The photosensitivity of Ge-doped silica fiber using 213nm, 150ps Nd:YAG radiation, is presented here for first time. Refractive index changes greater than 10-3 were measured in Bragg grating reflectors recorded in a low-Ge content fiber, using average intensities of ≈0.35GW/cm2. Grating growth curves for 213nm inscription wavelength are presented and discussed, in comparison with data obtained using 248nm excimer laser radiation. The experimental results presented denote that contrary to the recording using longer laser wavelengths and pulse durations, the grating inscription employing 213nm picosecond radiation is dominated by a two-photon absorption, which role becomes prominent in long-exposures.
© 2005 Optical Society of America
The exposure of Ge-doped silicate glasses to intense UV radiation, results in significant electronic and structural alterations in the material matrix which are translated to mutually correlated refractive index and absorption changes, compaction and stress induction. Depending on the Ge concentration and codopants, the laser wavelength, the exposure characteristics and other conditioning processes (i.e. hydrogenation), UV induced index changes may vary from 10-5 up to 10-2 . The UV photosensitivity of Ge-doped silicate glasses has been extensively exploited in the development of functional Bragg and long period grating devices in optical fibers and waveguides, that being the backbone components in numerous optical communications and sensing applications.
The photosensitivity of a low-Ge content silicate optical fiber, under 213nm 150ps Nd:YAG irradiation is presented here for first time. Optical fibers with GeO2-concentration lower than 10% have been exposed with 248nm , 193nm , 157nm  excimer lasers, but also recently to 800nm, 200fs Ti:sapphire laser radiation . Fiber-grating inscription using 213nm wavelength and picosecond pulse duration may include specific advantages, compared with the existing art. The use of short wavelength 213nm radiation, which provides an energy per photon of ≈5.8eV, may trigger two-photon processes in the fiber core and cladding, that are not directly dependent on highly absorbing dopants and defects. Picosecond laser pulses can be easily interfered through phase mask elements without the penalty of extreme spatial dispersion ; while fringe formation is straightforward using open interferometers without strict alignment requirements due to limited coherence length as that holds for femtosecond pulses. In addition, picosecond pulses lead to significantly higher intensities compared with those of nanosecond duration, accelerating the growth of non-linear effects.
The output of the fifth harmonic of a single longitudinal and transversal mode Nd:YAG laser (EKSPLA, Lithuania), was used for the performed experiments. The laser pulse was “in-cavity” compressed from 3ns to 150ps, using SBS (stimulated Brillouin scattering) mirror technique, and subjected to a two-stage amplification. The final output at 213nm was 7mJ in average, at 10Hz repetition rate, with a beam size of 5.5mm diameter, resulting in intensity of the order of 0.2GW/cm2. The coherence length of the above laser unit is longer than 2cm, allowing the straightforward formation of grating fringes using open interferometric setup.
An adjustable rectangular aperture (RA), having dimensions of 3.0mm×3.5mm, was used for selecting the optimum part of the laser beam (see Fig. 1). The rectangular beam was injected into an elliptical Talbot interferometer, consisting of a fused silica phase mask (PM) for beam splitting and two 45° degrees dielectric mirrors (M1 & M2) for beam refolding on the fiber plane (PF) . Optical path and pulse energy were equalized for both the interfering beams, ensuring high-contrast fringes on the interference plane. The grating period was set to be ≈528.5nm. A fused silica cylindrical lens (CL) of 750mm focal length was placed outside of the interferometer for adjusting the energy density at the position of the fiber. Furthermore, before the cylindrical lens, the laser beam was passed through an oscillating CaF2 plate (OP) of 5mm thickness, for horizontally slanting the laser beam over the grating area and subsequently averaging spatial irregularities. The above optical arrangement resulted in recording of slightly apodised gratings of 3.5±0.2mm approximate length. For comparison, gratings of the same length as the above were recorded using a phase mask setup and 248nm excimer laser radiation, of 34ns pulse duration, at 10Hz repetition rate. These exposures were performed using pulses of ≈360mJ/cm2 energy density.
