spie proceedings [spie spie photonics europe - brussels, belgium (monday 16 april 2012)] optical...

Fiber optic Brillouin distributed sensing using phase-shift

keying modulation techniques

Birgit Stillera, Min W. Leea, Duc Minh Nguyena, Jerome Haudenb, Alexandre Mottetb, Herve

Maillottea and Thibaut Sylvestrea

aDepartement d’Optique, Institut FEMTO-ST, Universite de Franche-Comte, Besancon,

France;bPHOTLINE Technologies, F-25001 Besancon, France;

ABSTRACT

In this work we demonstrate two new BOTDA sensing systems based on differential (DPSK) and quadrature(QPSK) phase-shift keying modulation techniques with enhanced performances. First we demonstrate Brillouinechoes distributed sensing (BEDS) with centimeter resolution using a single intensity DPSK modulator for thepump pulse. The optical π-phase pulse is directly generated at the end of an intensity pulse using DPSKtechnique. This allows an easy adjustment of the delay between the intensity and phase pulse and improves theoptical loss of the pump. The second technique uses an optical QPSK modulator (I & Q modulator) as a singlesideband (SSB) modulator. The advantage of I & Q modulator compared to dual-drive modulator lies on thehigh performance of carrier suppression of 55 dB as well as side-mode suppression of 40 dB at 1535nm. Besidesthe filter that chooses either the Stokes or anti-Stokes component before detecting the Brillouin response on thephotodiode is no more needed. By use of the I & Q modulator the performance of BOTDA using either Stokesor anti-Stokes component is shown and discussed.

Keywords: Stimulated Brillouin scattering, distributed fiber sensors, Brillouin optical time-domain analysis.

1. INTRODUCTION

Fiber optic distributed sensors have attracted interest worldwide over the last two decades for monitoring solu-tions in, e.g., civil engineering or petrol industry. One of the most popular techniques is Brillouin optical timedomain analysis (BOTDA) that enables distributed strain and temperature sensing with meter spatial resolu-tion.1, 2 Sensor lengths can go up to 100km with distributed Raman amplification and 2m spatial resolution3 andreaches 120km with 3m spatial resolution using optical pulse coding technique.4 However, the BOTDA systemis limited to one meter spatial resolution because of the acoustic decay time (10 ns). To overcome this spatialresolution limit different techniques have been proposed as using dark pulses reaching a 2 cm spatial resolution5

and Brillouin echo distributed sensing (BEDS)6 where short π-phase shift pulses instead of a intensity pulses areused. The remarkable advantage of this new distributed measurement technique is the high spatial resolution(up to 5 cm) with high contrast while conserving a narrow Brillouin gain spectrum. In this paper we presentfurther development of the BOTDA and BEDS technique to reach high lengths and high spatial resolution. Forthis purpose we use differential (DPSK) and quadrature (QPSK) phase-shift keying modulation techniques. Inthe first part of this paper we experimentally demonstrate a simplified architecture of a BEDS system withcentimeter spatial resolution. It is based on the DPSK-technique using a single Mach-Zehnder modulator togenerate a pump pulse and a π-phase-shifted pulse with an easy and accurate adjustment of delay.7 In order tohave a comparison, we show a distributed measurement of a 5cm-splice segment with our simplified setup andwith the conventional BEDS technique. Moreover a strain measurement over 5 cm has been effected to show thehigh spatial resolution. The second part presents a BOTDA system with a single sideband cw probe using aI & Q modulator. It provides a high carrier suppression of 55 dB as well as side-mode suppression of 40dB at1535nm. The distributed measurement of a 50km-long single mode fiber is presented and discussed with regardto the use of the Stokes- and anti-Stokes component in comparison to a dual-band probe.

