Precise tailoring of acoustic velocity in optical fibers ... - OSA Publishing

Precise tailoring of acoustic velocity in optical fibers by hydrogenation and UV exposure Fanting Kong* and Liang Dong Center for Optical Materials Science and Technology/Department of Electric and Computer Engineering, Clemson University, 91 Technology Drive, Anderson, SC, 29625, USA * [email protected]

Abstract: Tailoring of acoustic properties in solids has many potential applications in both acoustics, i.e. acoustic gratings and waveguides, and photon-phonon interactions, i.e. stimulated Brillouin scattering (SBS). One immediate application is in the area of SBS suppression in optical fibers. We demonstrate, for the first time, a post-processing technique where hydrogen is diffused in to a fiber core and then locally and permanently bonded to core glass by a subsequent UV exposure. It is discovered that local acoustic velocity can be altered by as much as ~2% this way, with strong potential for much further improvements with an increased hydrogen pressure. It is also found that the large change in acoustic velocity is primarily due to a reduction in bulk modulus, possibly as a result of network bonds being broken up by the addition of OH bonds. It is possible to use this technique to precisely tailor acoustic velocity along a fiber for more optimized SBS suppression in a fiber amplifier. Change in Brillouin Stokes frequency of ~320MHz at 1.064μm was observed. ©2012 Optical Society of America OCIS codes: (060.4370) Nonlinear optics, fibers; (190.5890) Scattering, stimulated; (060.3510) Lasers, fiber.

References and links 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

A. Kobyakov, S. Kumar, D. Q. Chowdhury, A. B. Ruffin, M. Sauer, S. R. Bickham, and R. Mishra, “Design concept for optical fibers with enhanced SBS threshold,” Opt. Express 13(14), 5338–5346 (2005). M. J. Li, X. Chen, J. Wang, S. Gray, A. Liu, J. A. Demeritt, A. B. Ruffin, A. M. Crowley, D. T. Walton, and L. A. Zenteno, “Al/Ge co-doped large mode area fiber with high SBS threshold,” Opt. Express 15(13), 8290–8299 (2007). S. Gray, A. Liu, D. T. Walton, J. Wang, M. J. Li, X. Chen, A. B. Ruffin, J. A. Demeritt, and L. A. Zenteno, “502 Watt, single transverse mode, narrow linewidth, bidirectionally pumped Yb-doped fiber amplifier,” Opt. Express 15(25), 17044–17050 (2007). P. D. Dragic, “SBS suppressed, single mode Yb-doped fiber amplifiers,” Proc. of OFC (2009) paper JThA10. M. D. Mermelstein, “SBS threshold measurements and acoustic beam propagation modeling in guiding and antiguiding single mode optical fibers,” Opt. Express 17(18), 16225–16237 (2009). B. J. Ward and J. Spring, “Finite element analysis of Brillouin gain in SBS-suppressing optical fibers with nonuniform acoustic velocity profiles,” Opt. Express 17(18), 15685–15699 (2009). M. D. Mermelstein, M. J. Andrejco, J. Fini, A. Yablon, C. Headley III, D. J. DiGiovanni, and A. H. McCurdy, “11.2dB SBS gain suppression in a large mode area Yb-doped optical fiber,” Proc. SPIE 6873, 68730N, 68730N-7 (2008). L. Dong, “Formulation of a complex mode solver for arbitrary circular acoustic waveguides,” J. Lightwave Technol. 28, 3162–3175 (2010). L. Dong, “Limits of stimulated Brillouin scattering suppression in optical fibers with transverse acoustic waveguide designs,” J. Lightwave Technol. 28, 3156–3161 (2010). K. Shiraki, M. Ohashi, and M. Tateda, “Suppression of stimulated Brillouin scattering in a fiber by changing core radius,” Electron. Lett. 31(8), 668–669 (1995). V. I. Kovalev and R. G. Harrison, “Suppression of stimulated Brillouin scattering in high-power singlefrequency fiber amplifiers,” Opt. Lett. 31(2), 161–163 (2006). S. L. Floch and P. Cambon, “Study of Brillouin gain spectrum in standard single-mode optical fiber at low temperature (1.4-370K) and high hydrostatic pressure (1-250Bars),” Opt. Commun. 219(1-6), 395–410 (2003). N. Shibata, K. Okamoto, and Y. Azuma, “Longitudinal acoustic modes and Brillouin gain spectra for GeO2doped core single mode fibers,” J. Opt. Soc. Am. B 6(6), 1167–1174 (1989).

