
NIR PL has been studied at room temperature in a 2.2 cm-long piece of PCF in transmission geometry and in preform rods (obtained after high temperature drawing of the sol-gel preform at about 2000°C) in transverse geometry. As an excitation sources tunable CW Ti:Sapphire laser (Coherent 899), frequency doubled CW Nd:YAG (Cobolt) and frequency doubled subnanosecond micro-chip Nd:YAG (Teem Photonics) were used. The PL measurements in the bulk samples were performed with the equipment described in the reference  and with an OSA (ANDO AO-6315A) in the case of PCF. The normalized PL spectra of PCF and preform rod at different pump WL’s in the region of 765–860 nm are reported in Figs. 5a and 5b, respectively.
One can observe that the shape and position of broadband NIR PL with a Full Width at Half Maximum (FWHM) of about 100 nm strongly depend on the excitation WL. For instance, in PCF the peak of NIR PL shifts from 1405 to 1432 nm when excitation WL is tuned from 775 to 860 nm. Similar behavior was observed in a bulk sample. We attribute this peculiarity to the complex structure of the absorption band peaked at 820 nm. Hence, it is suggested that the “short” WL’s mainly pump the energy level associated with the absorption band component at 785 nm giving rise to the PL component at 1380 nm while the “long” WL’s pump the levels associated with a long-WL absorption components giving rise to the PL component at 1425 nm. Though there is no apparent absorption bands in the range of 1000 – 1200 nm, pumping the PCF at 1080 nm (Yb fiber laser, KEOPSYS) also produces the NIR PL peaked at 1370 nm together with a strong green up-conversion visible with a naked eye. The latter result implies that the Bismuth centers in a silica subnetwork should strongly affect the efficiency of a fiber lasers based on aluminosilicate fibers and operating in the 1150 – 1200 nm WL region. Obviously, when pumped in the WL region of 1060 – 1090 nm, the absorption and up-conversion of the pump in a silica subnetwork lead to reduction of the laser efficiency. As a whole, the system seems to behave as a continuum of electronic states having a limited set of maxima in the distribution of density of states. So, any excitation WL produces NIR PL. This result once more underlines the complex structure of Bismuth-related PL even in a pure silica glass.
There exists in the literature one more contradiction in the assigning of PL band at 940 nm. While in references [3, 10, 11] this band was attributed to Bismuth centers “bonded” to Germanium, it was recently attributed to Bismuth centers in a silica subnetwork . Although in our recent work it was shown that PL band at 940 nm clearly appears at λexc = 532 nm in Bismuthdoped pure silica sol-gel preform sintered at 1300°C , we report in Fig. 6 a Time Resolved Spectrum (TRS) obtained in a preform rod (after heat treatment at 2060°C) under following conditions: 90° geometry, T = 10 K, time resolution 4 ns. In TRS, the above PL band appears as a most intense and short-lived together with 670 nm and 860 nm PL bands. Although at room temperature the redistribution of the initial intensities was observed, the position of bands and their decay times were similar to those at T = 10K . It is clear that the PL band at 940 nm has also been wrongly attributed to Bismuth “bonded” to Germanium in [3, 10, 11]. While in the work  the latter band was correctly assigned to the Bismuth centers in a silica subnetwork, the attribution of the absorption and PL band at 670 nm to centers in aluminosilicate subnetwork seems to be doubtful in view of the present result and results reported in . Nevertheless, at the present level of our knowledge we cannot exclude a possible superposition of absorption (PL) bands belonging to different subnetworks.
Though the investigation of the role of Germanium codoping in the optical properties of Bismuth-doped silica glass is out of the scope of present work and will be reported elsewhere, we would like to emphasize that the presence of Germanium in a silica network can modify the crystal (glass) field components interacting with Bismuth centers. For instance, the presence of Germanium should certainly lead to the lowering of environment symmetry of Bismuth centers that, in turn, should strongly influence the oscillator strengths of electronic transitions. But it does not means that there exists a direct coupling (or “bonding”) between Bismuth center and Germanium ions.