Contact e-mail: [email protected]

Optical Sensing and Detection II, edited by Francis Berghmans, Anna Grazia Mignani, Piet De Moor, Proc. of SPIE Vol. 8439, 843909 · © 2012 SPIE · CCC code: 0277-786X/12/$18 · doi: 10.1117/12.922581

Proc. of SPIE Vol. 8439 843909-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 08/18/2013 Terms of Use: http://spiedl.org/terms

Vπ Vπ

V

I

Vπ

Opti

cal

inte

nsi

ty

t

t

Vπ Vπ

V

I

t

t

2 Vπ

(a) (b)

Opti

cal

inte

nsi

ty

phase 0 phase 0 phase π

Im

Re

Im

Re

(i) (ii)

(iii) (iv)

(i) (ii)

(iii) (iv)

Figure 1. Use of a Mach-Zehnder interferometer modulator with (a) On Off keying and (b) Differential Phase Shift Keying.(i) Transfer function, (ii) optical output of modulator, (iii) RF input for bias, (iv) constellation diagram, respectively for(a) and (b).

2. DIFFERENTIAL PHASE SHIFT KEYING BASED BEDS

Distributed Brillouin sensing by help of the BEDS technique is also a pump-probe technique as BOTDA. For thisnew technique short optical π-phase pulses are used to enhance the spatial resolution without broadening of theBrillouin gain spectrum. Additionally, the π-phase pulse is created at the end of a long intensity pump pulse inorder to reduce the effect of pump depletion and other non linear effects.6 Therefore two modulators are insertedinto the setup on the side of the pump wave: an intensity modulator to create the long intensity pulse and aphase modulator to generate the short π-phase pulse. The measurements are obtained by taking the differenceof two successive measurements with and without the π-phase pulse which requires a sophisticate adjustment ofthe delay between the pulses. We present here a new concept to perform BEDS with only one modulator forthe pump. The optical π-phase pulse is directly generated using a single intensity modulator based on DPSKtechnique instead of a phase modulator. A long positive electrical pulse is followed by a short negative pulse andthis positive-negative shape pulse is applied to the modulator to generate an optical π-phase shift at the end ofthe long intensity pulse. This means that the long intensity pulse and the short π-phase pulses are generated bythe same intensity modulator, a phase modulator is needless.

2.1 Principle of differential phase shift keying

DPSK technique is widely used in telecom industry to transmit bit-sequence messages in form of π-phase shiftusing an intensity modulator, i.e. Mach-Zehnder interferometer modulator (MZI). In the working principle ofDPSK, optical phase can be shifted by modulating an MZI with an amplitude of double half-wave voltage (2 ·Vπ).The transfer function of an MZI is shown in Fig.(1a(i)). When an electrical pulse with amplitude Vπ , the bias,is applied on the MZI, as in Fig.(1a(iii)), we get from a minimum to a maximum of the transfer function. Thisinduces an intensity pulse to the light wave, as in Fig.(1a(ii)). The same intensity pulse would be obtainedswitching to the other maximum in the transfer function as depicted in Fig.(1b(ii)). Nevertheless, there is adifference in optical phase as it can be seen in the constellation diagram in Fig.(1a,b(iv)). Thus switching from−Vπ to Vπ induces a phase pulse almost without inducing any difference in optical intensity. But since the ”0” iscrossed by switching from −Vπ to Vπ, a short decay in intensity can be observe between the 0-phase and π-phaseintensity pulse.Also in our case, the MZI is driven at its minimum bias point by a positive voltage 0 < V < Vπ and successivelya negative voltage 0 > V > −Vπ, the relative phase-shift between the optical fields at the two voltages is π, orvice-versa.8 Therefore, when a negative pulse of Vπ (or a pulse of −Vπ) is applied to the modulator just after apositive pulse of Vπ at the minimum bias point as illustrated in Fig.(2 a), the output optical intensity remainsunchanged whilst the phase of the optical field at the negative pulse part is shifted by π with respect to thepositive pulse part. In order to generate such an electrical pulse, the positive and negative electrical pulses areseparately generated as shown in Fig.(2 b). Their peak-to-peak amplitude is 7.6V and applied to the MZI usedin the work (Photline MXPE series, Vπ = 6.2V, 20GHz bandwidth). The pulse durations are set to 30 ns and



0 5 10 15 20 25 30 35-1.5

0

0.5

1

1.5

Time (ns)

Outp

ut

(V)

(a)

-3.6

0

3.6

Relative optical phase: πTime

πV

π-V

(b)

(c)

Figure 2. (a) Schematic drawing of a positive/negative electrical pulse. (b) A upward (blue trace) and a downward (redtrace) electrical pulses are generated separately and applied to the MZI. The durations of the long and short intensitypulses are 30 ns and 500 ps, respectively. (c) The intensity pulses at the MZI output. The phase is shifted by π for 500 psat the pulse end.