#176563 - $15.00 USD Received 24 Sep 2012; revised 12 Nov 2012; accepted 14 Nov 2012; published 29 Nov 2012

(C) 2012 OSA

3 December 2012 / Vol. 20, No. 25 / OPTICS EXPRESS 27810

14. Y. Koyamada, S. Sato, S. Nakamura, H. Sotobayashi, and W. Chujo, “Simulation and designing Brillouin gain spectrum in single-mode fibers,” J. Lightwave Technol. 22(2), 631–639 (2004). 15. K. Shiraki, M. Ohashi, and M. Tateda, “SBS threshold of a fiber with Brillouin frequency shift distribution,” J. Lightwave Technol. 14(1), 50–57 (1996). 16. A. Evert, A. James, T. Hawkins, P. Foy, R. Stolen, P. Dragic, L. Dong, R. Rice, and J. Ballato, “Longitudinallygraded optical fibers,” Opt. Express 20(16), 17393–17401 (2012). 17. P. Dragic, T. Hawkins, P. Foy, S. Morris, and J. Ballato, “Sapphire-derived all-glass optical fibres,” Nat. Photonics 6(9), 629–633 (2012). 18. J. Stone, “Interactions of hydrogen and deuterium with silica optical fibers, as review,” J. Lightwave Technol. 5(5), 712–733 (1987). 19. R. M. Atkins, V. Mizrahi, and P. J. Lemaire, “Enhanced photo-induced refractive index changes in optical fibers via low temperature hydrogen loading,” Proc. of CLEO (1993) paper CPD20. 20. R. M. Atkins, P. J. Lemaire, T. Erdogan, and V. Mizrahi, “Mechanism of enhanced UV photosensitivity via hydrogen loading in germanosilicate glasses,” Electron. Lett. 29(14), 1234–1235 (1993). 21. P. J. Lemaire, R. M. Atkins, V. Mizrahi, and W. A. Reed, “High pressure H2 loading as a techqnue for achieving ultra-high UV photosneitivity and thermal sensitivity in GeO2 doped optical fibers,” Electron. Lett. 29(13), 1191–1193 (1993). 22. P. J. Lemaire, A. M. Vengsarkar, W. A. Reed, and D. J. DiGiovanni, “Thermal enhancement of UV photosensitivity in H2-loaded optical fibers,” Proc. of OFC (1995) paper WN1. 23. P. J. Lemaire, A. M. Vengsarkar, W. A. Reed, and D. J. DiGiovanni, “Thermal enhancement of UV photosensitivity in GeO2 and P2O5 doped optical fibers,” Appl. Phys. Lett. 66(16), 2034–2037 (1995). 24. B. I. Greene, D. M. Krol, S. G. Kosinski, P. J. Lemaire, and P. N. Saeta, “Thermal and photo-induced reaction of H2 with germanosilicate optical fibers,” J. Non-Cryst. Solids 168(1-2), 195–199 (1994). 25. 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). 26. G. Meltz and W. W. Morey, “Bragg grating formation and germanosilicate fiber photosensitivity,” Proc. SPIE 1516, 185–199 (1991). 27. I. Dajani, C. Vergien, C. Robin, and C. Zeringue, “Experimental and theoretical investigations of photonic crystal fiber amplifier with 260 W output,” Opt. Express 17(26), 24317–24333 (2009). 28. R. W. Tkach, A. R. Chraplyvy, and R. M. Derosier, “Spontaneous Brillouin scattering for single-mode optical fiber characterization,” Electron. Lett. 22, 1012–1013 (1986). 29. N. Shibata, R. G. Waarts, and R. P. Braun, “Brillouin-gain spectra for single-mode fibers having pure-silica, GeO2-doped, and P2O5-doped cores,” Opt. Lett. 12(4), 269–271 (1987). 30. R. Le Parc, C. Levelut, J. Pelous, V. Martinez, and B. Champagnon, “Influence of fictive temperature and composition of silica glass on anomalous elastic behaviour,” J. Phys. Condens. Matter 18(32), 7507–7527 (2006). 31. J. Kushibiki, T. C. Wei, Y. Ohashi, and A. Tada, “Ultrasonic microspectroscopy characterization of silica glass,” J. Appl. Phys. 87(6), 3113–3121 (2000). 32. M. D. Gallagher and U. L. Osterberg, “Ultraviolet absorption measurements in single-mode optical glass fibers,” Appl. Phys. Lett. 60(15), 1791–1793 (1992). 33. R. M. Atkins, “Measurement of the ultraviolet absorption spectrum of optical fibers,” Opt. Lett. 17(7), 469–471 (1992).