The fiber exposed was the GF1B Ge-doped photosensitive fiber, produced by Nufern, with numerical aperture (NA) of 0.13 . From the NA value an approximated GeO2 concentration of ≈5.5% was deduced. Also, from the NA value provided and the measurement of the core diameter using white light injection and optical microscopy, we estimated a core power confinement factor Γ of 0.82, approximately . The calculated Γ factor was used for the normalization of the obtained index change data.
The gratings recorded, were online monitored in transmission mode using the amplified spontaneous emission of an Er-doped fiber amplifier (EDFA) and an optical spectrum analyzer (OSA), operating at its maximum resolution of 0.05nm. For obtaining grating reflectivity measurements the back-port of a 50/50 fiber coupler (FC) was connected between the EDFA source and the inscribed fiber grating. Finally, isochronal annealing of the recorded gratings was performed at ambient environment, in a tube furnace of temperature stability better than 2 degrees, for isothermal intervals of 30mins duration. Ramping time between fixed temperatures was 15mins.
3. Results and discussion
Fiber gratings were inscribed using energy densities up to 50mJ/cm2 per pulse. Higher energy densities induced progressive cladding damage after exposure to few thousand of pulses, resulting in gratings with distorted spectra and reduced mechanical durability. Grating strength data and average index change extracted from Bragg wavelength shift, for exposure at 213nm, are presented in Fig. 2(a) and (b).
Average index changes Δnav≈8.5×10-4 were obtained after a two hours exposure of the fibre with a cumulative energy density of 2.9KJ/cm2. Nonetheless, the growth rates of both average and modulated index changes had not reached a plateau, indicating that the recording process was not saturated. The index modulation Δnmod estimated from the grating strength - see Fig. 2(a)- for such exposure conditions, was Δnmod≈4×10-4, assuming that the grating is uniform along its length. The reduced index modulation observed in the inscribed structures may be attributed to spatial irregularities  and slight elliptical polarization  of the recording laser beam. Both of the above experimental factors can significantly degrade UV induced index modulation contrast, resulting in lower diffraction efficiency fibre gratings, especially in the case of prolonged exposures. A maximum index change of Δnmax≈1.25×10-3 results from the sum of the average index change (Δnav) and the index modulation (Δnmod) figures.
The data of the average index change Δnav (see Fig. 2b) were fitted using a power law function, of the form:
where F energy density, N number of pulses, τ the duration of the pulse in nanoseconds and b rate constant to be evaluated . The parenthesis-enclosed term translates to radiation dose per unit area. The power law of Eq.1 has been extensively used for the description of compaction and defect rate formation in silicate glasses under ultra-short wavelength irradiation (γ- and x-rays) and particle/electron beam bombardment , providing b factors lying between 0.5 and 0.7, approximately. A b factor value of ≈0.66 was evaluated through our data analysis; figure which is in close agreement with those reported in the literature for compaction and defect formation in pure and Ge-doped silicate glasses under exposure to 193nm excimer laser radiation [10–12].
A transmission spectrum of a 3.5mm long grating recorded using 74440 pulses of 42mJ/cm2 energy density is presented in Fig. 3. The strength of the above grating is greater than -14.5dB, exhibiting reduced sidelobes at short wavelengths, signature of apodisation due to the beam scanning method adopted during recording. The distinct “sidelobe-like” structure observed for long wavelengths may be attributed to a stitching error inserted by significant beam irregularity, not being properly “washed out” during homogenization process. Furthermore, the homogenization process employed can be responsible itself for the creation of such spectral feature, since the beam scanning rate is slightly faster in the middle than in the edges of the grating length. That artifact of the scanning method may result in a grating strength spatial profile of peculiar shape. Similar irregular spectral peaks were observed on either sides of the Bragg notch for gratings recorded with different exposure conditions.
The temperature stability of the 213nm recorded gratings was studied by employing isochronal annealing at fixed intervals of 30mins, reaching temperatures up to 700°C. Results on the effect of annealing on the grating strength are presented in Fig. 4. For comparison, annealing results of a grating recorded with 248nm excimer laser are also appended. Both gratings exhibit quite similar thermal decay behavior within the limits of experimental error, reaching full erasure at 700C°.