500ps, respectively. In telecom, such pulses are applied to a dual-drive MZI.9 In our work, these long and shortpulses are applied to the DC and RF inputs of the MZI, respectively. An adaptation of 50Ω is made in the DCport and it provides a sufficient bandwidth for the 30-ns pulse. As the pulses are in form of intensity at theoutput of the MZI, the delay between the pulses is readily adjusted. Fig.(2 c) exhibits two intensity pulses atthe modulator output, which are generated by the electrical pulses seen in Fig.(2 b). A 30-ns intensity pulse isfollowed by a 500-ps intensity pulse with π-shifted phase. A short drop between the two pulses is seen due tothe fast transition from the positive pulse to the negative pulse. The 500-ps duration of the short pulse definesa spatial resolution of 5 cm in our DPSK-based BEDS system.

2.2 Experimental setup

This dual-phase state intensity pulse is used as the pump wave in the experimental setup shown in Fig.(3). Thesetup is nearly the same as BOTDA system but different electrical pulses are applied to the MZI of the pump.The CW emission of a distributed feedback (DFB) laser at 1550 nm is split into the pump and probe arms bya 50:50 tap coupler. On the pump side, an RF bias-T is used to combine the positive pulse and DC bias. TheDC bias of the MZI is adjusted to be at the minimum bias point. The optical pulses in Fig.(2 c) are amplifiedby an Erbium-doped fiber amplifier (EDFA) with a peak power of 1.3 W. They are then polarisation-scrambledby a polarisation scrambler to average out polarisation-sensitive Brillouin gain and injected into a fiber undertest via an optical circulator. In the probe side, the Brillouin frequency-shifted probe is generated by the secondMZI with a carrier suppression of 38 dB at 1550 nm by adjusting the DC bias. The probe power is set to 1mWby another EDFA. The probe is then injected into the fiber after an isolator. At the fiber output, a fiber Bragggrating (FBG) enables to select the Brillouin gain obtained by the pump at the Brillouin frequency shift (BFS).The gain is retrieved by a photodetector and then recorded as time traces by a digital oscilloscope. By sweepingthe frequency of the optical probe, the gain along the fiber is mapped as a function of frequency. In this way, allBFS variations along the fiber caused by temperature or strain can be easily detected.

2.3 Results

In order to demonstrate the performance of the concept developed in the work, a 2-m single-mode fiber (SMF)was spliced with a 1-m fiber with a high numerical aperture (HNA). The splice point is protected by a heat-shrinking fiber protective sleeve of 5 cm. For comparison, distributed measurements of the splice segment havebeen done in two BEDS systems: conventional BEDS system using a phase and an intensity modulator for thepump and DPSK-BEDS system using a single intensity modulator. Fig.(4 a) shows the distributed measurementobtained in the conventional BEDS system via a π-phase shifted pulse with a pulse width of 500 ps. It clearlyreveals the splice segment of 5 cm between the fibers at a frequency shift of 10.55GHz. The BFS of the HNA andSMF are 10.67GHz and 10.85GHz, respectively. Fig.(4 b) displays the mapping using DPSK-BEDS developedin the work. It also manifests clearly the 5-cm splice segment and the 10.55GHz frequency shift. Therefore,it is evident that our system demonstrates performances as good as the standard BEDS system. However, thecontrast of the measurement in BEDS is slightly better than that in DPSK-BEDS. This is due to the peak-to-peak amplitude of 7.2V which is only 1.16 · Vπ in DPSK-BEDS (limited by the pulse generator) where 2 · Vπ is