1. Introduction In the past decade, there have been tremendous developments in fiber lasers both in technology and commercialization. There are still significant challenges in further power scaling. For single frequency fiber lasers where there are strong interests for scientific and defense applications, stimulated Brillouin scattering (SBS) represents a major limit for further power scaling. There have been strong research efforts in techniques for SBS suppression in recent years. Much of it has been devoted towards engineering transverse acoustic properties for minimizing optical and acoustic mode overlap in order to suppress SBS [1–7]. SBS suppression up to ~11dB has been demonstrated with this technique [7]. Further suppression with this technique will be limited due to the impact of leaky acoustic modes in optical fibers [8,9]. Longitudinally tailoring of acoustic properties can significantly suppress SBS if local acoustic properties can be altered in a significant fashion. Variation of fiber diameter reported in [10] only weakly affects local Brillouin Stokes frequency shift in fibers with large effective mode areas. In this regime, both optical and acoustic waveguides involved are highly multimoded with large V value. The modal indexes have only a weak dependence on waveguide diameter in this regime. A temperature gradient naturally occurs in a fiber amplifier due to


(C) 2012 OSA


end-pumping schemes typically used, which leads to longitudinally varying acoustic properties and can be used for SBS suppressions [11]. Due to the constraint of coating damage at high temperatures, SBS suppression is also limited in this case. Assuming a maximum allowable temperature rise of 100°C, maximum SBS suppression is ~5dB using a Brillouin line-width of 50MHz and a rate of Brillouin frequency shift change with temperature of ~1.5MHz/K at ~1μm [12]. For SBS suppression beyond 10dB, Brillouin Stokes frequency shift needs to be changed by >500HMz. This can be achieved by compositional change in the order of several percent [13–15]. Furthermore, for optimal SBS suppression, local Brillouin frequency shift needs to vary in a fashion to ensure local SBS gain is near constant along a fiber amplifier. In a fiber amplifier where power increases significantly from input to output, this typically requires the local rate of change for the Brillouin frequency shift to increase towards the output of a fiber amplifier. Since power distribution in a fiber amplifier varies according to designs, a longitudinal SBS suppression scheme is preferably implemented in existing fiber once the power distribution along the fiber and details of the fiber are known. It also needs to be implemented in a highly controlled fashion so that local Brillouin frequency shift can be varied according to a design. It is worth noting that Brillouin frequency shift has been varied along a 28.5Km long fiber by as much as 350MHz by compositional change made during fiber fabrication [15]. Recently, significant compositional change over a much short fiber length scale of few meters, which is more appropriate for fiber amplifiers, has been demonstrated for SBS suppression [16]. It is, however, hard to implement highly controlled Brillouin frequency shift distribution along a fiber using compositional variation in the fiber fabrication. It is recently proposed to use glass with low Brillouin gain for SBS suppression [17]. The ability of the proposed materials to achieve low-loss efficient amplifiers and precise waveguide geometries are still to be proven. In this work, we present a new technique for SBS suppression. It is based on varying acoustic properties of an optical fiber along its length in a controlled manner. Molecular hydrogen is first diffused into the fiber by keeping the fiber in a pressurized hydrogen cell for well over a week. The hydrogen is then permanently locked into the fiber by an exposure to UV lights at 266nm. The amount of hydrogen which can be incorporated into the fiber core is dependent on the pressure of the hydrogen cell assuming that the fiber stays in the cell for long enough to allow sufficient diffusion [18] and is exposed to UV light for long enough to allow sufficient time for the reaction. This hydrogen loading technique is well known for photosensitization of optical fibers for fiber gratings [19–24]. To our knowledge, this is the first time that this technique is used for altering local acoustic properties of an optical fiber. The incorporated hydrogen takes the form of OH bonds and does create a weak absorption band at 950nm in the fiber core, besides additional stronger bands at ~1.25μm and ~1.4μm. The longer wavelength bands are far enough in wavelength to have any effect on the operation of ytterbium-doped fiber lasers. For typical ytterbium-doped fiber lasers claddingpumped at ~976nm, the absorption band at ~950nm at the level of OH in this work is too weak to have any significant effect on the laser operation [25]. The scheme works best for fibers with some germanium-doping in the core [19–21]. It can also work with germaniumfree fibers with appropriate thermal treatment [22–24]. We have demonstrated Brillouin Stokes frequency shift of ~320MHz, i.e. SBS suppression of ~8dB. Using heavier deuterium instead of hydrogen, the induced absorption bands will be moved to much longer wavelength, too far to affect the operation of ytterbium-doped fiber lasers. Using much higher pressure in loading cell, much higher SBS suppression is potentially possible. The demonstrated technique also enables the precise 3D structuring of acoustic properties in solids at spatial precision of optical wavelength for the first time. This may enable some interesting applications in acoustic signal and image processing.


