Stimulated Brillouin scattering and its dependences on strain and temperature in a high-delta opticalfiber with F-doped depresd inner cladding Weiwen Zou,Zuyuan He,Masato Kishi,and Kazuo Hotate
Department of Electronic Engineering,University of Tokyo,Tokyo113-8656,Japan
Received September7,2006;revid November27,2006;accepted December8,2006;
posted December11,2006(Doc.ID74831);published February15,2007 Stimulated Brillouin scattering(SBS)in a high-deltafiber with F-doped depresd inner cladding is studied through considering the interaction of acoustic and optical modes in thefiber.It is found that the number of acoustic modes in thefiber is reduced and the frequency spacing between neighboring modes is enlarged becau of the F doping.The dependences of SBS on strain and temperature are measured and compared for each acoustic mode to investigate the feasibility of discriminative nsing of strain and temperature by u of thefiber.©2007Optical Society of America
OCIS codes:290.5900,120.5820,060.2310,060.2270,060.2370.
Brillouin-bad distributedfiber optic nsors1–3at-tract a great deal of interest for their potential appli-c
ations in smart materials and smart structures. The nsors,however,suffer difficulty in distin-guishing respons to strain from respons to tem-perature change with a singlefiber.To overcome the difficulty,rearchers recently explored the utiliza-tion of two different Brillouin resonance peaks who frequencies show different dependences on strain and temperature infibers.So far,a large-effective-area nonzero-dispersion-shiftedfiber4and a photonic crystalfiber with a small core of high germanium doping5have been investigated.Relatively,the latter approach gives a higher strain–temperature accuracy becau the utilized cond peak is said to originate from the acoustic antiwaveguide in thefiber.5This method,however,is limited in application becau of the prence of multiple subpeaks with clo fre-quency spacing in the Brillouin gain spectrum(BGS), which are hard to parate in measurement.
In this Letter we investigate the stimulated Bril-louin scattering(SBS)in a high-delta opticalfiber with F-doped depresd inner cladding(F-HDF).Our simulation and measurement show that the BGS of the F-HDF has fewer acoustic modes and larger reso-nance frequency spacing between neighboring modes compared with a normal high-deltafiber(HDF)with the same core and cladding but without the inner cladding.
The F-HDF,supplied by Fujikura Ltd.,has a highly GeO2-doped core(radiusϳ3.65m),a ϳ1wt.%F-dope
d depresd inner cladding(radius
ϳ17m),and a pure-silica cladding(radius 62.5m).Compared with thefiber samples demon-strated in Ref.6,our F-HDF sample has a greater GeO2concentration with a maximum ofϳ24mol.%, and its F-doped region is part of the claddings.Figure 1(a)depicts the modeled refractive index profile(solid curve)and the acoustic velocity profile(dashed curve) that is deduced according to Ref.7.The profiles show that the depresd inner cladding for optical modes is an enhanced inner cladding for acoustic modes and then forms a cond acoustic waveguide with respect to pure-silica cladding.By using our newly propod two-dimensionalfinite-element modal analysis8on the profiles in Fig.1(a),we simulate the BGS in F-HDF(solid curve)and that in HDF(dotted curve) as shown in Fig.1(b).Here we assume that the F-HDF and HDF guide only the fundamental optical
LP01mode,although the average normalized fre-
quency v value9is estimated asϳ2.95.This assump-tion is reasonable for the F-HDF becau its de-presd F-doped region enhances the normalized
cutoff v value of the optical LP11mode,which can be understood qualitatively from Ref.9.Figure1(b) shows that,by the F-doping,the number of total acoustic modes is reduced(six modes in HDF to four
modes in F-HDF)and the frequency spacing between neighboring modes is enlarged(for example,ϳ320MHz in HDF between the cond-order mode and the third-order mode is enlargedϳ400MHz in F-HDF).According to the calculation of the four acoustic modes’effective pha velocities(5058,5271,
5470,and5645m/s,respectively)as marked in Fig.
1(a),all four acoustic modes(L01,L02,L03,and L04)
existing in F-HDF are located in the GeO2core region in terms of their effective pha velocities,but the higher-order mode is clor to the F-doped inner-cladding region.The difference in the BGS of F-HDF from that of HDF originates from the decrea of the average acoustic velocity in the claddings induced by the F-doped inner-cladding region,which results
播音主持培训
in
Fig.1.(Color online)(a)Modeled refractive index profile (solid curve)and the deduced acoustic velocity profile (dashed curve)in F-HDF,where⌬corresponds to the rela-tive difference of the refractive index and V l is the acoustic velocity.The marked points show the effective pha veloci-ties of different acoustic modes.(b)Simulated BGS in F-HDF(solid curve)and in HDF(dotted curve).