Amplified

DFB-Laser

1550 nm

Polarization controller

Fiber under test

Polarization

controller

EDFA

DC-supply

DC-supply

RF-Signal

Generator

FBG

Intensity Modulator

Photo Diode

Optical

Circulators

PU

MP

PRO

BE

Scrambler

ννBB

Optical

Isolator

Pump pulse

Fiber

Coupler 50:50

Pulse

GeneratorScopeTrigger

Intensity

Modulator

DC-supply + RF-Input for short

π phase pulses

DC-Input

for long

intensity pulses

π

Figure 3. Experimental setup of DPSK-BEDS system using a single Mach-Zehnder Interferometer modulator (MZI) forthe pump.

needed for the best performance. On the other hand, in BEDS a pulse amplitude of 5.1V is applied to the phasemodulator of which the Vπ is 5.9V. In this case, only Vπ is required for the best operation and the amplitudemeets almost this requirement (0.87 · Vπ).

2.4 Distributed strain measurement

We also tested our DPSK-BEDS system with centimeter resolution for strain monitoring. Therefore we have setup an elongation platform that can be controlled with high precision (10μm) by Labview. Therefore we fixedthe fiber with tape and marked the point where the fiber has been fixed to be sure that the fiber is stretchedwithout slipping under the tape. First we measured the BFS depending on applied strain on 5 cm of an SMF.The measurement was obtained by the DPSK-BEDS technique with a spatial resolution of 5 cm. The effect ofan elongation of 0 - 80μm on the BFS can be seen in Fig.(5) where (a) shows the analysis of the BFS for threedifferent elongations, (b) fitted data in order to estimate the strain coefficient and (c-e) the mapping of the 5 cmsection for 20μm, 50μm and 80μm over 5 cm. Since 5 cm corresponds to the spatial resolution of the system the

Fiber length (m)

Bri

llo

uin

fre

qu

ency

sh

ift

(GH

z)

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

10.9

10.85

10.8

10.75

10.7

10.65

10.6

10.55

10.5

Fiber length (m)

Bri

llo

uin

fre

qu

ency

sh

ift

(GH

z)

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

10.9

10.85

10.8

10.75

10.7

10.65

10.6

10.55

10.5

5 cm 5 cm

b)a)

Figure 4. Distributed measurements of HNA/SMF fibers showing a splice segment of 5 cm (a) in standard BEDS systemusing a phase modulator and (b) in our BEDS system based on DPSK technique.



0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.510.845

10.85

10.855

10.86

10.865

10.87

10.875

10.88

10.885

Fiber length (m)

Bri

llouin

fre

quen

cy (

GH

z)

20μm/5cm50μm/5cm80μm/5cm

0 200 400 600 800 1000 1200 1400 16000

5

10

15

20

25

30

35

40

45

50

Strain (μepsilon)

Fre

quen

cy s

hif

t (M

Hz)

MeasurementFit

2,97 MHz / 100 με

(a)

2.8 2.9 3 3.1 3.2 3.3

10.9

10.88

10.86

10.84

10.82

10.82.8 2.9 3 3.1 3.2 3.3 2.8 2.9 3 3.1 3.2 3.3

Bri

llouin

fre

quen

cy (

GH

z)

Fiber length (m) Fiber length (m) Fiber length (m)

(b)

(c) (d) (e)

5 cm

Figure 5. (a) Brillouin frequency shift versus fiber length and (c-e) Brillouin gain mapping for three different elongationsapplied to a 5-cm fiber section. (b) BFS dependent on applied strain