(C) 2012 OSA


2. Theory Hydrogenation technique was developed in the early nineties for the enhancement of photosensitivity in optical fibers for writing in-core fiber gratings [19–24]. The discovery of UV photosensitivity in optical fibers in combination with the development of the external writing technique [26] led to the possibility of writing grating with arbitrary period directly in the fiber core. Despite of the early interests being primarily in optical fiber sensors [26], the realization of possible strong filters and reflectors at any wavelengths directly written in fiber core quickly led to very strong interests for applications in the emerging wavelength-divisionmultiplexing systems in telecommunications as well as fiber lasers. Early fiber gratings were written in germanium-doped fiber core [19–21,26]. It was realized fairly early on that the germanium oxygen deficient centers (GODC) formed near germanium sites with a strong absorption band at ~240nm play a significant role in the UV photosensitivity. UV exposure can disassociate the GODC and lead to formations of other color centers. In the process, refractive index change is created via re-organization of the glass structure and local strain/stress effects. The possibility of hydrogen diffusion through optical fibers was known much early on [18]. In fact, hydrogen released from early coatings caused an increase of losses in optical fibers and significant problems in terms of long-term reliability of the newly installed optical fiber systems. Hydrogen, being a very small molecule, can diffuse through an optical fiber with a 125μm silica glass diameter with a diffusion time of around a week at room temperature. The absorption from molecular hydrogen caused the increase of loss in optical fibers in some early optical fiber systems. The molecular hydrogen, however, stays at the interstitial space in silica glass and can diffuse out given the right concentration gradient. The discovery that UV lights, when absorbed by the ODC, can cause molecular hydrogen to bond to nearby oxygen to form OH bond lead to the development of a very important technique capable of significantly increasing optical fiber photosensitivity for in-core fiber gratings [20]. The fibers are first place in hydrogen cell typically pressurized at hydrogen bottle pressure of ~150Bar for over a week to allow hydrogen molecules to diffuse through the optical fiber at room temperature. Much less time is needed if the hydrogen cell is heated. Fiber coating is then stripped to allow interfering UV beams to exposure the fiber from the side. The interfering UV beams create a periodic intensity pattern along the fiber core. Hydrogen molecules are bonded to the glass according to local UV intensity. The remaining hydrogen molecules are allowed to diffuse out of the optical fiber after the writing process. The possibility of altering local composition with addition of OH bonds leads to the feasibility of very strong refractive index change in fiber gratings [19–21]. Very high refractive index change has been demonstrated with a significant increase of pressure of the hydrogen cell [21]. In this case, the composition has been locally altered at few percent levels. It was later shown that similar effect can be achieved in fibers without germanium in the core when appropriate heating is used during the writing process [22,23]. The fact that composition can be changes at few percent levels is very interested for altering local acoustic properties in existing optical fibers. Acoustic velocity va is related to bulk modulus E and density ρ in the following way. va =

E

ρ

(1)

In SBS, acoustic wave vector is matched to the propagation constant difference between the forward and backward optical modes. The period Λ is related to the mode index n of the optical mode and optical wavelength λ in the following relation.