600OPTICS LETTERS/Vol.32,No.6/March15,2007
0146-9592/07/060600-3/$15.00©2007Optical Society of America
the cutoff of the fifth-͑L 05͒and sixth-order ͑L 06͒acoustic modes and then the increa of the fre-quency spacing.
The experimental tup for SBS measurement is depicted in Fig.2.The pump and the probe waves are equally divided from a 1.549-m distributed-feedback lar diode (DFB-LD).The pump wave is amplified with an erbium-doped fiber amplifier (EDFA)to ϳ20dBm and chopped at 8.3MHz for lock-in detection.The probe wave frequency has a downshift B from the pump wave through a single-sideband modulator (SSBM).To compensate for the loss in the SSBM,two additional EDFAs are in
rted before and after the SSBM,respectively.The probe power launched into the fiber under test (FUT),the F-HDF,is ϳ2.3dBm.A variable optical attenuator (VOA)is ud to prevent saturation of the photode-tector (PD).As shown in int A,a water bath of ±0.1°C accuracy is ud to control the temperature of the FUT.The FUT is coated with only a 250-m acry-late jacket,so that the coating’s influence on the tem-perature dependence can be neglected.The 4.74m FUT spliced to two 5cm standard single-mode fiber pigtails of an isolator and a circulator is handwound around a couple of drums with a diameter of 110mm to depress the influence of bending-induced birefringence.10The drums are inrted into the wa-ter bath and mounted on an x stage t for applying strain.
The typical BGS of the F-HDF measured at 25°C in the loo state is depicted in Fig.3,in which the simulated BGS is also drawn for comparison.The measured result is in good agreement with the simu-lation except for a few frequency discrepancies that are possibly due to the influence from the leaky acoustic modes 11not considered in the simulation.By controlling the microwave frequency for SSBM at each temperature or strain tting,we measured the BGS corresponding to the four acoustic modes (i.e.,L 01,L 02,L 03,and L 04modes).Then we fitted each to a Lorenzian function with an offt to find each reso-nance frequency B pk i .Figures 4(a)–4(d)depict the results measured at 25°C in the loo state as an ex-ample.The solid curves reprent the Lorenzian fits,which match ve
ry well with the experimental data.The resonance frequencies B pk i are 9.3930,9.7572,10.1539,and 10.5645GHz,respectively.
The resonance frequencies B pk i and B pk j of two different peaks (peak i and peak j )in BGS simulta-neously affected by the applied strain and tempera-ture change are governed by the following relation:
ͩ⌬B pk i ⌬B pk j
ͪ=ͩA i B i A j B j
ͪͩ⌬⑀⌬T
ͪ
,͑1͒
where A i ͑j ͒and B i ͑j ͒correspond to the coefficients of strain and temperature for peak i ͑j ͒,respectively.Here we introduce a coefficient difference ratio ␥Aij or ␥Bij to let A j =A i ͑1+␥Aij ͒and B j =B i ͑1+␥Bij ͒.From Eq.(1),we know that whether the strain and tempera-ture can be distinguished is determined only by the following condition:
␥Aij ␥Bij .
͑2͒
In fact,the Brillouin-bad nsor system posss a frequency measurement uncertainty ͑␦͒,which in-duces discrimination errors in strain ͑␦⑀ij ͒and tem-perature ͑␦T ij ͒.The strain and temperature errors can be estimated
bypie
store怎么读
Fig.2.(Color online)Experimental tup of SBS measure-ment.Int A,schematic control of the temperature and the strain on the FUT.EOM,electro-optic modulator;PC,polarization controller;VOA,variable optical attenuator;DAQ,data
acquisition.
Fig.3.(Color online)Typical BGS measured at 25°C in loo state (dotted curve)compared with the simulated BGS (solid curve)in the F-HDF.Bottom axis,measured BGS;top axis,simulated
BGS.
Fig.4.(Color online)Measured BGS (dots)at 25°C in loo state and Lorenzian fittings (solid curves)for (a)first-order ͑L 01͒,(b)cond-order ͑L 02͒,(c)third-order ͑L 03͒,and (d)fourth-order ͑L 04͒acoustic mode scattering.
March 15,2007/Vol.32,No.6/OPTICS LETTERS 601
␦⑀ij =
͉ͯ1+␥Bij ͉+1A i ͑␥Bij −␥Aij ͒ͯ␦,
␦T ij =
ͯ
͉1+␥Aij ͉+1
B i ͑␥Bij −␥Aij ͒
ͯ
␦,͑3͒
respectively,which show that the measurement error becomes smaller if the value of ͉␥Bij −␥Aij ͉is greater.Figure 5depicts the measured resonance frequency of each acoustic mode in the F-HDF as a function of strain and temperature.Note here that the tempera-ture dependence is measured in the loo state and that the strain dependence is measured at 25°C.By using least-squares linear fitting,we get the strain coefficient A i and temperature coefficient B i for each acoustic mode scattering.As summarized in Table 1,the difference ratios of ␥Aij and ␥Bij of the higher-order scatterings with respect to the first-order scat-tering satisfy condition (2)better,providing feasibil-ity to employ the fiber for discriminating the respon to strain from the respon to temperature by using the first-order scattering as the reference and higher-order (e.g.,the fourth-order)scattering as the cond peak,respectively.