graphs do not decline sharply at the end of the 5cm-section in Fig.(5 a,c-e). Nevertheless, it confirms well theperformance of the DPSK-BEDS system. As the elongation length increases, the BFS increases from 10.847 GHzto 10.865 GHz and 10.883 GHz. However, the figure reveals the elongation over more than 5 cm. We speculatethat the pure fiber may slightly slip inside the fiber coating by the strain. The BFS is displaced by 18 MHzat 600 με and 36 MHz at 1200 με. It is seen from the elongation measurements that the strain coefficient isobtained as 2.97MHz per 100με, with a strain accuracy of 20 με. This accuracy can be further improved byusing more frequency samples and higher pump and probe power. Since it is based on BEDS technique, oursetup can potentially reach spatial resolution down to 1 cm if a faster pulse generator is implemented.We have successfully demonstrated a new concept for BEDS with centimeter resolution based on differentialphase-shift keying technique. Our results clearly reveal the 5-cm splice segment between two fibers and itsBrillouin frequency shift with a spatial resolution of 5 cm. Our new concept enables distributed measurementswith centimetre resolution by simply adding a negative pulse on the pump in a BOTDA system. If a 100ps-pulsegenerator and a fast photo-detector are equipped in the setup, the resolution down to 1 cm can be made. Thereforeit simplifies conventional BEDS systems by using a single modulator for pump rather than two modulators andimproves the optical loss of the pump. It also allows an easy adjustment of the delay between two pulses.

3. BOTDA USING A QUADRATURE PHASE SHIFT KEYING MODULATOR

DPSK-BEDS technique improves the BEDS-setup with respect to optical loss on the pump wave and the ad-justment of the delay between intensity and phase pulses. Let us now have a look at the probe wave arm in thesetup (Fig.(3)). Due to the intensity modulator, which is driven by an RF signal generator at about 11GHzcorresponding to the BFS of the fiber, we obtain two sidebands at the output of the modulator: a down- and aup-shifted component corresponding to the Stokes and anti-Stokes components for spontaneous Brillouin scat-tering. Both, Stokes and anti-Stokes component (=down-shifted and up-shifted component), are launched intothe fiber and interact with the pump pulse. By help of the fiber Bragg grating we chose either the Stokes oranti-Stokes component and filter out residual pump light. The Stokes wave leads to an upward signal at the oscil-loscope whilst the Anti-Stokes wave produces a downward signal. The reason is that the anti-Stokes (up-shifted)wave transfers its energy to the pump (Brillouin loss process), whereas the Stokes wave receive energy from thepump pulse (gain process). But the selection takes place after the interaction of both waves in the fiber becauseStokes or anti-Stokes component is filtered out by the fiber Bragg grating. Other configurations launch either



the Stokes or anti-Stokes component into the fiber. Using Brillouin loss, hence the Anti-Stokes component, ledto a 32-km long sensor with 5m spatial resolution10 and using Brillouin gain to a 22-km sensing fiber with 10mresolution.11 The authors claim that pump depletion is the reason for this discrepancy. A technique to suppressStokes or anti-Stokes component to avoid pump depletion was proposed using a single sideband modulator suchas dual-drive modulator.12 A dual-drive modulator is a Mach-Zehnder interferometer modulator with 2 RF portsfor each interference arm. By applying an RF signal to both ports with a phase shift of π/2 between the ports,a single side-band can be obtained at the modulator output. This technique was adapted to Brillouin opticalcorrelation-domain analysis (BOCDA)13 and BOTDA based on Brillouin dynamic grating.14

In our work, we have adapted the technique to our BOTDA system using an optical QPSK (Quadrature Phase-Shift Keying) modulator (I & Q modulator). Such a modulator is also exploited in BOCDA.15 QPSK modulationis used to transmit twice more bit sequence messages than conventional PSK. Optical I & Q modulators canperform such a modulation on optical phase. In fact, it is an intensity modulator with two Mach-Zehnderinterferometers as shown in Fig.(6).

3.1 Experimental setup

A 11GHz RF signal is injected into two RF ports of the modulator with a π/2-shifted phase. Three DCbias must be adjusted to obtain a single side-band at the modulator output. The two DC bias for the twointerferometers are used to suppress the pump wave and the DC bias at the Y junction at the interferometeroutputs is used to unbalance the side bands. By adjusting the last DC bias, either Stokes or anti-Stokes signal canbe chosen. The advantage of I & Q modulator compared to dual-drive modulator lies on the high performanceof carrier suppression as well as side-mode suppression. The modulator used in the setup is provided by PhotlineTechnologies and has a carrier suppression of 55 dB with a side-mode suppression of 40 dB at 1535nm. Therefore,a centre-wavelength of 1535nm is used in the setup shown in Fig.(6) for the best performance of the modulator.As a single side band is used as the probe, no additional filter is needed before the photo-detector. Althoughthe I & Q modulator has a high optical insertion loss, the distributed measurement signal has a higher contrastthan that using a conventional intensity modulator and a fiber Bragg grating. In fact, the probe does not sufferfrom the optical loss due to the fiber Bragg grating. On the other hand, problems with this technique are thecomplicated adjustment of three DC-bias and the continuous DC-bias drift of the modulator. Hence, it has tobe regulated for each measurement.