λ = 2n Λ

(2)


(C) 2012 OSA


The Brillouin Stokes frequency shift fB is then the acoustic frequency and is given below. fB =

v a 2nv a 2n = = Λ λ λ

E

ρ

(3)

It can be seen in Eq. (3), Brillouin Stokes frequency shift can be significantly changed by change in density or bulk modulus. Incorporation of OH in silica glass does lead to addition of absorption bands. The fundamental band of OH and its first overtone locate at around 2.7μm and 1.4μm respectively. The first overtone has a side band located at ~1.25μm. All these bands are too far in wavelength to have any effect in the operation of ytterbium fiber lasers. The second overtone at ~950nm is very close to the typical pumping wavelength of 976nm for ytterbium. This band is typically ~1dB/km/ppm [25]. OH levels at several thousand ppm are used in this work, corresponding to losses of up to ~several dB/m at ~950nm and less than 0.1dB/m at 976nm. This should have negligible impact on the operation of ytterbium-doped fiber lasers. Higher order OH overtones are too weak to be considered here. We further anticipate the use of deuterium instead of hydrogen in the future. This increase of molecular weight has the benefit of moving all the absorption bands significantly towards longer wavelengths, too far away to affect the operation of ytterbium fiber lasers.

Fig. 1. SBS in a fiber amplifier with varying local Brillouin frequency shift.

For optimal SBS suppression in a fiber amplifier with significant power increase from input to output, the local Brillouin frequency shift fB needs to be arranged to obtain a reflected power which is uniform across frequency. Since the various frequency components are amplified at different part of the fiber, this is necessary so that the reflected power from any part of the fiber does not far exceed that from any other part. This is illustrated in Fig. 1. If the overall SBS gain is flat within the entire gain bandwidth, the SBS suppression factor can be estimated from total SBS gain bandwidth ΔΩ and Brillouin line width ΓB by ΔΩ/ΓB. To be able to tailor the distribution of Brillouin frequency shift along the fiber amplifier, precise control is necessary for any SBS suppression technique based on longitudinal acoustic designs. In our proposed approach, hydrogen-loaded fiber can be UV-exposed from the side. The local fB can be controlled by the total local UV exposure. In this case, fiber coating does need to be stripped off for the UV exposure. The fiber needs to be re-coated later. Since power distribution in many fiber amplifiers follows a near exponential curve, this requires an exponential distribution of fB for optimal SBS suppression. UV light may be launched into the fiber from the output end in this case. The absorption naturally dictates an exponential decay of the UV light along the fiber and, consequently, an exposure exponentially varying along the fiber. Assuming operating in the linear regime for the exposure, an exponential


(C) 2012 OSA


distribution of fB can be achieved this way. Coating does not need to be removed in this case. The end-launched UV exposure can significantly simplify the overall process. The level of UV absorption is required to be appropriate for the length of the fiber amplifier in this case. It is worth noting that the proposed technique can lead to a higher NA towards the output end of the amplifier. Adiabatic condition needs to be maintained to minimize any possible mode coupling along the fiber. 3. Experiments The fiber used in this experiment is a Thorlabs Panda polarization-maintaining (PM) fiber with a NA of 0.12 and MFD of 6.6μm at 980nm (PM980-XP). The cladding diameter is 125μm and coating diameter is 245μm with dual acrylate coating. Second mode cut-off wavelength is 920 ± 50nm. Nominal polarization mode cross talk is less than −30dB at 100m. The fiber is kept in hydrogen cell for typically well over a week at a pressure of ~150Bar. For the first set of the experiment, the fiber was stripped of coating and exposed to UV light from the side straight after being removed from the hydrogen cell. The laser used is a diodepumped frequency quadrupled YAG laser at 266nm from Newport Corporation (Hippo 2662C), which is capable of providing an average power of 2W at 50kHz repetition rate. Pulse width is 11ns. The beam was moved by a mirror mounted on a translation stage to scan along the fiber. The beam was focused with a cylindrical lens onto the fiber. Minimum of 10cm length of the fiber is uniformly exposed. A transmission measurement was then performed to characterize the OH peak at ~1.4μm. The absorption coefficient of OH at ~1.4μm was given to be 50.4dB/km/ppm by Keck et al, 61.9dB/km/ppm by Elliott et al and 60.6dB/km/ppm by Clasen et al. These data was summarized in [25]. Incorporated OH level was calculated using the coefficient of 50.4dB/km/ppm given by Keck in this work. Using the second order mode cut-off wavelength of 920nm, we could calculate the V value at 1.4μm is ~1.58. The fraction of power of the fundamental mode in the core can then be estimated to be ~44%. Assuming the OH is in the core, OH level was then calculated considering this overlap factor.