The frequency uncertainty ␦in our measurement is 0.1MHz according to a repeatability test.Bad on Eq.(3),the strain errors and the temperature errors in using the F-HDF for discriminative measurements are summarized in Table 1.For the cond-order acoustic mode,the errors are 55⑀and 2.4°C,re-spectively.The fourth-order acoustic mode gives the smallest discrimination errors becau of its greatest ͉␥Bij −␥Aij ͉.This is probably due to the different re-
sidual stress in the core region and the inner clad-ding induced during the fabrication of the fiber.12Therefore the higher-order acoustic mode located clor to the F-doped inner cladding shows a greater difference with respect to strain and temperature.However,for the measured F-HDF,the performance of the fourth-order acoustic mode scattering in dis-criminative measurement of strain and temperature is not satisfactory,becau its effective acoustic pha velocity is still located in the core region [e Fig.1(a)],although it is clor to the inner-cladding region compared with the cond-and third-order.Also,the Brillouin gain of the fourth-order is rela-tively lower (e Fig.3).This result suggests that fur-ther improvement is possible with properly designed core and inner-cladding regions to move the effective velocity of a higher-order acoustic mode (e.g.,cond-or third-order)into the inner cladding region and to enhance its Brillouin gain relative to the first-order acoustic mode.
In conclusion,we have investigated the SBS in an F-HDF and its dependences on strain and tempera-ture.Compared with the BGS in normal HDF,the acoustic modes in F-HDF are found to be modified by the F-doped inner cladding,resulting in fewer acous-tic modes appearing in the core region and a wider frequency spacing between neighboring modes.We also discusd the feasibility of using the F-HDF for discriminative measurement of strain and tempera-ture by utilizing behaviors
of the higher-order scat-terings that differ from the first-order one.The cur-rent performance is expected to be improved by properly designing the core region and the inner-cladding region.
The authors are grateful to Mr.Akira Wada of Fujikura Ltd.for providing the F-HDF sample.W.Zou’s e-mail address is u.-tokyo.ac.jp.References
1.K.Hotate and M.Tanaka,IEEE Photon.Technol.Lett.14,179(2002).
gelato2.K.Hotate and S.S.L.Ong,IEEE Photon.Technol.Lett.15,272(2003).
3.M.Nikles,L.Thevenaz,and P .Robert,Opt.Lett.21,758(1996).
4.C.C.Lee,P .W.Chiang,and S.Chi,IEEE Photon.Technol.Lett.13,1094(2001).
horn5.L.Zou,X.Bao,S.Afshar,and L.Chen,Opt.Lett.29,1485(2004).
6.N.Shibata,K.Okamoto,and Y.Azuna,J.Opt.Soc.Am.B 6,1167(1989).outstanding是什么意思
7.Y.Koyamada,S.Sato,S.Nakamura,H.Sotobayashi,and W.Chujo,J.Lightwave Technol.22,631(2004).8.W.Zou,Z.He,and K.Hotate,IEEE Photon.Technol.Lett.18,2487(2006).
美国留学利弊
9.M.Monerie,IEEE J.Quantum Electron.18,532(1982).
hala
10.R.Ulrich,S.C.Rashleigh,and W.Eickhoff,Opt.Lett.
5,273(1980).
11.S.Afshar,V .P .Kalosha,X.Bao,and L.Chen,Opt.
Lett.30,2685(2005).
12.Y.Park,K.Oh,U. C.Paek, D.Y.Kim,and C.R.a few
Kurkjian,J.Lightwave Technol.17,1823(1999).
Table 1.Strain and Temperature Coefficients and
Discriminative Measurement Errors for Various Acoustic Modes in F-HDF Parameters 1st Order 2nd Order 3rd Order 4th Order A i ͑MHz/⑀͒0.031640.031940.031900.0330B i (MHz/°C)
0.69570.78760.80610.8342␥A 1j —0.00960.00820.0430␥B 1j —0.13210.15880.1992͉␥B 1j -␥A 1j ͉—0.12250.15050.1562␦⑀1j ͑⑀͒—554544␦T 1j (°C)
—
2.4
1.9
1.8
Fig.5.(Color online)Resonance frequencies of different acoustic modes as a function of (a)strain and (b)tempera-ture.Solid lines,least-squares linear fits to data.Their slope rates,strain coefficients A i ,an
d temperature coeffi-cients B i are summarized in Table 1.
长颈鹿的英文602OPTICS LETTERS /Vol.32,No.6/March 15,2007