Amplified

DFB-Laser

1550 nm

Polarization controller

Fiber under test

Polarization

controller

EDFA

DC-supply

Pulse

GeneratorScopeTrigger

Intensity Modulator

Photo Diode

Optical

Circulator

PU

MP

PRO

BE

Scrambler

Optical

Isolator

Pump pulse

Fiber

Coupler 50:50

DC-supply

RF-Signal

Generator

Intensity Modulator

νB

DC-supply

DC-supply

50:50

phase-shifted π/2

Figure 6. BOTDA setup using a I & Q modulator.



195191 195201 195211 195221 195231

−80

−70

−60

−50

−40

−30

−20

−10

195191 195201 195211 195221 195231

−80

−70

−60

−50

−40

−30

−20

−10

Frequency (GHz) Frequency (GHz)

Op

tica

l o

utp

ut

po

wer

(d

Bm

)

Op

tica

l o

utp

ut

po

wer

(d

Bm

) (b)(a)

Pump

Stokes

Anti-Stokes Pump

Stokes

Anti-Stokes

Figure 7. Spectra of (a) the Stokes probe and (b) the anti-Stokes probe at the output of the 50 km SMF at 10.94 GHz.

3.1.1 Experimental results

In the following some results are shown. First, spectra of Stokes and anti-Stokes probes have been observed atthe output of the I & Q modulator. Fig.(7 a) shows a spectrum when the I & Q modulator is tuned to generatethe Stokes probe. The modulator can suppress the carrier wave by 55 dB and its single side-band suppressionratio is 40 dB at 10.94GHz from the pump frequency at 195211GHz (1535nm). These suppression ratios canensure a good quality of BOTDA measurements without a filter. A spectrum with the anti-Stokes probe is shownin Fig.(7 b). However the single side-band suppression ratio is 32 dB which is lower than that of the Stokes probewhich is due to the challenging adjustment of the I & Q modulator. Nevertheless this suppression ratio is enoughto perform a good quality of distributed measurement without filter. After the spectra observation, we haveperformed a distributed measurement of an SMF with 50km length using the I & Q modulator as shown inFig.(8). Although the loss of the SMF is only 0.2 dB/km, a significant loss occurs for long lengths like 50 km. Wehave measured the distributed Brillouin gain for the Stokes and anti-Stokes probes with a 3m-spatial resolutionand a 1MHz-frequency resolution. The distributed measurements using the Stokes and the anti-Stokes probesare shown in Fig.(8 a) and (8 b), respectively. By comparing the two figures, it is evident that the measurementwith the anti-Stokes probe shows a better quality than that with the Stokes probe. Fig.(8 b) clearly reveals the50 km length of the fiber despite of the fiber attenuation. The beginning of the fiber is very similar in bothcases. But the contrast at the end of the fiber is remarkably improved using the anti-Stokes probe. Fig.(9) showsBrillouin responses recorded by the oscilloscope. The blue trace in Fig.(9 a) represents the Brillouin responsewith the anti-Stokes (loss process), the red one is obtained by use of the Stokes component (gain process). Wecan observe in the figure the linear loss along the SMF mainly due to the fiber attenuation. In the beginningof the fiber, the gain amplitudes of the Stokes and the anti-Stokes probes are similar. However, at the end ofthe fiber, the Brillouin gain is merely noticeable for the Stokes wave whilst the gain for the anti-Stokes wavestill remains. The origin of this difference may be the pump depletion being higher in the Stokes probe thanthat in the anti-Stokes probe. Since the anti-Stokes involves Brillouin loss process, the energy of the anti-Stokes

(b)(a)Bri

llo

uin

fre

qu

ency

sh

ift

(GH

z) 10.98

10.9667

10.9533

10.94

10.9267

10.9133

10.9

10.98

10.9667

10.9533

10.94

10.9267

10.9133

10.9

Bri

llo

uin

fre

qu

ency

sh

ift

(GH

z)

0 10 20 30 40 50

Fiber length (km)

60 0 10 20 30 40 50

Fiber length (km)

60

Figure 8. Distributed measurements of a 50 km SMF using (a) the Stokes probe and (b) the anti-Stokes probe generatedby the I & Q modulator.