Fig. 2. Setup for Brillouin gain measurement.

The fiber was then measured for its Brillouin gain spectrum. The setup is shown in Fig. 2 and is very similar to what was used in [27]. It is effectively a pump and probe measurement with counter-propagating pump and probe beams. Pump and probe beams are from two NPRO single frequency lasers capable of providing ~100mW with a line width less than 5kHz. Two isolators are used to prevent feed backs into the NPRO lasers (JDSU). Both pump and probe beams propagate in the same polarization in the PM fiber. The pump is amplified further by an ytterbium-doped fiber amplifier to increase sensitivity of the system so that Brillouin gain spectrum can be measured for centimeter long fibers. A Faraday rotator combined with a Polarizing beam splitter is used to separate the amplified probe beam from the pump for spectrum measurement. The probe frequency is tuned over 16 GHz by thermally tuning the crystal in the NPRO generating the probe beam. A RF spectrum analyzer is used to measure the frequency separation between the pump and probe beams. The key motivation of


(C) 2012 OSA


using the configuration in Fig. 2 is its high sensitivity in measuring Brillouin gain. This enables that our Brillouin gain spectrum is measured in the low-gain spontaneous Brillouin regime with minimum bandwidth narrowing from any stimulated effects.

Fig. 3. Measured Brillouin gain spectrum for (a) 2.7m original fiber, (b) 0.3m treated fiber with 4150 wt ppm OH and 0.4m untreated fiber, (c) 0.1m treated fiber with 5600 wt ppm OH and 0.36m untreated fiber and (d) 0.1m treated fiber with 12500 wt ppm OH and 0.4m untreated fiber.

The measured Brillouin gain spectrum of the original fiber is shown Fig. 3(a), showing a single peak at 15.88GHz and a FWHM of 50MHz. The measured Brillouin gain spectrum of a 0.7m long fiber with 0.3m treated is shown in Fig. 3(b). The measured OH level is 4150 wt ppm in the treated section in this case. The Brillouin gain spectrum shows two peaks. The peak at 15.88GHz is from the untreated section of the fiber. The Brillouin frequency shift is lowered by 140MHz in the treated fibers. Figure 3(c) shows the case where 0.1m of a 0.46m fiber was subject to longer UV exposure time. The measured OH level is 5600 wt ppm with corresponding Brillouin frequency shift lowered by 180MHz. The UV exposure in Fig. 3(d) was further increased. The Brillouin frequency shift over the treated 0.1m section with measured OH level at 12500 wt ppm is lowered by 276MHz. The measured change in Brillouin frequency shift versus measured OH level in the treated fiber is shown in Fig. 4(a). The slope is around −2% per wt% of OH. If we can ignore the changes in the modal index n and bulk modulus E, the relative density change Δρ/ρ required for the observed −2% change in ΔfB/fB is −4%. If all the hydrogen molecules are incorporated in the glass in the form of OH, we can estimate the density change from the addition of hydrogen atoms in the glass. The measured OH level is 12500 wt ppm for the −2% change in ΔfB/fB. Considering that the oxygen atoms are already in the glass prior to the treatment, 12500 wt ppm OH corresponds to a density change of 0.074% in density. This is far too small to account for the required 4% change in density. Relative change in Brillouin frequency shift ΔfB/fB as a result of GeO2 doping is given as −0.82%, −0.33% and −0.46% per wt% respectively in [15,28,29]. Relative change in Brillouin frequency shift as a result of P2O5 doping is given as −0.32% per wt% in [29].