We believe that the underlying photosensitivity mechanism for UV induced index changes in low-Ge silicate fibres using 213nm radiation is not solely associated with defect exhaust that are related to germanium-oxygen-deficiency-centers (GODC). Instead, strong structural changes that take place in the core and less in the cladding region may further promote that photosensitive behaviour. The above assertion is primarily triggered by deep UV spectrophotometric measurements performed for 9% GeO2 fibre performs , which show that the absorption at 213nm is tenfold lower than the GODC absorption peak centered at 242nm wavelength; and threefold lower than the absorption at the 193nm band. Thus, the recording procedure using 213nm radiation, lacks in terms of linearly absorbed UV photons in the fiber core, compared with the recordings employing 248nm and 193nm wavelengths.
Compaction driven index changes associated with bond-cleaving in the Ge-doped glass matrix due to two-photon absorption may not be negligible for the case of 213nm picosecond radiation. The superposition of two photons of the questioned wavelength provides an energy quantum that lies well above the band gaps of pure silica and germanosilicate glasses. In addition, the two-photon absorption coefficient of fused silica, for 213nm radiation is 50×10-11cm/W, slightly higher than that measured for 193nm ; while the value for the Ge-doped core is expected to be much greater . Index changes due to UV induced compaction can steadily grow due to the significant two-photon absorption coefficient, but also due to the short pulse duration which promotes non-linear interactions. Borrelli and co-workers  have presented that glass matrix densification is a major contributor in the UV induced refractive index changes for 193nm excimer laser exposures of Ge-doped glasses. In another experiment, Albert et al  exposed low-defect, Ge-doped fibre to 193nm excimer laser radiation, obtaining index changes of the order of 10-3, with the estimated contribution of compaction being approximately the half of the maximum index change inscribed. Both of the above investigators examined Ge-doped glasses, which are relatively transparent, at the wavelength of exposure; this fact also holds for exposures using 213nm radiation. Thus, similar photosensitivity mechanism may apply for the case of 213nm exposures, where two-photon bond cleaving leads to color center formation and structural changes. Two-photon processes do not rely on highly absorbing pre-existing states, but they can occur for the majority of GeO2 and SiO2 population, allowing defect formation over long-term exposures.
An interesting point of the presented data is the annealing trend of 213nm gratings, which appears to be similar with that of 248nm excimer laser recorded grating. Such thermal erasure behavior has been reported before  for comparative annealing of fibre gratings inscribed by 248nm and 193nm excimer lasers, constituting an indication that the same color centers are created by the two wavelengths. In addition to the above, Fujimaki et al  exposed germanosilicated glass of various compositions to 222nm KrCl excimer laser radiation and subsequent electron spin resonance measurements revealed that Ge(1) and GeE’ are predominantly formed. So, it is possible that such color centers will be created in our case, contributing to the overall index changes attained through Kramers-Kronig transformation and induced compaction. Nonetheless, we cannot confirm such assumption by our experimental data and analysis. Color centers and structural changes solely associated with the SiO2 matrix also play role in the photosensitivity mechanism; however, their strength is expected to be manifold lower than those of GeO2 core .
The photosensitivity of a low-Ge fibre using 213nm, 150ps Nd:YAG radiation was presented. Significant index changes of 1.25×10-3 were inscribed in unsensitised low-Ge content silicate fibre (Nufern GF1B) using total exposures lower than 3kJ/cm2. The experimental results obtained indicate that the photosensitivity mechanism underlying the exposures using such wavelength includes a significant component correlated with matrix compaction. We are working towards the improving of our experimental setup, for achieving higher index modulations during UV recording, thus, writing stronger gratings. Exposure of other optical fibres using 213nm Nd:YAG radiation is in progress and emerging results will be reported soon.
Authors would like to acknowledge support from the Ultraviolet Laser Facility operating at FO.R.T.H. under the Improving Human Potential (IHP) -Access to Research Infrastructures Programme of the EC (contract Nos HPRI-CT-1999-00074 and HPRI-CT-2001-00139), as well as, Dr Alexandros Lappas (Solid State & Materials Chemistry Laboratory, FORTH-IESL) for providing annealing equipment. SP gratefully acknowledges Prof. Michail Zervas (SPI, UK) and Dr Sotirios Kanellopoulos (Polarmetrix, UK) for constructive discussions.