0 10 20 30 40 50

−0.05

−0.04

−0.03

−0.02

−0.01

0

0.01

0.02

0.03

0.04

0.05

Fiber length (km)

Inte

nsi

ty (

arb

.un

its)

(a)

0 10 20 30 40 50

−0.05

−0.04

−0.03

−0.02

−0.01

0

0.01

0.02

0.03

0.04

0.05

Fiber length (km)

Inte

nsi

ty (

arb

.un

its)

(b)

Figure 9. (a) Distributed measurements along the 50 km fiber using the Stokes probe (red trace) and anti-Stokes probe(blue trace), (b) distributed measurements along the 50-km fiber using the classical BOTDA setup. The double side-bandprobe is filtered out to select the Stokes signal (red trace) or the anti-Stokes signal (blue trace).

wave is transferred to the pump. Therefore it would be expected that the pump bears low depletion thanksto the anti-Stokes wave. Moreover, the Stokes probe is not present and the pump looses its energy mainly byfiber attenuation (10 dB for 50 km), but not much by Brillouin gain. For the Stokes case, the Stokes probe takesBrillouin gain continuously from the pump. This results in high pump depletion. As the anti-Stokes probe isnot present, the pump keeps loosing its energy by the fiber attenuation and depletion. The measurements showa high performance of our BOTDA setup using a I & Q modulator because commercial BOTDA sensors workwith a 30 km fiber length. We have made a comparison between BOTDA including the I & Q modulator andthe classical BOTDA with a filter. The 50 km distributed traces are shown in Fig.(9 b) for the Stokes (red trace)and the anti-Stokes probe (blue trace). The anti-Stokes traces in Figs.(9 a,b) reveals no significant difference interms of the signal contrast. Besides, the Stokes trace in Fig.(9 b) also exhibits gain at the end of the fiber. Asboth the Stokes and anti-Stokes probes are present in the classical BOTDA system, the pump looses its energyby the Stokes and the anti-Stokes transfers some of its energy to the pump. Therefore, the pump with the Stokesand anti-Stokes would suffer less from depletion than that with the Stokes only in the QPSK-BOTDA. Thecomparison also manifests that anti-Stokes can enhance the sensing range compared to Stokes. From the traceamplitudes, it also reveals that the signal amplitude is twice greater in the QPSK BOTDA setup than that inthe classical BOTDA setup. As such, a I & Q modulator provides an improvement of signal-to-noise ratio ondistributed measurements.

4. CONCLUSION

We successfully demonstrated a new distributed differential measurement technique using Brillouin echoes(BEDS) with π-phase-shift pulses. It is based on differential phase-shift keying (DPSK) using a single Mach-Zehnder modulator to generate a pump pulse and a short π-phase-shifted pulse with an easy and accurateadjustment of delay. With this simplified technique, we achieved centimeter spatial resolution when measuring asplice segment between two different fibers while reducing the optical loss for the pump pulse. The setup is sim-plified in comparison to a conventional BEDS system since two modulators are replaced by only one. We wouldlike to point out that this technique is especially interesting for the commercial use of BEDS because the setupis practically the same as for the BOTDA but with enhanced resolution just by using another electrical pulseformat which is applied to the intensity modulator. Moreover it allows for an easier adjustment of the intensityand phase pulse. A quadrature-phase shift keying (QPSK) modulator for the Brillouin probe has also beeninvestigated as a single-sideband modulator in the BOTDA system and compared to the previous dual-sidebandcase. The experiments show that using the anti-Stokes component can enhance the sensing range compared tousing the Stokes component. It also reveals a higher signal amplitude in the QPSK BOTDA setup than that inthe classical BOTDA setup, which improves the signal-to-noise ratio on the distributed measurement.