(C) 2012 OSA


Relative change in Brillouin frequency shift as a result of F doping is given as –2.8% per wt% in [15]. The theoretical prediction from Eq. (3) requires the slope of relative change in Brillouin frequency shift ΔfB/fB due to density change is −0.5% per 1% density change. If we can assume that one GeO2 simply replaces one SiO2 in the GeO2-doped silica core, we can convert the measured slope of −0.82%, −0.33% and −0.46% per wt% of GeO2 given in [15,28,29] respectively to −1.9%, −0.78% and −1.08% per 1% of density change. If we can deal with P2O5-doped silica similarly, we can convert the measured slope of −0.32% per wt% of P2O5 in [29] to −0.55 per 1% density change. There is a large spread in these data. These slope data, however, are not far from the predicted −0.5% per 1% density change given in Eq. (3), especially considering our over-simplified model for these conversions.

Fig. 4. Measured change in relative Brillouin frequency shift due to treatment versus measured OH level in the treated fiber, red solid line is from Eq. (3) assuming ΔE/E changing at −4% per wt% of OH, and (b) Ratio of Brillouin gain per unit length in the treated fiber over that of the original fiber versus measured change in Brillouin frequency shift.

As we have pointed earlier, density change is far too small to account for the measured change in Brillouin frequency shift in our case. Changes in modal index n and bulk modulus E must therefore be considered in this case. From the works on photosensitivity in [19–23], we know the refractive index will increase as a result of the hydrogen treatment. The relative increase in modal index is expected to be much less than 1%. This increase of refractive index will lead to an increase of fB, and partial cancellation of the observed reductions in fB. Bulk modulus was, however, observed to change by −12% per wt% OH in silica glass [30]. Ignoring any changes in mode index and density ρ, this along can lead to ΔfB/fB change by −6% per wt% OH. Using ultrasonic micro-spectroscopy, longitudinal acoustic velocity was measured to change by ~-3% per wt% OH in [31], corresponding to a bulk modulus change of ~-6% per wt% OH. Using a bulk modulus change of ~-4% per wt% of OH, a good fit to the measured data (see red solid line in Fig. 4(a)) is achieved. This bulk modulus change is slightly higher if we consider any positive change in modal index n. Density change can be neglected in our case. This level of change in bulk modulus is consistent with what was reported in [30,31] and sufficient to explain the large change in fB observed in this work. We, therefore, believe that the significant change in bulk modulus at higher OH levels is largely responsible for the observed significant change in Brillouin frequency shift. The reduction of bulk modulus may arise as a result of the breaking up of closed SiO2 tetrahedral network by the forming of the dangling OH bonds. In order to quantify any change in the Brillouin gain in the treated fiber, we plotted the ratio of Brillouin gain per unit length over that of the original fiber as relative Brillouin gain in Fig. 4(b). The ratio stays essentially around 1, indicating there is no measurable change in the level of Brillouin gain in the treated fibers.