The optical properties of Bismuth-doped pure silica sol-gel core photonic crystal fiber were investigated. It is shown that, whereas the glass composition studied here is very simple, all absorption and photoluminescence bands exhibit a complex structure. The present study puts in evidence that 1400 nm absorption band already observed in germano-silicate fibers is not associated with Germanium codoping but it is intrinsic to a silica network. The laser experiments using reported PCF and various pump WL’s are in progress and will be communicated in our upcoming paper.
We are grateful to K. Delplace for the assistance in the process of fiber drawing. The work was supported by the “Conceil Régional du Nord/Pas de Calais” and by the “Fonds Européen de Développement Economique des Régions” (FEDER) through the “Contrat de Projets Etat Region (CPER) 2007–2013”.
References and links
1. Y. Fujimoto and M. Nakatsuka, “Infrared luminescence from Bismuth-doped silica glass,” Jpn. J. Appl. Phys. 40, L279–L281 (2001). [CrossRef]
2. V. V. Dvoyrin, V. M. Mashinsky, and E. M. Dianov, “Efficient Bismuth-Doped Fiber Lasers,” IEEE J. Quant. Electron. 44, 834–840 (2008). [CrossRef]
3. S. Firstov, I. Bufetov, V. Khopin, A. Shubin, A. Smirnov, L. Iskhakova, N. Vechkanov, A. Guryanov, and E. Dianov, “2 W bismuth doped fiber lasers in the wavelength range 1300–1500 nm and variation of Bi-doped fiber parameters with core composition,” Laser Phys. Lett. 6, 665–670 (2009). [CrossRef]
4. M. Neff, V. Romano, and W. Lüthy, “Metal-doped fibres for broadband emission: Fabrication with granulated oxides,” Opt. Mat. 31, 247–252 (2008). [CrossRef]
5. I. Razdobreev, H. El Hamzaoui, V. Yu. Ivanov, E. F. Kustov, B. Capoen, and M. Bouazaoui, “Optical spectroscopy of Bismuth-doped pure silica fiber preform,” Opt. Lett. 35, 1341–1343 (2010). [CrossRef] [PubMed]
6. P. S. J. Russell, “Photonic-crystal fibers,” J. Lightwave Technol. 24, 4729–4749 (2006). [CrossRef]
7. O. Humbach, H. Fabian, U. Grzesik, U. Haken, and W. Heitmann, “Analysis of OH absorption bands in synthetic silica,” J. Non-Cryst. Solids 203, 19–26 (1996). [CrossRef]
8. I. Razdobreev, L. Bigot, V. Pureur, A. Favre, G. Bouwmans, and M. Douay, “Efficient all-fiber bismuth-doped laser,” Appl. Phys. Lett. 90, 031103 (2007). [CrossRef]
9. V. V. Dvoyrin, V. M. Mashinsky, L. I. Bulatov, I. A. Bufetov, A. V. Shubin, M. A. Melkumov, E. F. Kustov, E. M. Dianov, A. A. Umnikov, V. F. Khopin, M. V. Yashkov, and A. N. Guryanov, “Bismuth-doped-glass optical fibers - a new active medium for lasers and amplifiers,” Opt. Lett. 31, 2966–2968 (2006). [CrossRef] [PubMed]
10. I. A. Bufetov, S. V. Firstov, V. F. Khopin, O. I. Medvedkov, A. N. Guryanov, and E. M. Dianov, “Bi-doped fiber lasers and amplifiers for a spectral region of 1300–1470 nm,” Opt. Lett. 33, 2227–2229 (2008). [CrossRef] [PubMed]
11. I. Bufetov and E. Dianov, “Bi-doped fiber lasers,” Laser Phys. Lett. 6, 487–504 (2009). [CrossRef]
12. L. Bulatov, V. Mashinsky, V. Dvoyrin, E. Kustov, and E. Dianov, “Luminescent properties of bismuth centres in aluminosilicate optical fibres,” Quantum Electron. 40, 153–159 (2010). [CrossRef]
13. I. Razdobreev, CERLA-PHLAM, University Lille-1, Villeneuve d’Ascq 59655, France, and V. Yu. Ivanov, Institute of Physics of Poland Academy of Sciences, 32/46 al. Lotnikw, Warsaw 02668, Poland are preparing a manuscript to be called “Time resolved spectroscopy of Bismuth-doped silica glass.”