ACKNOWLEDGMENTS

We thank L. Thevenaz for helpful discussions. This work was funded by the Fond Europeen de DeveloppementRegional (FEDER) through the European Interreg IV A program, CD-FOM project.

REFERENCES

1. Horiguchi, T. and Tateda, M., “Optical-fiber-attenuation investigation using stimulated Brillouin scatteringbetween a pulse and a continuous wave,” Optics Letters 14(8), 408–410 (1989).

2. Nikles, M., Thevenaz, L., and Robert, P. A., “Simple distributed fiber sensor based on Brillouin gainspectrum analysis,” Optics Letters 21(10), 758–760 (1996).

3. Angulo-Vinuesa, X., Martin-Lopez, S., Nuno, J., Corredera, P., Ania-Castanon, J.-D., Thevenaz, L., andGonzalez-Herraez, M., “Hot spot detection over 100 km with 2 meter resolution in a Raman-assisted Brillouindistributed sensor,” in [Proc. SPIE ], 7753, 775309 (2011).

4. Soto, M. A., Bolognini, G., and Pasquale, F. D., “Long-range simplex-coded BOTDA sensor over 120kmdistance employing optical preamplification,” Optics Letters 36(2), 232–234 (2011).

5. Brown, A. W. and Colpitts, B. G., “Dark-pulse Brillouin optical time-domain sensor with 20-mm spatialresolution,” Journal of Lightwave Technology 25(1), 381–386 (2007).

6. Foaleng, S. M., Tur, M., Beugnot, J.-C., and Thevenaz, L., “High spatial and spectral resolution long-rangesensing using Brillouin echoes,” Journal of Lightwave Technology 28(20), 2993–3003 (2010).

7. Lee, M. W., Stiller, B., Hauden, J., Maillotte, H., Roch, C., Thevenaz, L., and Sylvestre, T., “Differ-ential phase-shift keying technique-based Brillouin echo-distributed sensing,” IEEE Photonics Technology

Letters 24(1), 79–81 (2012).

8. Charlet, G., “Progress in optical modulation formats for high-bit rate WDM transmissions,” IEEE Journal

of Quantum Electronics 12, 469–483 (2006).

9. Wen, Y. J., Nirmalathas, A., and Lee, D.-S., “RZ/CSRZ-DPSK and chirped NRZ signal generation usinga single-stage dual-electrode Mach-Zehnder modulator,” IEEE Photonics Technology Letters 16(11), 2466–2468 (2004).

10. Bao, X., Webb, D. J., and Jackson, D. A., “32-km distributed temperature sensor based on Brillouin lossin an optical fiber,” Optics Letters 18(18), 1561–1563 (1993).

11. Bao, X., Webb, D. J., and Jackson, D. A., “22-km distributed temperature sensor using Brillouin gain inan optical fiber,” Optics Letters 18(7), 552–554 (1993).

12. Loayssa, A., Hernandez, R., Benito, D., and Galech, S., “Characterization of stimulated Brillouin scatteringspectra by use of optical single-sideband modulation,” Optics Letters 29(6), 638–640 (2004).

13. Song, K.-Y. and Hotate, K., “Enlargement of measurement range in a Brillouin optical correlation do-main analysis system using double lock-in amplifiers and a single-sideband modulator,” IEEE Photonics

Technology Letters 18(3), 499–501 (2006).

14. Song, K. Y. and Yoon, H. J., “High-resolution Brillouin optical time domain analysis based on Brillouindynamic grating,” Optics Letters 35(1), 52–54 (2010).

15. Zou, L., Bao, X., Yang, S., Chen, L., and Ravet, F., “Effect of Brillouin slow light on distributed Brillouinfiber sensors,” Optics Letters 31(18), 2698–2700 (2006).



spie proceedings [spie spie photonics europe - brussels, belgium (monday 16 april 2012)] optical...

Documents