References and links
1. M. Douay, W.X. Xie, T. Taunay, P. Bernage, P. Niay, P. Cordier, B. Poumellec, L. Dong, J.F. Bayron, H. Poignant, and E. Delevaque, “Densification involved in the UV-based photosensitivity of silica glasses and optical fibers,” J. Lightwave Technol. 15, 1329–1342 (1997). [CrossRef]
2. J. Albert, B. Malo, K.O. Hill, F. Bilodeau, D.C. Johnson, and S. Theriault, “Comparison of one-photon and two-photon effects in the photosensitivity of germanium-doped silica optical fibers exposed to intense ArF excimer laser pulses,” Appl. Phys. Lett. 67, 3529–3531 (1995). [CrossRef]
3. K.P. Chen, P.R. Herman, J. Zhang, and R. Tam, “Fabrication of strong long-period gratings in hydrogen-free fibers wth 157-nm F2 laser radiation,” Opt. Lett. 26, 771–773 (2001) [CrossRef]
4. S.J. Michailov, C.W. Smelser, D. Grobnic, R.B. Walker, P. Lu, H. Ding, and J. Unruh, “Bragg gratings written in all-SiO2 and Ge-doped core fibers with 800nm femtosecond radiation and a phase mask,” J. Lightwave Technol. 22, 94–100 (2004) [CrossRef]
5. S. Pissadakis and L. Reekie, “An elliptical Talbot interferometer for fiber Bragg grating fabrication,” Rev. Sci. Instr. (submitted)
6. For more details on GF1B fiber specifications see Nufern website: http://www.nufern.com.
7. T. Erdogan, “Fiber grating spectra,” J. Lightwave Technol. 15, 1277–1294 (2004) [CrossRef]
8. S. Pissadakis, M.N. Zervas, L. Reekie, and J.S. Wilkinson, “UV interferometric ablation and structural modification for the fabrication of sub-micron scale periodic structures in hard optical materials,” in Proceedings of International Conference on Transparent Optical Networks (Institute of Electrical and Electronics Engineers, Wroclaw, 2004), pp. 313–318
9. H.J. Deyerl, N. Plougmann, J.B. Jensen, F. Floreani, H.R. Sorensen, and M. Kristensen, “Fabrication of advanced Bragg gratings with complex apodization profiles by use of the polarization control method,” Appl. Opt. 43, 3513–3522 (2004) [CrossRef] [PubMed]
10. R.E. Schenker and W.G. Oldham, “Ultraviolet-induced densification in fused silica,” J. Appl. Phys. 82, 1065–1071 (1997) [CrossRef]
12. N.F. Borrelli, D.C. Allan, and R.A. Modavis, “Direct measurement of 248- and 193-nm excimer-induced densification in silica germania waveguide blanks,” J. Opt. Soc. Am. B 16, 1672–1679 (1999) [CrossRef]
13. J.-L. Archambault, “Photorefractive gratings in optical fibres,” PhD thesis, University of Southampton (1994)
14. S.A. Slattery and D.N. Nikogosyan, “Two-photon absorption at 211 nm in fused silica, crystalline quartz and some alkali halides,” Opt. Commun. 228, 127–131 (2003) [CrossRef]
15. A. Dragomir, J.G. McInerney, D.N. Nikogosyan, and P.G. Kazansky, “Two-photon absorption properties of commercial fused silica and germanosilicate glass at 264nm,” Appl. Phys. Lett. 80, 1114–1116 (2002) [CrossRef]
16. J. Albert, K.O. Hill, D.C. Johnson, F. Bilodeau, S.J. Mihailov, N.F. Borrelli, and J. Amin, “Bragg gratings in defect-free germanium-doped optical fibers,” Opt. Lett. 24, 1266–1268 (1999) [CrossRef]
17. M. Fujimaki, T. Watanabe, T. Katoh, T. Kasahara, N. Miyazaki, Y. Ohki, and H. Nishikawa “Structures and generation mechanisms of paramagnetic centers and absorption bands responsible for Ge-doped SiO2 optical-fiber gratings,” Phys. Rev. B 57, 3920–3926 (1998) [CrossRef]
18. H. Hosono, H. Kawazoe, and J. Nishii, “Defect formation in SiO2:GeO2 glasses studied by irradiation with excimer laser light,” Phys. Rev. B 53, R11921–R11923 (1996) [CrossRef]