(C) 2012 OSA


Once we have established that a significant change in Brillouin frequency shift can be achieved using this technique, we investigated the feasibility of UV exposure from one end of the fiber. A 1.9m fiber was first loaded with hydrogen and then exposed to UV by launching the UV light into the first end of the fiber. The output from the second end was monitored and used to optimize launching condition. The fiber was exposed for 5hrs. The long exposure was to ensure saturation. The Brillouin gain spectrum was measured by launching the probe beam into the first end of the fiber. The fiber was then successively cut back from the second end. Brillouin gain spectrum was measured after each cut-back. The measured Brillouin gain spectra are normalized to the peak gain and plotted in Fig. 5(a). A curve fitting is performed for each of the Brillouin gain spectra by iteratively guessing the local Brillouin frequency shift while minimizing fitting error. Since we expect the slope in the Brillouin frequency shift distribution along the fiber is high towards the UV launching end, i.e. the first end, we expect the Brillouin gain spectrum is dominated by the part of the fiber where the Brillouin frequency shift changes slowly along the fiber, i.e. towards the second end. This can be clearly seen in Fig. 5(a). For the longer fiber length, the measured Brillouin gain spectrum is getting narrower and getting more and more like the Brillouin gain spectrum of the original fiber. Consequently, we expect a larger error in the obtained local Brillouin frequency shift towards the UV launching end by this method, i.e. the first end. This is especially true for longer fiber lengths. We therefore discarded the Brillouin frequency shift data obtained from the curve fitting over the half of the fiber towards the UV launching end. We only keep the entire data for the shortest three fiber lengths, i.e. 50cm, 60cm and 70cm. The Brillouin frequency shift along the fiber obtained by the curve fitting are shown in Fig. 5(b) for each of the fiber length measured. Also plotted is an exponential fit in solid black line. It can be seen that the measured Brillouin frequency shift distribution is fairly close to the expected exponential. The maximum change in Brillouin frequency shift is ~150MHz in this case. UV absorption is determined from the exponential fit to be ~1.3dB/cm at the UV laser wavelength of 266nm.

Fig. 5. (a) Normalized measured Brillouin gain spectra of a fiber UV-exposed by launching UV light into the first end of the fiber. The fiber was 1.9m long. It was then successively cutback from the second end to shorter lengths. Brillouin gain was measured for each length of fiber by launching probe into the first end of the fiber. The fiber lengths are shown. (b) The Brillouin frequency shift along the length of fiber obtained by curve fitting the measured Brillouin gain spectrum for each length of fiber. An exponential fit is also shown in solid black line.

The loss of pure silica core Schott fiber was measured in [32] to be ~0.5dB/cm. In germanium-doped fibers, the loss is much higher due the presence of absorption peaks at ~240nm due to GODC. The loss of a 3mol% germanium-doped optical fiber (Telecom Fiber) and a 10mol% germanium-doped optical fiber (AT&T Tethered Vehicle Fiber) were measured to be ~30dB/cm and ~100dB/cm at ~266nm respectively in [33]. Since our


(C) 2012 OSA


measured UV loss is close to that of the silica than germanium-doped silica glass, we took the image of the output ends with a CCD camera during the UV exposure. This is shown in Fig. 6(b), along with the cross section of the fiber when illuminated by a white light source in Fig. 6(a). It is interesting to see that the UV light is mostly guided around the two stress elements in the silica cladding glass. Stress elements are typically made of boron-doped silica with much higher thermal expansion coefficient than that of the surround silica glass. Once the stress elements and the surround silica cladding solidify during the draw, the boundary between the two materials is fixed. The silica glass will try to restrain any further contraction of stress elements during the subsequent cooling process. This can put the surround silica glass under strong compression once the fiber is cooled down to the room temperature. This compression can raise the refractive index of the silica cladding around the stress elements and forms waveguides.

Fig. 6. (a) Cross section of the fiber. (b) Image of UV light at the output of the 1.9m fiber.

4. Discussions and conclusions We have demonstrated a post-processing technique with hydrogenation to modify acoustic properties of an existing optical fiber. Change in Brillouin frequency shift as high as 320MHz has been demonstrated with hydrogen. This technique can be used for SBS suppression of ~8dB. We expect much higher SBS suppression is possible with an increased hydrogen pressure. Technique for significantly increasing the pressure of hydrogen has already been demonstrated in [21]. The pressure cell was kept at cryogenic temperature during filling. The cell was then sealed and allowed to return to room temperature. Significant increase of pressure was demonstrated. Our post-processing technique is also capable tailoring Brillouin frequency shift in a controlled manner along an existing optical fiber. This is critical for optimal SBS suppression. We have further demonstrated that an exponential change of Brillouin frequency shift along a fiber can be achieved by having UV launching into one end of the fiber for the exposure. We do expect local fiber NA to be raised proportional to local Brillouin frequency change by this technique. If adiabatic condition is maintained along the amplifier, this is not expected to cause beam quality degradation at the output. Acknowledgments This material is based upon work supported in part by the U. S. Army Research Laboratory and the U. S. Army Research Office under contract/grant number W911NF-10-1-0423 through a Joint Technology Office MRI program.


(C) 2012 OSA
